![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Journal of Virology, May 2008, p. 4573-4584, Vol. 82, No. 9
0022-538X/08/$08.00+0 doi:10.1128/JVI.02353-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.
Institute of Medical Microbiology and Hygiene, Molecular Microbiology and Gene Therapy Unit, University of Regensburg, 93053 Regensburg, Germany
Received 31 October 2007/ Accepted 21 February 2008
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Despite comprehensive analyses of PR-mediated processing events giving rise to infectious virions (40-44, 57), the mechanism stringently controlling spatiotemporal PR activation during the late phase of viral replication is not yet fully understood. Correct folding of the enzyme within the dimerized Gag-Pol precursors is a prerequisite to allow formation of a functional active site (32, 58). As dimerization events become more frequent with Gag-Pol polyproteins accumulating beneath the cytoplasmic membrane during virus assembly, increasing PR activity results in stepwise autoprocessing and subsequent cleavage of the viral precursors while particles are being readied for release (24).
These sequential processing steps driving maturation of Gag and Gag-Pol precursors have largely been clarified (10, 30, 43, 44, 57), and kinetic analyses revealed significantly differing hydrolysis rates for individual PR cleavage events (15, 25, 42, 43). It is widely accepted that cleavage of the Gag precursor is initiated at the carboxyl terminus of the spacer peptide p2, separating the MA-CA-p2 polypeptide from the NC-p1-p6 moiety (39, 43). Subsequently, MA and p6 are cleaved off the corresponding intermediates at an approximately 10-fold-lower rate (43). Finally, the rate-limiting release of the spacer peptides p2 and p1 is catalyzed, allowing CA and NC domains to adopt their final conformation (16, 57). While most cleavages within Gag are mediated by the mature PR in trans, initial processing steps within Gag-Pol are accomplished by the precursor-associated immature enzyme, for which intrinsic activity has repeatedly been demonstrated (26, 30, 46, 61). However, sequential proteolysis of Gag-Pol via clearly defined intermediates remains difficult to determine under ex vivo conditions, as PR-containing precursors are characterized by very short half-lives (28, 52).
Apart from sterical and biochemical processes governing PR activation, we and others have ascribed a role for the transframe domain p6* (p6pol) in the regulation of PR activity (38, 51, 63). Nuclear magnetic resonance analysis of recombinant p6* protein revealed a widely flexible structure (3, 22), and its presence directly upstream of the PR is reminiscent of the propeptides flanking eukaryotic aspartic proteases and has been proven to prevent folding of the PR to its dimeric structure (6). Indeed, we could provide evidence that recombinant p6* protein is a strong competitive inhibitor of PR activity in vitro, and we proposed a model in which the carboxyl-terminal tetrapeptide of p6* following release from the Pol precursor, blocks the substrate binding cleft of the PR, thereby delaying overall processing (37).
Nevertheless, the in vivo function of p6* is still not clearly understood, and our recent data indicate that p6* functional domains are confined to the highly conserved amino- and carboxyl-terminal regions, whereas the sequence context of the entire central part does not appear essential for productive viral replication (8, 38). It is noteworthy though that p6* itself is a substrate of PR and is therefore likely involved in stepwise autoactivation of the enzyme. Whereas functional relevance of amino-terminal p6* cleavage separating the Pol moiety from the NC domain has not yet been demonstrated, there is accumulating evidence that carboxyl-terminal cleavage of p6* is a prerequisite for complete activation of the PR (8, 38, 53). Furthermore, p6* has been shown to contain a conserved internal PR cleavage site between Phe8 and Leu9 (see Fig. 1) hydrolyzed early during precursor maturation (1, 7, 29, 40, 41, 64). Interestingly, natural polymorphisms of the p6* cleavage sites have been reported to impact PR release with considerable effects on susceptibility toward PR inhibitors (56).
|
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
and pGL-fsM1 has been described previously in detail (38). Frameshift modifications of the p6* mutants csM1, csM2, and csM3 were introduced into fs-wt sequence by site-directed mutagenesis using a QuikChange kit and primer pairs fs1, fs2, and fs3 (Table 1). All reporter constructs were verified by sequencing. Synthetic oligopeptides. The chromogenic substrate Lys-Ala-Arg-Val-Nle-Phe (p-NO2)-Glu-Ala-Nle-NH2 was obtained from Bachem Biochemica (Bubendorf, Switzerland), and all unmodified synthetic oligopeptides were purchased from G. J. Arnold (Gene Center, Munich, Germany). The chemical composition and purity of the synthetic oligopeptides were analyzed by electrospray ionization mass spectrometry (17). Concentrations of peptide dilutions were determined by Analytical Research and Services (University of Bern, Switzerland) by quantitative analysis of amino acid composition.
Peptide cleavage assay. Affinity-purified HIV-1 PR for peptide cleavage was purchased from Bachem Biochemica. Oligopeptides were cleaved at concentrations of 140 µM at 25°C in 0.1 M sodium acetate (pH 5.0), 4 mM EDTA, 5 mM dithiothreitol, and 0.3% dimethyl sulfoxide in a PCR cycler with a heated lid to avoid condensation. Reactions were started by adding 450 nM PR dimer and were stopped by adding trifluoroacetic acid at a total concentration of 0.1%. To exclude unspecific cleavage, samples were incubated with 5 µM PR inhibitor RO 31-5989 (49) from Roche (Mannheim, Germany). Reaction products were separated via a linear acetonitrile gradient (0 to 50% in 0.1% trifluoroacetic acid) within 20 min at a flow rate of 0.2 ml/min using a µRPC C2/C18 SC 2.1/10 column (Amersham Pharmacia Biotech) and were detected at 215 nm. Cleavage products were collected and analyzed by electrospray ionization mass spectrometry (17) and gas phase sequencing. Total substrate turnover was calculated using external standards from the integrated peak area.
Cell lines, transfections, and infections. H1299 human lung carcinoma and 293T human kidney epithelial cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 µg/ml streptomycin. The CD4-positive human cervix carcinoma cell line HeLa-CD4-LTR-β-gal (Magi cells) was maintained in Dulbecco's modified Eagle's medium containing 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin supplemented with 0.2 mg/ml G418 and 0.1 mg/ml hygromycin B. The HIV-1 permissive human T-cell lines CEM4 (P. R. Clapham, AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH) and MT-4 (27) were grown in RPMI 1640 supplemented with 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin.
Adherent H1299 and 293T cells were transfected with 30 µg of provirus plasmid DNA or with 3 µg of frameshift reporter constructs using the calcium phosphate precipitation technique, and cells and supernatants were harvested as described in Results. For analysis of replication kinetics, 5 x 106 logarithmically growing CEM4 or MT-4 suspension cells were transfected with 1 µg of provirus plasmid DNA using Fugene 6 from Roche according to the manufacturer's protocol. Samples were collected, and cultures were diluted with 1 volume of fresh medium every 48 h over a total period of 20 to 30 days.
Single-round infection of Magi cells was performed as previously described (38) with serial dilutions of virus-containing supernatants from transfected H1299 cells.
Quantification of HIV-1 capsid antigens. To determine the amount of HIV-1 capsid protein in culture medium, supernatants were clarified of cell debris (300 x g, 10 min), filtered through a 0.45-µm-pore-size filter and used in a CA-specific capture enzyme-linked immunosorbent assay (ELISA). Alternatively, particle-associated antigens were enriched by centrifugation of clarified supernatants through a 20% sucrose cushion. For quantification of CA, 96-well Maxi Sorb microtiter plates (Nunc, Wiesbaden, Germany) were coated overnight (4°C) with a 1:300 dilution (0.1 M carbonate buffer [pH 9.5]) of the CA-specific antibody M01 (Polymun, Vienna, Austria). After three washes (with phosphate-buffered saline [PBS] containing 0.05% Tween), the wells were incubated with the serially diluted samples (PBS containing 1% bovine serum albumin [BSA]) for 1 h at 37°C. Following six washes, the wells were incubated with a 1:20,000 dilution (PBS containing 1% BSA) of the biotinylated CA-specific antibody 37G12 (Polymun) for 1 h at room temperature. After 10 further washes, a 1:10,000 dilution of horseradish peroxidase-conjugated streptavidin (Roche) was added for 30 min at room temperature, and after 10 final washes, the plates were developed with 3,3',5,5'-tetramethylbenzidine (TMB) substrate (Becton Dickinson, Heidelberg, Germany). The reaction was stopped with 1 N H2SO4/well, and plates were measured at 450 nm. CA content was quantified using a calibration curve based on serial dilutions of a CA standard (Polymun).
Western blot analysis. To characterize cell-associated virus proteins, lysates of transfected H1299 cells were prepared and quantified as previously described (38), and 100 µg of total protein was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose. Capsid protein species were visualized with the CA-specific monoclonal antibody (MAb) 16/4/2 (59) and an alkaline phosphatase-labeled anti-mouse antibody (Bio-Rad), and proteins were detected with Nitro Blue Tetrazolium and 5-bromo-4-chloro-3-indolylphosphate (BCIP) solutions from Roche.
For analysis of virion-associated proteins, particles were purified from cell supernatants as previously described (38), and virus proteins were separated by SDS-PAGE and detected by immunoblotting with the CA-specific MAb 13/5 (59), the p17-specific MAb 3-H-7 (33), the polyclonal IN-specific rabbit antiserum 757 (D. P. Grandgenett, NIH AIDS Research and Reference Reagent Program), the polyclonal PR-specific sheep antiserum ARP413 (D. Bailey and M. Page, NIH AIDS Research and Reference Reagent Program), the polyclonal RT-specific antiserum 4F8 (P. Chandra, Gustav-Embden-Center of Biological Chemistry, University of Frankfurt, Germany), and polyclonal rabbit antisera raised against glutathione S-transferase (GST)-p6gag and GST-NC fusion proteins. Proteins were visualized by enhanced chemiluminescence using SuperSignal WestFemto (Pierce, Rockford, IL).
Frameshift luciferase reporter assay. For quantification of cell-associated luciferase activity, transfected cells were harvested and lysed as described previously (38), and 100 µg of total protein was used in a luciferase assay from Promega according to the manufacturer's protocol. Luciferase activity was measured by recording light emission for 10 s in a Lumat 9501 luminometer (Berthold, Bad Wildbach, Germany).
Growth competition assay. Supernatants of 293T cells transfected with provirus plasmid DNA were harvested 48 h later, and the amounts of CA in the supernatants were quantified by ELISAs. Equal CA amounts (500 ng) of wild-type (wt) and mutant particles were mixed at ratios of 1:1, 4:1, and 1:4 and used to infect 2 x 106 CEM cells. After 6 h, infected cultures were transferred to flasks and cultivated for 3 weeks in a total volume of 4 ml. Every 2 days, samples were collected, and half of the cultures were replaced with fresh cells. Infected cells were pelleted (300 x g, 10 min), and genomic DNA was extracted using the QIAamp DNA Mini kit (Qiagen). DNA was PCR amplified with primers S6f and S7r (Table 1). Purified PCR products were then subjected to DNA sequencing with primer S7r. Chromatograms were analyzed using Vector NTI 10.3.0 software from Invitrogen (Karlsruhe, Germany).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To screen for cleavage site variants capable of (i) completely blocking cleavage or (ii) altering hydrolysis rates, an algorithm was used to predict the probability for any sequence of eight amino acids to be cleaved by the HIV-1 PR (9). Those residues predicted to either block cleavage or alter cleavage kinetics according to the algorithm were introduced into the infectious provirus HX10. To block cleavage of the amino-terminal scissile bond (PN) of p6*, Arg in the P3' position was replaced by an aromatic Trp residue (54), yielding mutant csM1 (Fig. 1A). This mutation was associated with a nucleotide substitution turning the conserved slippery site UUUUUUA into UUUUUUU (Fig. 1B). However, this sequence has previously been shown to promote frameshifting in vitro (12). The amino-terminal p6* mutation also resulted in a Leu-to-Phe substitution at the P2' position of the overlapping NC-p1 cleavage site (60), which according to the employed algorithm (9) was not expected to prevent cleavability of this modified site.
On the basis of the previous findings that the introduction of a β-branched amino acid at position P1 prevents cleavage by the PR (45, 54), Phe8 was replaced by Val to block internal cleavage (PI) of p6* (Fig. 1B). As the corresponding csM2 mutation reduced base pairing in the upper stem region, possibly affecting overall stem-loop stability (4, 35), a second mutant (csM3) was generated in which base pairing was restored (Fig. 1B). This resulted in a Lys-to-Thr substitution within a variable p6* portion without altering the gag open reading frame.
Since carboxyl-terminal cleavage of p6* from the Pol precursor is an essential prerequisite for PR autorelease and activation (37, 53), we further substituted the conserved carboxyl-terminal tetrapeptide of p6* by four different amino acid combinations predicted to be cleaved according to the criteria of Chou et al. (9) to analyze potential effects of altered hydrolysis rates on subsequent virus maturation. The modified tetrapeptides encoded by the resulting virus mutants contained either two (csM6), four (csM4 and csM5), or five (csM7) amino acid substitutions compared to wt p6* (Fig. 1A). Again the gag reading frame remained intact, as overlapping p6gag residues have been proposed to be involved in packaging of envelope proteins (34).
In vitro characterization of peptides with cleavage site variations.
In order to determine the influence of cleavage site modifications on hydrolysis rates in vitro, synthetic oligopeptides comprising wt cleavage site sequences (PN, PI, and PC) or mutated cleavage site sequences (P1 to P7, corresponding to virus mutants csM1 to csM7, respectively) were incubated with recombinant HIV-1 PR for at least 24 h to guarantee quantitative turnover of the peptide substrates. To confirm that cleavage products were originating from PR activity, parallel samples were incubated in the presence of an HIV-1 PR-specific inhibitor. Cleavage products were separated by reverse-phase high-performance liquid chromatography and analyzed by mass spectrometry and sequencing (data not shown). As expected, cleavage of the modified amino-terminal cleavage site (PN) was efficiently blocked in peptide P1. However, this substrate was cleaved two amino acids downstream of the wt site, instead (Table 2). Likewise, the original internal cleavage site (PI) was blocked by the mutations in P2 and P3 and replaced by a new scissile bond emerging two amino acids upstream. Consistently, altered internal cleavage was also observed when GST-p6* full-length proteins harboring the corresponding P2 and P3 mutations were processed in trans (data not shown), excluding the possibility that the observed cleavage site modifications were restricted to the short oligopeptide substrates. Deviating from predictions by the algorithm ( values in Table 2), only two of the four carboxyl-terminal cleavage site (PC) variations were processed by the PR in substrates P5 and P6, whereas cleavage of this site was completely blocked in P4 and P7.
|
|
was generated allowing for constitutive full-length luciferase expression from the –1 frame (corresponding to a frameshift rate of 100%). In addition, we used the formerly described frameshift mutant fsM1 as a negative control to detect luciferase background activity (38). This mutant harbors a destroyed slippery site driving expression of luciferase exclusively from the 0 frame (corresponding to a 0% frameshift rate), which results in premature translation termination. To analyze frameshift modifications, H1299 cells were transfected with the corresponding reporter plasmids, and luciferase activity was quantified after 48 h and related to values determined for fs. As shown in Fig. 3B, frameshift-dependent luciferase expression from the reporter constructs fs-wt, fs-csM1, fs-csM2, and fs-csM3 ranged between 
4 and 8% and was therefore clearly distinguishable from fsM1-transfected cells, which gave luciferase activities around the mock-transfected level. Basically, all tested mutants were capable of driving frameshift-dependent luciferase expression. However, taking into account the limited sensitivity of this assay system, a slight increase in frameshifting could be observed for the stem-loop variants fs-csM2 (8.2%) and fs-csM3 (7.2%) in H1299 cells, whereas luciferase activity derived from expression of fs-csM1 (4.3%) was clearly diminished compared to fs-wt (6.9%). Similar frameshift ratios were obtained in multiple transfection experiments performed with different DNA preparations, basically excluding the possibility that differential luciferase expression resulted from variations in transfection efficiencies. Furthermore, comparable results were obtained in 293T cells (data not shown), indicating that the altered heptamer sequence in fs-csM1 was less efficient in promoting ribosomal slippage.
|
|
48-kDa product detected with the CA-specific antibody in csM4 and csM7 particles and which might thus represent a MA-CA-p2-NC-p1 intermediate (Fig. 4B, panel 4). The presence of a faint double band further suggests that the spacer peptide p1 has been partially released from this aberrant precursor. Whereas csM2, csM3, csM5, and csM6 mutants showed wt-like processing of NC proteins, csM1 virions contained additional NC-specific bands that were not detected in any other particle preparation. Since the csM1 mutation is associated with an amino acid substitution in the p1gag domain, the prominent 8-kDa band found in csM1 particles might represent a NC-p1 intermediate, suggesting retarded hydrolysis of the corresponding cleavage site. However, the identities of two additional NC-specific products of 16 to 20 kDa remain unknown, as these proteins were not recognized by any other antibody used in this study. Immunoblot analysis of csM1 particles with a p6gag-specific antibody (Fig. 4B, panel 5) did not reveal any differences compared to wt virions. Hence, the csM1-associated processing defect appears to be limited specifically to the NC protein.
To further examine packaging and maturation of Pol proteins, particle preparations were analyzed with IN-, RT-, and PR-specific antibodies. As shown in Fig. 4C, none of the p6* mutations interfered with packaging of Pol proteins into virus particles. However, processing patterns revealed differing levels of maturation. Along with the mature IN and RT species, we could clearly detect traces of the 98-kDa RT-IN precursor in csM4 and csM7 particles (Fig. 4C, panels 1 and 2), whereas all other virions contained exclusively the mature enzymes. Finally, a PR-specific antibody failed to identify the mature 11-kDa PR in csM4 and csM7 particles (Fig. 4C, panel 3), where a protein of approximately 17 kDa was detected instead. This product likely represents a p6*-PR precursor as verified with a p6*-specific antiserum (data not shown). These results strongly indicate that the carboxyl-terminal csM4 and csM7 mutations completely block cleavage between p6* and PR. Although this processing defect clearly affected maturation of the Gag precursor, quantitative cleavage of the Pol components was influenced to a minor extent. Interestingly, in addition to the mature PR, trace amounts of this p6*-PR intermediate were also observed in csM5 and csM6 particles. These data suggest that amino-terminal autorelease of the PR was not completely impaired but rather delayed in csM5 and csM6 viruses, with csM5-specific mutations causing the most striking effect. This observation is in accordance with the results obtained by in vitro cleavage of the corresponding peptides (Fig. 2).
In contrast, the amino-terminal (csM1) and internal (csM2 and csM3) p6* mutations did not visibly influence the release of the viral PR or otherwise affect the maturation of virus-associated Pol proteins yielding only the mature IN, RT, and PR species. Since we could not detect any aberrant p6*-specific cleavage intermediates in those particles probably owing to the narrow time frame in which p6* cleavages occur, these results make it tempting to speculate that the modified cleavage sites of corresponding peptides P1, P2, and P3 were either not used in the viral precursors or did not impact on Pol-specific maturation.
Influence of p6* cleavage site mutations on viral replication and infectivity. Next, the influence of p6* mutations on virus growth in permissive cells was studied following transfection of CD4-positive MT-4 lymphocytes with the various provirus clones. Virus production was followed over a period of 3 to 4 weeks by measuring the amount of total CA protein in culture supernatants. As depicted in Fig. 5A, wt viruses replicated to high titers in MT-4 cells peaking at days 10 to 12 posttransfection and followed by a rapid decline as cell populations died. p6* mutants csM2, csM3, csM5, and csM6 showed similar replication kinetics (csM3- and csM5-specific replication curves have been omitted for clarity), indicating that none of the underlying internal or carboxyl-terminal mutations had a major impact on viral growth in cultured cells. In contrast, replication of the csM1 mutant was significantly delayed and reached lower peak titers compared to wt viruses. This retarded phenotype was also observed in a second T-cell line (data not shown), confirming the above results. A reversion of the amino acid substitution to the wt or the emergence of compensatory mutations within the monitored period was widely excluded by PCR sequencing of the entire Gag-p6* region and part of the PR, suggesting that the delayed replication profile of these viruses was rather due to compromised frameshifting (Fig. 3B) and/or retarded NC-p1 cleavage (Fig. 4B, panel 4). As expected from deficient Gag processing, no productive replication of the csM4 (data not shown) and csM7 mutants in the analyzed lymphocytes was detectable within the monitored period.
|
Modifications in csM5 and csM6 mutants result in a loss of viral fitness when competing with wt virus. Although carboxyl-terminal csM5 and csM6 mutations showed significantly altered hydrolysis rates of the corresponding cleavage site in vitro (Fig. 2), resulting in a retarded autorelease of the PR (Fig. 4), we could not observe any severe effects of these mutations on viral replication or infectivity in the cell lines examined. This prompted us to further assess viral fitness of these candidates when challenged with wt virus. Therefore, a dual competition assay was designed, where CEM4 cells were infected by wt viruses mixed at different ratios with mutant csM5, csM6, or csM7. Spread of virus infection was monitored via quantification of integrated virus cDNA. Overlapping peaks at the modified p6* positions of the resulting chromatograms were then used to determine the relative amounts of wt versus mutant sequences in a sample. Taking into account the fact that peak heights were dependent on the base due to different kinetics in base incorporation during the sequencing reaction, the wt to mutant inoculation ratios of 1:1 or 1:4 were nicely reflected by the corresponding peak heights obtained at day 1 postinfection (see for example csM5 versus wt infection in Fig. 6). Surprisingly, csM7-specific peaks were found 1 day following infection, suggesting that csM7-derived cDNA was at least partially capable of integrating into target cell genomes. This observation is in accordance with the faint infection of Magi cells by csM7 viruses (Fig. 5B). However, no csM7-specific peaks could be detected beyond 5 days postinfection confirming the above experiments in lymphocytes, where no productive replication of this mutant was visible (Fig. 5A). To more precisely quantify the loss of viral fitness, ratios of wt versus mutant peak heights were calculated for each individual time point (Fig. 6). Accordingly, slopes of the resulting curves are indicative of the rate at which mutants are overgrown by wt viruses with the fastest shift determined for mutant csM7 (Fig. 6C). It is interesting, however, that at day 5 postinfection with 1:1 mixtures, there was a also a significant reduction of csM5- and csM6-specific peaks compared to wt peaks, indicating that wt viruses were more successful in the following infection rounds (Fig. 6A and B). Twenty-one days postinfection, wt virus was the dominant species in all samples even when mutant viruses were inoculated at a fourfold excess. These data clearly indicate that carboxyl-terminal p6* modifications in mutants csM5 and csM6, albeit allowing for amino-terminal release of the PR, result in a loss of viral fitness that becomes apparent when competing with wt virus.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mutagenesis of p6* in the context of an infectious provirus appears challenging, as parts of the pol reading frame are superimposed by cis-active RNA frameshift signals and functional Gag protein domains. Thus, giving priority to frameshift function, only a single mutation predicted to affect amino-terminal cleavage of p6* appeared reasonable without completely destroying the slippery site. Albeit we could prove functionality of this modified slippery site in a cell culture-based frameshift reporter assay, frameshift rates were clearly reduced compared to the wt level, confirming recent observations by others (14). Further analysis of this modification in the csM1 virus mutant revealed a clearly delayed replication profile in cultured lymphocytes associated with a 50% loss of infectivity in single-round replication. Furthermore, we observed NC-associated processing defects in csM1 particles likely associated with a substitution in the NC-p1 cleavage site introduced by amino-terminal p6* mutation. Albeit hydrolysis of the NC-p1 site has been reported to be rate limiting in virus maturation (16, 62), recent analyses by Coren et al. (10) have provided evidence that NC-p1 cleavage is not required for viral replication and infectivity. Together, these findings suggest that the aberrant csM1 phenotype is mainly due to compromised frameshifting at the altered slippery site.
Surprisingly, the modified amino-terminal p6* cleavage site, although resulting in a 240-fold-increased hydrolysis rate in vitro, did not influence the order and rate of PR-mediated precursor maturation. In contrast to our in vitro analyses, Chen et al. (7) did not observe amino-terminal cleavage of p6* when deploying a similar peptide cleavage assay and hypothesized that NC released from the Gag-Pol precursor might be extended by eight residues of p6*. However, our in vitro cleavage experiments have shown that substrate turnover was not completed until 24 h of incubation, suggesting that an observation period of 60 min selected by Chen et al. might not be sufficient to draw this conclusion. Furthermore, peptides representing the amino-terminal octapeptide of p6* were formerly isolated from HIV-1 virions (19), and mutations in the NC-p6* cleavage site have been implicated in compensating for the loss of viral fitness in PR inhibitor-resistant mutants (11), which clearly points to a biological function of amino-terminal p6* cleavage.
As internal cleavage of p6* between Phe8 and Leu9 is an early event in the Gag-Pol processing cascade (1, 40, 53), modification of this site in the viral context appeared particularly informative. However, the mutation introduced to block this cleavage site resulted in a shift of the scissile bond associated with a 5- to 10-fold-decreased hydrolysis rate. Surprisingly, this mutation did not influence correct processing of the viral precursor proteins or replication of csM2 and csM3 viruses in permissive T lymphocytes. Further confirming our observations, a comparable mutation did not impact viral infectivity, when analyzed in the context of stem-loop stability studies (21). These results indicate that the original cleavage site might be functionally replaced by the novel proximal site which has also been used in full-length GST-p6* proteins carrying the respective mutation. In a current model, the internal cleavage site is processed intramolecularly by the precursor-embedded PR which exhibits a substrate specificity different than that of the mature free enzyme (30, 40, 41, 61). Taking into account the fact that in vitro peptide cleavage was performed by a mature PR dimer in trans, we cannot completely exclude the possibility that the internal mutation might have different effects in the context of the Gag-Pol precursor.
Although carboxyl-terminal mutagenesis of p6* was carefully designed, cleavability predicted by the algorithm used (9) was achieved only for the amino acid combinations in P5 and P6. In contrast, cleavage was completely blocked in peptides P4 and P7 and in the respective virus mutants csM4 and csM7. Corroborating previous work by others and us, completely blocking carboxyl-terminal p6* cleavage resulted in immature viruses with impaired infectivity and severe Gag processing defects (38, 53). Interestingly, the amino-terminally extended PR species found in the corresponding virions was not capable of processing the MA-CA scissile bond, indicating that this cleavage can be carried out only by the mature enzyme. Even though the amino-terminal PR extension is likely to restrict accessibility of the Gag substrates within the condensed viral shell (53), similar observations have been made with uncondensed, in vitro-translated Gag-Pol and Gag precursors, rather suggesting that the precursor-embedded PR exhibits a cleavage site affinity or substrate specificity different from that of the mature enzyme (40).
In contrast, those carboxyl-terminal p6* modifications allowing peptide cleavage in vitro (P5 and P6) at 15- and 1.5-fold-lower hydrolysis rates than the wt site, which was reflected perfectly by the retarded release of the PR in csM5 and csM6 virions, did not result in Gag-specific processing defects. As traces of a PR-specific 17-kDa precursor were detected in those particles along with the mature 11-kDa species, the 17-kDa protein likely represents the very last PR intermediate. This implies that internal hydrolysis of p6* precedes carboxyl-terminal cleavage in virus particles, albeit the carboxyl-terminal site was processed at a 60-fold-higher hydrolysis rate in the in vitro cleavage assay. Hence, our data are concordant with a recently proposed model where p6* cleavage is initiated in cis at an internal site (Pi) that appears to be best accessible to the active site of the precursor-embedded PR (40, 41). However, as mature RT and IN species together with an RT-IN intermediate were present in csM4 and csM7 virions where amino-terminal cleavage of the PR was blocked, our results also indicate that amino-terminal release of the PR is not a prerequisite for removal of the carboxyl-terminal RT-IN moiety.
Whereas blockage of PR release severely affected maturation and infectivity of csM4 and csM7 particles with a potential impact on the natural processing cascade, the retarded PR release in mutants csM5 and csM6 did not visibly interfere with virus infectivity or replication on cultured cells, suggesting that altered cleavage rates at the PR amino terminus are well tolerated. These findings are concordant with former reports that viral infectivity is not significantly impaired unless PR activity is reduced by more than fourfold (50). Nevertheless, these mutants were rapidly overgrown by wt viruses in a more sensitive dual competition assay. It is noteworthy that the csM5 mutant, albeit showing a significantly delayed amino-terminal release of the PR, performed better in this assay than csM6, for which PR release was scarcely affected. Interestingly, csM6-specific p6* residues seemed to cause product inhibition in the in vitro cleavage assay, suggesting that compromised growth of the csM6 mutant in the competition assay might be due at least partially to secondary effects exerted by the released carboxyl-terminal p6* tetrapeptide (37). Additional support for this assumption has been given by analysis of purified p6* proteins an in vitro PR inhibition assay, where a p6* variant carrying the csM6-specific carboxyl terminus showed a slightly stronger inhibitory capacity than the wt protein did (unpublished data).
In sum, this study has provided further evidence that correct processing of p6* is essential to drive temporal and stepwise activation of the PR. However, our results have also shown that individual cleavage site variations of p6* did not interfere with viral replication in cultured cells as long as overall processing of the viral Gag and Pol precursors was guaranteed.
|
ACKNOWLEDGMENTS |
|---|
Part of the work was financed by DFG grant Wa873/1-4.
|
FOOTNOTES |
|---|
Published ahead of print on 5 March 2008.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
