|Home | About | Journals | Submit | Contact Us | Français|
Primate lentiviruses are composed of several distinct lineages, including human immunodeficiency virus type 1 (HIV-1), HIV-2, and simian immunodeficiency virus SIVagm. HIV-1 and HIV-2 have significant differences in the mechanisms of viral RNA encapsidation. Therefore, the RNA packaging mechanisms of SIVagm cannot be predicted from the studies of HIV-1 and HIV-2. We examined the roles of the nucleocapsid (NC) zinc finger motifs on RNA packaging by mutating the conserved zinc finger (CCHC) motifs, and whether SIVagm has a preference to package RNA in cis by comparing the RNA packaging efficiencies of gag mutants in the presence of a wild-type vector. Our results indicate that the SIVagm NC domain plays an important role in Gag-RNA recognition; furthermore SIVagm is distinct from the other currently known primate lentiviruses as destroying either zinc finger motif in the NC causes very drastic RNA packaging defects. Additionally, trans-packaging is a major mechanism for SIVagm RNA encapsidation.
In all known retroviruses, full-length viral RNAs are specifically encapsidated into the virions (Vogt, 1997). This specific encapsidation of the viral genomic RNA is based on interactions between the viral protein Gag and the packaging signal in the viral RNA (Berkowitz, Fisher, and Goff, 1996; Lever, 2000; Luban and Goff, 1991). The major packaging signals have been identified in many retroviruses; in most cases, the sequences are located at the 5′ untranslated region of the viral genome and often extend into the gag gene sequences (Adam and Miller, 1988; Aldovini and Young, 1990; Bender et al., 1987; Berkowitz et al., 1995a; Embretson and Temin, 1987; Kaye, Richardson, and Lever, 1995; Lever et al., 1989; McBride and Panganiban, 1996; Mougel and Barklis, 1997). Gag has at least three domains: matrix (MA), capsid (CA), and nucleocapsid (NC) (Leis et al., 1988; Vogt, 1997). Additionally, most viruses have other domains that are varied in size and location. For example, human immunodeficiency virus type 1 (HIV-1) has spacer peptide 1 (SP1, also known as p2) located between CA and NC domains, and SP2 (also known as p1), and p6 that are C-terminal to the NC domain. Of the Gag domains, NC is known to play an important role in RNA recognition (Berkowitz et al., 1995b; Certo et al., 1999; Kaye and Lever, 1998; Rein et al., 1994; Zhang and Barklis, 1995); the MA domains of some viruses are thought to also interact with nucleic acid (Katoh et al., 1991; Ott, Coren, and Gagliardi, 2005; Poon, Li, and Aldovini, 1998).
With the exception of spumaviruses, retroviral NC domains contain one or two conserved motifs with the C-X2-C-X4-H-X2-C (CCHC) sequence, which have similarity to some known cellular zinc finger motifs that bind zinc. The CCHC motif has been shown to bind zinc in a number of viruses (Bess et al., 1992; Green and Berg, 1989; Summers et al., 1992); furthermore, zinc binding is essential for the function of the motif, as a zinc-ejecting agent causes the loss of viral infectivity in HIV-1 (Rossio et al., 1998). The CCHC motif is critical to NC function; mutations that destroy the zinc-binding ability of the CCHC motif often lead to loss of viral infectivity and significant decrease of full-length viral RNA encapsidation (Dupraz et al., 1990; Gorelick et al., 1988; Gorelick et al., 1990; Kaye and Lever, 1999). Interestingly, varied effects have been observed when replacing the CCHC motif with other zinc finger motifs such as CCCC or CCHH. In murine leukemia virus (MLV), substituting CCHC with CCCC or CCHH did not affect viral RNA encapsidation, but these mutant viruses had other defects and were noninfectious (Gorelick et al., 1996). In HIV-1, substituting the CCHC/CCHC motifs with CCCC/CCHC, CCHC/CCCC, or CCCC/CCHH generated viruses that efficiently package viral RNAs. However, replacing the motifs with CCCC/CCCC or CCHH/CCCC resulted in viruses that did not efficiently package viral RNA (Gorelick et al., 1999b). Therefore, there are far more subtle features in the conserved CCHC sequences than the simple ability to bind zinc, and other zinc finger motifs may not be able to replace the authentic sequences. Additionally, for those viruses with two CCHC motifs, the individual motifs may have different contributions to various NC functions and they cannot replace each other’s function (Gorelick et al., 1993). This effect is illustrated in the analyses of the two CCHC motifs in HIV-1. Disruption of the N-terminal zinc finger motif had a more drastic effect in the loss of viral RNA encapsidation (Gorelick et al., 1990). Replacing the N-terminal CCHC motif with the C-terminal CCHC motif, thereby generating a virus with two C-terminal motifs, resulted in a significant decrease of viral RNA encapsidation and loss of viral infectivity. Replacing the C-terminal CCHC motif with that of the N-terminal motif resulted in a virus that packaged viral RNA but also had decreased infectivity (Gorelick et al., 1993). These studies indicated that the two CCHC motifs contribute distinctive functions to NC; the N-terminal zinc finger contributes more to RNA packaging. Interestingly, although preserving the CCHC motif(s) is critical for RNA packaging in most viruses, there are exceptions. For example, simian immunodeficiency virus strain mne (SIVmne), a different primate lentivirus, has two CCHC motifs in the NC domain. SIVmne does not require both NC zinc finger motifs for RNA packaging — destroying either motif does not significantly affect RNA packaging (Gorelick et al., 1999a). Therefore, different viruses may have varied requirements for the CCHC motifs in RNA packaging.
Full-length genomic RNA is both the genetic material packaged into virions and the template for the translation of Gag/Gag-Pol (Butsch and Boris-Lawrie, 2002). There are two contrasting mechanisms proposed for Gag-RNA interaction: the cis- and trans-packaging mechanisms. The cis-packaging mechanism proposes that Gag encapsidates the RNA from which it was translated, whereas the trans-packaging mechanism suggests that Gag does not have a strong preference to package the RNA from which it was translated. It has been shown that two separate pools of RNAs exist during the replication of some viruses, such as MLV (Levin et al., 1974; Levin and Rosenak, 1976). One pool of RNA serves as the template for Gag/Gag-Pol translation and the other pool of RNA serves as the substrate for packaging into newly produced virions (Levin and Rosenak, 1976). When two separate pools of RNA exist for translation and encapsidation, RNA packaging mechanisms must occur predominantly in trans. In other viruses, such as HIV-1 and HIV-2, there is no evidence currently supporting the presence of two separate pools of RNA for translation and encapsidation (Butsch and Boris-Lawrie, 2000; Dorman and Lever, 2000). Therefore, in these viruses, RNA encapsidation could occur via either the cis- or trans-packaging mechanism. Interestingly, viruses from the same genus may not use the same mechanism for RNA packaging. For example, HIV-1 Gag efficiently packages RNA in trans (Butsch and Boris-Lawrie, 2000; Kaye and Lever, 1999; Nikolaitchik et al., 2006); in contrast, HIV-2 Gag preferentially packages RNA in cis (Kaye and Lever, 1999).
Primate lentiviruses consist of two human pathogens, HIV-1 and HIV-2, and SIVs that infect many species of nonhuman primates in Africa. Most of the primate lentiviruses can be assigned to six phylogenetic lineages (Hahn et al., 2000; Salemi et al., 2003). Among them, three major lineages include 1) SIVcpz, and its zoonotic derivative, HIV-1 (Gao et al., 1999; Huet et al., 1990); 2) SIVsm, its zoonotic derivative, HIV-2, and other SIVs that infect macaques (Chen et al., 1996; Hirsch et al., 1989; Novembre et al., 1992); and 3) SIVagm (Muller, 2003; Soares et al., 1997). In addition to the important human pathogens, HIV-1 and HIV-2, much research interest has been focused on SIVagm because the virus-host interactions revealed interesting features of SIVagm infection. Much like HIV-1, SIVagm generates high viral loads in the infected host (African green monkey); however, in sharp contrast to HIV-1, SIVagm does not cause immunodeficiency in the infected animal (Broussard et al., 2001; Hirsch, 2004). Despite interest in the pathogenesis of the virus, very little is known about the replication mechanisms of SIVagm. In this report, we examined the molecular mechanisms of SIVagm RNA packaging. Most of the studies on mechanisms of RNA packaging in primate lentiviruses have been performed using HIV-1 or HIV-2 or viruses closely related to HIV-2. These studies revealed that there are significant differences between HIV-1 and HIV-2 RNA packaging mechanisms such as the aforementioned cis- and trans-packaging. Therefore, the mechanisms of SIVagm RNA packaging cannot be predicted from the HIV-1 and HIV-2 studies. In this report, we found that both zinc finger motifs of the NC domain are important in SIVagm RNA encapsidation, because destroying either motif caused a drastic decrease in RNA packaging. We also found that SIVagm Gag can efficiently package the viral genome in trans. We conclude that trans-packaging is the predominant mechanism for SIVagm Gag-RNA interaction.
To study the trans-acting elements important for SIVagm viral RNA packaging, we examined the roles of the CCHC zinc finger motifs in the NC domain of Gag. We generated SIVagm-based vectors derived from pTan, which contained the full-length viral genomes. Two modifications were made to generate pTan-IHSA: a frameshift mutation was introduced into env and a DNA fragment containing IRES-hsa was inserted into nef. The general structure of pTan-IHSA is shown in Fig. 1A. The unmodified vector contains the wild-type NC domain. A series of mutations were introduced into the NC zinc finger motifs by site-directed mutagensis; these mutations are shown in Fig. 1B. The conserved zinc finger motifs were either destroyed by changing the CCHC motif with SSHS or replaced with a different zinc finger motif such as CCCC or CCHH.
To examine the effects of NC mutations on viral replication, we cotransfected each vector with an HIV-1 Env-expressing plasmid, pIIINL(AD8)env, into 293T cells. Because 293T cells do not express CD4 and CCR5, receptor and coreceptor of the HIV-1 AD8 strain Env, reinfection does not occur in this system. Virus was harvested from transfected cells, a portion of the virus was used to infect human T cell line Hut/CCR5, and another portion was reserved for biochemical analyses. Viral titers generated by wild-type or NC mutant vectors were measured by the expression of the HSA marker in the target cells.
We measured particle production from the wild-type and the various mutants virus using the capsid (p27) capture assay. In three independent experiments, the p27 values generated from all of the mutants were generally within twofold of the wild-type viruses, indicating that none of the mutants had severe defects in virion production (data not shown). Generally, the wild-type vectors generated robust viral titers and infected at least 40% of the cells based on HSA marker expression (HSA+). To compare the infectivity of the viruses generated by the wild-type virus and the NC mutants, we measured the number of HSA+ cells in the infected cells, converted the percentage of infected cells into MOI, and then normalized the number to p27 amounts. Results summarized from three independent experiments are shown in Fig. 2A, with the wild-type titer defined as 100%. Eliminating either or both of the CCHC zing finger motifs by changing the sequence to SSHS abolished detectable viral titers. Similarly, replacing either CCHC motif with CCCC reduced the viral titers to undetectable levels. However, replacing the CCHC motifs with CCHH had variable effects. When the N-terminal CCHC was changed to CCHH (CCHH/CCHC), the resulting viruses generated very low, albeit detectable, amounts of infected cells (~0.6% of the wild-type titer). When the C-terminal CCHC was changed to CCHH (CCHC/CCHH), the resulting viruses had high virus titers; after normalizing the titers to the virus production (p27 values), we determined that this mutant generated virions with approximately 60% of the wild-type virus infectivity (Fig. 2A).
The amounts of SIVagm RNA packaged by various NC mutants were determined by quantitative real-time PCR of viruses produced from cells infected with an SIVagm vector. To avoid possible contamination of the transfected DNA in the virion RNA preparation, we generated pools of cells, each expressing proviruses from one of the NC mutants. The 293T cells were cotransfected with NC mutants, a helper construct expressing HIV-1 gag-pol but lack cis-acting elements essential for viral replication, and a plasmid expressing VSV G. We have observed that HIV-1 gag-pol can transduce SIVagm genome (Fu and Hu, unpublished results). The resulting viruses were harvested and used to infect fresh 293T cells to generate cell pools containing various NC mutants. Viruses generated from these cell pools did not have functional Env and could not reinfect the cells. We then harvested viruses from these cell pools, measured their virion production by p27 assay, and the amount of RNA packaged in the virion by real-time quantitative RT-PCR using primers and probe that annealed to gag of SIVagm. The amounts of RNA packaged were standardize to the p27 amounts; because NC mutant vector RNA were used both as substrate for viral proteins translation and for encapsidation, this method measured the efficiencies of RNA packaging by various NC mutants. Data from four independent sets of experiments are summarized in Fig. 2B. Abolishing either one of the CCHC motifs rendered the virus unable to efficiently package viral RNA. Altering the CCHC sequences to other zinc finger motifs had variable effects on RNA packaging. When either the N- or C-terminal CCHC was changed to CCCC, the resulting mutant proteins did not package viral RNA efficiently. However, all of the CCHH mutants were able to package some levels of viral RNA. Compared with the wild-type virus, the CCHH/CCHC mutant packaged at a fivefold reduced efficiency, whereas the CCHC/CCHH mutant packaged at levels similar to those of wild-type viruses.
Taken together, these data indicated that abolishing the zinc finger motif by replacing CCHC with SSHS, or substituting either one of the CCHC motifs with CCCC resulted in loss of viral RNA packaging and viral infectivity. Therefore, both of the NC zinc finger motifs play important roles in the selection of viral RNA during the processes of virus assembly and replication.
To determine the mechanism(s) of SIVagm RNA packaging, we compared the packaging efficiencies of two coexpressed SIVagm RNAs, one carrying wild-type gag and the other carrying a mutant gag, in a competition assay. If Gag proteins predominantly package the RNA from which they were translated, then the wild-type Gag proteins will predominantly package the RNA of the wild-type vector. Because the mutant Gag proteins cannot efficiently package viral RNA, the mutant vector RNA will not be efficiently packaged, thereby creating a strong bias to package wild-type vector RNA in this system. In contrast, if SIVagm Gag proteins do not preferentially package in cis, then the both the wild-type RNA and the RNA containing mutant gag gene should be packaged efficiently.
We generated vectors that contained wild-type or mutant gag; the structures of these vectors are shown in Fig. 3A. The general structures of these vectors are similar to those described in Fig 1. However, the vectors used to examine packaging mechanisms carry either a functional hsa or a functional thy; additionally, each vector contains an IRES and a mutated gfp. The two vectors containing wild-type gag genes are pTanH0G and pTanT6G, which carry a hsa and a thy marker gene, respectively. We also constructed three vectors containing gag mutations; all of these mutants carry thy markers (Fig. 3A). Vector pTanT6GS239* contains a stop codon at codon 239 of Gag, thereby creating a truncation in the CA domain. Additionally, we generated two vectors with mutations in the NC zinc finger motifs; each mutation severely inhibited the packaging of viral RNA into virions (Fig. 2B). The N-terminal CCHC motif in pTanT6G1SSHS was changed to SSHS, and the C-terminal CCHC motif in pTanT6G2CCCC was changed to CCCC.
In this experimental design, the expression of both wild-type and mutant vectors is required to generate virions containing the mutant vector RNA, whereas the expression of the wild-type vector is sufficient to produce particles encapsidating the wild-type RNA. Therefore, if the two viruses are not properly coexpressed in most of the cells, biased results could be generated. To ensure that most of the virus producer cells express both vectors, we generated cell lines expressing both proviruses and examined virions generated from these producer cells. The experimental protocol is illustrated in Fig. 3B. Producer cell lines containing a wild-type vector and a mutant vector were generated by sequentially infecting 293T cells at low MOIs, between 0.1 and 0.05. Low MOIs were used to avoid the presence of multiple proviruses from the same vector in a significant portion of the cells. Dually infected cells were enriched by sorting until more than 95% of the cells expressed both HSA and Thy markers. In general, each producer cell line contains at least 100,000 independent infection events. We then compared the cellular RNA expression, preference of RNA packaging into the virions, and viral titers generated by the wild-type and gag mutant vectors.
We generated a total of eight cell lines. Two cell lines were infected with two wild-type vectors, TanH0G and TanT6G. Six cell lines were infected with a wild-type vector, TanH0G, and a mutant vector, TanT6GS239*, TanT6G1SSHS, or TanT6G2CCCC; two independent cell lines were generated for each combination of vectors.
All of the vectors used in these experiments had a mutation in env; therefore, virions generated from producer cell lines cannot reinfect producer cells or infect new target cells. To compare the titers generated by the wild-type and mutant vectors, we transfected the producer cells with pIIINL(AD8)env plasmid, which expresses env of the AD8 strain of HIV-1. The resulting viruses were used to infect Hut/CCR5 cells; these cells were processed and analyzed by flow cytometry. Infected cells were detected by the expression of HSA or Thy markers; the numbers of infected cells were converted to MOIs as previously described (Rhodes et al., 2005). Results obtained from six sets of experiments are summarized in Fig. 4A.
Virus produced from cell lines infected with two wild-type viruses, one carrying the hsa marker and the other carrying the thy marker, generated equivalent HSA and Thy titers. Viruses produced from cell lines containing one wild-type vector expressing HSA and one mutant vector expressing Thy all had high Thy titers, although the Thy titers were slightly lower (~ twofold) than the HSA titers. Therefore, the mutant vector RNAs, despite the fact that they did not carry functional gag genes, were packaged and replicated efficiently.
Compared with wild-type vectors, the twofold lower mutant vector titers could have resulted from reduced cellular RNA expression of the mutant vectors or from the preferential RNA packaging of the wild-type vector. To distinguish these two possibilities, we performed RNA analyses of the cellular and virion RNAs. The gag genes in the wild-type and mutant vectors are identical in sequence except at the mutated sites. The sequence divergence in the mutated sites was used to distinguish wild-type and mutant vector RNAs and determine the ratios of these two RNAs by using an RT-PCR DNA sequencing technique. Cytoplasmic RNA from the producer cells and RNA from the virion generated by the producer cells were isolated. These RNA samples were reverse transcribed into DNA, and the resulting DNA was analyzed by sequencing. At the mutated sites, the wild-type and mutant vectors contained different sequences, which generated two different signals; the strength of each signal reflected the amount of the particular vector in the sample. By comparing the wild-type and mutant signals, we could determine the ratio of the two vector RNAs.
We compared the wild-type and mutant vector RNA ratios in the six cell lines containing one wild-type and one mutant vector. Data from at least five sets of experiments are summarized in Fig. 4B and 4C. Analyses of the cellular RNA demonstrated that in all of the cell lines, wild-type vector and gag mutants generated equivalent signals (Fig. 4B), indicating that these vectors were expressed at similar levels. We then analyzed the RNAs isolated from cell-free virions generated by these cell lines. In the virion samples, gag mutant vector RNA signals were approximately 1.7- to 3-fold lower than those of wild-type vector RNAs (Fig. 4C). These results were consistent with the observation that the mutant vector titers were approximately twofold lower than the wild-type vector titers. Separately, the SSHS/CCHC and the CCHC/CCCC mutant have close to two log decrease in RNA encapsidation (4.5%, and 0.9% of the wild type, respectively; Fig. 2B). However, when coexpressed with a wild-type vector, RNAs from these two mutants could compete very well with the wild-type vector RNA for encapsidation and were packaged about only about twofold lower than the wild-type RNAs. These results indicated that coexpression of a wild-type vector resulted in a 10- to 50-fold rescue of the mutant viral RNA encapsidation, which mainly come from trans-packaging. Therefore, trans-packaging is efficient and a major mechanism for SIVagm RNA encapsidation.
To assess whether the RNAs encoding wild-type gag and RNAs encoding mutant gag were well mixed and copackaged into the virions, we examined the recombination frequencies between two wild-type vectors, and between a wild-type vector and a mutant vector. In the system we used to study cis- and trans-packaging, each vector also carries a mutant, nonfunctional gfp. The inactivating gfp mutation in the viruses containing hsa is located at the 5′ end of the gfp gene, whereas the mutation in the virus carrying thy is located at the 3′ end of the gfp gene. The RNA from the wild-type vector and the RNA from the mutant vector may or may not copackage together. If the two RNAs from the same vectors are copackaged, the resulting viruses are homozygous virions; if the RNAs from the two different viruses are copackaged into the same virus particle, then heterozygous virions are formed. Recombination could occur during the reverse transcription of all viruses; however, a functional gfp can only be reconstituted by recombination events in the heterozygous viruses, because the two RNAs have different mutations in the gfp gene. If gfp reconstitution occurs frequently between a wild-type gag virus and a mutant gag virus, then their RNAs must be copackaged efficiently. Therefore, recombination can be used to indicate the efficiency of RNA copackaging and whether the viral RNAs from different viruses are well mixed in the virus producer cells.
We measured the frequencies of gfp reconstitution between two wild-type gag viruses and between one wild-type gag virus and a mutant gag virus. Results from at least three independent experiments are shown in Fig. 5. We found that the frequencies of gfp reconstitution were similar in all groups, indicating that wild-type gag virus RNA and mutant gag virus RNA were copackaged efficiently. This result also suggested that these two types of RNAs were well mixed and did not traffic to different parts of the cytoplasm.
In this report, we sought to gain an understanding of the SIVagm RNA packaging mechanism(s). We found that both CCHC motifs in the NC domain of Gag are important in RNA packaging, because destroying either motif severely reduced specific viral RNA encapsidation and viral infectivity. However, the second CCHC motif could be replaced by a CCHH motif without significantly affecting RNA packaging or viral infectivity. In contrast, replacing either motif with CCCC abolished RNA packaging specificity and viral infectivity. We also determined that SIVagm Gag proteins efficiently package RNA in trans; furthermore, RNAs encoding functional or nonfunctional Gag proteins were well mixed and copackaged together efficiently.
To our knowledge, the roles of two zinc fingers in the NC domain on RNA packaging have been studied in three other primate lentiviruses: HIV-1, SIVmac, and SIVmne. Destroying both zinc fingers led to decreased RNA packaging in HIV-2; however, the role of each finger has not been studied (Kaye and Lever, 1999). These viruses belong to three different branches of the primate lentiviruses (Fig. 6A). Comparison of the viral genomes illustrates that HIV-2, SIVmac, and SIVmne cluster closer together in the same lineage (Fig. 6A); comparing the strains used in the RNA packaging studies, these three viruses have more than 83% amino acid sequence homology in Gag and in NC (Fig. 6B). HIV-1 and SIVagm are more divergent than the cluster of HIV-2, SIVmne, and SIVmac (Fig. 6A); SIVagm shares 49% and 56% homology in Gag with HIV-1 and SIVmne, respectively, and 40% and 42% homology in the NC domain with HIV-1 and SIVmne, respectively (Fig. 6B). Our study indicates that the requirements for NC zinc finger motifs are different in SIVagm than in HIV-1, SIVmac, or SIVmne. In HIV-1, both zinc fingers appear to be important for RNA packaging, although the first finger is more important; one mutant with the destroyed second finger could still package 25% of the viral genome compared with wild-type virus (Gorelick et al., 1990). In SIVmne, either of the two fingers appear to be sufficient for RNA packaging; destroying one finger had little effect on RNA packaging but destroying both reduced specific packaging sevenfold (Gorelick et al., 1999a; Yovandich et al., 2001). In SIVmac, either finger has some role in RNA packaging, because destroying either finger decreased specific packaging fivefold and destroying both fingers had a 14-fold effect (Akahata, Ido, and Hayami, 2003). In SIVagm, both fingers are required for RNA packaging; destroying either finger had a similar 20–50-fold effect on RNA packaging (Fig. 2B). Therefore, SIVagm differs from SIVmne or SIVmac because both zinc fingers in SIVagm are required for RNA packaging. SIVagm also differs from HIV-1 because destroying either of the fingers has a similar level of reduction in RNA packaging. Therefore, even within the primate lentiviruses, there are significant differences in the roles of each zinc finger motif. Because the NC domain plays a critical role in RNA selection during virus assembly, the variations in the zinc finger motif requirements most likely also reflect the differences in the molecular interactions between Gag and viral RNA in these viruses.
The full-length retroviral genome serves as template for Gag translation and as genetic material for new virions. Different viruses may use different mechanisms to balance these two roles. For example, MLV separates the RNA into two pools, whereas HIV-1 and HIV-2 do not appear to do so. This difference may provide a basis for the different mechanisms used for RNA packaging. For those viruses that do not have distinct pools of RNA for each function, it is possible for RNA encapsidation to occur in cis. There are putative theoretical evolutionary advantages and disadvantages for each type of packaging mechanism. For example, in the cis-packaging mechanism, viral genomes encoding Gag proteins capable of RNA recognition are preferentially encapsidated. Therefore, the packaging mechanisms can serve as a process to select functional genomes. Conversely, by using cis-packaging mechanism, gag mutations could be selected against at two different levels. These mutants cannot generate infectious viruses, and their RNAs and information within the genome cannot be easily rescued by other functional viruses. Trans-packaging mechanism has the opposite properties; although gag mutants are not selected against at two different levels, the mechanism also does not select for RNA encoding functional gag. However, trans-packaging mechanism allows the mixing of genetic information via copackaging and recombination, thereby possibly increasing the genetic variation of the viral population, at least in the gag gene. Interestingly, of the primate lentiviruses studied, HIV-1 and SIVagm appear to package RNA efficiently in trans, whereas HIV-2 appears to package RNA predominantly in cis. It is intriguing that these distantly related viruses evolve to use different RNA packaging mechanisms; furthermore, the employment of different mechanisms raised the questions about their effects on virus replication and fitness. It is likely that future studies on the mechanisms of viral replication will lead to a better understanding of the selective advantages of the mechanisms they employ, which may also shed light on the mechanisms of virus-host interaction and viral pathogenesis.
Standard protocols were used to construct all of the plasmids (Sambrook, 1989). Site-directed mutagenesis was performed by overlapping PCR. The general structures of the plasmids were confirmed by restriction enzyme mapping, whereas regions generated by PCR were further characterized by DNA sequencing to avoid inadvertent mutations.
Plasmid pTan contains a full-length, infectious molecular clone of the tan1 strain of SIVagm (a generous gift from Fan Gao, Duke University) (Soares et al., 1997). This molecular clone carries functional gag-pol, env, vif, tat, rev, and nef, whereas vpr contains an in-frame stop codon. A 19-bp linker was inserted into the XhoI site to introduce an inactivating frameshift mutation in env thus generating pTan1-XMS. A DNA fragment containing the internal ribosomal entry site (IRES) sequences from encephalomyocarditis virus fused to mouse heat-stable antigen (hsa) was inserted into the BsiWI site in the nef of pTan1-XMS to generate pTan1-IHSA. Vectors pSIVagmH0G and pSIVagmT6G were derived from pTan1-XMS with markers inserted into the nef gene; markers include hsa, green fluorescence protein gene (gfp) (Chalfie et al., 1994), and mouse thy1.2 gene (thy) (Giguere, Isobe, and Grosveld, 1985). Vector pSIVagmH0G carries hsa-IRES-gfp, whereas pSIVagm-T6G carries thy-IRES-gfp; in both vectors, gfp genes were inactivated by mutations. To generate SIVagm vectors containing NC mutations, mutated DNA fragments (from site-directed mutagenesis PCR) were digested with HpaI and BspEI and cloned into pTan1-IHSA digested with HpaI and BspEI. Vector pTanT6GS237* was generated by inserting a linker with an in-frame stop codon into the SanDI site, thereby creating a stop codon at codon 237 of gag. Vectors pTanT6G2CCCC and pTanT6G1SSHS were generated by replacing the BspEI to Tth111I fragment of pSIVagmT6G with that of the mutants containing CCHC/CCCC and SSHS/CCHC mutations, respectively. For clarity, pSIVagmH0G and pSIVagmT6G are referred to as pTanH0G and pTanT6G in the following text, respectively.
Hut/CCR5 is a human T cell line modified to express CCR5, and 293T is a human kidney fibroblast cell line. Cell lines were grown at 37°C with 5% CO2, in Dulbecco’s modified Eagle’s medium (293T) or Roswell Park Memorial Institute-1640 medium RPMI (Hut/CCR5) supplemented with 10% fetal calf serum and antibiotics.
DNA transfection was performed using the calcium phosphate method (Sambrook, 1989). To generate pseudotyped viruses, pTan vector was transfected with pIIINL(AD8)env, which expresses CCR5-tropic HIV-1 Env (Freed, Englund, and Martin, 1995), or pHCMV-G, which expresses vesicular stomatitis virus G protein (VSV G) (Burns et al., 1993). Supernatants were harvested 36 to 48 hours later from the producer cells, clarified through a 0.22-μm or 0.45-μm filter, and either used directly for infection or stored at −80°C and thawed prior to infection. Virus infections were performed in the presence of polybrene, infected cells were stained with phycoerythrin-conjugated anti-HSA antibody (BD PharMingen) or allophycocyanin-conjugated anti-Thy1.2 antibody (eBiosciences). Flow cytometry analyses were performed using a FACSCalibur (BD Biosciences) and data acquired were analyzed using Flowjo software (Tree Star). Cell sorting was performed using a FACSVantage SE system with the FACSDiVa digital option (BD Biosciences).
To generate cell lines containing two SIVagm vectors, sequential infections were performed using low multiplicity of infection (MOI). HIV-1 helper construct was used to mobilize mutant SIVagm vectors. MOI was calculated from the number of infected cells in the total live cells based on Poisson distribution (Rhodes et al., 2005).
Virus production was monitored by the amount of CA (p27) measured by the SIV p27 ELISA kit (Retro Teck) or by reverse transcriptase (RT) activity measured by the standard protocol (Cheslock et al., 2000; Temin and Mizutani, 1970). Cytoplasmic RNA was isolated using a standard protocol (Sambrook, 1989). Virion RNA was isolated either as previously described (Fu, Gorelick, and Rein, 1994) or using Trizol reagent as recommended by the manufacturer (Invitrogen).
The amounts of SIVagm RNA packaged by various NC mutants were determined by quantitative real-time PCR of viruses produced from cells infected with an SIVagm vector. These cells were produced by infecting 293T cells with viruses derived from VSV G-pseudotyped SIVagm vector. Cell-free viruses produced from the infected 293T cells were collected 48 hours postinfection and viral RNA were isolated from virions by standard protocol (Fu, Gorelick, and Rein, 1994). The amounts of SIVagm RNA were determined by quantitative real-time RT PCR using primers and probe annealed to sequences in MA – primers (5′-GCGGGACACTCGGCACT-3′ and 5′-CGTTCGGGCGTAAGCGTC-3′) and probe (5′-TAMTCAGGGAGGAATTTGGACACATTTGAGAAATAMRA-3′). Plasmid DNA (pTan1-XMS) was used to generate standard curves for real-time RT-PCR detection.
The percentage of wild-type or mutant viral RNA expressed in cells or packaged in the virions was determined using an RT-PCR sequencing method as previously described (Nikolaitchik et al., 2006). Briefly, RT-PCR was performed using viral RNA as template, and the resulting DNA was analyzed by a sequencing reaction. The ratios between the wild-type and mutant RNAs were calculated based on the heights of the peaks of the sequencing reactions. Standard curves were generated by mixing varied ratios of wild-type and mutant plasmids and subjecting the mixtures to PCR and sequencing reactions.
We thank Anne Arthur for expert editorial help, Vinay K. Pathak for scientific input and discussions of this project, and Vinay K. Pathak and Robert Gorelick for critical reading and suggestions of this manuscript.
This research was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.