Analysis of HIV-1 CRF_01 A/E protease inhibitor resistance: structural determinants for maintaining sensitivity and developing resistance to atazanavir

Abstract

A series of HIV-1 protease mutants has been designed in an effort to analyze the contribution to drug resistance provided by natural polymorphisms as well as therapy-selective (active and non-active site) mutations in the HIV-1 CRF_01 A/E (AE) protease when compared to that of the subtype B (B) protease. Kinetic analysis of these variants using chromogenic substrates showed differences in substrate specificity between pretherapy B and AE proteases. Inhibition analysis with ritonavir, indinavir, nelfinavir, amprenavir, saquinavir, lopinavir, and atazanavir revealed that the natural polymorphisms found in A/E can influence inhibitor resistance. It was also apparent that a high level of resistance in the A/E protease, as with B protease, is due to it aquiring a combination of active site and non-active site mutations. Structural analysis of atazanavir bound to a pretherapy B protease showed that the ability of atazanavir to maintain its binding affinity for variants containing some resistance mutations is due to its unique interactions with flap residues. This structure also explains why the I50L and I84V mutations are important in decreasing the binding affinity of atazanavir.

Figure 1

Ribbon diagrams of HIV-1 proteases. Top left, pre-therapy HIV-1 subtype-B (B) protease; top right, B protease plus active site mutation V82F (B^V82F); middle left, pre-therapy HIV-1 CRF_01 A/E (AE) protease; middle right, AE protease plus active site mutation V82F (AE^V82F); bottom left, post-therapy A/E protease (AE-P); bottom right, AE-P protease with back mutation F82V (AE-P^F82V). The red sphere marks the V82F active site mutation. The green spheres mark the natural polymorphisms found in AE protease when compared to the B protease. The brown spheres mark the therapy acquired non-active site mutations in the AE-P protease.

Figure 2

Relative vitality values. (A) Relative vitality values for B, B^V82F, AE, and AE^V82F using B and K_i of IDV as reference. (B) Relative vitality values for AE, AE^V82F, AE-P, and AE-P^F82V using AE and K_i of IDV as reference.

Figure 3

(A) Stereo view of ATV in the active site. Conformation-1 (Conf1) (red) and Conformation-2 (Conf2) (green) are shown as sticks. Residues Val32(A), Val82(A), and Ile84(A) in chain A, and Val32(B), Val82(B), and Ile84(B) in chain B are shown as green sticks. 2FoFc electron density is drawn as a beige mesh at 1σ level. (B) Stick drawing of ATV Conf1 and 2 colored by B-values with red highest and blue lowest.

Figure 4

Ligplot two dimensional representation of hydrophobic interactions of ATV with active site residues for ATV in (panel A) Conf1 and (panel B) Conf2. Hydrophobic interactions with residues in the first monomer [e.g., Gly 49(A)] and the second monomer [e.g., Ile 50(B)] are shown as red dashed lines.

Figure 5

Stereo view of hydrogen bonding interactions of ATV for (A) Conf1 (green) and (B) Conf2 (aqua). Flap and active site residues are drawn as CPK sticks. Key residues described in the text are labeled. Hydrogen bonds are shown as yellow dashes. Water molecules are drawn as red spheres.

Figure 6

Interactions of residues Val32, Ile47, and Ile84 in inhibitor Conf1 (panel A) and Conf2 (panel B). In both panels, residues Ile84 and Ile47 (gray), Ile32(A) (blue), Val32(B) (orange), and inhibitor are modeled as ball and sticks.