Extended nucleocapsid protein is cleaved from the Gag-Pol precursor of human immunodeficiency virus type 1 -- Chen et al. 82 (3): 581 -- Journal of General Virology

The structural and catalytic proteins of human immunodeficiency virus type 1 (HIV-1) are translated as part of two polyproteins, Gag (Pr55) and GagPol (Pr160). The 20:1 ratio of Gag to GagPol is determined by a ribosomal frameshift (FS), which takes place at the N terminus of the short peptide p1 (also named SP2) (Jacks et al., 1988 ; Ratner et al., 1985 ). Both viral precursors are processed into the proteins of the mature virion by a virus-encoded protease (PR), which is synthesized as part of the GagPol precursor and located downstream of the p6^Pol protein. Leu is the N-terminal residue of the p6^Pol protein, located 8 amino acid residues downstream of the C terminus of the nucleocapsid (NC) protein (Phylip et al., 1992 ; Almog et al., 1996 ; Zybarth et al., 1994 ). The scissile site at the NC/p1 junction in the Gag precursor (ERQAN-FLGKI) was determined and shown to be a substrate for PR (Henderson et al., 1988 , 1992 ; Wondrak et al., 1993 ). However, due to the frameshift event, the amino acid sequence in this area is changed. The N terminus of p6^Pol was previously determined by Edman analysis (Almog et al., 1996 ), but it is not yet known whether PR cleaves the GagPol precursor at this site (DLAF-LQGK), or at the frameshift vicinity (Fig. 1). If NC protein is released from the GagPol precursor (NC*) by a sole cleavage at the N terminus of p6^Pol, then it should be 8 amino acid residues longer than the NC released from Gag.

(22K): Fig. 1. (A) The HIV-1 Gag and GagPR regions. The upper line describes the coding sequence of the gag gene. The dotted line shows the frameshift site (FS), which places the gag and pol genes into the same reading frame. Gag proteins p1 and p6 are not translated from this construct because of the frameshift mutation. The truncated GagPol polyprotein therefore contains the MA, CA, p2, the putative NC* protein (illustrated with a double box), p6Pol and PR. (B) DNA (in italics) and protein (boxed) sequences of HIV-1 at the frameshift vicinity. The Gag lane describes the sequence of the NC/p1 junction (the gag gene sequence). Lanes GagPol(FS1) and GagPol(FS2) show the frameshift site (fs). The bases a and t were inserted to generate the frameshift mutations (FS1 or FS2, respectively). The NC*/p6Pol junction is between the two frames in the GagPol reading frame. The 8 additional amino acids present in NC* are in bold. (C) Schematic representation of the wild-type (wt) and cleavage-site-mutated GagPR fragments. The mutated cleavage sites are marked with an X, wild-type and mutated segments were cloned into the pT5 plasmid. cs1 shows GagPR with a cleavage site mutation in the NC*/p6Pol junction. cs2 shows GagPR with a cleavage site mutation in the p6Pol/PR junction. cs1+2 shows cleavage site mutations in the NC*/p6Pol/PR junctions. DNA (in italics) and protein sequences of the mutated NC*/p6Pol and p6Pol/PR junctions are given to the right of the mutated constructs (cs1 and cs2, respectively). The Phe residue at p1' in both junctions was replaced by Ile in order to prevent the cleavage by PR. The frameshift site mutation is indicated (FS).

Here we report that: (i) abrogation of the cleavage sites at the N terminus of p6^Pol and the P6^Pol/PR junction results in the production of an NC*p6^PolPR fusion protein, suggesting that a single scissile site exists between NC and p6^Pol; (ii) synthetic peptides, homologous to the sequence at the NC and frameshift vicinity junctions, were resistant to PR digestion, while the peptide consisting of the NC*/p6^Pol site was hydrolysed under the same conditions; and (iii) NC* was identified with monoclonal anti-NC in bacteria expressing truncated GagPol precursor (GagPR) and in COS-7 cells expressing the frameshift-mutated GagPol. Taken together, these results suggest that the NC* released from the GagPol precursor is 8 amino acid residues longer than NC cleaved from the Gag precursor.

Previously it was suggested that the autoprocessing of HIV-1 Gag by PR is an ordered process; the p2/NC junction is the first to be cleaved from Gag, while the CA/p2 is the last site processed (Tritch et al., 1991 ; Pettit et al., 1994 ; Krausslich et al., 1995 ). Cleavage of Gag at p2/NC yields a fusion intermediate product consisting of membrane-associated protein (MA), capsid protein (CA) and p2 peptide (Tritch et al., 1991 ; Pettit et al., 1994 ; Krausslich et al., 1995 ; Vogt, 1996 ; Accola et al., 1998 ). If the cleavage at the p2/NC junction is indeed the first to take place in both polyproteins, then the release of PR from its precursor is not a prerequisite for the initiation of GagPol precursor processing. Previously, we and others showed that p6^PolPR and PRreverse transcriptase (RT) fusion proteins are enzymatically active (Kotler et al., 1992 ; Zybarth et al., 1994 ; Louis et al., 1994 ; Tessmer & Krausslich, 1998 ; Cherry et al., 1998 ). Here we support and extend these findings and report that HIV-1 PR is enzymatically active as an integral part of a larger fusion protein which includes NC*p6^PolPR.

Mammalian cells.
CV-1 and COS-7 cells were grown in Dulbeccos modified Eagle medium, supplemented with 10% foetal calf serum, glutamine and penicillin/streptomycin (Biological Industries, Kibbutz Beit Haemek, Israel). H9 cells were grown in RPMI 1640 supplemented with 10% foetal calf serum (Biological Industries).

Bacterial cells.
E. coli strain BL21 was used as a host for vectors expressing wild-type and mutated MACANCp6^PolPR (GagPR) constructs (Fig. 1C). E. coli strain DH5α was used to propagate and maintain all plasmids.

Plasmids and mutagenesis.
CANCp6^PolPR constructs were generated by PCR as previously described (Almog et al., 1996 ; Kotler et al., 1992 ). Briefly, an NcoI site was introduced into the DNA segments of HIV-1 BH10 strain that encode the N terminus of CA. The NcoI site contains an ATG that serves as an initiator for protein synthesis. A termination codon was introduced just downstream of Phe, the last codon of PR, at position 2125. To permit efficient translation through the Gag/Pol junction, frameshift mutations were introduced: either downstream of the A nucleotide (nt 1634) by inserting an additional base A (FS1), or downstream of the T nucleotide (nt 1631) by inserting an additional base T (FS2) (Fig. 1B). These single-base insertions put Gag and Pol into the same reading frame. The amplified sequences of HIV-1 containing the CAPR were cloned into pUC12N (Kotler et al., 1992 ; Vieira & Messing, 1982 ; Norrander et al., 1985 ) for propagation and sub-cloned into the pT5 plasmid (Studier et al., 1990 ) for expression, as previously described (Kotler et al., 1992 ).

A cleavage site mutation (cs) was introduced at the NC*/p6^Pol scissile site by a PCR reaction that replaced Phe with Ile, thus preventing hydrolysis at this site (Fig. 1C). The amplified DNA was cut with BglII and BamHI and inserted into pUC12N (CAPR). The double cleavage-site-mutated HIV-1 sequence (CANCcsp6^PolcsPR) was constructed by amplification of the same DNA using pUC12N (CANCp6^PolcsPR) as the template. The viral fragments were rescued from the pUC12N plasmids by cleavage with NcoI and SpeI and transferred to a pT5 vector cleaved with NcoI and XbaI restriction enzymes, respectively.

The MA-encoding fragment was generated by a PCR reaction and coupled to the wild-type and mutated pT5 (CAPR) to form the wild-type pT5 (GagPR) and the cleavage site mutants cs1 (NC*csp6^Pol), cs2 (p6^PolcsPR) and cs1+2 (NC*csp6^PolcsPR).

Expression of HIV-1 proteins in bacterial cells.
The pT5 plasmids containing wild-type and mutated GagPol fragments were used to transform E. coli BL21 cells, which contain an inducible T7 polymerase (Fuerst et al., 1986 ). Bacterial cells were grown to saturation in LB medium at 37 °C overnight, cultures were diluted 1:20 and grown at 37 °C to a density of 0·40·6 OD₆₀₀ units before IPTG was added to a final concentration of 0·4 mM to induce the T7 polymerase. Bacteria were harvested 2 h post-induction by centrifugation (3 min, 10000 r.p.m.), and pellets were re-suspended in 200 µl of Laemmli loading buffer.

Expression of HIV-1 GagPol fragments in CV-1 cells.
Monolayer cultures of 2x10⁶ CV-1 cells in 100 mm tissue culture dishes (Nunc) were infected with vaccinia virus expressing the T7 polymerase (vTF7-3) (Fuerst et al., 1986 ; Karacostas et al., 1993 ) at an m.o.i. of 10. Two hours post-infection cells were transfected by the CaPO₄ method (Demetrios & Welkie, 1984 ) with pT5 pDNAs containing wild-type HIV-1, or mutated DNA.

Expression of HIV-1 GagPol fragments in COS-7 cells.
Monolayer cultures of 2x10⁶ cells in 100 mm tissue culture dishes (Nunc) were co-transfected by the DEAE-dextran method with plasmids pSVGAGPOL-RRE-RFS5T (pSV^GAGPOL) or pSVGAGPOL-RRE-R (codes for HIV-1 GagPol with and without frameshift mutation) and pCMVrev (codes for the Rev protein) (Smith et al., 1993 ). These plasmids were kindly provided by Professor David Rekosh (University of Virginia, Charlottesville, VA, USA). Forty-eight hours post-transfection the cells were washed twice with ice-cold PBS+0·1 mM PMSF, harvested and re-suspended in Laemmli buffer.

Virus production and cell infection.
H9 cells were infected with HIV-1_IIIB. Viral spread in the culture was monitored by p24 assay (Organon Teknika). To obtain virus particles, medium of infected cells was centrifuged for 15 min at 10000 g, and supernatant was centrifuged through a 20% sucrose cushion (100000 g for 60 min). Pellets were re-suspended in PBS and stored in aliquots at -70 °C.

Detection of HIV-1 proteins.
Bacterial or CV-1 cell extracts were run electrophoretically on polyacrylamide gels, blotted onto nitrocellulose filters (0·2 µm, Gelman Sciences), which were treated with rabbit anti-PR, monoclonal anti-CA or anti-MA (kindly supplied by Dr S. M. Nigida, SAIC, NCI, Frederick, MD, USA), and developed by alkaline phosphatase (Sigma) or enhanced chemiluminescence (Sigma). For the detection of NC and NC* proteins, cell lysates (bacteria, COS-7) and HIV-1_IIIB virions were run on 16·5% or 20% Tricinepolyacrylamide gels, blotted onto PVDF filters and treated with monoclonal anti-NC (SAIC, NCI, Frederick, MD). To prevent loss of NC molecules during blotting, bovine haemoglobin (5 µg) was added to each slot and to the transfer buffer (0·25 mg/ml) as previously described (Gillespie & Gillespie, 1997 ).

Peptide cleavage assay.
Oligopeptides were synthesized and characterized as described before (Baraz et al., 1998 ; Friedler et al., 1999 ). Assays were performed in 10 µl reactions containing 50 mM sodium phosphate buffer, pH 5·6, 0·1 M NaCl, 10 nmol of decapeptide, and 10 pmol of purified HIV-1 PR. The assay mixtures were incubated for 60 min at 37 °C, and the reactions were stopped by addition of guanidine hydrochloride (6 M final concentration). The cleavage products were analysed by reverse-phase HPLC on a C₁₈ column (Vydac, 250x4 mm), using a gradient of 050% (v/v) acetonitrile in 0·1% aqueous trifluoroacetic acid at a flow rate of 1 ml/min, and absorbance at 220 nm was recorded.

HIV-1 protease inhibitor Ro 31-5989 (Roberts et al., 1990 ) was kindly supplied by Roche Products.

A single scissile site separates NC and p6^Pol
Truncated HIV-1 GagPol precursors with frameshift mutation, which adjusts pol and gag genes to a single reading frame (Fig. 1A), were expressed in E. coli BL21 cells. The products of the GagPR autoprocessing in bacteria are shown in Fig. 2(A). This polyprotein is cleaved into a NC*p6^PolPR (24 kDa) fusion protein, which is further cleaved into a p6^PolPR fusion protein (see also Almog et al., 1996 ). Because of internal cleavage sites located in the C terminus of the PR protein, several intermediate proteins of 1922 kDa are produced in bacterial cells.

(57K): Fig. 2. Immunoblot analysis of wild-type and mutated GagPR proteins and their processed products expressed in E. coli strain BL21. Cell lysates were fractionated on an SDSpolyacrylamide gel and viral proteins were detected with rabbit anti-PR (A), anti-CA (B) and anti-MA (C). Lysates from cells expressing PR or CANC and uninduced cells (BL21) were included as controls. Molecular markers (kDa) are indicated on the left. The presumptive processed polypeptides are indicated.

The cs2 mutation, which prevents the separation of PR from p6^Pol, allows the release of P6^PolPR fusion protein, and reduces the amount of intermediate products. Alteration of the cleavage at the N terminus of p6^Pol (cs1) reduces both the number and amount of the intermediate products and enables the release of 11 kDa PR. Prevention of the cleavage at both junctions (cs1+2) abolishes the autoprocessing of this polyprotein, suggesting that a single scissile site exists between NC* and p6^Pol. Had an additional cleavage site existed at the C terminus of NC, then a 17 kDa p6^PolPR protein should have been released. These results also suggest that NC is cleaved from the GagPol polyprotein downstream of the frameshift site at the N terminus of p6^Pol.

The autoprocessing results were further investigated by biochemical experiments. As previously described (Jacks et al., 1988 ), ribosomal frameshift occurs either as FS1 or FS2 (Fig. 1B). We synthesized decapeptides presenting the sequence at the frameshift vicinity (NC/FS1, NC/FS2) and peptides homologous to the wild-type and mutated NC/p6^Pol junction (Fig. 3). The decapeptides NC/FS1 and NC/FS2 were not cleaved by purified HIV-1 PR in a cell-free system (Fig. 3 AD). These peptides were not hydrolysed even in high concentrations of PR and after a longer incubation time, conditions which allow almost complete hydrolysis of the accessible NC*/p6^Pol peptide (data not shown). In contrast, the decapeptide consisting of the NC*/p6^Pol junction was cleaved by HIV-1 PR under the same conditions (Fig. 3E and F).

(19K): Fig. 3. Cleavage of decapeptides corresponding to the junction of NC and the frameshift area [NC/FS2 (A, B), NC/FS1 (C, D)], putative scissile site NC*/p6Pol downstream of the frameshift site (E, F) and the mutated NC*/p6Pol site [NC*csp6Pol (G, H)]. NC/FS1, Arg-Gln-Ala-Asn-Phe*Leu-Arg-Glu-Asp-Leu; NC/FS2, Arg-Gln-Ala-Asn-Phe*Phe-Arg-Glu-Asp-Leu; NC*/p6Pol, Glu-Asp-Leu-Ala-Phe*Leu-Gln-Gly-Lys-Ala; NCcsp6Pol, Glu-Asp-Leu-Ala-Ile*Leu-Gln-Gly-Lys-Ala. Ten nmol oligopeptides were incubated with 10 pmol of purified HIV-1 PR (B, D, F, H) or without (A, C, E, G) and incubated for 1 h at 37 °C. The reactions were terminated by the addition of guanidine hydrochloride (final concentration of 6 M). Cleavage products were analysed by reverse-phase HPLC on a C18 column (Vydac) using 050% (v/v) acetonitrile in 0·1% aqueous trifluoroacetic acid at a flow rate of 1 ml/min; absorbance at 220 nm was recorded. S, Substrate; P, product.

PR could not hydrolyse the decapeptide with the mutated NC*csp6^Pol junction, supporting the finding that the mutation inserted at the N terminus of p6^Pol prevents the cleavage at this site (Fig. 3G and H). These results show that HIV-1 PR cleaves at NC*/p6^Pol but not at the NC/FS junctions and support the finding that a sole cleavage site separates the NC from p6^Pol. Thus, the NC released from GagPol precursors is composed of an additional 8 amino acid residues which are not present in the homologue protein released from the Gag precursor.

NC*p6^PolPR fusion protein is enzymatically active
Expression of cs1+2 mutant in E. coli BL21 cells has shown that PR could release the NC*p6^PolPR fusion protein from the GagPR precursor (Fig. 2A). We were therefore interested to assess whether this fusion protein accurately cleaves at the other sites in the polyprotein. To this end we expressed the GagPR constructs in bacterial and mammalian cells and analysed the cleavage products by Western blots. Fig. 2(B, C) shows that prevention of the cleavage at NC*/p6^Pol, p6^Pol/PR or at both junctions does not interfere with the autocatalysis of the polyproteins, since CA and MA proteins were released. The p41 (MACAp2) protein is the first to be cleaved from the polyprotein. This intermediate product was completely processed in the cs1 and cs1+2 mutants but only partially cleaved in the wild-type and cs2 mutant. We therefore concluded that PR is enzymatically active as part of the NC*p6^PolPR (24 kDa) fusion protein, since other sites in the polyprotein were cleaved normally. Addition of 100 µg/ml of Ro 31-5989 inhibitor during IPTG induction (as described in Fig. 5) verified that the fused viral PR exclusively processed the GagPR polyprotein (not shown).

(46K): Fig. 5. (A) NC* protein expressed in bacteria (GagPR/BL21) versus NC protein in virus pellet (HIV-1 IIIB). E. coli BL21 cells transformed with the pT5 (GagPR) construct were induced by IPTG for 30 min and then harvested and resuspended in Laemmli buffer. HIVIIIB virions were pelleted from medium of H9-infected cells by ultracentrifugation through a 20% sucrose cushion. Cells were washed twice in ice-cold PBS+0·1 mM PMSF, harvested and resuspended in Laemmli buffer. Proteins were separated on a 16·5% Tricinepolyacrylamide gel, blotted onto a PVDF filter and treated with monoclonal anti-NC antibodies. (B) Inhibition of the autoprocessing of GagPR polyprotein expressed in bacterial cells by the protease inhibitor Ro 31-5989. E. coli BL21 cells expressing HIV-1 GagPR were treated with 0, 10, 100 and 1000 µM of the protease inhibitor Ro 31-5989 simultaneously with induction by IPTG. Cell induction and protein separation and treatment were as described in (A). (C) Immunoblot analysis of COS-7 cells. Cells were co-transfected with either plasmids encoding the GagPol polyprotein and rev protein (GagPol+Rev) or with the frameshift-mutated GagPol and rev (GagfsPol+Rev). As a negative control cells were also transfected with pCMVrev (Rev). Post-transfection (48 h) cells were washed twice with ice-cold PBS+0·1 mM PMSF, harvested and resuspended in Laemmli buffer. Proteins were separated on a 20% Tricinepolyacrylamide gel, blotted onto a PVDF filter and treated with monoclonal anti-NC antibodies. Size markers (kDa) and presumptive processed proteins are indicated.

CV-1 cells were transfected with pT5 (GagPR) constructs following infection with vaccinia virus expressing T7 polymerase (Fuerst et al., 1986 ; Karacostas et al., 1993 ). Both wild-type and cleavage-site-mutated GagPR underwent efficient processing, releasing MACAp2, and CA (Fig. 4A). Fig. 4(B) verifies that the cleavage site mutations prevent the release of PR, p6^Pol and/or NC, while the polyproteins underwent autoprocessing. Thus, NC*p6^PolPR is enzymatically active in eukaryotic cells. Moreover, as shown in bacterial cells (Fig. 2A), the accumulation of the 24 kDa fusion protein in eukaryotic cells (Fig. 4B, lane 4) indicates again that single cleavage is required to separate the NC* from the p6^Pol protein.

(27K): Fig. 4. Expression and processing of wild-type or cleavage site mutations of GagPR polyprotein in CV-1 cells. 2x106 cells were infected with vaccinia virus expressing the T7 polymerase (vTF7-3) and transfected 2 h later with constructs expressing either the wild-type or cleavage site mutants. Cells were harvested 48 h post-transfection, and lysates were Western blotted and treated with monoclonal anti-CA (A) or rabbit anti-PR (B). Lysates of cells transfected with only pT5 vector (pT5) or not transfected (mock) and lysate of bacteria expressing wild-type GagPR (GagPR/BL21) or the viral protease (PR) were included as controls.

Nucleocapsid protein released from GagPol is larger than NC
Expression of GagPR polyprotein in E. coli BL21 cells by pT5 vector yields equimolar amounts of Gag and Pol proteins, while the Gag proteins residing downstream of the frameshift are not translated (Fig. 1A). Truncated GagPR polyprotein undergoes autocleavage following expression in bacterial cells, releasing NC* protein. The migration coefficient of NC* expressed in bacterial cells was compared to NC presented in IIIB virions by SDSPAGE. Fig. 5(A) shows that the NC* protein released from GagPR migrates slower in the gel than the NC protein extracted from HIV-1_IIIB. The NC* protein is exclusively cleaved from GagPR polyprotein by HIV-1 PR, since the specific HIV-1 protease inhibitor Ro 31-5989 (Roberts et al., 1990 ) inhibits the release of this protein in a dose-dependent manner (Fig. 5B). The concentration of Ro 31-5989 required to inhibit PR in bacterial cells is higher than that used in eukaryotic cells, probably due to the high concentration of viral precursor in the bacteria, low permeability or instability of the drug in the bacterial culture.

To demonstrate the release of NC* protein from the GagPol precursor in eukaryotic cells, COS-7 cells were co-transfected with pSV^GAGPOL and pCMVrev plasmids (Smith et al., 1993 ). These constructs co-expressed the wild-type and frameshift-mutated GagPol and the Rev proteins. An NCp1 band, which was identified by specific anti-p1 antibodies (data not shown), is clearly distinguishable from the NC protein (Fig. 5C, lane 2). NC* protein is released from the frameshift-mutated GagPol (lane 3). As anticipated, this NC* protein (63 aa) migrates to midway between NC (55 aa) and NC+p1 (72 aa). NC* protein could not be detected in lane 2 because it is expressed at a ratio of 1:20 of NC and NC-p1, which are expressed by the pSV^GAGPOL construct. Taken together, these results demonstrate that the GagPol polyprotein is processed by the viral PR, yielding the NC* protein, which is 8 amino acids longer than the NC released from Gag.

The HIV-1 NC protein is cleaved from the Gag precursor at the p2/NC and NC/p1 junctions (Henderson et al., 1988 , 1992 ; Swanstrom & Wills, 1997 ). The ribosomal frameshift events, which take place at the N terminus of the p1 peptide, modify the NC/p1 cleavage site, making it non-cleavable by HIV-1 PR. We anticipate, therefore, that the NC* embedded in the GagPol precursors will be cleaved out from the 8 amino acid residues downstream of the C terminus of the NC protein, located in the Gag polyprotein. This cleavage takes place adjacent to the N terminus of the p6^Pol protein. The experiments described in this report show that the NC* protein indeed bears 8 additional amino acid residues at its C terminus which are absent in NC protein and that the NC*p6^PolPR is enzymatically active.

This assumption was supported by demonstrating that: (i) abrogation of the NC*/p6^Pol and NC/p6^Pol/PR cleavage sites causes accumulation of NCp6^PolPR fusion protein, indicating that a single cleavage site separates NC from p6^Pol; (ii) synthetic decapeptides homologous to the junctions at the C terminus of NC* (NC fs1 and NC fs2) are not cleaved by HIV-1 PR, while peptide identical to the NC*/p6^Pol junction is efficiently hydrolysed by the viral PR. These results strongly suggest that HIV-1 PR is unable to cleave at the C terminus of NC when it is part of the GagPol precursor; (iii) NC released from Gag polyprotein, extracted from mature virions, was found to be of lower molecular mass than NC* released from the GagPR expressed in bacterial cells; and (iv) NC* protein cleaved from the frameshift-mutated GagPol polyprotein expressed in COS-7 cells migrates in SDSPAGE between NC and NCp1. Based on the biochemical experiments and the analysis of the viral proteins cleaved from the polyproteins in bacterial and eukaryotic cells, we claim that during autoprocessing, the HIV-1 GagPol precursor releases an extended protein, NC*. The NC* protein could not be detected in the virions released from HIV-1-infected cells. One explanation may be that our test is not sufficiently sensitive to detect NC*, which is only about 5% of the NC. However, it also possible that NC* is not packed in the virions or that it is degraded before particle assembly. Elongated proteins released from GagPol in the frameshift vicinity were previously described in avian leukaemia sarcoma virus, where the PR is 12 amino acid residues longer than the PR translated in the Gag frame (Arad et al., 1995 ; Sedlacek et al., 1988 ).

The function of NC* is not yet known, nor is it clear whether NC* and NC fulfil the same functions. Computerized prediction of the secondary structure of the NC* protein reveals that the additional 8 amino acid residues elongate an α-helix structure from the Gag into the Pol domain. The α-helix structure provides flexibility and elasticity for the PR-bearing polyprotein. This prediction is in agreement with the model in which the p6^Pol region is employed as a flexible hinge between the Gag and Pol domains (Tessmer & Krausslich, 1998 ; Beissinger et al., 1996 ). Thus, it is plausible that elongation of NC enables the initiation of autoprocessing before the release of PR from GagPol.

In accord with this model, we demonstrate that NC*p6^PolPR is enzymatically active and thus the release of PR is not a prerequisite for the autoprocessing. Previous studies show that PR is enzymatically active as part of the p6^PolPR or PRRT fusion proteins (Kotler et al., 1992 ; Zybarth & Carter, 1995 ; Cherry et al., 1998 ; Tessmer & Krausslich, 1998 ). The enzymatic activity of NC*p6^PolPR could not be compared to that of mature PR, since it was assessed in the truncated GagPol context only. It could be that this fusion protein is less active than the mature enzyme, as previously shown (Zybarth et al., 1994 ; Zybarth & Carter, 1995 ; Louis et al., 1994 , 1999 ; Cherry et al., 1998 ; Tessmer & Krausslich, 1998 ). Since the fusion with NC* does not interfere with the autocleavage of the other viral protein, it is plausible that PR is active as part of the GagPol precursor, enabling autoprocessing which leads to virus maturation.

This work was carried out in the Peter A. Krueger Laboratory, with the generous support of Nancy and Lawrence Glick, and Pat and Marvin Weiss. We thank Roche Products Ltd for donating Ro 31-5989, Professor D. Rekosh (University of Virginia, Myles H. Thaler Center for AIDS Research, Charlottesville, VA, USA) for the pSVGAGPOL-RRE-RFS5T and pCMVrev plasmids, Professor B. Moss (Laboratory of Viral Diseases, NAID, NIH, Bethesda, MD, USA) for the vTF7-3 and Dr S. M. Nigida (SAIC Frederick, Division of Science Applications, International Corporation, Frederick, MD, USA) for supplying the monoclonal anti-MA and CA. This work was supported in part by grant #02730-28-RG from the American Foundation for AIDS Research (amfAR), by the Israel Science Foundation, the Israel Ministry of Health and by the CanadaIsrael Bi-national Foundation.

Footnotes

^b Since deceased.