Correspondence: Mette Burmølle, Section of Microbiology, University of Copenhagen, Universitetsparken 15, bygn 1, 2100 Copenhagen Ø, Denmark. Tel.: +45 40220069; e-mail: email@example.com
Horizontal gene transfer by conjugation has been reported to increase overall biofilm formation. Biofilm is considered a hot spot for plasmid transfer, and it has been found that social interactions during biofilm formation can increase the biomass. In this study, we demonstrate a contrast to previous studies by showing that the conjugative IncP-1 plasmid pKJK5 influences biofilm formation negatively. The results showed that a co-culture (Pseudomonas putida, Kluyvera sp., and Escherichia coli) formed significantly more biofilm than the strains did individually. When pKJK5 was inserted into P. putida, biofilm formation was significantly reduced compared with the co-culture without plasmid. A nonconjugative version of pKJK5 was also used, and the biofilm formation was restored. Visualization with the BioFlux 1000 facility showed that the presence of pKJK5-containing P. putida in the co-culture led to a changed biofilm structure, where the cells showed a higher tendency to attach to other cells rather than surfaces. This study thus indicates that the presence of conjugative plasmids in some species may decrease the surface-associated biofilm formation of a mixed co-culture by facilitating cell–cell attachment with reduced surface attachment as the consequence.
Bacteria have been found to form biofilm in most natural environments. Biofilm consists of a self-produced polymeric matrix composed mainly of exopolysaccharides, proteins, and nucleic acids in which bacteria are embedded (Sutherland, 2001) and exist as single or multispecies communities. Most naturally occurring biofilms consist of multiple species. This structure has several advantages for the bacteria, for example, increased resistance against antimicrobial agents when compared to bacteria in the planktonic state (Cowan et al., 2000). Social interactions found in biofilm have also been reported to cause an overall increase in biofilm biomass and in the overall resilience of the gathered species (Burmølle et al., 2006). Furthermore, biofilm has been suggested as a hot spot for horizontal gene transfer (HGT; Sørensen et al., 2005) because of the close proximity of the cells and the ability of the matrix to concentrate various chemical compounds as, for example, extracellular DNA and communication signals (Jefferson, 2004). This can however be influenced by many different factors. The HGT-mediated communal gene pool is available to the bacteria in the biofilm and allows them to adapt and propagate genetic traits intra- and interspecifically. It is thus highly relevant to study the impact of HGT on mixed biofilm and thereby connect these two well-investigated research areas.
The horizontal gene pool includes mobile genetic elements, such as plasmids, which can be transferred through cell contact (Thomas & Nielsen, 2005). Biofilm provides perfect conditions for mobile genetic elements because of the stable environment, which enables such elements to transfer and persist within the biofilm. Development of biofilm has conversely been suggested to be reinforced by conjugative plasmids (Ghigo, 2001; Reisner et al., 2006). Cell surface attachment has also been showed to be assisted by an IncF plasmid by inducing curli production in Escherichia coli (May & Okabe, 2008), and a mutually beneficial feedback mechanism between biofilm and plasmids has been proposed as a possibility (Madsen et al., 2012). The multispecies biofilm provides a broader host range for the plasmids (de la Cruz-Perera et al., 2013), and the gene shuffling accelerates microbial adaption to fluctuations in the environment. Exchanges between bacteria via HGT allow for many possible scenarios of interaction in the biofilm, and plasmids may interact differently with different hosts, which could induce various trade-offs to a positive connection between biofilm and plasmids. The conjugation process and the effect on the fitness of the bacteria harboring the plasmid can be complex and are believed to influence the plasmid host in different ways.
The aim of this study was to investigate the link between HGT and biofilm formation as earlier studies have shown a positive correlation between plasmid conjugation and biofilm formation. The conjugative IncP-1 plasmid pKJK5 was used as it has been shown not to possess any biofilm-promoting genes (Bahl et al., 2007a), making it possible to investigate the impact of conjugation on biofilm development along with the impact of the plasmid on the hosts engaged in biofilm development in a multispecies co-culture.
Materials and methods
Bacterial strains, plasmids, and media
The bacterial strains and plasmids used in this study are listed in Table 1. Luria–Bertani (LB) broth (Sambrook et al., 1989) was used for growth of the bacteria. Pseudomonas putida was selectively grown on minimal media (de Lipthay et al., 2000) supplemented with 0.4% butanol as carbon source. 15 g L−1 agar was used when solid media were needed. Bacterial cultures were grown at 30 °C on a rotary shaker at 250 r.p.m. Antibiotics (Sigma) used when needed were ampicillin (AMP, 100 μg mL−1); kanamycin (KAN, 50 μg mL−1); nalidixic acid (NAL, 100 μg mL−1); rifampicin (RIF, 100 μg mL−1); tetracycline (TET, 20 μg mL−1); streptomycin (STR, 100 μg mL−1); and trimethoprim (TMP, 20 μg mL−1). Phosphate-buffered saline (PBS) was used for washing cells and dilutions (Sambrook et al., 1989).
The role of the conjugative IncP-1 pKJK5 was investigated by introducing it into E. coli, Kluyvera sp., and P. putida as follows. Overnight cultures were grown separately at 30 °C on a shaker (250 r.p.m.). The strains were washed three times by centrifugation (5000 g, 5 min) and resuspended in LB in same volume. Donor and recipient were mixed and left at 30 °C for two hours. The cells were then washed three times in PBS, serial diluted, and plated on selective agar plates. When Kluyvera sp. was the recipient, 0.2-μm-pore-size-mixed cellulose ester filters (Advantec MFS, Inc., Pleasanton, CA) placed on LB agar plates were used instead. The plates were incubated overnight at 30 °C. Afterward the filter was vortexed in PBS and plated. Donor and recipient cells were also plated on the selective plates as negative controls.
Plasmid preparation of pMIB8 (Plasmid Mini AX kit; A&A Biotechnology) was performed, and the purified plasmid was introduced into E. coli and Kluyvera sp by electroporation. Another approach was used with P. putida; pMIB8 was first transferred by electroporation into E. coli S17.1 and then by conjugation into P. putida. All transformants were restreaked twice. RP4 was already present in P. putida in our in-house strain collection. The presence of the correct plasmids was further verified in the strains by plasmid purification followed by visualization on agarose gels (1%).
Pseudomonas putida, E. coli, and Kluyvera sp. with and without pKJK5 were grown in total volumes of 160 μL in triplicates for 13 h at 30 °C. Optical density (OD) was registered with an EL 340 microplate reader (OD590) every 15 min (with shaking prior to readings). The results were statistically analyzed by t-test with a confidence level of > 95% and further verified with cultures (100 mL, two replicates) grown with constant shaking and OD600 measurements every 20 min.
Quantification of biofilm formation with pKJK5
The biofilm formation of E. coli, Kluyvera sp., and P. putida, individually and co-incubated in mixed co-cultures with and without pKJK5, was assessed by a modified version of the classic crystal violet (CV) assay described by O'Toole & Kolter (1998). The modified version, described as a Calgary biofilm device, uses a lid that has 96 pegs that fits to a 96-well plate and allows biofilm formation on each peg (Ceri et al., 1999). The aggregated biofilm formation was measured by the CV assay. As CV binds to all negatively charged molecules, including nucleic acids, polysaccharides, and other components present on cell surfaces and in the biofilm matrix, this measurement represents an assessment of the whole biofilm. To ensure that the strains were at the same state at the beginning of each experiment, tubes with 100 μL aliquots of an overnight culture were stored in 20% glycerol at minus 80 °C and used as culture inoculum in 5 mL LB. The culture was incubated for 19.5–20 h. This procedure was followed for all strains used for this experiment to enable comparisons between the different biofilm formations. The cells were then washed as described above and adjusted to 0.15 OD by measuring at 600 nm and dilution in appropriate volumes of LB. A total volume of 150 μL was transferred to the wells in 96-well microtiter plate (NUNC, Roskilde, Denmark), and the lid (TSP; NUNC) was placed on top. The biofilm was then grown in LB for 24 h at 30 °C in a plastic bag. The staining and quantification were carried out by washing the pegs in washing trays with 55 mL PBS three times, staining for 20 min in 160 μL aqueous CV solution (1%), washing three times again, and moving the pegs to 200 μL ethanol (96%) for 20 min. The absorbance was measured using an EL 340 microplate reader (Bio-Tek Instruments) at 590 nm [for spectrophotometric measurements above 1.1, the CV–ethanol suspensions were diluted 1 : 1 with ethanol (96%)]. Absorbance measurements of wells with only media were subtracted from the measured values. A minimum of four replicates were included, and from these, the averages and standard errors were calculated. The results were evaluated for statistical verification by t-test with a confidence level of > 95%, to verify the significance of the findings.
Determining biofilm formation by BioFlux
The mono- and multispecies biofilm formation with different P. putida strains used in this study (+/− pKJK5) was analyzed in a BioFlux™ 1000 device (Fluxion Biosciences, South San Francisco, CA). The experiments were carried out by first preparing an overnight culture by washing and adjusting OD before loading it in a BioFlux 48-well plate prepared by being pre-exposed to LB medium. The cells were then allowed to settle for 30 min before the flow was started at 0.15 dyn cm−2. The plate was run for c. 24 h at 30 °C, while micrographs of the wells were acquired at the same location every 20 min to follow the growth.
The overall purpose of the presented study was to explore the ability of plasmid-free and plasmid-carrying cells to form biofilm and reveal potential synergistic or antagonistic effects when co-incubated. The biofilm formation was examined by the classical CV assay but refined using microtiter plates with pegs (Ceri et al., 1999). This assay measured the overall biofilm formation in a static, endpoint analysis assay, while the BioFlux facility allows continuous examination and visualization of the biofilm development under flow conditions, making these two biofilm assays very complementary. The three chosen bacteria differ in their physiology and their biofilm-forming ability, and it was known that they were able to harbor pKJK5.
When the conjugative IncP-1 plasmid pKJK5 was introduced into the strains Kluyvera sp. and E. coli, a small but statistically insignificant increase in their ability of biofilm formation on the pegs was observed compared with their respective, plasmid-free variants (Fig. 1a). In contrast, P. putida showed a general decrease in biofilm formation when carrying the plasmid (Fig. 1a). The growth rates of the strains with and without plasmid were similar (no significant differences), verifying that the observed trends were not caused by the impact of the plasmid on the growth rate.
In multispecies biofilm containing the three strains, significantly more biofilm was produced than the strains did individually (Fig. 1b). The impact of the presence of pKJK5 was found to highly depend on the bacterial host in the mixed biofilm: when P. putida contained pKJK5, a significant reduction in biofilm formation on pegs was observed in the CV assay (Fig. 2) compared with that of the plasmid-free co-culture. In the multispecies co-cultures where Kluyvera sp. or E. coli contained pKJK5, there were no profound effects on the biofilm formation (Fig. 2); however, a small general increase in biofilm biomass was observed for plasmid-containing Kluyvera sp. when compared to the plasmid-free variant. Thus, similar results for plasmid presence in P. putida were obtained for the 3-member community as for the single P. putida strains with and without plasmid. These results were further verified by the BioFlux analysis: multispecies biofilm containing P. putida without pKJK5 formed higher and denser layers of biofilm resulting in a more uniform appearance (Fig. 3). When the plasmid was inserted into P. putida, the biofilm formation did not display this structure, but rather a flat biofilm with nodes of denser concentrations and large, loosely attached cell aggregates were observed (Fig. 3). Both assays showed a clear tendency of reduced surface attachment of multispecies co-cultures where pKJK5-containing P. putida were present compared those containing the plasmid-free P. putida.
To further investigate the plasmid-associated reduction in biofilm biomass observed with P. putida, a nonconjugative version of pKJK5, pMIB8, was used to examine whether the observed phenomenon was associated with expression of the transfer region on the plasmid. pMIB8 contains a knockout in traF, which encodes an essential protease for pilus synthesis (Haase & Lanka, 1997; Bahl et al., 2007b). When this variant was inserted into the P. putida, its ability to form biofilm was significantly increased to a level resembling the plasmid-free variant (Fig. 2). These results indicate that the observed effect is connected to the conjugative system on the plasmid.
To further investigate the effects of different plasmids when present in P. putida in multispecies co-cultures, the experiment was repeated with the IncP-1 plasmid RP4. Similar to the observations for pKJK5, RP4 was observed to have a negative effect on biofilm formation of P. putida (Fig. 4).
The results obtained in this study indicate that the presence of conjugative plasmids influences biofilm formation with respect to biomass and structure. Both in single and mixed species biofilms, there was a significant negative effect of plasmid presence in P. putida on biofilm formation. Surprisingly, we found that the interactions between the host and the plasmid are not always advantageous for the host with regard to biofilm development on surfaces. The obtained results contrast earlier studies, which showed that conjugative plasmids had a beneficial role for biofilm formation (Ghigo, 2001; Reisner et al., 2006). These studies, however, focused on the impact of conjugative plasmid on E. coli biofilm formation, whereas we, in the present study, explore and compare different bacterial species. In fact, the most profound impact of the plasmids is observed when hosted by P. putida, which may explain the observed differences.
The influence of pKJK5 on the ability of P. putida to form biofilm was significantly improved when the nonconjugative version was inserted instead, indicating that the transfer system itself, most likely the conjugative pili, reduces the ability of the host (Pseudomonas) to form biofilm. This was supported by the fact that the RP4 plasmid showed results similar to pKJK5, which indicates that the observed plasmid-induced impact on Pseudomonas is widespread among different conjugative plasmids.
The three bacterial strains tested in this study formed significantly more biofilm when mixed together than individually. This is in accordance with previous findings (Burmølle et al., 2006) and emphasizes that synergistic interactions is a common feature in multispecies biofilm.
In the present study, profound structural differences in P. putida biofilm with and without pKJK5 were observed by use of the BioFlux 1000 device: with the plasmid present, the biofilm forms in more scattered, dense nodes compared to a thicker, more equally dense cell layer without. Based on these observations, we speculate that the presence of the conjugative plasmid could facilitate tight cell–cell attachment, favoring the formation of cell aggregates (flocs) over biofilm formation on a solid, abiotic surface. These aggregates may be loosely attached to the surface-associated biofilm, but escape this when exposed to high flow or mechanical disturbance (e.g. washing of pegs). We did, however, not observe visible differences in cell aggregation when comparing nonshaken cultures of P. putida with and without plasmid. The present study emphasizes the usefulness of combining the quantitative CV assay with methods facilitating structural studies of biofilm morphology; the BioFlux is well suited for this, as this flow system allows visualization of biofilm formation under development (Benoit et al., 2010), and several flow cells can be run in parallel with online monitoring under controlled conditions.
The plasmid-induced change in biofilm formation of P. putida was not observed for Kluyvera sp. and E. coli probably because the strains used in this study only produced small amounts of biofilm. Additionally, it is very likely that the impact of the plasmid is different in different hosts, so the reported observations for P. putida may differ in other biofilm-forming strains.
Conjugative plasmids impact biofilm formation differently depending on a variety of factors including the plasmid host. However, the accessory genes encoded on these plasmids often have more distinct adaptive effects on the biofilm. As examples, plasmid-associated resistance genes are able to rapidly spread in the biofilm resulting in enhanced resilience of the biofilm cells. Additionally, plasmid sequencing has revealed that gene operons mediating fimbriae expression are in several cases present on plasmids, thereby enhancing biofilm formation independently of the effects caused by conjugation (Burmølle et al., 2008, 2012; Madsen et al., 2012). This highlights the importance of biofilm and plasmid interactions.
In conclusion, some conjugative plasmids have the ability to affect biofilm structure and formation, but whether this is a promoting or restraining effect is here shown to depend on the plasmid host. The complex interconnections between plasmids and biofilm formation need to be further studied to reveal the involved mechanisms and their role in multispecies natural biofilms.
This research was supported by funding to S.J.S. by The Danish Council for Independent Research (Natural Sciences), The Danish Council for Independent Research (Technology and Production, ref no.: 09-090701, M.B.), and The Lundbeck Foundation (project j.nr R44-A4384, L.H.H.). H.L.R. was supported by a Novo Scholarship. The authors would like to thank Peter Nikolai Holmsgaard for help on the statistical analysis.