Fig 1.
C. trachomatis serovar E (Bour) PmpD protein linear map.
The total PmpD protein length is 1530 amino acids. The protein map shows functional domains which include a secretory sequence, the extracellular β-helix passenger domain (grey), the Middle domain (aquamarine), and the transmembrane β-barrel autotransporter (purple). The FxxN and GGA(I,L,V) tetrapeptide motifs are shown in blue and yellow, respectively. The motif repeats are concentrated in the extracellular passenger domain but can also be found, less frequently, in the C-terminal domain.
Fig 2.
The Pmp tetrapeptide repeats, GGA(I,L,V) and FxxN, fit in a larger repeat.
(A) The GGA(I,L,V) and FxxN motifs tend to alternate, with GGA(I,L,V) almost always followed by FxxN. (B) GGA(I,L,V) and FxxN repeats show a regular spacing. (C) Sequence logo for the longer repeat we identified. (D) Sequence logo for PF02415.
Fig 3.
Comparison of structure predictions for PmpC serovar E.
Structure predictions using TrRosetta (A), RoseTTAFold (B), AlphaFold2 (C), and ESMFold (D). Confidence scores (pLDDT for AlphaFold and ESMFold, RMSD for TrRosetta and RoseTTAFold) were mapped to the same blue-to-red scale using the empirical formula [55, 74].
Fig 4.
The Pmps include a characteristic C-terminal transmembrane β-barrel.
(A) The β-barrel for PmpG showing six “mortise-tenon joints”, in which an aromatic side chain facing into the lumen of the barrel domain locks into a void created by the presence of a glycine residue on the neighboring β-strand. Blue: aromatic side chains. Red: glycine residues. Shown in green is a 4-stranded β-sheet formed by β-hairpins in extracellular loops L4 and L5. The β-hairpin in L5 has been shown to be important for correct folding of the passenger domain in other autotransporters. Prediction of transmembrane β-strands using TMbed (B) and visualization of the hydrophobic residues (C) indicates that only the bottom 3-4nm of the β-barrel is predicted to be embedded in the Chlamydia outer membrane.
Fig 5.
Pmp passenger domains form a regular β-helical structure.
Pmp passenger domain structural models and nomenclature illustrating the parallel β-sheets (PB) and turns (T). (A) Family portrait of Pmp passenger domain models. From left to right: PmpA, B, C, D, E, F, G, H, and I. (B) PmpD passenger domain showing 22 complete β-helix coils. (C) PmpD passenger domain cross section between amino acid residues 615–641 shows how the repeat sequence forms the core of the β-helix. Note that the nomenclature used here follows the well-established numbering by Yoder et al., 1993 and Jenkins et al., 2001 [80, 81], rather than the one proposed by [45]. The parallel β-sheet PB1 is shown in yellow, PB2 in blue, PB3 in red. The GGA(I,L,V) and FxxN motifs are bolded—in this case we have GGAL at the transition from T3 to PB1, and FSRN at the transition from PB3 to T3. (D) π-stacking interactions between the conserved phenylalanines may help stabilize the core of the β-helix.
Fig 6.
The Pmp passenger domains are connected to the transmembrane β-barrel by a flexible hinge.
Coarse-grained molecular dynamics simulation of PmpE with the β-barrel domain embedded in a POPC bilayer shows that the passenger domain is connected by a flexible hinge between the β-barrel and the Middle domain.
Fig 7.
Pmp cleavage sites for serovar L2 Pmp proteins mapped to the serovar E passenger domain structures.
Pmp cleavage sites from serovar L2 were mapped by BLAST to the homologous locations on the serovar E PmPs, for purposes of illustrating their location in the side loops only. Note that these cleavage sites may not be present or active in the serovar E Pmps.
Fig 8.
B-cell and T-cell epitopes on Pmp structure.
(A) Top view of the PmpC and PmpD passenger domains, with experimentally validated B-cell epitopes. (B) Computationally predicted B-cell epitopes based on structure, using Discotope. From left to right: Pmp A, B, C, D, E, F, G, H, and I. (C) Computationally predicted MHC-II T-cell epitopes based on structure. (D) The number of MHC-II T-cell epitopes differs considerably across Pmps and HLA subtypes. Note that DRB3*02:02 is overrepresented, and some of the smaller Pmps have significantly more MHC-II epitopes than the larger ones, as indicated by the red highlighting. (E) The core binding motif for MHC-II allele HLA-DRB3*02:02 overlaps with the FxxN repeat of the β-helix.
Fig 9.
Proline-rich regions in the Pmp passenger domains may be involved in adhesion.
Proline-rich regions are located primarily in the side loops and the Middle domain. Sites with at least four prolines separated by 0–4 residues are highlighted in red and labeled with the sequence.
Fig 10.
Putative membrane-penetrating amino acids, predicted by DREAMM.
Predicted membrane-penetrating amino acids (highlighted in red) are located primarily in the side loops, and at the N terminal, as well as the transmembrane β-barrel (not shown).
Fig 11.
Areas of high sequence variation between serovars are located primarily in the side loops.
Highlighted in red are the amino acids with less than 70% sequence conservation based on a multiple sequence alignment across the C. trachomatis serovars by MAFFT.
Fig 12.
(A) The predicted oligomer structure for Pmp21-D forms an extended β-helix, joined N-terminal to N-terminal and C-terminal to C-terminal. (B) The predicted PmpA408-608:PmpD269-918 heterodimer structure for also forms an extended β-helix, with the N-terminal of PmpA408-608 meeting up with the C-terminal of PmpD269-918. (C) the predicted PmpE19-459 homodimer structure shows an antiparallel configuration. All structures are shown in Rainbow coloring, with N-terminal in blue, C-terminal in red.