iconstar paper   HIV Articles  
Back grey arrow rt.gif
Next Generation anti-HIV Agents - Capsid Inhibitors
  Large-Scale Functional Purification of Recombinant HIV-1 Capsid
Magdeleine Hung, Gilead Sciences Inc., Foster City, California, United States of America Anita Niedziela-Majka, Gilead Sciences Inc., Foster City, California, United States of America Debi Jin, Gilead Sciences Inc., Foster City, California, United States of America Melanie Wong, Gilead Sciences Inc., Foster City, California, United States of America Stephanie Leavitt, Gilead Sciences Inc., Foster City, California, United States of America Katherine M. Brendza, Gilead Sciences Inc., Foster City, California, United States of America Xiaohong Liu, Gilead Sciences Inc., Foster City, California, United States of America Roman Sakowicz * E-mail: roman.sakowicz@gilead.com Gilead Sciences Inc., Foster City, California, United States of America
Plos One March 5, 2013
We have used the purified capsid protein to characterize two known assembly inhibitors in our in-house developed polymerization assay and to measure their binding affinities

During human immunodeficiency virus type-1 (HIV-1) virion maturation, capsid proteins undergo a major rearrangement to form a conical core that protects the viral nucleoprotein complexes. Mutations in the capsid sequence that alter the stability of the capsid core are deleterious to viral infectivity and replication. Recently, capsid assembly has become an attractive target for the development of a new generation of anti-retroviral agents. Drug screening efforts and subsequent structural and mechanistic studies require gram quantities of active, homogeneous and pure protein. Conventional means of laboratory purification of Escherichia coli expressed recombinant capsid protein rely on column chromatography steps that are not amenable to large-scale production. Here we present a function-based purification of wild-type and quadruple mutant capsid proteins, which relies on the inherent propensity of capsid protein to polymerize and depolymerize. This method does not require the packing of sizable chromatography columns and can generate double-digit gram quantities of functionally and biochemically well-behaved proteins with greater than 98% purity. We have used the purified capsid protein to characterize two known assembly inhibitors in our in-house developed polymerization assay and to measure their binding affinities. Our capsid purification procedure provides a robust method for purifying large quantities of a key protein in the HIV-1 life cycle, facilitating identification of the next generation anti-HIV agents.

Human immunodeficiency virus (HIV) infections require a lifelong therapy. Despite the availability of highly effective anti-HIV drugs, the development of drug-resistant HIV-1 variants remains a major threat to the patient population. To overcome the emergence of drug resistance during current therapies, there is a constant need for the discovery of new classes of antiretroviralsto augment existing HIV treatment regimen.
HIV capsid protein (CA) represents such a target, with biologically validated importance in HIV life cycle but requiring a clinical proof of concept. During HIV virion maturation, CA is released from the Gag polyprotein by the viral protease [1]. Despite the presence of other structural and non-structural proteins in the maturing virion, ∼1,500 CA monomers assemble into a lattice of ∼ 250 CA hexamers interspersed with ∼12 pentamers that together form a distinct fullerene cone encapsulating the viral RNA. Upon entry of HIV particles into host cells, the CA core disassembles in a coordinated fashion to allow reverse transcription and subsequent integration of the reverse-transcribed viral genome into the host DNA. Stability of the CA core and the corresponding rate of core disassembly are essential for successful viral infection [2], [3]. A number of deleterious surface mutations in CA protein were reported, that alter the infectivity, replication and assembly of virions in vivo [4]. Disruption of proper CA core assembly during particle maturation and/or destabilization of the incoming CA core [5], which causes premature disassembly, may enable new antiviral strategies that target an essential part in the life cycle of the HIV virus. Currently, a number of small molecules with the potential to interfere with CA assembly pathway are under development [6]. Such efforts are critically dependent on the availability of large quantities of homogeneous and active CA protein. A typical in vitro CA assembly assay uses large quantities of recombinant CA protein at concentrations ranging from 60-200 μM. CA polymerization is monitored spectrophotometrically by measuring the increase in absorbance at λ350 nm due to the light scattering caused by polymerized tubular structures [7]-[9]. Due to the high concentration of protein required in the assay, gram quantities of CA are needed to support high-throughput screening efforts to identify inhibitors of the polymerization process. Large amounts of CA are also essential for structural studies to supplement rational drug discovery efforts. The existing published procedures describing CA purification methods are not conducive for such large scale purification efforts. One earlier report on recombinant CA purification by Gross et al (1997) relied on multiple rounds of ammonium sulfate precipitation to separate the CA proteins from E. coli contaminant proteins. Later protocols adopted differential ammonium sulfate fractionation as the first step coupled with either anion exchange chromatography using Tris buffer at pH 8.1 [10] or cation exchange chromatography in KMOPS buffer at pH 6.9 [11]. As there is a limit to the dynamic binding capacity of any chromatographic medium, a sizable volume of chromatography matrix would be required to capture double-digit gram quantities of CA protein. In addition, liters of buffer would be required for the column equilibration, loading and elution. Thus, a chromatography-based approach is not amenable to gram-scale protein purification in a traditional laboratory setting. Moreover, CA protein that is purified with differential ammonium sulfate fractionation or traditional chromatography steps is refined by exploiting its biochemical properties rather than its functional competency.
Here, we describe a novel CA purification method exploiting its innate ability to polymerize and depolymerize in vitro. This new protocol enables rapid purification of double-digit gram quantities of pure and functional HIV-1 CA protein without the need to separate the target protein on any chromatographic medium. The final product is highly enriched in polymerization-competent capsid protein, ensuring high functional activity. We have successfully used this protein to study wild-type capsid (WT CA) polymerization, to form covalent CA hexamers [12] and to monitor the effects of known CA inhibitors.
Here we describe a method to purify HIV-1 CA protein by exploiting its ability to reversibly polymerize. This method facilitates the purification of CA protein from E. coli lysate without investment in bulk chromatographic resins normally used for protein fractionation. Repeated cycles of polymerization/depolymerization had been previously employed for efficient purification of other polymeric proteins like actin or tubulin from animal tissues [26]. Moreover, the incorporation of two cycles of CA polymerization and depolymerization ensured that only functional CA was purified. Unlike column fractionation, CA protein purified by this novel purification scheme can be prepared at relatively high concentrations of ∼20 mg/mL without the need of downstream concentration steps as the protein can be directly solubilized in the desired type and volume of buffer and the purity is maintained at >98% level.
We can further extend this polymerization/depolymerization approach to the purification of large quantities of the CA 4Mu ([A14C/E45C/W184A/M185A] CA). The W184A and M185A mutations in the C-domain dimerization interface were reported to significantly decrease the ability of the CA protein to dimerize and consequently to impair cylinder formation in vitro [4], [21], [27]. However, it was also reported that the CA 4Mu monomers were able to form tubes at much higher protein concentration but with lower efficiency [28]. As observed from our assembly kinetics experiment for CA 4Mu (Figures 4C and 4D), the weakened dimer interface affinity could be compensated by the use of higher protein concentration (>75 μM) and the initial rates of polymerization at higher CA 4Mu concentrations were comparable to the rates of WT CA polymerization. Thus we were able to use the same functional purification scheme to isolate CA 4Mu from E. coli lysate by maintaining a significantly high protein concentration (>10 mg/mL) throughout the process. In addition, the ability of CA 4Mu to polymerize could be enhanced by the formation of disulfide bonds between the engineered cysteine residues in the amino-terminal domain (NTD). This was further supported by the observation that we could achieve complete dissolution of the polymerized tubes formed by the quadruple mutant only in the presence of reducing agent in resolubilization buffer, for example 200 mM β-mercaptoethanol. It is worth noting that CA proteins with only C-terminal domain (CTD) dimerization interface mutations (W184A/M185A CA) were unable to polymerize under identical conditions (data not shown) and we cannot adopt the functional purification scheme in isolation of this recombinant protein.
We showed that both the wild-type and quadruple mutant CA proteins purified with this novel protocol were functionally active in the polymerization assay and the assembly process was attenuated by two well-described polymerization inhibitors: CAI-4, a variant of CAI peptide and the small molecule inhibitor BM2. Although we obtained ∼ 1 log higher IC50 values than previously reported [13], [24], this could be attributed to differences in assay formats employed in the determination of compound inhibition on capsid assembly, such as the type and the amount of capsid proteins.
In addition, we utilized the purified WT CA as well as the assembled CA 4Mu hexamer to measure the binding affinity to the two inhibitors using SPR methods. Although both CAI-4 peptide and BM2 inhibited CA polymerization with similar single digit micromolar potencies, there was a 10-fold difference in their respective binding affinities to the wild-type protein as shown in Figure 6B and 6E. It would be interesting to see if there is any correlation of their binding affinities to the cellular potencies. Due to the lack of cell permeability of CAI peptide, there was no direct measurement of its in vivo efficacy. However, there was a reasonable correlation between the dissociation constants of small molecules binding to CANTD and their respective antiviral potencies (EC50 values) [13]. With the availability of large amount of well-behaved protein, the SPR assay could serve as an alternative, higher throughput tool to rank potency of capsid assembly inhibitors by characterizing their binding affinities to WT CA.
Interestingly, although both inhibitors display micromolar potencies in inhibition of CA polymerization and micromolar equilibrium dissociation constants for binding to wild-type CA, they do not bind to assembled hexamer protein. Based on the structures of CANTD and CACTD domains bound to BM2 and CAI respectively, as well as modeling of the complexes in the assembled CA, their binding would sterically hinder the formation of interface contacts that would be necessary for the formation of higher-order assembled CA structures [13], [24], [25]. Thus, it would not be surprising to see a loss in the abilities of BM2 and CAI to bind to the pre-formed hexamers as the newly formed intermolecular contacts in the hexameric structure might preclude accessibility of the binding pockets to the respective inhibitors [12]. Instead, the hexameric structure would present new interfaces for the interaction with novel classes of small molecules inhibitors. In fact, we have preliminary evidence that a series of CA assembly accelerators, which is represented by PF-3450074 [29], actually displays stronger affinity to the hexamer over wild-type CA monomer (data not shown). Binding of PF-3450074 to CANTD and wild-type CA have been previously determined by isothermal titration calorimetry with comparable single-digit micromolar affinities [29]. Although there is no published data of its binding to capsid hexamer, the model of PF-3450074 in complex with capsid hexamer indicates its potential role in modulating inter-subunit interactions [29]. Hence, the differential binding affinities towards WT CA and the 4Mu hexamer demonstrated by these two classes of CA assembly modulators as measured by the SPR assay could provide an insight into the mode of action of various classes of small molecules.
In conclusion, we have developed a fast and functional approach to purify large quantities of HIV CA protein using its assembly competency. The wild-type CA proteins can subsequently be used in a high-throughput screening campaign to identify new chemical matter that may potentially become the next generation of assembly inhibitors. The quadruple mutant CA can be assembled into a hexamer and employed in the SPR binding assay to evaluate the binding affinity of inhibitors towards hexameric CA structures as part of the mode of action characterization during drug discovery effort. This new purification method will provide a significant boost to the field of HIV CA research by enabling the generation of abundant starting material for biochemical, biophysical and structural studies.

  iconpaperstack View Older Articles   Back to Top   www.natap.org