Current Trends in GPCR Allostery.

G-protein-coupled receptors (GPCRs) are seven transmembrane proteins which are mostly located on the cell surface. More than 800 GPCRs are known to be present in our body. Of these, ~ 400 are olfactory or pheromone receptors and remaining GPCRs are devoted to mediate neuroendocrine and/or pharmacological stimuli ( Hauser et al. 2017 , 2018 ). GPCRs are classified as class A to F ( Krishnan et al. (2016) ). Class A or “rhodopsin-like” receptor family is comprised of hormone-, neurotransmitter-, and light-receptors ( Zhou 2019 ) that accounts for largest homologous group of GPCRs ( Fig. 1a ). Class B receptors, also known as “secretin receptor family,” consist of about ~ 70 receptors ( Hu et al. 2017 ). Unlike class A receptors, they have a long N-terminal extracellular domain which consists of about 120 residues. Class C consists of metabotropic glutamate receptors, GABA receptors, calcium-sensing receptors, and taste receptors. These receptors have a large extracellular N-terminal domain with approximately 600 residues to which endogenous ligands bind. Class D includes fungal mating pheromone receptors, class E includes cAMP receptors, and class F includes frizzled/smoothened receptors ( Hu et al. 2017 ). Crystal structures of the ligand-bound states are available for representative GPCRs from families A, B, C, and F ( Fig. 1b ). Based on classical pharmacological practice, the orthosteric pocket is the site on a GPCR where the endogenous ligand binds. Orthosteric sites lie in transmembrane region of most GPCRs with the exception of the glutamate family class C receptors ( Christopoulos et al. 2014 ). Representative structure from each class of GPCRs is shown in Fig. 1d – g . There are ~ 89 GPCRs for which the endogenous ligands are still unknown. They are classified at present as orphan GPCRs ( Davenport et al. 2013 ). Upon binding of the endogenous agonist, GPCRs are activated and recruit heterotrimeric G-proteins in order to mediate downstream signaling. Heterotrimeric G-protein subunits consist of three subunits, i.e. , G α , G β , and G γ . G α has four types, i.e. , G αs , G α12/13 , G αq , and G αi . G αs is classified into two subtypes- G αs and G αolf , G α12/13 into two subtypes—G α12 and G α13 , G αq into three subtypes—G αq , G α11 , and G α15/16 , and G αi into eight subtypes—G αo , G αi1 , G αi2 , G α3 , G αt1 , G αt2 , G αt3 , and G αz ( Oldham and Hamm 2008 ). Most activated GPCRs recruit one of these G-proteins. There are a few exceptions, for example, the angiotension type 2 receptor does not recruit G-proteins. Crystal structures reveal that most of the G-proteins bind to the receptor in the same orientation in which the C-terminal region of the G α primarily interacts with the GPCR. Multiple sequence alignment of G α -proteins reveals that C-terminal residues are conserved across the G-protein sequences and thereby suggesting their specificity towards cognate GPCRs. Till now, 421 receptor structures have been solved for 73 unique GPCRs ( Fig. 1b ) [ https://gpcrdb.org/ ]. Of these, 410 structures were solved with a native or pharmacological ligand and 19 structures are solved with an allosteric ligand. Definition for an allosteric ligand is one that interacts with the receptor outside the orthosteric ligand pocket and exerts experimentally measurable effects on the pharmacology and signaling of the GPCR. The site on a GPCR where the allosteric ligand binds then becomes the allosteric ligand pocket. Thus, there are no GPCR class-specific allosteric binding sites or class-specific allosteric modulators. About 34% of the Food and Drug Administration (FDA)-approved drugs are targeted to GPCRs ( Fig. 1c ). Therefore, GPCRs remain the primary target for most drugs ( Sriram and Insel 2018 ). The goal of designing a therapeutic drug is to modulate the function of a GPCR, and hence, it can target either the orthosteric or allosteric site of the GPCR. Till now, most of the drugs used in the clinics target the orthosteric site. The drugs targeting the allosteric site are limited at present but gain popularity due to their significant advantages. In this chapter, we will discuss the importance and need of allosteric drugs for GPCRs. Binding sites other than the endogenous ligand binding sites are known as allosteric sites and ligands which bind to the allosteric sites are known as allosteric ligands. Allosteric ligands can modulate the functions of the endogenous ligand. Based on the receptor function modulated by the allosteric ligand in the presence of endogenous ligand, allosteric ligands are classified as negative allosteric modulators (NAM), positive allosteric modulators (PAM), and silent allosteric modulator (SAM) ( Christopoulos et al. 2014 ; Kenakin and Strachan 2018 ). A negative allosteric modulator decreases the action of the endogenous ligand, a positive allosteric modulator increases the action of the endogenous ligand and a silent allosteric modulator does not alter the function of the endogenous ligand. Despite conservation of the architecture in the GPCR family, there are no distinct regions of GPCR to harbor a specific site for a PAM or a NAM. However, allosteric ligand binding to a defined allosteric site can elicit responses described as PAM or NAM. Based on these properties, an allosteric ligand can be used in different ways to combat the pathophysiology induced by GPCRs. The major disadvantages of orthosteric drugs are as follows: (i) conservation of sequence and properties in transmembrane regions of GPCRs are high across the family. This type of conservation is highest within subtypes of receptors that bind the same endogenous ligand as in the cases of adrenergic receptors (α, β), muscarinic receptors, serotonin receptors, and angiotensin receptors. Therefore, orthosteric sites are highly conserved and within the subtypes, they are almost identical since they are activated by the same endogenous ligand. (ii) In examples like muscarinic receptors, opioid receptors, or chemokine receptors which have many subtypes—with very high subtype sequence similarity and activated by the same endogenous ligand—developing specific orthosteric agonist or antagonist becomes a complicated issue. (iii) Orthosteric site-targeted drugs completely block the binding of endogenous ligand and consequently inhibit all important signals elicited by the endogenous ligands. The role played by multiple signals may be vital and inhibition of these may cause severe side effects in the long run. However, allosteric ligands are highly specific for a receptor and often minimize even subtype cross-specificity. Allosteric sites are less conserved and hence drugs can be very specific to the target receptor. Allosteric ligands for specific subtypes have been developed that overcome complex cross reactivity issues because allosteric sites differ between subtypes. Allosteric ligands (NAM or PAM) are widely used for targeting a specific subtype for activating (e.g., M 2 ) or inhibiting (e.g., β 2 ) the receptor signaling. Moreover, allosteric ligands also reduce the off-target activity and significantly reduce the undesirable side effects. Allosteric drug activity is completely reliant on the presence of the endogenous agonist. Endogenous ligands are produced when it is required by the cell. Therefore, these allosteric drugs have the potential to maintain activity dependence and both temporal and spatial aspects of endogenous physiological signaling. At the same time, it won’t have the risk of overdosing. Allosteric drugs do not compete with the endogenous agonist and thus do not require high binding affinity with the receptor ( Grover 2013 ; Abdel-Magid 2015 ; Wenthur et al. 2014 ). As of September 2020, 19 crystal structures of GPCRs with allosteric ligands have been deposited in the protein data bank. These structures are from nine different GPCRs. These structures reveal that the allosteric site in GPCRs can be located at any region, i.e. , extracellular, transmembrane or in the intracellular region ( Fig. 2a – c ). The structures were solved in different states, i.e., inactive, intermediate, and active states. The allosteric ligands consist of both NAM and PAM. Details of the GPCRs with their allosteric ligand are shown in Table 1 . The commonly used approaches for computer-aided drug design (CADD) are structure-based drug design (SBDD) and ligand-based drug design (LBDD). SBDD typically uses protein or RNA (targets) 3D structures, which identify the key interaction sites (active sites) for biological functions, and then the drugs are designed ( Yu and MacKerell 2017 ). As we know, the purpose of a drug is to influence the biological functions by binding to the target’s active site. LBDD adopts an indirect approach to design desired pharmacologically active compounds by studying the molecules that interact with the target of interest ( Acharya et al. 2011 ; Kurogi and Guner 2001 ). This method is used in the absence of a 3D structure of target molecule. This method is also useful in design of prodrugs to increase the specificity or bioavailability of the original drug molecules ( Balakrishnan 2006 ; Takakura and Hashida 1995 ; Tolle-Sander et al. 2004 ). Until 2000, only one GPCR structure was solved. Technological advancements in recent years facilitated the solving over 400 structures of GPCR ( Gacasan et al. 2017 ) and thus facilitating design of several allosteric ligands using structure-based methods. Identification of an allosteric site (binding site or active site) is the first and key step involved in SBDD. Wakefield et al. (2019) used several bioinformatics tools to analyze the tractable allosteric sites in GPCRs ( Wakefield et al. 2019 ). The methods they have adopted will be very useful in predicting the allosteric sites of GPCRs. Korczynska et al. (2018) used both the inactive (PDB ID code 3UON) and active structures (PDB ID code 4MQT) of M2 muscarinic acetylcholine receptor (M2 mAChR) to identify PAM and NAM compounds. Extensive molecular docking was used to identify these allosteric ligands ( Korczynska 2018 ). Graaf et al. (2011) using homology modeling and structure-based virtual screening algorithm against the glucagon receptor (GLR) and the glucagon-like peptide 1 receptor (GLP-1R) identified two compounds which can inhibit the glucagon response. One compound was found to bind to GLP-1R and potentiate the response to the endogenous GLP-1 ligand ( Graaf et al. 2011 ). This was an “accidental” discovery of an allosteric compound for GLP-1R. Recently, Lückmann et al. (2019) performed molecular dynamics (MD) simulation of fatty acid receptor 1 (FFAR1) after removing fatty acid. They observed that an unoccupied, solvent-exposed pocket closes and a major conformational change of the receptor occurs. They then mined for a compound using structure-based virtual screening to prevent the closing of this pocket. Finally, they have identified a compound which proved to be an ago-PAM ( Luckmann 2019 ). MD simulation remains an invaluable complementary tool in the modern drug discovery. Moreover, continuing development in this technique can provide the mechanistic, thermodynamic, and kinetic insight of protein–ligand interactions which then simplify the complex problem associated in GPCRs drug discovery ( Lamim Ribeiro and Filizola 2019 ). Further, MD simulation can also determine the allosteric communication between the extracellular region and intracellular G-protein/β-arrestin binding site ( Vaidehi 2016 ). Ligand-based drug discovery was prominently used in the discovery of drugs for GPCRs until 2012. Taylor et al. (2010) used tetrazole peptidomimetic which can stabilize the photoactivated state of rhodopsin to model pharmacophore and screen against various chemical databases. Three compounds were found to be stabilize rhodopsin in the meta-rhodopsin II state (MII). Interestingly, these compounds were not inhibited by binding of transducin to the photoactivated state of rhodopsin and they were assumed to be an allosteric ligand ( Taylor 2010 ). Butkiewicz et al. (2019) used multiple metabotropic glutamate (mGlu5) positive allosteric modulators to employ in silico quantitative structure–activity relationship (QSAR) modeling and subsequent ligand-based virtual screening. They prioritized a set of 63 potential PAMs identified from a library of over 4 million compounds. Finally, using medicinal chemistry optimization, they were able to discover PAM compound VU6003586 with a potency of 174 nM ( Butkiewicz 2019 ). There are few allosteric drugs in clinical use. Only four drugs have been approved for clinical use out of 365 drugs which are in clinical trials for 30 different GPCRs. Around 93% are in preclinical studies and 1–3% of the drugs are in Phase I-III clinical trial. The names of the receptors and the allosteric ligands with the stage of clinical trials are shown in Table 2 . Autoantibodies are also generated against GPCRs in similar fashion to autoimmune diseases. These autoantibodies can activate the GPCRs and cause several diseases which include endocrine diseases such as Graves’ disease, cancer, hypertension and heart diseases, Alzheimer’s diseases, and organ transplant rejection ( Cabral-Marques 2018 ; Schulze et al. 2005 ; Meyer 2018 ). The epitope of these autoantibodies is present on extracellular region of GPCRs suggesting that autoantibodies bind to the extracellular sites ( Xia and Kellems 2011 ). This binding of autoantibody can be blocked by pharmacological intervention. The drugs which can block binding of autoantibodies can prevent the diseases caused by autoantibodies. However, the allosteric ligand in this regard should be a silent allosteric modulator which does not influence the normal signaling by the autoantibody targeting GPCR. In summary, allosteric ligand application in GPCRs is a fast-growing field in receptor pharmacology. The number of GPCR crystal structures solved with allosteric ligands is also continuously increasing. Allosteric drugs have the potential to inhibit, activate, or maintain the normal signaling of the receptor which suggests that receptor signaling can be modulated based on the physiological requirement without blocking endogenous ligand binding. Currently, over 300 drugs are in preclinical trials and this number is expected to be double in few years.