We assessed the relative importance of two serine residues located near the top of transmembrane helix 5 of the human 5-HT 2A receptor, comparing the wild type with S5.43(239)A or S5.46(242)A mutations. Using the ergoline lysergic acid diethylamide (LSD), and a series of substituted tryptamine and phenethylamine 5-HT 2A receptor agonists, we found that Ser5.43(239) is more critical for agonist binding and function than Ser5.46(242). Ser5.43(239) seems to engage oxygen substituents at either the 4-or 5-position of tryptamine ligands and the 5-position of phenylalkylamine ligands. Even when a direct binding interaction cannot occur, our data suggest that Ser5.43(239) is still important for receptor activation. Polar ring-substituted tryptamine ligands also seem to engage Ser5.46(242), but tryptamines lacking such a substituent may adopt an alternate binding orientation that does not engage this residue. Our results are consistent with the role of Ser5.43(239) as a hydrogen bond donor, whereas Ser5.46(242) seems to serve as a hydrogen bond acceptor. These results are consistent with the functional topography and utility of our in silico-activated homology model of the h5-HT 2A receptor. In addition, being more distal from the absolutely conserved Pro5.50, a strong interaction with Ser5.43(239) may be more effective in straightening the kink in helix 5, a feature that is possibly common to all type A GPCRs that have polar residues at position 5.43.The serotonin 2A (5-HT 2A ) receptor is a member of the monoamine family A type GPCRs. It seems to play an essential role in cognition, memory, and consciousness (Nichols, 2004). Although no crystal structure exists for this or any of the monoamine GPCRs, advances in X-ray crystallography have provided structures for the dark-adapted inverse agonist form of bovine rhodopsin (Palczewski et al., 2000;Li et al., 2004), creating opportunities to develop homology models of general GPCR structure. Given sufficient sequence similarity/identity, the accuracy and reliability of comparative homology models is generally believed to be superior to de novo models (Baker and Sali, 2001) As an example of this approach, the in silico-activated homology model of the h5-HT 2A receptor developed in our laboratory has been used to predict several ligand-receptor interactions. Our results thus far have provided qualitative support for the model and have identified directions for further investigation of the structure-activity relationships of agonist ligands (Parrish et al., 2005;Braden et al., 2006; McLean et al., 2006a,b). Although such homology receptor models must be viewed with caution, they can often be validated by receptor mutagenesis experiments, along with complementary changes in ligand structure.To refine our homology model and provide empirical support for the receptor functional topography it defines, we extended results derived from previous mutagenesis studies Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org. doi:10....