Tissue factor pathway inhibtor-2 (TFPI-2), also known as matrix serine protease inhibitor or placental protein 5, contains three Kunitz-type inhibitory domains in tandem. A variety of cells including keratinocytes, dermal fibroblasts, smooth muscle cells, syncytiotrophoblasts, synoviocytes, and endothelial cells synthesize and secrete TFPI-2 into the extracellular matrix (ECM). Kunitz domain 1 (KD1) of TFPI-2 inhibits plasmin (Ki = 3 nM), trypsin (Ki = 13 nM), and FVIIa/TF (Ki = 1640 nM). We employed crystallography and molecular modeling approaches to elucidate the basis of the specificity of KD1 for plasmin versus trypsin or FVIIa/TF. Crystals of the complex of KD1 with bovine trypsin were obtained that diffracted to 1.8 Å and belonged to the space group P212121 with unit cell parameters, a=74.11, b=77.01, and c=125.42. Each asymmetric unit contained two KD1-trypsin complexes. The structure of KD1 thus obtained was then used in conjunction with the known structures of plasmin and FVIIa/TF to model the KD1-plasmin and KD1-FVIIa complexes. KD1 contained a hydrophobic core consisting of residues Leu-9 (BPTI numbering), Tyr-11, Tyr-22, and Phe-33. In all structures, Arg-15 (P1 residue) of KD1 interacted with Asp-189 (chymotrypsin numbering) at the bottom of the specificity pocket. A hydrophobic patch involving residues Leu-17, Leu-18, Leu-19, and Leu-34 of KD1 was identified to interact with a hydrophobic patch in plasmin and trypsin but not in FVIIa/TF. This complementary hydrophobic patch in plasmin consists of Phe-37, Met-39, Phe-41, and the carbon side chains of Gln-192 and Glu-141. In trypsin, it consists of Tyr-39, Phe-41, Tyr-151 and the carbon side chain of Gln-192. Furthermore, a basic patch involving Arg-98, Arg-173 and Arg-221 in plasmin was identified to interact with an acidic patch in KD1 consisting of residues Asp-10 and Glu-39. This electrostatic interaction is absent in trypsin and in FVIIa/TF. Moreover, Tyr-46 in KD1 can make H-bonds with Lys-61 and Arg-64 in plasmin as well as with Lys-60A in FVIIa/TF; however, these interactions are absent in trypsin. Further, Arg-20 of KD1 is important for making a H-bond with Glu-60 in plasmin, with Lys-60 through a water molecule in trypsin and with Asp-60 in FVIIa/TF. Cumulatively, the crystal structure and refined modeling data confirm our previous predictions and illustrate the molecular basis for preference of KD1 to inhibit plasmin versus trypsin or FVIIa/TF. KD1 interacts with plasmin through hydrophobic and electrostatic interactions whereas the electrostatic contacts are limited in trypsin. Notably, both the electrostatic and hydrophobic interactions are sparse in FVIIa/TF. Thus, both the crystal and modeled structures validate the differential effects of mutations in KD1 involving residues surrounding the P1 site, including D10A, L17Q, R20D, and F33A, reported earlier (Chand HS, Schmidt A, Bajaj SP and Kisiel W. JBC 279, 17500–17507, 2004). Knowledge gained from such studies may help in the development of a potent and specific TFPI-2 KD1 molecule that specifically inhibits plasmin without targeting other proteases. Such a molecule could have a large pharmacologic impact specifically in preventing tumor metastasis, retinal degeneration, and degradation of collagen in the ECM.

Author notes

Corresponding author

Sign in via your Institution