The Chlamydia trachomatis PmpD adhesin forms higher order structures through disulphide-mediated covalent interactions

Abstract

Chlamydia trachomatis (Ct) is the most common sexually transmitted bacterial pathogen, and the leading cause of infectious blindness worldwide. We have recently shown that immunization with the highly conserved antigenic passenger domain of recombinant Ct polymorphic membrane protein D (rPmpD) is protective in the mouse model of Ct genital tract infection, and previously, that ocular anti-rPmpD antibodies are elicited following vaccination. However, the mechanisms governing the assembly and structure-function relationship of PmpD are unknown. Here, we provide a biophysical analysis of this immunogenic 65 kDa passenger domain fragment of PmpD. Using differential cysteine labeling coupled with LC-MS/MS analysis, we show that widespread intra- and intermolecular disulphide interactions play important roles in the preservation of native monomeric secondary structure and the formation of higher-order oligomers. While it has been proposed that FxxN and GGA(I, L,V) repeat motifs in the Pmp21 ortholog in Chlamydia pneumoniae mediate self-interaction, no such role has previously been identified for cysteine residues in chlamydial Pmps. Further characterisation reveals that oligomeric proteoforms and rPmpD monomers adopt β-sheet folds, consistent with previously described Gram-negative bacterial type V secretion systems (T5SSs). We also highlight adhesin-like properties of rPmpD, showing that both soluble rPmpD and anti-rPmpD serum from immunized mice abrogate binding of rPmpD-coated beads to mammalian cells in a dose-dependent fashion. Hence, our study provides further evidence that chlamydial Pmps may function as adhesins, while elucidating yet another important mechanism of self-association of bacterial T5SS virulence factors that may be unique to the Chlamydiaceae.

Fig 1. Chlamydia trachomatis polymorphic membrane protein D (PmpD) schematic and elution profile.

(A) Ct PmpD displays the archetypal characteristic three-domain structure of Gram-negative bacterial type V autotransporters. Signal sequence (pink), passenger domain (blue) and β-barrel autotransporter (green) are depicted with the corresponding amino acid positions. The 65 kDa passenger domain fragment used in this study is depicted below, as detected by Kiselev et al. in infected cells [13]. (B) Soluble 65 kDa rPmpD elutes as a mixture of oligomeric species (‘O’) with substantially lower concentrations of soluble monomeric protein (‘M’). SDS-PAGE shows the chemical purity of the eluate (inset), and is representative of each collected fraction. Fractions B7-B3 (within dashed red lines) were pooled, and comprise the ‘oligomeric’ fraction used in all downstream experiments. The flow rate used was 0.3 ml/min, and the elution profile is representative of independent replicate purifications (n = 5). (C) rPmpD partitions exclusively in to the aqueous fraction following phase separation, indicative of predominantly hydrophilic properties of the passenger domain.

Fig 2. rPmpD monomeric and oligomeric forms show differing polydispersity and are comprised of β-sheet secondary structure.

rPmpD fractions were analysed using dynamic light scattering and circular dichroism. The polydispersity index is displayed above the frequency histograms of (A) oligomeric and (C) monomeric fractions. Regularization histograms showing % mass versus hydrodynamic radius (nm) for (B) oligomeric and (D) monomeric rPmpD fractions are also displayed, indicating the absence of larger molecular weight proteoforms in monomeric rPmpD following size exclusion chromatography. Soluble oligomeric (E) and monomeric (F) fractions of rPmpD were analysed using circular dichroism. Spectra for oligomeric (Peak ‘O’) and monomeric (Peak ‘M’) rPmpD fractions are indicative of predominantly β-sheet secondary structure interspersed with random coil elements, similar to the Pmp21 ortholog in C.pneumoniae [18].

Fig 3. Inter- and intramolecular disulphide bonding critically influence the biophysical properties of rPmpD.

(A) Oligomeric (Lanes 1–3) and monomeric (Lanes 4–6) rPmpD samples were incubated with loading buffer in the absence or presence of 0.5 M β-mercaptoethanol (β-ME). Native gels were also run with (B) oligomeric or (C) monomeric rPmpD samples incubated in the absence of SDS with differing concentrations of three reducing agents– 10 mM or 5 mM DTT (Lanes 1–2), 25 mM or 2.5 mM TCEP (Lanes 3–4) and 0.5M or 50 mM β-ME (Lanes 5–6). Non-reduced sample was run as a control (Lane 8). (D) oligomeric or (E) monomeric rPmpD were incubated in the presence of trypsin for 30 min, 5 mins, or 2 mins (Lanes 2–4, respectively). Undigested control samples were run in Lane 1 of each gel.

Fig 4. Differential alkylation of rPmpD and LC-MS/MS analysis.

PEAKS identified peptides for (A) non-reduced and (B) reduced rPmpD are indicated by blue lines are mapped onto the amino acid sequence of His₆-tagged rPmpD. Modified positions are annotated with coloured boxes and letters: c = carbamidomethylated cysteine (blue); b = methylthiolated cysteine (red); o = oxidised methionine (yellow). All cysteine-containing peptides show increased sampling of methylthiolated forms following reduction (B) suggesting that in the dominant proteoform(s) all 18 residues are likely disulfide-bonded. Methylthiolated residues at positions 124, 133, 143, 534 and 535 are identified by PEAKS exclusively post reduction (B), remaining unidentified in the non-reduced sample (A) (green boxes), suggesting that these cysteines are fastidiously disulphide-bonded, or remain non-solvent-exposed as peptides following digestion.

Fig 5. rPmpD-coated beads form adhesion clusters on the surface of mammalian cell lines.

Syrian hamster kidney (Hak) cell lines (Panels A-C) or murine fibroblast (McCoy) cell lines (Panels D-F) were incubated with naked (A,D), BSA-coated (B,E) or rPmpD-coated (C,F) beads. Large clusters of rPmpD-coated beads are observed on the surfaces of both Hak and McCoy cell lines (red arrows). BSA-coated and naked beads do not demonstrate this aggregative property on the cell surface. Images were acquired at 10x magnification (Scale bar = 20 μm).

Fig 6. Adhesion of rPmpD-coated beads to Hak cells is inhibited by soluble rPmpD and anti-rPmpD serum.

(A) Carboxylate-modified beads were coated with oligomeric or monomeric rPmpD (50 μg/ml) and compared to control (naked or 200 μg/ml BSA-coated beads) at two different bead dilutions. The number of rPmpD-coated beads adhered to 50 cells within triplicate fields of view was counted. Significantly enhanced adhesive capacity compared to naked or BSA-coated beads was observed, although no significant difference in adhesion was observed between beads coated with oligomeric and monomeric rPmpD proteoforms (B) Titration of oligomeric rPmpD (50 μg/ml-0.25 μg/ml) prior to bead coating shows a pronounced concentration-dependent reduction in adhesion. Significance values are reported relative to BSA-coated beads. (C) Soluble rPmpD competitor protein significantly reduces adhesion of rPmpD-coated beads in a dose-dependent manner, likely indicating competitive binding of putative host cell receptor(s). Significance is measured relative to pre-incubation with BSA only. (D) Heat-inactivated anti-rPmpD serum obtained from rPmpD-immunized mice abrogates adhesion of rPmpD-coated beads. Bead coating concentration of rPmpD varied from 50 μg/ml-2.5 μg/ml. Significance is measured relative to a 1:50 dilution of heat-inactivated pre-immune serum. All data are representative of 3 separate experiments, and presented as mean ± standard deviation. For all experiments, statistical significance was determined using a two-tailed t-test with GraphPad Prism 6. * = p≤0.05, ** p≤0.01, ***p≤0.001.