- Email: [email protected]
Secretory and GM1 receptor binding role of N-terminal region of LTB in Vibrio cholerae
Biochemical and Biophysical Research Communications 376 (2008) 770–774
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Secretory and GM1 receptor binding role of N-terminal region of LTB in Vibrio cholerae Pankaj V. Alone 1, Lalit C. Garg * Gene Regulation Laboratory, National Institute of Immunology, Aruna Asaf Ali Marg, New Delhi, Delhi 110067, India
a r t i c l e
i n f o
Article history: Received 1 September 2008 Available online 22 September 2008
Keywords: Heat labile enterotoxin Cholera toxin Secretion Periplasm CD spectra a1 helix
a b s t r a c t Heat labile enterotoxin from enterotoxigenic Escherichia coli is similar to cholera toxin (CT) and is a leading cause of diarrhea in developing countries. It consists of an enzymatically active A subunit (LTA) and a carrier pentameric B subunit (LTB). In the current study, we evaluated the importance of the N-terminal region of LTB by mutation analysis. Deletion of the glutamine (DQ3) residue and a substitution mutation E7G in the a1 helix region led to defects in LTB protein secretion. Deletion of the proline residue (DP2) caused a decrease in a helicity. The DP2 mutant affected GM1 ganglioside receptor binding activity without affecting LTB pentamer formation. Upon refolding/reassembly, the DP2 mutant showed defective biological activity. The single substitution mutation (E7D) strengthened the helix, imparting structural stability and thereby improved the GM1 ganglioside receptor binding activity. Our results demonstrate the important role of N-terminal a1 helix in maintaining the structural stability and the integrity of GM1 ganglioside receptor binding activity. Ó 2008 Elsevier Inc. All rights reserved.
Heat labile enterotoxin from enterotoxigenic Escherichia coli (ETEC) is the causative agent of diarrhea, a leading cause of mortality in developing countries. Nearly 800,000 deaths, mainly in children, are reported every year [1–3]. The enterotoxin is a heterohexameric protein consisting of an enzymatically active A subunit (LTA) and a carrier pentameric B subunit (LTB). The amino acid sequence of both LTA and LTB is more than 80% similar to cholera toxin (CT). The genes encoding the two LT subunits, eltA and eltB, are plasmid borne and are transcribed as a single polycistronic mRNA [4]. Both subunits are synthesized and exported to the periplasm in E. coli, where the B subunit during the process of pentamerization binds to the A subunit to form the holotoxin [5,6]. The expression of the B subunit of either CT or LT in engineered AB+ strain of Vibrio cholerae resulted in effective secretion of B subunit as a pentamer into the medium, although in engineered A+B strain of V. cholerae, the A subunit remained cell associated. This implies that both the subunits are assembled before being secreted and that the secretory machinery recognizes the structural domains within the B subunit in order to secrete them [7]. LTB is an important mucosal adjuvant and shows great promise for developing oral vaccines by coupling immunogenic peptides to its B subunit. The adjuvant activity of LTB is strongly linked to its
* Corresponding author. Fax: +91 11 26162125. E-mail addresses: [email protected], [email protected] (L.C. Garg). 1 Present address: University of California-Davis, School of Medicine, Davis, CA 95616, USA. 0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2008.09.066
receptor binding activity [8]. Recognition of the oligosaccharide portion of GM1 on the intestinal epithelial cell membrane is the first crucial step in the pathogenicity of LT or CT. The crystal structure of LT with lactose shows involvement of Glu51, Gln61, Ans90, and Lys91 residues in hydrogen bonding with lactose from the B monomer. Trp88 and Gly33 from the adjacent monomer (N + 1) are a part of the ganglioside binding cavity [9–11]. All the residues involved in GM1 binding have been found to be conserved in both LT and CT [12]. Unlike CT, which binds only to GM1 ganglioside receptors, LT binds to a wide variety of glycolipid and glycoprotein receptors [13,14]. LTs ability to recognize additional receptors may be due to differences in the primary amino acid sequence between LT and CT. A hybrid CT toxin made by swapping N-terminal 1–25 amino acids along with residues 75–83 with those of LT exhibited strong binding to intestinal glycosphingolipids [15]. The N-terminal end of LTB contains a a1 helix which is linked with the b5 strand by Cys9 and Cys86 disulfide bond and is not part of the core structure [9–11]. We have recently shown that the a1 helix imparts stability to the LTB, as deletion mutations within the a1 helix render LTB susceptible to degradation [16]. CTB is poorly secreted in V. cholera when E11K mutation is introduced at the N-terminal end [17]. Moreover, the fusion of 8, 12, or 24 residue peptides from GtfB.1 to the N-terminal end of CTB caused a progressive decrease in GM1 receptor binding affinity [18]. This suggests that the N-terminal end of LT and CT plays an important role in their functions and warrants further investigation. In our present work, we delineate mutations at the N-terminal end of LTB which are involved in its secretion and receptor binding.
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
Methods Conjugal transfer of mutant construct to V. cholerae JBK70 cells. Deletion of Proline 2 (DP2), deletion of glutamine 3 (DQ3), substitution of glycine for glutamic acid 7 (E7G) and substitution of aspartic acid for glutamic acid 7 (E7D) mutant constructs were designed as described in Alone et al. [16]. Mutant and wild type plasmids were conjugally transferred from DH5a cells to JBK70 (polyB+, AB) cells using E. coli pRK2013 cells as a helper strain [19]. Protein purification, reconstitution of holotoxin, and evaluation of its biological activity. The holotoxin as well as mutant and wild type (WT) LTB proteins were over-expressed into the supernatant from induced V. cholerae cultures and purified to homogeneity as described by Mekalonos et al. [20]. The purified holotoxin was either directly tested for biological activity on CHO cell line or was separated into the A and B subunits as described by Takeda et al. [21] before being tested for biological activity. The WT A subunit and the mutant B subunits were reconstituted in vitro by continuous dialysis with 10 mM Tris–Cl (pH 7.4) and 20 mM NaCl after denaturation in 6 M urea and 0.1 M propionic acid (pH 4.0) solution [22]. CHO (2 104) cells/ml were suspended in DMEM supplemented with 1% FCS and seeded into 24-well tissue culture plates. Fifty nanograms of in vitro reconstituted mutant and wild type holotoxin was added into wells and the change in morphology of CHO cells was observed under the microscope after a 16 h incubation at 37 °C with 5% CO2. Circular dichorism (CD) spectroscopy. CD spectra were obtained using a Jasco-700 spectropolarimeter at 25 °C. A protein concentration of 80 lM was maintained for CD spectra analysis. The cuvette path length was 1 cm for far-UV region measurements. Each sample was scanned 10 times and the average spectra of molar ellipticity were plotted against wavelength. GM1 ganglioside receptor binding activity by competitive-ELISA. Hundred microliters of GM1 (1 lg/ml) prepared in 50 mM carbonate buffer, pH 9.2, was coated on polypropylene ELISA plates. Nonspecific sites were saturated with 1% BSA in PBS. Biotinylated WT
A
C
771
LTB protein was mixed with different dilutions of mutant or WT LTB protein. The mixtures were added to GM1 coated plates and incubated for 1 h at 37 °C either in 10 mM or 80 mM PBS. The unbound protein was removed by washing five times with 0.05% Tween 20 in PBS. Hundred microliters of Streptavidin–HRP conjugate (1 U/well) was added and incubated for an additional 1 h. The unbound protein was washed five times with 0.05% Tween 20 in PBS, and then mixed with 0.05% orthophenylene diamine (OPD) in citrate phosphate buffer, pH 5.5, along with hydrogen peroxide (1 ll/ml). The reaction was terminated using 50 ll of 5 N H2SO4 and quantified by monitoring at A490. Results Effect of N-terminal a1 helix mutation on LTB secretion and pentamer formation Previously we reported that the deletion of DS4, DI5, DT6, DE7 residues in a1 helix causes the loss of LTB expression, due to its susceptibility to degradation [16]. In order to study the effect of mutation in the N-terminal a1 helix, the residues P2 and Q3, which lie at N00 and N0 -cap region of a1 helix, respectively, were deleted, while residue E7 was substituted with glycine or aspartic acid. The only charged residue E7 in the a1 helix was substituted with glycine (E7G) to destabilize the a1 helix, as the glycine residue has been shown to be a helix breaker [22]. The E7D substitution served as a positive control as the residue is naturally present in the CTB. Western blot analysis of the total cell extract using anti-LTB antibody showed expression of the mutant as well as WT LTB in V. cholerae cells (Fig. 1A). The mutant proteins were successfully exported into the periplasm of V. cholerae cells (Fig. 1B). It is known that upon assembly in the periplasm, the LTB pentamer can be secreted into the culture supernatant from V. cholerae cells [23]. The culture supernatant from the four mutants was analyzed for the presence of LTB protein (Fig. 1C) and it was found that the secretion of mutant LTB proteins DQ3 (lane 4) and E7G (lane 5) into the culture supernatant was severely curtailed as compared to WT (lane 1) and E7D mutant (lane 6).
B
D
Fig. 1. Analysis of LTB N-terminal a1 helix mutations on its secretion and pentamerization in Vibrio cholerae. V. cholerae (JBK 70) cells harboring WT or mutant LTB constructs were induced with 1 mM IPTG at A600 = 0.6 for 5 h. Total cell extract (A), periplasmic fraction (B) and culture supernatant (C) were analyzed by Western blotting using antiLTB polyclonal antibody. Lane 1, WT (+); lane 2, WT (); lane 3, DP2 (+); lane 4, DQ3 (+); lane 5, E7G (+); lane 6, E7D (+). + and denote cells grown in presence or absence of IPTG, respectively. Arrows indicates migration of LTB protein. Molecular weight in kDa is shown on the left. (D) Mutant LTB proteins were purified from V. cholerae culture supernatant as described in Methods. The non-denatured un-boiled purified proteins were resolved on a 15% SDS–PAGE. Lane 1, WT; lane 2, E7D; lane 3, DP2; lane 4, E7G and lane 5, DQ3. Arrow indicates migration of pentameric LTB. Molecular weight in kDa is shown on the left.
772
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
To test the impact of LTB mutations on pentamer formation, the mutant proteins were purified from V. cholerae culture supernatants and resolved on a 15% native SDS–PAGE (Fig. 1D). As compared to the WT, none of the mutations had an effect on pentamer formation. Since the mutants, DQ3 and E7G, have defective secretion, ample quantity of these proteins could not be purified to enable their further characterization by CD spectra and GM1 ganglioside binding activity. Mutations in the a1 helix of LTB cause changes in the secondary structure The CD spectra for DP2, E7D, and WT LTB were analyzed and are compared in Fig. 2. As compared to the WT LTB, a positive shift at [h]218 nm was observed in the DP2 mutant, indicating a change in its secondary structure, probably due to a weakening of the helicity of the protein. It has been observed consistently that proline prefers to stay at the N-cap position of helixes and is noted to promote helix formation [22]. In contrast to the DP2 mutant, the CD spectra of E7D showed a negative shift at [h]218 nm as compared to WT, suggesting strengthening in helicity. These results indicated that the deletion of Pro2 at the N-cap region of the a1 helix, effects larger changes on the secondary structure than does the chemically similar substitution mutation, E7D, in the helix. Analysis of GM1 ganglioside receptor binding activity of LTB mutants by competitive GM1-ELISA Pentameric LTB binds to the GM1 ganglioside receptor with strong affinity. To check if the mutations in the a1 helix have a direct effect on GM1 receptor binding activity, we employed a competitive GM1 binding-ELISA. Increasing concentrations of mutant proteins were incubated with a constant amount of biotinylated WT LTB protein and the mixtures were then reacted with immobilized GM1 ganglioside in 10 mM PBS. As shown in Fig. 3A, the LTA subunit by itself does not bind to GM1 ganglioside. At 10 mM PBS, the E7D mutant protein showed comparable affinity to that of WT LTB, while the
affinity for DP2 was reduced by 30%. However, when the ionic strength of the incubation buffer (PBS) was increased to 80 mM, a decrease in the affinity of WT and DP2 as compared to the E7D mutant protein was observed (Fig. 3B). The weakened helicity observed in the CD spectra (Fig. 2) and a consistent decrease in GM1 ganglioside binding affinity associated with DP2 mutant suggests that deleting Pro2 has a profound effect on the structural stability of the a1 helix, which may influence the conformation of the GM1 binding pocket and impair B subunit binding to GM1 ganglioside. In contrast, the E7D mutant has a more stable a1 helix as judged by CD spectra (Fig. 2) and also shows stable binding affinity to GM1 ganglioside receptor even at a higher ionic concentration. Defective biological activity of LTB DP2 mutant following in vitro refolding LTB serves as a carrier molecule for the catalytically active A subunit, and shows strong binding affinity for GM1 ganaglioside receptors on CHO cells. Due to the ADP-ribosylation activity of LTA, CHO cells undergo a change in morphology (cell elongation) in the presence of LT or CT [24]. We first tested the biological activity of 50 ng native holotoxin on CHO cells. Fig. 3C (left) shows the average percentage of cell elongation with native WT (98%, textured bar), DP2 (84%, black bar), E7D (99%, dark gray bar) and A subunit only (1%, light gray bar). While the A subunit by itself cannot stimulate CHO cell elongation, the WT and E7D holotoxins showed comparable biological activity. The DP2 mutant showed 16% reduction in biological activity, which may be related to its defective GM1 ganglioside receptor binding. To test the effect of mutation on the efficacy of in vitro refolding, the mutant proteins were denatured as described in Methods. The efficacy of refolding/reassembly was tested by scoring the percentage of CHO cells showing elongated morphology. With the reconstituted WT (textured bars) and E7D (dark gray bars) holotoxins 76% and 78% of the cells displayed an elongated morphology, respectively, whereas only 24% of the cells were elongated in the presence of the DP2 (black bars) holotoxin ( Fig. 3C (right)). Considering that 25% protein was unable to assemble properly upon refolding as seen for the WT or E7D mutant, the DP2 mutant showed significant reduction in biological activity, indicating that the DP2 deletion mutant was unable to refold and assemble properly. This may have affected the conformation of the GM1 receptor binding pocket due to defects in the a1 helix resulting in defective biological activity, whereas the E7D substitution mutant refolded as efficiently as WT.
Discussion
Fig. 2. Circular dichroism spectra of LTB mutants. CD spectra of purified WT, DP2 and E7D proteins was obtained at 25 °C on Jasco-700 spectropolarimeter using 1 cm path length cuvette. Molar ellipticity is plotted against wave length (200–260 nm). Each spectrum represents an average of 10 scans.
In order to delineate the role of N-terminal on LTB export and biological activity in V. cholerae cells, four mutant constructs (DP2, DQ3, E7G and E7D) were generated and studied. The mutant LTB proteins were effectively exported to the periplasm of V. cholera. However, the mutant proteins DQ3 and E7G were poorly secreted into the culture medium. This is possible if the mutations changed the domain conformation that interacts with the secretory apparatus of V. cholerae which has not yet been fully characterized. Similar evidence was put forth by Connell et al. , where they demonstrated that E11K mutant of CTB is poorly secreted in V. cholerae [17]. The crystal structure of LTB shows that the N-terminal Ala1 forms a hydrogen bond with Thr92 of an adjacent monomer [9], presumably contributing to stablize pentamer formation. Deletion of Pro2 or Gln3 would shorten the length of the N-terminus and could disrupt this hydrogen bond, thus affecting pentamerization. However, the deletion mutants, DP2 or DQ3 do not show any
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
A
773
B
Concentration in nmole
Concentration in nmole
C
Fig. 3. GM1 ganglioside receptor binding activity of LTB mutants. Competitive GM1-ELISA was carried out in 10 mM PBS (A) or 80 mM PBS (B) to study receptor binding affinity of the mutant proteins. Equal amounts of biotinylated WT LTB subunit was competed with different dilutions of mutant LTB as described in Methods. Each dilution was performed in triplicate. Absorbance was read at A490 nm. Mean absorbance of each dilution was plotted against the concentration of competitor. (C) Biological activity of mutant protein. CHO cells were cultured in DMEM medium supplemented with 1% FCS. Native holotoxin (left) or in vitro refolded holotoxin (right) was added at 50 ng concentration and the change in morphology of the CHO cells was recorded after 16 h incubation. Percentage of cells showing elongation is represented in each bar. WT (textured bar), DP2 (black bar), E7D (dark gray bar) and A subunit (light gray bar).
defect in pentamer formation (Fig. 1D). Interestingly, the CD spectra of DP2 revealed change in the secondary structure ( Fig. 2), though this change may not be strong enough to destabilize the pentamer as stability is imparted by extensive interactions with the adjacent subunit by multiple hydrogen bonds, and by hydrophobic and salt-bridge interactions [9]. The DP2 mutant protein shows a weakened helicity and its reduced binding to GM1 receptor indicates that the change in the a1 helix conformation might have altered its GM1 receptor binding properties. Consistent with the in vitro GM1 receptor binding data, the DP2 holotoxin mutant also shows defective biological activity as compared to the WT holotoxin. The attenuation of biological activity suggests that the a1 helix not only maintains structural stability but also helps in maintaining the conformation of the GM1 ganglioside receptor binding pocket. The crystal structure shows that the receptor binding residues Try88, Asp90 and Lys91 lie on the loop region between b5 and b6 and contribute to the receptor binding pocket of LTB [10]. While the a1 helix is away from these residues and does not contribute directly to receptor binding activity, the exposed
hydrophobic surface of b5 and b6 is masked by it [9]. It has been shown that the disruption of disulfide bond by chemical modification abolishes GM1 receptor binding in CTB, suggesting that the proper orientation of the a1 helix is essential to maintain the overall conformation of the CTB/LTB pentamer and hence essential for the receptor binding pocket of the protein [25]. LTB has a net 10 positive charges at physiological pH and it can be precipitated by hexametaphosphate, suggesting that the negative charged phosphate group neutralizes the positive surface charge on the protein [9,20]. At 80 mM PBS, the negatively charged phosphate group may be able to neutralize some of the surface charges. The change in the a1 helix conformation due to the weak helicity displayed by the DP2 mutant may have exposed the buried region between b5 and b6 to negatively charge phosphate ions causing further change in protein conformation around receptor binding pocket and may have exacerbated the GM1 receptor binding activity, although the displacement of key water molecules contributing to receptor binding cannot be ruled out. Similarly, the weak a1 helix may not adapt proper conformation while
774
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
protein folding/assembly when attempting to refold in vitro in the presence of the LTA chain, thus showing a significant defect in biological activity on CHO cell line. The E7D mutant showed a gain in function by showing remarkably stable binding to the GM1 ganglioside even at higher ionic strength buffer, consistent with strengthening in helicity as revealed by CD spectra. Although the crystal structure suggests that the side chain of Glu7 makes polar contact with main chain Gln3 residue [9], it is still unclear why the E7D substitution should change the a1 helix property since the side chain still retains its negative charge. Given that aspartic acid is naturally present in CTB, it implies that Asp7 would be the best evolutionary adoption for LTB in order to improve GM1 receptor binding stability in a higher ionic strength buffer. Previous studies have suggested a progressive decrease in GM1 receptor binding affinity of CTB by the fusion of 8, 12, or 24 peptide residues of GtfB.1 to the N-terminal end [18]. This could be due to the peptide’s interference with the GM1 ganglioside receptor which functions by blocking the binding pocket. However, our data indicates that the a1 helix is playing a direct and important role in the function of LTB by providing structural stability and thus contributes to the receptor binding pocket’s conformational stability. Thus, the present investigation has demonstrated that the N-terminal a1 helix is essential for the structural stability and function of LTB, as it maintains the conformation of the GM1 ganglioside binding pocket. The residues at the a1 helix may contribute to maintain specific conformation of the N-terminal domain optimum to recognize secretory apparatus of V. chorelae. Acknowledgments CSIR-India is acknowledged for research fellowship to P.V.A. Financial support from the DBT-India is acknowledged. Technical assistance provided by Ram Bodh and Ved Prakesh. References [1] M. Jertborn, C. Ahren, J. Holmgren, A.M. Svennerholm, Safety and immunogenicity of an oral inactivated enterotoxigenic Escherichia coli vaccine, Vaccine 16 (1998) 255–260. [2] F. Qadri, A.I. Khan, A.S. Faruque, Y.A. Begum, F. Chowdhury, G.B. Nair, M.A. Salam, D.A. Sack, A.M. Svennerholm, Enterotoxigenic Escherichia coli and Vibrio cholerae diarrhea, Bangladesh, 2004, Emerg. Infect. Dis. 11 (2005) 1104–1107. [3] E. Takahashi, Z. Sultan, S. Shimada, W.W. Aung, M.M. Nyein, K.N. Oo, G.B. Nair, Y. Takeda, K. Okamoto, Studies on diarrheagenic Escherichia coli isolated from children with diarrhea in Myanmar, Microbiol. Immunol. 52 (2008) 2–8. [4] W.S. Dallas, D.M. Gill, S. Falkow, Cistrons encoding Escherichia coli heat-labile toxin, J. Bacteriol. 139 (1979) 850–858. [5] T.R. Hirst, S.J. Hardy, L.L. Randall, Assembly in vivo of enterotoxin from Escherichia coli: formation of the B subunit oligomer, J. Bacteriol. 153 (1983) 21–26.
[6] H. Hofstra, B. Witholt, Heat-labile enterotoxin in Escherichia coli. Kinetics of association of subunits into periplasmic holotoxin, J. Biol. Chem. 260 (1985) 16037–16044. [7] T.R. Hirst, J. Sanchez, J.B. Kaper, S.J. Hardy, J. Holmgren, Mechanism of toxin secretion by Vibrio cholerae investigated in strains harboring plasmids that encode heat-labile enterotoxins of Escherichia coli, Proc. Natl. Acad. Sci. USA 81 (1984) 7752–7756. [8] J.J. Guidry, L. Cardenas, E. Cheng, J.D. Clements, Role of receptor binding in toxicity, immunogenicity, and adjuvanticity of Escherichia coli heat-labile enterotoxin, Infect. Immun. 65 (1997) 4943–4950. [9] T.K. Sixma, K.H. Kalk, B.A. van Zanten, Z. Dauter, J. Kingma, B. Witholt, W.G. Hol, Refined structure of Escherichia coli heat-labile enterotoxin, a close relative of cholera toxin, J. Mol. Biol. 230 (1993) 890–918. [10] T.K. Sixma, S.E. Pronk, K.H. Kalk, B.A. van Zanten, A.M. Berghuis, W.G. Hol, Lactose binding to heat-labile enterotoxin revealed by X-ray crystallography, Nature 355 (1992) 561–564. [11] T.K. Sixma, S.E. Pronk, K.H. Kalk, E.S. Wartna, B.A. van Zanten, B. Witholt, W.G. Hol, Crystal structure of a cholera toxin-related heat-labile enterotoxin from E. coli, Nature 351 (1991) 371–377. [12] E.A. Merritt, T.K. Sixma, K.H. Kalk, B.A. van Zanten, W.G. Hol, Galactose-binding site in Escherichia coli heat-labile enterotoxin (LT) and cholera toxin (CT), Mol. Microbiol. 13 (1994) 745–753. [13] S.L. Griffiths, D.R. Critchley, Characterisation of the binding sites for Escherichia coli heat-labile toxin type I in intestinal brush borders, Biochim. Biophys. Acta 1075 (1991) 154–161. [14] J. Holmgren, I. Lonnroth, L. Svennerholm, Tissue receptor for cholera exotoxin: postulated structure from studies with GM1 ganglioside and related glycolipids, Infect. Immun. 8 (1973) 208–214. [15] M. Backstrom, V. Shahabi, S. Johansson, S. Teneberg, A. Kjellberg, H. MillerPodraza, J. Holmgren, M. Lebens, Structural basis for differential receptor binding of cholera and Escherichia coli heat-labile toxins: influence of heterologous amino acid substitutions in the cholera B-subunit, Mol. Microbiol. 24 (1997) 489–497. [16] P.V. Alone, G. Malik, A. Krishnan, L.C. Garg, Deletion mutations in N-terminal alpha1 helix render heat labile enterotoxin B subunit susceptible to degradation, Proc. Natl. Acad. Sci. USA 104 (2007) 16056–16061. [17] T.D. Connell, D.J. Metzger, M. Wang, M.G. Jobling, R.K. Holmes, Initial studies of the structural signal for extracellular transport of cholera toxin and other proteins recognized by Vibrio cholerae, Infect. Immun. 63 (1995) 4091–4098. [18] M.T. Dertzbaugh, C.O. Elson, Reduction in oral immunogenicity of cholera toxin B subunit by N-terminal peptide addition, Infect. Immun. 61 (1993) 384–390. [19] A. Ghorpade, L.C. Garg, Efficient expression, processing and secretion of a biologically active mammalian protein by Vibrio cholerae, FEBS Lett. 387 (1996) 167–170. [20] J.J. Mekalanos, R.J. Collier, W.R. Romig, Purification of cholera toxin and its subunits: new methods of preparation and the use of hypertoxinogenic mutants, Infect. Immun. 20 (1978) 552–558. [21] Y. Takeda, T. Honda, S. Taga, T. Miwatani, In vitro formation of hybrid toxins between subunits of Escherichia coli heat-labile enterotoxin and those of cholera enterotoxin, Infect. Immun. 34 (1981) 341–346. [22] G.D. Fasman, Predication of Protein Structure and Principles of Protein Conformation, Plenum Press, New York, 1990. [23] R.J. Neill, B.E. Ivins, R.K. Holmes, Synthesis and secretion of the plasmid-coded heat-labile enterotoxin of Escherichia coli in Vibrio cholerae, Science 221 (1983) 289–291. [24] R.L. Guerrant, L.L. Brunton, T.C. Schnaitman, L.I. Rebhun, A.G. Gilman, Cyclic adenosine monophosphate and alteration of Chinese hamster ovary cell morphology: a rapid, sensitive in vitro assay for the enterotoxins of Vibrio cholerae and Escherichia coli, Infect. Immun. 10 (1974) 320–327. [25] D.S. Ludwig, R.K. Holmes, G.K. Schoolnik, Chemical and immunochemical studies on the receptor binding domain of cholera toxin B subunit, J. Biol. Chem. 260 (1985) 12528–12534.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Secretory and GM1 receptor binding role of N-terminal region of LTB in Vibrio cholerae Pankaj V. Alone 1, Lalit C. Garg * Gene Regulation Laboratory, National Institute of Immunology, Aruna Asaf Ali Marg, New Delhi, Delhi 110067, India
a r t i c l e
i n f o
Article history: Received 1 September 2008 Available online 22 September 2008
Keywords: Heat labile enterotoxin Cholera toxin Secretion Periplasm CD spectra a1 helix
a b s t r a c t Heat labile enterotoxin from enterotoxigenic Escherichia coli is similar to cholera toxin (CT) and is a leading cause of diarrhea in developing countries. It consists of an enzymatically active A subunit (LTA) and a carrier pentameric B subunit (LTB). In the current study, we evaluated the importance of the N-terminal region of LTB by mutation analysis. Deletion of the glutamine (DQ3) residue and a substitution mutation E7G in the a1 helix region led to defects in LTB protein secretion. Deletion of the proline residue (DP2) caused a decrease in a helicity. The DP2 mutant affected GM1 ganglioside receptor binding activity without affecting LTB pentamer formation. Upon refolding/reassembly, the DP2 mutant showed defective biological activity. The single substitution mutation (E7D) strengthened the helix, imparting structural stability and thereby improved the GM1 ganglioside receptor binding activity. Our results demonstrate the important role of N-terminal a1 helix in maintaining the structural stability and the integrity of GM1 ganglioside receptor binding activity. Ó 2008 Elsevier Inc. All rights reserved.
Heat labile enterotoxin from enterotoxigenic Escherichia coli (ETEC) is the causative agent of diarrhea, a leading cause of mortality in developing countries. Nearly 800,000 deaths, mainly in children, are reported every year [1–3]. The enterotoxin is a heterohexameric protein consisting of an enzymatically active A subunit (LTA) and a carrier pentameric B subunit (LTB). The amino acid sequence of both LTA and LTB is more than 80% similar to cholera toxin (CT). The genes encoding the two LT subunits, eltA and eltB, are plasmid borne and are transcribed as a single polycistronic mRNA [4]. Both subunits are synthesized and exported to the periplasm in E. coli, where the B subunit during the process of pentamerization binds to the A subunit to form the holotoxin [5,6]. The expression of the B subunit of either CT or LT in engineered AB+ strain of Vibrio cholerae resulted in effective secretion of B subunit as a pentamer into the medium, although in engineered A+B strain of V. cholerae, the A subunit remained cell associated. This implies that both the subunits are assembled before being secreted and that the secretory machinery recognizes the structural domains within the B subunit in order to secrete them [7]. LTB is an important mucosal adjuvant and shows great promise for developing oral vaccines by coupling immunogenic peptides to its B subunit. The adjuvant activity of LTB is strongly linked to its
* Corresponding author. Fax: +91 11 26162125. E-mail addresses: [email protected], [email protected] (L.C. Garg). 1 Present address: University of California-Davis, School of Medicine, Davis, CA 95616, USA. 0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2008.09.066
receptor binding activity [8]. Recognition of the oligosaccharide portion of GM1 on the intestinal epithelial cell membrane is the first crucial step in the pathogenicity of LT or CT. The crystal structure of LT with lactose shows involvement of Glu51, Gln61, Ans90, and Lys91 residues in hydrogen bonding with lactose from the B monomer. Trp88 and Gly33 from the adjacent monomer (N + 1) are a part of the ganglioside binding cavity [9–11]. All the residues involved in GM1 binding have been found to be conserved in both LT and CT [12]. Unlike CT, which binds only to GM1 ganglioside receptors, LT binds to a wide variety of glycolipid and glycoprotein receptors [13,14]. LTs ability to recognize additional receptors may be due to differences in the primary amino acid sequence between LT and CT. A hybrid CT toxin made by swapping N-terminal 1–25 amino acids along with residues 75–83 with those of LT exhibited strong binding to intestinal glycosphingolipids [15]. The N-terminal end of LTB contains a a1 helix which is linked with the b5 strand by Cys9 and Cys86 disulfide bond and is not part of the core structure [9–11]. We have recently shown that the a1 helix imparts stability to the LTB, as deletion mutations within the a1 helix render LTB susceptible to degradation [16]. CTB is poorly secreted in V. cholera when E11K mutation is introduced at the N-terminal end [17]. Moreover, the fusion of 8, 12, or 24 residue peptides from GtfB.1 to the N-terminal end of CTB caused a progressive decrease in GM1 receptor binding affinity [18]. This suggests that the N-terminal end of LT and CT plays an important role in their functions and warrants further investigation. In our present work, we delineate mutations at the N-terminal end of LTB which are involved in its secretion and receptor binding.
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
Methods Conjugal transfer of mutant construct to V. cholerae JBK70 cells. Deletion of Proline 2 (DP2), deletion of glutamine 3 (DQ3), substitution of glycine for glutamic acid 7 (E7G) and substitution of aspartic acid for glutamic acid 7 (E7D) mutant constructs were designed as described in Alone et al. [16]. Mutant and wild type plasmids were conjugally transferred from DH5a cells to JBK70 (polyB+, AB) cells using E. coli pRK2013 cells as a helper strain [19]. Protein purification, reconstitution of holotoxin, and evaluation of its biological activity. The holotoxin as well as mutant and wild type (WT) LTB proteins were over-expressed into the supernatant from induced V. cholerae cultures and purified to homogeneity as described by Mekalonos et al. [20]. The purified holotoxin was either directly tested for biological activity on CHO cell line or was separated into the A and B subunits as described by Takeda et al. [21] before being tested for biological activity. The WT A subunit and the mutant B subunits were reconstituted in vitro by continuous dialysis with 10 mM Tris–Cl (pH 7.4) and 20 mM NaCl after denaturation in 6 M urea and 0.1 M propionic acid (pH 4.0) solution [22]. CHO (2 104) cells/ml were suspended in DMEM supplemented with 1% FCS and seeded into 24-well tissue culture plates. Fifty nanograms of in vitro reconstituted mutant and wild type holotoxin was added into wells and the change in morphology of CHO cells was observed under the microscope after a 16 h incubation at 37 °C with 5% CO2. Circular dichorism (CD) spectroscopy. CD spectra were obtained using a Jasco-700 spectropolarimeter at 25 °C. A protein concentration of 80 lM was maintained for CD spectra analysis. The cuvette path length was 1 cm for far-UV region measurements. Each sample was scanned 10 times and the average spectra of molar ellipticity were plotted against wavelength. GM1 ganglioside receptor binding activity by competitive-ELISA. Hundred microliters of GM1 (1 lg/ml) prepared in 50 mM carbonate buffer, pH 9.2, was coated on polypropylene ELISA plates. Nonspecific sites were saturated with 1% BSA in PBS. Biotinylated WT
A
C
771
LTB protein was mixed with different dilutions of mutant or WT LTB protein. The mixtures were added to GM1 coated plates and incubated for 1 h at 37 °C either in 10 mM or 80 mM PBS. The unbound protein was removed by washing five times with 0.05% Tween 20 in PBS. Hundred microliters of Streptavidin–HRP conjugate (1 U/well) was added and incubated for an additional 1 h. The unbound protein was washed five times with 0.05% Tween 20 in PBS, and then mixed with 0.05% orthophenylene diamine (OPD) in citrate phosphate buffer, pH 5.5, along with hydrogen peroxide (1 ll/ml). The reaction was terminated using 50 ll of 5 N H2SO4 and quantified by monitoring at A490. Results Effect of N-terminal a1 helix mutation on LTB secretion and pentamer formation Previously we reported that the deletion of DS4, DI5, DT6, DE7 residues in a1 helix causes the loss of LTB expression, due to its susceptibility to degradation [16]. In order to study the effect of mutation in the N-terminal a1 helix, the residues P2 and Q3, which lie at N00 and N0 -cap region of a1 helix, respectively, were deleted, while residue E7 was substituted with glycine or aspartic acid. The only charged residue E7 in the a1 helix was substituted with glycine (E7G) to destabilize the a1 helix, as the glycine residue has been shown to be a helix breaker [22]. The E7D substitution served as a positive control as the residue is naturally present in the CTB. Western blot analysis of the total cell extract using anti-LTB antibody showed expression of the mutant as well as WT LTB in V. cholerae cells (Fig. 1A). The mutant proteins were successfully exported into the periplasm of V. cholerae cells (Fig. 1B). It is known that upon assembly in the periplasm, the LTB pentamer can be secreted into the culture supernatant from V. cholerae cells [23]. The culture supernatant from the four mutants was analyzed for the presence of LTB protein (Fig. 1C) and it was found that the secretion of mutant LTB proteins DQ3 (lane 4) and E7G (lane 5) into the culture supernatant was severely curtailed as compared to WT (lane 1) and E7D mutant (lane 6).
B
D
Fig. 1. Analysis of LTB N-terminal a1 helix mutations on its secretion and pentamerization in Vibrio cholerae. V. cholerae (JBK 70) cells harboring WT or mutant LTB constructs were induced with 1 mM IPTG at A600 = 0.6 for 5 h. Total cell extract (A), periplasmic fraction (B) and culture supernatant (C) were analyzed by Western blotting using antiLTB polyclonal antibody. Lane 1, WT (+); lane 2, WT (); lane 3, DP2 (+); lane 4, DQ3 (+); lane 5, E7G (+); lane 6, E7D (+). + and denote cells grown in presence or absence of IPTG, respectively. Arrows indicates migration of LTB protein. Molecular weight in kDa is shown on the left. (D) Mutant LTB proteins were purified from V. cholerae culture supernatant as described in Methods. The non-denatured un-boiled purified proteins were resolved on a 15% SDS–PAGE. Lane 1, WT; lane 2, E7D; lane 3, DP2; lane 4, E7G and lane 5, DQ3. Arrow indicates migration of pentameric LTB. Molecular weight in kDa is shown on the left.
772
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
To test the impact of LTB mutations on pentamer formation, the mutant proteins were purified from V. cholerae culture supernatants and resolved on a 15% native SDS–PAGE (Fig. 1D). As compared to the WT, none of the mutations had an effect on pentamer formation. Since the mutants, DQ3 and E7G, have defective secretion, ample quantity of these proteins could not be purified to enable their further characterization by CD spectra and GM1 ganglioside binding activity. Mutations in the a1 helix of LTB cause changes in the secondary structure The CD spectra for DP2, E7D, and WT LTB were analyzed and are compared in Fig. 2. As compared to the WT LTB, a positive shift at [h]218 nm was observed in the DP2 mutant, indicating a change in its secondary structure, probably due to a weakening of the helicity of the protein. It has been observed consistently that proline prefers to stay at the N-cap position of helixes and is noted to promote helix formation [22]. In contrast to the DP2 mutant, the CD spectra of E7D showed a negative shift at [h]218 nm as compared to WT, suggesting strengthening in helicity. These results indicated that the deletion of Pro2 at the N-cap region of the a1 helix, effects larger changes on the secondary structure than does the chemically similar substitution mutation, E7D, in the helix. Analysis of GM1 ganglioside receptor binding activity of LTB mutants by competitive GM1-ELISA Pentameric LTB binds to the GM1 ganglioside receptor with strong affinity. To check if the mutations in the a1 helix have a direct effect on GM1 receptor binding activity, we employed a competitive GM1 binding-ELISA. Increasing concentrations of mutant proteins were incubated with a constant amount of biotinylated WT LTB protein and the mixtures were then reacted with immobilized GM1 ganglioside in 10 mM PBS. As shown in Fig. 3A, the LTA subunit by itself does not bind to GM1 ganglioside. At 10 mM PBS, the E7D mutant protein showed comparable affinity to that of WT LTB, while the
affinity for DP2 was reduced by 30%. However, when the ionic strength of the incubation buffer (PBS) was increased to 80 mM, a decrease in the affinity of WT and DP2 as compared to the E7D mutant protein was observed (Fig. 3B). The weakened helicity observed in the CD spectra (Fig. 2) and a consistent decrease in GM1 ganglioside binding affinity associated with DP2 mutant suggests that deleting Pro2 has a profound effect on the structural stability of the a1 helix, which may influence the conformation of the GM1 binding pocket and impair B subunit binding to GM1 ganglioside. In contrast, the E7D mutant has a more stable a1 helix as judged by CD spectra (Fig. 2) and also shows stable binding affinity to GM1 ganglioside receptor even at a higher ionic concentration. Defective biological activity of LTB DP2 mutant following in vitro refolding LTB serves as a carrier molecule for the catalytically active A subunit, and shows strong binding affinity for GM1 ganaglioside receptors on CHO cells. Due to the ADP-ribosylation activity of LTA, CHO cells undergo a change in morphology (cell elongation) in the presence of LT or CT [24]. We first tested the biological activity of 50 ng native holotoxin on CHO cells. Fig. 3C (left) shows the average percentage of cell elongation with native WT (98%, textured bar), DP2 (84%, black bar), E7D (99%, dark gray bar) and A subunit only (1%, light gray bar). While the A subunit by itself cannot stimulate CHO cell elongation, the WT and E7D holotoxins showed comparable biological activity. The DP2 mutant showed 16% reduction in biological activity, which may be related to its defective GM1 ganglioside receptor binding. To test the effect of mutation on the efficacy of in vitro refolding, the mutant proteins were denatured as described in Methods. The efficacy of refolding/reassembly was tested by scoring the percentage of CHO cells showing elongated morphology. With the reconstituted WT (textured bars) and E7D (dark gray bars) holotoxins 76% and 78% of the cells displayed an elongated morphology, respectively, whereas only 24% of the cells were elongated in the presence of the DP2 (black bars) holotoxin ( Fig. 3C (right)). Considering that 25% protein was unable to assemble properly upon refolding as seen for the WT or E7D mutant, the DP2 mutant showed significant reduction in biological activity, indicating that the DP2 deletion mutant was unable to refold and assemble properly. This may have affected the conformation of the GM1 receptor binding pocket due to defects in the a1 helix resulting in defective biological activity, whereas the E7D substitution mutant refolded as efficiently as WT.
Discussion
Fig. 2. Circular dichroism spectra of LTB mutants. CD spectra of purified WT, DP2 and E7D proteins was obtained at 25 °C on Jasco-700 spectropolarimeter using 1 cm path length cuvette. Molar ellipticity is plotted against wave length (200–260 nm). Each spectrum represents an average of 10 scans.
In order to delineate the role of N-terminal on LTB export and biological activity in V. cholerae cells, four mutant constructs (DP2, DQ3, E7G and E7D) were generated and studied. The mutant LTB proteins were effectively exported to the periplasm of V. cholera. However, the mutant proteins DQ3 and E7G were poorly secreted into the culture medium. This is possible if the mutations changed the domain conformation that interacts with the secretory apparatus of V. cholerae which has not yet been fully characterized. Similar evidence was put forth by Connell et al. , where they demonstrated that E11K mutant of CTB is poorly secreted in V. cholerae [17]. The crystal structure of LTB shows that the N-terminal Ala1 forms a hydrogen bond with Thr92 of an adjacent monomer [9], presumably contributing to stablize pentamer formation. Deletion of Pro2 or Gln3 would shorten the length of the N-terminus and could disrupt this hydrogen bond, thus affecting pentamerization. However, the deletion mutants, DP2 or DQ3 do not show any
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
A
773
B
Concentration in nmole
Concentration in nmole
C
Fig. 3. GM1 ganglioside receptor binding activity of LTB mutants. Competitive GM1-ELISA was carried out in 10 mM PBS (A) or 80 mM PBS (B) to study receptor binding affinity of the mutant proteins. Equal amounts of biotinylated WT LTB subunit was competed with different dilutions of mutant LTB as described in Methods. Each dilution was performed in triplicate. Absorbance was read at A490 nm. Mean absorbance of each dilution was plotted against the concentration of competitor. (C) Biological activity of mutant protein. CHO cells were cultured in DMEM medium supplemented with 1% FCS. Native holotoxin (left) or in vitro refolded holotoxin (right) was added at 50 ng concentration and the change in morphology of the CHO cells was recorded after 16 h incubation. Percentage of cells showing elongation is represented in each bar. WT (textured bar), DP2 (black bar), E7D (dark gray bar) and A subunit (light gray bar).
defect in pentamer formation (Fig. 1D). Interestingly, the CD spectra of DP2 revealed change in the secondary structure ( Fig. 2), though this change may not be strong enough to destabilize the pentamer as stability is imparted by extensive interactions with the adjacent subunit by multiple hydrogen bonds, and by hydrophobic and salt-bridge interactions [9]. The DP2 mutant protein shows a weakened helicity and its reduced binding to GM1 receptor indicates that the change in the a1 helix conformation might have altered its GM1 receptor binding properties. Consistent with the in vitro GM1 receptor binding data, the DP2 holotoxin mutant also shows defective biological activity as compared to the WT holotoxin. The attenuation of biological activity suggests that the a1 helix not only maintains structural stability but also helps in maintaining the conformation of the GM1 ganglioside receptor binding pocket. The crystal structure shows that the receptor binding residues Try88, Asp90 and Lys91 lie on the loop region between b5 and b6 and contribute to the receptor binding pocket of LTB [10]. While the a1 helix is away from these residues and does not contribute directly to receptor binding activity, the exposed
hydrophobic surface of b5 and b6 is masked by it [9]. It has been shown that the disruption of disulfide bond by chemical modification abolishes GM1 receptor binding in CTB, suggesting that the proper orientation of the a1 helix is essential to maintain the overall conformation of the CTB/LTB pentamer and hence essential for the receptor binding pocket of the protein [25]. LTB has a net 10 positive charges at physiological pH and it can be precipitated by hexametaphosphate, suggesting that the negative charged phosphate group neutralizes the positive surface charge on the protein [9,20]. At 80 mM PBS, the negatively charged phosphate group may be able to neutralize some of the surface charges. The change in the a1 helix conformation due to the weak helicity displayed by the DP2 mutant may have exposed the buried region between b5 and b6 to negatively charge phosphate ions causing further change in protein conformation around receptor binding pocket and may have exacerbated the GM1 receptor binding activity, although the displacement of key water molecules contributing to receptor binding cannot be ruled out. Similarly, the weak a1 helix may not adapt proper conformation while
774
P.V. Alone, L.C. Garg / Biochemical and Biophysical Research Communications 376 (2008) 770–774
protein folding/assembly when attempting to refold in vitro in the presence of the LTA chain, thus showing a significant defect in biological activity on CHO cell line. The E7D mutant showed a gain in function by showing remarkably stable binding to the GM1 ganglioside even at higher ionic strength buffer, consistent with strengthening in helicity as revealed by CD spectra. Although the crystal structure suggests that the side chain of Glu7 makes polar contact with main chain Gln3 residue [9], it is still unclear why the E7D substitution should change the a1 helix property since the side chain still retains its negative charge. Given that aspartic acid is naturally present in CTB, it implies that Asp7 would be the best evolutionary adoption for LTB in order to improve GM1 receptor binding stability in a higher ionic strength buffer. Previous studies have suggested a progressive decrease in GM1 receptor binding affinity of CTB by the fusion of 8, 12, or 24 peptide residues of GtfB.1 to the N-terminal end [18]. This could be due to the peptide’s interference with the GM1 ganglioside receptor which functions by blocking the binding pocket. However, our data indicates that the a1 helix is playing a direct and important role in the function of LTB by providing structural stability and thus contributes to the receptor binding pocket’s conformational stability. Thus, the present investigation has demonstrated that the N-terminal a1 helix is essential for the structural stability and function of LTB, as it maintains the conformation of the GM1 ganglioside binding pocket. The residues at the a1 helix may contribute to maintain specific conformation of the N-terminal domain optimum to recognize secretory apparatus of V. chorelae. Acknowledgments CSIR-India is acknowledged for research fellowship to P.V.A. Financial support from the DBT-India is acknowledged. Technical assistance provided by Ram Bodh and Ved Prakesh. References [1] M. Jertborn, C. Ahren, J. Holmgren, A.M. Svennerholm, Safety and immunogenicity of an oral inactivated enterotoxigenic Escherichia coli vaccine, Vaccine 16 (1998) 255–260. [2] F. Qadri, A.I. Khan, A.S. Faruque, Y.A. Begum, F. Chowdhury, G.B. Nair, M.A. Salam, D.A. Sack, A.M. Svennerholm, Enterotoxigenic Escherichia coli and Vibrio cholerae diarrhea, Bangladesh, 2004, Emerg. Infect. Dis. 11 (2005) 1104–1107. [3] E. Takahashi, Z. Sultan, S. Shimada, W.W. Aung, M.M. Nyein, K.N. Oo, G.B. Nair, Y. Takeda, K. Okamoto, Studies on diarrheagenic Escherichia coli isolated from children with diarrhea in Myanmar, Microbiol. Immunol. 52 (2008) 2–8. [4] W.S. Dallas, D.M. Gill, S. Falkow, Cistrons encoding Escherichia coli heat-labile toxin, J. Bacteriol. 139 (1979) 850–858. [5] T.R. Hirst, S.J. Hardy, L.L. Randall, Assembly in vivo of enterotoxin from Escherichia coli: formation of the B subunit oligomer, J. Bacteriol. 153 (1983) 21–26.
[6] H. Hofstra, B. Witholt, Heat-labile enterotoxin in Escherichia coli. Kinetics of association of subunits into periplasmic holotoxin, J. Biol. Chem. 260 (1985) 16037–16044. [7] T.R. Hirst, J. Sanchez, J.B. Kaper, S.J. Hardy, J. Holmgren, Mechanism of toxin secretion by Vibrio cholerae investigated in strains harboring plasmids that encode heat-labile enterotoxins of Escherichia coli, Proc. Natl. Acad. Sci. USA 81 (1984) 7752–7756. [8] J.J. Guidry, L. Cardenas, E. Cheng, J.D. Clements, Role of receptor binding in toxicity, immunogenicity, and adjuvanticity of Escherichia coli heat-labile enterotoxin, Infect. Immun. 65 (1997) 4943–4950. [9] T.K. Sixma, K.H. Kalk, B.A. van Zanten, Z. Dauter, J. Kingma, B. Witholt, W.G. Hol, Refined structure of Escherichia coli heat-labile enterotoxin, a close relative of cholera toxin, J. Mol. Biol. 230 (1993) 890–918. [10] T.K. Sixma, S.E. Pronk, K.H. Kalk, B.A. van Zanten, A.M. Berghuis, W.G. Hol, Lactose binding to heat-labile enterotoxin revealed by X-ray crystallography, Nature 355 (1992) 561–564. [11] T.K. Sixma, S.E. Pronk, K.H. Kalk, E.S. Wartna, B.A. van Zanten, B. Witholt, W.G. Hol, Crystal structure of a cholera toxin-related heat-labile enterotoxin from E. coli, Nature 351 (1991) 371–377. [12] E.A. Merritt, T.K. Sixma, K.H. Kalk, B.A. van Zanten, W.G. Hol, Galactose-binding site in Escherichia coli heat-labile enterotoxin (LT) and cholera toxin (CT), Mol. Microbiol. 13 (1994) 745–753. [13] S.L. Griffiths, D.R. Critchley, Characterisation of the binding sites for Escherichia coli heat-labile toxin type I in intestinal brush borders, Biochim. Biophys. Acta 1075 (1991) 154–161. [14] J. Holmgren, I. Lonnroth, L. Svennerholm, Tissue receptor for cholera exotoxin: postulated structure from studies with GM1 ganglioside and related glycolipids, Infect. Immun. 8 (1973) 208–214. [15] M. Backstrom, V. Shahabi, S. Johansson, S. Teneberg, A. Kjellberg, H. MillerPodraza, J. Holmgren, M. Lebens, Structural basis for differential receptor binding of cholera and Escherichia coli heat-labile toxins: influence of heterologous amino acid substitutions in the cholera B-subunit, Mol. Microbiol. 24 (1997) 489–497. [16] P.V. Alone, G. Malik, A. Krishnan, L.C. Garg, Deletion mutations in N-terminal alpha1 helix render heat labile enterotoxin B subunit susceptible to degradation, Proc. Natl. Acad. Sci. USA 104 (2007) 16056–16061. [17] T.D. Connell, D.J. Metzger, M. Wang, M.G. Jobling, R.K. Holmes, Initial studies of the structural signal for extracellular transport of cholera toxin and other proteins recognized by Vibrio cholerae, Infect. Immun. 63 (1995) 4091–4098. [18] M.T. Dertzbaugh, C.O. Elson, Reduction in oral immunogenicity of cholera toxin B subunit by N-terminal peptide addition, Infect. Immun. 61 (1993) 384–390. [19] A. Ghorpade, L.C. Garg, Efficient expression, processing and secretion of a biologically active mammalian protein by Vibrio cholerae, FEBS Lett. 387 (1996) 167–170. [20] J.J. Mekalanos, R.J. Collier, W.R. Romig, Purification of cholera toxin and its subunits: new methods of preparation and the use of hypertoxinogenic mutants, Infect. Immun. 20 (1978) 552–558. [21] Y. Takeda, T. Honda, S. Taga, T. Miwatani, In vitro formation of hybrid toxins between subunits of Escherichia coli heat-labile enterotoxin and those of cholera enterotoxin, Infect. Immun. 34 (1981) 341–346. [22] G.D. Fasman, Predication of Protein Structure and Principles of Protein Conformation, Plenum Press, New York, 1990. [23] R.J. Neill, B.E. Ivins, R.K. Holmes, Synthesis and secretion of the plasmid-coded heat-labile enterotoxin of Escherichia coli in Vibrio cholerae, Science 221 (1983) 289–291. [24] R.L. Guerrant, L.L. Brunton, T.C. Schnaitman, L.I. Rebhun, A.G. Gilman, Cyclic adenosine monophosphate and alteration of Chinese hamster ovary cell morphology: a rapid, sensitive in vitro assay for the enterotoxins of Vibrio cholerae and Escherichia coli, Infect. Immun. 10 (1974) 320–327. [25] D.S. Ludwig, R.K. Holmes, G.K. Schoolnik, Chemical and immunochemical studies on the receptor binding domain of cholera toxin B subunit, J. Biol. Chem. 260 (1985) 12528–12534.