Computational docking of CRP onto GPVI. (A) Side view of the surface of D1, illustrating the predicted binding of CRP within the binding groove. Two distinct docking algorithms placed CRP within the groove in D1 adjacent to the C′E region (shown in yellow). (B) Model of the GPVI-CRP complex predicted by the PatchDock and FTDock algorithms. The docking prediction for CRP determined by each program is shown in red with the C′E loop labeled for reference. Three crystal structures of CRP variants with their PDB IDs are illustrated above the predicted GPVI-CRP complex to show that the spacing and orientation of the docked CRP triple helices (red) match the conserved approximately 1.4 nm (14 Å) spacing observed in the CRP crystal structures, which is itself similar to that observed in intact collagen fibers. Of note, although collagen fibers have circular cross-sections, within a blood vessel the average fiber diameter is very large (∼ 30-40 nm [∼300-400 Å]) relative to the spacing between individual collagen triple helices (∼ 1.4 nm [14 Å]), such that the surface recognized by GPVI would be approximately planar. (C) Electrostatic potential of GPVI D1 mapped onto its surface, with the PatchDock CRP model superimposed. (D) Stereoview of the residues in D1 that contact CRP in one or both of the docked models. Side chains are colored according to chemical properties: yellow are hydrophobic, green are polar, and blue are positively charged. (E) Residues implicated in collagen or CRP binding are shown in color on the GPVI surface after adding modeled N-glycans. Purple residues (V34, K41, K59, and N-glycan) have been implicated in both collagen and CRP binding,^25,26,48,49  green residues (L36, R60, R166) have been implicated in collagen binding only,^25,48  and light-blue residues (F91, R117, Y118, F120, R139, S164) were mutated with no effect on either collagen or CRP binding.²⁵