We previously described the DM deletion mutant (Fig. 1A) of HIV-2 which shows a major packaging defect 8. Recently, the formation of a stem, named stem B and located at the base of SL-1 (Fig. 1B), was shown to be required for efficient genome encapsidation and viral replication 14. The sequence involved in the formation of stem B is part of a 10 nt palindrome (pal), located within the packaging signal Psi (Ψ), and which has been suggested to play a role in RNA dimerisation in vitro 21. To investigate the requirement for this palindromic sequence in HIV-2 dimerisation in vivo, we mutated the first four bases of pal but maintained the bases involved in stem B formation (Fig. 1A, SM2). In HIV-2, initiation of dimerisation has been postulated to occur at the DIS, a palindrome located at the top of SL-1 (Fig. 1B) 2122. Interestingly, mutations of the HIV-1 DIS revealed that the sequence was not required for genomic RNA dimerisation in vivo or for viral replication in primary cells but was important for replication in T cells 2526. Hence, we decided to examine whether the HIV-2 DIS was required for genomic RNA dimerisation and viral replication by mutating the first three bases of the palindrome (Fig. 1A, SM1).

None of the above described mutations had a significant effect on protein production by the virus as judged by western blot analysis (data not shown, 8), even though the reverse transcriptase (RT) activities of the DM and SM2 mutants were slightly reduced compared to that of the wild type and SM1 mutant (data not shown), suggesting that the virus production is slightly lower for the two Psi mutants.

We analysed the effect of the mutations in the packaging signal (DM and SM2) and the DIS (SM1) on genomic RNA dimerisation. Virion RNA extracted from wild type and mutant HIV-2 was analysed by non-denaturing northern blot (Fig. 2A) and the percentage of dimer present in each sample was quantified by densitometry (Fig. 2B). RT activity was measured to load RNA from an equivalent amount of virus particles in each lane. Wild type HIV-2 RNA appeared mostly dimeric within the virion (80%, Fig. 2B), whereas viruses bearing mutations in the Psi region (DM and SM2) showed a significant defect in dimerisation, with around 30 to 35% dimer (Fig. 2B), in addition to a reduction in the apparent amount of RNA encapsidated (Fig. 2A). These results demonstrate that the Psi region contains a signal essential for efficient dimerisation in vivo and that at least part of it is mapped to the palindrome pal. Surprisingly, mutation of the DIS palindrome does not have an effect on dimerisation in vivo and 80% of the genomic RNA packaged appears dimeric (Fig. 2A and 2B). This result contradicts data obtained in an in vitro dimerisation assay, where short RNA transcripts harbouring the SM1 mutation did not dimerise efficiently (data not shown). However, this discrepancy reflects the importance of working in the context of the whole virus, where different long-range interactions and secondary and tertiary structures than those observed in vitro 101227 might influence the ability of the genome to form a dimer. Furthermore, a number of factors such as the Gag and nucleocapsid (NC) proteins, which have been shown to promote dimerisation 1027282930, are only present in the context of the virus.

Deletion of the DM sequence (Fig. 1A) was previously reported to cause a severe packaging defect in HIV-2 8. Northern blot analysis of the SM2 mutant revealed a possible defect in encapsidation (Fig. 2A), even though only four bases of the Psi region are substituted in this mutant. To investigate this further, we assessed the level of HIV-2 genomic RNA in the cytoplasm of transfected cells and pelleted virions using RNase Protection Assay (RPA, Fig. 3). Wild type and mutant genomic RNAs were detected by probing with KS2ΨKE (Fig. 3B) and the size of the protected fragments are shown in figure 3A. A specific riboprobe (KS2ΨEP) was used to measure plasmid DNA contamination (Fig. 3B). Finally, equal loading of cytoplasmic RNA was confirmed by probing for GAPDH mRNA (Fig. 3B). Packaging efficiencies, taken as the ratio of virion to cytoplasmic RNA of a mutant relative to the wild type, are reported in figure 3C. As observed in figure 2A, the SM2 mutant displayed some reduction in packaging (approx. 40% on average). Although this figure is lower than for the DM deletion mutant, which showed a 70% decrease in RNA encapsidation, statistical analysis revealed that the difference in the packaging efficiencies of these two mutants was not significant. By contrast, the SM1 mutation of the DIS did not affect HIV-2 RNA encapsidation (Fig. 3C).

Since the SM2 mutant exhibited a comparable reduction in the level of dimer as the DM mutant, but retained a greater portion of the HIV-2 packaging determinant and possibly encapsidated its genome more efficiently than the DM mutant, it was of interest to compare the replication kinetics of these two mutants (Fig. 4). Despite having no defect in RNA encapsidation or dimerisation, the SM1 mutant was included to verify that mutation of the DIS does not impair viral replication over a longer period of time. As shown on figure 4, viral spread in the DM and SM2 mutants was markedly reduced and the mutant viruses were not able to revert to a replication competent phenotype despite prolonged culture. The similar behaviour of the DM and the SM2 mutants suggest that a failure to encapsidate a dimeric genome rather than just a reduction in encapsidation might be responsible for the replication defect observed. Indeed, viral RNA dimerisation has been associated with viral infectivity in other retroviruses 16193132. The SM1 mutant virus replicated as efficiently as wild type virus, confirming that an intact DIS palindrome is dispensable for the establishment of a productive infection.

Although we now had evidence that mutation of the DM region, including pal, altered both genomic RNA encapsidation and dimerisation, it was important to know whether this alone was responsible for the very poor replication of the DM and SM2 viruses in T cells. Hence, we examined whether the mutant viruses were impaired in infectivity using the Ghost CCR5/CXCR4 reporter cell line which contains a stably transfected reporter cassette, consisting of an HIV-2 LTR driving expression of a green fluorescent protein (gfp) gene. Successful infection of these cells by HIV-2 leads to activation of the cassette by the newly synthesised Tat protein and can be detected by measuring GFP production in the infected cells. We compared the infectivity of VSV-G pseudotyped wild type and mutant virions. Both the DM and the SM2 mutants showed a strikingly lower level of infectivity when equivalent amounts of virus, as measured by RT activity, were used to infect the Ghost cells (Fig. 5A). As observed in T cells, the SM1 mutation had no effect on the ability of the virus to infect permissive target cells. The envelope-deleted viruses or VSV-G envelope had no effect on GFP expression when the corresponding plasmids were transfected individually into COS-1 cells (data not shown). The VSV-G envelope was used to pseudotype envelope-deleted HIV-2. This may influence the ability of the virus to enter and hence infect the target cells because of the postulated difference in entry mechanism using this envelope. However, a preliminary experiment was carried out using full-length wild type and DM mutant HIV-2. Hela cells stably transfected with a reporter cassette, consisting of an HIV-1 LTR driving expression of a beta-galactosidase gene, were used. Similar results as those observed in the Ghost CCR5/CXCR4 cells were obtained (data not shown), even though the high level of background staining rendered the determination of an exact infectious titre difficult.

The reduced infectivity observed could be due to the defective packaging of the DM and SM2 viruses and thus the delivery of fewer genomes to the Ghost cells. Therefore, we were concerned to ensure that the absolute quantity of genomic RNA delivered to the Ghost cells was similar for wild type and mutant viruses. Since we have shown that the DM deletion virus displayed a three fold reduction in RNA encapsidation (Fig. 3C), we augmented the DM virus input by three fold so that a comparable number of genomes would be delivered to the Ghost cells (Fig. 5B). We verified the level of HIV-2 genomic RNA in two viral inputs by semi-quantitative RT-PCR performed on serial dilutions of the virus inoculum (Fig. 5B, top panel). To ensure that the RNA extraction process did not result in a loss of RNA, leading to differences in genomic RNA levels, a GAPDH carrier RNA was added prior to the extraction and analysed in parallel by RT-PCR as an internal control (data not shown). Interestingly, targeting equivalent numbers of HIV-2 genomes to the Ghost cells did not lead to equivalent infectivity (Fig. 5B). With both inputs tested, infectivity of the DM virus was three to five fold lower than that of the wild type virus. These results show that the infectivity defect seen in figure 5A was not simply a consequence of reduced numbers of virus genomes entering the Ghost cells.

To investigate whether there was any contribution of RNA encapsidation and genome dimerisation to viral assembly in HIV-2, we purified virions produced by cells transfected with either wild type or DM deletion provirus, both envelope-deleted, and analysed them using electron microscopy and negative staining. Electron micrographs were examined by two independent observers and approximately two hundred particles were identified for each virus. Representative particles containing immature, mature and multiple cores are shown in figure 6A (i), (ii) and (iii), respectively. We observed that cells transfected with the DM proviral DNA produced significantly fewer mature particles as compared to those transfected with wild type provirus (Fig. 6B). Surprisingly, the loss of mature particles for the DM virus was not accompanied by an increase in the proportion of immature virions (Fig. 6B). However, there was a larger number of duplex cores in the DM deletion virions, although this difference does not appear to be statistically significant (Fig. 6B). Furthermore, the existence of multiple cores in mature virions has been documented previously for the related virus HIV-1 33. These results show that a mutation in the HIV-2 packaging signal that also affects genome dimerisation leads to a reduction in the number of mature particles and an abnormal proportion of particles containing more than one core, possibly explaining the loss of infectivity associated with this mutation. Finally, it was also notable that the average diameter of the mutant virions was larger than that of the wild type virions (Fig. 6C). This did not appear to solely result from an increased number of particles containing multiple cores, unlike in HIV-1, where particles containing two cores were significantly larger than those containing a single core 33. Taken together, these results suggest that genomic RNA encapsidation, genome dimerisation and virion particle morphology are extremely closely linked in HIV-2.

pSVR is an infectious proviral clone of HIV-2 ROD containing a simian virus 40 origin of replication 7. Restriction sites and nucleotide numbering, where given, are relative to the first nucleotide of the viral RNA. Proviral construct pSVRDM containing a 28 nt deletion in the 5' leader has been previously described 8. Mutations in the 5' leader were introduced by site-directed mutagenesis 50 into a subclone of HIV-2, pGRAXS 8, using primers 5' GGCGCGGGCCGACCAACCAAAGGCAGCGTGTGG 3' and its complement and 5' GGAACAAACCACGACCCTCTGCTCCTAGAAAGGCG 3' and its complement for SM1 and SM2, respectively. Sequences from the resulting subclones were introduced into the provirus by exchanging an AatII-XhoI fragment, generating proviral constructs pSVRSM1 and pSVRSM2. Proviral constructs pSVRΔNB and pSVRΔNBDM which contain a 550 nt deletion in the env gene (positions 6369 to 6927) have been previously described 8. pSVRΔNBSM1 and pSVRΔNBSM2 were generated by replacing the AatII-XhoI region of pSVRΔNB with the same region of pSVRSM1 and pSVRSM2 respectively. pCMV-VSVG contains the vesicular stomatitis virus (VSV) G glycoprotein gene in the context of pCDNA3 (Invitrogen). All plasmids based on HIV-2 proviral sequences were grown in TOPF'10 (Invitrogen) Escherichia coli at 30°C to avoid recombination. All other plasmids were grown in DH5α E. coli under standard conditions.

Plasmids used as template for production of riboprobes were constructed as follows. Plasmids KS2ΨKE and KS2ΨEP have been previously described 851. They contain HIV-2 sequences from positions 306 to 751 and (-107) to 306 of pSVR, respectively, cloned into the polylinker of the pBluescript KSII(+) transcription vector (Stratagene). A glyceraldehyde-3-phosphate dehydrogenase (GAPDH) riboprobe template was created by reverse transcriptase (RT)-PCR amplification of the sequence from position 61 to 346 of human GAPDH using 5' GGTGAAGGTCGGAGTCAACG 3' and 5' AATTAACCCTCACTAAAGGACTCCACGACGTACTC 3' primers. Total cytoplasmic RNA extracted from Jurkat T cells served as template for the reaction. In vitro transcription of the linearised plasmids and PCR fragment using T3 RNA polymerase yielded antisense riboprobes for use in the northern blot and RNase protection assay (RPA).

COS-1 simian epithelioid cells, obtained from the European Collection of cell culture (ECACC), were maintained in Dulbecco's modified Eagle's medium (DMEM, Gibco BRL) supplemented with 10% fetal calf serum and 1× penicillin/streptomycin (Invitrogen). Cells were transfected in 10 cm-diameter dishes by the DEAE-dextran method 52 with a total of 5 μg of DNA. Cells and supernatants were harvested 48 h later, and virus production was assessed with a RT assay 53. CD4⁺ human osteosarcoma (HOS) cells stably transfected with CCR5, CXCR4 and the green fluorescent protein (gfp) reporter gene under the control of HIV-2 LTR (designated Ghost CCR5/CXCR4), provided by the Centralised Facility for AIDS Reagents, NIBSC, UK, were maintained in DMEM supplemented with 10% fetal calf serum, 1× penicillin/streptomycin (Invitrogen), 100 μg/ml hygromycin (Invitrogen), 500 μg/ml geneticin (Gibco BRL) and 1 μg/ml puromycin (Sigma) as previously described 54. Jurkat T cells, provided by the Centralised Facility for AIDS Reagents, NIBSC, UK, were maintained in RPMI 1640 medium (Roswell Park Memorial Institute, Gibco BRL) supplemented with 10% fetal calf serum and 1× penicillin/streptomycin (Invitrogen).

Replication assays were performed as described previously 8 with the following modifications. Pelleted virus from two individual transfections were resuspended in 500 μl RPMI 1640 each and pooled. Subsequently, an amount of virus equivalent to 1 × 10⁷ cpm of RT activity was added to 1 × 10⁶ Jurkat cells in the presence of 10 μg/ml DEAE-dextran in one well of a 6-well culture plate. After 24 h at 37°C, cells were transferred into a 25 cm² flask containing 10 ml of media and virus replication was followed by measuring RT activity every 4 days. Cells were split 1/2 every 8 days.

Supernatants of COS-1 cells transfected with 5 μg envelope-deleted provirus and 5 μg VSV-G expressor plasmid, 10 μg envelope-deleted provirus alone or 10 μg VSV-G expressor alone as well as mock transfected cells were harvested as previously described 8 except that pelleted virus from two individual transfections were resuspended in 500 μl DMEM each and pooled. The RT activity of the resulting virus preparation was determined as described above and the amount of virus added to a 6-well plate of cells was normalised this way. Ghost CCR5/CXCR4 cells were seeded at 2 × 10⁵ cells per well in selection medium 24 h prior to infection. Cells were exposed to increasing amount of WT and mutant viruses in the presence of 8 μg/ml polybrene (Sigma) for 16 h, after which the virus-containing medium was replaced by fresh selection medium. At 72 h post-infection, cells were washed twice with cold PBS, trypsinised and fixed in 2% paraformaldehyde. GFP expression was measured by FACS analysis.

Viral RNA from normalised amounts of virions was isolated using the QiAamp viral RNA kit (Qiagen). 1 μg of in vitro synthesised GAPDH RNA (pos. 61 to 346) was added to each sample prior to extraction. The extracted RNA was treated with 5 U TURBO DNase (Ambion) for 1 h at 37°C, followed by heat inactivation for 10 min at 65°C. 10 μl of neat RNA, 1:10 or 1:100 dilution were used for RT-PCR using One step RT-PCR kit (Abgene) with HIV-2 Psi F 5' TAATACGACTCACTATAGGCTGAGTGAAGGC 3' and Psi R 5' AGGTACTTACCTTCACCC 3'. A control without RT was performed with 10 μl of neat RNA. Samples were visualised on a 2% agarose gel.

Cytoplasmic and virion RNAs were harvested with the RNAeasy mini kit (Qiagen) and the QIAamp viral RNA kit (Qiagen), respectively. The isolated RNA was treated with 5 U of TURBO DNase (Ambion) for 30 min at 37°C and extracted once with acid-buffered phenol-chloroform and once with chloroform. RNA was precipitated with ammonium acetate and ethanol and stored at -80°C.

KS2ΨKE biotinylated riboprobe was synthesised by in vitro transcription of linearised plasmid using T3 RNA polymerase (Promega), and purified with a G-50 column (Roche) prior to use in northern blot assays. Reagents for the northern blots were obtained from a commercially available kit (Ambion). Virion RNA input was normalised on RT activity, with an equivalent of 2.5 × 10⁶ cpm being the standard amount used per reaction. RNAs were subjected to non-denaturing electrophoresis on 0.8% LE-agarose (Ambion) gel. For size determination, biotinylated RNA Millennium markers (Ambion) were run in parallel. RNAs were transferred onto positively charged nylon membranes (Ambion) prior to incubation with 0.4 nM riboprobe for 16 h at 68°C. Detection was performed using the BrightStar Biodetect kit (Ambion). RNA was quantified by densitometry using the ImageJ software.

[α-³²P]-labelled KS2ΨKE riboprobe was synthesized by in vitro transcription of linearised plasmid using T3 RNA polymerase (Promega) and gel purified prior to use in RPA, as described previously 8. RNA inputs were normalised on concentration, for cytoplasmic RNA, and RT activity, for virion RNA. Typically, 2 μg of cytoplasmic RNA and an equivalent of 2.5 × 10⁶ cpm of viral RNA were used. For each experiment, a separate RPA was performed using the same RNA inputs but probing for viral plasmid DNA using a probe generated from plasmid KS2ΨEP. In addition, a probe for human GAPDH RNA was included in the reaction to control for variations in cytoplasmic RNA input. Any DNA contamination or variations in the GAPDH signal were accounted for when calculating encapsidation efficiencies, taken as the ratio of virion to cytoplasmic RNA of a mutant relative to the wild type.

Supernatants of COS-1 cells transfected with envelope-deleted WT and DM deletion mutant proviruses were harvested as previously described 8. Pelleted virions were resuspended in 10 μl of 100 mM NaCl, 50 mM MOPS buffer at 4°C for 24 h prior to negative staining. 0.8 μl of virion preparation was applied to a carbon coated grid and stained with a few drops of 1% uranyl acetate. Micrographs were recorded on a Philips EM208S at a nominal magnification of × 20,000.