&EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA/600/R-92/147
June 1992
Proceedings of the 4th
Investigators' Meeting for
EPA's Biotechnology-
Biological Control Agent
Risk Assessment
Research Program
-------
-------
EPA/600/R-92/147
June 1992
PROCEEDINGS OF THE 4th
INVESTIGATORS' MEETING FOR ERA'S
BIOTECHNOLOGY-BIOLOGICAL CONTROL AGENT
RISK ASSESSMENT RESEARCH PROGRAM
APRIL 8-12, 1991
Prepared by
James E. Harvey
Technical Resources, Inc.
Gulf Breeze Environmental Research Laboratory
Gulf Breeze, Florida 32561
Contract Number 68-03-3479
Project Officer
Richard B. Coffin
Microbial Ecology and Biotechnology Branch
Gulf Breeze Environmental Research Laboratory
Gulf Breeze, Florida 32561
GULF BREEZE ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
GULF BREEZE, FLORIDA 32561
Printed on Recycled Paper
-------
r
DISCLAIMER
The information in this document has been funded wholly or in part by the U.S.
Environmental Protection Agency under various cooperative agreements. It has been
subject to the Agency's peer and administrative review, and it has been approved for
publication as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
-------
TABLE OF CONTENTS
Acknowlegment
Preface
Introduction
Invited Special Presentations
European Biotechnology Research Presentations
Canadian Biotechnology Efforts
U.S. Department of Agriculture Biotechnology Overview
U.S. Environmental Protection Agency Presentations
Research Presentations
Session I: Environmental Exposure
Detection/Enumeration
Dispersal/Transport
Survival/Colonization
Gene Transfer
Session II: Environmental and Human Health Effects
Ecological Processes
Higher Organisms
Human Health
Session III: Risk Control
Mitigation
Field Releases
Process Containment
Contributor Index
Affiliation Index
Subject Index
Page
iv
v
1
3
4
15
20
22
35
37
61
81
102
129
131
169
207
223
225
228
242
247
249
251
111
-------
. ACKNOWLEDGMENT
This publication is a result of the patience and cooperation of the researchers
who contributed to the 1991 All-Investigators Meeting. We wish to acknowledge the
word processing and editing assistance of Ms. Maureen Stubbs, and Ms. Valerie
Coseo, Computer Science Corporation. We also wish to acknowledge the assistance
of Ms. Nancy Padgett,.Technical Resources, Inc., both at the meeting and in the
completion of this document.
IV
-------
PREFACE
The Office of Research and Development (ORD) in the U.S. Environmental
Protection Agency is responsible for providing EPA program offices with scientific and
technical information on biotechnology products intended for environmental release.
This includes researchon genetically engineered microorganisms (GEMs) and biological
control agents (BCAs). To accomplish this, ORD created a Matrix-managed program
for Biotechnology/BCA Risk Assessment and designated the Gulf Breeze Environmental
Research Laboratory as the lead laboratory. To assure coordination and promote
multi-disciplinary research approaches, the Matrix manager holds periodic meetings
attended by laboratory and headquarters scientists and cooperative agreement and
contract researchers. This document is a compilation of extended abstracts of the
research presented at the 1991 All-Investigators meeting> the fourth meeting of this
type. The meeting was held at the Holiday Inn Crowne Plaza, Crystal City, Virginia,
April 8-12. Extended abstracts submitted from specially invited guests representing
biotechnology efforts in Canada and several European countries reflect the
international collaboration theme of the meeting. These invited abstracts are followed
by scientific presentations from EPA laboratory, cooperative agreement and contract
personnel separated into the three program element research areas: exposure, effects
and risk control.
-------
-------
SECTION 1
INTRODUCTION
During the last two decades,
sophisticated scientific advances in
molecular biology have increased our
ability to change or combine the
inherited characteristics of micro-
organisms. The power of recombinant
DNA techniques to make extremely
precise alterations in the genetic
character (genotype) and expression
(phenotype) of organisms has created
both the promise of enormously useful
products as well as concern over
potential environmental release.
Among the potential industrial
products and applications of this
technology are microorganisms
designed to degrade toxic pollutants,
leach minerals, enhance oil recovery,
produce industrial chemicals, and act as
pesticides. These activities fall within
EPA's regulatory purview under the
Toxic Substances Control Act (TSCA)
and, for pesticidal applications, the
Federal Insecticide, Fungicide and
Rodenticide Act (FIFRA) as established
by the Coordinated Framework for
Regulation of Biotechnology (EPA,
1986). These regulations were
intended to protect the environment
from perturbation by industrial products
regardless of the scientific process of
their manufacture, and are therefore
applicable to the products of
biotechnology.
EPA regulators rely on the results of
research and development programs to
provide the scientific and technical
knowledge required for successful
regulatory policies. Risk assessment
methods for predicting the risk posed
by some chemical pesticides arid toxic
industrial chemicals, including fate and
effects data, are reasonably well
developed and have been incorporated
into regulatory policies.
The more dynamic properties
presented by microorganisms, including
replication, mutation and gene transfer,
coupled with the complexity of
molecular biological techniques create a
high degree of uncertainty when risk
assessment methods developed for
chemicals are applied to genetically
engineered microorganisms (GEMs) or
biological control agents (BCAs). BCAs
include microbial pest control agents
i.e. Bacillus thuringiensis derivatives,
and biochemical pest control agents,
such as insect growth regulators.
Concern over the potential risks
introduced by the environmental release
of biotechnological products and the
need to reduce uncertainty in assessing
risk led the Agency to establish the Bio-
technology/Biological Control Agent
Risk Assessment Research Program in
1985.
Participating ORD research labora-
tories and their expertise areas are
listed below:
Gulf Breeze, FL., Environmental Research
Laboratory (GBERL), Lead Laboratory, Fate
,and Effects of GEMs and BCAs in Marine
and Estuarine Ecosystems.
Corvallis, OR., Environmental Research
Laboratory (CERL), Fate, Effects, and
Transport of GEMs and BCAs in Terrestrial
Ecosystems.
Duluth, MN., Environmental Research
Laboratory (DERL), Fate and Effects of
MPCAs in Freshwater Environments.
-------
RTF, NC., Health Effects Research
Laboratory (HERD, Human Health Effects
from GEMs and MPCAs.
Cincinnati, OH., Risk Reduction Engineering
Laboratory (RREL), Control of Risk through
Containment and Decontamination.
Las Vegas, NV., Environmental Monitoring
Systems Laboratory (EMSU, Aerosol Dis-
persal of Biologicals, Model Development.
Also, the Office of Environmental
Processes and Effects Research in EPA
Headquarters at Washington, DC.,
serves as a liaison between ORD and
Program Offices.
ORD scientists have developed a
complementary extramural research
program fostering close interaction with
academic scientists in the fields of
genetics, biochemistry, ecology and
molecular biology. This cooperative
agreement process promotes col-
laboration between diverse disciplines
to solve research problems often
resulting in scientific advancement.
Over twenty-five universities in the U.S.
and abroad have contributed to EPA's
program and to scientific information
exchange. Scientists at EPA labora-
tories and under EPA cooperative
agreements have further facilitated
scientific exchange by participating in
national and international scientific
meetings, publishing journal articles,
and participating in research
coordination workshops.
The knowledge base for character-
izing the risk posed by GEMs/BCAs
continues to improve and has potential
to estimate and to reduce those risks.
The use of information developed for
assessing risk to create predictive
models and to develop control
strategies represents a shift in our
capabilities towards risk reduction.
Research designed to improve our
understanding of problems and to
identify promising solutions to those
problems are important and recognized
tools for reducing environmental risk.
EPA funded research provides an
important basis for continued
development of methods and tech-
niques needed for reliably evaluating
the behavior of microorganisms and
microbial communities.
The ecological processes mediated
by microorganisms are crucial to the
operation and health of ecosystems. If
these processes are affected by en-
vironmental stress resulting from the
introduction of novel microorganisms or
their toxic by-products, then the po-
tential for affecting the rest of the
ecosystem is significant. Continued
development of methods for evaluating
and improving the quality of eco-
systems is necessary if strategies for
assessing and reducing deleterious
changes in ecosystems health are to be
effective, timely, and relevant. The
research program's mission is to dev-
elop these methods and techniques and
to provide guidance in their appropriate
application for characterizing and
reducing risk to the environment.
Extended scientific abstracts
describing EPA funded research com-
prise most of this report, the fourth in a
continuing proceedings series. The
abstracts are grouped by broad
research areas of environmental
exposure, environmental and human
health effects, and control of risk
through mitigation and containment
strategies. These abstracts demon-
strate the Scientific direction and
accomplishments of EPA's Biotech-
nology/Biological Control Agent Risk
Assessment Program.
-------
SECTION 2
INVITED SPECIAL PRESENTATIONS
Information dissemination and technical information transfer are important facets
of EPA's Biotechnology Risk Assessment Program. To improve international under-
standing and foster collaborative research efforts, representatives from several
European countries, Canada, USDA, and EPA were invited to present an overview of
their biotechnology efforts. Extended abstracts of those presentations comprise this
section.
M. Hoefle, National Research Center for Biotechnology,
Braunschweig, Germany
P. Kearns, European Organization for Economic Cooperation and Development,
Paris, France ,
M. Reuss, Ministry of the Environment, Charlottenlund, Denmark
T. Mclntyre, Environment Canada, Hull, Quebec, Canada
D. MacKenzie, National Biological Impact Assessment Program, U.S. Department of
Agriculture, Washington, D.C.
E. Milewski, Office of Pesticides and Toxic Substances, U.S. Environmental
Protection Agency, Washington, D.C.
P. Sayre, Office of Toxic Substances, U.S. Environmental Protection Agency,
Washington, D.C. •
W. Schneider, Office of Pesticides Programs, U.S. Environmental Protection
Agency, Washington, D.C.
-------
OVERVIEW OF GESELLSCHAFT FUR BIOTECHNOLGISCHE FORSCHUNG
mbH/NATIONAL RESEARCH CENTER FOR BIOTECHNOLOGY (GBF)
Manfred Hoefle
Gesellschaft fur Biotechnolgische Forschung
Braunschweig, West Germany
The GBF (Gesellschaft fur
Biotechnolgische Forschung mbH/
Research Institute for Biotechnology
had its origins in the "Institute for
Molecular Biological Research (GMBS)"
started and funded by the Volkswagen
Foundation. Since the inception of
GBF, it has been financed by the
Federal Government, represented by the
Federal Ministry of Research and
Technology (BMFT) on the one hand,
and the State of Lower Saxony on the
other, to a ratio of 90:10. The articles
of the company incorporated in 1976
reinforce the structure of GBF as a
research-oriented national research
centre. The statutes governing the GBF
were amended in 1984. The GBF is a
member of the AGF (Association of
National Research Centres), and
employed a total of 590 staff at the end
of 1989.
The objective of the company is the
operation of a multidisciplinary research
centre in the area of biotechnology with
the aim of contributing to the solution
of public duties and tasks by means of
comprehensive biotechnological
research and development work. The
research and development activities at
the GBF are basically located within the
framework of the Federal Government's
"Applied Biology and Biotechnology"
programme. The GBF pursues
exclusively peaceful purposes,
publishing the results of scientific work.
The structure and operation of the
bio-pilot plant based on interdisciplinary
cooperation, facilitates the work on
research projects within the GBF as
well as in the form of cooperation in
working groups involving industry,
universities and other institutions.
Accordingly the GBF is devoted above
all to the following tasks:
* applications-oriented basic research
in the areas of microbiology, cell
biology, molecular biology, genetics,
biochemistry, natural product
chemistry, enzyme technology and
bioengineering,
* the development of new
pharmacological and technically
significant natural products and
research on their mode of action,
* developmentof new biotechnological
processes as a contribution to
maintaining an assured supply of
basic products for the chemicals,
pharmaceutical and foodstuff
sectors, as well as processes for the
reduction of environmental pollution,
* the transfer from the biotechno-
logical laboratory process to the
semi-industrial scale as a pre-
condition for the development of
tried industrial processes.
-------
* the support of external research
groups in the sectors of biology,
chemistry and medicine by the
provision of natural products from
enzymes and biomasses, which are
not available commercially,
* use of the scientific-technical
equipment and installations at the
GBF as a service to projects
performed by external research
groups and industry,
* participation in joint projects, in
particular within the framework of
the biotechnology programme of the
Federal Ministry of Research and
Technology (BMFT); and in addition,
participation in the main areas of
research of bioengineering together
with the universities of
Braunschweig, Hannover and
Gottingen,
* enhanced interdisciplinary training of
scientists, engineers and technicians
in the framework of further training
courses on a national and inter-
national basis, and
* strengthening of the national and
European biotechnology infra-
structure by the provision and
performance of scientific services
such as informatics (data bases), or
the analysis of biological molecules.
THE MAIN AREAS OF RESEARCH
AND PROJECTS OF THE GBF
«
In 1989 the research activities at the
GBF were undertaken under the
framework of the following four main
strategic research areas:
1. Biosynthesis and biocatalytics
2. Biomolecules and molecular design
3. Environmental biotechnology
4. Biochemical engineering
These involve the screening for new
materials as well as method-oriented
developments. A balance between
target-oriented basic research and
industrial development work is sought.
The projects were essentially
performed in the framework of
interdisciplinary cooperation by
numerous departments/researchgroups
at the GBF. The central facilities of the
bio pilot plant service unit and the
central computer group provided
support to these activities. Project
participation is clearly presented in the
matrix below. The scientific
organization of the GBF is detailed in
the organigram at the end of this
abstract.
-------
Scientific Organization of GBF
• a
Priority g
1
pi Essential Work B
M
| [ Secondary Activities |
1°
Main Areas 3. a
of Research Projects Q c
Btocatatysls UtfOxMmtm .Miero»s«oi«m«
BioHantbmialion of Nauial
Compound*, e*peciaiJyTerpenoide
Enzymas'o Reaction*
&3cen*or»
Enzyme* from Kgher Cell*
BlomolecUleS ProtetiOeeign *
Molecular
Design
ImmunoModuUtor*
Regulation* of Eukaryof c Gene*
Cel Cyde Dependent
Gen* Exprevaicn
Bone Growti Factor*
Environmental eidogioai D«bxiiiaifc<)
BIOlBChnOlODV olEn»****?»
Qptimtzalion and Modeling
ol FermentA&'cn Proce**«*
Dovm Stream Prcce**e*
Compukr Aoplicalion*
and Conkol Technique*
Material* lor
BioMchndogy
1
1
S
D
•
•
•
Mikrobal Secondary 1
Metaboltee 1
•
Miaobial 1
Translormationft 1
•
G«,«-c. Cell Biology 1
Genetic* |
•
n
n
M
•
n
TlS
a
a
•
H
•
,
•
n
if
•
Immuna Biology I
•
i Enzyme Technology Enzymo Technology I
' Nituril Product |
•
•
S
m
m
Natural Product ' 1
Chemitty 1
m
TJ
•
IMotocular Structure 1
flesearch I
a
H
•
•
! Biochemical Bk>ChemlC*l |
Engineering Engineering •
. •
E3
•1
ii
H
n
H
Cel Culture Technitjje 1
•
H
a
en
s
O
n
D
•
Material Science * I
a
o
a
n
•
1 Fermentation 1
Service Unit j
•
n
a
a
M
£
I
'" -?
•
•
•
•
•
5
The main research areas and projects of the QBF1989/90 and the participation of the departments and research groups In the projects.
The first projects of each research area Is the central project In this field, (oo means cooperation with two or more Industrial partners.)
6
-------
THE WORK OF OECD RELEVANT TO
THE ENVIRONMENTAL IMPLICATIONS OF BIOTECHNOLOGY
Peter E.W. Kearns
Organization for Economic Cooperation and Development,
Environment Directorate, Paris, 75017, France
THE OECD IN BRIEF
The Organization for Economic Co-
operation and Development (OECD) is
an inter-governmental organization
which is part of the system of Western
international institutions developed
after World War II.
OECD replaced the Organization for
European Economic Cooperation (OEEC)
which was founded in 1948 to
administer the Marshall Plan of the
United States.
There are 24-member countries
which are the'free-market democracies
of North America (USA & Canada),
Western Europe (e.g. the EC & EFTA
nations, Switzerland, etc.) and the
Pacific (Australia, New Zealand &
Japan). The Commission of the
European Communities also participates
in OECD work, as does Yugoslavia as a
"special status" country.
It is worth nothing that these account
for only 16% of the world's population
but produce two-thirds of its goods and
services, three-fifths of its exports, and
generate four-fifths of economic aid to
developing countries. ,
The basic aims of the OECD, as
described its founding Convention, are:
1) to achieve the highest sustainable
economic growth and employment,
2) to promote economic and social
welfare throughout the OECD area by
co-ordinating the policies of its Member
countries, and
3) to stimulate and harmonize its
Members' efforts in favor of developing
countries.
For those of you with an interest in
history, you might like to know that
following the signing of the Convention
on 30 September 1960, OECD .was
established on 30 September, 1961.
This year is the 30th anniversary.
HOW DOES THE OECD WORK?
Each of the Member countries
maintains a permanent delegation of
OECD which is headed by an ambas-
sador. The ambassador represents his
country on the Council, the supreme
body of the OECD which normally
meets about once a week. The
Chairman of the Council is the
Secretary-General of the OECD.
Once a year, the Council meets at
Ministerial level under the Chairmanship
of one of the national Ministers.
The Council, which operates on the
principle of consensus, produces
Decisions, which are legally binding on
Member countries. The Council also
makes Recommendations which are
more an expression of political will than
something binding.
The Council also approves the work
of numerous specialized committees,
-------
expert groups and working parties
which carry out the major part of
OECD's work.
To give you a flavour of the scope of
the work, a few of the main
Committees are:
Economic Policy,
Trade,
Financial Markets,
Energy Policy,
Scientific and Technological Policy,
Environment,
Agriculture,
and Management Committee of the
Special Programme on the Control of
Chemicals.
The common approach of public
policy of the OECD countries means
that discussions on a given issue
usually reach consensu^. It is then
possible to establish "rules of the
game" or codes of behavior, for a given
topic, to which the participants can
bind themselves. The OECD has given
rise to several such agreements, for
example, an agreement on the
movement of hazardous wastes over
borders.
OECD WORK ON BIOTECHNOLOGY
OECD work on Biotechnology really
began in 1982 with the recognition, by
the OECD Committee for Science and
Technology Policy, that Biotechnology
was an major emerging field from
which various policy issues would arise
with wide scientific and economic
impacts.
Consequently, a group of national
experts was convened to consider
future OECD work in this area. Their
report on "Biotechnology: International
Trends and Perspectives (OECD, 1982)"
is still, in many ways, a forward-looking
study of many of the issues associated
with use of biotechnology.
The authors recognized the need for
a unique multidisciplinary approach
integrating the development and
implementation of policies and technical
instruments in the areas of science and
technology research, industry, agri-
culture, education, environment,, and
health and worker safety.
As a follow-up to this activity, a
study on "Safety ,and Regulations in
Biotechnology" was undertaken. That
study led to a Recommendation of the
OECD Council concerning Safety
Considerations for Applications of
Recombinant DNA Organisms in
Industry, Agriculture and ,the
Environment which called for further
research to improve the prediction,
evaluation and monitoring the outcome
of application of, recombinant, DNA
organisms. This work also led to the
publication of "Recornbinant DNA
Safety Considerations (OECD, 1986)".
A new programme of work, began in
1988, is being carried out by the OECD
Directorate for Science, Technology and
Industry (DSTI) in co-oper.ation with the
Environment Directorate, under the
supervision of a group of National
Experts on Safety in Biotechnology.
The Group of National Experts
decided that part of the programme
should be the development of general ,
principles which would identify a
generic approach and guidelines for the
design of small-scale field research with
genetically modified plants and micro-
organisms. As this work evolved, the,
Group of National Experts recognized ,
8
-------
the need for developing criteria and test
methods for the environmental monitpr-
ing of genetically modified organisms in
"Good Developmental Practices for
Small Scale Field Research with
Genetically Modified Plants and Micro-
organisms: A Discussion Document
(OECD, 1990)".
The OECD Secretariat was called on
to analyze monitoring methods and data
requirements with a view of identifying
the most promising approaches for
conducting and evaluating the
deliberate release of genetically
engineered organism to the
environment.
As part of subsequent efforts to
prepare a publication, an "International
Survey on Biotechnology Use and
Regulations (OECD/ Environmental
monograph no. 991, 1990)", inform-
ation on monitoring programmes for the
environmental release of GMOs in the
Member countries was requested.
While this information was not included
in detail in the survey report, some
information was annexed.
Additional information is summarized
in BIOTRACK which is OECD's com-
puterized Pointer System on the Use of
Genetically Modified Organisms.
BIOTRACK contains summarized details
of over 260 releases of genetically
modified organisms in Member
countries.
As a further step on the OECD work
on monitoring releases of genetically
modified organisms , a workshop was
held in Copenhagen the first week of
December, 1990. This workshop was
an opportunity for experts from a
variety of member countries to
exchange views and experiences. The
report of the workshop is currently in
preparation and will include two main
parts. The first, "Principal Findings of
the Workshop", is comprised of three
sections: General Approaches to
Monitoring; Plants - Monitoring
Approaches and Methods; and Micro-
organisms - Monitoring Approaches and
Methods. The second part is
suggestions for future work and
priorities.
I have stressed our interest in
monitoring because that is the aspect
of the work that I feel is most relevant
to a meeting like this.
However, there are other relevant
biotechnology activities in the OECD
programme; for example, there is work
underway to elaborate Good Industrial
Large Scale Practices, and this work is
likely to result in a publication before
much longer. There is also work
underway on Food Safety and
Biotechnology.
-------
NATIONAL ENVIRONMENTAL RESEARCH INSTITUTE
DIVISION OF MARINE ECOLOGY AND MICROBIOLOGY
M. Reuss and T. Leser
Ministry of the Environment
National Environmental Research Institute
Divison of Marine Ecology and Microbiology
Charlottenlund, DK-2920, Denmark
PERSPECTIVES AND EXPECTATIONS
OF BIOTECHNOLOGY
Since genetic engineering started
accelerating in the mid 1970s there
have been great expectations of this
new technology. Both industry and
agriculture interests hoped that
biotechnology would soon be able to
solve a number of problems. However,
we now must acknowledge that the
speed of the development has not been
quite what we expected.
In spite of this, several genetically
engineered microorganisms have been
developed and are now exploited com-
mercially. So far, it has mostly been
industry that has benefitted from
biotechnology and the new organisms
have been well-known strains of
microorganisms, which have been used
in fermentation industry for many
years. The microbes have been
engineered to produce new or modified
products in fermentation. Several
Danish industries have been able to
implement the new technology.
LEGISLATION
In June, 1986, the Danish
parliament, passed the worlds first
comprehensive act regulating genetic
engineering. The act introduced tight
regulation from the start, partly on the
basis of previous experience with the
regulation of pollutants and partly
because this was a new area involving
possible risks. The act will be revised
as we acquire more insight into the
problems resulting from discharge and
release of genetically engineered
microorganisms.
In connection with the administration
of the act, the National Environmental
Research Institute's (NERI) job is to
provide the necessary professional
expertise needed by the National
Environmental Protection Agency in
order to evaluate the risks and decide
on the rules in this area.
NERI has participated in a group
working with the realization of the
OECD concept GILSP (Good Industrial
Large-Scale Practice). The idea of
GILSP is that certain organisms may be
categorized as low risk organisms and
used in ordinary production systems
where the discharge of organisms is
limited as much as possible. In 1990,
two directives were introduced on
genetically modified organisms which
are now being incorporated in Danish
legislation. So far, only contained use
of genetically engineered micro-
organisms has been approved in
Denmark. Approval has been given for
the use of the organisms in ferment-
ation plants' production and restrictions
have been imposed on the number of
10
-------
living cells in the effluents.
into nature.
The organisms used so far have been
strains that do not survive in the
environment. They have a very
reduced frequency of gene transfer
compared to the parental strains and
they do not produce substances that
constitute any environmental risks.
In the future however, an increasing
number of genetically engineered
microorganisms will be deliberately
released into the environment. These
organisms will be designed to carry out
various tasks in the environment, e.g.
biodegradation of recalcitrant organic
pollutants or biological pest control.
These organisms will have a much
larger potential of survival than the
organisms used for industrial
production. They will be able to persist
in the environment at least while they
perform the function for which they
were designed. Because of their new
traits some will interfere with ecological
processes such as mineralization and
nutrient cycling. Viewed in an
ecological perspective this type of
genetically engineered microorganisms
poses a risk of unintended ecological
impact on the environment. We have
no experience in assessing this kind of
risk as the release of genetically
modified organisms is a completely new
activity with no previous examples that
can help in identifying the potential
dangers.
The research at NERI has been
focused on the ecology of microbes and
their influence on ecological processes.
By experimenting in contained
ecological models (microcosms)
researchers have tried to elucidate what
happens to microorganisms released
DETECTION OF GENETICALLY
ENGINEERED MICROORGANISMS
To be able to study the fate of
genetically modified microorganisms in
the environment or in experiments in
the laboratory sensitive detection
techniques are needed.
The classical microbiological
methods are all based on identification
and enumeration of the organisms on
selective media. By composing media
with a very special content of essential
nutrients it is possible to select the type
of organisms growing on the media.
Very specialized media have been
made for detection of specific
microorganisms as well as general
media for the enumeration of total
number of bacteria in environmental
samples.
Genetically engineered micro-
organisms have very often been
endowed with antibiotic resistance that
can be used for detection. These
organisms can be grown in the
presence of concentrations of
antibiotics that inhibit the indigenous
microflora and when samples are plated
on media containing one or more
antibiotics only resistant organisms, i.e.
the genetically engineered strains, form
colonies.
Other traits such as special
temperature requirements or the ability
to degrade toxic organic compounds
may be useful for detection.
Plating on selective media is still the
most frequently used method of
detection and the one that new
techniques are compared with.
11
-------
However, selective plating has
several drawbacks. The plating and the
colony counting are very labor-
intensive. Many microorganisms grow
very slowly on the plates, and often the
results are only obtained after several
days.
The most serious drawback of
selective plating is, however, that not
all bacteria from environmental samples
are able to form colonies. Bacteria
which are stressed, e.g. when released
into waste water or into the environ-
ment, are able to change into a dormant
state. Bacteria in this state (viable-but-
nonculturable) are not dead as they
may be brought back to a culturable
form under certain circumstances.
They are however, unable to grow on
the selective media they would
otherwise grow on. The phenomenon
has first been described for pathogenic
bacterial strains that are stressed and
enter the viable-but-nonculturable state
when they are released into the
environment, but it seems now to be a
very general bacterial characteristic.
It is possible to reinduce the
culturable state for pathogenics again
through injection into a host-organism,
but the mechanisms governing this are
unknown.
When organisms are in a
nonculturable state they may be present
in the environment but undetectable
when using selective plating. Because
of the general problems concerning the
selective plating method and because
the development of biotechnology has
made specific detection of genetically
engineered bacterial strains necessary,
research has been directed towards the
development of alternative detection
methods. What these methods all have
in common is that they are direct,
meaning they do not imply cultivation
of the organism in the laboratory.
Direct detection methods are based on
immunological and molecular biological
principles. At NERI selective plating as
well as direct methods have been
implemented in the detection of
microbes.
Previously, an ELISA (Enzyme -
Linked Immunosorbant Assay) has been
implemented for the detection of bakers
yeast producing insulin in industrial
waste water. This method has been
used in monitoring the survival of yeast
in the field.
For the validation of terrestrial
microcosms (see below) a method for
detection of Enterobacter cloacae by
immunofluorescence has been
developed. In this analysis, specific
polyclonal antibodies react with cells of
E. cloacae. After treatment with fluoro-
chromconjugated secondary antibodies
cells can be counted with the aid of an
epifluorescence microscope only.
Detection of microorganisms based
on molecular biological methods has
focused on DNA hybridization. By the
use of a specific DNA probe for a 2,4-D
degrading Alcaligenes eutrophus strain,
it has been possible to study the
survival of the bacteria in soil
microcosm and water samples.
It is a necessary step for several
molecular biological approaches to the
study of microbial communities and for
the detection of specific micro-
organisms in nature to be able to
extract purified DNA from environ-
mental samples.
Extraction and purification of DNA
from environmental samples is not an
easy task. Lysis has to be effective on
12
-------
all bacteria present in the sample and
once extracted, the DNA is con-
taminated by various organic
substances. At NERI a great deal of
work has been put into, the imple-
mentation and the improvement of
methods for recovering high quality
DNA from complex environmental
samples.
MICROCOSMS
It is internationally agreed that
microcosms are an essential
experimental tool for research in
microbial ecology and for the risk
assessment of genetically engineered
microorganisms.
At the microbiological department of
NERI aquatic and terrestrial microcosm
(ecological models) have been
constructed. The models have been
designed, to contain parts of natural
ecosystems in such a way that
physical, chemical and biological
conditions are kept as natural as
possible. .
. The philosophy behind the structure
of the models is that because micro-
organisms in nature are controlled by a
complex set of determinants, complex
ecosystem models are essential to
studies o-f the ecology of genetically
engineered microorganisms and their
possible ecological effects.
Only little is known of the factors
controlling microbes. Recreating the
natural ecosystems in the laboratory as
far as is possible is the best way to
ensure that all decisive factors are
included in the experiment.
But experiments in microcosms
cannot stand alone. Experiments
designed to uncover causalities on a
more detailed level are needed.
The context for the design of the
microcosms has been to create an as
natural as possible model system in the
laboratory and it has therefore, been
important to investigate how well the
models simulate nature. Extensive
validations of the models have been
carried out.
The aquatic models have been used
to simulate the eutrophic Lake
Bagsvaerd and validation has been
made by comparing important
ecological variables measured in the
models to similar measurements in the
field. Microcosms have been run for up
to 4 months and an array of variables
have been measured: total bacterial
numbers determined by microscope and
by colony forming units on plates; the
numbers of fluorescent Pseudomonads;
changes of bacterial community
structure; bacterial production; primary
production; algal biomass and the
concentrations of inorganic nutrients.
The overall conclusion is that with
the actual microcosm design it is
possible to simulate important features
of aquatic ecosystems.
At the bacterial level there is a very
high concordance between measure-
ments made in the models and in the
lake. This is important as the models
are intended to be used for studies of
bacterial ecology.
The microcosm validations also
reveal that the model is able to function
satisfactorily through long experimental
periods (months) and that replication
between uniformly treated microcosms
is very high. Therefore, it will be
possible to study microbial populations
throughout many generations "in the
models and to discover long term
13
-------
ecological effects of the organisms'
presence and activities in the
environment.
The terrestrial microcosms have been
validated by conducting the s"ame
experiment in the field and in the
microcosms. A non-genetically engine-
ered strain of Enterobacter cloacae was
sprayed on barley and bean plants
growing in the field and in microcosms.
Survival of the bacteria was followed
for 7 weeks on the leaves of the plants
and in the soil. E. cloacae survived
equally well in the two systems and at
the end of the experiments the numbers
of organisms present in the field and in
the models was the same.
In the field experiment the numbers
of E. c/oacaethat could be detected by
selective plating decreased dramatically
immediately after spraying, while the
numbers detected by immuno-
fluorescence were unchanged. The
field release took place during a spell of
drought and the discrepancy between
the two detection methods may be
explained by stress suffered by the
bacteria that were sprayed into the dry
environment. This stress may have
induced a non-culturable state. After a
couple of weeks the number of cultur-
able cells increased and at the end of
the experiment were indistinguishable
from the numbers determined by the
direct method. In the microcosms none
of this was seen, probably because the
models did not simulate the dry period
and the rain. The conclusion about the
behavior of the terrestrial microcosms is
that they are able to simulate ecological
conditions that are of importance to
bacterial survival and activity, but that
extreme climatic conditions cannot-be
simulated.
CONCLUSION
.Through the extensive validation of
both types of microcosms we have
obtained important knowledge of some
of the factors determining the extent to
which ecological models resemble
natural ecosystems.
Experiments with genetically
engineered microorganisms which will
be carried out in the microcosms will
benefit from the thorough evaluations
of ecological parameters, by making
extrapolations to nature possible from
results obtained, in the models.
Extrapolations will be meaningful
because the model behavior and the
model limitations are well known.
In the near future the models will be
used in studies of the fate and the
possible ecological effects of genetically
engineered microorganisms in the
environment. This work will be done in
collaboration with international
scientific groups which work on the
same topics.
The molecular biological approaches
to detection will be an important and
integrated part of the future work and
one of the techniques that are going to
be implemented is the polymerase chain
reaction (PCR). By PCR it is possible to
improve the detection limit for DNA
hybridization to 1-10 cells per gram of
soil or millilitre of water.
Extensive research will be put into
the study of possible ecological effects
of the introduction of genetically
engineered microorganisms into the
environment.
14
-------
DEVELOPMENT OF REGULATORY AND ENVIRONMENTAL OVERSIGHT
FOR BIOTECHNOLOGY IN CANADA UNDER
THE CANADIAN ENVIRONMENTAL PROTECTION ACT
(PRESENTATION TRANSCRIPT)
T.C. Mclntyre
Biotechnology Section, Commercial Chemicals Branch
Environment Canada, Hull Quebec K1A OH3
INTRODUCTION
Environment Canada is currently in the
process of finalizing draft regulations for
biotechnology products under the
Canadian Environmental Protection Act.
This Act, passed in 1988, provides for a
comprehensive regulatory scheme to
control toxic substances at each stage in
their life cycles: from development and
manu-facture; through transport,
distribution,, use and storage; to their
ultimate safe disposal as wastes. The
purpose of my presentation this morning
is threefold:
i) Provide an update on the extent of
organizations currently carrying out
biotechnology R&D activities in
Canada, (Figure 1);
ii) Share the proposed methodological
approach for the regulation of
biotechnology products by
Environment Canada under the
Canadian Environmental Protection
Act; and finally
iii) Highlight ongoing in- house research
and development activities in support
of the regulatory oversight initiatives
of Environment Canada.
In discussing the environmental
components of the CEPA biotechnology
regulations this morning, I will briefly
examine three forces that greatly
influenced the structure and methodo-
logical approach proposed in our draft
regulations.
1) CEPA AND THE LEGISLATIVE
FRAMEWORK FOR ENVIRONMENTAL
PROTECTION
Our proposed biotechnology reg-
ulations are based on a number of
elements underlying the legislative
framework for CEPA that features:
CEPA AND THE LEGISLATIVE FRAMEWORK
FOR
ENVIRONMENTAL PROTECTION
* Strong prevention focus
* Life cycle approach to management of
toxic substances (cradle to grave)
* Shift in onus of proof to manufacturer/
importer
* Integration of environmental and health
reporting
* An ecosystem approach to environmental
protection
15
-------
N = 420
FIGURES IN PARENTHESIS ARE FOR 1988
Source ISTC
CANADA
Figure 1. Geographical distribution of organizations carrying out biotechnology
R&D activities in Canada 1990.
16
-------
CEPA AND THE LEGISLATIVE
FRAMEWORK FOR
ENVIRONMENTAL PROTECTION (CONT.)
* National Inspection plan and enforcement
policy based in criminal law
* Extensive public role in the administration
of the act
* Innovative mechanisms for federal-
provincial partnerships
* Enforceable standards for federal lands,
works and initiatives
* Environment economy linkages and
environmental quality objectives
2) Our draft regulations acknowledge
criteria developed at the national level as
essential components of a regulatory
framework for biotechnology. Some of
these include:
CRITERIA CONSIDERED AS ESSENTIAL
COMPONENTS OF A REGULATORY
FRAMEWORK FOR BIOTECHNOLOGY
* Engenders public confidence
* Makes economic sense
* Transparent to all affected stakeholders
* Facilitates industrial planning for
development
* Compatible with international approaches
* Flexible and accommodating of new
approaches
* Clarifies jurisdictional responsibilities
* Draws upon independent scientific advice
3) Our regulations acknowledge certain
working principles that have evolved in
the international community for the
regulation of biotechnology products that
include:
FEDERAL GOVERNMENT WORKING
PRINCIPLES FOR THE REGULATION OF
BIOTECHNOLOGY PRODUCTS IN CANADA
* Use of a standard definition of
biotechnology
* Buijd upon existing legislation where ever
possible
* Regulate product as opposed to process
* Assess biotechnology products on a case
by case basis
* Build upon internationally developed
guidelines and harmonization whenever
possible at both federal and international
levels
PART II Methodological Approach
General provisions of the
biotechnology regulations under the
Canadian Environmental Protection Act
include:
* Products assessed under the pro-
posed Regulations include micro-
organisms to be used in a variety of
applications such as, but not limited to,
bioremediation, mineral leaching,
degradation of chemicals, mining,
waste treatment, waste disposal,
chemical production, lignin degrad-
ation, microbial enhanced oil recovery,
biosensors, and energy production.
17
-------
* An interactive and staged approach
to notification and provision of
information is used, based on the
stages of development of the
product towards commercial
exploitation. The corresponding
information requirements and
assessment periods build on those
of previous stages of notification.
* Notification is required prior to
manufacture or importation for
contained uses beyond the scale of
laboratory research.
* Information will be assessed within
prescribed assessment periods.
* ' CEPA contains explicit provisions
for protecting confidential business
information.
* The waiver provisions of CEPA
provide flexibility for dealing on a
case-by-case basis with specific
products where the progression
through the stages is either not
appropriate or necessary, or for
waiving information requirements
that may not be applicable for the
product.
* A mechanism is proposed for listing
specific uses of microorganisms for
which notification would not be
required (Schedule XVI).
Categories of Notification
Five stages of development are
defined for notification.
For contained use, notification is
required:
* prior to research and development
beyond laboratory scale (Schedule XI)
* prior to commercial manufacture
(Schedule XII)
For environmental introduction,
notification is required:
* prior to small-scale field trials
(Schedule XIII)
* prior to large scale field trials
(Schedule XIV).
* prior to commercial production
(Schedule XV)
Information Requirements
Each of the five information
schedules specify the application
information requirements. Categories
of-information to be submitted include:
* identification and characterization of
the microorganism(s)
* data on environmental., fate and
effects on nontarget substances in
the environment
* human health safety testing data
* description of intended uses,
manufacturing methods and quality
control and quality assurance
procedures
* specifics on field trials including
location and procedures to be used
during the trial, where applicable
1.8
-------
PART III Biotechnology Research
In development of the CEPA
biotechnology regulations, we
recognized that a properly drafted and
administered regulatory framework can
be a catalyst and not an impediment to
technology development.
We recognize the need for flexibility
in addressing the needs of the dynamic
and expanding biotechnology industry
in Canada. The flexibility we are
building into the regulations is linked to
and dependent upon information
concerning both human health and the
environment. Environment Canada is
funding and continues to fund basic
research in microbial ecology
addressing such issues as:
* mechanisms for gene exchange in
the environment '
* development of methods for detec-
tion, monitoring and determination
of environmental fate -of
microorganisms, • '
i
* microbial ecological profiles on
specific genus of organisms
targeted for environmental release,
* environmental codes of good prac-
tice for contained research and de-
velopment involving biotechnology
* characterization and quantification
of emission streams from
• bioprocessing facilities, and
* • identification of factors/character-
istics that affect survival,
persistence, and dissemination of
microorganisms in the environment.
The data generated by our research
programs will not only assist us in our
regulatory program but will be a source
of valuable scientific information on the
behavior of microorganisms in the
environment for industry, academia,
and other government departments.
We constantly seek to promote
national and international cooperation
and information exchange in bio-
technology through our involvement in
a number of interdepartmental commit-
tees under the National Biotechnology
Strategy, the OCED, federal provincial
regulatory consultations, and more
recently, under a memorandum of
understanding (MOD) we have with the
Environmental Protection Agency in
United States.
We believe that we are establishing a
sound and responsive health and
environmental oversight structure for
biotechnology in Canada under the
Canadian Environmental Protection Act.
Conclusion
The design of a regulatory oversight
mechanism for any activity, particularly
one as novel .and exciting as bio-
technology will not be an easy task.
We are however, guided in our
endeavours by previous experiences in
the regulatory arena that have
demonstrated unequivically , that a
consensus building process involving all
stakeholders can be a powerful and
effective tool for the establishment of
an appropriate regulatory framework.
19
-------
USDA'S BIOTECHNOLOGY RISK ASSESSMENT RESEARCH GRANTS PROGRAM
D.R. MacKenzie
National Biological Impact Assessment Program
Cooperative State Research Service
U.S. Department of Agriculture, Washington, D.C. 20250
Section 1668 of the 1990 Food
Security Act (a.k.a., Farm Bill) directs
the Secretary of Agriculture to establish
a grants program in biotechnology risk
assessment research. The program is
to be funded as a 1 % set-aside of the
Department's biotechnology research
expenditures. This Section also
specifies necessary coordination with
EPA and consultation with Department
agencies in establishing the direction
and activities of the grants program.
The responsibilities for administering
the grants program have been assigned
to the National Biological Impact
Assessment Program. At the time of
this writing, discussions are underway
on the dimensions of the program, and
the mechanisms for funding. A
meeting with representatives from the
appropriate federal agencies and
interested public groups provided a
starting point for discussions on how
the program could best meet the spirit
of Section 1668.
One of the obvious intentions of the
program is to meet the needs of federal
regulatory agencies for sufficient
information to make informed, science-
based decisions on regulatory permits
for field testing with genetically
modified organisms. It has thus been
proposed that an advisory committee
solicit "wish lists" each year of
researchable questions that could be
incorporated into the annual request for
proposals. The received grant pro-
posals would then be "flagged" for
interest by the regulatory and research
agencies of the USDA. Subsequently,
a panel of independent peer scientists
would rank the proposals for scientific
merit. These rankings would then be
reviewed by the regulatory and research
agency representatives for interest and
appropriateness. The scientifically
meritorious, high-priority proposals
would then be recommended for
funding.
Through these mechanisms, a new
set of questions could be asked each
year, but the duration of the research
undertaken to answer those questions
would likely extend for several years to
obtain the necessary answers. For
instance, a research proposal might
look at seed dormancy factors of
genetically engineered canola. It might
be necessary to study the seed buried
in different types of soil over a period
of years. This reality would not limit
the program from asking new questions
each year, and then funding them for a
period of time appropriate to the
question.
Another recommendation from the
first meeting was that an annual
conference should be held with the
scientists receiving funding from the
program. This conference would be a
forum for the exchange of information,
and for annually reporting results to all
20
-------
interested parties. It has been
proposed that this annual meeting of
the USDA Biotechnology Risk
Assessment Research Grants recipients
be linked to the EPA's Biotechnology
Risk Assessment Research annual
meeting, for mutual benefit. This
proposal is now under discussion
between the two agencies.
What has not been decided as of this
writing is the definition of bio-
technology, which will be critically
important for determining the amount
of funds available to support the grants
program in risk assessment research.
A very broad definition of bio-
technology for the Department of
Agriculture would circumscribe about
$ 140,000,000, yielding at 1 % $1.4
million for risk assessment research. A
very narrow definition of biotechnology
research being conducted in the field by
the U.S. Department of Agriculture
would ,yield .$60,000 - a rather
insignificant sum.
Whatever the final outcome of the
mechanisms used to select and,support
biotechnology risk assessment
research, there is a clear and evident
commitment by 'the Department of
Agriculture to launch this program as a
responsible complement to its ongoing
investment in biotechnology research.
The solicitation for proposals should be
issued early summer 199.1 for funding
in early Fiscal Year 1992 (begins
October 1, 1991).
21
-------
THE EPA PERSPECTIVE ON BIOTECHNOLOGY
Elizabeth Milewski
Office of Pesticides and Toxic Substances
U.S. Environmental Protection Agency, Washington, D.C. 20460
For centuries, people have used living
organisms to produce useful substances
such as bread, cheese, beer, and wine,
and more recently to perform useful
functions such as the control of harmful
insects. Initially, use of the organisms
was empirical; however, over the
centuries people have developed
techniques to manipulate the genetic
materials of useful organisms to select
and enhance desired characteristics.
Recently, powerful new techniques in
molecular biology have been developed
that permit the manipulation of the
genetic materials of organisms in ways
that heretofore were not possible, such
as moving genetic material between
organisms that would not normally
exchange such material. Public
awareness and concern that use of
these new technologies may be as-
sociated with dramatic consequences
have resulted in a call for some type of
oversight to ensure adequate protection
of human health and the environment.
FEDERAL OVERSIGHT
The United States Environmental
Protection Agency (EPA) is one of
several United States agencies involved
in the regulation of biotechnology
products as" part of a coordinated
federal effort. During the past several
years, EPA has been developing the
specifics of its regulatory approach to
these products. In order to put EPA's
approach into perspective, it is helpful
to have an overview of the federal
concept encompassed within a co-
ordinated federal effort. In 1984,
.recognizing its responsibilities to
address issues raised by the use of
biotechnology, the United States
government formed an interagency
working group under the White House
Cabinet Council on Natural Resources
and the Environment. This group
examined existing laws and concluded
that for the most part, these laws
would adequately address regulatory
needs for biotechnology. A regulatory
matrix describing applicable laws and
responsible agencies was published in
the Federal Register on November 14,
1985 (50 Fed. Reg. 47174).
Subsequently, the "Coordinated
Framework for Regulation of
Biotechnology" was published on June
26, 1986 (51 Fed. Reg. 23302). As
used in the Coordinated Framework, the
term biotechnology is broadly defined
as "the application of biological
systems and organisms to technical and
industrial processes" (see "Proposal for
a Coordinated Framework for
Regulation of Biotechnology," 49 Fed.
Reg. 50856, December31, 1984). The
Coordinated Framework spells out the
basic federal philosophy for regulating
products of biotechnology, details the
network of agency jurisdiction over
both research and commercial products,
and includes statements of regulatory
policy from the agencies principally
responsible for such regulation. The
22
-------
United States approach to regulating
biotechnology is based on the
recognition that biotechnology
encompasses a large and varied
collection of techniques and activities,
and that the potential products resulting
from this technology will cover a wide
spectrum of uses. The Coordinated
Framework provides that biotechnology
products will be regulated in the United
States as are products of other
technologies -- that is, by the various
regulatory agencies on the basis of use.
Thus, many agricultural uses of
microorganisms, plants, and animals are
regulated by the Department of
Agriculture; foods and drugs are
regulated by the Food and Drug
Administration; microorganisms used as
pesticides are regulated by EPA under
the Federal Insecticide, Fungicide, and
Rodenticide Act (FIFRA), 7U.S.C. §136
et seq.; commercial uses of micro-
organisms not covered by other existing
authorities are regulated by EPA under
the Toxic Substances Control Act
(TSCA), 15 U.S.C. §2601 et seq.
Examples of uses covered by TSCA
include microorganisms used in metal
mining, degrading wastes, conversion
of biomass for energy, production of
proteins and enzymes for non-
pharmaceutical purposes, and non-
pesticidal agricultural applications such
as nitrogen fixation.
EPA REGULATION
As indicated above, an essential
characteristic of any regulatory scheme
for biotechnology is that is promotes
public confidence that the oversight
process will protect human health and
the environment. Similarly, the
scientific and regulated communities
must believe that the process is
credible, equitable, and sound. At the
same time, it is important that the
development and commercialization of
biotechnology products be allowed to
proceed in a reasonable and timely
manner. Accordingly, EPA's goals in
developing a regulatory scheme for
products of biotechnology are to (1)
protect human health and the
environment, (2) meet the public's need
to be assured that use of living
products does not present unacceptable
risk, (3) expedite the regulatory
process, (4) maintain flexibility in order
to respond to rapidly developing
knowledge and technology, and (5) to
accomplish this end without stifling
innovation and the development of
valuable hew products.
At this time, EPA's regulatory effort
involving these products is, in large
part, pursuant to its authority under
FIFRA and TSCA. FIFRA creates a
statutory framework under which EPA,
through a registration process,
regulates the development, sale,
distribution, and use of pesticides,
regardless of how these pesticides are
made or their mode of action. FIFRA,
therefore, covers natural and genetically
altered microbes that are used for
pesticide purposes. EPA has routinely
reviewed and registered natural
microbial pesticides, such as Bacillus
thuringiensis, for years.
A pesticide can be registered for use
only if the pesticide will not cause
unreasonable adverse effects to
humans or the environment. The
"unreasonable adverse effects" test
involves a weighing of the risks and
benefits of use of the pesticide. In
23
-------
order to demonstrate that a pesticide
will not cause unreasonable adverse
effects, an applicant seeking to register
the product must submit or cite data on
subjects such as product composition,
toxicity, environmental fate, and effects
on nontarget organisms.
Some of the information needed to
support registration must be developed
under actual field conditions, and such
preregistration field testing is authorized
under FIFRA through the experimental
use permit process. The results of
such early testing are a critical element'
in determining when and under what
circumstances a full registration may be
issued.
TSCA gives EPA jurisdiction over the
manufacturing,processing, importation,
distribution, use, and disposal of all
chemicals in commerce (or intended for
entry into commerce) that are not
specifically covered by other regulatory
authorities (e.g., substances other than
foods, drugs, cosmetics, and
pesticides). TSCA authorized EPA,
through a "screening" process, to
acquire information on chemical
substances and mixtures of chemical
substances in order to identify and
regulate potential hazards and
substantial exposures. Under TSCA,
EPA can require testing of any chemical
substance that may present an
unreasonable risk to human health or
the environment or which'is produced
in substantial quantities and may result
in substantial environmental release or
substantial human exposure.
TSCA's applicability to the regulation
of microbial biotechnology products is
based on the interpretation that living
organisms are chemical substances
under TSCA - an interpretation that
was embodied in the development of
the initial chemical inventory in 1979.
The basis for this interpretation is that
all substances, living and nonliving,
have a chemical foundation at the most
fundamental molecular level. As a
result of this interpretation,
microorganisms (except for those in
excluded use categories) are subject to
all provisions of TSCA.
The premanufacturing notice (PMN)
process implements TSCA's goal of
screening new substances before they
enter commerce in order to identify
those that pose potential hazards.
Manufacturers and importers of "new"
chemical substances are required to
submit data and other information to
EPA that will allow the Agency to
evaluate the potential risks of the
product. If a potentially unreasonable
risk is identified during the PMN
screening process, EPA is authorized to
[exert] comprehensive oversight of the
commercialization of the affected
product.
Because microorganisms introduced
into the environment have the ability to
replicate, increase in number, and
disseminate from the test site, EPA
feels that it is prudent to address
certain questions before some of these
microorganisms are released to the
environment. Accordingly, EPA's
regulatory approach encompasses
procedures designed to provide
sufficient oversight of the initial stages
of environmental testing of certain
microbial products of biotechnology. In
its section of the Coordinated
Framework, EPA indicated that, under
FIFRA and TSCA, ^intended to focus
regulatory emphasis on three categories
of microorganisms:
24
-------
1. microorganisms with "new"
characteristics or that are new to
the environment in which they
are intended to be used (and
whose behavior is therefore less
predictable),
2. microorganisms that have
hazardous components or
characteristics (and therefore a
recognized potential to cause
adverse effects in other
organisms), and
3. microorganisms that are used in
the environment (and therefore
have the potential for widespread
exposure).
Through its statement in the
Coordinated Framework, EPA made
certain provisions immediately effective
and indicated that rulemaking would be
required before other provisions could
be formally implemented. Where
provisions were not immediately
effective, EPA invited voluntary
compliance in the interim.
One of the most complex (and
contentious) issues for the EPA
biotechnology rulemaking process has
been identification of the scope of
organisms to be subject to regulatory
oversight. Because this is a cross-
cutting issue for all the agencies
included in the Coordinated Framework,
development of a "scope, definition"
was undertaken as a joint effort. The
result of this effort, "Principles for
Federal Oversight of Biotechnology:
Planned Introduction Into the
Environment of Organisms With
Modified Hereditary Traits", was
published on July 1, 1990, and serves
as guidance for EPA and the other
agencies during their rulemaking
processes.
In developing the Principles for
Scope, the agencies relied upon
"familiarity" as the basis for
determining whether oversight is
appropriate, and emphasized the role of
information in the determination of
"familiarity". (The appropriateness of
using familiarity based on solid
information as the basis for a flexible
regulatory process is clearly supported
by the 1989 National Academy of
Sciences report, "Field Testing
Genetically Modified Organisms".)
Specifically, the Principles for Scope set
forth the general directive that "to the
extent permitted by law, planned
introductions into the environment of
organisms with deliberately modified
hereditary traits should not be subject
to oversight... unless information
concerning the risk posed by the
introduction indicates that oversight is
necessary." Also included are
examples of organisms that would
"generally be excluded from oversight
because their introduction is considered
to be similar to previous safe intro-
ductions, or because other information
available regarding the risk posed by
such introductions (including knowledge
that existing practices or regulations
adequately address possible risk posed
by the introduction) makes such
oversight unnecessary." With the
Principles for Scope as guidance, and
utilizing provisions available under both
TSCA and FIFRA that allow for
flexibility in the regulatory process, EPA
is [well on its way to establishing the
regulatory program promised in the
Coordinated Framework]. The Agency
25
-------
is currently nearing completion of the
drafting process for its biotechnology
regulations under both TSCA and
FIFRA, and it is-anticipated that the
proposed rules will be published for
comment in the near future.
EPA REVIEW PROCESS
AND EXPERIENCE
Since 1984, EPA has reviewed
several small-scale tests using products
of biotechnology, many of which
involved release to the environment. A
scientific review process for such tests
has been developed that EPA believes is
credible, sound, and respected by the
scientific community and which allows
the public to feel confident that
potential human or environmental
impacts have been addressed. To date,
the review procedures, which are
briefly described below, have involved
case-by-case evaluation.
When a submission for small-scale
testing is received, TSCA or FIFRA staff
groups evaluate the submission and
develop a coordinated scientific
position. During this process, hazard
and exposure are addressed, potential
problems, issues, or significant
unanswered questions are identified,
and the likelihood of significant risk
from the proposed test is assessed.
Under both TSCA and FIFRA, intro-
agency workgroups then comment on
the positions developed by staff. If
appropriate, the submission and EPA's
scientific position are sent to other
federal agencies for comment.
Appropriate state regulatory agencies
are also contacted in order to alert them
to the submission, to discuss EPA's
assessment, and to ensure that the
federal and state positions are as
consistent as possible. For some
submissions, visits to the test sites are
conducted in order to evaluate actual
field conditions.
In order to obtain an independent
peer review of EPA's scientific position,
to address specific scientific questions
raised by staff, or to identify any
additional data that may be needed to
complete the risk assessment, the
submission and the EPA's scientific
evaluation may be sent to a group of
independent scientists. EPA has
established a specific advisory
committee, the Biotechnology Science
Advisory Panel (SAP) for these
independent peer reviews. Several
proposed field trails have been reviewed
by specially convened BSAC, SAP, or
joint BSAC/SAP subcommittees, and
these groups have consistently
endorsed the EPA assessments.
Public comment is considered an
important aspect of the reviews, and
for many proposals the public is
provided several opportunities to
comment during the review process.
In some cases, microorganisms that
are subject to TSCA or FIFRA are also
subject to the Federal Plant Pest Act or
other statutes administered by the
Animal and Plant Health Inspection
Service (APHIS) of the United States
Department of Agriculture. In such
situations, APHIS and EPA conduct
coordinated reviews. The two agencies
cooperate closely, alerting each other to
submissions, and sharing expertise and
information. This close cooperation
benefits both submitters and the
agencies.
At the conclusion of a review, EPA
determines whether the microorganism
26
-------
may be released into the environment.
Under both TSCA and FIFRA, this
decision is based on an assessment of
both the potential risks and the
potential benefits of the proposed use
of the microorganism. Should the
analysis indicate that use of the
microorganism may pose unreasonable
risks, EPA has authority under either
statute to impose restrictions on its
use.
Between the two programs, TSCA
and FIFRA,. EPA has reviewed
approximately 75 submissions, some of
'which have been "second-generation"
variations of previously approved tests.
While it is still too early to speak in
general terms or draw definitive
conclusions about the reviews, certain
observations can be made:
1. EPA's review process has been
effective in identifying and
addressing risk concerns.
2. Evaluation of the proposed tests
often requires different types of
background studies, data bases,
and scientific expertise than are
required for review of traditional
chemical products. Similarly,
monitoring for living
microorganisms once they have
been released to the environ-
ment is frequently more
complicated.
3. The issues giving rise to the
greatest concerns are often not
related to whether the
microorganism performs its
intended function, but rather to
other, nonintended impacts that
might occur.
4. Many (if not most) of the risk
concerns about these micro-
organisms have been related to their
potential impacts on the environment
and nonhuman, nontarget species; to
date, potential impacts on human
health have not been a significant
concern.
5. Monitoring data from some of the
tests have confirmed that under
actual use conditions, physical
containment of the test micro-
organisms may not be feasible or
possible. Accordingly, issues related
to spread of the test organisms from
the test site are appropriately
considered in most risk
assessments.
6. The small-scale tests reviewed thus
far have been found to pose no
significant risks. However, large-
scale or commercial use of some of
these microorganisms may raise
more significant concerns,
particularly regarding environmental
and ecological impacts, than those
addressed for small-scale tests.
As to nonscientific observations, it
has become increasingly clear that
public involvement is an important
element of most test programs
involving environmental release.
AA/ithout such involvement, public fear
of the unknown could derail the whole
process and prolong the testing through
extensive administrative and judicial
maneuvers. The impact of such
administrative and judicial activity can
be seen by contrasting the first and a
more recent EUP review of two
genetically engineered microbial
27
-------
pesticides. In the first, the time
elapsed between receipt of the EUP
application and the test was almost two
years. In between were two lawsuits
brought in federal and state courts to
stop the test, and numerous state and
local administrative proceedings. The
more recent EUP application, submitted
after more extensive preliminary public
outreach and presumably after the
public had gained more confidence in
EPA's review process, moved from
submission to field testing approx-
imately five months. There were no
lawsuits or supplemental administrative
proceedings involved with this review.
It appears that the more the public
knows and feels a part of and
comfortable with the process, the less
likely it is to obstruct the test. Hence,
many submitters have elected to shield
from public scrutiny only the barest
essentials about their microorganisms
and proposed tests.
EPA is using the experience gained
from its early test reviews, along with
public comments and advice from our
BSAC, in fashioning its proposed
regulations. Based on these sources,
EPA has bolstered confidence that its
basic approach as set out in the
Coordinated Frameworkand procedures
developed for individual test reviews
are adequate to accomplish the goals
for its biotechnology regulatory
program. Therefore, in many areas,
EPA's proposed regulations are
envisioned in much the same light as its
policy statement in the Coordinated
Framework.
To summarize, EPA's experiences
over the last few years with the
approach for regulating products of
biotechnology first articulated in 1984
and later included in the 1986
Coordinated Framework, demonstrate
that the process has met its major
goals:
* utilization of a systematic approach
for identifying and assessing
potential human or environmental
risks,
* incorporation of a peer review
mechanism that draws upon
independent scientists expert in
diverse areas of relevant knowledge,
* establishment of an open process
that allows for and invites public
participation, and
* approval of field tests, the results of
which confirm EPA's original
judgment of insignificant risk.
Notwithstanding these successes,
EPA recognizes that certain modifi-
cations to the current process are
probably appropriate in order to
accommodate in a timely manner the
number of tests anticipated for review
in the future. EPA is currently
developing proposed regulations that
will implement its modified approach.
However, based on its past experi-
ences, EPA is confident that its ultimate
regulatory program will protect the
public health and the environment from
unreasonable risks while allowing the
public to reap the benefits of scientific
and technological innovations. "-
28
-------
OFFICE OF TOXIC SUBSTANCES PURVIEWS AND RESEARCH NEEDS
Philip G. Sayre
Health and Environmental Review Division, Office of Toxic Substances,
U.S. Environmental Protection Agency, Washington, D.C. 20460
The Office of Toxic Substances
(OTS) implements the Toxic Substances
Control Act (TSCA), which applies to
commercial microbial products. The
microbial products reviewed under
TSCA are those that are specifically
excluded by other federal statutes, e.g.,
pesticides (which are regulated by the
EPA's Office of Pesticide Programs
(OPP) under FIFRA), foods, drugs, and
cosmetics which fall under the purview
of the Food and Drug Administration,
and nuclear materials regulated by the
Nuclear Regulatory Commission. The
products reviewed thus far by OTS fall
into two categories: (1) fermentation
applications for production of cellular
components such as enzymes to be
used in the production of ethanol,
detergents, and culture media; (2)
microbial products intended for
intentional environmental release such
as recombinant rhizobia for enhanced
nitrogen fixation in legumes. Other
applications which are subject to TSCA
include for example those for degrad-
ation of toxic wastes, metal mining,
desulfurization of coal, enhanced oil
recovery, and biomass conversion (for
reduction of solid waste). Products for
bioremediation of toxic compounds
-.such as organic solvents, crude and
refined petroleum wastes, and creosote
may account for a significant portion of
future submissions to OTS.
There are several differences in the
research needs of OTS and OPP, which
naturally follow from the current
interpretation of TSCA as it applies to
microbial products. First, a potentially
diverse group of microorganisms and
uses are subject to TSCA. Second, the
microorganisms reviewed by OTS are
not intended to be toxic or pathogenic,
as are microbial pesticides. Third, OTS
examines submissions for fermentation
applications of microorganisms. OPP
does not, although FIFRA provides for
ensuring quality control in the
manufacturing process. Fourth, the
OTS risk assessment process places
more emphasis on microbial fate
analysis, in addition to examining
human health and ecological effects
concerns. Concerns for microbial fate
necessitate research which leads to
better detection of microorganisms, and
better estimation of microbial dispersal
as a result of fermentation releases and
intentional environmental releases.
Although interest in research on
microbial fate will continue, the
research program is now shifting to
place a stronger emphasis on human
and ecological effects.
Currently, OTS is in the process of
developing tiered testing schemes for
both ecological effects and human
health. In order to implement these
testing schemes, additional protocols
beyond those developed for OPP, will
be needed. OTS has begun a joint
project with ORD to develop simple
screening-level tests to determine the
survival and competitiveness of
genetically engineered microorganisms
29
-------
in various environmental media.
Further, since exposure to OTS-
reviewed products often includes areas
beyond the agricultural environment,
protocols for effects testing should be
developed for different species of
organisms that are not only important
ecologically but are commonly found in
the general environment (such as some
ant species). Protocols may also need
to be developed which can discriminate
between effects of a microorganism
and its metabolites to address both
human health and ecological concerns;
such protocols are particularly relevant
to the evaluation of microorganisms
used for bioremediation of toxic
pollutants or recalcitrant compounds.
Due to the fact that many microbial
products reviewed under TSCA are not
designed to be toxic/pathogenic but
may play a role in other critical
interactions (e.g., mineral cycling), OTS
will continue its interests in test
systems such as microcosms which
examine a broader range of effects
concerns. Finally, OTS will continue to
seek support as it develops computer-
ized risk assessment tools such as
those for gene sequence analysis,
microbial taxonomy, and microbial fate
modelling.
30
-------
THE OFFICE OF PESTICIDE PROGRAMS: REGULATION OF
MICROBIAL PESTICIDES AND RESEARCH NEEDS
William R. Schneider
Office of Pesticide Programs
U.S. Environmental Protection Agency,
Washington, D.C. 20460
The Office of Pesticide Programs
(OPP) is responsible for regulating all
pesticides under the Federal Insecticide,
Fungicide, and Rodenticide Act (FIFRA)
and the Federal Food, Drug, and
Cosmetic Act (FFDCA). FIFRA defines
pesticides as any substance, or mixture
of substances, intended for preventing,
destroying, repelling, or mitigating any
pest, and specifically includes plant
regulators, defoliants, or desiccants.
Before the Agency can register a
pesticide, FIFRA requires the Agency to
have sufficient data to determine that
the pesticide, when used in accordance
with widespread and commonly
recognized practice, will not cause (or
significantly increase risk of)
unreasonable adverse effects to
humans of the environment.
The microbial pesticides include
bacteria, viruses, fungi, protozoa, and
algae. The first microbial pesticide was
registered in 1948 (Bacillus popilliae]
and 20 others have been registered to
date. Of those, 8 have been registered
in the last 2 years. Over one third of
the new pesticides registered recently
have been microbial pesticides. This is
indicative of the overall trend for
industry to turn to biological pesticides
to address the agricultural problems of
the 90's: insect resistance to chemical
pesticide and the need for safer
pesticides. These biologically derived
pesticides include the microbial
pesticides, non-toxic biochemicals such
as pheromones, and the newest and
largest group: plants that are genetically
engineered to product pesticidal
chemicals. Although our regulatory
system for transgenic plants is not yet
in place, we have reviewed over 50
small scale field tests through a
memorandum of understanding with the
U.S. Department of Agriculture Animal
and Plant Health Inspection Service.
Part 158 of 40 CFR 158 specifies
the data and information that must be
submitted to EPA to support regis-
tration of pesticides. Subdivision M of
the Pesticide Testing Guidelines (NTIS
# PB89-211676, July, 1989) provides
information relating to the data
requirements listed in 40 CFR 158.690
and 158.740, including conditions
under which each data requirement is
applicable, standards for acceptable
testing, information that should be
included in a test report, guidance on
evaluation and reporting of data, and
examples of protocols. In addition,
scientific publications and ORD
laboratory reports are cited in the
guidelines as useful information for
designing test protocols. OPP must be
able to evaluate these protocols since
40 CFR 160.120, the regulation for
Good Laboratory Practice Standards,
specifies that "each study shall have an
31
-------
approved written protocol....".
In addition to registering microbial
pesticides, the Agency must perform
risk assessments to evaluate
applications for Experimental Use
Permits (EUPs) as described in 40 CFR
172. Microbial pesticides must have
EUPs before being tested on cumulative
total of not more than 10 acres of land
or not more than one surface-acre of
water, or if to be used for food or feed
uses, or on water containing animals or
plants that my be used for food or feed.
In addition, some microbial pesticides
are subject to an interim policy,
published in the June 26, 1986
FEDERAL REGISTER (51 FR 23302).
This policy essentially states that the
Agency must be notified prior .to
conducting small scale field testing of
genetically altered or non-indigenous
microbial pesticides in order to allow
the Agency to determine if an EUP
should be obtained. Under this policy,
OPP has reviewed a number of interest-
ing constructs including ice enucleating
gene deletions in Pseudomonas, and
Bacillus thuringiensis toxin genes in
Pseudomonas and in Clavibacter xyli.
Indications are that most microbial
pesticide genetic engineering work in
the immediate future will involve toxin
genes inserted into baculoviruses. In
reviewing these notifications, we have
found that the majority of the risk
issues are the same as for the wild-
type, naturally occurring microbial
pesticides.
OPP uses a risk assessment
framework for microbial pesticides that
was derived from traditional chemical
assessment methods, i.e. the overall
risk is a function of both exposure and
hazard (the effects on non-target
species, including human). In order to
minimize the data requirements, OPP
has elected to require environmental
exposure data only if unacceptable
hazards are observed in non-target
species or short-term mammalian tests.
Although we have recommended
specific protocols for many tests in
Subdivision M, relatively few have been
performed and some tests have been
difficult to perform and evaluate.
Accordingly, OPP has identified
Subdivision M protocol development
and evaluation as its primary research
need. Some problems with the eco-
logical effects testing are: (1) We
generally test adult honeybees since
larvae are difficult to keep alive, yet
larvae are the most likely to be sus-
ceptible to biological insecticidal toxins.
(2) It has been difficult to maintain
Daphnia for a sufficient time to evaluate
potential pathogenicity. (3) Very few
species of beneficial insects are
available for testing. (4) Which aquatic
species are most at risk and are we
adequately protecting them? (5) Do the
protocols adequately reflect a maximum
hazard approach?
In addition to the actual testing
protocol development, there is a need
for information that allows us to
evaluate the study results. In many
cases, we do not have much inform-
ation on the role of pesticidal gene
products in the environment, even for
the most prevalent Bacillus thuringiensis
toxins. Bacillus thuringiensis is
ubiquitous in soil but not at levels that
appear to affect insects. Microcosm,
mesocosm, and/or field studies could
be utilized to study the relationship of
Bacillus thuringiensis and/or its many
toxins to the soil and aquatic eco-
32
-------
systems. Are there alternative soil
hosts? Do the toxins give Bacillus
thuringiensis a competitive advantage
over other soil microorganisms?
Potential human health effects of
pulmonary exposure to microbial
pesticides are one of our primary
concerns, since bacterial, fungal, and
viral pesticides are frequently broadcast
over large areas by aerial applications.
These extensive uses present pos-
sibilities for human epidemiology
studies and for relating experimental
test results with actual effects.
Health research has centered on the
tier 1 acute pulmonary exposure tests
for Subdivision M of the Pesticide
Assessment Guidelines. These tests
were developed, standardized and
evaluated for the assessment of the
fnfectivity primarily of naturally-
occurring agents. In these studies,
intratracheal and intranasal challenges
were demonstrated to be equivalent to,
if not better than, aerosol inhalation for
the assessment of pulmonary exposure
to infectious bacterial and viral agents.
In addition, results from EPA-sponsored
studies have shown that pulmonary
exposure of rodents to Bacillus
thuringiensis have shown adverse
toxicological effects at sufficiently high
dose levels. The extent of toxicity
observed appears related to differences
in the Bacillus thuringiensis strain
administered. While this is a significant
first step in methodology and protocol
development for testing guidelines,
relevant new technologies and
innovative methods must be in-
corporated to increase the breadth of
coverage of these tests for a wider
range of agents in the future. This can
be attained by obtaining, through
research, a better understanding of the
basis of these empirical tests. Areas of
additional inquiry should address the
microbial agent/cellular interactions,
factors influencing the reisolation and
identification of the microbial pest-
icides, the utility of these tests for
repeated pulmonary exposures, the
equivalence of intratracheal and intra-
nasal challenge results with aerosol
inhalation results when toxicity instead
of infectivity is the endpoint, and the
scientific underpinnings of new ap-
proaches so appropriate tests can be
developed to address problems such as
assessment of the potential for transfer
of deleterious genes to mammalian
tissues and cells by more innovative
and complex genetically altered agents
which can be expected to be developed
for such problems as assessment of the
potential for transfer of deleterious
genes to mammalian tissues and cells
by more innovative and complex
genetically altered agents which can be
expected to be developed in the future.
Microbial agents of greater anticipated
use should be identified and studied to
better understand any adverse cellular
responses or effects that might be pro-
duced in the pulmonary tests. In
particular, Baci/fusthuringiensis,bacu\o-
viruses, and Beauvaria bassianna have
high pulmonary exposure potential.
Additionally, as successful agents are
more utilized and exposure becomes
more frequent, components of microbial
pesticides of minor relevance to the
human health effects must be
reassessed.
33
-------
-------
SESSION I
ENVIRONMENTAL EXPOSURE RESEARCH
Detection and Enumeration
Scientists are using serological, biochemical, physiological, and genetic methods for
improving detection and enumeration of microorganisms under laboratory, microcosm,
and field conditions. Methods are being developed and refined for their sensitivity,
specificity, and reliability in a variety of terrestrial, aquatic, and air habitats. The
following extended abstracts summarize continuing research in detection and
enumeration of microorganisms of special interest.
Dispersal and Transport
Developing a better understanding of the movement of microorganisms through the
environment, either by the natural elements, or through associations with higher
organisms is crucial. The mechanisms and dynamics of transport among and within
various environmental components (air, soil, groundwater, plants, insects, and other
animals) are being studied.- This research includes development of mathematical
models and other aids for predicting transport and exposure. The following extended
abstracts summarize continuing research in the potential of microorganisms of special
interest to disperse or be transported from the site of release.
Survival and Colonization
Understanding the factors that affect the ability of microorganisms to survive and
colonize habitats under a variety of conditions is central to this research element.
Cellular, molecular, and environmental factors influencing survival or actual
multiplication are being identified and described.
Gene Transfer
This research provides test methods that describe conditions for determining the
frequency and probability of genetic exchange between microorganisms released to
the environment and their indigenous counterparts. Factors that affect and control
gene stability and rates of transfer in the environment are being evaluated.
35
-------
-------
MOLECULAR STUDIES OF MICROBIAL ECOSYSTEM PERTURBATIONS
D.A. Stahl
University of Illinois
Urbana, Illinois
INTRODUCTION
A central question in environmental
microbiology is: how do the activities
of microbial communities correlate with
the diversity of populations responsible
for those activities? This question
remains essentially unanswered
because of well recognized limitations
of pure culture isolation and existing
determinativeschemes of classification.
Pure culture isolation and phenotypic
characterization have not served to
adequately describe environmental
diversity. However, the need for pure
culture isolation is no longer a
prerequisite. Techniques of molecular
biology (including recombinant DNA,
nucleic acid hybridization, molecular
systematics, comparative sequencing
and the polymerase chain reaction
(PCR)) now provide additional tools for
characterizing natural populations. In
particular, comparative sequencing of
the ribosomal RNAs is serving to unify
microbial systematics and has provided
the foundation for the explicit
characterization of natural microbial
diversity.
We have developed a variety of
techniques using comparative rRNA
sequencing and hybridization for use in
the characterization of microbial
populations in natural communities.
These techniques and a phylogenetic
framework have served our ongoing
studies of microbial population
ecology; addressing questions of
community stability, competitive
exclusion, succession and synergism,
community response to perturbation,
and the relationship between
community structure and community
function. These questions are central
to microbial ecology. From a more
practical standpoint, the rational
manipulation of the activity of
environmental populations [e.g. via the
introduction of exogenous organisms
(or selected DNA encoded traits)]
should be based on an understanding of
population ecology. Similarly, rational
risk assessment is dependent upon this
understanding.
Our research with the EPA has
focused upon two "model" com-
munities to further develop comparative
molecular methodologies and establish
basic paradigms of microbial ecology.
These communities are the bovine
ruminal microbial community and those
represented by marine anaerobic
sediments. Studies of the ruminal
community have focused on the
principal fiber-digesting populations
whereas sediment studies have
examined the diversity and distribution
of sulfate-reducing bacteria in
relationship to their activity. This
abstract addresses only the ruminal
studies. Studies of the environmental
diversity and activities of sulfate-
reducing bacteria will be addressed by
Dr. R. Devereuxand M. Winfrey, whose
work builds upon our earlier studies of
the sulfate-reducing bacteria.
37
-------
METHODS
Comparative 16S rRNA Sequencing and
Oligonucleotide Probe Design
The determination of partial or near-
complete 16S rRNA sequences has
been the foundation for the population
studies. Initially we focused on the
now-recognized major fiber-digesting
populations of the rumen (the anaerobic
ruminal fungi, Butyrivibrio fibrisolvens,
species of Ruminococcus and
Fibrobacter). These studies emphasized
the above stated limitations of the
existing taxonomy; organisms classified
as strains of Fibrobacter succinogenes
and F. fntestina/is (previously
Bacteroides succinogenes) and
Butyrivibrio fibrisolvens were
demonstrated to encompass diversity at
the levels of genus or greater. This is
an observation common to other
natural communities . as well.
Population diversity (genetic diversity)
has generally been greatly under-
estimated. Thus, past studies of
environmental abundance and
distribution have most often obscured
the true population/genetic diversity.
Comparative 16S rRNA sequencing
has served both to reveal genetic
diversity and to provide sequence
information to design Oligonucleotide
probes for use in environmental studies.
Probes are either labeled with 32P or a
fluorescent dye. Radioactive probes
have been used for determining
population abundance, either of specific
species (or subspecies) and of larger,
phylogenetically coherent, groups (e.g.
eukaryotes or archaea). Fluorescent
probes have been used to identify
single cells, either directly in
environmental samples or in pure
culture.
The Fibrobacter Paradigm
The genus Fibrobacter provides the
paradigm for our general experimental
approach to defining the environ-mental
diversity of functionally comparable
populations of micro-organisms. Our
work with both the sulfate-reducing
bacteria and the fiber-digesting
populations of the bovine rumen has
shown that many functionally
comparable groups are phylogenetically
coherent. In general, a collection of
microorganisms that is phylogenetically
coherent can also be circumscribed by
regions of rRNA sequence (signature
sequences) common to, and unique to,
the group (a group could consist of a
subspecies, species, genus or larger
more inclusive assemblage). These
signature regions serve as target sites
either for Oligonucleotide probes (used
for identification and quantification) or
PCR primers (for selective sequencing).
Thus, it has been possible to fabricate
probes (and primer sets) to specific
functional target groups.
We have outlined the genetic
diversity within the genus Fibrobacter
by several genetic criteria, including
comparative 16S rRNA sequencing,
DNA similarity, restriction fragment
length polymorphism and Oligo-
nucleotide probe hybridization to both
pure culture isolates and total rumen
contents. The now recognized diversity
within the genus as inferred by
comparative 16S rRNA sequencing and
DNA similarity is displayed in Figure 1.
Oligonucleotide probes specific for the
genus, species and subspecies have
38
-------
NR9
- C1a
DR7
81 A3C S85
HM2
MM4
MB4
BL2
REH9-1
MC1
10%
Bactero/des fragllis
Flavobacterium heparinum
JG1
LH1
DR7
C1a
NR9
r- MC1
JB1
JA3C
r|ss5
Fibrobacter intestinalis
IGroup 4
> subsp. succinogenes
Fibrobacter succinogenes
,MM4
MB4 Group 3
.— i
ilVUVIt \
_IMB4 1(
IHM2 )
10%
B. fragilis
F. heparinum
Figure 1. Comparison of genetic relationships among Fibrobacter as inferred by
comparative 16S rRNA sequencina (bottom) and DMA simiisritw t+™\
sequencing (bottom) and DNA similarity (top).
39
-------
been used both for classification and
natural population studies (below).
Assurance that natural population
diversity has been reasonably-well
circumscribed comes from consistency
of quantification between the general
and specific probes: the sum of groups
quantified by hybridization to specific
probes (e.g. species and subspecies)
should be comparable to the value
determined,by use of a more general
probe (e.g. genus). In, addition, we
have recently constructed a "global"
panel of 66 reference organism nucleic
acids (representing the diversity of
archaea, bacteria and eukaryotes)
against which the specificity of each
probe will be evaluated.
RESULTS AND DISCUSSION
Community Stability and Community
Response to Perturbation
The overall goals of the research are
to define the major ruminal populations,
their constancy (between animals and
over time) and associations
(succession, competitive exclusion,
synergism). Only with appreciation of
general community structure and
population associations can more
specific questions be addressed. We
are establishing the -necessary
overview of community structure via
two interrelated studies. First, we are
establishing normal variation among
major ruminal populations and.second,
we are using this baseline of normal
variation to evaluate changes 'in
abundance of specific 'ruminal
populations following 'perturbation
(altered diet) and with normal diet
variation.
Long Term Population Stability
Studies of population abundance and
distribution among subspecies of
Fibrobacter have revealed unexpected
population stability. Of animals so far
inspected (five steer), only one
subspecies of Fibrobacter predominates
in a single animal at a given time.
Among the five steer examined over
extended time periods (months), either
the group 2 ^subspecies or F.
succinogenes subsp. succinogenes
predominated. Initial population studies
of a single animal demonstrated that F.
succinogenes subsp. succinogenes
predominated in that animal. A more
recent study of fo\jr animals (Figure 2)
demonstrated a "different subspecies
(group 2) to be stably associated with
each of these animals. Although
representatives of other species and
subspecies of Fibrobacter were;
present, their numbers remained low:
throughout the study period (2;
months).
Diel Population Variation
Another approach we are using to
identify population associations and
population successions is to follow
population changes through time
following feeding. A study set of four
animals was fed once daily at 11:00
AM and samples were taken at four
hour intervals over a three day period.
Initial populations^studies examined the
diversity and abundance of different
subspecies of Fibrobacter. Fibrobacter
numbers were depressed immediately
following feeding and reached their
greatest numbers[almost 20 hours after
feeding (as measured by total rRNA
40
-------
COW 1
COW 3
7
6 -
5
2
•E 3i
2
1 -\
0
20 40
DAYS
60
7
6-
5 -
1 3-
2 •
1 -
0
20 40 60
DAYS
COW 2
COW 4
7
6 .
5 -
4-
3 -
2
20 40
DAYS
60
7
6
5-
< 4-
2
1
0
20 40
DAYS
60
Figure 2. Long term ruminal population stability study demonstrating the relative
constancy and stability of populations of Fibrobactersuccinogenes among a study set
of four cows. The species-specific probe and universal probes were used to quantify
relative abundance of this species. Hybridization with the different subspecies-specific
probes has demonstrated that the group 2 subspecies (Figure 1} accounts for most of
the genus-specific hybridization throughout the study period.
41
-------
abundance or fraction of total ruminal
rRNA). This is consistent with this
group functioning to digest more
refractory plant cell wall polymers (and
retained with larger feed particles).
Continued studies will examine the
relationship of this population to other
major fiber-digesting populations (e.g.
ruminococci and the anaerobic ruminal
fungi) during the diel cycle. This should
point up synergism, competition and
succession among populations serving
a similar function role in the ruminal
community (fiber digestion).
Autecological Studies
We earlier demonstrated the use of
fluorescent oligonucleotide probes in
combination with fluorescence
microscopy to visualize individual
bacterial cells. Fluorescent probes have
been used to identify species and
subspecies of Fibrobacter, both in pure
culture and within rat cecal contents.
Continued studies will use fluorescent
probes in combination (each labeled
with a different fluorescent dye and
targeting Fibrobacter subspecies and
other fiber-digesting populations) to
examine the distribution of individual
fiber-digesting bacteria on the surfaces
of plant material. This analysis should
also serve to address issues of niche,
competition and synergism.
Genetic Isolation of Catabolic Traits
An outstanding question in
environmental microbiology is the
fluidity of gene transfer among
populations. The ribosomal RNA offers
an historical framework for evaluating
the stability of specific genes within
microbial populations. Specifically, if a
gene is stably associated with a given
population (i.e. not subject to lateral
transfer), then the genealogy of that
gene should correspond to the
genealogy of the microbial populations
that it is associated with.
In general, catabolic genes appear to
be more readily transferred between
populations. The catabolic genes of
greatest importance to the ruminal
microbial community are those that
participate in the hydrolysis of plant
cell wall material (e.g. poly-
saccharidases). We initially evaluated
the potential lateral transfer of two
polysaccharidases (an endo-glucanase
and a xylanase) among subspecies of
Fibrobacter. Both genes (as assessed
by heterologous hybridization) were
found to be widely distributed among
representatives of the genus. This
raised the question of whether this
distribution reflected lateral transfer or
conservation of structure. To address
this question, we sequenced the
homologous endoglucanse from three
Fibrobacter isolates (representing the
diversity of the genus).
The endoglucanse genes are highly
conserved relative to total genomic
similarity. Excluding the third codon
position from sequence comparisons,
the more highly conserved C-terminus
(ca. 900 nucleotides out of 2000)
shares approximately 75% sequence
similarity with the most distantly
related strains (S85 and DR7) and
approximately 92% with the closest
(A3C and S85). This compares with
total DNA similarity values of 5-7% and
65% (Figure 1). Conservation does not
appear to reflect lateral transfer of
these catabolic genes among sub-
42
-------
species. The relationships of the
different subspecies, as inferred by
sequence divergence of both the
endoglucanses and the 16S rRNAs are
consistent (Figure 3). This data
strongly suggests that this endo-
glucanase is a stable component of the
cell and also is a strong comparative'
argument that the enzymes partici-
pating in fiber digestion are closely
integrated. More generally, the
retrospective approach outlined here
could serve for the assessment of past
gene transfer events among genetically
distinct populations.
FUTURE WORK
Continued studies will continue to
use the nucleic acid samples obtained
from the study set of five animals to
further characterize ruminal population
diversity, stability, population
associations and response to per-
turbation. Initially, major fiber-digesting
populations will be examined in greater
detail. The long term goal is to establish
the basis for stable community
structure (reflected by long term
association of specific populations over
time and following community per-
turbation). Given that the study
animals all received the same diet, an
initial hypothesis is that specific
populations (each representing a subset
of a larger functionally comparable
collection of populations) are stabilized
by overall community architecture. If
so, this should be reflected by specific
associations between populations, for
example, between genetically distinct
populations of fibrobacter and
ruminococci.
REFERENCES (ruminal studies)
Amann, R.I., C. Lin; R. Key, L.
Montgomery, and D.A. Stahl. Diversity
among Fibrobacter isolates: Towards a
phylogenetic and habitat-based
definition of species (accepted to
System. Appl. Microbiol.).
Dore, J. and D.A. Stahl. Phylogeny of
anaerobic rumen Chytridiomycetes
inferred from small subunit ribosomal
RNA sequence comparisons (accepted
Can. J. Botany).
Krumholz, L.R., M.P. Bryant, W.J.
Brulla, J.L. Vicini, J.H. Clark and D.A.
Stahl. Quinella ova/is gen. nov., sp.
nov., a phylogenetic analysis (accepted
Int. J. Syst. Bacteriol.).
Stahl, D.A. and R. Amann. 1991.
Development and application of nucleic
acid probes in bacterial systematics.
in: E. Stackebrandtand M. Goodfellow
(Eds.), Sequencing and Hybridization
Techniques in Bacterial Svstematics.
John Wiley and Sons, Chichester,
England. *
Amann, R.!., I.E. Krumholz, and D.A.
Stahl. 1990. Fluorescent-
oligonucleotide probing of whole cells
for determinative, phylogenetic, and
environmental studies in microbiology.
J. Bacteriol. 172, 762-770.
Amann, R.I., B.J. Binder, R.J. Olson,
S.W. Chisholm, R. Devereux, and D.A.
Stahl. 1990. Combination of 16S
rRNA-targeted oligonucleotide probes
with flow cytometry for analyzing
mixed microbial populations. Appl.
Environ. Microbiol. 56: 1919-1925.
43
-------
Endoglucanase
— F. succinogenes S8S
- F. succinogenes A3c
• F. succinogenes REH9-1
• Fiintestinalis DR7
• C. therm. D
16S rRNA
70%
• DR7
• S8S
-A3C
— REH9-1
• C. petfringens
— C. botulinum
• C. barken
10%
Figure 3. Comparison of endoglucanaseand 16S rRNA sequence relationships among
Flbrobacter,
44
-------
Mannarelli, B.M., D.A. Stahl, and R.J.
Stack. 1990. Genetic relatedness
among strains of Butyrivibrio
determined from DNA analysis,
extracellular polysaccharide (EPS)
composition and 16S rRNA sequences.
Abstracts of the Annual Meeting of the
American Society for Microbiology.
Montgomery, L. Flesher, and D. Stahl.
Transfer of Bacteroides succinogenes
(Hungate) to Fibrobacter gen. nov. as
Fibrobacter succinogenes comb. nov.
and Fibrobacterintestinalissp. nov. Int.
J. Syst. Bacteriol. 38, 430-435
(1988).
Stahl, D.A., B.F. Flesher, H. Mansfield,
L. Montgomery. 1988. The use of
phylogenetically based hybridization
probes for studies of ruminal microbial
ecology. Appl. Environ. Microbiol. 54,
1079-1084.
Stahl, D.A. Phylogenetically-based
studies of microbial ecosystem
perturbations. 1988. In: American
Chemical Society Symposium Volume:
Biotechnology in Crop Protection. P.
Hedin (Ed.) pp. 373-390.
45
-------
USE OF 16S rRNA PROBES TO CORRELATE SULFATE REDUCER COMMUNITY
STRUCTURE WITH MERCURY METHYLATION
Michael R. Winfrey and Janet Winfrey
University of Wisconsin-La Crosse
Department of Biology and Microbiology
La Crosse, Wisconsin
INTRODUCTION
In order to assess the effect of
releasing genetically engineered
microorganisms (GEMs) in the environ-
ment on native microbial communities,
it is necessary to be able to detect
changes in the microbial community. A
potential method to evaluate microbial
ecosystem changes is to quantify
changes in microbial activities and
changes in microbial community
structure in response to a perturbation.
Although methods are available to
quantify most microbial activities,
severe limitations exist in the ability of
microbiologists to characterizemicrobial
communities. Traditional enumeration
and identification techniques require
cultivation of microorganisms, yet many
native microorganisms are difficult or
impossible to cultivate. This dilemma
has precluded any accuratedefinition of
microbial community structure in
natural habitats.
The use of 16S rRNA-targeted probes
to detect microbial groups offers a
method to overcome this limitation in
microbial ecology. Since the probes are
used on total RNA extracted from a
natural sample, this method eliminates
the need to cultivate the micro-
organisms and has the potential to
accurately define microbial community
structure.
Because of the wide diversity of
microbial populations in natural
communities, the use of this technique
as a risk assessment tool should target
microbial groups that play a vital role in
the community. In addition, an ideal
target group would change in
abundance and/or diversity in response
to any changes in the microbial
community. The sulfate-reducing
bacteria (SRB) represent such a
microbial group because they act as
terminal organisms in anaerobic food
chains and are essential to efficient
anaerobic decomposition in habitats
containing sulfate. By consuming
products of bacterial fermentations in
higher trophic levels, they allow
thermodynamically unfavorable
fermentations to occur and "pull" the
entire anaerobic food chain. Thus, a
change in the SRB population would be
detrimental to carbon decomposition in
the environment they inhabit, and
changes in other microbial groups that
provide the substrates for the SRB may
also alter the SRB population.
We used 16S rRNA-targeted probes
to groups of SRB to determine their
community structure in sediment from
northern Wisconsin lakes and correlated
these communities with mercury
methylation and lake pH. Since SRB
are known agents of mercury
methylation and methylmercury
bioavailability is enhanced in low pH
46
-------
lakes, it is possible that sulfate loading
due to acid-rain may play a role in the
enhanced methylmercury production by
stimulating methylation by sulfate
reducers. Thus, acidified lakes provide
an excellent system to evaluate the use
of rRNA-targeted probes to examine
changes in community structure in
response to an ecosystem perturbation.
METHODS
Sample Collection
Lake sediment and water samples
were collected from selected lakes
(ranging in pH from 4.6 to 7.1) in the
Northern Highlands Lake district of
Wisconsin. One of the lakes, Little
Rock Lake, is particularly useful in
studying SRB community structure
because it has been divided and one
basin is being artificially acidified with
sulfuric acid. Sediment samples (upper
5 cm.) were either hand collected by a
diver (for littoral sediments) or with a
peristaltic pump (for flocculent
profundal sediments). All sediments
were placed immediately on ice after
collection and either frozen or extracted
within 24 hours of collection.
Methylation assays were set up within
24 hours of collection. Water samples
were collected using a peristaltic pump.
A Pellicon tangential flow filter system
(Millipore) fitted with four 0.45 fjm filter
cassettes was used on site to
concentrate water down to 1 liter.
Lake water and concentrates were kept
on ice to minimize RNase activity.
Water concentrates were transported to
the lab on ice, centrifuged, and the
pellets frozen. Processing of water
concentrates was done within 24 hours
of collection. Strict anaerobic
technique was used during collection
and storage of anoxic sediment and
water samples.
Methylation Assays
Mercury methylation was measured
by spiking water or sediment samples
with 203Hg(ll) (1.0/yg as 203HgCI2) and
incubating 24 hours at in situ
temperature. Methylation was
quantified by extracting the radioactive
methylmercury produced followed by
scintillation counting.
RNA Extraction from Environmental
Samples
RNA in sediment samples was
initially extracted by placing 1 ml of
sediment in a 2-ml bead beating vial
containing 1.2 g baked zirconium beads
and 0.2 g acid washed polyvinyl-
polypyrrolidone (PVPP). Vials were
then filled with 50 mM sodium
acetate/10 mM EDTA buffer (pH 5.2)
containing p-mercaptoethanoland SDS.
The vial was shaken at high speed on a
Mini Beadbeater for six 1-min intervals
with a two minute incubation on ice
between each beating cycle. The
aqueous phase was separated by
centrifugation and extracted once with
an equal volume of phenol (equilibrated
with 50 mM sodium acetate/10 mM
EDTA (pH 5.2), once with phenol:
chloroform and twice with chloroform.
RNA was ethanol precipitated,
resuspended in RNase-free water, and
stored at -20°C. All extracts were
quantified by absorbance at 260 nm
and the purity was evaluated by
260/280 and 234/260 ratios.
47
-------
In order to quantify the amount of
RNA per volume of environmental
sample, we modified the procedure so
that the percentage of RNA recovered
in the extraction was quantitatively
known. To increase the aqueous
volume in the extraction vials, we
decreased the amount of PVPP by half
(to 0.1 g/vial) and decreased the
amount of zirconium beads from 0.5 ml
(1.2 g) to 0.5 g. The total aqueous
volume of an extraction mixture in a 2-
ml bead beating vial was calculated by
subtracting the volume of non aqueous
materials (PVPP, zirconium beads, and
sediment dry weight) from the total
volume of liquids in each tube. After
bead beating and centrifugation a
constant volume (900 //I) of aqueous
supernatant was removed. After
solvent extraction, we recovered a
constant volume (500 fj\) of the
aqueous extract. From the initial and
final aqueous volumes we calculated
the percent recovery of the RNA in the
sample.
Nucleic Acid Hybridizations and
Detection
RNA extracted from environmental
samples and reference standard RNAs
(from pure cultures of SRB) were
denatured for 10 minutes in 1.5%
glutaraldehyde at 30°C and applied to
nylon membrane with a slot blotter.
Amounts applied ranged from 1 to
1000 ng (based on A260) for environ-
mental samples and from 0.01 to 10 ng
for reference standards. Membranes
were baked at 80°C for 1 hour, then
prehybridized at 40°C for 4 hours in a
hybridization bag containing 0.1 ml
hybridization buffer per cm2 of
membrane. Universal (1400) or SRB-
specific oligonucleotide probes were
end labelled with polynucleotide kinase
and (gamma-32P)-ATP. SRB specific
probes used were: 687 (for
Desulfovibrio), 660 (for Desulfobulbus],
129 (for Desulfobacter], BTM (for
Desuffobacterium], 813 (for
Desulfosarcina, and Desulfococcus),
and 804 (which detects all organisms
detected by the 129, BTM and 813
probes). Probes were added to
prehybridized membranes at a
concentration of 2 X 106 cpm/ml
hybridization buffer. Each membrane
was hybridized at an optimized probe-
specific temperature for 16 to 18
hours. Membranes were washed 30
minutes at room temperature and 30
minutes at an optimized probe-specific
wash temperature in a 1 % SDS, 1X
SSC wash buffer. Air dried membranes
were placed in X-ray cassettes against
preflashed Kodak X-AR film, exposed at
-70°C (or room temperature for the
universal probe), and developed to
visualize hybridization signal.
Hybridization signal was quantified by
scanning densitometry coupled with a
peak integration software package.
Hybridization signals of environmental
samples were converted to ng RNA
based on the signals from hybridized
reference standards.
Quality Assurance
We initiated several quality assurance
protocols to evaluate factors that may
affect the amount of hybridization
signal produced from RNA extracted
from environmental samples. These
procedures included: (a) evaluating the
effect of varying degrees of
48
-------
glutaraldehyde denaturation on the
hybridization signal from environmental
RNA, (b) testing the hybridization of
SRB-specific probes to non SRB
negative controls containing RNA
equivalent to the amount of total RNA
present in environmental samples, and
(c) evaluating the presence of
components in environmental extracts
that alter the hybridization signal by
performing standard addition
experiments.
To evaluate the effect of
glutaraldehyde denaturation on the
•hybridization signal, environmental RNA
samples were denatured in 1.5, 3.0 or
4.5% glutaraldehyde and incubated in a
30°C heat block for intervals from 0 to
60 minutes. These samples were
applied to nylon membranes and
hybridized with the 1400 (universal)
probe and the 687 (Desulfovibrio spp.)
probe.
The amount of non-specific
hybridization of SRB-specific probes to
environmental RNA was determined by
preparing negative control RNA
extracted from grass, yellow perch
fillets, Clostridium perfringens, and
Pseudomonas aeruginosa. Extracts
from each species were combined in
equal amounts (based on A260) and
the mixture applied to membranes in
amounts equivalent to the amounts of
environmental RNA applied.
Matrix effects in environmental
extracts were examined by the addition
of reference standards to sediment and
water extract, or to extracts that had
been treated with RNase or with RNase
plus DNase. Samples were hybridized
with the universal probe and the slope
of standard curve in the standard
addition experiments was compared to
the slope obtained with reference
standards alone.
RESULTS AND DISCUSSION
Methods Development .and Quality
Assurance
We modified the extraction procedure
to allow quantitative calculation of the
total amount of RNA extracted from a
given volume of environmental sample.
Decreasing the amount of zirconium
beads or PVPP had no effect on the
recovery (based on A260) or purity
(based on A260/A280) of the extracted
RNA and allowed us to routinely
recover 29% of the aqueous extract in
sediment samples. This will allow
accurate calculation of the absolute
amount of SRB-specific 16S rRNA in
addition to the relative abundance
(percent of the total populations
accounted for by SRB). Absolute
amount of 16S rRNA is a better
parameter to describe community
structure since the total microbial
population varies significantly among
different habitats. One potential
limitation of the current extraction
procedure is that the effectiveness of
bead beating on quantitatively
disrupting all cells in the sample
population is unknown. This will affect
the accuracy of quantifying microbial
communities using 16S rRNA probes
and should be investigated in the
future.
Quality assurance protocols are
important in any quantitative assay and
should be developed and incorporated
into rRNA hybridization studies if the
results are to be used quantitatively or
semiquantitatively. We have begun to
49
-------
develop procedures to quantify factors
that may interfere with the hybridization
signal of SRB-specific probes. Since
glutaraldehyde denatures by covalently
bonding to the RNA molecule, the
degree of exposure to glutaraldehyde
could conceivably affect the resultant
hybridization signal. However, varying
the time and concentration of
glutaraldehyde used to denature RNA
showed no significant, effect on the
hybridization signal obtained from
sediment extract. Only the elimination
of glutaraldehyde resulted in a lower
hybridization signal.
Matrix interferences were observed in
standard addition experiments when
sediment extract was added to
reference standards. The addition of
10 and 50 ng (based on A260) of
sediment RNA to reference standards
decreased the slope of standard
addition curves (compared to curves
prepared with reference standards) by
about 24%. This indicates that the
sediment extract contains material that
interferes with the detection of 16S
rRNA, but the matrix effect can be
corrected for with traditional standard
addition procedures. When RNase
treated or RNase and DNase treated
sediment extract was added to
standards, the slope was not
significantly different from the slope of
the standards alone. This suggests that
the interfering matrix in the extract was
RNA.
We observed significant non-specific
hybridization with SRB-specific probes
when amounts of negative control RNA
comparable to amounts of
environmental sample were applied to
membranes. Since the hybridization
signal obtained from environmental
extracts is often low, correcting for non
specific hybridization by negative
controls is important if quantitative
estimates of SRB populations is to be
obtained.
SRB Community Structure in Wisconsin
Lakes
The relative abundance of the SRB
groups examined was similar in all lakes
and Desulfovibrio spp. comprised
greater than 66% of the total SRB
population in each lake. The total 16S
rRNA (based on the 1400 probe) was
positively correlated with lake pH
suggesting that total sediment biomass
decreases at low pH. Methylation was
active in all lake samples and was
negatively correlated with lake pH. The
relative abundance of all SRB groups
detected was negatively.correlated with
lake, pH although the correlation was
only significant (P < 0.10) with two of
the probes. This suggests that sulfate
reducers make up a larger portion of the
microbial population in low pH lakes,
but should be further evaluated with
larger sample sizes. Since SRB are
known to be active methylators of
mercury, this may account for the
inverse correlation observed between
sediment methylation and lake pH.
' Although the use of 16S rRNA-
targeted probes is in its infancy in
microbial ecology, our results indicate
that they show much promise in
defining microbial community structure
in natural habitats. Further work needs
to be done before truly quantitative
estimates of bacterial community
structure are available, but the use of
SRB-specific probes provide information
on the community structure of these
50
-------
key microorganisms in anaerobic
habitats that was previously
unobtainable. Thus, determination of
SRB community structure based on 16S
rRNA probes provides a useful risk
assessment tool to evaluate changes in
microbial communities.
FUTURE WORK
In the remainder of the current
funding period we will characterize the
SRB communities in the anoxic
hypolimnion of Little Rock lake.
Samples for this study were collected in
the summer of 1990 to follow the
development of SRB communities as
the anoxic hypolimnion built up, and in
the fall to monitor persistence of SRB
after the fall turnover when the water
column became aerobic. We will also
be sampling sediments from the
acidified and reference basin of Little
Rock Lake in a random design to allow
us to show any statistical differences in
SRB community structure and mercury
methylation between basins.
Future areas of work involving the
use of, 16S rRNA-targeted probes to
determine microbial community
structure should focus on further
methods improvement to allow more
quantitative measurements to be made,
and further evaluation of changes in
SRB community structure and activity
in response to environmental
perturbations.
PUBLICATIONS (abstract)
Winfrey, J., R. Devereux, and M. R.
Winfrey. 1991. Use of 16S rRNA-
targeted Probes to Correlate
Community , Structure of Sulfate-
reducing Bacteria with Mercury
Methylation in Freshwater Sediments.
American Society for Microbiology
Annual Meeting, Dallas, TX, May 5-9.
51
-------
APPLICATION OF DNA HYBRIDIZATION TO COMPARE
BULK DNA'S FROM DIFFERENT SOILS
L. Arlene Porteous and John L. Armstrong
U.S. Environmental Protection Agency
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
The distribution and persistence of
recombinant DNA in the terrestrial
environment has brought molecular
ecologists to develop techniques for
extracting DNA from soil samples. One
approach relies on recovery of intact
bacteria from soil before the DNA is
extracted. Another approach involves
direct- extraction of DNA without prior
removal of the cells from soil. Based on
these procedures, we developed a new
direct-extraction method tha| is small-
scale, rapid, and simple, yielding up to
20 jjg of DNA from a gram of soil.
Furthermore, we have applied DNA
hybridization of RFLP's and dot blots to
demonstrate differences between
DNA's extracted from different soils.
METHODS
Soil samples were collected near
Corvallis, Oregon: under an Oak tree, in
a sheep pen, from the rhizosphere
under grass. One gram of soil in a 50
ml plastic Oakridge tube was combined
with six ml of mixing buffer (0.5 M D-
sorbitol, 15% PEG 4000, 2% diethyl-
dithiocarbamic acid, 100 mM EDTA,
and 50 mM Tris-CI, pH 8.0) and vortex-
mixed for one minute. Then, 500 mg
of polyvinylpolypyrrolidone was added.
Lysis of bacteria was promoted with
100 jj\ lysozyme solution (50 mg/ml)
and fungal cell walls were degraded
with 120//I Novozym 234 solution (50
mg/ml). The sample was vortex-mixed
for 15 sec and incubated on ice for 1 -2
h. Next, 3.8 ml lysis buffer (4% SDS,
100 mM EDTA, proteinase K at 500
//g/ml, and 50 mM Tris-CI, pH 8.0)
were added. After mixing the contents,
the tube was returned to ice for 1-16 h.
The extract was then centrifuged at
5000 x g for 5 min. at 4°C and placed
in a sterile Oakridge tube on ice. To
recover additional DNA, the pellet was
resuspended in 3 ml wash buffer (100
mM EDTA, 50 mM Tris-CI, pH 8.0),
vortex-mixed, and centrifuged. The
previous .step with wash buffer was
repeated and both supernatant liquids
were combined. Then, 5 M potassium
acetate was added to a final
concentration of 0,5 M. After 1-2 h on
ice, the sample was centrifuged at 4°C
for 10 min. at 15,000 x g. The
supernatant liquid was mixed with two
volumes of 95% ethanol and
centrifuged at 10-15°C for 10 min. at
15,000 x g. The pellet was dried and
suspended in 1 ml buffer (0.01 M Tris-
CI, pH 8.0, and 0.001 M EDTA; TE).
Ten fj\ of each DNA extract was
electrophoresed for 1.5 h at 125 V in
0.7% agarose using a buffer consisting
of 0.04 M Tris-acetate (pH 8.0) and
0.001 EDTA (TAE), and stained for 30
min. in TAE containing 0.4//g ethidium
bromide per milliliter. Gels were
52
-------
photographed over UV light. DNA
extracts adjusted to a density of about
1.58 g/ml were purified by gradient
centrifugation using standard methods.
Extracts were further purified with
Geneclean Glassmilk (BIO 101, Inc.)
according to manufacturer's directions.
Restriction enzyme digests were
performed with approximately 0.5 jug
DNA and 2 fj\ of endonuclease in 20 //I
buffer. To determine if impurities in
DNA extracts inhibited restriction
enzyme activity, 0.2//g of pBR322 was
added to some reaction mixtures. After
incubating for 16-24 h at 37°C, more
enzyme was added for another hour.
DNA samples were electrophoresed,
transferred to a nylon membrane
(procedure by Schleicher and Schuell),
and hybridized with the following P-
labeled probes: nick-translated pBR322
DNA or random-primed labeled
chromosomal DNA from Pseudomonas
putida, Enterobacter cloacae and
Bacillus cereus, and the rRNA genes of
Escherichia coli (Bam H1 insert of
plasmid pKK3535). The final post-
hybridization washes for pBR322-
hybridized filters were performed at
56°C using 0.1X SSPE (18 mM NaCI,
1 mM NaPO4, pH 7.7, and 0.1 mM
EDTA) and 0.5% SDS. Southern
transfers of RFLP's hybridized with the
random-prime labeled probes were
washed in 2X SSC (0.3 M NaCI and 30
mM sodium citrate, pH 7) and 0.5%
SDS at 65°C. Autoradiograms were
prepared by standard methods.
Dot blots were prepared on nylon
membranes (Schleicher and Schuell)
according to the manufacturer's
instructions. Concentrations of serial
dilutions applied through a filtration
manifold (Bethesda Research Labs.)
included: 500 to 3.9 ng/dot for soil
DNA's; '10 to 0.08 ng/dot for
chromosomal DNA's of P. putida, E.
cloacae, and B. cereus; 1 to 0.008
ng/dot for salmon DNA and E. coif
rRNA genes. Membranes were
hybridized according to methods of
Boehringer Mannheim with random-
primed, digoxigenin labeled DNA probes
(P. putida, E. cloacae, B. cereus, and E
coli rRNA genes). Final post-
hybridization washes were performed
under highly stringent conditions using
0.1X SSC and 0.5% SDS at 68°C.
Autoradiograms were produced using
the chemiluminescent substrate, Lumi-
Phos 530 (Boehringer Mannheim).
RESULTS AND DISCUSSION
Soils from different sources were
used to demonstrate the effectiveness
of the DNA extraction method. The
method reported here yielded up to 20
fjg DNA from one gram of various
agricultural soils. In some cases (e.g.,
from the soil of a sheep pen), we
recovered up to 35 JJQ DNA in one
gram. Determinations were measured
by spectrophotometric absorbance
(wavelength, 260 nm). Accurate
readings depended on the complete
elimination of Nal and glassmilk
following the Geneclean purification
step. Due to the possible selectivity of
the extraction process, all of the DNAs
from a sample may not be retrieved.
For example, some species of DNA may
be firmly bound to soil components.
The degree of purity and restrict-
ability that resulted from the use of this
method was measured on agarose gels.
High molecular weight DNA, 20-25 kb
in size, was routinely observed
53
-------
following the extraction procedure.
This DNA was digested by restriction
enzymes so that the entire DNA sample
was fragmented into smaller pieces.
This observation suggested that the
DNA lacked residual, soil constituents
that inhibit endonuclease activity. This
was further demonstrated by the
appearance of completely restricted
pBR322 DNA that was added to digests
containing purified soil DNAs. The
resulting one band at 4.4 kb clearly
demonstrated complete restriction of
the plasmid to the linear form.
Southern transfers of these agarose
gels were probed with pBR322 in the
more sensitive method of DNA
hybridization. The autoradiograms
verified complete digestion with Eco Rl,
Sal I, and Bam HI (Bethesda Research
Labs.). These findings emphasized the
reliability of the method as an effective
means to purify DNA from different
types of soils.
Bulk DNA extracts from soils were
examined by RFLP analysis to determine
differences in the observed fragmented
band patterns between samples.
Distinguishable, isolated bands were
not detected in any of the soil extract
samples analyzed. The completely
fragmented DNA appeared as a wide
distribution of molecules ranging from
less than 0.125 kb to approximately 23
kb. We suggest that the lack of
discernible bands is indicative of the
presence of a wide variety of DNAs.
The work of other researchers confirms
the diversity of microorganisms found
in the soil and the resulting high
heterogeneity of their DNAs.
Discernible restriction fragments would
be seen if the DNA extract was
relatively homogeneous, e.g., a sample
containing DNA from a predominant
microorganism in large numbers.
Hybridization of electrophoresed soil
DNAs was observed when DNA
extracts were digested with Sail and
Notl (Promega) and probed with the E.
coli rRNA genes. In contrast, little or
no hybridization was observed when
the chromosomal DNA's of P. put/da, E.
cloacae, and B. cereus were used as
probes. As with visual analysis of the
RFLP's, discernable bands were not
seen following DNA hybridization with
any of the bulk DNA extracts examined.
Due to the lack of specific bands, dot
blots were utilized to determine
whether , quantitative differences
between samples were evident. Using
post-hybridization washes under
conditions of high stringency,
differences were detected. Different
samples demonstrated different degrees
of homology to a probe encoding the E.
coli rRNA genes. The extract from the
soil of the rhizosphere of grass
contained 8 pg of hybridizable DNA and
that from the sheep pen had 62.5 pg,
i.e., 0.0016% and 0.0125% of the
total DNA, respectively. When using
the P. put/da chromosomal DNA probe,
differences were also detected between
the two DNA extracts. Related DNA
recovered from the rhizosphere of grass
and the soil from a sheep pen
represented 0.06% and 0.25% of the
total DNA. Hybrids were not detected
when E. cloacae and B. cereus
chromosomal DNA's were used as
probes. Based on the sensitivity of the
method, we estimated that both
extracts contained <0.06% of DNA
related to E. c/oacaeDNA and <0.12%
of DNA related to B. cereus DNA, if
present.
54
-------
In summary, we describe a bulk soil
DNA extraction method that is rapid
and small-scale. The DNA extracts are
pure enough to be restricted and used
in hybridization experiments. Coupled
with the use of dot blot techniques,
specific DNA populations in soil DNA
extracts can be compared and
quantified.
FUTURE WORK
Plans for this project include: a) use
various probes to identify DNAs from
various "functional groups" of
microorganisms, b) enhance sensitivity
of detection with PCR methods, c)
assess efficiency of method for re-
covery of fungal, protozoan, plant
rhizosphere, and "microfauna" DNAs,
d) retrieve recombinant DNA after
introducing a specific GEM into a
terrestrial microcosm, e) develop
methods to recover bulk DNA from leaf
litter of tilled soil containing transgenic
plants, and f) demonstrate effects of a
GEM on population of DNA molecules
indigenous to soil.
PUBLICATION
Porteous, L. A., and J. L. Armstrong.
1991. Recovery of bulk DNA from soil
using a rapid, small-scale extraction
method. Current Microbiol. (accepted
for publication).
55
-------
COMPARISON OF xy/E GENE ACTIVITY
IN DIFFERENT MOLECULAR CONSTRUCTS
H.M. Abebe1, R.J. King1, S.E. Lindow2, K.A. Short3
and R.J. Seidier3
ManTech Environmental Technology Inc.1,
Environmental Research Laboratory
Corvallis, Oregon
Department of Plant Pathology,
University of California,
Berkeley, California2
U.S.Environmental Protection Agency
Environmental Research Laboratory
Corvallis, Oregon3
INTRODUCTION
The activity of the xy/E gene product,
2,3-catechol dioxygenase, is a useful
colorimetric tool for autecological
investigations. The yellow color (2-
hydroxymuconic semialdehyde)
produced due to the enzymatic
cleavage of catechol, a colorless
substrate, has been exploited to
enumerate target microorganisms
released into soil, lake water or sprayed
onto plant leaves. Thus far however,
the application of the xy/E marker gene
has been confined to plating and
spectrophotometric assays of sonicated
filter concentrated cells.
The objectives of the current study
were: a) to assess the stability and level
of expression of a plasmid-borne versus
chromosomally inserted xy/E marker
gene, b) to examine the effect of cell
density and growth phase on catechol
2,3-dioxygenase activity, c) to
determine the efficacy of xy/E gene
expression for direct determination of
cell numbers via a spectrophotometric
analysis of 2-hydroxymuconic
semialdehyde, and d) to establish the
sensitivity of the technique in detecting
target microbes by HPLC analysis.
MATERIALS AND METHODS
Cultures and media
Luria-Bertani (LB) agar was used for
plate counts. Liquid cultures Were
grown in LB broth. All cultures were
incubated at 30°C. Broth cultures were
maintained on a rotary shaker at 200
rpm. Appropriate antibiotics were added
when selective media were needed.
Construction of xv/E Gene Marked
Strains .:•
Most molecular biological techniques
were performed according to standard
procedures. A typical approach used
was similar to that for inserting the xy/E
gene into the iceC structural gene.
Plasmid plCE1.2, which contains the
iceC gene, was subjected to partial
digestion with sa/1 (Boehringer
Mannheim), followed by dephos-
phorylation with .calf intestinal
phosphatase (Boehringer Mannheim).
56
-------
Plasmid pR01733 was restricted with
xho\ (Boehringer Mannheim) and
electrophoresed in low melting
temperature agarose (International
Biotechnologies, Inc.). The 2.3 kb band
containing the xy/E structural gene was
excised and extracted. The 2,3 kb xho\
fragment from pR01733 was ligated
with the partially digested PICE1.2
DNA, and transformed into E. C0//DH5.
Colonies resistant to 50 //g ml-1
ampicillin were sprayed with 0.1M
catechol, and those colonies which
turned yellow were isolated. One of the.
x//E+ colonies that exhibited an Ice-
phenotype was selected and the
recovered plasmid was designated
pSEL2.
Marker Exchange
Selection for integration of non-
replicating but mobilizable plasmid.
Plasmid pSEL2-4 was mobilized into P.
syringae Cit7 which.contains the iceC
gene using pRK.2013 as , a -helper
plasmid. Colonies which fluoresced
under UV were selected, and tested for
xy/E activity by spraying with catechol*
Single xy/E+ Flu+ (colony growing on
King's B agar plates fluoresce upon
exposure to UV light) colonies were
restreaked onto KB amended with 100
JJQ ml"1,, rifampicin, 30 , VQ ml"1
kanamycjn,.and 100#g ml"1 ampicillin,
and tested for ice nucleation activity to
confirm the lce+ phenotype and thus
ensure that plasmid integration had
occurred but a second recombination
event had 'not. Strains designated mer
3, 12, 15, and 17 were used for
subsequent marker exchange steps.
Identification of Second Recombination
Events
Four different xy/E+ Flu+ lce +
colonies were inoculated into KB
amended with 100/yg ml"1 rifampicinto
107 cells ml"1, and cultured at 28°C for
about 12 hours to a cell density of over
10? cells-ml"1. Cultures were diluted
1:50 into fresh medium, and regrown
to 109 cells ml"1. This cycle of dilution
and regrowth was repeated 10 times.
Dilution series of the final culture were
plated onto KB amended with 100 //g
ml"1 rifampicin and grown 2 days at
28°C.
Approximately 104 colonies were
replica-plated onto KB containing 100
fjQ ml"1 rifampicin, 30 jjg ml"1
kanamycin, and 100//g ml"1 ampicillin.
About 1% of these colonies did not
grow in the presence of kanamycin and
ampicillin and were .thus hypothesized
to have undergone a second
recombination event in iceC distal
relative to xy/E from the first
recombinational event. Of the
kanamycin- and/or ampicillin-sensitive
colonies, about 9% were also xy/E +
and Ice-, suggesting that excision for
vector-derived sequences and
recombinational placement of the native
iceC 'gene with the xy/E-iceC fusion
construct had occurred. Five such
strains, designated hap 18, 45, 49, 69,
and 72 were obtained.
Stability and Expression of Genetic
Markers
Antibiotic and xy/E gene expression
and stability were assayed by spread
57
-------
plating onto selective and nonseiective
media. All cultures except P. syringae
Cit7 #69 colonies, were assayed for
expression of catechol 2,3-dioxygenase
activity after 48 hours of incubations at
30°C. P. syringae Cit7 #69 was
assayed after 72 hours of incubation.
White colonies, on selective and non-
selective agar plates, were enumerated
before they were sprayed with a 1 %
aqueous catechol solution for counting
xy/E gene harboring constructs.
Determination of Bacterial Growth
Cell density and bacterial growth
phases were determined by plate
counts and optical density
measurements. A 300 ml LB broth
amended with an appropriate
antibiotic(s) was inoculated, from an
over night culture, to a final cell
concentration of less than 106 cells/ml
and incubated at 30°C. Samples were
withdrawn for plating and optical
density measurements (spectronic 21
Milton Roy Co., Rochester, NY set to a
wave length of 600nm) at intervals 0,
15, 24, 39, 48, 63, 72 and 87 hours.
Spectrophotometric Enzyme Assays
Cells were washed twice with 10 ml
20 mM phosphate buffer, pH 7.2. The
pellet from the second wash was
resuspended in 100 mM phosphate
buffer, pH 7.5'. One milliliter of this cell
suspension was then treated with
catechol solution to a final
concentration of 0.05% and incubated
for 30 to 60 minutes at room
temperature. Catechol 2,3-dioxygenase
activity was then determined
spectrophotometrically (Beckman DU-
BS) by reading the increase in
absorbance at 375 nm. Where catechol
2,3-dioxygenase activity resulted in an
intense yellow ' coloration (high 2-
hydroxymuconic semialdehyde pro-
duction), serial dilutions of the product
was made to accommodate ranges of
Spectrophotometric specificity and
sensitivity.
HPLC Enzyme Assays
For HPLC analysis, cells were
washed twice with 20 mM phosphate
buffer, pH 7.2 and resuspended in the
same buffer. Final catechol concen-
tration and incubation time was 0.05%
and 30 to 60 minutes, respectively.
Following incubation however, the
samples were pelleted at 13000 rpm
for 10 minutes at 4°C using a
microfuge (Beckman Microfuge 11).
The supernatant was transferred into
LC automatic sampler 2 ml vials with
11 mm crimp caps (Hewlett-Packard
Co., Avondale, PA). The samples were
analyzed with an HP-1090 high
performance liquid chromatograph
(Hewlett-Packard Co., Avondale, PA)
coupled to a diode-array UV detector
and monitored at 375 and 277 nm
(lambda max for 2- hydroxymuconic
semialdehyde and catechol respect-
ively). Samples were separated by an
isocratic mobile phase (35% water with
0.1% acetic acid, 65% methanoDon an
HP C18 reverse phase micro-bore
column with C8 guard column at a flow
rate of 0.40 ml min"1. External catechol
reference standards (NBS certified)
were incorporated to bracket samples'
concentrationsand analyzed in triplicate
every tenth injection.
58
-------
RESULTS AND DISCUSSION
Marker instability, especially with
markers carried on a plasmid, is a major
problem in ecological studies since
under nonselective environments where
the trait borne on the plasmid need not
be expressed, the antibiotic-resistant
populations have been observed to
disappear as they are out competed by
the native populations. Because of the
importance of genetic markers , as
selectable traits for tracking/detecting
genetically engineered microorganisms
(GEMs), measuring their stable
expression or maintenance in the host
is a prerequisite for studies prior to field
releases.
We present here results of our study
in which a comparison of the
expression and stability of genetic
markers, and assay systems is made for
bacteria proposed for release into a
microcosm. Results of our study
suggest that the xy/E marker gene is
stably expressed in some strains but
not in others. Expression of xy/E gene
was or approached 100% in all colonies
of the seven strains when grown on
selective media. Except for E. coli
W311DHRalpha (PSEL2-4) and P.
aeruginosaPA4-032 (pRO1940::xy/E) all
strains which carried the xy/E gene on
a plasmid proved highly unstable with
respect to the simultaneous expression
of the xy/E and antibiotic resistant
phenotypes. Such instability could be a
function of the frequency of
segregation of plasmid-free cells and
the difference in growth rate between
plasmid-free and plasmid-containing
cells. In contrast, chromosomally
inserted marker genes, such as the xy/E
gene in the haploid (P. syringae
Cit7#69) and merodiploid (P. syringae
Cit7#17) colonies, consistently
maintained both the xy/E and
associated antibiotic marker gene
phenotype, on various selective and
nonselective growth media.
The activity of the 2,3-catechol
dioxygenase depends on the cell
density and also on the growth phase.
The activity of the enzyme was optimal
when the cell concentration was greater
than 108 CFU/ml and the cultures were
in the stationary phase. It was also
observed that 2,3-catecholdioxygenase
activity was strain specific and its
activity decreased in the late stationary
phase of the culture. This loss of
activity may be a function of the
depletion of iron which is known to
affect the specific activity of the
enzyme.
The spectrophotometric limit of
detection observed in this study
indicates that a sample needs to
contain a population density greater
than 107 CFU/ml of the marked bacteria
for direct enumeration without
cultivation on plates. This condition is
unlikely to occur in soil or in lake water.
Even under laboratory conditions and
with the use of a strain with vigorous
catabolic activity of catechol such as in
the merodiploid, the limit of detection
of the assay system could not be
improved. Additional limitation noted
was the instability of the enzyme
product, 2-hydroxymuconic semi-
aldehyde. Where large numbers of
samples are involved, the short half life
characteristic of the product limits its
application.
Although catechol dioxygenase
activity could not be detected
spectrophotometrically at whole cell
59
-------
concentrations below 107 CFU/ml,
HPLC analysis revealed that activity
was detectable with cell density as low
as 103 CFU/ml. The sensitivity attained
via HPLC analysis is comparable to the
spectrophotometric detection limit (103
cells ml"1 } observed with cell extracts
(supernatant of sonicated cells) of filter
concentrated lake water samples
reported by others.
Of the seven strains studied, five
carried the xy/E gene on a plasmid. Of
the five assayed for expression and
stability of the xy/E gene, only one
consistently expressed both xy/E and
antibiotic resistance phenotype. All
constructs bearing chromosomally-
inserted xy/E gene and antibiotic
markers expressed catechol 2,3-
dioxygenase activity and resistance to
antibiotics. The merodiploid (P. syringae
Cit7#17) showed superior catabolic
activity of the substrate and produced
the most 2-hydroxymuconic semi-
aldehyde. The apparent hyper-
expression of xy/E gene in this strain
facilitated HPLC mediated detection of
very low concentrations of marked
bacteria. Although our study suggested
that the xy/E-gene tagger system has
limitations for spectral assay of whole
cells (107 cells/ml), it also demon-
strated that the xy/E gene assay is very
useful for plate assays and is highly
sensitive (103 cells/ml) for detection of
merodiploid xy/E activity by HPLC
analysis. Our study also confirmed the
viability of the assay system as a stable
and sensitive analytical tool for the
study of anthropogenic strains. The
high degree of sensitivity afforded by
HPLC analysis presents an opportunity
useful in the autecological investigation
of released microbes.
FUTURE WORK
Future work will include a) increasing
the sensitivity of the assay system by
concentrating and stabilizing the
product, b) determining the factors
responsible for the differential
expression of xy/E gene in the haploid
versus the merodiploid, and c) testing
the assay system in a microcosm
setting using plant, lake water or soil.
PUBLICATIONS
This study was initiated this year and
there are no publications at this time.
60
-------
POTPOURRI OF BIOAEROSOL RESEARCH AT ERL-CORVALLIS IN 1990
Bruce Lighthart1, Brenda T. Shaffer2, and Balkumar Marthi2
U.S.Environmental Protection Agency1, and
ManTech Environmental Technology Inc.2,
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
Viable microbial bioaerosols
generated by many human (and natural)
activities are dispersed by the
atmosphere to unwanted downwind
locations. In order to estimate the near
source downwind risk associated with
the deposition of those viable bio-
aerosols, the following areas of
research are being pursued: increasing
bioaerosol sampling efficiency,
developing better tools to evaluate
bioaerosol generation potential, and
developing predictivesimulation models
of bioaerosol dispersal and deposition.
The models act as integrators of new
information generated with better tools
and sampler efficiencies.
The remainder of this report will
describe the progress in the past year
in: (1) increased bioaerosol sampling
efficiency, (2) tool development of an
automated dynamic aerosol torroid
(auto-DAT), (3) development of an
Environmental Bioaerosol Research
Chamber (EBARC), and (4) bioaerosol
simulation models. In the following,
bioaerosol specifically means bacterial
bioaerosols.
PROGRESS DESCRIPTIONS
Bioaerosol Sampling
The atmospheric environment is more
or less inhospitable for maintenance of
viability in bioaerosols. However, at
least some organisms in bioaerosols are
only partly damaged while airborne and
will not grow when cultured, although
they are viable. In an effort to
resuscitate some bioaerosols, betaine
has been added to culture media
increasing sampling efficiency in terms
of viable counts by up to 30% (Marthi
and Lighthart, 1990)
Catalase is another compound that
has been used to resuscitate stressed
bacteria. When we add catalase to our
bioaerosot recovery media for both pure
cultures and wastewater treatment
plant bioaerosols, sampling efficiency
increased up to 113% (Table 1).
Table 1. Resuscitation of aerosolized Pseudqmonas svringae or of wastewater
treatment plant (WWTRP) bacteria in 1000 units of catalase containing collection fluid
(as percent of 0 min. non-catalase control).
Catalase Contact
Time (minutes)
A. WWTP
0
30
60
B, Ps. syringae
0
30
60
Catalase
4
0.0
98.4
112.9
0.0
110.9
112.1
-
0.0
10.1
8.2
0,0
-8.5
•28.0
-------
A speculative explanation of the
catalase resuscitation mechanism is as
follows:
1. Many bacteria produce toxic
hydrogen peroxide.
2. Some bacteria produce
membrane associated catalase to
dissipate hydrogen peroxide.
3. On aerosolization, the bacterial
membrane is disrupted affecting
catalase function, peroxide ac-
cumulates, and nonculturability
occurs.
4. Exogenous catalase penetrates
damaged cells, catalyzes
dissipation of toxic hydrogen
peroxide, and subsequent
culturability occurs.
5. Upon addition of betaine (an
osmoprotectant) to aerosol
suspension fluid, the catalase
effect is reduced, indicating that
the betaine may negate
membrane disruption and
consequent need for exogenous
catalase.
Tool Development: Auto-DAT
Large, slowly rotating airtight drums
have been used to evaluate the survival
of airborne microbes (Figure 1). .In.
order to increase the precision of the
drum system, we have automated it
using a microcomputer. The auto-DAT
has just been used for the first time
with some of the resuscitation experi-
ments above. A protocol describing the
construction of the auto-DAT audits use
to evaluate the survival of bioaerosol is
forthcoming.
Drum
Bioaerosol
Input/Output
Figure 1. An Automated Dynamic Aerosol Torroid (Auto-DAT).
62
-------
Tool Development: EBARC
In an effort to determine some of the
conditions that would lead to bioaerosol
entrainment from plant surfaces
sprayed with microbial suspensions, the
EBARC was designed to give a set of
plants from one to many puffs of air. It
is thought that the accelerated force
(drag and thrust) occurring at the
leading edge of a puff wind field would
be the most effective mechanism to
force microbial release from plant
surfaces. Figure 2 (next page) is a
diagram of the EBARC showing the
rotating "lazy susan" with plants
passing through influent air streams
emitted from the influent air manifold.
Samples downwind from the ports
collect the entrained bioaerosol.
To date, we have used the EBARC to
evaluate the variation between
Andersen samplers (Table 2) and the
particle size distribution of bacteria size
fluorescent spheres sprayed onto plants
and subsequently blown off (Figure 3).
Preliminary data in Table 2 shows
that there are significant differences
between Andersen sampler stages.
Figure 3 (page 65) shows that many
bacteria-sized particles are released
from plants on larger rafting particles.
This is also found in nature, where
approximately 60% of the bacteria are
particles of _>. 7 //m aerodynamic
diameter.
Table 2. Comparison of 8 randomly selected Andersen samplers used in the EBARC to evaluate
BACILLUS SUBTILIS var. NIGER spores blown off oat leaves.
Andersen
Sampler
1*
2
3
4
5
6
7
8
Average
(CV)
6-
(0.65 Mm)
1.0
2.6
9.9
5.1
3.2
1.1
9.3
1.0
4.2+1.0
(23.8)
Stage Number-(Particle
5- 4-
(I'.Owm) (2.1 Mm)
16.2
14.3
5.0
12.6
13.1
13.8
10.8
15.3
12.6 + 0.9
(7.1)
0.4
11.8
10,3
7.4
10.7
11.0
10,4
9,7
9.0 + 0.5
(5.5)
Size), Percent
3-
(3.3 Mm) (4
20.7
10.8
12.3
8.8
10.1
9.7
13.7
9.5
12.0 + 0.6 8.
(5,0)
Total
2-
.7 Mm)
6.8
8.5
6.7
9.4
5.6
7.1
11.8
8.3
0 + 0.4
(5.0)
1-
(7.0 Mm)
54.8
51.9
55.7
56.7
57.3
57.3
44.0
56.2
54.3+1.5
(2.8)
Total Colonies
(Log 10)
2.97 + 0.04
2.78 + 0.15
2,84 + 0.14
3.04 + 0.05
2,96 + 0.04
3.01 +0,04
2.96 + 0.04
2.97+0.05
2.95 + 0.03
(1.0)
* 10 Replicates per sampler
63
-------
Roof
Roof
in
Influent
air i
Sampler Sampler
Stationary
, Rotating T platform
Side view
Figure 2. Environmental Bio-Aerosol Research Chamber (EBARG).
64
-------
ea
I
•s
^
100
90 -
80 -
70
60
50
40
30
20
10
1 pro spheres
6
Particle Size (pin)
Figure 3. Comparison of the particle size distribution of Bacillus subtilis var. niger spores and 1//m
fluorescent microspheres blown off oat plants after multiple wind puff treatments using the EBARC.
General Survival
Simulation Models:
Model
Using DAT determined death rate
values from ca. 70 literature cited cases
involving 17 bacterial species in which
gram reaction, relative humidity,
temperature, and aerosol age were the
experimental variables (Figure 4b), a
regression model was fit having R2-
value 0.94 (Figure 4a)(Lighthart, ,B.,
1989, Aerobiol. 5:138-144). Death
Rate Constant (log^g) = Constant -
Log10 (Aerosol Age) "+ temperature -
Gram Reaction * RH. A; less
demanding model with an R2 if 0.90
was fit to: Death Rate Constant (Iog10)
= Constant - Log10 (Aerosol Age)
These models may be inserted into
the dispersion model to give realistic
survival properties of the bioaerosol as
a function of temperature, relative
humidity, gram reaction over time in the
atmosphere.
Simulation Models: Trajectory Model
In an effort to obtain a more realistic
characterization how a polydispersed
bioaerosol might wander in the
atmosphere, the trajectory of various
size droplets containing various
numbers of microorganisms in
atmospheres of different relative
humidity, temperatures, and wind
speeds was, prepared (Lighthart et al.,
in press). The results of this effort are
shown in Figure 5a (simulated) where it
is seen that droplets of .<. 80 fj
diameter in a 50% relative humidity and
20 °C atmosphere, evaporate to the
bacterial residue of about 1 jjm before
they can fall to the ground and
therefore remain airborne becoming part
of the aeroplankton. Those greater
then 80 /ym fall to the ground.
Observations confirm the model
(Figures 5b,c observed).
65
-------
Figure 4. a) Response surfaces for aerosol age, evaporation function and bacterial
death rate constants for Gram positive (lower surface) and Gram negative (upper
surface) bacteria using the general death rate equation model. The two surfaces
result to be not parallel; b) enlarged picture of insert shown in fig. 4A.
66
-------
Simulated
if. it, it, n, H > M in '.si
Wl.ll
11(1.«
Ml',ii
MO.II.
I I
Sourc* PROXIMAL
Observed
iM («*>
8ourc» DISTAL
Figure 5. Trajectory model
67
-------
Simulation Models: Deposition Model
A bioaerosol dispersion model has
been developed and crudely compared
with observations made at the Tulelake,
CA, ice-minus bacterium spray-out
(Lighthart and Kim, 1989). recently, a
controlled spray of bacterial spores has
been carried out by the U.S. Army at
Dugway Proving Ground to provide
data to quantitatively evaluate reliability
of the bioaerosol dispersion model with
a good set of observations. Presently,
the model has been run using the
observed meteorological data (Figure
6); the statistical comparison between
simulated and observed will be
completed in the near future.
FUTURE RESEARCH
* Evaluate the effects of genetic
construct on airborne microbial
survival.
* Evaluate the airborne survival of
insect pathogens including MPCAs.
* Prepare a protocol to evaluate the
survival of airborne microorganisms.
* Modify the bioaerosol dispersion
model for use in a microcomputer.
REFERENCES
Marthi, B., and B. Lighthart. 1990.
Effects of betaine on enumeration of
airborne bacteria. Appl. Environ.
Microbiol. 56(5):1286-1289,
Lighthart, B. 1989. A statistical model
of laboratory death rate measurements
for airborne bacteria. Aerobiol. 5:138-
144.
Lighthart, B., B.T. Shaffer, B. Marthi,
and L. Ganio. In press. Trajectory of
aerosol droplets from a sprayed
bacterial suspension. Appl. Environ.
Microbiol. 57(4).
Lighthart, B., and J. Kim. 1989.
Simulation of airborne microbial droplet
transport. Appl. Environ. Microbiol.
55(9):2349-2355.
68
-------
M
X
M
!
k
/
A
1
^
X
D
X
2
X
T
X
c
<\
\* *3
^^
X
B3
X
C3
X
D3
. X
£3
X
n
X
83
>
* ^
X
A2
X
B2
X
C2
X
02
X
E2
X
M
X
02
H '
X
X
Al
• X
Bl
X
Cl
X
01
X
Bl
X
Tl
X
01
4
15m
9 OB
Om
X - Andersen Sampler, X - Aerodynamic Particle Sizer and
Andersen Sampler, H - High Volume Sampler, M - Meteorological
Tower, D - Aerosol Disseminator
Figure 6. Sampler Arrangement
69
-------
MEASURING ENTRAPMENT OF BACTERIA FROM THE PHYLLOSPHERE
Michael Walter1, Valerie Fieland1, Lisa Ganio1, and Ray Seidler2
ManTech Environmental Technology1
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
Monitoring the dispersal patterns of
genetically engineered microorganisms
(GEMs) from a release site is an
important aspect of understanding what
impact GEMs may have on the
environment. The possibility exists that
after application, GEMs could become
resuspended (entrained) from soil or
leaf surfaces and transported
downwind to colonize other locations.
Plant canopies as well as soil have
been shown to be sources of entrained
bacteria. The number of bacteria
entrained is influenced by the season of
the year, moisture content of the soil,
and type of plant canopy. In addition,
activities such as harvesting have been
known to influence the number of
entrained bacteria detected.
Measurements taken on days 1,2 and
4 following the application of
recombinant strains of Pseudomonas
indicated that low numbers of the GEM
were being entrained into the
atmosphere. The number of entrained
bacteria decreased with time after
application. In addition, entrained
bacteria were detected primarily on
gravity settling plates (GSPs), which are
150 mm petri dishes containing
selective agar. Mechanical air samplers
that proved very sensitive during the
actual applications of the GEM either
failed to detect entrained GEMs or
detected very few of them compared to
the GSPs.
It has been demonstrated that under
greenhouse conditions, bacteria can be
injured during, aerosolization and that
this injury can adversely effect their
subsequent survival. Entrainment of
bacteria results in their resuspension
into the atmosphere, therefore it is
possible that this process could cause
injury or death to GEMs after release, or
render them incapable of surviving in
competition with indigenous
microorganisms.
The objectives of this study were to:
A) develop a method of artificially
inducing bacterial entrainment from the
phyllosphere, B) compare sampling
efficiencies of GSPs, all glass impingers
(AGIs) and Andersen samplers to
determine which method of sampling
would be most sensitive in detecting
entrained bacteria, C) compare the
magnitude of entrainmentfrom different
species of host plants, and, D)
determine if the viability of bacteria is
affected during entrainment.
METHODS
Media and Cultures
All experiments were conducted
using a nonrecombinant strain of
Pseudomonas syringae TLP2, resistant
to 100 //g/ml of rifampicin (RIF).
70
-------
Several studies were conducted using
spores of Bacillus subtilis var. niger. P.
syringae were cultured and enumerated
using 100//g/ml of RIF in Luria-Bertani
broth and agar (LB-RIF) respectively. B.
subtilis spores were obtained from
spread plates on casitone agar (CAS)
incubated at 30°C for seven days.
Spores were recovered by scraping 10
plates with 10 ml of sterile distilled
water and suspending them in 100 ml
of distilled water.
Plant Application and Sampling
P. syringae were applied to three-
week-old beans (Phaseolus vulgaris
humilus cv. Bush Blue Lake) and three-
week-old oats (Avena sativa L. cv.
Cayuse) planted in 61 X 30 X 13 cm.
pulp trays grown in a greenhouse.
Plants were sprayed until run-off with
twice washed P. syringae following
incubation for 18 hours at 30°C. A
population of 1 x 108/ml was applied
using a CO2 backpack sprayer
pressurized to 36 psi and equipped with
a Tee-Jet nozzle #8004-SS. Plants
were allowed to dry a minimum of two
hours before sampling was initiated.
Bacterial populations were determined
by randomly sampling three bean leaves
or five oat stalks. Leaves were diluted
with 20 ml of phosphate buffer and
blended for one min. in a stomacher
blender. Samples were serially diluted
and plated on selective media and
incubated for 48 hours at 30°C.
Counts were standardized on a per
gram basis. Spore populations were
determined by heating at 80°C for 20
min. to heat-shock spores into
germination and kill undesired bacteria.
Samples were filtered using 0.45 /vm
filters and filters placed on CAS agar
plates.
Aerosol Samplers
Six-stage Andersen samplers
containing six agar plates filled with 20
ml of LB RIF were operated at 28.3
L/min. The AGIs were loaded with 20
ml of sterile 10 mM phosphate buffer
and run at 12.5 L/min. Following
sampling, the buffer was filtered
through 37 mm diameter, 0.45 //m
cellulosic filters, then transferred to the
appropriate agar. GSPs used in these
experiments consisted of 150 mm petri
dishes filled with 75 ml of LB-RIF agar.
Plates were uncovered immediately
prior to sampling and covered
immediately after sampling events. All
plates were incubated for 48 hours at
30°C and colonies counted.
Bacterial Entrainment
Bacterial entrainment was induced by
placing a 20 inch electric fan 0.5 m
from sprayed plants. Air flow through
the plants was measured at 4 m/sec
and entrainment was measured for 10
min. In all experiments except sampler
efficiency, plants were mounted on two
9 cm2 wood blocks. Experiments were
conducted in a 10 X 30 m greenhouse.
Sampler Efficiency
A two-level 122 X 183 cm. platform
was constructed to adjust the intakes
of the three samplers to the same
height. The upper level of the platform
consisted of three parallel rows of three
circular holes located 0.25, 0.6, and
1.0 m from sprayed plants. The AGI
71
-------
and Andersen samplers were placed on
the lower level of the platform with
intakes protruding through the holes
level with the height of the gravity
settling plates. Entrainment was
measured at 2, 24, and 120 hours after
application of the P. syringae using a
separate set of plants for each time
point. Results were compared using
Iog10 transformations of the number of
CPU observed on a per sampler basis.
Factors Affecting Bacterial Entrainment
Eight AGIs or two Andersen samplers
were placed at 0.25, 1.0, 2.0, and 3.0
m from inoculated plants. AGIs were
mounted in two groups of 4, at 26 cm.
and 53 cm. above the ground, each
group of four separated by 35 cm.
Andersen samplers were placed 35 cm.
apart on the soil surface at each of the
four distances.
Effect of Host Plant
The effect of host plant was
measured by comparing the magnitude
of bacterial entrainment from beans
with that of oats. Experiments were
conducted to measure total bacterial
populations being entrained from each
plant using AGIs, and difference in
particle size distributions were
measured with Andersen samplers.
Dilution vs Die-off
To determine if decreases in
entrained bacterial counts were due to
dilution or loss of viability, the number
of entrained vegetative bacteria were
compared with the number of entrained
spores of B. subtilis. Decreases in the
spore populations were due to dilution,
while decreases in the vegetative
bacterial populations were due to loss
of viability.
Time After Application
The effect of time after application
was measured by sampling bean plants
sprayed with P. syringae at 2, 24, and
120 hours after application using only
AGIs. In these experiments, all four
distances were sampled simultaneously,
since the aim of the experiment was to
detect the total number being entrained.
The number of bacteria entrained were
compared over time and related to the
population of bacteria found on the leaf
surface.
RESULTS AND DISCUSSION
Sampler Efficiency
In general Andersen samplers, AGIs
and GSPs had the same pattern of
response throughout the experiments.
In all cases, there were significant
effects due to both time after
application and distance from the
sprayed plants. On a per volume basis
(CFU/m3), there was no difference
between the number of entrained
bacteria detected in AGI and Andersen
samplers. AGIs exhibited the most
variable data over both distance and
time. However, variability tended to
increase for all three samplers as the
distance from the sprayed plants and
time after inoculation increased. No
entrained bacteria were detected on
GSPs beyond 1 m from plants.
72
-------
Factors Affecting Bacterial Entrainment
Over a distance of 3 m from sprayed
plants, both AGI and Andersen
samplers exhibited similar dispersal
patterns of entrained bacteria. Both
samplers detected substantial numbers
of bacteria 3 m from plants.
Significantly higher numbers of
entrained bacteria were detected 0.25
m from plants than at the other three
sampling distances. There was no
significant difference in the number of
entrained bacteria detected 1 to 3 m
from plants.
Graphical examination of the data
collected from AGI and Andersen
sampler? demonstrated that the
distributions of data were not the same.
This difference indicates that the two
populations being sampled are not
equivalent, and rendered it impossible
to make a direct comparison between
sampler types. It appears that the AGI,
which is a liquid based sampler, is
sampling actual numbers of bacteria
entrained, while the Andersen sampler,
which impacts bacteria directly onto an
agar plate, is sampling the number of
infective particles being entrained.
Effect of Host Plant
An average of 79% more P. syringae
were entrained from beans than from
oats. For each distance sampled, the
number of entrained bacteria from
beans was significantly higher than the
number of bacteria detected from oats.
, A significantly higher population of P.
syringae were detected on bean leaves
as compared with oat leaves. If the
number of bacteria entrained from each
plant is compared as a frequency of the
relative population on the plant, then
the amount of entrainment was actually
higher from oat plants than bean plants.
Bacteria entrained from oats followed
the same dispersal pattern as from
beans but with significantly higher
numbers of entrained bacteria detected
at 0.25 m from plants than from 1 to 3
m. Again, no difference was found in
the number of bacteria detected 1-3 m
from plants. There was a significantly
different particle size distribution from
each plant. The size of particles
entrained from bean plants tended to
decrease in size over distance while
those from oats tended to have a
particle size of 7 //m or greater.
Dilutions vs. Loss of Viable Count
The decrease in entrained spores was
significantly greater than the decrease
in vegetative P. syringae. At 0.25 m,
virtually the same number of spores and
vegetative cells were detected.
However, as the distance increased, the
number of vegetative P. syringae
decreased below the number of B.
subtf/is. This indicates that differential
loss of viable count was occurring. By
dividing the number of vegetative cells
by the number of spores detected at
each distance, it was possible to
generate a survival ratio over distance.
If the number of vegetative cells
decreased relative to the number of
spores at each distance, then the ratio
would become smaller over distance,
which is what was observed.
Therefore, loss of viable count is
occurring beyond a distance of 0.25 m
from plants.
73
-------
Time After Application
Significantly more bacteria were
entrained 2 hours after application than
24 and 120 hours after application.
However, the number of bacteria
entrained 24 hours after application
was not significantly different than the
number entrained 120 hours after
application. Bacterial populations on
plants reflected the decline in the
number of bacteria entrained to a
limited extent. The number of bacteria
detected 2 hours after application was
significantly higher than the number
detected at 24 hours, and these
populations were significantly higher
than those detected after 120 hours.
However, while plant populations
continued to decline from 24 to 120
hours, the numbers of bacteria being
entrained did not.
Results from these experiments
suggest that: A) GEMs are entrained
following leaf applications in response
to moderate amounts of wind, B) the
type of sampler used to detect bacterial
entrapment must be based on the type
of data desired, C) the magnitude of
bacterial entrainment may be dependent
on the host plant, D) the process of
entrainment may cause death or injury
to vegetative bacteria, and E) the
magnitude of bacterial entrainment may
not decrease over time.
Future research should be directed
toward using this method to measure
the colonization of plant surfaces by
entrained bacteria, identifying
environmental conditions most
influential on bacterial entrainment, and
conducting field experiments to validate
the greenhouse data.
PUBLICATIONS
M.V. Walter, V.J. Prince, B. Marthi, B.
Shaffer, L. Ganio, and R.J. Seidler.
1990. Measuring aerosol survival of
sprayed bacteria. Review of progress
in the biotechnology-microbial pest
control agent risk assessment program.
USEPA Office of Research and
Development, November 5-9, Corvallis,
OR.
M.V. Walter, V. Prince, L. Ganio, and
R.J. Seidler. 1990. A method to
measure bacterial entrainment from the
phyllosphere. Abstract. Annual
meeting of the American Society for
Microbiology May 13-17, Anaheim,
CA.
M.V. Walter, V.J. Fieland, L.M. Ganio,
and R.J. Seidler. 1991. Factors
affecting bacterial entrainment from the
phyllosphere. Abstract. Annual EPA
Biotechnology All Investigators
Meeting. April 8-11, Arlington, VA.
M.V. Walter, V.J. Fieland, L.M. Ganio,
and R.J. Seidler. 1991. Measuring
entrainment of bacteria from the
phyllosphere. Manuscript submitted to
Applied and Environmental Microbiology
3/91.
74
-------
UPDATE ON THE ASSESSMENT OF BACTERIAL DISPERSAL
FROM FOLIAGE BY RAIN
H. A. McCartney and Julie Butterworth
A.F.R.C. Institute of Arable Crops Research,
Rothamsted Experimental Station,
Harpenden, Hertsfordshire
United Kingdom
INTRODUCTION
Recent advances in genetic
engineering techniques have made
possible the modification of bacteria for
use as novel pest and disease control
agents in crop plants, or as potential
replacements for chemicals currently
used in agriculture. However, before
such agents are released into the
environment, possibly on a large scale,
it is important to know if they are likely
to have deleterious effects on non-
target organisms or plants.
The release of genetically engineered
microorganisms requires their
containment within the area of release,
and thus an assessment of their
potential for spread is of great
importance. Rain splash as a factor in
the spread of disease in crops has long
been recognized and it has been shown
that it can play a major part in the
spread of many fungal and bacterial
plant pathogens. However, little is
known of the efficiency of rain in
removing and dispersing bacteria from
leaf surfaces. A project to study the
potential of rain splash for removing
and distributing bacteria from the
foliage of crop plants is underway at
the Institute of Arable Crops Research,
Rothamsted Experimental Station.
METHODS
All experiments were done in the
Rain tower/wind tunnel facility at
Rothamsted. Mono-sized water drops
(simulating rain drops) were allowed to
fall 11m onto target leaves carrying
populations of bacteria. Rain drops of
between 2 and 5mm diameter could be
produced. Droplets, dispersed by
splash, were collected at different
distances from the targets and their
bacterial content assessed. The
bacterial content of water which
collected on, and subsequently ran off,
the leaves was also assessed.
Isolates of Pseudomonas syringae,
Klebsiella planticola and Bacillus subtilis
var. niger, all resistant to 100//g/ml of
the antibiotic rifampicin were used in
these experiments.
Two species of glasshouse grown
crop plants were used: Phaseolus
vulgaris cv. Prince, and Brassica napus
cv. Cobra. The plants were chosen
principally because of their different leaf
surface characteristics: P. vulgaris has
a hairy surface and no pronounced wax
cuticle, while B. napus has a waxy
cuticle, and is not hairy.
Bacteria were grown overnight in
nutrient broth (Oxoid) at 30°C to a
concentration of between 107 and 108
CFU/ml. The target plants were
sprayed, inside a closed perspex
75
-------
chamber, with the nutrient broth and
left for about 22 hours to allow the
bacterial population to stabilise. The
plants were removed from the chamber
about 6 hours before each experiment
to allow surface moisture to dry from
the leaves. Leaves were removed from
the sprayed plants and placed in an
overlapping layer on a horizontal nylon
mesh covered frame at the base of the
rain tower. The population of antibiotic
resistant bacteria on the target leaves
was assessed by taking a sample of
four leaves before exposing the rest to
splash. Each sample leaf was weighed
and placed in a sterile glass conical
flask containing 50ml of sterile quarter
strength Ringer's solution (Oxoid) and
washed on a flask shaker for 1 hour.
Samples of the wash water were plated
out on selective medium (N + R)
consisting of nutrient agar (Oxoid) with
100^/g/ml actidione to suppress fungal
growth, and 100//g/ml rifampicin as a
selective agent for the inoculated
bacteria. The plates were incubated at
30°C overnight and the bacterial
colonies counted.
The target leaves were exposed to
splash from artificial rain drops for 30
minutes. Immediately after exposure, a
further sample of four leaves was
removed and washed to assess the
numbers of inoculated bacteria
remaining on the leaves. A sample of
water from the rain generator was also
plated out onto N + R as a check for
contamination by antibiotic resistant
organisms.
Splash droplets were collected in
either sterile 9cm plastic Petri dishes
containing 20ml of N + R or in sterile
9cm polypropylene funnels supported in
sterile glass universals. The Petri
dishes were used to estimate the
numbers of bacteria-carrying droplets
splashed and the funnels to estimate
the total number of cells dispersed.
The collecting vessels were supported
at the same height as the leaves, at 8,
24, 40, 56, 72 and 88cm from the
edge of the target area. During an
experiment the two Petri dishes closest
to the target were replaced with fresh
plates at three minute intervals and the
third closest plate was replaced after
15 minutes. Four pre-weighed funnel/
universal assemblies were arranged
underneath the leaf support frame to
collect samples of the water running off
the leaves. Run off water was also
collected in plastic trays placed under
the target. The sum of the volume of
water collected in the trays and
funnel/universal assemblies was taken
as the total amount of run off water.
The Petri dish droplet samplers were
incubated at 3Q°C overnight and the
colonies growing on the plates counted.
The droplet collecting funnels were
rinsed with a known volume of sterile
distilled water which was added to the
collected water and samples of the
resultant bacterial suspension plated
out onto N + R medium. Samples of
run-off water were diluted as
appropriate and also plated out onto
N + R. The plates were incubated
overnight at 30°C and colonies
counted.
RESULTS
Twenty-five experiments were done
using the six different combinations of
two plant species, P. vulgaris and B.
napus and three bacteria species P.
syringae, K. planticola and B. subtilis.
76
-------
For all experiments the rain drop size
was 3mm and air temperature was
about 18°C with about 85% relative
humidity. The effect of rain drop size
was studied in a further 24 tests.
Three different rain drop sizes were
used: 2.4, 2.9 and 4.7mm diameter.
Eight experiments were done with each
drop size using B. subti/is on B. napus
as the target. Experiments were done
at about 21 °C and 75% relative
humidity.
Proportion of Bacterial Populations
Removed by Splash
There was a large variation in the
estimate of the proportion of bacteria
removed, by splash for replicate
experiments and no statistical
differences were found either between
bacteria species or between plants.
The variation in bacterial numbers
between individual leaves, used to.
assess pre- and post-splash bacterial
populations, was often large and
frequently the differences between pre-
and post-splash bacterial numbers were
not statistically significant. However,
the results suggest that up to 90% of
bacteria may be washed off the leaves.
Surprisingly, there was little effect of
drop size in removing B. subtilis cells
from B. napus leaves. It was estimated
that, for all drop sizes, about 75% of
the bacteria were removed during each
experiment.
Bacteria Removed bv Run-off
In all experiments substantial
numbers of bacteria were found in the
run-off water. In the first set of
experiments the numbers of bacteria
recovered per unit area of target was
frequently (70% of experiments) greater
than the estimated population density
at the start of the experiment,
suggesting that the method of leaf
washing may have underestimated the
initial populations. There was no
significant difference between the
proportion of bacteria in the run-off
water (Pr) for the two plant species and
the same organism. However, Pr for P.
syringae and K. planticola, was nearly
always greater than for B. subtilis. In
the second series of experiments the
value of Pr increased with drop size
(mean values of 0.27, 0.44 and 0.84
for 2.4, 2.9 and 4.7mm drops
respectively).
Dispersal by Splash
The numbers of inoculated bacteria
recovered on exposed agar plates or
from the funnel assemblies decreased
rapidly with distance from the target
leaves. The number of bacteria
deposited per square centimetre, D,
decreased exponentially with distance
from the target: D = A exp(-ax) where
A, a are constants. The equation has
the property that D decreases by half at
regular distances, d1/2 ( = 0.693/a),
from the source. The half distance,
dy2, is an easily visualised measure of
the deposition gradient.
Deposition gradients for both plates
and funnels for all tests were steep: dy2
ranged between about 5 and 17 cm.
For a given organism the values of d,/2
for funnels were similar for each plant
species: for P. syringae and K.
planticola the average value of dy2 was
about 7cm for B. napus but slightly
larger (»9cm) for P. vulgaris. Because
77
-------
the numbers of B. subtilis cells caught
in the funnels were small at distances
greater than about 20cm from the
target dy2 values were unreliable. The
values of dy2 for plates were similar
(di^«7cm) for K. planticola and B.
subtilis inoculated leaves, but, for P.
syringae on both plant species
deposition gradients were significantly
shallower (di/2=12cm). For both funnels
and plates mean (of 8 replicate
experiments) d i/2 values were largest for
4.7mm rain drops and smallest for the
2.4mm drops: 8.2, 7.9 and 5.1cm for
the funnels and 10.3, 7.5 and 7.2cm
for the plates for 4.7, 2.9 and 2.4mm
rain drops respectively.
Proportion Dispersed by Splash
The total numbers of bacteria
dispersed in. splash droplets per
centimetre width of the target was
estimated by numerically integrating the
deposit with distance from the edge of
the target. The numbers of bacteria
dispersed depended on the initial
population present on the leaves. In
nearly all cases, (80%), it was
estimated that less than 5% of the
initial inoculated bacterial population
was splashed. The average fraction of
bacteria splashed was about 1.5%
(range 0.02 - 8%). There appeared to
be little systematic difference in
proportions splashed between either
bacteria species or plant species,
except for B. subtilis where the mean
values for B. napus were significantly
larger than for P. vulgaris (2%
compared to 0.3%). In the second set
of experiments a greater proportion of
B. subtilis cells were splashed from
target leaves by large rain drops than
the smaller ones: about 1.7% for
4.9mm drops compared to 0.5 and
0.2% for 2.9 and 2.4mm drops.
, The number of cell carrying droplets
splashed was much smaller than the
number of cells for both P. syringae and
K. p/antico./a: between 5 and 200 times
as many P. syringae cells and between
2 and 50 K. planticola cells were
splashed than droplets. For B. subtilis
the numbers of droplets and cells
splashed were similar. This difference
between organisms was probably due
to differences in initial populations. The
initial populations of both P. syringae
and K. planticola were usually larger
than those of B. subtilis. The number
of cells carried in each splash droplet
will depend on the leaf population
density: larger densities will lead to
more cells per drop. Thus at high
densities the number of cells splashed
will be larger than the number of cell
carrying droplets splashed. As with the
cells larger rain drops splashed more
bacteria carrying drops than smaller
ones. About twice as many cells as
bacteria carrying droplets were
splashed for the three drop sizes.
Dispersal Rates with Time
The pattern of dispersal with time, as
indicated by the change in deposition
rate on the plates nearest the target,
was substantially different for th'e
experiments with P. syringae than for
the other two organisms. With P.
syringae the deposition rates decreased
almost linearly with time and for both
plant species were reduced by about
20%. over the 30 min. exposure period.
Deposition rates of the other two
organisms decreased roughly linearly
78
-------
with time when splashed from P.
vulgaris plants:.rates were reduced by
about 70% during the experiments. For
B. napus the decrease in deposition
rates.followed a concave pattern .with
rates at the end of the test only about
10% of those at the beginning,
suggesting that the available bacteria
were washed off B. napus leaves more
readily. There were also differences in
the changes in deposition rate with time
for the rain drop experiments. For the
4.7mm drops deposition rates,
decreased in a similar manner to the
first set of experiments, except that,the
differences between beginning and end
rates were smaller; about a 60%
reduction. With the other drop sizes
deposition rate initially, increased
reaching a maximum about 5 minutes
and 1-0. minutes from the start of ,the
experiment for the 2.9 and 2.4mm
droplets respectively. Deposition rates
then decreased and at the end of the
tests were about 45% of the maximum
value for the 2.9mm drops and about
60%'for,the 2.4mm drops.
DISCUSSION
Our experiments show that rain
splash can be an efficient mechanism
for removing bacteria from leaf
surfaces. Patterns of dispersal and the
distances travelled by bacteria in splash
droplets were similar to those which
have been reported for both fungal and
bacterial pathogens. Although we only
examined the effect of splash on the
removal of three species of bacteria
from two plant species, the. results
suggest that splash dispersal patterns
and splash processes may be similar for
a wide range of organisms and plants.
The proportions of the leaf populations
dispersed in splash droplets were small
(usually < 10%) as were the distances
travelled by most of the bacteria.
However, large numbers of cells were
found in the run-off water collected
from underneath the target area. Thus,
rain water running off leaves to the soil
around crop plants could be an
important source of bacteria for further
spread through travel in ground water,
or by secondary splashing.
Large rain drops appear to be more
effective in removing and splashing
bacteria from leaves than small ones.
However, bacteria were more quickly
removed by the large drops, suggesting
that for longer periods of rainfall drop
size may not be critical for the removal
(but not necessarily the dispersal) of
bacteria from leaves. Indeed, when the
proportions of bacteria removed during
the experiments were adjusted for the
volumes of rain falling on the target,
the small drops were just as efficient at
removing bacteria as the large ones.
These experiments show that rainfall
has the potential to remove and
disperse bacteria from leaf surfaces.
However, as studies of the dispersal of
fungal pathogens show, the splash
process is complex and its effective-
ness depends upon the nature of the
rainfall or irrigation and on the nature of
the crop or vegetation canopy. More
work is needed to answer questions
such as how does crop structure effect
bacteria spread? Which rainfall events
favour the dispersal of bacteria? Are
natural bacteria dispersed in a similar
manner to inoculated bacteria? How
efficiently do bacterial cells survive
dispersal? Wind may also have a role in
the spread of bacteria from leaf
79
-------
surfaces. Thus, an understanding of
the effects of the combination of wind
and rain to the removal and spread of
bacteria is needed before the potential
for dispersal of genetically manipulated
organisms can be quantified.
FUTURE WORK
As the co-operative agreement
terminates in July 1991 there is little
scope for expanding the studies much
further. However, experiments, to
examine the influence of environmental
factors such as relative humidity on
bacterial spread are planned. We also
intend to examine the potential for
longer distance transport by very small
airborne droplets. Funding to expand
these studies is being sought from
other sources.
PUBLICATIONS
Butterworth, J. and McCartney, H.A.
(1990) Dispersal of foliar bacteria by
rain splash (abstract). Journal of
Applied Bacteriology, 69, xix (meetings
supplement).
Butterworth, J. and McCartney, H.A.
The dispersal of bacteria from leaf
surfaces by water splash. Submitted to
Journal of Applied Bacteriology
(January 1991).
8.0.
-------
GENETIC AND MOLECULAR ANALYSIS OF SURVIVAL MECHANISMS
OF PSEUDOMONADS
Anne J. Anderson, Robin Buell, Jirasak Katsuwon, Chris Heck and Rob Zdor
Biology Department, Molecular Biology Program
Utah State University
Logan, Utah
INTRODUCTION
Colonization of plant roots by certain
isolates of fluorescent pseudomonads
promotes plant growth and suppresses
pathogens. These beneficial effects
have raised interest in the agricultural
value of crop inoculation using native or
genetically engineered strains to
improve quality and yield. In order to
assess some of the potential hazards
associated with release of genetically
engineered pseudomonads we have
researched basic processes involved in
colonization of plant roots by
pseudomonads. Colonization is
important in the attainment of effective
populations of the bacteria and in the
persistence and spread of organism in
the environment.
We have concentrated upon two
aspects of pseudomonad colonization in
the plant root environment: 1)
mechanisms of pseudomonad attach-
ment to the plant root involving plant
agglutinin, and 2) the role of catalase
and superoxide dismutase (SOD) in
pseudomonads in protection of the
bacteria against root produced activated
oxygen species.
METHODS, RESULTS and DISCUSSION
Mechanism of Attachment to the Plant
Root Involving Agglutinin
Certain isolates of Pseudomonas
put/da are agglutinated by plant surface
glycoproteins termed agglutinins.
Purification of the agglutinin has been
initiated to determine structure and as
a primary step in isolation of the plant
genes encoding agglutinin-active
components. We have examined wild
type agglutinable isolates, mutants
which lack the agglutination phenotype
(Agg") and agglutination complemented
mutants to probe the molecular and
genetic basis of the agglutination
phenotype in the bacterium.
A) The agglutinin is a IS-lectin
arabinogalactan protein
Agglutinin activity is detected in
water washes of intact plant roots.
Components in bean root washes have
been purified by passage through a
sizing filter to remove molecules of less
than 10,000 Mr, followed by
Concanavalin A and DEAE-Sephadex
chromatography. These approaches
demonstrate that agglutinin activity is
associated with multiple fractions. The
majority of the agglutinin activity elutes
without adsorption to both ConA and
DEAE-Sephadex. This nonadsorbed
material retains ability to react with
Yariv's reagent: a reaction indicative of
fc-lectin activity associated with
arabinogalactan proteins. IEF of the
nonadsorbed material fractionates the
81
-------
preparation into four bands of agglutinin
-active components with pi 7-8 and a
larger agglutinin-active heterogenous
fraction with a more basic pi.
A second line of evidence supports
the agglutinin to be an arabinogalactan
protein. Both crude and more purified
preparations react with a monoclonal
antibody MAC 207, which is derived to
an arabinose epitope of a membrane-
associated arabinogalactan protein from
carrot. This antibody was generated in
the laboratory of Dr. Keith Roberts,
John Innes Institute of Norwich
England, with whom collaborative
studies are being continued.
Agglutinin-active fractions which elute
without adsorption to ConA, but which
are retained or adsorbed to DEAE-
Sephadex react differently to two other
mononclonals derived by Robert's et al.
to pectin : thus these studies indicate
that differences in carbohydrate
structure between agglutinin-active
components exists. Data from studies
of Robert's et al. indicate that MAC
207 antibody-reactive materials are
widespread in the plant kingdom: these
data agree with our previous observ-
ations that extracts from a wide array
of plant species possess agglutinin
activity.
In collaboration with Dr. Schippers at
the University of Utrecht we have
surveyed several pseudomonad isolates
from wheat, tomato and potato for
agglutinability. Certain of these
European strains display agglutinability
with our crude and more purified
preparations of agglutinin from bean, as
well as with crude root wash
preparations from wheat, tomato and
potato prepared at Utrecht. These data
support that the agglutination
phenotype in pseudomonads has
relative little plant species specificity.
Additional collaborativestudies with our
Agg + wild type and Agg" mutants of P.
put/da have revealed that the
agglutination phenotype has little
influence on whether the bacteria
become endorhizosphere colonists or
remain on the rhizoplane.
B) Genetic and molecular studies of the
bacterial agglutination phenotype.
Genetic and molecular techniques
have been used to explore the nature of
the bacterial phenotype which is
responsible "for agglutination. Genetic
analysis has revealed two distinct loci
associated with the agglutinable
phenotype. One locus, termed aggA
has been analyzed in mutant Tn5 5123.
A second locus aggB is detected in Tn5
mutants 1104 and 6000.
The aggA locus is present in a 2.7
kbp EcoRI-Hind III subclone of parental
DNA. Deletion analysis using
exonucleases revealed only 1.45 kbp
was needed for complementation of the
Agg" mutant 5123 to the Agg +
phenotype. The presence of the
subclone in transconjugants of 5123
restored the agglutinability of the cells.
Adherence to intact bean roots surfaces
in a 15 minute exposure was also
restored to wild type level in the
transconjugants. Thus, a role of the
locus in binding, which was suggested
by previous studies with Agg" mutants,
was confirmed.
Sequence analysis of the 2.7 kbp
fragment reveals an open reading frame
on one strand of 1356 nucleotides
encoding a predicted 50,509 Da protein
and pi of 5.24- A consensus sequence
82
-------
for a ribosome binding site (AGGGAG)
was located 8bp upstream of the
projected ATG translational start site.
Portions of the canonical sequences
recognized by the 6 54 and 5 70 RNA
polymerase holoenzymes were present
5' of the ATG. The predicted protein
contains one membrane associated
helix, from residues 2-17, and there is
consensus for a prokaryotic cleavage
site located between residues 22 and
23.
Another reading frame was present
on the opposing strand of the 2.7 kbp
subclone but it lacked a ribosome
binding site, promoter sequences and
would use GTG as a potential
translational initiation site. No
translational stop codons were
observed in 86 bases 5 to the GTG
site. These data suggest that the
second ORF may be incomplete and not
be involved in the agglutination
phenotypes. A search of the NBRF and
GenEMBL data bases has revealed no
homologous sequences with either of
the predicted ORFs.
Hybridization analysis with the aggA
locus to the genomes of plant-
associated bacteria indicated that with
thirty isolates examined from bean,
wheat, tomato and potato, only the
agglutinable isolates demonstrated
strong to moderate hybridization. Thus
this locus has good specificity for
detection of the agglutinable
phenotype.
The second locus aggB from P.putida
also is involved in the agglutination
phenotype. Complementation of the
phenotype has not been achieved in
trans using cosmid clones from a wild
type P. put/da library. Marker exchange
studies with the locus aggB: Tn5
disrupted the Agg+ wild type
chromosome and the recombinants
were nonagglutinable. Preliminary data
suggest the presence of repetitive
sequences in the region of the insertion.
Molecular studies have demon-
strated that ceil envelope proteins are
associated with agglutination pheno-
type. Agg'mutants Tn5 5123, EMS
1202 and EMS 1236 possessed a
16,000 Mr protein band which was less
apparent in the Agg+ parental strain.
Fractionation of the cells indicates this
16,000 Mr protein is located in the
periplasm. Agg" mutants Tn5 4312,
Tn5 5123 and EMS 1236 displayed an
outermembrane protein band of Mr
42,000 which was not observed in the
parent. Transconjugants of Agg" 5123
which were restored to agglutinabilty
(Agg + ) by the presence of clone
pRKAGG201 bearing a wild type 2.7
kbp EcoRI -Hind III fragment did not
produce the 42,000 Mr protein but still
expressed the 16,000 Mr protein.
These data suggests that the 42,000
Mr protein is associated with the
agglutination phenotype.
Catalase and SOD Activities from Root
Colonizing Pseudomonads
The plant root is potentially a toxic
environment for microbes because of
the production of activated oxygen
species from the root surface.
Consequently, we have examined the
role of catalase and SOD in root
colonizing pseudomonads as a
protective mechanism to permit survival
in the rhizosphere. Each of the
pseudomonad isolates examined have
catalase and SOD activities. We have
83
-------
surveyed five beneficial isolates from
wheat, six isolates from sugar beet, the
bean colonizing isolate of P. putida and
isolates from plants grown at a mine
spoil site with acidic pH and high
copper content. The mine spoil site was
examined because of the potential of
heavy metals to catalyze
interconversions of activated oxygen
species which would perhaps put
oxidative stress on the bacteria. The
pseudomonads from the agricultural
soils had similar levels of catalase
whereas three of four mine spoil
isolates displayed three to ten fold more
activity. Multiple isozymes of catalase
of differing mobilities between isolates
were observed for cells grown to
stationary phase in rich medium,
whereas one or two isozymes of SOD
were obtained.
Studies with P. putida and P.
fluorescens isolates with beneficial
properties demonstrate that SOD is
rapidly increased in activity upon
exposure to bean roots and remains at
high level during root colonization.
Catalase activity is low in cells that
haveestablished colonization. However,
upon short term exposure in the
bacteria the activity transiently
increases, in part because of production
of a second isozyme and enhanced
levels of the normally expressed
isozyme. We are exploring the
differential functions of these catalase
isozymes. Catalase isozyme A is the
first isozyme produced during
logarithmic growth and has a
cytoplasmic location. Additional
isozymes, which are less anodic on
nondenaturing gels and are termed B,C
etc., are produced later in growth
phase. Catalase B is located in the
membrane envelope of P. putida.
Increased activities of catalase A and
production of catalase B and C occur in
logarithmic phase cells of P. putida
after exposure to H202. Thus these
changes in isozyme expression may be
important for the survival of the
bacterium to exogenous H2O2, a
situation occurring upon exposure to
plant roots.
It seems likely that catalase A and B
are products of different genes, rather
than being derived from one gene
product. The isozymes are differentially
inhibited by aminotriazole and organic
solvents. Also we can separately
mutate the activities in P. putida.
Mutant J1, obtained by EMS
mutagenesis of the parental strain,
lacks catalase A activity under liquid
growth conditions. Nondenaturing
electrophoresis indicated that catalase
A was absent, although catalase B and
C were produced during late growth
periods. However, mutant J1 was able
to colonize plant roots at wi|d type
rates. This ability may relate to
production of catalase A in J1 after
exposure to external sources of H202.
A similar induction of catalase A in J1
occurs on contact of J1 with plant
roots. Cells of J1 examined from the
root surface after twelve hours of
exposure displayed,both catalase B and
catalase A isozymes.
A further mutant J6 obtained by Tn5
insertion into J1 lacked all detectable
catalase activities when grown on liquid
growth media even to stationary phase.
Thus J6 appears to have a mutation
which prevents the production of
catalases B and C. This mutant only
colonized roots at levels much reduced
from that of the wild type. Also J6
84
-------
was reduced in its ability to survive
H202 treatments although it was able
to produce catalase A, after H202
exposure. These data suggest that
regulation of catalases B and C as well
as catalase A is important in the
survival of bacteria to H202 and on
plant roots.
We are exploring the altered gene
expression of SOD upon root coloniz-
ation by the generation of SOD
deficient mutants, through marker
exchange of a mutated gene into the
parental cell. Currently we have isolated
a fragment from a cosmid library of P.
putida which will complement the SOD
deficiency in a SOD-mutant of
Escherichia coli. The complemented
cells produce a SOD activity which
migrates on. nondenaturing gel
electrophoresis to the same location as
the Fe SOD from the pseudomonad cell.
This migration pattern is quite distinct
from that of the £.. coli SODs. We are
currently subcloning the fragment and
will perform transposon mutagenesis to
delete the SOD activity.
CONCLUSION
We have demonstrated that root
colonizing pseudomonad cells have
several levels of interaction with the
plant. Interaction between: a root
surface arabinogalactan protein and
surface features of root colonizing
bacteria is involved in strong
attachment and improved colonization
potential. The agglutination phenotype
in the bacteria is related.to discrete
genetic loci and we have identified one
locus as being a specific marker for the
agglutinability trait. Other genes in the
bacteria are also regulated upon root
colonization. These include genes for
catalase and SOD which encode
proteins involved in the survival of the
bacteria on the root surface.
FUTURE STUDIES
Our goals are to:
1) continue the purification of - the
agglutinin so that amino acid sequences
and glycosylation structure can be
determined. We will initiate progress
towards isolation of plant genes
involved in agglutinin synthesis.
2) isolate genes encoding catalase A
and FeSOD from P. putida. Mutated
genes will be inserted into the
chromosome of pseudomonads by
marker exchange to probe their
function. We also will examine the
regulation of the promoters of the
genes through gene fusion constructs.
3) examine the function of the promoter
regions for the two loci that are
involved in the agglutination phenotype
in the bacteria. We will use promoter
fusions to determine if expression of
the agglutination genes is regulated as
the bacteria colonize root surfaces.
PUBLICATIONS
Tari, P.M. and A.J. Anderson. 1988.
Fusarium wilt suppression and
agglutinability of Pseudomohas putida.
Appl. Environ. Microbiol 54:2037-
2041.
Katsuwon and A.J. Anderson. 1989.
Response of plant-colonizing
pseudomonads to hydrogen peroxide.
Appl. Environ. Microbiol 55:2985-
2989.
85.
-------
Anderson, A.J., and Katsuwon, J.
1990. Catalase and superoxide
dismutase of root-colonizing
saprophytic pseudomonads. Applied
and Environmental Miicrobiology.
56:3576-3583.
EPA Investigators Meeting, Corvallis,
November 13-17, 1989. Presentation:
Bacterial traits involved in root surface
colonization.
A.J. Anderson. 1990. Biological
Control of Sugar Beet Diseases. 161-
163. In: Proceedings of the University
of Idaho Winter Commodity Schools
Burley, Idaho Feb 15 1990, Caldwell
Idaho Feb 16 1990.
UCLA Conference April 1990. R. Buell
and A.J. Anderson. Genetic analysis of
adhesion of P. put/da. J. Cellular
Biochem. Suppl. 14E p. 314
in: G.Defargo, ed., Plant Growth
Promoting Rhizobacteria. 1991.
i) R.E. Zdor and A. J. Anderson,
Influence of root colonizing bacteria on
the defense responses of beans.
ii) J. Katsuwon and A. J. Anderson,
Survival strategies of root colonizing
pseudomonads
iii) C. R. Buell and A.J. Anderson,
Genetic analysis of agglutination in
Pseudomonas put/da
i-iii) presented at Rhizosphere meeting,
Interlaken October 1990.
EPA Investigators meeting April 8-12
1991 Presentation Genetic and
molecular analysis of survival
mechanisms of pseudomonads.
Buell, R. and A. J. Anderson. Genetic
analysis of a locus involved in
agglutination of pseudomonads (in
review J. Bacteriology)
Katsuwon, J. and A.J. Anderson.
Characterization of catalases from
fluorescent root colonizing
pseudomonads (in review Appl.
Environ. Bacteriol)
Zdor, R. and A. J. Anderson. Defense
responses in bean activated by root
colonization by fluorescent
pseudomonads (in review Plant and
Soil)
B. Lovic, C. Heck, J.J. Gallian and A. J.
Anderson. Biological control of sugar
beet pathogens (in review
Phytopathology)
Anderson, A. J., R. Buell, R. Whetten
and P. Tari. Cell surface properties and
agglutinability of Pseudomonas put/da
(in review Canadian Journal of
Microbiology)
C. R. Buell and A.J. Anderson. 1990.
Genetic analysis of agglutinability of
beneficial root colonizing
pseudomonads. J. Cellular Biochem.
Suppl. 14E, p.314.
86
-------
STABILITY OF FOREIGN DNA
IN THE FUNGUS Colletotrichum gloeosporioides
John L. Armstrong1 and Deborah L. Harris2
U.S. Environmental Protection Agency1
ManTech Environmental Technology Inc.2
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
Progress in the genetic engineering of
fungi has been rapid in recent years.
With these advances, researchers are
observing that the foreign DNA is often
unstable in the transformed fungi. It is
typical for the introduced DNA to be
lost or rearranged within the
chromosome. Also, transformants are
often relatively sickly as demonstrated
by their poor survival, abnormal
pigmentation, and strange colony
morphology. We selected the fungus
Colletotrichum gloeosporioides var.
aeschynemone (CGA) as a model
organism to study the factors that
determine the stability of transformed
marker genes in CGA sprayed on plants
and in microcosm chambers.
METHODS
Transformation of Protoplasts
and Detection of Foreign DNA
The fungus was cultured for 4 days
at 30°C on Torula yeast agar (15 g
Torula yeast, 15 g starch, 1 g K2HPO4,
0.5 g MgSO4'H20, 15 g agar/L) or in
YpSS broth. To prepare protoplasts,
spores were harvested from plates with
sterile water, centrifuged for 5 min. at
1000xg, suspended in a solution
consisting of 2 parts 1.2 M MgSO4-
Tris-CI (10 mM, pH 5.8) and 1 part 0.6
MgS04-Tris-CI (10 mM, pH 5.8.)
containing Novozym 234 (50 mg/ml),
and incubated for 5 hours at 32°C.
Protoplasts were centrifuged for 1 min.
at 2000xg and washed five times by
centrifugation in 1.2 M sorbitol-Tris-CI
(10 mM, pH 7.0) to remove nucleases
present in the Novozym. Cells were
resuspended in the sorbitol solution for
DNA transformation.
Cells were transformed with pBT.
This plasmid was made by inserting a
2.6 kb Sal I fragment that encodes the
p-tubulin gene from a Neurospora
crassa mutant (gene confers resistance
to benomyl) into the Sal I site of
pBS( + /-) (Stratagene). Twenty //g pBT
DNA were added to about 106 proto-
plasts. After adding an equal volume of
30% PEG in 10 mM CaCI2, the
suspension was incubated 30 min. at
25 °C. Protoplasts were centrifuged for
3 min. at SOOOxg, resuspended in 1.2
M sorbitol-Tris-HCI (10 mM; pH 7), and
spread on agar containing 0.8 M
sucrose. After 24 h, four ml volumes
of 1 % YpSS overlay agar containing 10
jjg benomyl/ml were layered on the agar
surfaces. These were incubated at
30°C for 5 days. Putative
transformants were restreaked three
times on the selective medium.
To extract DNA, fungi were cultured
in YpSS broth containing 10 /vg
87
-------
benomyl/ml and collected by filtration,
dried, and ground with a mortar-pestle.
The powder was suspended in a
hexadecyltri-methylammoniumbromide
(CTAB) solution (10% CTAB, 0.7 M
NaCl), heated 1 hour at 55°C, mixed
with an equal volume of
chloroform:isoamyl alcohol (24:1),
centrifuged 10 min., and the
supernatant liquid collected. An equal
volume of 1% CTAB-Tris-CI (50 mM,
pH 8.0) was added. After 30 min. at
25°C, the suspension was centrifuged
at 5000xg for 5 min. The pellet was
dissolved in cesium chloride solution
and the DNA purified by ultra-
centrifugation.
Presence of the p-tubulin gene in
transformants was confirmed by DNA
hybridization. DNA was restricted,
electrophoresed, transferred to a Nytran
membrane, and probed with the Sal I
fragment carrying the (J-tubulin gene.
The probe was labelled with 3^P by
random primer extension. The final
wash in the hybridization step was
performed for 15 min. at 65°C in 2 X
SSC and 0.1 % SDS. Hybridization was
detected by autoradiography.
Construction of Plasmid pBC
Plasmid pBC was prepared by first
restricting the Sph I site in the multiple
cloning site (MCS) of pBS( + /-) and
inserting a 2.2 kb Sph I fragment which
encoded the cutinase gene from C.
gloeosporioides. The direction of the
cutinase reading frame in this construct
was away from the single Eco Rl site in
the MCS. This DNA was then doubly-
digested with Eco Rl and Eco RV (site
about 1 kb from the Eco Rl site in the
MCS) and blunt end-ligated to eliminate
the 1 kb RI-RV fragment. This step
removed the Sal I site in the MCS and
shifted the unique Nar I site in the
cutinase gene to a position near the
middle of the fungus DNA insert. Next,
we restricted the Nar I site, removed
the tails, and inserted Sal I linkers.
Finally, the Sal I fragment carrying the
N. crassa p-tubulin gene was inserted
into the new Sal I site of the cutinase
sequence.
Microcosm Studies
Microcosm studies were performed in
a glass chamber (1 .Ox 0.75 x 0.75 m)
that was irradiated with a 1000-watt
Sylvania metal halide lamp (18 h/6
hours light/dark cycle). Crimson clover
plants were raised in wooden trays
filled with 6-7 cm. of soil. To spray
plants, CGA spores were washed off
Torula yeast agar plates with sterile
water, centrifuged twice in water, and
suspended at 2-5x107 spores/ml.
Plants and soil were sprayed with about
100 ml suspension/tray using a plastic
misting bottle. Three plant samples (2-
3 g/sample) were aseptically taken at
random from each chamber on each
sampling day, put in sterile plastic bags
with 20 ml water, and blended for 1
min. in a Stomacher blender (Model 80,
Tekmar Co.). Using a spatula, three
soil samples (1-2 g/sample) were
aseptically collected at random from the
top centimeter of soil and vortex-mixed
for 1 min. in 18 x 150 mm screw-
capped tubes containing 5 ml water and
glass beads. To enumerate colonies,
samples were diluted in water, spread
in duplicate on Martin's agar amended
with 100//g benomyl/ml, and incubated
for 36 hours at 30°C. Colony counts
88
-------
from a pair of plates were averaged for
calculations of colony forming units
(CFU).
RESULTS AND DISCUSSION
Transformation of Protoplasts
and Detection of Foreign DNA
We obtained benomyl® strains of
CGA following transformation with
plasmid pBT (A/, crassa p-tubulin gene
on Sal I insert in pBS(+ /-)). To date,
three putative transformants have been
analyzed. They tolerate up to 100 /yg
benomyl/ml and hybridize with the p-
tubulin gene probe. However, in the
presence of the antibiotic, only
transformant II-4 has the orange
pigmentation of the parent strain and
does not have retarded growth. This
isolate also grows without significant
inhibition in the presence of 300 //g
benomyl/ml. Based on DNA hybrid-
ization of Southern transfers, the
transforming DNAs in the three isolates
inserted into sites other than the
genomic P-tubulin sequences. We are
currently determining if the foreign DNA
changes the pathogenesis of the
transformants by inoculating northern
joint vetch plants (susceptible to
invasion by CGA) with spores from
transformants.
We are at the early stages of the
transformation studies and have not yet
transformed CGA with pBC (pBS(+/-)
carrying N. crassa (i-tubulin gene with
flanking C. gloeosporioides cutinase
sequences). As we acquire these
transformants, they will be compared
with isolates containing DNA from
plasmid pBT (pBS( + /-) carrying
Neurospora p-tubulin without the
flanking C. gloeosporioides cutinase
sequences). This will allow .us to
determine the influence of the
associated, homologous cutinase DNA
on a) frequency of p-tubulin gene
transformation, b) site of integration, c)
stability of the transforming DNA in the
chromosome, and d) persistence of the
benomyl phenotype.
Microcosm Studies
In the second aspect of our research,
soil/plant microcosms are being used to
compare survival of CGA transformants
with the nonrecombinant CGA parent.
We have already developed methods for
spraying spores on plants and
processing soil and plant samples. This
preliminary work was done by
inoculating crimson clover plants with
106 CFU/g of leaf. Populations dropped
about 100-fold by day 22. During the
same period, populations in the soil
decreased from about 1O4 to 10
CFU/g of soil.
Currently, microcosm experiments
are being used to compare populations
of the parent and transformant II-4
strains. When the CGA populations in
the microcosms drop to near-threshold
levels, we will culture isolates from leaf
and soil samples and extract the DNAs.
RFLPs of these DNAs will then be
analyzed by DNA hybridization with the
P-tubulin gene as a probe. This will
allow us to determine if the location of
the transformed gene changes during
exposure of the fungus to leaf surfaces
and soil.
v
89
-------
ACKNOWLEDGMENT
Dr. George Rohrmann's expert advice
on cloning strategy was crucial for the
successful construction of pBC.
FUTURE WORK
Plans for this project include: a)
determine genetic factors (e.g.,
homologous versus heterologous
integration) related to stability of
recombinant DNA in transformed fungi;
b) assess survival recombinant fungi in
the field; c) study persistence of
recombinant fungi in microcosms that
simulate on-going field conditions (RH,
air temperature, light intensity, and soil
moisture); d) use recombinant fungi to
enhance bioremediation and restoration
of polluted environments.
90
-------
PERSPECTIVES ON PLASMID STABILITY:
A STUDY OF THE EPA BENCHMARK PLASMIDS
Toai T. Nguyen and Richard E. Lenski
Department of Ecology and Evolutionary Biology
University of California, Irvine
Irvine, California
INTRODUCTION
Microbial degradation of toxic
pollutants released into the environment
by various human activities proceeds
through co-oxidation and / or
mineralization (Alexander, 1981).
Often resulting from non-specific
degradation by enzymes of common
pathways, co-oxidation produces
intermediates which can no longer be
broken down by enzymes with high
substrate specificity further down the
pathway and which accumulate in the
medium. In some cases, the
intermediates are even more toxic than
their precursors, which makes
co-oxidation quite undesirable. It
follows that co-oxidation is rather slow,
fails to completely degrade toxic
pollutants and cannot provide
microflora with energy or metabolites.
In contrast, through mineralization,
toxic pollutants are completely broken
down by microflora which can derive
energy and metabolites from them.
Thus microbial mineralization appears to
be a more attractive process to detoxify
the environment.
Although, in several documented
instances, microbial mineralizationcould
be shown to result from the combined
action of chromosomal and plasmid-
borne genes in the same strain (Ghosal
et al.; 1985), it is seldom observed that
a naturally occurring pure culture can
completely break down a target
compound. Rather, mineralization
usually results from complementary
action of members of a mixed culture
(Chatterjee et al., 1981). Through
intensive selection in the laboratory,
one can isolate a pure strain that can
use a target compound as sole energy
and carbon source. Often, such a
strain could subsequently be shown to
harbor a plasmid carrying all the genes
necessary for the mineralization
process.
By using selection to engineer a
strain that can metabolize a target
compound, one is faced with the
problem that the resulting strain might
require a much higher concentration of
substrate for growth than could be
found in the worst polluted site.
Compounding that problem, alternative
carbon sources readily available in the
environment might be used in prefer-
ence to the target compound, thus
defeating the purpose of the engineered
strain. Also, a strain developed in a
laboratory environment might lose its
competitiveness over indigent flora of a
polluted site, thus would be rapidly
outgrown before it has time to
complete its intended detoxification
task.
While making them more amenable to
various studies the plasmid location of
those genes of interest also increases
their likelihood to be lost. Our current
91
-------
research focuses on the problem of
plasmid loss from a population of
genetically engineered microorganisms.
With plasmid copy number restored to
its pre-division state shortly after cell
division, plasmid loss could result from
random unequal partitioning of copies
to daughter cells, or segregation. It
could also be accelerated by a higher
growth rate of the plasmid-free cells, as
compared to that of plasmid-bearing
ones, or selection.
Most often, plasmid loss results from
a combined effect of segregation and
selection. Knowing which phenomenon
plays a major role in plasmid loss helps
one decide on the most .appropriate
course of action to enhance plasmid
stability. If segregation is the major
cause of plasmid loss, one could
consider ways to increase copy number
to lessen the chance of loss by unequal
partitioning or to incorporate a function
that ensures equal partitioning. If a
plasmid reduces the growth rate of its
host, increasing its copy number or
incorporating a partitioning function
might have the opposite effect of
making plasmid-bearing cells even less
competitive, thus exacerbating the
problem.
As model plasmids, we used a series
of EPA benchmark plasmids (pR02313,
pRO2317, pRO2318, pRO2320 and
PRO2321; Zylstra et al., in press),
maintained in two bacterial model
backgrounds, Pseudomonas aeruginosa
strain PAO1c and P. putida strain
PRS2015. We determined that all the
above benchmark plasmids are more or
less unstable in either host background.
Whether segregation or selection is the
major cause of plasmid loss depends
very much on the plasmid, the host
background and the growth medium.
METHODS
Bacterial Strains and Plasmids
Dr. R. Olsen (Medical School, Univ.
of Michigan) kindly provided plasmid-
free P. aeruginosa PAO1c and PAOIc
strains bearing benchmark plasmids
PR02313, pR02317, pR02318,
PR02320 and pRO2321. Dr. S.
Cuskey (Environmental Research
Laboratory, Gulf Breeze, FL) kindly
provided plasmid-free P. putida
PRS2015 and PRS2015 strains bearing
the same set of the above benchmark
plasmids. To ensure true isogenicity,
plasmid DMA was isolated from plasmid
be.aring strains by the rapid isolation
technique (Maniatis et al., 1982) and
used to re-transform plasmid-free
strains by the CaCI2 technique (Mandel
and Higa, 1970). Transformants were
isolated, streak-purified on antibiotic
plates, grown to saturation in L broth
with antibiotic, mixed with an equal
volume of glycerol and frozen at -80
°C. Glycerol stocks of plasmid-free
strains were made in a similar way
.except no antibiotic was used. All
experiments were initiated with single
colonies streaked on antibiotic plates
from such glycerol stbcks.
Media and Antibiotics
LB broth contains per liter 10 g
Tryptone, 5 g Yeast Extract and 10 g
NaCI. LB agar is LB broth With 14 g
Bacto agar per liter. PG contains per
liter 30.7 mmoles K2HP04, 14.7
mmoies KH2PO4, 18.7 mmoles NH4CI,
1 mmole MgSO4, 2 micromoles FeSO4,
92
-------
2 mg thiamine hydrochloride and 1 g
glucose. Anhydrous ampicillin (Ap) is
dissolved in dimethyl sulfoxide at 100
g/l, stored at -80 °C, and used at a
final concentration of 500 mg/I.
Carbenicillin (Cb) and piperacillin (Pp)
are dissolved in water at 600 g/l and
100 g/l, stored at -80 °C and used a
final concentrations of 600 mg/I and 20
mg/1, respectively. Tetracycline (Tc) is
dissolved in methanol at 10 g/l, stored
at -80 °C, and used at a final
concentration of 80 mg/I for PAOIc or
10 mg/I for PRS2015 strains bearing
Tc- resistant plasmids.
Competition-type
Stability- and
Experiments
Stability- and competition-type
experiments were conducted in a
pairwise fashion. Essentially,
plasmid-bearing cells were grown ^to
saturation in LB supplemented with
antibiotic, harvested by centrifugation,
washed three times with an equal
volume of 8.5 g/l NaCI. Plasmid-free
cells were grown up similarly but
without antibiotic, and harvested in the
same manner. For a stability-type
experiment, plasmid bearing cells were
inoculated on day 0 as pure culture at
,QD550 - 0.05. A competition-type
experiment started with a culture of the
same OD550 but consisting of a 1:1
mixture of plasmid-bearing , and
plasmid-free cells. Cultures were
incubated at 32 °C, aerated with a 240
RPM gyratory motion. Every
subsequent 24 hours, cells were
subcultured at 1:100, as described
above, into fresh medium. This
regimen resulted in an equivalent of
Iog2(100)=6.64 generations per 24 h.
Each day, saturated cultures were
serially diluted in 8.5 g/l NaCI and
plated on LB agar to obtain single
colonies; the daily frequencies of
antibiotic resistant cells were
determined by toothpick streaking 100
or 200 colonies on plates of LB agar
supplemented with antibiotic.
Data analysis. According to Lenski
and; Bouma (1987), the rate of plasmid
loss could be described by the following
diff.erential equation:
[1]
dP
dt
= - uP- sP(1 - P)
where P is the frequency of plasmid-
bearing cells, u. the segregation rate and
s the selection coefficient. Equation [1]
can be integrated to yield:
[2] P =
• . . [U H
(u + s)Pr
-P0)]e
lu
sP
where P0 is the initial frequency of
plasmid-bearing cells. We computed
estimates of the segregation rate, u,
and selection coefficient, s, by least
squares non-linear regression (Dixon,
1985) fit to equation [2]. Through the
use of a dummy variable (Kleinbaum
and Kupper, 1978), the intercept, P0,
could be varied according to whether a
particular set of points came from a
stability- or competition-type
experiment. This allowed paired
stability- and competition-type
experiments "to be used simultaneously
in the estimation of either u alone
(model I, in which s is constrained to be
zero) or both u and s (model II). Partial
93
-------
E-tests were conducted to determine if
model II significantly improves the fit
over model I with the null hypothesis
being rejected only at p<0.05.
RESULTS AND DISCUSSION
For each pair of stability- and
competition-type experiments, we used
partial F-test to determine whether the
data points fit model II (segregation and
selection) significantly better than
model I (segregation alone). For each
combination of plasmid, host and
growth medium, we tabulated how
many pairs of experiments fall into the
above category. If over 5% of the total
pairs of experiments does, we conclude
that for that combination, selection
plays a significant role in plasmid loss.
The results are shown in Table T. The
TABLE 1. PROPORTIONS OF
SIGNIFICANT CONTRIBUTION OF
process leads to assembling the
host-plasmid-mediumcombinationsinto
two groups: Group A comprising those
with a pattern of instability fitting
model I (segregation alone); group B
comprising those with a pattern of
instability fitting model II (segregation
.and selection).
The above arrangement facilitates
comparison among member combina-
tions of a same group for their
instability. This could readily be done
with group A since it involves direct
comparison of the segregation rates.
However, the instabilities of member
combinations in group B result from an
interaction between segregation and
selection. To arrive a meaningful
comparison, we define a new quantity,
1, a loss factor in equation [3]:
EXPERIMENTS SHOWING
SELECTION IN PLASMID LOSS
Plasmid
PRO2313
PRO2317
pRO2318
pR02320
PR02321
pRO2313
PR02317
PR02318
pRO2320
pRO2321
pRO2313
PR02317
pRO2318
PR02320
PR02321
Experiments
Total with selection
as major cause
host: PRS2015
medium: PG
6
6
6
6
6
host: PRS2015
medium: LB
5
6
18
20
6
host: PAO1c
medium: PG
4
10
4
15
4
0
0
0
1
0
1
2
7
0
5
1
3
0
9
0
Percentage
0.0%
0.0%
0,0%
16.7%
0.0%
20.0%
33.3%
38.9%
0.0%
83.3%
25.0%
33.3%
0.0%
60,0%
0,0%
94
-------
dp
[3] I = -lim = -lim-u-s( 1 -P) = u + s
P->0 Pdt P->0
where u, s. and P are defined
previously.
Since the segregation rates and
consequently the loss factors as well,
are not normally distributed, we chose
non- parametric methods to conduct
our statistically analyses. We used
Kruskal-Wallis method to determine if
there was a significant difference
among the segregation rates of group A
and the loss factors of group B. In
both cases, the null hypothesis was
rejected at p<0.0005. We then
proceeded to carry out pairwise
two-tailed Mann-Whitney test within
each group to determine the relative
ranking in stability. In this case, the
null hypothesis was rejected at
p < 0.05. The results are summarized in
Table 2 (page 96).
Our analyses of the EPA benchmark
plasmids indicate that stability of a
plasmid-bearing strain is the result of a
complex interaction among plasmid,
host and growth medium. It is thus
very difficult to generalize and to
predict if a particular background would
support two plasmids equally well even
if they were derived from a common
precursor. For each particular
combination, it is therefore necessary
to determine empirically the pattern by
which a plasmid-bearing strain behaved.
FUTURE WORK
From our studies, we identified
several host-plasmid-medium combina-
tions with minimal instabilities. We
plan to use them in experiments to test
for the destabilizing effects of gene(s)
of environmental utility. We also
identified several combinations with
high instabilities. These would be ideal
backgrounds to test for the stabilizing
effects of various genetic functions.
REFERENCES
Alexander, M. 1981. Biodegradation
of Chemicals of environmental concern.
Science 211:132-138
Chatterjee, O.K., S.T. Kellog, D.R.
Watkins and A.M. Chakrabarty. 1981,.
Plasmids in the biodegradation of
chlorinated aromatic compounds. |Q:
S.B. Levy,, R.C. Clowes and E.L.
Koenig (eds.) Molecular Biology,
Pathogenicitv, and Ecology of Bacterial
Plasmids. Plenum Press, London.
Dixon, W.J. (ed.). 1985. BMDP
Statistical Software. University of
California Press, Berkeley. Ghosal, D.,
I.-S. You, O.K. Chatterjee and A.M.
Chakrabarty. 1985. Plasmids in the
degradation of chlorinated aromatic
compounds, in: D.R. Helinski, S.N.
Cohen, D.B. Clewell, D.A. Jackson and
A. Hollaender (eds.) Plasmids in
Bacteria. Plenum Press, New York and
London.
Kleinbaum, D. G., and L. L. Kupper.
1978. Applied regression analysis and
other multivariable methods. Duxbury
Press, North Scituate, Mass.
Lenski, R.E. and J.E. Bouma. 1987.
Effects of segregation and selection on
instability of plasmid pACYC184 in
Escherichia coli B. J. Bacteriol.
169:5314-5316.
95
-------
TABLE 2. HOST-PLASM ID-MEDIUM COMBINATIONS3
WITH DIFFERENT PATTERNS OF PLASMID LOSS
Host
Plasmid
Medium
(Generation-1)b
Group A
(segregation alone)
PRS2015
PRS2015
PRS2015
PRS2015
PAO1C
PRS2015
PAO1C
pRO2313
pRO2317
pRO2321
PRO2318
PRO2321
pRO2320
pRO2318
PG
PG
PG
PG
PG
LB
PG
0.002
0.002
0.002
0.003
0.044
0.053
0.15
Group B
(segregation and selection)
PRS2015
PRS2015
PRS2015
PRS2015
PAO1C
PRS2015
PAO1c
PAO1C
pRO2321
pRO2317
pRO2318
pRO2320
pRO2317
pRO2313
pRO2313
PRO2320
LB
LB
LB
PG
PG
LB
PG
PG
-0.011
-0.002
0.006
0.017
0.037
0.055
0.069
0.15
a Combinations showing no statistical difference in segregation rates or loss
factors are clustered together.
b Median segregation rate u for Group A and median loss factor I for Group B.
96
-------
Mandel, M. and A. Higa. 1970.
Calcium-dependent bacteriophage DNA
infection. J. Mol. Biol. 43:159-162.
Maniatis, T., E.F. Fritsch and J.
Sambrook. 1982. Molecular Cloning.
Cold Spring Harbor Laboratory, Cold
Spring Harbor, New York.
Zylstra, G.J., S.M. Cuskey and R.H.
Olsen. In press. Construction of
plasmids for use in risk assessment
research, in: M. Levin, R. Seidler and
P. Pritchard (eds.) Classical and
molecular methods to assess
applications of
environmental
microorganisms.
PUBLICATIONS
Lenski, R. E. Fitness and gene stability.
in: M. Levin, R. Seidler, M. Rogul and
H. Pritchard (eds.), Classical and
Molecular Methods to Assess
of
Environmental Applications
Microorganisms. In press.
Nguyen, T.T., M. Patel and R.E. Lenski.
1991. Reduced growth rates due to
plasmid carriage cause instability of
pACYC184 and derivatives in
Escherichia coli K12. Submitted to J.
Bacterioi.
Lenski, R.E. 1991. Quantifying fitness
and gene stability in microorganisms.
in: Ginzburg, L.R. (ed), Assessing
Ecological Risks of Biotechnology, p.
173-192. Butterworth-Heinemann.
Nguyen, T.T., J. Lanners and R.E.
Lenski. 1990. How evolution and the
pSC101 par locus stabilize plasmids, p.
115-118. in: R. Seidler (ed.), Review
of Progress in the Biotechnology-
Microbial Pest Control Agent Risk
Assessment Program. U.S.
Environmental Protection Agency.
Lenski, R.E. and T.T. Nguyen. 1989.
Fitness and the fate of genetically
engineered microorganisms, p. 108-
116. in: P.H. Pritchard, B.L. Jackson,
J.E. Harvey and S.M. Martin (eds.),
Integration of Research and Predictive
Model Development for Biotechnology
Risk Assessment. U.S. Environmental
Protection Agency.
Nguyen, T.T., J. Lanners and R.E.
Lenski. 1989. How evolution and the
pSC101 par locus stabilize plasmids.
An abstract submitted to the EPA
Biotechnology All-Investigators Meeting
14-16 November 1989, Corvallis,
Oregon.
Nguyen, T.T. and R.E. Lenski. 1989.
Stability of EPA benchmark plasmids:
Effects of host background and growth
medium, Submitted In Proceedings of
the Review. U.S. Environmental
Protection Agency.
Lenski, R.E. and T.T. Nguyen. 1988.
Stability of recombinant DNA and its
effects on fitness, p. S18-S20. in: J.
Hodgson and A.M. Sugden (eds.),
Planned Release of Genetically
Engineered Organisms (Trends in
Biotechnology/ Trends in Ecology and
Evolution Special Publication). Elsevier
Publications, Cambridge.
97
-------
MODELING THE FATE OF BACTERIA IN SURFACE WATERS
John P. Connolly1, Richard B. Coffin2 and Robin E. Landeck1
Environmental Engineering & Science Program1
Manhattan College
Riverdale, New York
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Gulf Breeze, Florida
INTRODUCTION
Potential ecological and public health
impacts resulting from the release of a
genetically engineered microorganism
(GEM) to the environment are
dependent on the fate of the organism
and the relationship between organism
density or activity and a particular
effect. A fundamental consideration in
this regard is the ability of the organism
to survive and compete in a natural
setting and possibly transfer the
engineered genetic trait to other
organisms. Prediction of the fate of the
organism and its engineered trait in
natural systems is a major component
of a quantitative risk assessment.
Surface water systems are an
important environmental compartment
with regard to organism fate because
they typically provide a suitable habitat
for a variety of organisms and they
provide a means of rapid transport.
Predicting the fate of a GEM in a
surface water system requires an
analysis of substrate and nutrient
inputs, the response of the GEM and
the indigenous community to these
inputs and the impact of predation as a
population control factor. A necessary
component of this analysis is the
competition for resources that defines
the potential of the GEM to invade the
community. The analysis must also
describe the transfer of engineered
genetic material between the GEM and
indigenous organisms.
The overall objective of the research
being conducted in this project is the
step-wise development of the
components of a framework to model
the movement and growth of
genetically-engineered bacteria that,
either purposely or inadvertently, have
been introduced to a surface water
system. A primary goal has been to
develop a modeling framework capable
of predicting the long-term behavior of
bacteria in surface water systems. We
have made significant progress towards
this goal. We have completed
development of a framework which has
been successfully applied to a
microcosm study of bacterial dynamics.
The second goal is to extend the
framework to include competition
between bacterial populations. This
involves the division of bacteria into a
minimum of two groups (indigenous
community and introduced population).
Recombinant bacteria will be included
as a third group if gene transfer is
considered to be a potentially
significant process. Achieving this
objective requires the definition and
quantification of the mechanisms of
resource competition. Existing theory
98
-------
and past laboratory studies will be used
to establish working hypotheses that
will be evaluated through laboratory
experimentation to be conducted at the
Gulf Breeze EPA.
Ultimately the framework must be
expanded to incorporate mechanistic
descriptions of gene transfer. In
ongoing and recently completed
research, models have been proposed
to describe this process in simple
laboratory systems. These formulations
provide the basis for development of a
complete biotechnology risk assess-
ment modeling framework. Such a
framework must then be evaluated
based on its ability to predict the fate of
engineered genetic material. This
would be accomplished by application
to laboratory systems in which the
organisms have been introduced. Such
laboratory data are being generated
through the biotechnology program at
the Gulf Breeze EPA and evaluation of
the full modeling framework could be
con-ducted in a future cooperative
agreement.
RESULTS AND DISCUSSION
Bacterial Dynamics Modeling
Framework. The primary goal of
developing a .mathematical model of the
microbial food web capable of
predicting bacterial dynamics in surface
water systems has involved the
following tasks:
1) a review of bacterial growth
kinetics in laboratory systems,
2) development of a methodology
for defining bacterial substrate,
3) formulation of the equation set
describing the dynamics of the
microbial food web, and
4) application of the modeling
framework to a natural system.
We have completed the first three of
these tasks and are nearing completion
of the fourth. The major conclusions of
this work follows.
The utilization of dissolved organic
carbon (DOC) by pelagic bacteria is
dependent on the both the compounds
comprising the DOC and environment
specific growth characteristics of the
bacteria. Laboratory studies have
shown that growth yield coefficients
are related both to the energy content
of the compound and the availability of
other required nutrients such as
nitrogen. Half-saturation constants
describing substrate or nutrient
limitation of growth rate appear to be a
function of environment, higher values
being associated with higher nominal
substrate levels.
To make a modeling analysis of
bacterial growth tractable it is
necessary to define a scheme for
classifying DOC according to its ability
to be used as substrate. We have
proposed three categories of DOC: L1,
L2 and R carbon, in order of decreasing
lability. The fractionation of carbon into
these categories may be accomplished
through a BOD assay in which breaks in
the oxygen utilization curve are used to
signify the exhaustion of first the L1
and then the L2 component.
Development of the BOD assay
procedure is continuing at the Gulf
Breeze EPA. A methodology has been
established which appears to have the
needed sensitivity and replicability.
Samples (unfiltered and innoculated
sterile filtered) are incubated at in situ
99
-------
temperature in the dark and oxygen
utilization is monitored at close (1-2
hour) intervals. A change in slope of
the oxygen utilization curve is used to
define exhaustion of L1 carbon. Total
oxygen utilization is used to establish
the degradable (L1 +L2) carbon. DOC
measurements establish the
oxygen:carbon stoichiometry and
bacterial counts (AODC) confirm the
exhaustion of substrate and cessation '
of growth. This methodology is being
applied to natural water samples,
samples spiked with various carbon
sources (e.g., algal exudate, sea grass
detritus) and will be applied to
terrestrial carbon in non-point runoff
and to treatment plant effluent.
Preliminary laboratory assay data
suggest that the L1 substrate pool is
small. The bacteria are able to readily
use this substrate at maximum growth
rates calculated to be about 4/d.
Because of the small size of this
substrate pool and its rapid use it is
exhausted in the assays within a time
scale of hours or even minut'es.
A review of experimental studies of
the fate of carbon produced through
primary production indicates that a
significant fraction of the DOC released
from phytoplankton through exudation,
cell lysis and sloppy feeding by grazers
would be classified as L1 substrate.
Further, a portion of the detrital POC
generated through grazing appears to
be rapidly solubilized to L1 DOC.
A comprehensive review of
zooplankton bioenergetics has been
used to establish equations defining
trophic transfer and loss of biomass
carbon. Allometric relationships
between consumption rate and body
weight and between respiration rate
and body weight are being used to
define parameter values for the three
grazer levels (nano-, micro- and
mesozooplankton) included in the
: framework.
Use of the L1-L2 substrate
categorization, routing phytoplankton
carbon to these classes in accordance
with published experimental results and
specification of bacteria and grazer
dynamics according to a Monod
formulation successfully modelled the
microbial dynamics of a microcosm
study.
Competition
The expansion of the modeling
framework to include GEMs entails
quantification of competition between
the GEM and the indigenous population
and-gene transfer between them. We
have focused on modeling competition
between bacterial groupings. Growth
of each group is defined using the
Monod equations. Competition results
through interaction with the substrate
pools as defined by growth rates, yield
coefficients and half-saturation
constants. In addition, differences in
predation, as defined 5by group specific
grazing rates, are also considered.
Coexistence of an introduced organism
and the indigenous community or the
isogenic counterpart of the GEM
depends on each of the coefficient
values and substrate specificity. It may
be necessary to model multiple
substrates to adequately define the
potential of the GEM to survive and to
invade the community.
100
-------
FUTURE WORK
We are continuing to review the
literature on zooplankton growth and
grazing. We have found that the
allometric relationship we used for
maximum ingestion rate underestimated
rates reported for microzooplankton.
We are now using an alternate
relationship with published growth rate
data to estimate the assimilation
efficiency of ingested carbon. Once
this has been completed we will
attempt to recalibrate the Bagsvaerd
microcosm model with the new
parameters.
A model of bacteria in the Delaware
River estuary has been formulated. It
consists of .the bacterial kinetic
equations which we previously
developed and empiricalspecificationof
phytoplankton dynamics. Observed
primary production rates define
substrate input to the system.
Observed phytoplankton biomass
concentrations define algal prey
concentrations for the zooplankton
grazers. Segmentation of the estuary
and bay has been completed and we
are currently calibrating transport over
the four month period being modeled
(2/18/85 to 5/25/85). Salinity data
from the five cruises conducted during
this period are being used as the
transport tracer. When this is
completed we will begin to model the
bacterial dynamics that have been
.observed in this system.
The laboratory experiments to
examine bacterial uptake of various
carbon sources (i.e., the substrate
utilization assay) has begun to yield
data. We will analyze these data to
determine the distribution of carbon
between operationally defined substrate
categories and bacterial growth on
these.
We will continue to examine
competition between bacteria to better
define the processes controlling survival
and growth of an introduced organism.
We will continue to prepare
manuscripts covering research
conducted during this project. The first
manuscript has just been completed
and a manuscript describing the
Bagsvard microcosm modeling study
will be started shortly.
An abstract • of a platform
presentation of our modeling work has
been submitted for consideration for
inclusion in, the- -Fifth-International
Workshop oh the Measurement of
Microbial Activities in the Carbon Cycle
in Aquatic Environments "Microbial
Ecology of Pelagic Environments" to be
held from August 18-23, 1991 in
Helsingor, Denmark;
PUBLICATION
Connolly, J.P., R. B. Coffin and R. E.
Landeck. 1991. Modeling Carbon
Utilization by Bacteria in Natural
Waters. To be published in Modeling
the Metabolic and Physiologic Activities
of Microorganisms (Christen Hurst,
Ed.), John Wiley, New York.
101
-------
MODELING TRANSDUCTION IN AQUATIC ENVIRONMENTS
Robert V. Miller1, Tyler A. Kokjohn2, and Gary S. Sayler3
Department of Microbiology1, Oklahoma State University
Stillwater, Oklahoma
Environmental Research Division2, Argonne National Laboratory
Argonne, Illinois
Center for Environmental Biotechnology3, University of Tennessee
Knoxville, Tennessee
INTRODUCTION
With the recent advances of genetic
engineering have come new questions
about genetic stability and exchange
among microbes in natural ecosystems.
We are using Pseudomonas aeruginosa
as a model organism to study viral-
mediated gene transfer (transduction) in
freshwater microbial populations. Our
studies have revealed a significant
potential for transduction of both
plasmid and chromosomal DNA in these
environments. Initially, we explored
three models for the source of
transducing particles in freshwater
environments: (a) cell-free lysates of
bacteriophages grown on an
appropriate DNA donor, (b) environ-
mental induction of bacteriophages
from a lysogenic DNA donor bacterium,
and (c) environmental induction of
bacteriophages from a lysogenic
recipient bacterium. The transfer of
plasmid and chromosomal DNA was
documented in each of these systems
in situ. The highest number of
transductants were routinely recovered
from systems where the recipient
bacterium was a lysogen, probably due
to the immunity imparted by the
resident prophage. Transduction was
observed in both the absence and
presence of the natural microbial
community. Reciprocal chromosomal
transduction was observed in chambers
inoculated with two lysogens.
Apparently, both primary infection of a
non-lysogen and prophage induction
from a lysogen can generate sufficient
numbers of transducing particles to
allow gene exchange to be observed.
For transduction to take place in this
system a unique sequence of events
must take place, (a) Phage virions must
be produced through spontaneous or
stress-stimulated induction of the
prophage from the lysogen. (b) These
viral particles must infect, propagate,
and lyse the plasmid-containing donor.
(c) Transducing particles produced
during this lytic infection must absorb
and transfer DNA to the remaining
lysogens. Hence, environmental
lysogens serve as both efficient sources
of transducing phages and as viable
recipients for transduced DNA.
The finding that transduction can
occur in natural environments is
significant as this mechanism of gene
transfer has been virtually ignored in
both the design and preliminary testing
of genetically engineered micro-
organisms for environmental release.
However, the ultimate question which
must be addressed in determining the
effects of environmental transduction is
102
-------
whether phage-mediated horizontal
gene transfer alters the probability or
equilibrium frequency of an introduced
genetic sequence in a natural bacterial
community. We are now turning our
attention to developing a model
transduction system whose
components are derived directly from
the environment and to determining
whether transduction is effective in
altering the makeup of the gene pool
available to natural populations of
bacteria. These studies will allow a
more accurate assessment of the real
potential for genetic exchange to affect
natural microbial populations.
METHODS
Our investigations of horizontal gene
transfer are conducted in both
laboratory and field formats and can be
divided into three stages. First, the
potential for transmission is evaluated
by standard genetic protocols. Second,
microcosm studies are conducted in the
laboratory to evaluate protocols to be
tested in situ. Third, field trials are
conducted in biological containment
chambers incubated at the freshwater
field site. This stage is used to validate
predictions made from the laboratory
simulation experiments. The potential
for transduction in freshwater P.
aeruginosa populations was evaluated
using a variant of the generalized
transducing phage F116 originally
designated DS1. For studies of
plasmid transfer, the Tra- Mob" plasmid
Rms149 is used. Plasmid transduction
was confirmed by molecular analysis.
Detailed methods have been published
(see below). Natural bacteriophages
were isolated from lake water collected
at several of our test sites. Samples
were used either directly or con-
centrated. Phages were identified by
their ability to plaque on one or more
indicator strains. Phages selected for
further studies were purified by glycerol
gradient ultracentrifugation. P.
aeruginosa strains were isolated from
lake water by means of selective media
and Pseudomonas Isolation Agar (Difco,
Detroit, Ml) and further characterized as
appropriate. In some experiments,
mixed populations of either bacterio-
phages or bacteria isolated from the
field sites were used. Lysogens were
identified by their ability to release
phage either spontaneously or following
exposure to UV radiation and by their
superinfection-immunity phenotype.
Continuous culture experiments are
conducted in a New Brunswick
Scientific (Edison, NJ) BioFlo
Chemostat. Culture medium consists
of Pseudomonas Minimal Medium Salts
without citrate supplemented with
yeast extract (10~5 to 10"7 g/ml).
RESULTS AND DISCUSSION
Developing a
Natural Transduction Model
While our research has demon-
strated the potential for transduction in
natural freshwater habitats, it is still
artificial as its components are well
characterized laboratory strains. We
are currently attempting to develop a
model system based on components
isolated directly from the environment.
To begin this study, we
demonstrated the occurrence of
bacteriophages, potential hosts, and
lysogens in the aquatic environment.
103
-------
We monitored these parameters in our
field sties for an eight month period.
During this period, we found total
bacterial counts rangee from 105 to
107 CFU/ml and Pseudomonas counts
from 5 x 102 to 5 x 104 CFU/mi. Titers
of bacteriophages capable of forming
plaques on laboratory strains of P.
aeruginosa ranged from 101 to 104
PFU/ml. When lysogeny was evaluated
using a laboratory strain of P.
aeruginosa as a indicator, between 1
and 7% of the Pseudomonas isolates
tested positive. However, we found
that approximately 45% of
Pseudomonas isolates from our field
site tested positive in colony
hybridization when probed with DNA
from a naturally occurring phage
isolate.
Second, we isolated transducing
phages from our environmental
samples. One of these, UT1, has been
studied in some detail. This phage was
isolated from Fort Loudoun Lake near
Knoxville, Tennessee where we have
carried out the majority of our in situ
transduction experiments. Although
this phage is virulent under laboratory
conditions, it establishes pseudo-
lysogeny in the environment. It is a
generalized transducing phage capable
of mediating the transfer of both
plasmid (Rms149) and chromosomal
DNA among P. aeruginosa. As might
be expected with a virulent phage, the
highest levels of transduction are
observed at very low MOIs (10"3
PFU/CFU). Using standard laboratory
protocols, both chromosomal and
plasmid transduction was detected at
frequencies as high as 10~4
transductants/PFU depending on the
MOI used.
Third, we have used UT1 to initiate
studies to determine the potential of the
natural microbial community present at
our field sites to act as recipients of
transduced plasmid DNA. Cells from
10 liters of lake water were concen-
, trated and used as a recipient pool for
transduction using cell-free lysates of
UT1 grown on a>P. aeruginosa strain
containing Rms149. Significant levels
of transduction were observed and
confirmed molecularly. These
observations have been made using
samples from several different field
sites collected at different times of
year. They suggest that the natural
microbial community at our field sites
contains organisms capable of acting as
hosts and recipients for plasmid
transduction.
For transduction to take place,
interaction between the transducing
phage and its host must take place.
We examined, the ability of bacterio-
phages to interact with their hosts at
low cell densities and under starvation
conditions. The attachment and
replication of three P. aeruginosa
bacteriophages were investigated under
conditions similar to those found in
nature. Attachment and replication of
bacteriophages were not impaired at
host-cell densities equal to or lower
«105 CFU ml"1} than those
frequently found in aquatic environ-
ments when the host cells were
physiologically competent to allow
phage growth. .Attachment to either
actively growing or starved cells was
not impaired in river water, indicating
that attachment is efficient in natural
freshwater habitats. However, the
replication of bacteriophages was
significantly altered .in. starved cells in
104
-------
river water; the latency period was
extended (broth, 70-110 min.; river
water, 110-240 min.), and the burst
size was reduced (broth, 27-65; river
water 5-7). The findings of this study
indicate that phages are likely to affect
microbial ecology significantly in
freshwater ecosystems.
We isolated strains of P.- aeruginosa
which were sensitive and resistant to
UT1 from Fort Loudoun Lake and used
them to monitor host-phage interactions
over 45 days in lakewater microcosms.
Temporal changes in host density,
phage-to-bacterium ratio (PBR), and the
appearance of apparent prophage
carriers within the host population were
analyzed. About 45% of sensitive
bacteria incubated with phage UT1
were pseudolysogenic within 12 hours
of incubation in natural lake water.
Phage UT1 appeared to stabilize the
density x>f host bacteria in lake water at
a level of 104 CFU/ml. Bacterial
coexistence with a mixed phage
population isolated from Fort Loudoun
Lake resulted in an oscillating
equilibrium with the PBR stabilizing at
about 3. The presence of extraneous
homoimmune phages appeared to be
detrimental to the stability of the
pseudolysogens, which were
maintained at a lower population
density than prophage-free cells in lake
water containing the mixed phage
population.
A Mathematical Model
for Transduction
We are currently developing a
mathematical model for transduction.
We have convincingly shown that
transduction can act to stabilize (and
even increase) the frequency of a
genotype in a population from which
that genotype would otherwise be lost.
Our research to date suggests that, in
populations of bacteria in which
transduction is occurring, the change in
the frequency of a transducable
genotype (Pht) as a function of
generation of growth "g" can be
described by the formula:
dPht/dg = t - s
Where "t" is the fraction of new cells
added to the "Pht" genotype due to
transduction and "s" is the fraction due
to selection.
FUTURE WORK
In the future, we plan to evaluate our
natural model system in situ by
introducing a plasmid donor (either a
lysogen or pseudolysogen) and allowing
the natural community to act as
recipients. We are currently developing
an appropriate contraselection for the
introduced donor which will allow us to
identify transductants in the natural
population. In addition, we are
determining the physiological, environ-
mental, and genetic factors which
regulate "t." Preliminary analysis of our
data suggest that "t" cannot be
described simply as a mass-action
phenomenon but contains parameters
which are highly dependent on the
physiological state of the hosts. We
will continue to explore these factors in
the immediate future. '
105
-------
PUBLICATIONS
.Kokjohn, T.A., G.S. Sayler, and R.V.
Miller. 1991. Attachment and
Replication of Pseudomonasaeruginosa
bacteriophages under conditions
simulating aquatic environments. J.
Gen. Microbiol. 137:661-666.
Kokjohn, T.A., and R.V. Miller. 1991.
Gene transfer in the environment:
transduction. JnJ.C. Fry and M.J. Day
(eds.), Release of Genetically
Engineered and Other Microorganisms.
Edward Arnold, London, in press.
Miller, R.V. 1991. Genetic Stability
and Fate Concerns on Environmental
Release, in: A.P. Hugenjoltz (ed.),
Contemporary Issues in Toxicology.
Proc. Twenty-Third Symp., Soc.
Toxicol. Canada. Montreal, in press.
Miller, R.V. 1991. Methods for the
evaluation of gene transfer in the
environment; general considerations.
in: M. Levin, R. Seidler, and M. Rogul
(eds.), Microbial Ecology: Principles.
Methods, and Application in
Environmental Biotechnology. McGraw-
Hill, N.Y., in press.
Miller, R.V. 1991. Transduction in
natural environments, in: M. Levin, R.
Seidler, and M. Rogul (eds.), Microbial
Ecology: Principles, Methods, and
Application in Environmental
Biotechnology. McGraw-Hill, N.Y., in
press.
Saye, D.J., and S.B. O'Morchoe.
1991. Evaluating the potential for
genetic exchange in natural freshwater
environments, in: M. Levin, R. Seidler,
and M. Rogul (eds.), Microbial Ecology:
Principles, Methods, and Application in
Environmental Biotechnology. McGraw-
Hill Pub.,.N,Y,, in press.
Miller, R.V., and G.S. Sayler. 1991.
Bacteriophage-host interactions' in
aquatic systems, in: E.M. Wellington
and J.D. VanElsas (eds.), Genetic
Interactions Between Microorganisms in
the Microenvironment. Univ.
Manchester Press, Manchester, in
press.
Ogunseitan, O.A., G.S/ Sayler, and
R.V. Miller. 1990. Dynamic interaction
of Pseudomonas aeruginosa and
bacteriophages in lake water. Microb.
Ecol. 19:171-185.
Simonson, C.S., T.A. Kokjohn, and
R.V. Miller. 1990. Inducible UV repair
potential of Pseudomonas aeruginosa
PAD. J. Gen. Microbiol. 136:1241-
1249.
Saye, D.J., O.A. Ogunseitan, G.S.
Sayler, and R.V. Miller. 1990.
Transduction of linked chromosomal
genes between Pseudomonas
aeruginosa during incubation in situ in a
freshwater habitat. Appl. Environ.
Microbiol. 56:140-145.
Miller, R.V. 1990. Increasing levels of
environmental mutagens: potential fbr
affecting viral evolution and
pathogenicity-a speculative review.
Environ. Carcinogen. Rev. 08:89-137.
Miller, R.V., and T.A. Kokjohn. 1990.
Microbiology and evolution of the recA
gene. Annu. Rev. Microbiol. 44: 365-
394!
106
-------
Miller, Robert V., Tyler A. Kokjohn, and
GaryS. Sayler. 1990. Genetic Transfer
in Freshwater Environments, p. 72-77.
in: Review of Progress in the
Biotechnology-Microbial Pest Control
Agent Risk Assessment Program. U.S.
EPA, EPA/600/9-90/029.
Levy, S.B., and R.V. Miller. 1989.
Gene Transfer in the Environment,
McGraw-Hill, NY.
Saye, D.J., and R.V. Miller. 1989.
Gene transfer in aquatic environments.
p. 223-254. In S.B. Levy and R.V.
Miller (eds.), Gene Transfer in the
Environment. McGraw-Hill, NY.
Miller, R.V., and S.B. Levy. 1989.
Horizontal gene transfer in. relation to
environmental release of genetically
engineered microorganisms, p. 405-
420. In: S.B. Levy and R.V. Miller
(eds.), Gene Transfer in the
Environment* McGraw-Hill, NY.
107
-------
ENVIRONMENTALLY INDUCED GENETIC INSTABILITY IN MICROORGANISMS
Tyler A. Kokjohn1 and Robert V. Miller2
Environmental Research Division1, Argonne National Laboratory, Argonne, Illinois
Department of Microbiology2, Oklahoma State University, Stillwater, Oklahoma
INTRODUCTION
Organisms must maintain the
integrity of their genetic material if
species survival is to be assured. It is
now apparent that microorganisms
possess DNA repair systems which,
although differing in capabilities and
potentials, act to repair damage
resulting from cell metabolism or
environmental insult.
Most of the data available concerning
DNA repair has utilized the enteric
bacterium Escherichia coli under
conditions which probably will rarely, if
ever, occur in natural environments.
However, some important principles
have been revealed by these studies. It
is clear that bacterial cells suffering
DNA damage will undergo repair
processes. In some cases, these repair
systems are mutagenic. Further, the
efficiency and activity of DNA repair
systems will be different under differing
growth conditions.
From work already performed, there
can be no doubt that genetic systems
exist in the eubacteria which are
induced by environmental stress
resulting in enormously increased rates
of mutagenesis of the genetic material.
The existence of such potentials for
genetic instability complicates the task
of risk assessment for environmental
releases of genetically engineered
microorganisms (GEMs) and raises
questions concerning the stability of
introduced genotypes in natural
ecosystems. We have been utilizing
Pseudomonas aeruginosa as a model
system to study stress-induced genetic
alterations in bacterial cells. The use of
this model will allow a more accurate
assessment of risk associated with
GEM releases in the future.
METHODS
The frequency of mutation-containing
cells in a population of P. aeruginosa
was determined for cells incubated in
growth medium in the laboratory and
for cells placed in
in biological containment chambers and
incubated either in the laboratory or in
situ at our established field sites. The
chambers utilized in these studies are
permeable to gases and allow the
transmittance of solar UV radiation.
They were filled with sterilized lake
water from our field sites and
inoculated with genetically well
characterized P. aeruginosa strains.
Chambers incubated in situ were placed
at locations receiving solar UV radiation
for significant periods of the day or in
locations substantially shielded from
sunlight.
Cells were recovered from chambers
by plating on one-tenth strength YEPG
agar. Viable counts were determined
for each sampling period. In order to
determine the frequency of mutation-
containing cells in the population,
samples (0.1 ml) were plated on 0.1 X
YEPG agar and incubated for 48 hours
108
-------
to recover cells. The cells were then
replica plated onto various selective
media. Cells were screened for
streptomycin resistance, rifampicin
resistance and, in some cases,
resistance to infection by the virulent
phage UT1. The prptocols for
determination of drug resistance have
been modified to include another round
of replica plating on nonselective
medium before recovered cells are
plated on media containing antibiotics.
Individual drug resistant clones used for
further experimentation were carefully
streaked for isolation. Extra replica
plating steps are required to avoid
artifactual contributions to the apparent
mutation rate of cells that are
physiologically adapted to starvation
conditions but, contain no genetic
mutations to drug resistance.
The ability of cells existing under
starvation conditions to resist stress
was determined by laboratory experi-
mentation. Cells were grown in 0.1 X
YEPG at room temperature to stationary
phase. The cells were diluted 1:10 in
sterile water from a field site and
incubated at room temperature for
various periods of time. After a period
of such starvation, the cells were
irradiated with UV, and the percentage
survival was determined for several
doses of UV radiation.
Induction of therecA gene product of
P. aeruginosa and £. co//was examined
under several growth conditions using
the Western Blotting technique. Cells
were grown in Luria Broth, low-strength
medium (0.01 X YEPG) or starved as
described above. They were then
irradiated with UV light. Subsequent to
UV exposure, the cells were incubated
in the dark and samples were taken
periodically to be used for Western
Blotting. These samples were
electrophoresed on denaturing
polyacrylamide gels and the proteins
transferred to nitrocellulose
membranes. These membranes were
subjected to Western analysis using a
biotintylated-streptavidin-horseradish
perox id ase-conju gated antibody
detection system. Anti-rec/4 E. coli
antibody was used as the primary
antibody.
Fusions of a promoterless P-
galactosidase gene to stress-inducible
genes of P. aeruginosa PAO were
constructed using a modified Tn3
transposon. This transposon will insert
into target DNA yielding transcriptional
fusions of the disrupted gene and P-
galactosidase.
For prophage induction experiments,
F116L or D3 lysogens of P. aeruginosa
were placed in chambers and incubated
in the laboratory or in situ at sites
receiving solar UV irradiation. The ratio
of infectious centers (1C) to total
colony-forming units (CFU) was
determined to detect induction of the
prophages. Cell concentrations were
determined as described above. ICs
were enumerated by dilution in 0.85%
NaCI and plating with phage-sensitive
indicator strains. In some laboratory
experiments, lysogens of E. coli were
treated in a similar manner.
For starvation induction experiments,
a D3 lysogen of P. aeruginosa was
grown to early exponential phase in
Luria broth. The cells were harvested
by centrifugation and suspended in
sterile river water to induce a putative
stringent response. 1C determinations
were made at various times after
medium shift.
109
-------
RESULTS AND DISCUSSION
Mutation Frequency
Our original data suggesting
increased rates of mutagenesis in situ
were based on the reversion
frequencies of amino acid auxotrophs.
Although no apparent selection for the
prototrophs could be detected, we were
concerned that unknown factors were
acting to increase the reversion
frequency. Therefore, we have spent
some time evaluating our procedures
for determination of mutation
frequency. The experimental protocols
which we are now utilizing reflect
experience gained from our field
experiments. Specifically we have
found that expression time after
sampling the cells is required to detect
mutation events. It has proven best to
recover cells on low-strength medium
(0.1 X YEPG) and apply selection later
by replica plating onto appropriate
media. The use of the replica plating
technique allows a one-to-one mutation
to resultant colony detection and allows
testing of identical populations for the
frequency of several different mutant
phenotypes.
We have modified our screening
procedures for detection of drug
resistant mutants to include a longer
recovery time before selection is
applied. Cells undergoing starvation
stress apparently adapt physiologically
in such a way that their apparent drug
resistance is increased. This change is
not necessarily genetic in nature. We
have clearly been able to isolate
Rifampicin-resistant mutants of at least
two distinct classes. By transductional
analysis we have determined one class
is resistant only to Rifampicin, the other
class is resistant to several drugs
simultaneously. The exact frequency of
production of Rifampicin-resistant
mutations subsequent to stress
application is unknown at present. Our
previous estimates were confounded by
artifactual contributions from adapted
cells not containing true genetic
alterations.
Reversion of amino acid auxo-
trophiessubsequentto starvation stress
has proven to be reproducible. Under
conditions of no apparent selection
advantage for revertants, we find rapid
production of prototrophic cells. It is
unknown if these are true revertants,
second site suppressor mutations, or
expression of normally cryptic genes.
Several different mutant phenotypes
have been examined. Screening for
rifampicin resistance (Rifr) has proven
to be ideal since Rifr bacteria are
infrequent at our field sites. The use of
streptomycin resistance is also
convenient. Mutation to phage
resistance is also an excellent marker.
For these studies, we have used the
extremely virulent environmental phage
UT1. This phage is ideal- since it is
absorbed by and lyses even starved P.
aeruginosa cells. Other virulent
laboratory phages were unacceptable
since changes in growth condition
resulted in declined virulence or
reproduction potential.
Enhanced Resistance to UV Stress
The effects of starvation on the
response of bacterial cells to DNA
damage are dramatic. The physio-
logical adaptations required for viability
maintenance result in a tremendously
110
-------
enhanced ability to resist UV ir-
radiation. Preliminary experiments
suggest that the genes required for
DNA repair (recA, uvrABC, etc. in E
co//) are expressed at enhanced,
constitutive levels. Western Blotting
experiments performed on both E. coli
and P. aeruginosa support this
conclusion.
Induction of Prophages
Laboratory and field experiments
have also demonstrated that starvation
of E. coli and P. aeruginosa lysogens
greatly affects DNA-damage induction
of resident prophages. Preliminary
experiments using the temperate P.
aeruginosa bacteriophage F116 have
suggested that conditions encountered
in natural aquatic environments may
lead to enhanced activation of lytic
phage production in these cells. One
important question is whether this is a
general characteristic of all phages.
Our initial experiments have used three
UV-inducible temperate phages: D3 and
F116 of P. aeruginosa and of E. coli.
Field experiments have demonstrated
no increased induction of D3 prophages
from lysogens of P. aeruginosa exposed
to significant, long term solar irradiation
as measured by the frequency of ICs in
the population. Laboratory investi-
gations have confirmed that starved
cells lose the ability to support UV
induction of prophage. This
phenomenon has been observed for
both E. coli and P. aeruginosa and has
been termed "aptitude" by Andre
Lwoff. Significantly, we find that
starved lysogens of P. aeruginosa are
capable of the spontaneous production
of phage particles for long periods. In
contrast, lysogens of E. coli rapidly
become inviable under the same growth
conditions.
Stress-lnducible Expression
of P. aeruginosa Genes
To aid in the quantification of levels
of expression of DNA-damage inducible
(din] genes under environmental
conditions such as starvation and
exposure to solar UV, we have
constructed and partially characterized
transcriptional fusions of various P.
aeruginosa PAO genes to (5-galacto-
sidase (/acZ) which are inducible by
exposure to DNA damaging agents.
These genes are clearly expressed at
greatly increased levels subsequent to
UV irradiation of Rec+ P. aeruginosa
and are inducible by other agents which
damage DNA as well in this species.
They are not inducible when transferred
into E. coli. The molecular mechanism
of this anomaly is currently under
investigation.
Plasmid-Encoded UV Resistance
Laboratory studies of the UV-
resistance plasmid R2 have revealed
that it encodes a true /-ec/4-dependent,
DNA damage-inducible repair system
which is highly mutagenic. Laboratory
and field studies of P. aeruginosa cells
containing the UV-resistance plasmid
R2 have demonstrated an enhanced
mutagenesis activity under environ-
mental conditions which result in
exposure to solar-UV radiation. An
understanding of the contribution of
these plasmids to genetic instability of
natural populations of bacteria and their
effects on novel genetic sequences
111
-------
introduced into these environments by
the release of GEMs is absolutely vital
in risk and fate assessment.
FUTURE WORK
We will continued to investigate the
existence of a hypermutable state in
cells undergoing starvation stress.
Auxotrophic revertants will be mapped
to better characterize their nature. We
will continue to examine the expression
of UV resistance plasmids exemplified
by R2.
PUBLICATIONS
Kokjohn, T.A., G.S. Sayler, and R.V.
Miller. 1991. Attachment and
Replication of Pseudomonas aeruginosa
bacteriophages under conditions
simulating aquatic environments. J.
Gen. Microbiol'. 137:661-666.
Kokjohn, T.A., and R.V. Miller. 1991.
Gene transfer in the environment:
transduction. In: J.C. Fry and M.J. Day
(eds.), Release of Genetically
Engineered and Other Microorganisms.
Edward Arnold, London, in press.
Miller, R.V. 1991. Increasing levels of
environmental mutagens: Potential for
affecting viral evolution and
pathogenicity -- a speculative review.
Environ. Carcinogen. Rev. C8:89-137.
Miller, R.V. 1991. Genetic Stability
and Fate Concerns on Environmental
Release, in: A.P. Hugenjoltz (ed.),
Contemporary Issues in Toxicology.
Proc. Twenty-Third Symp. Soc. Toxicol.
Canada, Montreal, in press.
Simonson,; C.S., T.A. Kokjohn, and
R.V. Miller. 1990. Inducible UV repair
potential of Pseudomonas aeruginosa
PAO. J.Gen.Microbiol. ,136:1241 -1249.
Miller, R.V., T.A. Kokjohn, and G.S.
Sayler. 1990. Environmental and
molecular characterization of systems
which affect genome alteration in
Pseudomonas, pp. 252-268. in: S.
Silver, A.M. Chakrabarthy, B. Iglewski,
and S. Kapland (eds.), Pseudomonas:
Biotransformations. Pathoaenesis. and
Evolving Biotechnology. Amer. Soc.
Microbiol., Washington, D.C.
Miller, R.V. ,1990. Estimating the
stability of gene inserts: Microbial
systems, p. 61-72. in: D. Mahon (ed.),
First Consultative Workshop on
Assessing Safety in Foods Derived
Through Biotechnology. Health and
Welfare Canada, Ottawa.
Miller, R.V., and T.A. Kokjohn. 1990:
General Microbiology of recA: Environ-
mental and Evolutionary Significance.
Annu. Rev. Microbiol. 44: 365-394.
Kokjohn, T.A., and R.V. Miller. 1990.
Environmentally induced genetic
instability in microorganisms, p. 110-
114. in: Review of Progress in the
Biotechnoloav-Microbial Pest Control
Agent Risk Assessment Program. U.S.
Environmental Protection Agency
A/600/9-90/029
Kokjohn, T.A. 1989. Transduction:
mechanism and potential for gene
transfer in the environment, p. 73-99.
in: S.B. Levy and R,V. Miller (eds.),
Gene Transfer in the Environment.
McGraw-Hill, N..Y.
112
-------
GENOMIC PLASTICITY AND CATABOLIC POTENTIAL OF
Pseudomonas cepacia
T.G. Lessie, A. Ferrante, A. Byrne, M.S. Wood, and H.-P. Cheng
Department of Microbiology
University of Massachusetts
Amherst, Massachusetts
INTRODUCTION ,
The genome of the bacterium
Pseudomonas cepacia contains a large
number of insertion sequences. The
abilities of these elements to mediate
genomic rearrangements and activate
the expression of neighboring genes
have been implicated in the evolution of
novel catabolic functions and appear to
underlie the extraordinary nutritional
versatility and adaptability' of this
bacterium. We have been particularly
interested in the roles of IS elements in
the recruitment of foreign genes for
new degradative pathways as well as
their capacity to be transferred to and
transpose and activate gene expression
in other bacteria.
In order to gain information about the
relationship of P. cepacia IS elements
and insertion sequences from other
bacteria we have determined the
nucleotide sequences of \S401, \S402,
\S406, and \S407 and compared them
to elements listed in. the GenBank and
EMBL data bases. The four elements
examined had been shown to promote
rearrangements of a cryptic plasmid
present in P. cepacia 249 or to activate
the expression of the lac genes of
Tn357 when this transposon was
introduced into P. cepaciaon the broad-
host-range plasmid 6GC91.14. The
data obtained has also been used to
construct \S402 and IS407 variants
carrying the trimethoprim resistance
(TpR) gene of pR388, which are being
used to examine factors influencing
transpositional activity.
To gain information about the
organization of the P. cepacia genome
we have undertaken the construction of
a physical map of the chromosome of
strain 249. Our aims are to examine
the distribution of IS elements and
genes encoding key catabolic enzymes.
Such information should provide
insights into the bases for the unusual
catabolic potential of this bacterium.
METHODS
Analyses Of the Nucleotide Sequences
of Selected IS Elements
Insertion sequences were cloned
from element-containing derivatives of
pRP1 and pGC91.14 into the
sequencing vector pBLUEKSP and
inserts of suitable length for sequence
analysis were generated by subcloning
DNA fragments or by creating nested
deletions. Nucleotide sequence data
was obtained by the dideoxynucleotide
procedure of Sanger using ccc-DNA as
template. The data were compiled and
analysed on a MicroVAX computer
using programs provided by the
Genetics Computer Group at the
University of Wisconsin.
113
-------
Construction of Marked IS Elements
The nucleotide seque/ices of \S402
and \S407 were examined to identify
unique restriction enzyme sites located
outside open reading frames and
terminal inverted repeat sequences that
might be important for transposition.
Two such sites were identified: a Dra\\
site within \S402 and a Dra\ site within
\S407. A 0.7-kb trimethoprim
resistance (TpR) cassette from pR388
was ligated into these sites to form the
derivatives \S402.2 and \S407.1, which
each carry the dihydrofolate reductase
(fol) gene of pR388, and accordingly
confer high level resistance to
trimethoprim.
Physical Mapping of the Chromosome
of P. ceoacia 249
Bacteria were immobilized in agarose
plugs and lysed under alkaline
conditions by treatment with SDS and
proteinase K. The plugs were washed
in TE buffer and the chromosomal DNA
was digested in situ with restriction
enzymes such as Pac\, Afl\\, Xba\, and
Dra\, which cleaved the P. cepacia
chromosome into a relatively small
number of high molecular weight
fragments. The agarose plugs were
transferred into wells on agarose gels,
and the DNA fragments were resolved
by pulsed-field gel electrophoresis using
a CHEF gel apparatus obtained from
Owl Scientific Co., Cambridge, MA.
Yeast chromosomes and concatamers
of coliphage lambda DNA were used as
size standards. Neighboring Pad
fragments were identified by Southern
hybridization experiments using as DNA
probes Af/ll and Xba\ fragments which
overlapped the pertinent Pac\ sites.
RESULTS AND DISCUSSION
Relationship between Elements from
P. ceoacia and other Bacteria
The results of our nucleotide
sequence analyses indicate that \S401
and \S407 are members of the IS3
family of insertion sequences, whereas
\S402 and \S406 are distinct from
previously described elements. The
results indicate that at least some of
the IS elements present in P. cepacia
have been widely dispersed among
bacteria. Table 1 compares some of
the characteristics of \S401 and \S407
and other IS3-like elements. \S407,
\S476, and IS/?7 comprise a subgroup
of the IS3 family. All three of these
elements produced 4-bp duplications of
target DNA upon insertion. IS407 was
most closely related to \S476. The
deduced amino acid sequences of the
major open reading frames of these two
elements exhibited between 72 and
82% similarity. \S407 was typical of
the majority of IS3-like elements in that
it generated 3-bp target duplications.
\S40J exhibited a high degree of
homology with IS57 and with a region
of the Ti plasmids of Agrobacterium
tumefaciens important for tumor
induction in plants. In this context it is
of interest that certain Ti plasmids
confer ability to utilize octopine, a
compound which supports rapid growth
of P. cepacia. It is possible that \S401
might have been involved in acquisition
of octopine utilization genes.
114
-------
Table 1. Characteristics and sources of IS3-like elements
Element
IS407
\S476
\SR1
\S401d
\S51
\S3411
IS3
\S911
Size
(bp)
1236
1225
1260
1316
1311
1309
1258
1250
IRsa
(bp)
12
26
13
26
26
27
39
27
Target
repeats'3
4
4
4
3
3
3
3
3
Percent0
Hornology
100
62
55
44
40
39
38
38
Source
Pseudomonas cepacia
Xanthomonas campestris
Rhizobium lupini
P. cepacia
Pseudomonas syringae
E. coll
E. coli
Shigella disenteriae
Terminal inverted repeats
Directly repeated duplications of target site DNA
% Overall nucleotide sequence homology with IS407
\S401 was most closely related to IS57 (65% homology)
Transposition of \S402 and \S407
Variants in E. coli
Donor strains containing pBluescript
vectors carrying \S402.2 and \S407,1
(variants conferring resistance to
trimethoprim) as well as IncPI plasmids
conferring resistance to tetracycline
were mated with E. coli and P. cepacia
recipients, Transconjugants containing
cointegrate plasmids were obtained
readily. \S402.2, a variant with the
TpR gene of pR388 inserted into the
unique Drall site of \S402, was ligated
into pTG.LI 33, an IncPI plasmid
temperature sensitive with respect to
its replication to form pTGL154. E. coli
derivatives bearing pTGL154, when
propagated at 42°C, gave rise to TpR
transposants lacking the TcR marker of
the donor plasmid. The results suggest
that \S402 and \S407 can transpose in
E. coli, We are analysing the putative
tranposants to confirm that this is the
case.
MgcrQrgstriction Analysis of the
P. cepacia Genome
The restriction enzymes Pac\, Dra\,
and Ase\ cleaved the P. cepacia
chromosome into 6, 41, and 35
fragments, respectively. The sums of
the molecular weights of the fragments
indicated that the size of the P. cepacia
chromosome is 4.9 Mb. Thus the
unusual degradative abilities of this
bacterium are not a consequence of its
having a large genome than that of less
versatile bacteria. Prototrophic and
auxotrophic derivatives were identified
in which large deletions resulted in
115
-------
fusion of neighboring Pad fragments.
Afl\\ and Xba\ fragments containing
Pad sites have been used in Southern
hybridization experiments to define
linkages between Pad fragments. All
six Pad fragments were shown to
comprise a single linkage group. We
have carried out Southern hybridization
experiments to determine the
distribution of \S401, \S406, \S407,
\S408, and \S415. All copies of these
elements were clustered within a
limited region of the genome comprised
by the two smallest Pad fragments.
Analyses of the distributions of
additional elements and of essential
genes should reveal whether such
clustering reflects the location of
essential genes in other regions of the
chromosome or perhaps organization of
elements into domains where their
transposition can be regulated by
localized changes in DNA topology --
e.g. alterations in extent of
supercoiling.
FUTURE WORK
Our long term goal is to understand
the factors underlying the unusual
adaptability of P. cepacia. We are
particularly interested in examining the
roles of IS elements in the evolution of
catabolic pathways. Towards this end
we will extend the Pad restriction map
we have generated to include more data
about the distribution of IS elements as
well as the locations of genes encoding
various biosynthetic and catabolic
pathway enzymes. A major aim is
understand how P. cepacia maintains
relatively stable phenotypes despite its
potential to undergo frequent IS
element dependent rearrangements of
its genome. Our working hypothesis is
that transpositional activity is triggered
in response to environmental stress --
e.g. prolonged starvation for nutrients
and exposure to toxic agents. To
examine whether transposition -is
regulated and gain information about
functions required for transposition we
are constructing strains whose
chromosomes contain IS elements
carrying a TpR resistance casette.
These will be used to obtain data about
the influence of various physiological
conditions on transposition. (The
transposition assay entails transfer of
broad-host-range plasmids from the
pertinent strains and screening of
transconjugants for acquisition of the
TpR determinant of the marked
elements). It should also be possible to
use such strains to gain information
about the factors governing exchange
of P. cepacia elements in complex
bacterial populations.
PUBLICATIONS
A. Ferrante, and T.G. Lessie. 1991.
Nucleotide sequence of \S402, a
transposable gene-activating element
from Pseudomonas cepacia. Gene. In
press.
Wood, M.S., A. Byrne, and T.G. Lessie.
1991. Characteristics of \S406 and
\S407, two lac gene activating
elements from Pseudomonas cepacia.
Submitted for publication.
Wood, M.S., C. Lory, and T.G. Lessie.
1990. Activation of the lac genes of
Tn951 by insertion sequences from
Pseudomonas cepacia. J. Bacteriol.
172: 1719-1724.
116
-------
Lessie, T.G., M.S. Wood, A. Byrne, and
A. Ferrante. 1990. Transposable gene
activating elements in Pseudomonas
cepac/a. • p. 279-291. in: S. Silver,
A.M. Chakrabarty, B. Iglewski, and S.
Kaplan (ed.), Pseudomonas:
Biotransformations. Pathogenesis, and
Evolving Biotechnology. Am. Soc.
Microbiol., Washington, D.C.
117
-------
STUDIES ON CONJUGAL TRANSFER OF PLASMIDS FROM GEMs TO
INDIGENOUS AQUATIC BACTERIA
Tamar Barkay
U.S. Environmental Protection Agency
Environmental Research Laboratory
Gulf Breeze, Florida
INTRODUCTION
The horizontal spread of recombinant
genes from genetically engineered
microbes may magnify ecologicalrisks
because their stability and expression
are likely to increase in microorganisms
indigenous to aquatic and terrestrial
environments. Two experimental
approaches have been used for
evaluation of gene transfer in the
environment: i. A retrospective
approach comparing clonal identity with
plasmid patterns of bacterial pop-
ulations (Selander and Levin, Science
210:545-547(1979)). ii. The addition
of genetically marked donor and
recipient strains to environmental
samples and the scoring of recombinant
strains. Due to problems posed by
background indigenous microbes this
last approach employs either environ-
mental enclosures or sterilized
environmental samples. Both ap-
proaches indicate that gene exchange
does occur, albeit at low frequencies, in
the environment (Levy and Miller (eds.),
Gene Transfer in the Environment.
McGraw-Hill Publishing Co. (1989)).
However, the use of marked laboratory
strains may underestimate frequencies
of transfer because the strains are not
adapted to life in the test environment.
Thus, methods are needed to detect in
situ horizontal gene transfer between
indigenous microbes. Here, I describe
an experimental approach for the
detection of gene transfer from an
engineered organism to microbes
indigenous to aquatic environments.
The approach is based on the
assembly of catabolic pathways when
the donor and recipient genes are
brought together through the formation
of recombinant strains. If genes
encoding part of the pathway are
placed on a genetic element whose
transfer is followed, and the indigenous
community is acclimated to carry out
the remaining reactions of the same
pathway, then strains with new
catabolic capabilities would emerge
when indigenous microbes receive and
replicate the transferred genes. I have
used the assembly of merB (encoding
organomercurial lyase) on broad host
range conjugal plasmids, and merA
(encoding mercuric reductase); to
detect gene transfer from a donor
pseudomonad strain to Hg(ll)
acclimated aquatic microbes by the
ability of transconjugants to grow in
presence of an organomercury
compound, phenyl-mercuric acetate
(PMA). This is based on a prior
demonstration that Hg(ll) acclimated
microbial communities are enriched for
microbes 'that reduce Hg(il) by the
activity of the mercuric reductase
enzyme and carry merA genes (Barkay
et al., Appl. Environ. Microbiol.
56:1695-1701 (1990)).
118
-------
MATERIALS AND METHODS
Construction of conjugal broad host
range plasmids containing a /we/B-plant
DNA cassette: Plasmids, pGTE15 and
pGTE22, containing the merB gene of
the broad spectrum mercury resistance
plasmid R831 and a 350 base pairs (bp)
sequence of the plant napin gene,
bracketed by multi cloning sites of the
pUC plasmids were prepared. The
/nerB-plant DNA cassette was excised
and cloned into appropriate cloning
sites of the following broad host range
plasmids, R388 (IncW), RP1 (IncP) and
pKM101 (IncN) to create plasmids
pGTE16, pGTE26 and pGTE25, res-
pectively (Figure 1, page 120).
Conjugal transfer of parental and
recombinant plasmids between E. coli
and pseudomonad strains was tested
by the filter conjugation test as
described by Simon et al., (Bio-
technology 1:784-791 (1983)).
Conjugal transfer of merB-plant DNA
plasmids to aquatic microbes: The
experimental procedure employed is
outlined in Figure 2 (page 121). Fresh-
water bacteria were acclimated to Hg(ll)
using conditions described by Barkay
(Appl. Environ. Microbioi. 53:
2725-2732 (1987)). P. aeruginosa
containing the test plasmid (approx-
imately 107 cells/ml) were added to
acclimated water samples, and samples
were filtered through 0.2 jj nitro-
cellulose filters using low vacuum
pressure. This treatment did not injure
either donor or acclimated microbes.
Following a 4 hour incubation (a period
demonstrated to allow for conjugal
transfer), the cells were suspended in 1
ml broth, Hg(ll) was added to 0.1 //M
and suspensions were incubated for 30
minutes prior to plating on selective
media. This incubation in presence of
a sublethal concentration of Hg2+ was
essential for expression of mer and
subsequent selection of PMA resistant
transconjugants. Conjugation with
microbes from control communities
(incubated in the laboratory in the
absence of Hg ) were used as
negative controls. Colonies grown on
PMA plates (10//g/ml) were confirmed
as transconjugants by i. transfer to
medium containing antibiotics (Tp - tri-
methoprim (1000//g/ml), Cb - carben-
icillin (600 //g/ml), and Kn - kanamycin
(600//g/ml), for pGTE16, pGTE25 and
pGTE26, respectively), and ii. hybrid-
ization with the plant DNA probe
(described by Barkay ibid).
RESULTS AND DISCUSSION
Evaluation of enrichment (Table 1)
indicates that whereas Hg(ll) resistant
bacteria were selected during the 2
days acclimation period, selection for
PMA resistance did not occur. Such
selection was anticipated because merB
is a part of the mer operon that
includes merA, the gene enriched
during acclimation (Barkay et al., 1989.
Appl. Environ. Microbiol. 55:
1574-1577). Coselection would have
resulted in a high background of PMA
resistant bacteria and a possible
masking of transconjugants.
Results of conjugation between
PA01(pGTE16) and indigenous Hg(ll)
are presented in (Table 2). Presumptive
transconjugants were present at 10 to
103 CFU/ml whereas background PMA
resistance was below detection. Thus,
there is at least two orders of
magnitude difference between the
119
-------
Hlndill
EcoRl
Hlndill
EcoRl dlgttt
Hlndill
Km
2.3kb
_ _.x B«lhHI
EcoRI\frigrn»nt «"8««<
., ,: E!
Figure 1. Construction of Broad Range Plasmids with a mer B-Plant DMA Cassette
120
-------
Freshwater samples
£ ~3-£
+25Q|Lig/L
Control
NoHg2*
y 2 days Incubation at 30°C
Enumeration of total, HgR and PMAR CFU/ml.
20 ml water sample 2x106 cells of P. aeruginosa PA01 (merS-plant DNA plasmid)
: filter - 0.22^ Type GS filter
1 4 hours incubation at 30°C ;
Cells suspended in 1ml 1/2PCB
HgCI2 added toO.
.' •'-• --''" 'i
incubation at 30°C for 30 minutes
Plating on: 1/2PCA with PMA (25 \iM) -+* Transconjugants
172PCA with Hg2+ (25 ^M) -»- Potential recipients
1 /2PCA with antibiotics •' r^- Donors
2 days • 30°C
Confirm conjugation by:
i. Growth in presence of antibiotics
ii. Hybridization with a eukaryotlc DNA probe
Figure 2. Conjugal Gene Transfer From GEMs to Aquatic Bacteria
121
-------
Table 1. Enrichment of Hg(ll) and PMA resistant heterotrophic bacteria after exposure
to 250 M9/L Hg(ll) for two days at 30 • C.
Exp.1
CFU/ml
Exp. 2
Day 0
Day 2
Total
Hg2 + -R
PMA-R
Total
Hg2+-R
PMA-R
+ Hg(ll) Control
3.3x1 04
4.8x1 02
2.X10'2
4.0x1 05 3.2x1 04
8.6x1 04 ' 1.0x102
<10.0 <0.01
+ Hg(ii)
5.4x1 04
1.2x101
<0.01
4.7x106
8.0x1 02
0.9
Control
6.7x1 04
<1.0
<0.01
Table 2. Results of conjugation experiments between PA01(pGTE16) and indigenous
Hg(ll) resistant indigenous bacteria.
Donor
"Recipient"
(Hg2+-R)
Transcon.
Background
(PMA-R,
no donor)
Efficiency
(trans./donors)
+ Hg(ll)
3.1x106
ND2
6.0x1 02
<1.25
1.9x10'4
(< 4.0x1 0'7)3
Control
4.0x1 07
ND
<1.25
<1.25
+ Hg(ll)
7.8x1 06
9.5x1 03
3.5x1 02
<1.42
3.6x1 Q-2
Control
1.2x107
ND
ND
ND
recipient
1 Hg(ll) - indigenous flora from exposed microcosms; Control - indigenous flora from
microcosms not exposed to Hg(ll).
2 ND - not determined
3 appearance of background PMA-R/donor
122
-------
number of transconjugants and
background PMA resistant organisms,
allowing for unmasked detection of
transconjugants. Efficiency of transfer
was at the range of 10~4 to 10~5
transconjugants/donor. In comparison,
the efficiency of pGTEt 6 transfer to an
E. coli HMS174 that contained
pACYC184::Tn501 (encoding Hg(II)
resistance and compatible with
pGTE16) was 3x10"3 per donor. When
the number of transconjugants is
related to the number of potential
recipients (in Exp. II, Table 2), it
appears that 3% of the Hg(ll) resistant
strains received pGTE16. Thus,
experiments using standard laboratory
strains may overestimate efficiencies of
transfer. All PMA resistant
transconjugants were found to be
resistant to Tp at 1000 //g/ml and they
hybridized with the plant DNA probe,
confirming that they were formed by
conjugation between PA01 and
indigenous bacteria. When the number
of transconjugants is related to the
number of potential recipients (in Exp.
II, Table 2), it appears that 3% of the
Hg(ll) resistant strains received
pGTE16, indicating that there is a
substantial potential for transfer'of this
IncW plasmid to indigenous flora.
In summary, the utility of catabolic
gene assembly as a tool for detection of
conjugal transfer of recombinant
plasmidsto indigenous micro-organisms
was demonstrated under optimal
conditions. This approach is currently
applied to bacterial populations in
environmental samples. This experi-
mental system will allow determinations
of the effect of environmental
parameters on transfer of conjugal
plasmids.
123
-------
THE FATE, STABILITY & MOVEMENT OF FOREIGN DNA IN FILAMENTOUS
FUNGI: AN ENVIRONMENTAL STUDY
Martin B. Dickman1 & John F. Leslie2
Department of Plant Pathology
University of Nebraska
Lincoln, Nebraska1
Department of Plant Pathology
Kansas State University
Manhattan, Kansas2
INTRODUCTION
Genetically engineered fungi (GEF)
offer great potential as tools with which
to study the fate of recombinant DNA
{rDNA) in the environment. While
techniques have been developed to
detect and monitor genetically
engineered bacteria, progress has been
less rapid with fungi. Presently there is
no eukaryotic model for studying the
effects resulting from the introduction
of rDNA into an ecosystem via a fungal
vector. As simple, but true,
eukaryotes, GEFs provide a unique
opportunity for studies of strain
development and use under field
conditions. Prior to environmental
release, however, detailed risk
assessments are needed to design
rational strategies for safe release.
We have been using both molecular
and genetic techniques to study the
impact of GEFs in the ecosystem. In
order to predict the potential impact of
a GEF following introduction into the
environment, it is critical to evaluate the
consequences of foreign DNA insertion
into the organism. We have focused
our studies on the stability of such
foreign sequences following meiosis
and mitosis, and are using sequences
with and without significant homology
to the recipient's genome to determine
if the origin of the foreign sequences
plays a significant role in determining
stability.
For our studies we have been using
fungal strains identified as Gibberella
fujikuroi mating population "A"
(Sawada) Ito in Ito & K. Kimura
(asexual stage Fusarium moniliforme
Sheldon). This organism can be readily
manipulated under laboratory
conditions, and sexual crosses
completed with ease in a 4-6 week time
frame. Numerous markers, including
resistance to hygromycin B and
benomyl®, as well as auxotrophs and
spore color variants, are available in
strains that we have developed. Also,
this fungus can be divided into over
1000 different vegetative compatibility
groups (VCGs). At least 10 genes
underlie this VCG system and they
determine whether stable heterokaryons
can be formed and appear to limit the
degree to which asexual transfer of
genetic material can occur. We have
also developed a reliable DNA-mediated
gene transfer system for this fungus.
These advantages make this organism
a natural candidate for studies of the
fate of foreign DNA in the environment.
124
-------
Our overall objectives are: a.) to
determine conditions for stable
maintenance of rDNA in filamentous
fungi and identify factors that influence
genetic stability, and b.) to determine
parameters affecting genetic transfer of
rDNA within and between fungal
populations.
Specifically we will: a), determine if
copy number affects the stability of the
foreign DNA through meiosis and
mitosis, and b). Determine if rDNA can
be passed within and between
populations of G. fujikuroi through
asexual or sexual genetic exchange.
METHODS
Strains and Plasmids
All Gibtrerella fujikuroi strains used in
this study belong to mating population
"A" (variety moniliformis) and were
fertile as males. Most of the strains
were also female fertile. Strains were
stored in 15% glycerol at -80°t:;
vegetative cultures were maintained on
complete medium. For hygromycin
resistance studies, the vector pBHM-1
containing the hygromycin B
phosphotransferase resistance gene
from E. coli fused to promoter
sequences from Cochliobolus
heterostrophus was used. For studies
of benomyl resistance the Neurospora
crassa altered p-tubulin gene in the
vector pBT6, was used. In addition,
we have cloned as mutant P-tubulin
gene from G. fujikumi (see below).
Cultural Practices ,-. .
.Resistance to the amino glycoside
antibiotic hygromycin B was scored
three days following transfer from
complete medium to minimal medium
plus 100 mg/L of the drug. Plates
containing hygromycin B were made
fresh and not kept for more than two
weeks because the drug's inhibitory
properties in the media decrease with
time. Resistance to benomyl® was
done at a concentration of 1 mg/L.
Strains carrying the nicl mutation
require nicotinic acid and were grown
on MM supplemented with 10 mg/L
nicotin-amide. Strains carrying the
pdxl mutation require pyridoxine and
were grown on MM supplemented 10
mg/L pyridoxine H.CI. Sexual crosses
were made on carrot agar, and single
random ascospores recovered using a
micro-manipulator. Mutants and trans-
formants were purified through the
isolation and subculturing of uni-
nucleate microconidia.,
Mutant Induction
A benomyl resistant mutant of G.
fujikuroi was obtained by UV irradiation
of fungal spores. ,
DNA Manipulations
Fungal DNA was extracted using a
"rniniprep" procedure that we have
modified for G. fujikuroi. Large and
small scale plasmid preparations were
made using the alkaline lysis procedure.
For Southern analyses, DNA samples
were digested to completion with
appropriate restriction enzymes,
separated on agarose gels, and
transferred to nitrocellulose mem-
branes. DNA probes were isolated and
purified from agarose gels and radio-
labeled by nick translation. Blots were
125
-------
autoradiographed with Kodak X-
(OMAT) film and Dupont Cronex
intensifying screens at -70°C.
Transformation Procedures
Cells were transformed to either
hygromycin or benomyl resistance via a
whole cell transformation procedure
adapted for G. fujikuroi. This procedure
is relatively inefficient with respect to
the number of transformants per //g of
DNA, however it is rapid and
technically simple.
RESULTS AND DISCUSSION
Recovery of Benomyl Resistant Mutants
and Cloning of G. fuiikuroi /?-Tubulin
To isolate the p-tubulin gene, a
plasmid library in pUC119 was
constructed from the benomyl resistant
mutant. Using highly conserved
sequences for (5-tubulin, primers were
designed and the polymerase chain
reaction (PCR) was used to generate a
probe from total genomic DNA of the
mutant. Following the screening of the
library a 2.1 kb clone was selected and
mapped. Constitute expression of the
gene was indicated by RNA blots and a
transcript of 1.9 kb was observed.
Validation of the functionality of this
gene was demonstrated by trans-
formation of wildtype G. fujikuroi to
benomyl resistance.
Meiotic Stability of Transformed Strains
of G. fuiikuroi
To date, stability studies have
focused on the hygr gene under meiotic
conditions. Crosses with single copy
transformants segregated hygr: hygs in
a 1:1 manner consistent with that
expected for a Mendelian locus in a
haploid organism. Multiple-copy
transformants, however, segregated
hygr: hygs in a 1:2 manner that was
not consistent with Mendelian
expectations for a chromosomal
marker, even though two unrelated
auxotrophic nuclear genes were
segregating normally. Segregation
ratios in crosses in which hygr was
introduced via the male parent did not
differ significantly from crosses in
which the transformed strain served as
the female parent. Some of the
sensitive progeny from the crosses with
the multi-copy transformants carried
hygr sequences. When these pheno-
typically sensitive progeny were
crossed with a wild-type strain that
carried no hygr sequences, some of the
progeny were phenotypically hygr.
Some progeny from some crosses were
more resistant to hygromycin than were
their sibs or the transformant strains
that served as their parents.
Transformants passaged through a
maize plant only rarely segregated
progeny with high levels of resistance.
The mechanism underlying these
genetic instabilities is not clear, but
may involve unequal crossing-over
and/or methylation.
FUTURE WORK
Foreign DNA can be inserted into
fungal chromosomes in three basic
patterns: a single copy at a single site,
multiple copies at a single site, and
multiple copies at multiple sites. We
are presently constructing G. fujikuroi
126
-------
multiple copies at multiple sites. We
are presently constructing G. fujikuroi
strains with each of these different
patterns with respect to the hygr,
Neurospora benr, and G. fujikuroi benr
genes. We will analyze these different
genomic configurations with respect to
fitness characters, stability and
transmission capability. We will cross
these transformed, strains with our
auxotrophically marked strains to
determine if the origin of the sequence
and/or the amount of sequence
homology between the genomic DNA
and the foreign sequence affect the
stability of the DNA. We will also
further examine the structure of the
second generation progeny with
"reactivated" resistance to determine if
the configuration of the foreign DNA
has been altered in these strains as
compared to th'e original transformant.
We will begin studies to determine if
strains that belong to the same or dif-
ferent VCGs are capable of exchanging
foreign DNA ihtroduced via trans-
formation. Strains marked with
traditional genetic markers (auxotrophs
and spore color mutants) will be paired
with the foreign DNA included in the
auxotrophic strain. Selection will be
based on the concurrent recovery of
resistant prototrophic colonies fol-
lowing growth under conditions that
favor the formation of heterokaryons.
Strains that differ at 0, 1, and 10 of the
loci underlying the VCG phenotypes will
be used to provide a range of possible
inhibitions to the asexual genetic
transfer process.
PUBLICATIONS
Leslie, J. F., and Dickman, M. B.
1991. Fate of DNA encoding
hygromycin resistance following
meiosis in transformed strains of
Gibberella fujikuroi. Applied and
Environmental Microbiology 57{5): in
press.
127
-------
-------
SESSION II
EFFECTS RESEARCH
Ecological Processes
Researchers are evaluating the effects of introduced microorganisms on
structural/functional aspects of freshwater, marine and terrestrial ecosystems.
Improvements in the procedures that are developed in these ecological
investigations may be incorporated into new testing methods. Further work is
being conducted on smaller components of aquatic and terrestrial systems not
directly related to ecosystem-level effects. This work includes pesticide degrading
microorganisms and biological control agents such as insect growth regulators and
microbial pest control agents.
Higher Organisms
A wide variety of microorganisms may be used in biotechnology and information
on their lexicological, behavioral, pathogenic, and histopathological effects on
higher organisms is unavailable. This research investigates the genetic and
molecular basis of infectivity and pathogenicity, and is developing methods that
determine the host range of microbial pest control agents.. Experimental results will
be used to develop protocols for testing effects of microorganisms on beneficial
invertebrate and vertebrate species.
Human Health
Researchers are investigating the toxicological, pathological and cytogenic
effects of exposing humans to selected recombinant and non-recdmbinant
microorganisms using mammalian cells and surrogate organisms. Different
exposure routes are evaluated in terms of measurable effects.
129
-------
-------
USE OF RIBOSOMAL rRNA SEQUENCES TO CHARACTERIZE DIVERSITY AND
STABILITY OF MICROBIAL POPULATIONS
Richard Devereux
Technical Resources, Inc.
Environmental Research Laboratory
Sabine Island
Gulf Breeze, Florida
INTRODUCTION
An explicit characterization of
natural microbial community diversity
has been elusive. In large part, this has
been due to the constraints pure culture
methodologies have imposed upon
environmental microbiology - pure
cultures needed to be representatively
obtained from the natural population
for subsequent laboratory identifi-
cation and biochemical character-
ization. This approach was recognized
to be limited in that only 10% (at best)
of the bacteria observed in an
environmental sample by direct
counting could be obtained in
laboratory pure cultures. Further,
biochemical identifications of such pure
cultures may have been imprecise and
have lead to erroneous misclass-
ifications in the past. Assessing the
potential risk imposed upon natural
microbial populations and the ecological
processes they mediate requires a
greater degree of resolution than
offered by isolation of pure cultures.
Recent advances in the application
of molecular techniques now permit
long-standing issues in microbial
ecology to be addressed. In particular,
small subunit ribosomal RNA (16S-like
rRNA) sequence comparisons facilitate
the direct measure of natural microbial
population diversity. Population
f diversity of sulfate-reducing bacteria
(SRB) in anaerobic estuarine sediments
is being explored through 16S rRNA
sequences. SRB have direct roles in
the cycling of sulfur and the terminal
oxidation of organic matter. Currently,
SRB population diversity is being
measured with 16S rRNA-targeted
hybridization probes. These studies are
directed towards defining populations
of SRB and assessing population
stability in relation to ecological
processes.
METHODS
Phvlogenetic Probes
Oligonucleotide probes were
redesigned such that occurrence of.
mixed, or multiple nucleotides at a
common nucleotide position, were
eliminated. This was found to provide
more precise measurements of relative
rRNA abundance for hybridizations to
membrane bound nucleic acids. Probes
for specific SRB groups were labeled at
their 5' ends using gamma 32P-ATP and
polynucleotide kinase. Additionally,
fluorescent dye-conjugated bacterial
domain- and delta subdivision-specific
Oligonucleotide probes were prepared
for direct fluorescent microscopic
identification of bacterial cells obtained
from marine sediments.
131
-------
Analysis of Natural SRB Populations
Nucleic acids were extracted
directly from fractions of sediment
cores by mechanical disruption (bead-
beater lysis) in the presence of sodium
acetate/EDTA buffer pH 5.2, sodium
dodecyl sulfate, polyvinylpoly-
pyrollidone, and beta mercaptoethanol.
Lysates were extracted with phenoU
phenol:chloroform, and chloroform.
Nucleic acids were precipitated with
ethanol. These were then applied to
nylon membranes and hybridized with
the SRB rRNA specific probes. Rates of
sulfate reduction in the sediment core
fractions were determined by isotopic
assay. Measurements of sulfate and
sulfide in the sediment porewater was
by standard methods.
Preparation of a Sediment Bacterial
Fraction
Sediment bacterial fractions were
obtained for whole cell hybridizations
with fluorescent dye-conjugated
oligonucleotide probes. Sandy
sediment samples (3.0 g wet weight)
were placed into 50 ml flasks with 3
mm glass beads and 0.15 M NaCI. The
contents of the flasks were vigorously
shaken for 30 min. then allowed to
settle for 15 min. A 3 ml aliquot was
withdrawn from the upper portion of
the settled suspensions. DAPI stained
cells and cells hybridizing with the
rRNA probes were determined.
Efficiencies of cell recoveries were
obtained by adding known amounts of
sulfate-reducing bacteria to washed,
sterile sand and to natural sediment
samples. For enumeration of sulfate-
reducing bacteria, preparation of cell
fractions were prepared under an-
aerobic conditions. The bacterial
fractions were passed through 8.0 //M
filters to remove large particles.
Filtrates were treated with combina-
tions of SRB basal salts medium,
nutrients, and nalidixic acid to stimulate
production of ribosomes and enhance
the detection of SRB.
RESULTS AND DISCUSSION
SRB Population and Function Profiles
in Sediment Cores
Concentrations of sulfate were 15
mM in the surface water and in
sediment porewater to a depth of 4 cm.
Porewater concentrations of sulfate
decreased from 10 mM at the 4-5 cm.
depth interval to less than 5 mM in the
14-16 cm, depth interval. Porewater
concentrations of sulfide were 0.1 to
0.2 mM in the 4-10 cm. depth intervals
and increased markedly to 0.8 mM 12
to 16 cm. below the water-sediment
interface. The peak in activity of
sulfate reduction (3.5 mmol/ml/hr)
occurred in the 2-3 cm. interval.
Desulfovibrio 16S rRNA was the
most abundant SRB rRNA and peaked
at the 3-4 cm. interval. Stratification of
the SRB population with sediment
depth was apparent. Amounts of 16S
rRNA of fatty acid- and/or acetate-
utilizing SRB (encompassing four SRB
phylogenetic groups) decreased in the
sediment fractions haying the highest
amounts of Desulfovibrio 16S rRNA.
The peak, in activity of sulfate
reduction did not occur within the same
fraction as the predominant peak in
Desulfovibrio rRNA. Acetate has
previously been shown by other
132
-------
investigators to be the primary
substrate for sulfate reduction in
estuarine sediments. Thus, the peak
sulfate reduction .rate was expected to
coincide with a peak in the rRNA
abundance of acetate-utilizing SRB.
This relationship was not seen in the
results and may be accounted for by
several possibilities. Additional groups
of SRB may exist in nature that are not
represented in the culture collection
that not are being accounted for by the
probes. Also, since the data represent
one sampling time, it is possible that a
shift in the activity of sulfate reduction
had occurred and that the population
had yet to respond by an increase in
rRNA abundance. Finally, accounting
for energy released by the reduction of
sulfate by different electron donors
should be considered in evaluation of
the community structure-function
relationship. Hydrogen, for example, as
an electron donor for sulfate reduction
yields considerably more energy than
acetate.
Whole Cell Hybridizations
of Natural Populations
Laboratory grown SRB cells were
recovered from washed, sterile
sediment samples with efficiencies
approaching 90%. For fixed sediment
samples, cells hybridizing with the
bacterial domain probe were 16% of
the cells observed by DAPI staining.
Cell counts hybridized with the delta
subdivision-specific probe were 42% of
the counts hybridized with the bacterial
domain probe. Delta subdivision cell
counts increased to 59% and 49% of
the bacterial domain counts when
samples were incubated in medium
with or with nalidixicacid, respectively.
When samples were incubated in
medium containing molybdate, and
inhibitor of sulfate reduction, delta
subdivision cell counts were only 11 %
of the bacterial domain counts. The
results demonstrate that fluorescent
dye labeled rRNA probes could be used
with natural samples and that specific
bacterial populations can be stimulated
to enhance their detection.
FUTURE WORK
Research planned will address the
community structure-function relation-
ship of SRB populations. PCR amplifi-
cation and sequencing of 16S rRNA
genes from natural populations will be
used to determine natural diversity not
represented in the culture collections.
Natural populations will be character-
ized during temporal sampling to
determine base line variations in
populations with respect to sulfate
reduction rates and turnover rates of
substrates. Laboratory sediment
systems will be used to control
parameters, such as availability of
specific electron donors, to alter SRB
populations. Hybridization of
fluorescent labeled probes sediment
bacteria as a means to increase the
number of samples which can be
analyzed will be explored.
PUBLICATIONS
Amann, R.I.,. B.J. Binder, R.J, Olsen,
S.W. Chisholm, R. Devereux, and D.A.
Stahl. 1990. Combining 16S rRNA-
targeted oligonucleotide probes with
flow cytometry for analyzing mixed
microbial populations. Appl. Environ.
133
-------
Microbiol. 56:2012-2020.
Devereux, R., S.-H. He, C.L. Doyle, S.
Orkland, D.A. Stahl, J. LeGall, and
W.B. Whitman. 1990. Diversity and
origin of Desulfovibrio species:
phylogenetic definition of a family. J.
Bacteriol. 172:3609-3619.
Stahl, D.A., R. Devereux, R.I. Amann,
B. Flesher, C. Lin, and J. Stromley.
1990. Ribosomal RNA based studies of
natural microbial diversity and ecology.
pp. 669-673. In: T. Hattori et al. (eds.),
Recent Advances in Microbial Ecology.
Proceedings of the 5th International
Symposium on Microbial Ecology.
Japan Scientific Societies Press.
Devereux, R. and D.A. Stahl. 1991.
Phylogeny of sulfate-reducing bacteria
and a perspective for analyzing their
natural communities, in: J.M. Odom
and R. Singleton, Jr. (eds), Sulfate-
Reducinq Bacteria: A Contemporary
Perspective. Brock-Springer series in
Contemporary Bioscience. (in editorial
review).
134
-------
HABITAT SPECIFIC DIFFERENCES IN PERSISTENCE AND EFFECTS OF
INTRODUCED CELLULOLYTIC BACTERIA USED AS SURROGATES FOR GEMS
Thomas L. Bott and Louis A. Kaplan
Stroud Water Research Center
Academy of Natural Sciences
Avondale, Pennsylvania
INTRODUCTION
These studies have three primary
objectives: a) to determine the
persistence of introduced bacteria in
natural benthic aquatic communities, b)
to test several community and
ecosystem level response variables in
order to pinpoint those most useful for
detecting effects of the introduced
organism, and c) to evaluate the utility
of mesocosms for assessing the fate
and effects of genetically engineered
microorganisms (GEMs) in natural
systems. We have worked with
surrogates for a GEM with enhanced
cellulosedegradingcapability. Cellulose
"superdegrader" GEMs and their
products are of industrial interest and
may find application in the conversion
of biomass to fuel. The use of a
surrogate expedites field studies,
although we recognize that a surrogate
will not mimic a GEM in all
physiological and ecological respects.
Our research focuses on stream
ecosystems since they often receive
waste discharges and non-point source
inputs. White Clay Creek, the study
stream, drains a protected rural
watershed in southeastern (Chester
Co.) Pennsylvania.
We screened over 20 isolates of
aerobic cellulolytic bacteria from culture
collections by assaying their lytic
activity on pure cellulose, ability to
degrade algal detritus and leaf litter,
and their growth rates and ranked them
according to their performance in these
tests. Fluorescent antisera (FA) were
produced against the most active
isolates, Cellulomonas flavigena (NRC
2403), C. fimf (NRRL B402), C. sp.
(NRC 2406), and C. uda (NRRL B404)
to enable their detection in benthic
communities. Preliminary studies of the
population dynamics of C. sp. (NRC
2406), C. uda (NRRL B404), and C. sp.
CS1-1 (an organism that cross reacted
with C. uda antiserum) in streambed
sediments were conducted in 2.95 L
microcosms.
C. sp. (NRC 2406) was selected for
initial studies in larger (35 L) mesocosm
streams. The first experiment,
conducted during the summer months
indicated that the isolate could persist
for several weeks in sediments and
benthic algal growths, although
densities declined fairly rapidly from
high post-inoculation levels. Effects on
benthiccommunity metabolism (primary
productivity and community respir-
ation), associated changes in DOC, and
bacterial productivity were not
pronounced, High densities of FA
stained cells were found in algal
growths in both an inoculated and a
control mesocosm 3 months after
introduction. We have also found cells
135
-------
that stained with C. uda (NRRL B404)
and C. sp. (NRC 2406) antisera in leaf
packs and streambank soil from White
Clay Creek. While there may be a
natural source of FA stained cells in
White Clay Creek, our observations
suggested habitat specific differences in
persistence deserved study. The
experiment described below focused on
population dynamics of the introduced
organism in three different micro-
habitats; algae, sediments, and leaves,
and on alteration of the rate of
degradation of leaf litter.
METHODS
Six stream mesocosms (2.23 m
long x 0.203 m wide x 0.127 m deep),
designed to simulate a slow run habitat
in White Clay Creek, were constructed
of vinyl coated steel and Plexiglas end
plates. They were filled with 2 cm. of
gravel and coarse sand from White Clay
Creek, over which 40 plastic trays
(0.10 m square x 0.038 m deep)
containing surface sediments from
White Clay Creek were placed. The
bottoms of the trays were removed and
replaced with 400 jim mesh nylon
screening to allow exchange of water,
nutrients, dissolved gases and
organisms within the sediments.
Stream water (35 L) was recirculated in
each system; 900 ml/min of new
stream water was added continuously
to replace the entire volume
approximately every 40 min. Water
depth approximated 1.5 cm. Water
returned to White Clay Creek was first
filtered through cartridge filters (which
were autoclaved before discarding) and
irradiated with ultraviolet light, a
treatment found to effectively kill the
study bacteria. The streams, housed in
a greenhouse, were immersed in water
jackets continuously replenished with
White Clay Creek stream water to
maintain ambient stream water
temperatures. The systems were
separated by clear plastic sheeting
suspended from a wooden frame to
hinder spread of organisms between
microcosms.
Population and community struct-
ural and functional parameters
(including densities of FA stained cells,
total bacteria, primary productivity,
community respiration, ATP, chlorophyll
a, and cellulasesactivity) were measured
by sampling sediments and filamentous
algae (dominated by Cladophora sp.)
prior to addition of the isolate. Just
before inoculation of the isolate, packs
(approximately 5 g each) of tulip poplar
(Liriodendron tulipifera) leaves that had
been previously leached and redried to
obtain weights, were placed in the
mesocosms. The surrogate was placed
in the water in two mesocosms,
maintaining two others as uninoculated
controls. Cells for inoculation were
grown at 20°C on Solka-Floc in basal
salts medium for 5 days with agitation.
After allowing the cellulose to settle,
cells were harvested, resuspended in
autoclaved stream water, and added to
duplicate mesocosms. After
recirculating the inoculum for
approximately 72 h with no stream
water addition, the systems were
returned to flow-through mode and 4 h
later, an intensive (3-5 day) study of
densities of the surrogate and total
bacteria in surface sediments, leaf
packs, and algae was initiated. Then,
at approximately 10-14 day intervals
the response variables noted above
136
-------
were measured, and samples for leaf
pack chemical composition and weight
loss were collected. Light,
temperature, and selected water
chemical characteristics were
monitored. Samples of sediment and
Cladophora sp. were obtained by filling
plastic rings (1.15 cm. i.d. x 0.2 cm.
deep) with sample. Leaf packs were
sampled by coring replicate samples
(1.10 cm. i.d. x approximately 2 cm.)
with a cork borer.
Primary productivity and community
respiration of sediments and
Cladophora sp. were estimated from
dissolved oxygen changes in flowing
water microcosms from which eco-
system P/R status was determined.
Leaf pack respiration rates were
determined in the same systems. For
bacterial enumeration, formalin fixed
samples were sonicated for 45 sec. at
30 W with 0.1 M dibasic ammonium
phosphate as a surfactant, centrifuged
once (412 x g for 5 min at 4°C) in
30% glycerol, and twice more in 60%
glycerol, Aliquots of supernatant fluids
were filtered onto 0.2 |im pore size
Nuclepore filters and stained with FA or
DAP! ,for epifluorescence microscopic
counts of the target cell or total
bacteria on from 6-10 or 3 replicate
samples, respectively. Algal biomass
was assessed from chlorophyll a
determination, and total viable biomass
from ATP, each on 5 aliquots.
Cellulase activity was determined using
the Cellulose Azure (CA) assay; data
were corrected for color contributed by
samples using a two wavelength
correction procedure for leaf packs or a
regression of color against sample
weight for sediments. Leaf pack
decomposition was determined from
weight loss and cellulose content.
Cellulose was obtained by combustion
at 450°C for 4 h following sequential
extraction of non-cell-wall material,
hemicellulose, and lignin using
detergent solutions and permangenate
oxidation.
RESULTS AND DISCUSSION
Epifluorescence microscopic
counts indicated that the C. sp. (NRC
2406) inoculum contained 1.97 ± 0.15
x 109 cells/ml, and viable counts from
spread plates indicated 4.77 ± 0.68 x
109 colony forming units/ml. Using the
average of these estimates, we cal-
culated that the initial density in the
inoculated mesocosms was approxi-
mately 4.57 x 107 cells/ml of water.
Six water changes after exposure,
the densities (ce!ls/cm3) associated
with sediments, Cladophora sp., and
leaf packs in the duplicate inoculated
mesocosms averaged 1.06 ± 0.16 x
106, 4.32 ± 2.32 x 106, and 2.07 ±
0.97 x 106, respectively. Based on
sample weights, the densities (cells/g
dry weight) were 8.06 ± 4.16 x 105,
3.52 ± 2.83 x 107, and 1.30 ± 6.34
x 107, for those sample types,
respectively. Given the densities of
total bacteria in these samples, which
ranged between 1 x 108 and 1 x 1010
in all three sample types, the isolate
composed maximally 0.17% and
0.44% of the total bacterial community
in Cladophora in each inoculated
stream, 0.35 % and 0.49 % of in leaf
packs, and only 0.04% and 0.04% of
the sediment bacterial communities in
each of the duplicate inoculated
mesocosms. FA stained cell densities
(cells/g dry weight) in the duplicate
137
-------
control mesocosms averaged only 3.8
± 17.1 x 104 and 4.58 ± 4.20 x 106
in sediments and Cladophora sp.
respectively and the organism was
never detected in leaf packs.
Densities decreased rapidly from
post inoculation maxima in sediments,
and slowly in leaf packs, with an
intermediate rate of decline in
Cladophora. Regression analyses of
cell densities on both a dry weight and
a volumetric basis indicated half lives
(T1/2) of 13.5 (r2 = 0.829, 0.778),
1.5 (r2 = 0.599, 0.602), and 8.8 (r2
= 0.875, 0.989) days for the isolate in
those habitats, respectively. While
densities were near background
(control) levels in sediments and
Cladophora after 10 and 30 days
respectively, FA stained cells were
found in a leaf pack at 70 days.
Effects on community respiration
and primary productivity of sediment
were not statistically significant
(Before-After, Control-Impact test,
BACI, a = 0.05). These measures
were made on Cladophora samples less
frequently, and although rates were
much higher the isolate had no
pronounced effect on either respiration
or primary productivity. Similarly, there
were no strong effects on chlorophyll a
concentrations and assimilation ratios
and hourly P/R ratios were in a similar
range in both inoculated and control
mesocosms throughout the study. Leaf
pack respiration per gram dry weight
sample was higher on packs exposed to
the isolate (0.450 ± 0.205 vs. 0.367
± 0.181, n = 6), but the difference was
only statistically significant at the a =
0.10 level (paired sample t-test). There
was overlap in ATP values between
streams and no striking difference
between data collected before and after
exposure.
Cellulase activity tended to be
slightly higher overall in the inoculated
streams for each sample type.
However, the difference was statist-
ically significant only for sediment
samples (paired t test, a = 0.05) and
since it occurred both before and after
inoculation the effect cannot be
attributed to introduction of the isolate.
Correlation analyses including all
sample types indicated that isolate
densities whether expressed on a
volumetric or dry weight basis, were
weakly but significantly correlated with
cellulase activity (r= 0.40 and 0.32, a
= 0.01 and 0.05 respectively). The
relationship was highly significant for
sediment samples (r = 447; a = 0.0.1,
n = 32) although not for leaf packs,
perhaps because other cellulolytic
bacteria and fungi were more numerous
there.
Regression analyses indicated that
leaf packs lost weight at a similar rate
in both the experimental (T1/2 = 99
days) and control (T1/2 = 115 days)
systems, and predicted,slightly less
cellulose (68.9%) in exposed leaf packs
than in unexposed (75.6%). Sample
sizes were small however, and these
indications require confirmation. While
other cellulolytic bacteria and fungi
were present in the packs, correlation
analyses showed that both weight loss
and cellulose content of leaf packs
were weakly correlated (r= 0.45 and
0.50, a = 0.08 and 0.10) with FA + cell
densities and those densities expressed
as a percentage of the total bacterial
flora (r = 0.64, a = 0.05, n = 11).
138
-------
FUTURE WORK
A more detailed study of FA
population dynamics in leaf packs and
effects on leaf pack decomposition has
been performed, including a comparison
of those parameters in mesocosms and
the field. Samples from that experi-
ment are still being processed. More
work with Cladophora or other algae
with a cellulose cell wall is needed.
Obviously, for an introduced
organism to affect community or
ecosystem properties the impact must
be either immediate and acute or the
organism must be able to persist and
exert an effect over time. Work to date
indicates the need to understand in
greater detail the factors controlling
bacterial population dynamics in natural
systems, including the physiological
and ecologically relevant characteristics
of the organism as they interact with
ecosystem properties such as nutrient
supply and grazing rates. The
application of rRNA probes in those
studies is desirable, to increase the
specificity of identification of the isolate
in natural samples. Finally, we
anticipate the extension of these
studies to work with a GEM.
PUBLICATIONS
Bott, T.L. and L.A. Kaplan. In revision.
Selection of surrogates for a genetically
engineered microorganism with
cellulolytic capability for ecological
studies in streams. Canadian Journal of
Microbiology.
Bott, T.L., and L.A. Kaplan. 1991,
Persistence of introduced cellulolytic
bacteria and cellulose decomposition in
leaves, algae, and sediments measured
in mesocosm streams. Abstracts of the
91st Annual Meeting of the American
Society for Microbiology, Dallas TX,
May 5 - 9, 1991.
Bott, T.L., and L.A. Kaplan. 1990.
Cellulolytic bacteria as surrogates for a
GEM: Microcosm studies of
persistence and effects in streambed
sediments. Abstracts of the 90th
Annual Meeting of the American
Society for Microbiology, Los Angeles,
CA, May 13 - 17, 1990.
Bott, T.L., and L.A. Kaplan. 1990.
Cellulolytic bacteria as surrogates for
genetically engineered microorganisms:
Microcosm studies of persistence and
effects in streambed sediments, pp.
139 - 143. jn; Review of Progress in
the Biotechnology-Microbial Pest
Control Agent Risk Assessment
Program. EPA/600/9-90/028.
EPA, Washington.
U.S.
Bott, T.L. and L.A. Kaplan. 1990.
Beyond chemical toxicity: Stream
microcosms as systems for the study of
persistence and effects of genetically
engineered microorganisms (GEMs).
Technical Information Workshop-
Experimental Ecosystems: Applications
to Ecotoxicology. Publications of the
North American Benthological Society.
Bott, T.L., and L.A. Kaplan. 1989.
Selection of a surrogate for a GEM with
enhanced cellulose degrading capability
for studies in streams. Abstracts of the
89th Annual Meeting of the American
Society for Microbiology, New Orleans,
LA, May 14- 18, 1989.
139
-------
FIELD CALIBRATION OF SOIL-CORE MICROCOSMS FOR EVALUATING
FATE AND EFFECTS OF GENETICALLY ENGINEERED MICROORGANISMS
IN TERRESTRIAL ECOSYSTEMS
Harvey Bolton, Jr., and James K. Fredrickson
Battelle, Pacific Northwest Laboratory
Richland, Washington
INTRODUCTION
Microcosms are an attractive option
for obtaining preliminary data on the
fate and ecological effects issues
associated with the release of
genetically engineered microorganisms
(GEMs) into the environment, because
tests and microorganisms can be
contained within the laboratory.
Historically, microcosms have been
used to assess the fate, transport, and
toxicity of chemicals and pollutants.
Recently/ intact soil-core microcosms
incubated in the laboratory were used
to evaluate the fate and ecological
effects of transposon mutants of
Azospirillum lipoferum and a wheat
(Triticum aestivum) -root colonizing
Pseudomonas sp. However, before
microcosms can be used as a standard
tool for biotechnology risk assessment,
they must be compared with the field to
ensure that they are effective field
models.
The principal objective of this
project was to calibrate soil-core
microcosms with their field counter part
for microbial fate and ecosystem
structural and functional properties and
the resultant effect -of introduced
microorganisms on these properties. In
the first year of this project (October
1988 to September 1989), the fate of
rhizobacterium Pseudomonas sp. RC1
in soil and on the wheat rhizoplane was
determined as a function of time in
intact soil-core microcosms at ambient
temperature (22°C) and in a growth
chamber with temperature fluctuations
that simulated average conditions i.n the
field, field lysimeters, and field plots
(Bolton et al., 199la). The effect of
the introduced RC1 on ecosystem
structural and functional properties as
well as a comparison between the four
systems was also conducted the first
year (Bolton et al., 1991 b').
In the second year of this project
(October 1989 to September 1990)
two experiments were conducted with
soil-cores as microcosms located in a
growth chamber and as lysimeters in
the field. First, RC1 was used to
determine ho.w year-to-year field vari-
ability influenced the calibration of
microcosms with the field. Second, the
Gram-positivebacterium Streptomyces
lividans TK24 was used "to allow'us to
compare microcosms as field 'surro-
gates for both Gram-negative (RC1) and
Gram-positive (TK24) bacterial survival
and ecological effects. The complete
results of this project are contained in
the final report (Bolton et al. 1991c).
MATERIALS AND METHODS
The bacterial strains used in this
project included a spontaneous
rifampicin resistant mutant.(100 mg/L)
of Pseudomonas sp. strain RC1 (RC1)
140
-------
grown on Sands rif50 medium and
Streptomyces lividans TK24 (TK24),
which was selectively grown on
starch-casein mineral salts medium with
carbenicillin (25 nng/L), streptomycin
(50 mg/L), nystatin (50 mg/L), Benomyl
(50 mg/L), and cycloheximide (500
mg/L).
Intact soil cores for use as
microcosms and field lysimeters (17.5
cm. diameter, 60 cm. length) were
obtained from the field site in
southeastern Washington state using a
steel coring device. The surface soil
(top 15 cm.) was removed from all
cores and combined, mixed, and sieved
to pass a 2-mm screen. The RC1 study
had two soil treatments including
inoculation with 2.3 x 107 colony
forming units (CPU) of RC1/g dry soil
and uninoculated. The surface soil for
the separate TK24 study was amended
with 1.0% cellulose and either
inoculated with 2.5 x 106 CPU of TK24
spores/g dry soil or left uninoculated.
All treatments were brought to a final
soil moisture content of 16% with tap
water.
The microcosms were incubated in
a growth chamber under a cycling
temperature regime that approximated
the average daily minimum temperature
during the dark cycle and the average
daily high temperature during the
photoperiod at the Hanford Site. The
lower temperature limit in the growth
chamber was 5°C. The field lysimeters
consisted of intact soil cores within
polyethylene pipe placed back into the
soil at the field site. The bottoms of
the microcosms and field lysimeters
were covered with polyester
monofilament cloth (160 mesh) to
prevent root (for RC1) and soil (for
TK24) loss from the core.
The mineralization of 14C-labelled
cellulose in microcosms and lysimeters
was determined in the TK24 study in
minitubes (4.3 cm. diameter, 24 cm.
length PVC pipe) with the 14CO2
evolved determined as a function of
time.
All treatments were replicated six
times. The RC1 study was seeded with
five "Daws" winter wheat seeds in
each soil core on October 25, 1989
that were later thinned to two wheat
seedlings. The TK24 study was started
on November 2, 1989. Microcosms
and field lysimeters were routinely
watered with tap water to a final soil
moisture content of 16%.
The survival of RC1 and TK24 in
surface soil was determined as a
function of time by plating onto their
selective media.
For the RC1 study, the populations
of RC1, Pseudomonads, and total
aerobic heterotrophs on the wheat
rhizoplane were enumerated at the
three-leaf stage of growth on seedlings
and at the boot stage of growth on
roots from three sections (top: 0 to 15
cm., middle: 15 to 35 cm., and bottom:
35 to 55 cm.). The biomass of wheat
shoots and soil dehydrogenase activity
in surface (0 to 15 cm. depth) soil were
determined at the three-leaf and boot
stage of wheat growth.
For the TK24 study, surface soil
samplings occurred in late November,
early March, early June, and early
September. The surface soil was
sampled at these four times for
populations of total aerobic hetero-
trophs and Actinomycetes, the soil
enzyme activities of dehydrogenase and
b-glucosidase, and 14C-labelled
141
-------
microbial biomass.
RESULTS AND DISCUSSION
The decline in the population of RC1
in surface soil was similar in the
microcosms and the field. However,
the population of RC1 in surface soil
was significantly higher in the
microcosms than the field at the
three-leaf stage of wheat growth (5.9
log CFU/g > 4.5 log CFU/g), while the
population decreased to the same value
at the boot stage (4.2 and 3.6 log
CFU/g for microcosms and the field).
The rate of decline was similar to the
rate found in the first year study
(Bolton et al. 1991 a), suggesting that
small differences in inoculum level (7.4
and 7.7 log CFU RC1/g in this and the
previous study) and year-to-year
climatic variability in the field did not
greatly influence the rate of decline of
RC1 in soil. There was no difference in
the colonization of wheat roots by RC1
at the three-leaf stage in microcosms or
the field, similar to first year results.
The population of RC1 on the
rhizoplane at the boot stage of wheat
was different from that of the first year
with significantly higher populations
associated with the microcosms than
the field. Also, the distribution of RC1
on the rhizoplane as a function of depth
in the microcosms was different from
the field with the population of RC1
generally higher in the microcosms than
the field.
Pseudomonad and total aerobic
heterotrophrhizoplanepopulationswere
very similar in the microcosms and
field, and similar to the previous year's
results. Shoot biomass and dehydro-
genase activity of soils were no
different in the microcosm and the field
at the three-leaf stage of wheat growth,
but were lower and higher, respect-
ively, in microcosms versus the field at
the boot stage. These results were
different from those of the previous
year where the growth chamber and
the field had a similar final biomass.
The previous year the soil dehydro-
genase activity was significantly higher
in the microcosms than the field at the
three-leaf stage of wheat growth. Soil
dehydrogenase activity at the boot
stage sampling was not determined the
first year. Year-to-year variability was
found in comparing this study with
previous years (Bolton et al., 1991 a,
1991b) including differences in the
colonization of the rhizoplane by RC1 at
the boot stage and wheat biomass
production at the boot stag. However,
for the majority of parameters, results
from microcosms and the field were
similar in both years.
Soil populations of TK24 declined
less than 2 log units over the ten-month
study. This slow decline in the soil
population of TK24 compared to RC1 is
not surprizing because TK24 is a
gram-positive spore-forming bacterium
and was added to soil as spores. There
were no differences between the
microcosms and the field for the soil
population of TK24 until after 45 weeks
when the population was 1 log unit
higher in the microcosms. At the initial
sampling in the fall, soil dehydrogenase
activity was greater in the microcosms
than the field and inoculation with
TK24 depressed activity. Dehydro-
genase activities were higher in
microcosms than in the field during the
winter and summer, but there were no
differences at the spring sampling. Soil
142
-------
B-glucosidase activity was higher in
microcosms than the field during the
winter and summer, but were the same
during the fall and spring. Aerobic
heterotroph and Actinomycete popula-
tions were higher in microcosms than in
field soils during winter and summer,
while the field had larger populations in
spring. The rate of C-cellulose
mineralization was similar in the
microcosms and the field until early
May when there was a flush of 14C02
in the field. The 14C-labelled soil
microbial biomass was significantly
larger in the microcosm than the field,
while inoculation with TK24 had no
effect-.
In summary, the microcosms
simulated the field quite well for
microbial fate and effects. Exceptions
include a flush of 14C02 from the field
soils in early May and the higher
populations of TK24 in microcosm soils
at the summer sampling, both pf which
could be attributed to differences in
temperature and water regimes. The
differences in microbial populations and
enzyme activities at the winter and
summer sampling points were attribut-
ed to the inability of the growth
chamber to mimic temperature
extremes in the field during these
periods.
The results from studies conducted
the second year of this project suggest
that microcosms can accurately
simulate the field with respect to the
fate and effects on ecosystem
structural and functional properties of
both gram-negative and gram-positive
introduced micro-organisms. However,
better controls of environmental
variables including temperature and
moisture will be necessary to more
closely simulate the field for future use
of microcosms for risk assessment.
FUTURE WORK
This project was completed and a
final report submitted to EPA-Corvallis
in March 1991.
PUBLICATIONS
Bentjen, S. A., H. Bolton, Jr., J. K.
Fredrickson, and D. J. Workman.
1990. Field calibration of intact
soil-core microcosms inoculated with
Streptomyces lividans TK24. American
Society for Microbiology Abstracts, pp.
315.
Bolton, Jr., H., J. K. Fredrickson, S. A.
Bentjen, D. J. Workman, S. W. Li, and
J. M. Thomas. 1991 a. Field
calibration of soil-core microcosms :
Fate of a genetically altered
rhizobacterium. Microb. Ecol. in press.
Bolton, Jr., H., J. K. Fredrickson, J. M.
Thomas, S. W. Li, D. J. Workman, S.
A. Bentjen, and J. L. Smith. 1991b.
Field calibration of soil-core
microcosms: Ecosystem structural and
functional comparisons. Microb. Ecol.
(in press).
Bolton, Jr., H., J. K. Fredrickson, D. J.
Workman, S. A. Bentjen, and S. W. Li.
1990. Field calibration of microcosms
to assess the fate and ecological effects
of a genetically altered rhizobacterium.
American Society for Microbiology
Abstracts pp. 314.
143
-------
Bolton, Jr., H., J. K. Fredrickson, S. A.
Bentjen, D. J. Workman, S. W. Li, and
J. M. Thomas. 1991c. Field
calibration of soil-core microcosms for
evaluating fate and effects of
genetically engineered microorganisms
in terrestrial ecosystems. Final Report.
PNL-7667, UC-403. National Technical
Information Service, Springfield, VI.
Fredrickson, J. K., S. A. Bentjen, H.
Bolton, Jr., S. W. Li, and P. Van Voris.
1989. Fate of Tn5 mutants of root
growth-inhibiting Pseudomonas sp. in
intact soil-core microcosms. Can. J.
Microbiol. 35:867-873.
Fredrickson, J. K., H. Bolton, Jr., S. A.
Bentjen, K. M. McFadden, S. W. Li, and
P. Van Voris. 1990. Evaluation of
intact soil-core microcosms for
determining potential impacts on
nutrient cycling by genetically
engineered microorganisms. Environ.
Toxicol. Chem. 9:551-558.
Fredrickson, J. K., H. Bolton, Jr., S. A.
Bentjen, D. J. Workman, and S. W. Li.
1990. Fate and ecosystem effects of a
root-colonizing Pseudomonas sp. and
Streptomyces/ividansTK.24 in soil-core
microcosms and field trials. European
Environmental Research Organization -
Molecular Microbial Ecology Workshop,
May 1, 1990.
Fredrickson, J. K., H. Bolton, Jr., D. J.
Workman, and S. A. Bentjen. 1989.
Field calibration of microcosms for
evaluating the fate and effects of
genetically engineered rhizobacteria.
American Society of Agronomy,
Agronomy Abstracts pp. 215.
Fredrickson, J. K., P. Van Voris, S. A.
Bentjen, and H. Bolton, Jr. 1990.
Terrestrial microcosms for evaluating
the environmental fate and risks
associated with the release of
chemicals or genetically engineered
microorganisms to the environment.
Haz. Assess. Chem. 7:157-202.
Fredrickson, J. K. H. Bolton, Jr., and G.
Stotsky. 1991. Methods for
evaluating the effects of GEMs on
nutrient cycling processes. Methods
for Microbial Ecology, McGraw-Hill, NY,
in press.
Fredrickson, J. K. and C. Hagedorn.
1991. Methods for evaluating the
effects of GEMs and MCPAs on
ecological processes. Methods for
Microbial Ecology, McGraw-Hill, NY, in
press.
144
-------
THE EFFECT OF A GENETICALLY ALTERED BACTERIUM
ON NITROGEN TRANSFORMATION RATES, INDIGENOUS M1CROBIOTA
AND MICROBIAL BIOMASS IN A XERIC SOIL
L. K. Gander1, E. R. Ingham2, J. D. Doyle1, and C. W. Hendricks3.
ManTech Environmental Technology, Inc.1,
U.S. Environmental Protection Agency3,
Environmental Research Laboratory
Dept. of Botany and Plant Pathology,
Oregon State University2
Corvallis, Oregon
INTRODUCTION
Genetically engineered micro-
organisms (GEMs) released into the
environment could influence microbial
populations and processes in soil, by
altering nutrient availability through
competition, removal of substrate(s),
and/or addition of metabolites. The
presence of GEMs may also alter
predation patterns and immobilization
and mineralization processes, affecting
soil tilth and fertility. Thus, an
understanding of the interrelationship
between GEM function, differences in
environment (e.g., soil type and
microbial community), and 'soil
processes is critical.
We chose to study Pseudomonas
putida PP0301(pR0103), engineered to
degrade the herbicide 2,4-dichloro-
phenoxyacetate (2,4-D); an ability not
found in its plasmidless, parental strain
(PP0301). Previous work with PP0301
(pR0103) demonstrated that total
fungal propagule numbers were
depressed (both in the absence and
especially in the presence of 2,4-D) in a
low carbon soil inoculated with this
GEM. Our study sought to both repeat
this effect, and to examine other
ecological endpoints not previously
investigated.
In this study, we determined the
dynamics of the indigenous populations
of bacteria (active and total; nitrifying
and denitrifying populations); fungi
(active and total); protozoa, and
nematodesover an 89-day period in the
following treatments: (a) unamended
and uninoculated soil, (b) 2,4rD-
amended, uninoculated soil, (c) PP0301
inoculated, unamended soil, (d)
PP0301 inoculated, 2,4-D-amended
soif, (e) PP0301(pR0103) inoculated,
unamended soil, and (f) PPO301
(pR0103) inoculated, 2,4-D-amended
soil. Additionally, biomass estimates,
concentrations of inorganic nitrogen
species, respired carbon (CO2), and pH
were determined.
MATERIALS AND METHODS '
Sail
The Xeric/Aridic Frigid soil used in
this study was collected in March 1990
from the same site (Millican Limited Use
Area, near Bend, OR), and character-
ized and stored in the same manner as
that (obtained in July 1987) used in
previous work.
145
-------
Bacterial Cultures
Pseudomonas putida PP0301
(pR0103) and its parental strain
PP0301 were derived previously from
P. putida PP0300 (ATCC 17514).
These strains are resistant to nalidixic
acid (500 jug/ml). Plasmid pR0103
encodes resistance to both tetracycline
(25 //g/mL) and mercury (25 /vg/mL),
and the constitutive degradation of 2,4-
D to chloromaleylacetate. Both strains
were maintained on either brain heart
infusion or tryptone nutrient agar
supplemented with the appropriate
selective agents. The 'presence of the
plasmid in the GEM was confirmed by
DNA extraction and agarose gel
electrophoresis prior to the start of the
experiment. Inocula of both bacterial
cultures were prepared as in previous
work.
Soil Preparation and Microcosms
Sieved (2mm-mesh) soil was
adjusted to the -33 kPa water tension,
and left to equilibrate at 4°C, with daily
mixings for 72 hours. On Day 0, the
soil was inoculated with ca. 10^ CFU/g
oven-dry equivalent (ODE) soil and
amended with 2,4-D (500 //g/g ODE
soil) where appropriate. Following
ample mixing, 60 g ODE soil was
placed into each of 144 pint-size Mason
jars and incubated at 25°C in a light-
free chamber. Each Mason jar was
closed with an air-tight lid fitted with a
rubber septum to permit headspace
sampling for C02 analysis. Carbon
dioxide measurements were conducted
by gas chromatography (Hewlett-
Packard model 5840; thermal
conductivity detector; Porapak Q
(80/100 mesh) column; helium gas
mobile phase).
Sampling Description
During the 89-day study period,
soils were sampled eight times and C02
analysis was performed 23 times. To
prevent the accumulation of metabolic-
ally inhibitory amounts of CO2 (i.e., _>.
3-4%), all jars were periodically allowed
to off-gas for 1-2 minutes in a moist
chamber. At each soil-sampling period,
three replicate jars per treatment were
processed for: % moisture; pH; 2,4-D
and dichlorophenol (DCP) concentra-
tions (by HPLC); inorganic nitrogen
analysis; enumeration of the indigenous
bacteria (nitrifying, denitrifying, active,
and total), the introduced bacteria, the
active and total fungi (including hyphal
length measurements), the protozoa,
and the nernatodes; and microbial
biomass estimations. Only one jar per
treatment was subsampled for C02
analysis, and enough jars were used
such that, except for the last three
sampling periods, none were sampled
for C02 on more than one occassion.
Enumeration of Indigenous Bacterial and
Fungal Populations
Bacteria and fungi were enumerated
both by viable plate counts (soil extract
agar + cycloheximide (100//g/mL), and
Martin's medium, respectively) and by
direct counts. Nitrifying and denitrifying
bacterial populations were estimated by
standard MPN techniques. Ten-fold
(v/v) serial dilutions of each replicate
were prepared in either 1 mM
potassium phosphate buffer, pH 7.2
(for nitrifier and denitrifier enumerations
146
-------
only) or sterile tap water (for all other
estimates, including protozoa).
Aliquots of 1-5 mi_ were removed from
the first soil dilution and prepared as
suspension-agar films for estimates of
active (fluorescein diacetate stained)
and total (phase contrast microscopy)
fungi and for hyphal length
determinations by standard methods.
Active bacterial numbers and diameters
were determined by the same method,
under oil immersion. Total numbers of
bacteria were determined using fluo-
rescein isothiocyanate direct counts.
Diameters of fungal hyphae and
bacteria were measured to convert
estimates of bacterial numbers or
fungal volumes into biomass, using the
average values of 0.41 g biomass/cm3
hyphae and 0.33 g/cm3 bacteria.
and
Enumeration of Protozoa
Nernatodes
Protozoa were enumerated by an
IVIPN technique and characterized as
flagellates, amoebae, or ciliates.
Nematode numbers were determined by
extracting 5 gram subsamples of soil in
Baermann extraction funnels.
Inorganic Nitrogen Determinations
Soil aliquots of 15 grams/sample
were extracted in 30 mL of 0.5 M
K2SO4 and filtered by gravity-flow.
Filtrates were brought to final volumes
of 30 mL, preserved in 0.2% (v/v)
H2SO4, and stored at 4°C prior to
colorimetric analysis for ammonium and
nitrate species by an autoanalyzer
(Alpkem Rapid Flow Analyzer: RFA-
300; methods A303-S021 and A303-
S170, respectively). Nitrite, unstable
under these conditions, was oxidized to
nitrate, and expressed as part of the
sum of the two species.
Statistical Comparisons
Statistical comparisons (ANOVA
and correlation coefficients) were
performed using the Statistical Package
for the Social Sciences (SPSS; Nie et.
al. 1975. 2nd ed., McGraw-Hill, NY).
Two-way analysis of variance was
performed using time and treatment as
the main effects. Only those main or
interactive effects found significant by
the F-test at P < 0.05 were consider-
ed. The mean separation test used was
Least Significant Differences calculated
for a significance level of P < 0.01.
RESULTS
Concentrations of 2.4-D and DCP
During the 89-day study, no
reduction of 2,4-D concentration was
observed in any of the amended soils.
Extraction efficiencies of 2,4-D were
ca. 36% of the added material. Also,
no DCP was detectable in 2,4-D-
amended soil inoculated with the GEM.
Soil Moisture. pH, and Respiration
In spite of the periodic opening of
jars to relieve accumulating CO2 levels,
soil moisture did not vary significantly
during the study. Soil pH values did
not vary widely over time within the
treatments, although all of the 2,4-D-
amended soils were more alkaline
(perhaps reflecting concurrent
reductions in nitrifying bacterial
populations and increases in NH4 +
147
-------
levels). Furthermore, within the 2,4-D-
amended soils, those inoculated with
either the GEM or • the parent were
significantly higher in pH than the
uninoculated soil. All 2,4-D-amended
treatments showed lower rates of CO2
evolution.
Funaal Activity. Hvohal Length and
Number of Prooagules
During the first week of the study,
some contrasting changes in bacterial
and fungal numbers occurred in
treatments with introduced bacteria,
i.e., as fungal numbers increased,
bacterial numbers decreased, and vice
versa. However, these population
changes were transient, and in no case
was a prolonged GEM-specific effect
observed. Contrary to previous
findings with PPO301(pR0103), no
decline in fungal numbers was seen in
soil inoculated with the GEM. In fact,
the numbers of fungal propagules were
significantly higher in 2,4-D-amended
soils inoculated with either the GEM or
the parent than in any other treatments,
when enumerated by the viable plate
count procedure. Counts of the active
fungi suggested, however, that these
increases were only transient (possibly
reflecting sporulation). Both direct
counts of total fungi and total hyphal
length measurements' revealed no
strong differences between any of the
treatments (2,4-D-amended or not). In
general, the hyphal lengths of active
fungi were decreased in all soils
amended with 2,4-D. .
Introduced Bacteria
The numbers of both the GEM and
parent strains ranged between approx-
imately 106 to 107 CFU/g ODE soil for
most of the study. Occassion-ally there
was a transient decrease in the number
of both strains to 105. On Day 89, the
numbers of both of the introduced
bacteria in the unamended treatments
were undetectable. The numbers of
introduced bacteria in soil amended
with 2,4-D remained constant,
suggesting that the presence of the
herbicide gave these organisms a
selective advantage. At each soil
sampling period, the GEM was pheno-
typically distinguishable from the
parental strain.
Indigenous Bacteria. Protozoa, and
Nematodes
No GEM effects, including
remediation of 2;4-D effects, were
observed. Bacterial biomass estimates
for total bacteria and (especially) active
bacteria, as well as direct counts, were
reduced in the presence of 2,4-D. The
numbers of total, nitrifying, and
denitrifying bacteria were also reduced
in the presence of 2,4-D.
Nematode numbers were only
affected by 2,4-D on Day 89. Neither
amelioration of this latent 2,4-D effect
nor significant selection for nematode
species were observed in GEM (or
parent) inoculated treatments.
The community structure of
protozoa was modified in soils
inoculated with the GEM. Two species
of amoebae appeared in greater
numbers in GEM-inoculated soils
relative to soils inoculated with the
parental strain, regardless of 2,4-D
amendment. However, protozoal
numbers were reduced in all 2,4-D-
148
-------
amended soils.
Nitrogen Pools
Nitrogen transformations in both
parent- and GEM-inoculated soils were
affected by the presence of these
organisms; however, no GEM-specific
effects were seen. Relative to the
uninoculated, unamended control,
NH4 + levels were higher in both parent-
and GEM-inoculated soils, regardless of
2,4-D amendment. After Day 10, this
was also true for (NO2~ + N03") levels
in unamended soils. The former might
be attributed to nitrogen supplied by
the introduced populations of bacteria,
but more probably reflects reduced
nitrification in the 2,4-D treatments as
nitrifying bacteria were suppressed.
The levels of (NO2~ + N03") were
consistently reduced in all 2,4-D-
amended soils. However, the levels of
NH4+ were consistently increased by
2,4-D in these soils.
DISCUSSION
Contrary to earlier work with
PPO301(pRO103), we saw no GEM-
mediated influences on 2,4-D
degradation or numbers of fungi. Yet,
the GEM maintained its phenotypic
traits throughout the experiment, and
the pR0103 plasmid has been shown
to be stable in PP0301. Given the
similarities (e.g., the same soil source
and handling; manner of both addition
and extraction/analysis from soil, as
well as concentration of 2,4-D; size and
preparation of inocula, etc.) between
these experiments, these findings are
surprising. In earlier studies, addition
of 500 ppm 2,4-D resulted in the
recovery of 66% of the added
chemical, while we recovered only 36%
of the added 2,4-D on Day 0 and in
subsequent samplings. Short et, al.
(1991) found a steady decline of 2,4-D
from 500 ppm to <100 ppm in 53
days. They attributed this to the
degradation of 2,4-D (and subsequent
appearance of the metabolite DCP) by
the added GEM.
We believe our findings may reflect
changes in Millican soil, which caused
the extraction of 2,4-D to be
compromised in our study. Of the
differences that we are aware by
chemical analysis between Millican soil
collected in 1987 and in 1990, perhaps
two are most significant. Grazing
pressure, or at least cattle usage
patterns, on this land may have
changed. Inorganic nitrogen (i.e., NO3"
and NH4 + ).levels were 2-3 times lower
in soil collected in 1990. Also, this soil
had a higher organic content (1.8 ±_
0.2% in 1990 vs. 1.3 ±_ 0.1% in
1987).
An increased (albeit low) level of
organic matter may have reduced the
extraction efficiency (and availability to
the GEM) of 2,4-D by promoting
increased soil binding and incorporation
of the compound (e.g., to humic acid).
We never recovered more than 200
ppm 2,4-D. Short et al. (1991) found
that as 2,4-D decreased below 150
ppm, the concentration of DCP
decreased as well. Although this may
reflect the instability of DCP, another
explanation could be that a minimal titer
is required to activate the catabolism of
2,4-D by PPO301(pRO103). In our
study, no DCP was produced, no 2,4-D
was removed, and no DCP effects were
observed. We speculate that not
149
-------
enough 2,4-D was available for the
GEM to be activated.
FUTURE WORK
We hypothesize that if 2,4-D were
available (i.e., extractable) in Millican
soil to PPO301(pRO103) in sufficient
concentrations, the GEM-mediated
effects of 2,4-D degradation, DCP
accumulation, and decreased fungal
populations would occur. We plan to
test this in the series of short
experiments (using Millican soil
collected in 1990), detailed below.
Unfortunately, our reserves of Millican
soil collected in 1987 are depleted.
However, prior to the initiation of this
work, we will verify that our present
strains of PPO301(pRO103) have the
ability to degrade 2,4-D in culture.
(a) Titration study of 2,4-D iri Millican
soil.
Objective: To determine that
amount of 2,4-D recoverable as 500
ppm.
(b) Comparative extraction of 2,4-D
(amount recoverable Millican soil as
500 ppm) from three different
"soils": low carbon (Millican); high
carbon (Thatuna -Naff); no carbon
(silica sand).
Objective: To verify that soil
composition affects the extraction
efficiency of 2,4-D.
(c) Small scale repeat of present study,
but using the 2,4-D amount which
can be recovered as 500 ppm.
Objective: To determine if higher
levels of 2,4-D amendment will
result in the GEM-mediated effect of
depressed fungi.
PUBLICATIONS .
Ingham, E. R., Gander, L. K., Doyle, J.
D. and C. W. Hendricks. 1991.
Assessing interactions between a
genetically engineered microorganism,
target toxic chemicals and the soil
foodweb. (manuscript in progress).
Doyle, J. D., Short, K. A., Stotzky, G.,
King, R. J. and R. H. Olsen. 1991.
Ecologically significant effects of
Pseudomonas put/da PPO301 (pR0103),
genetically engineered to degrade 2,4-
dichlorophenoxyacetate on microbial
populations and processes in soil. Can.
J. Microbiol. Accepted.
.Short, K. A., Doyle, J. D., King, R. J,,
Seidler, R. J., Stotzky, G. and R. H.
Olsen. 1991. Effects of 2,4-
dichlorophenol, a metabolite of a
genetically engineered bacterium and
2,4-dichlorophenoxyacetate on some
microorganism-mediated ecological
processes in soil. Appl. Environ.
Microbiol. 57: 412-418.
Short, K. A., Seidler, R. J., and R. H.
Olsen. 1990. Survival and degradative
capacity of Pseudomonas put/da
induced or constitutively expressing
plasmid-mediated degradation of 2,4-
dichlorophenoxyacetate in soil. Can. J.
Microbiol. 36: 821-826.
150
-------
EFFECTS OF A LIGNIN-DEGRADING RECOMBINAIMT Streptomyces ON
MICROBIAL ACTIVITY AND NUTRIENT CYCLING IN SOIL
Don L. Crawford1, Zemin Wang1, Jack D. Doyle2, Harvey Bolton, Jr.3,
James K. Fredrickson3, Steve A. Bentjen3, and Charles W. Hendricks4
Department of Bacteriology and Biochemistry1
Institute for Molecular and Agricultural Genetic Engineering
University of Idaho, Moscow, Idaho
ManTech Environmental Technology, Inc.2
Environmental Research Laboratory, Corvallis, Oregon
Battelle Pacific Northwest Laboratory3, Richland, Washington
U.S. Environmental Protection Agency4
Environmental Research Laboratory, Corvallis, Oregon
INTRODUCTION
This research was a collaborative
effort between the University of Idaho,
the Battelle .Pacific Northwest
Laboratory, and EPA's Environmental
Research Laboratory-Corvallis. The
research was designed to provide
information on the magnitude, extent,
and mechanisms by which a
recombinant lignin-degradation
enhanced Streptomyces strain effected
the integrity of a soil ecosystem, the
need for the study stemmed from
previous findings in D.L. Crawford's
laboratory at the University of Idaho
(EPA Cooperative Agreement CR-
815300-01-05) showing that
recombinant S. lividans strains
expressing a plasmid-encoded gene
coding for bacterial lignin peroxidase
(1), when released into soil, transiently,
but significantly affected the rate of
carbon mineralization in the soil as
measured by an increased rate of C02
evo.lution from the soils (2). The effect
wtas particularly significant in nonsterile
soils containing recently deposited
lignocelulosic residues. It was found
that this carbon mineralization effect
resulted from increased production of
lignin peroxidase by the recombinant in
soil. It was hypothesized that this
resulted in a transient increase in the
rate of lignin depolymerization, which in
turn more quickly opened the lignin
barrier and temporarily stimulated
lignocellulose decomposition by the
native cellulolytic soil microflora. The
most affected lignin was likely' the
relatively undegraded lignin recently
deposited in the soil. In effect, the
recombinant transiently relieved a rater
limiting, step in carbon cycling in soil,
that is lignin depolymerization. This in
turn temporarily made more cellulosic
polysaccharides available to the non-
ligniaolytic, cellulolytic soil microbial
flora. This resulted in a short term
spike in microbial metabolic activity.
In addition to carbon mineralization
effects, this genetically engineered
microorganism (GEM) might have other
effects on the soil ecosystem. The
objective of the just completed' multi-
investigator project was to determine
"what those effects, if any, were.
Ecological parameters examined
included 1) degradation of low
molecular weight aromatic and/or
151
-------
chloroaromatic compounds in the soil,
2) incorporation of organic carbon into
soil humus, 3) effects on other
biogeochemical cycles including the
nitrogen, sulfur, and p'hosphorus
cycles, and, 4) effects on total
microbial biomass, specific microbial
populations, and the activities of soil
microbial enzymes within the soil
environment. Results showed that
while no long term environmental
effects on these parameters apparently
resulted from the release of the GEM,
certain statistically significant, but
transient effects were measured.
These findings confirmed the previous
observations of Crawford's group, but
extended the findings of measurable
transient effects to several other
components of the soil ecosystem.
METHODS
The soil used was from a single
batch of a Palouse silty loam rich in
organic matter^(3). This soil was air-
dried, sieved, and split into equal lots,
which were distributed to investigators
and then stored at 4°C until used.
14C-lignin and 14C-cellulose labeled
lignocelluloses were prepared from
poplar and chemically characterized as
previously described (4). The wild-type
microbial strain used was Streptomyces
lividans TK23, a prototroph expressing
chromosomally encoded resistance to
spectinomycin. The recombinant strain
was Streptomyces lividans TK23.1
which expresses plJ702.LP encoding a
chromosomal lignin peroxidase gene of
Streptomyces viridosporus T7A. this
plasmid also encodes thiostrepton
resistance as a selectable marker.
Maintenance and characterization of the
strains was as described previously
(1,2). Enumeration of the
Streptomyces in soil was monitored via
plate counts onto selective media, using
procedures similar to those described
previously (2,3). Maintenance and
expression of plJ701.LP by the
recombinant cells in soil was also
monitored as previously described
(2,5). C02 evolution from sterile and
nonsterile control soils and from soils
inoculated with the recombinant or
wildtype strain was monitored as
described previously (2,3). 14C02
evolution and incorporation of 14C into
soil humic and fulvic acids was
measured as described previously (6,7).
In humification experiments,
incorporation of 14C-lignocellulose
derived 14C into the different fractions
of soil organic matter was examined by
extracting and analyzing the soil for14C
labeling of the soil fulvic acids, humic
acids, and humans (7). Soil microbial
biomass (C and 14C-labeled) was
quantified by the soil chloroform
fumigation technique (8), while total
soil carbon (C and 14C) was determined
by soil oxidation. Soil inorganic (NH4-N
and NOg-N) determinations were by the
method of Keeney and Nelson (9), soil
biomass nitrogen by the soil chloroform
fumigation technique (8), and total soil
nitrogen by Dumas combustion. Soil
enzyme assays included those for acid
and alkaline phosphatase, dehydro-
genase, (J-glucosidase, peroxidase,
arylsulfatase, and cellulase. These
assays were carried out as described
previously (1,10-13). In addition, in
some experiments soluble phenolics,
pH, and water content of the soils were
monitored. In some experiments
involving unamended soils or soils
152
-------
amended with lignocellulose (un-
inoculated or inoculated with the
wildtype or recombinant Streptomyces
strains), specific cultivable, aerobic
microbial populations were monitored
using plate counts, and included total
bacteria, total fungi, cellulose
decomposers, chitin utilizing bacteria,
total spore-forming bacteria, nitrifying
bacteria, and denitrifying bacteria. The
media used for these enumerations
have been described (3,9,10,13-16).
The experiments were carried out in
"Master jar" microcosms (17). All
experiments were done in 3 or more
replicates with values reported as
averages ± standard deviations.
Additional statistical analyses were also
performed on all data, and each
experiment was repeated at least two
times.
RESULTS
The objective of this research was
to evaluate the effects of release of S.
//V/tfansTK23.1 (expressing plJ702.LP)
on microbiological activities and
biogeochemical cycling or carbon and
nitrogen in soil. While some repro-
ducible, statistically significant transient
effects were indeed observed, among
the environmental parameters examin-
ed, this GEM appears to have had no
detrimental long term (> 60-90 days)
effects on the soil ecosystem.
The GEM specific enhancement'of
short term CO2 evolution rates from
soils inoculated with strain TK23.1, as
compared to control soils and those
inoculated with strain TK23 (2), was
confirmed.
For nonsterile soils inoculated with
the wildtype or recombinant strain, and
supplemented with 14C-lignin labeled
lignocellulose, the rate of 14C02
evolution from the soils was not
significantly different from the ' rate
observed for nonsterile uninoculated
soils over a 30 day periods. Thus,
neither the parent or GEM measurably
altered the net rate of lignin mineral-
ization. 14C02 evolution rates from
soils supplemented with 14C-cellulose
labeled lignocellulose were significantly
greater for soils inoculated with TK23.1
as compared to TK23; however, the
rate for TK23.1 inoculated soils was
not significantly different from the
uninoculated control soil.
After 30 days, incorporation of 14C
into the humus fractions of these same
soils was " determined. In soils
supplemented with 14C-lignih
lignocellulose, the amount of 14C-humic
acid recovered was greater for soils
inoculated with the GEM (TK23.1) than
with the parental strain (TK23), and the
value was higher for TK23-inoculated
soil than for nonsterile, uninoculated
soil. However, by 60 days this
difference had disappeared, such that
the control soil value was greater then
the TK23.1 value, which was still
greater than the TK23 value. After 30
days, the amount of 14C recovered in
fulvic acid fractions was somewhat
lower for soils inoculated with TK23
and TK23.1 as compared to the control
soil; however, after 60 days the soils
inoculated With TK23.1 contained
somewhat more 14C-fulvic acids that
either the TK23-inoculated or
uninoculated soils. For soils
supplemented with 14G-cellulose
labeled lignocellulose, after 30 days the
amount of 14C incorporated into soil
humic acids was significantly greater
153
-------
for TK23.1-inoculated soil as compared
to TK23-inoculated soil, and the value
for TK23-inoculated soil was greater
than for the uninoculated soil.
However, after 60 days there were no
significant differences between the
three soils. After 30 days the level of
14C incorporation into fulvic acids was
higher for TK23.1-inoculated soils than
for the TK23-inoculated or control soils.
Again, however, the values were not
significantly different after 60 days.
In the water content, soil pH,
soluble soil phenolics, and TK23/
Tk23.1 population studies, it was
observed that the water content of the
soil remained stable throughout the
experiments. The presence of
lignocellulose in the soil'affected soil
pH, which rose about 0.5 units over the
first 35 days of the experiment before
stabilizing. In unamended soil there
appeared to be a transient increase in
pH due to presence of the GEM. In
lignocellulose amended soil, the GEM
appeared to significantly increase soil
pH relative to controls. Thus, the
effects of lignocellulose ori the other
measured variables could be related to
this pH effect. The survival of both
TK23 and TK23.1 was similar through-
out the study. The numbers of the GEM
in unamended soil remained greater
than those in lignocellulose amended
soil. After 14 days, the numbers of the
GEM in lignocellulose amended soil
approximated those of the parental
strain in both unamended and ligno-
cellulose amended soil. Thus, the
presence of lignocellulose did not give
the GEM a selective advantage over the
parental strain. There was no signif-
icant effect of the GEM or parental
strain on the total phenolics content of
the soil in either amended or un-
amended soils over the time of the
experiment (90 days).
The enzyme studies showed some
transitory GEM specific effects.
Greater arylsulfatase activity was
present in soil amended with ligno-
cellulose as compared to unamended
soil, regardless of whether the soils
were inoculated or not inoculated.
There was a transient (3 day)
enhancing effect of the GEM on
arylsulfatase activity in both
unamended and amended soils. There
was an initial increase in acid
phosphatase activity in all of the soils,
but the activity than leveled off for the
remainder of the study. There was no
measurable effect of the GEM on acid
phosphatase activity in either amended
or unamended soils. In general a
greater amount alkaline phosphatase
activity was detected in soils amended
with lignocellulose than in unamended
soils, regardless of whether the soil
was inoculated or not inoculated.
There was typically a small increase in
activity which then leveled off for the
remainder of the study. Alkaline
phosphatase activity was, however,
transiently enhanced during the first
week of incubation in GEM-inoculated
soils amended with lignocellulose.
Similarly, soil dehydrogenase was
transiently increased in GEM inoculated
soils. Soil peroxidase activity was not
significantly affected by the GEM,
although it is likely that peroxidase was
tightly bound to lignin-derived organic
matter in the soil and therefore not
f
easily measured by the techniques
employed.
In all soils, over time there was
typically a slight increase in total
154
-------
bacterial numbers followed by a
decline. By day 91, however, the
numbers of culturable bacteria had
generally stabilized. There were
measurable increases in bacterial
numbers, as compared to the other
soils, in soils amended with ligno-
cellulose and inoculated with either the
GEM (TK23.1) or the parental strain
(TK23). However, bacterial numbers in
the soil amended with lignocellulose
and inoculated with TK23.1, were
transiently lower than for those in
lignocellulose amended soil inoculated
with TK23. When the numbers of
added bacteria (TK23 and TK23.1)
were subtracted from the numbers of
total bacteria, the numbers of total
bacteria in soils inoculated with TK23.1
were transiently lower than those in
uninoculated, unamended soils or un-
amended soils inoculated with TK23.
As expected, bacterial numbers in soils
amended with lignocellulose were
generally higher than in unamended
soils.
In all soils from time zero there was
typically an initial 10-fold increase in
culturable fungal propagules, followed
by a decline and stabilization over time.
Relative to controls, the numbers of
fungal propagules were transiently
depressed in the presence of the GEM
in soils amended with lignocellulose.
This may be seen in the result of
increased competition form the more
active bacterial microflora.
In general, from time zero there was
an initial increase in the numbers of
spore-forming bacteria, followed by a
decline, subsequent recovery, and then
stabilization of numbers. In unamended
soil inoculated with the GEM, the
numbers of spore-formers were
significantly greater then those in soil
inoculated with the parental strain, but
only between days 14 and 35. In soil
amended with lignocellulose and
inoculated with the GEM, the number of
spore-formers were fewer than in
amended soil inoculated with the
parental TK23 strain, but only on
sampling days 14. 35, and 63. By day
63, the numbers of spore-formers were
lower in soil amended with ligno-
cellulose than in the unamended soil.
The numbers of added bacteria (TK23
or TK23.1) had essentially no
measurable effect on the numbers of
spore-formers.
As would be expected, there were
some general differences in number of
cellulose decomposing bacteria in
unamended soils versus those amended
with lignocellulose. However, in
general no differences were seen in the
numbers of cellulose-degraders in soil
amended with lignocellulose and
inoculated with the GEM as compared
with those that were uninoculated or
inoculated with the parental strain.
There was, however, a long term
difference in the numbers of cellulose
degraders in unamended soils
inoculated with the GEM as compared
to unamended uninoculated soil or
unamended soil inoculated wi the
parent. When the total numbers of
added bacteria were subtracted from
the numbers of cellulose-degraders, the
effect due to the GEM in uninoculated
soil disappeared, suggesting that the
addition of TK23.1 to unamended soil
may affect the numbers of cellulose-
degrading bacteria in that one
treatment.
Again, there were some differences
between populations of chitin utilizing
155
-------
bacterial populations in soil amended
with lignocellulose as compared to
unamended soil. There was a transient
effect of the GEM in unamended soil,
but only on one sampling day (day 14).
This effect remained statistically valid
when the number of added bacteria
were subtracted from the numbers of
chitin utilizing bacteria.
In general, the numbers of nitrifying
and denitrifying bacteria were not
significantly affected by introduction of
the GEM or parental strain into any of
the soils, and no appreciable effects on
nitrogen cycling were observed.
Overall, the results show that a few
statistically significant effects on the
measured environmental parameters
were observed during the course of this
research. In all cases, the effects
tended to be small and transient.
These results generally support our
hypothesis that the GEM transiently
alters the rate of lignin depolymerization
in the soil and, thereby, temporarily
stimulates the saprophytic soil micro-
flora. The data, particularly from the
microbial populations and soil enzyme
activity studies, support this conclusion
and extend our original observation
concerning the CO2 mineralization
effect (2). The transitory nature of the
measurable effects, and also the humif i-
cation study, support the conclusion
that these effect are of little, if any,
significance to the long term stability
and health of this soil ecosystem.
REFERENCES
1) Wang, Z., B.H. Bleakley, D.L.
Crawford, G. Hertel, and F. Rafii.
1990. Cloning and expression of a
lignin peroxidase gene from
Streptomyces viridosporus in
Streptomyces lividans. J. Biotechnol.
13: 131-144,
2) Wang, Z., D.L.Crawford, T.S.
Magnuson, B.H. Bleakley, and G.
Hertel. 1991. Effects of bacterial
lignin peroxidase on organic carbon
mineralization in soil using recombinant
Streptomyces strains. Can. J.
Microbioi. 37: In Press.
3) Wang, Z., D.L. Crawford, A.L.
Pometto III, and F. Rafii. 1989.
Survival and effects of wild-type,
mutant, and recombinant Streptomyces
in a soil ecosystem. Can. J. Microbiol.
35: 535-543.
4) Crawford, R.L. and D.L. Crawford.
1988. 14C-[lignin]-lignocelluloses and
14C-milled wood lignins: preparation,
characterizations and uses.
Enzymol. 161B: 18-31.
Meth.
5) Rafii, F. and D.L. Crawford, 1988.
Transfer of conjugative plasmids and
mobilization of a nonconjugative
plasmid between Streptomyces strains
on agar and in soil. Appl. Environ.
Microbiol; 54: 1334-1340.
6) Crawford, D.L., R.L. Crawford, and
A.L. Pometto III. 1977. Preparation of
specifically labeled 14C-lignin and 14C-
cellulose labeled lignocelluloses and
their decomposition by the microflora of
soil. Appl. Environ. Microbiol. 33:
1242-1251.
7) Paul, E.A. and F.E. Clark. 1989.
Soil Microbiology and Biochemistry.
Academic Press, Inc. N.Y. pp. 91-114.
156
-------
8) Jenkinson, D.S. and D.S. Powlson.
1976. The effects of biocidal
treatments on metabolism in soil. V. A
method for measuring soil biomass.
Soil. Biol. Biochem. 8: 209-213.
9) Keeney, D.R. and D.W. Nelson.
1982. Nitrogen-inorganic forms, pp
643-698. In, A.L. Page, R.H. Miller,
and D.R. Keeney (eds.), Methods of
Soil Analysis. Part 2. Chemical and
Microbiological Properties, 2nd Edition.
Amer. Soc. Agron. Madison, Wl.
10) Tabatabai, M.A. 1982. Soil
enzymes, pp. 903-947. In, A.L. Page,
R.H. Miller, and D.R. Keeney (eds.).,
Methods of Soil Analysis. Part 2.
Chemical and Microbiological
Properties, 2nd Edition. Amer. Soc.
Agron., Madison, Wl.
11) Bartha, R. and L. Bordeleau. 1969.
Cell-free peroxidases in soil. Soil Biol.
Biochem. 1: 139-143.
12) Sichak, P.S. and A.L. Dounce.
1986. Analysis of the peroxidatic
mode of action of catalase. Arch.
Biochem. Biophys. 219: 286-295.
13) Martin, J.P. 1950. Use of acid,
Rose Bengal, and streptomycin in the
plate method for estimating soil fungi.
Soil Sci. 69: 215-232.
14) Hsu, S.C. and J.L. Lockwood.
1975. Powdered chitin agar a selective
medium for enumeration of
actinomycetes in water and soil. Appl.
Microbiol. 29: 422-426.
15) Scales, P.M. 1916. A new
method for precipitating cellulose for
cellulose agar. Zentralbl. f. Bakt. II, 44:
661. .
16) James, N. 1958. Soil extract in
soil microbiology. Can. J. Microbiol. 4:
363-370.
17) Stotsky, G. 1965. Microbial
respiration, p. 1550-1572. In, C.A.
Black, D.D. Evans, L.E. Ensminger, J.L
White, and F.E. Clark (eds), Methods of
Soil Analysis, Amer. Soc. Agron.,
Madison, Wl.
ANTICIPATED FUTURE RESEARCH
It would be of interest to examine
the long term (e.g., a year or more) fate
of the GEM (TK23.1) and its plasmid in
several different soil types. More
importantly, since humification is a
slow process, it would be of interest to
examine the effects of the GEM on the
humification process over much longer
periods of time than could be done in
the present project. Effects not seen
over 30 days might show up over
longer period of time. Finally, this
project has shown that the procedures
employed are sufficiently sensitive to
detect effects if they occur. Thus, it
would be of value to utilize these
procedures to evaluate environmental
effects of other GEMs in soil
ecosystems.
PUBLICATIONS
Crawford, d.L., Z. Wang, H. Bolton, Jr.,
J.D. Doyle, J.K. Fredrickson, S.A.
Bentjen, and C.W. Hendricks. 1991.
Effects of a recombinant lignin-
degrading Streptomyces on microbial
activity and nutrient cycling in soil.
157
-------
MECHANISMS OF EFFECTS OF RECOMBINANT Streptomyces
ON THE CARBON CYCLE IN SOIL
Don L. Crawford, Zemin Wang, and Bruce H. Bleakley
Department of Bacteriology and Biochemistry
Institute for Molecular and Agricultural Genetic Engineering
University of Idaho
Moscow, Idaho
INTRODUCTION
In our recently completed EPA
cooperative agreement CR-815300-01 -
05, we carried out research to elucidate
the mechanism by which a recombinant
Streptomyces strain, genetically
engineered to overproduce an
extracellular lignin peroxidase,
transiently enhanced the rate of carbon
turnover in soil. The gene encoding the
lignin peroxidase was obtained from the
chromosome of the lignin-solubilizing
actinomycete Streptomyces
viridosporus T7A and was cloned into
Streptomyces lividans TK64 in plasmid
plJ702 (1). This enzyme catalyzes the
oxidative depolymerization of lignin
during lignocellulose degradation and is
involved in the ligning solubilization
process of S. viridosporus (1,2). The
need for the present study stemmed
from our previous observation that
certain genetic variants of S.
viridosporus temporarily enhanced the
rate of carbon turnover in soil, as
measured by CO2 evolution rates from
soils inoculated with the variants (3).
The enhancement was particularly
significant when the variants were
released into nonsterile soils amended
with lignocellulose. Evidence implicated
overproduction of lignin peroxidase by
the variants as a possible mechanism
for the effect, since they characteristic-
ally, but transiently overproduced lignin
peroxidase as the result of spontaneous
chromosomal gene amplifications.
However, because the variations were
unstable, the mechanisms of their
effect on carbon turnover rates-could
not be easily studied. The availability
of a recombinant lignin peroxidase-
overproducing S. lividans TK64 strain
(1), however, made such a study
feasible, and it also raised the question
of how such a genetically engineered
bacterium might effect the soil carbon
cycling if it were released into the
environment. The objectives of this
past year's reseach were to elucidate
the genetic and chemical bases for the
carbon mineralization effect, using the
recombinant S. lividans strains [strain
TK64.1 expressing lignin peroxidase
gene ALip-P3 in plasmid plJ702.LP (2)].
TK64.1 was inoculated into soil, and its
effects on carbon turnover were
monitored and compared to control
soils and soils inoculated with the
nonrecombinant parental S. lividans
TK64 strain. In addition to monitoring
carbon mineralization effects, we also
examined how release of the
recombinant strain affected the rate of
incorporation of lignocellulosic carbon
into soil humus fractions.
158
-------
METHODS
The parental S. lividans TK64 and
recomblnant S. lividans TK64.1 strain
expressing lignin peroxidase-encoding
plasmid plJ702.LP have been described
previously (1,2). A gene encoding the
ALiP-P lignin peroxidase of S.
viridosporus (2) was cloned on a 4.1 kb
fragment of chromosomal DNA into
multi-copy vector PIJ702 to construct
plJ702.LP(1).
Peroxidase activity and H202
concentrations were measured as
described by Ramachandra et al. (2).
Plasmid isolations and character-
izations, Southern hybridizations, and
other molecular analyses of DNA
preparations were performed as
described by Wang eta/. (3,4). Growth
of wild-type and recombinant
Streptomyces strains in soil,
Streptomyces inoculation procedures,
selective enumeration procedures,
measurement of carbon mineralization,
and the Palous silt loam soil used in the
study have been described by Wang et
al. (3,4). . Preparation of 14C-lignin and
14C-celluiose labeled lignocelluloses
was as previously (5,6), and incor-
porated of 14C into soil humic and
fulvic acids was monitored by
extraction of the acids from soil and
analysis by liquid scintillation counting
after combustion to C02.
RESULTS
When recombinant strain S, lividans
TK64.1, which expresses a lignin
peroxidase gene in plasmid plJ702.LP
(1), was released into a rich organic soil
in, flask scale laboratory experiments,
the rate of CO2 evolution from the soil
was temporarily enhanced (4). The
enhancement was shown to be lignin
peroxidase specific, and significant only
in nonsterile soils still containing an
active microbial flora. The increase in
carbon mineralization resulted from
increased production of lignin per-
oxidase by the recombinant strain in
soil. This resulted in a transitory
increase in the rate of lignin depolymer-
ization, which in turn temporarily
stimulated lignocellulosedegradation by
the native soil microflora. The affected
lignin was likely the relatively un-
degraded lignin that had been recently
deposited in the soil. In effect, the
recombinant strain- transiently relieved
a rate-limiting step in carbon cycling in
soil, that is lignin depolymerization.
The increase in lignin depolymerization
rate in turn temporarily made more
cellulosic polysaccharides available to
the celluloytic, non-ligninolytic soil
microbial population. This resulted in a
temporary spike in the metabolic
activity of the microflora. While the
effect was transitory, this is the first
report of a genetically engineered
microorganism (GEM) having a measur-
able effect on a soil biogeochemical
cycle. Studies of the incorporation of
carbon into the humus fractions of the
soil showed that the GEM-caused en-
hancement of soil organic carbon
mineralization probably did not
significantly effect the soil humification
process. The rate of incorporation of
organic carbon into soil humic and
fulvic acids was not significantly altered
by introduction of the GEM. This
conclusion was based upon results
showing that when a rich organic soil
inoculated with the recombinant strain
was amended with 14C-Iignin or 14C-
159
-------
cellulose labeled lignocellulose,
incorporation of 14C into soil humic and
fulvic acid fractions over a 60 day
incubation period was not significantly
different as compared to control soils or
those inoculated with the non-
recombinant parental strain. Thus, the
overall effect of this GEM on the soil
ecosystem, at the level of carbon
turnover, appears to be measurable, but
short term. The observed temporary
spike in carbon mineralization probably
has little if any long term effect on soil
humification or fertility.
REFERENCES
1) Wang, Z., B.H. Bleakley, D.L.
Crawford, G. Hertel, and F. Rafii.
1990. Cloning and expression of a
lignin peroxidase gene from
Streptomyces viridosporus in
Strepomyces lividans. J. Biotechnol.
13: 131-144.
2) Ramachandra, M., D.L. Crawford,
and G. Hertel. 1988. Characterization
of an extracellular lignin peroxidase of
the lignocellulolytic actinomycete
Strep torn yces viridosp or us. A p p I.
Environ. Microbiol. 54: 3057-3063.
3) Wang, Z., D.L. Crawford, A.L.
Pometto III, and F. Rafii. 1989.
Survival and effects of wild-type,
mutant, and recombinant Streptomyces
in a soil ecosystem. Can. J. Microbiol.
35: 535-543.
4) Wang, Z., D.L. Crawford, T.S.
Magnuson, B.H. Bleakley, and G.
Hertel. 1991. Effects of bacterial
lignin peroxidase on organic carbon
mineralization in soil using recombinant
Streptomyces strains.
Microbiol. In press.
Can. J.
5) Crawford, R.L. and D.L. Crawford.
1988. 14C-[lignin]-lignocelluloses and
14C-milled wood lignins: prepration,
characterization, and uses. Meth.
Enzymol. 161B: 18-31.
6) Crawford, D.L., R. L. Crawford, and
A.L. Pometto III. 1977. Preparation of
specifically labeled 14C-lignin and 14C-
cellulose labeled lignocelluloses and
their decomposition by the microflora of
soil. Appl. Environ. Microbiol. 33:
1242-1251.
ANTICIPATED FUTURE RESEARCH
It is not currently anticipated that
additional research will be carried out
on this project. It would be of interest
to study the long term (e.g., a year)
fate of the plasmid bearing recombinant
S. lividans TK64.1 in soil, particularly
with regard to maintenance and
expression of the lignin peroxidase-
encoding plasmid. It would also be
valuable -to extend the humification
effects study out to a year or more.
Humification is a relatively slow
process, and statistically significant
effects of the GEM might not be
observed over the short periods of time
covered in the present research.
PUBLICATIONS
Wang, Z., D.L. Crawford, A.L. Pometto
III, and F. Rafii. 1989. Survival and
effects of wild-type, mutant, and
recombinant Streptomyces in a soil
ecosystem. Can. J. Microbiol. 35:
535-543.
160
-------
Wang,Z., B.H. Bleakley, D.L. Crawford,
G. Hertel, and F. Rafii. 1990. Cloning
and expression of a lignin peroxidase
gene from Streptomyces viridosporus in
Stfepomyces lividans. J. Biotechnol.
13: 131-144.
Wang, Z., D.L. Crawford, T.S.
Magnuson, B.H. Bleakley, and G.
Hertel. 1991. Effects of bacterial
lignin peroxidase on organic carbon
mineralization in soil using recombinant
Streptomyces strains.. Can. J.
Microbiol. |n press. (April).
Bleakiey, B.H. and D.L. Crawford.
1991.;Plasmid exchange and hetero-
trophic activity associated with
colonization of artificial soil aggregates
by Streptomycetes. Submitted for
Publication (April 1991).
Abstracts and Presentations
Crawford, D.L. 1989. Effects of
recombinant Streptomyces oh a soil
ecosystem. Presented, EPA
Biotechnology Risk Assessment All
Investigators Meeting and Peer Review.
EPA Corvallis Environmental Research
Laboratory, November 14-16.
Crawford, D.L. 1989. Mechanisms of
effects of recombinant Streptomyces
on the carbon cycle in soil. Presented,
89th Annual Meeting, American Society
for Microbiology. New Orleans, LA.
May 14-18.
Crawford, D.L. 1990. Effects of
recombinant Streptomyces in soil.
Presented, 90th Annual Meeting,
American Society for Microbiology,
Anaheim, CA. May 13-17.
161
-------
LABORATORY OBSERVATIONS ON THE INHIBITION OF SOIL FUNGI BY MPCAs
OF THE GENUS Pseudomonas
H.M. Abebe1, V.P. Fieland1, and R.J. Seidler2
ManTech Environmental Technology Inc1
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
Biotechnological methods have
stimulated interest in the use of
microorganisms to ease the environ-
mental impact posed by anthropogenic
toxic chemicals. One aspect of the
studies conducted in our laboratory
over the last several years, involved an
examination of the genetics, ex-
pression, and metabolic controls of 2,4-
D and phenoxyacetic acid degradation
in fluorescent pseudomonads. In one
xenobiotic model system, the use of a
deregulated operon (continuously
expresses the ability to biodegrade)
has been proposed to stimulate the
bioremediation of phenoxyacetic acid
and 2,4-Dichlorophenoxyacetic acid
(2,4-D). However, before these kinds of
modified strains (genetically engineered
microorganisms) can be field tested, it
has been proposed that a series of test
methodologies be developed to evaluate
the ecological consequences of their
releases. Such methodologies need to
be accurate, reliable, economical and
easily applicable for appropriate risk
assessment studies.
In a series of ecological studies
designed to evaluate the consequences
of 2,4-D biodegradation on microbial
ecosystems, we demonstrated an un-
expected accumulation of a metabolic
intermediate (2,4-dichlorophenol) that
reduced cell numbers of indigenous soil
fungi. Similarly, in control experiments
where 2,4-D amendments were not
made, we found evidence that the fluo-
rescent pseuddmonad, Pseudomonas
putida PP0301 also caused a decline in
the number of fungal propagules
present in nonsterile soil. Although
fluorescent pseudomonads are
important (agronomically) in the control
of fungal plant diseases such as
damping off of cotton arid take all of
wheat, very few attempts have been
noted to study the effects of these
bacteria on beneficial fungi.
The present study represents a
follow up of our initial observations and
involves a more general assessment of
the interactions between bacteria and a
variety of fungi, including beneficial
fungi of economic importance. The
initial series of fungal inhibition assay
experiments have been conducted using
agar plates and sterile soil. In the latter
case, bacteria-fungi antagonism in the
sterile soil was assessed by replica
plating onto agar plates.
METHODS
Bacterial Cultures and Media
Cultures of P. putida PPO301, with
and without plasmids, P. fluorescens
Pf-5 and VT2KTR21), and P.
162
-------
aeruginosa PA01C, with and without
plasmids, were used in this study.
Luria-Bertani (LB) agar was used for
plate counts. Liquid cultures were
grown in Brain Heart Infusion (BHI)
broth. Broth cultures were propagated
on a rotary shaker at 200 rpm.
Appropriate antibiotics were added
when selective media were needed.
Unless otherwise stated, cultures were
incubated at 30°C.
Fungal Cultures and Media
Thirty nine fungal isolates
representing free-living soil forms,
mycorrhizae, microbial pest control
agents, and biodegraders of xenobiotics
were used in this study. Except for the
ectomycorrhizae where MMN slants
were used, all fungal cultures were
maintained on potato-dextrose agar
slants at 4°C. Fungal liquid cultures
were grown in MMN broth on a rotary
shaker (200 rpm) at 25-30°C.
Fungal Inhibition Assays
In vitro bacteria-fungus plate assays
were performed on agar plates
consisting of media of different
compositions and included potato
dextrose agar (PDA) (Difco
Laboratories, Detroit, Mich.), MMN,
medium 523 and Kanner's medium.
Fluorescent pseudomonads were
patched along the perimeter of the agar
plates and incubated overnight at
30°C. Ten fj\ of fungal suspension or a
plug from the leading edge of a 5 day
old fungal culture, grown on MMN
broth or PDA, was placed at the center
of the agar plate previously patched
with the test bacteria. The plates were
then incubated at 25-30°C for 5 to 14
days. For mycorrhizal fungi, longer
incubation time (30-40 days) was
necessary before scoring for inhibition
could be made. The degree of fungal
inhibition by fluorescent pseudomonads
was expressed relative to a control
(fungus grown in the absence of
fluorescent pseudomonads) by
measuring the diameter of fungal
growth over time.
Assay of Fungal Inhibition in Sterile Soil
Sixty five grams of Willamette River
bottom soil was placed in a glass petri
dish and autoclaved three times (by
tyndallization) each for 2 hours at 121
C and 15 psi. The sterile soil was
inoculated with 20 ml of a washed cell
suspension ( 24-48 old culture) such
that the final bacterial density was 108
CFU/gram of sterile soil. Controls were
not inoculated with the test bacteria. A
free-living fungal (S4) inoculum was
prepared by culturing in MMN broth and
incubating for 5 days at 30°C and 200
rpm on a rotary shaker. After the
.bacteria were inoculated into sterile soil
in petri dishes and were incubated for
60-75 hours at 30°C, an aliquot of 0.1
ml of 8.4x103 CFU/ml of the fungus S4
was centrally inoculated per petri dish.
Each treatment was replicated at least
in triplicate. Fungal spread or growth
was monitored by replica plating onto
Martins medium agar plates. Bacterial
growth in the soil was assessed by
serial dilution plating on selective LB
agar plates.
163
-------
RESULTS AND DISCUSSION
All 39 isolates of fungi representing
diverse origins and properties were
evaluated for potential sensitivity (in
plate assays) to eight strains of Gram
negative bacteria. A scoring system of
minus (-) to four plus (4 + ) range was
used to denote the extent of inhibition.
A complete inhibition was assigned a
four plus where as insensitivity was
denoted with a minus sign.
A large variation was noted in the
extent of fungal inhibition induced by
the various bacterial strains and
different media. However, on medium
523, virtually all bacterial strains
exhibited maximum inhibition (4 + ) of
fungal growth, i.e., no fungal growth
occurred over the 7-14 days of incub-
ation. The antagonism was due to
diffusible substances expressed in both
the presence and absence of added
ferric chloride.
Susceptibility of ectomycorrhizaeto
the bacteria was evaluated on medium
MMN. Great differences were noted in
the patterns of inhibition. In general,
Pf-5 and PA013 demonstrated the
greatest levels of inhibitory effects. P.
fluorescens VT21(TR21) was the least
inhibitory. This diversity in the extent
of inhibitory activities suggests that the
different bacterial strains produce
different type(s) of inhibitory
substances.
Antifungal activity was detected in
sterile soil inoculated with the free-
living soil fungus (S4) and various
strains of P. fluorescens and P.
aeruginosa. Several significant features
were derived from these experiments.
P. aeruginosa strongly inhibited the
growth of the fungus in sterile soil
medium. Growth inhibition was noted
over the 25 day observation period.
Maximum growth of the fungus oc-
curred in the control soil (without the
bacterium) and reached the perimeter of
the plate within 8 days of incubation.
PA013 (pR0103) exhibited significantly
less fungal inhibition implying the
plasmid somehow interfered with
production of the inhibitory
substance(s). Numbers of the Pf-5
inoculum declined 10-fold by day 15.
This decline, in bacterial numbers
coincided with slightly greater spread of
the S4 fungus. On the other hand, S4
growth in the presence of VT21 (TR21)
was only slowed and by day .21
covered as much of the spi'l. as the
uninoculated control following a 50-fold
reduction of bacterial numbers in the
soil.
FUTURE WORK
Studies are underway to develop
assays using nonsterile soil and plants
to evaluate the interaction of
pseudomonads with indigenous soil
microflora, including fungi. Studies will
be conducted to ascertain the type(s) of
chemicals produced in the rhizosphere
environment that alter the composition
of the ecosystem. Fungus-bacterium
antagonism and its autecological
significance will, be evaluated using
fungus and fluorescent pseudomonads
both marked with stably expressed
antibiotic labels.
PUBLICATIONS
This study was initiated this fiscal
year and there are no publications at
this time.
164
-------
EFFECTS OF ENVIRONMENTAL FACTORS ON BACTERIAL CONJUGATION
M. A. Gealt, T. A. Khalil and S. Selvaratnam
Drexel University
Department of Bioscience and Biotechnology
Philadelphia, PA 19104
INTRODUCTION
The intentional release of bacteria to
foreign habitats is common. The
agricultural use of bacteria, e.g.,
Rhizobia to enhance nitrogen fixation,
has resulted in the introduction of high
concentrations of reactor grown
bacteria directly into the natural
environment. The enhancement of
bioremediation capability of indigenous
bacte'ria by augmentation with bacteria
containing specific gene sequences has
been proposed. Pathogens contained
within the effluents from sewage
treatment facilities are used for soil
emendation. There are many proposals
for, and have been some tests of, the
release of bacteria modified by genetic
engineering techniques. In addition to
scheduled releases of bacteria, many
possibilities exist for the unscheduled
release resulting from breaches in
industrial reactor vessels or transport
from release sites by foraging animals.
Many bacteria from scheduled or
unscheduled releases will end up in
either agricultural or industrial
wastewater. Wastewater also is a
likely place for genetic interactions to
take place because it contains a high
bacterial titer, and, during the treatment
process, the bacteria will be exposed to
different environmental conditions,
including aerobic, anaerobic, etc.,
which may enhance or curtail gene
transfer. Conjugation may occur using
wastewater-resident conjugative
plasmids; it is also possible for non-
conjugative plasmids to participate in
conjugal transfer by the mobilization
process.
Conjugation requires the presence
of specific genes, e.g., the tra genes of
conjugative plasmids and the mob
genes of mobilizable non-conjugative
plasmids. In addition to the presence
of these genes needed for conjugation
or mobilization, bacteria must be able to
produce functioning gene products in
their environment. In order to predict
the possible escape of gene sequences
from released bacteria to indigenous
populations it is necessary to know
how specific components of an
environment affect the transcriptional
(and translatio'nal) control mechanisms
for the production of these products.
We report here the effect of different
factors on transcription of tra and mob
as might occur in wastewater.
METHODS
Bacteria and Plasmids
E. coli chi1784 that contains the
plasmid R100-1 was used to determine
inhibition of conjugation by exogenous
compounds, with E. coli chi1997,
which is plasmidless, serving as the
recipient. E. co//HB101 was used as
host for plasmids F, R100-1, RP4, R6K,
N3, RA1, R144, R27, and colEI-amp.
165
-------
£. co// DH5alpha was used as host of
plasmid pTK31. This reporter plasmid
contained traP promoter sequences of
plasmid pFL114 fused to the facZ of
plasmid pMC1403. This construct en-
abled us to monitor the effect on the
transcription through the traP promoter
of chemicals coincubated with the
bacteria.
The bacteria were grown on
primarily on L-broth supplemented with
appropriate antibiotics. Some
conjugation experiments were
performed in sterile synthetic
wastewater. Compounds tested for
their effect on conjugation or beta-
galactosidase activity were added to
actively growing cultures. Beta-
galactosidaseactivity was monitored by
method of Miller (1972) in which cells
were perme'abilized, then induced with
o-nitrophenol-beta-D-galactoside
(ONPG). The developing yellow color
was monitored at A420-
Transcription of the mob Genes
In order to determine if the
mobilization specific mob genes were
being transcribed, total RNA was
extracted from the cells and probed by
slot blot with oligonucleotides specific
for either the mob2 or the mob3 (and
mob5) gene sequences. Comparisons
were made of extracts from bacteria
containing either a conjugative plasmid
(F, R100-1, RP4, R6K, N3, R144, RA1
or R27), the colEI-amp plasmid, or both
of these plasmids. Prior to binding to
the membrane, extracts treated with
either DNase or RNase to destroy one
nucleic acid or the other. The absence
of DNA or RNA was confirmed by gel
electrophoresis.
RESULTS AND DISCUSSION
Effect of Environmental Factors on
Conjugation
The chemical composition of
wastewater can have significant impact
on the probability of gene transfer by
conjugation. Significant inhibition (>3
logs) was exhibited in the presence of
millimblar concentrations of Zn2 + ,
Fe3 + , and SDS, which are relatively
common (especially of Fe3 + , which is
often added to enhance precipitation).
It has been proposed that the effect of
zinc is directly on the receptor of the
pilus while SDS affects the structural
integrity of the pilus. Significant
inhibition, but at much higher concen-
trations, was exhibited by Ca2 + ,
Mg2 + , Na+and K + . No inhibition was
observed with Triton X-100.
To see if any of these inhibitors
specifically affected the transcription of
traP, the major regulatory gene of the
tra operon of the derepressed IncFII
plasmid R100-1, a fusion was made of
the promotor of traP and the /acZ beta-
galactosidase reporter gene. In this
way the effect of external compounds
on the transcription of traP was
monitored. Comparisons were made
between the effect of compounds on
lacZ. genes located on this fusion
plasmid with the lacZ. located other
plasmids or on the chromosome (using
ONPG as an inducer). All constructs
showed similar inhibition of beta-
galactosidase transcription 'as a
function of temperature and Zn2 +
concentration. Fe +, however, showed
decreased induction of beta-galacto-
sidase from the traP fusion, but an
increase in the amount of enzyme
166
-------
induced when facZ was under its
natural promoter. Therefore, at least
part of the decrease in conjugation
frequency which was observed when
ferric ions were present was probably
due to the inhibition of traP
transcription.
Induction of mob Genes
Some of the dissemination of genes
in natural environments is due to the
mobilization of non-conjugative
plasmids. It is possible that some
environmental elements will specifically
affect mobilization of these non-
conjugative plasmids, but not the
functions on conjugative plasmids. We
used ColEI, a common non-conjugative
plasmid, to analyze these interactions.
Becausethisnaturally-occurringplasmid
had no easily selectable gene, we used
a modified ColEI that contains the beta-
lactamase gene of plasmid pBR322.
This ColEI-amp plasmid is mobilizable
because it contains mob genes as well
as a bom site, the origin of transfer
during transfer to a recipient cell. In
particular, ColEI has the mob2, 3 and 5
genes. The mob5 gene is wholly
contained within the mob3 gene
sequences. Under conditions in which
ColEI is the sole plasmid in a strain
grown in monoculture, there was only
trace synthesis of RNA from the mob
sequences. When a conjugative
plasmid, however, was introduced into
this cell containing the mob sequences,
transcription initiated. This
transcription depended solely on the
presence of the conjugative plasmid and
was not dependent on the presence of
a potential recipient cell. The induction
was found to occur with several
plasmids from different incompatibility
groups (F [IncF], N3 [IncN], RA1
[IncA/C], RP4 [lncP]I, R6K [IncX], R27
[IncH], R100-1 [IncFII], R144
[Inclalpha]). The induction by such a
variety of conjugative plasmids
suggests that there is a common, and
perhaps universal, product of the
conjugative plasmids which induces the
synthesis of the mob gene products.
FUTURE WORK
Effect of Environmental Factors on mob
Induction
In a manner similar to the effect of
Fe3+ on transcription at the traP
promoter, it is possible that various
environmental components would have
an effect on induction of mob. We will
analyze the effect of various
compounds common to wastewater on
the mobilization of ColEI and on mob
transcription. This would aid in
predicting probability of conjugative
plasmids mobilizing non-conjugative
plasmids.
Isolation and Characterization of mob
Effector
Although eight conjugative
plasmids have shown the capability of
inducing mob function, the universality
of this induction by plasmids from
different incompatibility groups has yet
to be completely determined. We will
analyze frequency of colEI mobilization
for conjugative plasmids studied and
correlate this mobilization with the
inducton of mob gene transcription.
The induction of mob genes by
plasmids from several different
167
-------
incompatibility groups suggest that
there may be a common effector
molecule. It would be of great interest,
therefore, to know whether there is
similarity in sequence between all such
inducing plasmids. A common
sequence for such a regulator gene
would be of interest to (a) quantitatively
and specifically monitor by hybridization
analysis the presence of the total
population of conjugative plasmids
without knowing which plasmids were
present and (b) assess the amount of
conjugative activity which is going on in
wastewater and other natural
environments by the amount of active
product (RNA or protein) which cells
are synthesizing in situ.
Evaluation of Metabolic State
Necessary for Conjugation
The synthesis of proteins necessary
for conjugation may be a function of
the metabolic state of the cell. It'is
unknown whether all conjugation
occurs in the initial period following
release. It may be possible that
following 24 hours in wastewater no
more conjugation will occur. In this
case the risk of release is immediate,
but has a short half life. Alternatively
the intention of releasing plasmid
containing bacteria with the possibility
of the gene transfer into the native
population would be limited.
PUBLICATIONS
Gealt, M. A. 1991. Gene transfer in
waste water, in: M. A. Levin, R. J.
Seidler and M. Rogul (eds.), Microbial
Ecology: Principles, Methods and
Applications. McGraw-Hill: NY (in
press).
Gealt, M. A. 1988. Recombinant DNA
plasmid transmission to indigenous
organisms during waste treatment.
Water Sci. Tech. 20: 179-184.
Khalil, J. and M. A. Gealt. 1987.
Effect of exogenous compounds on the
mobilization of plasmids in synthetic
wastewater. Can. J. Microbiol. 33:
733-737.
Mancini, P., S. Fertels, D. R. Nave, and
M. A. Gealt. 1987. Mobilization of
pHSV106 from Escherichia coli in a
laboratory waste treatment facility.
Appl. Environ. Microbiol. 53: 665- 671.
168
-------
TEST PROCEDURES FOR ASSESSING HAZARDS OF
MICROBIAL PEST CONTROL AGENTS TO FRESHWATER FISH
Virginia M. Snarski
U.S. Environmental Protection Agency
Environmental Research Laboratory
Duluth, Minnesota
INTRODUCTION
Widespread concern over the
environmental safety of chemical
pesticides has stimulated interest in the
development and use of microbial pest
control agents (MPCAs). Selected
because of their host . specificity,
MPCAs (bacteria, fungi, viruses, and
protozoa) are generally regarded as
safer alternatives to some environ-
mentally-hazardouschemicalpesticides.
However, as living organisms, MPCAs
pose concerns not shared by chemicals,
ie., their survival, multiplication, and
dissemination in the environment and
the potential to cause infection or
disease in nontarget species.
As mandated under the Federal
Insecticide, Fungicide, and Rodenticide
Act (FIFRA), EPA requires testing as
described in Subdivision M of the
Pesticide Assessment Guidelines to
address potential ecological and human
health effects prior to the registration of
a MPCA. These requirements include
determining the safety of MPCAs to
nontarget aquatic species, including
fish.
While test protocols are available for
assessing effects of microbial agents on
fish, basic information is needed on the
factors influencing the extent and kinds
of interactions between microbes and
fish to better understand and predict
potential effects and environmental
exposure of nontarget fish populations,
The studies presented here investigated
the interactions between fish and the
registered MPCA, Bacillus thuringiensis
subsp. israelensis (Bti) under laboratory
conditions.
METHODS
Single species exposures were
conducted with two commercial Bti
spore-crystal formulations, Vectobac-G
(Abbott Laboratories) and Mosquito
Attack (Reuter Laboratories). Exposure
concentrations used ranged from 104 to
106 CFU/ml, based on spore counts,
and were at and above recommended
field application rates.
Four species of freshwater fish
were used: fathead minnows,
Pimephales promelas, bluegill sunfish,
Lepomis macrochirus, and two
salmon ids, rainbow trout,
Oncorhynchusmykiss, and brook trout,
Salvelinus fontinalis.
Bti formulation was added to the
water and fish were exposed under
static or static renewal conditions for
periods of from one hour to 30 days.
Following exposure fish were removed,
rinsed in clean water, and sampled to
determine the extent of Bti accumu-
lation or transferred to clean water to
study clearance of Bti.
169
-------
Fish samples, either whole body
(WB) or dissected tissues, were
homogenized, generally pasteurized
(65°C for 20 min.) to reduce back-
ground vegetative cells, and plated into
tryptic soy agar (TSA). Typical .Bti
colonies were counted following a 24
hour incubation at 32°C.
To determine if germination and
colonization occurred within the fish,
homogenates were subdivided and
plated with and without prior
pasteurization onto a selective,
differential media, PEMBAC (developed
by Dr. Allan Yousten, VA Polytechnic
Institute and State Univ., Blacksburg,
VA). Comparison of total Bti (non-
pasteurized samples) and Bti spore
(pasteurized) counts was used to
determine if germination occurred.
RESULTS AND DISCUSSION
Concentration-dependent dissolved
oxygen depletion with both formu-
lations, believed to result from
stimulation of indigenous aquatic
bacteria by nutrients in the formu-
lations, resulted in mortality of all fish
within 24 hours at and above water
concentrations of 2-3x106 CFU/ml;
partial mortality of trout occurred at
5x105 CFU/ml. No direct adverse
effects of Bti on fish were observed.
Rapid accumulation of Bti spores by
all fish tested was measured following
addition of either formulation to their
water. No differences in responses
were detected to the two Bti formu-
lations. Plate counts on dissected guts
were nearly identical to WB counts in
all species showing that ingestion was
the major exposure route. Skin swabs
and gills had counts orders of
magnitude lower than gut samples. For
example, fathead minnows whole body
and gut samples contained 105 to 10
CPU/fish while gill samples were 102 to
103 CFU/gill.
Differences in relative accumulation
of Bti spores by species was related to
feeding habits and food size prefer-
ences. The omnivorous fathead
minnows actively ingested the small
formulation particles resulting in WB
counts that were two to four orders of
magnitude higher than counts in the
bluegill and trout.
Upon transfer to clean water, rapid
elimination of Bti spores occurred
through fecal elimination. Maximum
loss occurred within the first 24 hours
after transfer; near complete elimination
occurred in 4 to 10 days if reingestion
did not occur.
A decrease in WB spore counts
with increasing exposure time and the
presence of spores in feces for over a
week after spore counts on tissue
homogenates were negative, led to
studies to investigate germination
within the fish. Comparisons of total
Bti counts and spore counts of
subsamples of tissue homogenates
using PEMBAC agar demonstrated that
germination of spores occurred within
the fish gut. Attempts to identify
vegetative cells within these samples
were not successful. It may be that
even though spores germinated (as
shown by loss of heat-resistance), they
did not continue through outgrowth to
form vegetative cells but formed spores
in feces via the process of microcycle
sporulation.
No evidence of increases in total
counts with time were shown suggest-
ing that colonization of the fish guts did
170
-------
not occur. Results of these studies
showing rapid accumulation of spores
directly from the water and short-term
retention of Bti suggests that fish may
have signficant infuence of the
persistance and dissemination of Bti,
and possibly other microbes, within an
aquatic ecoysystem.
FUTURE WORK
We are planning additional
laboratory studies on fish-MPCA
interactions, including work with
fungus (possibly Lagenidiumgiganteum)
and other bacterial species, including
non-sporeformers (possibly with several
species of Pseudomonas). Plans are
underway to evaluate additional
endpoints ( h istolog ical and
physiological/immunological) for our
effects testing procedures. Studies
comparing laboratory-derived data with
that from small scale field studies will
also be conducted in conjunction with
other members of ERL-Duluth's
research program.
In addition, studies will be begun to
develop procedures to study the
influence of stressors of environmental
and physiological origin on the
susceptibility of fish to microbial
exposure. As part of these procedures,
positive controls for infectivity/
pathogenicity endpoints will be
developed by including exposures to
known fish pathogens.
PUBLICATION^
Snarski, V. M. 1990. Interactions
between Bacillus thuringiensis subsp.
israelensis and fathead minnows,
Pimephalespromelas Rafinesque, under
laboratory conditions. Appl. Environ.
Microbiol., 56: 2618-2622.
171
-------
INTERACTIONS OF NON-TARGET INVERTEBRATES WITH
Bacillus thuringiensis var. israelensis IN NATURAL PONDS
Richard L. Anderson1, Eric Mead2, Lyle J. Shannon3, and David M. Janssen3,
U.S. Environmental Protection Agency1
American Scientific International2
Environmental Research Laboratory-Duluth,
Duluth, Minnesota
Biology Department3
University of Minnesota-Duluth
Duluth, Minnesota
INTRODUCTION
. A goal of the MPCA research
program at ERL-D is to improve the
confidence of predicting effects in
natural systems using laboratory
derived data. Our research has three
projects. Each of these is represented in
this section. One project is to develop
and evaluate single species test
systems. Research has been conducted
with fish and invertebrates exposed to
either fungi or bacteria spore-based
microbial pest control agents (MPCAs).
The second project is to develop and
evaluate multiple species/microcosm
tests. The third project is to conduct
fate and effect studies in natural
systems treated with the MPCA. The
natural system study is the topic of this
presentation.
Few studies exist which are
specifically designed to evaluate
non-lethal interactions of non-target
invertebrates with MPCAs in natural
aquatic systems. Most studies were
designed to measure only effects on
either target or non-target animals. We
propose to develop general information
on how an added microorganism
interacts with both the target and
non-target animals. The objective of
this project was to determine, in
temporary pools treated to control
mosquitoes, whether or not non-target
animals accumulated and retained the
spores of Bti. •
METHODS , ,.
Site and Application
• The study was conducted in three
pools located in a wooded area near
Duluth, MN. Two pools were located in
an open grassy area, while the third
pool was more enclosed by trees. The
area contains many small depressions
(2-10 m in diameter) that fill with snow
melt and rain in the spring and usually
remain wet for up to 2 months. During
the summer the pools were occasionally
filled with run-off from rain. All pools
contained emergent cattails and all
were heavy mosquito producers.
Three pools were divided with
Scrimweave, a rip-stop -woven
polyolefin material. Wooden walkways
were constructed to provide access to
inner pool areas to enable sampling
without disturbing the pool sediments.
One side of each pond was treated
with Vectobac, a formulation consisting
of corncob particles covered with
172
-------
Bacillus thuringiensis var israelensis
(Bti). The amount added to each pond
was specifically calculated for each site
with a target of 104 Bti spores per ml
of pond water. The formulation is
designed to disperse the agent
throughout the water. The particles
float for a short time before sinking.
Sampling
Samples of water and invertebrates
were collected from both sections of
each pool. The aquatic animal samples
were collected by sweeping the water
with a fine mesh net and transferring
the animals to a sterile glass jar to be
transferred to the laboratory for
analysis. Emerging insects were
captured and transferred to the
laboratory in sterile glass jars.
At the laboratory, invertebrates
from the aquatic samples were removed
from the jar and separate species were
placed in sterile glass test-tubes. The
animals were washed to remove any
loosely held Bti by removing the pool
water with a pipet, then rinsed with
sterile water. Each sample was rinsed
twice, after the water had been
removed. A sample may consist of a
single large individual, or for smaller
animals such as daphnids, a pool of up
to 10 animals. After rinsing twice, all
samples were frozen in 5 ml of sterile
water. In all cases, the values were
adjusted to reflect the number of Bti
spores (as Colony Forming Units, CPU)
per individual.
Each sample to be processed was
thawed and poured into a sterile plastic
bag. The sample tube was rinsed with
5 ml of sterile water which was added
to the bag and the sample was blended
in a Stomacher™. The sample was
removed from the bag with a sterile
pipet, heat shocked, cooled and
pour-plated at dilutions of 10~2 and
10~4 with Tryptic Soy Agar.
The Bti content of emerged insects
was measured after returning the
insects to the laboratory. Individuals are
placed .in a Stomacher bag, frozen,
thawed and blended in sterile water
with the Stomacher. The blended
sample is heat-shocked and pour plated
on Tryptic Soy Agar.
RESULTS AND DISCUSSION
The concentration of Btim samples
of water, and in insects, zooplankton,
snails and emergent insects were
measured. Six taxa of insects, 3 taxa of
zooplankton and 1 taxon of snails were
collected with sufficient frequency to
provide data throughout the summer.
Four hundred and forty-nine samples
were processed. Of that total, 104
samples were adult insects. Accumu-
lation differed for each taxa and the
extent of the accumulation might have
been affected by pool characteristics.
Insects
Uptake of Bti is rapid in insects.
Insects sampled soon after the
application showed the highest
concentrations. Within several days,
many of the insect taxa had lost most
of their accumulated ,Bti. However,
there were some samples that
measured uncharacteristically higher
than samples taken before or after that
specific sample. The insects collected
from the pools were predaceous,
therefore ingestion of prey that
173
-------
contained Bti might account for the
higher values.
Zooplankton
Zooplankton also rapidly ac-
cumulated Bti. The highest concen-
tration in any of the animal samples
was found in Lynceus, (Conchostraca:
clamshrimp) which rapidly accumulated
up to 30,000 CPU/ animal. The
daphnids in pool 1 contained only a
small amount of Bti when the four-hour
sample was taken but by 24 hours the
average count was about 2000 spores
per animal. The amount of Bti in the
daphnids decreased through the next
two sampling periods until at day 7 the
content was about 300 spores per
animal. At 14 days post-treatment the
count had decreased to about 100
spores per animal. The pool dried up
before the next scheduled sampling
date.
Pool 3 was deeper and larger than
pool 1 and samples could be taken
throughout the season. The rate and
extent of uptake by daphnids in this
pool was much different than was
measured in pool 1. The accumulation
did not show a peak but remained, flat
throughout the monitoring period. The
content never did exceed 300 - 400
spores per animal in this pool.
The differences in uptake between
daphnids in each pool may be due to
pool structure. Pool 1 was shallow, less
than 1 m in depth, while pool 3 was
deeper and contained about 3-times as
much water volume. In both pools,
although the Bti was measured in the
water it ultimately accumulated on the
surface of the sediment. In the
shallower pool the animals would have
been exposed both in the water and by
contact with the sediment. In the
deeper pool, those daphnids in the
water column would have less contact
with the higher concentrations
associated with the sediment. The
consistency in values after 7 days may
indicate a minimum value for
zooplankton which accumulated Bti.
Lynceus and other zooplankton are
detritus or filter feeders. These forms
would easily collect Bti either from the
water or from feeding on the surface of
the sediment. Within 6 to 8 days most
zooplankton had lost over 80% of the
accumulated Bti. A low concentration
of Bti was found in these animals the
remainder of the season.
The last of the aquatic forms
analyzed in this study were Planoribidae
snails. The concentration of Bti found in
these animals did not increase rapidly
soon after the application and did not
decrease as was demonstrated in other
animals. This may be related to feeding
practices or contact with sediment
surfaces. If feeding was on or near a
cob particle, high Bti concentrations
would be expected.
Adult Insect
Adult insects were also collected
and analyzed. Early in the season, the
samples were largely chironomids and
mosquitoes. Later in the season, most
adults were odonates, either dragonflies
or damselflies. A few caddisflies were
also collected and analyzed. Only one
chironomid sample contained Bti and,
as was expected, none of the
mosquitoes contained Bti. Bti was
found in a few damselfly adults but the
greatest concentrations were found in
174
-------
dragonflies. Dragonfly content ranged
from less than 100 to more than 1000
CPU per individual. Most of the samples
contained between 100 and 500 CPU
per individual. The highest concen-
trations were found in dragonfly
exuviae which were up to 10 times
higher than in the emergent adult.
The adult insect collections show
that the Bti spore could move from the
site of application either in or on the
bodies of some emerging insects. The
data also show that the amount carried
by each insect is highly variable and
that not all emerging insects contain the
spore. Odonates may contain Bti
because of their predatory feeding
habits. Direct ingestion of particles
may also result in Bti accumulation. A
study this year will begin to evaluate
some of these issues.
FUTURE
An integral part of the risk
assessment of microbial pest control
agents must be to evaluate the
predictive value of laboratory data. This
data begins to describe how inverte-
brates can affect distribution of an
MPCA in natural systems. This year's
uptake and distribution research will
include another application to the
ponds. Special emphasis will be placed
on the movement of the spore out of
the system by emerging insects.
Long-term plans are to use the divided
pond method to evaluate other MPCAs
and to understand how the invertebrate
and vertebrate populations affect the
MPCA survival and distribution.
PUBLICATIONS
Nestrud, Lori B. and Richard L.
Anderson. 1990. Protocols for
Exposing .Freshwater Fish and
Invertebrates to Fungi used as Pest
Control Agents. Environmental
Research Laboratory-Duluth, Duluth,
MM. (Report describes techniques for
exposing and measuring effects on 8
species of invertebrates and one fish
species)
175
-------
FIELD VALIDATION OF LABORATORY MICROCOSMS USING
Bacillus thuringiensis var. israelensis
L.J. Shannon1, D.M. Janssen1, and R.L. Anderson2
Biology Department1
University of Minnesota-Duluth
U.S. Environmental Protection Agency2
Environmental Research Laboratory-Duluth
Duluth, Minnesota
INTRODUCTION
The potential for negative
environmental effects resulting from the
release of genetically engineered
organisms (GEMs) is a source of
concern for both regulators and the
public. Unfortunately, it is not yet
possible to predict the survival or
ecological effects of an introduced
microorganism released into the
environment. Single species laboratory
exposures can measure direct toxicity
or pathogenicity of a new micro-
organism toward a given species, but
they provide little information on the
critical questions of survival and
ecological effects. More complex tests
are needed to predict effects at the
community and ecosystem level.
Microcosms, contained model
ecosystems, should be important tools
for answering these ecological
concerns. A variety of terrestrial and
aquatic microcosms have been used in
chemical testing but their use for
evaluating GEMs has been limited. This
reflects a shortage of validated
protocols. Although many microcosm
procedures have been proposed by
various investigators, lack of
information about sensitivity, repeat-
ability, and similarity to the "real world"
has precluded their widespread use.
Indeed, none of these procedures
should be used until we can demon-
strate that (1) they are sensitive enough
to detect effects, (2) they give the
same answer twice, and (3) they ac-
curately mirror events in natural
systems.
We have developed two aquatic
microcosm protocols of differing
complexity. Both have performed well
in lab tests with Bacillus thuringiensis
var. israelensis (Bti). We are now in the
process of duplicating lab tests in the
field. Preliminary analyses of field data
are encouraging. Although the two
microcosms showed differences in their
ability to model the natural ponds
(calibration), both correctly predicted
that Bti would survive for months,
would kill target organisms (mosquito
larvae) but would cause little ecological
impact.
METHODS
Microcosms
MFC microcosms: Our original
microcosms were modifications of
systems originally developed for
chemical testing by John Leffler. After
extensive testing of these "mixed flask
culture" (MFC) microcosms we
developed a protocol for evaluating
176
-------
survival and ecological effects of
microorganisms added to aquatic
communities. These microcosms are
small (1-Liter), easily maintained
systems that can support a variety of
bacteria, protozoa, algae, zooplankton
and insect species. The biotic
community in these systems is a
species assemblage collected from the
wild and allowed to co-adapt in the
laboratory.
- Core microcosms: While the MFC
protocol performed adequately in tests
with Bti, it was not a close surrogate
model of a natural pond. Among other
differences it lacks a complex organic
sediment, is highly eutrophic and
wholly aerobic.
As an alternative, we constructed
new microcosms from intact sediment
cores. Cores were collected from a dry
temporary pond known to contain large
populations of zooplankton, insects,
-and molluscs. We anticipated that
resting stages of most of these
organisms would be present in the dry
cores. Cores 75mm in diameter and 40
;mm deep were removed with a PVC
corer and transferred to wide-mouthed
glass canning jars (Ball Corp.). The
rehydrated jars were held in a
Percivaltm Growth chamber at 20°C on
a 12/12 hour - light/dark cycle.
Field Sites
Field tests were conducted at a
remote location owned by the Duluth,
MN International - Airport. Three
temporary ponds (vernal pools) were
used in the field tests. These three
were selected on the basis of their
similar water chemistries and
populationsof zooplankton, protozoans,
snails, and insects (including
mdsquitos). Each pond was sub-
divided with a 1-M high barrier
"T"IV A
constructed of Scrimweave , a
rip-stop woven polyplefin material.
Wooden walkways were constructed
over each pond to provide better access
and permit sampling without disturbing
the sediments of the pond. One side of
each divided pond was treated with
Vectobactm (corn cob grits treated with
Bti) to yield a water concentration of
approximately 104 Bti spores per ml.
This material was broadcast across the
pond with a spreader constructed from
a one-gallon plastic bottle.
Test Endpoints . '
The same test end points were used
in both lab and field tests.
Fate of Bti
Survival of Bti was monitored daily
for the first two weeks after appli-
cation, and weekly thereafter. Bti
density was determined in both water
and sediment by heat-shocking 65°C
and plating appropriate dilutions on
Tryptic Soy Agar (TSA). Both total
colonies and typical Bti colonies were
counted. Plates that had suspicious Bti
colonies were bioassayed after the
plates were exposed at 25°C for three
days to allow germination of spores.
In addition to water and sediment
samples, animals were periodically
collected and analyzed for Bti content.
For large organisms (e.g. tadpoles,
insects, snails) only one individual was
used per sample. Smaller organisms
(e.g., zooplankters) were analyzed in
groups of five, so that Bti concen-
177
-------
trations would be sufficient for
detection. These groups were washed
by suspending them in distilled water,
shaking vigorously and filtering through
a sterile membrane filter. Tadpoles and
snails were then homogenized in 5 ml
of sterile water using a Tissumizer
(Tekmar Corp., Cincinnati, OH). Other
organisms were ground in 5ml of water
using a Ten Broek manual tissue
grinder. All ground samples were then
pasteurized for 20 minutes at 65 °C.
Cooled samples were then diluted, on
TSA agar and incubated overnight at
30°C.
The duration of Bti toxicity to
mosquito larvae was determined with
daily bioassays on water samples
collected from microcosms and ponds.
Bioassays were done in the lab in a
series of small glass vials containing 5
ml sample and 5 Aedes atropalpus
larvae. The bioassays were read at 18
hours and held for another 24 hours to
observe any delayed killing activity.
Bioassays for each pond were run a
week prior to treatment and daily until
the samples were no longer toxic.
Ecological Effects
A suite of physical, chemical and
biological measurements were made
both to characterize the systems and to
detect effects. These are grouped into
functional (related to processes within
an ecosystem) and structural
(populations) variables. Functional
variables included the following
measurements:
Carbon Cycling
Dissolved Organic Carbon (DOC)
Total Organic Carbon (TOC)
Nitrogen Cycling
Ammonia (NH3)
Nitrate + Nitrite
Total Nitrogen
Phosphorus Cycling
Ortho-Phosphate
Total Phosphate
Primary Production
Diel Oxygen Gain
Community Respiration
Diel Oxygen Loss
Structural variables included
measurements of populations of
bacteria, protozoa, rotifers, algae,
cladocerans, copepods, ostracods, and
macroinvertebrates. Temperature and
pH were also measured routinely.
RESULTS AND DISCUSSION
Calibration
MFC vs Core Microcosms: The core
microcosms developed into fully
functioning, .biologically diverse
systems. They showed several differ-
ences from typical MFC microcosms.
For comparative purposes we contrast-
ed events in the developing cores to
events in a typical MFC microcosm
experiment. The MFC nutrient data are
from experiments run at ERL-D by Frank
Stay.
Communities were slower to
develop for cores than in MFC micro-
cosms, probably due to the lower initial
nutrient levels. The MFC protocol used
a nutrient medium to "jump-start" the
systems, while in the cores, nutrients
were initially tied up in organic matter
178
-------
and gradually released by decom-
position processes. Nitrates were
reduced to near zero in MFC systems
within four weeks, but phosphates
remained high. Primary production was
initially very high in MFC systems and
decreased over time, while primary
production in the cores continued to
increase over time. In response to the
initial algal bloom in MFC systems a
largezooplanktoncommunity developed
but later crashed. The zooplankton
community developed much more
slowly in core microcosms. As
organisms died off there was an
increase in ammonia in MFC systems
that did not occur in the cores. Finally,
the pH of the two systems was quite
different, ranging between 8 and 9 in
MFC systems and between 5 and 6 in
core systems.
Laboratory vs. Field: Temporary
pools are biologically and physically
more variable habitats than laboratory
microcosms. Species composition,
seasonal changes and recurring dry
periods are some obvious differences.
The water levels in these pools are
often low and are sensitive to daily and
diel temperature changes. At all sites,
daily low water temperatures were
within a few degrees of 10°C while
high temperatures ranged from 15°C to
32°C depending on the air temperature.
This is in contrast to laboratory
microcosms where water level is
carefully controlled and water
temperature is maintained at 20°C.
Despite these fluctuations, pH in the
pools showed little day-to-day change.
Since the diel change was so small we
measured only morning pH in 1990,
and found a mean value of 6.4. This
compares with mean pH values of 8.75
and 9.6 for morning and afternoon,
respectively, in MFC microcosms. Core
microcosms behave much more like the
ponds and showed little difference
between morning and evening.
Core microcosms were much closer
surrogates of natural ponds than MFC
systems. Nutrient levels, primary
production, and community respiration,
and the diversity of invertebrate biota in
core systems were close to those seen
in natural ponds. Although the
temporary pond fauna were far more
diverse than those of the microcosm
communities, the major taxa were
similar in both. The zoopiankton
community of the microcosms and field
sites contained cladocerans, cyclopoid
and harpacticoid copepods, and
ostracods. Both microcosms and field
sites also supported several species of
coleopterans, chironomids, and aquatic
oligochates, as well as snails (Lymnaea
sp.) and clams (Sphaeriidae}. Some
predaceous insect taxa (e.g., Dytiscid
beetle larvae and Chaoborus larvae)
were deliberately removed from
microcosms to prevent excessive
predation on the zooplankton
community. Other insect taxa (e.g.,
dragon fly nymphs), Eubranchipoda
(conchostracans) and tadpoles (Hyla
crucifer) were present in ponds but not
in microcosms.
Validation
Bti Survival: Btispores survived for
extended periods in both laboratory
microcosms and field sites. While most
spores disappeared from the water
column within a few days, levels in the
sediments remained stable for most of
each test.
179
-------
Although spores persisted, toxicity
rapidly disappeared in both lab and
field. In the field, sediment samples
lost toxicity to mosquito larvae between
9 and 12 days after application.
Similarly samples from both micro-
cosms lost toxicity between 7 and 12
days after application. This loss of
larvacidal activity is due to the
inactivation or microbial degradation of
the parasporal protein crystal
associated with Bti spores. Despite the
loss of toxicity, spores from
microcosms and field sites could
produce new toxic crystals when
cultured on TSA agar.
The patterns of Bti uptake by
invertebrates were similar in lab and
field. The amount of Bti accumulated
varied among the organisms tested.
Filter feeding planktonic organisms
accumulated the smallest amounts
while the bottom-feeding snails
accumulated very large amounts. While
there were some differences in rates of
uptake and losses between lab and
field, the most important finding is that
we could confirm in both lab and field
that the organisms had been exposed to
the test agent.
Ecological Effects: The effects on
the target organism (mosquitoes) were
obvious in both lab and field tests. In
each case they were killed within 24
hours after application. The effects on
non-target organisms and functional
variables were not so obvious. In fact,
our first task in comparing effects on
microcosm and field communities has
been to develop appropriate statistical
procedures for analysis of field data.
We detected no significant
ecological effects in either lab or field.
We did become more aware of the
difficulty of detecting effects .in the
field. In the laboratory we are can
control environmental conditions and
manipulate microcosms to reduce
variability. Since this is not possible in
the field, replicate field sites show
much greater variability than replicate
microcosms. This makes it much more
difficult to detect effects in the field.
We are currently investigating several
procedures for analyzing such data.
In an attempt to deal with this high
variability we investigated the use of an
analysis of covariance (ANCOVA). In
this procedure we adjusted experi-
mental means for the influence of the
covariate factors. In this case, depth of
the pond and water temperature were
the most significant covariates. We
conducted ANCOVAs for each day of
the exposure, and found no consistent
effects that could be attributed to Bti
treatment.
We have not yet completed all
summation and analysis of data
collected during the 1990 pond
exposures. All but one set of ponds
dried up by mid-July. With heavy rains
in August and September all the ponds
re-filled, however, and we were able to
collect additional samples in October.
We are still analyzing these and a few
earlier samples so our conclusions
remain incomplete at this time. At this
point we do not see any significant
effects caused by Bti in the field.
SUMMARY
Microcosm tests provide fate,
direct effects and ecological data that
can be used as a surrogate for outdoor
testing. In this report we have
discussed the strength of the link
180
-------
between microcosms and field
performance. On the basis of current
data we can draw several conclusions:
1. Both the MFC and the core
microcosm model events were
observed in natural systems. The core
microcosm is more closely correlated
with the field than the MFC.
2. We were able to monitor the fate
and effects of Bti in both lab and field
and recorded similar results:
a. Bti spores persisted for the
duration of the test.
b. Bti toxicity persisted for 7 to 12
days.
c. Non-target animals accumulated
Bti. The extent of the accumulation
and rate of loss of Bti were species-
specific.
d. Target animals were killed in both
systems.
e. Non-target animals were not
sensitive in either system.
Tests in a single laboratory or field
ecosystem will not provide definitive
answers to the questions of survival
and ecological effects of a particular
introduced microorganism. Survival
and effects are determined by a
complex series of abiotic (e.g., pH,
temperature, redox potential, structure
and composition of the substrate) and
biotic factors (e.g., the presence or
absence of competitors, predators, prey
or host species). Clearly, tests in one
type of ecosystem cannot provide
results generalizable to all ecosystems.
This presents a dilemma, since it is
obviously not possible to run all the
tests needed to evaluate a micro-
organism under every combination of
abiotic and biotic factors.
The core procedure may provide an
alternative to running multiple tests.
The system is flexible enough to use
biota and sediment from any local site.
Survival and effects questions could
then be focused on a particular
ecosystem of concern, rather than on a
generalized generic aquatic system.
The questions of field calibration and
validation will be further clarified as we
complete additional tests.
PUBLICATIONS
Shannon, L.J., T.E. Flum, R.L.
Anderson, and J.D. Yount, 1989.
Adaptation of the mixed flask culture
microcosm for testing the survival and
effects of introduced microorganisms.
pp. 224 - 242. in: U.M. Cowgill and
L.R. Williams, (eds.), Aquatic
Toxicology and Hazard Assessment:
12th Volume. ASTM STP 1027. Amer.
Soc. for Testing and Materials,
Philadelphia, PA.
Shannon, L.J. and R.L. Anderson. Use
of the mixed flask culture microcosm
protocol to estimate the survival and
effects of microorganisms added to
freshwater ecosystems, in: M. Levin
(ed.), Methods in Microbial Ecology.
Am. Soc. of Microbiology (in press)
Stay, F.S., T.E. Flum, L.J. Shannon,
and J.D. Yount. 1989. An Assessment
of the precision and accuracy of SAM
and MFC microcosms exposed to
toxicants, in: U.M. Cowgill and L.R.
Williams, (eds.), Aquatic Toxicology
and Hazard Assessment: 12th Volume,
ASTM STP 1027. Amer. Soc. for
Testing and Materials, Philadelphia, PA.
181
-------
Shannon, L.J., T.E. Flum, and J.D.
Yount. 1987a Draft protocol for a
mixed flask culture aquatic microcosm
toxicity test. Output No. 7667A 36 pp.
Shannon, L.J., M.C. Harrass, J.D.
Yount, C.T. Walbridge. 1986. A
comparison of mixed flask culture and
standardized laboratory model
ecosystems for toxicity testing, pp
135-157. in: John Cairns, Jr. (ed.),
Community Toxicity Testing, ASTM
STP 920. American Society for Testing
and Materials, Philadelphia, PA.
182
-------
CURRENT INVESTIGATIONS ON MICROSPORIDIAN TEST SYSTEMS
W.S. Fisher, J.W. Fournie, C.L. McKenney, Jr. and D.P. Middaugh
U.S. Environmental Protection Agency
Environmental Research Laboratory
Gulf Breeze, Florida
INTRODUCTION
Since 1982, Gulf Breeze Environ-
mental Research Laboratory (GB/ERL)
has supported a request by the Office
of Pesticide Programs to develop
testing protocols for fate and effects
testing of microbial pest control agents
(MPCAs). This has included description
of single and multi-species test systems
and toxicity and pathogenicity testing
protocols for freshwater and marine
nontarget organisms.
It was recognized that MPCA test
systems must be capable of containing
the MPCA and providing reproducible
data, must be reasonable in cost and
size, must maintain nontarget test
organisms from a variety of phyletic
groups, and should represent as closely
as possible the conditions of a natural
system. Additionally, it should have
the capacity for testing all microbial
groups of MPCAs, including viruses,
bacteria, protozoa and fungi. Multi-
species test systems have the
advantage of testing several species
with one exposure and may demon-
strate interactions among nontargettest
species.
Single and multi- species test
systems employed at GB/ERL quickly
evolved into the present totally-
enclosed aquarium that incorporates an
undergravel filter system using artificial
substratum and an external ultraviolet
light. Several different marine and
fresh water nontarget organisms, includ-
ing fishes, grass shrimp, bivalve
molluscs and aquatic plants, have been
held in these systems and challenged
with a variety of MPCAs. Represent-
atives of all four MPCA groups were
initially tested, including a baculovirus
Autographa californica, a bacterium
Bacillus thuringiensis, a protozoan
Nosema cuneatum and a fungus
Lagenidium giganteum. Positive
controls demonstrated characteristic
infections and pathogenicity in target
species; however, no signs of infection
or pathogenicity were found in non-
target species using histology,
serology, and electron microscopy.
More recently, GB/ERL research has
focused on microsporidian MPCAs.
Microsporidians are small unicellular
organisms, all of which are .obligate
intracellular parasites with a unique
mode of infecting host cells. Currently,
there is one registered microsporidian,
Nosema locustae, developed for control
of rangeland grasshoppers and other
requests have been received by the
Office of Pesticides Programs. Future
registration requestsfor microsporidians
will probably increase as their pesticidal
capabilities become better known.
There is a relative abundance of new
information on the biology of micro-
sporidians: Many are now known to
have indirect life cycles, multiple
sporulation sequences, vertical as well
as horizontal transmission and a
183
-------
broader host specificity than previously
believed.
One of the first microsporidians
tested at Gulf Breeze was the mosquito
pathogen Nosema algerae, which has a
relatively simple direct life cycle.
Development of N. algerae as an MPCA
was originated in the 1970's when
limited field tests were performed for
it's use as a mosquito control agent.
Currently, Edhazardia aedis is being
investigated in collaboration with the
USDA's Insects Affecting' Man and
Animals Research Laboratory for
possible use in the control of container-
breeding mosquitos. This micro-
sporidian is an interesting model
because it has a more complex life
cycle than N. algerae, exhibiting
multiple sporulation sequences and the
capability of both vertical and horizontal
transmission. Nontarget testing for this
MPCA has been expanded to include
fish embryo/larval assays and mysid
shrimp life-cycle assays.
METHODS
Nosema algerae
A series of single species assays
were conducted on freshwater and
estuarine grass shrimp (Palaemonetes
kodiakensis and P. pugio), marine
rotifers (Brachionus plicatilis), and
inland silversides (Menidia beryllina)
employing the aquarium system
described above. Exposure was water-
borne except for grass shrimp, which
were also exposed by intrahemocoelic
injection or gavage.
Edhazardia aedis
Single species tests with grass
shrimp, P. pugio, and mosquito fish,
Gambusia affinis, were conducted in
the aquarium systems described above.
Grass shrimp were exposed via water
or gavage. Gambusia affinis were fed
infected or uninfected mosquito larvae
once a day for 7 days then uninfected
artificial food for 7 days. Samples were
taken eight hours after the first feeding
and then at the end of the experiment
to examine gill, liver, stomach and
reproductive tissues for presence of E.
aedis.
Exposure systems were developed
to test estuarine mysids Mysidopsis
bahia to E. aedis through an entire life
cycle. Newly released young mysids
were exposed to E. aedis spores (1000
mL"1} at optimal (20 ppt) salinity for 1
week and monitored for survival. Adult
mysids were exposed for 4 weeks
under low-salinity (7 ppt) test
conditions and monitored for survival
and reproductive effort.
Fish embryo/larval toxicity assays
were adapted to MPCA testing for
water-borne exposures. Embryonic
inland silversides> Menidia beryllina,
were exposed to E. aedis spores in
moderately-hard freshwater (hardness
100mg/L as CaCoS). Single blastula-
stage fish embryos were placed in each
of 30 individual tissue culture tubes
with spore concentrations of 0, 10,
100 and 1000 mL"1. Tubes were
sealed with teflon lined caps and
embryos incubated at 25°C with a
14:10 photoperiod until death or
hatching occurred.
184
-------
RESULTS AND DISCUSSION
Nosema algerae
Infections did not develop in
gavaged grass shrimp, but did develop
in those that received intrahemocoelic
injections. Infected tissues included the
gills, antennal gland, eyes, skeletal
muscle, heart, and gonads. Proof of
infection was demonstrated ultra-
structurally by the presence of mature
spores and developmental stages in
infected tissues. Infections did not
occur in the marine rotifer after
ingestion of spores or in inland
silversides fed marine rotifers
containing ingested spores. Due to the
extreme exposure measures, infection
of nontarget organisms in nature is not
indicated by these results; however
they do demonstrate that infections can
be identified and characterized using
the diagnostic techniques available.
Edhazardia aedis
Exposure of E. aedis spores to grass
shrimp did not cause any mortalities nor
were any microsporidian life stages
detected in sampled tissues. Eight
hours after feeding of infected larvae,
E. aedis was found in the guts of fish,
G. affinis, demonstrating ingestion with
the infected mosquito larvae. However,
none were detected at the end of the
experiment, one week after fish were
removed from the mosquito larvae diet.
This indicated that colonization and/or
infection of mosquito fish did not occur.
Initial exposures of newly released
young mysids to E. aedis spores
indicated no mortality (> 90% survival)
after 1 week under optimal salinity and
temperature conditions. Exposures of
adult mysids for four weeks under
conditions of low salinity stress (7 pptj
showed reduced mysid survival (from
80% to 54%) and inhibited reproduc-
tion (the total young produced by three
females was reduced from 9 to 3).
Based on these results, a series of
studies have been initiated to examine
the interaction of stress conditions (10
p.pt salinity and temperatures of 20, 25,
and 30°C) survival and reproductive
capacity of M: bahia exposed through
an entire life cycle.
Results from a. preliminary test with
M. beryllina indicated that E. aedis
spores cause an increase in embryo
mortality (3% control vs. 27%
exposed) and deformed larvae at
hatching. This effect did not appear to
be concentration dependent; rather, all
exposure levels, of E. aedis showed
approximately, the same result.
Preliminary observations of exposed
embryos indicated that the chorion of
developing embryos may serve as an
attachment surface for E. aedis spores.
The mode of pathogenicity/toxicity
remains unknown.
FUTURE WORK
Research at GB/ERL on micro-
sporidians will continue to expand the
number of MPCAs tested and determine
toxicity, infectivity and pathogenicityon
nontarget aquatic organisms. Emphasis
will be placed on microsporidian species
with different and more complex life
cycles, such as Amblyospora and
Parathelohania, which have
intermediate hosts in their life cycles.
The results obtained with E. aedis
on fish larvae will be pursued and
185
-------
mechanisms of toxicity or pathogenicity
will be elucidated. The adaptation of
this fish embryo/larval toxicity assay to
MPCA testing appears extremely
productive and other MPCAs will be
examined using this procedure.
Development of the mysid test
system is highly appropriate for MPCAs
developed for mosquito control.
Mysidopsis bahia is an estuarine
arthropod that is closely related to
mosquitos and can be reared through
it's entire life cycle in the laboratory. It
is a representative of the dominant
aquatic, non-insect arthropods (the
Crustacea), and plays a crucial role in
the trophodynamics of the highly
productive estuarine ecosystem.
Mysids are the first nontarget test
organism to be used at GB/ERL to
explore the effects of environmental
stress on pathogenicity and toxicity of
MPCAs. Eventually, exposure condi-
tions will be altered to determine MPCA
effects under stressful environments for
several nontarget species. Stressors
will include both natural environmental
factors (salinity and temperature) and
chemical pollutants. It is expected that
these types of stress are prevalent in
nature and more accurately reflect
conditions of susceptibility.
PUBLICATIONS
Couch, J. A., T. W. Duke, S. S. Foss,
and K. T. Perez 1986. Enclosed
systems for testing microbial pest
control agents. Proc. workshop at
ERL/Gulf Breeze, sponsored by U. S.
EPA, Office of Pesticides Programs,
Washington, D. C.
Couch, J. A., S. S. Foss, and L. A.
Courtney 1985. Evaluation for risks of
an insect virus, bacterium, and
protozoan .to a nontarget, estuarine
crustacean. Report EPA 600/X-85/290,
U. S. EPA, Gulf Breeze FL.
Couch, J. A., S. M. Martin, G.
Tompkins, and J. Kinney 1984. A
simple system for the preliminary
evaluation of infectivity and
pathogenesis of insect virus in a
nontarget estuarine shrimp. J,
Invertebr. Pathol. 43:351-357.
Couch, J. A. and K. R. Rao 1983.
Biorational Workshop. Report EPA-
600/X-83-054, Gulf Breeze ERL, Gulf
Breeze FL.
Foss, S. S., L. A. Courtney, J. W.
Fournie and D. V. Lightner 1989.
Nontarget testing of microbial pest
control agents in aquatic systems.
Report EPA/GOO/x-89/387, 21pp.
Fournie, J. W., S. S. Foss and J. A.
Couch 1988. A multispecies system for
evaluation of infectivity and
pathogenicity of microbial pest control
agents in nontarget aquatic species.
Dis. Aquat. Org. 5:63-70.
Fournie, J. W., S. S. Foss, L. A.
Courtney and A. H. Undeen 1990.
Testing of insect microsporidians
(Microspora: Nosematidae) in nontarget
aquatic species. Dis. Aquat. Org,
8:137-144.
186
-------
DEVELOPMENT OF TEST METHODS TO ASSESS FATE OF MICROBIAL PEST
CONTROL AGENTS AND THEIR EFFECTS ON
NONTARGET AQUATIC ORGANISMS
Fred J. Genthner, Geraldine M. Gripe and Doug P. Middaugh
U.S. Environmental Protection Agency
Environmental Research Laboratory
Sabine Island
Gulf Breeze, Florida
INTRODUCTION
Microbial pest control agents
(MPCAs) are regulated by the Environ-
mental Protection Agency. One respon-
sibility of EPA is to prescribe tests for
determining the potential ecological
risks of MPCAs to nontarget organisms.
To fulfill the Agency's need for
ecosystem effects testing of MPCAs, a
fully contained, single species test
system was developed to determine
whether exposure of a nontarget
aquatic invertebrate to an MPCA will
result in infectivity, toxicity or
pathogenicity. This system was field
calibrated to assess whether MPCAs
will colonize, persist or germinate in the
animal. MPCAs representing a
vegetative bacterial cell, Pseudomonas
fluorescens; and both a fungal,
Colletotrichum gloeosporioides (Cga),
and a bacterial spore, Bacillus
sphaericus, were used for field
calibration.
In addition, we have successfully
adapted standard chemical toxicity test
methods for assessing nontarget
effects by exposing the estuarine
crustacean, Mysidopsis bahia, and
embryos of the inland silverside fish,
Menidia beryllina to Beauveria bass/ana,
an entomopathogenic fungus.
METHODS
Single Species Test System
American oysters, Crassostrea
virginica, were exposed in enclosed 57
liter aquaria filled with filtered (2 /vm)
seawater from Santa Rosa Sound.
Oysters were fed with a mixture of 3
marine algae 3 times a week.
To field calibrate the elimination
rates of MPCAs, exposed oysters were
thoroughly cleaned and divided into 2
groups. One group was placed in wire
cages suspended in 6-8' of water (6
inches above the bottom) in Santa Rosa
Sound. The remaining oysters were
placed in an enclosed 57 liter aquarium
equipped with aeration, and a pump
driven, recirculating system (1 £/min)
which allowed water to pass first
through an external cartridge filter of
crushed coral then through a UV
sterilization unit. Both the exposure
and UV-clearance tanks were sealed
with a plexiglas lid. Sampling ports and
inlets for air and water were sealed
with silicon stoppers.
Viable counts of the MPCAs were
performed on aquaria water and oyster
samples at regular intervals. Sections
of whole oysters were taken for histo-
logical examination. Tissue samples
187
-------
were fixed in Bouin's solution,
dehydrated, embedded in paraffin, and
sectioned at 6 //m. Sections were
stained with Harris' hematoxylin and
eosin, Brown and Brenn tissue Gram
stain or periodic acid shift.
Nontarget Effects
ASTM recommendations for acute
exposures were used in toxicity tests
exposing M. bahia with viable spores of
B. bassiana. Spores were obtained
from cadavers of infected Heliothis zea
to assure pathogenicity. Inoculum was
prepared by scraping conidiospores
from the surface of 11 day old plates.
Spores were either treated with Triton
X-100 (0.03%) or left untreated before
exposure of the animals. Animals (10
per bowl, 2 bowls per spore density)
were exposed for 96 hours at spore
densities of 2 X 106, 1 X 106, 1 X 105,
1 X 104, 1 X 103 per ml. Autoclaved
controls treatments were also
performed. Beauvericin (Sigma), an
secondary metabolite produced by this
fungus, was also tested with M. bahia.
Single blastula stage embryos of the
inland silverside, M. beryllina, were
placed in each of 120 Leighton tissue
culture tubes and exposed to B.
bassiana at densities of .1 X 106, 1 X
105 and 1 X 104 conidiospores per ml.
Embryos were incubated at 25 C under
a 14L:10D photoperiod. Observations
for effects were made daily.
RESULTS AND DISCUSSION
Single Species Test System
Neither toxicity, pathogenicity, or
infectivity was observed in oysters
exposed to B. sphaericus, Cga or P.
fluorescens. Germination of B.
sphaericus of Cga spores was not
observed in the oyster tissue. The
rates of clearance of these MPCAs from
the oyster were compared in exposed
oysters relayed in the UV-tank or in
Santa Rosa Sound. With all MPCAs
tested the rates of clearance from
oysters in the sound versus the UV-
clearance tank were similar.
The rates of clearance, however,
between MPCAs varied widely.
Oysters were exposed to either B.
sphaericus and P. fluorescens for 2
weeks. B. sphaericus spores were
undetectable in oyster tissue at 5
weeks. In contrast, P. fluorescens was
still detectable in oyster tissue at 49
days. Oysters, exposed to Cga spores
for 3 days, contained 1.7 X 105 spores
per gram dry wt. These spores were
rapidly cleared from the oyster and
were undetectable in oyster tissue at 9
days of clearance.
Exposure plates, placed in various
locations during the laboratory tests,
indicated a completely sealed system.
The clearance rates of MPCAs from
oysters in the fully contained system
were similar to the ''clearance rates in'
the field. Thus, this system will be of
value before an environmental release
has been deemed safe in determining
whether, or how long, genetically
engineered MPCAs or other micro-
organisms will persist in nontarget
aquatic invertebrates and whether,
exposure will result in toxicity,
infectivity or pathogenicity.
Nontarget Effects '.
In an initial toxicity test with M:
188
-------
bahia, B. bassfana spores were treated
with Triton X-100 to aid in dispersal.
This test demonstrated no toxicity of
the spores up to and including 1 X 106
per ml test water. However, a sub-
sequent exposure of test animal to
these spores without the wetting agent
indicated that with this particular strain,
mortality was associated with high
particulate density of the spores (as low
as 1 X 106 per ml) an not with spore
viability. Testing of beauvericin, an
antibiotic produced by some other
strains of this fungus, has indicated
some toxicity.
While responses of the inland
silverside embryos to B. bassiana has
been variable among the three tests
performed to date, in all instances
pathogenicity has been observed in the
form of rupture of the chorion of
developing embryos followed by death,
FUTURE WORK
The fully contained exposure and
clearance system will continue to be
field calibrated using a viral and
protozoan MPCA. Testing B. bassiana,
its secondary metabolites, and other
MPCAs will proceed using M. bahia.
Additional exposures of inland silverside
embryos are planned to determine the
reason(s) for variability of responses
among tests.
PUBLICATIONS
Yousten, A.A., E.F. Benfield, R.P.
Campbell, S.S. Foss, and F.J.
Genthner. 1991. Fate of Bacillus
sphaericus 2362 spores following
ingestion by nontarget invertebrates. J.
Invert. Pathol. (In Press).
Genthner, F.J., S.S. Foss, and R.P.
Campbell. An enclosed system for
testing effects of microbial pest control
agents on nontarget aquatic
invertebrates. (In Preparation).
Genthner, F.J. and D.P. Middaugh.
Effects of the entomopathogenic
fungus, Beauveria bassiana, on inland
silversides, Menidia beryllina, embryos
(In Preparation).
189
-------
FATE OF Bacillus sphaericus MICROBIAL PEST
CONTROL AGENT IN THE ENVIRONMENT
Allan A. Yousten and Ernest F. Benfield
Biology Department
Virginia Polytechnic Institute and State University
Blacksburg, Virginia
INTRODUCTION
Some strains of the bacterium
Bacillus sphaericus have been found to
be pathogenic for mosquito larvae.
This pathogenicity is caused by the
production by the bacteria of protein
toxin which is lethal upon ingestion by
larvae of many species of mosquito.
Unlike the toxin produced by Bacillus
thurfngiensis subsp. israelensis (Bti),
the toxin of B. sphaericus is not active
against black fly larvae. The gene for
the B. sphaericus toxin has been cloned
and sequenced from several strains of
the bacteria and has been found to
differ among strains by only a few
amino acids. However, the strains do
differ somewhat in their toxicity for
different species of mosquitoes and
these differences probably lie in the
slight differences in " amino acid
sequence in the toxins. The toxin is
synthesized in the bacterial cell at the
time of sporulation and it accumulates
in the cell as a parasporal inclusion
body or "crystal". The inclusion body
contains primarily two proteins, and
both of these are required for toxicity to
larvae. Thus, the toxin of B. sphaericus
is a "binary toxin".
In its role as a mosquito larvicide,
viable B. sphaericus spores along with
the attached parasporal bodies will be
delivered into the aquatic environment.
Since the metabolism of B. sphaericus
is different than that of Bti, the
ecological niche in which it may survive
or grow in the aquatic environment may
also differ. One salient feature that has
attracted attention is the apparent
ability of B. sphaericus to persist in the
mosquito larval feeding zone for longer
than Bti. This effect has not been
observed in all field trials, but it has
been reported often enough to make
this bacterium an attractive larvicide in
situations where frequent application is
not practical. The reasons for this
larvicidal persistence are unclear.
Our studies are examining fate of
the bacterial spores during their
interaction with the aquatic environ-
ment as well as the fate of spores
ingested by certain non-target
invertebrates.
METHODS
Bacterial Growth
B. sphaericus 2362-7, a spon-
taneous rifampicin-resistant mutant of
B. sphaericus was used in all experi-
ments. Spores were produced by
growth in nutrient broth-0.05% yeast
extract-mineral salts (Mn2 + , Mg +,
Ca2 + ) broth (NYSM) with shaking at
30°C for 48 hours. Spores were
washed and held as a distilled water
suspension at 4°C. Spore numbers
were determined by heating 2 ml of a
190
-------
sample at 80°C for 12 min. and plating
on NYSM agar containing 50 yt/g/ml
rifampicin and 0.002% cycloheximide.
Midge Larvae
Seventy five, fourth instar
Chironomus riparius larvae were placed
into 200 ml sterile distilled water in a
23 cm. X 23 cm. dish at 23°C. Three-
tenths ml Tetramin suspension was
added along with B. sphaericus 2362-7
spores at a final concentration of 9.6 X
I06/ml. After 5 hours of feeding, the
larvae were rinsed 3 times in 100 ml
sterile distilled water. Five larvae were
weighed, homogenized in a ground
glass tissue grinder, and a spore count
performed as described. Forty five
spore-fed larvae were placed into 200
ml water containing Tetramin but
without spores. Groups of five were
homogenized at intervals to obtain
spore counts.
To collect larval feces, twenty five
spore-fed larvae were placed into a
funnel blocked at the narrow end with
screen. The funnel outlet led into a
latex rubber tube which dipped into an
ice water bath. A small amount of
Tetramin was added to the sterile
distilled water in the funnel. After 17
hours, the 50-ml of water in the funnel
and tube were removed and the funnel
and tube were rinsed with 50 ml water.
The combined volumes were centri-
fuged (27,200 x g, 15 min), the
'resulting pellet containing 2362-7
spores suspended in sterile distilled
water, and a spore count performed.
Thirty-one snails, Helisoma trivo/vis,
were placed into 200 ml sterile distilled
water in a 23 cm. X 23 cm. dish at
23°C. B. sphaericus 2362-7 spores
were added to a final concentration of
9.6 X I06/ml. After grazing for 5 hours,
the snails were rinsed 3 times with 50
ml sterile water and placed into 300 ml
sterile water containing a piece of
lettuce. Two snails were immediately
removed, the bodies separated from the
shell, the foot dissected away, and the
remaining tissue homogenized. A spore
count was performed on the
homogenate. The remaining snails
were placed in fresh sterile water daily
for the first three days after feeding on
spores and twice weekly thereafter.
Lettuce was provided as food
throughout the experimental period.
Three of the spore-fed snails were
placed in a funnel apparatus described
above for fecal collection. These snails
were also moved to fresh water daily
for three days and twice weekly
thereafter to prevent reingestion of
feces that might not pass through the
screen at the bottom of the funnel.
Feces were collected for 17 hours, the
collection water centrifuged (27,200 x
g, 15 min.), and the resulting pellet
suspended in sterile distilled water for
spore counts and bioassay.
Oysters
Four, 112-liter glass aquaria were
used for spore feeding tests; two for
inoculation and two for controls.
Aquaria were equipped with under-
gravel filters covered with crushed coral
and filled with artificial seawater at a
salinity of 10 parts per thousand. Sixty
cleaned oysters were placed into each
test aquarium. B. sphaericus spores
191
-------
were added and the aquaria sealed.
Oysters were fed with a mixture of 3
marine algae every third day. To
determine the spore burden in oysters,
samples of digestive, gill and remaining
whole oystertissues were homogenized
as previously described.
To determine spore elimination
rates, 20 oysters were removed from
each inoculated aquarium after 28
days, cleaned, and placed into a
flow-through seawater tank. Four
oysters were removed every week for 5
weeks and a spore count was
performed on the whole oyster tissue.
Fecal material was obtained directly
from the shell surface. When
recognizable, fecal material was
collected from the tank bottom using a
pipette. Fecal material was placed into
a test tube, and the excess liquid was
removed from the surface. The tube
was centrifuged (7,700 X g, 10 min.),
and the supernatant discarded.
Toxicity of Snail and Oyster Feces
Feces from snails and oysters were
collected as described above. Fecal
suspensions were bioassayed for
toxicity to second instar larvae of Culex
quinquefasciatus.
Stonefly Larvae
Paragnetina media (Walker) larvae
were placed individually into chambers
containing 400ml tap water at 20°C.
Midge larvae were fed on/ B. sphaericus
spores overnight, rinsed 3X, and 3
larvae placed into each stonefly
chamber. Only stoneflies that had
eaten 3 midges within 4 hours were
used in the depuration experiment.
Immediately after feeding on midges
and at intervals thereafter, the stonefly
larvae were rinsed, appendages
dissected away, bodies weighed, and
homogenized. Homogenates were
plated for spore counts. Beginning 24
h after having ingested spore-laden
midges, the water in each stonefly
chamber was changed and this was
repeated on alternate days. On the
days on which the water was not
changed, each stonefly larva was fed 1
midge larva (not containing spores).
The water changes alternating with
midge feeding were continued
throughout the experiment.
Crane Fly Larvae
Tipula abdominalis larvae held
individually in 400 ml chambers in the
dark at I3°C were each fed 1 yellow
poplar conditioned leaf disc that had
been soaked in B. sphaericus spore
suspension. Each disc had approxi-
mately 5xl06 spores adhering to it.
Twenty four hours after consuming the
spore-laden disc, the water in each
larval chamber was changed and a
fresh leaf disc lacking spores was fed.
The water was changed and a leaf disc
fed at 48 h intervals thereafter for the
duration of the experiment. At intervals
groups of 3 larvae were removed, the
guts dissected out, weighed, homogen-
ized, and spore counts performed on
the homogenates.
RESULTS AND DISCUSSION
Indigenous Microflora
None of the animals or their feces
used in this study contained microflora
192
-------
which survived heating at 80°C and
grew on NYSMRC selective medium in
numbers which interfered with
enumeration of B. sphaericus,
Midge Larvae
Elimination (depuration) of spores
from fourth instar larvae was followed
for 4 days following ingestion of
spores. Larvae examined immediately
after feeding averaged 1.2 X 10°
spores/mg larval dry weight. This
number declined to 7.7 X I02
spores/mg after 4 days at which time
the larvae began to pupate. The total
viable bacterial count on the fourth day
was approximately equal to the spore
count. The spore content of feces
declined in parallel to body spore
content.
Larvae that were reared, allowed to
pupate, and emerge as adults in water
initially containing I X I07 or 4.3 X I03
spores/ml, had an average of 8.2 X I03
and 68 spores in/on the bodies (15 and
18 flies, respectively). At the
conclusion of the experiments, the
culture water contained 9.5 X I06 and
8.4 X I02 spores/ml. Control flies
reared in water lacking added spores
averaged 4 spores/fly (14 animals) and
no spores/fly (10 animals), respectively.
Snails
Depuration of spores from snails
was followed for 49 days. There was
a rapid decline (about 2 logs) in the
body burden of spores during the first 7
days followed by. an approximately
constant number of spores (between 2
X I01 and 2 X I02/mg) for an extended
period. At 49 days the body spore
content declined to less than IOO
spores/mg for the first time. As with
midge larvae, the total viable bacterial
count approximated the spore count.
The spore content of feces declined
even more rapidly than that of the
body, but it also leveled out/and feces
containing low numbers of spores
continued to be produced for the
remainder of the experiment.
To eliminate the possibility that the
persistence of spores in snails was a
result of reingestion of feces containing
spores, snails were allowed to feed in
9.6 X I06 spores/ml for 23 hours which
resulted in a body content of 4.9 X I06
spores/snail. These snails were placed
in a sterile fecal collection funnel which
was drained and rinsed with sterile
water twice daily before being refilled
with sterile water. After 7 and 14 days,
the snails were transferred to another
sterile funnel. After 21 days, the feces
from the snails were collected and the
body burden of spores determined. The
snails averaged 1.7 X I04 spores/snail
and the feces contained 2.3 X I03
spores/snail.
The LC50 of the spore/toxin
preparation fed to snails was 1.2 X I02
spores/ml. This declined 66% to 3.6 X
I02 spores/ml in feces.
Oysters
Oysters placed in spore-containing
water and then transferred to flowing
seawater, decreased in spore Content
from 2.5 X I06 spores/g dry tissue to
5.0 X I03 in 21 days and had eliminated
the spores in 37 days. Histological
examination showed no indication of
spore germination or of vegetative
bacilli in the gut. Oyster feces were
193
-------
non-toxic when fed to mosquito larvae,
apparently having been detoxified by
passage through the oyster.
Stonefly Larvae
Ingestion and clearance of spores in
a predator-prey relationship was
examined with stonefly larvae
(predator) and spore-carrying midge
larvae (prey). The midges were rapidly
consumed by the stoneflies and B.
sphaericus spores were found to be
present in the stoneflies at levels
approximately equal to the sum of the
spores present in the midges. The
spore number declined rapidly and
spores were below the level of
detection by 7 days.
Crane Fly Larvae
In contrast to rapid depuration seen
in midge and stonefly larvae, crane fly
larvae retained a large number of spores
for an extended period of time. The
animals had a mean number of 4.5 X
I05 spores/gut immediately after eating
a leaf disc. This number decreased to
8.5 X I04 after 4 weeks. This was the
highest level of spore retention
observed among any of the animals
studied. Spores were present in the
feces for the entire period, but the
content dropped more rapidly than did
the spore content of the animal gut.
Total viable plate counts done at 1 day
and 2 weeks after feeding on spores
approximately equaled the spore count.
This suggests that at least on those
times there was not a large population
of vegetative cells growing in the
animal. Such a population could have
continuously produced spores to
replace those .lost in the feces. We
cannot exclude the presence of a small
vegetative population approximately
equal to the spore population that
would not be detected given the
sensitivity of plate counts. The crane
fly has an unusually alkaline foregut and
a possible relationship between this
phenomenon and spore retention or
production is being investigated. Also,
more a exact location of the spores in
this anjmal is being determined.
Summary
Spores of the MPCA, B. sphaericus,
were consumed but largely or complete-
ly eliminated from the gut of 2 insects
(midge,, stonefly) and 1 mollusk
(oyster). A second mollusk (snail)
eliminated most of the spores but
retained a low number for an extended
period of time. A third insect, crane
fly, retained a significant number of
spores for up to I month. All animals
tested passed viable spores in their
feces. Midges reared in spore-
containing water emerged as adults
carrying spores in/on their bodies. The
mosquitocidal toxin of B. sphaericus
was partially destroyed by passage
through the snail and completely (or
nearly completely) destroyed by
passage through the oyster.
FUTURE WORK
We will continue to examine the
effect of different gut conditions on
survival of B. sphaericus spores. This
will focus on the highly alkaline gut of
the crane fly where extended survival
was observed and move to another leaf
shredder, Pteronarcys proteus, which
194
-------
has a slightly acidic gut. We will also
examine a micro-crustacean, Daphnia,
for the first time. The possible
proliferation of B. sphaericus in
nontarget cadavers produced when the
animals died of other (non B.
sphaericus toxin) causes will be
investigated. These interactions of
spores with nontarget animals will be in
addition to other studies of the
interaction of spores with physical
factors in the aquatic environment.
PUBLICATIONS
Yousten, A.A., E.F. Benfield, R.P.
Campbell, S.S. Foss, and F:J.
Genthner. Fate of Bacillus sphaericus
2362' spores following ingestion by
nontarget invertebrates. J. Invertebr.
Pathol. (in press).
Genthner, F.J., R.P. Campbell, S.S.
Foss, E.F. Benfield, and A.A. Yousten.
1990. Fate of Bacillus sphaericus 2362
spores following ingestion by nontarget
invertebrates. Abst. Annu. Mtg. Am.
Soc. Microbiol. p. 254.
195
-------
PROTOCOL DEVELOPMENT FOR TESTING THE EFFECTS OF
BACTERIAL PESTICIDES ON BENEFICIAL BEETLES
Rosalind James1 and Bruce Lighthart2
ManTech Environmental Technology, Inc.1
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
Microbial pest control agents
(MPCAs) are increasing in use because
of their specificity and low mammalian
toxicity and pathogenicity; however,
they are a potential hazard to non-
target, beneficial insects. MPCAs could
be used more effectively in integrated
pest management if their impact on
non-target insects was better under-
stood. Non-target insects are eco-
logically and economically important in
a variety of ways, including: as natural
predators and parasites of pests, as
introduced biological pest control
agents of both insect and plant pests,
as food for wildlife, and as
decomposers. Subdivision M of the
Pesticide Testing Guidelines calls for
assessing the potential hazard of
MPCAs on several orders of terrestrial
non-target insects including Coleoptera
(beetles). Such tests need to be
standardized because the results of the
tests may be affected by the conditions
of the bioassay. To develop a standard
assay for testing the effects of bacterial
pesticides on beneficial insects, the
effect of temperature and dietary
stress, and of larval instar on the
susceptibility of Hippodamia con vergens
(the convergent lady beetle) to a weak
bacterial pathogen was tested. H.
convergens is a major predator of
aphids and occurs throughout most of
the United States and as such is ah
important biological control agent.
METHODS
H. convergens was reared on pea
aphids, Acyrthosipon pisum, at 25°C,
70% RH. Pseudomonas fluorescens
was used as the pathogen (i.e. the
surrogate MPCA). It is a bacterium that
was isolated from the haemolymph of
dead adult beetles during a time when
there was high morality in the
laboratory colony. The relative
susceptibility of different larval instars
of H. convergens to P. fluorescens was
compared by determining the dose
response of each instar at 25°C on an
aphid diet. First instar larvae were used
to determine the effects of diet and
temperature stress on susceptibility.
Since temperature and the insect's diet
can potentially affect the bacterial
virulence as well as insect suscept-
ibility, the treatments were given to the
insects in 24 hour pulses before the
insects were exposed to the pathogen.
Three different diets (water, 5%
sucrose, and aphids) were used at each
of two temperatures (25° and 30°C),
thus there were six treatments.
The experiment was set up to
determine the probit dose response of
the pathogen on the beetle. Insects
196
-------
were exposed to the pathogen by
dipping them in the appropriate dose of
the bacteria. Five or six doses were
used per treatment with twenty insects
per dose for a total of 100-120 insects
per treatment. There were two
replicates of each treatment. After
insects were dosed, they were
maintained at 25°C, 70% RH, 14:10
hours light:dark. Mortality was
recorded on day six. Both replicates
were combined for the probit analysis
of mortality.
RESULTS AND DISCUSSION
The mean LC50 for the first, second
and third instars was 7.9x109 cfu/ml.
There were no significant differences in
the slopes or intercepts of the probit
lines for these age groups. However,
the last instar was 40 times less
susceptible with an LC50 of 3.2x1011
cfu/ml. The slope of the probit line was
not significantly different from the other
instars but the intercept was (p.
-------
convergens, and to develop protocols
for beneficial Diptera, such as the
Tachinidae. The methods used for
laboratory testing of the protocols will
be similar to that used for this protocol.
However, future work is planned for
laboratory and field studies used to
determine the appropriateness of the
LCg0 endpoint. Some of the questions
we would like to address are: What
does a laboratory LC50 mean when it
comes to field applications? Might
some other endpoint be more approp-
riate? We would also like to monitor
the effects of specific MPCAs on non-
target insects under field conditions.
The results of laboratory and field tests
could then be incorporated into a model
for predictive value.
PUBLICATIONS
James, R.R. and Lighthart, B. 1990.
Bioassay for testing the lethal effects of
bacterial pathogens on the predatory
beetle Hippodamia convergens Gue.
(Coleoptera: Coccinellidae). National
Technical , Information Service
publication no. PB91-127795.
Springfield, VA.
198
-------
ASSESSING HOST SPECIFICITY OF FUNGAL MPCAs
TO THE BENEFICIAL WASP Trichogramma pretiosum
O.K. Sewall1 and B. Lighthart2
ManTech Environmental Technology Inc.1
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
The registration process for
Microbiological Pest Control Agents
(MPCA's) requires data on nontarget
effects to representatives of key
beneficial insect species, e.g.,
Trichogramma pretiosum Riley, a
lepidopteran egg parasite.
Trichogramma spp. are recognized as
important, both as naturally occurring
and commercially released natural
enemies of pest Lepidoptera. The
laboratory evaluation of MPCAs for
nontarget effects is generally
recognized as the first step in
determining the risk a pest control
agent may pose to natural enemies.
Although methods have been developed
to evaluate chemical pesticide effects
on Trichogramma spp., bioassay's
evaluating entomogenous fungal
pathogenicity to Trichogramma spp.
have reported only negative effects.
This study was undertaken to
develop bioassay procedures to
evaluate pathogenicity of entomo-
genous fungi, compare specific and
nonspecific fungi, and evaluate factors
that may affect the expression of fungal
virulence to T. pretiosum or the
susceptibility of T. pretiosum to fungal
pathogens.
METHODS
Insects Used
A colony of T. pretiosum was
maintained on U-V killed Cabbage
Looper eggs, Trichoplusia ni (Hubner),
at 25 ± 1C0, 75% RH and with continu-
ous light (40w fluorescent).
Fungi Tested
The life form of the fungi tested
was conidia. Pure stocks of Beauveria
bass/ana and Metarhizium anisopliae
conidia, without mycelia or agar, were
obtained by inoculating Sabouraud
Maltose Agar + 1 % yeast extract
(SMA +Y) plates, lined with a sterile
disc of 200 mesh nylon screening.
After 21 days incubation, plates were
dried and conidia were scraped from
the screening with a sterile spatula and
stored dry (20-40mg portions) at-70°C
until needed. B. bass/ana was passed
through T. pretiosum by inoculating
ten wasps with enough conidia
(108/ml) to ensure 100% mortality by
72 hours when incubated at 100% RH
and 25°C. As wasps died, they were
moved to a fresh vial, incubated until
covered with conidia, seven days, and
ten new wasps were added to the vial.
199
-------
These steps were repeated five times.
Conidia on cadavers from the last set
were washed with 1.0% T-80, and a
stock of conidial inoculum was
generated as described above. To
ensure maximum pathogenicity of the
host specific fungus Nomureae rileyi,
conidia were collected directly from
infected T. nf cadavers and used
immediately.
Inoculum Preparation
Approximately 20mg of conidia
were suspended in 3 ml of sterile
0.10% (v/v) Tween 80 (T-80) with an
all glass tissue grinder and processed
until clumps of conidia were no longer
visible. A dilution sequence was
prepared with a final T-80 concentra-
tion = 0.01%. Concentrations of
conidia were based on viable conidia
and estimated with drop plate counts.
Exposure to pathogen
Because of the small size and
relative delicateness of Trichogramma
spp. special techniques were developed
to handle and expose wasps to
pathogens which minimized risk of
pathogen escape while allowing
efficient handling of the large numbers
of T. pretiosum needed for dose
response tests. To sort and handle,
wasps were anesthetized on a cold
plate (3-10°C). Medium to large
females (0.5-0.75mm long) were used
for the bioassay. Wasps were
transferred in groups of 75 to
confinement chambers of an immersion
apparatus. Wasps were exposed to
pathogens by filling the immersion
apparatuses with conidial suspensions.
Apparatuses containing wasps were
then dried on a vacuum manifold with
down stream air filtered to remove
entrained conidia. Wasps from each
apparatus were divided into three
groups of 20 wasps each, transferred
to clear one ounce plastic cups supplied
with a streak of 50% fructose syrup,
sealed with snap over plastic lids, and
incubated inverted in humidity
chambers. To evaluate the effect of
relative humidity on the expression of
virulence by B. bass/ana and M.
anisopliae, cups with wasps were either
supplied with four droplets (0.010 ml
ea.) of distilled water and incubated in
a humidity chamber with saturated
KH2PO4 (= rapid saturated air) or
water droplets were omitted and cups
were incubated in chambers with
distilled water (= delayed saturated
air).
Mortality was evaluated on days 3
and 4 after exposure to fungi.
Estimates of the reported dose
response parameters and pairwise
comparisons of dose response slope
functions were made with .probit
software (SAS Version 6.06).
Conidia adhering to wasps were
enumerated by using drop plate counts
on SMA + Y plates. Counts were
prepared using ten wasps from each
dilution in a ten-fold dilution sequence
and homogenized in 1 ml of 0.1 % T-80
with tissue grinders. Duplicate plates
were prepared directly from each
dilution with no further dilution. Plates
were counted at 48 hours for B.
bass/ana and at 72 hours for N. rileyi.
200
-------
RESULTS AND DISCUSSION
The bioassays differentiated
between host specific and nonspecific
pathogenic fungi. Of the three species
of fungi with which T. pretiosum was
challenged, only N. rileyi failed to
produced mortality. B. bassiana levels
of virulence were significantly different
for the isolates tested. Virulence of a
given isolate of B. bassiana maintained
continuously on artificial media was
significantly increased when passed
through T. pretiosum. Estimates of
conidia adhering to wasps after
immersion indicated that the lack of A/.
/•//ey/pathogenicity to T. pretiosum was
not due to rate of conidia attachment to
wasp cuticle but to other factor(s). M.
anisopliae and B. bassiana while both
pathogenic to T. pretiosum, expressed
different levels of virulence to T.
pretiosum when saturation of air in the
test vials was rapid versus delayed.
When saturation of air was delayed the
virulence of these fungi were nearly
identical. When saturation of air was
rapid the virulence (LC50) of both
pathogens increased (= decrease in
LC50) with M. anisopliae having a
significantly lower LC50 ( = more
virulent). However, the dose response
slope function for B. bassiana increased
by 64% while the slope for M,
anisopliae decreased by 74%. T.
pretiosum longevity and susceptibility
to B. bassiana was not affected by
withholding food, withholding water, or
by exposure to temperature extremes
(5° or 30°C for 24 hours). These
results indicate that the nontarget
effects of entomogenous fungi on T.
pretiosum are largely affected by innate
differences of the fungi tested and the
environmental factors, relative humidity
and temperature, which affect the
expression of virulence by the fungus.
For quantitative comparisons of
different species, isolates, or strains of
entomopathogenic fungi the culture
history, original isolate source and
environmental conditions adequate, for
expression of the pathogen's virulence
must be rigorously documented and
controlled. The shifts and differences in
the dose response slope functions
observed when saturation of air was
achieved rapidly suggest that
comparisons between pathogens for
relative virulence and host susceptibility
must go beyond the classic point
comparisons (LC50 fiducial limits) and
include the other dose response
functions e.g., slope. This bioassay will
discern between host specific and
nonspecific pathogenic fungi and
produce quantitative results to compare
nonspecific fungal MPCAs for
pathogenicity to T. pretiosum.
FUTURE WORK
Future research is to be focused on
developing test procedures to evaluate
pathogenicity of MPCAs to adult and
larval honey bees (Apis mellifera L.) and
effects on colony health and product-
ivity. Our goals are to (1) develop a
microbiologically defined test system
using microcosms and mini-hives to
evaluate the effects of MPCA's on
specific life stages, e.g., workers
versus larvae, and the colony and (2)
evaluate the effects of biotic and abiotic
environmental factors on bee suscept-
ibility. Parameters proposed to evaluate
effects include, but are not limited to,
worker longevity, larval survivorship,
201
-------
food consumption and storage, and the
growth rate of the colony. Pending
development of the microcosm, the
system will be validated for detecting
effects using known life stage specific
honey bee pathogens. This will be
followed by evaluations of environ-
mental factors on susceptibility of bee
life stages to pathogens.
PUBLICATION
Sewall, O.K. and B. Lighthart. 1990.
Standard practice for conducting fungal
pathogenicity tests on the Lepidopteran
egg parasite Trichogramma pretiosum
(Hymenoptera: Trichogrammatidae).
NTIS Acces. # PB90-263849/AS. 40
PP-
202
-------
AN OVERVIEW OF PROTOCOL DEVELOPMENT FOR
AVIAN PATHOGENICITY TESTS
A. Fairbrother1 and P. Buchholz2
U.S. Environmental Protection Agency1
Environmental Research Laboratory „
ManTech Environmental Technology, Inc.
Corvallis, OR 97333
2
INTRODUCTION
Microbial pest control agents
(MPCAs) are microorganisms applied to
agricultural and silvicultural
environments to control proliferation
and dissemination of insect or plant
pests. During the past four years, we
have been working to develop a set of
standardized protocols that can be used
to determine the pathogenicity of
microorganisms in nontarget avian
species. The mallard (Anas
platyrhynchos) and the northern
bobwhite (Co/fnus virgin/anus) were
chosen as the test species since both
are required by the U.S. Environmental
Protection Agency for routine wildlife
toxicity testing during the registration
process of chemical pesticides. These
protocols are used primarily by the
Office of Pesticide Programs when
prospective registrants propose new
microbial pest control agents {MPCAs).
The protocols are written such that
could be used to determine patho-
genicity of any microorganism, either
wildtype or genetically engineered;
however, they lack requirements for
determining potential gene exchange
between the introduced microbe and
endogenous gut microflora.
Three routes of exposure of the
birds to MPCAs were evaluated: oral,
intravenous (I.V.), and respiratory.
Three classes of microorganisms, a
virus, fungus, and bacterium, were
used to test protocols developed for
oral and I.V. exposure tests. A
bacterium and a virus were used to
evaluate the respiratory protocol. This
report reviews our development of the
protocols with emphasis on respiratory
protocols developed during the last year
and a half and a study by our col-
laborator, Dr. Matsumoto, at Oregon
State Universtity to develop a sensitive
method for determining if birds are
being exposed to Bacillus
thurfnginensis.
METHODS
Oral and Intravenous Protocols
Ten-day-old chicks and ducklings
were used in the tests because they
lack full immunocompetency until 20-
25 days of age, and in order to
maximize the number of replicates. The
test microbes used were Autographa
californica nuclear polyhedrosis virus as
the viral agent; Metarrhizium anisopliae
as the fungal agent; Salmonella
pullorum as the bacterial agent. For the
oral exposure route, birds were given a
single oral inoculation by insertion of a
gavage needle down their esophagus,
and injecting a known volume and
concentration of the organism. The
203
-------
intravenous route is similar, with the
microbial injection inoculated into the
right jugular vein. The dose
administered to the birds was 104
times the insect LC50 which is 100
times the commercial application rate
and represents a potential worst case
scenario. Forty birds were inoculated
with live test pathogen, forty with
inactivated pathogen, and another forty
with diluent only. Birds were monitored
for 28 days post inoculation with
weekly recording of body weights,
weekly collection of blood samples for
serum chemistry analyses and antibody
response assays, collection of fecal
samples at specified intervals to
determine if organism is shed in the
feces, and histopathological
examination of all tissues showing
gross lesions or that are collected from
birds that die during the experiment.
Salmonella pullorum is known to be
pathogenic in poultry and was chosen
as a positive control. LD50 studies
were conducted for both species.
Respiratory Protocol
In collaboration Dr. Crystal Driver
and colleagues at Battelle Pacific
Northwest Laboratories, a head-only
exposure chamber was developed for
aerosol exposures of 10-day-old quail.
An initial study was conducted to
determine the deposition of particles in
the respiratory tract of quail after
intratracheal instillation and aerosol
exposure of polystyrene beads ranging
in size from 0.4 to 20 microns.
Subsequent tests were conducted to
test the procedures for either intra-
tracheal (I.T.) or aerosol exposure
routes. A non-motile, non-pathogenic
E. coli was chosen to represent the
bacterial agent. Fifteen minutes after
exposure, birds were euthanized and
bacterial reisolation attempted from
nasal turbinates, trachea, lungs,
syrinxes, and air sacs. A similar study
was conducted with Autographs
californica NPV, but technical problems
with bacterial contamination interfered
with attempts to reisolate the organism
from tissues. --..,-.
ELISA Development For Mallard
Antibodies
A commercial suspension of
Bacillus thuringiensis var. israelensis
(Bti) was purchased and the toxin
proteins were purified. Bacteria were
sonicated and incubated with trypsin
and proteinase K to solubHize the
proteins. The supernatant was washed
twice with ammonium sulfate and
purified on a SephadexG150 column.
Four protein peaks were produced.
Peak 4 was purified on DEAE Bio-Gei A,
and separated on SDS-PAGE electro-
phoresis into three peaks: , A 58
KiloDalton (KD) peak, and 19 or 20 KD
peak, and another 19 KD peak. The 19
KD peak, previously shown to cause
hemolysis of mouse erythrocytes/ was
used for the ELISA and coated onto the
plates at 0.2 ng/ml concentration.
Experimental mallards were
exposed to Bti by aerosolization. Birds
were placed inside the carrier crate and
covered with a plastic bag. The
bacteria was nebulized into the bag for
one hour which resulted in 145 ml of
the suspension being used. Ten birds
received a 1:10 dilution of the
commercial preparation, ten birds
received a 1:100 dilution, and five birds
204
-------
served as negative controls. Birds were
allowed to inhale the contaminated air
for 45 minutes. They were exposed six
times at 14-day intervals with blood
samples collected every 28 days during
this period. Antibodies to the Bti toxin
were measured by the ELISA.
RESULTS AND DISCUSSION
Oral and Intravenous Protocols
The baculovirus, Autographs
californica NPV, and the fungus,
Metarrhizium anisolpiae, were
nonpathogenic to the birds when
administered orally or I.V. inoculation.
They caused no detectable alterations
in the physiological parameters
measured. The studies with S.
pullorum showed that 10-day-old quail
are very susceptible to pullorum
disease, with 65-100% mortality;
however, no mallards died or exhibited
signs of morbidity. Blood chemistry
values and antibody response assays of
the quail were inconclusive due to high
mortality rate, small volumes of serum
obtained and few birds that survived
long enough to seroconvert. Mallards
inoculated orally had significant
alterations in serum chemistry values;
antibody titers were detected as early
as 1 week post inoculation. Bacteria
were isolated from ail tissue of quail at
necropsy and from feces collected 6
days post inoculation. Attempts to
isolate the organism from mallard fecal
samples were unsuccessful and
bacteria was isolated only from liver
tissue of four birds. Tissue samples
examined histopathologically showed
the quail to be severely affected. Gross
and histopathological examination of
mallard tissues were normal. Use of S.
pullorum as a positive control
demonstrated that: 1) the protocols are
written such that lethal and sublethal
effects can be detected; 2) there can
be significant differences in suscept-
ibility between mallards and bobwhite;
3) S. pullorum is an effective positive
control for bacterial tests with bobwhite
but not for mallards.
Respiratory Protocol
Particle size distribution in the
respiratory tract of young quail differed
significantly between I.T. and aerosol
inoculation methods. The I.T. route
deposited almost twice as many
particles in the respiratory tract
compared to the aerosol exposure, and
particles were much larger in size. No
particles greater than 0.8 microns
lodged in any air sacs following aerosol
exposure whereas I.T. instillation
deposited 1O micron particles in the
clavicular air sac and 5 micron particles
in the thoracic air sacs. The
intratracheal inoculation method caused
sneezing and coughing by the birds
which resulted in contamination of the
workers and the exterior of the bird.
From these preliminary studies, it was
concluded that aerosol exposure is a
safer, more realistic exposure route
than intratracheal instillation.
ELISA Development for Mallard
Antibodies
Antibody production was observed
when antibodies were measured against
the 19 KD protein but the test lacked
the required specificity. Preinoculation
205
-------
samples and control birds also showed
positive antibody response. When an
ELISA was developed using the 58 KD
protein as the antigen, the specificity
increased significantly. Antibody titers
decreased rapidly following a single
exposure to Bti. These experiments
demonstrate the potential for use of the
ELISA as a tool to monitor sentinel
mallards exposed to Bti applications.
FUTURE STUDIES
The oral and intravenous exposure
protocols require testing with a
protozoan to determine if and where
difficulties might arise when working
with this class of microorganism. In
addition, we plan to develop a positive
control system with a protozoan that is
pathogenic in birds.
The respiratory protocol requires
further testing with a pathogenic
bacterium (e.g. S. pullorum), a virus,
fungus, and protozoan. Mallards
should be incorporated into the
respiratory tests to determine if there
are differences in particle deposition
patternsdueto anatomical/physiological
differences in respiratory tract.
Work will continue with the mallard
ELISA test. It is anticipated to employ
this test with free-ranging birds this
summer or fall once an area is
established where Bti is being applied.
The oral, intravenous, and
respiratory protocols developed test for
lethality and pathogenicity. The next
step is to develop protocols that test
whether an MPCA would cause reprod-
uctive impairment. The studies will
include a chronic feeding exposure
study, and evaluation of contamination
of embryonated eggs via eggshell pene-
tration of microorganisms. Fertility,
teratogenesis, hatchabiltiy, and egg
shell quality will be major endpoints
measured.
PUBLICATIONS
Buchholz> P. and A. Fairbrother. 1991.
Pathogenesis of Salmonella pullorum in
northern bobwhite and mallards. Avian
Diseases (in press).
Driver, C., L. Smith, J. Briant, P. Van
Voris, A. Fairbrother, and P. Buchholz.
1990. Laboratory test methods of
exposure of microbial pest control
agents by the respiratory route to
nontarget avian species. U.S.
Environmental Protection Agency,
Environmental Research Laboratory,
Corvallis, OR. EPA 600/3-90/007.
Fairbrother, A. and P. Buchholz. 1990.
Laboratory test methods of exposure by
oral and intravenous routes of microbial
pest control agents to nontarget avian
species. U.S. EPA Environmental
Research Laboratory, Corvallis, OR.
EPA 600/3-90/002.
Fairbrother, A. and P. Buchholz. 1990.
Synthesis Report: Laboratory test
methods for exposure of birds to
microbial pest control agents. U.S.
EPA Environmental Research
Laboratory, Corvallis, OR. EPA 600/3
90/079.
206
-------
ROUTE SPECIFICITY OF THE TOXICITY OF
THE Bacillus thuringiensis subsp. israelensis 28 KILODALTOIM PROTEIN
C. Y. Kawanishi and M. E. Mayes
U.S. Environmental Protection Agency
Health Effects Research Laboratory
Research Triangle Park, North Carolina
INTRODUCTION
Repeated human exposure to elevat-
ed levels of natural and genetically
altered microbial pesticides and
biotechnology agents increases the
probability of occurrence of detrimental
human health effects. This is partic-
ularly true of agents that are broadcast
or applied on a large scale in the
environment. Thus, agents or factors
with such potential must be investi-
gated to aid in the risk assessment
process.
The polypeptides of the parasporal
crystals of Bacillus thuringiensis subsp.
israelensis (Bti) have been reported to
produce a variety of undesirable bio-
logical effects such as cytolysis,
hemolysis, neurotoxicity, etc., as well
as mammalian toxicity. Our previous
studies implicated the solubilized 28
kDa polypeptide of the parasporal
crystal as the cause of mammalian
effects including toxicity and hemolysis
of various mammalian erythrocytes
(Mayes et al., 1989). Monoclonal
antibody affinity-chromatography
purified 28 kDa polypeptide was
demonstrated to cause a profound and
lethal hypothermia and bradycardia in
mice challenged intraperitoneally with
this protein. The 28 kDa polypeptide
was found to possess minimal insect-
icidal activity (Held et al., 1986).
To more fully assess the potential
for detrimental effects in humans,
studies were carried out to ascertain
the effects of different routes of
challenge with the solubilized Bti
parasporal crystal 28 kDa polypeptide in
rats. Additionally, changes in serum
parameters after intraperitoneal chal-
lenge with the protein mixture were
monitored. Results are presented in this
report.
METHODS
Rats were challenged intra-
peritoneally (IP) and intratracheally (IT)
as described for mice by Mayes et al.
(1989) and Sherwood et al. (1988),
respectively. Sprague Dawley rats were
dosed intravenously (IV) by injection
into a lateral tail vein and sub-
cutaneously (SC) into the area of the
back. Peroral treatment (PO) was
accomplished with the use of a gavage
needle. A range of dosages were
utilized for the'various routes. Mice
were observed for 14 days for clinical
signs and mortality.
Studies of the effects of solubilized
Bti parasporal crystal polypeptides on
blood parameters were undertaken.
One hundred gram rats were challenged
IP with an 8 //g/g mixture of crystal
polypeptides and, two hours later,
blood was collected by cardiac
207
-------
puncture. Serum was obtained by
allowing the blood to clot in Eppendorf
tubes for 20 min. at 0°C and then
centrifuged. Serum was analyzed with
a Centrifichem400 automated analyzer.
RESULTS
Results showing the effects of the
maximum administered dose by various
routes {Table 1) indicate that only IP
injection of rats caused mortality. An IP
injection LD50 of 1.95 //g £?/solublized
parasporal crystal protein/g body
weight was estimated from the
complete data set. Preliminary results
with a limited number of animals
challenged IV, IT, SC and PO indicate
that there was no mortality. SC
injection, however, caused localized
cutaneous necrosis at the site of
administration. The area of skin
involved was proportional to the
magnitude of the dose.
Table 1. Toxicity to rats of solubilized Bacillus thuringiensis subsp. israelensis crystal
polypeptides by different routes of administration.
Route of
Administration Maximu
intratracheal 10 p
m Dose Mortality
g/g o/e
intravenous ' 21 Mg/g 0/6
per os
9 M$
3/g o/e
subcutaneous8 9 ^g/g 0/12
intraperitoneal g ^c
]/g 21/39b
a An increased lesion size was correlated with increased dosage.
b LD50 = 1.95 0/17 »g/g (p.0.05).
208
-------
The changes in blood parameters
indicate effects in numerous organ
systems (Table 2). Serum levels of
albumin, blood urea nitrogen,
creatihine, creatinine phosphokinase,
glucose; and phosphate were signifi-
cantly elevated in rats challenged IP
with Bti. Serum levels of sodium were
sigificantly decreased in the same rats.
Parameters which were not significantly
affected were: alkaline phosphatase,
alanine transaminase, total bilirubin,
total protein, amylase, total CO2 ,
lactate dehydrogenase, albumin/
globulin, anion gap, calcium, potassium,
and chloride.
DISCUSSION
The 28 kDa protein of. the Bti
parasporal crystal is toxic IP to a
number of vertebrate species (Mayes,
et al., 1990). Consequently, it is all the
more surprising that the present results
indicate that it is lethal to rats only
when administered IP. The estimated
median lethal dose of 1.95y/g/g by this
route is close to that reported for Swiss
Webster mice (2.3 //g/g) (Mayes et al.,
1989). SC injection, however, pro-
duces localized cutaneous necrosis at
the site of administration. The lack of
effects by the IV, IT and PO routes
Table 2. Changes of serum chemistry parameters in rats injected intraperitoneally with
solubilized polypeptides from Bacillus thuringiensis subsp. israelensis;
Target Organ System
Parameter Heart Kid
ney Pancreas Other
Albumin f
BUN t
Creatinine T
CPK f
Glucose
Phosphate
Sodium
t
t
: : :"." ...".*. . ":
209
-------
remain unexplained. It is interesting to
note that although Bti produces
dramatic hemolysis and cytolysis in
vitro, preliminary evidence in our
studies show no measurable hemolysis
in vivo after IP challenge (Mayes et al.,
1989).
FUTURE WORK
Future work will attempt to explain
the mechanisms involved in the route
specific toxicity of Bti. Factors that
may effect the activity of the 28 kDa
protein when challenged by routes
other than IP will be investigated. This
should help to clarify the risk to human
health of the Bti 28 kDa protein.
PUBLICATIONS
Held, G.A., Huang, Y-S., and
Kawanishi, C.Y. 1986. Effect of
removal of the cytolytic factor of
Bacillus thuringiensis subsp. israelensis
on mosquito toxicity. Biochem.
Biophys. Res. Comm., 141 (3),
937-941.
Mayes, M.E., Held, G.A., Lau, C.,
Seely, J.C., Roe, R.M., Dauterman,
W.C. and Kawanishi, C. Y. 1989.
Characterization of the mammalian
toxicity of the crystal polypeptides of
Bacillus thuringiensis subsp. israelensis.
Fund. Appl. Toxicol. 13, 310-322.
Mayes, M.E., Kallapur, V.L.,
Dauterman, W.C., Roe, R.M., Edens,
F.W.,and Kawanishi, C.Y. 1990.
Comparative toxicology of the
delta-endotoxin of Bacillus thuringiensis
subsp. israelensis in the mosquito, rat
and quail. Abstract for the
Entomological Society of America, New
Orleans, Louisiana, Decembers, 1990.
Sherwood, R.L., Thomas, P.T.,
Kawanishi, C. Y. and Fenters, J.D.
1988. Comparison of Streptococcus
zooepidemicus and influenza virus
pathogenicity in mice by three
pulmonary exposure routes. Appl. and
Environ. Microbiol. 54:1744-1751.
Sherwood, R.L., Byrne, M.J.,
Kawanishi, C.Y. and Sjoblad, R. 1991.
Comparison of hemolytic and toxic
components of Bacillus cereus, B.
thuringiensis variety israelensis (Bti) and
B. thuringiensis variety kurstaki (Btk).
Abstract for the American Society of
Microbiology, Dallas, Texas, May 5-9,
1991.
210
-------
CLEARANCE OF ENVIRONMENTAL PSEUDOMONADS FROM
CD-1 MICE FOLLOWING INTRANASAL EXPOSURE
S. E. George, M. J. Kohan, D. A. Whitehouse, and L. D. Claxton
U.S. Environmental Protection Agency
Health Effects Research Laboratory
Research Triangle Park, North Carolina
INTRODUCTION
Even though environmentally
released microorganisms are considered
safe from a health standpoint, there is
always a potential for the occurrence of
adverse health effects. Exposure may
occur through ingestion, inhalation, or
dermal contact with the microbial
product. Previous research in our
laboratory has investigated the ability of
engineered microorganisms to survive
in the intestinal tract and compete with
the normal flora of mice. Translocation
of these microbes to the spleen and
liver, which can be indicative of a
possible systemic infection, has also
been examined. In these previous
studies, the microorganisms were
introduced by gavage.
The current study investigates
potential health effects associated with
intranasal (i.n.) exposure to engineered
microorganisms. Effects monitored
include a) morbidity, b) mortality, and
c) survival of the dosed strains in the
lungs, nasal washing, and intestinal
tract.
METHODS
Bacterial Strains and Preparation of
Dosing Suspension
Overnight (16-hour) yeast ex-
tra ct-trypton e cultures of Pseudomonas
aeruginosa strain ACS69 (Chatterjee et
al., 1982, Mol. Gen. Genet., 188,279)
and P. cepacia strain AC1100 (Kilbane
et al., 1982, Appl. Environ. Microbiol.,
44,72) were concentrated in phosphate
buffered saline (PBS). Subsequent
dilutions were made in PBS where
indicated.
Intranasal Exposure
Fasting (16 hour) strain CD-1 male
mice were anesthetized with methoxy-
fluorane and dosed i.n. with 50 fj\ of
the mic'robial suspension. At 3 hours,
and 1, 2, 5, 7, 10, and 14 days after
treatment, animals were sacrificed and
the lungs and intestinal tract (small and
large intestines, cecum) removed and
placed in buffer. The nasal cavity was
washed with PBS. Animals and tissues
were concurrently weighed.
Enumeration of Dosed Strains
Homogenized tissues were diluted
and plated onto Pseudomonas isolation
agar (PIA) with (AC869) or without
(AC1100) 50 jjg/ml kanamycin. Plates
were incubated at 30 °C for 48
(AC869) or 72 (AC1100) hours.
Colony forming units (CPU) were
determined and representative colony
types isolated for antibiotic sensitivity
determinations to insure source
identification.
211
-------
Statistical Analysis
Pulmonary Survival
A two-way analysis of variance was
performed to determine what effect the
treatment had on lung and animal
weights. A value was considered
significant if P<0.05.
RESULTS AND DISCUSSION
Morbidity and Mortality
P. aeruginosa strain AC869 had an
LD50 of 2.7 x 107 CPU. When mice
were dosed with ~109 CPU of strain
AC869, mortality occurred within 24 to
36 hours. Slight morbidity occurred in
animals treated with ~107CFU. These
animals had ruffled fur, weight loss,
and were lethargic. A statistically
significant (P<0.05) increase in lung
weights occurred in these animals as
well. No effects were observed when
mice were administered ~ 103 CPU.
When mice were treated with ~ 108
CPU of P. cepaciastrain AC1100, slight
morbidity was observed. The decrease
in body weights was accompanied with
an increase in lung weights. Animals
also were lethargic and had ruffled fur
but recovered within 5 days after
treatment.
The slight morbidity observed in
treated animals is probably due to
endotoxin (lipopolysaccharide), a
component of gram negative bacterial
cells. Mortality in strain
AC869-treated animals may occur
through endotoxin shock possibly
coupled with pathogenicity factor(s).
Strain AC869 also may have more
endotoxin than strain AC1100.
Strain AC869 was detectable at
•high levels in the nasal wash and lungs
of animals treated with ~ 109 CPU prior
to animal death. When treated with
~ 107 and ~ 103 CPU, clearance of the
dosed strain was dose dependent.
When administered ~103 CPU, strain
AC869 was not detectable in the lungs
after 3 hours, but was present in the
nasal washings intermittently for 14
days. Even though there are no overt
health effects associated with this
dose, the nasal cavity may serve as a
reservoir for future intestinal or
pulmonary inoculations. The ~107
CPU dose was cleared from the lungs
after day 7 but was evident in the nasal
washings at day 14.
• Unlike strain AC869, the ~108
CPU dose of P.- cepacia strain AC1100
was cleared from the lung after day 7
and from the nasal cavity after day 2.
Even though strain AC1100 is not
detectable in the nasal wash, it may
still be present at numbers below the
detection limit of the method used.
Intestinal Survival
Strain AC869 was easily detected
in the Gl tract within 3 hours of
treatment. In the mice treated with the
highest- dose (~109 CPU), recovery
values remained constant until 100%
mortality occurred. At the end of 14
days, strain AC869 was still detectable
in mice treated with ~ 107 CPU in all 3
sections of the Gl tract. However, the
strain was cleared from the cecum and
large intestine after 3 hours and never
212
-------
detected in the small intestine when
mice were administered ~ 103 CPU.
P. cepacia strain AC1100 was
detectable in all three regions of the Gl
tract at 1 day post treatment, cleared
from the small intestine by day 2, and
not detectable in the cecum and large
intestine after day 2. When dosed by
gavage, strain AC1100 is not recovered
from the Gl tract even at 3 hours after
treatment. When dosed i.n., mucus
may coat the bacteria and protect it
from adverse conditions in the stomach
and small intestine.
The Gl tract is a reservoir for
potential pathogens. Endotoxin can
promote translocation from the Gl tract
to the mesenteric lymph nodes, spleen,
and liver, causing a systemic infection.
The treated strains may be harbored in
the Gl tract or alter the normal flora
producing future adverse health effects.
FUTURE WORK
Intranasal exposure to other
biotechnology agents will be
investigated in endotoxin sensitive and
resistant mice. Translocation to the
mesenteric lymph nodes, spleen, and
liver will be determined. Alterations in
normal flora will be examined.
Morbidity, mortality, clearance, and
translocation in antibiotic-treated
animals also will be investigated.
Pathogenicity factors associated with
Pseudomonas spp. will be studied.
PUBLICATIONS
George; S.E., M.J, Kohan, D.A.
Whitehouse, J.P. Creason, C.Y.
Kawanishi, R.L. Sherwood, and L.D.
Claxton. Distribution, clearance, and
mortality of environmental
pseudomonads in mice upon intranasal
exposure. Appl. Environ. Microbiol.,
submitted.
213
-------
EFFECT OF ORGANOCHLORINE COMPOUNDS ON THE BIOACTIVATION OF
2,6-DINITROTOLUENE IN FISCHER-344 RATS
S.E. George, R.W. Chadwick, M.J. Kohan, and J.C. Allison
U.S. Environmental Protection Agency
Health Effects Research Laboratory
Research Triangle Park, North Carolina
INTRODUCTION
Exposure to environmentally
released microorganisns may occur
during production, application, or use of
biotechnology products. There also is
a potential for exposure to the target
environmental contaminant and its
microbial metabolites that arise during
bioremdiation. Therefore, it is
important to understand the effects of
exposure to these hazardous chemicals
and their metabolites in conjunction
with microorganism use.
The intestinal tract contains
enzymes that can transform pro-
mutagens and procarcinogens to their
mutagenic and carcinogenic metabo-
lites. Key enzymes involved in this
process include nitroreductase and
p-glucuronidase. For example, 2,6-
dinitrotoluene (DNT), a hepato-
carcinogen, has been reported to
require Gl tract bioactivation to express
genotoxicity.
In this study, Fischer 344 rats were
orally treated with the organo-chlorine
compounds, pentachlorophenol (PCP),
2,4,5-trichlorophenoxyacetic acid
(2,4,5-T), or Aroclor 1254.
2,6-DNT-induced urine mutagenicity,
DNA adduct formation, and key Gl tract
enzymes were assayed to determine if
alterations in Gl tract flora and/or
metabolism occur due to exposure of
these compounds.
METHODS
Animal Exposure
Weanling male Fischer 344 rats
were dosed p.o. with 20 mg/kg PCP,
54 mg/kg 2,4,5-T, or 25 mg/kg Aroclor
1254 daily for 5 weeks. At 1,2,4, and
5 weeks of treatment, rats were
administered 75 mg/kg 2,6-DNT and
placed into metabolism cages for urine
collection.
Urine ^Collection. Preparation.
Mutagenicitv Bioassav
and
Twenty-four-hour urine samples
from animals treated with the peanut oil
vehicle, the organochlorine compound,
2,6-DNT plus peanut oil, or the organo-
chlorine compound plus 2,6-DNT were
collected over dry ice. Urine .was
treated with p-glucuronidase and
sulfatase and then concentrated on
C-18 columns. The resulting methanol
extract was dried under N2 and stored
at -80°C until bioassay. Samples were
diluted in DMSO and tested for
mutagenicity in a .microsuspension
bioassay using Salmonella typhimurium
strain TA98 without metabolic
activation. At week 5, urine was
fractionated by HPLC (C-18) and
mutagenicity bioassay performed
directly on the fractions that were
solvent exchanged into DMSO.
214
-------
Gl Tract Enzyme Analysis
Six rats treated with the
organochlorine compound and 6 rats
that received the peanut oil vehicle
were sacrificed at 2 and 4 weeks of
treatment and the small intestine,
cecum, and large intestine were
removed and placed into reduced
buffer. The tissues were homogenated
under CO2 and placed into an anaerobic
chamber where aliquots were pipetted
into the reaction vessels. The reaction
vials were stoppered, removed from the
anaerobic chamber, and 25 //I of the
enzyme substrate mixture (in DMSO,
20 mg/ml p-nitro.phenyl-p-D-glucur-
onide,3,4-dichloronitrobenzene)added.
The reaction mixtures were incubated
for 1 hour at 37°C with gentle mixing.
The reaction was terminated and
products (yo-nitrophenol, p-glucuron-
idase; 3,4-dichloro-aniline, nitro-
reductase) extracted, derivatized, and
analyzed.
DNA Adduct Analysis
DNA was isolated from frozen livers
and digested to mono-nucleotides with
micrococcal endonuclease and spleen
phospho-diesterase. P1 nuclease was
used to select for adducted nucleotides.
The DNA was labeled with (y-32P) ATP
and T4 polynucleotide kinase. The
labeled DNA adducts were resolved on
PEI-cellulose TLC plates.
RESULTS AND DISCUSSION
Pentachlbrophenoj Effects
By week 4 of PCP treatment, a
decrease in nitroreductase activity
coincided with toxic urine in the
bioassay. Five weeks after PCP
treatment, urine again was collected
and fractionated by HPLC to separate
the toxic components from the
mutagens. Animals that received both
2,6-DNT and PCP had more mutagenic
urine than animals that received
2,6-DNT alone. Peanut oil controls and
PCP only treated rats voided non-
mutagenic urine. PCP potentiated the
formation of 2,6-DNT-derived DNA
adducts whereas none were observed
in peanut oil controls or rats treated
with PCP only.
Arochlor 1254 Effects
2,6-DNT induced urine geno-
toxicity was elevated in Aroclor-treated
rats at 1,2,4, and 5 weeks after treat-
ment. -This finding corresponded with
either a decrease in small intestinal
nitroreductase activity or an elevation in
small intestinal (J-glucuronidaseactivity.
Again, at week 5, elevated DNA adduct
formation corresponded with urine
genotoxicity.
2,4,5-TrichloroDhenoxyacetic Acid
Effects
In contrast to PCP and Aroclor
1254 effects, 2,4,5-T treatment
reduced urine 2,6-DNT-induced geno-
toxicity 1 week after treatment with
levels returning to control values at
weeks 2 and 4. No differences in DNA
adducts were observed after 5 weeks
but small intestinal nitro-reductase
activity- was decreased at week 1 and
p-glucuronidase activity reduced at
week 2. Unlike PCP and Aroclor 1254,
2,4,5-T is not an inducer of the hepatic
215
-------
P450 monooxygenase enzymes. From
our studies, it appears that the
following criteria must be met to
elevate 2,6-DNT-induced urine
genotoxicity and hepatic DNA adduct
formation: a) small intestinal
nitroreductase activity reduced and/or
p-glucuronidase activity elevated, and
b) the compound must induce hepatic
monooxygenase enzymes.
FUTURE WORK
Future studies include examining the
effects of a biodegradative micro-
organism with its target compound on
the key Gl tract enzymes. Similar
studies involving microbial metabolites
of the compound of interest may be
included.
Other future work includes
evaluating the. effect of cresote and
petroleum products on 2,6-DNT-
induced urine genotoxicity, Gl enzymes,
and hepatic DNA adduct formation.
Genotoxic metabolite identification will
be done as well.
PUBLICATIONS
Chadwick, R.E., S.E. George, J. Chang,
M.J. Kohan, J.P. Dekker, J.E. Long,
and M.C. Duffy. 1990. Comparative
enzyme activity and activation of the
promutagen 2,6- dinitrotoluene in male
CD-1 mice and Fischer 344 rats.
Cancer Lett. 52:133-119.
Chadwick, R.E., S.E. George, J. Chang,
M.J. Kohan, and J.P. Dekker. 1991.
Potentiation of the genotoxicity of 2,6-
dinitrotoluene by pretreatment of
Fischer 344 rats with penta-
chlorohpenol. Pesticide Biochem.
Physiol. 39:168-181.
Chadwick, R.W., S.E. George, J.
Chang,,M.J. Kohan, J.P. Dekker, J.C.
Allison, J.E. Long, M.C. Duffy, and L.R.
Forehand. 1991. Some effects of age,
species difference, antibiotics and
toxicant exposure, on intestinal enzyme
activity and genotoxicity. Environ.
Toxicol. Chem., submitted.
George, S.E., R.W. Chadwick, M.J.
Kohan, and J.P. Dekker.
Pentachlorophenol effect on the
activation of 2,6-dinitrotoluene to
genotoxic urinary metabolites in CD-1
mice: A comparison of Gl enzyme
activities and urine mutagenicity.
Environ. Molec. Mutagen., submitted.
George, S.E., R.W. Chadwick, J.J.
Chang, M.J. Kohan, J.P. Dekker, and
Y. Hayes. 2,4,5-Trichlorophenoxy-
acetic acid influence on 2,6-dinitro-
toluene-induced urine genotoxicity in
the Fischer 344 rat: Effect on Gl
microflora and enzyme activity. Fund.
Appl. Toxicol., submitted.
216
-------
ENVIRONMENTALLY RELEASED MICROORGANISMS
AND THE ENTERIC MICROFLORA
W. Dobrogosz,1 Y. Lin,1 M. Fiuzat,1
E. George,2 and L. Claxton2
Department of Microbiology1
North Carolina State University
Raleigh, North Carolina
U.S. Environmental Protection Agency2
Health Effects Research Laboratory
Research Triangle Park, North Carolina
INTRODUCTION
The release of natural occurring or
genetically engineered microorganisms
for bioremediational purposes has
prompted questions concerning their
effects on the environment and human
health. The Health Effects Research
Laboratory (HERL) of the Environmental
Protection Agency (EPA) is responding
to these questions. The long term
objective of our collaborative research
in this regard is to determine what
effects bioremediational microorganisms
(pseudomonads in particular) have on
the "normal microflora." of the mouse.
If bioremediational microorganisms are
able to disrupt the "normal microflora"
to any significant extent, they could
compromise it's regulatory and
protective role in the gastrointestinal
tract, thereby enhancing the host
animal's susceptibility to an assortment
of diseases.
Attempts will be made to induce
pseudomembraneous colitis in CD-1
mice by oral administration of
Clostridium difficile coincident with
alterning their protective "normal
microflora" by antibiotic tratments.
Using this in vivo model as a guideline,
we will determine if the protective
effects of the "normal microflora" can
be altered or disrupted by oral
administration of ' bioremediational
microorganisms.
The first phase of this research has
involved development of appropriate
analytical systems andmethodologies.
Described in this report are (a) bur
attempts to determine if an Oxyrase™
technology for anaerobic incubations
can be use instead of anaerobic hoods,
(b) development of culture media for
enumeration of C. difficle and C.
perfringens using the Oxyrase™
system,, and (c) isolation and
enumeration of Lactobacillus reuteri
(purportedly a member of the protective
"normal microflora") in the mouse Gl
tract.
METHODS
Growing, Enteric Anaerobes Using
Oxvrase '
™
Oxyrase is a partially purified,
sterilized suspension (0.2 microns or
smaller) of a microbial (E. coli)
membrane system that, in the presence
of a suitable hydrogen donor, removes
dissolved oxygen rapidly and com-
pletely from aqueous and semisolid
217
-------
environments. It is commercially
availablefrom OXYRASE, INC (Ashland,
OH 44805) and was used according to
the manufacuter's instruction. Broth
cultures were supplemented with 0.1
unit of Oxyrase ml"1 and 30 mM
sodium lactate. Agar pour plates were
supplemented with 0.3 units ml"1 of
Oxyrase, 30 mM lactate and overlayed
with 15-20 ml of 2% agar to slow
diffusion of oxygen in the air to the
nutrient layer containing the culture.
For agar spread plates the 2% overlay
agar contained 0.3 units ml"1 of
Oxyrase, 30 mM lactate, and 20 mM
phosphate buffer at ,pH of 7.5. All
incubations were carried out at 37°C in
a conventinal incubator.
Growth of Clostridium difficile (Strains
7698 and 10463) and C. oerfrinaens
(NCTC 8798)
Stock cultures wee maintained in
Cooked Meat Medium and inocula
prepared using BHI broth. C.
perfrlngens was enumerated using
Trypticase Soy Cycloserine (TSC) agar
counting black colonies after 8 h at
37°C. C. difficile was enumerated
using TSC agar supplemented with
Cefoxitin (16 //g ml) and counting
black colonies after 18 h at 37°C. C.
perfrlngens does not grow in the latter
medium, and C. difficile colonies do not
appear in 8 h in the former medium.
Determination of Total Anaerobes in the
Mouse Gastrointestinal (Gl) Tract
Tissue extracts were macerated and
diluted in a sterile gelatin-salts solution.
The preparations were then spread on
Brucella Agar (supplemented with
sheep blood, vitamin K, and hemihKfor
incubation in the anaerobic hood, or
pour-plated in the same medium when
using the Oxyrase.
Isolation and Enumeration of
Lactobacillus reuteri
Lactobacillus Selection (LBS) agar
medium (pH = 5.4) was used to
enumerate the total gut population of
lactobacilli. The sub-population of L.
reuteri colonies on appropriately diluted
plates is idenified and enumerated
based on the ability of L. reuteri
colonies to convert glycerol under
anaerobic conditions to a metabolic
intermediate, 3-hydroxypropion-
aldehyde, which has potent anti-
microbial activity.
RESULTS AND DISCUSSION
Results obtained to date indicate
that Oxyrase can be used as an
alternative methodology to isolate and
grow enteric anaerobes. Pure cultures
of a relatively oxygen-tolerant anaerobe
(C. perfringens) and an oxygen-
intolerant anaerobe (C. difficile) were
observed to grow in comparable
manner using either the Oxyrase™
method an anaerobic chamber. This
was observed for both liquid and agar
cultures.
Similar results were obtained when
the two methods were used to
determine the total number of
anaerobes present in (a) the entire
mouse (Gl) tract or (b) the cecum/colon
segments of the tract. Also, we were
able to determine that the CD-1 mice
which will be used later in this research
have very low background populations
218
-------
of either C. difficile-like or C.
perfringens-tike bacteria resident in their
Gl tract.
L. reuteri has been found in the Gl
tract of all animals examined to date.
This species is believed to play an
important role as a member of the
regulatory and protective "normal
microflora." Our initial attempts to
isolate this species from CD-1 mice
have been unsuccessful. An
examination of "pet store purchased"
mice, on the other had revealed the
expected complement of L. reuteri ce\\s
in their Gl tract. These results indicate,
not surprisingly, that rearing methods
can effect the composition of an
animals "normal microflora."
FUTURE WORK
The Oxyrase anaerobic technology
will be applied toward accomplishing
the goals of the proposed research
which are:
(a) to develop an in vivo model disease
system, (i) to study and monitor the
protective role of the "normal
microflora" against disease, and (ii)'to.
evaluate the influence of environ-
mentally released microorganisms on
this protective role.
(b) to determine if an in vitro model
can be developed and used for these
purposes.
219
-------
In Vitro SURVIVAL AND COMPETITION OF ENVIRONMENTAL PSEUDOMONADS
WITHIN A HUMAN FECAL FLORA CULTURE
G.M. Nelson1, L.D. Claxton2, and S.E. George2
Environmental Health Research and Testing, Inc.1
U.S. Environmental Protection Agency2
Health Effects Research Laboratory
Research Triangle Park, North Carolina
INTRODUCTION
Investigation of potential adverse
health effects due to the environmental
release of microorganisms is needed.
Previous work in this laboratory has
examined the colonization and competi-
tion potentials of environmental
pseudomonads with the intestinal flora
of CD-1 mice both in vivo and in vitro.
The current research is a comple-
mentary study investigating the ability
of some of these same Pseudomonas
strains to survive and compete in
culture with human fecal microbiota.
METHODS
Veal infusion broth enriched with
yeast extract, vitamin K1, and hemin
served as the in vitro culture medium.
Human fecal flora, obtained using
Anaerobic Culturette Collection and
Transport Systems (Becton Dickinson
Microbiology Systems), were grown
overnight in enriched veal infusion broth
and used as the culture inoculum.
Pseudomonas aeruginosa strains BC16
(isolated from a product for PCB-
degradation), PAMG (a mouse intestinal
isolate), and AC869 (a 3,5-dichloro-
benzoate degrader) were grown
overnight in yeast extract-tryptone
broth and concentrated in VPI dilution
buffer for culture inoculation. A 0.1 ml
aliquot of one competitor strain was
used to inoculate the culture.
At 12 hours, a 0.06 ml aliquot of
the culture was removed and used to
inoculate a new culture. The new
culture was incubated for 1.2 hours and
the process repeated for a total of 4
transfers (5 cultures). The competitor
strain and fecal microbiota populations
from the first, third, and fifth cultures
were enumerated by spread-plating 0.1
ml of the appropriate dilutions onto
selective media at pre-selected time
intervals.
. RESULTS AND DISCUSSION
All three competitor strains grew
anaerobically in pure culture to a level
of 108-109 organisms per milliliter.
Although Pseudomonas spp. are
aerobes, they can use nitrate as a
terminal electron acceptor. In culture
with human fecal flora, all three strains
declined in number, but were still
present in low numbers in culture 5.
P. aeruginosa AC869 appeared to be
the best survivor of the strains tested.
Survival of the competitor strains was
comparable in culture with human or
mouse flora.
Differences between the human
and mouse enumerated flora popu-
lations were minimal. The total aerobic
and anaerobic counts were higher in
220
-------
mouse fecal cultures than in the
corresponding human fecal cultures for
the 6 hour timepoint, perhaps indicating
faster growth of the mouse cultures.
No effect of strain addition on the
human fecal flora populations was
detected for strain BC16. Statistical
tests have not yet been completed to
determine if strains PAMG or AC869
alter the balance of the flora
populations.
FUTURE WORK
Flora populations and enzyme
activities for single-stage, 2-stage, and
2-stage with cell-recycling continuous
cultures will be compared to determine
which most closely mimics the in vivo
situation. Survival and competition of
environmentally relevant strains will
then be tested in continuous culture.
Rodent and human in vitro results then
can be combined with rodent in vivo
results in a parallelogram approach for
extrapolation of potential human health
effects,
PUBLICATIONS
Nelson, G.M., L.D. Claxton, J.P.
Creason, and S.E. George. 1991. A
toxicological screen to determine in
vitro survival and competition of
environmentally applied pseudomonads
with mouse fecal microbiota.
Environmental Toxicology and
Chemistry, 10:597-608.
221
-------
-------
SESSION III
RISK CONTROL RESEARCH
Decontamination and Mitigation
Investigators are exploring a variety of risk reduction strategies, including the
development of genetically altered strains for conditional lethal control of survival and
gene exchange and procedures for physically and chemically decontaminating field
sites. These procedures will be evaluated under a variety of environmental conditions.
Field Releases
ORD scientists develop criteria for evaluating containment and monitoring
strategies. Test methods are based on the actual application of recombinant and
surrogate organisms to specially designed experimental release sites representing
different environmental habitats. Methods for detection, monitoring distribution,
dispersal, and dislodging characteristics, and for proper containment of
microorganisms during field releases are being evaluated.
Process Containment
Researchers produce the appropriate scientific information to develop protocols for
evaluating the causes of emissions from biological process facilities, and control and
reduction of risks associated with these emissions. Engineering and cost models for
assessing risks associated with the manufacture, and use in manufacture of
genetically engineered microorganisms in large-scale fermentation facilities will be
produced. Issues dealing with process equipment design, decontamination
technology, worker exposure and protection, and loss prevention techniques will be
considered. ,
223
-------
-------
THE USE OF LETHAL BACTERIAL GENES TO LIMIT THE SURVIVAL
OF INTENTIONALLY RELEASED GENETICALLY ENGINEERED MICROORGANISMS
Wade H. Jeffrey1, Stephen M. Cuskey2,
Richard B. Coffin2, and Michael Reagin1
Technical Resources, Inc.1 and
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Gulf Breeze, Florida
INTRODUCTION
Concerns about the release of
genetically engineered microorganisms
to the environment may be assuaged if
the released strains contain a mechan-
ism whereby they may be killed after
completion of their appointed task(s).
This mechanism should, ideally, be
contained within the organism and
should not hinder the survival of the
released strain prior to the application
of the lethal signal, i.e. the strains
should have conditional lethality. A
potentially lethal gene putatively
involved with plasmid maintenance, kilA
from plasmid RK2, has had its regu-
lation altered so that expression is
dependent on the presence of an added
compound. The native upstream
regulatory sequences were removed
from kilA and replaced with the TOL
plasmid lower pathway operator-
promoter OP2 to form plasmid pEPA88.
Cells containing this plasmid and an
appropriate regulatory gene should be
killed in the presence of an added OP2
inducer such as benzoic acid.
This report contains the results of
tests on Pseudomonas aeruginosa
PA01 containing pEPA88. Effects on
growth, lethality, and the previously
uncharacterized mechanism of action of
kilA were examined.
METHODS
P. aeruginosa PA01 was trans-
formed with pE.PA88 and incubated in
a minimal salts medium in the presence
and absence of benzoate. Several
parameters were examined including
increases in culture turbidity, cell
survival (CPU's), 3H-glutamate uptake,
and the differences in these parameters
in different ceil size fractions. A
second lethal plasmid was constructed
which placed kilA down stream of the
inducible tac promoter (pWHJ63). This
plasmid was transformed into E. coli
JM109 and the effect of kilA expres-
sion on DNA synthesis was determined.
RESULTS
Microscopic examination of PA01
(pEPA88) cultures incubated in
benzoate minimal medium showed two
distinct morphological forms.
Numerous long, filamentous cells were
observed as well as a significant
number of normal size cells. Filaments
appeared after a 2 hour incubation with
benzoate and remained until the
termination of the experiments at 24
hours. Filaments were not formed in
samples which did not contain
benzoate. Filaments could be separated
from single cells by differential filtration
225
-------
through 3 ;/m pore-size filters.
Microautoradiography of cells labelled
with 3H-glutamate indicated that
filaments were metabolically active.
After 4 hours, 5 % of the biomass (jjg
protein/ml) was in the greater than 3
//m fraction (filaments). By 8 hours and
12 hours, 60% and 70% of . the
biomass,. respectively, was in the
greater than 3 //m fraction. By 24
hours, 50% of the biomass was
filaments. At 4 hours, 17% of the 3H-
glutamate was taken up by the greater
than 3 //m fraction. At 8 hours, 45%
of the uptake was by filaments. By 12
hours, 87% of the uptake was in the
larger fraction and 83% was taken up
by filaments after 24 hours exposure to
benzoate. Biomass specific rates of
glutamate uptake were 3 to 5 times
greater for filaments at all time points.
In contrast to metabolic activity
measurements, all of the colony
forming organisms were found to be in
the less than 3 /jm fraction. This
indicates that while the filaments were
more active, they were incapable of cell
division and subsequent colony
formation.
Filaments and single cells were also
observed in E. coli cells containing
pWHJ63 after induction of kilA by
IPTG. Filaments began forming within
an hour. Microscopic examination
indicated that the number of filaments
remained constant after a 2 hour
exposure to IPTG. 3H-Thymidine in-
corporation by these cells indicated that
filaments continued to synthesize DNA.
DISCUSSION
The use of kilA has proven inapprop-
riate as a biological control agent for
Pseudomonas aeruginosa PA01. Two
types of response have been observed
in cells containing induced kilA genes.
The first is a biostatic effect
characterized by normal sized cells
which, while not actively growing,
remain metabolically active. The
addition of an alternative carbon source
provides growth substrate and these
cells begin to grow and divide and will
form colonies when plated on nutrient
agar. The second lethal response is
noted by the production of long
filamentous cells. Metabolic activity
remains very high in these cells, yet
these cells do not grow and form
colonies when presented with alter-
native growth substrates. The lethal
mechanism of kilA has not been
previously characterized. Our results
indicate that lethality is not caused by
an inhibition of protein nor DNA
synthesis. The production of filaments
and the inability of them to divide and
form colonies indicates that kilA
functions by inhibition of daughter cell
separation. The reasons for the two
types of response are not known.
Plasmid minipreps indicate that those
cells which do grow maintain intact
pEPA88 plasmids so that loss of the
plasmid construct cannot account for
loss of activity.
PUBLICATIONS
Cuskey, S.M., and A.B. Sprenkle.
1988. Benzoate dependent induction
from the OP2 operator-promoter region
of the TOL plasmid (pWWO) in the
absence of known regulatory genes. J.
Bacteriol. 170: 3742-3746.
226
-------
Cuskey, S.M. 1990. Lethal genes in
biological containment of released
microorganisms, in: M.J. Day and J.C.
Fry (eds.), Environmental Release of
Genetically Engineered and Other
Microorganisms.
Publishers.
Edward Arnold
Cuskey, S.M. 1990. Biological
containment of genetically engineered
microorganisms. in: M. Levin, R.
Seidler, and P. Pritchard (eds.), Guide
to Environmental Microbiology.
McGraw-Hill
Jeffrey, W.J., S.M. Cuskey, R.B.
Coffin, and M. Reagin. Conditional
expression of RK2 kilA gene in
Pseudomonas aeruginosa can cause a
non-lethal inhibition of growth. To be
submitted to J. Bacteriol.
Jeffrey, W.J., S.M. Cuskey, P.J.
Chapman, and R.H. Olsen. Cloning and
characterization of Pseudomonas genes
involved in benzoate catabolism:
Isolation of a chromosomal DNA
fragment able to substitute for xylS in
the activation of the TOL lower
pathway promoter. To be submitted to
J. Bacteriol.
227
-------
FIELD RELEASE OF A GENETICALLY-ALTERED BACULOVIRUS
WITH A LIMITED SURVIVAL CAPACITY
H. Alan Wood and Patrick R. Hughes
Biological Control Program
Boyce Thompson Institute
Ithaca, New York
INTRODUCTION
Although baculovirus epizootics are
often the major factor in natural control
of lepidopterous insects, there has been
limited commercial usage of these viral
insecticides. One of the primary
problems with using baculoviruses as
microbial pesticides is that target
insects are not killed quickly and
continue to damage crops for 3 to 10
days. To overcome this problem
baculoviruses can now be genetically
engineered to enhance their
pathogenicity. New genes can be
inserted into viral genomes which will
effect a fast death.
The insertion of new genes into
baculoviruses is technically easy, and
many commercial and academic labora-
tories are currently conducting
programs aimed at producing
genetically-enhanced viral pesticides.
However, a major problem which must
be addressed prior to the release of
these products is the environmental
consequences, e.g., persistence,
spread, genetic stability, gene transfer
potential, and displacement of natural
virus populations (Wood, in press;
Wood and Granados, • in press - see
publications list). In order to address
these issues, we have developed and
tested a release strategy by which the
engineered baculovirus pesticides could
have enhanced pathogenicity but low
survival and displacement potential in
nature.
The strategy for this construction is
to delete the 'polyhedrin gene and
replace it with a foreign gene whose
product will enhance pathogenicity.
The polyhedrin gene product is required
for occlusion of virus particles within
polyhedra and is therefore essential for
survival and transmission .in nature.
Therefore, the infectivity of the altered
virus is stabilized by occlusion within
polyhedra produced by wild-type virus.
This is accomplished by co-infection of
cells with both the wild-type and
altered virus. Late in the replication
cycle, the polyhedrin produced by the
wild-type virus occludes both virus
types. Persistence of the altered virus
in the virus population is determined by
subsequent levels of co-infection and
co-occlusion.
This field release project is being
conducted to test our laboratory based
model under natural conditions. The
release was conducted with an altered
Autographs californica nuclear
polyhedrosis virus (AcMNPV) which
had a deleted polyhedrin gene but no
foreign gene inserted. The altered virus
was co-occluded with the wild-type
AcMNPV. Investigations are being
performed to evaluate the spread of
virus in nature and the survival potential
of the altered virus within the wild-type
virus population.
228
-------
METHODS AND RESULTS
During the field application phase of
this project (1989), three applications
of polyhedra containing 49% genetic-
ally-altered and 51% wild-type
Autographs californica nuclear
polyhedrosis virus (AcMNPV) particles
were made on a quarter acre cabbage
plot in Geneva, NY. The cabbage
plants were seeded with a total of
4500 Trichoplusia ni (cabbage looper)
larvae which were all infected and killed
by the virus. Polyhedra samples were
purified from,individual larvae. The viral
DNA was purified, digested with the
restriction enzyme £coR1 and fraction-
ated, by agarose gel electrophoresis.
Based on molar ratios of the £coR1 I
fragments (altered virus genome has a
smaller I fragment), it was determined
that the progeny virus produced in
these larvae contained 42% genetically-
altered virus. In addition, it was
estimated that approximately 1 x 1014
progeny polyhedra were deposited in
the field.
The questions asked in 1990 were:
1) how much of biologically-active virus
(which contained 42% altered virus)
was present in the soil in the spring of
1990, 2) with this as the starting
inoculum how much occluded, altered
virus would be produced and 3) what.
was the distribution of virus in the field
plot. The two approaches taken to
answer these questionswere soil
sampling and larval seeding.
In 1990, the test site was replanted
with cabbage in precisely the same
location and manner as in 1989. The
cabbage plants were seeded with T. ni
larvae 4 times during the growing
season. Several thousand larvae were
inspected in the field without evidence
of viral infections. When half of the
larvae started to pupate, a total of
1600 larvae were brought into the
laboratory and examined for virus
infections. Less than 1 % of the test
larvae became infected, indicating that
only a small amount of virus in the soil
was deposited on the plant tissues.
As a second evaluation procedure,
soil samples were taken from the same
97 locations in the application area as
were sampled in the spring of 1989
prior to the release. In order to
quantitate the amount of virus in the
test site, polyhedra were extracted from
the soil samples using a combination of
differential centrifugation and detergent
treatments. The amount of biologically
active virus was determined by feeding
one-gram soil extracts to each of 30
larvae from each location (2910
samples). Based on control data, the
soil extraction efficiency was 25%.
Therefore each insect was fed the
amount of virus present in 0.25 grams
of soil. The percent death from these
bioassays was used to determine the
number of biologically-active polyhedra
per gram of soil.
Based on infectivity of the soil
extracts, the average number of
polyhedra detected in soil samples from
the quarter-acre application area was
1219 _+_ 2269 polyhedra per gram of
soil. The average number of polyhedra
in the soil samples from the buffer zone
(1.75 acre area surrounding the
application area) was 274 +_ 404. The
high standard deviations are the result
of high variation between samples not
within replicated samples.
Based on laboratory studies, it was
predicted that the biological stability of
229
-------
polyhedra containing wild-type arid
altered virions would be equal to
polyhedra containing only the wild-type
virus. In order to evaluate this, soil
samples were taken from a control plot
which had been treated identically in all
respects to the application site except
that the polyhedral inocula contained
only wild-type virions. The average
number of polyhedra in this control site
in the spring of 1990 was 1151 +_
1942 polyhedra per gram of soil. This
value is very close to the number of
active polyhedra found in the
application area. It is, therefore,
concluded that occlusion of altered
virions does not effect the stability of
polyhedra under laboratory or field
conditions.
Having established the level of active
virus deposition in the field and the .
percentage of altered virus (based on
introduction data), we are now
determining how much altered virus
replication has resulted from this level
and type of inoculum. To this end,
infected larvae from the quantitation
procedure discussed above are being
further analyzed. The polyhedra from
the individual larvae have been purified
and the percentage of occluded, altered
virus is being determined by DNA
restriction analysis. As previously
demonstrated, the concentration of
inocula as well as the percent altered
virus present in the inocula determine
the amount of altered virus in progeny
polyhedra. It is considered that larvae
feeding on leaf tissue contaminated
with soil would ingest less than 0.25
grams of soil, and therefore the
inoculum level in this test can be
considered high in terms of natural
contamination processes.
We are now in the process of
completing our analysis of the amount
of altered virus present in the polyhedra
isolated from the individual larval
samples. The present data show that
approximately one-third of the
polyhedra samples contain altered virus.
The range in concentration is 12-57%
altered virus. The data set is
incomplete at present, however the
average amount of occluded, altered
virus is approximately 20%.
FUTURE INVESTIGATIONS
Questions for 1991:
A. How many biologically-active
polyhedra remain in the test site
after one year?
B. Using soil extracts as inocula, what
quantity of altered virus will be
occluded in progeny polyhedra?
The original prediction was that from
1990 to 1991 there would be a
decrease in both values. If a decrease
occurs,, then the co-infection/co-
occlusion process used in this study
would be of use to ensure that a
released engineered baculovirus would
slowly be removed from natural virus
populations(nonpersistent). According-
ly, the engineered virus could not
displace natural virus populations based
on its competitive disadvantage.
If the number of biologically-active
polyhedra do not decline between 1990
and 1991 and the amount of occluded,
altered virus remains the same as in
1990, the co-infection/co-occlusion
strategy does not provide the environ-
mental attributes for which it was
230
-------
designed. Accordingly, new
approaches will be needed to ensure
environmental safety following releases
of genetically-enhanced viral pesticides.
In the absence of biological contain-
ment, engineered baculoviruses may
pose a multitude of potential problems.
In the spring of 1991, we will again
take soil samples from the same 97 test
site locations as used in 1989 and
'1990. As performed in 1990, thirty
larvae will be fed with extracts from
one-gram soil samples (1/4 gram based
on extraction efficiency) from each site.
Based on the percent infection of
larvae, the concentration of biologically-
active polyhedra will be determined and
compared with the 1990 data.
The dead larvae from the soil assays
will then be processed to determine the
amount of occluded, altered virus
present in polyhedra. The polyhedra
from each larva will be processed and
analyzed separately. The percent
altered virus and standard deviation will
be compared with the 1990 data.
PUBLICATIONS
Wood, H.A., P.R. Hughes, N. van Beek
and M. Hamblin. 1990. An ecologically
acceptable strategy for the use of
genetically engineered baculovirus
pesticides, in: A.B. Borkovec and E.P.
Master (eds.), Insect Neurochemistry
and Neurophysiology 1989. Humana
Press, New Jersey, p. 285.
Hamblin, M., N.A.M. van Beek, P.R.
Hughes and H.A. Wood. 1990. Co-
occlusion and persistence of a
baculovirus mutant lacking the
pblyhedrin gene. Applied and Environ.
Microbiol. 56:3057-3062.
Wood, H.A. and R.R. Granados. 1991.
Genetically engineered baculoviruses as
agents for pest control. Ann. Rev.
Microbiol. 45:69-87.
Wood, H.A. Development of
genetically-enhanced faaculovirus
pesticides, in: K.M. Maramorosh (ed,),
Biotechnology for Biological Control of
Pests and Vectors. CRC Press, Boca
Raton, FL. In press.
231
-------
EPIPHYTIC FITNESS GENES AND PHENOTYPIC ADAPTATION
Steven E. Lindow
Department of Plant Pathology
University of California
Berkeley, California
INTRODUCTION
The fate and effects of micro-
organisms introduced into the open
environment are usually predicted from
an understanding of their biology and
ecology studied under controlled
conditions, usually in culture. Studies
we have made of introduced recombi-
nant bacteria reintroduced into field
plots indicated that the strains we used
exhibited considerable phenotypic
plasticity. That is, the behavior of Ice"
Pseudomonassyringae strains grew and
survived with different efficiencies
depending on the conditions under
which they had originated. For
example, we had observed that cells
harvested from broth cultures were less
capable of surviving on leaves surfaces
after inoculation than were cells grown
under solid agar conditions. If we are
to accurately predict the growth and
survival of bacteria under natural
conditions we must be aware of the
factors such as phenotypic plasticity
which may effect the behavior of cells
while in nature. Similarly, we need to
know more of the nature of the habitats
occupied by bacteria after introduction
into the open environment such as
plants since the growth and survival of
the introduced bacteria will be dictated
by how well their adaptations match
the physical and chemical environment
presented to them by the plant. We
have therefore initiated studies to
determine the range of differences in
epiphytic fitness exhibited by a given
genotype of bacterium that is exposed
to different environments prior to
inoculation onto plant surfaces. Since
bacteria that occupy specific habitats
such as leaf surfaces are rather unique
and exhibit superior ability to grow and
survive in such habitats we expect that
they posses unique traits that allow
them to exploit such habitats. Little is
known of the fitness traits that bacteria
employ to exploit natural habitats such
as leaf surfaces. If such fitness traits
were better known, it should be
possible to better define the range of
habitats that an introduced bacterium
would colonize based on its possession
such attributes. Similarly, it should be
possible to reduce the fitness of
introduced organisms by eliminating
habitat-specific fitness traits. We have
therefore initiated studies to determine
novel fitness determinants in the
epiphytic bacterium, Pseudomonas
syringae.
METHODS
Phenotypic Plasticity and
Nature of Epiphytic Habitats
' Pseudomonas syringae strain MF714
was grown for 12 hours either in Kings
Medium B (kB) broth or agar at 28°C
and harvested or inoculated onto bean
plants 24 hours prior to use. Bacterial
232
-------
cells were removed from the surface of
plant by sonication for 5 minutes in
distilled water and concentrated by
centrifugation. Bacterial cells form
each of these sources were suspended
in distilled water at a concentration of
107 cells/ml and sprayed onto bean
plants under field conditions or in a
growth chamber with different
temperature and relative humidity
conditions. Bacterial populations were
determined every 2 to 8 hours for up to
4 days after inoculation. Viable
bacterial populations were determined
by dilution plating onto KB containing
100 /yg/ml rifampicin. The proportion
of the viable cells that were culturable
was determined by microscopic exam-
ination of cells recovered form leaf
surfaces at different times that were
exposed for 12 hours to solutions of
dilute yeast extract containing 50 //g/ml
nalidixic acid. The relative exposure of
bacteria of different genotypes and of
cells of a given genotypes exposed to
different environmental conditions on
leaf surfaces was determine by
quantifying the ratio of cells surviving
exposure to a topical treatment of 15%
hydrogen peroxide or exposed to a flux
of UV irradiation of 2000 erg/mm2.
Determination of Epiphytic Fitness
Determinants
Erwinia herbicola strain BRT89 was
subjected to transposon mutagenesis
with Tn5 delivered on the plasmid
PDGS61 for which negative selection
could be imposed. This plasmid
consisted on pLAFRS containing Tn5
and the levan sucrase gene from
Bacillus subtilis. Strains containing the
levan sucrase gene are killed when
grown on media containing 5%
sucrose. Tn-5 mutants were then
inoculated onto bean plants and
subjected to alternating periods of
wetness and drying. The final
population size of individual mutants on
leaves after 73 hours of alternating
environmental conditions was estimated
by a measurement of the freezing
temperature of individual leaves cooled
slowly after immersion in tubes of ice
nucleus-free water. The relationship
between population size of lce+ strain
BRT98 and freezing temperature of
leaves was determined. Genomic
regions of Pseudomonas syringae strain
B728a identified by insertion of Tn-5
that confer epiphytic fitness were
identified in a library of genomic clones
of this strain using the flanking regions
of cloned DNA containing Tn-5 from
the mutants. Subcloning experiments
on the genomic clones were followed
by complementation studies to
determine the minimum size of DNA
necessary to confer the stress tolerance
phenotype identified using these
mutants.
RESULTS AND DISCUSSION
Pseudomonas syringae strains
differed greatly in their behavior after
application onto leaf surfaces as a
function of the physiological condition
of the cells prior to inoculation. Cells
harvested from broth cultures died
much more rapidly upon application to
plant surfaces than cells harvested from
agar plates. Broth-grown cells also
required up to twice as long to resume
growth after inoculation on leaf
surfaces. In contrast, cells of P.
syringae that were recovered from the
233
-------
surface of bean leaves exhibited up to
1000-fold greater survival than cells
harvested from broth culture after
application to leaves. Cells recovered
from leaves also resumed growth on
leaf surfaces as much as twice as fast
as even cells harvested from agar
surfaces. These results indicate that
bacteria exhibit considerablephenotypic
plasticity. It is likely the pattern of
gene expression in cells in natural
habitats such as leaves are much
different than that in culture.
Predictions of microbial behavior upon
field releases must take into account
the great deal of phenotypic plasticity
that microbes appear to possess.
Bacterial survival in natural habitats
such as leaf surfaces appears to be
greatly influenced by their modification
of the habitat. The fraction of cells of
P. syringae that survived on leaf
surfaces was directly "proportional to
the concentration of cells initially
applied to the leaf. While up to 10% of
the cells applied as a suspension of 109
cells/ml survived during the first 48
hours after inoculation on leaves, only
about 0.005 of the cells applied as a
suspension of 103 cells/ml survived
during this period. Application of 109
dead cells in conjunction with 103 live
cells increased the survival of the 103
live cells to about 10%. The rate of
immigration of cells from sites of field
releases thus could greatly effect the
efficiency with which they survive the
transport from the field site and become
established in an adjacent site.
The natural habitats of similar
bacteria appear to differ greatly. For
example, the ability of cells of different
P. syringae strains to become
established in "protected sites" on
leaves differed greatly and differed
greatly form other phylloplane residents
such as Erwinia herbicola and
Xanthomonas maltiphilia. The fraction
of P. syringae cells that could not be
killed by surface sterilization of leaves
ranged from less than 0.2% to as much
as 24% of the total bacterial microflora
on a leaf. Most non-pathogenic
bacterial species such as E. herbicola
and X. maltiphilia did not exhibit more
than 1 % of the cells in such "protected
sites". These great differences in
habitat preference are unexpected and
indicate that such preference should be
carefully established for recombinant
organisms to be released into the open
environment for which a prediction of
survival and subsequent behavior is
desired. The eradication of'bacteria
having greatly different habitat
preferences also may differ greatly and
should be further investigated.
A novel leaf freezing assay is being
used to assess the epiphytic fitness of
Tn-5 -induced mutants of E. herbicola.
Bean plants are being inoculated with
individual transposon mutants of this
species and subjected to alternating
period of wetness and dryness. The
final population size of the lce+ E
herbicola strain used in this study is
proportional to the freezing temperature
of leaves colonized by this strain.
Mutants deficient in their ability to grow
or survive on leaf surfaces are being
identified by a reduction in the
temperature at which leaves colonized
by such a mutant strain freeze.
Transposon mutants of P. syringae
which exhibit a reduced ability to
survive the stresses of drying on leaf
surfaces have been identified and are
being characterized. Most of these
234
-------
mutants do not exhibit a deficiency in
any phenotype such as osmosensitivity
that can be associated with stress
tolerance. The genomic determinants
of one such mutant that conditions the
ability to survive on leaves as surface
moisture drys from the leaf has been
localized to a 6.5kb region. This region
will compliment the reduced survival
exhibited by the mutant strain. Fusions
of this gene with a promoterless ice
nucleation gene are being made to
further localize the position and size of
the genes involved in stress tolerance
and to permit the assessment of the
transcriptional activity of this region
under natural conditions and under
conditions of specific environmental
stresses. The identification of such
novel stress-tolerance genes should
permit their identification in other
bacterial to be released into the open
environment and thus to indicate their
potential for tolerating the stresses at
field sites; Alternatively, such genes
could be specifically inactivated to
biologically contains organisms at field
sites by making them incapable of
surviving specific environmental
stresses. ,
PUBLICATIONS
Kinkel, L.L., and S.E. Lindow. 1989.
The role of competitive interactions in
bacterial survival and establishment on
the leaf surface, pp.634-638. in: T.
Hattori, Y. Ishida, Y. Maruyama, R.Y.
Morika, and A. Uchida, (eds.), Recent
Advances in Microbial Ecology. Japan
Sci. Soc. Press.
Lindow, S. E. 1990. Design and results
of field trials of Ice" recombinant
Pseudomonas syringae strains, pp. 61-
69. In: J. Marois and J. Bruhning
(eds.), Risk Assessment in Agricultural
Biotechnology: Proceedings of the
International Conference. U. of CA.,
Oakland, CA.
Kinkel,-L.L..& Lindow, S.E., 1990.
Spatial distributions of Pseudomonas
syringae strains on potato leaves.
Phytopathology 80:1030.
Wilson, S.E. and S.E. Lindow. 1990.
Phenotypic plasticity affecting epiphytic
survival in Pseudomonas syringae.
Phytopathology 80:1058.
Kinkel, L.L. , M. Wilson, and S.E.
Lindow. 1990. Sampling phylloplane
populations: Distributional effects on
sample design. Phytopathology
80:1030.
Lindow, S.E. 1991. Determinants of
epiphytic fitness in bacteria, in.: S.S.
Hirano and J. Andrews (eds.),
Microbiology of the Phyllosphere.
Springer-Verlag, New York, (in press).
235
-------
COMPARISON OF SURVIVAL ON THE PHYLLQPLANE OF BACTERIA
RELEASED IN GREENHOUSE AND FIELD ENVIRONMENTS
Katharine Donegan1, John Armstrong2, Carl Matyac1,
and Ray Seidler2
ManTech Environmental Technology, Inc.1
U.S. Environmental Protection Agency2
Environmental Research Laboratory
Corvallis, Oregon
INTRODUCTION
Because of current concerns
regarding the release of genetically
engineered microorganisms (GEMs) into
the environment, the fate, survival, and
effects of many GEMs will need to be
evaluated in small-scale releases
performed in controlled, contained
environments. In our research, the use
of greenhouses for predicting the
results of field releases, and the
influence of bacterial species, plant
species and environmental conditions
on bacterial survival in the greenhouse
and the field were investigated.
Small-scale releases in controlled
environments have been recommended
to evaluate the effectiveness of
genetically engineered microorganisms
(GEMs) and to detect any potential
adverse effects prior to their release
into the environment. Microcosms have
been used as controlled environments
in which to conduct small-scale
releases for the evaluation of the fate,
survival and effects of GEMs in aquatic
and terrestrial ecosystems. The use of
microcosms in such studies has been
criticized, however, because of their
limited size and lack of biological,
chemical and physical complexity
relative to the natural environment.
Greenhouses can provide an
environment intermediate between the
microcosm and the field for studying
the survival and effects of GEMs prior
to field release. Greenhouses allow a
larger scale rele'ase into a more complex
environment than a microcosm yet
retain some of the containment and
control advantages of a microcosm.
As has been done with microcosm
studies, greenhouse releases need to be
evaluated for their ability to predict the
results of field releases. In addition., the
influence of bacterial species and plant
species, and also meteorological
conditions, on the ability of greenhouse
releases to simulate field releases needs
to be investigated. The goal of our
research was to address these issues.
METHODS :
Plant Growth Conditions.. and
Experimental Design
The experiment was conducted once
in the greenhouse and twice in the
field. In each trial, .Erwinia herbicola
(NalR), Klebsiella planiicola (RifR), and
Pseudomonas syringae (RifR) were
sprayed 'on bean plants (Phaseolus
vulgaris] and on oat plants (Avena
sativa) for a total of six plant-bacterium
236
-------
treatments. Treatments were arranged
in a split-strip plot experimental design.
Five replicate plots of beans and of oats
were randomly assigned and planted in
the greenhouse and in the field. In the
greenhouse, seeds were planted
directly in the soil floor. Each oat and
bean replicate plot was divided in thirds
and randomly assigned a bacterial
treatment for a total of five plot
replicates per plant-bacterium treat-
ment. The area of each treatment plot
that was sprayed with a bacterial
species was 5 x 10 ft in the field and 2
x 5 ft in the greenhouse. Unsprayed
areas were maintained between plant-
bacterium treatment plots to prevent
cross-contamination.
Application of Bacteria to Plants
All bacterial suspensions for
greenhouse and field sprays were
prepared from 18 h cultures grown in
Luria-Bertani (LB) broth with the
appropriate antibiotics. The cells were
washed 3 times in 0.01 M phosphate
buffer (pH 7.2) by centrifuging for 10
min. at 5000 rpm and resuspended in
sterile water to ca. 2x10' CFU/ml.
Bacteria were applied when bean plants
were beginning to flower and oat plants
were forming seed head. Plants were
uniformly sprayed until run-off with a
CO2 pressurized backpack sprayer.
Sample Collection and Processing
Plant samples were collected in
sterile bags on days 1,3,7, 10, 14 and
21 post-application. In the greenhouse
and field II experiments, samples were
also collected two hours post-
application. Bean samples consisted of
3 leaves per plot that were collected at
•low, middle and high levels of the plant
canopy. Oat samples consisted of 3
plants per plot, with seed heads and
roots removed. A total of 15 samples
per plant-bacterium treatment was
collected per sample day. Samples
were immediately taken to the
laboratory where they were weighed
and processed by adding 20 ml of 0.01
M phosphate buffer (pH 7.2) and
treating 1 minute in a Stomacher Lab-
Blender. Samples were diluted in 0.01
M phosphate buffer (pH 7.2) and plated
on LB agar containing 500 //g/ml
nalidixic acid or 100 //g/ml rifampicin
and 100 //g/ml cycloheximide and
incubated at 30°C for 24 hours.
Monitoring Environmental Conditions
Meteorological data were collected
during the greenhouse and field
experiments with a 21 x Micrologger.
Temperature, relative humidity, and
incident light were measured every
minute and averaged per hour.
Data Analysis
Plate counts were converted to
logarithmic colony forming units per
gram of plant tissue (log CFU/g) and
used in the Statistical Analysis System
(SAS Institute) in a stepwise backward
elimination regression procedure.
Separate regression curves for each
experiment with each bacterial species
were fit in the same model by using
classification variables. Patterns of
bacterial survival were compared by
testing the hypothesis that the intercept
237
-------
and slope of the regression line for each
bacterial genus was equal for the green-
house, field 1, and field 2 experiments.
RESULTS AND,DISCUSSION
Despite the application of equivalent
bacterial concentrations in the trials,
bacterial populations after only one day
post-application weresignificantly lower
in the field 1 and field 2 trials as
compared to the greenhouse trial. In
the greenhouse and field 2 trials, two
hour post-application samples were also
collected; bacterial populations in the
field were on the average 10-fold lower
than in the greenhouse. Substantially
lower light levels occurred in the
greenhouse trial than in the field trials.
The initial large decrease in bacterial
populations in the field 1 and field 2
experiments may have been caused by
exposure to ultra-violet light. Solar
radiation is an important environmental
factor in the field that is not accurately
represented in most greenhouse and
microcosm simulations of field releases.
Because of the greater decline
during the first day in bacterial levels in
the field trials as compared to the
greenhouse trial, the intercepts of the
regression lines for the greenhouse trial
were all significantly greater than those
of the field trials. The slopes of the
regression lines, however, were not
statistically different between the
greenhouse trial and field trial 1 for the
oat P. syringae treatment and the oat
and bean E. herbicola treatments.
Survival was very different among
the three bacterial species. In addition,
the extent of agreement in results
between the greenhouse and field trials
was influenced by bacterial species.
The K. planticola treatment always
showed a decline in population and
reached the lowest bacterial level,
whether on oat or bean plants or in the
greenhouse or field. In contrast, the P.
syringae and E. herbicola treatments
displayed variable responses. Because
E. herbicola and P. syringae commonly
occur on the oat and bean phylloplane,
unlike K. planticola which is primarily a
soil inhabitant, their potential for
survival and growth was higher than K.
planticola but more variable due to a
greater influence from meteorological
conditions.
In the greenhouse experiment,
populations for all three bacterial
species were significantly greater on
oat plants than on bean plants. In the
field 1 and field 2 experiments,
however, no significant differences in
bacterial survival between plant species
were observed. These results
emphasize the need in simulated
releases to evaluate survival of a
bacterial species on the particular plant
species on which it will be released in
the field. Due to the initial decrease in
bacterial-levels in the field trials, the
greenhouse experiment accounted for
the maximum bacterial population levels
observed in the field experiments in all
but one treatment. Our results suggest
that greenhouse experiments will be
more successful in predicting trends in
bacterial populations for field releases
than actual population levels and that
the predictive success of greenhouse
experiments will vary with plant
species, bacterial species, and
environmental conditions.
238
-------
FUTURE WORK
Our results indicate that
meteorological conditions significantly
influence the survival of bacteriaapplied
to plants in field environments.
Consequently, the use of environmental
chambers or "mesocosms" that
successfully reproduce the different
meteorological factors of field
environments will be required for
accurate prediction of field releases.
Additional research evaluating the
individual and collective influence of
such variables as temperature, relative
humidity, incident radiation, leaf
moisture and rain fall on survival of
applied microorganisms will be
necessary. A statistical and experi-
mental assessment of conducting such
experiments is under discussion.
PUBLICATIONS
Donegan, K., J. Armstrong, C. Matyac
and R. Seidler. 1990. Comparison of
survival on the phylloplane of bacteria
released in greenhouse and field
environments. U.S. EPA Report
600/3-90/085.
239
-------
NATIONAL ECOLOGICAL FIELD RESEARCH STATION
Linda D. Stetzenbach, Mark P. Buttner, and James R. Meldrum
Environmental Research Center
University of Nevada, Las Vegas, Nevada
The U.S. Environmental Protection
Agency environmental biotechnology
research program is charged with the
development of methods to measure
the impact of microorganisms released
to the environment. The techniques
and data generated by this research
program are used by the Office of Toxic
Substances and Office of Pesticide
Programs in making decisions on the
human and ecological safety of
microorganisms under consideration for
release into the environment. EPA
sponsored biotechnology research
programs have focused on the
detection, fate, transport, survival,
genetic exchange, effects, and
monitoring methodology and design for
field-released microorganisms in
terrestrial ecosystems. The techniques
and methods developed, to date
however, are products of experiments
conducted in laboratory, greenhouse,
and microcosm environments. Many of
the methods now being proposed for
use in tracking and enumerating
microorganisms such as DNA-
hybridization (probes), polymerase
chain reaction (PCR), and bulk recovery
of DNA directly from soil samples, have
yet to be tested over long periods of
time under field conditions. While
useful to the program offices and the
biotechnology industry, none of the
results have been validated under field
conditions. The lack of a field research
facility has rendered it impossible to
integrate the many facets involved in
the survival of microorganisms in the
environment and has hindered the
development of predictive risk
assessment expertise. The develop-
ment of permanent National Ecological
Field Research Stations at which long-
term experiments could be conducted
would be a direct benefit to EPA and
would be useful to EPA-funded
researchers and biotechnology industry
research scientists as well.
A potential site for a pilot field
research facility is available at the EPA
experimental farm on the northeast
corner of the Nevada Test Site (NTS) in
southern Nevada. The NTS is approx-
imately 15,000 square miles of
dedicated land used by the Department
of Energy located 90 miles north of Las
Vegas with a commute from the
Environmental Monitoring Systems
Laboratory (EMSL-LV) on U.S. Highway
95 of under two hours. The farm was
originally used for radiation monitoring
studies on plants and cattle but has
been idle for several years. The farm
site is comprised of 16 acres of flat
agricultural fields with a gentle slope to
the south. Electrical power, telephone
service, and a well are in operation and
only one small laboratory space is
currently used on an infrequent basis by
the Nuclear Radiation Division of the
EMSL-LV for monitoring of range cattle
and deer. The fields, remaining building
space, and hookup site for research
trailers would be available to bio-
technology researchers. The farm
240
-------
offers a secure site, removed from
encroachment by development or
intrusion by vandals. Low cost living
and dining facilities, machine shop, and
field engineering by the NTS on-site
contractor are available within 1.5 miles
of the farm. Extensive meteorological
data are available indicating a climate
with a full range of seasons. The
maximum summer daytime temperature
is customarily 95-96°F (approximately
10-15°F cooler than Las Vegas). Light
snows are not uncommon in the winter
although the average maximum daytime
high temperature is above 60°F. High
winds may blow across the site during
the year but calm periods are common
in the early morning.
Initial experiments at this pilot
facility could include field validation of
laboratory, greenhouse, and microcosm
studies using non-recombinant bacteria
and fungi. Such experiments would
provide useful data to the program
offices on the survival, dispersal, and
fate of field-applied microorganisms and
on the effectiveness of field monitoring
methods. These experiments at a pilot
site would also illustrate the advantages
of the field research site concept.
While no one site could be expected
to meet every possible environmental
conditions, the NTS farm offers the
unique opportunity of an existing EPA
facility located in a remote yet
accessible area within easy commuting
distance to an EPA monitoring
laboratory. The facilities at the farm
are both generic and flexible to
accommodate a wide range of testing
parameters and conditions. This pilot
facility would provide a baseline from
which additional field research stations
could be patterned.
241
-------
OVERVIEW OF BIOTECHNOLOGY ENGINEERING
RISK MANAGEMENT PROGRAM
J. Burckle
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio
BACKGROUND
The biotechnology engineering sub-
program is a small component of the
Agency's Biotechnology Risk Asses-
sment Research Program, which seeks
to characterize risk through methods
development, data collection and
analyses, and mathematical modeling.
It is comprised of research projects in
three broad categories: environmental
exposure, environmental effects, and
risk control. The engineering risk
management research activities, a
major component of risk control,
addresses three major concerns:
Mechanisms of accidental or
deliberate release of the modified
genome or organism from the site of
production (e.g., as in spills or
effluents),
* Availability and effectiveness of
containment controls or destruction
techniques, and
Worker exposure, particularly due to
aerosols.
Potential commercial applications of
biotechnology are divided into two
loosely defined categories; (1)
contained product manufacturing
processes and (2) uncontained and
semi-contained processes. Our
approach has been to address these
two applications individually, with our
first and primary efforts devoted to
contained product manufacturing as
requested by the OTS.
In 1988, the first phase of the
biotechnology assessment effort was
completed, providing a state-of-the-art
review entitled, "Final Report on
Biosafety in Large-Scale rDNA
Processing Facilities." This effort was
undertaken to document existing
information regarding (1) the potential
source of hazards associated with
large-scale fermentation facilities based
on recombinant DNA, and (2) to
identify biosafety practices to manage
such risks. The report consists of four
volumes which deal with the following
subjects: a summary of applicable
regulations, a review of biological
characteristics of a large variety of
microorganisms, unit operations and
equipment used in production facilities,
releases and containment, and
methodologies for assessing worker
exposure.
242
-------
CURRENT PROJECTS
Assessment of Filter Containment of
Bioaerosols
K. Willeke, Aerosol Research and
Respiratory Protection Laboratory,
Department of Environmental Health,
University of Cincinnati; Cincinnati, OH,
M. Lehtimaki, Technical Research
Center of Finland; R. Wyza, Battelle
Memorial Institute, OH.
The OTS has requested field
verification of information used to
estimate the effectiveness of filters
to remove bioaerosols from
fermenter off-gas. Bioaerosols are
generated in significant numbers in
fermenters by the combination of
Theological properties of fermenter
broths, mechanical agitation, and
introduction of air/oxygen for
growth and CO2 by microbial
activity. Techniques are being
developed to employ off-the-shelf
equipment to characterize the
particle size distribution (by number
frequency) using light-scattering.
When operation is satisfactory, the
instrumentation will be used in
conjunction with the previous
project in the field to: (1) assess the
concentration of bioaerosol in the
fermenter headspace and attempt to
relate the concentration to operating
conditions; (2) assess the amount of
aerosol challenging the control filter;
(3) determine if the apparatus has
sufficient sensitivity to directly
measure the filter efficiency; and (4)
test the apparatus for its ability to
identify sources of bioaerosol
leakage into the workplace.
Containment Technology and Validation
Techniques
G. Leaver, G.A.L. DeCosemo, and I.W.
Stewart, Warren Springs Laboratory,
U.K.; R.H. Cumming, Teeside
Polytechnic; A.M. Cottam, Health and
Safety Executive, U.K.; AI-Hassan,
Birmingham University, Birmingham,
Alabama.
In the United Kingdom, the Warren
Spring Laboratory under the
sponsorship of the Department of
Trade and Industry, is developing
techniques to evaluate process
equipment containment. We have
completed an agreement to support
this work which addresses the
following areas: . :
a. Development and Validation of
Biological Fugitive Emission
Measurement. This project invest-
igates techniques for sampling,
storage, and rapid assay of samples
for biologically active materials.
Procedures are being developed for
optimum positioning and frequency
of sampling for several techniques.
b. Containment Aspects of Bio-
process Equipment. Research is
being conducted to develop standard
methods for measuring containment
of bioprocess unit operations and
equipment components using
various leak detection techniques so
that a breach of containment can be
detected before the introduction of
viable biological materials. Lab-
oratory studies are being conducted
to evaluate specific methodologies
and equipment. In-plant studies will
be used to evaluate applicability to
real work-place situations.
243
-------
Evaluation of Alternative Methods of
Assessment of Bioaerosols in Large-
Scale rDNA Manufacturing Facilities
K. Martinez and P. Jensen, National
Institutes of Occupational Safety and
Health, Cincinnati, OH.
In preparation for conducting in-
plant tests, we have evaluated
commercial bioaerosol samplers
designed for workplace sampling.
This effort has provided the
information needed to select the
most cost-effectivesampling system
and design of a quality assurance
program. This work is being
performed for EPA by NIOSH
experts who have conducted
laboratory experiments on a variety
of commercial samplers to determine
sampling efficiency and influence on
bacterial viability during sampling.
Evaluation of the Release of Bioaerosol
from a Fermenterin a Larae-ScalerDNA
Processing Facility
K, Willike, Kettering Institute, University
of Cincinnati, OH.
Bioaerosols were sampled at the gas
off-take in the top of a fermenter to
determine approximate particle size
distribution and droplet
concentrationsexiting the fermenter.
FUTURE ACTIVITIES
Formation of Bioaerosols
Research into the formation of
bioaerosols is a continuing need. The
investigation of the formation of
bioaerosols needs'to be extended to the
formation during separation processes
to describe the formation of bioaerosols
in the fermenter headspace based on
typical operating conditions,
Efficiencies of Processes to Inactive
Microorganisms
Research into the kill/inactivation
efficiencies of various physical and
chemical treatments commonly used for
both sporeforming and nonsporeforming
microorganisms is needed. Although
limited data on inactivation efficiencies
are available, experiments were
^conducted in demand-free water and do
not represent the typical conditions in
fermenters.
Release Applications of Microorganisms
As more biotechnology PMN's are
reviewed, information concerning
release scenarios for field tests and
other non-contained uses covered by
TSCA will be needed. An investigation
of the potential use scenarios for
release-applications of microorganisms
should focus on the method of appli-
cation, potential occupational
exposures, protocols for monitoring
workers during non-contained uses, and
efficiencies of methods of controlling
releases to minimize release to non-
target areas of the environment.
PUBLICATIONS
Cottam, A.N. Risk assessment and
control in biotechnology. In Preparation.
Gumming, R.H., arid D. Brown. 1991.
Process validation and safety in
244
-------
biotechnology. Institution of Chemical
Engineers, Symposium Series No. 124.
DeCosemo, G.A.L., I.W. Stewart, W.D.
Griffiths, and J.S. Deans. The
assessment of airborne micro-
organisms. In Preparation.
Jensen, P.A., W.F. Todd, G.N. Davis,
and P.V. Scarpino. 1990. Sampling
efficiencies of eight bioaerosol sampling
instruments challenged with an aerosol
of free bacteria. National Institute for
Occupational Safety and Health,
Division of Physical Sciences and
Engineering, Murray State University,
Department of Occupational Safety and
Health, Murray, KY; University of
Cincinnati, Department of Civil and
Environmental Engineering; and U.S.
EPA, Cincinnati, OH.
Leaver, G., et al. Integrity of
fermentation equipment - containment
and sterility. In Preparation.
Leaver, G. 1991. Measuring and
monitoring containment in bioprocess
equipment. Institution of Chemical
Engineers, Symposium Series No. 124,
pp. 349-361.
Martinez, K.F., and W.F. Todd. The
sampling efficiencies of alternative
bioaerosol sampling devices as
determined using a physical surrogate
aerosol. National Institute for
Occupational Safety and Health, and
U.S. EPA Report, Cincinnati, OH.
Pilacinski, W., M.J. Pan, K.W.
Szewczyk, M. Lehtmaki,and K. Willeke.
1990. Aerosol release from aerated
broths. Biotechnology and
Bioengineering 36:970-973
Pan, M.J., W. Pilacinski, K.W.
Szewczyk, U. Krishnan, and K. Willeke.
1991 Size characteristics of particles
released from broths. Applied
Occupational Environmental Hygiene 6,
(7): 604-607.
Stewart, I.W., and J.S. Deans.
Containment testing of cell disrupters.
In Preparation
Stewart, I.W., and T.T. Salusbury.
Evaluation and comparison of
environmental samplers and particle
monitors for bioprocessing plants. In
Preparation.
Wickramanayake,C.B., B.W. Vigon, D.
Evers, M.F. Arthur, J. Kasler, A.
Gavaskar, S. Clark, N.G. Reichenback,
and V. Kogan, 1988. Biosafety in large-
scale rDNA processing facilities.
Battelle Institute-Columbus Division and
U.S. EPA Report.
Williams, R, and D. Keir. A
methodology for hazard assessment of
high integrity containments. Atomic
Energy Agency Technology, Safety
Reliability Directorate Report. In
Preparation.
245
-------
BIOTECHNOLOGY QUALITY ASSURANCE GUIDELINES MANUAL
Amy Smiecinski and Linda D. Stetzenbach
Environmental Research Center
University of Nevada, Las Vegas
The Environmental Protection
Agency requires every monitoring and
measurement project to have a written
and approved quality assurance (QA)
plan. The purpose of the QA plan
requirement is to ensure that the data
will yield scientifically sound and
unbiased conclusions related to the
projects principle hypothesis. Quality
assurance principles and practice have
been previously published and guidance
documents have been written to assist
the physical science resegrcher- in
preparation of QA project plans but
these publications were written with
the emphasis on chemical research
projects and do not specifically address
many of the QA issues unique to
biological and biotechnology-based
research projects. Sampling,
measurement, and quality control
methods for biological and
biotechnological parameters become
increasingly more complex due to
concerns for the viability of the
organisms under study and biological/
biotechnological scientists are often
confused by the terminology used in
QA discussions owing to their
unfamiliarity with this emerging area.
The purpose of this document is to
describe and clarify QA and quality
control (QC) responsibilities prescribed
by the EPA for biotechnology
researchers. This document is intended
to be used as a guidance document and
is not to serve as a template by the
research scientist. Nor is' this
document intended to be an exclusive
source of information on quality
assurance considerations for bio-
technology researchers. It is designed
to provide a logical connection between
EPA QA policy requirements and the
research scientist focusing on
biotechnological projects. It is intended
to assist the scientist in achieving an
understanding of QA and successfully
implementing a quality assurance
project plan (QAPP).
This manual will include discussion
of the elements of a QAPP with
particular emphasis on issues en-
countered with GEMs monitoring inthe
environment and incorporating bio-
logical and biotechnolgical research
activities not discussed in the available
QA guidance literature. Comprehensive
disucssions of QA principles, practical
applications, and definitions are also
included.
246
-------
CONTRIBUTOR INDEX
Abebe, H.M., 56, 162
Allison, J.C., 214
Anderson, Richard L., 172, 176
Anderson, Anne, 81
Armstrong, J.L., 52, 87, 236
Barkay, Tamar, 118
Benfield, Ernest F,, 190
Bentjen, Steve A., 151
Bleakley, Bruce H., 158
Bolton, Harvey, Jr., 140, 151
Bott, Thomas L., 135
Buchholz, Phyllis S., 203
Buell, Robin, 81
Burckle, John, 242
Butterworth, Julie, 75
Buttner, Mark P., 240
Byrne A., 113
Chad wick, R.W., 214
Cheng, H.-P., 113
Claxton, Larry D., 211, 217, 221
Coffin, Richard B., 98, 225
Connolly, John P., 98
Crawford, Don L., 151, 158
Gripe, Geraldine M., 187
Cuskey, Stephen M., 225
Devereux, Richard, 131
Dickman, Martin B.; 124
Dobrogosz, W., 217
Donegan, K., 236
Doyle, J.D., 145, 151
Fairbrother, Anne, 203
Ferrante, A., 113
Fieland, Valerie, 70, 162
Fisher, W.S., 183
Fiuzat, M., 217
Fournie, Jack, 183
Fredrickson, James K., 140, 151
Gander, L.K., 145
Ganio, Lisa, 70
Gealt, Michael A., 165
Genthner, Fred, 187
George, Elizabeth, 211,214, 217, 221
Harris, D., 87
Heck, Chris, 81
Hendricks, C.W., 145, 151
Hoefle, Manfred, 4
Hughes, Patrick R., 228
Ingham, E.R., 145
James, Rosalind D., 196
Janssen, David M., 172, 176
Jeffrey, Wade H., 225
Kaplan, Louis A., 135
Katsuwon, Jirasak, 81
Kawanishi, C.Y., 207
Kearns, Peter E.W., 7
Khalil, T.A., 165
King, R.J., 56
Kohan, Michael J., 211, 214
Kokjohn, Tyler A., 102, 108
Landeck, Robin, 98
Lenski, Richard E., 91
Leser, T., 10
Leslie, John F., 124
Lessie,T.G., 113
Lighthart, Bruce, 196, 199
Lin, Y., 217
Lindow, Steven E., 232
MacKenzie, David R., 20
Marthi, Balkumar, 61
Matyac, C., 236
Mayes, M.E., 207
McCartney, H. Alastair, 75
Mclntyre, T.C., 15
McKenney, C.L., 183
Mead, Eric, 172
Meldrum, James R., 240
Middaugh, D.P., 183, 187
Milewski, Elizabeth, 22
Miller, Robert V., 102, 108
Nelson, Gail M,, 221
Nguyen, Toai T., 91
Porteous, L.A., 52
247
-------
Reagin, Michael, 225
Reuss, M., 10
Sayler, Gary S., 102
Sayre, Philip G., 29
Schneider, William R., 31
Seidler, Ramon J., 56, 70, 162, 236
Selvaratnam, S., 165
Sewall, O.K., 199
Shaffer, Brenda, 61
Shannon, Lyle J., 172, 176
Short, K.A., 56
Smiecinski, Amy, 246
Snarski, Virginia M., 169
Stahl, David A., 37
Stetzenbach, Linda D., 240, 246
Walter, Michael V., 70
Wang, Zemin, 151, 158
Whitehouse, Douglas A., 211
Winfrey, Janet, 46
Winfrey, Michael R., 46
Wood, M.S., 113
Wood, H. Alan, 228
Yousten, Allan A., 190
Zdor, Rob, 81
248
-------
AFFILIATION INDEX
A.F.R.C. Institute of Arable Crops Research, Rothamsted Experimental Station,
Harpenden, Hertsfordshire, United Kingdom, 75
American Scientific International, Duluth Environmental Research Laboratory, 172
Argonne National Laboratory, Illinois, 102, 108 • '•-,..
Battelle Pacific Northwest Laboratory, Washington, 140, 151 , * '.
Boyce Thompson Institute, New York, 228 • .- •
Denmark Ministry of the Environment, 10 ' ,
Drexel University, 165
Environment Canada, 15 ;
Environmental Health Research and Testing, Inc., U.S. Environmental Protection
Agency (EPA) Health Effects Research Laboratory, 221
European Organization for Economic Cooperation and Development, 7
German National Research Center for Biotechnology, 4
Manhattan College, 98
ManTech Environmental Technology, Inc., U.S. EPA Corvallis Environmental Research
Laboratory, 56, 61, 70, 87, 145, 151, 162, 196, 199, 203, 236
North Carolina State University, 217
Kansas State University, 124
Oklahoma State University, 102, 108
Oregon State University, 145
Stroud Water Research Center, Academy of Natural Sciences, Pennsylvania, 135
Technical Resources, Inc., U.S. EPA Gulf Breeze Environ. Research Lab, 131, 225
Teeside Polytechnic Health and Safety Executive, U.K., 242
249
-------
University of California-Berkeley, 56, 232
University of California-Irvine, 91
University of Cincinnati, 242
University of Idaho-Moscow, 151, 158
University of Illinois-Urbana, 37
University of Massachusetts, 113
University of Minnesota-Duluth, 172, 176
University of Nebraska, 124
University of Nevada Environmental Research Center, 240, 246
University of Tennessee, Center for Environmental Biotechnology, 102
University of Wisconsin-La Crosse, 46
U.S. Department of Agriculture, National Biological Impact Assessment Program, 20
U.S. EPA, Corvallis Environmental Research Laboratory, 52, 56, 61, 70, 87, 145,
151,162,196,199,203,236
U.S. EPA, Duluth Environmental Research Laboratory, 169, 172, 176
U.S. EPA, Gulf Breeze Environmental Research Laboratory, 98, 118, 183, 187, 225
U.S. EPA, Health Effects Research Laboratory, 207, 211, 214, 217, 221
U.S. EPA, Office of Pesticides and Toxic Substances, 22
U.S. EPA, Office of Pesticides Programs, 31
U.S. EPA, Office of Toxic Substances, 29
U.S. EPA, Risk Reduction Engineering Laboratory, 242
Utah State University, 81
Virginia Polytechnic Institute and State University, 190
250
-------
SUBJECT INDEX
16SrRNA, 38, 46, 131
2,4-dichlorophenoxyacetate (2,4-D), 12, 145, 162
2,4,5-trichlorophenoxyacetic acid (2,4,5-T), 214
2,6 dinitrotoluene, 214
32P, 38, 48, 88, 131
Actinomycete, 141
Acyrthosipon pisum, 196
Agglutinins, 81
Alcaligenes eutrophus, 12
Algae, 135
Amblyospora, 185
Anas platyrhynchos, 203
Ancova (analysis of covariance), 180
Animal and Plant Health Inspection Service (APHIS), 26
Apis mellifera, 201
Aroclor 1254, 214
Autographa californica, 183, 203, 228
Avena sativa, 236
Azospirillum lipoferum, 140
Bacillus cereus, 53
Bacillus popilliae, 31
Bacillus subtilis, 71, 75, 233
Bacillus sphaericus, 187, 190
Baculovirus, 183, 228
Beauveria bassiana, 33, 187, 199
Bioaerosols, 61, 243
Biosafety, 242
Biotrack System, 8
Bovine Rumen Microbial Communities, 37
Brachionus plicatilis, 184
Brassica napus, 75
Butyrivibrio fibrisolvens, 38
Canadian Environmental Protection Act, 15
Carbon Cycling, 100, 151, 158, 178
Cellulomonas fimi, 135
Cellulomonas flavigena, 135
Cellulomonas uda, 135
Chaoborus (larvae), 179
Chironomus riparius, 179, 191
Cladocerans, 179
Cladophora sp., 136
Clavibacter xyli, 32
Clostridium difficile, 217
Clostridium perfringens, 49, 217
Cochliobolus heterostrophus, 125
Coleoptera, 179, 196
Colinus virginianus, 203
Colletotrichum gloeosporioides, 88, 187
Conchostraca, 174
Conjugation (bacterial), 118, 165
Coordinated Framework for Regulation of Biotechnology, 22
251
-------
Copepods, 179
Core Microcosm, 177
Crassostrea virginica, 187
Culex quinquefasciatus, 192
Daphnia sp., 32, 195
Desulfobacter, 48
Desulfobulbus, 48
Desulfococcus, 48
Desulfosarcinia, 48
Desulfovibrio, 48, 132
Diptera, 198
Dissolved Organic Carbon, 99, 135
Dytiscid (beetle larvae), 179
Edhazardia aedis, 184
Enterobacter cloacae, 12, 53
Enzyme Linked Immunoabsorbant Assay (ELISA), 12, 206
Erwinia herbicola, 233, 236
Escherichia coli, 53, 85, 108, 165, 204, 217, 225
Eubranchipoda, 179
Federal Insecticide, Fungicide and Rodenticide Act (FIFRA), 23, 31, 169
Fibrobacter sp., 38
Fungi, 87, 124, 145, 162
Gambusia affinis, 184
Gibberela fujikuroi, 124
Heliothis zea, 188
Helisoma trivolvis, 191
Hippodamia convergens, 196
Hyla crucifer, 179
Insertion sequences (IS), 113
kilA, 225
Klebsiella planticola, 75, 236
lac, 111
Lactobacillus reuteri, 217
Lagenidium giganteum, 183
Lepidoptera, 199
Lepomis macrochirus, 169
Liriodendron tulipifera, 136
Lynceus, 174
Menidia beryllina, 184, 187
Mer genes, 118
Mercury , 46, 118
Mesocosm, 135, 172, 176
Metarhizium anisopliae, 199, 203
Methylmercury, 46, 118
Mice (CD-1), 207, 211,217, 220
Mixed Flask Culture (MFC) Microcosm, 176
Mysidopsis bahia, 184, 187
National Biological Impact Assessment Program (USDA), 20
National Ecological Field Research Station, 240
National Environmental Research Institute (of Denmark) NERI, 10
Nematodes, 147
Neurospora crassa, 87, 125
Nitrogen Cycling, 149, 178
Nomureae rilryi, 200
252
-------
Nosema algerae, 184
Nosema cuneatum, 183
Nosema locustae, 183
Oncorhynchus mykiss, 169
Organization for Economic cooperation and Development (OECD), 7
Ostracods, 179
Oxyrase, 217
Palaemonetes kodiakensis, 184
Palaemonetes pugio, 184
Paragnetina media, 192
Parathelohania, 185
Pentachlorophenol (PCP), 214
Phaseolus vulgaris, 71, 75, 236
Phosphorus Cycling, 149, 178
Pimephales promelas, 169
Planoribidae, 174
Plasmids (benchmark), 92
Polymerase Chain Reaction (PCR), 14, 37
Primary Production, 178
Protozoa, 147
Pseudomonas aeruginosa, 49, 59, 102, 108, 118, 164, 211, 220, 225
Pseudomonas cepacia, 113, 211
Pseudomonas fluorescens, 85, 164, 187, 196
Pseudomonas putida, 53, 81, 145, 162
Pseudomonas syringae, 57, 69, 70, 75, 232, 236
Pteronarcys proteus, 194
Quality assurance, 246
Rats (Fisher 344, Sprague Dawley), 207, 214
Recombinant DNA processing facilities, 242
Ruminococcus, 38
Salmonella pullorum, 203
Salmonella typhimurium, 214
Salvelinus fontinalis, 169
Scrimweave, 172, 177
Streptomyces viridosporus, 152, 158
Streptomyces lividans, 140, 151, 158
Siilfate-reducing bacteria (SRB), 37, 46, 131
Tachinidae, 198
Tipula abdominalis, 192
Toxic Substances Control Act (TSCA), 23
Transduction, 102
Transformation, 127
Trichogramma pretiosum, 199
Trichoplusia ni, 199, 229
Triticum aestivum, 140
UV (stress), 110
Vectobac, 169, 172, 177
Vegetative compatibility group (VGC), 124
Worker exposure, 242
Xanthomonas maltiphilia, 234
xylE, 56
Zooplankton, 100, 173, 179
253
*U.S.COVERNMEVTPRINTINGOFFICE:1992 -ste -003/60061
-------
-------
-------
m
-o
£
g
CD
n
€/» -n O
W CD g
O ffl £>!
^SL
3-i1
~" CO
"05"
o o m c
5' CD 3 3.
CO
CD
O> 3 en
00 CD fi
33 >
CD CQ
« CD
&
o'
32 SL«<
III
So*
§.:§£
8g*
3 » o
EQ.T3
' P5"
°- S 3
, lit
S S>< B
w r- -.. —
0)
------- |