Journal of Molecular Biology
Volume 284, Issue 2, 27 November 1998, Pages 449-461
Journal home page for Journal of Molecular Biology

Regular article
The molecular mechanism of pneumolysin, a virulence factor from Streptococcus pneumoniae1

https://doi.org/10.1006/jmbi.1998.2167Get rights and content

Abstract

Pneumolysin, a member of the thiol-activated cytolysin family of toxins, is a virulence factor from the Gram-positive bacterium Streptococcus pneumoniae. The toxin forms large oligomeric pores in cholesterol-containing membranes of eukaryotic cells. A plethora of biochemical and mutagenesis data have been published on pneumolysin, since its initial characterization in the 1930s. Here we present an homology model of the monomeric and oligomeric forms of pneumolysin based on the recently determined crystal structure of perfringolysin O and electron microscopy data. A feature of the model is a striking electronegative surface on parts of pneumolysin that may reflect its cytosolic location in the bacterial cell. The models provide a molecular basis for understanding the effects of published mutagenesis and biochemical modifications on the toxic activity of pneumolysin. In addition, spectroscopic data are presented that shed new light on pneumolysin activity and have guided us to hypothesise a detailed model of membrane insertion. These data show that the environment of some tryptophan residues changes on insertion and/or pore formation. In particular, spectroscopic analysis of a tryptophan mutant, W433F, suggests it is the residue mainly responsible for the observed effects. Furthermore, there is no change in the secondary structure content when the toxin inserts into membranes. Finally, the basis of the very low activity shown by a pneumolysin molecule from another strain of S. pneumoniae may be due to the movements of a key domain-domain interface. The molecular basis of pneumolysin-induced complement activation may be related to the structural similarity of one of the domains of pneumolysin to Fc, rather than the presumed homology of the toxin to C-reactive protein as previously suggested.

Introduction

The thiol-activated cytolysins (TACs) are a group of membrane-damaging toxins produced by a wide variety of Gram-positive bacteria (for a recent review, see Tweten, 1995). They are called thiol-activated as modification of a key cysteine residue causes toxin inactivation, although site-directed mutagenesis studies have shown this thiol group not to be essential to function (Saunders et al., 1989). More than 20 members have been characterized and, based on serological cross-reactivity and sequence comparisons, these toxins are believed to share a common mode of action, i.e. binding of the monomeric, water-soluble form of the toxin to the host cell membrane via the cholesterol receptor, followed by the generation of large oligomeric transmembrane pores larger than 150 Å in diameter causing cell lysis (Tweten, 1995). These toxins are highly specific for cholesterol, exhibiting a Kd of approximately 10−9M (Ohno-Iwashita et al., 1988). Eight TACs have been sequenced and they share sequence identities ranging from 40 to 83%, suggesting that they all share a common three-dimensional (3D) fold as well as a common mode of action. An 11 residue stretch of sequence, which we term the Trp-rich motif, represents the longest stretch of sequence identity within the TAC family. The motif contains the key cysteine residue and three tryptophan residues and plays a vital role in the mechanism of pore formation, although its precise role is unclear (Tweten, 1995).

We recently determined the 2.7 Å resolution crystal structure of perfringolysin O (PFO), the TAC from Clostridium perfringens(Rossjohn et al., 1997). The toxin is an unusually thin, rod-shaped molecule consisting predominantly of β-sheet structure, reminiscent of the structures of other membrane-damaging toxins such as aerolysin (Parker et al., 1994), Staphyloccus aureus α-haemolysin (Song et al., 1996) and anthrax protective antigen (Petosa et al., 1997). However, despite this gross similarity, these toxins are not structurally similar. A model of the oligomer was proposed based on the structure, electron microscopy images and biochemical data which led to suggestions of how the toxin might form pores in membranes (Rossjohn et al., 1997).

The bacterium Streptococcus pneumoniae is a major human pathogen causing diseases such as pneumonia, meningitis, bacteraemia and otitis media Paton et al 1991, Paton 1996. The emergence of drug-resistant pneumococci and the poor efficacy of the pneumococcal polysaccharide vaccine has prompted the search for alternative chemotherapeutic and vaccine targets (Paton, 1996). S. pneumoniae produces a number of protein virulence factors of which pneumolysin (PLY) is one. PLY is a 53 kDa protein consisting of 471 amino acid residues (Walker et al., 1987). Pneumolysin was first characterised circa the 1930s. Since that date, a wealth of biochemical and mutagenesis studies have been undertaken to elucidate the mechanism of this toxin. It is unique amongst the TACs in that it is cytosolic and only released upon autolysis of the bacterium. Studies have indicated that PLY plays a direct role in the virulence of the pneumococcal bacterium. Immunization of mice with a defective version of the toxin conferred protection against challenge by virulent pneumococci (Paton et al., 1991) and PLY-negative pneumococcal strains display a reduced virulence in animal models (Berry et al., 1989). Pure, sub-lytic concentrations of PLY have been shown to significantly inhibit polymorphonuclear leucocyte respiratory burst, chemotaxis and bactericidal activity (Paton & Ferrante, 1983). In addition, PLY has been shown to activate the classical complement pathway, in the absence of PLY-directed antibodies Paton et al 1984, Mitchell et al 1991. In vivo, this would stimulate an inflammatory response, a characteristic and key component of a pneumococcal infection. This complement activity was reported to arise from a region of PLY sharing sequence identity to human C-reactive protein (Shrive et al., 1996), a plasma protein involved in the acute phase response to infection or injury. There is in vivo evidence that both functions have a role in pneumococcal disease (Rubins et al., 1996).

Here we present homology models of PLY in its monomeric and oligomeric states, based on the crystal structure of PFO (Rossjohn et al., 1997) and other data. The PLY structure reveals a number of novel features in comparison to that of PFO, including a marked electronegative patch that encompasses the top of domain 1. The pneumolysin homology model provides a molecular basis for understanding the effects of the numerous site-directed mutants and biochemical modifications on the toxic activity of PLY. In addition, spectroscopic data are presented that corroborate the proposed molecular mechanism. These data show that the environment of some tryptophan residues changes on insertion and/or pore formation. The structure of PLY reveals that it is not structurally related to the C-reactive protein. Instead, it is proposed that domain 4 of PLY, by virtue of its structural homology to Fc (in particular its aggregated state), may provide an explanation as to how PLY can activate the classical complement pathway.

Section snippets

The PLY monomer

An homology model of PLY has previously been constructed (Sowdhamini et al., 1997) based on the crystal structure of another pore-forming toxin, aerolysin (Parker et al., 1994). The sequence identity between aerolysin and PLY is only 16.6%, but modelling was undertaken because electron microscopy data suggested that PLY was an elongated molecule containing four domains like aerolysin (Morgan et al., 1994), and the homology score for the sequence of PLY compared to proteins of known structure

Mechanism of membrane insertion

The PLY oligomer generates functional pores. However, there are no hydrophobic patches on either the surface of the molecule or at the domain interfaces of the PLY model. At first sight this suggests that PLY could not partition into the membrane. A similar dilemma was found with the PFO structure. A plausible mechanism for membrane insertion has been proposed for PFO (Rossjohn et al., 1997). Briefly, cholesterol is proposed to bind close to the Trp-rich motif, causing the motif to be

Construction of the homology model

The homology model of PLY was based on the 2.7 Å resolution crystal structure of PFO (Rossjohn et al., 1997). The sequences of PLY and the TACs were aligned using AMPS and displayed using ALSCRIPT (Barton, 1993). The PLY model was constructed using the HOMOLOGY module of Insight II (Molecular Simulations Inc., San Diego, U.S.A.) on a Silicon Graphics Indigo 2 Maximum Impact Workstation. The initial model was built in two stages: (1) identification of significant regions of sequence identity

Acknowledgements

We thank Dr Barbara Bolgiano for supplying the CONTIN fits and Dr Bill McKinstry and Ms Susanne Feil for their contributions towards the structure determination of PFO. J.R. is an Australian Research Council Postdoctoral Fellow. M.W.P. is an Australian Research Council Senior Research Fellow and acknowledges support from the Australian National Health and Medical Research Council. Work in Leicester was supported by the United Kingdom Medical Research Council, the Royal Society, and the

References (46)

  • A.T Brünger

    XPLOR: A System for X-ray Crystallography and NMR

    (1993)
  • The CCP4 suiteprograms for protein crystallography

    Acta Crystallog. sect. D

    (1994)
  • C Chothia et al.

    The relation between the divergence of sequence and structure in proteins

    EMBO J.

    (1986)
  • J.R de los Toyos et al.

    Functional analysis of pneumolysin by use of monoclonal antibodies

    Infect. Immun.

    (1996)
  • J.P Derrick et al.

    Crystal structure of a streptococcal protein G domain bound to an Fab fragment

    Nature

    (1992)
  • T.E Ferrin et al.

    The MIDAS display system

    J. Mol. Graph.

    (1988)
  • J.P Hennessey et al.

    Information content in the circular-dichroism of proteins

    Biochemistry

    (1981)
  • J Hill et al.

    Amino acids in pneumolysin important for hemolytic activity identified by random mutagenesis

    Infect. Immun.

    (1994)
  • E.G Hutchinson et al.

    PROMOTIF-A program to identify and analyze structural motifs in proteins

    Protein Sci.

    (1996)
  • T.A Jones et al.

    Improved methods for building models in electron density maps and the location of errors in these models

    Acta Crystallog. sect. A

    (1991)
  • Y.E Korchev et al.

    A conserved tryptophan in pneumolysin is a determinant of the characteristics of channels formed by pneumolysin in cells and planar lipid bilayers

    Biochem. J.

    (1997)
  • R.A Laskowski et al.

    PROCHECKa program to check the stereochemical quality of protein structures

    J. Appl. Crystallog.

    (1993)
  • R Lüthy et al.

    Assessment of protein models with three-dimensional profiles

    Nature

    (1992)
  • Cited by (0)

    1

    Edited by J. Thornton

    2

    Present address: T. J. Mitchell, Division of Infection and Immunity, Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow, G12 8QQ, U.K.

    View full text