LEDGF/p75 Is Essential for Nuclear and Chromosomal Targeting of HIV-1 Integrase in Human Cells*

We have reported that human immunodeficiency virus type 1 (HIV-1) integrase (IN) forms a specific nuclear complex with human lens epithelium-derived growth factor/transcription co-activator p75 (LEDGF/p75) protein. We now studied the IN-LEDGF/p75 interaction and nuclear import of IN in living cells using fusions of IN and LEDGF/p75 with enhanced green fluorescent protein and far-red fluorescent protein HcRed1. We show that both the N-terminal zinc binding domain and the central core domains of IN are involved in the interaction with LEDGF/p75. Both domains are essential for nuclear localization of IN as well as for the association of IN with condensed chromosomes during mitosis. However, upon overexpression of LEDGF/p75, the core domain fragment of IN was recruited to the nuclei and mitotic chromosomes with a distribution pattern characteristic of the full-length protein, indicating that it harbors the main determinant for interaction with LEDGF/p75. Although the C-terminal domain of IN was dispensable for nuclear/chromosomal localization, a fusion of the C-terminal IN fragment with enhanced green fluorescent protein was found exclusively in the nucleus, with a diffuse nuclear/nucleolar distribution, suggesting that the C-terminal domain may also play a role in the nuclear import of IN. In contrast to LEDGF/p75, its alternative splice variant, p52, did not interact with HIV-1 IN in vitro and in living cells. Finally, RNA interference-mediated knock-down of endogenous LEDGF/p75 expression abolished nuclear/chromosomal localization of IN. We conclude, therefore, that the interaction with LEDGF/p75 accounts for the karyophilic properties and chromosomal targeting of HIV-1 IN.

The human immunodeficiency virus (HIV), 1 the causative agent of AIDS, belongs to the Lentiviridae genus of retrovi-ruses. The early steps of HIV replication include reverse transcription of the diploid viral RNA genome into a doublestranded linear DNA replica and integration into a host cell chromosome. Reverse transcription takes place in the cytoplasm of the infected cell and results in the formation of a compact and stable preintegration complex (PIC), containing the viral reverse-transcribed genome and a number of virionderived and cellular proteins. HIV and other lentiviruses are able to productively infect non-dividing, terminally differentiated cells, a feature distinguishing them from oncoretroviruses, which require cell division for productive infection (1)(2)(3)(4). Previous work has characterized the nuclear import of HIV-1 PICs as an active, energy-dependent process (5), yet its mechanism has remained a puzzle. The determinants of HIV nuclear import that have been suggested so far are: the nuclear localization signals (NLSs) of the viral matrix (MA), Vpr, and integrase (IN) proteins, and the central DNA flap (for reviews see Refs. 6 -8). The latter has been reported to be essential for nuclear import of HIV PICs and viral replication (9). Although the effect of the central DNA flap appears to be viral strain-and host cell-dependent (10,11), its insertion in HIV-derived lentiviral vectors clearly augments transduction efficiency (12) and nuclear import (13). The karyophilic properties of MA and its role in HIV nuclear import (14) are unclear (15)(16)(17). Vpr is also not strictly required for HIV replication and DNA integration in non-dividing cells (15,18,19). It seems plausible that the DNA flap, MA, and Vpr, albeit redundant, exert additive and/or inter-dependent effects on HIV nuclear import. IN, on the other hand, is an attractive candidate for the role of the PIC import factor. (i) It is essential for the viral replication and spread of infection in primary cells and most T-cell lines (20); (ii) it is present in the PIC; (iii) its karyophilic properties have been demonstrated by many groups (for references see below). Unfortunately, mutations in IN have pleiotropic effects on viral replication, including alterations in viral particle morphology, defects in reverse transcription and integration (21), confounding a detailed genetic analysis of its functions.
HIV-1 IN is a 32-kDa protein, initially produced as part of the Gag-Pol precursor polyprotein and released after cleavage by the viral protease during maturation of the virion. IN is responsible for the catalysis of the insertion of the viral DNA into the host cell chromosome (for reviews see Refs. [22][23][24]. Like all retroviral INs, HIV-1 IN is composed of three domains. The N-terminal domain harbors an HHCC-type zinc binding domain and has been implicated in the multimerization of the protein (25). The core domain contains the catalytic site and possesses structural elements necessary for sequence-specific recognition of the viral long terminal repeat (26). The arginine/ lysine-rich C-terminal domain of IN also contributes to the multimerization of the protein (27) and is thought to be involved in DNA binding. In addition, binding to DNA has been shown to induce oligomerization of HIV-1 IN in vitro (28). When expressed in or microinjected into human cells, HIV-1 IN accumulates in the nuclei (17, 29 -33). During mitosis, IN stably associates with condensed chromosomes (30). Recent studies with digitonin-permeabilized cells have shown that nuclear import of HIV-1 IN can occur in the absence of cytosolic extracts, requires ATP hydrolysis, and is GTPase Ran-independent (34). Similar observations have been made on the nuclear import of the NLS receptor, importin ␣ (35). Therefore, the virus either does not rely on the classical importin-and Randependent nuclear import mechanism or is able to take advantage of an alternative pathway.
Recently, we have shown that in human cells HIV-1 IN forms a specific nuclear complex with lens epithelium-derived growth factor/transcription co-activator p75 (LEDGF/p75) (36). Recombinant LEDGF/p75 protein effectively promoted HIV-1 IN strand transfer activity in vitro. Of note, LEDGF/p75 was found to be up-regulated in HIV-infected cells (37). All these observations suggested that LEDGF/p75 could play a role in retroviral DNA integration. Although its precise cellular function remains elusive, several reports have implicated LEDGF/ p75 in gene expression and cellular stress response (38 -40). LEDGF/p75 has been reported to be a DNA-binding protein, with specificity for stress response DNA elements (40). According to the abundant mouse and human LEDGF/p75 mRNA-derived expressed sequence tags in GenBank TM , LEDGF/p75 is expressed at all stages of development in a variety of organs and tissues, including skin, bone, thymus, brain, mammary gland, testis, and embryonic and hematopoietic stem cells. A second protein product, p52, can be produced from the same gene as LEDGF/p75 as a result of alternative splicing of the pre-mRNA (38,41). At least in vitro, p52 was found to be a more general and stronger transcriptional coactivator than LEDGF/p75. In addition, the in vitro interaction of p52 with the ASF/SF2 splicing factor suggests that p52 might play a role in both transcription and splicing (42).
In  (30) or its fragments were PCR-amplified using Pfu DNA polymerase (Stratagene) and an appropriate pair of primers (see below). Each primer contained either a recognition site for XhoI (sense primers) or for HindIII (antisense primers). In-frame stop codons were included in the antisense primers. The primers used were: C s , 5Ј-GGGGGCTC-GAGCAGACTGCAGAAGCAGATCACC; C as , 5Ј-GGGGAAGCTTGGAC-TTAGTCCTC; N s , 5Ј-GGGGGCTCGAGCAGATTCCTGGACGGCATT-GAC; N as , 5Ј-GGGGAAGCTTACATAGCCTCGCC; Core s , 5Ј-GGGGGC-TCGAGCAGACACGGGCAGGTTGATTGC; Core as , 5Ј-GGGGAAGCTT-ACTCTTTGGTCTGG. The PCR fragments were subcloned between the XhoI and HindIII restriction sites of the pEGFP-C2 vector (Clontech pEGFP-IN s /Nt (using the primers N s and N as ), pEGFP-IN s /core (using Core s and Core as ), pEGFP-IN s /Ct (using C s and C as ) coding the Nterminal, the central, and the C-terminal IN domain, respectively. The H12N mutant of the synthetic gene was engineered by PCR with the primers H12N (5Ј-GACGGCATTGACAAGGCTCAGGAGGAGA-ACGAGAAGTACCACTC) and T3 (5Ј-AATTAACCCTCACTAAAGGG) using Pwo DNA polymerase (Roche Applied Science) and pIN s (30) as the template. The second PCR was performed on the resulting amplicon with the primers N s and T3; the final PCR product was digested with XhoI and HindIII and subcloned into pEGFP-C2 to obtain pEGFP-IN s (H12N). To obtain pHcRed1-IN s , for expression of HIV-1 IN fused to the C terminus of the far-red fluorescent protein, HcRed1 (43), the XcmI/EcoRI restriction fragment of pIN s (30) was cloned between the XhoI and EcoRI sites of pHcRed1-C1 (Clontech) after treatment of the XcmI and the XhoI termini of the DNA fragments with T4 DNA polymerase.
To generate HcRed1-labeled LEDGF/p75, the BamHI/EcoRI fragment of pCP6H75 (36), spanning the LEDGF/p75 open reading frame, was subcloned between the BglII and EcoRI sites of pHcRed1-C1. The coding sequences of HcRed1 and LEDGF/p75 were placed into frame via BspEI restriction and mung bean nuclease (Invitrogen, Groningen, The Netherlands) digestion followed by re-ligation to obtain the plasmid pHcRed1-p75. To generate an analogous fusion between p52 and HcRed1, the XhoI/EcoRI fragment of pHcRed1-p75 was replaced by the XhoI/EcoRI fragment of pKB-Nat52 (see below), resulting in pHcRed1-p52. The plasmid pEGFP-p75 expressing EGFP-tagged human LEDGF/ p75 protein was obtained by inserting the BamHI/EcoRI fragment of pCP6H75 (36) between the BglII and EcoRI sites of pEGFP-C2. Prior to ligation the BamHI and BglII termini of the DNA fragments were filled in, using T4 DNA polymerase. All plasmid constructs used in this work were verified via sequence analysis to confirm absence of mutations.
To create pGM-INH12N-6H, for bacterial expression of the C-terminally His 6 -tagged IN H12N mutant, a DNA fragment containing the HIV-1 IN open reading frame with the mutation was engineered in two consecutive PCR reactions. First, pKB-IN6H was used as the template and a PCR was performed with the forward primer 5Ј-GATAAGGC-CCAAGAAGAAAATGAGAAATATCACAG and the INSalI primer (see above). The resulting amplicon was used as template in the second reaction with the primer 5Ј-ATATACATATGTTTTTAGATGGAATA-GATAAGGCCCAAG and the INSalI primer. The final PCR fragment was digested with NdeI and SalI and cloned into the pET-20b(ϩ) vector.
The constructs pCP-Nat75 and pKB-Nat52 were used for bacterial expression of non-tagged LEDGF/p75 and p52 proteins, respectively. A DNA fragment, containing the LEDGF/p75 open reading frame, was amplified using pCP-6H75 (36) as a template and 5Ј-TGACTCGCGAT-TTCAAACC and 5Ј-CCGCGAATTCTAGTTATCTAGTGTAGAATCC-TTC as the primers. To obtain the pCP-Nat75 plasmid, this PCR fragment was digested with EcoRI and inserted between the NdeI and EcoRI sites of the pRSETB vector (Invitrogen). Prior to ligation, the NdeI terminus of the vector DNA was filled in using T4 DNA polymerase. To produce pKB-Nat52, a DNA fragment containing the p52 open reading frame was constructed by PCR using the primers 5Ј-TGACTCGCGATTTCAAACC and 5Ј-GGCGAATTCTACTGTAGATTA-CATGTTGTTGGTGCTCAGTTTCCATTTGTTCC. The resulting fragment was digested with EcoRI and cloned between the EcoRI and NdeI sites of pRSETB.
Cell Culture and Transfections-HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with GlutaMAX TM , 10% fetal calf serum, and 0.02 mg/ml gentamicin (Invitrogen) at 37°C and 5% CO 2 in a humidified atmosphere. Cells were seeded the day before transfection in 8-well LabTek chambered coverglass cuvettes (VWR International, Leuven, Belgium). Transfection of HeLa cells was performed at ϳ80% confluency using Lipofectamine 2000 reagent (Invitrogen) with 0.36 g of plasmid, following the instructions from the manufacturer. Double transfections were accomplished following the same protocol except that transfections were performed with 0.18 g of each plasmid.
For transient LEDGF/p75 knock-down experiments, HeLa cells were transfected with small interfering RNA (siRNA) synthetic duplexes using the Gene Silencer transfection reagent (Gene Therapy Systems) or co-transfected with siRNA plus the pEGFP-IN s plasmid using Lipo-fectamine 2000 according to established protocols (45). Fetal calf serum was added to the medium at 5 h after transfection. The effect of the RNA interference was studied 48 -60 h after transfection.
Western Blotting and Indirect Immunofluorescence-For Western blot detection, protein extracts separated in 11% SDS-PAGE gels were electroblotted onto polyvinylidene difluoride membranes (Bio-Rad, Nazareth, Belgium). The polyclonal anti-GFP antibody was purchased from Invitrogen; the mouse anti-LEDGF p75/p52 antibody was from BD Biosciences (Erembodegem, Belgium). The polyclonal rabbit anti-HIV-1 IN antibody has been described previously (30). Secondary horseradish peroxidase-or alkaline phosphatase-conjugated goat anti-mouse and goat anti-rabbit antibodies were from Dako Diagnostics (Leuven, Belgium). Detection was carried out using ECLϩ chemiluminescent horseradish peroxidase substrate (Amersham Biosciences) or with 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium chromogenic substrate for alkaline phosphatase. The broad range prestained protein marker mix (New England Biolabs, Hertfordshire, United Kingdom) was used for estimation of the molecular weights.
Indirect immunofluorescent detection of endogenous LEDGF/p75 was performed as previously described (36). Cells grown in LabTek II glass chamber slides (VWR International) were fixed by incubation with 4% formaldehyde in PBS for 10 min, washed with PBS, and permeabilized/fixed with ice-cold methanol. The cells were then blocked in 10% fetal calf serum, 20 mM ammonium chloride, and PBS for 30 min and incubated with monoclonal anti-LEDGF p75/p52 antibodies (used at a dilution of 1:300), followed by Alexa-488-or Alexa-568-conjugated goat anti-mouse antibody (Molecular Probes, Leiden, The Netherlands). The nuclear DNA was stained with 5 M ToPro3 iodide (Molecular Probes).
Laser Scanning Microscopy and Image Analysis-Confocal microscopy was performed using an LSM 510 unit (Zeiss, Zaventem, Belgium). SYTO 17 (Molecular Probes) was used to stain DNA of live cells. Prior to image acquisition, cells were washed with serum-free OptiMEM (Invitrogen). All two-and three-color images with a resolution of 1024 ϫ 1024 pixels were acquired in the multi-track mode. EGFP was excited at 488 nm (AI laser), HcRed1 at 543 nm (HeNe laser), and SYTO 17 at 633 nm (HeNe laser). After the main dichroic beam splitter (HFT 488/543 for EGFP and HcRed1, HFT 488/543/633 for EGFP and SYTO 17) the fluorescence signal was divided by a secondary dichroic beam splitter (NFT 635 VIS or NFT 543) and detected in the separate channels using the appropriate filters: BP 505-530 (for EGFP), LP560 (for HcRed1), and LP650 (for SYTO 17). In this set-up, no cross-talk between the green and red channels was observed.
The colocalization of fluorescently-tagged proteins in the nucleus was quantified using the Image-Pro Plus version 4.5 software (Media Cybernetics, Carlsbad, CA) and expressed in terms of the correlation coefficient (r p ) (same as Pearson's r correlation) (46).
and S2 represent the signal intensities of pixels in the first and second channel, respectively; ͗S1͘ and ͗S2͘ are the average intensity of the first channel and second channel, respectively. The correlation coefficient is a value between Ϫ1 and ϩ1, with Ϫ1 corresponding to negative correlation between images and ϩ1 corresponding to a total overlap of the images from the two channels. It reflects similarity of image patterns and does not depend on intensities of the images. Recombinant Proteins-Non-tagged LEDGF/p75 and p52 proteins were produced from the plasmids pCP-Nat75 and pKB-Nat52, respectively, in the Endo I-free PC1 Escherichia coli host strain (E. coli B, BL21(DE3), ⌬endA::Tc R , pLysS) (47). Expression was induced in LB medium at 29°C by addition of 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside. Cells harvested 3 h after induction were disrupted using a French press in 450 mM NaCl, 30 mM Tris, pH 7.0. The supernatant obtained by centrifugation of the lysate was passed through a 1-ml HiTrap heparin column (Amersham Biosciences, Uppsala, Sweden) to capture LEDGF/p75 or p52, and the protein was eluted by a linear gradient of NaCl concentration in 30 mM Tris, pH 7.0. The fractions containing LEDGF/p75 or p52 were pooled and further purified by cation exchange chromatography on a 1-ml HiTrap SP Sepharose column (Amersham Biosciences).
To produce C-terminally His 6 -tagged wild type HIV-1 IN, PC1 E. coli cells harboring pKB-IN6H were grown in LB medium to an optical density of 0.8 and induced by addition of 0.5 mM isopropyl-1-thio-␤-Dgalactopyranoside, at 29°C for 3 h. The protein was purified essentially as described for N-terminally tagged HIV-1 IN (48). In brief, cells were lysed using a French press in 1 M NaCl, 7. Synthetic siRNA Duplexes-The LEDGF/p75-specific siRNA duplexes L1 and L3 were designed according to Elbashir et al. (45) and were synthesized by Xeragon (Germantown, MD) and Dharmacon Research (Lafayette, CO), respectively. The GFP22 (49) siRNA (Xeragon) and the Scramble II duplex (SDII) (Dharmacon) were used as noninterfering controls. All siRNA duplexes were purchased in pre-annealed form and were deprotected and treated according to the instructions from the manufacturer. The L1, L3, and SDII duplexes contained a pair of 3Ј-terminal deoxythymidine (TT) overhangs each. The sequences of the siRNA duplexes and locations of their target sites within LEDGF/p75 mRNA are shown in Fig. 9A. , and merge (right panels). p52 and IN show a discrete nuclear staining during interphase (upper three panels); however, their nuclear patterns are distinct form each other. During mitosis (lower three panels), both proteins are concentrated on the condensed chromosomes, but no clear colocalization can be observed. C, numeric analysis of colocalization. The degree of colocalization was quantified as described under "Experimental Procedures" and expressed in terms of Pearson's correlation coefficient (r p ). Average values and standard deviations obtained from minimum 10 cells for each condition are shown. The correlation coefficients, determined for the HcRed1-p52/EGFP-IN and HcRed1-p52/EGFP-p75 pairs, are approximately equal to that for co-expressed HcRed1 and EGFP-p75 (r p ϭ 0.32 Ϯ 0.09). This can be viewed as a negative control for colocalization. HcRed1-p75 and EGFP-IN co-localize with an r p of 0.67 Ϯ 0.11, which is very close to the r p value for the HcRed1-p75/EGFP-p75 pair (0.69 Ϯ 0.07) (positive control). D, recombinant LEDGF/p75 forms a stable complex with His 6 -tagged HIV-1 IN in vitro, whereas p52 does not. The pull-down assay was performed as described under "Experimental Procedures." The specific buffer conditions, i.e. NaCl concentrations and presence or absence of MgCl 2 in the reactions are indicated above the gels. Recombinant LEDGF/p75 (lanes 3-7) or p52 (lanes 10 -14) was incubated with His 6 -tagged HIV-1 IN, and the complexes were recovered on a Ni 2ϩ -chelating resin. In the control samples (lanes 2 and 9), recombinant IN was omitted. Lanes 1 and 8 reflect protein input in the reactions; BSA, His 6 -IN, LEDGF/p75, or p52 were loaded in the same amounts as were present in the binding reactions. The respective positions of LEDGF/p75, BSA, p52, and His 6 -IN are indicated on the left side of the gel; the gel was stained using Coomassie R250. The same samples were analyzed by Western blotting with anti-LEDGF p75/p52 antibody, for a more sensitive detection of the p52 protein.

Colocalization of Fluorescently
staining of FLAG-tagged IN and LEDGF/p75 showed that the two proteins co-localize in the nucleus. For studies in intact cells, we constructed vectors for expression of EGFP-and HcRed1-tagged human LEDGF/p75 and HIV-1 IN (Figs. 1A and 3A, respectively). Western blot analysis of HeLa cells transfected with the fusion constructs using anti-LEDGF p75/ p52 antibodies (Fig. 1B) or anti-HIV-1 IN (Fig. 3B) and anti-EGFP (data not shown) revealed specific bands at the expected positions. Separately expressed full-length HIV-1 IN and LEDGF/p75, fused to EGFP or HcRed1, showed nuclear localization and distribution patterns in HeLa cells, in agreement with previous reports (36,50). When HcRed1-IN was co-expressed with EGFP-p75 ( Fig. 2A) or EGFP-IN together with HcRed1-p75 (Fig. 4A), a clear nuclear colocalization was observed, suggesting that fusing both binding partners to the fluorescent proteins did not abolish their interaction. As a quantitative measure of colocalization, we determined the correlation coefficient (r p ) (same as Pearson's r correlation; see Ref. 46) for the distributions of the fluorescently tagged IN and LEDGF/p75 proteins obtained from the green and red channels (Fig. 2C). The correlation coefficient for EGFP-IN and HcRed1-p75 in the nucleus was determined to be 0.67 Ϯ 0.11, which is very close to the r p value for the colocalization of EGFP-p75 and HcRed1-p75 (0.69 Ϯ 0.07), which served as a positive control (Fig. 2C). In contrast, when EGFP-p75 fusion and free HcRed1 were co-expressed, neither colocalization nor cross-talk between the green and the red channels was observed, and a low correlation coefficient (r p ϭ 0.32 Ϯ 0.09) was obtained (Fig. 2C).
p52 Does Not Interact with HIV-1 IN-A second protein, p52, can be expressed from the same gene as LEDGF/p75 as a result of alternative splicing of the pre-mRNA (38,41). p52 (333 amino acids) and LEDGF/p75 (530 amino acids) share 325 N-terminal amino acid residues (Fig. 1A). The endogenous levels of p52 in HeLa, HEK-293, or CEM cells were too low to be detected by Western blotting (data not shown), although, at least in the case of recombinant proteins, the sensitivity of detection of p52 was higher than that of LEDGF/p75 (compare lanes 1 and 8 on the blot shown in Fig. 2D). The fact that p52 The distribution of the HcRed1-p52 fusion protein was as shown previously for EGFP-tagged p52 (50) and was clearly different from the distribution of EGFP-IN (Fig. 2B). The mean correlation coefficient for HcRed1-p52 and EGFP-IN colocalization in the nucleus was determined to be 0.32 Ϯ 0.09, which is close to the r p value for our negative controls (Fig. 2C). In addition, HcRed1-p52 did not show colocalization with EGFP-p75 (r p ϭ 0.30 Ϯ 0.12), when the two fusions were co-expressed in HeLa cells (Fig. 2C).
Direct interaction between IN and LEDGF/p75 was confirmed in vitro by a His 6 tag IN pull-down assay with Ni 2ϩ -chelating agarose (Fig. 2D) (Fig. 3A). Transient transfection of HeLa cells resulted in efficient expression of the fusion proteins with predicted molecular masses, as observed by Western blotting with anti-IN antibodies (Fig. 3B) or anti-GFP antibodies (data not shown). Fusions were visualized by laser scanning microscopy (Figs. 4 -7). An overview of the specific properties of the different IN deletion mutants in terms of their intracellular distribution and interaction with mitotic DNA and LEDGF/p75 is given in Table I. Transient expression of EGFP-IN resulted in a specific irregular nuclear distribution pattern, characteristic for IN (Fig.  4A). Strikingly, the absence of the C-terminal DNA binding  a Accumulation of the EGFP-fused protein in the nucleus in comparison to the cytoplasm. ϩϩϩ, predominantly nuclear; ϭ, evenly distributed throughout the whole cell; ϩϩ/Ϫ, present in both nucleus and cytoplasm, more concentrated in the nucleus; ϭ/ϩϩ, nuclear accumulation is observed upon overexpression of LEDGF/p75. b Colocalization with mitotic DNA and LEDGF/p75 is observed upon overexpression of HcRed1-p75. The degree of concentration in the nucleus and the colocalization with LEDGF/p75 and condensed chromosomes are dependent on the expression levels of LEDGF/p75. domain of IN did not have any significant effect on the intracellular distribution. The observed nuclear pattern for the EGFP-IN/⌬C deletion mutant was the same as for the fulllength protein (Fig. 4B). The presence of the other IN domains, however, did play a crucial role in IN intracellular localization. Although EGFP-IN/Nt showed a dispersed cytoplasmic-nuclear localization pattern (Fig. 4D), the EGFP-IN/Ct was exclusively nuclear and concentrated in nucleoli (Fig. 4C), as confirmed by indirect immunostaining using anti-nucleolin antibody (data not shown). Although the predicted molecular mass of EGFP-IN/Ct protein (37.5 kDa for a monomer) is smaller than the exclusion limit of the nucleopore complex (45-60 kDa) (for review, see Ref. 51), taking its diffuse nuclear distribution into account, these data suggest the absence of a specific intranuclear binding partner and favor the idea of an active nuclear import mechanism for EGFP-IN/Ct. Both EGFP-IN and EGFP-IN/⌬C showed clear colocalization with co-expressed HcRed1-p75 (Fig. 4, A and B). The distinct intracellular distribution patterns observed for EGFP-IN/Ct and EGFP-IN/Nt correlated with the absence of colocalization with HcRed1-LEDGF/p75 (Fig. 4, C and D, respectively). Whereas the nuclear localization of full-length EGFP-IN and EGFP-IN/⌬C fusions appeared to be more pronounced in cells expressing HcRed1-p75 (data not shown), no significant changes in the intracellular distribution pattern of EGFP-IN/Nt and EGFP-IN/Ct were observed upon co-expression of HcRed1-p75.
Both HIV-1 IN and LEDGF/p75 have been shown to be associated with condensed chromosomes during mitosis (30,50). In agreement with these reports, fusions of full-length HIV-1 IN and LEDGF/p75 to fluorescent proteins displayed strong chromosomal staining in mitotic HeLa cells when expressed together (Fig. 5A) or separately (data not shown). Unexpectedly, the absence of the C-terminal DNA binding domain in the EGFP-IN/⌬C fusion did not abolish or diminish its association with chromosomes (Fig. 5B). In addition, the EGFP-IN/Ct protein, containing only the C-terminal domain, did not display detectable chromosomal localization, suggesting that the C terminus of IN is not involved in binding to chromosomal DNA (Fig. 5C). Because EGFP-IN/Nt displayed a diffuse distribution in mitotic cells (Fig. 5D), the N-terminal domain is essential but not sufficient for the chromosomal association.  (Fig. 6A). EGFP-IN/⌬N was mostly nuclear, although even at low expression levels, it was readily detectable in the cytoplasm (Fig. 7A) IN (IN H12N ) on the binding to LEDGF/p75 in vitro. In our His 6 tag integrase pull-down assay, a stable complex between recombinant IN carrying the H12N mutation and LEDGF/p75 could not be detected (Fig. 8A), suggesting that the structural fold of the zinc binding domain is important for the interaction with LEDGF/p75. Next, we introduced the same mutation into the EGFP-IN fusion construct (Fig. 3A). The expression level of EGFP-IN H12N in HeLa cells was similar to that of the wild type EGFP-IN protein (Fig. 8B). Remarkably, however, the mutant lost karyophilic properties and was distributed in a diffuse manner throughout the whole cell (Fig. 8C). As was observed for EGFP-IN/core and EGFP-IN/⌬N, EGFP-IN H12N became recruited to the nucleus upon overexpression of HcRed1-p75, and colocalized with the latter (Fig. 8D). In addition, co-expression of HcRed1-p75 also restored the chromosomal localization of EGFP-IN H12N in mitotic cells (data not shown).
Depletion of Endogenous LEDGF/p75 Affects Nuclear Accumulation of HIV-1 IN-To prove the importance of LEDGF/p75 for nuclear import of HIV-1 IN, we used RNA interference to transiently knock-down endogenous LEDGF/p75. After transfection of HeLa or HEK-293 cells with synthetic LEDGF/p75 mRNA-specific siRNA duplexes, ϳ80 -95% depletion in LEDGF/p75 was observed 60 -72 h after transfection (Fig. 9B). The knock-down of LEDGF/p75 did not cause a change in cell morphology or affect the growth of the transfected cells as compared with the cells treated with non-interfering siRNA. Accordingly, a stable knock-down of LEDGF/p75 in human cell lines using hairpin siRNA has been reported to be compatible with cell survival and proliferation (56). However, knock-down of endogenous LEDGF/p75 resulted in a striking defect in nuclear accumulation of EGFP-IN in HeLa cells. EGFP-IN appeared diffusely distributed in cells co-transfected with an LEDGF/p75-specific siRNA (Fig. 9C). Although EGFP-IN was still detectable in the nucleus, the typical irregular distribution pattern was lost. Moreover, in mitotic cells, EGFP-IN was no longer concentrated on the chromosomes (Fig. 9D). Although depletion of LEDGF/p75 varied from cell to cell, indirect immunostaining using anti-LEDGF/p75 antibody confirmed the correlation between knock-down of LEDGF/p75 and the alteration in EGFP-IN distribution (data not shown). We also noticed that depletion of LEDGF/p75 was accompanied by a decrease in the EGFP-IN expression levels, as reflected by the fact that higher input laser power was necessary for efficient detection of EGFP signal after knock-down of LEDGF/p75 (data not shown). A similar loss of the nuclear/chromosomal accumulation of HIV-1 IN concomitant with a decrease in IN levels upon LEDGF/p75 knock-down was also observed in stable cell lines expressing FLAG-tagged IN (data not shown). In addition, incubation of the siRNA-transfected cells in the presence of 5 nM leptomycin B, a potent inhibitor of the CRM1mediated nuclear export (57), for 3-6 h did not cause nuclear re-entry of EGFP-IN (data not shown). Hence, CRM1-mediated nuclear export is probably not involved in the re-distribution of IN upon knock-down of LEDGF/p75. its structural domains. According to their intracellular distribution, the mutants can be subdivided in two families (Table I). The first family is composed of the EGFP fusions that displayed the irregular nuclear distribution pattern as seen with wild type IN and retained the ability to colocalize with LEDGF/p75 throughout the cell cycle. It includes the full-length IN, the ⌬C and ⌬N deletion mutants, the core domain fragment, and the full-length IN H12N mutant. The latter three mutants required overexpression of LEDGF/p75 to display nuclear/chromosomal accumulation and colocalization with LEDGF/p75. The second family consists of the mutants that showed a distinct intracellular distribution and no colocalization with LEDGF/p75. plain the observed phenotype. Of note, Woodward et al. (58) have recently reported that the zinc-binding domain of feline immunodeficiency virus IN is essential for its nuclear localization. It remains to be seen whether feline immunodeficiency virus IN interacts with and requires LEDGF/p75 for its nuclear localization in feline and human cells.

Nuclear and Chromosomal Localization of IN Deletion Mutants Correlates with Their Colocalization with LEDGF/p75-
A second protein product, p52, is expressed from the same gene as LEDGF/p75 as a result of alternative splicing of the pre-mRNA (38,41). It appears to be much less abundant and ubiquitous than LEDGF/p75, which is indirectly reflected by the scarcity of the available human and mouse p52 mRNA-derived expressed sequence tags in GenBank TM . In agreement with Nishizawa et al. (50), we observed a marked difference in the intranuclear distributions of the p52 and LEDGF/p75 proteins. Although p52 had a speckled distribution pattern, LEDGF/p75 appeared more heterogeneous in the nucleoplasm. The distinct nuclear distribution patterns probably reflect different binding partners and functions in the cell. At the amino acid sequence level, p52 shares 325 N-terminal residues with LEDGF/p75. Because p52 did not interact with HIV-1 IN, the C-terminal 205 residue fragment of LEDGF/p75 is likely to harbor the site of interaction with IN.
LEDGF/p75 Is Essential for the Nuclear Accumulation of HIV-1 IN-Transient transfection typically results in a wide range of expression levels, varying from cell to cell. Initially, we considered cells with moderate to low EGFP-IN expression levels as more representative for the in vivo situation, whereby only a limited number of IN molecules are present in the infected cell. In these cells, EGFP-IN was predominantly present in the nucleus. However, in cells that produced higher levels of EGFP-IN, the nuclear accumulation was less pronounced. Overexpression of HcRed1-p75 restored nuclear accumulation of EGFP-IN even at very high expression levels of the latter (data not shown). Hence, when IN overtitrates the endogenous levels of LEDGF/p75, the fraction that is free of the binding partner becomes diffusively distributed. This dependence on LEDGF/p75 expression levels was much more pronounced with the EGFP-IN/core, EGFP-IN/⌬N, and EGFP-IN H12N mutants. Even more compelling evidence that LEDGF/ p75 is essential for nuclear import of IN was provided by our siRNA experiment. The karyophilic properties of IN and its association with chromosomes were completely abolished upon knock-down of endogenous LEDGF/p75. C-terminal fragment, although intrinsically karyophilic, does not associate with and is not trapped on specific nuclear structures, suggesting an active nuclear import mechanism. We speculate that the C-terminal domain may have a role in HIV-1 IN nuclear import, although its karyophilic property seems to be masked in the context of the full-length protein. Interestingly, deletion of the N-terminal domain of IN, but not the mutation of zinc-binding His-12, could expose the NLS. Our results also indicate that the C-terminal domain of HIV-1 IN is not essential for binding to chromosomal DNA and therefore, probably, is not involved in integration site selection. However, it remains possible that, within the viral PIC, the C-terminal domain has a distinct configuration and might play a role in nuclear trafficking and binding to target DNA.
What Is the Mechanism of HIV-1 IN Nuclear Import?-The first mechanism proposed for HIV-1 IN nuclear import implicated the classical importin ␣/␤ pathway (33). A putative bipartite NLS within the C-terminal half of the protein was put forward. However, the mutational disruption of the suggested NLS sequence did not abolish nuclear localization of EGFP-IN fusions (61). A reduction in nuclear accumulation was observed when the putative NLS was mutated in the context of FLAGtagged HIV-1 IN protein (62). These apparently conflicting data might have resulted from the differences between the fusion peptides used. In particular, addition of the negatively charged FLAG tag sequence to the C terminus of HIV-1 IN increases the proportion of protein in the cytosolic fraction (data not shown). In addition, some mutations might have disrupted folding or multimerization of IN resulting in reduced binding to endogenous LEDGF/p75. The involvement of the classical importin ␣/␤ in the nuclear import of HIV-1 IN has been challenged recently by Depienne et al. (34), who showed that, in semi-permeabilized cells, the protein can be imported via a novel importin-and Ran-independent mechanism. What can we learn from the lack of nuclear accumulation of fulllength HIV-1 IN in the absence of LEDGF/p75 protein? The results of our LEDGF/p75 knock-down experiments argue against the involvement of a cellular nuclear import factor other than LEDGF/p75, directing HIV-1 IN into the nucleus. However, three alternatives must be considered. First, the observed re-distribution of IN from nucleus to cytoplasm upon LEDGF/p75 knock-down could be mediated by an active nuclear export of IN molecules that are no longer associated with LEDGF/p75. Although we cannot completely refute this hypothesis, the general CRM1-dependent nuclear export mechanism is probably not involved, because incubation of cells with leptomycin B did not cause nuclear re-entry of EGFP-IN and FLAG-tagged HIV-1 IN in the absence of LEDGF/p75 (data not shown). Second, the pool of cellular IN, not associated with LEDGF/p75, may be subjected to a post-translational modification that would disrupt its nuclear import. Instability of HIV-1 IN in human cells through ubiquitination and subsequent proteasome-dependent degradation of HIV-1 IN has been previously suggested (63). Of note, knock-down of LEDGF/p75 leads to a reduction of IN expression, likely resulting from the proteasome activity, as the effect is sensitive to specific proteasome inhibitors. 2 Other types of post-translational modifications, such as sumoylation or phosphorylation, often implicated in regulation of nuclear import should also be considered (64). Finally, LEDGF/p75 might be necessary for the correct folding of HIV-1 IN. Whether LEDGF/p75 is directly involved in active nuclear import of IN or plays a more passive role, for example, by trapping the imported IN within the nuclear structure, is currently under investigation.
Although the precise role of the LEDGF/p75 protein in the viral life cycle is yet to be unraveled, its specific interaction with HIV-1 IN and its role in nuclear accumulation of IN leaves little doubt that it is a genuine cofactor for HIV replication. Experiments to determine the role of LEDGF/p75 in HIV PIC nuclear import and DNA integration are currently under way. According to one plausible model, LEDGF/p75 is an integral part of the viral PIC stabilizing the IN multimer and targeting the PIC to the nucleus/chromatin. Alternatively, LEDGF/p75 could be a chromosome-bound receptor for the PIC, helping to direct HIV DNA integration into the transcriptionally-active regions of the chromosome. A similar mechanism has been implicated in targeting of yeast retrotransposons. Thus, integration of the Ty5 element is selectively directed to heterochromatin by the specific interaction between Ty5 integrase and Sir4p (65,66). Furthermore, tethering of retroviral INs to specific DNA-binding proteins promotes integration close to the corresponding recognition sites in vitro (67,68). According to both models, LEDGF/p75 would provide the PIC with a target DNA binding/selecting capacity, and, as a result of its association with the general transcription machinery, could account for the marked selectivity of HIV-1 DNA integration into active genes (37). According to one preliminary report (69) and our observations (data not shown), HIV-1 replication is severely impaired upon knock-down of LEDGF/p75 in HeLa-CD4 cells. A novel target for antiretroviral therapy might thus be emerging.