Direct Participation of Sam68, the 68-Kilodalton Src-Associated Protein in Mitosis, in the CRM1-Mediated Rev Nuclear Export Pathway

293T, Jurkat, CEM, and BOSC23 cells were purchased from the American Tissue Culture Collection and maintained in either Dulbecco's modified Eagle's medium (293T and BOSC23 cells) or RPMI 1640 medium (Jurkat and CEM cells) supplemented with 10% fetal bovine serum at 37°C with 5% CO₂. Human peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation with Ficoll-Hypaque, as described previously (55), and cultured in complete RPMI 1640 medium containing 10% fetal calf serum. Cell transfections were performed by the calcium phosphate precipitation method, which usually gives rise to 80% or higher transfection efficiency in 293T and BOSC23 cells. pcDNA3 was used to equalize the amount of DNA transfected among all transfections throughout the studies. pCMVβGal was included to normalize variations of transfection efficiency among all transfections involving the chloramphenicol acetyltransferase (CAT) reporter gene assay, as previously described (23, 24).

The sources of plasmids used in the studies were as follows: HIVΔenv and HIVΔenvΔrev were from M. Emerman (34), pSam68.HA from T. Ishidate (29), pSG5.Sam68 from S. Courtneidge (37), pGFP.Rev from R. Brack-Werner (38), pc.Rev and M10 from B. Cullen (40), pMX and pMX-IL-3R from T. Kitamura (52), pRRE-CAT and pPRE-CAT from T. Hope (27), and pSV2-CAT and pCTE-CAT from J. Roth (60). pcDNA3, pDsRed1-C1, and pEGFP-C1 were purchased from Clontech. pSam68.RFP and pRev.RFP have been described previously (35). The standard PCR cloning technique was used to construct the following expression plasmids (with the templates, primers, and cloning backbones shown in parentheses and the restriction enzyme sites underlined): pAs.Sam68a (pSG5.Sam68, 5′-TAA TGA ATT CAT GCA GCG CCG GGA CGA C-3′ and 5′-GGA ATT CTT AAT AAC GTC CAT ATG GGT G-3′, pcDNA3), pCRM1.RFP (pc3.CRM1, 5′-CGG GGT ACC ATG CCA GCA ATT ATG ACA ATG-3′ and 5′-CGC GGA TCC TTA ATC ACA CAT TTC TTC TGG-3′, pDsRed1-C1), and pCRM1.GFP (same template and primers as for pCRM1.RFP, with pEGFP-C1 as the cloning backbone). pAs.Sam68b was constructed by cloning the entire Sam68 insert (∼1.7 kb) from pSG5.Sam68 by EcoRI digestion, followed by ligation of the insert into pcDNA3 at the EcoRI site in a reverse orientation. pMX.As-Sam68 was constructed in a similar manner, except that a different backbone, i.e., pMX, was used. All recombinant plasmids were verified by sequencing, and functional activities of pRev.GFP and pM10.GFP were confirmed by both an RRE-mediated reporter gene assay and an in trans complementation assay.

Cells were washed twice with ice-cold phosphate-buffered saline and then harvested with a cell harvester. Cell pellets were resuspended in 2 volumes of whole-cell lysis buffer (10 mM NaHPO₄, 150 mM NaCl, 1% Triton X-100, 0.1% sodium dodecyl sulfate, 0.2% sodium azide, 0.5% sodium deoxycholate, 0.004% sodium fluoride, 1 mM sodium orthovanadate) and incubated on ice for 10 min. Whole-cell lysates were obtained by centrifugation and removal of cell debris. Cell lysates (25 μg of protein) were electrophoretically separated on sodium dodecyl sulfate-10% polyacrylamide gels and analyzed by immunoblotting using antibodies against Sam68 (Santa Cruz). The blots were then probed with appropriate horseradish peroxidase-conjugated secondary antibodies and visualized with the ECL system (Amersham). Relative levels of protein expression were determined by densitometric scanning of the blots using the housekeeping gene actin as an internal standard.

NL4-3 virus stocks were made by transfection of 293T cells with HIV-1 NL4-3 proviral DNA as previously described (1). Cell culture supernatants were collected 48 h after medium change, filtered, and saved as NL4-3 virus stocks. Replication-competent xenotropic murine leukemia virus (MLV) BV2 stocks were prepared as described previously (39). pMX-based viruses were prepared in a similar manner, except that a different cell line, i.e., the packaging cell line BOSC23, was used, as described elsewhere (56). For infection, cells were cultured on 10-cm-diameter plates and infected with viruses as indicated, in the presence of 8 μg of Polybrene/ml at 37°C for 2 h. Cells were then washed and supplemented with fresh medium. Viral infection and replication were determined by measuring reverse transcriptase (RTase) activity as described elsewhere (24). To assay the RTase activity of the BV2 or pMX virus, we adopted a similar protocol, except that magnesium was replaced with manganese in the RTase assay buffer.

Fluorescence images were captured by using a digital video imaging microscope system consisting of an Axiovert 200 M microscope with a 40× 1.4 UVF objective and an Axiovision video camera (all from Carl Zeiss) and were analyzed on an IDE Imaging Workstation (Tegraf).

To determine the functional significance of Sam68 protein in HIV-1 gene expression, we down-modulated constitutive Sam68 expression and investigated its effect on HIV-1 production. We took advantage of an antisense RNA expression vector-based strategy which has been widely and successfully used to study the functions of a number of genes (for a review, see reference 73). We made two Sam68 antisense constructs that expressed different lengths of the Sam68 cDNA in a reverse orientation in the context of a pcDNA3 backbone. 293T cells were transfected with these two antisense RNA expression plasmid DNAs. Western blot analysis showed that transient expression of these two anti-Sam68 antisense RNAs effectively down-regulated Sam68 protein expression by 67 to 69% (Fig. 1a). To determine whether there was any adverse effect on these cells transfected with anti-Sam68 antisense RNAs, we performed [³H]thymidine uptake, cell cycle, and cell viability analyses. However, no changes in cell proliferation, cell cycling time, or cell survival were detected in these transfected cells during the experiments, i.e., 2 to 7 days (data not shown).

We then transfected 293T cells with HIV-1 NL4-3 proviral DNA, in combination with or without anti-Sam68 antisense RNA expression vectors, and monitored viral production in these cells. The results showed that expression of either anti-Sam68 antisense RNA construct inhibited HIV-1 production by 80 to 86% (Fig. 1b). Inhibition of viral production was correlated with the amount of Sam68 antisense RNA expression vector DNA transfected (Fig. 1c). To ascertain that the inhibitory effect of Sam68 down-modulation on HIV-1 production is not due to a defect in virus budding in cells transfected with an anti-Sam68 antisense RNA expression vector, we also determined intracellular virus production levels in transfected 293T cells. The results showed a similar inhibitory effect (data not shown), further suggesting that transient HIV-1 gene expression and virus production were inhibited by down-modulation of constitutive Sam68 expression.

We then determined whether down-modulation of constitutive Sam68 expression would also affect HIV-1 replication in human T lymphocytes and PBMCs. To achieve a higher level of anti-Sam68 antisense RNA expression, we utilized a retrovirus-based gene delivery system, pMX, which has been shown to be very efficient at introducing foreign genes into a variety of highly refractory cell lines and primary cells (52). We cloned the 1.7-kb Sam68 cDNA in a reverse orientation into the retroviral pMX vector (pMX.As-Sam68) and prepared the viruses using the BOSC23 packing cell line (56). We transduced human T lymphocytes, Jurkat and CEM cells, and human PBMCs with the MX.As-Sam68 viruses expressing anti-Sam68 antisense RNA. The transduction efficiency was determined to be about 26% by using pMX-IL-3R as the control for interleukin-3 receptor expression (52). We then infected the transduced cells with HIV-1 NL4-3 and monitored the cells for HIV-1 replication. During the first 6 days postinfection, cells transduced with MX control viruses containing no Sam68 insert exhibited a kinetics typical of productive HIV-1 replication (Fig. 2a). In contrast, cells transduced with MX.As-Sam68 viruses produced little or no virus (Fig. 2a). It has been documented that down-modulation of constitutive gene expression by the antisense RNA technology is a result of double-stranded RNA formation between mRNA and complementary antisense RNA, so that mRNA is not accessible for translation (for a review, see reference 73). On the other hand, it has also been known that formation of double-stranded RNAs may give rise to a nonspecific inhibitory effect on viral replication (13, 62). Thus, we included as an infection control a recombinant replication-competent, Rev-independent simple retrovirus, MLV strain BV2 (39), in order to determine whether double-stranded RNA formation between Sam68 mRNA and anti-Sam68 antisense RNA resulted in a nonspecific antiviral effect. No differences were noted in MLV BV2 viral replication between cells transduced with MX viruses and cells transduced with MX.As-Sam68 viruses (Fig. 2b), suggesting that a nonspecific antiviral effect of double-stranded RNAs is not responsible for the inhibition of HIV-1 replication by anti-Sam68 antisense RNA. In agreement with our previous findings that adequate Sam68 expression is required for productive HIV-1 gene expression (35), these results demonstrated that down-modulation of constitutive Sam68 expression inhibited HIV-1 replication in human T lymphocytes and PBMCs.

The apparent discrepancy between the transduction efficiency (26%) and the inhibition efficiency (98%) (Fig. 2a) raised a possibility that the inhibitory effect on HIV-1 gene expression resulting from down-modulation of constitutive Sam68 expression was not cell autonomous. To address this possibility, we transduced T lymphocytes (Jurkat cells) with the same amount of pMX.As-Sam68 virus expressing anti-Sam68 antisense RNA as that used for Fig. 2a, collected the cell culture supernatant 48 h after transduction, and determined its effects on HIV-1 replication in Jurkat cells that were infected with HIV-1 NL4-3. We also included pMX and pMX.GFP viruses as transduction controls. The transduction efficiency was 24.6% based on green fluorescent protein (GFP) expression in pMX.GFP virus-transduced cells, which was close to that previously obtained. The results showed that, like pMX viruses and mock transduction (conditioned medium), the supernatant from pMX.As-Sam68-transduced cells failed to exhibit any inhibitory effect on HIV-1 replication in Jurkat cells (Fig. 2c). We also tested supernatants collected at 0, 6, 12, 24, and 72 h after MX viral transduction and found no effects (data not shown). These results together suggest that inhibition of HIV-1 gene expression by down-modulation of Sam68 expression was cell autonomous. Furthermore, these results also validated the findings of earlier studies that infection with MLV or MLV-based retroviruses promotes or facilitates coinfection with HIV-1 (8, 9, 71).

Since our previous studies (35) and studies by others (58, 68) have shown that Sam68 interaction with Rev may modulate Rev activity, we then determined whether inhibition of HIV-1 gene expression and viral replication by down-modulation of constitutive Sam68 expression is due to a decrease in Rev activity. We first performed an in trans complementation assay in the context of the HIV-1 genome. We transfected 293T cells with HIV.ΔenvΔrev, containing both env and rev deletions (4), and with either an expression vector for Rev, an expression vector for anti-Sam68 antisense RNA, or both. We also included HIV.Δenv (34), containing an env deletion but an intact rev gene, as a control. As expected, deletion of the rev gene from the HIV-1 genome, i.e., HIV.ΔenvΔrev, resulted in a minimal level of HIV-1 production, which was fully compensated for in trans by expression of a Rev expression plasmid (Fig. 3). Anti-Sam68 antisense RNA expression decreased Rev-mediated HIV-1 production by 64 and 82%, which appeared to be correlated with the amount of anti-Sam68 antisense RNA expression vector DNA transfected (Fig. 3). However, increasing Rev expression did not abolish the inhibitory effect on HIV-1 production in cells transfected with a larger amount of anti-Sam68 antisense RNA expression vector DNA (Fig. 3). These results demonstrated that down-modulation of constitutive Sam68 expression inhibited Rev activity, further supporting the notion that Sam68 expression is required for Rev function (35).

To determine the inhibitory specificity of Sam68 down-modulation on Rev function, we expressed several known Rev-independent reporter DNAs in the presence of an anti-Sam68 antisense RNA expression vector. These included a Mason-Pfizer monkey virus RNA constitutive transport element (CTE)-dependent CAT reporter (CTE-CAT) (60), a hepatitis B virus posttranslational regulatory element (PRE)-dependent CAT reporter (PRE-CAT) (54), and the simian virus 40 promoter-driven SV2-CAT reporter (60). We also transfected the Rev-dependent reporter RRE-CAT (54) as a control. As expected, Rev expression transactivated RRE-CAT expression by 54.4% (Fig. 4a) but exhibited little or no effect on CTE-, PRE-, or SV2-CAT expression (Fig. 4b through d). The results also showed that coexpression of anti-Sam68 antisense RNA resulted in a 44.9% decrease in Rev-mediated RRE-CAT expression (Fig. 4a) but in little or no inhibition of CTE-, PRE-, or SV2-CAT expression, regardless of whether Rev was expressed or not (Fig. 4b through d). These results together demonstrated that inhibition of Rev-dependent HIV-1 gene expression by Sam68 down-modulation was due to a preferential inhibition of Rev function.

Extensive studies have shown that Rev nuclear export is mediated by CRM1 (2, 16, 19, 32, 49, 53). Our previous studies have demonstrated that Sam68 expression promotes Rev nuclear export (35), suggesting that Sam68 may be also involved in the Rev nuclear export pathway. Thus, we decided to determine the relationship between Sam68 expression-induced Rev nuclear export and CRM1-mediated Rev nuclear export. Although our previous studies have shown that Sam68 is expressed only in the nucleus (35), we first wanted to ascertain whether Sam68 protein itself can be shuttled between the cytoplasm and the nucleus. We transfected 293T cells with the Sam68.RFP expression plasmid, in which Sam68 was tagged with red fluorescent protein (RFP), and then treated the cells with actinomycin D (ATD), a transcriptional inhibitor that is known to be capable of inducing accumulation of nucleocytoplasmic shuttling proteins in the cytoplasm (30, 59, 60). We also included as controls in the transfection the Rev.GFP and CRM1.RFP expression plasmids, which were tagged with GFP and RFP, respectively. We also fixed the cells in 2% paraformaldehyde and stained the cells with 4′,6′-diamidino-2-phenylindole (DAPI) (blue fluorescence) to locate the nucleus. As expected, Rev was localized in the nucleus (Fig. A), while CRM1 was mainly localized at the nuclear membrane, with some in the nucleus and in the cytoplasm (Fig. B). When the transfected cells were treated with ATD, both Rev and CRM1 were considerably shifted into the cytoplasm (Fig. D and E). However, ATD treatment resulted in little or no cytoplasmic localization of Sam68 (Fig. F), which remained within the nucleus (Fig. C). Taken together with our previous finding that Sam68 alone was not able to substitute for Rev in the Rev-responsive element (RRE)-mediated gene expression (35), these results demonstrated that Sam68 itself is not a nucleocytoplasmic shuttling protein.

We then took advantage of an CRM1-specific inhibitor, leptomycin B (LMB), which is known to prevent CRM1 from binding to Rev and thus to disrupt Rev nuclear export (2, 16, 19, 32, 50, 75), to determine whether LMB treatment would inhibit Sam68 expression-induced Rev nuclear export. We transfected 293T cells with both Rev.GFP and Sam68.RFP and then treated the cells with LMB. In agreement with our previous studies (35), the results showed that coexpression of Sam68 and Rev relocalized each from the nucleus (Fig. A and C) onto the nuclear membrane, which resulted in Rev nuclear export (Fig. A). Colocalization analysis showed that Sam68 was almost completely and exclusively colocalized with Rev on the nuclear membrane (Fig. C). Surprisingly, LMB treatment completely abolished Sam68 expression-induced Rev nuclear export, as Rev was retained in the nucleus (Fig. D) and Sam68 mainly remained on the nuclear membrane (Fig. E). Moreover, LMB treatment to some degree prevented colocalization of Sam68 and Rev (Fig. F) compared to that in the absence of LMB treatment (Fig. C). These results together demonstrated that Sam68 expression-induced Rev nuclear export was CRM1 dependent, suggesting that Sam68 may function as a previously unrecognized cofactor in the CRM1-mediated Rev nuclear export pathway.

To further characterize the roles of Sam68 in the CRM1-mediated Rev nuclear export pathway, we first determined whether CRM1 bound directly to Sam68. For this analysis we used immunoprecipitation combined with Western blotting of both constitutive proteins and transiently expressed proteins, as well as the yeast two-hybrid protein-protein interaction assay. Nevertheless, we were unable to detect any direct interaction between CRM1 and Sam68 in any of these experiments (data not shown). Although the NES of Rev has been demonstrated to be directly involved in the interaction of Rev with both Sam68 (35) and CRM1 (16, 19, 49), we did not detect any adverse effect of Sam68 expression on CRM1 binding to Rev, or vice versa (data not shown). These data suggest that Rev binding to Sam68 and to CRM1 may occur independently or simultaneously.

To determine the functional relationship between Sam68 and CRM1 in the Rev nuclear export pathway, we investigated the effect of Sam68 down-modulation on the CRM1-Rev interaction. We transfected 293T cells with Rev.GFP and CRM1.RFP in the absence or presence of the anti-Sam68 antisense RNA expression vector pAs-Sam68a (Fig. 1a). We also treated the transfected cells with LMB as a control. In agreement with the findings of previous studies (60, 79), the results showed that coexpression of CRM1 and Rev relocalized each to the nuclear membrane and led to Rev nuclear export (Fig. 6A and B). As expected, LMB treatment disrupted the CRM1-Rev interaction, resulting in no Rev nuclear export (Fig. 6D) and little colocalization of Rev and Sam68, as they were localized in distinct nuclear compartments (Fig. 6D and E). In contrast, down-modulation of Sam68 expression by expression of the anti-Sam68 antisense RNA expression vector As-Sam68a resulted in little or no Rev nuclear export (Fig. 6G) and exclusively nuclear localization of both Rev and CRM1 (Fig. 6G and H). Moreover, unlike LMB treatment (Fig. 6F), down-modulation of Sam68 expression also resulted in exclusive colocalization of Rev and Sam68 within the nucleus (Fig. 6I). Taken together, these results demonstrated that down-modulation of Sam68 expression blocked the CRM1-mediated Rev nuclear export pathway, suggesting that Sam68 expression-induced Rev nuclear export and CRM1-mediated Rev nuclear export were mutually dependent. In addition, these results also raised the possibility that Sam68 may serve to transport the export complex containing CRM1 and Rev from the nucleus to an area in close proximity to the nuclear membrane (the NPC).