Identification and Characterization of the Chromatin-binding Domains of the HIV-1 Integrase...

doi:10.1016/j.jmb.2006.04.073 J. Mol. Biol. (2006) 360, 760–773

Identification and Characterization of theChromatin-binding Domains of the HIV-1 IntegraseInteractor LEDGF/p75

Manuel Llano, Maria Vanegas, Noelle Hutchins, Daniah ThompsonSharon Delgado and Eric M. Poeschla⁎

Molecular Medicine Program,Mayo Clinic College ofMedicine, Rochester,MN 55905, USA

Abbreviations used: EBNA1, Epstantigen 1; FIV, feline immunodeficiefluorescent protein; eGFP, enhancedhepatoma-derived growth factor; Himmunodeficiency virus type 1; HMgroup A1; HRP-2, Hepatoma-derivefactor-related protein 2; IN, integrasepithelium-derived growth factor pantibody; NLS, nuclear localization s40; RNAi, RNA interference; shRNATRCB, Triton-resistant chromatin biE-mail address of the correspondi

[email protected]

0022-2836/$ - see front matter © 2006 E

Depletion of the transcriptional co-activator LEDGF/p75 by RNA interfer-ence alters the genome-wide pattern of HIV-1 integration, reducingintegration into active genes, reducing integration into LEDGF/p75-regulated genes, and increasing integration into G+C-rich sequences.LEDGF/p75 is also able to act as a molecular tether linking HIV-1 integraseprotein to chromatin, a phenomenon likely to underlie the integration sitedistribution effects. The LEDGF/p75 integrase-binding domain has beenestablished but the domain or domains responsible for the chromatin-binding component of tethering are unknown. Here, we identify andcharacterize these domains. Complementary methods were used to assesscondensed and uncondensed chromatin, and to determine the stringency ofchromatin binding. Immuno-localization analyses revealed that an N-terminal PWWP domain and its β-barrel substructure are needed forbinding to metaphase chromatin. However, the PWWP domain isinsufficient to transfer metaphase chromatin binding to green fluorescentprotein, which requires addition of a downstream charged region (CR1).Biochemical analysis showed that full-length LEDGF/p75 resists Triton X-100 extraction from chromatin. To transfer Triton-resistant chromatinbinding to green fluorescent protein, PWWP-CR1 is necessary but notsufficient. Further inclusion of a tandem pair of AT-hooks in combinationwith at least one of two identified downstream charged regions (CR2 orCR3) is needed. Deletion of just the PWWP or the AT-hook domain fromfull-length LEDGF/p75 reduced Triton-resistant chromatin binding, whiledeletion of both elements abolished it, underscoring their dominant andcooperative role. The results establish a molecular mechanism for LEDGF/p75-mediated tethering of HIV-1 integrase to chromatin.

© 2006 Elsevier Ltd. All rights reserved.

*Corresponding author
Keywords: LEDGF/p75; HIV; integrase; chromatin; molecular model
ein-Barr virus nuclearncy virus; GFP, greenGFP; HDGF,IV-1, humanGA1, high mobilityd growthe; LEDGF/p75, lens75; mAb, monoclonalignal; NP-40, nonidet, short hairpin RNA;nding.ng author:

lsevier Ltd. All rights reserve

Introduction

Lens epithelium-derived growth factor/p75(LEDGF/p75) is a transcriptional coactivator inthe hepatoma-derived growth factor (HDGF)family.1,2 A specialized role in regulating stressresponse gene transcription has been proposed.3–6Differential splicing of the LEDGF mRNA resultsin two proteins, p52 and p75, which have differentC termini.1,2 LEDGF/p75, which is 530 amino acidresidues in length, co-localizes with chromatin.7–10The protein has recently attracted special attentionbecause of its interaction with the integrase (IN)protein of human immunodeficiency virus type 1

d.

mailto:[email protected]

http://dx.doi.org/10.1016/j.jmb.2006.04.073

761Chromatin-binding Determinants in LEDGF/p75

(HIV-1),8–18 which was identified initially by theco-immunoprecipitation experiments of Debyserand colleagues.8 The interaction is mediated byan IN-binding domain (IBD) located in the C-terminal region of the protein (residues 340–417;see Figure 1(a) for diagram).13,14 Because of thisinteraction, intracellular trafficking of IN is deter-mined by LEDGF/p75, and a model has emergedin which LEDGF/p75 can function as a moleculartether, linking IN to chromatin.9,10 Stable RNAiknockdown of LEDGF/p75 untethers IN, produc-ing a striking phenotypic shift from an exclusivelynuclear location to an exclusively cytoplasmicone.10 IN thus appears to lack an autonomousnuclear localization signal (NLS); nuclear localiza-tion depends instead on the LEDGF/p75 NLS.11,14Moreover, subsequent investigation has demon-strated that LEDGF/p75 protects HIV-1 IN pro-tein from proteasomal degradation.12 IN proteinwas revealed to be otherwise highly unstable incells.12 LEDGF/p75 has also been detected in theHIV-1 pre-integration complex, which is theuncoated, post-entry form that is the precursor tointegration into the host genome.10 Each of thesephenomena has been corroborated with the INprotein of a second lentivirus, FIV, and eachappears to be lentivirus-specific, since none is

Figure 1 (legend

exhibited by the IN proteins of murine or avianretroviruses.10,12,15One other HDGF family member, HRP-2, contains

the IBD.13,14 This protein also binds to HIV-1 IN andcan mediate IN nuclear import in an IBD-dependentmanner.13,14 In distinct contrast to LEDGF/p75,however, HRP-2 is not a chromatin-bound proteinand does not tether IN to chromatin.14 The fiveremaining members of the HDGF family (LEDGF/p52, HDGF, HRP-1, HRP-3, HRP-4) lack the IBD andhence do not interact with IN.Potential roles of LEDGF/p75 and/or HRP-2 in

the viral life-cycle are currently a focus of intensiveinvestigation. It was proposed that LEDGF/p75may influence trafficking of the pre-integrationcomplex to particular genomic sites10,13,14 and mayenhance its stability.12 The first of these hypotheseswas confirmed recently.18 Depletion of LEDGF/p75produced three main effects: reduced HIV-1 inte-gration into active transcription units; reducedintegration in LEDGF/p75-regulated genes; andincreased integration into G+C-rich sequences.18 Inthe same cell lines used in these analyses, IN-to-chromatin tethering is fully disrupted.10,12,14Since IN catalyzes viral integration and is obli-

gately associated with the pre-integration complexuntil the required DNA cleaving and joining

on next page)

Figure 1. Domain transfer experiments. (a) Schematic representation of LEDGF/p75, showing the two previouslyestablished elements (NLS and IBD, which are depicted in green), as well as elements mapped functionally in the presentwork (other colors). Amino acid numberings from N-terminus: PWWP, 1-93; CR1, 94-142; NLS, 146-156; AT-hooks, 178-198; CR2, 208-266; CR3, 267-325; IBD, 340-417. CR1, CR2 and CR3 contain 39, 43 and 55% charged residues, respectively.(b–d), Confocal fluorescence microscopic images of L cells expressing various LEDGF/p75-eGFP fusion proteins. (b)LEDGF/p75-eGFP, LEDGF/p751-325-eGFP and LEDGF/p75326-530-eGFP; (c) LEDGF/p751-176-eGFP and LEDGF/p751-207-GFP; (d) LEDGF/p751-93-eGFP, LEDGF/p751-142-eGFP, and Gallus gallus LEDGF/p751-182-eGFP. Data in photomicro-graphs here and in subsequent figures represent consistently observed phenotypes recorded from at least 50 mitoses orfrom at least 100 interphase cells. Unless otherwise stated, 95% or more of cells displayed the phenotype of the shownimages.

762 Chromatin-binding Determinants in LEDGF/p75

reactions occur, the tethering phenomenon is hy-pothesized to underlie the role of LEDGF/p75 indetermining HIV-1 integration site distributions.18 Apressing issue in this regard is to determine howLEDGF/p75 interacts with chromatin. While thedomain mediating the LEDGF/p75-to-IN compo-nent of lentiviral IN tethering is now clearly

established to be the IBD,13,14 the domain ordomains mediating the other half of the tetheringmechanism, i.e. LEDGF/p75-to-chromatin binding,are unknown. Recurring functional modules havebeen identified within chromatin-binding proteins.For example, bromodomains bind acetylated lysineresidues in the N-terminal tails of histone proteins


and chromodomains bind to methylated lysineresidues.19 Sequence analysis using protein data-bases suggests the region comprising the first 93amino acid residues (Figure 1(a)) has homology to‘‘PWWP domains’’, so-named for a relativelyconserved proline-tryptophan-tryptophan-prolinemotif found near the N terminus.20 The role ofsuch domains is unknown in the majority ofproteins, but participation in mediating protein–protein and/or DNA–protein interactions has beenproposed.21–23 In Dnmt3 methlytransferases, thePWWP domain may facilitate chromatin interactionduring both mitosis and interphase.22 In addition,the PWWP domain mediates metaphase chromo-some association of the mismatch repair proteinMSH6 and the histone methyltransferase NSD2/WHSC.22 Four PWWP domain structures have beenreported.19,23–25 The core structure is an N-terminalfive stranded anti-parallel β-barrel (in LEDGF/p75,residues 1–63) with several additional C-terminal α-helices (LEDGF/p75 residues 64–93).24,25 The fourexisting structures are quite similar in the β-barrelregion, but vary significantly in the adjacent helices.The other potential LEDGF/p75 chromatin-bind-

ing element suggested by computer-assisted analy-ses is a pair of tandemly situated AT-hook sequences(178PKRGRPAATEVKIPKPRGRPK198). AT-hooksare predominantly basic motifs with a core se-quence, RGRP, that is highly conserved across phylafrom bacteria to mammals. AT-hooks are necessaryand sufficient to bind to the minor groove of A+T-rich DNA, are found exclusively in DNA-interactingnuclear proteins, and are thought to co-regulatetranscription by modifying the architecture of DNAto enhance the accessibility of promoters to tran-scription factors.26–28 From one to 15 AT-hooks canbe found within a single protein. Like LEDGF/p75,Epstein-Barr virus (EBV) nuclear antigen 1 (EBNA-1) has two AT-hooks. These are functionally critical,as they are required for EBNA-1 binding to cellularchromosomes; the resulting tethering of latent EBVgenomes to chromosomes facilitates viral replicationand partitioning.29 Cooperative activity may occurbetween AT-hooks and other DNA-binding activi-ties in proteins and affinity is augmented by DNAcontacts flanking the core sequence, particularlypolar residues.26,27 Nevertheless, AT-hooks may bemain effectors of chromatin binding. In the cellularprotein high-mobility group A1a (HMGA1a), forexample, mutation of its three AT-hook motifs fullydisrupts chromatin binding.30Here, we have systematically identified and

characterized the chromatin-binding regions inLEDGF/p75 by using immuno-localization andfluorescent protein fusion studies as well as com-plementary biochemically-based chromatin-bindingassays that assess stringency of binding. Weemployed RNAi-generated cells that lack immuno-blot-detectable endogenous LEDGF/p7510 in orderto permit unambiguous analyses without the con-founding presence of endogenous protein, which ishighly expressed and chromatin-bound. Mutantsinclude multiple fusions of candidate LEDGF/p75

domains to green fluorescent protein (GFP). In orderto assure a complete analysis, specific mutants ofunfused LEDGF/p75 were also used, with proper-ties scored using antibody-based detection. Anequally important aspect of the complementarymorphological and biochemical approaches used istheir application to the physiologically relevantproperty of binding to cellular chromatin ratherthan binding to DNA alone.We assign chromatin binding to an ensemble of

modular domains. The PWWP domain plays acentral and required role. However, it is notsufficient, and additional domains in both the Nterminus and the C terminus of the protein,including but not limited to the AT-hook domain,are either required for binding or augment bindingaffinity. By establishing the molecular mechanismfor the LEDGF/p75–chromatin interaction, theseresults complete a model for LEDGF/p75-mediatedIN tethering.

Results

Identification of transferable LEDGF/p75domains responsible for chromosome binding

LEDGF/p75 is a 530 amino acid residue protein.Two domains have been identified, the NLS and theIBD (Figure 1(a)). In the present study, mappingstudies were initiated by fusing segments ofLEDGF/p75 to enhanced GFP (eGFP) and deter-mining subcellular distributions in cells havingstable RNA interference (RNAi) that eliminatesdetectable endogenous LEDGF/p75 (L cells, seeMaterials and Methods). First, N and the C-terminalportions (residues 1–325 and 326–530, respectively)were fused to the eGFP N terminus. As assessed byexamining condensed metaphase or anaphase chro-matin in mitotic cells, residues 1–530 and 1–325targeted eGFP to chromatin in all cells analyzed, butthe C-terminal segment did not (Figure 1(b)). Thus,the N-terminal 325 amino acid residues appeared toharbor the major chromosome-binding apparatus.Two different protein modules suggested by in silicoanalyses to reside in this segment, the PWWPdomain and the pair of AT-hooks, were next targetedfor analysis. eGFP fusions having residues 1–176,which contain the PWWP domain and the NLS, andresidues 1–207, which additionally contain the AT-hook motifs, were evaluated (Figure 1(c)). In allGFP-positive cells, both bound to chromosomeswith a pattern similar to the 1–325 fusion, suggest-ing that the AT-hooks are not required for chromo-some binding.To map the minimal region capable of transferring

chromatin binding, we analyzed in the same fashionthe β-barrel subdomain (amino acid residues 1–63),the full PWWP domain (residues 1–93), and asegment with the PWWP domain plus a down-stream patch of relatively polar amino acids(residues 1–142; 38.7% of ressidues 94–142 are


charged). Only the 1–142 segment bound eGFP tochromosomes (LEDGF/p751-142-eGFP, Figure 1(d)).Over 90% of the LEDGF/p751-142-eGFP-positivecells had the phenotype of the top cell in the middlepanel of Figure 1(d), with occasional remainingcells showing some fluorescence elsewhere, as inthe bottom cell. Note that LEDGF/p751-142-eGFPlacks the NLS (Figure 1(a)). These results suggestedthat the PWWP domain and the adjoining 49residue charged region, which we designate CR1,may form a main chromosome-binding domain inLEDGF/p75.The PWWP domain is well conserved within the

seven-member HDGF family, including the INinteractor HRP-2. However, the latter proteinlacks CR1. HRP-2 also does not bind to chromo-somes in assays identical with those in Figure 1(b)–(d), and it does not tether IN to chromatin.14 Thisdifferent phenotype supports the relevance of CR1for chromosome association. CR1 is very wellconserved between mammalian LEDGF/p75 pro-teins (94% identity). Homology drops significantlyin a comparison of human LEDGF/p75 with eitherthe chicken of Xenopus protein, where this regionhas 42% identity with human LEDGF/p75. Theregion is also larger in the chicken protein, with 89versus 49 residues. Nevertheless, there is conserva-tion of overall charge density (38.7% and 39.3%charged residues in human and chicken proteins,respectively). Therefore, we fused the N-terminal182 residues of chicken LEDGF/p75 to eGFP (i.e.,the PWWP domain plus 89 additional residues).Unlike LEDGF/p751-142-eGFP, chicken LEDGF/p751-182-eGFP was not tethered (Figure 1(d); 71%of GFP-positive cells had the phenotype shown; in29%, some trace of chromosome binding could bedetected but the large majority of the protein wasnon-bound and distributed diffusely). Thus, chick-en LEDGF/p751-182-eGFP resembles human HRP-214 in its pan-cellular distribution during mitosisand its diffuse nuclear distribution in interphase.These results suggest that specific determinantswithin human LEDGF/p75 CR1, rather thanpolarity per se, determine its contribution tochromosome binding.

The PWWP domain, with an intact β-barrel, isrequired for chromatin binding of LEDGF/p75

Having analyzed these domains as individualsegments in test fusions, we then performed thecomplementary analysis of asking whether thePWWP domain is required for chromosome associ-ation in the context of full-length LEDGF/p75.LEDGF/p75PWWP-, a truncation mutant lackingresidues 1–93, was expressed stably in L cells (Figure2(a)). It had a diffuse nuclear pattern in interphasebut never associated with chromosomes during celldivision (Figure 2(a)-ii). Furthermore, a LEDGF/p75 protein deleted only in the β-barrel sub-domain (residues 1–63) had the same phenotype(Figure 2(a)-iii), indicating that this sub-domain isessential for chromosome binding.

The PWWP domain is necessary for chromatincapture of NLS-mutant LEDGF/p75

LEDGF/p75 has a single NLS, at residues 146–15611,14 (see Figure 1(a)). LEDGF/p75.NLS-, whichis alanine-substituted at four critical NLS residues(149RKRK152), fails to undergo nuclear import and isfound in the cytoplasm of L cells after transienttransfection.14 However, subsequent passage oftransfected cells through mitosis results in chroma-tin association, and in stable transformants, theNLS-deleted protein is actually constitutively nu-clear and completely chromatin-bound during celldivision.14 Thus, stable nuclear residence resultsfrom capture of LEDGF/p75.NLS- by chromatinduring mingling of nuclear and cytosolic contents atcell division and retention of the protein in thischromatin-bound state in daughter cells.14 Here, weintroduced the same NLS mutation into LEDGF/p75PWWP- and generated similar stable L celltransformants (Figure 2(a)-iv). Expression of pre-dicted proteins was verified by Western blotting(Figure 2(b)). In contrast to LEDGF/p75.NLS-, thedouble mutant was always exclusively cytoplasmic(Figure 2(a)-iv; this panel can be compared toFigures 3(a)-iv and 5 of Vanegas et al.14). It wasalso always excluded from metaphase chromatinduring mitosis (data not shown). Thus, chromatincapture of LEDGF/p75 is PWWP domain-depen-dent, and the NLS in the protein acts to mediatenuclear import but not chromatin binding.

Biochemical analyses of chromatin binding

To characterize chromatin binding of LEDGF/p75further, we used an alternative, biochemical assaythat enables us to assess stringency of binding incells and to measure binding by scoring co-localization with condensed metaphase or anaphasechromosomes in mitotic cells, and to asses chroma-tin binding in the uncondensed DNA of the majorityinterphase population of cells. Nuclear proteins canbe attached to chromatin or other nuclear structuresor exist in a soluble form within the nucleoplasm.Chromatin-binding assays assessing complete cellpopulations make use of the resistance of chromatin-bound proteins to extraction with non-ionic deter-gents such as TritonX-100.22,31,32 Disruption ofTriton-resistant chromatin binding (TRCB) requiresfurther treatment of the Triton-insoluble fractionwith DNase and salt; the latter measures do notsolubilize other non-chromatin-bound, Triton-insol-uble proteins, such as those comprising the cyto-skeleton or the nuclear matrix.To apply and validate this approach in the

present study, we began by analyzing L cells co-transfected with LEDGF/p75 and HIV-1 IN. Con-sistent with its previously determined co-localiza-tion with chromatin in L cells,10 LEDGF/p75 wasstrongly chromatin-bound and remained exclusive-ly with the Triton-insoluble fraction (fraction P1,Figure 3(a)-i; see Materials and Methods forfractionation details). LEDGF/p75 was released

Figure 2. PWWP domain: requirement for chromosome binding and chromatin capture. A LEDGF/p75-specificmAb was used in (a) and (b). (a) Confocal immunofluorescence microscopic images of L cells stably expressingdifferent LEDGF/p75 mutants. (i) LEDGF/p75; (ii) LEDGF/p75PWWP-; (iii) LEDGF/p75βbarrel-; (iv) NLS-mutantLEDGF/p75PWWP-. (b) Immunoblotting of mutants analyzed in (a).


completely by DNase and salt treatment (fractionS2, Figure 3(a)-i). HIV-1 IN co-fractionated withLEDGF/p75, consistent with the known cellular co-localization of these proteins and the ability ofLEDGF/p75 to tether IN to chromatin (Figure 3(a)-i, bottom panel). In addition to LEDGF/p75 andHIV-1 IN, further validation was carried out byanalyzing the sub-cellular distribution of a panel ofcontrol proteins with established sub-cellular loca-tions (Figure 3(a)-ii). eGFP, which does not interactwith chromatin, has a pan-cellular distribution; it isdistributed solely to the Triton-soluble fraction (S1).A second control, the cytosolic protein α-tubulin,was also found exclusively in S1. In contrast,proliferating cell nuclear antigen (PCNA; auxiliaryprotein of DNA polymerase delta), was distributedboth in the Triton-soluble (S1) and Triton-resistant(P1) fractions, and was extracted efficiently withDNase and salt, which is in agreement withpreviously established data showing that approx-imately 40% of PCNA is bound tightly to chroma-

tin at the sites of DNA replication.33 In clearcontrast to PCNA, histone H1 was recovered onlyin the Triton-resistant fractions. Histone H1 wasextracted more efficiently with DNase and salt (S2)than with the RIPA buffer (P1 and T). However,treatment with DNase and salt did not extract p62,a component of the nuclear pore complex that,although Triton-resistant (found in P1), was recov-ered completely in the P2 rather than S2 fraction aspreviously reported.34 Finally, we determined thefractionation properties of the NLS-mutantLEDGF/p75PWWP- in the stable L cell transformantsdescribed above. In agreement with the microscopydata (Figure 2(a)-iv), this protein was foundexclusively in S1 (Figure 3(a)-iii).

Contributions of AT-hooks and chargeddomains

Using this fully validated assay, we carried outdomain-specific analyses. LEDGF/p75-eGFP fusions

Figure 3. Chromatin binding assays. L cells were used to study TRCB of LEDGF/p75 and LEDGF/p75-eGFP fusions.(a) Fractionation patterns of LEDGF/p75,NLS-mutant LEDGF/p75PWWP- andHIV-1 IN in a chromatin-binding assay. Cellfractions are defined in Materials and Methods and summarized at left of panel. (i) LEDGF/p75 and HIV-1 IN were co-transfected into L cells and the same samples were immunoblotted for LEDGF/p75 (top panel in i) and for IN (bottompanel in i). Both proteins co-segregate into chromatin-bound fractions (P1, S2), which is consistentwith their co-localizationwith chromatin as previously determined by immunofluorescence in L cells.10 LEDGF/p75 displays identical fractionationin the absence of co-expressed IN (data not shown). (ii) Fractionation patterns of control proteins. eGFP, α-tubulin,proliferating cell nuclear antigen (PCNA), histoneH1, andp62were evaluated. (iii) fractionation of LEDGF/p75PWWP-, NLS-.Cells are the same stable cell line shown in the photomicrograph of Figure 2(a)-iv. Near-confluent cultures with 10% or lessmitotic cells were lysed in i–iii. T, total cell lysate before subcellular fractionation.Neg, cell lysate from cells transfectedwithan empty plasmid vector. (b) Fractionation patterns of LEDGF-eGFP fusion proteins.


were used first, to permit direct comparison ofbiochemical data to intracellular localization pheno-types; this also allowed scoring of the differentfragments in the same immunoblot. Consistent withthe immunofluorescence data of Figure 1(b), LEDGF/p751-325-eGFP distributes completely to P1 and S2(Figure 3(b)), i.e., identical with full-length LEDGF/p75 (Figure 3(a)), indicating that residues 1–325 aresufficient for Triton-resistance.Further experiments were then performed to

investigate the individual contributions of protein

modules within LEDGF/p751-325-eGFP that couldmediate TRCB: the PWWP domain, the two AT-hooks, and three charged regions (CR1 plus twoadditional regions that we designate CR2 and CR3;see Figure 1(a) for a diagram). LEDGF/p751-176-eGFP (PWWP domain, CR1 and the NLS),LEDGF/p751-207-eGFP (includes AT-hooks aswell), LEDGF/p751-266-eGFP (further includesCR2) were tested. All of these proteins have theLEDGF/p75 NLS and localized exclusively to thenucleus in interphase cells as expected (Figure 1,


and data not shown). Despite binding to meta-phase chromatin (Figure 1(c)), LEDGF/p751-176-eGFP was Triton-soluble, indicating that chromatinassociation is weak in the absence of the AT-hook-CR2-CR3 region (Figure 3(b)). However, inclusionof the AT-hooks, as in LEDGF/p751-207-eGFP,imparted a strong shift towards a predominantlychromatin-bound phenotype (note the prominenceof the S2 fraction, and the minor S1 band).Inclusion of CR2, in LEDGF/p751-266-eGFP, led torecovery only in P1 and S2 (Figure 3(b)). We alsoestablished that the AT-hook domain of LEDGF/p75 does not possess autonomous chromatinbinding activity, since fusion of residues 177–207to eGFP did not result in any detectable chromatinassociation (data not shown).These biochemical results thus significantly

advance the conclusions obtainable by immuno-fluorescence and GFP-fusion morphologies, byshowing that the AT-hooks and additional polardomains augment chromatin binding. Full TRCBcommensurate with that of the wild-type pro-tein requires the PWWP domain, the AT-hooksand CR2, with the additive effects of theseelements disclosed by combining the differential-ly stringent immunofluorescence and biochemicalassays.To further understand the relative contributions

of these elements to TRCB and control forpotential compensatory effects of the C-terminalportion, we studied several additional deletionmutants of our eGFP fusions. Deletion of justthe PWWP domain (residues 1–93) from LEDGF/p751-325-eGFP completely abrogated Triton-resis-tance (Figure 4), supporting its central role.However, selective deletion of just the pair of AT-hooks from the 1–325 or 1–266 fusions reducedTriton resistance only partially (Figure 4), support-ing a secondary role for this element. In contrast,selective removal of either CR2 (Figure 4) or CR3(i.e. LEDGF/p751-266-eGFP, Figure 3(b)) preservedTriton-resistance, supporting a more minor role forthese polarized regions. Nevertheless, combined

interpretation of Figure 3(b) and Figure 4 showsthat either CR2 or CR3 must be present tomaintain full TRCB.

Role of C-terminal residues

In isolation, the LEDGF/p75 C-terminal domain(amino acid residues 326–530) cannot mediatechromatin interaction (Figure 1(b)). This result isconsistent with the modular tethering model thathas emerged from current work focused on LEDGF/p75 as an IN-to-chromatin tethering protein, inwhich the C-terminal IBD binds IN, while the N-terminal regions bind chromatin. However, suchexperiments do not test whether the 326–530fragment can facilitate in any way the chromatin-binding activity of the more N-terminal domains.Since charged residues in the three CR domainswere implicated by our data, we considered thepossibility that additional charged regions in the Cterminus may modulate binding. We thereforeexamined the biochemical properties of a LEDGF/p75-eGFP fusion deleted in residues 177–325, whichin effect inserts the 326–530 C terminus intoLEDGF/p751-176-eGFP. In interesting contrast toLEDGF/p751-176-eGFP (Figure 3(b)), LEDGF/p75Δ177-325-eGFP distributed predominantly in P1and S2 with a comparatively less prominent S1 band(Figure 5(a)). While this result suggests that regionsin the C-terminal domain of LEDGF/p75 mightbolster chromatin binding, LEDGF/p75Δ177-325-eGFP distributed to the bound fractions to a lesserextent than LEDGF/p751-325-eGFP, which maysuggest a less significant capacity than that of theAT-hooks and CR2-CR3. Considering charge alone,the 205 amino acid residues of the C-terminalsegment (326–530) resemble CR2-CR3 in containing37.5% polar residues, which can be grouped intofour relatively discrete patches, with most chargedsurface outside the IBD (Figure 5(b)). When weselectively deleted the IBD (340–417) from LEDGF/p75Δ177-325-eGFP, the protein became completelyTriton-soluble (Figure 5(a)). In agreement with this

Figure 4. Chromatin bindingassays with eGFP fusions havingdiscrete domain deletions. See Fig-ure 3(a) legend and diagram formethodological details. Proteinswere detected with an anti-eGFPmAb.

Figure 5. Analysis of C-termi-nus. (a) Chromatin binding assays.Proteins were detected with an anti-eGFP mAb. (b) Surface accessibilityplot. I-IV: clusters of charged resi-dues. The IBD (residues 340–417) isindicated.


observation, deletion of residues 178–418 fromfull-length LEDGF/p75 produced an identical phe-notype (Figure 5(a)). Taken together, these findingssupport an inference that the IBD region, but notresidues 418–530, may participate in augmentingchromatin interaction. Consistent with the pheno-types of the other mutants, Triton-resistance wasalso abolished when the PWWP domain was deletedfrom LEDGF/p75Δ177-325-eGFP (Figure 5(a)). Sincethe normal cellular function of the C terminus ofLEDGF/p75 is unknown (e.g. whether an en-dogenous protein interacts with the IBD), it couldbe that protein–protein interactions with one ormore other chromatin-interacting proteins are atplay in the effects observed with C-terminal domainmanipulations.

Biochemical analyses of full-length LEDGF/p75proteins

To further establish relevance of these function-ally mapped modules within the full-length proteinand to rule out artifact due to the GFP moiety, weperformed TRCB assays with LEDGF/p75 proteinsbearing targeted deletions. Somewhat surprisingly,

selective deletion of one or both AT-hooks fromLEDGF/p75 (LEDGF/p75AThook-) yielded no effecton Triton-resistance (Figure 6) or on metaphasechromatin binding in immunofluorescence assays(data not shown). Similarly, glycine substitution ofthe critical arginine residue following the centralglycine in each AT-hook,30 alone or in combination,had no effect in either assay (data not shown).Deletion of the IBD (LEDGF/p75IBD-) also did notimpair LEDGF/p75 TRCB (Figure 6). In contrast tothe complete Triton solubility of PWWP domain-deleted LEDGF/p751-325-eGFP (Figure 4), we foundthat LEDGF/p75PWWP- was partially Triton-resis-tant (Figure 6), a result that supports our earlierinference that determinants downstream of residue325 contribute to chromatin binding. However,deletion of the PWWP domain and the AT-hooksfrom full-length LEDGF/p75 abolished Tritonresistance (Figure 6) and metaphase chromatinassociation (data not shown), confirming thatthese two protein modules are the main effectorsof chromatin binding. In contrast, deleting the IBDfrom the PWWP-minus protein had no suchadditive effect (Figure 6). Taken together, theseresults support a dominant role for the PWWP

Figure 6. Triton-resistant chromatin binding ofLEDGF/p75 deletion mutants (PWWP domain, AT-hookdomain, IBD). Proteins were detected with a LEDGF/p75-spcific mAb.


domain with an important accessory role for theAT-hook domain.

Discussion

LEDGF/p75 has attracted interest because of itschromatin-tethering and proteasome-shieldingeffects on HIV integrase and its role in HIVintegration targeting. Chromatin binding is centralto IN tethering and undoubtedly to the still rathervaguely characterized normal role of LEDGF/p75 inmodulating cellular gene expression, but its molec-ular determinants have been unknown. In thepresent work, we identify and characterize compre-hensively the distinct protein modules involved. Theresults establish a molecular mechanism for thetethering of HIV-1 IN to chromatin and provide abasis for further investigation of the normal role ofthis protein in the cell. The most basic feature of themodel is the protein’s modular structure. PrimarilyN-terminal regions mediate binding to chromatin ina cooperative manner, while the C terminus attachesto IN via the IBD. Functional N-terminal domainsidentified here are the PWWP domain, AT-hooks,CR1, CR2 and CR3.The N-terminal PWWP domain plays a dominant

role. This domain was necessary for Triton-resis-tance in bulk cell populations having primarilyuncondensed chromatin, and for binding to con-densed chromosomes during mitosis. The PWWPdomain contains two subdomains. The N-terminalβ-barrel, formed by the first 63 residues of LEDGF/p75, was found to be an essential subdomain. Aprecedent is the β-barrel subdomain of the Dnmt3bmethyltransferase, which is critical for chromatinbinding and the DNA methylation activity of thisprotein. A single missense mutation (S282P) in the

human Dnmt3b PWWP domain β-barrel causescomplete loss of chromatin targeting capacity, whichresults clinically in immunodeficiency, centromericheterochromatin instability, and facial anomalies(ICF) syndrome.22Although necessary, the LEDGF/p75 PWWP

domain was not sufficient to transfer metaphasechromatin binding to a test protein, eGFP. Rather,the CR1 domain was also required. While CR1 iswell-conserved in mammalian LEDGF/p75 ortho-logs (94% identity), the avian and amphibianLEDGF/p75 proteins diverge considerably in thisregion (89 versus 49 residues, less than 50% identitywith the mammalian proteins). These differencessuggest a different functional role for CR1 in theseother classes of Chordata. In support of this concept,we found that human LEDGF/p75 PWWP-CR1 wassufficient for chromatin tethering of eGFP in theimmunofluorescence assay, while the chickenPWWP-CR1 was not. The insufficiency of thehuman PWWP domain alone is also fully consistentwith our previous observation that HRP-2, whichcontains the entire PWWP domain but not CR1, isnot chromatin-bound and does not tether IN.14Chromatin capture of LEDGF/p75.NLS-, an infor-mative phenomenon recently identified in L cells,14was also found in the present studies to be PWWPdomain-dependent. In cycling cells, a majority ofnuclear LEDGF/p75.NLS-appears to be what isretained post-mitosis in daughter cells via thePWWP-dependent tethering pathway. We speculatethat chromatin retention of pre-formed LEDGF/p75through completion of cell division may help toguarantee continuous positioning of the protein infunctionally relevant chromatin sites. The NLS in theprotein functions to ensure adequate nuclearLEDGF/p75 in non-dividing cells but it is not asignificant determinant of chromatin association.Note that, in contrast to the partial TRCB ofLEDGF/p75PWWP-, LEDGF/p75PWWP-,NLS- displaysno TRCB (compare Figure 3(a)-iii with the top panelof Figure 6). As LEDGF/p75PWWP-,NLS- is constitu-tively cytoplasmic, while LEDGF/p75.NLS- is con-stitutively chromatin-bound even in stable cell-lines(Figure 2(a, i–iv) and Vanegas et al.14), this differencein the TRCB assay implies that the role of the NLShere is to provide nuclear residence rather thanchromatin binding, and that regions outside thePWWP domain that mediate partial TRCB cannotsuffice to mediate stable chromatin capture duringmetaphase.The chromatin-binding assays we performed,

which use detergent extraction to assess the strin-gency of binding in predominantly interphase cells,provide key data not obtainable by using immuno-fluorescence or GFP fusions, as they revealed thatother protein modules participate in LEDGF/p75chromatin binding. The AT-hook domain plays animportant but auxiliary role. The cooperativeinteraction between this domain and the PWWPdomain is underscored by the analyses of combina-torial deletion mutants. Deleting the PWWP domainalone (LEDGF/p75PWWP-) only partially reduced


TRCB (Figure 6), despite fully disrupting metaphasechromatin association (Figure 2(a)), suggesting thatthere is greater binding of this mutant to interphasechromatin than condensed mitotic chromatin. De-leting one or both AT-hooks from full-lengthLEDGF/p75 had no effect, either in the immunoflu-orescence assay or the TRCB assay, and transfer ofthis domain to a test protein conferred no chromatinbinding. However, deleting both the PWWP domainand the AT-hooks abolished TRCB completely(Figure 6). In this respect, LEDGF/p75 appears toresemble some other proteins, where AT-hookmotifs have been found to associate functionallywith larger DNA-binding domains, providing addi-tional contacts in the minor grove to augment theactivity of the dominant chromatin-bindingdomain.27 This pattern is not always the case. Forexample, in contrast to their secondary role inLEDGF/p75, mutation of the critical arginineresidues in all three of the AT-hook motifs presentin HMGA1a blocks chromatin binding duringinterphase and mitosis.30 Similarly, two AT-hookmotifs located at the amino terminus of EBNA1 arerequired for its association with the cellularchromosome.29 The participation of CR2 and CR3are also consistent with other examples whereflanking charged regions influence AT-hook domainfunction.35,36Since AT-hooks interact with A/T residues in

the minor groove of DNA molecules, the chroma-tin interaction role identified for the AT-hookmotifs in LEDGF/p75 could be related to therecent finding that RNAi-depletion of LEDGF/p75shifts the genome-wide pattern of HIV-1 integra-tion to G+C-rich sequences.18 Since IN-to-chroma-tin tethering is fully disrupted by the sameRNAi,10 it is likely that the tethering phenomenonis related mechanistically to effects on integrationsite distributions.When LEDGF/p75 binding was attenuated by

deletion of the central group of auxiliary elements(AT-hooks, CR2, CR3) from eGFP fusion proteins, anaugmenting effect of the IBD was detectable. Whilethis result must be interpreted with reserve becauseof the artificial juxtaposition of normally non-contiguous domains, and because the IBD displayedno autonomous chromatin-binding activity, it ispossible that role of the IBD in cells might involvemodulation of the affinity of LEDGF/p75 forchromatin, perhaps mediated by interaction of theIBD with an endogenous protein.In addition to influencing viral integration site

distributions,18 potential related roles of LEDGF/p75 could be to serve as a cofactor for IN catalysis orto enhance pre-integration complex or IN proteinstability during steps preceding integration. L cellshave no detectable LEDGF/p75 protein but doexpress a small fraction of the normal amounts ofLEDGF/p75 mRNA.10 If a small residuum ofLEDGF/p75 protein exists in these cells, it wouldbe inconsequential to the trafficking and subcellularlocalization phenotypes observed when subclonedHIV-1 and FIV IN proteins are expressed, while in

the case of the much less abundant pre-integrationcomplex in a nascently infected cell, the stoichiom-etry might permit functionally important interactionof IN with LEDGF/p75. If this is the case,eradication of residual LEDGF/p75 could help tobetter define the virological role of the protein,which could extend beyond the modulation ofintegration site preferences.

Materials and Methods

Cell culture and transfections

L cells express no immunoblot-detectable endogenousLEDGF/p75, while S cells express normal levels of theprotein.10 The current abbreviation is adopted here;12,14 Lcells were originally designated si1340 cells and S cellswere designated siScram cells.10 Both lines were derivedfrom 293T human embryonic kidney cells by stableexpression of pol III-promoted short hairpin RNAs(shRNA). S cells express a scrambled sequence shRNA,while L cells express an shRNA that targets a 21 ntsequence in the p75-specific C-terminal portion of theLEDGF/p75 mRNA; this shRNA is highly effective.10Cells were grown in Dulbecco’s modified Eagle’s medium(Gibco BRL) with 10% (v/v) fetal calf serum, penicillin,and streptomycin. Transfections were performed by thecalcium phosphate co-precipitation method with a total of1 μg of DNA per well in a six-well plate or 1μg of DNA perchamber in a two-chamber LabTek II glass chamber slide(Nalge Nunc Naperville, IL). Briefly, cells were transfectedat 24 h after being plated in 2 ml of medium at 0.45×106cells/well or 1 ml of medium at 0.8×105 cells/chamber.The medium was replaced 14–16 h later. Cells wereharvested or used for indirect immunofluorescence 40–48 h after transfection. Stable cell-lines expressingLEDGF/p75 mutants were derived from L cells bytransfection of linearized plasmid DNA followed byselection in 3 μg/ml of puromycin.

Plasmid construction

Plasmids were constructed by standard recombinantcloning techniques and all changes were verified by DNAsequencing. All plasmids used in the present workincorporate the relevant segment of a previously con-structed LEDGF/p75 cDNA version (LEDGF/p75siMut)in which we rendered the mRNA resistant to the stableanti-p75 RNAi in L cells by introducing seven synony-mous mutations at the shRNA-binding site.10 Generationof the IBD deletion (residues 340–417) in LEDGF/p75 hasbeen described.14 For eGFP fusions, amplicons wereinserted into peGFP-N1 (Clontech) between BglII orXbaI, and SalI. PCR was also performed as needed,using LEDGF/p75AThook- or LEDGF/p75Δ340-417(also referred to as LEDGF/p75IBD-)14 templates andfragments. LEDGF/p75PWWP- was generated with pri-mers 5′ TATAAGATCTATGTCAAGTCAACAGG and 5′-TATATTCTAGACCTAGTTATCTAGTGTAG followed byinsertion between BglII and XbaI in pEFIRES-P.37 LEDGF/p75AThook- has a deletion in residues 177–201 and wasgenerated by overlap extension PCR with primers5′-TAATCTAAAACAGCCCTGTCCTTCAGAGAG5′-GGACAGGGCTGTTTTAGATTAACAGATGCTG-TTGC


5′- GTTTATGGGAGATAGATAACAATCC5′-GATCACGGAATCTCCTTCACCAACC

The PCR product was then introduced into LEDGF/p75 asan NsiI fragment. LEDGF/p75 1-142-eGFP was generatedwith primers5′-TATAAGATCTATGACTCGCGATTT5′-TATAGTCGACCGGAGTAGTTATGTCLEDGF/p75 1-62-eGFP was generated with primers5′-TATAAGATCTATGACTCGCGATTT5′-TATAGTCGACGAGTAAGGAAATATATCCLEDGF/p751-176-eGFP was generated with primers5′-TATAAGATCTATGACTCGCGATTTCAAACC5′-TATAGTCGACCACTTTTAGATTAAC

Other LEDGF/p75 fragment amplicons were subsequent-ly inserted in-frame with both genes, between BamHI andSalI. Using this strategy, residues 266–325 of LEDGF/p75were amplified with5′-TATAGTCGACGAAAACAGGGGTTAC5′-TATAGGATCCTCAGTTTCCATTTGTTC

to delete the AT-hooks and CR3 from LEDGF/p751-325-eGFP. For LEDGF/p75Δ177-325-egfp, primers5′-TATAGTCGACATGCAGCAGAATAAAG5′-TATAGGATCCCGGGGCCCGTTATCTAGTG

were used to amplify the 326–530 segment from LEDGF/p75 or LEDGF/p75Δ340-417. For the Δ177-418 construct,5′- TATAGTCGACGATGTTGTATAACAAG5′-TATAGGATCCCGGGGCCCGTTATCTAGTG

were used and the amplicon was inserted in-frame withLEDGF1-176. To delete the PWWP domain from LEDGF/p75Δ177-325-eGFP (IBD+ and −), the 1–176 segment wassubstituted with residues 93–176, which were amplifiedwith5′-TATAAGATCTATGTCAAGTCAACAGG5′-TATAGTCGACCACTTTTAGATTAAC

To delete the PWWP domain from LEDGF/p751-325-eGFPresidues 93–325 were amplified with5′ TATAAGATCTATGTCAAGTCAACAGG5′-TATAGTCGACTCAGTTTCCATTTGTTC

LEDGF1-207-eGFP was constructed with primers5′-TATAAGATCTATGACTCGCGATTTCAAACC5′-TATAGTCGACTCTGAAGGACAGGGCTG

For LEDGF/p751-266-eGFP and LEDGF/p751-325-eGFP(both AT-hook + or −), primers5′-TATAAGATCTATGACTCGCGATTTCAAACC5′-TATAGTCGACTTAGCTAAATTTTTCCTor 5′-TATAAGATCTATGACTCGCGATTTCAAACC5′ TATAGTCGACTCAGTTTCCATTTGTTC

were used respectively. Construct LEDGF/p751-207-eGFPwas used to generate other fusion proteins by insertingSalI and Bam HI sites between the LEDGF/p751-207 andeGFP frames. To delete CR2 from LEDGF/p751-325-eGFPresidues 266–325 were amplified using primers5′-TATAGTCGACGAAAACAGGGGTTAC5′-TATAGGATCCTCAGTTTCCATTTGTTC

To generate the fusion of chicken LEDGF1-182 to the eGFPN terminus, PCR was performed using a chicken LEDGF/p75 cDNA template,13 a kind gift fromA. Engelman (DanaFarber Cancer Institute), with5′-TATAAGATCTATGAGCCGAGATTTTAAACC5′-TATAGTCGACGGAATAGATGCATCATTTG

followed by insertion between BglII and SalI in peGFP-N1.

Chromatin binding assay

The method of He et al.38 was used, with modifica-tions described by Kannouche et al.32 Briefly, 36 h aftertransfection cells were lysed for 15 min on ice in coldCSK I buffer (10 mM Pipes, (pH 6.8), 100 mM NaCl,

1 mM EDTA, 300 mM sucrose, 1 mM MgCl2, 1 mMDTT) supplemented with 0.5% (v/v) Triton X-100,protease inhibitors (Roche Complete Mini) and 1 mMphenylmethylsulfonyl fluoride. One-tenth of the lysate(total fraction, T) was mixed with RIPA buffer (150 mMTris–HCl, (pH 8.0), 150 mM NaCl, 0.5% DOC, 0.1%(w/v) SDS, 1% (v/v) NP-40). The remaining cell lysatewas divided into two equal portions, which werecentrifuged at 500g at 4 °C for 3 min. The supernatants(S1 fraction), which contains Triton-soluble proteins,were further analyzed. One of the pellets, which containchromatin-bound, nuclear matrix-bound, and insolubleproteins, was resuspended in RIPA buffer (the P1fraction). The second was resuspended in CSK II buffer(10 mM Pipes (pH 6.8), 50 mM NaCl, 300 mM sucrose,6 mM MgCl2, 1 mM DTT), treated with DNase for30 min followed by extraction with 250 mM NH2SO4 for10 min at 25 °C. The sample treated with DNase and saltwas then centrifuged at 1200g for 6 min at 4 °C and thesupernatant (S2 fraction, containing DNase-releasedchromatin-associated proteins) and pellet (P2, containinginsoluble, cytoskeletal, and nuclear matrix proteins)were collected. P2 was also resuspended in RIPA buffer.All fractions were analyzed by immunoblotting.

Immunoblotting

Samples of protein (20 μg) were separated by SDS-PAGE and transferred overnight to, polyvinylidenedifluoride (PVDF) membranes. Membranes wereblocked with 10% (w/v) milk powder in Tris-bufferedsaline (TBS) and probed with primary monoclonalantibodies (mAb) diluted in 5% milk powder in TBSplus 0.05% (v/v) Tween 20 for 2 h at room temperature.Membranes were washed in TBS with 0.1% Tween 20and incubated with a horseradish peroxidase-conjugat-ed rabbit anti-mouse IgG diluted 1/5000 in 5% milkwith 0.05% Tween. Bound antibodies were detectedwith the ECL reagent (Amersham Pharmacia Biotech).Primary mAbs used were anti-Myc (Covance clone9E10), diluted 1/500 to detect Myc epitope-tagged HIVIN, anti-GFP (Clontech clone JL-8), diluted 1/5000 todetect eGFP fusion proteins and a monoclonal LEDGF/p75-specific antibody (BD Transduction Laboratories),diluted 1/500.

Laser scanning confocal microscopy

For GFP fusions, L cells were fixed 40 h aftertransfection with 4% (v/v)paraformaldehyde, stainedwith 4′,6-diamidino-2-phenylindole (DAPI), and imagedby laser scanning confocal fluorescence microscopy asdescribed.10,14 Indirect immunofluorescence was per-formed with a primary monoclonal antibody to LEDGF/p75 as described.10,14 Briefly, cells grown in LabTek IIchamber slides were fixed with 4% formaldehyde in PBSfor 10 min at 37 °C, washed with PBS, and thenpermeabilized with ice-cold methanol for 2 min. Fixedcells were blocked in 10% fetal calf serum, 20 mMammonium chloride, and PBS for 30 min at roomtemperature, then incubated with the appropriate anti-bodies, followed by Alexa-488-conjugated goat anti-mouse antibody (Molecular Probes, Eugene, Oregon)and imaged by laser scanning confocal fluorescencemicroscopy. Nuclear DNA was stained with DAPI(Molecular Probes). At least 50 mitotic cells and 100interphase cells were analyzed to determine sub-cellulardistributions.


Acknowledgements

We thank P. Fuentes-Prior for helpful suggestions,A. Engelman and P. Cherepanov for a chickenLEDGF/p75 cDNA, S. Hobbs for pEFIRESp, E.Kelly for assistance with plasmid construction, andM. Peretz for general technical assistance.

References

1. Ge, H., Si, Y. & Roeder, R. G. (1998). Isolation ofcDNAs encoding novel transcription coactivators p52and p75 reveals an alternate regulatory mechanism oftranscriptional activation. EMBO J. 17, 6723–6729.

2. Ge, H., Si, Y. & Wolffe, A. P. (1998). A noveltranscriptional coactivator, p52, functionally interactswith the essential splicing factor ASF/SF2.Mol. Cell 2,751–759.

3. Sharma, P., Singh, D. P., Fatma, N., Chylack, L. T., Jr &Shinohara, T. (2000). Activation of LEDGF gene bythermal-and oxidative-stresses. Biochem. Biophys. Res.Commun. 276, 1320–1324.

4. Nakamura, M., Singh, D. P., Kubo, E., Chylack, L. T., Jr& Shinohara, T. (2000). LEDGF: survival of embryonicchick retinal photoreceptor cells. Invest. Ophthalmol.Vis. Sci. 41, 1168–2275.

5. Shinohara, T., Singh, D. P. & Fatma, N. (2002). LEDGF,a survival factor, activates stress-related genes. Prog.Retin. Eye Res. 21, 341–358.

6. Ganapathy, V., Daniels, T. & Casiano, C. A. (2003).LEDGF/p75: a novel nuclear autoantigen at thecrossroads of cell survival and apoptosis. Autoimmun.Rev. 2, 290–297.

7. Nishizawa, Y., Usukura, J., Singh, D. P., Chylack, L. T.,Jr & Shinohara, T. (2001). Spatial and temporaldynamics of two alternatively spliced regulatoryfactors, lens epithelium-derived growth factor(ledgf/p75) and p52, in the nucleus. Cell Tissue Res.305, 107–114.

8. Cherepanov, P., Maertens, G., Proost, P., Devreese, B.,Van Beeumen, J., Engelborghs, Y., De Clercq, E. &Debyser, Z. (2003). HIV-1 integrase forms stabletetramers and associates with LEDGF/p75 protein inhuman cells. J. Biol. Chem. 278, 372–381.

9. Maertens, G., Cherepanov, P., Pluymers, W., Busschots,K., De Clercq, E., Debyser, Z. & Engelborghs, Y. (2003).LEDGF/p75 is essential for nuclear and chromosomaltargeting of HIV-1 integrase in human cells. J. Biol.Chem. 278, 33528–33539.

10. Llano, M., Vanegas, M., Fregoso, O., Saenz, D. T.,Chung, S., Peretz, M. & Poeschla, E. M. (2004).LEDGF/p75 determines cellular trafficking of di-verse lentiviral but not murine oncoretroviral inte-grase proteins and is a component of functionallentiviral pre-integration complexes. J. Virol. 78,9524–9537.

11. Maertens, G., Cherepanov, P., Debyser, Z., Engel-borghs, Y. & Engelman, A. (2004). Identification andcharacterization of a functional nuclear localizationsignal in the HIV-1 integrase (IN) interactor LEDGF/p75. J. Biol. Chem. 279, 33421–33429.

12. Llano, M., Delgado, S., Vanegas, M. & Poeschla, E. M.(2004). LEDGF/p75 prevents proteasomal degrada-tion of HIV-1 integrase. J. Biol. Chem. 279, 55570–55577.

13. Cherepanov, P., Devroe, E., Silver, P. A. & Engelman,A. (2004). Identification of an evolutionarily con-served domain in human lens epithelium-derived

growth factor/transcriptional co-activator p75(LEDGF/p75) that binds HIV-1 integrase. J. Biol.Chem. 279, 48883–48892.

14. Vanegas, M., Llano, M., Delgado, S., Thompson, D.,Peretz, M. & Poeschla, E. (2005). Identification of theLEDGF/p75 HIV-1 integrase-interaction domain andNLS reveals NLS-independent chromatin tethering.J. Cell Sci. 118, 1733–1743.

15. Busschots, K., Vercammen, J., Emiliani, S., Benarous,R., Engelborghs, Y., Christ, F. & Debyser, Z. (2005). Theinteraction of LEDGF/p75 with integrase is lentivirus-specific and promotes DNA binding. J. Biol. Chem. 280,17841–17847.

16. Emiliani, S., Mousnier, A., Busschots, K., Maroun, M.,Van Maele, B., Tempe, D. et al. (2005). Integrasemutants defective for interaction with LEDGF/p75are impaired in chromosome tethering and HIV-1replication. J. Biol. Chem. 280, 25517–25523.

17. Cherepanov, P., Sun, Z. Y., Rahman, S., Maertens, G.,Wagner, G. & Engelman, A. (2005). Solution structureof the HIV-1 integrase-binding domain in LEDGF/p75. Nature Struct. Mol. Biol. 12, 526–532.

18. Ciuffi, A., Llano, M., Poeschla, E., Hoffman, C.,Leipzig, J., Marshall, H., Shinn, P., Ecker, J. &Bushman, J. (2005). A role for LEDGF/p75 in targetingHIV DNA integration. Nat. Med. 11, 1287–1289.

19. Slater, L. M., Allen, M. D. & Bycroft, M. (2003).Structural variation in PWWP domains. J. Mo. Biol.330, 571–576.

20. Stec, I., Wright, T. J., van Ommen, G. J., de Boer, P. A.,van Haeringen, A., Moorman, A. F. et al. (1998).WHSC1, a 90 kb SET domain-containing gene,expressed in early development and homologous toa Drosophila dysmorphy gene maps in the Wolf-Hirschhorn syndrome critical region and is fused toIgH in t(4;14) multiple myeloma. Hum. Mol. Genet. 7,1071–1082.

21. Stec, I., Nagl, S. B., van Ommen, G. J. & den Dunnen,J. T. (2000). The PWWP domain: a potential protein-protein interaction domain in nuclear proteinsinfluencing differentiation? FEBS Letters, 473, 1–5.

22. Ge, Y. Z., Pu, M. T., Gowher, H., Wu, H. P., Ding, J. P.,Jeltsch, A. & Xu, G. L. (2004). Chromatin targeting ofde novo DNA methyltransferases by the PWWPdomain. J. Biol. Chem. 279, 25447–25454.

23. Qiu, C., Sawada, K., Zhang, X. & Cheng, X. (2002). ThePWWP domain of mammalian DNA methyltransfer-ase Dnmt3b defines a new family of DNA-bindingfolds. Naure Struct. Biol. 9, 217–224.

24. Sue, S. C., Chen, J. Y., Lee, S. C., Wu, W. G. &Huang, T. H. (2004). Solution structure and heparininteraction of human hepatoma-derived growthfactor. J. Mol. Biol. 343, 1365–1377.

25. Nameki, N., Tochio, N., Koshiba, S., Inoue, M.,Yabuki, T., Aoki, M. et al. (2005). Solution structureof the PWWP domain of the hepatoma-derivedgrowth factor family. Protein Sci. 14, 756–764.

26. Reeves, R. & Nissen, M. S. (1990). The A.T-DNCorre-sponding author. A-binding domain of mammalianhigh mobility group I chromosomal proteins. A novelpeptide motif for recognizing DNA structure. J. Biol.Chem. 265, 8573–8582.

27. Aravind, L. & Landsman, D. (1998). AT-hook motifsidentified in a wide variety of DNA-binding proteins.Nucl. Acids Res. 26, 4413–4421.

28. Siddiqa, A., Sims-Mourtada, J. C., Guzman-Rojas, L.,Rangel, R., Guret, C., Madrid-Marina, V. et al. (2001).Regulation of CD40 and CD40 ligand by the AT-hooktranscription factor AKNA. Nature, 410, 383–387.


29. Sears, J., Ujihara, M., Wong, S., Ott, C., Middeldorp, J.& Aiyar, A. (2004). The amino terminus of Epstein-Barr Virus (EBV) nuclear antigen 1 contains AT hooksthat facilitate the replication and partitioning of latentEBV genomes by tethering them to cellular chromo-somes. J. Virol. 78, 11487–11505.

30. Harrer, M., Luhrs, H., Bustin, M., Scheer, U. & Hock,R. (2004). Dynamic interaction of HMGA1a proteinswith chromatin. J. Cell Sci. 117, 3459–3471.

31. Mendez, J. & Stillman, B. (2000). Chromatin associa-tion of human origin recognition complex, cdc6, andminichromosome maintenance proteins during thecell cycle: assembly of prereplication complexes in latemitosis. Mol. Cell Biol. 20, 8602–8612.

32. Kannouche, P. L., Wing, J. & Lehmann, A. R. (2004).Interaction of human DNA polymerase eta withmonoubiquitinated PCNA: a possible mechanism forthe polymerase switch in response to DNA damage.Mol. Cell, 14, 491–500.

33. Bravo, R. & Macdonald-Bravo, H. (1987). Existence oftwo populations of cyclin/proliferating cell nuclear

antigen during the cell cycle: association with DNAreplication sites. J. Cell Biol. 105, 1549–1554.

34. Davis, L. I. & Blobel, G. (1986). Identification andcharacterization of a nuclear pore complex protein.Cell, 45, 699–709.

35. Aulner, N., Monod, C., Mandicourt, G., Jullien, D.,Cuvier, O., Sall, A. et al. (2002). The AT-hook proteinD1 is essential for Drosophila melanogaster develop-ment and is implicated in position-effect variegation.Mol. Cell Biol. 22, 1218–1232.

36. Reeves, R. (2001). Molecular biology of HMGAproteins: hubs of nuclear function. Gene, 277, 63–81.

37. Hobbs, S., Jitrapakdee, S. & Wallace, J. C. (1998).Development of a bicistronic vector driven by thehuman polypeptide chain elongation factor 1alphapromoter for creation of stable mammalian cell linesthat express very high levels of recombinant proteins.Biochem. Biophys. Res. Commun. 252, 368–372.

38. He, D. C., Nickerson, J. A. & Penman, S. (1990). Corefilaments of the nuclear matrix. J. Cell Biol. 110,569–580.

Edited by J. O. Thomas

(Received 3 January 2006; received in revised form 19 April 2006; accepted 30 April 2006)Available online 17 May 2006

Identification and Characterization of the Chromatin-binding Domains of the HIV-1 Integrase...

Documents

Transcript of Identification and Characterization of the Chromatin-binding Domains of the HIV-1 Integrase...