Antibiotics Promote Escherichia coli-Pseudomonas aeruginosa Conjugation through Inhibiting Quorum Sensing

ABSTRACT

The effect of antibiotics on horizontal gene transfer (HGT) is controversial, and the underlying mechanism remains poorly understood. Here, using Escherichia coli SM10λπ as the donor strain, which carries a chromosomally integrated RP4 plasmid, we investigated the effect of antibiotics on conjugational transfer of a mobilizable gentamicin (Gm) resistance plasmid. The results showed that an exposure to gentamicin that restricted the survival of recipient cells significantly enhanced SM10λπ-Pseudomonas aeruginosa PAO1 conjugation, which was attenuated by a deficiency of lasI-rhlI, genes associated with the generation of the quorum sensing signals N-acyl homoserine lactones (AHLs) in PAO1, or the deletion of the AHL receptor SdiA in SM10λπ. Subsequent mechanistic investigations revealed that a treatment with Gm repressed the mRNA expression of lasI and rhlI in PAO1 and upregulated traI expression in SM10λπ. Moreover, PAO1 treated with other quorum sensing (QS)-inhibiting antibiotics such as azithromycin or chloramphenicol also showed a conjugation-promoting ability. On the other hand, when using non-AHL-producing E. coli strain EC600 as the recipient cells, the promoting effect of Gm on conjugation could not be observed. These data suggest that AHL-SdiA contributes to the effectiveness of antibiotics on plasmid conjugation. Collectively, our findings highlight the HGT-promoting effect of antibiotics and suggest quorum sensing as a promising target for controlling antibiotic resistance dissemination. These findings have implications for assessing the risks of antibiotic use and developing advisable antibiotic treatment protocols.

INTRODUCTION

Horizontal gene transfer (HGT) mediated by plasmid conjugation plays a significant role in the spread of antibiotic resistance and pathogenicity determinants. Conjugation appears to be a universally conserved process. It requires the elaboration of a pilus in the donor cell that is assembled by a type IV secretion system (T4SS), comprising the transferosome encoded by the mating pair formation (Mpf) genes, and the Dtr system, comprising DNA processing and transfer process genes (1, 2). Thus, conjugation is regarded as a function of the donor cell, which identifies a suitable recipient cell and transports the DNA into it in an active donor-dependent manner (3). However, the introduction of novel genes in a preexisting well-tuned genetic background is also a source of genetic conflict (4). The effect of recipients in conjugation has been studied (5, 6) but will not be addressed in this paper.

The selective pressures caused by increases in the use and misuse of antibiotics in medicine and animal feedstuffs account for the spread of antibiotic resistance genes (7). There is a substantial body of evidence suggesting that the subinhibitory concentrations of antibiotics may significantly increase the frequency of horizontal transfer both under experimental conditions and in the guts of animals (8). For instance, the conjugative transfer of a tetracycline resistance plasmid in Staphylococcus aureus was enhanced 100- to 1,000-fold by exposure to subinhibitory concentrations of beta-lactam (9). The preincubation of donor Bacteroides cells with subinhibitory tetracycline accelerated the conjugal transfer of CTnDOT (10, 11). The subinhibitory concentrations of ciprofloxacin and erythromycin measured in hospital sewage can increase conjugative transfer frequencies of gentamicin resistance plasmids from staphylococci (12). In the guts of gnotobiotic rats, the presence of tetracycline or erythromycin resulted in a higher transfer rate of conjugative transposon Tn916 or plasmid pLFE1, respectively (13, 14). However, there are also converse statements that do not support the idea that antibiotics promote conjugation (9, 15). One report investigating the mechanism has shown that the SOS response mediates antibiotic-induced conjugation (16). An extensive investigation of the bacterial response to antibiotics will help to disclose the elaborate regulatory network of HGT.

A number of bacterial pathogens are capable of conducting bacterial “cell-to-cell” communication with the same and/or different species via diffusible chemical compounds, a process commonly referred to as quorum sensing (QS) (17, 18). In Pseudomonas aeruginosa, an opportunistic nosocomial pathogen that causes severe infections, especially in patients with compromised immune systems, the most commonly described signaling molecules are N-acyl-homoserine lactones (AHLs), including N-(3-oxododecanoyl)-l-homoserine (3-oxo-C12-HSL) and N-butanoyl-homoserine lactone (C4-HSL). Our previous studies revealed that 3-oxo-C12-HSL or C4-HSL produced by P. aeruginosa inhibited interspecies conjugation by activating SdiA in Escherichia coli, suggesting an inhibitory function of QS on conjugation. However, whether QS is under the control of antibiotics and further regulates conjugation is still unknown. In this study, the effect of antibiotics on the conjugative transfer of resistance genes carried by plasmids was observed. We found that exposure to quorum sensing inhibitory antibiotics resulted in a significant increase in the conjugation of the mobilizable shuttle.

RESULTS

Exposure to antibiotic stress enhanced conjugation ability of PAO1.

In our SM10λπ-PAO1 and SM10λπ-EC600 conjugation models, PAO1 and EC600 are sensitive to gentamicin (Gm). Therefore, when PAO1 and EC600 are under the stress of exposure to Gm, they are more likely to acquire a Gm resistance gene through conjugative transfer of pUCP24T from SM10λπ (Fig. 1). To explore whether Gm induced SM10λπ-PAO1 or SM10λπ-EC600 conjugation, we first determined the appropriate concentration of Gm that might threaten the growth of PAO1 and EC600. As shown in Fig. 2A, when the concentration of Gm gradually increased, the growth of PAO1 and EC600 was dramatically inhibited at concentrations of approximately 0.25 μg/ml and 2 μg/ml, respectively. Lower or higher concentrations of Gm either hardly affected PAO1 or EC600 growth or showed germicidal efficacy, respectively. Accordingly, the following tests revealed that treatment with 0.3 μg/ml Gm for 6 h enhanced SM10λπ-PAO1 conjugation, while lower or higher concentrations of Gm (such as 0.03 or 3 μg/ml, respectively) had no such effect (Fig. 2B). Intriguingly, for SM10λπ-EC600 cocultures, the conjugation efficiency was not elevated but was reduced by Gm (Fig. 2B). These data suggest there is a promoting effect of a given antibiotic on the bacterial conjugative transfer of resistance genes in a recipient-dependent manner.

FIG 1.

Scheme showing the experimental design to test the effect of antibiotics on plasmid conjugation. (A) SM10λπ-PAO1 conjugation model. (B) SM10λπ-EC600 conjugation model. Donor and recipient cells were mixed 1:1 and mated in LB broth with or without Gm for 6 h. Cells were then resuspended and spread on plates containing 30 μg/ml Gm plus 100 μg/ml AMP for SM10λπ-PAO1 or 30 μg/ml Gm plus 50 μg/ml Rif for SM10λπ-EC600 transconjugants. See Materials and Methods for details.

FIG 2.

Antibiotics promoted SM10λπ-PAO1 but not SM10λπ-EC600 conjugation. (A) Effect of gentamicin (Gm) on the growth of PAO1 or EC600. PAO1 or EC600 was cultured with serial dilutions of Gm at 37°C for 6 h. (B) Effect of Gm on SM10λπ-PAO1 or SM10λπ-EC600 conjugation. Donor SM10λπ and recipient PAO1 or EC600 cells (0.5 × 10⁷ CFU/ml each) were mated in the presence of indicated concentrations of Gm at 37°C for 6 h. Values are means ± SEMs from at least three independent experiments; **, P < 0.01.

AHL-SdiA mediated the antibiotic-induced conjugation.

We previously disclosed that quorum sensing AHL molecules produced by P. aeruginosa inhibited SM10λπ-PAO1 conjugation by activating SdiA of SM10λπ, while EC600 lacks AHL synthase and does not produce AHLs (19). Therefore, we explored whether AHL-SdiA determined the effectiveness of antibiotic-induced SM10λπ-PAO1 and SM10λπ-EC600 conjugation. Using an sdiA deletion strain (SM10λπ ΔsdiA), we found that the presence of Gm promoted both SM10λπ-PAO1 and SM10λπ ΔsdiA-PAO1 conjugation. However, compared to that in the control group (SM10λπ-PAO1), the promoting effect of Gm on SM10λπ ΔsdiA-PAO1 conjugation was much diminished (Fig. 3A). Similarly, the deletion of lasI or rhlI in PAO1 (PAO1 ΔlasI or PAO1 ΔrhlI, respectively) also attenuated Gm-induced SM10λπ-PAO1 conjugation (Fig. 3B).

FIG 3.

Antibiotics promoted SM10λπ-PAO1 conjugation via a mechanism dependent on AHL-SdiA signaling. (A) Deficiency of sdiA in SM10λπ attenuated Gm-induced SM10λπ-PAO1 conjugation. PAO1 cells were mated with SM10λπ or SM10λπ ΔsdiA (deficiency of sdiA) cells in LB broth with Gm (0.3 μg/ml) or without (vehicle) for 6 h. (B) Deficiency of lasI or rhlI in PAO1 attenuated Gm-induced SM10λπ-PAO1 conjugation. SM10λπ cells were mated with PAO1, PAO1 ΔlasI or PAO1 ΔrhlI (deficiency of lasI or rhlI, respectively) cells in LB broth with Gm (0.3 μg/ml) or without (vehicle) for 6 h. (C) Gm repressed lasI-rhlI expression in PAO1 and promoted traI expression in SM10λπ. SM10λπ and PAO1 cocultures were treated with 0.3 μg/ml Gm for 6 h, followed by real-time PCR analysis. The rpoD genes of PAO1 and SM10λπ were used as an internal control for lasI-rhlI and traI, respectively. *, P < 0.05; **, P < 0.01.

Moreover, repressing traI expression in the donor cells was a critical mechanism behind the inhibitory effect of AHLs on conjugation. Therefore, we measured the expression of lasI, rhlI, and traI in SM10λπ and PAO1 mixed cultures. A real-time PCR analysis showed that treatment with 0.3 μg/ml Gm repressed the mRNA expression of lasI and rhlI in PAO1 and upregulated traI expression in SM10λπ (Fig. 3C), which was attenuated by the deletion of sdiA in SM10λπ (see Fig. S1 in the supplemental material).

These results suggest the promoting effect of antibiotics on SM10λπ-PAO1 conjugation is partially mediated by AHL-SdiA.

QS-inhibiting antibiotics increased conjugation ability of PAO1.

To validate the role of QS in antibiotic-regulated conjugation, the positive association between reduced QS levels and the increased conjugation ability of PAO1 was further confirmed by using 3 other antibiotics, namely, ampicillin (AMP), azithromycin (AZM), and chloramphenicol (CHL), whose effects on QS in P. aeruginosa have been explored (20). The MICs of AZM and CHL were found to be 128 μg/ml and 64 μg/ml, respectively, while the addition of AMP (256 μg/ml) hardly affected the growth of PAO1 (see Fig. S2). Further investigation revealed that one-fourth MICs of AZM and CHL inhibited lasI and rhlI expression and promoted SM10λπ-PAO1 conjugation, but AMP had no such effect (Fig. 4A and B).

FIG 4.

Antibiotics inhibited quorum sensing and promoted conjugation. (A) AZM and CHL repressed lasI-rhlI expression. (B) AZM and CHL promoted SM10λπ-PAO1 conjugation. For (A) and (B), PAO1 was treated with 256 μg/ml AMP, 32 μg/ml AZM, or 16 μg/ml CHL, followed by real-time PCR or conjugation analysis. (C) Schematic overview of the antibiotic-lasI-rhlI-HSL-SdiA-traI pathway. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

On the basis of the evidence provided here and in our previous study, we suggest the following regulatory pathway. Antibiotics suppress lasI-rhlI expression and attenuate AHL signaling in P. aeruginosa, which impairs the activity of AHL-SdiA chemical signaling to inhibit the transcription of traI in donor E. coli cells and in turn promotes E. coli-P. aeruginosa conjugation (Fig. 4C).

DISCUSSION

The rate of HGT has increased tremendously over the past few decades due to the selective pressure derived from human antibiotic use. Of the three canonical mechanisms for HGT, conjugation is thought to play a major role in the worldwide spread of antibiotic resistance (7). Many publications have focused on characterizing the impact of antibiotics on the transfer of conjugative plasmids. However, the conclusions are not comparable (9, 12–15, 21). A possible explanation for these divergent results could be the differences in conjugation models (donor and recipient cells) and types and concentrations of antibiotics, as well as the methods used to detect transconjugants. Here, using more than one conjugation model and antibiotic, we found that different concentrations of Gm, AZM, or CHL showed distinct effects on SM10λπ-PAO1 or SM10λπ-EC600 conjugation. For Gm, the plasmid encoded a resistance gene, but for the other two, no resistance gene was present. Thus, the effect seemed to be independent of the resistance genes carried in the donor. However, our findings confirmed that the contribution of antibiotics to the promotion of HGT was not only affected by the types and concentration of antibiotics but also relied on the donor and recipient cells.

The underlying mechanisms for antibiotic-induced HGT via conjugation remain largely unknown. However, two broad ways are proposed, namely, (i) through increasing the transfer frequency by inducing global cellular responses of host cells and the expression of conjugation machinery and (ii) by a growth advantage of transconjugants (13, 15). We speculate that the increase in SM10λπ-PAO1 transconjugants resulted not only from a growth advantage during selective pressure but also from a direct effect on the conjugation on the basis of the following evidences. First, in our SM10λπ-PAO1 and SM10λπ-EC600 conjugation models, both PAO1 and EC600 are exposed to Gm that may restrict their survival, meaning that the selective pressures for the growth advantage of transconjugants, if present, were identical. However, the number of resistance plasmid-carrying transconjugants of EC600 did not increase. Second, we found that AHL-SdiA partially determined the effectiveness of antibiotic-induced SM10λπ-PAO1 or SM10λπ-EC600 conjugation. In SM10λπ and PAO1 mixed cultures, the exposure to Gm repressed the mRNA expression of lasI and rhlI in PAO1 and upregulated conjugation-associated traI expression in SM10λπ, indicating that both host cellular responses and conjugation machinery were involved in antibiotic-induced conjugation. Nevertheless, we do not know the underlying mechanisms that enabled the antibiotics to inhibit QS. The results shown in Fig. 2 indicated that at the antibiotic concentration used, the population of PAO1 recipient cells decreases significantly, and this decrease might have caused the lower transcription of lasI and rhlI shown in Fig. 3. However, we calculated the numbers of recipient and transconjugant PAO1 cells in the SM10λπ and PAO1 cocultures and found that transconjugant numbers were much lower than the numbers of recipient cells. Meanwhile, the growth restriction effect of Gm on PAO1 disappeared in PAO1-SM10λπ coculture system (see Fig. S3 in the supplemental material).

It has been suggested that antibiotics play a regulatory role in nature at low concentrations unlike the lethal concentrations used in clinical therapy. Sub-MICs of antibiotics are potent modulators of transcription and change physicochemical properties of microorganisms, including the generation of unintended effects on conjugational transfer in various bacteria (22, 23). Here, our data suggest that a low concentration of Gm represses quorum sensing in P. aeruginosa and finally facilitates conjugation, which can be attenuated by the deletion of lasI or rhlI. Nevertheless, the blocking of AHL-SdiA signaling did not abolish Gm-induced conjugation absolutely (Fig. 2), indicating the existence of other molecular mechanisms that jointly regulate this course. Future studies are needed to illuminate these.

Although plasmid-driven conjugation has provided an important mechanism of bacterial evolution, it also places a burden on the host. How host bacteria minimize the metabolic cost while maximizing the benefits of conjugation remains a fascinating question. Our previous work showed that donor SM10λπ and recipient PAO1 cells cooperated to inhibit conjugation via the quorum sensing system and its downstream pathway LasI/RhlI-HSL-SdiA-TraI. In the present study, we found that under the stress of exposure to antibiotic, P. aeruginosa downregulated lasI-rhlI mRNA expression to enhance conjugative transfer and acquire resistance genes from E. coli. Our findings reveal a novel mechanism by which host cells balance the burden and benefit from conjugation.

In conclusion, our findings disclose the significance of antibiotics in regulating HGT and the underlying mechanisms. These findings have implications for assessing the risks of antibiotic use and developing advisable antibiotic treatment protocols.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The bacterial strains and plasmids used in this study were described in our previous work (19, 24). Bacteria were grown in Luria-Bertani (LB) medium or on LB plates containing 1.5% agar unless otherwise indicated. If required, antibiotics were added to LB plates at the following final concentrations: ampicillin (AMP), 100 μg/ml; gentamicin (Gm), 30 μg/ml; and rifampin (Rif), 50 μg/ml.

MIC determinations.

MIC values were determined by the broth microdilution method in polystyrene microtiter plates (no. 3599; Costar) according to CLSI protocol M07-A8 using cation-adjusted Mueller-Hinton broth (CAMHB) (no. 11865; Oxoid). MICs were interpreted visually after incubating at 37°C for 16 h.

Conjugation experiments.

E. coli SM10λπ(pUCP24T), in which the RP4 plasmid is integrated in the chromosome, was used as the source of donor cells. PAO1 or EC600 were used as the source of recipient cells. The mobilizable plasmid pUCP24T that contains a gene cassette (aacC1) conferring gentamicin resistance is not able to transfer on its own but can transfer by using the conjugative apparatus of E. coli SM10λπ. For mating experiments, equal amounts of donor and recipient cells (0.5 × 10⁷ CFU/ml, counted using the Sysmex UF-1000i automated urine particle analyzer; Tokyo, Japan) were mixed in 200 μl LB with or without the indicated antibiotics at 37°C in a 96-well plate. After 6 h of mating, the cultures were vigorously mixed, and 30-μl aliquots of each conjugation mixture were spread on selective plates. The numbers of transconjugant colonies were counted after an overnight incubation at 37°C.

Real-time PCR.

Total RNA was extracted using RNAiso Plus reagent (TaKaRa, Dalian, Liaoning, China). Reverse transcription (1 μg of total RNA) was performed with the PrimeScript RT reagent kit (TaKaRa, Dalian, Liaoning, China). The cDNA was subjected to quantitative PCR (qPCR) on a ViiA 7 Dx system (Applied Biosystems, Foster, CA, USA) using SYBR green qPCR master mixes (TaKaRa, Dalian, Liaoning, China). The expression levels of target genes were normalized to that of the internal control gene (rpoD) using the 2^-ΔΔCt method. The sequences of the primers are listed in Table S1.

Statistical analysis.

Data are expressed as the means ± standard errors of the means (SEMs) from at least three independent experiments. The differences between groups were analyzed using the Student t test when two groups were compared or a one-way analysis of variance (ANOVA) when more than two groups were compared. All analyses were performed using GraphPad Prism, version 5 (GraphPad Software, Inc., San Diego, CA, USA). All statistical tests were two-sided; P values of <0.05 were considered statistically significant.