Gelatinase Is Important for Translocation of Enterococcus faecalis across Polarized Human Enterocyte-Like T84 Cells
http://www.100md.com
感染与免疫杂志 2005年第3期
Division of Infectious Diseases, Department of Internal Medicine
Center for the Study of Emerging and Re-emerging Pathogens
Department of Microbiology and Molecular Genetics, University of Texas Houston Medical School, Houston, Texas
ABSTRACT
Previously, in our laboratory, we established a two-chamber system to study translocation of Enterococcus faecalis across monolayers of polarized human colon carcinoma-derived T84 cells. By using the same system in the present study, we now show that disruption of gelE of strain OG1RF, which also has a polar effect on the cotranscribed sprE, as well as disruption of its regulatory system (fsrA, fsrB, and fsrC) resulted in a loss of detectable translocation by E. faecalis OG1RF; these mutants lost gelatinase (GelE) and serine protease (SprE) production by standard assay. A gelE deletion mutant of OG1RF (GelE– SprE+) also showed that significantly reduced translocation and complementation with the gelE gene (pTEX5438) in trans restored gelatinase and translocation, demonstrating that gelatinase is important for E. faecalis translocation. Complementation of fsrA, fsrB, and fsrC mutants with all three fsr genes also resulted in production of gelatinase and translocation. Furthermore, introduction of fsr genes into two non-gelatinase-producing E. faecalis isolates, the well-characterized laboratory strain JH2-2 and a human-derived fecal isolate, TX1322 (both of which have gelE but not fsrA or fsrB, are gelatinase negative, and do not translocate), resulted in gelatinase production by these strains and restored translocation across T84 monolayers, while transformation with pTEX5438 (gelE) showed little or no translocation and no detectable gelatinase, confirming the importance of both fsr and gelatinase for E. faecalis translocation. The importance of gelatinase production was also corroborated among 20 E. faecalis human isolates (7 fecal, 7 endocarditis, and 6 urine isolates), which showed translocation by all gelatinase-positive isolates but little to no translocation for gelatinase nonproducers. These results indicate that gelatinase is important for the successful in vitro translocation of E. faecalis across human enterocyte-like T84 cells.
INTRODUCTION
Enterococci are gram-positive bacteria that inhabit the intestine of many animals including humans and have also been used as probiotic agents and as starters for cheese production. However, in the past few decades, these organisms have become some of the leading causes of nosocomial infections, probably due to the widespread use of antibiotics in hospitals and resistance of enterococci to multiple antimicrobials (11). In the genus Enterococcus, Enterococcus faecalis is the organism most commonly isolated from patients with enterococcal infections (10). Although the routes of enterococcal infections are not yet well understood, Wells et al. and Runkel et al. have previously presented evidence that E. faecalis can translocate across mouse and rat intestinal tracts and reach other sites (18, 29). More recently, Krueger et al. reported that by orally feeding mice with antibiotics and E. faecalis, these organisms were found in livers, spleens, and mesenteric lymph nodes (7). Those researchers also assessed aggregation substance and binding substance in the mouse model of extraintestinal translocation and found that these factors did not appear to be required for this process (7), although others have found aggregation substance important for internalization by intestinal epithelial cell lines HT29, T84, and HCT-8 (19, 28). A similar observation was made by Vinderola and colleagues, who found that when mice were orally fed with 104 CFU of E. faecalis per day for 7 consecutive days, translocation of this organism to the liver was observed at days 2, 5, and 7 (26).
In order to study the translocation of E. faecalis and factors involved in this process, we previously established an in vitro model to mimic this process by using human colon carcinoma-derived T84 cells (31). This model involves a two-chamber system with a permeable support separating the two chambers. The T84 cells are grown on the permeable support to form an epithelial monolayer (which differentiates and shows structural resemblance to the native intestine), E. faecalis cells are added to the upper chamber, and the translocated bacterial cells are recovered from the lower chamber (31). In that study, we found that the commonly used E. faecalis strain OG1RF, unlike Escherichia coli DH5, was able to translocate across T84 monolayers, and by using this in vitro model, we were able to compare translocation of OG1RF to that of mutants and found that the epa gene cluster, previously shown to be important for E. faecalis virulence in a mouse peritonitis model (30), is important for this process (31).
In our initial study, we examined 14 E. faecalis human isolates and found considerable differences in their abilities to translocate among these isolates (31). We subsequently examined our results from a survey of 215 E. faecalis isolates (17) and noted that although the assay used was rather crude and the growth conditions were different from those used for translocation, the results suggested a correlation between gelatinase activity and translocation capability of E. faecalis. That is, 6 out of 6 gelatinase-positive isolates translocated across T84 monolayers in more than 63% of the transwells, 4 out of 7 gelatinase-negative isolates did not show translocation in any of the transwells, 3 out of 7 gelatinase-negative isolates showed translocation in less than 25% of the transwells, and 1 isolate showed weak gelatinase activity and no translocation (unpublished observations). In the present study, we applied the two-chamber T84 model and established the importance of gelatinase activity for translocation capability of E. faecalis.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. Bacterial isolates and plasmids used in this study are listed in Table 1 and include 13 clinical isolates (7 endocarditis and 6 urine isolates), 7 human fecal isolates, OG1RF (23), and JH2-2 (5); other than OG1RF and JH2-2, these strains are different from those studied previously (31). E. faecalis isolates were grown at 37°C in brain heart infusion broth or agar (Difco Laboratories, Sparks, Md.). The concentrations of antibiotics used for selection of recombinant strains were 300 μg of erythromycin/ml for E. coli and 2,000 μg of kanamycin/ml and 10 μg of erythromycin/ml for E. faecalis.
T84 cell translocation. Growth and maintenance of T84 cells, the preparation of bacteria for translocation, and translocation experiments were performed according to methods described previously (31). For clinical isolates, because of some variation in translocation by OG1RF seen with different passages of T84 cells used over the span of the experiments, the results for strains showing translocation are presented as the percentage of CFU relative to results obtained with OG1RF determined concurrently. For the strains which had shown no translocation at all in pilot studies, we plated the whole 1-ml volume of the bottom chamber at 8 h, and for other strains, 100 μl out of 1 ml was plated.
Assays for gelatinase activity. For screening of colonies, the production of gelatinase in E. faecalis strains was performed by using Todd-Hewitt (Difco Laboratories) agar plates containing 3% gelatin (gelatin-TH) (15). After overnight incubation at 37°C, colonies with opaque zones around them were considered to be gelatinase activity positive.
For assaying gelatinase activity in culture supernatants, a previously described procedure was performed with a slight modification (12). Azocoll (Azo dye-impregnated collagen, 0.25 g, <50 mesh; Calbiochem, Darmstadt, Germany) was washed in 50 ml of 50 mM Tris-HCl buffer (pH 7.8) containing 1 mM CaCl2, kept standing for 90 min at 37°C, and then centrifuged at 1,500 x g for 10 min, discarding the buffer. The residue was resuspended in 50 ml of the same buffer, and 0.5-ml aliquots were transferred into a 1.5-ml Eppendorf tube. Tubes were incubated at 37°C for 15 min on a shaker, and then 25 μl of culture supernatant (see below) was added to each tube containing preincubated Azocoll. The mixture was incubated for 4 h at 37°C on a shaker and then centrifuged at 1,500 x g for 5 min, followed by measurement of absorbance at 540 nm.
The preparation of bacteria for the gelatinase activity assays was the same as the preparation of bacteria for translocation. Approximately 8 x 107 bacteria in 0.5 ml of tissue culture medium (31) were added to confluent T84 cells that had been grown for 4 to 5 days in 12-well cell culture plates (Corning Incorporated, Corning, N.Y.) and were then incubated at 37°C, 5% CO2. After 6 h of incubation, about 0.5 ml was removed to a 1.5-ml Eppendorf tube and then centrifuged at 1,500 x g for 5 min, and the supernatant was used for gelatinase activity assay; some samples were incubated for up to 24 h. Bacteria were also grown in tissue culture medium (31) alone and processed similarly.
Cloning of gelE and complementation of TX5264. A 1,767-bp DNA fragment containing the gelE gene and its promoter region was amplified by using the primers 5'-GGCGAATTCGCTATGGTATTG (forward primer) and 5'-GCGGGATCCTCATTCATTGACCAGAACA (reverse primer) and cloned into the shuttle vector pAT18 (Eryr) (25) to produce plasmid pTEX5438. The plasmid was then transformed into the gelE deletion mutant TX5264, and Eryr transformants were plated onto the gelatin-TH plates to determine the effect of complementation on the production of gelatinase.
Introduction of gelE and fsr genes into JH2-2 and TX1322. The plasmids pTEX5438, which contains gelE, and pTEX5249, which contains fsrA, fsrB, and fsrC genes but only the first 395 bp of gelE, were transformed into TX1322 and JH2-2, respectively, and Eryr transformants were plated onto gelatin-TH plates to verify gelatinase activity.
Statistical analysis. Analysis of variance (ANOVA) with Bonferroni's posttest was performed using GraphPad (San Diego, Calif.) Prism software version 4.00 for comparing multiple continuous variables, and Fisher's exact test was used for categorical results comparing GelE+ and GelE– human isolates.
RESULTS
Effect of gelE and fsr genes on translocation of E. faecalis across T84 monolayers. The fsr-gelE/sprE locus is shown in Fig. 1A. Six mutants of this locus generated in OG1RF, TX5240 (fsrA disruption mutant), TX5241 (fsrB disruption mutant), TX5242 (fsrC disruption mutant), TX5266 (fsrB deletion mutant), TX5128 (gelE disruption mutant), and TX5264 (a gelE deletion mutant), all previously shown to be gelatinase negative in the gelatin-TH plate assay, were compared with wild-type OG1RF in the transcytosis assay. At least two individual experiments were performed (4 wells each) with all strains, and consistent results were obtained. In the experiments shown in Fig. 1B, OG1RF cells were not detected in the lower chamber at 0 h, were detected in the lower chamber of 11 of 12 wells at 6 h, and were detected in all 12 wells at 8 h; this degree of translocation was consistently seen throughout all experiments with OG1RF in the present work and previously (31). The fsr disruption mutants (TX5240, TX5241, and TX5242), the fsrB deletion mutant (TX5266), and the gelE disruption mutant (TX5128) were not recovered from the lower chamber of any of the 12 wells at 6 or 8 h (Fig. 1B), while the gelE deletion mutant TX5264 (GelE– SprE+), which produces a "superactive" form of SprE (6), was detected in the lower chambers of 3 out of 12 wells at 6 h and 8 out of 12 wells at 8 h, although the number of bacteria translocated across T84 monolayers was significantly lower than that for wild-type OG1RF (P < 0.001) (Fig. 1B).
Effect of complementation on translocation of E. faecalis across T84 monolayers. To further explore the importance of GelE for translocation, the gelE deletion mutant TX5264 was complemented with the whole-length gelE gene, resulting in TX5439. In concurrently performed experiments, TX5264 again showed significantly less ability to translocate across T84 monolayers than wild-type OG1RF at 6 and 8 h (Fig. 1B), while complemented cells (TX5439), which were gelatinase positive, were detected in the lower chambers of all 8 wells at 6 and 8 h (Fig. 1B).
We also tested our disruption mutants after complementation by pTEX5249, which contains the fsr genes (Fig. 1A). Each of the complemented fsr mutants (TX5244, TX5245, and TX5246) which tested gelatinase positive was able to translocate across T84 monolayers. TX5244 cells were detected in the lower chambers of 6 of 8 wells at 6 h and all 8 wells at 8 h, and TX5245 and TX5246 cells were detected in the lower chambers of 8 wells at 6 and 8 h, although the CFU of translocated TX5244 were fewer than those of wild-type OG1RF (P < 0.001) (Fig. 1B). As expected, the gelE disruption mutant TX5128 was still gelatinase negative by plate assay after introduction of the fsr locus and did not show translocation across T84 monolayers (data not shown).
Gelatinase activity of gelE/fsr mutants and their complemented derivatives. Since the growth and detection of gelatinase activity on gelatin-TH plates may not reflect what is occurring in the translocation experiments, we also tested gelatinase activity of E. faecalis grown in tissue culture medium with T84 cells and with a more quantitative method described previously by Nakayama et al. (12) (Fig. 1B). None of the above-described disruption mutants showed detectable gelatinase activity, consistent with the results from plate assay. The complemented fsrA mutant (TX5244) showed less gelatinase activity than OG1RF (optical density at 540 nm [OD540] of 1.1 ± 0.8 for TX5244 compared to 2.3 ± 0.2 for OG1RF), while complemented fsrB and fsrC mutants (TX5245 and TX5246) showed a level of gelatinase activity similar to that of OG1RF (Fig. 1B). The gelE deletion mutant TX5264 (GelE– SprE+) showed a much lower level of proteolytic activity than OG1RF (an OD540 of 0.1 ± 0.1 for OG1RF compared to an OD540 of 2.3 ± 0.2 for TX5264); after complementation of TX5264 by gelE (pTEX5438), production of gelatinase was fully restored (Fig. 1B). For all these OG1RF derivatives, gelatinase activity correlated with translocation capability across T84 monolayers (Fig. 2).
Effect of providing gelE or fsr genes on translocation of JH2-2 and TX1322 across T84 monolayers. We next transformed pTEX5249 containing fsr genes into two gelatinase phenotype-negative and gelE genotype-positive strains, JH2-2 and TX1322, neither of which showed translocation across a T84 monolayer at 6 or 8 h (Fig. 3). The resulting strains, TX5443 and TX5441, were gelatinase positive, but the gelatinase produced as measured by OD was less than that of OG1RF (an OD540 of 1.7 ± 0.2 and 1.9 ± 0.4 for TX5443 and TX5441 compared to an OD540 of 2.5 ± 0.2 for OG1RF) (Fig. 3). During translocation, no TX5443 cells were detected in lower chambers at 6 h, but translocation was observed with 6 of 8 wells at 8 h, although the CFU in the lower chamber were only 101 to 102, which is significantly lower than that of OG1RF (P < 0.001) (Fig. 3). For JH2-2 complemented with the fsr genes (TX5441), bacteria were detected in the lower chambers of 8 of 12 wells at 6 h, with all 12 wells showing translocation at 8 h; the CFU in the lower chamber reached 102 to 104, which is also significantly lower than that for OG1RF (P < 0.001) (Fig. 3). We also transformed pTEX5438 (gelE gene) into TX1322 and JH2-2 to generate TX5442 and TX5440. For TX5442, no bacteria cells were detected in the lower chamber at 6 or 8 h (Fig. 3); for TX5440, only 1 well of 12 showed translocation at 6 h and 3 of 12 wells showed translocation at 8 h (Fig. 3). No gelatinase activity was detected with TX5440 and TX5442 at 6 h incubation (Fig. 3), consistent with the absence of fsr. All these results further confirmed that E. faecalis transcytosis is enhanced by gelatinase production and activity.
Survey of translocation of clinical and fecal isolates of E. faecalis across T84 monolayers. Twenty E. faecalis human isolates were tested in the two-chamber transcytosis system (Fig. 4). All of these strains grew well in T84 medium (ca. 109 CFU in the upper chamber at the end of the experiments). The five fecal strains that were gelatinase positive showed translocation, while the two gelatinase-negative fecal strains did not show translocation. Among the endocarditis strains, the four gelatinase-positive strains showed translocation, and the three gelatinase-negative strains did not show translocation; the translocation of gelatinase-positive strain TX0040 cannot be seen in the figure due to the low percentage compared to that of OG1RF (0.2% at 6 h and 0.8% at 8 h). Among urine strains, the three gelatinase-positive strains showed translocation, two gelatinase-negative strains did not show translocation, and one gelatinase-negative strain showed translocation in a single well of 8 wells. The strains (TX0018, TX0040, and TX0614) which showed less translocation than OG1RF also produced less gelatinase than OG1RF. However, TX1317 and TX0231, which showed 17 and 22 times more translocation than OG1RF at 6 h, did not show significantly higher gelatinase activity than OG1RF. TX1341 showed low gelatinase activity, but the strain performed almost the same as OG1RF in the transcytosis assay. It was also observed that one of the human isolates, TX0231, which showed a high level of translocation and produced a moderate level of gelatinase activity under translocation conditions, did not show detectable gelatinase activity in the absence of T84 cells (an OD540 of 1.8 ± 0.7 for growth in the presence of T84 cells compared to an OD540 of 0.0 ± 0.0 for growth in tissue culture medium alone), while all the other strains showed a similar level of gelatinase activity with or without T84 cells.
DISCUSSION
An extracellular proteolytic enzyme of E. faecalis was reported 40 to 50 years ago and was later shown to be a Zn-metalloprotease (1, 2, 4, 21); this enzyme also appeared to be present in some of the first known clinical isolates of enterococci, which were classified in the early 1900s as "Streptococcus faecalis var. zymogenes," referring to their production of proteolytic and hemolytic activity (20). This "gelatinase," purified by Makinen et al. in 1989, was shown to be able to degrade a number of substrates including Azocoll, gelatin, insulin B-chain, human endothelin, collagen, E. faecalis sex pheromone-related peptides, and several mammalian bioactive peptides (8, 9). More recently, gelatinase was also shown to cleave misfolded surface proteins of E. faecalis as well as fibrin, to reduce E. faecalis supernatant pheromone levels, and to affect E. faecalis chain length, possibly by activation of an autolysin (27). Gelatinase has been a factor of interest to our laboratory and has been shown by our group and others to be important for virulence in a mouse peritonitis model (23), the nematode worm Caenorhabditis elegans (22), and a rabbit endophthalmitis model (3). Downstream and cotranscribed with gelE (24) is sprE, encoding a serine protease, and upstream are fsrA, fsrB, and fsrC, which are homologs of agrA, agrB, and agrC of Staphylococcus aureus (14, 15). The fsr genes form a quorum-sensing system which appears to be autoregulated and to positively regulate the expression of gelatinase and serine protease (14, 15).
To study the possible contribution of gelatinase to E. faecalis translocation across human intestine, we first examined the previously constructed gelE mutants (22, 23) in our in vitro translocation model. We prolonged the translocation time from 6 to 8 h, since results from the latter time point showed less variation among the Transwells. The resistance across the membrane remained high at 8 h (data not shown), indicating that the T84 monolayer was still intact. We also used a more quantitative method described previously by Nakayama et al. (12), which measures proteolytic cleavage of Azocoll, and assayed bacteria grown in the same medium and conditions as those used for translocation, and with T84 cells. Our results showed that the gelE disruption mutant TX5128 (GelE– SprE+) was consistently unable to translocate across T84 monolayers, while the gelE deletion mutant TX5264 (SprE+) showed significantly decreased translocation compared to that of wild-type OG1RF; complementation of the gelE deletion mutant in trans with whole-length gelE and its upstream promoter region restored translocation. These results indicate that gelE is important for E. faecalis translocation. Subsequent analysis of gelatinase activity, as measured by proteolysis of Azocoll, indicated a clear correlation between gelatinase activity of bacteria grown in tissue culture medium plus T84 cells and translocation: the gelE disruption mutant showed no detectable activity against Azocoll, the gelE deletion mutant showed a minimal amount of activity (OD540 of 0.1 ± 0.1 compared to 2.3 ± 0.3 for OG1RF), and complementation fully restored activity (OD540 of 2.7 ± 0.1). The residual proteolytic activity shown by the gelE deletion mutant is likely due to production of a superactive serine protease in this mutant (6), the activity of which is different from that of SprE in wild-type OG1RF due to different processing (6).
Since the expression of gelE is positively regulated by fsr genes, we also examined our previously constructed fsr mutants and complementation derivatives (14, 15). None of the three fsr disruption mutants or the fsrB deletion mutant showed translocation or detectable gelatinase activity. Complementation of fsrB and fsrC disruption mutants with the fsr-containing plasmid pTEX5249 resulted in restoration of their gelatinase activity and translocation, while complementation of the fsrA disruption mutant partially restored its translocation, which may be due to its lower gelatinase activity (OD540 about half of that of OG1RF). Overall, the results with different OG1RF derivatives not only indicate that gelatinase is critical for E. faecalis translocation across T84 monolayers but also demonstrate a correlation between the Azocoll-related proteolytic activities and translocation capabilities among these strains (Fig. 2). Although the reason for the different levels of gelatinase activity in the different complementation derivatives is not clear, previous analysis of the expression of fsr genes revealed that fsrB and fsrC are cotranscribed from a promoter upstream of fsrB while a separate promoter is used for fsrA, consistent with fsrB and fsrC behaving the same in complementation. Why providing the fsr genes in multicopy does not fully restore the defect of the fsrA mutant is not known.
We also studied two gelatinase nonproducers of E. faecalis: TX1322 and JH2-2. The two strains were previously shown to have the gelE gene but to lack fsrA and fsrB, and the latter strain was previously found to be unable to translocate T84 monolayers (31). A 23.9-kb deletion involving fsrA and fsrB was first described by Nakayama et al. in E. faecalis urine isolates obtained from a Japanese hospital (13), and this region was later found by our group, using colony hybridization and PCR, to be absent in 60 E. faecalis isolates (including TX1332 and JH2-2) from different clinical and geographical sources, which account for ca. 28% of the E. faecalis isolates tested by us (17). Neither TX1322 nor JH2-2 showed detectable gelatinase activity or translocation. Introduction of fsr into these strains resulted in successful translocation, which can be explained by activation of gelatinase production by providing the Fsr function. When compared with OG1RF, the two fsr-complemented isolates showed lower levels of translocation, as well as lower levels of gelatinase activity, than OG1RF. Introduction of gelE without fsr into the two gelatinase nonproducers had no effect on their gelatinase activity under the conditions tested, consistent with gelatinase production and activity being tightly controlled by the Fsr system in these isolates. However, JH2-2 provided with pTEX5438 showed a low degree of translocation; the difference in translocation of TX1322 (pTEX5438) compared to that of JH2-2 (pTEX5438) may suggest a level of gelatinase activity not detectable in our assay or that some other factor(s) (present in JH2-2 but not in TX1322) is assisting in the process.
To determine if our findings could be extended to other E. faecalis strains, 20 human isolates (which are different from the ones tested previously) were tested for gelatinase activity and in the translocation model, and the results showed a positive correlation between the two phenotypes: 12 of 12 gelatinase-positive isolates were able to translocate, while the gelatinase-negative isolates were either unable to translocate at all (7 of 8) or translocated at a very low level (1 of 8). When combined with the translocation results obtained from the previous study (31) and our subsequent determination of their gelatinase activity (17), the capability of translocation of the two groups (GelE+ versus GelE–) is significantly different, with translocation occurring in a total of 167 out of 185 transwells at 6 h for GelE+ strains compared to a total of 7 out of 128 transwells for GelE– strains (P < 0.001 by Fisher's exact test). On the other hand, the level of translocation by gelatinase-positive human isolates varied considerably, which cannot be completely explained by the level of gelatinase activity, suggesting that although gelatinase is critical for E. faecalis translocation, other factor(s) are also important in this process. We also noticed, consistent with our previous study (31), that fecal isolates showed greater translocation than isolates from other sources, although we did not find an obvious difference in the percentage of GelE+ or in the amount of gelatinase produced by isolates from different sources in the present study, nor did we find obvious differences in our previous work (17). This observation may suggest that the fecal isolates (which reside in the human intestine) are enriched in attributes which, together with gelatinase, facilitate their translocation across human intestine, and after leaving the intestine and entering other sites of the human body, these attributes are not needed and are thus lost or repressed. However, the number of isolates tested is too limited to do more than pose this hypothesis. Interestingly, in the present study, we found that one of the human isolates, TX0231, showed a high level of translocation but, when grown in tissue culture medium, produced gelatinase activity only in the presence of T84 cells, indicating induction of gelatinase activity by T84 cells in this isolate. This observation raises the possibility that gelatinase activity of E. faecalis can be activated by interaction with certain in vivo sites or condition(s).
The mechanism by which gelatinase enhances translocation is unknown. Because of the prominent chaining of gelE mutants (27), we assayed a previously constructed autolysin mutant (16) as well as mutants in a putative regulatory gene, all showing long cell chains similar to those of the gelE mutants, and found that these mutants had a translocation capability similar to that of OG1RF (data not shown), indicating that the chaining effect per se does not cause the decrease in translocation. Whether gelatinase augments translocation from the intestinal tract in vivo, in the milieu of fecal contents, is also not known. Since we showed that two distinct strains (JH2-2 and TX1332) that are GelE– but gelE+ are still capable of producing gelatinase under certain conditions (e.g., providing fsr genes), it is also possible that some in vivo conditions may induce gelatinase production, enabling fsr-lacking, gelE-containing strains, which comprise about 28% of E. faecalis isolates (17), to translocate.
In conclusion, the present study showed that the E. faecalis gelatinase, which is secreted outside of E. faecalis cells and can potentially interact with E. faecalis cells as well as host cells and tissues, is critical for translocation of E. faecalis across T84 monolayers in vitro. Future studies that address functions that may be assigned to this protein as well as those looking at alternative regulation of gelE may help our understanding of the role(s) of this protease in E. faecalis infections.
ACKNOWLEDGMENTS
The present work was supported by grant NIH R37 AI47923 from the Division of Microbiology and Infectious Diseases, NIAID, to B.E.M.
REFERENCES
1. Bleiweis, A. S., and L. N. Zimmerman. 1964. Properties of proteinase from Streptococcus faecalis var. liquefaciens. J. Bacteriol. 88:653-659.
2. Cano, F. R., I. A. Casas, and L. N. Zimmerman. 1971. Purification and characterization of a protease produced by Streptococcus faecalis var. liquefaciens. Prep. Biochem. 1:269-282.
3. Engelbert, M., E. Mylonakis, F. M. Ausubel, S. B. Calderwood, and M. S. Gilmore. 2004. Contribution of gelatinase, serine protease, and fsr to the pathogenesis of Enterococcus faecalis endophthalmitis. Infect. Immun. 72:3628-3633.
4. Grutter, F. H., and L. N. Zimmerman. 1955. A proteolytic enzyme of Streptococcus zymogenes. J. Bacteriol. 69:728-732.
5. Jacob, A. E., and S. J. Hobbs. 1974. Conjugal transfer of plasmid-borne multiple antibiotic resistance in Streptococcus faecalis var. zymogenes. J. Bacteriol. 117:360-372.
6. Kawalec, M., J. Potempa, J. L. Moon, J. Travis, and B. E. Murray. 2005. Molecular diversity of a putative virulence factor: purification and characterization of isoforms of an extracellular serine glutamylendopeptidase of Enterococcus faecalis with different enzymatic activities. J. Bacteriol. 187:266-275.
7. Krueger, W. A., S. Krueger-Rameck, S. Koch, V. Carey, G. B. Pier, and J. Huebner. 2004. Assessment of the role of antibiotics and enterococcal virulence factors in a mouse model of extraintestinal translocation. Crit. Care Med. 32:467-471.
8. Makinen, P. L., D. B. Clewell, F. An, and K. K. Makinen. 1989. Purification and substrate specificity of a strongly hydrophobic extracellular metalloendopeptidase ("gelatinase") from Streptococcus faecalis (strain 0G1-10). J. Biol. Chem. 264:3325-3334.
9. Makinen, P. L., and K. K. Makinen. 1994. The Enterococcus faecalis extracellular metalloendopeptidase (EC 3.4.24.30; coccolysin) inactivates human endothelin at bonds involving hydrophobic amino acid residues. Biochem. Biophys. Res. Commun. 200:981-985.
10. Murray, B. E. 1990. The life and times of the enterococcus. Clin. Microbiol. Rev. 3:46-65.
11. Murray, B. E., and G. M. Weinstock. 1999. Enterococci: new aspects of an old organism. Proc. Assoc. Am. Physicians 111:328-334.
12. Nakayama, J., Y. Cao, T. Horii, S. Sakuda, A. D. Akkermans, W. M. de Vos, and H. Nagasawa. 2001. Gelatinase biosynthesis-activating pheromone: a peptide lactone that mediates a quorum sensing in Enterococcus faecalis. Mol. Microbiol. 41:145-154.
13. Nakayama, J., R. Kariyama, and H. Kumon. 2002. Description of a 23.9-kilobase chromosomal deletion containing a region encoding fsr genes which mainly determines the gelatinase-negative phenotype of clinical isolates of Enterococcus faecalis in urine. Appl. Environ. Microbiol. 68:3152-3155.
14. Qin, X., K. V. Singh, G. M. Weinstock, and B. E. Murray. 2001. Characterization of fsr, a regulator controlling expression of gelatinase and serine protease in Enterococcus faecalis OG1RF. J. Bacteriol. 183:3372-3382.
15. Qin, X., K. V. Singh, G. M. Weinstock, and B. E. Murray. 2000. Effects of Enterococcus faecalis fsr genes on production of gelatinase and a serine protease and virulence. Infect. Immun. 68:2579-2586.
16. Qin, X., K. V. Singh, Y. Xu, G. M. Weinstock, and B. E. Murray. 1998. Effect of disruption of a gene encoding an autolysin of Enterococcus faecalis OG1RF. Antimicrob. Agents Chemother. 42:2883-2888.
17. Roberts, J. C., K. V. Singh, P. C. Okhuysen, and B. E. Murray. 2004. Molecular epidemiology of the fsr locus and of gelatinase production among different subsets of Enterococcus faecalis isolates. J. Clin. Microbiol. 42:2317-2320.
18. Runkel, N. S., F. G. Moody, G. S. Smith, L. F. Rodriguez, Y. Chen, M. T. Larocco, and T. A. Miller. 1993. Alterations in rat intestinal transit by morphine promote bacterial translocation. Dig. Dis. Sci. 38:1530-1536.
19. Sartingen, S., E. Rozdzinski, A. Muscholl-Silberhorn, and R. Marre. 2000. Aggregation substance increases adherence and internalization, but not translocation, of Enterococcus faecalis through different intestinal epithelial cells in vitro. Infect. Immun. 68:6044-6047.
20. Sherman, J. M. 1937. The streptococci. Bacteriol. Rev. 1:3-97.
21. Shugart, L. R., and R. W. Beck. 1964. Purification and activity of proteinase of Streptococcus faecalis var. liquefaciens. J. Bacteriol. 88:586-590.
22. Sifri, C. D., E. Mylonakis, K. V. Singh, X. Qin, D. A. Garsin, B. E. Murray, F. M. Ausubel, and S. B. Calderwood. 2002. Virulence effect of Enterococcus faecalis protease genes and the quorum-sensing locus fsr in Caenorhabditis elegans and mice. Infect. Immun. 70:5647-5650.
23. Singh, K. V., X. Qin, G. M. Weinstock, and B. E. Murray. 1998. Generation and testing of mutants of Enterococcus faecalis in a mouse peritonitis model. J. Infect. Dis. 178:1416-1420.
24. Su, Y. A., M. C. Sulavik, P. He, K. K. Makinen, P.-L. Makinen, S. Fiedler, R. Wirth, and D. B. Clewell. 1991. Nucleotide sequence of the gelatinase gene (gelE) from Enterococcus faecalis subsp. liquefaciens. Infect. Immun. 59:415-420.
25. Trieu-Cuot, P., C. Carlier, C. Poyart-Salmeron, and P. Courvalin. 1991. Shuttle vectors containing a multiple cloning site and a lacZ gene for conjugal transfer of DNA from Escherichia coli to gram-positive bacteria. Gene 102:99-104.
26. Vinderola, C. G., M. Medici, and G. Perdigon. 2004. Relationship between interaction sites in the gut, hydrophobicity, mucosal immunomodulating capacities and cell wall protein profiles in indigenous and exogenous bacteria. J. Appl. Microbiol. 96:230-243.
27. Waters, C. M., M. H. Antiporta, B. E. Murray, and G. M. Dunny. 2003. Role of the Enterococcus faecalis GelE protease in determination of cellular chain length, supernatant pheromone levels, and degradation of fibrin and misfolded surface proteins. J. Bacteriol. 185:3613-3623.
28. Waters, C. M., C. L. Wells, and G. M. Dunny. 2003. The aggregation domain of aggregation substance, not the RGD motifs, is critical for efficient internalization by HT-29 enterocytes. Infect. Immun. 71:5682-5689.
29. Wells, C. L., R. P. Jechorek, and S. L. Erlandsen. 1990. Evidence for the translocation of Enterococcus faecalis across the mouse intestinal tract. J. Infect. Dis. 162:82-90.
30. Xu, Y., K. V. Singh, X. Qin, B. E. Murray, and G. M. Weinstock. 2000. Analysis of a gene cluster of Enterococcus faecalis involved in polysaccharide biosynthesis. Infect. Immun. 68:815-823.
31. Zeng, J., F. Teng, G. M. Weinstock, and B. E. Murray. 2004. Translocation of Enterococcus faecalis strains across a monolayer of polarized human enterocyte-like T84 cells. J. Clin. Microbiol. 42:1149-1154.(Jing Zeng, Fang Teng, and)
Center for the Study of Emerging and Re-emerging Pathogens
Department of Microbiology and Molecular Genetics, University of Texas Houston Medical School, Houston, Texas
ABSTRACT
Previously, in our laboratory, we established a two-chamber system to study translocation of Enterococcus faecalis across monolayers of polarized human colon carcinoma-derived T84 cells. By using the same system in the present study, we now show that disruption of gelE of strain OG1RF, which also has a polar effect on the cotranscribed sprE, as well as disruption of its regulatory system (fsrA, fsrB, and fsrC) resulted in a loss of detectable translocation by E. faecalis OG1RF; these mutants lost gelatinase (GelE) and serine protease (SprE) production by standard assay. A gelE deletion mutant of OG1RF (GelE– SprE+) also showed that significantly reduced translocation and complementation with the gelE gene (pTEX5438) in trans restored gelatinase and translocation, demonstrating that gelatinase is important for E. faecalis translocation. Complementation of fsrA, fsrB, and fsrC mutants with all three fsr genes also resulted in production of gelatinase and translocation. Furthermore, introduction of fsr genes into two non-gelatinase-producing E. faecalis isolates, the well-characterized laboratory strain JH2-2 and a human-derived fecal isolate, TX1322 (both of which have gelE but not fsrA or fsrB, are gelatinase negative, and do not translocate), resulted in gelatinase production by these strains and restored translocation across T84 monolayers, while transformation with pTEX5438 (gelE) showed little or no translocation and no detectable gelatinase, confirming the importance of both fsr and gelatinase for E. faecalis translocation. The importance of gelatinase production was also corroborated among 20 E. faecalis human isolates (7 fecal, 7 endocarditis, and 6 urine isolates), which showed translocation by all gelatinase-positive isolates but little to no translocation for gelatinase nonproducers. These results indicate that gelatinase is important for the successful in vitro translocation of E. faecalis across human enterocyte-like T84 cells.
INTRODUCTION
Enterococci are gram-positive bacteria that inhabit the intestine of many animals including humans and have also been used as probiotic agents and as starters for cheese production. However, in the past few decades, these organisms have become some of the leading causes of nosocomial infections, probably due to the widespread use of antibiotics in hospitals and resistance of enterococci to multiple antimicrobials (11). In the genus Enterococcus, Enterococcus faecalis is the organism most commonly isolated from patients with enterococcal infections (10). Although the routes of enterococcal infections are not yet well understood, Wells et al. and Runkel et al. have previously presented evidence that E. faecalis can translocate across mouse and rat intestinal tracts and reach other sites (18, 29). More recently, Krueger et al. reported that by orally feeding mice with antibiotics and E. faecalis, these organisms were found in livers, spleens, and mesenteric lymph nodes (7). Those researchers also assessed aggregation substance and binding substance in the mouse model of extraintestinal translocation and found that these factors did not appear to be required for this process (7), although others have found aggregation substance important for internalization by intestinal epithelial cell lines HT29, T84, and HCT-8 (19, 28). A similar observation was made by Vinderola and colleagues, who found that when mice were orally fed with 104 CFU of E. faecalis per day for 7 consecutive days, translocation of this organism to the liver was observed at days 2, 5, and 7 (26).
In order to study the translocation of E. faecalis and factors involved in this process, we previously established an in vitro model to mimic this process by using human colon carcinoma-derived T84 cells (31). This model involves a two-chamber system with a permeable support separating the two chambers. The T84 cells are grown on the permeable support to form an epithelial monolayer (which differentiates and shows structural resemblance to the native intestine), E. faecalis cells are added to the upper chamber, and the translocated bacterial cells are recovered from the lower chamber (31). In that study, we found that the commonly used E. faecalis strain OG1RF, unlike Escherichia coli DH5, was able to translocate across T84 monolayers, and by using this in vitro model, we were able to compare translocation of OG1RF to that of mutants and found that the epa gene cluster, previously shown to be important for E. faecalis virulence in a mouse peritonitis model (30), is important for this process (31).
In our initial study, we examined 14 E. faecalis human isolates and found considerable differences in their abilities to translocate among these isolates (31). We subsequently examined our results from a survey of 215 E. faecalis isolates (17) and noted that although the assay used was rather crude and the growth conditions were different from those used for translocation, the results suggested a correlation between gelatinase activity and translocation capability of E. faecalis. That is, 6 out of 6 gelatinase-positive isolates translocated across T84 monolayers in more than 63% of the transwells, 4 out of 7 gelatinase-negative isolates did not show translocation in any of the transwells, 3 out of 7 gelatinase-negative isolates showed translocation in less than 25% of the transwells, and 1 isolate showed weak gelatinase activity and no translocation (unpublished observations). In the present study, we applied the two-chamber T84 model and established the importance of gelatinase activity for translocation capability of E. faecalis.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. Bacterial isolates and plasmids used in this study are listed in Table 1 and include 13 clinical isolates (7 endocarditis and 6 urine isolates), 7 human fecal isolates, OG1RF (23), and JH2-2 (5); other than OG1RF and JH2-2, these strains are different from those studied previously (31). E. faecalis isolates were grown at 37°C in brain heart infusion broth or agar (Difco Laboratories, Sparks, Md.). The concentrations of antibiotics used for selection of recombinant strains were 300 μg of erythromycin/ml for E. coli and 2,000 μg of kanamycin/ml and 10 μg of erythromycin/ml for E. faecalis.
T84 cell translocation. Growth and maintenance of T84 cells, the preparation of bacteria for translocation, and translocation experiments were performed according to methods described previously (31). For clinical isolates, because of some variation in translocation by OG1RF seen with different passages of T84 cells used over the span of the experiments, the results for strains showing translocation are presented as the percentage of CFU relative to results obtained with OG1RF determined concurrently. For the strains which had shown no translocation at all in pilot studies, we plated the whole 1-ml volume of the bottom chamber at 8 h, and for other strains, 100 μl out of 1 ml was plated.
Assays for gelatinase activity. For screening of colonies, the production of gelatinase in E. faecalis strains was performed by using Todd-Hewitt (Difco Laboratories) agar plates containing 3% gelatin (gelatin-TH) (15). After overnight incubation at 37°C, colonies with opaque zones around them were considered to be gelatinase activity positive.
For assaying gelatinase activity in culture supernatants, a previously described procedure was performed with a slight modification (12). Azocoll (Azo dye-impregnated collagen, 0.25 g, <50 mesh; Calbiochem, Darmstadt, Germany) was washed in 50 ml of 50 mM Tris-HCl buffer (pH 7.8) containing 1 mM CaCl2, kept standing for 90 min at 37°C, and then centrifuged at 1,500 x g for 10 min, discarding the buffer. The residue was resuspended in 50 ml of the same buffer, and 0.5-ml aliquots were transferred into a 1.5-ml Eppendorf tube. Tubes were incubated at 37°C for 15 min on a shaker, and then 25 μl of culture supernatant (see below) was added to each tube containing preincubated Azocoll. The mixture was incubated for 4 h at 37°C on a shaker and then centrifuged at 1,500 x g for 5 min, followed by measurement of absorbance at 540 nm.
The preparation of bacteria for the gelatinase activity assays was the same as the preparation of bacteria for translocation. Approximately 8 x 107 bacteria in 0.5 ml of tissue culture medium (31) were added to confluent T84 cells that had been grown for 4 to 5 days in 12-well cell culture plates (Corning Incorporated, Corning, N.Y.) and were then incubated at 37°C, 5% CO2. After 6 h of incubation, about 0.5 ml was removed to a 1.5-ml Eppendorf tube and then centrifuged at 1,500 x g for 5 min, and the supernatant was used for gelatinase activity assay; some samples were incubated for up to 24 h. Bacteria were also grown in tissue culture medium (31) alone and processed similarly.
Cloning of gelE and complementation of TX5264. A 1,767-bp DNA fragment containing the gelE gene and its promoter region was amplified by using the primers 5'-GGCGAATTCGCTATGGTATTG (forward primer) and 5'-GCGGGATCCTCATTCATTGACCAGAACA (reverse primer) and cloned into the shuttle vector pAT18 (Eryr) (25) to produce plasmid pTEX5438. The plasmid was then transformed into the gelE deletion mutant TX5264, and Eryr transformants were plated onto the gelatin-TH plates to determine the effect of complementation on the production of gelatinase.
Introduction of gelE and fsr genes into JH2-2 and TX1322. The plasmids pTEX5438, which contains gelE, and pTEX5249, which contains fsrA, fsrB, and fsrC genes but only the first 395 bp of gelE, were transformed into TX1322 and JH2-2, respectively, and Eryr transformants were plated onto gelatin-TH plates to verify gelatinase activity.
Statistical analysis. Analysis of variance (ANOVA) with Bonferroni's posttest was performed using GraphPad (San Diego, Calif.) Prism software version 4.00 for comparing multiple continuous variables, and Fisher's exact test was used for categorical results comparing GelE+ and GelE– human isolates.
RESULTS
Effect of gelE and fsr genes on translocation of E. faecalis across T84 monolayers. The fsr-gelE/sprE locus is shown in Fig. 1A. Six mutants of this locus generated in OG1RF, TX5240 (fsrA disruption mutant), TX5241 (fsrB disruption mutant), TX5242 (fsrC disruption mutant), TX5266 (fsrB deletion mutant), TX5128 (gelE disruption mutant), and TX5264 (a gelE deletion mutant), all previously shown to be gelatinase negative in the gelatin-TH plate assay, were compared with wild-type OG1RF in the transcytosis assay. At least two individual experiments were performed (4 wells each) with all strains, and consistent results were obtained. In the experiments shown in Fig. 1B, OG1RF cells were not detected in the lower chamber at 0 h, were detected in the lower chamber of 11 of 12 wells at 6 h, and were detected in all 12 wells at 8 h; this degree of translocation was consistently seen throughout all experiments with OG1RF in the present work and previously (31). The fsr disruption mutants (TX5240, TX5241, and TX5242), the fsrB deletion mutant (TX5266), and the gelE disruption mutant (TX5128) were not recovered from the lower chamber of any of the 12 wells at 6 or 8 h (Fig. 1B), while the gelE deletion mutant TX5264 (GelE– SprE+), which produces a "superactive" form of SprE (6), was detected in the lower chambers of 3 out of 12 wells at 6 h and 8 out of 12 wells at 8 h, although the number of bacteria translocated across T84 monolayers was significantly lower than that for wild-type OG1RF (P < 0.001) (Fig. 1B).
Effect of complementation on translocation of E. faecalis across T84 monolayers. To further explore the importance of GelE for translocation, the gelE deletion mutant TX5264 was complemented with the whole-length gelE gene, resulting in TX5439. In concurrently performed experiments, TX5264 again showed significantly less ability to translocate across T84 monolayers than wild-type OG1RF at 6 and 8 h (Fig. 1B), while complemented cells (TX5439), which were gelatinase positive, were detected in the lower chambers of all 8 wells at 6 and 8 h (Fig. 1B).
We also tested our disruption mutants after complementation by pTEX5249, which contains the fsr genes (Fig. 1A). Each of the complemented fsr mutants (TX5244, TX5245, and TX5246) which tested gelatinase positive was able to translocate across T84 monolayers. TX5244 cells were detected in the lower chambers of 6 of 8 wells at 6 h and all 8 wells at 8 h, and TX5245 and TX5246 cells were detected in the lower chambers of 8 wells at 6 and 8 h, although the CFU of translocated TX5244 were fewer than those of wild-type OG1RF (P < 0.001) (Fig. 1B). As expected, the gelE disruption mutant TX5128 was still gelatinase negative by plate assay after introduction of the fsr locus and did not show translocation across T84 monolayers (data not shown).
Gelatinase activity of gelE/fsr mutants and their complemented derivatives. Since the growth and detection of gelatinase activity on gelatin-TH plates may not reflect what is occurring in the translocation experiments, we also tested gelatinase activity of E. faecalis grown in tissue culture medium with T84 cells and with a more quantitative method described previously by Nakayama et al. (12) (Fig. 1B). None of the above-described disruption mutants showed detectable gelatinase activity, consistent with the results from plate assay. The complemented fsrA mutant (TX5244) showed less gelatinase activity than OG1RF (optical density at 540 nm [OD540] of 1.1 ± 0.8 for TX5244 compared to 2.3 ± 0.2 for OG1RF), while complemented fsrB and fsrC mutants (TX5245 and TX5246) showed a level of gelatinase activity similar to that of OG1RF (Fig. 1B). The gelE deletion mutant TX5264 (GelE– SprE+) showed a much lower level of proteolytic activity than OG1RF (an OD540 of 0.1 ± 0.1 for OG1RF compared to an OD540 of 2.3 ± 0.2 for TX5264); after complementation of TX5264 by gelE (pTEX5438), production of gelatinase was fully restored (Fig. 1B). For all these OG1RF derivatives, gelatinase activity correlated with translocation capability across T84 monolayers (Fig. 2).
Effect of providing gelE or fsr genes on translocation of JH2-2 and TX1322 across T84 monolayers. We next transformed pTEX5249 containing fsr genes into two gelatinase phenotype-negative and gelE genotype-positive strains, JH2-2 and TX1322, neither of which showed translocation across a T84 monolayer at 6 or 8 h (Fig. 3). The resulting strains, TX5443 and TX5441, were gelatinase positive, but the gelatinase produced as measured by OD was less than that of OG1RF (an OD540 of 1.7 ± 0.2 and 1.9 ± 0.4 for TX5443 and TX5441 compared to an OD540 of 2.5 ± 0.2 for OG1RF) (Fig. 3). During translocation, no TX5443 cells were detected in lower chambers at 6 h, but translocation was observed with 6 of 8 wells at 8 h, although the CFU in the lower chamber were only 101 to 102, which is significantly lower than that of OG1RF (P < 0.001) (Fig. 3). For JH2-2 complemented with the fsr genes (TX5441), bacteria were detected in the lower chambers of 8 of 12 wells at 6 h, with all 12 wells showing translocation at 8 h; the CFU in the lower chamber reached 102 to 104, which is also significantly lower than that for OG1RF (P < 0.001) (Fig. 3). We also transformed pTEX5438 (gelE gene) into TX1322 and JH2-2 to generate TX5442 and TX5440. For TX5442, no bacteria cells were detected in the lower chamber at 6 or 8 h (Fig. 3); for TX5440, only 1 well of 12 showed translocation at 6 h and 3 of 12 wells showed translocation at 8 h (Fig. 3). No gelatinase activity was detected with TX5440 and TX5442 at 6 h incubation (Fig. 3), consistent with the absence of fsr. All these results further confirmed that E. faecalis transcytosis is enhanced by gelatinase production and activity.
Survey of translocation of clinical and fecal isolates of E. faecalis across T84 monolayers. Twenty E. faecalis human isolates were tested in the two-chamber transcytosis system (Fig. 4). All of these strains grew well in T84 medium (ca. 109 CFU in the upper chamber at the end of the experiments). The five fecal strains that were gelatinase positive showed translocation, while the two gelatinase-negative fecal strains did not show translocation. Among the endocarditis strains, the four gelatinase-positive strains showed translocation, and the three gelatinase-negative strains did not show translocation; the translocation of gelatinase-positive strain TX0040 cannot be seen in the figure due to the low percentage compared to that of OG1RF (0.2% at 6 h and 0.8% at 8 h). Among urine strains, the three gelatinase-positive strains showed translocation, two gelatinase-negative strains did not show translocation, and one gelatinase-negative strain showed translocation in a single well of 8 wells. The strains (TX0018, TX0040, and TX0614) which showed less translocation than OG1RF also produced less gelatinase than OG1RF. However, TX1317 and TX0231, which showed 17 and 22 times more translocation than OG1RF at 6 h, did not show significantly higher gelatinase activity than OG1RF. TX1341 showed low gelatinase activity, but the strain performed almost the same as OG1RF in the transcytosis assay. It was also observed that one of the human isolates, TX0231, which showed a high level of translocation and produced a moderate level of gelatinase activity under translocation conditions, did not show detectable gelatinase activity in the absence of T84 cells (an OD540 of 1.8 ± 0.7 for growth in the presence of T84 cells compared to an OD540 of 0.0 ± 0.0 for growth in tissue culture medium alone), while all the other strains showed a similar level of gelatinase activity with or without T84 cells.
DISCUSSION
An extracellular proteolytic enzyme of E. faecalis was reported 40 to 50 years ago and was later shown to be a Zn-metalloprotease (1, 2, 4, 21); this enzyme also appeared to be present in some of the first known clinical isolates of enterococci, which were classified in the early 1900s as "Streptococcus faecalis var. zymogenes," referring to their production of proteolytic and hemolytic activity (20). This "gelatinase," purified by Makinen et al. in 1989, was shown to be able to degrade a number of substrates including Azocoll, gelatin, insulin B-chain, human endothelin, collagen, E. faecalis sex pheromone-related peptides, and several mammalian bioactive peptides (8, 9). More recently, gelatinase was also shown to cleave misfolded surface proteins of E. faecalis as well as fibrin, to reduce E. faecalis supernatant pheromone levels, and to affect E. faecalis chain length, possibly by activation of an autolysin (27). Gelatinase has been a factor of interest to our laboratory and has been shown by our group and others to be important for virulence in a mouse peritonitis model (23), the nematode worm Caenorhabditis elegans (22), and a rabbit endophthalmitis model (3). Downstream and cotranscribed with gelE (24) is sprE, encoding a serine protease, and upstream are fsrA, fsrB, and fsrC, which are homologs of agrA, agrB, and agrC of Staphylococcus aureus (14, 15). The fsr genes form a quorum-sensing system which appears to be autoregulated and to positively regulate the expression of gelatinase and serine protease (14, 15).
To study the possible contribution of gelatinase to E. faecalis translocation across human intestine, we first examined the previously constructed gelE mutants (22, 23) in our in vitro translocation model. We prolonged the translocation time from 6 to 8 h, since results from the latter time point showed less variation among the Transwells. The resistance across the membrane remained high at 8 h (data not shown), indicating that the T84 monolayer was still intact. We also used a more quantitative method described previously by Nakayama et al. (12), which measures proteolytic cleavage of Azocoll, and assayed bacteria grown in the same medium and conditions as those used for translocation, and with T84 cells. Our results showed that the gelE disruption mutant TX5128 (GelE– SprE+) was consistently unable to translocate across T84 monolayers, while the gelE deletion mutant TX5264 (SprE+) showed significantly decreased translocation compared to that of wild-type OG1RF; complementation of the gelE deletion mutant in trans with whole-length gelE and its upstream promoter region restored translocation. These results indicate that gelE is important for E. faecalis translocation. Subsequent analysis of gelatinase activity, as measured by proteolysis of Azocoll, indicated a clear correlation between gelatinase activity of bacteria grown in tissue culture medium plus T84 cells and translocation: the gelE disruption mutant showed no detectable activity against Azocoll, the gelE deletion mutant showed a minimal amount of activity (OD540 of 0.1 ± 0.1 compared to 2.3 ± 0.3 for OG1RF), and complementation fully restored activity (OD540 of 2.7 ± 0.1). The residual proteolytic activity shown by the gelE deletion mutant is likely due to production of a superactive serine protease in this mutant (6), the activity of which is different from that of SprE in wild-type OG1RF due to different processing (6).
Since the expression of gelE is positively regulated by fsr genes, we also examined our previously constructed fsr mutants and complementation derivatives (14, 15). None of the three fsr disruption mutants or the fsrB deletion mutant showed translocation or detectable gelatinase activity. Complementation of fsrB and fsrC disruption mutants with the fsr-containing plasmid pTEX5249 resulted in restoration of their gelatinase activity and translocation, while complementation of the fsrA disruption mutant partially restored its translocation, which may be due to its lower gelatinase activity (OD540 about half of that of OG1RF). Overall, the results with different OG1RF derivatives not only indicate that gelatinase is critical for E. faecalis translocation across T84 monolayers but also demonstrate a correlation between the Azocoll-related proteolytic activities and translocation capabilities among these strains (Fig. 2). Although the reason for the different levels of gelatinase activity in the different complementation derivatives is not clear, previous analysis of the expression of fsr genes revealed that fsrB and fsrC are cotranscribed from a promoter upstream of fsrB while a separate promoter is used for fsrA, consistent with fsrB and fsrC behaving the same in complementation. Why providing the fsr genes in multicopy does not fully restore the defect of the fsrA mutant is not known.
We also studied two gelatinase nonproducers of E. faecalis: TX1322 and JH2-2. The two strains were previously shown to have the gelE gene but to lack fsrA and fsrB, and the latter strain was previously found to be unable to translocate T84 monolayers (31). A 23.9-kb deletion involving fsrA and fsrB was first described by Nakayama et al. in E. faecalis urine isolates obtained from a Japanese hospital (13), and this region was later found by our group, using colony hybridization and PCR, to be absent in 60 E. faecalis isolates (including TX1332 and JH2-2) from different clinical and geographical sources, which account for ca. 28% of the E. faecalis isolates tested by us (17). Neither TX1322 nor JH2-2 showed detectable gelatinase activity or translocation. Introduction of fsr into these strains resulted in successful translocation, which can be explained by activation of gelatinase production by providing the Fsr function. When compared with OG1RF, the two fsr-complemented isolates showed lower levels of translocation, as well as lower levels of gelatinase activity, than OG1RF. Introduction of gelE without fsr into the two gelatinase nonproducers had no effect on their gelatinase activity under the conditions tested, consistent with gelatinase production and activity being tightly controlled by the Fsr system in these isolates. However, JH2-2 provided with pTEX5438 showed a low degree of translocation; the difference in translocation of TX1322 (pTEX5438) compared to that of JH2-2 (pTEX5438) may suggest a level of gelatinase activity not detectable in our assay or that some other factor(s) (present in JH2-2 but not in TX1322) is assisting in the process.
To determine if our findings could be extended to other E. faecalis strains, 20 human isolates (which are different from the ones tested previously) were tested for gelatinase activity and in the translocation model, and the results showed a positive correlation between the two phenotypes: 12 of 12 gelatinase-positive isolates were able to translocate, while the gelatinase-negative isolates were either unable to translocate at all (7 of 8) or translocated at a very low level (1 of 8). When combined with the translocation results obtained from the previous study (31) and our subsequent determination of their gelatinase activity (17), the capability of translocation of the two groups (GelE+ versus GelE–) is significantly different, with translocation occurring in a total of 167 out of 185 transwells at 6 h for GelE+ strains compared to a total of 7 out of 128 transwells for GelE– strains (P < 0.001 by Fisher's exact test). On the other hand, the level of translocation by gelatinase-positive human isolates varied considerably, which cannot be completely explained by the level of gelatinase activity, suggesting that although gelatinase is critical for E. faecalis translocation, other factor(s) are also important in this process. We also noticed, consistent with our previous study (31), that fecal isolates showed greater translocation than isolates from other sources, although we did not find an obvious difference in the percentage of GelE+ or in the amount of gelatinase produced by isolates from different sources in the present study, nor did we find obvious differences in our previous work (17). This observation may suggest that the fecal isolates (which reside in the human intestine) are enriched in attributes which, together with gelatinase, facilitate their translocation across human intestine, and after leaving the intestine and entering other sites of the human body, these attributes are not needed and are thus lost or repressed. However, the number of isolates tested is too limited to do more than pose this hypothesis. Interestingly, in the present study, we found that one of the human isolates, TX0231, showed a high level of translocation but, when grown in tissue culture medium, produced gelatinase activity only in the presence of T84 cells, indicating induction of gelatinase activity by T84 cells in this isolate. This observation raises the possibility that gelatinase activity of E. faecalis can be activated by interaction with certain in vivo sites or condition(s).
The mechanism by which gelatinase enhances translocation is unknown. Because of the prominent chaining of gelE mutants (27), we assayed a previously constructed autolysin mutant (16) as well as mutants in a putative regulatory gene, all showing long cell chains similar to those of the gelE mutants, and found that these mutants had a translocation capability similar to that of OG1RF (data not shown), indicating that the chaining effect per se does not cause the decrease in translocation. Whether gelatinase augments translocation from the intestinal tract in vivo, in the milieu of fecal contents, is also not known. Since we showed that two distinct strains (JH2-2 and TX1332) that are GelE– but gelE+ are still capable of producing gelatinase under certain conditions (e.g., providing fsr genes), it is also possible that some in vivo conditions may induce gelatinase production, enabling fsr-lacking, gelE-containing strains, which comprise about 28% of E. faecalis isolates (17), to translocate.
In conclusion, the present study showed that the E. faecalis gelatinase, which is secreted outside of E. faecalis cells and can potentially interact with E. faecalis cells as well as host cells and tissues, is critical for translocation of E. faecalis across T84 monolayers in vitro. Future studies that address functions that may be assigned to this protein as well as those looking at alternative regulation of gelE may help our understanding of the role(s) of this protease in E. faecalis infections.
ACKNOWLEDGMENTS
The present work was supported by grant NIH R37 AI47923 from the Division of Microbiology and Infectious Diseases, NIAID, to B.E.M.
REFERENCES
1. Bleiweis, A. S., and L. N. Zimmerman. 1964. Properties of proteinase from Streptococcus faecalis var. liquefaciens. J. Bacteriol. 88:653-659.
2. Cano, F. R., I. A. Casas, and L. N. Zimmerman. 1971. Purification and characterization of a protease produced by Streptococcus faecalis var. liquefaciens. Prep. Biochem. 1:269-282.
3. Engelbert, M., E. Mylonakis, F. M. Ausubel, S. B. Calderwood, and M. S. Gilmore. 2004. Contribution of gelatinase, serine protease, and fsr to the pathogenesis of Enterococcus faecalis endophthalmitis. Infect. Immun. 72:3628-3633.
4. Grutter, F. H., and L. N. Zimmerman. 1955. A proteolytic enzyme of Streptococcus zymogenes. J. Bacteriol. 69:728-732.
5. Jacob, A. E., and S. J. Hobbs. 1974. Conjugal transfer of plasmid-borne multiple antibiotic resistance in Streptococcus faecalis var. zymogenes. J. Bacteriol. 117:360-372.
6. Kawalec, M., J. Potempa, J. L. Moon, J. Travis, and B. E. Murray. 2005. Molecular diversity of a putative virulence factor: purification and characterization of isoforms of an extracellular serine glutamylendopeptidase of Enterococcus faecalis with different enzymatic activities. J. Bacteriol. 187:266-275.
7. Krueger, W. A., S. Krueger-Rameck, S. Koch, V. Carey, G. B. Pier, and J. Huebner. 2004. Assessment of the role of antibiotics and enterococcal virulence factors in a mouse model of extraintestinal translocation. Crit. Care Med. 32:467-471.
8. Makinen, P. L., D. B. Clewell, F. An, and K. K. Makinen. 1989. Purification and substrate specificity of a strongly hydrophobic extracellular metalloendopeptidase ("gelatinase") from Streptococcus faecalis (strain 0G1-10). J. Biol. Chem. 264:3325-3334.
9. Makinen, P. L., and K. K. Makinen. 1994. The Enterococcus faecalis extracellular metalloendopeptidase (EC 3.4.24.30; coccolysin) inactivates human endothelin at bonds involving hydrophobic amino acid residues. Biochem. Biophys. Res. Commun. 200:981-985.
10. Murray, B. E. 1990. The life and times of the enterococcus. Clin. Microbiol. Rev. 3:46-65.
11. Murray, B. E., and G. M. Weinstock. 1999. Enterococci: new aspects of an old organism. Proc. Assoc. Am. Physicians 111:328-334.
12. Nakayama, J., Y. Cao, T. Horii, S. Sakuda, A. D. Akkermans, W. M. de Vos, and H. Nagasawa. 2001. Gelatinase biosynthesis-activating pheromone: a peptide lactone that mediates a quorum sensing in Enterococcus faecalis. Mol. Microbiol. 41:145-154.
13. Nakayama, J., R. Kariyama, and H. Kumon. 2002. Description of a 23.9-kilobase chromosomal deletion containing a region encoding fsr genes which mainly determines the gelatinase-negative phenotype of clinical isolates of Enterococcus faecalis in urine. Appl. Environ. Microbiol. 68:3152-3155.
14. Qin, X., K. V. Singh, G. M. Weinstock, and B. E. Murray. 2001. Characterization of fsr, a regulator controlling expression of gelatinase and serine protease in Enterococcus faecalis OG1RF. J. Bacteriol. 183:3372-3382.
15. Qin, X., K. V. Singh, G. M. Weinstock, and B. E. Murray. 2000. Effects of Enterococcus faecalis fsr genes on production of gelatinase and a serine protease and virulence. Infect. Immun. 68:2579-2586.
16. Qin, X., K. V. Singh, Y. Xu, G. M. Weinstock, and B. E. Murray. 1998. Effect of disruption of a gene encoding an autolysin of Enterococcus faecalis OG1RF. Antimicrob. Agents Chemother. 42:2883-2888.
17. Roberts, J. C., K. V. Singh, P. C. Okhuysen, and B. E. Murray. 2004. Molecular epidemiology of the fsr locus and of gelatinase production among different subsets of Enterococcus faecalis isolates. J. Clin. Microbiol. 42:2317-2320.
18. Runkel, N. S., F. G. Moody, G. S. Smith, L. F. Rodriguez, Y. Chen, M. T. Larocco, and T. A. Miller. 1993. Alterations in rat intestinal transit by morphine promote bacterial translocation. Dig. Dis. Sci. 38:1530-1536.
19. Sartingen, S., E. Rozdzinski, A. Muscholl-Silberhorn, and R. Marre. 2000. Aggregation substance increases adherence and internalization, but not translocation, of Enterococcus faecalis through different intestinal epithelial cells in vitro. Infect. Immun. 68:6044-6047.
20. Sherman, J. M. 1937. The streptococci. Bacteriol. Rev. 1:3-97.
21. Shugart, L. R., and R. W. Beck. 1964. Purification and activity of proteinase of Streptococcus faecalis var. liquefaciens. J. Bacteriol. 88:586-590.
22. Sifri, C. D., E. Mylonakis, K. V. Singh, X. Qin, D. A. Garsin, B. E. Murray, F. M. Ausubel, and S. B. Calderwood. 2002. Virulence effect of Enterococcus faecalis protease genes and the quorum-sensing locus fsr in Caenorhabditis elegans and mice. Infect. Immun. 70:5647-5650.
23. Singh, K. V., X. Qin, G. M. Weinstock, and B. E. Murray. 1998. Generation and testing of mutants of Enterococcus faecalis in a mouse peritonitis model. J. Infect. Dis. 178:1416-1420.
24. Su, Y. A., M. C. Sulavik, P. He, K. K. Makinen, P.-L. Makinen, S. Fiedler, R. Wirth, and D. B. Clewell. 1991. Nucleotide sequence of the gelatinase gene (gelE) from Enterococcus faecalis subsp. liquefaciens. Infect. Immun. 59:415-420.
25. Trieu-Cuot, P., C. Carlier, C. Poyart-Salmeron, and P. Courvalin. 1991. Shuttle vectors containing a multiple cloning site and a lacZ gene for conjugal transfer of DNA from Escherichia coli to gram-positive bacteria. Gene 102:99-104.
26. Vinderola, C. G., M. Medici, and G. Perdigon. 2004. Relationship between interaction sites in the gut, hydrophobicity, mucosal immunomodulating capacities and cell wall protein profiles in indigenous and exogenous bacteria. J. Appl. Microbiol. 96:230-243.
27. Waters, C. M., M. H. Antiporta, B. E. Murray, and G. M. Dunny. 2003. Role of the Enterococcus faecalis GelE protease in determination of cellular chain length, supernatant pheromone levels, and degradation of fibrin and misfolded surface proteins. J. Bacteriol. 185:3613-3623.
28. Waters, C. M., C. L. Wells, and G. M. Dunny. 2003. The aggregation domain of aggregation substance, not the RGD motifs, is critical for efficient internalization by HT-29 enterocytes. Infect. Immun. 71:5682-5689.
29. Wells, C. L., R. P. Jechorek, and S. L. Erlandsen. 1990. Evidence for the translocation of Enterococcus faecalis across the mouse intestinal tract. J. Infect. Dis. 162:82-90.
30. Xu, Y., K. V. Singh, X. Qin, B. E. Murray, and G. M. Weinstock. 2000. Analysis of a gene cluster of Enterococcus faecalis involved in polysaccharide biosynthesis. Infect. Immun. 68:815-823.
31. Zeng, J., F. Teng, G. M. Weinstock, and B. E. Murray. 2004. Translocation of Enterococcus faecalis strains across a monolayer of polarized human enterocyte-like T84 cells. J. Clin. Microbiol. 42:1149-1154.(Jing Zeng, Fang Teng, and)