United States
                         Environmental Protection
                         Agency
       National Health and Environmental
       Effects Research Laboratory
       Gulf Breeze, FL 32561
                         Research and Development
       EPA/600/S-98/005
May 1998
 &EPA      ENVIRONMENTAL
                        RESEARCH  BRIEF
       In Situ Bioremediation of Trichloroethylene Using Burkholderia cepacia G4 PR1:
               Analysis of Microbial Ecology Parameters for Risk Assessment

                                      Richard A. Snyder1
 Abstract
 The introduction of bacteria into aquifers for bioremediation
 purposes requires monitoring of the persistence and activ-
 ity of microbial populations for efficacy and risk assessment
 purposes. Burkholderia cepacia G4 PR1 constitutively ex-
 presses a toluene ortho- monooxygenase (torn) that aero-
 bically mineralizes TCE. Groundwater and sediment from
 a potential release site have been used in laboratory mi-
 crocosms to develop predictive models for the response of
 this organism.  In sterile systems, PR1 maintains stable
 populations for extended periods. In non-sterile systems,
 the bacterium is eliminated concomitant with an increase
 in bacterivores. The half life for the organism in non-sterile
 systems increases logarithmically with increasing initial in-
 oculation density above 1 x 106 PR1 ml-1. Below this level
 of inoculation, the half life of PR1  increases with decreas-
 ing inoculation density. The inflection point corresponds to
 a numerical response threshold for bacterivores.  In col-
 umn systems designed to mimic aquifer flow, repeated
 pulses of PR1 build up bacterivore populations reducing
 the half life of the bacterium for subsequent additions. Ad-
 dition of 0.5 nmM TCE in the elution stream results in pro-
 longed survival of PR1.  The results suggest that abiotic
 factors are not limiting to the bacterium in the target aqui-
 fer, but rapid losses from native bacterivores will occur.
 1 Center for Environmental Diagnostics and Bioremediation
 Department of Biology, University of West Florida
 11000 University Pkwy, Pensacola, FL 32514.
 Mention of trade names or commercial products does not constitute endorsement
or recommendation of use.
 Introduction
 The application of biotechnology to solve biological and
 ecological problems is becoming widespread in medicine
 and agriculture. The potential for use of genetically engi-
 neered  or altered  microorganisms (GEMS)  for
 bioremediation of aquifers contaminated with toxic materi-
 als is high, despite the problems of physical access to this
 environment to define the ecological parameters for opti-
 mal microorganism function and to determine the extend of
 any risk involved in releasing non-native microorganisms
 into unconfined groundwaters.  In  natural communities,
 bacteria are subject to starvation, competition, predation
 and viral infection pressures in addition to limits imposed
 by physical and chemical conditions. It is these selective
 forces that maintain bacterial populations at relatively stable
 numbers in the environment. For successful bioremediation
 with introduced microorganisms (bioaugmentation), selec-
 tive pressures must be altered within a target zone to main-
 tain the high densities and activity required for treatment. If
 the goal of the introduction is activity through growth of the
 GEM rather than using GEM biomass as a static reagent,
 then the carrying capacity and potential niche space within
 the target environment for the GEM must be considered.
 Beyond a treatment zone where niche space may be al-
 tered for GEM survival, persistence  and spread of GEMs
 and their genes require integration of the foreign bacterium
 (or genes) and  displacement of species from existing mi-
 crobial communities that have been  derived by the selec-
tive pressures unique to that environment.

                        *gO Printed on Recycled Paper

-------
Analyses of population dynamics, trophic interactions and
microorganism transport are critical for the application of
successful bioaugmentation and require means to monitor
both organisms and activity. Risk assessment associated
with the introduction of non-native bacteria also requires
a careful  monitoring of the survival  and dispersal of
released microorganisms and altered  genes.  Although
releases of non-native or recombinant bacteria have not
been reported to result in adverse environmental effects
to date, there is a responsibility to ensure that released
microorganisms  will  be constrained  by the selective
pressures of  the  target environment. Monitoring  of
released populations of microorganisms thus aids in the
final evaluation of both success and safety.

The use of chloroethenes including  trichloroethylene
(TCE)   has  led  to  an  extensive  contamination  of
groundwater  resources in the United States.  In situ
bioremediation of this contaminant may be possible by the
aerobic microorganism Burkholderia  cepacia  G4.  B.
cepacla G4 PR123 and PR131 constitutively express a
toluene ortftomonooxygenase (torn) due to a secondary
transposition  of  Th5 sequence into a TCE  degradative
plasmid (TOM) (Nelson etal., 1986; Shields etal., 1992).
TnSmutagenesis also confers kanamycin (km) resistance
to the organisms. PR123 contains an IS50R in the TOM23C
plasmid. PR131 contains a single Tn5 in the chromosome
and an IS50R  in the TOM31C plasmid.   The  IS50R
elements in the plasmid of both strains are at nearly the
same locations and are thought to be responsible for the
constitutive expression of the torn gene.

This paper presents laboratory analysis of the behavior of
Burkholderia cepacia G4 PR1  (PR1) in simulated aquifer
conditions to address both function and fate questions for
use  of  this  microorganism  in  remediating  TCE
contaminated ground water. The laboratory analysis was
targeted for a release at the "Borden Aquifer," a shallow
sandy  aquifer located on the Canadian Forces Base
Borden, Alliston, Ontario.

Procedures

 Media,  Growth, and  Cell Counts.  PR1 was
 maintained as a frozen stock and on a basal salts medium
 (BSM) with 20 mM lactate and 1.5% agar in petri plates for
working cultures. For experimentation, cells were grown
 in batch cultures in BSM 20 mM lactate, glucose, phenol,
 m-cresol, or phthalic acid.  A commercial spring water,
 Georgia Mountain Water (GMW) was used in place of a
 limited supply of Borden aquifer water where bulk use was
 required. Cell cultures were  routinely checked for TOM
 activity using the  TFMP  assay (described below).
Enumeration of total CPU in aquifer samples was done
using Difco R2A agar.

Direct counts of total bacteria and protists in formalin or
Lugol's fixed samples were performed using the DNA
fluorochrome DAPI (Porter & Feig, 1980; Pomeroy, 1984).
Minimum numbers for direct counts of bacteria were 10
microscope fields and 200 cells. Minimum numbers for
protist DAPI counts were 10 fields and 100 cells.  Plate
counts of culturable bacterial cells were  performed by
spread plates (50-100  /2\)  of serial  dilutions. Where
possible, plate counts within 10-100 CFU/plate were used.

Tracking Methods.  Four methods of tracking  and
confirming PR1 have been developed.
1) Selective plating was based on phthalic acid, phenol,
and cresol utilization and the presence of kanamycin (km)
resistance associated with Tn5.
2) A monoclonal antibody specific to PR1 LPS has been
used for direct  fluorescent  counts of  PR1   from
suspensions and for identifying colonies on filter blots of
agar plates as previously described (Winkler et al., 1995).
3) A colorimetric reaction (triflouromethyl phenol, TFMP
to triflouroheptadienoic acid, TFH A) was used to assay for
the toluene monooxygenase enzyme that degrades TCE
(Shields et al., 1991). This reaction  was used for both
activity  determinations  in  cell  suspensions  and  for
identifying colonies on plate blots. For cell suspensions,
absorbance  at 600 nm (A600) was  determined for 1  ml.
Cells are then pelleted and resuspended in 1 ml of 1.0 mM
TFMP in 10 mM Tris-CI pH 8.5.  The suspension  was
incubated in a 25ml Erlenmeyer flask at 30°C for 20 min.
Cells were repelleted in a 1.5 ml Eppendorf tube to
transfer the cell-free supernatant to a microcuvette and
read absorbance  at 386 and 600 nm.  To base the
reaction on protein, a conversion factor of A600 x 0.290
provided an estimate of mg/ml protein. To convert to [M
TFHA, absorbance at 385 nm was multiplied by 0.0269.
For  distinguishing between colonies on agar plates, filter
paper with colony blots was soaked with 1.0 mM TFMP in
10mM Tris-CI pH 8.5, and any color change recorded.
4) A nucleic acid thermocyler amplification assay using
primers  unique to the Tn5 insertion  site  has  been
developed and tested. The primers were designed based
on the assumption that the insertion points would be
unique for PR1.  This assay was performed for  both
extracted DNA from sediment samples and as an assay
for whole cells in ground water suspension.

 Protist  MPN.  PR1  was grown on BSM lactate km
 medium to stationary growth phase, pelleted, and washed
free of metabolites and resuspended to an optical density
 (OD) of 0.2 @ 480 nm on an HP  spectrophotometer.

-------
 Using  Corning  24-well culture  plates,  1  ml  of  the
 suspension of bacteria was placed into each, with 8 wells
 per dilution of  sample and six dilutions  (10"1 to 10"6).
 Approximately 1 ml subsamples of Borden aquifer material
 were removed from core material of the saturated zone
 and  placed in  pre-weighed 15 ml disposable  Falcon
 centrifuge tubes.  Each sample tube then received 10 ml
 of filter sterilized Borden aquifer water (BAW).  Dilution
 tubes (5 each) for each subsample received 9 ml of BAW.
 Tubes with the soil/water sample were vortexed well for
 approximately 1  min and allowed to stand for approxi-
 mately 20 sec to allow larger sand grains to sediment.
 Eight 1 ml portions of the diluted sediment were distributed
 to 8 wells in the 24-well plate to establish  a 10'1 dilution
 with a final prey density of 0.1 OD. This was repeated for
 serial dilutions made from the initial 10~1 suspension.
 Initial dilution  tubes containing sediment were dried and
 weighed to obtain the  gram dry weight (gdw) of each
 sample.

 Plates were examined every two or three days starting on
 day 4 and extending through day 28. Plates were viewed
 on an inverted microscope (Zeiss) and scored positive or
 negative  for  flagellates,  ciliates,  naked  and  testate
 amoebae.

 Viruses.  An  attempt was  made to  isolate viruses
 capable of infecting PR1 from Borden aquifer sediment.
 Sediment  was sterilized with  chloroform  vapor  under
 vacuum for 1  week to eliminate native microorganisms,
 then allowed to air dry for 2 days.  PR1 was grown on
 BSM 20 mM lactate medium, washed and resuspended in
 sterile spring water to an OD of 1.0. The PR1 suspension
 (10 ml)  was placed into a Falcon 15 ml centrifuge tube
 with 10 mM lactate and approximately 2 ml of the air-dried
 sterile sediment.   This  sediment/PRI  suspension was
 mixed on a rotatory wheel for 2 hours, followed by a 1
 hour settling time. The supernatant was then mixed 50:50
 with BSM 10 mM lactate with 0.75% agar and poured over
 BSM 10 mM 1.5% agar in petri plates. These pour plates
 were monitored for 13 days for clearing zones (plaques)
 indicating lysis of PR1.

 Groundwater Survival.  Survival in sterile and non-
 sterilized Borden ground water was determined by plate
 counts  (Phthalate/Km),  4',6-diamidino-2-phenylindole
 (DAPI) counts and direct immunofluorescence counts
 (fmab) and was  previously  reported  for a  30-day
 incubation  (Winkler et al., 1994).  Mab blots of plate
counts and DAPI counts of this presterilized treatment
were subsequently obtained after a 7 month duration.
  Anaerobic Effects.  Survival of Pm deprived of
  oxygen was monitored in suspensions of cells in sterile
  ground water bubbled  with  0.2 ^m  filtered N2  gas.
  Incubations were maintained in a constant temperature
  bath and continuously agitated with  magnetic stir bars.
  Temperature   and   oxygen  concentrations   were
  continuously monitored using Nessler oxygen probes and
  meters connected to a 386 MS-DOS computer fitted with
  a data acquisition card.  DAPI direct counts and CFU
  response were determined over time.

  Extinction  Rates  in   Sediment  Slurries.
  Sediment slurries were set up with approximately 1 0 g wet
 weight of Borden aquifer sediment and 30 mis of Borden
 aquifer water in 125 ml flasks or 100 ml serum bottles.
 Repeated experiments were conducted at various initial
 densities of PR1 ranging from 9 x 103 to 9 x 10s cells ml'1
 of slurry. A typical experiment compared uninoculated
 controls, flasks receiving nutrient addition only, PR1 only,
 and PR1 plus a nutrient.  Other experiments compared
 different levels of inoculated PR1 only. Slurries were kept
 at constant temperature (1 8 and 1 0°C) and mixed at each
 sampling point.  Numbers of PR1 were monitored over
 time by plate counts, confirmed by monoclonal antibody or
 TFMP blots.  DAPI counts  of both total bacteria  and
 protists were recorded.

 Extinction rate constants were determined by regression
 analysis as the slope of natural log-transformed loss data.
 Extinction  rate  constants  (k) were  determined  by
 regression  analysis as  the slope of  linear portions of
 natural log-transformed cell loss data using the formula.
                       t,-t
These estimates included both growth  and predation
losses of the bacterium.  Dividing these extinction rate
constants by 0.693 (natural  log of 2) provides a time
interval for reduction of the population by half (half life).
These half life  estimates were  then  plotted  as an
inoculation-density dependent  function  and fitted to a
predictive regression model,  providing an estimator for
survival times of PR1 in the aquifer.

Growth responses of the bacterivorous protist assemblage
in Borden sediment were also determined from cell count
data and growth rate constants calculated similarly as the
extinction rates. Growth rate constants were plotted as a
function of inoculated PR1 density.

-------
PR1 Response  in Flow  Columns.  Sediment
columns were established to mimic the groundwater flow
at the Borden site. A chromatography column was fitted
with cut GC vials closed with Teflon septa to provide
sampling ports. Teflon tubing and fittings were used for all
flow lines containing TCE. Tygon tubing connected to a
constant temperature recirculating bath was used to jacket
the column and maintain it at 15°C. Flow was controlled
at the column outflow and set to the flow rate in the target
environment (2 cm day"1). GMW was used as the eluant.
Cells were added by syringe pump. An overflow ensured
constant supply of eluant, or cell suspension, to the top of
the sediment. Pore water samples were taken by syringe
(200-0 /ul  each)  and used  for  plate counts and direct
counts of bacteria and protists at each sampling port.

Typically, an inoculum of two void volumes at 10s cells ml"1
was added to the column, followed by continued flow of
GMW or GMW + 0.5 ^M TCE. Samples (200 ywl) from
each port were taken by syringe from the center of the
sediment to  avoid edge effects. Samples were split for
plate counts and preserved for direct counts. Samples for
TCE analysis where taken on alternate days to avoid large
disturbances of the flow field by removal of large volumes.

Potential Host Range for the TOM Plasmid.
Twenty-four random isolates on R2A medium from Borden
aquifer sediment were  used to determine the potential
spread of the plasmid from PR131.  Transformations were
attempted by standard direct filter matings.   Antibiotic
sensitivity  of the isolates and PR1 were determined to
provide for  selective  plating  of  donor, recipient and
transconjugant. Overnight cultures of PR131 were grown
in BSM-20 mM lac.  Potential recipients were grown in
R2A broth (Difco).  Five mis of donor and recipient were
pelleted separately.  Each pellet was resuspended in R2A
broth to A^ OD=1.0.  Viability of each was checked by
spread plate. One ml of each culture was transferred to a
5ml test tube and mixed.  The mixed suspension was
filtered onto  0.2 //m membrane filters with a presterilized
swinex and sterile syringe.  The filter was then placed on
an R2A agar plate. The filter was incubated at 30°C
overnight.   The filter  was  then  transferred to  a tube
containing  1ml R2A broth and  vortexed.   Resulting
suspension was plated on selective media. Selectivity of
media was checked by plating  an aliquot of donor and
recipient.    Putative transconjugants were tested for
positive TFMP reaction, indicating the presence of the
TCE-degrading  toluene  monooxygenase,   TCE
mineralization and presence of the TOM plasmid by the
nucleic acid  amplification assays.
Transport.  PR1 and transconjugants were tested for
their  relative transport  potential  in  Borden  aquifer
sediment.  Cells were grown on R2A broth medium and
harvested at stationary growth phase. Cells were pelleted
and  washed  free of  medium  and metabolites   and
adjusted to 0.1 OD @ 480 nm.  Columns were set up
using 2 ml glass syringes filled with sterilized Borden
sediment packed under water saturation and vibration.
Flow was controlled at the column outflows and set to the
flow rate in the target environment (2 cm day"1). One void
volume of cell suspension was added to the columns and
chased by sterilized Borden aquifer water.  Fractions of 3
drops each were collected at the outflow in 96-well tissue
culture plates. Fractions were serially diluted and plated
on selective media. Output concentrations were plotted
as a percentage of the inflow concentration to determine
relative transport propensity for the different bacteria.

Results and Discussion

Thermocyler   Amplification   of   TOM
Sequences. This assay provided a definitive test for
the presence of the PR1  degradative plasmid.  Lower
detection limits determined by dilution of cells in Borden
water and sediment slurries yielded approximately 1 x 102
and 1 x  103 cells ml"1, respectively.

MPN for PR1  Predators.  MPN data for PR1-
consuming bacterivores from four separate subsamples of
Borden aquifer material indicated numerous protists in the
aquifer  system. Ciliates were  rare, but recorded at a
mean (± standard deviation) MPN value of 0.678 ± 0.255
gdw"1. The ciliate observed is apparently an undescribed
species of Hymenostome and is the first report of a ciliate
from aquifer sediments. Means and standard deviations
for other protists were: flagellates, 1.71 x 104± 2.13 x 104
gdw"1; naked amoebae, 2.20 x 103 ± 1.49 x 103 gdw1; and
testate amoebae 3.22 x 102 ± 1.77 x 102 gdw"1, indicating
a  prevalence of potential PR1  predators in  the target
environment.    Attempts were made  to isolate  PR1
viruses, but none were recovered.

Groundwater Survival. PR1 was recoverable after
7 months incubation in sterilized aquifer water, but was
quickly eliminated in non-sterile water. Response of PR1
to anaerobic conditions indicated little effect of oxygen
deprivation on PR131 culturability through 2 days at 20°C
and  through  25 days at 15°C,  although at  reduced
numbers. Direct counts with the DNA fluorochrome DAPI
remained   constant  through   the  incubation  period
indicating  the cells and DNA of PR1  remained intact
despite a loss of culturability. This finding is significant in
that  most  aquifers   are  anaerobic,  and  prolonged

-------
 persistence of the bacterium beyond an aerobic treatment
 zone  will  not  be limited  by  anaerobiasis.    These
 preliminary experiments suggested that abiotic factors of
 the  target  environment would not be limiting  to PR1
 survival, but  biological  interactions were potentially
 important.

 Potential Host Range of TOM Plasmid. Out of
 24 random isolates, 10 (42%) were positive for TOM31C
 transfer by detection  of TOM target sequences.  Tom
 enzyme activity was present  in  80% (8 of 10) of the
 transconjugants as indicated by a positive transformation
 of trifluoromethyl phenol (TFMP) to trifluoroheptadianoic
 acid (yellow product) and mineralization of TCE.  These
 data indicate a wide potential host range within the target
 environment for TOM, highlighting the need to ensure
 tracking  capability for both  the  organism  and  its
 associated genetic elements.

 Transport of PR1 and Transconjugants. Five
 out of seven positive transconjugants had greater relative
 transport through Borden sediment material than did PR1.
 These data suggest  that  plasmid transfer within the
 aquifer may be a major factor in determining the spread of
the novel DNA within the system.  Detailed  analyses of
transport phenomena associated with PR1  and  Borden
 aquifer material  have  been  presented   elsewhere
 (Lawrence  & Hendry, 1996,1998; Hendry et al., 1997).

 Extinction Rates  in  Sediment Slurries,  in
aquifer sediment slurries,  numbers of native  bacteria
 isolated on ph/km medium varied below an upper limit of
approximately 5 x 105 CPU  ml'1.  Direct epifluorescence
 microscopy counts of bacterivorous protists were below 8
x 102 ml"1 in unamended incubations.  Where PR1 was
added, selective plate counts, as confirmed by monoclonal
antibody blots, decreased to zero concomitant with a rise
in protist numbers. CPU of phthalate-utilizing  bacteria
increased  following  protist   mineralization  of  PR1,
suggesting that  general enrichment of the  system will
increase population levels  of other bacteria.  None of
these isolates,  however, were found to have acquired
TOM plasmid from PR1  in any of these experiments.
Addition of  substrate with PR1 (phenol or phthalate) had
no significant effect on PR1  survival.

Total protist population response increased with the
decrease in PR1-at inoculation from 1x10s to 5x108 cells
ml"1.  Maximum protist response did not increase when the
inoculation  density was raised from 5 x 108to  1 x 109 cells
ml"1. However, maximum protist population response was
sustained for a  longer duration, and precipitous loss of
PR1 does not occur at the 1 x 109 cells ml'1 inoculation
density until after 22 days of incubation.

Extinction rates fit well to a logarithmic curve, providing an
estimated life span for PR1 within a treatment zone, and
indicate a threshold of 1 x 108 ml"1 to 2 x 108 ml"1, above
which further increases in survival rate with increased
inoculation density is  minimal. In addition, extinction to a
lower threshold of 1x 107  bacteria ml"1, the level of PR1
considered necessary for significant TCE mineralization
activity, is shown in  Figure 1.  This  predictive model
provides an estimated cycling time  for maintaining high
densities within a treatment zone.
                y = 2.3069 + 8.7028e-Q8x R= 0.98832
        0 10"   1 10B  2 10   3 10   4 10   5 10

                Inoculation Density PR1  ml'1
       Figure 1. A predictive model for maintaining PR1
       populations above a 1 x 107 cell ml'1 threshold.
       6-r
                
-------
The lowest survival rates occurred at inoculation densities
around 1 x 106 PR1 ml"1. Inoculation densities above and
below this level resulted in increased survival times. This
phenomenon is shown in Figure 2. The inflection point of
lowest  PR1  survival  coincides  with  a threshold of
response for bacterivores, as indicated by the numerical
response curve for Borden aquifer protists grown on PR1
shown as Figure 3. This growth response of the native
bacterivore assemblage has its greatest range of increase
at 10s PR1 ml"1 and above. Thus, inoculation at or above
the inflection point stimulates bacterivore excystment and
numerical  response,  accelerating  bacterial  losses.
Inoculation below this point, as would occur as a plume of
bacteria leaving a treatment zone enters the downstream
aquifer,   would  result  in  bacterial cell  losses being
restricted to encounters with existing active bacterivores,
and survival rates are extended. Given that the densities
of native bacteria  in the ground  water are below the
inflection point,  most bacterivores in the unamended
aquifer  are  probably  either encysted or just meeting
maintenance energy requirements. At low PR1 densities,
predator-prey contact rates in the sediment slurries would
have  been conceivably higher than  contact rates  that
would occur in the undisturbed  aquifer, making these
estimates an upper limit to  removal rates in the natural
system.
     0.45-


     0.4-


     0,35-


     0.3-


     0.25-
        i

     0,2-


     0.15
-y x -0.095211 + 0.056452log(x) R= 0.84192
              2 10*
     4 10s    6 10*
    Initial Bacteria ml"'
8 108
                                             1 10"
       Rgure 3.  Bacterivore numerical response from
       sediment slurries.
PR1 Response  in Flow Columns.  With  an
inoculum of two void volumes at 1x 10s cells ml"1, a
population above 107 CFU ml"1 pore water was maintained
for 5 days at the top of a column. Native bacterivores
responded  to the  addition of PR1 to the column  by
increasing protist numbers in proportion to the decrease
of PR1 cells.  A linear decrease in PR1  numbers with
distance  through the column was observed prior to
bacterivore impacts.  By days 8 and 10, the combined
effect of predation and elution decreased PR1 numbers in
the upper portion of the column.  By day 15, the pulse of
PR1 was eliminated at the upper and lower portions of the
column, leaving residual cells in the central portion. Most
of the  bacterivores form resistant cysts on  sediment
surfaces when  food  is not available, so a decrease in
protist numbers overtime is likely due to re-encystment of
these organisms after depletion of PR1.

Data from subsequent runs demonstrated accelerated
losses,  presumably due to  an  increased reservoir of
encysted bacterivores.  A run  with  TCE  added to the
inflow had greater persistence of PR1, with a sustained
population above 107 CFU ml"1 through 9 days of elution.
TCE breakthough did not occur until after the sampling at
15 days of elution. Half-life estimates for data collected at
the topmost sampling port indicate that survival rates are
equal to or lower than the survival rates determined for the
sediment slurries (Figure 2). The relatively slow transport
of PR1  through Borden sediment limits the effects of
elution on the disappearance of cells from the sediment,
but repeated inoculation increases bacterivore populations
and lowers survival rates (CM and CHI points). Addition of
TCE to the column (CIV)  appeared  to have  inhibitory
effects on bacterivores and  increased  PR1 survival to
similar rates as found in the slurries where bacterivore
impact was affected by the limitation of growth  response
to the  inoculated  bacteria.   The  toxicity of  TCE to
bacterivores may restrict the effect of the built-up reservoir
of encysted bacterivores. If this scenario is correct, then
as TCE concentrations are reduced, bacterivore impact
would increase, limiting the dispersal of PR1 in the non-
contaminated parts of the system.

This study has  shown that the abiotic conditions of the
aquifer do not appear to be detrimental to survival of this
organism, but  that biological interactions may be of
primary concern. Predation has often been demonstrated
to affect the removal of  allochthonous  bacteria  and
certainly is a factor where inoculation is  above the lower
growth thresholds for bacterivores.  Competition is also
often   cited   as  the cause of  inoculated   bacteria
disappearance. Competitive ability for available growth
substrates and  resistance to starvation would be strong
determinants of microorganism displacement and survival,
although the ability to integrate cooperatively into existing
microbial communities may be equally important. In this
investigation, competition was not investigated directly,
although the  increased survival  rate observed  for
inoculation  below  bacterivore excystment and  growth
thresholds  suggests  that  competition  may  be  less
important than predation. Competition between bacteria
may also be manifest in the ability to attach and colonize

-------
surfaces where growth  substrates are  more  readily
obtained, especially in habitats with low organic content.
Surface colonization and availability of micropore spaces
may also provide a refuge from bacterivores.

This work has sought to apply ecological principles in
defining the interactions that determine persistence and
activity of GEMS in the environment. Growth substrate
availability,  competition,  and  predation  are dominant
forces that  control bacterial densities and community
structure in natural habitats. The reactions of introduced
microorganisms to these  selective forces will determine
their persistence and activity, especially when the target
concentrations of bacteria  are above normal thresholds for
natural systems.

Acknowledgements

Technical assistance  was  provided by  Wendy  S.
Steffensen, John Millward, Sheree Enfinger, and Angela
Andrews.   This summary is  distilled from research
supported   by USEPA  Cooperative   Agreement
#CR822568-01-Oto RASatthe University of West Florida.

References Cited

Nelson, M.J.K., Montgomery,  S.O., O'Neill, E.J., and
Pritchard,  P.H.   1986.   Aerobic  metabolism  of
trichloroethylene by a bacterial  isolate. Appl. Environ.
Micmbiol. 52:383-384.

Shields,  M.S.,  S.O.  Montgomery, S.M.  Cuskey, P.J.
Chapman,   and P.H.  Pritchard.  1991.   Mutants  of
Pseudomonas  cepacia G4 defective in catabolism of
aromatic compounds and trichloroethylene. Appl. Environ.
Micmbiol. 57: 1935-1941.

Shields, M.S. and Reagin, M.J.  1992. Selection of a
Pseudomonas  cepacia  strain  constitutive  for  the
degradation of trichloroethylene. Appl. Environ. Microbiol.
58:3977-3983.

Winkler, J., K.N. Timmis, and R.A. Snyder. 1995. Tracking
the response of Burkholderia cepacia G4  5223 PR1 in
aquifer microcosms.  Appl. Environ.  Microbiol. 61: 448-
455.

-------
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
     BULK RATE
POSTAGE & FEES PAID
         EPA
   PERMIT No. G-35
Official Business
Penalty for Private Use
3300
EPA/600/S-98/005

-------