United States
Environmental Protection
Agency
Environmental Research
Laboratory
Corvalhs OR 97333
Research and Development EPA/600/9-90/029
Environmental Research
Laboratory
Gulf Breeze FL 32561
July 1990
Review of
Progress in the
Biotechnology-
Mi crobia I Pest Control Agent
Risk Assessment Program
-------
EPA/600/9-90/029
July 1990
REVIEW OF PROGRESS IN THE BIOTECHNOLOGY-
MICROBIAL PEST CONTROL AGENT RISK ASSESSMENT PROGRAM
Prepared by
ORD Biotechnology Risk Assessment Program
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97333
GULF BREEZE, FLORIDA 32561
U S Environmental Protection Agency
Region 5, Library (PL42J)
77 West Jackson Boulevard, 12tn rioor
Chicago, tL 60604-3590
-------
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.
n
-------
TABLE OF CONTENTS
Acknowledgement iv
Preface v
Introduction 1
Scientific Abstracts
Session I Special Presentations 11
Session II Detection/Enumeration 19
Session III Dispersal/Transport 33
Session IV Gene Exchange 55
Session V Survival/Colonization ... 95
Session VI Environmental Effects 131
Session VII Higher Organism Effects 169
Session VIII Risk Control 195
Session IX Human Health 203
Index of Contributors 223
m
-------
ACKNOWLEDGEMENT
This publication is a result of the patience and cooperation of many who
prepared and edited this text. We give special thanks to Dr. Ray Seidler,
Corvallis Environmental Research Laboratory Biotechnology Team Leader, who
organized, coordinated, and implemented the all investigators' meeting and the
compilation of this report. We wish to acknowledge the efforts of Dr. Robert
Frederick, Office of Environmental Processes and Effects Research, EPA Head-
quarters; Dr. Richard Anderson, Duluth Environmental Research Laboratory
Biotechnology Team Leader; and Dr. Stephen Cuskey, Gulf Breeze Environmental
Research Laboratory Acting Matrix Manager. We thank Kiki Alexander, Nancy
Lanpheare, Valerie Prince, Brenda Shaffer, Ann Hairston, and Kelly Donegan of
NSI Technology Services Corporation for typing, proofing, and editing camera
ready copy. We are especially appreciative of the reviewers who, through two
and one-half days, listened to and evaluated some 50 technical presentations.
The reviewers were Drs. Ellis Kline, Jay Grimes, Marvin Rogul, Phyllis Martin,
and Peter Hartel.
A list of relevant EPA-sponsored publications is appended to each
abstract. In the interest of space, the huge numbers of abstracts, presenta-
tions, and early publications have often been trimmed. We apologize to all
authors who were affected by this decision.
We wish to dedicate this document to the memory of Dr. Stephen Cuskey
whose untimely passing affected us deeply and whose contributions to this
program were numerous.
IV
-------
PREFACE
When biotechnology research proliferated in the 1970s, the National
Institutes of Health initially assumed the role of regulation of the medical
applications that dominated the early discoveries. However, as commercial
applications diversified, an interagency unit was established under the
Council of Natural Resources and the Environment to implement biotechnological
regulation. In 1984, a coordinated regulatory framework was developed, and
EPA became responsible under the Toxic Substances Control Act (TSCA) and the
Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) for regulating
environmental biotechnology products and evaluating the environmental impact
of the deliberate or accidental release of Genetically Engineered Microor-
ganisms (GEMS).
In November 14-16, 1989, all investigators of research projects funded
under the EPA Office of Research and Development (ORD) Biotechnology Risk
Assessment Program assembled at the U.S. EPA Corvallis Environmental Research
Laboratory for their third review of the state of research, unresolved issues,
and future directions. Expanded abstracts presented at the meeting form the
basis of this report. The meeting provided a forum for the peer review of
research performed by EPA scientists and extramural cooperators supported by
the Program. A panel of scientific peers reviewed and offered analysis of the
EPA-funded research, relevancy of research emphasis, and adequacy of research
facilities. Their summary comments are included in this document. Fifty-
three abstracts were presented during nine sessions that covered: development
of recombinant microorganisms, detection and enumeration, dispersal and
transport, gene exchange, effects of higher organisms, survival and coloniza-
tion, molecular, physiological, and ecological approaches to the determination
of environmental effects of GEMS, and risk control.
Investigators within the program discussed development and refinement of
a wide variety of test organisms, genetic constructs, and methods. In
noteworthy developments since the last all investigators' meeting, researchers
reported significant progress in gene transfer and in use of environmental
simulators (microcosms) for risk assessments. Further, research results
indicated for the first time that a GEM can induce non-transient changes on
microbe-mediated ecological processes in non-sterilized soil. The successful
completion of protocols to evaluate the impact of bacterial MPCAs on benefi-
cial aquatic insects also was reported.
-------
ORD BIOTECHNOLOGY RISK ASSESSMENT PROGRAM
The Biotechnology/MPCA Risk Assessment Research Program is supported by
six laboratories and an EPA Headquarters liaison. These laboratories and
team leaders are listed below:
Dr. P. Hap Pritchard (Matrix Manager)
Environmental Research Laboratory/ORD
Sabine Island
Gulf Breeze, FL 32561
Dr. Robert Frederick (Headquarters Liaison)
Office Environmental Processes Effects Research
U.S. EPA
401 M Street, SW
Washington, DC 20460
Dr. Richard Anderson
Environmental Research Laboratory/ORD
6201 Congdon Blvd.
Duluth, MN 55804
Dr. Ramon Seidler
Environmental Research Laboratory/ORD
200 S.W. 35th Street
Corvallis, OR 97333
Dr. Steve Hern
Environmental Monitoring Systems Laboratory/ORD
Las Vegas, NV 89114
Dr. John Burkle
Risk Reduction Engineering Laboratory/ORD
Cincinnati, OH 45268
Dr. Larry Claxton
Health Effects Research Laboratory/ORD
Research Triangle Park, NC 27711
-------
INTRODUCTION
Background of EPA's Research Program
In the 1970s biotechnology became popularized with news that scientists
had developed techniques for combining fragments of DNA from different
organisms. These new techniques are significantly different from the directed
evolution of crops and animal husbandry previously practiced for centuries.
It is now possible to alter the genetic composition of organisms in a precise
and purposeful fashion. A new science of genetic engineering began from these
techniques and has led to a major new industry, biotechnology. Applications
of biotechnology are going beyond contained research laboratories, but the
impact of genetic engineering on commercialization of recombinant microbes is
in its infancy.
Seven environmental releases of genetically engineered microorganisms
(GEMs) have taken place in the U.S. The agents have included bacteria and one
virus. Many of these releases have involved microbial pest control agents
while two dealt with enhanced nitrogen fixation on legumes and use of a marker
gene for evaluating microbial detection, survival, and dispersal. The
magnitude of the potential number of products being developed can only be
estimated from the more than 3,000 patents that have been issued involving
microbes in the period 1980-84. Furthermore, the patent office has a backlog
of several thousand additional biotechnology patent applications not included
in the above number.
Initially, the National Institutes of Health assumed the role of regula-
tion of biotechnology research in the 1970s, since most of,-the early discov-
eries dealt with medical applications. However, as commercial applications
diversified, another mechanism for providing regulation was implemented. In
1984 an interagency group under the White House Cabinet Council on Natural
Resources and the Environment was formed. A coordinated regulatory framework
for biotechnology products was developed and EPA assumed responsibility under
the Toxic Substances Control Act (TSCA) and the Federal Insecticide,
Fungicide, and Rodenticide Act (FIFRA) for regulating environmental release of
microbial pest control agents and recombinant microorganisms.
Evaluating the impact of deliberate or accidental releases of GEMs
presents a high degree of uncertainty compared to chemicals. Regulators were
familiar with the fate and effects of chemicals. However, microbial biotech-
nology products have dynamic properties not predictable by chemical fate and
effects models, i.e., replication, mutation, gene transfer. Thus, certain
issues were targeted early for evaluation as part of the risk assessment
process for GEMs:
• Non-target effects of an introduced pathogen
• Extent of genetic exchange between GEMs and indigenous organisms
• Possible disruption of biochemical processes such as nutrient cycling
t Long-term effects via evolutionary processes
1
-------
In 1985 EPA's Office of Research and Development (ORD) implemented an
integrated Biotechnology Risk Assessment Research Program responsive to the
Agency needs. The program develops methods from research for risk assessment
in these areas: environmental exposure, environmental effects, and risk
control of GEMs and microbial pest control agents (MPCAs) intended for
environmental release (or having a high probability of release, either
deliberate or accidental). Research groups from ORD laboratories at
Corvallis, Duluth, Gulf Breeze, Cincinnati, and Research Triangle Park meet
annually to develop project-level research objectives. The objectives
consider emerging research issues, new experimental methods, and input from
peer and program reviews. The laboratory research coordinators identify
Agency needs and formulate funding mechanisms to accomplish the research.
The biotechnology risk assessment research program can be divided into
several areas based on its overall mission and objective of the program.
These areas are as follows:
t Development of methods for the detection and enumeration of GEMs and
MPCAs
t Identification of factors that regulate microbial survival and coloniza-
tion in the environment
• Assessment of the stability of the recombinant DNA and its potential for
transfer to indigenous microorganisms
• Determination of factors involved in the dispersal and transport of
microorganisms
• Development of predictive models for dispersal
t Development of methods for the detection of environmental perturbations
(ecological effects, pathogenicity) resulting from the introduction of
microorganisms
• Development of methods to control, mitigate, and decontaminate GEM
releases in an environmental context.
The Research Program
The research program seeks to characterize risk through methods develop-
ment, analyses, and mathematical modeling. Investigations of three broad
areas of the research program considered significant to EPA's regulatory
functions are: environmental exposure, environmental effects, and risk
control.
Environmental Exposure Studies
Detection and Enumeration. Scientists are using serological,
biochemical, physiological, and genetic methods for improving detection and
enumeration of microorganisms under laboratory, microcosm, and field condi-
tions. Methods are being developed and refined for their sensitivity,
-------
specificity, and reliability in a variety of terrestrial, aquatic, and air
habitats. Certain plants, animals, and aquatic systems are of special
interest.
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.
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 animals) are being
studied. This research includes development of mathematical models and other
aids for predicting transport and exposure.
Gene Exchange. 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.
Environmental Effects Studies
Ecological Effects. Researchers are evaluating the effects of introduced
microorganisms on structural/functional aspects of the ecosystem. Improve-
ments in the procedures that are developed in these ecological investigations
may be incorporated into new testing methods.
Higher Organism Effects. A wide variety of microorganisms may be used in
biotechnology and information on their toxicological, behavioral, pathogenic,
and histopathological effects on higher organisms is unavailable. This
research investigates the genetic and molecular bases of infectivity, patho-
genicity, 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 verte-
brate species.
Risk Control Studies
Field Release. ORD scientists develop criteria for evaluating contain-
ment and monitoring strategies. Test methods are based on the actual applica-
tion of recombinant and surrogate organisms to specially designed experimental
release sites representing different environmental habitats. Methods for
detection, for monitoring distribution, dispersion, and dislodging charac-
teristics, and for proper containment of microorganisms during field releases
are being evaluated.
Control. 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 condi-
tions.
The biotechnology risk assessment research program has just completed its
fifth year. The program has evolved significantly each year in emphasis and
scope. In its early stage, research emphasized environmental exposure
components that examined conventional and molecular technologies for detection
and enumeration. Significant improvements in bacterial detection techniques
were achieved. During the second and third years, investigators initiated
studies to evaluate the use of laboratory microcosms in GEM survival, coloni-
zation, and genetic exchange. In the third and fourth years, studies were
initiated to develop methods for evaluating ecological effects induced by
GEMs. More recently, research has demonstrated that mathematical models can
be used to predict spatial and temporal parameter effects on GEM locations.
Peer Reviews
Since the inception of the biotechnology risk assessment program in 1985,
peer reviews have been held almost yearly. These activities insure the
maintenance of the highest scientific quality and integrity. These meetings
have also provided a valuable forum for exchange of ideas for scientists and
regulators, both within the United States and abroad.
An all investigators meeting coupled with an official peer review of the
biotechnology program was conducted in May 1987. The peer review panel was
comprised of seven prominent scientists in the areas of microbial genetics,
biotechnology, molecular biology, microbial ecology, microbial pest control
agents, and terrestrial and aquatic microbiology. In October 1988, and again
in November 1989, all investigators meetings were held. This document
presents Abstracts of presentations given at the last meeting held in
Corvallis, Oregon, November 14-16, 1989. Over 90 individuals representing
five countries participated in this meeting. The goals of the meeting, as in
previous all investigators meetings, were two-fold. The meeting provided a
forum for discussions among scientists for evaluating the status of each
other's work and for discussing topics for future efforts. The meeting also
provided a forum for conducting a peer review of the Agency's Biotechnology
Risk Assessment Program regarding progress, relevancy of major research
topics, and an evaluation of in-house facilities for conducting this research.
Fifty-nine presentations were given over the three-day period. The meetings
were divided into nine sessions:
Session I. Special Presentations
Five speakers representing regulatory agencies in the United States and
Canada presented their views on the development of recombinant microorganisms
and the status of regulatory guidelines. Two of these speakers provided
Abstracts.
-------
Session II. Detection/Enumeration
Four presentations illustrated the importance of viable/non-culturable
microorganisms, difficulties in the enumeration of aerosolized microorganisms,
and the importance of DNA probes in detecting recombinant sequences.
Session III. Dispersal/Transport
Eight presentations featured the modeling of airborne microorganisms and
modeling the fate of bacteria in surface waters.
Session IV. Gene Exchange
Eight presentations described gene transfer mechanisms of bacteria in
fresh and marine waters, soil and plants, as well as gene transfer mechanisms
in fungi.
Session V. Survival/Colonization
Nine presentations were delivered on diverse ecosystem test systems
involving the colonization of marine, freshwater, plants, and soil habitats by
GEMs.
Session VI. Environmental Effects
Nine reports described molecular, physiological and ecological approaches
to evaluating any potential adverse environmental effect caused by GEMs.
Session VII. Higher Organism Effects
Eight reports described safety protocols for MPCA effects on beneficial
insects, avian species, aquatic invertebrates, and vertebrate organisms.
Session VIII. Risk Control
Two reports were presented. One dealt with the ability of conditionally
lethal genes to limit GEM survival. The second presented results on the
control of microorganisms under field conditions by physical and chemical
treatments.
Session IX. Human Health
Six reports that dealt with the survival of and potential adverse human
health effects caused by GEMs in several model and test systems were
presented.
-------
These presentations have been encapsulated as abstracts in the body of
this document. Observations directly applicable to risk assessment of GEMs
presented in the abstracts include:
t concerns over viable but non-culturable recombinant microbes
t the first demonstration of long-term ecological effects from a GEM
metabolite that is lethal for certain soil fungi
t the demonstration of gene exchange by transduction and transformation,
involving recombinant DNA in the environment
• the successful completion of protocols to evaluate the impact of
bacterial MPCAs on beneficial aquatic insects
Summary of Panel Review Comments
The 1989 All Investigators Meeting provided a unique opportunity to
obtain an overview of EPA-supported research relating to Biotechnology Risk
Assessment. The meeting prompted a lively exchange of experiences and ideas,
and it encouraged interaction among the participants from within and from
outside the EPA. The meeting enabled those present to learn not only of the
status of studies relating to the development of biotechnology risk assessment
in this country, but also the status of similar programs in other countries,
through participation by scientists from Britain, Canada, Belgium, and Japan.
An introductory presentation series of five speakers (USA and non-USA) gave
the audience an up-to-date view of the rapid expansion that is taking place in
the environmental applications of biotechnology. Some of the key points
addressed were: a) the need for protocols to extrapolate from microcosms to
field tests, both large- and small-scale; b) the need for a better data base
for proper modeling and predictability; and c) a discussion of the baselines
being considered for protocols/guidelines for release in the agricultural
environment of genetically engineered microbes, plants, and animals.
The panel reviewed research by the Office of Environmental Processes and
Effects laboratories at Corvallis, Duluth, and Gulf Breeze and identified the
following as highlights of the presentations:
t The demonstrated efficacy of several DNA extraction/restriction
digestion/DNA probe hybridization protocols indicates the procedures are
closer to a usable methodology in environmental risk assessment.
• It has been established that the condition of viable but non-culturable
bacteria is generally applicable to and may impact the recovery of GEMs.
t The EPA ORD research program is to be complimented for encouraging new
approaches to design and- analysis of aerosols and microbial dispersal.
These investigations should continue.
t The research which first demonstrated that bacteria can become genetic-
ally transformed in the environment was sponsored by EPA.
-------
• The use of naturally-occurring plasmids with selectable markers has been
key to the analysis of gene transfer In the environment.
t Virus-mediated exchange of genetic material in a natural system has been
verified through EPA-sponsored research.
• EPA should continue its development of test protocols for pathogenic
microbial pest control agents.
• Specific genes have been discovered in bacteria that control colonization
of plant surfaces.
• Environmental effects of GEMs are subtle, unpredictable, and difficult to
measure.
t The use of suicide genes offers great promise for containment of biologi-
cal agents.
t Physical and chemical techniques for containment of biological agents
have considerable potential.
Panel Recommendations
• It is important to evaluate the potential of GEMs to become viable-non-
culturable.
• Genetic fingerprinting (RFLPs) should be further evaluated as a way to
characterize changes in communities.
• Further study is needed to establish a rapid, less labor-intensive method
for direct DMA extraction from soil and water microbes.
• There are too few studies dealing with transport effects involving fungi.
It is important for this work to be done with organisms that have a high
probability for release.
• It is important to correlate microcosm research with field calibrations.
• The effect of environmental conditions on microbial survival in aerosols
needs to be continued.
• It is important to know more about the biological activity of cell-free
DMA in the environment.
• Verification of viral-mediated gene transfer should continue using
microbial strains indigenous to the environment.
t EPA should continue development of safety protocols for MPCAs involving
representatives from the major arthropod groups, with emphasis on insects
since they are the most common host for MPCAs.
-------
• Microcosms should be developed as multiple-species tests for safety
testing of MPCAs.
• EPA needs to establish a cooperative association with ornithologists,
aerobiologists, entomologists, ichthyologists, and agronomists to
determine appropriate test animals.
• Investigations defining factors produced by plant colonizing microbes may
lead to improved understanding of microbial/plant association in the
environment.
• Research on the effects of MPCAs on mycorrhizal fungi, N2-fixers,
protozoan grazers, and predator-prey should be conducted.
• Use of agronomically important microbial associations in effects studies
should be encouraged.
FACILITIES REVIEW
The panel met with team leaders from three EPA laboratories (Corvallis,
Duluth, and Gulf Breeze) to discuss resources for research at each of these
locations. Included in these discussions were considerations of space,
equipment, and professional and support personnel. The present condition at
each laboratory was described, plans for the future were presented, and
requirements for implementation of these plants were reviewed. The panel took
time during the meeting to tour the EPA facilities at Corvallis, and on other
occasions members of the panel have visited in an official capacity the
laboratories at Gulf Breeze and Duluth.
The panel is of the opinion that the Corvallis, Gulf Breeze, and Duluth
laboratories have made considerable progress in establishing scientifically
credible and productive programs of research that relate directly to the EPA
Biotechnology/MPCA Risk Assessment mission. However, these programs appear to
have expanded beyond the limits of available resources. Program growth at
each of the three locations has been in response to ORD, OEPER, and program
office needs to generate information needed for decisions pertaining to the
products of biotechnology. Continued progress in these laboratories is
essential for the EPA to meet its responsibilities in an area that must have
attention, especially if this country is to compete successfully in a world
that recognizes biotechnology as important to economic development.
-------
SCIENTIFIC ABSTRACTS
-------
-------
SESSION I
SPECIAL PRESENTATIONS
11
-------
-------
RESEARCH NEEDS BASED ON INDUSTRIAL DEVELOPMENT
Morris A. Levin
U.S. EPA and University of Maryland
College Park, MD 20742
This paper is divided into two parts: (1) a description of the growth
and expectations of the biotechnology industry in terms of possible effect on
EPA research needs, including a survey of releases to date; and (2) an
estimate of research needs based on discussions with personnel in Office of
Toxic Substances (OTS), Office of Pesticide Programs (OPP), U.S. Department of
Agriculture (USDA), and biotechnology companies.
Biotechnology Industry: Growth and Expectations
Table 1 is a summary of a survey of biotechnology patents issued between
1980 and 1984 which could result in applications to EPA. Over 3000 patents
were issued during the five-year period (during which time the patent office
had a 7000-patent backlog which was not included in the survey). The impact
of these patents on the Agency is just beginning.
Figure 1 depicts the first impact, showing the number of reviews
conducted at EPA from 1983 through the first half of 1989. The last point was
estimated by averaging data for all years and estimating for the period May-
December. The rate at which applications are being received is increasing
exponentially. Table 2 identifies the reviews conducted by OPP. As can be
seen, 42% of the reviews involved Bacillus thurinqiensis, either as host or
donor. Actual recombinant DNA technology was involved in the fewest cases.
OTS, under TSCA, reviewed 14 strains of Rhizobium or Bradvrhizobium.
In addition, EPA has collaborated on the review of 23 engineered plants
with USDA, mostly tobacco and tomato. Interestingly, the Canadian government
has authorized 39 field trials with engineered plants, mostly canola and flax.
Research Needs
Discussions with OPP and OTS staff as well as industry and academic
scientists resulted in agreement as to which areas of research were most
needed. These fall into five general categories involving research on:
genetic issues, phenotype determinations, competitiveness issues, and impact
of microbes on environment and one very specific area involving methodology --
use and value of microcosms.
Genetic issues include investigating the role of mobile genetic elements,
the value and appropriateness of using antibiotic resistance genes as markers,
and stability of the recombinant genome in the environment.
The stability issue crosses into the general topic of phenotypic deter-
mination. Research issues raised in this topic area include defining the
impact of environmental parameters on expression of introduced genetic
material, estimating environmental impact on genetic transfer (rates, proba-
13
-------
bility of occurrence), probability of expression of transferred genes in other
genera or species, and the probability of pleiotropic effects.
There was intense interest in developing methodology, estimating competi-
tive ability, and understanding the role of environmental parameters in
affecting the outcome of competitive situations. The need for easily
available and well organized background ecological data was repeatedly
mentioned, especially in terms of the need to clearly identify the host and
donor microbes.
These needs were re-emphasized in a committee report on mobile genetic
elements to EPA's Biotechnology Science Advisory Committee at a meeting on
December 20, 1989.
The need for a better understanding of the role of environmental para-
meters on the survival and persistence of the recombinant organism was
stressed repeatedly. The lack of ability to clearly identify effects which
could be monitored and quantified was also stressed.
Common to the last three issues and stressed heavily in all discussions
is the need for research in the area of microcosms and growth chambers in
terms of relating data between different types of microcosms and relating
micro- and mesocosm data to field situations. The issue of using small-scale
field data to predict the effects of large-scale applications can be seen as
an extension of this need.
14
-------
Table 1. Microbial Patent Search*
Number of
Area Genera
Agricultural chemistry 8
Conversion of biomass 14
Industrial chemical production 14
Monitoring/measurement/biosensor 7
Energy 4
Polymer/macromolecule production 16
Enhanced oil recovery 9
Waste/pollutant degradation 17
Mining/metal recovery 10
* 1980-1984.
Table 2. GEM Reviews
Host Source/Basis
B. thurinqiensis Other B.t.
Mutagenesis
Fusion
T. harzanium Fusion
Mutagenesis
P. fluorescens Deletion
B.t.
P. svringae Deletion
B.t.
C. xvlii B.t.
C. gloeosporiodes Mutation
I. viridae Mutation
P. aureofaciens E. coli
Baculovirus Deletion
Patents
37
809
528
36
217
787
25
541
80
Applications
10
5
3
3
3
2
1
1
1
15
-------
o»
48
42
36
30
24
18
12
6
0
REVIEWS
03
c
n>
19S3
19S4
1985 1986
YEAR
1987
1988
1989
-------
UPDATE ON USDA RESEARCH GUIDELINES FOR GEMS, PLANTS, AND ANIMALS
Daniel D. Jones
Office of Agricultural Biotechnology
U.S. Department of Agriculture
Washington, D.C. 20460
The U.S. Department of Agriculture is in the process of developing
proposed Guidelines for Research with Genetically Modified Organisms Outside
Contained Facilities. The USDA Guidelines are based in general on the concept
of the National Institutes of Health (NIH) Guidelines for Research Involving
Recombinant DNA Molecules, but they will incorporate features unique to
agricultural research. The USDA Guidelines will contain provisions for
agricultural research on genetically engineered microorganisms (GEMs), plants,
and animals.
The primary source of scientific input for the USDA Guidelines is a
Federal advisory committee called the Agricultural Biotechnology Research
Advisory Committee (ABRAC). The ABRAC currently consists of 7 academic
scientists, 3 private sector scientists, 1 government scientist, 1
scientist/attorney, and 1 public interest group representative.
It is proposed that the conditions under which research with a genetic-
ally modified organism can be conducted safely should be assessed relative to
the conditions required for safely conducting research with the unmodified or
wild-type organism. USDA, upon recommendation by the ABRAC, has.developed a
proposed process under the Guidelines for determining experimental conditions
for the safe performance of agricultural research involving genetically
modified organisms.
The steps in the proposed process are: a) determine the level
of safety concern for the unmodified organism; b) determine the effect of the
genetic modification on safety (i.e., whether it increases, decreases, or has
no effect on safety); c) determine the level of safety concern for the
modified organism; d) determine the confinement level appropriate to the
particular level of safety concern for the modified organism and develop a
safety protocol to meet this level of confinement; and e) determine the
appropriate organizational level of safety review.
The level of safety concern referred to above currently ranges over a
scale from 1 to 5. Level 1 includes organisms whose ecological attributes in
the accessible environment are well understood, and for which it can be
determined that the unmodified organism has virtually no potential for adverse
effects on human health or on managed or natural ecosystems. Level 5 includes
organisms whose ecological attributes may cause predictably high adverse
effects on human health or on managed or natural ecosystems and for which no
feasible types of confinement will allow safe conduct of research outside
contained facilities. Levels 2, 3, and 4 are then arrayed between these two
extremes.
17
-------
The ABRAC has formed a Working Group on the Classification of Unmodified
Organisms. The charge of the Working Group was to develop a science-based
framework for Step 1 above for classifying unmodified organisms according to
risk that is flexible enough to change as new information on the risk of
specific organisms is obtained.
The Working Group developed a generic framework for classifying organisms
according to safety. The steps in the process are: a) describe the access-
ible environment; b) describe the relevant attributes of the organism in the
accessible environment; c) evaluate the relative importance of specific
attributes. Explain why some attributes are more or less important in the
context of the proposed test; and d) select a level of safety concern for the
unmodified organism. Provide the rationale for assigning the organism to a
level of safety concern from 1 through 5.
The Working Group identified attributes of organisms that should be
considered in determining levels of safety concern. These included
pest/pathogen status, potential of the organism to establish itself in the
accessible environment, ecological relationships with other organisms,
potential for introduction of genetic change in natural or managed popula-
tions, and potential for control of the organism.
As a pilot exercise, members of the Working Group developed 12 detailed
examples of how the above-described process might be applied to plants,
animals, and microbiological agents. The microbiological examples included
Pseudomonas fluorescens 2-79, Butvrivibrio fibrisolvens. soybean mosaic virus,
and foot and mouth disease virus. Using this process, the Working Group
members tentatively assigned B. fibrisolvens to Level of Safety Concern 1, P.
fluorescens 2-79 and soybean mosaic virus to Level 2, and foot and mouth
disease virus to Level 5.
Current plans are to include a description of this organism classifica-
tion process in a handbook for agricultural researchers entitled Agricultural
Biotechnology: Introduction to Field Testing to be published in the near
future. The handbook will contain chapters on the field testing responsi-
bilities of principal investigators, institutions, and USDA, as well as the
regulatory requirements of the various Federal agencies which may relate to
field testing. In addition, it will contain informational chapters on the
National Environmental Policy Act, confinement of genetically modified
organisms, public relations, and societal issues. The organism classification
process is intended to be flexible enough to evolve as scientific information
on field testing genetically modified organisms grows.
18
-------
SESSION II
DETECTION/ENUMERATION
19
-------
-------
VIABLE BUT NON-CULTURAL PHENOMENON IN RELATIONSHIP TO STARVA-
TION/SURVIVAL /INJURY," AND STRATEGIES FOR SURVIVAL OF
BACTERIA IN THE ENVIRONMENT
R. R. Colwell,1'2 I. T. Knight,1'2 C. Somerville,1'2 B. Conde,2
and J. Byrd2
Department of Microbiology1
University of Maryland
College Park, MO 20742
Center of Marine Biotechnology
University of Maryland
Baltimore, MD 21202
INTRODUCTION
One of the major limitations to research in microbial ecology has been
the inability to isolate and grow in culture the vast majority of bacteria
which occur in nature. The occurrence of non-culturable bacteria has long
been known, but the nature of the phenomenon has not been determined. Colwell
et al. reported that some pathogenic bacteria lost the ability to grow on
laboratory media after incubation in oligotrophic ocean water or microcosms
for short periods of time (less than 1 day to 3 weeks), while cell numbers
changed little by direct microscopic counts. The implication of these
observations is that pathogens surviving in the environment may not be
detected by standard methods. Grimes and Colwell investigated the survival of
pathogenic Escherichia coli in membrane chambers submerged in semitropical
ocean water, and reported that viable cells could be detected by the direct
viable count (DVC) method long after they could be cultured on laboratory
media. Introduction into a ligated rabbit ileal loop led to the recovery of
cells which grew in culture and demonstrated plasmid profiles identical to
those of the original inocula. Results of this study suggested that water-
borne pathogens which elude detection in the laboratory may retain their
pathogenicity, and may be "revived" to the culturable state by animal passage.
Tamplin and Colwell evaluated the effect of salinity and organic concentration
on the production of enterotoxin by Vibrio cholerae in microcosms. Results
showed that at both high (10 g/1) and low (1 g/1) concentrations of organic
nutrient, toxin production increased with salinity, in the range of 0 to 25
ppt. Thus, bacteria not only survive exposure to the marine environment,
previously believed to lead to rapid die-off, but retain important properties,
including potential pathogenicity.
It is possible to detect bacteria in a habitat by the presence of
specific gene sequences. In fact, the ability of aquatic bacteria to transfer
genetic information horizontally implies that important genetic traits may be
propagated in the environment even after the demise of the original host.
Therefore, tasks such as tracking genetically engineered microorganisms (GEMs)
in the ocean should not be by cell detection alone, but should be supplemented
by detection of the relevant genes. We have recently developed a procedure by
which total cellular RNA and DNA, including plasmid DNA, can be isolated from
relatively large aquatic samples without the need for cell culture. Nucleic
acids purified by this method are suitable for targets of DNA/DNA or DNA/RNA
21
-------
probes, restriction digestion and sub-cloning procedures, or templates for
direct sequence determination. Isolation of ribosomal RNAs by this method
make it possible to identify the bacterial component of an aquatic habitat
without growing the cells, and the isolation of mRNA may make it possible not
only to detect species present, but the genes being expressed. Direct
isolation of nucleic acids from environmental samples is complemented by
molecular characterization of ecologically important genes.
Detection of specific gene sequences in the environment by DMA probes
may be complicated by the relative small number of cells of the parent
organism and large volume of water, with the result that the gene may be
present in very low concentration. We have attempted to overcome these
difficulties by developing a probe method which uses the sensitivity of DNA
oligomers (approximately 20-30 bp) and adds a long tail of radiolabeled or
otherwise tagged DNA to enhance detection.
Another aspect of monitoring bacteria in the environment is the
detection of culturable organisms which occur in such low numbers that their
presence may be missed by standard microbiological tests. This points to
another weakness of standard bacteriological assays, and that is the small
sample size which can be evaluated. We have attempted to avoid the
limitations of sample size by concentrating the biomass from relatively large
samples of seawater onto Sterivex (Millipore) membrane filters. This type of
filter serves as a convenient culture vessel to which selective or enrichment
media can be added, allowing culturable cells which are present in very low
numbers to grow to large cell numbers. These tests are not only more
sensitive, but are rapid and simple, and will improve detection of indicator
organisms in the environment.
Direct Detection of Salmonella in Estuaries Using a DNA Probe
Salmonella is a prime example of a water and shellfish transmitted
pathogen which is difficult to culture from environmental samples, although
many different culture media and enrichment regimes have been proposed. One
of the factors contributing to this difficulty is the ability of Salmonella
spp. to enter a viable but non-culturable state after lengthy exposure to
river and seawater, under ambient conditions of temperature and low nutrient
concentration. Other investigators have described "sublethal injury" subse-
quent to exposure to the aquatic environment, when the organism remains
culturable but the efficiency of recovery in culture is decreased. Culture
methods for detection and enumeration of Salmonella spp. in aquatic systems
are not only unreliable, but tedious as well, requiring incubation in many
enrichment media for several days before presumptive salmonellae can be
isolated.
DNA probes specific for Salmonella spp. have been developed by several
workers. There is a commercially available probe (Gene-Trak Inc., Framingham,
Massachusetts) in a kit for detecting Salmonella spp. in food. Advantages of
DNA probe assays and immunoassays include their high specificity and rapid
assay time, reducing the total time for positive identification from several
days to 1-1/2 days. Direct application of the available DNA probe methods to
analysis of environmental water samples yield only mixed success because the
22
-------
numbers of Salmonella spp. in most waters is generally below the detection
limits of the assays, necessitating enrichment of the sample by culture.
Since the efficiency of culturing Salmonella spp. from aquatic samples is low,
detection by methods which rely upon culture will result in underestimations
of the number of salmonellae present.
As stated above, we have developed a simple, rapid method for
concentrating bacteria in water samples and extracting their nucleic acids.
The nucleic acid extracts were of sufficient quantity and purity for molecular
cloning and hybridization with DNA and RNA probes. We have applied the DNA
extraction method to obtain target DNA directly from filter-concentrated water
for hybridization with a DNA probe for Salmonella spp. Detection of
Salmonella spp. in water and treated wastewater is thereby accomplished, with
no culture procedures required.
Salmonella spp. were cultured from Sterivex-concentrated samples at five
sites and from unconcentrated samples at only one site. There was no single
broth and agar medium combination which was superior for isolating Salmonella
spp. Of the 18 isolates recovered, 10 were isolated from RVN broth, 6 were
isolated from RV broth and only two were isolated from OS broth. TSBG and
TSBG-S agars proved equally effective plating media, but only two of the 18
isolates were recovered on XN agar. The DNA probe hybridized with colony
blots of all 18 of the isolates but did not hybridize with any presumptive
Salmonella colonies which were later identified as non-salmonellae by API 20E
tests.
Probe specificity and sensitivity. The Gene-Trak probe proved to be
highly specific for the genus Salmonella under the hybridization and wash
conditions used. The probe did not cross-hybridize with any of the strains
tested but did hybridize with all the Salmonella we tested.
Dot blots of serially diluted nucleic acids, prepared from four species
of Salmonella, bound probe DNA quantitatively. The correlation coefficients
for each species and the slope of the regression line for each plot were as
follows: S. paratvphi A, 0.964 and 10.3 CPM/ng; S. arizonae. 0.998 and 14.6
CPM/ng; S. tvphimurium. 1.00 and 20.7 CPM/ng; and S. salamae. 0.999 and 15.2
CPM/ng. The mean + standard deviation of CPM probe bound per ng DNA was 15.2
+ 4.27.
The heterologous control DNA (E. coli ATCC 11303 and C. freundii ATCC
8090) on some dot blots bound probe DNA sufficiently to yield CPM levels above
background with perceptible darkening of the autoradiographs. The mean CPM
bound to the heterologous control dots and their standard deviations were used
to generate 99% and 99.9% confidence intervals for each blot, the upper limits
of which were used to determine the detection limits for the assay. If the
amount of probe bound (CPM) by the DNA sample was greater than the upper limit
of the 99.9% confidence interval of the amount bound by the heterologous
control DNA, then the sample was scored positive. If the probe bound by the
sample was less than the upper limit of the 99.9% interval but greater than
the upper limit of the 99% confidence interval, the sample was scored
plus/minus. All DNA samples which bound probe at CPM levels less than the
23
-------
upper limit of the 99% confidence interval of the heterologous controls were
scored negative.
The detection limits (i.e., the ng Salmonella DNA required for a
positive score) for each dot blot were calculated by dividing the CPM value
which defined the upper boundaries of the 99.9% confidence limit of the
heterologous control DNA by the mean CPM probe bound per ng of the S.
tvohimurium control DNA. The detection limits of the dot blots ranged from 7
to 15 ng, meaning that this was the minimum amount of the Salmonella
chromosomal DNA required for detection with the probe.
When heterologous DNA was applied to membranes in amounts approaching
the DNA binding capacity of the membrane (100 g) the amount of probe bound in
the hybridization assay did not increase (data not shown). The heterologous
control background signal was, therefore, considered independent of amount of
DNA applied to the membrane.
SUMMARY
The phenomenon of "starvation response" has been documented for marine
bacteria (ability of cells exposed to sea water, where salinity is elevated,
temperature low, and nutrient concentration minimal, to undergo reduction in
size and decrease in cell metabolism). Reduction in viable cell count and
cell size occurs. AODC and plate counts show discrepancy between cells
counted by plating and direct staining. The observation that nalidixic acid
induces cell elongation of bacteria considered "dead", caused us to pursue the
phenomenon of "viable but non-culturable cells." These cells actively
respire, taking up radio-labeled substrate, as well as demonstrate inducible
enzyme production and maintain plasmids. Experiments with animal models show
Vibrio cholerae and enteropathogenic £. coli regain culturability after animal
passage (ileal loop inoculation). Such findings have been confirmed with
human volunteers, employing a viable but non-culturable vaccine strain of
Vibrio cholerae. Data suggest many gram negative bacteria can enter the
dormant-like, viable but non-culturable state, akin to spore formation, for
survival under conditions inimical to active metabolism and replication.
Preliminary molecular genetic evidence for SpoO genes in V. cholerae, E. coli
and A. caviae has been obtained, but the experiments are not yet complete.
A method for direct detection of Salmonella spp. in water was developed
using a commercially available DNA probe. Particulate DNA was extracted from
500-1500 ml water samples collected from New York Harbor and Chesapeake Bay
and used as a substrate for a Salmonella-specific DNA probe in dot-blot
assays. The method detected salmonellae in water samples from 12 of 16 sites,
including 6 sites where salmonellae could not be cultured. The specificity of
the probe was evaluated and cross-hybridization, although negligible, was used
to set detection limits for the assay. Salmonella DNA bound the probe quanti-
tatively and from these results Salmonella DNA in the total particulate DNA in
environmental samples could be estimated. The data obtained in this study
indicate that Salmonella spp. often are not detected in water samples using
culture methodology, even when they may be present in significant numbers and
in a viable but non-culturable state.
24
-------
ANTICIPATED FUTURE WORK
The genetic basis, i.e. regulation, of the viable but non-culturable
phenomenon will be investigated employing gene probes coding for sporulation
in Bacillus and related species to determine the correlation, if any, with
"dormancy" in gram negative bacteria. Studies are also underway to determine
environmental parameters controlling non-culturability after exposure to in
situ conditions, as well as those factors associated with "recoverability,"
recently established in our laboratory for £. coli and Camovlobacter
(Helicobacter) pylori.
REPRESENTATIVE PUBLICATIONS
Col well, R.R., and D. Santavy. 1988. New Developments in Marine Biotechnol-
ogy. Proc. 8th International Conference on Global Impacts of Applied Micro-
biology (GIAM-VIII) and International Conference on Applied Biology and
Biotechnology (INCABB), Hong Kong, Aug 1-5, 1988.
Colwell, R.R. 1989. Developing a data base for the release of genetically
engineered organisms. Symposium on Environmental Issues in Biotechnology. I.
Tracking Genetically Engineered Organisms in the Environment, Division of
Environmental Chemistry 197th National American Chemical Society Meeting,
Dallas, Texas, April 11, 1989.
Knight, I.T., J.J. Byrd and R.R. Colwell. 1988. Detection of specific
genetic elements in the aquatic environment. Program Abstr. 1st Intn'l
Symposium on Marine Molecular Biology, Abstr. No. 54.
Somerville, C.C., I.T. Knight, W.L. Straube, and R.R. Colwell. 1988, A
simple, rapid method for the direct isolation of nucleic acids from the
aquatic environment. Program Abstr. 1st Intn'l Conference on Release of
Genetically Engineered Microorganisms, Abstr. No. 40, p. 20.
Somerville, C.C., I.T. Knight, W.L. Straube, and R.R. Colwell. 1988.
Probe-directed, polymerization-enhanced detection of specific gene sequences
in the environment. Program Abstract 1st Intn'l Conference on Release of
Genetically Engineered Microorganisms, Abstr. No. 70, p. 36.
Roszak, D.B., D.J. Grimes, and R.R. Colwell. 1984. Viable but non-recover-
able stage of Salmonella enteritidis in aquatic systems. Can. J. Microbiol.
30:334-338.
Roszak, D.B., and R.R. Colwell. 1987. Metabolic activity of bacterial cells
enumerated by direct viable count. Appl Environ. Microbiol. 53:2889-2983.
Roszak, D.B., and R.R. Colwell. 1987. Survival strategies of bacteria in the
natural environment. Microbiological Rev. 51:365-379.
Somerville, C.C., I.T. Knight, W.L. Straube, and R.R. Colwell. 1989. Simple,
rapid method for direct isolation of nucleic acids from aquatic environments.
Appl. Environ. Microbiol. 55:548-554.
25
-------
RECOVERY OF BULK DNA FROM SOIL USING A
RAPID, SMALL-SCALE EXTRACTION METHOD
J.L. Armstrong and L.A. Porteous
U.S. EPA
Environmental Research Laboratory
Corvallis, OR 97333
INTRODUCTION
The distribution and persistence of recombinant DNA in the terrestrial
environment has stimulated molecular ecologists to develop techniques for
extracting DNA from soil. One approach relies on recovery of intact bacteria
from the soil before the DNA is extracted. Another approach involves the
direct-extraction of DNA without prior removal of the cells from soil. We
describe a new direct-extraction method that is small-scale, rapid, and
simple, which yields up to 200 ug of DNA from a gram of soil.
METHODS
Soil samples were collected near Corvallis, Oregon: under an oak tree;
in a sheep pen; from the rhizosphere under grass; from the rhizosphere under
clover. 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-Cl, pH 8) and vortex-mixed
for one minute at 20-25°C. Then, 500 mg of polyvinylpolypyrrolidone was
added. Lysis of bacteria was promoted with 100 ul lysozyms solution (50
mg/ml) and fungal cells walls were degraded with 120 ul Novozym 234 solution
(50 mg/ml; Novo Biolabs, Bagsvaerd, Denmark). The sample was vortex-mixed at
20-255C 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 ug/ml, and 50 mM tris-Cl, pH 8) were
added. The tube was slowly inverted to mix the contents and returned to ice
for 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 washed twice by centrifugation in 3 ml wash buffer (100 mM EDTA, 50
mM tris-Cl, pH 8), vortex-mixed for 2 sec, and inverted until mixed. The
pooled DNA extract was centrifuged at 15,000 x g for 5 min and the supernatant
was transferred to a sterile Oakridge tube and held at 20-25°C. Then, 5M
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 at 20-25°C 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-Cl, pH 8, and 0.001 M EDTA; TE).
DNA extracts were further purified with Geneclean Glassmilk (BIO 101,
Inc., La Jolla, CA) by adding three volumes of saturated Nal to 100 ul of DNA
extract and 10 ul Glassmilk and processing further according to the manufac-
turer's directions. The DNA-Glassmilk pellet was washed three times by
diffusion with Geneclean NEW wash solution (see manufacturer's instructions)
to avoid shearing of high molecular weight DNA. The DNA was eluted twice from
the Glassmilk by gentle agitation using 50 ul TE.
26
-------
When DNA extracts were purified by CsCl centrifugation, one gram CsCl and
30 ul ethidium bromide (10 mg/ml) were added to each ml of DNA solution
following the potassium acetate precipitation step (density 1.56-1.58 g/ml),
and centrifuged in a Beckman TL-100 using a TLV100 rotor.
For restriction enzyme digestions of DNA, approximately 0.5-2 ug DNA were
incubated for 16-24 h at 37°C with 10 units of either Bam HI or Sal I in
appropriate buffers. DNA samples were electrophoresed for in 0.7% agarose
using a buffer consisting of 0.04 M tris-acetate, pH 8, and 0.001 EDTA.
RESULTS AND DISCUSSION
DNA appeared to be only partially purified at the potassium acetate-
ethanol precipitation step since it was not restricted by Bam HI (indicated by
persistence of the 23-25 kb band). However, DNA purified through the
Geneclean process, or CsCl density gradient centrifugation, or a combination
of both was digested (conversion of the 23-25 kb band into many faintly
visible fragments). Controls without enzymes showed that nucleases in the
impure DNA were not active or absent. We recommend CsCl and Geneclean in
combination to produce the purest DNA. However, since DNA could be recovered
from some soil samples with the Geneclean method alone, we suggest the
ultimate choice depends on the soil.
The method reported here has yielded between 75 to 200 ug DNA from one
gram soil samples, as determined by spectrophotometric absorbance (wavelength,
260 nm). We do not know whether we retrieve all of the DNA from a sample.
The extraction process may be selective. For example, some species of DNA may
be firmly bound to soil components.
In summary, there are several advantages to the method we describe. It
is rapid and 8 samples can be processed in 24 h using Glassmilk or in 48 h if
both CsCl and Glassmilk are used. It is small-scale, since small glassware
items and small volumes of reagents are used. It is simple, since the
ultracentrifugation step can be omitted when DNA is purified with the
Geneclean method alone.
FUTURE WORK
Our plans for this project include:
o Assess the efficiency of DNA recovery with known microorganisms
o Identify the types of soil DNA using specific probes
o Extract bulk DNA from leaves and the rhizosphere
o Extract DNA from a microcosm after spraying a GEM
o Characterize soils using RFLPs
o Demonstrate changes in soil RFLPs after adding a GEM
o Quantify the extraction method
o Increase sensitivity of detection with PCR technology.
27
-------
PUBLICATION
Porteous, L. A., and J. L. Armstrong. 1989. Recovery of bulk DNA from soil
using a rapid, small-scale extraction method. Submitted to Appl. Environ.
Microbiol.
28
-------
UTILIZATION OF THE OSMOPROTECTANT, BETAINE, TO INCREASE THE
EFFICIENCY OF ENUMERATION OF AEROSOLIZED BACTERIA
B. Marthi1 and B. Lighthart2
NSI Technology Services Corporation1 and
U.S. EPA2
Environmental Research Laboratory
Corvallis, OR 97333
INTRODUCTION
Cellular adaptation to osmotic stress (osmoregulation) is mediated
through small molecular weight organic compounds such as glycine betaine,
proline, proline betaine, taurine trehalose and N-acetylglutaminylglutamine
amide. These molecules accumulate in, and are transported from, cells during
osmotic stress and balance the osmotic strength of the cytoplasm with that of
the environment, thus reducing damage due to dehydration.
Studies with bacteria have demonstrated the osmoprotective effects of
betaine, and similar compounds which function by preventing dehydration and
stabilizing enzyme activity in solutions of high ionic strength. Betaine
accumulation and subsequent transport is enhanced at high osmotic levels,
whereas such activity is reduced at low osmotic levels. Betaine also restores
colony-forming ability in osmotically stressed bacteria growing in sea water
and other saline environments.
This study evaluated the effects of betaine on airborne bacteria, as
airborne bacteria may be affected by osmotic stress caused by loss of cellular
water, and addition of betaine may protect them from such stress.
METHODS
Air samples were obtained from three locations: an activated sludge
treatment tank from the Albany, OR Municipal Wastewater Treatment Plant
(WWTP); the second floor roof of the U.S. Environmental Protection Agency's
Western Fish Toxicology Station laboratory building in Corvallis, OR (WFTS);
and, an unobstructed area at a local farm, south of Corvallis, OR. Airborne
bacteria were collected directly on MacConkey's Agar (MAC) and/or Trypticase
Soy Agar (TSA) plates with and without 2 mM betaine in slit samplers or in All
Glass Impingers (AGIs) containing 0.01M phosphate buffer with and without 2 mM
betaine.
To determine the concentration of betaine that produced the maximum
effect on colony forming ability, a suspension of Pseudomonas svringae (ca. 1
x 10 CFU/mL) was sprayed into a greenhouse, collected in AGIs, and enumerated
on Luria-Bertani (LB) agar plates containing 0 mM (control), 2 mM, 5 mM, 10
mM, 15 mM, or 20 mM betaine.
All data obtained from these studies was subjected to analysis of
variance (ANOVA) using the Statistical Analysis System.
29
-------
RESULTS AND DISCUSSION
Results from all three locations showed a statistically very significant
increase (p<0.1) in counts of airborne bacteria when 2 mM betaine is added to
enumeration media as compared to a non-betaine containing control (21.6% to
61.3%). This effect was independent of the extramural sampling sites, as
very similar results were obtained from three very different locations.
Betaine incorporation also significantly increases the numbers of pigmented
bacteria.
The presence of betaine in both the collection fluid and the enumeration
medium very significantly increased the colony-forming ability of airborne
bacteria from the WWTP. Betaine addition to either the AGIs or the medium
alone showed an effect that is lower than when betaine was present
simultaneously in the AGIs and the medium. This indicates that betaine may
have an additive effect, and may be able to facilitate recovery of specific
sub-populations of stressed organisms otherwise unable to survive the short
(20 min) incubation in the collection fluid.
The effect of betaine was concentration-dependent, with the highest
degree of recovery seen at concentrations of 2-5 mM. At higher concentrations
(10-20 mM) the effect of betaine was reduced, suggesting that there may be a
critical threshold concentration of betaine beyond which its activity is
reduced.
This study showed that addition of a simple organic compound like betaine
can significantly increase the colony forming abilities of airborne bacteria.
Therefore, it may be possible to increase efficiencies of recovery and
enumeration media by relatively simple modifications. This, in turn, can lead
to the development of protocols that can be used to monitor environmental
releases of airborne microorganisms, including genetically engineered
microorganisms (GEMs).
PUBLICATIONS
Marthi, B., and B. Lighthart. 1990. Effects of betaine on the enumeration of
airborne bacteria. Appl. Environ. Microbiol. (in press).
Marthi, B., M. Walter, V. Prince, R. Seidler. Dilution versus loss of viable
counts in aerosolized bacteria. Abstract, Annual Meeting, American Society
for Microbiology, May 13-18, 1989. New Orleans, LA. Manuscript in prepara-
tion.
Walter, M., V. Prince, B. Marthi, R. Seidler. Effects of aerosol-induced
stress on post-aerosol survival of bacteria. Abstract, Annual Meeting,
American Society for Microbiology, May 13-18, 1989. New Orleans, LA.
Manuscript in preparation.
30
-------
EVALUATION OF METHODS FOR DETECTING DNA
SEQUENCES IN THE SOIL ENVIRONMENT
Batty H. Olson, Paul A. Roche!le, and Simon J. Ford
Environmental Microbiology and Genetics
Social Ecology
University of California
Irvine, CA 92717
INTRODUCTION
The aims of this research were to compare hybridization techniques for
the detection of DNA sequences in environmental samples, to evaluate probe
labeling techniques, and to develop a direct extraction method for the
detection of DA in soil. Important considerations when developing techniques
are sensitivity, reproducibility, specificity and ease of application. The
detection of genes encoding mercury resistance was used as the model system
with which to evaluate these techniques. Mercury was chosen because the
biochemistry and genetics of resistance were well understood and a number of
gene probes were readily available.
METHODS
The probes used were merR (675 bp) from Tn501, merA (1 kb) and merB (457
bp) from the broad spectrum resistance operon of pDU1358, and a Gram positive
probe (G+, 6.4 kb) containing the entire mer operon of p!258. Three methods
of labeling probes with 32P were employed: nick translation produced probes
with a specific activity 106 dpm/ug DNA while random priming and in vitro
transcription routinely gave a specific activity of 109 dpm/ug DNA. Because
of its simplicity and high specific activity, random priming is now the
labeling method of choice in the laboratory.
RESULTS AND DISCUSSION
Two methods of confirming cell lysis were compared: a) hybridization
with a riboprobe; and b) staining with methylene green. With 70% (n=80) of
isolates which hybridized with the G+ probe, the riboprobe and methylene green
stain both showed lysis. However, in 29% of isolates the methylene green
detected DNA whereas the riboprobe did not. DNA was never detected by the
riboprobe when methylene green showed no lysis. Therefore, methylene green is
now routinely used.
Using an SDS based lysis method 85% of Pseudomonas spp. (n=233), 76% of
Flavobacterium spp. (n=62), 70% of Acinetobacter spp. (n=20), and 70% of
Moraxella spp. (n=23) were lysed. Using sodium hydroxide to lyse colonies on
nylon fiLters, 86% (n=4750) were lysed, as determined by methylene green
staining.
Gene divergence and probe homology are important factors to consider when
working with environmental samples. For example, 32-96% of mercury resistant
isolates had DNA homologous with the merA probe under conditions of 95%
31
-------
stringency, and 46-85% of isolates resistant to organomercury probed positive
with the merB probe, depending on the site from which samples were collected.
Thus, up to 68% of mercury resistant isolates did not hybridize with the merA
probe. Divergence of the mercuric reductase gene probably accounts for a
large part of this non-hybridization.
A novel method was developed for the recovery and analysis of DNA in
environmental samples and was used to detect DNA sequences homologous with the
mer probes in river sediments, activated sludge and mercury contaminated
soils. The method involved casting 0.25 g of environmental sample in an
agarose cube (1 cm x 1 cm x 1 cm) and then lysing the bacteria within the cube
by treatment with acetone, lysozyme, and alkaline SDS, followed by treatment
with RNase and Proteinase K. The agarose cube was than sliced, weighed and
incorporated into a conventional horizontal agarose gel and subjected to
electrophoresis overnight. The DNA could then be examined by Southern
analysis. Less than 24 h was required to produce hybridization ready
membranes. The method could detect 1 pg of DNA in 0.25 g of soil and plasmid
DNA could be detected in 1 x 104 bacteria in 0.25 g of soil.
32
-------
SESSION III
DISPERSAL/TRANSPORT
33
-------
-------
ASPECTS OF SIMULATED VIABLE DROPLET AIRBORtfE TRANSPORT
Bruce Lighthart
U.S. EPA
Environmental Research Laboratory
Corvallis, OR 97333
INTRODUCTION
In the past, large-scale aerial dispersal of microorganisms has been
estimated by using atmospheric dispersion models based on temporal and spatial
averaging as defined in Gaussian plume models. The averaging process elimin-
ates the detail necessary to adequately delineate the deposition pattern of a
microbial aerosol near its source.
The purpose of this presentation is: (a) to describe a droplet disper-
sion model for viable microbes that is useful for predicting and detecting
with relatively high spatial resolution near-source downwind dispersion
patterns; (b) to describe a general regression model for airborne bacterial
death rate as a function of time, relative humidity (RH), and temperature.
METHODS
Droplet Dispersion Model
The droplet model was separated into five submodels: aerosol generation,
dispersion, deposition, and microbial death. Any or all w»re computed at each
time-step in the trajectory of a droplet.
To determine how well this model fit the survival of airborne microbes
under natural conditions, data obtained from the field spray of Pseudomonas
syringae at Tulelake, California, was utilized.
Death Rate Model
Previously published death rate data for 15 species was utilized to
generate this model. A regression equation for the death rate of airborne
bacteria as a function of time, relative humidity (RH) and temperature was
developed. A three-dimensional plot relating the death rate constant, the
evaporation function, and the aerosol age was generated. The death rates of
both Gram positive and Gram negative organisms were compared.
RESULTS AND DISCUSSION
Droplet Dispersion Model
The sprayout of an aqueous polydispersed aerosol of Pseudomonas syringae
on a quiescent summer day in Tulelake, CA, showed a pattern of deposition
about the source with three and possibly four downwind deposition plumes.
During the observation time, the wind shifted from a relatively strong
northeast wind to a weaker erratic southwest wind. Except for an easterly
35
-------
plume, the wind conditions accounted for the plume patterns. At the time of
the wind shift, the simulated sprayout had ceased, but the aerosol droplets
produced had progressed far downwind. During this downwind progression, the
southern plume had a more easterly dispersal pattern. It is thought that the
mathematical wind-filtering procedure obscured the earlier southwesterly
episodes that might have occurred during the sprayout event.
There are two general uses for microbial dispersion models: (a) to
predict prior to release where microbial droplets might disperse, and (b) to
determine where microbial droplets have dispersed after release. The
predictive models have been the focus of past research and are the Gaussian
type used in large systems, while determinative models are undergoing develop-
ment and represent the aerosol-particle type used in small-scale (and large-
scale) systems. The particle type also can be used for prediction if repre-
sentative wind, temperature, and humidity are known.
A practical application of the mathematical model would be to prepare a
strategy to limit the downwind drift from a sprayout. For example, simulated
droplets generated from a sprayer can be divided into those whose trajectories
intersect the ground surface; those >50 urn in diameter, about 50% of the
droplets from the simulated nozzle; and those <50 Urn in diameter, whose
trajectory contributed to the dispersion cloud. Another application would be
to locate the sample site dependent upon the predicted deposition pattern
generated by the model with known input meteorological conditions and death
rate constant of the applied agent.
Death Rate Model
This regression model presents airborne bacterial death rate as a
function of time, RH and temperature (the "evaporation function"). The model
was generated from published data and had a very good fit (R2 » 0.94). The
death rates of Gram positive bacteria were consistently lower than those of
Gram negative bacteria under the conditions tested.
FUTURE WORK
Future applications of the droplet dispersion model would be its use in
real time to predict the aerosol dispersal of a biological mitigant sprayed
out but not contaminating a field where an experimental bacterium has been
applied. Experiments are now in progress to validate the model using data
from actual sprayouts, and also to determine death rate kinetics of airborne
bacteria.
PUBLICATIONS
Lighthart, B., and A.S. Frisch. 1976. Estimation of viable airborne microbes
downwind from a point source. Appl. Environ. Microbiol. 31:700-704.
Lighthart, B., and J. Kim. 1989. Simulation of airborne microbial droplet
transport. Appl. Environ. Microbiol. 55:2349-2355.
36
-------
Seidler, R.J., and S. Hern, 1988. Special report: release of ice-minus
recombinant bacteria, p.83. Environmental Research Laboratory, U.S. Environ
mental Protection Agency, Corvallis, OR.
37
-------
THE ASSESSMENT OF BACTERIAL DISPERSAL FROM FOLIAGE BY RAIN SPLASH
H. Alastair McCartney and Julie Butterworth
Institute of Arable Crops Research
Rothamsted Experimental Station
Harpenden, Hertsfordshire
United Kingdom AL5 2JQ
INTRODUCTION
The potential benefits to agriculture ensuing from the use of genetically
engineered bacteria as novel pest and disease control agents or as replace-
ments for chemical agents is great. However, before such organisms can be
used commercially it must be shown that their application will not cause
unwanted effects elsewhere. Therefore this project seeks to quantify the
efficiency of natural rainfall and irrigation systems in dispersing foliar
bacteria and to determine the spatial scales of transport involved.
The project is only in its preliminary stages, however, we outline below
the approaches being adopted to quantify the potential that wind and rain have
for dispersing bacteria.
METHODS
The approach adopted is to study bacteria dispersal by simulating the
effects of wind and rain using the Rain Tower/Wind Tunnel (RTWT).facility at
Rothamsted. Briefly, to simulate rain, water drops are released from the top
of the Rain Tower and allowed to fall 11 m onto target plants or leaves below.
The resulting splash droplets are collected at different distances from the
target and the fraction of the bacteria on the target carried by the droplets
assessed. The fraction dispersed in runoff water and ejected into the air is
also measured. Experiments can be done in wind speeds ranging from 0 to about
8 ms'1.
As natural rain contains a wide range of droplet sizes the full droplet
spectrum is difficult to simulate in the laboratory. Thus we have chosen to
use mono-sized drops and to identify the drop size range responsible for
dispersal. The drops are produced from a single row of 20 hypodermic needles
spaced 4 cm apart. The needles can be changed to produce drops from 2 to 5.2
mm in diameter. The wetted target area at the foot of the Rain Tower is about
30 cm deep over the width of the tower. Sterile water is used during experi-
ments and the rain generator is regularly sterilized by washing through with
bleach solution.
Targets can either be whole plants or removed leaves. When leaves are
used they are supported on a nylon frame which can be set at different angles
to the horizontal. Before exposure to splash the bacteria content of the
target leaves is determined by washing a subsample of the leaves in 50 ml of
sterile Ringer's solution for about 1 hour (experimentally determined optimum
time). The bacteria content of the washing water is then measured, by plating
38
-------
out on nutrient agar, to determine the numbers of bacteria per leaf and thus
on the target.
RESULTS AND DISCUSSION
Due to the short time since the funding of this project, there are as yet
no appreciable results to report.
39
-------
MEASURING POST AEROSOL SURVIVAL OF SPRAYED BACTERIA
Michael V. Walter,1 Valerie J. Prince,1 Balkumar Marthi,1
Brenda Shaffer, Lisa Ganio,1 and Ramon J. Seidler2
NSI Technology Services Corporation1
U.S. EPA2
Environmental Research Laboratory
Corvallis, OR 97333
INTRODUCTION
The use of aerosols to apply genetically engineered microorganisms (GEMs)
has raised concerns regarding the dispersal and subsequent survival of these
organisms in unintended locations. While field data exist describing aerial
dispersal of fungal spores, there is little or no information regarding
aerosol dispersal of bacteria. Likewise, little is known about what effect
aerosolization has on the subsequent survival of vegetative bacterial cells
following aerosolization.
Such data is of relevance to both the biotechnology industry and
regulatory agencies. Bacteria which are unable to withstand the stress of
aerosolization, or are more prone to injury incurred during aerosol ization
would be less likely to survive in competition with the indigenous populations
encountered in the environment. Therefore, bacterial survival following
aerosolization (post aerosol survival) would determine the application
strategies from an industrial point of view, and the monitoring strategies
from a regulatory stand point.
Another cause for concern about the release of GEMs into the natural
environment is their dispersal by wind after they have established populations
in the soil and phyllosphere (entrainment). Little is known about the
entrainment of indigenous bacteria, and virtually nothing is known about the
entrainment of GEMs.
It was the objective of research investigating survival of bacteria
following aerosolization to: a) develop and test a method to determine if
aerosolization can impair the subsequent survival of bacteria; b) identify
what environmental conditions can influence post aerosol survival and, c)
determine if post aerosol survival is specific to different genera of
bacteria.
Research investigating the entrainment of bacteria from soil and the
phyllosphere was intended to: a) develop a method which can be used to induce
bacterial entrainment from the phyllosphere and soil; b) investigate the
effects of leaf moisture on entrainment; and, c) determine if bacterial
entrainment can be detected over a prolonged period of time.
Cultures and Media
Experiments for both post aerosol survival and entrainment studies were
conducted using nonrecombinant strains of either Pseudomonas syrinqae,
40
-------
spontaneously resistant to 100 ug/ml of rifampicin (Rif) or Erwinia herbicola
spontaneously resistant to 500 ug/ml of nalidixic acid (Nal) which served as
target bacteria in both of these studies. Target microbes were grown in LB
broth containing the appropriate antibiotic at 30°C for 18 hrs with shaking at
280 rpm. Cultures were washed three times in 10 mM phosphate buffer (pH 7.2)
and resuspended in the same buffer. Target organisms were enumerated using LB
agar containing the appropriate antibiotic and counted after incubation at
30°C for 48 hrs.
Post Aerosol Survival
Target bacteria were sprayed for two minutes at a pressure of 36 psi
using a C02 stainless steel sprayer (R & D Sprayers, Opelousas, LA). Sprays
were conducted in a 100 X 10 m greenhouse. The air was sampled for an
additional 18 minutes following the spray using all glass impinger air
samplers (AGIs). The number and distance of AGIs from the point of release
were as follows: 2 AGIs located at 1 m and 6 AGIs located at each of 3, 5,
10, and 15 m. Replicate samplers were located approximately 10 mm apart. In
addition, 2 control samplers into which the target bacteria were inoculated
directly from the spray tank were located 3 m up wind from the spray site.
After sampling, target bacterial populations in AGIs were enumerated
either by serial dilution or by filtration through .45 urn, 37 mm diameter
filters which were placed on selective LB agar plates. AGIs were placed in a
water bath at 30°C and sampled every two hours over a 6 hr period.
A linear regression of the base 10 logarithm of CPUs observed for each
distance from the site of release, over the six-hour sampling period, was
conducted. The regression included indicator variables for different environ-
mental conditions as independent variables. The rate of change of the log
CPUs over time for different environmental conditions, as measured by the
coefficient of time in the regression model, was compared using T-tests.
Bacterial Entrainment
Experiments designed to measure entrainment of bacteria from the phyllo-
sphere were conducted using beans and oats. Oats were planted in three beds,
10 m long. Beans were planted in two beds 15 m. In both cases, beds were
planted on a north-south axis. The beds were divided into two sections of 5 m
each. Each 5 m section was subdivided into three subsections which denoted
the moisture condition of the leaf surface at the time entrainment was
induced, either dry, or following misting of the plants for five minutes. The
third section was sampled prior to misting and was not subjected to entrain-
ment and was used to measure the effect of the artificial wind on bacterial
population. On the eighth day of the experiment, all plants were misted prior
to inducement of entrainment to determine if differences between the two
treatments could be explained on the basis of leaf surface moisture condition.
Air flow during the spray was north to south. The north five meter
section was sprayed directly until run-off with L. svrinqae using a CO,
sprayer at a pressure of 36 psi. Plants in the adjacent 5 m section were
inoculated by bacteria which drifted down wind.
41
-------
Bacterial entrainment was induced using a 17 inch window fan which was
placed approximately 500 nun from the plants. Plants were blown for 5 min and
entrained bacteria were collected using 150 m petri dishes containing agar
supplemented with 100 ug/ml of Rif. Petri dishes were placed in two rows of
three from about 100 mm to 450 mm from the plants which gave a sampling
surface area of 0.106 m2.
Leaf populations were determined by sampling 5 oat leaves or 3 bean
leaves selected at random from each subsection and placed in sterile stomacher
bags (Tekmar Scientific, Cincinnati, OH). Samples were diluted in 20 ml of 10
mM phosphate buffer and blended for one minute in a stomacher blender
(Tekmar). Serial dilutions were plated onto L8 agar containing Rif. All
plates were incubated for 48 hr at 30°C and colonies counted.
RESULTS AND DISCUSSION
Post-Aerosol Survival
Target bacteria from control samplers and samplers located 1 m from the
site of the spray had slopes for the time variable equal to zero, indicating
that there was no change in populations over the six hour period. Target
bacteria from samplers located 3 m or greater from the site of the spray
demonstrated negative rates of change which were significantly different from
zero (P<0.05). This indicated that the populations declined over the six hour
period following the release at rates which were significantly greater than
bacteria which were not sprayed or traveled only 1 m. Therefore, it would
appear that aerosolization can impair subsequent survival of bacteria beyond a
certain distance.
The magnitude of this impaired survival appears to be related to the
environmental conditions under which the bacteria are aerosolized. For
example, bacteria released at 12°C and 76% RH declined over the six hour
period at rates ranging from -0.01 In cfu/hr at 3 m to -0.05 In cfu/hr at 15
m. Bacteria released at 27°C and RH of 40% decreased at rates ranging from
-0.36 to -0.50 In cfu/hr at 3 and 15 m, respectively. The rate of decline of
bacteria released under conditions of 12°C, 76% RH as measured by the slope of
the time variable in the regressions were significantly higher (P<0.05) than
the rate of decline of bacteria released at 27°C and 40% RH.
The post aerosol survival of bacteria which were released at 27°C and 80%
RH ranged from -0.05 at 3 m to -0.06 In cfu/hr at 15 m. However, bacteria
released at the same temperature, but at an RH of 40% declined at rates
ranging from -0.36 at 3 m to -0.50 at 15 m, which were significantly different
(P<0.05) from rates at 27°C and 80% RH. Therefore, it would appear that the
RH has more of an impact on post aerosol survival than does temperature.
42
-------
Table 1. Correlation between rates of decline in colony forming units and
relative humidity.
Rate of Decline, as log (cfu)/hr
Temp. R,H. at 3 m at 15 m
12°C 76% -.01 -.05
27°C 40% -.36 -.50
27°C 80% 0.05 -.06
Bacterial Entrainment
Bacteria sprayed directly onto oat plants demonstrated survival which was
comparable to the survival of bacteria which were sprayed directly onto bean
plants. Populations on oats decreased from approximately 1 X 10"/g leaf on
the day of the spray to 1 X 105/g leaf on day 8. The populations observed on
beans decreased from about 1 X 107/g leaf on the day of the spray to 2 X 105/g
leaf on day 8 of the experiment.
Target bacterial populations observed on plants which where drift-
inoculated during the spray showed different survival trends in comparison to
those which where directly inoculated. On oats, the target population
declined from 1 X 102/g on the day of the spray to undetectable levels on the
day after the spray. On beans, the target population declined from about 1 X
10s/g leaf on the day of the spray, to undetectable levels within 48 hrs. The
possibility exists that these organisms were injured during aerosolization
impairing their subsequent survival. This observation is supported by results
which indicated that aerosolized bacteria demonstrated reduced survival
compared with bacteria not aerosolized or aerosolized distances less than 1 m
as described in post aerosol survival experiments.
Entrainment of bacteria from directly inoculated oats was observed
throughout the eight days of the experiment. Higher numbers of entrained
bacteria were observed from wet oat leaves than from dry oat leaves. In both
cases, the numbers of entrained bacteria did not appear to decrease substan-
tially over the six days of the experiment. On day eight, when all plants
were misted with water prior to inducement of entrainment, the number of
entrained bacteria detected from moist plants increased from about 70 cfu/.106
m2 to over 200 cfu/.106 m2.
Entrainment was greatest from beans on the day of the spray at about 70
cfu/.106 m2 for wet plants, and 200 cfu/106 m2 from dry plants. Low numbers
of entrained bacteria were detected throughout the 8 days of the experiment
from wet plants. However, the number of entrained bacteria from dry plants
decreased dramatically from 200 on day 0, to 70 on day 1, to undetectable on
day 6. However, on day 8, when plants were tested for entrainment when wet,
low numbers were detected.
Entrained bacteria from drift-inoculated plants were also observed at
numbers which ranged from 1 to 5 times less than directly inoculated plants
over the eight days of the experiment. This is surprising since populations
on the leaves of both oats and beans became undetectable by day 2 of the
43
-------
experiment. Entrainment was higher from oats than beans. In fact entrained
bacteria were detected from oats throughout the 8 days of the experiment,
while beans became undetectable by day 6. However, on day 8, low numbers were
detected after plants were misted.
FUTURE WORK
Future research in bacterial entrainment will be directed towards: a)
developing more sensitive sampling methods, b) comparing the entrainment of
applied bacteria with that of the indigenous bacteria, c) measuring the effect
of environmental conditions on the magnitude of entrainment, d) determining
the vertical and horizontal patterns of entrainment, and e) investigating if
entrained bacteria can colonize other environments.
44
-------
DEVELOPMENT OF HALF-LIFE RATE CONSTANTS FOR AEROSOLIZED
BACTERIA SUBJECTED TO ENVIRONMENTAL CONDITIONS
Linda D. Stetzenbach and Brian D. Green
Environmental Research Center
University of Nevada
Las Vegas, NV 89154
INTRODUCTION
The U.S. Environmental Protection Agency is charged with the responsi-
bility to register and issue permits for the commercial use and release of
genetically engineered microorganisms (GEMs) under the Toxic Substances
Control Act (TSCA) and microbial pest control agents (MPCAs) under the Federal
Insecticide, Fungicide, and Rodenticide Act (FIFRA). This process involves
analysis of the potential impact to the environment resulting from the release
of GEMs or MPCAs. An assessment of the viability of the organisms following
release and the extent of the dispersal of the released agent are significant
factors in this analysis.
The U.S. EPA Environmental Research Laboratory in Corvallis is currently
developing transport models to assist in determining potential impact follow-
ing aerosol release and subsequent dispersal of microbial agents. These
models are designed to include viability of the released organism. Survival
rates of GEMs and non-recombinant parental strains subjected to aerosolization
and environmental stresses, however, have not been determined. Studies to
determine the survival rates (half-life) of microorganisms,would assist in the
development of these transport models.
Laboratory-based survival studies, therefore, have been initiated with
selected bacteria species termed "benchmark strains" to determine the half-
life rate constants under varying environmental conditions.
METHODS
A biological aerosol vertical chamber has been designed and constructed
by the environmental Monitoring Systems Laboratory in Las Vegas to subject
microorganisms to ranges of environmental stresses following aerosolization.
Bacterial suspensions of a known source strength are aerosolized using a
nebulizer that produces a droplet size range of 1 to 10 urn in diameter. The
aerosol cloud is then directed into a circular holding chamber comprised of
six settling columns. When the sampling ports at the bottom of the holding
chamber are opened, the droplets settle through columns ranging in length from
4 cm to 120 cm. The bacteria in the droplets.are subjected to selected ranges
of temperature and relative humidity and RH is determined by the length of
column and the settling velocity of the droplet. The bacteria are collected
at the bottom of the columns into sampling dishes containing a phosphate
collection buffer. Currently data are being collected on the survival of a
rifampicin resistant strain of Pseudomonas syrinoae. Seven additional
bacteria species have also been selected for study. These species include
rifampicin resistant: Pseudomonas aeruginosa. Pseudomonas cepacia.
45
-------
Pseudomonas aureofaciens. Klebsiella planticola. and Enterobacter cloacae and
nalidixic acid resistant Pseudomonas aeruoinosa and £. cloacae. Spores of
Bacillus subtilis are to be co-aerosolized with the vegetative bacteria cells
to account for physical loss of organisms (i.e., wall effects of the settling
columns).
Organisms are enumerated using several techniques. A subsample of the
phosphate buffer is processed for total counts using a Coulter ZM particle
counter calibrated to enumerate particles of a selected size range. Total
counts are also enumerated using epifluorescence microscopy with acridine
orange stained subsamples (AODC). Viable counts using spread plate culture
techniques are performed on culture media supplemented with rifampicin or
nalidixic acid as appropriate. Audit blank samples, collected without the
introduction of a bacterial suspension into the bioaerosol chamber, are also
analyzed using the same procedures.
RESULTS AND DISCUSSION
Procedures for the operation of the bioaerosol chamber and analysis
protocols for the electronic particle counter, AODC, and spread plate analyses
have been established to generate quality assured data. Data of total numbers
of organisms from the replicate sampling ports have shown recovery of cells
without a significant difference according to port location or column length.
Variability of viable bacterial counts between sampling port locations,
however, has been noted. These preliminary data also indicate that the
filtered carrier air may be a factor in the survival of the aerosolized cells
prior to settling in the columns.
Ongoing trials with Pseudomonas svrinqae are currently being conducted
with the settling chamber conditions at 27°C, 18°C, and 12°C with 80% and 50%
RH to determine half-life rate constants. These data will then be incorp-
orated into the ERL-Corvallis bacteria transport models.
FUTURE WORK
The eight "benchmark" non-recombinant bacterial strains will be aerosol-
ized and subjected to the three temperature and two humidity values described
above to determine the half-life rate constants for each strain. Similar
chamber studies could also be conducted on recombinant organisms to provide
survival data for transport models.
46
-------
EVALUATION OF AEROSIOLOGICAL SAMPLERS FOR THE RECOVERY OF
MICROORGANISMS RELEASED TO THE ENVIRONMENT BY AEROSOL
Mark P. Buttner and Linda D. Stetzenbach
Environmental Research Center
University of Nevada
Las Vegas, NV 89154
INTRODUCTION
Microorganisms may be introduced into the environment for a variety of
beneficial applications, including increasing agricultural productivity,
degradation of pollutants and enhanced extraction of minerals. Monitoring the
fate and transport of released genetically engineered microorganisms (GEMs) is
essential to assess the potential impact of these organisms in the environ-
ment. Currently, there are insufficient data on the relative efficiency of
aerobiological sampling equipment to design a valid monitoring scheme for the
recovery of aerosol released GEMs. The purpose of this research is to
evaluate air monitoring methodologies for the tracking of GEMs released to the
environment via aerosol. This includes the evaluation of the relative
efficiency of both active and passive aerobiological sampling equipment and
various enumeration methods for potential use in future field release monitor-
ing efforts. These data can then be used for the detection of aerosolized
microorganisms and minimize the uncertainty in data encountered in previous
field releases.
METHODS
Preliminary trials were conducted at EMSL-Las Vegas in a greenhouse
facility measuring 9 meters by 6 meters. Three active (forced air flow)
samplers and two passive sampling methods have been compared to date. The
three active samplers, the Andersen six-stage impacter sampler, an all-glass
impinger (AGI-30) sampler and the Spiral Air Systems (SAS) sampler were co-
located on platforms and operated at the manufacturer's recommended flow rate
of 28 1pm, 12 1pm, and 128 1pm, respectively, using Millipore vacuum pumps.
Gravity plates and sentinel plants (oat plants in pots) were also located on
the platforms as passive (gravitational settling) sampling methods. Duplicate
sets of the five samplers were placed on benches 8 meters from the sprayer. A
rifampicin resistant strain of Pseudomonas syrinqae, one of right "benchmark"
bacterial strains, was aerosolized at a concentration of 107 cells/ml. Spores
of Bacillus subtilis were used as an indicator of physical loss during
aerosolization. The spores were mixed with P. syrinqae at a final concentra-
tion of 105/ml immediately prior to spraying. The organisms were aerosolized
using a backpack field sprayer fitted with a feeJet 8004 nozzle and operated
at 35 psi for 2 minutes. Temperature and relative humidity were recorded
before each spray trial. Andersen, AGI-30 and SAS samplers were operated for
5, 10 and 1 minute(s), respectively, in order to sample equivalent volumes of
air. Passive samplers remained in place for time periods of 10, 30 or 60
minutes from the initiation of the spray.
47
-------
King's medium B agar supplemented with 100 ug/ml of both rifampicin and
cycloheximide incubated at 28°C was used for selective recovery of P. syrinqae
and nutrient agar with cycloheximide (100 ug/ml) incubated at 37°C was used
for recovery of B. subtilis spores. Andersen, SAS and gravity plate samplers
were enumerated directly, and the active sampler data converted to CFU/m air.
AGI-30 samplers contained 20 ml phosphate buffer (0.01 M) as the collection
medium). Cells and spores were enumerated from the collection buffer by
spread plate, acridine orange direct count (AODC) and electronic particle
counter EPC) methods. Recovery of cells from oat plants consisted of harvest-
ing 5 g of plant material, placing the sample in sterile bags containing 50 ml
phosphate buffer (0.01 M) and homogenizing the leaves for 1 minute using a
Tekmar stomacher. The extract was then spread onto agar plates.
RESULTS AND DISCUSSION
In preliminary trials, viable B. subtilis were recovered from all
samplers using a spray concentration of 105 spores/ml. Viable counts from
Andersen, AGI-30 and SAS samplers were all ca. 10* CFU/m3. Approximately 102
CPU were counted on both 30 and 60 minute gravity plates, and ca. 102 CFU/g
fresh wt. (below the reportable detection limit of <30 colonies per plate)
were recovered from oat plants.
Approximately 102 to 103 CFU/m3 of P. svringae were collected with the
Andersen samplers. No viable organisms were recovered from co-located AGI-30
samplers (detection limit >4.5 x 10* CFU/m3). Total counts (AODC) from the
AGI samplers, however,m were 10* cells/m3. These data indicate an average of
<0.1% of P. syrinqae cells collected were viable. Data from passive samplers
also indicated low viability for P. syrinqae in these triaJs. On 30 minute
gravity plates, a mean of 2 colonies were counted, and a mean of <1 colony was
counted on 60 minute plates. No viable organisms were counted on leaf
washings from plants (detection limit 3.0 x 103 CFU/g fresh weight).
The data from these preliminary sprays indicate a physical loss of
approximately 1 order of magnitude from sprayer (cells/ml) to sampler (total
cells/m3) of both P. syrinqae and B. subtilis. The low viability of aerosol-
ized P. svrinqae remains unexplained at this time and future experiments are
being designed to address this issue.
FUTURE WORK
Plans for upcoming trials include lowering detection limits of AGI and
sentinel plants by concentrating samples, determining the viable non-
culturable fraction of cells, and testing replicate samplers at various
distances from the spray source. Future experiments will extend the sampler
comparison to the "benchmark" bacterial species.
48
-------
MODELING THE FATE OF BACTERIA IN SURFACE WATERS
John P. Connolly,1 Robert V. Thomann,1
Richard B. Coffin,2 and Robin Landeck1
Environmental Engineering and Science Program1
Manhattan College
Riverdale, NY 10471
Technical Resources, Inc.2
U.S. EPA
Gulf Breeze Environmental Research Laboratory
Gulf Breeze, FL 32561
INTRODUCTION
The proposed release of a genetically-engineered microorganism (GEM) to
the environment engenders questions regarding resulting ecological and public
health impacts. Our experience with chemicals (e.g., pesticides) indicates
that it is highly likely that a GEM introduced to the environment will be
transported beyond the site of application. Whether that transport results in
an adverse effect is dependent on the fate of the organism in the environment
and the relationship between organism density or activity of 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. Predict-
ing 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 a GEM to invade the community. The
analysis must also describe the transfer of engineered genetic material
between the GEM and indigenous organisms.
The theory of bacterial population dynamics is well developed and models
exist for simple laboratory systems. However, quantitative extrapolation to
the field has not been demonstrated. A necessary first step in the develop-
ment of a risk assessment framework is the development and testing of a
modeling framework to describe bacterial population dynamics in natural water
systems. It is the purpose of our cooperative agreement to formulate and test
a framework for predicting bacterial population density in natural water
systems. At this point our work has been restricted to natural populations
and has focused on bacterial growth kinetics and the relationship between
bacteria and phytoplankton.
49
-------
METHODS
Modeling Framework for Bacterial Growth and Death
Growth of bacteria and their uptake of substrate is classically described
using the Monod equation. Death is determined by a non-predatory loss rate
(i.e., respiration and loss of viability) and a predation (and/or parasitism)
rate. The equation describing these mechanisms is:
j? - umf(S,N,02)B - (kd+k,P) B
where B = bacterial biomass
um - maximum growth rate
f - limitation factor
S - substrate carbon concentration
N = concentration of additional, potentially limiting, nutrients
O2 » oxygen concentration
P - predator concentration
kd - non-predatory loss rate
kp » predation rate
The limitation factor f accounts for the reduction in growth rate due to
substrate carbon, nutrient or oxygen limitation. For any one of these effects
the limitation factor is given by the Monod expression. For substrate the
equation is:
(2) f-_J
where Km - Michael is half-saturation constant
The carbon and other nutrient concentrations are controlled by the rate
at which these constituents enter the system being modeled from external
sources, the rate at which they are produced in the system through algal
excretion and grazing (i.e., internal sources), the rate at which they are
transported through the system by water flow and dispersion and the rate at
which they are taken up by the bacteria. Uptake by the factor is defined by
the growth rate and a carbon or nutrient yield coefficient:
(3) dS m UB
dt YB
where u » net growth rate - umf(S,N,02)
Y - yield coefficient
The predation term included in equation (1) defines the loss of bacteria
to the next trophic level. The equation defining the grazer population (Z)
is:
50
-------
(4) dl.
dt
where a - assimilation efficiency of the predator
G - grazing rate
Kx - half-saturation constant for grazing rate limitation
Kdl - death rate of the grazers
The assimilation efficiency defines the fraction of grazed bacteria that is
converted to biomass. The remaining fraction is assumed to be excreted to the
substrate pools. The predation term of equation (1) is defined from equation
(4) as
(5)
RESULTS AND DISCUSSION
Analysis of Bacteria-Substrate Interactions
Yield Coefficients
We have compiled several hundred growth yield values encompassing greater
than seventy compounds. Analysis of these data has indicated that values are
relatively constant for any one substrate. Significant differences exist
between substrates, although substrates of similar structure appear to have
similar yields. The highest yields are observed for the amino acids.
Alcohols, alkanes and sugars all have similar yields. Yields on DOC and POC
from plant materials are generally much lower. Consistent with previous work,
we found that growth yields were linearly related to the heat of combustion of
the substrate for substrates with heats of combustion less than about 11
kcal/gC. Above this value the yields were approximately constant.
Substrate Definition and Utilization
As an initial hypothesis we have assumed that carbon can be separated
into labile, refractory and non-degradable components. Evidence of a split
between labile and refractory substrate components was noted in the BOD assays
being conducted at ERL-GB on water from Santa Rosa Sound and Range Point.
To estimate parameter values of a Monod model considering two substrate
components, we began by analyzing published studies of bacterial utilization
of substrates that may be classified as either labile (e.g., glucose) or
refractory (e.g., secondary effluent). Significant differences in bacterial
utilization of the two carbon sources were determined. As expected, maximum
growth rate and growth yield are much higher for glucose than for secondary
effluent. The half-saturation constants are also much different, with the
glucose value being about two orders of magnitude less. Respiration rates are
about the same, reflecting the bacterial rate of endogenous metabolism. These
values provide some guidance for the application of this model to carbon
sources that contain both labile and refractory components.
51
-------
We have applied the model to data from Santa Rosa Sound 600 experiments
by dividing the total degradable substrate calculated from the ultimate BOD
into labile and refractory components. The labile component was set by
visually estimating the BOO at the break in oxygen utilization presumed
representative of depletion of the labile substrate. Oxygen utilization and
bacterial numbers from experiments with filtered and unfiltered water were
successfully fit with the same parameter values.
The coefficient values used for the labile and refractory components are
similar to those found for our earlier fits to labile and refractory substrate
sources. To further test the model we computed the BOD for an independent
experiment with Santa Rosa Sound water in which the oxidizable carbon appeared
to be much more labile. Assuming that all of the oxidizable carbon was labile
and using the coefficient values determined previously the model fits the data
fairly well.
Our analysis of the BOD data suggests that it may be feasible to model
substrate by using oxygen utilization as a measure of substrate and separating
the oxidizable substrate into labile and refractory components. Monod
kinetics appear to be capable of simulating the bacteria and substrate
dynamics, although more rigorous tests are needed in which the influence of
grazing is considered.
Bacteria-Phytoplankton Relationship
As part of the overall effort on developing modeling frameworks for
predicting the fate of genetically engineered bacteria in aquatic systems, a
more fundamental understanding of the relationships between: the bacteria and
phytoplankton is necessary. Such an understanding is facilitated by the
development of generalized models that incorporate the principal features of
the interactions between the heterotrophic bacteria and the phytoplankton.
This work is devoted towards the development of such generalized, generic
(i.e., non site-specific) models.
The model utilizes four state variables: heterotrophic bacteria carbon,
phytoplankton carbon, dissolved organic carbon (DOC) and particular organic
carbon (POC). Phytoplankton release DOC upon cell death and lysis (extra-
cellular release is not included). The DOC pool receives external carbon
inputs, is reduced by the uptake of the DOC by the bacteria for growth (with
suitable inclusion of the bacterial yield) and is replenished by bacterial
cell lysis. The POC pool also receives external input and a fraction of the
recycled phytoplankton cells. Phytoplankton growth and death kinetics are
included. Primary production is the internal carbon source. The model
currently does not include any predation by zooplankton of the phytoplankton
or by protozoa of the bacteria. In addition, nutrient limitation of the
primary production and any organic carbon substrate limitation on bacterial
growth are not included. A simple completely mixed body of water is used.
Finally, a steady state condition is assumed, equivalent to a type of overall
summer, average condition.
The model indicates that the secondary bacterial carbon production is
related linearly to the primary production (net after settling loss) and to
52
-------
the external carbon input. The slope of the line relating the secondary and
primary production provides an estimate of the global bacterial yield coeffi-
cient. The published empirical relationships between primary and secondary
production indicate a yield coefficient of about 0.2, which is consistent with
direct measurements of yield.
The model has been successfully calibrated to coupled measurements of
bacterial and phytoplankton biomass that have been summarized by Cole et al.
It is seen that the model indicates a plateau in the production of bacterial
biomass as phytoplankton biomass increases. This is a consequence, in this
model, of the reduction in primary production (the carbon input for the
bacteria) as the biomass increases due to the effect of self-shading.
The model also indicates that as the allochthonous carbon source
increases, the dependence of the bacterial biomass and production on the
phytoplankton decreases and eventually becomes essentially independent of the
phytoplankton. A system such as the Potomac estuary which has a relatively
high carbon input (approximately 3000 mg C/m2-d) would be expected to show
little relationship between bacterial numbers and phytoplankton biomass.
This generic model has helped to elucidate some of the reasons for the
observed empirical relationships repeated by others between bacterial biomass
and production and phytoplankton biomass and production. The shallow slope
between bacterial and phytoplankton biomass is attributed to first approxima-
tion to the non-linear interaction of available light for primary production
and the phytoplankton biomass. The effect of increasing allochthonous carbon
loads is to shift the entire relationship of bacterial biomass and production
upward until at high loading (e.g., 3000 mg C/m2-d), the bacterial abundance
becomes independent of the phytoplankton biomass.
FUTURE WORK
We are in the process of compiling and analyzing the available data on
substrate and nutrient limitation of bacterial growth. We will evaluate these
data in a manner similar to that done for growth yield, so as to establish
appropriate ranges for the various substrates and nutrients.
Additional work will be conducted to determine the sensitivity of the
steady-state bacteria-phytoplankton model to the various parameters as well as
a more detailed examination of the individual water bodies from which the
original data set of Cole et al. was drawn. Finally, an expansion of the
model to include the effect of bacterial biomass on nutrient recycling and the
subsequent impact on phytoplankton biomass is planned.
We have recently obtained field data from the Chesapeake Bay from Tuttle
et al. which we will begin to analyze in the hopes of establishing a calibra-
tion data set to rigorously evaluate our modeling framework.
Laboratory experiments to examine bacterial uptake of phytoplankton
exudate are being conducted at the Gulf Breeze lab. These data will be
analyzed and further experiments designed to establish bacterial growth
53
-------
kinetics on the labile and refractory components of the natural carbon
sources.
54
-------
SESSION IV
GENE EXCHANGE
55
-------
-------
TRANSPOSABLE GENE-ACTIVATING ELEMENTS IN Pseudomonas ceoacia
T.G. Lessie, A. Byrne, A. Ferrante, and M.S. Wood
Department of Microbiology
University of Massachusetts
Amherst, MA 01003
INTRODUCTION
We have identified a family of transposable gene-activating elements in
Pseudomonas ceoacia which appears to play an important role in the evolution
of catabolic pathways in this bacterium. We believe that the extraordinary
degradative capacity and adaptability of P. ceoacia is related to the
abilities of these elements a) to promote frequent genomic rearrangements and
b) to increase the expression of neighboring genes. Insertion elements were
isolated on the basis of their capacity to turn on the expression of foreign
genes introduced on broad-host-range plasmids and recruit them for novel
catabolic functions. Other experiments have demonstrated that P. ceoacia
insertion sequences activate gene expression in such diverse bacteria as:
Acetobacter xvlinum. Zvmomonas mobilis. Pseudomonas aeruoinosa. Pseudomonas
fluorescens. Acinetobacter calcoaceticus. and Escherichia coli. Thus these
elements may have wide application in engineering different bacteria for
specific purposes of bioremediation.
To gain information about the mechanism of insertion-sequence-dependent
gene activation and about the factors governing element transposition we are
focusing on the detailed characterization of three insertion sequences:
IS402, IS406, and IS407. We have determined the complete nucleotide sequences
of all three elements, and are attempting to define regions of each important
for transposition and for the activation of neighboring genes. We are
particularly interested in determining whether transposition is triggered in
response to environmental stress and whether the element-dependent activation
of neighboring genes is due to the presence of outwardly-directed promoters at
their ends.
METHODS
Our approach has involved isolation of gene activating elements by
selection for increased expression of the bla and lac genes of plasmids pRPl
and pGC91.14. Activation of the pRPl bla gene by upstream insertion of P.
ceoacia IS elements increased 0-lactamase levels sufficiently to restore
growth of Bla" mutants on penicillin. Likewise* activation of the lac genes
of pGC91.14 conferred ability to utilize lactose.
We have cloned the activated bla and lac genes and their associated
gene-activated elements into high copy number £. coli vectors, and determined
the sites of element insertion by restriction mapping and nucleotide sequence
analysis. For the latter experiments we used Bluescript vectors and
constructed sets of nested deletions extending into the IS-element-containing
DNA inserts. Supercoiled DNA was sequenced by the dideoxynucleotide chain
termination procedure using Sequenase (modified T7 DNA polymerase) to extend
57
-------
oligonucleotide primers. Nucleotide sequence data were analyzed using
programs supplied by the Computer Genetics Group of the University of
Wisconsin.
RESULTS AND DISCUSSION
We have determined the complete nucleotide sequences of the bla gene
activating element IS402 (0.9-kb) and the lac gene-activating elements IS406
(1.3-kb) and IS407 (1.2-kb). These three elements had low guanosine +
cytosine (G+C) contents (between 59 and 64% G+C) compared to total P. ceoacia
DNA (71% G+C). Each of the elements had terminal inverted repeat sequences,
and produced short duplications of target DNA at their sites of insertion,
consistent with their designation as insertion sequences. Pertinent
characteristics of IS402, IS406, and IS407 are summarized in the following
table.
Characteristics of IS402, IS406, and IS407
Feature
IS402
Element
IS406
IS407
Duplicated target DNA
Length of element
Percent guanosine + cytosine
Terminal inverted repeats
Major open reading frames(ORFs)
Number
Sizes
3 bp
914 bp
61%
21 bp
3
633 bp
441 bp
264 bp
7 bp
1,368 bp
64%
41 bp
1,164 bp
363 bp
255 bp
4 bp
1,256 bp
59%
12 bp
4
549 bp
282 bp
261 bp
We have compared the nucleotide sequences of the above three elements
with those compiled in the EMBL and Genbank collections. Neither IS402 nor
IS407 exhibited homology with known insertion sequences. IS406 on the other
hand was clearly related to members of the Tn21 family of transposons (Tn21,
TnSOl, and Tn2501). Homology was due primarily to a common 10-bp core region
within the terminal inverted repeats of the elements.
Having determined the potential coding sequences within the three
elements it should now be possible to define which specify functions related
to transposition. We are in the process of constructing derivatives of each
of the elements carrying drug resistance markers. For this purpose we are
58
-------
attempting to Insert tetracycllne and trimethoprim resistance genes at sites
located outside the major ORFs. This will allow us to monitor their transpo-
sition and subsequently determine the effects of deleting or otherwise
altering each of the ORFs on this process.
Sequence analysis of IS407 revealed the presence of an outwardly-
directed £. coli-like promoter at one of its ends. This promoter is
presumably responsible for the activation of neighboring genes. Insertion of
IS406 upstream of the lac genes of the Tn951 element on pGC91.14 created a new
£. col11-1 ike promoter. We are in the process of carrying out primer extension
experiments to define if the IS406- and IS407-associated promoters are active
in vivo. Similar experiments will be carried out with IS402 to determine if
its capacity to activate gene expression is related to the presence of one or
more mobile promoters. This element is of particular interest, since it has
been show to activate gene expression in both orientations.
ANTICIPATED FUTURE WORK
One of our major goals is to examine the influence of various factors on
the transposition of the P. ceoacia insertion sequences. We are particularly
interested in determining whether environmental stress triggers transposition.
Our aim is to resolve the paradox of how the bacteria maintain a relatively
constant genotype vis a vis the marked capacity to undergo insertion-sequence-
dependent genomic rearrangements. Our working hypothesis is that events such
as insertion sequence-dependent inversions, deletions, and replicon fusions
are subject to regulation by some sort of global control system. The con-
struction of marked IS elements will permit us to monitor transposition under
different conditions and resolve whether or not it increases in response to
various environmental stimuli such as nutrient deprivation, uv-light, heat
stress, and exposure to toxic agents.
We are also initiating experiments to gain information about the
structure of the P. cepacia chromosome. We are particularly interested in
determining whether the elements we have identified are located in AT-rich
islands of the chromosome. We plan to construct macrorestriction maps of the
chromosomes of P. cepacia strains 249 and 64 using alternating field electro-
phoresis to resolve large DNA fragments. We have on hand a number of strains
whose mutant phenotypes appear to be due to large deletions. These should
prove useful in ordering fragments and determining the locations of particular
genes.
PUBLICATIONS
Lessie, T.G., M.S. Wood, A. Byrne, and A. Ferrante. 1990. Transposable gene
activating elements in Pseudomonas ceoacia. p 279-291 In S. Silver, A.M.
Chakrabarty, B. Iglewski, and S. Kaplan (eds.), Pseudomonas:
Biotransformations, Pathogenesis, and evolving biotechnology. American
Society for Microbiology, Washington, DC. In Press.
59
-------
Wood, M.S., C. Lory, and T.G. Lessie. 1990. Activation of the lit genes of
Tn951 by insertion sequences from Pseudomonas ceoacia. J. Bacteriology.
April 1990.
Wood, M.S., A. Ferrante, A. Byrne, and T.G. Lessie. 1989. Characterization
of transposable gene-activating elements from Pseudomonas cepacia. Abstr.
Annual Meet. Amer. Soc. Microbiol. H-143:193.
60
-------
PARASEXUALITY AS A MECHANISM FOR GENE TRANSFER
IN Colletotrichum qloeosporioides
Rosy J. Chacko, G.J. Weidemann, and D.O. TeBeest
Department of Plant Pathology
University of Arkansas
Fayetteville, AR 72701
INTRODUCTION
The form-genus, Colletotrichum, includes several species that cause
anthracnose on a variety of plants. Some of the parasitic species are being
investigated as potential mycoherbicides by researchers around the world.
Several members of this form-genus lack a known sexual stage. Parasexual
processes have long been considered as a possible mechanism for natural
diversity among conidial fungi.
Colletotrichum qloeosporioides (Penzig.) Sacc. f. sp. aeschvnomene (Cga)
was registered in 1982 as the bioherbicide Col lego. It is being used success-
fully in the rice and soybean growing regions of Arkansas to control the
leguminous weed, Aeschvnomene virqinica (L.) B.S.P. (northern jointvetch).
The current project is designed to help understand the potential for genetic
exchange between Coll ego and related Colletotrichum species under natural
conditions. The principal objectives are the following:
o Determine the occurrence of genetic exchange between laboratory
generated auxotrophic and or toxin resistant strains^of Cga.
o Explore the possibility of inter-strain exchanges between Cga and
related anamorphic forms such as C. qloeosporioides f. sp. .lussiaeae
(Cgj) and C. malvarum (Cm).
o Estimate the rate of heterokaryosis, recombination frequencies and
genetic stability of the strains generated by somatic fusion.
METHODS
Field isolates Cga 662 and Cgj 053 were mutagenized with
nitrosoguanidine to develop auxotrophic and/or toxin resistant laboratory
strains. Complementation studies were conducted by plating spores of
compatible mutant strains on Fries minimal media at a 1:1 ratio. For the
initial experiments, a total of 2 x 105 spores was seeded on each plate; in
other tests, the spore concentrations were reduced to 1 x 105 or 2 x 10 .
Spore suspensions of isolates that did not germinate on minimal media were
supplemented with small amounts of their nutritional requirements to enhance
spore germination.
Attempts were made to force heterokaryosis between complementary strains
of Cga as well as strains of Cga and Cgj. Heterokaryotic colonies developed
as crusty, slow-growing sectors on minimal media plates in 3 to 5 weeks.
Presumptive heterokaryons from successful pairings were transferred to fresh
61
-------
minimal plates for continued growth and analysis. Spores collected from
heterokaryotic colonies were tested on minimal media as well as minimal media
supplemented with all possible nutritional requirements to determine segrega-
tion patterns and to select possible diploids or recombinant genotypes.
Northern jointvetch seedlings were inoculated with conidial suspensions
to determine pathogenicity of mutant Cga strains and to conduct preliminary in
vivo tests.
RESULTS AND DISCUSSION
Heterokaryotic colonies were formed by 3 out of 11 compatible pairings
of Cga mutants. The successful pairings are the following:
710 (pur-) x 712 (leu-)
846 (met- or cys-) x 712 (leu-)
885 (pur-, met-) x 881 (leu-, ino-)
Compared to the parental strains, the heterokaryons exhibited slower growth
rates, altered morphology and limited sporulation. Most pairings yielded
heterokaryons at a rate of 1 x 10"5 although some pairings yielded up to 3 x
10"4 heterokaryotic colonies. Spore analysis of sporulating heterokaryons
yielded only parental genotypes.
Heterokaryons from 710 x 712 pairings segregated as 90%, isolate 710 and
10%, isolate 712. Heterokaryons from 846 x 712 pairings and 885 x 881
pairings also yielded both parental isolates in most cases although the
proportion of parental genotypes recovered from individual ..heterokaryons was
variable. Heterokaryotic isolate R119 from 885 x 881 pairing was used to
conduct large scale spore dilution tests on minimal media. Although 1 x 10
spores were plated on minimal media, no diploids or recombinants were
observed.
Plant inoculation tests indicated that strains 710, 712 and the hetero-
karyons from 710 x 712 pairings were non-pathogenic. Strain 846 is pathogenic
to northern jointvetch. Preliminary in vivo pairings of 846 x 712 yielded
only 846 spores upon re-isolation from lesions.
Inter-strain pairings of compatible Cga and Cgj mutants were also
conducted. Although eight sets of Cga x Cgj pairings were tested, hetero-
karyosis was not observed in any of the pairings.
Studies conducted so far suggest that genetic exchange between mutant
Cga strains is limited. Heterokaryons from successful pairings show reduced
growth and sporulation suggesting reduced fitness under natural conditions.
ANTICIPATED FUTURE WORK
Intra-strain complementation studies conducted so far, used Cga mutants
derived from strain 662, originally isolated from Arkansas. Additional
strains, isolated from Louisiana, are being mutagenized at the present time.
Intra-strain complementation studies will be repeated using compatible mutants
62
-------
derived from Cga isolates of different geographic locations to determine
whether intra-strain incompatibility exists. If such pairings show higher
rates of heterokaryosis spore analysis will be repeated to determine
recombination tendencies. Cga mutants from the Louisiana collections will
also be paired with compatible Cgj mutants to study interstrain
heterokaryosis.
PUBLICATIONS
Weidemann, G.J., D.O. TeBeest, and R.J. Chacko. 1989. Genetic exchange
through heterokaryosis and mitotic recombination in Colletotrichum species.
Review of Progress In Biotechnology-Microbial Pest Control Agent Risk
Assessment. EPA/600/X-89/130. pp. 130-132.
Chacko, R.J., G.J. Weidemann, and D.O. TeBeest. 1989. Heterokaryosis in
Colletotrichum qloeosporioides f. sp. aeschvnomene. Mycol. Soc. Amer.
Newsletter 40(1):25 (Abstr.).
Chacko, R.J., G.J. Weidemann, and D.O. TeBeest. 1989. Heterokaryosis among
subspecies of Colletotrichum qloeosporioides. Phytopathology 79:1208.
63
-------
STUDIES ON CONJUGAL TRANSFER OF PLASMIDS
FROM GEMS TO INDIGENOUS AQUATIC BACTERIA
Tamar Barkay
U.S. EPA
Environmental Research Laboratory
Gulf Breeze, FL 32561
INTRODUCTION
The horizontal transfer of recombinant DNA from introduced species to
indigenous microbes may result in the establishment of new genes in organisms
that are better fitted to survive and reproduce in the environment. Such
events may have harmful consequences if the foreign genes are expressed and
alter the ecology of the affected microbial community. The well characterized
bacterial mer operon has been used to study how the spread of mer genes
affects the response of aquatic microbial communities to Hg2* stress.
METHODS
The effect of conjugal transfer of the mer operon of transposon 501
(mer(501)) on acclimation of aquatic microbial communities to Hg could be
delineated by comparing the response of communities to which mer(Tn501) is
introduced on a conjugal plasmid with that of communities to which mer(501) is
introduced on a tra negative mutant of the same plasmid. A role for conjugal
gene transfer in the ecology of Hg2* exposed communities would be suggested by
increased rates of acclimation in the community where gene.transfer occurred
in comparison with the community where conjugation is prohibited. These
experimental designs necessitated the development of approaches to describe
and measure acclimation of aquatic communities to Hg2* and the development of
appropriately marked strains to serve as genetically engineered microorganisms
(GEMs).
RESULTS AND DISCUSSION
Acclimation of Aquatic Microbial Communities to Hq2*
A series of microcosm studies showed that: (a) exposed communities
responded after a lag period, lasting between 6 and 12 hours (depending on the
source of the water sample), by volatilizing Hg2*; (b) the major product of
volatilization was identified as Hg°; (c) bacteria were the actively volatil-
izing component in the community; (d) this response was of ecological signifi-
cance mostly in saline environments because of strong non-biological volatil-
ization of Hg2* in freshwater environments; (e) the mer operon of TnSOl is the
predominant gene mediating these response in aquatic communities although in
many cases, most noticeably in bacteria from saline environments, the homology
with mer genes was detected only under low stringency of hybridization. These
microcosm findings were confirmed by the characterization of microbial
communities from a mercury polluted stream in the vicinity of Oak Ridge,
64
-------
Tennessee. Thus, a model GEM consisting of fflejr(Tn501) in a salt tolerant
donor strain 1s appropriate to study the role of conjugal gene transfer on the
ecology of mercury stressed microblal communities in saline environments.
Construction of mer Plasmids and Donor Strains
1. Plasmids
The wide host range IncW plasmid R388 was selected as a model for a
conjugative plasmid. Insertion of a restriction fragment containing
mer(Tn501), but not the transposition functions, generated plasmid pEPA83. In
addition, a 350 bps plant ONA was cloned in pEPA83 to enable detection of
transconjugants over indigenous Hg2+ resistant organisms. A non-conjugative
derivative of pEPA83 is plasmid pEPASl that was prepared by cloning the mer
fragment and the plant DNA into plasmid pR02317, a derivative of R388 with a
deletion of the transfer genes.
2. Donor strain
Pseudomonas fluorescens PF015 was selected as a donor strain because of
its survival in seawater (no decline in viable or direct viable counts was
noted after 7 days of incubation in sterile seawater). A conditional lethal
system consisting of the hok gene controlled by the promoter of the lower
pathway of the toluene degradation genes from the TOL plasmid is currently
introduced to the chromosome of PF015. This system was shown to effectively
kill PF015 when present on a plasmid together with a xvlS gene, and when
benzoate (0.05%) was provided in the growth medium.
ANTICIPATED FUTURE WORK
Upon completion of strain construction the following directions will be
followed: a) the transfer rates of pEPA83 to indigenous microbes will be
determined; b) the rate of acclimation to Hg2* stress after the introduction
of pEPA83 and pEPASl to estuarine and marine communities will be determined;
c) a similar approach will attempt to look at the role of plasmid
mobilization, transduction and transformation in acclimation to Hg*2.
PUBLICATIONS
Levin, M.A., R. Seidler, A.M. Bourquin, J.R. Fowle III, and T. Barkay. 1987.
EPA developing methods to assess environmental release. Bio/Technology 5:38-
Trevors, J.T., T. Barkay, and A.W. Bourquin. -1987. Bacterial gene transfer
in soil and aquatic environments: A review. Can. J. Microbiol. 33:191-198.
Barkay, T. 1987. Adaptation of aquatic microbial communities to Hg2* stress.
Appl. Environ. Microbiol. 53:2725-2732.
65
-------
Ogram, A., G.S. Sayler, and T. Barkay. 1987. The extraction and purification
of microbial DNA from sediments. J. Mlcrobicl. Methods 7:57-66.
Genthner, F.J., P. Chatterjee» T. Barkay, and A.M. Bourquin. 1988. Capacity
of aquatic bacteria to act as recipients for plasmid DNA. Appl. Environ.
Microbiol. 54:115-117.
Barkay, T., and G.S. Sayler. 1988. Gene probes as a tool for the detection
of specific genomes in the environment. Aquatic Toxicology and Hazard
Assessment: 10th Volume ASTM STP 971, W.J. Adams, G.A. Chapman, and W.G.
Landis (eds.), American Society for Testing and Materials, Philadelphia, PA.
pp. 29-36.
Barkay, T., and P. Pritchard. 1988. Adaptation of aquatic microbial communi-
ties to pollutant stress. Microbiol. Sci. 5:165-169.
Liebert, C., and T. Barkay. 1988. A direct viable counting method for
measuring tolerance of aquatic microbial communities to Hg2>. Can. J.
Microbial. 34:1090-1095.
Barkay, T., D. Chatterjee, S. Cuskey, R. Walter, F. Genthner, and A. Bourquin.
1989. Bacteria and the Environment. In: A Revolution in Biotechnology, J.
Marx (ed.). Cambridge University Press, Cambridge, New Rochelle, Melbourne,
Sydney, pp. 94-102.
Barkay, T., C. Liebert, and M. Gillman. 1989. The environmental significance
of the potential for mer(Tn21) mediated reduction of Hg2* to Hg° in natural
water. Appl. Environ. Microbiol. 55:1196-1202.
Barkay, T., C. Liebert, and M. Gillman. 1989. Hybridization of DNA probes
with whole community genome for detection of genes that encode microbial
responses to pollutants: mer genes and Hg2* resistance. Appl. Environ.
Microbiol. 55:1574-1577.
Summers, A.O., and T. Barkay. 1989. Metal resistance genes in the environ-
ment. In: Gene Transfer in the Environment, S. Levy and R. Miller (eds.).
McGraw-Hill Publishing Company, New York. pp. 287-308.
66
-------
NATURAL PLASMIO TRANSFORMATION IN A MARINE Vibrio
John H. Paul,1 Marc E. Frischer,1 Made H. Jeffrey,1'2 and Gregory J. Stewart3
Department of Marine Science1
University of South Florida
St. Petersburg, FL 33701
Current address: U.S. EPA2, Gulf Breeze, FL 32561
Biology Department3
University of South Florida
Tampa, FL 33201
INTRODUCTION
Natural transformation is the process of uptake of extracellular DNA and
the maintenance and expression of genes encoded in this DNA. It is distinct
from transformation induced by chemical, enzymatic, or electroporation
techniques in that it is a natural physiologic process of the cell. Where
known, the genes for this process are chromosomally encoded, making it a
heritable cellular trait. Our original EPA cooperative agreement, entitled
"The Role of Extracellular DNA in the Dissemination of Recombinant DNA in
Aquatic Environments," asked this fundamental question: Can extracellular DNA
be a conduit for gene transfer in aquatic environments? Our approach to
answering this question was first to determine the persistence of extra-
cellular DNA in aquatic environments and secondly to determine if aquatic
(particularly marine) bacteria were transformable.
To determine the survival of extracellular DNA, we have examined the
turnover time of dissolved DNA in the water column of oligotrophic and
eutrophic freshwater environments, and found similar turnover times (ca. 10
h), irrespective of nutrient status. We have shown that genetically engin-
eered bacteria released into aquatic environmental microcosms produced
extracellular DNA and that natural microbial populations took up released
material. By studying DNA uptake mechanisms of natural populations of
estuarine and freshwater bacteria we have learned that these bacteria have
uptake sites specific for nucleic acids capable of recognizing molecules as
short as dinucleotides, and distinct from nucleotide, nucleoside, nucleobase,
phosphomonoester, and inorganic phosphate binding sites. These results
indicate that natural populations of bacteria have evolved mechanisms for
uptake of extracellular DNA, which is abundant in marine and freshwater
environments. We have also developed methodologies to concentrate and purify
dissolved DNA for direct probing and have recently detected the ribulose 1,5-
biphosphate carboxylase gene in dissolved DNA by the polymerase chain
reaction.
The purpose of the current study was to investigate natural transforma-
tion with plasmid and chromosomal DNA in marine bacterial isolates and natural
microbial populations.
67
-------
METHODS
Filter transformation assays were performed with logarithmically-
growing bacterial cultures as previously described. Sediment column
transformation microcosm experiments were performed in 10 cc disposable
syringes loaded with 3 to 5 cc subtidal sediment from North Shore Park, St.
Petersburg, Florida. Plasmid pEPA53 was a gift from Dr. Steve Cuskey, U.S.
EPA, Gulf Breeze, Florida. All plasmid constructs were made by standard
methods, and plasmid multimers were made in vitro by single restriction enzyme
digestion at a unique site followed by ligation.
RESULTS AND DISCUSSION
Three types of transforming DNA have been investigated: a) "homo-
logous" chromosomal DNA; b) non-homologous plasmid DNA; and c) non-homologous
transposon-containing plasmid DNA.
In every form of natural transformation described to date there is a
requirement for homology between transforming DNA and other DNA in the cell
such that homologous recombination mechanisms might incorporate the exogenous
sequences into one of the cellular replicons. Therefore, to study transforma-
tion with homologous DNA, transformation with chromosomal DNA has been
performed with Pseudomonas stutzeri. We have shown that Pseudomonas stutzeri
strain ZoBell is an excellent model for gene transfer in marine sediments.
This transfer process is DNase I sensitive, indicating natural transformation
as the mode of transfer. We have also shown that transformation occurs in
non-sterile sediments to which recipient cells and DNA are added. Further-
more, we have shown that naturally-occurring populations of marine sediment
bacteria can be transformed to antibiotic resistance if those sediments are
provided with DNA preparation made from a heterogenous population of spontan-
eous antibiotic resistant mutants from sediments. Again, the sensitivity of
this process to DNase I indicates natural transformation.
We have also shown a correlation between the amount of DNA added to
sterile sediments and the frequencies of transformation for recipient cells.
We have shown that marine sediments tightly bind DNA in a form that resists
transfer of the markers to recipients when those sediments are loaded with low
concentrations of DNA. If, however, the sediments are first loaded with calf-
thymus DNA we find transformation detected at lower concentrations of
resistant-marker DNA, suggesting that the sterile sediments first bind DNA in
tight association that precludes uptake by competent cells. Only after these
sediments are "saturated" with DNA is that DNA available for transformation,
and at that point one detects a correlation between DNA concentration and
frequency of transformation.
Although rifampin chromosomal DNA preparations from marine bacteria
might contain sequences homologous to the endogenous microbial DNA,
verification of transformation is difficult owing to the lack of identifiable
sequences in this type of DNA. Specific probes to the transforming DNA which
arises from spontaneous mutation cannot be easily made that will hybridize
with transformants. Therefore, we have employed broad range host plasmid DNA
with specific markers as transforming DNA. To insure homology, we constructed
68
-------
multimeric forms of the plasmid, such that internal homology is supplied with
the transforming DNA. We have discovered how to make multimers of the broad
host range plasmid, pKT230, and used these to transform the isolate, Vibrio
DI-9. Verification of transformation was obtained by molecular probing with
the kanamycin resistance gene probe derived from the plasmid. Frequencies of
transformation were quite low (10 yrecipient), and transformation was
detected in filter transformation and sediment column assays (see below).
Although not funded, we have performed studies with this isolate to gain basic
knowledge of marine bacterial transformation mechanisms. Additionally, these
studies will enable us to have a positive control organism to seed or spike
environmental samples in field transformation assays. Curing the transformed
Vibrio DI-9 of its plasmid resulted in the selection of a high frequency of
transformation (HfT) phenotype. This HfT strain transformed with pkT 230
multimers at frequencies of ca. 6 x 10'8 of 1.5 times greater than the wild
type. However, when we transformed the HfT strain with multimers of a broad
host range plasmid containing Tn5 (pQSRSO), transformation frequencies of 10"
to 10"* were obtained, or 2 to 4 orders of magnitude greater than the wild
type. Verification of transformation has been obtained with a probe
construction from a portion of the pQSRSO plasmid. The nature of the physio-
logic change which results in the HfT phenotype is currently being investi-
gated. It appears that spontaneous Hft mutants that arise in the Vibrio
microbial population are selected during transformation; but more work is
needed. Additionally, the role of Tn5 in this high frequency phenotype is not
understood. Perhaps the transposition genes allow illegitimate recombination
and thereby circumvent the need for homologous sites for recombination.
A fundamental problem in studying transformation in the environment is
that it is an extremely infrequent process. Even with pure cultures,
frequencies are on the order of 10"4 to 10~9 transformants/recipient. Such
frequencies preclude detection of transformants by direct (i.e., epifluor-
escence) procedures, since under best conditions, 1 in 10,000 cells viewed
would be a transformant, and under the worst, 1 in 109 cells. We have
attempted to detect transformants in natural populations directly by use of
thymidine incorporation. Thymidine incorporation assays were performed with
antibiotic selective conditions.
Studies were conducted in estuarine, coastal, oceanic, and offshore
water column environments, the coral surface microlayer, and sediment samples
from the Bahamas. In 21 of 31 experiments, transformation was indicated by
phenotypic expression by the microbial community as indicated by [3H]thymidine
incorporation in the presence of specific antibiotics. These were compared to
controls which received calf thymus DNA in the place of transforming DNA.
Verification of transformants by molecular probing of DNA extracted from these
communities could not be obtained. We have therefore modified our assay to
detect transformants by liquid selection and probing of community DNA prepara-
tions. A second approach is the isolation of mRNA to detect expression of the
transforming DNA genes. These studies are currently underway.
Lastly, we are investigating the possibility of colorimetric detection
of transformants by using genes that encode for colored product formation. We
have used the x^LE gene of the toluene degrading plasmid, pWWO, which encodes
for catechol dioxygenase. This enzyme converts catechol to the bright yellow
69
-------
2-hydroxvmucon1c semi aldehyde. There are no cofactors for this reaction and
levels of expression are extremely high. We have subcloned this gene from the
plasmld pEPA53 into both pQSRSO and pKT230 to yield pYELl-37 and pYEL2,
respectively. Both of these plasmids have been used to successfully transform
the cured marine Vibrio strain, WJT-l-C, and transformants were detected both
by yellow color of colonies and by molecular probing. We have not as yet made
multimers of these plasmids, but both contain two copies of the xvIE gene, and
thus may contain sufficient homology for transformation. We have also
obtained V1013, a pKT230 derivative containing xvIE and temperature-sensitive
lambda promoters, which we are purifying for use in transformation studies.
Studies with laboratory isolates indicated that water column or liquid
transformation occurred much less frequently than filter or sediment transfor-
mation. We have employed sediment columns to demonstrate transformation with
marine isolates with homologous rifampin-resistant chromosomal DNA and plasmid
multimers. We have also detected what appears to be transformation of natural
sediment bacterial populations with both chromosomal ONA and the Tn5-contain-
ing plasmid, pQSRSO, by extracting sediment bacteria and filtering them on
sterile membrane filters. Frequencies of transformation with rifampicin-
resistant chromosomal DNA were only 1/2 order of magnitude greater than with
calf thymus DNA, but considerably greater with pQSRSO multlmers (1-1.5 orders
of magnitude greater). Apparent transformation of the ambient microbial
population was sensitive to DNase I treatment. Attempts to verify transform-
ants have to date been unsuccessful. We plan to use the colorimetric detec-
tion method described above with natural populations, shortly.
PUBLICATIONS
Paul, J.H., M.F. DeFlaun, and W.H. Jeffrey. 1988. Mechanisms of DNA utiliza-
tion by estuarine microbial populations. Appl. Environ. Mlcrobiol. 54:1682-
1688.
Paul, J.H., and S.L. Pichard. 1989. Specificity of cellular DNA-binding
sites of microbial populations in a Florida reservoir. Appl. Environ.
Microbiol. 55:2798-2801.
Paul, J.H., W.H. Jeffrey, A.W. David, M.F. DeFlaun, and L.H. Cazares. 1989.
The turnover of extracellular DNA in eutrophic and oligotrophic freshwater
environments of southwest Florida. Appl. Environ. Microbiol. 55:1823-1888.
Paul, J.H., and A.W. David. 1989. Production of extracellular nucleic acids
by genetically altered bacteria in aquatic microcosms. Appl. Environ.
Microbiol. 55:1865-1869.
DeFlaun, M.F., and J.H. Paul. 1989. The concentration and detection of
exogenous gene sequences in dissolved DNA from aquatic environments. Microb.
Ecol. 18:21-28.
Jeffrey, W.H., J.H. Paul, and G.J. Stewart. Natural transformation of a
marine Vibrio sp. by plasmid DNA. Appl. Environ. Microbiol. (submitted).
Stewart, G.J., and C.D. Sinigalliano. Binding exogenous DNA to marine
sediments and the effect of DNA/sediment binding on natural transformation of
70
-------
Pseudomonas stutzeri strsln ZoBall 1n sediment columns. Appl. Environ.
Mlcrobiol. (submitted).
Stewart, 6.J., and C.D. SinigalHano. Detection of horizontal gene transfer
by natural transformation in native and Introduced species of bacteria in
marine and synthetic sediments. Appl. Environ. Mlcrobiol. (submitted).
Stewart, G.J., and C.D. S1nigall1ano. Detection and characterization of
natural transformation in the marine bacterium Pseudomonas stutzeri strain
ZoBell. Arch, of Microbiol. (in press).
SinigalHano, C., and G.J. Stewart. Natural transformation occurs between
soil and marine isolates of Pseudomonas stutzeri. J. Bacteriol. (submitted)
71
-------
GENETIC TRANSFER IN FRESHWATER ENVIRONMENTS
Robert V. Miller,1 Tyler A. Kokjohn,1 and Gary S. Sayler2
Department of Biochemistry and The Program in Molecular Biology1
Loyola University Chicago
Maywood, IL 60153
Department of Microbiology and The Program in Ecology2
University of Tennessee
Knoxville, TN 37932
INTRODUCTION
The proposed release of genetically engineered microorganisms (GEMs) to
attack problems of environmental pollution and for agricultural uses has
raised concerns over the ultimate fate of these organisms and their engineered
genes. At present, efforts are being made to limit the range of ecological
niches available to GEMs. Horizontal gene transfer of recombinant molecules
to natural populations of bacteria would nullify these efforts.
We have been using Pseudomonas aeruqinosa as a model organism to study
genome alteration in freshwater microbial populations and have demonstrated a
significant potential for horizontal gene transmission by both transduction
and conjugation. The finding that transduction can occur in natural environ-
ments is significant as this mechanism of gene transfer has been virtually
ignored in both the design and preliminary testing of GEMs for environmental
release. In this report, we will discuss general protocols, for measuring
transduction in aquatic habitats and describe some of our current research
designed to address a number of questions important to the evaluation of the
significance of transduction in determining the fate of novel gene sequences
in the environment.
METHODS
Our investigations of horizontal gene transfer are conducted in both
laboratory and field formats and can be divided into three stages: a) The
potential for transmission of specific DNA sequences is evaluated by standard
genetic protocols; b) Laboratory experiments are conducted ir microcosms to
evaluate protocols to be tested in situ; c) Field trials are conducted in
biological containment chambers incubated at the freshwater field site.
The potential for chromosomal and plasmid transduction was evaluated
using a variant of the generalized transducing phage F116 originally
designated DS1. For studies of plasmid transfer, we utilized the Tra" Mob"
plasmid Rmsl49. Plasmid transduction was confirmed by analysis of the
restriction endonuclease digestion patterns of extrachromosomal DNA isolated
from the presumptive transductants.
Natural bacteriophages were isolated from lake water collected at
several test sites. Samples were used either directly or concentrated.
Phages were identified by their ability to plaque on one or more indicator
72
-------
strains. Phages selected for further studies were purified by glycerol
gradient ultracentrifugatlon. P. aeruqlnosa strains were Isolated from lake
water by means of selective madia and Pseudomonas Isolation Agar and further
characterized as appropriate. In some experiments, mixed populations of
either bacteriophages 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 Scien-
tific BioFlo Chemostat1. Culture medium consists of Pseudomonas Minimal
Medium Salts supplemented with yeast extract to provide a source of carbon and
energy.
RESULTS AND DISCUSSION
Horizontal Gene Transfer Mediated by Transduction
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) environmental induction of bacterio-
phages from a lysogenic DNA donor bacterium, and c) environmental induction of
bacteriophages from a lysogenic recipient bacterium. In laboratory simula-
tions and in field trials at high cell densities, transduction of both plasmid
and chromosomal DNA was observed in each of these models. In field trials
where cell concentrations were reduced to environmentally appropriate cell
densities (104 CFU/ml), transduction of plasmid DNA was detected only when the
recipient bacterium was a lysogen.
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 the lytic infection of the donor must
absorb and transfer DNA to the
lysogen. Hence, environmental lysogens serve as both efficient sources of
transducing phages and as viable recipients for transduced DNA.
Donor-to-recipient ratio was found to affect the level of transductants
recovered. The frequency of transduction was independent of donor-cell
concentration suggesting that the efficiency of lytic infection and production
of phages is unaffected by cell density. Transduction frequency was dependent
on the relative abundance of recipient cells and was maximized at an inter-
mediate donor-to-recipient ratio (20:1). This relationship may result from
differences in the phage-to-bacterium ratio (PBR) in the various test chambers
due to the relative concentrations of lysogens (recipients) and non-lysogens
(donors).
Transduction was observed in both the absence and presence of the
natural microbial community. The presence of the natural community resulted
in a rapid decrease in the recovery of both transductants and introduced
donors and recipients. Where observed, the ratio of transductants to parental
73
-------
strains was similar in sterile and natural lake water. The natural community
appears to reduce our ability to detect transductants and not the absolute
potential for gene transfer by this mechanism.
Chromosomal transduction was measured in microcosms and field trials
between a lysogen and non-lysogen and between two lysogens. Transduction of
single chromosomal alleles was observed in all cases. In chambers containing
a lysogen and a non-lysogen, each strain could act as both a donor and as a
recipient of transduced DNA. Reciprocal exchange of alleles was also 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.
Co-transduction of linked chromosomal markers was observed at
frequencies similar to those obtained in standard laboratory transduction
protocols. However, co-inheritance of unlinked markers was not observed
indicating that multiple transduction events in a single cell are too
infrequent in natural environments to be detected by our methods.
We have shown that a number of factors have significant effects on the
detection of transductants and should be considered when designing model
systems to evaluate the potential for transduction in a natural freshwater
habitat. When these factors are taken into account, we feel that the best
qualitative test for transduction in natural environments is the measurement,
in sterilized water, of transduction of a Tra" Mob" plasmld containing at
least two selectable genetic markers. We feel this is true for several
reasons. First, our data indicate that the presence of the natural community
affects the ability to detect transductants more than the absolute ability for
transduction to take place. Second, because the plasmid contains several
genetic markers, co-transduction frequencies approach 100%. Third, the
ability to physically detect plasmid DNA in the presumptive transductants
increases the confidence that true transduction events have been detected.
A Natural Transduction Model
Our research has demonstrated the potential for horizontal gene transfer
mediated by transduction in natural freshwater habitats. The model system we
have developed is still artificial as its components are well characterized
laboratory strains. We are currently attempting to develop a system whose
components are derived directly from the environment. Only by continuing to
improve the design of our models to more accurately represent the natural
situation will we be able to assess the real potential for genetic exchange
among natural microbial populations.
EVALUATION OF THE LEVELS OF BACTERIA AND BACTERIOPHAGES AT OUR FIELD SITES--
One important question in evaluating the significance of transduction in
natural ecosystems is the determination of the occurrence of bacteriophages,
potential hosts, and lysogens in the environment. We have monitored these
parameters in one of our field sites for an eight-month period. During this
period, we found total bacterial counts to range from 10 to 10 CFU/ml and
Pseudomonas counts from 5 x 102 to 5 x 104 CFU/ml. Titers of bacteriophages
capable of plaquing on laboratory strains of P. aeruqinosa ranged from 5
74
-------
PFU/ml to 2 x 104 PFU/ml. When lysogeny was evaluated using a laboratory
strain of P. aeruqinosa as an 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 hybridiza-
tion when probed with DNA from a naturally occurring phage isolate.
ISOLATION OF NATURALLY OCCURRING TRANSDUCING PHAGES--
In an attempt to establish a natural transduction model, we have
isolated several transducing phages from our field sites. One of these, UT1,
was of particular interest and is being studied in some detail. This phage
was isolated from Fort Loudoun Lake near Knoxville, TN, where we have carried
out the majority of our vn situ transduction experiments. This phage is a
generalized transducing phage capable of mediating the transfer of both
plasmid (Rmsl49) and chromosomal DNA among P. aeruqinosa. Using standard
laboratory protocols, both chromosomal and plasmid transduction has been
detected using this phage. In addition, we have preliminary evidence that UT1
infects and transduces Pseudomonas ceoacia as well as P. aeruqinosa. The
frequency of this inter-species transduction is in the range of 10~l0
transductants/PFU. A comparison of the effectiveness of UT1 and F116L to
mediate transduction in situ will be quite informative.
INTERACTION OF BACTERIA AND BACTERIOPHAGES IN FRESHWATER ENVIRONMENTS--
For transduction to take place, interaction between the transducing
phage and its host must take place. We are investigating the potential for
such interaction in several ways. First, the ability of bacteriophages to
interact with their hosts at low cell densities and under starvation
conditions, are being examined. We found that productive phage-host
interactions can take place when cell concentrations are 10? CFU/ml or lower.
This is certainly well below the concentrations of Pseudomonas found in many
freshwater environments. In addition, productive interaction can take place
with bacteria grown and incubated in river water. While the latency periods
are usually longer, significant bursts of phage are produced. It appears that
the relative nutrient limitations experienced by bacteria in natural
freshwater environments is not a deterrent to the infection and growth of some
bacteriophages.
Second, the persistence and interaction between strains of P. aeruqinosa
newly isolated from our field site and UT1 were monitored over 45 days in lake
water microcosms. Interaction in the absence and presence of natural
bacterial and bacteriophage communities were evaluated. Under these growth
conditions, UT1 will establish a pseudolysogenic state in a significant
fraction of the bacterial community. After prolonged incubation the bacterial
density stabilized at approximately 104 CFU/ml. Phage densities stabilized
and were maintained throughout the incubation.
Third, we have been examining the kinetics of lysogeny establishment in
mixed populations of lysogens and non-lysogens. Our preliminary studies
suggest that lysogenization is relatively rapid in a growing population and
can occur even under nutritional conditions found in the natural habitat.
75
-------
PLASMID TRANSDUCTION TO A NATURAL AQUATIC COMMUNITY--
We have Initiated studies to determine the potential of the natural
microbial community present at our field sites to act as recipients of
transduced plasmid DNA mediated by bacteriophages, F116L and UT1. In
preliminary studies, cells from lake water were concentrated and used as a
recipient pool for transduction using cell-free lysates of the transducing
phages grown on a P. aeruqinosa strain containing Rmsl49 which codes for
carbenicillin (Cb) and streptomycin (Sm) resistance. While few Cbr Smr clones
were observed in lake water, the frequency of these colonies was increased as
much as 100 fold following transduction by either of the phages. This
suggests that the natural microbial community at our field sites contains
organisms capable of acting as hosts and recipients for plasmid transduction.
In the future, we hope to evaluate these processes in situ by
introducing a lysogenic plasmid donor into our containment chambers containing
the natural community as recipients. These studies require the development of
an appropriate contraselection for the introduced donor. In collaboration
with Dr. S. Cuskey (Gulf Breeze Laboratory), we are currently developing such
a suicide donor.
Modeling of The Potential for Transduction to Stabilize a Foreign Genotype
Within an Established Microbial Community
The ultimate question to be addressed in determining the effects of
environmental transduction is whether phage-mediated horizontal gene transfer
alters the probability of the permanent addition of a genetic sequence into a
natural bacterial community. We have begun a series of experiments using
continuous culture techniques aimed at addressing this question. Initially,
we are establishing a community of plasmid-free organisms (recipient) which in
some experiments are lysogens and in others are not. When the population is
in equilibrium, non-lysogenic cells containing the Tra" Mob' plasmid Rmsl49
(donor) are introduced. Total, donor, and recipient cell concentrations, the
concentration of bacteriophages, the frequency of the plasmid and the
occurrence of transductants are monitored. Comparison of communities where
transduction is either possible or not possible should allow us to develop
models for evaluation of the consequences of transduction to the fate of an
introduced novel ONA sequence.
ANTICIPATED FUTURE WORK
The development of natural models for transduction and continuous
culture studies are in the preliminary stages. They will continue for the
remainder of the project. We also will initiate studies on the effects of
suspended particles on transduction.
76
-------
REPRESENTATIVE PUBLICATIONS
Ogunseitan, Oladele A., Gary S. Sayler, and Robert V. Miller. 1990. Dynamic
interaction of Pseudomonas aeruginosa and bacteriophages in lake water.
Microbial Ecology (in press).
Miller, Robert V., and Tyler A. Kokjohn. 1990. Microbiology of recA:
Environmental and Evolutionary Significance. Annual Reviews of Microbiology
(in press).
Kokjohn, Tyler A., and Robert V. Miller. 1990. Gene transfer in the environ-
ment: transduction. In J. C. Fry and M. J. Day (eds.), Release of Genetic-
ally Engineered and Other Microorganisms. Edward Arnold, London (in press).
Miller, Robert V. 1990. Methods for the evaluation of gene transfer in the
environment; general considerations, in M. Levin, R. Seidler, and P.M.
Pritchard (eds.), Guide to Environmental Microbiology. McGraw-Hill, New York
(in press).
Miller, Robert V. 1990. Transduction in natural environments, in M. Levin,
R. Seidler, and P.M. Pritchard (eds.), Guide to Environmental Microbiology.
McGraw-Hill, New York (in press).
Saye, Dennis J., 0. A. Ogunseitan, G. S. Sayler, and Robert V. Miller. 1990.
Transduction of linked chromosomal genes between Pseudomonas aeruginosa during
incubation i_n situ in a freshwater habitat. Applied and Environmental
Microbiology 56:140-145.
Levy, Stuart B., and Robert V. Miller (editors). 1989. Gene Transfer in the
Environment. McGraw-Hill, New York. 440 pages.
Saye, Dennis J., and Robert V. Miller (editors). 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, New York.
Miller, Robert V., and Stuart 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, New York.
O'Morchoe, Susan B., 0. Ogunseitan, Gary S. Sayler, and Robert V. Miller.
1988. Conjugal transfer of R68.45 and FP5 between Pseudomonas aeruginosa in a
freshwater environment. Applied and Environmental Microbiology 54:1923-1929.
77
-------
NATURAL TRANSFORMATION IN MARINE SYSTEMS
Gregory J. Stewart
Department of Biology
University of South Florida
Tampa, FL 33620
INTRODUCTION
The major focus of this project has been to determine the potential for
horizontal gene transfer by natural transformation in the marine environment.
While Dr. Paul has concentrated on studies of natural transformation of
plasmid DNA, our laboratory has focused on the uptake and expression of
chromosomally-linked genes. The goals we have addressed are as follows:
o To determine whether natural transformation occurs in the marine
environment.
o To develop an environmental simulation that will allow us to measure
horizontal gene transfer by natural transformation.
o To identify bacteria from the marine environment that are physiolog-
ically capable of natural transformation.
o To develop one or more model systems for transformation in the marine
environment.
RESULTS AND DISCUSSION
Determination of Natural Transformation in the Marine Environment
Initially we isolated a number of marine heterotrophic bacteria, gener-
ated antibiotic resistant variants, and used these variants as sources of
transforming DNA. In this analysis we determined that a high percentage of
the marine isolates tested (16%) was capable of transformation. We know that
at least two genera of bacteria were included among these transforming
isolates, the genera Pseudomonas and Vibrio. We also assessed the potential
for transformation in both water column and in sediments and found that while
transformation could not be detected in water column simulations, it was
detectable in sediment simulations. These results suggested to us that at
least marine sediments were capable of supporting natural transformation.
Development of an Environmental Simulation for Transformation in Marine
Environments
The results of the preliminary experiments discussed above encouraged us
to focus our efforts on sediment simulations. Sediment columns were prepared
by packing sterile, 10 cc disposable syringes with 5 cc of sterile or non-
sterile sediment. Sediments were loaded first with recipient cells, allowed
to incubate for one hour and were then loaded with transforming DNA. Columns
were incubated for 18 hours at 28°C. Frequencies of transformation were
78
-------
standardized to total recipient cells. Samples were plated directly from the
column flow through, from a high salt wash of the column, and from the
bacterial populations tightly associated with the sediments. Significant
numbers of transformants were only detected in the latter fraction. Transfor-
mation frequencies were in the range of 10-100 times that for spontaneous
resistance, and transformation frequencies were reduced to spontaneous rates
by the addition of DNase to the column suggesting that the mechanism of
conversion to antibiotic resistance was mediated by natural transformation
when sterile sediments were spiked with strains of Pseudomonas stutzeri.
Frequencies of transformation were reduced 7-10 fold when recipients were
added to non-sterile sediments. These transformation frequencies were
considerably lower than those for sterile sediments, but were still
significantly higher than frequencies for spontaneous resistance. As was true
for transformation in sterile sediments, treatment of columns with DNase I
reduced frequencies of transformation to spontaneous rates.
We have also conducted research to determine the relationship of DNA
binding to sediments and the availability of the bound DNA for transformation.
In studies with sterile sediments, 2-3X more DNA was required to saturate
natural transformation than was required for filter transformation. Based on
the hypothesis that sterile sediments were capable of binding a certain amount
of DNA, we pre-loaded sediments with calf-thymus DNA, then added rifampin-
resistant, transforming DNA. Under these conditions, transformation frequen-
cies saturated at about 1.0 ug per cc sediment, a value close to that
predicted from filter transformation studies. We then loaded sediments with
different amounts of rifampin-resistant DNA, collected the flow-through and
high salt washes, and determined the total amount of DNA recovered. We also
determined transformation frequencies for recipient cells under these
conditions. These studies indicated that when DNA was added to sterile
sediments at concentrations less than 2.5 ug per cc sediment, the majority of
the DNA was retained by the sediment column and the frequencies of transforma-
tion were sub-saturating. When DNA was added to sediments at 5.0 g per cc
sediment, the majority of the DNA was recovered in the wash, the sediments
seemed to retain about 2.0-2.5 ug DNA per cc, and transformation occurred at
saturating frequencies. These studies suggest that marine sediments may have
a capacity to bind DNA at a specific concentration, and that binding of the
DNA to the sediment reduces its availability for transformation. Only when DNA
is added at a concentration sufficiently above this saturation level is it
biologically significant for natural transformation.
Development of a Model System for Natural Transformation Studies in Marine
Systems
We have focused our efforts toward this goal on one particular
bacterium, Pseudomonas stutzeri strain ZoBelU chosen because it is a marine
isolate of a species previously shown to transform at a relatively high
frequency. Our initial laboratory studies have established that this isolate
is also capable of natural transformation. We have shown that the bacterium
transforms on filters and in sediment columns, and we have evaluated a number
of environmental factors that influence induction of competence. We have also
shown that the frequencies of transformation for this strain are sufficiently
high to serve as a sensitive indicator organism for natural transformation in
79
-------
marine systems. We have established that the bacterium transforms purified,
exogenous DNA, and that it also transforms by cell contact transformation,
thus protecting the DNA from long term exposure to nucleases. We have
established that the strain transforms a series of different resistance
markers. We have been successful in growing the bacterium on a minimal medium,
thus allowing the use of auxotrophic markers as well as resistance loci.
Finally, we have demonstrated transformation of chromosomal markers between
soil and marine isolates of this bacterium. This allow us the luxury of using
many of the existing, genetically distinct strains derived from the soil
isolate. It also validates the importance of the information known for the
soil strain.
Development of a Model System for Natural Transformation in Marine
Environments
Based on the work of the past three years we feel that marine sediments
are one likely environment for gene transfer by natural transformation. We
also feel that P. stutzeri strain ZoBell offers an interesting model for
analyzing natural transformation in sediment simulations.
Other Related Results
Our work with non-sterile sediments has involved the addition of
rifampin resistant ONA from a heterogeneous population of spontaneous
resistant mutants. These mutants were isolated by direct selection, and were
derived from isolates from marine sediments. A modification of the sediment
column was employed where the sediments were gathered onto a membrane filter.
This filter was placed on artificial sea water-based medium. After the
addition of the heterogeneous DNA the filters were incubated overnight and the
bacteria were resuspended from the filters. Frequencies of transformation
were approximately three-times that of no-DNA controls (spontaneous rates).
The increased frequency was lowered to spontaneous resistance levels by
treatment with DNase I. These results were reproducible. Even though the
levels of transformation were only slightly higher than spontaneous rates, the
sensitivity of the transfer to DNase I strongly suggests that natural
transformation occurred. We are currently attempting to identify some of the
transformed species.
PUBLICATIONS
Stewart, 6.J., and C.D. Sinigalliano. Detection and characterization of
natural transformation in the marine bacterium Pseudomonas stutzeri. strain
ZoBell. Archives of Microbiology (submitted).
Sinigalliano, C.D., and G.J. Stewart. Natural transformation occurs between
soil and marine isolates of Pseudomonas stutzeri. Journal of Bacteriology
(submitted).
Stewart, G.J., and C.D. Sinigalliano. Detection of horizontal gene transfer
by natural transformation in native and introduced species of bacteria in
marine and synthetic sediments. Applied and Environmental Microbiology
(submitted).
80
-------
Stewart, G.J., and C.D. Sinigalliano. Binding of exogenous DNA to marine
sediments and the effect of OHA/sediment binding on natural transformation of
Pseudomonas stutzeri strain ZoBell in sediment columns. Applied and Environ-
mental Microbiology (submitted).
Coughter, J.P., and G.J. Stewart. 1989. Genetic exchange in the environ-
ment. Journal of Microbiology 55:15-22.
Stewart, G.J. 1989. The mechanism of natural transformation. In S. Levy and
R. Miller (eds.), Gene Transfer in the Environment. McMillan Publishing Co.
Stewart, G.J. Gene Transfer in the Environment: Transformation. In M.J.
Day and J.C. Fry (eds), Environmental Release of Genetically Engineered and
Other Microorganisms. Edward Arnold Publishing, London (submitted).
81
-------
GENE TRANSFER BY CONJUGATION, TRANSDUCTION, AND TRANSFORMATION IN SOIL
G. Stotzky
Department of Biology
New York University
New York, NY 10003
INTRODUCTION
The transfer of genetic information among bacteria by conjugation (cell-
to-cell contact), transduction (via a bacteriophage), and transformation
(uptake of "naked" DNA by an intact cell) has been demonstrated in a wide
spectrum of gram-negative and gram-positive bacteria in numerous laboratories
(i.e., in pure culture). In contrast, there is sparse information on the
occurrence of such transfer in soil and other natural habitats. Any assess-
ment of the risks related to the release of genetically modified organisms
(GMOs) to the environment for agricultural, degradation of toxic wastes, and
other purposes must consider the probability of transfer of the novel genes in
the GMOs to other microorganisms present in the recipient environment,
especially to autochthonous microbes, which are probably more adapted to the
recipient environment than the introduced GMOs. Consequently, information on
the probability and frequency of transfer of genetic information among
bacteria in soil and other natural habitats by the three methods of transferr-
ing DNA is urgently needed.
For unknown reasons, conjugation has been considered to be the major
method for the potential transfer of genetic information among bacteria in
soil and other habitats. However, transduction and transformation may be as
important as, if not more so than, conjugation as methods of gene transfer in
natural habitats. For example, the packaging of bacterial DNA in a bacterio-
phage probably represents an evolutionary survival strategy for the genetic
material, and such viruses may serve as reservoirs for exogenous genes.
Furthermore, the survival of viruses, including bacteriophages, in soils and
waters has been shown in this laboratory to be enhanced by their absorption on
clay minerals and other particulates. Consequently, novel genetic information
in an introduced GMO could be incorporated into the genetic material of a
phage and could persist in soil and other habitats longer than the GMO itself.
Such persistence would be undetected (i.e., cryptic) in the absence of a
bacterial host susceptible to infection and capable of supporting the repro-
duction of the phage. However, if an appropriate host were subsequently
infected, the novel DNA could be rapidly dispersed among susceptible bacteria;
i.e., a single non-defective phage that was able to accommodate the extra DNA
and with a burst size of 100 could transfer the novel DNA to approximately 10s
bacteria after only four lytic cycles. Hence,- there is the potential for a
novel gene, whether located in the chromosome or in a plasmid of a GMO, to
persist undetected and then to reappear unexpectedly in soil and other
habitats if the transfer of the gene occurred by transduction. Moreover,
observations that bacteriophages multiply in soil and other natural habitats
(obviously, with the involvement of appropriate host bacteria) emphasize the
potential for transducing phages to transfer genes in situ.
82
-------
Transformation has been considered to be a relatively unimportant
mechanism for the transfer of genetic information in soil and other habitats,
apparently because of the presumed susceptibility of "naked" DNA to microbial
degradation. However, the adsorption of DNA on clay minerals and other
surfaces, which may protect the DNA against degradation; the possible survival
and uptake of transforming DNA in microcolonies within microhabitats; the
ability of some species to incorporate heterospecific as well as homospecific
DNA; and the routine use of transformation to incorporate genetic information
including recombinant plasmids, into bacteria i_n vitro suggest that this
mechanism of gene transfer could also be important in situ. Moreover, the
persistence of transforming DNA in natural habitats, as the result of binding
on particulates, may also perpetuate cryptic genes in these habitats. The
long-term persistence of cryptic genes that could subsequently reappear in
multiplying cells must be considered in risk assessment. However, the
occurrence and frequency of transformation has not been studied sufficiently
in non-sterile soil and other natural habitats.
The purposes of this project were to determine: a) the frequency of gene
transfer by conjugation, transduction, and transformation in soil; and b) the
importance of the physicochemical characteristics in soil, especially the clay
mineral content, on survival and gene transfer.
METHODS
Most of these studies were conducted in a "minimal soil microcosm," which
consisted of test tubes containing 2 g of soil (sterile or non-sterile) at the
-33 kPa water tension. Replicate tubes were sacrificed for analysis at
various times of incubation (at 25 ± 2°C) after the addition of the GMOs
(including phage lysates and lysogenic bacteria) and appropriate recipients.
The soils were analyzed (using e.g., selective media, DNA probes, restriction
enzymes, electrophoresis, auxotrophy, and temperature-induction) for the
numbers of recipient, donor (including phages), and recombinant bacteria. All
data were statistically analyzed. Details of the methods used have been
published (e.g., Stotzky, 1989; Stotzky et al., 1989; and other papers in the
bibliography that relate to different methods for the transfer of genetic
information).
RESULTS AND DISCUSSION
Gene Transfer bv GMOs and Effect of Phvsicochemical Characteristics in Soil
o Frequencies of gene transfer.
conjugation -- chromosomal genes: 10"5 to 10"4
plasmid-borne genes: - 10"7 to 10"5
transduction: 10"5 to 10°
transformation: 10"7 to 10"6
o Transfer of chromosomal genes was enhanced by the presence of the clay
mineral, montmorillonite, but not of the clay mineral, kaolinite.
83
-------
o Transfer of chromosomal genes was greater at neutral than at acidic pH
levels.
o Montmorillonite probably enhanced the transfer of chromosomal genes by
increasing the growth of the donors and recipients, thereby providing
more mating pairs and a greater probability of contact for conjugation;
moreover, montmorillonite also maintained the pH of the microhabitats at
higher levels, as the result of its ability to scavenge protons and
replace them with basic cations from the exchange complex, thereby
promoting gene transfer.
o Intra- and intergeneric conjugal transfer occurred with the broad host-
range plasmid, RP4, but not with narrow host-range plasmids; transfer of
the latter occurred in sterile soil.
o Plasmid transfer was enhanced by montmorillonite, nutrient additions, and
optimum water content (i.e., the -33 kPa water tension).
o In the intergeneric transfer of RP4, the recipient P. aeruqinosa was
always ready to accept the plasmid from the donor £. co_M, but the donor
appeared to lose the ability to transfer the plasmid after residence in
soil.
o Gene transfer by transduction was unaffected by clay minerals but
enhanced, to a limited extent, by nutrient additions.
o Survival of transducing bacteriophages was enhanced by montmorillonite.
o Frequencies of transduction were higher with phage lysates than with
lysogens.
o Susceptibility to inactivation of different coliphages by a mixture of
soil microbes was: T5 > PI > Tl > lambda BF23 > T6; phage F116L of P.
aeruqinosa was the most resistant of the phages evaluated.
o Recovery of transducing phage PI from soil was enhanced by the use of
skim milk as the extractant (60% recovery vs. 10% with distilled water).
o The sensitivity and specificity of differentially-labeled DNA probes for
phage PI were evaluated; the highest sensitivity was obtained with a 32P-
labeled probe, but the detection of false positives was greater with this
probe than with the other probes; the probes prepared by biotin-nick
translation, biotin-photolabel, and antigen-chemical linkages were
equally sensitive and specific, and although less sensitive than the P-
labeled probe, they gave fewer false positives and were safer, cheaper,
and stable for longer periods.
o Transformation of plasmid-borne genes conferring resistance to erythro-
mycin or kanamycin but not of chloramphenicol-resistance genes (neither
plasmid- nor chromosomal-borne) was observed in non-sterile soil, whereas
all genes were transferred in sterile soil.
84
-------
o These results were obtained in non-sterile soil; in sterile soil,
frequencies of transfer by all three methods of transfer were several
orders of magnitude higher, again emphasizing the major role of the
indigenous microbiota in the survival of and gene transfer by GMOs;
moreover, phage F116L of P. aeruqinosa PA01 incorporated plasmid RP4 into
the virion and transduced the plasmid to cells of PA01 in sterile soil.
Preliminary studies in vitro indicated that the presence of montmorillon-
ite reduces the transformation of Bacillus subtilis by transforming DNA. This
reduction did not appear to be entirely the result of the binding of the DNA
on the clay but may have resulted from a competition between the clay and DNA
for DNA-binding sites on the recipient cells. The binding of DNA on clay
appeared to be pH-dependent, and at physiological pH values, the majority of
the DNA was not tightly bound, as most could be desorbed by water and buffers.
The DNA adsorbed on and then desorbed from clay did not lose its transforming
ability. Even more important, the small amount of DNA that was tightly bound
on the clay and could not be desorbed was still capable of transforming cells
when the clay-DNA complex was added to competent cells. This preliminary
observation supports the suggested possibility of cryptic genes persisting in
soil.
FUTURE WORK
The following types of studies would be conducted to provide additional
data from which EPA could develop realistic criteria on which to base risk
assessments for the release of GMOs to the environment.
Conjugation
Studies similar to those that have been conducted with E. coli. other
enterics, and P. aeruqinosa would be conducted with species and strains -f
GMOs that have a high probability of being released to the environment. M:re
emphasis would be placed on studying the effects of additional physicochemical
factors of soil that might affect the transfer of chromosomal- and plasmid-
borne genes and the survival of the GMOs and indigenous microbes to which the
novel genes may have been transferred. Such studies would eventually be
conducted in more complex soil microcosms.
Transduction
A portion of these studies would be a continuation of our current studies
on transduction by phage PI of E. coli in soil. Inasmuch as we have defined
this system and have developed a highly specific and sensitive DNA probe for
PI (using P-, biotin, and antigenic sulfone-labeling), we would conduct some
additional studies in soil with Pi-sensitive mutants of species of
Enterobacter and Klebsiella. which are typical soil bacteria, to estimate the
relevance of the data obtained with E. coli.
Transformation
We would continue the current in vitro studies with a variety of "clean"
(i.e., containing either their natural mixed cation complement or made
85
-------
homoionic to different cations) and "dirty" (i.e., coated with polymeric
oxyhydroxides of Fe, Al, or Mn) clays to determine the mechanisms of binding
of DNA and how the physicochemical characteristics of the clays affect
binding. The transforming ability of these various clay-DNA complexes would
be evaluated.
Concurrently, studies on transformation in non-sterile soil would be
conducted. Initial studies would be conducted with genetically-defined
strains of B. subtilis that carry both auxotrophic and antibiotic-resistance
markers on the chromosome and on small antibiotic-resistance plasmids.
Preliminary studies in vitro and in sterile soil have shown that these strains
can be transformed at adequately detectable frequencies, and methodologies for
their use in non-sterile soil would be developed. Soils that differ in
various physicochemical characteristics would be used in these studies, to
determine those characteristics that most affect transformation.
PUBLICATIONS
Devanas, M.A., D. Rafaeli-Eshkol, and G. Stotzky. 1986. Survival of plasmid-
containing strains of Escherichia coli in soil: effect of plasmid size and
nutrients on survival of hosts and maintenance of plasmids. Curr. Microbiol.
13:269-277.
Devanas, M.A., and G. Stotzky. 1986. Fate in soil of a recombinant plasmid
carrying a Drosophila gene. Curr. Microbiol. 13:279-283.
Devanas, M.A., and G. Stotzky. 1988. Survival of genetically engineered
microbes in the environment: effect of host/vector relationship. In Develop-
ments in Industrial Microbiology, G.E. Pierce (ed.). Society of Industrial
Microbiology 29:287-296.
Devanas, M.A., and G. Stotzky. 1989. The metabolic state of donor and
recipient cells affects conjugation frequencies and plasmid transfer to
indigenous bacteria in soil. Abstr. Ann. Mtng. Am. Soc. Microbiol., p.
353.
Glaser, D., G. Stotzky, and L. Watrud. 1987. Prospects for containment of
genetically engineered bacteria. In Prospects for Physical and Biological
Containment of Genetically Engineered Organisms, J.W. Gillett (ed.). Ecosys-
tems Research Center Report 114, Cornell University, Ithaca, NY. pp. 31-49.
Krasovsky, V.N., and G. Stotzky. 1987. Conjugation and genetic
in Escherichia coli in sterile and non-sterile soil. Soil Biol.
19:631-638.
recombination
Biochem.
Lee, G.H., and G. Stotzky. 1989. Transformation is a mechanism of gene
transfer in soil. Abstr. Wind River Conf. Genetic Exchange.
Stotzky, G. 1986. Influence of soil mineral colloids on metabolic processes,
growth, adhesion, and ecology of microbes and viruses. J_n Interactions of
Soil Minerals with Natural Organics and Microbes, P.M. Huang and M. Schnitzer
(eds.). Soil Science Society of America, Madison, MI. pp. 305-428.
86
-------
Stotzky, 6. 1989a. Gene transfer among bacteria in soil, in Gene Transfer
in the Environment. S.8. Levy and R.V. Miller (eds.). McGraw-Hill, NY. pp.
165-222.
Stotzky, G. 1989b. Microorganisms and the growth of higher plants in lunar
derived "soils." In Lunar Base Agriculture: Soils for Plant Growth, D.W.
Ming and D.L. Henninger (eds.). American Society of Agronomy, Madison, WY.
pp. 131-138.
Stotzky, G., and H. Babich. 1984. Fate of genetically-engineered microbes in
natural environments. Recombinant DNA Technical Bulletin 7:163-188.
Stotzky, G., and H. Babich. 1986. Survival of, and genetic transfer by,
genetically engineered bacteria in natural environments. In Advances in
Applied Microbiology, A.I. Laskin (ed.). Academic Press, NY. pp. 193-138.
Stotzky, G., M.A. Devanas, and L.R. Zeph. 1988. Behavior of genetically
engineered microbes in natural environments and their potential use in in situ
reclamation of contaminated sites. In Biotechnologische In-situ-Sanierung:
Kontaminierter Stadorte, Z. Filip (ed.). Gustav Fischer Verlag, Stuttgart.
pp. 293-343.
Stotzky, G., M.A. Devanas, and L.R. Zeph. 1989a. Methods of studying bac-
terial transfer in soil by conjugation and transduction. U.S. Environmental
Protection Agency Protocol Document, EPA-600/3-89/042. 169 pp.
Stotzky, G., M.A. Devanas, and L.R. Zeph. 1989b. Methods for studying
bacterial gene transfer in soil by conjugation and tr^.nsduetion. In Advances
in Applied Microbiology, S.L. Neidleman (ed.). Academic Press, NY (in press).
Stotzky, G., M.A. Devanas, and L.R. Zeph. 1989c. Gene transfer by conjuga-
tion and transduction in soil. Abstr. 5th Intern. Symp. Microbial Ecol.,
Kyoto, Japan, p. 87.
Zeph, L.R., X. Lin, and G. Stotzky. 1989. Comparison of radioactive versus
non-radioactive DNA probes for the detection of target DNA. Abstr. Ann. Mtng.
Am. Soc. Microbiol., p. 342.
Zeph, L.R., M.A. Onaga, and G. Stotzky. 1988. Transduction of Escherichia
coli by bacteriophage PI in soil. Appl. Environ. Microbiol. 54:1731-1737.
Zeph, L.R., and G. Stotzky. 1989. Use of biotinylated DNA probe to detect
bacteria transduced by bacteriophage PI in soil. Appl. Environ. Microbiol.
55:661-665.
87
-------
EXPLORING GENETIC INSTABILITY IN TRANSFORMED FUNGI
J. Armstrong,1 D. Harris,2 P. Rygiewicz,1 and N. Fowles2
U.S. EPA1
Environmental Research Laboratory
NSI Technology Services Corporation2
Corvallis, OR 97333
INTRODUCTION
With the likelihood that recombinant fungi will soon be released into the
ecosystem, EPA recognizes a parallel need for the development and evaluation
of methods to detect and identify specific fungi in the environment. Our
research related to these needs with the following objectives: a) evaluate
DNA probe hybridization with restriction fragment length polymorphisms (RFLPs)
as a means to identify fungi, and b) study the recombination-related factors
that determine stability of recombinant marker genes in transformed fungi.
Mycorrhizal fungi are good model organisms for studies of RFLPs as tools
for identification, since the mycobiont of ectomycorrhizae is difficult to
identify when based on mycorrhizae morphology or extramatrical hyphae. By
using a probe encoding rRNA genes from Coprinus. we were able to distinguish
between different isolates of the same species. This work was a first step to
evaluate DNA probe methods in the detection and identification of species and
isolates of the ectomycobiont.
Although RFLPs offer one approach to detection and identification of
fungi released into the environment, it would also be convenient to have fungi
which have been transformed with marker genes, e.g., antibiotic resistance
genes. However, researchers are observing that recombinant DNA is often
unstable in the transformed fungi. It is typical for transformants to lose
the transformed DNA or the introduced DNA is rearranged within the chromosome.
Also, transformants are often relatively sickly as demonstrated by their poor
survival. We have selected the fungus Colletotrichum gloeosporioides var,
aeschvnomene (Cga) as a model organism to study the factors that determine the
stability of transformed marker genes.
METHODS
RFLPs of Mycorrhizal Fungi
DNA was extracted from hyphae of mycorrhizal (Laccaria bicolor. L.
laccata. Pisolithus tinctorius, and Hebeloma crustuliniforme) and saprophytic
(Schizophvllum commune and Coprinus cinereus) fungi. These DNAs were digested
with Eco Rl, Pvu II, Bam HI, or Xho I, and electrophoresed. Southern
transfers of the RFLPs were hybridized with 32P-labeled plasmid pCcl (encodes
rRNA genes from C. cinereus) and autoradiographed.
88
-------
DNA Transformation of Cqa
Cga was cultured on Torula yeast agar for 4 d at 30"C. Spores were
harvested from the plates by washing with sterile water, centrifuged for 5 min
at lOOOxg, suspended in 0.6 H MgS04 containing Novozym 234 (20 mg/ml), and
incubated for 5 h at 30°C. Protoplasts were then centrifuged at 2000xg and
resuspended in 1.2 M sorbitol and washed five times by centrifugation to
remove nucleases that were present in the Novozym. They were finally
suspended in electroporation buffer (10% mannitol, 10 mM HEPES, pH 7). About
10 ug of DNA were added and electroporated into Cga with a Cell-Porator (BRL).
After 24 h regeneration in yeast-starch medium, the electroporated fungi were
plated on Torula yeast agar containing hygromycin B (Hyg B; 500 ug/ml) and
incubated at 30°C. T-ansformants were isolated from these plates and
purified. We are currently transforming Cga with two plasmids encoding Hyg
BR. Plasmid pHlS has a promoter from Cochliobolus and pHLl has a promoter
from listilago.
RESULTS AND DISCUSSION
RFLPs of Mvcorrhizal Fungi
Digestions of the mycorrhizal fungus DNAs with the four enzymes produced
RFLPs from the five genera of Basidiomycetes which lacked fragments that were
characteristic of Basidiomycetes. The differences among all the RFLPs made
them distinguishable to genus. Next, we hybridized the RFLPs with labeled C.
cinereus rDNA (gene encoding rRNA in pCcl). When RFLPs produced from Pvu II-
digested DNA were hybridized with probe DNA, the bands were unique at the
genus level for the five fungi, but did not allow discrimipation of two L.
bicolor isolates. However, when DNA from the five genera and the two L.
bicolor isolates were digested with Xho I and Eco RI, the hybridization
patterns were dissimilar. Additional cultures of ].. bicolor and i. laccata
were studied to determine if RFLPs would be characteristic for the genera and
species. Non-probed RFLPs produced with Xho I, Bam HI, and Eco RI were
isolate-specific. Also there appeared to be no predominant bands that were
common to either species. As with non-probed RFLPs, those hybridized with
pCcl showed no consistent patterns for the Laccaria genus or for either
species. Collectively, RFLPs were isolate-specific.
These experiments illustrate how analyses of RFLPs offers a way to
detect, identify and study the ecology of mycorrhizal fungi. For example, the
fingerprint of L. bicolor 2027 was markedly different from the other L.
bicolor isolates, suggesting that it be considered for reclassification.
Another example of the utility of the methods was demonstrated by L,. bicolor
S238, which was originally accessioned and distributed as i. laccata and later
reclassified as L. bicolor (Greg Mueller, personal communication). This
reclassification is supported by the qualitative similarities among the Xho-
produced RFLPs from L.. bicolor S238 and the other L- bicolor isolates. The
extensive variation in RFLPs in the assortment of mycorrhizal fungi isolates
we studied suggests the techniques will not permit us to identify these fungi
at the species level.
89
-------
Thus, DNA probe methods should be useful in the identification of
mycorrhizal fungi in the field under certain conditions. For example, one or
more known isolates in field samples would be distinguishable when their ONA
fingerprints are previously ascertained and. are distinct from those of
indigenous organisms. To identify the mycobiont to genus or species, probes
will be needed at each taxonomic level.
DNA Transformation of Cqa
Since the research on introduction of marker genes in fungi is in the
early phases of the work, we have no data concerning the comparative stability
and activity of the Hyg BR genes in the transformed Cga strains we have
obtained. However, we have noticed that they typically have altered morphol-
ogy and pigmentation, have a slower growth rate than the nonrecombinant
parent, have poor survival, and frequently lose the transformed antibiotic
resistance phenotype. These observations have lead us to question the role
that homologous and heterologous integration plays in stabilizing transformed
DNA. Our anticipated future work will focus on this question.
FUTURE WORK
Our plans encompass two general areas of study. One is the production of
various Hyg BR strains of Cga, which will be studied for their stability and
growth in laboratory media as compared to the parent strain. Two of these
strains will be transformants carrying the resistance gene with either the
Ustilaqo or the Cochliobolus promoter. A third strain will be Cga transformed
with a plasmid carrying the Hyg BR gene flanked by portions of the cutinase
gene from a C. qloeosporioides strain isolated from papaya.-fruit. This
construction will allow us to determine if stability of integrated DNA is
enhanced by the presence of DNA common to the transformed organism. We will
determine if the presence of the homologous DNA influences the integration
event. The second phase of our research will involve comparisons of the
survival of these strains and the persistence of the transformed DNA after the
fungi are sprayed into terrestrial microcosms containing either peas or
crimson clover.
PUBLICATION
Armstrong, J.L., N.L. Fowles, and P.T. Rygiewicz. 1989. Restriction fragment
length polymorphisms distinguish ectomycorrhizal fungi. Plant and Soil 116:1.
90
-------
PLASMID EXCHANGE AMONG AGROBACTERIA ASSOCIATED WITH AGRICULTURAL PLANTS
Joyce Loper,1 Michael Kawalek,2 Virginia Stockwell,2 and Larry Moore2
USOA-ARS1
Horticultural Crops Research Laboratory
Department of Plant Pathology2
Oregon State University
Corvallis, OR 97331
INTRODUCTION
Genetic exchange between a microbial agent introduced into the environ-
ment and the indigenous microflora is well known. However, most studies
evaluating conjugal plasmid transfer between microbes in the environment have
focussed on plasmids constructed in laboratories or host organisms that are
not presently utilized as microbial inoculants. The present studies focus on
plasmid exchange between Aqrobacterium radiobacter strain K84, an EPA
registered agent for the biological control of crown gall disease, and
Aqrobacterium tumefaciens. a ubiquitous soil bacterium which causes the crown
gall disease of plants. Dr. Alan Kerr first demonstrated in 1968 that the
tumor-inducing plasmids (pTi) of A. tumefaciens are conjugative, transferring
at relatively high frequencies between Aqrobacterium strains in crown gall
tissue. Plasmid exchange among Agrobacteria continues to be an excellent
model system for the study of conjugal plasmid transfer in the environment.
Aqrobacterium radiobacter strain K84 produces an antibiotic, agrocin 84,
with specific toxicity against certain A. tumefaciens strains. Agrocin 84
production is encoded on the plasmid pAgK84, which is indigenous to strain
K84. The plasmid pAgK84 also encodes immunity of the host bacterium to
agrocin 84. Previous work in Dr. Steven Farrand's laboratory at the Univer-
sity of Illinois at Urbana has demonstrated that pAgK84 is a conjugative
plasmid. The sensitivity of A. tumefaciens strains to agrocin 84 is encoded
on the tumor-inducing plasmids (pTi), which also confer virulence to the host
bacterial cell.
Plasmid exchange between A. radiobacter and virulent A. tumefaciens
strains has great economic significance. The transfer of p~AgK84 to A.
tumefaciens will confer agrocin immunity to the virulent strain. Similarly,
pTi transfer to strain K84 will confer virulence to the biocontrol strain. In
both cases, a virulent strain with immunity to biocontrol will result. An
alarming report from Greece in 1978 stated that the biological control
activity of strain K84 had broken down due to the transfer of pAgK84 from
strain K84 to virulent Aqrobacterium strain(s) in crown gall tissue. However,
subsequent attempts to detect such strains following the field release of
strain K84 in other parts of the world were not successful. Such attempts,
which relied on the phenotypic characteristics associated with the plasmids
for their detection, were extremely labor-intensive and thus restricted to the
evaluation of less than one thousand isolates.
The present study utilizes transposon tagging and DNA probes as methods
for the specific and sensitive detection of Aqrobacterium plasmids. Anti-
91
-------
biotic resistance markers and serological techniques are used for the detec-
tion of host bacterial strains. The application of these plasmid-dependent
and piasmid-independent detection methods facilitated the study of plasmid
exchange among Agrobacteria in the laboratory, microcosm, and field.
METHODS
Plasmid Exchange in Culture
Strain B49c is a virulent A. tumefaciens strain isolated from crown gall
tissue of apple in Washington state. Strain B49c was selected for this study
because it is not sensitive to agrocin 84 nor to biocontrol by strain K84, and
can coexist with strain K84 in culture and in crown gall tissue. Strains with
resistance to streptomycin (Sm) or rifampicin (Rif) were designated B49c-Sm
and K84-Sm, or B49c-Rif and K84-Rif, respectively. Strain B49c(pTiB49c::Tn5)
and K84(pAgK84::Tn5) were mated with strain K84-Rif and B49c-Rif, respec-
tively, as follows. Donor and recipient strains were adjusted to 0.5 O.D.^g
(approximately 107 cfu/ml). 100 ul of a mixed suspension (1:1 donor:-
recipient) was spotted onto the surface of Mannitol Glutamate Yeast Extract
agar (MGY) supplemented with 3 mM nopaline and 10 uM agrocinopine. Plates
were incubated for 24 h at 27°C. Serial dilutions of cells scraped from the
plate surface were spread on MGY agar supplemented with rifampicin (Rif) (100
ug/ml), kanamycin (Km) (50 ug/ml) and streptomycin (Sm) (25 ug/ml). Plates
were incubated at 27 C for 4-5 days before colonies were enumerated. The
presence of plasmids in Rif, Kmr, Smr transconjugants was confirmed by
alkaline lysis plasmid purification and gel electrophoresis.
Plasmid Exchange in Microcosms
Bacterial suspensions (0.5 OD540 ^ were prepared from 5 day cultures
grown on MGY medium supplemented with appropriate antibiotics. Plant wounds
were inoculated with 20 ul of a mixed (1:1) cells suspension of donor and
recipient strains. Host plants were tomato (cv Bonny Best), sunflower (cv
Teddy Bear), Bryophyllum (Kalanchoe daigremontiana). apple (cv Mark root-
stock), and cherry (cv Mazzard rootstock). Plants were grown at 25 C with a
16 hr photoperiod. At 4, 6, 8, and 10 weeks, galls were harvested from
inoculated plants, and sampled by grinding. Aliquots of gall extract dilu-
tions were spread on MGY-Rif, MGY-Km-Sm (50 ug/ml), and MGY-Rif-Km-Sm, to
enumerate recipient, donor and transconjugant colonies, respectively. The
identity of ten representative transconjugants from each mating was confirmed
by serological techniques, as described previously, and by gel electro-
phoresis.
Plasmid Exchange in the Field
Cherry seedlings were inoculated in 1988 and 1989 by dipping wounded root
systems into mixed suspensions of donor strain K84-Sm and recipient strain
B49c-Rif before planting in field trials located at the Botany Farm at Oregon
State University, Corvallis, OR. Galls from inoculated cherries were
harvested as described above. Aliquots (100 ul) of gall extract dilutions
were spread on MGY-Rif or MGY-Sm (500 ug/ml) to enumerate recipient and donor
colonies, respectively. Galls with populations of the recipient strain B49c
92
-------
of at least 10 cfu/g fresh weight gall tissue were analyzed for the presence
of transconjugant colonies (strain B49c harboring pAgK84). Gall extracts were
spread on ten MGY-Rif plates to give a final density of 5000 cfu/plate.
Colony blots of the ten plates were probed with 2P-labeled agrocin-biosyn-
thesis genes cloned in pUC8. Autoradiograms were exposed at -80 C for 7 days
prior to development.
RESULTS AND DISCUSSION
Plasmid Exchange in Culture
Plasmid pTiB49c::Tn5 transfer from strain B49c to strain K84 was observed
at mean frequency of approximately 10"6 transconjugants/recipient (3.4 X 10"6
and 6.9 X 10"7 in two subsequent experiments). Plasmid pAgK84::Tn5 transfer
from strain K84 to strain B49c was observed at a mean frequency of approxi-
mately 10'5 (1.7 X 10'5 and 9.2 X 10'6).
Plasmid Exchange in Microcosm Experiments
Transfer of pAgK84::Tn5
Populations of strain B49c-Rif and K84(pAgK84::Tn5) ranged between 106
and 10s cfu/g crown gall tissue from 4 to 10 weeks following inoculation of
tomato plants. Transconjugants (B49c-Rif(pAgK84::Tn£)) were detected in
tomato galls at all sampling dates from 4 to 10 weeks at a frequency 10"2 to
10'4 transconjugant/recipient. Transconjugants were detected in only three of
the ten sunflower galls that were sampled over the ten week period. No
transfer of pAgK84::Tn5 was observed in galls of cherry or,bryophyllum. Opine
analysis of individual galls from each plant sampled provided no obvious
explanation for the presence or absence of transconjugants in the galls of
this experiment. However, transconjugants were detected only in galls with
populations of the recipient strain B49c-Rif of at least 106 cfu/g. The
population size of the donor strain K84 ranged from 104 to 107 in galls where
transconjugants were detected. Thus, detection of transconjugants was
dependent on a large population size of the recipient, but was relatively
independent of the donor population size in gall tissue.
Transfer of pTiB49c::Tn5
No transfer of pTiB49c::Tn5 was detected over the ten week period
following inoculation of plants with B49c(pTiB49c::Tn5) and K84-Rif. The
population size of B49c(pTiB49c::Tn5) was similar to that of B49c-Rif,
described above, in galls of all plant hosts. The population size of K84-Rif
was similar to that of K84(pAgK84::Tn5), described above, in galls of all
plant hosts. The population size of the recipient strain K84-Rif never
exceeded 10 cfu/g gall tissue. The relatively small population size of the
recipient strain in gall tissue may explain the lack of transconjugant
detection in this experiment.
93
-------
Plasnrid Exchange in the Field
The population size of recipient strain B49c-Rif and donor strain K84-Sm
appeared adequate for detection of transconjugants in galls retrieved from
cherry trees in the field, based on the microcosm experiments discussed above.
Preliminary results from colony blot experiments using P-labeled agrocin
biosynthesis genes as a probe suggest the presence of one to three transcon-
jugants in one cherry gall. The population size of K84-Sm and B49c-Rif in
this cherry gall were 1.2 X 105 and 2.4 X 106 cfu/g gall tissue, respectively.
Experiments are in progress to confirm the identity of the colonies as B49c-
Rif (pAgK84) by serological techniques, agarose gel electrophoresis, and
Southern analysis.
The Aqrobacterium system provides an excellent model for the development
of methods to study the conjugal transfer of indigenous plasmids among
microbes in t^e environment. Transposon tagging and DNA hybridization coupled
with serologi-al techniques and antibiotic-resistance markers have allowed the
:ransconjugants in laboratory, microcosm, and possibly field
The present study, addressing a problem of great economic and
significance, exemplifies the power of the implemented methods
of DNA exchange among microbes in native or agricultural
detection of
experiments.
environmental
for the study
habitats.
PUBLICATIONS
Kawalek, M.D., J.E. Loper, L.W. Moore, and V.S. Stockwell. 1989. Influence
of the Aarobacterium radiobacter strain K84 nopaline plasmid, pAt84b, on
conjugal transfer of plasmids between Aarobacterium strains-. Phytopathology
(in press) (abstract).
Stockwell, V.O., L.W. Moore, M.D. Kawalek, and J.E. Loper. 1989a. Field
evaluation of antibiotic-resistant mutants of Aorobacterium spp. Phytopathol-
ogy (in press) (abstract).
Stockwell, V.O., L.W. Moore, M.D. Kawalek, and J.E. Loper. 1989b. Field
population dynamics of Aqrobacterium spp. associated with cherry. Phytopath-
ology (in press) (abstract).
94
-------
SESSION V
SURVIVAL/COLONIZATION
95
-------
-------
BACTERIAL TRAITS INVOLVED IN ROOT SURFACE COLONIZATION
Anne J, Anderson, Robin Buell, Branco Lovic, and Jlrasak Katsuwon
Biology Department
Utah State University
Logan, UT 84322-5305
INTRODUCTION
Roots are colonized by a range of microorganisms including fluorescent
pseudomonads. Certain pseudomonads are being considered for release into
agricultural use because of their potential to improve plant growth and
welfare. Growth promotion has been attributed to direct effects on plant
metabolism and/or indirect consequences of the suppression of soil-borne
pathogens. The aggressive ability of the isolates to compete for colonization
of root surfaces may be an important facet of these beneficial effects. The
pseudomonads are provided with a nutrient source from the root exudates and
are at the right location to affect root metabolism and to deter root patho-
gens. Consequently, our work concerns the mechanisms involved in root coloni-
zation. Understanding the mechanisms that permit survival at the root surface
may shed light on the risks that are involved with GEM release and permit
"safer" strains to be constructed. The research has centered in two areas of
survival at the root surface: a) attachment mechanisms; b) protection against
host defenses.
RESULTS AND DISCUSSION
Attachment Mechanisms
The aggressive and competitive colonization of the root surface by
pseudomonads may involve discrete mechanisms of attachment. We had previously
demonstrated that cells of a beneficial isolate of Pseudomonas putida were
agglutinated by glycoproteins termed agglutinins from bean or soybean root
surfaces. This agglutination phenomenon contributed to short term binding of
the cells to the root surface and was beneficial in long term root coloniza-
tion from inoculated seeds and suppression of root pathogens.
Investigations of the extent of the agglutination phenomenon involved
studies with two other crops for which beneficial effects of pseudomonads have
been identified in the field, wheat and sugar beet. A survey of fluorescent
pseudomonads obtained from the roots of field-grown sugar beet indicated the
proportion of agglutinable isolates was higher in rhizoplane strains than
those associated with the rhizosphere. Several isolates were self agglutin-
able. Examination of wheat and sugar beet root washes revealed the presence
of agglutinin for wild type Agg* P. outida. These same root washes did not
agglutinate Agg- mutant forms of P. putida. which were selected because of
their lack of agglutination with bean agglutinin. Agglutinin was also
detected for Agg but not Agg" P. putida isolates in root washes of weedy
species. Consequently, this adherence mechanism may act in hosts which are
non-target organisms for released pseudomonads.
97
-------
Biochemical studies involving the bacterial surface factors involved in
agglutination have Indicated that expression of the agglutination phenotype is
highly regulated. The phenotype 1s suppressed in minimal medium but expressed
in rich medium or during growth on root wash components. Lack of agglutina-
tion is not correlated with altered ability to bind to hydrophobic or hydro-
philic surfaces or change in net surface charge. IPS structure is not altered
between Agg" mutants and the Agg* wild type but there are differences in total
protein and outer membrane components. However, there is no component that is
altered in common for all of the Agg" mutants. Agglutination does depend on
some proteinaceous factor on the bacterial cell surface as indicated by loss
of agglutinability after trypsin treatment.
Genes involved in determining agglutination phenotype in the bacterium
are being isolated. One Tn5 Agg" mutant 5123 is restored to the Agg* state by
a wild type clone containing a 2.7 kb EcoRI-Hind III insert. This same clone
enhances agglutinability in the wild type but does not restore agglutinability
to other Agg" mutants. However, this clone does cross hybridize with genomic
material from other pseudomonads, including both agglutinable and non-
agglutinable root colonizers and plant pathogens which are non-agglutinable.
Other clones were obtained by hybridization to sequences which flank the
transposon insertion site in a second Tn5 Agg mutant 1104. These clones
inhibited the agglutination ability of the wild type isolate and failed to
restore the Agg phenotype to 1104 or other Agg" mutants.
Protection Against Host Defenses
The root is a hostile environment for a microbe because of the
production of potentially toxic phenolics and other defense mechanisms
including the generation of active oxygen species. Consequently, in examining
the strategies that bacteria use in survival at the root surface, we centered
on the role of enzymes that could protect the bacterium from active oxygen
species produced by the plant root. These enzymes are catalase (CAT) and
superoxide dismutase (SOD).
We have demonstrated that root colonizing pseudomonads produce CAT and
SOD in isozymic forms. The difference in isozyme pattern for CAT between root
colonizing pseudomonads could be used as a tool for identification of a
specific strain. Both CAT and SOD are regulated in terms of activity levels
and isozyme expression dependent on growth medium and growth phase. Exposure
of the bacteria to H202 also changed the level of CAT activity and isozyme
pattern.
Preliminary data suggest that regulatory mechanisms may affect CAT and
SOD activities during the process of root colonization by the bacteria.
Growth in culture on components in root exudates enhanced CAT activity.
Growth on plant roots promoted SOD activity, the activity being higher in
rhizoplane than rhizosphere cells. The importance of enzymes to cope with
active oxygen species to the root colonizing bacteria is further suggested by
the failure of a mutant of P. outida. deficient in catalase, to colonize bean
roots.
98
-------
ANTICIPATED FUTURE STUDIES
Our goals are to: a) characterize the binding site of the agglutinin on
the bacterium; and b) sequence the structural and regulatory regions of genes
that are involved in agglutinability.
We feel that the agglutination phenotype is associated with survival in
the root environment and that its further examination will reveal information
concerning environmental sensing systems in the bacterial cell. The promoter
analysis may provide sequences useful for generation of a suicide construct
which would cause a GEM to self destruct in the absence of the host plant.
Our goal is to isolate the genes from root colonizing pseudomonads that
encode CAT and SOD. Because of the regulation of these genes in the root
environment, promoter analysis will be conducted. Additional mutants lacking
in CAT or SOD will be selected and their behavior during root colonization
examined. The factors from the root which condition altered regulation of CAT
and SOD will be investigated.
PUBLICATIONS
Katsuwon, K., and A.J. Anderson. 1988. Resistance of the root colonizing
bacterium Pseudomonas outida to activated oxygen species. Phytopathology
78:1591 (Abstract).
Tari, P.M., and A.J. Anderson. 1988. Fusariurn wilt suppression and agglutin-
ability of Pseudomonas putida. Appl. Environ. Microbiol. 54:2037-2041.
Katsuwon, J., and A.J. Anderson. 1989. Response of plant-colonizing pseudo-
monads to hydrogen peroxide. App. Env. Microbiol. 55:2985-2989.
99
-------
FATE OF Bacillus sohaericus: A MICROBIAL
PEST CONTROL AGENT IN THE ENVIRONMENT
Allan A. Yousten and Ernest F. Benfield
Biology Department
Virginia Polytechnic Institute and State University
Blacksburg, VA 24061
INTRODUCTION
Bacillus sphaericus is an aerobic, spore-forming bacterium found in high
numbers in soil. Some strains of this species have been found to be
pathogenic for mosquito larvae. This pathogenicity is caused by the produc-
tion by the bacteria of a protein toxin which is lethal upon ingestion by
larvae of many mosquito species. Unlike the better known toxin produced by
Bacillus thurinqiensis subsp. israelensis (Bti) the toxin of fi. sphaericus is
not active against black fly larvae. The amino acid sequence of the B.
sphaericus toxin has been determined, and it is not similar to the toxins of
B. thurinqiensis. However, like the Bti toxin it is formed at the time of
bacterial sporulation and is found as a parasporal inclusion body located
adjacent to the spore in the sporangium, fi. sphaericus also differs from Bti
in that the former bacteria do not metabolize most carbohydrates as sources of
carbon and energy for growth, and this characteristic undoubtedly has an
influence in determining the ecological niches where the bacteria can survive.
Laboratory studies and field trials indicate that fi. sphaericus may be
developed as a useful mosquito larvicide. In this role both spores and toxic
parasporal inclusions will be sprayed on the water to be consumed by filter
feeding larvae. One characteristic which has attracted most interest is the
apparent ability of this bacterium to maintain control of larval populations
in water for a longer period of time than Bti, typically a few weeks vs. a few
days for Bti. The reasons for this extended persistence are unclear. The
studies which we have undertaken examine spore survival in the laboratory and
in the field as well as the fate of spores ingested by certain non-target
animals in the laboratory. Spore survival is measured by the ability of the
spores in a sample to survive heating and to produce a colony on selective
media. Failure to do this could be due to loss of the ability to germinate,
germination (and by definition) loss of heat resistance, or by destruction of
the spore (or resulting vegetative cell) by some physical or biological
influence. At the present time we do not distinguish between these
alternatives but rather are determining the conditions under which they may
occur.
METHODS
A spontaneous, rifampicin-resistant mutant of fi. sphaericus 2362 was
used in all studies. Spores were produced by growth in nutrient broth-0.05%
yeast extract-mineral salts mix broth (NYSM) with shaking at 30°C for 48 h.
Spores were washed and held as a distilled water suspension at 4°C. Spore
numbers were determined by heating of 2 ml at 80°C for 12 min and plating on
NYSM agar containing 50 ug/ml rifampicin and 0.002% cycloheximide.
100
-------
The effect on spore dormancy of extended incubation in the laboratory
has been studied by suspending spores in 0.05M MES (pH 6.0), MOPS (pH 7.0) and
HEPES (pH 8.0) buffers. The buffered spore suspensions were incubated at 15°
25°, and 35°. The buffers lacked added carbon or nitrogen sources and were
incubated static in closed bottles. Spore dormancy was also studied using
suspensions prepared in water collected from a local pond. Both filtered
(0.45 urn) and unfiltered water were used to prepare 0.05M MOPS buffer (pH 6.5)
or were left unbuffered at the initial pH of 8.4. Spore suspensions in
filtered and unfiltered water at both pHs were incubated, static, in closed
bottles at 15° and 30°, Bottles were vigorously shaken before sampling for
spore counts. Organic carbon in the water was determined at the initiation of
the laboratory experiments and, weekly, in field experiments by use of a
Dohrmann DC-10 carbon analyzer.
The effect of aeration during incubation was studied by preparing spore
suspensions in autoclaved and non-autoclaved pond water (buffered at pH 7.5
with 0.05M MOPS) and incubating in closed bottles or shaken (80 rpm) flasks at
25°.
Maintenance of spore dormancy in the field was examined by placing spore
suspensions in dialysis bags and submerging the bags to about 0.5 meter in
ponds. The bags were removed weekly, sampled, and the remaining contents
placed in new bags and returned to the water.
The fate of spores in laboratory-reared midge larvae was determined by
allowing 3rd-4th instar larvae to feed on a spore suspension mixed with
Tetramin fish food for 6 h. Larvae were removed from the food, washed 3x, and
placed in fresh food without spores. Immediately after feeding and at 1 day
intervals for 5 days, 5 larvae were collected, washed in sterile water, and
homogenized in sterile water. The spores remaining in/on the larvae were
determined by plating. Background counts to determine the number of spores
adhering to the larval surface were done using larvae killed by a 2 sec dip in
boiling water. The dead larvae were submerged in spore suspension for 6 h,
washed 3x in sterile water, homogenized, and plated.
RESULTS AND DISCUSSION
Laboratory Studies of Spore Dormancy
Spores suspended in buffers lacking added carbon or nitrogen sources at
pHs 6, 7, and 8 at 15°, 25°, and 35° germinated less than 1 log after 237 days
of incubation. Percent germination ranged from 28% to 78% with slightly
higher germination at the higher temperatures but with no indication that pH
affected germination.
Spores suspended in closed, unshaken bottles in filtered pond water
demonstrated less than one log of germination in 238 days whether incubated at
pH 6.5 or 8.4 or at 15° or at 30°. However, when unfiltered (non-sterile), pH
8.4 water was used to suspend the spores, there was a drop of 92% and 94% at
15° and 30°, respectively. At pH 6.5 there was a drop of 38% at 15° and 86%
at 30°. Most germination occurred in the first 100 days. Germination was
favored in non-filtered water at 30°.
101
-------
In comparing the effect of shaking vs. static conditions on spore
germination in autoclaved and non-autoclaved pond water, it was found that
after 35 days little germination had occurred in autoclaved water whether it
was shaken or not. However, in non-autoclaved water a 90% drop in spore
number occurred in the static spore suspension and a 60% drop in the shaken
suspension. Phase contrast microscopic observation confirmed these plating
results. In static, autoclaved water, only 10% of the spores had become phase
dark (indicating germination) whereas 75% of the spores incubated statically
in non-autoclaved water were phase dark. Vegetative cells were not present,
indicating a lack of spore outgrowth. The toxic parasporal bodies associated
with germinated spores appeared to be intact. Spore germination was favored
by non-sterile water and somewhat by static vs. shaken conditions. The
viability of the phase-dark (germinated) spores will be determined. The
reason(s) for more rapid germination in non-autoclaved water are as yet
unknown.
Field Studies of Spore Dormancy
Placing spore suspensions in dialysis bags allowed them to be exposed in
the pond water to dissolved substances of about molecular weight 10,000 or
less but to be retrievable for counting and determination of dormancy. The
drop in spore numbers observed in the suspensions under these conditions was
greater and more rapid than seen in the laboratory studies. In two ponds a
drop of about 3 logs in spore number was seen in 42 days. During this time
period the temperature ranged from 20° to 29° in one pond and from 18° to 31°
in the second. The pH ranged from 7.40 to 8.27 in the first pond and 7.62 to
8.55 in the second. The level of dissolved organic carbon ranged from 2.85 to
5.57 mg/L in the first and 5.07 to 9.53 mg/L in the second. The dialysis
membrane was found to remain intact for 2-3 weeks under these conditions, but
the suspensions were placed into new bags weekly.
Fate of Spores Fed to Non-Target Invertebrates
Mosquito larvae which consume B. sphaericus spores and toxic paraspores
die in 1 or 2 days. This is followed by proliferation of the bacteria in the
larval cadaver. Each cadaver contains about 105 to 106 spores after 3 or 4
days as the cadaver begins to disintegrate. In contrast, midge larvae, which
are not susceptible to the toxin, contained about 105 to 106 spores after
feeding on a spore suspension for 6 h. In 3 to 4 days they had eliminated
most of these spores. It is unknown if the spores were digested, if they
germinated and the vegetative cells were digested, or if the spores (and
possibly vegetative cells) were passed out in the feces. These possibilities
are being investigated. In the future, spores will be fed to snails feeding
on biofilms and to two species of aquatic insects with vastly different gut
pHs (6 vs. 12). Spore survival in the guts and feces will be determined.
102
-------
Fate of Bacillus sphaericus Microbial Pest Control Agent
in the Estuarine Environment
Fred Genthner1, Bob Campbell2, Susan Martin2, Steve Foss3, and Jack FournieS
Microbial Ecology and Biotechnology Branch,
technical Resources, Inc. and 3Pathobiology
Branch
U.S. Environmental Protection Agency
Environmental Research Laboratory
Gulf Breeze, Florida
Introduction
The development of Bacillus sohaericus as an alternative to chemical
pesticides or the traditional biological control agent g. thurinqiensis var.
israelensis (B.t.i.),for mosquito control holds a great deal of promise. The
majority of research on B. sphaericus has been concerned with its
effectiveness as a mosquito larvicide. Relatively little is known about the
environmental fate and effect of B. sohaericus endospores which are present in
preparations of the toxin. It is unlikely that B. sphaericus will behave in
the same manner as the more extensively studied B.t.i due to differences in
metabolism and the capacity of B. sphaericus to deliver toxin to mosquito
larvae for longer time periods than does B.t.i.
B. sphaericus will likely be applied to both freshwater and estuarine
environments since mosquito breeding sources range from fresh to brackish
water. Therefore, our objective is to study persistence of B. sohaericus
endospores in the estuarine environment to provide a comprehensive data base
for predicting their fate in the aquatic environment. Spore persistence was
determined by the ability of spores in the sample to form colonies on solid
medium. By these procedures, a reduction in spore number could, therefore,
result from either destruction of the spore, loss of its ability to germinate,
or loss of heat resistance through germination. Recognizing these
alternatives, the goal of this research is to determine which conditions
affect spore persistence.
Additionally, the infectivity, toxicity, and pathogenicity of B.
sphaericus in a nontarget organism, the American oyster, Crassostrea
virqinica. will be assessed in single species test systems.
103
-------
Methods
B. sphaericus. 2362-7, obtained from Dr. A. Yousten, was used for all
studies. This strain was a spontaneous rifampicin resistant derivative of
strain 2362. Nutrient broth-0.05% yeast extract-mineral salts medium (NYSM)
was used for liquid culture. Solid medium (NYSMRC) contained 0.8% agar, 0.05
mg/ml rifampicin, and 0.002% cyclohexamide. Spore suspensions were produced
from 48 hr cultures which were incubated with shaking at 30*C. Spore counts
were performed by heating a 2ml sample at 80*C for 12 min, sonicating for 1
min to disperse clumped spores, and plating on NYMSRC.
Flask studies were performed to examine both physical and biological
factors which affect spore persistence. Spores (« 2 X 106 per ml) were added
to 100 ml volumes of estuarine water contained in 125 or 250 ml flasks.
Unless otherwise indicated, flasks were incubated with slow shaking (50 rpm).
Sterile synthetic seawater was used to examine the effects of salinity
(10, 20 and 30 °/oo), temperature (4, 15 and 30 *C) dissolved organic carbon
(5, 50, and 500 mg/L) on spore persistence. Glucose, a nonutilizable carbon
source for B. sphaericus, and casamino acids, a utilizable carbon source for
B. suhaericus were the dissolved organic carbon substances tested. The effect
of pH was not examined because estuarine water is normally buffered around pH
8.0. Natural estuarine water and detritus (obtained from Santa Rosa Sound,
Pensacola, Florida) were used to examine the effect of particulates and
predation on spore persistence. Organic carbon in the water was determined
before inoculation and during the tests using a Carlo Erba Instruments NA1500
nitrogen/carbon analyzer.
The fate and effect of B. sphaericus on a nontarget organism were studied
in a single species test systems which held the American oyster (Crassostrea
virginica). For the chronic exposure study, 20 gallon aquaria, each
containing 30 oysters, were used. Half the tanks were inoculated with spores
(= 7 X 10 per ml) while the remaining aquaria served as controls.
Persistence of the spores was monitored in both the bulk water and in oyster
tissue during the 28-day study. Each week histological examinations of the
oyster tissue were performed. Cross sections of the digestive gland and gills
were taken, stained with hematoxylin, eosin and/or the Brown and Brenn tissue
gram stain, and examined by light microscopy.
Results and Discussion
After incubation for 128 days in synthetic sea salts (20 °/oo salinity)
at 4, 15 and 30°C, spore counts in all flasks declined 2 logs. Therefore, no
effect of temperature on spore persistence was observed during 54 days.
However, during the first 35 days of incubation lower spore counts were
obtained in the 30*C flasks suggesting that a certain population of spores was
particularly sensitive to high temperature.
Salinity had no effect on spore persistence. Declines of 2 logs in the
spore count were observed for all salinities (10, 20 and 30 °/oo) after a 131
day incubation at 30°C.
104
-------
High amounts of dissolved organic carbon in synthetic seawater ammended
with 500 mg/L glucose or casamino acids or in natural seawater ammended with
detritus to contain 21.5 mg/L dissolved organic carbon favored spore
persistence after incubations of at least 80 days.
The effects of predators and particulates on spore persistence were
assessed in natural estuarine water. After a 15 day incubation no differences
in spore counts were observed in flasks which contained either filtered (0.2jj)
or unfiltered estuarine water with or without eukaryotic inhibitors. Thus,
predation did not appear to affect spore persistence.
Chronic exposure of oysters to B. sphaericus in the single species test
system resulted in a 99% decrease in spore numbers in the water column over
the 28 day exposure. Conversely, no decrease in the numbers of spores per
gram of oyster tissue was observed throughout the exposure period. We were
unable to definitively identify spores or vegetative cells of fi. sphaericus in
histological sections of oyster tissue. No pathological effects due to B.
sphaericus spores were observed in the oysters.
Anticipated Future Work
The single species test will be repeated with the following
modifications: every three days during the exposure the oysters will be fed a
mixture of three marine algae. In the previous chronic exposure test the
oysters were not fed and starvation may have resulted in abnormal function.
In the previous test the entire oyster was sampled for the spore count. Spore
counts will be performed in future on oyster gills and digestive gland
separately to determine where spores localize. In addition to light
microscopy, electron microscopy will be used to examine thin sections of
oyster tissue for the presence of spores. Should germination of B. sphaericus
spores occur in oyster tissue the identity of vegetative cells can be
confirmed with immunogold staining which uses an antiserum prepared against a
unique protein which covers the surface of this bacterium. A final
modification of the single species test will be a depuration phase. Oysters
will be removed from the chronic exposure tank and placed into flowing sea
water. Samples of oyster tissue taken during depuration will be able to
assess the rate of clearance of the spores from the animal.
105
-------
COLONIZATION AND SURVIVAL IN BIOFILMS
Madilyn Fletcher, Marjorie M. Cowan, and Tessie M. Warren
Center of Marine Biotechnology
Maryland Institute of Biotechnology
University of Maryland
Baltimore, MD 21202
INTRODUCTION
The purpose of this project is to investigate the persistence of
genetically engineered bacteria in aquatic biofilm communities through the use
of laboratory model biofilm systems. In natural environments, biofilm
communities are more stable than planktonic populations on both temporal and
spatial scales. The organisms are immobilized in an extracellular slime
matrix and interact among themselves in a complex manner which may involve a
large degree of co-dependence, thus fostering stability of the community.
This study aims to obtain fundamer'al information on how genetically
engineered bacteria may become introduced into biofilm communities, and
demonstrate some of the specific bacterial properties, both adhesive and
physiological, which can enhance or interfere with the ability of recombinant
microorganisms to colonize or persist in biofilms.
METHODS
Over the past year, experiments have addressed three principal
questions: a) Do the adhesive properties of bacteria, as measured by in vitro
adhesion assays, relate to their ability to colonize mature model biofilms?
b) How do nutrient conditions affect the composition of the model biofilm
communities? c) How do the colonization abilities of Pseudomonas fluorescens,
H2 Tn5, adhesion mutants in mature biofilm communities compare with those of
the wild type strain? To address these questions, two experimental approaches
have been used.
The first experimental system is a series of laboratory microcosms with
which we study long-term (ca. 30 days) colonization by combinations of
bacterial species. Each microcosm is a pulse feed system containing 800 ml
liquid volume and glass coverslip surfaces suspended and held in plastic "0"
rings. The liquid in the vessel is recirculated at a rate of 93 ml/hr, and
100 ml of liquid volume is replaced with fresh medium (0.01% peptone/0.007%
yeast extract in a carbonate buffered minimal salts solution; PYE) every 3
days. The system is mixed by gentle air sparging from the bottom of the
vessel. Test bacteria are successively inoculated into each system. A
variety of combinations of bacteria and inoculation schedules have been used.
However, a typical inoculation schedule is a) inoculation with a coryneform
and Aeromonas hvdrophila on day-1; b) inoculation with one test species on
day-9; c) inoculation with a second test species on day-15.
106
-------
The following measurements are being made:
(1) Numbers of each organism in the bulk phase are regularly
determined by viable counts (samples taken during pulse feeding).
(2) At days 9, 13, 15, and 28, 6 cover slips are removed and treated
as follows. Three are stained with acridine orange (AO) and numbers of
attached cells counted by computer enhanced microscopy. Three are treated
with a Tween/EGTA solution (1 hr with shaking at 150 rpm) for removal of
cells. The solution is then plated out to determine viable counts of removed
cells. The cover slips are retrieved and used for an AO direct count of
remaining attached cells.
From these measurements the following information is obtained:
The rate of surface colonization (approximate numbers of attached cells,
percentage area coverage, and degree of clumping) by the first organisms and
their numbers in the bulk phase are determined by computer-enhanced microscopy
and viable counts, respectively.
The rates of surface colonization by the test organisms are evaluated by
viable counts after removal of attached cells with Tween/EGTA. Numbers of the
different organisms have been determined by antibiotic selection on differen-
tial medium and by different colony morphologies where applicable. It has not
been possible to distinguish them on the surface by microscopic observation,
and we are currently devising a method to do this.
The "strength" of biofilm adhesion to the surface is-indirectly
evaluated by determining the cells which remained attached to the surface
after treatment with Tween/EGTA.
These experimental microcosms have been used to determine a) the
relationship between adhesive abilities qf given strains and transposon
mutants and the colonization ability in biofilms, and b) the effect of
nutrient concentration on community composition.
The second experimental system is 75-ml batch cultures in 250-ml flasks,
which has been used to refine methods or to obtain results of a more prelimi-
nary nature to be tested subsequently in jthe microcosms. These batch cultures
are inoculated with test bacteria using inoculation schedules similar to those
used for microcosms, and they are pulse-fed. They differ from the microcosms
primarily in their mixing characteristics, but require less time and effort.
RESULTS AND DJISCUSSION
i
The microcosm and batch culture experiments have generated a large
amount of data, indicating various trendsi and relationships among biofilm
colonizers. The principal results at this stage are discussed below.
a) The relationship between adhesion ability, as measured by in vitro
assay, and colonization in biofilms. The numbers and persistence in biofilms
of two bacteria, a highly adhesive Xanthomonas maltophilia and poorly adhesive
107
-------
A. hvdrophila. were found to be related to their adhesive abilities. In these
experiments, a typical microcosm inoculation schedule was as follows:
inoculation with the coryneform and A. hvdrophila on day 1 in duplicate
microcosms, inoculation with X. maltophilia on day 8 in one of the duplicates,
and inoculation with Pseudomonas fluorescens. H2, on day 15 in both
microcosms, followed by sampling for a further 11 days to evaluate
colonization. The presence of the highly adhesive Xanthomonas resulted in
increased coverage of the attachment surface and in a greater resistance of
the biofilm to disruption in the EGTA/Tween solution used to disperse
biofilms.
Table 1. Effect of presence of X. maltophilia on biofilms produced by A.
hydrophila. coryneform, and P. fluorescens.
DAY CELLS IN CONTACT WITH GLASS* % SURFACE COVERAGE
+ X. malt. - X. malt. + X. malt. - X. malt.
1 (A. h., c.)
8 X.m. 9 7 3.5 2.9
15(P.f_l.) 46 26 18.8 10.8
17 77 30 31.6 12.2
23 130 - 51.8
27 81 22 32.9 8.4
* 106 cells/232 mm2. Determined by dividing total surface area covered by
area of one cell.
In contrast, experiments with Aeromonas have illustrated another
important trend relating adhesiveness with biofilm colonization. This strain
is poorly adhesive according to i_n vitro assays, and similarly it is to a
large extent displaced by P. fluorescens H2-416, when subsequently inoculated.
Thus, these experiments demonstrate that the adhesive ability of
different strains can have an influence on their ability to colonize and
persist in mixed-species biofilm communities. However, adhesive ability is
not the only property influencing colonization success, as was demonstrated in
experiments testing the effect of nutrient concentrations. The trends
identified by the experiments described above did not persist at all nutrient
concentrations, as discussed in the following section.
b) Nutrient concentration influences biofilm concentration. The
experiments described in section 1 all utilized a medium comprising 0.01%
peptone and 0.007% yeast extract (PYE). Experiments were then extended to
test the effect of media containing 0.01, 0.1, and 10 times the original
concentration. The coryneform and Aeromonas were used initially to inoculate
the microcosm, and P. fluorescens H2 was inoculated 14 days later. As
expected, total biofilm numbers progressively decreased with 0.1 and 0.01 PYE,
but were unchanged with 10-fold PYE concentration. However, the relative
proportions of the different species altered with nutrient concentration. The
coryneform was dominant at the two lowest nutrient concentrations, with both
Aeromonas and Pseudomonas ca. almost an order of magnitude lower in numbers.
However, with an increase in nutrients to the standard PYE concentration,
Pseudomonas numbers increased to match those of the coryneform, and with a
108
-------
further increase to ten-fold PYE, Aeromonas numbers increased to match both
the coryneform and Pseudomonas. The results indicate that not only are
adhesive abilities important in determining biofilm community composition, but
also probably the ability of different bacterial constituents to compete for
scarce nutrients or alternatively to take advantage of nutrients in ample
supply. For example, the coryneform appears to have been able to be the
dominant colonizer when nutrients were comparatively scarce, whereas the
growth of Aeromonas increased more than the other bacteria in response to the
highest PYE concentration.
c) Biofilm colonization by transposon adhesion mutants. The ability
of Tn5 adhesion mutants to colonize surfaces in batch culture experiments is
consistent with their adhesion ability, as determined by in vitro assays
(Table 2). Thus, these experiments support data from microcosm experiments
(Table 1 above) indicating that adhesive ability of bacterial strains is
significant in determination of biofilm composition.
Table 2.
Coverage in vitro
Mutant added (cells/sq mm) % change adhesion
on Day 9 Day 9 Day 15 in coverage index
H2-2 1.9 x 105 4.4 x 104 -76 1.20
H2-3 1.9 x 105 1.2 x 10s -36 1.31
H2-4 1.5 x 105 5.2 x 10s +338 1.49
H2-6 1.8 x 105 5.2 x 10s +287 1.49
* (A490 mutant) / (A49Q Parent)
ANTICIPATED FUTURE WORK
Over the next year, work will be concentrated in two principal areas.
First, we shall address the relationship between biofilm composition and
physiological characteristics of the colonizing bacteria. Substrate uptake
characteristics of the strains will be individually assessed using radio-
tracers in pure culture batch experiments. Subsequently, the assimilation of
substrate by bacteria in the biofilms will be measured to determine to what
extent substrate utilization is influenced by the biofilm microenvironment or
characterized by close proximity of neighboring bacteria and a polymeric
intercellular matrix. At the same time, the colonization of biofilms by
adhesion mutants of £. fluorescens will be studied in detail to confirm the
significance of adhesive ability in biofilm community composition and deter-
mine the bacterial surface chemistries which contribute to colonization
success.
109
-------
ENVIRONMENTALLY INDUCED GENETIC INSTABILITY IN MICROORGANISMS
Tyler A. Kokjohn and Robert V. Miller
Department of Biochemistry and The Program in Molecular Biology
Loyola University of Chicago
Maywood, IL 60153
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 environ-
mental insult.
Most of the data available concerning DNA repair has utilized the
enteric bacterium Escherichia coli under conditions which probably 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 vary 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 micro-
organisms (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.
aeruqinosa was determined for cells incubated in growth medium in the labora-
tory and for cells placed in biological containment chambers followed by
incubation 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. aeruqinosa strains. Chambers inducted in situ were placed at locations
receiving solar UV radiation for significant periods of the day or in
locations substantially shielded from sunlight.
Cells incubated in the chambers were recovered 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 placed on 0.1 X YEPG agar and incubated for
110
-------
48 h to recover cells. The cells were then repllea-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 ability of cells existing under starvation conditions to resist
stress was determined by laboratory experimentation. Cells were grown in 0.1
X YEPG at room temperature to stationary phase. The cells were diluted one to
ten 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 the recA gene product of P. aeruginosa and £. coli 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 electro-
phoresed on denaturing polyacrylamide gels and the proteins transferred to
nitrocellulose membranes. These membranes were subjected to Western analysis
using a biotinylated-streptavidin-horseradish peroxidase-conjugated antibody
detection system. Anti-recA £. coli antibody (gift of S. Kowalczykowski,
Northwestern University Medical School, Chicago, Illinois) was used as the
primary antibody.
Fusions of a promoterless b-galactosidase gene to stress-inducible genes
of P. aeruqjnosa PAO were constructed using a modified Tn3-transposon. This
transposon will insert into target DNA yielding transcriptional fusions of the
disrupted gene and b-galactosidase.
For prophage induction experiments, F116L or D3 lysogens of P.
aeruginosa were placed in chambers and incubated in the laboratory or vn 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% NaCl and plating with phage-sensitive indi-
cator strains. In some laboratory experiments lambda lysogens of £. coli were
treated in a similar manner.
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 determina-
tion of mutation frequency. The experimental protocols which we are now
utilizing have been modified from our original proposal and reflect experience
111
-------
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.
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 adsorbed by and lyses even starved P.
aeruqinosa cells. Other virulent laboratory phages were unacceptable since
changes in growth conditions resulted in declined virulence or reproduction
potential.
Experiments accomplished to date suggest that P. aeruqinosa cells
incubated under conditions mimicking those of natural ecosystems will produce
a higher frequency of mutation-containing cells in their populations. At
present, it is not clear what factor or factors cause this increase.
Enhanced Resistance to UV Stress
The effects of starvation on the response of bacterial cells to DNA
damage are dramatic. The physiological adaptations required for.viability
maintenance result in a tremendously enhanced ability to re-sist UV irradia-
tion. Preliminary experiments suggest that the genes required for DNA repair
(recA. uvrABC. etc. in E. coli) are expressed at enhanced, constitutive
levels. Western Blotting experiments performed on both E. coli and P.
aeruqinosa support this conclusion.
Induction of Prophages
Laboratory and field experiments have also demonstrated that starvation
of £. coli and P. aeruqinosa lysogens greatly affects DNA-damage induction of
resident prophages. Preliminary experiments using the temperate P. aeruqinos-
a bacteriophage, F116, have suggested that conditions encountered in natural
aquatic environments may lead to enhanced activation of lytic phage production
in these cells. We are currently investigating this phenomenon in greater
detail. 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. aeruqinosa and lambda of E. coli. Field experi-
ments have demonstrated no significant, long-term solar irradiation as
measured by the frequency of ICs in the population. While the frequency of
spontaneous ICs is not reduced in populations of D3 lysogens incubated under
starvation conditions for extended periods, £. coli lysogens of prophage
lambda are severely impaired in their ability to produce a burst of phage
subsequent to UV irradiation if they have been starved.
112
-------
Stress-Induclble Expression of P. aeruginosa Genes
To aid in the quantification of levels of expression of DNA-damage
indueible (din) genes under environmental conditions, such as starvation and
exposure to solar UV, we have constructed and partially characterized trans-
criptional fusions of various £. aeruglnosa PAO genes to b_-galactosidase
(lacZ) which are inducible by exposure to ONA damaging agents. These genes
are clearly expressed at greatly increased levels subsequent to UV irradiation
of the cell. The fusions are inducible by other agents which damage DNA as
well.
Plasmid-Encoded UV Resistance
Laboratory studies of the UV-resistance plasmid R2 have revealed that it
encodes a true rec,A-dependent, ONA damage-inducible repair system which is
highly mutagenic. Laboratory and field studies of P. aeruqinosa cells
containing the UV resistance plasmid R2 have demonstrated an enhanced
mutagenesis activity under environmental 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 introduced into these environments by the release of
GEMs is absolutely vital in risk and fate assessment.
ANTICIPATED FUTURE WORK
A major emphasis in the future will be placed on repetition of the field
experiments on mutation frequency using the protocols developed over the past
year. Such repetition is essential if accurate statistical models of stress-
induced genetic change are to be developed. Field experiments will be
conducted on a year-round basis to estimate effects of changing temperature
and solar UV intensity.
Experiments will be conducted using the cloned din gene-b-galactosidase
fusions of P. aeruqinosa. These experiments will reveal the levels of
expression of the din genes under the conditions of nutrient limitations and
stress present in natural aquatic ecosystems. The use of gene fusion technol-
ogy will allow a correlation to be made between environmental conditions and
the level of expression of the din genes.
The expression of UV-resistance plasmid genes will be studied in the
laboratory and field. An action spectrum of the solar radiation wavelengths
capable of simulation expression of the mutagenesis activity of the plasmid
will be determined. In collaborations with Drs. T. Barkay (Gulf Breeze
Laboratory) and M. Day (University of Wales) we have identified numerous heavy
metal-resistance plasmids which code for enhanced UV resistance. These
plasmids will be further analyzed as described for plasmid R2.
113
-------
PUBLICATIONS
Kokjohn, T.A., and R.V. Miller. 1990. Gene transfer in the environment:
transaction, in J.C. Fry and M.J. Day (eds.), Release of Genetically
Engineered and Other Microorganisms, Edward Arnold, London (in press).
Miller, R.V., T.A. Kokjohn, and G.S. Sayler. 1990. Environmental and
molecular characterization of systems which affect gene alteration in
Pseudomonas. In S. Silver, A.M. Chakrabarthy, B. Igluski, and S. Kapland
(eds.), Pseudomonas '89: Biotransformations, Pathogenesis, and Evolving
Biotechnology, American Society for Microbiology, Washington, D.C. (in press).
Miller, R.V., and T.A. Kokjohn. 1990. General Microbiology of recA:
Environmental and Evolutionary Significance. Annual Review of microbiology
(In press).
Kokjohn, T.A. 1989. Transduction: mechanism and potential for gene transfer
in the environment, pp. 73-99. in S.B. Levy and R.V. Miller (eds.), Gene
Transfer in the Environment, McGraw-Hill, New York.
Simonson, C.S., T.A. Kokjohn, and R.V. Miller. 1990. UV repair potential of
Pseudomonas aeruqinosa PAO. Journal of General Microbiology. (In press).
114
-------
HOW EVOLUTION AND pSClOl file STABILIZE PLASMIDS
Toai T. Nguyen, Jacques Lanners, and Richard E. Lenski
Department of Ecology and Evolutionary Biology
University of California
Irvine, CA 92717
INTRODUCTION
In order to standardize future research on genetically engineered
microorganisms (GEMs), a series of EPA benchmark plasmid vectors (pR02313,
pR02317, pR02318, pR02320, and pR02321) were constructed. Before these
vectors could be used in experiments to assess the stability of biodegradative
genes in GEMs, their own stability properties were studied in detail. We
observed that all the above EPA benchmark plasmids were unstable to various
degrees, depending on the bacterial host and the growth medium. They were
less stable in Pseudomonas aeruqinosa PAOlc than in P. outida pRS2015 and less
stable when grown in rich medium than when grown in minimal salts medium.
Because pR02320 was the least stable plasmid of the five, any event
stabilizing it would be easily detected. Furthermore, since pR02320 confers
resistance to tetracycline, which does not involve deactivation of the
antibiotic, bacteria could be forced to carry pR02320 for an extended period
of time in media containing an appropriate concentration of tetracycline so
that plasmid stability could be enhanced through evolution. Presently, we
report that after maintaining a population of PAOlc bearing pR02320 in L broth
containing tetracycline for 930 generations by periodic sub-culturing, we
successfully isolated a deletion derivative of pR02320 which was more stable
than the parental plasmid, when the host was growing in an antibiotic-free
medium.
A plasmid could also be made stable by the insertion of a stabilizing
function, such as the pSClOl par locus. With the intent to use pSClOl par to
stabilize recombinant plasmids containing genes coding for important bio-
degradative functions, we investigated how pSClOl par exerted its stabilizing
effects on heterologous replicons such as plasmids derived from pACYC184, in
Escherichia coli K12 strain PM191 (thv* recA dra thr leu thi drm) under
different culture conditions. Presently, we report that only under conditions
of growth at 4-5 hour generation time in supplemented minimal salts medium,
the pSClOl par locus could stabilize a heterologous replicon by a novel
mechanism, namely by enhancing the growth rate of the bacterial host bearing
the recombinant plasmid.
METHODS
Isolation of Evolved Host and Evolved Plasmid
The evolved PAOlc strain (PAOlc-2-140) and pR02320 plasmid (pR02320-2-
140) were obtained as follows. A single PA01c/pR02320 colony was used to
found a population of plasmid-bearing cells. This population was propagated
by periodic sub-culturing (see below) at 1:100 dilution every 24 hours. At
115
-------
the end of the 140-day period, a plasmid-free isolate of the evolved host
(PAOlc-2-140) was identified by testing single colonies for growth on L agar
medium but not on L agar medium containing 80 ng/£ Tc and evolved pR02320
plasmid (pR02320-2-140) was isolated from a plasmid-bearing isolate.
Construction of Plasmid pMPll
pMPll was constructed and propagated in E. coli B JB11. Essentially,
pACYC184 DNA was first restricted with EcoRI and Aval to liberate a 1326 base
pair (bp) fragment; then products of the digest were incubated with Klenow
fragment and deoxyribonucleoside triphosphates to fill in the staggered ends;
next the DNA fragments were incubated with T4 DNA ligase; and finally plasmid
DNA was transformed into JB11. Plasmid DNA was isolated from transformants
resistant to tetracycline and sensitive to chloramphenicol and analyzed by
restriction enzymes to identify the appropriate construct. Plasmid-free PM191
was transformed with pMPll and stored in 40% glycerol at -80°C.
To ensure true isogenicity, plasmid DNA was isolated from plasmid-
bearing strains and re-introduced into plasmid-free PM191 and the
reconstructed strains were stored in 40% glycerol at -80°C. All experiments
were started with single colonies isolated from those stocks.
Stability and Competition Experiments
The frequencies of plasmid-bearing cells were monitored for 40-120
generations of growth in antibiotic-free medium using two types of experiments
run in parallel: stability and competition. In the stability experiments, a
single clone of plasmid-bearing cells was used to found a population. In the
competition experiments, a 1:1 mixture of plasmid-bearing and isogenic
plasmid-free cells was used to found a population. Both types of experiments
can be performed under a regimen of either periodic subculture or continuous
culture in chemostats.
Experiments run by way of periodic subculture consisted of inoculating
antibiotic-free liquid medium with a serial dilution of saturated culture from
the previous growth period. At each sub-culturing, samples were serially
diluted and plated for single colonies on antibiotic-free medium, and 100 to
200 colonies were subsequently streaked on L agar containing Tc to determine
the frequency of plasmid-bearing cells.
Continuous cultures were maintained in chemostats. The volume of each
chemostat was measured at the end of the run and the generation time calcu-
lated using the average flow rate monitored throughout the run. The frequen-
cies of plasmid-bearing cells were determined as before at regular intervals.
All liquid cultures and agar plates were incubated at 37°C.
116
-------
Data Analysis
Data were analyzed by least squares non-linear regressions fit to the
following equation:
[u + s(l - P0)]e(u*s)t + sP
0
where P^ is the initial fraction of plasmid-bearing cells and t is the elapsed
time of growth in antibiotic-free medium, to estimate segregation rates, u,
and selection coefficients, S.
RESULTS AND DISCUSSION
Stabilization of Plasmids by Evolution
We determined that after growing for 930 generations in LB containing
tetracycline, the plasmid-bearing strain (PA01c/pR02320)-2-140 did not lose
its plasmid as readily as the unevolved one in stability and competition
experiments. We reconstructed four combinations of plasmid-bearing strains
using unevolved and evolved plasmid-free hosts (PAOlc and PAOlc-2-140), and
unevolved and evolved plasmids (pR02320 and pR02320-2-140) . We determined
that the evolved plasmid and not the host genome was responsible for the
stabilizing effect. Further restriction analyses showed that the evolved
plasmid had a 2.2 x 103 base pair deletion. Comparing the unevolved and
evolved plasmid sizes, the relative energetic gain for not synthesizing the
extra DNA present in the unevolved plasmid could not account for all the gain
in plasmid stability. The deletion probably removed an unnecessary function
which had proved costly for the host to maintain.
Stabilization of Plasmids by pSClOl par
We conducted stability and competition experiments for pACYC184, a p!5A-
derived replicon; pMPll, constructed by deleting a fragment from pACYC184; and
pPM31, in which the same deletion as in pMPll was replaced by a fragment
containing pSClOl par. With periodic subculture at 1:106 dilution in LB every
12 hours (condition A), we observed very low segregation rate for all three
plasmids in stability experiments. However these plasmids were unstable in
competition experiments (i.e., these plasmids were readily lost from the
population if the cultures were seeded with isogenic plasmid-free bacteria),
thus indicating that their instabilities were caused mostly by selection.
Under condition A, pSClOl par (in pPM31) seemed not to have any effect on the
stability of plasmids. With periodic subculture at 1:104 dilution in supple-
mented minimal salts medium (PG) every 24 hours (condition B), all three
plasmids were unstable, chiefly due to selection. However pMPll and pPMSl
were significantly more stable than their parental plasmid, pACYC184. Under
condition B, the deletion in pMPll seemed to reduce the intensity of selection
against plasmid carriage. In continuous culture in PG with generation times
of 2-3 hours (condition C), although all three plasmids were unstable, pPM31
seemed to be more stable than either pACYC184 or.pMPll. Under condition C,
117
-------
pPM31 was stabilized to some degree by the pSClOl par insert through the
reduction of plasmid carriage burden. In continuous culture in PG with
generation times of 4-5 hours (condition D), the loss of pACYC184 was caused
by selection; pMPll was somewhat stable; and pPM31 was stabilized by a growth
rate advantage of plasmid-bearing cells. Under condition D, pSClOl par
stabilized pPM31 by increasing the growth rate of its host bacterium.
ANTICIPATED FUTURE WORK
We plan to further characterize the deletion in pR02320-2-140 that
caused this plasmid to be more stable than its progenitor, in order to
elucidate the underlying process(es). To take advantage of the enhanced
stability of pR02320-2-140, we will use this plasmid in the stability study of
the chloro-catechol degradative gene cluster (clc). We also plan to study the
effectiveness of the plasmid stabilizing activity of the pSClOl par locus in
Pseudomonas spp. and to use this locus to stabilize plasmid vectors such as
pR02320, or recombinant plasmids containing important biodegradative genes.
PUBLICATIONS
Lenski, R.E., and J.E. Bouma. 1987. Effects of segregation and selection on
instability of plasmid pACYC184 in Escherichia coli B. J. Bacteriol.
169:5314-5416.
Lenski, R.E., and T.T. Nguyen. 1988. Stability and 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.
118
-------
FIELD RELEASE OF A GENETICALLY ENGINEERED BACULOVIRUS
WITH A LIMITED SURVIVAL CAPACITY
H. Alan Wood, Patrick R. Hughes, and Martha Hamblin
Boyce Thompson Institute for Plant Research
Ithaca, NY 14850
INTRODUCTION
Baculoviruses are ecologically attractive alternatives to synthetic
chemical insecticides because they are target specific and have been shown to
pose no environmental or health hazards. However the commercial efficacy of
viral pesticides is limited because viruses kill insects more slowly then
chemical pesticides. Several laboratories are attempting to overcome this
problem through genetic engineering. By the insertion and expression of new
pesticidal genes in baculoviruses, the new gene product(s) will cause a
cessation of feeding or rapid death of insect pests.
The field release of engineered baculoviruses will have to take into
consideration health, ecological and environmental issues. Accordingly we
have developed and are evaluating under laboratory and field conditions a
strategy whereby these new genetically engineered viruses can be released into
the environment in a safe and effective manner. Most baculoviruses have a
nonessential gene coding for polyhedrin protein which crystallizes around the
virus particles, thereby protecting them from inactivation. The polyhedrin
gene can be removed and replaced with foreign genes which are expressed during
replication. This is the basis of the widely used baculovirus expression
vector system, and it can easily be used to insert pesticidal gene(s) into
baculoviruses. Such an engineered virus would be environmentally safe because
in the absence of polyhedrin protein occlusion the virus could not persist in
environment. However, the naked virus particles are so unstable that they
could not be delivered to the field in an active form.
In order to stabilize the infectivity of a polyhedrin-minus virus and
limit its persistence in the environment, we have developed a co-infection:co-
occlusion procedure. Insect cells are co-infected with an engineered virus
which has a deletion of the polyhedrin gene and an unaltered wild type virus.
The polyhedrin protein produced by the wild type virus co-occludes both the
wild type and the engineered virus particles.
METHODS
The virus used in this study was the Autoqrapha californica nuclear
polyhedrosis virus (AcMNPV). The virus was propagated in Trichoplusia nj_
larvae as well as T. rn. and Spodoptera frugiperda tissue culture cells.
Using standard genetic engineering techniques, the region of the AcMNPV
genome which contained the polyhedrin gene was cloned. Base deletions were
made in the 5' region of polyhedrin gene, and this mutated fragment was
introduced into the virus through recombination between the viral and mutated
fragment DNAs.
119
-------
Polyhedra containing co-occluded wild type and polyhedrin-minus virus
particles were initially obtained by co-infection of tissue culture cells and
then mass propagated in T. ni larvae. The amounts of occluded, polyhedrin-
minus virus present in polyhedra from individual larvae were determined by DMA
restriction fragment profiles. The polyhedra were purified, virus released
and viral DNA extracted. The viral DNA was digested with EcoRI and fraction-
ated by agarose gel electrophoresis. The percentage of polyhedrin-minus virus
was determined from the ratios of the complete (wild type) and truncated
(polyhedrin-minus) EcoRI fragments.
Trichoplusia rn. larvae were fed polyhedra containing varying ratios of
both virus types at dosages ranging from LD,. to 100 x the LD100. Polyhedra
were purified from individual insects, and the amount of co-occluded,
polyhedrin-minus virus determined.
RESULTS AND DISCUSSION
The laboratory studies have been conducted to evaluate the persistence of
co-occluded, polyhedrin-minus AcMNPV. The data showed that persistence of the
polyhedrin-minus virus in the virus population requires high inoculum levels
of polyhedra containing equal amounts of the two virus types. These dosage
levels and ratio of virus types could not be maintained under natural condi-
tions. At low dosages, a majority of the larvae become infected with only the
wild type or engineered virus. Therefore, as the viruses are passed from
insect to insect, there is a low probability of co-infection of individual
insects and cells. In the absence of occlusion the polyhedrin-minus virus
loses 100% of its infectivity within 3 days at room temperature in the
laboratory.
A field release to evaluate the persistence of a co-occluded, polyhedrin-
minus AcMNPV was initiated during the summer of 1989. Polyhedra containing
equal amounts of wild type and polyhedrin-minus virus particles were applied
to a quarter acre circular cabbage plot which was seeded with T. n_j (cabbage
looper) larvae. The application site was surrounded by a three-quarter acre
buffer zone of cabbage plants. Three applications were made with a backpack
sprayer at dosages of 3 x 1012 polyhedra per acre.
During each application, spray drift (virus dispersal) was measured.
Open petri plates containing insect diet were located symmetrically around the
application site during the applications. Immediately after spraying, the
dishes were covered and brought to the laboratory where ten neonate host
larvae were placed on each plate to detect the presence of virus deposition.
This assay could detect from 50 to 100 polyhedra^per plate. Virus deposition
was detected up to 130 feet from the application area.
When the first dead larva was detected following each of the applica-
tions, 200 larvae were collected and placed individually on diet in order to
evaluate the percent infection. One hundred percent infection was obtained
following each application. Accordingly, the secondary inoculum produced in
the remaining larvae in the field is calculated to be more than 101 polyhedra
per acre. The polyhedra from these larval samples will now be analyzed to
determine the amount of polyhedrin-minus virus present.
120
-------
During the next two years, the entire test site will be replanted with
cabbage plants and seeded with cabbage looper larvae. The polyhedra in the
soil are expected to contaminate the foliage leading to infection and produc-
tion of succeeding generations of virus polyhedra. Insect samples will be
analyzed to determine the amount of occluded, engineered virus produced in
each generation. In addition, polyhedra will be extracted from the soil and
analyzed.
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 pesti-
cides. Proceedings of the International Conference on Insect Neurochemistry
and Neurophysiology. Humana Press (in press).
Wood, H.A., and A.M. Shelton. 1990. Genetically-Modified Viruses and
Microbes: Environmentally Safe Alternatives to Chemical Pesticides. The
World and I (in press).
121
-------
CHARACTERIZATION OF EPIPHYTIC FITNESS GENES
Steven E. Lindow, G. Andersen, M. Wilson, and N.J. Panopoulos
Department of Plant Pathology
University of California
Berkeley, CA 94720
INTRODUCTION
The overall objective of these studies is to predict the survival and
behavior of applied bacterial strains and subsequent generations of these
cells in natural environments. Preliminary information available indicated
that the physiological state of cells released into the open environment had a
large effect on their subsequent growth and/or survival . Mutant strains were
also available which had altered epiphytic fitness in different environmental
conditions. Studies have been initiated, therefore, to further elucidate
important phenotypes and physiological adaptations present in bacteria which
allow them to survive harsh environmental conditions. This work is justified
because predictions of the environmental fate of released recombinant micro-
organisms will normally be based on studies of these organisms in culture or
in simplified environments such as microcosms. We must be assured that the
traits bacteria might express that would be pertinent to its eventual fate in
the environment will also be expressed in these simplified laboratory
settings.
The objectives of this study were:
1. Determine the influence of pre-adaptation of cells on their location and
survival on leaves subsequent to their inoculation.
2. Identify phenotypes conferred by genes affecting survival during environ-
mental stress and determine the conservation of these genes or phenotypes
in different bacterial genera.
RESULTS AND DISCUSSION
Preliminary studies had indicated that bacterial cells cultivated on
solid or liquid bacteriological culture media exhibited poor survival when
exposed to plants either as a dilute sprayed solution or when transported to
plants as dry aerosol particles. We hypothesized that adaptations of bacteria
have evolved for survival on leaf surfaces and other rigors of the physical
environment encountered during their growth and survival on leaves, or when
transported to new habitats such as soil or water, and that these traits are
not efficiently expressed in bacteriological culture media but instead only on
leaf surfaces. Studies have been undertaken to verify these preliminary
results and to further identify physiological mechanisms involved in epiphytic
fitness that might be expressed on leaf surfaces but that are not expressed in
culture. Recent studies conducted in our laboratory have now verified that
cells of P. svrinqae strains Cit7 and TLP2 that have been cultured on solid
agar surfaces behave differently from those cultured on liquid media, and from
those cells collected directly from plant surfaces, when reinoculated onto the
122
-------
surface of plants. Cells grown on solid agar surfaces and collected from
plants nearly always exhibit an enhanced survival rate when the plants on
which they were inoculated were allowed to dry, or are exposed to high
temperatures, low relative humidities, or both. These data confirm the
preliminary studies. Studies have been initiated to ascertain what physiolog-
ical differences exist between laboratory-cultured cells and cells grown on
plant surfaces. Preliminary studies now indicate that cells grown on plant
surfaces exhibit a lower sensitivity to ultraviolet light, a subtly reduced
sensitivity to dehydration under laboratory conditions, and possibly an
increased rate of mobility responses (flagellar motility). These studies will
be repeated within the next two months and further delimitation of the
conditions under which strains exhibit altered response to physical environ-
ments, or behavioral effects such as motility, between laboratory- cultured
and plant-cultured cells will be conducted.
At the time of the initiation of this work, genes had been identified
within Pseudomonas svrinqae strain B728a which confer survival on leaf
surfaces exposed to harsh environmental conditions (Tn5-induced mutants of P.
syringae with reduced ability to survive when wet leaves of plants such as
bean were allowed to dry were identified). Studies have been initiated to
identify the physical conditions operative during the differential responses
of mutant strains and the parental strain of P. syringae. Mutant and parental
strains of P. syrinqae strain B728a have been exposed to different relative
humidities on filters and will soon be exposed to different matric potentials
on filters placed on the surfaces of agar plates containing polyethylene
glycol. The results of these studies have not yet been fully analyzed, but
some differences between mutant strains and the parental strain B728a in
survival of low water potentials have been noted. To better enable an
evaluation of the temporal, spatial and physiological stimulants of the
expression of these genes, which appear to be involved in stress tolerance of
P. syrinqae on leaves, the genes are being further characterized. At the time
of the initiation of this Cooperative Agreement, it was unknown whether stress
tolerance genes were found within a single region of the genome of P. syrinqae
or were dispersed widely throughout the chromosome. To evaluate this, the
genome of P. syrinqae was cut with restriction enzymes (such as Xbal) with 8-
base-pair recognition sequences that produce large restriction fragments (in a
range from 10 to 100 kb in length). These restriction fragments were sepa-
rated by pulsed-field gel electrophoresis and the presence of Tn5 within
different-sized fragments was verified by Southern blot analysis using
radiolabeled Iambda-Tn5 as a probe in DNA hybridizations. Preliminary
evidence indicates that several different regions of the P. syrinqae genome
are involved in the stress tolerance response. Individual restriction
fragments (Sau3A partial and EcoRI complete digests) of mutant strains
containing Tn5 are being isolated, and corresponding flanking sequences are
being identified in a cosmid (pLAFR3) library, of the £. svrinqae B728a
parental strain. This will allow the identification of native sequences
involved in stress tolerance. Ice" derivatives of P. syrinqae strain B728a
are also being constructed by site-directed mutagenesis using cloned ice
nucleation genes to produce an Ice" derivative suitable for studies involving
a promoterless ice nucleation gene (pTn3-Spice), which will be used as a probe
for promoter activity of the unidentified stress tolerance genes. This will
aid in identifying the physiological conditions under which they are
123
-------
expressed, and therefore hopefully the functions that they confer which lead
to stress tolerance.
Genes in £. herbicola (a common leaf-surface inhabitant with a habitat
which is likely to be distinct from that of P. svrinqae) are being identified
by random Tn5 mutagenesis followed by screening of Tn5 mutants for differen-
tial survival and/or growth on leaf surfaces exposed to alternating favorable
and unfavorable physical conditions. This has required the development of a
delivery system for Tn5 into the chromosome of £. herbicola. A suitable
"suicide" plasmid vector has now been constructed in our laboratory to
efficiently introduce Tn5 into the chromosome of £. herbicola. The plasmid
pLAFR3 containing the cloned levan-sucrase gene from Bacillus subtilis has
been constructed. E. herbicola containing plasmids expressing levan-sucrase
activity are killed when cells are grown on medium containing high levels of
sucrose. Therefore selection against the retention of plasmids containing the
levan-sucrase gene is possible by simple growth on sucrose medium. We have
found that pLAFRS containing Tn5 and the levan-sucrase gene is very unstable
in E. herbicola and that Tn5 transposes at a rather high frequency (greater
than 10° per generation). Studies are underway to identify the frequency of
generation of auxotrophic mutants within E. herbicola by this Tn5 suicide
delivery system and to ascertain the randomness of the insertion of Tn5 into
the chromosome of E. herbicola strain BRT9S. Simultaneous to conducting these
genetic steps, studies have been made of the growth dynamics of E. herbicola
strain BRT98 on bean plants under alternating wet and dry conditions in a
growth chamber. Tn5 mutant strains are now being assessed for their relative
growth on bean leaves compared to parental strains following a cycle of two or
more wet and dry periods on beans under growth chamber conditions- when
population sizes are estimated by a leaf-freezing assay.
124
-------
ENVIRONMENTAL FATE OF Aqrobacterium radiobacter K84
RELEASED IN AGRICULTURAL FIELDS
L. Moore,1 V. Stockwell,1 J. Loper,2 and M. Kawalek1'2
Department of Botany and Plant Pathology1
Oregon State University
USDA Agricultural Research Service
Horticultural Crops Research Laboratory
Corvallis, OR 97331
INTRODUCTION
Aqrobacterium radiobacter K84 is an EPA-registered microbial pesticide
used successfully in numerous parts of the world for biological control of
crown gall disease of plants. The success of this biological control is
correlated with the production of agrocin 84, a specific antibiotic against
the crown gall pathogen, A. tumefaciens. Genes coding for synthesis of
agrocin 84 are located on a conjugative plasmid present in the wild-type
strain and can be transferred under laboratory conditions to A. tumefaciens.
There is one report of this genetic transfer occurring in a field habitat,
resulting in a breakdown of the biological control.
In addition to investigating the possibility of horizontal gene transfer
from K84 to a recipient, more information was needed about the persistence and
dissemination of K84, or a genetically modified K84, in soil, on target and
nontarget plants, and in surface and ground water associated with the agroeco-
system. Ecological information about survival and dissemination of K84 has
been difficult to obtain because of the lack of reliable methods for detection
and monitoring.
Objectives of this research included development of new or improved
diagnostic methods to study wild type K84 in the environment. The focus was
on more specific, sensitive techniques for detection, isolation, and identifi-
cation of the target organism among the myriad of other soil microbes, namely
the use of a) a strain-specific antiserum to K84 and b) DNA probes of two
fragments from the agrocin biosynthetic region. The efficacy and limitations
of these methods for ecological studies were compared to traditional methods
of selective media, antibiotic-resistant ("marked") strains, and serology
while investigating the environmental fate of K84 in agricultural fields.
METHODS
Seroloqy
Strain-specific polyclonal antibodies to A. radiobacter K84 for use in
immunoisolation and identification experiments were tested for titer by
microprecipitation and specificity by double diffusion and immunofluorescence
microscopy.
125
-------
Immunoisolation
Preliminary experiments were conducted to determine if K84 could be
isolated selectively from aqueous bacterial suspensions with antibodies.
Sterile wooden tongue depressors were coated with polyvinyl buterol, and
polyclonal antibodies to K84 (1:250 dilution) were adsorbed onto the plastic-
coated sticks. The immunoisolation sticks were submerged into a suspension of
K84 of known concentration for 1 hr and removed. Then, the bacterial suspen-
sion was diluted to determine the number of bacterial cells remaining in the
suspension. The sticks were rubbed onto the agar surface of MGY plates to
confirm adherence of the bacteria to the antibody-coated sticks.
Immunoblotting for Identification of K84 Colonies
Immunoblotting is a technique where bacterial colonies are attached to a
solid support, reacted with antibodies made to the target strain, and then
reacted with a secondary antibody marked with an enzyme. After staining for
the marker enzymes, the target organism is cultured on isolation plates and
can be identified, enumerated, and reisolated.
Field Experiments (1988)
Field experiments were planted in May 1988 to compare the survival and
biological activity of wild-type strains of Aqrobacterium tumefaciens and A.
radiobacter. K84 with antibiotic-resistance marked mutants of the same
strains on Mazzard cherry seedlings. Plots were located at Ephrata, WA, Moses
Lake, WA, Aurora, OR, and Corvallis, OR, and harvested during October.
The number of infected and non-infected cherry seedlings in each plot was
recorded. Roots were examined for the presence of galls, and the relative
size and number of galls on each root system was recorded. As the seedlings
were removed from the soil, root samples were collected from each block at
each site, except for the plot at Moses Lake, WA, to determine the rhizo-
sphere populations of the introduced bacteria. Root samples were packed on
ice in a cooler and transported back to Corvallis for processing. The tap
root was separated from the fibrous roots, rinsed briefly, weighed, and
agitated for 30 minutes in a known volume of sterile peptone-phosphate buffer.
The suspension was dilution-plated onto semi-selective medium (Kerr 1A for
Biovar I or Kerr 2E for Biovar II strains) with some of the medium amended
with antibiotics (100 ug/ml rifampcin, 300 ug/ml naladixic acid, or 500 ug/ml
streptomycin) for specific recovery of antibiotic-resistant marked strains.
Populations were determined after incubation of plates for 7 to 10 days at
25C.
Field Experiments (1989)
Field experiments were initiated at three locations to examine the
survival and spread of the target bacterium, A. radiobacter K84, in the
environment.
126
-------
Survival of Strain K84 on Cherry Roots
Roots of mazzard cherry seedlings were inoculated with 108 cfu/ml of
strain K84 resistant to 500 ug/ml streptomycin sulfate. Seedlings were
planted at the Botany and Plant Pathology Farm near Con/all is, OR, in 5/89.
Destructive samples were taken over time to monitor rhizosphere population
size of K84 on cherry seedlings.
Survival of Strain K84 in Fallow Soil
Fallow soil plots at the Botany and Plant Pathology Farm were inoculated
in 5/89 with strain K84 resistant to streptomycin or K84 resistant to 100
ug/ml rifampicin, so that the final concentration of bacteria in the soil plot
was 1 X 105 cfu/g dry weight soil. Soil samples were extracted with
phosphate-buffered peptone and dilution plated on Kerr 2E medium amended with
the appropriate antibiotics. Near the end of the summer when populations
decreased below detectable levels by standard dilution plating techniques,
aliquots were either enriched for 24 hours in MGY broth then dilution plated
or extracted with wooden tongue depressors coated with strain-specific
polyclonal antibodies against strain K84 to capture cells of K84 from the
samples.
Survival of Strain K84 on Annual Ryegrass Roots
Plots seeded with annual ryegrass were inoculated in 5/89 with either
strain K84 smR or K84 rfR/ so that the final concentration of bacteria added
was 1 X 105 cfu/g dry weight soil. Root samples were extracted with
phosphate-buffered peptone and dilution plated on Kerr 2E medium amended with
the appropriate antibiotics.
Movement of K84 Through the Soil
In the 1988 field tests, K84 could not be detected in fallow soil any
further than 5 cm from a cherry seedling to which it had been inoculated at
planting time. This may have been due to the inability of the strain to
survive in fallow soil or the lack of sensitivity of our detection methods.
To determine if K84 could be detected in soil at greater than 5 cm from an
inoculated cherry seedling, this year's experiment was modified to include a
trap plant. Since our earlier work demonstrated good survival of K84 on grass
roots, annual ryegrass was planted in concentric rings at increasing distances
(5, 10, 20, 40, 80, and 160 cm) around an inoculated cherry seedling to act as
a biological trap for the bacterium as it moves through the soil. Mazzard
cherry seedlings inoculated with 108 cfu/ml K84 smR were planted at the center
of the rings in 5/89, and the soil between each ring was kept fallow.
Ryegrass samples were removed periodically from each ring and the roots
assayed for the presence of strain K84.
127
-------
RESULTS AND DISCUSSION
Immunoisolation
When immunoisolation sticks were incubated for one hour in aqueous
suspensions of K84 (102, 103, and 104 CFU/ml), there was a 10 to 100 fold
decrease in the number of cells remaining in the suspension. Mixed
bacterial populations in soil suspensions posed a problem, because nontarget
bacterial cells were physically adsorbed to the sticks, and the sticks had to
be streaked to antibiotic-amended media to reisolate the target strain.
Methods to improve the procedure are being examined.
Immunoblotting for Identification of K84 Colonies
Based on fluorescence microscopy, whole bacterial cells from agar
colonies adhered best when blotted to loosely woven Whatman #1 filter paper
and Whatman 3MM chromatography paper. Whole cells and colonies washed off the
other papers and nitrocellulose sheets. The type of medium that the bacteria
were grown on affected the adherence of the cells to the papers.
The study of marker enzymes and stains to determine their suitability for
immunoblotting whole K84 cells showed that cells immunolabeled with alkaline
phosphatase reacted quickly (within 20 min) with the reagents BCIP/NBT--
colonies were easily seen as dark blue-black spots. However, non-labeled
colonies also reacted slowly with these reagents, and after 6 hours they were
indistinguishable from the labeled colonies. Alkaline phosphatase is poorly
suited for this technique because of the high background activity from
endogenous alkaline phosphatase present in these bacteria.
The use of peroxidase as a marker enzyme avoids some of the non-specific
background-staining problems. However, the CNT stain frequently cited in the
literature for use with peroxidase reacted poorly with the immunoblots. After
1 hour incubation, the blots were only faintly visible as gray-blue spots, and
small colonies (less than 2 mm dia.) were difficult to see. In contrast, the
stain AEC reacted well with the immunoblots in the presence of the marker
enzyme, peroxidase. Background staining was minimal with the red-colored
compound, and small colonies (1 mm dia) were seen easily on the blots.
Although very good progress has been made towards developing a more accurate
staining procedure, tests are still needed on the specificity of the method
and optimization of antibody concentrations and incubation times.
Harvest of Field Plots (1988)
The background incidence (water-treated controls) of crown gall at all
locations was decreased by the wild-type biological control agent A.
radiobacter K84. Average gall size was also reduced by K84 at all locations
with the exception of Moses Lake, WA. Background incidence of crown gall and
average tumor size was effectively decreased also at 3 locations by the
streptomycin-resistant mutant of strain K84; however, gall incidence at
Ephrata, WA, was higher on trees treated with K84sm compared to the applica-
tion of wild type K84. At the Oregon sites approximately 104 cfu/g root
tissue of K84sm were recovered from the rhizosphere, whereas at Ephrata, WA,
128
-------
only 200 cfu/g root tissue were recovered. The decreased ability of K84sm to
control crown gall at the Ephrata, WA, site may be related to the observation
that high populations of the strain were not sustained on root surfaces at
that site, or the presence of naturally occurring agrocin-resistant pathogens
in the soil.
In general, differences in survival and pathogenicity of A. tumefaciens
strains seemed to be related more to location and biovar grouping rather than
an effect of antibiotic resistance markers. Biovar I strains performed poorly
at the sites located in the Willamette Valley in Oregon, i.e., disease
incidence and average size of tumors from strains C58, B6 and their anti-
biotic-resistant derivatives were lower compared to that found in Washington.
The two biovar II agrobacteria strains used in the field studies, B49C (A.
tumefaciens) and K84 (A. radiobacter) and their antibiotic-resistant deriva-
tives performed well at all sites. The rifampcin-resistant strain of B49C
performed nearly as well as the wild-type strain of the pathogen. The
streptomycin-resistant strain of K84 performed comparably to the wild-type
strain of the biological control agent. Rifampicin (100 ug/ml) and strepto-
mycin (500 ug/ml) were the most effective of the antibiotic-resistance markers
used as few strains of agrobacteria naturally-resistant to these chemicals
were recovered from water-treated control seedlings. In contrast, naladixic
acid (300 ug/ml) is not recommended for future field recovery studies as
numerous strains of agrobacteria resistant to this antibiotic were recovered
from water-treated cherry seedlings.
Field Experiments (1989)
Survival of Strain K84 on Cherry Roots
During the first week of growth in the field, the population size of the
target bacterium, K84, decreased on roots of cherry seedlings by one log unit,
but stabilized during the next two weeks at approximately 1 X 106 cfu/g root.
The population size of K84 on roots remained stable the following four months
at approximately 105 cfu/g root. The target bacterium appears to colonize a
niche in the rhizosphere which provides for fairly long-term survival.
Survival of Strain K84 in Fallow Soil Plots
The population sizes of antibiotic-resistant strains of A. radiobacter
K84 (marked with spontaneous resistance to streptomycin and rifampicin) were
similar over the growing season in fallow soil. In contrast to roots, the
strains maintained high populations in fallow soil for only 1 week after
application and then decreased about 5 log units over the following 15 weeks.
By 16 weeks after inoculation of soil plots, K84 smR was detected in only 2 of
9 samples by dilution plating. However, immunoisolation and enrichment
plating of samples increased detection to 7 out of 9 samples. K84 rf* was
also found sporadically (3 of 9 samples) in dilution-plated soil samples
sixteen weeks after inoculation of plots. Treatment of samples by enrichment
culture increased detection of these strains slightly to 4 of 9 samples.
Treatment of samples with immunoisolation sticks increased detection to 6 of 9
samples. Thus, the immunoisolation procedure increased the sensitivity of
detection of the target bacterium in soil over traditional methods.
129
-------
Survival of Strain K84 on Annual Rvegrass Roots
Strains K84 smR and K84 rf* maintained a steady population size between
104 to 105 cfu/g on roots of annual ryegrass for four months after inocula-
tion. The rhizosphere population sizes on annual ryegrass in the field plots
were similar to that observed in studies conducted in the greenhouse. The
population size of the rifampicin- and streptomycin-marked strains on annual
ryegrass roots were similar also, which indicates that either antibiotic-
resistance marker can be used as a suitable standard of comparison to newer
detection methods for ecological studies of strain K84 under field conditions.
Movement of K84 Through the Soil
K84 was detected on grass seedlings 5 cm away from the point source of
inoculum one and two weeks after planting K84 sm" inoculated cherry seedlinas
at the center of rings of grass. Five weeks after planting, strain K84 was
found 10 cm away from the seedling and in one case, at a distance of 20 cm
from the cherry plant. Eight and twelve weeks after planting, strain K84 was
detected in ryegrass rings 40 cm distal to the inoculated seedlings. In
contrast, last field season, five centimeters was the maximum distance we
detected K84 in soil around inoculated seedlings. Importantly, the use of
annual ryegrass as a biological trap plant increased our ability to trace this
bacterium through the environment.
130
-------
SESSION VI
ENVIRONMENTAL EFFECTS
131
-------
-------
MOLECULAR STUDIES OF MICROBIAL ECOSYSTEM PERTURBATIONS
David A. Stahl1'2 and Richard Devereux1'3
Department of Veterinary Pathobiology1
Department of Microbiology2
University of Illinois
Urbana, IL 61801
U.S. EPA3
Environmental Research Laboratory
Gulf Breeze, FL 32561
INTRODUCTION
One critical area relating to the release of genetically engineered
microorganisms is the response of the resident community. Can an exogenous
species perturb indigenous communities of bacteria, altering the ecological
processes they execute? Assessment of process effects requires some
fundamental advances in microbial ecology. These include, the ability to
describe complex microbial communities in detail, an understanding of the
relationship between community composition and the ecological processes
mediated by the community, and sensitive methods to monitor and characterize
subtle shifts in community composition.
The tradition of community studies in microbial ecology is one of
isolating and characterizing pure cultures of bacteria obtained from the
environment. However, it is recognized that only about 10% of direct
bacterial counts in an environmental sample can be brought into culture. Once
in culture, the isolates are often difficult to identify unambiguously. The
ribosomal RNAs, particularly 16S rRNA, provide an avenue which can circumvent
these limitations. 16S rRNA can be extracted directly from natural microbial
communities and the sequence information contained therein related to the 16S
rRNA sequences of known, well characterized bacterial strains. We have been
developing a 16S rRNA-based assessment that utilizes hybridization probes and
comparative sequencing to describe specific phylogenetic entities within
natural microbial communities.
The research has focused on the characterization of two complex
microbial communities whose members are difficult to isolate and characterize;
the bacterial communities of the bovine rumen and the sulfate-reducing
bacteria in anaerobic marine sediments.
METHODS
Phvloqenetic Probes
Sequences of 16S rRNAs from rumen bacteria and sulfate-reducing bacteria
were determined (ca. 100). The sequences were aligned with previously
determined 16S rRNA sequences and searched for nucleotide tracts to serve as
target sites for synthetic oligonucleotide hybridization probes. The probes
133
-------
were labeled at their 5' ends using gamma 32P-labeled ATP and polynucleotide
kinase. More recently, fluorescent dye-conjugated probes have been fabricated
for identification of single cells by microscopy using a technique of whole
cell hybridization.
Extraction of Nucleic Acids from Natural Populations
An improved rRNA extraction protocol has been developed. Nucleic acids
are extracted directly from anoxic marine sediment or rumen fluid biomass by
mechanical disruption (bead beating) in the presence of sodium acetate/EDTA
buffer at pH 5.2, polyvinylpolypyrrolidone, sodium dodecyl sulfate, and
beta-mercaptoethanol. The resulting lysates are then extracted with phenol,
phenol:chloroform, and chloroform. Nucleic acids are recovered from the final
aqueous phase by overnight precipitation with ethanol. This procedure yields
5-8 micrograms nucleic acid per gram sediment and electrophoresis of the
nucleic acids on polyacrylamide gels demonstrates the presence of high
molecular weight rRNA.
Community Composition Determinations
16S rRNA Hybridization Probes
Composition of microbial communities was determined by hybridizations
with the phylogenetically based 16S rRNA probes. Nucleic acids extracted from
the natural community and from reference organisms were applied to nylon
membranes and hybridized with specific or universal rRNA probes. Hybridiza-
tion signals were quantitated by densitometry. Intensities of hybridization
signals obtained from the natural population nucleic acids,were compared to
intensities of hybridization obtained with reference nucleic acids. The
contribution of 16S rRNA from a phylogenetic assemblage to the total 16S rRNA
of the community was expressed as relative abundance normalized to hybridiza-
tion of the universal probe.
Fluorescent Probes
We have optimized whole cell hybridization with fluorescent probes.
Specific fiber-digesting and sulfate-reducing bacteria have been micro-
scopically observed in natural samples.
Group-Specific Cloning and Sequencing
The polymerase chain reaction (PCR) has been used to selectively amplify
sulfate-reducing bacteria-specific 16S rRNA sequences for cloning and
sequencing.
Sulfate Reduction Rate Measurements
Sediment cores were taken from the lagoon adjacent to the EPA facility
in Gulf Breeze and fractionated by depth. The functional capacity of the
sulfate-reducing bacterial community at each depth was measured using fulfate
reduction assays that followed the formation of hydrogen sulfide from S
labeled sodium sulfate.
134
-------
RESULTS AND DISCUSSION
Phvloqenetic Probes
Kingdom, "group", genus, species, and subspecies probes have been
developed for sediment and rumen studies. Probes for five phylogenetic
assemblages of sulfate-reducing bacteria were developed: BTR, Desulfobacter
spp.; BTM, Desulfobacterium spp.; BUL, Desulfobulbus spp.; 825, Desulfovibrio
spp. related to D. desulfuricans; and 818, for Desulfococcus multivorans
related spp.
For ruminal studies, genus, species and subspecies-specific probes for
principal fiber digesting species of the rumen (e.g., fibrobacter and
ruminococci) and kingdom-specific probes (archaebacterial, eubacterial and
eukaryotic) have been developed.
Community Structure Measurements
The sulfate-reducer specific probes provide specificity and sensitivity
for detection of the species present at 10 cells/g sediment as determined by
cell seeding experiments. In slurried sediment, without added cells, all
groups of sulfate-reducing bacteria were detected except for Desulfovibrio
species.
Community Structure and Function Relationships
Sulfate reduction rate measurements and community composition were
determined for three sediment cores. In each core there was a distinct peak
of activity near or at the sediment water interface. Coincident with the peak
activity was an increase in the abundance of 16S rRNA from species which
hybridize to the 818 probe. The profile for Desulfobacter and Desulfobac-
terium species 16S rRNA was evenly distributed throughout the core.
Desulfobulbus spp. 16S rRNA was detected only in the top three fractions of
each core, while no Desulfovibrio spp. 16S rRNA was detected in any fraction.
The results demonstrated a particular community at each depth and a relation-
ship between the sulfate-reduction activity and community composition. Since
each group of sulfate-reducing bacteria detected by the probes are physiolog-
ically distinct, it was possible to relate the function and community
structure. In this case, the species detected by the 818 probe are higher
carbon chain fatty acid utilizers so it was likely that at the time of
sampling the sulfate-reducing bacterial community was driven by the avail-
ability of fatty acids.
Probing and comparative sequence analysis of environmentally derived
sequences (relative to the physiology of pure culture isolates in the 16S rRNA
sequence collection) suggests that physiological activities, and therefore
community function, can be predicted by 16S rRNA molecular criteria.
135
-------
ANTICIPATED FUTURE WORK
Relationships between the community structure and function of the
sulfate-reducing community will be further explored. It will be necessary to
determine baseline variation in natural populations before the effect of an
introduced species on the sediment communities can be evaluated. Results of a
study concerning the long-term stability of ruminal microbial populations
(inter- and intra-animal) are now being analyzed. Sediment studies will focus
on spatial and temporal variation. The next phase of the studies will measure
perturbations of the community caused by environmental change resulting from
an introduced species. In all these studies the reliability of the
measurements will be enhanced by the development of additional probes.
PUBLICATIONS
Stahl, D.A. 1988. Phylogenetically-based studies of microbial ecosystem
oerturbations. In American Chemical Society Symposium Volume: Biotechnology
-or Crop Protection. P. Hedin, J.J. Menn, and R.M. Hollingworth (eds.),
American Chemical Society, Washington, D.C. 373-390.
Stahl, D.A., B. Flesher, H.R. Mansfield, and L. Montgomery. 1988. Use of
phylogenetically based hybridization probes for studies of ruminal microbial
ecology. Appl. Environ. Microbiol. 54:1079-1084.
Montgomery, L., B. Flesher, H.R. Mansfield, and D.A. Stahl. 1988. Transfer
of Bacteroides succinogenes to Fibrobacter gen, nov. as Fibrobacter
succinogenes comb, nov. and Fibrobacter intestinalis sj). nov. Int. J. Syst.
Bacteriol. 170:430-435.
Devereux, R., M. Delaney, F. Widdel, and D.A. Stahl. 1989. Natural relation-
ships among sulfate-reducing bacteria. J. Bacteriol. 171:6689-6695.
Stahl, D.A. , R. Devereux, R.I. Amann, B. Flesher, C. Lin, and J. Stromley.
1989. Ribosomal RNA based studies of natural microbial diversity and ecology.
Proc. 5th Int. Symp. Microbial Ecol. 669.673.
Amann, R., L. Krumholz, and D.A. Stahl. 1990. Fluorescent DNA probing of
whole cells for determinative, phylogenetic and environmental studies in
microbiology. J. Bacteriol. 172:762-770.
Stahl, D.A., and R. Amann. Development and application of nucleic acid probes
in bacterial systematics. In Sequencing and Hybridization Techniques in
Bacterial Systematics. E. Stackebrandt and M. Goodfellow (eds.). John Wiley
and Sons, Chichester, England (in editorial review).
Devereux, R., and D.A. Stahl. 1989. Diversity of sulfate-reducing bacteria.
Ann. Meeting Am. Soc. Microbiol., New Orleans, La.
Stahl, D.A. 1989. Comparative rRNA sequencing and extant microbial
diversity. Ann. Meeting Am. Soc. Microbiol., New Orleans, La.
136
-------
HETEROTROPHIC MICROBIAL DYNAMICS IN AQUATIC MICROCOSMS:
DESIGN CONSIDERATIONS AND FIELD VALIDATION
R. Coffin,1 Niels Kroer,2 and Niels Jorgensen3
TRI, U.S. EPA1
Environmental Research Laboratory
Gulf Breeze, FL 32561
Ministry of the Environment2
Division of Marine Ecology and Microbiology
Copenhagen, Denmark
Royal Veterinary and Agricultural University3
Copenhagen, Denmark
INTRODUCTION
Release of genetically engineered microorganisms (GEMs) to the environ-
ment raises questions regarding ecological and public health impacts. A
fundamental consideration is the ability of the organism to survive and
compete in a natural setting and possibly transfer the engineered trait to
other organisms. Prediction of the fate of the organism and its new gene in
natural systems is a major component of a risk assessment.
Environmental studies designed to assess the risk associated with
introduction of GEMs to the environment are limited by the complexity of the
environment and by the potential for environmental damage. Therefore, a small
scale system that can be manipulated in the laboratory is needed. Aquatic,
site-specific microcosms have been used to study the fate of pollutants,
however only simple chemical parameters are used to assess the effect on
community metabolism. In addition, aquatic microcosms have been used to study
phytoplankton. However, none of these microcosm studies have examined the
ecology of bacteria. Obviously, microbial ecology in microcosms needs to be
compared with the field before these systems may be used for risk assessment.
METHODS
To examine the important design components of aquatic microcosms, a
complex system designed at the National Food Agency of Denmark (LST) and a
simple system used at the Environmental Research Laboratory, Gulf Breeze
(ERL/GB) were compared using water samples from a eutrophic Danish lake.
Briefly, the LST system consists of three units, a 216 liter phytoplankton
tank, a 27 liter herbivore tank, and a 1 liter tank containing sediment cores.
The phytoplankton tank is illuminated with up to 12 Phillips TL33 fluorescent
lights, while the herbivore and benthic components are maintained in the dark.
All units are temperature controlled. Water is pumped between the units with
a peristaltic pump at a rate that yields a 10 day residence time in the algal
tank.
137
-------
The EPA microcosm consists of a 27 liter pyrex cylinder in which a
sediment core is placed. The cylinder is illuminated and temperature control
is similar to the 1ST microcosm. This system is mixed gently so that the
sediment surface is not visibly disturbed. Water in both microcosms was
replaced with untreated lake water at a rate of 10% per day. Three replicate
LST and 4 replicate ERL/GB microcosms were studied simultaneously.
Parameters measured in the microcosms and the lake were chosen to
provide a broad overview of microbial trophic interactions. These parameters
include phytoplankton production, chlorophyll a, bacterial abundance,
bacterial production, grazing on bacteria, microflagellate abundance, micro-
flagellate production, amino acid concentration, amino acid uptake, and
leucine incorporation. Samples were taken three times a week for 24 days.
These comparisons were repeated twice.
RESULTS AND DISCUSSION
In comparison to the field, phytoplankton production and biomass in both
microcosms were similar for 10-14 days and then decreased. Bacterial
parameters, including biomass, production, amino acid turnover and leucine
incorporation were consistent in both microcosms and the field. Microflag-
ellate biomass replicated the biomass in the lake for 10-14 days and then
decreased. It is surprising that bacteria in the microcosms continued to
track the field throughout the experiment in the absence of a consistent
supply of organic matter from phytoplankton. Results from previous studies
indicate that phytoplankton are the dominant source of bacterial carbon in the
microcosm and show that microcosms may support extensive cell wall growth. We
believe that bacteria maintained a biomass and activity similar to the field
throughout the experiment because total phytoplankton production of the
microcosm was similar to the field when attached and planktonic primary
production were considered.
Varability within a microcosm system was greater in the ERL/GB system
than the LST system. The ratio of water volume to sediment area was 1900,
5400, and 71 for the field, LST, and ERL/GB microcosms, respectively. We
believe that the high variation in the ERL/GB microcosm results from over-
representation of the sediment. During both experiments one microcosm
especially deviated from the other replicates. After the 24-day incubation
period, the sediment of these systems was examined and clams were found in the
divergent microcosm.
ANTICIPATED FUTURE WORK
Recent studies have provided extensive data relating heterotrophic
microbial ecology in field samples with various microcosm systems. Future
work will trace the fate of bacteria added to microcosms in relation to
indigenous bacterial populations in microcosms. This work will focus on the
role of substrate supply and predation in controlling populations.
138
-------
CELLULYTIC BACTERIA AS SURROGATES FOR A GENETICALLY
ENGINEERED MICROORGANISM: MICROCOSM STUDIES OF
PERSISTENCE AND EFFECTS IN STREAMBED SEDIMENTS
Thomas L. Bott and Louis A. Kaplan
Stroud Water Research Center
The Academy of Natural Sciences
Avondale, PA 19311
INTRODUCTION
These studies address two concerns relative to genetically engineered
microorganisms (GEMs) in the environment. The first is whether a bacterial
population with a salient property of a GEM will persist in natural communi-
ties, possibly displacing other populations, and affect community and
ecosystem properties. By assessing several response variables, the studies
are designed to pinpoint those of greatest utility in detecting effects of the
organism. Second, experiments will be conducted both in microcosm streams and
a natural stream in order to assess the equivalence of microcosm and field
data, and thereby assess the utility of stream microcosms for assessing the
fate and effects of GEMS in natural systems.
We are working with surrogates for a GEM with enhanced cellulose degrad-
ing capability. Cellulose "superdegrader" GEMs are of interest to the food
and paper industries and in the conversion of biomass to fuels. While a
surrogate will not mimic a GEM in all physiological and ecological respects,
the use of the surrogate allows field studies with the organism and microcosm
validation as preludes to studies with a GEM whose safety has been adequately
insured. In nature, enhanced cellulose degradation could affect rates and
pathways of carbon cycling. Our research focuses on stream ecosystems, since
they often receive waste discharges and non-point source inputs. Since there
is close interaction between streams and their watersheds, GEMs could be
introduced into streams during storm runoff, during applications on cropland,
or in groundwater. White Clay Creek, the study stream, drains a protected
rural watershed in southeastern (Chester County) Pennsylvania.
METHODS
Over 20 isolates of aerobic cellulolytic organisms were obtained from
culture collections worldwide and assayed for cellulolytic activity on
cellulose, dead algae, leaf litter and for growth rates. Isolates were ranked
within experiments from most to least active and an overall mean rank was
calculated for each isolate. The isolates considered best candidates for
experiments were: Cellulomonas flaviqena (NRC 2403), C. fimi (NRRL B402), C.
sp. (NRC 2406), C. uda (NRRL B404), and Cellvibrio gilvus (ATCC 13127).
Antisera produced against them were conjugated with a fluorescent tag (FITC or
DTAF) to enable determination of the densities and dynamics of the target
populations. We have scaled our research effort concerning the persistence of
the isolate and effects on benthic community and ecosystem properties into
three phases: studies in 2.95 L respirometers, studies in 35 L microcosm
streams, and comparative studies in microcosms and field sites. We have
139
-------
selected approximately 15 response variables to assess potential effects of
the isolates on benthic communities and ecosystems.
Respirometer Experiments
Trays of sediment from a microcosm stream were transferred to replicate
respirometers, one per chamber. Cells of an isolate were added to stream
water in some respirometers; others were maintained as uninoculated controls.
Sediments were inoculated by recirculating the water for 48-72 h, after which
new stream water was added to turn over the respirometer volume every 40 min.
Samples of surface sediments (top 2-3 mm) and periphyton were collected after
inoculation and at intervals thereafter for up to a month for enumeration of
cells of the isolate and total bacteria.
Bacterial Enumeration
Formalin fixed samples were sonicated for 45 sec at 30 W (some with 0.1M
dibasic ammonium phosphate as a surfactant). Aliquots of sonicated materials
were filtered and stained with FA, DAPI, or Rhodamine 123 for epifluorescence
microscopic counts (EMC) of the isolate, total, or active bacteria, respec-
tively. Cells were separated from sediments by centrifugation once (412 xg
for 5 min at 4°C) in 30% glycerol, and twice more in 60% glycerol. Cells
associated with the pellet remaining after centrifugation were also enumerated
for some samples.
Microcosm Stream Experiments
Six stream microcosms (2.23 m long x 0.203 m wide x Q..127 m deep) were
constructed and 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 with bottoms
of 400 urn mesh nylon screening) containing surface sediments from White Clay
Creek were placed. The stream water (35 L) is recirculated; 900 ml/min of new
stream water is added to replace the entire volume approximately every 40 min.
Water depth approximates 1.5 cm. Water returned to White Clay Creek is
filtered through cartridge filters and irradiated with ultraviolet light, a
treatment found to effectively kill the study bacteria. The streams, housed
in a greenhouse, are immersed in water jackets continuously replenished with
White Clay Creek stream water to maintain near-ambient stream water tempera-
tures. The systems are separated by clear plastic sheeting to prevent spread
of organisms between microcosms.
Sediments and filamentous algae (dominated by Cladophora sp.) were
sampled over a 3-4 day period prior to addition of the isolate to establish
population and community structural and functional parameters in each micro-
cosm. Structural measures included EMCs of the isolate, total, and active
bacteria; algal biomass (chlorophyll a); and total viable biomass (ATP).
Functional measures included bacterial productivity (from [3H]thymidine
incorporation) and algal primary productivity and community respiration (from
dissolved oxygen changes). In current studies cellulose decomposition is
being determined using the Cellulose Azure assay and detritus (leaf pack or
dead Cladophora sp.) decomposition rates from weight loss and changes in
cellulose content. The isolate was added to the water in duplicate micro-
140
-------
cosms, maintaining duplicate microcosms as uninoculated controls. After
recirculating the water for approximately 72 h, the system was returned to
flow-through mode. After the water in the systems was turned over 6 times an
intensive study of densities of the surrogate in surface sediment and algae
was carried out for 3-5 days. Samples were also taken for determination of
total and active bacterial numbers, chlorophyll a, and ATP. Then, approxi-
mately 10-14 days after inoculation, the entire suite of response parameters
was measured. Physical, water chemical, and sediment conditions were moni-
tored.
RESULTS AND DISCUSSION
Respirometer Experiments
In the first experiment C. uda (NRRL B404) was added to stream water to
provide 3.39 x 107 cells/ml, 75% of which were viable according to plate
counts. Water temperatures ranged from 1.5 to 14.0°C and averaged 6.3 ± 3.4°C
(x ± s.d.). The chlorophyll a content of the periphyton averaged 0.058 +
0.010 mg/cm2 initially, and 0.049 + 0.026 mg/cm2 15 days later. Sampling
pressure, coupled with low temperature, changes in invasion rate and flow rate
shifted the dominant algae from Cladophora sp. to diatom species.
Approximately 2.0% of the organisms inoculated into the respirometer
became associated with the benthos. C. uda declined in number but persisted
for at least 4 weeks. Densities jn sediment samples (cell/cm3) decreased from
approximately 4.28 + 0.91 x 107 (x ± s.d., n - 3 respirometers) to 4.49 + 2.12
x 105 (n * 2). Total bacterial densities averaged 1.94 ± 0.47 x 109 cells/cm3
(n « 17 estimates). Thus, C. uda comprised approximately 2.2% of the total
community initially and 0.2% of the community 4 weeks later. C. uda, or
cross-reacting organisms in the natural flora were present in some
experimental respirometers before inoculation and in the uninoculated control
microcosms, but at densities that were 3 to 27 times lower than in the
inoculated systems.
In a second experiment, conducted at 11-17°C, the persistence of C. uda
was compared with that of C. sp. (NRC 2406). Sediments in duplicate respiro-
meters were seeded with an isolate and after 2 days the stream water was
replenished continuously to provide a turnover every 40 min. The two isolates
persisted for nearly a month in the sediment and periphyton microbial communi-
ties. Nevertheless, C. uda sample, means decreased over an order of magnitude
from 7.89 ± 3.62 x 10r cells/cm3 (x ± s.d., n = 2 respirometers) to 2.74 ±
1.15 x 105 cells/cm3, over 27 days while those for C. sp. decreased from 1.61
± 1.13 x 107 cells/cm3 to 1.89 ± 0.57 x 106 cells/cm3 over the same period.
However, C. sp. sample densities showed an intermittent, but distinct,
increase in number from 3.86 ± 1.94 x 106 cells/cm3 (n = 2) to 1.31 ± 1.04 x
10 cell/cm3 (n = 2),which may have been due to either growth or a more patchy
distribution than C. uda. Total bacterial densities averaged 1.87 ± 1.34 x
109 cells/cm3; thus isolates maximally composed 0.41% and 0.86% of the
communities.
141
-------
C. sp. (CS1-1} cross-reacted with antisera against C. uda allowing its
population dynamics to be followed with that antisera. Sample means from
duplicate respirometers suggested intermittent growth from a mean post-
inoculation density of 2.89 ± 1.64 x 106 cells/cm3 to 1.33 ± 0.32 x 107
cells/cm3 in 5 days at 12-1S°C, followed by a decline to 1.91 ± 0.91 x 10b
cells/cm3 over the next 20 days.
Microcosm Experiment
C. sp. (NRC 2406) was inoculated into duplicate microcosms to achieve
densities of 1.1 and 1.5 x 108 cells/ml, values approximately 4 times higher
than in the first respirometer experiment. However, densities in the sediment
were approximately 3-6 times lower than in that experiment. Maximum densities
in the two treated microcosms were 6.54 and 1.49 x 106 cells/cm3. FA positive
organisms were detected in the control microcosms, but densities were approxi-
mately fifteen times lower, averaging only 2.53 and 3.14 x 105 cells/cm . FA
positive organisms were also detected in one experimental stream in one of
three samples taken before innoculation, yielding an average density of only
4.23 x 10 cells/cm3. Over the next two weeks densities in the sediment
samples from the inoculated microcosms declined to between 5.09 and 6.77 x 10s
eel Is/cm3, which approached the background levels in the controls. Water
temperatures approximated 20°C and differences in water flow characteristics
and environmental patchiness between microcosms and the respirometers
undoubtedly affected these results.
Densities of C. sp. in Cladophora growths that covered portions of
sediment in the inoculated microcosms four days after inoculation were 1.40
and 4.56 x 106 cells/cm3. Approximately 10 days later these values declined
to 4.94 and 3.49 105 cells/cm. These values were close to the average
density of 5.6 + 3.1 x 105 FA positive cells in samples of algae in the
control microcosms. FA positive cells were detected before inoculation in the
algae growing in all systems.
The inoculated microcosms were examined for the presence of FA positive
organisms approximately 3 months later. FA positive cells were detected in
algae in both of them, and also in the two control microcosms. FA positive
organisms were not detected in the preliminary checks of sediment samples from
the experimental microcosms, but they were found unequivocally in one of the
controls. We are working to clarify whether these are cells of the isolate or
cross-reacting organisms. However, we have detected FA positive organisms in
White Clay Creek.
Primary productivity of the sediment communities ranged from 0.07 to
0.31 mg DO T1 h"1 before inoculation and from 0.15 to 3.39 mg DO I"1 h"1 after
inoculation (means of duplicate treatment and control streams on each of three
days). Algal productivity averaged 3.90 to 5.19 mg DO I"1 h'1 in the treatment
and control microcosms, respectively, before inoculation. After inoculation
the values were 5.01 and 3.77 mg DO I'1 h"1 in the treatment and control
microcosms, respectively. Sediment community respiration ranged from 0.21 to
0.39 mg DO I"1 h'1 prior to inoculation and from 0.15 to 0.43 mg DO T1 h"1
after inoculation. Respiration associated with algae prior to inoculation was
0.55 and 0.72 mg DO I"1 h"1 in the treatment and control microcosms,
142
-------
respectively, and 0.41 and 0.28 mg DO I"1 h'1 in those systems, respectively,
after inoculation. Tests of the productivity and respiration data using the
BACI analysis showed that the surrogate never affected these benthic
processes. The data were normalized for incident PAR and retested, and again
no significant differences were found.
Bacterial productivity for the inoculated microcosms after exposure to
the surrogate bracketed the values for the control microcosms, being 2691 and
9455 cells cm"2 h"1 in the inoculated systems and 5435 and 8032 cells cm"2 h"1
in the controls. The data from measures before inoculation also overlapped.
Occurrence of the Organism in White Clay Creek
We have begun to screen habitats in White Clay Creek for FA positive
organ-isms. They have been detected in low densities in sediments, in leaf
packs in higher densities, but not in moss, at locations near the laboratory
and farther upstream.
ANTICIPATED FUTURE WORK
Results to date indicate that isolates can persist for periods of
several weeks to months in sediments and benthic algal growth although
densities declined fairly rapidly from high inoculation levels. However,
moderately high densities have been discovered in a few distinct natural
habitats. Effects on community metabolism and associated changes in DOC, or
on bacterial productivity have not been pronounced. Given these findings,
more attention is being, and will be, given to the occurrence of the isolates
in specific microhabitats, to rates of cellulolysis, and alterations of rates
of detritus decomposition in future studies. The extension of these studies
to work with a GEM possessing enhanced cellulolysis is anticipated. That work
would deal with displacements of native organisms or added surrogates for the
GEM as well as effects on community and ecosystem structure and function. We
are also working with a collaborating researcher to obtain gene probes for the
cellulase gene. If we are successful, we will use the probe along with the FA
in studies of the distribution and dynamics of the isolates.
PUBLICATIONS
Bott, T.L., and L.A. Kaplan. 1989. Selection of a surrogate for a GEM and
enhanced cellulose degrading capability for studies in streams. Abstracts of
the 89th Annual Meeting of the American Society of Microbiology, New Orleans,
LA, May 14-18, 1989.
143
-------
EFFECTS OF MICROBIAL PESTICIDES ON MYCORRHIZAL SYMBIOSES
Robert G. Linderman.1 Robert P. Griffiths,2 Joyce E. Loper,1
Timothy C. Paulitz,1 and Bruce A. Caldwell2
USDA1
Agricultural Research Service
Horticultural Crops Research Laboratory
Department of Microbiology2
Oregon State University
Corvallis, OR 97331
INTRODUCTION
The widespread use of pesticides in modern agriculture and forestry has
prompted significant regulatory activity, especially in controlling new
chemicals coming on the market. The strict regulation and threat of chemical
over-use has stimulated rapid research and development of alternative pest
control measures, including biological pesticides. This, in turn, has
prompted the question of the impact such microorganisms would have on the
natural balances of other organisms in the environment. One major and
commonly recognized natural component of the soil ecosystem is the nearly
ubiquitous mycorrhizal symbiosis that occurs between plant roots and specific
soilborne fungi. This symbiosis has a significant impact on the health and
well-being of most plants, both in natural and cultivated ecosystems. Thus,
the risk of damage by both chemical and biological pesticides to the
mycorrhizal symbiosis needs to be critically assessed.
The objectives of this research are to develop methods for evaluation of
the impact of naturally-occurring and genetically-engineered biological
control agents (BCAs) on mycorrhizae. The described methods will identify and
characterize bacterial or fungal BCAs with deleterious effects on various
stages of mycorrhizal development that can be used as standards for evaluating
microbial pesticides in the future. Those stages are a) spore germination, b)
early stages of root colonization and vegetative growth, c) overall root
colonization and development of external hyphae and rhizomorphs, and d) host
plant growth enhancement under nutrient stress conditions. In addition, the
research will attempt to elucidate the mechanisms of any observed adverse
reactions.
METHODS
Ectomvcorrhizae and Ericoid Mycorrhizae
The effect of BCAs on colonization of 2-month-old Douglas-fir seedling
roots by ectomycorrhizal fungi was determined in pairing studies wherein
inoculum of ectomycorrhizal fungi and fungal or bacterial BCAs was added to
soil. The mycorrhizal fungi used were Rhizopogon vinicolor. Hebeloma
crustuliniforme, Cenococcum qeophilum, and Laccaria laccata. The fungal BCAs
tested were Gliocladium virens, Trichoderma harzianum, and Talaromvces flavus,
and the bacterial BCAs tested were Pseudomonas fluorescens. Enterobacter
cloacae. Enterobacter aeroqenes, Aqrobacterium radiobacter, Alcaligenes sp.,
144
-------
Streotomvces sp., and Bacillus subtilis. Plants were incubated in insulated
root temperature boxes at 20°C. The seedlings were fertilized weekly with
modified Long Ashton Nutrient Solution low in nitrogen). Harvested plants
were evaluated for mycorrhizal colonization, root and shoot dry weights, and
the relative amount of rhizomorphs present. In most cases, rhizosphere
population size of the biocontrol organisms was monitored at each harvest by
dilution plating on selective media.
In vitro experiments pairing ericoid or ectomycorrhizal fungi against all
fungal and bacterial BCAs were conducted to test for antagonism. The pairings
were done by placing the two test organisms 3 cm apart on either potato
dextrose agar (PDA) or modified Melin-Norkrans (MMN) agar plates, incubating
at 20°C, and examining daily for zones of inhibition or pigment production
of or by either organism. These mycorrhizal fungi were also screened for
siderophore production.
To test the hypothesis that mycorrhizal fungi may be less sensitive to
inhibitors from BCAs, seven antifungal antibiotics (Nystatin, Filipin Complex,
Griseofulvin, Puromycin, Antimycin A, Valinomycin, and Kasugamycin) were
tested at four concentrations in challenge experiments against three phyto-
pathogenic fungi, (Fusarium oxysporum. Phytophthora cinnamomi. and Rhizoctonia
fragariae ; four ectomycorrhizal fungi (Cenococcum qeophillum, Hebeloma
crustuliniforme, Rhizopogon vinicolor, and Laccaria laccata); and two ericoid
mycorrhizal fungi (Qidiodendron griseum, and Hvmenoscyphes ericae).
Vesicular-Arbuscular Mycorrhizae (VAM)
The effects of bacterial BCAs on early VAM colonization by Glomus
etunicatum added to soil were tested using Pseudomonas putida A12, N1R, R20;
P. fluorescens 3882, 3871 and 2-79; Bacillus subtilis; Streptomvces sp.:
Alcaligenes sp.; Agrobacterium radiobacter, and Enterobacter cloacae. Three
weeks after planting, plants were harvested and mycorrhizal colonization was
evaluated. Experiments were also designed to confirm biological control
activity of the BCAs against Pythium damping-off disease of cucumber under
the soil and growth conditions used in mycorrhizal experiments.
Isolates of the fungal BCAs Gliocladium and Trichoderma were added to
Pythium-infested soil. After 8 days, cucumber seeds were sown in the soils,
and percent emergence and shoot dry weight of surviving plants was recorded.
Experiments were conducted on effects of BCAs on germination of spores of
Glomus placed on membrane filters buried in soil. The BCAs were either added
to the soil or to the membranes and incubated for various times up to 7 days.
Then the spores were stained with Trypan Blue and percent germination deter-
mined. We determined the appropriate host and phosphorus combination for
consistent plant growth promotion in response to plant inoculation with VA
mycorrhizal fungi. Two host plants, onion alia green pepper, and three
mycorrhizal fungi, Glomus mosseae. G. intraradices, and G. deserticola. were
screened in a low phosphorus sand soil mix. Biological control agents were
added by banding a layer of either fungal inoculum or a layer of vermiculite
saturated with a bacterial suspension above the G. mosseae infested soil.
145
-------
This inoculation method allowed the BCAs to establish in the plant rhizosphere
before the roots contacted the mycorrhizal fungus spores.
RESULTS AND DISCUSSION
Ectomycorrhizae and Ericoid Mycorrhizae
In our initial studies, we ran numerous short studies to determine if
biocontrol fungi and bacteria would persist and control plant disease in the
same plant systems in which mycorrhizal fungi grow and colonize roots. Our
results demonstrated that the BCAs survived in the rhizosphere and persisted
over the time necessary for mycorrhizae formation (2-4 weeks).
Some fungal BCAs adversely affected colonization by Rhizooogon and
Cenococcum when compared to the controls. Shoot and root dry weights of
Rhizopogon plants were reduced by Gliocladium virens. Some of the bacterial
BCAs actually appeared to enhance ericoid mycorrhizae (EM) colonization of
Douglas-fir roots by Rhizopogon. but not by Hebeloma or Laccaria. Seedling
root and shoot dry weights were significantly different from the controls in
some mycorrhizal fungus-BCA combinations.
In in vitro antagonism experiments, an actinomycete (H68-4) isolated from
the ectomycorrhizal rhizomorphs of Hvsterangium setchellii inhibited the
growth of the pathogenic fungus Phellinus weirii. While all ectomycorrhizal
fungi eventually overgrew the actinomycete H68-4, radial extension of
Rhizopogon vinicolor and Laccaria laccata on agar medium was reduced and
distorted. Both ericoid mycorrhizal fungi, Hvmenoscvphus ericae and
Oidiodendron griseum. inhibited growth of H68-4. The response of fungal
pathogens (Fusarium oxvsporum £ol, Phvtophthora cinnamomi Pc-6, and
Rhizoctonia solani Rs J-l) to H68-4 were in part dependent on the media used.
Of the in vitro pairings of ectomycorrhizal or ericoid mycorrhizal fungi
conducted to date, no growth inhibition or pigment production at the locus of
contact has been observed. This result is in sharp contrast to the striking
inhibition of pathogens tested against the same biocontrol agents.
Assays of soils from around Douglas-fir seedlings inoculated with the EM
fungi Cenococcum geophilum, Rhizopogon vinicolor. or Hebeloma crustuliniforme
showed no differences in phosphatase levels between controls and treatments
with the fungal BCAs Tf-1, Gv-P or Wt-6. A second study including the EM
fungus Laccaria laccata and the bacterial BCAs Pseudomonas fluorescens (Pf-5
3871), Enterobacter cloacae (Encl 1157), E. aerogenes (Enae B8) and Bacillus
subtilis (Q4000) produced similar results. These results indicate that the
application of the BCAs did not adversely effect soil phosphatase levels.
We conducted experiments to determine if mycorrhizal fungi are more
resistant to antifungal antibiotics that other fungi. We reasoned that if
mycorrhizal fungi are sensitive to these antibiotics, they should be affected
by antibiotic-producing organisms. Preliminary results suggest that the
sensitivity of ectomycorrhizal fungi to the antifungal antibiotics is similar
to that of the fungal pathogens.
146
-------
VA Mvcorrhizae
None of the BCAs had a significant effect on spore germination other than
a slight delay in germination caused by one isolate of Pseudomonas.
We have developed reproducible methods to measure the effect of BCAs on
VAM symbiosis. This method assumes a quantitative relationship between the
inoculum density of mycorrhizal fungus in soil and mycorrhizal colonization of
the host root. This relationship can be described by regression analyses. If
a BCA inhibits VAM symbiosis, the quantitative relationship between inoculum
density and VAM colonization can be defined by statistical analyses. None of
the bacterial BCAs tested gave significant decreases in early root coloniza-
tion of cucumber by G. etunicatum. Mycorrhizal root colonization data,
plotted and analyzed by linear regression, revealed no effect of bacterial
BCAs on G, etunicatum.
We have developed procedures for evaluating the possible deleterious
effects of introduced biocontrol agents on the establishment and beneficial
performance of mycorrhizal fungi in symbiotic association with plant roots.
These procedures address effects on spore germination, root colonization, and
plant growth enhancement by VA mycorrhizal fungi. For ericoid mycorrhizal
fungi, we have as yet only evaluated interactions (possible inhibition) on
agar media. For ectomycorrhizal fungi, we have evaluated interactions on agar
media, in soil, and on plant roots by observing effects of BCAs on root
respiration, mycorrhizal colonization, and plant growth enhancement. We have
clearly shown that both the biocontrol agents and mycorrhizal fungi are able
to function in the test systems. We have seen only minor inhibition of
mycorrhizal symbioses from introduced biocontrol agents. Ectomycorrhizal
fungus colonization of Douglas-fir roots was inhibited in some cases, but not
in others. The subtle levels of inhibition observed in our studies suggest
that our methods are sensitive, and will be effective in detecting strongly
detrimental candidate biocontrol agents.
FUTURE WORK
Work will continue on perfecting the Douglas-fir/EM system by reducing
the concentration of Hebeloma and Laccaria inocula which will allow moderate
levels of colonization thereby increasing the sensitivity of the assay system
to detect the effects of BCAs on mycorrhizal symbiosis. In addition, we are
working on cranberry/ericoid mycorrhizal system to evaluate the effects of
BCAs in that system.
In the onion/VAM system, a plant host, mycorrhizal fungus, and phosphorus
fertilization level was chosen to define a system where consistent plant
growth promotion in response to mycorrhizal inoculation was achieved.
Currently, the sensitivity of the test system is being maximized by inoculat-
ing soils with a range of G. mosseae spores, concentrations and by establish-
ing rhizosphere populations of BCAs prior to root contact with mycorrhizal
spores. The onion-G. mosseae-low phosphorus system will be used in future
studies evaluating the affects of BCAs on mycorrhizae-induced plant growth
promotion.
147
-------
Work will continue on the pure culture media filtrate experiments to
screen for the effects of BCAs on the mycorrhizae we are able to grow on
culture media.
PUBLICATIONS
Caldwell, B.A., R.P. Griffiths, R.G. Linderman, and J.E. Loper. 1989.
Production of siderophore-like iron chelators by ericoid and ectomycorrhizal
fungi. In Proc. XIV Beltsville Symposium on "The Rhizosphere and Plant
Growth," May 8-11, 1989 (poster).
Linderman, R.G. 1990. Mycorrhizal interactions in the rhizosphere. Jn
Proceedings of XIV Beltsville Symposium on "The Rhizosphere and Plant Growth,"
D. Keister and G. C. Papavisas (eds.).
Linderman, R.G., T.C. Paulitz, N.J. Hosier, R.P. Griffiths, J.E. Loper, B.A.
Caldwell, and M.D. Henkels. 1989. Evaluation of the effects of biocontrol
agents on mycorrhizal fungi. In Proc. XIV Beltsville Symposium on "The
Rhizosphere and Plant Growth." May 8-11, 1989 (poster).
Paulitz, T.C., and R.G. Linderman. 1988. Interactions between VA mycorrhizal
fungi and fluorescent pseudomonads in the rhizosphere. Phytopathology
(abstract).
Paulitz, T.C., and R.G. Linderman. 1989a. Interactions between mycorrhizae
and biocontrol agents. HortScience (abstract).
Paulitz, T.C., and R.G. Linderman. 1989b. Interactions between fluorescent
Pseudomonads and VA mycorrhizal fungi. New Phytologist (in press).
148
-------
EFFECTS OF RECOMBINANT STREPTOMYCES ON A SOIL ECOSYSTEM
Don L. Crawford, Zemin Wang, and Bruce H. Bleakley
Department of Bacteriology and Biochemistry
Institute for Molecular and Agricultural Genetic Engineering (IMAGE)
University of Idaho
Moscow, ID 83843
INTRODUCTION
In our EPA-funded research program, we are studying the fate and effects
of recombinant Streotomvces after their release into soil environments. These
recombinants express cloned genes on self-transmissible or non-transmissible
plasmid vectors. In comparisons with nonrecombinant wild-type cultures of the
same strains, we have characterized how the recombinant Streptomyces survive
and at what frequencies they transfer plasmids to other Streptomyces in soil.
The research has shown that while survival of recombinant Streptomyces in soil
is strain specific, recombinants often survive as well as their nonrecombinant
parents. They also tend to maintain recombinant plasmids and exchange them
readily with other soil Streptomvces. by mechanisms including conjugation and
conjugational mobilization. Environmental factors such as soil temperature,
moisture, and nutrient availability markedly affect the frequency of plasmid
transfer observed.
Our current research has involved elucidation of the mechanism by which
certain released recombinant Streptomyces lividans strains affect the carbon
cycle in soils. In some of our previous work, we observed-that one recom-
binant, Streotomvces lividans TK23-3651, significantly affected the short term
(30 days) mineralization rate of soil organic carbon. After its release into
soil, the rate of soil organic carbon mineralization increased, particularly
in non-sterile soils amended with lignocellulose. The significant enhancement
of CO, evolution rate, however, was transient. Additional studies showed that
the strain was unstable in soil. Here, we report on our studies of the
mechanism for the transitory enhancement of carbon turnover rates in soils
where S. lividans TK23-3651 was released.
METHODS
The wild-type and recombinant Streptomvces strains and plasmids used in
this research have been developed in this laboratory. Lignin peroxidase-
expressing recombinant S. lividans strains were isolated by shotgun-cloning
and extracellular H^Oj and peroxidase assays were performed using standard
techniques. Protoplast formation, protoplast regeneration, plasmid transfor-
mation, and plasmid characterization procedures were carried out as we
previously described. Growth of wild-type and recombinant Streptomvces
strains, Streptomvces soil inoculation procedures, selective Streptomyces
enumeration procedures, measurement of carbon mineralization, and the Palouse
silt loam soil used have been previously described.
149
-------
RESULTS AND DISCUSSION
We have demonstrated that strains such as £. lividans TK23-3651 are
transitory over-producers of extracellular hydrogen peroxide and peroxidases.
They lose these enhanced activities after inoculation into soil and after 1-3
transfers on laboratory media. Thus, new strains must be continually
constructed in order to study this phenomenon. Additional study has shown
that the temporarily enhanced activities probably result from unstable gene
duplications within the chromosome of S_. lividans. The gene sequences within
the duplicated DNA are transitorily amplified in expression. This phenomenon
is apparently fairly common in Streptomvces and is observed when cells are
protoplasted and then reverted to vegetative cells on regeneration media.
Thus, when we protoplast our S. lividans strains prior to performing plasmid
transformations, we generate such amplified genetic variants.
We have also shown that we can mimic the C02 mineralization effect over
short periods of time (2-4 days) in non-sterile soils by supplementing them
with a purified lignin peroxidase from Streptomvces viridosporus T7A.
Addition of H202 alone does not significantly affect C02 evolution from the
soil, and peroxidase addition in combination with peroxidase give results
similar to those observed with peroxidase addition alone. Therefore, we have
hypothesized that the enhancement of soil organic carbon mineralization rates
previously observed after release of strains such as TK23-3651 (see above)
resulted from its transitory over-production of extracellular peroxidases. We
believe that the mechanism by which the peroxidase affects carbon turnover
rate involves the activity of the peroxidases on lignified soil organic
matter. This in turn affects a rate-limiting step in carbon turnover in soil,
that is, lignin depolymerization. If the lignin depolymerization rate
increases, recalcitrant, highly lignified organic matter in the soil becomes
more available to the non-1ignin-degrading microflora. This may be the reason
why we observe the greatest enhancement of carbon mineralization rates by
strains such as TK23-3651 when they are released into non-sterile soils
amended with lignocellulose. Our hypothesis also implies that lignin peroxi-
dases are expressed by members of the soil microflora at levels that limit the
overall biodegradation rate.
To further test the hypothesis, we shotgun-cloned the ALipP3 lignin
peroxidase gene, on a 4 kb fragment of chromosomal DNA from S. viridosporus
T7A, into several non-1ignin-degrading S. lividans strains (TK64, TK23, and
TK24) using plasmid vector pIJ702. pIJ702 is a stable, strongly-expressed,
high copy number plasmid in S. lividans. Clones expressing the ALip-P3 gene
are stable in the presence or absence of plasmid-encoded selection pressure
(thiostrepton resistance), and these clones produce large amounts of extra-
cellular lignin peroxidase. We have recently released selected of these
lignin peroxidase over-expressing recombinants (e.g., strains TK64.1, TK24.1,
and TK23.1, all expressing pIJ702.LP) into sofl. We do observe enhanced rates
of carbon mineralization from soils containing these strains. The level of
enhancement is strain specific and also varies with the nutritional status of
the soil. However, the pattern of increased C02 evolution from the soils
inoculated with these recombinants is similar to that previously observed with
the unstable strains.
150
-------
Based upon these results, it appears that our hypothesis on the mechanism
of the carbon mineralization effect is correct. The results also show that
upon release some genetically engineered microorganisms will measurably affect
the environment. The instances where measurable affects will occur are
probably predictable in most cases, as they were in the present instance. The
actual environmental risks of releasing GEMS such as our recombinant S.
lividans strains remain to be assessed. All of our work has been done in
laboratory simulations, and none of our experiments have been carried out past
30 days.
FUTURE WORK
The fact that we have recombinant Streptomvces which affect the rate
of carbon turnover in soil, and that the effect is related to their over-
production of lignin peroxidases, is an important finding . However, it does
not establish if there is an actual risk to the environment if these GEMS are
released. Therefore, further research should emphasize risk assessment.
Questions that need to be asked include: a) What are the short and long term
effects of these recombinants on overall soil microbial activity?; b) Are
there long term changes in soil microfloral extracellular enzyme levels
(cellulases, peroxidases, xylanases, etc.) after release of these recombin-
ants?; c) Will the plasmid-encoded lignin peroxidase gene present in released
recombinants transfer to other members of the soil microbial community, and
upon expression in those organisms result in long term elevated lignin
peroxidase levels in the soil?; and d) Can measurable changes in the atmos-
pheric or soil Carbon Cycle be detected over the short or long term in
environments where these recombinants have been released? We will address
each of these topics in our continuing research.
PUBLICATIONS
Bleakley, B.H., and D.L. Crawford. 1989. The effects of varying moisture and
nutrient levels on the transfer of a conjugative plasmid between Streptom.yces
species in soil. Can. J. Microbiol. 35:544-549.
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 the environment:
Lignin-peroxidase-expressing recombinant Streptomvces can effect the carbon
cycle in soil. To be Presented, 90th Annual Meeting, American Society for
Microbiology, Anaheim, CA. May 13-17.
Crawford, D.L., Z. Wang, B.H. Bleakley, and F. Rafii. 1990. Cloning and
expression of a lignin peroxidase gene from Streptomyces viridosporus in
Streptomvces lividans. J. Biotechnol. (in press).
Rafii, F., and D.L. Crawford. 1988. Gene transfer among Streptomyces. In
Gene Transfer in the Environment, S.B. Levy and R.V. Miller (eds.). McGraw-
Hill Publ. Co., New York. pp. 309-345.
151
-------
Rafii, F., and D.L. Crawford. 1989. Donor/recipient interactions affecting
plasmid transfer among Streptomvces species: A conjugative plasmid will
mobilize non-transferable plasmids in soil. Curr. Microbiol. 19:115-121.
Wang, Z., and D.L. Crawford. 1990. Effects of genetically engineered micro-
organisms in the environment: Lignin peroxidase-expressing recombinant
Streotomvces can effect the rate of carbon mineralization in an
agricultural soil (in preparation).
Wang, Z., D.L. Crawford, A.L. Pometto III, and F. Rafii. 1989. Survival and
effects of wild-type, mutant, and recombinant Streptomvces in a soil
ecosystem. Can. J. Microbiol. 35:535-543.
152
-------
Pseudomonas putida PP0301(pR0103), GENETICALLY
ENGINEERED TO DEGRADE 2,4-DICHLOROPHENOXYACETIC
ACID AFFECTED MICROBE-MEDIATED PROCESSES IN SOIL
J.D. Doyle,1 K.A. Short,3'4 R.J. King,1 G. Stotzky,5
R.H. Olsen,4 and R.J. Seidler3
NSI Technology Services Corporation1
National Research Council2
U.S. EPA3
Environmental Research Laboratory
Corvallis, OR 97333
University of Michigan4
Ann Arbor, MI 48109
New York University5
New York, NY 10004
INTRODUCTION
To evaluate methodologies previously developed at New York University and
EPA-CERL for detecting the effects of genetically engineered microorganisms
(GEMs) on microbe-mediated ecological processes in soil, Pseudomonas putida
strains PP0301 and PP0301(pR0103) were inoculated individually into non-
sterile soil. P. putida PP0301(pR0103) was chosen as the test organism, as it
contains the catabolic genes necessary for the degradation of 2,4-dichloro-
phenoxyacetic acid (2,4-D) to chloromaleylacetate. These genes are plasmid-
borne, and are expressed constitutively. The ability of PP0301(pR0103) to
degrade 2,4-D to an intermediate compound could affect the ecological homeo-
stasis of the soil environment. As £. putida is a common inhabitant of soil,
this GEM is an ideal organism for assessing methods for detecting the effects
of GEMs on microbe-mediated ecological processes in soil.
METHODS
The soil (Millican) used is classified taxonomically as Xeric/Aridic
Frigid and contains 68.1% sand, 29.4% silt, 2.5% clay, and 2.26% organic
matter, with a pH of 6.2 and a cation-exchange capacity of 10.6 meq/100 g
soil. Soil passing through a 2-mm mesh screen was adjusted to the -33 kPa
water tension and amended with 1% glucose (w/w), 500 ppm 2,4-D, glucose + 2,4-
D, or left unamended.
Two studies were conducted. In the first study, P. putida was inoculated
at 105 CFU/g soil. The inoculum was increased to 107 CFU/g soil in the second
study. Respiration (C02 evolution), activity of soil enzymes (arylsulfatase,
acid and alkaline phosphatase, and dehydrogenase), species diversity (total
bacteria, spore-forming bacteria, fungi, and chitin utilizers), disappearance
of 2,4-D, and survival of P. putida strains were evaluated in each study.
153
-------
RESULTS AND DISCUSSION
The response patterns of the uninoculated controls were similar in the
two studies. The addition of the strains of P. putida perturbed some
ecological processes relative to the uninoculated controls. The effects of
PP0301 were similar in both studies. In contrast, differences were observed
between the first and the second studies in soil inoculated with
PP0301(pR0103). In the first study, the effects of the GEM were similar to
those of the parental strain. In the second study, the effects of the GEM
appeared to be related, in part, to its ability to degrade partially 2,4-D. No
degradation of 2,4-D was observed in the first study or with PP0301 in either
study.
The presence of 2,4-D did not inhibit or stimulate the populations of
either the parental or GEM strains. The overall trends for the parental and
the GEM strains were similar in both studies: there was an initial decline in
numbers during the first few days followed by an increase and subsequent
stabilization of bacterial numbers. No treatment differences were observed
between the survival patterns of these bacterial strains.
In the second study, C02 evolution was suppressed during the early phase
of the incubation in soil inoculated with the GEM and amended with glucose +
2,4-D in comparison to soil amended with only glucose. No similar suppression
occurred in either uninoculated soil or in soil inoculated with PP0301. The
numbers of fungi declined rapidly in 2,4-D-amended soil inoculated with
PP0301(pR0103). This decline did not occur in 2,4-D-amended soil that was
uninoculated or inoculated with the parental strain. Moreover, fungal numbers
in the unamended soil inoculated with the GEM declined more rapidly than in
the unamended soil inoculated with PP0301. The numbers of total bacteria in
unamended soil inoculated with the GEM and soil amended with 2,4-D and
inoculated with the GEM were lower than similarly treated soils inoculated
with the parental strain, for days 12-39 of the experimental period. There
was no effect of the GEM relative to the parental strain on populations of
spore-formers or chitin utilizers. There were differences in the dehydrogen-
ase activities between soil inoculated with the GEM and soil inoculated with
PP0301 or not inoculated. After Day 6 of the incubation, dehydrogenase
activity increased in soil inoculated with the GEM and amended with 2,4-D or
glucose + 2,4-D. This increase in activity was not observed in soil inocu-
lated with the parental strain. The increase in dehydrogenase activity
appeared to be correlated with changes in CO, evolution from soil inoculated
with the GEM and amended with glucose + 2,4-D. No effects of the GEM on
arylsulfatase activity was apparent. In contrast, unamended soil inoculated
with the PP0301(pR0103) had a lower acid and alkaline phosphatase activity
then did unamended soil inoculated with PP0301.
These studies showed that the genetically engineered bacterium, P. putida
PP0301(pR0103), can cause changes in microbe-mediated ecological processes in
soil that are distinct from those induced by the parental strain. Hence, the
methods developed were appropriate for detecting GEM-induced perturbations of
microbe-mediated processes in soil. Further, there appeared to be a minimum
inoculum density of PP0301(pR0103) necessary for producing detectable ecologi-
cal effects. There were no detectable differences in effects between the
154
-------
parental and GEM strains when both were inoculated at 105 CFU/g soil. More-
over, 2,4-D was not metabolized at this inoculum density. When inoculated at
107 CFU/g soil, the GEM had a more marked effect on microbe-mediated soil
processes than did the parental strain. Additionally, 2,4-D was metabolized
in soil inoculated with the GEM, but not in soil inoculated with the parental
strain. This is the first study to show that a GEM can induce non-transient
changes in microbe-mediated ecological processes in non-sterile soil.
PUBLICATIONS
Doyle, J., K. Short, R. King, R. Seidler, R. Olsen, and G. Stotzky. 1989.
Pseudomonas outida PP0301(pR0103), genetically engineered to degrade 2,4-D,
affected microbe-mediated processes in soil. Abstr. Ann. Meet. Northwest
Branch Am. Soc. Microbiol.
Doyle, J., K. Short, and G. Stotzky. 1989. Effects of Pseudomonas putida
PP0301(pR0103), genetically engineered to degrade 2,4-dichlorophenoxyacetic
acid, on microbe-mediated processes in soil. Abstr. Ann. Meet. Am. Soc.
Microbiol. Q140.
155
-------
THE EFFECT OF 2,4-DICHLOROPHENOXYACETATE AND A GENETICALLY ENGINEERED
MICROORGANISM METABOLITE, 2,4-DICHLOROPHENOL, ON SOIL MICROBIAL ECOLOGY
K.A. Short,1'4 J.D. Doyle,2 R.J. King,2 G. Stotzky,4
R.H. Olsen,"5 and R.J. Seidler3
National Research Council1
NSI Technology Services Corporation2
U.S. EPA3
Environmental Research Laboratory
Corvallis, OR 97333
Department of Biology4
New York University
New York, NY 10004
Department of Microbiology and Immunology5
University of Michigan
Medical School
Ann Arbor, MI 48109
INTRODUCTION
At the 1988 All Investigators Meeting, Short, Olsen and Seidler demon-
strated the capacity of the genetically engineered microorganism (GEM),
Pseudomonas putida PP0301(pR0103), to degrade 2,4-dichlorophenoxyacetate (2,4-
D) in an Oregon agricultural soil. The microbiota indigenous to this soil was
also capable, approximately one week after the application-of 500 ppm 2,4-D,
of degrading 2,4-D. No metabolites of 2,4-D were observed in uninoculated
soil or in soil inoculated with the GEM and amended with 2,4-D.
In the present study, an Oregon desert soil was amended with 500 ppm 2,4-
D. The soil is classified as a Xeric/Aridic Frigid soil, and is approximately
68% sand, 29% silt, and 3% clay and contains 2% total organic matter. We
assessed the capacity of the GEM and the capacity of the microbiota, indigen-
ous to this desert soil, to degrade 2,4-D. The effect of 2,4-D on some
microbe-mediated processes of this soil was also studied. The objectives of
this research were to: a) assess and develop laboratory-based methods for
determining the effect of a GEM on microbe-mediated processes in soil; b)
evaluate these methods with P. putida PP0301(pR0103) and determine the
capacity of the indigenous microbiota and the GEM to degrade 2,4-D in soil; c)
assess the effect of 2,4-D on microbe-mediated processes; d) identify the
primary metabolite of 2,4-D observed to accumulate in the Xeric/Aridic Frigid
soil; and e) determine the toxicity of this metabolite to fungi.
METHODS
The bacterial strains used were P. putida PP0301 and PP0301(pR0103).
Plasmid pR0103 is expressed constitutively and encodes for the degradation of
2,4-D to chloromaleylacetate via a five step reaction. In order, the metabo-
lites of this pathway are 2,4-dichlorophenol, 3,5-dichloro-muconic acid, 2-
chloro-4-carboxymethylene but-2-enolide and chloromaleylacetate.
156
-------
RESULTS AND DISCUSSION
The addition of 500 ppm 2,4-D, to the uninoculated soil, had a marked
effect on soil microbial ecology. All activities of the enzymes measured-
dehydrogenase, alkaline phosphatase, acid phosphatase, and arylsulfatase, as
well as gross metabolic activity (C02 evolution) were lower in 2,4-D-amended
than in unamended soil. The reduced enzyme activities were coincident with
the lower numbers of total bacteria and chitin degraders in 2,4-D amended
soil, relative to unamended soil. The numbers of spore-forming bacteria and
fungi, however, did not appear to be effected by the addition of 2,4-D.
Degradation of 2,4-D occurred only in the presence of the GEM. The
addition of PP0301(pR0103) (approximately 5 X 106 cells/g 2,4-D-amended soil)
resulted in the degradation of 2,4-D from an initial concentration of 500 ppm
to a final concentration of 100 ppm. Concurrent with this degradation, a
metabolite of 2,4-D accumulated to > 50 ppm, and soil fungal propagules in
soil declined approximately 500-fold. From these observations we hypothesized
that the GEM metabolite was specifically toxic to fungi. Gas chromato-
graphic/mass spectrophotometric analyses of liquid and soil cultures, inocu-
lated with the GEM, indicated that the metabolite was 2,4-dichlorophenol (2,4-
DCP). The toxicity of 2,4-DCP to fungi was confirmed by two different
experiments with five morphologically distinct fungal cultures isolated from
the experimental soil. In the first study, sterile soil was amended with 25,
50, or 100 ppm 2,4-DCP and inoculated with each of the five fungal isolates.
The inoculated soil was periodically replica-plated onto Martin's agar to
measure the rate of growth of the fungi. Increasing concentrations of 2,4-DCP
inhibited the spread of the fungi and was lethal in some cases. In the second
study, the relative toxicity of 2,4-D and 2,4-DCP to fungi-was determined.
Different concentrations of 2,4-D or 2,4-DCP were incorporated into unsolidi-
fied Martin's agar, poured into petri plates and allowed to gel. Each of the
fungal cultures were then spread-plated onto the various media. Less than 50
ppm 2,4-DCP was lethal to all of the fungal isolates, whereas there was
minimal toxicity (68% recovery relative to a control) with 200 ppm 2,4-D.
The results of this study indicated that: a) neither the indigenous
microbiota of the xeric soil nor the parental, P. putida PP0301, were able to
degrade 2,4-D; b) the addition of 2,4-D to the uninoculated xeric soil
affected microbial diversity, soil enzyme activity, and gross metabolic
activity; c) the GEM metabolized 2,4-D in this soil, resulting in the accumu-
lation of 2,4-DCP; and d) the accumulation of the metabolite, 2,4-DCP, was
primarily responsible for the observed decrease (approximately 500-fold) in
numbers of soil fungal propagules.
PUBLICATIONS
Doyle, J., K. Short, R. King, R. Seidler, R. Olsen, and G. Stotzky. 1989.
Pseudomonas putida PP0301(pR0103), genetically engineered to degrade 2,4-D,
affected microbe-mediated processes in soil. Abstracts, 1989 Meeting of the
Northwest Branch of the American Society for Microbiologists, Seattle, WA
(submitted).
157
-------
Doyle, J,, K. Short, and G. Stotzky. 1989. Effects of Pseudotnonas outida
PP0301(pR0103), genetically engineered to degrade 2,4-D, on microbe-mediated
ecological processes in soil. Abstracts, 1989 Meeting of the American Society
for Microbiologists, Miami Beach, FL.
Short, K.A., J. Doyle, R. King, G. Stotzky, and R.J. Seidler. 1990. Effect
of 2,4-dichlorophenoxyacetate and a GEM metabolite, 2,4-dichlorophenol, on
microbe-mediated processes in the soil. Abstracts, 1990 Meeting of the
American Society for Microbiologists (submitted).
Short et al. 1990. A GEM degrader of phenoxy-herbicides affected soil
ecology. Abstracts, 1990 Meeting of the American Society for Microbiologists,
Anaheim, CA (submitted).
158
-------
FIELD CALIBRATION OF SOIL-CORE MICROCOSMS FOR EVALUATING ECOSYSTEM
EFFECTS AND FATE OF GENETICALLY ENGINEERED MICROORGANISMS
Harvey Bolton, Jr. and James K. Fredrickson
Battelle Pacific Northwest Laboratory
Richland, WA 99352
INTRODUCTION
Laboratory microcosms are potential tools for evaluating the risk
associated with the release of Genetically Engineered Microorganisms (GEMs)
into the environment. Laboratory microcosms are attractive for determining
the risks associated with the release of GEMs because tests and microorganisms
can be contained. Recently, intact soil-core microcosms have been used to
evaluate the fate and ecological effects of transposon mutants of Azospirillum
lipoferum and a deleterious wheat root growth-inhibiting Pseudomonas sp.
However, to enhance the utility of microcosms as a biotechnology risk assess-
ment tool, they must be first calibrated with field observations in order to
be effective field models.
Pseudomonas sp. rhizobacteria that are aggressive root colonizing
organisms that can reach high populations on both the rhizoplane and various
plant residues were selected for the initial field calibration study. There
were two major objectives to this comparison of microcosms with field observa-
tions. First, determine the fate of root-colonizing bacteria added to soil
that was planted to winter wheat (Triticum aestivum L.) in field plots, field
lysimeters, and intact soil-core microcosms incubated in a-growth chamber and
the laboratory. Second, calibrate ecosystem structure and function in
microcosms and the field, since these can also serve as indicators of
ecosystem impact.
METHODS
Field Site, Microcosms. Inoculation, and Locations
The field site was located in southeastern Washington State. The site
receives 19.5 cm annual precipitation with a mean annual temperature of 10°C.
Intact soil-core microcosms (17.5 cm diameter, 60 cm length), field lysimeters
(17.5 cm diameter, 60 cm length), and field plots (17.5 cm diameter, open
ended rings) contained four treatments including: a) surface soil (0 to 15 cm
depth) inoculation with Pseudomonas sp. RC1 (RC1), a spontaneous rifampicin-
resistant (100 ug ml"1) mutant, and a soil amendment of 1% (w/w) alfalfa; b)
inoculated and unamended; c) uninoculated and amended; and d) uninoculated and
unamended. Soil incubation and planting of winter wheat at four locations
including: a) microcosms incubated at ambient laboratory temperature (22°C)
(ambient microcosms); b) microcosms incubated in a growth chamber (chamber
microcosms) with temperature fluctuations that simulated average field values;
c) field lysimeters; and d) field plots occurred on October 29, 1988. The
alfalfa amended soil received 8.2 log CFU of RC1 g"1 dry soil, while the
unamended soil received 7.7 log CFU g"1 dry soil.
159
-------
Fate of Pseudotnonas sp. RC1 and Ecological Effects
RC1 surface soil (0 to 15 cm depth) populations were determined over time
by dilution and plating on Sands agar containing 50 ug ml"1 rifampicin (Sigma)
(Sands-rif ). Wheat roots were sampled at the three-leaf and boot stage of
wheat growth for enumeration of rhizoplane populations. At the three-leaf
stage, one wheat seedling with intact roots was removed from the ambient
microcosms, growth chamber microcosms, and field lysimeters and plots on 18,
45, and 144 d after planting, respectively. At boot stage, which was reached
on 67, 110, and 205 d after planting for the ambient microcosms, growth
chamber microcosms, and field plots and lysimeters, respectively, soil cores
were destructively sampled. At the boot stage sampling, the intact soil-core
from each treatment was divided into three sections (top: 0 to 15 cm; middle:
15 to 35 cm; and bottom: 35 to 55 cm) with wheat roots sampled from each
depth.
RC1 ^hizoplane poc ations were determined at the three-leaf and boot
stage by Dilution plating on Sands-rif50. At the three-leaf stage, rhizoplane
populations of total, fluorescent, and non-fluorescent pseudomonads, and total
aerobic heterotrophs were also determined. Tryptic soy agar plates were also
used to calculate a Shannon-Weaver species diversity index. Distinct colony
types were treated as separate species, while the number of each colony type
was treated as the number within those species. Wheat shoot biomass was
determined at the three-leaf and boot stage. Soil dehydrogenase activity in
the surface soil (0 to 15 cm depth) was determined at the three-leaf stage of
wheat growth.
RESULTS AND DISCUSSION
Soil (0 to 15 cm depth) populations of RC1 declined more rapidly in the
order: ambient microcosm > chamber microcosm > field plot = field lysimeter.
The alfalfa amendment initially increased the surface soil population of RC1,
but RC1 populations declined with time. When soil populations of RC1 were
compared at the three-leaf stage of wheat growth, the alfalfa amended soil had
populations that were 0.5 to 1.4 log units higher than the unamended soil.
Although RC1 soil populations declined more rapidly in the ambient microcosms,
at the three-leaf stage of wheat growth, RC1 soil populations were less than 2
log units different at the four locations despite a 125 d difference from the
start of the experiment to the three-leaf stage. At the boot stage of wheat
growth, the soil populations in the growth chamber were 1 log unit higher than
the three other locations. We hypothesize the growth chamber temperature and
moisture conditions were more favorable to the introduced microorganism in
comparison to either the field or the ambient microcosms.
There was no difference in wheat root colonization by RC1 between the
different location treatments at the three-leaf stage. The mean across all
treatments was 8.7 log CPU g"1 dry root. These results indicate that intact
soil-core microcosms may simulate microbial fate in the field quite well, even
though they may not duplicate field conditions. At the boot stage sampling,
wheat rhizoplane populations of RC1 decreased with an increase in depth. The
wheat roots sampled from the top and bottom of the cores had similar popula-
tions at the four locations, but there was greater variation in the middle
160
-------
depth. This difference with depth demonstrates the importance of quantifying
bacterial populations at different soil depths and not relying on a single
composite sample for evaluating bacterial survival and fate.
RC1 competed favorably with native pseudomonads and other heterotrophic
bacteria that colonized the rhizoplane, with the RC1 population comprising 98%
and 40% of the total pseudomonad and bacterial rhizoplane populations,
respectively. Inoculation with the non-fluorescent pseudomonad RC1, signifi-
cantly reduced the percent of fluorescent pseudomonads on the wheat rhizo-
plane, in comparison to the control, from 24% to 1%.
The Shannon-Weaver index of species diversity, based on heterotrophic
colony diversity, was significantly higher in the growth chamber microcosms
(2.7) than the field lysimeters (2.3), field plots (2.1) or ambient microcosms
(2.0). Inoculation with RC1 had no effect, while the alfalfa amendment
significantly increased the index from 2.1 to 2.5. Soil conditions in the
growth chamber soil-core microcosm were believed to be more conducive to
bacterial proliferation than the three other locations because of a more
favorable temperature regime.
Soil dehydrogenase activity at the three-leaf stage of wheat growth was
significantly higher in the chamber microcosm and with alfalfa amendment,
while RC1 inoculation had no effect. The alfalfa addition increased soil
dehydrogenase approximately three times that of the unamended soil as would be
expected from the addition of readily decomposable organic material. Wheat
shoot biomass at both the three-leaf and boot stage was significantly lower in
the ambient microcosms than the three other locations. The alfalfa amendment
significantly reduced shoot biomass at the three-leaf stage- and was likely due
to the addition or simulation of plant pathogens or saprophytes with the
alfalfa amendment. Several of the seedlings in the alfalfa amended treatment
were stunted and chlorotic. These plants were able to overcome this initial
stunting and over time their biomass increased to a value significantly
greater than the control, likely due to a greater supply of N from alfalfa
mineralization.
The results from the first year of this project indicate that soil-core
microcosms can be used to predict the fate of introduced microbial populations
and the structure of indigenous soil and rhizosphere microbial populations in
the field. For many of the ecosystem parameters measured in this study, the
agreement between microcosm and the field data was much closer than a temporal
reference point such as plant phenology (or degree-days) which was used for
comparison. There were some unexpected results with regards to differences in
the behavior of the growth chamber microcosms and the ambient laboratory
microcosms relative to the field. In general, the growth chamber microcosms
were more favorable for introduced and indigenous microbial populations as
well as for plant growth. We hypothesize that this is due to the more
favorable temperature and moisture conditions which were maintained in the
growth chamber in comparison to either the field or the ambient laboratory
microcosms. This emphasizes the need to control and/or monitor key environ-
mental variables such as moisture and temperature in order to better calibrate
microcosm behavior with the field.
161
-------
FUTURE WORK
The second year of our field calibration study will consist of two parts.
First, a continuation of the Pseudomonas sp. RC1 study, and second, we will
use Streptomyces lividans TK24 with different experimental parameters studied
as described below.
Pseudomonas so. RC1 (RC1) Field Calibration Study
A second year field calibration experiment will be conducted with field
lysimeters and growth chamber microcosms inoculated or uninoculated with RC1
to obtain two years of field data. The year to year variability for tempera-
ture and moisture in the field necessitates at least two years of data for a
valid field-growth chamber comparison. Measurements will again focus on
bacterial fate and microbial structure and function on the wheat rhizoplane.
Streotomvces lividans TK24 (TK24) Field Calibration Study
We will initiate an additional field calibration study using TK24
introduced into soil-core microcosms and field lysimeters. TK24 was chosen as
a contrast to RC1. It is a gram-positive spore forming bacterium that belongs
to a class of microorganisms important in soil carbon cycling. TK24 has been
a recipient for genetically engineered plasmids in gene transfer experiments
and some ecological effects work has been conducted in the laboratory.
Therefore, we will concentrate on the fate of TK24 and C cycling processes as
calibration (and effects) endpoints. Treatments will include growth chamber
microcosm and field lysimeters inoculated or uninoculated with TK24 and 1%
cellulose added. Microplots that contain a spike of C-14 labeled cellulose
will be included so that an in-depth understanding of C cycling rates and pool
sizes can be obtained to compare growth chamber microcosms with field
lysimeters.
PUBLICATIONS
Bolton, H., Jr., and J.K. Fredrickson. 1988. Field calibration of soil-core
microcosms for evaluating fate and effects of genetically engineered micro-
organisms in terrestrial ecosystems. U.S. EPA, Biotechnology Risk Assessment
Research Review, Baltimore, MD. October 11-14, 1988.
Bolton, H., Jr., J.K. Fredrickson, S.A. Bentjen, D.J. Workman, S.W. Li, and
J.M. Thomas. 1989. Field calibration of soil-core microcosms: I. Microbial
fate of genetically engineered microorganisms (in preparation).
Bolton, H., Jr., J.K. Fredrickson, J.M. Thomas, S.W. Li, D.J. Workman, and
S.A. Bentjen. 1989. Field calibration of soil-core microcosms: II. Effects
of genetically engineered microorganisms on ecosystem structure and function
(in preparation).
Fredrickson, J.K. 1989. Terrestrial microcosm and field evaluations of the
fate and effects of genetically altered rhizosphere bacteria. Symposia on
GEMs in the environment at the American Society for Microbiology Annual
Meeting. New Orleans, LA.
162
-------
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. 1989. Nutrient efflux from intact soil-core microcosms for
evaluating the impacts of genetically engineered microorganisms on ecological
processes. Environ. Tox. Chem. (in press).
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.
Fredrickson, J.K., P. Van Voris, S.A. Bentjen, and H. Bolton, Jr. 1989.
Terrestrial Microcosms for evaluating the environmental fate and risks
associated with the release of chemicals or genetically engineered micro-
organisms to the environment. In J. Saxena (ed.), Hazard Assessment of
Chemicals -- Current Developments 7:157-202. Hemisphere Publishing Co., New
York, NY.
163
-------
EFFECTS OF INTRODUCING GENETICALLY ENGINEERED MICROORGANISMS
ON MICROBIAL COMMUNITY DIVERSITY AND USE OF CONDITIONALLY
LETHAL (SUICIDAL) MICROORGANISMS TO LIMIT RISK
Ronald M. Atlas, Michael H. Perlin, and Asim K. Bej
Department of Biology
University of Louisville
Louisville, KY 40292
INTRODUCTION
To examine the potential ecological effect of deliberately releasing
genetically engineered microorganisms (GEMs) into the environment, the 2,4,5-
trichlorophenoxyacetic acid (2,4,5-T) degrader Pseudomonas ceoacia AC1100 was
introduced into soil microcosms. Periodically, microorganisms were enumer-
ated, isolates were selected at random and characterized, and cluster analyses
were performed to estimate microbial taxonomic diversities. DNA was extracted
from the microcosms and the DNA reannealing kinetics were measured so that C t
plots could be used to estimate the genetic diversities of the soil microbial
communities.
In addition to their potential effect on diversity of the microbial
community to which they are released, a separate concern with GEMs is that
once released into the environment, they cannot be recalled. One approach for
containing GEMs released into the environment is to create GEMs that are
potentially suicidal, that is, organisms containing conditionally lethal
genes. The key to one of the more interesting prospective-conditional suicide
systems is the hok gene, which codes for a small polypeptide (Hok) that causes
loss of cell membrane potential and rapid death of cells in which it is
overproduced. Molin et al. fused the hok gene to the Escherichia coli trp
promoter to create a conditional suicide system such that in the absence of
tryptophan there is no active represser, leading to over-expression of hok and
death of the cells; they also fused the hok gene to the invertible fimA
promoter, which controls periodic expression of type 1 fimbriae in E. coli so
as to limit population size but not to eliminate directly the modified E.
coli. In studies conducted in vitro, both the trp promoter-hok and the fimA
promoter-hok constructs were shown to be capable of conditional killing of
cells and, thereby, of limiting population size. The studies by Molin et al.
were designed for the conditional maintenance of a strain rather than for
conditional death. We fused the hok gene to the lac promoter of E. coli and
tested the ability of the construct to trigger the death of cells by the
introduction of an inducer both in vitro and in soil microcosms. We also
examined the effect of a counter selective pressure on survival following Hok
induction.
Besides E. coli, we examined the effectiveness of a suicide vector
containing hok in pseudomonads which are likely candidates for deliberate
releases of GEMs.
164
-------
METHODS
Bacterial Strains. Growth Media, and Plasmids
The bacterial strains used in this study were as follows: E. coli JM101
(sufiE, thl. delta lac-proAB. [F', traD 36, proAB. laclq. lacZ delta M15], U.S.
Biochemical, Cleveland, OH); £. coli JM83 (ara, lac-pro rpsL (SmR), thi.
080dl_acZM15); P. putida mt-2 (cure), P. cepacia AC1100.
The plasmids used were: plasmid pTZ19u (pUC19 replicon, AmpR, Lac*,
ParB', U.S. Biochemical, Cleveland, OH); plasmid pPR633 (pBR322 replicon,
AmpR, Lac", ParB*, S. Molin, Copenhagen, Denmark); and pVDtac24 (Ampr, lacZ.
1aclq. mob, rnc, oriV. tac). Plasmid pTZ19u (U.S. Biochemical, Cleveland,
OH), which contains a multiple cloning site within the lacZ gene, was used as
a vector for plasmid construction. Plasmid pPR633 was used as a source of the
hok gene. Plasmid pVDtac_24, which contains multiple cloning sites within the
lacZ' gene, was used as a vector for plasmid construction. Plasmid pPR633 was
used as the source of hok gene.
Soil
A silt loam soil from Corvallis, Oregon, was used for microcosm experi-
ments. Soil properties, determined by standard methods for soil analysis,
were: organic content = 6.6%; sand 28%; silt 44%; clay 28%. Microbiological
properties of the soil were: viable heterotrophs, determined by plating
serial dilutions onto 0.1 strength trypticase soy agar and incubating at 28°C
for 48 h = 1.3 x 106 CFU/g dry wt; carbenicllin resistant heterotrophs,
determined by plating serial dilutions onto 0.1 strength tirypticase soy agar
containing 200 ug/tnl carbenicillin and incubating at 28°C for 48 h = 0 CFU/g
dry wt.
Construction and Cloning of the Suicide Vector (pBAP19h) Containing hok Gene
Partial digestion of plasmid pPR633 with AceII and SSP! restriction
enzymes was used to generate a series of blunt end fragments. DNA fragments
ranging between 400 and 600 bp, as determined by comparison with lambda
digested with Hindlll size standard DNA, were isolated from Seaplaque (0.7i)
agarose gel. The fragments included the hok gene (493 bp), but excluded sok
and the hok promoter. The hok gene-containing fragment was cloned into Smal-
digested pTZ19u vector treated with bacterial alkaline phosphatase to yield
pBAP19h. Alternatively, the vector pVDtac24 was digested with Xbal and the
ends were filled in with the Klenow fragment of DNA polymerase I. The treated
vector was then ligated to the fragments from pPR633 using T4 DNA ligase (U.S.
Biochemical, Cleveland, OH) as described by Bej et al., to yield plasmid
pBAP24h. Transformation of E. coli with pBAP19h and Pseudomonas putida mt-2
with pBAP24h was performed using standard methods.
Induction of hok in P. putida and P. cepacia was carried out as for E.
coli except the cells were grown at 30°C with or without 500 mg per ml
carbenicillin (CB) and cells were plated onto TYE agar containing 500 mg per
ml CB. In addition, 5 mM IPTG was added to the culture for hok induction.
165
-------
Also, the £. outida transformants were maintained on Pseudomonas isolation
agar containing 1000 ug carbenicillin per ml.
Hok Gene Induction In vitro
The induction of hok gene was determined by growing E. coli JM101
containing pBAP19h plasmids in 20 ml 2 x YT broth with or without 200 ug/ml
carbenicillin.
Hok Induction in Soil Microcosms
To examine hok induction in soil microcosms, E. coli JM101 containing
pBAP19h plasmid was grown to early log phase in 2 x YT broth containing 100
ug/ml carbenicillin. Cells were centrifuged and washed twice with 0.7%
sterile saline solution. Cells were resuspended in distilled water such that
a final cell concentration of 1 x 107 cells/g could be achieved by adding
sufficient solution to reach 60% soil water holding capacity. Cells were
added to 12 replica microcosms, each containing 50 g of soil. The microcosms
consisted of flower pots with an opening in the bottom for drainage. The soil
in one set of microcosms had been pre-sterilized by autoclaving at 121°C for
30 min. The soil in another set of microcosms was supplemented with nutrient
broth by substituting nutrient broth for distilled water in the cell
suspensions used for inoculation. A third set of microcosms contained non-
sterile soil and no nutrient addition. Microcosm controls, containing soil
not receiving bacterial inoculation, were also included. Microcosms were
incubated at 28°C.
One gram samples were collected from the central and peripheral regions
of each microcosm 2 and 24 h after addition of the bacteria. Serial dilutions
in 0.1 M phosphate buffer (pH 7.6) were plated in duplicate onto TYE agar
containing 200 ug/ml carbenicillin. Immediately after collecting the 24 h
sample, IPTG was added to a final concentration of 1.5 mg/ml to half of the
microcosms of each set. An equal volume of sterile distilled water was added
to the remaining half of the microcosms. To achieve as uniform a distribution
of IPTG in the soil as possible, half of the total IPTG solution was added to
the top of the soil and the other half was placed into a petri-plate beneath
the soil; the IPTG containing solution was drawn into the soil by capillary
action; an air current was maintained over the microcosms by using a fan so as
to maximize evaporative transport of IPTG into the soil. Additional samples
were collected at 36 and 48 h, that is, 12 and 24 h after addition of IPTG,
and serial dilutions were plated as described above. Colony hybridizations
were performed to identify colonies containing the hok gene.
RESULTS AND DISCUSSION
C0t Plots Measuring Diversity After Introduction of a GEM
For microbial soil communities into which the 2,4,5-T degrader P. ceoacia
AC1100, we examined diversity using reannealing kinetics of DNAs from the
communities at various time intervals after introduction. We found that P.
ceoacia AC1100 persisted during the 6 week experiment but dominated the
community only during the first weeks after its introduction. Disturbance due
166
-------
to the introduction of P. cepacia AC1100 increased the genetic diversity of
the soil microbial community. Genetic diversity increased most dramatically
after the introduction of the GEM in soils treated with 2,4,5-T.
Induction of hok Gene in Soil Microcosms
The model suicide vector (pBAP19h), designed for the potential contain-
ment of GEMs, was made by constructing a plasmid with the hok gene, which
codes for a lethal polypeptide, under the control of the lac promoter. The
vector plasmid also codes for carbenicillin resistance. In the absence of
carbenicillin, induction of the hok gene in vitro caused elimination of all
detectable cells containing the suicide vector pBAP19h-free cells of the
culture survived and grew exponentially. In the presence of carbenicillin,
however, cells containing pBAP19h initially declined after induction of hok.
but then multiplied exponentially. The surviving cells still had a fully
functional hok gene and had apparently developed resistance to the action of
the hok polypeptide. Thus, high selective pressure against the loss of the
suicide vector led to a failure of the system. Soil microcosm experiments
confirmed the ability of a suicide vector to restrict the growth of a GEM in
the absence of selective pressure against the loss of the plasmid, with 90-99%
elimination of hok-bearing cells within 24 h of hok induction. However, some
pBAP19h-bearing cells survived in the soil microcosms after hok induction.
These surviving cells contained an active hok gene but were not capable of
normal growth even after elimination of the hok gene; it appears that a
mutation that made them Hok resistant also reduced their capacity for membrane
functions needed for energy generation and exponential cell growth. Thus, the
model suicide vector was shown to be functional in soil as well as in vitro.
However, this system is not yet adequate to provide a fail.safe system for
containment of genetically engineered microorganisms.
Since the lac promoter does not function efficiently in pseudomonads a
hybrid trp-lac promoter, called tac, was created which functions well in most
of the Pseudomonads and all E. coli tested. Plasmid vector pVDtac24 has such
a promoter. Also the vector pBDtac_24 used in the construction of pBAP24h is
that laclq. which is required to suppress the expression of the hok gene.
After transformation with pBAP24h, growth and survival of the host organism
and retention of pBAP24Ji were examined for Pseudomonas putida, the genetic-
ally-engineered 2,4,5-T-degrader P. cepacia AC1100 and, for comparison,
Escherichia coli. In Pseudomonas species, hok induction caused elimination of
greater than 60% of cells carrying pBAPtac24h within 5 h of induction compared
to complete elimination in E. coli within 3 h. The results indicate that the
conditional lethal ("suicide") system pBAP24h can kill both E. coli and
Pseudomonas species, providing a mechanism of containment of genetically-
engineered microbes deliberately released into the environment; however, this
system is less effective in Pseudomonas spp.
The in vitro experiment in the absence of carbenicillin clearly estab-
lished the potential of the model suicide vector pBAP19h to restrict the
movement, proliferation, or persistence of the genetic information stored on
the same plasmid. The soil microcosm experiments confirmed the ability of a
suicide vector to restrict the growth of a GEM, although they did not estab-
lish an absolute margin of safety. However, the in vitro experiment in the
167
-------
presence of carbenicillin showed that this suicide vector was not a fail safe
system for containment. Thus, while the constructed suicide vector has the
potential for containment, more restrictive suicide vectors will be necessary
to ensure environmental safety if such vectors are used to contain GEMs.
The "suicide vector" pBAP24h showed differential stability and efficiency
of killing pseudomonads compared to E. coli. Plasmid stability was high for
E. coli and hok* cells were effectively eliminated after induction. In
contrast, in the pseudomonads tested the plasmid was relatively unstable,
being lost during multiple transfers on non-selective media. Hok induction in
the pseudomonads showed less dramatic killing than in £. coli. In E. coli and
in P. putida. when there was antibiotic selection for the pBAP24h plasmid,
hok-containinq survivors were present by 4 h, and at least in the case of £.
coli, the percentage increased thereafter. Survivors of hok induction may
represent: hok-resistant mutants, hok' mutants (e.g., the E. coli with
deletions in the pBAP24h plasmid), surviving cells which have lower hok gene
dosage, or cells "adapted" to the Hok protein which are phenotypically Hok-
resistant.
PUBLICATIONS
Atlas, R.M., A.K. Bej, R.J. Steffan, and M.H. Perlin. 1988. Approaches for
monitoring and containing genetically engineered microorganisms released into
the environment. Hazardous Waste Treatment by Genetically Engineered or
Adapted Organisms. Washington, D.C.
Atlas, R.M., A.K. Bej, R.J. Steffan, and M.H. Perlin. 1989a. Approaches for
monitoring and containing genetically engineered microorganisms released into
the environment. Hazardous Waste and Hazardous Materials (in press).
Atlas, R.M., A. Bej, R.J. Steffan, and M.H. Perlin. 1989b. Environmental
monitoring and containment of genetically engineered microorgan- ms. Meetings
of the American Chemical Society. Dallas, TX.
Bej, A.K., M.H. Perlin, and R.M. Atlas. 1988. Model suicide vector for
containment of genetically engineered microorganisms. Appl. Environ. Micro.
54:2472-2477.
Bej, A.K., M.H. Perlin, and R.M. Atlas. 1989. "Suicide" systems for the
containment of genetically engineered microorganisms: model conditional
lethal systems for E. coli and Pseudomonas using hok. Molecular Genetics of
Bacteria and Phages. Cold Spring Harbor, NY.
168
-------
SESSION VII
HIGHER ORGANISM EFFECTS
169
-------
-------
SAFETY PROTOCOLS FOR EVALUATING THE EFFECTS OF MICROBIAL
PEST CONTROL AGENTS ON NON-TARGET, BENEFICIAL ARTHROPODS
Rosalind R. James1 and Bruce Lighthart2
NSI Technology Services Corporation1
U.S. EPA2
Environmental Research Laboratory
Corvallis, OR 97333
INTRODUCTION
Subsection M of the Pesticide Assessment Guidelines in the Federal
Insecticide, Fungicide and Rodenticide Act (FIFRA) requires testing of a
microbial pest control agent (MPCA) for toxicity or pathogenicity to nontarget
beneficial insects and mites. Tests are required for predacious mites,
neuropterans (lace wings), coleopterans (beetles), hemipterans (true bugs),
and parasitic hymenopterans (wasps) and dipterans (flies), and honey bees.
Interim protocols for these tests have been developed for all of the above
insect groups as well as for litter arthropods and non-target butterflies
(Lepidoptera). Final protocols for bioassays have been developed after
laboratory tests. An important strategy considered for the bioassays was that
of maximizing the susceptibility of the arthropods to MPCAs. Some environ-
mental conditions were found to increase the susceptibility of the insect to
the pathogen independent of the effect of the condition on the microorganism.
Final protocols have been developed for the predacious mite Metaseiulus
occidental is and the predacious neuropteran Chrysoperla carnea and are in
progress for the parasitic hymenopteran Trichogramma pretiosum and the
predatory coleopteran Hippodamia converqens.
METHODS
First, a test arthropod species must be identified for each of the
species categories. Each insect must have the following characteristics: fit
E.P.A. subsection M guidelines, be widely available (preferably, commercially
available), be culturable in the laboratory, and have known pathogens avail-
able to use in testing the bioassay.
Once a stable culture of the arthropod is developed, acceptable methods
of exposing it to the MPCA need to be developed. The method used depends on
the size and habits of the insect or mite and the type of microorganism used
(i.e., virus, fungi, protozoan, or bacteria).
Preliminary experiments were then carried out to evaluate the effect of
environmental conditions on the susceptibility of the insect. Four conditions
were tested: temperature, relative humidity, age, and nutrition. Tests are
developed for both adults and larvae, gender will be included as a condition
for the adult test (larval gender cannot be determined).
For H. converqens (the convergent lady beetle), each factor was first
tested independently using Pseudomonas fluorescens. a weak bacterial pathogen
of the beetle. First instar larvae (up to 24 hours old) were exposed to one
171
-------
of five temperatures (10, 20, 25, 30, 35'C) or five humidities (40%, 70%, 80%,
96% RH) for eighteen hours before exposure to the pathogen. Larvae were
exposed to the pathogen by dipping them in a known concentration of bacterial
suspension for ten seconds. Control larvae were treatejd in the same manner
except they were not exposed to the pathogen. The effect of nutrition was
tested by giving newly emerged first instar larvae one of four diets (water
only, 5% sucrose in water, 5% sucrose + 1% casein in water, or aphids) for
twenty four hours before being exposed to the pathogen. The aphid diet is
normally used to rear the beetles. Larvae do not survive for more than two or
three days on the other diets, so after 24 hours, all treatments were main-
tained on aphids. Control larvae were treated in a similar manner except they
were not exposed to the pathogen.
For each of the experiments, mortality was compared after six days (there
is little change in mortality after day five). Each environmental condition
was carried out before the insects were exposed to the pathogen to separate
the effects of the condition on the susceptibility of the insect from its
effects on the pathogenicity of the microbe. To test the effect of insect
development on susceptibility to the pathogen, the LC50 of each larval instar
was also determined. This series of experiments was used to identify which
conditions acted as stresses on the insect.
Identified stress factors are then to be used in a fractional factorial
experiment to evaluate the effects of, and interactions between, the stresses.
Two levels of each stress factor ("on" and "off") are used and analyzed as a
multi-way anova. Lethal concentration^ (LC50) or Lethal dosage50 (LDcq) levels
are used in the anova since they are more sensitive than the % mortality used
in the above experiments.
RESULTS AND DISCUSSION
For H. convergens, an increase in the mortality of insects exposed to the
pathogen occurred at both high and low temperatures. The effect of tempera-
ture is only weakly significant (P<0.08), but there was also a large amount of
variation in the data generated using % mortalities. Diet also affected
mortality due to the pathogen; the more complete the diet the beetles had, the
lower their mortality. However, the combination of protein (casein) and sugar
(sucrose) seemed to have had an effect little different from sugar alone. The
susceptibility of larvae greatly decreased with age, where the LC50 of first
instar larvae was five log doses lower than the LC50 of fourth (last) instar
larvae. Humidity had an effect different from the other conditions in that
high and low humidities decreased susceptibility. In other words, those
humidities that were other than optimum for rearing the insect did not act as
a stress on the insect.
The factorial design will include two factors: temperature at two levels
(25° and 35°C) and nutrition at three levels (starved (water only), sugar fed
(5% sucrose), and aphid fed). For each condition and each combination of
conditions, the LC50 of the pathogen on first instar larvae will be determined
and the LC50 values will then be analyzed in a two-way anova. Adult beetles
will similarly be treated and tested except gender will be included in the
analysis.
172
-------
FUTURE WORK
After completion of the bioassay protocols, future research is antici-
pated to move beyond the laboratory and focus on development of methods to
evaluate any negative impact of MPCA applications to sites with non-target,
beneficial arthropods. This research would focus on how to monitor applica-
tion sites, how to determine whether any negative effects seen are due to
direct or indirect effects of the MPCA or due to other factors, and how to
mitigate the MPCA if necessary.
PUBLICATIONS
Donegan, K., and B. Lighthart. 1989. Effect of several stress factors on the
susceptibility of the predatory insect, Chrysoperla carnea (Neuroptera:
Chrysopidae), to the fungal pathogen Beauveria bassiana. Journal of Inverte-
brate Pathology 53:79-84.
Lighthart, B., D. Sewall, and D.R. Thomas. 1988. Effect of several stress
factors on the susceptibility of the predatory mite, Metaseiulus occidental is
(Acari: Phytoseiidae), to the weak bacterial pathogen Serratia marcescens.
Journal of Invertebrate Pathology 52:33-42.
173
-------
USE OF Bacillus thurinqlensis kurstaki FOR SPRUCE BUDWORM
CONTROL: EFFECTS ON NON-TARGET LEPIDOPTERA
Jeffrey C. Miller and Brian Scaccia
Department of Entomology
Oregon State University
Corvallis, OR 97331-2907
INTRODUCTION
Microbial pest control agents (MPCAs) are a primary means for controlling
insect pests in the context of biological control. In general, the use of an
MPCA is targeted for a particular pest species. For instance, large-scale use
of Bacillus thurinqiensis Berliner var. kurstaki (Bkt) against the gypsy moth
and spruce budworm has been commonly employed over forested habitat. However,
non-target species that are related to the target pest may also be effected in
an adverse manner. For instance, field application of Btk (three treatments
in a single season) for the gypsy moth in western Oregon reduced species
richness and larval abundance for up to a two-year period within a guild of
native, non-target Lepidoptera feeding on oak.
Many concerns need to be addressed regarding the use of MPCAs, particu-
larly in large-scale programs and in the advent of the emergency of genetic-
ally altered organisms. Among some of the concerns are what the impact of an
MPCA would be on non-target populations that have important functions in food
webs, the fate of endangered species, and the conflict with other biological
control agents, such as Lepidoptera on weeds. Also, studies on ecological
effects of standard MPCAs are needed for reference to compare with data sets
concerning ecological effects involving genetically engineered MPCAs.
The research data presented here were derived from an investigation with
the objective of quantifying the effects of a single Btk treatment, targeted
for the spruce budworm, on a guild of native, non-target leaf-feeding
Lepidoptera.
METHODS
The study was conducted during the spring and summer of 1989 and focused
on the guild of immature Lepidoptera (caterpillars) that feed on the foliage
of Ceanothus velutinus Doug!., tobacco brush. The study sites were 50 km
south of Estacada, Clackamas County, Oregon. The area is on the western slope
of the Cascade Mountain range at 1000-1200 m elevation where the mean maximum
temperature in July is 24-28°C and precipitation ranges from 160-200 cm per
year, mostly between December and March.
Four plots were established. Two of the plots were located within a 4000
ha region treated with Btk during late June. The sample plots were treated
June 26. The Btk was applied from a helicopter at the rage of 8 BIU per 2.8
liters of water per 0.4 ha. The other two plots were located outside the
spray zone by at least 2 km. Because of heterogeneity in features of topo-
graphy and plant community, the plots were matched as pairs using the criteria
174
-------
of: a) physical aspects of the habitat, b) size of £. velutinus plants, and
c) the number of C. velutinus plants within patches of 100 x 100 m. Each
matched pair of pTots was represented by a treated and untreated site and was
sampled between June 21 and August 25. Each treated plot was approximately 3-
5 km from a matched untreated plot.
The caterpillar fauna was sampled by shaking foliage over a 75 x 75 cm
sheet for a timed interval of 30-45 sec. Sampling was continued until a total
of 180-220 sec of sampling effort was achieved at each plot. The samples were
analyzed to determine species richness and individual abundance of all
immature leaf-feeding Lepidoptera.
The statistical analysis of the data was treated in two ways. First, the
total number of immature Lepidoptera and the number of species were assessed
between the matched treated and untreated plots with a Student's t-test. Then
the data among the plots were pooled for an 'overall' analysis to compare
between treatments using a Chi-square test. Levels of significance were
established at P < 0.05.
RESULTS AND DISCUSSION
Species Composition
A total of 32 species belonging to nine families were collected during
the study. The most commonly encountered groups of species were in the
Geometridae, Noctuidae, Lycaenidae, and Gelechiidae. The most abundant
species were two geometrids, Drepanulatrix falcataria (Pack.) and Eupithecia
sp.; a lycaenid, Satyrium saepium saepium (Boisduval); and.a gelechiid,
Chirnidess sp. The highest number of species was observed during July and
early August, within 24-45 days after the Btk application in the treated
plots.
Comparison of treatments regarding species richness indicated that Btk
significantly affected the number of species of Lepidoptera on C. velutinus.
While the number of species in the untreated plots increased from 15 in the
pre-spray sample to 19 in the first post-spray sample, the number of species
in the treated plots decreased from 15 to 10. The second post-spray sample,
on August 5, contained significantly more species (21) in the untreated plots
than in the treated plots (12). No significant differences were observed in
species richness in the third sample, 65 days after the Btk application.
Larval Abundance
A total of 1029 immature Lepidoptera were collected from all plots.
Overall, 65% of the larvae were collected in the untreated plots. However, if
only the first two sample dates following treatment are compared, than 76%
(n=489) of the larvae were collected in the untreated plots.
The differences in the overall numbers of larvae collected between the
pre-spray samples and the first post-spray samples further demonstrated the
impact of Btk on non-target Lepidoptera. While the number of larvae per 100
sec sample effort in the untreated plots increased from 75.2 in the pre-spray
175
-------
to 80.1 in the first post-spray sample, the number of larvae in the treated
plots decreased from 75.5 to 14.9. The second post-spray sample, on August 5,
contained significantly more larvae, 28.3, in the untreated plots than in the
treated plots, 14.3.
An interesting effect was observed in the sample taken 65 days after the
Btk application. Differences in larval abundance showed a significant
increase in the treated plots. The cause for this increase in the treated
plots is unknown but could be related to 1) changes in competition among
species of the herbivore guild, 2) predators could have left the lower density
sites allowing a resurgence, or 3) plant physiology-herbivore nutrition was
altered.
The data developed in this study addressed a set of effects on non-target
organisms which are case-specific but nonetheless pertinent to at least two
important issues.
Lepidoptera as an Ecological Resource
Populations of Lepidoptera serving as biological control agents of
noxious weeds may be at risk. A reduction in their population density could
result in an increase in the seed set and populations of their host plant.
Secondly, under certain conditions, a reduction in caterpillar abundance could
negatively affect the population dynamics of some birds. Thirdly, rare or
endangered species of Lepidoptera could be at risk of extinction if their
populations were restricted to areas encompassed by pest control programs,
even those using Btk as the control agent.
Historical Data Base for MPCA Evaluation
A feasible field assessment protocol is needed to study concerns about
impacts on non-target organisms for existing MPCAs and genetically altered
MPCAs, including those incorporated into the genome of a plant. A data base
documenting the effects of a "standard" MPCA may then be used as a reference
for assessing the impact of that MPCA after genetic engineering. Ideally,
additional studies on non-target populations will be able to incorporate an
array of relevant field conditions involving the considerations and concerns
that occurred during this study.
176
-------
THE USE OF EXPERT SYSTEMS AND DATABASE TECHNOLOGY IN ESTIMATING
MICROBIAL PEST CONTROL AGENT IMPACT ON NON-TARGET ARTHROPODS
B. A. Croft and R. H. Messing
Department of Entomology
Oregon State University
Con/all is, OR 97331
INTRODUCTION
The United States Environmental Protection Agency (U.S. EPA) has framed
many of its regulatory decisions on pesticides in terms of a "Risk Assess-
ment/R-sk Management" paradigm. In our presentation, we will discuss the data
requirements and especially their integration for pesticide risk assessment.
Our examples are taken from impacts of microbial pathogens on terrestrial non-
target beneficial arthropods, and more specifically the natural enemies of
pest insects and mites. Use of these systems as examples should bring out
some of the data requirements for risk assessment that apply generally to
other non-target groups (e.g., avian, mammal systems), as well. For data
integration, we will discuss and demonstrate an expert support system, which
uses knowledge sources of data bases, models, expert opinion, and "rules of
thumb."
Risk assessment logic and knowledge bases are similar from one organismal
group to another or from one general risk assessment sub-activity to another
(e.g., for hazard identification, organismal response assessment, exposure
assessment, and risk characterization). They all include a problem solving
logic path, a knowledge base, the opportunity for user input (i.e., a user
interface), and lastly an output of risk assessment prediction. In the
remainder of our presentation, we would like to illustrate how these data
elements and processes can be integrated and viewed, generally. First, we
will discuss the use of database and expert systems technology as a means to
accomplish this integration. We will then demonstrate a prototype system that
estimates the risk of impact of microbial pathogens on non-target, terrestrial
arthropod natural enemies. The specific product is called NERISK.
METHODS
The basic interfaces of an expert system are a chain of reasoning or
decision logic and an inference engine which allows the overall systems to
interact with a user in a question-advise mode. The inference engine also
often keeps track of the user's question sequencing, his use of the expert's
advice, and even allows him to give input to the designer on ways to improve
the system. With certain user prompts, the inference engine fires off or
retrieves certain elements of the knowledge base (rules, facts, relationships)
or information types such as models, database files, expert opinion summaries,
communications, etc.
The NERISK program has been developed in association with our colleagues:
Kevin Currans, Ray Drapek, Karen Theiling, and Bruce Lighthart of OSU or the
U.S. EPA, Corvallis. It uses database and modeling technology, as well as
177
-------
expert opinion and testing protocols to evaluate the impacts of microbial
pesticides. The SELCTV database (a database management system with 30
information fields for over 500 records on microbials, but over 15,000 records
on toxics, in general. Theiling), is in an advanced stage of development and
serves as a major information source for NERISK. The modeling technology used
in NERISK focuses on describing the epizootic dynamics of an insect host (pest
usually), a MPCA, and one or more natural enemies of the pest arthropod. A
set of interim test protocols for microbial pesticides on five insect natural
enemies (developed by B. Lighthart) is another component of the system
accessible from NERISK, which provides additional information for MPCA
evaluation.
RESULTS AND DISCUSSION
The System Components of NERISK
A flow chart of the general operation of the expert system and its
components and the introductory screen for the program are shown in Figs. 1
and 2, respectively. In Fig 1, an overview of the system construction and the
logic behind the organization of its various parts is illustrated. The
perspective was designed primarily for U.S. EPA regulatory agents who must
predict risk in order to make decisions regarding registration of new
compounds. The flow chart illustrates the basic two-fold nature of the NERISK
system, as a "traditional" expert system using rule-based forward chaining,
and as an interface, integrator, and interpreter of diverse knowledge sources.
Starting with an application providing biological and chemical informa-
tion regarding the proposed compound, and with a realistic.management frame-
work providing economic, environmental, and legal context, the user applies
NERISK to frame the problem, define the components, and access the knowledge
sources. For each knowledge source these is a direct route of entry for the
experienced user and a more "user-friendly" front-end program for the novice.
Output from the database or the model can be channeled into the expert-rule
framework. Output from any component can be used to generate graphs or
reports, or can be checked against standard assessment protocols to determine
if relevant and sufficient test procedures have been carried out (i.e., to
provide the basic application information). The NERISK integrator then
prioritizes and integrates all program output to give the user a single
quantitative estimate of risk assessment.
The Resources of NERISK
A MANUAL, serves to organize the encoded information into discrete
modules ("Chapters"), and allows access to various component features of
NERISK. The SELCTV database, interim test protocols, and the epizootic
simulation model can be accessed through this mechanism.
A GUIDED-SESSION leads the user, by a system of question and prompts,
through the chain of reasoning which an expert might employ in evaluating
MPCAs. User input defines the problem and helps guide the logic path to the
desired evaluation.
178
-------
A NON-GUIDED SESSION allows the experienced user to take shortcuts and
pick out specific information modules for analysis.
ACCESSORIES include tools such as a DICTIONARY (defines new or unfamiliar
terms), REFERENCES (provides full literature citations to key publications),
CHAPTER INDEX (lists all information modules in the program), and a SCRATCHPAD
(allows the user to take notes during a work session for later analysis, and
also automatically records in memory key data which are incorporated into the
quantitative risk assessment).
An INTEGRATOR assimilates data from the scratchpad, allows the user to
prioritize both biological and resource importance values, and presents a
weighted average risk assessment value.
In addition, an important feature of the program is the "WHY" function,
which allows the user to question the justification or reasoning for any
statement or invocation rule. The program responds with an explanation of the
logic behind the rule. The user may also leave messages within the program to
the "expert" or systems designer for periodic troubleshooting, suggested
improvements, etc.
PUBLICATIONS
Croft, B.A., ed. 1989. Arthropod Biological Control Agents and Pesticides.
Wiley Intersci. J. Wiley and Sons, New York, N.Y. 736 pp.
Croft, B.A. 1989. Standardized assessment methods. Chap. 5. p. 101-126.
In Arthropod Biological Control Agents and Pesticides, ed, B.A. Croft. J.
Wiley and Sons, New York, N.Y. 736 pp.
Croft, B.A., and J.L. FTexner. 1989. Microbial pesticides. Chap. 11. p.
269-303. In Arthropod Biological Control Agents and Pesticides, ed. B.A.
Croft. J. Wiley and Sons, New York, N.Y. 736 pp.
Croft, B.A., R.H. Messing, and J. Hendricks. 1988. Data integration for risk
assessment: A case study of microbial pathogen impact on terrestrial arthro-
pods, in The Integration of Research and Predictive Model Development in
Biotechnology Risk Assessment. EPA Wkshp. Proc. ORD/OTS, Washington, D.C.
455 pp.
Croft, B.A., and Karen Theiling. 1989. Pesticide effects on arthropod
natural enemies: a database summary. Chap. 2. p. 17-46. In Arthropod
Biological Control Agents and Pesticides. B.A. Croft (ed.). J. Wiley and
Sons, New York, N.Y. 736 pp.
Flexner, J.L., B.A. Croft, and B. Lighthart. 1986. The effects of microbial
pesticides on non-target, beneficial arthropods. Agric., Ecosyst. and
Environ. 16:203-254.
Messing, R.H., and B.A. Croft. 1990a. NERISK: An expert system for predict-
ing the effects of pesticides on arthropod natural enemies. Acta Hort. (in
press).
179
-------
Messing, R.H., and B.A. Croft. 1990b. NERISK users manual. Public. O.S.U.
Dept of Entomol. 40 pp. (Contact B.A.C. for copies)
Messing, R.H., B.A. Croft, and K. Currans. 1989. Assessing pesticide risk to
arthropod natural enemies using expert system technology. A.I. Applic. Nat.
Res. Mngmt. 3:1-12.
Theiling, K. 1987. The SELCTV Database: The Susceptibility of Arthropod
Natural Enemies of Agricultural Pests to Pesticides. M.S. Thesis. Oregon
State University, Con/all is, OR. 170 pp. (B.A.C., major professor).
Theiling, K., and B.A. Croft. 1987. The SELCTV Database. Version 1.1.
Copyright O.S.U., Software Product., Dept. Entomol., O.S.U. Corvallis, OR.
Theiling, K., and B.A. Croft. 1988a. Pesticide effects on arthropod natural
enemies: A database summary. Agric., Ecosys. and Environ. 21:191-218.
Theiling, K., and B.A. Croft. 1988b. Toxicity, selectivity and sublethal
effects of pesticides on arthropod natural enemies: A database summary.
Abst. Symp. Evaluation of Pesticide Side-Effects on Non-target Invertebrates.
Proc. XVIII Intern. Cong. Entomol. Vancouver, B.C. p. 464.
Theiling, K.M., and B.A. Croft. 1989. Toxicity, selectivity and sublethal
effects of pesticides on arthropod natural enemies: A data-base summary.
Cpt. 11. p. 213-232. J_n Pesticides and Non-target Invertebrates. P.C.
Jepson (ed.). Intecept Publ. Dorset, U.K. 240 pp.
Theiling, K.M., and B.A. Croft. 1990. Systems manual for-SELCTV and REFER
databases and SELCTV data management system program. Public. O.S.U. Dept. of
Entomol. 60 pp. (Contact B.A.C. for copies).
180
-------
SAFETY PROTOCOLS FOR EVALUATING HAZARDS OF MICROBIAL
PEST CONTROL AGENTS TO NON-TARGET AVIAN SPECIES
Anne Fairbrother1 and Phyllis S. Buchholz2
U.S. EPA1
Environmental Research Laboratory
NSI Technology Services Corporation2
Corvallis, OR 97333
INTRODUCTION
Microbial pest control agents (MPCAs) are microorganisms applied to the
environment to control proliferation and spread of agricultural or silvicul-
tural insect, arthropod, or plant pests. We have been working for the past
three years to develop standardized protocols that can be used to determine
the pathogenicity of proposed MPCAs to nontarget avian species. The mallard
(Anas platvrhvnchosl and bobwhite (Colinus viroinianus) were selected as test
species as they are used routinely by U.S. EPA for evaluation of effects of
chemical pesticides on wildlife. Three routes of exposure of the birds to
MPCAs were evaluated: oral, intravenous (IV), and respiratory. This report
discusses only work done to evaluate oral and IV routes; the respiratory
exposure methods will be discussed in a separate report. Microorganisms
representing three classes of MPCAs were used to test the protocols.
Autoqrapha californica nuclear polyhedrosis virus was used as a representative
viral agent; Metarrhizium anisopliae was used for a fungal agent; and
Salmonella pullorum (a known avian pathogen) was selected as a representative
bacterium.
METHODS AND RESULTS
Initial work focused on establishing suitable husbandry conditions for
the 1 to 40 day-old chicks and ducklings used in the tests. Brooder
dimensions and temperature specifications were written into the protocols.
Attention was given to the need for appropriate animal facilities for contain-
ment of potentially pathogenic microorganisms. Feed and water systems were
developed to determine the feasibility of monitoring consumption as an
indication of sublethal infection or stress. It was determined that the
variability in measurements was too great to detect any pathogen-induced
reduction in either feed or water consumption by these species. Therefore,
this requirement was dropped from the protocols. Body weight measurements,
serum chemistry parameters, antibody production, and gross and histopathology
were included in the final protocols as they appear to be sensitive indi-
cators of infection.
Appropriate dosing methods were examined. A decision was made to use
gelatin capsules for oral dosing rather than gavage needles in order to reduce
the amount of agent regurgitated. It was determined that IV inoculation tech-
niques could be successfully accomplished, even for small (20 gram) quail
chicks. Both species of birds were inoculated orally and intravenously with
live organisms to verify that appropriate doses could be achieved and admin-
istered. The viral and fungal agents were nonpathogenic to the birds. They
181
-------
caused no detectable changes in the physiological parameters measured, no
antibodies were detected, and no organisms were recovered from the tissues or
feces. The bacterium, Salmonella pullorum. is a known pathogen of domestic
chickens and was extremely virulent in the young quail, causing mortality of
significant numbers with viable organisms recoverable from feces and tissues
of all birds, even those that survived 14 days post-exposure. The mallard
ducklings were more resistant to infection and none died following inocula-
tion. Gross lesions were observed on lungs of ducklings euthanized 14 days
post-exposure. Organisms were isolated from lungs, spleen, heart, and
pancreas of these ducklings. Mallards developed detectable antibody titers
within one week of inoculation with the bacterium. Antibody determinations
for quail were inconclusive due to small volumes of serum and few birds that
survived long enough to seroconvert.
DISCUSSION
These experiments showed that the protocols are suitably designed to
demonstrate pathogenicity of a microorganism in the test bird species.
Although the two birds require substantially different husbandry practices,
the study with Salmonella pullorum emphasized the necessity for examining
pathogenicity of proposed MPCAs in at least two bird species. The interim
protocols were revised to reflect the necessary additional information on
husbandry of the two species and changes in requirements of endpoints to be
measured.
FUTURE WORK
The oral and intravenous exposure methods have been completed and await
peer review comments. Respiratory exposure methods are under development
through an extramural cooperative agreement and standardized protocols will be
written within the year.
PUBLICATION
Fairbrother, A., M. Craig, K. Walker, and D. O'Loughlin. 1989. Effects of
age, sex, and reproductive status on normal serum chemistry of mallards (Anas
platvrhvnchos). J. Wildl. Dis. 26(l):(in press).
182
-------
RESPIRATORY EXPOSURE OF AVIAN WILDLIFE TO AIRBORNE MICROBIAL
PEST CONTROL AGENTS: ASSESSMENT METHODOLOGIES
3
Crystal J. Driver,1 Linda G. Smith,1 James K. Briant,1
Ann E. Jarrell,1 Anne Fairbrother,2 and Phyllis Buchholz
Battelle Pacific Northwest Laboratory1
Richland, WA 99352
U.S. EPA2
Environmental Research Laboratory
NSI Technology Services Corporation3
Corvallis, OR 97333
INTRODUCTION
Regulatory agencies now face the task of assessing the safety of deliber-
ate releases of biological control agents (including genetically engineered
microbes) into natural systems. Prior to authorization for use of a new
microbial pest control agent (MPCA), safety of these agents to non-target
avian wildlife must be determined. Wildlife may be exposed to MPCAs through
inhalation either during or following a spray event. Because the avian
respiratory tract differs fundamentally from that of the mammalian system,
inhalation of MPCAs by birds, their deposition in the respiratory tract and
resultant pathogenicity cannot be predicted from mammalian respiratory models.
Information on particle distribution in the avian respiratory system and
protocols for assessing the pathogenicity of inhaled MPCAs in birds are
needed.
Respiratory exposure can be simulated by aerosol inhalation or by
intratracheal (IT) instillations of MPCAs. IT instillations are unrealistic
in that they bypass the filtration and impaction that occurs at the nasal
operculum and in the nasal turbinates, and introduce a bolus of solution with
abnormal pressure into the respiratory airways. However, containment of the
microbe is greater by IT methods, than by the whole body exposures commonly
used for aerosol inoculations of pathogenic microbes. In addition to contain-
ment problems with whole body exposures, aerosol generation over a wide range
of particle sizes represented by microbes (0.4 to 2.0 urn diameters) is
technically difficult, and economically and spatially costly. The purpose of
this study was to develop cost-effective protocols for respiratory exposure
and pathogenicity testing of MPCAs in non-target avian wildlife and to compare
their efficacy with regard to hazard assessment and worker safety.
METHODS
A small economical inhalation exposure system suitable for head-only
exposure of 10-day-old northern bobwhite (Colinus virqiananus) chicks was
constructed. Respiratory rate and tidal volume of the chicks as well as chick
holding-tube temperature were monitored. Exposure chamber temperature and
humidity and aerosol characteristics were also measured during each exposure.
Optimum exposure conditions for the chicks was determined during mock aerosol
exposures using distilled water. Particle distribution in the respiratory
system of young birds was characterized by exposing 10-day-old chicks to
183
-------
aerosols containing fluorescent polystyrene microspheres. Birds were sacri-
ficed by halothane overdose to inhibit post-mortem movement of respiratory
cilia. Nasal turbinates were flushed with saline. Quantification of the
particle distribution in the lower respiratory tract was accomplished by
infusing standard plate agar which was cooled to the point of solidifying
(viscous) into the airways through the trachea. The infused birds were then
placed on ice to solidify the agar. Intact air sacs, trachea, and lungs were
removed and digested enzymatically. Microspheres were enumerated by micro-
scopy and fluorometry. Particle size deposition was confirmed by histologic
examination of the respiratory tract. The exposure protocols were tested by
exposing 10-day-old chicks to a viable organism. Ornithine negative-motility
negative E. coli. by IT instillation or by aerosol exposure. Control birds
were exposed to sterile media. Birds were sacrificed 15 minutes post-
exposure. Positive controls were instilled post mortem with 2 ml of the E.
coli inoculum. Chicks were covered in adhesive drape, skinned and rinsed with
disinfectant :o eliminate external (e.g., feather) contamination of the
respiratory svstem. External contamination was monitored by swabbing the
head, beak, body and subcutaneous tissue prior to disinfection. Subcutaneous
surfaces were sampled again following disinfection. Air sacs and trachea of
all birds were sampled for MPCAs by swabbing. Because the syrinx is an
important point of infection due to the restricted air flow through its
structure, the syrinx was removed and cultured separately from the trachea and
lungs. The lung was sampled through a dorsal incision by removing a section
of the lung lateral to the thoracic vertebrae. Nasal turbinates were flushed
with saline. Swab and tissue samples were transferred to thioglycolate broth
for 24 hours. The enriched samples were then plated on tryptic soy agar (with
5% sheep red blood cells). Developing colonies were subplated onto MacConkeys
agar and confirmed as motility negative ornithine negative --£. coli in motility
indole ornithine media.
RESULTS AND DISCUSSION
An aerosol exposure chamber was constructed that generates the full
spectrum of MPCA particle sizes (viral to fungal spore). Quantitative
evaluation of fluorescent microsphere delivery to the exposure ports under
exposure conditions showed that the system delivers particles ranging in size
from 0.4 to 20 urn uniformly and in suitable concentrations. Larger sizes (10
and 20 urn) are transported through the chamber with efficiency similar to that
of the smaller particles (0.4 to 5 urn). Relative humidity is controlled at
50% which ensures that aerosol particles are of proper aerodynamic size at the
exposure ports. The chamber is economical, compact and easily decontaminated.
Respiratory physiology of the chicks as well as body temperature and chamber
and holding-tube temperatures can be monitored during exposure. A temperature
of 32 to 35°C inside of the occupied bird holding-tubes resulted in a steady
respiratory rate similar to that of unrestrained chicks at rest (but not
sleeping) in their home cage. A relative humidity of 50% provided comfort and
adequate heat exchange.
Head only exposure and containment of the bird in the holding-tube until
sampling greatly reduced human exposure to potential pathogens. Only 10% of
the head swabs from chicks exposed in the chamber were uncontaminated. Beak,
body and subcutaneous samples were uncontaminated. Intratracheal instilla-
184
-------
tions, by comparison, offered greater opportunity for human exposure. Body
feathers were contaminated in 25% of the IT exposed birds, while the head and
beak swabs were contaminated in over 50% of the chicks. Swabs of the subcu-
taneous tissue were contaminated in 100% of the IT instilled chicks.
Particle distribution in the respiratory tract of the chicks was greatly
influenced by the method of exposure. Few particles greater than 2 urn in
diameter were found in the air sacs and lungs of birds exposed to an aerosol
of microspheres. However, particles >5 urn were commonly found in the lungs
and cervical and thoracic air sacs of IT-exposed chicks. The nasal operculum
effectively filtered out large particles such that the nasal turbinates and
trachea of aerosol exposed birds did not contain the larger particles (>5 urn)
introduced into the trachea of IT exposed birds. Because of its size (2 urn in
length), the distribution of the E. coli was uniform throughout the respira-
tory tract of both the IT and aerosol exposed birds. However, 90% of the
chicks exposed to the bacterial aerosol had syrinxes positive for the E. coli,
whereas only 10% of the syrinxes of IT instilled chicks were contaminated.
Apparently the gentle force at which the IT bolus is instilled does not allow
the bacteria to deposit at this area of restricted flow. The differences in
particle size distribution and in the deposition of particles at critical
infection points in the avian airways have significant impact on the useful-
ness of IT instillations for exposure assessments of field applications of
MPCAs. More realistic assessments can be economically and safely obtained
using the exposure chamber developed for this study.
185
-------
TEST PROTOCOLS TO DETERMINE THE TOXICITY
OF MICROBES TO AQUATIC INVERTEBRATES
Richard L. Anderson1 and Eric Mead2
U.S. EPA1
Environmental Research Laboratory
American Scientific International2
Duluth, MN 55804
INTRODUCTION
The microbial pest control agent (MPCA) project at ERL-Duluth is directed
toward developing an understanding of relationships between the agent and
invertebrates and fish, ecological processes and transport through the aquatic
system. Our research is conducted as a part of the Ecosystem Response Branch
at the laboratory and is integrated with research on mathematical modeling,
effects anc ecological processes research being conducted with chemical
pesticides.
The MPCA research is divided into two sub-projects; one is to develop
single species test guidelines and protocols and the other is to examine
effects and expression in laboratory microcosms and in natural systems treated
with a microorganism to control an insect pest. Both sub-projects examine
issues of exposure, routes of administration, test animal age and test length,
which are important factors in microorganism registration for pest control.
This abstract describes a portion of the research directed toward developing
single species test guidelines and protocols. Another abstract by Shannon and
Anderson describes the microcosm and natural system research.
In our single species test development, we have conducted exposures and
measured either uptake or loss of the registered agent Bacillus thurinqiensls
var israelensis (Bti) in 13 species of invertebrates. The list includes three
daphnid, a copepod, fairy shrimp, two amphipod, 5 insect and a snail species.
Test guidelines have been developed for the daphnids and the copepod and will
be developed for the other species. This report will highlight our work with
the caddisfly, Brachycentrus americanus. The larvae of this insect are
commonly found in trout streams and would be exposed if these streams were
treated with Bti or some other microorganism to control blackfly populations.
Our approach is to examine insects exposed through ingestion of formu-
lated particles in a water exposure and through ingestion of target insects
killed with Bti. Since blackflies were not available, mosquitoes were used as
the target animals.
METHODS
Animals
The Trichopteran Brachycentrus was selected as a non-target insect, Aedes
atropalpus mosquito as a target insect and Bacillus thurinqiensis var
israelensis as the agent. All experiments were conducted in the laboratory
186
-------
with caddisflies collected from a stream, mosquitoes from a laboratory culture
and formulated Bti.
Test Procedures
The insects used in these tests were collected from the Blackhoof River,
a trout stream in Carlton County, Minnesota. The insects were collected by
picking individuals from stones in the river, placing them in glass jars and
transporting them to the laboratory in river water. In the laboratory, the
insects were placed in flowing Lake Superior water and allowed to acclimate to
the new conditions before being used in the experiments.
Analytical Procedures
The spore content of the caddisflies and mosquitoes were determined with
similar methods. Each animal was washed with sterile water, transferred to a
glass tissue homogenizer and disrupted, in water, until all recognizable form
disappeared. The total sample was removed with a pipet and an aliquot was
diluted to provide the unpasteurized values. The remainder was pasteurized to
obtain a spore count. All counts were made after 24 hours of incubation on
tryptic soy agar.
Exposure Procedures
Three sets of procedures were developed for exposing the caddisflies.
One set exposed caddisflies in static conditions and the second in flow-
through conditions. The third procedure exposed Brachvcentrus through
ingestion of contaminated mosquitoes.
The static exposure was conducted in beakers containing aqueous mixtures
of the formulated compound. To assure that the animals had ingested Bti, a
control set of cold-immobilized animals were exposed in ice-water to determine
surface adsorption of Bti. For the flow-through tests, an exposure chamber
that allowed a simulation of actual exposure conditions in natural systems was
developed and tested. The mosquitoes were exposed to the formulated mixtures
in static conditions, and upon death, were rinsed and fed to the caddisflies
in the flow-through system.
RESULTS AND DISCUSSION
Our research exposed both mosquitoes and caddisflies. We expected and
found that mosquitos ingest a large amount of Bti spores. We were surprised
that the mosquito ingested much more than the amounts reported for other
mosquito species. Aedes atropalpus larvae are larger than other mosquitoes
and that may account for the greater accumulation. The accumulation may
actually be higher than we measured, since the spores germinate in dying and
dead mosquitoes. Iced mosquito exposures showed that about 450 spores were
adsorbed per mosquito. This is about 1% of the total adsorbed/ingested during
a warmwater exposure.
The content of the warmwater exposed caddisfly is about 150 x 104 spores
per animal, which is greater than that found in the mosquitoes (4.4 x 104
187
-------
spores). Iced caddisflies adsorbed 33 x 102 spores, which is greater than the
amount adsorbed by the mosquito. This may also be related to size, since the
caddisfly is larger than the mosquito. Caddisflies, fed mosquito larvae,
retained only a small amount, about 1% of the content of the mosquito.
In our flow-through experiments, caddisflies exposed for a short time in
a flow-through system accumulated Bti rapidly from the water and the content
increased after the water was clear of Bti. The additional uptake may be due
to ingestion of adsorbed Bti on the substrate or on the caddisfly case.
There are two forms of exposure in our laboratory experiments: in water
or by feeding on a contaminated insect. The results showed that exposure from
water mixtures of Bti can result in accumulation to the non-target animal.
The amount ingested from a water exposure by the non-target animal was greater
than that ingested by the target animal. We are continuing studies to
determine the uptake and retention of the Bti spores by other invertebrates.
Our investigations also include an examination of the retention of toxicity
after passage through the gut of the non-target animal. These studies, along
with culture and rearing studies not reported in this abstract, will provide a
basis for testing guidelines for invertebrates not commonly tested in the
laboratory.
188
-------
TRACKING Bacillus thurinoiensis var. israelensis (Bti)
IN LABORATORY MICROCOSMS AND NATURAL SYSTEMS
Lyle J. Shannon1 and Richard L. Anderson2
Department of Biology1
University of Minnesota, Duluth
U.S. EPA
Environmental Research Laboratory
Duluth, MN 55804
INTRODUCTION
The ability to manipulate a microorganism's genetic substance offers
opportunities to benefit many aspects of human health and well-being. The
applications for novel organisms appear boundless; new uses are continually
being reported in manufacturing, pharmaceutical, agricultural, mining, and
pollution control industries. Coupled with these potential benefits, however,
is a concern about adverse effects on human welfare and environmental quality.
Indeed, experience with exotic species which have successfully exploited new
niches underscores a need for caution.
Newly developed organisms must be tested to determine their potential
effects on the environment. Since the ability of any new organism to survive
in the wild is unknown, it should first be evaluated in a contained laboratory
test prior to release into the natural environment. To be of value, labora-
tory test systems must be reasonable models of the natural environment, and
accurate predictors of the survival and ecological effects .-of an introduced
microorganism.
We have developed a freshwater microcosm test protocol that has worked
well in chemical testing and shows promise as a microbial test system. It has
been used to test 15 chemicals including pesticides and industrial organic
compounds. In tests with Bacillus thurinqiensis var. israelensis (Bti), a
microbial pest control agent (MPCA) registered for mosquito and blackfly
control, we found that this protocol provided repeatable data on both survival
and ecological effects. Such information is of no value, however, unless it
accurately predicts events in the wild. To determine this microcosm's
predictive capabilities, we began a series of small scale field tests with
Bti.
METHODS
Microcosms
Microcosms are constructed in one-liter pyrex beakers. Each beaker
contains a defined nutrient medium, a sand sediment and a diverse community of
bacteria, fungi, protozoa, algae, rotifers, planktonic and benthic grazers and
is covered with a 150 mm x 15 mm petri dish cover. The cover reduces intro-
duction or loss of microorganisms while providing gas exchange. The micro-
cosms' small size permits a large number to be held in a small space, reduces
concern about containment and increases the statistical power by increasing
189
-------
replication. The starting community "seed" is derived from a variety of
natural sources that are allowed to "co-adapt" in the laboratory. After
adding the "seed" and completing a development phase, the microcosms are ready
to be tested.
To insure that a highly susceptible host was present, 40 second-instar
mosquito larvae (Aedes aegypti or Aedes atropalpus) were added to microcosms.
The presence of a target organism served as a positive control to verify the
toxicity of Bti.
Field Sites
The field sites selected were all small (<4 m diameter) temporary
woodland pools. Such sites are important mosquito breeding areas and are
typical of locations likely to be treated with Bti. Four sites were treated
with commercial formulations of Bti. The first was a small rock-lined pool on
the ERL-D property. We divided this pool into two zones with a floating
curtain. The other sites were located in a forested area two miles north of
Duluth. Two small pools of approximately 1 m diameter and 30 cm deep were
used in 1988 for pilot studies to determine the feasibility of monitoring the
abundance and toxicity of Bti released into the wild. In 1989 two larger
sites (Shannon pools) were selected. Both pools were roughly oval in shape.
The first (the Bti treated site) was approximately 3 m long and 2 m wide. The
second (an untreated control site) was approximately 5 m by 2 m. Both sites
were shallow with a maximum water depth of 12-14 cm. In both years we waited
until the pools developed abundant populations of insect and zooplankton fauna
before beginning the tests. Since this typically occurs in the latter stages
of the wet cycle, we were limited to short (approximately one week) test
exposures before the pools went dry.
Three preparations of Bti spores were tested in three microcosm experi-
ments. Two of these, Vectobac™ (Abbott Laboratories lot #52-009 BR) and
Mosquito Attack™ (Reuters Laboratories) were commercially formulated
products. The third preparation was a pure culture of washed Bti spores grown
in our laboratory.
Ecosystem effects were determined through regular measurements of produc-
tion, respiration, pH and P/R ratio. Community and population effects were
determined with counts of macroinvertebrate and zooplankton populations.
Survival of Bti was monitored by pasteurizing samples of water, sediment, and
aquatic fauna at 65°C for 20 minutes, and plating on Tryptic Soy Agar.
Animals were prepared for analysis by rinsing three times in sterile distilled
water. 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 concentrations would be sufficient for
detection. Tadpoles and snails were then homogenized in 5 ml of sterile water
using a Tissumizer (Tekmar Corp., Cincinnati, Ohio). Other organisms were
ground in 5 ml of water using a Ten-Broek manual all-glass tissue grinder.
All ground samples were pasteurized for 20 minutes at 65°C. Cooled samples
were then diluted, placed on TSA and incubated overnight at 30°C.
190
-------
RESULTS AND DISCUSSION
Microcosm Tests
In our first microcosm test, commercially formulated Bti spores
(Vectobac™; Abbott Laboratories) were added to microcosms at initial
densities of 10*, 104 and 106 spores per ml. These spores persisted at levels
near their inoculation densities for the duration of a six-week monitoring
period.
At exposure levels of 104 spores per ml and above, target organisms
(i.e., mosquito larvae and some chironomids) were killed within 24 hours.
Other organisms were not affected. We noted, however, that these non-target
organisms accumulated Bti rapidly following their initial exposure. This Bti
was then lost gradually over the four-week test period.
Although Bti spores persisted in the microcosms, toxicity to target
organisms disappeared. Water and sediment samples bioassayed with first
instar mosquito larvae, Aedes atropalpus or A. aegypti. consistently showed a
complete loss of toxicity between 7 and 14 days post-application.
At high exposure levels (106 spores/ml) microcosms showed brief increases
in primary production and community respiration. Similar brief responses were
noted in other ecosystem variables (pH, production/respiration ratio). We
believe that these effects were caused by the degradation of nutrients
associated with the formulation, rather than direct effects of Bti.
Field Tests
Temporary pools were biologically and physically more variable habitats
than the laboratory microcosms. Despite large fluctuations in water tempera-
ture, water level, water chemistry and photoperiod, these sites supported a
zooplankton community quite similar to that of the microcosms. Both contained
a cladoceran species, a cyclopoid copepod species, and an ostracod species.
The field sites contained a large number of insect species as well as snails
(Lymnaea sp.) and tadpoles of the spring peeper (Hvla crucifer).
In all field sites, Bti spore densities in the sediments increased
rapidly after application. These densities persisted for the duration of the
tests. Water column densities of Bti were variable depending on the formula-
tion used and fluctuations in water level.
Mosquitos in the field sites were killed within 24 hours, while non-
target animals were not affected. Analysis of dead mosquitoes confirmed the
presence of Bti. We were able to collect seven non-target taxa from the pools
with sufficient frequency to measure their uptake of Bti. Three zooplankton
taxa (Daphnia pulex. Cyclops sp., and an ostracod), two insect taxa (Dytiscid
larvae and Hydrophilid adults), a snail (Lvmnaea sp), and spring peeper
tadpoles (Hvla crucifer) were analyzed. Zooplankton and dytiscids picked up
Bti quickly, reaching a peak within 1 hour. Within 24 hours they had lost 90%
of the Bti load. Hydrophilid adults had their highest levels of Bti at 8
hours post-application. None were collected at one hour, so we don't know
191
-------
whether they would have been higher at that time. Snails reached their peak
level at 1 hour, tadpoles at 8 hours. Both of these larger taxa lost Bti more
slowly than either zooplankton or insects. After 120 hours individuals of
each taxon still contained over 5,000 spores per animal. As the ponds dried,
some of the tadpoles metamorphosed into adult frogs. Tests on adult frogs
collected at 120 hours showed they still retained Bti, but densities had
dropped to 50-60 spores per animal.
As in the microcosms, Bti lost its toxicity within 9-12 days post-
application. The loss of toxicity in both microcosms and pools may be related
to either destruction or inactivation of the toxic parasporal protein crystal.
When Bti spores from the test systems were recultured in the laboratory,
crystals were again produced and toxicity returned. No ecosystem level
effects were seen in the field.
In summary, microcosms appear to be a reasonable surrogate for temporary
pools. In the repeated experiments, the survival and effects of Bti in
microcosms showed the same patterns observed in the field. Specifically these
are:
o Target animals were killed in both systems.
o The persistence of toxicity was similar, about 9 to 13 days, in the
microcosms and in the pools. This time is similar to published data.
o Non-target animals were not affected in either system.
o Non-target animals in both systems accumulated Bti through adsorption or
ingestion.
o The extent of the accumulation and rate of loss of Bti was species-
specific.
o There were no effects on ecosystem variables at the 104 spores/ml level
tested in the field.
FUTURE WORK
While the microcosms used in these tests were reasonably good predictors
of the fate and effects of Bti, we feel that they can be improved. The
foremost modification involves the sediment layer, which at present consists
of sand and accumulated detritus. We want to evaluate the use of natural,
more complex sediments. In addition to increasing the physical and chemical
sophistication of the microcosms, this would provide a more complex microbial
community. It might also be used to develop site-specific versions of the
microcosm tailored toward predicting effects at a particular location.
Secondly, we need to further characterize the microbial community of the
microcosms and its role in major ecosystem processes. By expanding our
knowledge of this area, we can add tests to our current monitoring protocol
which would be more sensitive to detecting perturbations at the microbial
192
-------
level. Once this has been accomplished, we will begin a more thorough lab-to-
field comparison.
Finally, we need to test organisms other than spore-formers. The
persistence of Bti spores in our tests was not unexpected, given the stability
of bacterial spores. Testing other microorganisms will permit a more compre-
hensive evaluation of how well these microcosms can predict events in the
field.
PUBLICATIONS
Shannon, L.J, T.E. Flum, R.E. Anderson, and J.O. Yount. 1989. Adaptation of
the mixed flask culture microcosm for testing the survival and effects of
introduced microorganisms. Jn Aquatic Toxicology and Hazard Assessment: 12th
Volume, ASTM STP 1027, U.M. Cowgill and L.R. Williams (eds.). Amer. Soc. for
Testing and Materials, Philadelphia, PA. pp. 224-242.
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
Aquatic Toxicology and Hazard Assessment: 12th Volume, ASTM STP 1027, U.M.
Cowgill and L.R. Williams (eds.). Amer. Soc. for Testing and Materials,
Philadelphia, PA. pp. 189-203.
193
-------
-------
SESSION VIII
RISK CONTROL
195
-------
-------
THE USE OF LETHAL BACTERIAL GENES TO LIMIT SURVIVAL OF
INTENTIONALLY RELEASED GENETICALLY ENGINEERED MICROORGANISMS
Stephen M. Cuskey
U.S. EPA
Environmental Research Laboratory
Gulf Breeze, FL 32561
INTRODUCTION
Concerns about the intentional release of genetically engineered micro-
organisms to the environment may be assuaged if the released strains contain a
mechanism 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 applica-
tion of the lethal signal, i.e., the strains should have a conditionally
lethal phenotype. I have taken two potentially lethal bacterial genes
putatively involved in plasmid maintenance, kilA from plasmid RK2 and hok from
Rl, and altered their regulation so that expression is dependent on the
presence of an added compound. This was accomplished by removing the native
regulatory sequences upstream from the kilA and hok structural genes and
replacing them with the OP2 operator-promoter region from the TOL plasmid,
pWWO, forming plasmids p£PA88 (OP2-directed kilA) and pEPA112 (OP2-directed
hok). Cells containing either of these plasmids and an appropriate regulatory
gene should be killed in the presence of an added OP2 inducer such as benzoic
acid.
In addition to their potential utility in biological containment of
intentionally released genetically engineered microorganisms, lethal genes may
also be used in gene transfer experiments. Normally, gene transfer determin-
ations in the laboratory involve well-character!zed donor and recipient cells.
Conditions are usually set so that donors and recipients are selected against
and the only cells that grow on the selection medium are recipients that have
received DNA from the donors. This is necessary because transfer events,
which may be relatively rare, may be masked by growth of the donor cells which
also contain the selected trait. This is more difficult when dealing with a
heterologous recipient population as is found in environmental samples.
Recipients are often made resistant to a particular antibiotic which is
included in the selection medium, killing the donor cells. This is not
possible, obviously, when using environmental bacteria as recipients.
Inclusion of conditionally lethal constructs in the chromosomes of donor cells
would allow for donor counter-selection by inclusion of a normally innocuous
compound in the selection medium.
I report here on the construction and testing of cells containing
conditionally lethal constructs under a variety of genetic and environmental
conditions in the laboratory. I also report on preliminary attempts to design
and construct conditionally lethal counter-selection transposons for use in
gene transfer experiments and also on attempts to determine the lethal
mechanism of action of kilA which has not been previously reported.
197
-------
RESULTS AND DISCUSSION
Pseudosnonas aeruqinosa strain PA01 cells transformed with plasraid pEPA88
were incubated in a minimal salts solution in the presence and absence of
benzoate and surviving cell numbers were determined over 24 hours. This
strain contains a chromosomal locus (designated benR) which can be substituted
for xvlS in benzoate-dependent action of expression of genes downstream from
the TOL plasmid OP2 promoter region. Therefore, cells transformed with pEPA88
should not survive in benzoate-containing media. Cell number determinations
revealed, however, that cell numbers did not decrease upon incubation in
benzoate minimal medium, neither did they increase (PA01 can utilize benzoate
as a sole source of carbon and nitrogen.) This suggests that, under these
conditions, kilA causes a non-lethal inhibition of growth. Addition of an
alternate carbon source resulted in growth of the culture after a short lag,
supporting this hypothesis. Cells containing either the cloned xvlS or benR
regulatory genes on a multi-copy plasmid in addition to pEPA88 showed a
decrease in numbers of approximately two orders of magnitude upon incubation
in benzoate minimal, showing that kilA was lethal under these conditions and
suggesting that the "lethal" effects of kilA may be dose-dependent.
Microscopic examination of PA01(pEPA88) cultures incubated in benzoate
minimal medium showed the presence of two distinct morphological forms:
numerous long, filamentous as well as single cells forms. The filamentous
cells appeared after about 2 hours of incubation and were not seen in cultures
incubated in benzoate-free minimal medium. The filaments were separated from
the single cells by differential filtration and the ability of each fraction
to take up a radio!abeled substrate was determined. Results, obtained both by
counting radioactivity in each exposed fraction and by microautoradiography,
showed that the single cells were more proficient at uptake than the
filaments, suggesting that the latter cells were less metabolically active.
These results suggest that kilA works by inhibiting daughter cell separation
at cell division and that this may be eventually lethal to the cell. kilA
activity may be dose-dependent, whereby cells receiving a low dose may be
inhibited but not killed.
Plasmid pEPA95 contains the hok gene downstream from the OP2 operator-
promoter region. Escherichia coli cells transformed with this plasmid are
selected at normal rates on solid medium but do not grow on media containing
0.05% benzoate. This suggests that this construct may be useful in donor
counter-selection experiments. The lethal construct was sub-cloned to a
broad-host-range plasmid forming pEPA112 which was transformed to an environ-
mental isolate, Pseudomonas fluorescens PF015 containing the xvlS regulatory
gene on a compatible plasmid. These cells are also sensitive to the presence
of benzoate and may be useful as donors in gene transfer determinations. For
stability, both xvlS and the OP2-directed hok genes were placed on different
transposon derivatives of Tn5 forming plasmidS, pEPA124 and pEPA134, which are
currently being inserted into the chromosomes of various bacterial strains.
198
-------
ANTICIPATED FUTURE WORK
In the future, I plan to investigate the apparent dose-dependent lethal-
ity of kilA and further investigate its mechanism of action. In addition,
OP2-directed expression of both kilA and hok is inefficient in that numerous
survivors of benzoate addition are seen. I plan to utilize other promoters
and regulatory genes in construction of biological control cassettes as well
as other lethal genes to maximize lethality in released strains. I also plan
to test various constructs inserted into the chromosomes of potential donors
in gene transfer determinations to test this strategy in donor counter-
selection when using a heterologous recipient population.
PUBLICATIONS
Cuskey, S.M., and A.B. Sprenkle. 1988. Benzoate-dependent induction from the
02 operator-promoter region of the TOL plasmid (pWWO) in the absence of known
plasmid regulatory genes. J. Bacteriol. 170:3742-3746.
Cuskey, S.M., W.H. Jeffrey, P.J. Chapman, and R.H. Olsen. 1990. Cloning and
characterization of Pseudomonas genes involved in benzoate catabolism:
isolation of a chromosomal DNA fragment able to substitute for xvlS in
activation of the TOL lower pathway promoter. J. Bacteriol. (submitted).
Cuskey, S.M. 1990. Lethal genes in biological containment of released
microorganisms. In Environmental release of genetically engineered and other
microorganisms, M.J. Day and J.C. Fry (eds.). Edward Arnold (submitted).
Cuskey, S.M. 1990. Biological containment of genetically-engineered micro-
organisms. In Guide to environmental microbiology, M. Levin, R. Seidler, and
P. Pritchard (eds.). McGraw-Hill (in press).
Zylstra, G., S.M. Cuskey, and R.H. Olsen. 1990. Construction of plasmids for
use in risk assessment research. In Guide to environmental microbiology, M.
Levin, R. Seidler, and P. Pritchard (eds.). McGraw-Hill (in press).
199
-------
PHYSICAL AND CHEMICAL CONTROL OF RELEASED MICROORGANISMS
K. Donegan,1 R. Seidler,2 and C. Matyac1
NSI Technology Services Corporation1
U.S. EPA2
Corvallis, OR 97333
INTRODUCTION
An important consideration of the environmental release of a genetically
engineered microorganism (GEM) is the capability for reduction or elimination
of GEM populations once their function has been completed or if adverse
environmental effects are observed. Decontamination methods employed in past
releases of GEMs have included burning or tilling, or removal from the field
and subsequent autoclaving, of sprayed plants, and the application of
biocides. In this study the treatments of burning and biocides, alone and in
combination with tilling, were evaluated.
METHODS
Twenty-four plots of Blue Lake bush beans, 1.5 x 3 m in dimension and
separated by 1.25 m, were sprayed with 1 x 108 cfu/ml of Erwinia herbicola
(resistant to nalidixic acid). After 5 days one of 8 decontamination methods
was applied to each plot for a total of 3 plot replicates per treatment. The
treatments used were: (1) control, (2) control + till, (3) burn, (4) burn +
till, (5) Kocide (cupric -vdroxide), (6) Kocide + till, (7) Agri-strep
(streptomycin sulfate), and (8) Agri-strep + till. The till treatment was
applied with a rototiller, the burn treatment involved propane torching of the
leaves and removal of the remainder of the plant, and the two biocides, Kocide
and Agri-strep, were applied at the recommended field rate of 100 ppm until
runoff. Plots were sampled 4, 7, 11, 18, 21 25, and 33 days after application
of the bacteria to determine the effect of the decontamination treatments on
bacterial survival. Plots with the control, Kocide, and Agri-strep treatments
were sampled by the collection of 3 leaves/sample with triplicate samples.
Leaf samples were processed for 1 minute in 20 ml 0.01M phosphate buffer in a
stomacher-blender, diluted in phosphate buffer, plated on LB + nalidixic acid
+ cycloheximide media and incubated overnight at 30°C. The control + till,
Kocide + till, and Agri-strep + till plots were sampled by the collection of 5
soil cores/plot which were mixed and sub-sampled in triplicate by the removal
of ca. 5 g of leaf material. These leaf samples were then processed as
described above for the non-tilled treatments. Five soil core samples/plot
were also collected for the burn and burn + till treatments. Because the
plots with the burn and burn + till treatments contained only ashes, leaf
material could not be sub-sampled, and the 3 sjib-samples instead consisted of
10 g of soil. These soil samples were processed by 5 minutes of shaking in 95
ml of 0.01M phosphate buffer and plated identically to the other treatments.
Plate counts were converted to log cfu/g and analyzed with the SAS statistical
program for blocking effects and best fit of regression lines. T-tests of
slopes from regression lines were performed to compare treatments to the
control and to other treatments.
200
-------
RESULTS AND DISCUSSION
The burn, burn + till, control + till, kocide + till, and streptomycin +
till treatments all differed significantly from the control treatment. The
burn and burn + till treatments, however, were the only treatments which
significantly reduced bacterial populations whereas all other treatments with
tilling actually produced an increase in the number of bacteria.
Burning, alone or in combination with tilling, produced a dramatic and
persistent reduction in bacterial populations.
Tilling, alone or in combination with the biocide treatments, stimulated
a significant increase in bacterial populations which persisted for several
weeks. Most likely the decay of the tilled plants produced an increase in
available nutrients which promoted growth of the saprophytic £. herbicola.
The biocides, Kocide and Agri-strep, were ineffective in producing a
decline in bacterial numbers which significantly differed from that of the
control treatment. Although these compounds have been successfully used for
control of several plant pathogens, in this experiment, probably due to the
concentration used or the species of the target microorganism, their
bactericidal activity was minimal.
The results of this study indicate that decontamination methods currently
proposed for field releases of GEMs, such as tilling and application of
biocides, may not be effective for all microbial species or with all crops,
and that the development of new methods, or the modification of present
methods, may be necessary.
ANTICIPATED FUTURE WORK
The effects of these treatments on other species of bacteria and on fungi
will be evaluated. Different application procedures and chemical dosage rates
of the methods will also be explored.
201
-------
-------
SESSION IX
HUMAN HEALTH
203
-------
-------
DEVELOPING A BIOTECHNOLOGY HEALTH EFFECTS
PROGRAM FOR THE TWENTY-FIRST CENTURY
L.D. Claxton, C. Kawanishi, and S.E. George
Health Effects Research Laboratory
U.S. EPA
Research Triangle Park, NC 27711
Large scale environmental releases of biotechnology products are pre-
dicted in the near future. Although the greatest risks from these activities
are anticipated to be ecological effects, the Environmental Protection Agency
is responsible for examining the potential human health hazard of these
activities. While these are not expected to be likely, because known human
pathogens will not be approved for use, it is reasonable to anticipate that a
finite risk does exist. The Office of Health Research (OHR) in the Office of
Research and Development (ORD) is working to improve EPA's capabilities to
protect against this possibility by thoroughly identifying possible risks,
prioritizing them into those thought to be significant and those thought to be
trivial, and developing recommendations for research and risk assessment based
on this analysis. A health subcommittee of the Biotechnology Science Advisory
Committee (BSAC) has been formed to provide guidance and feedback on the
recommendations. Once developed by OHR, the plan will be reviewed by the
Science Advisory Board (SAB) and subsequently revised as appropriate and
implemented.
As background to the development of the research strategy, a survey of
past and current efforts to systematically define health risks associated with
the environmental release of biotechnology products was conducted. It is
reported in a document entitled "A Survey of Research Efforts Associated with
the Health Effects of Microorganisms Used in the Field of Biotechnology." The
goal was to identify the extent to which the potential health risks had been
identified and resolved through work done by EPA and other federal agencies,
academia, and organizations. The report demonstrates that there are some
ongoing efforts by EPA (e.g., microbial pesticide control agent research) and
by others on food, drugs, agricultural and energy products both in the United
States (i.e., the NIH Office of Recombinant DNA Activities, Food and Drug
Administration, National Institute of Occupational Safety and Health, Depart-
ment of Defense, Department of Energy, and Department of Agriculture) and
abroad (i.e., Canada, United Kingdom, European Communities, Organization for
Economic Cooperation and Development, and World Health Organization).
The document examines both the scientific and regulatory aspects of the
issue. The analysis reveals that studies involving biotechnology-related
organisms have focused on ecological, food, drug, pollution control, and other
areas not directly related to detecting and evaluating human health effects.
In part this stems from the recognition that the introduction of non-
indigenous organisms into novel ecosystems can result in serious adverse
ecological consequences. EPA and other groups associated with the environ-
mental release of biotechnology products, therefore, focus their efforts on
ecological effects rather than on health effects. However, our analysis does
not lead us to the conclusion that health effects should be ignored. In the
205
-------
unlikely event that an environmental release did cause human disease, signifi-
cant adverse impacts would occur. The Microbial Pest Control Agent Sub-
division M Testing Guidelines describe data that must be collected for EPA
analysis prior to approval for pesticidal use. While the test requirements
provide for the evaluation of the proposed MPCAs under "worst case" scenarios
in mammalian species, many uncertainties are associated with these tests and
their ability to measure parameters important for human pathogenicity or
toxicity. EPA needs a systematic analysis of the issues surrounding health
risk from environmental release of biotechnology products as a basis for
decision-making. The report, therefore, provides the state-of-the-science and
identifies relevant issues.
Although biotechnology health research needs can be grouped in several
ways, the Survey placed research needs into 4 major categories:
1. infectivity/pathogenicity questions,
2. allergic reaction issues,
3. toxicity, and
4. moaulation effects.
Because of the regulatory role of the EPA which attempts to anticipate and
prevent adverse health effects through testing, a fifth category must be added
to the four identified in the survey. This category, "regulatory protocol
development and validation," is a logical extension of all four basic research
categories.
Table 1 illustrates, to some degree, how identified research programs
complement each other. Some programs (e.g., U.S. EPA) havt targeted the
examination of microorganisms used by biotechnology industries. Other
programs are more interested in basic research associated with the various
health effects categories. For example, programs within the NIH are inter-
ested in developing basic knowledge about the mechanisms of infectivity and
pathogenicity of known pathogens. The extremely useful information developed
in these NIH programs will be useful to EPA scientists who develop pathogenic-
ity screening tests for engineered organisms. NIH scientists, however, are
unlikely to examine their methods and information for each newly
developed microorganism. By examining the Survey, it is apparent that EPA
must play a role in each of these research categories.
A companion document that is now being prepared will contain the prelim-
inary proposal of OHR's research strategy and identify the issues we think
should be studied through our health research program at EPA. The presentation
will provide illustrations of the type of efforts now contemplated.
206
-------
Table 1. Research Roles and Activities by Organizations Within the United
States That Examine the Health Effects of Microorganisms
U.S. Health
Research Infectivity/
Proqrams1 PathoqenicitY
U.S. EPA
NIOSH
USDA
NIH
DOD
DOE
Non-government
O2'*
P*
M*
0
0*
U
0/P*
Allergic
Responses
P*
0/P*
M*
0
7
U
0/P*
Toxicitv
0*
0/P*
?
6
0/P*
p*
0/P*
Nodulation
Effects
0*
?
?
6/M
?
U
0/P*
Regulatory
Protocols
0*
0*
U
U
U
U
U
1 Organizational abbreviations are as follows: NIOSH, National Institute of
Occupational Safety and Health; USDA, U.S. Department of Agriculture; NIH,
National Institutes of Health; DOD, Department of Defense; DOE, Department of
Energy.
2 0 - Ongoing efforts; P - Potential research area; U - Unlikely research
area; ? » Unknown; M » Minor or occasional efforts.
* * Associated with microorganisms used for Biotechnology purposes.
PUBLICATIONS
Chadwick, R.E., S.E. George, J. Chang, M.J. Kohan, J.P. Dekker, J.E. Long, and
M.C. Duffy. Comparative enzyme activity and activation of the promutagen 2,6-
dinitrotoluene in male C011 mice and Fischer 344 rats. Cancer Lett.
(submitted).
Chadwick, R.W., S.E. George, J. Chang, M.J. Kohan, J.P. Dekker, J.E. Long, and
M.C. Duffy. Altered GI enzyme activity in CD11 mice and F1344 rats after 2
and 4 weeks of pentachlorophenol treatment. Submitted for presentation at
Society for Toxicology Meeting, February 1990.
Claxton, L.D., S.E. George, and C.Y. Kawanishi. 1988. The health effects
research laboratory biotechnology research program: An overview summary.
Proceedings of the Biotechnology Risk Assessment All Investigators Research
Review (in press).
Claxton, L., S.E. George, D. DeMarini, and J. iewtas. 1985. Biotechnology
health effects research. ORD Innovative Research Highlights September 1985:p.
3.
Claxton, L.D., S.E. George, and C.Y. Kawanishi. The Health Effects Research
Laboratory's Biotechnology Program: A Background Document. In review by the
Health Subcommittee of the Biotechnology Science Advisory Committee.
207
-------
Dekker, O.P., R.W. Williams, G.R. Lambert, S.E. George, M.J. Kohan, and R.W.
Chadwick. Bioassay directed fractionation of the urinary metabolites from
Fischer 344 rats treated with 2,6-dinitrotoluene and pentachlorophenol.
Submitted for presentation at Society for Toxicology Meeting, February 1990.
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 metabo-
lites: A comparison of GI enzyme activities and urine mutagenicity in CD11
mice (in preparation).
George, S.E., R.W. Chadwick, M.J. Kohan, and J.P. Dekker. 1989. Effect of
pentachlorophenol on 2,6-dinitrotoluene-induced urine mutagenicity and
influence on intestinal enzymes in the male CD11 mouse. Seventh Annual
Meeting Program and Abstracts, Genotoxicity and Environmental Mutagen Society,
Raleigh, NC, p. 20.
George, S.E., R.W. Chadwick, M.J. Hohan, and J.P. Dekker. Potentiation of
2,6-dinitrotoluene genotoxicity by pentachlorophenol in CD11 mice and Fischer
344 rats. Submitted for presentation at Society for Toxicology Meeting,
February 1990.
George, S.E., and M.J. Kohan. Antibiotic-induced alterations in the
intestinal microbiota of the CD-I male mouse. Lab. Anim. Sci. (submitted).
George, S.E., M.J. Kohan, D.B. Walsh, and L.D. Claxton. 1987. Methodology
for evaluating potential human health effects of microorganisms that degrade
hazardous wastes. Proceedings Third Annual Symposium on Solid Waste Testing
and Quality Assurance 1:1751192.
George, S.E., M.J. Kohan, D.B. Walsh, and L.D. Claxton. 1989. Acute coloni-
zation study of polychlorinated biphenyl degrading pseudomonads in the mouse
intestinal tract: comparison of single and multiple exposures. Environ.
Toxicol. Chem. 8:123-130.
George, S.E., M.J. Kohan, D. B. Walsh, G.M. Nelson, D.A. Whitehouse, and L.D.
Claxton. 1988. In vivo and in vitro intestinal survival and competition of
environmental Pseudomonas species. Proceedings of the Biotechnology Risk
Assessment All Investigators Research Review (in press).
George, S.E., M.J. Kohan, D.B. Walsh, A.G. Stead, and L.D. Claxton. 1989.
Polychlorinated biphenyl degrading pseudomonads: Survival in the mouse
intestines and competition with the normal microbial flora. J. Toxicol.
Environ. Health. 26:19-37.
George, S.E., M.J. Kohan, D.W. Whitehouse, and. L.D. Claxton. Antibiotic
influence on environmental Pseudomonas spp. survival in the intestinal tract
and translocation to the spleen and liver (in preparation).
George, S.E., M.J. Kohan, D.B. Walsh, and L.D. Claxton. 1987. Survival and
competition of polychlorinated biphenyl degrading pseudomonads in the mouse
intestines. Abst. Ann. Meet. Amer. Soc. Microbiol. 87:284.
208
-------
George, S.E., M.J. Kohan, D.B. Walsh, and L.O. Claxton. 1987. Methodology
for evaluating potential human health effects of microorganisms that degrade
hazardous wastes. World Conference on Hazardous Wastes, Budapest, Hungary.
#367.
George, S.E., M.J. Kohan, D.B. Walsh, and L.D. Claxton. 1988. Single and
multiple dose exposure models for evaluating the survivability of genetically
altered bacteria among intestinal flora. First International Conference on
the Release of Genetically Engineered Microorganisms, Cardiff, Wales, UK, p.
X / •
George, S.E., D.B. Walsh, and L.D. Claxton. 1988. Influence of ampicillin on
the survival and competition potential of polychlorinated biphenyl degrading
pseudomonads in the mouse intestine. Abst. Ann. Meet. Amer. Soc. Microbiol.
88:299.
George, S.E., D.B. Walsh, A.G. Stead, and L.D. Claxton. Effect of ampicillin
treatment on the survival and competition of polychlorinated biphenyl degrad-
ing pseudomonads in mouse intestines. Fund. Appl. Toxicol. (in press).
George, S.E., D.A. Whitehouse, and L.D. Claxton. 1989. Genotoxicity of 3,5-
dichlorobenzoate biodegradation metabolites. Abst. Ann. Meet. Amer. Soc.
Microbiol. 89:358.
Kohan, M.J., S.E. George, M.H. George, J.E. Gallagher, and R.W. Chadwick. DMA
adduct formation in CD11 mice and Fischer 344 rats treated with 2,6-dinitro-
toluene and pentachlorophenol. Submitted for presentation at Society for
Toxicology Meeting, February 1990.
Nelson, G., S. George, and L. Claxton. Use of in vitro survival and competi-
tion methods to evaluate toxicological potentials of environmentally applied
pseudomonads (in preparation).
Nelson, G., S. George, and L. Claxton. 1989. Use of in vitro survival and
competition methods to evaluate toxicological potentials of environmentally
applied pseudomonads. Abst. Ann. Meet. Amer. Soc. Microbiol. 89:352.
209
-------
ENVIRONMENTAL PSEUDOMONAD SURVIVAL IN THE GI TRACT AND TRANSLOCATION
TO THE SPLEEN AND LIVER IN ANTIBIOTIC-TREATED MICE
S. Elizabeth George, Michael J. Kohan,
Douglas A. Whitehouse, and Larry D. Claxton
Health Effects Research Laboratory
Genetic Toxicology Division
U.S. EPA
Research Triangle Park, NC 27711
INTRODUCTION
The deliberate release of microorganisms into the environment has
prompted questions about human health. In order to investigate the potential
of environmental strains to cause adverse health effects, methods development
is under way to evaluate these strains in an in vivo rodent model. Initially,
the effect of environmentally released microorganism survival in the GI tract
and competition with the intestinal microbiota has been investigated. If the
normal microbiota of the intestinal tract are altered, harbored pathogens can
multiply and express their pathogenic traits or potential pathogens can
colonize. Antibiotic treatment of mice has been shown to increase the
colonization potential of Pseudomonas spp. Promotion of a systemic infection
by translocation from the intestinal tract to the mesenteric lymph nodes,
spleen, and liver is also associated with antibiotic therapy.
In this study, a method is described that evaluates the effects of
several antibiotics on the survival and translocation of Pseudomonas strains
that were isolated from a commercial environmental product. The micro-
organisms are similar to those that may be released into the environment in
high concentrations.
METHODS
All chemicals used in this study were of reagent grade and obtained
commercially. The bacterial strains, P. aeruqinosa strain BC16 and P.
maltophilia strain BC6, were isolated as described previously from a commer-
cial product. P. aeruqinosa strain PAMG was isolated from a mouse intestinal
homogenate. Strain PAMG was designated as the "positive control" due to its
intestinal origin. Pseudomonad enumeration was done on Pseudomonas isolation
agar (PIA) (Difco, Detroit, MI) supplemented with 30 ug/ml HgCl2 (P.
aeruqinosa strain BC16), 50 ug/ml kanamycin (P. maltophilia strain BC6), or
both antimicrobial agents (P. aeruqinosa strain PAMG).
Sixty-day-old strain CD11 male mice (Charles River Laboratories,
Kingston, Stone Ridge, NY) were dosed by gavage individually with 1 mg of each
antibiotic twice daily for 3 days. On day 4, animals were gavaged indi-
vidually with 109 CPU of each microorganism and antibiotic treatment (1 mg)
continued daily for the duration of the experiment. At 3, 24, 48, 72, and 120
hours after the bacterial dose, animals were sacrificed, the intestines
removed and homogenized, and microorganisms enumerated as described pre-
viously.
210
-------
Microbial resistance to clindamycin, kanamycln, rifampicin, and spectino-
mycin was determined to better understand the effect of antibiotic resistance
on intestinal survival. P. maltophilia strain BC6 was resistant to the four
antibiotics at 10 and 50 ug/ml. Strain BC16 was resistant to rifampicin (10
ug/ml only), clindamycin, and spectinomycin at both concentrations. However,
strain BC16 was sensitive to kanamycin and 50 ug/ml rifampicin. P. aeruqinosa
strain PAM6 had an identical resistance profile as strain BC16 except that it
was resistant to kanamycin.
RESULTS AND DISCUSSION
P. aeruqinosa strain BC16 was detectable five days after dosing in
clindamycin-treated and untreated mice. Kanamycin, rifampicin, and spectino-
mycin suppressed survival. Strain BC6, which is resistant to rifampicin, was
recovered after 5 days from the GI tract of rifampicin-treated mice. Strain
BC6 did not survive the experimental period in any of the other treatment
groups even though it was resistant to all four antibiotics. On the other
hand, the mouse intestinal isolate, strain PAMG was present in the GI tract
from animals in all antibiotic treatment groups at the end of five days.
Translocation to the spleen was observed only in spectinomycin-treated
mice. Strain BC6 was recovered three hours after dosing and strain PAMG was
present at 72 hours. P. aeruqinosa translocation to the spleen has been
observed by other laboratories. Hentges et al. reported translocation of a
clinical P. aeruqinosa strain to the spleen in streptomycin-treated mice one
week after dosing with 108 CPU. The kanamycin results supported those of
Hentges et al. who found no translocation to the spleen in kanamycin-treated
animals.
Translocation to the liver occurred in all antibiotic treatment groups
dependent on the dosed strain. £. aeruqinosa strains BC16 and PAMG trans-
located to livers in animals treated with all antibiotics. Strain BC6 was
detectable in the livers of clindamycin-treated and untreated mice. There was
no direct relationship between antibiotic-resistance and translocation.
However, clindamycin, which eliminates the predominant obligately anaerobic
Gram-negative rods and lactobacilli in the GI tract of CD-I mice (George,
unpublished data) promoted translocation more frequently. This observation
coincides with the findings of Van Furth and Guiot. They observed that
reduction in anaerobes, more so than facultative species, enhances the
introduction of "foreign" microorganisms in the GI tract. Therefore, because
the normal flora are altered, the potential for translocation is increased.
FUTURE WORK
Several avenues of future research have been proposed. The effects of
different antibiotics on GI tract survival and translocation will be continued
on individual strains as well as product mixtures. Health effects due to the
pulmonary exposure to these microorganisms will be investigated. Another area
of research includes human exposure monitoring methods. Finally, the effect
due to exposure to the environmentally released microorganisms and their
xenobiotic metabolites will be examined.
211
-------
AN In Vitro TOXICOLOGICAL SCREEN TO EVALUATE POTENTIAL
ENVIRONMENTAL PSEUDOMONAD HEALTH EFFECTS
Gail M. Nelson,1 S. Elizabeth George,2 and Larry D. Claxton2
Environmental Health Research and Testing1
Health Effects Research Laboratory2
Genetic Toxicology Division
U.S. EPA
Research Triangle Park, NC 27711
INTRODUCTION
Alteration of the normal intestinal microbiota can result in disease.
Previous in vivo work in this laboratory examined the colonization and
competition potential of Pseudomonas spp. isolated from a commercial product
for PCS degradation with the intestinal flora of CD11 mice. The objective of
this study is to develop an in vitro system to evaluate the colonization and
competition capabilities of environmental strains with the intestinal micro-
biota. J_n vitro results obtained from both rodent and human systems then can
be used together with rodent in vivo results in a parallelogram approach for
extrapolation of potential human in vivo health effects.
METHODS
The environmental strains used in this study are the same as those
studied previously in vivo and include one P. maltophilia strain (BC6) and
three P. aeruqinosa strains (BC16, BC17, and BC18) isolated from a commercial
environmental product. In addition, one P. aeruqinosa strain isolated from a
mouse intestinal homogenate (PAMG) and two Escherichia coli control strains
(23559 and 39566) were tested. Freshly voided fecal material was obtained
from male CV-1 mice. Veal infusion broth enriched with yeast extract, vitamin
K1 all hemin served as the in vitro culture medium.
A tier testing approach was used. Initially, competitor strains were
tested for their ability to survive in anaerobic pure culture. Next, their
survival was monitored in the presence of mouse fecal microbiota. Five-
culture serial transfer experiments then were performed to monitor the
survival and competition potential of each strain with the fecal microbiota.
Fecal microbiota populations were monitored at pre-selected time intervals in
the first, third, and fifth serial transfer cultures. Enumerated flora
include the total aerobic/facultative and total anaerobic/facultative micro-
organisms, the obligately anaerobic Gram-negative rods, the lactose
fermentors, and the anaerobic and facultative Gram-positive microorganisms.
Finally, one competitor strain was tested for survival and competition With
the fecal microbiota in a long-term culture, or anaerobic microcosm.
RESULTS AND DISCUSSION
All strains tested survived for 48 hr in anaerobic pure culture.
Pseudomonad strains maintained their inoculum levels for at least 12 hr, and
E. coli strains exhibited growth.
212
-------
The presence of mouse fecal microbiota did not greatly affect the
persistence of the competitor strains in vitro. Those strains that maintained
their population levels for at least 12 hr in pure culture also maintained
their numbers in the presence of the fecal microbiota. £. coli strain 39566
did not exhibit growth in mixed culture, but did grow in pure culture.
E. coli strains survived serial transfer through five cultures, although
in declining numbers. Environmental Pseudomonas spp. survived transfer
through three cultures but were not detected after four transfers. The mouse
intestinal isolate, P. aeruqinosa PAMG, was detected in culture 5 in one of
two experiments. Strain PAMG was the only competitor strain to significantly
alter the normal microbiota populations, causing a decrease in the obligately
anaerobic Gram-negative rods.
P. aeruqinosa strain BC16 was the only competitor tested in the anaerobic
microcosm portion of the experiment. Results were similar to those obtained
with serial transfer experiments. Strain BC16 declined in number after its
introduction into the microcosm and was not detectable after 72 hr.
In conclusion, the environmental strains tested were poor competitors
with the mouse fecal microbiota. E. coli is a normal constituent of the mouse
fecal flora, and as expected, the f. coli control strains survived longer in
serial transfer than the environmental strains. The mouse intestinal isolate,
strain PAMG also survived longer than the environmental strains. This is in
agreement with previous in vivo results and indicates that strain PAMG is
better adapted for intestinal tract survival than the environmental
pseudomonads.
FUTURE WORK
Future research directions include further testing of the environmental
strains in long-term continuous culture (anaerobic microcosm) studies,
complementary studies using human microbiota, and examination of the effects
of antibiotics in vitro.
213
-------
HEALTH EFFECTS ASSESSMENT OF PULMONARY EXPOSURE TO BIOTECHNOLOGY AGENTS
C.Y. Kawanishi,1 S.E. George,1 and R.L. Sherwood2
Health Effects Research Laboratory1
U.S. EPA
Research Triangle Park, NC 27711
I IT Research Institute2
Chicago, IL 60616
INTRODUCTION
A primary route of human exposure to many biotechnology agents, both
naturally occurring and genetically altered, is pulmonary. While in the
majority of cases the interactions are expected to be innocuous, the potential
for detrimental health effects must be assessed. "Ms is s. :ifically the
case with biotechnology agents that are broadcast to or a. ied on a large
scale in the environment. Thus, one of the data r.^uiremenij for preregis-
tration testing of microbial pesticides is the Acute Pulmonary Toxicity/Patho-
genicity Test to assess the human health effects potential of the agent. This
alternative to the Acute Aerosol Inhalation Test utilizes intranasal (IN) or
intratracheal (IT) challenge with the agent. This test was recently developed
and standardized utilizing bacterial and viral pathogens of laboratory
animals. During the evaluation of these protocols, a bacterial agent,
Bacillus thuringiensis (Bt), utilized in several EPA-registered microbial
pesticides, was used as a negative control. Unexpected mortality was observed
among these CD1 mice challenged with sporulating preparations. Presented are
the results of research on the mechanism of the observed pathogenicity.
Additionally, while these tests were developed mainly for naturally occurring
microbial pesticidal agents, their preliminary application to a genetically
altered agents is described.
METHODS
For IN and IT challenge Bt was recovered from frozen stock, grown on
trypticase soy agar, harvested, quantified, and diluted to the desired
concentrations in 0.1% peptone. Heat-killed preparations were obtained by
autoclaving.
Challenged mice were placed in a glass desiccator jar and lightly
anesthetized with methoxyflurane. For IN a 0.05 ml of a bacterial suspension
was instilled into the nostrils with a 0.5-ml syringe with a 26 gauge,
3/8-inch needle. For IT the preparation were injected at the bottom of the
trachea. Immediately after dosing, lungs were removed from 2 treated
mice/dosage groups and homogenized and plated on trypticase soy agar to
quantify dose. Remaining mice were observed for 14 days for clinical signs
and mortality. Suspensions of 0.77, 1.43 and 3.14 urn diameter latex beads
were prepared and applied in a similar manner to assess the role of pulmonary
blockage in the death of animals treated by IN or IT. The largest diameter
beads, 3.14 urn have volumes and surface areas comparable to the vegetative
cells of Bt.
214
-------
RESULTS AND DISCUSSION
Experimental results demonstrate that latex beads given IN do not cause
mortality until doses approach 5X109 particles per animal. Pulmonary
blockage, therefore, appears not to contribute to mortality caused by IN
challenge with lower doses of Bt. The rapidity of the onset of mortality (1-2
days) after IN and IT challenge appear to eliminate infectivity as the cause
of death. Experiments also suggest that mortality is associated with the
vegetative cells rather than the spores of the bacilli. The mortality can be
prevented by autoclaving bacterial preparations. The three subspecies of Bt
tested, kurstaki. israelensis and aizawai. were all capable of causing
mortality by IN challenge, fi. subtilis. however, did not affect GDI mice at
similar concentrations. The weight of evidence indicates, therefore, that
mortality resulting from IN and IT challenge with preparations of Bt is due to
a specific heat-labile factor associated with the vegetative cells.
Utilization of the IN protocol with similar concentrations of a genetic-
ally altered strain of Pseudomonas aeruginosa. BC16, developed for use in
bioremediation, did not result in mortality. However, challenged mice showed
transient rough coats for several days possibly indicating endotoxin (IPS)
effects.
These results demonstrate the general utility of the IN and IT challenge
methods for investigating the mechanism of pulmonary effects of biotechnology
agents. Future studies will characterize the protocols to further enhance the
utility of these methods in other animal species and with other types of
biotechnology agents such as fungi. The methods will also be specifically
adapted to address questions pertaining specifically to the health effects of
genetically altered agents.
PUBLICATIONS
Sherwood, R.L., Mega, W.M., Kawanishi, C.Y., and Sjoblad, R. 1989. Microbial
size and concentration effects on the pulmonary toxicity/pathogenicity test
for microbial pesticides. Abstract for the American Society of Microbiology,
New Orleans, LA.
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.
215
-------
CELL CULTURE TESTS FOR VIRAL BIOTECHNOLOGY AGENTS
P. Hartig and M. Cardon
NSI Technology Services Corporation
Research Triangle Park, NC 27709
INTRODUCTION
Safety considerations regarding insect viruses and their use in pest
control have been reviewed and have been the subject of several symposia (ASM
1975 ISBN 0-914826-OZ-Z, EPA 1978 EPA-60Q/9-78-026). Based on these assess-
ments and specifically on the recommendations of an American Institute for
Biological Sciences panel, the "Pesticide Assessment Guidelines Subdivision M:
Microbial Pesticides" were promulgated in 1978 and revised in 1989. These
tests are organized on a tier progression approach. Among the tier one tasts
are mammalian cell culture tests for the assessment of the potential effects
of viral pesticidal agents on human health. Jji vitro testing was designed to
assess potential for: a) Overt infection, b) Persistent, latent, or abortive
infections, c) Transformation, and d) Toxicity. This abstracted report
provides a detailed description of the methodology used in assessing toxicity.
The 1978 Subdivision M guidelines suggest various methods for the
assessment of the potential deleterious effects of viral pesticides. However,
none of these methods has been validated for use in the assessment of the
possible effects invertebrate viruses may have on human cells. The questions
regarding the appropriateness of the current testing procedures were
compounded by the criteria for successful completion of any. given test. The
methods suggested in the guidelines required the test virus have no effect in
the test system. It was assumed a virus which produced no effect in the test
system was safe. However, the absence of visible viral induced effects could
simply result from the improper running of the test procedure. Thus, it
became necessary to establish detailed and validated testing protocols to
guard against such an event. The procedures described here have been vali-
dated utilizing agents known to produce deleterious effects in human cells.
They were subsequently employed to evaluate an insect virus (AcNPV) of great
scientific and industrial interest for its potential effects on human cells.
Excerpts of these studies are contained within this document. In total, these
studies validate test procedures which can be performed to comply with
Subdivision M, as well as to provide information to support the risk assess-
ment of the virus (AcNPV) which has received considerable attention
domestically and which has already been genetically engineered and deliber-
ately released abroad.
METHODS
American Type Culture Collection (Rockville, MD) provided Human Foreskin
Fibroblasts, Hs 27 (CRL 1643), WI38 (CCL Z5), CV11 (CCL70) and A6 Xenoous
laevis cells (CCL 102). Autograoha californica nuclear polyhedrosis virus
(AcNPV) isolate E2 and SF21 cells were a gift from Dr. Max Summers (Texas ASM
University, College Station, TX). AcNPV was plaque purified prior to use.
Vero cells were supplied by Dr. Bruce Casto (Environmental Health Research and
216
-------
Testing, RTF, NC). Primate cell lines were maintained in EMEM (Eagle's
Minimum Essential Medium with Earle's Salts, 10% fetal bovine serum (FBS),
Penicillin 10,000 U/dl, kanamycin 0.01 gm%). Cultures were maintained at
37°C, 5% CO,, and 100% humidity. A6 cells were maintained as described by J.
Brusca et al. SF21 cells were maintained in TMN-FH (K C Biological Inc.,
Lenexa, KS), 10% FBS, incubated at 27°C, 100% humidity. AcNPV was produced by
inoculating SF21 cells in exponential growth at an MOI of 5. Seventy two hr
post-infection cultures were clarified by centrifugation and the supernatants
stored at 4°C. Plaque assays were performed on SF21 cells using a 1.5%
agarose overlay. Inactivated AcNPV was prepared by adding 4'-aminomethyltri-
oxsalen-HCl (Lee Biomolecular Research Lab. Inc., San Diego, CA), 10 ug/ml to
AcNPV preparations then exposing virus to 360-380 nM light by setting dishes
on a standard UV transilluminator, 6 min., 4°C, gentle rocking. AcNPV titer
was reduced 3 logs by this treatment. Control samples of SF21 conditioned
TNM-FH were also treated. Gradient purified AcNPV was prepared by pelleting
clarified AcNPV preparations, 250,000 x g, 30 min, 4P, resuspending virus in
gradient buffer (0.1M Tris, 0.001M EDTA, pH 6.4) and isolating the visible
band from quasi-equilibrium 30-60% W/W sucrose gradients (sw 27 rotor, 2 hr,
25%, 4°C). Control "bands" were harvested from similar areas of virus free
gradients. Gradient purified virus, control "bands", and aliquots of
unpurified AcNPV were dialyzed, 2 x 24 hr, against 2000 x their volume with
Hank's buffered salt solution.
Toxicity assay: Two hundred cells were seeded per 60 mm dish. Twenty
four hr post seeding media was removed and replaced with 0.5 ml test fluid,
incubated under conditions appropriate for cell growth, rocked every 15 min,
for 1 hr, 5 ml medium added, incubated 12 days. Cultures were rinsed with
methanol, stained with 1% crystal violet, rinsed, dried, and colonies counted.
All comparisons utilized the Student's T test.
Inhibition of Reproductive Survivability (IRS) assays were undertaken to
define the toxic effects of AcNPV. Individual cells were exposed to AcNPV and
their ability to form colonies subsequent to that exposure was evaluated.
Care was taken to insure that the highest concentration of virus contained as
little invertebrate cell culture medium and cell debris as possible (e.g.,
minimum dilutions in EMEM were 1:5). Control cultures were inoculated with
EMEM alone, or EMEM containing the same concentration of SF21 conditioned
medium as in the virus treatment. Psoralen inactivated virus was also
evaluated for toxicity. Concentration of inactivated virus was determined by
PFU prior to inactivation. Dilutions of psoralen treated SF21 conditioned
medium were included to quantitate the toxic effect of psoralen alone and to
determine if psoralen can act synergistically with SF21 metabolites to produce
toxicity. High concentrations (2 x 106 PFU/dish) of ACNPV proved cytotoxic
to all three primate cell lines and slightly toxic to A6 cells. AcNPV
significantly reduced colony forming efficiencies (CFE) of CV-1, WI38, and HF
cells. Relative to controls, CFE's were 1.2, 3.6, and 1.3%, respectively.
AcNPV reduced CFE of A6 cells to 70% of controls. The toxic effects of AcNPV
were reduced with dose. HF cells were less sensitive to AcNPV than were WI38
and CV-1 cells. WI38 and CV-1 cells were significantly effected by as little
as 2 x 104 or 2 x 105 PFU/dish, respectively. Surprisingly, psoralen-
inactivated AcNPV produced greater toxicity than the stock virus regardless of
the cell line tested. High concentrations of conditioned invertebrate media
217
-------
were cytotoxic to WI38 and HF cells, reducing their colony forming efficien-
cies to 28 and 66% of controls. EMEM containing 20% SF21 cell conditioned
medium was not toxic to CV-1 cells (CFE 118%).
RESULTS AND DISCUSSION
To test the possibility that the toxicity was due to cellular metabolites
produced during infection, rather than a constitutive component of the
virions, AcNPV was separated from cellular components by gradient purification
and/or dialysis, and tested. Neither purification procedure eliminated the
toxic effect.
Historically, the IRS assay has been utilized as a preliminary step in
the evaluation of chemicals suspected of other biological effects. In the
present study IRS was utilized to assess the toxicity of AcNPV to vertebrate
cells. The results demonstrate the following advantages, a) Colonies, clonal
survivors of a treated cell, need not be detached and can be quantified
directly, b) small numbers of cells (ca. 200) are exposed and neutralization
and detoxification are reduced to minimal levels, c) variability is minimized
because all test cultures and controls are seeded with an equivalent number of
cells derived from the same dilution and are evaluated relative to each other,
and d) the fixed, stained and dehydrated culture dishes can be archived and
provide a permanent record of the experiments.
The toxicity of AcNPV preparations were not removed by gradient purifica-
tion and dialysis. This suggests that either the components of budded AcNPV
are innately toxic, or toxic components of the preparations are so tightly
associated with AcNPV as to co-purify with the virions during gradient purifi-
cation. Cross-linking of AcNPV DNA with psoralen and UV failed to reduce
toxicity and in fact, appears to have enhanced the toxicity. This result
suggests that abortive infections do not play a significant role in the
observed toxicity.
The sensitivity of IRS assays have allowed documentation of the toxicity
of budded AcNPV preparations to human cells but the significance of this low
cytotoxicity is unclear. It is doubtful whether cells of vertebrate species
will encounter budded virus at these high concentrations. Thus, the relevance
of this toxicity to pesticide health hazard evaluation remains unascertained.
However, AcNPV has recently been frequently utilized as an expression vector
for gene products of medical utility. In this application, the toxic compo-
nents of AcNPV preparations may have some relevance and should be defined if
viral expression vectors are used for synthesis of certain types of medical
products.
The present study supports the non-persistence of the AcNPV genome in
mammalian cells. However, it documents a low-cytotoxicity to vertebrate cells
not detected in earlier investigations. Inclusion of the IRS assay in the
revised Subdivision M should lead to improved toxicity testing for both
naturally occurring and genetically altered viral pesticides.
218
-------
PUBLICATIONS
Hartig, P.C. Evaluation and Standardization of Mammalian Cell Culture Test
Protocols for Viral Pesticidal Agents Required in Subdivision M of the
Pesticide Assessment Guidelines. PRIORITY RESEARCH PRODUCT: Task 82941E104.
March 1989.
Hartig, P.C., M.A. Chapman, G.G. Hatch, and C.Y. Kawanishi. 1989. Insect
Virus: Assays for Toxic Effects and Transformation Potential in Mammalian
Cells. Applied and Environmental Microbiology 55:1916-1920.
219
-------
GENETIC AND TOXICOLOGICAL STUDIES OF THE Bacillus
thurinaiensis subsp. israelensis 28 kDa PROTEIN
Gary A. Held
Health Effects Research Laboratory
U.S. EPA
Research Triangle Park, NC 27711
INTRODUCTION
Bacillus thurinqiensis subsp. israelensis is a entomopathogenic gram-
positive bacterium which synthesizes a proteinaceous, parasporal inclusion
body during sporulation. This inclusion body, commonly referred to as the
crystal, is toxic primarily to mosquito and blackfly larvae. The inclusion is
composed primarily of four proteins, two with a molecular weight of about 135
kDa, one of about 70 kDa and the last of about 28 kDa. Solubilized crystals
of B. thurinqiensis subsp. israelensis are cytolytic to a wide variety of both
insect and mammalian cells. In addition, if the solubilized crystals are
administered by injection they are lethal to mice and rats. Intact crystals
of B. thurlnqiensis subsp. israelensis are relatively nontoxic to mice,
however, various strains of B. thurinqiensis including B_. thurinqiensis subsp.
israelensis are being subjected to genetic engineering as well as being used
as a source of toxin genes for genetic engineering of other organisms. Since
the toxic proteins in engineered organisms may be in a different form than in
wild-type B. thurinqiensis. it is important to examine the mechanisms of
toxicity to provide confidence in their safety.
Most of the insecticidal toxins of B. thurinqiensis strains are found on
large plasmids. In addition to containing the genes for the toxins, at least
some of these plasmids code for genes required for transfer of plasmids to
other bacteria. This increases the possibility that new traits introduced
into engineered B. thurinqiensis will be transferred into other bacteria.
These considerations indicate that an assessment of potential risks to human
health require examination of both the toxicological and the genetic proper-
ties of bacterial pesticides.
METHODS
Crystals of B. thurinqiensis were prepared by washing sporulated cultures
(grown in modified GYS medium) several times in water followed by centrifuga-
tion through 15-40% NaBr gradients. NaBr was removed by several washes with
deionized water. Crystals were solubilized by incubation at pH 11.7 for 5
hours at room temperature. Protein was quantitated by measuring absorbance at
280 nm. The protein composition of solutions was analyzed by SDS-polyacryl-
amide gel electrophoresis (0.75 mm thick, 10-12%). The gels were run at a
constant power of 6-8 watts/gel. The protein bands were visualized by silver
staining. Crystal proteins were fractionated by immunoaffinity chromatography
using monoclonal antibody specific for the 28 kDa peptide. Bound 28 kDa
protein was eluted from the column with 50 mM Tris-HCl buffer (pH =» 10.5)
containing 0.5 M NaCl. Mosquitocidal activity was determined by using ten
third-instar Aedes aegypti larvae per well of a 24 well microtiter plate
220
-------
containing 2.5 ml of water. Hemolytic activity was assessed by measuring the
amount of hemoglobin (415 nm absorbing material) released from sheep red blood
cells. Mouse LD.0 values were determined by intraperitoneal injection of
solubilized crystal proteins. Mortality was recorded 24 hours after treat-
ment.
Plasmid DNA was prepared from mid-exponential phase cells (grown in LB
medium) by alkali lysis. After lysis the preparation was extracted with
phenol in the presence of 3% NaCl. Genomic DNA was prepared by digestion of
cells with lysozyme followed by lysis with SDS. The preparation was then
extracted multiple times with phenol and dialyzed. Bi-directional Southern
blots were hybridized at 65°C in 6x SSC. The final wash was at 65°C in O.lx
SSC. Plasmid curing and tests of plasmid stability were carried out by growth
42°C.
RESULTS AND DISCUSSION
Immunoaffinity column chromatography of solubilized B. thurinqiensis
subsp. israelensis crystals resulted in a bound fraction containing purified
28 kDa protein and a flow-through fraction containing the remainder of the
crystal proteins. The purified 28 kDa protein retained the hemolytic activity
observed in the unfractionated solubilized crystals, however, the flow-through
fraction did not have detectable hemolytic activity. In contrast, the
mosquitocidal activity of the 28 kDa protein was 30-35 times lower than the
unfractionated crystals. The flow-through fraction retained about 90% of the
mosquitocidal activity of the unfractionated crystals. These experiments
suggested that the mammalian toxicity observed with solubilized crystals may
be due to the 28 kDa protein. The LD50 observed in mice using purified 28 kDa
protein was 0.27 g/kg. Injection with unfractionated solubilized crystals
resulted in an LD50 of 2.33 mg/kg. The flow-through fraction resulted in no
mortality. Treatment of mice with the unfractionated solubilized crystals
also resulted in significant hypothermia, a reduction in heart rate, and the
formation of hemorrhages in several internal organs, most notably the jejunum.
Purified 28 kDa protein was about 4-5 times more active than the unfraction-
ated crystals in inducing these effects. The flow-through fractions possessed
no detectable activity. Taken together these results indicate that the 28 kDa
protein is responsible for the mammalian toxicity observed with B.
thurinqiensis subsp. israelensis crystals and has little role in~mosquitocidal
activity.
The genes responsible for the synthesis of the crystal proteins of B.
thurinqiensis subsp. israelensis are located on the 75 MDa plasmid. Strain
87051202 has been cured of this plasmid by growth at elevated temperature,
however, this strain continues to synthesize a crystal. Although, this
crystal is smaller than the crystal in the parental strain it contains all the
proteins normally present and retains full activity against both mosquitoes
and sheep red blood cells. The synthesis of the crystal is stable after
growth at elevated temperatures which cause the parental strain to lose the
ability to synthesize the crystal at a high rate. Hybridization experiments
using cloned segments of the 75 MDa plasmid have isolated a 1.3 kb Hindlll
fragment that appears to contain the sequences involved in the insertion of
the 75 MDa plasmid into a new genetic location, most probably the chromosome.
221
-------
These hybridization experiments also indicate that sequences homologous to the
1.3 kb Hindlll fragment occur in several locations on the 75 MDa plasmid.
PUBLICATIONS
Held, G.A., Y-S. Huang, and C.Y. Kawanishi. 1986. Effect of removal of the
cytolytic factor of Bacillus thurinqiensis subsp. israelensis on mosquito
toxicity. Biochem. Biophys. Res. Commun. 126:961-965.
Held, G.A., Y-S Huang, and C.Y. Kawanishi. 1990. Characterization of the
parasporal inclusion of Bacillus thurinqiensis subsp. kvushuensis. J.
Bacteriol. (in press).
Mayes, M.E., G.A. Held, C. Lau, J.C. Seely, R.M. Roe, W.C. Dauterman, and C.Y.
Kawanishi. 1989. Characterization of the mammalian toxicity of the crystal
polypeptides of Bacillus thurinqiensis subsp. israelensis. Fund. Appl.
Toxicology 13:310-322.
222
-------
INDEX OF CONTRIBUTORS
Page
Anderson, Anne J 97
Andersen, G 122
Anderson, Richard L 186, 189
Armstrong, J.L 26, 88
Atlas, Ronald M 164
Barkay, Tamar 64
Bej, Asim K 164
Benfield, Ernest F 100
Bleakley, Bruce H 149
Bolton, Harvey, Jr 159
Bott, Thomas L 139
Briant, James K 183
Buchholz, Phyllis S 181, 183
Buell, Robin 97
Butterworth, Julie 38
Buttner, Mark P 47
Byrd, J 21
Byrne, A 57
Caldwell, Bruce A 144
Cardon, M 216
Campbell, Bob 103
Chacko, Rosy J 61
Claxton, Larry 0 205, 210
Coffin, Richard B 49, 137
Colwell, R.R ". 21
Conde, B 21
Connolly, John P 49
Cowan, Marjorie M 106
Crawford, Don L 149
Croft, B.A 177
Cuskey, Stephen M. 197
Devereux, Richard 133
Donegan, K 200
Doyle, J.D 153, 156
Driver, Crystal J 183
Fairbrother, Anne 181, 183
Ferrante, A 57
Fletcher, Madilyn 106
Ford, Simon J 31
Foss, Steve 103
Fournie, Jack 103
Fowles, N 88
Fredrickson, James K 159
Frischer, Marc E 67
Ganio, Lisa 40
Genthner, Fred 103
George, S. Elizabeth 205, 210, 214
Green, Brian D 45
223
-------
INDEX OF CONTRIBUTORS (continued)
Pace
Griffiths, Robert P 144
Hamblin, Martha 119
Harris, D 88
Hartig, P 216
Held, Gary A 220
Hughes, Patrick R 119
James, Rosalind D 171
Jarrell, Ann E 183
Jeffrey, Wade H 67
Jones, Daniel D 17
Jorgensen, Niels 137
Kaplan, Louis A 139
Katsuwon, Jirasak 97
Kawalek, Michael 91
Kawalek, M 125
Kawanishi, C.Y 205, 214
King, R.J 153, 156
Knight, T.T 21
Kohan, Michael J 210
Kokjohn, Tyler A 72, 110
Kroer, Niels 137
Landeck, Robin 49
Lanners, Jacques 115
Lenski, Richard E 115
Lessie, T.G ". 57
Levin, Morris A 13
Lighthart, Bruce 29, 35, 171
Linderman, Robert G 144
Lindow, Steven E 122
Loper, Joyce E 91, 125, 144
Lovic, Branco 97
Marthi, Balkumar 29, 40
Martin, Susan 103
Matyac, C 200
McCartney, H. Alastair 38
Mead, Eric 186
Messing, R.H 177
Miller, Jeffrey C 174
Miller, Robert V 72, 110
Moore, Larry 91, 125
Nelson, Gail M 212
Nguyen, Toai T " 115
Olson, Betty H 31
Olsen, R.H 153
Olsen, R.H 156
Panapoulos, N.J 122
Paul, John H 67
Paul Hz, Timothy C 144
224
-------
INDEX OF CONTRIBUTORS (continued)
Pace
Perl in, Michael H 164
Porteous, L.A 26
Prince, Valerie J 40
Rochelle, Paul A 31
Rygiewicz, P 88
Sayler, Gary S 72
Scaccia, Brian 174
Seidler, Ramon J 40, 153, 156, 200
Shaffer, Brenda 40
Shannon, Lyle J 189
Sherwood, R.L 214
Short, K.A 153, 156
Smith, Linda G 183.
Somerville, C 21
Stahl, David A 133
Stetzenbach, Linda D 45, 47
Stewart, Gregory J 67, 78
Stockwell, Virginia 91, 125
Stotzky, G 82, 153, 156
TeBeest, D.O. 61
Thomann, Robert V 49
Wang, Zemin 149
Walter, Michael V 40
Warren, Tessie M ,. 106
Weidemann, G.J 1 61
Whitehouse, Douglas A 210
Wilson, M 122
Wood, H. Alan 119
Wood, M.S 57
Yousten, Allan A 100
225
------- |