Defining the COX Inhibitor Selectivity of NSAIDs: Implications for Understanding Toxicity

Kathleen M Knights; Arduino A Mangoni; John O Miners


Expert Rev Clin Pharmacol. 2010;3(6):769-776. 

In This Article

COX Catalysis & Structure

COX is a heme-containing bifunctional protein that sequentially catalyzes two reactions. The first reaction involves cyclooxygenation of the endogenous substrate arachidonic acid to yield the hydroperoxy endoperoxide PGG2. Subsequent reduction of the hydroperoxyl moiety of PGG2 results in formation of PGH2. The latter peroxidase reaction occurs in an adjacent, but spatially distinct, site within the COX catalytic domain. Further metabolism of PGH2 by various PG synthetases results in formation of PGD2, PGE2, PGF, PGI2 (prostacyclin) and TXA2 (Figure 1). PGI2 and TXA2 are biologically active but unstable and are rapidly broken down nonenzymatically, forming the less active 6-keto-PGF and inactive TXB2, respectively. For extensive reviews refer to.[3–5]

To further appreciate the unique catalytic aspects of COX and the pharmacology of NSAIDs it is essential to consider the relevant structural features of COX-1 and COX-2. Functional COX-1 and COX-2 proteins contain 576 and 587 amino acids, respectively, and within the same species COX-1 and COX-2 share 60–65% amino acid sequence identity. The 3D structure of both has been determined by x-ray crystallography.[6,7] COX enzymes are homodimers. Each monomer comprises three domains: a N-terminal EGF-like domain, a membrane-binding domain and a C-terminus catalytic domain that comprises approximately 480 amino acids (~80% of the protein) and contains the two distinct catalytic sites – the COX and the peroxidase active sites. The EGF-like domain appears to be the dimerization domain holding the monomers together through hydrophobic interactions, hydrogen bonding and salt bridges.[3]

Unlike many proteins that have transmembrane spanning domains, the membrane-binding domain of the COX enzymes comprises four a-helices that create a hydrophobic surface, which integrates into a single leaflet of the lipid bilayer (monotopic insertion) on the luminal side of the endoplasmic reticulum (ER) and the nuclear envelope.[3] The catalytic domain comprises the COX and the peroxidase active sites. The peroxidase active site is located in the vicinity of the heme, near the protein surface. The helices of the membrane-binding structure create the entrance to a hydrophobic channel that provides the access route for arachidonic acid released from the hydrophobic core of the lipid bilayer. This narrow dead-end channel forms the COX active site. Within the active site, two important amino acids at the apex of the channel, Tyr385 and Ser530, are crucial for activity (Figure 2). Reaction of an endogenous oxidant (e.g., hydroperoxide) with the heme generates a higher oxidation state of the heme, which then oxidizes Tyr385 to a tyrosyl radical. COX proteins are inactive until the tyrosyl radical is generated and hence this reaction is the rate-determining step in the initiation of the COX catalytic cycle. During the course of the oxygenation of arachidonic acid the tyrosyl radical is initially reduced to tyrosine, but is regenerated in the last step of the catalytic cycle by the peroxy radical precursor to PGG2. Regeneration of the tyrosine radical is essential for continued oxygenation of arachidonic acid.[4] Thus, the presence of an endogenous oxidant, such as hydroperoxide, in the incubation milieu is critical for the measurement of COX activity in vitro.

Figure 2.

Isoleucine to valine substitution opens up the hydrophobic 'side' pocket in COX-2 relative to COX-1. Note the wider channel opening in COX-2 and the narrowing caused by Arg120 and Tyr355 in COX-1. Tyr385 and Ser530 are essential for COX activity. Ser530 is irreversibly acetylated by aspirin.

Other crucial amino acids in COX proteins include Arg120 and Tyr355, which together form a narrow constriction in the channel towards the bottom of the COX active site. Modeling studies of arachidonic acid in the COX active site indicate that arachidonic acid lies in an L-shaped configuration with the ω-end of the fatty acid near the top of the active site. The 13-pro(S) hydrogen of arachidonic acid locates adjacent to Tyr385 and a pairing interaction occurs between the carboxylate group and the guanidinium group of Arg120.[8] Mutation of Arg120 significantly alters COX activity of COX-1 but has much less effect in COX-2, suggesting that the pairing interaction between the carboxylate group and Arg120 plays a greater role in arachidonic acid binding in COX-1 compared with COX-2.[9]

COX-2 closely resembles COX-1 with the exception that the COX-2 active site accommodates larger chemical structures owing to substitution of the isoleucine at position 523 in COX-1 with a valine in COX-2. The loss of a single methyl group arising from the isoleucine to valine substitution is sufficient to open up a secondary internal hydrophobic side pocket in COX-2 that increases the volume of the active site by approximately 25%.[10] This enlarged COX site explains the recognition by COX-2 of bulkier substrates, including arachidonylethanolamide and 2-arachidonylglycerol.[11] There are also other subtle changes that result in a wider channel opening (~17% greater) in COX-2.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.