opioid Receptor Models — Bautista, Asher, and Caq)enter 113 



Figure 2. Complex of morphine witli the mii model. Panel A, view from the side showing placement of ligand in the 

 receptor. Panel B, view from extracellular space showing placement of ligand in the helical bundle. Panel C, close-up 

 of ligand and residues in the receptor. Hydrogen bonds are shown as dotted lines. 



had the same hydrogen bonds listed and, in 

 addition, a hydrogen bond to the side chain of 

 SerSSl in some but not all entries of the da- 

 tabase. Overall, the complexes, very stable, 

 maintained the same network of hydrogen 

 bonds. 



DISCUSSION 



To evaluate the fitness of the models, they 

 were compared to experimental data in the lit- 

 erature. The literature review was not exhaus- 

 tive but surveyed some of the more important 

 findings of others to determine if the receptor 

 models developed in our study were in accor- 

 dance with experimental data. 



In one computational study of the human 

 opioid receptors, ab initio models were devel- 

 oped (Strahs and Weinstein 1997). These 

 models were developed before the crystal 

 structure of bovine rhodopsin was solved and 

 therefore used ideal helices and packed them 

 based on the diffraction map of bovine rho- 

 dopsin (Schertler et al. 1993). The heUces 



were oriented based on the calculated hydro- 

 phobic moment that oriented the most hydro- 

 phobic side to the lipid bilayer. The residues 

 Fukuda et al. (1995) predicted to be inward 

 facing and that varied across the opioid recep- 

 tors are listed in Table 5. They reasoned that 

 since the position varied and it was predicted 

 to be inward facing, it might be involved in 

 ligand selectivity. Of the 14 residues listed, 

 only three are inward facing in our models: 

 AlalSl, Val364, and Tq3382. Many are orient- 

 ed towards other helices or the lipid bilayer. 

 The validity of using a hydrophobic moment 

 calculation can be questioned since the data 

 presented by Befort et al. (1999) disagree with 

 models based on experimentally validated 

 models. Of the three residues that are inward 

 facing only one is largely different. The vari- 

 ation at AlalSl is minimal, either alanine or 

 valine. Neither substitution would greatly im- 

 pact either charge or size at that position and 

 would therefore not account for binding dif- 

 ference in the opioid family. Val364 is likewise 



