• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Sep 7, 2010; 107(36): 15910–15915.
Published online Aug 23, 2010. doi:  10.1073/pnas.1002346107
PMCID: PMC2936642

Structure-based modeling of the functional HIV-1 intasome and its inhibition


The intasome is the basic recombination unit of retroviral integration, comprising the integrase protein and the ends of the viral DNA made by reverse transcription. Clinical inhibitors preferentially target the DNA-bound form of integrase as compared with the free protein, highlighting the critical requirement for detailed understanding of HIV-1 intasome structure and function. Although previous biochemical studies identified integrase residues that contact the DNA, structural details of protein–protein and protein–DNA interactions within the functional intasome were lacking. The recent crystal structure of the prototype foamy virus (PFV) integrase–viral DNA complex revealed numerous details of this related integration machine. Structures of drug-bound PFV intasomes moreover elucidated the mechanism of inhibitor action. Herein we present a model for the HIV-1 intasome assembled using the PFV structure as template. Our results pinpoint previously identified protein–DNA contacts within the quaternary structure and reveal hitherto unknown roles for Arg20 and Lys266 in DNA binding and integrase function. Models for clinical inhibitors bound at the HIV-1 integrase active site were also constructed and compared with previous studies. Our findings highlight the structural basis for HIV-1 integration and define the mechanism of its inhibition, which should help in formulating new drugs to inhibit viruses resistant to first-in-class compounds.

Keywords: HIV/AIDS, integrase, integration, drug resistance, raltegravir

The integration of the linear DNA molecule made during reverse transcription into a cell chromosome is an essential step in the retroviral lifecycle. The key players in retroviral DNA integration are the integrase (IN) protein and the LTR end regions of the reverse transcript. Together with other viral and cellular factors, they comprise a high-molecular-weight preintegration complex (PIC) wherein IN catalyzes two successive chemical reactions. During 3′ processing, IN hydrolyzes two or three nucleotides from one or both DNA ends, exposing invariant CA dinucleotides. In the nucleus, IN catalyzes DNA strand transfer using the reactive 3′-OHs to cut phosphodiester bonds on opposing strands of target DNA, at the same time joining the LTR ends to the chromosome. The resulting strand transfer reaction intermediate is repaired by cellular enzymes to yield the integrated provirus (see ref. 1 for a comprehensive review).

Retroviral INs and other nucleic acid metabolizing enzymes, including RNase H, RuvC, and bacterial and eukaryotic transposases such as Tn5 and Mos1, respectively, belong to the polynucleotidyl transferase superfamily (24). Common features among these proteins include an RNase H-fold adopted by their catalytic domains and divalent metal ion-dependent bimolecular nucleophilic substitution (SN2) reactions at nucleic acid phosphodiester bonds mediated by carboxylate side chains of active-site residues. The retroviral INs minimally harbor two functional domains in addition to their catalytic core domain (CCD), the N-terminal domain (NTD) that folds into a helical bundle coordinating a Zn atom via invariant His and Cys residues and the C-terminal domain (CTD), which adopts an SH3 fold and binds nonspecifically to DNA (reviewed in ref. 1).

The integration activity of PICs extracted from infected cells persists when the complexes are treated with high salt (5, 6), and the term “intasome” was coined to describe the PIC substructure that catalyzes integration in vitro (6, 7). Integration can furthermore be reconstituted from purified components, minimally requiring IN protein and DNA mimics of the viral LTR ends (1). Integration in vitro proceeds through a series of functional IN–DNA complexes. An initial stable synaptic complex (SSC) consists of an IN tetramer and two DNA end substrates (8). 3′ Processing converts the SSC to the cleaved donor complex (CDC), which is functionally analogous to cell-extracted PICs because these invariably harbor processed termini (9, 10). Subsequent target DNA capture and integration generates the strand transfer complex (STC) (8, 11), which shares key attributes of viral PICs, including resistance to high-ionic-strength conditions. Herein, “intasome” refers to CDCs derived from virus infection or assembled from purified components.

Because of its essential nature, HIV-1 IN has been extensively studied as an antiviral drug target, and raltegravir (RAL), which preferentially inhibits DNA strand transfer activity, was licensed for the clinic in 2007 (12). Other compounds in clinical trials, typified by elvitegravir (EVG), likewise inhibit DNA strand transfer activity, defining the IN strand transfer inhibitor (INSTI) class of antiviral drugs (13). Although the 3D structure of each IN domain as well as NTD-CCD (14) and CCD-CTD (15) two-domain fragments have been solved, to date there is no structure for the active HIV-1 intasome. Because INSTIs preferentially bind to and inhibit the IN–DNA complex as compared with free IN (16), the elusive 3D structure would predictably play an important role in developing second-generation INSTIs to inhibit viruses resistant to front-line compounds. Available IN fragment structures have revealed highly variable interdomain (NTD-CCD and CCD-CTD) linker conformations (14, 15, 1719), and DNA binding is moreover required for the active site to assume its functional state. Consequently, intasome modeling based on partial HIV-1 IN structures has been less than straightforward and resulted in numerous, nonconverging results (2025). We recently reported the X-ray crystal structure of the prototype foamy virus (PFV) intasome together with chemically inhibited structures in the presence of RAL or EVG (26). Taking advantage of the relatively close relationship among retroviruses, we construct here a set of realistic models for the HIV-1 intasome in its active and INSTI-inhibited forms.

Results and Discussion

HIV-1 Intasome Construction.

The previous two-domain NTD-CCD (14) and CCD-CTD (15) crystal structures have been widely used to construct HIV-1 IN–DNA models. Some approaches relied on results of protein–DNA cross-linking (20, 27) or protein modification (24) experiments as guides for DNA placement. Others used the Tn5 transpososome (3), which until recently was the most closely related nucleoprotein complex crystal structure, as a modeling template (21, 23, 28). Because PFV IN is much more closely related to HIV-1 IN than is Tn5 transposase, we have modeled the HIV-1 intasome on the basis of the recent PFV X-ray crystal structure (26).

The crystallographic asymmetric unit contained a PFV IN dimer bound to one molecule of substrate DNA, and a pair of symmetry-related dimers formed an oblong tetramer with two synapsed DNA ends (26). Within the asymmetric unit, a near fully resolved IN monomer (residues 10–374) contacted the DNA, whereas the N-terminal extension domain (NED; see Fig. S1A for a PFV/HIV-1 IN structure-based alignment), NTD, and CTD of the second monomer were unseen in electron density maps (26). Structure determination of MnCl2-soaked crystals furthermore revealed two metal ions coordinated at the DNA-bound active site. The HIV-1 model was constructed in a step-wise fashion starting from the Mn2+-bound PFV asymmetric unit [Protein Data Bank code 3L2S] and two-domain HIV-1 IN crystal structures (14, 15). The resulting DNA-bound IN tetramer was optimized for stereochemistry and energy minimized as described in Materials and Methods.

Protein Interactions Within the HIV-1 Intasome.

Previous HIV-1 CCD crystal structures revealed globular dimers with the two sets of active-site residues separated by ≈35 Å on opposing surfaces. Because 5 bp, roughly equivalent to 18 Å in ideal B-form DNA, separates the points of attack on target DNA phosphodiesters during strand transfer, it was unlikely that this pair of active sites catalyzed the integration of both viral DNA ends (2). Inspection of the intasome structure (26) and HIV-1 model (Fig. 1) reveals the answer to this long-standing riddle. The common CCD-CCD dimer interface is maintained but occurs between the outer and inner monomer within each half of the IN tetramer (blue and green protomers in Fig. 1). It is the inner subunits of the tetramer (green and cyan in Fig. 1A) that make all visible DNA contacts and thus harbor the two functional active sites—the outer CCDs (blue and yellow) provide an apparent supportive role(s) to the overall structure. The DNA-contacting subunits moreover adopt an extended conformation, with the CTD positioning between the CCD and NTD (Fig. 1B). Because of lack of electron density for the analogous regions of PFV (26) and two-domain HIV-1 CCD-CTD (15) INs, we note that structural information for the C-terminal 18 amino acids (HIV-1 IN residues 271–288) is not available in our model.

Fig. 1.
Architecture of the HIV-1 intasome. (A) The inner subunits of the IN tetramer, comprising residues 1–270 and engaged with viral DNA, are green and cyan; outer IN CCDs (residues 56–202) are blue and yellow. The reactive and nontransferred ...

Relatively long α helices extending from the common CCD dimer to each CTD were observed in the HIV-1 IN CCD-CTD crystal structure (15). Other CCD-CTD structures, however, did not possess extended α helical linker regions. In the analogous Rous sarcoma virus construct, for example, the linkers adopted extended variable conformations (18). These results suggested considerable flexibility of CCD-CTD linker regions within IN deletion constructs. A dramatic increase in the sensitivity of Arg199 to small modifying reagents upon viral DNA binding moreover confirmed flexibility in this region of full-length, active HIV-1 IN (24). Consistent with these observations, the CCD-CTD as well as NTD-CCD linker regions adopt extended conformations within the PFV structure (26) and HIV-1 model (Fig. 1).

The CCD of each inner monomer engaged with the reactive DNA terminus at its active site interacts with the NTD of the other inner monomer in trans (green CCD and cyan NTD in upper portion of Fig. 1A). The X-ray crystal structure of the two-domain NTD-CCD fragment from maedi-visna virus IN revealed critical intermolecular interactions between NTD residues Glu11 and Glu25 and CCD residues Lys188 and Lys190, respectively (19). Salt bridges between corresponding HIV-1 positions Glu11–Lys186 and Asp25–Lys188 were notably seen in the initial NTD-CCD structure, although at that time it was unclear these formed intermolecularly (14). E11K and K186E mutations each similarly disrupted HIV-1 IN tetramerization and intasome function, with a mixture of E11K and K186E proteins supporting partial restoration of multimerization and IN activities (19). This functionally validated NTD-CCD interface is notably conserved in the PFV structure (26) and HIV-1 model (Fig. S2). Previously observed hydrogen (H)-bond contacts between the side chains of CCD residues Gln164 and Arg187 and the backbones of NTD residues Lys14 and Tyr15, respectively (14, 19), also persist in the model (Fig. S2).

IN–DNA Interactions.

Residues within each domain of the inner HIV-1 IN monomers as well as their linker regions contact the DNA in our model (Fig. S1A). Table S1 summarizes these and compares them with the contacts observed in the PFV crystal structure. Also tabulated are results from previous biochemical experiments that implicated HIV-1 residues in DNA binding. We note that whereas some experiments identified specific amino acid–base interactions (20, 2931), others did not discern precise DNA (22, 24) or IN (25, 30, 32) contact points.

For the majority of HIV-1 IN amino acids that contact DNA, the analogous PFV IN residue interacts with its substrate, although notably some PFV interactions are absent owing to different interdomain linker lengths or absence of the NED in lentiviral IN (Fig. S1A and Table S1). As observed in the PFV structure (26), the majority of HIV-1 IN residues contact the nontransferred DNA strand (Table S1). Likewise, the majority of amino acids interact with the DNA backbone, although some base contacts are observed, implicating these in defining specificity. Sequence-specific interactions include the main chain carbonyl of Gly149, which H-bonds with G4 of the nontransferred strand (Fig. S3A). This interaction is maintained in both structures because Gly218 in PFV IN, equivalent to Gly149 in HIV-1 (Fig. S1A), likewise contacts G4 of the nontransferred PFV strand (Table S1) (26). Because of the invariance of G4 among retroviral and retrotransposon LTRs and the conservation of Gly or similar small nonpolar Ala at this IN position (33, 34), we speculate this contact does not tolerate the presence of a bulky side chain and moreover may be conserved among a diverse set of intasome complexes.

The side chain of PFV IN residue Arg222 interacts with T5 and C6 of the nontransferred strand (26), whereas its HIV-1 counterpart, Ser153, also contacts bases in the nontransferred strand (G4 and C5 in this case). Results based on chimera avian sarcoma virus–HIV-1 IN activities (22) and NMR chemical shift spectra (35) previously implicated Ser153 in viral DNA binding. PFV IN residue Asn348 contacts T2 of the nontransferred strand, whereas the analogous HIV-1 residue, Glu246, interacts with nontransferred strand position T3 (Fig. S3A and Table S1). Glu246 was previously shown to interact with viral DNA, initially to A7 of the nontransferred strand (20) although subsequently in a less specific manner (24).

Because of the relatively large number of previous studies, many of the other contact residues in our model were previously implicated, either directly or indirectly, in DNA binding (15, 22, 24, 27, 29, 30, 32, 35, 36) (Table S1). Some unique contacts were nevertheless noted, and potential roles for Asn18, Arg20, and Lys211 in IN function were probed by correlating 3′ processing and DNA binding activities of site-directed mutant proteins. Arg228 (32) and Lys266 (25, 30, 32), which reside within peptides that cross-linked to DNA, were additionally targeted because specific roles in DNA binding and IN function were untested. Arg262 (36), Lys219, Arg263, and Arg269 (24), which are known to contact DNA, were included because systematic comparison of mutants in DNA binding and activity assays were lacking. Because the final energy-minimized model placed Lys219 approximately 7 Å from the DNA, this contact was excluded from Table S1. However, consistent with previous results (24), Lys219 approached within 4 Å of the DNA backbone during Molecular Dynamics simulations.

DNA binding was assessed via covalent IN–DNA complex formation after UV irradiation and polyacrylamide gel electrophoresis (Fig. 2A). Under these conditions, IN bound a U5 DNA substrate approximately 8-fold more efficiently than a base composition-matched, scrambled control sequence. IN controls included active-site mutant D64A and DNA binding defective K156E/K159E (29). As expected, D64A effectively bound DNA under conditions that precluded K156E/K159E binding (Fig. 2 A and B).

Fig. 2.
IN DNA binding and 3′ processing activities. (A) DNA binding assay. Left: Representative gel loaded with reactions conducted in the absence of IN or containing WT, D64A, or K156E/K159E (EE) IN and U5 or sequence nonspecific DNA; Right: quantified ...

The previously undescribed IN mutant proteins displayed wide spectra of 3′ processing and DNA binding activities (Fig. 2B). As anticipated from previous studies (24, 36), K219E, R262E, R263E, and R269E harbored 3′ processing and DNA binding defects (Fig. 2B). Protein–protein cross-linking in the absence of DNA was conducted to assess potential mutational affects on inherent IN multimerization (19). Because the efficiencies of K219E, R262E, R263E, and R269E tetramerization were similar to the WT (Fig. S4), impaired catalysis correlated with defective DNA binding in these cases. K211E harbored approximate 2- to 2.5-fold DNA binding and 3′ processing defects (Fig. 2B). Because this mutant failed to effectively tetramerize (Fig. S4), it is unclear whether defective multimerization or DNA binding predominantly caused the K211E catalytic defect. Next to K156E/K159E, R228E and K266E were the most impaired for catalysis and DNA binding (Fig. 2B). R228E failed to effectively tetramerize, whereas K266E multimerized similar to the WT (Fig. S4). We therefore conclude that Lys266 contributes significantly to viral DNA binding and IN catalysis. Finding that the substitution of Glu perturbed function more so than Ala at positions 262, 263, 266, and 269 is consistent with the modeled interactions of these residues with the DNA backbone (Table S1).

R20E harbored an approximate 3-fold processing defect yet bound DNA as efficiently as the WT. R20A by contrast was highly defective for both activities (Fig. 2B). Because R20A formed tetramers as efficiently as the WT (Fig. S4), we conclude its DNA binding and catalytic defects specifically correlate. Arg20 is predicted to H-bond with nontransferred strand residues A9 and G10 (Fig. S5 and Table S1). Ala at this position would effectively negate potential base contacts, whereas the Glu side chain could potentially H-bond with C10 of the transferred strand (Fig. S5). Accordingly, and in contrast to the results obtained with residues predicted to interact with the DNA backbone, R20A was significantly more defective than R20E for 3′ processing and DNA binding (Fig. 2B). The same situation moreover applied to Asn18, predicted to H-bond with nontransferred strand residue G8 (Table S1). N18D displayed an approximate 5-fold processing defect yet, like R20E, it bound DNA as the WT (Fig. 2B). Considering that N18G displayed similar 2- to 2.5-fold 3′ processing and DNA binding defects, the lack of a side-chain at this position disrupted DNA binding more so than the electronegative Asp substitution.

HIV-1 Intasome Active Site.

A theme common to transpososome (3, 4) and retroviral intasome (26) structures is the unpairing of the donor DNA duplex at the enzyme active site, which is probably important to unveil the scissile phosphodiester bond for 3′ processing and resulting 3′-OH nucleophile for subsequent DNA strand transfer. Accordingly, the terminal adenine (A17) of the reactive HIV-1 DNA strand in the model is completely separated from its T3 complement on the nontransferred strand, whereas the first three bases (A1, C2, and T3) of the nontransferred strand loop around the back of CCD α4 helix and active site, emerging to make extensive contacts with the CTD and CCD-CTD linker (Fig. S3A). As predicted from previous biochemical analyses (22, 27, 2931, 35), many of the CCD–DNA contacts involve the active site loop (residues 140–148) and α4 helix (Figs. S1A and S3A).

Consistent with a two metal ion reaction mechanism (26, 37), two Mn2+ ions (Fig. S3A, labeled A and B) are coordinated by the side chain carboxylates of the catalytic triad in the modeled active site. During DNA strand transfer, metal ion B engaged by Asp64 and Glu152 would activate the 3′-OH of the reactive DNA strand, whereas metal ion A, coordinated by Asp64 and Asp116, would destabilize the scissile phosphodiester group in target DNA. By superimposing the Cα atoms of the Tn5 transpososome and PFV intasome Asp-Asp-Glu (DDE) active site residues, we previously noted marked conservation in the positioning of donor DNA 3′-OH ends and divalent metal ions (26). Incorporating the Mos1 transpososome (4) into the structural overlay reinforces the universal mechanism at the heart of these DNA recombination machines (Fig. S3B).

INSTI Binding and Mechanism of Inhibition.

A common feature of clinical INSTIs is coplanar oxygen atoms reminiscent of an original diketo acid pharmacophore (12, 13), which has been proposed to chelate the essential divalent metal ions at the IN active site and effectively shield their availability for DNA strand transfer (38). Predecessor diketo acids moreover displayed affinity for binding to the intasome as compared with free IN, highlighting a role for the viral DNA end in forming the complete drug interaction site (16). Because of the importance of discerning the underlying mechanism of drug binding, numerous studies have modeled INSTIs at the HIV-1 IN active site, the most recent of these focusing on RAL and EVG (28, 39, 40).

PFV intasome activity, as well as cell infection by PFV, is inhibited by RAL and EVG (41). PFV intasome crystals soaked in the presence of RAL or EVG, each with MgCl2 or MnCl2, led to the determination of drug-bound structures (26). These findings afforded an optimal platform from which to model RAL and EVG binding at the HIV-1 IN active site. The resulting interactions were mediated via two conserved features of each partner: coplanar RAL and EVG oxygen atoms engage the divalent metal ions, whereas their halobenzyl moieties stack against the penultimate cytosine (C16) of the reactive DNA strand, forcing the 3′-OH of the terminal A17 nucleotide away from the active-site carboxylates and divalent metal ions (compare the drug-bound models in Fig. 3 B and C with the drug-free state in Fig. 3A). These results are consistent with the demonstrated importance of the terminal reactive adenosine for the effective binding of a related INSTI (42). We conclude that the dramatic displacement of the 3′-OH nucleophile from the active site upon drug binding explains the mechanism of inhibition.

Fig. 3.
INSTI interactions with the HIV-1 intasome. (A) View of the IN active site in the committed, drug-free state. Amino acid side chains and nucleotide bases within 4 Å of both drugs are labeled with the same coloring scheme used in Fig. S3A. (B and ...

We note that the induced fit mode of RAL and EVG binding to divalent metal ions and donor DNA bases was achieved only through modeling these salient components from the corresponding PFV crystal structures. Concordantly, automated docking of RAL or EVG onto the Mg-bound intasome (Fig. 3A) using Autodock 4.2 (43) under conditions that assumed flexibility in receptor and binding ligands failed to achieve the drug binding mode observed through manual docking. We can only assume that the rather dramatic rearrangement in adenosine positioning upon INSTI engagement lies outside the myriad of possibilities sampled by automated docking programs. Accordingly, comparing the positioning of RAL and EVG in our models with results of recent INSTI docking studies (28, 39, 40) reveals strikingly different binding modes.

Mechanism of HIV-1 Resistance to RAL and EVG.

A select number of amino acids in addition to the DDE active-site residues contact the drugs in our models: Gln146 and Arg231 interact with each compound, RAL additionally contacts Asn117, Tyr143, Asn144, and Pro145, whereas EVG is engaged by Cys65 (Fig. 3). Each drug additionally interacts with three of the four bases that constitute the invariant LTR DNA end (Fig. S1B): C16 and A17 of the reactive DNA strand and G4 of the nontransferred strand (Fig. 3 AC).

Numerous mutations recovered from ex vivo virus evolution experiments and/or patients confer resistance to RAL and/or EVG (see refs. 13 and 44 for recent overviews). Individual amino acid changes are defined as primary determinants if they confer ≥10-fold increases in drug resistance, and secondary mutations have been identified that work with primary changes to increase overall resistance. Substitutions of Gln148, Asn155, or Thr66 confer primary resistance to both drugs (45, 46). Mutations at Tyr143 confer resistance to RAL (47), whereas changes at Glu92, Phe121, Pro145, Gln146, Ser147, and Val151 are observed in EVG-resistant strains (45, 48) (Fig. S6 A and B, residues painted cyan). Sites of secondary resistance changes are shown as salmon color in Fig. S6. Clearly, only a limited number of resistance mutations arise at residues that directly contact the drugs in our models: Tyr143 in the case of RAL, and Gln146 for EVG (compare Fig. S6 A and B with Fig. 3 B and C, respectively). Our analysis therefore highlights numerous instances in which a mutant side chain that predictably will not contact the drug nevertheless confers significant resistance to it. At least in some of these cases, the mutation is likely to affect the positions of the key drug contacts—the viral DNA end and/or coordinated metal ions. Conceivably, subtle detuning of these crucial ligands will significantly reduce the affinity of the intasome–drug interaction while at the same time retain sufficient IN activity to enable integration during HIV-1 infection. Occurrence of secondary mutations might predominantly compensate for loss of enzymatic function instilled by primary resistance changes to increase the replication fitness of drug-resistant strains (49).


HIV-1 Intasome Structure and Function.

Several features of our model, which shed light on molecular details important for HIV-1 integration, are supported by experimental data. Critical intermolecular NTD-CCD interactions (19) were maintained in the final energy-minimized structure (Fig. S2), and results reported here moreover validate important roles for Arg20 and Lys266 in DNA binding and IN function. Because mutations at Ser230, Asp232, or Arg263 can confer INSTI resistance (13), the CTD was expected to approach the CCD in the functional nucleoprotein complex, and our results indeed highlight the juxtaposition of these residues in the vicinities of the RAL and EVG binding sites (Fig. S6 and Table S1).

Some previously observed specificity interactions, for examples between Gln148 and Tyr143 with A1 (30) or Gln148 with C2 (31) of the nontransferred DNA strand, were not observed here. These differences may very well be attributable to experimental design. Some cross-linking substrates (30, 31) harbored the pGTOH dinucleotide that is removed during 3′ processing and absent from our model, and its presence could conceivably position the 5′ end of the nontransferred strand closer to Tyr143 and/or Gln148 (Fig. 3 and Figs. S3A and S6). Although a precleaved DNA modified at C2 also cross-linked with Gln148 (31), in this case a used alkanethiol tether may have provided sufficient flexibility. Consistent with this interpretation, the adjacent residue in the CCD flexible loop, Ser147, interacted with the neighboring T3 position of the nontransferred strand in our model (Fig. S3A and Table S1). Lys159 was reported to interact with the adenine base of the terminal reactive nucleotide (29), and although close by, we note that this side chain in the HIV-1 model and analogous PFV IN residue Lys228 contacts the adjoining DNA backbone (Fig. S3A and Table S1).

The functional organization of IN protomers and DNA ends in the intasome model is fully consistent with results of complementation experiments that indicated that the NTD functioned in trans to the active site, whereas the CTD could be supplied in cis or in trans (50, 51). Notably, these early experiments used conditions under which only one DNA end instead of two predominantly became integrated, implying that the intasomal tetramer may be the protagonist of single-end in addition to concerted HIV-1 integration. The IN–DNA complex that mediates single-end integration is far less stable than the intasome (8), suggesting that DNA end synapsis is crucial for proper assembly and/or stability. The results of some experiments have indicated that an HIV-1 IN dimer can suffice to process a DNA end (52, 53), whereas we and others have concluded that processing is likely to occur within the context of the tetramer (8, 19, 54). If 3′ processing occurred before synapsis, the missing NTD of the top blue protomer in Fig. 1 would have to assume the position filled by the cyan NTD in the intact intasome. The cyan NTD would then have to displace this blue NTD upon tetramerization and DNA end synapsis.

INSTI Binding Mechanism and Inhibition.

The mechanism of INSTI-IN binding has been difficult to ascertain in the absence of intasome structural information. The underlying features of the drugs, coplanar oxygen atoms and aromatic rings that tend to be halogenated (12, 13, 42), had suggested that coordination of critical divalent metal ions via pharmacophore oxygens underscores the mechanism of inhibition (38), and our results indeed highlight these interactions between RAL and EVG with the intact PFV (26) and HIV-1 (Fig. 3) IN active sites. Until recently, the aromatic elements of these compounds were proposed to interact with a protein component(s) of the intasome (23). Because of the new structural information, we however speculate that an additional common and critical theme among INSTIs is an interaction between the aromatic features of the drugs and nucleotides that constitute the viral DNA end, specifically A17 and C16 of the reactive DNA strand and G4 of the nontransferred strand. Because of a strict requirement for these bases during retroviral integration (1) and as-of-yet lack of evidence for LTR mutations that confer INSTI resistance, next-generation compounds should take advantage of these invariant design features of the HIV-1 integration machine.

Materials and Methods

Model Building and Validation.

The secondary structure-based alignment of PFV and HIV-1 INs (26) was used to guide the construction of the HIV-1 intasome (see also Fig. S1A). RAL and EVG were modeled at the resulting IN active site using corresponding drug-bound PFV structures (26). In vitro integration and cross-linking assays were used to validate IN-DNA contacts; see SI Materials and Methods for salient details.

Supplementary Material

Supporting Information:


This work was supported by UK Medical Research Council Grant G0900116 (to P.C.), US National Institutes of Health Grant AI070042 (to A.E.), and the Harvard University Center for AIDS Research, a National Institutes of Health-funded program (P30AI060354) that is supported by multiple National Institutes of Health Institutes and Centers (National Institute of Allergy and Infectious Diseases, National Cancer Institute, National Institute of Mental Health, National Institute on Drug Abuse, National Institute of Child Health and Human Development, National Heart, Lung, and Blood Institute, and National Center for Complementary and Alternative Medicine).


The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1002346107/-/DCSupplemental.


1. Engelman A. Reverse transcription and integration. In: Kurth R, Bannert N, editors. Retroviruses: Molecular Biology, Genomics and Pathogenesis. Norfolk, United Kingdom: Caister Academic Press; 2010. pp. 129–159.
2. Dyda F, et al. Crystal structure of the catalytic domain of HIV-1 integrase: Similarity to other polynucleotidyl transferases. Science. 1994;266:1981–1986. [PubMed]
3. Davies DR, Goryshin IY, Reznikoff WS, Rayment I. Three-dimensional structure of the Tn5 synaptic complex transposition intermediate. Science. 2000;289:77–85. [PubMed]
4. Richardson JM, Colloms SD, Finnegan DJ, Walkinshaw MD. Molecular architecture of the Mos1 paired-end complex: The structural basis of DNA transposition in a eukaryote. Cell. 2009;138:1096–1108. [PMC free article] [PubMed]
5. Farnet CM, Bushman FD. HIV-1 cDNA integration: Requirement of HMG I(Y) protein for function of preintegration complexes in vitro. Cell. 1997;88:483–492. [PubMed]
6. Wei S-Q, Mizuuchi K, Craigie R. A large nucleoprotein assembly at the ends of the viral DNA mediates retroviral DNA integration. EMBO J. 1997;16:7511–7520. [PMC free article] [PubMed]
7. Chen H, Wei S-Q, Engelman A. Multiple integrase functions are required to form the native structure of the human immunodeficiency virus type I intasome. J Biol Chem. 1999;274:17358–17364. [PubMed]
8. Li M, Mizuuchi M, Burke TRJ, Jr, Craigie R. Retroviral DNA integration: Reaction pathway and critical intermediates. EMBO J. 2006;25:1295–1304. [PMC free article] [PubMed]
9. Miller MD, Farnet CM, Bushman FD. Human immunodeficiency virus type 1 preintegration complexes: Studies of organization and composition. J Virol. 1997;71:5382–5390. [PMC free article] [PubMed]
10. Chen H, Engelman A. Asymmetric processing of human immunodeficiency virus type 1 cDNA in vivo: Implications for functional end coupling during the chemical steps of DNA transposition. Mol Cell Biol. 2001;21:6758–6767. [PMC free article] [PubMed]
11. Pandey KK, et al. Inhibition of human immunodeficiency virus type 1 concerted integration by strand transfer inhibitors which recognize a transient structural intermediate. J Virol. 2007;81:12189–12199. [PMC free article] [PubMed]
12. Summa V, et al. Discovery of raltegravir, a potent, selective orally bioavailable HIV-integrase inhibitor for the treatment of HIV-AIDS infection. J Med Chem. 2008;51:5843–5855. [PubMed]
13. McColl DJ, Chen X. Strand transfer inhibitors of HIV-1 integrase: Bringing IN a new era of antiretroviral therapy. Antiviral Res. 2010;85:101–118. [PubMed]
14. Wang J-Y, Ling H, Yang W, Craigie R. Structure of a two-domain fragment of HIV-1 integrase: Implications for domain organization in the intact protein. EMBO J. 2001;20:7333–7343. [PMC free article] [PubMed]
15. Chen JC-H, et al. Crystal structure of the HIV-1 integrase catalytic core and C-terminal domains: A model for viral DNA binding. Proc Natl Acad Sci USA. 2000;97:8233–8238. [PMC free article] [PubMed]
16. Espeseth AS, et al. HIV-1 integrase inhibitors that compete with the target DNA substrate define a unique strand transfer conformation for integrase. Proc Natl Acad Sci USA. 2000;97:11244–11249. [PMC free article] [PubMed]
17. Chen Z, et al. X-ray structure of simian immunodeficiency virus integrase containing the core and C-terminal domain (residues 50-293)—an initial glance of the viral DNA binding platform. J Mol Biol. 2000;296:521–533. [PubMed]
18. Yang Z-N, Mueser TC, Bushman FD, Hyde CC. Crystal structure of an active two-domain derivative of Rous sarcoma virus integrase. J Mol Biol. 2000;296:535–548. [PubMed]
19. Hare S, et al. Structural basis for functional tetramerization of lentiviral integrase. PLoS Pathog. 2009;5:e1000515. [PMC free article] [PubMed]
20. Gao K, Butler SL, Bushman F. Human immunodeficiency virus type 1 integrase: Arrangement of protein domains in active cDNA complexes. EMBO J. 2001;20:3565–3576. [PMC free article] [PubMed]
21. Wielens J, Crosby IT, Chalmers DK. A three-dimensional model of the human immunodeficiency virus type 1 integration complex. J Comput Aided Mol Des. 2005;19:301–317. [PubMed]
22. Chen A, Weber IT, Harrison RW, Leis J. Identification of amino acids in HIV-1 and avian sarcoma virus integrase subsites required for specific recognition of the long terminal repeat ends. J Biol Chem. 2006;281:4173–4182. [PMC free article] [PubMed]
23. Chen X, et al. Modeling, analysis, and validation of a novel HIV integrase structure provide insights into the binding modes of potent integrase inhibitors. J Mol Biol. 2008;380:504–519. [PubMed]
24. Zhao Z, et al. Subunit-specific protein footprinting reveals significant structural rearrangements and a role for N-terminal Lys-14 of HIV-1 Integrase during viral DNA binding. J Biol Chem. 2008;283:5632–5641. [PMC free article] [PubMed]
25. Michel F, et al. Structural basis for HIV-1 DNA integration in the human genome, role of the LEDGF/P75 cofactor. EMBO J. 2009;28:980–991. [PMC free article] [PubMed]
26. Hare S, Gupta SS, Valkov E, Engelman A, Cherepanov P. Retroviral intasome assembly and inhibition of DNA strand transfer. Nature. 2010;464:232–236. [PMC free article] [PubMed]
27. Alian A, et al. Catalytically-active complex of HIV-1 integrase with a viral DNA substrate binds anti-integrase drugs. Proc Natl Acad Sci USA. 2009;106:8192–8197. [PMC free article] [PubMed]
28. Savarino A. In-Silico docking of HIV-1 integrase inhibitors reveals a novel drug type acting on an enzyme/DNA reaction intermediate. Retrovirology. 2007;4:21. [PMC free article] [PubMed]
29. Jenkins TM, Esposito D, Engelman A, Craigie R. Critical contacts between HIV-1 integrase and viral DNA identified by structure-based analysis and photo-crosslinking. EMBO J. 1997;16:6849–6859. [PMC free article] [PubMed]
30. Esposito D, Craigie R. Sequence specificity of viral end DNA binding by HIV-1 integrase reveals critical regions for protein-DNA interaction. EMBO J. 1998;17:5832–5843. [PMC free article] [PubMed]
31. Johnson AA, et al. Integration requires a specific interaction of the donor DNA terminal 5′-cytosine with glutamine 148 of the HIV-1 integrase flexible loop. J Biol Chem. 2006;281:461–467. [PubMed]
32. Heuer TS, Brown PO. Mapping features of HIV-1 integrase near selected sites on viral and target DNA molecules in an active enzyme-DNA complex by photo-cross-linking. Biochemistry. 1997;36:10655–10665. [PubMed]
33. Khan E, Mack JP, Katz RA, Kulkosky J, Skalka AM. Retroviral integrase domains: DNA binding and the recognition of LTR sequences. Nucleic Acids Res. 1991;19:851–860. [PMC free article] [PubMed]
34. Terzian C, Ferraz C, Demaille J, Bucheton A. Evolution of the Gypsy endogenous retrovirus in the Drosophila melanogaster subgroup. Mol Biol Evol. 2000;17:908–914. [PubMed]
35. Hobaika Z, et al. Specificity of LTR DNA recognition by a peptide mimicking the HIV-1 integrase alpha4 helix. Nucleic Acids Res. 2009;37:7691–7700. [PMC free article] [PubMed]
36. Lutzke RA, Plasterk RH. Structure-based mutational analysis of the C-terminal DNA-binding domain of human immunodeficiency virus type 1 integrase: Critical residues for protein oligomerization and DNA binding. J Virol. 1998;72:4841–4848. [PMC free article] [PubMed]
37. Nowotny M, Gaidamakov SA, Crouch RJ, Yang W. Crystal structures of RNase H bound to an RNA/DNA hybrid: substrate specificity and metal-dependent catalysis. Cell. 2005;121:1005–1016. [PubMed]
38. Grobler JA, et al. Diketo acid inhibitor mechanism and HIV-1 integrase: Implications for metal binding in the active site of phosphotransferase enzymes. Proc Natl Acad Sci USA. 2002;99:6661–6666. [PMC free article] [PubMed]
39. Barreca ML, Iraci N, De Luca L, Chimirri A. Induced-fit docking approach provides insight into the binding mode and mechanism of action of HIV-1 integrase inhibitors. ChemMedChem. 2009;4:1446–1456. [PubMed]
40. Perryman AL, et al. A dynamic model of HIV integrase inhibition and drug resistance. J Mol Biol. 2010;397:600–615. [PMC free article] [PubMed]
41. Valkov E, et al. Functional and structural characterization of the integrase from the prototype foamy virus. Nucleic Acids Res. 2009;37:243–255. [PMC free article] [PubMed]
42. Langley DR, et al. The terminal (catalytic) adenosine of the HIV LTR controls the kinetics of binding and dissociation of HIV integrase strand transfer inhibitors. Biochemistry. 2008;47:13481–13488. [PubMed]
43. Morris GM, et al. AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J Comput Chem. 2009;30:2785–2791. [PMC free article] [PubMed]
44. Metifiot M, Marchand C, Maddali K, Pommier Y. Resistance to integrase inhibitors. Viruses. 2010;2:1347–1366. [PMC free article] [PubMed]
45. Kobayashi M, et al. Selection of diverse and clinically relevant integrase inhibitor-resistant human immunodeficiency virus type 1 mutants. Antiviral Res. 2008;80:213–222. [PubMed]
46. Fransen S, et al. Loss of raltegravir susceptibility by human immunodeficiency virus type 1 is conferred via multiple nonoverlapping genetic pathways. J Virol. 2009;83:11440–11446. [PMC free article] [PubMed]
47. Delelis O, et al. Impact of Y143 HIV-1 integrase mutations on resistance to raltegravir in vitro and in vivo. Antimicrob Agents Chemother. 2010;54:491–501. [PMC free article] [PubMed]
48. Shimura K, et al. Broad antiretroviral activity and resistance profile of the novel human immunodeficiency virus integrase inhibitor elvitegravir (JTK-303/GS-9137) J Virol. 2008;82:764–774. [PMC free article] [PubMed]
49. Delelis O, et al. The G140S mutation in HIV integrases from raltegravir-resistant patients rescues catalytic defect due to the resistance Q148H mutation. Nucleic Acids Res. 2009;37:1193–1201. [PMC free article] [PubMed]
50. van Gent DC, Vink C, Groeneger AA, Plasterk RH. Complementation between HIV integrase proteins mutated in different domains. EMBO J. 1993;12:3261–3267. [PMC free article] [PubMed]
51. Engelman A, Bushman FD, Craigie R. Identification of discrete functional domains of HIV-1 integrase and their organization within an active multimeric complex. EMBO J. 1993;12:3269–3275. [PMC free article] [PubMed]
52. Faure A, et al. HIV-1 integrase crosslinked oligomers are active in vitro. Nucleic Acids Res. 2005;33:977–986. [PMC free article] [PubMed]
53. Guiot E, et al. Relationship between the oligomeric status of HIV-1 integrase on DNA and enzymatic activity. J Biol Chem. 2006;281:22707–22719. [PubMed]
54. Bosserman MA, O'Quinn DF, Wong I. Loop202-208 in avian sarcoma virus integrase mediates tetramer assembly and processing activity. Biochemistry. 2007;46:11231–11239. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...