Allosteric β-propeller signalling in TolB and its manipulation by translocating colicins
2009; Springer Nature; Volume: 28; Issue: 18 Linguagem: Inglês
10.1038/emboj.2009.224
ISSN1460-2075
AutoresDaniel A. Bonsor, Oliver Hecht, Mireille Vankemmelbeke, Amit Sharma, Anne Marie Krachler, Nicholas G. Housden, Katie J Lilly, Richard James, Geoffrey R. Moore, Colin Kleanthous,
Tópico(s)RNA and protein synthesis mechanisms
ResumoArticle20 August 2009free access Allosteric β-propeller signalling in TolB and its manipulation by translocating colicins Daniel A Bonsor Daniel A Bonsor Department of Biology, University of York, York, UKPresent address: Boston Biomedical Research Institute, 64 Grove Street, Watertown, MA 02472, USA Search for more papers by this author Oliver Hecht Oliver Hecht Centre for Molecular and Structural Biochemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich, UK Search for more papers by this author Mireille Vankemmelbeke Mireille Vankemmelbeke School of Molecular Medical Sciences, Institute of Infection, Inflammation and Immunity, Centre for Biomolecular Sciences, University of Nottingham, Nottingham, UK Search for more papers by this author Amit Sharma Amit Sharma Department of Biology, University of York, York, UK Search for more papers by this author Anne Marie Krachler Anne Marie Krachler Department of Biology, University of York, York, UK Search for more papers by this author Nicholas G Housden Nicholas G Housden Department of Biology, University of York, York, UK Search for more papers by this author Katie J Lilly Katie J Lilly Department of Biology, University of York, York, UK Search for more papers by this author Richard James Richard James School of Molecular Medical Sciences, Institute of Infection, Inflammation and Immunity, Centre for Biomolecular Sciences, University of Nottingham, Nottingham, UK Search for more papers by this author Geoffrey R Moore Geoffrey R Moore Centre for Molecular and Structural Biochemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich, UK Search for more papers by this author Colin Kleanthous Corresponding Author Colin Kleanthous Department of Biology, University of York, York, UK Search for more papers by this author Daniel A Bonsor Daniel A Bonsor Department of Biology, University of York, York, UKPresent address: Boston Biomedical Research Institute, 64 Grove Street, Watertown, MA 02472, USA Search for more papers by this author Oliver Hecht Oliver Hecht Centre for Molecular and Structural Biochemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich, UK Search for more papers by this author Mireille Vankemmelbeke Mireille Vankemmelbeke School of Molecular Medical Sciences, Institute of Infection, Inflammation and Immunity, Centre for Biomolecular Sciences, University of Nottingham, Nottingham, UK Search for more papers by this author Amit Sharma Amit Sharma Department of Biology, University of York, York, UK Search for more papers by this author Anne Marie Krachler Anne Marie Krachler Department of Biology, University of York, York, UK Search for more papers by this author Nicholas G Housden Nicholas G Housden Department of Biology, University of York, York, UK Search for more papers by this author Katie J Lilly Katie J Lilly Department of Biology, University of York, York, UK Search for more papers by this author Richard James Richard James School of Molecular Medical Sciences, Institute of Infection, Inflammation and Immunity, Centre for Biomolecular Sciences, University of Nottingham, Nottingham, UK Search for more papers by this author Geoffrey R Moore Geoffrey R Moore Centre for Molecular and Structural Biochemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich, UK Search for more papers by this author Colin Kleanthous Corresponding Author Colin Kleanthous Department of Biology, University of York, York, UK Search for more papers by this author Author Information Daniel A Bonsor1, Oliver Hecht2, Mireille Vankemmelbeke3, Amit Sharma1, Anne Marie Krachler1, Nicholas G Housden1, Katie J Lilly1, Richard James3, Geoffrey R Moore2 and Colin Kleanthous 1 1Department of Biology, University of York, York, UK 2Centre for Molecular and Structural Biochemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich, UK 3School of Molecular Medical Sciences, Institute of Infection, Inflammation and Immunity, Centre for Biomolecular Sciences, University of Nottingham, Nottingham, UK *Corresponding author. Department of Biology (Area 10), University of York, Heslington, PO Box 373, York, YO10 5YW, UK. Tel.: +44 0 1904 328820; Fax: +44 0 1904 328825; E-mail: [email protected] The EMBO Journal (2009)28:2846-2857https://doi.org/10.1038/emboj.2009.224 Correction(s) for this article Allosteric β-propeller signalling in TolB and its manipulation by translocating colicins16 September 2009 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The Tol system is a five-protein assembly parasitized by colicins and bacteriophages that helps stabilize the Gram-negative outer membrane (OM). We show that allosteric signalling through the six-bladed β-propeller protein TolB is central to Tol function in Escherichia coli and that this is subverted by colicins such as ColE9 to initiate their OM translocation. Protein–protein interactions with the TolB β-propeller govern two conformational states that are adopted by the distal N-terminal 12 residues of TolB that bind TolA in the inner membrane. ColE9 promotes disorder of this 'TolA box' and recruitment of TolA. In contrast to ColE9, binding of the OM lipoprotein Pal to the same site induces conformational changes that sequester the TolA box to the TolB surface in which it exhibits little or no TolA binding. Our data suggest that Pal is an OFF switch for the Tol assembly, whereas colicins promote an ON state even though mimicking Pal. Comparison of the TolB mechanism to that of vertebrate guanine nucleotide exchange factor RCC1 suggests that allosteric signalling may be more prevalent in β-propeller proteins than currently realized. Introduction The outer membrane (OM) of Gram-negative bacteria is an asymmetric bilayer composed of an inner layer of phospholipids and an outer layer of lipopolysaccharide (LPS) that serves an essential barrier function (Nikaido, 2003). Many questions remain unanswered concerning the OM, in particular how it is maintained and renewed in the face of continual environmental assault and how its assembly is coordinated during cell growth and division (Ruiz et al, 2006). It has long been known that the Tol system is involved in maintaining the integrity of the OM, but in what capacity and by what mechanism(s) have remained largely unknown (Lazzaroni et al, 2002; Cascales et al, 2007). In this work, we uncover an allosteric signal based on conformational transitions in the β-propeller protein TolB that lies at the heart of Tol function in Escherichia coli. The Tol system (also referred to as Tol–Pal) is ubiquitous in Gram-negative bacteria and is organized typically in the form of two operons: one encoding tolQ, tolR and tolA, the products of which are all inner membrane (IM) proteins, and one encoding tolB and pal (Sturgis, 2001). Pal (or peptidoglycan associated lipoprotein) resides in the inner leaflet of the OM and binds to the peptidoglycan layer, an interaction that is mutually exclusive of its interaction with TolB (Bouveret et al, 1995, 1999). Deletion of any of the tol genes leads to periplasmic contents leaking to the extracellular environment, formation of OM blebs or ruffles, reduction in the amount of LPS on the cell surface, cell division defects and sensitivity to large antibiotics, such as vancomycin, and detergents, such as SDS, that are normally excluded from the cell (Lazzaroni et al, 1989; Webster, 1991). With the exception of pal, all tol deletions also render cells resistant to filamentous bacteriophages (such as f1, fd and M13) and a wide range of antibacterial colicins, including the pore-formers ColN, ColE1, ColA and the nuclease toxins ColE2–E9 (Cascales et al, 2007). The tol genes have also been implicated in bacterial pathogenesis; tolA is upregulated during biofilm formation in Pseudomonas aeruginosa, whereas tolB is required for virulence in Salmonella typhimurium and Vibrio cholerae and its expression is regulated by Cl− ions in clinical isolates of Burkholderia cenocepacia in cystic fibrosis patients (Heilpern and Waldor, 2000; Whiteley et al, 2001; Tamayo et al, 2002; Bhatt and Weingart, 2008; Cameron et al, 2008). The function of the Tol system has remained enigmatic despite almost two decades of work largely because of the pleiotropic nature of mutations and deletions. Two possible functions have emerged: involvement in cell envelope biogenesis and/or as a tether in cell division that maintains the appropriate juxtaposition of the two membranes relative to the peptidoglycan layer (Cascales et al, 2007; Gerding et al, 2007). Both functions are built around two key features of the Tol system. First, that it is a transperiplasmic network spanning the two membranes of the bacterium, and, second, that it is coupled to the proton motive force (pmf). Traversal of the periplasm is through TolA, which has a C-terminal globular domain (TolAIII) connected to a long helical stalk domain (TolAII) that is anchored to the IM by a single transmembrane helix (TolAI) (Levengood et al, 1991; Levengood-Freyermuth et al, 1993). Coupling to the pmf is through TolQ and TolR, which associate with TolA in the IM and are functional and likely structural homologues of ExbB and ExbD of the Ton system (Kampfenkel and Braun, 1993; Cascales et al, 2001). ExbB and ExbD drive pmf-dependent entry of ligands such as iron–siderophore complexes and vitamin B12 through the plugged pores of 22-strand β-barrel OM receptors through TonB (Wiener, 2005). TonB spans the periplasm and contacts specific N-terminal sequences in nutrient receptors called TonB boxes (Braun and Endriß, 2007; Postle and Larsen, 2007). Both ExbB/ExbD and TolQ/TolR are related to the flagellar stator proteins MotA and MotB emphasizing their role in linking the pmf to protein conformational changes (Cascales et al, 2001; Germon et al, 2001). In the case of the Tol system, the pmf is reported to influence the interaction of TolA with Pal at the OM, but for what purpose is unclear (Cascales et al, 2000). This work sets out to address how the OM complex of TolB and Pal communicates with TolA in the IM. Through a combination of crystallography, nuclear magnetic resonance (NMR), isothermal titration calorimetry (ITC), cross-linking and in vivo assays, we describe a novel signal transduction mechanism in the bacterial periplasm that is centred on a disorder–order transition in TolB. We show how a conformational switch in TolB is the means by which the effects of Pal binding to TolB are communicated to TolA and how tol-dependent colicins exploit this switch to expedite their entry into cells. The role we describe for Pal runs counter to the accepted modus operandi for this assembly (Cascales et al, 2007), and questions the presumed transmembrane nature of the Tol system. Finally, we draw comparisons with another β-propeller protein that points to a general signal transduction mechanism for this class of protein. Results Resolving the N-terminus of TolB Several X-ray structures of TolB have been reported, including two structures of TolB in isolation (Figure 1A), TolB bound to Pal (Figure 1B) and TolB bound to a fragment of the translocation domain of ColE9 (Abergel et al, 1999; Carr et al, 2000; Loftus et al, 2006; Bonsor et al, 2007). TolB is a 44-kDa protein composed of two domains, an N-terminal α/β domain and a C-terminal six-bladed β-propeller to which both Pal and ColE9 bind (Loftus et al, 2006; Bonsor et al, 2007). Although informative of the conformational changes experienced by TolB as a result of its protein–protein interactions, these structures raise an important question concerning the N-terminus of the protein, which has only been structurally resolved in the TolB–Pal complex. As a periplasmic protein TolB carries a Sec-dependent signal sequence that is cleaved during entry to the periplasm, so, appropriately, all constructs used to determine its structure thus far have had this sequence (residues 1–21) deleted. However, the N-terminus of these constructs did not match precisely that of processed TolB in vivo, most often including an N-terminal methionine or a polyhistidine tag. (Note also that amino-acid numbering in earlier reports and pdb depositions have not taken account of the signal sequence, which is herein included in all TolB numbering.) We, therefore, re-determined the structure of the TolB–Pal complex at 1.86 Å resolution, using TolB purified from E. coli periplasmic extracts (see Materials and methods; Supplementary Table S1 in Supplementary data) to resolve the wild-type N-terminus. Figure 1.Structural changes in TolB suggest a conformational signal operates in the Tol system. (A) Crystal structure of unliganded TolB (pdb, 1c5k). *indicates where density for the polypeptide chain begins, 12 amino acids are missing presumed unstructured. (B) 1.8 Å crystal structure of the TolB–Pal complex highlighting how the N-terminal 12 residues of TolB (green) become ordered on binding Pal. (C) Molecular surface of the domain–domain interface of unliganded TolB, as in (A), showing how the 'proline gate' (Pro415) shuts off access of the N-terminal residues to the surface. (D) Molecular surface of the domain–domain interface of TolB in the latest TolB–Pal structure showing opening of the proline gate and structural resolution of TolB residues Glu22–Ser33. Download figure Download PowerPoint The root mean square deviation of the TolB–Pal complex relative to that determined earlier for the cytoplasmically expressed complex was 0.3 Å. Hence, TolB purified from periplasmic extracts behaves identically to that purified from the cytoplasm with the exception that the N-terminal residue Glu22 is now resolved in the structure (Figure 1D). In this analysis, we do not describe the interface of the TolB–Pal complex, as this has been documented earlier (Bonsor et al, 2007). Instead, we focus on structural changes to the TolB N-terminus that have yet to be detailed. Several loops and propeller β-strands in TolB move as a result of Pal binding. These cause the latching or 'Velcro' strand of the β-propeller, which conjoins the first and last propeller blades (Neer and Smith, 1996), to move away from the domain–domain interface carrying with it a proline residue (Pro415; Figure 1C and D). The movement of this 'proline gate', which in unliganded TolB is within van der Waals distance of Tyr190, reveals a canyon on the TolB surface between its N- and C-terminal domains that constitute a binding site for the 12 N-terminal TolB residues (Glu22–Ser33). The N-terminus forms a helical half-turn and an anti-parallel β-sheet against this surface (Figure 1B), stabilized by 12 hydrogen bonds, 9 to the β-propeller latching strand and a further 3 with residues in the N-terminal domain. Further stabilization of the bound N-terminus occurs through van der Waals interactions of several hydrophobic residues (Val23, Ile25, Val26 and Ile27) with docking sites in the canyon (Figure 1D). In total, the self-association of the N-terminus with the body of TolB buries ∼1700 Å2 accessible surface area, equivalent to three-quarters of the total accessible surface area buried at the TolB–Pal interface. We next investigated the functional importance of the TolB N-terminus. The TolB N-terminus is essential for Tol function and colicin toxicity There is currently no in vitro biochemical assay for Tol function and so the effects of mutations are typically assayed in vivo through complementation of the tol phenotype in deletion strains. An E. coli tolB deletion strain (JW5100) was transformed with a pBAD vector under the control of AraC (arabinose inducible) into which had been cloned wild-type tolB targeted to the periplasm by its own signal sequence. To test OM integrity, growth of this complemented strain was assessed on rich media agar plates in the presence of inducer with or without 2% SDS. Growth was compared with a vector control and two truncations: one in which four residues had been deleted from the N-terminus (TolB Δ22−25), the other in which the entire N-terminal sequence was deleted (TolB Δ22−33). All cells were viable in the absence of SDS, but only cells transformed with wild-type tolB grew in the presence of SDS indicating that both truncations generated a tol phenotype (Table I). In a series of control experiments, we verified that N-terminal TolB truncations were targeted to the periplasm and processed correctly and that the deletions had no effect on protein structure, determined by far UV-CD spectroscopy of the purified proteins, which were essentially indistinguishable from wild type (data not shown). We conclude that the N-terminus of TolB is critical to Tol function and that its removal does not grossly perturb protein structure. Table 1. The N-terminal residues of TolB are required for Tol function in vivo No SDS (105 CFU/ml) 2% SDS (105 CFU/ml) tolB− 5.1 0 TolB 9.9 7.5 TolB Δ22–25 6.3 0 TolB Δ22–33 3.3 0 E. coli JW5100 cells were transformed with pDAB17 encoding wild-type TolB or derived plasmids encoding TolB Δ22–25 and TolB Δ22–33 all containing the TolB secretion signal, grown on LB agar with or without 2% w/v SDS and the number of colony forming units (CFU) determined. JM5100 cells transformed with the vector (pBAD24) was included as a tolB− control. Most Gram-negative bacteria are equipped to release bacteriocins during times of stress as competitive agents that target and kill neighbouring strains of the same species (Riley and Kirkup, 2004). Colicins are specific for E. coli and share a similar domain architecture, comprising a central receptor-binding domain flanked by an N-terminal (T-) domain involved in translocation across the OM and a C-terminal cytotoxic domain (Cascales et al, 2007). The T-domain dictates the route colicins take through the periplasm, with group A colicins commandeering the Tol system and group B colicins the Ton system. ColE9 is a group A colicin that elicits cell death through metal-dependent endonuclease cleavage of the bacterial genome by an enzymatic domain related to the HNH family of homing endonucleases and the apoptotic DNase CAD (Kleanthous et al, 1999; Walker et al, 2002). Translocation of ColE9 across the E. coli OM is dependent on both a functional Tol system and specific interaction with the TolB β-propeller domain (Loftus et al, 2006). We tested strains expressing the N-terminal TolB deletion constructs for ColE9 sensitivity and found that both were completely resistant towards the colicin (Figure 2A). We also tested the ability of these strains to promote dissociation of the colicin-immunity protein complex. Nuclease colicins, such as ColE9, are co-synthesized with a high-affinity-immunity protein (Kd∼10−14 M), which prevents suicide of the producing organism (Kleanthous and Walker, 2001). We and others have shown this complex is dissociated at the E. coli cell surface during translocation in a pmf-dependent manner that requires an intact Tol assembly (Duche et al, 2006; Zhang et al, 2008; Vankemmelbeke et al, 2009). Using Alexa594-labelled Im9 bound to ColE9, we found that E. coli cells expressing either TolB Δ22−25 or TolB Δ22−33 in the periplasm exhibited significantly reduced Im9 release at the cell surface (Figure 2B). We conclude that the N-terminus of TolB is required both for release of the colicin-bound immunity protein at the cell surface and colicin cytotoxicity even though the colicin makes no direct contact with this region of TolB. Figure 2.The N-terminal residues of TolB are required for colicin toxicity and immunity protein release. (A) Sensitivity of E coli JW5100 cells transformed with plasmid pDAB17 encoding wild-type TolB or derived plasmids encoding TolB Δ22−25 and TolB Δ22−33 towards a serial dilution of the endonuclease colicin ColE9 (see Materials and methods for details). Zones of clearance indicate colicin activity against the strain. Only wild-type TolB cells are ColE9 sensitive. (B) Im9 release from the ColE9–Im9 complex is compromised in E coli JW5100 cells expressing the TolB N-terminal deletion mutants. Data show triplicate measurements for the release of Alexa-594-labelled Im9 at the cell surface from a ColE9S−S–Im9 complex in which cell entry of the colicin was initiated by the reduction of an inactivating disulphide bond across the receptor-binding domain (Penfold et al, 2004). Download figure Download PowerPoint The N-terminus of TolB is the TolA-binding site Earlier yeast two-hybrid experiments and an E. coli suppressor screen have identified an interaction between TolAIII and the N-terminal α/β domain of TolB, but no specific binding site has yet been defined (Dubuisson et al, 2002; Walburger et al, 2002). Indeed, there has been no biochemical or biophysical characterization of the isolated TolAIII–TolB complex, reflecting the difficulty in showing any interaction between the proteins in vitro. Lakey and co-workers, for example, reported that no interaction between a domain TolAII–III construct and TolB could be detected by ITC, although in this instance binding may have been masked by an N-terminal polyhistidine tag on TolB (Gokce et al, 2000). In trying to establish whether TolB is capable of binding TolAIII in vitro, we adopted initially a chemical cross-linking approach using formaldehyde in a bid to detect weakly bound complexes. We found that purified proteins incubated at a concentration of 10 μM yielded cross-linked adduct detectable by Coomassie staining (data not shown), but more clearly visualized by western blotting using antibodies raised against TolA or TolB (Figure 3A). Importantly, no cross-linking between TolAIII and TolB was detectable for either of the TolB N-terminal truncations, Δ22−25 and Δ22−33. The interaction between TolB and TolAIII was verified using ITC, although these experiments were hampered by the limited solubility of TolB purified from periplasmic extracts. Although binding isotherms were incomplete, they were nevertheless sufficient to indicate that binding is endothermic, and so entropically driven at 20°C, pH 7.5 and weak, Kd ∼40 μM (Figure 3B; Table II). The specific nature of the TolB–TolAIII interaction was again confirmed by the two N-terminal truncations neither of which showed evidence of binding in ITC experiments (Figure 3B). In summary, chemical cross-linking and ITC show that the N-terminal 12 residues of TolB constitute most, if not all the TolAIII-binding epitope, herein referred to as the TolA box, and that the two proteins form a weak complex in vitro. Through a suppressor screen of TolAIII mutants Lazzaroni and co-workers earlier identified TolB Asp120 as potentially involved in binding TolAIII (Dubuisson et al, 2002). However, this residue is on the opposite face of the TolA box identified in this study, and so the effect observed in the earlier study is most likely indirect. Moreover, Asp120 experiences no structural changes as a result of Pal binding, which we show below is a key regulator of the TolB–TolA interaction. Figure 3.The N-terminal residues of TolB are required for TolAIII binding. (A) Formaldehyde (F; 1%) cross-linking reactions, analysed by western blotting using anti-TolA (left) and anti-TolB (right) antibodies, of purified TolB (WT), TolB Δ22−25 and TolB Δ22−33 incubated with TolA III (each at 10 μM). +, cross-linked; −, untreated. (B) Raw ITC data (top panel) and integrated heats (lower panel) for wild-type TolB and TolB truncation mutants Δ22−25 (grey triangles) and Δ22−33 (open circles) at a cell concentration of 60 μM binding TolAIII in 50 mM Hepes buffer pH 7.5, containing 50 mM NaCl. Proteins had earlier been loaded with Ca2+ ions, which bind within the β-propeller of TolB (see Materials and methods). TolAIII binding is abolished by the deletion of just four amino acids from the TolB N-terminus. Download figure Download PowerPoint Table 2. Thermodynamic parameters for binary and ternary TolB protein–protein interactions obtained by ITC Proteins ΔH (kcal/mol) ΔS (cal/K/mol) N Kd TolB+T-domain −20.6 (±0.04) −38.6 (±0.2) 0.93 (±0.01) 125 (±4) nM TolB Δ22–25+T-domain −23.0 (±0.07) −43.8 (±0.2) 0.93 (±0.01) 28 (±1) nM TolB Δ22–33+T-domain −23.6 (±0.02) −45.7 (±0.1) 0.94 (±0.01) 25 (±1) nM TolB+Pala −9.47 (±0.10) 2.7 (±0.5) 0.98 (±0.00) 38 (±3) nM TolB Δ22–25+Pala −7.20 (±0.11) 6.0 (±0.4) 0.91 (±0.02) 313 (±15) nM TolB Δ22–33+Pala −7.52 (±0.06) 4.8 (±0.3) 0.91 (±0.02) 337 (±18) nM TolB+TolAIII 2.9 (±0.02) 30.0 (±0.1) 1.00 (±0.04) 43 (±2.1) μM TolB Δ22–25+TolAIII NH TolB Δ22–33+TolAIII NH TolB-T-domain+TolAIII −0.58 (±0.03) 20.4 (±0.3) 1.16 (±0.04) 13 (±1) μM TolB-Pal+TolAIII NH TolA+Palb NH NH, no heats detected. All data were collected at 20°C unless otherwise stated and in 50 mM Hepes buffer at pH 7.5 containing 50 mM NaCl and CaCl2 (see Materials and methods for details). In all cases, the cell compartment contained wild-type or mutant TolBs purified from the E. coli periplasm or TolB in complex with Pal or ColE9 T-domain (typically at a concentration of ∼60 μM). Data presented were obtained by fitting integrated heats to 1:1 binding models using Origin software and show the means of two independent observations. Errors are the standard errors of the mean. a Data collected at 30°C because no heats were detectable at 20°C. b Experiments carried out at both 20 and 30°C. Opposing modulation of the TolB–TolA interaction by Pal and ColE9 To translocate into E. coli, ColE9 competitively recruits TolB from its complex with Pal using a 16-residue intrinsically unstructured TolB-binding epitope (TBE) that is part of its 30 kDa T-domain. Structural comparisons of the TBE bound to TolB with that of the present TolB–Pal complex reveals that although the colicin and Pal bind at the same site on the β-propeller, their impact on TolB differs dramatically (Figure 4). ColE9 does not cause any gross change in TolB structure, whereas Pal induces long-range conformational changes that result in ordering of the TolA box. Closer inspection of the two interfaces reveals a simple yet sophisticated stealth mechanism used by the colicin to ensure conformational ordering of the TolA box does not occur. First, the ColE9 TBE does not contact regions of TolB that induce ordering of the TolA box, as opposed to Pal in which residues on the outer periphery of the β-propeller domain move up to 5 Å to embrace it (Figure 4A). Second, the colicin inserts a tryptophan sidechain into a pocket that becomes occluded in the Pal-bound form of TolB (Figure 4), thereby blocking any conformational changes. Consistent with this interpretation, mM concentrations of tryptophan weaken the TolB–Pal complex by ∼30-fold (Bonsor et al, 2007). Figure 4.Structural basis for differential stabilization of TolA box residues by Pal, but not ColE9 binding to the β-propeller domain of TolB. (A) 16-residue ColE9 TBE (cyan ribbon) in complex with TolB, pdb code 2ivz. (B) Pal (purple) in complex with TolB, this work. For clarity, only Pal helices in contact with TolB are indicated. In both panels, TolB is shown as van der Waals surface and coloured according to Cα movements experienced by residues that are involved in protein–protein interactions: black, 2.5 Å. The stippled side-chain shown in (A) is that for Trp46 of the ColE9 TBE that slots into a pocket on the TolB surface. This side-chain has been superimposed on the TolB–Pal complex in (B) to highlight how the pocket into which Trp46 inserts becomes partially occluded, and so an equivalent interaction is not possible. Constriction of the pocket is due to conformational changes in TolB that ultimately result in sequestration of the TolA box to the TolB surface (B, orange). The TolA box is not resolved in the ColE9-bound structure and so is not depicted in (A). Download figure Download PowerPoint We reasoned whether these differential effects on TolB could impact on its ability to recruit TolAIII through the TolA box, and so this was investigated by formaldehyde cross-linking and ITC. Figure 5 shows the effect of Pal and ColE9 T-domain binding on TolAIII–TolB cross-linking. In this experiment, TolB purified from periplasmic extracts was first complexed with either Pal (Figure 5A) or ColE9 T-domain (Figure 5B) and the products of cross-linking to TolA compared with reactions lacking Pal or ColE9. The data show that while Pal reduces greatly TolB–TolAIII cross-linking, it does not abolish it or that of a ternary complex containing Pal–TolB–TolAIII, although this is also in low yield. Critically, ColE9 T-domain enhances the TolB–TolAIII cross-link and produces more of the ternary ColE9–TolB–TolAIII complex relative to that of Pal. These data suggest that group A colicins promote association of TolB with TolAIII, enhancing both binary and ternary complexes, whereas Pal disfavours their formation. Hence, ColE9 and Pal exert opposing effects on the equilibrium of the TolB–TolAIII complex. Figure 5.Differential effects of Pal and ColE9 TBE on TolB–TolAIII cross-linking. Purified TolB and TolAIII (10 μM each) were incubated together and cross-linked with formaldehyde (F), as described in Materials and methods, in the presence or absence of Pal (10 μM), (A) or the presence or absence of ColE9 T-domain (10 μM), (B). Products of these reactions were analy
Referência(s)