Modeling a Ligand Bound State of the 5HT2C Receptor

Most 3D homology models of therapeutically relevant GPCRs are conceived to design, from a training set of reference compounds, in situ fingerprints of interaction able to forecast the structure and efficiency of new potent drugs. Achieving the best prediction of structure-based qualitative 3D pharmacophores or quantitative structure-activity models requires a reliable correlation between in vitro and in silico amino acid residues found to be specifically bound by high-affinity ligands. Moreover, the binding mode of relevant ligands inside the predicted binding cavities needs a prior study of the structural features associated with the pharmacological profile of ligands. At a pharmacological point of view, receptors conform to the rules of thermodynamic equilibrium between the inactive (R) and active states (R*), which structurally means a gradient of conforma-tional states between these two limit snapshot structures. Computational MD calculations enable the simulation of the collective motion of atoms in such mac-romolecular systems and result in the lowest-energy conformational states of receptor-ligand complexes theoretically corresponding to the ligand-induced activation state of the receptor.

6.3.1 Investigation of Antagonist Binding

Many strategies have been adopted by authors in order to study binding modes of 5-HT2C reference compounds whose associated structures and experimental data are represented in Table 6.3. Since templates used for homology modeling are either a ground state of RHO or ADRB2 cocrystallized with an inverse agonist ligand, both consequently not in an agonist-induced active state, resultant 3D models are presumed to be in an inactive state.

6.3.1.1 Identification of Ligand Binding Sites and Docking of Antagonists

In this way, Rashid et al. as well as Farce et al. have directly worked from the optimized and validated structure of the apo-5-HT2C to dock therein sarpogrelate (Rashid et al. 2003) and agomelatin (Millan et al. 2003) antagonist, respectively,

Table 6.3 Structure and affinity of reference 5-HT2C ligands Agomelatin antagonist pKi 6.2 Sarpogrelate antagonist pKi 7.4

(continued)

Table 6.3 (continued)

Ritanserin antagonist pKi 9.6 Serotonin agonist pKi 7.8

Table 6.3 (continued)

Ritanserin antagonist pKi 9.6 Serotonin agonist pKi 7.8

Metergolin inverse agonist pKi 8.5 Azepinoindol agonist pKi 7.9

whereas modifications have been performed by Bray et al. before the docking of the methiothepin (Knight et al. 2004), metergolin (Knight et al. 2004), and ritanserin (Knight et al. 2004) reference antagonists in the ligand binding site. Indeed, in the later case, they were searching the best binding site for each studied ligand by virtually mutating all bulky nonpolar residues (leucine, isoleucine, valine, phely-lalanine, and tryptophane) to alanine. Then the reshaped ligand binding site was divided into a set of overlapping regions in which were sampled different ligand conformations by docking calculations. For each of the 50 ligand conformations, the bulky non-polar residues were reversed to their original form then the geometry of ligands and closest residues was refined by energy minimization. Structures with the lowest total energy of the ligand-protein complex and with the best contacts between the respective functional groups of the ligands and the residues in the binding site were finally selected.

Once the ligand binding site has been identified in the apo-5-HT2C 3D models by virtually probing the solvent-accessible surface, the other authors have therein performed the docking calculations of the relevant ligands within a sphere radius. In these cases, flexible docking is commonly used and based on iterative generations of sampled conformations of ligands for preferential electrostatic and hydro-phobic interactions with the amino acid residues of the protein. Bray et al. took on a systematic search of docking poses by an incremental building of ligand fragments (Kuntz 1992), whereas the others employed stochastic-oriented programs based on a genetic algorithm (Goodsell et al. 1996; Jones et al. 1995a; Jones et al. 1995b). In our ADRB-based 3D model of apo-5-HT2C we have manually preposi-tioned SB-228357, one potent member of the ((aryl)carbamoyl)indolin series (Bromidge et al. 1997; Bromidge et al. 1998; Bromidge et al. 2000; Bromidge et al. 1999), found to actually be the most selective 5-HT2C antagonists and voluminous enough to target the maximal spots of interactions in the vicinity of the binding site. A conformational sampling of the receptor-ligand complex has been then performed by simulated annealing. This molecular dynamics procedure, consists in ten heating cycles at 700 K, each followed by cooling steps down to 300 K, and permits to catch the lowest-energy conformations of the receptor-ligand complex. Annealing was simulated on all atoms of the E2 loop and the ligand as well as the side chains of TM amino acid residues involved in interaction with the ligand. Receptor-ligand complexes with the lowest total energy matching the two critical electrostatic (Asp3.32) and aromatic (Phe6.51 or Phe6.52) anchor sites were finally selected.

6.3.1.2 Analysis of 5-HT2C-Antagonists Interactions

The receptor-ligand interactions brought out from all docking procedures in the four 3D models are summarized in the Table 6.4 and are distributed according to van der Waals (VdW) and aromatic (p-p stacking) as hydrophobic interactions or electrostatic interactions gathering hydrogen bonds and salt bridges, respectively, H-bond and ionic in Table 6.4. The common fingerprint of interactions among all 3D models corresponds to the previously discussed patterns Asp3.32 and one member of the aromatic cluster (Phe5.47, Trp6.48, Phe6.51, Phe6.52), although not involved in metergolin and methiothepin interactions. Since these are conserved residues among GPCR-type monoamine receptors, specific 5-HT2C spots of interactions seem to be located over all the TM bundle with hydrophobic interaction from Trp3.28, Leu3.31, Val3.33, Phe3.35, Phe5.38, and Asn6.55 and polar interactions from Ser3.36 and Tyr7.43, all of which are described more than once in at least two different models. Other single interactions are coming from the relatively different chemical scaffolds of metergolin and methiothepin, which have additional hydrophobic interactions with TM4 hydrophobic residues (Ile4.60 and Pro4.61). Specific observations associated to metergolin and methiothepin could be conferred by their particular pharmacological profile of inverse agonist, which would induce a binding to another intermediate activation state of the receptor. Otherwise, a benzyl group of the more

Table 6.4

Interactions of 5-HT2C antagonists

Sarpogrelate (Rashid's

Agomelatin

Interactions of 5-HT2C antagonist

model)

(Farce's model)

TM2

Ala2.64(113)

El

Trp120

TM3

Trp3.28 (130) Leu3.31(133)

Aromatic VdW

Asp3.32(134)

Ionic

H bond

Val3.33(135)

VdW

VdW

Phe3.35(137)

Ser3.36(138)

H bond

H bond

Ser3.39(141)

Ile3.40(142)

VdW

TM4

Ile4.60(189) Pro4.61(190)

E2

Arg195 Asn204 Thr205 Asn210

H bond

TM5

Phe5.38(214) Val5.39(215) Ser5.43(219) Phe5.47(223) Ile5.49(225)

Aromatic

TM6

Trp6.48(324)

VdW

Aromatic

Phe6.51(327)

VdW

Aromatic

Phe6.52(328)

Aromatic

Aromatic

Asn6.55(331)

Ile6.56(332)

TM7

Asn7.36(351) Val7.39(354) Tyr7.43(358)

H bond

6 Homology Modeling of 5-HT2C Receptors

113

Bray's model

SB-228357

Ritanserin Metergolin

Methiothepin

(Renault's model)

VdW

H bond

Aromatic

Ionic Ionic

Ionic

H bond

Aromatic H bond

VdW

VdW

VdW

VdW

H bond

H bond

H bond

H bond

Aromatic

VdW

VdW VdW

H bond

VdW

H bond

Aromatic

VdW

VdW

H bond

H bond

weighty and voluminous sarpogrelate is implied in an aromatic interaction with the buried Phe5.47 residue. Only 3D models constructed in our lab highlighted amino acid residues of E2 loop that are directly involved in interactions with the ligands (Fig. 6.4). As previously reported, this loop was experimentally found to bring the receptor subtype selectivity within the class A GPCR family. In these models, it appears that only hydrogen bonds are involved. If such a bond establishes only once between agomelatin and Asn204, there are four hydrogen bonds between SB-228357 and E2 residues, Arg195, Asn204, Thr205, and Asn210. Moreover, E1 loop is also extensively represented in our ADRB-based model by interaction between the donor nitrogen of Trp120 indole and the acceptor nitrogen of the pyridyl group.

Defeat Drugs and Live Free

Defeat Drugs and Live Free

Being addicted to drugs is a complicated matter condition that's been specified as a disorder that evidences in the obsessional thinking about and utilization of drugs. It's a matter that might continue to get worse and become disastrous and deadly if left untreated.

Get My Free Ebook


Post a comment