Drug Resistance Mutations Alter Dynamics of Inhibitor-Bound HIV-1 Protease

Abstract

Under the selective pressure of therapy, HIV-1 protease mutants resistant to inhibitors evolve to confer drug resistance. Such mutations can impact both the dynamics and structures of the bound and unbound forms of the enzyme. Flap+ is a multidrug-resistant variant of HIV-1 protease with a combination of primary and secondary resistance mutations (L10I, G48V, I54V, V82A) and a strikingly altered thermodynamic profile for darunavir (DRV) binding relative to the wild-type protease. We elucidated the impact of these mutations on protein dynamics in the DRV-bound state using molecular dynamics simulations and NMR relaxation experiments. Both methods concur in that the conformational ensemble and dynamics of protease are impacted by the drug resistance mutations in Flap+ variant. Surprisingly this change in ensemble dynamics is different from that observed in the unliganded form of the same variant (Cai, Y. et al. J. Chem. Theory Comput.2012, 8, 3452-3462). Our comparative analysis of both inhibitor-free and bound states presents a comprehensive picture of the altered dynamics in drug-resistant mutant HIV-1 protease and underlies the importance of incorporating dynamic analysis of the whole system, including the unliganded state, into revealing drug resistance mechanisms.

Figure 1

Snapshots of DRV-bound WT (blue) and Flap+ (red) protease complex conformation at the end of each 100 ns MD simulation.

Figure 2

RMSF values of the Cα atoms (Å) for each residue in WT (blue) and Flap+ (red) HIV-1 protease averaged over ten 100 ns MD simulations.

Figure 3

Distribution in percent of distances in Å between alpha carbons of the flaps, 80s loop, and the active site in WT (blue) and mutant Flap+ (MT, red) HIV-1 protease calculated over ten 100 ns trajectories. The h value is 1 for statistically significant differences between WT and Flap+ according to rank sum analysis (see Methods for details).

Figure 4

NMR relaxation data for WT (filled circles) and Flap+ (open triangles) HIV-1 protease in DRV-bound state. Data were acquired at a ¹⁵N Larmor frequency of 61 MHz at 20 °C. Two data points are displayed for some residues due to the loss of degeneracy between the two subunits on binding the asymmetric inhibitor DRV.

Figure 5

Order parameters of backbone N–H bonds from MD simulations (blue, green, red lines for 1, 10, and 50 ns time windows, respectively) and NMR experiments (black circles) for WT and Flap+ HIV-1 protease. There are two data points displayed for most residues for experimentally determined NMR order parameters, as the degenerate resonances cannot be unequivocally assigned to chains a or b due to the homodimeric nature of the protease.

Figure 6

Conformational exchange due to motions in the ms−μs time scale in unliganded WT (filled circle) and Flap+ (open triangle) HIV-1 protease. (A) The partial F statistic comparing the fits of individual residues to no-exchange and exchange models and (B) R_ex, the exchange contribution to transverse relaxation. High F statistic and R_ex values indicate residues undergoing conformational exchange in the ms−μs time scales.

Figure 7

Residues displaying statistically significant differences between WT and Flap+ protease dynamics in apo (left) and complex (right) forms. (a) MD order parameters. Blue indicates higher flexibility for WT (smaller order parameters) and red higher flexibility for Flap+ protease. (b) RMSF of Cα atoms in MD simulations. Magenta and cyan indicate higher fluctuations for WT and Flap+, respectively.