Adenosine Receptor Modeling: What Does the A2A Crystal Structure Tell Us?

1

Adenosine receptor modeling: what does the A2A crystal structure tell us?

Diego Dal Ben*, Catia Lambertucci, Gabriella Marucci, Rosaria Volpini, Gloria Cristalli

School of Pharmacy, Medicinal Chemistry Unit, University of Camerino, Via S. Agostino 1, 62032 Camerino, Italy

Abstract. For a long time, there have been no experimentally determined structural data for any adenosine receptor (AR) and the only approach available for making structure/function correlations about these proteins has been homology modeling. While the early attempts to model these receptors followed the crystallization of bacteriorhodopsin, the cryo-microscopy studies of bovine and frog rhodopsin, and the modeling of a Cα-template for the TM helices in the rhodopsin family of GPCRs, the crystallization of bovine rhodopsin by Palczewski was of extreme importance as it first provided the crystal structure of an eukaryotic GPCR to be used as template for more realistic homology models. Since then, rhodopsin-based modeling became the routine approach to develop AR structural models that proved to be useful for interpretation of site-directed mutagenesis data and for molecular docking studies.

The recently reported crystal structures of the adrenergic beta1 and beta2 receptors only partially confirmed the structural features showed by bovine rhodopsin, raising a question about which template would have been better for further modeling of ARs. Such question remained actually not-answered, due to the publication in late 2008 of the crystal structure of human adenosine A2A receptor (AA2AR). Since its publication, this structure has been used for ligands docking analysis and has provided a high similarity template for homology modeling of the other AR subtypes. Still, the AA2AR crystal structure allows to verify the hypotheses that were made on the basis of the previously reported homology modeling and molecular docking.

Keywords: adenosine receptors, docking, homology modeling, molecular modeling, mutagenesis, structure-based drug design.

INTRODUCTION

Adenosine receptors (ARs) are four G protein-coupled receptors (GPCRs) belonging to a specific cluster of the class A family (also named rhodopsin family or family I [1]). They are referred to as adenosine A1, A2A, A2B, and A3 receptors (AA1R, AA2AR, AA2BR, and AA3R, respectively) [2] and are activated by extracellular adenosine (Ado). Each AR has unique pharmacological profile, tissue distribution, and effector coupling. Mainly, activation of AA1R and AA3R inhibits adenylyl cyclase (AC) activity through Gi and stimulates phospholipase C (PLC) through Gβγ subunits, while activation of AA2AR and AA2BR stimulates AC activity through Gs. With the exception of the AA3R, the existence of AR subtypes in various tissues had been realized before their cloning as results of pharmacological characterizations [3]. Cloning of the four AR subtypes has made easier the understanding of several aspects of AR activity at a molecular level. Considering the overall protein structure, the recently reported crystal structure of human AA2AR [4] confirms that ARs display the topology typical of GPCRs, as they have in common a central core domain consisting of seven transmembrane helices (TM1-7) composed of 20–27 amino acids and largely α-helical. TM domains are also slightly bent and linked by three intracellular (IL1, IL2, and IL3) and three extracellular (EL1, EL2, and EL3) loops. There is also a short helix VIII that runs parallel to the cytoplasmic surface of the membrane. While two

Cysteine residues (one in TM3 - 3.25 - and one in EL2) are conserved in most GPCRs and form a disulfide link critical for the packing and for the stabilization of a restricted number of conformations of these seven TMs, the AA2AR crystal structure presents four disulfide links at extracellular level. AA1R, AA2BR, and AA3R are very similar considering the number of amino acids composing their primary structure, with the length of the cloned AR sequences varying from 324 to 328 residues for the AA1R, from 332 to 340 for the AA2BR, and from 314 to 320 for the AA3R [5-7]. AA2AR sequence length varies from 409 to 412 [8] and this is due to an extension of the C-terminal region, which consists of about 120 residues. Among the four AR subtypes N-terminal extracellular, C-terminal intracellular, and IL domains differ in length and function and each of these domains provides very specific properties to these receptor proteins. The carboxyl-terminal tails of the AA1R, AA2BR, and AA3R, but not AA2AR, possess a conserved Cysteine that may putatively provide a site for receptor palmitoylation and allow the formation of a fourth intracellular loop. Considering overall sequence identity at the amino acid level, the four human AR subtypes share about 46% (average) amino acid sequence identity, with higher similarity between AA2AR and AA2BR (58%) and only 37% between AA2BR and AA3R subtypes. The sequence identity is about 57% within the TM domains, with higher conservation again between AA2AR and AA2BR (about 70%).

2

Fig. (1). Sequence alignment of the four human AR subtypes. Transmembrane (TM), intracellular loop (IL), extracellular loop (EL), and C-terminal (C-TERM) domains are indicated; * symbols indicate sequence identity in all the four subtypes; C letters indicate cysteine residues involved in disulfide bridges as observed in hAA2AR crystal structure [4] and cysteine residues in the other ARs when conserved in all the four subtypes; 1 numbers indicate the cysteine pairs involved in the four disulfide bridges as observed in hAA2AR crystal structure; R letters indicate the x.50 position for each TM domain with the numbering system suggested by Ballesteros and Weinstein [11]; bold letters indicate residues analysed with mutagenesis studies described within this review. A sequence alignment of human ARs is presented in Fig. (1). The highest residues conservation is found in the upper part of AR TM domains and this data suggests a common mechanism for ligand recognition for the four subtypes. On the other hand, specific variable amino acids give unique pharmacological features to the various subtypes of the ARs. For example, there is a 90% sequence identity among the residues between the AA2AR and AA2BR in ligand binding region but the AA2BR has a significantly lower affinity for Ado and prototypical nucleoside agonists [9]. Different affinities of the four ARs for the same ligands could be related to substitutions of aminoacids composing the binding

pocket but also residues distal to the ligand, which contribute to the formation of meta binding sites to which the ligand binds in its path to the principal binding site. These elements have been the topics of a number of molecular modeling studies on ARs reported to date. In particular, the attempts to rebuild molecular models of ARs were aimed at explaining or guiding mutagenesis studies, at understanding the molecular basis of receptor-ligand recognition, and at taking advantage of the chemical-physical properties of the AR subtypes binding site to design potent and selective ligands. Considering mutation analyses carried out for the ARs, some of these studies employed site-directed mutagenesis to target

3

individual amino acids while other studies were based on the construction of chimeric receptors composed of varying amounts of sequence from two parent wild-type ARs. Most frequently, these mutated ARs were transiently expressed in mammalian cells and their pharmacological profiles determined in ligand-binding analysis, or in some studies, via functional assays. It must be remembered that perturbations in ligand-binding affinity observed upon amino acid substitution may arise from direct or indirect effects on the ligand-receptor interaction. In the first case the targeted amino acid actually interacts with the ligand under study, while an indirect effects occurs if the mutation alters overall protein architecture and influences the interaction of the ligand with receptor in a region far from that in which the substitution has been made. In fact, the observation that a receptor mutation modulates agonist affinity may not simply reflect disruption of the ligand-binding pocket, but rather, a modification in the ability of a GPCR to undergo conformational changes associated with agonist binding and G-protein coupling [10]. In the first part of this review, a historical overview on the structural studies on ARs will be reported, considering even their role for site directed mutagenesis and drug design analyses. In the second part, AA2AR crystal structure will be introduced as comparison term for the data predicted in silico, to evaluate the ability of computational approaches in simulating three-dimensional models of AR subtypes. This X-ray structure will be also analysed for its use as template for the upcoming AR models. Within the text, amino acid residues will be indicated with their position within the respective AR subtype and also with the numbering system suggested by Ballesteros and Weinstein [11]. According to this method, each residue in a TM segment has a unique identifier composed of the helix number (1 to 7) and a position number. After defining the most conserved residue in each TM segment as position 50, the other amino acids are identified with increasing numbers toward the TM C-terminus, and decreasing numbers toward the TM N-terminus. For example, residue Cys77 of the AA2AR is also designated 3.25, indicating that it is twenty-five residues upstream from the conserved Arg in TM3. This nomenclature allows easier identification of the residues within the seven-helix bundle, as well as simpler association of equivalent residues of different GPCRs or of the same receptor in different species. MOLECULAR MODELING SUPPORT IN AR STUDY For several years, there have been no experimentally determined structural data for any AR and the only approach available for making structure/function correlations about these receptors has been homology modeling. This computer-assisted method is aimed at obtaining a 3D model of a protein with an unknown structure by the use of the structures of related proteins as templates in the modeling procedure. The technique is based on the assumption that, during their evolution, the

three-dimensional structure of the proteins has been more conserved than their individual sequences [12]. As consequence, if the structure of a member of a protein family is known, a sequence alignment of a series of homologue proteins can be converted into three-dimensional models. The accuracy and consistency of the protein model depends on the availability of a high-resolution structure of the 3D template and on the sequence similarity of the protein to be modeled respect to the one or those used as template. The evolution of the field of GPCRs modeling, including ARs, has depended on the availability of suitable molecular receptor templates. In fact, while it is quite simple to deduce the amino acid sequence of a protein from the DNA sequence of the gene encoding it, determining the three-dimensional molecular structure of proteins has proven to be more complex, especially for membrane proteins. As consequence, atomic resolution crystal structures of soluble proteins have been reported in a rapidly increasing number over the last years, while such result has not been achieved for membrane proteins, which have proven to be extremely difficult to crystallize for two main factors: the amphipathic nature of their surface, with a hydrophobic area in contact with membrane phospholipids and polar surface areas in contact with the aqueous phases on both sides of the membrane; the very low concentrations in tissues of the majority of membrane proteins. These technical difficulties complicate not only experimental X-ray diffraction but also NMR structure determination of GPCRs and the 3D structure of most of these receptors is still unknown [13,14]. Early AR models The first attempts to depict structural features of ARs with computational tools consisted basically in ligand-based approaches and were made by comparing and superimposing known ligands to obtain some “negative images” of receptors binding site. Such superimpositions were made by maximising the theoretical ligand-receptor binding interactions consisting in steric, electrostatic, hydrophobic, and hydrogen bonding complementarities between the bound ligand and receptor cavity. In the case of AA1R, various antagonists were compared with respect to their minimum energy conformation and molecular electrostatic potential (MEP), with the aim at clarifying key factors for AA1R affinity [15]. Analogue analyses were carried out by superimposing Ado and xanthine derivatives [16-18]. Each model assumed a common binding site for agonists and antagonists. These studies proposed also the “N6-C8” superimposition model, based on the hypothesis that the C8 substituent of xanthines binds to the same region of the receptor as the N6-substituent of Ado derivatives. The results of these studies were further considered and improved for the analysis of 3D requirements to achieve AA1R affinity and for the development of a pharmacophoric map of AA1R agonists and antagonists [19]. Ligand-based approaches to build pseudo-receptor models were used

4

again for AA1R [20] but also for other AR subtypes [21,22]. Pre-bovine rhodopsin AR models Early structural data of GPCRs came from low-resolution projection maps from cryo-electron microscopy experiments. In a paper published in 1975, Henderson and Unwin demonstrated that bacteriorhodopsin (bacRho), the light-driven proton-translocating pump from the purple membrane of Halobacterium salinarium (previously known as H. halobium), has seven TM domains that were interpreted as α-helices [23]. The idea of a 7TM α-helical architecture of GPCRs, initially based on bacRho that doesn’t act through a G-protein, was confirmed in the following years by the successful cloning of monoamine and peptide class A GPCRs showing all the characteristic 7TM signatures in hydrophobicity plot analyses. Strong support for the 7TM architecture of GPCRs came from Henderson’s group in 1993, which reported an electron projection map of visual rhodopsin, a chromoprotein in the retina acting via the transducin G-protein; the results of this study demonstrated the α-helical nature of the 7TM domains, arranged in a slightly different way respect those in bacRho [24]. New hypotheses about the structure of these membrane proteins came from cryo-microscopy studies of rhodopsin indicating the existence and the relative disposition of 7TM segments [25]. Like for other GPCRs, AR modeling was based on assumed structural homology with bacRho. This assumption was accepted despite the lack of any functional or sequence homology of bacRho with GPCRs. The structure of bacRho was resolved to 3.5 Å using electron cryo-microscopy (pdb code: 1BRD, Fig. (2) [26]) and it showed the seven α-helical TM domains pattern forming a funnel like cavity with the diameter of the extracellular side being larger than that of the intracellular side. Furthermore, the all-

trans retinal ligand covalently bound to Lys216 residue was present in this cavity.

Fig. (2). The bacRho structure resolved by Henderson using electron cryo-microscopy (pdb code 1BRD [26]). Transmembrane (TM) domains are indicated. Retinal is showed with Space-filling style. The atomic coordinates of the 7TM domains directed the modeling of the ARs, with a protocol similar to that described for other G protein-coupled receptors [27]. Briefly, AR models were generated by initially “mutating” the bacRho aminoacids to the corresponding AR residues and then by optimising the structures through rotation of helices, energy minimisation, and molecular dynamics. In 1992 IJzerman et al. published a canine AA1R model [28] and this was the first AR model built using the atomic coordinates of bacRho as template.

Fig. (3). Molecular structure of a set of compounds mentioned in this review, grouped as agonists (above) and antagonists (below).

5

This study revealed a rather distinct partition between hydrophobic and hydrophilic regions in the pore formed by the seven amphipathic α-helices. It was also highlighted the role in ligand binding of two Histidine residues, one located in TM6 and one in TM7. This result was in agreement with the first mutagenesis data on ARs, published in the same year and related to the characterization of bovine AA1R. In this work, His251 (6.52) and His278 (7.43) proved to be important for ligands binding, and their mutation to Leucine led to decrease of ligands affinity, especially in the case of His278 (7.43) [29]. The pH role in ligands binding confirmed the importance of the two Histidine residues in a test performed for AA2AR [30]. The canine AA1R model [28] was used for docking analysis of the potent

and AA1R-selective agonist N6-cyclopentylAdo and the antagonist 1,3-dipropyl-8-cyclopentylxanthine (DPCPX, Fig. (3)). This receptor model (and the ones built analogously) has some evident limits due to the similarity of its structure to the molecular architecture of bacRho and to the highly speculative nature of docking strategy, with limited knowledge from biochemical studies to guide it. In the following years more experimental evidences were made available. In 1994, two mutagenesis studies made on AA1R suggested a role of residue 270 (7.35, Isoleucine or Methionine in bovine/canine receptor) in the binding of N6-adenine-substituted compounds and the importance of residue 277 (7.42, Threonine in human/bovine receptor, Serine in canine receptor) in recognition and interaction of Ado ribose moiety [31,32]. In the same year, the building of rat AA2AR [33] and AA3R [34] models was carried out with analogue procedure respect to the canine AA1R [28]. In the first case, the receptor model was built using bacRho as template and it was employed for docking analysis of the agonist 2-(cyclohexylmethylenehydrazino)adenosine (SHA174) and the antagonist 8-(3-chlorostyryl)caffeine (CSC) (Fig. (3)) [33]. In the case of rat AA3R analysis, the model was built onto the previously developed structure of canine AA1R. The putative ligand binding region was suggested as contained between TM3, TM6, and TM7 and it was explored by docking analysis of several antagonists. The model helped in probing the role of a Histidine residue (6.52) present in AA1R/AA2AR subtypes but absent in AA3R where it is substituted by Serine [34]. The successive published AR models were built on the basis of overall characteristics of low resolution electron density map (EDM) of rhodopsin [24] and of rhodopsin family GPCR model proposed by Baldwin [35]. In 1995, Kim et al. published results of site-directed mutagenesis experiments on the AA2AR [36]. Ala-scanning revealed the essential role of some residues for ligand interaction, in particular Phe182 (5.43), Asn253 (6.55), Ile274 (7.39), and Ser281 (7.46). The key role of His250 (6.52), Ser277 (7.42), and His278 (7.43) was also confirmed. These studies were guided by an AA2AR model that was developed starting from the identification of the seven TM domains with the aid of Kyte–Doolittle hydrophobicity and Emini surface probability parameters [37]. The TM helices

were then individually built, minimised, and finally grouped together to form a helical bundle matching the overall characteristics of the rhodopsin EDM and the Baldwin GPCR template. The obtained structure (deposited as theoretical model to the Protein Data Bank with the code 1MMH) was used for a manual docking analysis of the agonist NECA (structure reported in Fig. (3)), whose binding pose was located in the region defined by the above listed residues (see Fig. (4)) [36].

Fig. (4). A. AA2AR model in complex with agonist NECA developed by Kim et al. [36] and deposited in PDB with the code 1MMH. Main residues involved in binding site description and ligand interaction are displayed and indicated. B. Schematic view of receptor-ligand interaction. In 1996 and 1997 several results of mutagenesis studies were reported related to AR residues involved in direct or indirect interaction with ligands or playing a role in allosteric regulation. It was observed that the conserved Glu1.39 (Glu16 in AA1R, Glu13 in AA2AR) is critical for agonist but not antagonist binding and molecular modeling studies suggested a proton transfer mechanism with His278 (7.43) and an indirect effect to agonist interaction with this residue [38,39]. Furthermore, a possible role of Glu13 (1.39) and His278 (7.43) residues in allosteric regulation of AA2AR was lately proposed [40]. Val84 (3.32) substitution was proved to be allowed only with other hydrophobic residues like Leucine, while

6

substitution of His250 (6.52) with other aromatic residues or Asparagine appeared to maintain (if not improve) agonist potency [41]. It was observed also that Thr88 (3.36) is important for agonists binding, even if when this residue is mutated the receptor function is preserved. Gln89 (3.37) was suggested to play an indirect role in ligand binding, while Ser281 (7.46) mutation to Asparagine showed to improve agonists affinity [42]. Additional studies were carried out to analyse the role of loop residues of AA2AR, revealing the importance of Glu151 and Glu169 (EL2) for ligands binding. Cys262 (EL3) mutation to Glycine proved to be not detrimental for ligands binding, hence demonstrating that the eventual disulfide link provided by this residue is not critical for receptor function [43]. Deletion of 95 residues from AA2AR C-terminal segment showed to attenuate desensitization (by phosphorylation) grade, with analogue effect given by mutation of Thr298 and Ser305 residues [44]. A different approach was employed to evaluate the role of Thr277 (7.42) residue in AA1R. In this case, a thermodynamic study on agonists, partial agonists, and antagonists binding to this subtype was carried out on both wild type and Thr277Ala (7.42) mutant receptors and the results confirmed the critical role of this residue for agonists and partial agonists binding [45]. These data served to direct future studies aimed at model refinement and further computational studies tried to interpret mutagenesis data in function of ligands docking pose. In 1998, the original AA3R model reported by van Galen et al. [34] was improved in the description of ligand/receptor interactions by introduction of the cross-docking methodology. This technique is aimed at simulating the reorganization of the native receptor structure induced by ligand coordination [46,47]. More in detail, AA3R model was built analogously respect to [36], by calculating Kyte-Doolittle hydrophobicity and Emini surface probability parameters for AA3R and by modeling the AA3R TM domains using rhodopsin as template. After the docking of the ligands into the binding site, the 7TM conformations were sampled (by manual adjustments) in the presence of each docked ligand to explore possible ligand-induced rearrangements of the receptor. This method gave good results in predicting local structural changes induced by ligand in the receptor binding site. The presence of antagonist within the receptor produced a simultaneous adjustment in the orientation of TM3, TM5, TM6, and TM7 [46]. Using this approach, several AA3R models were built and refined and also an AA2BR model was reported [46,48-50]. The AA3R model was also used for docking analysis of a chiral AA3R antagonist [49]. Among the last AR models built before the publication of bovine rhodopsin (bRho) crystal structure we can mention a human AA1R structure reported in 1998 by Bianucci et al. [51] and then used for further docking studies [52] and a bovine AA1R model developed using as templates the frog rhodopsin [25] and the GPCR template provided by Baldwin. The latter receptor model was used for docking analysis of AA1R antagonists following a pharmacophore study of the same

compounds [53]. In 1999-2001 several mutational analyses were reported, often accompanied by modeling studies. An example is a work related to the development an AA1R model for agonists docking study [54]. The AA2AR model previously reported by Kim et al. [36] was used as template. In this study, it was observed that Pro25Leu (1.48) mutation and substitution of Leu88 (3.33), Thr91 (3.36), and Gln92 (3.37) with Alanine reduce ligands affinity, in particular in the case of N6-unsubstituted Ado derivatives, the same for Pro86Phe (3.31) [54]. In the same study it was seen also that Gly14Thr (1.37) mutation increases agonists affinity and it was lately suggested a constitutively active form of this mutant receptor [55]. Substitution of Leu65 (2.60) with Phenylalanine proved to have no effect on ligands binding, but the mutation of this residue and of Ile69 (2.64) with smaller or hydrophilic residues were lately demonstrated to affect the receptor interaction with ligands [56]. Further studies on AA1R were related to the evaluation of the role of the Cysteine residues of AA1R. Results showed that the GPCRs conserved Cys80 (3.25) and Cys169 (EL2) are critical for receptor function and interaction with ligands, while Cys85 (3.30) could play some role in ligand interaction [57]. Finally, analysis of the Thr277Ala (7.42) mutation by using an allosteric enhancer suggested an allosteric role for this residue [58]. Additional studies described the positive effect of Asn273Tyr (7.36) mutation for the affinity of substituted Ado derivatives for AA2BR [59], the role of AA3R Thr318 and Thr319 (C-term) in phosphorylation and desensitization [60], and the one of AA3R Arg108 (3.50) and Ala229 (6.34) in coupling with G protein [61]. It can be noticed that the early structural studies on ARs were related mainly to mutagenesis data, while only in the late ‘90s the receptor models were employed for drug discovery aims. bRho-based AR models The crystallization of bRho (Fig. (5)) and its publication in 2000 (pdb code: 1F88 [62]) represented a milestone for AR computational studies, as it demonstrated beyond any doubt the 7TM structure of G protein-coupled receptors assumed several years earlier and provided a much better template for molecular modeling of ARs than the previous projection maps. Since then, a number of 3D structures of rhodopsin were solved through both X-ray diffraction (pdb codes: 1HZX [63], 1L9H [64], and 1U19 [65]) and NMR structure determination (pdb codes: 1JFP [66] and 1LN6 [67]) and for several years they were used as templates for building of ARs structures through homology modeling techniques [68-71]. bRho structure presents some key features, like the presence of the ground-state chromophore, 11-cis-retinal, which maintains the protein structure in the inactive conformation. Furthermore, Cys110 (3.25) and Cys187 (EL2) form a disulfide link resulted conserved in most GPCRs. Another relevant feature of bRho crystal structure is the presence of intra- and extra-cellular loops (the latter ones being completely solved) in addition to the N-terminal domain.

7

Fig. (5). Crystal structure of bRho published in 2000 (pdb code: 1F88 [62]). A. Global view of the membrane protein; TM domains are indicated; 11-cis-retinal ligand is displayed with Space-filling style. B. Particular of the extracellular region; N-term and EL2 domains are indicated; 11-cis-retinal and disulfide link formed by Cys110 (3.25) and Cys187 (EL2) are displayed with ball and sticks style.

The higher accuracy of bRho structural features provided by the crystal structures and the higher similarity of bRho to AR sequences if compared to bacRho allowed a more correct prediction of ARs structure. As consequence, better structural models were provided even to evaluate previous mutagenesis results that put in evidence even the possible role of EL2 residues in ligand binding. Furthermore, a number of docking studies and drug design analyses of AR ligands based on bRho-based models were reported since then. An additional improvement of AR models quality was the introduction of more sophisticated molecular dynamics experiments in structure refinement phases, in some cases considering also the presence of membrane bilayer models. Even post-docking analyses were often carried out with energy minimization and dynamics experiments. The early bRho-based AR models were built with similar approaches respect to the previous ones, by calculating Kyte-Doolittle hydrophobicity and Emini surface probability parameters for the studied AR and by modeling only the TM domains using bRho as template. Some AA1R, AA2BR, and AA3R models consisting basically in the TM domains were reported in the following years and were employed for docking analysis of agonists and antagonist compounds [72-80]. One of these studies reported an analysis of protonation state of His6.52 and His7.43 residues as critical factor for the definition of the binding site chemical-physical properties. The analysed model was used also to depict the interaction with nucleoside agonists starting from the definition of possible ribose binding modes [81,82]. Further AA3R models were similarly built, but with the addition of the EL2 segment [83,84]. In one of these studies the neoceptor concept was proposed by Jacobson et al. as method to evaluate ligand-target interaction and as an approach for the development of engineered receptors for therapeutic applications [83]. These models were used also as guide for mutagenesis studies and for the evaluation of roles of residues His95 (3.37), Lys152 (EL2), Trp243 (6.48), Leu244 (6.49), Ser247 (6.52), Asn250 (6.55), and His272 (7.43) [84]. An improved AA3R model was finally built with the addition of all the intra- and extra-cellular loops and used for docking analysis of agonists and antagonists [85]. This structure was submitted as theoretical model to the Protein Data Bank with the code 1O74 (then replaced by the model 1OEA). The rebuilding of the binding mode of the agonist 2-chloro-N6-(3-iodobenzyl)adenosine-5′-N-methylcarboxamide (Cl-IBMECA, Fig. (3)) to the AA3R (as described in the article) is shown in Fig. (6). The ligand was located in a pocket between TM3, TM5, TM6, and TM7, with the N6-N7 atoms pointing towards the space between TM5 and TM6 and the N3-ribose atoms interacting with TM3 and TM7. Residues in ligand proximity were: Leu91 (3.33), Thr94 (3.36), His95 (3.37), Gln167 (EL2), Phe168 (EL2), Phe182 (5.43), Phe239 (6.44), Trp243 (6.48), Leu246 (6.51), Asn250 (6.55), Ser271 (7.42), His272 (7.43), and Asn274 (7.45). The docking of nucleoside agonists required a movement of Trp243 (6.48) side chain that was not necessary for nucleoside antagonists and that

8

was hence considered as a possible component in the overall conformational change of ARs induced by agonist binding [84]. The rotation of Trp243 (6.48) side chain was observed even during molecular dynamics experiments reported few years later, where binding modes of an agonist and an inverse agonist were compared [86]. The model was used in the following years for further docking analyses of AA3R agonists and antagonists [87-95] even aimed at studying a possible activation pathway of the AA3R [91]. A mutational analysis guided by this model allowed to find that the AA3R residues Asn30 (1.50), Asp58 (2.50), and Asn274 (7.45) are involved in sodium regulation [96]. The role of Gln167 (EL2) in ligand interaction suggested by this model was evaluated by mutational analysis and the results suggested a possible involvement of this residue in interaction with Ado C2 moiety [97]. The model was used finally to build an AA3R homodimer on the bases of bRho dimer [98] theoretical model. The contact area between the two monomers was found at TM4-TM5 level, that is TM4 and TM5 in one monomer contacted with TM5 and TM4 in the other monomer, respectively [99]. An AA1R model was published as target structure for docking analysis of selective agonist and antagonist. This model was built with a different approach consisting firstly in a multiple alignment of the bRho and the four AR subtypes sequence, secondly in the building

of the AA1R TM domains by systematically mutating all the bRho residues to the corresponding AA1R aminoacids. Final refinement with molecular mechanics tools and addition of the EL and IL segments completed the protocol [100]. The same approach was followed to develop theoretical models of AA2AR [101] and AA2BR [102]. Binding mode of Ado derivatives resulted similar to the ones previously proposed by Kim et al. [36] and IJzerman et al. [33] for AA2AR; ligands were deeply inserted in the binding site, with the presence of several hydrophilic interactions between ribose moiety and Ser7.42-His7.43 residues and an between N6 amino group and Asn6.55. Mutagenesis studies on AA2BR published in those years led to the identification of residues whose mutation proved to take to different basal activity and response to agonists and antagonists. It was also hypothised a silent behaviour of AA2BR wt and a consequent low affinity for its ligands [103,104]. An improved study on AA2BR was reported some years later in which the receptor model was inserted in a hydrated phospholipid bilayer and refined with molecular dynamics techniques. Docking of xanthine derivatives as AA2BR antagonists followed by dynamics study was performed using this model. The results allowed to evaluate the role of binding site residues and to depict the function of Thr89 (3.36), Ser92 (3.39), His251 (6.52), and Asn282 (7.45) in ligands interaction [105].

Fig. (6). AA3R model developed by Gao et al. [85] and deposited in PDB with the code 1OEA. A. Global view of the model and ligand binding mode and location. B. Rebuilding of agonist Cl-IBMECA docking pose and (C) schematic description of ligand interaction with binding site.

9

The AA1R model reported by Giordanetto et al. [106] was used for docking analysis of agonist compounds. The binding mode suggested for these molecules resulted quite different respect to the one previously proposed by Gao et al. for AA3R [85] as these molecules presented again the ribose moiety deeply inserted in the receptor pocket, but with the adenine N6-N7 atoms pointing towards TM7 domain and not towards TM5-TM6 segments like in the above mentioned study. Further bRho-based AA1R modeling studies were then

reported, which described the building of bovine AA1R models consisting in just the TM domains [107] or in the full receptor sequence [108], with refinement procedures even including molecular dynamics approaches with membrane bilayer models [109]. All these bovine AA1R models were used for docking analysis of agonist or antagonist compounds. The building of AA2AR-D2R (dopamine D2 receptor) model [110] provided an interesting insight into a GPCR heterodimer involving an AR.

Fig. (7). AA2AR model developed by Kim et al. [111] and deposited in PDB with the code 1UPE. A. Global view of the model and ligand binding mode and location. B. Rebuilding of antagonist CGS15943 binding pose and (C) schematic description of interaction with binding site. D. Rebuilding of agonist NECA binding pose and (E) schematic description of interaction with binding site.

10

Differently respect to the previously reported AA3R homodimer [99], both AA2AR and D2R were singularly built on modified bRho templates, and then the two receptor models were docked to each other (rigid body protein docking protocol) to evaluate the possible dimer interfaces. Two possible complexing modes were found, the first one presenting TM5, TM6, and TM7 of D2R in contact with TM4 of AA2AR, the second one resembling the intradimer contact model proposed for bRho by Liang et al. [98]. Following the modeling of AA3R [85], the same research group reported the building of an AA2AR model developed with the same procedure. Even in this case the model was used for docking analysis of nucleoside agonists and antagonists and additionally for a mutational analysis as support of modeling study. Neoceptor hypothesis for Thr88Asp (3.36) mutation and its validation were also reported. Like the previous AA3R model, the AA2AR structure was submitted as theoretical model to the Protein Data Bank with the code 1UPE [111]. The rebuilding of the binding mode of the agonist adenosine-5′-N-ethylcarboxamide (NECA) and the antagonist 9-chloro-2-(2-furanyl)-(1,2,4)triazolo(1,5-c)quinazolin-5-amine (CGS15943) in the AA2AR binding site (as described in the article) is shown in Fig. (7). The ligands were placed in a site between TM3, TM5, TM6, and TM7, with the agonist NECA having a different orientation than Cl-IBMECA in Gao et al. AA3R model [85]. In fact, the agonist N6-N7 atoms pointed towards the EL2 region, the N3-ribose atom were oriented towards the intracellular environment, and the modified ribose group interacted with TM3 and TM7. Residues in ligand proximity were: Leu85 (3.33), Leu87 (3.35), Thr88 (3.36), Gln89 (3.37), Ser91 (3.39), Leu167 (EL2), Phe168 (EL2), Asn181 (5.42), Phe182 (5.43), Val186 (5.47), Phe242 (6.44), Trp246 (6.48), Leu249 (6.51), His250 (6.52), Asn253 (6.55), Ser277 (7.42), His278 (7.43), Asn274 (7.45), and Ser275 (7.46). Several of these residues were already found as defining the binding interaction scheme for NECA agonist in Kim et al. model [36]. The model was used in the following years for further docking analysis of AA2AR ligands [112,113] and for the building of an AA2AR homodimer on the above mentioned AA3R dimer developed by Kim et al. [99]. The AA2AR dimer was employed for the binding simulation of a poly(amidoamine) (PAMAM) dendrimer containing several copies of AA2AR agonist [114]. Several modeling studies on AA3R were reported by Moro and co-workers. The receptor building procedure was improved in several steps starting from a classical bRho-based protocol illustrated in [115] then updated to a so-called “ligand-based homology modeling” (LBHM) method. This approach consists basically in building of a temporary receptor model, docking of a ligand, and use of the receptor-ligand complex as template to build a final model that results improved in binding site definition [116]. These studies were always combined with antagonists docking analysis [117-126], but at the same time they allowed to refine the AA3R binding site and to well assess the structural requirements for ligands affinity.

Additional development of ligand-receptor interaction analysis was achieved by combining structure- and ligand-based approaches [127] and applying receptor-based 3D-QSAR techniques [69]. These studies were also at the base of the development of an AA2AR model [128]. Further AA3R models were reported as target for docking analysis of agonists [129,130] and antagonists [131]. A pharmacophore-based AA3R model [132] was built and refined considering structural features of AA3R ligands and then used for docking analysis of AA3R antagonists [133], even in combination with 3D-QSAR analysis [134]. The comparative analysis of the four AR subtype models by Ivanov et al. [135] completed an ideal route of bRho-based modeling by carrying out docking analysis of AR agonists and by suggesting explanations for the differences in ligands activity. The models proposed in this study were used for further docking analysis, in particular in the case of AA1R [136] and AA2BR [137,138]. The above reported AR modeling studies were analysed in a number of review articles describing the different adopted approaches and the obtained results [71,139,140], while some of these reviews focused their attention on single subtypes like for example the AA3R [68,69,115]. In some cases the modeling approaches and results were described as functionalised to drug discovery topics [141-143]. Further reviews described GPCR modeling techniques and results before and after the publication of bRho crystal structure [144-152], even in this case focused on drug design [153-159] but also on supramolecular organization of GPCRs [160-164]. Adrenergic receptors-based AR models In late 2007-mid 2008 the crystal structures of human adrenergic beta2 (β2AR) and turkey adrenergic beta1 (β1AR) receptors were reported [165-169] and they are represented in Fig. (8). The human β2AR structures reported by Cherezov et al. [166] and Hanson et al. [168] were highly engineered with the introduction of a segment of T4 Lysozyme (T4L) in the IL3 region to reduce conformational heterogeneity and facilitate crystal nucleation. Both structures were resolved in complex with a partial inverse agonist, Carazolol [166] or Timolol [168], respectively, with the ligand binding position similar to the retinal one in bRho crystal structure. These structures present the structurally conserved TM helices confirming the common motif of class A GPCRs, but in contrast to bRho their conformation is more open in particular at level of the extracellular ligand binding region. Further key differences respect to bRho are also observed, particularly in the EL2 section containing two disulfide links and an extra helix segment. The β2AR crystal structure reported by Rasmussen et al. [165] was crystallized in complex with a Fab (fragment generated from monoclonal antibody Mab5) binding the IL3 domain. This structure is well resolved at IL segments, but it lacks the extracellular regions.

11

Fig. (8). Crystal structures of human β2AR and turkey β1AR. From left to right: β2AR engineered with insertion of T4L segment and in complex with Carazolol (pdb code 2RH1 [166]), β2AR in complex with Fab (pdb code 2R4S [165]), β2AR engineered with insertion of T4L segment and in complex with Timolol (pdb code 3D4S [168]), and β1AR in complex with Cyanopindolol (pdb code 2VT4 [169]). The four crystal structures are displayed in “parallel” with a schematic representation of membrane to highlight the lacking of the extracellular regions in 2R4S structure. The turkey β1AR structure published by Warne et al. [169] in complex with the antagonist Cyanopindolol presents several mutations and segments deletion. The ligand binding mode and location are similar to those seen for β2AR in complex with Carazolol and Timolol, as well as the short helix within EL2 segment and two disulfide links. A comparative study of bRho- and β2AR-based modeling of rat AA1R and AA2AR was reported [170] that underlined the differences between bRho and β2AR templates like TM helixes packing, highly different EL2 fold, and different number of disulfide links. The effect of these features was a different ability of the bRho- and β2AR-based AR models to bind a set of analysed antagonists, in particular a different orientation of the residues in the binding pocket and a different binding location and orientation of the ligands in the bRho- and β2AR-based models. The published crystal structures of adrenergic receptors raised a lot of interest as they made available new templates for GPCR modeling, in particular in the case of receptors presenting higher similarity with adrenergic receptors if compared with bRho. On the other hand, the fact that the majority of the recent AR modeling studies were focused on ligands design and the remarkably different binding site regions in adrenergic receptors and bRho could represent critical factors for the following template choice among bRho or adrenergic receptors.

HUMAN AA2AR CRYSTAL STRUCTURE The publication in late 2008 of the human AA2AR crystal structure [4] solved this problem providing the experimental structure of one AR and a high similarity 3D template for the other three subtypes. The receptor was engineered with the introduction of a T4L segment in the IL3 region and was crystallised in complex with the antagonist 4-(2-[7-amino-2-(2-furyl) [1,2,4]-triazolo[2,3-a][1,3,5]triazin- 5-yl amino]ethyl) phenol (ZM241385, Fig. (9); ligand structure is showed in Fig. (3)). AA2AR structure partially confirms the features showed by adrenergic receptor crystals, presenting an open receptor conformation at extracellular level and an EL2 section pointing towards the extracellular environment. On the other hand, this segment lacks a defined secondary structural element and presented three disulfide links with EL1. An additional disulfide link is present within EL3 segment. The ZM241385 ligand lies perpendicular to the membrane plane, with the main receptor-ligand interactions given as H-bonds between Glu169 (EL2) carboxy and Asn253 (6.55) amido groups and ZM241385 scaffold. Additional key interaction is a π-stacking with Phe168 (EL2) aromatic ring. An H-bond network is provided by co-crystallized water molecules and connects ligand scaffold and binding site residues.

12

Fig. (9). Crystal structures of human AA2AR engineered with insertion of T4L segment and in complex with ZM241385 (pdb code 3EML [4]). A. Global vision of AA2AR crystal structure; ZM241385 ligand is indicated. B. H-bond network provided by crystallized water molecules and connecting ligand scaffold and binding site residues. C. The four disulfide links present in the EL domains; each link is indicated by a number according to Fig. (1). D. The three AA2AR residues providing the most relevant interaction with ZM241385, with the indication of the H-bonds. AA2AR-based evaluation of modeling studies and interpretation of mutagenesis data The binding site of AA2AR is located differently compared to the position of the bRho and β2AR pockets and only some TM positions are conserved in receptor-ligand interaction for all three GPCR (Fig. (10)). In particular, bRho-retinal contacts are given mainly by residues in TM3, TM5, TM6, TM7, and the central region of EL2 segment. β2AR-Carazolol interactions are given as well by residues in TM3, TM5, TM6, TM7, and additionally the C-term region of EL2. AA2AR-ZM241385 contacts are provided mainly by residues in C-term region of EL2, TM6, and partially TM3, with an additional participation of residues in EL3. The dissimilar receptor-ligand interaction pattern showed by the three GPCRs is a consequence of the ligand position and orientation and of the EL2 fold. In bRho crystal structure (in which the ligand retinal is covalently bound to TM7) the EL2 segment works as cover of the binding site that results small and narrow and almost isolated from the external environment. As consequence, the definition of bRho-based ARs binding site was often a task as there was limited space in these models to accommodate bulky ligands like for example nucleoside

agonists. As already described, the early bRho-based AR models consisted basically in the TM domains and the EL segments were not considered; this approach led obviously to a limited description of AR structural features and of receptor-ligand interaction, but it paradoxically carried to a definition of an externally opened binding site that allowed a less constrained orientation of ligands. In other studies, de novo folding simulation of EL segments (in particular the EL2 one) on the TM domains gave as result an AR binding site slightly more open respect to the case in which the original bRho EL2 conformation was taken as template. The correction of the binding pocket size with approaches like the above cited cross-docking or ligand-based homology modeling methods allowed an easier insertion of ligands. In some cases, analogue results were obtained by manually adjusting binding site residues side chains. An analysis of the docking studies performed in bRho-based AR models allows to see that there was high variability in ligands orientation within the binding pocket, even considering the docking poses of the same compound in different studies. In addition, the binding position and orientation presented by ZM241385 were only partially predicted by previous modeling studies.

13

Fig. (10). Comparison of bRho, β2AR, and AA2AR crystal structures. A. Sequence alignment of the three crystal structures indicated by the respective pdb codes 1F88 [62], 2RH1 [166], and 3EML [4]; T4L segments in β2AR, and AA2AR crystal structures were removed before the alignment; bold letters indicate residues in each GPCR found in 4.5 Å proximity from the respective co-crystallised ligand; “star”, “two dots” and “one dot” symbols indicate receptor-ligand proximity in all three, two out of three, and only one GPCR, respectively. B. Comparison of ZM241385 (light grey), Carazolol (medium grey), and Retinal (dark grey) binding mode after superimposition of AA2AR, β2AR, and bRho crystal structures; only AA2AR structure is displayed. Concerning the evaluation of residues involved in receptor-ligand interaction and/or receptor function, Table 1 reports a list of AR residues analysed with mutagenesis study and/or found as playing a role in ligand interaction or binding site definition in the published modeling studies. Furthermore, a map of AR residues studied with mutational analysis built on the AA2AR crystal structure is displayed in Fig. (11). It can be seen that some TM3 residues (corresponding to 3.32, 3.33, and 3.36 positions) were suggested as important for ligand interaction in a number of studies, even considering AR models built before bRho crystal structure publication. Still, 3.37 and 3.39 positions were indicated as important for ligand interaction by some studies without correspondence in AA2AR crystal structure (Fig. (11)). Considering again TM3 domain, the DRY motif located in the C-terminal part of the helix is conserved among GPCRs and it was very often used as alignment constrain in the preliminary steps of bRho-based AR modeling. This aminoacids triplet was observed in β2AR crystal structure to be involved in interaction with residues located within IL1, IL2, and TM5 (C-terminal part) domains. AA2AR crystal structure confirms this data presenting an H-bond network between Asp101 (3.49), Arg102 (3.50), and Thr41 (IL1). Additional interactions are observed with Tyr112 (IL2). Asp101 (3.49), Arg102 (3.50), and Thr41 (IL1) were analysed with mutagenesis and the results highlighted the importance of the presence of hydrophilic residues in these positions. Interestingly, Tyr103 (3.51) mutation to Phenylalanine led to decreased of agonist activity, revealing a possible key role for the p-hydroxy group in

Tyr3.51 side chain. This residue can be observed in AA2AR structure to coordinate two Arginines one of which located in the TM5 segment (Fig. (11)). About the EL2 segment, bRho-based AR models presented several residues in ligand proximity. Some of these residues were located in the N-terminal part of EL2 domain, hence the derived AR models showed as well some ligand-binding residues in the same position (See Table 1). AA2AR structure highlights the role of C-terminal region on EL2 in ligand interaction and in particular the function of Phe168 and Glu169; these residues (and the corresponding Phe171-Glu172 in AA1R, Phe173-Glu174 in AA2BR, and Phe168-Val169 in AA3R) were already reported by previous AR modeling studies as providing ligand interactions. It can be seen that the mutation of some of EL2 residues has an effect on agonists or antagonists binding, without being in ligand proximity. Conversely, AA2AR crystal structure put in evidence the role of Phe168 (EL2) in coordinating the ligand scaffold through the formation of a π-stacking, while mutagenesis studies were focused only on the surrounding residues even in the case of the other AR subtypes. The 5.42 and 5.43 positions were highlighted by a number of modeling simulations, while AA2AR structure confirms only the role of Asn181 (5.42) in coordinating water molecules in the binding pocket. TM6 domain presents the highest conservation by comparing bRho, β2AR, and AA2AR crystal structures involved in ligand interaction; considering the AR models, 6.48, 6.51, 6.52, and 6.55 residues were suggested as important for ligands binding.

14

Fig. (11). A. Location of residues studied with mutational analysis indicated on the AA2AR crystal structure. Light and dark spheres indicate residues mutated for only one or for more AR subtypes, respectively. B. Indication of Glu13 (1.39) and His278 (7.43) position and interaction. C. Location and H-bond networking of Asp101 (3.49), Arg102 (3.50), and Tyr193 (3.51) composing the so-called E/DRY motif conserved in GPCRs. D. Location of Thr88 (3.36), Gln89 (3.37), and Ser91 (3.39) residues. E. TM5 and TM6 residues in ligand proximity and analysed with mutagenesis. It must be noted that 6.48 position corresponds to a conserved Tryptophan residue suggested as a possible key component in the conformational change of ARs induced by agonist binding [84,86]. On the other hand, 6.52 position corresponds to Histidine in AA1R, AA2AR, and AA2BR and Serine in AA3R, suggesting a role for this residue in providing particular chemical-physical conditions related to AA3R ligands selectivity. Asn6.55 is a conserved aminoacid located in the core of the binding site. Its side chain contains an amide function that was predicted to provide H-bond acceptor (carbonyl) and donor (amino) features to interact with ligands, the first with ligands amino function like for example the N6 amino group of adenine or Ado derivatives, the second with polar oxygens or nitrogens present in the binding compound. The AA2AR crystal structure confirms this hypothesis, as the Asn253 (6.55) is observed to interact with ZM241385 scaffold using both chemical functions as H-bonding groups. EL3 residues were not predicted in playing some role in ligand interactions in the bRho- and β2AR-based AR

models, while AA2AR crystal structure presents three aminoacids (His264, Ala265, and Leu267) in ZM241385 proximity. Finally, 7.39, 7.42, 7.43, 7.45, and 7.46 positions were suggested by modeling studies as relevant for ligand interaction. This hypothesis is only partially confirmed by AA2AR structure as only Ile274 (7.39) is found in ligand proximity. The observation that the conserved Glu1.39 and His7.43 are critical for agonist but not antagonist binding and the suggestion of a possible interaction between these two residues [38-40] are both confirmed at the light of AA2AR crystal structure, since these two residues are close to each other but in a region separated from ZM241385 antagonist (Fig. (11)). It must be considered that 7.42 and 7.43 positions (Threonine/Serine and Histidine, respectively) were often linked to the interaction with Ado derivatives ribose moiety and this hypothesis was tested by a recent docking study performed on the AA2AR crystal structure and reported by Ivanov et al. [171] (see Fig. (12)).

Table 1. AR residues analysed with mutagenesis study and/or found as interacting with ligand or composing binding site in the published AR modeling studies.

Residue / Mutational results Early models Xray bRho-based models β2AR-based models

AA2AR Xray struct

AA2AR-based models

1.36 AA2BR V11I: no variation [59]

AA1R G14T: increased agonists affinity [54,55] 1.37

AA2BR A12T: no variation [59]

AA1R E16A: reduced agonists affinity [39] E16Q: reduced antagonists affinity [39] 1.39

AA2AR E13Q: reduced agonists affinity [38,40] [170,172]

15

1.46 [81,82]

1.48 AA1R P25L: reduced agonists affinity [54]

1.50 AA3R N30A: decreased affinity [61,96]

1.54 AA1R I31C: no variation [54]

AA2BR N36D: no variation [103] IL1

AA2BR T42A: decreased agonists affinity [103]

2.41 AA1R C46A/S: no variation [57]

2.45 AA1R S50A: no variation [54]

2.46 [170]

2.49 [170]

AA1R D55A: increased agonists affinity [39] 2.50

AA3R D58N: no variation [96] [81,82] [170]

2.51 AA2BR V54L: decreased agonists affinity [103]

2.53 [170] [170]

2.55 AA2BR L58V: no variation [59]

2.56 AA2BR F59L: loss of affinity [59]

2.57 [170]

2.58 [106,170] [170]

2.60 AA1R L65F: no variation [54] L65T: decreased agonists affinity [56] [170] [170]

2.61 [170] [170] [173,174]

2.64 AA1R I69S/T/V: decreased agonists affinity [56] [170] [173,174]

2.65 [170] [172]

3.25 AA1R C80A/S: no detectable binding [57]

3.26 [101]

3.27 AA1R M82F: no variation [57]

3.28 [111,170,172] [172]

3.29 [172]

AA1R C85A: no variation [57] C85S: reduced agonists affinity [57] 3.30

AA3R C88F: decreased agonists affinity [61]

AA1R P86F: reduced agonists affinity [54] 3.31

AA2BR F84L/S: decreased agonists affinity [103]

AA1R V87A: reduced antagonists affinity [54] 3.32

AA2AR V84L: marginal variation [41] [48,49,53] [89,102,106,113,143,170] [170] [175]

3.33 AA1R L88A: reduced affinity [54] [53,54] [79,82,89,101,105,106,109,111,134,170] [170] [4] [171,173-175]

3.34 [129,130,134]

AA1R T91A: reduced affinity [54]

AA2AR T88A/R/S: reduced affinity [42] T88D/E: loss of agonists binding [111]

3.36

AA3R T94A: decreased affinity [96]

[36,50,53,54]

[76-82,85,88-92,101,102,105-109,111,112,117,119,120,123,124,126,131,135,136,138,143,170]

[170] [171,172,176]

AA1R Q92A: reduced affinity [54]

AA2AR Q89A/D: increased affinity [42,111] 3.37

AA3R H95A: decreased affinity [84,85]

[54] [83-85,106,109,111,112,117,119-128,132,133,170,172]

[170,172] [172]

AA1R S93A: no variation [39]

AA2AR S90A: marginal variation [2] 3.38

AA2BR S91G: no variation [103]

AA1R S94A: no detectable binding [39] S94T: reduced antagonists affinity [39] 3.39

AA2AR S91A: marginal variation [42] [36,50] [79,90,100-102,105,135] [170]

3.40 [106]

3.43 [170]

16

3.49 AA3R D107K/N/R: marginal variation [61,96]

3.50 AA3R R108E/H/K/N: marginal variation [61]

3.51 AA3R Y109F: decreased agonists affinity [61]

4.43 AA1R A125K: no variation [54]

4.49 AA1R C131A/S: no variation [57]

4.53 AA1R S135A: no variation [2]

4.56 [111,170]

4.59 AA1R T141A: no variation [2]

4.62 AA1R F144L: no variation [54]

AA3R K152A: decreased antagonists affinity [84] [133]

AA2AR K150 [135]

AA2AR E151A/D/Q: loss of affinity [43]

AA1R W156 [170]

AA2AR E161A: increased antagonists affinity [43]

AA2BR S165 [135,137]

AA1R I167 [170]

AA2AR V164 [170]

AA1R K168 [78,136,170]

AA2BR K170 [105,172]

AA3R S165 [117,119,126]

AA1R C169A: no detectable binding [57] [170]

AA1R E170 [78] [170]

AA2AR L167 [111,113,143,170]

AA3R Q167A/E/R: decreased affinity [97] [84,85,87-95,117,119,120,126,131,135]

[176]

AA1R F171 [170] [170]

AA2AR F168 [111,143,170] [4] [171-174]

AA2BR F173 [172] [172] [172]

AA3R F168 [85,88-92,94,95,120,122,126] [175,176]

AA1R E172 [170] [170]

AA2AR E169A: loss of affinity [43] E169Q: no variation [43] [101] [170] [4] [171-174]

AA2BR E174 [172] [172] [172]

AA3R V169 [175,176]

AA2AR D170K: no variation [43]

AA3R S170 [119]

AA2AR V172 [101]

AA3R M172 [176]

EL2

AA2AR P173R: no variation [43]

5.37 [106]

5.38 [92,106] [4] [171,175,176]

5.39 [143,170]

5.41 AA2AR F180A: marginal variation [36] [106]

5.42 AA2AR N181S: reduced agonists affinity [36] [79,87,91,93,94,105,111-113,136,138,143,170,172] [4] [171,172]

AA2AR F182A: loss of affinity [36] F182Y/W: reduced agonists affinity [36] 5.43

AA3R F182A: decreased antagonists affinity [96] [46,48-50]

[79,83,85,87-92,95,105,111,113,122,127,133-135,143,170]

5.46 [170] [170]

5.47 [111,119,121,136,143]

6.34 AA3R A229E: no variation [61]

17

6.40 [170]

6.44 [85,105,111,134,136] [170]

6.47 [48,49] [73,74,108,136]

6.48 AA3R W243A/F: decreased antagonists affinity [84]

[79,82-85,88,91,93,94,105-108,111,113,117,127,132-136,143,170]

[170] [4] [171-175]

6.49 AA3R L244A: no variation [84]

6.51 [82,94,101,105,106,111,134,170] [170] [4] [171,173-175]

AA1R H251L: no variation [29]

AA2AR H250A: loss of affinity [36] H250F/Y: decreased agonists affinity [36] H250N: increased agonists affinity [41]

6.52

AA3R S247A: no variation [84]

[33,34,36,46,48-50,52,53]

[87-89,91,92,95,101,102,105,106,108,109,112,113,120-131,135,170]

[170,172] [4] [171,172]

AA2AR N253A: loss of affinity [36] 6.55

AA3R N250A: loss of affinity [84] [33,36,46,48-50]

[76-80,83-85,87-95,100-102,105,107,111-113,117,120,122,127,129-135,137,138,143,170,172]

[170,172] [4] [171-176]

AA1R C255A/S: no variation [57] 6.56

AA2AR C254A: marginal variation [36]

6.58 AA3R I253 [87] [176]

6.59 AA2AR F257A: loss of affinity [36] [105,132]

AA1R C260A/S: no variation [57]

AA1R C263A/S: no variation [57]

AA2AR C262G: no variation [43,57]

AA3R E258 [176]

AA2AR H264 [4] [171]

AA2AR A265 [4] [171]

EL3

AA2AR L267 [4]

7.35 [82] [170] [4] [171,175]

7.36 AA2BR N273Y: no variation [59] [170] [4] [172-176]

7.39 AA2AR I274A: loss of affinity [36] [79,90,91,100,106,111,135,136,170] [170] [4] [171,173-175]

7.40 [170]

AA1R T277A/S: reduced affinity [32,177] 7.42

AA2AR S277A/C: reduced agonists affinity [36,42] S277N/E/T: no variation [111]

[36,48-50,53,54]

[73,74,79-82,84,85,88-92,100-102,106,107,111,112,127,130,135,136,143]

[170,172] [171,172,176]

AA1R H278L: loss of affinity [29,178]

AA2AR H278A: loss of affinity [36] H278D/E: no variation [111] H278Y: decreased affinity [40]

7.43

AA3R H272E: decreased affinity [83,84]

[33,34,48-50,54]

[76-85,88-93,100,106,107,111,112,117,119,127,129,130,133-136,138,143,170,172]

[170] [171,172,176]

7.45 AA3R N274A: decreased affinity [96] [73,74,76,77,83,85,102,105,135] [170]

7.46 AA2AR S281A: loss of affinity [36] S281T: increased affinity [36] S281N: increased agonists affinity, decreased antagonists affinity [42]

[33,46,48,50] [73,79,81,100,127] [170]

7.49 [74,105]

7.53 AA3R Y282F: decrease of agonists affinity [61] [170]

AA2AR crystal structure-based AR models The AA2AR structure was used as target of docking studies of agonists and antagonists [171] and the results were compared with analogue studies carried out using the bRho- and the β2AR-based AA2AR models previously developed by Kim et al. [111] and Yuzlenko

et al. [170], respectively. Further docking studies were carried out with adenine based antagonists [174]. The AA2AR structure was also used as 3D template to develop rat AA2AR [173], human AA2BR [172], and human AA3R models [175,176] that were used as target for docking studies of agonists or antagonists and for

18

comparative analysis with previously built bRho- and β2AR-based analogue models. The orientation of some key residues like Asn6.55 or the conserved Phenylalanine in EL2 (Phe168 in AA2AR) worked as guide for ligands docking and in fact the results of these studies showed high conservation of receptor-ligand interaction pattern among AR subtypes. Still, AA2AR crystal structure-based AR modeling requires to well consider some structural features presented by AA2AR crystal structure. As first point, it must be remembered that while all bRho-based models (containing EL segments) were built with the presence of one disulfide link between TM3-EL1 and EL2, one of the structural features presented by the AA2AR crystal structure is the presence of three disulfide links between EL1 and EL2 and one within EL3. Hence, AA2AR-based modeling of the other three AR subtypes has to analyse the possibility to introduce additional disulfide links if and where possible. The human ARs alignment reported in Fig. (1) shows that the AA1R subtype presents the Cys80 (3.25)-Cys169 (EL2) and the Cys260 (EL3)-Cys263 (EL3) pairs, corresponding in AA2AR to Cys77 (3.25)-Cys166 (EL2) and Cys259 (EL3)-Cys262 (EL3) pairs, respectively. This data should suggest the presence of two disulfide links in AA1R. Mutagenesis studies demonstrated that AA1R Cys80 (3.25) and/or Cys169 (EL2) mutation to Alanine or Serine leads to no detectable binding of agonists or antagonists. On the other hand, analogue mutation of the EL3 Cysteines has no effect on ligands affinity [57], hence suggesting a non-critical role for the EL3 disulfide link. The AA2BR presents the Cys78 (3.25)-Cys171 (EL2) pair, corresponding to Cys77 (3.25)-Cys166 (EL2) in AA2AR. AA2BR Cys72 (EL1) matches with the AA2AR Cys71 (EL2) that takes part to another disulfide link (indicated with number 1 in Fig. (1) alignment). The counterpart Cysteine residue (Cys159 in AA2AR EL2 domain) doesn’t correspond to an identical aminoacid in AA2BR but it can be noticed that two Cysteines (Cys166 and Cys167) are present in the analogue region of AA2BR and one of these residues could work as disulfide link counterpart. Finally, AA3R subtype presents a residue conservation only for the Cys83 (3.25)-Cys166 (EL2) pair, corresponding again to Cys77 (3.25)-Cys166 (EL2) in AA2AR. A second concern comes from the evidence suggesting that the engineering of AA2AR may have moved the receptor conformation toward the activated state, as indicated by a significantly increased agonist affinity as compared to the wild type receptor while the antagonist Ki values are at wild-type AA2AR level [4,179]. An indirect proof is the docking study reported by Ivanov et al. [171] (see Fig. (12)) in which the natural agonist Ado was docked into the AA2AR crystal structure binding site with no need to significantly alter the residues side chain orientation. Hence, the fact that a high affinity antagonist bound AA2AR structure is able to well accommodate the agonist in the binding site raises a question if this crystal structure can be used “as it is” for both agonists and antagonists docking or if it needs further validation. The early AA2AR-based studies seem encouraging at least in

providing comparable receptor-ligand contact patterns that anyway resemble the original interaction scheme presented by the co-crystallized ZM241385 ligand.

Fig. (12). A. Rebuilding of Ado docking pose accordingly to Ivanov et al. [171]. The AA2AR-agonist interaction highlights the additional roles of Ser277 (7.42) and His278 (7.43) respect to the receptor-antagonist binding pattern. B. Docking studies on AA2AR-based rat AA2AR model [173]. Superimposition of original binding pose of ZM241385 (black) to the docking poses of two adenine-based derivatives (white and grey). The π-stacking between Phe163 (EL2, not showed) and the ligands scaffold is conserved, as well the H-bond interaction between the ligands free amino group and Glu164 (EL2) and Asn248 (6.55). It is possible to notate that one adenine derivative (grey) conserves the same scaffold orientation of ZM241385, while the other one (white) is oriented in opposite way. Still, the two adenine derivatives provide a nitrogen atom (N7 for the first derivative, N1 for the second one) able to give H-bond interaction with Asn248 (6.55) amino group in side chain. Finally, it must be considered the presence of co-crystallised water molecules in the AA2AR crystal structure, which partially fill the binding site and provide an H-bond network coordinating the receptor-ZM241385 interaction. Docking studies performed on AA2AR or on AA2AR-based AR models demonstrated only partial ability of docking tools to reproduce the ZM241385 binding mode in the absence of these water molecules, while the addition of docking constrains (i.e. H-bond with Asn6.55 amido group) helped to improve the

19

results. The presence of water molecules during docking analysis could help only in the case of ligands with high similarity to ZM241385, but it could be detrimental in the case of compounds with significantly different scaffolds. Hence, an accurate prediction of the position and the role of water molecules inside a binding pocket is still difficult and the use of crystallised solvent molecules can be applied only in limited cases. An alternative approach was tried by performing a docking analysis in the absence of solvent followed by the reintroduction of co-crystallized water molecules (with the exclusion of the ones giving clashes with compound atoms) during the energetical refinement of docking poses [173,174]. CONCLUSION For several years, molecular modeling studies of ARs (and the other class A GPCRs) were based on the crystal structure of bacRho or on the 3D structures of bRho (obtained by electron cryo-microscopy and in 2000 by X-ray crystallography). While the early models consisted basically on the TM domains, the following studies led to improvement of the receptors structure by progressive insertion of intra- and extra-cellular domains and through refinement stages time by time more sophisticated. In those years, structure–activity relationships and site-directed mutagenesis studies were not only target topics but also tools to refine bacRho- and then bRho-based AR models. The recently published crystal structures of adrenergic receptors and of human AA2AR confirm the helical arrangement similarity among the GPCRs, which can be explained in function of structural similarity of the interacting systems like G proteins or receptor kinases. They also indicate a high variability in the GPCRs binding site leading to receptor diversity and ligand selectivity. A comparative analysis of bRho-based AR models with the AA2AR crystal structure could lead to the conclusion that the modeling results deviated significantly from the experimental data, in particular in predicting the binding compound position and orientation, the receptor EL domains fold, and the possible roles of water molecules in receptor-ligand interaction. Despite the structural differences between in silico and crystal structure, we have evidenced how the residues playing a key role for AA2AR crystal structure in ligand interaction or in binding site definition were already predicted to play such role in the bRho-based studies, confirming the validity of homology modeling approaches. This data is supported by several docking and 3D-QSAR studies on ARs that allowed to design potent and often AR subtype selective ligands. Even if it can be discussed the different utility for rational drug design of native and extensively engineered GPCRs, without doubt the AA2AR crystal structure represents a breakthrough in X-ray field and provides a powerful tool to understand structural and functional properties of ARs.

ACKNOWLEDGEMENT This work was supported by University of Camerino (Fondo di Ricerca di Ateneo), Italian Ministry of Health (Progetto Ordinario Neurolesi, grant RF-CNM-2007-662855), and Italian Ministry of University and Research (MIUR: grant PRIN 2006). REFERENCES [1] Fredriksson, R.; Lagerstrom, M. C.; Lundin, L. G.;

Schioth, H. B. The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints. Mol. Pharmacol. 2003, 63, 1256-1272.

[2] Fredholm, B. B.; IJzerman, A. P.; Jacobson, K. A.; Klotz, K. N.; Linden, J. International Union of Pharmacology. XXV. Nomenclature and classification of adenosine receptors. Pharmacol. Rev. 2001, 53, 527-552.

[3] Fredholm, B. B.; Arslan, G.; Halldner, L.; Kull, B.; Schulte, G.; Wasserman, W. Structure and function of adenosine receptors and their genes. Naunyn-Schmiedeberg's Arch. Pharmacol. 2000, 362, 364-374.

[4] Jaakola, V. P.; Griffith, M. T.; Hanson, M. A.; Cherezov, V.; Chien, E. Y.; Lane, J. R.; IJzerman, A. P.; Stevens, R. C. The 2.6 angstrom crystal structure of a human A2A adenosine receptor bound to an antagonist. Science 2008, 322, 1211-1217.

[5] Libert, F.; Van Sande, J.; Lefort, A.; Czernilofsky, A.; Dumont, J. E.; Vassart, G.; Ensinger, H. A.; Mendla, K. D. Cloning and functional characterization of a human A1 adenosine receptor. Biochem. Biophys. Res. Commun. 1992, 187, 919-926.

[6] Pierce, K. D.; Furlong, T. J.; Selbie, L. A.; Shine, J. Molecular cloning and expression of an adenosine A2B receptor from human brain. Biochem. Biophys. Res. Commun. 1992, 187, 86-93.

[7] Salvatore, C. A.; Jacobson, M. A.; Taylor, H. E.; Linden, J.; Johnson, R. G. Molecular cloning and characterization of the human A3 adenosine receptor. Proc. Natl. Acad. Sci. U. S. A. 1993, 90, 10365-10369.

[8] Furlong, T. J.; Pierce, K. D.; Selbie, L. A.; Shine, J. Molecular characterization of a human brain adenosine A2 receptor. Brain Res. Mol. Brain Res. 1992, 15, 62-66.

[9] Volpini, R.; Costanzi, S.; Vittori, S.; Cristalli, G.; Klotz, K. N. Medicinal chemistry and pharmacology of A2B adenosine receptors. Curr. Top. Med. Chem. 2003, 3, 427-443.

[10] Colquhoun, D. Binding, gating, affinity and efficacy: the interpretation of structure-activity relationships for agonists and of the effects of mutating receptors. Br. J. Pharmacol. 1998, 125, 924–947.

[11] Ballesteros, J. A.; Weinstein, H. Integrated Methods for the Construction of Three-Dimensional Models and Computational Probing of Structure-Function Relations in G Protein-Coupled Receptors. Methods Neurosci. 1995, 25, 366-428.

[12] Costanzi, S.; Vincenzetti, S.; Cristalli, G.; Vita, A. Human cytidine deaminase: a three-dimensional homology model of a tetrameric metallo-enzyme inferred from the crystal structure of a distantly related dimeric homologue. J. Mol. Graph. Model. 2006, 25, 10-16.

[13] Kristiansen, K. Molecular mechanisms of ligand binding, signaling, and regulation within the superfamily of G-protein-coupled receptors: molecular modeling and

20

mutagenesis approaches to receptor structure and function. Pharmacol. Ther. 2004, 103, 21-80.

[14] Lomize, A. L.; Pogozheva, I. D.; Mosberg, H. I. Structural organization of G-protein-coupled receptors. J. Comput. Aided Mol. Des. 1999, 13, 325-353.

[15] van Galen, P. J.; van Vlijmen, H. W.; IJzerman, A. P.; Soudijn, W. A model for the antagonist binding site on the adenosine A1 receptor, based on steric, electrostatic, and hydrophobic properties. J. Med. Chem. 1990, 33, 1708-1713.

[16] Peet, N. P.; Lentz, N. L.; Meng, E. C.; Dudley, M. W.; Ogden, A. M.; Demeter, D. A.; Weintraub, H. J.; Bey, P. A novel synthesis of xanthines: support for a new binding mode for xanthines with respect to adenosine at adenosine receptors. J. Med. Chem. 1990, 33, 3127-3130.

[17] Quinn, R. J.; Dooley, M. J.; Escher, A.; Harden, F. A.; Jayasuriya, H. A Computer Generated Model of Adenosine Receptors Rationalising Binding and Selectivity of Receptor Ligands. Nucleosides Nucleotides Nucleic Acids 1991, 10, 1121-1124.

[18] Dooley, M. J.; Quinn, R. J. The three binding domain model of adenosine receptors: molecular modeling aspects. J. Med. Chem. 1992, 35, 211-216.

[19] Dooley, M. J.; Kono, M.; Suzuki, F. Theoretical structure-activity studies of adenosine A1 ligands: requirements for receptor affinity. Bioorg. Med. Chem. 1996, 4, 923-934.

[20] Bondavalli, F.; Botta, M.; Bruno, O.; Ciacci, A.; Corelli, F.; Fossa, P.; Lucacchini, A.; Manetti, F.; Martini, C.; Menozzi, G.; Mosti, L.; Ranise, A.; Schenone, S.; Tafi, A.; Trincavelli, M. L. Synthesis, molecular modeling studies, and pharmacological activity of selective A1 receptor antagonists. J. Med. Chem. 2002, 45, 4875-4887.

[21] Mager, P. P.; Reinhardt, R.; Richter, M.; Walther, H.; Rockel, B. Molecular simulation of 8-styrylxanthines. Drug Des. Discov. 1995, 13, 89-107.

[22] Mager, P. P.; Richter, M.; Reinhardt, R. Molecular simulation applied to 2-(N'-alkylidenehydrazino)- and 2-(N'-aralkylidenehydrazino)adenosine A2 agonists. Eur. J. Med. Chem. 1995, 30, 15-25.

[23] Henderson, R.; Unwin, P. N. Three-dimensional model of purple membrane obtained by electron microscopy. Nature 1975, 257, 28-32.

[24] Schertler, G. F.; Villa, C.; Henderson, R. Projection structure of rhodopsin. Nature 1993, 362, 770-772.

[25] Unger, V. M.; Hargrave, P. A.; Baldwin, J. M.; Schertler, G. F. Arrangement of rhodopsin transmembrane alpha-helices. Nature 1997, 389, 203-206.

[26] Henderson, R.; Baldwin, J. M.; Ceska, T. A.; Zemlin, F.; Beckmann, E.; Downing, K. H. Model for the structure of bacteriorhodopsin based on high-resolution electron cryo-microscopy. J. Mol. Biol. 1990, 213, 899-929.

[27] Hibert, M. F.; Trumpp-Kallmeyer, S.; Bruinvels, A.; Hoflack, J. Three-dimensional models of neurotransmitter G-binding protein-coupled receptors. Mol. Pharmacol. 1991, 40, 8-15.

[28] IJzerman, A. P.; Van Galen, P. J.; Jacobson, K. A. Molecular modeling of adenosine receptors. I. The ligand binding site on the A1 receptor. Drug Des. Discov. 1992, 9, 49-67.

[29] Olah, M. E.; Ren, H.; Ostrowski, J.; Jacobson, K. A.; Stiles, G. L. Cloning, expression, and characterization of the unique bovine A1 adenosine receptor. Studies on the ligand binding site by site-directed mutagenesis. J. Biol. Chem. 1992, 267, 10764-10770.

[30] Askalan, R.; Richardson, P. J. Role of histidine residues in the adenosine A2a receptor ligand binding site. J. Neurochem. 1994, 63, 1477-1484.

[31] Tucker, A. L.; Robeva, A. S.; Taylor, H. E.; Holeton, D.; Bockner, M.; Lynch, K. R.; Linden, J. A1 adenosine receptors. Two amino acids are responsible for species differences in ligand recognition. J. Biol. Chem. 1994, 269, 27900-27906.

[32] Townsend-Nicholson, A.; Schofield, P. R. A threonine residue in the seventh transmembrane domain of the human A1 adenosine receptor mediates specific agonist binding. J. Biol. Chem. 1994, 269, 2373-2376.

[33] IJzerman, A. P.; van der Wenden, E. M.; van Galen, P. J.; Jacobson, K. A. Molecular modeling of adenosine receptors. The ligand binding site on the rat adenosine A2A receptor. Eur. J. Pharmacol. 1994, 268, 95-104.

[34] van Galen, P. J.; van Bergen, A. H.; Gallo-Rodriguez, C.; Melman, N.; Olah, M. E.; IJzerman, A. P.; Stiles, G. L.; Jacobson, K. A. A binding site model and structure-activity relationships for the rat A3 adenosine receptor. Mol. Pharmacol. 1994, 45, 1101-1111.

[35] Baldwin, J. M. The probable arrangement of the helices in G protein-coupled receptors. EMBO J. 1993, 12, 1693–1703.

[36] Kim, J.; Wess, J.; van Rhee, A. M.; Schoneberg, T.; Jacobson, K. A. Site-directed mutagenesis identifies residues involved in ligand recognition in the human A2A adenosine receptor. J. Biol. Chem. 1995, 270, 13987-13997.

[37] Kyte, J.; Doolittle, R. F. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 1982, 157, 105-132.

[38] IJzerman, A. P.; Von Frijtag Drabbe Kunzel, J. K.; Kim, J.; Jiang, Q.; Jacobson, K. A. Site-directed mutagenesis of the human adenosine A2A receptor. Critical involvement of Glu13 in agonist recognition. Eur. J. Pharmacol. 1996, 310, 269-272.

[39] Barbhaiya, H.; McClain, R.; IJzerman, A.; Rivkees, S. A. Site-directed mutagenesis of the human A1 adenosine receptor: influences of acidic and hydroxy residues in the first four transmembrane domains on ligand binding. Mol. Pharmacol. 1996, 50, 1635-1642.

[40] Gao, Z. G.; Jiang, Q.; Jacobson, K. A.; IJzerman, A. P. Site-directed mutagenesis studies of human A2A adenosine receptors: involvement of glu13 and his278 in ligand binding and sodium modulation. Biochem. Pharmacol. 2000, 60, 661-668.

[41] Jiang, Q.; Lee, B. X.; Glashofer, M.; van Rhee, A. M.; Jacobson, K. A. Mutagenesis reveals structure-activity parallels between human A2A adenosine receptors and biogenic amine G protein-coupled receptors. J. Med. Chem. 1997, 40, 2588-2595.

[42] Jiang, Q.; Van Rhee, A. M.; Kim, J.; Yehle, S.; Wess, J.; Jacobson, K. A. Hydrophilic side chains in the third and seventh transmembrane helical domains of human A2A adenosine receptors are required for ligand recognition. Mol. Pharmacol. 1996, 50, 512-521.

[43] Kim, J.; Jiang, Q.; Glashofer, M.; Yehle, S.; Wess, J.; Jacobson, K. A. Glutamate residues in the second extracellular loop of the human A2A adenosine receptor are required for ligand recognition. Mol. Pharmacol. 1996, 49, 683-691.

[44] Palmer, T. M.; Stiles, G. L. Identification of an A2A adenosine receptor domain specifically responsible for mediating short-term desensitization. Biochemistry 1997, 36, 832-838.

[45] Dalpiaz, A.; Townsend-Nicholson, A.; Beukers, M. W.; Schofield, P. R.; IJzerman, A. P. Thermodynamics of

21

full agonist, partial agonist, and antagonist binding to wild-type and mutant adenosine A1 receptors. Biochem. Pharmacol. 1998, 56, 1437-1445.

[46] Moro, S.; Li, A. H.; Jacobson, K. A. Molecular modeling studies of human A3 adenosine antagonists: structural homology and receptor docking. J. Chem. Inf. Comput. Sci. 1998, 38, 1239-1248.

[47] Moro, S.; Guo, D.; Camaioni, E.; Boyer, J. L.; Harden, T. K.; Jacobson, K. A. Human P2Y1 receptor: molecular modeling and site-directed mutagenesis as tools to identify agonist and antagonist recognition sites. J. Med. Chem. 1998, 41, 1456-1466.

[48] Li, A. H.; Moro, S.; Forsyth, N.; Melman, N.; Ji, X. D.; Jacobson, K. A. Synthesis, CoMFA analysis, and receptor docking of 3,5-diacyl-2, 4-dialkylpyridine derivatives as selective A3 adenosine receptor antagonists. J. Med. Chem. 1999, 42, 706-721.

[49] Jiang, J.; Li, A. H.; Jang, S. Y.; Chang, L.; Melman, N.; Moro, S.; Ji, X.; Lobkovsky, E. B.; Clardy, J. C.; Jacobson, K. A. Chiral resolution and stereospecificity of 6-phenyl-4-phenylethynyl- 1,4-dihydropyridines as selective A3 adenosine receptor antagonists. J. Med. Chem. 1999, 42, 3055-3065.

[50] Kim, Y. C.; de Zwart, M.; Chang, L.; Moro, S.; von Frijtag Drabbe Kunzel, J. K.; Melman, N.; IJzerman, A. P.; Jacobson, K. A. Derivatives of the triazoloquinazoline adenosine antagonist (CGS 15943) having high potency at the human A2B and A3 receptor subtypes. J. Med. Chem. 1998, 41, 2835-2845.

[51] Bianucci, A. M.; Bigi, M. U.; Biagi, G.; Giorgi, I.; Livi, O.; Scartoni, V. A 3D model of the human A1 adenosine receptor. An evaluation of the binding free-energy with ligands. Drug Des. Discov. 1998, 15, 149-156.

[52] Bianucci, A. M.; Biagi, G.; Coi, A.; Giorgi, I.; Livi, O.; Pacchini, F.; Scartoni, V.; Lucacchini, A.; Costa, B. Bioisosterism, enantioselectivity, and molecular modeling of new effective N6- and/or N(9)-substituted 2-phenyl adenines and 8-aza analogs: Different binding modes to A1 adenosine receptors. Drug Dev. Res. 2001, 54, 52-65.

[53] Da Settimo, F.; Primofiore, G.; Taliani, S.; Marini, A. M.; La Motta, C.; Novellino, E.; Greco, G.; Lavecchia, A.; Trincavelli, L.; Martini, C. 3-Aryl[1,2,4]triazino[4,3-a]benzimidazol-4(10H)-ones: a new class of selective A1 adenosine receptor antagonists. J. Med. Chem. 2001, 44, 316-327.

[54] Rivkees, S. A.; Barbhaiya, H.; IJzerman, A. P. Identification of the adenine binding site of the human A1 adenosine receptor. J. Biol. Chem. 1999, 274, 3617-3621.

[55] de Ligt, R. A.; Rivkees, S. A.; Lorenzen, A.; Leurs, R.; IJzerman, A. P. A "locked-on," constitutively active mutant of the adenosine A1 receptor. Eur. J. Pharmacol. 2005, 510, 1-8.

[56] Xie, K. Q.; Cao, Y.; Zhu, X. Z. Role of the second transmembrane domain of rat adenosine A1 receptor in ligand-receptor interaction. Biochem. Pharmacol. 2006, 71, 865-871.

[57] Scholl, D. J.; Wells, J. N. Serine and alanine mutagenesis of the nine native cysteine residues of the human A1 adenosine receptor. Biochem. Pharmacol. 2000, 60, 1647-1654.

[58] Kourounakis, A.; Visser, C.; de Groote, M.; IJzerman, A. P. Differential effects of the allosteric enhancer (2-amino-4,5-dimethyl-trienyl)[3-trifluoromethyl) phenyl]methanone (PD81,723) on agonist and antagonist binding and function at the human wild-type and a

mutant (T277A) adenosine A1 receptor. Biochem. Pharmacol. 2001, 61, 137-144.

[59] Beukers, M. W.; den Dulk, H.; van Tilburg, E. W.; Brouwer, J.; IJzerman, A. P. Why are A2B receptors low-affinity adenosine receptors? Mutation of Asn273 to Tyr increases affinity of human A2B receptor for 2-(1-Hexynyl)adenosine. Mol. Pharmacol. 2000, 58, 1349-1356.

[60] Palmer, T. M.; Stiles, G. L. Identification of threonine residues controlling the agonist-dependent phosphorylation and desensitization of the rat A3 adenosine receptor. Mol. Pharmacol. 2000, 57, 539-545.

[61] Chen, A.; Gao, Z. G.; Barak, D.; Liang, B. T.; Jacobson, K. A. Constitutive activation of A3 adenosine receptors by site-directed mutagenesis. Biochem. Biophys. Res. Commun. 2001, 284, 596-601.

[62] Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima, H.; Fox, B. A.; Le Trong, I.; Teller, D. C.; Okada, T.; Stenkamp, R. E.; Yamamoto, M.; Miyano, M. Crystal structure of rhodopsin: A G protein-coupled receptor. Science 2000, 289, 739-745.

[63] Teller, D. C.; Okada, T.; Behnke, C. A.; Palczewski, K.; Stenkamp, R. E. Advances in determination of a high-resolution three-dimensional structure of rhodopsin, a model of G-protein-coupled receptors (GPCRs). Biochemistry 2001, 40, 7761-7772.

[64] Okada, T.; Fujiyoshi, Y.; Silow, M.; Navarro, J.; Landau, E. M.; Shichida, Y. Functional role of internal water molecules in rhodopsin revealed by X-ray crystallography. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 5982-5987.

[65] Okada, T.; Sugihara, M.; Bondar, A. N.; Elstner, M.; Entel, P.; Buss, V. The retinal conformation and its environment in rhodopsin in light of a new 2.2 Ǻ crystal structure. J. Mol. Biol. 2004, 342, 571-583.

[66] Yeagle, P. L.; Choi, G.; Albert, A. D. Studies on the structure of the G-protein-coupled receptor rhodopsin including the putative G-protein binding site in unactivated and activated forms. Biochemistry 2001, 40, 11932-11937.

[67] Choi, G.; Landin, J.; Galan, J. F.; Birge, R. R.; Albert, A. D.; Yeagle, P. L. Structural studies of metarhodopsin II, the activated form of the G-protein coupled receptor, rhodopsin. Biochemistry 2002, 41, 7318-7324.

[68] Dal Ben, D.; Lambertucci, C.; Vittori, S.; Volpini, R.; Cristalli, G. GPCRs as therapeutic targets: a view on adenosine receptors structure and functions, and molecular modeling support. J. Iranian Chem. Soc. 2005, 2, 176-188.

[69] Moro, S.; Spalluto, G.; Jacobson, K. A. Techniques: Recent developments in computer-aided engineering of GPCR ligands using the human adenosine A3 receptor as an example. Trends Pharmacol. Sci. 2005, 26, 44-51.

[70] Moro, S.; Deflorian, F.; Bacilieri, M.; Spalluto, G. Homology modeling as attractive tool to inspect GPCR structural plasticity. Curr. Pharm. Des. 2006, 12, 2175-2185.

[71] Martinelli, A.; Tuccinardi, T. Molecular modeling of adenosine receptors: new results and trends. Med. Res. Rev. 2008, 28, 247-277.

[72] Baraldi, P. G.; Cacciari, B.; Moro, S.; Romagnoli, R.; Ji, X.; Jacobson, K. A.; Gessi, S.; Borea, P. A.; Spalluto, G. Fluorosulfonyl- and bis-(β-chloroethyl)amino-phenylamino functionalized pyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine derivatives: irreversible antagonists at the human A3 adenosine receptor and molecular modeling studies. J. Med. Chem. 2001, 44, 2735-2742.

22

[73] Baraldi, P. G.; Cacciari, B.; Moro, S.; Spalluto, G.; Pastorin, G.; Da Ros, T.; Klotz, K. N.; Varani, K.; Gessi, S.; Borea, P. A. Synthesis, biological activity, and molecular modeling investigation of new pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine derivatives as human A3 adenosine receptor antagonists. J. Med. Chem. 2002, 45, 770-780.

[74] Maconi, A.; Pastorin, G.; Da Ros, T.; Spalluto, G.; Gao, Z. G.; Jacobson, K. A.; Baraldi, P. G.; Cacciari, B.; Varani, K.; Moro, S.; Borea, P. A. Synthesis, biological properties, and molecular modeling investigation of the first potent, selective, and water-soluble human A3 adenosine receptor antagonist. J. Med. Chem. 2002, 45, 3579-3582.

[75] Pastorin, G.; Da Ros, T.; Spalluto, G.; Deflorian, F.; Moro, S.; Cacciari, B.; Baraldi, P. G.; Gessi, S.; Varani, K.; Borea, P. A. Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine derivatives as adenosine receptor antagonists. Influence of the N5 substituent on the affinity at the human A3 and A2B adenosine receptor subtypes: a molecular modeling investigation. J. Med. Chem. 2003, 46, 4287-4296.

[76] Costanzi, S.; Lambertucci, C.; Vittori, S.; Volpini, R.; Cristalli, G. 2- and 8-alkynyladenosines: conformational studies and docking to human adenosine A3 receptor can explain their different biological behavior. J. Mol. Graph. Model. 2003, 21, 253-262.

[77] Cristalli, G.; Costanzi, S.; Lambertucci, C.; Taffi, S.; Vittori, S.; Volpini, R. Purine and deazapurine nucleosides: synthetic approaches, molecular modelling and biological activity. Farmaco 2003, 58, 193-204.

[78] Kehraus, S.; Gorzalka, S.; Hallmen, C.; Iqbal, J.; Muller, C. E.; Wright, A. D.; Wiese, M.; Konig, G. M. Novel amino acid derived natural products from the ascidian Atriolum robustum: identification and pharmacological characterization of a unique adenosine derivative. J. Med. Chem. 2004, 47, 2243-2255.

[79] Tuccinardi, T.; Ortore, G.; Manera, C.; Saccomanni, G.; Martinelli, A. Adenosine receptor modelling. A1/A2A selectivity. Eur. J. Med. Chem. 2006, 41, 321-329.

[80] Franchetti, P.; Cappellacci, L.; Vita, P.; Petrelli, R.; Lavecchia, A.; Kachler, S.; Klotz, K. N.; Marabese, I.; Luongo, L.; Maione, S.; Grifantini, M. N6-Cycloalkyl- and N6-bicycloalkyl-C5'(C2')-modified adenosine derivatives as high-affinity and selective agonists at the human A1 adenosine receptor with antinociceptive effects in mice. J. Med. Chem. 2009, 52, 2393-2406.

[81] Gutierrez-de-Teran, H.; Centeno, N. B.; Pastor, M.; Sanz, F. Novel approaches for modeling of the A1 adenosine receptor and its agonist binding site. Proteins 2004, 54, 705-715.

[82] Gutierrez-de-Teran, H.; Pastor, M.; Centeno, N. B.; Aqvist, J.; Sanz, F. Comparative analysis of putative agonist-binding modes in the human A1 adenosine receptor. ChemBioChem 2004, 5, 841-849.

[83] Jacobson, K. A.; Gao, Z. G.; Chen, A.; Barak, D.; Kim, S. A.; Lee, K.; Link, A.; Rompaey, P. V.; van Calenbergh, S.; Liang, B. T. Neoceptor concept based on molecular complementarity in GPCRs: a mutant adenosine A3 receptor with selectively enhanced affinity for amine-modified nucleosides. J. Med. Chem. 2001, 44, 4125-4136.

[84] Gao, Z. G.; Chen, A.; Barak, D.; Kim, S. K.; Muller, C. E.; Jacobson, K. A. Identification by site-directed mutagenesis of residues involved in ligand recognition and activation of the human A3 adenosine receptor. J. Biol. Chem. 2002, 277, 19056-19063.

[85] Gao, Z. G.; Kim, S. K.; Biadatti, T.; Chen, W.; Lee, K.; Barak, D.; Kim, S. G.; Johnson, C. R.; Jacobson, K. A. Structural determinants of A3 adenosine receptor activation: nucleoside ligands at the agonist/antagonist boundary. J. Med. Chem. 2002, 45, 4471-4484.

[86] Hallmen, C.; Wiese, M. Molecular dynamics simulation of the human adenosine A3 receptor: agonist induced conformational changes of Trp243. J. Comput. Aided Mol. Des. 2006, 20, 673-684.

[87] Jung, K. Y.; Kim, S. K.; Gao, Z. G.; Gross, A. S.; Melman, N.; Jacobson, K. A.; Kim, Y. C. Structure-activity relationships of thiazole and thiadiazole derivatives as potent and selective human adenosine A3 receptor antagonists. Bioorg. Med. Chem. 2004, 12, 613-623.

[88] Tchilibon, S.; Kim, S. K.; Gao, Z. G.; Harris, B. A.; Blaustein, J. B.; Gross, A. S.; Duong, H. T.; Melman, N.; Jacobson, K. A. Exploring distal regions of the A3 adenosine receptor binding site: sterically constrained N6-(2-phenylethyl)adenosine derivatives as potent ligands. Bioorg. Med. Chem. 2004, 12, 2021-2034.

[89] Ohno, M.; Gao, Z. G.; Van Rompaey, P.; Tchilibon, S.; Kim, S. K.; Harris, B. A.; Gross, A. S.; Duong, H. T.; Van Calenbergh, S.; Jacobson, K. A. Modulation of adenosine receptor affinity and intrinsic efficacy in adenine nucleosides substituted at the 2-position. Bioorg. Med. Chem. 2004, 12, 2995-3007.

[90] Gao, Z. G.; Duong, H. T.; Sonina, T.; Kim, S. K.; Van Rompaey, P.; Van Calenbergh, S.; Mamedova, L.; Kim, H. O.; Kim, M. J.; Kim, A. Y.; Liang, B. T.; Jeong, L. S.; Jacobson, K. A. Orthogonal activation of the reengineered A3 adenosine receptor (neoceptor) using tailored nucleoside agonists. J. Med. Chem. 2006, 49, 2689-2702.

[91] Kim, S. K.; Gao, Z. G.; Jeong, L. S.; Jacobson, K. A. Docking studies of agonists and antagonists suggest an activation pathway of the A3 adenosine receptor. J. Mol. Graph. Model. 2006, 25, 562-577.

[92] Cosyn, L.; Palaniappan, K. K.; Kim, S. K.; Duong, H. T.; Gao, Z. G.; Jacobson, K. A.; Van Calenbergh, S. 2-triazole-substituted adenosines: a new class of selective A3 adenosine receptor agonists, partial agonists, and antagonists. J. Med. Chem. 2006, 49, 7373-7383.

[93] Melman, A.; Wang, B.; Joshi, B. V.; Gao, Z. G.; Castro, S.; Heller, C. L.; Kim, S. K.; Jeong, L. S.; Jacobson, K. A. Selective A3 adenosine receptor antagonists derived from nucleosides containing a bicyclo[3.1.0]hexane ring system. Bioorg. Med. Chem. 2008, 16, 8546-8556.

[94] Priego, E. M.; Perez-Perez, M. J.; von Frijtag Drabbe Kuenzel, J. K.; de Vries, H.; IJzerman, A. P.; Camarasa, M. J.; Martin-Santamaria, S. Selective human adenosine A3 antagonists based on pyrido[2,1-f]purine-2,4-diones: novel features of hA3 antagonist binding. ChemMedChem 2008, 3, 111-119.

[95] Wei, J.; Li, H.; Qu, W.; Gao, Q. Molecular docking study of A3 adenosine receptor antagonists and pharmacophore-based drug design. Neurochem. Int. 2009, 55, 637-642.

[96] Gao, Z. G.; Kim, S. K.; Gross, A. S.; Chen, A.; Blaustein, J. B.; Jacobson, K. A. Identification of essential residues involved in the allosteric modulation of the human A3 adenosine receptor. Mol. Pharmacol. 2003, 63, 1021-1031.

[97] Duong, H. T.; Gao, Z. G.; Jacobson, K. A. Nucleoside modification and concerted mutagenesis of the human A3 adenosine receptor to probe interactions between the 2-position of adenosine analogs and Gln167 in the second

23

extracellular loop. Nucleosides Nucleotides Nucleic Acids 2005, 24, 1507-1517.

[98] Liang, Y.; Fotiadis, D.; Filipek, S.; Saperstein, D. A.; Palczewski, K.; Engel, A. Organization of the G protein-coupled receptors rhodopsin and opsin in native membranes. J. Biol. Chem. 2003, 278, 21655-21662.

[99] Kim, S. K.; Jacobson, K. A. Computational prediction of homodimerization of the A3 adenosine receptor. J. Mol. Graph. Model. 2006, 25, 549-561.

[100] Ivanov, A. A.; Baskin, I. I.; Palyulin, V. A.; Zefirov, N. S. Molecular modeling the human A1 adenosine receptor and study of the mechanisms of its selective ligand binding. Dokl. Biochem. Biophys. 2002, 386, 271-274.

[101] Ivanov, A. A.; Baskin, II; Palyulin, V. A.; Zefirov, N. S. Molecular modeling of the human A2A adenosine receptor. Dokl. Biochem. Biophys. 2003, 389, 94-97.

[102] Ivanov, A. A.; Baskin, I. I.; Palyulin, V. A.; Baraldi, P. G.; Zefirov, N. S. Molecular modelling of the human A2B adenosine receptor and an analysis of the binding modes of its selective ligands. Mendeleev Commun. 2002, 12, 211-212.

[103] Beukers, M. W.; van Oppenraaij, J.; van der Hoorn, P. P.; Blad, C. C.; den Dulk, H.; Brouwer, J.; IJzerman, A. P. Random mutagenesis of the human adenosine A2B receptor followed by growth selection in yeast. Identification of constitutively active and gain of function mutations. Mol. Pharmacol. 2004, 65, 702-710.

[104] Li, Q.; Ye, K.; Blad, C. C.; den Dulk, H.; Brouwer, J.; IJzerman, A. P.; Beukers, M. W. ZM241385, DPCPX, MRS1706 are inverse agonists with different relative intrinsic efficacies on constitutively active mutants of the human adenosine A2B receptor. J. Pharmacol. Exp. Ther. 2007, 320, 637-645.

[105] Ivanov, A. A.; Baskin, II; Palyulin, V. A.; Piccagli, L.; Baraldi, P. G.; Zefirov, N. S. Molecular modeling and molecular dynamics simulation of the human A2B adenosine receptor. The study of the possible binding modes of the A2B receptor antagonists. J. Med. Chem. 2005, 48, 6813-6820.

[106] Giordanetto, F.; Fossa, P.; Menozzi, G.; Schenone, S.; Bondavalli, F.; Ranise, A.; Mosti, L. Exploring the molecular basis of selectivity in A1 adenosine receptors agonists: a case study. J. Comput. Aided Mol. Des. 2003, 17, 39-51.

[107] Cappellacci, L.; Franchetti, P.; Pasqualini, M.; Petrelli, R.; Vita, P.; Lavecchia, A.; Novellino, E.; Costa, B.; Martini, C.; Klotz, K. N.; Grifantini, M. Synthesis, biological evaluation, and molecular modeling of ribose-modified adenosine analogues as adenosine receptor agonists. J. Med. Chem. 2005, 48, 1550-1562.

[108] Giorgi, I.; Bianucci, A. M.; Biagi, G.; Livi, O.; Scartoni, V.; Leonardi, M.; Pietra, D.; Coi, A.; Massarelli, I.; Nofal, F. A.; Fiamingo, F. L.; Anastasi, P.; Giannini, G. Synthesis, biological activity and molecular modelling of new trisubstituted 8-azaadenines with high affinity for A1 adenosine receptors. Eur. J. Med. Chem. 2007, 42, 1-9.

[109] Tuccinardi, T.; Schenone, S.; Bondavalli, F.; Brullo, C.; Bruno, O.; Mosti, L.; Zizzari, A. T.; Tintori, C.; Manetti, F.; Ciampi, O.; Trincavelli, M. L.; Martini, C.; Martinelli, A.; Botta, M. Substituted pyrazolo[3,4-b]pyridines as potent A1 adenosine antagonists: synthesis, biological evaluation, and development of an A1 bovine receptor model. ChemMedChem 2008, 3, 898-913.

[110] Canals, M.; Marcellino, D.; Fanelli, F.; Ciruela, F.; de Benedetti, P.; Goldberg, S. R.; Neve, K.; Fuxe, K.; Agnati, L. F.; Woods, A. S.; Ferre, S.; Lluis, C.; Bouvier,

M.; Franco, R. Adenosine A2A-dopamine D2 receptor-receptor heteromerization: qualitative and quantitative assessment by fluorescence and bioluminescence energy transfer. J. Biol. Chem. 2003, 278, 46741-46749.

[111] Kim, S. K.; Gao, Z. G.; Van Rompaey, P.; Gross, A. S.; Chen, A.; Van Calenbergh, S.; Jacobson, K. A. Modeling the adenosine receptors: comparison of the binding domains of A2A agonists and antagonists. J. Med. Chem. 2003, 46, 4847-4859.

[112] Jacobson, K. A.; Ohno, M.; Duong, H. T.; Kim, S. K.; Tchilibon, S.; Cesnek, M.; Holý, A.; Gao, Z. G. A Neoceptor Approach to Unraveling Microscopic Interactions between the Human A2A Adenosine Receptor and Its Agonists. Chem. Biol. 2005, 12, 237-247.

[113] Parenti, M. D.; Fioravanzo, E.; Mabilia, M.; Gallo, G.; Ciacci, A. Induced fit and pharmacophore generation approach applied to A2A adenosine receptor antagonists. ARKIVOC 2006, VIII, 74-82.

[114] Ivanov, A. A.; Jacobson, K. A. Molecular modeling of a PAMAM-CGS21680 dendrimer bound to an A2A adenosine receptor homodimer. Bioorg. Med. Chem. Lett. 2008, 18, 4312-4315.

[115] Moro, S.; Deflorian, F.; Spalluto, G.; Pastorin, G.; Cacciari, B.; Kim, S. K.; Jacobson, K. A. Demystifying the three dimensional structure of G protein-coupled receptors (GPCRs) with the aid of molecular modeling. Chem. Commun. (Camb.) 2003, 2949-2956.

[116] Moro, S.; Deflorian, F.; Bacilieri, M.; Spalluto, G. Ligand-based homology modeling as attractive tool to inspect GPCR structural plasticity. Curr. Pharm. Des. 2006, 12, 2175-2185.

[117] Colotta, V.; Catarzi, D.; Varano, F.; Calabri, F. R.; Lenzi, O.; Filacchioni, G.; Martini, C.; Trincavelli, L.; Deflorian, F.; Moro, S. 1,2,4-triazolo[4,3-a]quinoxalin-1-one moiety as an attractive scaffold to develop new potent and selective human A3 adenosine receptor antagonists: synthesis, pharmacological, and ligand-receptor modeling studies. J. Med. Chem. 2004, 47, 3580-3590.

[118] Pastorin, G.; Bolcato, C.; Cacciari, B.; Kachler, S.; Klotz, K. N.; Montopoli, C.; Moro, S.; Spalluto, G. Synthesis, biological and modeling studies of 1,3-di-n-propyl-2,4-dioxo-6-methyl-8-(substituted) 1,2,3,4-tetrahydro [1,2,4]-triazolo [3,4-f]-purines as adenosine receptor antagonists. Farmaco 2005, 60, 643-651.

[119] Catarzi, D.; Colotta, V.; Varano, F.; Lenzi, O.; Filacchioni, G.; Trincavelli, L.; Martini, C.; Montopoli, C.; Moro, S. 1,2,4-Triazolo[1,5-a]quinoxaline as a versatile tool for the design of selective human A3 adenosine receptor antagonists: synthesis, biological evaluation, and molecular modeling studies of 2-(hetero)aryl- and 2-carboxy-substituted derivatives. J. Med. Chem. 2005, 48, 7932-7945.

[120] Lenzi, O.; Colotta, V.; Catarzi, D.; Varano, F.; Filacchioni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.; Varani, K.; Marighetti, F.; Morizzo, E.; Moro, S. 4-amido-2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-ones as new potent and selective human A3 adenosine receptor antagonists. synthesis, pharmacological evaluation, and ligand-receptor modeling studies. J. Med. Chem. 2006, 49, 3916-3925.

[121] Pastorin, G.; Da Ros, T.; Bolcato, C.; Montopoli, C.; Moro, S.; Cacciari, B.; Baraldi, P. G.; Varani, K.; Borea, P. A.; Spalluto, G. Synthesis and biological studies of a new series of 5-heteroarylcarbamoylaminopyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidines as human A3 adenosine receptor antagonists. Influence of the

24

heteroaryl substituent on binding affinity and molecular modeling investigations. J. Med. Chem. 2006, 49, 1720-1729.

[122] Morizzo, E.; Capelli, F.; Lenzi, O.; Catarzi, D.; Varano, F.; Filacchioni, G.; Vincenzi, F.; Varani, K.; Borea, P. A.; Colotta, V.; Moro, S. Scouting human A3 adenosine receptor antagonist binding mode using a molecular simplification approach: from triazoloquinoxaline to a pyrimidine skeleton as a key study. J. Med. Chem. 2007, 50, 6596-6606.

[123] Colotta, V.; Catarzi, D.; Varano, F.; Capelli, F.; Lenzi, O.; Filacchioni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.; Pugliese, A. M.; Pedata, F.; Schiesaro, A.; Morizzo, E.; Moro, S. New 2-arylpyrazolo[3,4-c]quinoline derivatives as potent and selective human A3 adenosine receptor antagonists. synthesis, pharmacological evaluation, and ligand-receptor modeling studies. J. Med. Chem. 2007, 50, 4061-4074.

[124] Colotta, V.; Catarzi, D.; Varano, F.; Lenzi, O.; Filacchioni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.; Traini, C.; Pugliese, A. M.; Pedata, F.; Morizzo, E.; Moro, S. Synthesis, ligand-receptor modeling studies and pharmacological evaluation of novel 4-modified-2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives as potent and selective human A3 adenosine receptor antagonists. Bioorg. Med. Chem. 2008, 16, 6086-6102.

[125] Bolcato, C.; Cusan, C.; Pastorin, G.; Spalluto, G.; Cacciari, B.; Klotz, K. N.; Morizzo, E.; Moro, S. Pyrazolo-triazolo-pyrimidines as adenosine receptor antagonists: Effect of the N-5 bond type on the affinity and selectivity at the four adenosine receptor subtypes. Purinergic Signal. 2008, 4, 39-46.

[126] Colotta, V.; Lenzi, O.; Catarzi, D.; Varano, F.; Filacchioni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.; Pugliese, A. M.; Traini, C.; Pedata, F.; Morizzo, E.; Moro, S. Pyrido[2,3-e]-1,2,4-triazolo[4,3-a]pyrazin-1-one as a new scaffold to develop potent and selective human A3 adenosine receptor antagonists. Synthesis, pharmacological evaluation, and ligand-receptor modeling studies. J. Med. Chem. 2009, 52, 2407-2419.

[127] Moro, S.; Braiuca, P.; Deflorian, F.; Ferrari, C.; Pastorin, G.; Cacciari, B.; Baraldi, P. G.; Varani, K.; Borea, P. A.; Spalluto, G. Combined target-based and ligand-based drug design approach as a tool to define a novel 3D-pharmacophore model of human A3 adenosine receptor antagonists: pyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine derivatives as a key study. J. Med. Chem. 2005, 48, 152-162.

[128] Michielan, L.; Bacilieri, M.; Schiesaro, A.; Bolcato, C.; Pastorin, G.; Spalluto, G.; Cacciari, B.; Klotz, K. N.; Kaseda, C.; Moro, S. Linear and nonlinear 3D-QSAR approaches in tandem with ligand-based homology modeling as a computational strategy to depict the pyrazolo-triazolo-pyrimidine antagonists binding site of the human adenosine A2A receptor. J. Chem. Inf. Model. 2008, 48, 350-363.

[129] Dal Ben, D.; Lambertucci, C.; Taffi, S.; Vittori, S.; Volpini, R.; Cristalli, G.; Klotz, K. N. Molecular modelling study of 2-phenylethynyladenosine (PEAdo) derivatives as highly selective A3 adenosine receptor ligands. Purinergic Signal. 2006, 2, 589-594.

[130] Volpini, R.; Dal Ben, D.; Lambertucci, C.; Taffi, S.; Vittori, S.; Klotz, K. N.; Cristalli, G. N6-methoxy-2-alkynyladenosine derivatives as highly potent and selective ligands at the human A3 adenosine receptor. J. Med. Chem. 2007, 50, 1222-1230.

[131] Colotta, V.; Capelli, F.; Lenzi, O.; Catarzi, D.; Varano, F.; Poli, D.; Vincenzi, F.; Varani, K.; Borea, P. A.; Dal

Ben, D.; Volpini, R.; Cristalli, G.; Filacchioni, G. Novel potent and highly selective human A3 adenosine receptor antagonists belonging to the 4-amido-2-arylpyrazolo[3,4-c]quinoline series: molecular docking analysis and pharmacological studies. Bioorg. Med. Chem. 2009, 17, 401-410.

[132] Tafi, A.; Bernardini, C.; Botta, M.; Corelli, F.; Andreini, M.; Martinelli, A.; Ortore, G.; Baraldi, P. G.; Fruttarolo, F.; Borea, P. A.; Tuccinardi, T. Pharmacophore based receptor modeling: the case of adenosine A3 receptor antagonists. An approach to the optimization of protein models. J. Med. Chem. 2006, 49, 4085-4097.

[133] Da Settimo, F.; Primofiore, G.; Taliani, S.; Marini, A. M.; La Motta, C.; Simorini, F.; Salerno, S.; Sergianni, V.; Tuccinardi, T.; Martinelli, A.; Cosimelli, B.; Greco, G.; Novellino, E.; Ciampi, O.; Trincavelli, M. L.; Martini, C. 5-amino-2-phenyl[1,2,3]triazolo[1,2-a][1,2,4]benzotriazin-1-one: a versatile scaffold to obtain potent and selective A3 adenosine receptor antagonists. J. Med. Chem. 2007, 50, 5676-5684.

[134] Baraldi, P. G.; Preti, D.; Tabrizi, M. A.; Romagnoli, R.; Saponaro, G.; Baraldi, S.; Botta, M.; Bernardini, C.; Tafi, A.; Tuccinardi, T.; Martinelli, A.; Varani, K.; Borea, P. A. Structure-activity relationship studies of a new series of imidazo[2,1-f]purinones as potent and selective A3 adenosine receptor antagonists. Bioorg. Med. Chem. 2008, 16, 10281-10294.

[135] Ivanov, A. A.; Palyulin, V. A.; Zefirov, N. S. Computer aided comparative analysis of the binding modes of the adenosine receptor agonists for all known subtypes of adenosine receptors. J. Mol. Graph. Model. 2007, 25, 740-754.

[136] Palaniappan, K. K.; Gao, Z. G.; Ivanov, A. A.; Greaves, R.; Adachi, H.; Besada, P.; Kim, H. O.; Kim, A. Y.; Choe, S. A.; Jeong, L. S.; Jacobson, K. A. Probing the binding site of the A1 adenosine receptor reengineered for orthogonal recognition by tailored nucleosides. Biochemistry 2007, 46, 7437-7448.

[137] Adachi, H.; Palaniappan, K. K.; Ivanov, A. A.; Bergman, N.; Gao, Z. G.; Jacobson, K. A. Structure-activity relationships of 2,N6,5'-substituted adenosine derivatives with potent activity at the A2B adenosine receptor. J. Med. Chem. 2007, 50, 1810-1827.

[138] Ivanov, A. A.; Wang, B.; Klutz, A. M.; Chen, V. L.; Gao, Z. G.; Jacobson, K. A. Probing distal regions of the A2B adenosine receptor by quantitative structure-activity relationship modeling of known and novel agonists. J. Med. Chem. 2008, 51, 2088-2099.

[139] Jacobson, K. A.; Kim, S. K.; Costanzi, S.; Gao, Z. G. Purine receptors: gpcr structure and agonist design. Mol. Interv. 2004, 4, 337-347.

[140] Costanzi, S.; Ivanov, A. A.; Tikhonova, I. G.; Jacobson, K. A. Structure and Function of G protein-coupled receptors studied using sequence analysis, molecular modeling and receptor engineering: adenosine receptors. Front. Drug Des. Discov. 2007, 3, 63-79.

[141] Moro, S.; Gao, Z. G.; Jacobson, K. A.; Spalluto, G. Progress in the pursuit of therapeutic adenosine receptor antagonists. Med. Res. Rev. 2006, 26, 131-159.

[142] Cristalli, G.; Cacciari, B.; Dal Ben, D.; Lambertucci, C.; Moro, S.; Spalluto, G.; Volpini, R. Highlights on the development of A2A adenosine receptor agonists and antagonists. ChemMedChem 2007, 2, 260-281.

[143] Cristalli, G.; Lambertucci, C.; Marucci, G.; Volpini, R.; Dal Ben, D. A2A adenosine receptor and its modulators: overview on a druggable GPCR and on structure-activity relationship analysis and binding requirements of

25

agonists and antagonists. Curr. Pharm. Des. 2008, 14, 1525-1552.

[144] Muller, G. Towards 3D structures of G protein-coupled receptors: a multidisciplinary approach. Curr. Med. Chem. 2000, 7, 861-888.

[145] Gershengorn, M. C.; Osman, R. Minireview: Insights into G protein-coupled receptor function using molecular models. Endocrinology 2001, 142, 2-10.

[146] Shacham, S.; Topf, M.; Avisar, N.; Glaser, F.; Marantz, Y.; Bar-Haim, S.; Noiman, S.; Naor, Z.; Becker, O. M. Modeling the 3D structure of GPCRs from sequence. Med. Res. Rev. 2001, 21, 472-483.

[147] Archer, E.; Maigret, B.; Escrieut, C.; Pradayrol, L.; Fourmy, D. Rhodopsin crystal: new template yielding realistic models of G-protein-coupled receptors? Trends Pharmacol. Sci. 2003, 24, 36-40.

[148] Bosch, L.; Iarriccio, L.; Garriga, P. New prospects for drug discovery from structural studies of rhodopsin. Curr. Pharm. Des. 2005, 11, 2243-2256.

[149] Kristiansen, K. Molecular mechanisms of ligand binding, signaling, and regulation within the superfamily of G-protein-coupled receptors: molecular modeling and mutagenesis approaches to receptor structure and function. Pharmacol. Ther. 2004, 103, 21-80.

[150] Fanelli, F.; De Benedetti, P. G. Computational modeling approaches to structure-function analysis of G protein-coupled receptors. Chem. Rev. 2005, 105, 3297-3351.

[151] Paiva, A. C.; Oliveira, L.; Horn, F.; Bywater, R. P.; Vriend, G. Modeling GPCRs. Ernst Schering Found. Symp. Proc. 2006, 23-47.

[152] Shi, L.; Javitch, J. A. A role for information collection, management, and integration in structure-function studies of G-protein coupled receptors. Curr. Pharm. Des. 2006, 12, 1771-1783.

[153] Poulsen, S. A.; Quinn, R. J. Adenosine receptors: new opportunities for future drugs. Bioorg. Med. Chem. 1998, 6, 619-641.

[154] Flower, D. R. Modelling G-protein-coupled receptors for drug design. Biochim. Biophys. Acta 1999, 1422, 207-234.

[155] Klabunde, T.; Hessler, G. Drug design strategies for targeting G-protein-coupled receptors. ChemBioChem 2002, 3, 928-944.

[156] Becker, O. M.; Shacham, S.; Marantz, Y.; Noiman, S. Modeling the 3D structure of GPCRs: advances and application to drug discovery. Curr. Opin. Drug Discov. Devel. 2003, 6, 353-361.

[157] Patny, A.; Desai, P. V.; Avery, M. A. Homology modeling of G-protein-coupled receptors and implications in drug design. Curr. Med. Chem. 2006, 13, 1667-1691.

[158] Lundstrom, K. Latest development in drug discovery on G protein-coupled receptors. Curr. Protein Pept. Sci. 2006, 7, 465-470.

[159] Goddard, W. A., 3rd; Abrol, R. 3-Dimensional structures of G protein-coupled receptors and binding sites of agonists and antagonists. J. Nutr. 2007, 137, 1528S-1538S; discussion 1548S.

[160] Filizola, M.; Weinstein, H. The study of G-protein coupled receptor oligomerization with computational modeling and bioinformatics. FEBS J. 2005, 272, 2926-2938.

[161] Fanelli, F.; De Benedetti, P. G. Inactive and active states and supramolecular organization of GPCRs: insights from computational modeling. J. Comput. Aided Mol. Des. 2006, 20, 449-461.

[162] Reggio, P. H. Computational methods in drug design: modeling G protein-coupled receptor monomers, dimers, and oligomers. AAPS J. 2006, 8, E322-336.

[163] Franco, R.; Casado, V.; Cortes, A.; Ferrada, C.; Mallol, J.; Woods, A.; Lluis, C.; Canela, E. I.; Ferre, S. Basic concepts in G-protein-coupled receptor homo- and heterodimerization. ScientificWorldJournal 2007, 7, 48-57.

[164] Vohra, S.; Chintapalli, S. V.; Illingworth, C. J.; Reeves, P. J.; Mullineaux, P. M.; Clark, H. S.; Dean, M. K.; Upton, G. J.; Reynolds, C. A. Computational studies of Family A and Family B GPCRs. Biochem. Soc. Trans. 2007, 35, 749-754.

[165] Rasmussen, S. G.; Choi, H. J.; Rosenbaum, D. M.; Kobilka, T. S.; Thian, F. S.; Edwards, P. C.; Burghammer, M.; Ratnala, V. R.; Sanishvili, R.; Fischetti, R. F.; Schertler, G. F.; Weis, W. I.; Kobilka, B. K. Crystal structure of the human β2 adrenergic G-protein-coupled receptor. Nature 2007, 450, 383-387.

[166] Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmussen, S. G.; Thian, F. S.; Kobilka, T. S.; Choi, H. J.; Kuhn, P.; Weis, W. I.; Kobilka, B. K.; Stevens, R. C. High-resolution crystal structure of an engineered human β2-adrenergic G protein-coupled receptor. Science 2007, 318, 1258-1265.

[167] Rosenbaum, D. M.; Cherezov, V.; Hanson, M. A.; Rasmussen, S. G.; Thian, F. S.; Kobilka, T. S.; Choi, H. J.; Yao, X. J.; Weis, W. I.; Stevens, R. C.; Kobilka, B. K. GPCR engineering yields high-resolution structural insights into β2-adrenergic receptor function. Science 2007, 318, 1266-1273.

[168] Hanson, M. A.; Cherezov, V.; Griffith, M. T.; Roth, C. B.; Jaakola, V. P.; Chien, E. Y.; Velasquez, J.; Kuhn, P.; Stevens, R. C. A specific cholesterol binding site is established by the 2.8 Ǻ structure of the human beta2-adrenergic receptor. Structure 2008, 16, 897-905.

[169] Warne, T.; Serrano-Vega, M. J.; Baker, J. G.; Moukhametzianov, R.; Edwards, P. C.; Henderson, R.; Leslie, A. G.; Tate, C. G.; Schertler, G. F. Structure of a β1-adrenergic G-protein-coupled receptor. Nature 2008, 454, 486-491.

[170] Yuzlenko, O.; Kiec-Kononowicz, K. Molecular modeling of A1 and A2A adenosine receptors: comparison of rhodopsin- and beta2-adrenergic-based homology models through the docking studies. J. Comput. Chem. 2009, 30, 14-32.

[171] Ivanov, A. A.; Barak, D.; Jacobson, K. A. Evaluation of Homology Modeling of G-protein-coupled receptors in light of the A2A adenosine receptor crystallographic structure. J. Med. Chem. 2009, 52, 3284–3292.

[172] Sherbiny, F. F.; Schiedel, A. C.; Maass, A.; Muller, C. E. Homology modelling of the human adenosine A2B receptor based on X-ray structures of bovine rhodopsin, the β2-adrenergic receptor and the human adenosine A2A receptor. J. Comput. Aided Mol. Des. 2009, 23, 807-828.

[173] Volpini, R.; Dal Ben, D.; Lambertucci, C.; Marucci, G.; Mishra, R. C.; Ramadori, A. T.; Klotz, K. N.; Trincavelli, M. L.; Martini, C.; Cristalli, G. Adenosine A2A receptor antagonists: new 8-substituted 9-ethyladenines as tools for in vivo rat models of Parkinson's disease. ChemMedChem 2009, 4, 1010-1019.

[174] Lambertucci, C.; Antonini, I.; Buccioni, M.; Dal Ben, D.; Kachare, D. D.; Volpini, R.; Klotz, K. N.; Cristalli, G. 8-Bromo-9-alkyl adenine derivatives as tools for developing new adenosine A2A and A2B receptors ligands. Bioorg. Med. Chem. 2009, 17, 2812-2822.

[175] Lenzi, O.; Colotta, V.; Catarzi, D.; Varano, F.; Poli, D.; Filacchioni, G.; Varani, K.; Vincenzi, F.; Borea, P. A.;

26

Paoletta, S.; Morizzo, E.; Moro, S. 2-Phenylpyrazolo[4,3-d]pyrimidin-7-one as a new scaffold to obtain potent and selective human A3 adenosine receptor antagonists: new insights into the receptor-antagonist recognition. J. Med. Chem. 2009, 52, 7640-7652.

[176] Tosh, D. K.; Chinn, M.; Ivanov, A. A.; Klutz, A. M.; Gao, Z. G.; Jacobson, K. A. Functionalized congeners of A3 adenosine receptor-selective nucleosides containing a bicyclo[3.1.0]hexane ring system. J. Med. Chem. 2009, 52, 7580-7592.

[177] Olah, M. E.; Stiles, G. L. The role of receptor structure in determining adenosine receptor activity. Pharmacol. Ther. 2000, 85, 55-75.

[178] Dawson, E. S.; Wells, J. N. Determination of amino acid residues that are accessible from the ligand binding crevice in the seventh transmembrane-spanning region of the human A1 adenosine receptor. Mol. Pharmacol. 2001, 59, 1187-1195.

[179] Mustafi, D.; Palczewski, K. Topology of class A G protein-coupled receptors: insights gained from crystal structures of rhodopsins, adrenergic and adenosine receptors. Mol. Pharmacol. 2009, 75, 1-12.