Created by: Julia Moy

The HIV integrase in complex with inhibitor (PDB ID: 4NYF) from Homo sapien is a polynucleotidyl transferase protein complexed with the transferase inhibitor BI 224436. HIV integrase is an enzyme responsible for the integration of HIV viral DNA into the host cell genome (1,2). HIV integrase binds to viral DNA and cleaves a dinucleotide from each 3’-end in the cytoplasm. The preintegration complex of integrase and viral DNA then enters the nucleus and covalently links the 3’-ends of the viral DNA to the 5’-ends of the host cell DNA (1). Integrase is a transferase protein because the linkage between the viral and host DNA is called a strand transfer. The BI 224436 inhibitor of HIV integrase is of special importance for its potential to provide new antiretroviral therapies. Existing drugs for HIV target the other two enzymes crucial to the viral replication of HIV, reverse transcriptase and HIV protease, but HIV integrase is a new target for HIV drug therapies and therefore very important in potentially curing HIV and AIDS (1,2).

HIV integrase was crystallized with the single nucleotide substitution of Lys-185 for Phe-185 (2). This substitution of a hydrophilic residue for a hydrophobic residue greatly increased the solubility of the protein, allowing for its concentration and crystallization. The HIV integrase containing the Lys-185 substitution was expressed in E. coli and purified by Ni affinity chromatography. The protein was crystallized by the siting drop vapor diffusion method over several weeks and the crystalline structure was determined by x-ray diffraction (2). 

The catalytic activity of HIV integrase depends on its structure. The molecular weight and isoelectric point of HIV integrase was calculated with Expasy to be 36981.37 daltons and 6.68, respectively (3). HIV integrase as modeled here is a dimer of two nearly identical domains. The dimeric nature allows HIV integrase to attach to DNA at both domains for strand processing and transfer. The domains consist of a central five-strand beta pleated sheet and six alpha helices (1,4). The two domains interact at an interface away from the solvent connected by multiple salt bridges and hydrogen bonds. The side chain of Lys-185 of one HIV integrase monomer hydrogen bonds to the carbonyl oxygen of Ala-105 of the other monomer, connecting them into a dimer (1). It is known that the active form of HIV integrase for 3’-end processing and strand transfer is a multimer, but it is not yet known whether the functional unit is dimeric or higher order. In the dimer model, each monomer processes a viral DNA strand, and then the complex is likely to rearrange to position the processed 3’-ends of viral DNA and link to the 5’-ends of the host cell DNA (1). The linkage between viral and DNA strand occurs via nucleophilic attack of the phosphodiester bonds of the host cell DNA by the hydroxyl groups at the 3’-end of the viral DNA at the active site of each of the two subunits (1,4). Divalent metal ions are essential for integrase activity and there are two Cd2+ ions per domain that act as cofactors close to the active site. Parallel studies of HIV integrase and related retroviruses, specifically avian sarcoma and leucosis virus, have helped advance the understanding of the biochemical mechanism of integration (4). In the presence of divalent metal ions, retroviral integrases of different types catalyze both the processing of the 3’-ends of viral DNA and strand transfer to a DNA substrate that mimics viral DNA in experiments. These experiments have shown that, like other integrases, the mechanism of HIV integrase is likely a one-step transesterification with chiral inversion in the integration product (4).  

The three amino acid residues Asp-64, Asp-116 and Glu-152 seem to define the active site of the enzyme and be responsible for catalytic activity (1,4). These enzymes follow the retroviral Asp,Asp-35-Glu  motif and likely coordinate the two divalent metal ions involved in catalysis. The Asp,Asp-35-Glu motif is observed in the integrase proteins of retroviruses and retrotransposons and it has been shown that mutation of the motif causes loss of enzymatic function (1,4). Many polynucleotidyl transfer proteins have this motif of three acidic residues in the active site that catalyze strand transfer (1). 

HIV integrase has tertiary structure similarities to the HIV-1 reverse transcriptase ribonuclease H domain (PDB: 1HRH) (1). However, the secondary structures have structural differences. The first four strands of the beta sheet of integrase superimpose well on ribonuclease H (RNase H), and alpha-1 of integrase is displaced by 6 Angstrom relative to alpha-1 of RNase H. Alpha-2 is a three-turn helix in RNase H and a one-turn helix in HIV integrase. Alpha-3 is similar in both structures, but alpha-4 is very different and alpha-5 and alpha-6 are present in HIV integrase but not in RNase H (1). Despite this lack of similarity in sequence, HIV integrase and RNase H have a great similarity in two catalytic residue positions. Two of the conserved acidic residues in the active site of HIV integrase, Asp-64 and Asp-116 superimpose very well on two catalytic residues of RNase H, Asp-443 and Asp-498 (1). The third catalytic residue of HIV integrase, Glu-152, does not superimpose well. The two aspartates are most likely the most highly conserved residues of the active sites (1). 

(2S)-tert-butoxy[4-(4-chlorophenyl)-2-methylquinolin- 3-yl]ethanoic acid (BI 224436) is a noncatalytic site inhibitor complexed with the HIV integrase protein.BI 224436 binds to HIV integrase distal from the active site and inhibits the 3’-end processing activity of the enzyme via allosteric inhibition (2,5). The BI 224436 inhibitor binds to the allosteric site and causes a conformational change that has not yet been crystallized. The active site cannot process the 3’-ends of viral DNA because the DNA substrate cannot bind to the active site’s changed conformation. BI 224436 and other allosteric HIV-1 integrase inhibitors interfere with the binding of lens epithelium-derived growth factor protein (LEDGF/p75) (PBD ID: 1Z9E) that binds its N-terminal domain to viral DNA and its C-terminal domain to integrase, enhancing the integration by increasing the processing of viral DNA by HIV integrase (4,5).  Studies of the mechanism of inhibition indicate that LEDGF-binding inhibitors effect integrase oligomerization (5). Allosteric inhibitors of HIV integrase are a new addition to clinically approved active site strand transfer inhibitors. Strand transfer inhibitors include Raltegravir and Elvitegravir and inhibit HIV integrase by binding to the preintegration complex of integrase and viral DNA and displacing the reactive 3’-end of the viral DNA and removing the essential divalent cations near the active site (6). A single tablet antiretroviral regimen for HIV includes the anti-HIV integrase strand transfer inhibitor Elvitegravir and two nucleoside reverse transcriptase inhibitors Emtricitabine and Tenofovir (6). 

HIV in complex with inhibitor BI 224436 was compared to similar structures using the Position-Specific Iterated Basic Local Assignment Search Tool (PSI-BLAST) and the Dali server. PSI-BLAST compares proteins’ primary structure and uses gaps in sequence to assign an E value to indicate level of similarity. An E value below 0.5 indicates a significant similarity (7). The Dali server compares proteins’ tertiary structure and assigns Z scores to indicate similarity. A Z score greater than 2 indicates a significant similarity (8). Using PSI-BLAST and the Dali server, the HIV-1 integrase catalytic core domain A128T mutant complexed with allosteric inhibitor (PDB ID: 4GVM) was found for comparison. The HIV-1 integrase catalytic domain A128T mutant has a low E value of 2e-106 and a high Z score of 23.2, indicating large similarities in both primary and tertiary structure (7,8). The HIV complex with inhibitor BI 224436 was compared to a mutant protein from the same organism because no similar proteins from other organisms were found using either database. The difference between the two human HIV integrase proteins further illuminates the effects of integrase inhibitors.

HIV-1 integrase catalytic domain A128T mutant has a substitution of Thr-128 for Ala-128 in the binding pocket of LEDGF/p75 at the dimer interface of HIV integrase (9). The mutation of Thr for Ala has been shown as the primary mechanism for resistance to allosteric HIV-1 integrase inhibitors such as the BI 224436 inhibitor complexed with the non-mutated HIV integrase protein. The A128T substitution changes the position of the inhibitor at the integrase dimer interface but does not effect the hydrogen bonding between integrase and LEDGF/p75, indicating that allosteric HIV integrase inhibitors work by targeting integrase multimerization rather than integrase-LEDGF/p75 binding (9). The inhibitor’s changed position makes it unable to target integrase multimerization and thus the A128T substitution results in significant resistance to allosteric HIV-1 integrase inhibitors. The discovery of HIV integrase A128T can enable the development of second generation allosteric HIV-1 integrase inhibitors with decreased potential to select for drug resistance (9).