A Role for KH Domain Proteins (Sam68-like Mammalian Proteins and Quaking Proteins) in the Post-transcriptional Regulation of HIV Replication*

Overexpression of Sam68 functionally substitutes for, as well as synergizes with, human immunodeficiency virus type 1 (HIV-1) Rev in RRE (Rev response element)-mediated gene expression and virus replication. In ad-dition, COOH-terminal deletion and/or point mutants of Sam68 exhibit a transdominant negative phenotype for HIV replication. Sam68 is a member of KH domain family that includes SLM-1, SLM-2 (Sam68 like mammalian); and QKI-5, QKI-6, and QKI-7 (mouse quaking) proteins. The objective of this study was to examine the effects of these KH family proteins on RRE- and CTE (constitutive transport element of type-D retrovirus)-mediated transactivation. We now report that SLM-1 and SLM-2 proteins, which are the closest relatives of Sam68, marginally enhanced RRE-mediated transactivation, while QK isoforms that are distant relatives of Sam68 had no effect. Interestingly, these proteins

The Rev protein of human immunodeficiency virus (HIV) 1 facilitates the nuclear export of unspliced or singly spliced viral mRNA (1). Rev has been shown to be a shuttle protein (2), comprised of a basic, nuclear localization sequence (3) and a leucine-rich nuclear export sequence (4). The basic domain also constitutes an RNA-binding domain that specifically interacts with the cognate target sequence, RRE. Substitutions of leucine residues within nuclear export sequence yielded a mutant Rev protein (RevM10) with a dominant negative phenotype (3). RevM10 has been shown to confer human CD4 cells with antiviral resistance in cell culture (5) and preferential survival in HIV-1-infected patients (6,7). Recently, the Revnuclear export sequence was shown to bind the nuclear export receptor CRM-1, a member of the importin-␤ family (8). This interaction is functionally relevant, since leptomycin B (LMB), a drug that disrupts the complex formation of Rev, CRM-1, and RanGTP, also inhibits the nuclear export of Rev-nuclear export sequence conjugates (8). CRM-1 probably bridges the indirect interaction of Rev with members of the nucleoporin family (9) such as CAN/Nup214 proteins (10). Additional cellular proteins that bind Rev (11)(12)(13)(14)(15) and/or RRE (16,17) have been identified, which either positively or negatively modulate Rev activity. Recently, one such cellular protein that functionally substitutes for Rev has been identified as Sam68 (18). Sam68 binds to RRE in vitro and in vivo, and functionally replaces as well as synergizes with HIV-1 Rev in RRE-mediated gene expression and virus replication (18). Sam68 mutants deleted in the carboxyl terminus show a dominant negative phenotype in HIV replication (18). Furthermore, a single amino acid (P439R) substitution in the COOH-terminal domain of Sam68 also confers a transdominant negative phenotype (19). Overexpression of Sam68 also activated CTE-regulated HIV gag gene expression in human cells as well as in quail cells in the presence of human Tap (20).
Sam68 is a target of the c-Src tyrosine kinase (21,22). It contains an hnRNP-K homology domain (KH domain) (23) that mediates RNA binding and protein-protein interaction (24). The KH domain has also been reported in several RNA-binding proteins such as GRP33 (25), fragile X mental retardation gene product FMR-1 (26), and the Caenorhabditis elegans germ linespecific tumor suppressor GLD-1 (27), and as recently reported, other proteins such as SLM-1, SLM-2 (28), QKI-5, QKI-6, and QKI-7 (29,30). In this study, we have investigated the role of KH proteins in the post-transcriptional regulation of HIV replication. Here, we report that KH proteins (SLM-1, SLM-2, QKI-5, QKI-6, and QKI-7), that have partial homology with Sam68 failed to transactivate RRE-directed reporter gene expression independent of Rev, but are able to enhance Rev transactivation on RRE. These effects were sensitive to LMB, but insensitive to olomoucine. Our results provide the first direct evidence that other members of KH proteins are also involved in post-transcriptional regulation of HIV gene expression.
Cells, Transfections, and Chloramphenicol Acetyltransferase (CAT) Assays-The 293T, HeLa, and COS cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. In general, between 1 and 3 g of DNA was transfected into (1 ϫ 10 5 ) 293T cells using FuGENE 6 (Roche Molecular Biochemicals) according to the manufacturers protocols. To normalize the transfection efficiencies, we have used 0.1 g of pcDNA-Lac Z expression vector as an internal control. PcDNA3 plasmid was used to equalize the amount of DNA for each transfection. For treatment with soluble inhibitors, 18 h after transfection, cells were incubated in medium containing 2 nM LMB, and/or 75 M olomoucine. Forty-eight hours post-transfection, cells were harvested, washed with phosphate-buffered saline, and then re-suspended in 50 -100 l of 0.25 M Tris, pH 7.8. The cell extracts, CAT assays, and separation of reaction products were performed as previously described (18). Fold of trans-activation was quantitated by scintillation counting of the products separated from the reaction.
Northern Analysis-We have fractionated the nuclear and cytoplasmic fractions as described by Maniatis (32). Using Ultraspec TM (Biotecx Laboratories, Inc.) reagent, total or fractionated RNA was extracted from the 293T cells transfected with different plasmids shown in Figs. 2 and 5 by FuGENE 6 as described previously (19). Approximately 20 g of RNA was separated on 1% agarose formaldehyde gels by electrophoresis and blotted onto nitrocellulose filters. The filters were hybridized with 32 P-labeled CAT cDNA and detected by autoradiography. Blots were re-hybridized with control ChoA cDNA. Ribosomal RNA (28 S and 18 S) was also used to assess the integrity of the RNA and for RNA loading controls (data not shown).
In Vitro Transcription and RNA Gel Mobility Shift Assay-Plasmid pcRRE was constructed by inserting PCR-amplified HIV-1 (HXB-2) RRE sequence into HindIII and BamHI cloning site of pcDNA3 (Invitrogen). [ 32 P]UTP-labeled RRE RNA were synthesized by in vitro transcription with T7 RNA polymerase according to the protocols (Promega) using BamHI linearized pcRRE plasmid as template. RNA-protein binding reactions were carried out at room temperature in a total volume of 30 l in the binding buffer containing 60 mM NaCl, 12 mM Hepes (pH 7.9), 12 mM dithiothreitol, and 50 units of RNasin. Typically, 1 ϫ 10 4 cpm of 32 P-labeled RNA and 100 ng of protein were used. The binding reaction was allowed to proceed for 15 min at room temperature and then the mixture was electrophoresed on a 4.5% nondenaturing PAGE and then subjected directly to autoradiography.
Co-immunoprecipitation Assay-293T cells (6 well dish) were cotransfected with pCMV128 (1 g) alone and with wild-type and/or mutant pRev (3 g). Forty-eight hours later, cell extracts were prepared by lysing the cells in 1 ml of 0.65% Nonidet P-40 lysis buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.8, 1.5 mM MgCl 2 , 0.65% Nonidet P-40). For the detection of RNA free interaction, cell extracts were treated with 10 g of RNase per ml for 30 min at 37°C. Cell lysates were precleared with normal rabbit serum by incubating overnight with 20 l of normal rabbit serum-conjugated agarose beads at 4°C. The resultant lysate from the cells co-transfected with pCMV128 and pRev was divided into four parts. Each part was mixed with 40 l of protein A-and G-agarose beads plus 10 l of anti-QKI-5, QKI-6, Rev, and/or normal rabbit IgG. After incubating at 4°C overnight, the beads were washed 3 times with lysis buffer and immunoprecipitates were suspended in SDS buffer, boiled for 5 min, and analyzed by Western blot using Rev antibody.
P24 Antigen Capture Assay-293T cells were co-transfected with CTE-gag expression vector (33) with KH proteins and/or HIV-1 rev(Ϫ) proviral DNA (34) and KH proteins expressing vectors plus Rev expression vector. Forty-eight hours post-transfection, cell-free supernatants were collected and subjected to p24 antigen assay (Coulter).

Effects of KH Proteins on RRE-and CTE-mediated Reporter Gene Expression-
We have previously demonstrated that overexpression of Sam68 functionally substitutes for, as well as synergizes with HIV-1 Rev in RRE-mediated gene expression and virus replication. Amino acid sequence alignment suggests that SLM-1, SLM-2 (28), QKI-5, QKI-6, and QKI-7 (29,30) have partial homology with Sam68 and belong to the KH domain family (Fig. 1). Because of the sequence homology among these proteins, we have investigated the role of KH proteins in the post-transcriptional regulation of HIV replication. To explore a potential role of KH family proteins in RRE-and CTE-mediated transactivation, we examined the effect of exogenously expressed SLM-1, SLM-2, QKI-5, QKI-6, and QKI-7 under the control of the CMV promoter on a RRE-regulated reporter gene expression in transient co-transfection assays. As shown in Fig. 2A, Sam68, SLM-1, and SLM-2 induced 5-, 2.5-, and 1.7-fold of RREmediated CAT reporter gene expression, respectively, while Rev expression yielded a 10-fold increase in these studies (Fig. 2B). In contrast, QKI-5, QKI-6, and QKI-7 did not transactivate RRE-mediated gene expression. We next investigated whether an increased CAT activity was also reflected at the levels of RRE-CAT mRNAs. We isolated total RNA from 293T cells transfected with the plasmids shown in Fig.  2A and performed Northern analysis using CAT gene as a probe. Northern blot analysis demonstrated (Fig. 2B) that an increase of RRE-CAT mRNA (5-and 3-fold) was observed from the cells co-transfected with Sam68 and SLM-1 but not SLM-2, QKI-5, QKI-6, and QKI-7 expression plasmids. Thus, the increase in mRNA levels ( Fig. 2B) was comparable with the CAT activity shown in Fig. 2A. These results suggest that Sam68 may have involved in the stability and/or export of mRNAs.
We previously reported that Sam68 also efficiently activated HIV-gag expression from a CTE-reporter gene (20). To assess the effect of other KH proteins on CTE-mediated gene expression, we co-transfected expression vectors for these proteins with CTE-gag reporter gene (33) and measured the p24 antigen expression in cell-free supernatants 48 h post-transfection. As a positive control, we used Sam68. Here, we demonstrate that Sam68, but not other KH proteins, significantly enhanced CTE-mediated gag expression (Fig. 3). These results suggest that among the KH proteins tested, only Sam68 efficiently activated RRE and CTE-mediated reporter gene expression.
Effects of KH Proteins on Rev Activation of RRE-To determine whether these KH proteins also enhance Rev activation of RRE-directed CAT reporter gene expression, we co-transfected expression vectors for these proteins independently with Rev into 293T cells. The expression of Rev yielded a 10-fold increase in CAT activity, while co-expression of Sam68 with Rev resulted in 45-fold increase of CAT activity over the basal level (Fig. 4). Interestingly, all the KH proteins significantly enhanced Rev activation of RRE-mediated gene expression (25-35-fold). We next investigated whether an increased CAT activity was also reflected at the levels of cytoplasmic RRE-CAT RNA. For these studies, we have isolated RNA from the cytoplasmic and nuclear fractions of 293T cells transfected with various expression plasmids shown in Fig. 5 and performed Northern analysis using 32 P-labeled CAT gene as a probe. Northern blot analysis demonstrated that Sam68, like Rev, induced the accumulation of unspliced cytoplasmic CAT RNA (Fig. 5A). Co-expression of Sam68 and Rev resulted in a synergistic increase in the accumulation of unspliced cytoplasmic CAT RNA. Similarly, SLM-1, SLM-2, and QK proteins also efficiently enhanced the Rev-mediated nuclear export of un-spliced CAT mRNA into the cytoplasm (Fig. 5, A and B). Thus, the increase in CAT activities of KH proteins and Rev correlated with the RNA levels shown in Fig. 5.
This effect of the KH proteins was also apparent in a virus rescue assay. We co-transfected 293T cells with a rev(Ϫ) proviral DNA plasmid (34), various KH domain protein expression vectors and a Rev expression vector, and measured the expression of p24 antigen in the cell-free supernatants. Expression of Rev alone yielded 18 ng/ml p24 expression, while co-expression of KH proteins and Rev resulted in 2-3.5-fold (36 to 67 ng/ml) of p24 antigen expression increase over Rev (Fig. 6). These results indicate that KH proteins enhance Rev activation of RRE-directed gene expression in the context of a provirus.
Effects of KH Proteins on RRE-mediated Gene Expression in Various Cells-To determine whether the effect of KH family proteins is specific to 293T cells, we co-transfected various expression plasmids into COS and HeLa cells and investigated the consequence of overexpressed KH proteins on RRE-mediated CAT gene expression (Table I) Forty-eight hours post-transfection cell extracts were prepared and subjected to CAT enzyme assays and quantification of CAT activity as described under "Experimental Procedures." B, RRE-CAT RNA expression as determined by Northern blot analysis. 293T cells (1 ϫ 10 6 ) were co-transfected with RRE-CAT (0.25 g) alone and plus Rev (0.05 g), Sam68, SLM-1, SLM-2, QKI-5, QKI-6, and/or QKI-7 (0.5 g) expression plasmids using FuGENE 6 transfection reagent. Forty-eight hours posttransfection, total RNA was isolated and subjected to Northern hybridization using the 32 P-labeled CAT cDNA as a probe. For internal RNA controls, blots were rehybridized with labeled ChoA cDNA. US, unspliced CAT RNA. cell lines while QK proteins had no effect (data not shown). However, all KH proteins tested enhanced Rev activation of RRE-mediated reporter gene expression in COS and HeLa cells, consistent with the results with 293T cells. These results suggest that the effects of KH proteins on RRE-mediated gene expression are not cell-type specific.
Interaction of KH Proteins with RRE and Rev-Since unlike Sam68, the other KH proteins tested did not enhance RREdirected reporter gene expression independently of Rev, we propose that these proteins may not interact with RRE directly. To test this hypothesis, we employed gel mobility shift assay to assess an in vitro binding of QKI-5, and QKI-6 proteins to 32 P-labeled RRE RNA. Labeled RRE RNA bound to Sam68, but not to QKI-5 and QKI-6 proteins (Fig. 7A). As a negative control, we have used GST, gp120, and QKI-5kt4 proteins, which also failed to bind to RRE RNA. QKI-5kt4 is a mutant of QKI-5 that contains a single amino acid change from glutamic acid to glycine at position 48 in the QUA1 domain, which leads to a loss of QKI self-interaction (30). These results suggest that direct binding of Sam68 to RRE RNA is important for its effect on the transactivation of RRE-mediated gene expression independently of Rev (Fig. 2).
Since KH proteins enhance Rev activation of RRE-mediated gene expression, we propose that KH proteins may interact with Rev. To assess whether QKI-5 and QKI-6 proteins interact with Rev in vivo, 293T cells were co-transfected with Rev and CMV128 expression vectors and total cellular proteins from these cells were immunoprecipitated with QKI-5 and QKI-6 antibodies. The precipitates were subjected to Western blot analysis using antibodies to Rev. Normal rabbit IgG was used as a negative control. To rule out the possibility that the interaction between Rev and KH proteins is bridged by specific or nonspecific cellular RNAs, we also performed in vivo binding assays with cell extracts that had been treated with RNase. Rev was detected in both the nuclease-treated and untreated immune complexes with antibodies to QKI-5 and QKI-6, but not with control IgG (Fig. 7B) indicating that RNA does not bridge the interaction between QK proteins and Rev.
To further characterize the specific interaction between Rev and KH domain proteins in vivo, we analyzed the association of QKI-5 with various carboxyl-terminal mutants of Rev (31) shown in Fig. 8A by co-immunoprecipitation assays. For these studies, we used antibodies against QKI-5 to immunoprecipitate interacting mutant Rev proteins from 293T cells transfected with plasmids expressing wild-type and M18, M20, and M21 Rev mutants. The precipitates were subjected to Western blot analysis using antibodies to Rev. M18, M20, and wild-type  125 g), Sam68, SLM-1, SLM-2, QKI-5, QKI-6, QKI-7 (0.25 g), and/or Rev (0.025 g) expression plasmids using the FuGENE 6 transfection reagent. pcDNA3 was used to equalize the amount of DNA to 2.5 g for each transfection. Forty-eight hours post-transfection cell extracts were prepared and subjected to CAT enzyme assays and quantification of CAT activity as described previously (18). Rev, but not M21, co-immunoprecipitated with QKI-5 (Fig. 8B). All Rev mutant proteins were expressed at comparable levels, as detected by anti-Rev antibodies (Fig. 8C). To determine the functional relevance of these interactions, 293T cells were cotransfected with RRE-CAT and wild-type and mutant Rev constructs with or without QKI-5 expression vectors. Forty-eight hours post-transfection, cell extracts were made and subjected to CAT assays as described previously (18). M18 and M20, but not M21, were functional in activating RRE-mediated CAT gene expression. As shown above, QKI-5 only interacted with functional Rev (M18, M20, and wild type) but not the nonfunctional M21 mutant. In addition, they only increased RREmediated CAT gene expression in the presence of wild-type Rev or functional Rev mutants (Fig. 8D). Taken together, these results strongly suggest that the in vivo interaction of QKI-5 and Rev is specific and functionally relevant.
The Effect of KH Proteins on Rev Activation Is hCRM-1 De-pendent-Previously, we demonstrated that the activation of RRE RNA by Sam68, in contrast to Rev, was hCRM-1 independent (18), since it was resistant to LMB treatment. Furthermore, we showed that Sam68, but not Rev, was inhibited by olomoucine, an inhibitor of Cdc2 kinase (35). To determine whether the increase in Rev activation by KH proteins is still mediated by hCRM-1, we assessed the effect of LMB and olomoucine on these activities. LMB inhibited 50 to 58% of the activity of Rev in the presence or absence of KH proteins. In contrast, olomoucine did not have any effect on these activities (Table II). As a positive control, we have used Sam68 and Rev. These results suggest that the non-Sam68 KH proteins only enhanced the CRM-1-dependent, Rev-mediated RNA export pathway.

DISCUSSION
Sam68 is a member of the KH domain family proteins that include SLM-1, SLM-2 (28), and mouse quaking genes QKI-5, QKI-6, and QKI-7 (36). Although most of the biochemical and molecular analyses of these genes have been carried out, the role of these genes in vivo has not yet been defined. We now extend our earlier studies on the effects of Sam68 on RREmediated gene expression and HIV replication to these additional members of the KH domain protein family.
We demonstrate here that, unlike Sam68, the other KH proteins tested failed to significantly activate RRE-directed reporter gene expression in the absence of Rev (Fig. 2A). Sequence alignment of these proteins suggests that SLM-1 and SLM-2 are more related to Sam68 than QKI-5, QKI-6, and QKI-7, with the greatest homology in the GSG and KH domain, and lesser homology in the COOH-terminal region (Fig. 1A). Furthermore, Sam68 and SLM-1 but not SLM-2 are the substrates for c-Src kinase (28), and the heterodimerization ability of SLM-1 with Sam68 is greater than SLM-2 (28). Our results showed that SLM-1 is more efficient than SLM-2 in transactivating RRE-mediated gene expression (Fig. 2) and both are far less active than Sam68. The COOH-terminal domains of QKI-5, QKI-6, and QKI-7 proteins have no sequence homology FIG. 6. KH proteins enhance Rev activity in rev(؊) rescue assay. 293T cells were co-transfected with rev(Ϫ) proviral DNA (0.2 g), Sam68 and KH proteins (0.25 g), and/or Rev (0.025 g) expression plasmids as indicated, using FuGENE 6 transfection reagent. PcDNA3 was used to equalize the amount of DNA to 2 g for each transfection. Forty-eight hours post-transfection, the cell-free supernatants were collected and subjected to p24 antigen capture assay (Coulter).

FIG. 7. Interaction of KH proteins (QKI-5 and QKI-6) with RRE RNA and Rev.
A, QK proteins do not bind to RRE in vitro. Approximately, 100 ng of GST, gp120, GST-Sam68, QKI-5kt4, QKI-5, and/or QKI-6 proteins were incubated at room temperature for 10 min in a binding buffer containing 60 mM NaCl, 12 mM Hepes (pH 7.9), 12 mM dithiothreitol, and 50 units of RNasin. 32 P-Labeled RNA (1 ϫ 10 4 cpm) was added to a final volume of 30 l, and the incubation was carried out at room temperature for an additional 15 min. The reaction mixture was electrophoresed on a 4.5% nondenaturing polyacrylamide gel using the 1 ϫ Tris borate-EDTA (TBE) buffer. The gel was dried and exposed to x-ray film. B, in vivo binding of KH proteins to Rev. QKI-5 and QKI-6 antibodies were used to immunoprecipitate interacting components from the lysates prepared from the 293T cells transfected with pRev (3 g) and pCMV128 (1 g). Nonimmune IgG and Rev-IgG was used as negative and positive controls, respectively. For the detection of RNAfree interaction, cell extracts were treated with 10 g of RNase per ml for 30 min at 37°C. The immunoprecipitates were subjected to SDS-PAGE followed by Western blot analysis using Rev antibodies. with Sam68 (Fig. 1B). They also failed to bind RRE RNA (Fig.  7A), and failed to transactivate RRE-mediated reporter gene expression directly (Fig. 2A). These results are consistent with our previous demonstration that deletion of COOH-terminal region of Sam68 abolished its binding to RRE RNA as well as transactivation activity (18). Recently, QKI-7 was shown to fail to bind homopolymeric RNAs (37). We also demonstrated that the increase in CAT activities was comparable with the increase in CAT mRNA (Fig. 2). These results suggest that the function of Sam68 may be at the level of up-regulating the steady-state levels of RRE-containing mRNAs. We also demonstrated that unlike Sam68, the KH proteins tested did not have any effect on CTE-mediated gene expression (Fig. 3). These results suggest that Sam68 is unique among KH family members in its ability to specifically activate RRE-and CTE-mediated transactivation. Interestingly, all KH domain proteins tested enhanced Rev activation of RRE-mediated gene expression in two different assay systems ( Fig. 4 and 6). This enhancement may be mediated by interaction between Rev and the KH domain proteins, since QKI-5 and QKI-6 proteins do not bind RRE RNA in vitro, but complex with Rev in vivo ( Figs. 7 and 8). Similarly, SLM-2, which is more closely related to Sam68 than the QK proteins, also failed to bind RRE RNA (data not shown) and marginally activated RRE-mediated gene expression independently of Rev ( Fig. 2A). The increases in CAT activity also correlated with increases in cytoplasmic CAT mRNAs (Figs. 4 and 5). These results suggest that these proteins are involved in increasing the stability or export of unspliced mRNAs, rather than their translational efficiency. In fact, QKI-6 was reported to act as a translational repressor (38). LMB, a drug that interferes with the CRM-1-dependent nuclear export pathways, inhibits Rev but not Sam68 activation of RRE-dependent gene expression (18). In contrast, olomoucine, an inhibitor of Cdc2 kinase, specifically inhibits Sam68, but not Rev activation (18). Our data presented here showed that the increase in Rev activity due to the non-Sam68 KH proteins was inhibited by LMB to the same extent as Rev itself, and not at all inhibited by olomoucine (Table II). These results suggest that phosphorylation of KH proteins by Cdc2 kinase is not important for the observed enhancing effect of Rev function.
Sam68, SLM-1, SLM-2, and QKI-5 are predominantly nuclear proteins (39), while QKI-6 and QKI-7 are mostly cytoplasmic. Furthermore, QKI-5 was shown to shuttle between the nucleus and cytoplasm (30). Although Sam68 does not seem to be a shuttle protein, it has been reported to re-localize to the cytoplasm of cells infected with poliovirus (40). Recently, it was reported that insulin stimulation promotes the re-localization of Sam68 from the nucleus to the cytoplasm (41). Since the KH proteins can interact with each other (24,28) as well as Rev in vivo (Fig. 7), it is conceivable that they can traffic between the nucleus and cytoplasm through a piggy-back mechanism. Alternatively, different KH proteins might co-operate with Rev in the two cellular compartments at different stages of nuclear export. In light of these finding, it would be of interest to determine whether the functional domains are interchangeable among the KH proteins, and if the COOH-terminal domain mutants of the other KH proteins also exhibit transdominant negative phenotype for HIV replication. Additionally, it would be important to determine which of the nuclear pore proteins associate with Sam68 or other KH proteins and play a role in the export of HIV mRNA.
Acknowledgments-We thank Dr. Stephane Richard for SLM-1 and SLM-2 expression vectors, Dr. Michael Malim for Rev antibodies and Rev mutant plasmids, and Dr. Minoru Yoshida for LMB. We also thank Dr. Thomas Holland for helpful suggestions and Dr. Keshamouni Venkateshwar for assistance in preparing the manuscript.

FIG. 8. The interaction of QKI-5 and Rev in vivo is functionally relevant.
A, carboxyl-terminal Rev mutants. We have used the Rev mutants that were described previously (31). B, association of QKI-5 and Rev mutants in vivo. 293T cells (1 ϫ 10 6 ) were co-transfected with pCMV128 (1 g) alone and plus pQKI-5 (2 g) in the presence and absence of wild-type and M18, M20, and M21 Rev mutant expression vectors (2 g) using FuGENE 6 transfection reagent. QKI-5 antibodies were used to immunoprecipitate interacting components from the lysates of 293T cells. The immunoprecipitates were subjected to SDS-PAGE followed by Western blot analysis using Rev antibodies. C, expression of Rev mutants. To analyze the expression of M18, M20, and M21 Rev mutants, cell lysates (from B) before immunoprecipitation were subjected to Western blot analysis using Rev antibodies. D, QKI-5 enhances the activities of functional Rev mutants on RRE-mediated transactivation. 293T cells (1 ϫ 10 5 ) were co-transfected with RRE-CAT (0.125 g), pRev, pM18, pM20, and pM21 (0.025 g) in the presence and absence of QKI-5 expression vector (0.25 g) using the FuGENE 6 transfection reagent. pcDNA3 was used to equalize the amount of DNA to 2.5 g for each transfection. Forty-eight hours post-transfection cell extracts were prepared and subjected to CAT enzyme assays and quantification of CAT activity as described previously (18). Arrow denotes the position of Rev.