Virion packaging determinants and reverse transcription of SRP RNA in HIV-1 particles

Transcription

1 Nucleic cids Research, 27, Vol. 35, No. 21 Published online 24 October 27 doi:1.193/nar/gkm816 Virion packaging determinants and reverse transcription of SRP RN in HIV-1 particles hunjuan Tian 1,2, Tao Wang 1, Wenyan Zhang 1,2 and Xiao-Fang Yu 1, * 1 epartment of Molecular Microbiology and Immunology, Johns Hopkins loomberg School of Public Health, altimore, M 2125, US and 2 Jilin University, Jilin, hina Received ugust 22, 27; Revised and ccepted September 18, 27 STRT iverse retroviruses have been shown to package host SRP (7SL) RN. However, little is known about the viral determinants of 7SL RN packaging. Here we demonstrate that 7SL RN is more selectively packaged into HIV-1 virions than are other abundant Pol-III-transcribed RNs, including Y RNs, 7S RN, U6 snrn and cellular mrns. The majority of the virion-packaged 7SL RNs were associated with the viral core structures and could be reverse-transcribed in HIV-1 virions and in virus-infected cells. Viral Pol proteins influenced trnlys,3 packaging but had little influence on virion packaging of 7SL RN. The N-terminal basic region and the basic linker region of HIV-1 Np7 were found to be important for efficient 7SL RN packaging. lthough lu RNs are derived from 7SL RN and share the lu RN domain with 7SL RN, the packaging of lu RNs was at least 5-fold less efficient than that of 7SL RN. Thus, 7SL RNs are selectively packaged into HIV-1 virions through mechanisms distinct from those for viral genomic RN or primer trnlys,3. Virion packaging of both human cytidine deaminase POE3G and cellular 7SL RN are mapped to the same regions in HIV-1 N domain. INTROUTION Retroviruses package two copies of viral genomic RN per viral particle. The selective packaging of viral genomic RN is mediated by the specific interaction between sequences in the viral RN (c) and the nucleocapsid (N) domain of Gag molecules (1,2). lthough packaging of viral genomic RN is essential for virus infectivity, viral genomic RN is dispensable for virus assembly, which is mediated by the viral structural protein Gag. However, RNs of either viral or cellular origin are believed to be important for retroviral particle assembly (3,4). In addition to viral genomic RN, retroviruses also contain abundant copies of small RN molecules ranging in size from 4S to 7S (1,2). mong these small RNs, the trns used by various retroviruses, and particularly the primer trns, have been well characterized (1,2). Primer trns are selectively packaged through an interaction with viral reverse transcriptase (5,6). In the case of HIV-1, trnlys,3 is also selected by means of an interaction between the capsid domain of Gag and trnrs, which forms a complex with trnlys,3 (7). Other small RNs in retroviral particles that have been characterized include 7SL RN (8 12), 5S rrn (9), and U6 snrn (9,13). recent study of Moloney murine leukemia virus (MuLV) observed that 7SL RN and viral genomic RN were similarly enriched in MuLV virions (9). Several other cellular RNs, including Y1 RN, Y3 RN, 1 RN, 5S rrn and U6 snrn, were also found to be packaged with an efficiency similar to that of 7SL RN (9). n earlier study detected three major species (7S, 5S and trns) of small cellular RNs in HIV-1 virions (14). lthough 7SL RN has been detected in HIV-1 virions (1,12), packaging of other Pol-III-transcribed RNs into HIV-1 virions has not been well studied. The viral determinants for the packaging of various cellular small RNs are also poorly defined. In the present study, we have shown that 7SL RN is packaged into HIV-1 particles at a much higher efficiency than are Y RNs, 7S RN or U6 snrn. lthough lu RN was derived from 7SL RN and shares the lu domain with 7SL RN, packaging of lu RNs was at least 5-fold less efficient than that of 7SL RN. The majority of the virion-associated 7SL RN molecules were associated with viral core structures. Viral Pol and Env proteins, as well as viral genomic RN, were dispensable for the packaging of 7SL RN. lthough both the M and N domains of HIV-1 Gag polyproteins have been shown to interact with nucleic acids, we found that the N domain, but not the M domain or L domain (p6), played a critical role in *To whom correspondence should be addressed. Tel: ; Fax: ; xfyu@jhsph.edu The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First uthors ß 27 The uthor(s) This is an Open ccess article distributed under the terms of the reative ommons ttribution Non-ommercial License ( by-nc/2./uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

2 Nucleic cids Research, 27, Vol. 35, No mediating 7SL RN packaging. The N-terminal basic region and the basic linker region of HIV-1 N, but not the zinc finger motifs, were important for 7SL packaging. MTERILS N METHOS Plasmid constructs The HIV-1 constructs Pr-, PolEnv, Gag expression vectors pggins, pns and pp2lz have been previously described (15). 2FS- was generously provided by r Shan en. HIV-1Gag-myc, SrcMGag-myc, Gag- FP432, Gag-FP411, Gag-FP46 and Gag-FP395 were generously provided by r Paul Spearman. Gag411N8 and Gag411Z1 were generated from Gag- FP411. The HIV-1 N mutant construct pr was a generous gift of r Robert Gorelick. ntibodies and cells The following antibodies were used for this study: anti-h mouse monoclonal antibody (Mab; ovance, at. #MMS-11R-1), anti-myc mouse mouse Mab (Sigma, at. #M 5546), pooled HIV-1+ human sera, anti-n antibody and anti-human ribosomal P antigen antibody (Immunovision, at. #HP-1). nti-p24 Mab (at# 1513) and anti-gp12 antibody (at# 288) were obtained from the IS Research Reagents Program, ivision of IS, NII, NIH. 293T cells were maintained in ulbecco s modified Eagle s medium (MEM, Invitrogen) with 1% fetal bovine serum and gentamicin (5 mg/ml) (-1 medium) and passaged upon confluence. Transfections, virus and VLP purification, and virion-associated RN extraction N transfection was carried out using Lipofectamine 2 (Invitrogen) as recommended by the manufacturer. Viruses in cell culture supernatants were cleared of cellular debris by centrifugation at 3 r.p.m. for 15 min in a Sorvall RT 6 centrifuge and filtration through a.2-mm-pore size membrane (Millipore). Virus particles were then concentrated by centrifugation through a 2% sucrose cushion by ultracentrifugation at 1 g for 2 h at 48 in a Sorvall Ultra8 ultracentrifuge. Viral pellets were resuspended in lysis buffer (PS containing 1% Triton X-1 and complete protease inhibitor cocktail [Roche] and RNase inhibitor [New England iolabs]). Viral lysates were analyzed by immunoblotting, or virionassociated RN was extracted with Trizol (Invitrogen) according to the manufacturer s instructions. ell and virion-associated RNs were extracted and analyzed by qrt-pr. For sucrose density gradient purification of HIV-1 particles, viral pellets as described above were dissolved in PS solution and loaded on top of 2 to 6% stepwise sucrose gradients. Samples were subjected to ultracentrifugation using a SW41 rotor in the Sorvall iscovery ultracentrifuge at 22 r.p.m. for 16 h at 48. fter centrifugation, 12 fractions (.85 ml each) were collected from the top of the gradient and analyzed by immunoblotting using anti-p24 Mab. Fractions 3 to 8 were also analyzed for viral genomic RN and virionassociated cellular RNs. Samples (1 ml) from each fraction were left untreated, treated with 8 mg/ml RNase (Roche, ) in the absence of detergent, or treated with 8 mg/ml RNase plus.1% Triton X-1 at 378 for 3 min. Reactions were stopped by the addition of 1 mm ET plus 1 ml Trizol, and RNs were extracted according to the manufacturer s instructions. Immunoblot analysis ells were collected 48 h after transfection. ell and viral lysates were prepared as previously described (16). ells (1 1 5 ) were lysed in 1 loading buffer (.8 M Tris, ph 6.8, with 2.% SS, 1% glycerol,.1 M dithiothreitol, and.2% bromophenol blue). The samples were boiled for 1 min, and proteins were separated by SS-PGE. Proteins were transferred onto two separate nitrocellulose membranes by passive diffusion for 16 h, producing identical mirror-image blots. Membranes were probed with various primary antibodies against proteins of interest. Secondary antibodies were alkaline phosphatase-conjugated anti-human, anti-rabbit, antimouse, or anti-goat IgG (Jackson Immunoresearch, Inc), and staining was carried out with 5-bromo-4-chloro- 3indolyl phosphate (IP) and nitroblue tetrazolium (NT) solutions prepared from chemicals obtained from Sigma. Quantitative real-time PR (qrt-pr) RN samples were derived from cell lysates or viral lysates and treated with Nase by incubation in 1 ml of EP-treated water with 1 RQ1 RNase-free Nase uffer, l ml RQ1 RNase-free Nase (Promega) and 4 U RNase inhibitor (New England iolabs) for 3 min at 378. Nase was inactivated by the addition of 1 ml of RQ1 Nase Stop Solution and incubated at 658 for 1 min. RN was reverse-transcribed using random primers and the Multiscribe reverse transcriptase from the High apacity cn rchive it (pplied iosystems) according to the manufacturer s instructions. The cn was either undiluted or serially diluted in EP-treated water before input into the qrt-pr reaction to ensure that amplification was within the linear range of detection. The I 7 sequence detection system (pplied iosystems) was used for qrt-pr amplifications. ll primers were synthesized by Invitrogen, and fluorescent-tagged probes were synthesized by pplied iosystems. garose gel analysis was used to verify that each primer pair produced single amplicons, and the identities of the PR products were verified by cloning and sequencing. qrt-pr was performed using either Taqman fluorescent probes or SYR Green methods. For the Taqman method, each 2 ml reaction contained 1 ml of each forward- and reverse-specific primers (1 mm), 1 m of fluorescent TaqMan probes (5 mm), 1 ml of 2 Universal Taqman PR Master Mix, 4 ml of RNasefree water, and 3 ml of template cn. The reactions were carried out under the following conditions: 58 for 2 min,

3 729 Nucleic cids Research, 27, Vol. 35, No for 1 min, 4 cycles of 958 for 15 s and 68 for 1 min. The target sequences were amplified using the following primer pairs and probes: Y1 RN: forward, 5 -GGTGGTGGGTGT G-3 ; reverse, 5 -GTGTGTGGT- 3 ; and probe 5 FM-TGTTGTTGTGTT -TMR- 3 ; Y3 RN: forward, 5 -GGTGGTGGTGGT- 3 ; reverse, 5 -GGTGTGTGG-3 ; and probe 5 FM-GTTGTT-T MR- 3 ; 7SL RN: forward, 5 -TGGGTGTGT G-3 ; reverse, 5 -TTTGGT-3 ; and probe 5 FM-TTTGGTGT TMR -3 ; HIV RN: forward, 5 -TGTGTGGTTGTT GTGT-3 ; reverse, 5 -GGTTGGTGGGG -3 ; and probe 5 FM-GTGGGGG GG-TMR- 3 ; beta-actin RN: forward, 5 -TTGTG TTG-3 ; reverse, 5 -GGGGT TTGTGG-3 ; and probe: 6FM-TGT TGTTGGT-TMR-3. The primer/probe set specific for 3G was a predesigned TaqMan gene expression assay (pplied iosystems assay, identification number: Hs222415). For the SYR Green method, each 2 ml reaction contained 1 ml of each forward- and reverse-specific primer (1 mm), 1 ml of2syr Green PR Master Mix, 5 ml of RNase-free water, and 3 ml of template cn. The reactions were performed with the following conditions: 58 for 2 min, 958 for 1 min, 4 cycles of 958 for 15 s and 68 for 1 min, followed by a dissociation protocol. Single peaks in the melting curve analysis indicated specific amplicons. The target sequences amplified by the SYR Green method used the following primer pairs: 5S RN: forward, 5 -TTGGTTGGT -3 ; reverse, 5 -GGGT-3 ; Y4 RN: forward, 5 -GGTGGTGTGGTG TG-3 ; reverse, 5 -GGTTTTGGTG GG-3 ; Y5 RN: forward, 5 -GTTGGTGGTGTTGT GGGT-3 ; reverse, 5 -GGTGTGG G-3 ; trn-phe: forward, 5 -GGTGTGTTG GGG-3 ; reverse, 5 -TGGTGGGG-3 ; trn-lys,3: forward, 5 -GGGTGTGT G-3 ; reverse, 5 -TGGGGGG-3 ; GPH: forward, 5 -G TTTGG GT-3 ; reverse, 5 -TGTTGTTTTG G-3 ; The copy numbers of the target cns in the qrt-pr assay were determined by using a standard curve of 1-fold serial dilutions of non-linearized plasmid N containing the target sequence (ranging from 5 or 1 copies to or 1 6 copies). bsolute RN copy numbers were calculated by using standard dilution curves of plasmids containing the target sequence. If the template cn was diluted before input into the reaction, the copy number of the target transcript was adjusted by the dilution factor. The sensitivity of the assay or limit of detection was determined by the lowest copy number that was consistently amplified within the linear portion of the standard curve. The detection limit ranged from 5 to 5 copies per reaction for the various primer sets. The qrt-pr assay detected 3G, Y1, Y3, Y4, 7SL and GPH at 5 copies; 5S and trn-lys,3 at 1 copies; lu1, trn-phe and Y5 at 5 copies per reaction. ll standards at varying concentrations were amplified linearly over a range of at least 5 orders of magnitude, and the correlation coefficients (R 2 ) were greater than.99. opy numbers of transcripts were calculated by using standard dilution curves of plasmids containing the target sequence. etection of 7SL RN reverse transcription in HIV-1 virions and infected cells 293T cells were transfected with NL4-3 or control pcn3.1. ulture supernatant was clarified by lowspeed centrifugation, filtered through a.22-mm membrane, and sedimented by ultracentrifugation over a cushion of 2% sucrose at 28 r.p.m. using an SW28 rotor for 2 h. The viral pellets were re-suspended in 2 ml of endogenous RT buffer (4 mm Tris-Hl, ph 8., 1 mm Mgl 2, 6 mm l, 2 mm TT,.2% Triton X-1) containing no dntp or.2 mm dntp. Samples were incubated at 378 for16 h. The reaction mixtures were stopped by adding 1% SS at 18 for 1 min. The reaction products were extracted by adding chloroform/ phenol and precipitated with isopropanol, washed with ethyl alcohol, and dissolved in water for PR or qrt-pr analysis. To detect 7SL N in HIV-1-infected cells, a total of 1 6 MGI cells were infected with NL4-3 viruses (equivalent to 1 mg of p24) or negative control supernatant for 14 h. ells were then washed with PS, dissolved in lysis buffer (.5% Triton X-1, 5 mm Tris-Hl, ph 7.5, 1 mm Nal,1 mm ET) for 3 min at 378, and centrifuged at 3 g for 1 min, and the pellets (containing nucleus and chromosomal N) were discarded. The N sample in the supernatant was extracted with chloroform/phenol, precipitated with ethanol (3- to 4-fold), and dissolved in water for qrt-pr analysis. RESULTS The majority of 7SL RN molecules in HIV-1 virions are associated with the viral cores n earlier study has identified three major species of small cellular RNs (7S, 5S and 4S trns) in purified HIV-1 virions (14). Recently, 7SL RN has also been detected in HIV-1 virions (1,12); however, the mechanism of 7SL packaging into HIV-1 virions is still unclear. It is possible that 7SL RNs passively diffuse into budding particles as a result of the high local concentration at the site of virus assembly/budding. If this were the case, one would not expect 7SL RNs to be enriched in any subviral structures. To address this question, we determined the localization of 7SL RN in mature HIV-1 virions. Purified HIV-1 virions were subjected to brief treatment with the non-ionic detergent Triton

4 Nucleic cids Research, 27, Vol. 35, No gp12 Virus+ etergent 2% Fraction 1 (soluble proteins) 2 RTp66 Pr55 Gag gp41 INp32 p24 6% 3 (viral cores) Mp17 Np Fraction HIV-1 RN 12% 1% 8% 6% 4% 2% % 7SL RN 12% 1% 8% 6% 4% 2% % Fraction Fraction Figure 1. ssociation of 7SL RN molecules with viral cores. Purified HIV-1 virions were treated briefly with detergent to separate cores from viral membranes, as illustrated in () and described previously (17). () nalysis of core-associated viral proteins. HIV-1 core-enriched fraction 3 contained HIV-1 INp32, Np7, GagPr55 and p24. Viral membrane-associated proteins gp12, gp41 and Mp17 were mostly found in the soluble fraction 1. s previously described (18,19), significant amounts of p24 and RTp66 also disassociated from the cores. ( and ) ssociation of HIV-1 viral genomic RN () and 7SL RN molecules () with viral cores. RNs in fractions 1, 2 and 3 were extracted and analyzed by qrt-pr using specific primers for HIV-1 RN or 7SL RN. Results represent means standard deviations of at least three independent experiments. X-1 to remove the viral membrane and isolate viral cores (Figure 1) as previously described (17). The majority of the viral-membrane-associated proteins, such as the Env proteins gp12 and gp41 and the viral Map17 proteins (Figure 1, lane 1), were separated from the core structure (lane 3). In contrast, most of the Np7 and INp32 were located inside the cores (Figure 1, lane 3). onsistent with the observations of Forshey et al. (18) and Tang et al. (19), we also found that some p24 and RTp66 molecules could be separated from the core structures (Figure 1). s expected, HIV-1 genomic RN was located mainly in the cores (Figure 1, lane 3). Similarly, the majority of the 7SL RNs were also detected with the cores (Figure 1, lane 3). These results suggest that 7SL RNs are likely to be recruited into HIV-1 virions by interacting with viral components inside the core structures. Selective packaging of 7SL RNs into HIV-1 virions lthough packaging of 7SL RN and trns into HIV-1 virions has been studied (6,1,12), the packaging of other Pol-III-derived RNs into HIV-1 virions has not been fully characterized. To examine the packaging of Pol-III-derived RNs into HIV-1 particles, HIV-1 virions in the culture supernatants of NL4-3-transfected 293T cells were separated from cell debris by filtration and ultracentrifugation through a 2% sucrose cushion, then further purified on sucrose density gradients (2 6% wt/ vol). s a control for non-specific secretion of cellular RNs in the culture supernatants, culture supernatants from mock-transfected 293T cells were also prepared sideby-side. Sucrose-gradient-purified HIV-1 virions were pelleted by ultracentrifugation, and virion-associated RNs were analyzed by real-time PR (qrt-pr) using specific primers for various Pol-III RNs. Packaging of various Pol-III RNs within HIV-1 virions was indeed detected, and there was only minimal non-specific secretion of most of these RNs into the culture supernatant in the absence of HIV-1 virions (Figure 2). For example, the level of particle-associated 7SL RN in purified HIV-1 virions was more than 5-fold higher than that of the mock control (Figure 2). High copy numbers of 7SL RN and 5S rrn were detected in HIV-1 virions (Figure 2), consistent with the earlier observation of 7S and 5S RN species in purified HIV-1 virions (14). The packaging of various RNs could potentially be influenced by the relative abundances of these RNs in

5 7292 Nucleic cids Research, 27, Vol. 35, No. 21 opies of Pol-III RNsin HIV-1 virions MO HIV-1 7SL Y1 Y3 Y4 Y5 5S lu GPH Relative efficiencies of Pol-III RN packaging into HIV-1 virions 1.% 1.% 1.% 1.%.1%.% 7SL Y1 Y3 Y4 Y5 7S Lys,3 U2 U6 5S lu GPH ß-actin Figure 2. nalysis of the packaging of cellular RNs into HIV-1 virions. () etection of 7SL RN and 5S rrn in purified HIV-1 virions. HIV-1 virions in the culture supernatants of NL4-3-transfected 293T cells were purified by sucrose density gradient centrifugation. Virion-associated RNs were analyzed by qrt-pr using specific primers for Y1, Y3, Y4, Y5 RNs, 5S rrn, GPH mrn, lu RNs or 7SL RN. Similar samples from mock-transfected 293T cells were prepared and analyzed side-by-side to determine non-virusmediated secretion of various RNs. The copy numbers of RNs detected in HIV-1 virions from 6 pg of p24 equivalent culture supernatant are shown. () Relative efficiency of virion packaging of RNs, as analyzed by qrt-pr. The efficiency of RN packaging was evaluated by comparing the ratio of the number of copies of a particular RN in HIV-1 virions to the number of copies of the same RN in virus-producing cells. The efficiency of 7SL RN packaging was set to 1%. Results represent means standard deviations of at least three independent experiments. virus-producing cells. We therefore compared the relative efficiency of virion packaging of the various cellular RNs. For comparison purposes, the efficiency of 7SL RN packaging (copies of 7SL RN in HIV-1 virions/copies of 7SL RN in virus-producing cells) was set to 1%. Selective enrichment of primer trnlys,3 and 5S rrn within HIV-1 virions was also observed (Figure 2). Virion packaging of U6 snrn, 7S RN, Y1, Y3, Y4 and Y5 RNs was much less efficient than that of 7SL RN (Figure 2). bundant cellular mrns such as b-actin and GPH mrns were also packaged much less efficiently into HIV-1 virions than was 7SL RN (Figure 2). Since the intracellular abundance of the various RNs was taken into consideration, the inefficient packaging of these RNs when compared to 7SL RN could not be attributed to their lower intracellular levels. In MuLV, a 7SL RN-derived SINE RN, 1 RN, was packaged as efficiently as 7SL RN (9). lu RNs are also derived from 7SL RN and share the lu domain structures with it (2). However, lu RNs were also packaged into HIV-1 virions at least 5-fold less efficiently than was 7SL RN (Figure 2). While 7SL RN was clearly packaged within HIV-1 virions (Figure 2), HIV-1 virion-associated 5S rrn was only a few-fold higher than the mock control (Figure 2). To further investigate whether 5S rrn was indeed packaged into HIV-1 virions, sucrose-gradient-purified HIV-1 virions were pelleted by ultracentrifugation, and virion-associated RNs were analyzed by real-time PR (qrt-pr) using specific primers for 7SL RN or 5S rrn. Fractions 3 8 contained detectable HIV-1 virions, with the peak occurring in fractions 5 7, as determined by analyzing viral p24 (Figure 3), viral genomic RN (Figure 3) or 7SL RN (Figure 3). Surprisingly, both HIV-1 virions and samples from mock-transfected cells contained 5S rrn peaks that co-migrated with p24 (Figure 3). Virionassociated 7SL RNs (Figure 3E) and 5S rrn (Figure 3F) were resistant to RNase treatment unless the viral membrane was first disrupted with the detergent Triton X-1, indicating that these RNs were packaged within viral particles. In repeated experiments, the level of HIV-1 virion-associated 5S rrn was always a few-fold higher than the mock control, suggesting that some 5S rrn was indeed packaged into HIV-1 virions. These results also suggest that 293T cells secreted some unidentified structures that contained 5S rrn and had a density similar to that of HIV-1 virions. To further examine the packaging of 7SL RN and 5S rrn into HIV-1 virions, HIV-1 (NL4-3)-infected 4+ Jurkat T cells were established. HIV-1 virions in the culture supernatants of NL4-3-infected Jurkat cells were separated from cell debris by filtration and ultracentrifugation through a 2% sucrose cushion, then further purified on sucrose density gradients (2 6% wt/vol). s a control for non-specific secretion of cellular RNs in the culture supernatants, culture supernatants from uninfected Jurkat cells were also prepared side by side. Sucrose-gradient-purified HIV-1 virions were pelleted by ultracentrifugation, and virion-associated RNs were analyzed by real-time PR (qrt-pr) using specific primers for 7SL RN, 5S rrn and HIV-1 genomic RN. Fractions 3 8 contained detectable HIV-1 virions, with the peak occurring in fraction 5, as determined by analyzing viral p24 (Figure 4) and viral genomic RN (Figure 4). Packaging of 7SL RN (Figure 4) or 5S rrn (Figure 4) into HIV-1 virions produced from HIV-1-infected Jurkat cells, but not uninfected Jurkat cells, was clearly observed. HIV-1 Gag N zinc-binding domains, viral genomic RN, Pol and Env proteins are not essential for 7SL RN packaging With the exception of primer trnlys,3, the determinants of Pol-III RN packaging into HIV-1 virions are poorly defined. The influence of the viral Pol and Env proteins on 7SL RN packaging was first examined using the constructs Pr-, 2FS- and PolEnv. The construct Pr- contains an active site mutation in the protease gene and expresses the Gag, Gag-Pol and Env proteins (Figure 5). The 2FS- construct contains mutations that destroy the frameshifting sequences in the gag coding region that are essential for the generation of Gag-Pol

6 Nucleic cids Research, 27, Vol. 35, No k Pr55 Gag p HV-1 genomic RN Mock 7SL RN NL MO 5S rrn NL43 E SL RN Untreated Rnase treated RNase+Triton F S rrn Untreated RNase treated RNase+Triton Figure 3. haracterization of 7SL RN and 5S rrn in HIV-1 virions. () NL4-3 viruses were produced from transfected 293T cells and purified by sucrose-density gradient centrifugation. Viral proteins in each fraction were analyzed by immunoblotting using the anti-p24 antibody. Virion-associated RNs were extracted from fractions 3 to 8 of the sucrose-density gradient and analyzed by qrt-pr using HIV-1 viral RN- (), 7SL RN- () or 5S rrn- () specific primers. Samples from mock-transfected 293T cells were also prepared side-by-side and analyzed to detect the non-specific secretion of 7SL RN or 5S rrn. Virion-associated 7SL RN (E) or 5S rrn molecules (F) were resistant to RNase treatment unless the virions were pretreated with Triton X-1. Results represent means standard deviations of at least three independent experiments. proteins (Figure 5). The construct PolEnv contains a deletion of most of the protease, the whole RT, and the IN coding sequences and an internal deletion of the Env coding sequence that abolished most of the Env protein expression (Figure 5). VLPs from transfected 293T cells were separated from cell debris by filtration and ultracentrifugation through a 2% sucrose cushion. s a control for non-specific secretion of cellular RNs into the culture supernatants, supernatants from mock-transfected 293T cells were also prepared side by side. Pelleted particles were analyzed by immunoblotting using a monoclonal antibody against p24 (Figure 5). s expected, Pr- particles contained both Pr55 Gag and Pr16 Gag-Pol, while 2FS- and PolEnv particles contained Pr55 Gag but not Pr16 Gag-Pol (Figure 5). Particle-associated RNs were isolated and analyzed by qrt-pr using specific primers for 7SL RN or trnlys,3. Packaging of 7SL RNs was clearly detected in Pr-, 2FS- and PolEnv particles (Figure 5). Non-specific secretion of 7SL RN from the mocktransfected 293T cells was less than 9.% of that for the Pr- particles (Figure 5). Packaging of 7SL RNs was essentially comparable in Pr-, 2FS- and PolEnv particles (Figure 5). In contrast, packaging of trnlys,3 was reduced in 2FS- and PolEnv particles when compared to that in Pr- particles (Figure 5), consistent with previous reports (5,6) that viral Pol proteins influence the packaging of trnlys,3. The role of viral genomic RN in the packaging of 7SL RNs was also examined. We compared the level of virion-associated 7SL RNs in PolEnv particles and in virus-like particles (VLP) containing only Gag molecules (pggins) (15). The PolEnv construct contains all the 5 RN sequences upsteam of the gag coding sequence, which are important for viral genomic RN packaging (Figure 6). The Gag amino acid sequence of pggins is identical to the Gag of PolEnv. However, pggins lacks all the 5 viral RN sequences, including the viral RN packaging sequences upstream of the Gag coding region (Figure 6). lthough VLP produced by pggins contained less than 1% of the viral RN present in

7 7294 Nucleic cids Research, 27, Vol. 35, No. 21 k SL RN p24 Mock NL HIV-1 genomic RN 5S rrn Mock NL4-3 Figure 4. Packaging of 7SL RN and 5S rrn into HIV-1 virions produced from infected 4+ Jurkat T cells. () NL4-3 viruses were produced from infected 4+ Jurkat T cells and purified by sucrose-density gradient centrifugation. Viral proteins in each fraction were analyzed by immunoblotting using the anti-p24 antibody. Virion-associated RNs were extracted from fractions 3 to 8 of the sucrose-density gradient and analyzed by qrt-pr using HIV-1 viral RN- (), 7SL RN- (), or 5S rrn- () specific primers. Samples from mock-transfected 293T cells were also prepared side-by-side and analyzed to detect the non-specific secretion of 7SL RN or 5S rrn. Results represent means standard deviations of at least three independent experiments. particles produced by PolEnv (data not shown), packaging of 7SL RNs was readily detected in VLP produced by pggins (Figure 6). When the amount of Gag was normalized, the level of 7SL RNs was similar in VLP produced by pggins and in particles produced by PolEnv (Figure 6). Thus, VLP containing the Gag molecules alone could efficiently package 7SL RN, suggesting that the Gag molecule but not viral genomic RN contains a major determinant for the packaging of 7SL RN. Furthermore, packaging of viral genomic RN into HIV-1 mutant pr viruses (21), with mutations of the cysteine residues in the zinc fingers of the HIV-1 N domain (Figure 6), was reduced by about 87% when compared to that of NL4-3 viruses (Figure 6). However, these mutations had little effect on 7SL RN packaging into these N mutant viruses, when compared to the parental HIV-1 NL4-3 viruses (Figure 6E). Thus, the zinc fingers of HIV-1 Gag N domain and the viral genomic RN do not play a major role in mediating 7SL RN packaging. The N-terminal basic region and the basic linker region of the HIV-1 Gag N domain are required for 7SL RN packaging To examine which Gag domain(s) is(are) required for 7SL RN packaging, the HIV-1 full-length Gag construct (pggins), the Gag construct missing p6 and p1 (pns) and the Gag construct missing N (P2LZ) were used (Figure 7). P2LZ contains a deletion of N, which is replaced by the leucine zipper (LZ) domain of the yeast GN4 (Figure 7), and is assembly-competent (15,22). Viral particles were produced from transfected 293T cells and separated from cell debris by filtration and ultracentrifugation through a 2% sucrose cushion. Particle-associated RNs were extracted and analyzed by qrt-pr using 7SL RN-specific primers. 7SL RNs were detected in both GagINS and NS VLPs. When the amount of Gag was normalized (Figure 7), there was no significant difference in the levels of these RNs in the two types of VLPs (Figure 7). On the other hand, 7SL RNs were poorly packaged into P2LZ VLPs lacking N (Figure 7). Thus, the p6 domain of HIV-1 Gag is not essential for the packaging of 7SL RNs, but the N domain is apparently important. The M domain of HIV-1 Gag has also been reported to mediate RN binding (23). To examine whether the M domain of Gag is required for 7SL RN packaging, HIV-1 Gag VLP with or without M (Figure 7) were produced from transfected 293T cells. The membrane targeting function in the M deletion mutant was restored by the v-src myristylation signal (24). Viral particles were separated from cell debris by filtration and ultracentrifugation through a 2% sucrose cushion (Figure 7E). Particle-associated RNs were extracted and analyzed by qrt-pr using 7SL RNspecific primers. 7SL RNs were packaged into Gag VLPs containing M and Gag VLPs lacking M (Figure 7F). 7SL RN packaging was only slightly reduced in Gag VLPs lacking M (Figure 7F). Thus, the N domain is more important than the M domain for the packaging of 7SL RNs. To further map the region of the N domain that is required for 7SL RN packaging, HIV-1 Gag-FP expression vectors (25) containing various -terminal N sequence truncations (Figure 8) were transfected into 293T cells. Gag particles were isolated and analyzed

8 Nucleic cids Research, 27, Vol. 35, No PS ψ rev Pr- LTR gag tat * pol env PS ψ rev Pol Env LTR gag pol tat env LTR LTR PS ψ rev pol 2FS- LTR gag tat * env LTR Mock Pr- Pol Env 2FS- Pr16 Gag-Pol Pr55 Gag 16% 14% 12% 1% 8% 6% 4% 2% % 7SL RN trnlys,3 12% 1% 8% 6% 4% 2% % Mock Pr- Pol Env 2FS- Mock Pr- Pol Env 2FS- Figure 5. Pol or Env is not required for 7SL RN packaging. () Illustration showing Pr- (which expresses viral Env proteins and uncleaved Gag and Gag-Pol proteins), 2FS- (which expresses viral Env proteins and uncleaved Gag proteins) and PolEnv (which only expresses uncleaved Gag proteins) constructs. () Immunoblot analysis of Pr-, 2FS- or PolEnv particles using an anti-p24 antibody. etection of 7SL RN () or trnlys,3 () in Pr-, 2FS- or PolEnv particles by qrt-pr using 7SL RN- or trnlys,3-specific primers. Packaging of 7SL RN or trnlys,3 in Pr- particles was set to 1%. Results represent means standard deviations of at least three independent experiments. by immunoblotting using an anti-p24 antibody (Figure 8). Particle-associated RNs were extracted and analyzed by qrt-pr using primers specific for 7SL RN. The packaging of 7SL RNs into Gag432 (containing full-length N) particles was set to 1%. Gag411, which contains the -terminal zinc finger deletion, was able to package 7SL RN efficiently when compared to Gag432 (Figure 8). However, removal of the basic linker region of the N domain (Gag45) significantly reduced 7SL RN packaging (Figure 8). The N-terminal basic region, the first zinc finger, and the basic linker region (Gag411) are the minimal determinants for 7SL RN packaging. To examine the role of each of these three domains in 7SL RN packaging, several N mutant constructs were generated and compared (Figure 9). Viral particles were produced from transfected 293T cells (Figure 9). Particle-associated RNs were extracted and analyzed by qrt-pr using primers specific for 7SL RN. The packaging of 7SL RNs into Gag411 particles was set to 1%. eletion of the N-terminal basic region (Gag411N8) or removal of the basic linker region of the N domain (Gag45) significantly reduced 7SL RN packaging (Figure 9). eletion of the first zinc finger (Gag411ZF1) had a lesser effect on 7SL RN packaging (Figure 9). Thus, these data indicate that the N-terminal basic region and the basic linker region of the N domain play an important role in mediating 7SL RN packaging in HIV-1. HIV-1 Gag interacts with 7SL RN in virus-producing cells To determine whether HIV-1 Gag interacts with 7SL RNs and therefore mediates its virion packaging, we transfected the HIV-1 Gag-myc or untagged Gag expression vector into 293T cells. Gag-myc, but not untagged Gag, was immunoprecipitated with the anti-myc antibody (Figure 1). RN samples of Gag-myc co-precipitated with the anti-myc antibody from transfected 293T cells were analyzed by qrt-pr using primers specific for 7SL RN. ell lysates from Gag-transfected 293T

9 7296 Nucleic cids Research, 27, Vol. 35, No. 21 Pol Env LTR PSψ rev gag pol tat env LTR GGINS MV gagins poly 12% 1% 8% 6% 4% 2% % MO Pol Env GGINS NL4-3 N F IQGNFRNQRTV R IQGNFRNQRTV E G N F E S G S G H I Zn E G W N RPR G G H I G S S W N S RPR G Zn G HQ M E G HQ S TERQN M TERQN 15% 125% 1% 75% 5% 25% MO HIV-1 pr % HIV-1 7SL Figure 6. HIV-1 Gag, but not viral genomic RN, is required for 7SL RN packaging. () Illustration showing PolEnv (which package viral RN) and pggins constructs. () etection of 7SL RN in PolEnv and pggins particles. Virion-associated RNs were extracted from purified virions and analyzed by qrt-pr using HIV-1 viral RN- or 7SL RN-specific primers. () Illustration of N sequences in NL4-3 or pr constructs. () HIV-1 N zinc finger mutant viruses (pr653-47) are defective for viral genomic RN packaging but still package 7SL RN. Results represent means standard deviations of at least three independent experiments. cells, which could not be immunoprecipitated by the antimyc antibody, were also examined side-by-side as the negative control. specific interaction of 7SL RN with Gag-myc, as compared to the control Gag sample, was clearly detected (Figure 1). Thus, 7SL RN is likely packaged into HIV-1 virions through an interaction with HIV-1 Gag molecules during virus assembly. 3G interacts with 7SL RN in HIV-1 virions The virion packaging of 7SL RN, as well as that of 3G, requires N, suggesting that 7SL RN could mediate 3G packaging. To examine whether 3G interacts with 7SL RN in released HIV-1 virions, we immunoprecipitated 3G-H in purified HIV-1 virions with an anti-h antibody conjugated to agarose beads (Figure 11). RNs co-precipitated with 3G-H were analyzed by qrt-pr using primers specific for HIV-1 genomic RN, trnlys,3, or 7SL RN. s a control, HIV-1 virions lacking 3G-H were also examined side by side to detect non-specific binding of these virionassociated RNs to the assay system (set as 1). specific interaction between 7SL RN and 3G-H was observed; this interaction was approximately 18-fold higher than that observed for control HIV-1 virion samples lacking 3G-H (Figure 11). Interaction of 3G-H with viral genomic RN was also observed (Figure 11). The level for trnlys,3 co-precipitated with 3G-H in purified HIV-1 virions was about 5-fold higher than that for the control HIV-1 virion sample lacking 3G-H (Figure 11), suggesting that there may be a weak interaction between 3G-H and trnlys,3 in HIV-1 virions. 7SL RN could be reverse-transcribed in HIV-1 virions Various cellular RNs packaged into retroviruses have been shown to be reverse-transcribed (8,13,26,27). To determine whether virion-packaged 7SL RN could be reverse-transcribed in HIV-1 virions, HIV-1 particles

10 Nucleic cids Research, 27, Vol. 35, No GGINS NS P2LZ Mp17 p24 p2 Np7 p1 p6 LZ MO GGINS NS P2LZ 12% 7SL RN 1% 8% 6% 4% 2% % MO GGINS NS P2LZ HIV-1GG-myc Mp17 p24 p2 Np7 p1 p6 myc Src MGG-myc V-Src myc E MO GG-myc Src M GG-myc F 12% 1% 8% 6% 4% 2% % 7SL RN 7SL RN MO GG-myc Src M GG-myc Figure 7. The N domain of HIV-1 Gag is important for 7SL RN packaging. () Schematic representation of HIV-1 -terminal truncation mutant constructs. () The full-length HIV-1 Gag (pggins), Gag lacking p6 and p1 (pns), or Gag lacking p6, p1 and Np7 (pp2lz) particles were produced from transfected 293T cells and analyzed by immunoblotting using the anti-p24 antibody. () Influence of N deletion on 7SL RN packaging. RNs associated with pggins, pns or pp2lz particles were extracted and analyzed by qrt-pr. Packaging of 7SL RN in pggins particles was set to 1%. () Schematic representation of the HIV-1 M deletion mutant and its parental constructs. (E) The full-length HIV-1 Gag-myc or or SrcMGag-myc particles were produced from transfected 293T cells and analyzed by immunoblotting using the anti-p24 antibody. (F) Influence of M deletion on 7SL RN packaging. RNs associated with HIV-1 Gag-myc or SrcMGag-myc particles were extracted and analyzed by qrt-pr. Packaging of 7SL RN in HIV-1 Gag-myc particles was set to 1%. Results represent means standard deviations of at least three independent experiments. were produced from transfected 293T cells. We then subjected purified HIV-1 virions to detergent-permeabilization and endogenous RT reactions, followed by PR analysis using 7SL-specific primers that could detect fulllength or near-full-length 7SL N. 7SL N was clearly detected in the HIV-1 virion sample after endogenous RT reaction in the presence of dntp (Figure 12, lane 1) but not in the absence of dntp (Figure 12, lane 2). Reverse transcription of 7SL RN in HIV-1 virions was also analyzed by qrt-pr using HIV-1 and 7SL N-specific primers. s controls we used samples from mock-transfected 293T cells and HIV-1 virion samples without the addition of dntps. Neither HIV-1 N nor 7SL N was detected in samples from mock-transfected 293T cells (Figure 12). Low levels of 7SL N and HIV-1 N were detected in HIV-1 virions without the addition of dntps (Figure 12), presumably as a result of a low level of reverse transcription in released HIV-1 virions (add ref.). Significant increased levels of 7SL N and HIV-1 N were detected in endogenous RT-treated HIV-1 virion samples (Figure 12). The PR N products were sequenced and confirmed to be 7SL or HIV-1 sequences. To further examine whether virion-packaged 7SL RN could be reverse-transcribed during virus infection, we infected MGI cells (4+ HeLa cells) with NL4-3 virions. ytoplasmic N samples from uninfected or HIV-1 infected MGI cells were extracted and analyzed by qrt-pr using 7SL- or HIV-1-specific primers. 7SL N was clearly detected in the HIV-1 infected MGI cell extracts when compared to those from uninfected MGI cells (Figure 12). Thus, these data

11 7298 Nucleic cids Research, 27, Vol. 35, No. 21 HIV-1 N G E G H E G I HQ N G M Zn Zn F N W IQGNFRNQRTV* *RPRG* TERQN* Gag411 IQGNFRNQRTVFNGEGHINRPRG Gag411 N8 IQGNF FNGEGHINRPRG Gag411 Z1 IQGNFRNQRTV RPRG Gag45 IQGNFRNQRTVFNGEGHIN MO GG45 GG411 GG411 N8 GG411 Z1 GG391 GG45 GG411 GG432 MO 12% 12% 7SL RN 1% 8% 6% 4% 2% % GG391 GG45 GG411 GG432 MO Relative abundance 1% 8% 6% 4% 2% % MO GG45 GG411 GG411 N8 GG411 Z1 Figure 8. Identification of minimal regions in the HIV-1 Gag N domain that are important for 7SL RN packaging. () Schematic representation of Gag-FP constructs. sterisks indicate the location of the stop codons in the N truncation constructs. () Immunoblot analysis of Gag-FP particles. Gag-FP particles were produced from transfected 293T cells and analyzed by immunoblotting using the antip24 antibody. () Influence of N deletion on 7SL RN packaging. RNs associated with various Gag-FP particles were extracted and analyzed by qrt-pr using 7SL RN-specific primers. The packaging of 7SL RN within Gag-FP particles containing full-length N domain (Gag432) was set to 1%. Results represent means standard deviations of at least three independent experiments. Figure 9. omparison of the role of the N-terminal basic region, first zinc finger and the basic linker region of HIV-1 Gag N in 7SL RN packaging. () Schematic representation of N deletion Gag-FP constructs. () Immunoblot analysis of N mutant particles. Gag-FP particles were produced from transfected 293T cells and analyzed by immunoblotting using the anti-p24 antibody. () Influence of N deletion on 7SL RN packaging. RNs associated with various N mutant particles were extracted and analyzed by qrt-pr using 7SL RN-specific primers. The packaging of 7SL RN within Gag411 particles was set to 1%. Results represent means standard deviations of at least three independent experiments. indicate that virion-packaged 7SL RN can be reversetranscribed during HIV-1 infection. ISUSSION In addition to viral genomic RN, which accounts for >5% of the total RN mass, small cellular RNs of 4S to 7S have been detected in diverse retroviruses (1,2). These small RNs are more abundant on a molar basis than the viral genomic RN, and some have been identified as trns, 7SL RN, 5S rrn and U6 snrn in RSV (28), MuLV (5,9,29 31), FeLV (32), visna virus (33) and HIV-1 (1,14). With the exception of primer trns, the viral determinants of Pol-III RN packaging into retroviruses are poorly understood. Our study has demonstrated that the N domain, and particularly the basic linker region of this domain, plays an important role in mediating 7SL RN packaging into HIV-1 virions. Viral genomic RN and the viral structural proteins Pol and Env were dispensable for the packaging of 7SL RNs, consistent with previous reports (1,12). The majority of the virion-packaged 7SL RNs were associated with viral cores. Furthermore, interactions of 7SL RNs with HIV-1 Gag proteins could be detected in virus-producing cells (Figure 1). These data suggest that 7SL RNs are not passively included into budding particles but are instead packaged through interactions with HIV-1 Gag molecules during virus assembly. Our data further demonstrate that virion-packaged 7SL RN could be reverse-transcribed in HIV-1 virions and in HIV-1-infected cells. Reverse transcription of cellular RN packaged into retroviruses such as mrn (26,27), U6 snrn (13), VL3 RN (34,35) and 7SL RN (8) has been previously reported. Primers annealing to the ends of 7SL sequences amplified an approximately 3-bp PR product (confirmed to be 7SL sequence), suggesting full-length or nearly full-length 7SL RN was being reverse-transcribed. Future study will be required to determine whether 7SL RN was reversetranscribed as a single N product using unidentified primer(s) or as a product of viral/7sl recombination. The N domain of retroviral Gag molecules plays important roles in viral genomic RN packaging, RN dimerization and annealing of primer trn to the viral genomic RN through RN chaperoning activity (1,2,36 39). Results presented here indicate that HIV-1 N is important for the packaging of both viral genomic RN and 7SL RNs. Mutations of cysteine residues in the zinc fingers of the HIV-1 N domain largely abolished viral genomic RN packaging. However, these N mutations had little effect on the packaging of 7SL RNs.

12 Nucleic cids Research, 27, Vol. 35, No GG GG-myc opies of Gag bound 7SL RN GG GG-myc anti-myc Figure 1. Interaction of 7SL RN with HIV-1 Gag. () HIV-1 Gagmyc, or Gag was expressed in transfected 293T cells. ell lysates were prepared and immunoprecipitated with the anti-myc Mab, and Gagmyc, but not Gag, was precipitated as analyzed by immunoblotting using the anti-myc antibody. () RNs were extracted from co-precipitated samples and analyzed by qrt-pr using primers specific for 7SL. Non-specific binding of 7SL RN to the assay system was represented by the Gag sample. Results represent means standard deviations of at least three independent experiments. These data indicate that the packaging mechanism of viral genomic RN is distinct from that of 7SL RN packaging. y analyzing the packaging of 7SL RN into a series of -terminal truncation and internal deletion mutants of the HIV-1 N domain, we were able to demonstrate that the N-terminal basic region and the basic linker region, but not the N-terminal or -terminal zinc finger, was the most critical for 7SL RN packaging. Further analysis will be required to identify the amino acids in the basic linker region of the HIV-1 N domain that are critical for this packaging process. The N-terminal basic region and the basic linker region of HIV-1 N have also been proposed to be critical for viral genomic RN packaging, virus production and virion stability (4 43). The role of 7SL RN in virus production, virion stability, and viral infectivity requires further investigation. Various Pol-III-derived RNs are packaged into retroviruses, but their packaging mechanisms are apparently different from that of 7SL RNs. Primer trn packaging requires viral reverse transcriptase (5,6) and, at least in the case of HIV-1, an interaction between trnlys synthetase and the viral capsid domain (6,7). In contrast, the packaging of 7SLRN is more dependent on the N domain of HIV-1 Gag. The apparent differences between retroviruses in terms of the packaging efficiencies of the various Pol-III-derived RNs are also worth noting. In the case of MuLV, Y1 and Y3 RNs are packaged as efficiently as 7SL RN (9). On the other hand, Y RN packaging into HIV-1 virions was much less efficient than that of 7SL RN. The packaging efficiency of U6 snrn and 7S RN into MuLV particles is only moderately lower than that of 7SL RN (9). However, in HIV-1, the packaging of U6 snrn or 7S RN was approximately 1-fold less efficient Relative RN binding to 3G-H in HIV-1 virions NL4-3 Vif HIV-1 Vif virion IP: anti-h 1 2 NL4-3 Vif +3G-H 3G-H NL4-3 Vif NL4-3 Vif+3G-H 7SL HIV-1 LYS PHE Figure 11. Interaction of 7SL RN with 3G-H in HIV-1 virions. () Immunoprecipitation and immunoblot analysis of 3G-H from HIV-1Vif virions produced from transfected 293T cells in the presence or absence of 3G-H. 5 mg NL4-3Vif was transfected into 293/3G-H or 293 cells in T175 flasks. HIV-1 virions in the culture supernatants were purified by sucrose-density gradient centrifugation. Virion lysates were immunoprecipitated with the anti-h antibody conjugated to agarose beads and analyzed by immunoblotting using the anti-h antibody. () RNs were extracted from co-precipitated samples and analyzed by qrt-pr using primers specific for 7SL RN, HIV-1 genomic RN, trnphe or trnlys,3. The RN detected in the control NL4-3Vif sample lacking 3G-H was set to 1%. Results represent means standard deviations of at least three independent experiments. than that of 7SL RN. It has been proposed that lu RNs were originally derived from 7SL RNs (2). lthough lu RNs share the lu RN domain with 7SL RN (2), we found that the packaging of lu RNs into HIV-1 virions was much less efficient than that of 7SL RN. n interesting observation from the current study is that the virion packaging of both human cytidine deaminase 3G and cellular 7SL RN mapped to the HIV-1 N domain. lthough it is still controversial whether HIV-1 viral RN plays an important role in mediating virion packaging of 3G (17,44), many groups have reported that HIV-1 Gag can mediate efficient 3G packaging in the absence of viral genomic RN (15,17,44 5). Previous studies have observed that the packaging of 3G proteins into Gag particles that contain a leucine zipper domain from the yeast GN4 in place of the HIV-1 N domain is much less efficient than that of Gag particles containing an intact N domain (15,45 5). Gag particles containing the leucine zipper domain also packaged 7SL RN significantly less efficiently than did Gag particles containing an intact N domain (Figure 7). Furthermore, we have demonstrated that the N-terminal basic region and the basic

13 73 Nucleic cids Research, 27, Vol. 35, No. 21 HIV-1+dNTP HIV-1 7SL 4bp 3bp 2bp SL 1 5 HIV dntp + opy number opy number 15 HIV HIV dntp + opy number 7SL HIV-1 + opy number HIV-1 HIV-1 + Figure 12. Reverse transcription of 7SL RN in purified HIV-1 virions. () etection of near-full-length 7SL N by PR from HIV-1 virions subjected to endogenous RT reaction. Purified HIV-1 virions (NL4-3) were partially permeabilized with detergent and incubated in the presence (lane 1) or absence (lane 2) of dntps. Samples were then analyzed by PR using 7SL-specific primers. () etection of 7SL N by qrt-pr from HIV-1 virions subjected to endogenous RT reaction. Purified HIV-1 virions (NL4-3) were partially permeabilized with detergent and incubated in the presence (lane) or absence (lane) of dntps. Samples were then analyzed by qrt-pr using 7SL- or HIV-1-specific primers. The copy numbers of RNs detected in HIV-1 samples from 2 ng of p24 equivalent virions are shown. () etection of 7SL N by qrt-pr from HIV-1- infected cells. MGI cells (1 6 ) were infected with 1 mg of p24 equivalent NL4-3 virions. ytoplasmic N samples from uninfected or HIV-1- infected MGI cells were extracted and analyzed by qrt-pr using 7SL- or HIV-1-specific primers. The copy numbers of Ns detected from samples equivalent to 6 HIV-1 virion-infected MGI cells are shown. Results represent means standard deviations of at least three independent experiments. linker region of HIV-1 N domain are important for 7SL RN packaging (Figure 8). It is interesting that these regions were also identified as being important for efficient 3G packaging (15,25). We have observed that 7SL RN is one of the most abundant RNs that interacts with 3G in HIV-1 virions (Figure 11). ollectively, these observations support a role for 7SL RN in mediating 3G packaging into retroviruses, including HIV-1. Studies mapping the regions in 7SL RN that are important for 3G and HIV-1 Gag binding/virion packaging are currently underway. NOWLEGEMENTS We thank Paul Spearman, Robert Gorelick and Shan en for critical reagents and eborah Mclellan for editorial assistance. We also thank Elana Ehrlich for technical assistance and thoughtful discussions. This work was supported by a grant from the NIH (I62644), a grant from the Johns Hopkins enter for IS Research (FR), and funding from the National Science Foundation of hina (NSF ) and heung ong Scholars Program Foundation of the hinese Ministry of Education to X-F.Y. Funding to pay the Open ccess publication charges for the article was provided by a grant from the NIH (I62644). onflict of interest statement. None declared. REFERENES 1. erkowitz,r., Fisher,J. and Goff,S.P. (1996) RN packaging. urr. Top. Microbiol. Immunol., 214, Linial,M.L. and Miller,.. (199) Retroviral RN packaging: sequence requirements and implications. urr. Top. Microbiol. Immunol., 157, ampbell,s. and Vogt,V.M. (1995) Self-assembly in vitro of purified -N proteins from Rous sarcoma virus and human immunodeficiency virus type 1. J. Virol., 69,