The facultative human pathogen Vibrio cholerae can be isolated from estuarine and aquatic environments. V. cholerae is well recognized and extensively studied as the causative agent of the human intestinal disease cholera. In former centuries cholera was a permanent threat even to the highly developed populations of Europe, North America, and the northern part of Asia. Today, cholera still remains a burden mainly for underdeveloped countries, which cannot afford to establish or to maintain necessary hygienic and medical facilities. Especially in these environments, cholera is responsible for significant mortality and economic damage. During the last three decades, intensive research has been undertaken to unravel the virulence properties and to study the epidemiology of this significant human pathogen. More recently, researchers have been elucidating the environmental lifestyle of V. cholerae. This review provides an overview of the current knowledge of both the host- and environment-specific physiological attributes of V. cholerae.
1 Historical perspective
Accounts of a cholera-like disease go back to the times of Hippocrates and Buddha and perhaps even earlier . However, the modern history of cholera began in 1817. At that time an epidemic outbreak was reported in India, which subsequently spread across the Indian continent and was then defined as the first pandemic outbreak of cholera disease in southeast Asia. During the 19th century six cholera pandemics took place, ending in 1923 and affecting mostly the continents located in the southern hemisphere, as well as North America and Europe [1, 2]. In 1961, the seventh pandemic began in Indonesia, then spread to the Indian subcontinent and the Middle East, then moved on to Africa in the 1970s and finally reached South America in the early 1990s [3–6].
The etiologic agents responsible for cholera and many other infectious diseases (e.g. tuberculosis, anthrax, diphtheria, tetanus and typhoid fever) were identified during the 19th century. In 1883 Robert Koch demonstrated that cholera is produced by a bacterium that he referred to as ‘comma(-shaped) bacteria’ , later designated Vibrio cholerae.
The early pandemics preceded Koch's discovery of the infectious cause of cholera, and thus have never been associated with a specific strain variant. The fifth and sixth pandemics, however, have been shown to have been caused by the V. cholerae serogroup O1 of biotype ‘classical’ (see below), while the seventh pandemic was caused by serogroup O1 biotype ‘El Tor’ . In 1992, a serogroup conversion event(s) led to the emergence of the new V. cholerae serogroup O139, resulting in a large epidemic in Bangladesh and India [8–10]. Since then O139 strains are endemically found in these areas and cases have been reported all over south-east Asia. This epidemic has already been described as the eighth pandemic ; however, O1 El Tor strains also still persist in these areas.
As a result of continuing scientific and medical effort directed toward combating cholera, major improvements in medical treatment as well as a better understanding of the molecular processes involved in the virulence of V. cholerae have been achieved. However, despite all efforts, new variants of this pathogen are obviously still evolving that are able to circumvent established immunity and to resist clinical and hygienic prophylactics.
2 V. cholerae and its environment
V. cholerae is a motile, Gram-negative curved rod that belongs to the family Vibrionaceae. About 200 recognized O serogroups are known, however only serogroup O1 and the newly emerged O139 have been associated with severe disease and cholera pandemics. In contrast, intestinal and/or extraintestinal infections with non-O1 and -O139 serogroups or non-toxigenic O1 strains are rarely found and seem to be of little clinical significance [11–14].
It has been shown that V. cholerae, including serogroup O1 and O139 pandemic strains, exist as natural inhabitants of aquatic ecosystems [15–18], thus making them facultative human pathogens. Within the marine environment they attach to surfaces provided by plants, filamentous green algae, copepods (zooplankton), crustaceans, and insects [16, 19]. Interestingly, there is a correlation of cholera outbreaks and the seasonal occurrence of algal blooms [16, 18, 20]; however, there is no direct evidence that such events lead to an enrichment of toxigenic V. cholerae strains responsible for cholera epidemics. Non-O1 and non-O139 strains are more frequently isolated from rivers and estuarine areas than O1 and O139 strains, and interestingly most environmental O1 strains are non-toxigenic . This led to the hypothesis that environmental strains that acquire the respective virulence genes, which are harbored on genetic mobile elements (see below), may then be enriched by intestinal conditions [22, 23]. Recent evidence  demonstrates that non-toxigenic environmental strains can be converted by phage transduction with cholera toxin (CT)-encoding phage CTXφ, and this event could conceivably also take place in the gastrointestinal environment, yielding new detectable toxigenic strains.
V. cholerae O1 is found associated with marine organisms, and it has been shown that these strains can bind to chitin , and in addition can then acquire acid tolerance . Recently, one factor involved in adherence to the chitin of zooplankton was identified as the mannose-sensitive hemagglutinin (MSHA) type IV pili. These pili are expressed and assembled by O1 El Tor and O139 strains; in contrast O1 classical strains carry the pili (msh) genes but do not assemble functional MSHA pili  (Fig. 1). Additional uncharacterized chitin binding proteins have been reported to be present in V. cholerae strains , and this has also been shown for other Vibrio species like V. alginolyticus. V. cholerae also possess secreted chitinase enzyme(s) which are probably needed to utilize chitin (homopolymer of N-acetylglucosamine) as a carbon source [24, 29]; five hypothetical chi gene products were identified in the genome sequence ). These findings are suggestive of a close association of V. cholerae with chitin structures in the environment (e.g. zooplankton such as copepods).
MSHA-mediated attachment of V. cholerae to zooplankton. A: Shown is a light micrograph of an adult Daphnia pulex with a length of about 2.5 mm. The arrow indicates the position of the antenna structure, an exoskeleton structure which is shown at higher magnification in panels B and C. In panel B D. pulex was colonized by V. cholerae O139 (GFP-expressing); the attached bacteria are visualized as bright green spots. The D. pulex in panel C were incubated with GFP-expressing V. cholerae O139 mshA mutants, indicating a role for MSHA in zooplankton colonization. (Reprinted with permission from R. Taylor ).
A recent report revealed that environmental non-O1 and non-O139 V. cholerae strains can also be isolated from chironomid egg masses (Chironomus sp., Diptera). These egg masses were collected from waste stabilization ponds and it was found that they also serve as a favorable rich nutrient niche for V. cholerae strains .
3 Mechanisms involved in environmental survival
In general it is assumed that association of V. cholerae with zooplankton or egg masses may provide some protection from the relatively harsh environmental conditions faced in the aquatic habitats of planktonic bacterial cells. A variety of microorganisms, including V. cholerae, are able to build up large multicellular structures on solid surfaces known as biofilms. The biofilm provides a microenvironment that favors survival and persistence due to increased resistance to various stresses (e.g. chlorine, antibiotics).
It has been demonstrated that V. cholerae O1 El Tor and O139 are both able to form a three-dimensional biofilm an abiotic surfaces [32–34] (Fig. 2). Biofilm formation is likely to be important for the life cycle of V. cholerae, facilitating environmental persistence within natural aquatic habitats during interepidemic periods.
Steps in V. cholerae biofilm formation. The planktonic bacterium was visualized by transmission electron microscopy (bar=1 μm), the attached cells and microcolony were visualized by scanning electron microscopy (bar=2 μm), and the biofilm micrograph represents a vertical section through a 20-μm biofilm taken by confocal scanning laser microscopy (bar=10 μm). (Reprinted with permission from P. Watnick ).
Biofilm formation by both O1 El Tor and O139 strains is dependent on the expression of an exopolysaccharide (VPS) [32, 34, 35]. Strains containing defects in one of the vps genes necessary for exopolysaccharide synthesis can colonize an abiotic surface but fail to make a three-dimensional biofilm, indicating that the exopolysaccharide is used as the ‘glue’ to build a mature biofilm [33, 34]. Expression of the VPS exopolysaccharide causes a rugose colony phenotype, and provides enhanced chlorine resistance and also phage resistance [32, 36, 37]. Interestingly, expression of the VPS exopolysaccharide inhibits intestinal colonization in the infant mouse cholera model, suggesting that the expression of a factor that enhances environmental persistence actually decreases virulence . Rugose colonies (i.e. expressing the VPS exopolysaccharide) appear in low numbers as phase variants, and their appearance increases during carbon starvation [32, 38]. The lack of flagellar synthesis causes high-level VPS exopolysaccharide expression (at least in some strains, e.g. O139), indicating that flagellar synthesis is coupled to VPS expression. This effect is specifically due to the lack of a flagellum, rather than a lack of motility, suggesting that the loss of the flagellum may be a developmental cue during biofilm formation .
The VPS exopolysaccharide constituents have been identified, and galactose was identified among the carbohydrate residues [32, 35]. Recent attempts to investigate the Gal pathway in V. cholerae prompted us to generate and characterize a galU mutant, encoding the UDP-glucose-4-epimerase. We found that the galU mutant has lost the ability to produce the VPS exopolysaccharide, and is defective for biofilm development . These results are consistent with a requirement for the synthesis of UDP-galactose via UDP-glucose for the biosynthesis of the VPS exopolysaccharide.
In addition to the VPS exopolysaccharide, V. cholerae O1 El Tor requires the MSHA type IV pilus to form a normal biofilm , but the O139 strain does not . The V. cholerae O1 El Tor also requires flagellar motility for biofilm formation, while non-flagellated O139 strains can eventually form three-dimensional biofilms, apparently because of high level VPS exopolysaccharide expression [33, 34]. Watnick et al. have postulated a three-step model in V. cholerae biofilm formation (see Fig. 2). Flagellar-driven motility is important for the initial attachment step, followed by microcolony formation assisted by the MSHA pili. The final step is VPS exopolysaccharide expression, which allows the three-dimensional architecture of the mature biofilm. A non-flagellated O139 strain bypasses the need for motility presumably because it already expresses VPS exopolysaccharide, which allows for microcolonies to form in solution that eventually settle onto the solid surface. It is not clear why the O139 strain does not require MSHA pili, but it may perhaps be due to the fact that O139 strains are encapsulated, unlike the O1 El Tor (see below).
In addition to biofilm formation, V. cholerae possess the remarkable ability of being able to switch into a viable, but non-culturable (VNC) state in response to nutrient deprivation [39, 40]. Such a dormant state could serve as a survival strategy, which can be resuscitated upon receiving the appropriate signals, e.g. extreme change of environmental conditions such as the transition from the aquatic system upon ingestion into the human intestinal environment [6, 39]. During interepidemic periods, epidemic V. cholerae strains have been postulated to exist in the VNC form, with unknown environmental conditions serving to resuscitate these cells back to free-living, virulent organisms [18, 41]. If this ecological niche serves as a common reservoir for toxigenic V. cholerae then such strains should be highly related. In contrast to this assumption, isolated epidemic strains are quite different, based on ribotyping and different phenotypic and genetic properties (as reviewed recently ), hence a clonal origin of a VNC reservoir seems less likely. In general, the existence of the VNC state is in debate, since there is no direct conclusive information about the underlying molecular processes or the genetic factors involved.
4 Transmission and identification of toxigenic V. cholerae
Poor sanitation practices in highly populated areas harboring endemic toxigenic strains are the source of occasional outbreaks due to contamination of drinking water and/or improper food preparation. Contaminated water with free-living V. cholerae cells are probably the main origin of epidemics, followed to a lesser extent by contaminated food, especially seafood products like oysters, crabs, and shellfish (for review see [42, 43]).
For example, a massive outbreak of O1 El Tor V. cholerae-mediated cholera took place in the South American continent and caused over 750 000 cases and 6500 deaths . The exact origin of the epidemic El Tor strain remains unknown, but the epidemic originated on the west coast of Peru. The spread of the epidemic strain was mainly caused by contaminated drinking water supplies, followed by fecal contamination of the marine environment, which led to contamination of seafood .
There have been reports of bottled water contamination with toxigenic V. cholerae. The evidence suggested that carbonated water was safer than non-carbonated water, due to the organism's poor acid tolerance . Subsequently, pH neutral food was found to serve as a good carrier for toxigenic V. cholerae strains . In fact, it was shown that toxigenic O1 V. cholerae can survive refrigeration and freezing in food supplies shipped internationally, therefore an epidemic strain may travel far from its original endemic location .
Isolation of V. cholerae from the extraintestinal environment or stool samples of patients can be achieved by taking advantage of the organism's fast generation time, and tolerance to alkaline conditions (pH 8.5). Thus, isolation generally includes growth in alkaline peptone water medium, which provides a competitive advantage to Vibrionaceae. Also frequently used are blood -and chocolate-based or Müller–Hinton and MacConkey agar media. Subsequently, a selective growth medium and identification scheme routinely used for Vibrionaceae is growth on thiosulfate–citrate–bile salts–sucrose medium. This medium is highly selective for Vibrio spp., and the ability of the organism to ferment the sucrose in this medium (e.g. V. cholerae) results in yellow colonies. Additional testing of oxidase activity will lead to the identification of Vibrio spp. (for additional information see ). In contrast to the halophilicity of other Vibrio species, V. cholerae is one of the few Vibrionaceae that can grow in the absence of NaCl, so this is a further defining characteristic.
Once a strain has been identified as V. cholerae by these and several additional tests, typing of V. cholerae is achieved by slide agglutination with O1 and/or O139 antisera. This differentiates the strain as either O1, O139, or non-O1/O139, based on the structure of the lipopolysaccharide and capsule for O139 (Fig. 4). Mostly the serogroups O1 and O139 are associated with epidemic disease and further classification is necessary. In case of the O1 serogroup, three different serotypes or antigen variations were recognized termed ‘Inaba’, ‘Ogawa’, and a minor serotype ‘Hikojima’ . In addition, O1 strains can be classified in two different biotypes, called O1 ‘El Tor’ and O1 ‘classical’ determined by physiological properties, such as polymyxin B resistance, number of CT-encoding genes, hemolysin activity and the presence of the mannose-sensitive hemagglutinin (for review see ).
LPS and capsule structure of serogroup O1 and O139 V. cholerae. The deduced structure of V. cholerae serogroup O1 (left) and O139 (right); shown are the different O antigen and capsule structures, and the similar core-OS and lipid A structures [139, 151] (carbohydrate and fatty acid compositions are indicated). The few chemical modifications found in O139 but not O1 core-OS are indicated with ‘?’. The O-antigen of V. cholerae O1 consists of a homopolymer of approximately 18 perosamine (4-NH2-4,6-dideoxymannose) residues consisting of dideoxy-phosphomannose substituted with tetronate . The O-antigen biosynthesis is catalyzed by the gene products encoded by the rfb gene cluster (see recent review ). V. cholerae O139 O-antigen has a different carbohydrate composition, which is also presented as a capsule [138, 139].
5 V. cholerae infections and the outcome of cholera
Infection normally starts with the oral ingestion of food or water contaminated with V. cholerae (Fig. 3). Subsequently the bacteria must pass through and survive the gastric acid barrier of the stomach, then penetrate the mucus lining that coats the intestinal epithelia. In human volunteer studies, the infectious dose was determined to be fairly high, and varied depending on the inocula conditions (ranging from 106 to 1011 colony-forming units) . This high dose is probably needed because of the acid sensitivity of V. cholerae cells, which are exposed to low pH in the gastric compartment . The surviving bacteria adhere to and colonize the intestinal epithelial cells, eventually producing the CT and causing cholera symptoms . The primary site of V. cholerae colonization is the small intestine. During the transition from the aquatic environment to the human body the bacterial cells are exposed to a series of changes, such as temperature, acidity, and osmolarity. They also must survive in the intestinal environment, which contains growth inhibitory substances like bile salts and organic acids, and also factors of the innate immune system, such as complement secreted by intestinal epithelial cells  and defensins produced by Paneth cells . Therefore, V. cholerae has developed the ability to survive, colonize, and express virulence factors such as CT in spite of, and possibly in response to, harsh environmental conditions.
Infection cycle of V. cholerae. Shown is a simplified depiction of the infection cycle of the cholera-causing bacteria V. cholerae.
After V. cholerae secretes the CT, the symptoms of cholera will occur. Cholera is reflected by massive watery diarrhea of such prodigious volume that hypotensive shock and death can occur within 12 h of the first symptoms . The rate of body fluid loss is greatest in the jejunum with up to 20 l per day in adults . This tremendous water loss is mainly mediated by the action of the CT (for review see ). CT is a hetero-oligomer consisting of one A subunit in association with five B subunits. The B subunit is responsible for specific binding to the GM1 ganglioside receptor of epithelial cells [59, 60]. Upon binding, the A subunit is translocated into the host cell cytosol, where it is activated by thiol dependent reduction, probably by thiol:protein disulfide oxidoreductases . Only the resulting nicked A1 subunit possesses an ADP-ribosylating activity that targets the host cell G-protein Gsα. ADP-ribosylated Gsα in turn permanently activates adenylate cyclase activity, leading to increased levels of intracellular cAMP. cAMP inhibits active sodium chloride absorption and increases chloride and bicarbonate secretion (for review see [58, 62]). This results in passive water loss, leading to a marked decrease in intravascular volume, hypotension and hypoperfusion of critical organs, and in severe cases death ensues with a high mortality rate (>20%) (for review see ).
6 Dissecting the molecular aspects of V. cholerae infections
Due to extensive molecular and genetic analyses, the important virulence factors and the virulence-responsive regulatory systems of V. cholerae have been elucidated [62–64]. Although no animal model replicates all aspects of V. cholerae infection in humans, several animal models have been extremely useful in the study of cholera pathogenesis. Two particular animal models have been widely used. The infant mouse colonization assay has accurately predicted the colonization behavior of V. cholerae strains in human volunteer studies, while the rabbit ileal loop model has been useful to measure the effect of CT in vivo, namely fluid accumulation within the intestine.
Recently the complete genome sequence of the two chromosomes of V. cholerae O1 El Tor was made available to the scientific community , which will help to facilitate the further dissection of cholera pathogenesis. Cholera is the result of a multifactorial process, consisting of V. cholerae colonization, coordinate expression of virulence factors, and toxin action. V. cholerae also acquire virulence genes in a step-by-step process, in which a pathogenicity island and a temperate bacteriophage play important roles. In the following paragraphs, an overview will be provided of recent discoveries and an outlook toward future developments.
6.1 Colonization factors
Several V. cholerae gene products have been shown to be important for colonization of the small intestine, primarily in the infant mouse and the adult rabbit model of cholera. These include the toxin-coregulated pili (TCP) , accessory colonization factors (ACFs) , mannose–fucose-resistant hemagglutinin , TolC , regulatory proteins (e.g. ToxR/ToxS, TcpP/TcpH, and ToxT) [65, 69, 70], outer membrane porins , biotin and purine biosynthetic genes , an iron-regulated OMP protein IrgA, , the O-antigen of the lipopolysaccharide (LPS) [74–76] and attributes of the LPS core region .
The TCP, a type IV pilus, is considered to be the most important colonization factor because it has been demonstrated to be required for intestinal colonization both in animal models and in human volunteer studies [65, 77]. The TCP is thought to be a polymer composed of a single structural subunit, TcpA, and also serves as the receptor for the CTXφ bacteriophage (see below; [23, 78]). The genes required for TCP synthesis, including tcpA, as well as the acf genes and the genes encoding the transcriptional activators ToxT and TcpP (see below), are located on a ∼40-kb Vibriopathogenicity island (VPI [22, 79]). Despite its essential role in colonization, TCP appears to not directly bind to intestinal epithelia, but rather seems to facilitate microcolony formation (i.e. bacterial–bacterial interaction) on the epithelial cell surface . Other factor(s) may be involved in adhesion to the intestinal epithelium, however, no specific V. cholerae adhesins or specific mucosal receptors have been identified yet.
6.2 The CT-converting phage CTXφ
CT, while being critical for the symptoms of cholera, appears to play no direct role in intestinal colonization . However, experimental data obtained from deletion analysis of ctxAB mutants suggests that CT activity may enhance growth in the intestinal environment through its destruction of epithelial cells . In contrast, there is no experimental evidence showing any benefit of CT activity for V. cholerae outside its human host. The genes encoding CT are actually encoded within a single-stranded filamentous phage, termed CTXφ. The phage contains two regions, ‘RS2’ and the ‘core’. RS2 encodes genes involved in integration, replication and regulation of CTXφ, whereas the core region is composed of the CT operon (ctxAB), as well as the zot, ace, cep, and orfU genes (Table 1) . zot and ace were originally believed to encode additional toxins [83, 84], but are now recognized along with orfU and cep as genes involved in CTXφ morphogenesis (Table 1) . The CTXφ is frequently found integrated into the chromosome in tandem arrays of prophage DNA, and this arrangement is important for excision and transmission of the virion . The CTXφ integrates into the attRS element located in the large V. cholerae chromosome; the CTXφ can also replicate as a double-stranded RF form in the cytoplasm . Interestingly, in the RF form CT expression is dependent on phage induction and independent of the virulence gene expression cascade (see below) .
Relevant genes associated with virulence and environmental properties in V. cholerae
Virulence (v)/environmental (e)
M13 gene VI homologue formerly ‘accessory enterotoxin’
accessory colonization factors, function unknown
aldehyde dehydrogenase ToxT-activated
M13 gene VIII homologue formerly ‘core-encoded pili’
CT subunits A, B
flagellar transcriptional regulator
iron-regulated outer membrane protein
type IV pili (mannose-sensitive hemagglutinin)
outer membrane porins
M13 gene III homologue
regulation, integration, replication
‘repeats in toxin’ toxin, cross-links cellular actin
toxin co-regulated pili major subunit (type IV pili)
transmembrane regulatory proteins
transmembrane regulatory proteins
virulence transcriptional activator
LPS core oligosaccharide synthesis
M13 gene I homologue formerly ‘zonula occludens toxin’
6.3 Surface compounds
LPS represents the most abundant of exposed molecules in the outer membrane of Gram-negative bacteria, and contributes to barrier function [87, 88]. During infection, V. cholerae cells are exposed to a series of changes, such as temperature, acidity, osmolarity and exposure to antibacterial agents and innate immune system components. The outer membrane effectively prevents the entry of noxious compounds into the cell, helps to evade recognition by host agents, and may facilitate colonization. LPS and capsule structures represent important virulence factors and typing determinants of pathogenic V. cholerae strains (Fig. 4). Mutant analysis for O1 and O139 strains lacking O-antigen as well as O139 strains lacking the capsule showed that they are significantly attenuated for intestinal colonization in the mouse model [37, 76, 89–92]. Moreover, mutations that affect the biosynthesis of LPS core oligosaccharide (core-OS) also reduce intestinal colonization [37, 93]. Recently, we have found that the core-OS-encoding gene clusters (wav genes) of various environmental and clinical V. cholerae isolates have varying genes and organization, and can be differentiated into distinctive sub-groups. One particular wav gene cluster is highly associated with pathogenic isolates, suggesting a contribution of this particular core-OS to virulence .
In addition to LPS, the outer membrane proteins OmpU and OmpT contribute to V. cholerae virulence. These porins were suspected to be involved in virulence because their expression is regulated by ToxR, which also regulates CT and TCP (see below) ; ToxR activates the transcription of ompU and represses the transcription of ompT [96–98]. OmpU is more protective (compared to OmpT) against the bactericidal effects of bile salts and other anionic detergents [99, 100]. Switching the ToxR-dependent regulation of these porins (i.e. ToxR activation of ompT and repression of ompU) leads to increased bile sensitivity, reduced CT and TCP expression and reduced intestinal colonization . These results can be explained by the differential permeabilities of the porins affecting the flux of molecules (e.g. bile salts) across the outer membrane and thereby altering the induction of the virulence cascade (see below). Recent evidence supports OmpT having a higher permeability than OmpU to negatively charged molecules such as bile salts . V. cholerae is also able to mount an organic acid tolerance response, and this has been shown to be mediated by OmpU [102–104].
Flagellar-dependent motility has been linked to V. cholerae virulence, but the connection between motility and pathogenesis has been elusive. Non-motile mutants were shown to have decreased virulence in some animal models but not others [105–107]. Importantly, in human volunteer studies motility appears to play a role in virulence [108, 109]. Recent studies suggest that the transcriptional activity of a flagellar regulatory protein, FlrC, affects the ability of V. cholerae to colonize the intestine, indicating a possible link between virulence and motility . Normal transcription of ctx within the intestine is dependent on motility and chemotaxis, suggesting that V. cholerae must navigate to the correct site within the intestine before inducing virulence factor expression .
6.4 Additional virulence factors
A variety of additional putative virulence factors have been characterized that either do not affect colonization or have not been evaluated for effects on colonization; these factors may be involved in other aspects of cholera pathogenesis. These include protein factors such as neuroaminidases, secreted proteases, hemagglutinin (HA) protease and a hemolysin–cytolysin (as reviewed by Kaper et al. ). Recently, a gene encoding a ‘repeats in toxin’ (RTX)-type protein, encoded by rtxA, was discovered linked to the CT prophage . The RTX toxin exhibits cross-linking activity on cellular actin leading to cell rounding in tissue culture, and G-actin appears to be the toxin target . The role of this toxin in cholera is unclear because classical strains appear to lack functional RTX.
7 Transcriptional regulation of virulence factor expression
Extensive study of the molecular basis of cholera pathogenesis has revealed that expression of the most important virulence factors, namely CT and TCP, is coordinated by a unique regulatory system (Fig. 5). This regulatory system consists of three different transcriptional activators, two (ToxT and TcpP) located within the VPI, and the third, ToxR, located within the ancestral Vibrio chromosome [22, 79, 114]. The AraC-like transcriptional activator ToxT directly activates both the ctx and tcp gene clusters, as well as additional genes (e.g. aldA, tagA) [115, 116]. V. cholerae strains lacking ToxT make no CT or TCP and are avirulent . Transcription of the toxT gene is controlled by the two transmembrane transcriptional activators ToxR and TcpP [117, 118]. Recent studies suggest that both proteins bind to the toxT promoter, with ToxR binding distally and acting as a ‘scaffold’ protein that allows TcpP binding adjacent to the RNA polymerase binding site .
The regulatory cascade that controls ctx and tcp transcription. Induction of CT and TCP expression in V. cholerae occurs in three basic steps (at least in vitro): (1) AphA and AphB activate transcription of tcpPH, (2) TcpPH and ToxRS (localized in the cytoplasmic membrane; ‘IM’) activate transcription of toxT, and (3) ToxT activates transcription of the genes encoding TCP (tcp) and CT (ctx), as well as additional genes (acfA, tagA, aldA, tcpI). Arrows within genes indicate direction of transcription. Within the laboratory, inducing environmental signals primarily stimulate step 1, while some repressing signals affect step 3. ToxRS alone modulates outer membrane porin expression by activating the transcription of ompU (‘+’) and repressing the transcription of ompT (‘−’). OmpU (shown as a porin trimer in the outer membrane; ‘OM’) is predicted to be less permissive for passage of certain substances like bile through the outer membrane, compared to OmpT. The tcp, toxT, and acf genes are located on a horizontally acquired pathogenicity island (VPI; not drawn completely nor to scale). The ToxT-activated tcpI, tagA and aldA genes lie upstream of tcpPH within the VPI. The ctx genes are found on the lysogenic filamentous bacteriophage CTXφ which integrates into the attRS site in the chromosome. The toxRS, ompU, ompT, aphA and aphB genes reside in the ancestral Vibrio chromosome.
The expression of CT and TCP is induced within the intestinal environment, and only occurs under specific culture medium compositions (e.g. peptone yeast NaCl, NaHCO3; thiosulfate citrate yeast; casamino acids yeast extract peptone; and yeast extract peptone water medium). Likewise, transcription of toxT occurs only under these same virulence-‘inducing’ conditions, consistent with ToxT expression controlling virulence factor induction . Environmental signals such as osmolarity, pH and temperature affect toxT transcription in vitro, but the true intestinal signals that stimulate its transcription in vivo remain unknown. The signals that stimulate toxT transcription actually appear to stimulate tcpP transcription , and thus this virulence-regulatory cascade is initiated by TcpP expression, which is regulated by two additional activators, AphA and AphB [122, 123]. Recent evidence has demonstrated that these two activators bind to the tcpP promoter and synergistically activate its transcription . Expression of AphA and AphB, whose genes appear to be located in the ancestral Vibrio chromosome, is independent of the in vitro virulence factor-inducing conditions , and so the inducing environmental conditions likely modulate their activity. aphA and aphB mutants are defective for tcpP and toxT transcription, and subsequent intestinal colonization and CT expression, consistent with these gene products occupying the top level of the known virulence cascade ; it is not yet known what other genes are regulated by AphA and AphB besides tcpP on the VPI. Levels of ToxR protein, like AphA and AphB, also appear to remain constant under various in vitro conditions , consistent with ToxR continuously occupying the toxT promoter and acting as a ‘scaffold’ protein that facilitates TcpP binding once it is expressed, leading to toxT transcription and subsequent virulence factor expression. This implies that ToxR continuously tethers the toxT promoter to the cytoplasmic membrane, but as yet there is no direct evidence for this.
Although ToxT appears to be synthesized in a transcriptionally active form, certain environmental signals (e.g. bile, temperature) can inhibit ToxT transcriptional activity . This observation led to the suggestion that the virulence cascade occurs in two temporally and spatially distinct steps within the intestinal environment: first toxT is transcribed within the lumen of the intestine in the presence of bile but the ToxT protein remains in an inactive state, until the bacteria penetrate the mucus lining and arrive at the cell surface, where lower bile concentrations would allow for ToxT-dependent expression of CT and TCP.
Camilli and coworkers have shown that ctx and tcp transcription within the intestine differs from their transcription under laboratory ‘inducing’ conditions. Most interestingly, ctx transcription was dependent on TCP expression within the intestine (unlike in the laboratory) , suggesting that the bacteria must colonize the intestinal cell surface to receive a specific signal (perhaps lower bile concentrations?) that leads to robust CT and TCP expression. Also interesting to note is that full ctx transcription within the intestine required toxR but not tcpP (unlike in the laboratory), suggesting that the true V. cholerae virulence cascade may differ significantly from that elucidated through in vitro laboratory experiments, and thus demonstrating the importance of studying pathogenesis in the context of a living host rather than in a test tube.
8 Evolution and transmission of virulence factors
It has become clear that V. cholerae virulence is the result of a series of horizontal gene transfers that allowed a benign marine bacterium to rapidly evolve into a dangerous human pathogen (summarized in Fig. 6). This transition can be achieved by the acquisition of two transmissible elements, the VPI, which contains the TCP cluster necessary for intestinal colonization, and phage CTXφ, which encodes the CT that leads to the symptoms of cholera. The VPI, which is found in pathogenic strains and sporadically in environmental strains [22, 126], has been suggested to be a filamentous bacteriophage . This element can clearly move by some means from strain to strain, perhaps through phage transduction events or co-mobilization along with the self-transmissible replicon SXT (; M. Waldor, personal communication). After V. cholerae acquires the VPI and expresses TcpA, it becomes susceptible to CTXφ transduction, because this phage uses the TcpA as its receptor . The CTXφ has been found in environmental (non-O1) V. cholerae as well as in V. mimicus strains [129, 130]. Because tcp expression is induced in the small intestine, toxigenic conversion by CTXφ can occur there also. It is noteworthy that the ctx genes on CTXφ are transcribed by the same regulatory factors that are responsible for TCP expression, and thus the phage appears to be ‘choosing’ a bacterium that is able to express CT.
Horizontal gene transfer events important for V. cholerae virulence. Shown are the possible events of horizontal gene transfer, which led to the acquisition of important virulence factor-encoding genes. See text for more information.
The regulatory cascade responsible for CT and TCP expression (described above) is a fascinating mixture of ancestral and recently acquired elements. AphA and AphB, which appear to initiate the virulence cascade by stimulating tcpP transcription, are probably located in the ancestral Vibrio chromosome, and yet they both recognize and activate the tcpP promoter, located on the VPI. TcpP then interacts with ToxR, which is encoded in the ancestral Vibrio genome, to activate the transcription of the toxT gene, located on the VPI. Finally ToxT activates the transcription of tcp and acf genes, located on the VPI, and the ctx genes, located on the CTXφ. One can hypothesize that the elements located on the recently acquired VPI are usurping specific Vibrio regulatory factors because they respond to signals relevant to the intestinal environment. Presumably these complex interactions allow the integration of multiple environmental signals into the virulence cascade.
V. cholerae continues to evolve as a human pathogen. For example, the new serogroup O139 emerged as a result of newly acquired LPS biosynthesis-encoding genes (rfb). The high level of relatedness between O139 and O1 El Tor strains [131–133] indicates that epidemic O139 strains are derived from O1 El Tor strains, mainly through the acquisition of the genes encoding synthesis of the O139 antigen [89, 134, 135]. It appears that the genes encoding the O1 antigen have been replaced by genes with characteristic features of group 4 capsule gene clusters [136, 137], leading to the synthesis of the LPS-linked O139 antigen and an unlinked O139 antigen capsule [138, 139]. This new variant was able to circumvent established immunity in human populations that had been previously exposed to O1, and has led to large outbreaks in 1992/93 in Bangladesh and India. Considering that other non-O1 and non-O139 environmental V. cholerae strains have been found with both VPI and CTXφ, it is still not clear why only those two serogroups have been associated with pandemic disease.
9 Control of cholera: conceptions of cholera vaccines
Beside hygienic and sanitary control measures and cholera surveillance, one of the main efforts at combating cholera epidemics is directed towards the development and use of modern vaccine strategies (for an overview see Tacket et al. ). Cholera is predicted to have a high potential for successful prevention by vaccination . Analysis of a large number of Indian residents demonstrated that a primary episode of cholera disease provided significant protection against a second episode . Efficient protection is dependent on the biotype: infection with the classical biotype shows more conserved protection against different serotypes (Inaba, Ogawa, and Hikojima) of classical strains, and El Tor-derived protection is more labile against different El Tor isolates . It was also found that naturally acquired immunity lasts for at least 3 years, whereas longer immunity depends on the individual . Antigenic components that contribute to the protective immune response are CT, LPS, flagella, fimbriae and outer membrane proteins. Immunity induced by anti-CT alone, however, is short-lived and not as effective as antibody responses to the whole bacterium .
In former vaccination strategies, acellular (inactivated holotoxins) or whole cell (phenol-inactivated) preparations only provided short-lived protection and proved to be reactogenic (i.e. causing some symptoms of disease). For example, in a large clinical trial in Bangladesh (reviewed in ), a vaccine consisting of a mixture of CT subunit B and killed whole bacterial cell cocktails or killed cells alone conferred about 85% and 58% protection, respectively, against cholera for the first 6 months and declined over 36 months to about 50% for both vaccine types.
Modern cholera vaccine development, based on attenuated live vaccine strains (for review see ), takes several biological parameters into account. (i) What type of immunity needs to be stimulated to provide protection against cholera? (ii) What virulence factors must be removed from the vaccine strain to eliminate disease symptoms? (iii) What factors have to be expressed to achieve efficient colonization and protective immunity? And finally, (iv) how can the reacquisition of virulence factors by horizontal transfer (phage conversion) be prevented in a live vaccine strain? For the future development of live attenuated oral vaccine strategies these questions are of fundamental importance since the maintenance of intestinal colonization enhances the risk of acquiring phage-transmitted CT.
In current approaches these questions are addressed in the following manner. (i) Colonization of V. cholerae occurs in the upper small intestine and is sensed by the local mucosal immune system. Because the organisms are non-invasive, the transcytotic activity of the M cells is of foremost importance to present antigens to the underlying immune cells of the lymph system. If specific IgA production is induced, the secretion of these antibodies into the intestinal lumen ensures the immune response against the colonizing strain. For this reason an oral vaccination strategy with a live strain, which mimics natural infection, is important to induce intestinal mucosal immunity. (ii) Because the primary symptoms of cholera are caused by CT, it was originally assumed that V. cholerae deleted for ctx (or just the toxic ctxA portion) would make a good vaccine. However, such strains still caused significant disease symptoms (reactogenic) in humans, indicating the presence of additional virulence factors [145–147]. Therefore, investigations are under way to address the contribution of other factors to reactogenicity, e.g. the HA protease . (iii) The expression of TCP provides efficient colonization and the overexpression of the CT B subunit (the non-toxic portion) seems to produce good antigenicity, although the issue of reactogenicity still needs to be resolved. (iv) Since TCP is both critical for colonization and also serves as the CTXφ receptor, a ctx vaccine strain could undergo toxigenic conversion and thus the question of vaccine safety has become an important issue. Waldor and colleagues have described a clever technique to protect live vaccines against toxigenic conversion by the CTXφ, namely the provision of the CTXφ repressor protein RstR within the vaccine strain, which provides immunity against subsequent CTXφ infection .
It is beyond the scope of this review to comprehensively cover all the research into CT, toxin regulation and V. cholerae pathogenesis. The characterization of the CT itself had a huge impact on the understanding of basic mechanisms of cellular signaling and of the function of a class of cell surface glycoconjugates known as gangliosides. During the past decade of cholera research two important findings have illuminated the complexity of V. cholerae: first, the characterization of a novel serogroup (O139) which circumvented established immunity against O1, and second, the discovery that CT genes are encoded on a bacteriophage genome. These two characteristics of V. cholerae make vaccine development more complicated, and illustrate how bacterial pathogens can rapidly evolve through the acquisition of horizontally transferred genetic elements.
Recently, researchers have begun to appreciate and study the environmental interepidemic persistence mode of V. cholerae. Our understanding of biofilm formation and the VNC state is still in its infancy, but this environmental phase of the V. cholerae infectious cycle is critical for the establishment of endemic pathogenic strains in various water sources, and serves as the inoculum that initiates cholera pandemics. Also, horizontal gene transfer that leads to new epidemic strains likely occurs in this environment, and therefore cannot be ignored. A better understanding of biofilm formation and the VNC state should lead to the identification of strategies to forecast and prevent cholera outbreaks and hopefully eradicate V. cholerae from community water supplies in developing countries.
For contributing to this review we thank Ronald Taylor and Paula Watnick. We also thank Stefan Schlör for critical reading of the manuscript. Our part of the work was founded by the BMBF Grant 01KI8906 and NIH AI43486.
Blake, P.A. (1994) Historical perspectives on pandemic cholera. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsik, O., Eds.), pp. 293–295. American Society for Microbiology Press, Washington, DC.
Swerdlow, D.L. and Isaäcson, M. (1994) The epidemiology of cholera in Africa. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsik, O., Eds.), pp. 297–307. American Society for Microbiology Press, Washington, DC.
Tauxe, R., Seminario, L., Tapita, R. and Libel, M. (1994) The Latin American epidemic. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsik, O., Eds.), pp. 321–344. American Society for Microbiology Press, Washington, DC.
Albert, J.B., Siddinque, A.K., Islam, M.S., Faruquwe, A.S.G., Anzarzzaman, M., Faruque, S.M., Sack, R. Large outbreak of clinical cholera due to Vibrio cholerae non-O1 in Bangladesh,. Lancet. 341, 1993. 704
(1993) Emergence of novel strain of Vibrio cholerae with epidemic potential in southern and eastern India. Lancet 341, 703–704.
Rodrigue, D.C., Popovic, T. and Wachsmuth, K. (1994) Nontoxigenic Vibrio cholerae O1 infections in the United States. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsvik, O., Eds.), pp. 69–76. American Society for Microbiology, Washington, DC.
Morris, J.G. (1994) Non-O group 1 Vibrio cholerae strains not associated with epidemic disease. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, I.K., Blake, P.A., Olsvik, O., Eds.), pp. 103–116. American Society for Microbiology, Washington, DC.
et al. (1996) Nontoxigenic Vibrio cholerae 01 serotype Inaba biotype El Tor associated with a cluster of cases of cholera in southern India. J. Clin. Microbiol. 34, 1114–1117.
(1999) Vibrio cholerae O1 El Tor: Identification of a gene cluster required for the rugose colony type, exopolysaccharide production, chlorine resistance, and biofilm formation. Proc. Natl. Acad. Sci. USA 96, 4028–4033.
Kaysner, C.A. and Hill, W.E. (1994) Toxigenic Vibrio cholerae O1 in food and water. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsvik, O., Eds.), pp. 27–39. American Society for Microbiology, Washington, DC.
Tison, D.L. (1999) Vibrio. In: Manual of Clinical Microbiology (Murray, P.R., Baron, E.J., Pfaller, M.A., Tenover, F.C. and Yolken, R.H., Eds.), pp. 497–506. American Society for Microbiology, Washington, DC.
Bradford, A.K., Cheryl, A.B. and Wells, J.G. (1994) Isolation and identification of Vibrio cholerae O1 from fecal specimens. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmut, K., Blake, P.A. and Olsvik, O., Eds.), pp. 3–25. American Society for Microbiology, Washington, DC.
Bennish, M.L. (1994) Cholera: pathophysiology, clinical features, and treatment. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsik, O., Eds.), pp. 229–255. American Society for Microbiology, Washington, DC.
Kaper, J.B., Fasano, A. and Trucksis, M. (1994) Toxins of Vibrio cholerae. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsvik, O., Eds.), pp. 145–176. American Society for Microbiology, Washington, DC.
(1994) Multiple regulatory systems in Vibrio cholerae pathogenesis. Trends Microbiol. 2, 37–38.
Ottemann, K.M. and Mekalanos, J.J. (1994) Regulation of cholera toxin expression. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, I.K., Blake, P.A. and Olsvik, O., Eds.), pp. 177–185. American Society for Microbiology, Washington, DC.
(2000) Altered expression of the ToxR-regulated porins OmpU and OmpT diminishes Vibrio cholerae bile resistance, virulence factor expression, and intestinal colonization. Proc. Natl. Acad. Sci. USA 97, 10220–10224.
(1988) A novel suicide vector and its use in construction of insertion mutations: osmoregulation of outer membrane proteins and virulence determinants in Vibrio cholerae requires toxR. J. Bacteriol. 170, 2575–2583.
(1981) Role of chemotaxis in the association of motile bacteria with intestinal mucosa: chemotactic responses of Vibrio cholerae and description of motile non-chemotactic mutants. Infect. Immun. 34, 215–221.
(2000) A region of the transmembrane regulatory protein toxR that tethers the transcriptional activation domain to the cytoplasmic membrane displays wide divergence among Vibrio species. J. Bacteriol. 182, 526–528.
(1996) Differential expression of the ToxR regulon in classical and El Tor biotypes of Vibrio cholerae is due to biotype-specific control over toxT expression. Proc. Natl. Acad. Sci. USA 93, 7991–7995.
Levine, M.M. and Tacket, C.O. (1994) Recombinant live cholera vaccines. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsvik, O., Eds.), pp. 395–415. American Society for Microbiology, Washington, DC.
Holmgren, J., Osek, J. and Svennerholm, A.M. (1994) Protective oral cholera vaccine based on a combination of cholera toxin B subunit and inactivated cholera vibrios. In: Vibrio cholerae and Cholera: Molecular to Global Perspectives (Wachsmuth, K.I., Blake, P.A. and Olsvik, O., Eds.), pp. 415–424. American Society for Microbiology, Washington, DC.
(2000) Association of protease activity in Vibrio cholerae vaccine strains with decreases in transcellular epithelial resistance of polarized T84 intestinal epithelial cells. Infect. Immun. 68, 6487–6492.