Human immunodeficiency virus type 1 (HIV-1) entry into cells is mediated by the viral envelope glycoproteins (Env) that decorate the surface of the virion (reviewed in [1, 2]). The Env complex is organized into trimers, and consists of gp120, which is positioned on the surface of the viral membrane and presents the receptor binding surfaces of Env, and gp41, which is occluded by gp120 and anchors Env to the viral membrane. The primary point of contact between Env and the host cell is between gp120 and cellular CD4. This leads to structural rearrangements in gp120 that expose the binding site for a cellular coreceptor, which is either of the chemokine receptors CCR5 or CXCR4. Binding of gp120 to coreceptor leads to further structural changes in Env that expose the fusion peptide of gp41 that enables virus-cell fusion and entry of the virion core into the cell.
CCR5 comprises seven transmembrane helices that form a central hydrophobic pocket. Extending from the helices are the chemokine binding domains, comprising the N-terminus and three extracellular loops (ECL1, 2 and 3). Current models of gp120 binding to CCR5 suggest that the crown of the gp120 V3 loop interacts principally with the ECL2 region, while the gp120 bridging sheet that is formed between the C1, C2 and C4 domains of gp120 after CD4 binding, and the stem of the V3 loop interact with the CCR5 N-terminus [3–6]. Sulfation of tyrosine residues in the CCR5 N-terminus, particularly those at position 10 and 14, has been shown to be important for HIV-1 entry [7, 8].
CCR5 antagonists belong to a relatively new class of HIV-1 antivirals known as entry inhibitors. They act by binding to the hydrophobic pocket in CCR5 formed by the transmembrane helices [9–14], which leads to structural alterations in the ECLs such that they are no longer recognized by gp120 [10, 13–15]. The structure of the N-terminus does not appear to be affected by the binding of CCR5 antagonists to CCR5 . Thus, CCR5 antagonists do not competitively block the binding of Env to CCR5, rather they are allosteric inhibitors of HIV-1 entry.
Maraviroc (MVC) [16, 17] is the only CCR5 antagonist that has been licensed for clinical use as an HIV-1 antiretroviral therapy . Other CCR5 antagonist HIV-1 inhibitors that are no longer being pursued for clinical development due to lack of clinical efficacy and/or adverse side effects include vicriviroc (VVC) , aplaviroc (APL)  and TAK-779 . HIV-1 can develop clinical resistance to CCR5 antagonists by two routes. The first pathway is through emergence of minor CXCR4-using HIV-1 variants that were not detected in plasma prior to initiation of CCR5 antagonist therapy [22–29]. The second pathway to resistance is via continued use of CCR5, which is characterized not by shifts in IC50 (as would be expected for a competitive inhibitor), but rather by reductions in the maximal percent inhibition (MPI) [15, 28, 30–39]. Reductions in MPI are due to the resistant virus developing the ability to bind to the antagonist-modified form of CCR5 [34, 35]. The plateau height of the MPI is also informative, with viruses that display a relatively high MPI (>80%) having relatively inefficient usage of the antagonist-modified form of CCR5 compared to unmodified CCR5, and viruses that display a relatively low MPI (<20%) having relatively efficient usage of the antagonist-modified form of CCR5 compared to unmodified CCR5.
Multiple studies have demonstrated that adaptive changes in the gp120 V3 loop are almost always responsible for the ability of resistant viruses to utilize the antagonist-modified form of CCR5 [15, 32–34, 37, 40–42], although mutations in gp41 conferring this ability have also been observed [43–45]. No signature pattern of common amino acid mutations has been identified that would predict resistance to CCR5 antagonists, such as that which has been noted for HIV-1 resistance to the fusion inhibitor enfuviritide. Rather, the V3 loop alterations responsible for resistance to CCR5 antagonists appear to be strain specific [31, 37, 44, 46, 47].
The mechanisms underlying the ability of viruses with resistance to CCR5 antagonists to recognize the antagonist-modified form of CCR5 are incompletely understood. Most studies suggest that resistant viruses develop an increased reliance on the CCR5 N-terminus, most likely signaling a shift in gp120 binding to a region of CCR5 not modified by the binding of antagonist [15, 32, 42, 48–50]. However, a recent study of VVC resistant strains showed heterogeneity in the ability of the resistant viruses to interact with VVC-modified CCR5, suggesting differing resistance mechanisms . There is also evidence to suggest that resistant viruses must still interact with the drug modified ECLs, although the degree to which this interaction is required may be strain specific [15, 51, 52].
The vast majority of studies of HIV-1 resistance to CCR5 antagonists, however, have been conducted on resistant viruses that were generated in vitro[34, 37, 53, 54], or on viruses isolated from subjects failing therapy by CCR5 antagonists that are no longer being pursued for clinical development, such as VVC, APL and TAK-779 [30, 32, 33, 42]. In contrast, there is a paucity of data on MVC-resistant viruses arising in vivo[15, 55]. Since different CCR5 antagonists cause subtly different structural alterations in CCR5 [14, 15], it is unlikely that mechanisms of HIV-1 resistance to VVC, APL and TAK-779 can completely recapitulate those to MVC.
In this study we characterized the genetic, structural and functional properties of MVC-resistant Envs derived from two subjects who developed phenotypic resistance in vivo, with comparison to Envs derived from the same subjects prior to commencement of MVC. By measuring the MPIs in different cell types and by conducting quantitative Env-CCR5 affinity profiling studies, we show that the magnitude of MVC resistance was highly divergent in these subjects, and was characterized by either a highly efficient or a very inefficient ability of the resistant Envs to engage MVC-bound CCR5. Despite highly divergent resistance phenotypes and a lack of common gp120 resistance mutations, both resistant Envs exhibited an almost superimposable pattern of dramatically increased reliance on sulfated tyrosine residues in the CCR5 N-terminus, and on histidine residues in the CCR5 ECLs. Furthermore, neither of the MVC-resistant Envs displayed broad cross resistance to other CCR5 antagonists. Together, our results suggest that clinical resistance to MVC may involve divergent Env phenotypes and different genetic alterations in gp120, but the molecular pathway to resistance for the Envs studied here appears to be related.