Class A Gpcrs 1631 Sequence Homology

The X-ray structures of bacteriorhodopsin and bovine rhodopsin have been used as templates for the construction of homology models of GPCR targets. However, because both proteins are only distantly related to GPCRs of drug-discovery interest, and bacteriorhodopsin is not a GPCR, the use of their X-ray structures as templates is challenging. For example, PSI-Blast [18] sequence alignments indicate that bacteriorhodopsin shares only 17%, 12%, 12%, 14%, and 14% sequence identity with the human p1 adrenergic receptor (p1AR), the human p2 adrenergic receptor (p2AR), the human A2a adenosine receptor, the human dopamine D2 receptor, and the human metabotropic glutamate receptor 2 (mGluR2), respectively. Bovine rhodopsin's sequence identity percentages relative to the same GPCRs marginally improve to 20%, 19%, 21%, 21%, and 17%, respectively. Interestingly, bacteriorhodopsin and bovine rhodopsin share only 17% sequence identity. The percentage identity values derived by the use of PSI-Blast for these selected examples are likely to represent upper limits, because PSI-Blast alignments tend to emphasize individual residue matches. Thus, the percentage identity values may differ when sequence alignments are performed manually while incorporating the knowledge of conserved motifs or biophysical data. Also, one would expect that the percentage identity would be lower in the loop regions.

The low levels of identity and overall homology create uncertainty in aligning [12] GPCR sequences of interest with that of bovine rhodopsin or bacte-riorhodopsin. If an alignment is in error by even a single residue, the resulting model may have an inappropriately sized binding cavity (preventing a meaningful fitting of ligands) or have displaced critical residues so they are no longer capable of interacting with ligands. Such a model would be unusable for drug design, and without an adjustment to the alignment, the model cannot generally be corrected by relaxation afforded by molecular dynamics or other computational techniques.

0 0

Post a comment