VIROLOGY 219, 29–36 (1996) ARTICLE NO. 0219

Extensive Regions of pol Are Required for Efficient Human Immunodeficiency View metadata, citation and similar papersPolyprotein at core.ac.uk Processing and Particle Maturation brought to you by CORE provided by Elsevier - Publisher Connector CAROLINE QUILLENT, ANDREW M. BORMAN,1 SYLVIE PAULOUS, CHARLES DAUGUET, and FRANC¸ OIS CLAVEL2

Unite´ d’Oncologie Virale, CNRS URA 1157, De´partement SIDA et Re´trovirus, Institut Pasteur, Paris, France Received November 15, 1995; accepted February 9, 1996

Human Immunodeficiency type 1 particle maturation is dependent upon proteolytic cleavage of the gag and gag-pol precursors by the pol-encoded viral . We have investigated the importance of domains of pol other than the protease for particle maturation and gag proteolytic processing. Truncations of the gag-pol polyprotein precursor of HIV-1 were created by deleting segments of the coding region or by introducing stop codons in the region of an HIV-1 infectious molecular clone. In these mutants, the protease coding sequence was left intact. Particles produced by all of the mutants displayed abnormal morphologies and impaired proteolytic processing of gag. The severity of particle morphology abnormalities and of gag polyprotein processing impairment appeared to be affected both by the size and by the position of the deletions in pol, suggesting that the integrity of several pol domains within the gag-pol precursor is required for optimal protease activation and particle maturation. Additionally, cotransfection of a deletion mutant with wild-type provirus led to a marked reduction in the titer of infectious virus, suggesting that truncated gag-pol precursors can interfere with wild-type virus assembly and maturation. ᭧ 1996 Academic Press, Inc.

INTRODUCTION a distended, spherical structure, identical to that resulting from the expression of Pr55gag alone, rather The production of infectious human immunodeficiency than the condensed cone-shaped electron-dense core virus type 1 (HIV-1) particles is an ordered process that that is typical of mature particles (Gelderblom et al., 1987; involves sequential macromolecular interaction and pro- Kaplan et al., 1993; Kohl et al., 1988). It has been sug- cessing events. Initial assembly at the cell surface, bud- gested that only full gag-pol expression can promote ding, and release of viral particles can be obtained by normal particle maturation (Ross et al., 1991): when a expression of the Pr55gag precursor polyprotein alone truncated recombinant consisting only of gag-protease (Gheysen et al., 1989; Royer et al., 1992; Shioda and is expressed, maturation occurs only in a fraction of the Shibuta, 1990). Nevertheless, a subsequent maturation resultant particles, with a marked reduction in Pr55gag step, which leads to organization of the final structural processing (Gheysen et al., 1989; Ross et al., 1991). components of the particle into a functional nucleopro- Similarly, we have described the effects of a frameshift tein replicative complex, occurs shortly after budding and mutation in the reverse transcriptase (RT) region of the requires proteolytic cleavage of the 55-kDa gag precur- HIV-1 genome on particle morphology. These mutant par- sor by the viral protease (Crawford and Goff, 1985; Gott- ticles, which possessed protease but not RT or integrase, linger et al., 1989; Kohl et al., 1988; Lœb et al., 1989). The had a typically immature morphology lacking the con- protease is part of the larger gag-pol precursor whose densed nucleoid (Quillent et al., 1993). Spontaneous re- synthesis results from an infrequent ribosomal frame- version of this mutation in RT resulted in the production shifting event near the gag-pol junction (Jacks et al., of mature viral particles. These results and those of Ross 1988). HIV protease activity is dependent on dimerization et al. (1991) suggested that the requirements for efficient of monomeric subunits present on gag-pol polyproteins protease activation and complete maturation of the viral (Pearl and Taylor, 1987). It is presumably this require- particle were not confined to the gag-protease region ment for dimerization which prevents early protease acti- alone. Here, we set out to further examine the possible vation and polyprotein cleavage prior to virus budding involvement of more distal regions of pol in viral assem- (Krausslich and Wimmer, 1988). When the protease is bly, maturation, and proteolytic activity. Toward this aim, mutated, or in the presence of HIV protease inhibitors, a series of deletion mutations were constructed to dis- noninfectious viral particles are produced which display rupt regions of pol while leaving protease intact. We found that Pr55gag processing and subsequent particle 1 Present address: Unite´ de Virologie Mole´culaire (CNRS URA 1966), maturation was proportionally affected by the reduction Institut Pasteur. in size of the gag-pol precursor, suggesting that several 2 To whom correspondence and reprint requests should be ad- domains in the RT and integrase are required for optimal dressed. Fax: (33) 1 40 61 34 65. E-mail: [email protected]. activity of the viral protease.

0042-6822/96 $18.00 29 Copyright ᭧ 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology 30 QUILLENT ET AL.

MATERIALS AND METHODS (Amersham). Quantification of the revealed viral was performed using NIH Image 1.54. DNA constructs The infectious HIV-1 proviral molecular clone pNL43 Electron microscopy of WT and mutant viral particles (Adachi et al., 1986) served as the parental control for Forty-eight hours after transfection, HeLa cells were all experiments. For the construction of plasmid pNLbb, washed twice with phosphate-buffered saline, removed plasmid pNL43 was digested to completion with MscI, from culture dishes by scraping, pelleted by low-speed the large fragment was purified and the ends were centrifugation, and prepared for transmission electron closed by ligation. For the pNLRT3 and pNLdK mutations, microscopy as described previously (Quillent et al., 1993). the MscI–MscI small fragment from pNL43 was sub- cloned into a modified Bluescript vector which had been Interference experiments deleted for sequences (including the polylinker) between the two PvuII sites and has an inserted single MscI site. The potential of mutant gag-pol products to interfere The resulting plasmid containing the MscI fragment was with maturation of wild-type particles was evaluated in then digested either with PvuII (for pNLRT3) or KpnI (for cotransfection experiments. HeLa cells were transfected pNLdK), and each of the large fragments was purified following the calcium-phosphate coprecipitation tech- and religated. The resulting truncated MscI fragments nique with different proportions of wild-type and mutant were then excised from the Bluescript vector and were plasmid DNA. The resulting particle stocks were har- reinserted into a pNLbb vector digested with MscI. Prote- vested 48 hr after transfection, measured for their HIV-1 ase mutant pNLdBcl was generated by BclI digestion p24 content on the DuPont ELISA test. These stocks were of pNL43 grown in Dam- Escherichia coli, filling of the subsequently evaluated for their HIV content by cohesive ends with Klenow polymerase, and blunt end Western blotting as described and for their infectivity on ligation. Mutant pNLint (Clavel et al., 1989) was created the HeLa-CD4-LTR-LacZ P4 indicator cell line. The P4 by oligonucleotide-directed mutagenesis to introduce indicator cells were infected with equal volumes of cul- two tandem stop codons at residues 14 and 15 of the ture supernatant in the presence of 20 mg/ml DEAE dex- integrase coding region. tran for 24 hr. Cells were then washed, fixed, and stained with X-gal as described previously (Charneau et al., Cells 1994). Infectivity was scored as the average number of HeLa cells were maintained as subconfluent mono- blue cells per microscopic field. In separate experiments, layers in Dulbecco’s MEM (DMEM) with 10% fetal calf CEMx174 cells were transfected by the DEAE-dextran serum. The P4 indicator cell line (HeLa CD4 LTR-Lac Z) method (Yu et al., 1992) with 3 mg of either pNL43 or were cultured in DMEM with 10% fetal calf serum and pNLRT3 plasmids or with 3 mg of both of the plasmids. 500 mg/ml G418 as described previously (Charneau et Following transfection, the cultures were monitored for al., 1994). H9 and CEM-x174 cells were maintained in virus production by p24 ELISA. RPMI-1640 supplemented with 10% fetal calf serum. RESULTS Western analysis of viral proteins Figure 1 shows a schematic representation of the dif- The mutant and pNL43 were introduced into ferent mutants used in this study. Mutant pNLdBcl intro- HeLa cells by transfection using the calcium phosphate duces a frameshift at amino acid residue 60 of the prote- coprecipitation technique. Mock-transfected control cul- ase and can neither synthesize a functional protease nor tures were analyzed alongside in all experiments. Viral any other pol-encoded product. Plasmid pNLbb bears a particle-associated proteins were prepared from filtered 1932-bp deletion in the pol , retaining the original cell-free supernatants harvested 48 hr after transfection reading frame and sparing only the first 24 amino acids by ultracentrifugation at 100,000 rpm (TL100, Beckman) of RT and the last 179 amino acids of the integrase. for 10 min. Viral pellets were resuspended in PBS and Plasmid pNLRT3 also retained the pol reading frame quantified for HIV p24 antigen by ELISA. Equal quantities downstream of the 1044-bp deletion, which spared the of p24-containing virus lysates (as measured by HIV p24 first 253 amino acids of RT and only removed the first 38 ELISA) were subjected to SDS–polyacrylamide gel elec- amino acids of integrase. Mutant pNLdK bears a 328-bp trophoresis using 10% gels. After semi-dry transfer to deletion in the RNase H domain of RT, retaining the first Hybond C super (Amersham), the proteins were exposed 427 amino acids of RT. Since this deletion is not in phase to a mixture of several gag-specific mouse monoclonal with the pol open reading frame, the end of RT and the antibodies, provided by Hybridolab, Institut Pasteur. Blots integrase cannot be synthesized in this mutant virus. Fi- were developed by enhanced chemi-luminescence nally, mutant pNLint, with two stop codons at the start (ECLWestern Blotting, Amersham), after incubation with of the integrase coding region, produces RT but no inte- horseradish peroxidase-conjugated anti-mouse antibody grase. The length in amino acids of the gag-pol precur-

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology pol REGIONS REQUIRED FOR HIV PROCESSING AND MATURATION 31

FIG. 1. Schematic map of pNL43 and of the five mutant proviruses. For each proviral construct, the boundaries of the deletions are presented together with the corresponding restriction sites, indicated by arrows. Under each proviral diagram, the different open reading frames are indicated as overlapping rectangles. Shaded rectangles represent regions of gag-pol that are not affected by the mutations and that are part of the resulting precursor gag-pol protein. Broken line open rectangles are regions of the gag-pol precursor that are removed as a consequence of the mutations. The details of the construction procedures is presented under Materials and Methods. The sizes in amino-acids of the truncated gag-pol precursors are indicated in the results section. sors from the different studied here should there- 2F. This type of mature particle was only seen in cultures fore be 1435 for wild-type pNL43, 547 for pNLdBcl, 792 transfected with the wild-type pNL43 plasmid. Fully im- for pNLbb, 1087 for pNLRT3, 1012 for pNLdK, and 1160 mature particles were characterized by either a closed for pNLint. Apart from pNLdBcl, none of the mutations ring or an open ring morphology, with a layer of dense affected the protease coding sequence itself. material located under the viral membrane and no sign The different proviral constructs were introduced into of nucleoid condensation at the center of the particle. HeLa cells by calcium-phosphate precipitation transfec- Although found in most of our viruses, these morpholo- tion, and the viral particles produced by transfected cells gies were the only types oberved in the pNLdBcl (not were examined by transmission electron microscopy. shown), pNLbb, and pNLRT3 transfection cultures (Figs. The results of particle morphology analysis for the differ- 2A and 2B). The proportion of these particle types was ent mutants are shown on Fig. 2 and on Table 1. Particles also significantly increased in the pNLdK and pNLint cul- were considered mature when displaying a characteris- tures. Abnormal particles with an empty or an apparently tic electron-dense, conical nucleoid, as shown on Fig. incomplete nucleoid were also observed; in some of

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology 32 QUILLENT ET AL. (C and (A, B, and D), 40,000 1 Negative staining transmission electron micrographs of HIV-1 particles produced after transfection pol. (F). 1 Virion morphology is affected by deletions in FIG. 2. of HeLa cell monolayers with pNL43 (F), pNLbb (A), pNLRT3 (B), pNLdK (C), or pNLint (D and E). Magnifications were 20,000 1 E), and 50,000

AID VY 7837 / 6a14$$7837 04-02-96 13:57:05 vira AP: Virology pol REGIONS REQUIRED FOR HIV PROCESSING AND MATURATION 33

TABLE 1 Quantification of the Effects of pol Deletions on Virion Morphology

Mature Immature

Condensed, Excentered Empty nucleoid, cone-shaped dense material, with little or no nucleoid empty nucleoid dense material Closed ring Open ring pNL4-3 75% 7.5% 0% 10% 7.5% pNLbcl 0% 0% 0% 100% 0% pNLbb 0% 0% 0% 64.5% 35.5% pNLrt3 0% 0% 0% 65% 35% pNLdk 0% 25% 30.5% 41.7% 2.8% pNLint 0% 38.9% 16.7% 38.9% 5.5%

Note. HIV-I particles produced by HeLa cells transfected with the indicated proviral constructs were examined by transmission electron microscopy. A total of 100 to 150 particles were examined in each transfection. Mature particles with a condensed cone-shaped nucleoid were as presented on Fig. 2F. Immature particles were separated in four categories according to their morphology: particles with excentered dense material and an empty nucleoid were as presented on Fig. 2E, or as the bottom particle on Fig. 2C; particles with an empty nucleoid but no observable dense material were as shown on Fig. 2C, top two particles; immature closed ring particles were as the top three particles of Fig. 2B; immature open ring particles were similar to the bottom particle of Fig. 2B. them, the dense material normally found within the the p24:p55 ratio shown on Fig. 3B, derived from the nucleoid was found as a condensed, excentrically lo- densitometric quantification of the corresponding bands cated mass under the surface of the particle. This mor- on the blot shown on Fig. 3A. In pNLdBcl, virtually no phology was found to be mostly typical of pNLint (Figs. cleavage of Pr55gag could be detected. In pNLbb and 2D and 2E), although it was also found in the pNLdK pNLRT3 virions, the majority of the gag-related proteins cultures (Fig. 2C, bottom) and to some extent in the wild- remained as unprocessed Pr55gag or as the p41gag pro- type cultures. In some other particles, there appeared to cessing intermediate. In the case of pNLdK , the pro- be no trace of condensed material in addition to the cessing defect was intermediate, and for the pNLint virus, empty or incomplete nucleoid (Fig. 2C, top). This mor- which appears to be closest to attaining a mature particle phology, which could represent an intermediate between morphology, protein processing was found almost com- the excentered type and the fully immature, ring-shaped plete, with a yield of mature p24 approaching parental types, was mostly observed in the pNLdK cultures. Over- levels and only slightly more p41gag intermediate than all, there appeared to be a gradation in the severity of the pNL43 virions. These results revealed that deletions the alterations in particle morphology within the studied in the HIV-1 gag-pol precursor result in a defect in gag mutants: particles produced by pNLdBcl, pNLbb, and processing that appeared approximately proportional to pNLRT3 were fully immature; pNLint ad pNLdK particles the extent of the deletion, suggesting a proportional im- were a mixture of immature and of aberrant morphology pairment of the function of the viral protease. Indeed, the types, with always a more severe extent of alterations in gag protein profile of the pNLbb and pNLRT3 mutants pNLdK particles. was indistinguishable from those exhibited by particles It has been reported that morphological defects, very produced by cells treated with high doses of A-77003 similar to those observed here on the basis of transmis- (Abbott), an inhibitor of the HIV protease (data not sion electron microscopy, can be induced by treatment shown). Although the total amount of gag antigen as of HIV-infected cultures with an inhibitor of the HIV-1 measured by ELISA was identical in all six samples ana- protease (Kaplan et al., 1993). Thus, we examined the lyzed on the western blot presented in Fig. 3A, it ap- profiles of particle-associated viral proteins in order to peared that significantly more gag protein was actually analyze any possible gag processing defects for these analyzed for the pNLRT3 and pNLbb virions. This finding mutants. Viral particles produced after transfection with confirms previous observations that p24, when present pNL43 showed essentially completely processed protein as part of incompletely cleaved gag precursors, is ineffi- profiles (Fig. 3A). Although a very small proportion of ciently recognized by some p24-specific monoclonal anti- unprocessed p55gag remained associated with these bodies in elisa (Sarubbi and Denaro, 1993). particles, the vast majority of the polyprotein was cleaved Our findings that the function of protease is affected to mature p24. All of the mutant viruses showed disrupted by deletions in pol that spare the protease domain itself cleavage profiles upon Western analysis (Fig. 3A). The prompted us to evaluate whether the resulting truncated processing defects in the mutants can be ordered in precursors could exert a trans-inhibitory effect on wild- terms of severity, as can best be seen on the plot of type virus gag processing. When CEMx174 cells were

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology 34 QUILLENT ET AL.

DNA was increased. We also wished to avoid measuring any effect on infectivity that would result from the dilution of wild-type virus by mutant particles. Therefore, the in- trinsic infectivity of the total particle mix produced by each transfected culture was measured after infection of the target cells by equal volumes of supernatant rather than after standardization based on equal amounts of p24. As shown on Fig. 4B, the P4 titer of viruses from the different cotransfection cultures was found inversely

FIG. 3. Proteolytic cleavage of gag precursors is impaired by dele- tions in pol. (A) Western-blot analysis of the gag proteins recovered in mutant particles. Particle-associated proteins (standardized to 20 ng of p24) produced after transfection of HeLa cells with the provirus denoted above each lane were analyzed by SDS–polyacrylamide gel electrophoresis (10% acrylamide) and subjected to Western blot analy- sis as described in the text. Samples from mock-transfected HeLa cells were analyzed in parallel (Mock lane).The sizes of the HIV-specific proteins are given in kDa on the right-hand side of the panel. (B) Quantification of the gag processing defect in the mutants. The different protein bands on the autoradiogram on A were quantified using NIH image 1.54. The ratio of the p24 to p55 signals was calculated and is presented for each of the mutants studied. FIG. 4. Mutant truncated gag-pol precursors can interfere with the infectivity of wild-type HIV-1. (A) Effect of cotransfection of pNLRT3 on the infectivity of pNL43 in CEM cells. Three micrograms of the indicated transfected by the DEAE-dextran method with equal pro- proviral constructs were transfected into CEM cells by the DEAE-dex- portions of pNL43 and pNLRT3, a strong inhibition of tran method, and production of HIV-1 p24 in the corresponding culture supernatants was monitored regularly. (B) Effect of cotransfection of the propagation of the infection was obtained (Fig. 4A). pNLRT3 on the infectivity of pNL43 in a single cycle infectivity assay. Similarly, HeLa cells were cotransfected with pNL43 and Five micrograms of pNL43 were cotransfected with increasing amounts pNLRT3 in different proportions, and the resulting virus of pNLRT3 into HeLa cell monolayers. Equal volumes of supernatant was analyzed for infectivity on the P4 indicator cell line, from the cotransfected cultures were used to infect P4 cells (HeLa- which allows simple and precise evaluation of HIV infec- CD4-LTR-LacZ) (Charneau, et al., 1994). After incubation and X-gal staining, blue cells were counted in a total of 40 microscopic fields, tivity on the basis of a single round of replication. In this and the results were expressed as the average number of blue cells experiment, the amount of wild-type proviral DNA in each per field. The numbers above each bar are the ratios of the amounts of transfection was kept constant, and the amount of mutant pNL43 to pNLRT3 (indicated in micrograms) used in the transfections.

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology pol REGIONS REQUIRED FOR HIV PROCESSING AND MATURATION 35 proportional to the amount of mutant DNA in the transfec- could have affected the kinetics or the timing of gag tion. Such a reduction of the infectivity of particles from cleavage by the protease, resulting in an improperly or- a wild-type genome by coexpression of truncated gag- dered sequence of maturation-related events. The effect pol demontrates that the mutant gag-pol precursor is of our mutations on gag processing and particle morphol- readily incorporated in virions, and that it can interfere ogy was determined to some extent, but not solely, by with the function of the protease present on full-length the size of the truncations of the gag-pol polyprotein gag-pol precursors. precursor. In mutant pNLRT3, for example, the size of the truncation is comparable to that in pNLdK, although DISCUSSION NLRT3 exhibits a much greater gag processing impair- ment than pNLdK. In this case, the processing defect The present study has confirmed and extended previ- could relate more to the position of the truncation relative ous observations that the requirements for efficient pro- to the protease domain, suggesting that domains that lie teolytic cleavage of the Pr55gag protein and thus particle closer to the protease in the RT region are more im- maturation are not confined to the gag and protease portant than more C-terminal domains. The more drastic regions alone (Quillent et al., 1993; Ross et al., 1991). All effect of the pNLRT3 deletion could also be partly ex- changes which we made in the pol gene (including a plained by the described enhancement of protease activ- mutant that possessed all of gag-pol except integrase) ity in trans by RT (Goobar-Larsson et al., 1995). Neverthe- affected both particle morphology and gag processing. less, since clear, gradual effects on gag processing and Although we cannot conclude that the abnormal mor- particle morphology were also observed for mutations phologies observed were solely due to the disrupted pro- affecting the RNase H and/or the integrase coding re- cessing of the gag precursor, the finding that very similar gions, one must conclude that several domains in pol defects could be achieved by treating a wild-type virus must be kept intact for optimal processing of gag by the with a protease inhibitor suggests that defective cleav- protease. age must account for a large part of our morphological The activity of the HIV-1 protease requires dimerization defects. In particular, the aberrant morphology patterns of protease subunits present on separate gag-pol pro- that we observed for the NLint and NLdK mutations, in teins, followed by autocatalytic cleavage and release of which only a minimal gag processing defect was ob- free, functional, dimeric protease (Krausslich, 1991, served, was very reminiscent of what has been described 1992). The mutations described here could have an effect following treatment with small doses of a protease inhibi- on any of the events leading to the activation of the tor (Kaplan et al., 1993). The same type of morphological protease. First, the gag-pol truncations could change the alterations of HIV virions has been also recently reported overall conformation of the gag-pol precursor, in effect for different integrase mutants, in which, however, no modifying the ability of the protease domain to dimerize obvious defect in gag polyprotein processing was de- or cleave itself out of the polyprotein. Alternatively, the tected (Engelman et al., 1995). We can only speculate on deletions could directly affect dimerization of the gag-pol the mechanisms through which a minimal impairment of precursor, through the removal of dimerization domains protease function would result in such radical morpho- in RT or integrase (Goff, 1990; Jones et al., 1992). Under logical alterations. One likely possibility is that beyond a this hypothesis, dimerization of the protease as part of threshold of protease function impairment, the nucleoid the gag-pol precursor would require the assistance of is incompletely assembled due to a lack of mature gag several other dimerization sites within the polyprotein. It products, allowing RNA and nucleocapsid contents to can be envisioned that such multiple dimerization events leak toward the periphery of the particle. Interestingly, within pol would be important essentially in the context very similar morphology types were observed in particles of the budding particle. Encapsidation of gag-pol precur- produced by an HIV-1 RNA mutant in which RNA packag- sors is dependent upon interactions of the gag domains ing was reduced as a result of a short deletion in the that are part of gag-pol with gag precursor molecules 5؅ untranslated region of the genome. In this mutant, (Park and Morrow, 1992). Since the number of gag-pol nevertheless, we could detect no significant gag protein precursor molecules at the site of particle budding is processing defect (Clavel and Orenstein, 1990). greatly exceeded by that of gag precursors (Jacks et al., There is little doubt that the deletions described here 1988), strong pol-pol interactions could be required to and the resulting truncations of the gag-pol precursor secure their dimerization. had a direct effect on the ability of the protease to cleave Finally, we have found that the expression of truncated gag and/or gag-pol polyprotein precursors. The pattern gag-pol precursors could significantly interfere with the of gag precursor accumulation in our mutants was very maturation of wild-type particles. This demonstrated that similar to what is seen following treatment of HIV-in- the truncated precursors can be normally incorporated fected cells with protease inhibitors, indicating a global into particles and suggested that a precise balance be- decrease in protease function, rather than an inability to tween gag proteins and full-length gag-pol precursors cleave at particular sites. Alternatively, the truncations must be attained to acheive proper maturation of the

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology 36 QUILLENT ET AL. particle. In this respect, it would be interesting to evalu- Jacks, T., Powers, M. D., Masiarz, F. R., Luciw, P. A., Barr, P. J., and ate whether overexpression of gag alone in HIV-infected Varmus, H. E. (1988). Characterization of ribosomal frameshifting in HIV-1 gag-pol expression. Nature 331, 280–283. cells could interfere with particle maturation, and Jones, K. S., Coleman, J., Merkel, G. W., Laue, T. M., and Skalka, A. M. whether this could prove a means to protect cell popula- (1992). Retroviral integrase functions as a multimer and can turn over tions from spreading HIV infection. catalytically. J. Biol. Chem. 267, 16037–16040. Kaplan, A. H., Zack, J. A., Knigge, M., Paul, D. A., Kempf, D. J., Norbeck, ACKNOWLEDGMENTS D. W., and Swanstrom, R. (1993). Partial inhibition of the HIV-1 prote- ase results in aberrant virus assembly and the formation of non We thank Luc Montagnier for his support and Katherine M. Kean for infectious viral particles. J. Virol. 67, 4050–4055. her interest in this work. We are indebted to Claudine Axler for her Kohl, N. E., Emini, E. A., Schleif, W. A., Davis, L. J., Heimbach, J. C., help in preparing the electron microscopy samples. This work was Dixon, R. A., Scolnick, E. M., and Sigal, I. S. (1988). Active human supported in part by the Agence Franc¸aise de Recherches sur le SIDA immunodeficiency virus protease is required for viral infectivity. Proc. (ANRS). C.Q. and A.M.B. are recipients of ANRS fellowships. Natl. Acad. Sci. USA 85, 4686–4690. Krausslich, H. G. (1991). Human immunodeficiency virus proteinase REFERENCES dimer as component of the viral polyprotein prevents particle assem- bly and viral infectivity. Proc. Natl. Acad. Sci. USA 88, 3213–3217. Adachi, A., Gendelmann, H. E., Koenig, S., Folks, T., Willey, R., Rabson, Krausslich, H. G. (1992). Specific inhibitor of human immunodeficiency A., and Martin, M. A. (1986). Production of acquired immunodefi- virus proteinase prevents the cytotoxic effects of a single-chain pro- ciency syndrome-associated in human and non-human teinase dimer and restores particle formation. J. Virol. 66, 567–572. cells transfected with an infectious molecular clone. J. Virol. 59, 284– Krausslich, H. G., and Wimmer, E. (1988). Viral proteinases. Annu. Rev. 291. Biochem. 54, 701–754. Charneau, P., Mirambeau, G., Roux, P., Paulous, S., Buc, H., and Clavel, Lœb, D. D., Hutchison III, C. A., Edgell, M. H., Farmerie, W. G., and F. (1994). HIV-1 reverse transcription: A termination step at the center Swanstro¨m, R. (1989). Mutational analysis of human immunodefi- of the genome. J. Mol. Biol. 241, 651–662. ciency virus type 1 protease suggests functional homology with Clavel, F., Hoggan, M. D., Willey, R. L., Strebel, K., Martin, M. A., and aspartic proteinases. J. Virol. 63, 111–121. Repaske, R. (1989). Genetic recombination of Human Immunodefi- Park, J., and Morrow, C. D. (1992). The nonmyristylated Pr160gag-pol ciency Virus. J. Virol. 63, 1455–1459. polyprotein of human immunodeficiency virus type 1 interacts with Clavel, F., and Orenstein, J. (1990). A mutant of HIV-1 with reduced Pr55gag and is incorporated into viruslike particles. J. Virol. 66, 6304– RNA packaging and abnormal particle morphology. J. Virol. 64, 5230– 6313. 5234. Pearl, L. H., and Taylor, W. R. (1987). A structural model for the retroviral .Crawford, S., and Goff, S. P. (1985). A deletion mutation in the 5؅ part . Nature 329, 351–354 of the pol gene of Moloney blocks proteolytic Quillent, C., Dumey, N., Dauguet, C., and Clavel, F. (1993). Reversion processing of the gag and pol polyprotein. J. Virol. 53, 899–907. of a polymerase-defective integrated HIV-1 genome. AIDS Res. Hum. Engelman, A., Englund, G., Orenstein, J. M., Martin, M. A., and Craigie, 9, 1031–1037. R. (1995). Multiple effects of mutations in Human Immunodeficiency Ross, E. K., Fuerst, T. R., Orenstein, J. M., O’Neill, T., Martin, M. A., and Virus type 1 integrase on viral replication. J Virol. 69, 2729–2736. Venkatesan, S. (1991). Maturation of human immunodeficiency virus Gelderblom, H. R., Hausmann, E. H. S., Ozel, M., Pauli, G., and Koch, particles assembled from the gag precursor protein requires in situ M. A. (1987). Fine structure of human immunodeficiency virus (HIV) processing by gag-pol protease. AIDS Res. Hum. Retroviruses 7, and immunolocalization of structural proteins. Virology 156, 171– 475–483. 176. Royer, M., Hong, S. S., Gay, B., Cerutti, M., and Boulanger, P. (1992). Gheysen, D., Jacobs, E., de Foresta, F., Thiriart, C., Francotte, M., Thines, Expression and extracellular release of human Immunodeficiency D., and De Wilde, M. (1989). Assembly and release of HIV-1 precursor virus type 1 Gag precursors by recombinant baculovirus-infected Pr55gag virus-like particles from recombinant baculovirus-infected cells. J. Virol. 66, 3230–3235. insect cells. Cell 59, 103–112. Sarubbi, E., and Denaro, M. (1993). Epitope mapping of a monoclonal Goff, S. P. (1990). Retroviral reverse transcriptase: Synthesis, structure, antibody which binds HIV-1 Gag and not the Gag-derived proteins. and function. J. Acquired Immune Defic. Syndr. 3, 817–831. FEBS Lett. 335, 335–337. Goobar-Larsson, L., Luukkonen, B. G. M., Unge, T., Schwartz, S., Utter, Shioda, T., and Shibuta, H. (1990). Production of human immunodefi- G., Strandberg, B., and Oberg, B. (1995). Enhancement of HIV-1 pro- ciency virus (HIV)-like particles from cells infected with recombinant teinase activity by HIV-1 reverse transcriptase. Virology 206, 387– viruses carrying the gag gene of HIV. Virology 175, 139– 394. 148. Gottlinger, H. G., Sodroski, J. G., and Haseltine, W. A. (1989). Role of Yu, X., Yuan, X., Matsuda, Z., Lee, T. H., and Essex, M. (1992). The capsid precursor processing and myristoylation in morphogenesis matrix protein of human immunodeficiency type 1 is required for and infectivity of human immunodeficiency type 1. Proc. Natl. Acad. incorporation of into mature virions. J. Virol. 66, 4966– Sci. USA 86, 5781–5785. 4971.

AID VY 7837 / 6a14$$$101 04-02-96 13:57:05 vira AP: Virology