Search

Efficient and accurate models to predict the fitness of a sequence would be extremely valuable in protein design. We have explored the use of statistical potentials for the coevolutionary fitness landscape, extracted from known protein sequences, in conjuction with Monte Carlo simulations, as a tool for design. As proof of principle, we created a series of predicted high-fitness sequences for three different protein folds, representative of different structural classes: the GA (all-α) and GB (α/β) binding domains of streptococcal protein G, and an SH3 (all-β) domain. We found that most of the designed proteins can fold stably to the target structure, and a structure for a representative of each for GA, GB and SH3 was determined. Several of our designed proteins were also able to bind to native ligands, in some cases with higher affinity than wild-type. Thus, a search using a statistical fitness landscape is an remarkably effective tool for finding novel stable protein sequences.

We have systematically validated the activity and inhibition of a HIV-1 protease (PR) variant bearing 17 mutations (PR(S17)), selected to represent high resistance by machine learning on genotype-phenotype data. Three of five mutations in PR(S17) correlating with major drug resistance, M46L, G48V, and V82S, and five of 11 natural variations differ from the mutations in two clinically derived extreme mutants, PR20 and PR22 bearing 19 and 22 mutations, respectively. PR(S17), which forms a stable dimer (10 nM), is ∼10- and 2-fold less efficient in processing the Gag polyprotein than the wild type and PR20, respectively, but maintains the same cleavage order. Isolation of a model precursor of PR(S17) flanked by the 56-amino acid transframe region (TFP-p6pol) at its N-terminus, which is impossible upon expression of an analogous PR20 precursor, allowed systematic comparison of inhibition of TFP-p6pol-PR(S17) and mature PR(S17). Resistance of PR(S17) to eight protease inhibitors (PIs) relative to PR (Ki) increases by 1.5-5 orders of magnitude from 0.01 to 8.4 μM. Amprenavir, darunavir, atazanavir, and lopinavir, the most effective of the eight PIs, inhibit precursor autoprocessing at the p6pol/PR site with IC50 values ranging from ∼7.5 to 60 μM. Thus, this process, crucial for stable dimer formation, shows inhibition ∼200-800-fold weaker than that of the mature PR(S17). TFP/p6pol cleavage, which occurs faster, is inhibited even more weakly by all PIs except darunavir (IC50 = 15 μM); amprenavir shows a 2-fold increase in IC50 (∼15 μM), and atazanavir and lopinavir show increased IC50 values of42 and70 μM, respectively.

The escape mutant of HIV-1 protease (PR) containing 20 mutations (PR20) undergoes efficient polyprotein processing even in the presence of clinical protease inhibitors (PIs). PR20 shows >3 orders of magnitude decreased affinity for PIs darunavir (DRV) and saquinavir (SQV) relative to PR. Crystal structures of PR20 crystallized with yttrium, substrate analog p2-NC, DRV and SQV reveal three distinct conformations of the flexible flaps and diminished interactions with inhibitors through the combination of multiple mutations. PR20 with yttrium at the active site exhibits widely separated flaps lacking the usual intersubunit contacts seen in other inhibitor-free dimers. Mutations of residues 35–37 in the hinge loop eliminate interactions and perturb the flap conformation. Crystals of PR20/p2-NC contain one uninhibited dimer with one very open flap and one closed flap, and a second inhibitor-bound dimer in the closed form showing six fewer hydrogen bonds with the substrate analog relative to wild type enzyme. PR20 complexes with PIs exhibit expanded S2/S2′ pockets and fewer PI interactions arising from coordinated effects of mutations throughout the structure, in agreement with the strikingly reduced affinity. In particular, insertion of the large aromatic side chains of L10F and L33F alters intersubunit interactions and widens the PI binding site through a network of hydrophobic contacts. The two very open conformations of PR20 as well as the expanded binding site of the inhibitor-bound closed form suggest possible approaches for modifying inhibitors to target extreme drug resistant HIV.

Purification and in vitro protein-folding schemes were developed to produce monodisperse samples of the mature wild-type HIV-2 protease (PR2), enabling a comprehensive set of biochemical and biophysical studies to assess the dissociation of the dimeric protease. An E37K substitution in PR2 significantly retards autoproteolytic cleavage during expression. Furthermore, it permits convenient measurement of the dimer dissociation of PR2E37K (elevated Kd ∼20 nM) by enzyme kinetics. Differential scanning calorimetry reveals a Tm of 60.5 for PR2 as compared with 65.7°C for HIV-1 protease (PR1). Consistent with weaker binding of the clinical inhibitor darunavir (DRV) to PR2, the Tm of PR2 increases by 14.8°C in the presence of DRV as compared with 22.4°C for PR1. Dimer interface mutations, such as a T26A substitution in the active site (PR2T26A) or a deletion of the C-terminal residues 96–99 (PR21–95), drastically increase the Kd (>105-fold). PR2T26A and PR21–95 consist predominantly of folded monomers, as determined by nuclear magnetic resonance (NMR) and size-exclusion chromatography coupled with multiangle light scattering and refractive index measurements (SMR), whereas wild-type PR2 and its active-site mutant PR2D25N are folded dimers. Addition of twofold excess active-site inhibitor promotes dimerization of PR2T26A but not of PR21–95, indicating that subunit interactions involving the C-terminal residues are crucial for dimer formation. Use of SMR and NMR with PR2 facilitates probing for potential inhibitors that restrict protein folding and/or dimerization and, thus, may provide insights for the future design of inhibitors to circumvent drug resistance.

No drug has been targeted specifically for HIV-2 (human immunodeficiency virus type 2) infection despite its increasing prevalence worldwide. The antiviral HIV-1 (human immunodeficiency virus type 1) protease (PR) inhibitor darunavir and the chemically related GRL98065 and GRL06579A were designed with the same chemical scaffold and different substituents at P2 and P2′ to optimize polar interactions for HIV-1 PR (PR1). These inhibitors are also effective antiviral agents for HIV-2-infected cells. Therefore, crystal structures of HIV-2 PR (PR2) complexes with the three inhibitors have been solved at 1.2-A resolution to analyze the molecular basis for their antiviral potency. Unusually, the crystals were grown in imidazole and zinc acetate buffer, which formed interactions with the PR2 and the inhibitors. Overall, the structures were very similar to the corresponding inhibitor complexes of PR1 with an RMSD of 1.1 A on main-chain atoms. Most hydrogen-bond and weaker C–H…O interactions with inhibitors were conserved in the PR2 and PR1 complexes, except for small changes in interactions with water or disordered side chains. Small differences were observed in the hydrophobic contacts for the darunavir complexes, in agreement with relative inhibition of the two PRs. These near-atomic-resolution crystal structures verify the inhibitor potency for PR1 and PR2 and will provide the basis for the development of antiviral inhibitors targeting PR2.

The HIV-1 protease enzyme is indispensable for viral maturation, making it a major potential target of anti-HIV therapy. Its role is to split newly formed Gag and Gag-Pol polyproteins to produced finished structural and functional proteins — itself included. The early events in the autoprocessing of HIV-1 protease, in which a precursor dimer is thought to undergo intramolecular cleavage, have now been visualized using NMR spectroscopy and paramagnetic relaxation enhancement. This reveals that although primarily monomeric, the protease is also present as transient encounter complexes that occupy a wide range of orientations relative to the mature dimer. The N-terminal region makes transient intra- and intersubunit contacts with the substrate binding site, allowing autocleavage to occur when the correct dimer orientation is sampled by the encounter complex. HIV-1 protease processes the Gag and Gag-Pol polyproteins into mature structural and functional proteins. The mature protease is only active as a dimer, with catalytic residues contributed by each subunit. The precursor of the active protease undergoes 'maturation' via the intramolecular cleavage of a dimeric species, but it is not clear how this cleavage reaction occurs. The early events in N-terminal auto-processing were visualized using NMR spectroscopy, and it was determined that the precursor forms transient, lowly populated dimeric encounter complexes that occupy a wide range of orientations relative to the mature dimer. The N-terminal region makes transient intra- and intersubunit contacts with the substrate binding site, enabling auto-cleavage to occur when the correct dimer orientation is sampled by the encounter complex ensemble. HIV-1 protease processes the Gag and Gag-Pol polyproteins into mature structural and functional proteins, including itself, and is therefore indispensable for viral maturation1,2. The mature protease is active only as a dimer3,4,5 with each subunit contributing catalytic residues6. The full-length transframe region protease precursor appears to be monomeric yet undergoes maturation via intramolecular cleavage of a putative precursor dimer5,7,8,9,10,11, concomitant with the appearance of mature-like catalytic activity7,9. How such intramolecular cleavage can occur when the amino and carboxy termini of the mature protease are part of an intersubunit β-sheet located distal from the active site is unclear. Here we visualize the early events in N-terminal autoprocessing using an inactive mini-precursor with a four-residue N-terminal extension that mimics the transframe region protease precursor5,12. Using paramagnetic relaxation enhancement, a technique that is exquisitely sensitive to the presence of minor species13,14,15,16, we show that the mini-precursor forms highly transient, lowly populated (3–5%) dimeric encounter complexes that involve the mature dimer interface but occupy a wide range of subunit orientations relative to the mature dimer. Furthermore, the occupancy of the mature dimer configuration constitutes a very small fraction of the self-associated species (accounting for the very low enzymatic activity of the protease precursor), and the N-terminal extension makes transient intra- and intersubunit contacts with the substrate binding site and is therefore available for autocleavage when the correct dimer orientation is sampled within the encounter complex ensemble.

N-Terminal self-cleavage (autoprocessing) of the HIV-1 protease precursor is crucial for liberating the active dimer. Under drug pressure, evolving mutations are predicted to modulate autoprocessing, and the reduced catalytic activity of the mature protease (PR) is likely compensated by enhanced conformational/dimer stability and reduced susceptibility to self-degradation (autoproteolysis). One such highly evolved, multidrug resistant protease, PR20, bears 19 mutations contiguous to sites of autoproteolysis in retroviral proteases, namely clusters 1-3 comprising residues 30-37, 60-67, and 88-95, respectively, accounting for 11 of the 19 mutations. By systematically replacing corresponding clusters in PR with those of PR20, and vice versa, we assess their influence on the properties mentioned above and observe no strict correlation. A 10-35-fold decrease in the cleavage efficiency of peptide substrates by PR20, relative to PR, is reflected by an only ∼4-fold decrease in the rate of Gag processing with no change in cleavage order. Importantly, optimal N-terminal autoprocessing requires all 19 PR20 mutations as evaluated in vitro using the model precursor TFR-PR20 in which PR is flanked by the transframe region. Substituting PR20 cluster 3 into TFR-PR (TFR-PR(PR20-3)) requires the presence of PR20 cluster 1 and/or 2 for autoprocessing. In accordance, substituting PR clusters 1 and 2 into TFR-PR20 affects the rate of autoprocessing more drastically (>300-fold) compared to that of TFR-PR(PR20-3) because of the cumulative effect of eight noncluster mutations present in TFR-PR20(PR-12). Overall, these studies imply that drug resistance involves a complex synchronized selection of mutations modulating all of the properties mentioned above governing PR regulation and function.

The envelope glycoprotein gp41 mediates the process of membrane fusion that enables entry of the HIV-1 virus into the host cell. Strong lipid affinity of the ectodomain suggests that its heptad repeat regions play an active role in destabilizing membranes by directly binding to the lipid bilayers and thereby lowering the free-energy barrier for membrane fusion. In such a model, immediately following the shedding of gp120, the N-heptad and C-heptad helices dissociate and melt into the host cell and viral membranes, respectively, pulling the destabilized membranes into juxtaposition, ready for fusion. Post-fusion, reaching the final 6-helix bundle (6 HB) conformation then involves competition between intermolecular interactions needed for formation of the symmetric 6 HB trimer and the membrane affinity of gp41's ectodomain, including its membrane-proximal regions. Our solution NMR study of the structural and dynamic properties of three constructs containing the ectodomain of gp41 with and without its membrane-proximal regions suggests that these segments do not form inter-helical interactions until the very late steps of the fusion process. Interactions between the polar termini of the heptad regions, which are not associating with the lipid surface, therefore may constitute the main driving force initiating formation of the final post-fusion states. The absence of significant intermolecular ectodomain interactions in the presence of dodecyl phosphocholine highlights the importance of trimerization of gp41's transmembrane helix to prevent complete dissociation of the trimer during the course of fusion.

We previously reported a series of antibodies, in fragment antigen binding domain (Fab) formats, selected from a human non-immune phage library, directed against the internal trimeric coiled-coil of the N-heptad repeat (N-HR) of HIV-1 gp41. Broadly neutralizing antibodies from that series bind to both the fully exposed N-HR trimer, representing the pre-hairpin intermediate state of gp41, and to partially-exposed N-HR helices within the context of the gp41 six-helix bundle. While the affinities of the Fabs for pre-hairpin intermediate mimetics vary by only 2 to 20-fold between neutralizing and non-neutralizing antibodies, differences in inhibition of viral entry exceed three orders of magnitude. Here we compare the binding of neutralizing (8066) and non-neutralizing (8062) antibodies, differing in only four positions within the CDR-H2 binding loop, in Fab and single chain variable fragment (ScFv) formats, to several pre-hairpin intermediate and six-helix bundle constructs of gp41. Residues 56 and 58 of the mini-antibodies are shown to be crucial for neutralization activity. There is a large differential (≥ 150-fold) in binding affinity between neutralizing and non-neutralizing antibodies to the six-helix bundle of gp41 and binding to the six-helix bundle does not involve displacement of the outer C-terminal helices of the bundle. The binding stoichiometry is one six-helix bundle to one Fab or three ScFvs. We postulate that neutralization by the 8066 antibody is achieved by binding to a continuum of states along the fusion pathway from the pre-hairpin intermediate all the way to the formation of the six-helix bundle, but prior to irreversible fusion between viral and cellular membranes.

Dimerization is indispensible for release of the human immunodeficiency virus protease (PR) from its precursor (Gag-Pol) and ensuing mature-like catalytic activity that is crucial for virus maturation. We show that a single-chain Fv fragment (scFv) of a previously reported monoclonal antibody (mAb1696), which recognizes the N-terminus of PR, dissociates a dimeric mature D25N PR mutant with an enhanced dimer dissociation constant (K(d)) in the sub-micromolar range to form predominantly a monomer-scFv complex at a 1:1 ratio, along with small (5-10%) amounts of a dimer-scFv complex. Enzyme kinetics indicate a mixed mechanism of inhibition of the wild-type PR, which exhibits a K(d)10nM, with effects both on K(m) and k(cat) at an scFv-to-PR ratio of 10:1. ScFv binds to the N-terminal peptide P(1)QITLW(6) of PR and to PR monomers with dissociation constants of ≤30 nM and ~100 nM, respectively. Consistent with an ~400-fold increase in the dissociation of the antibody (K(Ab)) on even addition of an acetyl group to P(1) of the peptide, the antibody fails to inhibit N-terminal autoprocessing of the PR from a model precursor (at ~5 μM). However, subsequent to this cleavage, it sequesters the PR, thus blocking autoprocessing at its C-terminus. A second monoclonal antibody [PRM1 (human monoclonal antibody to PR)], which recognizes part of the flap region (residues 41-47) of the mature PR and its precursor, does not inhibit autoprocessing and ensuing catalytic activity. However, its failure to recognize drug-resistant clinical mutants of PR may be beneficial to monitor the selection of mutations in this region under drug pressure.

Self-cleavage at the N terminus of HIV-1 protease from the Gag-Pol precursor (autoprocessing) is crucial for stabilizing the protease dimer required for onset of mature-like catalytic activity, viral maturation, and propagation. Among nine clinical protease inhibitors (PIs), darunavir and saquinavir were the most effective in inhibiting wild-type HIV-1 group M precursor autoprocessing, with an IC 50 value of 1–2 μM, 3–5 orders of magnitude higher than their binding affinities to the corresponding mature protease. Accordingly, both group M and N precursor–PI complexes exhibit T m s 17–21 °C lower than those of the corresponding mature protease–PI complexes suggestive of markedly reduced stabilities of the precursor dimer–PI ensembles. Autoprocessing of group N (natural variant) and three group M precursors bearing 11–20 mutations associated with multidrug resistance was either weakly responsive or fully unresponsive to inhibitors at concentrations up to a practical limit of approximately 150 μM PI. This observation parallels decreases of up to 8 × 10 3 -fold (e.g., 5 pM to 40 nM) in the binding affinity of darunavir and saquinavir to mature multidrug resistant proteases relative to wild type, suggesting that inhibition of some of these mutant precursors will occur only in the high μM to mM range in extreme PI-resistance, which is an effect arising from coordinated multiple mutations. An extremely darunavir-resistant mutant precursor is more responsive to inhibition by saquinavir. These findings raise the questions whether clinical failure of PI therapy is related to lack of inhibition of autoprocessing and whether specific inhibitors can be designed with low-nM affinity to target autoprocessing.

Using a human non-immune phage library comprising more than 10 9 functional human antibody specificities in Fab format, we have been able to select a set of eight monoclonal Fabs targeted against diverse epitopes of the ectodomain of gp41 from HIV-1. The antigens used for panning the antibodies comprised two soluble, disulfide-linked, trimeric polypeptides derived from gp41, N CCG -gp41 and N35 CCG -N13. The former comprises an exposed trimeric coiled-coil of the N-helices of gp41 fused in helical phase to the minimal thermostable ectodomain of gp41, while the latter comprises only the trimeric coiled-coil of N-helices. The selected Fabs were probed by Western blot analysis against four antigens: N CCG -gp41, N35 CCG -N13, N34 CCG (a smaller version of N35 CCG -N13), and the minimal thermostable ectodomain core of gp41 in its six-helix bundle conformation (6-HB). Three classes of Fabs were found: class A (two Fabs) interact predominantly with the 6-HB; class B (four Fabs) interact with both the 6-HB and the internal trimeric coiled-coil of N-helices; and class C (two Fabs) interact specifically with the internal trimeric coiled-coil of N-helices. The IC 50 values for the Fabs, expressed as bivalent mini-antibodies, ranged from 6 μg/ml to 60 μg/ml in a quantitative vaccinia virus-based reporter gene assay for HIV-1 envelope-mediated cell fusion using the envelope from the HIV-1 T tropic strain LAV. The two most potent fusion inhibitors belonged to class B. This panel of Fabs provides a set of useful probes for studying HIV-1 envelope-mediated cell fusion and may serve as a basis for developing Fab-based anti-HIV-1 therapeutics.