Wzi is an outer membrane lectin that underpins group 1 capsule assembly in Escherichia coli

Abstract

Many pathogenic bacteria encase themselves in a polysaccharide capsule that provides a barrier to the physical and immunological challenges of the host. The mechanism by which the capsule assembles around the bacterial cell is unknown. Wzi, an integral outer-membrane protein from Escherichia coli, has been implicated in the formation of group 1 capsules. The 2.6 Å resolution structure of Wzi reveals an 18-stranded β-barrel fold with a novel arrangement of long extracellular loops that blocks the extracellular entrance and a helical bundle that plugs the periplasmic end. Mutagenesis shows that specific extracellular loops are required for in vivo capsule assembly. The data show that Wzi binds the K30 carbohydrate polymer and, crucially, that mutants functionally deficient in vivo show no binding to K30 polymer in vitro. We conclude that Wzi is a novel outer-membrane lectin that assists in the formation of the bacterial capsule via direct interaction with capsular polysaccharides.

Figure 1

Crystal Structure of Wzi (A) The X-ray crystal structure of Wzi reveals a β-barrel fold. Wzi is oriented with the extracellular face at the top. The extracellular loops are colored from red to blue according to primary sequence, and the periplasmic N-terminal helical bundle is colored magenta. Sheets β3, β4, and β5 are labeled. (B) Top view of extracellular loops gradient colored according to primary sequence. Loops with asterisks indicate the loops that were deleted for assay; the helical bundle has been removed for clarity. (C) Bottom view of periplasmic turns, gradient colored according to primary sequence. (D) Structural alignment of Wzi (white) with its closest structural homolog PflBenF, a BenF-like outer-membrane protein from Pseudomonas fluorescens (red, PDB ID code 3JTY) (Sampathkumar et al., 2010). The alignment was generated using PDBeFold (Krissinel and Henrick, 2004) and extracellular loops and helices have been removed for clarity. Structural figures were created with UCSF Chimera (Pettersen et al., 2004). See also Figures S1, S2, and S4.

Figure 2

Analysis of Wzi Mutants (A) Colonies of wild-type K. pneumoniae 889/50 (wtKp, serotype O1:K20) and CWG874 (wzi::cat) were indistinguishable on LB agar plates but were different from those of a capsule-deficient strain, CWG171 (K. pneumoniae O1:K⁻, strain KD2 (McCallum et al., 1989)). (B) After centrifugation, wild-type K. pneumoniae 889/50 formed a diffuse pellet (red arrow), indicative of a coherent capsule. The pellet from CWG874 was more compact, similar to the pellet from the unencapsulated CWG171 (blue arrow). Shown here as examples, expression of WziΔL8 and WziFLAG, but not WziΔL7, could produce a cell pellet phenotype identical to that of the wild-type strain. (C) Immunoblotting of cell lysates of CWG874 expressing Wzi mutants probed with anti-His₅ antibodies (QIAGEN). Three times more sample was loaded for the helix-deletion mutants relative to the other samples, due to reduced expression levels. The blot was further contrast enhanced to visualize the fainter bands of the helix mutants. A summary of the cell pelleting results for each Wzi mutant is listed below the blot; “+” represents a rescue of the wild-type cell pellet phenotype. See also Tables S1 and S2.

Figure 3

Wzi Binds to K30-Derived Oligosaccharides (A) Segments of the K30 CPS repeat unit, comprising two different tetrasaccharides (N1 and N2), were tested for binding. A third analogous tetrasaccharide (A1, right) was also tested. (B) Steady-state affinity curve of immobilized Wzi for polymeric K30 measured by SPR. (Inset) SPR sensorgram of K30 concentration series titrated over Ni-NTA-immobilized Wzi. (C) Steady-state affinity curve of immobilized Wzi for monomeric K30N2. (Inset) SPR sensorgram of a K30N2 concentration series titrated over Ni-NTA-immobilized Wzi. (D) Steady-state affinity curve of immobilized WziΔL8 for K30 polymer. (Inset) SPR sensorgram of K30 polymer concentration series titrated over Ni-NTA-immobilized WziΔL8. (E) Vacuum electrostatic rendering of the extracellular surface of Wzi. The location of each extracellular loop is labeled. Boxed loops indicate the loops that were deleted and tested for binding to K30. Figure constructed using the PyMOL Molecular Graphics System (Version 1.5.0.3, Schrödinger, LLC). See also Table S1 and Figures S3 and S5.

Figure 4

Phylogenetic Tree of Wzi Homologs The Wzi hits, identified by BLAST searches of completed bacterial genomes (12-2011) and refined by secondary-structure prediction, clustered into groups A–C. The Wzi hits in Groups A and B are all from organisms belonging to the Gammaproteobacteria class, except for Verrucomicrobiae bacterium DG1235, from the Verrucomicrobia phylum. Group C is more diverse, containing organisms from the following phyla: Acidobacteria, Nitrospirae, Thermodesulfobacteria, Deferribacteres, and Chlorobi, as well as many from Proteobacteria (Alpha, Delta, and Gamma classes). This tree has been abbreviated for clarity; branches with multiple hits of the same genus were represented by the top hit of that branch, and bootstrap values were removed. See also Table S4.

Figure 5

The Role of Wzi Lipid-linked components of the polysaccharide capsule are synthesized in the cytoplasm and extended by the Wzy-dependent pathway. Wza translocates polymerized capsular components en masse to the outer membrane. It is proposed that Wzi serves as an initial tethering point, capturing translocated polysaccharide that serves as a nucleation point for further secreted CPS. (Inset) E. coli with the K30 capsule.