PMC:2724173 / 1404-14776 JSONTXT

{"project":"2_test","denotations":[{"id":"18304577-15701523-62526544","span":{"begin":5411,"end":5412},"obj":"15701523"},{"id":"18304577-10094626-62526545","span":{"begin":5512,"end":5514},"obj":"10094626"},{"id":"18304577-8126725-62526546","span":{"begin":6269,"end":6271},"obj":"8126725"},{"id":"18304577-11551174-62526547","span":{"begin":7734,"end":7736},"obj":"11551174"},{"id":"18304577-17064953-62526548","span":{"begin":8048,"end":8050},"obj":"17064953"},{"id":"18304577-17496915-62526548","span":{"begin":8048,"end":8050},"obj":"17496915"},{"id":"18304577-15757514-62526549","span":{"begin":8185,"end":8186},"obj":"15757514"},{"id":"18304577-16091038-62526550","span":{"begin":8786,"end":8788},"obj":"16091038"},{"id":"18304577-16091038-62526551","span":{"begin":8909,"end":8911},"obj":"16091038"},{"id":"18304577-8377180-62526552","span":{"begin":9478,"end":9480},"obj":"8377180"},{"id":"18304577-15572779-62526553","span":{"begin":9539,"end":9541},"obj":"15572779"},{"id":"18304577-12894216-62526554","span":{"begin":9912,"end":9914},"obj":"12894216"},{"id":"18304577-17060447-62526555","span":{"begin":10061,"end":10063},"obj":"17060447"},{"id":"18304577-16086183-62526556","span":{"begin":10294,"end":10296},"obj":"16086183"},{"id":"18304577-17485524-62526556","span":{"begin":10294,"end":10296},"obj":"17485524"},{"id":"18304577-11115111-62526557","span":{"begin":10511,"end":10512},"obj":"11115111"}],"text":"Shigella flexneri is a Gram-negative bacterial pathogen that causes human bacillary dysentery resulting in over a million deaths annually worldwide. The pathogenicity of Shigella sp. is dependent on a complex macromolecular machine, the type 3 secretion system (T3SS), that delivers into host cells a set of effector proteins required for invasion. The Shigella sp. T3SS consists of structural components of the injection machinery, secreted proteins, chaperones and regulators, all of which are encoded by approximately 25 genes located in the mxi, spa and ipa operons on a large 230 kb plasmid.1–3\nThe delivery of effectors into host cells involves secretion, the crossing of both bacterial membranes via the basal body, and translocation, the passage through the eukaryotic cell membrane. Following assembly of the external needle, the proteins secreted via the T3SS fall into two main categories: translocators and effectors. Upon host cell contact, translocators assemble into the host cell membrane, forming a pore complex, or translocon, that triggers the subsequent export of effectors.4 Since translocators must be secreted before effectors, so that effectors will be exported directly into host cells instead of the extracellular milieu, pathogens require mechanisms to ensure hierarchical and temporal control over their secretion. Although the exact mechanisms underlying these processes are not clearly established, several cytoplasmic and inner-membrane proteins have been identified that recognize secretion substrates and respond to specific signals to ensure that structural and sensing components (needle subunits and pore proteins) are secreted first, and that virulence proteins are not secreted before contact with a host cell.\nBlockage of effector secretion before host cell contact is mediated, in part, by a protein that has been proposed to act as either a physical impediment to the entrance to the secretion apparatus,5 or as a gatekeeper that determines substrate hierarchy.6 Across bacterial species, this protein (known as MxiC in Shigella sp.) possesses only weak sequence homology and in some species exists as two separate polypeptide chains (e.g., in Yersinia sp. the homologue consists of YopN and TyeA).7 Despite this, distinct functional homologies can be identified across species. Functional knock-outs of members of this family have no effect on needle formation or stability but significantly reduce or abolish the secretion of translocators.8–11 In addition, in several species these mutations also result in enhanced secretion of effector proteins.9–14 This differential effect on translocator and effector secretion suggests that these proteins have a role in T3SS discrimination between secreted proteins involved in translocation and proteins that have effector function.\nThere are, however, several differences between the members of this family. Most notably, activation of type 3 secretion in Yersinia sp. results in the secretion of YopN, while TyeA remains in the bacterial cytoplasm.15,16 The dissociation of YopN and TyeA has been proposed as a mechanism for the regulation of secretion but clearly cannot be a conserved mechanism in those species where the homologue is a single polypeptide chain.17 In the crystal structure of the Y. pestis YopN-TyeA complex, the close proximity of the C terminus of YopN with the N terminus of TyeA suggested that a single polypeptide encoding both proteins could maintain the same overall structure.5 In order to confirm this, we have determined and refined the structure of the S. flexneri homologue, MxiC, in three distinct crystal forms. The molecular architecture and movement of the domains of MxiC compared with that of YopN-TyeA highlights conformational differences that may play an important role in T3SS signaling.\n\nProteolytic susceptibility of the N terminus of MxiC\nA full-length construct of MxiC (residues 1–355, MxiCFL) was purified by nickel-affinity chromatography followed by size-exclusion chromatography and revealed that MxiCFL elutes at a volume less than that of a monomer, but greater than that of a dimer (Fig. 1a). This result, combined with dynamic light-scattering data that revealed a major species with a larger than expected hydrodynamic radius (Rh ∼ 3.8 nm), suggested that MxiCFL does not possess a globular structure and may possess an elongated fold and/or be partially disordered. In crystallization trials, MxiCFL initially yielded three different crystal forms (two distinct P21 and one P43212) that diffracted only to 3.5–3.9 Å resolution. Selenomethionine-labeled MxiCFL yielded crystals (P43212 with a different cell) that diffracted to 3.2 Å resolution and were used for phasing by the multiple-wavelength anomalous diffraction (MAD) method (Table 1). Preliminary model building into these maps revealed that the first 70–80 residues were poorly ordered and not visible in the electron density. In addition, we observed that, in solution, the N terminus was susceptible to proteolytic degradation (Fig. 1b). Limited proteolysis with subtilisin followed by N-terminal sequencing and mass spectrometry revealed several degradation products resulting from cleavage of the N terminus up to residue 64. This proteolytically sensitive region of MxiC is equivalent to the region of YopN (32–76) that was shown to bind its chaperone and was disordered in the absence of this chaperone.5 Furthermore, in vivo, the stable expression and efficient secretion of YopN requires its chaperone.18 The proteolytic susceptibility of the N-terminal region of MxiC suggests that chaperone binding may act to protect MxiC from degradation via a similar mechanism. To date, a chaperone for MxiC has not been identified. In order to improve the quality of MxiC crystals, a shortened construct encompassing residues 74–355 (MxiCNΔ73) was expressed, purified (Fig. 1a) and subjected to crystallization trials.\n\nThe structure of MxiC\nReductive methylation of MxiCNΔ73 yielded crystals that diffracted to higher resolution: 2.85 Å and 2.5 Å in space groups P212121 and P222, respectively (Table 1). MxiC is an elongated rod-shaped molecule with a long axis of 86 Å (Fig. 2a). It is composed of three domains, each possessing a four-helix X-bundle fold (Fig. 2b).19 The first and last domains consist only of the X-bundle motif while the central domain also possesses a bent helix (α5) that is packed against domain 1. It is the first two domains of MxiC that are equivalent to YopN while the third domain of MxiC, equivalent to TyeA, is connected to domain 2 via a ten-residue linker. This linker acts only to tether the domains and does not have any major structural role, allowing the equivalent regions of MxiC and YopN–TyeA to adopt similar folds. There are a total of seven independent MxiC molecules in the crystallographic asymmetric units of the three refined crystal forms (see Table 1). These structures reveal that, although the fold of each domain is maintained in all structures (rmsd over Cα atoms of domains 1, 2 and 3 are 0.5 Å, 0.9 Å and 0.8 Å, respectively), there is some flexibility at the interfaces between domains resulting in a “wobble” of the terminal domains about the central domain (rmsd over all Cα atoms of 1.4 Å; Fig. 2c). The elongated shape of MxiC means that the most distal regions undergo the greatest displacement while the more central interdomain interfaces undergo minimal change.\nThe structural similarity between the domains of MxiC suggests that they might be the product of gene duplication events. Such internal repetition arising via intragenic duplication and recombination events has been a successful stratagem throughout evolution for enlargement of the available surface area.20 This, combined with the elongated shape of MxiC, provides the maximal exposure of surface area and considerable binding interfaces suitable for large substrates, linear peptides or multiple partners. This feature is typical of scaffolding proteins that act to recruit multiple proteins and enhance signaling.21,22\nPallen et al. used multiple sequence alignments to identify members of the MxiC/YopN–TyeA family from divergent Gram-negative species;7 a structure-based sequence alignment of MxiC and YopN–TyeA (Fig. 3a) differs in the N-terminal region from the alignment obtained in that work, and highlights the need to be cautious when interpreting conservation based on sequence alignments alone. The combination of structural and sequence alignment information identifies that most of the highly conserved residues are involved in maintaining the correct fold of these proteins. In support of this, mutations of YopN that constitutively block secretion replace highly conserved residues that are buried in the interfaces between domains of YopN.23 Ferracci et al. suggested that these mutations are likely to cause part of YopN to become unfolded or less well-ordered.23 Interestingly, mapping of sequence conservation onto the structure across the more closely related Salmonella, Burkholderia and Shigella species reveals a hydrophobic patch on the surface of the central domain consisting of residues Leu222, Met226, Gly239, Leu242 and Leu245 (Fig. 3b, red circle). This region is buried by crystal contacts in all seven independent molecules of MxiC in the P212121, P222 and P43212 crystal forms, suggesting that this region may be a “hot spot” for protein–protein interactions.\nA structural homology search using the DALI algorithm,24 and the Secondary Structure Matching algorithm of MSDfold,25 was unable to identify any protein, other than YopN, possessing multiple X-bundle motifs. However, two classes of proteins were identified that possess X-bundle motifs similar to the individual domains of MxiC. The structural homologues belonged to the Bcl-2 family of apoptosis regulators, specifically the Caenorhabditis elegans homologue Ced9 (DALI Z-score 4.3, 1ohu)26 and the programmed cell death proteins involved in inhibition of protein synthesis, specifically the MA3 domain of Pdcd4 (DALI Z-score 3.9, 2nsz).27 Although the significance of these similarities is uncertain, structural mimicry of apoptotic factors is a common theme seen in viral and bacterial pathogens and represents an important defense against the host immune response.28,29 This structural similarity may be of functional relevance, as MxiC has been identified as a protein that is secreted by the T3SS under conditions that mimic those encountered by bacteria during infection of a host.3\n\nComparison with the YopN–TyeA complex\nAlthough the individual domains of MxiC and YopN–TyeA adopt similar folds (rmsd over Cα atoms of the first and second domains of MxiC with the equivalent regions of YopN of 3.9 Å and 3.2 Å, respectively, and the third domain of MxiC with TyeA of 0.8 Å), the arrangement of these domains results in a different overall shape for these molecules (Fig. 3b). This structural rearrangement differs from the “wobble” seen between different MxiC molecules and is particularly noticeable in one orientation (Fig. 3b, top panel). This view highlights the straight conformation adopted by MxiC and the relative curvature of the YopN–TyeA complex. The different position and orientation of the first domain of MxiC may be due, in part, to the missing N-terminal portion, as the YopN–TyeA structure possesses an additional helix at its N terminus.\nIt seemed likely that the major difference between these structures would be the arrangement of the domain equivalent to TyeA as this is a separate polypeptide in the Yersinia structure. Surprisingly, the most striking difference instead involves the long á-helix (α9) in the central domain of MxiC, which is straight in all of our MxiC structures, but possesses a sharp kink in YopN when complexed with TyeA (Fig. 3b). It is the straightening of this helix that results in a reorientation of the C-terminal domain of MxiC, compared with TyeA, and opens one face of the molecule. As the sequence at the hinge is not conserved, and all MxiC molecules possess the straight helix conformation, these structures may represent genuine differences between species. Alternatively, it may represent a conformational switch that, in this case, was captured due to the very different pH for the YopN–TyeA structure (pH 10.5) compared with the MxiC structures (pH 6.5–7.5).\nDespite the lack of sequence conservation on the surface of these structures, an analysis of the surface electrostatics reveals that the face that is curved in YopN–TyeA and open in MxiC possesses a conserved negatively charged patch (Fig. 3c). This patch spreads across one face of the C-terminal half of the molecule (displayed in red on the right-hand side of the surface shown in Fig. 3c). The helices that line this face undergo minor repacking and can accommodate the large movements of the surrounding domains. The conserved patch on this face suggests a role for this region in interactions with partner proteins. If this is a binding face for another component of the T3SS, it is interesting to note that this interface is conserved despite it being intramolecular in one homologue and intermolecular in another.\n\nProtein Data Bank accession numbers\nThe atomic coordinates and structure factors have been deposited at the RCSB Protein Data Bank with accession codes 2vix (P212121), 2vj4 (P222) and 2vj5 (P43212)."}