Structural inferences for Cholera toxin mutations in Vibrio cholerae

Abstract

Cholera is a global disease that has persisted for millennia. The cholera toxin (CT) from Vibrio cholerae is responsible for the clinical symptoms of cholera. This toxin is a hetero-hexamer (AB₅) complex consisting of a subunit A (CTA) with a pentamer (B₅) of subunit B (CTB). The importance of the AB₅ complex for pathogenesis is established for the wild type O1 serogroup using known structural and functional data. However, its role is not yet documented in other known serogroups harboring sequence level residue mutations. The sequences for the toxin from different serogroups are available in GenBank (release 177). Sequence analysis reveals mutations at several sequence positions in the toxin across serogroups. Therefore, it is of interest to locate the position of these mutations in the AB₅ structure to infer complex assembly for its functional role in different serogroups. We show that mutations in the CTA are at the solvent exposed regions of the AB₅ complex, whereas those in the CTB are at the CTB/CTB interface of the homo-pentamer complex. Thus, the role of mutations at the CTB/CTB interface for B₅ complex assembly is implied. It is observed that these mutations are often non-synonymous (e.g. polar to non-polar or vice versa). The formation of the AB₅ complex involves inter-subunit residue-residue interactions at the protein-protein interfaces. Hence, these mutations, at the structurally relevant positions, are of importance for the understanding of pathogenesis by several serogroups. This is also of significance in the improvement of recombinant CT protein complex analogs for vaccine design and their use against multiple serogroups.

Background

Vibrio cholerae is the cause of a waterborne disease affecting thousands of life every year [1]. The outbreak in October, 2010, in Haiti demonstrates the global issue of cholera and resulted in approximately 1,000 deaths within a month [2]. Cholera is an acute diarrheal disease caused by the gram-negative bacterium, Vibrio cholerae. There are more than 200 serogroups of Vibrio cholerae present in the natural environment [3]. However, two serogroups, O1 (widespread with El Tor and classical biotypes) and O139 (colonizes few regions of Asia) have been associated with the epidemics and pandemics of cholera during the last 25 years [4– 6]. The O1 (El Tor biotype ¯ Ogawa serotype) serogroup is responsible for the recent outbreak in Haiti [7]. The symptoms of cholera are caused by the secretion of an entero-toxin called cholera toxin (CT) [8– 9] which is encoded by virulence factor genes; ctxA and ctxB [10–11]. These genes are acquired from a lysogenic filamentous bacteriophage (CTXφ) through CTXφ DNA integration into the host Vibrio cholerae genome [12– 14]. It should be noted that the incidence of cholera outbreaks with serogroups other than O1/O139 (collectively referred as non O1/non O139) has also been recorded [5, 10,15– 17]. These strains are responsible for the sporadic outbreaks [18– 22]. It is known that the virulent factors for non-(O1/O139) are different from the O1/O139 strains [5, 23, 24]. However, non-(O1/O139) strains with ctxA and ctxB genes also have been observed [25–28].

CT, also known as choleragen, is a hetero-hexameric AB5 complex in structure (Figure 1) [29–31] and is composed of an enzymatic A subunit (CTA) and a cell targeting B subunit (CTB) [32– 34]. The enzymatically activated A subunit catalyzes adenylate cyclase to cause massive excretion of electrolytes from bowel [35, 36]. However, the homo-pentamer B subunit is mandatory for pathogenesis because of its vital role in binding to receptors of target cells [37–39]. The B complex binds to the intestinal epithelium and the A molecule then detaches and enters the cell via endocytosis. The A molecule then goes onto ribosylating the Gs alpha-subunit of G proteins that results in constitutive production of cAMP. The result is excretion of bicarbonate, chloride, potassium, and sodium ions as well as water from these cells [40]. Thus, the AB5 complex assembly is critical for pathogenesis. The virulence factors in both O1/O139 and non-(O1/O139) strains have been identified [8, 9, 16, 17, 24, 28, 41]. It should be noted that information on the diarrheagenic potential of non-(O1/O139) is limited. The effect of mutations in the toxin from all known serogroups is not available. Therefore, it is of importance to describe the virulence factors in both O1/O139 and non-(O1/O139). This is possible with the help of known structural complexes available in Protein databank (PDB). A comprehensive analysis of the AB₅ CT structures from PDB describing the nature of A and B₅ interface has been documented elsewhere [42]. Here, we describe the significance of mutations in CT among serogroups based on their residue positional occurrence (either at solvent exposed or interface regions of the AB₅ complex).

Figure 1.

Structural model of a cholera toxin (CT). CT is a hetero-hexameric complex (AB₅) consisting of CTA (194 residues A1 and 46 residues A2) and CTB (103 residues) pentamer with D, E, F, G and H chains.

Materials and Methodology

CT sequence dataset

We created a dataset of 27 CTA (O1: 14; O139: 5; non-O1/O139: 8) and 165 CTB (O1: 121; O139: 37; non-O1/O139: 7) sequences as available from GenBank (release 177; year 2010 [43] using the procedure outlined in Figure 2. The number of sequences in the datasets is stated in Table 1 (see supplementary material). There are more CTB sequences than CTA sequences suggesting a higher frequency of mutations in CTB. Some partial sequences have been included in the dataset due to the non-availability of their full-length sequences in GenBank. In addition, these partial sequences also harbored mutations compared to wild type sequences.

Figure 2.

Creation of sequence dataset for CTA and CTB. A sequence dataset of CTA and CTB was derived from GenBank (release 177) using KEYWORD search as illustrated in the flowchart. The KEYWORD search “cholera toxin” resulted in 1257 hits. This set consists of 27 CTA sequences, 165 CTB sequences according to GenBank description and available annotations. The remaining 1065 sequences with descriptions such as secretion protein, cholera toxin transcriptional activator, ADP-ribosylation factor, GNAS complex, dopamine receptor, Pertusis toxin, Shiga-like toxin and the like are eliminated from the dataset. Thus, a CT sequence dataset of 192 sequences (Table 1 in supplementary material) consisting of 27 CTA and 165 CTB was created. The CTA and CTB sequences are included in the dataset as available in the GenBank. The biased availability on the amount of CTA and CTB sequences in GenBank is attributed to the likely observation of frequent mutations in CTB.

Multiple Sequence Alignment (MSA) of CTA and CTB

MSA is performed using ClustalX 2.0.12 [44] with the substitution matrix PAM 80. A gap-opening penalty of 10 and extension-penalty of 0.2 were used for the alignment. Sequences of CTA and CTB with known structures (PDB ID: 1XTC [45]) belonging to the O1 classical 569B strain were used as reference sequences in this alignment. The alignment was used to identify mutations in CTA (Figure 3) and CTB (Figure 4) among the different serogroups. Mutations were identified at six residue positions (7, 28, 112, 134, 163 and 222) in CTA (Figure 3) and at 13 residue positions (3, 7, 13, 15, 18, 22, 25, 34, 46, 47, 52, 60 and 94) in CTB (Figure 4) among O1/O139 and non- (O1/O139) strains.

Figure 3.

MSA for the CTA subunit of different serogroups. The MSA was performed using the wild type O1 classical strain sequence with known structure (PDB ID: 1XTC) as reference. The position specific mutations among the available CTA sequences (27) with reference to the classical sequence are indicated using dark shades. CT is an AB₅ hetero-hexamer and hence, the CTA/CTB interface residues in CTA are indicated using light shades.

Figure 4.

MSA for the CTB subunit of different serogroups. The MSA was performed using the wild type O1 Classical strain with known structure (PDB ID:1XTC) as reference. The position specific mutations among the available CTB sequences (165) with reference to the Classical sequence are indicated using dark shades. B₅ is a homo-pentamer and hence, the CTB/CTB interface residues in B₅ are indicated using light shades. It should be noted that the mutated residues at the CTB/CTB interfaces in B₅ are highlighted using both dark and light shades at their corresponding position specific residues.

CT structures

The formation of the AB₅ complex is critical for pathogenesis. This is achieved through the formation of B₅ and AB₅ complexes. The B₅ complex is formed through the assembly of 5 monomeric B subunits arranged in a circle with a central groove in the first stage. This results in an assembly with each B subunit juxtaposed on either side with two other B subunits with a stable interface as shown in Figure 5. Mutations in the B subunit and the potential occurrence at the CTB/CTB interface influence the formation of the B₅ complex. The formation of AB5 complex occurs through the interaction of CTA and B₅ complex. Thus, mutations in either CTA or CTB among the different serogroups have effect at the interface of CTA/ CTB complex.

Figure 5.

Structural model of CTB/CTB interfaces in B₅. B₅ is a homopentamer and each CTB subunit (D) is juxtaposed by two other CTB units on either side (E and H). Thus, the D subunit creates two different types of interfaces (D-E and D-H) on either side. This subsequently results in two different “position specific interacting” patterns in sequence for subunit D.

Interface residues

Interface residue positions were identified using the change in solvent accessible surface area (ASA) upon complex formation from a monomer state to a dimer state both within B₅ complex and between CTA/CTB. ASA is calculated using an algorithm developed by Lee and Richard (1971) implemented in the software SURFACE RACER with a probe radius of 1.4 Å [46]. We identified the interface residues between CTA/CTB complex and within the B₅ complex in respective serogroups using the procedure described elsewhere [42]. In this procedure, interface residue positions were identified using ASA analysis of subunits in the AB₅ structural complexes.

Mapping mutated residue positions to structures

The structures for AB₅ and B₅ complexes of the wild type O1 strain are available at the PDB. It is of interest to infer structural effect caused by the mutations in other known serogroups. It is well known that homologous sequences have similar structures and they differ only in side chain details. Therefore, mapping of mutated residue positions from MSA (Figure 3 and Figure 4) to known structural regions (exposed, buried, interface) provide the opportunity to identify mutations at the interface of CTA/CTB and within B₅ (Figure 6). This approach identified mutations (at six residue positions such as 7, 28, 112, 134, 163 and 222) in CTA that are located at structurally solvent exposed regions of the complex (Figure 6a). It also helped to locate several mutations (seven residue positions such as 3, 15, 25, 34, 47, 52 and 60) in CTB that are at the B5 homo-pentamer subunit interfaces (Figure 6b and Figure 6c).

Figure 6.

Representation of mutated residue positions in serogroups to interface residues in CT complex as a function of their residue position identified using ΔASA measure.

1. Mapping of CTA mutations to CTA/CTB interface residues in CTA (Please refer to Figure 1 for the visual illustration of CTA/CTB interface).
2. Mapping of CTB mutations to CTB (D subunit)/CTB (E subunit) interface residues (Please refer to Figure 5 for the visual illustration of D-E interface).
3. Mapping of CTB mutations to CTB (D subunit)/CTB (H subunit) interface residues (Please refer to Figure 5 for the visual illustration of D-H interface).

It should be noted that mutated residue positions are mapped on to corresponding interface residue positions in all the three cases (a), (b) and (c).

Structural 3D visualization of mutated residue positions in serogroups

We used Discovery Studio Visualizer (v2.5.5.9350) to illustrate the mutated residue positions in CTA (Figure 7a) and CTB (Figure 7b) among the serogroups. The mutated residue positions at the interface of CTA/CTB (Figure 8a) and with B5 (Figure 8b) is also shown.

Figure 7.

Structural models of CTA (a) and CTB (b) subunits with known mutations among archived serogroups. We used the structure with PDB entry (1XTC) for generating this visual using the freeware Discovery studio from Accelrys Inc.

1. A total of 6 unique mutations thus observed among the known CTA sequences (Table 2 in supplementary material) from several serogroups are shown at their corresponding 6 residue positions using the Corey-Pauling-Kultun (CPK) residue model representation.
2. Fourteen unique mutations thus observed among the known CTB sequences (Table 3 in supplementary material) from several serogroups are shown at their corresponding 13 residue positions using the CPK residue model representation.

Figure 8.

Structural models of CTA (a) and CTB (b) subunits with known mutations at the respective structural interfaces or solvent accessible regions in the complex among archived serogroups. We used the structure with PDB entry (1XTC) for generating this visual using the freeware Discovery studio from Accelrys Inc.

1. A total of 6 unique mutations thus observed among the known CTA sequences (Table 4 in supplementary material) from several serogroups are shown at their corresponding 6 residue positions using the CPK residue model representation. All of these 6 mutated positions are present at the solvent exposed regions of CTA in both monomer and CTA/CTB complex state.
2. A total of 7 out of 14 unique mutations thus observed among the known CTB sequences (Table 4 in supplementary material) from several serogroups are shown at their corresponding 7 (3, 15, 25, 34, 47,52 and 60) out of the 13 residue positions using the CPK residue model representation are at the CTB/CTB interfaces in the B₅ complex.

Results

Table 1 describes the dataset of CTA and CTB sequences retrieved from GenBank (release 177; year). The dataset consists of CTA and CTB sequences from O1 (El Tor, Classical, Matlab), O139 and non-(O1/O139) serogroups. We compared the CT sequences for O139 and non-(O1/O139) with the wild type Classical O1 serogroup. Figure 3 and Figure 4 show the results of MSA for CTA (27 sequences) and CTB (165 sequences), respectively. The wild type O1 Classical sequence with known structure (PDB ID: 1XTC) from strain 569B was used as reference in the alignment. The alignment is showed only for sequences with mutations (7 CTA and 52 CTB mutants) to the wild type reference sequence (Figure 3 and Figure 4). The mutations observed from the MSA of known CTA and CTB sequences are summarized in Table 2and (see Table 3), respectively. The mutations in CTA are found at 6 residue positions (7, 28, 112, 134, 163 and 222) among serogroups in the dataset. The mutations in CTB sequences are at 13 residue positions (3, 7, 13, 15, 18, 22, 25, 34, 46, 47, 52, 60 and 94) in the dataset.

Table 2 shows that the I222Y mutant is in the O139 strain (4260B) and the other six are in non-(O1/O139) strains. Data also shows that all strains except for strains B (2 positional mutations) and J31W (3 positional mutations) have only one positional mutation (Table 2 and Figure 3). Similarly, mutations were seen at one position in 18 strains (O1: 12; O139: 6), at two positions in 26 strains (O1: 14; O139: 7; non-(O1/O139): 5), at 3 positions in 6 O1 strains, 4 positions and 6 positions in one non-(O1/O139) strain for CTB (Figure 4 and Table 3). The non-(O1/O139) serogrouped J31W strain carried the maximum number of mutations in CT (CTA and CTB). Thus, the position specific mutations for CTA and CTB sequences were observed.

The availability of CT structure (PDB ID: 1XTC) provides an opportunity to map position-specific mutations in different serogroups to its structural preference (solvent exposed, buried, interfaces). Therefore, the significance of these mutations in the formation of the AB₅ assembly could be subsequently inferred. The mutations (dark shades as background) in the CTA and the CTB are shown in Figure 3 and Figure 4 along with corresponding interface residues (light shades). This helps to relate the consequence of mutations in structure. CT is an AB₅ complex (Figure 1) consisting of several layers of subunit protein interfaces formed by non-covalent interactions. Therefore, it is of interest to map the mutations in serogroups to their structural positions (interior, interface, surface).

Protein-Protein interfaces are formed between A and B₅ as well as within B₅. The mutations in A and B will potentially affect A/B₅ interface ( Figure 1). B₅ is a homo-pentamer and each B subunit is juxtaposed with similar CTB units on either side (Figure 5). Similarly, mutations within B will possibly affect the formation of B₅ such that the D-E and D-H interfaces are affected (Figure 5). Nevertheless, these interfaces should be translated into sequence positions using ΔASA in solved CT structures as described in Figure 6. Moreover, Figure 6 maps the mutations in CTA to their occurrence at the CTA/CTB interface (Figure 6a) and in CTB to their possible occurrence at the D-E and D-H interfaces (Figure 6b and 6c) in B₅. This comparison helps to identify the presence of mutated residues (Figure 7) in CTA (Figure 7a) and CTB (Figure 7b) at their respective CTA/CTB (Figure 8a) interface and CTB/CTB (Figure 8b) interfaces. The 6 mutations (R7W, S28N, E112G, V134G, G163R and I222Y) in CTA are positioned at the solvent exposed regions of the subunit (Figure 6) with no mutations at the CTA/CTB interface. However, it should be noted that the S28N and I222Y mutation were closely located to the CTA/CTB interface (Table 4). The role of these mutations in CT complex assembly is of interest. A number of mutations in the CTB sequence are positioned at the CTB/CTB interfaces unlike the mutations in CTA sequence. The mutations at residue positions 3, 15, 25, 34, 47, 52 and 60 are within CTB/CTB interfaces (Figure 8). The nature of amino acid mutations in CTA and CTB among O1/O139 and non-(O1/O139) serogroups are given in (see Table 4).

Discussion

Choleragen (CT) and Choleragenoid (CTB) have been used as cholera vaccine candidates [47]. A number of subunit vaccine candidates using CTA ((S63K, R192G, R192N) [48], (I16A or V72Y, I16A+Y68S, V72Y+Y68S) [49], (V53D, V53E, V53Y, S63K, V97K, V97Y, Y104K, Y104D, Y104S, P106S) [50]) mutants and CTB recombinants have been developed in addition to heat killed attenuated Vibrio cholerae as vaccines. Sequence and structural studies of CT offer tremendous opportunity for the improvements in vaccine candidate design and development. The presence of CT epitypes [51] and heterogeneity in CTB subunit [52] also need to be considered from a vaccine perspective. A vaccine for cholera must target O1, O139 as well as non-O1 and non-O139 strains to have effective control over cholera outbreaks. Moreover, different serogroups of non-(O1/O139) strains (with ctxAB genes [25, 41, 53, 54]) and newly emerging Vibrio cholerae strains (O1 Matlab [55–57], O1 El Tor with altered CTB [58, 59]) must be taken into consideration in cholera vaccine design. Hence, comparison studies on CTA and CTB sequences from various Vibrio cholerae serogroups provide insights in developing an effective toxin analog for vaccine design against multi serogroups.

A number of sequence comparison studies show CT sequence homology among various Vibrio cholerae serogroups. Recently, Kumar et al. (2009) documented a new CT variant of the Vibrio cholerae O1 El Tor biotype isolated from Orissa (India) [60]. The study highlighted a novel mutation (H20N) in CTB and the presence of altered CTB of the Classical biotype in the El Tor clinical isolates. Raychoudhuri and team (2009) conducted a study to attest the replacement of El Tor biotype ctxB allele by Classical biotype ctxB allele in O1 strains [61]. A study by Ansaruzzaman and colleagues (2004) reported H18Y and T47I substitutions in CTB of El Tor strain and these sequences are similar to CTB of Classical biotype [62]. Previously, a study demonstrated the emergence of new El Tor strains with a modified Classical biotype CT [60]. Thus, a dataset of sequences (Table 1) for CTA and CTB representing diverse serogroups isolated at various periods of time from a variety of sources and locations available in GenBank (release 177) is created for this study. The nature of mutations (Table 4) among the serogroups is presented for CTA (Table 2) and CTB (Table 3) sequences. Several studies have demonstrated the effects of site directed mutations in CTA as well as in CTB subunits for the wild type O1 strain (Table 5 see Table 5). Manufactured site directed mutants leading to decrease or loss in toxicity has been reported for CTA (R7K, R11K, I16A, R25G, E29H, S68Y + V72Y, E112Q, F223D) and CTB (R35D, H57A, L77D, I74D, T78D). Thus, the role of site directed mutants in the loss of toxicity is known for the wild type O1 strain. Therefore, it is important to evaluate the effect of mutations caused by natural selection pressure among serogroups.

The multiple sequence alignment (MSA) of these sequences showed mutations in CTA (at six residue positions such as 7, 28, 112, 134, 163 and 222) and CTB (at 13 residue positions such as 3, 7, 13, 15, 18, 22, 25, 34, 46, 47, 52, 60 and 94) among O1/O139 and non-(O1/O139) strains. The effects of these mutations in the formation of a clinically functional cholera toxin (AB₅ hetero-hexamer) are of significant importance. Reports describing the emergence of new serogroups with novel mutations in CTA and CTB are available. However, studies on the effects of mutations in CT relative to CTA/CTB-pentamer interface (Figure 1) and within CTB/CTB interfaces (Figure 5) are not yet available. Here, we present results of a comprehensive analysis of mutations in CTA and CTB sequences from several serogroups (Table 4).

This mutational data is presented relative to A/B5 and CTB/CTB interfaces for AB5 assembly to understand its functions. Data suggest the presence of mutations in CTA (Figure 8a) and CTB (Figure 8b) at the solvent exposed, interior, subunit interface regions of the complex. The mutations (at 6 residue positions such as 7, 28, 112, 134, 163 and 222) in CTA are located at structurally solvent exposed regions of the subunit. However, several mutations (7 residue positions such as 3, 15, 25, 34, 47, 52 and 60) in CTB are at the B₅ homo-pentamer subunit interfaces. Thus, the role of these mutations in CTA and CTB towards the assembly of AB₅ CT among the O1/O139 and non (O1/O139) strains is inferred from this study (Figure 7 and Figure 8). It should also be noted that some of these mutations (polar to non-polar or vice versa) are largely non-synonymous (causing physical and chemical property shift) in nature and have potential effect on protein-protein interactions of the CT subunits affecting AB₅ formation (Table 5). Thus, data presented in Table 4 is allinclusive, updated, relevant and specific for several known serogroups. This is of significance towards the improvement of recombinant CT protein complex analogs for vaccine design against multi serogroups.

Conclusion

The structural role of cholera toxin in pathogenesis is known for the wild type O1 strain. It was of interest to document its role in other known serogroups showing mutations with the wild type. We described the structural location of such mutations in the known serogroups to infer its functional role. We documented that mutations in CTA are at the solvent exposed regions of the AB₅ complex, while those in CTB are at the CTB/CTB interface of the homopentamer complex. It is observed that these mutations are also nonsynonymous (i.e. polar to non-polar or vice versa) in property. Thus, the effect of these mutations in the AB₅ assembly is inferred. It is also of global importance to quantify precisely the structural effects caused by these mutations. The resulting data is relevant in designing a recombinant CT protein complex analog for vaccine design against multiple serogroups. Coupled to these analyses, it may be stated also that from a clinical perspective, the task of enhancing oral cholera vaccines entails reducing bacterial and Giardia infection and improving diet [63].

Supplementary material

Data 1

[64–72]