Comparison of prostaglandin H synthase isoform structures using limited proteolytic digestion.

Prostaglandin H synthase (PGHS) catalyzes a key step in the biosynthesis of a variety of bioactive lipid mediators. The two known isoforms (PGHS-1 and -2) share about 60% amino acid identity, but exhibit distinct interactions with substrates, activators, and inhibitors. Ovine PGHS-1 has previously been shown to have a distinctive protease-sensitive site near Arg277; cleavage by trypsin, chymotrypsin, or proteinase K produces fragments of 33 and 38 kDa and loss of activity. The ovine PGHS-1 crystal structure shows Arg277 located in an exposed loop structure; homology modeling predicts similar loop structures for both human isoforms (hPGHS-1 and -2). We have used limited proteolytic digestion of recombinant hPGHS-1 and hPGHS-2 to probe their structures. Incubation of hPGHS-1 with either trypsin or proteinase K produced 33- and 38-kDa fragments and loss of activity. In contrast, incubation of hPGHS-2 with the same proteases led to cleavage of only a 2- to 3-kDa fragment, with no decrease in activity. Immunoblotting with site-specific antibodies demonstrated that the cleaved fragment originated from the hPGHS-2 C-terminus. Similar immunoblotting experiments indicated that trypsin did not attack the ovine PGHS-1 C-terminus. Mutagenesis was used to replace Pro263 of hPGHS-2 (corresponds to Arg277 of ovine PGHS-1) with arginine, inserting a potential trypsin site. Incubation of this P263R hPGHS-2 mutant with either trypsin or proteinase K resulted in cleavage near the C-terminus and retention of activity, just as with wild-type hPGHS-2. A peptide containing residues 259-268 of the P263R mutant was cleaved by trypsin at the same rate as a peptide corresponding to hPGHS-1 residues 272-281, demonstrating that the sequence differences were not responsible for the lack of tryptic cleavage at residue 263 in the hPGHS-2 mutant. Preincubation of hPGHS-2 with graded levels of guanidinium HCl before incubation with proteinase K did not produce large proteolytic fragments, indicating that the hPGHS-2 loop region was not selectively unfolding. The results point to two regions of significant structural difference between PGHS-1 and -2: the Arg277 loop, which is protease-sensitive in PGHS-1 but protease-resistant in PGHS-2, and the C-terminus, which is protease-sensitive in PGHS-2 but not in PGHS-1.