• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Structural Biology and Structure-Based Inhibitor Design of Cholera Toxin and Heat-Labile Enterotoxin

Posted on: Thursday, 18 November 2004, 03:00 CST

Abstract

Structural biology studies on cholera toxin and the closely related heat-labile enterotoxin from enterotoxigenic Escherichia coli over the past decade have shed light on the mechanism of toxin action at molecular and atomic levels. Also, components of the extracellular protein secretion apparatus that translocate the toxins across the outer membrane are being investigated. At the same time, structure-based design has led to various classes of compounds targeting different toxin sites, including highly potent multivalent inhibitors that block the toxin receptor-binding process.

2004 Elsevier GmbH. All rights reserved.

Keywords: Cholera toxin; Heat-labile enterotoxin; Structure- based drug design; Type II secretion system; Multivalent inhibitor

Contents

Contents

Introduction

Cholera toxin (CT) from Vibrio cholerae and heat-labile enterotoxin (LT) from enterotoxigenic Escherichia coli (ETEC) are two closely related heterohexameric AB5 toxins (Spangler, 1992; Merritt and Hol, 1995). CT is the causative agent of the life- threatening disease cholera, while LT causes infant and travelers' diarrhea. Together, CT and LT are responsible for thousands of deaths each year worldwide, especially among young children in developing countries (Widdus, 2001; WHO, 2002). There are currently no effective prophylactics to prevent these toxin-caused diarrhea, nor vaccines with long-lasting protective effects. Although oral rehydration can significantly reduce the fatality rate, it is very labor intensive and requires supplies of clean water. This makes the task of disease control difficult because cholera outbreaks usually occur in areas with contaminated water sources. Therefore, development of effective therapeutics against CT- and LT-caused diarrhea would be significantly beneficial worldwide.

CT and LT are structurally and functionally very similar toxins, sharing over 80% sequence identity in both A and the B subunits. The toxin A and B peptide chains are encoded by two genes that overlap by one base, and are synthesized in the cytosol. Then both the A and B subunits of the toxin are translocated across the inner membrane into the periplasm of the respective bacterium, where one A subunit and five B subunits assemble into one hetero-hexameric holotoxin. The toxins are actively secreted across the outer membrane by the extracellular protein secretion (Eps) and the general secretion pathway (Gsp) export apparatuses of V. cholerae and ETEC, respectively, which are examples of the type II bacterial protein secretion system (Sandkvist, 2001; Tauschek et al., 2002). Once in the lumen of the gastrointestinal tract of the human host, the B pentamer portion of the toxins recognizes the receptor ganglioside GM1 presented on the host cell surface and triggers endocytosis. Trafficking of the toxins inside the host cell is a fascinatingly complicated and not yet fully understood process (Lencer et al., 1995; Kahn et al., 2002; Lencer and Tsai, 2003). After translocation across the ER membrane, the enzymatic Al fragment of the A subunit of CT or LT enters the cytosol, where it modifies the alpha subunit of the stimulatory G protein (G^sub sα^) through an NAD- dependent ADP-ribosylation reaction. ADP ribosylation of Arg201 of G^sub sα^ locks the G protein in its GTP-bound form, which continually stimulates adenylate cyclase to produce cAMP. The resulting elevated levels of cAMP cause dramatic efflux of ions and water from the host, leading to watery diarrhea.

In this review, we briefly summarize our work over the past 12 years on solving the crystal structures of CT and LT in many variant forms, as well as the machineries that are associated with toxin trafficking. Using structural biology, we aim to provide insights into the mechanisms of action of the two toxins at the molecular level. At the same time, our research will provide a structural basis for the design of highly potent inhibitors that can block toxin action in various ways. These include inhibition of the ADP- ribosylation reaction catalyzed by the A subunit, blocking the assembly of the holotoxin, as well as preventing the receptor- binding process of CT and LT. We hope that the efforts in structural biology and structure-based drug design will eventually lead to novel therapeutics for the treatment, or more likely, prevention of CT- and LT-caused diarrhea.

The holotoxin structure

The crystal structure of the LT holotoxin was initially reported in 1991 (Sixma et al., 1991a, b). It allowed for the first time a close examination of the unique AB^sub 5^ architecture of this class of toxins at the molecular level (Fig. 1). The toxin contains a symmetrical pentamer of five identical B subunits surrounding a central pore. The A subunit is composed of two functional parts: a wedge-shaped A1 domain with the enzyme-active site and an elongated A2 domain consisting of an α-helix plus a "tail" that extends through the pore formed by the B pentamer. The interactions between the A and B subunits are almost entirely mediated by the A2 domain and the B pentamer pore. These initial structures of the Al domain represent an inactive enzyme, since a fully active toxin A1 domain requires severing of the covalent connections between A1 and A2, which is accomplished by cleavage of the peptide chain and reduction of a disulfide bond between these two entities (Mekalanos et al., 1979). The structure of LT in complex with lactose (Sixma et al., 1992) revealed that the GM1-binding sites of the toxin are located at the "bottom" of the B pentamer, opposing the A subunit (Fig. 1). The three-dimensional architecture of this LT holotoxin was later found to be conserved among a much larger family of AB^sub 5^ bacterial toxins (Merritt and Hol, 1995). The holotoxin structure also points to three potential target areas for inhibitor design (Fig. 1): blocking the enzyme-active site located at the Al domain; inhibiting holotoxin assembly by interrupting A2-B pentamer interactions; and preventing receptor binding to the bottom of the B pentamer.

Fig. 1. A snapshot of the holotoxin, with key sites for potential drug design indicated.

The A-subimit: mechanism of activation of enzymatic activity

Ever since the first LT holotoxin structure was elucidated, it has been a challenge to unravel the mechanism for the ADP- ribosylation activity of the A subunit of CT and LT. Because only the inactive form of the wild-type holotoxin could be crystallized, it was apparent that there was no room for the substrates to be accommodated in the active-site cleft conformation seen in the LT holotoxin structures (Sixma et al., 1991a, b, 1993). Attempts to co- crystallize the holotoxin with its substrates or substrate analogs in hope to reveal an altered binding-site conformation also failed (van den Akker et al., unpublished results). However, we have determined that proteolytic cleavage of the A subunit between residues 192 and 195, known to be required for activation of the toxin, does not by itself induce a conformational change (Merritt et al., 1994a). This is in agreement with the two-step activation (nicking and reduction) requirement/mechanism observed for CT and LT (Mekalanos et al., 1979).

Structural studies of various A subunit mutants of LT provided more indirect evidence that eventually led to a partial model for toxin activation and catalytic activity. Of particular importance is the structure of the R7K mutant. This mutant structure is essentially identical to wild-type LT, except that in the R7K structure the A:47-56 loop is mobile (van den Akker et al., 1995). Residues in this loop were subjected to mutagenesis experiments which demonstrated the importance of this loop in catalysis (Feil et al., 1996). These findings led van den Akker et al. (1995) to propose a series of conformational changes that would lead to Al activation, starting with proteolytic nicking and disulfide bond reduction at a location ~20 [Angstrom] away from essential catalytic residues. This hypothesis suggested that several key residues distant from the active site might be crucial for the activation pathway. It has inspired the group of Dr. Randall Holmes in Denver to carry out extensive mutagenesis studies leading to the discovery of a constitutively active variant of CT. The structure of this variant holotoxin, together with a redelcrmination of the wild-type CT structure, provides a new platform for furthering our understanding of the toxin activation process (O'Neal et al., 2004).

Towards inhibition of toxin assembly

The CT and LT holotoxins are assembled in the periplasm of V. cholcrae and ETEC by pentamerizalion of B subunits around the C- terminal tail of the A2 domain of the A subunit, although B pentamers can form in the absence of A subunit. More importantly, holotoxin cannot be assembled from preformed B pentamers with the A subunit without an unfolding and refolding process (Hardy et al., 1988). If one could design a small molecule that can enter the bacterial periplasm and bind to the pore region of the B subunits during assembly, it would compete with the A subunit for this site and thereby inhibit holotoxin formation, regardless whether the compound promotes or inhibits B pentamer formation. Direct interactions between the A subunit and the B pentamer are remar\kably sparse in LT and CT, making assembly inhibition an attractive strategy. Our analysis of several AB^sub 5^ holotoxin structures suggested that a hydrophobic area encircling the top of the B pentamer pore might play a role in positioning the A subunit during holotoxin assembly (van den Akker et al., 1996). In this region, the side chains of F223 and Y226 of the A subunit interact with the B pentamer pore. Using the positions and orientations of F223 and Y226 as a starting pharmacophore, we searched databases of commercially available chemicals and identified potential candidate inhibitors. One of the candidate compounds, 3-methylthio-1,4- diphenyl-1H-1,3,4-triazolium (MDT), co-crystallized with the LT B pentamer (Hovey et al., 1999). Structural analysis indicated that not less than three MDTs bind in the pore of toxin B pentamer more or less as targeted. Therefore, future structure-based designs of "pore-plug" inhibitors can be based on the MDT co-crystal structure. In addition, a reproducible toxin assembly assay was reported recently (Tinker et al., 2003). This will aid in future study of MDT derivatives and other compounds aimed at inhibiting the assembly of CT and LT holotoxins.

Components of the type TI secretion system

After assembly of the holotoxin in the periplasm, the toxin is translocated in the folded state across the outer membrane by the Eps export apparatus of proteins in V. cholerae (Sandkvist et al, 2000; Sandkvist, 2001) and the recently discovered, closely related Gsp assembly of proteins in ETEC (Tauschek et al., 2002). About 14 proteins form a multimillion-Dalton secretion machinery spanning the inner and outer membrane. We have recently succeeded in solving the structure of the cytoplasmic secretion ATPase EpsE from V. cholerae (Robien et al., 2003). The EpsE structure in the presence and absence of AMPPNP appeared to be essentially identical. EpsE consists of two domains (the N and C domains) (Fig. 2) with relatively little mutual contact. The latter domain comprises three subdomains: the ATP-binding domain, a metal-binding domain and a C- terminal extension. The function of the metal-binding domain, in which four cysteines surround a bound metal ion, is still to be determined. In crystalline form, EpsE molecules are arranged into an extended helical structure, which has been reported in a significant number of related proteins, yet many of these proteins also form hexameric assemblies with cyclic point group symmetry (see (Robien et al., 2003) for references). A hexameric model of EpsE could be constructed in which the metal-binding domains of the six subunits face the cytoplasm and the six much more conserved N-terminal domains are close to the cytoplasmic surface of the inner membrane. Since the first 90 residues of the EpsE were deleted for structure determination, yet they are required for interactions with other components of the Eps apparatus (Sandkvist et al., 2000), much remains to be done to fully understand the functioning of EpsE within the context of this sophisticated secretion machinery. This is of course only the beginning of understanding how the entire export machinery recognizes CT and LT in the periplasm and secretes these proteins through a pore across the outer membrane.

Fig. 2. Structure of V. cholerae EpsE, a cytoplasmic ATPase involved in secretion of CT.

The B-subunit: receptor binding and structure-based inhibitor design

Structural studies of the receptor-binding sites on the B pentamer of CT and LT have so far received the most attention, mainly due to extensive efforts focused on the structure-based design of a large number of receptor-binding antagonists. Because endocytosis of CT and LT is preceded by receptor binding of ganglioside GM1 on the host cell surface, antagonists of the receptor-binding process will be good leads for the development of prophylactics against toxin-caused diarrhea. Shortly after the elucidation of structures of the LT AB^sub 5^ holotoxin alone and in complex with the disaccharide lactose (Sixma et al., 1991a, 1992), we solved the structure of CT B pentamer in complex with the oligosaccharide head group of its native receptor GM1 (GM1-OS) at very high resolution (Merritt et al., 1994b). These structures revealed that galactose, the terminal sugar of GM1-OS, binds very specifically at a buried pocket of the receptor-binding site. However, the rest of the binding site is very shallow and lacks well- defined hydrophobic pockets that can be exploited using traditional structure-based drug design approaches to arrive at potent inhibitors.

Hence, we set out to explore the use of galactose as an "anchor" for inhibitor design by making chemical modifications and extensions from a galactose molecule. Initial studies were carried out by screening commercially available galactose derivatives. This led to the discovery of m-nitrophenyl-α-D-galaclopyranoside (MNPG) as better inhibitor with 100-fold improvement of affinity over the starting galactose (Minke et al., 1999). MNPG therefore served as a better "anchor" for further inhibitor design (Pickens et al., 2002). So far, more than 100 galactose or MNPG derivatives have been synthesized in our laboratories, and low micromolar inhibitors for CT and LT receptor binding have been obtained. A large number of these compounds have been co-crystallized with the CT or LT B pentamer (for a partial list, see Fig. 3). In addition, molecular dynamics simulations can reproduce the crystallographically observed ability of molecules like MNPG to displace canonical water molecules, and the inability of closely related compounds such as MCPG (w-carboxy-phenyl-α-D-galactopyranoside) to do the same; further analysis of these properties by free energy perturbation calculations points to delicate balances in partial desolvations to be a crucial factor (Tan and Verlinde, unpublished results). These crystallographic and simulation studies will form the basis of future structure-based design of potent receptor-binding antagonists.

Fig. 3. Chemical structures of a partial list of inhibitors co- crystallized with toxin B pentamers.

The most dramatic improvement in receptor-binding antagonist design was achieved with a multivalent approach, taking advantage of the symmetrical arrangement of five identical binding sites on the B pentamer of the toxins. Our basic concept to achieve significant gains in affinity of ligands towards the B pentamer has been a modular approach of "Finger-Linker-Core", where a pentavalent "Core" is extended by five long (and, so far, very flexible) "Linkers" with each Linker ending in a "Finger" that occupies the GM1-OS-binding site. This concept allows for independent variation and optimization of the three modules. Such an approach has been proven to be spectacularly successful and, in addition, has led to new fundamental insights into factors affecting the affinity of precisely defined multivalent compounds for multivalent targets. Affinity gains, by five and six orders of magnitude, for large pentavalent and decavalent inhibitors of CT and LT B pentamers have been realized using our modular approach (Fan et al., 2000; Merritt et al., 2002; Zhang et al., 2002). Two of these mega-ligands, with molecular weights of 6.4 and 10.6kDa, have been observed in crystal structures in complex with toxin B pentamers (Merritt et al., 2002; Zhang et al., 2002). We demonstrated that the effective length of the "linkers" is a critical factor in governing the affinity of multivalent ligands for the B pentamer (Fan et al., 2000), and also that improving the affinity for individual fingers leads to a concomitant improvement of pentavalent ligands (Merritt et al., 2002). Our crystallographic studies showed a part of the linker in a pentavalent ligand-B pentamer complex remained close to the protein surface (Merritt et al., 2002) and that the decavalent inhibitor resulted in a "sandwich arrangement" of two B pentamers facing each other bridged by the ligand (Fig. 4) (Zhang et al., 2002). Insights gained from these structures are the foundations for new strategies of multivalent ligand design; for example, incorporation of functional groups in the linker to enhance favorable interactions to the surface of CT or LT B pentamers.

Fig. 4. A decavalent ligand bound to toxin B pentamer, forming a discrete sandwich structure.

Conclusions

The structural biology and structure-based drug design studies outlined above have provided detailed insight into many steps of the mode of action of these sophisticated heterohexameric toxins, and laid the foundation for the design and development of specific antagonists of the modes of action of the toxins. Yet, only a small part of the complex life cycle of CT and LT has been unraveled so far. In particular, the assembly of the toxins in the periplasm, the secrets of the type II secretion system, and the precise mode of interaction of the toxins with a plethora of proteins of the human host while traveling through the target cells (Kahn et al., 2002; Lencer and Tsai, 2003) remain a tremendous challenge for further understanding of this pair of toxins while providing new opportunities for intervention strategies.

Acknowledgements

We thank the National Institutes of Health for financial support of our work (AI44954 to E. Fan; GM54618 to C. Verlinde; and AI34501 to W. Hol). C. O'Neal is a Howard Hughes Medical Institute predoctoral fellow. We are grateful to current and past members of our laboratories who have made contribution to the work described here, particularly in structural biology: Titia Sixma, Focco van den Akker, Ingeborg Feil, Bianca Nerenberg, Misol Ahn, Jan Abendroth, Steve Sarfaty, Stewart Turley, and Francis Athappilly; in inhibitor design: Wendy Sanderson; and in chemical synthesis: Zhen Hou, Feng Hong, Ajit Ghosh, Zhenfa Zhang and Jiyun Liu. We also like to thank our numerous collaborators including: Bernard Witholt, R\ino Rappuoli, Joseph Martial, Randall Holmes, Michael Jobling, Timothy Hirst, Maria Sandkvist and Michael Bagdasarian.

References

Fan, E., Zhang, Z., Minke, W.E., Hou, Z., Verlinde, C.L.M.J., Hol, W.G.J., 2000. High affinity pentavalent ligands of Escherichia coli heat-labile enterotoxin by modular structure-based design. J. Am. Chem. Soc. 122, 2663-2664.

Feil, I.K., Reddy, R., de Haan, L., Merritt, E.A., van den Akker, F., Storm, D.R., Hol, W.G., 1996. Protein engineering studies of A- chain loop 47-56 of Escherichia coli heat-labile enterotoxin point to a prominent role of this loop for cytotoxicity. MoI. Microbiol. 20, 823-832.

Hardy, S.J., Holmgren, J., Johansson, S., Sanchez, J., Hirst, T.R., 1988. Coordinated assembly of multisubunit proteins: oligomerization of bacteria] enterotoxins in vivo and in vitro. Proc. Natl. Acad. Sci. USA 85, 7109-7113.

Hovey, B.T., Verlinde, C.L., Merritt, E.A., Hol, W.G., 1999. Structure-based discovery of a pore-binding ligand: towards assembly inhibitors for cholera and related AB^sub 5^ toxins. J. Mol. Biol. 285, 1169-1178.

Kahn, R.A., Fu, H., Roy, C.R., 2002. Cellular hijacking: a common strategy for microbial infection. Trends Biochem. Sci. 27, 308-314.

Lencer, W.I., Tsai, B., 2003. The intracellular voyage of cholera toxin: going retro. Trends Biochem. Sci. 28, 639-645.

Lencer, W.I., Constable, C., Moc, S., Jobling, M.G., Webb, H.M., Ruston, S., Madara, J.E., Hirst, T.R., Holmes, R.K., 1995. Targeting of cholera toxin and Escherichia coli heat labile toxin in polarized epithelia: role of COOH-terminal KDEE. J. Cell Biol. 131, 951-962.

Mekalanos, J.J., Collier, R.J., Romig, W.R., 1979. Enzymic activity of cholera toxin. II. Relationships to proteolytic processing, disulfide bond reduction, and subunit composition. J. Biol. Chem. 254, 5855-5861.

Merritt, E.A., Hol, W.G.J., 1995. AB^sub 5^ toxins. Curr. Opin. Struct. Biol. 5, 165-171.

Merritt, E.A., Pronk, S.E., Sixma, T.K., Kalk, K.H., van Zanten, B.A., Hol, W.G., 1994a. Structure of partially-activated E. colt heat-labile enterotoxin (LT) at 2.6 [Angstrom] resolution. FEBS Lett. 337, 88-92.

Merritt, E.A., Sarfaty, S., van den Akker, F., L'Hoir, C., Martial, J.A., Hol, W.G., 1994b. Crystal structure of cholera toxin B-pentamer bound to receptor GM1 pentasaccharide. Protein Sci. 3, 166-175.

Merritt, E.A., Zhang, Z., Pickens, J.C., Ahn, M., Hol, W.G., Fan, E., 2002. Characterization and crystal structure of a high-affinity pentavalent receptor-binding inhibitor for cholera toxin and E. coli heat-labile enterotoxin. J. Am. Chem. Soc. 124, 8818-8824.

Minke, W.E., Roach, C., Hol, W.G., Verlinde, C.L., 1999. Structure-based exploration of the ganglioside GM1 binding sites of Escherichia coli heat-labile enterotoxin and cholera toxin for the discovery of receptor antagonists. Biochemistry 38, 5684-5692.

O'Neal, C.J., Amaya, E.I., Jobling, M.G., Holmes, R.K., Hol, W.G., 2004. Crystal structures of an intrinsically active cholera toxin mutant yield insight into the toxin activation mechanism. Biochemistry 43, 3772-3782.

Pickens, J.C., Merritt, E.A., Ahn, M., Verlinde, C.L., Hol. W.G., Fan, E., 2002. Anchor-based design of improved receptor-binding antagonists of cholera toxin and Escherichia coli heat-labile enterotoxin that display multiple binding modes. Chem. Biol. 9, 215- 224.

Robien, M.A., Krumm, B.E., Sandkvist, M., Hol, W.G., 2003. Crystal structure of the extracellular protein secretion NTPase EpsE of Vibrio choleras. J. Mol. Biol. 333, 657-674.

Sandkvist, M., 2001. Type II secretion and pathogenesis. Infect. Immun. 69, 3523-3535.

Sandkvist, M., Bagdasarian, M., Howard, S.P., 2000. Characterization of the multimeric Eps complex required for cholera toxin secretion. Int. J. Med. Microbiol. 290, 345-350.

Sixma, T.K., Pronk, S.E., Kalk, K.H., Wartna, E.S., van Zanten, B.A., Witholt, B., Hol, W.G., 1991a. Crystal structure of a cholera toxin-related heat-labile enterotoxin from E. coli. Nature 351, 371- 377.

Sixma, T.K., Pronk, S.E., Terwisscha van Schellinga, A.C., Aguirre, A., Kalk, K.H., Vriend, G., Hol, W.G.J., 1991b. Native non- isomorphism in the structure determination of heat labile enterotoxin (LT) from E. coll. In: Wolf, W., Evans, P.R., Leslie, A.G.W. (Eds.), Isomorphous Replacement and Anomalous Scattering. SERC Daresbury Lab., Warrington, UK, pp. 133 -140.

Sixma, T.K., Pronk, S.E., Kalk, K.H., van Zanten, B.A., Berghuis, A.M., Hol, W.G., 1992. Lactose binding to heat-labile enterotoxin revealed by X-ray crystallography. Nature 355, 561-564.

Sixma, T.K., Kalk, K.H., van Zanten, B.A., Dauter, Z., Kingma, J., Witholt, B., Hol, W.G., 1993. Refined structure of Escherichia coli heat-labile enterotoxin, a close relative of cholera toxin. J. Mol. Biol. 230, 890-913.

Spangler, B.D., 1992. Structure and function of cholera toxin and the related Escherichia coli heat-labile enterotoxin. Microbiol. Rev. 56, 622-647.

Tauschek, M., Gorrell, R.J., Strugnell, R.A., Robins-Browne, R.M., 2002. Identification of a protein secretory pathway for the secretion of heat-labile enterotoxin by an enterotoxigenic strain of Escherichia coli. Proc. Natl. Acad. Sci. USA 99, 7066-7071.

Tinker, J.K., Erbe, J.L., Hol, W.G., Holmes, R.K., 2003. Cholera holotoxin assembly requires a Hydrophobie domain at the A-B5 interface: mutational analysis and development of an in vitro assembly system. Infect. Immun. 71, 4093-4101.

van den Akker, F., Merritt, E.A., Pizza, M., Domenighini, M., Rappuoli, R., Hol, W.G., 1995. The Arg7Lys mutant of heat-labile enterotoxin exhibits great flexibility of active site loop 47-56 of the A subunit. Biochemistry 34, 10996-11004.

van den Akker, F., Sarfaty, S., Twiddy, E.M., Connell, T.D., Holmes. R.K., Hol, W.G., 1996. Crystal structure of a new heat- labile enterotoxin, LT-IIb. Structure 4, 665-678.

WHO, 2002. World Health Organization, [http:// www.who.int/emc/ diseases/cholera/wer7731.pdf].

Widdus, R., 2001. Public-private partnerships for health: their main targets, their diversity, and their future directions. Bull. World Health Org. 79, 713-720.

Zhang, Z., Merritt, E.A., Ahn, M., Roach, C., Hou, Z., Verlinde, C.L., Hol, W.G., Fan, E., 2002. Solution and crystallographic studies of branched multivalent ligands that inhibit the receptor- binding process of cholera toxin. J. Am. Chem. Soc. 124, 12991- 12998.

Erkang Fan(a), Claire J. O'Neal(b), Daniel D. Mitchell(a), Mark A. Robien(a), Zhongsheng Zhang(a), Jason C. Pickens(b), Xiao-Jian Tan(a), Konstantin Korotkov(a), Claudia Roach(a,c), Brian Krumm(a,c), Christophe L.M.J. Verlinde(a), Ethan A. Merritt(a), Wim G.J. Hol(a,c),*

a Department of Biochemistry, Biomolecular Structure Center, University of Washington, Box 357742, Seattle WA 98195, USA

b Department of Chemistry, University of Washington, Seattle, WA 98195, USA

c Howard Hughes Medical Institute, University of Washington, Seattle, WA 98195, USA

* Corresponding author. Department of Biochemistry, Biomolecular Structure Center, University of Washington, Box 357742, Seattle WA 98195, USA. Tel.: +206-685-7044; fax: +206-685-7002.

E-mail address: wghol@u.washington.edu (W.G.J. Hol).

Copyright Urban & Fischer Verlag Oct 2004

Source: International Journal of Medical Microbiology

More News in this Category