Polymorphic human prostaglandin H synthase-2 proteins and their interactions with cyclooxygenase substrates and inhibitors

Abstract

The cyclooxygenase (COX) activity of prostaglandin H synthase-2 (PGHS-2) is implicated in colorectal cancer and is targeted by nonsteroidal anti-inflammatory drugs (NSAIDs) and dietary n-3 fatty acids. We used purified, recombinant proteins to evaluate the functional impacts of the R228H, E488G, V511A and G587R PGHS-2 polymorphisms on COX activity, fatty acid selectivity and NSAID actions. Compared to wild-type PGHS-2, COX activity with arachidonate was ∼20% lower in 488G and ∼20% higher in 511A. All variants showed time-dependent inhibition by the COX-2-specific inhibitor (coxib) nimesulide, but 488G and 511A had 30-60% higher residual COX activity; 511A also showed up to 70% higher residual activity with other time-dependent inhibitors. In addition, 488G and 511A differed significantly from wild type in Vmax values with the two fatty acids: 488G showed ∼20% less and 511A showed ∼20% more discrimination against eicosapentaenoic acid. The Vmax value for eicosapentaenoate was not affected in 228H or 587R, nor were the Km values or the COX activation efficiency (with arachidonate) significantly altered in any variant. Thus, the E488G and V511A PGHS-2 polymorphisms may predict who will most likely benefit from interventions with some NSAIDs or n-3 fatty acids.

Figure 1

COX kinetic parameters for PGHS-2 wild type (WT) and variants with AA (open circles) and EPA (filled circles). Values for V_max relative to the rate with 100 µm AA and no HOOH (a) and K_m (b) represent the average ± s.d. for triplicate determinations. Asterisks indicate values significantly different from wild type (P < 0.05).

Figure 2

Kinetics of COX inhibition in hPGHS-2 wild type (WT) and variants during preincubation with 2µm nimesulide. Aliquots of each protein (10 ⁻⁸ m) were preincubated with (filled symbols) or without (open symbols) inhibitor for the indicated lengths of time before the surviving COX activity (i.e., enzyme not in the EI′ form) was assayed by injection of 100 µm AA. The values represent averages ± s.d. for triplicate measurements.

Figure 3

Kinetics of re-equilibration of tightly bound nimesulide complexes of hPGHS-2 wild type (WT) and the E488G and V511A variants upon 1500-fold dilution. The diluted samples were subsequently incubated for the indicated lengths of time before the COX activity was assayed by injection of 100 µm AA. Activities were normalized to control values obtained by dilution into buffer already containing 100 µm AA. Averages ± s.d. for two separate experiments are presented. Each set of data points was fitted to a single exponential equation with a 25 s lag.

Figure 4

Crystallographic structure of murine PGHS-2 (6cox.pdb), with views focused on Val511 (a) and Glu488 (b). Residue numbers are for human PGHS-2, with ovine PGHS-1 equivalents in parentheses. Other highlights include heme, bound inhibitor (SC-558), the tyrosyl radical sites at Tyr371(385) and Tyr490(504), the coxib specificity determinant at Val509(523), and the Arg442(456) and Tyr481(495) residues that interact with Glu488.

Scheme 1

Mechanistic scheme for interaction of PGHS (E) with COX inhibitor (I).