Its time to recycle a paper I wrote for my Pharmcology course last semester. (P.S. Never write a receptor-based drug discovery paper for a clinical aspect-oriented Pharmacology course.)
Since the discovery of the adrenergic system of receptors, there have been numerous attempts to develop drugs that can modulate their effects. The earlier attempts were mostly synthetic approaches based on the structures of endogenous ligands like epinephrine and norepinephrine. Over the years our understanding of the ligand binding sites and the general structure of the receptor has improved considerably and therefore development of drugs for specific receptors is possible now. The x-ray crystallography structure of b2 adrenergic receptor was resolved recently and this provides a very good insight into the structure of the receptor and its differences from other GPCRs.
G-protein coupled receptors (GPCRs):
These are seven transmembrane receptors associated with heterotrimeric guanine nucleotide-binding proteins (G-proteins) that mediate a number of cellular responses to hormones and neurotransmitters. There are more than 800 different G-protein coupled receptors identified so far and a number of them are potential drug targets for a wide range of diseases.
Multiple phylogenetic analyses of GPCRs have revealed that they can be commonly divided into five families on the basis of their sequences and structural similarities. GRAFS classification system - Rhodopsin (Family A), Secretin (Family B), Glutamate (Family C), Adhesion (Family D) and Frizzled/Taste2 (Family E).
The general structure of GPCRs includes the characteristic seven membrane-spanning a-helical segments connected by alternating 3 intracellular loops and 3 extracellular loops. The ligand binds to the binding site on the extracellular surface resulting in conformational changes in the final intracellular loop. This conformational change allows for the binding of specific G-protein present on the inner surface of the membrane to the intracellular loop followed by the activation of the G-protein (replacement of GDP for GTP). The subunits of the activated G-protein can then trigger a cascade of intracellular signaling pathways that result in the necessary cellular responses. Many GPCRs can activate multiple signaling pathways by coupling with different G-proteins and different ligands.
According to the widely accepted, Ternary Complex Model, the receptor exists in two states – R state and R* state. The receptor stays in the R state in the absence of the agonist and in R* state (the activated state) in the presence of the agonist. However, recent studies have shown that a ‘basal activity’ or ‘constitutive activity’ exists even in the absence of a ligand. It is hypothesized that an equilibrium exists between the active state and the inactive state in the absence of a ligand and this equilibrium is shifted to active state (R*) on binding to an agonist. Studies show that mutations at certain positions in the transmembrane segments result in increased basal activity which shows some proof for this hypothesis, although we do not have a complete understanding of these conformations.
Thus according to this modified model, there are four possible types of ligands –
Full Agonists, Partial agonists, Antagonists and Inverse agonists. Full agonists exhibit maximal receptor stimulation whereas partial agonists are unable to elicit full activity. Neutral antagonists have no effect on receptor activation and Inverse agonists decrease the baseline receptor activity or ‘constitutive activity’.
Numerous biophysical studies suggest that a specific ligand stablizes a distinct receptor conformation and has distinct efficacies for different signaling pathways. This implies that a ligand can act as an agonist for one signaling pathway and simultaneously act as an antagonist to a different pathway on binding to the same receptor.
GPCRs are extensively targeted for the treatment of numerous diseases and therefore resolving their structures is vital for the development of novel and subtype-specific drugs. Owing to their structural flexibility and instability, only 6 GPCR structures have been resolved so far - b1, b2 adrenergic receptors, Adenosine receptor, Rhodopsin receptor, Dopamine (D3) receptor and Chemokine receptor (CXCR4).
b2 adrenergic receptors:
b2 Adrenergic receptors (b2-AR) are a part of the Rhodopsin family and are a good example for ligand-binding GPCRs. b2-AR are located almost throughout the body and are mainly implicated in cardiovascular and bronchial therapy. They are also clinically important in glaucoma treatment, smooth muscle relaxation (especially uterine relaxation, GIT relaxation and ciliary relaxation).
b2-AR interacts with both the stimulatory G-protein (Gs) and the inhibitory G-protein (Gi) and also MAPK. They can also activate MAPK through the b-Arrestin pathway (Figure 1). The Gs pathway activates the Adenylyl cyclase, which produces cAMP. The cAMP release results in smooth muscle relaxation and thus has implications in the treatment of asthma, inhibition of uterine contraction in premature labor and in glaucoma. The Gi pathway is regulated by PKA-mediated receptor phosphorylation and occurs only when the Gs pathway is blocked. b-arrestins are mainly considered to be regulatory proteins that cause receptor desensitization and internalization. But studies show that when certain inverse agonists bind to the receptor, b-arrestins can cause MAPK activation. So, it was found that individual ligands favor different pathways due to their varying affinities to each conformation of the seven transmembrane helices.
b2-AR-specific drugs are required in order to prevent unwanted side effects especially cardiovascular effects through b2-ARs. The most commonly used b2-AR specific drugs are – salbutamol, salmeterol, terbutaline etc. A number of agonists and antagonists for β-adrenergic receptors have been synthesized and tested for almost 50 years, and about 155 adrenergic drugs have been developed. Of these agents about 32 are considered specific to b2-AR and have been widely used clinically.
Crystal structure of b2 adrenergic receptor:
X-ray crystallography has been an important tool to understand the structures of cellular proteins and receptors. Bovine rhodopsin was the first GPCR structure to be resolved by X-ray crystallography in the year 2000. But due to the low natural abundance and structural flexibility or instability of most GPCRs, no significant progress was made in solving other structures. Ligand-binding GPCRs such as b2-AR show large number of conformational changes after ligand binding and therefore are very difficult to crystallize. Therefore, using targeted protein-engineering methods, mutants were developed to increase stability and crystallization of b2-AR. The following modifications were made –
· Intracellular loop 3 (ICL3) is considered to interact with the G-protein after the ligand binds to the receptor. This loop shows large physical movements thereby causing difficulties during crystallization. Therefore, ICL3 loop was truncated and replaced with the T4 Lysosome sequence that could restrict large movements (Figure A).
· The b2-AR-T4L mutants were cloned in Sf9 insect cells to improve yields.
· This mutant receptor was crystallized in the presence of Carazolol, a b2-AR-specific partial inverse agonist. This ligand favors the inactive conformation and proved to improve stability of the receptor.
But the resulting structure may have been compromised due to the limited movement of the seven transmembrane segments. Therefore, another method was developed using a monoclonal antibody (Mab5) that recognizes that native ICL3 (Figure A). The Fab segment of the antibody binds to the ICL3 segment forming the b2-AR–Fab complex. This did not alter the agonist-induced conformational changes, and ligand-binding affinities of the wild-type b2-AR. On comparing the two structures, they were found to be almost similar except for the positions of the extracellular loops and the higher basal activity seen in b2-AR-T4L receptors. Therefore, either structure can be used as a template for In-silico screening studies.
In 2008, another b2-AR crystal structure was resolved but this time while binding to Timolol (a partial inverse agonist) along with 2 molecules of cholesterol. This structure provided additional details about the binding sites of cholesterol molecules to the GPCR and also its effect on receptor stability when present in the plasma membrane. It could also provide insights into the trafficking mechanisms of b2-AR through cholesterol sequestration.
More recently in Jan 2011, the same lab from Stanford has published the structures of an agonist-bound b2-AR in both high affinity (bound to the G-protein) and low affinity (in the absence of G-protein) states. The b2-AR when bound with the Gs protein shows higher agonist affinity. Therefore, to resolve this high-affinity state structure, a camelid antibody fragment called Nanobody (Nb) was used. The Nb showed Gs-like binding properties and stabilized the high-affinity state during crystallization. The crystal structure of the b2-AR-T4L in the presence of an agonist (BI-167107) was obtained when stabilized with Nb80. For resolving the unstable low-affinity structure, an agonist (Procaterol) was covalently bound using a linker. These structures could reveal vital details about the differences in conformation between the antagonist- and agonist-bound GPCR.
In silico screening for drugs:
Virtual screening in general involves the assessment of a large database of compounds and the identification of potential ligands for specific target. There are two main types of methodologies – Ligand-based screening and Receptor-based screening.
a) Ligand-based screening involves the identification of new ligands using the 2D structures of previously known ligands. It produces very few hits as only related compounds from the same or similar classes can be obtained. Since only the 2D structures are being considered, therefore, inaccurate hits maybe obtained. Another ligand-based technique is the use of a 3D structure of the pharmacophore as the template through which some of the problems of 2D structure screening can be avoided.
b) Receptor-based screening or Structure-based drug discovery involves the use of computational docking modules to assess a large database against an X-ray crystal structure of the receptor. Most molecular docking studies include the process of assessing the various conformations of the ligand against the rigid target. Based on the binding force calculations and potential potency values, each ligand is scored and ranked. Recently developed docking software like GLIDE, DOCK, AUTODOCK, etc also take into consideration the flexibility of both the ligand and the receptor (template). Thus using these novel modules, better screening of compound databases can be performed.
Requirement of Structure-based drug discovery (SBDD):
Currently, all of the agents that are being used clinically were developed using ligand-based analoging techniques and not using the SBDD studies. Structure-based drug discovery techniques could potentially enable us to optimize selectivity, duration of action as well as to combine β2-agonist activity with desirable ancillary actions (e.g. anti-oxidant activity, calcium channel block, etc.). It also provides us with the possibility to design molecules that interact with multiple subtypes but show subtype selectivity. The most important advantage of SBDD is the potential to find newer classes of ligands whose pharmacophore is different from the regular catecholamine motif. These compounds could not be considered as possibilities during earlier studies.
Virtual screening results and potential drugs:
So far, a couple of studies have been published that used the Carazolol-bound b2-AR structure for small molecule screening. One study from Stanford University used the DOCK program to screen about 972,608 molecules from the ZINC database and individual ranks were assigned to each ligand. From the top 500, 25 compounds from 4 classes were selected and studied using radioligand-binding experiments. These were then compared with the known adrenergic ligands (8063 molecules) from the WOMBAT database. Most compounds were found to be similar to known adrenergic agents; although, few were novel chemotypes that have not been explored before. But the most interesting result observed from these studies was that all the compounds from the screen were either antagonists or inverse agonists and none were agonists.
Similarly in the second study from Lundbeck research, the GLIDE module was used to screen 400,000 compounds from proprietary database and about 4 million compounds from commercial databases. The physically available ones of the top 150 compounds from each database were selected and studied. These included a number of known antagonists such as carvedilol; thus validating the X-ray structure of the receptor and the in-silico screening program.
Recent studies in heart failure have shown that although agonists enhance acute effects, chronic use results in decreased signaling owing to desensitization mechanisms. In case of antagonists, they can potentially increase signaling on chronic use but cause severe side effects in acute conditions. In contrast, inverse agonists inhibit basal signaling initially and with chronic use up-regulate the receptors thus can be used as effective therapeutics.
The above studies provide a large list of inverse agonists of different affinities, pharmacokinetics and toxicity profiles which can be further studied for developing into drug candidates. It is clear that structure-based drug discovery using X-ray crystal structures are effective in the discovery of novel subtype specific drugs.
Problems faced in X-ray structure-based drug discovery:
The X-ray structures provide considerable advantages relative to the rhodopsin-based homology models using which a number of SBDD was being carried out earlier. But we still face a number of problems for precise determination of active ligands. One of the biggest obstacles with SBDD is the static nature of the X-ray crystal structures. It has been observed that the binding pocket of avian β1-AR bound to cyanopindolol and human β2-AR bound to carazolol are identical owing to a high conservation of binding-site contact residues. However, subtype-specific binding affinities can be observed for both β1- and β2-AR. These differences are due to the subtype-specific conformational preferences in distant residues, which in turn influence the amino acid spatial positions at the binding site.
Thus the static template of the receptor can only provide the data with respect to that particular conformation which is being used as the template. The above virtual screening studies using the carazolol-bound β2-AR crystal structure as a template identified number of new β2-AR ligands showing high affinities; however, most of the compounds exhibited inverse agonistic or antagonistic activity. Agonists could not be obtained from the screen using an inverse-agonist bound receptor template.
The GPCRs are currently the most widely targeted proteins for therapeutics and the X-ray crystal structures of these complex receptors has given the opportunity to study them in extreme detail and potentially develop novel, more potent drugs for existing and newer disease conditions. The primary intent for resolving the X-ray crystal structure of Beta-2 AR was to study and understand the structural differences and also the conformational mechanisms of signal transduction. But these structures have also been proven to be an excellent resource for drug discovery research.
Docking studies have revealed that most of the highly ranked ligands are inverse agonists or antagonists when screened against the inactive state of the receptor. The recently published agonist-bound structures may have to be used for obtaining potential potent agonists. A recent study has shown that by virtually modifying the binding site parameters of the inactive state receptor, the screen can also identify agonists and partial agonists. Such improvements allow us to potentially use any one of these structures to discover a wider range of subtype-specific and potent ligands.
Other future directions could include the use of dynamic images from other biophysical methods along with the information from static X-ray structures. To study the conformational changes and the rates of interconversion between these states we need to develop other time-dependent biophysical methods. Developments in fluorescence and NMR spectroscopy may help us to understand GPCR dynamics. Using such novel methods, more effective virtual drug discovery efforts can be undertaken and potentially develop novel, potent drugs for less cost and in less time.
1. Audet M and Bouvier M (2008) Insights into signaling from the β2-adrenergic receptor structure. Nat. Chem. Biol. 4, 397.
2. Bond RA and IJzerman AP (2006) Recent developments in constitutive receptor activity and inverse agonism, and their potential for GPCR drug discovery. Science 27, 92–96
3. Fredriksson R, Lagerstrom MC, Lundin LG and Schioth HB (2003) The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints. Mol. Pharmacol. 63, 1256–1272.
4. Hanson MA, Cherezov V, Roth CB, Griffith MT, Jaakola VP, Chien EYT, Velasquez J, Kuhn P, and Stevens RC (2007) A Specific Cholesterol Binding Site Is Established by the 2.8 Å Structure of the Human β2-Adrenergic Receptor. Science 318, pp. 1258–1265.
5. Kolb P, Rosenbaum DM, Irwin JJ, Fung JJ, Kobilka BK, Shoichet BK. (2009) Structure-based discovery of β2-adrenergic receptor ligands. Proc. Natl. Acad. Sci. U.S.A. 106 (16): 6843-6848.
6. Rasmussen SGF, Choi HJ, Rosenbaum DM, Kobilka TS, Thian FS, Edwards PC, Burghammer M, Ratnala VRP, Sanishvili R, Fischetti RF, Schertler GFX, Weis WI and Kobilka BK (2007) Crystal structure of the human β2 adrenergic G-protein-coupled receptor. Nature 450, 383–387.
7. Rosenbaum DM, Cherezov V, Hanson MA, Rasmussen SGF, Thian FS, Kobilka TS, Choi HJ, Yao XJ, Weis WI, Stevens RC and Kobilka BK (2007) GPCR engineering yields high-resolution structural insights into β2-adrenergic receptor function. Science 318, 1266–1273
8. Rosenbaum DM, Rasmussen SG and Kobilka BK (2009) The structure and function of G-protein-coupled receptors. Nature 459, 356.
9. Rubenstein LA, Zauhar RJ and Lanzara RG, (2006) Molecular dynamics of a biophysical model for beta2-adrenergic and G protein-coupled receptor activation, J Mol Graph Model 25 (4), pp. 396–409.
10. Sabio Ma,b, Jones Kb and Topiol Sa,b (2008) Use of the X-ray structure of the β2-adrenergic receptor for drug discovery. 18:1598–1602.aPart 2: Identification of active compounds. Bioorg Med Chem Lett 18:5391–5395.b