SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Human pathogenic Bartonella henselae cause cat scratch disease and vasculoproliferative disorders. An important pathogenicity factor of B. henselae is the trimeric autotransporter adhesin Bartonella adhesin A (BadA) which is modularly constructed and consists of a head, a long and repetitive neck-stalk module with 22 repetitive neck/stalk repeats and a membrane anchor. The BadA head is crucial for bacterial adherence to host cells, binding to several extracellular matrix proteins and for the induction of vascular endothelial growth factor (VEGF) secretion. Here, we analysed the biological role of the BadA stalk in the infection process in greater detail. For this purpose, BadA head-bearing and headless deletion mutants with different lengths (containing one or four neck/stalk repeats in the neck-stalk module) were produced and functionally analysed for their ability to bind to fibronectin, collagen and endothelial cells and to induce VEGF secretion. Whereas a head-bearing short version (one neck/stalk element) of BadA lacks exclusively fibronectin binding, a substantially truncated headless BadA mutant was deficient for all of these biological functions. The expression of a longer headless BadA mutant (four neck/stalk repeats) restored fibronectin and collagen binding, adherence to host cells and the induction of VEGF secretion. Our data suggest that (i) the stalk of BadA is exclusively responsible for fibronectin binding and that (ii) both the head and stalk of BadA mediate adherence to collagen and host cells and the induction of VEGF secretion. This indicates overlapping functions of the BadA head and stalk.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bartonella henselae is a Gram-negative, facultative intracellular and slow-growing bacterium. In immunocompetent patients, B. henselae can cause cat scratch disease, which is characterized by a self-limiting lymphadenopathy. Infection of immunosuppressed patients can result in tumorous proliferations of endothelial cells (ECs) in the skin and internal organs, denoted as bacillary angiomatosis and peliosis hepatis (Relman et al., 1990; Anderson and Neuman, 1997). It has been shown that the ability to induce vasculoproliferations upon a B. henselae infection correlates in vitro and in vivo with the activation of hypoxia inducible factor (HIF)-1, the key transcription factor involved in angiogenesis (Pugh and Ratcliffe, 2003) and results in the subsequent secretion of vasculoproliferative cytokines [e.g. vascular endothelial growth factor (VEGF)] (Kempf et al., 2001; 2005).

One of the best characterized pathogenicity factors of B. henselae is Bartonella adhesin A (BadA), an adhesin of enormous size (∼ 240 nm in length, monomer: 328 kDa, trimer: ∼ 1000 kDa) (Riess et al., 2004). BadA is crucial for the adherence of B. henselae to ECs and to extracellular matrix (ECM)-components like collagen, laminin or fibronectin (Fn) under static and dynamic conditions (Müller et al., 2011). BadA expression also correlates with the activation of HIF-1 and the subsequent secretion of proangiogenic cytokines (Kempf et al., 2001; 2005; Riess et al., 2004). BadA homologues are also present in other Bartonella species and include the variably expressed outer-membrane proteins (Vomps) of B. quintana (Zhang et al., 2004) or Bartonella repeat protein A (BrpA) of B. vinsonii (Gilmore et al., 2005). Likethe prototypic Yersinia adhesin A (YadA) of Yersinia enterocolitica, BadA belongs to the class of trimeric autotransporter adhesins (TAAs) which represent important virulence factors of Gram-negative pathogens [e.g. NadA of Neisseria meningitidis (Comanducci et al., 2002), Hia of Haemophilus influenzae (St. Geme and Cutter, 2000), Burkholderia pseudomallei adhesion protein A (BpaA) (Edwards et al., 2010), ubiquitous surface proteins A1 and A2 (UspA1, UspA2) of Moraxella catarrhalis (Hoiczyk et al., 2000) or XadA from the plant pathogen Xanthomonas oryzae pv. oryzae (Ray et al., 2002)].

All TAAs share a similar modular organization consisting of head, neck, stalk and membrane anchor elements leading to a ‘lollipop’-like surface structure (Hoiczyk et al., 2000; Linke et al., 2006). The C-terminal membrane anchor is conserved in all TAAs, contains the autotransport activity and defines the family (Meng et al., 2006; Wollmann et al., 2006). The N-terminal signal sequence of BadA is followed by the head which is composed of three domains: a N-terminal YadA-like head repeat, a Trp ring and a GIN domain. The Trp ring and GIN domain of BadA showed high structural similarities to domains from the H. influenzae adhesin Hia (Szczesny et al., 2008). The BadA head is linked to the anchor by a long and highly repetitive neck-stalk module rich in coiled coils varying significantly in length in different B. henselae isolates (Riess et al., 2007). The modular structure of BadA leads to the hypothesis of a domain–function relationship in which certain domains are responsible for different biological functions. This has already been proposed for the YadA head which has a prominent role in the infection process of Y. enterocolitica (Roggenkamp et al., 2003). Earlier experiments revealed that a truncated BadA mutant lacking 21 of the neck stalk repeats had no Fn-binding capacity whereas other biological functions (host cell adhesion, induction of VEGF secretion) were not influenced (Kaiser et al., 2008). This leads to the assumption that the head is the most biologically active part of BadA. However, the question whether the stalk itself is involved in Fn binding or just acts as a spacer needed to bridge the gap between the bacterial and the host cell surfaces has not yet been answered.

Here, we analysed the function of the BadA stalk in B. hensleae by performing in-frame deletion mutagenesis of badA, leading to the expression of a range of truncated BadA variants with different lengths. Like wild-type bacteria, the headless B. henselae badA-/pF12, expressing only a short part of the stalk domain on the surface, mediated adhesion to collagen and host cells and induced VEGF secretion upon host cell infection. Moreover, B. henselae badA-/pF12 and B. henselae badA-/pHN2F12, a F12 version of BadA containing the BadA head, bound to Fn revealing that the stalk domain of BadA is responsible for Fn binding. Summarized, our results demonstrate that the BadA stalk harbours the Fn-binding capacity of BadA and shares some properties originally assigned to the BadA head of B. henselae.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Cloning and expression of BadA mutants in B. henselae

To determine the role of the BadA stalk in the infection process in greater detail, in-frame deletion mutagenesis of badA was performed by overlap extension PCR, subsequent cloning into the pBBR1MCS-5 and expression of BadA mutants from this plasmid in the BadA-negative transposon mutant B. henselae badA- (BadA F12, BadA N23). The generation of BadA variant HN2F12 by in-frame deletion turned out to be not feasible for unclear reasons. Therefore, this construct was synthesized according to the original badA sequence of B. henselae Marseille (Genebank accession no. DQ665674.1) and cloned into the pBBR1MCS-5 by GenScript (Piscataway, NJ, USA) resulting in the strain B. henselae Marseille badA-/pHN2F12. These efforts resulted in truncated BadA (BadA HN2F12) which lacks 18 of the 22 neck-stalk repeats (Fig. 1) as well as the headless variants of BadA HN2F12 and BadA HN23 [lacking 21 of the 22 neck-stalk repeats (Kaiser et al., 2008)] designated as BadA F12 and BadA N23 respectively. Details of cloning procedures are given in Experimental procedures and in Tables 1 and 2.

image

Figure 1. A. Schematic diagram and domain organization of the truncated BadA stalk mutants. BadA of B. henselae Marseille wild type consists out of the head, the stalk including the repetitive neck-stalk repeats and the membrane anchor. Truncated BadA HN23 has been constructed by deleting the indicated 21 neck-stalk repeats (aa 470–2850; Kaiser et al., 2008), whereas the longer BadA HN2F12 was constructed by deleting 18 neck-stalk repeats (aa 489–2478). BadA N23 and BadA F12 represent variants of BadA HN23 and BadA HN2F12 each lacking the head. Domains were annotated using the daTAA tool (Szczesny and Lupas, 2008) and drawn to scale (according to sequence length). B. Representative structures for domains found in BadA. The structures are the YadA head (PDB ID: 1PH9) of Y. enterocolitica (Nummelin et al., 2004), the BpaA head (3LAA) of B. pseudomallei (Edwards et al., 2010), the TRP ring and GIN domain of the BadA head (3DX9) of B. henselae (Szczesny et al., 2008), the BpaA FGG domain (3LAA) of B. pseudomallei (Edwards et al., 2010) and the Hia translocation domain (3EMO) of H. influenzae (Meng et al., 2006). The structures are presented in cartoon rendering showing secondary structural elements, and the colouring of the various structural motifs matches the colouring of the schematic diagram in A. The depicted structural motifs of 3LAA occur in the inverse order in the original structure to that shown in the figure.

Download figure to PowerPoint

Table 1.  Bacterial strains used in this study.
StrainCharacteristicsReference or source
B. henselae  
 Wild typeB. henselae Marseille, patient isolate, early passage, BadA-positiveDrancourt et al. (1996), Riess et al. (2004)
 badA-B. henselae Marseille, EZ::TN<KAN2>-transposon mutant, BadA-negative, KmrRiess et al. (2004)
 badA-/pHN23B. henselae badA- containing pHN23, Kmr, GmrKaiser et al. (2008)
 badA-/pHN2F12B. henselae badA- containing pHN2F12, Kmr, GmrThis study
 badA-/pN23B. henselae badA- containing pN23, Kmr, GmrThis study
 badA-/pF12B. henselae badA- containing pF12, Kmr, GmrThis study
E. coli  
 TOP 10Host strain used for cloningInvitrogen (Darmstadt, Germany)
 DH5αHost strain used for cloningInvitrogen
Table 2.  Plasmids and primers used in this study.
PlasmidsCharacteristicsReference or source
pBBR1MCS-5Broad host range vector, GmrKovach et al. (1995)
pCR-Blunt II TOPOBlunt end topoisomerase vector, KmrInvitrogen
pN23pBBR1MCS-5 containing the 1.9 kb badA fragment (BadA N23) from pPK8, GmrThis study
pHN23pBBR1MCS-5 containing the 3.1 kb badA fragment (BadA HN23) from pPK3, GmrKaiser et al. (2008)
pHN2F12pBBR1MCS-5 containing the 4.3 kb badA fragment (BadA HN2F12, carrying an in-frame deletion of aa 489–2478), GmrThis study
pPK5pCR-Blunt II TOPO containing a 3.0 kb badA fragment carrying an in-frame deletion of aa 53–2481 (BadA F12), KmrThis study
pPK8pCR-Blunt II TOPO containing a 1.9 kb badA fragment carrying an in-frame deletion of aa 53–2850 (BadA HN23), KmrThis study
pF12pBBR1MCS-5 containing the 3.0 kb badA fragment (BadA F12) from pPK5, GmrThis study
pTR14pBluescript II KS containing B. henselae Marseille badARiess et al. (2004)
PrimersSequenceReference or source
badA f10CTGAATTTAGAGAGTGTAAGCKaiser et al. (2008)
badA r15TTTTTCGTAGAAACAAGAGACCKaiser et al. (2008)
badA r24GTATGAATCTGTCCCGCAAGATTCGAAGCCAATThis study
badA f28TTCGAATCTTGCGGGACAGATTCATACAATCGGThis study
badA r28xTCTACATGCGTAAGCGCAAGATTCGAAGCCAATThis study
badA f32xCTTCGAATCTTGCGCTTACGCATGTAGAGAATGGThis study

To prove the expression and surface localization of the BadA mutants, transmission electron microscopy was performed (Fig. 2), revealing that B. henselae wild type expressed a hairy-appearing dense layer of the long BadA (total length: ∼ 240 nm) on the bacterial surface. Here, electron-dense structures on the tip of the long stalk were clearly visible most probably representing the BadA head (Kaiser et al., 2008) whereas B. henselae badA- showed no BadA expression. B. henselae badA-/pHN23 expressed the strongly truncated BadA HN23 located on the outer membrane (total length: ∼ 24 nm) as described earlier (Kaiser et al., 2008). Similarly to B. henselae wild type, B. henselae badA-/pHN2F12 showed a hair-like structure spanning from the outer membrane to an electron-dense globular structure, although the length of the fibres were significantly reduced (total length: ∼ 40 nm). B. henselae badA-/pF12 mutant, a headless variant of HN2F12, expressed structures of similar length but missing the electron dense head on the tip of the adhesin. Expression of BadA N23 construct lacking almost the whole stalk and, moreover, the head domain, correlated with the expression of a very short layer of hair-like fibres without the electron dense head (total length: ∼ 9 nm). Summarized, these data demonstrate that all of the herein described BadA constructs (BadA HN2F12, BadA N23 and BadA F12) were expressed on the surface of the complemented BadA-deficient B. henselae transposon mutant, demonstrating the proper autotransporter function of the BadA constructs as an indispensable prerequisite for further functional experiments.

image

Figure 2. Surface expression of BadA stalk mutants of B. henselae analysed by transmission electronmicroscopy. On B. henselae wild type, the typical dense surface layer of long ‘hairy’ BadA (∼ 240 nm) is visible which is missing on the surface of B. henselae badA-. The different length of the truncated BadA variants can be detected by the dense layers of the ∼ 24-nm-long BadA HN23 and the ∼ 40-nm-long BadA HN2F12 each expressed on the surface of B. henselae. Note the electron-dense structures at the tips of the adhesins representing the heads. The corresponding mutants of HN23 and HN2F12 lacking the head (N23 and F12) are also expressed on the surface of B. henselae badA- as a dense layer (length of N23: ∼ 9 nm, length of F12: ∼ 40 nm). A 2.4-fold magnification of the adhesins is given in the boxes. Scale bar: 0.3 µm.

Download figure to PowerPoint

Adherence of B. henselae BadA stalk mutants to collagen and fibronectin

BadA expression is crucial for the interaction of B. henselae with ECM compounds like collagen or Fn (Riess et al., 2004; Müller et al., 2011). To investigate the role of the BadA stalk in collagen binding, B. henselae strains were exposed to collagen-G-coated coverslips and bacterial adherence was analysed by 4′,6-diamidin-2′-phenylindoldihydrochlorid (DAPI) staining and subsequent confocal laser scanning microscopy (CLSM). Results revealed that the BadA-expressing strains B. henselae wild type, B. henselae badA-/pHN23, B. henselae badA-/pHN2F12 and, surprisingly, B. henselae badA-/pF12 showed a significantly higher adherence to collagen-coated coverslips than B. henselae badA- (see Fig. 3A). When using a quantitative software analysis tool (ImageJ), B. henselae wild type showed a ∼ 22.5-fold, B. henselae badA-/pHN23 a ∼ 12.2-fold, B. henselae badA-/pHN2F12 a ∼ 15.3-fold and B. henselae badA-/pF12 a ∼ 19.8-fold increased adherence ratio compared to B. henselae badA-. The collagen binding ability of B. henselae badA-/pN23 was comparatively low and similar to that of B. henselae badA- (∼ 2.5-fold). Moreover, all bacterial mutant strains (except B. henselae badA- and B. henselae badA-/N23) showed autoagglutination, a typical feature of many TAA-expressing bacteria.

image

Figure 3. Binding of B. henselae expressing truncated BadA stalk mutants to collagen and fibronectin. A. Collagen-G-coated coverslips were incubated with 1.0 × 107 bacteria, washed, stained with DAPI and analysed by CLSM. Note the collagen binding of BadA HN23-, HN2F12- and F12-expressing mutants which is not detectable in B. henselae badA- and B. henselae badA-/pN23. Scale bar: 20 µm. B. Collagen-G-bound bacteria (30 independent microscopic fields) were quantified by counting the pixels of fluorescent bacteria using the ImageJ software (relative fluorescence). *Significant difference between B. henselae wild type and B. henselae badA- (P ≤ 0.01); **significant difference between B. henselae badA- and B. henselae badA-/pHN23, B. henselae badA-/pHN2F12 and B. henselae badA-/pF12 (P ≤ 0.01). C. Bacteria grown on fibronectin-containing CBA were harvested and BadA-bound fibronectin (240 kDa) was detected by immunoblotting using specific anti-fibronectin antibodies. Note that B. henselae wild type, B. henselae badA- pHN2F12 and the headless B. henselae badA-/pF12 but not the shorter BadA mutants B. henselae badA-/pHN23 and B. henselae badA-/pN23 bind fibronectin.

Download figure to PowerPoint

To further analyse the Fn binding of the BadA stalk mutants, B. henselae were grown on Fn containing Columbia blood agar (CBA), harvested and bacteria-bound Fn was detected by immunoblotting using whole bacterial lysates and anti-Fn-antibodies. Experiments confirmed the Fn binding of B. henselae wild type which was missing in B. henselae badA- and B. henselae badA-/pHN23, as previously described (Riess et al., 2004; Kaiser et al., 2008). These results served as robust internal controls. Interestingly, mutants with a longer stalk (B. henselae wild type, B. henselae badA-/pF12 and B. henselae badA-/pHN2F12) bound Fn (see Fig. 3B). In summary, experiments elucidating the interaction of BadA with components of the ECM suggest that the stalk but not the head of BadA is crucially involved in Fn binding of B. henselae indicating that the BadA stalk itself mediates Fn binding whereas collagen binding can be mapped to both the stalk and head of BadA.

Adherence to ECs of B. henselae BadA stalk mutants

Expression of BadA is essential for the adherence of B. henselae to host cells (Riess et al., 2004). To assess the ability of the BadA stalk in the binding of B. henselae to host cells, human umbilical venous ECs (HUVECs) were infected with the B. henselae BadA deletion mutants. Adherent bacteria were stained 30 min after infection by B. henselae specific antibodies and analysed by CLSM. Data revealed that B. henselae wild type, B. henselae badA-/pHN23, B. henselae badA-/pHN2F12 and B. henselae badA-/pF12 adhered by a significantly higher amount to ECs than B. henselae badA- (see Fig. 4). Compared to B. henselae badA-, B. henselae wild type showed a ∼ 7.0-fold higher adherence. The adherence rate of B. henselae badA-/pHN23 was ∼ 5.0-fold that of B. henselae badA-, and adherence rates of B. henselae badA-/pHN2F12 and of B. henselae badA-/pF12 were ∼ 8.6-fold and ∼ 7.0-fold higher when compared with B. henselae badA-. In contrast, B. henselae badA-/pN23 was strongly diminished in its binding capacity to ECs, comparable to B. henselae badA- (∼ 1.0-fold). These data suggest that both head and stalk act as structurally independent but biologically synergestic binding partners of B. henselae to ECs.

image

Figure 4. Endothelial cell adherence of B. henselae expressing truncated BadA stalk mutants. A. HUVECs were infected with B. henselae wild type, B. henselae badA- and B. henselae badA- expressing BadA HN23, BadA HN2F12, BadA F12 or BadA N23. Adherent bacteria were labelled by Cy2-conjugated mouse anti-B. henselae antibodies (green signal); filamentous actin was stained with TRITC-labelled phalloidin (red signal). Note the endothelial cell adherence of B. henselae wild type, B. henselae badA-/pHN23, B. henselae badA-/pHN2F12 and B. henselae badA-/pF12 which is not detectable for B. henselae badA- and B. henselae badA-/pN23. Scale bar: 47 µm. B. Bacterial adherence was quantified by counting adherent bacteria per cell on 100 individual endothelial cells. *Significant difference between B. henselae wild type and B. henselae badA- (P ≤ 0.01); **significant difference between B. henselae badA- and B. henselae badA-/pHN23, B. henselae badA-/pHN2F12 and B. henselae badA-/pF12 (P ≤ 0.01).

Download figure to PowerPoint

Induction of VEGF secretion by B. henselae BadA stalk mutants

The B. henselae-induced proangiogenic host cell response is due to a BadA-dependent activation of HIF-1 resulting in the secretion of proangiogenic cytokines like VEGF (Riess et al., 2004; Kempf et al., 2005). To verify whether the BadA stalk is involved in the secretion of VEGF, HeLa-229 cells were infected with B. henselae stalk mutants and VEGF levels were quantified from cell culture supernatants 24 h after infection by elisa. B. henselae wild type and B. henselae badA-/pHN23 but not B. henselae badA- triggered the secretion of VEGF, as described earlier, and these strains served as robust internal controls (see Fig. 5). B. henselae badA-/pHN2F12 and B. henselae badA-/pF12 but not B. henselae badA-/pN23 triggered the secretion of VEGF, indicating that both the BadA head and the stalk are involved in the angiogenic reprogramming of host cells.

image

Figure 5. Secretion of the proangiogenic cytokine VEGF after infection with B. henselae expressing BadA stalk mutants. VEGF secretion upon infection of HeLa-229 cells was quantified by analysing supernatants 24 h after infection. *Significant difference between B. henselae wild type and B. henselae badA- infected cells (P < 0.01); **Significant difference between B. henselae badA- and B. henselae badA-/pHN23, B. henselae badA-/pHN2F12 and B. henselae badA-/pF12 infected cells (P < 0.01). Negative control: non-infected cells; positive control: desferrioxamin (DFO; 200 µM).

Download figure to PowerPoint

Collagen-binding capability of B. henselae does not correlate with BadA stalk length under dynamic flow conditions

In vivo, interaction of bacteria with ECM components or host cells often occur under dynamic conditions (e.g. in the bloodstream) and bacterial binding is influenced by continuous shear stress (Mairey et al., 2006; Mikaty et al., 2009). To determine whether the length of the stalk (in the presence of the head domain) influences the ability of B. henselae to bind to collagen under dynamic conditions, we used a recently established flow chamber system (Müller et al., 2011) and compared the degree of binding of B. henselae wild type (total length: ∼ 240 nm), B. henselae badA-/pHN2F12 (total length: ∼ 40 nm), B. henselae badA-/pHN23 (total length: ∼ 24 nm) with that of B. henselae badA- (negative control). Bacterial suspensions containing a defined amount of the particular B. henselae strains were pumped under constant flow conditions (shear stress: 1.0 dyne cm−2) through collagen-G-coated flow chambers (see Fig. 6A). After 15 min, adherent bacteria were stained with DAPI and visualized by CLSM. In these assays, the B. henselae BadA stalk mutants demonstrated a significantly higher adherence to collagen than B. henselae badA- (see Fig. 6B). Remarkably, differences in the amount of bound B. henselae BadA stalk mutants of different lengths did not appear. This observation suggests that the length of the stalk does not play a crucial role in the collagen binding of BadA under shear-stress conditions and that the binding capability of the head on its own is strong enough to mediate collagen adherence in the dynamic flow model.

image

Figure 6. Stalk length-dependent adherence of B. henselae expressing BadA stalk mutants to collagen under dynamic flow conditions. A. Schematic picture of the experimental setting. Collagen-G-coated multichannel slides were exposed to 1.0 × 108 bacteria per millilitre under constant flow conditions for 15 min (shear stress 1 dyne cm−2). B. Adherent bacteria were stained with DAPI and analysed by CLSM. Scale bar: 47 µm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

BadA of B. henselae belongs to the class of TAAs which represent important pathogenicity factors in many Gram-negative bacteria (Linke et al., 2006). All TAAs share a common and characteristic modular organization consisting of head, neck-stalk and membrane anchor elements. In the genus Bartonella, several BadA homologues are present in both human and animal pathogenic species (e.g. Vomp A-D from B. quintana or BrpA from B. vinsonii). TAAs of Bartonella spp. are characterized by highly repetitive neck-stalk modules. These modules vary in their number between species (Kaiser et al., 2011) and even within the species B. henselae (Riess et al., 2007). In the case of B. henselae Marseille (used in this study), BadA consists of 22 neck-stalk repeats. Earlier experiments with the truncated B. henselae badA-/pHN23 strain suggested that a long and repetitive coiled-coil stalk is dispensable for autotransport through the pore built by the trimerized membrane anchor, for adherence of B. henselae to collagen and HUVECs and the induction of an angiogenic reprogramming. These results indicated that the head is the biologically most active part of BadA (Kaiser et al., 2008). Remarkably, B. henselae expressing BadA HN23 did not bind Fn; however, the interaction of B. henselae with this component of the ECM seems of importance in the process of EC adherence (Riess et al., 2004).

To investigate whether the Fn-binding region of BadA is located in the neck-stalk element or whether binding is mediated exclusively by the head with a neck/stalk element working as a ‘spacer’ to maintain the optimum distance needed for correct binding, B. henselae BadA deletion mutants with neck/stalk elements of different length (BadA HN2F12, BadA HN23) and the corresponding ‘headless’ mutants (BadA F12, BadA N23) were compared in their biological functions. Our data revealed that the stalk itself clearly binds to Fn as this function was present in both B. henselae badA-/pHN2F12 and B. henselae badA-/pF12. Furthermore, similarly to B. henselae badA-/pHN23 and B. henselae badA-/pHN2F12, the stalk-expressing but headless deletion mutant B. henselae badA-/pF12 exhibited adherence to collagen and ECs. Moreover, B. henselae badA-/pF12 was also able to induce the secretion of the proangiogenic cytokine VEGF. These findings suggest that, beside the BadA head, the BadA stalk itself can assume an important and redundant role in the infection process of B. henselae. Finally, experiments performed under dynamic flow conditions revealed that the stalk length itself has no obvious influence on the binding affinity of the adhesion under shear stress, as both B. henselae badA-/pHN23 and B. henselae badA-/pHN2F12 showed a similar collagen-binding phenotype.

The stalks of TAAs are highly repetitive, fibrous structures with a high content in coiled coils. They display enormous differences in their length even between orthologues of the same species (Riess et al., 2007). Stalks are thought to act as a spacer between the bacterial cell surface and the head of the adhesin to bring the head close to its binding partner (Linke et al., 2006). Hence, the head has been proposed to be the primary binding domain of the adhesin. However, involvement of stalk elements in the biological binding process of TAAs has been demonstrated several times in the last few years. For instance, the UspAs of Moraxella catarrhalis have been demonstrated to bind Fn and laminin with the N-terminal part (including the head) (Tan et al., 2005; 2006; Conners et al., 2008). Another binding partner of UspA1 is the carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM 1) expressed on epithelial cells (Hill and Virji, 2003; Conners et al., 2008). The UspA1 binding site for CEACAM 1 is located in a central coiled-coil part of the stalk of UspA1. Stalk-dependent biological binding can also be attributed to the atypical TAA Haemophilus adhesin A (HadA) of H. influenzae which exhibits no typical N-terminal globular domain. HadA mediates adherence of H. influenzae to ECM compounds and host cells. Here, the N-terminal part of the stalk plays a crucial role in the binding activity (Serruto et al., 2009). Apart from the missing head, the architecture of HadA is thought to be very similar to the architecture of NadA (Serruto et al., 2009).

These examples demonstrate that there is no exclusive role for the head in the biological binding processes of TAAs. Obviously, the stalk can also play an important role in the biological function of TAAs beyond structural and static functions. We suggest that the stalk, at least in BadA, represents a supplementary element of the head and ensures a redundant biological function in terms of adhesion. Therefore, the function of the BadA stalk seems not only to bring the head into close contact to its binding partners, but also to harbour redundant binding sites for collagen and host cells and, moreover, by binding to Fn, to add an additional adhesive function not covered by the head. In this context, it is interesting to look in detail at the structure of the FGG domain that is repeated 14 times in the BadA stalk. In the TAA BpaA from Burkholderia pseudomallei, FGG domains are also present, and the structure of one has been solved (Edwards et al., 2010) (see Fig. 1). FGG domains are coiled coils, but they are interrupted by an insertion that forms an unpaired beta-strand (actually three, one per monomer). Such unpaired beta-strands have been implied as binding interfaces in various bacterial surface proteins, e.g. for OmpX of Escherichia coli (Vogt and Schulz, 1999). Fn, on the other hand, consists of domains full of unpaired beta-strands (Pankov and Yamada, 2002); this is especially true for the type III repeat, an Ig-like domain with two flattened beta-sheets that expose their edges. It is tempting to assume that pairing of beta-strands is the molecular basis for the interaction of the BadA stalk with Fn, which in the future could potentially be shown using co-crystallization strategies. Notably, other yet unresolved structures of the BadA stalk might also play a role in Fn binding.

Due to the enormous length of the BadA stalk, Fn-binding domains will reside far away from the BadA head. We speculate that the BadA head acts as a first binding initiator which is then assisted by the stalk during the adherence process. This could explain the enormous length of the stalk: a long neck/stalk module could confer more flexibility to the stalk resulting in stronger adherence to its binding partners as steric clashes are avoided. A further advantage of a long BadA stalk could be the generation of a larger binding area (containing specific binding domains or motifs). How the whole stalk or, respectively, particular neck-stalk repeats of B. henselae badA-/pHN2F12 contribute to the biological functions of BadA, still remains unclear. Furthermore, it has to be realized that bacteria expressing a very short stalk connected to a head (B. henselae badA-/pHN23) exhibit a binding behaviour (to collagen, ECs) similar to that of B. henselae expressing a longer ‘headless’ stalk (B. henselae badA-/pF12), strengthening again the suggestion that the neck/stalk module seems to play a supporting but important role in the biological function of BadA.

Interestingly, other biological functions of TAA stalks have been described which have not been analysed in B. henselae yet. In particular, it has been demonstrated that the binding of complement factor H and the resulting serum resistance of Y. enterocolitica is located in the stalk of YadA (Roggenkamp et al., 2003; Biedzka-Sarek et al., 2008). In YadA, the binding of factor H is not restricted to a specific binding site but is due to several conformational and discontinuous sites of the YadA stalk. Sharing the same coiled-coil structure as YadA, the BadA stalk raises the question of a role for BadA in the potential serum resistance of B. henselae. A further immunomodulatory function of TAAs has been shown for the E. coli Ig-binding protein EibD: here, the stalk binds IgG and IgA immunoglobulins via their respective Fc fragments at different sites in the stalk region (Leo et al., 2011).

It has been proposed earlier that the head is the most crucial part of BadA and is needed for most of the biological functions of BadA (Kaiser et al., 2008). Now, we can argue for an additional and specific role of the stalk in the biological functions mediated by BadA. Which exact binding motifs of the BadA head and stalk are the structural basis for, e.g. binding to ECM compounds, host cells and the induction of VEGF secretion, needs to be analysed further. Such experiments could be done by generating further domain and subdomain deletion mutants of BadA despite the expected technical challenges resulting from the enormous length of BadA and its highly repetitive nature.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and growth conditions

Bartonella henselae was grown in a humidified atmosphere at 37°C and 5% CO2 on CBA. E. coli was grown in Luria–Bertani (LB) broth at 37°C. Antibiotics were used at the following concentrations: kanamycin 30 µg ml−1, gentamicin 10 µg ml−1. Bacterial stock suspensions were produced by harvesting bacteria from agar plates after 5 days of culture and freezing at −80°C in LB broth containing 20% glycerol. Serial dilutions were performed from the frozen stocks and colony-forming units were counted to determine the number of viable bacteria per aliquot. Bacteria used in this study are summarized in Table 1.

Construction and cloning of badA mutants

In-frame deletion of badA elements was performed by overlap extension PCR. Here, the reverse primer of the first fragment contained a short 5′-overhang complementary to the starting sequence of the second fragment. The forward primer of the second fragment, respectively, contained a short 5′-overhang complementary to the ending sequence of the first fragment. An overview of primers and plasmids is given in Table 2.

To construct the f12 deletion mutant, a 873 bp fragment encoding the putative badA promoter and the first four amino acid residues (aa) of badA (aa 1–51 head region; primers: badA f10 and badA r25) and a 2121 bp fragment of badA encoding aa 2482–3082 (stalk region beginning with the last four aa of neck #20, including FGG domain #12 and membrane anchor; primers: badA f28 and badA r15) were amplified by PCR using Phusion polymerase (Finnzymes, Espoo, Finland). pTR14 was used as template DNA. To construct n23 (headless deletion variant of hn23), pPK3 was used as template DNA. A 873 bp fragment encoding the badA promoter and the first four aa of badA (aa 1–51 head region; primers: badA f10 and badA r28x) and a 1007 bp fragment of badA encoding aa 2851–3082 (stalk region beginning with neck #23 and membrane anchor; primers: badA f32x and badA r15) were amplified by PCR. Both fragments of the respective mutants were fused in a third PCR reaction by using the primers badA f10 and badA r15. The fusion constructs (badA F12: 2994bp; badA n23: 1880 bp) were ligated into pCR-Blunt II TOPO (resulting in the plasmids pPK5 and pPK8) and elektroporated into E. coli TOP 10. The proposed sequence of the construct was proven by DNA sequencing (GATC, Konstanz, Germany).

The plasmids pPK5 and pPK8 were digested with EcoRI, inserts were ligated into the broad host range vector pBBR1MCS-5 (resulting in plasmids pF12 and pN23) and amplified in E. coli DH5α. For expression of these constructs in B. henselae, pF12 and pN23 were electroporated in a BadA negative transposon mutant B. henselae badA- resulting in badA-/pF12 and badA-/pN23. PCR and DNA manipulations were performed according to standard protocols (Kaiser et al., 2008).

The synthesis and cloning of hn2f12 deletion mutant (consisting of the head region encoding aa 1–488 and the stalk region beginning with the FGG domain #12 encoding aa 2479–3082) into pBBR1MCS-5 via EcoRI resulting in pHN2F12 were carried out by GenScript (Piscataway, NJ, USA). Gene synthesis, instead of construction via overlap extension PCR and cloning, was performed as the above described protocols remained unsuccessful for this particular construct for unclear reasons.

Secondary structure analysis of TAA domains

The structural figures of TAA domains were prepared using the VMD software as described (Humphrey et al., 1996).

Transmission electron microscopy

For ultrastructural studies, bacterial colonies grown on CBA were fixed with 2.5% glutaraldehyde in PBS for 20 min at ambient temperature and kept for 20 h at 4°C. Fixed cells were covered with 2% agarose and blocks containing single colonies were cut out. After postfixation with 1% osmium tetroxide in 100 mM phosphate buffer pH 7.2 for 1 h on ice, these blocks were rinsed with aqua bidest, treated with 1% aqueous uranyl acetate for 1 h at 4°C, dehydrated through a graded series of ethanol and embedded in Epon. Sections were stained with 1% aqueous uranyl acetate and lead citrate and analysed in a Philips CM10 electron microscope at 60 kV using a 30 µm objective aperture as described earlier (Kaiser et al., 2008).

Determination of collagen binding of B. henselae

To analyse the binding of B. henselae to collagen under static conditions, bacteria were resuspended in PBS with a concentration of 1.0 × 107 bacteria per millilitre. Bacteria were allowed to adhere for 30 min onto collagen-G-coated coverslips (each coated with 1 ml of a 10 µg ml−1 collagen-G/PBS dilution for 24 h at room temperature) after centrifugation (5 min, 300 g). Subsequently, coverslips were washed extensively with PBS and bacteria were fixed with 3.75% paraformaldehyde (PFA). Adherent bacteria were stained with DAPI and adherence was detected microscopically (see below). Bacterial adherence was quantified by counting the pixels gained by the fluorescent bacteria from 30 microscopic independent fields from three different coverslips using the ImageJ software (for details see: http://rsbweb.nih.gov/ij/).

To analyse the collagen binding of B. henselae under dynamic flow conditions, multichannel slides (µ slide VI0.4 flow kit, Ibidi, Martinsried, Germany) were coated with collagen-G (10 µg ml−1, 24 h) and a bacterial suspension with a concentration of 1.0 × 108 bacteria per millilitre (in PBS) was pumped with a shear stress of τ = 1 dyne cm−2 for 15 min at room temperature through the channels for 15 min using syringes (Braun, Melsungen, Germany) and pumping perfusors (MTS, Schweinfurt, Germany). Multichannel slides were washed with PBS, and adherent bacteria were fixed with 3.75% PFA, stained with DAPI and analysed via CLSM (see below). All bacterial strains used in this study were tested in parallel in the same multichannel slide as described recently (Müller et al., 2011).

Detection of fibronectin-binding of B. henselae and BadA immunoblotting

Bartonella henselae-bound Fn (originating from CBA) was assessed by immunoblotting as described (Riess et al., 2004). Fn-containing CBA grown bacteria were harvested in PBS, adjusted to an OD600 of 1.0 and lysed in SDS sample buffer. The samples were separated by 8% SDS-PAGE. Membranes were incubated with monoclonal anti-Fn Abs (Becton Dickinson, Heidelberg, Germany) as described (Riess et al., 2004).

Culture and infection of HUVECs and HeLa-229 cells

Adherence experiments were preformed using HUVECs. Cells were cultured in EC growth medium (PromoCell, Mannheim, Germany) as described (Kempf et al., 2000). Briefly, 1.0 × 105 cells were seeded onto collagen-G-coated coverslips for 18 h. Infection experiments were performed with 1.0 × 107B. henselae (MOI: 100) in cell culture media without antibiotics and foetal calf serum (FCS, Sigma-Aldrich, Deisenhofen). After infection, bacteria were centrifuged onto cultured cells for 5 min at 300 g and non-adherent bacteria were removed after 30 min by extensively washing with pure EC growth medium. Bacterial adherence was determined by CLSM as described (Riess et al., 2004).

Induction of VEGF secretion by B. henselae was analysed using HeLa-229 cells. Cells were cultured in RPMI-1640 supplemented with 10% FCS as described (Kempf et al., 2000). For elisa experiments, 1.0 × 105 cells were seeded in 12-well plates and bacteria were centrifuged onto cultured cells for 5 min at 300 g. Infection experiments were performed with 5.0 × 107B. henselae (MOI: 500) in cell culture media without antibiotics and without FCS to avoid unspecific cytokine secretion. Uninfected cells were used for negative control, desferrioxamine (DFO, 200 µM, Sigma-Aldrich) was used as positive control.

Immunostaining and CLSM

For microscopic analysis of bacterial adhesion onto host cells, HUVECs were seeded onto collagen-G-coated coverslips (1.0 × 105 cells). After infection with B. henselae strains for 30 min, non-adherent bacteria were removed by washing with prewarmed cell culture medium (without supplements) and cells were fixed by adding 3.75% PBS-buffered PFA. Bacteria were stained with mouse polyclonal antibodies (Ab) raised against B. henselae (Kaiser et al., 2008) for 45 min followed by incubation with a carbocyanine-2 (Cy2)-conjugated secondary anti-IgG Ab (Dako, Hamburg, Germany) for 1 h. For actin staining, cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich) in PBS for 15 min, washed and incubated with tetramethyl rhodamine iso-thiocyanate (TRITC)-labelled phalloidin (Sigma-Aldrich) for 1 h. Fluorescence was detected by a Leica DM IRE 2 confocal laser scanning microscope (Leica, Wetzlar, Germany) using three different fluorochromes representing the green (FITC), red (TRITC) and blue (DAPI) channels. Images were digitally processed with Photoshop 11.0.2 (Adobe Systems, Mountain View, CA). Bacterial adherence to ECs was quantified by counting cell-adherent bacteria on 100 individual ECs per bacterial strain (from three different coverslips).

Quantification of VEGF

Concentrations of secreted VEGF were measured from HeLa-229 cell culture supernatants by elisa (R&D systems, Wiesbaden, Germany) according to the manufacturer's instructions.

Statistical analysis

All experiments were performed at least three times and yielded comparable results. Differences between mean values of experimental and control groups were analysed by the paired Student's t-test. A value of P < 0.01 was considered statistically significant.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The authors thank Linda Stolz (Frankfurt) and Andrea Schäfer (Tübingen) for expert technical support and Tanja Riess for continuous discussion. This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 766) to V.A.J.K. and D.L. and a FEMS Advanced Fellowship to J.C.L.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Anderson, B.E., and Neuman, M.A. (1997) Bartonella spp. as emerging human pathogens. Clin Microbiol Rev 10: 203219.
  • Biedzka-Sarek, M., Salmenlinna, S., Gruber, M., Lupas, A.N., Meri, S., and Skurnik, M. (2008) Functional mapping of YadA- and Ail-mediated binding of human factor H to Yersinia enterocolitica serotype O:3. Infect Immun 76: 50165027.
  • Comanducci, M., Bambini, S., Brunelli, B., Adu-Bobie, J., Arico, B., Capecchi, B., et al. (2002) NadA, a novel vaccine candidate of Neisseria meningitidis. J Exp Med 195: 14451454.
  • Conners, R., Hill, D.J., Borodina, E., Agnew, C., Daniell, S.J., Burton, N.M., et al. (2008) The Moraxella adhesin UspA1 binds to its human CEACAM1 receptor by a deformable trimeric coiled-coil. EMBO J 27: 17791789.
  • Drancourt, M., Birtles, R., Chaumentin, G., Vandenesch, F., Etienne, J., and Raoult, D. (1996) New serotype of Bartonella henselae in endocarditis and cat-scratch disease. Lancet 347: 441443.
  • Edwards, T.E., Phan, I., Abendroth, J., Dieterich, S.H., Masoudi, A., Guo, W., et al. (2010) Structure of a Burkholderia pseudomallei trimeric autotransporter adhesin head. PLoS ONE 5: e12803.
  • Gilmore, R.D., Jr, Bellville, T.M., Sviat, S.L., and Frace, M. (2005) The Bartonella vinsonii subsp. arupensis immunodominant surface antigen BrpA gene, encoding a 382-kilodalton protein composed of repetitive sequences, is a member of a multigene family conserved among Bartonella species. Infect Immun 73: 31283136.
  • Hill, D.J., and Virji, M. (2003) A novel cell-binding mechanism of Moraxella catarrhalis ubiquitous surface protein UspA: specific targeting of the N-domain of carcinoembryonic antigen-related cell adhesion molecules by UspA1. Mol Microbiol 48: 117129.
  • Hoiczyk, E., Roggenkamp, A., Reichenbecher, M., Lupas, A., and Heesemann, J. (2000) Structure and sequence analysis of Yersinia YadA and Moraxella UspAs reveal a novel class of adhesins. EMBO J 19: 59895999.
  • Humphrey, W., Dalke, A., and Schulten, K. (1996) VMD: visual molecular dynamics. J Mol Graph 14: 3338.
  • Kaiser, P.O., Riess, T., Wagner, C.L., Linke, D., Lupas, A.N., Schwarz, H., et al. (2008) The head of Bartonella adhesin A is crucial for host cell interaction of Bartonella henselae. Cell Microbiol 10: 22232234.
  • Kaiser, P.O., Riess, T., O'Rourke, F., Linke, D., and Kempf, V.A. (2011) Bartonella spp.: throwing light on uncommon human infections. Int J Med Microbiol 301: 715.
  • Kempf, V.A., Schaller, M., Behrendt, S., Volkmann, B., Aepfelbacher, M., Cakman, I., et al. (2000) Interaction of Bartonella henselae with endothelial cells results in rapid bacterial rRNA synthesis and replication. Cell Microbiol 2: 431441.
  • Kempf, V.A., Volkmann, B., Schaller, M., Sander, C.A., Alitalo, K., Riess, T., et al. (2001) Evidence of a leading role for VEGF in Bartonella henselae-induced endothelial cell proliferations. Cell Microbiol 3: 623632.
  • Kempf, V.A., Lebiedziejewski, M., Alitalo, K., Wälzlein, J.H., Ehehalt, U., Fiebig, J., et al. (2005) Activation of hypoxia-inducible factor-1 in bacillary angiomatosis: evidence for a role of hypoxia-inducible factor-1 in bacterial infections. Circulation 111: 10541062.
  • Kovach, M.E., Elzer, P.H., Hill, D.S., Robertson, G.T., Farris, M.A., Roop, R.M., et al. (1995) Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene 166: 175176.
  • Leo, J.C., Lyskowki, A., Hattula, K., Hartmann, M.D., Schwarz, H., Butcher, S.J., et al. (2011) The structure of E. coli IgG-binding protein D suggests a general model for bending and binding in trimeric autotransporter adhesins. Structure 19: 10211030.
  • Linke, D., Riess, T., Autenrieth, I.B., Lupas, A., and Kempf, V.A. (2006) Trimeric autotransporter adhesins: variable structure, common function. Trends Microbiol 14: 264270.
  • Mairey, E., Genovesio, A., Donnadieu, E., Bernard, C., Jaubert, F., Pinard, E., et al. (2006) Cerebral microcirculation shear stress levels determine Neisseria meningitidis attachment sites along the blood-brain barrier. J Exp Med 203: 19391950.
  • Meng, G., Surana, N.K., St Geme, J., 3rd, and Waksman, G. (2006) Structure of the outer membrane translocator domain of the Haemophilus influenzae Hia trimeric autotransporter. EMBO J 25: 22972304.
  • Mikaty, G., Soyer, M., Mairey, E., Henry, N., Dyer, D., Forest, K.T., et al. (2009) Extracellular bacterial pathogen induces host cell surface reorganization to resist shear stress. PLoS Pathog 5: e1000314.
  • Müller, N.F., Kaiser, P.O., Linke, D., Schwarz, H., Riess, T., Schäfer, A., et al. (2011) Trimeric autotransporter adhesin-dependent adherence of Bartonella henselae, Bartonella quintana and Yersinia enterocolitica to matrix components and endothelial cells under static and dynamic flow conditions. Infect Immun 79: 25442553.
  • 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 left-handed parallel beta-roll. EMBO J 23: 701711.
  • Pankov, R., and Yamada, K.M. (2002) Fibronectin at a glance. J Cell Sci 115: 38613863.
  • Pugh, C.W., and Ratcliffe, P.J. (2003) Regulation of angiogenesis by hypoxia: role of the HIF system. Nat Med 9: 677684.
  • Ray, S.K., Rajeshwari, R., Sharma, Y., and Sonti, R.V. (2002) A high-molecular-weight outer membrane protein of Xanthomonas oryzae pv. oryzae exhibits similarity to non-fimbrial adhesins of animal pathogenic bacteria and is required for optimum virulence. Mol Microbiol 46: 637647.
  • Relman, D.A., Loutit, J.S., Schmidt, T.M., Falkow, S., and Tompkins, L.S. (1990) The agent of bacillary angiomatosis. An approach to the identification of uncultured pathogens. N Engl J Med 323: 15731580.
  • Riess, T., Andersson, S.G., Lupas, A., Schaller, M., Schäfer, A., Kyme, P., et al. (2004) Bartonella adhesin A mediates a proangiogenic host cell response. J Exp Med 200: 12671278.
  • Riess, T., Raddatz, G., Linke, D., Schäfer, A., and Kempf, V.A. (2007) Analysis of Bartonella adhesin A expression reveals differences between various B. henselae strains. Infect Immun 75: 3543.
  • Roggenkamp, A., Ackermann, N., Jacobi, C.A., Truelzsch, K., Hoffmann, H., and Heesemann, J. (2003) Molecular analysis of transport and oligomerization of the Yersinia enterocolitica adhesin YadA. J Bacteriol 185: 37353744.
  • St. Geme, J., 3rd, and Cutter, D. (2000) The Haemophilus influenzae Hia adhesin is an autotransporter protein that remains uncleaved at the C terminus and fully cell associated. J Bacteriol 182: 60056013.
  • Serruto, D., Spadafina, T., Scarselli, M., Bambini, S., Comanducci, M., Hohle, S., et al. (2009) HadA is an atypical new multifunctional trimeric coiled-coil adhesin of Haemophilus influenzae biogroup aegyptius, which promotes entry into host cells. Cell Microbiol 11: 10441063.
  • Szczesny, P., and Lupas, A. (2008) Domain annotation of trimeric autotransporter adhesins – daTAA. Bioinformatics 24: 12511256.
  • Szczesny, P., Linke, D., Ursinus, A., Bär, K., Schwarz, H., Riess, T.M., et al. (2008) Structure of the head of the Bartonella adhesin BadA. PLoS Pathog 4: e1000119.
  • Tan, T.T., Nordstrom, T., Forsgren, A., and Riesbeck, K. (2005) The respiratory pathogen Moraxella catarrhalis adheres to epithelial cells by interacting with fibronectin through ubiquitous surface proteins A1 and A2. J Infect Dis 192: 10291038.
  • Tan, T.T., Forsgren, A., and Riesbeck, K. (2006) The respiratory pathogen Moraxella catarrhalis binds to laminin via ubiquitous surface proteins A1 and A2. J Infect Dis 194: 493497.
  • Vogt, J., and Schulz, G.E. (1999) The structure of the outer membrane protein OmpX from Escherichia coli reveals possible mechanisms of virulence. Structure 7: 13011309.
  • 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: 39.
  • Zhang, P., Chomel, B.B., Schau, M.K., Goo, J.S., Droz, S., Kelminson, K.L., et al. (2004) A family of variably expressed outer-membrane proteins (Vomp) mediates adhesion and autoaggregation in Bartonella quintana. Proc Natl Acad Sci USA 101: 1363013635.