Crystal Structure of the C2 Domain of Lactadherin and Computational Docking of Phosphatidyl-L-Serine.

Lactadherin is a phosphatidyl-L-serine (Ptd-L-Ser)-binding protein that decorates the membranes of milk fat globules. Lactadherin is also secreted by macrophages, engaging Ptd-L-Ser on the membranes of apoptotic cells and hastening clearance by the macrophages. Sequence homology between the lectin-like C1 and C2 domains of lactadherin and those of factor VIII and factor V correlates with the capacity of lactadherin to compete for sites on Ptd-L-Ser-containing membranes and function as a potent anticoagulant. The C2 domain of lactadherin contains a major Ptd-L-Ser binding motif. Solution studies have shown that binding of the phospho-L-serine head group exhibits stereospecificity, consistent with the existence of one or more specific binding sites on the protein. Neither the three-dimensional structure of the lactadherin C2 domain nor the identity of specific residues that interact with Ptd-L-Ser have been determined. We have solved the crystal structure of the bovine lactadherin C2 domain (Lact-C2) at 1.67 Å resolution. The β-barrel protein core, composed of eight anti-parallel strands, is homologous with the previously published structures of the factor VIII C2 domain (fVIII-C2) and factor V C2 domain (fV-C2). The loops on the lower surface of Lact-C2 display four water-exposed hydrophobic amino acids, reminiscent of the membrane-interacting residues of fVIII-C2 and fV-C2. These hydrophobic amino acids, Trp26, Leu28, Phe31, and Phe81, form a diamond-shaped hydrophobic patch on the lower surface of Lact-C2. This hydrophobic patch is well positioned to interact extensively with the phospholipid membrane. The longest loop, predicted to participate in membrane binding, has a W-like shape that contrasts with the simpler shapes adopted by the corresponding loops of fV-C2 and fVIII-C2. Furthermore, in Lact-C2 glycine residues within putative membrane-interacting loops appear likely to provide conformational flexibility that may facilitate Lact-C2 binding to the membrane. To investigate potential Ptd-L-Ser binding sites on the protein, computational docking studies were performed, utilizing the program AutoDock. A Ptd-L-Ser analog with acetyl moieties replacing both acyl chains was utilized as the docking ligand. This approach avoided spurious identification of non-specific binding sites for acyl chains, which would naturally be immersed in a membrane bilayer. We identified two candidate Ptd-L-Ser binding sites that could engage the hydrophilic head group of Ptd-L-Ser in a stereoselective manner. The binding pocket for the better of the two solutions is not homologous with the Ptd-L-Ser binding sites previously proposed for fVIII-C2 or fV-C2. These results provide the basis for a testable hypothesis rationalizing the efficient phospholipid binding of lactadherin vs. factor VIII and factor V and the mechanism for Ptd-L-Ser stereoselectivity.