Intracellular Trafficking and Interactions of the HIV-1 Tat Protein

VIROLOGY

252, 126±136 (1998) VY989400

ARTICLE NO.

Intracellular Trafficking and Interactions of the HIV-1 Tat Protein Roland H. Stauber and George N. Pavlakis1 Human Retrovirus Section, ABL-Basic Research Program, NCI-FCRDC, Frederick, Maryland 21702-1201 Received June 15, 1998; returned to author for revision July 24, 1998; accepted August 27, 1998

Fusions of the human immunodeficiency virus type 1 (HIV-1) transactivator protein Tat to the green fluorescent protein (GFP) were used to study the intracellular localization, trafficking, and interactions of Tat in human cells. Tagging Tat with GFP did not change its nuclear localization or ability to act as a transactivator. Tat-GFP expressed at low levels was found in the nucleus, whereas overexpression resulted in nucleolar accumulation. A Tat-GFP hybrid protein containing in addition the HIV-1 Rev nuclear export signal (NES) localized predominantly to the cytoplasm. This shuttle protein, Tat-GFP-NES, transactivated the HIV-1 long terminal repeat. Thus a Tat molecule being only transiently present in the nucleus is active and nucleolar accumulation of Tat is not prerequisite for function. A coexpression assay previously used to define protein interaction domains in the HIV-1 Rev protein [R. H. Stauber, E. Afonina, S. Gulnik, J. Erickson, and G. N. Pavlakis (1998a). Virology 251, 38±48.] indicated that Tat exists predominantly as a monomer and does not form stable multimers with B23 in living cells. Using a heterokaryon fusion assay, we found that Tat-GFP was able to shuttle between the nucleus and the cytoplasm. Tat therefore has the potential to perform functions in the nucleus as well as in the cytoplasm.

Gibellini et al., 1998). In addition, it was suggested that Tat enhances also translation efficiencies (Cullen, 1986; Rosen et al., 1986; Wright et al., 1986) and modulates the cellular immune response (Viscidi et al., 1989; Seeger et al., 1997; Conant et al., 1998). The precise mechanisms by which nuclear Tat can perform these different functions and the role of cellular Tat co-factors are not fully characterized. By mutational analysis, two functional domains have been identified in the 16-kDa two-exon Tat protein. These include an aminoterminal transcriptional activation domain and a highly basic region that is involved in specific TAR binding and nuclear/nucleolar localization. Tat proteins with mutations in the basic regions display a transdominant (TD) phenotype (Pearson et al., 1990; Modesti et al., 1991). Since Tat was proposed to dimerize at least in vitro (Frankel et al., 1988a,b) TD Tat and wild-type Tat were suspected to form heterodimers inhibiting function (Pearson et al., 1990). Protein fusions to the green fluorescent protein (GFP) are powerful tools not only to study dynamic processes (for review, see Misteli and Spector, 1997) but also to probe protein interactions in living cells (Stauber et al., 1998a). Therefore, we used Tat-GFP hybrid proteins to address the intracellular localization, trafficking, and interactions of Tat. This study also underlines similarities and differences between the nuclear/nucleolar viral proteins Rev and Tat and helps to understand how Tat can perform additional functions besides transactivation.

INTRODUCTION Transcription of the human immunodeficiency virus type 1 (HIV-1) is regulated by the virus-encoded Tat protein and by cellular factors that interact with enhancer and promoter regions in the HIV-1 long terminal repeat (LTR) (for detailed reviews see, Peterlin, 1990; Jones and Peterlin, 1994; Gaynor, 1995; Jones, 1997; Pavlakis, 1997). Activation of HIV-1 gene expression by Tat is dependent on a cis-acting RNA stem loop structure, the transactivation response element (TAR), present at the 59 end of all viral RNAs. Tat is believed to function by interacting with and directing components of the cellular transcription machinery to the vicinity of the HIV-1 promoter (Towatari et al., 1998) and by stimulating the processivity of RNA polymerase II (Keen et al., 1996; Mavankal et al., 1996; Cujec et al., 1997a,b; Mancebo et al., 1997; Wei et al., 1998). Studies using HIV-1 clones deleted in the tat gene indicated that Tat appears to be also important for efficient reverse transcription and hence may function during preintegration (Harrich et al., 1997). Effects of Tat have been proposed to affect also uninfected cells. Several studies indicated that Tat is released from HIVinfected cells, endocytosed, and targeted to the nucleus of uninfected cells, where it can exert a series of pleiotrophic effects (Ensoli et al., 1990, 1993; Zauli et al., 1995;

1 To whom reprint requests should be addressed at Human Retrovirus Section, ABL-BRP, NCI-FCRDC, P.O. Box B, Frederick, MD 217011201. E-mail: [email protected].

0042-6822/98

126

Tat NUCLEOCYTOPLASMIC TRAFFICKING

127

FIG. 1. Schematic representation of Tat and Tat-GFP expression plasmids. The expression is under the control of the CMV early promoter, and the constructs contain the bovine growth hormone polyadenylation signal (polyA). pTat encodes the two exon Tat of HIV-1 (NL4±3). pTat-NES expresses a fusion of Tat with the nuclear export signal (NES) of Rev. pTat-GFP encodes a fusion between Tat and the complete GFP protein. pTat-GFP-NES expresses a hybrid protein between Tat and a concatemer of two GFP proteins linked by the Rev NES. GFP and NES are not drawn to scale.

RESULTS Nucleolar accumulation of HIV-1 Tat-GFP is caused by overexpression The Tat-GFP fusion protein (Fig. 1) allowed us to study the localization and trafficking of the HIV-1 Tat protein in living cells. Tat-GFP contained amino acids 1±86 of the HIV-1 Tat protein (strain NL4±3) and the complete GFP protein. Confocal laser scan microscopy (CLSM) of living HeLa cells, transfected with pTat-GFP, revealed that TatGFP was localized in the nucleus, concentrating at the nucleolus (Fig. 2A). Expression of untagged Tat by transfection of pTat resulted in similar nucleolar/nuclear localization as analyzed by indirect immunofluorescence. Cotransfection of pB37R that is transcribed into an RNA containing the Tat-binding TAR element did not cause detectable changes in Tat-GFP localization (data not shown). The use of a enhanced GFP mutant (Stauber et al., 1998b) allowed us to study the localization of Tat-GFP also at low expression levels. In several stable Tat-GFP HeLa cell lines, we observed a heterogeneous nuclear Tat-GFP localization (Fig. 2B). In cells stably expressing higher levels of Tat-GFP, nucleolar accumulation could still be detected, whereas in low level expressing cells, Tat-GFP was distributed equally throughout the nucleoplasm. The fluorescence signal in the stable cell lines was on average 25 times lower compared to transient transfections, and Tat protein concentrations were under the detection level by indirect immunofluorescence (data not shown). To verify that the stable Tat-GFP cell lines express a biologically active protein, the cells were transfected with the LTR-driven Gag expression plasmid pB37 together with Rev. As a positive control, the previously described cell line HLtat, producing untagged Tat protein, was included. The levels of Tat transactivation in the stable HeLa Tat-GFP or Tat cell lines, respectively, were similar, as determined by quantitation of the pro-

duced p24 Gag protein (data not shown). Therefore, the localization of Tat-GFP observed in the stable cell lines is not an artifact of a biologically inactive protein, and nucleolar accumulation of Tat-GFP depends on the expression level of Tat-GFP. The HIV-1 Rev nuclear export signal overrides the nuclear retention of Tat The localization of a protein within the cell is the net result of a variety of dynamic processes. To change the localization of Tat without introducing mutations or deletions, we used a novel strategy by fusing the strong nuclear export signal (NES) of the HIV-1 Rev protein to the C terminus of Tat-GFP (Fig. 1). In contrast to the nuclear/nucleolar localization of Tat-GFP, the steadystate distribution of Tat-GFP-NES was predominantly cytoplasmic, indicating that the NES of Rev was dominant over the nuclear retention of Tat (Fig. 2C). Similar localization of Tat-GFP-NES was observed in other cell lines tested (data not shown). A functional Tat-GFP protein does not have to accumulate at the nucleolus for transactivation The localization of Tat in the stable Tat-GFP cell lines already indicated that nucleolar accumulation was not prerequisite for Tat function. The predominantly cytoplasmic shuttle protein Tat-GFP-NES represented an alternative approach to test this hypothesis. Thus we compared the ability of the different Tat-GFP proteins to transactivate HIV-1 LTR directed gene expression. We measured the expression of HIV-1 gag from the subgenomic expression vector pB37R, which is dependent on Rev and Tat activity. HeLa cells were transfected with pB37R and pBSRev together with different amounts of the Tat expression plasmids, and Gag production was quantitated in cytoplasmic extracts (Fig. 3). The presence of either

128

STAUBER AND PAVLAKIS

FIG. 2. Localization of Tat-GFP fusion proteins in living human cells. In transiently transfected cells (2 mg of pTat-GFP),Tat-GFP is accumulating at the nucleolus, whereas Tat-GFP-NES is cytoplasmic (A and C). In stable HeLa cell lines expressing low levels of Tat-GFP, the hybrid protein is distributed almost equally throughout the nucleus (B). FIG. 4. The HIV-1 Tat protein exists predominantly as a monomer in living cells. The presence of Tat or Tat-NES did not change the localization of Tat-GFP-NES or Tat-GFP, respectively. HeLa cells were cotransfected with pTat-GFP-NES (2 mg) and pTat (4 mg) or pTat-GFP (2 mg) together with pTat-NES (4 mg), respectively, and analyzed by CLSM. The same cells are shown in pictures (A)/(B) or (C)/(D). Tat and Tat-NES were detected by immunofluorescence using a polyclonal Tat antiserum followed by staining with a rhodamine labeled secondary antibody (B and D). The GFP signal was used to localize Tat-GFP-NES and Tat-GFP (A and C).

Tat NUCLEOCYTOPLASMIC TRAFFICKING

129

FIG. 6. Overexpression of the nucleolar protein B23 does not change the localization of Tat-GFP-NES. HeLa cells were cotransfected with pTat-GFP-NES (2 mg) and pB23 (4 mg) and analyzed by CLSM. The same microscopic field is shown in (A) and (B). B23 was detected by immunofluorescence (red) using a B23 specific monoclonal antibody followed by staining with a rhodamine conjugated secondary antibody (B). The GFP signal (green) was used to localize Tat-GFP-NES (A).

Tat, Tat-GFP or Tat-GFP-NES, stimulated the expression of Gag protein ;100-fold. As a control, cotransfection of pB37R together with pBSRev did not lead to Gag expression in the absence of Tat (Fig. 3, no Tat). In additional

transfection experiments using as low as 50 ng of plasmid DNA, we could not detect significant differences in transactivation by Tat-GFP and Tat-GFP-NES (data not shown). Equal transfection efficiencies were ensured by

FIG. 3. Tat-GFP and Tat-GFP-NES are active in HIV-1 LTR transactivation. HeLa cells were transfected with the indicated amounts of the different Tat expression vectors, together with 0.2 mg of pBSRev, 2 mg of the gag-expressing plasmid pB37R, and 2 mg of pCMV-GFPsg25. Two days later the cells were harvested and intracellular Gag production was measured by a p24gag antigen capture assay. Duplicate plates were used and the results were averaged. Similar results were obtained in two independent experiments. Transfection efficiencies were controlled quantitating GFP as described.

130

STAUBER AND PAVLAKIS

adding the plasmid pCMV-GFPsg25 to the transfection mix and quantitating the GFP signal in cytoplasmic extracts. The fluorescence resulting from the low amount of produced Tat-GFP or Tat-GFP-NES protein was not significantly higher than the background autofluorescence and 36-fold lower than the signal emitted by the coexpressed GFP (data not shown). This allowed us to use GFP expression as a transfection control in our assay. Comparison of the amounts of produced proteins indicated similar expression levels for pTat-GFP and pTatGFP-NES. In addition, by Western blot analysis using a polyclonal Tat antiserum, we could not detect proteolytic cleavage of Tat-GFP liberating untagged Tat which could account for the transactivation observed (data not shown). Therefore, the GFP-tag does not interfere with Tat's ability to transactivate, and a Tat molecule that is only transiently present in the nucleus appears to be functional in transactivation. HIV-1 Tat does not form stable multimers in living cells The biological activity of Tat-GFP-NES indicated that the conformation of Tat in the hybrid protein was comparable to untagged Tat. Thus we could test if the HIV-1 Tat protein has the potential to form stable multimers in living cells. Our assay was based on the change of the distribution of a tagged protein caused by the interaction with another protein localized in a different subcellular compartment. We already successfully applied this approach to define domains important for HIV-1 Rev oligomerization (Stauber et al., 1998a). For Rev, cytoplasmic Rev mutants could be localized to the nucleolus by coexpression of wild-type Rev, which accumulated at the nucleolus. We reasoned that if Tat exists predominantly as a multimer, it should be possible to colocalize the cytoplasmic shuttle protein Tat-GFP-NES to the nucleolus by Tat±Tat interactions by coexpression of untagged nucleolar Tat. However, cotransfection of pTat-GFPNES even with an excess of pTat, expressing untagged Tat, did not result in nuclear/nucleolar colocalization of green Tat-GFP-NES (Fig. 4A). The presence of unlabeled nuclear/nucleolar Tat in the same cell was verified by a Tat polyclonal antiserum in combination with a rhodamine-coupled secondary antibody (Fig. 4B). Since GFP fluorescence was not inactivated by the fixation procedure, it could be used to localize the Tat-GFP-NES fusion protein. Tat-GFP fluorescence was more sensitive than antibody staining under these conditions because only high levels of Tat-GFP-NES were stained in the cytoplasm by the Tat antiserum (4B). To control for the possibility that Tat was relocalized into the cytoplasm by interaction with Tat-GFP-NES, we cotransfected a twofold excess of pTat-NES together with pTat-GFP. pTatNES encodes a Tat-NES fusion protein without the GFPtag (Fig. 1), and the cytoplasmic localization was visual-

ized by immunofluorescence (Fig. 4D). Figure 4C illustrates that the nucleolar/nuclear localization of TatGFP was not changed by the presence of the cytoplasmic Tat-NES protein in the same cell. We concluded that Tat does not form stable multimers and, in contrast to Rev, exists predominantly as a monomer in living human cells. The HIV-1 Tat protein can shuttle between the nucleus and the cytoplasm Because Tat was suspected to perform additional functions in the cytoplasm, we investigated if Tat was capable of nucleo-cytoplasmic trafficking. We used a standard cell fusion assay in which Tat-GFP-transfected HeLa cells were fused to untransfected cells using PEG (Afonina et al., 1998). After PEG treatment the plasma membranes of the individual cells start to fuse, resulting in a multinuclear syncytium. If Tat-GFP is constantly exported from and imported into the donor nucleus, it will be imported also into the acceptor nuclei present in the syncytium over time. In Fig. 5A, a single cell expressing Tat-GFP is surrounded by a layer of nonexpressing cells. Even long exposure times of the CCD camera did not reveal detectable amounts of Tat-GFP in the neighboring cells (data not shown). Figures 5A±5D show the same microscopic field at various time points after fusion. Tat-GFP was easily detectable in the closest acceptor nuclei/nucleoli as early as 30 min postfusion, and after 2 h Tat-GFP was almost equally distributed throughout the syncytium. De novo protein synthesis was blocked by cycloheximide in the fusion experiments. Incubation of Tat-GFP expressing cells at 4°C for as long as 3 h, shortly after fusion, did not result in trafficking of Tat-GFP into the surrounding nuclei. Likewise, depleting endogenous ATP-pools by treatment with sodium azide and 2-deoxyglusose also blocked Tat-GFP nuclear export. As a control, a nonshuttling GFP hybrid protein of similar size (;40 kDa), RevM10BL-GFP, was not exported in our fusion assay (data not shown). Our results indicate that the shuttling of Tat is active and not caused unspecifically by passive diffusion. Tat does not form stable complexes with B23 in vivo Marasco et al. (19940 reported that Tat colocalized with the nucleolar protein B23. Because B23 was reported to be a shuttle protein (Borer et al., 1989), it could function as a carrier in the nuclear export of Tat. To test if Tat and B23 form stable complexes, a similar approach as described to examine Tat±Tat interactions was used. Figure 6A demonstrates that cotransfection of pTat-GFPNES with an excess of pB23, encoding B23, did not change the cytoplasmic localization of Tat-GFP-NES. Likewise, immunofluorescence demonstrated high levels of B23 in the nucleolus but not in the cytoplasm in the presence or absence of the cytoplasmic Tat-GFP-NES

Tat NUCLEOCYTOPLASMIC TRAFFICKING

131

FIG. 5. Nucleo-cytoplasmic trafficking of Tat-GFP in living cells after cell fusion. HeLa cells expressing Tat-GFP were fused by PEG to untransfected cells, and images were taken at the indicated time points postfusion using a CCD camera. (A) Single cell expressing Tat-GFP surrounded by a layer of nonproducing cells. (B) The same microscopic field shown 30 min postfusion. Tat-GFP is clearly detectable in the closest acceptor nuclei/nucleoli. (C) Sixty minutes after fusion. (D) After 2 h Tat-GFP is almost equally distributed throughout the syncytium. To block new protein synthesis, cycloheximid was present throughout the experiment. Exposure time was slightly increased in (D) to compensate for the decrease of the GFP-signal caused by the export of Tat-GFP into the acceptor nuclei. Similar export kinetics could be observed in at least three independent experiments.

protein (Fig. 6B). This finding together with the nuclear but not nucleolar localization of biologically active TatGFP in stable cell lines argues against a stable association of Tat and B23 essential for function. Tat-GFP-NES does not mediate RNA export Expression of Gag from the subgenomic construct pB37R, containing the TAR-structure and the Rev RNA binding site RRE, is dependent on Tat transactivation

and Rev-mediated nucleo-cytoplasmic RNA transport. If Tat remains stably associated with the TAR structure during or after transcription, we hypothesized that a Tat protein that shuttles efficiently should be active in RNA transport. Since Tat-GFP-NES is a TAR±RNA binding protein that shuttles via the Rev export pathway, we tested if it can replace the Rev/RRE axis. Cotransfection experiments of pB37R together with high amounts of pTat-GFP-NES or Tat-GFP, respectively, indicated that neither Tat-GFP-NES nor Tat-GFP was

132

STAUBER AND PAVLAKIS

FIG. 7. Tat-GFP and Tat-GFP-NES cannot replace Rev function. HeLa cells were transfected with the indicated amounts of the different Tat expression vectors, together with 2 mg of the gag-expressing plasmid pB37R, 2 mg pCMV-GFPsg25, in the presence or absence of 0.2 mg pBSRev. Intracellular Gag production was measured by a p24gag antigen capture assay as described. Duplicate plates were used and the results were averaged. Transfection efficiencies were controlled by quantitating GFP.

able to rescue Gag expression in the absence of Rev (Fig. 7). DISCUSSION The present study demonstrates that fusions of GFP to the HIV-1 Tat protein result in functional hybrid molecules. In contrast to the HIV-1 Rev-GFP protein, which was almost completely concentrated at the nucleoli (Stauber et al., 1995), substantial amounts of Tat-GFP could be detected in the nucleoplasm. The high proportion of nuclear Tat indicated that the affinity of Tat for the nucleolus was not as high as Rev's. This was supported by the finding that adding the nuclear export signal (NES) of Rev to Tat resulted in a predominantly cytoplasmic protein. The steady-state distribution of Tat-NES was cytoplasmic, suggesting nuclear export being more efficient than nuclear retention. Adding a strong NES to nuclear proteins represents a novel strategy to influence protein localization to study the influence of localization on function. It also underlines that protein localization is not static but controlled by the fine tuning of signals mediating nuclear import, export, or nuclear and cytoplasmic retention. In stable cell lines expressing low levels of Tat-GFP, the protein was almost equally distributed throughout the nucleoplasm. During HIV-1 infection, the levels of Tat are most likely even lower, suggesting that nucleolar accumulation is not important for function but facilitated by overexpression. In addition, transacti-

vation by Tat-GFP-NES demonstrated that a Tat protein which was only transiently present in the nucleus and did not accumulate at the nucleolus was able to interact successfully with the cellular transcription machinery. Although, we could not detect significant differences in transcriptional activity comparing low amounts of Tat versus Tat-NES in transactivation assays it is likely that nuclear Tat is advantageous, especially at the early stages of the HIV-1 life cycle. Ongoing experiments will address the potential of Tat-NES in the context of the complete virus during infection. Since Tat stimulates the processivity and not primarily the assembly of the transcription complex (Peterlin, 1990; Jones and Peterlin, 1994; Gaynor, 1995; Jones, 1997) we are currently addressing if other viral transactivators have to accumulate in the nucleus for function. On the analogy of the HIV-1 Rev protein, we found that Tat was also able to shuttle between the nucleus and the cytoplasm although, with a three- to fourfold slower transport rate in our cell fusion assay (R. Stauber, unpublished observations). The transport kinetic of Tat is comparable to the shuttling of other trafficking proteins (Dobbelstein et al., 1997; Roth et al., 1998; Sandri-Goldin, 1998). Therefore Tat represents another nuclear viral protein that is constantly shuttling between different cellular compartments. Recently, specific signals mediating nuclear export have been identified (for reviews see (GoÈrlich and Mattaj, 1996; Nakielny and Dreyfuss, 1997;

Tat NUCLEOCYTOPLASMIC TRAFFICKING

Nigg, 1997)). Analysis of Tat revealed no amino acid motifs homologous to the known NESs. Testing the N- or C-terminal half of Tat in a heterologous export system did not lead to the identification of a linear export signal (R. Stauber, unpublished observations). However, there are a variety of fast shuttling proteins like steroid hormone receptors (Guichon-Mantle et al., 1991; Madan and DeFranco, 1993) or RNA binding proteins (Lee et al., 1996; Afonina et al., 1998) where no export signals have been defined so far. We suggest that the NES of Tat is relatively weak and thus could use an alternative export pathway than the ``strong'' leucin-rich NES. Alternatively, Tat could be transported ``piggy-backed'' by Tat-interacting proteins. A potential candidate was B23, since it was reported to shuttle (Borer et al., 1989) and to localize to the same intranuclear compartment (Marasco et al., 1994). In our cotransfection assay, Tat and B23 did not form strong complexes in living cells. The slow trafficking rate of B23 (.16 h) reported by Borer et al. (1989) argues also against a predominant role of B23 in Tat's nucleocytoplasmic transport. Recently, Li (1997) reported interaction between B23 and Tat and cytoplasmic relocalization of Tat by cotransfection with a hybrid protein between b-galactosidase (bgal) and a part of B23. However, it has to be clarified if the relocalization of Tat was specific or was caused by the disturbance of the nucleolar structure by overexpression of the bgal-B23 hybrid protein which would result in cytoplasmic accumulation of Tat due to its shuttling activity. In addition, we observed that in stable cell lines Tat-GFP being not concentrated at the nucleolus was yet functional. We reasoned that the nucleolus, especially 5S-RNA, as suggested for Tat and Rev (Giel-Pietraszuk et al., 1997; Lam et al., 1998), or B23 may represent secondary binding sites for Tat when other primary nuclear binding sites are saturated. Demonstrating that Tat is not confined to the nucleus helps to understand how Tat can perform additional functions besides transcriptional activation in the cytoplasm. Among those, it was suggested that Tat enhances translation efficiency (Cullen, 1986; Rosen et al., 1986; Wright et al., 1986; Hauber et al., 1987; Braddock et al., 1993) and is necessary for efficient reverse transcription (Huang et al., 1994; Harrich et al., 1997). The latter could be mediated by incorporation of cytoplasmic Tat into the virion. In addition, it was demonstrated that Tat is efficiently released from HIV-1 infected cells (Ensoli et al., 1993; Zauli et al., 1995; Gibellini et al., 1998). Thus the shuttling of Tat would provide a continuous source to maintain extracellular secretion. To stimulate transcription, Tat has to bind to the TAR structure and subsequently to interact with factors stimulating the processivity of the RNA polymerase II holoenzyme (RNAPolII) (Jones, 1997). An indirect approach to test whether Tat remains tightly associated with TAR was to assay if Tat could function in RNA export via Tat/TAR

133

interaction. In analogy to the bacteriophage MS2 system (Venkatesan et al., 1992), the Tat/TAR interaction should also be able to replace the Rev/RRE axis (reviewed in Felber, 1997). Adding the efficient Rev NES to Tat-GFP ensured that the shuttling protein Tat-GFP-NES was targeted to the appropriate Rev export pathway. However, the interaction of Tat-GFP-NES with TAR could not substitute for Rev/RRE interaction. This is in agreement with data from experiments using the HIV-1 Tev protein. Tev which contains TAR-binding activity (by the Tat portion) and shuttling activity (mediated by the Rev portion) was also unable to substitute for Rev (D'Agostino, unpublished observations). These findings suggest that Tat is not bound efficiently to the TAR structure during or after transcription. Alternatively, since Tat exists predominantly as a monomer, multimerization might be prerequisite to mediate RNA export as demonstrated for Rev (Stauber et al., 1998a; Thomas et al., 1998). Recently, Keen et al. (1997) demonstrated that the TAR±RNA does not remain associated with the Tat-RNAPII complex and a high-affinity TatTAR complex exists only transiently during the initial phase of transcription (Wei et al., 1998). To achieve RNA export by a Tat-Rev hybrid protein, Luo et al. (1993) had to use an array of 12 TAR elements. This set-up ensured that even after transcription sufficient Tat-Rev hybrid molecules remained attached to the multiple TAR elements to mediate RNA export. Reports on Tat as a metal-linked dimer were based on studies using prokaryotic produced Tat protein (Frankel et al., 1988a,b). However, Waszak et al. (1992) were able to purify an active monomeric recombinant Tat protein from Escherichia coli. We observed that cotransfection of Tat with Tat-GFP-NES or Tat-GFP together with Tat-NES, respectively, did not change the localization of the tagged Tat proteins. This provided evidence that Tat exists predominantly as a monomer in living human cells. In contrast to Tat, cytoplasmic localized Rev mutants could be colocalized to the nucleolus upon coexpression of wild-type Rev, demonstrating the feasibility of this in vivo approach (Stauber et al., 1998a). However, we can not completely exclude the possibility that Tat exists transiently as a loose oligomer. The data from Rice and Chan (1991) who showed that Tat is a monomer in cell extracts are also in agreement with our finding. Demonstrating that Tat is predominantly a monomer in vivo favors the model that trans-dominant (TD) Tat mutants inhibit the activity of wild-type Tat by sequestering cellular cofactors (Modesti et al., 1991; Carroll et al., 1992) rather than by the formation of inactive heterocomplexes as suggested for TD Rev (Hope et al., 1992; Stauber et al., 1995, 1998a; Szilvay et al., 1995). Visualizing the nucleocytoplasmic shuttling of Tat explains how even cytoplasmic Tat mutants can be TD (Orsini and Debouck, 1996). In summary, we have demonstrated that fusions to GFP can be used to study the structure, function, and interactions of the HIV-1 Tat protein. Our results suggest

134

STAUBER AND PAVLAKIS

that Tat exits predominantly as a monomer and nucleolar accumulation of Tat is not necessary for transactivation. A Tat molecule being present only transiently in the nucleus appears to be sufficient to stimulate the processivity of the RNA-PolII machinery. Tat is able to shuttle between the nucleus and cytoplasm, suggesting that Tat has the potential to perform additional functions in the cytoplasm which makes it again an attractive target for antiviral approaches. MATERIALS AND METHODS Cells and transfections The human cell lines HeLa and 293 (Stauber et al., 1995) were used for microscopic studies and to measure LTR directed transactivation by Tat. HLtat is a HeLaderived cell line that constitutively produces Tat protein (Schwartz et al., 1990). Cells were transfected by the calcium phosphate coprecipitation technique, and cytoplasmic extracts were prepared as described (Stauber et al., 1995). The production of intracellular HIV-1 Gag protein was measured by a p24gag antigen capture assay (Cellular Products) as described (Stauber et al., 1995). To control for transfection efficiencies, 2 mg of the GFPexpressing plasmid pCMV-GFPsg25 (Stauber et al., 1998b) was cotransfected with the indicated plasmids. The GFP signal in cytoplasmic extracts was quantitated in a Cytofluor II plate reader (Perseptive Biosystems, MA) as described (Stauber et al., 1998b). Recombinant plasmids pBsrev produces HIV-1 Rev from the HIV-1 LTR promoter (D'Agostino et al., 1995). Plasmid pB37R contains the HIV-1 59 LTR promoter, the p37gag coding region, RRE330, and the HIV-1 39 LTR (Stauber et al., 1995). This construct expresses gag in a Tat- and Rev-dependent manner. Plasmid pCMV-GFPsg25 encodes a mutated GFP protein under the control of the early CMV promoter (Stauber et al., 1998b). To generate pTat-GFP, the coding region of the HIV-1 two exon Tat (strain NL4±3) was amplified by PCR using appropriate primers containing NheI restriction sites. The PCR product was digested with NheI and cloned into the NheI cut vector pCMVGFPsg25. In the construct pTat, the GFP open reading frame (ORF) was removed by digesting pTat-GFP with EcoRV and HpaI. To generate a Tat protein containing the Rev nuclear export signal (NES), complementary oligonucleotides coding for the Rev NES (amino acids PLQLPPLQRLTLN), were annealed and cloned into the EcoRV/ HpaI-digested pTat-GFP after removal of the DNA fragment encoding GFP. The resulting construct pTat-NES produces a Tat protein fused to the HIV-1 Rev NES. pTat-GFP-NES expresses a Tat-GFP hybrid protein containing the Rev NES. To obtain this plasmid, the GFP ORF of pCMV-GFPsg50 (Stauber et al., 1998b) was PCR am-

plified and cloned into the NheI site of pCMV-GFPsg25. This resulted in the plasmid pCMV-GFP50±25, which codes for a fusion of two GFP proteins connected by a linker sequence. This linker contained AflII and KpnI restriction sites in which the Rev NES was cloned using appropriate complementary oligonucleotides. The resulting construct pCMV-GFP50-RevNES-25 was digested with PmlI, and the DNA fragment containing the Rev NES was cloned into the PmlI-digested plasmid pTat-GFP. Plasmid pB23 encodes the nucleolar protein B23 under the control of the early CMV promoter and was constructed by digesting pF25-B23-GFP with EcoRV/HpaI, thereby removing the DNA fragment containing the GFP ORF. The plasmid pBsrevM10BL-GFP, coding for a HIV-1 Rev protein lacking the NES, was described earlier (Stauber et al., 1995). All plasmids were verified by sequence analysis as described (Stauber et al., 1995). Cell fusion assay HeLa cells transfected with pTat-GFP were seeded 24 h posttransfection with a fivefold excess of untransfected HeLa cells. The following day, cells were washed with PBS, and 2 ml of PEG (PEG-4000, 50% in PBS) (GIBCO BRL) was added for 2 min. Cells were extensively washed with PBS and further incubated at 37°C. All solutions were prewarmed to 37°C prior to use. To block new protein synthesis cycloheximide (25 mg/ml) was added 30 min before fusion and was present during the experiment. Cells were observed under phase contrast and fluorescent illumination and scanned for fusion involving one GFP-Tat donor cell and nonexpressing surrounding acceptor cells. For energy depletion, sodium azid (10 mM) and 2-deoxxglucose (6 mM) (Sigma) in glucose-free DMEM were added to the cells 30 min prior to fusion and were present during the experiment. Microscopy Cells were seeded into coated 50-mm glass-bottom dishes (MatTek Corp., Ashland, MA) in phenol-red-free DMEM and transfected 24 h later with the indicated plasmids. One day later, the cells were analyzed by a Zeiss LSM 410 Micro System in the confocal mode. For GFP excitation, an argon/krypton laser at 488-nm wavelength was used and emitted fluorescence was detected with a 515- to 540-nm band-pass filter. For rhodamine red excitation, a 568-nm helium/neon laser was used, and fluorescence was detected with a 590-nm cut-off filter. Normaski images were made using a 543-nm green laser and appropriate polarized lenses. Alternatively, cells were observed with an inverted fluorescence microscope (Zeiss Axiovert 135). GFP signals were obtained with a FITC-fluorescence filter set (Zeiss O9, excitation 450±490 nm, beam splitter 510 nm, emission filter .520 nm). Twelve-bit black-and-white images were captured using a digital CCD camera (Pho-

Tat NUCLEOCYTOPLASMIC TRAFFICKING

tometrix, AZ). Image analysis and presentation were performed using the IPLab Spectrum software (Scanalytics, Vienna, VA). The nuclear signal was obtained by measuring the pixel intensity in the nucleus. All pixel values were measured below saturation limits. Generation of stable Tat-GFP expressing cell lines HeLa cells were transfected with the plasmid pTatGFP by the calcium phosphate method. Two days after transfection single cell FACS sorting was performed into 96-well plates as described (Stauber et al., 1998b). Positive, Tat-GFP-expressing, single-cell clones were identified using an inverted fluorescence microscope and further expanded in G418-containing DMEM (500 mg/ml). Indirect immunofluorescence Transfected cells were fixed and permeabilized as described (Stauber et al., 1995). The cells were incubated with a polyclonal rabbit anti-Tat antiserum (Felber et al., 1990) (diluted 1:100 in PBS) or a monoclonal mouse anti-B23 antibody (undiluted hybridoma supernatant) for 1 h. Appropriate secondary rhodamine conjugated antiIgG antibodies (KRL) (diluted 1:200 in PBS) were added for 1 h, and cells were analyzed by microscopy. Extensive washings with PBS were performed after each step throughout the procedure. ACKNOWLEDGMENTS We thank J. Hauber and B. Felber for suggestions and discussions, P. Carney for technical assistance, J. Resau and R. Hudson for confocal microscopy, and H. Busch and M. Neumann for supplying the anti-B23 monoclonal antibody and the plasmid pF25-B23-GFP, respectively. This research was sponsored in part by the National Cancer Institute, DHHS, under contract with ABL. R.S. is supported by the DKFZ AIDSStipendium Programm. The contents of this publication do not necessarily reflect the views or policies of the Department of Health and Human Services nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

REFERENCES Afonina, E., Stauber, R., and Pavlakis, G. N. (1998). The human poly-A binding protein 1 shuttles between the nucleus and the cytoplasm. J. Biol. Chem. 273(21), 13015±13021. Borer, R. A., Lehner, C. F., Eppenberger, H. M., and Nigg, E. A. (1989). Major nucleolar proteins shuttle between nucleus and cytoplasm. Cell 56, 379±390. Braddock, M., Powell, R., Blanchard, A. D., Kingsman, A. J., and Kingsman, S. M. (1993). HIV-1 TAR RNA-binding proteins control TAT activation of translation in Xenopus oocytes. FASEB J. 7(1), 214±222. Carroll, R., Peterlin, B. M., and Derse, D. (1992). Inhibition of human immunodeficiency virus type 1 Tat activity by coexpression of heterologous trans activators. J. Virol. 66(4), 2000±2007. Conant, K., Garzino-Demo, A., Nath, A., McArthur, J. C., Halliday, W., Power, C., Gallo, R. C., and Major, E. O. (1998). Induction of monocyte chemoattractant protein-1 in HIV-1 Tat-stimulated astrocytes and elevation in AIDS dementia. Proc. Natl. Acad. Sci. USA 95(6), 3117±3121. Cujec, T. P., Cho, H., Maldonado, E., Meyer, J., Rainberg, D., and Peterlin, M. B. (1997a). The human immunodeficiency virus transac-

135

tivator protein Tat interacts with the RNA polymerase II holoenzyme. Mol. Cell. Biol. 17(4), 1817±1823. Cujec, T. P., Okamoto, H., Fujinaga, K., Meyer, J., Chamberlin, H., Morgan, D. O., and Peterlin, M. B. (1997b). The HIV transactivator TAT binds to the CDK-activating kinase and activates the phosphorylation of the carboxy-terminal domain of RNA polymerase II. Genes Dev. 11, 2645±2657. Cullen, B. R. (1986). Trans-activation of human immunodeficiency virus occurs via a bimodal mechanism. Cell 46, 973±982. D'Agostino, D. M., Ciminale, V., Pavlakis, G. P., and Chieco-Bianchi, L. (1995). Intracellular trafficking of the human immunodeficiency virus type 1 Rev protein: Involvement of continued rRNA synthesis in nuclear retention. AIDS Res. Hum. Retroviruses 11, 1063±1072. Dobbelstein, M., Roth, J., Kimberly, W. T., Levine, A. J., and Shenk, T. (1997). Nuclear export of the E1B 55-kDa and E4±34-kDa adenoviral oncoproteins mediated by a rev-like signal sequence. EMBO J. 16, 4276±4284. Ensoli, B., Barillari, G., Salahuddin, S. Z., Gallo, R. C., and Wong-Staal, F. (1990). Tat protein of HIV-1 stimulates growth of cells derived from Kaposi's sarcoma lesions of AIDS patients. Proc. Natl. Acad. Sci. USA 87(9), 3479±3483. Ensoli, B., Buonaguro, L., Barillari, G., Fiorelli, V., Gendelman, R., Morgan, R. A., Wingfield, P., and Gallo, R. C. (1993). Release, uptake, and effects of extracellular human immunodeficiency virus type 1 Tat protein on cell growth and viral transactivation. J. Virol. 67(1), 277±287. Felber, B. K. (1997). Regulation of mRNA expression in HIV-1 and other retroviruses. In ``mRNA Metabolism and Post-Transcriptional Gene Regulation,'' pp. 323±340. Wiley-Liss, Inc, New York. Felber, B. K., Drysdale, C. M., and Pavlakis, G. N. (1990). Feedback regulation of human immunodeficiency virus type 1 expression by the Rev protein. J. Virol. 64, 3734±3741. Frankel, A. D., Bredt, D. S., and Pabo, C. O. (1988a). Tat protein from human immunodeficiency virus forms a metal-linked dimer. Science 240, 70±73. Frankel, A. D., Chen, L., Cotter, R. J., and Pabo, C. O. (1988b). Dimerization of the tat protein from human immunodeficiency virus: A cysteine-rich peptide mimics the normal metal-linked dimer interface. Proc. Natl. Acad. Sci. USA 85, 6297±6300. Gaynor, R. B. (1995). Regulation of HIV-1 gene expression by the transactivator protein Tat. In ``Transacting Functions of Human Retroviruses'' (I. S. Y. Chen, H. Koprowski, A. Srinivasan, and P. K. Vogt, Eds.), pp. 51±77. Springer-Verlag, Berlin. Gibellini, D., Bassini, A., Pierpaoli, S., Bertolaso, L., Milani, D., Capitani, S., La Placa, M., and Zauli, G. (1998). Extracellular HIV-1 Tat protein induces the rapid Ser133 phosphorylation and activation of CREB transcription factor in both Jurkat lymphoblastoid T cells and primary peripheral blood mononuclear cells. J. Immunol. 160, 3891±3898. Giel-Pietraszuk, M., Barciszewska, M. Z., Mucha, P., Rekowski, P., Kupryszewski, G., and Barciszewski, J. (1997). Interaction of HIV Tat model peptides with tRNA and 5S rRNA. Acta Biochim. Pol. 44(3), 591±600. GoÈrlich, D., and Mattaj, I. W. (1996). Nucleocytoplasmic transport. Science 271, 1513±1518. Guichon-Mantle, A., Lescop, P., Christin-Maitre, S., Loosfelt, H., PerrotApplanat, M., and Milgrom, E. (1991). Nucleocytoplasmic shuttling of the progesterone receptor. EMBO J. 10, 3851±3859. Harrich, D., Ulich, C., Garcia-Martinez, L. F., and Gaynor, R. (1997). Tat is required for efficient HIV-1 reverse transcription. EMBO J. 16(6), 1224±1235. Hauber, J., Perkins, A., Heimer, E. P., and Cullen, B. R. (1987). Transactivation of human immunodeficiency virus gene expression is mediated by nuclear events. Proc. Natl. Acad. Sci. USA 84, 6364±6368. Hope, T. J., Klein, N. P., Elder, M. E., and Parslow, T. G. (1992). transdominant inhibition of human immunodeficiency virus type 1 Rev occurs through formation of inactive protein complexes. J. Virol. 66(4), 1849±1855. Huang, L. M., Joshi, A., Willey, R., Orenstein, J., and Jeang, K. T. (1994).

136

STAUBER AND PAVLAKIS

Human immunodeficiency viruses regulated by alternative transactivators: Genetic evidence for a novel non-transcriptional function of Tat in virion infectivity. EMBO J. 13(12), 2886±2896. Jones, K. A. (1997). Taking a new TAK on Tat transactivation. Genes Dev. 11, 2593±2599. Jones, K. A., and Peterlin, M. B. (1994). Control of RNA initiation and elongation at the HIV-1 promoter. Annu. Rev. Biochem. 63, 717±743. Keen, N., Churcher, M., and Karn, J. (1997). Transfer of Tat and release of TAR RNA during activation of the human immunodeficiency virus type-1 transcription eleongation complex. EMBO J. 16, 5260±5272. Keen, N. J., Gait, M. J., and Karn, J. (1996). Human immunodeficiency virus type-1 Tat is an integral component of the activated transcription -elongation complex. Proc. Natl. Acad. Sci. USA 93, 2505±2510. Lam, W.-C., Seifert, J. M., Amberger, F., Graf, C., Auer, M., and Millar, D. P. (1998). Structural dynamics of HIV-1 Rev and its complexes with RRE and 5S-RNA. Biochemistry 37, 1800±1809. Lee, S. L., Henry, M., and Silver, P. A. (1996). A protein that shuttles between the nucleus and the cytoplasm is an important mediator of RNA export. Genes Dev. 10, 1233±1246. Li, Y.-P. (1997). Protein B23 is an important factor for the nucleolar localization of the human immunodeficiency virus protein Tat. J. Virol. 71, 4098±4102. Luo, Y., Madore, S. J., Parslow, T. G., Cullen, B. R., and Peterlin, B. M. (1993). Functional analysis of interactions between Tat and the transactivation response element of human immunodeficiency virus type 1 in cells. J. Virol. 67, 5617±5622. Madan, A. P., and DeFranco, D. (1993). Bidirectional transport of glucocorticoid receptor across the nuclear envelope. Proc. Natl. Acad. Sci. USA 90, 3588±3592. Mancebo, H., Lee, G., Flygare, J., Tomassini, J., Luu, P., Zhu, Y., Blau, C., Hazuda, D., Price, D., and Flores, O. (1997). P-TEFb kinase is required for HIV Tat transcriptional activation in vivo and in vitro. Genes Dev. 11, 2633±2644. Marasco, W. A., Szilvay, A. M., Kalland, K. H., Helland, D. G., Reyes, H. M., and Walter, R. J. (1994). Spatial association of HIV-1 tat protein and the nucleolar transport protein B23 in stably transfected Jurkat T-cells. Arch. Virol. 139(1±2), 133±154. Mavankal, G., Du, S. I., Oliver, N., Sigman, D., and Gaynor, R. B. (1996). HIV-1 and HIV-2 Tat proteins specifically interact with RNA polymerase II. Proc. Natl. Acad. Sci. USA 93, 2089±2094. Misteli, T., and Spector, D. L. (1997). Application of the green fluorescent protein in cell biology and biotechnology. Nat. Biotech. 15, 961±964. Modesti, N., Garcia, J., Debouck, C., Peterlin, M., and Gaynor, R. (1991). Trans-dominant Tat mutants with alterations in the basic domain inhibit HIV-1 gene expression. New Biol. 3, 759±768. Nakielny, S., and Dreyfuss, G. (1997). Nuclear export of proteins and RNAs. Curr. Opin. Cell Biol. 9, 420±429. Nigg, E. A. (1997). Nucleocytoplasmic transport: Signals, mechanisms and regulation. Nature 386, 779±787. Orsini, M. J., and Debouck, C. M. (1996). Inhibition of human immunodeficiency virus type 1 and type 2 Tat function by transdominant Tat protein localized to both the nucleus and cytoplasm. J. Virol. 70(11), 8055±8063. Pavlakis, G. N. (1997). The molecular biology of human immunodeficiency virus type 1. In ``AIDS: Biology, Diagnosis, Treatment and Prevention'' (V. T. DeVita, S. Hellman, and S. A. Rosenberg, Eds.), Vol. 4, pp. 45±74. Lippincott-Raven Publishers, Philadelphia. Pearson, L., Garcia, J., Wu, F., Modesti, N., Nelson, J., and Gaynor, R. (1990). A transdominant tat mutant that inhibits tat-induced gene expression from the human immunodeficiency virus long terminal repeat. Proc. Natl. Acad. Sci. USA 87, 5079±5083. Peterlin, B. M. (1990). Transcriptional regulation of HIV. Hum. Retroviruses 119, 153±162. Rice, A. P., and Chan, F. (1991). Tat protein of human immunodeficiency virus type 1 is a monomer when expressed in mammalian cells. Virology 185, 451±454.

Rosen, C. A., Sodroski, J. G., Chun Goh, W., Dayton, A. I., Lippke, J., and Haseltine, W. A. (1986). Post-transcriptional regulation accounts for the trans-activation of the human T-lymphotropic virus type III. Nature 319, 555±559. Roth, J., Dobbelstein, M., Freedman, D. A., Shenk, T., and Levine, A. J. (1998). Nucleo-cytoplasmic shuttling of the hdm2 oncoprotein regulates the levels of the p53 protein via a pathway used by the human immunodeficiency virus rev protein. EMBO J. 17, 554±564. Sandri-Goldin, R. M. (1998). ICP27 mediates HSV RNA export by shuttling through a leucine-rich nuclear export signal and binding viral intronless RNAs through an RGG motif. Genes Dev. 12, 868±879. Schwartz, S., Felber, B. K., Benko, D. M., FenyoÈ, E. M., and Pavlakis, G. N. (1990). Cloning and functional analysis of multiply spliced mRNA species of human immunodeficiency virus type 1. J. Virol. 64, 2519±2529. Seeger, M., Ferrell, K., Frank, R., and Dubiel, W. (1997). HIV-1 Tat inhibits the 20S proteasome and its 11S regulator-mediated activation. J. Biol. Chem. 13, 8145±8148. Stauber, R. H., Afonina, E., Gulnik, S., Erickson, J., and Pavlakis, G. N. (1998a). Analysis of intracellular trafficking and interactions of cytoplasmatic HIV-1 Rev mutants in living cells. Virology 251, 38±48. Stauber, R., Gaitanaris, G. A., and Pavlakis, G. N. (1995). Analysis of trafficking of Rev and transdominant Rev proteins in living cells using Green Fluorescent Protein fusions: Transdominant Rev blocks the export of Rev from the nucleus to the cytoplasm. Virology 213, 439±449. Stauber, R. H., Horie, K., Carney, P., Hudson, E. A., Tarasova, N. I., Gaitanaris, G. A., and Pavlakis, G. N. (1998b). Development and applications of enhanced green fluorescent protein mutants. BioTechniques 24, 462±471. Szilvay, A. M., Brokstad, K. A., Kopperud, R., Haukenes, G., and Kalland, K.-H. (1995). Nuclear export of the human immunodeficiency virus type 1 nucleocytoplasmic shuttle protein Rev is mediated by its activation domain and is blocked by transdominant negative mutants. J. Virol. 69(6), 3315±3323. Thomas, S. L., Oft, M., Jaksche, H., Casari, G., Heger, P., Dobrovnik, M., Bevec, D., and Hauber, J. (1998). Functional analysis of the human immunodeficiency virus type 1 Rev protein oligomerization interface. J. Virol. 72, 2935±2944. Towatari, M., Kanei, Y., Saito, H., and Hamaguchi, M. (1998). Hematopoietic transcription factor GATA-2 activates transcription from HIV-1 long terminal repeat. Aids 12(3), 253±259. Venkatesan, S., Gerstberger, S. M., Park, H., Holland, S. M., and Nam, Y. S. (1992). Human immunodeficiency virus type-1 Rev activation can be achieved without Rev-responsive element RNA if Rev is directed to the target as a Rev/MS2 fusion protein which tethers the MS2 operator RNA. J. Virol. 66, 7469±7480. Viscidi, R. P., Mayor, K., Lederman, H. M., and Frankel, A. D. (1989). Inhibition of antigen-induced lymphocyte proliferation by Tat protein from HIV-1. Science 246, 1606±1608. Waszak, G. A., Hasler, J. M., McQuade, T. J., Tarpley, W. G., and Deibel, M. R. (1992). Purification of an active monomeric recombinant HIV-1 trans-activator. Protein Expr. Purif. 3, 126±133. Wei, P., Garber, M. E., Fang, S.-M., Fischer, W. H., and Jones, K. A. (1998). A novel CDK9-associated C-type cyclin interacts directly with the HIV-1 Tat and mediates its high-affinity, loop-specific binding to TAR RNA. Cell 92, 451±462. Wright, C. M., Felber, B. K., Paskalis, H., and Pavlakis, G. N. (1986). Expression and characterization of the trans-activator of HTLV-III/LAV virus. Science 234, 988±992. Zauli, G., La Placa, M., Vignoli, M., Re, M. C., Gibellini, D., Furlini, G., Milani, D., Marchisio, M., Mazzoni, M., and Capitani, S. (1995). An autocrine loop of HIV type-1 Tat protein responsible for the improved survival/proliferation capacity of permanently Tat-transfected cells and required for optimal HIV-1 LTR transactivating activity. J. Acquir. Immune Defic. Syndr. Human Retrovirol. 10, 306±316.