Background:  CCR5-tropic HIV-1 variants, selected during prolonged serial passage of virus in the presence of HIV co-receptor antagonists, appear to use compound-bound receptor to infect target cells. Previously published studies suggest that all small molecule CCR5 antagonists bind to a similar region of the receptor, raising the possibility of generalized class-resistance.

Methods:  We have located the binding site of UK-427,857 and structurally related compounds within CCR5, using a panel of site-directed mutants and a 3-dimensional model based on homology with bovine rhodopsin. We also measured their potency against R5 viruses, including CCR5-tropic variants of primary HIV-1 isolates that are highly resistant to UK-427,857 (427^res).

Results:  All the antagonists bind CCR5 in a pocket formed by the trans-membrane helices and extracellular loop 2 (ECL2), centered on an ionic interaction with E283 and a hydrophobic interaction with Y108. The compounds have low nanomolar potency against the R5 lab-adapted HIV-1 strain, Ba-L, and are active against the wild type R5 primary isolates, CC1/85 and RU570. Compounds that contain a triazole moiety were inactive against 427^res variants of CC1/85 and RU570. In contrast, compounds where the triazole is replaced by an imidazopiperidine retained activity.

Conclusions:  Substitution of a key functional group in a series of structurally related HIV co-receptor antagonists leads to biologically significant changes in the way 427^res viruses interact with compound-bound CCR5. It appears that subtle differences in the occupation of the binding pocket, in particular around the ECL2 interface, enable some compounds to block replication of 427^res strains. Encouragingly, our data indicate that resistance to an HIV co-receptor antagonist will not necessarily lead to drug-class resistance.

Keywords: CCR5; resistance; entry