The ability of Sam68ΔC to selectively suppress expression of the 9 kb and 4 kb classes of HIV-1 mRNAs suggested that there is some unique feature that renders them susceptible to repression. Cellular mRNAs use the Tap export pathway, while HIV-1 9 and 4 kb RNAs are incompletely spliced and contain sequences preventing their export by Tap 2627282930313233. These incompletely spliced HIV-1 RNAs are exported from the nucleus via the interaction of HIV-1 Rev with the host protein Crm1 5634. Two possible explanations for repression of the 9 kb and 4 kb HIV-1 RNAs by Sam68ΔC are immediately apparent: either they contain unique RNA sequences recognized by Sam68ΔC or export via the Crm1 pathway marks the viral RNA for inhibition. To address this question directly, we examined the ability of Sam68ΔC to inhibit expression of HIV-1 Gag RNAs utilizing different nuclear export elements; the constitutive transport element (CTE) from Mason-Pfizer Monkey virus that interacts directly with Tap (Gag-CTE), or the RRE that requires Rev and Crm1 (Gag-RRE) (Fig. 1a) 35. Gag RNA generates a 55 kDa polyprotein that is subsequently processed by the viral protease into matrix (p17), capsid (p24), nucleocapsid (p9) and p6. Gag expression was measured by anti-Gag western blot in which the p55 precursor and p24 are detected. GagRRE expression is dependent upon Rev and is reduced to baseline by Sam68ΔC (Fig. 1b,c). Parallel western blots demonstrated that Sam68ΔC does not markedly alter Rev levels (Fig. 1b). In contrast to the Gag-RRE reporter, Sam68ΔC had no significant effect on expression from Gag-CTE (Fig. 1b,c). This demonstrates that it is not the Gag sequence, but rather the RRE and/or the Crm1 export pathway that dictates inhibition by Sam68ΔC.

Previously, we reported that Sam68ΔC expression induced accumulation of both unspliced viral RNA and Sam68ΔC in bundles at the outer periphery of the nucleus 21. As shown in Fig. 2a, in the absence of Sam68ΔC, HIV-1 env RNA is distributed throughout the cytoplasm. The formation of the perinuclear bundles upon co-expression of Sam68ΔC, as seen in Fig. 2b, suggested that Sam68ΔC might be sequestering the RNA from the translational apparatus. If true, disruption of these complexes should restore translation of the viral RNAs. We examined the effect of various agents which disrupt the cytoskeleton on the integrity of the Sam68ΔC/viral RNA perinuclear bundles, and the expression of viral proteins (Fig. 2c–f)). To minimize secondary drug effects, the minimum amount of each drug required to depolymerize its target in one hour was used (data not shown). While treatment of cells with either nocodazole or colcemid to disrupt microtubules had no effect on localization of viral RNA or Sam68ΔC (Fig. 2c,d), disassembly of the microfilaments by treatment with cytochalasin D or latrunculin B resulted in dispersal of both viral RNA and Sam68ΔC throughout the cytoplasm (Fig. 2e, f). The distribution of the viral RNA which has been released from the perinuclear bundles is similar to that seen in the absence of Sam68ΔC (Fig. 2a).

Next we questioned whether the released viral RNA was translated. Cells were transfected with Gag-RRE and Rev expression vectors in the presence or absence of Sam68ΔC. Forty-eight hours post-transfection, cells were treated with the drugs as before and synthesis of viral protein was monitored by incubation in the presence of ³⁵S methionine. Cells were lysed, the ³⁵S-labelled Gag immunoprecipitated, the immunoprecipitates run out on SDS-PAGE gels, and the gels exposed to a phosphorscreen overnight (Fig. 2g). The 55 kDa Gag polyprotein (p55) was not detected in the immunoprecipitates from the mock transfected cells but only in immunoprecipitates from cells transfected with Gag-RRE and Rev (Fig. 2g, lanes 1–5 versus 6). Therefore, the immunoprecipitation was specific and the p55 signal could be used as a measure of Gag-RRE translation. None of the drugs had any effect on the level of Gag expression in the absence of Sam68ΔC indicating that they had no significant effect on translation (Fig. 2g, lanes 1–6). None of the drugs induced expression of Gag in the presence of Sam68ΔC (Fig. 2g, lanes 7–12) despite latrunculin B/cytochalasin D shifting the subcellular distribution of both Sam68ΔC and the viral RNA. Therefore, integrity of the Sam68ΔC/viral RNA perinuclear bundles is not essential for translational repression. This observation suggests that Sam68ΔC inhibits viral RNA translation, not by changing its subcellular distribution, but blocking interaction with the translational apparatus. This effect could be achieved either through alterations in viral RNA structure or composition of the mRNP.

To confirm that Sam68ΔC was acting at the level of translation, cytoplasmic extracts were fractionated on linear sucrose gradients (Fig. 3). Fractions collected were subsequently analyzed for the presence of Sam68ΔC, ribosomal protein L26, actin mRNA and unspliced HIV env RNA. As shown in Fig. 3b, Sam68ΔC is predominately found at the top of the gradient but a significant amount is present in the heavier fractions consistent with previous observations of an association between Sam68 and ribosomes 36. Addition of EDTA to disrupt polysomes resulted in a shift in Sam68ΔC distribution to the top of the gradient, suggesting that there might be interaction of Sam68ΔC with polysomes. Parallel analysis of actin mRNA (Fig. 3c) revealed that the bulk of this RNA is found within heavy polysomes in the presence or absence of Sam68ΔC. EDTA addition induced a shift in distribution to lighter gradient fractions consistent with an association with polysomes. In contrast, unspliced HIV-1 env RNA underwent a shift in distribution from being predominately in the polysome fraction to predominately in the mRNP/monosome fraction in the presence of Sam68ΔC (Fig. 3d). The distribution of unspliced env RNA in the presence of Sam68ΔC significantly overlaps with that seen upon addition of EDTA consistent with a selective inhibition of translation of this mRNA.

Sam68 contains a number of well-defined domains that mediate specific RNA binding (KH-domain), non-specific RNA binding (RGG-boxes) or protein-protein interactions (proline-rich domains and tyrosine-phosphorylation sites) 9. We made a number of deletions of Sam68ΔC to define a minimal inhibitory mutant (Fig. 4a,b). As shown by western blots (Fig. 4b), all mutants were equally expressed and subsequent immunofluorescence microscopy confirmed that all were localized to the cytoplasm (data not shown). Previous analyses have determined that the region spanning amino acids (a.a) 269–321 is essential for the inhibitory property of Sam68ΔC 37. In our work, Sam68ΔCmin, spanning amino acids 14 to 300 with an internal deletion of amino acids 45 to 54 (encompassing an RGG box), was the minimal construct that retained significant inhibitory activity (Fig. 4b). In contrast, deletion of the first 28 (Sam68Δ28ΔC) or last 70 (Sam68:5–262) amino acids of Sam68ΔC resulted in a loss of inhibitory activity (Fig. 4b). These observations indicate that domains at the N- and C-terminus of Sam68ΔC are essential to its inhibitory activity.

Splicing, cleavage, and polyadenylation are tightly coupled events within the nucleus. However, in the case of HIV-1 RNA, numerous forms of viral RNA are generated, some of which have failed to undergo one or more of these steps. This is essential for the production of the HIV-1 structural proteins and therefore, the HIV-1 lifecycle. The HIV-1 env reporter, pgTat, expresses gp160/120 from the unspliced, cleaved mRNA (see Fig. 4a) 11. Previously, we showed that full-length Sam68 increases the amount of unspliced, cleaved pgTat mRNA available for polyadenylation, export and translation into gp160/120 11. We have also shown that restoring nuclear accumulation of Sam68ΔC by addition of a heterologous nuclear localization signal (NLS) converted the protein into a stimulator of Rev function comparable to full length Sam68 21. In light of these results, we wanted to assess whether Sam68ΔC inhibition is due to alterations in the abundance or processing of env RNA.

We examined the extent of splicing and cleavage of pgTat RNA by RNase protection assay (RPA). As illustrated in Fig. 4a, the RPA probe used in this analysis spans both the 3' splice site and the polyadenylation site, yielding four RNase protection products: unspliced, uncleaved (US-UC); unspliced, cleaved (US-C); spliced, uncleaved (S-UC); and spliced, cleaved (S-C) (Fig. 4a). Sam68ΔC did not cause any marked change in the amount of US-C RNA that could account for the loss of gp160/120 expression but a reduction in levels of S-C RNA was consistently observed (Fig. 4c) that might reflect effects on either viral RNA splicing or S-C RNA stability. In contrast, the mutant, Sam68:5–262, increased the abundance of US-C RNA (Fig. 4c). The stimulation of cleavage of unspliced RNA by Sam68:5–262 is consistent with the known activity of full length Sam68 in promoting RNA polyadenylation previously reported by our laboratory 11 and define that different domains of Sam68 are required for inhibitory versus stimulatory activity.

De-adenylation of mRNA is a common form of translational repression 38. To assess changes in the polyadenylation state of env mRNAs in the presence of Sam68ΔC, total RNA was extracted and the polyadenylated RNA isolated using oligo(dT)₂₅ beads. Given that cleavage and polyadenylation are tightly coupled processes, all cleaved RNAs are expected to have a poly(A) tail and bind to the oligo(dT)₂₅ column. Appearance of cleaved RNA in the poly(A)- fraction would be indicative of a de-adenylation event. Analysis of the poly(A)- and poly(A)+ fractions by RPA revealed that the fractionation was successful; uncleaved env RNAs (US-UC and S-UC) being predominantly in the poly(A)- fraction and cleaved versions (US-C and S-C) in the poly(A)+ fraction (Fig. 5a). Sam68ΔC did not shift the distribution of the unspliced RNAs between the fractions, indicating that the US-C form of pgTat RNA retained a poly(A) tail of sufficient size to bind to the column (Fig. 5a).

To assay whether there were any significant changes in poly(A) tail length, we used a

r

a

c

e

p

a

t

PABP1 has been shown to be an important promoter of mRNA translation through its direct interaction with eIF4G and resultant indirect interactions with the eIF4E cap binding protein 4041. Therefore, we examined PABP1 association with pgTat mRNA by RNP immunoprecipitation (RIP). PABP1 levels were not affected by expression of the Sam68 proteins and there was no change in epitope availability since precipitation of PABP1 was similar for each sample (Fig 6a). RNA analysis determined that pgTat RNAs were specifically precipitated with PABP1 as compared to control rabbit IgG RIPs (Fig. 6b). The S-C form of pgTat was the major isoform pulled down with anti-PABP-1 antibody but US-C RNA was also detected in the immunoprecipitates. Unexpectedly. some immunoprecipitation of both US-UC and S-UC was also detected particularly in the presence of the Sam68Δ28ΔC and Sam68:5–262. The reason for this is unclear but might reflect an indirect interaction of PABP1 with these RNAs (that lack a polyA tail). PABP1 association was quantitated by comparing the ratio of US-C to S-C pgTat RNA in the RIP compared to the input (Fig. 6c). Sam68Δ28ΔC and Sam68:5–262 had no significant effect on US-C RNA association with PABP1. Sam68ΔC and Sam68ΔCmin significantly reduced US-C RNA association with PABP1 (p < 0.05, Fig. 6b,c). Expression of Sam68ΔC also had no effect on the immunoprecipitation of RNPs containing the S/C form of pgTat RNA, consistent with Sam68ΔC being able to discriminate between the US-C and S-C forms of the viral RNA.

To confirm that Sam68ΔC induced similar changes in PABP1 association in the context of the HIV-1 provirus, the experiment was repeated using a replication inactive form of the virus (HxBruR^-/RI^-). Sam68ΔC strongly reduced both p24 and gp160/120 expression from this construct (Fig. 7). However, the mutants have slightly different effects on the provirus than on the env reporter construct pgTat. While Sam68:5–262 enhances expression of gp160/120 and p24, Sam68Δ28ΔC and Sam68ΔCmin have little effect. None of the mutants had any effect on Rev or PABP1 protein levels (top band, Fig 7, Fig. 8a). PABP1 was immunoprecipitated uniformly in all samples (Fig. 8a). The proviral RNA associated with PABP1 was analyzed using an RPA probe that spans the 5' splice site within Gag, yielding 2 protection products. One protection product corresponds to the 9 kb and the other to the spliced forms (2 and 4 kb classes) of HIV-1 mRNAs. As translation of both the 9 kb (p24) and 4 kb (gp160/120) classes of viral RNAs are affected by Sam68ΔC, the band corresponding to the 2 and 4 kb classes could not be used for quantitation. Relative PABP1 association was quantitated by the ratio of 9 kb RNA to actin RNA in the RIP compared to the input. Sam68ΔC significantly decreased PABP1 association with HIV-1 9 kb RNA relative to pcDNA, while Sam68:5–262 significantly increased PABP1 association with HIV-1 9 kb RNA (p < 0.01, Fig. 8b,c). Both Sam68Δ28ΔC and Sam68ΔCmin showed slight reductions in PABP1 association with 9 kb viral RNA but these were not found to be significant even at a higher p-value (p < 0.05, Fig. 8b,c). Therefore, similar to the results seen with the pgTat vector, changes in viral protein expression are correlated with extent of interaction of the corresponding RNA with PABP1.

The following constructs have been previously described: SV-Hygro, SV-H6Rev, CMVmyc3xterm 53, pDM128 54, pgTat 55, Bl-env-HindIII 56, Sam68, and Sam68ΔC 21. Gag-CTE and Gag-RRE plasmids were provided by Dr. K. Boris-Lawrie, Ohio State University. HxBruR^-/RI^- was provided by Eric Cohen, Universite de Montreal. Sam68Δ28ΔC was generated by PCR using primers: 5'-CGGAATTCCCCTCGGTGCGTCTGA-3' and 5'-CGGGATCCAGTCACAGTGG-CACCTCT-3'. The amplicon was digested with EcoRI and BamHI and ligated into CMVmyc3xterm. Sam68:5–262 was made by EcoRI and EcoRV digest of Sam68ΔC and ligated into the EcoRI and SmaI sites of CMVmyc3xterm. Sam68ΔCmin was made in a stepwise fashion. First, Sam68:5–300 was generated by PCR using primers 5'-CCATTAACGCAAATGGGCGGTA-3' and 5'-CCGCTCGAGAACAGG-TGGAGG-3'. The amplicon was digested with EcoRI and XhoI and ligated into Sam68ΔC. To generate internal deletions, Quickchange mutagenesis (Stratagene) was carried out using Sam68:5–300 as a template: Δ14: 5'-GGGGAATTCGAGAAGATCGGGCCGCAGCTGGC-3' and 5'-GCAGCTGCGG-CCCGATCTTCTCGAATTCCC-3'; Δ(45–54): 5'-GCTTCCTCAC-CGGCCCGCTCGGGCCTCGCCC-3' and 5'-GGGCGAGGCCCGAGCGGGCCGG-TGAGGAAGC-3'. Bl-actin was generated by PCR from cDNA using primers: 5'-GCTACGAGCTGCCTGACG-3' and 5'-TCCTTCTGCATCCTGTCG-3'. The amplicon was cloned into the EcoRV site of Bluescript. Bl-SD-Gag was amplified from HxBruR^-/RI^- using: 5'-CGCGGATCCGAAGTAGTGTGTGCCCGTCT-3' and 5'-CCCAAGCTTCCCTGCTTGCCCATACTATA-3'. The amplicon was digested with BamHI and HindIII and ligated into Bluescript.

HeLa and 293T cells were maintained in Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% fetal bovine serum (FBS), 50 μg/mL gentamycin sulfate and 2.5 μg/mL amphotericin B. Vectors were introduced to 293T cells by calcium phosphate transfection 57. Vectors were introduced to HeLa cells by Fugene 6 transfection reagent (Roche) following manufacturer's protocol. Cells were harvested two days post-transfection.

The following antibodies were used: mouse anti-myc (Invitrogen), mouse anti-tubulin (Sigma), mouse anti-gp120AK (courtesy of H. Schaal), mouse anti-p24 (clone 183-1112-5C), rabbit anti-L26 (Cell Signaling Technology) and rabbit anti-human PABP1 (aa 462–633) (courtesy of N. Sonenburg, McGill University). The following secondary antibodies were used: donkey anti-rabbit IgG conjugated HRP, donkey anti-mouse IgG conjugated HRP, donkey anti-rabbit IgG conjugated Texas red, donkey anti-mouse IgG conjugated Texas red (Jackson Immuno Research), protein G conjugated HRP (Molecular Probes), and sheep anti-digoxigenin conjugated FITC (Boehringer Mannheim).

6 × 10⁵ 293T cells were transfected as follows: 0.5 μg GagRRE, 0.1 μg SV-Hygro or SV-H6Rev, and 2.0 μg pcDNA3.1 or Sam68ΔC. For Rev-independent expression cells were transfected with 2.0 μg GagCTE, and 2.0 μg pcDNA3.1 or Sam68ΔC. Cells were harvested in whole cell lysis buffer (150 mM NaCl, 10 mM Na₂HPO₄, 1% Triton X-100, 0.1% SDS, 0.2% sodium azide, 0.5% sodium deoxycholate, 1 mM sodium orthovanadate), fractionated on SDS-PAGE gels and transferred to PVDF membrane (Pall Life Sciences). Bound antibodies were detected using Western Lightning (Perkin Elmer). For quantitation of western blots, films were scanned and analyzed using Imagequant software. Signals for both p55 and p24 were combined for analysis and data normalized using the tubulin signals. Quantitation shown was generated from three independent experiments.

3 × 10⁵ HeLa cells were transfected with 0.25 μg SV H6Rev, 1.25 μg pgTat and 5.0 μg pcDNA3.1 or Sam68ΔC. 48 hours post-transfection, cells were treated with fresh media containing no drug, colcemid (0.1 μg/mL), nocodazole (1.0 μg/mL), cytochalasin D (0.5 μg/mL), or latrunculin B (0.5 μg/mL) for 2 hours prior to fixation. In situ hybridization was performed as previously described 56. Digoxigenin-labeled Env-HindIII probe, antisense to HIV-1 env mRNA, was made with 10 × digoxigenin RNA labeling kit (Roche). Myc-tagged proteins were detected with monoclonal anti-myc antibody (Invitrogen) as previously described 58. Immunofluorescence was detected using a Leica DMR microscope at either 400× or 630× magnification.

6 × 10⁵ HeLa cells were transfected with 0.2 μg SV-H6Rev, 1.0 μg Gag-RRE and 4.0 μg pcDNA3.1 or Sam68ΔC. 48 hours post-transfection, cells were treated with drugs for 2 hours, as outlined above. Cells were then labeled with 100 μCi of ³⁵S- methionine for 4 hours. Cells were harvested in whole cell lysis buffer then diluted with 3 volumes IPP150 (150 mM KCl, 10 mM Tris-HCl pH 7.5, 0.1% NP40, 0.1% sodium azide). The precleared lysates were incubated with anti-p24 antibody for 1 hour at 4°C, then immunoprecipitated with Gammabind plus sepharose (GE Healthcare) for 1 hour. Beads were washed with IPP150 buffer and the immunoprecipitated protein was run out on 10% SDS-PAGE. Labeled proteins were detected following exposure to a Phosphor Imager screen.

3 × 10⁶ 293T cells were transfected with 0.4 μg SV-Hygro or SV-H6Rev, 2.0 μg pgTat and 8.0 μg of pcDNA3.1, Sam68ΔC or mutants thereof. 48 hours post-transfection 25% of the cells were harvested in whole cell lysis buffer for protein analysis, and 75% of the cells were harvested for RNA isolation 59. RNase protection assays (RPA) were performed as previously described 11. To monitor the polyadenylation status of the pgTat mRNA 10 μg total RNA was selected using oligo(dT)25 beads according to manufacturers directions (Dynal Biotech). The selected RNA was then input into the RNase protection assay using Bl-Tat X/X probe 11. RACE-PAT (random amplification of cDNA ends-polyadenylation test) cDNA was synthesized using an anchor primer (5'-CTCGCCGGACACGCTGAACTTTTTTTTTTTTTTTTTTTT-3') with MMLV-RT (Invitrogen). The cDNA generated was then used to generate both spliced and unspliced amplicons using the forward spliced (5'-AGCGGAGACAGCGACGAAGAG-3') or unspliced (5'-CGACCTGGATGGAGTGGGACA-3') and the reverse primer (5'-CTCGCCGGACACGCTGAAC-3'). Amplification used the following cycle parameters: 94°C, 1 minute; 66°C, 1 minute; 72°C, 2 minute; for 30 cycles. Amplicons were fractionated on native PAGE gels and detected by exposure to Phosphor Imager screen. For proviral RNA analysis, 3 × 10⁶ 293T cells were transfected with 2.0 μg HxBruR^-/RI^- and 8.0 μg of Sam68ΔC or mutants thereof. The probes used for RNase protection assay were Bl-actin and Bl-SD-Gag.

Polysome analysis was performed as described by Li et al. 60 with the only modification being the addition of cytochalasin D (0.5 μg/ml) to the media 1 h prior to harvest. 0.5 ml fractions were collected following centrifugation and analyzed for protein or RNA distribution. For protein analysis, 5 μg/ml BSA was added to 200 μl of each fraction and proteins precipitated by addition of an equal volume of 40% TCA. After incubation at 4 C for 1 h, precipitate was collected by centrifugation, washed 5 times with acetone and resuspended in whole cell lysis buffer. Proteins were fractionated on SDS-PAGE gels and blotted onto PVDF membranes for detection of indicated proteins. In the case of RNA, following proteinase K/phenol-chloroform purification, RNA was ethanol precipitated, resuspended and treated with Turbo DNase 1 as per manufacturer's instructions (Ambion). RNA was then used for quantition of actin and env RNA by real time QRT-PCR using the following primer sets: actin forward 5'-GAG CGG TTC CGC TGC CCT GAG GCA CTC-3', actin reverse 5'-GGG CAG TGA TCT CCT TCT GCA TCC TG-3', env forward 5'-CAA CAA TGG GTC CGA GAT CTT-3', env reverse 5'-AGC TCC TAT TCC CAC TGC TC-3'. QPCR reactions were run in duplicate for each probe and gradient fraction.

3 × 10⁶ 293T cells were transfected with 0.4 μg SVH6Rev, 2.0 μg pgTat and 8.0 μg of pcDNA3.1, Sam68ΔC or mutants thereof. Whole cell lysates were generated by a modification of the protocol of Siomi et al 61. 48 hours post-transfection, cells were washed twice with 1 × PBS and once with Buffer A (110 mM potassium acetate, 2 mM magnesium acetate, 2 mM DTT, 10 mM HEPES pH 7.5). The cell pellet was resuspended in 400 μl Buffer B (10 mM potassium acetate, 2 mM magnesium acetate, 2 mM DTT, 5 mM HEPES pH 7.5, 20 μM cytochalasin D) and incubated on ice for 10 min. Cells were disrupted by passage through 25 gauge needle, 5 times. KCl was added to 100 mM final concentration. Cells debris was pelleted by centrifugation at 1,500 rpm. Supernatant was adjusted to 0.5% NP40 and incubated on ice for 20 min. Insoluble material was pelleted by centrifugation at 14,000 rpm and supernatant retained for use in subsequent immunoprecipitations. 1/42 of the supernatant was retained for analysis of the RNA input and 1/42 for analysis of protein input. For each immunoprecipitation, 10/42 of the soluble fraction was diluted with 10 volumes RIPA buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 1% NP40, 1 mM EDTA ph 7.5, 0.5% sodium deoxycholate, 0.05% SDS) and 100 u RNaseOUT (Invitrogen). The sample was precleared by incubation with Gammabind plus sepharose (GE Healthcare) for 1 h at 4°C. Supernatant was incubated overnight at 4°C with the indicated antibody. Immune complexes were collected by addition of Gammabind plus sepharose and incubation at 4°C for 1 h. The resin was washed 5× with 1 ml of RIPA buffer and then resuspended in 100 μl of RIPA. RNA was extracted from 150 μl of the resin slurry 59 and 50 μl of slurry was used for protein analysis following addition of 50 μl 2× dissociation buffer. Isolated RNA was subsequently used in RPA assays as outlined previously.

The data are presented as mean +/- one standard deviation. Data were compared using Student's t-test. * indicates a p-value < 0.05. ** indicates a p-value < 0.01.