Article

• The EMBO Journal (2006) 25, 5241 - 5249
• doi:10.1038/sj.emboj.7601402

Published online: 2 November 2006

• Subject Categories:

  • Membranes and Transport | Microbiology and Pathogens

Bacterial outer membrane secretin PulD assembles and inserts into the inner membrane in the absence of its pilotin

Ingrid Guilvout¹, Mohamed Chami², Andreas Engel², Anthony P Pugsley¹ and Nicolas Bayan^1,3

1. Molecular Genetics Unit and CNRS URA2172, Institut Pasteur, Paris, France
2. ME Müller Institute, Biozentrum, University of Basel, Basel, Switzerland
3. CNRS UMR8619, Université de Paris Sud, Orsay, France

Correspondence to:

Anthony P Pugsley, Molecular Genetics Unit, Institut Pasteur, 25, rue du Dr Roux, 75724 Paris Cedex 15, France. Tel.: +33 1 4568 8494; Fax: +33 1 4568 8960; E-mail: max@pasteur.fr

Received 20 July 2006; Accepted 29 September 2006

• Keywords:

  • chaperone,
  • outer membrane,
  • pilotin,
  • secretin,
  • type II secretion

Introduction

The secretin PulD is the only integral outer membrane component of the type II secretion system (T2SS) or secreton by which the Gram-negative bacterium Klebsiella oxytoca secretes the amylolytic enzyme pullulanase (d'Enfert et al, 1989). Thus, PulD probably forms the conduit through which pullulanase crosses the outer membrane. Like other secretins involved in type II (Bitter et al, 1998; Brok et al, 1999) and type III (Crago and Koronakis, 1998; Burghout et al, 2004) secretion, type IV pilus biogenesis (Bitter et al, 1998; Collins et al, 2004) and filamentous bacteriophage secretion (Opalka et al, 2003), PulD forms multimers, in this case, a dodecamer (Nouwen et al, 1999; Chami et al, 2005). When examined by electron microscopy, PulD dodecamers resemble a ca 15 nm deep cylinder with a central plug (Chami et al, 2005). Secretin multimers do not dissociate in SDS, and some, including PulD, do not even dissociate at 100°C. Unlike classical outer membrane proteins that are extracted by at least some nonionic detergents, only the ionic, denaturing detergent SDS (Hardie et al, 1996a; Bitter et al, 1998) and zwitterionic detergents such as ZW3-14 at high salt concentrations solubilize PulD (Nouwen et al, 1999).

A dodecameric core domain of PulD, PulD-C, obtained by limited proteolysis with trypsin (Nouwen et al, 2000; Chami et al, 2005) and comprising amino acids 262–616 of the 660 amino-acid polypeptide, with a single nick after amino acid 297, includes the integral outer membrane- and plug-forming regions (Chami et al, 2005). The region N-terminal to this domain, the N domain, constitutes part of the outer membrane-distal wall of the cylinder that plunges deep into the periplasm (Nouwen et al, 1999, 2000; Chami et al, 2005). These features are consistent with a model in which secretins form a closed outer membrane channel that opens to allow the selective release of specific exoproteins such as pullulanase from the periplasmic compartment. The driving force for secretion could be the piston-like motion of an elongating and retracting pilus-like structure, the pseudopilus, of which PulG is the major component in K. oxytoca (Köhler et al, 2004; Chami et al, 2005).

The region C-terminal to the C domain of PulD, the S domain, is the binding site for a lipoprotein chaperone, the pilotin PulS (Daefler et al, 1997), which is specifically required to protect PulD from proteolysis and to direct it to the outer membrane (Hardie et al, 1996a, 1996b). Twelve copies of PulS copurify with the PulD dodecamer during purification (Nouwen et al, 1999, 2000). In the absence of PulS, PulD associates with the inner membrane and induces the phage shock response, resulting in the appearance of large amounts of PspA protein (Hardie et al, 1996a, 1996b). Other Klebsiella-derived proteins are not required for PulD multimerization and outer membrane localization in Escherichia coli, but secretin assembly in the T2SSs of some other bacteria requires ancillary factors, notably the ExeA and ExeB proteins in Aeromonas hydrophila (Ast et al, 2002) and the ExeB-like protein OutB in Erwinia chrysanthemi (Condemine and Shevchik, 2000), the lipoproteins Tgl and PilW in Myxococcus xanthus and Neisseria meningitidis, respectively (Carbonnelle et al, 2005; Nudleman et al, 2006), and the general outer membrane insertion factor Omp85 in N. meningitidis (Voulhoux et al, 2003; reviewed in Bayan et al, 2006).

Peptide insertions in the C domain of PulD disrupt its ability to form multimers (Guilvout et al, 1999). Knowledge of the precise junction between the N and C domains (Chami et al, 2005) allowed us to amplify, clone and express DNA encoding only the C and S domains fused to the PulD signal peptide and the first 14 amino acids of PulD that are required for its export. We report studies of the multimerization, targeting and membrane insertion of this PulD fragment (hereafter referred to as PulD-CS) and of the complete PulD multimer.

PulD-CS is targeted to the outer membrane together with PulS

Two forms of PulD-CS were generated in this study, one without additional amino acids and one with a His tag inserted in the S-domain (PulD-CShis), as in the previously-characterized full-length PulDhis (Chami et al, 2005) (see Materials and methods). PulD-CS and PulD-CShis exhibited identical properties and were used interchangeably. Total cell extracts of E. coli producing PulD-CS together with PulS were treated with phenol to dissociate any PulD-CS multimers (Hardie et al, 1996a) and examined by SDS–PAGE (on a 10% acrylamide gel) and immunoblotting. A single immunoreactive band that was absent from control cells without PulD-CS was detected at ca 46 kDa (not shown). Analysis of membrane fragments separated by floatation through sucrose gradients revealed PulD-CS predominantly in the dense, outer membrane fractions (Supplementary Figure S1A), as is full-length PulD (Hardie et al, 1996a; Supplementary Figure S1B). Thus, PulD-CS localizes to the outer membrane in cells producing PulShis. However, PulD-CS was unable to substitute for PulD to promote pullulanase secretion (not shown).

To determine whether the properties of PulD were affected by removal of the N domain, membranes containing PulD-CS were treated with a range of detergents under different conditions. Four conditions lead to quantitative extraction of dodecameric PulD from E. coli envelope preparations: the ionic detergent SDS, the zwitterionic detergent N-tetradecyl N,N-dimethyl-3-ammonio-1-propanesulfonate (ZW3-14) with 0.25 M NaCl at pH 7.5 or 8 or without salt at pH 10, and simultaneous action of the nonionic detergent octylpolyoxyethylene (octylPOE) and lysozyme (25 g/ml) at pH 10. Lysozyme cannot replace NaCl for extraction by ZW3-14 at pH 7.5 or 8, and high salt does not permit PulD extraction by octylPOE with lysozyme at pH 7.5–8 (Supplementary Figure S2 and data not shown).

PulD-CS in cell envelopes with PulS was extracted exclusively under conditions that also permitted PulD extraction; that is, ZW3-14 with NaCl at pH 7.5 and octylPOE with lysozyme at pH 10, but not ZW3-14 without NaCl at pH 7.5, or octylPOE without lysozyme at pH 10 or at pH 7.5 (Table I). Thus, the properties of PulD that contribute to its unusual requirements for extraction from the outer membrane are entirely determined by the C (and S) domains.

Table 1: Extraction of PulD and PulD-CS by different detergents from strains with and without PulS

View full table

PulD-CS forms PulD-like multimers in the outer membrane

The amount of PulD-CS monomer detected by immunoblotting was considerably lower when the cell extracts were not treated with phenol, suggesting the presence of PulD-CS multimers with properties similar to those of PulD (Hardie et al, 1996a). PulD-CS multimers were not detected by immunoblotting, however (data not shown). Therefore, PulD-CShis was produced in E. coli together with PulS, solubilized in ZW3-14 with NaCl, purified by cobalt affinity chromatography and examined by SDS–PAGE and Coomassie blue staining. Substantial amounts of a slow-migrating protein were detected (indicated by an arrow in Figure 1, lane 3). This band migrated faster than PulDhis (compare lanes 1 and 3 in Figure 1) but did not react with anti-PulD (not shown).

Figure 1.

SDS–PAGE analysis of PulDhis and PulD-CShis purified by cobalt affinity chromatography from envelopes of E. coli cells with and without PulS or PulShis. The strains used were PAP105(pCHAP585) (with PulS) or PAP5198 (without PulS). PulDhis and PulD-CShis multimers ((PulDhis)12 and (PulD-CShis)12, respectively) were dissociated into monomers (PulDhis and PulD-CShis) with phenol. Proteins were separated on a 4–15% acrylamide gradient gel and stained with Coomassie blue. The positions of molecular size markers (kDa) are shown on the left. Arrows indicate the multimeric forms of PulDhis and PulD-CShis. Note that PulDhis and PulD-CShis migrate more slowly than expected from their mass (68.7 and 45.6 kDa) in this electrophoresis system.

View full figure (45 KB)

Phenol treatment of the purified protein prior to SDS–PAGE caused these multimers to disappear and generated a band (Figure 1, lane 7) that migrated faster than monomeric PulDhis (compare lanes 5 and 7 in Figure 1) and reacted with PulD antibodies (not shown). We conclude that PulD-CShis forms large, defined multimers and that PulD antibodies (Hardie et al, 1996a) react with PulD-CShis only when it is dissociated.

To determine whether the PulD-CShis multimers are structurally similar to C domain dodecamers obtained by proteolysis, negatively stained samples of purified PulD-CShis complexes were examined by transmission electron microscopy. Abundant ring-like structures almost indistinguishable from those of trypsin-resistant PulD-C dodecamers (Nouwen et al, 1999; Chami et al, 2005) were observed (Figure 2A). These particles were 14 nm in diameter and had a stain-filled 7 nm diameter cavity, indicating that, like PulD (Chami et al, 2005), they are dodecameric. Many cup-and-saucer-like side-views of individual PulD-CS particles were observed (Figure 2A), in contrast to the low amounts of almost exclusively 'cup-to-cup' double complexes previously found in PulD-C (Chami et al, 2005). This change in behaviour suggests that amino acids present in PulD-CS but absent from PulD-C obtained by proteolysis (i.e., the S domain and additional amino acids at the N-terminus) prevent double dodecamer formation. These data indicate that the N domain influences neither the outer membrane targeting nor the multimerization of PulD.

Figure 2.

Electron microscopy of purified PulD-CShis and PulDhis particles negatively stained with uranyl acetate. (A) PulD-CShis purified from membranes of E. coli PAP105 with PulS; inset, averaged images of two major classes: top views (n=171) and side views (n=233). (B) PulD-CShis purified from membranes of E. coli PAP5198 without PulS; inset, averaged images of two major classes: top views (n=177) and side views (n=300). (C) PulDhis purified from E. coli PAP5198 without PulS; inset, averaged images of two major classes: top views (n=272) and side views (n=253). The scale bar corresponds to 50 nm and the inset baseline corresponds to 25 nm.

View full figure (171 KB)

PulS is not required to form PulD or PulD-CS multimers

Previous studies indicated the formation of a PulD complex in the absence of PulS (Hardie et al, 1996b), but it was not characterized and could correspond to heterogeneous aggregates. PulD-CShis was partially degraded in the absence of PulS (not shown), as reported previously for PulD (Hardie et al, 1996b). A protease-deficient strain (PAP5198) was used to reduce this problem. PulD-CShis produced by this strain without PulS was extracted from the cell envelope with ZW3-14 with NaCl, purified by cobalt affinity chromatography and examined by SDS–PAGE. Data in Figure 1 revealed the presence of a PulD-CShis multimer that co-migrated with PulD-CShis produced by cells with PulS (lanes 3 and 4). Thus, PulD-CShis forms discrete multimers in the absence of PulS. Likewise, PulDhis migrated as a discrete multimeric band in the absence of PulS (Figure 1, lane 2). When dissociated by phenol (lane 8), PulD-CShis without PulS migrated as two bands. N-terminal sequencing of each band revealed a unique sequence (EEF) corresponding to the N-terminus of the short N-terminal region of PulD that is retained in this construct, indicating that the faster-migrating band results from proteolysis near the C-terminus.

Electron microscopy of negatively stained samples of PulD-CShis and PulDhis complexes from envelopes of the strain without PulS revealed multimers that were virtually indistinguishable from those of the same protein produced in a strain with PulS (Figure 2B and C). The diameters of the PulD-CShis particles were identical (14 nm) to those of the same particles purified from the outer membrane. The external diameter of the PulDhis particles was larger (16 nm), as expected from previous studies (Nouwen et al, 1999; Chami et al, 2005).

PulS prevents mislocalization of PulD-CS and PulD to the inner membrane

Immunoblotting of phenol-treated fractions obtained after floatation sucrose gradient centrifugation of envelopes from the protease-deficient strain revealed PulD-CS and full-length PulD monomers predominantly in the inner membrane fraction in the absence of PulS (Supplementary Figure S1C and D). To determine whether these monomers were created by phenol dissociation of PulD-CS and PulD multimers, respectively, proteins were examined by SDS–PAGE without phenol extraction and stained with Coomassie blue. In both cases, a distinct, slow-migrating band was observed (Figure 3) that was absent from control cells without PulD or PulD-CS. Furthermore, these multimers were found exclusively in the same fractions as the authentic inner membrane proteins SecG (detected by immunoblotting) and PspA (detected by protein staining), while outer membrane porins and OmpA were found exclusively in the outer membrane fraction by Coomassie blue staining (Figure 3). These data rule out the possibility that PulD-CS or PulD aggregate when PulS is absent, since aggregated proteins remain at the bottom of the centrifuge tube, or that the inner membrane fraction is contaminated with outer membranes. Furthermore, nonaggregated secretin-like particles identical to those purified from total membranes were found exclusively in the inner membrane fractions from these gradients (not shown). Thus, PulD-CS and PulD multimers both localize to the inner membrane in the absence of PulS.

Figure 3.

Inner membrane localization of PulD (upper panel) and PulD-CS (lower panel) multimers produced in the absence of PulS and DegP detected by separating membranes in sucrose floatation gradients. Fractions from the gradients were loaded onto 7.5 or 10% acrylamide–SDS gels and stained with Coomassie blue (note that PulD-CS multimers were detected using the 7.5% acrylamide gel whereas PulD multimers were detected in the stacking gel). SecG was detected by immunoblotting with specific antibodies. The outer membrane fractions contain proteins OmpF, OmpC and OmpA, whereas inner membrane fractions contain PspA and SecG. The band migrating close to the position of PspA but in the outer membrane fraction is not PspA.

View full figure (116 KB)

To determine whether the PulD and PulD-CS multimers were intimately associated with inner membrane vesicles, they were analysed directly by electron microscopy of negatively stained, large inner membrane vesicles obtained in a cell disrupter. Approximately 5% of the vesicles displayed a few secretin-like rings (Figure 4) that were clearly separated and completely absent from inner membrane vesicles without secretin. Some smaller inner membrane vesicles obtained by disrupting cells in a French pressure cell contained one or two secretin particles (Supplementary Figure S3), confirming that the particles seen in the larger vesicles (Figure 4) were not physically linked. The diameter of the secretin particles was close to that of the purified particles. The stain-filled centre of the barrel has a diameter of 7 nm, as in the purified particles. The vesicles containing secretin rings were clearly different from outer membrane sheets (Supplementary Figure S4).

Figure 4.

Electron microscopy of inner membrane vesicles purified from protease-deficient E. coli producing PulD (A) or PulD-CShis (B) without PulS. Arrows indicate ring-like particles corresponding to secretin particles. The scale bar corresponds to 100 nm.

View full figure (219 KB)

Localization of PulD-CS to the inner membrane does not change its solubilization properties

Data in Figure 4 indicate the PulD-CS multimers are inserted into the inner membrane in the absence of PulS. To validate this interpretation, cells of strain PAP5198 producing PulD-CShis without PulS were resuspended in different concentrations of urea, which should release peripheral membrane proteins but not integral inner membrane proteins, and then disrupted in a French pressure cell. Addition of urea prior to disruption should enable it to access both inner and outer surfaces of the inverted inner membrane vesicles that form under these conditions. Total cell proteins and proteins in the soluble (supernatant) and sedimented (pellet) fractions separated by ultracentrifugation were extracted with phenol and examined by SDS–PAGE and immunoblotting. The peripheral inner membrane protein PspA (see below) was extracted efficiently by 4 M urea and partially by 2 M urea, whereas the integral inner membrane protein SecG was not (Figure 5). Only small amounts of PulD-CShis were extracted by urea, indicating that it behaves like an integral inner membrane protein. Urea (4 M) was also unable to extract PulDhis from the inner membranes of cells without PulS (data not shown). Inspection of membranes after extraction with 4 M urea did not reveal any change in the numbers or distribution of secretin particles (not shown). Only detergents that extracted PulD or PulD-CShis from the outer membrane extracted PulD-CShis from the inner membrane of cells without PulS (Table I).

Figure 5.

Extraction of PulD-CShis, PspA and SecG with urea. Samples were treated with phenol and then subjected to SDS–PAGE and immunoblotting with specific antibodies. Samples were derived from the same amount of starting material. Size markers (kDa) are indicated on the left.

View full figure (37 KB)

PulD-CS induces the phage shock response

The phage shock response, characterized by the massive production of the 26 kDa peripheral membrane protein PspA, is induced following permeabilization of the E. coli inner membrane (Darwin, 2005). Production of PulD-CS in the absence of PulS caused the appearance of a prominent 26 kDa inner membrane protein (Figure 3) that we identified as PspA (Brissette et al, 1990) by immunoblotting (Figure 6). Thus, insertion of PulD-CS into the inner membrane induces the phage shock response. Variants of PulD with substantially reduced ability to form multimers due to the insertion of a 24 amino-acid peptide in the C domain (Guilvout et al, 1999) did not induce PspA production when produced without PulS (two examples are shown in Figure 7), whereas variant D640TEV, which formed multimers efficiently (Guilvout et al, 1999), induced PspA production (Figure 7).

Figure 6.

Induction of PspA caused by production of PulD or PulD-CS in the absence of PulS. Proteins from pooled inner membrane fractions (Supplementary Figure S1) were separated by SDS–PAGE and immunoblotted with PspA-specific antibodies.

View full figure (21 KB)

Figure 7.

PspA induction by PulD derivatives with linker insertions in the C or S domains. Total cell extracts were analysed by SDS–PAGE with or without phenol treatment. Immunoblots were developed with antibodies against PulD and PspA. All extracts are derived from the same amount of starting material.

View full figure (43 KB)

The induction of the phage shock response would be consistent with PulD-induced partial dissipation of the transmembrane electrochemical potential () (Kleerebezem et al, 1996). To test this idea, the of E. coli cells producing PulD or PulD-CS without PulS was measured by their ability to accumulate ³H-triphenyl phoshonium (TPP⁺) (Possot et al, 1997). Production of PulD or PulD-CS (encoded by the high copy number pUC18-derived plasmids pCHAP3671 and pCHAP3711 under lacZp control) in the absence of PulS consistently caused a modest (ca 10%) drop in the (not shown). These plasmids could not be introduced into strain MC4100degP pspAkm F'lacI^q1 (lacking DegP and PspA), despite the presence of the lacI^q1 allele to repress lacZp promoter activity. However, this strain could be transformed with the low copy number plasmid pCHAP362 in which pulD is also under lacZp control (d'Enfert et al, 1989). When grown in liquid cultures, this strain showed a substantial drop in 2 h after IPTG-induced PulD production (Table II). This drop in is probably insufficient to kill the cells, implying that the drops further when PulD production is continued or that other events, such as futile ATP hydrolysis to restore the , caused cell death.

Table 2: Measurement of in cells of strain MC4100 F'lacI^q1 degP with or without pspAKm producing pCHAP362-encoded PulD

View full table

The C-terminal half of the outer membrane secretin PulD (PulD-CS) is an autonomous multimerization and membrane insertion module. The N-terminal half of the polypeptide does not influence either of these properties. The pilotin protein PulS, which binds to the S domain, is not required for correct multimerization but targets secretin to the outer membrane since, in its absence, apparently normal PulD multimers assemble in the inner membrane.

Four independent lines of evidence indicate that PulD and/or PulD-CS inserts into, rather than associates with the inner membrane in the absence of PulS. First of all, very little PulD-CS or PulD is stripped from inner membrane vesicles by urea at concentrations that remove most of the peripheral membrane protein PspA. Second, only detergent extraction conditions that release PulD-CS from the outer membrane (of cells with PulS) can release it from the inner membrane (of cells without PulS). Indeed, the conditions required to extract PulD and PulD-CS are much more drastic than those needed to solubilize other outer membrane proteins, suggesting that factors other than merely dissolving the lipids around the secretin complex and protecting the hydrophobic regions thereby exposed are involved.

Third, electron micrographs revealed PulD-CS or PulD complexes in around 5% inner membrane vesicles. Large vesicles often contained several secretins that were not in direct contact and were in the same orientation, indicating that they are not micro-aggregates and are inserted into the membrane, rather than lying on the surface. Small vesicles (obtained in a French pressure cell) contained only single secretin particles (data not shown), indicating that the limited clustering of secretin particles in large vesicles does not reflect their physical association.

Finally, the induction of the phage shock response by cells producing PulD (or PulD-CS) without PulS is consistent with the formation by the former of a channel in the inner membrane that decreases the across this membrane (Kleerebezem et al, 1996). Although a high level production of export-defective outer membrane proteins or defects in the Sec protein export pathway can also induce high level PspA production, (Kleerebezem and Tommassen, 1993), the underlying mechanisms remain unclear. Induction of the phage shock response by mislocalized PulD cannot be explained by a defect in PulD export across the inner membrane by the Sec system, since PulS probably affects only PulD transport to the outer membrane. PulD has a plug in the centre of the oligomer (Chami et al, 2005) but electrophysiological data indicate that a small conductance can be detected in lipid bilayers containing PulD when a voltage is applied across the membrane (Nouwen et al, 1999). The electrical potential across the inner membrane might cause these channels to open, allowing protons to leak into the cell. The massive amounts of PspA protein produced might prevent complete dissipation of the membrane potential (Hankamer et al, 2004), allowing the cells to remain viable despite the presence of secretin channels in the inner membrane (Model et al, 1997). This would be entirely consistent with the observed lethality of PulD in cells lacking PspA and PulS.

The fact that unusually harsh conditions are required to solubilize PulD from both the outer and inner membrane (see above) suggests that it might have an atypical membrane anchor or an unusually strong interaction with membrane lipids. Like other outer membrane proteins, PulD could be anchored in the membrane by antiparallel amphipathic -strands. PulD has rather less -strand structure than other outer membrane proteins, consistent with the fact that a large proportion of the protein is exposed on the periplasmic side of the membrane (Chami et al, 2005), as in the trimeric TolC protein (Koronakis et al, 1997). In this scenario, multimerization and membrane insertion and organization could be similar to that of the heptameric Staphylococcus aureus haemolysin (Montoya and Gouaux, 2003). Alternatively, like the E. coli Wza protein involved in capsular polysaccharide excretion, which forms rings that are superficially similar to PulD dodecamers (Beis et al, 2004; Dong et al, 2006), the transmembrane segments of PulD could be amphipathic -helices. Since the structures of PulD-CS and PulD in the inner and outer membranes are apparently identical, it seems reasonable to assume that the same -strands or -helices span the lipid bilayer irrespective of the membrane in which PulD is inserted.

Different secretins require different proteins for multimerization and outer membrane insertion (reviewed in Bayan et al, 2006). Formation of stable PilQ multimers in Neisseria requires the lipoprotein PilW (Carbonnelle et al, 2005). PilW is not needed for PilQ outer membrane association, however (Carbonnelle et al, 2005). Likewise, Tgl, which is related to PilW, is required for multimerization of the PilQ secretin in Myxococcus xanthus (Carbonnelle et al, 2005; Nudleman et al, 2006). In A. hydrophila, multimerization of the T2SS secretin ExeD requires the inner membrane proteins ExeA (an ATPase) and ExeB (Ast et al, 2002). The pullulanase T2SS does not have Tgl/PilW homologues, and the ExeB homologue, PulB, is not required for pullulanase secretion (Possot et al, 2000). Thus, PulD differs from PilQ and ExeD in that dodecamer formation occurs in the absence of all other components of the system of which it is part. Furthermore, PulS differs from PilW and Tgl in that it is required to target its cognate secretin to the outer membrane, and not for multimerization. It seems likely, therefore, that secretins such as PilQ require additional, as yet unidentified factors to modulate multimerization and to target them to the outer membrane.

Omp85/YaeT, an essential outer membrane protein, facilitates insertion of integral outer membrane proteins and porin trimerization (Voulhoux et al, 2003; Doerrler and Raetz, 2005; Ruiz et al, 2005; Werner and Misra, 2005; Wu et al, 2005). N. meningitidis PilQ fails to multimerize in the absence of Omp85 (Voulhoux et al, 2003), suggesting that Omp85 and PilW have overlapping functions (Bayan et al, 2006). However, since PulD forms dodecamers in the inner membrane in the absence of PulS, and, therefore, without contact with outer membrane protein YaeT (Omp85), it seems reasonable to assume that PulD multimerization and membrane insertion can be independent of YaeT and, indeed, of any other protein. This raises the question of why PulD and PulD-CS insert exclusively into the inner membrane, rather than into both membranes, when PulS is absent. A likely explanation is that PulS binds rapidly to PulD monomers emerging from the inner membrane to prevent their multimerization, and that the lipoprotein sorting pathway (Tokuda and Matsuyama, 2004) leads them immediately to the outer membrane. PulS binding to the outer membrane receptor LolB could trigger a conformational change in the PulD–PulS complex that allows PulD multimerization and insertion. The naturally constitutive expression of the pulS gene (d'Enfert and Pugsley, 1989) ensures that PulS is present before PulD synthesis is switched on by maltodextrins in the culture medium (d'Enfert et al, 1987). In the absence of PulS, PulD multimers would assemble in the inner membrane by default because they would be too large to diffuse through the peptidoglycan (Demchick and Koch, 1996). Production of plasmid-encoded PulD causes low level induction of the phage shock response even in the presence of PulS (Hardie et al, 1996a, 1996b), suggesting that a small amount of PulD is incorrectly localized despite the presence of its pilotin.

These observations raise the question of whether other outer membrane proteins can or do insert into the inner membrane and the role that Omp85/YaeT plays in outer membrane targeting. Misrouting of many outer membrane proteins to the inner membrane could be more detrimental than it is for secretins. For example, the creation of a porin channel in the inner membrane would almost certainly be lethal. At least three interlocking factors appear to prevent such events. First, the structures of intermediates in their folding pathway might enable porins to interact at an early stage with components exclusively located in the outer membrane, such as lipopolysaccharide (de Cock et al, 1990, 1996). Second, chaperones such as Skp (Chen and Henning, 1996; Schäfer et al, 1999; Harms et al, 2001) might prevent their premature folding (Bulieris et al, 2003) or illicit association with the inner membrane (Mogensen et al, 2005). In their absence, porins would aggregate or be degraded. Third, Omp85/YaeT and associated outer membrane proteins might be absolutely required for the final stages of their membrane insertion and folding (Voulhoux et al, 2003; Doerrler and Raetz, 2005; Wu et al, 2005); hence, ensuring that even if porins do associate with the inner membrane, they cannot insert in a functional, pore-forming configuration. Misrouting of porins to the inner membrane cannot be studied in vivo because of the associated lethality, but their insertion into proteoliposomes might provide a way to test the role of membrane proteins, lipopolysaccharide and chaperones. We note with interest that bacterial outer membrane porins can be reconstituted into lipid bilayers without outer membrane lipopolysaccharide (Kleinschmidt et al, 1999) and that routing of a mitochondrial outer membrane porin to the endoplasmic reticulum leads to the assembly of a functional form of this -barrel protein in the Golgi membranes of eukaryotic cells (Buettner et al, 2000).

Plasmid constructions

To construct plasmids encoding PulD-CS, pCHAP3671 (pUC18 carrying the complete pulD gene) (Guilvout et al, 1999) was cleaved with EcoRI and BglII. The fragment corresponding to pulD was then cleaved with ApoI and PasI. The fragments encoding amino acids 1–42 and 259–634 were purified and ligated in the presence of oligonucleotides 5'-AATTTCTCGACCGCCAGCAGGCGACC-3' and 5'-CTGGGTCGCCTGCTGGCGGTCGAGA-3'. The resulting fragment was ligated back into pCHAP3671 cleaved with EcoRI and BglII to give pCHAP3711. To create pCHAP3713, carrying a 6 His codon insert, pCHAP3711 was cleaved with EcoRI and HpaI and the liberated fragment coding for the N-terminal region of PulD-CS was ligated into pCHAP3678 (pSU18 encoding PulD-his) (Chami et al, 2005) cleaved with the same enzymes. The EcoRI–HindIII fragment from pCHAP3678 and pCHAP3713 were subcloned into pUC18 to give pCHAP3715 (encoding PulD-his) and pCHAP3714 (PulD-CShis), respectively.

Strains, other plasmids and growth conditions

E. coli K-12 strain PAP105 (Guilvout et al, 1999) was used for plasmid construction and verification. Strain PAP105 carrying pCHAP3516 (pulD) (Daefler et al, 1997) and pCHAP5506 (encoding PulS-his) (Chami et al, 2005) was used for experiments on the extraction of PulD from the outer membrane. In other cases, plasmid pCHAP585 (Guilvout et al, 1999) was used as a source of pulS. Strain PAP105(pCHAP1226) carrying all pul genes except pulD in the chromosome) (Possot et al, 2000) was used for complementation assays. E. coli K-12 strain SF120 lacking DegP, OmpT and Ptr protease (Meerman and Georgiou, 1994) and carrying F' lacI^q1 Tn10 (strain PAP5198) was used to examine proteins in the inner membrane. Derivatives of E. coli K-12 strain MC4100 (araD139 lacU169 relA1 rpsL150 thi mot flb-5301 deoC ptsF25 rbsR) carrying pspAkm (M Russel) and/or degP (N Sassoon and J-M Betton) together with F' lacI^q1 Tn10 and pCHAP362 (lacZp-pulD; Cm^R, (d'Enfert et al, 1989)) were used to measure the . Plasmids carrying TEV insertions in pulD were described elsewhere (Guilvout et al, 1999). Bacteria were grown in LB medium (Miller, 1992) containing appropriate antibiotics (100 g/ml ampicillin, 25 g/ml chloramphenicol) at 30°C with vigorous aeration. Isopropyl-thio--galactoside (IPTG) (0.5 mM) was used to induce expression of genes under lacZp control. Maltose (0.4%) was used to induce expression of pul genes other than pulS (which is constitutively expressed).

SDS–PAGE and immunoblotting

Proteins dissolved in buffer containing 2.5% SDS, heated to 100°C for 5 min and separated by SDS–PAGE in 7.5, 10 or 12% acrylamide or on 4–15% acrylamide gradient gels (BioRad) using Tris–HCl-glycine buffers. After electrophoresis, proteins were either fixed and stained with Coomassie blue or were electrotransferred onto nitrocellulose membranes and reacted with PspA, SecG or PulD-specific antibodies and then with horseradish peroxidase–coupled secondary antibodies (Amersham). Bound antibodies were detected by enhanced chemiluminescence (Amersham). Phenol extraction was performed as previously (Hardie et al, 1996a). For N-terminal sequence analysis, proteins were transferred onto PVDF membranes (Millipore).

Membrane fractionation and extraction

Unless otherwise indicated, bacteria were broken in a French pressure cell (Aminco) at 1600 bar. A cell disrupter (Constant Systems) was sometimes used at 1200 bar to obtain larger vesicles. In both cases, membranes separated from the soluble fraction by centrifugation at 180 000 g for 30 min were fractionated by floatation through centrifuged sucrose gradients (Robichon et al, 2005). Membranes were extracted with detergents for 60 min at room temperature with constant mixing. Soluble material was separated from insoluble material by centrifugation for 30 min at 180 000g. For extraction of inner membrane proteins with urea, bacteria were resuspended in 50 mM Tris–HCl buffer (pH 7.5) containing 1 mM EDTA or in the same buffer containing 2 M or 4 M urea for 15 min before being disrupted in the French press. Membranes were then separated from soluble proteins as above.

Purification

PulDhis-PulS and PulD-CShis-PulS complexes were purified by cobalt affinity chromatography from enriched outer membrane fractions as previously (Chami et al, 2005). PulD-CShis and PulDhis were similarly purified from total envelope fractions or from inner membrane vesicles purified by sucrose gradient fractionation, as above.

Electron microscopy and image processing

The inner and outer membrane vesicles purified by sucrose gradient fractionation were pelleted at 250 000 g and the pellets were resuspended in 20 mM Tris–HCl buffer (pH 7.5) containing 200 mM NaCl to remove the sucrose. Vesicles were sometimes incubated in 4 M urea for 30 min to remove surface-bound material and then centrifuged and resuspended as above. Aliquots (5 l) of samples were adsorbed onto glow-discharged 200-mesh carbon coated grid and stained with 2% (W/V) uranyl acetate. The micrographs were recorded at an accelerating voltage of 100 kV and a magnification of 50 000, using a Hitachi 7000 electron microscope. Purified secretin particles were examined by negative staining as previously (Nouwen et al, 1999, 2000). All micrographs were recorded on Kodak SO-163 film.

Reference-free alignment was performed on manually selected particles from digitized electron micrographs using the EMAN image processing package (Ludtke et al, 1999). After using a reference-free alignment procedure, particle projections were classified by multivariant statistical analysis. The class averages with the best signal-to-noise ratio were selected.

We are grateful to Marjorie Russel and Bill Wickner for antibodies against PspA and SecG, respectively, George Georgiou for the protease-depleted E. coli, Marjorie Russel for the pspA mutant, Jean-Michel Betton and Nathalie Sassoon for the degP mutant, Jacques d'Alayer (Institut Pasteur Microsequencing Platform) for N-terminal sequence analysis, Hervé Remigy and Marco Gregorini of the ME Müller Institute at the Biozentrum for fruitful discussions and to the members of the Molecular Genetics Unit of the Institut Pasteur for their constant support and interest. This work was supported, in part, by grants from the EC (HPRN-2000-00075), the French ANR (NT05-3-41560), the Swiss National Research Foundation (Grant 3100-059415), the ME Müller Foundation, and the Swiss National Center of Competence in Research (NCCR) 'Structural Biology'.

Ast V, Schoenhofen I, Langen G, Stratilo C, Chamberlain M, Howard S (2002) Expression of the ExeAB complex of Aeromonas hydrophila is required for the localization and assembly of the ExeD secretion port multimer. Mol Microbiol 44: 217–233 | Article | PubMed | ChemPort |

Bayan N, Guilvout I, Pugsley AP (2006) Secretins take shape. Mol Microbiol 60: 1–4 | Article | PubMed | ChemPort |

Beis K, Collins RF, Ford RC, Kamis AB, Whitfield C, Naismith JH (2004) Three-dimensional structure of Wza, the protein required for translocation of group 1 capsular polysaccharide across the outer membrane of Escherichia coli. J Biol Chem 279: 28227–28232 | Article | PubMed | ISI | ChemPort |

Bitter W, Koster M, Latijnhouwers M, de Cock H, Tommassen J (1998) Formation of oligomeric rings by XcpQ and PilQ, which are involved in protein transport across the outer membrane of Pseudomonas aeruginosa. Mol Microbiol 27: 209–219 | Article | PubMed | ISI | ChemPort |

Brissette JL, Russel M, Weiner L, Model P (1990) Phage shock protein, a stress protein of Escherichia coli. Proc Natl Acad Sci USA 87: 862–866 | Article | PubMed | ChemPort |

Brok R, Van Gelder P, Winterhalter M, Ziese U, Koster AJ, de Cock H, Koster M, Tommassen J, Bitter W (1999) The C-terminal domain of the Pseudomonas secretin XcpQ forms oligomeric rings with pore activity. J Mol Biol 294: 1169–1179 | Article | PubMed | ISI | ChemPort |

Buettner R, Papoutsoglou G, Scemes E, Spray DC, Dermietzel R (2000) Evidence for secretory pathway localization of a voltage-dependent anion channel isoform. Proc Natl Acad Sci USA 97: 3201–3206 | Article | PubMed | ChemPort |

Bulieris PV, Behrens S, Holst O, Kleinschmidt JH (2003) Folding and insertion of the outer membrane protein OmpA is assisted by the chaperone Skp and by lipopolysaccharide. J Biol Chem 278: 9092–9099 | Article | PubMed | ISI | ChemPort |

Burghout P, Van Boxtel R, Van Gelder P, Ringler P, Muller SA, Tommassen J, Koster M (2004) Structure and electrophysiological properties of the YscC secretin from the type III secretion system of Yersinia enterocolitica. J Bacteriol 186: 4645–4654 | Article | PubMed | ChemPort |

Carbonnelle E, Helaine S, Prouvensier L, Nassif X, Pelicic V (2005) Type IV pilus biogenesis in Neisseria meningitidis: PilW is involved in a step occurring after pilus assembly, essential for fibre stability and function. Mol Microbiol 55: 54–64 | Article | PubMed | ChemPort |

Chami M, Guilvout I, Gregorini M, Rémigy H, Müller SA, Valerio M, Engel A, Pugsley AP, Bayan N (2005) Structural insights into the secretin PulD and its trypsin resistant core. J Biol Chem 280: 37732–37741 | Article | PubMed | ISI | ChemPort |

Chen R, Henning U (1996) A periplasmic protein (Skp) of Escherichia coli selectively binds a class of outer membrane proteins. Mol Microbiol 19: 1284–1294 | Article

Collins RF, Frye SA, Kitmitto A, Ford RC, Tonjum T, Derrick JP (2004) Structure of the Neisseria meningitidis outer membrane PilQ secretin complex at 12 Å resolution. J Biol Chem 279: 39750–39756 | Article | PubMed | ChemPort |

Condemine G, Shevchik V (2000) Overproduction of the secretin OutD suppresses the secretion defect of an Erwinia chrysanthemi outB mutant. Microbiology 146: 639–647 | PubMed | ISI | ChemPort |

Crago AM, Koronakis V (1998) Salmonella InvG forms a ring-like multimer that requires the InvH lipoprotein for outer membrane localization. Mol Microbiol 30: 47–56 | Article | PubMed | ISI | ChemPort |

d'Enfert C, Chapon C, Pugsley AP (1987) Export and secretion of the lipoprotein pullulanase by Klebsiella pneumoniae. Mol Microbiol 1: 107–116 | Article | PubMed | ChemPort |

d'Enfert C, Pugsley AP (1989) Klebsiella pneumoniae pulS gene encodes an outer membrane lipoprotein required for pullulanase secretion. J Bacteriol 171: 3673–3679 | PubMed | ChemPort |

d'Enfert C, Reyss I, Wandersman C, Pugsley AP (1989) Protein secretion by Gram-negative bacteria: characterization of two membrane proteins required for pullulanase secretion by Escherichia coli K-12. J Biol Chem 264: 17462–17468 | PubMed | ChemPort |

Daefler S, Guilvout I, Hardie KR, Pugsley AP, Russel M (1997) The C-terminal domain of the secretin PulD contains the binding site for its cognate chaperone, PulS, and confers PulS dependence on pIV^f1 function. Mol Microbiol 24: 465–475 | Article | PubMed | ISI | ChemPort |

Darwin AJ (2005) The phage-shock-protein response. Mol Microbiol 57: 621–628 | Article | PubMed | ChemPort |

de Cock H, Hendriks R, de Vrije T, Tommassen J (1990) Assembly of an in vitro synthesized outer membrane porin into its stable trimeric configuration. J Biol Chem 265: 4646–4651 | PubMed | ChemPort |

de Cock H, van Blokland S, Tommassen J (1996) In vitro insertion and assembly of outer membrane protein PhoE of Escherichia coli K-12 into the outer membrane. Role of Triton X-100. J Biol Chem 271: 12885–12890 | Article | PubMed | ChemPort |

Demchick P, Koch AL (1996) The permeability of the wall fabric of Escherichia coli and Bacillus subtilis. J Bacteriol 178: 768–773 | PubMed | ChemPort |

Doerrler WT, Raetz CR (2005) Loss of outer membrane proteins without inhibition of lipid export in an Escherichia coli YaeT mutant. J Biol Chem 280: 27679–27687 | Article | PubMed | ISI | ChemPort |

Dong C, Beis K, Nesper J, Brunkan-LaMontague A, Clarke B, Whitfield C, Naismith J (2006) Wza, the translocon for E. coli capsular polysaccharides, defines a new class of membrane proteins. Nature (in press)

Guilvout I, Hardie KR, Sauvonnet N, Pugsley AP (1999) Genetic dissection of the outer membrane secretin PulD: are there distinct domains for multimerization and secretion specificity? J Bacteriol 181: 7212–7220 | PubMed | ISI | ChemPort |

Hankamer BD, Elderkin SL, Buck M, Nield J (2004) Organization of the AAA(+) adaptor protein PspA is an oligomeric ring. J Biol Chem 279: 8862–8866 | Article | PubMed | ChemPort |

Hardie KR, Lory S, Pugsley AP (1996a) Insertion of an outer membrane protein in Escherichia coli requires a chaperone-like protein. EMBO J 15: 978–988 | PubMed | ISI | ChemPort |

Hardie KR, Seydel A, Guilvout I, Pugsley AP (1996b) The secretin-specific, chaperone-like protein of the general secretory pathway: separation of proteolytic protection and piloting functions. Mol Microbiol 22: 967–976 | Article | ISI | ChemPort |

Harms N, Koningstein G, Dontje W, Muller M, Oudega B, Luirink J, de Cock H (2001) The early interaction of the outer membrane protein PhoE with the periplasmic chaperone Skp occurs at the cytoplasmic membrane. J Biol Chem 276: 18804–18811 | Article | PubMed | ISI | ChemPort |

Kleerebezem M, Cielaard W, Tommassen J (1996) Involvement of stress protein PspA (phage shock protein A) of Escherichia coli in maintenance of the protonmotive force under stress conditions. EMBO J 15: 162–171 | PubMed | ChemPort |

Kleerebezem M, Tommassen J (1993) Expression of the pspA gene stimulates efficient protein export in Escherichia coli. Mol Microbiol 7: 947–956 | Article | PubMed | ISI | ChemPort |

Kleinschmidt JH, den Blaauwen T, Driessen AJ, Tamm LK (1999) Outer membrane protein A of Escherichia coli inserts and folds into lipid bilayers by a concerted mechanism. Biochemistry 38: 5006–5016 | Article | PubMed | ISI | ChemPort |

Köhler R, Schäfer K, Müller S, Vignon G, Driedrichs K, Philippsen A, Ringler P, Pugsley AP, Engel A, Welte W (2004) Structure and assembly of pseudopilin PulG. Mol Microbiol 54: 647–664 | Article | PubMed | ChemPort |

Koronakis V, Li J, Koronakis E, Stauffer K (1997) Structure of TolC, the outer membrane component of the bacterial type I efflux system, derived from two-dimensional crystals. Mol Microbiol 23: 617–626 | Article | PubMed | ISI | ChemPort |

Ludtke SJ, Baldwin PR, Chiu W (1999) EMAN: semiautomated software for high resolution single-particle reconstructions. J Struct Biol 128: 82–97 | Article | PubMed | ISI | ChemPort |

Meerman HJ, Georgiou G (1994) Construction and characterization of a set of E. coli strains deficient in all known loci affecting the proteolytic stability of secreted recombinant proteins. Biotechnology (NY) 12: 1107–1110 | Article | ChemPort |

Miller JH (1992) A short course in bacterial genetics. A laboratory manual and Handbook for Escherichia coli and related bacteria. Cold Spring Harbor: Cold Spring Harbor Press

Model P, Jovanovic G, Dworkin J (1997) The Escherichia coli phage-shock-protein (psp) operon. Mol Microbiol 24: 255–261 | Article | PubMed | ChemPort |

Mogensen JE, Kleinschmidt JH, Schmidt MA, Otzen DE (2005) Misfolding of a bacterial autotransporter. Protein Sci 14: 2814–2827 | Article | PubMed | ChemPort |

Montoya M, Gouaux E (2003) Beta-barrel membrane protein folding and structure viewed through the lens of alpha-hemolysin. Biochim Biophys Acta 1609: 19–27 | PubMed | ISI | ChemPort |

Nouwen N, Ranson N, Saibil H, Wolpensinger B, Engel A, Ghazi A, Pugsley AP (1999) Secretin PulD: association with pilot protein PulS, structure and ion-conducting channel formation. Proc Natl Acad Sci USA 96: 8173–8177 | Article | PubMed | ChemPort |

Nouwen N, Stahlberg H, Pugsley AP, Engel A (2000) Domain structure of secretin PulD revealed by limited proteolysis and electron microscopy. EMBO J 19: 2229–2236 | Article | PubMed | ISI | ChemPort |

Nudleman E, Wall D, Kaiser D (2006) Polar assembly of the type IV pilus secretin in Myxococcus xanthus. Mol Micro 60: 16–29 | ChemPort |

Opalka N, Beckmann R, Boisset N, Sion MN, Russel M, Darst S (2003) Structure of the filamentous phage pIV multimer by cryo-electron microscopy. J Mol Biol 325: 461–470 | Article | PubMed | ISI | ChemPort |

Possot O, Letellier L, Pugsley AP (1997) Energy requirement for pullulanase secretion by the main terminal branch of the general secretory pathway. Mol Microbiol 24: 457–464 | Article | PubMed | ISI | ChemPort |

Possot O, Vignon G, Bomchil N, Ebel F, Pugsley AP (2000) Multiple interactions between pullulanase secreton components involved in stabilization and cytoplasmic membrane association of PulE. J Bacteriol 182: 2142–2152 | Article | PubMed | ISI | ChemPort |

Robichon C, Vidal-Ingigliardi D, Pugsley AP (2005) Depletion of apolipoprotein N-acyltransferase causes mislocalization of outer membrane lipoproteins in Escherichia coli. J Biol Chem 280: 974–983 | PubMed | ChemPort |

Ruiz N, Falcone B, Kahne D, Silhavy TJ (2005) Chemical conditionality a genetic strategy to probe organelle assembly. Cell 121: 307–317 | Article | PubMed | ISI | ChemPort |

Schäfer U, Beck K, Müller M (1999) Skp, a molecular chaperone of gram-negative bacteria, is required for the formation of soluble periplasmic intermediates of outer membrane proteins. J Biol Chem 274: 24567–24574 | Article | PubMed | ISI | ChemPort |

Tokuda H, Matsuyama S (2004) Sorting of lipoproteins to the outer membrane in E. coli. Biochim Biophys Acta 1694: 1–9 | Article

Voulhoux R, Bos MP, Geurtsen J, Mols M, Tommassen J (2003) Role of a highly conserved bacterial protein in outer membrane protein assembly. Science 299: 262–265 | Article | PubMed | ISI | ChemPort |

Werner J, Misra R (2005) YaeT (Omp85) affects the assembly of lipid-dependent and lipid-independent outer membrane proteins of Escherichia coli. Mol Microbiol 57: 1450–1459 | Article | PubMed | ChemPort |

Wu T, Malinverni J, Ruiz N, Kim S, Silhavy TJ, Kahne D (2005) Identification of a multicomponent complex required for outer membrane biogenesis in Escherichia coli. Cell 121: 235–245 | Article | PubMed | ISI | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising