Allosteric effects of chicoric acid on human immunodeficiency virus type 1 integrase

Abstract

Human immunodeficiency virus (HIV) integrase (IN) is an essential viral protein involved in the integration of the viral DNA into the host genome. Although having a specific catalytic function, it is apparent from mutagenesis studies that IN is pleiotropic and affects the viral life cycle at multiple points other than integration. Compounds that bind to allosteric sites on IN typically disrupt its ordered multimerization and stalls the viral life cycle at various points. Chicoric acid (CA) is a well-known IN inhibitor, however, its mechanism does not follow a conventional active site inhibition, yet it presents with antiviral activity. We thus, hypothesised that CA has an allosteric inhibitory mechanism. To test the hypothesis we aimed to determine alloteric effects of CA on HIV IN. Site directed mutagenesis was used to develop IN mutants resistant to conventional IN inhibitors and these proteins were purified. These were used in an enzyme-linked immunosorbent assay (ELISA) for comparative resistance profiling of CA and raltegravir (RAL). The ELISA also compared magnesium (Mg2+) and manganese (Mn2+) dependent differences on CA inhibition and to determine the importance of order of addition of assay components. An AlphaScreen assay was developed to test for the disruption of the IN/LEDGF interaction. Crosslinking assays and size exclusion chromatography (SEC) was performed to determine the multimeric state of IN in the presence of CA. Surface Plasmon resonance (SPR) was used to confirm binding of CA to IN and determine the kinetics. In silico docking onto the IN catalytic core (CCD) structures was used to identify a possible binding mode of CA at the allosteric binding site. Resistance profiling depicted a clear distinction between CA and RAL. IN resistance mutants: INQ148H, INN155H, ING140S/Q148H and INE92Q/N155H, showed a fold change in IC50 (FCIC50) of 3.42, 1.09, 1.43 and 2.90 for CA respectively. While the same mutants showed an FCIC50 of 947.99, 392.44, 1262.41 and 583.92 for RAL respectively. Additionally, cooperativity trends between the compound’s inhibition profiles were different. It was concluded that CA and RAL have different resistance profiles. Metal dependent differences show that the IN soluble mutant is resistant to CA inhibition only in the presence of Mg2+. This indicates that metal specific

structural differences may play a role in resistance. Order of addition indicated that if DNA was present after CA incubation with IN the inhibition was cooperative, while if DNA was added before incubation with CA the inhibition was non-cooperative. This posits that CA may bind to a single site, possibly only the allosteric site, when DNA is already present in the IN active site. When DNA is absent CA binds to multiple sites, possibly both the active site and allosteric sites. The AlphaScreen indicated that the IN/LEDGF interaction is disrupted by CA with IC50 of 237nM (±27nM). Also the disruption is dependent on the order of addition of assay components. Multimerization of IN was increased in the presence of CA as shown by crosslinking assays as well as SEC. Not only did CA induce multimerization of free IN, a nonreducing gel electrophoresis indicated that virus assembled in the presence of CA had increased multimerization of IN, similar to a control. SPR indicated that CA binds to the full length IN with similar kinetics compared to the catalytic core domain (INCCD) indicating that this domain likely contains the CA binding site. The kinetics indicated a slow kon and koff rate for CA binding. These slow binding kinetics are indicative of a high barrier to resistance. Finally, in sillico molecular docking of CA to the active site as well as the allosteric site indicated that it potentially has a dual binding mode on the IN apo-enzyme. The results show that CA has allosteric effects on IN and may provide a good pharmacophore for further development of allosteric IN inhibitors.