Accessory mutations balance the marginal stability of the HIV‐1 protease in drug resistance

The HIV‐1 protease is a major target of inhibitor drugs in AIDS therapies. The therapies are impaired by mutations of the HIV‐1 protease that can lead to resistance to protease inhibitors. These mutations are classified into major mutations, which usually occur first and clearly reduce the susceptibility to protease inhibitors, and minor, accessory mutations that occur later and individually do not substantially affect the susceptibility to inhibitors. Major mutations are predominantly located in the active site of the HIV‐1 protease and can directly interfere with inhibitor binding. Minor mutations, in contrast, are typically located distal to the active site. A central question is how these distal mutations contribute to resistance development. In this article, we present a systematic computational investigation of stability changes caused by major and minor mutations of the HIV‐1 protease. As most small single‐domain proteins, the HIV‐1 protease is only marginally stable. Mutations that destabilize the folded, active state of the protease therefore can shift the conformational equilibrium towards the unfolded, inactive state. We find that the most frequent major mutations destabilize the HIV‐1 protease, whereas roughly half of the frequent minor mutations are stabilizing. An analysis of protease sequences from patients in treatment indicates that the stabilizing minor mutations are frequently correlated with destabilizing major mutations, and that highly resistant HIV‐1 proteases exhibit significant fractions of stabilizing mutations. Our results thus indicate a central role of minor mutations in balancing the marginal stability of the protease against the destabilization induced by the most frequent major mutations.


| INTRODUCTION
The HIV-1 protease plays an essential role in HIV replication by cleaving newly synthesized polyprotein chains at several places into functional protein components of the virus. The cleavage of the polyprotein chain occurs at a tunnel-shaped active site enclosed by the two identical chains of the dimeric HIV-1 protease. 1,2 In its unbound state, the HIV-1 protease adopts a semiopen conformation that enables the entry of the polyprotein chain into the active-site tunnel. The binding of the polyprotein substrate induces a change to the closed conformation of the protease in which substrate cleavage occurs. 3,4 Because of its essential role for HIV replication, the HIV-1 protease is a major target in AIDS therapy. Nine drugs approved for highly active antiretroviral therapies are inhibitors of the HIV-1 protease. 1,5 These protease inhibitors bind in the active-site tunnel and, thus, prevent substrate binding.
During therapy, mutations of the HIV-1 protease can lead to resistance to protease inhibitors. The development of resistance typically involves several mutations. 1,6 Mutations that usually occur first and individually reduce the susceptibility 7 to one or several protease inhibitors are called "major" mutations. 8 Mutations that occur later and individually do not substantially affect the susceptibility to inhibitors are called "minor" mutations. At present, 23 major mutations and 50 minor mutations involved in resistance development have been classified. 8 Major mutations of the HIV-1 protease involved in drug resistance are predominantly located in the active-site tunnel or at the dimer interface and can directly interfere with drug binding by changing the shape of the active-site tunnel. 1,5 Minor mutations, in contrast, are predominantly located distal to the active-site tunnel.
A central question is how minor mutations contribute to resistance development. Structural investigations of mutant proteases indicate that some minor mutations may indirectly affect the active-site tunnel via a propagation of structural changes from the distal site of mutation to the active site, 5,9 via coordinated structural rearrangements of multiple mutations, 10 or via coordinated changes in the structural dynamics that may affect the balance between substrate cleavage and drug binding. 11 Other minor mutations have been shown to increase the thermal stability of the HIV-1 protease and to compensate stability decreases caused by major mutations. 12 Highly drug-resistant mutants of the HIV-1 protease with a large number of mutations can be more stable than the wildtype (WT). 13,14 The stability of the WT protease dimer depends on the monomer concentration and ranges from 4 to 10 kcal/ mol for micro-to millimolar concentrations, 15,16 which are typical stability values of marginally stable proteins. [17][18][19][20][21] Here, we systematically investigate mutation-induced stability changes ΔΔG of the HIV-1 protease via calculations with the program Rosetta. 22,23 The Rosetta prediction accuracies for ΔΔG exceed the accuracies of other programs, [23][24][25] in particular for mutations of smaller amino acids into larger amino acids, which frequently occur in the HIV-1 protease. Our ΔΔG calculations indicate that the most frequent major mutations destabilize the HIV-1 protease, and that roughly half of the frequent minor mutations are stabilizing. We find that the stabilizing minor mutations are often correlated with destabilizing major mutations, and that highly resistant HIV-1 proteases exhibit significant fractions of stabilizing mutations. These results indicate a central role of minor mutations in maintaining the marginal stability of the HIV-1 protease.

| METHODS
We have chosen the high-resolution and high-quality structure 2PC0 26 for the stability calculations in the unbound state of the HIV-1 protease. The resolution of this structure is 1.4 Å, which is significantly higher than the resolution of other unbound structures. For the stability calculations in the bound state, we have chosen the three high-resolution and high-quality structures 4EJD, 27 4EJK, 27 and 4E43 28 with the same sequence as 2PC0. We have performed the stability calculations with the recommended Rosetta protocol 16, which includes limited backbone flexibility. 23 Backbone flexibility appears to be of particular importance for reliable predictions of stability changes for mutations of smaller amino acids into larger amino acids, 23,29 which frequently occur in the HIV-1 protease. Rosetta results are multiplied with a factor 0.57 kcal/mol to obtain units of kcal/mol. 23   in the absence of substrate or drug molecules. The common major mutations V82A, I84V, and L90M, which are associated with resistance against several approved inhibitors, 7 lead to a decrease of the melting temperature T m relative to the WT. This decrease in the melting temperature reflects a thermal destabilization and is significantly more pronounced for the double mutant I84V-L90M, compared to the single mutants I84V and L90M. In contrast, double mutants that contain the major mutation I84V and one of the minor mutations L10I, L63P, and A71V have melting temperatures ΔT m close to the WT value, which indicates that these minor mutations compensate the thermal destabilization induced by I84V. The double mutant I84V-V77I exhibits a melting temperature close to the single mutant I84V.
In Figure 1, the melting-temperature changes ΔT m measured by Chang and Torbett are plotted against stability changes ΔΔG calculated with Rosetta 23 for the open conformation of the HIV-1 protease, which is the ground-state conformation in the unbound state.
The double-mutant A71V-I84V is excluded from the figure because of steric overlaps in Rosetta, which occur also for the single mutant A71V. These steric overlaps indicate that the reconfiguration of the backbone and/or neighboring sidechains induced by the "small-tolarge" mutation A71V is beyond the scope of the recommended Rosetta protocol for calculating stability changes. 23 Table 1, and (C) for the frequent minor mutations of Table 2. Blue bars indicate stabilizing mutations with ΔΔG ≤ −0.5 kcal/mol. The mutation-induced stability changes ΔΔG in the unbound state have been calculated for the high-resolution and high-quality X-ray structure with protein data bank code 2PC0. 26 The 1427  3.2 | Roughly half of the frequent minor mutations are stabilizing Tables 1 and 2 summarize the calculated stability changes for all major and frequent minor mutations of the HIV-1 protease. 8 Table 1 and the frequent minor mutations of Table 2 Figure 2B has a shape that is rather similar to the shape of the distribution in Figure 2A, with the caveat that the distribution of Figure Table 2, which is a fraction of 50%.
This fraction of stabilizing frequent minor mutations is significantly larger than the fraction of stabilizing mutations in Figure 2A Table 3 lists all correlated pairs of major and frequent minor mutations X and Y with ϕ ≥ 0.2. Besides the ϕ coefficients, Table 3 includes the stability changes ΔΔG X and ΔΔG Y for the single mutations X  Table 3, the sum ΔΔG X + ΔΔG Y of the stability changes for the single-residue mutants does not deviate from the stability change ΔΔG XY of the double mutants by more than ± 0.5 kcal/mol, which is within the numerical accuracy of our Rosetta calculations. The stability changes ΔΔG X and ΔΔG Y for the single mutants X and Y thus appear to be additive within numerical accuracy.
The correlations of major and frequent minor mutations of Table 3 are highly significant with P-values smaller than 10 −100 , but typically do

| Highly resistant HIV-1 proteases exhibit significant fractions of individually stabilizing mutations
In standard essays, 35,36 the resistance of protease mutants to inhibitors is inferred from the half-maximal inhibitory concentration IC 50 at which virus replication is inhibited by 50%. The fold resistance is the ratio of In Figure 3A,  Table 1). These stabilizing minor mutations are often correlated with the most frequent, destabilizing major mutations (see Table 3) and, thus, appear to play a central role in maintaining the marginal stability of the HIV-1 protease. Highly resistant HIV-1 proteases exhibit significant fractions of stabilizing mutations (see Figure 3B).