EPA-600/4-79-075
NASA Technical Memorandum 80130
INVESTIGATION OF EFFECTS OF TEMPERATURE,
SALINITY, AND ELECTRODE DESIGN ON THE
PERFORMANCE OF AN ELECTROCHEMICAL COLIFORM
DETECTOR
DAVID C, GRANA
NOVEMBER 1979
Langley Research Center
Hampton, Virginia 23665
-------
INVESTIGATION OF EFFECTS OF TEMPERATURE, SALINITY,
AND ELECTRODE DESIGN ON THE PERFORMANCE OF AN
ELECTROCHEMICAL COLIFORM DETECTOR
David C. Grana
National Aeronautics and Space Administration
Langley Research Center
Hampton, Virginia 23665
IAG-D7-0053
-------
PREFACE
The Federal Water Pollution Control Act Amendments of 1972
charter the Environmental Protection Agency (EPA) to provide a
water quality surveillance system for monitoring the quality of
navigable waters, ground waters, the contiguous zone, and the
ocean. This Act also specifically charges the EPA to utilize
the resources of the National Aeronautics and Space Administra-
tion (NASA), to the extent practical, to provide such a system.
Accordingly, the Director, Surveillance and Analysis Division,
EPA Region II, initiated a joint effort with NASA to evaluate
the use of NASA's new technique for rapidly detecting coliforms
in the coastal waters adjoining the New York Bight. The goal
of this effort was to aid EPA in developing technology for an
operational system to monitor total and fecal coliforms in waters
that lie within the jurisdiction of Region II. The following
tasks were identified to meet the program objectives:
TASK 1 - 8-CHANNEL ELECTROCHEMICAL COLIFORM DETECTOR UNITS
Establish a data base using laboratory and field evaluation
of the unit in order to compare with standard methods for total
coliform measurements. One unit was furnished to the EPA
Region II Laboratory, Edison, New Jersey, for a 2-week shipboard
evaluation in the New York Bight area. NASA provided personnel
to operate the unit and analyze sensor data during the evaluation,
EPA personnel performed standard method tests. Continuation of
this effort and subsequent tasks delineated in this plan were
at the discretion of EPA upon completion of evaluation of the
2-week test results.
TASK 2 - FECAL COLIFORMS MEASUREMENT STUDIES
Determine the optimum detector design for measuring fecal
coliforms in saline waters for operational systems. This task
was implemented as follows:
(a) Research grants were initiated to study the effects
of temperature and salinity on the electrochemical/organism
interface performance. EPA provided funding for this task.
In addition, EPA implemented this study under the technical
guidance of NASA.
(b) NASA performed concurrent in-house studies to aid the
grantees in delineating the effects of temperature and salinity
on fecal coliforms.
-------
(c) The results of the studies given in Task 2(a) and 2(b)
were used to define the optimum sensor design for monitoring fecal
coliforms in saline water.
TASK 3 - EVALUATE REMOTE SAMPLING PLATFORMS
Remote sampling platforms to measure total and fecal coliforms
in situ were evaluated and information was transmitted to a base
station. This evaluation was performed in conjunction with EPA
in local waters adjacent to the NASA Langley Research Center.
The task was implemented as follows:
(a) Assembled and checked out platforms, sensors, electronics
and base station unit.
(b) Performed 1-month test to monitor total and fecal
coliforms in water adjacent to the NASA Langley Research Center.
Test results were analyzed and documented, and appropriate
design modifications implemented, as required.
(c) Additional tests were implemented as specified jointly
by EPA and NASA. EPA provided support to deploy and retrieve the
platform and personnel to perform standard method tests. Test
site location within EPA Region II area was Caven Point, New
Jersey. NASA provided the base station equipment in order to
communicate with the platform.
TASK 4 - EVALUATE REMOTE SAMPLING PLATFORM IN THE EPA REGION II
AREA
NASA refurbished one sampling platform and delivered it to
EPA. The platform sensor was modified for demonstration purposes
to incorporate the capability to measure fecal coliforms in
saline water. This platform was evaluated by EPA and NASA.
NASA provided personnel to operate and maintain the unit for a
2-week demonstration test. EPA provided support and personnel to
perform standard method tests. Test site location within the
New York Bight was determined by EPA subject to the NASA speci-
fied design constraints of the platform. NASA provided the base
station equipment in order to communicate with the platform.
Additional evaluation was at the discretion of EPA with consulta-
tion by NASA.
Tasks 1, 3, and 4 were implemented under EPA Interagency
Agreement IAG-D6-0930 and are reported in references 2, 3, and 4.
This report describes the results of task 2 and is in fulfillment
of the Interagency Agreement between NASA Langley Research Center
and the Environmental Monitoring and Support Laboratory (EMSL),
Las Vegas (EPA-IAG-D7-0053).
-------
The work reported herein was accomplished through two
grants: Virginia Institute of Marine Science was awarded a
grant to evaluate salinity and temperature as stress factors
affecting the enumeration of fecal conforms by the
electrochemical detection of molecular hydrogen. A second
grant was awarded the University of Virginia for optimization
of waterborne coliform sensor for saline water.
iv
-------
ABSTRACT
This report presents the results of two research programs
to determine the optimum detector design for measuring fecal
coliforms in saline waters for operational systems. One program
was concerned with the effects of temperature and salinity on
endpoint detection time and the other, the interaction between
electrode configurations and the test organisms.
Test results show that the endpoint response time is
related to salinity and seawater temperature; however, these
results can be minimized by the correct choice of growth media.
Electrode configurations were developed from stainless steel,
Parlodion-coated stainless steel and platinum that circumvented
problems associated with the commercial redox electrodes.
-------
FIGURES
Number Page
1 Experimental setup for performing hydrogen response
measurements .-."...- ..25
2 Strip«chart tracing of hydrogen response curve for
1.9 x 10° cells/100 ml ... o ...... , . .26
3 Relationship between inoculum size and length of
lag period ........... 27
4 Relationship between inoculum size and the time
of endpoint response in TSB at 44.5°C when
inocula (urinary track isolate) were pregrown
in TSB at 35°C ......_ 28
5 Relationship between inoculum size and the time of .
endpoint response in TSB at 44.5°C when inocula
(urinary tract isolate) were pregrown in M-9
medium at 35OG . . . •„ 29
6 Relationship between inoculum size and the time of
endpoint response in EC at 44.5°C when inocula
(urinary tract isolate) were pregrown in TSB
at 350C ....-.."......_ o 30
7 Relationship between inoculum size and the time of
endpoint response in EC at 44.5°C when inocula
(urinary tract isolate) were pregrown in M-9
medium at.35°C ....... -31
8 Comparison of viability of E. coli at 20°C and 30°C
at various salinities 32
9 Effect of starvation at 20°C in seawater of various
salinities on the endpoint response (ER) of E. coli
inoculated into either EC or TSB at 44.5°C . . .33
10 Effect of starvation at 30 C in seawater of various
salinities on the endpoint response (ER) of E.coli
inoculated into either EC or TSB at 4405°C . . .34
11 Effect of starvation at 20°C in seawater (10 ppt
and 25 ppt) on the endpoint response (ER) of
E. coli inoculated into either EC or TSB at
35
vi
-------
Number Page
12 Effect of starvation at different temperatures in
seawater of various salinities on the enumeration
of E. coli (urinary tract isolate) immediately
folTowing a tenfold dilution into either EC or
TSB at 44.5°C0 . . . . . 36
13 Effect of starvation at 20°C in seawater (10 ppt
and 25 ppt) on the enumeration of E. coli
immediately following dilution into either EC
or TSB at 44.5°C '. . 37
14 Growth of E. coli in EC and TSB, 44.5°C after
Oj,2 and £ days starvation at 20°C in seawater
(25 ppt) . . 38
15 Viability of E. coli at various temperatures in
seawater (25 ppt) „ . . . „ „ . . . „ 39
16 Effect of starvation at 20°C in seawater (25 ppt)
on the endpoint response (ER) of E. coli
inoculated into either EC or TSB at 4~4T5~°C ... 40
17 Effect of starvation at 35°C in seawater (25 ppt)
on the endpoint response (ER) of E. coli
inoculated into either EC or TSB at 44.5 C ... 41
18 Effect of starvation at 2°C in seawater (25 ppt)
on the endpoint response (ER) of E. coli
inoculated into either EC or TSB at 44.5 C ... 42
o
19 Effect of starvation at 2 C in seawater (25 ppt)
on the enumeration of E_. coli immediately follow-
ing a tenfold dilution into either EC or TSB at
44.5°C . . . . o 43
20 Growth of E. coli in EC or TSB, 44.5°C after 20
and 67 hours starvation at 20°C in seawater
(25 ppt) 44
21 Relationship between inoculum size and the time of
response in half-strength Medium A-l when inocula
were pregrown in TSB at 35°C . 45
22 Relationship between inoculum size and the time of
endpoint response in half-strength EC at 44.5°C
when inocula (urinary tract isolate) were pre-
grown in TSB at 35°C . 46
vii
-------
SECTION 1
INTRODUCTION
The principal biological indicator of water pollution is the
presence of coliforms. In the 14th edition of Standard Methods
for the Examination of Water and Wastewater (ref. 1), the approved
methods for coliform analysis are the most probable number (MPN)
and membrane filtration (MF) tests. Both laboratory tests are
laborious and time-consuming, requiring 24-96 hours to complete
the MPN test, and 24 hours for the MF test. In view of the
increased interest in United States water resources and the ever
present threat of fecal contamination, increased emphasis has
been placed on the development of automated in situ systems,
rather than laboratory methods, to meet the extensive surveillance
required to maintain coliform standards.
Recently, Wilkins et al. (ref. 2) described a technique for
detecting bacteria based on the time of hydrogen (Ho) evolution.
The principle of the electrochemical method is outlined in figure 1.
Briefly, it consisted of a reference electrode and a platinum
electrode connected to a strip^chart recorder. A typical dose response
curve consisted of a lag period, a period of rapid buildup in
potential due to hydrogen, and a period of decline in potential
(figs. 2 and 3). A linear relationship was established between
inoculum size and the time hydrogen was detected (lag period); it
was shown that, for example, one bacterial cell could be detected
in 7 hours (fig. 4). Further studies indicated that the electro-
chemical method could also be used to detect fecal coliforms
(ref. 3), and the electrodes were readily adaptable to an in situ
system. In an extension of these studies, an automated system
for the in situ monitoring of coliforms in water was developed to
circumvent many of the current problems of enumerating coliforms,
viz., sample acquisition, transport, trained laboratory technicians,
and the 24 to 96 hours to process a sample. In order to adequately
test the in situ concept for monitoring coliforms, a two-phase
field evaluation program was conducted. Phase one was conducted
in fresh and estuarine waters in the vicinity of the Langley
Research Center, Hampton, Virginia, and the second phase was
conducted at Caven Point, New Jersey, in conjunction with the
Environmental Protection Agency (EPA). Reference 4 describes a
pilot model of the in situ monitoring system and presents the
results of the field evaluations.
A number of problems were identified during laboratory and
field tests of the coliform detector. Among the problems were the
effect of salt concentration on detection time and the occasional
-------
instability of the commercial electrodes, For example, when the
test system was evaluated using estuarine and freshwater samples,
differences in detection times between stock- cultures and envir-
onmental coliforms were observed. Wilkins and Boykin (fef.. 3)
suggested that a portion of the coliform population present in
the aquatic environment may be stressed with consequent delay in
detection times due to exposure of sensitive cells to selective
media and elevated incubation temperatures. Data resulting from
studies using coliforms exposed to freezing (ref. 32, 32 ), chlor-
ination (ref, 20), and freshwater environments (ref. 7, 16, and 19)
suggested that injured bacterial cells may remain undetected by
routine enumeration procedures.
The instability of the electrode systems occasionally manifested
itself as either a sudden or gradual departure from the baseline
voltage level. The sudden change in voltage levels generally
occurs at the beginning of the measurement and gradually reaches
a steady-state about 1 hour after inoculation. Both instability
problems have been minimized by appropriate designs in the
electronics; however, a more reliable electrode would greatly
simplify the measurement and cost.
This report presents the results of each task separately.
Section 4 contains the results of the evaluation of salinity and
temperatures as stress factors. Section 5 contains the results of
the electrode sensor investigation.
-------
SECTION 2
CONCLUSIONS
SALINITY AND TEMPERATURE STRESS FACTORS
The ability of an electrochemical detection method to predict
viable numbers of fecal coliforms was evaluated under laboratory
conditions with respect to seawater adjusted to various salinities
and temperatures. The viability of an Escherchia coli isolate
as measured by the spread-plate technique utilizing nonselective
media was unaffected after 12 weeks exposure at 2°C and 25 ppt
salinity. At higher temperatures (15-30°C), both the total
decrease in cell numbers as well as the rates of die-off were
greater than at 2°C0 There was little apparent difference in
viability across the temperature range 15-30°C, Viability was
observed to be inversely related to salinity over the range 10-30
ppt. Stress was measured using the electrochemical detection
method (ECDM), and defined as the difference between the predicted
endpoint response time (ER) calculated from a standard curve and
the observed ER time. Seawater of higher salinities generally
produced greater stress. With respect to temperature, stress was
greater at 2QOC than at 30°C, while at 2°C stress occurred after
a prolonged period of starvation. Delayed ER times were attributed
to (1) a reduction in viable cells upon inoculation of starved
bacteria into media at 44.5°C and/or (2) an extended lag phase
prior to logarithmic growth. Medium A-l was superior to EC for
enumeration of fecal coliforms in estuarine water samples by
the ECDM method. ER times occurred sooner and the results were
more predictable with the former medium.
STABILITY OF ELECTRODE SYSTEMS
The experiments to date have resulted in the selection of
inexpensive materials which can function as a microbe sensing
probe for use in culture media. If the probe can be manufactured
at a reasonable cost with good reproducibility, the stainless-steel,
Parlodion-coated stainless steel system could be used in a variety
of applications.
-------
SECTION 3
RECOMMENDATIONS
SALINITY AND TEMPERATURE AS STRESS FACTORS
In order for ECDM to predict coliform levels accurately, it
would be advisable to perform additional experiments to determine
selective media and temperatures most effective for the recovery
of injured coliforms. Our data indicated that Medium A-l would
be preferential to EC at 44.5 C. Moreover, it is essential to
have accurate standard curves developed using test media at the
same nutrient and selective agent concentrations and temperatures
as would be used in obtaining ER data from field samples. It is
also imperative that low cell concentrations in the range 10~1 to
10^ cells/ml be used in developing these standard curves. Standard
curves developed during this study varied depending upon the media
used, previous cultural history of the organisms, and nutrient
and/or selective agent concentration in the media. In addition,
there were suggestions that inocula of different volumes, although
diluted to the same cell density, could affect ECDM results.
Additional experiments should also be conducted to determine
the physiological condition of E. coli (or other bacteria)
exposed to the environment under a variety of conditions. The
sensitivity of this method for monitoring stress should be
examined. A comparison could be made with other techniques
capable of detecting sublethal stress, i.e., adenylate charge,
enzyme analyses, or growth on selective versus nonselective
media.
This investigation has posed additional questions. For
example, what role does cryptic growth play in maintaining the
apparent viability of starved cultures? Postgate and Hunter
(ref, 27) demonstrated that denser populations of E. aerogenes
survived longer than sparser ones. Greater stress and viability
losses might be observed if starvation experiments were conducted
using cell inocula in the range 10^ to 1CH cells/ml. Such
information would be valuable in further assessing the ECDM as a
method of monitoring stress. Viability and stress levels should
be compared using cells starved under hypoosmotic versus hyper-
osmotic conditions to evaluate whether or not the hyperosmotic
environment encountered in these studies provided protection to
the exposed cells. The degree of leakage of cell constituents
under various conditions of salinity should be determined.
-------
STABILITY OF ELECTRODE SYSTEMS
Based upon results of the electrode system investigations, it
is recommended that (1) the impedance of the test equipment
should be more closely matched to that of the commercial electrodes,
and (2) further tests should be conducted on stainless steel and
other metals that show promise of functioning as bacteria-sensing
probes.
As a corollary to the work on stainless-steel probes, NASA
Langley Research Center has conducted tests on platinum wire as
one candidate for the bacteria probe. The results of these tests,
published in reference 38, show that a linear relationship was
established between inoculum size and response time.
-------
SECTION 4
EVALUATION OF SALINITY AND TEMPERATURE AS STRESS FACTORS AFFECTING
THE ENUMERATION OF FECAL COLIFORMS BY THE ELECTROCHEMICAL METHOD
EXPERIMENTAL PROCEDURES
Organisms
Fecal coliforms employed in this study were isolated by enrich-
ment in lactose broth at 35°C for 24 hours followed by inocula-
tion into EC broth at 44.5°C for 24 hours. Isolates were obtained
from tubes which produced gas at the elevated temperature. All
isolates were identified using API-20E (Analytab Products, Inc.,
Plainview, New York), a miniaturized multiple test system for
identification of enteric bacteria. API profile recognition
numbers for Escherichia coli isolates were 5 044 552, human fecal
isolate; 5 044 562, human urinary tract infection isolate;
and 1 144 562, estuarine water isolate obtained from a sample
collected from the York River, Gloucester Point, Virginia. Stock
cultures were maintained on trypticase soy agar (TSA) at 4°C.
During the course of experimentation, isolates were subcultured
on TSA three times or less.
Cell Preparation
Inocula for seawater survival experiments were obtained from
exponential phase cultures unless otherwise specified. Isolates
were grown at 35°C in either trypticase soy broth (TSB) or M-9
minimal medium which consisted of 6g Na2HP04, 3g KH2p04> O.Sg.NaCl,
Ig NH4C1, 0.13g MgSo4'7H20 and 5 ml glycerol in 1 liter distilled
water. The latter two components were sterilized separately and
added to the remaining components of minimal medium after auto-
claving. The final pH of the minimal medium was 7.0. Bacteria
were harvested by centrifugation (1000 x g) at room temperature
for 10 minutes and washed three times in phosphate buffered saline
(PBS) at pH 7.2. Cells were resuspended to an optical density
of 0.2-0.3 (540 nm) in seawater of selected salinities and
temperatures.
Survival Experiments
A tenfold dilution of the final cell suspension was made
into 90 ml seawater previously equilibrated to the test temperature
in 125 ml sterile screw-capped flasks. Seawater at specified
-------
salinities was prepared by diluting aged filtered ocean water
(35 ppt) with reverse osmosis/glass distilled water and
sterilized by autoclaving. Samples of inoculated seawater were
removed at various time intervals for enumeration by spread
plating on TSA and determination of endpoint response (ER) in EC
and TSB at 44.5°C. Colony counts were made after approximately
24 hours incubation at 35 C0 Total exposure time to seawater
prior to 0 hour sampling ranged from 2-5 minutes while exposure
time to media at 44.5°C prior to dilution for plate counts was
less than 1 minute. All dilutions were performed using PBS at
room temperature unless otherwise specified.
Endpoint Response (ER) Determination
The experimental design for detecting ER time consisted of a
test tube (25 mm by 100 mm) containing two platinum electrodes
and 18 ml medium to which 2 ml E. coli seeded seawater was added.
Tubes were fitted with a No. 4 rubber stopper containing elec-
trodes of grade A platinum-alloy wire, 24-gauge (0.508 mm)
(Englehard Industry, Carteret, New Jersey). Electrodes were
inserted into slits made in the stopper, and the stopper bound
with wire and/or epoxy to prevent electrode slippage. Electrodes
were designed so that the ratio of their lengths below the surface
of the medium was 1:4. Sterilization was by flaming over an
alcohol lamp. The longer electrode was connected to the negative
terminal and the shorter one to the positive terminal of a strip-
chart recorder (Model 194, Honeywell Industrial Division, Fort
Washington, Pennsylvania, or Model SR-204, Heath Company, Benton
Harbor, Michigan), Recorders were set at a chart speed of 10
minutes/inch and operated at 0.5 volt (Honeywell) or 1.0 volts
(Heath) full scale. ER times were measured as the time elapsed
between challenge and the initial increase in potential difference,
Environmental Samples
Estuarine water samples were collected from the York River
at Gloucester Point or from its nearby tributaries which included
Yorktown Creek, Sarah Creek, as well as the Northwest and North-
east branches of Sarah Creek. Sampling sites at all localities
were subject to varying degrees of pollution from domestic
sewage and boats. Water samples were collected in sterile 500 ml
Erlenmeyer flasks and processed within 60 minutes. Simultaneous
temperature and salinity measurements were made. Parallel fecal
coliform enumerations were performed using a five-tube most-
probable-numer (MPN) technique (ref. 1) with lactose enrichment
at 35°C for 48 hours with subsequent inoculation of positive
tubes into EC at 44.5°C for 24 hours, or direct inoculation
into Medium A-l (ref. 5) at 44.5°C for 24 hours. Corresponding
ER determinations were made by addition of a 100 ml sample into
each of two 250 ml Erlenmeyer flasks containing 100 ml single
strength EC or Medium A-l prewarmed to 44.5°C. Flasks were
fitted with rubber stoppers containing the appropriate sterile
platinum electrode configuration as specified above.
-------
Standard Curves
Endpoint response times were determined in EC, TSB, and
Medium A-l at 44.5°C for various inocula. Tenfold dilutions of
an E. cpli suspension (urinary-tract isolate) prepared as for
survival experiments were made in 20°C seawater (25 ppt) and
inoculated into test media. Exponential cultures were pregrown
in either TSB or M-9 media. The relationship between inoculum
size and ER time was determined by linear least-squares regression
analysis, and confidence belts for the linear regression were
established at the 95 percent confidence level (ref. 31).
Data Analysis
During the course of experiments, designed to assess the
effect of seawater exposure on ER time, the following data were
collected: coliforms/ml seawater, coliforms/ml medium following
a tenfold dilution from seawater and ER time in the respective
media. Standard curves of ER versus coliforms/ml medium were
analyzed by linear-regression techniques. The ER, predicted from
the linear-regression line and based on the number of cells/ml
medium calculated to have resulted from a tenfold dilution from
seawater, was subtracted from the observed ER. This relationship
of observed ER minus predicted ER as a functi.on of time in sea-
water was graphically presented with 95 percent confidence limits
for y. Stress was defined as the time delay between observed
ER and predicted ER, and was considered significant when this
value exceeded the 95 percent confidence limits.
8
-------
SECTION 5
STABILITY OF ELECTRODE SYSTEMS FOR COLIFORM DETECTORS
ANALYSIS OF COMMERCIAL ELECTRODE FAILURES
The following nine "defective" combination electrodes were
provided by NASA to the University of Virginia (UVA). They
were reported to have one of two defects--erratic baseline or
no response.
Electrode No. Performance
RSP-1 erratic baseline
RSP-5 erratic baseline
RSP-7 erratic baseline
RSP-8 erratic baseline
RSP-9 erratic baseline
RSP-16 no response
RSP-17 no response
A-l no response
2 no response
Experiments were performed by comparing at least two electrodes
known to function properly with a number of the NASA problematic
electrodes. Typical experiments were carried out by preparing
a mother culture of trypticase soy broth (TSB) which had been
recently inoculated with a low concentration (10 /ml), of
Pseudpmonas aeruginosa. The mother culture was then split into
3-ml samples added to a sterile test tube in which the
electrode had been placed. The electrbdes(up to eight) were
then monitored for approximately 24 hours.
All of the electrodes were presterilized by immersion
in boiling water for a period of 10 minutes.
The following is a summary of this first set of experiments
-------
Electrode No. Performance (NASA) Performance (UVA)
RSP-1 erratic baseline OK
RSP-5 erratic baseline OK
RSP-7 erratic baseline erratic baseline
RSP-8 erratic baseline OK
RSP-9 erratic baseline OK
RSP-16 no response erratic baseline
RSP-17 no response erratic baseline
A-l no response OK
2 no response no response
These results indicate that a fundamental disagreement
exists between University of Virginia (UVA) procedures/apparatus
and those at NASA for at least five of the nine electrodes. For
those which displayed "erratic baseline" in UVA experiments, the
deviation was much less severe (10-40 mV) than that reported at
NASA. Only electrode 2 failed completely, in agreement at NASA
and UVA. On examination, the reason for this electrode's failure
was apparent in that there was a large leak of KC1 at the rate
of about 0.2 ml/minute. When arrested, this electrode performed
favorably.
In view of the fact that with this one exception, the UVA
experiments gave either less severe problems or no problems at
all (in five cases), attempts were made to compare the differences
in the UVA equipment and that at NASA. Communications with
Dr. Judd Wilkins at NASA indicated there were no significant
procedural differences in either the microbiology or electrode
preparation procedures. However, it was learned recently that
there is a fundamental difference in the test equipment: namely,
the input impedance of the electronics. The NASA equipment has
an input impedance of approximately lO^ whereby UVA's is about
10 . Independent of this effort, Wilkins found that by lowering
the impedance of the NASA system many, but not all, of the
electrode problems were reduced.
7 UVA has recently proceeded with an examination of RSP-7, 16,
and 17, which still did not perform properly. Reducing the
concentration of electrolyte in the electrode from saturated KCl
to 0.9 percent NaCl diminished the possibility of salt clogging
the fritt. This seemed to straighten out the baseline, but also
10
-------
reduced the overall response about 25 percent. This change from
a saturated calomel electrode (SCE) to a normal calomel electrode
(NCE) should only change the baseline of the reference electrode
from +0.226 to a slightly less positive value according to the
Nernst equation. Since the platinum electrode moves in the
same direction with micro-organism growth (in the negative
direction), the switch from a SCE to a NCE is consistent with a
slightly smaller response.
DEVELOPMENT OF A COST-EFFECTIVE DETECTION PROBE
This part of the report describes the series of experiments
performed between October 1, 1977, and March 6, 1978, at the
Applied Electrochemistry Laboratory, University of Virginia.
The goals of these experiments were
To develop an electrochemical probe with the following
features:
a. Capable of detecting organism growth in standard
culture media.
b. Potentially low enough in cost to be disposable
if manufactured in quantity.
Variables investigated:
1. Membrane type and application
2. Electrode materials
3. Media
4. Microbe
All development tests were carried out on an electrode system
with one exposed surface and another (reference) electrode coated
with a membrane.
Membrane Selection
Experiments 1 through 14: using TSB and P. aeruginesa.-
All variables were held constant with the exception of membrane
type and application. Membranes studied were agar, 7 percent
Parlodion dissolved in acetone, 7 percent Parlodion dissolved
in amyl acetate, and a natural fruit membrane. Agar as a membrane
was initially investigated by covering an electrode placed in
the bottom of a test :tube. This test-tube application showed
good response to microbe growth; but as a thin surface coating
on stainless-steel wire*, it was not conducive to microbial
detection when used in a stainless-steel wire--stainless-steel
needle coaxial electrode arrangement.
*earlier electrode research by Stoner has shown that 304 stain-
less steel is a suitable electrode material.
11
-------
The 7 percent Parlodion (dissolved in acetone) membranes
were studied on the basis of applying the material to the stainless-
steel wire surface to give maximum response and reliability when
used in the coaxial configuration. Membranes were dried at room
temperature (approximately 24°C). Note that all membranes were
single application to stainless-steel wire.
Seven percent Parlodion (dissolved in amyl acetate) was
investigated. It was shown that this membrane, applied as a
thin coat and dried at approximately 70° C, was better than all
other membrane materials tested.
Experiments 15 through 19: Membrane Optimization Using
TSB ancT P. aeruginosa.- Tnis set of experiments was performed to
evaluate the optimum method for producing a uniform membrane
(7 percent Parlodion dissolved in amyl acetate) coating on the
stainless-steel wire. A single application with only the tip
of the stainless-steel wire dipped a second time to cover any
sharp edges proved to be more effective than applying either
an all-over single coat or double coat of membrane material.
The reason was twofold:
1. A double coating increased resistance between wire
and needle thus making probe impedance too high.
2. A single coating did not sufficiently cover burrs at
the end tip of the wire used in some probes thus causing shorting
of that probe.
Pretreatment of the stainless-steel wire surface was also
studied. Acid roughening and sanding of the wire indicated no
significant increase in detection sensitivity or baseline
stability.
At this point, the following assembly protocol was produced.
If followed, this assembly procedure gave consistently reproduc"-
ible results for all systems tested to date. It must be pointed
out that these experiments were performed in test tubes with
seeded cultures.
Preparation of Disposable Needle Electrodes-
Materials: 18 gauge stainless-steel disposable hypodermic
needle; 0.014-inch diameter stainless-steel wire (hypo-
grade); Parlodion (cellulose nitrate); amyl acetate; cotton.
1. Cut stainless-steel wire into 3-inch lengths (if available
use V-portion of wire snippers to reduce rough edges at tip).
2. Soak in acetone for 10 or more minutes — then wipe off.
12
-------
3. Soak in alcohol (Ethyl) for 10 or more minutes—wipe off.
4. Premake membrane solution: 7 percent cellulose nitrate
by weight into amyl acetate (assume that amyl acetate density
is 1 gm/cc).
5. Dip cleaned wires into membrane solution. Pull out
very slowly. A slow velocity will allow the surface tension
of the solution to pull excess solution off the wire.
6. Dry wires by hanging them in an oven at 65-75°C for
15 minutes.
7. Redip the tips (forming small bead at end to assure
coverage of sharp edges) and let dry 15 minutes in the oven.
8. Insert membrane coated wire into the syringe end of
the hypodermic needle being careful not to bend or scrape
the wire on the needle. When the wire comes to the sharpened
end of the needle, stop and bend remaining wire over syringe
end of the needle.
9. Insert small piece of cotton inside syringe end of
needle to secure wire and allow venting.
All the remaining experiments were performed with probes
made according to the assembly procedure described.
Experiments 20 through 24, 30. Using TSB and P. aeruginosa.-
Experiments were conducted to determine type of electrode materia1
most applicable in manufacturing electrode detection unit (i.e.,
coaxial wire and needle assembly). Optimum electrode materials
were found to consist of straight 28-gauge stainless-steel wire
(0.014-inch diameter) and 18-gauge 1 1/2-inch, stainless-steel
disposable needles. Commercially prepared coaxial tubing was
also tested and showed poor adaptability to this process.
Problems were encountered with these experiments until the
coated center wire was obtained in stiffened and straightened
form. Soft wire or wire which had been wound on spools had a
tendency to short-out against the needle.
Experiments 27, 28, 31, 32, 33: System Evaluations Using
Various Media and P. aeruginosa.- These experiments were designed to
evaluate the detection system using various media that are used
in clinical BCB. Columbia broth with 10 percent sucrose and
certoid, thioglycolate, and brain-heart infusion were evaluated.
All had 0(+) blood added (10 percent by volume). The Columbia
and thioglycolate broths were also evaluated without the addi-
tion of blood. Results from all of these experiments showed good
baseline stability with good, clear deflections in output poten-
tial indicating microbe detection. Experiments using a 10 percent
13
-------
blood addition achieved baseline stability in less time than
experiments studied without the blood addition.
Experiment 34: Using Four-Wire Electrodes.- This experiment
was designed to show the characteristic behavior of 304 stainless-
steel wire in both a coated (with 7 percent Parlodion) and
non-coated condition when placed in Columbia broth that has
been inoculated with P e^ aeruginosa. There were two coated wires and
two non-coated wires, all of which were placed into the same
solution. One coated wire was selected as a common reference for
the remaining test electrodes When tested with the other coated
wire, the results indicate that both wires experienced the same
changes thus producing no net change in potential. Testing the
coated reference wire wiTth respect to the non-coated wires
produced the characteristic detection deflection in potential.
Both responded at the same time and with the same level of
potential change.
14
-------
SECTION 6
RESULTS
Linear regression analysis of standard curve results indicated
that both the previous cultural history of the inoculum and the
medium used to measure endpoint response affected ER detection
times. Using TSB as the test medium to measure ER times, cells
pregrown in either TSB or M-9 gave standard curves with similar
slopes, correlation coefficients, and standard errors of the
estimate (figs. 4 and 5). In contrast, using EC as the test
medium, TSB pregrown cells gave a standard curve with a smaller
correlation coefficient and a higher standard error of the
estimate (fig. 6). When transferred into EC, both M-9 and TSB
pregrown cells produced standard curves with steeper slopes and
greater ER time intervals compared with cells transferred into
TSB (figs. 6 and 7).
A series of experiments were conducted to evaluate E. coli
viability in seawater and post-starvation behavior in various
test media as a function of salinity and temperature. E. coli
(urinary tract) growing exponentially in M-9 was inoculated into
seawater of various salinities. The greatest rate of decline
in viable counts occurred during the first 2 days, rates of which
increased directly with salinity (fig. 8). Although two experi-
ments each were conducted at 2Qoc and 30°C, no clear effect of
temperature with respect to viability was evident. Average
E. coli decreases after 2 days were 0.1, 0.6, 1.1, and 1.9 log
units at 20°C and 0.3, 0.7, 1.2, and 1.7 log units at 3QOC in
10, 15, 25, and 30 ppt, respectively.
When the data were analyzed with respect to ER determinations,
stress as previously dejEined. was evident as a function of salinity,
temperature, and test media (figs. 9 and 10). In starvation
studies conducted at 20°C, stress was more pronounced in EC
than TSB and increased with increasing salinity. Generally,
maximum stress developed after 2 days of starvation and, there- ,
after, decreased. Cells starved at 30°C showed little evidence
of stress except at 30 ppt salinity when transferred to EC.
A similar study was conducted using an E. coli strain isolated
from estuarine water, pregrown in M-9, and starved in 10 ppt
and 25 ppt seawater at 20°C. This isolate was more refractory
to starvation and salinity effects than the preceding isolate
which was obtained from a human host. After 15 days in seawater,
there was no decrease in viable cell count in 10 ppt seawater,
whereas a 0.9 log unit reduction had occurred in 25 ppt
15
-------
seawater. Stress resulting from starvation at either salinity in
terms of prolonged ER time was not observed (fig. 11).
A study was conducted to determine if cell death in the test
media contributed to the increases in ER seen under different
conditions. The observed log number of E. cpli in the test
medium minus the predicted E. coli number calculated from the
viable count in seawater was plotted against time of starvation.
With respect to the urinary tract isolate, the greatest differences
between observed and anticipated bacterial levels occurred in EC
as opposed to TSB (figs.!2A and 12B) and were generally accompanied
by prolonged ER values (figs. 9 and 10). In contrast, the
estuarine water isolate, which showed no evidence of starvation
stress also showed little if any reduction in cell number upon
exposure to EC or TSB (fig. 13). Although the data suggested that
delayed ER times were related to cell death upon exposure primarily
to the selective medium EC at the elevated temperature, all the
results could not be explained solely on this basis. For instance,
cell death in EC after exposure to 30 ppt seawater at 20°C
and 30 C was approximately the same after 5 and 8 days starvation
whereas significantly prolonged ER times were observed only at
the earlier time interval. Therefore, growth in EC and TSB was
examined after various exposure periods to 25 ppt seawater at
20 C. Figure 14 illustrates that upon inoculation into EC,
starved cells evidenced -a decline in viable count for approximately
1 hour with the lag period extended an additional hour after
2 days but not after 8 days starvation. Similar effects were not
observed in TSB nor was there an apparent effect on growth rate in
either medium.
A series of experiments were initiated to evaluate the impact
of prestarvation history of the inoculum on viability in seawater
and on the post-starvation behavior in test media at 44.5 C. The
urinary tract isolate was grown under the following set of condi-
tions: (a) exponential phase in minimal medium (M-9), (b) sta-
tionary phase in a rich medium (TSB), and (c) exponential phase
in a rich medium (TSB). Treated cultures were then inoculated
into 25 ppt seawater at different temperatures. Exponential
phase ceils from M-9 and stationary phase cells from TSB showed
similar viability profiles (fig. 15). Maximum cell death
occurred during the first 3 days of starvation and ranged from
0.7 to 1.2 log units. In contrast, exponentially grown TSB cells
were more sensitive to starvation with a reduction of 2.3 log
units over a 3-day period at 20 C and 4.2 log units after 1 day
at 35°C (fig. 15C).
Similarly, cells grown exponentially in TSB, followed by
starvation in seawater, showed an aberrant post-starvation
behavior in test media at 44.5°C. After starvation in 25 ppt
seawater at 20°C, TSB-grown cells evidenced greater die-off when
introduced into either TSB or EC at 44.5°C than was previously
16
-------
observed with M-9 grown bacteria (figs. 12A and 12D). Use of
PBS at room temperature as diluent for serial dilution and
plating may have contributed to this apparent die-off. For
the cells starved at 35 C, TSB prewarmed to 35°C was used as
diluent, and this apparent cell die-off was not observed (fig. 12C).
Stress as measured by an increase in observed minus predicted ER
time was more obvious in the TSB pregrown culture starved at
20°C than at 35°C (figs. 16 and 17). This stress occurred earlier
than in the M-9 pregrown cultures, as soon as 5 hours after
exposure to seawater. The degree of sublethal stress at 35°C
was probably obscured by the rapid loss of viability.
The effects of cold-temperature starvation on viability and
sublethal stress were determined for an E. coll isolate obtained
from human feces. It should be noted that this culture differed
from the other isolates in having been maintained on artificial
media approximately 6 months prior to experimentation while other
isolates were used immediately after isolation with laboratory
maintenance not exceeding 2 months. The fecal isolate was grown
in M-9 and inoculated in 25 ppt seawater at.2°C. After a
3-week starvation period, no die-off was evident with the maximum
reduction of 0.4 log units occurring by 12 weeks. However, ER
times increased dramatically, especially in TSB indicating the
existence of sublethal stress (fig. 18). The increase in ER
represented increased lag time and/or decreased growth rate since
very little cell die-off occurred when the starved cells were
introduced into test media at 44.5°C (fig. 19). Growth of this
same fecal isolate in EC and TSB following starvation in
25 ppt seawater at 20°C revealed a longer lag time in TSB as
starvation progressed (fig. 20)„ In contrast, growth curves
of the urinary-tract isolate following starvation revealed increased
lag times in EC (fig. 14). These data illustrated that length
of exposure to artificial media was an important aspect to be
considered in conducting seawater survival experiments.
Standard curves were developed for use in detecting stress
in field isolates (figs. 21 and 22). For these studies, one
part TSB pregrown inoculum was added to one part single-strength
Medium A-l or EC at 44.5°C. The ER times in Medium A-l were
shorter than those in EC for similar inoculum levels with time
differences decreasing as the inoculum size increased. Introduc-
tion- of a larger volume of inoculum, resulting in dilution of
selective agents as well as reduction in the medium temperature
probably contributed to more consistent results and earlier
ER values as compared to previous standard curves in EC using
TSB grown inocula (fig. 6).
Similar results were noted when field samples were diluted 1:1
into single-strength test media. The ER times occurred 2 to 11
hours sooner in A-l than in EC with a mean difference of approxi-
mately 5 hours. When compared to the standard curve, field samples
17
-------
inoculated into A-l gave ER values extended by mean values of 0.6
and 0.9 hour for samples collected from waters with mean tempera-
tures of 14°C and 29°, respectively. Since the standard error
of the estimated y(Sy) at tQQ5 for the A-l standard curve was
+0.7 hour these results did hot indicate that the cells had been
greatly stressed at either temperature range.
Corresponding field samples processed in EC produced ER
times with mean deviations from the predicted times of +1.65 hours
for samples taken from cooler waters (X = 14°C) and -1.78 hours
for samples taken from warmer (X = 29°C) waters. Since the
standard error of the estimated y for the standard curve in
EC was +1.14, and since the number of sampling points was
limited, the question as to whether or not fecal coliforms in
our environmental samples were stressed remains unresolved.
DISCUSSION
Fecal coliforms introduced into the estuarine environment
are exposed to various physiochemical and biological fc.ctors
which contribute to viability losses. For example, dilution,
adsorption to particulates (ref. 12), sedimentation, sunlight
(refs. 6 and 13), organics (refs. 14, 24, 26, 30, 35), salinity
(ref. 9), temperature (refs. 12, 25, 36), dissolved oxygen
(ref. 12), heavy metal toxicity (ref. 17), algae (ref. 21),
bacteria (refs. 18, 22, 23, 29), bacteriophage (ref. 10), and
protozoa (refs. 11, 22), have all been discussed. In addition,
researchers have suggested that a portion of the coliform popula-
tion may be sublethally stressed following exposure to the aquatic
environment (refs. 7, 39).
Studies at the Virginia Institute of Marine Science have
been limited to the effects of three variables--salinity, ,
temperature, and previous cultural history on viability and stress
of fecal coliforms. Although the data showed that viability of
E_. coli was inversely related to salinity, a similar relationship
with respect to temperature was not as evident. Cells at 15,
20, 30, and 35°C showed similar viability profiles, although at
35°C, there was a continued loss of viability after the initial
sharp decline in bacterial numbers. A negligible loss of
viability .was demonstrated after 85 days of exposure to 2°C
seawater (25 ppt). Cryptic growth may have occurred thus
obscuring viability losses since the initial bacterial levels
in seawater in all of these experiments ranged from approximately
106 to 5 x 107 cells/ml.
The literature suggests that Enterobacter (as Aerobacter)
aerogens exposed to either distilled water or buffer solutions
at low temperatures displays loss of permeability control which
results in leakage of cellular constituents with a progressive
- 18 -
-------
and rapid loss of viability (refs. 33 and 34). In the absence of
pronounced cell death in our experiment at 2°C, it could be
hypothesized that the hyperosmotic environment provided by
25 ppt seawater protected chilled cells by decreasing the
diffusion rate of cellular constituents.
The most dramatic differences between viability profiles
were observed to be a function of cultural history. Cells
harvested during exponential phase from a rich medium (TSB)
were more sensitive to seawater than either exponential cells
from minimal media M-9, or stationary phase cells from TSB.
Although it is tenuous to extrapolate laboratory results to the
estuarine environment, our findings are of value to those
concerned with estuarine modeling studies involving the determina-
tion of coliform die-off coefficients. The previous history of
coliforms introduced to the estuary could affect survival. For
example, the physiological fate of 13. coli derived from untreated
sewage or other rich organic millieux, as opposed to immediate
storm water runoff, may differ when discharged into the estuarine
environment.
Little actually is known about the physiological changes
that occur in E. coli•in the estuarine environment. There is
no doubt that the organisms develop different levels of sublethal
stress depending upon the conditions encountered. Reports in
the literature suggest that difficulties in analyzing water
samples using membrane filter techniques rftay be due to the
fragility of sublethally stressed organisms. These organisms
may undergo mechanical destruction during the filtration
process or be killed by selective media used in the enumeration
procedure (ref. 15). New techniques have been devised to deal
with sublethally stressed organisms—some involved a period of
resuscitation in a nonselective media (refs. 7 and 37); others
use a sandwich technique involving nonselective and selective
media (refs. 28 and 32).
The electrochemical detection method (ECDM) assays the time
required for the inoculum to become metabolically active and
reach a critical cell density. Wilkins and Boykin (ref. 3)
suggested that ECDM might provide a rapid and inexpensive method
for monitoring coliform levels in the estuarine environment.
However, it was not clear how the presence of stressed cells would
affect the reliability of the ECDM results. The VIMS studies
were concerned with what conditions of salinity, temperature, or
previous cultural history would produce the highest levels
of sublethal stress as determined by ECDM. In addition, estuarine
water samples were collected for analysis by ECDM and by standard
MPN techniques to determine if stress in environmental coliforms
could be detected.
19
-------
It was obvious from the data collected that ECDM could detect
stress induced in laboratory coliform populations. Stress was
particularly evident in cells exposed to 25 and 30 ppt seawater
at 20°C. At 3QOC, stress was detected only in those cells
exposed to the highest salinity (30 ppt). When kept in seawater
at 2oc (25 ppt), cells developed significant stress only after a
long exposure even though there was little loss in viability.
Pronounced stress was also observed in samples of E_. coli pregrown
in a rich medium. The data suggested that prolonged ER times may
be due to immediate cell die-off as the starved cells encounter the
test media at 44.5OC and/or an increase in lag time.
When a limited number of environmental samples were analyzed
by the ECDM method, stress was not apparent. All of the VIMS
samples were collected in the immediate vicinity of a known
point pollution source. The organisms sampled were likely to
have had a short residence time in the estuarine environment
and this could explain the absence of pronounced stress.
20
-------
REFERENCES
1. American Public Health Association, Standard Methods for
the Examination of Water and Wastewater. Fourteenth
Edition, 1975.
2. Wilkins, Judd R., Glenn E. Stoner, and Elizabeth H. Boykin.
Microbial Detection Method Based on Sensing Molecular
Hydrogen. Applied Microbiology, vol. 27, no. 5, May 1974,
pp. 949-952.
3. Wilkins, Judd R., and Elizabeth H. Boykin. Electrochemical
Method for Early Detection and Monitoring of Coliforms. J.
American Water Works Assoc., vol. 68, no. 5, May 1976,
pp. 257-263.
4. Grana, David C., and Judd R. Wilkins. Description and
Field Test Results of an In situ Coliform Monitoring
System. NASA TP-1334, January 1979.
5. Andrews, W0 H., and M. W. Presnell. Rapid Recovery of
Escherichia coli from Estuarine Water. Appl. Microbiol.
23:521-523, WU.
6. Anson, A. E., and G. C. Ware. Laboratory Studies on the
Effect of the Container on the Mortality of E. coli in
Sea Water. Water Res. 9: 895-899, 1975.
7. Bissonnette, G. K.,J. J. Jezeski, G. A. McFeters, and
D. G. Stuart. Influence of Environmental Stress on
Enumeration of Indicator Bacteria from Natural Water.
Appl. Microbiol. 29: 186-194, 1975.
8. Braswell, J. R., and A. W. Hoadley. Recovery of Escherichia
coli from Chlorinated Secondary Sewage. Appl. Microbiol.
2TT328-329, 1974.
9. Carlucci, A. F., and D. Pramer. An Evaluation of Factors
Affecting the Survival of Escherichia coli in Sea Water.
II. Salinity, pH, and Nutrients. A"ppl. Microbiol. 8:247-250,
1960a.
10. Carlucci, A. F., and D. Pramer. An Evaluation of Factors
Affecting the Survival of Escherichia coli in Sea Water.
IV. BacteriophageSo Appl. Microbiol.~S72~54-256 , 1960b.
21
-------
11. Enzinger, R. M. , and R. C0 Cooper, Role of Bacteria and
Protozoa i-n the Removal of Escherichia coli from Estuarine
Waters. Appl. Environ. Microbiol. 31:758-763, 1976.
12. Faust, Maria A., A. E. Aotaky, and M. L. Hargadon. Effect
of Physical Parameters on the In situ Survival of Escherichia
coli MC-6 in an Estuarine Environment. Appl. Microbiol.
3T5T500-806, 1975.
13. Gameson, A. L. H. , and J. R. Saxon. Field Studies on
Effect of Daylight on Mortality of Coliform Bacteria.
Water Res. 1: 279-295, 1967.
14. Gerba, C. P, , and J0 S. McLeod. Effect of Sediments on
the Survival of Escherichia coli in Marine Waters. Appl.
Environ. Microbiol. 32 : 114-l20~7~1976 .
15. Hoadley, A. W. Effect of Injury on the Recovery of Indicator
Bacteria on Membrane Filters. In Symposium on the Recovery
of Indicator Organisms Employing Membrane Filters,
R. Bordner, C. Frith, and J. Winter, eds . , Environmental
Protection Agency, Cincinnati, Ohio. 1977, pp. 34-40.
16. Hoadley. A. W. , and C „ M. Cheng. The Recovery of Indicator
Bacteria on Selective Media. J. Appl. Bacteriol. 37:45-57,
1974.
17. Jones, G. E. Effect of Chelating Agents on the Growth of
Escherichia coli in Sea Water. J. Bacteriol. 3: 483-499,
18. Ketchum, B. H. , C. L. Carey, and M. Briggs. Preliminary
Studies on the Viability and Dispersal of Coliform Bacteria
in the Sea. In Limnological Aspects of Water Supply and
Waste Disposal, Publ. Amer. Assoc. Adv. Sci. , 1949, pp. 64-73,
19. Klein, D0 A. and S. Wu. Stress: A Factor to be Considered
in Heterotrophic Microorganism Enumeration from Aquatic
Environments. Appl. Microbiol. 27:429-431, 1974.
20. Milbauer, R. , and N. Grossowicz. Reactivation of Chlorine-
Inactivated Escherichia coli. Appl. Microbiol. 7:67-70, 1959.
21. Mitchell, R. Factors Affecting the Decline of Non-Marine
Microorganisms in Sea Water. Water Res. 2:535-543, 1968.
22. Mitchell, R. Destruction of Bacteria and Viruses in Sea
Water. J. Sanitary Engineering Division. Proceedings of
the Am. Soc. Civil Engineers. 1971, pp. 425-432.
22
-------
23. Mitchell, R., and J. C. Morris. The Fate of Intestinal
Bacteria in the Sea. In S. H. Jenkins (ed.), Advances in
Water Pollution Research. Pergamon Press, London, 1969,
pp. 811-821.
24. Moebus, K0 Factors Affecting Survival of Test Bacteria in
Sea Water: Marine Bacteria, Test Bacteria and Solid
Surfaces. Helgolander wiss. Meeresunters. 23:271-285, 1972.
25. Nusbaum, I., and R. M. Carver. Survival of Coliform Organisms
in Pacific Ocean Coastal Waters. Sewage Ind. Wastes 27:
1383-1390, 1955.
26. Orlob, G. T. Viability of Sewage Bacteria in Sea Water.
Sewage Ind. Wastes 28: 1147-1167, 1956.
27. Postgate, J0 R., and J. R0 Hunter. The Survival of Starved
Bacteria. J. Gen. Microbiol. 29:233-263, 1962.
28. Rose, R. E., E. E. Geldreich, and W. Litsky. Improved
Membrane Filter Method for Fecal Coliform Analysis.
Appl. Microbiol. 29:532-536, 1975.
29. Rosenfeld, W. D., and C. E. ZoBell. Antibiotic Production
by Marine Organisms. J. Bacteriol. 54:393-398, 1947.
30. Savage, H0 P,, and N. B. Hanes. Toxicity of Sea Water to
Coliform Bacteria. J. Water Pollut. Control. 43:854-861,
1971.
31. Simpson, G., A. Roe, and R. C. Lewontin. Quantitative
Zoology. Chapter 11, Correlation and Regression. Harcourt,
Brace, and World, Inc., New York, New York. 1960.
32. Speck, M. L., B. Ray, and R. B. Read, Jr. Repair and
Enumeration of Injured Coliforms by a Plating Procedure.
Appl, Microbiol. 29: 549-550, 1975.
33. Strange, R. E., and F. A. Dark. Effect of Chilling on
Aerobacter aerogenes in Aqueous Suspension. J. Gen. Micro-
biol. 29: 719-370, 1962.
34. Tempest, D0 W., and R. E. Strange. Variation in Content
and Distribution of Magnesium, and its Influence on Survival
in Aerobacter aerogenes Grown in a Chemostat. . J. Gen. Micro-
biol., 44: 273-279, 1966.
35. Vaccaro, R. F.f M. P. Briggs, C. L. Carey, and B. W. Ketchum.
Viability of Escherichia coli in Sea Water. Am. J. Publ.
Hlth, 40: 1257-1266, 1950.
•I. ; 23
-------
360 Vasconcelos, G. J., and R. G. Swartz. Survival of Bacteria
in Sea Water using a Diffusion Chamber Apparatus In situ.
Appl. Environ. Microbiol. 31:913-920, 1976.
37. Warseck, M0, B0 Ray, and M. L. Speck. Repair and Enumera-
tion of Injured Coliforms in Frozen Foods. Appl. Microbiol.
26:919-924, 1973.
38. Wilkins, J. R. Use of Platinum Electrodes for the Electrd-
chemical Detection of Bacteria. Appl. and Environ.
Micro. 36: 683-689, 1978.
24
-------
ELECTRODES-TEST
TUBE ASSEMBLY
STRIP-CHART RECORDER
BUFFER AMPLIFIER
WATER BATH
FIGURE 1, - EXPERIMENTAL SETUP FOR PERFORMING HYDROGEN RESPONSE MEASUREMENTS,
-------
S3
Rapid hydrogen
buildup period
i ii i i
24
incubation time, hrs
Figure 2. - Strip-chart tracing of hydrogen response curve for
1.9 X10° cells/100 ml
-------
o
JD
JO
TO
Symbols:
O Escherichia coli
D Enterobacter aerogenes
O Serratia marcescens
• Citrobacter freundii
• Citrobacter intermedium
+ Proteus mirabilis
O
45 6
Lag time, hrs
7
8
Figure
3. - Relationship between inoculum size and length of lag period.
(Correlation coefficients for the lines fitted by the method of
least souares.) ! 27 _
-------
8
3
O
01
\n
c
O
QL
CO
0)
e:
o
Q.
Log Number £. col 1/ml Media
Figure 4 Relationship between Inoculum size and the time of endpolnt response In TSB at
44.5°C when Inocula (urinary tract Isolate) were pregrown 1n TSB at 35°C. A
linear least-squares regression calculation gave a correlation coefficient of
-0.99 (N = 8) with an Intercept of 7.39 and a slope of -0.98. The standard
error of estimate at the 95 percent confidence leyel was ± 0.37. B28-
-------
^
iiiiiijiiiiiiiiiiiii'iiri
c ^. _^ j j „ _ .,
|^ 1 1 :
i __ T *f
:" 1 1 1 !: 1 1 1 1 : 1 1 i : 1 1 1 : 1 1 1 1 1 1 it 1 1 : i :: 1 1 1 1 1 1
h ' "i-
. ™ '
ri i
"S > t
7 . * r :, _ $ X
' i [ 1 1 1 . . i . i ? i . . i . ± - 1 1 . - . 1 1 1 1 1 1 1 1 1 : 1 1 1 1 1 ±
^ >
s s
"j!r S bi
NJ 5
*" S L.
j S
6_ m . _ . _ E S . . fc _ .. ........
................ j^-.-.b,. — .. — ........
. *" s
r" w k.
r* * . ^
s
^— ^ ^T" . 5i
51— ' f- *t ^
^j i..-_,.^..-~, .)— ^ ^~i
T * >.
ik A 1 .A. «A_ h S-s-
(/> --- - .-+-- _.{^r_-
-• N
O " ~ ' " "
_ _. .._..- . -.- T
c ........ -
__._.-. --i --••) .-.--. -i (-]-
... ..-,.._ ... - M. .
2 : if
• ^^^^^^B^^H
ffifflffflfM^fflJftfflasffl
n " ~ L ' -j___|-j_
[" — r j- --- T ij "in ~" —
ji . . .p _ . . ip _(i£ 1 :E 3 —
: - • : : : — i - - • ~ r j±
II:II:::::I:::II:::IIHII:I::III:J ci::i::±t
HI i;. i in. — ..I i n. j -------- -j=
-....-..-.-_.-_--._.-... — ------ -j __._j_
ii."iii.iii-i-iii.-in n iiiiiiinii~"iiit
i
21 I
™t*
"3L
it IE
.....£.. .-__£.
T""" -to. - = .=
... It
Ji -t
......... . . . ... .. ......It. . - - 1C
...III. Ill IIHIIIItllllt
i
" T T" " 31
T
. . I*. I .1C
± : . .1 .1 t
ii;:iii;iiiiiiiii:i::::::::::::i:±i:iiiii:i:±
-1
£ ^^L*
i -4-4—
^ l£jf-
TJ^
tl IE! 3-JC
Zc 3jC
"T* "-i-XZ
i IjC JL JXjL"
III jr^J^f^
i np DC£DC
Ixl 331X1
i" J^ "^J jLjCTf!
JU. -!— J- J—JnL,
ZL£ IXIJDC
nzxi IJLJLJU
3Z 333—
i nn ^OC]C
i j^ ijO™E
i JuTTT"
ill niijC
^c ^^^ct
3l ZSZ
3— -3Z
HL
-3-
3Z
— L. pJ= ^C
HC
IJI
nC
L_ IJ »A.
vXe
1- * *X«
31
!^I ^ . . _ .. ..IL ....__ — jJI — 3l
_!_.........- »^ — - -£4- - » - -4^
\ "T[ T^
_)_ JCi j_ Ju
^ 31
N 1
e ^ Zl_
1 1 : i ? ! ; 1 1 " i ; 1 1 1 1 1 1 1 1 1 : : 1 1 1 : : 1 1 1 1 : 1 1 1 1 1 1 1 1 : i - it :
~-«r .- «... ................... ......
.. -. ^ . ,. - ..X.__ .L^
\ *\ \. J
i. « t- — !
.' \ „ _ . , 4 r U" - „ - . .. . - - - .. . j
5:-j:::::.:K:±S!::i::::::::::::::::::i::"3
|-::;SM-;^:|;;;;;;-;p;;|
-j~ ' jr - - -T-r-j- 4 4- - - -h - - - : • : 5 - : - : - 5; ^ : - : - ±
j:j. .. _ .tj-j.- -j_y. --"--;> — J£J5 ;:::*viii..i
0 12 3456 78
Log Number £, coli/ml Media
Figure 5V Relationship between inoculum size and the time of endpoint response in TSB at
44.5°C when inocula (urinary tract isolate) were pregrown in M-9 medium at
35°C. A linear least-squares regression calculation gave a correlation coeffi-
cient of -0.98 (H = 22) with an intercept of 8.07 and a slope of -1.05. The
standard error of estimate at the 95 percent confidence level was + 0,54, -29
-------
O
O)
O
CX
10
-------
in :.::: i i ::::::::
1 u . , x
"
*
9. .. ;i :LIII : ::::::::
^- ---*--- "H-s ------
k
V i
. • * "
\
U_
T" * *
^""^ , "* ^
t/> . .. _ _ . . . ~j -J _fS - . _
3 7 : i 1:1 :i:i: "iiki in
i *
0) ..--. ""[I" -----
I/) i
c " " " "" "LJ1" * ~ - - - -
Q . -j. _ . ^_ +. . _,
CL 0 ' _._/
• '- 2 1
UJ ,
4b _ . ""~ ;_
.... . . . — . .. . . .
^
- - — - -i
a j . _ i _. _
— ----- — — - • -— ~f—
-
-_- - _ ... ..
2 — :h-
|™ "
_.._.. - -j..
........... — ,_,..._.
- - - - - •• - --i - — — -
[..._., .- — -.- - - -•
l .__. ^_,__
1 tnr-.-j---. ^r
l ., ... ....... — j. — ». . . - - -
___ ...... .- — j. j
— - - • -[ r "
- - - • - - j - - _p -
. _ . JTITI. ErainL. . 1
1 1 1 " ii : in lit
" "' TT -+r !' -^
. 1 J
i
I I " " ."*":' T IT L .
, ------- -; -1 -(->••
. t [ i . 1_ .. _
------ | t v .-•--••-
S I I
i J u— II ~4>
p. j J
"t !
::: s ":::!: : :::" if :':::*:
S i
Ij
S S ^ J_J
s 1 J
Li_ ^ -1
X
i U"
> i
^ ^
k S« 1
— .. V^J ' vj-- •
_^-_. ........ ^ --J- --«,-•-
"T ~1 i N
I ^
1 "^ L
"^ ^
"t "T i J
"" i
lj^_
i 1" lit — zr -
• :: IT T ± .
i_J_
. - t
.......... ^« .. . | ,.
\ i
- - * «------- ,...-. -|
.
i
i
— 1 -• ±T""""- '- H
T -i
* T " .rr ' 7_.
• ---••-• - -t ••, '11 — '^
—^ - -tU.. . "1-1— » - - -T- -|— \
[4 j . -i-ff - (•-•- i -f "T" - • 1 • -
.... J. ..^ 1. f+ . ..j-4 (.. . 4- -,.
.,..-. -j-p-r- ..j-- -t-- f- • ; -
- - - J+ ]• ±'Tt tti"i T-A " -i
f - - .• ]• -j-t- -r
*•-••• T± T' '1 1 T" 'jri' 1 _ :" ; r
T X -U 44^- 4-4--i4- 4-
-jr ~ _r 'T i j T J. 1 1 I { j
±t±t.a tl a .U liLJ i^il i. UJ
... . .
_ . ~|- ..
4
L?
>
,
S -4
V
is "
_. - - - s
S i_
( ^
_J__
C-T
~ T " :
!
"""t "
"i-|*- •
—4
4"!+.:
..|_ .
i
."t -
1
j
s
1
I
!
i
11
ISJ
I
wt- ^
N ^
S
...,.>- ........
. 31 r . . 4-L .
. : ; r _ _
. S. _ . .
^ i
\ tr ' ^
1 S> J
j i . ^
'. '.. . T '--'£.'- "• '.- \ '. ---:
~ ' ~ ' " ^ - . - i - - -f-
~ ~-~." „ ".LA:: s ~ i .
::::n:::::::;::::is:
- f- T-ff ----, <
s
LLiiii" .0,1 li.Hl J±LL
I . •* r Tr
k-.--..._ . ....... -4-1-
::::::::::::::": :~ "J
t ::±
___..-. _.. _ JQL
DC
™r"
it
........... _ . .
1_ * "
III
IJL
III
III
IJC
..„„
1 1 1 1 1 1 1. .. it
. _ ±r
DIE
f ±r
T^ "
'
f
"T"
^
^
L ^
t- b r J
. _._ _.>. -- --....If
S I]
LLI M 1 KIIM IN 1 Ml ii 1 1 1
1
8
Figure
Log Number £. coll /ml Media
Relationship between inoculum size and the time of endpoint response in EC at
44.5°C when inocula (urinary tract isolate) were pregrown in M-9 medium at
35°C. K linear least-squares regression calculation gave a correlation coeffi-
cent of -0.97 (N = 20) with an intercept of 9.41 and a slope of -1.25. .The
standard error of estimate at the 95 percent confidence level was + 0.73.
-------
i.
V
4->
to
10
(U
I/O
UJ|
t.
o
A—A Expt. 1
A—A Expt. 2
D D Expt. 1
33 •—• Expt. 2
20°C
30°C
6.0
8 9
Figure 8. Comparison of viability of E. coli at 20°C and 3QOc at various
salinitieso Prior to starvation the isolate (urinary tract)
was pregrown in M-9 medium.at 35°C0 Salinity: A=10 ppt.
B=15 ppt? C= 25 ppt, and D = 30 ppt0
32
-------
10
3
o
•a
01
-t->
o
•r*
•o
0)
CL
(/I
3
C
•o
at
01
.Q
O
Figure
8
1
Days in Seawater
Salinity
0
A
n
-n
-M
10
15
25
30
ppt
ppt
ppt
•ppt
Effect of starvation at 20°C in seawater of various salinities on the
endpoint response (ER) of £. coli inoculated into either EC or TSB at
44.5°C. Prior to starvation the isolate (urinary tract) was pregrown
in M-9 medium at 35°'C. Viable counts determined using PBS diluent at
room temperature.
33
-------
3
O
•o
O
•r-
•o
L.
-------
10
0 5
Days in Seawater
10
15
Salinity
O O 10PP±_
° D 25 ppt .
Figure 11. Effect of starvation at 20°C in seawater (10 ppt and 25 ppt)
on the endpoint response (ER) of E. coli inoculated into
either EC or TSB ,at 44.5°C. Prior to starvation the
isolate (estuarine water) was pregrown in M-9 medium at 35°C.
Viable counts determined using PBS diluent at room temperature,
35
-------
0)
UJ|
T3
U
•r—
TJ
0)
>
-------
•o
O)
UJ|
•a
a;
4J
o
•o
-------
(O
•r-
•o
-------
i.
-------
80 0 20
Hours in Seawater
80
Figure 16; Effect of starvation at 20°C in seawater (25 ppt) on the;endpdint
response (ER) of E.. coli inoculated into eith'er EC or TSB at 44.5°C.
Prior to starvation the isolate (urinary tract) was pregrown 1n TSB
at 35°C. Viable counts '1n both experiments determined using PBS at
room temperature. Expt. 1 D a and Expt. 2 a 9 .
40
-------
s.
o
•a
-------
3
O
TJ
•V
T3
a>
in
-1
0 10 20
40
60
80 0 20
Days in Seawater
40
60
vieure 18. Effect^ of starvation at 2°C in seawater (25 pp£) on the endpointt
" Q- " ' " xnnx 'f w _ _ -i • i ." T'~* J J_j «. j *-l. -T-I •CT' ^T— Tt'TJ «*- 7T7T~ ^T
response (ER) of E. "coll"'iiio'cuTated into either EC or TSB at
Prior to starvation tRe isolate (feces) was pregrown in M-9 medium at
35°C. Viable counts determined using PBS diluent at room temperature,
42
-------
10
20
30
40 50
Days in Seawater
60
80
Figure
Effect pf^ starvation at 2°C_ in seawater (25._pjpt)_on the enumeration -
of E. coli iramediateiy following a tenfold dilution into either EC or
TSB at 44.5°C. Viable counts determined using PBS diluent at room
temperature. ._ .
43
-------
8.0 t
10
•o
Ol
'' ' J *"!"' n ' ' •!L! I_" "
•f trtir-: r 1 T^ -/Tfe' H tWOT
:iiStiiilH.JjjnN--ffig:
—-h •••!—K--t-t-i—: T-J-TJ I--t-f•—"I i i-H1 •
oXi; - J. Hi-.jini ...n.. .iiU u.—
Pitffipfflai
it: ilf-f ^ + Xj.jl4.;.^4.|
Time 1n Seawater ^;.;ix
O o 20 h
• 67 h i-H
: . .!....: • i... : i [ .•,.i. . .. i
•h: 'iiri:i:-!. = ; ;!.•;•!' •!-f
0 I 2 3 4
Hours Incubation
Figure 20.
Growth of E. coli in EC and TSB, 44.5°C after 20 and
67 hours starvation at 20°C in seawater (25 ppt). Prior
to starvation the isolate (feces) was pregrown in M-9
medium at 35OC.
44
-------
H " i i H4 ' : ! • W-i- •!•! H • -f
r^r.-li ha' LI r. .Th-|-_nj..j:llT.
H'H ''• -F-M-H-: H! if iKlff
i I ' I ' I I II I I I
ffi:pM±^^4[-f-
^•w m$i i ji h.
i:j±j-J>mjij.t J-tJ.
••; j4 ••'--' ••• H-f-j-1-, :. ! r i'
V i .1 ' --!-.-•• . \ ,44-V-' : > •• (•
Figure 21. Relationship between inoculum size and the time of endpoint response in
half-strength Medium A-l when inocula were pregrown in TSB at 35°C. A
linear least-squares regression calculation gave a correlation coeffi-
cient of -0.97 (N = 12) with an intercept of 6.71 and a slope of -0.84.
The standard error of estimate at the 95 percent confidence level was ±
0.69,
45
-------
12
11,
10
FH
~ 8
t
tt-tt
nz:X3''n.7.',"T!T.
ffl:
XLi±
•tie
:ito
• rr ,-|-
zEtP
^:
-)- t~l- -
:o-:.u:
tit
i~i"i"T"!~i l~f TTTTTT't ! *"!~ri'
* , ^4^t^tFta-{f>.ri-
2 7.|LL«^;EaMl^:
a Pffl-pt^fm^-"
c 6
-t
T\r
•vrh
ItLM^fLiMll
;: "M.icha";•."HA!yr- tv IT
grq-rgrTTttbi^b ;^|^fI ^Swtl"-
Jli-^it
The
'.
4 jp
,ffl
33
M r-v-rr-r
te
ss
r T
1 , ^V IT1 ; IT' I T I ( f ' i i
"Hxi-qrnr-fT-w-tt-^i
jt^VrfitN-iillhptt^
i . ; ' I , • , i
-vrj-1-:1;-!!!!!'!':'.;-
3.j±LL:.Hr,^.Ui:.].;|:
2 j**mmm
1;-.-,-' !ri-H
' a..:.'.
i. : i :.i i.;.'....: '.j.u ! . i i.! i I-.:. l-l-u-'. .;.-:•' ' '--j1 •.! U- v.. . | i.:v. >
i .._!_... :.;.L.L. u.-4 -j. i i.,.'.« ;.; a. i. i., i L... .; ... i : ;-.,.'. • . ;_|.-|...,., •-. -j. ,.\ ..
j,-....,.. j-|-|-f -::1 -{ j'|- •[-[ •rn |_pp_ji[' ,_!.i < J_Li. i 1.'.'. i j ;\ ' L_r\'L
|r'^^---;:|:|il|:r[J i:jj j;iJj:i±{-iT'±n :'• fv t.l"l I i U li^FE}-:-^
4::
K
-trr.
it
u
-r, H
m-i-
-v-r-,-4-
Si
Et
f-U
1
iii
H-l-l
1
:qq-
•VT
• T--
at
:Jt
n-
uttt
:j-r.r:
.ij_i--
TTft
--T-1
«
f.; !:::. ±.fi.id
3 45 6 7
Log Number £. coli/ml Media
Figure 22 Relationship between inoculum size and the time of endpoint response In
half-strength EC at 44.5°C when inocula (urinary tract isolate) were
pregrown in TSB at 35°C. A linear least-squares regression calculation
gave a correlation coefficient of -0.99 (N = 9) with an intercept of
12.17 and a slope of -1.91. The standard error of estimate at the 95
percent confidence level was +1.14.
~" - 46 -
-------
1. Report No. NASA TM-80130
EPA-600/4-79-075
2. Government Accession No.
3. Recipient's Catalog No.
4. Title and Subtitle
Investigation of Effects of Temperature, Salinity
and Electrode Design on the Performance of an
Electrochemical Coliform Detector
5. Report Date
November 1979
6. Performing Organization Code
7. Author(s)
David C. Grana
8. Performing Organization Report No.
10. Work Unit No.
9. Performing Organization Name and Address
National Aeronautics and Space Administration
Langley Research Center
Hampton, VA 23665
11. Contract or Grant No.
EPA-IAG-D7-0053
12. Sponsoring Agency Name and Address
U.S, Environmental Protection Agency--Las Vegas, NV
Office of Research and Development
Environmental Monitoring Systems Laboratory
Las Vegas, NV 89114
13. Type of Report and Period Covered
Technical Memorandum
14. Sponsoring Agency Code
EPA/600/07
15. Supplementary Notes
This report was published by NASA Langley, with EPA and NASA numbers, under an
interagency agreement.
16. Abstract
This report presents the results of two research programs to determine
the optimum detector design for measuring fecal coliforms in saline
waters for operational systems. One program was concerned with the effec
of temperature and salinity on endpoint "response time, and the other,
the interaction between electrode configurations and the test organisms.
Test results show that the endpoint response time is related to
salinity and seawater temperature; however, these results can be minimize
by the correct choice of growth media. Electrode configurations were
developed from stainless steel, Parlodion-coated stainless steel and
platinum that circumvented problems associated with the commercial
redox electrodes.
17. Key Words (Suggested by Author(s))
Microbiology
Coliforms
Water quality measurement
Electrochemical sensor
Electrochemistry
Biological detection
18. Distribution Statement
Unclassified - Unlimited
Subject Category 51
(7D 15B)
19. Security Classif. (of this report)
Unclassified
20. Security Classif. (of this page)
Unclassified
21. No. of Pages
53
22. Price*
$5.25
* For sale by the National Technical Information Service, Springfield, Virginia 22161
------- |