JOURNAL OF BACTERIOLOGY, May 2006, p. 3582–3588 0021-9193/06/$08.00⫹0 doi:10.1128/JB.188.10.3582–3588.2006 Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Vol. 188, No. 10
Synergistic Effects in Mixed Escherichia coli Biofilms: Conjugative Plasmid Transfer Drives Biofilm Expansion Andreas Reisner,1* Brigitte M. Ho ¨ller,1 Søren Molin,2 and Ellen L. Zechner1 Institut fu ¨r Molekulare Biowissenschaften, Karl-Franzens-Universita ¨t Graz, Universita ¨tsplatz 2, A-8010 Graz, Austria,1 and Center for Biomedical Microbiology, BioCentrum-DTU, Bldg. 301, Technical University of Denmark, DK-2800 Lyngby, Denmark2 Received 30 November 2005/Accepted 27 February 2006
Bacterial biofilms, often composed of multiple species and genetically distinct strains, develop under complex influences of cell-cell interactions. Although detailed knowledge about the mechanisms underlying formation of single-species laboratory biofilms has emerged, little is known about the pathways governing development of more complex heterogeneous communities. In this study, we established a laboratory model where biofilm-stimulating effects due to interactions between genetically diverse strains of Escherichia coli were monitored. Synergistic induction of biofilm formation resulting from the cocultivation of 403 undomesticated E. coli strains with a characterized E. coli K-12 strain was detected at a significant frequency. The survey suggests that different mechanisms underlie the observed stimulation, yet synergistic development of biofilm within the subset of E. coli isolates (n ⴝ 56) exhibiting the strongest effects was most often linked to conjugative transmission of natural plasmids carried by the E. coli isolates (70%). Thus, the capacity of an isolate to promote the biofilm through cocultivation was (i) transferable to the K-12 strain, (ii) was linked with the acquisition of conjugation genes present initially in the isolate, and (iii) was inhibited through the presence in the cocultured K-12 strain of a related conjugative plasmid, presumably due to surface exclusion functions. Synergistic effects of cocultivation of pairs of natural isolates were also observed, demonstrating that biofilm promotion in this system is not dependent on the laboratory strain and that the described model system could provide relevant insights on mechanisms of biofilm development in natural E. coli populations. function of a biofilm can be changed dramatically by modulation of the conditions applied. In summary, in vitro studies of single-species biofilms reveal that biofilm formation is a highly complex phenomenon that proceeds via numerous pathways depending on the strain, the nutritional status, and other environmental conditions. Given the complexity observed for single-species biofilms, it is reasonable to expect that the formation and function of heterogeneous surface-attached consortia, present in many environments, consisting of multiple, genetically distinct strains and species, will involve additional mechanisms and degrees of complexity. Based on a limited number of mixed-species model systems investigated thus far, competition for nutrients and commensal metabolic networks are certainly important driving forces for biofilm structure and dynamics (4, 16). Increased cell densities enable quorum-sensing-based group behavior in single-species biofilms, thus, chemical signaling across species borders is likely to add another level of complexity to the regulation of biofilm functions (24). Furthermore, the expression of different adhesins, their cognate receptors, and exopolymeric substances components by individual cell types within the community can contribute to overall biofilm formation. Indeed, in oral biofilms, the best understood multispecies biofilm, a combination of these mechanisms appears to govern coaggregation among resident species leading to sequential development of dental plaque (11). Similarly, it has been shown in studies with Escherichia coli K-12 (9) and Lactococcus lactis (13) that horizontal gene transfer within the biofilm community can introduce additional dynamics due to an enhanced expression of clumping factors during plasmid transmission, which also promotes biofilm formation by their new hosts. In conclusion, information derived from in vitro studies
Prokaryotic biofilms have a profound impact on a diverse range of ecosystems, and formation of these surface-associated communities is a fundamental aspect of microbial ecology and pathogenesis (5, 25). Over the last two decades, structural and phenotypic properties of single-species biofilms formed by gram-positive and gram-negative model organisms have been characterized under numerous, defined in vitro conditions. Benchmark studies aimed at identifying chromosomal and episomal genes that contribute to biofilm formation demonstrated the involvement of certain cell surface components, such as fimbriae, outer membrane proteins, and exopolymeric substances, and contributed to the concept that each species employs a unique genetic program to regulate biofilm development (21). Recently, however, this view has become increasingly challenged by evidence for multiple pathways contributing to biofilm formation by a single species (22, 26) and the absence of evidence for a universal gene expression pattern within global transcriptome analyses of biofilms formed by the same strain (2, 12). Moreover, it has become apparent that genetic diversity within a given species can support biofilm formation, although the repertoire of stimulating factors involved is not universally shared among strains (27). The diversity of experimental biofilm systems employed thus far has limited the potential for comparison between results obtained in different laboratories, yet that plurality has also prompted the important realization that environmental conditions dictate biofilm features (23). It follows that the structure and * Corresponding author. Mailing address: Institut fu ¨r Molekulare Biowissenschaften, Karl-Franzens-Universita¨t Graz, Universita¨tsplatz 2, A-8010 Graz, Austria. Phone: 43-316-380-5624. Fax: 43-316-3809898. E-mail:
[email protected]. 3582
SYNERGISTIC EFFECTS IN E. COLI BIOFILMS
VOL. 188, 2006
of single-species biofilms will not be directly transferable to consortia containing multiple genetically distinct species or strains. Therefore, a clear need exists for the development and investigation of model systems that provide knowledge about the adhesive properties of bacteria in more complex environments. E. coli remains an excellent model organism for gram-negative bacteria. Colonization of abiotic surfaces by E. coli has been studied extensively in vitro, and the complexity of factors that contribute to this phenotype is well documented (8, 31). However, in most in vivo situations, E. coli has to compete and interact with other bacterial species and with other E. coli strains. In the case of the human intestine, one or more resident and several transient E. coli strains are present at any given time, depending on the host’s exposure to E. coli in the environment (1, 19). The effect of interactions between genetically distinct E. coli strains on biofilm formation is not known. The aim of this study was to investigate whether synergistic effects of cocultivation of different E. coli strains on biofilm formation can be detected in a simple in vitro system. To gain insights into the significance and variety of these putative synergistic effects, we have initiated a broad survey where biofilm formation of 403 different E. coli isolates was monitored in the absence or presence of a laboratory K-12 strain. We found that induction of biofilm formation through cocultivation of strains can be observed at a significant frequency. Our initial analysis indicates that one synergistic effector of biofilm formation by natural E. coli isolates is the process of conjugative plasmid transfer. MATERIALS AND METHODS Bacterial strains and media. E. coli MG1655Str (15) is a spontaneous streptomycin-resistant mutant of E. coli MG1655. The E. coli Reference (ECOR) collection (20) was provided by Thomas S. Whittam (Michigan State University, East Lansing, Mich.). A total of 331 human isolates of E. coli used in this study were described previously (27). Bacterial strains were grown in Luria-Bertani (LB) medium or agar containing 5 g NaCl per liter (3). If required, antibiotics were added at following final concentrations: streptomycin sulfate, 100 g ml⫺1; kanamycin sulfate, 50 g ml⫺1. Plasmids. Plasmid pLG272 (10) is a derivative of the natural IncI␣ plasmid ColIb-P9 and was kindly provided by Brian M. Wilkins (University of Leicester, United Kingdom). Transposon Tn5 inserted in the cib gene inactivates production of colicin Ib and confers resistance to kanamycin. Plasmid pAR183 was constructed by -Red-mediated recombination as previously described (28). A tetRA cassette was amplified from pTP802 (17) using primers DblaDcas (5⬘-TG TCACGCTCGTCGTTTGGTATGGCTTCATTCAGCTCCGGTT-GGTATTT CACACCGCATAGC-3⬘) and ar027 (28). Recombination of homologous ends at the 5⬘ end of the oligonucleotides with the natural IncFII plasmid R1 resulted in replacement of the region from bla to cat by tetRA. Correct insertion was verified by PCR using primers homologous to regions adjacent to the recombination sites. Screen for synergistic effects in biofilms. E. coli isolates were grown overnight in 96-well stock plates (U-bottom; BioSterilin) containing 100 l LB medium per well (37°C, 200 rpm). After 20 h, 96-well test plates were prepared containing 150 l of 37°C LB medium or 150 l of an overnight culture (ONC) of E. coli MG1655Str freshly diluted to an optical density at 600 nm (OD600) of 0.03. Test plate cultures were inoculated from the stock plate of E. coli isolates with a 96-pin replicator (Boekel Scientific). Test plates were transferred to large plastic bags to avoid evaporation of medium and incubated at 37°C for 24 h without shaking. Coculture assays containing two E. coli isolates were inoculated in series from two independent stock plates to test plates containing 150 l of 37°C LB medium. Biofilm formation of the E. coli MG1655Str strains isolated from induced biofilms was assayed exactly as described above, except that the stock plates for overnight cultures were inoculated with single colonies of the MG1655Str isolates. All other biofilm assays were performed as described above, except that
3583
test strain ONCs adjusted to an OD600 of 3.0 were filled into the stock plate. Each assay was performed in duplicate with two parallel replicas. Quantification of biofilm formation in vitro. Biofilm formation was assayed by staining the polystyrene-attached cells with crystal violet (CV) semiautomatically utilizing microtiter plate washer AW1 (Anthos Lab). Following removal of bacterial suspensions and two washes with 200 l of 0.9% NaCl solution, surfaceattached cells were incubated with 160 l of 0.1% CV for 15 min. After two washes with 200 l of 0.9% NaCl solution, surface-bound CV was extracted by addition of 180 l of ethanol (96%). Absorbance measurements (595 nm) obtained with microplate reader model 550 (Bio-Rad Laboratories, Inc.) were normalized by values obtained from wells containing sterile medium. Statistical analysis. Paired t tests (one sided) were used to define two subpopulations of E. coli isolates that gave significantly higher A595 values after dissolution of biofilm-bound CV following coculture with MG1655Str compared to (i) the monoculture of the isolate or (ii) the sum of A595 values from monocultures of both isolate and MG1655Str. One-sided t tests were also used to evaluate induction of biofilm formation of reisolated MG1655Str in mono- and coculture or inhibition of biofilm formation due to the presence of a conjugative plasmid in the cocultivated MG1655Str strain. Sigmaplot 2004 (version 9.0; SyStat Software, Inc.) was used to create graphics. Strain proliferation in mixed biofilms. To assess the proliferation of individual strains in biofilms formed during cocultivation, cells were mechanically removed from the polystyrene surface prior to addition of CV using a Q-Tip soaked with 50 l of 0.9% NaCl solution. The Q-Tip was transferred to a polypropylene tube containing 400 l of sterile 0.9% NaCl solution. After vigorous mixing for 2 min, total CFU of the harvested biofilms were determined following serial dilution and growth on LB agar. CFU of MG1655Str in the suspension was determined using LB agar supplemented with streptomycin sulfate. CV staining of the processed microtiter wells confirmed the efficiency of biofilm removal. Multiplex PCR detection of conjugative transfer genes. Presence of conjugative plasmids belonging to the incompatibility group IncI␣ was evaluated by amplifying conserved regions of conjugative transfer genes traJ and pilS. Genespecific primers ar076 (5⬘-GCGAATTCAGTCTATTAGTGACAACAGC-3⬘) and ar077 (5⬘-GCGGATCCTGTTTGGGAGCTACGTATG-3⬘) facilitated amplification of a 775-bp region containing pilS. ar080 (5⬘-CGGGATCCGCTTAT AGGCACTGAA-3⬘) and ar082 (5⬘-CTGTATCGTCCTGTCAACC-3⬘) produced a 303-bp amplicon containing the 5⬘ end of traJ. Primer pair ar083 (5⬘-C AGCAGCCGCGGTAATAC-3⬘) and ar084 (5⬘-CCGTCAATTCATTTGAGTT T-3⬘) allowed amplification of a 408-bp region of chromosomal 16S rRNA genes and was used as an internal positive control. A single bacterial colony of each test strain was suspended in 50 l sterile water, and cells were lysed at 95°C for 10 min. One microliter of the bacterial lysate provided template DNA for PCRs containing 1⫻ PCR buffer, 0.13 mM deoxynucleoside triphosphates, 0.67 M of each oligonucleotide, and 0.4 U of Dynazyme II polymerase (Finnzymes) in a total volume of 15 l. Reaction mixtures were incubated at 94°C for 2 min, followed by 35 cycles of 20 s at 94°C, 30 s at 58°C, 45 s at 72°C, and a final extension for 5 min at 72°C. If the presence of traJ or pilS was indicated by multiplex PCR, the result was confirmed by another PCR using only the two primer pairs ar076-ar077 and ar080-ar082 per reaction. Detection of F-like conjugative transfer genes traA and finO was performed as previously described (27).
RESULTS Induction of biofilm formation by cocultivation of E. coli isolates with E. coli K-12. A high-throughput microdish biofilm model system was applied to determine stimulatory effects of cocultivation of natural E. coli isolates with E. coli K-12 on biofilm formation in vitro. A total of 403 E. coli strains of different origin were included in this study. Seventy-two of these belong to the E. coli reference (ECOR) collection, a representative subset of the E. coli population (20). A total of 331 E. coli isolates originated from feces of healthy or diarrhea-afflicted children, urine from men diagnosed with urinary tract infection, or blood of hospitalized patients with bacteremia. The latter 331 strains are particularly suitable, as they have been analyzed extensively for their ability to form pure culture biofilms (27). LB medium was selected for growth, because we observed in the prior study that the majority of these E. coli isolates formed poor single-species bio-
3584
REISNER ET AL.
J. BACTERIOL.
TABLE 1. Prevalence of induction of biofilm formation by cocultivation among test strains Strain source
No. of strains tested
No. of strains meeting the following criteriona ACo ⬎ AMono
ACo ⬎ (AMono ⫹ AK-12)
ECOR Bacteremia UTI in men Feces—healthy Feces—diarrhea
72 90 68 105 68
33 (46) 33 (37) 32 (47) 53 (50) 38 (56)
11 (15) 11 (12) 6 (9) 14 (13) 14 (21)
Total
403
189 (47)
56 (14)
a
Percentages of strains relative to the number of tested strains are shown in parentheses. ACo, A595 of a given strain after coculture with MG1655Str; AMono, A595 of a given strain after monoculture; AK-12, A595 of strain MG1655Str.
films in this medium in vitro (27), and therefore we reasoned that synergistic effects by cocultivation would be more readily discernible. Test strains were grown in the presence or absence of a two- to threefold excess of E. coli K-12 MG1655Str. After 24 h, biofilm formation was assessed by staining the attached cells with crystal violet (CV). Absorbance monitored at 595 nm (A595) following dissolution of CV was used as a measure for biofilm formation. Paired t tests were performed with the observed A595 values to identify cocultured test strains that formed increased biofilm compared to the monocultures (Table 1). Interestingly, a substantial proportion of strains (189; 47%) exhibited significantly stronger biofilm production in coculture compared to the monoculture of the E. coli isolate (P ⬍ 0.05). Similar results were obtained after a prolonged incubation time of 48 h, indicating that this phenotype is stable over time (data not shown). The synergistic effect of cocultivation was particularly remarkable for a subset of 56 (14%) test strains, since the biofilms formed during cocultivation of these strains with E. coli MG1655Str exceeded the sum of A595 values observed for the two inoculated strains in pure culture (P ⬍ 0.05). The distribution of A595 values for this subset of 56 strains (Fig. 1A) in the presence or absence of MG1655Str reveals that most test strains formed poor biofilms when cultivated alone (means ⫾ standard errors [M ⫾ SE], 0.108 ⫾ 0.014), whereas the additional presence of MG1655Str in the culture led to a profound induction of biofilm formation (M ⫾ SE, 0.806 ⫾ 0.066). MG1655Str alone produced marginal biofilm (M ⫾ SE, 0.155 ⫾ 0.017). We conclude from this initial screen that synergistic effects of cocultivation on biofilm formation can be monitored at considerable frequency in a simple in vitro system. The origin of test strains had no significant impact on the frequency of the observed induction phenotype (2 test, P ⬎ 0.05). The ability to induce biofilm formation in coculture was acquired by the cocultivated K-12 strain. To analyze the nature and variety of mechanisms underlying the induction phenomena, we focused on the subset of 56 test strains that displayed strongest induction in the initial screen. We predicted that if different mechanisms were operating in the population, we would also detect a different quantitative distribution of the two strains in the biofilms. To evaluate this hypothesis, we investigated the prevalence of the test strain and MG1655Str
in the induced biofilms following cocultivation. Surprisingly, in 49 (87%) induced biofilms MG1655Str comprised more than half of the living cells in the biofilms. On average, 74.5% of the cells isolated from the biofilms were identified as MG1655Str (Fig. 1B). This observation indicated that the ability of MG1655Str to form stable cell-surface or cell-cell interactions was stimulated during cocultivation with nearly all of the 56 test strains. We then investigated whether this improved ability was lost or maintained after reisolation of MG1655Str from the induced biofilms. We randomly chose three MG1655Str colonies, isolated and subcultured independently from each coculture investigated, and subsequently tested the capacity of these strains to form biofilm in LB medium. A total of 39 (23%) of the 168 tested MG1655Str isolates produced significantly higher A595 values in a single-species biofilm than the original MG1655Str strain (P ⬍ 0.05) (Fig. 2A). These improved MG1655Str variants originated from 17 (30%) of the 56 original cocultures. However, the observed increase in biofilm formation exhibited by these MG1655Str derivatives (in monoculture) was less than twofold (M ⫾ SE, 1.43 ⫾ 0.19) over the untreated MG1655Str and thus less pronounced than the level of induction observed during cocultivation in the original screen. Therefore, we asked whether growth of the MG1655Str isolates in the presence of untreated MG1655Str would increase the observed effect. All 168 MG1655Str isolates were coinoculated with the original MG1655Str. Biofilm formation of cocultures was compared to a biofilm formed by a pure culture of original MG1655Str inoculated with the same total number of cells as the tested cocultures. A remarkable number (133; 79%) of the tested MG1655Str isolates exhibited a significant induction of biofilm formation (P ⬎ 0.05) (Fig. 2B). Moreover, MG1655Str
FIG. 1. (A) Fifty-six E. coli isolates promote induction of biofilm formation in coculture with E. coli MG1655Str. Box-and-whisker diagrams depict the distribution of A595 values observed for 56 cocultures compared to the A595 distribution of monocultures of test strains and MG1655Str. (B) MG1655Str is dominant in the majority of induced biofilms. The prevalence of living MG1655Str cells in the biofilms formed after cocultivation of 56 test strains with an initial two- to threefold excess of MG1655Str is shown as a box-and-whisker diagram. Boxes range from the 25th to 75th percentile and are intersected by the median line and the broken mean line. Whiskers extending below and above the box range from the 10th to the 90th percentile, respectively. Outliers are indicated.
VOL. 188, 2006
SYNERGISTIC EFFECTS IN E. COLI BIOFILMS
3585
FIG. 2. Biofilm formation of MG1655Str strains isolated from coculture-induced biofilms. Three MG1655Str strains isolated from each of the 56 induced biofilms were grown in monoculture (A) or in the presence of untreated MG1655Str (B) for 24 h before biofilm formation was quantified by CV staining. For simplicity, only A595 results (M ⫾ SE) from one representative isolate per induced biofilm are shown. Reference lines indicate biofilm formation obtained when the original MG1655Str strain was subjected to the same assay conditions. Asterisks denote significant differences between biofilm formation of the MG1655Str reisolate and the original MG1655Str strain (P ⬍ 0.05). Detection of at least one conjugative transfer gene (traA, finO, traJ, or pilS) in each MG1655Str isolate is indicated by shaded bars.
derivatives exhibiting this phenotype originated from 52 (93%) of the 56 induced biofilms of the original cocultures. The average level of biofilm induction (M ⫾ SE, 5.32 ⫾ 2.79) achieved by cocultivation of MG1655Str isolates with the original MG1655Str was comparable to levels reached in the original screen (Fig. 1A). We conclude that most MG1655Str strains isolated from the original mixed biofilms had acquired a heightened ability to promote biofilm formation. This ability was exhibited at high frequency in coculture with untreated MG1655Str. Induction of biofilm formation during cocultivation is linked to conjugative plasmid transfer. The transmissibility of the capacity to induce biofilm formation in coculture suggested a significant role for horizontal gene transfer in the phenomenon. We evaluated next the potential contribution of conjugative plasmid transfer to the induction phenomena observed during cocultivation. F-like conjugative plasmids comprise several incompatibility groups (IncFI to IncFV) and, together with IncI␣ plasmids, are the most abundant self-transmissible plasmids among natural isolates of E. coli (14, 18, 29). PCR amplification of genes involved in F-like (finO and traA) or IncI␣ (traJ and pilS) conjugative pilus elaboration was used to indicate the likely presence of a conjugative plasmid in the mixed biofilm. The prevalence of natural F-like and IncI␣ plasmids was assessed among the original 403 E. coli isolates as well as among the 168 MG1655Str derivatives recovered from the original coculture-induced biofilms. The parent MG1655Str is devoid of plasmid-associated genes. Interestingly, the multiplex PCR analysis of all 403 E. coli isolates revealed that the conjugative transfer genes were more frequently detected among the subset of 56 isolates exhibiting coculture-induced biofilm formation than in the remaining 347 E. coli strains (89% versus 49%; 2 test, P ⬍ 0.05). Moreover, PCR results from the 168 MG1655Str derivatives isolated from induced biofilms demonstrated that conjugative transfer genes normally encoded by F-like plasmids were detectable in 18
(46%) of the 39 MG1655Str isolates that formed stronger biofilms in pure culture than the original MG1655Str. IncI␣ plasmid-specific genes were detected in an additional four (10%) of these strains. Among the 133 MG1655Str isolates that had exhibited a biofilm-promoting effect during cocultivation, F-like and IncI␣-specific conjugative transfer genes were observed at higher frequencies of 53% and 28%, respectively. Together, plasmid-carried genes were detected in 22 (56%) of the 39 MG1655Str isolates that had acquired an enhanced capacity for biofilm formation in pure culture and in 105 (79%) strains from the subset of 133 MG1655Str isolates that gave improved biofilm formation following cocultivation with the original MG1655Str. These 22 and 105 strains originated, respectively, from 11 (20%) and 39 (70%) of the 56 cocultures selected at the start of the study (Fig. 2). In conclusion, the majority of the tested K-12 former inhabitants of cocultureinduced biofilms had acquired genes that are normally associated with conjugative plasmids. Consistent with these findings, plasmid genes detected in the MG1655Str isolates were also amplified from the E. coli isolates present in the original coculture. These data clearly indicated that gene transfer occurs in the mixed E. coli biofilm but did not distinguish whether the observed biofilm expansion should be implicated as a likely cause or consequence of conjugative DNA transfer. Repressed IncI␣ and IncFII conjugative plasmids induce biofilm formation of K-12 strains during cocultivation. If the synergistic biofilm development was due to the process of conjugative transfer, then the elaboration of plasmid-specified cell surface pili was likely to be involved in the enhancement. Conjugation machineries universally require sex pili, but from natural IncI␣ and F-like plasmids their level of expression is typically repressed (by one or more orders of magnitude compared to that of mutated laboratory derivatives). Our PCR analysis suggested that many of the natural E. coli isolates carried one or several conjugative plasmids, but these remain uncharacterized. At present, we can only speculate that these
3586
REISNER ET AL.
FIG. 3. Natural IncFII and IncI␣ plasmids repressed for pili synthesis induce biofilm formation of K-12 strains during cocultivation. E. coli MG1655Str with or without conjugative plasmids pAR183 (IncFII) and pLG272 (IncI␣) were grown in separate microtiter wells (lanes 1 to 3) or in coculture (lanes 4 and 5) with plasmid-free MG1655Str for 24 h at 37°C before biofilm formation was quantified by CV staining. Standard errors of mean values obtained from three independent experiments are shown. The significance of increase in biofilm formation due to cocultivation is indicated (P ⬍ 0.005).
are both heterogeneous and subject to diverse and variable regulation of their conjugation genes. Accordingly, we chose to next assess whether two paradigm IncFII- and IncI␣-repressed plasmids with well-characterized fertility phenotypes would stimulate biofilm formation in the coculture system. pAR183 and pLG272 are slightly modified versions of the native F-like factor R1 and the IncI␣ prototype ColIb-P9, respectively. We found that cocultivation of MG1655Str carrying plasmids repressed for pili expression with a plasmid-free MG1655Str host indeed resulted in a significant induction of biofilm formation. Using the same relative excess of plasmid-free strain as in prior experiments, biofilm formation after 24 h was about three- to fivefold improved compared to the sum of A595 values observed for the two inoculated strains in pure culture (Fig. 3). Thus, the observed stimulatory effect with known repressed plasmids was comparable to the induction found in the cocultures involving natural E. coli isolates. Presence of conjugative IncFII or IncI␣ plasmids in MG1655Str inhibits biofilm induction during cocultivation with natural E. coli isolates. Based on our findings, we proposed that the high frequencies of transfer of plasmid genes during initial cocultivation of the natural E. coli isolates with MG1655Str were not a secondary effect of cocultivation but instead were responsible for the associated biofilm induction. To investigate this hypothesis, we took advantage of a known characteristic of the conjugation process known as exclusion. The molecular mechanisms of exclusion are poorly understood, but the functional outcome of these systems is an inhibition of conjugation between cells that already possess related conjugative plasmids. We predicted that if the process of conjugative gene transfer itself was important to the enhancement of biofilm formation in coculture, then the additional presence of a related plasmid in the mixed biofilm would have a negative effect on biofilm induction. On the other hand, if the observed gene transfer occurred as a consequence of the robust biofilm
J. BACTERIOL.
FIG. 4. Inhibition of coculture-induced biofilm formation by conjugative plasmids in MG1655Str. Fifty-six E. coli isolates were grown in 96-well plates in the presence of plasmid-free MG1655Str and MG1655Str carrying conjugative plasmid pAR183 (A) or pLG272 (B) for 24 h at 37°C before biofilm formation was quantified by CV staining. To reveal the impact of the presence of the reference plasmid on biofilm formation during cocultivation, A595 values obtained for each isolate in the presence of plasmid-free MG1655Str were divided by A595 values monitored in the presence of the plasmid-carrying strain. The ratio was designated the inhibition factor. Open circles above the gray reference line represent strains that gave significantly reduced biofilm formation in the presence of the plasmid-carrying MG1655Str (P ⬍ 0.05). Filled circles above the gray reference line indicate nonsignificant inhibition. Results are categorized based on the presence in the tested E. coli isolates of genes typically found on conjugative F-like (traA and finO) (A) and IncI␣ (traJ and pilS) (B) plasmids.
formation due to other factors, then exclusion due to a related plasmid should not play a role in the ultimate expansion of the biofilm. The 56 E. coli strains that displayed strongest synergistic effects in the initial screen were cocultured with a two- to threefold excess of plasmid-free MG1655Str or MG1655Str carrying conjugative plasmid pAR183 or pLG272, respectively, and biofilm formation was evaluated after 24 h. To determine potential inhibitory effects by the presence of the additional plasmid in the mixed biofilm, A595 values obtained during coculture with plasmid-free MG1655Str were divided by the A595 values monitored after cocultivation with the plasmid-carrying MG1655Str. In the absence of an effect, this ratio would equal 1. Ratios greater than 1, indicating diminution of the biofilm in the presence of the additional plasmid, were defined as an inhibition factor. We found that the presence of the IncFII plasmid pAR183 significantly reduced coculture-induced biofilm formation in 21 (38%) of the 56 cocultures (P ⬍ 0.05). In accordance with the specificity of the exclusion principle, inhibition of biofilm induction by the IncFII plasmid was only discernible if the E. coli isolate present in the coculture carried F-like conjugation genes (Fig. 4). Presence of the IncI␣ plasmid pLG272 resulted in a comparable effect on coculture biofilms, since 24 (43%) were affected (P ⬍ 0.05). The specificity of inhibition was less pronounced than that observed with pAR183 but was significant (2 test, P ⬍ 0.05): IncI␣-related transfer genes were detected in 12 (50%) of the E. coli isolates present in the affected cocultures compared to 5 (15%) in the 32 nonaffected cocultures. In conclusion, induction in coculture could be disrupted by the additional presence of known IncI␣ or F-like
SYNERGISTIC EFFECTS IN E. COLI BIOFILMS
VOL. 188, 2006
plasmids in the K-12 strain in accordance with the specific presence of F-like or IncI␣ transfer genes in the uncharacterized strain. Failure of the affected cocultures to develop strong biofilm argues against the notion that gene transfer in this system is an indirect consequence of favorable cell densities provided by a developing biofilm and instead is involved in promoting its expansion. Synergistic effects on biofilm formation are not restricted to cocultivation with MG1655Str. In the initial screen for coculture-induced biofilm formation and all subsequent coculture experiments, we utilized the laboratory strain MG1655Str as one of the two strains present during cocultivation. Because we could not exclude the possibility that our findings were significantly biased by the partner strain background, we evaluated the dependence of the induction phenomena on the cocultured strain. We chose eight of the E. coli isolates that possessed and transferred the ability to induce biofilm formation in coculture with E. coli K-12 and cocultivated these strains in a 1:1 ratio with MG1655Str and 20 randomly chosen E. coli isolates that originally did not exhibit this phenotype. To avoid a potential disruption due to exclusion phenomena, we selected only strains that did not yield positive PCR tests for any of the conjugative transfer genes traA, finO, pilS, and traJ. We found that for 47 (29%) of the 160 cocultures, A595 values obtained following cocultivation exceeded the sum of A595 values observed for the two inoculated strains in pure culture. In 17 of the cocultures, the induction effect was even stronger than that in the corresponding coculture of the biofilm-inducing strain with MG1655Str (data not shown). These data indicate that biofilm formation was indeed induced in a substantial proportion of the rather diverse cocultures of natural E. coli isolates and imply that the results obtained for E. coli MG1655Str are relevant for natural settings where genetically diverse E. coli strains interact and compete. DISCUSSION In the majority of cases, the surface-associated multicellular communities found in a wide variety of natural and pathogenic ecosystems are formed in the presence of multiple diverse species and genetically distinct strains. In recent years, wellcontrolled in vitro biofilm model systems have revealed a diversity of molecular mechanisms contributing to development and maturation of single-species biofilms. The mechanisms underlying the biofilm development in the presence of these multispecies consortia are expected to involve even higher degrees of complexity; however, our understanding of mixedspecies biofilms is hampered by the limited number of model systems that have been applied to date. The goal of this study was to test the capacity of a simple in vitro model to reveal factors contributing to the formation of more complex biofilm communities. The suitability of this approach to high-throughput analyses was demonstrated with a systematic survey of a large collection of E. coli isolates for synergistic effects on biofilm formation caused by cocultivation with an E. coli K-12 reference strain. Utilization of this collection of E. coli strains isolated from humans was reasonable, given that the capacity of these strains to form laboratory biofilms in monoculture has been characterized in detail (27). We found that synergistic effects of cocultivation can be
3587
observed at considerable frequency. To begin analyzing the variety of mechanisms causing the stimulation, we focused on the 56 isolates (14%) that exhibited the strongest induction in coculture. The results indicate that one substantial contributor to the promotion phenomenon was lateral gene transfer. Transmission of natural conjugative plasmids was directly linked to the induction effect for the majority (70%) of this group. In these cases we found (i) that the capacity to promote biofilm formation during cocultivation was transferred from the natural E. coli isolates to the cocultured laboratory strain, (ii) that the E. coli K-12 strain acquired conjugation genes present initially in the isolate, and (iii) that the biofilm induction could be inhibited specifically by the presence of related conjugative plasmids in the K-12 strain. These results support and expand the present view that conjugative plasmids encode an important pathway for E. coli biofilm development. In a groundbreaking study, Ghigo reported that conjugative plasmids belonging to various incompatibility groups and expressing different types of conjugative pili mediated a stimulatory effect on E. coli K-12 biofilm formation. The experiments employed Pyrex slides irrigated with glucose-supplemented minimal medium (9). Using a similar medium, we subsequently showed that the reported stimulation mediated by constitutive pilus-expressing mutants of natural IncFI and IncFII plasmids on E. coli K-12 strains was discernible in a flow chamber biofilm system. Under these conditions, expression of F-like conjugative pili was sufficient to overcome the requirement for other cellular factors known to be key determinants of E. coli K-12 biofilm development (26). A plasmid-mediated effect on monoculture K-12 biofilms has additionally been detected using a polystyrene surface after static growth in diverse media (27). Nonetheless, the relevance of all these observations for natural E. coli isolates remained unclear when the latter study, employing the same 331 E. coli isolates used in the present investigation, failed to reveal an association between the presence of conjugative F-like plasmids and strong biofilm formation (27). Our current results provide the first evidence for a role of conjugative plasmids in biofilm formation of natural E. coli isolates. However, detection of enhanced biofilm formation with natural strains would appear to require the opportunity for plasmid transmission presented by cocultivation with a recipient organism. Nevertheless, the nature of the mechanisms operating in coculture remains a matter of speculation. The elaboration of conjugative pili was implicated in promotion of biofilm formation of E. coli K-12 hosts harboring a characterized IncFI plasmid with high levels of pilus expression (9). The absence of biofilm promotion in monoculture suggests that the native conjugative plasmids involved in these observations are repressed for pili expression. Remarkably, 87% of the biofilms induced by cocultivation were comprised predominantly of the K-12 strain, implying an important involvement for transconjugant cells, or their progeny, in the developing biofilm. The biofilm-promoting activity of these cell types may be connected with increased pilus expression, as suggested earlier (9). The concept of a transient period of derepression of pilus synthesis in newly formed transconjugants has been raised by studies investigating transfer kinetics of natural conjugative plasmids in laboratory systems (6, 7, 30). Further investigation of the molecular mechanisms of plasmid-driven enhancement of E. coli biofilm formation may be expected to
3588
REISNER ET AL.
provide important insights not only into the process of biofilm development but also into this intriguing aspect of conjugative plasmid biology. Are there other mechanisms involved in biofilm stimulation unrelated to horizontal gene transfer? Although transmission of conjugative F-like and IncI␣ plasmids appears to be responsible for the majority of observed synergistic strain combinations in the analysis reported here, it needs to be emphasized that not all E. coli isolates in the inducer subset yielded positive PCR results utilizing primers specific for these conjugation systems. This leaves open not only a potential influence of additional classes of conjugative plasmids but may also indicate additional unrelated mechanisms. We are currently investigating these possibilities. In the accompanying study utilizing the same set of 331 human E. coli isolates, we found that biofilm-forming capabilities of the genetically distinct E. coli isolates are dependent on growth medium composition, highlighting the importance of environmental factors on single-species biofilm formation (27). It is important to note, therefore, that the coculture-induced biofilm formation described in this study is not limited to LB medium but is also observable in minimal medium supplemented with glucose or Casamino Acids (A. Reisner, unpublished data). In conclusion, screens such as the one described here may prove useful for identification of mechanisms that stimulate biofilm formation in cultures containing genetically distinct bacterial strains. Effects revealed in simple environments may provide the basis for relevant hypotheses applicable to more complex systems. For example, large conjugative plasmids are prevalent among the Enterobacteriaceae and are obviously common in human E. coli isolates (27). It is conceivable that in clinical environments where a diversity of bacterial species and clones colonize a hydrated surface, such as catheter-associated infections, that the process of conjugative plasmid transfer accelerates the accumulation of biomass on these surfaces. The frequent association of antibiotic resistance, including the emerging problem of extended-spectrum -lactam production, and other virulence determinants with these plasmids further implies important challenges for successful therapies. While there are obvious limitations to a direct extrapolation of in vitro observations to natural E. coli biofilms, the utility of this approach is to provide initial insights into synergistic effects, the relevance of which can then be evaluated in successively complex in vitro or in vivo model systems. ACKNOWLEDGMENTS This work was generously supported by the Austrian FWF through grants P16722-B12 (to E.L.Z.) and Schro ¨ dinger grant J2250-B04 (to A.R.). T. S. Whittam, B. M. Wilkins, F. Scheutz, B. Olsen, and P. Ulleryd are gratefully acknowledged for providing plasmids or strains. REFERENCES 1. Adlerberth, I., F. Jalil, B. Carlsson, L. Mellander, L. A. Hanson, P. Larsson, K. Khalil, and A. E. Wold. 1998. High turnover rate of Escherichia coli strains in the intestinal flora of infants in Pakistan. Epidemiol. Infect. 121:587–598.
J. BACTERIOL. 2. Beloin, C., and J. M. Ghigo. 2005. Finding gene-expression patterns in bacterial biofilms. Trends Microbiol. 13:16–19. 3. Bertani, G. 1951. Studies on lysogenesis. I. The mode of phage liberation by lysogenic Escherichia coli. J. Bacteriol. 62:293–300. 4. Christensen, B. B., J. A. Haagensen, A. Heydorn, and S. Molin. 2002. Metabolic commensalism and competition in a two-species microbial consortium. Appl. Environ. Microbiol. 68:2495–2502. 5. Costerton, J. W., Z. Lewandowski, D. E. Caldwell, D. R. Korber, and H. M. Lappin-Scott. 1995. Microbial biofilms. Annu. Rev. Microbiol. 49:711–745. 6. Cullum, J., J. F. Collins, and P. Broda. 1978. The spread of plasmids in model populations of Escherichia coli K12. Plasmid 1:545–556. 7. Freter, R., R. R. Freter, and H. Brickner. 1983. Experimental and mathematical models of Escherichia coli plasmid transfer in vitro and in vivo. Infect. Immun. 39:60–84. 8. Ghigo, J. M. 2003. Are there biofilm-specific physiological pathways beyond a reasonable doubt? Res. Microbiol. 154:1–8. 9. Ghigo, J. M. 2001. Natural conjugative plasmids induce bacterial biofilm development. Nature 412:442–445. 10. Howland, C. J., C. E. Rees, P. T. Barth, and B. M. Wilkins. 1989. The ssb gene of plasmid ColIb-P9. J. Bacteriol. 171:2466–2473. 11. Kolenbrander, P. E. 2000. Oral microbial communities: biofilms, interactions, and genetic systems. Annu. Rev. Microbiol. 54:413–437. 12. Lazazzera, B. A. 2005. Lessons from DNA microarray analysis: the gene expression profile of biofilms. Curr. Opin. Microbiol. 8:222–227. 13. Luo, H., K. Wan, and H. H. Wang. 2005. High-frequency conjugation system facilitates biofilm formation and pAM1 transmission by Lactococcus lactis. Appl. Environ. Microbiol. 71:2970–2978. 14. Mainil, J. G., F. Bex, P. Dreze, A. Kaeckenbeeck, and M. Couturier. 1992. Replicon typing of virulence plasmids of enterotoxigenic Escherichia coli isolates from cattle. Infect. Immun. 60:3376–3380. 15. Miranda, R. L., T. Conway, M. P. Leatham, D. E. Chang, W. E. Norris, J. H. Allen, S. J. Stevenson, D. C. Laux, and P. S. Cohen. 2004. Glycolytic and gluconeogenic growth of Escherichia coli O157:H7 (EDL933) and E. coli K-12 (MG1655) in the mouse intestine. Infect. Immun. 72:1666–1676. 16. Molin, S., T. Tolker-Niesen, and S. K. Hansen. 2003. Microbial interactions in mixed-species biofilms, p. 206–221. In M. Ghannoum and G. A. O’Toole (ed.), Microbial biofilms. ASM Press, Washington, D.C. 17. Murphy, K. C., K. G. Campellone, and A. R. Poteete. 2000. PCR-mediated gene replacement in Escherichia coli. Gene 246:321–330. 18. Niida, M., S. Makino, N. Ishiguro, G. Sato, and T. Nishio. 1983. Genetic properties of conjugative R plasmids in Escherichia coli and Salmonella isolated from feral and domestic pigeons, crows and kites. Zentralbl. Bakteriol. Mikrobiol. Hyg. Ser. A 255:271–284. 19. Nowrouzian, F., B. Hesselmar, R. Saalman, I.-L. Strannegard, N. Aberg, A. E. Wold, and I. Adlerberth. 2003. Escherichia coli in infants’ intestinal microflora: colonization rate, strain turnover, and virulence gene carriage. Pediatr. Res. 54:8–14. 20. Ochman, H., and R. K. Selander. 1984. Standard reference strains of Escherichia coli from natural populations. J. Bacteriol. 157:690–693. 21. O’Toole, G., H. B. Kaplan, and R. Kolter. 2000. Biofilm formation as microbial development. Annu. Rev. Microbiol. 54:49–79. 22. O’Toole, G. A. 2003. To build a biofilm. J. Bacteriol. 185:2687–2689. 23. Parsek, M. R., and C. Fuqua. 2004. Biofilms 2003: emerging themes and challenges in studies of surface-associated microbial life. J. Bacteriol. 186: 4427–4440. 24. Parsek, M. R., and E. P. Greenberg. 2005. Sociomicrobiology: the connections between quorum sensing and biofilms. Trends Microbiol. 13:27–33. 25. Parsek, M. R., and P. K. Singh. 2003. Bacterial biofilms: an emerging link to disease pathogenesis. Annu. Rev. Microbiol. 57:677–701. 26. Reisner, A., J. A. Haagensen, M. A. Schembri, E. L. Zechner, and S. Molin. 2003. Development and maturation of Escherichia coli K-12 biofilms. Mol. Microbiol. 48:933–946. 27. Reisner, A., K. A. Krogfelt, B. Klein, E. L. Zechner, and S. Molin. In vitro biofilm formation of commensal and pathogenic Escherichia coli strains: impact of environmental and genetic factors. J. Bacteriol. 188:3572–3581. 28. Reisner, A., S. Molin, and E. L. Zechner. 2002. Recombinogenic engineering of conjugative plasmids with fluorescent marker cassettes. FEMS Microbiol. Ecol. 42:251–259. 29. Sherley, M., D. M. Gordon, and P. J. Collignon. 2003. Species differences in plasmid carriage in the Enterobacteriaceae. Plasmid 49:79–85. 30. Simonsen, L. 1990. Dynamics of plasmid transfer on surfaces. J. Gen. Microbiol. 136:1001–1007. 31. Van Houdt, R., and C. W. Michiels. 2005. Role of bacterial cell surface structures in Escherichia coli biofilm formation. Res. Microbiol. 156:626–633.
