Title: The ‘P-usher’, a novel protein transporter involved in fimbrial assembly and TpsA secretion
Abstract: Article2 October 2008free access The ‘P-usher’, a novel protein transporter involved in fimbrial assembly and TpsA secretion Ségolène Ruer Ségolène Ruer UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Search for more papers by this author Geneviève Ball Geneviève Ball UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Search for more papers by this author Alain Filloux Corresponding Author Alain Filloux UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Division of Cell and Molecular Microbiology, CMMI, Imperial College London, London, UK Present address: Division of Cell and Molecular Biology, CMMI, Imperial College London, South Kensington Campus, London, SW7 2AZ UK Search for more papers by this author Sophie de Bentzmann Corresponding Author Sophie de Bentzmann UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Search for more papers by this author Ségolène Ruer Ségolène Ruer UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Search for more papers by this author Geneviève Ball Geneviève Ball UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Search for more papers by this author Alain Filloux Corresponding Author Alain Filloux UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Division of Cell and Molecular Microbiology, CMMI, Imperial College London, London, UK Present address: Division of Cell and Molecular Biology, CMMI, Imperial College London, South Kensington Campus, London, SW7 2AZ UK Search for more papers by this author Sophie de Bentzmann Corresponding Author Sophie de Bentzmann UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France Search for more papers by this author Author Information Ségolène Ruer1, Geneviève Ball1, Alain Filloux 1,2,3 and Sophie de Bentzmann 1 1UPR9027 CNRS, Systémes membranaires et pathogénicité chez Pseudomonas aeruginosa, LISM, IBSM, Marseille, France 2Division of Cell and Molecular Microbiology, CMMI, Imperial College London, London, UK 3Present address: Division of Cell and Molecular Biology, CMMI, Imperial College London, South Kensington Campus, London, SW7 2AZ UK *Corresponding authors: Division of Cell and Molecular Biology, CMMI, Imperial College London, South Kensington Campus, London SW7 2AZ, UK. Tel.: +44 0 2075 949 651; Fax: +44 0 2075 943 069; E-mail: [email protected] UPR9027 CNRS, LISM, IBSM, 31 chemin Joseph Aiguier, 13402 Marseille cedex 20, France. Tel.: +33 491 164 125; Fax: +33 491 712 124; E-mail: [email protected] The EMBO Journal (2008)27:2669-2680https://doi.org/10.1038/emboj.2008.197 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info We identified a new bacterial transporter, the Pseudomonas aeruginosa CupB3 protein, which is an outer membrane usher involved in pili assembly. In CupB3, the usher domain has fused during evolution with a POTRA (polypeptide-transport-associated)-like domain found in TpsB transporters of two-partner secretion systems. In TpsBs, the POTRA captures the TpsA passenger, which is then transported across the outer membrane through the TpsB β-barrel. We named CupB3 a ‘P-usher’ for POTRA-like domain-containing usher. We showed that CupB3 assembles CupB1 fimbrial subunits into pili and secretes CupB5, a TpsA-like protein. The CupB3 usher domain has the function of a TpsB β-barrel in CupB5 translocation. We revealed that the POTRA-like domain is neither essential for CupB1 fimbriae assembly nor for cell surface exposition of CupB5, but is crucial to coordinate bona fide transport of CupB1 and CupB5 through the usher domain. The P-usher defines a novel transport pathway involving a molecular machine made with old spare parts. Introduction Targeting of proteins to subcellular compartments is a major issue in cell biology. This process involves molecular machines for translocation across biological membranes. In Gram-negative bacteria, proteins are released in the milieu owing to secretion pathways (Kostakioti et al, 2005). Common features are shared between secretion systems and molecular machines assembling appendages at the bacterial surface. Type II secretion resembles type IV pili assembly (Filloux, 2004), type III secretion resembles flagellar assembly (Cornelis, 2006) and type IV secretion resembles assembly of conjugative F pili (Christie and Cascales, 2005). The type II–IV secretion machines are complex macromolecular systems, whereas the type V secretion system is simple (Henderson et al, 2004). The type Va consists of a unique protein called autotransporter, whereas the type Vb involves two distinct polypeptides, TpsA and TpsB. TpsA is usually an adhesin, such as the filamentous haemagglutinin (FHA) from Bordetella pertussis (Coutte et al, 2003), which is transported at the surface. This transport involves the TpsB protein, which is a β-barrel outer membrane protein. Recognition between TpsA and TpsB occurs through an interaction between the secretion signal located at the N terminus of TpsA and the N-terminal periplasmic domain of TpsB (Hodak et al, 2006). The TpsB transporters belong to a superfamily of proteins, including the Neisseria meningitidis Omp85 outer membrane protein (Voulhoux et al, 2003) or the eukaryotic Toc75 membrane protein (Schleiff and Soll, 2005), which are required for the insertion of β-barrel membrane proteins into the outer membrane of bacteria or plastids, respectively. These proteins have a C-terminal transmembrane domain forming a β-barrel and an N-terminal region containing 1–5 POTRA (polypeptide-transport-associated) domains, which mediate protein—protein interactions (Gentle et al, 2005). In many cases, the type Vb secretion system, also known as two-partner secretion (TPS) system, transports bacterial adhesins at the cell surface (Coutte et al, 2003). Bacterial adhesive properties are also conferred by cell surface pili, of which the chaperone—usher-dependent assembly of fimbriae is a remarkable example (Dodson et al, 2001). The assembly process relies on an outer membrane protein called usher, which is a pore-forming β-barrel (Henderson et al, 2004; Li et al, 2004). This channel is used by pilin subunits, which polymerize into fimbriae on the bacterial surface. The pilins are targeted to the usher owing to their association with a periplasmic chaperone (Hultgren et al, 1991; Sauer et al, 2004). The opportunistic pathogen Pseudomonas aeruginosa is equipped with at least five genes encoding TpsA-like proteins and three loci encoding a full set of components belonging to the chaperone—usher pathway (Cup) (Vallet et al, 2001; Filloux et al, 2004a, 2004b). We described the function of two P. aeruginosa Cup systems, CupB and CupC (Ruer et al, 2007). Expression of the cupB and cupC genes is controlled by a two-component regulatory system, namely the RocS1-R-A1 system (Kulasekara et al, 2005). On top of genes encoding Cup components, the cupB gene cluster contains the cupB5 gene, which encodes a protein with similarities to TpsA proteins such as FHA from B. pertussis, or HxuA from Haemophilus influenzae (Vallet et al, 2001). This is unusual as no obvious molecular mechanistic relation has ever been shown between these two systems. Yet a C-terminal motif was found in proteins of the Omp85 family, which is closely related to a motif identified for several usher proteins (Reumann et al, 1999). We have no clue on whether cupB5 encodes a cell surface protein and if yes, which secretion mechanism is involved in its transport. These are the two main questions that we addressed in this study and which brought us to the characterization of a novel type of bacterial transporter. Results The cupB5 gene encodes a TpsA-like protein The cupB gene cluster contains six genes encoding an usher (cupB3), a pilin (cupB1), an adhesin (cupB6) and two chaperones (cupB2 and cupB4) (Figure 1A). The cupB5 gene encodes a protein of 1018 residues sharing similarities to TpsA proteins (Jacob-Dubuisson et al, 2001; Henderson et al, 2004). No gene encoding a TpsB transporter was identified in the cupB cluster. Figure 1.Characteristic of CupB5 and other TpsAs. (A) Genetic organization of the cupB gene cluster. The name of the genes is indicated above the arrow and the function is indicated underneath. A code is used for similar functions. (B) A gene cluster encoding a TPS system. The name of the gene is indicated above and the PA number is indicated underneath. TpsB4 is the transporter and TpsA4 is the passenger protein. (C) Domain organization of the CupB5 and TpsA4 proteins. CupB5 and TpsA4 display a haemagglutinin domain (Haemagg_act, bricked boxes) at the N terminus and repeated Glug domains (dotted boxes) at the C terminus. The size in amino acids (aa) of each protein is indicated on the right. (D) Amino-acid sequence alignment of the ‘TPS’ domains of several TpsAs including FHA (B. pertussis), HMW1 and HxuA (H. influenzae), TpsA4 and CupB5 (P. aeruginosa). The NPNL motif in FHA (black) or NPNGI/V motif in all TpsAs is indicated with bold characters. The downstream NTNG motif in HMW1, together with the motif NVA/GG in HxuA, CupB5 and TpsA4, is also indicated with bold characters. The position of the residue is indicated at the end of each line. Download figure Download PowerPoint The CupB5 protein contains a putative 53-amino acids (aa)-long signal peptide reminiscent of unusually extended signal peptide in TpsAs (type Vb) and autotransporters (type Va) (Desvaux et al, 2006), which may exceed 50 aa in length, such as FHA from B. pertussis (Lambert-Buisine et al, 1998). Five tpsA—tpsB gene clusters were identified in the PAO1 genome (Jacob-Dubuisson et al, 2001; Filloux et al, 2004a). CupB5 is an orphan TpsA presenting high similarity (44%) with TpsA4 (Figure 1B). CupB5, similar to TpsA4, contains a haemagglutination activity domain (Haemagg_act) at the N terminus (between aa 53 and 167) (Kajava et al, 2001) (Figure 1C). This domain is found near the N terminus of several TpsAs designated adhesins, FHAs or haem/haemopexin-binding proteins. In CupB5, 13 repeats of Glug domains (17 for TpsA4) are found between aa positions 585 and 973. The Glug domain is found in many protein-containing Haemagg_act domain (Poulsen et al, 1996). TpsAs have an N-terminal secretion signal, or ‘TPS’ domain, which is responsible for the interaction with the TpsB transporter, as shown between FHA and the N terminus of the FhaC transporter (Hodak et al, 2006). In FHA, the ‘TPS’ domain structure presents a β-helix fold (Clantin et al, 2004; Hodak et al, 2006), and secondary structure prediction within the N terminus of CupB5 revealed the presence of a series of β-strands (data not shown). Two conserved motifs, NPNL and NPNGI, have been shown to be important for Serratia marcescens TpsA (ShlA) secretion (Schönherr et al, 1993). In FHA, the NPNGI motif is critical for FHA—FhaC interaction, whereas the NPNL motif is not required for interaction but needed for FHA secretion (Hodak et al, 2006). In HMW1, the NPNL motif is lacking; however, the NPNGI motif and a downstream NTNG motif are required for transport (Grass and St Geme, 2000; Yeo et al, 2007). The CupB5 protein possesses the NPNGV motif (aa 138–142), but lacks the upstream NPNL motif (Figure 1D). This is not unusual for a TpsA protein as HMW1 or HxuA in H. influenzae and TpsA4 of P. aeruginosa are lacking this motif (Jacob-Dubuisson et al, 2001; Yeo et al, 2007). Instead, a NVGG motif is found in CupB5 and TpsA4, or a NVAG motif in HxuA, which align with the HMW1 NTNG motif (Figure 1D). In conclusion, based on in silico analyses, we suggest that CupB5 possesses the characteristics of a TPS substrate. CupB5 is a cell surface-exposed protein We showed that RocS1 overproduction resulted in cupB gene expression and CupB1 fimbriae assembly (Kulasekara et al, 2005; Ruer et al, 2007). The pMMBrocS1 plasmid was thus introduced into the strain PAO1ΔpilAΔfliC (PAOΔΔ) or a cupB5 mutant (PAOΔΔΔcupB5), bacteria were grown on agar plates, scraped and extracts were used for immunoblotting using CupB5 antibodies. The CupB5 protein was readily detected in PAOΔΔ cell extracts (C), but was lacking in the cupB5 mutant (Figure 2A). By performing shearing of bacterial surface appendages (Ruer et al, 2007), large quantities of CupB5 were recovered in the sheared fraction (SF) (Figure 2A), indicating that CupB5 is cell surface exposed. The occurrence of CupB5 at the surface is not resulting from a partial cell leakage as intracellular OTCase (ornithine carbamoyltransferase) (cytoplasmic) and DsbA (periplasmic) proteins were not found in the SF (data not shown). Figure 2.Cell surface exposition of CupB5. (A) Cell extracts (C) or material in sheared fractions (SFs) of P. aeruginosa strains carrying pMMBrocS1 were prepared. Proteins are separated on an SDS gel, blotted on nitrocellulose and CupB5 is revealed using specific antibodies. CupB5 was detected in C and SF extracts from PAOΔΔ but not in the cupB5 mutant derivative (PAOΔΔΔcupB5). (B) Material in SFs of P. aeruginosa strains carrying pMMBrocS1 was prepared and samples were separated as in (A). Coomassie-stained SDS gel and mass spectrometry reveal the presence of CupB5 and/or TpsA4 in the PAOΔΔ strain or isogenic mutants PAOΔΔΔtpsB4 or PAOΔΔΔcupB3. TpsA4 is marked with an asterisk (*), whereas CupB5 is marked with **. (C) Material in sheared fractions (SFs) of P. aeruginosa strains carrying pMMBrocS1 was prepared and the presence of CupB5 was verified by immunoblotting (upper panel). CupB5 was not found in SF from the cupB3 mutant (PAOΔΔΔcupB3). The presence of CupB1 and CupC1 proteins in SF from different P. aeruginosa strains was assessed (lower panel). The CupB1 and CupC1 bands were recovered from a Coomassie-stained gel and analysed by mass spectrometry. In all panels, molecular weight standards are indicated on the left (kDa) and proteins of interest are indicated with arrows. Download figure Download PowerPoint CupB5 export is CupB3 dependent CupB5 exhibits high similarity to the P. aeruginosa TpsA4 protein (Figure 1C and D). We investigated whether the TpsB4 transporter (Figure 1B) could be used for cell surface exposition of CupB5. We engineered a tpsB4 mutant strain (PAO1ΔΔΔtpsB4) in which we introduced pMMBrocS1. SFs of the tpsB4 mutant were analysed on Coomassie-stained gels, and mass spectrometry analysis revealed that CupB5 is still exposed, whereas the TpsA4 protein is missing (Figure 2B). We concluded that CupB5 does not use TpsB4 for transport to the cell surface although TpsB4 transports TpsA4. Furthermore, we found that CupB5 transport is CupB3 usher dependent, as CupB5 was not found in the SF of a cupB3 mutant (Figure 2C, upper panel). The lack of CupB5 transport is specifically observed in the cupB3 mutant, as CupB5 is still surface exposed in the parental strain or the mutant lacking the CupC3 usher, PAO1ΔΔΔcupC3 (Figure 2C, upper panel). CupB5 secretion was restored when the cupB3 mutation was trans-complemented by introduction of the cupB3 gene (pBBRcupB3) (Figure 2C, upper panel). The presence of CupB1 and CupC1 into P. aeruginosa SFs was assessed by mass spectrometry analysis on proteins excised from Coomassie-stained gels (Figure 2C, lower panel). CupB1 was recovered into SF from the parental strain and the cupC3 mutant but not from the cupB3 mutant. CupB1 exposition was observed in the cupB3 mutant on trans-complementation with pBBRcupB3. CupC1 was present in SF of all tested strains except for the cupC3 mutant. Finally, CupB3-dependent transport of CupB5 appears specific as the TpsA4 protein is cell surface exposed in the cupB3 mutant (Figure 2B). CupB5 is associated with CupB1 fimbriae Using transmission electron microscopy coupled with immunogold-labelling and CupB5 antibodies, we analysed a series of P. aeruginosa strains. We observed that CupB5 was located distal from the PAO1ΔΔ bacterial cell surface (Figure 3A and B) and the gold beads are seen within the CupB1 fimbrial material surrounding the bacteria. In a cupB3 mutant, CupC1 fimbriae but not CupB1 fimbriae are assembled, and no CupB5 labelling is observed (Figure 3C). This result confirms that CupB5 export is CupB3-dependent. The specific association of CupB5 with CupB1 fimbriae was also observed when no CupC1 fimbriae are assembled as seen by using a cupC3 mutant (data not shown). Finally, CupB5 labelling was totally absent in a cupB5 mutant (data not shown). Interestingly, in a P. aeruginosa mutant that does not assemble CupC1 and CupB1 fimbriae (PAO1ΔΔΔcupC3ΔcupB1), CupB5 is close to the bacterial surface (Figure 3D), a localization also observed in a single cupB1 mutant (PAO1ΔΔΔcupB1) (data not shown). Figure 3.Cell surface localization of CupB5, in P. aeruginosa strains carrying pMMBrocS1, using transmission electron microscopy and immunogold labelling. (A, B) PAO1ΔΔ. (C) PAO1ΔΔΔcupB3 lacking the CupB3 ‘P-usher’. (D) PAO1ΔΔΔcupC3ΔcupB1 lacking both CupC1 and CupB1 fimbriae. Magnification in (A, C) is × 50 000; (B, D) is × 140 000. Download figure Download PowerPoint Our results suggest that the TpsA-like CupB5 and the pilin CupB1 use the CupB3 usher to reach the cell surface. Moreover, the CupB5 protein remains associated with the CupB1 fimbriae at a distal location from the cell surface. In the absence of CupB1 fimbriae, the CupB5 protein is transported to the cell surface in a CupB3-dependent manner, but is not found at a distal location, suggesting that CupB5 and CupB1 transport and assembly might be coordinated through the CupB3 usher. CupB3 is a POTRA-like-containing usher protein Because of the unique feature of CupB3, we investigated whether lessons could be learned from its primary sequence or predicted secondary structure. We noticed that the 844-aa-long CupB3 protein, as it is annotated at http://www.pseudomonas.com, does not contain a putative signal peptide. We found, 246 bp upstream from the proposed GTG start codon of cupB3, an alternative ATG start codon. The newly defined cupB3 gene encodes a 926-aa-long protein with a predicted signal peptide encompassing residues 1–16 (SignalP). The CupB3 usher is larger than the CupA3 (872 aa) or CupC3 (839 aa) ushers (Figure 4A). The aa sequence comparison between the three ushers revealed that CupB3 has an additional N-terminal domain, which lies between aa 17 and 99 (Figure 4A). Figure 4.POTRA domain characterization. (A) Schematic representation of the CupA3, CupB3 and CupC3 domains. The aa position bordering each domain is indicated. Black bars represent signal peptides and striped bars usher domains. The POTRA-like domain of CupB3 is indicated with a grey box. (B) The POTRA domains of ShlB from S. marcescens, TpsB1—TpsB5 and CupB3 from P. aeruginosa, PFL_1466 from P. fluorescens and the fourth POTRA domain of YaeT from E. coli have been aligned. Conserved amino acids (Sanchez-Pulido et al, 2003) are highlighted in red and blue. The additional residues conserved between CupB3 and TpsB4 are in purple. Amino-acid position within the sequence of each protein has been indicated in green for the first and last residue of each domain. (C) The secondary structures predicted by PSIPRED are shown for the POTRA domains of CupB3, TpsB4 and YaeT (fourth). The regions (Pred for secondary structure and AA for amino acids) corresponding to β-strands are in red (E), whereas the region corresponding to α-helices are in blue (H). The conserved motif GY has been indicated with bold characters and is lacking in CupB3. Download figure Download PowerPoint TpsB transporters contain an N-terminal region, called POTRA domain, which may provide a chaperone-like function for the TpsA passenger (Yang and Braun, 2000). As CupB3 serves for the transport of the TpsA-like CupB5 protein, we investigated whether the N-terminal extension of CupB3 has similarity to POTRA domains. Alignments of the P. aeruginosa TpsBs (Jacob-Dubuisson et al, 2001) with the POTRA domain of ShlB, a canonical S. marcescens TpsB, showed that they all have a putative POTRA including two conserved glycine residues (Figure 4B) (Sanchez-Pulido et al, 2003). We aligned the N terminus of CupB3 with the POTRA domains and found a match between aa positions 19 and 90. We observed that the putative CupB3 POTRA domain lacks the second conserved glycine, whereas it retained the first one (Figure 4B). The lack of this residue in POTRA domains is not unique and was reported in HXB1 from H. influenzae or FtsQ from Escherichia coli (Sanchez-Pulido et al, 2003). The 3D structure of the POTRA domain from the E. coli YaeT protein shows a tandem α-helices bordered with β-strands (β—α—α—β—β) (Kim et al, 2007). The secondary structure of all POTRA domains presented in Figure 4B complied with the β1—α1—α2—β2—β3 structure, except for the proposed CupB3 POTRA domain (Figure 4C), which is predicted to be β1—α1—α2—α3—β2. However, based on overall similarities, we proposed that CupB3 contains a POTRA-like domain between aa positions 19 and 90, and further assessed its function. The CupB3 POTRA-like domain coordinates CupB1 and CupB5 transport We engineered truncated versions of the cupB3 gene (Figure 5A), which encode proteins lacking the N-terminal (aa 20–80; CupB3ΔPN), the C-terminal (aa 37–97; CupB3ΔPC), the central region (aa 37–80; CupB3ΔPI) or the whole POTRA-like domain (aa 20–97; CupB3ΔP). Each cupB3-truncated genes was cloned in the broad host range pBBRMCS-3 plasmid and conjugated into the P. aeruginosa cupB3 mutant containing pMMBrocS1. The cell surface localization of CupB5 was investigated in the cupB3 mutant trans-complemented with either the cupB3-truncated genes or the full-length cupB3 gene (Figure 5B). Surprisingly, although CupB5 was not cell surface exposed in a cupB3 mutant, it could be recovered in the SF on introduction of the cupB3 gene or any of the cupB3-truncated alleles (Figure 5B, upper panel). It suggested that the POTRA-like domain of CupB3 is not required for CupB5 transport to the cell surface. In contrast, we observed that none of the cupB3-truncated genes was able to restore CupB1 fimbriae assembly in the cupB3 mutant (Figure 5B, lower panel). The recovery of the CupB1 subunits in the SF was observed only when using the parental strain or the cupB3 mutant trans-complemented with the full-length cupB3 gene (Figure 5B, lower panel). In all cases, recovery of the CupC1 subunit in the SF was not affected (Figure 5B, lower panel). CupC1 was found in variable amounts and mass spectrometry analysis on the corresponding bands did not reveal the presence of CupB1 degradation products or of another protein. These observations unexpectedly suggested that the POTRA-like domain of CupB3 is required for CupB1 assembly but not for transport of the TpsA-like CupB5 protein. Figure 5.Functionality of POTRA-truncated CupB3 usher. (A) Schematic representation of CupB3-truncated ushers. The CupB3 usher was truncated of the POTRA-like domain between aa 20–97 (CupB3ΔP) or aa 37–80 (CupB3ΔPI) or aa 20–80 (CupB3ΔPN) or aa 37–97 (CupB3ΔPC). The colour code is similar as in Figure 4A. (B) Analysis of sheared fractions obtained from the PAOΔΔ strain and derivative mutants cupB3 (PAOΔΔΔcupB3) or cupB3cupB5 (PAOΔΔΔcupB3ΔcupB5) carrying the pMMBrocS1 plasmid. Where indicated, the strains also carried a plasmid containing the cupB3 gene or one of the truncated cupB3 alleles. Upper panel, immunoblot with antibodies directed against CupB5. Lower panel, Coomassie-stained gel. The positions of CupB5, CupB1 and CupC1 proteins are indicated. The molecular weight standards (kDa) are on the left. (C) Crucial function of the G28 residue in the CupB3 POTRA-like domain. Sheared fractions obtained from the PAOΔΔ strain and derivative cupB3 mutant (PAOΔΔΔcupB3) carrying the pMMBrocS1 plasmid were analysed. Where indicated, the strains also carried a plasmid containing the cupB3 gene or one of the cupB3 alleles encoding the CupB3 derivatives, CupB3G28R or CupB3G28W. Upper panel, immunoblot using antibodies directed against CupB5. Lower panel, Coomassie-stained gel. The positions of CupB5, CupB1 and CupC1 proteins are indicated. The molecular weight standards (kDa) are on the right. Download figure Download PowerPoint We investigated whether lack of CupB1 assembly through the POTRA-defective CupB3 ushers might be caused by an incomplete translocation of CupB5. We thus engineered a cupB5 deletion in the cupB3 mutant strain. Interestingly, trans-complementation of the cupB3cupB5 mutant, with the cupB3 allele encoding CupB3 lacking its POTRA-like domain (cupB3ΔP), readily restored CupB1 fimbriae assembly (Figure 5B, lower panel). This observation indicates that the POTRA-like domain is not essential for CupB1 transport. In contrast, the POTRA-like domain might be required for a bona fide translocation of the TpsA-like CupB5 protein, a crucial event for subsequent CupB1 assembly. A conserved glycine residue within the POTRA-like domain is essential for CupB3 function Amino-acid sequence alignment between POTRA domains revealed the presence of a conserved glycine residue (Sanchez-Pulido et al, 2003). In the CupB3 POTRA-like domain, the conserved glycine is at position 28 (G28) (Figure 4B). We examined whether this residue could have a crucial function in CupB3 function. We engineered cupB3 alleles by substituting the codon corresponding to G28 with an arginine (G28R) or a tryptophan (G28W) codon. These substitutions yielded the cupB3G28R and cupB3G28W genes, respectively. The resulting CupB3 derivatives were unable to restore CupB1 fimbriae assembly in a cupB3 mutant but could support surface exposition of CupB5 (Figure 5C). Moreover, in the double cupB3cupB5 mutant, thus in the absence of CupB5, production of CupB3G28W or CupB3G28R restored CupB1 fimbriae assembly (data not shown). This phenotype is strictly similar to the one observed when cupB3 or cupB3cupB5 mutants are complemented with cupB3 deleted for the region corresponding to the POTRA domain (cupB3ΔP) (Figure 5B). This result confirms that the G28 residue is likely part of a POTRA-like domain. CupB1 and CupB5 are not cell surface exposed in the absence of the CupB4 chaperone The POTRA-like domain of CupB3 does seem to be required for CupB5 translocation but not for its targeting. We investigated whether the CupB4 chaperone could be involved in CupB5 targeting. The CupB4 chaperone is encoded by the cupB4 gene located upstream from the cupB5 gene (Figure 6A). A cupB4 mutant was engineered and shearing experiments were performed. Interestingly, neither CupB1 fimbriae nor the CupB5 protein could be recovered from the cell surface of the cupB4 mutant as compared with the parental strain (Figure 6B, lower and upper panels, respectively). We concluded that the CupB4 chaperone is required, directly or indirectly, for the cell surface exposition of CupB5 and CupB1. Figure 6.CupB4- and TpsBPF-dependent transport of CupB5. (A) Comparison between the genetic organization of the cupB gene cluster of P. aeruginosa (PA) and the PFL_1462—PFL_1469 gene cluster of P. fluorescens (PF). The genes encoding homologous components (function written underneath) are motif-encoded. (B) Analysis of sheared fractions obtained from the PAOΔΔ strain and the isogenic cupB4 mutant (PAOΔΔΔcupB4) carrying the pMMBrocS1 plasmid. The lower panel is a Coomassie-stained gel, whereas the upper panel is an immunoblot revealed with antibodies against CupB5. The proteins of interest are indicated with arrows and the molecular weight standards (kDa) are on the right. (C) PFL_1466/TpsB-dependent transport of P. aeruginosa CupB5. Analysis of sheared fractions obtained from the PAOΔΔ strain and the isogenic cupB3 mutant (PAOΔΔΔcupB3), which contains, or not, a plasmid encoding the P. fluorescens PFL_1466 gene (pBBRtpsBPF) or its derivative encoding a POTRA-truncated TpsB protein (pBBRtpsBPFΔP). Coomassie-stained gel (lower panel) and immunoblot using CupB5 antibodies (upper panel). The proteins of interest are indicated with arrows. The molecular weight standards (kDa) are on the right. Download figure Download PowerPoint The Pseudomonas fluorescens cupB-like cluster encodes a TpsB transporter In the genome of Pseudomonas fluorescens, a cupB-like cluster (PFL_1462—PFL_1469) has been identified (Nuccio and Baümler, 2007) (Figure 6A). The gene cluster organization in P. fluorescens is different as compared with the P. aeruginosa cupB cluster, because it contains a gene encoding the TpsB transporter PFL_1466. We cloned the tpsB (PFL_1466) gene in pBBR1MCS-3, yielding pBBRtpsBPF. We showed that introduction of tpsBPF into the P. aeruginosa cupB3 mutant allowed cell surface exposition of CupB5, which is recovered in SF (Figure 6C, upper panel). In contrast, it could not restore CupB1 fimbriae assembly (Figure 6C, lower panel). PFL_1466 does contain a POTRA domain (Figure 4B). We assessed whether the POTRA is required for CupB5 transport, by engineering a gene encoding a POTRA-truncated version of PFL_1466, namely TpsBPFΔP. When the cupB3 mutant was trans-complemented with tpsBPFΔP, no CupB5 could be recovered in the SF (Figure 6C, upper panel). Thus, transport of P. aeruginosa CupB5 through the P. fluorescens TpsB is dependent