It has been reported that somatostatin receptor subtypes 4 and 5 would be high-impact templates for homology modeling if their 3D structures became available. They have generated a homology model of the somatostatin receptor subtype 4 (sst4), using the newest active state β2 adrenoreceptor crystal structure, and subsequently docked a variety of agonists into the model-built receptor to elucidate the binding modes of reported agonists. Using experimental restraints, they were able to explain observed activity profiles. They propose two binding modes that can consistently explain findings for high-affinity agonists and reason why certain structures display low affinities for the receptor.
G protein coupled receptors (GPCRs) are integral membrane proteins targeted by an estimated 27% of the current rule-of-five-compliant experimental and marketed drugs. Besides being the drug targets of many pharmaceuticals, GPCRs also represent an opportunity for future therapeutic intervention based on the druggable genome.The GPCR superfamily is divided into three main families: A, B, and C, which share no sequence homology with one another, although their structures are similar. The largest is family A (class I or rhodopsin-like) and accounts for almost 85% of the GPCR genes. Over half of the class I GPCRs is predicted to encode olfactory targets, while the remaining are bound to known endogenous compounds or are classified as orphan receptors. Furthermore, the rhodopsin family is subdivided into four groups, α, β, γ, and δ, and 13 sub-branches. Even though there is no overall correlation between phylogeny and the types of ligands that bind to the rhodopsin family, a few general receptor clusters can be observed. Family B (class II) corresponds to the secretin/adhesion receptor family, and family C (class III) is the metabotropic glutamate/pheromone protein family.
GPCRs are involved in signaling cascades that regulate several cellular and physiological functions. The ligands that bind to these receptors range from ions to photons, peptides, proteins, lipids, and nucleotides. Activation-induced by agonists causes changes in the relative orientation of the transmembrane helices (TMs), leading to a wider intracellular surface with an exposure of those intracellular amino acids that are crucial for signal transduction. Inverse agonists and antagonists bind to a number of different sites, but prevent the TM helix reorientation. Given that a conformational equilibrium between active and inactive states exists for receptors, it is believed that the agonist-bound conformation of GPCRs will shift the equilibrium toward the active receptor states, with helices undergoing significant movements during activation. Specifically upon agonist binding, the GPCRs undergo conformational changes, leading to regulation of the associated G protein (Gαβγ) by stimulating the GDP–GTP exchange. The resulting GTP-bound α subunit dissociates from the β and γ subunits and subsequently modulates several intracellular effectors. Effectors such as adenylyl cyclases, phospholipase C, and phosphoinositide-3 kinases in turn regulate the production of secondary messengers and impact intracellular signaling proteins or target functional proteins.
Sequence alignment of sst4 with β2-adrenoceptor (PDB code 3P0G). Yellow highlights identical residues. Green indicates similar residues. Blue shows conservative residues. TM domains are labeled.
Structure determination for these systems has been challenging due to their heterologous expression. Recombinant production of GPCRs is still a matter of trial and error, while their folding and stability, when expressed in their native system, are complex and not yet fully understood processes. Furthermore, although milligram amounts of certain GPCRs are attainable, the majority of them are still either produced at very low levels or not at all.
The first crystal structure of a GPCR appeared 11 years ago and corresponded to bovine rhodopsin with the covalently bound endogenous chromophore cis-retinal in the inactive state. Because rhodopsin can be easily obtained in high quantities, several other rhodopsin crystal structures have been published since then. However, rhodopsin is a light-activated GPCR, and therefore such a template is inadequate for any homology modeling of ligand-activated GPCRs. The first non-rhodopsin crystal structure was reported for the β2 adrenergic receptor bound to a partially inverse agonist, carazolol, seven years later. Other non-rhodopsin structures were later resolved: the β2 adrenergic receptor bound to cholesterol and a partially inverse agonist timolol, the mutant version of the β-1 adrenergic receptor with cyanopindolol the A2A adenosine receptor with the high-affinity antagonist ZM241385,the human dopamine D3,and CXCR4 chemokine receptors. It should be noted that the chemokine receptor belongs to the γ group of rhodopsin-like receptors, whereas the others belong to the α group. The CXCR4 structure has a different topology and binding pocket, which in turn could provide a better template for GPCR homology modeling of the γ group.
(A) Putative model of sst4 (green) superimposed with template (orange). (B) Overlay of the binding site residues. The only differences are Ser 207, Tyr 308, and Asn 312 of 3P0G versus Gly 183, His 258, and Ile 262 of sst4, respectively.
While the above structures can be useful as templates for the antagonist-bound conformations of GPCRs, the trans-retinal/rhodopsin complex structure is still evasive, as are agonist-bound GPCR conformations, with few exceptions reported this year. Notable examples of informative activated crystal structures include a ligand-free, but with a G protein bound, opsin which is believed to be in an active conformation and has provided a better understanding of the requirements for full activation. Also, three agonist-bound structures of the thermostable turkey β1 agonist-bound adrenergic receptor, the human A2A agonist-bound adenosine receptor, and the human β2 agonist bound active state adrenoreceptor were deposited in the PDB this year. Notably, β2 was the first GPCR ever to be crystallized with an agonist and a peptide resembling the G protein, whereas the other two are only agonist-bound. Even more recently, the β2 adrenergic receptor–Gs protein complex was released.
Given the preceding discussion, it becomes apparent that additional GPCR crystal structures are needed for drug design and in order to investigate more direct templates for model-building of other closely related GPCR classes. However, with the latest reported agonist-bound structures, studies on agonist-receptor activation and agonist discovery programs have become more feasible.
Binding mode I. The violet grid indicates the TM2/TM7 hydrophobic cavity formed by Ala71, Val67 in TM2, Val259, Ile262, and Leu263 in TM7. The cyan grid depicts the aromatic cavity formed by Phe175 and Phe239. Residues involved in hydrogen bonding are shown in red and corresponding distances in yellow dotted lines.
The somatostatin receptors are rhodopsin-like GPCRs, consisting of five subtypes (sst 1–5). They mediate the inhibitory effects of somatostatin on secretion and proliferation, depending on receptor subtype and tissue localization. Most importantly, sst4 has been recognized as an ideal therapeutic target for Alzheimer’s disease, due to its high expression in neocortical and hippocampal areas, which are significantly affected by amyloid β accumulation. Since there is no crystal structure, studies on sst4 ligand binding and activation have been scarce. There are, however, reports that aspartic acid in TM3 is essential for ligand binding.
Somatostatin (somatotropin release-inhibiting factor, SRIF) is a hormone peptide, which is normally expressed as a tetradecapeptide (SRIF-14) or an N-terminally extended form (SRIF-28). Both SRIFs contain a disulfide bond between cysteines at positions 3 and 14 which stabilize the structure. SRIF peptidic structure–activity relationship (SAR) studies indicated that the core residues Trp8 and Lys9 are the essential binding sites for all somatostatin receptors. Furthermore, alanine scanning studies carried out by Lewis et al. demonstrated that Trp8 and Lys9 were essential for binding to all sst subtypes, whereas Phe6 is specifically important for sst4 activation.
(A) Representative examples of compounds consistent with binding mode I. The red surface represents the TM2/TM7 hydrophobic cavity. Cyan encodes for the aromatic phenylalanines 175 and 239. (B) Putative binding of compound 21 (purple surface) superimposed with binding mode I (yellow surface).
Due to poor oral bioavailability and rapid degradation by peptidases, SRIF is not viable for therapeutic application. Consequently, the development of orally effective, metabolically stable non-peptide compounds has been the focus of many research groups. On the basis of SAR data, compound series have been designed to mimic Trp8, Lys9, and Phe6. Ankersen et al. were the first to report a non-peptide ligand with a 6 nM Ki and over 100-fold selectivity toward sst4. Many other compounds have also been designed to mimic SRIF with varying binding affinities, ranging from 0.7 nM to 10 μM Ki values. However, it is not clear or obvious why such structurally similar compounds have so diverse binding affinities and how the receptor can also accommodate substantially diverse ligands. Are there distinct binding modes that could possibly explain observed assay data? Is there a possibility that high-affinity compounds adopt more specific modes with an increased number of binding and/or stabilizing interactions, as opposed to low-affinity molecules? To address these questions, they generated a homology model of sst4 and docked a number of reported compounds into the model-built structure. Inspection of the resultant receptor–agonist complexes led us to propose two partially overlapping, but not identical, binding modes for the high- and low-affinity sst4 agonists, which are presented here. It should also be pointed out that the high affinity binding they are proposing is consistent with experimental data for residues thought to be critical for molecular recognition at the sst4 level. Finally, a virtual screening experiment was carried out to validate the effectiveness of the model and its ability to retrieve actives seeded in a compound collection of 996 decoys.
Binding mode II. The violet grid shows the TM2/TM7 hydrophobic cavity and cyan, the aromatic cavity formed by Trp171, Phe175, Phe239, and Tyr240. Hydrogen bonds are also depicted.
Representative low affinity compounds (purple surface) superimposed with binding mode II (white surface).
A Structure-Based Approach to Understanding Somatostatin Receptor-4 Agonism (sst4) Zhaomin Liu, A. Michael Crider, Daniel Ansbro, Christina Hayes, and Maria Kontoyianni Journal of Chemical Information and Modeling 2012 52 (1), 171-186 DOI: 10.1021/ci200375j