The major outer sheath protein (Msp) of Treponema denticola has a bipartite domain architecture and exists as periplasmic and outer membrane-spanning conformers

Abstract

The major outer sheath protein (Msp) is a primary virulence determinant in Treponema denticola, as well as the parental ortholog for the Treponema pallidum repeat (Tpr) family in the syphilis spirochete. The Conserved Domain Database (CDD) server revealed that Msp contains two conserved domains, major outer sheath protein(N) (MOSP(N)) and MOSP(C), spanning residues 77 to 286 and 332 to 543, respectively, within the N- and C-terminal regions of the protein. Circular dichroism (CD) spectroscopy, Triton X-114 (TX-114) phase partitioning, and liposome incorporation demonstrated that full-length, recombinant Msp (Msp(Fl)) and a recombinant protein containing MOSP(C), but not MOSP(N), form amphiphilic, β-sheet-rich structures with channel-forming activity. Immunofluorescence analysis of intact T. denticola revealed that only MOSP(C) contains surface-exposed epitopes. Data obtained using proteinase K accessibility, TX-114 phase partitioning, and cell fractionation revealed that Msp exists as distinct OM-integrated and periplasmic trimers. Msp(Fl) folded in Tris buffer contained slightly less β-sheet structure than detergent-folded Msp(Fl); both forms, however, partitioned into the TX-114 detergent-enriched phase. CDD analysis of the nine Tpr paralogs predicted to be outer membrane proteins (OMPs) revealed that seven have an Msp-like bipartite structure; phylogenetic analysis revealed that the MOSP(N) and MOSP(C) domains of Msp are most closely related to those of TprK. Based upon our collective results, we propose a model whereby a newly exported, partially folded intermediate can be either processed for OM insertion by the β-barrel assembly machinery (BAM) or remain periplasmic, ultimately forming a stable, water-soluble trimer. Extrapolated to T. pallidum, our model enables us to explain how individual Tprs can localize to either the periplasmic (e.g., TprK) or OM (e.g., TprC) compartments.

Fig 1

(A) Domain structure of Msp predicted by the CDD server (54). (B) CD spectra of Msp^Fl, E. coli OmpG, Msp^N, and Msp^C. Msp^Fl (5 μM) and Msp^C (5 μM) were folded in DDM buffer; OmpG (5 μM) was folded in 50 mM NaCl, 10 mM Tris (pH 7.5), and 0.2% n-octyl-β-d-glucopyranoside, and Msp^N (5 μM) was folded in 50 mM Tris (pH 7.5) 50 mM NaCl.

Fig 2

Msp^Fl and Msp^C, but not Msp^N, are amphiphilic and possess channel-forming activity. (A) A total of 10 μg each of Msp^Fl, Msp^N, and Msp^C was phase partitioned in TX-114 and stained with GelCode Blue (Thermo Scientific) following SDS-PAGE. Molecular mass standards (kDa) are indicated on the right. (B) Liposomes were reconstituted with 10 μg each of Msp^Fl, Msp^N, and Msp^C followed by sucrose density gradient ultracentrifugation. Following SDS-PAGE, fractions were subjected to immunoblotting with antisera directed against Msp^Fl. Lanes: top fractions (TF) contain liposome-incorporated material, whereas middle and bottom fractions (MF and BF, respectively) contain unincorporated material. Molecular mass standards (kDa) are indicated on the right. (C) Quenching of Tb(DPA)₃³⁻ encapsulated in LUVs following incubation (30 min) with 100 nM E. coli OmpF, Msp^Fl, Msp^N, and Msp^C in 50 mM Tris (pH 7.5) and 100 mM NaCl supplemented with 5 mM EDTA. Each bar represents the mean ± standard error of the mean (SEM) from three independent experiments. P values of <0.05 (Student's t test) were considered significant.

Fig 3

Msp is expressed in extremely high copy number in T. denticola, but only Msp^C is surface exposed. (A) Quantitative immunoblot analysis of Msp expressed in T. denticola. T. denticola lysates (2.0 × 10⁶ organisms) were immunoblotted with anti-Msp^Fl antiserum; a standard curve generated from densitometric values obtained for graded amounts of Msp^Fl was used to determine the copy number of Msp per cell. Molecular mass standards (kDa) are indicated on the left. (B) Intact T. denticola or organisms treated with TX-100 (0.05%) were encapsulated in gel microdroplets and probed with rat antisera against Msp^N, Msp^C, or anti-T. denticola flagella (45). Antibody binding was detected with goat anti-rat Alexa Fluor 488 (green) or Alexa Fluor 594 (red) conjugates.

Fig 4

Limited accessibility of Msp to surface proteolysis in T. denticola. Immunoblot analysis of Msp, detected using anti-Msp^N (A) or anti-Msp^C (B) antiserum in motile treponemes (1.0 × 10⁸ organisms/lane) treated for 1 h with graded concentrations of proteinase K (PK). Molecular mass standards (kDa) are shown on the left. The arrow in panel A designates the ∼25-kDa degradation product reactive with anti-Msp^N antiserum. (C) PK accessibility of Msp and two periplasmic controls (TroA and FlaA) in intact and detergent lysozyme-treated organisms incubated with (+) or without (−) 50 μg/ml of PK. Each lane represents 1.0 × 10⁶ T. denticola immunoblotted with rat antisera directed against Msp^Fl, TroA, or flagella. Molecular mass standards (kDa) are indicated on the left of panels A and B.

Fig 5

T. denticola contains both hydrophilic and amphiphilic forms of Msp, but only the latter is PK accessible in intact organisms. Motile T. denticola without (A) or with (B and C) PK treatment (50 μg/ml, 1 h) was subjected to TX-114 phase partitioning followed by SDS-PAGE and immunoblotting with antibodies to Msp^Fl, Msp^N, and Msp^C (panels A, B, and C, respectively). The arrowhead in panel B indicates an ∼25-kDa degradation product detected only with anti-Msp^N antiserum. Lanes: whole cells (WC), TX-114 insoluble material (Ins), aqueous (Aq), and detergent-enriched (Det) phases. Molecular mass standards (kDa) are indicated on the left of each panel.

Fig 6

T. denticola membrane and soluble fractions contain trimeric forms of Msp. (A) Membrane and soluble fractions of T. denticola (MF and SF, respectively) were phase partitioned with TX-114 followed by SDS-PAGE and immunoblot analysis using anti-Msp^Fl antiserum. Molecular mass standards (kDa) are indicated on the right. (B) Membrane and soluble fractions were separated by SDS-PAGE without (−) and with (+) boiling followed by immunoblot analysis with anti-Msp^Fl antiserum. Molecular mass standards (kDa) are indicated on the left. (C) BN-PAGE showing trimer formation by Msp^Fl. Molecular mass standards (kDa) are indicated on the left.

Fig 7

Characterization of a water-soluble form of Msp^Fl. (A) Far-UV CD spectra of Msp^Fl folded in DDM (5 μM) and Tris buffer (3 μM). (B) TX-114 phase partitioning of water- and detergent-soluble forms (WS and DS, respectively) of Msp^Fl. Molecular mass standards (kDa) are indicated on the right.

Fig 8

Predicted architectural and phylogenetic relationships between Msp and Tpr proteins within the OM. (A) CDD analysis of the nine Tpr proteins predicted to be OMPs (35). (B) Phylogenetic relationships between Msp and the eight Tprs. The phylogenetic analysis was carried out using the ClustalX (46) program using 10,000 bootstrap trials. The phylogenetic tree was viewed with the help of TreeView32 software. Bootstrap values (%) are indicated for the major branch points in the tree.

Fig 9

Proposed two-pathway model for generation of amphiphilic, OM-inserted and soluble, periplasmic Msp conformers. Msp precursor is exported across the cytoplasmic membrane (CM) by the Sec translocon. Once within the periplasm, the unfolded precursor acquires some secondary structure, becoming the “intermediate form.” The intermediate form can either be chaperoned via Skp (Tde2602) to the POTRA arm of BamA (Tde2601) for OM assembly (25) and trimerization (OMP conformer) or fold into a soluble trimeric periplasmic conformer. We hypothesize that in the intermediate form, a stretch of amino acids at the extreme C terminus of MOSP^C (i.e., the signature sequence, shown in orange) containing the BamA recognition signal is unfolded but can become a β-strand once in contact with the POTRA1 domain of BamA. The resultant block to entry into the OMP assembly pathway drives the intermediate form toward the alternative, periplasmic pathway. The OMP conformer is depicted as trimerizing via the MOSP^C domains based on our prior analysis of TprC (40). Presently, we have no data as to the regions of Msp that mediate trimerization of the periplasmic conformer. The components of the Bam complex in T. denticola, other than BamA, are unidentified (25) and, therefore, not shown.