Evidence of Interactions between the Nucleocapsid Protein NCp7 and the Reverse Transcriptase of HIV-1*

The human immunodeficiency virus (HIV-1) nucleocapsid protein NCp7 containing two CX 2CX 4HX 4C-type zinc fingers was proposed to be involved in reverse transcriptase (RT)-catalyzed proviral DNA synthesis through promotion of tRNA3 Lys annealing to the RNA primer binding site, improvement of DNA strand transfers, and enhancement of RT processivity. The NCp7 structural characteristics are crucial because mutations altering the finger domain conformation led to noninfectious viruses characterized by defects in provirus integration. These findings prompted us to study a putative RT/NCp7 protein-protein interaction. Binding assays using far Western analysis or RT immobilized on beads clearly showed the formation of a complex between NCp7 and RT. The affinity of NCp7 for p66/p51RT was 0.60 μm with a 1:1 stoechiometry. This interaction was confirmed by chemical cross-linking and co-immunoprecipitation of the two proteins in a viral environment. Competition experiments using different NCp7 mutants showed that alteration of the finger structure disrupted RT recognition, giving insights into the loss of infectivity of corresponding HIV-1 mutants. Together with structural data on RT, these results suggest that the role of NCp7 could be to enhance RT processivity through stabilization of a p51-induced active form of the p66 subunit and open the way for designing new antiviral agents.

The HIV-1 1 nucleocapsid protein NCp7 is a highly basic protein containing two zinc fingers of the CX 2 CX 4 HX 4 C type ( Fig. 1) found in tight association with the dimeric RNA genome in the retrovirus core (1). In vivo, NCp7 is required for the protection of the genome against cellular nucleases and is involved in genomic RNA packaging and morphogenesis of virus particles (review in Ref. 2). Most of these functions are related to its well demonstrated high affinity for singlestranded nucleic acids (3). NMR studies have demonstrated that the folded CCHC boxes of NCp7 are in spatial proximity, whereas the N-and C-terminal sequences remain flexible (4,5). Mutations inducing modifications in the general conformation of the protein, such as the replacement of His 23 by Cys, Pro 31 by Leu, or Trp 37 by a nonaromatic residue led to more or less important defects in RNA packaging and virus core morphology (6 -11), which seem hardly reconcilable with the complete loss of infectivity of the mutated viruses. One possible explanation could be that changes in NCp7 structure hinders one essential NCp7-dependent step of virus life cycle such as reverse transcription and provirus synthesis (12)(13)(14). This process is catalyzed by the p66/p51 heterodimeric virion-associated reverse transcriptase (RT). This enzyme exhibits RNAand DNA-dependent DNA polymerase activities and an RNase H activity, which are achieved by the p66 polypeptide chain (review in Ref. 15). In vitro, HIV-RT shows an unusual low processivity for a replicative enzyme, suggesting that additional factors are required for efficient viral DNA synthesis in vivo.
In vitro, NCp7 has been shown to activate the annealing of the replication primer tRNA 3 Lys at the initiation site of reverse transcription (16,17) and the 5Ј-3Ј viral DNA strand transfer, leading to provirus formation (18 -21). NCp7 was shown to reduce nonspecific reverse transcription (18,21,22) and to enhance the efficiency and processivity of RT (19,(23)(24)(25)(26), suggesting that it could activate reverse transcription through direct interaction with the enzyme (18, 20, 22-24, 26, 27). Accordingly, NCp7 was shown to be capable of re-establishing strand transfer efficiency and RNase H activity of a defective RT mutant (14).
In this study, RT and NCp7 have been shown for the first time to form a 1:1 complex characterized by an affinity of 6.0 ϫ 10 Ϫ7 M. The structure-activity study has emphasized the critical role of the zinc finger domain in the complexation, giving insights into the loss of virus infectivity of mutants with structurally altered NCp7. Co-immunoprecipitation of NCp7 and RT in a viral environment and cross-linking experiments suggest that NCp7 could enhance RT processivity through stabilization by its zinc finger domain of the p51-induced active form of p66 RNase H domain (28). Moreover, the N-terminal part of NCp7 could compensate for the absence in RNase H domain of HIV-1 p66 RT of the sequence present in the Escherichia coli RT (29). The critical role of the complex during viral DNA synthesis suggests that inhibition of this protein-protein interaction could be an interesting way for designing new antiviral agents.
Virus Production-HIV-1 (NL43 strain) was produced by transfecting plasmid pNL43 into 293T cells using the calcium phosphate precipitation standard procedure. This virus was pseudotyped with the vesicular stomatotitis virus G glycoprotein by co-transfection of vesicular stomatotitis virus G glycoprotein-encoding plasmid. Enzyme-linked immunosorbent assay (kit from E. I. du Pont de Nemours & Co.) was carried out to quantify p24 content of the viral stock. The infectivity of the virus was determined on HeLa cells as described (33). After ultracentrifugation (50,000 ϫ g), the virus was resuspended in 200 l of phosphate buffer, pH 7.5, containing 150 mM NaCl and 0.5% Triton. The concentrations of viral proteins in the sample were 30 g of p24/ml, 10 g of NCp7/ml, and 3.5 g of RT/ml.
Far Western Blot Analysis-2 g of purified HIV-1 p66/p51 RT (Worthington, Freehold, New Jersey) were immobilized onto a Hybond-C super nitrocellulose membrane (Amersham Pharmacia Biotech) in 100 l of TBS (25 mM Tris, 100 mM NaCl, 0.1 mM DTT, 3 mM KCl, pH 7.4) for 3 h at room temperature. The membrane was treated with SuperBlock blocking buffer (Pierce) for 3 h at room temperature to reduce nonspecific interactions and then incubated overnight at 4°C either with NCp7, (a-D)NCp7, (12-53)NCp7, or Vpr (2.1 M) or without proteins as a control in 4 ml of 5% dry milk in TBS buffer containing 0.1% Tween 20. After three washes in TBS-Tween and a 45-min incubation in superblocker, blots were revealed by incubation, first, with 1/15,000 anti-NCp7 mAb (HH 3 for NCp7 and (a-D)NCp7, 2B10 for (12-53)NCp7) or 1/1000 anti-Vpr mAb for 3 h in 5% dry milk TBS-Tween followed by treatment with peroxidase-conjugated anti-mouse antibody for 1 h. These antibodies were not able to cross-react with RT. Complexes were revealed by the ECL method (Amersham Pharmacia Biotech) with peroxidase substrate incubation.
Alternatively, 2 g of HIV-RT were loaded on a 15% SDS-polyacrylamide gel electrophoresis (PAGE) and then transferred onto nitrocellulose membrane. The blot was incubated with 2.1 M NCp7 or without NCp7 as control following the procedure described above using HH 3 mAb against NCp7.
NCp7/RT Binding Studies Using RT Immobilized on Sepharose Beads-HIV-1 RT was covalently linked to CNBr-activated Sepharose 4B beads (Amersham Pharmacia Biotech) using the standard procedure described by the supplier. Typically, 100 mg of CNBr-activated Sepharose beads were swollen and washed in 20 ml of HCl (1 mM), then equilibrated in coupling buffer (0.1 M NaHCO 3 , 500 mM NaCl, pH 8.3), thus providing 350 l of gel, which was then diluted in 350 l of coupling buffer. The batch was divided in two equal parts. 40 g of HIV-1 RT in 200 l of coupling buffer were added to the first batch, and only coupling buffer was added to the second one. After a night at 4°C, excess ligand was washed away, and remaining unbound sites were inactivated in 1 M 1,3-diaminopropane, pH 8, during 2 h at room temperature. The two bead batches were washed with 3 5-min cycles of alternating pH using 0.1 M acetate buffer (500 mM NaCl, pH 4) and 0.1 M Tris-HCl buffer (500 mM NaCl, pH 8). The efficiency of RT immobilization was checked by detecting the quantity of unbound proteins remaining in the supernatant using Western blot analysis and calculated to be superior to 98% (data not shown), suggesting that 0.23 g (2 ϫ 10 Ϫ12 mol) of RT has been captured by l of beads. Then, 20 l of these beads corresponding to about 4.6 g of protein (3.9 ϫ 10 Ϫ11 mol) were equilibrated in TBS, 0.1% Tween 20, 0.1% Nonidet P-40 and incubated with increasing concentrations of NCp7 from 0.1 up to 4.2 M in 100 l of the same buffer for 5 h at room temperature. After centrifugation (6500 rpm), the beads were washed twice with the previous cold buffer at 4°C. NCp7 bound to immobilized RT or nonspecifically to the beads was recovered by heating at 80°C for 5 min in 30 l of Laemmli buffer (50 mM Tris, 10% glycerol, 2% SDS, 0.05% bromphenol blue, 200 mM DTT). Collected samples were loaded on a 20% SDS-PAGE, transferred onto nitrocellulose, and analyzed by Western blot using HH 3 mAb in order to reveal NCp7. The effect of NaCl concentration on NCp7/RT complex formation was measured using the same procedure, except that the beads were incubated with a constant concentration of NCp7 (0.9 M) in TBS-Tween-Nonidet P-40 buffer containing 100 to 500 mM NaCl. When (12-53)NCp7 or its derivatives were used as compet-itors, two concentrations of these peptides (4 and 20 M) were preincubated for 90 min with immobilized RT before the addition of 0.4 M NCp7. In this case, the monoclonal antibody used (HH 3 ) was selective for NCp7 and was unable to cross-react with (12-53)NCp7 or its derivatives. The affinity and stoichiometry of the NCp7/RT complex were determined by quantification of NCp7 bound to immobilized RT or nonspecifically bound to the beads using a standard curve of pure NCp7 and a Bio-profile imager (Vilber-Lourmat). Nonspecific binding was subtracted from the total binding, thus enabling us to calculate the concentrations of bound and free NCp7. The binding parameters were calculated using the Scatchard equation from Enzfit software. Results are means of three independent experiments performed in duplicate.
Cross-linking of RT and NCp7 with 3,3Ј-Dithiobis[sulfosuccinimidyl Propionate] (DTSSP)-RT and NCp7 were cross-linked with a homobifunctional N-hydroxysuccinimide ester-conjugated reagent DTSSP (Pierce), leading to covalent bonds between amino groups of both proteins. After incubation of 1 g of RT and 4.6 g of NCp7 (40 equivalents) for 1 h at room temperature in 10 l of phosphate-buffered saline (3 mM KCl, 1.5 mM KH 2 PO 4 , 140 mM NaCl, 8 mM Na 2 HPO 4 , pH 7.5), DTSSP was added (2 or 5 mM) for 30 min. The cross-linking reaction was stopped by incubating the mixture for 30 min in 50 mM Tris, 6 mM glycine buffer. Proteins boiled in Laemmli buffer without DTT were resolved by electrophoresis on a 12% SDS-PAGE, transferred onto nitrocellulose membrane, and analyzed by two successive Western blots using, first, rabbit polyclonal antibody against RT (a generous gift from S. Litvak) and then, after dehybridation in 0.1 M CH 3 COOH, a mouse mAb against NCp7 (HH 3 ) as described above.
Immunoprecipitations from HIV-1 (NL43 Strain) Crude Extracts-50 l of crude extract obtained from 293T transfected cells were rocked with 50 l of ascite fluid containing mAb against NCp7 (HH 3 or 2B10) overnight at 4°C. Then, 25 l of protein G-Sepharose beads (Amersham Pharmacia Biotech) equilibrated in 80 mM Tris, 80 mM NaCl, pH 8.0, 1% Nonidet P-40 were added, and the suspension was incubated for 4 h at 4°C. Beads were collected by centrifugation (10,000 ϫ g for 5 s) and washed at 4°C twice with the previous buffer and once with a low salt buffer (80 mM Tris, pH 8.0, 0.1% Nonidet P-40, 0.05% sodium deoxycholate). Beads were boiled in Laemmli buffer. Proteins were resolved by electrophoresis on a 15% SDS-PAGE and analyzed by two successive Western blots in order to reveal RT and NCp7 using the previously mentioned procedure.

Evidence for a Molecular Interaction between HIV-1 RT and
Nucleocapsid Protein NCp7-Interactions between NCp7 and heterodimeric p66/p51RT or monomeric p66RT and p51RT forms alone were investigated in vitro by far Western blot analysis using RT samples immobilized on nitrocellulose membrane under nondenaturating ( Fig. 2A) or denaturating conditions (Fig. 2B). The formation of a complex was evidenced using a mAb directed against NCp7 or its derivatives, which is unable to cross-react with native RT. As depicted in Fig. 2A, NCp7 (1st lane) and its central zinc-fingered domain, (12-53)NCp7 (3rd lane), interact with the heterodimeric RT, whereas (a-D)NCp7 (2nd lane), in which both zinc fingers have been re-placed by two Gly-Gly linkers (Fig. 1), was found unable to bind RT. In contrast, even after a long exposure time, no signal could be detected when Vpr, another basic HIV-1 protein (31), was used instead of NCp7 ( Fig. 2A, 4th lane). Likewise, p6, the second maturation product of p15, was found unable to recognize RT (data not shown). Fig. 2B shows that NCp7 recognized both subunits of RT, independently. This interaction is not critically dependent on salt concentration, because only a slight difference was observed when binding experiments were performed using buffers containing 100 up to 500 mM NaCl (Fig.  2C).
The affinity of NCp7 for RT was measured by means of an affinity test based on the immobilization of p66/p51RT onto Sepharose beads through the formation of covalent bonds between amino groups of RT and CNBr-preactivated sites on the resin. The beads were incubated with increasing concentrations of NCp7 from 0.1 up to 4.2 M or without NCp7 as control. After washes, the bound NCp7 molecules were released from the beads by denaturation and analyzed by successive 20% SDS-PAGE, nitrocellulose membrane transfer and Western blot analysis with NCp7 mAb (32) . Fig. 3A, which represents the total binding of NCp7 on immobilized RT, confirmed the formation of a dose-dependent RT/NCp7 complex, previously characterized by far Western analysis. The nonspecific binding of NCp7 (Fig. 2C, 1st lane) was significantly reduced by adding 1,3-diaminopropane to the RT substituted resin, which blocked the remaining CNBr-activated groups and provided positive charges at the surface of the beads, resulting in electrostatic repulsion of the positively charged NCp7. The specific saturation curve and the binding parameters were determined from Scatchard representation. The calculated apparent affinity K D was 0.60 M (Ϯ 0.07) for a 1:1 stoichiometry (n ϭ 0.86 Ϯ 0.04), corresponding to about 1 NCp7 molecule for 1 molecule of the heterodimer p66/p51RT (Fig. 3, B and C).
Structure-Activity Relationships-Since it appears from the experiments with (a-D)NCp7 ( Fig. 2A) that the N-and Cterminal parts of NCp7 are not critical for RT recognition, the contribution of the zinc finger domains was investigated by measuring the direct binding of (12-53)NCp7 and its various mutants, W 16 F 37 (12-53)NCp7, L 37 (12-53)NCp7, and C 23 (12-53)NCp7, to RT using surface plasmon resonance experiments. One concentration (1.6 M) of each peptide was injected to 6000 response units of RT immobilized on the sensorchip in HBS buffer (10 mM Hepes, 50 mM NaCl, 0.005% p20, pH 7.4). As shown on Fig. 4A, (12-53)NCp7 was able to recognize RT. The inversion of aromatic residues in W 16 F 37 (12-53)NCp7 induced a 50% decrease of RT binding as compared with wild-type (12-53)NCp7. Moreover, the replacement of Trp 37 by Leu or His 23 by Cys dramatically disturbed the RT recognition, because in both cases, only 20% of wild-type binding was retained. These results were confirmed by competition experi-ments using immobilized RT on Sepharose beads. For this purpose, the beads were preincubated with two different concentrations of (12-53)NCp7 or various mutants. Wild-type NCp7 was then added at 0.4 M, and the collected samples were analyzed by Western blot as above in order to quantify the amount of NCp7 bound to RT. The monoclonal antibody used (HH 3 ) interacts selectively with NCp7 and is unable to crossreact with the competing NCp7 derivatives. The addition of 10 equivalents of (12-53)NCp7 induced a complete disappearance of the NCp7 band (Fig. 4B, lanes 3-4), supporting the importance of the finger domain in RT recognition. Moreover, the structure of the zinc finger domain appears important for the interaction with RT because W 16 F 37 (12-53)NCp7 appeared to be able to decrease NCp7 binding only at 50 equivalents (Fig.  4B, lanes 9 -10), whereas L 37 (12-53)NCp7 (lanes 5-6) and C 23 (12-53)NCp7 (lanes 7-8) were unable to impede RT/NC recognition.
Cross-linking of RT and NCp7 with DTSSP-The interaction between RT and NCp7 was confirmed by cross-linking experiments using 2 and 5 mM DTSSP, which cross-links compounds in close spatial proximity by formation of covalent amide bonds with amino groups of interacting proteins (Fig. 5, A and B). Therefore, NCp7 and RT were preincubated to allow the complex to be preformed. Then DTSSP was added. After quenching the reaction with Tris-glycine buffer, samples were analyzed by two successive Western blots using antibodies directed either against RT (Fig. 5A) or against NCp7 (Fig. 5B). As expected, this resulted in the formation of various species with apparent molecular masses of 59, 74, 117, and 134 kDa that reacted with antibodies directed against RT (Fig. 5A). The band at 117 kDa corresponds to the coupling of the p51 and p66 subunits. When   FIG. 1. Primary sequences of NCp7 and (a-D)NCp7. (a-D)NCp7 was obtained by replacing both zinc fingers by two Gly-Gly linkers located between Lys 14 and Arg 24 and between Gly 35 and Thr 50 , respectively (30). The sites of His 23 3 Cys, Trp 37 3 Leu and Phe 16 , Trp 37 3 Trp 16 , Phe 37 mutations are indicated.

FIG. 2. Interaction of NCp7 and RT using far Western blot analysis.
A, 2 g of p66/p51 RT were immobilized on nitrocellulose membrane and incubated overnight at 4°C in TBS buffer with NCp7, (a-D)NCp7, (12-53)NCp7, or Vpr (2.1 M). Detection was performed using mAbs directed either against NCp7 (HH 3 or 2B10) or against Vpr, followed by the addition of peroxydase-conjugated anti-mouse antibody. B, 2 g of p66/p51 RT were loaded on 15% SDS-PAGE, transferred onto nitrocellulose membrane, and subjected to far Western blot analysis with 2.1 M NCp7 as probe. C, p66/p51 RT was immobilized on CNBrpreactivated Sepharose beads and incubated with 0.9 M NCp7 in TBS buffer containing increasing concentrations of NaCl from 100 to 500 mM. After washes, bound NCp7 was released through denaturation of beads in Laemmli buffer. Samples were loaded on 20% SDS-PAGE and subjected to Western blot analysis using antibodies directed against NCp7 (HH 3 ). The 1st lane represents the nonspecific binding of NCp7 on the beads at 100 mM NaCl.
anti-NCp7 mAb was used (Fig. 5B), NCp7 could be detected on the previously mentioned adducts migrating at the 59-, 74-, and 134-kDa positions. The first and second compounds correspond to the covalent binding of NCp7 to the p51 and p66 subunits of RT, respectively, and the third one to the binding of two molecules of NCp7 to p66/p51 RT. Interestingly, the p66/ NCp7 cross-linked compound is the most abundant adduct. Weak bands at 16, 24, and 32 kDa, corresponding to dimers, trimers, and tetramers of NCp7, were observed.
Co-immunoprecipitation of RT/NCp7 Complex-The existence of a RT/NCp7 complex was confirmed by co-immunoprecipitation using a crude extract from 293T cells transfected by pNL43 HIV-1 plasmid. The viral preparation was treated with HH 3 or 2B10 mouse monoclonal antibody. Antigen-antibody complexes were isolated by affinity absorption to protein G-Sepharose beads and subsequent elution with an SDS-containing buffer. After gel electrophoresis and transfer onto nitrocellulose membrane, the immunoprecipitated proteins were probed with a rabbit polyclonal antibody against RT (Fig. 6A) then, after dehybridation, with a mouse mAb against NCp7 (Fig. 6B). Significative amounts of NCp7 were retrieved in both cases (Fig. 6B, lanes 1-2), whereas no nonspecific binding on the beads was observed (Fig. 6B, lane 3). Moreover, when using HH 3 mAb directed against the C-terminal part of NCp7, both subunits, but principally p66, co-immunoprecipitated (Fig. 6A,  lane 1), whereas in the case of 2B10 recognizing the first zinc finger of NCp7, essentially p51 was found (Fig. 6A, lane 2). The light and heavy chains of immunoprecipitating mAbs were also stained (Fig. 6B), because these antibodies and the probing secondary anti-NCp7 antibody were both elicited in mice.

DISCUSSION
The role of the nucleocapsid protein NCp7 in HIV-1 viral replication is not as yet well defined. In addition to its involvement in dimerization and packaging of genomic RNA and in virus morphogenesis (2), several studies suggested that NCp7 may function as a key element of the reverse transcriptionally active ribonucleoprotein complex. Thus, in vitro, NCp7 was shown (i) to promote annealing of the tRNA 3 Lys to the primer binding site of HIV-1 genome (16,17), (ii) to accelerate minusstrand DNA transfers during proviral DNA synthesis (18 -21), and (iii) to enhance processivity and RNase H activity of the reverse transcriptase (23)(24)(25)(26). Most of these functions are related to the chaperone activity of NCp7 (34), which shows a high affinity for nucleic acids, in particular for single-stranded RNA. Nevertheless, a direct interaction between NCp7 and RT FIG. 3. Quantitative analysis of NCp7/RT binding. A, total binding of NCp7 on immobilized RT. p66/p51 RT was immobilized on CNBrpreactivated Sepharose beads and incubated at room temperature for 5 h with increasing concentrations of NCp7 from 0.1 to 4.2 M in TBS, 0.1% Nonidet P-40, 0.1% Tween 20. Unbound NCp7 molecules were removed by centrifugation followed by two successive washes at 4°C. Bound NCp7 molecules were recovered by heating in Laemmli buffer and resolved by Western blot analysis using mAb directed against NCp7 (HH 3 ). B, specific saturation curve. Nonspecific binding of NCp7 on the beads was subtracted from the total binding for each NCp7 concentration, providing the specific binding of NCp7 on RT, which was quantified using a standard curve of pure NCp7. Results are means (ϮS.D.) of three independent experiments performed in duplicate. C, Scatchard representation of NCp7/RT binding. The calculated binding parameters are: K D ϭ 6.0 Ϯ 0.7 ϫ 10 Ϫ7 M, n ϭ 0.86 Ϯ 0.04.
FIG. 4. RT/NCp7 recognition: zinc finger structure-activity relationships. A, direct binding of (12-53)NCp7 and its mutants to RT using surface plasmon resonance technology. The same concentration (1.6 M) of (12-53)NCp7, W 16 F 37 (12-53)NCp7, L 37 (12-53)NCp7, or C 23 (12-53)NCp7 was injected to 6000 response units of RT immobilized on the sensorchip in HBS buffer at 25°C. Each sensorgram represents one typical experiment. B, immobilized RT was preincubated or not with (12-53)NCp7 and its derivatives (10 and 50 equivalents) for 1.5 h before the addition of 0.4 M wild-type NCp7 in TBS-Tween. Unbound proteins were removed by centrifugation followed by two successive washes with TBS-Tween at 4°C. Bound NCp7 was recovered after denaturation of complexes in Laemmli buffer. Detection of NCp7 was carried out by Western blot analysis using antibody selective for wildtype NCp7 (HH 3 ). Lane Cϩ corresponds to specific binding of NCp7 to immobilized p66/p51RT, and lane CϪ corresponds to the same experiment with beads without RT.
In this work, we clearly demonstrate that NCp7 is able to bind in vitro in a dose-dependent manner the heterodimeric form of RT as well as both subunits independently. This interaction seems to be essentially hydrophobic since increasing NaCl concentrations have little effect on NC/RT complex formation. Under our experimental conditions, this interaction appears to be specific since Vpr, another small and positively charged retroviral protein, and p6 were unable to recognize RT. The binding constant K D was found to be 6.0 Ϯ 0.7 ϫ 10 Ϫ7 M for a 1:1 NCp7-p66/p51RT complex. This affinity is in agreement with both the mM concentration of NCp7 in the virion and the 1/40 RT/NCp7 ratio (36). Because NCp7 interacts in vitro with both subunits of RT, the binding site of NCp7 on the p66/p51 heterodimer could be located at the region connecting the two subunits.
Direct binding experiments with various NCp7 fragments demonstrated that the flexible N-and C-terminal parts of NCp7 are not involved in NCp7/RT complexation, since the finger domain-deficient (a-D)NCp7 did not recognize RT. In contrast, the central region, encompassing both zinc fingers and the basic linker RAPRKKG, seems to be crucial. The importance of the conserved structure of zinc finger domains (4,37,38) in NCp7/RT binding was confirmed by surface plasmon resonance and competition experiments. Thus, although the structure of W 16 F 37 (12-53)NCp7 3 was found close to that of (12-53)NCp7, this mutant was twice less active than NCp7 to bind RT. The replacement of Trp 37 by Leu in L 37 (12-53)NCp7, a modification that does not perturb the zinc finger folding 3 but impairs NCp7-induced annealing of tRNA 3 Lys to the primer binding site (17), strongly reduced the inhibitory effect of this mutant, suggesting that Trp 37 is important for NCp7/RT interaction (Fig. 4A, lanes 5 and 6). The mutation of His 23 to Cys 23 (12-53)NCp7 induces a rearrangement of the residues around the zinc ion in the first finger and increases the distance between the two zinc fingers (6). This mutant was also unable to impede RT/NCp7 recognition (Fig.  4A, lanes 7 and 8). From these data, it can be to concluded that at least in vitro, the interaction between NCp7 and RT requires the central zinc finger domain of NCp7. It is therefore tempting to explain the complete loss of infectivity observed with mutants characterized by conformational modifications of NCp7 by a defect in retrotranscription rather than by modifications in virus morphology or reduction in packaged genomic RNA (6,12,13).
Another striking result is that NCp7 is able to recognize both subunits of RT, although in chemical cross-linking experiments, the covalent binding of NCp7 occurred preferentially with p66. On the other hand, immunoprecipitation experiments performed on the supernatant of cells transfected by HIV-1 pNL43 plasmid using two different antibodies against NCp7 resulted in precipitates containing NCp7 and either p66 or p51 subunit. When HH 3 recognizing the C-terminal part (52-67) of NCp7 was used, NCp7 seems to interact only with p66. On the other hand, when the 2B10 mAb covering an epitope located in the first zinc finger (19 -27) of NCp7 was 3 N. Morellet, unpublished data. FIG. 7. Proposed mechanism of recognition between NCp7 and RT. NCp7, captured from the structure of NCp7 complexed with SL3, has been located at the connection site between p66 and p51 subunits of RT (adapted from Katz and Skalka (15)) with the N-terminal zinc finger facing p66 and the C-terminal one directed toward p51. This accounts for the results presented here. The two zinc fingers could be considered as a hinge element, stabilizing the p51-induced active form of p66 RNase H domain. The N-terminal helical part of NCp7 has been disposed in the position that it takes in the structure of the complex between NCp7 and the double-stranded stem part of SL3 (43).

FIG. 5. Chemical cross-linking of NCp7 and RT by DTSSP.
NCp7 and RT were preincubated in phosphate buffer 1 h at room temperature, then the cross-linking agent DTSSP was added at a final concentration of 2 or 5 mM. After 30 min, the reaction was stopped by the addition of Tris-glycine buffer. Samples were then analyzed by two successive Western blots using a polyclonal antibody specific of RT (A) followed by dehybridation and subsequent revelation using a monoclonal antibody directed toward NCp7 (B).
FIG. 6. Co-immunoprecipitation of NCp7/RT complex from HIV-1 viral extract. RT and NCp7 were co-immunoprecipitated from supernatant of 293T cells transfected by pNL43 HIV-1 plasmid with monoclonal antibody recognizing NCp7: lanes 1, HH 3 ; lanes 2, 2B10. Samples were developed by Western blot analysis. Detection was performed with either polyclonal antibody directed against RT (A) or monoclonal antibody directed against NCp7 (B). Lanes 3 represent the nonspecific binding of RT and NCp7 on the protein G-Sepharose beads. used, the nucleocapsid protein was found to bind only p51. This suggests that NCp7 could interact at the interface between the two subunits with the N-terminal zinc finger in proximity of p66 and the C-terminal zinc finger directed toward p51 (Fig. 7).
NCp7 has been shown to enhance the global processivity of reverse transcription by accelerating RT-catalyzed DNA strand transfer reactions through modulation of RNase H activity (20,23). The crystal structure of the RNase H domain of HIV-1 p66-RT shows that the protein folding is similar to that of E. coli RNase H despite a low (24%) amino acid homology between both proteins (29). But in contrast to E. coli RNase H, the HIV-1 RNase activity is critically dependent of the presence of the other domains of RT. In the viral enzyme, this could be related to the C-terminal region of HIV-1 RNase H encompassing the helix ␣E, which was found flexible and unstructured by NMR (39) and crystallographic (29) studies. Accordingly, the addition of purified p51 to the isolated domain led to the reconstitution of HIV-1 RNase H activity in vitro (28). Therefore, the biological relevance of the present results could be that the zinc finger domains of NCp7, by interacting with RT at the interface between the two subunits, reinforce the required structuration of the catalytic HIV-1 RNase H domain ensured by p51 subunit, thus improving its activity. This hypothesis is consistent with the model proposed by Cameron et al. (14), since their experiments showed that NCp7 was able to restore strand transfer efficiency and RNase H activity of a defective RT mutant containing a deletion in the C-terminal part of p51 found in spatial proximity of the RNase H domain.
On the other hand, a highly basic helix/loop region present in E. coli enzyme is missing in the HIV-1 domain. Interestingly, the introduction of this sequence in wild-type HIV-1 RNase H induced enzyme activity (40). In E. coli, this basic helix ␣C, QWIHNWKKR, has been proposed to contribute to the proper binding and positioning of the double-stranded RNA/DNA hybrid substrate, thus improving the catalytic process (41,42). Moreover, the basic N-terminal sequence of NCp7 is flexible in solution but was recently found to interact with the doublestranded RNA stem of the SL3 domain of genomic HIV-1 RNA under a helical form (43). Therefore, in agreement with the present results, it is tempting to propose that NCp7 could improve HIV-1 RNase H activity through RNA/DNA recognition by means of its N-terminal domain and stabilization of p66/p51 heterodimer by the zinc finger domain. Finally, this study provides an interesting new approach for the rational design of antiviral agents.