COX Receptors and GPCRs as Model Systems for Studying

Effect of Structure on Specificity

Kenneth O'Dell


            COX 1 and 2 receptors are involved in the production of essential prostaglandins that maintain the gastrointestinal lining. The structural difference between COX 1 and COX 2 is minute enough to serve as an example of acute substrate specificity. On the other hand, G-protein coupled receptors (GPCRs) are the gate keepers in intercellular communication. These receptors are activated by the binding of a specific extracellular substrate. GPCRs consist of a large superfamily and possess diverse structural differences to account for the need for different substrates in order to prevent mixed signals in cellular communication. By studying the structure of COX receptors and GPCRs, comparisons can be made between the levels of substrate specificity. The structural difference can be narrowed down in COX receptors to one amino acid; isoleucine in COX-2 is substituted with valine in COX-1 at two locations at the active site. Specificity of GPCRs is still not well elucidated. However, the structural factors which affect GPCR specificity cannot be narrowed down to one amino acid difference. Instead, it results from a cooperation of several factors, such as intramolecular interactions, covalent modifications, and structural flexibility. Ultimately, COX receptors and GPCRs help serve as model systems for studying the effect of a receptor's structure on its specificity.


            Proteins are the molecular machines of life with diverse functions from sensing light to serving as a chemical bridge to allow cells to communicate. Being multi-cellular organisms, humans have specialized cells which require a higher level of intercellular communication to maintain life. Of course, with such complex functions comes a myriad of signal molecules and other substrates which react with various proteins within our cells. Protein receptors are highly specialized to have varying degrees of specificity for different substrates. The substrate specificity of cyclooxygenase 1 and 2 varies greatly from that of  G-protein coupled receptors and a comparative study of these two receptors can help clarify the regulation of specificity by structure.


COX 1 and 2

            The cyclooxygenase receptors are responsible for production of prostaglandins (PGs) which are crucial to the maintenance of the gastrointestinal lining and kidney function2. Cyclooxygenase 1 is present in a constitutive form, and thus a steady amount of PGs necessary for functions such as maintaining gastrointestinal lining are produced regularly. Along with COX-1, a second type of cyclooxygenase receptors exists called COX-22. Unlike that of COX-1, COX-2 activity is induced by pro-inflammatory proteins known as cytokines, which are responsible for both nonspecific and specific immune responses2. In the case of COX-2, the immune responses are nonspecific such as inflammation.  COX receptor structures are homodimers consisting of two identical monomers as shown in Figure 114.           







Figure 1: Basic crystallography showing the two dimerized monomers. Perioxidase site is above the heme prosthetic group (shown in red) while the oxidation site is near the membrane domain (shown in yellow)14.

Each monomer of COX receptors has two catalytic functions: an oxidizing function, followed by a reduction function. The COX (oxidation) site is found on the side of the monomer near the endoplasmic reticulum membrane while the POX (catalytic unit called peroxidase for reduction) site is located on the opposite end of the monomer4. As shown in Figure 2, the oxidation process involves the addition of two oxygen molecules to the substrate, Arachidonic Acid (AA), forming a peroxidase with a 5 carbon ring called PGG2 (Prostaglandin-G2)3. The POX site reduces PGG2 to PGH2 (Prostaglandin-H2) via the addition of two hydrogen atoms. PGH2 is a precursor of many prostaglandins4.


Figure 2: Oxidative and reductive catalytic functions of COX monomers3.

            Based on research results, it has been suggested that only one monomer can perform catalytic activity at any given time and substrate binding to one monomer causes a conformational change in its partner monomer implying reciprocal allosteric regulation between them4. It is possible that the specificity for Arachidonic Acid (AA) may be regulated by this conformational change.

            Inhibition of COX-1 receptors can lead to deleterious side effects such as gastrointestinal ulcers and bleeding due to the restriction of the passive biosynthesis of necessary prostaglandins that maintain the gastrointestinal lining and kidney function2. Prostaglandins are also mediators of vasodilation because they serve as substrates to receptors that control muscle constriction2. They promote relaxation of the blood vessels which ultimately dilates them2. The dilation of the blood vessels constitutes inflammation and swelling which is perceived as a painful sensation by the individual. Many pharmacological benefits of pain killers can be improved by selectively inhibiting COX-2 receptors because many of the NSAIDs (non-steroidal anti-inflammatory drugs) in use such as aspirin inhibit both COX-1 and COX-2. The deleterious side effects of the NSAIDs come from preventing the steady biosynthesis of mandatory PGs by inhibiting COX-1 even though they may effectively decrease the inflammation, swelling and the pain sensation by inhibiting COX-2 simultaneously 2.

            When COX-2 was discovered, it was found to be extremely identical with COX-1 and a 63% homology existed between them. However, there were a few differences in the amino acid sequences at the entrance to the active site and one within the active site 12. Copying sequences from COX-1 and overlaying them over COX-2 using molecular modeling techniques allowed a better understanding of which amino acid differences lead to a substantial change in specificity2. In this case, one amino acid within the active site was determined to be the main factor in producing the difference in specificity between COX-1 and COX-21. This determination was made by a molecular modeling experiment via point mutagenesis (selectively replacing a specific amino acid with another)1.  It was found that COX-2 had the amino acid valine at residues 434 and 523 whereas COX-1 had isoleucine at the same locations2. As shown in Figure 3, valine's side chain is smaller with a structure of R(CH)(CH3)2 whereas isoleucine's side chain contains an additional -CH2 group ['R(CH)(CH3)(CH2)(CH3)] 2. What does COX-2 gain from the valine substitution of isoleucine? It simply allows for more space in the active site.



Figure 3: 3-D structures of valine and isoleucine.

            In the absence of an extra -CH2 group, there is less steric hindrance for the binding of an inhibitor molecule (e.g. NSAID) and thus a wider range of inhibitors bind to COX-2 compared to COX-1. COX-2 specific NSAIDs are the target of research and pharmacological interests. Overall, this small variation in a single amino acid structure can lead to a big difference in drug selectivity between COX-1 and COX-2 making the inhibition of inducible, pro-inflammatory PGs synthesis possible without the inhibition of essential PGs synthesis. In the pursuit of this goal, the new generation NSAIDs including Celecoxib and MK-966 that show no signs of gastrointestinal damage are synthesized. One study showed that these COX-2 selective inhibitors helped relieve pain after dental surgery in some volunteers2. Subsequent studies aim to test COX-2 selective inhibitors on colon cancer and Alzheimer's which are known to be two contributors to inflammation2. SC-588 is another COX-2 selective inhibitor found to help eliminate inflammatory effects and has shown no side effects such as gastrointestinal damage1. Although the tested drugs were used on select patients, several side-effects of COX-2 inhibitors have been hypothesized2. One includes inhibition of cell growth and repair since the same inflammatory cytokines are also associated with cells that release growth factors2.


            GPCRs constitute a large family of membrane proteins which are of significant pharmacological value since almost 50% of the drugs on the market target these receptors and development of new drugs would be facilitated with a better understanding of their structure10. Being membrane proteins, GPCRs prove to be difficult in structure determination. Part of this difficulty is due to their complexity as well as how the membrane layer helps to maintain the delicate structure of the GPCR5. GPCR structure makes it hard to crystallize due to the conformational flexibility in its extracellular loops. Furthermore, analysis of the crystal structure is typically not fully indicative of how the proteins would function in vivo since GPCRs are dynamic and their structures shift to allow them to conform to a variety of substrates5.

            Figure 4 shows the universal structure of GPCRs consisting of 7-transmembrane helices (TMH), an active extracellular loop which binds to signaling molecules, and an intracellular loop which serves as the signal transducer and interacts with an internal heterotrimeric G-protein9. Helices are the most common structure found in the transmembrane space due to their adaptation of hydrophobic surfaces which allow them to diffuse and exist in the lipid bilayer9. Another property that unifies the superfamily of GPCRs is the mechanical function of the helices. Due to the helices’ rigidity, they are ideal for mechanical function serving as a rotator, a lever, or a piston9. The highly dynamic extracellular loop undergoes structural changes when it binds to a substrate and sends this conformational change down through the helices which serve to relay the signal mechanically similar to piston-like, lever-like, or rotational movement9. A combination of movements such as the rotation of a curved or bent helix may be used to achieve a large displacement at the intracellular end9.

            Although the structure of GPCRs can be considered universal, they constitute a very large family of proteins and have 6 subfamilies12. Serving as a true display of its wide range of specificity, GPCRs within the same subfamily may have sequence similarities up to 20% or higher and dissimilarities as small as 3% which suggests a strong homology within the subfamilies yet a wide range of diversity in terms of structure versus function12. A comparative analysis of structural similarities within subfamilies was conducted by Worth et al12, and the results clearly showed the similarities within subfamilies even though each independent receptor still has distinct functions such as light perception in rhodopsin or sympathetic nervous system mediation12.



Figure 4: Topology of Subfamily A GPCRs. Superimposition of 5 template structures12.

        In addition to the seven trans-membrane helices, there is also an intracellular eighth helix which plays a role in receptor activation by G-protein receptor kinases (GRK)7. One example of the helix 8's function can be seen in the throtropin-releasing hormone receptor (TRHR), a class A GPCR, where a conserved positively charged site facilitates phosphorylation (the addition of a phosphate group) by a kinase which in turn activates the receptor. (Conversely, palmitoylation, addition of a fatty acid, palmitate, occurs at the end of TMH 7 and serves to keep TRHR in its inactive, non-constitutive form6.) Altogether, this is an example of how a particular module found in one class of GPCRs may have a universal function within that class, as in the case with helix 8 and phosphorylation7. Function may vary due to the lack of conserved motifs between the classes of GPCRs. To further exemplify similarities in structure regardless of large differences in homology, research has been done to model class B GPCRs using class A templates13. Little is known about class B GPCRs and thus comparative studies between class A and class B are difficult due to the lack of homology, however it is possible to identify the analogous motifs by searching the specific locations in class B GPCRs based on the locations of their analogs in class A13.


            COX receptors and GPCRs regulate their own levels of specificity and this has been taken advantage of in pharmacology. GPCRs, the super family of membrane proteins, are responsible for many processes of signal transduction and cellular communication. The active loops of GPCRs are dynamically shifting structures and some receptors have large active sites which can allow substrates with a wide range of sizes and thus have a broader level of specificity (selection of substrates is also based on its binding affinity to the receptor). GPCRs  have subtypes of receptors within subfamilies with their own specificity which can be taken advantage of if inhibiting a certain subtype is needed, yet a current problem in GPCR agonists/antagonists is having low specificity. Targeting certain subtypes is difficult since some drugs will readily bind to various receptor subtypes. In contrast, the COX-1 and 2 receptors have a difference in substrate specificity that has been used to develop specific drugs (such as celecoxib) for COX 2. GPCRs such as the cholinergic receptors have a level of specificity employed for the development of many quaternary ammonium muscle relaxants that mimic acetylcholine's structure, but as discussed earlier, selectivity may not be specific enough for the cholinergic subtypes. GPCR specificity is still not fully elucidated and it may result from the cooperation of factors such as interactions between the transmembrane helices, intracellular post-translation modification, and dynamic structure of the extracellular active loop6, 7, 11.  Determination of the interplay between these factors should be sought to fully understand the degree of specificity in GPCRs. In conclusion, there is a very high level of specificity in COX receptors (a mere removal of a -CH2- group to reduce steric collisions) in comparison to GPCRs (a broad group of proteins whose specificity cannot be narrowed down to one amino acid substitution) which still require further research to identify the hidden sources of receptor subtype specificity.



1. James K. Gierse, et al. A Single Amino Acid Difference between Cyclooxygenase-1 (COX1) and -2 (COX-2) Reverses the Selectivity of COX-2 Specific Inhibitors. The Journal of Biological Chemistry 1996, 271(26): 15810-15814.

2. J. R. Vane, et al. Cyclooxygenases 1and 2. Annu. Rev. Pharmacol. Toxicol. 1998, 38: 97-120.

3. Kurumbail, R; Kiefer, JR; Marnett, LJ. Cyclooxygenase enzymes: Catalysis and inhibition. Current Opinion in Structural Biology. 2001, 11 (6): 752–60. 

4. Sharma, Narayan P, et al. Asymmetric Acetylation of the Cyclooxygenase-2 Homodimer by Aspirin and Its Effects on the Oxygenation of Arachidonic, Eicosapentaenoic, and Docosahexaenoic Acids. Molecular Pharmacology. 2010, 77(6): 979-86

5. Buchen, Lizzie. Opioid receptors revealed. Nature. 2012, 482: 383

6. Du D, Raaka BM, Grimberg H, Lupu-Meiri M, Oron Y, Gershengorn MC. Carboxyl tail cysteine mutants of the thyrotropin-releasing hormone receptor type 1 exhibit constitutive signaling: role of palmitoylation. Molecular Pharmacology. 2005, 68: 204–209.

7.  Gehret, Austin U., et al. Role of Helix 8 of the Throtropin-Releasing Hormone Receptor in Phosphorylation by G Protein-Coupled Receptor Kinase. NCBI. 2010, vol. 77(2): 288-97

8.  Klabunde, Thomas, Hessler, Gerhard. Drug Design Strategies for Targeting G-Protein-Coupled Receptors. ChemBioChem. 2002, 3(10): 928-44 

9.  Sterne-Marr, Rachel, et al. G Protein-coupled Receptor Kinase 2/Gao/11 Interaction. The Journal of Biological Chemistry. 2003, 278: 6050-58

10. Stevens, Raymond C., et al. The GPCR network: a large-scale collaboration to determine human GPCR structure and function. Nature. 2013,12(1): 25-34

11. Williamson, Mike. How Proteins Work. New York: Garland Science, Taylor & Francis Group, 2012. 313-15.

12. Worth, Catherine L., Kleinau, Gunnar, Krause, Gerd. Comparative Sequence and Structural Analyses of G-Protein-Coupled Receptor Crystal Structures. PLOS one. 2009. Open web             access at          

13. Vohra, Shabana, et al. Similarity between class A and class B G-protein-coupled receptors exemplified through calcitonin gene-related peptide receptor modelling and mutagenesis studies. Royal Society Publishing. 2012, 10(79)

14. Simmons, Daniel L., Botting, Regina M., Hla, Timothy. Cyclooxygenase Isozymes: The Biology of Prostaglandin Synthesis and Inhibition. Pharmacological Review. 2004, 56(3): 387-437.