Portal de Periódicos da CAPES

Dissection of β-barrel outer membrane protein assembly pathways through characterizing BamA POTRA 1 mutants of Escherichia coli

2010; Wiley; Volume: 77; Issue: 5 Linguagem: Inglês

10.1111/j.1365-2958.2010.07280.x

1365-2958

Drew Bennion, Emily S. Charlson, Eric R. Coon, Rajeev Misra,

Cellular transport and secretion

Molecular MicrobiologyVolume 77, Issue 5 p. 1153-1171 Free Access Dissection of β-barrel outer membrane protein assembly pathways through characterizing BamA POTRA 1 mutants of Escherichia coli Drew Bennion, Drew Bennion School of Life Sciences, Arizona State University, Tempe, AZ 85287, USASearch for more papers by this authorEmily S. Charlson, Emily S. Charlson School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USASearch for more papers by this authorEric Coon, Eric Coon School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA College of Medicine, University of Arizona, Tucson, AZ 85724, USASearch for more papers by this authorRajeev Misra, Corresponding Author Rajeev Misra School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA E-mail rajeev.misra@asu.edu; Tel. (+1) 480 965 3320; Fax (+1) 480 965 6899.Search for more papers by this author Drew Bennion, Drew Bennion School of Life Sciences, Arizona State University, Tempe, AZ 85287, USASearch for more papers by this authorEmily S. Charlson, Emily S. Charlson School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USASearch for more papers by this authorEric Coon, Eric Coon School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA College of Medicine, University of Arizona, Tucson, AZ 85724, USASearch for more papers by this authorRajeev Misra, Corresponding Author Rajeev Misra School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA E-mail rajeev.misra@asu.edu; Tel. (+1) 480 965 3320; Fax (+1) 480 965 6899.Search for more papers by this author First published: 25 August 2010 https://doi.org/10.1111/j.1365-2958.2010.07280.xCitations: 92AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Summary BamA of Escherichia coli is an essential component of the hetero-oligomeric machinery that mediates β-barrel outer membrane protein (OMP) assembly. The C- and N-termini of BamA fold into trans-membrane β-barrel and five soluble POTRA domains respectively. Detailed characterization of BamA POTRA 1 missense and deletion mutants revealed two competing OMP assembly pathways, one of which is followed by the archetypal trimeric β-barrel OMPs, OmpF and LamB, and is dependent on POTRA 1. Interestingly, our data suggest that BamA also requires its POTRA 1 domain for proper assembly. The second pathway is independent of POTRA 1 and is exemplified by TolC. Site-specific cross-linking analysis revealed that the POTRA 1 domain of BamA interacts with SurA, a periplasmic chaperone required for the assembly of OmpF and LamB, but not that of TolC and BamA. The data suggest that SurA and BamA POTRA 1 domain function in concert to assist folding and assembly of most β-barrel OMPs except for TolC, which folds into a unique soluble α-helical barrel and an OM-anchored β-barrel. The two assembly pathways finally merge at some step beyond POTRA 1 but presumably before membrane insertion, which is thought to be catalysed by the trans-membrane β-barrel domain of BamA. Introduction Assembly of β-barrel outer membrane proteins (OMPs) in Escherichia coli is facilitated by the Bam complex, comprised of five proteins: BamABCDE, of which BamA and BamD are essential (Doerrler and Raetz, 2005; Werner and Misra, 2005; Wu et al., 2005; Malinverni et al., 2006; Sklar et al., 2007a). BamA is a member of the Omp85 family of proteins that are essential for OMP biogenesis in bacteria, mitochondria and chloroplast (Gentle et al., 2004; Knowles et al., 2009). BamA folds into two distinct domains: an N-terminal soluble domain assembled from five POTRA (polypeptide translocation associated) domains and a C-terminal trans-membrane domain assumed to fold as a β-barrel (Misra, 2007). High-resolution structures of the BamA POTRA domains have been determined (Kim et al., 2007; Gatzeva-Topalova et al., 2008; Knowles et al., 2008). Experimental data have indicated the involvement of POTRA domains in assisting β-barrel OMP assembly (Kim et al., 2007; Knowles et al., 2008). It has been proposed that POTRA–OMP interactions may be mediated through β augmentation, in which a β-strand of the POTRA domain is augmented by the addition of a β-strand from the substrate (Kim et al., 2007; Knowles et al., 2008). These initial interactions could then lead to the assembly of the entire OMP β-barrel, followed by its insertion into the outer membrane presumably with the help of the β-barrel domain of BamA. Prior to interacting with the Bam complex, the nascent mature OMP polypeptides recently released into the periplasm by the Sec machinery must stay assembly competent. For this, most β-barrel OMPs rely on a major periplasmic chaperone SurA (Lazar and Kolter, 1996; Rouvière and Gross, 1996; Vertommen et al., 2009; this study). However, SurA is not essential despite a severe defect of OMP biogenesis in a surA null mutant, indicating the existence of alternative but less efficient pathways for keeping OMPs assembly competent (for reviews, see Mogensen and Otzen, 2005; Ruiz et al., 2006a; Knowles et al., 2009). It has been shown that SurA interacts with BamA (Sklar et al., 2007b; Vuong et al., 2008), and this interaction presumably results in the handing over of SurA-bound OMPs to BamA for assembly and insertion into the outer membrane. Genetic analysis revealed that deletion of the POTRA 1 domain does not disrupt the interactions of BamA with the BamBCDE lipoproteins, but drastically reduces the level of β-barrel OMPs (Kim et al., 2007). Therefore, it is conceivable that the BamA ΔPOTRA 1 mutant is defective in its ability to receive SurA bound to OMPs. Aberrant OMP assembly induces an envelope stress response, which activates sigma E and Cpx regulons (Gerken et al., 2010). The members of these regulons include genes that encode for OMP assembly factors, such as the Bam complex proteins and SurA, a major periplasmic protease DegP, and small regulator RNAs that inhibit translation of OMP mRNA (for reviews, see Raivio and Silhavy, 1999; Johansen et al., 2006). The collective actions of these factors help reduce envelope stress by enhancing folding/degradation of OMPs and inhibiting their synthesis. Suppressor analysis has been a valuable tool for identifying genes involved in the biogenesis of outer membrane components. One of the genes identified through this analysis was lptD (imp) (Sampson et al., 1989). It is now known that LptD (Imp) is involved in LPS transport (Bos et al., 2004). Suppressors of an imp allele, imp4213, led to the isolation of null mutations in bamB (Eggert et al., 2001) and a mutation affecting the POTRA 3 domain of BamA (Ruiz et al., 2006b). In this study we sought suppressors of a different imp allele than previously used, in an attempt to isolate novel mutations in genes involved in outer membrane biogenesis. One such suppressor mutation identified a novel allele of bamA, bamA66, which resulted in the deletion of a single residue from the POTRA 1 domain. Cells expressing BamAΔR64 from a low-copy-number plasmid displayed a dramatic defect in the biogenesis of most β-barrel OMPs except that of TolC. Data showed that BamAΔR64 is severely defective in its ability to interact with SurA. Consistent with this, the POTRA 1 domain of BamA was shown to interact with SurA in close proximity to R64, the POTRA 1 residue that was deleted in BamAΔR64. Analysis of BamAΔR64 and a POTRA 1 deletion mutant, BamAΔR36-K89, allowed the dissection of two distinct β-barrel OMP assembly pathways, one of which is dependent on SurA–POTRA 1 interactions and is followed by most β-barrel OMPs, while the other, which is independent of SurA–POTRA 1 interactions, is followed by TolC. BamA's independence from SurA suggests that it too belongs to the TolC pathway. Both pathways ultimately merge, and in the case of TolC past POTRA 1, for BamA-dependent insertion into the outer membrane. Results Isolation and initial characterization of BamAΔR64 The imp208 allele of lptD (imp) confers hypersensitivity towards a variety of noxious compounds (Sampson et al., 1989). This hypersensitivity phenotype was used to isolate suppressor mutations by growing cultures on medium containing bacitracin. Although the majority of suppressor mutations conferring bacitracin resistance identified null alleles of either yfgM or bamB or missense mutations within imp208 (E. Charlson, E. Coon and R. Misra, unpubl. data), in revertant number 66 the suppressor mutation was found to be genetically linked to bamA. DNA sequence analysis of the entire bamA gene from this revertant revealed a deletion of the R64 codon within the α2 helix of the POTRA 1 domain of BamA. This mutant bamA allele is named bamA66 and the protein is referred to as BamAΔR64. In an lptD+ (imp+) background, bamA66 displays a slight sensitivity to vancomycin and rifampin and causes a moderate reduction (20–30%) in OMP levels, but is synthetically lethal with the depletion of BamB, conditionally lethal in the absence of DegP or SurA; i.e. no growth at or above 37°C and 40°C respectively. All of these observations point to an OMP biogenesis defect of BamAΔR64 that is exacerbated in the absence of other OMP assembly factors. Low expression of BamAΔR64 confers acute phenotypes To further assist in the investigation of BamAΔR64's defects, we introduced the chromosomal bamA66 allele into a low-copy number plasmid pZS21-bamA6His (Kim et al., 2007). The presence of a 6xHis tag at the N-terminal end of BamA is useful for affinity purification to study protein–protein interactions and the detection of BamA. Additionally, the low level constitutive expression of bamA from pZS21 is driven from the tet promoter, thus making BamA synthesis from this construct impervious to σE-mediated regulation. This uncoupling of σE-mediated upregulation of BamA synthesis during envelope stress proved to be important in dissecting BamAΔR64's functional defect (see below), which otherwise remains masked to some degree when BamAΔR64 is expressed from its native, σE-regulated, chromosomal locus. Construction of strains harbouring pZS21-bamA and pZS21-bamA66 plasmids required transformation into a ΔbamA strain expressing BamA from an arabinose-inducible PBAD promoter of a pBAD33-bamA plasmid. Subsequent curing of pBAD33-bamA was readily achieved from the ΔbamA (pZS21-bamA) strain at 37°C; however, curing of pBAD33-bamA from the ΔbamA (pZS21-bamA66) strain was only possible at 30°C. Due to the ts phenotype of the ΔbamA (pZS21-bamA66) strain, all subsequent experiments were carried out at 30°C unless otherwise stated. Effects of BamAΔR64 on β-barrel OMP biogenesis The acute conditional lethal phenotype of bamA66 when expressed from pZS21 indicated a severe OMP biogenesis defect associated with the reduced synthesis of BamAΔR64. To assess the extent of the OMP biogenesis defects we analysed envelope proteins enriched for integral membrane proteins, including OMPs, by two-dimensional isoelectric focusing/SDS-PAGE (sodium dodecyl sulphate polyacrylamide gel electrophoresis). Images of individual gels from bamA+ and bamA66 protein samples in Fig. 1A and B, respectively, were overlaid in Fig. 1C to aid in direct relative comparison. A total of 30 protein spots (including isoforms labelled as a, b or c), numbered in Fig. 1A and B, and listed in Fig. 1E, were identified by MS/MS analysis, of which 20 belong to 15 different OMP species (Fig. 1C). The level of fourteen OMPs, including BamAΔR64, was down significantly in the bamA66 (pZS21) mutant (Fig. 1C and D). Strikingly, intensities of the three TolC spots increased 6.0-, 6.9- and 4.9-fold (P-values of 0.0009, 0.009, and 0.000; n = 2) in the bamA66 (pZS21) strain compared with the bamA+ (pZS21) strain. For better visualization of BamA and LptD (Imp) spots, this region of the gel was enlarged in Fig. 1D. Note that the LptD (Imp) spot is readily detectable by Coomassie blue staining in wild-type strain samples where BamA is expressed from its normal chromosomal copy but only weakly when it is expressed from pZS21 (Fig. 1D). We believe that this difference is due to reduced BamA expression from pZS21 (Fig. 1D), which in turn lowers OMP levels in general. The migration of certain OMPs, including BamA, OmpA, LamB and TolC, as multiple spots during two-dimensional gel electrophoresis is a commonly observed phenomenon, and such heterogeneity is attributed to the existent of differently charged isoforms (Molloy et al., 2000) or insufficiently denatured species (Berven et al., 2003). Figure 1Open in figure viewerPowerPoint Two-dimensional (2D) gel electrophoresis analyses of envelope fractions enriched for integral membrane proteins.A and B. Coomassie blue stained gels containing protein samples from strains expressing bamA+ (A) and bamA66 (B) from pZS21.C. Individual 2D gel images from (A) and (B) are overlaid to show relative protein abundance. Gel images containing protein samples from wild-type and bamA66 strains are coloured green and magenta respectively. Overlapping areas within a spot common in both wild-type and mutant appear black; predominance of green indicates that the corresponding protein spot is more abundant in wild-type than mutant, and predominance of magenta indicates the opposite.D. A montage of individual Coomassie blue stained 2D gels containing envelope samples from wild-type and mutant strains analysed in (A and B). Areas outlined by green show a close-up view of gel regions encompassing BamA and LptD (Imp) spots. All numbered and labelled protein spots were identified using LC/MS/MS and are listed in (E). Isolation of envelopes and enrichment of integral membrane proteins were carried out as described in Experimental procedures. We also fractionated envelopes by sucrose density gradients and determined OMP levels by Coomassie blue staining and Western blots (Figs S1 and S2). In the parental strains (pZS21-bamA+) the major β-barrel OMPs – OmpC, OmpF and OmpA – were present in fractions with buoyant density of 1.226 g cm3−1, which are typical values for the wild-type outer membrane (Figs S1 and S2A). However, in the mutant strain (pZS21-bamA66) these proteins peaked in fractions of a much lower buoyant density (1.181 g cm3−1) owing to a profound defect in the biogenesis of β-barrel OMPs (Figs S1 and S2A). TolC and LptD (Imp) were analysed by Western blots from peak outer membrane fractions. The level of TolC in the outer membrane fractions of the bamA66 (pZS21) mutant was substantially greater than that present in the outer membrane fractions of the wild-type bamA (pZS21) strain (Fig. S2B). Considerably lower levels of LptD (Imp) were detected in the outer membrane fractions of bamA66 (pZS21) envelopes (Fig. S2C). To evaluate relative effects of bamA66 (pZS21) and ΔsurA on OMPs, we also analysed OMPs from wild-type and ΔsurA strains grown at 30°C by 2D gels and sucrose density gradients and found the effects of ΔsurA on OMPs to be similar to that exerted by bamA66 (pZS21) (Figs S1 and S3). The migration of several proteins as multiple spots during isoelectric focusing (Fig. 1) and their presence in multiple fractions during sucrose gradient analysis (Figs S1 and S2) made it somewhat difficult to assess the extent of bamA66 effect on various OMPs. To circumvent this, we performed a side-by-side comparison of TolC and BamA with four other β-barrel OMPs using equal amounts of proteins from purified cell envelopes (Fig. 2). Consistent with the data presented in Figs 1 and S2, TolC levels were higher in the bamA mutant (pZS21-bamA66) than in the parental strain (pZS21-bamA+) (Fig. 2A; lanes 1 and 2). In contrast to elevated TolC levels, the levels of LamB, OmpA, OmpC and OmpF in the bamA mutant decreased significantly while no changes in the level of AcrA, a control inner membrane protein, were observed (Fig. 2A and B; lanes 1 and 2). Interestingly, an increase in TolC level was no longer apparent when the mutant BamAΔR64 protein was expressed from the arabinose-inducible promoter of pBAD33 or synthesized from its native chromosome promoter (Fig. 2A and B; lanes 3–4 and 5–6). Thus, the severity of the BamAΔR64's effect on OMPs was connected to its expression levels, which were lowest from pZS21 and highest from pBAD33 (Fig. 2C). Together, these analyses showed that low expression of a mutant BamA protein negatively affected the biogenesis of several β-barrel OMPs, including BamA, LamB, LptD (Imp), OmpA, OmpC and OmpF, but not that of TolC. OMP data from surA+ and ΔsurA strains are shown in Fig. 2, Figs S1 and S3. Although the overall effects of ΔsurA on OMPs were similar to that of bamA66 (pZS21), the level of BamA, expressed from its native chromosomal location, went up in the ΔsurA strain compared with the isogenic surA+ strain (Fig. 2C; lanes 7–8). Figure 2Open in figure viewerPowerPoint Comparative analysis of selected envelope proteins from wild-type, bamA66 and ΔsurA strains grown at 30°C. Wild-type BamA and BamAΔR64 were expressed from pZS21 (low level constitutive expression), pBAD33 (induced with 0.1% arabinose) or from the native chromosomal location. Lanes 7 and 8 contain envelope samples prepared from surA+ and ΔsurA strains respectively. Proteins were detected either by Western blot analysis using antibodies shown in the figure (A and C) or by SDS(urea)-PAGE analysis followed by Coomassie blue staining (B). AcrA serves as both an envelope preparation and gel loading control. Reverse transcription quantitative PCR analyses to assess gene expression levels We quantified the levels of tolC mRNA, through RT-qPCR, to determine whether transcription or a post transcription process in TolC biogenesis is enhanced when the biogenesis of other β-barrel OMPs is severely impaired in the bamA66 (pZS21) mutant (1, 2, S1 and S2). The expression of tolC remained unchanged in either mutant (Fig. 3), indicating that increased TolC levels in the mutant backgrounds reflect benefit TolC incurs at a post-transcriptional step of its biogenesis. In contrast to tolC, the expression of ompF and lamB, which are indirectly regulated by the activated σE system, was significantly impaired in the mutant strain (Fig. 3). Expression of the two σE-regulated genes, lptD (imp) and degP, went up considerably in the bamA66 (pZS21) mutant, with no significant changes in the expression of bamA from the pZS21 plasmid (Fig. 3). Since the transcription of bamA66 is similar to wild-type bamA when expressed from pZS21, the low BamAΔR64 levels seen in 1, 2 are likely due to a defect in some post-transcriptional step, such as protein folding or assembly. It is important to note that despite increased expression of lptD (imp) in the bamA66 (pZS21) mutant, the levels of LptD (Imp) were significantly down (Fig. S2C), indicating that the assembly of LptD (Imp) depends on BamA POTRA 1. A parallel analysis in a surA mutant background showed that when gene expression was affected, the effects were somewhat milder than that seen in a bamA66 (pZS21) background (Fig. 3). Figure 3Open in figure viewerPowerPoint An examination of relative gene expression in the bamA66 (pZS21) and ΔsurA mutants using quantitative real-time PCR analysis. RNA was isolated from mid-log phase grown cultures at 30°C. Relative quantification of gene transcripts was performed in triplicate using the method and both ndh and ftsL as reference genes. Relative fold-changes in gene expression and error bars representing standard deviation are shown (n = 2). Interestingly, the expression of cpxP, a gene regulated by the CpxAR two-component signal transduction system, went up in the mutant BamA background (Fig. 3). This is consistent with our recent observation that the CpxAR two-component system is not only activated in response to aberrant β-barrel OMP assembly, but is necessary to reduce envelope stress under these conditions (Gerken et al., 2010). Thus, an increase in degP expression is likely the result of the combined effect of activated σE and CpxAR regulons. Distinguishing the effects of bamA66 (pZS21) on OMP assembly from synthesis The above analysis indicated that the assembly, but not the synthesis of BamAΔR64 and LptD (Imp), was compromised in a bamA66 (pZS21) background. However, since the synthesis of OmpF and LamB was drastically inhibited in the mutant BamA background (Fig. 3), reduced OmpF and LamB levels in 1, 2 (lanes 1 and 2) reflect a sum total of defects in synthesis and assembly. In order to distinguish the effects of bamA66 (pZS21) on OmpF and LamB assembly from their synthesis, we synthesized these proteins free from any native regulation by expressing them under the transcription/translation control of the pBAD24 plasmid. To verify the σE-independent expression of OmpF and LamB from pBAD24-ompF and pBAD24-lamB, these plasmids were transformed into wild-type and ΔrseA backgrounds. The synthesis of most β-barrel OMPs produced under their native transcription/translation control is expected to be severely inhibited in the absence of RseA, the negative regulator of σE. Indeed, the levels of chromosomally expressed LamB, OmpA, OmpC and OmpF were dramatically reduced in a ΔrseA background (Fig. 4A). In contrast, no significant change in OmpF and LamB levels were observed when they were expressed from pBAD24 plasmids (Fig. 4A). Note that pBAD24-ompF and pBAD24-lamB also lack OmpR and CpxR binding sites, and hence the activated forms of these regulatory proteins during envelope stress cannot influence plasmid-borne OmpF and LamB synthesis. Figure 4Open in figure viewerPowerPoint The effects of ΔrseA and bamA66 (pZS21) on OmpF and LamB protein levels when expressed from the arabinose PBAD promoter contained on the pBAD24 vector.A. To determine ompF and lamB sensitivity to σE-mediated downregulation, the levels of OmpF and LamB, expressed from either their native chromosomal location (left panel) or from the pBAD24 plasmid (middle and right panels respectively), were evaluated in wild-type (WT) and ΔrseA (Δ) backgrounds. The chromosomal status of ompF and lamB is shown below each panel. For Western blot analysis, 5 ml of LB containing ampicillin was inoculated with overnight grown cultures and grown at 37°C. Upon reaching an OD600 of approximately 0.4, 0.001% arabinose (w/v, final) was added to all cultures for 30 min. After induction, cultures were chilled on ice and whole-cell lysates were analysed by SDS-PAGE. Proteins were detected by Western blots using appropriate polyclonal antibodies.B and C. Expression of ompF (B) and lamB (C) from the chromosome or pBAD24 in backgrounds expressing wild-type (WT) or mutant BamAΔR64 (ΔR64) from pZS21. Methods were followed as described in (A) with the exceptions that whole-cell lysates were obtained from cultures grown at 30°C and inoculated with 0.001% or 0.005% arabinose for 45 or 60 min (B and C respectively). The pBAD24-phoA construct was used as an induction control and GroEL as a sample loading control. Plasmids expressing OmpF and LamB free of σE-mediated downregulation were introduced in the bamA66 (pZS21) background deleted of the chromosomal copy of ompF or lamB, so that plasmid-borne OmpF and LamB levels could be monitored independent from their chromosomally expressed levels. OmpF and LamB expression from plasmids was optimized so as to closely match the levels produced from the chromosomal genes. In the bamA66 (pZS21) background, the plasmid-borne OmpF and LamB levels were reduced by over threefold (Fig. 4B and C). Expectedly, the effects of bamA66 (pZS21) on these proteins were less severe than when they were expressed under their native transcription/translation control elements, because in addition to assembly their synthesis was inhibited by the activated σE regulon (Fig. 3). Lanes 1 and 2 of Fig. 4B and C show a combined assembly and synthesis defect in a bamA66 (pZS21) background, resulting in 25- and 12-fold reduction in OmpF and LamB levels respectively. As a control, we examined the expression of a periplasmic protein, PhoA, produced under the identical transcription/translation control of pBAD24 as that of pBAD24-ompF and pBAD24-lamB. No appreciable effect of bamA66 (pZS21) on plasmid-borne PhoA was observed in either a bamA66 (pZS21) ΔompF or bamA66 (pZS21) ΔlamB background (Fig. 4B and C), indicating that the negative effect of bamA66 (pZS21) was specific to the β-barrel OMPs, OmpF and LamB, and was not caused by their reduced expression from the plasmid replicon. These results showed that although the bulk of the effect of bamA66 (pZS21) on the chromosomally expressed OmpF and LamB proteins is inflicted at their synthesis, the rendering of OMP synthesis immune to such regulatory inhibition exposes the residual effect of BamAΔR64 on OmpF and LamB assembly. From this analysis we conclude that the assembly of OmpF and LamB, like that of LptD (Imp), depends on the wild-type POTRA 1 domain of BamA. Biogenesis of TolC depends on BamA even in a bamA66 (pBAD33) background It was intriguing that while the biogenesis of fourteen different β-barrel OMPs was severely impaired in a bamA66 (pZS21) background, TolC levels went up significantly (1, 2 and S2). This at first appears at odds with our earlier observations showing that TolC assembles in a BamA/BamD-dependent manner (Werner and Misra, 2005; Malinverni et al., 2006). Why then was TolC biogenesis not impaired in the mutant BamAΔR64 background? We considered the possibility that TolC, which assembles independent of SurA, may also bypass the POTRA 1 domain of BamA but rely on other POTRA domains and/or the β-barrel domain of BamA for insertion into the outer membrane. If so, TolC's biogenesis will meet the same fate as other β-barrel OMPs when BamAΔR64 is depleted from the cell. Since BamAΔR64 from pZS21-bamA66 is constitutively expressed at low levels, we opted to use a ΔbamA strain whose survival is dependent on bamA66 expression from the arabinose-inducible promoter of pBAD33. An overnight culture of ΔbamA (pBAD33-bamA66) grown at 30°C in LB (0.1% arabinose) was diluted 1:100 in LB supplemented with or without arabinose and growth was resumed at 30°C. When the cultures reached early stationary growth phase (OD600 of 1.0) they were diluted 1:50 with the same respective fresh medium and growth was resumed (Fig. 5A). Throughout the growth period, aliquots of bacterial samples were withdrawn and OMPs were analysed by Western blot from cell amounts equalized based on OD600. TolC and LamB were readily detected from cells grown in the presence of arabinose (Fig. 5B). However, LamB levels from BamAΔR64-depleted cells (i.e. grown in the absence of arabinose) collected at 3, 4 and 5 h after the first dilution dropped dramatically (Fig. 5B). In contrast to LamB, no significant reduction in TolC levels was noted during the same time points of BamAΔR64 depletion. It should be noted that because LamB synthesis, unlike that of TolC, is subject to inhibition by the activated σE regulon (Fig. 3), a drop in LamB levels prior to that of TolC was to be expected. When BamAΔR64 was further depleted after the second dilution, LamB became undetectable and TolC levels also dropped dramatically (Fig. 5B; hours 9 to 11 without arabinose). Since at these later time points cells ceased to grow without arabinose, a drop in TolC level is unlikely due to dilution of TolC in newly born cells. Biogenesis of OmpC, OmpF and OmpA followed the same trend as that of LamB (data not shown). The levels of AcrA, an inner membrane-anchored periplasmic protein, did not decrease during the entire growth period. Figure 5Open in figure viewerPowerPoint The effects of BamAΔR64 depletion on cell growth and TolC/LamB levels.A. Cells deleted of chromosomal bamA and harbouring pBAD33-bamA66 were grown in LB at 30°C with (circles) and without (squares) arabinose (0.1%, w/v). After the first dilution from overnight cultures, OD600 was measured from samples taken every hour.B. Western blot analysis to detect TolC, LamB and AcrA from equal number of cells (based on OD600) withdrawn at time points indicated by downward arrows in (A). An open downward arrow at the 3 h time point indicates that only the sample from cells grown without arabinose was analysed. AcrA serves as a sample loading control. These results reiterated our previous conclusions that the biogenesis of TolC, like other β-barrel OMPs studied here, is ultimately dependent on BamA though the dependence of the two groups of OMPs on BamA appears to be different: OmpF, LamB, and LptD (Imp) require wild-type POTRA 1 for their assembly while TolC does not. However, subsequent to the POTRA 1 step, the two assembly pathways appear to converge perhaps at the point of BamA's β-barrel domain, which may aid in the insertion of all β-barrel OMPs. A BamA POTRA 1 deletion mutant produces similar albeit a more severe phenotype than BamAΔR64 Though the expression of BamAΔR64 from a low-copy-number plasmid (pZS21) had a severe negative effect on most β-OMPs, TolC biogenesis was not inhibited. It is possible that despite the low levels of BamAΔR64, it retains some residual POTRA 1 activity that is sufficient for TolC assembly but not that of other β-barrel OMPs. We therefore constructed a POTRA 1 deletion mutant (BamAΔR36-K89) lacking R36 to K89 residues of the POTRA 1