Cterminal amino acid residues of the trimeric ... - Wiley Online Library

Molecular Microbiology (2010) 78(4), 932–946 䊏

doi:10.1111/j.1365-2958.2010.07377.x First published online 29 September 2010

C-terminal amino acid residues of the trimeric autotransporter adhesin YadA of Yersinia enterocolitica are decisive for its recognition and assembly by BamA mmi_7377 932..946

U. Lehr,1† M. Schütz,1*† P. Oberhettinger,1† F. Ruiz-Perez,2 J. W. Donald,3 T. Palmer,3 D. Linke,4 I. R. Henderson5 and I. B. Autenrieth1 1 Institute for Medical Microbiology und Hygiene, University Hospital Tübingen, Elfriede-Aulhornstr. 6, Tübingen D-72076, Germany. 2 Center for Vaccine Development, University of Maryland School of Medicine, Baltimore, MD 21201, USA. 3 College of Life Sciences, University of Dundee, Dundee DD1 5EH, UK. 4 Max-Planck-Institute for Developmental Biology, Department Protein Evolution, Spemannstr. 35-39, Tübingen D-72076, Germany. 5 School of Immunity and Infection, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK.

Summary The Bam complex is a highly conserved multiprotein machine essential for the assembly of b-barrel outer membrane proteins. It is composed of the essential outer membrane protein BamA and four outer membrane associated lipoproteins BamB–E. The Yersinia enterocolitica Adhesin A (YadA) is the prototype of trimeric auotransporter adhesins (TAAs), consisting of a head, stalk and a b-barrel membrane anchor. To investigate the role of BamA in biogenesis of TAAs, we expressed YadA in a BamA-depleted strain of Escherichia coli, which resulted in degradation of YadA. Yeast-two-hybrid experiments and immunofluorescence studies revealed that BamA and YadA interact directly and colocalize. As BamA recognizes the C-terminus of OMPs, we exchanged the nine most C-terminal amino acids of YadA. Substitution of the amino acids in position 1, 3 or 5 from the C-terminus with glycine resulted in DegP-dependent degradation of YadA. Despite degradation all YadA proteins assembled in the outer membrane. In summary we Accepted 26 August, 2010. *For correspondence. E-mail monika. [email protected]; Tel. (+49) 7071 29 81527; Fax (+49) 7071 29 5440. †These authors contributed equally to this study.

© 2010 Blackwell Publishing Ltd

demonstrate that (i) BamA is essential for biogenesis of the TAA YadA, (ii) BamA interacts directly with YadA, (iii) the C-terminal amino acid motif of YadA is important for the BamA-dependent assembly and differs slightly compared with other OMPs, and (iv) BamA and YadA colocalize.

Introduction The translocation of proteins across the outer membrane (OM) is a task accomplished by several secretion systems in Gram-negative bacteria (Kostakioti et al., 2005). Among them is the autotransporter pathway, which appears to be the most basic system since much of the machinery for mediating their transport across the outer lipid bilayer is intrinsically encoded in members of this protein family (Henderson et al., 2004). All autotransporter proteins contain three functional domains (Henderson et al., 1998): the N-terminal Sec-signal sequence, the passenger domain and the translocation unit at the C-terminus. The Sec-signal sequence targets the protein to the inner membrane for its Sec-dependent export into the periplasmic space and is then cleaved off. The passenger domain confers diverse effector functions to the autotransporters, once they are inserted in the bacterial OM. The translocation unit forms a b-barrel and thus anchors the protein to the OM (Desvaux et al., 2004; Kostakioti et al., 2005). Over recent years the existence of a subfamily of autotransporter proteins has been established, named the trimeric autotransporter adhesins (TAAs). A structural characteristic of TAAs is the formation of homotrimeric complexes, which when assembled appear similar to those of conventional autotransporters (Cotter et al., 2005; Linke et al., 2006). Members of the TAA protein family have adhesive functions in a-, b- and gproteobacteria (Linke et al., 2006). One of the best characterized TAAs is YadA of Yersinia enterocolitica, which is a human pathogen causing a number of diseases ranging from enterocolitis, acute enteritis and mesenteric lymphadenitis to autoimmune disorders (Bottone, 1997). The passenger domain of YadA contains an N-terminal head domain, which is a left-handed b-roll (Nummelin

YadA is recognized and assembled by the BAM complex 933

et al., 2004) and is responsible for autoagglutination and binding to host cell matrix proteins (el Tahir and Skurnik, 2001). The head domain is connected to the coiled-coil of the stalk domain (Alvarez et al., 2010), which serves as a spacer between the head domain and the bacterial OM. The translocation domain of YadA is a 12-stranded b-barrel called the membrane anchor. The membrane anchor is connected to the stalk domain by a-helices, which presumably pass through the pore formed by the oligomerized anchor as a coiled-coil before adopting their a-helical conformation (Wollmann et al., 2006). Being a homotrimer, each YadA monomer contributes the same portion to the whole structure of the complex (Nummelin et al., 2004; Koretke et al., 2006; Wollmann et al., 2006). Although many details are known regarding the structure of TAAs and conventional autotransporters, their mechanism of passenger domain translocation across the OM is still unclear. Several models have been proposed which attribute OM translocation entirely to the translocation domain, however only the ‘hairpin model’ in which the translocation domain secretes the passenger domain, starting with its C-terminus, remains creditable (Pohlner et al., 1987; Oomen et al., 2004). Other models involve BamA (formerly YaeT/Omp85), which is a highly conserved protein required for the OM assembly of b-barrel proteins (Voulhoux et al., 2003; Oomen et al., 2004; Voulhoux and Tommassen, 2004; Gentle et al., 2005; Wu et al., 2005). Indeed, the requirement of BamA was recently shown for the monomeric autotransporters IgA protease (Neisseria meningitidis), IcsA and SepA (Shigella flexneri ), AIDA-I (Escherichia coli ) and BrkA (Bordetella sp.) and more recently for the autotransporter haemoglobin protease (E. coli ) (Voulhoux et al., 2003; Jain and Goldberg, 2007; Sauri et al., 2009). Moreover also direct interaction between BamA and the EspP autotransporter (E. coli ) has been demonstrated (Ieva and Bernstein, 2009; Ruiz-Perez et al., 2009). However, while the requirement for BamA in the assembly of the conventional autotransporter b-barrel can easily be envisaged, a mechanism for trimeric autotransporter assembly where each monomer donates only part of the b-strands for the barrel is hard to devise. Nevertheless, true homology between b-barrels of monomeric and trimeric autotransporters suggest that the mechanism of insertion may share similarities (Remmert et al., 2010). BamA of E. coli consists of an integral OM b-barrel protein with five N-terminal domains residing in the periplasm (Voulhoux et al., 2003). In other species, the number of N-terminal periplasmic domains may vary (Arnold et al., 2010). These periplasmic domains are called polypeptide-transport-associated (POTRA) domains (Sanchez-Pulido et al., 2003). BamA forms a complex with four peripheral OM lipoproteins (BamB–E) and these proteins act in concert to correctly assemble

OMPs into the OM. While detailed mechanistic understanding of how the BAM complex functions is missing, it is clear that the POTRA domains are necessary for the substrate recognition by BamA (Kim et al., 2007; Knowles et al., 2008). Moreover, POTRA domains could interact with peptides mimicking nascent unfolded OMPs, implying the recognition of alternating hydrophobic and hydrophilic amino acids in the sequence of b-barrels (Knowles et al., 2008). Furthermore, unfolded OM proteins were able to modulate channel properties of BamA in vitro. Replacement of a C-terminal Phe, which is highly conserved at the C-terminus of OM proteins (Struyve et al., 1991), and the amino acid at the penultimate position, changed this interaction (Robert et al., 2006). In order to determine whether BamA is required for the biogenesis of trimeric autotransporter proteins, we expressed YadA in a BamA-depleted E. coli strain. We demonstrate for the first time localization of BamA within the cell envelope and a direct interaction between BamA and a TAA. To characterize the biogenesis of YadA in more detail, we introduced single amino acid exchanges into the C-terminus of YadA. Finally, we determined the effects of these C-terminal amino acid substitutions on YadA OM insertion, transport and surface display.

Results BamA is essential for the biogenesis of YadA Recently it was shown that BamA is required for the assembly of monomeric autotransporter proteins in bacterial OMs (Voulhoux et al., 2003; Jain and Goldberg, 2007). Based on this finding, we hypothesized that BamA also might be involved in the biogenesis of TAAs such as YadA. To address this question, we first created a BamA conditional depletion strain in an E. coli MC4100A background (see Experimental procedures). The ability of this strain to grow in the presence of arabinose or glucose was determined; the strain grew well in the presence of arabinose (E. coli BamA+), but failed to grow on subculturing into glucose (E. coli BamA-; Fig. 1A). Depletion of BamA in the presence of glucose was determined by Western blotting. This revealed that BamA was completely depleted after 2 h of growth (Fig. 1B). Therefore the impact of BamA on biogenesis and autotransport of YadA was determined 2 h after initiating BamA depletion. Subsequently, an anhydrotetracycline (AHTC)-inducible expression vector containing the yadA0:8 gene was transformed into E. coli MC4100A wild type and the corresponding BamA depletion strain. It is possible to analyse biogenesis of YadA in E. coli as Y. enterocolitica and E. coli BamA proteins are 79% identical, and the identity increases to 88% (similarity 95%) when comparing only the POTRA domains which are supposed to harbour the

© 2010 Blackwell Publishing Ltd, Molecular Microbiology, 78, 932–946

934 U. Lehr et al. 䊏

Fig. 1. YadA expression in E. coli wild type (wt) or BamA-mutant E. coli strain under conditions promoting BamA depletion (BamA-) or BamA rescue (BamA+). A. Bacteria were grown overnight in medium containing arabinose and diluted in the appropriate medium next day. Growth curves are shown of E. coli wild type or BamA-mutant E. coli strains cultured in LB medium supplied with glucose (BamA-) or arabinose (BamA+). After 3 h of growth, the cultures were diluted 1:50 in order to keep them in the exponential growth phase. The strain grown under BamA- conditions fails to grow on subculturing. B. Depletion of BamA upon growth in medium containing glucose. After 2 h of growth BamA is no more detectable in Western blot. DegP levels are shown to demonstrate equal loading. C. Levels of YadA and BamA in OM fractions of bacterial cells. Bacteria were grown for 2 h under BamA depletion or rescue conditions, and then expression of YadA was induced. OM fractions were prepared 0, 30, 90 and 150 min after the induction of YadA expression. YadA and BamA protein were detected by Western blot analysis. D. Immunofluorescence staining pattern and protein levels of YadA and BamA in bacteria with or without BamA expression. Bacteria were fixed before, 5 min or 2 h after the induction of YadA expression and stained for immunofluorescence microscopy using antibodies against YadA and BamA. All pictures were taken with identical exposure times. Exposure times were chosen such that also BamA- bacteria could be seen due to background signal in order to show that there were bacteria present. E. Localization of YadA and BamA at the bacterial surface. Cells were fixed 5 min or 2 h after YadA induction. Co-immunostaining of YadA (green) and BamA (red) was carried out on all bacteria. Pictures were digitally processed to enhance the staining pattern.



species–specific recognition site. Moreover, the Cterminal motif of BamA, which is necessary for selfrecognition, is 100% identical (Fig. S1). Growth of these strains was determined in medium supplied with glucose for BamA depletion or arabinose for BamA rescue, and BamA levels were determined. Growth rates were comparable for both strains and conditions until 5 h after the onset of BamA depletion. To rule out that AHTC entrance is hindered in BamA-depleted cells, we used an AHTCinducible plasmid carrying the coding sequence of a colicin derivative (Arnold et al., 2009), a protein unrelated to BamA. As protein levels of ColicinS4 R1-His after induction were comparable in all samples independently from the presence of BamA, we assume that AHTC can effectively enter BamA-depleted cells (Fig. S2). To test if the presence of BamA has an influence on YadA biogenesis, we first grew the bacteria for 2 h in the presence of glucose or arabinose respectively. Then, OM fractions were prepared 0, 30, 90 and 150 min after induction of YadA expression. We could detect YadA in all samples after YadA expression had been induced in the E. coli wild type and in the bamA-mutant E. coli strain growing under BamA rescue conditions. In contrast, YadA was absent in BamA-depleted E. coli (Fig. 1C). The apparent reduction in BamA signal in OM preparations over time is due to technical reasons and cannot be seen in whole-cell lysates. With ongoing YadA protein production the cells tend to form aggregates and the efficiency of lysis is reduced. Therefore, the overall loss of protein is greater in samples from late time points. As we loaded equal volumes of OM preparation, it might look like BamA is less present in samples prepared after 90 and 150 min. To investigate these findings in more detail, immunofluorescence microscopy of whole bacterial cells was performed. After 2 h of induction, YadA was observed in an intensive, typical OM staining in E. coli wild type and the BamA-rescued mutant strain whereas the BamAdepleted E. coli strain displayed a much weaker YadA immunostaining (Fig. 1D). Immediately after the induction of YadA, the immunofluorescence staining revealed a dotted staining pattern on the bacterial surface in all strains. A comparable pattern was observed in all examined strains stained for BamA, independently of YadA expression. After prolonged induction of YadA expression in E. coli wild type and in the BamA-rescued strain, the intensity of the BamA staining signals increased (Fig. 1D). Colocalization and direct interaction of BamA and YadA To test whether BamA was directly responsible for YadA assembly we performed colocalization studies and yeasttwo-hybrid (Y2H) analyses. To address whether BamA and YadA proteins colocalize, double-immunofluorescence

stainings of whole bacterial cells were carried out. The data revealed that YadA and BamA fluorescence signals overlap immediately after the induction of YadA. Upon prolonged expression, YadA was found distributed all over the bacterial surface, whereas BamA still displayed a punctate staining pattern (Fig. 1E). However, at present we cannot rule out fixation artefacts as bacteria had to be fixed for this staining. To further explore whether BamA interacted directly with YadA we performed Y2H studies (Fig. 2). The constructs applied to the Y2H screen are shown in Fig. 2A. By means of Y2H experiments, we provide direct evidence of interaction between the YadA C-terminal domain including the entire b-barrel domain (aa 362–422) and the BamA protein as manifested by the ability to rescue growth on minimal medium (Fig. 2B). This interaction was further supported by the ability to drive expression of the lacZ reporter gene to levels significantly higher than the negative control and similar to those of the positive control (Fig. 2C). Taken together, we conclude that BamA is necessary and essential for the biogenesis of the TAA YadA, and that the fluorescence signals derived from the staining of YadA and BamA at least overlap in E. coli immediately after the onset of YadA expression and show direct interaction in Y2H. Furthermore, the immunostaining indicates that YadA can be found all over the cell envelope (Grosskinsky et al., 2007; Schütz et al., 2010). Likewise, BamA is localized in a punctuate manner over the whole bacterial surface. Variation of the C-terminus of YadA involved in the interaction with BamA leads to increased degradation of YadA The C-terminus of b-barrel OM proteins plays a crucial role in substrate recognition by BamA (Robert et al., 2006). As YadA biogenesis depends on BamA we asked whether the YadA C-terminus fits the postulated requirements for interaction of BamA and OMPs. Alignments of the C-terminal sequences of Y. enterocolitica and E. coli OMPs and the resulting average motifs are shown in Fig. 3A. The Y. enterocolitica YadA C-terminal sequence (GSSDVMYNASFNIEW) follows the anti-parallel b-strand pattern of typical OMPs where every alternative position is occupied by a hydrophobic or aromatic residue but differs somewhat from the average motif, whereas Y. enterocolitica and E. coli average motifs are practically identical. As Phe at the ultimate position of the PhoE C-terminus was shown to be recognized by BamA (Robert et al., 2006) we changed the ultimate amino acid in YadA to Phe (W422F) and the penultimate residue Glu to the positively charged amino acids Lys and Arg (E421K, E421R) and the hydrophobic residue Leu (E421L) respectively. The nine amino acids framing the most C-terminal b-strand of



Fig. 2. Yeast-two-hybrid analysis of the interaction of the YadA membrane anchor domain with BamA, the periplasmic chaperones SurA and Skp and the Bam complex component BamC. A. Constructs employed in Y2H experiments. The bamA, surA, skp and bamC genes lacking the N-terminal signal sequence were cloned downstream of haemagglutinin (HA)/myc epitopes in pGBKT7 and pGADT7. The C-terminal portion of YadA (aa 362–422) comprising the four beta-strands building up the trimeric pore in the assembled trimer was cloned into both pGBKT7 and pGADT7. The structure and functional domains of full-length YadA with amino acid co-ordinates are shown at the bottom. B. Growth phenotype of the yeast AH109 strain transformed with the two-hybrid system constructs. The yeast strain harbouring the construct expressing the Gal4 DNA-BD-YadA-b or Pet-b (Knowles et al., 2008) proteins was transformed with Gal4 AD-Chaperone or BamA protein constructs and evaluated in minimal medium for protein–protein interaction. Only yeast harbouring both the plasmids encoding the bait and the prey protein can grow on minimal medium without Leucine and Tryptophan. If bait and prey protein show interaction, growth on minimal medium additionally lacking Histidine and Adenine is allowed due to reporter gene expression. C. The protein–protein interactions were quantified by b-galactosidase assays. Due to reporter gene activation if bait and prey protein do interact, b-galactosidase is expressed and can be quantified. The b-galactosidase assay values represent averages of three independent measurements. L/T = minimal medium without Leucine and Tryptophan. L/T/H/A = minimal medium without Leucine, Tryptophan, Histidine and Adenine.



YadA that displayed reduced transport efficacy (Grosskinsky et al., 2007). To determine whether degradation of YadA carrying C-terminal amino acid substitutions occurs in the periplasm, W422G was expressed in an E. coli strain lacking DegP, the major protease in the E. coli periplasm (Werner and Misra, 2005). The experiments were carried out with bacteria grown at 27°C as the DegPdeficient E. coli strain does not grow well at 37°C. In the absence of DegP, the expression of YadA W422G protein was comparable to YadA wild type (Fig. 4C). YadA W422G was significantly better expressed and was detected even as trimeric protein in E. coli wild-type strain grown at 27°C compared with 37°C which can be explained by the fact that the protease activity of DegP is optimized at this temperature while at 27°C YadA may fold more efficiently and DegP protease activity is reduced. YadA carrying C-terminal amino acid substitutions co-fractionates with the OM of E. coli As all YadA proteins were expressed at least to some extent in E. coli MC4100A, we wondered whether the

Fig. 3. Amino acid sequence alignment of the C-terminal signature sequence of OM proteins from Y. enterocolitica 8081 and E. coli CFT 073. A. OM proteins from Y. enterocolitica 8081 and E. coli CFT 073 were identified using HHOmp (Remmert et al., 2009). Sequences were considered to be a b-barrel OM protein down to a cut-off of 95%. Patterns were computed using MEME (Bailey and Elkan, 1994). The Y. enterocolitica YadA C-terminal sequence (GSSDVMYNASFNIEW) differs significantly from the average motif, whereas Y. enterocolitica and E. coli average motifs are practically identical. B. Alignment of the C-termini of selected OMPs from Y. enterocolitica and E. coli and an overview of YadA harbouring single C-terminal amino acid exchanges is depicted.

the YadA b-barrel (Y414-W422) were additionally substituted each by a Gly residue (Fig. 3B). Expression of the C-terminally altered YadA proteins was examined in E. coli MC4100A wild type after treatment with AHTC for 2 h. In Western blots of whole-cell lysates most of the proteins could be detected, F418G and especially W422G were hardly detectable. The protein level of YadA I420G was reduced compared with YadA wild type (Fig. 4A). In overloaded gels, F418G and also W422G could be detected (Fig. 4B). Interestingly, YadA S417G turned out to be even more stable than the wild-type protein with dominant trimer formation (Fig. 4A). In YadA E421G no significant trimer formation could be observed. Previous work from this laboratory demonstrated that DegP is involved in degradation of mutated

Fig. 4. Expression of YadA wild type and YadA carrying single amino acid substitutions in E. coli wild type. A. Whole-cell lysates were prepared 2 h after the induction of YadA expression and subjected to SDS-PAGE and Western blot analysis. B. Low amounts of YadA F418G, I420G and W422G are present in whole-cell lysates. After overloading of SDS gels, YadA trimers and monomers can be detected. C. Protein level of YadA W422G is restored in E. coli DdegP. E. coli wild type and E. coli DdegP were grown at 37°C and 27°C (wt) or 27°C (DdegP) respectively.



Fig. 5. Surface display of YadA wild type and YadA carrying C-terminal amino acid substitutions in E. coli wild-type strain. A. Bacteria were fixed 2 h after the induction of YadA and were stained for immunofluorescence microscopy. Pictures were taken with constant exposure times. B. In living bacteria the accessibility of the YadA proteins for the protease trypsin was investigated. Following incubation with trypsin (designated ‘+’) or without trypsin (designated ‘-’), whole-cell lysates were prepared. Western blot analysis was carried out using antibodies directed against YadA (upper blot) and BamA (lower blot).

C-terminally modified versions of YadA are associated with the bacterial OM. To answer this, OM fractions of bacteria expressing the YadA proteins were prepared. Western blot analysis of the obtained fractions revealed the presence of YadA in all strains expressing mutated versions or the YadA wild type (Fig. S2). Equal loading of the OM fraction was confirmed by staining BamA in the samples, and contamination by periplasmic proteins was ruled out by staining for the periplasmic maltose binding protein which turned out to be absent from this protein fraction (data not shown). Our data indicate that even YadA F418G and W422G, which show greatly reduced presence in the OM, co-purify with the OM fraction. However, as it cannot be excluded that we co-fractionated periplasmic aggregates and non-inserted membraneassociated YadA with the OM, we performed additional assays to ensure that YadA trimers are indeed inserted into the OM. Transport across the OM is not impaired by C-terminal modification of YadA In order to further analyse surface display of YadA, we performed immunofluorescence experiments. Expression

of YadA wild type and most of the YadA versions including Y414G, N415G, A416G, S417G, N419G, E421G, E421K, E421L, E421R and W422F for 2 h yielded a typical ringlike OM staining pattern with fluorescence intensities comparable to that obtained with YadA wild type, whereas YadA F418G, I420G and W422G obtained a strikingly weaker fluorescence signal (Fig. 5A). This finding is in accordance with the low production of these YadA proteins compared with YadA wild type as demonstrated by Western blot analysis (Fig. 4A and Fig. S3). To demonstrate surface display by a second method, the tryptic accessibility of the YadA proteins was assessed. Whole bacterial cells were incubated with trypsin and then subjected to Western blot analysis. Several bands of less molecular weight than trimeric or monomeric YadA protein were detected with all YadA versions indicating their tryptic digestion (Fig. 5B). As a control, staining of BamA was performed. BamA was not degraded by trypsin treatment. This confirmed OM integrity after trypsin digestion of whole bacterial cells (Fig. 5B). Taken together, these data suggest that despite reduced transport and surface display of YadA mutants, all YadA mutants were transported to and exposed on the outer bacterial membrane. © 2010 Blackwell Publishing Ltd, Molecular Microbiology, 78, 932–946


(Fig. 6B). We found that in E. coli expressing YadA carrying C-terminal amino acid substitutions the detected protein was largely present in the trimeric form. Taken together, this indicates that (i) trimer formation occurred in all of the YadA versions, but (ii) most of the proteins were severely affected in stability of trimers. Our results demonstrate that exchange of the hydrophobic residues at position 1, 3 and 5 from the C-terminus of YadA with glycine leads to proteolytic degradation in the periplasm of E. coli. Periplasmic accumulation and consequently the DegP-dependent degradation of YadA are presumably caused by the inefficient substrate recognition by BamA. Although some C-terminal substitutions severely affect the trimeric stability of YadA, these proteins are not degraded suggesting that trimeric stability of YadA and recognition of YadA by BamA depend on different features.

Fig. 6. Trimer stability. A. Trimer destabilization in YadA harbouring C-terminal amino acid exchanges using urea. Whole-cell lysates containing 6 M urea were incubated for 10 min at 85°C and subjected to SDS-PAGE and Western blot analysis. B. Trimeric YadA in C-terminally modified YadA versions are present in E. coli wild type when whole-cell lysates were prepared for SDS-PAGE and Western blot analysis without heating the samples.

Amino acid substitutions in the C-terminus of YadA may lead to altered trimer stability In addition to the periplasmic degradation of distinct YadA proteins, our Western blots revealed varying trimeric stability of the C-terminally modified YadA versions (Fig. 4A and B and Fig. S3). Wild-type YadA is known to form stable trimers which cannot be disrupted by heating in SDS sample buffer (Roggenkamp et al., 2003). This results in a band of approximately 200 kDa for trimeric YadA and several bands of intermediate size in Western blots. For each of the YadA proteins, SDS-PAGE with heated samples led to different ratios of their trimeric to monomeric state. Especially amino acid replacements at position 420 and 421 had severe effects on trimeric stability, as no or significantly reduced amounts of trimeric protein could be detected (Fig. 4 and Fig. S3). To investigate the stability of YadA trimers in more detail, whole-cell lysates were destabilized in 6 M urea at 85°C. In most of the C-terminally modified proteins trimers were disrupted to monomers comparable to the YadA wild-type protein. Only the trimeric YadA W422F was as stable as the YadA wild-type trimer; YadA S417G showed even enhanced trimer stability (Fig. 6A). In addition, we examined unheated samples of whole-cell lysates by Western blot in order to address whether at semi-native conditions, trimer formation of YadA can be detected

Decreased protein stability does not influence the autoagglutination properties of YadA In order to find out whether the E. coli strains expressing YadA protein with C-terminal amino acid substitutions are affected in their ability to confer autoagglutination which is a major characteristic of YadA (Bolin et al., 1982; Skurnik et al., 1984), we monitored autoagglutination in culture tubes and by microscopy. With exception of YadA F418G, I420G and W422G producing strains, all YadA proteins mediated rapid autoagglutination (Fig. 7, upper panel). Microscopic examination of the cultures confirmed that strains producing YadA F418G, I420G and W422G were present as single bacteria, but did not form aggregates. We tested if the non-agglutination phenotype of these strains could be rescued by lowering the growth temperature to 27°C (Fig. 7, lower panel). At 27°C the periplasmic protease DegP which is responsible for the degradation of insertion-hampered OM proteins exerts reduced protease activity. Additionally, at lower temperatures YadA folding might be improved. Lowering the temperature enabled YadA F418G and I420G producing strains to form agglutinates. Only YadA W422G producing E. coli did not clump. These results demonstrate that most YadA proteins carrying C-terminal amino acid substitutions can mediate autoagglutination supporting the view that in all strains, except the one producing YadA W422G, critical quantities of trimeric YadA protein are assembled, transported and exposed at the bacterial OM.

Discussion Trimeric autotransporter adhesins are non-fimbrial adhesins which have been demonstrated to be important virulence factors in Gram-negative bacteria. The molecular basis of assembly and autotransport of these molecules



Fig. 7. Autoagglutination mediated by wild type and YadA with C-terminal amino acid exchanges. After 2 h of induction, cultures were allowed to settle for 2.5 h. Bacterial sedimentation and clearance of the bacterial culture was recorded with a digital camera (small inserts show photos of the tubes) and the sediment was investigated microscopically by phase contrast microscopy.

are largely unknown. Recent discoveries demonstrated that a core complex, designated as b-barrel assembly machinery, BamA, plays a crucial role in folding and membrane insertion of OM proteins (Knowles et al., 2009). In the present study we investigated the role of BamA for assembly and autotransport of YadA, the prototype of non-fimbrial adhesins, the TAAs. In particular we addressed the role of C-terminal amino acid residues of YadA that might be involved in recognition of BamA. The most striking finding of our study is that BamA is essentially required for efficient biogenesis, membrane insertion and surface display of YadA, as in BamA-depleted bacteria YadA was degraded by the periplasmic protease DegP. Moreover, immediately after the onset of YadA expression, both BamA and YadA appear to colocalize at the bacterial OM, and a direct interaction of YadA and BamA appears to be necessary for biogenesis and surface display of YadA. In order to define the interaction between YadA and BamA in more detail, we exchanged single amino acids in the C-terminus of YadA by glycine which also resulted in the degradation of YadA. A striking effect was seen when aromatic amino acids located at the positions 1 and 5 from the C-terminus were changed to glycine. Replacement of the hydrophobic amino acid at

position 3 from the C-terminus resulted in slight protein degradation as well. Stability of YadA trimers was influenced by single amino acid exchanges of most of the nine C-terminal residues. Nevertheless, despite these changes and less efficacy in expression, membrane insertion and transport, all YadA protein versions were integrated into and transported across the OM. Requirement of BamA for autotransport of YadA The requirement of BamA in the biogenesis of YadA was predicted by two models which attempt to explain the mechanisms underlying autotransport. Oomen et al. (2004) were the first to propose a model which includes the assistance of BamA in the OM translocation of autotransporter proteins (‘model A’). The authors hypothesized that BamA not only integrates b-barrels of autotransporters into the OM, but also translocates the passenger domain of autotransporter proteins across the OM; the C-terminal b-barrel being the signal targeting the protein to the BamA complex. A modification of this model (‘model B’) was put forward later, stressing the fact that the structures of all autotransporters solved until today show a-helices traversing the membrane region inside a © 2010 Blackwell Publishing Ltd, Molecular Microbiology, 78, 932–946


b-barrel. In model B, the passenger domain of autotransporters passes through the b-barrel of the membrane domain while or after the b-barrel is stabilized in a widened form by BamA, allowing even the secretion of partially folded domains which was reported for some conventional autotransporters (Dautin and Bernstein, 2007). In the study presented here, BamA depletion led to the degradation of YadA, demonstrating the requirement of BamA in the biogenesis of YadA. In addition, exchange of the most C-terminal Trp by Gly caused degradation of the resulting YadA W422G. A recent study showed a direct interaction between BamA and the C-terminus of its substrate proteins. Especially the most C-terminal amino acid residue was shown to be decisive for this interaction (Robert et al., 2006). Therefore, YadA W422G likely interacts poorly with BamA and therefore is targeted for degradation. The restoration of W422G protein production in a bacterial strain lacking the periplasmic protease DegP supports this conclusion. In fact, insufficient OM insertion leads inevitably to accumulation of proteins in the periplasm triggering activation of DegP. Therefore YadA appears to be integrated into the OM upon recognition by and via the help of BamA. Regarding the translocation of autotransporter passenger domains across the OM, both models A and B suggest this occurs during the period of interaction between autotransporters and BamA. Using immunofluorescence microscopy we were able to show colocalization of YadA with BamA at the OM immediately after the induction of YadA expression. Effects of YadA trimerization on interaction with BamA A characteristic feature of YadA is its trimerization. As each of the YadA monomers contributes four b-strands to the b-barrel of the membrane anchor, trimerization of the membrane anchor very likely is a prerequisite for insertion of YadA into the OM and the transport of the passenger domain to the surface. This raises the question as to whether the degree of trimeric stability of YadA affects the interaction with BamA. The various C-terminal YadA amino acid substitutions generated in this study resulted in variable trimeric stability of YadA. The most severe effects were seen upon substitution of the amino acids S417 or E421. While S417G displayed increased trimeric stability compared with the YadA wild type, substitution of E421 decreased trimeric stability. Great stability of the YadA trimer seems to be beneficial, as at position 417 Gly is highly conserved across TAAs and a Ser at this position is present only in few TAAs. This finding is also supported by recent data from our lab, where we demonstrated that a Y. enterocolitica strain expressing YadA with reduced trimer stability, though to wild-type levels, is severely

attenuated in the mouse model (Schütz et al., 2010). Nevertheless, both YadA S417G and E421G were transported across the OM and resulted in a functional surface display of YadA, as shown by immunofluorescence microscopy, tryptic digestion and autoagglutination assays. YadA F418G, I420G and W422G, which were obviously degraded in the periplasm by DegP due to an impaired interaction with BamA, also were affected in their trimeric stability. In previous work we investigated the role of a conserved glycine (G389) residue in the membrane anchor of YadA by exchanging it with amino acids of increasing side-chain size (Grosskinsky et al., 2007). In these YadA proteins the decreased trimeric stability was closely associated with the increasing side-chain size of the amino acid replacing the glycine. Moreover, YadA G389substituted protein was degraded in the periplasm by DegP. Intriguingly, YadA G389-substituted versions display a similar phenotype as the C-terminal modified YadA proteins F418G, I420G and W422G suggesting that the trimeric stability of YadA may well influence interaction with BamA. This could also contribute to ensure that monomers of different TAAs do not trimerize in one cell. Substrate recognition by BamA The amino acid sequence of OM b-barrels is constituted of alternating hydrophobic and hydrophilic amino acids in the region of the b-sheets. Despite of the multitude of hydrophobic amino acids, in most b-barrel proteins a Phe is found at the ultimate position of the C-terminus (Struyve et al., 1991). The importance of this Phe was investigated by in vitro studies with PhoE with regard to substrate recognition by BamA, and was found to play a decisive role in this process. Moreover, in the same study the amino acid at the position penultimate to the C-terminus was also found to be important, as a positively loaded amino acid was able to inhibit the interaction with BamA (Robert et al., 2006). Based on these results, the authors suggested that a signature sequence in OM proteins is responsible for their interaction with BamA. They hypothesized that this speciesspecific C-terminal signature motif consists of a Phe or Trp at the C-terminal position, and hydrophobic residues at positions 3, 5, 7 and 9 from the C-terminus (Robert et al., 2006; Bos et al., 2007). In our study we exchanged each single amino acid residue from position 1 to 9 in the C-terminus of YadA by Gly. As a control we exchanged the most C-terminal Trp to Phe resulting in W422F, the evolutionarily preferred amino acid in E. coli (Struyve et al., 1991). This substitution in YadA obviously did not influence the biogenesis or trimeric stability of the protein and therefore YadA appears to be recognized by BamA in a similar way as E. coli proteins.



However, we found severe degradation of YadA proteins modified at the positions 1 (W422G) and 5 (F418G) from the C-terminus. Minor protein degradation was seen upon amino acid exchange at the position 3 from the C-terminus (I420G). As the nine most C-terminal amino acids of YadA form one b-sheet in the YadA b-barrel (Wollmann et al., 2006), BamA substrate recognition may depend on distinct amino acids in the C-terminal region of this b-sheet. Therefore, we conclude that BamA may recognize primarily hydrophobic amino acids at the positions 1 and 5 from the C-terminus. A hydrophobic amino acid at position 3 from the C-terminus also seems to positively affect the recognition process. To examine the BamA signal motif in more detail, we investigated the effects of amino acid exchanges at the position 2 from the C-terminus. In YadA wild type the negatively loaded amino acid Glu is found at this position. It has been shown previously for the neisserial protein PorA that a positively loaded amino acid at this position interferes with substrate recognition by BamA (Robert et al., 2006). In order to find out whether this is also true for YadA, we exchanged this amino acid with Arg (E421R) and Lys (E421K). Both amino acid substitutions influenced the trimeric stability of the YadA proteins; however, degradation was not observed. This indicates that positively loaded amino acids at the position penultimate to the C-terminus of YadA do not influence the recognition by BamA. In order to disrupt the pattern of alternating hydrophobic and hydrophilic amino acids at the C-terminus of the YadA b-barrel, we exchanged E421 by the hydrophobic amino acid Leu (E421L). As was observed for YadA E421R and E421K, YadA E421L displayed decreased trimeric stability compared with YadA wild type. Moreover, YadA E421L was integrated and translocated efficiently across the OM, implying efficient interaction with BamA. Remarkably, YadA F418G and W422G were not degraded completely in the bacterial periplasm. Low levels of these proteins were assembled in the OM. This is in accordance to what was previously observed when neisserial PorA was expressed in E. coli (Robert et al., 2006). Based on this and other findings, Bos et al. (2007) suggested the existence of an alternative, less efficient signal motif targeting b-barrel proteins to the OM, a hypothesis that is supported by our data. In conclusion, we have shown that YadA requires and interacts with BamA for its OM assembly; whether this is also true for other members of the TAA protein family remains to be shown in future studies. Furthermore, we have demonstrated that BamA does not recognize the whole C-terminal b-sheet of TAAs, but distinct amino acids at the positions 1, 3 and 5 from the C-terminus of YadA. Future studies will have to demonstrate which part of BamA actually interacts with which amino acid residues in monomeric autotransporters or TAAs.

Experimental procedures Generation of a BamA depletion strain The bamA gene with its stop codon and ribosome binding site was amplified with primers bamAEcoF and bamABglR, digested with EcoRI and BglII and cloned into pQE60 that had been similarly digested, to give pQEyaeT. To place bamA under control of the araBAD promoter, the PBAD promoter together with its regulator, araC, were amplified from pBAD18 (Guzman et al., 1995) using primers BADBglAtt and BADEco, digested with AatII and EcoRI, and cloned into similarly digested pQEyaeT to give plasmid pQEArayaeT. The method of Simons et al. (1987) was subsequently used to move the araC +, PBAD, bamA + allele into the lambda attachment site (attB) of the E. coli chromosome. However, since plasmid pRS552 can only be used for the insertion of cloned DNA of up to 2.9 kb onto the E. coli chromosome, it was first necessary to construct a modified version of pRS552. This was achieved following digestion of pRS552 with EcoRV and StuI, which removed approximately 4 kb of lac DNA followed by religation, to give plasmid pRS552-l to permit cloned DNA of up to 7 kb to be successfully incorporated into the attB site. The araC +, PBAD, bamA + allele was excised from pQEArayaeT by digestion with BglII and cloned into pRS552-l that had been pre-digested with BamHI. This construct was used to transfer the entire arabinose-regulated bamA allele to the lambda attachment site (attB) on the chromosome of E. coli host strain MC1061 according to Simons et al. (1987). The attB::(araC +, PBAD, bamA +) allele, now 100% linked to a kanamycin resistance marker, was then moved by P1 transduction to the chromosome of MC4100A, an arabinose-resistant isolate of MC4100 (Ize et al., 2002). For deletion of the native bamA gene, a method based on the l-redirect system was employed (Datsenko and Wanner, 2000). Briefly 500 bp of DNA upstream of bamA including its start codon was amplified with BamABamup and BamAHindup, digested with BamHI and HindIII and cloned into pBluescript that had been previously digested with the same enzymes. Into this was cloned an apramycin resistance cassette, flanked by FRT sites, as a HindIII–ApaI fragment [amplified with primers ApraHindIII and ApraApaI using plasmid pIJ773 (Gust et al., 2003) as template]. Finally into this construct the downstream part of the bamA deletion, covering the final 21 codons of the gene and 750 bp of downstream DNA, was cloned as an ApaI–KpnI fragment, amplified using primers BamAApadown and BamAKpndown. Using primers BamABamup and BamAKpndown, the whole insert was amplified by PCR and transformed by electroporation into strain MC4100A attB::(araC +, PBAD, bamA +) harbouring plasmid pIJ790, following the method of Datsenko and Wanner (2000), with the exception that 0.2% arabinose was included throughout the procedure to ensure that the arabinose-controlled copy of bamA was expressed. After selection for apramycinresistant colonies at 37°C, and confirmation by PCR analysis that replacement of native bamA by the marked deletion had occurred, the apramycin cassette was subsequently ‘flipped out’ to leave a non-polar scar sequence as described (Datsenko and Wanner, 2000). © 2010 Blackwell Publishing Ltd, Molecular Microbiology, 78, 932–946


Table 1. Primers used in this study. Primer name

Sequence of forward primer

Y414G N415G A416G S417G F418G N419G I420G E421G W422G E421L E421K E421R W422F bamAEcoF bamABglR BADBglAtt BADEco BamABamup BamAHindup ApraHindIII ApraApaI BamAApadown BamAKpndown YadA-BANdeIfwd YadA-BBamHIrev YaeT-fullNdeIfwd YaeT-fullBamHIrev

5′-CGGATGTCATGGGCAATGCATCATTT-3′ 5′-GATGTCATGTACGGTGCATCATTTAATATC-3′ 5′-GATGTCATGTACAATGGATCATTTAATATCG-3′ 5′-CATGTACAATGCAGGATTTAATATCGAG-3′ 5′-GATGTCATGTACAATGCATCAGGTAATATCGAGTGGTAATATCAT-3′ 5′-CAATGCATCATTTGGTATCGAGTGGTAATATC-3′ 5′-GCATCATTTAATGGCGAGTGGTAATATC-3′ 5′-GCATCATTTAATATCGGGTGGTAATATCATTTAG-3′ 5′-AATGCATCATTTAATATCGAGGGGTAATATCATTTAGAAGTTAAC-3′ 5′-GCATCATTTAATATCCTGTGGTAATATCATTTAG-3′ 5′-GCATCATTTAATATCAAGTGGTAATATCATTTAG-3′ 5′-TACAATGCATCATTTAATATCCGGTGGTAATATCATTTAGAAGTT-3′ 5′-AATGCATCATTTAATATCGAGTTTTAATATCATTTAGAAGTTAAC-3′ 5′-GGAATTCAGGAAGAACGCATAATAACG-3′ 5′-CGAGATCTTTACCAGGTTTTACCGATGTTAAAC-3′ 5′-GCGCGACGTCAGATCTTTATGACAACTTGACGGC-3′ 5′-GCGCGAATTCCCCGGGTACCGAGCT-3′ 5′-GCGCGGATCCGGTAAAGCGATTGGTTTTG-3′ 5′-GCGCAAGCTTCATGGTTATTATGCGTTC-3′ 5′-GCGCAAGCTTATTCCGGGGATCCGTCGACC-3′ 5′-GCGCGGGCCCTGTAGGCTGGAGCTGCTTC-3′ 5′-GCGCGGGCCCTTCAAAAAGTACGATGGAGACAAG-3′ 5′-GCGCGGTACCGGTTAACCATGAACGTAATGTGACC-3′ 5′-GCGCCAACTATTAATTTGTTCCAGCCATATGGTGTGG-3′ 5′-CCGGTTACCGGATCCTTACCACTCGATATTAAATGATGC-3′ 5′-GCGCCAATTCATATGGCTGAAGGGTTCGTAGTGAAA-3′ 5′-CCGGTTACCGGATCCTTACCAGGTTTTACCGATGTTAA-3′

For site-directed mutagenesis pairs of primers were used where the reverse primers (not shown) are identical in sequence to the forward primers but complement reverse. The positions of single amino acid substitutions are underlined.

Generation of anti-BamA antibodies To raise antibodies against BamA, DNA constructs encoding amino acids 1–424 (containing POTRA domains 1–5) of E. coli BamA were cloned into the pQE70 expression vector and overexpressed in E. coli M15pREP4 strain (Qiagen) to produce C-terminal His-tagged fusion proteins. Protein lacking the N-terminal signal peptide was purified from the periplasmic compartment of cells induced at 37°C for 4 h in Luria–Bertani broth. Purification was based on nickel affinity, Q sepharose ion exchange and Superdex 75 gel filtration chromatography and the monomeric and folded states were determined by size exclusion chromatography, analytical ultracentrifugation and small angle X-ray scattering, as described previously (Knowles et al., 2008). Rabbit antibodies were raised against the folded POTRA1–5 by Eurogentec (Belgium) using their 28-day Super Speedy polyclonal antibody protocol.

DNA constructs To generate single amino acid substitutions in the C-terminus of YadA we used a previously generated construct of the expression vector pASK-IBA2 (IBA GmbH, Göttingen, Germany). Here, the yadA gene from Y. enterocolitica serotype O:8 was inserted under the control of an AHTC-inducible promoter. Using the HerculaseR II Fusion DNA Polymerase in the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, California) and the primer pairs listed in Table 1,

point mutations were introduced into yadA0:8. Subsequently, all mutants were verified by DNA sequencing.

Bacterial strains and growth conditions Escherichia coli K-12 derivative MC4100A was transformed with pASK-IBA2 expression vectors containing wild type or mutant yadA0:8. The plasmid containing yadA0:8 wild type was also introduced into the corresponding BamA depletion strain. Bacterial strains were grown at 37°C in Luria–Bertani (LB) broth with 100 mg ml-1 ampicillin. Additionally, BamAmutant strains were supplied with 0.2% glucose for BamA depletion or 0.2% arabinose (w/v) for BamA rescue and 50 mg ml-1 kanamycin. Overnight cultures were resuspended in fresh medium and diluted to an OD600 of 0.05. Bacterial cultures were grown for 2 h at 37°C before AHTC was added to a final concentration of 40 ng ml-1. If not stated otherwise, bacteria were then allowed to express wild type or YadA carrying amino acid substitutions for another 2 h. Generation of E. coliBL21Omp8DdegP containing the pASK-IBA2 expression vector with the yadA0:8 gene was described earlier (Grosskinsky et al., 2007). With this strain, bacterial cultures were diluted to an optical density at 600 nm of 0.1 from overnight cultures and grown at 27°C, as growth is inhibited at 37°C due to the absence of degP.

Sample preparation for Western blot analysis For preparation of whole-cell lysates bacterial pellets were resuspended in SDS sample buffer to obtain 5 ¥ 106 bacteria



per ml and incubated for 10 min at 95°C. In order to assess the trimeric stability of wild type or YadA carrying single amino acid replacements, urea was added to whole-cell lysates to a final concentration of 6 M and the samples were incubated for 10 min at 85°C. Surface localization of proteins was determined by incubating live bacteria with 0.1 mg ml-1 Trypsin (1 mg ml-1, Applichem) in PBS for 30 min at 37°C. Digestion was stopped by addition of 0.2 mg trypsine inhibitor (Applichem) and whole-cell lysates were made. The preparation of unheated samples was performed as described previously (Grosskinsky et al., 2007).

Western blot analysis Proteins resolved by SDS-PAGE were transferred onto nitrocellulose membranes. The membranes were blocked overnight with Western wash (150 mM NaCl, 10 mM Tris, 0.5% Tween-20, pH 7.4) – 5% milk powder (w/v) at 4°C. Blots were probed with the purified IgG fraction of polyclonal rabbit antiYadA, anti-BamA (1:5000) or anti-DegP sera (1:10 000) and a peroxidase-conjugated secondary anti-rabbit antibody (diluted 1:10 000; Dianova, Hamburg, Germany). BamA sera had been pre-adsorbed against PFA-fixed and BamAdepleted E. coli before.

Preparation of OM fractions Preparation of OMs was carried out using 50 ml of bacterial culture. Cells were harvested and resuspended in 250 ml of resuspension buffer (0.2 M Tris, 1 M Sucrose,1 mM EDTA, pH 8). After addition of 250 mg of lysozyme (20 U mg-1, MSB) and 1.6 ml of water the samples were incubated for 15 min at room temperature. Protoplasts were lysed in 2.5 ml of lysis buffer (2% Triton X-100, 50 mM Tris, 10 mM MgCl2, pH 8) and released DNA was digested by addition of 50 mg of DNase I (10 mg ml-1, Roche). OMs were pelleted by centrifugation at 40 000 g for 1 h. After three washing steps with water the membranes were resuspended in SDS sample buffer.

Immunofluorescence microscopy For immunofluorescence stainings 1 ¥ 107 bacteria in PBS were centrifuged on collagen-coated coverslips and fixed with 4% paraformaldehyde (PFA) in PBS (w/v) at room temperature. For BamA stainings, bacterial cell walls were permeabilized in 0.5% Triton X-100/PBS (v/v). Rabbit serum directed against BamA was pre-adsorbed against PFA-fixed and BamA-depleted bacteria. Stainings were performed using polyclonal rabbit antibodies directed against YadAO:8 (diluted 1:200) or BamA (diluted 1:250, pre-adsorbed) and a 1:200 dilution of a Cy2-conjugated secondary anti-rabbit antibody (Dianova, Hamburg, Germany). Antibodies were diluted in 5% fetal calf serum in PBS (v/v) and incubated at room temperature for 1 h in a dark chamber. Bacterial DNA was stained with 0.01 mg ml-1 4′,6-diamidino-2-phenylindole (DAPI). For double stainings of YadA and BamA, polyclonal rabbit anti-YadA0:8 antibodies were labelled using the AlexaFluor 488 Protein Labelling Kit (Invitrogen, Carlsbad, California). After fixation and permeabilization of 2 ¥ 107 bacteria on collagen-coated coverslips, the cells were first incu-

bated overnight at 4°C with pre-adsorbed rabbit anti-BamA antiserum using a 1:100 dilution. Then an AMCA-conjugated anti-rabbit antibody (Dianova, Hamburg, Germany) in a 1:150 dilution and the AlexaFluor 488-labelled anti-YadA0:8 antibodies in a 1:150 dilution were applied successively for 1 h at room temperature in a dark chamber. Finally, coverslips were mounted. Fluorescence images were obtained with a DMRE fluorescence microscope (Leica, Wetzlar, Germany).

Autoagglutination Autoagglutination was investigated in bacterial cultures macroscopically and microscopically. Bacteria were grown as described, then the cultures were incubated at room temperature without agitation for 2.5 h. Rapid clearance of the medium and aggregated bacteria at the bottom of the tubes were recorded with a digital camera. For the microscopic determination of autoagglutination, 5 ml of the aggregates were pipetted onto a slide, sealed with a coverslip and examined under phase contrast. Images were obtained with a DMRE fluorescence microscope (Leica, Wetzlar, Germany).

Yeast-two-hybrid analyses The bamA gene without its signal sequence was amplified from E. coli MG100 using appropriate primer YaeTfullNdeIfwd and YaeT-fullBamHIrev. The PCR product was digested with NdeI/BamHl and cloned in the pGKBT7 and pGADT7 Y2H plasmids (Clontech) previously digested with the same enzymes. The portion of yadA encoding the b-barrel (YadA-b) was amplified using primers YadABANDeIfwd and YadA-BBamHIrev. The amplification product was digested with NdeI and BamHl restriction enzymes and cloned into the Y2H plasmids previously digested with the same enzymes. PCR cycling conditions were as follow: denaturing 95°C/5 min, followed by 30 cycles of denaturing 94°C/ 2 min, annealing 60°C/30 s and extension 68°C/2 min, and a final cycle of 68°C/10 min. The GAL4-based two-hybrid system contained the DNA-binding domain (BD) in pGBKT7 vector and the activation domain (AD) in pGADT7 vector. Analyses of protein–protein interactions using the Y2H system were conducted as described previously, following standard procedures for the Matchmaker two-hybrid system-3 (Clontech, Mountain View, CA) (Knowles et al., 2008). Briefly, AH109 yeast cells were transformed and plated onto SD minimal medium lacking leucine and tryptophan (SD-L/T, Clontech) for the initial selection. After 2–3 days, several transformants were picked with a sterile toothpick and inoculated in minimal SD medium lacking leucine, tryptophan, histidine and adenine (SD-L/T/H/A, Clontech) to select clones containing protein interactions. Negative controls included single or dual transformants run in the same assay. Protein–protein interactions were quantified using the yeast b-Galactosidase assay kit (Pierce, Rockford, IL) following the manufacturer’s procedures.

Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft (DFG) within the Collaborative Research Center © 2010 Blackwell Publishing Ltd, Molecular Microbiology, 78, 932–946


SFB 766. We would like to thank Hans-Georg Koch (Freiburg) for providing YidC antibodies, Thomas Arnold (MPI Tübingen) for providing the colicin expression vector and Tanja Griesinger for perfect technical assistance.

References Alvarez, B.H., Gruber, M., Ursinus, A., Dunin-Horkawicz, S., Lupas, A.N., and Zeth, K. (2010) A transition from strong right-handed to canonical left-handed supercoiling in a conserved coiled-coil segment of trimeric autotransporter adhesins. J Struct Biol 170: 236–245. Arnold, T., Zeth, K., and Linke, D. (2009) Structure and function of colicin S4, a colicin with a duplicated receptorbinding domain. J Biol Chem 284: 6403–6413. Arnold, T., Zeth, K., and Linke, D. (2010) Omp85 from the thermophilic cyanobacterium Thermosynechococcus elongatus differs from proteobacterial Omp85 in structure and domain composition. J Biol Chem 285: 18003–18015. Bailey, T.L., and Elkan, C. (1994) Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc Int Conf Intell Syst Mol Biol 2: 28–36. Bolin, I., Norlander, L., and Wolf-Watz, H. (1982) Temperature-inducible outer membrane protein of Yersinia pseudotuberculosis and Yersinia enterocolitica is associated with the virulence plasmid. Infect Immun 37: 506– 512. Bos, M.P., Robert, V., and Tommassen, J. (2007) Functioning of outer membrane protein assembly factor Omp85 requires a single POTRA domain. EMBO Rep 8: 1149–1154. Bottone, E.J. (1997) Yersinia enterocolitica: the charisma continues. Clin Microbiol Rev 10: 257–276. Cotter, S.E., Surana, N.K., and St Geme, J.W., 3rd (2005) Trimeric autotransporters: a distinct subfamily of autotransporter proteins. Trends Microbiol 13: 199–205. Datsenko, K.A., and Wanner, B.L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 97: 6640–6645. Dautin, N., and Bernstein, H.D. (2007) Protein secretion in gram-negative bacteria via the autotransporter pathway. Annu Rev Microbiol 61: 89–112. Desvaux, M., Parham, N.J., and Henderson, I.R. (2004) The autotransporter secretion system. Res Microbiol 155: 53–60. Gentle, I.E., Burri, L., and Lithgow, T. (2005) Molecular architecture and function of the Omp85 family of proteins. Mol Microbiol 58: 1216–1225. Grosskinsky, U., Schütz, M., Fritz, M., Schmid, Y., Lamparter, M.C., Szczesny, P., et al. (2007) A conserved glycine residue of trimeric autotransporter domains plays a key role in Yersinia adhesin A autotransport. J Bacteriol 189: 9011–9019. Gust, B., Challis, G.L., Fowler, K., Kieser, T., and Chater, K.F. (2003) PCR-targeted Streptomyces gene replacement identifies a protein domain needed for biosynthesis of the sesquiterpene soil odor geosmin. Proc Natl Acad Sci USA 100: 1541–1546. Guzman, L.M., Belin, D., Carson, M.J., and Beckwith, J. (1995) Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol 177: 4121–4130.

Henderson, I.R., Navarro-Garcia, F., and Nataro, J.P. (1998) The great escape: structure and function of the autotransporter proteins. Trends Microbiol 6: 370–378. Henderson, I.R., Navarro-Garcia, F., Desvaux, M., Fernandez, R.C., and Ala’Aldeen, D. (2004) Type V protein secretion pathway: the autotransporter story. Microbiol Mol Biol Rev 68: 692–744. Ieva, R., and Bernstein, H.D. (2009) Interaction of an autotransporter passenger domain with BamA during its translocation across the bacterial outer membrane. Proc Natl Acad Sci USA 106: 19120–19125. Ize, B., Gerard, F., Zhang, M., Chanal, A., Voulhoux, R., Palmer, T., et al. (2002) In vivo dissection of the Tat translocation pathway in Escherichia coli. J Mol Biol 317: 327– 335. Jain, S., and Goldberg, M.B. (2007) Requirement for YaeT in the outer membrane assembly of autotransporter proteins. J Bacteriol 189: 5393–5398. Kim, S., Malinverni, J.C., Sliz, P., Silhavy, T.J., Harrison, S.C., and Kahne, D. (2007) Structure and function of an essential component of the outer membrane protein assembly machine. Science 317: 961–964. Knowles, T.J., Jeeves, M., Bobat, S., Dancea, F., McClelland, D., Palmer, T., et al. (2008) Fold and function of polypeptide transport-associated domains responsible for delivering unfolded proteins to membranes. Mol Microbiol 68: 1216–1227. Knowles, T.J., Scott-Tucker, A., Overduin, M., and Henderson, I.R. (2009) Membrane protein architects: the role of the BAM complex in outer membrane protein assembly. Nat Rev Microbiol 7: 206–214. Koretke, K.K., Szczesny, P., Gruber, M., and Lupas, A.N. (2006) Model structure of the prototypical non-fimbrial adhesin YadA of Yersinia enterocolitica. J Struct Biol 155: 154–161. Kostakioti, M., Newman, C.L., Thanassi, D.G., and Stathopoulos, C. (2005) Mechanisms of protein export across the bacterial outer membrane. J Bacteriol 187: 4306–4314. Linke, D., Riess, T., Autenrieth, I.B., Lupas, A., and Kempf, V.A. (2006) Trimeric autotransporter adhesins: variable structure, common function. Trends Microbiol 14: 264–270. Nummelin, H., Merckel, M.C., Leo, J.C., Lankinen, H., Skurnik, M., and Goldman, A. (2004) The Yersinia adhesin YadA collagen-binding domain structure is a novel lefthanded parallel beta-roll. EMBO J 23: 701–711. Oomen, C.J., van Ulsen, P., van Gelder, P., Feijen, M., Tommassen, J., and Gros, P. (2004) Structure of the translocator domain of a bacterial autotransporter. EMBO J 23: 1257–1266. Pohlner, J., Halter, R., and Meyer, T.F. (1987) Neisseria gonorrhoeae IgA protease. Secretion and implications for pathogenesis. Antonie Van Leeuwenhoek 53: 479–484. Remmert, M., Linke, D., Lupas, A.N., and Soding, J. (2009) HHomp – prediction and classification of outer membrane proteins. Nucleic Acids Res 37: W446–W451. Remmert, M., Biegert, A., Linke, D., Lupas, A.N., and Soding, J. (2010) Evolution of outer membrane {beta}-barrels from an ancestral {beta}{beta} hairpin. Mol Biol Evol 7: 1348– 1358. Robert, V., Volokhina, E.B., Senf, F., Bos, M.P., Van Gelder, P., and Tommassen, J. (2006) Assembly factor Omp85



recognizes its outer membrane protein substrates by a species-specific C-terminal motif. PLoS Biol 4: e377. Roggenkamp, A., Ackermann, N., Jacobi, C.A., Trülzsch, K., Hoffmann, H., and Heesemann, J. (2003) Molecular analysis of transport and oligomerization of the Yersinia enterocolitica adhesin YadA. J Bacteriol 185: 3735–3744. Ruiz-Perez, F., Henderson, I.R., Leyton, D.L., Rossiter, A.E., Zhang, Y., and Nataro, J.P. (2009) Roles of periplasmic chaperone proteins in the biogenesis of serine protease autotransporters of Enterobacteriaceae. J Bacteriol 191: 6571–6583. Sanchez-Pulido, L., Devos, D., Genevrois, S., Vicente, M., and Valencia, A. (2003) POTRA: a conserved domain in the FtsQ family and a class of beta-barrel outer membrane proteins. Trends Biochem Sci 28: 523–526. Sauri, A., Soprova, Z., Wickstrom, D., de Gier, J.W., Van der Schors, R.C., Smit, A.B., et al. (2009) The Bam (Omp85) complex is involved in secretion of the autotransporter haemoglobin protease. Microbiology 155: 3982–3991. Schütz, M., Weiss, E.M., Schindler, M., Hallstrom, T., Zipfel, P.F., Linke, D., and Autenrieth, I.B. (2010) Trimer stability of YadA is critical for virulence of Yersinia enterocolitica. Infect Immun 78: 2677–2690. Simons, R.W., Houman, F., and Kleckner, N. (1987) Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene 53: 85–96. Skurnik, M., Bolin, I., Heikkinen, H., Piha, S., and Wolf-Watz, H. (1984) Virulence plasmid-associated autoagglutination in Yersinia spp. J Bacteriol 158: 1033–1036. Struyve, M., Moons, M., and Tommassen, J. (1991) Carboxyterminal phenylalanine is essential for the correct assembly of a bacterial outer membrane protein. J Mol Biol 218: 141–148.

el Tahir, Y., and Skurnik, M. (2001) YadA, the multifaceted Yersinia adhesin. Int J Med Microbiol 291: 209–218. Voulhoux, R., and Tommassen, J. (2004) Omp85, an evolutionarily conserved bacterial protein involved in outermembrane–protein assembly. Res Microbiol 155: 129– 135. Voulhoux, R., Bos, M.P., Geurtsen, J., Mols, M., and Tommassen, J. (2003) Role of a highly conserved bacterial protein in outer membrane protein assembly. Science 299: 262–265. Werner, J., and 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. Wollmann, P., Zeth, K., Lupas, A.N., and Linke, D. (2006) Purification of the YadA membrane anchor for secondary structure analysis and crystallization. Int J Biol Macromol 39: 3–9. Wu, T., Malinverni, J., Ruiz, N., Kim, S., Silhavy, T.J., and Kahne, D. (2005) Identification of a multicomponent complex required for outer membrane biogenesis in Escherichia coli. Cell 121: 235–245.

Supporting information Additional supporting information may be found in the online version of this article. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
