A conserved domain of previously unknown function in Gap1 mediates protein-protein interaction and is required for biogenesis of a serine-rich streptococcal adhesin

Abstract

Fap1-like serine-rich proteins are a new family of bacterial adhesins found in a variety of streptococci and staphylococci that have been implicated in bacterial pathogenesis. A gene cluster encoding glycosyltransferases and accessory Sec components is required for Fap1 glycosylation and biogenesis in Streptococcus parasanguinis. Here we report that the glycosylation-associated protein, Gap1, contributes to glycosylation and biogenesis of Fap1 by interacting with another glycosylation-associated protein, Gap3. Gap1 shares structural homology with glycosyltransferases. The gap1 mutant, like the gap3 mutant, produced an aberrantly glycosylated Fap1 precursor and failed to produce mature Fap1, suggesting that Gap1 and Gap3 might function in concert in the Fap1 glycosylation and biogenesis. Indeed, Gap1 interacted with Gap3 in vitro and in vivo. A Gap1 N-terminal motif, within a highly conserved domain of unknown function (DUF1975) identified in many bacterial glycosyltransferases, was required for the Gap1-Gap3 interaction. Deletion of one, four and nine amino acids within the conserved motif gradually inhibited the Gap1-Gap3 interaction and diminished production of mature Fap1 and concurrently increased production of the Fap1 precursor. Consequently, bacterial adhesion to an in vitro tooth model was also reduced. These data demonstrate that the Gap1-Gap3 interaction is required for Fap1 biogenesis and Fap1-dependent bacterial adhesion.

Fig. 1. Fap1 expression by gap1 and gap3 mutants of S. parasanguinis

Protein extracts prepared from wild type S. parasanguinis (lane 1), fap1 (lane 2), gap1 (lane 3), gap3 (lane 5), and gap1gap3 double mutant (lane 7) as well as their complemented strains, gap1−/+gap1 (lane 4), gap3-/+gap3 (lane 6) and gap1gap3−/+gap1gap3 (lane 8) were subjected to Western blotting analyses using peptide-specific mAbE42 (A), glycan-specific mAbD10 (B) and F51 (C) antibodies.

Fig. 2. Subcellular distribution of Fap1

Comparison of the subcellular distribution of Fap1 in wild type bacteria and the gap1 mutant (A). Subcellular fractions including culture supernatants (lanes 1 and 4), cell wall (lanes 2 and 5) and cytosolic fractions (lanes 3 and 6), were prepared from wild type bacteria (lanes 1–3) and the gap1 mutant (lanes 4–6) and then subjected to Western blotting analysis to determine the Fap1 distribution (upper panel) using mAbE42 and the presence of cytosolic-specific protein Tpx (bottom panel) using anti-Tpx antibody. Comparison of the subcellular distribution of Fap1 in a gtf mutant and a gap1gtf double mutant (B). Subcellular fractions, including cell wall (lanes 1 and 2) and culture supernatants (lanes 3 and 4) were prepared from the gtf1 mutant (lanes 1 and 3) and the gap1 gtf1 double mutant (lanes 2 and 4) and then subjected to Western blotting analyses to determine the Fap1 distribution (upper panel) using mAbE42 and to determine FimA expression (lower panel) using anti-FimA antibody.

Fig. 3. Gap1 and Gap3 interaction determined by yeast two-hybrid (Y2H) experiments (A) and in vitro GST pull-down assays (B)

A. Y2H assay. Yeast strain AH109 cells were transformed with different combinations of plasmid pairs: pBD-p53 and pAD-T (positive control) (1); pBD-Lam and pAD-T (negative control) (2); pBD-Gap3 and pAD-Gap1 (3); pBD-Gap1 and pAD-Gap3 (4); pBD-Gap1 and pAD (5); pAD-Gap1 and pBD (6); pBD-Gap3 and pAD (7) and pAD-Gap3 and pBD (8). The transformed cells were plated onto SD-LT medium to select transformants carrying plasmid pairs (left panel) and onto SD-LTHA medium to select plasmid pairs carrying specific interacting partners (right panel). B. In vitro GST pull-down assays. Equal amounts of purified GST, GST-Gap3 and GST-Gap1 fusion proteins immobilized on glutathione Sepharose 4B beads were incubated separately with in vitro translated c-Myc-Gap1 (left panel) or c-Myc-Gap3 (right panel). The resulting protein complexes were analyzed by Western blot using anti-c-Myc antibody. Inputs represent in vitro translated c-Myc-Gap1 (left panel) and c-Myc-Gap3 (right panel) respectively.

Fig. 4. Mapping Gap1 and Gap3 interaction domains in Gap1

A series of Gap1 truncated mutants were constructed in pBD (A). Full-length Gap1 (1) and its truncated derivatives (2–13) were co-transformed with Gap3-pAD into yeast AH109, plated on SD-LT medium, and resulting colonies streaked onto SD-LTHA medium (B) to select transformants carrying specific interacting partners. Co-transformants with pBD-p53/pAD-T and pBD-Lam/pAD-T were used as a positive (+) and a negative (−) control, respectively. Mapping Gap1 interaction domains by in vitro GST pull-down assays (C). Equal amounts of GST fusion proteins immobilized on glutathione Sepharose 4B beads were separately incubated with in vitro translated c-Myc-Gap3. Interactions of GST, GST-Gap1 and GST-Gap1 truncated fusion proteins with c-Myc-Gap3 were evaluated by Western blotting analyses using anti-c-Myc antibody. Input represents in vitro translated c-Myc-Gap3. Alignment of amino acid sequences from a minimal interaction motif in Gap1 and its homologs (D). The defined minimal interaction motif in Gap1 of S. parasanguinis was compared with similar regions from the Gap1 homologs of S. gordonii, S. agalactiae, S. pneumoniae and S. aureus. The N-terminal region of Gap1 harbors a conserved domain of unknown function (DUF1975) which is also identified in a variety of prokaryotic α-glucosyltransferases.

Fig. 4. Mapping Gap1 and Gap3 interaction domains in Gap1

A series of Gap1 truncated mutants were constructed in pBD (A). Full-length Gap1 (1) and its truncated derivatives (2–13) were co-transformed with Gap3-pAD into yeast AH109, plated on SD-LT medium, and resulting colonies streaked onto SD-LTHA medium (B) to select transformants carrying specific interacting partners. Co-transformants with pBD-p53/pAD-T and pBD-Lam/pAD-T were used as a positive (+) and a negative (−) control, respectively. Mapping Gap1 interaction domains by in vitro GST pull-down assays (C). Equal amounts of GST fusion proteins immobilized on glutathione Sepharose 4B beads were separately incubated with in vitro translated c-Myc-Gap3. Interactions of GST, GST-Gap1 and GST-Gap1 truncated fusion proteins with c-Myc-Gap3 were evaluated by Western blotting analyses using anti-c-Myc antibody. Input represents in vitro translated c-Myc-Gap3. Alignment of amino acid sequences from a minimal interaction motif in Gap1 and its homologs (D). The defined minimal interaction motif in Gap1 of S. parasanguinis was compared with similar regions from the Gap1 homologs of S. gordonii, S. agalactiae, S. pneumoniae and S. aureus. The N-terminal region of Gap1 harbors a conserved domain of unknown function (DUF1975) which is also identified in a variety of prokaryotic α-glucosyltransferases.

Fig. 5. Identification of in vivo interaction between Gap1 and Gap3 in S. parasanguinis by co-immunoprecipitation assays

Gap1 (Δ182–190)-GFP/pVPT and Gap1-GFP/pVPT were transformed into wild type S. parasanguinis to generate GFP-tagged strains Gap1 (Δ182–190)-GFP and Gap1-GFP. Cell lysates prepared from these transformed strains were subjected to immunoprecipitation with anti-GFP antibody. Cell lysates from the gap3 mutant (lane 1), Gap1 (Δ182–190)-GFP (lane 2) and Gap1-GFP (lane 3) as well as the corresponding immunoprecipitated samples (lanes 4 and 5) were subjected to Western blotting analysis using Gap3- (upper panel) and GFP-specific (lower panel) antibody.

Fig. 6. Effect of Gap1-Gap3 interaction on Fap1 glycosylation

Full-length Gap1 and a series of truncated Gap1 constructs cloned in E. coli-Streptococcal shuttle vector pVPT-GFP were used to transform a gap1 insertional mutant. Cell lysates prepared from wild type bacteria (lane 1), fap1 (lane 2), gap1 (lane 3) and the gap1 mutant transformed with the full-length Gap1 (lane 4) and with the truncated Gap1 constructs, Gap1Δ182 (lane 5), Gap1Δ182–185 (lane 6), Gap1Δ182–190 (lane 7) and Gap1Δ182–200 (lane 8) were subjected to Western blot analyses using Fap1 peptide-specific mAb E42 (A), glycan-specific mAb F51 (B) and anti-GFP antibody (C).

Fig. 7. Effect of the Gap1 interaction domain on bacterial adhesion

Wild type S. parasanguinis FW213, fap1, gap1, the gap1 complemented strain and three Gap1 interaction motif deletion mutants (Δ182, Δ182-185 and Δ182-190) were labeled with ³H-thymidine and incubated with SHA beads. The ability of bacteria to attach to the SHA beads was determined by the number of bacteria bound to the SHA versus the total input bacteria.