Domain- and nucleotide-specific Rev response element regulation of feline immunodeficiency virus production

Abstract

Computational analysis of feline immunodeficiency virus (FIV) RNA sequences indicated that common FIV strains contain a rev response element (RRE) defined by a long unbranched hairpin with 6 stem-loop sub-domains, termed stem-loop A (SLA). To examine the role of the RNA secondary structure of the RRE, mutational analyses were performed in both an infectious FIV molecular clone and a FIV CAT-RRE reporter system. These studies disclosed that the stems within SLA (SA1, 2, 3, 4, and 5) of the RRE were critical but SA6 was not essential for FIV replication and CAT expression. These studies also revealed that the secondary structure rather than an antisense protein (ASP) mediates virus expression and replication in vitro. In addition, a single synonymous mutation within the FIV-RRE, SA3/45, reduced viral reverse transcriptase activity and p24 expression after transfection but in addition also showed a marked reduction in viral expression and production following infection.

Introduction

Feline immunodeficiency virus (FIV) is a lentivirus, similar to human (HIV), simian (SIV), and bovine (BIV) immunodeficiency viruses, which causes immunodeficiency and neurological disease in its natural host, the domestic feline, similar to that of HIV in humans (Diehl et al., 1996; Pedersen, 1993; Power, Zhang, and van Marle, 2004). The genomic structure of FIV is similar to HIV and other lentiviruses (Olmsted et al., 1989; Pedersen et al., 1987; Sodora et al., 1994) but the FIV genomic sequence is distantly related to HIV-1, HIV-2, and SIV sequences (Bachmann et al., 1997; Browning et al., 2001; Olmsted et al., 1989). Indeed, FIV is among the non-primate lentiviruses which are associated with an AIDS-like disease (Diehl et al., 1996; Pedersen, 1993) and thus offers a natural model for HIV/AIDS research (Elder et al., 1998; Pistello and Bendinelli, 2008). Advancing the understanding of the molecular mechanisms underlying FIV infection and virulence might provide new insights into the development of effective interventions in terms of both antiretroviral compounds and vaccines for lentivirus infections (Elder et al., 2008; Johnston and Fauci, 2007).

It is widely assumed that lentivirus structural protein expression and genomic RNA production are mediated by the binding of the Rev protein to a highly structured cis-acting RNA element known as the Rev responsive element (RRE) (Daly et al., 1989; Felber et al., 1989; Lesnik, Sampath, and Ecker, 2002; Tang, Kuhen, and Wong-Staal, 1999). The function of Rev in the HIV-1 life cycle includes: (i) shuttling of viral structural protein mRNAs from nucleus to cytoplasm (Li et al., 2005); (ii) interacting with the RRE to suppress mRNA splicing (Barksdale and Baker, 1995; Pfeifer et al., 1991); (iii) increasing stability and translation level of the incompletely spliced viral mRNAs (Arrigo and Chen, 1991; Felber et al., 1989; Hadzopoulou-Cladaras et al., 1989; He, Henao-Mejia, and Liu, 2009); and (iv) enhancing viral RNA encapsidation (Brandt et al., 2007).

Despite extensive studies of Rev expression and function in lentivirus infections, the contribution of structural features of the RRE to viral infection is less well understood, especially with regard to its function in the full-length viral genome (Elder et al., 2008; Hope, 1999; Nekhai and Jeang, 2006; Strebel, 2003; Suhasini and Reddy, 2009). The HIV-1 RRE secondary structure was recently shown to exhibit a predominant stem-loop that contains several smaller branched stem-loop regions (Watts et al., 2009). Indeed, the HIV-1 RRE structure is difficult to study in its genomic context because the RRE overlaps the env region, thus restricting the extent to which mutagenesis studies can be performed without potentially affecting envelope function. In contrast, the putative FIV RRE overlaps the 3′ end of the env but is largely located outside of the env open reading frame (Phillips et al., 1992; Zou et al., 1997), thus permitting mutagenesis-driven investigations within this region of the virus.

Based on the degree of sequence divergence of envelope protein among the FIV isolates, the worldwide populations of FIV are classified into five subtypes or clades (Bachmann et al., 1997; Pecon-Slattery et al., 2008a; Pecon-Slattery et al., 2008b; Pecoraro et al., 1996; Sodora et al., 1994). While the nucleotide sequences of computer-predicted RREs exhibit sequence similarities ranging from ~52% to ~87% among selected different isolates upon comparison with the prototypic FIV strain, Petaluma (Supplementary Fig. 1), the principal secondary RNA stem-loop structure within the FIV RRE appears highly conserved across different strains of FIV (Supplementary Fig. 1), and is likely vital for viral function. However, the precise contribution of the FIV RRE structure, in particular the functional impact of sequence diversity within this element, remains unclear.

Herein, we present a detailed analysis of the predicted six-stem-loop containing secondary RNA structure within the RRE of a variant of FIV Petaluma, FIV-Ch. The functional relevance of specific features within the secondary structure of the RRE of this strain was investigated using a compensatory mutagenesis strategy in which specific nucleotide residues in each stem were substituted leading to disruption or restoration of base pair(s) with corresponding effects on RNA helical stability and ensuing viral gene expression, viral release and replication. The present studies, using both the FIV CAT-RRE reporter system and the full-length FIV-Ch, reveal that viral gene expression, production and replication are differentially affected by the stability of stem structures within the RRE, highlighting a critical RRE domain lying within the central region of the stem-loop structure. These studies provide additional insight into the common structural features that are critical for RRE function.

Results

Computational analysis of the RRE structure of FIV-Ch

Like HIV-1, FIV-Ch expresses a Rev protein that is encoded by a spliced transcript from regions flanking the env gene. In contrast to HIV, the putative FIV RRE overlaps the 3′ end of the env open reading frame and beginning of the LTR (Fig. 1A). An Mfold program-based analysis (Mathews et al., 1999; Zuker, 2003) predicted that the secondary RNA structure of the RRE region in the FIV-Ch strain possessed two major stem-loops. This predicted RRE structure included a single elongated hairpin termed stem-loop A (SLA), which contains six smaller stem-loop sub-domains, and its neighbouring branched stem-loop termed stem-loop B (SLB) (Fig. 1B). The terminal nucleotides of the env sequence were located within an internal loop A4 (LA4) (Fig. 1C) of SLA separating stem A4 (SA4) and SA5. Comparison with other FIV strains’ RRE disclosed that SLA was highly conserved structurally (Supplementary Fig. 1), while SLB varied widely depending on the individual viral strain (data not shown). Given the consistency and similarity of the RRE SLA across different FIV strains, its role was investigated in viral expression and release by means of a mutagenesis approach.

Fig. 1.

FIV-Ch genome and computer-predicted secondary structure of FIV-Ch RRE. (A) Schematic representation of genome organization and relative positions of open reading frames (ORF) and RRE RNA of FIV-Ch. The FIV-Ch genome is represented linearly with ORF regions depicted as boxes; the rev gene is depicted as a dotted box, the LTR is depicted by a solid thick line, the RRE RNA depicted as empty box. (B) The computer -predicted secondary structure of RRE contains two hairpins: stem-loop A (SLA) and stem-loop B (SLB). The end of the env ORF is indicated by an arrow. Nucleotides are numbered by their position within the full-length proviral (FIV-Ch) plasmid. (C) Secondary structure of RRE SLA in this study. The stop codon “UGA” of the env gene present in LA4 is denoted in gray, italic lettering.

FIV CAT-RRE activity

To ensure the impact of the RRE structure on viral production was independent of any flanking RNA sequences, in particular env-mediated effects, the RRE fragment of FIV-Ch was cloned into a pUC-L7C-CAT vector to generate a FIV CAT-RRE reporter system together with a control plasmid in which the entire RRE was cloned in the antisense orientation (Fig. 2A). The CAT gene and the FIV-Ch RRE were flanked by retroviral splice donor and acceptor sites (Fig. 2A). After co-transfecting this construct together with the wild-type FIV-Ch or cloned rev plasmid into mammalian cells, the functional activity of the Rev and RRE interaction is proportional to the level of CAT activity in the target cells (Powell et al., 1997; Zou et al., 1997). The Mfold analysis predicted that the cloned RRE fragment formed the same secondary structure within the CAT system as observed in the full-length FIV genome (Fig. 1B). In addition, the FIV-Ch rev gene was also cloned and transfected into CrFK cells (Fig. 2B). Transfection of the cloned rev gave rise to increased levels of Rev protein compared with transfection of cells with the wild-type full-length FIV-Ch clone.

Fig. 2.

Diagrammatic representation of the CAT-RRE reporter system (Phillips et al., 1992). (A) In the pRRE(+) construct, splice acceptor (SA) and splice donor (SD) sites flank a CAT gene and a FIV-Ch RRE which is in the sense orientation. In the pRRE(−) construct, FIV-Ch RRE was inserted in an antisense orientation. (B) The cloned rev plasmid and FIV-Ch plasmid were transfected into CrFK cells, respectively. Cell lysate was harvested at 24 hours and analyzed. The western blot confirms the expression of the cloned rev gene using expression of Rev encoded by FIV-Ch virus as a control.

Trinucleotide or binucleotide compensatory mutations in each of the six stems within SLA of the RRE were created to define the biological functions of each stem within the secondary RNA structure. Wild-type and mutated RRE clones were concurrently transfected with plasmids containing either full-length FIV-Ch or merely the FIV rev. Trinucleotide mutations which disrupted base pairing in the stem of SA1 (SA1/78C, SA1/79C) caused reductions in CAT expression (<50%) that were restored to near wild-type levels following reversion to a sequence which restored base pairing (Fig. 3F). The introduction of similar disruptive substitutions in SA2 (SA2/81C, SA2/82C) (Fig. 3E), SA3 (SA3/84C, SA3/85C) (Fig. 3D), SA4 (SA4/87C, SA4/88C) (Fig. 3C) and SA5 (SA5/70C, SA6/71C) (Fig. 3B), respectively, resulted in even greater reduction of CAT activity (<5-40%) relative to the wild-type FIV-RRE, while restoration of base pairing with compensatory mutations (Fig. 3) permitted variable recovery of CAT activity depending on the individual clone. The bi-residue mutations introduced into SLA6 (SA6/73C, SA6/74C) caused less reduction of CAT activity than was observed for other mutated stems and function was largely restored by re-establishing base pairing (SA6/75C) (Fig. 3A). Accordingly, while complete elimination of SLA5 and 6 (SLA5/91C) caused a marked reduction in CAT activity (~80%, Fig. 3B), deletion of only SLA6 (SLA6/90C) had a more subtle effect (Fig. 3A). Taken together, the current results underscore the importance of maintaining base-pairing interactions in the stems of SA1, 2, 3, 4 and 5 to ensure FIV RRE function, whereas the structural integrity of SA6 was less significant.

Fig. 3.

Detection of secondary structure of SLA in the CAT-RRE reporter system. Each stem of SLA within the FIV-Ch RRE in the CAT-RRE reporter system was modified with a compensatory mutagenesis and/or deletion strategy for functional assessment of the secondary structure of the RRE. The CAT-RRE plasmids were co-transfected with the FIV-Ch proviral plasmid (FIV-Ch) or the cloned FIV rev plasmid (Rev). Expression concentrations of the CAT reporter protein were analyzed by ELISA. For quantification of CAT expression, the concentration of wild-type FIV-Ch CAT was set at 100% and mutant concentration were normalized to this value for (A) SA6, (B) SA5, (C) SA4, (D) SA3, (E) SA2 and (F) SA1. Relative CAT protein expression levels are presented as histograms. The values represent the means of results from duplicate transfections across two separate experiments.

Functional analysis of RRE within the context of the full-length FIV-Ch genome

Most studies of lentiviral Rev/RRE-mediated posttranscriptional regulation have focused on the rev gene and/or dissected its RRE and determined the ensuing effects on viral production (Felber et al., 1989; Fischer et al., 1999; Gallego et al., 2003; Kjems et al., 1992; Malim et al., 1989a; Malim et al., 1989b; Mishra et al., 2006; Olsen et al., 1990; Wilkinson et al., 2000; Zuker, 2003). Having demonstrated the structural importance of each stem in the FIV CAT-RRE assay, the possible structural effects of the RRE within its natural context were investigated further in the full-length FIV-Ch genome. To examine the relative contribution of each stem sub-domain within SLA, wobble positions mapping to SA1, 2, 3, and 4 within the FIV env coding region were substituted and compensatory mutations were then introduced at positions involved in base pairing with the wobble positions, thus allowing interrogation of the effects of RNA secondary structure in these regions of the RRE without altering the amino acid sequence resulting from translation of the overlapping env. Furthermore, the binucleotide and trinucleotide mutations tested in the FIV CAT-RRE assay were also examined in the context of the full-length FIV-Ch genome.

To define the identity and structural significance of an “A₈₇₈₅ A₈₉₂₈” mismatch in SA1, the “A₈₇₈₅” residue was substituted for a “G” to generate a “G₈₇₈₅ A₈₉₂₈” mismatch (SA1/54); the “A₈₉₂₈” residue was also substituted for “U” that removed the “A₈₇₈₅ A₈₉₂₈” internal loop and generated an “A₈₇₈₅ U₈₉₂₈” base pair (SA1/55). Both synonymous mutants resulted in a ~50% down regulation of viral p24 expression and supernatant reverse transcriptase (RT) activity (Fig. 4A). To delineate further the conformation of SA1 that might be important for FIV expression, single base pair (synonymous) disruptive mutations (SA1/56 and SA1/57) and triple adjacent base pairs disruptive mutations (SA1/78, SA1/79, SA1/80) (non-synonymous) were introduced and/or then recovered with compensatory mutations. Mutations which disrupted the base pairing at these sites in SA1 stem resulted in both decreased RT activity in supernatant and p24 immunoreactivity (IR) in cell lysates. Restoration of base pairing permitted recovery of SA1 function (Fig. 4A). When similar mutational strategy was performed to examine the functional importance of SA2 (Fig. 4B), SA3 (Fig. 5A), SA4 (Fig. 5B), SA5 (Fig. 6A), and SA6 (Fig. 6B), similar findings were observed in that synonymous single nucleotide or trinucleotide stem-disrupting mutations reduced p24 and RT abundance, with full or partial restoration of activity following restoration of base pairing by compensatory mutation. In addition, the structural significance of an “A” bulge present within SA4 was examined by removing the bulge through inserting a “U” at its opposite position (Fig. 5B; SA4/69). These studies revealed that the structural feature of the “A” bulge was of limited importance for viral expression. In concordance with the previous findings indicating that SA6 was dispensable for RRE activity in the context of the FIV CAT-RRE assay system, mutations disrupting the base pairing within SA6 had only modest effects on p24 and RT activity when compared with the other stems. Furthermore, while complete deletion of SL5 and SL6 (Fig. 6B, SLA5/91) almost completely abolished p24 expression and RT activity, deletion of SLA6 alone had only a modest effect with reduction in p24 levels and RT activity to 70% and 42% of wild-type, respectively (Fig. 6B, SLA6/90). These results confirmed the dispensability of SLA6 for RRE activity and demonstrated the critical role of SA1-5 in this RRE model. Interestingly, while base-pairing disruptive mutations resulted in reduction of p24 expression and RT activity, there was a corresponding enhancement of FIV Rev expression (Fig. 4A, B; 5A, B; 6A, B).

Fig. 4.

Mutational analysis of SA1 and SA2 within the RRE of the full-length FIV-Ch. CrFK cells in 24 well plates were transfected with full length FIV-Ch molecular clones (1.25ug/well) containing RRE mutations in SA1 and SA2. The cell supernatants and proteins in the cell lysates were harvested 24 hours post-transfection. The released virus was assessed by RT activity assay in the cell supernatants and expression of the viral structural protein p24 was assessed by western blot assay. Western blot membranes were probed with a monoclonal anti-p24 antibody and polyclonal rabbit anti-Rev antibody. β-actin was used as a loading control. The values are obtained from triplicate experiments that were normalized to wild-type FIV-Ch levels (set at 100%). Immunoreactivity (IR) represents p24 expression measured on western blot. (A) Depiction of tested RRE mutants with modifications in SA1 indicated. Without changing the amino acid (aa) sequence of Env, SA1/54, SA1/55 mutants tested the nucleotide identity and/or structural importance of an “A A” internal loop, and SA1/56, SA1/57 mutants test the importance of a “G C” base pair. Triple base pair compensatory mutations test importance of SA1, in which two aa of the Env were concurrently changed by nucleotide modification. (B) Depiction of RRE mutations within SA2. Single base pair and triple base pair compensatory mutations, without and with changes to the amino acid sequence of Env, respectively, were designed to test the importance of SA2 for FIV replication. Error bars show standard deviations.

Fig. 5.

Mutational analysis of SA3 and SA4 within the RRE of the full-length FIV-Ch. (A) Depiction of RRE mutations within SA3. Single base pair and triple base pair compensatory mutations, without and with changes to the amino acid sequence of Env, respectively, were designed to test the importance of SA3 for FIV replication. (B) Depiction of RRE mutations within SA4 as in panel A. In addition, the structural importance of an “A” bulge was tested by inserting a “U” on its opposite side in the predicted RRE secondary structure to create an “A U” base pair and consequently remove the “A” bulge.

Fig. 6.

Mutational analysis of SA5 and SA6 within the RRE of the full-length FIV-Ch. (A) Depiction of RRE mutations within SA5. Triple base pair compensatory mutational analysis, without changing the amino acid sequence of Env, was designed to test the importance of SA5 for FIV replication. (B) Depiction of RRE mutations within SA6. Compensatory mutational analyses of two “A U” base pairs within SA6 were performed without changing the amino acid sequence of Env, which test the importance of SA6 for FIV replication. In addition, SA6 and SA5 deletion were also carried out to test importance of the two top stem-loops.

Secondary structure rather than a putative antisense protein modulates FIV replication in vitro

It has been posited that FIV expresses an 103 amino acid antisense protein (ASP) (Briquet et al., 2001), encoded by an antisense transcript in which the start codon of the ASP initiated at a nucleotide complementary to the “U₈₈₁₄” in SA3 of the FIV RRE (Fig. 7). To determine whether the structural features of the RRE and/or the putative ASP play a role in mediating FIV expression, the “C₈₈₁₂ G₈₉₀₃” base pair containing the start codon of the putative ASP and the wobble nucleotide “C₈₈₁₂” for env was mutated within the FIV-Ch genome. The start codon was removed or restored by a compensatory mutation while ensuring that all introduced mutations were synonymous within env. All mutations which disrupted base pairing within SA3, with or without silencing the ASP start codon (Fig. 7B, SA3/47, 65, 45), resulted in reduced p24 expression and RT activity (>50% and >70% reduction, respectively) compared with the wild-type FIV-Ch clone. However, compensatory mutations which restored the base pairing within SA3 but maintained silencing of the ASP start codon largely increased RT activity and p24 expression to wild-type levels (Fig. 7B, SA3/48, 46). Hence, a clear correlation between viral gene expression and RRE secondary structure was evident but the hypothesized ASP did not appear to exert measurable effects on viral production.

Fig. 7.

Viral RNA structure-based mutagenesis to evaluate a putative FIV encoded antisense protein (ASP). Compensatory mutational analysis was performed on a single “C G” base pair within SA3 of RRE which represents the 3^rd nucleotide of the putative ASP start codon. This mutation was generated such that it did not alter the predicted amino acid sequence of the overlapping env. (A) A schematic figure shows the positions of the env gene, RRE RNA and the putative ASP ORF. The putative ASP was predicted to be a 103 aa hydrophobic protein (Briquet et al., 2001). (B) Analysis of RRE mutants with modifications in SA3 within the full length FIV-Ch virus context as in Figs. 4, 5, and 6. A series of mutants were generated which tested the effect of disruption of base pairing within SA3 and restoration of base pairing by compensatory mutations, concurrently with or without a silenced ASP start codon. “+” and “−” represent activated and silenced start codon of ASP, respectively.

Structural features of the RRE are critical to FIV expression and replication

A daughter FIV-Ch clone with a base-pair disrupting single nucleotide mutation within SA3, FIV-SA3/45 (Fig. 7B), was selected for further examination of structural and functional relevance in the context of an in vitro viral infectivity assay. Both mutant and parent wild-type viruses were initially transfected into CrFK cells to generate viruses for subsequent infection. Feline T-lymphoblastoid cells (MYA-1) were infected, based on matched RT activity/vol, with FIV-SA3/45 or FIV-Ch, respectively. The FIV-SA3/45 clone showed a marked reduction in corresponding env (Fig. 8A), gag (Fig. 8B) and pol (Fig. 8C) transcript abundance at days 4, 8 and 12 post transfection compared with the wild-type parent FIV-Ch. These investigations were extended by examining p24 expression in cell lysates at each time point (Fig. 8D), which disclosed that like the transcript abundance, p24 immunoreactivity was high for the wild-type FIV-Ch at all time points but the FIV-SA3/45 mutant displayed a lower, albeit consistent, increase in expression over time. These results were confirmed by showing that the RT activity for the wild-type parent FIV-Ch virus peaked at days 7 and 8 but was sustained until day 12. In contrast, the daughter clone FIV-SA3/45 caused a consistent rise in RT activity reaching approximately 50% of the wild-type virus’ levels by day 12. Importantly, the RRE region of FIV-SA3/45 RNA collected from day 12 was sequenced and the results from 20 randomly selected clones showed that none had reverted to the wild-type sequence or incurred complementary mutations which restored SA3 base pairing (data not shown). Similar results were observed in CrFK cells harvested 2 and 4 days following transfection with FIV-SA3/45 or wild-type FIV-Ch molecular clones (Supplementary Fig. 2). Hence, the daughter mutant virus, FIV-SA3/45, was also infectious but exhibited diminished viral gene expression and replicative capacity. Given that the single nucleotide substitution in FIV-SA3/45 was synonymous with respect to env, only a structural modification in the RRE could be responsible for the differences from the wild-type virus observed in viral infectivity assays. These latter findings underscored the critical role for structural integrity within the RRE in the viral life cycle.

Fig. 8.

Secondary structure of RRE affects FIV replication. CRFK-transfection generated wild-type FIV-Ch and FIV-ST3/45 viruses were normalized with RPMI medium (15% FBS, 1% pen/strep) to the same RT activity level (~9000 cpm) and subsequently used to infect a feline T-lymphoblastoid cell line (MYA-1) (800 μl virus solution + 200 μl of ~2×10⁶ cells per well) in 24 well plates for a time course experiment. At 6 hours post-infection, supernatants were removed and 1 ml/well of fresh RPMI medium was added. On day 2 post-infection, 170μl/well of MYA-1 cell supernatant was harvested for the RT activity assay and 200μl/well of fresh medium was added. Sampling was repeated daily. A mock control was performed in parallel. At days 4, 8 and 12, total RNA and proteins were extracted from MYA-1 cells for viral mRNA semi-quantitative real time RT-PCR and p24 western blot assay, respectively. (A-C) show viral env, gag, and pol transcript levels, respectively. (D) Viral p24 expression corresponding to FIV-Ch and FIV-SA3/45 on day 4, 8 and 12 is shown. (E) Likewise, RT activity of FIV-Ch and FIV-SA3/45 strains are displayed for the entire 12 day time course with marked differences between viruses.

Discussion

The present studies defined the functional and structural relevance of a computationally-predicted RNA secondary structure of the FIV RRE by mutational analyses. Minimal changes in the primary RRE (RNA) sequence exerted substantial effects on viral expression and infectivity. These data also revealed that the formation of base-paired stem regions within the highly structured RRE, except stem A6, were critical for its function in the FIV CAT-RRE system and in the context of the full length virus in terms of viral structural gene expression, release and replication. Mutagenesis of one or more nucleotide positions within predicted stem regions of the FIV RRE sequence, which might alter secondary structure, led to reduced viral expression depending on the domain and nature of the mutation(s). In most circumstances, co-variation mutations which would be predicted to restore base-pairing in these regions restored viral function, thus supporting our FIV-Ch RRE model. These mutagenesis-related effects were not necessarily associated with the number of modified nucleotide positions. In fact, a single synonymous mutation (FIV-SA3/45) caused a marked reduction in viral expression, release and infectivity.

Importantly, it is possible that the effects of the mutations introduced to disrupt base pairing within particular stem regions in this study resulted in changes to the RRE secondary structure that extended outside of the particular stem-loop region targeted. To explore this concern, the Mfold predicted secondary structures and free energies of the mutated full-length RRE clones generated in this study were examined. This analysis revealed that the disruptive mutations within stem A6 or complete deletion of SLA6 only affected the predicted secondary structure of the RRE above SLA5. Conversely, disruptive mutations in the other stems within the RRE gave rise to several secondary structure predictions that were significantly different from the wild-type, even in regions distal from the mutation site. These findings may explain the greater functional impact of SA1-5 mutation compared to SA6 seen in our assays. Mutants in which base-pairing was restored within the SLA stems by the introduction of compensatory mutations were also subjected to Mfold analysis. The predicted secondary structures and free energy of RRE mutants SA1/80(ΔG = −52.00), SA2/83(ΔG = −51.80), SA3/86 (ΔG = −52.00) SA4/89(ΔG = −52.40), SA5/72(ΔG = −51.80), SA6/75(ΔG = −52.20) were comparable to that of the wild-type RRE (ΔG = −52.20), thus the moderately reduced functionality observed in some of these mutants is likely unrelated to stability.

Interestingly, Rev expression was inversely correlated with the extent of viral expression and release in the present assays. These results were to be expected given the proposed involvement of Rev in the negative regulation of viral mRNA splicing. Mutations within the RRE structure that interfere with CAT and viral protein expression likely mediate their effects by interfering with Rev RRE binding, leading to greater nuclear retention of the viral mRNAs, more complete viral mRNA splicing, and a greater proportion of multiple spliced viral mRNA species reaching the cytoplasm for translation. Since the rev message itself is multiply spliced, there will likely be a greater amount of rev mRNA available for translation, which in turn could lead to more Rev protein production, as evidenced in the current studies.

The RRE has been identified in most lentiviruses although its secondary structure varies widely depending on the individual virus. Indeed, its relative contribution to viral replication is also virus (and viral strain) specific despite its conserved function (Lesnik, Sampath, and Ecker, 2002; Seelamgari et al., 2004). It was previously shown that HIV and FIV Revs have no cross function on their RREs (Kiyomasu et al., 1991). The predicted RRE structures within different FIV strains showed similar secondary structural features (Supplementary Fig. 1), with a predominant hairpin domain (SLA) that was highly conserved in that all FIV strains contained a long unbranched hairpin with multiple stems and loops with similar sizes and locations (Phillips et al., 1992; Zou et al., 1997). Moreover, the amino acid sequence of FIV-Oma Rev protein has approximately 80% identity with the FIV-PPR Rev in the arginine rich basic domain near the C-terminus (Zou et al., 1997), which is thought to be a RRE binding motif in other lentiviruses (Hope et al., 1990; Ihm et al., 2009). Our calculations also suggested that the predicted amino acid sequences of Rev of FIV Petaluma and FIV-Ch strain share over 75% identity with that of the FIV-PPR Rev. Regardless of the exact nature of the Rev binding site within the RRE, its overall structural conservation among different FIV strains and the current functional observations call attention to the RRE as a pivotal determinant of viral expression.

It has been reported that HIV-1 viral structural gene mRNA and protein expression is dependent on the Rev protein (Felber et al., 1989; Fukumori et al., 1999; Hadzopoulou-Cladaras et al., 1989; Perales, Carrasco, and González, 2005). Several studies have indicated that the binding of Rev to the RRE motif and the interaction of the nuclear export signal (NES) domain of Rev with host cell Crm1 protein are necessary for efficient cytoplasmic export, translation and encapsidation of RRE-containing mRNAs (Askjaer et al., 1998; Brandt et al., 2007; Elfgang et al., 1999; Malim et al., 1991). The present results, derived from experiments involving in vitro infection of a feline T cell line (MYA-1), demonstrated a reduction in the expression of viral structural genes such as gag, pol and env at the mRNA and protein levels and reduced viral release after infection with the FIV-SA3/45 mutant (containing a single base-pair disruption in SA3) compared to the wild-type FIV-Ch. A plausible mechanism for the suppression of structural gene expression could be that the single residue change within the RRE structure interfered with its binding to Rev, resulting in the nuclear retention of the viral mRNAs. The nuclear retention of the transcripts in turn might lead to diminished expression of the mutant viral structural mRNAs and as a consequence, reduced production of viral structural proteins and viral release.

Virus-encoded antisense RNAs and proteins are well recognized for some herpes and retroviruses but their expression by lentiviruses (HIV-1 and FIV) has been more contentious (Arnold et al., 2006; Briquet et al., 2001; Briquet and Vaquero, 2002; Landry et al., 2007; Vanhee-Brossollet et al., 1995). Indeed, the HTLV-I encoded ASP, termed HBZ, has been described and verified by several groups (Arnold et al., 2006; Gaudray et al., 2002; Kuhlmann et al., 2007; Matsumoto et al., 2005; Mesnard, Barbeau, and Devaux, 2006) while ASP expression by HIV-1 and FIV has also been reported (Briquet and Vaquero, 2002; Tagieva and Vaquero, 1997; Vanhee-Brossollet et al., 1995). In the present study, FIV antisense RNA was detectable in FIV infected cell lines and animal tissue (data not shown). However, the experimental constructs in which the ASP start codon (located within SA3) was removed but base-pairing within SA3 was preserved gave rise to p24 expression and RT activity levels that were equivalent to wild-type (see Fig. 7B, SA3/48). These results further suggest that the integrity of RRE secondary structure, rather than the putative ASP, controls FIV expression. Moreover, these latter observations were reinforced by the failure to detect ASP immunoreactivity in cells infected with the wild-type virus and the absence of an anti-ASP serological response in FIV-infected adult animals (data not shown). However, these observations do not definitively exclude the potential expression of ASP and the possibility that ASP plays a role in FIV infection in vivo.

The specific contribution of the RRE to virulence is difficult to surmise but the present studies point to a key role for the RRE in FIV infection. Like HIV-1, FIV replication and its ability to infect and deplete lymphocytes are key determinants of virulence (Hosie et al., 2002; Sutton et al., 2005; Tompkins and Tompkins, 2008). Given these common properties among lentiviruses together with a shared secondary structure responsible for transport of mRNAs, it appears the RRE might be critical for lentivirus virulence. Nonetheless, in vivo studies remain to be performed to confirm the functional importance of the RRE structure to virulence. Both HIV-1 and FIV exhibit complex RREs but high resolution mutagenesis studies have not been performed for the HIV RRE, as described in the present studies of the FIV SLA. The SLB domain was not examined herein but preliminary findings suggest that this region is not as critical for viral infection (data not shown). Future studies will need to address the effects of RRE mutagenesis during in vivo FIV infection, and also determine the precise location of the Rev binding site and interacting host proteins to glean a more in-depth appreciation of RRE effects on FIV infection.

Treatment of HIV/AIDS has improved dramatically over the past decade but there is an ongoing need to improve and extend current antiretroviral regimens in large part because of growing drug resistance as well as the emergence of new strains of HIV-1 (Jain and Mondal, 2008; Kiertiburanakul and Sungkanuparph, 2009; Zdanowicz, 2006). The effectiveness of the mutagenesis studies in terms of interrupting FIV gene expression and replication raises the possibility of targeting individual RRE sub-domains as a therapeutic strategy. Several locations within the predicted stems of SA1-5 had strong inhibitory effects on viral gene expression and RT activity upon mutagenesis and thus future in vitro and in vivo studies might target these domains using oligonucleotides as a therapeutic approach. Indeed, previous studies suggest that disrupting Rev-RRE interactions might be a potential route for future antiretroviral development (Koeller and Wong, 2000; Ward, Rekosh, and Hammarskjold, 2009).

Conclusion

These studies highlight the importance of structural properties of the RRE in terms of controlling FIV viral structural gene expression and replication while also underscoring the region’s potential significance as a future molecular target for antiretroviral therapies with sequence-specific tools. Mutations within the RRE might also represent opportunities for the generation of attenuated virus strains, which could be used in future vaccine development.

Materials and methods

Full-length proviral plasmid construction

Construction of an infectious full-length FIV-chimera (FIV-Ch) proviral plasmid has been described previously in which the FIV Petaluma env was replaced with the cloned FIV V1CSF env (Genbank accession number: GQ420651.) (Johnston et al., 2000). All oligonucleotide primers used in this study are presented in Table 1. Modifications based on predicted RRE sequence were introduced into the FIV-Ch proviral construct by inserting overlap extension polymerase chain reaction (PCR) products between SpeI and BstBI restriction sites or using the Quikchange II XL Site-Directed Mutagenesis Kit (Stratagene). Primer pairs 64F/64R, 66F/66R, 98F/98R were used to construct clones HN55, HN57 and HN79, respectively, according to Quikchange Mutagenesis Kit instructions. Briefly, thermal cycling reactions were performed with the template of the FIV-Ch plasmid for 1 cycle (95°C for 1 min), 18 cycles (95°C for 50 sec, 55°C for 50 sec, and 68°C for 16 min), and 1 cycle (68°C for 18 min). Other sequence modifications were performed using an overlap-extension PCR method. Briefly, two flanking master primers, PHN36F-9 (forward) and PHN36F-4 (reverse), and two internal primers, which introduced the mutation(s) of interest, were used with FIV-Ch or mutated FIV-Ch plasmid as template to generate two fragments for each mutant for the first PCR (Heckman and Pease, 2007). The two overlapping fragments extracted from each mutant were then fused together in a subsequent extension reaction and amplified by secondary PCR with outside primers PHN36F-9 and PHN36F-4. Secondary PCR generated fragments were digested with SpeI and BstBI and subsequently used to replace the corresponding wild-type SpeI/BstBI fragment in FIV-Ch. The PCR-derived regions introduced into constructs were sequenced completely to ensure that only the modifications of interest were present. Details of the introduced modifications into the predicted RRE structure of FIV-Ch are presented in the respective figures.

TABLE 1.

PCR primers used in this study

Construct / gene	Primersa	Sequenceb	Sensec
	PHN36F-4	TCGTAAACAGTCCCTAGTCCATAAGC	−
	PHN36F-9	GTTCTGGCAACCCATCAAGAAGC	+
	PHN92-1FM	GGCGGCCCAGATCTGATATCATCGATGAGTGTTATTATTGA TTTTATGTTTAC	+
	PHN92-1R	TTCGAGCTCGGTACCCCTTTCTTCTTTCTTCTTCTTC	−
	PHN92-3F	GGGTACCGAGCTCGAATTCGAGCTCGGTACCC	+
	PHN91-2RM	AGGTTTTTTAAAGCAAGTAAAACCTC	−
	PHN92-1OF	GGCGGCCCAGATCTGATATCATCGATCTTTCTTCTTTCTTCTT CTTCTTTGTC	−
	PHN92-1OR	TTCGAGCTCGGTACCCGAGTGTTATTATTGATTTTATG	+
SA1/54	PHN63F	ACAAATGGA G TTGAGGAGAAATGGTAGG	+
	PHN63R	ATTTCTCCTCAA C TCCATTTGTGGTTG	−
SA1/55	PHN64F	GAATCCATTTCGAATCAA T TCAAACTAATAAAGTATGTATTG	+
	PHN64R	CAATACATACTTTATTAGTTTG A ATTGATTCGAAATGGATTC	−
SA1/56	PHN65F	ACAAATGGAATT A AGGAGAAATGGTAGG	+
	PHN65R	ATTTCTCCT T AATTCCATTTGTGGTTG	−
SA1/57	PHN66F	GAATCCATTTCGAAT G AAATCAAACTAATAAAGTATGTATTG	+
	PHN66R	CAATACATACTTTATTAGTTTGATTT C ATTCGAAATGGATTC	−
SA2/59	PHN68F	AGAAA C GGTAGGCAATGTGGCATGTC	+
	PHN68R	ATTGCCTACCGTTTCTCCTCAATTCC	−
SA2/60	PHN69F	CATATGAATCC T TTTCGAATCAAATC	+
	PHN69R	GATTTGATTCGAAA A GGATTCATATG	−
SA2/61	PHN70F	CATATGAATCC G TTTCGAATCAAATC	+
	PHN70R	GATTTGATTCGAAA C GGATTCATATG	−
SA3/62	PHN71F	TAGGCAATG C GGCATGTCTGAAAAAG	+
	PHN71R	GACATGCC G CATTGCCTACCATTTC	−
SA3/63	PHN72F	GGAAGGTATGTC T TATGAATCCATTTC	+
	PHN72R	GAAATGGATTCATA A GACATACCTTCC	−
SA3/64	PHN73F	GGAAGGTATGTC G TATGAATCCATTTC	+
	PHN73R	GAAATGGATTCATA C GACATACCTTCC	−
SA3/45	PHN36F-1	GGTAGGCAATGTGG A ATGTCTGAAAAAG	+
	PHN36F-2	CTTTTTCAGACAT T CCACATTGCCTACC	−
SA3/46	PHN36F-8	CGAAATGGATTCATATGA A ATACCTTCCTC	−
SA3/47	PHN36F-7	CGAAATGGATTCATATGA G ATACCTTCCTC	−
SA3/48	PHN36F-5	GGTAGGCAATGTGG G ATGTCTGAAAAAG	+
	PHN36F-6	CTTTTTCAGACATCC C ACATTGCCTACC	−
	PHN36F-7	CGAAATGGATTCATATGA G ATACCTTCCTC	−
SA3/65	PHN74F	TAGGCAATGTGG G ATGTCTGAAAAAG	+
	PHN74R	GACAT C CCACATTGCCTACCATTTC	−
SA4/66	PHN75F	GTCTGAAAAAGA A GAGGAATGATGAAG	+
	PHN75R	CTTCATCATTCCTC T TCTTTTTCAGAC	−
SA4/67	PHN76F	CTGAGTTCTTC G CTTTGAGGAAGG	+
	PHN76R	CCTTCCTCAAAG C GAAGAACTCAG	−
SA4/68	PHN77F	CTGAGTTCTTC T CTTTGAGGAAGG	+
	PHN77R	CCTTCCTCAAAG A GAAGAACTCAG	−
SA4/69	PHN78F	CTGAGTTCTTCC T CTTTGAGGAAGG	+
	PHN78R	CCTTCCTCAAAG A GGAAGAACTCAG	−
SA5/70	*PHN79F	GGAATGATGAAGT TAG TCAGACTTATTTTAT	+
	*PHN79R	AAATAAGTCTGA CTA ACTTCATCATTCC	−
SA5/71, 72	*PHN80F	ATTTTATAAGGGA CTA ACTGTGCTGAG	+
	*PHN80R	CAGCACAGT TAG TCCCTTATAAAATAAG	−
SA6/73	*PHN81F	CTCAGACT AT TTTTATAAGGGAGATAC	+
	*PHN81R	CTCCCTTATAAAA AT AGTCTGAGATAC	−
SA6/74	*PHN82F	CAGACTTATTTTA AT AGGGAGATACTG	+
	*PHN82R	CAGTATCTCCCT AT TAAAATAAGTCTG	−
SA6/75	*PHN83F	CAGACT AT TTTTAATAGGGAGATACTG	+
	*PHN83R	CTCCCT AT TAAAAATAGTCTGAGATAC	−
SA1/78, 80	*PHN97F	ACAAATGGAA AAC AGGAGAAATGGTAGG	+
	*PHN97R	ATTTCTCCT GTT TTCCATTTGTGGTTG	−
SA1/79	*PHN98F	GAATCCATTTCGAAT GTT ATCAAACTAATAAAGTATGTATTG	+
	*PHN98R	CAATACATACTTTATTAGTTTGAT AAC ATTCGAAATGGATTC	−
SA2/81	*PHN99F	AGAA TAC GTAGGCAATGTGGCATGTC	+
	*PHN99R	ATTGCCTAC GTA TTCTCCTCAATTCC	−
SA2/82, 83	*PHN100F	CATATGAATC GTA TTCGAATCAAATC	+
	*PHN100R	GATTTGATTCGAA TAC GATTCATATGAC	−
SA3/84	*PHN101F	GCAATGTGG GTA GTCTGAAAAAGAGGAGG	+
	*PHN101R	GAC TAC CCACATTGCCTACCATTTC	−
SA3/85, 86	*PHN102F	GGAAGGT TAC TCATATGAATCCATTTC	+
	*PHN102R	ATTCATATGA GTA ACCTTCCTCAAAGGG	−
SA4/87	*PHN103F	GTCTGAAAAAGA CCT GGAATGATGAAG	+
	*PHN103R	ATCATTCC AGG TCTTTTTCAGACATG	−
SA4/88, 89	*PHN104F	CTGAGTTCT AGG CTTTGAGGAAGG	+
	*PHN104R	CCTCAAAG CCT AGAACTCAGCAC	−
SLA6/90	*PHN105F	GAAGTATCTCAGACGGGAGATACTGTG	+
	*PHN105R	CACAGTATCTCCCGTCTGAGATACTTC	−
SLA5/91	*PHN106F	AATGATGAATGTGCTGAGTTCTTCCC	+
	*PHN106R	GAACTCAGCACATTCATCATTCCTCCTC	−
FIV pol	FIV-pol 2F	TACTTCTAGAGAAGCCTGGGAATC	+
	FIV-pol 2R	CTGCTTTTCCTAGCTTTCTACCTCC	−
FIV gag	FIV-gag 1F	GGTTATTTGCGATTTACAAGAAAGAAGAG	+
	FIV-gag 1R	GGACACCATTTTTGGGTCAAGTGC	−
FIV env	FIV-env F	TTGTCAAAGAACACAGAGTCAGCC	+
	FIV-env R	CCATCTACATCTAATTCTAAACCTTGC	−
GAPDH	GAPDH-F	AGCCTTCTCCATGGTGGTGAA	+
	GAPDH-R	CGGAGTCAACGGATTTGG	−
FIV-Ch rev	PHN108-1F	GGGCCCGGGATCCACCGGTCGCCACCATGGCAGAA GGATTTGCAGCC	+
	PHN108-1R	CTTCTTCTTTGTCTTTTCCTTTTACCTGCATTTCCTTC TTCCAG	−
	PHN108-2F	CTGGAAGAAGGAAATGCAGGTAAAAGGAAAAGACA AAGAAGAAG	+
	PHN108-2R	GGCGGATACCCGCGGCCGCCTAGTCCATAAGCATTCT TTCTATTTCTTC	−

^aAsterisk refers to the primer used to construct modifications in both proviral plasmid and CAT-RRE cDNA.

^bViral sequences are underlined. Mutated nucleotides are italicized, bold and without underline.

^cRefers to the sense of the oligonucleotide in reference to the plus-sense viral RNA

FIV-Ch CAT-RRE reporter plasmids construction

A FIV-Ch CAT-RRE expression plasmid was generated based on the previously described construct FIV-PPR CAT-RRE (Phillips et al., 1992). Modifications were introduced into the RRE of FIV-Ch CAT-RRE plasmid by using overlap extension PCR-based mutagenesis and standard cloning techniques (Heckman and Pease, 2007). Briefly, to construct FIV-Ch CAT-RRE, a 382-bp fragment from FIV-Ch (bases 8655-8992) containing the predicted RRE was amplified by PCR using primers PHN92-1FM (containing BgIII site) and PHN92-1R; another 206-bp fragment containing XhoI was amplified from the template FIV-PPR CAT-RRE by PCR using primers PHN92-3F and PHN91-2RM; and the two generated fragments were fused together in a subsequent extension reaction and amplified by secondary PCR with outside primers PHN92-1FM and PHN91-2RM. The PCR product was digested with BgIII and XhoI and cloned into pUC-L7C-CAT vector, which was linearized with the same enzymes. This generated a FIV plasmid that contained the predicted RRE of FIV-Ch in the sense orientation. Similarly, the FIV Ch-RRE in the antisense orientation was cloned as a control into the plasmid using two primer pairs of PHN92-1OF/PHN92-1OR and PHN92-3F/PHN91-2RM, yielding a FIV-Ch-RRE (−). For construction of the sequence modifications in the RRE, two outside master primers PHN92-1FM and PHN91-2RM and two internal primers (in Table 1 with asterisk), which introduced the corresponding mutations into the RRE derived from the full-length FIV-Ch clone, were used with the FIV-Ch CAT-RRE template to generate two overlapping fragments by PCR. The two fragments were fused together in a subsequent overlap extension reaction and amplified by secondary PCR with outside primers PHN92-1FM and PHN91-2RM. The PCR fragments digested with BgIII and XhoI were ligated into pUC-L7C-CAT vector that was digested with the same restriction enzymes.

Proviral plasmid transfection and virus infection

Transfection of proviral clones was performed in Crandle feline kidney (CrFK) cells (ATCC). CrFK cells were cultured in MEM medium supplemented with 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 μg/mL) and subsequently seeded into 24 well plate at 2 × 10⁵ cells/well followed by incubation at 37 degrees Celsius with 5% CO₂ for 24 hours to achieve >90% confluency. 1.25 μg of proviral plasmid DNA and 2.5 μl of Lipofectamine 2000 (Invitrogen) in 250 μl of Opti-MEM medium (Gibco) were transfected into each well according to the manufacturer’s protocols. At 24 hours post-transfection, culture supernatants were collected for viral reverse transcriptase (RT) activity assay and cellular proteins were extracted with 100 μl/well cell lysis buffer for western blot analysis. Feline T-lymphoblastoid cells (MYA-1) seeded in 24 well plates at 2 × 10⁶/well in 200 μl were infected with 800 μl of reverse transcriptase-normalized FIV viruses generated from day 3 CrFK transfections and supernatants were sampled at various time points for RT measurement. All assays were performed in triplicate.

Reverse transcriptase assay

RT activity in culture supernatants was assayed as previously described (Johnston et al., 2000). 8 μl of culture supernatant was incubated with 32 μl of reagent buffer containing [α-32P] dTTP for 2 hours at 37°C. 30 μl reaction mixes were spotted on pencil labelled DE-81 paper squares (Whatman International, Ltd.). Papers were air-dried again for 30 min and washed 5 times for 10 min with gentle shaking in 100 ml 2XSSC and twice for 1 min in 50 ml 95% ethanol. Papers were air-dried for 30 min and RT levels were measured by liquid scintillation counting (TRI-CARB 2100TR, PACKARD, USA). All assays were performed in triplicate.

Western blot analysis

The expression of FIV p24 and Rev proteins in transfected CrFK cells or infected MYA-1 cells was assessed by western blot and p24 levels were subsequently quantified by band densitometry. Briefly, 15 μl of centrifugation cleared cell lysates were resolved by 12% SDS-PAGE, transferred onto a nitrocellulose membrane, and blocked with 10% milk in TBST (25 mM Tris-buffered saline and 0.1% Tween 20) for 1 hour. For FIV p24 protein detection, a nitrocellulose membrane was probed with a anti-FIV p24 (clone PAK3-2C1) mouse antibody at a 1:200 dilution in blocking solution with 5% milk overnight at 4 C° and subsequently with secondary HRP-conjugated antibody, at a 1:5000 dilution in blocking solution with 5% milk for 1 hour at room temperature. For Rev protein detection, a 1:500 dilution of rabbit anti Rev sera (Huisman et al., 2008) as the primary antibody and HRP-conjugated goat anti-rabbit IgG as secondary antibody were used to probe membranes. The membranes were then developed on film (Kodak, CAT 870 1302). Bands corresponding to host protein β-actin and viral protein p24 were quantified by using Quantity One software (Bio-Rad) and normalized against the β-actin. All assays were performed in triplicate.

Cloning of FIV-Ch rev

FIV-Ch is a strain that contains the FIV-V1CSF envelope gene on a background of FIV-Petaluma. Based on a high level of sequence identity of the putative FIV-Ch Rev gene (98.7%) with the FIV Petaluma Rev gene (151-amino acid) (UniProtKB: P20885-1), the predicted FIV-Ch rev gene was cloned by using an overlap strategy as above described. Briefly, primer pairs of PHN108-1F/PHN108-1R and PHN108-2F/PHN108-2R were used to generate two fragments of the rev gene with template of FIV-Ch for the first PCR. The two generated fragments were fused together in a subsequent overlap extension reaction and amplified by secondary PCR with outside primers PHN108-1F and PHN108-2R, generating an entire rev gene fragment flanked with BamH I and Not I restriction sites. The entire rev gene fragment was subsequently digested with BamH I and Not I and ligated into a pEGFP-N1 plasmid vector digested with the same enzymes (thus excising the egfp ORF).

CAT-RRE plasmid transfection and CAT ELISA assay

FIV CAT-RRE and FIV-Ch rev gene cloned or FIV-Ch wild-type proviral plasmids were co-transfected into CrFK cells as described above. For the co-transfection, 0.12 μg FIV-Ch rev gene cloned DNA plasmid or 1.25 μg of FIV-Ch proviral plasmid DNA (served as the Rev provider) together with 0.5 μg of CAT-RRE plasmid DNA and 2.5 μl of Lipofectamine 2000 (Invitrogen) in 250 μl of Opti-MEM medium (Gibco) were used for transfection of CrFK cells in 24 well plate as above described. Twenty-four hours post-transfection, CrFK cellular proteins were extracted by using 180 μl/well of CAT ELISA Kit (Roche) lysis buffer. To quantify CAT expression, 7μl of centrifugation cleared cell lysate was diluted with 143 μl of lysis buffer and CAT ELISA analysis was carried out with preparation of CAT enzyme standards according to the manufacturers’ recommended protocol. Measurement of the absorbance of each sample at 405 nm (reference wavelength: approx. 490 nm) was performed at 40 min using a microplate (ELISA) reader. The exact CAT concentration (ng/ml) for the calibration standards was calculated and CAT concentration of transfected samples was then determined by plotting the observed absorbance values on the standard calibration curve. The values were normalized to wild-type FIV-Ch levels (set to 100%). All assays were performed in duplicate and repeated a minimum of two times.

Real time RT-PCR assay

Total cellular RNA was isolated from transfected CrFK cells or cultured FIV infected or non-infected MYA-1 cells with TRIzol reagent (Gibco). First-strand cDNA was synthesized using Superscript II (Invitrogen) and 500ng/reaction of DNase-digested total RNA for subsequent RT-PCR assay as described previously (Johnston et al., 2000). The prepared first-strand cDNA was diluted 1:1 with sterile water and 5μl were used per PCR. The primers used in real-time PCR for FIV gag, pol and env gene quantification and host GAPDH gene are provided in Table 1. Semi-quantitative analysis was performed by monitoring in real time the increase of the fluorescence of SYBR Green dye on a Bio-Rad I-Cycler IQ detection system. A threshold cycle value for each gene of interest was determined as previously reported (Power et al., 2003). All data were normalized to GAPDH mRNA levels using the delta-delta Ct method and viral mRNA expressed as relative fold change (RFC) compared with control uninfected samples. All assays were performed in triplicate.

RNA secondary structure analysis

RNA secondary structures of RRE of FIV were predicted by computer analysis using Mfold version 3.2 (Mathews et al., 1999; Zuker, 2003).

Supplementary Material

01

Fig. 1S. Computer-predicted RRE secondary structures of 5 FIV isolates: FIV-Ch, PPR, Subtype C, Oma and EU117992. The relative RRE sequence identitiy values were obtained from comparison of RREs of the 5 FIV isolates with the prototypic FIV strain, Petaluma (set at 100%).

02

Fig. 2S. Secondary structure of RRE affects FIV expression during transfection. Wild-type FIV-Ch and the FIV-SA3/45 mutant were transfected into CrFK cells while a mock control was performed in parallel. At day 2 (48 hours) and day 4 (96 hours), supernatants, total RNA and proteins were extracted from the cultured CrFK cells for measurement of viral RT activity, mRNAs (by semi-quantitative real-time RTPCR) and p24 (by western blot), respectively. (A) These latter studies showed that viral env, gag, pol mRNA were comparatively reduced for the mutant FIV-SA3/45. (B) Western blots showed viral p24 expression of FIV-Ch and FIV-SA3/45 strains with a reduction for the latter virus. (C) The relative RT activity was increased for FIV-Ch compared with FIV-SA3/45 strains. All studies were performed as triplicate experiments that were normalized to those of wild-type FIV-Ch levels (set at 100%). Error bars show standard deviations.

Abbreviations

RRE

rev response element

FIV

feline immunodeficiency virus

SLA

stem-loop A

SA

stem A

ASP

antisense protein

RT

reverse transcriptase

p24 IR

p24 immunoreactivity

HIV

human immunodeficiency virus

SIV

simian immunodeficiency virus

BIV

bovine immunodeficiency virus

CAT

chloramphenicol acetyltransferase

CrFK

crandle feline kidney

MYA-1

feline T-lymphoblastoid cells

FRC

relative fold change