I. Introduction
II. General Structure
III. Membrane-binding and EGF-like domains
IV. Cyclooxygenase Site
V. Peroxidase Site
VI. References
Prostaglandin H2 synthase (PGHS) is a dimeric membrane enzyme responsible for catalyzing the committed step in prostaglandin synthesis. Prostaglandins are lipid mediators involved in physiological processes including platelet aggregation, modification of vascular tone, and inflammation. Pharmacological control of PGHS activity is therefore central to the treatment of fever, inflammation, and heart disease, and has also shown potential in the prevention of colon cancer and Alzheimer’s disease (2).
The synthesis of prostaglandin H2 (PGH2) by PGHS takes place in two steps, occurring at spatially separate active sites, a cyclooxygenase (COX) site and a peroxidase site. In the first step, arachidonic acid is oxygenated at the COX active site to form hydroperoxide prostaglandin G2 (PGG2), which is reduced at the peroxidase active site to the alcohol PGH2. The two reactions are linked by the catalytic residue Tyr385. The peroxidase reaction converts this residue to a tyrosyl radical species, which is necessary for activation of the COX reaction cycle (2). Reaction mechanism
Pharmacological modification of PGHS activity is currently limited to inhibition of the COX reaction by non-steroidal anti-inflammatory drugs (NSAIDs), such as aspirin and ibuprofen; no peroxidase-specific agents have yet been developed (5). However, because the peroxidase reaction can continue to operate and produce free radical species even after termination of the COX reaction, such agents are of medical interest. In this crystal structure, the peroxidase site is occupied by a molecule of glycerol (used as a cryoprotectant during X-ray diffraction), representing the first view of a ligand in this active site, and potentially providing insight into peroxidase-inhibitor design (2).
Prostaglandin H2 synthase is a homodimer, composed of two 70 kDa monomers< >. Each monomer contains three structural domains< >: an EGF-like domain, a membrane-binding domain, and a catalytic domain, which contains both the COX and peroxidase active sites. The protein is made up largely of α helices, with the exception of the EGF-like domain, which has a two-stranded, anti-parallel β-sheet structure. In addition, each monomer contains three sites of glycosylation< >, at Asn68, Asn144, and Asn410. Glycosylation has been shown to be necessary for correct protein folding, but its role in PGHS function is currently unclear (2).
In the 2.0 Å resolution crystal structure of ovine PGHS-1 described here, α-Methyl-4-biphenylacetic acid< >—a defluorinated analog of the non-steroidal anti-inflammatory drug (NSAID) flurbiprofen—is shown bound to the COX active site. The peroxidase site is occupied by the cryoprotectant glycerol< > (2).
The membrane binding domain is composed of four alpha helices< >. These interact with eight molecules of a detergent, beta-octyl glucoside (BOG) < > . The binding of the protien to BOG is representative of the binding of this protien to membrane lipids. The distance the lipids of a membrane would penetrate into the protein is difficult to determine, but this model suggests that they bind weakly to the face of the binding domain. The EFG-like domain is defined by beta sheets near the membrane binding domain< > which are held together by three intradomain disulphide bonds between cysteine residues, and connected to the main body of the enzyme by an additional disulfide bond between cysteine residues 37 and 159< > (2).
The COX active site is found within the catalytic domain, at the end of a long, hydrophobic channel that extends from the membrane-binding domain and allows diffusion of arachidonic acid/inhibitor into the active site< > (2,6). A smaller channel also runs into the COX active site, nearly perpendicular to the main channel. This channel is lined with well-defined solvent molecules, and is thought to allow solvent release as substrate enters the active site (2,3)< >.
α-Methyl-4-biphenylacetic acid is bound to the COX active site< >, occupying the same position as other previously studied NSAIDs (2). Inhibitor binding occurs through a network of interactions, including two hydrogen bonds between the carboxylate group of the inhibitor and Arg120< >, another hydrogen bond between the inhibitor carboxylate and the hydroxyl group of Tyr355< >, and a salt bridge connecting Arg120 to Glu524< > (5). This same set of interactions has been shown to be necessary for binding of the COX substrate arachidonic acid, thus NSAIDS are considered competitive binding inhibitors (4).
The distal end of the inhibitor (opposite the carboxylate group) is positioned below the active site roof (residues Phe381, Leu384, Tyr385, and Trp387)< >, and alongside Leu352, Ser530, and Met522< > (5).
The catalytic residue Tyr385< > is positioned at the end of the main COX channel, surrounded by the hydrophobic residues Phe381, Leu384, and Trp387< > (2,5). A link between Tyr385 and Ser530 is formed by a hydrogen-bonding water molecule positioned between the two residues< > (2). When arachidonic acid is bound to the active site, its 13-pro-S-hydrogen is positioned for removal by the tyrosyl radical of Tyr385, which initiates conversion of arachidonic acid to PGG2 (4,6).
The peroxidase site is positioned in a cleft defined by 4 helices surrounding a heme cofactor < >. The proximal face of the heme cofactor is packed against 2 helices and lines one side of the cleft. The heme cofactor is not bound covalently to PGHS and there are relatively few protein-heme interactions. One arm of the heme H-bonds with an Asn 382 , and to a Thr 212 < >. Other interactions are weak, Van derWaals interactions. The heme iron is coordinated on the proximal side with a nitrogen of a His 388 < >.
The coordination with a His is conserved across the heme dependent peroxidases, however, in this peroxidase, the bond length is much longer than normal. Data shows a relationship between heme-nitrogen bond length and reduction potential. In PGHS, when 30% glycerol is added, the reduction potential increases from -167 mV to -52 mV, suggesting that interactions on the distal face of the heme affect the proximal bond length (2).
In this structure, glycerol is bound in the peroxidase site on the distal side of the heme group< >. The glycerol lies between Gln 203 and His 207 < >. It is too far from the heme to be considered a sixth ligand, but prevents water ligation. Val 291 and Leu 294 extend over glycerol backbone, forming a pocket for glycerol to pack into< >. Glycerol occupies the space we would expect substrate to bind in, so the interactions between the glycerol and PGHS could be useful in designing peroxidase inhibitors (2).
1. Brookhaven Protein Data Bank (http://www.nist.rcsb.org/pdb/). PDB: 1Q4G.
2. Gupta, K., Selinsky, B.S., Kaub, C.J., Katz, A.K., and Loll, P.J. 2004. The 2.0 Angstrom Resolution Crystal Structure of Prostaglandin H2 Synthase. J. Mol. Biol. 335: 503-518.
3. Loll, P.J. "Re: Solvent Release Channel in PGHS-1." Email to Patrick J. Loll. 24 November, 2004.
4. Rowlinson, S.W., et. al. 2003. A Novel Mechanism of Cyclooxygenase-2 Inhibition Involving Interactions with Ser530 and Tyr385. J. Biol. Chem. 278: 45763-45769.
5. Selinsky, B.S., Gupta, K., Sharkey, C.T, and Loll, P.J. 2001. Structural Analysis of NSAID Binding by Prostaglandin H2 Synthase. Biochemistry. 40: 5172-5180.
6. Smith, W.L., DeWitt, D.L., and Garavito, R.M. 2000. Cyclooxygenase: Structural, Cellular, and Molecular Biology. Annual Rev. Biochem. 69: 145-182.