- Email: [email protected]
Role of exopolysaccharide, the rugose phenotype and VpsR in the pathogenesis of epidemic Vibrio cholerae
FEMS Microbiology Letters 230 (2004) 105^113
www.fems-microbiology.org
Role of exopolysaccharide, the rugose phenotype and VpsR in the pathogenesis of epidemic Vibrio cholerae Mohammed H. Rashid a;b , Chythanya Rajanna a , Dalin Zhang a , Vincenzo Pasquale c , Laurence S. Magder a , Afsar Ali a , Stefano Dumontet c , David K.R. Karaolis a; a
Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore, MD 21201, USA b Department of Microbiology, University of Dhaka, Dhaka, Bangladesh c Dipartimento di Produzione Vegetale, Universita della Basilicata, Potenza, Italy Received 2 October 2003; received in revised form 14 November 2003; accepted 17 November 2003 First published online 6 December 2003
Abstract Vibrio cholerae, the causative agent of cholera can produce an exopolysaccharide (EPS). Some strains can also phenotypically switch from a smooth to a ‘rugose’ phenotype characterized by small wrinkled colonies, overproduction of EPS, increased biofilm formation in vitro and increased resistance to various stressful conditions. High frequency switching to the rugose phenotype is more common in epidemic strains than in non-pathogenic strains, suggesting EPS production and the rugose phenotype are important in cholera epidemiology. VpsR up-regulates Vibrio polysaccharide (VPS) genes and the synthesis of extracellular EPS (VPS). However, the function of VPS, the rugose phenotype and VpsR in pathogenesis is not well understood. We report that rugose strains of both classical and El Tor biotypes of epidemic V. cholerae are defective in the in vitro production of extracellular collagenase activity. In vivo studies in rabbit ileal loops suggest that VpsR mutants are attenuated in reactogenicity. Intestinal colonization studies in infant mice suggest that VPS production, the rugose phenotype and VpsR have a role in pathogenesis. Our results indicate that regulated VPS production is important for promoting in vivo biofilm formation and pathogenesis. Additionally, VpsR might regulate genes with roles in virulence. Rugose strains appear to be a subpopulation of cells that might act as a ‘helper’ phenotype promoting the pathogenesis of certain strains. Our studies provide new insight into the potential role of VPS, the rugose phenotype and VpsR in the pathogenesis of epidemic V. cholerae. = 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. Keywords : Vibrio cholerae; Cholera ; Virulence ; Pathogenesis; Rugose; Bio¢lm; VpsR; Exopolysaccharide ; In vivo; Colonization; Reactogenicity
1. Introduction Cholera is an epidemic diarrheal disease of humans that results in signi¢cant morbidity and mortality [1,2]. Due to the epidemic nature and severity of the disease, high transmissibility, high death-to-case ratio and ability to persist in water supplies, cholera is a public health concern. Vibrio cholerae, the bacterium causing cholera, is an organism that can persist in water supplies. Some strains have also adapted to human pathogenesis through their ability to colonize the proximal small intestine and release toxins. However, the factors involved in the emergence, patho-
* Corresponding author. Tel. : +1 (410) 706 4718; Fax : +1 (410) 706 4581. E-mail address : [email protected] (D.K.R. Karaolis).
genesis and persistence of epidemic V. cholerae are not fully understood. V. cholerae can produce an exopolysaccharide (EPS). Some strains can also switch from a smooth (EPS o¡) to a ‘rugose’ (EPS on) phenotype characterized by small wrinkled, also termed rugose, colonies [3^6]. VpsR is known to up-regulate expression of the vps genes involved in the biosynthesis of the Vibrio polysaccharide (VPS) [7]. We recently found that V. cholerae can switch at high frequency to the rugose hyperbio¢lm phenotype and that high frequency switching occurs more commonly in epidemic strains than in non-pathogenic strains suggesting that EPS and the rugose phenotype are important in the epidemiology of V. cholerae [6]. We also recently reported a molecular study on the identi¢cation of genes involved in the switch to rugose phenotype identifying novel genes not previously shown to have roles in the transition between the phenotypes [8]. We identi¢ed genes with roles in
0378-1097 / 03 / $22.00 = 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/S0378-1097(03)00879-6
FEMSLE 11329 29-12-03
Cyaan Magenta Geel Zwart
106
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
LPS synthesis, amino acid synthesis, an N-acetylmuramoyl amidase (amiB) and a novel potential regulatory gene (rocS) predicted to encode a protein containing a GGDEF domain. Proteins with GGDEF domains are found throughout the bacterial kingdom [9]. Although their function is not well understood, proteins with GGDEF domains are thought to be associated with levels of cyclic dinucleotide (c-di-GMP) signaling molecules [10^14]. In addition to defects in bio¢lm formation, VPS production and the rugose phenotype, the rocS and amiB mutants were defective in motility [8]. It appears that the switch to the rugose phenotype is a highly regulated and coordinated process and studies on these mechanisms are ongoing. It has been proposed that production of VPS, bio¢lm formation and the rugose phenotype might have an important role in the survival of V. cholerae in various environments [6,15,16]. Rugose V. cholerae strains show hyperbio¢lm formation in vitro, are highly chlorine resistant and are more resistant to UV light, extremes of acid and alkali and complement mediated lysis than are smooth strains [5,6,15,17]. Epidemic V. cholerae O1 rugose variants can cause £uid accumulation in the rabbit ileal loop model [17] and are capable of producing clinical cholera in humans [5]. However, epidemic O139 strains are typically capsulated and a previous study reported that mutation of £agellar genes in a serogroup O139 strain resulting in a rugose variant caused a defect in colonization [18]. Therefore, the role in virulence of extracellular VPS, the rugose phenotype and VpsR is still not well understood. In this study, we tested the hypothesis that VPS, the rugose phenotype and VpsR promote the pathogenesis of epidemic V. cholerae strains.
2. Materials and methods 2.1. Bacterial strains used The strains used in this study are shown in Table 1. Before use, V. cholerae strains were subcultured onto thiosulfate citrate bile salts sucrose agar (TCBS ; Oxoid) to verify purity and con¢rmed by serogrouping using polyclonal antisera (Difco).
Table 1 Bacterial strains used in this study Strain
Phenotype
Reference
N16961 smooth N16961 rugose NCTC 6585 smooth NCTC 6585 rugose DK562 DK581 DK589
wild-type El Tor biotype strain hyperbio¢lm variant of N16961 wild-type classical biotype strain hyperbio¢lm variant of NCTC 6585 N16961 Tn5: :vpsR N16961 Tn5: :vpsR N16961 Tn5: :vps(VC920)
[35] [6] [6] [6] [8] [8] [8]
FEMSLE 11329 29-12-03
2.2. Cholera toxin (CT) assay We studied the level of CT production in smooth and rugose strains of V. cholerae strain N16961 and in two independent VpsR mutants (DK562 and DK581) [8] grown under AKI conditions [19,20]. Brie£y, a fresh colony of each strain was inoculated into 15-ml tubes containing 4 ml AKI broth (1.5% bacto peptone, 0.4% yeast extract, 0.086 M NaCl and adjusted to pH 8.0 by 1 M NaHCO3 just prior to use) [19,20]. For VpsR mutants, neomycin (50 Wg ml31 ) was included in the culture medium. The cultures were incubated overnight with shaking at 200 rpm at 30‡C. The following day, 100 Wl was inoculated into a 15-ml tube containing 10 ml AKI broth (antibiotics were added where appropriate) and then incubated statically at 30‡C until OD600 = 0.3^0.4 (approximately 3 h). The cultures were then transferred into a 250-ml £ask and incubated with shaking at 30‡C and 250 rpm for 18 h and the following day 3.6 ml was removed and centrifuged at 8000 rpm for 5 min. Cell-free culture supernatants were tested for CT by GM-1 ELISA (enzyme-linked immunosorbent assay). Brie£y, Immulon 2HB (‘U’ bottom) microtiter plates (Fisher) were coated with GM1 ganglioside (100 Wl per well at 1 Wg ml31 in phosphate-bu¡ered saline (PBS)) and incubated overnight at 4‡C. Plates were washed three times with PBS, 200 Wl per well of 5% inactive fetal calf serum/PBS was added and incubated for 1 h at 37‡C and washed with PBS/ 0.05% Tween-20. Serial dilutions of CT (Sigma) were included to establish a standard curve while PBS served as a negative control. Plates were incubated for 2 h at 37‡C, washed with PBS/0.05% Tween-20 and 100 Wl per well of rabbit anti-CT antiserum (diluted 1:3200 in PBS/Tween-20 with 1% fetal calf serum) was added and the plates incubated for 1 h at 37‡C. Plates were then washed with PBS/ Tween-20 and 100 Wl of goat anti-rabbit IgG alkaline phosphatase conjugate (Kirkegaard and Perry Laboratories) diluted 1:5000 in PBS with 1% fetal calf serum was added to each well and incubated for 1 h at 37‡C. Plates were washed with PBS/Tween-20 and 100 Wl of Sigma Fast p-nitrophenyl phosphate substrate tablet diluted in sterile MiliQ water was added to each well and incubated for 45 min at 37‡C. The reaction was stopped by adding 50 Wl of 3 N NaOH and the OD405 nm determined on a plate reader (Spectra MAX250). Results were based on three independent samples for each strain (prepared on three di¡erent days), which were then tested six times. Data was analyzed by using a one-way analysis of variance or t-test. 2.3. Enzyme assay for collagenase and elastase activity V. cholerae strains were inoculated in brain heart infusion (BHI) broth and incubated overnight at 37‡C with shaking. Cultures were centrifuged and supernatants passed through 0.22-Wm ¢lters. Collagenase activity was
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
determined using the EnzChek collagenase kit (Molecular Probes) with collagen type I from bovine skin and collagen type IV from human placenta as substrates. Elastase activity was determined using the EnzChek elastase kit (Molecular Probes) with DQ elastin (soluble bovine neck ligament elastin) as the substrate. 2.4. Rabbit ileal loop studies The potential role of EPS, the rugose phenotype and vpsR in pathogenesis was examined using the rabbit ileal loop model of infection [21]. In these studies, three New Zealand white male rabbits (2 kg) (Covance Research Products) were used. Wild-type smooth and rugose variants of N16961, a vpsR mutant (DK562), and a vps mutant (DK589) were inoculated from frozen glycerol stocks onto LB agar plates. A single colony of each strain was then inoculated into 4 ml AKI broth (containing antibiotics where appropriate) and incubated at 37‡C overnight at 250 rpm. Following overnight incubation, the OD600 was adjusted to 1.2 with fresh AKI broth. From this culture, 100 Wl was transferred into into 250-ml £asks containing 10 ml AKI broth (containing antibiotics where appropriate) and incubated for approximately 3 h at 37‡C at 250 rpm until OD600 = 1.8. From this culture, 1 ml was centrifuged at 8000 rpm for 4 min, the pellet washed twice with 1.5 ml PBS and then transferred into a tube containing 10 ml PBS (OD600 V0.13^0.16). From this culture, 3 ml was removed and added to 7 ml PBS, from which the CFU ml31 was determined by plate count to be in the range of 4^8U107 ml31 . Prior to surgery, rabbits were fasted for 24 h and fed only water ad libitum. Rabbits were anesthetized, a laparotomy was performed, and the small intestine was tied o¡ into ¢ve loops (5 cm) with interloops (1 cm) between each main loop. Into each loop, using a 25-gauge needle, 1 ml was inoculated containing 107 cells of either a smooth strain, a rugose strain, vpsR, vps strains or PBS as a control. The intestine was returned to the peritoneal cavity, the incision closed, and the rabbits returned to their cage and given water ad libitum. Rabbits were sacri¢ced at 18 h, the peritoneal cavity opened and the small intestine removed. The loops were examined macroscopically, the £uid in each loop was removed and measured and an aliquot removed and plated onto appropriate antibiotic plates to determine V. cholerae counts (CFU ml31 ). Colonies ere subcultured onto TCBS agar to con¢rm their identity as V. cholerae. The £uid accumulation ratio (FAR) was determined by measuring the £uid (ml) in the loops and dividing by the length (cm) of the loop. 2.5. Infant mouse colonization studies To determine whether EPS, the rugose variant of V. cholerae N16961 and vpsR have a role in intestinal colo-
FEMSLE 11329 29-12-03
107
nization, multiple co-infection assays based on the method of Baselski and Parker [22] were performed using the suckling infant mouse model. A single colony of either the smooth strain, rugose strain or the vpsR mutant (DK581) was inoculated into a test tube containing 3 ml LB broth and incubated at 37‡C overnight at 250 rpm. The overnight cultures of each strain were diluted 1:1000 in fresh LB broth (5 Wl culture in 5 ml LB broth) and either used alone for individual assays or 1 ml was mixed in a 1:1 ratio with the another strain to be tested in coinfection assays. Each culture (individual and mixed) contained 50 Wl Evans blue food coloring. Five-day-old suckling CD-1 mice were inoculated orally using a 25-g syringe containing 50 Wl of culture (individual strains or mixtures) representing 105 CFU con¢rmed by plate count. Following incubation (5 or 24 h), the mice were sacri¢ced and their small and large intestines were removed, placed in 2 ml PBS, mechanically homogenized using a Virsonic 60 (Virtis) blender, serially diluted and plated onto LB agar supplemented with polymyxin B 50 U (and neomycin 50 Wg ml31 where appropriate) to enumerate V. cholerae CFU. Colonies were con¢rmed as V. cholerae by subculturing onto TCBS agar. The competitive index was calculated by dividing the rugose variant or vpsR mutant output number by the wild-type output number. Strains with an equivalent ability to colonize are expected to have a competition index of approximately 1.0.
3. Results and discussion Other than V. cholerae, a rugose-like phenotype with small wrinkled colonies and its associated properties has been reported in several species including Salmonella enterica Enteritidis [23], S. enterica typhimurium [24], Pseudomonas aeruginosa [25] and Enterobacter sakazakii [26]. As such, it is now becoming increasingly recognized that rugose strains might have an important role in V. cholerae and in several other species suggesting that these ‘variants’ might be more important and common than previously thought. Our recent ¢nding of conditions that promote high frequency switching to the rugose phenotype in V. cholerae and that switching appears to be more common in epidemic strains than in non-pathogenic strains suggests VPS production and the rugose phenotype is important in the epidemiology of cholera [6]. To better understand the molecular basis of switching, we reported the identi¢cation of genes involved in the switch to the rugose phenotype [8]. Although rugose variants of V. cholerae are known to be highly resistant to various stressful conditions in vitro and show hyperbio¢lm formation in vitro, the role of VPS, the rugose phenotype and VpsR in pathogenesis is not well understood. The studies reported here were performed in order to better understand the potential role of VPS, the rugose phenotype and VpsR in pathogenesis using a variety of in vitro and in vivo models of infection.
Cyaan Magenta Geel Zwart
108
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
3.1. CT production by smooth and rugose strains Although rugose strains can cause £uid accumulation in rabbit ileal loops and diarrhea in human volunteers, suggesting that they are virulent [5,17], the level of CT produced by rugose hyperbio¢lm variants compared to typical smooth cells has not been examined. In our studies, although extracellular VPS is increased in rugose strains, no di¡erence was found between smooth and rugose strains of N16961 or between wild-type N16961 and two independent VpsR mutants (DK562 and DK581) under the conditions tested. These results suggest that extracellular CT production is una¡ected in rugose strains and that VpsR does not regulate CT. 3.2. Studies on collagenase and elastase production activity The extracellular matrix (ECM) serves as a sca¡old to stabilize tissue structure. Matrix metalloproteinases such as collagenase and gelatinase that digest collagen and gelatin (denatured collagen) are potential factors in several microbial species a¡ecting the structure of basement membranes of eukaryotic cells. Studies of proteolytic enzymes in Vibrio spp. have shown that collagenase activity is a virulence factor for host infection and pathogenesis [27]. Recent studies in Vibrio vulni¢cus found that collagenase activity is rare [28,29] while in V. cholerae expression of elastase activity is common among clinical and environmental strains from di¡erent sources [30]. Previous studies have also shown that V. cholerae produces proteases that have a potential role in tissue injury and disintegration of ECM by being able to process and activate human matrix metalloproteases [31,32]. We assessed the ability of smooth and rugose epidemic V. cholerae O1 strains (including classical and El Tor strains) for their ability to produce hydrolytic extracellular proteases (collagenase and elastase activity) that like in other species might be associated with pathogenicity. We measured enzyme activity by comparing the £uorescence produced by the degradation of £uorescent collagenase and elastase substrates using a puri¢ed enzyme provided with the respective kits as a positive control. Table 2 Collagenase and elastase activity by smooth and rugose strains of V. choleraea Strain
Collagenase I
N16961 smooth N16961 rugose NCTC 6585 smooth NCTC 6585 rugose
152 4 57 0
Collagenase IV 86 0 28 20
Elastase 17 13 0 0
a
Results are based on the analysis of overnight culture supernatants grown in BHI that were passed through 0.22-Wm ¢lters. All results are expressed as percent of positive controls and blank. Positive controls with speci¢c inhibitors have been subtracted from the data. We subtracted positive controls with enzyme inhibitor in order to eliminate bias due to non-speci¢c proteases.
FEMSLE 11329 29-12-03
Our studies showed that the rugose strains of both V. cholerae N16961 (El Tor biotype) and NCTC 6585 (classical biotype) are defective in the production of both collagenase I and collagenase IV activity compared to their smooth cell variants (Table 2). In addition, we found that smooth strain N16961 produced approximately three-fold more collagenase I and collagenase IV activity than smooth strain NCTC 6585. Interestingly, while we detected similar levels of elastase activity in smooth and rugose El Tor N16961 strains, we did not detect any elastase activity in the classical NCTC 6585. These observations suggest that rugose variants of both strains (and biotypes) are attenuated in the production of collagenase activity and that classical strain NCTC 6585 is defective in elastase activity. Since V. cholerae produces many extracellular enzymes and proteases, several of which have multiple functions depending on the substrates chosen for study, it is di⁄cult to determine whether a single protein is responsible for the observed di¡erences in collagenase activity between smooth and rugose strains and in elastase activity di¡erences between the biotypes. Our results indicate that rugosity and the production of the extracellular protease (collagenase activity) are inversely linked and since several proteases have been linked with virulence, these results are consistent with a potential defect in virulence expression in rugose strains compared to smooth strains. 3.3. Studies in the rabbit ileal loop model Rugose V. cholerae O1 strains have been shown to cause £uid accumulation in the rabbit ileal loop model [17] and to be capable of producing clinical cholera in humans, showing that these variants are pathogenic [5]. However, the role of VPS, rugose strains and VpsR in pathogenesis is not well understood. In these studies, we determined whether VPS, the rugose phenotype, and vpsR have a role in cholera pathogenesis in vivo using the rabbit ileal loop model of infection. We did not see any obvious di¡erence in FAR between the smooth strain, rugose strain, vps and vpsR strains in 18-h loops (Table 3). This result is consistent with our earlier ¢nding that there is no di¡erence in CT between the strains. Plating of £uid contents from ileal loops did not reveal any obvious phenotypic switching between the smooth and rugose phenotypes. These ¢ndings suggest that if there is phenotypic switching occurring in vivo, it is at a low frequency and below the detection limit for methods used here. However, although there was no difference in the total number of V. cholerae CFU cultured from the intestinal £uid of each loop, gross examination of the 18-h loops revealed di¡erences in pathology and reactogenicity (i.e. serosal hemorrhage and necrotic areas) between strains (Fig. 1, colour ¢gure). Loops containing wild-type smooth and rugose strains showed obvious scattered regions of serosal hemorrhage (redness), necrotic
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
109
Table 3 Studies in rabbit ileal loops showing the role of the rugose strain and vpsR in pathogenesis Sample
E¡ect on rabbit ileal loops
N16961 smooth N16961 rugose DK562 (vpsR) DK589 (vps)
19.6 17.5 17.0 24
Volume (ml)
Length (cm) 9.6 9.0 8.4 10.9
areas and reactogenicity. Ileal loops inoculated with the vps mutant DK589 (defective in VPS) appeared to be slightly less reactogenic than its wild-type smooth parent. Our ¢nding that this vps mutant is slightly less reactogenic than its smooth parent is more likely to be explained if VPS production has a role in pathogenesis. However, it is possible that locus VC0920, encoding a putative glycosyl transferase, has additional roles in virulence as previous studies by us and others have shown that some genes in V. cholerae have multiple roles and are involved in several processes including LPS production and the rugose phenotype [8,33]. Most noticeably, however, the vpsR mutant (DK562) was consistently less reactogenic than its smooth wild-type parent and showed considerably less tissue damage and pathology. Although there was no di¡erence in CFU obtained from £uid, enumeration of bacterial cells adhering to intestinal tissue was not performed, making di¡erences in reactogenicity unable to be inferred from any potential di¡erences in colonization. The observation
FAR (volume/length)
Tissue damage
2.0 1.9 2.0 2.2
++ +++ 3 3/+
of even less reactogenicity in loops containing the vpsR mutant compared to loops containing its parent is consistent with ¢ndings of the vps mutant and supports a hypothesis that VPS and vpsR have a role in pathogenesis. It is important to note that the defect in switching to the rugose phenotype in the vpsR mutant could be restored in vitro by a vpsR plasmid [8]. The relative lack of reactogenicity in this VpsR mutant compared to its parent leads us to suggest that vpsR is important in the pathogenesis of cholera and not only regulates VPS biosynthesis and the rugose phenotype, but might also regulate other genes with roles in pathogenesis and might form part of a novel virulence regulon in V. cholerae. 3.4. The infant mouse colonization model Although studies have shown that the rugose phenotype promotes bio¢lm formation and cell survival in vitro, we explored whether VPS, rugose strains and VpsR have a
Fig. 1. Colour ¢gure showing gross examination of rabbit ileal loops inoculated with V. cholerae. Note the reduction in pathology and reactogenicity in the loop containing the vspR mutant DK562 compared to the wild-type smooth strain N16961. These results were consistent in three rabbits following 18 h incubation.
FEMSLE 11329 29-12-03
Cyaan Magenta Geel Zwart
110
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
Table 4 Colony counts and co-infection studies of V. cholerae in suckling infant mice Strains
Smooth Rugose VpsR Rugose/smooth VpsR/smooth VpsR/rugose
5h
24 h
(CFU ml)
CI
(CFU ml)
CI
ND ND ND 4.2U104 /2.2U105 9.7U103 /1.8U104 5.4U105 /4.3U104
ND ND ND 0.19 (n = 7) 0.53 (n = 8) 12.55 (n = 8)
2.28U107 9.54U106 3.70U106 5.1U106 /2.3 U107 2.1U106 /2.2 U106 1.1U107 /1.1 U106
ND (n = 10) ND (n = 14) ND (n = 17) 0.22 (n = 27) 0.95 (n = 19) 10.0 (n = 22)
CI = competitive index; n = number of mice in group; ND = not done.
role in vivo by studying intestinal colonization using the suckling infant mouse model. Although we found no obvious di¡erence in the level of TcpA production between smooth and rugose strains of V. cholerae O1 strain N16961 (data not shown), we hypothesized that there might be di¡erences in the level of colonization in vivo between the strains given that rugose strains show hyperbio¢lm formation in vitro [6,15,16]. In these studies, we compared the ability of smooth strains, rugose strains and a vpsR mutant (DK581) to colonize 5-day-old infant mice. Strains were inoculated orally either alone or in standard co-infection assays. While plating of the individual initial inoculums before inoculation into mice showed that there was no di¡erence in the average counts between the strains (smooth = 3.5 U105 ml31 ; rugose = 4.8U105 ml31 ; vpsR = 4.0U105 ml31 ), and multiple in vitro co-infection assays showed no di¡erence between strains following 5 h and 24 h coinfection (data not shown), we observed signi¢cant di¡erences in the intestinal colonization ability between the strains (Table 4 and Fig. 2A,B).
(P = 0.0004; Kruskal^Wallis test) (Table 4 and Fig. 2B). The results showed that smooth strains colonize V10-fold better than rugose strains. We also found that a vpsR mutant (VPS-negative and unable to switch to the rugose phenotype) is signi¢cantly attenuated in colonization compared to both the wild-type smooth and rugose strain. These results suggest that excessive (rugose) VPS production might inhibit colonizing ability. However, the results
3.4.1. Intestinal colonization studies (5 h) No obvious ( 6 two-fold) di¡erence was found between the smooth and vpsR strain in 5 h co-infection assays. However, the smooth strain colonized ¢ve-fold better than the rugose strain in co-infection studies. The vpsR mutant colonized 12-fold better than the rugose strain in co-infection studies. Interestingly, the colonization rate for both the smooth strain and the vpsR mutant was 12- and 55-fold higher, respectively, when these strains were coinfected with the rugose strain compared to their colonization rate in the smooth/vpsR co-infection assay. These results suggest that VPS production and the rugose phenotype has an e¡ect on the intestinal colonization of V. cholerae. 3.4.2. Intestinal colonization studies (24 h) To determine the level of colonization between the strains during later stages of infection, we performed multiple 24 h individual and co-infection assays. Initially, we determined the colonization levels of individual strains. We found a statistically signi¢cant di¡erence in the level of colonization between all strains when given alone
FEMSLE 11329 29-12-03
Fig. 2. Infant mouse colonization studies of V. cholerae N16961 smooth strain, rugose, and vpsR mutant after 24 h. A: Colonization rates after 5 h incubation. B: Colonization rates after 24 h incubation. Colony counts (CFU ml31 ) represent the mean values obtained where n represents the number of mice used in that study.
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
also suggest that the inability to produce VPS, such as in the vpsR mutant, results in an even greater defect in the colonization rate. These infant mice intestinal colonization studies support our ileal loop data whereby a vpsR mutant appears to be less reactogenic, possibly due to a defect in colonization, and indicates that VPS, the rugose phenotype and vpsR have a role in pathogenesis. We then performed 24 h co-infection assays with the smooth strain, rugose strain and vpsR mutant (Table 4 and Fig. 2B). We found no di¡erence in colonization rates between a smooth and vpsR mutant (P = 0.36; Wilcoxon signed rank test). Consistent with the 5 h co-infection assays, we found that the smooth strain has a signi¢cantly ¢ve-fold greater colonization rate than the rugose strain (P = 0.0034 ; Wilcoxon signed rank test). These results are consistent with Watnick et al. [18], who reported that a serogroup O139 £agellar mutant exhibiting the rugose phenotype shows a decrease in colonization compared to its wild-type smooth parent. Also, consistent with 5 h coinfection assays, the vpsR mutant colonized 10-fold better than rugose strains in co-infection assays (P 6 0.0009; Wilcoxon signed rank test). Interestingly, although the reason is not well understood, we found a decrease in the colonization rates of the smooth and rugose strain when co-infected with the VpsR mutant compared to when these strains were inoculated alone (P = 0.006 and P = 0.012, respectively; Kruskal^Wallis test). Importantly, and again consistent with 5 h assays, the smooth strain had 10-fold better colonization ability when it was co-infected with the rugose strain compared to when it was coinfected with the vpsR mutant. Likewise, the vpsR mutant colonized ¢ve-fold better in co-infection with the rugose strain compared to when it was inoculated alone (P = 0.027; Wilcoxon test). We suggest that these ¢ndings are consistent with a hypothesis that the ability to produce VPS in a carefully regulated manner is required for e⁄cient colonization and that the rugose phenotype, at least in part as a result of VPS production, might promote the colonization of certain strains. In these studies, our hypothesis was that VPS, the rugose phenotype and VpsR have a role in virulence. Our interpretation of the data is that the results cannot be explained only by the fact that the rugose strain, possibly as a result of VPS production masking a colonization factor, is defective in colonization compared to smooth strains. If this were the case, the vpsR mutant (defective in VPS) would be expected to colonize as well as wild-type smooth strains when each of these strains is inoculated alone. In addition, we would not expect to ¢nd the colonization rate of a vpsR mutant to be increased V¢ve-fold in the presence of rugose strains compared to when the vpsR mutant is co-infected with the smooth strain. Rather, our data strongly indicate that VPS production is actually required for e⁄cient intestinal colonization. Given the observed e¡ects in pathogenesis between VPS overproduction (rugose) and VPS underproduction (vpsR mutant),
FEMSLE 11329 29-12-03
111
we predict that there should be some regulatory mechanism that controls the timing and quantity of VPS produced in vivo. As such, we are performing further studies on the regulatory system and signal molecules controlling VPS production and bio¢lm formation. In a model presented here, controlled VPS production appears to be important for intestinal colonization and rugose strains appear to promote the virulence of certain ‘smooth’ strains. In this model, smooth strains can produce VPS in a regulated manner and are expected to colonize better than rugose strains that produce excessive VPS production. If some VPS is required for e⁄cient colonization, a vpsR mutant (defective in VPS production) would be expected to colonize less than wild-type smooth and rugose strains. Consistent with this model are our ¢ndings that the wild-type smooth strains co-infected with the rugose strain colonize better than when they are co-infected with the vpsR mutant. In addition, the vpsR mutant also colonizes better in the presence of rugose cells than in the presence of wild-type smooth strains or when inoculated alone. Switching from the smooth to rugose phenotype in vivo is not implicit in this model, however, it is possible that some switching occurs in vivo at a low frequency and this event might further promote the colonization of certain strains. Therefore, although the rugose strain colonizes less than the wild-type smooth strain but better than a vpsR mutant, our results not only suggest that VPS is an important factor for colonization but also imply that rugose strains might under some conditions act as a ‘helper’ phenotype for certain smooth strains in their colonization of the intestine.
4. Conclusions It has been shown that VPS production and the switch to the rugose phenotype is linked to the type II general extracellular protein secretion pathway which is also involved in secretion of CT and hemolysin [34]. Although the regulatory pathway involved is not clear, our ¢ndings suggest that production of the potential virulence-associated collagenase and elastase activity is inversely linked to the production of VPS and the rugose phenotype. Based on these in vitro proteolytic assays, rugose strains appear to be less pathogenic in this property than wild-type smooth strains. VpsR belongs to a family of proteins with homology to signal transduction two-component regulatory systems [7]. These regulatory systems are important in the sensing of and responding to di¡erent environments. Our results strongly suggest that VpsR has a role in the pathogenesis of cholera and that in addition to regulating vps genes and VPS production, VpsR might regulate additional genes with roles in virulence. The factors involved in the emergence, pathogenesis and persistence of epidemic V. cholerae are not fully under-
Cyaan Magenta Geel Zwart
112
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
stood but are being studied by us. Until now, the nature of the biological role of VPS and the rugose phenotype in V. cholerae has not been well identi¢ed but we and others have proposed some role in environmental survival. While it is possible that VPS and the rugose phenotype have a role in the environment, we propose a new model in which VPS production and the rugose phenotype might have a role in pathogenesis. In this model, careful timing and control of VPS production is required for e⁄cient intestinal colonization and rugose strains might potentially act as a ‘helper’ phenotype in promoting the pathogenesis of certain strains. Our ¢ndings provide a new understanding of the role of VPS, the rugose phenotype and VpsR in epidemic V. cholerae and might be relevant to other species.
Acknowledgements M.H.R. is part of a joint program between the University of Dhaka and the University of Maryland and wishes to acknowledge the assistance of Chowdhury Ahsan at the University of Dhaka. We thank Guard-Petter for useful discussions. This work was supported by NIH grant AI45637. D.K.R.K. is a recipient of a Burroughs Wellcome Fund Career Award in the Biomedical Sciences.
References [1] Pollitzer, R. (1959) Cholera. Monogr. Ser. 43. World Health Organization, Geneva. [2] Kaper, J.B., Morris Jr., J.G. and Levine, M.M. (1995) Cholera. Clin. Microbiol. Rev. 8, 48^86. [3] White, P.B. (1938) The rugose variant of vibrios. J. Pathol. Bacteriol. 46, 1^6. [4] White, P.B. (1940) The characteristic hapten and antigen of rugose races of cholera and El Tor vibrios. J. Pathol. Bacteriol. 50, 160^164. [5] Morris Jr., J.G., Sztein, M.B., Rice, E.W., Nataro, J.P., Losonsky, G.A., Panigrahi, P., Tacket, C.O. and Johnson, J.A. (1996) Vibrio cholerae O1 can assume a chlorine-resistant rugose survival form that is virulent for humans. J. Infect. Dis. 174, 1364^1368. [6] Ali, A., Rashid, M.H. and Karaolis, D.K.R. (2002) High-frequency rugose exopolysaccharide production in Vibrio cholerae. Appl. Environ. Microbiol. 68, 5773^5778. [7] Yildiz, F.H., Dolganov, N.A. and Schoolnik, G.K. (2001) VpsR, a member of the response regulators of the two-component regulatory systems, is required for expression of vps biosynthesis genes and EPSETr -associated phenotypes in Vibrio cholerae O1 El Tor. J. Bacteriol. 183, 1716^1726. [8] Rashid, M.H., Rajanna, C., Ali, A. and Karaolis, D.K.R. (2003) Identi¢cation of genes involved in the switch between the smooth and rugose phenotypes of Vibrio cholerae. FEMS Microbiol. Letts. 227, 113^119. [9] Bateman, A., Birney, E., Durbin, R., Eddy, S.R., Howe, K.L. and Sonnhammer, E.L. (2000) The Pfam protein families database. Nucleic Acids Res. 28, 263^266. [10] Ausmees, N., Mayer, R., Weinhouse, H., Volman, G., Amikam, D., Benziman, M. and Lindberg, M. (2001) Genetic data indicate that proteins containing the GGDEF domain possess diguanylate cyclase activity. FEMS Microbiol. Lett. 204, 163^167.
FEMSLE 11329 29-12-03
[11] Amikam, D. and Benziman, M. (1989) Cyclic diguanylic acid and cellulose synthesis in Agrobacterium tumefaciens. J. Bacteriol. 171, 6649^6655. [12] Ross, P. et al. (1990) The cyclic diguanylic acid regulatory system of cellulose synthesis in Acetobacter xylinum. Chemical synthesis and biological activity of cyclic nucleotide dimer, trimer, and phosphothioate derivatives. J. Biol. Chem. 265, 18933^18943. [13] Ross, P., Mayer, R. and Benziman, M. (1991) Cellulose biosynthesis and function in bacteria. Microbiol. Rev. 55, 35^58. [14] Mayer, R. et al. (1991) Polypeptide composition of bacterial cyclic diguanylic acid-dependent cellulose synthase and the occurrence of immunologically crossreacting proteins in higher plants. Proc. Natl. Acad. Sci. USA 88, 5472^5476. [15] Yildiz, F.H. and Schoolnik, G.K. (1999) Vibrio cholerae O1 El Tor: identi¢cation of a gene cluster required for the rugose colony type, exopolysaccharide production, chlorine resistance, and bio¢lm formation. Proc. Natl. Acad. Sci. USA 96, 4028^4033. [16] Watnick, P.I. and Kolter, R. (1999) Steps in the development of a Vibrio cholerae El Tor bio¢lm. Mol. Microbiol. 34, 586^595. [17] Rice, E.W. et al. (1993) Vibrio cholerae O1 can assume a ‘rugose’ survival form that resists killing by chlorine, yet retains virulence. Int. J. Environ. Health Res. 3, 89^98. [18] Watnick, P.I., Lauriano, C.M., Klose, K.E., Croal, L. and Kolter, R. (2001) The absence of a £agellum leads to altered colony morphology, bio¢lm development and virulence in Vibrio cholerae O139. Mol. Microbiol. 39, 223^235. [19] Iwanaga, M. and Yamamoto, K. (1985) New medium for the production of cholera toxin by Vibrio cholerae O1 biotype El Tor. J. Clin. Microbiol. 22, 405^408. [20] Iwanaga, M., Yamamoto, K., Higa, N., Ichinose, Y., Nakasone, N. and Tanabe, M. (1986) Culture conditions for stimulating cholera toxin production by Vibrio cholerae O1 El Tor. Microbiol. Immunol. 30, 1075^1083. [21] De, S.N. and Chatterjee, D.N. (1953) An experimental study of the mechanism of action of V. cholerae on the intestinal mucous membrane. J. Pathog. Bact. 66, 559^562. [22] Baselski, V.S. and Parker, C.D. (1978) Intestinal distribution of Vibrio cholerae in orally infected infant mice: kinetics of recovery of radiolabel and viable cells. Infect. Immun. 21, 518^525. [23] Petter, J.G. (1993) Detection of two smooth colony phenotypes in a Salmonella enteritidis isolate which vary in their ability to contaminate eggs. Appl. Environ. Microbiol. 59, 2884^2890. [24] Anriany, Y.A., Weiner, R.M., Johnson, J.A., De Rezende, C.E. and Joseph, S.W. (2001) Salmonella enterica serovar Typhimurium DT104 displays a rugose phenotype. Appl. Environ. Microbiol. 67, 4048^ 4056. [25] D’Argenio, D.A., Calfee, M.W., Rainey, P.B. and Pesci, E.C. (2002) Autolysis and autoaggregation in Pseudomonas aeruginosa colony morphology mutants. J. Bacteriol. 184, 6481^6489. [26] Farmer, J.J., Asbury III, M.A., Hickman, F.W., Brenner, D.J. and Group, E.S. (1980) Enterobacter sakazakii: a new species of ‘Enterobacteriaceae’ isolated from clinical specimens. Int. J. Syst. Bacteriol. 30, 569^584. [27] Kim, S.K., Yang, J.Y. and Cha, J. (2002) Cloning and sequence analysis of a novel metalloprotease gene from Vibrio parahaemolyticus 04. Gene 283, 277^286. [28] Moreno, M.L. and Landgraf, M. (1998) Virulence factors and pathogenicity of Vibrio vulni¢cus strains isolated from seafood. J. Appl. Microbiol. 84, 747^751. [29] Rodrigues, D.P., Ribeiro, R.V. and Hofer, E. (1992) Analysis of some virulence factors of Vibrio vulni¢cus isolated from Rio de Janeiro, Brazil. Epidemiol. Infect. 108, 463^467. [30] Guhathakurta, B., Sasmal, D., Pal, S., Chakraborty, S., Nair, G.B. and Datta, A. (1999) Comparative analysis of cytotoxin, hemolysin, hemagglutinin and exocellular enzymes among clinical and environmental isolates of Vibrio cholerae O139 and non-O1, non-O139. FEMS Microbiol. Lett. 179, 401^407.
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113 [31] Okamoto, T. et al. (1997) Activation of human matrix metalloproteinases by various bacterial proteinases. J. Biol. Chem. 272, 6059^ 6066. [32] Maeda, H., Okamoto, T. and Akaike, T. (1998) Human matrix metalloprotease activation by insults of bacterial infection involving proteases and free radicals. Biol. Chem. 379, 193^200. [33] Nesper, J., Lauriano, C.M., Klose, K.E., Kapfhammer, D., KraiM, A. and Reidl, J. (2001) Characterization of Vibrio cholerae O1 El Tor galU and galE mutants: in£uence on lipopolysaccharide structure, colonization, and bio¢lm formation. Infect. Immun. 69, 435^445.
FEMSLE 11329 29-12-03
113
[34] Ali, A., Johnson, J.A., Franco, A.A., Metzger, D.J., Connell, T.D., Morris, J.G.J. and Sozhamannan, S. (2000) Mutations in the extracellular protein secretion pathway genes (eps) interfere with rugose polysaccharide production in and motility of Vibrio cholerae. Infect. Immun. 68, 1967^1974. [35] Levine, M.M., Black, R.E., Clements, M.L., Nalin, D.R., Cisneros, L. and Finkelstein, R.A. (1981) In: Acute Enteric Infections in Children. New Prospects for Treatment and Prevention (Holme, T., Holmgren, J., Merson, M.H. and Mollby, R., Eds.), pp. 443^459. Elsevier, Amsterdam.
Cyaan Magenta Geel Zwart
www.fems-microbiology.org
Role of exopolysaccharide, the rugose phenotype and VpsR in the pathogenesis of epidemic Vibrio cholerae Mohammed H. Rashid a;b , Chythanya Rajanna a , Dalin Zhang a , Vincenzo Pasquale c , Laurence S. Magder a , Afsar Ali a , Stefano Dumontet c , David K.R. Karaolis a; a
Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore, MD 21201, USA b Department of Microbiology, University of Dhaka, Dhaka, Bangladesh c Dipartimento di Produzione Vegetale, Universita della Basilicata, Potenza, Italy Received 2 October 2003; received in revised form 14 November 2003; accepted 17 November 2003 First published online 6 December 2003
Abstract Vibrio cholerae, the causative agent of cholera can produce an exopolysaccharide (EPS). Some strains can also phenotypically switch from a smooth to a ‘rugose’ phenotype characterized by small wrinkled colonies, overproduction of EPS, increased biofilm formation in vitro and increased resistance to various stressful conditions. High frequency switching to the rugose phenotype is more common in epidemic strains than in non-pathogenic strains, suggesting EPS production and the rugose phenotype are important in cholera epidemiology. VpsR up-regulates Vibrio polysaccharide (VPS) genes and the synthesis of extracellular EPS (VPS). However, the function of VPS, the rugose phenotype and VpsR in pathogenesis is not well understood. We report that rugose strains of both classical and El Tor biotypes of epidemic V. cholerae are defective in the in vitro production of extracellular collagenase activity. In vivo studies in rabbit ileal loops suggest that VpsR mutants are attenuated in reactogenicity. Intestinal colonization studies in infant mice suggest that VPS production, the rugose phenotype and VpsR have a role in pathogenesis. Our results indicate that regulated VPS production is important for promoting in vivo biofilm formation and pathogenesis. Additionally, VpsR might regulate genes with roles in virulence. Rugose strains appear to be a subpopulation of cells that might act as a ‘helper’ phenotype promoting the pathogenesis of certain strains. Our studies provide new insight into the potential role of VPS, the rugose phenotype and VpsR in the pathogenesis of epidemic V. cholerae. = 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. Keywords : Vibrio cholerae; Cholera ; Virulence ; Pathogenesis; Rugose; Bio¢lm; VpsR; Exopolysaccharide ; In vivo; Colonization; Reactogenicity
1. Introduction Cholera is an epidemic diarrheal disease of humans that results in signi¢cant morbidity and mortality [1,2]. Due to the epidemic nature and severity of the disease, high transmissibility, high death-to-case ratio and ability to persist in water supplies, cholera is a public health concern. Vibrio cholerae, the bacterium causing cholera, is an organism that can persist in water supplies. Some strains have also adapted to human pathogenesis through their ability to colonize the proximal small intestine and release toxins. However, the factors involved in the emergence, patho-
* Corresponding author. Tel. : +1 (410) 706 4718; Fax : +1 (410) 706 4581. E-mail address : [email protected] (D.K.R. Karaolis).
genesis and persistence of epidemic V. cholerae are not fully understood. V. cholerae can produce an exopolysaccharide (EPS). Some strains can also switch from a smooth (EPS o¡) to a ‘rugose’ (EPS on) phenotype characterized by small wrinkled, also termed rugose, colonies [3^6]. VpsR is known to up-regulate expression of the vps genes involved in the biosynthesis of the Vibrio polysaccharide (VPS) [7]. We recently found that V. cholerae can switch at high frequency to the rugose hyperbio¢lm phenotype and that high frequency switching occurs more commonly in epidemic strains than in non-pathogenic strains suggesting that EPS and the rugose phenotype are important in the epidemiology of V. cholerae [6]. We also recently reported a molecular study on the identi¢cation of genes involved in the switch to rugose phenotype identifying novel genes not previously shown to have roles in the transition between the phenotypes [8]. We identi¢ed genes with roles in
0378-1097 / 03 / $22.00 = 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/S0378-1097(03)00879-6
FEMSLE 11329 29-12-03
Cyaan Magenta Geel Zwart
106
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
LPS synthesis, amino acid synthesis, an N-acetylmuramoyl amidase (amiB) and a novel potential regulatory gene (rocS) predicted to encode a protein containing a GGDEF domain. Proteins with GGDEF domains are found throughout the bacterial kingdom [9]. Although their function is not well understood, proteins with GGDEF domains are thought to be associated with levels of cyclic dinucleotide (c-di-GMP) signaling molecules [10^14]. In addition to defects in bio¢lm formation, VPS production and the rugose phenotype, the rocS and amiB mutants were defective in motility [8]. It appears that the switch to the rugose phenotype is a highly regulated and coordinated process and studies on these mechanisms are ongoing. It has been proposed that production of VPS, bio¢lm formation and the rugose phenotype might have an important role in the survival of V. cholerae in various environments [6,15,16]. Rugose V. cholerae strains show hyperbio¢lm formation in vitro, are highly chlorine resistant and are more resistant to UV light, extremes of acid and alkali and complement mediated lysis than are smooth strains [5,6,15,17]. Epidemic V. cholerae O1 rugose variants can cause £uid accumulation in the rabbit ileal loop model [17] and are capable of producing clinical cholera in humans [5]. However, epidemic O139 strains are typically capsulated and a previous study reported that mutation of £agellar genes in a serogroup O139 strain resulting in a rugose variant caused a defect in colonization [18]. Therefore, the role in virulence of extracellular VPS, the rugose phenotype and VpsR is still not well understood. In this study, we tested the hypothesis that VPS, the rugose phenotype and VpsR promote the pathogenesis of epidemic V. cholerae strains.
2. Materials and methods 2.1. Bacterial strains used The strains used in this study are shown in Table 1. Before use, V. cholerae strains were subcultured onto thiosulfate citrate bile salts sucrose agar (TCBS ; Oxoid) to verify purity and con¢rmed by serogrouping using polyclonal antisera (Difco).
Table 1 Bacterial strains used in this study Strain
Phenotype
Reference
N16961 smooth N16961 rugose NCTC 6585 smooth NCTC 6585 rugose DK562 DK581 DK589
wild-type El Tor biotype strain hyperbio¢lm variant of N16961 wild-type classical biotype strain hyperbio¢lm variant of NCTC 6585 N16961 Tn5: :vpsR N16961 Tn5: :vpsR N16961 Tn5: :vps(VC920)
[35] [6] [6] [6] [8] [8] [8]
FEMSLE 11329 29-12-03
2.2. Cholera toxin (CT) assay We studied the level of CT production in smooth and rugose strains of V. cholerae strain N16961 and in two independent VpsR mutants (DK562 and DK581) [8] grown under AKI conditions [19,20]. Brie£y, a fresh colony of each strain was inoculated into 15-ml tubes containing 4 ml AKI broth (1.5% bacto peptone, 0.4% yeast extract, 0.086 M NaCl and adjusted to pH 8.0 by 1 M NaHCO3 just prior to use) [19,20]. For VpsR mutants, neomycin (50 Wg ml31 ) was included in the culture medium. The cultures were incubated overnight with shaking at 200 rpm at 30‡C. The following day, 100 Wl was inoculated into a 15-ml tube containing 10 ml AKI broth (antibiotics were added where appropriate) and then incubated statically at 30‡C until OD600 = 0.3^0.4 (approximately 3 h). The cultures were then transferred into a 250-ml £ask and incubated with shaking at 30‡C and 250 rpm for 18 h and the following day 3.6 ml was removed and centrifuged at 8000 rpm for 5 min. Cell-free culture supernatants were tested for CT by GM-1 ELISA (enzyme-linked immunosorbent assay). Brie£y, Immulon 2HB (‘U’ bottom) microtiter plates (Fisher) were coated with GM1 ganglioside (100 Wl per well at 1 Wg ml31 in phosphate-bu¡ered saline (PBS)) and incubated overnight at 4‡C. Plates were washed three times with PBS, 200 Wl per well of 5% inactive fetal calf serum/PBS was added and incubated for 1 h at 37‡C and washed with PBS/ 0.05% Tween-20. Serial dilutions of CT (Sigma) were included to establish a standard curve while PBS served as a negative control. Plates were incubated for 2 h at 37‡C, washed with PBS/0.05% Tween-20 and 100 Wl per well of rabbit anti-CT antiserum (diluted 1:3200 in PBS/Tween-20 with 1% fetal calf serum) was added and the plates incubated for 1 h at 37‡C. Plates were then washed with PBS/ Tween-20 and 100 Wl of goat anti-rabbit IgG alkaline phosphatase conjugate (Kirkegaard and Perry Laboratories) diluted 1:5000 in PBS with 1% fetal calf serum was added to each well and incubated for 1 h at 37‡C. Plates were washed with PBS/Tween-20 and 100 Wl of Sigma Fast p-nitrophenyl phosphate substrate tablet diluted in sterile MiliQ water was added to each well and incubated for 45 min at 37‡C. The reaction was stopped by adding 50 Wl of 3 N NaOH and the OD405 nm determined on a plate reader (Spectra MAX250). Results were based on three independent samples for each strain (prepared on three di¡erent days), which were then tested six times. Data was analyzed by using a one-way analysis of variance or t-test. 2.3. Enzyme assay for collagenase and elastase activity V. cholerae strains were inoculated in brain heart infusion (BHI) broth and incubated overnight at 37‡C with shaking. Cultures were centrifuged and supernatants passed through 0.22-Wm ¢lters. Collagenase activity was
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
determined using the EnzChek collagenase kit (Molecular Probes) with collagen type I from bovine skin and collagen type IV from human placenta as substrates. Elastase activity was determined using the EnzChek elastase kit (Molecular Probes) with DQ elastin (soluble bovine neck ligament elastin) as the substrate. 2.4. Rabbit ileal loop studies The potential role of EPS, the rugose phenotype and vpsR in pathogenesis was examined using the rabbit ileal loop model of infection [21]. In these studies, three New Zealand white male rabbits (2 kg) (Covance Research Products) were used. Wild-type smooth and rugose variants of N16961, a vpsR mutant (DK562), and a vps mutant (DK589) were inoculated from frozen glycerol stocks onto LB agar plates. A single colony of each strain was then inoculated into 4 ml AKI broth (containing antibiotics where appropriate) and incubated at 37‡C overnight at 250 rpm. Following overnight incubation, the OD600 was adjusted to 1.2 with fresh AKI broth. From this culture, 100 Wl was transferred into into 250-ml £asks containing 10 ml AKI broth (containing antibiotics where appropriate) and incubated for approximately 3 h at 37‡C at 250 rpm until OD600 = 1.8. From this culture, 1 ml was centrifuged at 8000 rpm for 4 min, the pellet washed twice with 1.5 ml PBS and then transferred into a tube containing 10 ml PBS (OD600 V0.13^0.16). From this culture, 3 ml was removed and added to 7 ml PBS, from which the CFU ml31 was determined by plate count to be in the range of 4^8U107 ml31 . Prior to surgery, rabbits were fasted for 24 h and fed only water ad libitum. Rabbits were anesthetized, a laparotomy was performed, and the small intestine was tied o¡ into ¢ve loops (5 cm) with interloops (1 cm) between each main loop. Into each loop, using a 25-gauge needle, 1 ml was inoculated containing 107 cells of either a smooth strain, a rugose strain, vpsR, vps strains or PBS as a control. The intestine was returned to the peritoneal cavity, the incision closed, and the rabbits returned to their cage and given water ad libitum. Rabbits were sacri¢ced at 18 h, the peritoneal cavity opened and the small intestine removed. The loops were examined macroscopically, the £uid in each loop was removed and measured and an aliquot removed and plated onto appropriate antibiotic plates to determine V. cholerae counts (CFU ml31 ). Colonies ere subcultured onto TCBS agar to con¢rm their identity as V. cholerae. The £uid accumulation ratio (FAR) was determined by measuring the £uid (ml) in the loops and dividing by the length (cm) of the loop. 2.5. Infant mouse colonization studies To determine whether EPS, the rugose variant of V. cholerae N16961 and vpsR have a role in intestinal colo-
FEMSLE 11329 29-12-03
107
nization, multiple co-infection assays based on the method of Baselski and Parker [22] were performed using the suckling infant mouse model. A single colony of either the smooth strain, rugose strain or the vpsR mutant (DK581) was inoculated into a test tube containing 3 ml LB broth and incubated at 37‡C overnight at 250 rpm. The overnight cultures of each strain were diluted 1:1000 in fresh LB broth (5 Wl culture in 5 ml LB broth) and either used alone for individual assays or 1 ml was mixed in a 1:1 ratio with the another strain to be tested in coinfection assays. Each culture (individual and mixed) contained 50 Wl Evans blue food coloring. Five-day-old suckling CD-1 mice were inoculated orally using a 25-g syringe containing 50 Wl of culture (individual strains or mixtures) representing 105 CFU con¢rmed by plate count. Following incubation (5 or 24 h), the mice were sacri¢ced and their small and large intestines were removed, placed in 2 ml PBS, mechanically homogenized using a Virsonic 60 (Virtis) blender, serially diluted and plated onto LB agar supplemented with polymyxin B 50 U (and neomycin 50 Wg ml31 where appropriate) to enumerate V. cholerae CFU. Colonies were con¢rmed as V. cholerae by subculturing onto TCBS agar. The competitive index was calculated by dividing the rugose variant or vpsR mutant output number by the wild-type output number. Strains with an equivalent ability to colonize are expected to have a competition index of approximately 1.0.
3. Results and discussion Other than V. cholerae, a rugose-like phenotype with small wrinkled colonies and its associated properties has been reported in several species including Salmonella enterica Enteritidis [23], S. enterica typhimurium [24], Pseudomonas aeruginosa [25] and Enterobacter sakazakii [26]. As such, it is now becoming increasingly recognized that rugose strains might have an important role in V. cholerae and in several other species suggesting that these ‘variants’ might be more important and common than previously thought. Our recent ¢nding of conditions that promote high frequency switching to the rugose phenotype in V. cholerae and that switching appears to be more common in epidemic strains than in non-pathogenic strains suggests VPS production and the rugose phenotype is important in the epidemiology of cholera [6]. To better understand the molecular basis of switching, we reported the identi¢cation of genes involved in the switch to the rugose phenotype [8]. Although rugose variants of V. cholerae are known to be highly resistant to various stressful conditions in vitro and show hyperbio¢lm formation in vitro, the role of VPS, the rugose phenotype and VpsR in pathogenesis is not well understood. The studies reported here were performed in order to better understand the potential role of VPS, the rugose phenotype and VpsR in pathogenesis using a variety of in vitro and in vivo models of infection.
Cyaan Magenta Geel Zwart
108
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
3.1. CT production by smooth and rugose strains Although rugose strains can cause £uid accumulation in rabbit ileal loops and diarrhea in human volunteers, suggesting that they are virulent [5,17], the level of CT produced by rugose hyperbio¢lm variants compared to typical smooth cells has not been examined. In our studies, although extracellular VPS is increased in rugose strains, no di¡erence was found between smooth and rugose strains of N16961 or between wild-type N16961 and two independent VpsR mutants (DK562 and DK581) under the conditions tested. These results suggest that extracellular CT production is una¡ected in rugose strains and that VpsR does not regulate CT. 3.2. Studies on collagenase and elastase production activity The extracellular matrix (ECM) serves as a sca¡old to stabilize tissue structure. Matrix metalloproteinases such as collagenase and gelatinase that digest collagen and gelatin (denatured collagen) are potential factors in several microbial species a¡ecting the structure of basement membranes of eukaryotic cells. Studies of proteolytic enzymes in Vibrio spp. have shown that collagenase activity is a virulence factor for host infection and pathogenesis [27]. Recent studies in Vibrio vulni¢cus found that collagenase activity is rare [28,29] while in V. cholerae expression of elastase activity is common among clinical and environmental strains from di¡erent sources [30]. Previous studies have also shown that V. cholerae produces proteases that have a potential role in tissue injury and disintegration of ECM by being able to process and activate human matrix metalloproteases [31,32]. We assessed the ability of smooth and rugose epidemic V. cholerae O1 strains (including classical and El Tor strains) for their ability to produce hydrolytic extracellular proteases (collagenase and elastase activity) that like in other species might be associated with pathogenicity. We measured enzyme activity by comparing the £uorescence produced by the degradation of £uorescent collagenase and elastase substrates using a puri¢ed enzyme provided with the respective kits as a positive control. Table 2 Collagenase and elastase activity by smooth and rugose strains of V. choleraea Strain
Collagenase I
N16961 smooth N16961 rugose NCTC 6585 smooth NCTC 6585 rugose
152 4 57 0
Collagenase IV 86 0 28 20
Elastase 17 13 0 0
a
Results are based on the analysis of overnight culture supernatants grown in BHI that were passed through 0.22-Wm ¢lters. All results are expressed as percent of positive controls and blank. Positive controls with speci¢c inhibitors have been subtracted from the data. We subtracted positive controls with enzyme inhibitor in order to eliminate bias due to non-speci¢c proteases.
FEMSLE 11329 29-12-03
Our studies showed that the rugose strains of both V. cholerae N16961 (El Tor biotype) and NCTC 6585 (classical biotype) are defective in the production of both collagenase I and collagenase IV activity compared to their smooth cell variants (Table 2). In addition, we found that smooth strain N16961 produced approximately three-fold more collagenase I and collagenase IV activity than smooth strain NCTC 6585. Interestingly, while we detected similar levels of elastase activity in smooth and rugose El Tor N16961 strains, we did not detect any elastase activity in the classical NCTC 6585. These observations suggest that rugose variants of both strains (and biotypes) are attenuated in the production of collagenase activity and that classical strain NCTC 6585 is defective in elastase activity. Since V. cholerae produces many extracellular enzymes and proteases, several of which have multiple functions depending on the substrates chosen for study, it is di⁄cult to determine whether a single protein is responsible for the observed di¡erences in collagenase activity between smooth and rugose strains and in elastase activity di¡erences between the biotypes. Our results indicate that rugosity and the production of the extracellular protease (collagenase activity) are inversely linked and since several proteases have been linked with virulence, these results are consistent with a potential defect in virulence expression in rugose strains compared to smooth strains. 3.3. Studies in the rabbit ileal loop model Rugose V. cholerae O1 strains have been shown to cause £uid accumulation in the rabbit ileal loop model [17] and to be capable of producing clinical cholera in humans, showing that these variants are pathogenic [5]. However, the role of VPS, rugose strains and VpsR in pathogenesis is not well understood. In these studies, we determined whether VPS, the rugose phenotype, and vpsR have a role in cholera pathogenesis in vivo using the rabbit ileal loop model of infection. We did not see any obvious di¡erence in FAR between the smooth strain, rugose strain, vps and vpsR strains in 18-h loops (Table 3). This result is consistent with our earlier ¢nding that there is no di¡erence in CT between the strains. Plating of £uid contents from ileal loops did not reveal any obvious phenotypic switching between the smooth and rugose phenotypes. These ¢ndings suggest that if there is phenotypic switching occurring in vivo, it is at a low frequency and below the detection limit for methods used here. However, although there was no difference in the total number of V. cholerae CFU cultured from the intestinal £uid of each loop, gross examination of the 18-h loops revealed di¡erences in pathology and reactogenicity (i.e. serosal hemorrhage and necrotic areas) between strains (Fig. 1, colour ¢gure). Loops containing wild-type smooth and rugose strains showed obvious scattered regions of serosal hemorrhage (redness), necrotic
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
109
Table 3 Studies in rabbit ileal loops showing the role of the rugose strain and vpsR in pathogenesis Sample
E¡ect on rabbit ileal loops
N16961 smooth N16961 rugose DK562 (vpsR) DK589 (vps)
19.6 17.5 17.0 24
Volume (ml)
Length (cm) 9.6 9.0 8.4 10.9
areas and reactogenicity. Ileal loops inoculated with the vps mutant DK589 (defective in VPS) appeared to be slightly less reactogenic than its wild-type smooth parent. Our ¢nding that this vps mutant is slightly less reactogenic than its smooth parent is more likely to be explained if VPS production has a role in pathogenesis. However, it is possible that locus VC0920, encoding a putative glycosyl transferase, has additional roles in virulence as previous studies by us and others have shown that some genes in V. cholerae have multiple roles and are involved in several processes including LPS production and the rugose phenotype [8,33]. Most noticeably, however, the vpsR mutant (DK562) was consistently less reactogenic than its smooth wild-type parent and showed considerably less tissue damage and pathology. Although there was no di¡erence in CFU obtained from £uid, enumeration of bacterial cells adhering to intestinal tissue was not performed, making di¡erences in reactogenicity unable to be inferred from any potential di¡erences in colonization. The observation
FAR (volume/length)
Tissue damage
2.0 1.9 2.0 2.2
++ +++ 3 3/+
of even less reactogenicity in loops containing the vpsR mutant compared to loops containing its parent is consistent with ¢ndings of the vps mutant and supports a hypothesis that VPS and vpsR have a role in pathogenesis. It is important to note that the defect in switching to the rugose phenotype in the vpsR mutant could be restored in vitro by a vpsR plasmid [8]. The relative lack of reactogenicity in this VpsR mutant compared to its parent leads us to suggest that vpsR is important in the pathogenesis of cholera and not only regulates VPS biosynthesis and the rugose phenotype, but might also regulate other genes with roles in pathogenesis and might form part of a novel virulence regulon in V. cholerae. 3.4. The infant mouse colonization model Although studies have shown that the rugose phenotype promotes bio¢lm formation and cell survival in vitro, we explored whether VPS, rugose strains and VpsR have a
Fig. 1. Colour ¢gure showing gross examination of rabbit ileal loops inoculated with V. cholerae. Note the reduction in pathology and reactogenicity in the loop containing the vspR mutant DK562 compared to the wild-type smooth strain N16961. These results were consistent in three rabbits following 18 h incubation.
FEMSLE 11329 29-12-03
Cyaan Magenta Geel Zwart
110
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
Table 4 Colony counts and co-infection studies of V. cholerae in suckling infant mice Strains
Smooth Rugose VpsR Rugose/smooth VpsR/smooth VpsR/rugose
5h
24 h
(CFU ml)
CI
(CFU ml)
CI
ND ND ND 4.2U104 /2.2U105 9.7U103 /1.8U104 5.4U105 /4.3U104
ND ND ND 0.19 (n = 7) 0.53 (n = 8) 12.55 (n = 8)
2.28U107 9.54U106 3.70U106 5.1U106 /2.3 U107 2.1U106 /2.2 U106 1.1U107 /1.1 U106
ND (n = 10) ND (n = 14) ND (n = 17) 0.22 (n = 27) 0.95 (n = 19) 10.0 (n = 22)
CI = competitive index; n = number of mice in group; ND = not done.
role in vivo by studying intestinal colonization using the suckling infant mouse model. Although we found no obvious di¡erence in the level of TcpA production between smooth and rugose strains of V. cholerae O1 strain N16961 (data not shown), we hypothesized that there might be di¡erences in the level of colonization in vivo between the strains given that rugose strains show hyperbio¢lm formation in vitro [6,15,16]. In these studies, we compared the ability of smooth strains, rugose strains and a vpsR mutant (DK581) to colonize 5-day-old infant mice. Strains were inoculated orally either alone or in standard co-infection assays. While plating of the individual initial inoculums before inoculation into mice showed that there was no di¡erence in the average counts between the strains (smooth = 3.5 U105 ml31 ; rugose = 4.8U105 ml31 ; vpsR = 4.0U105 ml31 ), and multiple in vitro co-infection assays showed no di¡erence between strains following 5 h and 24 h coinfection (data not shown), we observed signi¢cant di¡erences in the intestinal colonization ability between the strains (Table 4 and Fig. 2A,B).
(P = 0.0004; Kruskal^Wallis test) (Table 4 and Fig. 2B). The results showed that smooth strains colonize V10-fold better than rugose strains. We also found that a vpsR mutant (VPS-negative and unable to switch to the rugose phenotype) is signi¢cantly attenuated in colonization compared to both the wild-type smooth and rugose strain. These results suggest that excessive (rugose) VPS production might inhibit colonizing ability. However, the results
3.4.1. Intestinal colonization studies (5 h) No obvious ( 6 two-fold) di¡erence was found between the smooth and vpsR strain in 5 h co-infection assays. However, the smooth strain colonized ¢ve-fold better than the rugose strain in co-infection studies. The vpsR mutant colonized 12-fold better than the rugose strain in co-infection studies. Interestingly, the colonization rate for both the smooth strain and the vpsR mutant was 12- and 55-fold higher, respectively, when these strains were coinfected with the rugose strain compared to their colonization rate in the smooth/vpsR co-infection assay. These results suggest that VPS production and the rugose phenotype has an e¡ect on the intestinal colonization of V. cholerae. 3.4.2. Intestinal colonization studies (24 h) To determine the level of colonization between the strains during later stages of infection, we performed multiple 24 h individual and co-infection assays. Initially, we determined the colonization levels of individual strains. We found a statistically signi¢cant di¡erence in the level of colonization between all strains when given alone
FEMSLE 11329 29-12-03
Fig. 2. Infant mouse colonization studies of V. cholerae N16961 smooth strain, rugose, and vpsR mutant after 24 h. A: Colonization rates after 5 h incubation. B: Colonization rates after 24 h incubation. Colony counts (CFU ml31 ) represent the mean values obtained where n represents the number of mice used in that study.
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
also suggest that the inability to produce VPS, such as in the vpsR mutant, results in an even greater defect in the colonization rate. These infant mice intestinal colonization studies support our ileal loop data whereby a vpsR mutant appears to be less reactogenic, possibly due to a defect in colonization, and indicates that VPS, the rugose phenotype and vpsR have a role in pathogenesis. We then performed 24 h co-infection assays with the smooth strain, rugose strain and vpsR mutant (Table 4 and Fig. 2B). We found no di¡erence in colonization rates between a smooth and vpsR mutant (P = 0.36; Wilcoxon signed rank test). Consistent with the 5 h co-infection assays, we found that the smooth strain has a signi¢cantly ¢ve-fold greater colonization rate than the rugose strain (P = 0.0034 ; Wilcoxon signed rank test). These results are consistent with Watnick et al. [18], who reported that a serogroup O139 £agellar mutant exhibiting the rugose phenotype shows a decrease in colonization compared to its wild-type smooth parent. Also, consistent with 5 h coinfection assays, the vpsR mutant colonized 10-fold better than rugose strains in co-infection assays (P 6 0.0009; Wilcoxon signed rank test). Interestingly, although the reason is not well understood, we found a decrease in the colonization rates of the smooth and rugose strain when co-infected with the VpsR mutant compared to when these strains were inoculated alone (P = 0.006 and P = 0.012, respectively; Kruskal^Wallis test). Importantly, and again consistent with 5 h assays, the smooth strain had 10-fold better colonization ability when it was co-infected with the rugose strain compared to when it was coinfected with the vpsR mutant. Likewise, the vpsR mutant colonized ¢ve-fold better in co-infection with the rugose strain compared to when it was inoculated alone (P = 0.027; Wilcoxon test). We suggest that these ¢ndings are consistent with a hypothesis that the ability to produce VPS in a carefully regulated manner is required for e⁄cient colonization and that the rugose phenotype, at least in part as a result of VPS production, might promote the colonization of certain strains. In these studies, our hypothesis was that VPS, the rugose phenotype and VpsR have a role in virulence. Our interpretation of the data is that the results cannot be explained only by the fact that the rugose strain, possibly as a result of VPS production masking a colonization factor, is defective in colonization compared to smooth strains. If this were the case, the vpsR mutant (defective in VPS) would be expected to colonize as well as wild-type smooth strains when each of these strains is inoculated alone. In addition, we would not expect to ¢nd the colonization rate of a vpsR mutant to be increased V¢ve-fold in the presence of rugose strains compared to when the vpsR mutant is co-infected with the smooth strain. Rather, our data strongly indicate that VPS production is actually required for e⁄cient intestinal colonization. Given the observed e¡ects in pathogenesis between VPS overproduction (rugose) and VPS underproduction (vpsR mutant),
FEMSLE 11329 29-12-03
111
we predict that there should be some regulatory mechanism that controls the timing and quantity of VPS produced in vivo. As such, we are performing further studies on the regulatory system and signal molecules controlling VPS production and bio¢lm formation. In a model presented here, controlled VPS production appears to be important for intestinal colonization and rugose strains appear to promote the virulence of certain ‘smooth’ strains. In this model, smooth strains can produce VPS in a regulated manner and are expected to colonize better than rugose strains that produce excessive VPS production. If some VPS is required for e⁄cient colonization, a vpsR mutant (defective in VPS production) would be expected to colonize less than wild-type smooth and rugose strains. Consistent with this model are our ¢ndings that the wild-type smooth strains co-infected with the rugose strain colonize better than when they are co-infected with the vpsR mutant. In addition, the vpsR mutant also colonizes better in the presence of rugose cells than in the presence of wild-type smooth strains or when inoculated alone. Switching from the smooth to rugose phenotype in vivo is not implicit in this model, however, it is possible that some switching occurs in vivo at a low frequency and this event might further promote the colonization of certain strains. Therefore, although the rugose strain colonizes less than the wild-type smooth strain but better than a vpsR mutant, our results not only suggest that VPS is an important factor for colonization but also imply that rugose strains might under some conditions act as a ‘helper’ phenotype for certain smooth strains in their colonization of the intestine.
4. Conclusions It has been shown that VPS production and the switch to the rugose phenotype is linked to the type II general extracellular protein secretion pathway which is also involved in secretion of CT and hemolysin [34]. Although the regulatory pathway involved is not clear, our ¢ndings suggest that production of the potential virulence-associated collagenase and elastase activity is inversely linked to the production of VPS and the rugose phenotype. Based on these in vitro proteolytic assays, rugose strains appear to be less pathogenic in this property than wild-type smooth strains. VpsR belongs to a family of proteins with homology to signal transduction two-component regulatory systems [7]. These regulatory systems are important in the sensing of and responding to di¡erent environments. Our results strongly suggest that VpsR has a role in the pathogenesis of cholera and that in addition to regulating vps genes and VPS production, VpsR might regulate additional genes with roles in virulence. The factors involved in the emergence, pathogenesis and persistence of epidemic V. cholerae are not fully under-
Cyaan Magenta Geel Zwart
112
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113
stood but are being studied by us. Until now, the nature of the biological role of VPS and the rugose phenotype in V. cholerae has not been well identi¢ed but we and others have proposed some role in environmental survival. While it is possible that VPS and the rugose phenotype have a role in the environment, we propose a new model in which VPS production and the rugose phenotype might have a role in pathogenesis. In this model, careful timing and control of VPS production is required for e⁄cient intestinal colonization and rugose strains might potentially act as a ‘helper’ phenotype in promoting the pathogenesis of certain strains. Our ¢ndings provide a new understanding of the role of VPS, the rugose phenotype and VpsR in epidemic V. cholerae and might be relevant to other species.
Acknowledgements M.H.R. is part of a joint program between the University of Dhaka and the University of Maryland and wishes to acknowledge the assistance of Chowdhury Ahsan at the University of Dhaka. We thank Guard-Petter for useful discussions. This work was supported by NIH grant AI45637. D.K.R.K. is a recipient of a Burroughs Wellcome Fund Career Award in the Biomedical Sciences.
References [1] Pollitzer, R. (1959) Cholera. Monogr. Ser. 43. World Health Organization, Geneva. [2] Kaper, J.B., Morris Jr., J.G. and Levine, M.M. (1995) Cholera. Clin. Microbiol. Rev. 8, 48^86. [3] White, P.B. (1938) The rugose variant of vibrios. J. Pathol. Bacteriol. 46, 1^6. [4] White, P.B. (1940) The characteristic hapten and antigen of rugose races of cholera and El Tor vibrios. J. Pathol. Bacteriol. 50, 160^164. [5] Morris Jr., J.G., Sztein, M.B., Rice, E.W., Nataro, J.P., Losonsky, G.A., Panigrahi, P., Tacket, C.O. and Johnson, J.A. (1996) Vibrio cholerae O1 can assume a chlorine-resistant rugose survival form that is virulent for humans. J. Infect. Dis. 174, 1364^1368. [6] Ali, A., Rashid, M.H. and Karaolis, D.K.R. (2002) High-frequency rugose exopolysaccharide production in Vibrio cholerae. Appl. Environ. Microbiol. 68, 5773^5778. [7] Yildiz, F.H., Dolganov, N.A. and Schoolnik, G.K. (2001) VpsR, a member of the response regulators of the two-component regulatory systems, is required for expression of vps biosynthesis genes and EPSETr -associated phenotypes in Vibrio cholerae O1 El Tor. J. Bacteriol. 183, 1716^1726. [8] Rashid, M.H., Rajanna, C., Ali, A. and Karaolis, D.K.R. (2003) Identi¢cation of genes involved in the switch between the smooth and rugose phenotypes of Vibrio cholerae. FEMS Microbiol. Letts. 227, 113^119. [9] Bateman, A., Birney, E., Durbin, R., Eddy, S.R., Howe, K.L. and Sonnhammer, E.L. (2000) The Pfam protein families database. Nucleic Acids Res. 28, 263^266. [10] Ausmees, N., Mayer, R., Weinhouse, H., Volman, G., Amikam, D., Benziman, M. and Lindberg, M. (2001) Genetic data indicate that proteins containing the GGDEF domain possess diguanylate cyclase activity. FEMS Microbiol. Lett. 204, 163^167.
FEMSLE 11329 29-12-03
[11] Amikam, D. and Benziman, M. (1989) Cyclic diguanylic acid and cellulose synthesis in Agrobacterium tumefaciens. J. Bacteriol. 171, 6649^6655. [12] Ross, P. et al. (1990) The cyclic diguanylic acid regulatory system of cellulose synthesis in Acetobacter xylinum. Chemical synthesis and biological activity of cyclic nucleotide dimer, trimer, and phosphothioate derivatives. J. Biol. Chem. 265, 18933^18943. [13] Ross, P., Mayer, R. and Benziman, M. (1991) Cellulose biosynthesis and function in bacteria. Microbiol. Rev. 55, 35^58. [14] Mayer, R. et al. (1991) Polypeptide composition of bacterial cyclic diguanylic acid-dependent cellulose synthase and the occurrence of immunologically crossreacting proteins in higher plants. Proc. Natl. Acad. Sci. USA 88, 5472^5476. [15] Yildiz, F.H. and Schoolnik, G.K. (1999) Vibrio cholerae O1 El Tor: identi¢cation of a gene cluster required for the rugose colony type, exopolysaccharide production, chlorine resistance, and bio¢lm formation. Proc. Natl. Acad. Sci. USA 96, 4028^4033. [16] Watnick, P.I. and Kolter, R. (1999) Steps in the development of a Vibrio cholerae El Tor bio¢lm. Mol. Microbiol. 34, 586^595. [17] Rice, E.W. et al. (1993) Vibrio cholerae O1 can assume a ‘rugose’ survival form that resists killing by chlorine, yet retains virulence. Int. J. Environ. Health Res. 3, 89^98. [18] Watnick, P.I., Lauriano, C.M., Klose, K.E., Croal, L. and Kolter, R. (2001) The absence of a £agellum leads to altered colony morphology, bio¢lm development and virulence in Vibrio cholerae O139. Mol. Microbiol. 39, 223^235. [19] Iwanaga, M. and Yamamoto, K. (1985) New medium for the production of cholera toxin by Vibrio cholerae O1 biotype El Tor. J. Clin. Microbiol. 22, 405^408. [20] Iwanaga, M., Yamamoto, K., Higa, N., Ichinose, Y., Nakasone, N. and Tanabe, M. (1986) Culture conditions for stimulating cholera toxin production by Vibrio cholerae O1 El Tor. Microbiol. Immunol. 30, 1075^1083. [21] De, S.N. and Chatterjee, D.N. (1953) An experimental study of the mechanism of action of V. cholerae on the intestinal mucous membrane. J. Pathog. Bact. 66, 559^562. [22] Baselski, V.S. and Parker, C.D. (1978) Intestinal distribution of Vibrio cholerae in orally infected infant mice: kinetics of recovery of radiolabel and viable cells. Infect. Immun. 21, 518^525. [23] Petter, J.G. (1993) Detection of two smooth colony phenotypes in a Salmonella enteritidis isolate which vary in their ability to contaminate eggs. Appl. Environ. Microbiol. 59, 2884^2890. [24] Anriany, Y.A., Weiner, R.M., Johnson, J.A., De Rezende, C.E. and Joseph, S.W. (2001) Salmonella enterica serovar Typhimurium DT104 displays a rugose phenotype. Appl. Environ. Microbiol. 67, 4048^ 4056. [25] D’Argenio, D.A., Calfee, M.W., Rainey, P.B. and Pesci, E.C. (2002) Autolysis and autoaggregation in Pseudomonas aeruginosa colony morphology mutants. J. Bacteriol. 184, 6481^6489. [26] Farmer, J.J., Asbury III, M.A., Hickman, F.W., Brenner, D.J. and Group, E.S. (1980) Enterobacter sakazakii: a new species of ‘Enterobacteriaceae’ isolated from clinical specimens. Int. J. Syst. Bacteriol. 30, 569^584. [27] Kim, S.K., Yang, J.Y. and Cha, J. (2002) Cloning and sequence analysis of a novel metalloprotease gene from Vibrio parahaemolyticus 04. Gene 283, 277^286. [28] Moreno, M.L. and Landgraf, M. (1998) Virulence factors and pathogenicity of Vibrio vulni¢cus strains isolated from seafood. J. Appl. Microbiol. 84, 747^751. [29] Rodrigues, D.P., Ribeiro, R.V. and Hofer, E. (1992) Analysis of some virulence factors of Vibrio vulni¢cus isolated from Rio de Janeiro, Brazil. Epidemiol. Infect. 108, 463^467. [30] Guhathakurta, B., Sasmal, D., Pal, S., Chakraborty, S., Nair, G.B. and Datta, A. (1999) Comparative analysis of cytotoxin, hemolysin, hemagglutinin and exocellular enzymes among clinical and environmental isolates of Vibrio cholerae O139 and non-O1, non-O139. FEMS Microbiol. Lett. 179, 401^407.
Cyaan Magenta Geel Zwart
M.H. Rashid et al. / FEMS Microbiology Letters 230 (2004) 105^113 [31] Okamoto, T. et al. (1997) Activation of human matrix metalloproteinases by various bacterial proteinases. J. Biol. Chem. 272, 6059^ 6066. [32] Maeda, H., Okamoto, T. and Akaike, T. (1998) Human matrix metalloprotease activation by insults of bacterial infection involving proteases and free radicals. Biol. Chem. 379, 193^200. [33] Nesper, J., Lauriano, C.M., Klose, K.E., Kapfhammer, D., KraiM, A. and Reidl, J. (2001) Characterization of Vibrio cholerae O1 El Tor galU and galE mutants: in£uence on lipopolysaccharide structure, colonization, and bio¢lm formation. Infect. Immun. 69, 435^445.
FEMSLE 11329 29-12-03
113
[34] Ali, A., Johnson, J.A., Franco, A.A., Metzger, D.J., Connell, T.D., Morris, J.G.J. and Sozhamannan, S. (2000) Mutations in the extracellular protein secretion pathway genes (eps) interfere with rugose polysaccharide production in and motility of Vibrio cholerae. Infect. Immun. 68, 1967^1974. [35] Levine, M.M., Black, R.E., Clements, M.L., Nalin, D.R., Cisneros, L. and Finkelstein, R.A. (1981) In: Acute Enteric Infections in Children. New Prospects for Treatment and Prevention (Holme, T., Holmgren, J., Merson, M.H. and Mollby, R., Eds.), pp. 443^459. Elsevier, Amsterdam.
Cyaan Magenta Geel Zwart