HIV-1 Integrase Inhibitors with Modifications That Affect Their Potencies against Drug Resistant Integrase Mutants

Abstract

Integrase strand transfer inhibitors (INSTIs) block the integration step of the retroviral lifecycle and are first-line drugs used for the treatment of HIV-1/AIDS. INSTIs have a polycyclic core with heteroatom triads, chelate the metal ions at the active site, and have a halobenzyl group that interacts with viral DNA attached to the core by a flexible linker. The most broadly effective INSTIs inhibit both wild-type (WT) integrase (IN) and a variety of well-known mutants. However, because there are mutations that reduce the potency of all of the available INSTIs, new and better compounds are needed. Models based on recent structures of HIV-1 and red-capped mangabey SIV INs suggest modifications in the INSTI structures that could enhance interactions with the 3'-terminal adenosine of the viral DNA, which could improve performance against INSTI resistant mutants. We designed and tested a series of INSTIs having modifications to their naphthyridine scaffold. One of the new compounds retained good potency against an expanded panel of HIV-1 IN mutants that we tested. Our results suggest the possibility of designing inhibitors that combine the best features of the existing compounds, which could provide additional efficacy against known HIV-1 IN mutants.

Keywords: inhibition; integrase; mutant; potency; strand transfer; susceptibility.

Figure 1

Chemical structures of the INSTIs. (A) The chemical structures of the clinically relevant INSTIs. (B) The chemical structures of our 4-amino-1-hydroxy-2-oxo-1,8-naphthyridine-containing compounds are shown.

Figure 2

Antiviral activities of the new compounds against RAL-resistant mutants. The EC₅₀ values were determined using a vector that carries WT HIV-1 IN and the RAL-resistant mutants in a single round infection assay. The potencies of DTG, 4d, 4f, and 5′g have been previously reported.^, The previously reported data are shown to simplify comparisons with the data for the new compounds 6u, 6v, 6w, and 5j. To better illustrate the higher EC₅₀ values, the y-axis was broken between 150 and 200 nM. Error bars represent the standard deviations of independent experiments, n = 4, performed in triplicate. The graph has a maximum value of 550 nM.

Figure 3

Antiviral activities of DTG, 4d, 4f, 5′g, and the new compounds against mutations in the connecting loop (β4-α2) near the active site. The EC₅₀ values were determined using a vector that carries the IN mutant in a single round infection assay. The y-axis was broken between 350 and 400 nM to better demonstrate the higher EC₅₀ values. Error bars represent the standard deviations of independent experiments, n = 4, performed in triplicate. EC₅₀ values shown in the graph have a maximum value of 1500 nM.

Figure 4

Mutations in the β5-α3 loop affect the antiviral potencies of the new compounds. The EC₅₀ values were determined using a vector that carries the IN mutant in a single round infection assay. The y-axis was broken between 100 and 150 nM to better show the higher EC₅₀ values. Error bars represent the standard deviations of independent experiments, n = 4, performed in triplicate. The EC₅₀ values shown in the graph have a maximum value of 550 nM.

Figure 5

Antiviral potencies of the new compounds against IN with mutations in the C-24 terminal domain. The EC₅₀ values were determined using a vector that carries the IN mutant in a single round infection assay. The y-axis was broken between 100 and 150 nM to better illustrate the higher EC₅₀ values. Error bars represent the standard deviations of independent experiments, n = 4, performed in triplicate. The EC₅₀ values shown in the graph have a maximum value of 650 nM.

Figure 6

Modeling of 6v in the HIV-1 intasome. (A) Compounds 4f, 6v, and Buffer MES were superimposed in the active site of the HIV-1 intasome. The binding of 6v (green) was predicted by docking it onto the structure of 4f (silver) in the active site of the HIV-1 intasome. (B) van der Waals (VDW) surface representation of MES (yellow and red). Water molecules observed in the HIV intasome apo (PDB ID: 6PUT, cyan) and PFV-6v (red) structures that are within 4 Å from compound 6v (green) are depicted. Red dashed circles indicate clashes between the waters observed in the HIV apo structure and PFV bound 6v structure, whereas the arrows point to waters that overlap with the sulfonyl group of MES. The protein backbone (depicted in light orange), Mg²⁺ cofactors (green), the penultimate cytosine of the viral DNA (light brown), and active site DDE motif (white) are labeled. The surface envelope of the unprocessed 3′ vDNA end (brown mesh) and VDW surface representation of MES (yellow and red) are also labeled.

Figure 7

Modeling of 5′g in the HIV-1 intasome. The binding of 5′g (yellow) from the PFV crystal structure (PDB ID: 5MMA) was superimposed onto the structure of HIV-1 IN in complex with 4d (PDB ID: 6PUY) represented by its surface (white density). The penultimate cytosine of the viral DNA end white density) is labeled along with the Mg²⁺ cofactors (green) and catalytic residues of the IN active site (light gray). Water molecules from HIV apo intasome within 4 Å away from 5′g (cyan) and red dashed circles are depicted to reveal clashes between the waters of apo HIV-1 intasome structure and the binding of 5′g into the active site of the HIV-1 intasome.

Figure 8

Modeling 5j into the active site of the HIV-1 intasome. (A) 5j (magenta) is docked onto the structure of 4d bound to the HIV-1 intasome, which is represented by its surface (white density). Two different rotameric conformations of the terminal adenine at the end of the viral DNA (dA21a and dA21b, labeled cream) are shown, along with the penultimate cytosine of the viral DNA end (dC20, surface map and labeled in light gray), the Mg²⁺ cofactors (green), and catalytic residues of the IN active site (gray). Water molecules (cyan) that lie in close proximity to 5j are labeled, and red dashed circles are depicted to reveal clashes between the waters of apo HIV-1 intasome and the binding of 5j into the active site of the HIV-1 intasome. The assigned water molecules from the PFV-5j intasome structure (red) and black dashed lines indicate waters within 4 Å distance from the polar groups of 5j. An arrow indicates a conserved water molecule between both models (PFV-5j and HIV-apo). Bound MES is also depicted (gray). (B) Cluster of potential interactions involving the hydroxyl group at 5-position of 5j (magenta) with the terminal adenine (dA21a, gold) and water molecules from HIV-1 intasome apo (cyan) in the active site of the HIV-1 intasome. Red dashed circles are depicted to reveal clashes involving these water molecules. Water molecules from PFV-5j model within 4 Å from 5j (red) are also depicted.

Figure 9

Model of the binding of INSTIs to the HIV-1 IN mutant G118R. BIC (blue), 5′g (yellow), 5j (magenta), and 6v (green) are superimposed in the active site of the HIV-1 intasome. Rotamer possibilities for the mutant G118R are depicted to show the potential steric hindrance of this mutation on the binding of INSTIs. Mg²⁺ ions (green), penultimate cytosine of the viral DNA (gray), IN residues G118 and Y143 (dark and light gray, respectively), and IN mutant G118R (mild gray) are labeled.