Mobilization of Broad Host Range
Plasmid from Pseudomonas putida to
Established Biofilm of Bacillus
azotoformans. I. Experiments
D. L. Beaudoin,1 J. D. Bryers,2 A. B. Cunningham,3 S. W. Peretti4
1Department of Chemical, Bio and Materials Engineering, Arizona State
University, Tempe, Arizona 85287; e-mail: diane@asuvax.eas.asu.edu
2Center for Biomaterials, University of Connecticut Health Center, 263
Farmington Avenue, Farmington, Connecticut 06030
3Center for Biofilm Engineering, Montana State University,
Bozeman, Montana 59717
4Department of Chemical Engineering, North Carolina State University,
Raleigh, North Carolina 27695
Received 30 December 1996; accepted 3 July 1997
Abstract: A strain of Pseudomonas putida harboring
plasmids RK2 and pDLB101 was exposed to a pure cul-
ture biofilm of Bacillus azotoformans grown in a rotating
annular reactor under three different concentrations of
the limiting nutrient, succinate. Experimental results
demonstrated that the broad host range RSF1010 deriva-
tive pDLB101 was transferred to and expressed by B.
azotoformans. At the lower concentrations, donor medi-
ated plasmid transfer increased with increasing nutrient
levels, but the highest nutrient concentration yielded the
lowest rate of donor to recipient plasmid transfer. For
transconjugant initiated transfer, the rate of transfer in-
creased with increasing nutrient concentrations for all
cases. At the lower nutrient concentrations, the fre-
quency of plasmid transfer was higher between donors
and recipients than between transconjugants and recipi-
ents. The reverse was true at the highest succinate con-
centration. The rates and frequencies of plasmid transfer
by mobilization were compared to gene exchange by ret-
rotransfer. The initial rate of retrotransfer was slower
than mobilization, but then increased dramatically. Ret-
rotransfer produced a plasmid transfer frequency more
than an order of magnitude higher than simple mobili-
zation. © 1998 John Wiley & Sons, Inc. Biotechnol Bioeng 57:
272–279, 1998.
Keywords: biofilm; plasmid transfer; conjugation; retro-
transfer
INTRODUCTION
The potential for using biotechnology in the treatment of
toxic waste in the United States is considerable. More than
24,000 industrial plants in the United States generate over
260 million tons of hazardous waste per year (EPA, 1993)
of which 25–50% can be biologically treated (McCormick,
1985). Biotechnology has allowed the creation of geneti-
cally engineered microorganisms (GEMs) capable of de-
grading single compounds or mixtures of similar chemicals
more completely and efficiently than natural bacterial iso-
lates (Haugland et al., 1990). Before these bioengineered
organisms can be released into the environment, the fate and
effects of novel microorganisms and their genetically al-
tered plasmid or chromosomal DNA on the natural ecosys-
tem must be assessed. In particular, the survival, establish-
ment, and growth of recombinant organisms, and the trans-
fer of their plasmid DNA in these environments must be
investigated before the government will permit the release
of recombinant organisms.
It has been shown that the survival and genetic transfer
abilities of microorganisms depend on several factors: the
nature of the bacterial host and cloning vector, the final
ecological niches of the organisms, the transmissibility of
the recombinant DNA (rDNA) to other bacteria, and the
selective advantage that the DNA can confer to the host in
order to survive harsh environments (Cruz–Cruz et al.,
1988; Stotzky and Babich, 1984). Most bacteria found in
nature are attached to surfaces and grow as surface-bound
microcolonies. These surface-attached cells and their extra-
cellular polymeric substances (EPSs) form what is called a
biofilm. It had been assumed that plasmid transfer in natural
environments would not occur because suspended bacterial
population densities are usually lower than those employed
in laboratory experiments. However, biofilms found in natu-
ral environments can support dense and active microbial
communities. Many of these biofilm ecosystems are stable
and resist newly introduced microorganisms from coloniz-
ing them. This has been shown to hold true even when the
newly introduced species was originally one of the indig-
enous population (Freter et al., 1983).
Because little information is available on the dynamics of
gene transfer to and between biofilm organisms, improved
knowledge of these gene transfer patterns will contribute to
the understanding of the effect of GEMs in natural ecosys-
tems. This is especially important because bacteria isolatedCorrespondence to: D. Beaudoin
© 1998 John Wiley & Sons, Inc. CCC 0006-3592/98/030272-08
from the environment have been shown to possess plasmids
that are conjugative (Bender and Cooksey, 1986; Diels et
al., 1989; Focht et al., 1996; Fry and Day, 1990; Guiney and
Lanka, 1989; Hill et al., 1992; Lilley et al., 1992; Shoe-
maker et al., 1992; Wickham and Atlas, 1988) or have been
shown to mobilize other plasmids (Hill et al., 1992; Mer-
geay et al., 1987). Angles et al. (1993) found transfer of the
broad host range plasmid RP1 to be higher in biofilms than
in the aqueous phase. However, in their experiments they
used a recipient strain that had been cultured and optimized
for plasmid transfer and they measured plasmid transfer by
replica plating, which has been shown to give artificially
high numbers (Walter et al., 1991).
This study investigates and quantifies the transfer of a
plasmid by mobilization in a biofilm as a function of am-
bient nutrient level and the mechanism of transfer. Esti-
mates of the rates and kinetics of transfer are made for each
condition. In the companion article to this work, theoretical
models of biofilm accumulation are coupled with the kinetic
expression for plasmid transfer to give a more quantitative
description of the dynamics of plasmid transfer in biofilm
systems.
MATERIALS AND METHODS
Strains and Plasmids
The donor strain used in all experiments was Pseudomonas
putida PB2440 (Bagdasarian et al., 1981). This Gram-
negative soil bacterium was chosen because it has been
widely characterized for use in biodegradation due to its
metabolic capabilities. This strain was made nalidixic acid
resistant by plating onto Luria–Bertani (LB) agar containing
50 mg/mL nalidixic acid and selecting spontaneous mutants.
For these experiments the recipient strain was selected
based on a different cell morphology than the donors, the
ability to express and maintain the broad host range plasmid
RSF1010, and an inability to metabolize lactose. Bacillus
azotoformans, a large Gram-positive rod (1 × 3–10 mm)
originally isolated from soil, fulfilled these criteria (Pichi-
noty et al., 1978).
The plasmid being transferred in these experiments,
pDLB101, combines the broad host range and mobilization
abilities of RSF1010 and the stability and selection markers
of pTKW106 (see Fig. 1). RSF1010’s mode of replication
renders it independent of any host replication functions,
unlike other plasmids (Scholz et al., 1985). This Gram-
negative plasmid has been shown to be stably maintained,
replicated, and expressed in several Gram-positive species
(Gormley and Davies, 1991; Hermans et al., 1991; Mermet–
Bouvier et al., 1993; Nesvera et al., 1994). The inducible
expression system is based on the lacZ gene that codes for
the production of the marker protein, b-galactosidase. This
tightly regulated expression vector originally derived from
plasmid pMJR1750 (Stark, 1987) includes the strong tac
promoter and, for complete repression, the lacIQ gene. The
plasmid also contains the hok/sok (formerly parB) locus that
eliminates the presence of plasmid-free bacteria generated
by plasmid segregational instability. Thus, plasmid
pDLB101 provides a system from which transfer can be
readily observed and serves as an upper limit for the poten-
tial for plasmids to transfer to indigenous organisms.
Because plasmid pDLB101 is only capable of mobiliza-
tion and not conjugation, a helper plasmid must be provided
to facilitate transfer. The broad host range plasmid RK2
served as this helper plasmid in the experiments (Thomas,
1989).
The strains were grown on a minimal salts medium
supplemented with casamino acids, a carbon source, and a
trace metal solution. Formulation for the medium was per
liter distilled water: 2.68 g Na2HPO4 z 7 H2O, 0.98 g
KH2PO4, 0.30 g MgSO4 z 7 H2O, 0.50 g NH4Cl, 1.00 g
NH4SO4, 0.10 g NaH2PO4, and 0.13 g K2HPO4. These com-
ponents were autoclaved for 20 min. After cooling, the fol-
lowing presterilized solutions were added: 600 mL 8% (w/v)
casamino acids, 500 mL 5% (w/v) succinate, 250 mL 10%
(w/v) lactose, and 200 mL trace metal solution (2.20 g
ZnSO4 z 7 H2O, 5.54 g CaCl2, 5.06 g MnCl2 z 4 H2O, 4.79 g
FeSO4 z 7 H2O, 1.10 g NH4Mo7O24 z 4 H2O, 1.57 g
CuSO4 z 5 H2O, 1.60 g CoCl2 z 6 H2O, 0.05 g H3BO3, and
50.0 g EDTA/L dH2O).
Growth rates and Monod kinetic parameters were deter-
mined in independent experiments (data not shown). Sus-
pended cultures of all three strains were grown in the above
media across the range of succinate concentrations em-
ployed in these experiments. Concentrations were deter-
mined from colony counts on LB plates. All measurements
Figure 1. Restriction map of mobilizable plasmid pDLB101 used in
transfer experiments.
BEAUDOIN ET AL.: PLASMA TRANSFER IN BIOFILM SYSTEMS. I 273
were done in triplicate. It was assumed that growth rates in
biofilms and in suspensions were the same even though
there is conflicting evidence supporting and refuting this.
Reactor Description
Biofilms were cultivated in a rotating annular reactor (see
Fig. 2) comprising two concentric cylinders creating a small
annular gap separating the two. This reactor design provides
a high surface area to volume ratio, thereby promoting bio-
film growth and minimizing suspended cell growth. The
outer cylinder remains stationary while the inner one is
rotated to affect a desired surface velocity or shear stress.
Twelve polycarbonate removable slides were machined to
fit into the wall of the outer cylinder for sampling.
To determine the effect of nutrient concentration, and
concomitantly bacterial growth rate, on plasmid transfer, the
succinate concentration in the media delivered to the rotat-
ing annular reactor was varied. For the base case, 25 mg/mL
(1×) succinate was used. The other cases used either 12.5
mg/mL (0.5×) or 75 mg/mL (3×) succinate. The type of
plasmid transfer was also varied. Mobilization, in which the
conjugative helper plasmid RK2 is transferred with the mo-
bilizable pDLB101 following after, was the primary mode
of plasmid transfer examined. However, in one retrotransfer
experiment the donor cells (P. putida PB2440) contained
only pDLB101 and the recipient cells (B. azotoformans)
harbored the conjugative plasmid RK2. During retrotransfer
B. azotoformans should transfer RK2 to PB2440 and then
pDLB101 could be transferred back to B. azotoformans.
Experimental Protocol
The reactor was inoculated with 500 mL of an overnight
culture of the recipient strain (B. azotoformans) grown in
mineral salts medium at 30°C. The suspended culture was
allowed to grow in batch mode for 6 h. At this point the
reactor was switched to continuous mode and fresh nutrients
were delivered at a flow rate to affect a dilution rate of
1.0/h. This dilution rate exceeds the maximum growth rate
of the cells (m 4 0.365/h), thereby minimizing suspended
growth in the reactor. After 3 days of biofilm accumulation,
the biofilm and reactor effluent were sampled and the nu-
trient flow was stopped. Immediately after sampling, 500
mL of P. putida donor cells (3 × 109 cells/mL) were added
to the reactor and allowed to contact the recipient biofilm
for 3 h before being washed out by resuming continuous
influent nutrient media flow, giving the donor cells ad-
equate time to contact the biofilm and attach themselves.
Both the biofilm and effluent were then sampled periodi-
cally over the next 4 days.
Analytical Techniques
At each sampling time, a Plexiglas slide was removed from
the reactor to collect a sample of the biofilm; a sample of the
reactor effluent was also taken. A portion (28 cm2) of the
biofilm was scraped from the slide into 8 mL of sterile
buffer and resuspended by vigorous vortexing for 2 min.
Liquid samples, both the resuspended biofilm sample and
reactor effluent sample, were fixed with 2 mL of 20 mg/mL
1-ethyl-3-(3-dimethylaminopropyl-)-carbodiimide HCl
(EDC, Sigma Chemical Co.) in phosphate buffer and
stained with 4,6-diamidino-2-phenylindole (DAPI, Sigma)
to indicate total cell numbers and with fluorescein di-b-
galactopyranoside (FDG, Sigma) to indicate cells express-
ing b-galactosidase.
For DAPI staining, 900 mL of fixed, suspended cells
(serially diluted if necessary) were added to 100 mL (10
mg/mL) DAPI in a microfuge tube. The contents were vor-
texed briefly to mix, then allowed to stand 20 min at room
temperature. At this time, the stained cells were filtered onto
a 25-mm Poretics 0.2-mm pore size black PCTE membrane
and washed by filtering 4 mL of sterile distilled H2O. The
filter was placed on a drop of immersion oil on a micro-
scope slide, topped with a coverslip, and a minimum of 10
grids were counted using an Olympus BH-2 microscope
with epifluorescence illumination (100-W mercury lamp).
An Olympus U excitation filter cubic unit with excitation
filter (UG-1), a dichroic mirror (DM400), and a barrier filter
(L420) were used to visualize the DAPI-stained cells.
For FDG staining, 980 mL of fixed, suspended cells (se-
rially diluted if necessary) were added to 20 mL of stock
FDG solution (50 mg/mL) in a microfuge tube. The contents
were vortexed briefly to mix then allowed to stand 30 min
at room temperature. The cells were then filtered, washed,
and counted as above.
The remainder of the biofilm on the slide (6 cm2) was
cryoembedded and cryosectioned via the method of Yu et
al. (1994), then stained and examined under the microscope
to visually determine the spatial location of bacterial cells
containing and expressing the pDLB101 plasmid. First a
slide was removed from the reactor and placed flat on a
brick of dry ice. A thick layer of Tissue-Tek OCT com-
Figure 2. Rotating annular reactor used for growing biofilms in plasmid
transfer experiments.
274 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 3, FEBRUARY 5, 1998
pound (Miles Incorporated) was poured on top of the bio-
film, embedding the biofilm and freezing it completely
(sample turns opaque white). The biofilm was removed
from the slide by gently bending the polycarbonate slide.
The sample was placed OCT-side down onto the block of
dry ice, another thick layer of OCT was poured over the
exposed biofilm side, and the sample was allowed to freeze.
Samples were wrapped in aluminum foil and stored at
−70°C until ready to cryosection. Frozen sections were
placed in a cryostat (Reichert–Jung Cryocut 1800, Leica)
operated at −19°C for 15 min before sectioning to allow
thermal equilibration. After equilibration, the samples were
cut into 5-mm slices and collected on positively charged
microscope slides. To each biofilm slice a 10-mL aliquot of
one of the following staining solutions was added: 1.0 mg/
mL DAPI in 4.0 mg/mL EDC fixative buffer or 1.0 mg/mL
FDG in 4.0 mg/mL EDC fixative buffer. P. putida (0.5 × 1
mm) and B. azotoformans (1 × 10 mm) are easily distin-
guished by relative cell sizes. FDG stains b-galactosidase in
only those cells containing a plasmid expressing lacZ ac-
tivity. Both stains were allowed to penetrate the biofilm
sample slice for 10 min. Excess solution was then absorbed
onto a Kimwipe tissue. Each biofilm sample slice was then
covered with a coverslip and examined using an epifluores-
cence microscope.
Statistical Analysis
Each sample was counted under the microscope on a mini-
mum of 10 grids. The numbers plotted represent the average
values and the error bars showing the standard deviation.
RESULTS AND DISCUSSION
There are two possible mechanisms of plasmid transfer in
the experiments above. Initially, only the donors are capable
of transferring their DNA to the recipient cells. However, as
the number of transconjugants increases, it is expected that
the frequency of plasmid transfer from the transconjugants
to the recipients also increases. It has been shown that the
rate of plasmid transfer in suspension can be described
mathematically by k1DR (or k2TR) where k1 and k2 are
kinetic constants and D, R, and T represent, respectively, the
concentrations of donors, recipients, and transconjugants in
the suspension (cells/mL) (Levin et al., 1979; MacDonald et
al., 1992). In biofilm systems, however, there is contact only
between individual layers of cells (see Fig. 3A) and not
between all cells in the biofilm volume. Therefore, the rate
expression for plasmid transfer in biofilms should use the
number of donors, recipients, or transconjugants per unit
biofilm surface area instead of per unit volume.
Effect of Nutrient Concentration
In these experiments the donor strain was P. putida PB2440
(RK2, pDLB101), the recipient strain was B. azotoformans,
and the resulting transconjugant was B. azotoformans (RK2,
pDLB101). As mentioned above, three different levels of
succinate were fed to the biofilm reactor: the base case (25
mg/mL), 3× succinate (75 mg/mL), and 0.5× succinate (12.5
mg/mL).
As stated previously, at short times plasmid transfer oc-
curs exclusively between donor cells and the recipients.
Growth of the transconjugants is assumed to be small com-
pared to their formation by plasmid transfer. By plotting the
initial transconjugant concentrations over time, the rate of
plasmid transfer between donors and recipients can be de-
termined from the slope of this plot. A change in the plas-
mid transfer rate would indicate a shift from donor-
mediated plasmid transfer to transfer initiated by the trans-
conjugants and transconjugant growth.
Figure 4 shows the short time transconjugant concentra-
tions measured for each of the three succinate concentra-
tions. As the plot indicates, there is very little difference in
the rate of plasmid transfer for either the base case or 0.5×
succinate. For the 3× succinate case, however, the rate of
plasmid transfer from donors to recipients decreases signifi-
cantly. The transfer rates for these cases are given in Ta-
ble I.
At longer times, it is speculated that the majority of plas-
mid transfer is carried out between transconjugant cells and
the recipients (see Fig. 3B). During this time, cell growth is
also a major factor in the net accumulation of transconju-
Figure 3. (A) Layering nature of bacterial species in biofilm for plasmid
transfer. The shaded cells are donors; open cells are recipients. (B) Distri-
bution of donor, recipient, and transconjugant cells in biofilm microcolo-
nies resulting from plasmid transfer between an established biofilm and a
released donor. The shaded cells are donors, open cells are recipients, and
striped cells are transconjugants.
BEAUDOIN ET AL.: PLASMA TRANSFER IN BIOFILM SYSTEMS. I 275
gants. By subtracting the rate of transconjugant cell growth
(measured in independent experiments, data not shown)
from the net accumulation of transconjugants, an estimate of
the rate of plasmid transfer between transconjugants and
recipients can be made.
The measured concentrations of transconjugants over the
course of the entire experiment for all three succinate con-
centrations are shown in Figure 5A. Figure 5B shows the
approximate transconjugant concentrations if cell growth is
subtracted from each case. The rate of plasmid transfer from
transconjugants to recipients is slightly higher for 3× suc-
cinate compared to the base case, with the 0.5× succinate
case not producing any further transconjugants via plasmid
transfer, probably due to consumption of the limited nutri-
ents for maintaining cell growth. These values are an order
of magnitude lower for the base case and an order of mag-
nitude higher for 3× succinate compared to the transfer rates
measured for donor initiated transfer.
According to MacDonald et al. (1992), plasmid transfer is
an energy intensive process, with higher transfer rates oc-
curring at higher nutrient concentrations in batch suspen-
sion. In a well-mixed batch system, the cells divide rapidly,
leading to a higher frequency of collisions between cells. In
a biofilm system, however, this is not necessarily true. In the
microcolony structure of a biofilm, rapidly dividing cells
may actually inhibit the formation of efficient mating pairs
by pushing neighboring cells away from each other. The
time between a donor cell and recipient cell coming in con-
tact and establishing a mating pair may be interrupted by
cell division pushing them apart. It is hypothesized that this
may account for the large decrease in the donor to recipient
transfer rate for 3× succinate because these cells are in the
outer region of the microcolony and have a large amount of
free space for microcolony expansion. This is not seen in the
inner region of the microcolony where transconjugant ini-
tiated transfer occurs. This region has less space available
for cells to be pushed apart, so disruption of the mating pairs
may be minimized.
MacDonald et al. (1992) also found that transconjugants
Table I. Measured plasmid transfer rates for donors and transconjugants
for various influent nutrient concentrations.
Succinate
concentration
(mg/mL)
Rate of plasmid transfer
(cells/cm2/h)
Donor mediated Transconjugant mediated
0.025 3.0E + 04 4.5E + 03
0.0125 3.4E + 04 0
0.075 7.0E + 02 5.6E + 03
Figure 4. Early transconjugant concentration profiles resulting from
plasmid transfer between P. putida PB2440 (RK2, pDLB101) and B. az-
otoformans using three different succinate concentrations: (j) 25 mg/mL
succinate, (m) 12.5 mg/mL succinate, and (d) 75 mg/mL succinate.
Figure 5. Transconjugant concentration profiles resulting from plasmid
transfer between P. putida PB2440 (RK2, pDLB101) and B. azotoformans
using three different succinate concentrations. (A) Measured concentra-
tions; (j) 25 mg/mL succinate, (m) 12.5 mg/mL succinate, and (d) 75
mg/mL succinate. (B) Approximate concentrations when cell growth is
subtracted: (—) 25 mg/mL succinate, (z z z) 12.5 mg/mL succinate, and
(- - -) 75 mg/mL succinate.
276 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 3, FEBRUARY 5, 1998
are equally if not more efficient at plasmid transfer than
donors. In these experiments, this theory holds true for the
3× succinate case but not for the base case or for 0.5×
succinate. This discrepancy in donor versus transconjugant
initiated transfer indicates that plasmid transfer may not be
just a strain related phenomenon, but also a result of the
nutrient level. DAPI and FDG images of the biofilms re-
sulting from each of these nutrient levels (photos not
shown) indicate that the 1× and 0.5× succinate cases pro-
duce a denser biofilm than the 3× succinate case based on
the intensity of the fluorescence. These dense microcolonies
may affect the transport and consumption of nutrients. Re-
duced local nutrient concentration in the microcolony inte-
rior due to mass transfer limitations may create gradients in
cell growth rates, reducing the cells’ ability to transfer plas-
mid DNA, especially because piliation and donor ability are
known to be functions of a cell’s growth phase (Brinton,
1965).
Effect of Retrotransfer
During retrotransfer, the recipient, B. azotoformans, harbors
the conjugative RK2 plasmid and the donor, P. putida, con-
tains plasmid pDLB101. Results of retrotransfer and mobi-
lization at 1× succinate are plotted in Figures 6 and 7. As
Figure 6 indicates, the rate of formation of transconjugants
during the first few hours of contact is initially slower for
retrotransfer compared to mobilization, but at 2 h retrotrans-
fer rates dramatically increase and surpass direct mobiliza-
tion rates. This change in the transfer rate may be attributed
to the time required for RK2 to be transferred to and rep-
licated in PB2440 before pDLB101 is transferred back to B.
azotoformans. The exponential increase in transconjugant
cell numbers after the initial 2 h of retrotransfer is likely a
result of transconjugant growth. Plasmid transfer rates for
these two cases are given in Table II. These results are in
agreement with retrotransfer experiments carried out by
Perkins et al. (1994). In their experiments they found a
45–60 min lag before detecting transconjugants.
During retrotransfer there can only be exchange between
the donor, PB2440 (pDLB101), and the recipient, B. azoto-
Figure 6. Early transconjugant concentration profiles resulting from
plasmid transfer between P. putida PB2440 (RK2, pDLB101) and B. az-
otoformans and between P. putida PB2440 (pDLB101) and B. azotofor-
mans (RK2). (d) Normal conjugation and (m) retrotransfer.
Table II. Measured plasmid transfer rates for donors and transconjugants
for simple mobilization and retrotransfer.
Rate of plasmid transfer
(cells/cm2/h)
Donor mediated Transconjugant mediated
Mobilization 3.0E + 04 4.5E + 03
Retrotransfer 1.2E + 04 0
Figure 7. Transconjugant concentration profiles resulting from plasmid
transfer between P. putida PB2440 (RK2, pDLB101) and B. azotoformans
and between P. putida PB2440 (pDLB101) and B. azotoformans (RK2)
using 25 mg/mL succinate. (A) Measured concentrations: (d) normal con-
jugation and (m) retrotransfer. (B) Approximate concentrations when cell
growth is subtracted out: (- - -) normal conjugation and (—) retrotransfer.
BEAUDOIN ET AL.: PLASMA TRANSFER IN BIOFILM SYSTEMS. I 277
formans (RK2). Resulting transconjugants, B. azotoformans
(RK2, pDLB101), and recipients both harbor the conjuga-
tive RK2 plasmid. Surface exclusion prevents two cells har-
boring the same conjugative plasmid from forming a mating
pair. Once all of the recipient cells in contact with donors
have become transconjugants, plasmid transfer stops and
the transconjugant population increases only due to cell
growth. Figure 7 shows the measured transconjugant con-
centrations for mobilization and retrotransfer over the
course of the entire experiment as well as the transconjugant
concentrations if cell growth is neglected. As expected, after
some initial plasmid transfer from the donors to recipients,
the rate of retrotransfer goes to zero.
CONCLUSIONS
Experiments show that plasmid transfer and subsequent
gene expression do occur when a GEM is released into an
established biofilm community, even between Gram-
negative and Gram-positive organisms. This means that
special attention should be paid to the types of genes placed
on plasmids used for bioremediation.
Previous research has shown that the rate of plasmid
transfer in suspension is higher for higher nutrient concen-
trations and when transfer is initiated by the transconjugants
as opposed to the original donors (MacDonald et al., 1992).
This study indicates that during the initial contact between
donors and recipients in a biofilm, plasmid transfer appears
to be inhibited by high concentrations of the limiting carbon
source. However, this may be a result of the nature of the
biofilm matrix and not a physiological response. Rapidly
dividing cells may actually inhibit the formation of efficient
mating pairs by pushing neighboring cells away from each
other before a donor cell and recipient cell can establish a
mating pair. At longer times, when plasmid transfer is be-
tween transconjugants and recipients, the anticipated corre-
lation between higher nutrient concentrations (more actively
growing cells) and faster rates of plasmid transfer is seen. In
these experiments, the rate of plasmid transfer from trans-
conjugants to recipients appears to be slower than that for
donor initiated transfer at the lower nutrient concentrations.
This latter result may be an artifact of the biofilm matrix and
not a function of the bacterial species.
Many of the bacteria in nature harbor conjugative plas-
mids, making retrotransfer of mobilizable plasmids a sig-
nificant mechanism of genetic exchange. In measuring plas-
mid transfer resulting from retrotransfer in biofilms, the
extent of transfer was limited by the number of donors in
contact with recipients. However, the transconjugants con-
tinued to grow and multiply, yielding transconjugant con-
centrations similar to those measured for mobilization.
Therefore, if the ultimate intent is to restrict the distribution
of a phenotype in a specific ecosystem, then it is imperative
that plasmids do not contain genes for any of the transfer
functions.
A companion article to this work combines a theoretical
model of biofilm accumulation with a kinetic expression for
plasmid transfer and compares the data presented above
with model simulations. This model includes attachment
and detachment of cells between the bulk fluid phase and
the biofilm matrix, the rates of mass transport of nutrients
from the bulk fluid, the net rate of conversion of these
substrates into new biomass, and bulk biomass convection
due to growth. Results of these simulations are in excellent
agreement with the experimental results.
The authors wish to acknowledge the assistance of Gayle Callis
in cryosectioning the samples for analysis. This work was sup-
ported by the Center for Biofilm Engineering at Montana State
University–Bozeman, a National Science Foundation Engineer-
ing Research Center (NSF Cooperative Agreement EEC-
8907039).
References
Angles, M. L., Marshall, K. C., Goodman, A. E. 1993. Plasmid transfer
between marine bacteria in the aqueous phase and biofilms in reactor
microcosms. Appl. Environ. Microbiol. 59: 843–850.
Bagdasarian, M., Lurz, R., Ruckert, B., Franklin, F. C. H., Bagdasarian,
M. M., Frey, J., Timmis, K. N. 1981. Specific-purpose plasmid clon-
ing vectors II: Broad host range, high copy number, RSF1010-derived
vectors, and a host-vector system for gene cloning in Pseudomonas.
Gene 16: 237–247.
Bender, C. L., Cooksey, D. A. 1986. Indigenous plasmids in Pseudomonas
syringae pv. tomato: Conjugative transfer and role in copper resis-
tance. J. Bacteriol. 165: 535–541.
Brinton, C. C. 1965. The structure, function, synthesis and genetic control
of bacterial pili and a molecular model for DNA and RNA transport in
Gram-negative bacteria. Trans. NY Acad. Sci. 27: 1003–1054.
Cruz–Cruz, N. E., Toranzos, G. A., Ahearn, D. G., Hazen, T. C. 1988. In
situ survival of plasmid-bearing and plasmidless Pseudomonas aeru-
ginosa in pristine tropical water. Appl. Environ. Microbiol. 54:
2574–2577.
Diels, L., Sadouk, A., Mergeay, M. 1989. Large plasmids governing mul-
tiple resistances to heavy metals: A genetic approach. Toxicol. Envi-
ron. Chem. 23: 79–89.
EPA 1993. Biennial RCRA hazardous waste report: Executive summary.
National Center for Environmental Publications and Information,
Washington, D.C.
Focht, D. D., Searles, D. B., Koh, S.-C. 1996. Genetic exchange in soil
between introduced chlorobenzoate degraders and indigenous biphe-
nyl degraders. Appl. Environ. Microbiol. 62: 3910–3913.
Freter, R., Brickner, H., Fekete, J., Vickerman, M. M., Carey, K. E. 1983.
Survival and implantation of Escherichia coli in the intestinal tract.
Infection Immunity 39: 686–703.
Fry, J. C., Day, M. J. 1990. Plasmid transfer and the release of genetically
engineered bacteria in nature: A discussion and summary. In: J. C. Fry
and M. J. Day (eds.), Bacterial genetics in natural environments. Chap-
man & Hall, London.
Gormley, E., Davies, J. 1991. Transfer of plasmid RSF1010 by conjugation
from Escherichia coli to Streptomyces lividans and Mycobacterium
smegatis. J. Bacteriol. 173: 6705–6708.
Guiney, D. G., Lanka, E. 1989. Conjugative transfer of IncP plasmids. In:
C. M. Thomas (ed.), Promiscuous plasmids of Gram-negative bacteria.
Academic Press Limited, London.
Haugland, R. A., Schlem, D. J., Lyons, R. P., Sfema, P. R., Chakrabarty,
A. M. 1990. Degradation of the chlorinated phenoxyacetate herbicides
2,4-dichlorophenoxyacetic acid and 2,4,5-trichlorophenoxyacetic acid
by pure and mixed bacterial cultures. Appl. Environ. Microbiol. 56:
1357–1362.
Hermans, J., Martin, C., Huijberts, G. N. M., Goosen, T., de Bout, J. A. M.
1991. Transformation of Mycobacterium aurum and Mycobacterium
278 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 3, FEBRUARY 5, 1998
smegmatis with the broad host-range Gram-negative cosmid vector
pJRD215. Mol. Microbiol. 5: 1561–1566.
Hill, K. E., Weightman, A. J., Fry, J. C. 1992. Isolation and screening of
plasmids from the epilithon which mobilize recombinant plasmid
pD10. Appl. Environ. Microbiol. 58: 1292–1300.
Levin, B. R., Stewart, F. M., Rice, V. A. 1979. The kinetics of conjugative
plasmid transmission: Fit of a simple mass action model. Plasmid 2:
247–260.
Lilley, A. K., Bailey, M. J., Day, M. J., Fry, J.C. 1992. Natural transfer
plasmids from sugarbeet rhizosphere. In: Sixth international sympo-
sium on microbial ecology. Barcelona, Spain.
MacDonald, J. A., Smets, B. F., Rittmann, B. E. 1992. The effects of en-
ergy availability on the conjugative-transfer kinetics of plasmid RP4.
Water Res. 26: 461–468.
McCormick, D. 1985. One bug’s meat. Bio/Technology 3: 429–435.
Mergeay, M., Lejeune, P., Sadouk, A., Gerits, J., Fabry, L. 1987. Shuttle
transfer (or retrotransfer) of chromosomal markers mediated by plas-
mid pULB112. Mol. Gene Genet. 209: 61–70.
Mermet–Bouvier, P., Cassier–Chauvat, C., Marraccini, P., Chauvat, F.
1993. Transfer and replication of RSF1010-derived plasmids in several
cyanobacteria of the genera Synechocystis and Synechococcus. Curr.
Microbiol. 27: 323–327.
Nesvera, J., Hochmannova, J., Patek, M., Sroglova, A., Becvarova, V.
1994. Transfer of the broad-host-range IncQ plasmid RSF1010 and
other plasmid vectors to the Gram-positive methylotroph Brevibacte-
rium methylicum by electrotransformation. Appl. Microbiol. Biotech-
nol. 40: 864–866.
Perkins, C. D., Davidson, A. M., Day, M. J., Fry, J. C. 1994. Retrotransfer
kinetics of R300B by pQKH6, a conjugative plasmid from river epi-
lithon. FEMS Microb. Ecol. 15: 33–44.
Pichinoty, F., Durand, M., Job, C., Mandel, M., Garcia, J.-L. 1978. Etude
morphologique, physiologique et taxonomique de Bacillus azotofor-
mans. Can. J. Microbiol. 24: 608–617.
Scholz, P., Haring, V., Scherzinger, E., Lurz, R., Bagdasarian, M. M.,
Schuster, H., Bagdasarian, M. 1985. Replication of the broad host
range plasmid RSF1010. In: D. R. Helinski, S. N. Cohen, D. B.
Clewell, D. A. Jackson, and A. Hollaender (eds.), Plasmids in bacteria.
Plenum, New York.
Shoemaker, N. B., Wang, G., Salyers, A. A. 1992. Evidence for natural
transfer of a tetracycline resistance gene between bacteria from the
human colon and bacteria from the bovine rumen. Appl. Environ.
Microbiol. 58: 1313–1320.
Stark, M. J. R. 1987. Multicopy expression vectors carrying the lac repres-
sor gene for regulated high-level expression of genes in Escherichia
coli. Gene 51: 255–267.
Stotzky, G., Babich, H. 1984. Fate of genetically-engineered microbes in
natural environments. Recomb. DNA Technol. Bull. 7: 163–188.
Thomas, C. M. 1989. Promiscuous plasmids of Gram-negative bacteria.
Academic Press, London.
Walter, M. V., Porteous, L. A., Fieland, V. P., Seidler, R. J., Armstrong,
J. L. 1991. Formation of transconjugants on plating media following in
situ conjugation experiments. Can. J. Microbiol. 37: 703–707.
Wickham, G. S., Atlas, R. M. 1988. Plasmid fluctuations in bacterial popu-
lations from chemically stressed soil communities. Appl. Environ. Mi-
crobiol. 54: 2192–2196.
Yu, F. P., Callis, G. M., Stewart, P. S., Griebe, T., McFeters, G. Q. 1994.
Cryosectioning of biofilms for microscopic examination. Biofouling 8:
85–91.
BEAUDOIN ET AL.: PLASMA TRANSFER IN BIOFILM SYSTEMS. I 279