Cadaverine covalently linked to peptidoglycan is required for interaction between the peptidoglycan and the periplasm-exposed S-layer-homologous domain of major outer membrane protein Mep45 in Selenomonas ruminantium

Abstract

The peptidoglycan of Selenomonas ruminantium is covalently bound to cadaverine (PG-cadaverine), which likely plays a significant role in maintaining the integrity of the cell surface structure. The outer membrane of this bacterium contains a 45-kDa major protein (Mep45) that is a putative peptidoglycan-associated protein. In this report, we determined the nucleotide sequence of the mep45 gene and investigated the relationship between PG-cadaverine, Mep45, and the cell surface structure. Amino acid sequence analysis showed that Mep45 is comprised of an N-terminal S-layer-homologous (SLH) domain followed by α-helical coiled-coil region and a C-terminal β-strand-rich region. The N-terminal SLH domain was found to be protruding into the periplasmic space and was responsible for binding to peptidoglycan. It was determined that Mep45 binds to the peptidoglycan in a manner dependent on the presence of PG-cadaverine. Electron microscopy revealed that defective PG-cadaverine decreased the structural interactions between peptidoglycan and the outer membrane, consistent with the proposed role for PG-cadaverine. The C-terminal β-strand-rich region of Mep45 was predicted to be a membrane-bound unit of the 14-stranded β-barrel structure. Here we propose that PG-cadaverine possesses functional importance to facilitate the structural linkage between peptidoglycan and the outer membrane via specific interaction with the SLH domain of Mep45.

FIG. 1.

Amino acid sequence profiles of Mep45 and determination of the protease-accessible site. (A) Amino acid sequence alignment of the N-terminal SLH domain of Mep45 with the corresponding region of P100 and SomB. Asterisks denote highly and specifically conserved residues found in SLH-bearing outer membrane proteins. (B and C) SDS-PAGE of proteinase K or trypsin treatment of intact cells (B) and crude envelopes (C) of S. ruminantium. Intact cells (3 mg) or crude envelopes (1 mg) were incubated with 100 μg/ml proteinase K or trypsin at 37°C for 30 min. Gels were stained with Coomassie brilliant blue. The bands designated by the arrowheads are the fragments of Mep45 produced by the protease treatment. Abbreviations: IC, intact cells; CE, crude envelopes; P, proteinase K-treated sample; T, trypsin-treated sample. (D) Schematic representation of the topology of Mep45. The locations of proteolytic sites of proteinase K and trypsin are indicated as black and open triangles, respectively. The N-terminal SLH domain protrudes into the periplasmic space, and the C-terminal β-strands span the outer membrane, except for at least one region exposed to the external cell surface, which is the protease-sensitive site.

FIG. 2.

In vitro peptidoglycan-binding assay. (A) Western immunoblot of Mep45 binding to wild-type S. ruminantium peptidoglycan. Solubilized Mep45 (3 μg) was mixed with or without 1 mg of purified peptidoglycan (PG). The peptidoglycan-associated Mep45 (ppt) and non-peptidoglycan-associated Mep45 in the supernatant fraction (sup) were detected by using anti-Mep45 antiserum. (B) Effect of the PG-cadaverine on the interaction between Mep45 and peptidoglycan. Relative amounts of peptidoglycan-associated Mep45 were quantified densitometrically following immunoblotting. Solubilized Mep45 (5 μg) was incubated with 0.4, 1.2, or 2.0 mg of purified peptidoglycan from (i) wild-type S. ruminantium cells (WT [closed bars]), (ii) S. ruminantium cells cultivated in the presence of 10 mM DFMO (open bars), and (iii) the E. coli lpo mutant (EC [shaded bars]). The values represent the ratio of the amount of Mep45 after binding to peptidoglycan over the initial amount (mean ± standard deviation [SD] of triplicate experiments).

FIG. 3.

Functional analysis of the SLH domain of Mep45. (A) Preparation of Mep45 with the SLH domain truncated (ΔSLH-Mep45). Lane 1, trypsin-digested crude envelopes; lane 2, purified ΔSLH-Mep45 preparation. The arrowhead indicates the ΔSLH-Mep45 bands. (B) Peptidoglycan-binding assay of ΔSLH-Mep45. Solubilized ΔSLH-Mep45 (100 μg/ml) was incubated with 0.4, 1.2, or 2.0 mg of peptidoglycan from wild-type S. ruminantium cells. (C) Preparation of recombinant SLH domain fused with glutathione S-transferase (G-SLH). Lane 1, cell extracts of E. coli BL21(DE3) expressing G-SLH; lane 2, purified G-SLH. The arrowhead indicates the G-SLH band. (D) Peptidoglycan-binding assay of G-SLH of (i) wild-type S. ruminantium cells (WT [closed bars]), (ii) S. ruminantium cells cultured in the presence of 10 mM DFMO (open bars), and (iii) the E. coli lpo mutant (EC [shaded bars]). In panels B and D, anti-glutathione S-transferase antiserum was used to detect G-SLH and relative amounts of peptidoglycan-associated G-SLH were represented as the amount relative to initial amount of G-SLH added to the reaction mixture (mean ± SD of triplicate experiments).

FIG. 4.

The structural constituent of the peptidoglycan required for interaction with SLH domain. The assay for G-SLH binding to peptidoglycan was performed in the presence of free cadaverine (A) or peptide A or the PG fragment (B). The free cadaverine, peptide A, or PG fragment was exogenously added to the reaction mixture containing G-SLH and 1 mg of peptidoglycan from wild-type S. ruminantium, and the relative amount of G-SLH bound to the peptidoglycan was measured.

FIG. 5.

Electron micrograph of cell surface structures of wild-type S. ruminantium cells (A) and PG-cadaverine-deficient cells (B). Arrowheads indicate the sites where the outer membrane (OM) and peptidoglycan are connected. IM, inner membrane. The numbers of the connection sites per 400 nm of cell surface were counted from 40 cells each. The difference between the wild-type cell and PG-cadaverine-deficient cell was significant (P < 0.05 according to Student's t test). Bar, 100 nm.

FIG. 6.

Structural analysis of Mep45. (A) SDS-PAGE analysis of Mep45 under heated or unheated conditions. 45 k and 37 k, 45 kDa and 37 kDa, respectively. (B) Subunit structure of Mep45 determined by cross-linking experiment using glutaraldehyde or BS³. Mep45 oligomers were detected by immunoblotting using anti-Mep45 antiserum.

FIG. 7.

Proposed model of Mep45 in the outer membrane. Shown are PRED-TMBB program predictions of transmembrane strands and the topology of Mep45. Transmembrane strands are shaded. The amino acid sequence with predicted high propensity to form an α-helical coiled-coil structure is underlined. Aromatic residues located near the ends of the transmembrane strand contacting the membrane surfaces are boxed. Proteinase K- and trypsin-sensitive cleavage sites are indicated as open and closed triangles, respectively.