Structural and Mutational Analysis of Canonical Loop Residues In Protease Nexin 2 Involved In Factor XIa and Trypsin Inhibition.

Factor XIa (FXIa) activates FIX and is regulated by platelet-secreted protease nexin 2 (PN2) that contains a Kunitz-type protease inhibitor (KPI) domain. Trypsin is regulated by basic pancreatic trypsin inhibitor (BPTI). The primary and tertiary structures of trypsin and the catalytic domain of FXIa are highly homologous, and KPI and BPTI are nearly identical structurally. We have previously identified two loop structures (loops 1 and 2) in the KPI domain of PN2 that interact with residues in the FXIa catalytic domain. Based on the structure of the FXIa/KPI complex crystal structure, residues within loops 1 and 2 were mutated for experiments examining the inhibition of FXIa and trypsin.

Results show that the loop-1-region P1 site residue Arg¹⁵ of PN2KPI plays a major role in FXIa inhibition by protruding into the S1 specificity pocket of FXIa. Ala mutation at this site renders PN2KPI non-inhibitory for both FXIa and trypsin. BPTI has Lys¹⁵ at the P1 site. BPTI inhibits both FXIa and trypsin significantly less effectively than PN2KPI. PN2KPI-R15K lost FXIa inhibitory activity, whereas BPTI-K15R substantially gained affinity for FXIa. Like FXIa, trypsin preferred BPTI-K15R showing a significant enhancement in affinity. Thus, a major determinant of the inhibitory activity of PN2KPI and BPTI against FXIa and trypsin is the P1 residue, with Arg being preferred over Lys for both inhibitors and both proteases. In addition, loop 1 residues Pro¹³ and Arg²⁰ make important contributions to both FXIa and trypsin inhibition as demonstrated by significantly elevated K_i values for Ala mutations (P13A, and R20A) at these sites. In contrast, Ser¹⁹ makes no significant contribution to inhibition of either FXIa or trypsin whereas Met¹⁷ makes a significant contribution to the inhibition of trypsin, but not FXIa. In loop 2, only Phe³⁴ is identified as a residue making significant contributions to the inhibition of both FXIa and trypsin, since the PN2KPI-F34A mutant displayed reduced inhibitory activity for both FXIa (6-fold) and for trypsin (3-fold). To rationalize these findings, we examined the crystal structures of the FXIa(catalytic domain)/PN2KPI complex, and the trypsin/BPTI complex. Structurally, the PN2KPI loop-1, P1-site residue Arg¹⁵ makes a complex primary interaction with Asp¹⁸⁹ of both FXIa and trypsin. Disruption of this site by R15A mutation renders PN2KPI non-inhibitory because it preempts salt bridge interactions from two nitrogen atoms of the guanidinium group of Arg¹⁵ with Asp¹⁸⁹ and Gly²¹⁸ in FXIa. In addition, the Arg¹⁵ carbonyl oxygen forms hydrogen bonds with main-chain nitrogen atoms of one of the catalytic triad residues, Ser¹⁹⁵, and with Gly¹⁹³. The other important interaction in FXIa/PN2KPI or trypsin/BPTI is hydrophobic, between PN2KPI-Phe³⁴ and FXIa-Tyr¹⁴³ and between BPTI-Val³⁴ and trypsin-Tyr¹⁵¹. This intermolecular interaction is further strengthened by an intramolecular interaction in which the side chain of Phe³⁴ packs closely with the side chain of Met¹⁷ within PN2KPI, altogether forming a strong hydrophobic patch in FXIa-PN2KPI and trypsin-PN2KPI. PN2KPI-F34A disrupts both inter- and intramolecular hydrophobic interactions, leading to discernable reductions in affinity for both FXIa and trypsin. Despite occupying extreme positions in the autolysis loops, (¹⁴³YRKLRDKI¹⁵¹ in FXIa and ¹⁴³NTKSSGTSY¹⁵¹ in trypsin), Tyr^143/151 residues still orient themselves in close proximity to Phe³⁴. Thus, loop-1 residues of PN2KPI establish complex ionic interactions that play a major role, which is supplemented by the loop-2 residue, Phe³⁴ (in PN2KPI) or Val³⁴ (in BPTI), which establish hydrophobic interactions with residues in FXIa and trypsin leading to very high-affinity enzyme-inhibitor complexes.

No relevant conflicts of interest to declare.