HMG CoA Reductase
Created by Kate Roche
3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoA reductase) (pdb ID=1HWK) serves an important role in the cardiovascular health of humans by catalyzing the committed step in cholesterol biosynthesis. This step in the pathway involves the reductive deacylation of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA). The protein uses four electrons to reduce HMG-CoA to mevalonic acid and coenzyme A (CoA), as shown in Figure 1, using two molecules of NADPH as electron donors (Istvan & Deisenhofer, 2001, 1160). Mevalonic acid is the precursor for a class of molecules known as isoprenoids, and products resulting from this biosynthetic pathway include heme and farnesyl-pyrophosphate as well as cholesterol. The role of HMG-CoA reductase in the regulation of cholesterol levels in the blood results in the protein having clinical importance and makes the protein a target of pharmaceutical research. Coronary artery disease affects as many as fourteen million people in the United States, and elevated cholesterol levels are a primary risk factor in developing the disease (Istvan & Deisenhofer, 2001, 1160). The protein’s role in cholesterol synthesis makes it an effective target for drugs that can inhibit it. Drugs collectively known as the statins competitively bind to the active site of the protein and inhibit HMG-CoA reductase. This reduces the overall cholesterol levels in the blood and combats the effects of coronary artery disease (Istvan et al., 819).
The structure of the catalytic portion of HMG-CoA reductase explains the binding of the protein to its substrate HMG-CoA (pdb ID= 1DQ8) and also the mechanism of the catalytic breakdown of HMG-CoA. The catalytic portion, which has a molecular weight of 204,458.2 Da and an isoelectric point (pI) of 7.73 as provided by the ExPASy Database, is made up of four identical subunits that form a tetramer (Expasy, 1). ADP is bound to the protein to induce crystallization (Istvan & Deisenhofer, 2001, 1164). The catalytic portion includes residues 426-888 for each of the four subunits (Istvan et al., 820) while residues 1-339 reside in the membrane of the endoplasmic reticulum (Istvan & Deisenhofer, 2000, 9). In the catalytic portion of the protein, the protein subunits are arranged into two dimers, each with two bipartite active sites formed by residues from both monomers (Istvan & Deisenhofer, 2000, 11). Each monomer has three unique domains that make up its tertiary structure: a helical amino-terminal domain (N-domain) consisting of residues 460-527, a large domain (L-domain) consisting of residues 528-590 and 694-872, and a small domain (S-domain) consisting of residues 592 to 682. The L-domain has a 27-residue alpha helix as the central element, while the S-domain has a central antiparallel four-stranded β-sheet (Istvan & Deisenhofer, 2000, 10).
These domains play an integral role in the binding of HMG-CoA and NADPH. The molecule can bind four molecules of both HMG-CoA and NADPH. HMG-CoA binds to the L-domain, while NADPH primarily binds to the S-domain. The two domains are connected by a loop consisting of residues 682-694 known as the cis-loop, which is stabilized by interactions between monomers and is part of the HMG-binding site (Istvan & Deisenhofer, 2000, 10). To initiate the binding of HMG-CoA, interactions between the enzyme and the substrate occur in the L-domain of one monomer while Tyr-479 of the neighboring monomer forms a “hydrophobic shield” over the adenine base that closes the binding pocket (Istvan & Deisenhofer, 2000, 14). Lys-691 stabilizes charged intermediates, and Ser-684, Asp-690, and Lys-692 also participate in the reduction of HMG-CoA (Istvan, 2003, 4). The binding of NADPH puts His-866 within hydrogen bonding range of HMG-CoA to serve as a proton donor (Istvan & Deisenhofer, 2000, 16).
The structure of HMG-CoA reductase also explains binding of one of its ligands, atorvastatin, and the inhibition of the protein by the competitive binding of the class of drugs known as the statins. All statins contain an HMG-like moiety bound to a hydrophobic group that can range from extremely hydrophobic, as in cerivastatin (pdb ID= 1HWJ), to only slightly hydrophobic, as in rosuvastatin (pdb ID= 1HWL) (Istvan, 2003, 5). Statins can additionally be separated into two groups based on their structure. Type 1 inhibitors, such as compactin (pdb ID= 1HW8), contain a decalin ring structure, and type 2 inhibitors, such as atorvastatin (pdb ID= 1HWK), have different central ring structures (Istvan, 2002, 29). Statins competitively bind HMG-CoA reductase at the binding site for HMG-CoA. For all statins, the HMG-like moiety forms ionic and polar interactions with the enzyme, and the flexibility of the enzyme and rearrangement of carboxy-terminal residues allows for the hydrophobic groups to fit between helices in the L-domain (Istvan, 2003, 7). Certain statins also form unique interactions with the protein. Type 1 statins bind through interactions with the decalin ring with a helix as well as through hydrogen bonding, and type 2 statins bind through interactions between their fluorophenyl groups and Arg-590. Atorvastatin, the ligand of the protein of interest, also forms a hydrogen bond between a carbonyl oxygen and Ser-565 (Istvan, 2003, 7). As a result of these interactions, statins sterically prevent HMG-CoA from binding by blocking the active site.
The structure of HMG-CoA reductase also provides insight into the intramolecular interactions in the formation of quaternary structure as well as the regulation of the protein. The overall secondary structure contains 37% α-helix and 20% β-sheet, and both of these structural elements are involved in the oligomerization of the protein. All three domains of each monomer contribute to the dimerization of the protein. Amino acids of the L-domains of two adjacent monomers form a β-sheet with nine inter-monomer hydrogen bonds (Istvan et al., 821). This region includes a conserved sequence element, ENVIGX3I/LP, which is the same in both class I and class II proteins (Istvan et al., 821). Two helices from the helical subdomain of the L-domain of each monomer also fold together to form a buried 4-helix bundle (Istvan et al., 821). The 310-helix of the N-domain of one monomer interact with a helix of the L-domain of another, and the S-domain’s β-sheet interacts with two helices in the L-domain of an adjacent monomer, with a salt bridge forming between Arg-595 and Glu-730 to stabilize the interactions (Istvan et al., 821). Hydrogen bonds between Glu-700 and Glu-700 on neighboring monomers, the packing of the β-sheet of the L-domain of one monomer against the S-domain of another, and hydrophobic interactions between the dimers lead to the formation of the tetramer structure (Istvan et al., 822). The protein’s structure also explains the regulation of HMG-CoA reductase. The levels of cholesterol in the blood regulate the activity of the enzyme through control of transcription and translation, degradation of existing enzymes, and through phosphorylation by protein kinases (Istvan et al., 819). Phosphorylation of Ser-872 results in a decreased affinity for NADPH and decreased activity of the enzyme (Istvan & Deisenhofer, 2000, 16).
The catalytic portion of HMG-CoA reductase (HMGR) in Pseudomonas mevalonii (pbd ID= 1QAX), a class II HMGR, shares some structural and functional similarities with the catalytic portion of human HMGR, but the sequences of the two proteins do not share extensive similarities. A PSI-BLAST search determines primary structure similarities by comparing a query protein to other known protein sequences and assigning them an “E value” based on sequence homology. A search with human HMG-CoA reductase as a query showed no significant sequence similarities between the human and Pseudomonas proteins because a significant E value is less than .05 and the E value between these two proteins is large (BLAST, 1). Class I and class II HMGR proteins generally only show around 14-20% similarity in primary structure (Istvan, 2002, 28). The results of a Dali server search (Z=28, rmsd=2.2), which compares tertiary structure elements and assigns proteins a “Z score” for similarity, as well as superimposition of the two proteins and a side-by-side view, show that the proteins share significant tertiary structure similarities since a significant Z score is greater than 2 (Dali, 1). While both enzymes catalyze the same reaction, the bacterial HMGR catalyzes it in reverse, synthesizing HMG-CoA from mevalonate and CoA (Istvan et al., 820).
The proteins have many key differences and similarties in structure and sequence. Human HMGR has eight transmembrane α-helices, while class II HMGRs are soluble and cytoplasmic (Istvan et al., 819). Both proteins have a dimeric active site with residues contributed by both monomers, and the folding in the catalytic domains is similar. Functionally similar residues in both proteins also contribute to CoA binding despite differences in amino acid sequence (Friesen, 2), and some key catalytic residues, including His-381 in the bacterial HMGR, are conserved in both classes of HMGR (Tabernero, 7167). The quaternary structures of the proteins differ in that Pseudomonas protein monomers associate to form a hexamer instead of the tetramer formed by human HMG-CoA reductase (Istvan, 2001, 747). Some elements of secondary structure that contribute to the tertiary and quaternary structure, such as the β-sheet of the L-domain, are conserved (Istvan, 2001, 747). One of the most striking differences between the two proteins is the absence of the cis-loop in the HMGR of Pseudomonas, but the class II protein compensates by replacing the catalytically necessary lysine of the cis-loop (Lys-691) with a lysine on a neighboring monomer (Lys-267) (Istvan, 2001, 748).