INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

A common theme emerges when one considers the diverse and sometimes harsh parameters encountered by bacteria during the course of their life cycles: microbes are experts at adaptation and survival within often tumultuous environments. Additionally, when one considers that a single bacterial species may encounter a broad spectrum of environments during the course of its life cycle, it is not surprising that bacteria have developed complex regulatory systems and stress adaptation mechanisms to best capitalize on each environment encountered. One particularly drastic change in environment is experienced by facultative pathogenic bacteria as they transition from their natural reservoir to the host organism. For instance, Vibrio cholerae is a gram-negative bacterium that naturally exists within an aquatic reservoir and infects human beings. Within the aquatic environment, V. cholerae is found in close association with planktonic species (13). This pathogen enters its human host through contaminated food and water, passes through the gastric acid barrier, and colonizes the relatively sterile environment of the small intestine, where it produces cholera toxin (reviewed in reference 20). Recently, a system involved in adaptation to low pH that likely plays a role in survival within the host gastrointestinal tract was described (17).

The ability to adapt to changing environments and to mount a general stress response has been the focus of intense study in many bacterial species. The ability to alter gene expression patterns via alternative sigma factors often plays an important role in the adaptation process (12). Specifically, ^S, which is encoded by rpoS, has been shown to play a key role in the adaptive processes of a diverse group of bacterial species (2, 10, 11, 12, 14, 21). RpoS has been best characterized in Escherichia coli, where it has been shown to control a regulon consisting of at least 30 genes. These genes are expressed upon entry into stationary phase and are also involved in the general stress response, which is required for survival upon exposure to starvation conditions, low pH, and oxidative stress (reviewed in reference 12).

In pathogenic bacterial species, the role of RpoS in the general stress response and virulence is varied. For example, it was recently shown that Legionella pneumophila requires RpoS for growth within an amoeboid species that serves as its natural reservoir. However, unlike in E. coli, RpoS seems to play no role in the growth phase-dependent stress responses of L. pneumophila (10). RpoS has been shown to be critical for stress response in Yersinia enterocolitica in a temperature-dependent manner (2). RpoS is a critical component of low-pH survival of Shigella flexneri (21) as well as Salmonella enterica serovar Typhimurium (7). The role that RpoS plays in virulence of each of these pathogenic organisms is also varied. For example, in animal models, an rpoS mutant of Y. enterocolitica was not attenuated in virulence, whereas rpoS mutants of S. enterica serovar Typhimurium showed significant attenuation (2, 7). In fact, the 50% lethal dose of an rpoS Salmonella strain is 1,000-fold higher than that of the wild type (7).

The rpoS orthologue of V. cholerae was recently identified and shown to play a crucial role in survival under a variety of stressful situations, including exposure to hydrogen peroxide, hyperosmolarity, and nutrient deprivation (22). In contrast to Salmonella, however, the V. cholerae RpoS was not required for colonization of the murine intestine. A separate study tested the role of RpoS in the ability of V. cholerae to mount an adaptive stress response, known as the acid tolerance response, upon exposure to acidic conditions (17). Once again, in contrast to the results with Salmonella, RpoS was not found to play a role in survival of V. cholerae under acidic conditions. In the present study, competition assays were used to analyze the ability of the rpoS strain to colonize the suckling mouse small intestine. Here we show that in contrast to previous reports, an rpoS null strain of V. cholerae is attenuated in its ability to colonize the small intestine.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strain and plasmid construction. All strains and plasmids used in this study are listed in Table 1. pDSM747 was constructed by PCR amplification of a 2,611-bp fragment from the chromosome of C6709-1, using Taq polymerase and primers RpoS1 (5'-GCGAGATCTGTTGAACCTGTCGGTAA-3') and RpoS-R-mut (5'-GGTGTTGATGGATGAGAT-3'). The PCR product was ligated directly into pGEMT (Promega) and then liberated by digestion with SalI/SphI. The resulting fragment was subcloned into similarly digested pCVD442 (6). All plasmid integration mutations were made using pGP704 (18), while deletions were made using pCVD442. All plasmids used for construction of insertion mutations were mobilized into V. cholerae from E. coli SM10pir as previously described (17), and all integration mutations were subsequently verified by Southern blot analysis.

Growth conditions. All strains were maintained at 80°C in Luria-Bertani (LB) broth containing 30% glycerol. All strains were grown at 37°C in LB broth with the following exception: pFY7, encoding ampicillin resistance, was cured from DSM-V506 by growth at 42°C. After growth at 42°C, DSM-V506 was plated on LB agar (LB), and colonies were replica plated to LB supplemented with ampicillin. Ampicillin-sensitive colonies were colony purified on LB and retested for ampicillin sensitivity. Ampicillin and streptomycin were used at concentrations of 100 µg ml¹. Counterselection of pCVD442 was accomplished by plating on LB lacking NaCl but supplemented with 10% sucrose followed by growth at 30°C. All growth curves, whether single or competitive, were done by diluting overnight cultures into LB broth. Single-strain growth assays were done with a 200-fold dilution, while competitive assays used a 1,000-fold dilution of each of the appropriate strains. CFU were determined by serial dilution and plating on LB supplemented with X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside; 40 µg ml¹) and/or the appropriate antibiotic.

Frequency of plasmid insertion. The frequency of plasmid insertion was determined at the rpoS, iviVI, and orf3 loci in the following manner. E. coli SM10pir (donor) was used for mobilization of each suicide construct into V. cholerae C6709-1 (recipient). Overnight cultures of donor and recipient were mixed together at a 1:1 (vol/vol) ratio, and 50 µl was spotted onto dry LB plates and incubated at 37°C for 4 h. The titer of each donor and recipient strain in the overnight cultures was concurrently determined by serial dilution and plating on LB. After 4 h, the mating mixture was excised from the agar plate, resuspended in 2 ml of LB, and then plated on LB supplemented with ampicillin and streptomycin to select for transconjugants. The frequency of plasmid integration at each locus was calculated by division of the total number of transconjugants by the total number of recipient cells per mating.

Reversion of the rpoS mutation. A DNA fragment that includes the entire 1,008-bp rpoS coding sequence, as well as 1,009 bp of upstream and 594 bp of downstream sequence, was amplified from C6709-1 and cloned into pCVD442 (6), resulting in the allelic exchange vector pCVD442::'nlpD-rpoS-mutS' (named pDSM-747). pDSM-747 was mobilized into DSM-V491 and DSM-V717 as previously described (17), and subsequent double-crossover products were isolated as previously described (6). Replacement of rpoS by the wild-type rpoS allele in the sucrose-resistant and ampicillin-sensitive strains was confirmed by Southern blot analysis.

In vitro and in vivo competition assays and peroxide exposure assays. Hydrogen peroxide killing assays (22) and competition assays (17) were conducted as previously described. The output ratio from each in vivo and in vitro competition was corrected for any deviations in the inoculum ratio from a value of 1:1.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

An rpoS null derivative of V. cholerae C6709-1 is defective for intestinal colonization. Strain DSM-V382, which contains a plasmid insertion within the V. cholerae rpoS coding sequence, was previously constructed to assess the phenotype of an rpoS mutant in response to acid exposure (17). DSM-V382 exhibited enhanced sensitivity to hydrogen peroxide exposure, as expected for an rpoS strain (22), but was fully competent for survival upon exposure to low pH (17). Nevertheless, we hypothesized that RpoS might play an important role in the establishment and persistence of V. cholerae infections since it has been shown to be essential for full virulence of S. enterica serovar Typhimurium, which also infects its host via an oral route. To address this hypothesis, DSM-V382 was coinoculated with the fully virulent, isogenic, LacZ strain AC-V168 into 5-day-old suckling CD1 mice. In this competition assay, DSM-V382 was attenuated in its ability to successfully colonize the suckling mouse small intestine, as indicated by its in vivo competition index of 0.2 (Table 2). Strains with an equivalent ability to colonize would be expected to show a competition index of approximately 1.0. Thus, the V. cholerae rpoS strain was fivefold attenuated in colonizing the suckling mouse intestine. A second rpoS strain with a different plasmid insertion mutation, DSM-V490, was found to have the same defect in intestinal colonization (Table 2).

View larger version (31K): [in this window] [in a new window]   FIG. 1.   (A) Southern blot analysis of the rpoS region. Chromosomal DNA harvested from the indicated strains was digested with HindIII and hybridized with a probe specific for nlpD, which lies immediately upstream of the rpoS coding sequence. Lane C6709-1 shows the presence of the wild-type rpoS HindIII fragment, while lane DSM-V491 shows a smaller band due to an in-frame deletion of the rpoS coding sequence. Three and two independent isolates of reverted DSM-V506 and DSM-V491, respectively, are shown. In each case, the rpoS band is absent and the wild-type rpoS band is regained. (B) Hydrogen peroxide sensitivity, shown as percent survival after 30 min of exposure to hydrogen peroxide as described in Materials and Methods. In each of the three data sets, a C6709-1 wild-type control is shown, as the three experiments were performed on different days.

rpoS strains are viable in the absence of second-site suppressor mutations. Since our results differed from published reports of the role of rpoS in colonization of the suckling mouse, we considered the possibility that second-site suppressor mutations might be required for the viability of an rpoS strain and that such mutations might have effects on virulence. To address this possibility, quantitative mating assays were conducted to determine the frequency of plasmid disruption at the rpoS locus in comparison with two nonessential genes.

The rpoS phenotype can be complemented in trans. To demonstrate that the rpoS mutation was responsible for the decreased competition index, we complemented the rpoS mutation by providing rpoS in trans using pFY7, a low-copy-number plasmid containing the entire rpoS promoter and coding region (22). This construct was previously shown to complement an rpoS mutation during in vitro growth under nutrient and oxidative stress test conditions (22). pFY7 was mobilized into DSM-V491 (rpoS) to generate strain DSM-V506. DSM-V506 regained wild-type level hydrogen peroxide resistance (Fig. 1B), indicating that functional RpoS was being produced. In competition assays with C6709-1, DSM-V506 showed an in vivo competition index of 1.2 (Table 2), establishing that the low-copy-number plasmid containing rpoS complements the in vivo colonization defect.

View larger version (16K): [in this window] [in a new window]   FIG. 2.   (A) Growth kinetics of DSM-V506 and C6709-1. For each strain, overnight cultures were diluted 1:200 into fresh LB and grown with aeration at 37°C. At the indicated times, the OD600 was read for each and plotted as a function of time. (B) In vitro competition growth kinetics of DSM-V506 and DSM-V583. A 1:1 mixture of overnight cultures of each strain was diluted 1:1,000 into fresh LB. The titer of each strain was determined at the indicated time points by plating on LB supplemented with X-Gal. In addition, the OD600 was read at each time point. The percentage of each colony type found at each time point was used to extrapolate the OD600 of each strain based on the total OD600.

The rpoS phenotype is revertible. Since interpretation of the intestinal colonization experiments utilizing pFY7 plasmid complementation is confounded by the ability of strains containing this plasmid to outcompete wild-type strains in vitro, we decided to revert the rpoS deletion by allelic exchange. This was done in the original rpoS strain, DSM-V491, and in the pFY7-containing strain DSM-V506, in order to confirm that phenotypes seen throughout the course of these experiments were not due to the presence of second-site mutations acquired spontaneously. The various strain constructions made in this study are depicted in Fig. 3.

View larger version (33K): [in this window] [in a new window]   FIG. 3.   Genetic manipulations of V. cholerae C6709-1 presented in this report. Strains of equivalent genotype and phenotype are in boxes. The box with a solid line represents the presence of a wild-type chromosomal copy of rpoS and wild-type growth and virulence; the box with a dashed line represents loss of rpoS function, sensitivity to H2O2 in vitro, and reduced colonization in vivo; the box with alternating short and long dashes represents the presence of a plasmid expressing rpoS in trans, resistance to H2O2, increased competitive fitness in vitro, and wild-type colonization.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The intestinal pathogen V. cholerae has a complicated life cycle that includes growth within an aquatic environment, oral ingestion by human hosts, passage through the low pH environment of the stomach, colonization within the small intestine, and subsequent dissemination in the cholera stool back into its aquatic niche. The study of environmentally induced gene regulation of V. cholerae has primarily focused on the production and regulation of virulence factors within the host environment (16, 20). Recently, however, the role of the stationary-phase sigma factor RpoS in regulation of genes required for surviving a variety of environmental stresses was demonstrated (22). In addition, the V. cholerae RpoS was found to positively regulate expression of at least 25 different genes upon entry into stationary phase (22). Similar but diverse roles for RpoS in response to environmental stresses have been demonstrated in a variety of bacterial species, including E. coli, S. flexneri, Y. enterocolitica, L. pneumophila, and S. enterica serovar Typhimurium (2, 3, 10, 21).

The role of RpoS in colonization and virulence is as diverse as the many pathogenic bacterial species in which it has been studied. A potential role for V. cholerae RpoS in the colonization and survival within a murine model of cholera was previously considered by two separate studies. In each case, it was concluded that RpoS plays no role in colonization within the suckling mouse model of cholera (15, 22). Data presented here suggest that RpoS is indeed important for intestinal colonization in this model by the El Tor biotype clinical isolate C6709-1. Specifically, we found that rpoS mutant derivatives of C6709-1 exhibit a four- to fivefold decrease in colonization compared to the wild type. This in vivo phenotype was fully complemented by a plasmid expressing rpoS in trans and by restoration of the rpoS allele on the V. cholerae genome.

We found that the provision of rpoS in trans from a low-copy-number plasmid resulted in the ability of the complemented strain to outcompete the wild-type strain in vitro. Other studies have noted that providing rpoS in trans from plasmids often results in aberrant complementation phenotypes, which likely result from overproduction or altered regulation of RpoS (10, 22). Curiously, growth curve analysis of the complemented strain showed the same growth kinetics as the wild type. We do not know the nature of the competitive advantage, but it is interesting to speculate that the complemented strain is able to acquire nutrients more efficiently than the wild-type strain. This phenomenon would be similar to the growth advantage in stationary phase (GASP) phenotype that has been demonstrated to arise spontaneously in E. coli (8, 9, 23). The GASP phenotype is often caused by mutations in rpoS, but there are no reported cases that are the result of increased rpoS expression. Rather, the GASP rpoS mutations usually result in decreased RpoS function (8). An alternative hypothesis is that the complemented strain has acquired the ability to produce and release a compound that is toxic to the wild-type bacteria. However, attempts to mimic this killing phenomenon using sterile culture supernatants of the complemented strain have been unsuccessful.

There are multiple explanations for the reported differences in the involvement of rpoS in V. cholerae colonization of suckling mice (15, 22), the most obvious and perhaps most likely being that of different strain usage by the various groups. While our mutation was constructed in the El Tor biotype strain C6709-1, an epidemic isolate from Peru in 1991, Klose and Mekalanos (15) constructed mutations in the Classical biotype strain O395. Multiple instances of differences in gene regulation and expression between the two biotypes of V. cholerae have been noted (19, 20), though none have addressed rpoS. The study by Yildiz and Schoolnik (22) was conducted using the El Tor biotype strain 92A1552, a clinical isolate from Latin America in 1992 (F. H. Yildiz, personal communication). It has previously been noted that different strains of E. coli show different phenotypes as a result of mutations or polymorphisms in rpoS (12). Other explanations include variations in the methodology of the competition assay and possible differences between the litters of mice in the different studies. Regardless, this work demonstrates a role for rpoS in V. cholerae C6709-1 colonization of the suckling mouse model.

What might be the role of the V. cholerae RpoS in colonization of the murine intestine? RpoS has been shown to function as both a positive and a negative regulator of expression of as many as 41 different proteins in V. cholerae (22). This RpoS-dependent regulation functions not only in the stationary phase but also in the exponential phase of growth (22), suggesting that RpoS is involved in expression of gene products which benefit the growing cell. Though the dynamics of colonization in suckling mice are not fully understood, after transit through the stomach and transit through the upper portion of the small intestine, V. cholerae encounters a niche that is permissive for colonization (1). Indeed, after an initial decline in bacterial number, V. cholerae undergoes a rapid growth phase that results in large numbers of progeny cells accompanied by fluid accumulation in the intestinal lumen (4). Therefore, RpoS may serve either of two functions: to aid in surviving environmental stresses encountered in vivo and/or to aid in optimizing the rapid growth phase that occurs subsequent to intestinal colonization.