A Novel Intein-Like Autoproteolytic Mechanism in Autotransporter Proteins

https://doi.org/10.1016/j.jmb.2010.06.068Get rights and content

Abstract

Many virulence factors secreted by pathogenic Gram-negative bacteria are found to be members of the autotransporter protein family. These proteins share a common mechanism by which they exit the periplasm, involving the formation of a 12-stranded β-barrel domain in the outer membrane. The role of this barrel in the secretion of the N-terminal passenger domain is controversial, and no model currently explains satisfactorily the entire body of experimental data. After secretion, some autotransporter barrels autoproteolytically cleave away the passenger, and one crystal structure is known for a barrel of this type in the postcleavage state. Hbp is an autotransporter of the self-cleaving type, which cuts the polypeptide between two absolutely conserved asparagine residues buried within the barrel lumen. Mutation of the first asparagine residue to isosteric aspartic acid prevents proteolysis. Here we present the crystal structure of a truncated Hbp mutant carrying the C-terminal residues of the passenger domain attached to the barrel. This model mimics the state of the protein immediately prior to separation of the passenger and barrel domains, and shows the role of residues in the so-called “linker” between the passenger and β domains. This high-resolution membrane protein crystal structure also reveals the sites of many water molecules within the barrel. The cleavage mechanism shows similarities to those of inteins and some viral proteins, but with a novel means of promoting nucleophilic attack.

Introduction

Autotransporters are virulence factors secreted by pathogenic Gram-negative bacteria.1, 2, 3, 4, 5 They carry a cleavable N-terminal signal sequence that directs them to the periplasm by the Sec pathway,2 but possess a unique mechanism of crossing the outer membrane into the medium. For many years, autotransporter proteins represented apparently the most elegant and straightforward solution to the problem of moving a protein across the outer membrane, since it was believed that they encoded within one polypeptide both the protein to be secreted (a globular N-terminal region dubbed the “passenger”) and the channel through which it passes from the periplasm (a C-terminal β-barrel domain)6 (Fig. 1). In this so-called “hairpin model”, the passenger domain emerges C-terminal end first, and the translocator must carry two strands of polypeptide chain passing each other.6 The simplicity proved more apparent than real, however, since it has now been shown that the essential outer membrane protein Omp85 is required for autotransporter secretion in Neisseria.4, 7 A homologue, recently renamed BamA, plays a required role in outer membrane protein folding in Escherichia coli.8 Since autotransporter secretion is not powered by ATP or any electrochemical gradient, folding at the cell surface was believed to pull the passenger through the pore, but various groups have shown that improperly folded mutant autotransporters may be presented at the cell surface.9, 10, 11 The autotransporter secretion mechanism has been of great commercial interest, leading to different proteins being attached to translocators in attempts to express desirable proteins; however, this work has been hampered by poor understanding of the mechanism.12 Several structures of autotransporter domains, including one of the related trimeric autotransporters, have been published recently13 (Table 1). The first crystal structure of a barrel domain, NalP from Neisseria, shows a simple 12-stranded barrel with short loops between strands and an N-terminal α-helix that passes through the barrel to present the N terminus at the cell surface.14 The pore size of the NalP barrel (roughly 10 Å across) appears quite incompatible with any folded structure passing through. Hbp is a member of a sub-group of autotransporters called SPATEs (Serine Protease Autotransporters of Enterobacteriaceae), which autoproteolytically cut the passenger from the barrel, but not using their trypsin-like domain at the N terminus of the passenger. The crystal structure of the passenger of Hbp has been solved,15 allowing the design of mutants carrying disulfide bonds that cross-link either residues close in sequence or residues on different domains within it. Cross-linking residues 110–348 blocked transport, whereas cross-linking residues 712–718 did not.16 This second result, together with the small pore of the β-barrel, makes the hairpin model highly unlikely, unless the barrel is in some expanded state. A second barrel structure has been solved from a SPATE called EspP.17 The barrel domain, purified from bacterial membranes, is the remnant missing the passenger protein. The EspP barrel shows a 12-stranded structure strongly reminiscent of the NalP model, but with only a truncated α-helix in the lumen, tilted so that its axis is roughly perpendicular to that of the pore (Fig. 2b). The crystal structure of a complete autotransporter called EstA, which has an unusually small passenger domain, has been described recently; the barrel resembles that of NalP.18

Absence of the passenger domain does not prevent correct insertion of the barrel domain into the outer membrane, and biochemical studies of EspP missing much of the passenger have helped reveal details of the cutting mechanism, suggesting it to be related to one previously characterized in viral proteins.19 In order to study the mechanism of Hbp secretion and autoproteolysis, we have constructed a noncleavable Hbp mutant, missing almost the entire passenger domain between the signal peptide and the cutting site. By replacing the first asparagine of the invariant asparagine pair straddling the cut site, proteolytic activity is prevented. We have isolated the protein from membranes and solved its crystal structure. The isosteric mutation from asparagine to aspartic acid gives a picture of the protein immediately before cleavage and a view of interactions made by the C terminus of the passenger domain with the barrel at this frozen moment in time. Our results complement and extend earlier crystallographic data from other proteins and show that the cleavable autotransporters do not form a single helix extending through the barrel prior to cleavage, as inferred from the NalP and EspP models. Comparison of the EspP barrel and the Hbp model shows changes in the hydrogen-bonding pattern that are consistent with the known stabilization of the barrel domain upon cleavage. The high resolution of the X-ray data allows us to observe many water molecules within the barrel and also shows how the acyl intermediate is stabilized in a completely different manner from those described in other families of proteases.

Section snippets

Overall structure

Wild-type Hbp is expressed as a 1377-residue preprotein with a signal sequence extending from residues 1 to 52. The crystal structure of the passenger (from residues 53 to 1100) has been solved previously. That model [Protein Data Bank (PDB) entry 1wxr] is numbered from 1, so that the residue numbering is out of step by 52 with the preprotein. Here we compare the model of the barrel with mutagenesis experiments carried out by Kostakioti and Stathopoulos, who numbered the preprotein from 1.20 To

Conclusion

The mutations of Arg1105 in Hbp and Asp1120 in EspP demonstrate the sensitivity of autoproteolysis to electrostatic changes in the vicinity of the cleavage site.17, 19 This obliges us to interpret our structure with caution, since an extra negative charge has been introduced; however, given that the model provides immediate explanations for the published phenotypes of many Hbp and EspP mutants it has substantial biochemical support. The Asp 1100 side chain makes two hydrogen bonds, both to

Plasmid construction

All constructs were made by PCR using the plasmid pACYC184-Hbp (carrying the full-length Hbp gene) as template DNA. The signal peptide of Hbp was cloned into pET22b using NdeI and NcoI to create pET22-NT. Site-directed mutagenesis at the cutting site was performed. The regions encoding residues 1075–1100 and 1100–1377 were amplified using the following pairs of primers:

  • 5′-GGGGCCATGGCACGTAACGACGGCCAG-3′ (NcoI site underlined)

  • 5′-GCGTTTGTTCAGGTTGTCAACTTCAGTGATGAAG-3′ (mutated base underlined)

Acknowledgements

We are grateful to Dr. Peter van Ulsen and Prof. Joen Luirink for helpful comments on the manuscript, and to Drs. Jose Caaveiro and Bernadette Byrne for helpful discussions on crystallization.

References (37)

  • DingY. et al.

    Crystal structure of a mini-intein reveals a conserved catalytic module involved in side chain cyclization of asparagine during protein splicing

    J. Biol. Chem.

    (2003)
  • VelardeJ.J. et al.

    Hydrophobic residues of the autotransporter EspP linker domain are important for outer membrane translocation of its passenger

    J. Biol. Chem.

    (2004)
  • HendersonI.R. et al.

    Type V protein secretion pathway: the autotransporter story

    Microbiol. Mol. Biol. Rev.

    (2004)
  • DesvauxM. et al.

    Type V protein secretion: simplicity gone awry?

    Curr. Issues Mol. Biol.

    (2004)
  • DautinN. et al.

    Protein secretion in Gram-negative bacteria via the autotransporter pathway

    Annu. Rev. Microbiol.

    (2007)
  • PohlnerJ. et al.

    Gene structure and extracellular secretion of Neisseria gonorrhoeae IgA protease

    Nature

    (1987)
  • VoulhouxR. et al.

    Role of a highly conserved bacterial protein in outer membrane protein assembly

    Science

    (2003)
  • OliverD.C. et al.

    A conserved region within the Bordetella pertussis autotransporter BrkA is necessary for folding of its passenger domain

    Mol. Microbiol.

    (2003)
  • Cited by (62)

    • BamA is required for autotransporter secretion

      2020, Biochimica et Biophysica Acta - General Subjects
      Citation Excerpt :

      Two sub-types, type Va and Vc, are composed of an N-terminal passenger domain (which usually contains the virulence factor [2]) and a C-terminal β-barrel translocon [3–8]. Some autotransporters contain an α-helical linker which connects the β-barrel to its passenger domain [3,5,6,9]. The name autotransporter originally comes from the understanding that they require no accessory factors nor external energy to secrete across the outer membrane [10,11], although more recently it has been recognized that insertion of the β-barrel translocon is mediated by the β-barrel assembly machinery (BAM) complex [12,13].

    View all citing articles on Scopus
    1

    Present address: F. Kawai, Department of Biology and Biochemistry, University of Houston, Houston, TX 77204, USA.

    View full text