Glucan Binding Protein C of Streptococcus mutans Mediates both Sucrose-Independent and Sucrose-Dependent Adherence

Abstract

The high-resolution structure of glucan binding protein C (GbpC) at 1.14 Å, a sucrose-dependent virulence factor of the dental caries pathogen Streptococcus mutans, has been determined. GbpC shares not only structural similarities with the V regions of AgI/II and SspB but also functional adherence to salivary agglutinin (SAG) and its scavenger receptor cysteine-rich domains (SRCRs). This is not only a newly identified function for GbpC but also an additional fail-safe binding mechanism for S. mutans Despite the structural similarities with S. mutans antigen I/II (AgI/II) and SspB of Streptococcus gordonii, GbpC remains unique among these surface proteins in its propensity to adhere to dextran/glucans. The complex crystal structure of GbpC with dextrose (β-d-glucose; Protein Data Bank ligand BGC) highlights exclusive structural features that facilitate this interaction with dextran. Targeted deletion mutant studies on GbpC's divergent loop region in the vicinity of a highly conserved calcium binding site confirm its role in biofilm formation. Finally, we present a model for adherence to dextran. The structure of GbpC highlights how artfully microbes have engineered the lectin-like folds to broaden their functional adherence repertoire.

Keywords: Streptococcus gordonii; Streptococcus mutans; antigen I/II; fibrillar; glucan binding protein; lectin-like fold; microbial adherence; polyproline type II helix; salivary agglutinin; sucrose-dependent adhesion; sucrose-independent adhesion.

FIG 1

Primary structure features of GbpC and the extent of the recombinant fragments used in this study. The equivalent of AgI/II's V region in GbpC is highlighted.

FIG 2

(A) Crystal structure of GbpC^111–522 with the lectin-like fold; the calcium binding site is highlighted. Three residues, Ser347, Glu360, and Asn349, form bipyramidal geometrical coordination for the calcium binding site. The loop region of residues 410 to 418 is highlighted in dark blue, and the potential movement of this loop region based on the B factors (see Table S2 in the supplemental material) is denoted by the dotted red line. (B) A closer examination of the A-P interaction on GbpC and AgI/II shows that there is a register shift in the interactions of the AxYxAx(L/V) motif on the A region and the PxxP motif on the P region that we had described earlier (3).

FIG 3

The figure shows the structural superposition of the V regions of AgI/II (light pink), SspB (light green), and GbpC (light blue) and their structure-based sequence alignment with alpha-helices (purple cylinders), strands (beige arrows), and the PPII helix (saffron). The residues of the conserved calcium-binding site are highlighted (red dots on the sequence). The loop regions that hover over the calcium-binding site which can be imagined to emanate from two sides of a palm are highlighted in dark pink for AgI/II (residues 584 to 590) and dark blue for GbpC (residues 410 to 418; shaded red boxes in sequence). SspB does not display either of these two loop regions.

FIG 4

Interaction of GbpC with 1,000-, 5,000-, 9,400-, 38,800-, and 70,800-Da dextran was determined using an Auto ITC 200. GbpC adheres to dextran with micromolar affinity. The stoichiometry reported here is the ratio of GbpC to dextran binding. The shaded portions of the table are values for model fits in the reverse mode in Origin, which displays the various parameters calculated for the reversed binding. In preliminary reversed ITC experiments with dextran in the cell and GbpC in the syringe, we observed very similar values (see Fig. S4 in the supplemental material). In similar experiments with AgI/II and SspB, dextran did not show any specific interaction (Fig. S4). K_d, dissociation constant; N, stoichiometry.

FIG 5

The effect of dextran on the adherence of GbpC to immobilized SRCR₁ (iSRCR₁) was analyzed using surface plasmon resonance. These results indicate that there is a significant dose-dependent reduction in the amount of GbpC that binds to iSRCR₁. While the affinities remain unchanged (see Table S1 in the supplemental material), the presence of dextran in close proximity to the SRCR binding surface could produce steric hindrance, thus precluding the adherence of GbpC. R_max, maximum analyte binding capacity. ***, P < 0.001.

FIG 6

Competition assays between GbpC and AgI/II against immobilized SAG. SAG was immobilized to a CM5 chip, and either buffer (no inhibitor) or a competitive inhibitor (AgI/II or GbpC) was passed to saturate SAG receptor sites. Afterwards, the competing protein (labeled at the bottom) was passed through, and the maximal binding was measured. GbpC inhibited the binding of AgI/II (left) by 25.1% ± 10%, whereas AgI/II inhibited GbpC binding (right) by 60.3 ± 6.7%. All experiments were carried out in duplicate; error bars indicate means ± standard deviations. In control experiments, where either AgI/II or GbpC was immobilized, neither one displayed any measurable adherence to the other (data not shown). RU, response units. *, P < 0.05.

FIG 7

(A) Stereo diagram of the interaction of BGC1 and BGC2 in the vicinity of the Ca²⁺ (yellow) binding site. The superposed native structure and its displaced water molecules are shown in magenta. (B) Surface plots highlighting the binding pocket, where for GbpC the loop region of residues 410 to 418 is highlighted in magenta. For AgI/II and SspB these binding pockets are very different and would pose steric hindrances for the glucose molecules, thus explaining the propensity for GbpC to specifically adhere to dextran.

FIG 8

Based on positions of BGC1 and BGC2 in the complex crystal structure, we have modeled dextran (6 glucose units) into the binding site. (A) Interactions of this modeled dextran with GbpC. (B) The loop region of residues 410 to 418, highlighted in purple, hovers over the Ca²⁺ site, which hosts the dextran (glucan) binding site. This loop region is highly flexible in the structure and could open and close to accommodate the glucose molecules thus locking into place. (C) The model for the interaction of GbpC with dextran (glucose lattice) based on ITC and crystal structure.

FIG 9

Effects of the loop region of GbpC and AgI/II on biofilm formation of S. mutans. Wild-type S. mutans and the loop region deletion mutants (GbpC^410–418Δ and AgI/II^584–590Δ) (left) and their complemented strains (right) were evaluated for biofilm formation using crystal violet staining. The results show that the loop region of residues 410 to 418 of GbpC plays an important role in biofilm formation. Biofilm formation defects caused by GbpC and AgI/II deficiency can be recovered to wild-type levels by complementation.

FIG 10

Effects of GbpC and AgI/II deficiency on dextran dependent bacterial aggregation (DDAG) of S. mutans. Wild type (WT) S. mutans and various gbpC and AgI/II gene (pac) mutant variants were evaluated for the dextran-dependent bacterial aggregation by monitoring bacteria remaining in suspension at OD_{600 nm}. The higher the OD reading, the lower was the bacterial aggregation. AgI/II appears to have a more profound effect on aggregation than GbpC. Particularly, the result of deleting the loop region of residues 584 to 590 of AgI/II (AgI/II^584–590Δ) indicated that it is involved in DDAG, whereas the loop region of residues 410 to 418 of GbpC is not. The deletion of the N-terminal region (residues 1 to 224) of GbpC resulted in a significant reduction in bacterial aggregation, while the double mutant AgI/II and GbpC^1–224Δ had the largest effect on bacterial aggregation.