Molecular design of Mycoplasma hominis Vaa adhesin

Abstract

The variable adherence-associated (Vaa) adhesin of the opportunistic human pathogen Mycoplasma hominis is a surface-exposed, membrane-associated protein involved in the attachment of the bacterium to host cells. The molecular masses of recombinant 1 and 2 cassette forms of the protein determined by a light-scattering (LS) method were 23.9 kD and 36.5 kD, respectively, and corresponded to their monomeric forms. Circular dichroism (CD) spectroscopy of the full-length forms indicated that the Vaa protein has an alpha-helical content of approximately 80%. Sequence analysis indicates the presence of coiled-coil domains in both the conserved N-terminal and antigenic variable C-terminal part of the Vaa adhesin. Experimental results obtained with recombinant proteins corresponding to the N- or C-terminal parts of the shortest one-cassette form of the protein were consistent with the hypothesis of two distinct coiled-coil regions. The one-cassette Vaa monomer appears to be an elongated protein with a axial shape ratio of 1:10. Analysis of a two-cassette Vaa type reveals a similar axial shape ratio. The results are interpreted in terms of the topological organization of the Vaa protein indicating the localization of the adherence-mediating structure.

Fig. 1.

Schematic drawing of constructs and coiled-coil regions. (A) Modular structure of native category 5 Vaa of M. hominis 5941 (Vaa₅₉₄₁) and category 3 Vaa of M. hominis 4195 (Vaa₄₁₉₅). The signal peptide, module I, is cleaved off in the mature Vaa, and the N-terminal cysteine (C27) is lipid-modified. Modules II and VI are highly conserved in all the Vaa types. The C-terminal part of Vaa₅₉₄₁, comprising modules II' and VII, is highly variable. The structure of Vaa₄₁₉₅ resembles that of Vaa₅₉₄₁ with the exception of an additional C-terminal cassette, module V. The modules V and VII are exchangeable cassettes. The regions with tryptophan encoded by UGA codons are shown with bars above the schematic drawings of Vaa₅₉₄₁ and Vaa₄₁₉₅. (B) Output from Coilscan using a window of 28 (solid line), 21 (dashed line), or 14 (dotted line) residues. The diagram shows the probability for the presence of coiled-coil regions in Vaa from M. hominis 4195. The probability of coiled coil is high throughout most of the native Vaa₄₁₉₅. Similar coiled-coil regions were observed for the corresponding part of Vaa₅₉₄₁. A threshold value of 0.5 is shown as a horizontal dashed line. (C) The full-length recombinant category 3 Vaa (rVaa3), category 5 Vaa (rVaa5), and the N- and C-terminal rVaa fragments called rVaaN and rVaaC, respectively. The rVaaN and the rVaaC/C^δ proteins contain the N-terminal and C-terminal parts of category 5 Vaa, respectively. A fusion peptide of 6.5 kD (LIC-tag + thrombin site and monoclonal antibody 35.2 epitope, see Materials and Methods) is present in the N-terminal part of the rVaa3 and rVaa5 proteins, and the LIC-tag (5 kD) is fused to the rVaaN, rVaaC, and rVaaC^δ proteins. The arrows indicate the thrombin-cleavage site for removal of LIC-tag in rVaa5 and rVaa3. The numbering corresponds to native Vaa.

Fig. 2.

Gel filtration, light scattering, and circular dichroism of rVaa. (A) Gel filtration of rVaa5^T and rVaa3^T (left and right panels, respectively) using a TSK-gel G3000SW column. The plots show single, monodisperse peaks. (B) Chromatogram for 90° light scattering (dashed line) contrasted with the refractive index (RI) signal (solid line). A peak was observed at V₀ (8.6 mL) for both proteins in the light-scattering detector. This was not protein as judged from the UV and RI detectors, but probably dust particles. A second peak observed for both detectors comigrated with the UV peak. (C, left) CD spectra of the rVaa5^T (thick line) and rVaa3 (thin line) constructs. (Right) CD spectra of rVaaN (dashed line), rVaaC (thin solid line), and the stoichiometric rVaaN:C mixture (thick solid line).

Fig. 3.

Secondary structure prediction of Vaa using the Jpred² server (Cuff et al. 1998). The N-terminal part and a multiple alignment of cassette sequences from Vaa were subjected to secondary structure prediction, and the results were superimposed on the rVaa5^T sequence. The prediction is shown in letter code (H, α-helix; E, β-sheet; C, coil) and schematically (tube, α-helix; arrow, β-sheet; line, coil). Above the schematic illustration the heptad repeats revealed by Coilscan are shown with probabilities (p). The two conserved prolines (corresponding to P59 and P89 in native Vaa) are shown in boldface. A stutter (boxed) is observed at R123, and a break in α-helicity corresponding to an indel region is observed at D162 in the rVaa5^T sequence. The fusion-peptide part of rVaa5^T is underlined. Shading and numbering is according to Figure 1 ▶.

Fig. 4.

Native and SDS-PAGE of the rVaaN, rVaaC, and rVaaN:C fragments. (A) Nondenaturing gel of the rVaaN (lane 2), rVaaC (lane 3), and a mixture of rVaaN and rVaaC (lane 4) proteins, showing a new band slightly above the position of the rVaaC band (lane 4). Lane 4 has two bands because of excess amounts of rVaaN. The BSA standard is shown in lane 1, and BSA monomer and dimer bands are also observed. (B) The dominating bands from lanes 2–4 of A were excised from the nondenaturing gel, boiled in SDS-sample buffer, and used for SDS-PAGE. The top band extracted from lane 4 of A (rVaaN:C) contained two protein subunits as revealed by SDS-PAGE (lane 1) corresponding to the rVaaN (lane 2) and rVaaC (lane 3) proteins with molecular masses of 13.5 kD and 22 kD, respectively.

Fig. 5.

Model of the one-cassette Vaa protein. The numbering of modules is according to Figure 1 ▶, and the shading is according to Figures 1 and 3 ▶ ▶. (Not to scale.)