Science of the Total Environment 493 (2014) 872–882

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Demonstrating plasmid-based horizontal gene transfer in complex

environmental matrices: A practical approach for a critical review
Xavier Bellanger, Hélène Guilloteau, Sébastien Bonot 1, Christophe Merlin ⁎
Université de Lorraine and CNRS, LCPME, UMR 7564, 15 Avenue du Charmois, F-54500 Vandoeuvre-lès-Nancy, France

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• We examined three methods for as-

sessing plasmid transfer in environ-
mental matrices.
• Isolation of transconjugants by cul-
ture is impaired by background
antibioresistances.
• Detection of plasmid transfer by fluo-
rescence is unsuitable for environmen-
tal samples.
• Molecular-based methods allow de-
tecting rare successful plasmid transfer
events.
• Eukaryotic predation both promotes
and limits plasmid transfer in natural
communities.

a r t i c l e i n f o a b s t r a c t

Article history: Plasmid-based dissemination of antibiotic resistance genes in environmental microbial communities is a matter of
Received 30 April 2014 concern for public health, but it remains difficult to study for methodological reasons. In this study, we used the
Received in revised form 16 June 2014 broad host range plasmid pB10 to compare and to point out the main drawbacks of the three different approaches
Accepted 17 June 2014
currently used to evaluate plasmid transfer in natural communities. Culture-based selection of transconjugants
Available online 5 July 2014
appeared to be compromised by high prevalence of antibiotic resistances among natural communities, unless
Editor: D. Barcelo high loads of initial pB10-donor inocula were used. Fluorescence-based detection of transconjugants reached a
dead-end consequently to the narrow host range of bacteria expressing fluorescent proteins from a genetically
Keywords: modified pB10 plasmid, in addition to the relatively high background level of fluorescence exhibited by some en-
Plasmid transfer vironmental matrices. The molecular-based approach was the only one to provide a mean to detect rare plasmid
pB10 transfer events following a low but realistic initial pB10-donor inoculation. Whatever the method, culture-based
Antibiotic resistance dissemination or molecular-based, the detection of successful transfer events in a given environmental matrix seemed to be
Environmental bacterial communities linked to the initial stability of the donor inoculum. Depending on the matrix considered, eukaryotic predation
Eukaryotic predation
plays a significant role in either limiting or promoting the plasmid transfer events.
© 2014 Elsevier B.V. All rights reserved.

⁎ Corresponding author. Tel.: +33 3 83 68 22 30; fax: +33 3 83 68 22 33. 1. Introduction

E-mail addresses: xavier.bellanger@univ-lorraine.fr (X. Bellanger),
helene.guilloteau@univ-lorraine.fr (H. Guilloteau), bonot@lippmann.lu (S. Bonot),
christophe.merlin@univ-lorraine.fr (C. Merlin).
The great promises once fueled by the discovery of antibiotics
1
Present address: Centre de Recherche Public Gabriel Lippmann, Département EVA, are now severely altered by our increasing failure to treat commonly
41 rue du Brill, 4422 Belvaux Luxembourg. curable infectious diseases. Such failure results from the still rising

http://dx.doi.org/10.1016/j.scitotenv.2014.06.070
0048-9697/© 2014 Elsevier B.V. All rights reserved.
 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882
occurrence of antibiotic resistances in bacteria following the produc-
tion and the consumption of antibiotics in therapeutics and animal
husbandry (Davies, 2007; Davies and Davies, 2010; Aminov, 2010).
The loss in therapeutic potential associated with the dissemination of
antibiotic resistances among pathogens has reached such a critical
point that some reports are already describing a near-return to pre-
antibiotic era (Levy and Marshall, 2004; Roberts and Simpson, 2008).
The relationship between antibiotic consumption and dissemination
of antibiotics was acknowledged long ago and, so far, much emphasis
has been given to the “hospital” context where antibiotics could act as
selective agent. However, more recently, it has become obvious that
(i) antibiotics still exert effects on bacteria at low and environmentally
relevant concentrations (Yim et al., 2006), and (ii) hot spots of antibiotic
resistance gene transfer are likely to be found in various environments
able to sustain high cell density (Dröge et al., 1999, 2000; Molin and
Tolker-Nielsen, 2003; Aminov, 2011; Rizzo et al., 2013).
To date, most of our knowledge regarding the dissemination of anti-
biotic resistances in the environment relies on retrospective evidences:
(i) the increasing occurrences of antibiotic resistances in various
receiving environments (Knapp et al., 2010), (ii) the absence of congru-
ence between the phylogeny of the transferred genes and the phyloge-
ny of the corresponding hosts (Sørensen et al., 2005; Lal et al., 2008),
(iii) the loss of synteny between closely related genomes resulting
from the insertion of acquired DNA in the host genome (Dobrindt
et al., 2004), and finally (iv) the association of the acquired resistance
genes with mobile genetic elements (MGEs) such as plasmids, phages,
transposons and integrons (Partridge, 2011). Despite being obvious
when analyzing acquired DNA sequences from bacterial isolates, the
transfer of antibiotic resistance genes in complex environments remains
difficult to demonstrate when it comes to work with natural bacterial
communities. Yet, demonstrating directly the effectiveness of gene
transfer in natural communities is a prerequisite to precisely identify
environmental hot spots of antibiotic resistance gene dissemination,
as well as environmental parameters driving the evolution of bacteria
towards more resistant microorganisms.
It is now well understood that antibiotic resistant genes can spread
in microbial communities thanks to genetic structures promoting their
mobility (Partridge, 2011). So far, most of our experiences have relied
on the transfer of conjugative elements (e.g. plasmids) that are be-
lieved to play a predominant role in the adaptation of bacteria to anti-
biotics pressure (Sørensen et al., 2005; van Elsas and Bailey, 2002;
Thomas, 2000; Bennett, 2008). To date, three kinds of approaches
have been used to monitor the transfer of known conjugative ele-
ments, mainly plasmids, in environmental matrices: culture-based,
fluorescence-based and molecular-based approaches (Rizzo et al.,
2013). Culture-based methods consist in using selective media to
enumerate transconjugants after introduction of donor bacteria in en-
vironmental samples (Sørensen et al., 2004; van Elsas et al., 2004).
These approaches have been used for more than three decades,
allowing the identification of several factors affecting plasmid transfer
efficiency. Nevertheless, they remain quite limited when indigenous
bacterial populations are used as recipients since less than 1% of envi-
ronmental bacteria are believed to be culturable, and only a fraction of
them are able to express the selective marker transferred (Sørensen
et al., 2005). Later on, alternative methods making use of plasmids ge-
netically modified with fluorescent protein genes have been developed
for the fluorescent detection of transconjugants in complex environ-
ments (Sørensen et al., 2005; Geisenberger et al., 1999). These fluores-
cence approaches advantageously provide information regarding the
localization where the transfers take place in complex community
structures (Geisenberger et al., 1999; Musovic et al., 2010; Dahlberg
et al., 1998), but they are constrained by the need to genetically modify
the element studied. More recently, several teams, including ours, have
simultaneously developed a molecular biology-based approach, which
makes use of quantitative PCR (qPCR) to monitor the dissemination of
a given plasmid in the vastness of microbial community DNAs (Bonot
873
and Merlin, 2010; Merlin et al., 2011; Haug et al., 2010, 2011; Wan
et al., 2011). Practically, it consists in inoculating microcosms of envi-
ronmental samples with a donor bacterium and monitoring the evolu-
tion of both the plasmid and the initial host DNAs over time. Because
conjugative transfer is an intercellular form of DNA replication, the plas-
mid to donor DNA ratio increases when the plasmid disseminates by
conjugation into the indigenous population. This method provides the
advantage of considering gene transfer on a wide range of possible
indigenous recipient bacteria, culturable or not, but it requires the de-
sign of very specific sets of primers and probes for the non-ambiguous
quantification of donor and plasmid.
All these methods have been used with various levels of success to
pinpoint particular environments or parameters for their propensity
to support the transfer of MGEs. Still, as shown here, demonstrating
the transfer of MGEs in environmental matrices is often a difficult and
tedious task, with no guarantee of success when it comes to work
with complex environments hosting already installed microbial com-
munities. In this report, we used the broad host range plasmid pB10,
which is highly efficient for conjugative transfer, to revisit each of
these methodologies, and to highlight major drawbacks associated
with working with complex environmental matrices. Plasmid transfer
was studied using environmental microbial communities maintained
in microcosms as recipients. Special attention was given to the initial
stability of the donor bacteria and the repercussions it could have on
the detection of plasmid transfer events. The influence of eukaryotic
predation on the persistence of the donor bacteria, and then on plasmid
transfer, was explored.
2. Materials and methods
2.1. Bacterial strains and media
Bacterial strains used in the present study are described in Table 1.
Bacteria were routinely cultured at 30 °C under agitation (160 rpm)
in LB medium (LB Broth Miller, DifcoTM) supplemented with the appro-
priate antibiotic for selection. Unless stated otherwise, antibiotics were
used for selection at the following concentrations: Amoxicillin (Amx):
30 μg.mL−1; Chloramphenicol (Chl): 20 μg.mL−1; Kanamycin (Kan):
100 μg.mL− 1; Nalidixic acid (Nal): 20 μg.mL− 1; Rifampicin (Rif):
100 μg.mL− 1; Streptomycin (Str): 10 μg.mL− 1; Sulfamethoxazole
(Sul): 150 μg.mL−1; Tetracycline (Tet): 10 μg.mL−1. The concentrations
of antibiotics were empirically adjusted on E. coli for being selective
when the genetic determinant for resistance is present (normal growth)
and counter selective when it is absent (no growth). For solid media,
agar was added at 15 g.L−1.
2.2. Construction of gfp-tagged pB10 plasmid derivatives
Plasmid pB10 was mutagenized by inserting a mini-Tn5-gfp transpo-
son using a procedure close to the one reported by Christensen et al.
(1998). An insertion library of mini-Tn5-gfp, conferring resistance
to kanamycin, was obtained in Cupriavidus metallidurans CM124
(AE815(pB10)) using a triparental mating with two other Escherichia
coli strains, the donor JB120 and the helper HB101(pRK600). Insertion
clones were recovered on LB Rif (100 μg.mL−1) Kan (1 mg.mL−1), and
pooled in 10 mM MgSO4. The pool was further used as donor in a bipa-
rental mating with the recipient strain E. coli CM125. Transconjugants
displaying a pB10 plasmid enlarged by a copy of mini-Tn5-gfp
(i.e. pB10::mini-Tn5-gfp) were selected on LB Nal (20 μg.mL− 1) Kan
(100 μg.mL−1), and were further checked for the antibiotic resistances
normally encoded by pB10. One of these transconjugants (strain
CM192), for which plasmid restriction analyses revealed a mini-Tn5-
gfp insertion in the vicinity of pB10 oriV (supplementary material), was
retained for this study. Plasmid pB10::mini-Tn5-gfp was further intro-
duced into various Proteobacteria species with a biparental mating be-
tween the donor strain E. coli CM125(pB10::mini-Tn5-gfp) and a set of
874 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882

Table 1
Bacterial strains and plasmids used in this study.

Name/species Genotypes or characteristics References

Plasmids
pB10 wild type IncP-1 plasmid isolated from activated sludge Schlüter et al., 2003
pB10::mini-Tn5-gfp gfp-tagged derivative of pB10 this study
pJBA28 mobilizable suicide plasmid for mini-Tn5-gfp delivery Andersen et al., 1998
pRK600 ChlR, ColE1, RK2 tra genes; helper plasmid for triparental matings Kessler et al., 1992
Cupriavidus metallidurans
AE815 plasmid free and RifR derivative of wild type strain CH34 Springael et al., 1993
CM124 AE815(pB10), RifR, AmxR, StrR, SulR, TetR This study
Cupriavidus necator
JMP228::lacIq JMP228 with a mini-Tn5-lacIq insertion, RifR De Gelder et al., 2005
CM224 JMP228::lacIq (pB10::mini-Tn5-gfp) This study
Escherichia coli
DH5α ϕ80lacZΔM15 recA1endA1gyrA96 (Nal R) thi-1 Sambrook et al., 1989
hsdR17(r-Km+ K ) supE44 relA1 deoR Δ(lacZYAargF) U169
DH5α(pB10) Nal R, AmxR, StrR, SulR, TetR Schlüter et al., 2003
HB101(pRK600) recAthi pro leu hsdRM1StrR; plasmid pRK600 ChlR Kessler et al., 1992
JM109 recA1 endA1 gyrA96 (NalR) thi hsdR17 supE44 Miller, 1992
relA1Δ(lac-proAB)/F'[traD36proAB + lacIqlacZΔM15]
JB120 MV1190(λpir)[pJBA28], KanR Andersen et al., 1998
Δ(lac-proAB) Δ(srl-recA)306::Tn10 [F’ traD36 proAB
lacIq (lacZ)M15] thisupEλpir
MG1655 Sequenced λ− and F− derivative of strain K-12 Blattner et al., 1997
CM125 NalR derivative of MG1655 This study
CM192 CM125(pB10::mini-Tn5-gfp) This study
Pseudomonas putida
SM1443 KT2442 with a mini-Tn5-lacIq insertion, RifR Christensen et al., 1998
CM237 SM1443(pB10::mini-Tn5-gfp) This study
Pseudomonas aeruginosa
PAO1 Wild type Holloway and Morgan, 1986
CM210 PAO1::lacIq(pB10::mini-Tn5-gfp) This study
Sinorhizobium meliloti
RM1021::lacIq RM1021 with a mini-Tn5-lacIq insertion, RifR De Gelder et al., 2005
CM217 RM1021::lacIq(pB10::mini-Tn5-gfp) This study

AmxR, ChlR, KanR, Nal R, RifR, Str R, Sul R,TetR: resistance to Amoxycillin, Chloramphenicol, Kanamycin, Nalidixic acid, Rifampicin, Streptomycin, Sulfamethoxazole, Tetracycline, respectively.

RifR recipient strains namely: E. coli JM109, P. aerginosa CM210, P. putida
SM1443, S. melioti RM1021::lacIq, C. metallidurans AE815, C. necator
JMP228::lacIq (Table 1), with a selection on LB Tet (10 μg.mL− 1) Rif
(100 μg.mL−1). For microscopic observation (epifluorescence micro-
scope Olympus BX 51, λex = 488 nm, λem = 515 ± 15 nm), bacteria
were previously cultured on LB plate amended (or not) with 400 μM
IPTG (isopropyl-β-D-thiogalactoside). For bacterial hosts expressing a
poor level of fluorescence, plasmid pB10::mini-Tn5-gfp was transferred
back into E. coli CM125 by conjugation in order to confirm that no mu-
tation had altered the expression of the gfp reporter.
2.3.2. Plate mating for strain construction
Plasmid donor and recipient bacteria were cultured overnight in LB
at 30 °C under selective conditions. Bacterial cells were washed twice
by centrifugation and dispersed in one volume of 10 mM MgSO4. Fifty
microliters of each suspension were mixed directly on top of an LB
plate and incubated overnight at 30 °C. Cells from the mating mixture
were collected, dispersed and diluted in 10 mM MgSO4, and spread
onto selective plates for selecting transconjuants. For the mobilization
of non-conjugative plasmid, a helper strain providing the transfer func-
tion was also added to the mixture in a triparental mating.

2.3. Bacterial plate mating 2.4. Environmental matrices and microcosms

2.3.1. Biparental filter mating for conjugation studies
Donor (E. coli DH5α(pB10)) and recipient bacteria (C. metallidurans
AE815) were cultured overnight in LB at 30 °C under selective condi-
tions. Bacterial cells were washed twice by alternating centrifugation
(10,000 g for 2 min at 20 °C) and cell pellets dispersion in one volume
of 10 mM MgSO4. Fifty microliters of each cell suspension were then
mixed and spotted onto a 0.45 μm nitrocellulose filter (HA millipore
filters, Ref HAWGO47A0) disposed on an LB plate that had been pre-
warmed at the desired mating temperature. After incubation for a
defined length of time at either 22 °C (the operating temperature
for the microcosm experiments) or 30 °C (the optimal growth tem-
perature for C. metallidurans), the bacterial mixtures were recovered
by vortexing the mating filters in microtubes containing 900 μL of
10 mM MgSO4. Bacteria were serially diluted in 10 mM MgSO4 and
100 μL aliquots were plated on LB supplemented with either Tet
(selection of donor and transconjugants), Rif (selection of recipient and
transconjugants), or Rif and Tet (selection of transconjugants), for viable
cell counts. Transfer frequencies were expressed as transconjugants per
recipient-type cells.
2.4.1. Environmental samples
Sediments from Moselle River (France) were described previously
(Bonot and Merlin, 2010). Cow (Prim’Holstein) manure was repeatedly
sampled in a dairy farm close to Nancy, in the Eastern part of France.
Fresh samples were obtained in a collector receiving fresh manure
continuously scraped from the barn. Manure characteristics obtained
from one of the samples gave: solid material: 11.6 % (mass); pH: 7.45;
conductivity: 10.97 mS.cm−1; dissolved O2: 0.28 mg.L− 1. Activated
sludge and effluent were sampled in a slaughterhouse wastewater
treatment plant (WWTP) dedicated to calves. Activated sludge samples
were collected in the biological basin and presented the following
characteristics: pH = 7.0; Dissolved O2 = 0.1 mg.L−1; total suspended
solids (TSS) = 9.9 g.L−1; volatile suspended solids (VSS) = 4.3 g.L−1.
Effluents were recovered from an automatic sampler operating for 24
h and presented the following characteristics: pH = 7.1; Dissolved O2
= 3.8 mg.L−1; TSS = 3.8 g.L−1; VSS = 0.1 g.L−1. The environmental
samples collected were used to conduct plasmid transfer experiments
with E. coli DH5α(pB10) as donor and indigenous bacterial communi-
ties as recipient.
 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882
Background of antibiotic resistant bacteria for Amx, Tet, Sul, and Str
in natural communities was estimated as follows. Approximately, 20 g
of environmental sample was resuspended in 20 mL of 10 mM
MgSO4, and checked at 200 rpm for 2 h at 30 °C. The particles were
allowed to settle for 30 min and the liquid phases were serially diluted
in 10 mM MgSO4 before being plated on a selective medium. The total
viable counts were evaluated by plating on LB medium. Indigenous bac-
teria naturally exhibiting the resistance profile encoded by pB10 were
enumerated in two steps: first a selection on LB supplemented with
Tet and Amx, and second a screening for additional resistances towards
Sul and Str.
2.4.2. Isolation of pB10 transconjugant in manure
Freshly sampled manure was slightly diluted with sterile water (one
fourth of the volume), homogenized, and distributed in 5 mL aliquots.
Each aliquot received a 400 μL inoculum of E. coli DH5α(pB10) cell sus-
pension in sterile water with a final concentration ranging from 2⋅104 to
2⋅109 cfu.mL−1. The mixtures were homogenized once more and incu-
bated at 30 °C for 24 h. They were then serially diluted in 10 mM MgSO4
and 100 μL of each dilution were plated on LB for the enumeration of
total culturable bacteria, or LB Tet Amx for a pre-selection of putative
pB10 bearing bacteria. For each assay, ca. 100 TetR AmxR colonies
were randomly chosen and screened for resistance to Tet, Amx, Sul,
Str, and Nal. Bacteria displaying the resistance profile of plasmid pB10
and/or E. coli DH5α(pB10) were subjected to molecular analyses.
Crude DNA extracts were obtained by boiling 1 colony in 100 μL of
water for 5 min. The extracts were used for PCR tests using highly spe-
cific primer sets for either plasmid pB10 or E. coli DH5α DNA, and addi-
tional stringency was gained by quantitative PCR (qPCR) using specific
Taqman® probes as previously described (Bonot and Merlin, 2010).
The absence of inhibitor in reactions was confirmed by running addi-
tional tests with the extracts spiked with defined quantities of pure E.
coli DH5α(pB10) DNA.
2.4.3. Microcosm operation and sampling
Activated sludge, WWTP effluents, and cow manure were assessed
for their propensity to support the transfer of plasmid pB10 in micro-
cosms. The slaughterhouse WWTP sludge or effluent microcosms
consisted of 800 mL environmental samples maintained in 1 L Schott
bottles at 20 °C–25 °C without agitation. Transfer experiments started
by inoculating the donor bacteria E. coli DH5α(pB10) (Schlüter et al.,
2003) at a final concentration of 2⋅105 cfu.mL−1. At intervals, 50 mL
samples were collected from the microcosms for molecular biology
analyses. For cow manure microcosms, 200 mL of freshly sampled ma-
nure was slightly diluted with 70 mL of sterile distilled water before
being inoculated with E. coli DH5α(pB10) at a final concentration
of 107 cfu.mL−1. The mixtures were homogenized and distributed in
25 mL fractions in 50 mL conical tubes. The tubes were maintained
caps unscrewed at 30 °C and were sacrificed at intervals for molecular
biology analyses. For assessing the influence of eukaryotic predation
on the transfer of pB10, the microcosms received an initial treatment
of cycloheximide at either 0, 0.1, or 1 mg.mL−1.
The E. coli DH5α(pB10) inocula were prepared by overnight culture
in LB medium supplemented with 10 μg.mL−1 of tetracycline (at 30 °C
under agitation at 160 rpm). Cells were washed twice by centrifugation
(10,000 g for 2 min at 20 °C) in 10 mM MgSO4 before being inoculated
at a given concentration. For each plasmid transfer experiment, sets of
non-inoculated control microcosms were systematically run in parallel
in order to rule out any natural occurrence or resurgence of non-specific
pB10-like DNA sequences in the subsequent qPCR analyses. Additional
sets of control microcosms were routinely run by inoculating replicate
microcosms with naked DNA instead of live E. coli DH5α(pB10)
cells. In all cases, the levels of DNA detected by qPCR after inoculation
with naked DNA were far below those obtained for the cognate micro-
cosms inoculated with the corresponding cells (by at least 2 logs
for equivalently sized inocula), therefore ruling out the incidence of
875
any possible DNA release from the donor bacteria in the phenomena
observed.
2.5. Molecular biology analyses
2.5.1. Total DNA extraction from environmental samples
Total community DNAs were extracted from 500 mg (wet weight) of
environmental sample (sediment or sludge) following the procedure
described by Porteous et al. (1997) later modified by Bonot et al.
(2010). For liquid WWTP effluents, community DNAs were extracted
from cell pellets collected by centrifugation of 25 mL samples using the
“Wizard® Genomic DNA Purification Kit” (Promega) according to the
recommendations delivered by the manufacturer. DNA concentration
and purity were estimated spectrophotometrically based on absorbance
readings at 230, 260 and 280 nm, using a BioPhotometer™ (Eppendorf)
and according to the manufacturer’s procedure: A260 was used for the
concentration calculation while ratios A260/280 and A260/230 were used
for estimation of purity. The absence of residual inhibitors in the DNA ex-
tracts was checked by qPCR after spiking known quantities of target DNA
and/or by serially diluting the template prior amplification.
2.5.2. Quantitative PCR assays
The relative amounts of pB10 plasmid and E. coli DH5α genome in
community DNAs were estimated by qPCR as described previously
(Merlin et al., 2011). Specific TaqMan® probes and primers have already
been described in great details by Bonot and Merlin (2010). Briefly,
quantitative PCR was performed in triplicate using a “StepOnePlus™
Real-Time PCR System” (Applied Biosystems®) with thermocycling
conditions set as follows: 2 min at 50 °C, then 10 min at 95 °C followed
by 45 cycles of 15 s at 95 °C and 1 min at 60 °C. Quantifications were
carried out from 25 ng of total community DNA using a “TaqMan®
Universal PCR Master Mix, noAmpErase® UNG” qPCR kit (Applied
Biosystems®) as recommended by the manufacturer, with 800 nM of
each primer and 300 nM of TaqMan® probe, in a 25 μL reaction volume.
Standards and controls were prepared as previously described (Merlin
et al., 2011). Most of the time, the evolution of either pB10 plasmid or
E. coli DH5α DNA over time appeared to follow a log-normal function
with a decrease rate (expressed in “Log per day”) that could be calcu-
lated from an exponential fitting. All reactions were carried out in trip-
licate and data were normalized and expressed as a number of copies
per μg of community DNA. The high specificity of the primer/probe
sets for detecting pB10 or E. coli DH5α DNAs was systematically checked
by verifying the absence of a qPCR amplification signal with total DNA
extracted from non-inoculated environmental matrices (control).
3. Results
3.1. Plasmid pB10 as model for studying gene transfer
This work resulted from a study aiming to evaluate the propensity of
environmental matrices to support the transfer of plasmid pB10, here
used as a model. Plasmid pB10 was initially chosen for its characteris-
tics: (i) it is a natural plasmid isolated from activated sludge (Dröge
et al., 2000), thus behaving normally in terms of regulation, (ii) it
belongs to the very ubiquitous group of broad host range IncP-1β
plasmids and therefore has the potential to invade a wide range of bac-
terial species (Adamczyk and Jagura-Burdzy, 2003), (iii) it carries resis-
tance genes for Amx, Str, Sul, and Tet, thus providing four means for
selection/screening, (iv) it is a fully sequenced element (Schlüter et al.,
2003), thus allowing the use of molecular approaches, and finally (v)
it transfers by conjugation with a relatively high efficiency as illustrated
in Fig. 1. Here, 12 independent biparental filter matings between E. coli
DH5α(pB10) and C. metallidurans AE815 were carried out independent-
ly, at 22 °C and 30 °C, and for different lengths of time. At 30 °C, the op-
timal growth temperature for the recipient, the relative amount of
recipient receiving pB10 ranges from 0.1% to 80% for mating durations
876 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882

100 In the following experiment, we used such a culture-based approach

30°C to assess the transfer of plasmid pB10 in manure microbial communi-
(transconjugants per recipient)

10-1 22°C ties. E. coli DH5α(pB10) was used as initial plasmid donor and the effect
of the inoculum size was studied concomitantly. Cow manure was sam-
Transfer frequency

pled twice at five-month intervals, at the same location, and aliquots

10-2
were inoculated independently with different concentrations of donor
bacteria ranging from 2⋅104 to 2⋅109 cfu.mL−1. After 24 h at 30 °C, pu-
10-3 tative pB10 bearing bacteria were pre-selected on LB-Tet-Amx and
further screened for resistance to Sul, Str, and Nal in order to identify
pB10 and E. coli DH5α typical antibiotic resistance profiles. Positive
10-4
candidates were further confirmed as E. coli DH5α (or not) and as
hosting plasmid pB10 (or not) by PCR using highly specific primer
10-5 sets, and final confirmation was obtained using qPCR where the higher
1h 2h 3h 4h 5h 24h stringency provided by the use of specific Taqman® probes could
Mating duration remove a few false positives (ca. 20%). The results obtained from two
independent pB10 transfer assays in manure are presented in Fig. 3.
Fig. 1. Transfer of pB10 in biparental matings. The donor E. coli DH5α(pB10) and the
recipient C. metallidurans AE815 were mated for different lengths of time at either 30 °C
or 22 °C.
A
1010
Total count (LB) Donor at T0
ranging from 1 to 24 h, respectively, which clearly demonstrates both
Donor at Day 1 Transconjugants

Cell count (cfu.mL-1)

the high velocity and the high yield of pB10 transfer between two com-
108
patible partners.

3.2. Evaluating pB10 transfer in environmental communities by 106

culture-based approaches
104
Determining the efficiency of transfer of pB10 in a biparental mating
is nothing more complicated than having a robust selection to enumer-
ate the donor, the recipient and the transconjugant cells independently 102
in a mating mixture. Such selection is generally more difficult to achieve #0 #4 #5 #6 #7 #8 #9
when it comes to work with environmental matrices as natural commu- Mating test (Manure M1)
nities can exhibit relatively high levels of background resistance to-
wards the antibiotics used for selection. As illustrated in Fig. 2, the B
1010
relative amount of bacteria exhibiting the antibiotic resistance profile
of pB10 ranges from c.a. 0.02% to 7% of the culturable bacteria in a few
Cell count (cfu.mL-1)

natural communities tested, even though none displayed the pB10 sig- 108
nature at the molecular level. This clearly points out at least two techni-
cal difficulties to evaluate the transfer of plasmids in environmental
106
communities using culture-based methods: first, the antibiotic resis-
tance profile, by itself, is not enough to differentiate background resis-
tance from the appearance of transconjugants and, therefore, does 104
Total count (LB) Donor at T0
not signify that transfer has occurred; second, the identification of
Donor at Day 1 Transconjugants
transconjugants should combine a robust selection procedure and addi-
102
tional sets of screening, including specific PCR tests for instance, which #0 #4 #5 #6 #7 #8 #9
also leads to a noticeable increase of the workload as shown next. Mating test (Manure M2)

109 C
Total count (LB) 100
108
Viable cell count (cfu.mL-1)

AmxR, TetR, SulR, StrR count

107 10-1
Transfer frequency

y = 5E-09x0.9926
6 R2 = 0.94568
10 (0.05%) (7.19%) 10 -2

105 (0.02%) (0.56%) 10-3

104 (3.58%)
10-4
103 (0.10%) Repeat M1
(0.04%)
102 10-5 Repeat M2
1 -6
10 10
100 102 104 106 108
100
-1
M1 M2 AS RS E S1 S2 Donor Count at Day 1 (cfu.mL )
Environmental matrices
Fig. 3. Transfer of pB10 in manure communities. A and B: Enumeration of donor and
Fig. 2. Background resistance in environmental communities. Numbers in parentheses transconjugant cells in 18 independent mating experiments involving 9 initial concentra-
indicate the relative abundance of bacteria showing resistance profiles as encoded by plas- tions of donor cells and 2 different manure samples (M1 and M2). C: Relationship between
mid pB10 for Amx, Tet, Sul, and Str. Environmental matrices tested: M1 and M2: manure; the transfer frequency (transconjugant per indigenous cell) and the stability of the donor
AS: activated sludge; RS: recirculation sludge; E: WWTP effluent; S1 and S2: river sediments. (enumerated at Day 1).
 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882
Interestingly, the two repeats carried out differed significantly with re-
spect to the stability of the donor cells over the 24-h incubation periods.
The first repeat (Fig. 3A) showed a 5-log decrease of the E. coli
DH5α(pB10) count compared to the inoculum, and the recovery
of transconjugants was relatively weak. Conversely, in the second re-
peat (Fig. 3B), the E. coli donor strain DH5α(pB10) appeared to be far
more stable, with only a 1- to 2-log decrease, and the recovery of
transconjugants was relatively high. In light of these experiments, it
seems that the frequency of transfer is directly related to the amount
of donor cells able to maintain in the environmental matrices, which
itself depends on both the size of the initial inoculum and the stability
of the donor over time (Fig. 3C).
The latter experiment highlighted two important points. First
the screenings of putative transconjugants have to be conducted in
depth to ascertain the acquisition of the plasmid. For the experiment
presented in Fig. 3, approximately 1000 isolates were screened
independently (i) on 5 antibiotics and (ii) at the molecular level, to re-
veal the presence of either the pB10 or the E. coli DH5α signature.
Transconjugants could be enumerated this way, only because they rep-
resented a significant proportion of the multiresistant isolates, thus
making the screening fruitful. If the background of resistant bacteria
was to be higher among natural communities, then more isolates should
have been screened to have a chance of isolating transconjugants, thus
making the approach almost impossible to carry out for practical
reasons. This also means that, to be effective, a culture-based approach
necessarily implies the use of high loads of donor cells in order to
have a chance to isolate a successful transfer event. In other words,
such a kind of approach is not really appropriate for the identification
of rare successful transfer events. Second, it should be noted that a
great variability was observed from one repeat to another although
the manure was sampled twice (but independently) at the same
site. This is likely to reflect the fluctuation of important parameters
that can easily vary over time, such as the presence and the quantity
of predators, as well as the structure of the microbial communities, for
instance. In the meantime, this raises the question of the significance
of replicate experiments when they are carried out on a same (split) en-
vironmental sample. Identifying trends and driving parameters in gene
transfer should rather take into account the sample variability in terms
of matrix permissiveness. In this respect, comparing the behavior of dif-
ferent environmental matrices, even as unique repeats, might be more
informative than wrongly repeated experiments that present the risk
of simply averaging the outcome of a particular case.
3.3. Using gfp-tagged pB10 for fluorescence-based plasmid transfer
monitoring
Over the past 20 years, a few genetic tools have been developed
allowing the introduction of fluorescent protein genes in plasmid se-
quences for in situ monitoring of plasmid transfer by confocal fluores-
cence microscopy-based approaches. Despite implying the modification
of the plasmid sequence, and therefore working with a non-natural plas-
mid, we attempted to use such approaches with the aim of monitoring
plasmid pB10 transfer in complex environmental communities. First,
plasmid pB10 was randomly mutated with a mini-Tn5 derivative
carrying a gfp gene controlled by a modified Plac promoter (PA1/04/03),
which led to the isolation of the E. coli strain CM192 (Table 1), carrying
plasmid pB10::mini-Tn5-gfp. Theoretically, such a construction should
be used with a donor strain carrying a lacIq mutation in order to avoid
background levels of gfp expression. Upon plasmid transfer into a
recipient bacterium, the absence of LacI repressor is supposed to
promote the zygotic induction of gfp expression, finally making the
transconjugants fluorescent (Sørensen et al., 2005). First, the ability of
plasmid pB10::mini-Tn5-gfp to promote the fluorescence of its host
was evaluated by introducing the construction into a set of bacterial
species/strains representative of three proteobacterial classes (α-, β-,
and γ-Proteobacteria) that are believed to be the preferred hosts for
877
IncP-1 plasmids (Suzuki et al., 2010). The different bacteria hosting
pB10::mini-Tn5-gfp, harboring or not a lacIq gene, were then observed
by epifluorescence microscopy after having been induced or not
by IPTG. In the wild type lacI derivative E. coli CM192, the gfp gene
was expressed in both IPTG-induced and -non induced cells (Fig. 4A
and B), while in the lacIq strain E. coli JM109(pB10::mini-Tn5-gfp) the
fluorescence was solely observed with IPTG-induced cells (Fig. 4C).
This demonstrated the absolute necessity of using a lacIq allele in a
donor strain for a good control of the PA1/04/03 promoter, and thus
to prevent background fluorescence in the absence of inducers. The in-
troduction of plasmid pB10::mini-Tn5-gfp in Sinorhizobium meliloti
RM1021::lacIq, Pseudomonas putida SM1443, Pseudomonas aeruginosa
PAO1::lacIq, and Cupriavidus necator JMP228::lacIq, all harboring a lacIq
allele, led to various levels of fluorescence upon IPTG induction, from
strong in P. putida to very weak in C. necator and even no signal in
P. aeruginosa (not shown), which already indicates a strong species
dependence for gfp expression (Fig. 4D, E, and F). Furthermore, the in-
troduction of plasmid pB10::mini-Tn5-gfp in C. metallidurans AE815,
for which the IPTG induction is dispensable (no lacI gene), also gave a
weak fluorescence (Fig. 4G). The weakness of the fluorescence observed
for P. aeruginosa and C. metallidurans was coherent with the poor lac
promoter expression reported for these bacteria compared to E. coli
(Ribeiro-dos-Santos et al., 2010; Baumberg et al., 1980). In comparison,
the microscopic observation of activated sludge with the same setting
gave an autofluorescence with an intensity comparable to the one
observed for C. metallidurans AE815(pB10::mini-Tn5-gfp) (Fig. 4G
and H). Seeing (i) the strong species-dependence for expressing the
gfp gene, and (ii) the non-negligible autofluoresence of some of the
environmental matrices (here activated sludge), fluorescence-based
approaches did not appear to be fully adequate for monitoring pB10
transfer in complex environmental matrices.
Fluorescence-based approaches have already proven to be useful
for in situ monitoring of plasmid transfer in complex structures such
as biofilms (Haagensen et al., 2002; Król et al., 2011). It even allowed
studying plasmid acquisition at individual cell levels in structured com-
munities, and estimating the influence of parameters as subtle as the
distance between mating partners, their relative orientation, or the
relative cell length of the recipient, for instance (Seoane et al., 2011).
Additionally, apart from being culture independent, the fluorescence-
based approaches present the advantage of being sensitive enough
to enable the identification of rare fluorescent transconjugants in the
vastness of “dark communities”. Nevertheless, these approaches have
mainly been used in complex but pure in vitro systems such as flow
cell biofilms, solely involving the mating partners, and where the proper
expression of the gfp could be tested independently beforehand. As a
matter of fact, we showed here that the fluorescence emitted is limited
to certain bacterial hosts, which restricts the range of possibilities espe-
cially when studying plasmid transfer in complex natural communities
of unknown microbial composition. This is without counting on the
background fluorescence of the environmental matrix that may outbal-
ance the poor expression of gfp in some transconjugants and further
limit the use of fluorescence-based approaches for these particular
applications.
3.4. Detecting rare successful transfer events by molecular-based approaches
Getting around the limitations associated with the detection of rare
transconjugants in natural bacterial communities led us to develop
an alternative methodology based on a molecular biology approach
(Bonot and Merlin, 2010; Merlin et al., 2011). Conjugative transfer of
plasmids is a replicative process, leading to an increasing plasmid
to donor genome ratio that can be monitored by qPCR directly on com-
munity DNA. Because of the high sensitivity of the qPCR, transfer is ex-
pected to be detectable even if rare, and much lower donor inoculum
can be used in order to minimize its impact on the behavior of the indig-
enous microbial community studied. Additionally, this approach also
878 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882

(pB10::mini-Tn5-gfp)

(pB10::mini-Tn5-gfp)
A(f) no IPTG A(pc) E(f) + IPTG E(pc)

S. meliloti CM217
E. coli CM125
(pB10::mini-Tn5-gfp)

(pB10::mini-Tn5-gfp) (pB10::mini-Tn5-gfp)
B(f) + IPTG B(pc) F(f) + IPTG F(pc)

C. metallidurans AE815 C. necator CM224

E. coli CM125
(pB10::mini-Tn5-gfp)

C(f) + IPTG C(pc) G(f) no IPTG G(pc)

E. coli JM109
(pB10::mini-Tn5-gfp)

D(f) D(pc) H(f) H(pc)

P. putida SM1443

+ IPTG Activated sludge

Fig. 4. Fluoresence of pB10::mini-Tn5-gfp bearing bacteria. E. coli CM125(pB10::mini-Tn5-gfp) contains a wild type lacI allele, C. metallidurans AE815 does not possess any lacI gene,
all the other strains are provided with lacIq allele.(f) fluorescence microscopy; (pc) phase contrast image.

offers the advantage of overcoming the limitations associated with both,
the poor culturability of environmental microbes (often considered to
be less than 1%), and the limited host-range in which selective markers
are expressed for the specific selection of transconjugants.
In this study, the levels of pB10 and its initial host E. coli DH5α were
monitored in three different environmental matrices maintained in mi-
crocosms (manure, activated sludge, and WWTP effluents) following
the initial inoculation of the donor bacteria (Fig. 5). Except for the
WWTP effluent microcosms, the relative abundance of E. coli DH5α
and plasmid pB10 decreased at the same pace indicating a simple
donor loss (Fig. 5A1, B1, and C1). Although both parameters decreased
in the WWTP effluent microcosms, the loss of E. coli DH5α was sharper
than the one of pB10, which led to a 25-fold increase in the plasmid to
donor ratio, meaning that pB10 had transferred (Fig. 5D1). In most of
the transfer experiments carried out, the relative abundance of E. coli
DH5α(pB10) started to decrease immediately after inoculation. This
led us to hypothesize that the absence of pB10 transfer could be linked
to a poor initial stability of the donor bacteria, either due to antagonistic
effects from the indigenous bacterial communities and/or due to preda-
tion by protozoa. The latter was evaluated in transfer experiments using
microcosms treated beforehand with a eukaryotic inhibitor. In a first set
of experiments, microcosms were set up with freshly sampled farm ma-
nure that received a single cycloheximide amendment (0, 0.1, or
1 mg.mL−1) before being inoculated with E. coli DH5α(pB10) donor
cells. Fig. 5A shows that cycloheximide strongly reduced the initial
loss of E. coli DH5α(pB10) following inoculation, which moved from
a 2-Log loss in the control experiment to a 1-Log loss and finally a re-
duced loss when cycloheximide was amended at 0.1 and 1 mg.mL−1,
respectively. For both the 0 and 0.1 mg.mL−1 cycloheximide amended
microcosms, the decline of plasmid pB10 paralleled the one of E. coli
DH5α DNA, which indicates the loss of the donor bacteria without
plasmid transfer (Fig. 5A1 and A2). However, in the 1 mg.mL−1
cycloheximide-amended microcosm, the donor bacteria declined con-
tinuously while pB10 remained relatively stably detected in the com-
munity DNA (Fig. 5A3). This resulted in an increasing plasmid to
donor ratio indicating that plasmid transfer had occurred. Surprisingly,
the experiment was repeated on manure sampled at the same location
but at a different period and led to different results (Fig. 5B). In this sec-
ond experiment, the 1 mg.mL− 1 cycloheximide amendment did not
have any observable effects neither on the stability of the donor bacteria
nor on the promotion of pB10 transfer. This demonstrates once again
that relatively rare transfer events could be difficult to reproduce
using similar environmental matrices obtained from two independent
samplings of the same environment. This also shows that an important
parameter such as predation may not always drive the stability of the
donor bacteria and that alternative parameters, such as competition
with indigenous bacteria should also be considered.
The effects of cycloheximide amendments were also evaluated on
activated sludge and WWTP effluent microcosms (Fig. 5C and D). In
both cases, cycloheximide significantly reduced the initial loss of the
donor bacteria. Nevertheless, despite this improved persistence, the
pB10 plasmid to E. coli DH5α ratio remained stable over the course of
the experiments signifying that, contrary to what was observed with
the first manure experiment (Fig. 5A), cycloheximide did not favor the
transfer of pB10. This was even unexpected considering the fact that,
as mentioned earlier, pB10 transfer was observed in the untreated efflu-
ent microcosm (Fig. 5D1). Thus, if eukaryotic predation appears to play
a significant role regarding the initial persistence of the donor bacteria
E. coli DH5α(pB10), it cannot alone explain the lack of transfer in non-
permissive environmental matrices. To the contrary, in certain instances
limiting the predation had a negative effect on pB10 transfer, as seen
with the WWTP effluents. To the best of our knowledge, if the influence
 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882 879

Fig. 5. Transfer of pB10 in environmental communities. Stability of pB10 (squares) and E. coli DH5α (triangles) in manure (A and B, two samples collected independently), activated
sludge (C) and WWTP effluent (D) microcosms, treated or not with cycloheximide (0, 0.1, and 1 mg.mL−1). Each data point is an average of 3 values obtained from triplicate qPCRs.
Error bars represent standard deviation.

of predation is well recognized (Hahn and Höfle, 2001; Rønn et al.,
2002), it remains poorly documented and far from being fully under-
stood regarding plasmid transfers. The experiment presented in Fig. 5
showed that, depending on the population considered, opposite effects
could take place when using eukaryotic inhibitors. This suggests that
eukaryotic predation may influence bacterial plasmid transfers in two
different ways. First by grazing on donor bacteria, eukaryotic predation
may actively reduce the chance of observing a rare transfer event.
This explains why such rare events became observable upon cyclohex-
imide treatment in the first manure microcosm experiment (Fig. 5A).
880 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882
On the other hand, eukaryotic predation is also known for its role in
maintaining the dynamics of the bacterial population (Huws et al.,
2005). Indeed, predator grazing limits the overgrowth of bacterial
population and therefore indirectly contributes to maintain a growing
state that has been demonstrated to be of prime importance for plasmid
transfer (Seoane et al., 2011). Consequently, predation may promote
or limit plasmid transfer depending on whether grazing mainly contrib-
utes to dynamize the bacterial population or to bring down the level of
donors, both phenomena being non exclusive.
Because it is highly specific and it does not rely on gene expression,
the qPCR approach we developed definitively got rid of the limitation
associated with the culture-based selection of transconjugants. Still,
considering an average bacterial genome size of 2 Mbp, a qPCR carried
out with 25 ng of community DNA, and an average copy number of
10 plasmid copies per cell for IncP plasmids, measuring 10 copies of
pB10 brings the detection limit to one plasmid bearing bacterium out
of ~ 107 bacterial genomes screened. This may not appear very high
but, on the other hand, the absence of background makes the data
100% reachable for evaluating horizontal gene transfer as opposed to
culture-based approaches. With the molecular approach, horizontal
gene transfer should be thought slightly differently than usual and
should rather be regarded as a DNA sequence invasion into a communi-
ty metagenome. If plasmid conjugation is concerned, the invasion can
be considered as effective when the plasmid sequence is maintained
in the indigenous community metagenome, which then appears as an
increasing plasmid to donor ratio over time. Considering this, a major
drawback comes with the fact that the implantation of the plasmid se-
quence may be missed by the experimenter simply because the plasmid
to donor ratio does not increase significantly in the event of a very low
level of transfer combined to a too stable persistence of the donor host
in the indigenous population. Ideally, a quick disappearance of the
donor host allows a better evaluation of the plasmid persistence in the
indigenous community metagenome. Nevertheless, as discussed above
this donor disappearance also has to allow time for the plasmid transfer
to occur.
4. Discussion
4.1. Studying plasmid transfer in environmental matrices
Despite the ever-increasing retrospective evidences showing the
dissemination of antibiotic resistance genes in the environment
(Aminov, 2011), detecting an antibiotic resistance in an environmental
isolate does not necessarily mean that it has been acquired by horizontal
gene transfer in the environment of concern. Actually, demonstrating
gene transfer is a difficult matter when it comes to deal with environ-
mental microbial populations, even for plasmids that have been repeat-
edly pointed as highly proficient for promoting gene dissemination.
Plasmid-mediated gene dissemination in environmental communities
has been studied in several ways depending on the global experimental
strategy used and the fundamental question asked. Typically, the exog-
enous isolation of plasmids from environmental communities has been
used for characterizing the mobile fraction of a studied phenotype, like
a community resistome for instance. In such a case, the indigenous
community is used as donor while a selectable recipient is added to
the environmental matrix for isolating transconjugants with acquired
phenotypic traits such as antibiotic resistances. Further analysis of the
transconjugants, including plasmid extraction, RFLP, or replicon typing,
allows characterizing the mobile genetic elements initially involved
in the transferred phenotype (Smalla et al., 2000; Oliveira et al., 2012).
Alternatively, when it comes to study the permissiveness of an environ-
mental community to support the transfer of a known mobile element,
the experimental strategy would rather consist in inoculating the indig-
enous community with a donor bacterium hosting a plasmid of recog-
nizable features. Next, the identification of transconjugants, resulting
from the transfer of the plasmid into either a provided recipient
(Götz and Smalla, 1997; Top et al, 1990) or the indigenous bacteria
(van Elsas et al., 2004; Merlin et al., 2011), should determine the level
of successful transfer events that reflects the permissiveness of an envi-
ronmental community in a given environmental matrix.
To date, demonstrating the transfer of a known plasmid, delivered
by a donor inoculum into indigenous bacteria, has involved three
main approaches (culture-based, fluorescent-based, or molecular-
based), each of which having its own biases and technical limitations,
as illustrated in this study with plasmid pB10. Culture-based ap-
proaches mainly rely on the isolation of transconjugants based on the
acquisition of a particular resistance phenotype. The relatively high
background levels of resistances in environmental samples make the
culture-based selection of transconjugants difficult unless high loads
of donor bacteria are introduced in the environmental samples in
order to outcompete naturally resistant bacteria. Moreover, because
transconjugants are indistinguishable from naturally resistant indige-
nous bacteria, additional screenings are necessary to ascertain the
presence of the plasmid. These screenings could consist in RFLP analyses
on plasmid extracts, southern blot hybridization with plasmid specific
DNA probe (Götz and Smalla, 1997; Fulthorpe and Wyndham, 1992),
or specific PCR typing (Götz and Smalla, 1997; Bonot and Merlin,
2010). Despite involving tedious screening tasks and being focused on
the poorly culturable fraction of the indigenous bacteria, culture-based
approaches remain the only mean to identify the bacterial actors
supporting plasmid transfer. If they have provided the basis of our cur-
rent knowledge on horizontal gene transfer in the environment, we be-
lieve these approaches still suffer from the unrealistic levels of donor
inoculum used when it comes to providing a transfer frequency. In
this respect, fluorescent-based approaches could have been interesting
alternatives as they rely on the production of a fluorescent protein
that is naturally absent from the microbial communities. Nevertheless,
the background level of fluorescence associated with some environmen-
tal matrices, combined with the narrow expression range of the fluores-
cent genes, strongly limits the possibility of monitoring plasmid transfer
using fluorescent-based approaches in complex environments. This
is even particularly troublesome considering the fact that even using
(i) a narrow set of bacterial species, all belonging to the preferred host
class for IncP-1 plasmids (Proteobacteria; Suzuki et al., 2010), and (ii) a
Plac-gfp expression system also originating from the proteobacterial
system, the fluorescence emission displayed severe strain/species-
dependent limitations. These drawbacks have initially motivated us to
explore a molecular-based approach that finally appeared sensitive
enough to detect rare transfer events under low (but realistic) donor in-
oculum size. Theoretically, it remains limited on a quantitative point
of view. Indeed, because it solely relies on the fact that the plasmid
shows a better persistence than its donor host, it does not make the
difference between repetitive transfer events and the growth of
transconjugants. And, since it is analyzed at the level of a community
metagenome, further analyses of transconjugants are not possible. On
the other hand, it allows detecting rare successful transfer events that
should then be interpreted as the invasion and possibly the persistence
of a plasmid sequence in the vastness of the community metagenomes.
In conclusion, we believe this molecular approach to be particularly
valuable at least for two reasons. First, because it can tolerate low
amounts of donor inocula, it can be seen as poorly invasive, thus
allowing the matrix effect to be studied without interference. Second,
despite focusing on rare transfer events, and being poorly quantita-
tive, it remains more exhaustive as it also considers the non-culturable
part of the microbial community, which represents at least 99% of its
content.
4.2. Parameters affecting plasmid transfer
This study also illustrates the great variability inherent to working
with environmental matrices. In spite of this variability, important
parameters, such as the stability of the initial donor inoculum and
 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882
the participation of eukaryotic predation, could emerge from different
environmental matrices. Still, predation cannot, by itself, fully explain
the stability of the donor or the transfer of the plasmid. If plasmid-
mediated gene transfer is a major process for the evolution of bacteria
towards more resistant forms, it also has a cost. The acquisition of a plas-
mid necessarily implies additional physiological burden for the new
host with the maintenance of plasmid DNAs, the production of plasmid
proteins, and the expression of the plasmid accessory functions such
as antibiotic resistances. Altogether, this paradoxically reduces the
fitness of the new host in non selective environments, which should
act against plasmid dissemination, unless compensatory mutations
occur (Harrison and Brockhurst, 2012; Maisnier-Patin and Andersson,
2004; Lenski, 1998). Antibiotic resistance silencing has been shown to
be one of those mutational events beneficial to the host. If such a
phenomenon is to be proficient during the dissemination of pB10, it
could partly explain the poor recovery of transconjugants by culture.
But, on the other hand, for the experiments presented here, resistance
silencing could not explain the poor recovery of donor bacteria, since
(i) they are already pB10-adapted, (ii) their recovery varies from one
experiment to the other (Fig. 3) while using a same inoculum size,
and (iii) abrupt losses of the donors have also been observed using
culture-independent molecular means (Fig. 5). If the fitness cost has
to play a major role, it is then probably during the early period following
the acquisition of the plasmid in newly formed transconjugants. The fit-
ness cost associated with antibiotic resistant plasmids, including those
of the IncP-1 group, has been shown to be variable and depend on the
bacterial species/strain or even the plasmid considered (De Gelder
et al., 2007; Røder et al., 2013; Humphrey et al., 2012). Therefore,
it can be assumed that, by defining its fitness cost level upon plasmid
acquisition, the microbial composition of a community pre-sets its per-
missiveness. If such is the case, communities already adapted to main-
taining some families of plasmids should have better fitness while
supporting the transfer of cognate plasmids, as they should already
display ad hoc compensatory mutations. In this respect, the comparison
of different environmental communities, supporting or not the transfer
of pB10, for their contents in IncP-1 plasmids shall soon provide some
indication as to whether a high load of a given plasmid family reflects
the propensity of the community to be invaded by a plasmid belonging
to the same family.
Acknowledgments
This work was supported by the Agence Nationale pour la Recherche
(ANR ECOTECH 2009, project DEFIVIANDE) and the Zone Atelier
Moselle. We wish to thank Younes Bou Orom (Master 2) for his par-
ticipation, and the members of the EU Cost Action “TD0803: Detecting
evolutionary hot spots of antibiotic resistances in Europe (DARE)” for
fruitful discussions.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.scitotenv.2014.06.070.
References
Adamczyk M, Jagura-Burdzy G. Spread and survival of promiscuous IncP-1 plasmids. Acta
Biochim Pol 2003;50(2):425–53. [Pub med ID:12833168].
Aminov RI. A brief history of the antibiotic era: Lessons learned and challenges for the
future. Front Microbiol 2010;1:134. http://dx.doi.org/10.3389/fmicb.2010.00134.
Aminov RI. Horizontal gene exchange in environmental microbiota. Front Microbiol 2011;
2:158. http://dx.doi.org/10.3389/fmicb.2011.00158.
Andersen JB, Sternberg C, Poulsen LK, Bjørn SP, Givskov M, Molin S. New unstable variants
of green fluorescent protein for studies of transient gene expression in bacteria. Appl
Environ Microbiol 1998;64(6):2240–6. [PMC106306].
Baumberg S, Cornelis G, Panagiotakopoulos M, Roberts M. Expression of the lactose trans-
poson Tn951 in Escherichia coli, Proteus and Pseudomonas. J Gen Microbiol 1980;
119(1):257–62. http://dx.doi.org/10.1099/00221287-119-1-257.
881
Bennett PM. Plasmid encoded antibiotic resistance: Acquisition and transfer of antibiotic
resistance genes in bacteria. Br J Pharmacol 2008;153(Suppl. 1):S347–57. http://dx.
doi.org/10.1038/sj.bjp.0707607.
Blattner FR, Plunkett III G, Bloch CA, Perna NT, Burland V, Riley M. The complete genome
sequence of Escherichia coli K-12. Science 1997;277:1453–74. http://dx.doi.org/10.
1126/science.277.5331.1453.
Bonot S, Merlin C. Monitoring the dissemination of the broad-host-range plasmid pB10
in sediment microcosms by quantitative PCR. Appl Environ Microbiol 2010;76(1):
378–82. http://dx.doi.org/10.1128/AEM.01125-09.
Bonot S, Courtois S, Block JC, Merlin C. Improving the recovery of qPCR-grade DNA from
sludge and sediment. Appl Microbiol Biotechnol 2010;87:2303–11. http://dx.doi.
org/10.1007/s00253-010-2686-0.
Christensen BB, Sternberg C, Andersen JB, Eberl L, Møller S, Givskov M, et al. Establish-
ment of new genetic traits in a microbial biofilm community. Appl Environ Microbiol
1998;64:2247–55. [PMCID: PMC106307].
Dahlberg C, Bergström M, Hermansson M. In situ detection of high levels of horizontal
plasmid transfer in marine bacterial communities. Appl Environ Microbiol 1998;
64(7):2670–5. [PMCID: PMC106442].
Davies J. Microbes have the last word. A drastic re-evaluation of antimicrobial treatment
is needed to overcome the threat of antibiotic-resistant bacteria. EMBO Rep 2007;8:
616–21. http://dx.doi.org/10.1038/sj.embor.7401022.
Davies J, Davies D. Origins and evolution of antibiotic resistance. Microbiol Mol Biol Rev
2010;74:417–33. http://dx.doi.org/10.1128/MMBR.00016-10.
De Gelder L, Vandecasteele FP, Brown CJ, Forney LJ, Top EM. Plasmid donor affects host
range of promiscuous IncP-1beta plasmid pB10 in an activated-sludge microbial
community. Appl Environ Microbiol 2005;71(9):5309–17. http://dx.doi.org/10.
1128/AEM.71.9.5309-5317.2005.
De Gelder L, Ponciano JM, Joyce P, Top EM. Stability of a promiscuous plasmid in different
hosts: No guarantee for a long-term relationship. Microbiology 2007;153(Pt. 2):
452–63. http://dx.doi.org/10.1099/mic.0.2006/001784-0.
Dobrindt U, Hochhut B, Hentschel U, Hacker J. Genomic islands in pathogenic and envi-
ronmental microorganisms. Nat Rev Microbiol 2004;2(5):414–24. http://dx.doi.org/
10.1038/nrmicro884.
Dröge M, Pühler A, Selbitschka W. Horizontal gene transfer among bacteria in terrestrial
and aquatic habitats as assessed by microcosm and field studies. Biol Fertil Soils 1999;
29:221–45. http://dx.doi.org/10.1007/s003740050548.
Dröge M, Pühler A, Selbitschka W. Phenotypic and molecular characterization of conjugative
antibiotic resistance plasmids isolated from bacterial communities of activated sludge.
Mol Gen Genet 2000;263(3):471–82. http://dx.doi.org/10.1007/s004380051191.
Fulthorpe RR, Wyndham RC. Involvement of a chlorobenzoate-catabolic transposon,
Tn5271, in community adaptation to chlorobiphenyl, chloroaniline, and 2,4-
dichlorophenoxyacetic acid in a freshwater ecosystem. Appl Environ Microbiol
1992;58(1):314–25. [PMCID: PMC195209].
Geisenberger O, Ammendola A, Christensen BB, Molin S, Schleifer KH, Eberl L. Monitoring
the conjugal transfer of plasmid RP4 in activated sludge and in situ identification
of the transconjugants. FEMS Microbiol Lett 1999;174(1):9–17. http://dx.doi.org/10.
1111/j.1574-6968.1999.tb13543.x.
Götz A, Smalla K. Manure enhances plasmid mobilization and survival of Pseudomonas
putida introduced into field soil. Appl Environ Microbiol 1997;63(5):1980–6.
[PMCID: PMC1389163].
Haagensen JA, Hansen SK, Johansen T, Molin S. In situ detection of horizontal transfer of
mobile genetic elements. FEMS Microbiol Ecol 2002;42(2):261–8. http://dx.doi.org/
10.1111/j.1574-6941.2002.tb01016.
Hahn M, Höfle MG. Grazing of protozoa and its effect on populations of aquatic bacteria.
FEMS Microbiol Ecol 2001;35:113–21. http://dx.doi.org/10.1111/j.1574-6941.2001.
tb00794.x.
Harrison E, Brockhurst MA. Plasmid-mediated horizontal gene transfer is a coevolu-
tionary process. Trends Microbiol 2012;20(6):262–7. http://dx.doi.org/10.1016/j.tim.
2012.04.003.
Haug MC, Tanner SA, Lacroix C, Meile L, Stevens MJ. Construction and characterization
of Enterococcus faecalis CG110/gfp/pRE25*, a tool for monitoring horizontal gene
transfer in complex microbial ecosystems. FEMS Microbiol Lett 2010;313:111–9.
http://dx.doi.org/10.1111/j.1574-6968.2010.02131.
Haug MC, Tanner SA, Lacroix C, Stevens MJ, Meile L. Monitoring horizontal antibiotic
resistance gene transfer in a colonic fermentation model. FEMS Microbiol Ecol 2011;
78:210–9. http://dx.doi.org/10.1111/j.1574-6941.2011.01149.
Holloway BW, Morgan AF. Genome organization in Pseudomonas. Annu Rev Microbiol
1986;40:79–105. http://dx.doi.org/10.1099/00207713-47-1-132.
Humphrey B, Thomson NR, Thomas CM, Brooks K, Sanders M, Delsol AA, et al. Fitness
of Escherichia coli strains carrying expressed and partially silent IncN and IncP1
plasmids. BMC Microbiol 2012;12:53. http://dx.doi.org/10.1186/1471-2180-12-53.
Huws SA, McBain AJ, Gilbert P. Protozoan grazing and its impact upon population dy-
namics in biofilm communities. J Appl Microbiol 2005;98:238–44. http://dx.doi.org/
10.1111/j.1365-2672.2004.02449.x.
Kessler B, de Lorenzo V, Timmis KN. A general system to integrate lacZ fusions into the
chromosomes of gram-negative eubacteria: Regulation of the Pm promoter of the
TOL plasmid studied with all controlling elements in monocopy. Mol Gen Genet
1992;233:293–301.
Knapp CW, Dolfing J, Ehlert PA, Graham DW. Evidence of increasing antibiotic resistance
gene abundances in archived soils since 1940. Environ Sci Technol 2010;44:580–7.
http://dx.doi.org/10.1021/es901221x.
Król JE, Nguyen HD, Rogers LM, Beyenal H, Krone SM, Top EM. Increased transfer of a mul-
tidrug resistance plasmid in Escherichiacoli biofilms at the air-liquid interface. Appl
Environ Microbiol 2011;77(15):5079–88. http://dx.doi.org/10.1128/AEM.00090-11.
Lal S, Cheema S, Kalia VC. Phylogeny vs genome reshuffling: Horizontal gene transfer.
Indian J Microbiol 2008;48(2):228–42. http://dx.doi.org/10.1007/s12088-008-0034-1.
882 X. Bellanger et al. / Science of the Total Environment 493 (2014) 872–882
Lenski RE. Bacterial evolution and the cost of antibiotic resistance. Int Microbiol 1998;
1(4):265–70. [PMID: 10943373].
Levy SB, Marshall B. Antibacterial resistance worldwide: Causes, challenges and
responses. Nat Med 2004;10(12 Suppl.):S122–9. http://dx.doi.org/10.1038/nm1145.
Maisnier-Patin S, Andersson DI. Adaptation to the deleterious effects of antimicrobial
drug resistance mutations by compensatory evolution. Res Microbiol 2004;155(5):
360–9. http://dx.doi.org/10.1016/j.resmic.2004.01.019.
Merlin C, Bonot S, Courtois S, Block JC. Persistence and dissemination of the multiple-
antibiotic-resistance plasmid pB10 in the microbial communities of wastewater
sludge microcosms. Water Res 2011;45:2897–905. http://dx.doi.org/10.1016/j.
watres.2011.03.002.
Miller JH. A short course in bacterial genetics: A laboratory manual and handbook for
Escherichiacoli and related bacteria. Plainview, N.Y: Cold Spring Harbor Laboratory
Press; 1992. http://dx.doi.org/10.1002/jobm.3620330412.
Molin S, Tolker-Nielsen T. Gene transfer occurs with enhanced efficiency in biofilms and
induces enhanced stabilisation of the biofilm structure. Curr Opin Biotechnol 2003;
14(3):255–61. http://dx.doi.org/10.1016/S0958-1669(03)00036-3.
Musovic S, Dechesne A, Sørensen J, Smets BF. Novel assay to assess permissiveness of a
soil microbial community toward receipt of mobile genetic elements. Appl Environ
Microbiol 2010;76(14):4813–8. http://dx.doi.org/10.1128/AEM.02713-09.
Oliveira CS, Lázaro B, Azevedo JS, Henriques I, Almeida A, Correia A. New molecular
variants of epsilon and beta IncP-1 plasmids are present in estuarine waters. Plasmid
2012;67(3):252–8. http://dx.doi.org/10.1016/j.plasmid.2011.11.002.
Partridge SR. Analysis of antibiotic resistance regions in Gram-negative bacteria. FEMS
Microbiol Rev 2011;35(5):820–55. http://dx.doi.org/10.1111/j.1574-6976.2011.00277.
Porteous LA, Seidler RJ, Watrud LS. An improved method for purifying DNA from soil for
polymerase chain reaction amplification and molecular ecology applications. Mol
Ecol 1997;6:787–91. http://dx.doi.org/10.1046/j.1365-294X.1997.00241.
Ribeiro-dos-Santos G, Biondo R, Quadros Ode F, Vicente EJ, Schenberg AC. A metal-
repressed promoter from gram-positive Bacillus subtilis is highly active and metal-
induced in gram-negative Cupriavidus metallidurans. Biotechnol Bioeng 2010;
107(3):469–77. http://dx.doi.org/10.1002/bit.22820.
Rizzo L, Manaia C, Merlin C, Schwartz T, Dagot C, Ploy MC, et al. Urban wastewater treat-
ment plants as hotspots for antibiotic resistant bacteria and genes spread into the
environment: A review. Sci Total Environ 2013;447:345–60. http://dx.doi.org/10.
1016/j.scitotenv.2013.01.032.
Roberts L, Simpson S. Drug resistance. Sci (Spec issue) 2008;321:355–69.
Røder HL, Hansen LH, Sørensen SJ, Burmølle M. The impact of the conjugative IncP-1 plas-
mid pKJK5 on multispecies biofilm formation is dependent on the plasmid host. FEMS
Microbiol Lett 2013;344(2):186–92. http://dx.doi.org/10.1111/1574-6968.12175.
Rønn R, McCaig AE, Griffiths BS, Prosser JI. Impact of protozoan grazing on bacterial com-
munity structure in soil microcosms. Appl Environ Microbiol 2002;68(12):6094–105.
http://dx.doi.org/10.1128/AEM.68.12.6094-6105.2002.
Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: A laboratory manual. 2nd ed. Cold
Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press; 1989
Schlüter A, Heuer H, Szczepanowski R, Forney LJ, Thomas CM, Pühler A, et al. The 64
508 bp IncP-1β antibiotic multiresistance plasmid pB10 isolated from a waste-
water treatment plant provides evidence for recombination between members of dif-
ferent branches of the IncP-1β group. Microbiology 2003;149:3139–53. http://dx.doi.
org/10.1099/mic.0.26570-0.
Seoane J, Yankelevich T, Dechesne A, Merkey B, Sternberg C, Smets BF. An individual-
based approach to explain plasmid invasion in bacterial populations. FEMS
Microbiol Ecol 2011;75(1):17–27. http://dx.doi.org/10.1111/j.1574-6941.2010.
00994.
Smalla K, Heuer H, Götz A, Niemeyer D, Krögerrecklenfort E, Tietze E. Exogenous isolation
of antibiotic resistance plasmids from piggery manure slurries reveals a high preva-
lence and diversity of IncQ-like plasmids. Appl Environ Microbiol 2000;66(11):
4854–62. http://dx.doi.org/10.1128/AEM.66.11.4854-4862.2000.
Sørensen SJ, Oregaard G, De Lipthay J, Kroer N. Plasmid transfer in aquatic environments.
In: Kowalchuk GA, de Bruijn FJ, Head IM, Akkermans AD, van Elsas JD, editors.
Microbial ecology manual. 2nd ed. the Netherlands: Kluwer Academic Publishers;
2004. p. 1081–108.
Sørensen SJ, Bailey M, Hansen LH, Kroer N, Wuertz S. Studying plasmid horizontal transfer
in situ: A critical review. Nat Rev Microbiol 2005;3(9):700–10. http://dx.doi.org/10.
1038/nrmicro1232.
Springael D, Kreps S, Mergeay M. Identification of a catabolic transposon, Tn4371, carrying
biphenyl and 4-chlorobiphenyl degradation genes in Alcaligeneseutrophus A5. J Bacteriol
1993;175:1674–81. [PMCID: PMC203961].
Suzuki H, Yano H, Brown CJ, Top EM. Predicting plasmid promiscuity based on genomic
signature. J Bacteriol 2010;192(22):6045–55. http://dx.doi.org/10.1128/JB.00277-10.
Thomas CM. The horizontal gene pool. Amsterdam, NL: Harwood Academic Publishers;
2000. http://dx.doi.org/10.1046/j.1365-2540.2001.0902b.
Top E, Mergeay M, Springael D, Verstraete W. Gene escape model: Transfer of heavy
metal resistance genes from Escherichia coli to Alcaligenes eutrophus on agar plates
and in soil samples. Appl Environ Microbiol 1990;56(8):2471–9. [PMID: 2206101].
van Elsas JD, Bailey MJ. The ecology of transfer of mobile genetic elements. FEMS Microbiol
Ecol 2002;42:187–97. http://dx.doi.org/10.1111/j.1574-6941.2002.tb01008.
van Elsas JD, Boersma H, Smit E. Detection of bacterial conjugation in soil. In: Kowalchuk
GA, de Bruijn FJ, Head IM, Akkermans AD, van Elsas JD, editors. Microbial
ecology manual. 2nd ed. the Netherlands: Kluwer Academic Publishers; 2004.
p. 1141–50.
Wan Z, Varshavsky J, Teegala S, McLawrence J, Goddard NL. Measuring the rate of conju-
gal plasmid transfer in a bacterial population using quantitative PCR. Biophys J 2011;
101:237–44. http://dx.doi.org/10.1016/j.bpj.2011.04.054.
Yim G, Wang HH, Davies J. The truth about antibiotics. Int J Med Microbiol 2006;296(2-3):
163–70. [PMID: 16503195].