Synergistic activation of transcription by the mutant and wild-type minimal transcriptional activation domain of VP16.

VP16 activates transcription by stimulating initiation, and for this function the aromatic residue at position 442 within its activation domain is critical. Recent studies have suggested that VP16 also stimulates transcriptional elongation. It has been shown that VP16 can activate transcription tethered downstream of the transcription start site to RNA. Here, we analyze the synergistic activation features of hybrid VP16 fusion proteins when tethered simultaneously to RNA downstream of the start site and to DNA upstream of a promoter in order to investigate its role in postinitiation control of transcription. Upon targeting the VP16 activation domain simultaneously to both DNA and RNA, high levels of transcriptional synergism is observed. Importantly, a transcription-defective VP16 minimal activation domain (amino acids 413-453) mutated at critical residue 442 (phenylalanine --> proline) maintained synergism, when bound to RNA, with the DNA-bound wild-type VP16 minimal activation domain. Targeting of this "functionally defective" VP16 minimal activation domain via RNA and an intact activation domain via DNA allowed us to uncover a postinitiation activity for VP16 not previously detected in DNA targeting studies. We suggest that, in addition to stimulating initiation, VP16 also acts at a postinitiation step involving residues other than the critical residue at position 442 within the same 41-amino acid minimal activation domain (amino acids 413-453) required for initiation.

VP16 activates transcription by stimulating initiation, and for this function the aromatic residue at position 442 within its activation domain is critical. Recent studies have suggested that VP16 also stimulates transcriptional elongation. It has been shown that VP16 can activate transcription tethered downstream of the transcription start site to RNA. Here, we analyze the synergistic activation features of hybrid VP16 fusion proteins when tethered simultaneously to RNA downstream of the start site and to DNA upstream of a promoter in order to investigate its role in postinitiation control of transcription. Upon targeting the VP16 activation domain simultaneously to both DNA and RNA, high levels of transcriptional synergism is observed. Importantly, a transcription-defective VP16 minimal activation domain (amino acids 413-453) mutated at critical residue 442 (phenylalanine 3 proline) maintained synergism, when bound to RNA, with the DNA-bound wild-type VP16 minimal activation domain. Targeting of this "functionally defective" VP16 minimal activation domain via RNA and an intact activation domain via DNA allowed us to uncover a postinitiation activity for VP16 not previously detected in DNA targeting studies. We suggest that, in addition to stimulating initiation, VP16 also acts at a postinitiation step involving residues other than the critical residue at position 442 within the same 41-amino acid minimal activation domain (amino acids 413-453) required for initiation.
VP16 protein is a component of the herpes simplex virus virion and in addition to its structural role selectively activates transcription of the viral immediate early genes during lytic infection by interacting with host cellular proteins that bind to specific enhancer elements located upstream of each immediate early gene (reviewed in Thompson and McKnight (1992)). The C-terminal 78-amino acid region of VP16 (amino acids 413-490) is highly acidic, and when fused to the DNA binding domain of the heterologous yeast protein GAL4 can activate transcription potently from target promoters containing GAL4binding sites (Sadowski et al., 1988;Cousens et al., 1989). Mutational analysis of the C-terminal activation domain of VP16 suggests that this domain can be subdivided into the proximal subdomain extending from amino acids 413-456 and the most C-terminal distal subdomain spanning amino acids 457-490. It is concluded, from studies using mammalian cells, that both subdomains are required for overall efficient transcriptional activation by VP16 (Regier et al., 1993;Walker et al., 1993). However, the proximal subdomain (amino acids 413-456) of VP16 constitutes the minimal activation domain and, when fused to GAL4 DNA binding domain, can activate transcription from a minimal promoter, retaining about 50% of the activity of the full-length domain Greaves and O'Hare, 1989). Studies have shown that the aromatic residue phenylalanine at position 442 (Phe 442 ) is critical for the function of this truncated activation domain rather than its net negative charge Regier et al., 1993). Mutation of Phe 442 to proline (F442P) within this minimal activation domain abolished transcriptional activity in both in vivo and in vitro transcription assays even when tethered to multiple DNA-binding sites highlighting the critical role of this residue and subdomain in the activation process Regier et al., 1993;Walker et al., 1993;Jiang et al., 1994). This region of VP16 interacts in vitro with components of the basal transcription complex, TFIID 1 (Stringer et al., 1990) and TFIIB . Mutation of phenylalanine to proline (F442P) disrupted the interaction with both TFIIB and TBP in vitro (Ingles et al., 1991;. It is, therefore, believed that VP16 stimulates transcription by interactions with TFIID and/or TFIIB to increase assembly of the preinitiation complex (Choy and Green, 1993;Chi et al., 1995). VP16 has also been shown in vitro to bind basal transcription factors, TFIIH (Xiao et al., 1994) and human positive coactivator 4 (PC4) (Ge and Roeder, 1994), and mutation at position 442 (F442P) substantially reduced these interactions.
The distal most C-terminal 39-amino acid region has been shown to constitute an independent domain and to have the ability to activate transcription, but poorly compared to the proximal region (Regier et al., 1993;Walker et al., 1993). Interestingly, this region of VP16 was found in vitro to directly bind dTAF II 40 and hTAF II 32, subunits of Drosophila and human TFIID, respectively, both of which in turn bind to TFIIB, suggesting a ternary interaction among the activator VP16, the coactivator TAF, and the basal transcription initiation factor TFIIB (Goodrich et al., 1993;Klemm et al., 1995). Thus, this subdomain may also contribute to VP16's transcriptional activity by stimulating initiation of transcription. Recent studies have suggested that, in addition to stimulating transcription initiation, VP16 also plays a role in transcription elongation (Yankulov et al., 1994;Krumm et al., 1995).
Here, in order to investigate VP16's role in postinitiation control of transcription, we have taken an independent in vivo functional approach and have analyzed the synergistic activation features of VP16 when tethered simultaneously upstream of a promoter to DNA and downstream of the transcription start site to RNA. Cullen and colleagues reported earlier that the VP16 activation domain could activate transcription from the human immunodeficiency virus-1 long terminal repeat (HIV-1 LTR) when tethered to a promoter-proximal RNA element (Tiley et al., 1992). We have constructed a modified HIV-1 LTR reporter plasmid, pG6RRE (see Fig. 1A) which allowed us to target fusion protein derivatives of VP16 simultaneously to RNA and DNA in transfection experiments in mammalian cells. We find that RNA-bound VP16 can cooperate with DNAbound VP16 to activate transcription synergistically. Furthermore, the minimal activation domain of VP16 (amino acids 413-453) was found to be sufficient for synergistic activation of transcription via RNA and DNA. Activation was abolished when the DNA-bound VP16 minimal activation domain was mutated at critical residue 442 (F442P). However, the RNAbound VP16 minimal activation domain mutated at critical residue 442 maintained synergism with DNA-bound VP16. Our studies suggest that the 41-amino acid minimal activation domain of VP16 (amino acids 413-453), which plays an important role in transcription initiation, also acts at a postinitiation step to increase rates of transcription. Furthermore, this postinitiation activity of the minimal activation domain does not involve the critical residue at position 442 required for initiation. Thus, VP16 activates transcription by a sequence of at least two major events. The first event involves stimulation of initiation upon interaction of both the proximal and distal subdomains of VP16 activation domain with components of the basal transcription initiation machinery at the promoter. Next, distinct determinant(s) within the proximal 41-amino acid activation region of VP16 activation domain act at a postinitiation step to enhance rates of transcription leading to the overall potent activation characteristic of VP16.

MATERIALS AND METHODS
Cell Culture and Transient Transfection Assays-Jurkat cells were grown in RPMI 1640 medium supplemented with 10% fetal calf serum and antibiotics in tissue culture flasks. HeLa and CV1 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and antibiotics on 100-mm Corning tissue culture plates. All cells were incubated at 37°C in a humidified atmosphere containing CO 2 .
Jurkat cells were transfected with plasmid DNA using the DEAEdextran method (Sambrook et al., 1989). Briefly, 1 ϫ 10 7 cells were washed with phosphate-buffered saline and then incubated with 10 g of target reporter and 5 g of effector plasmids in 2 ml of serum-free RPMI 1640 medium containing 300 g of DEAE-dextran/ml for 2 h at 37°C. Carrier DNA (pGem7Zfϩ; Promega) was added where necessary to maintain the same total amount of DNA in every transfection. Subsequently cells were pelleted and resuspended in 10% dimethyl sulfoxide (made in phosphate-buffered saline) and incubated for 2 min. Cells were then washed twice with phosphate-buffered saline, and RPMI supplemented with 10% fetal calf serum was added. Cells were harvested 48 h after transfection, cell lysates prepared, and chloramphenicol acetyltransferase (CAT) assays performed on 25 g of lysate.
Transfection of Hela and CV1 cells were performed using the DEAEdextran method as described earlier (Ghosh et al., 1993). CAT assays were performed by the solvent partition method (Neumann et al., 1987) where CAT activity is expressed as the rate of formation of 3 H-labeled acetylated chloramphenicol and is normalized to the amount of protein in the lysates. Protein concentrations were determined by the Bio-Rad protein assay kit (Bio-Rad). Transfection efficiency was examined by cotransfection of RSV-luciferase reporter plasmid, and luciferase assays were performed using the luciferase assay systems kit (Promega).
Plasmid Constructions-The reporter plasmid pG6RRE is a derivative of plasmids G6(Ϫ119)deltaTAR (Southgate and Green, 1991) and pSLLIIB/CAT (Tiley et al., 1992) and contains HIV-1 LTR sequences from positions Ϫ119 to ϩ80 (Sanchez-Pescador et al., 1985) linked in cis to the CAT gene. Upstream of the NF-B sequences at Ϫ119 are introduced six synthetic GAL4 DNA-binding sites (Southgate and Green, 1991). Downstream, the apical essential region of the TAR element (ϩ30 to ϩ35) is replaced by sequences that introduced the 29-nucleotide SLIIB high affinity Rev-binding site and the last pyrimidine residue (ϩ25) of the TAR bulge is deleted resulting in a nonfunctional TAR element while the Rev response element (RRE) (SLIIB) sequence forms an in vivo target for the HIV-1 Rev RNA-binding protein. pG6RRE was created first by digesting G6(Ϫ119)deltaTAR (HIV-1 LTR sequence with TAR element deleted and six GAL4-binding sites placed upstream of the promoter at Ϫ119) with XhoI and HindIII, filling in to create blunt ends, and religating the plasmid, resulting in the removal of the PvuII site located in between the XhoI and HindIII restriction sites upstream of the GAL4 DNA-binding sites to give G6(Ϫ119)deltaTAR(ϪPvuII). This plasmid was next digested with PvuII to remove all sequences located between the PvuII site 3Ј of the TATA element in the LTR and the PvuII site within the downstream CAT coding sequences and replaced with a PvuII fragment derived from pSLIIB/CAT which restored HIV-1 LTR and CAT coding sequences and now also introduced the TAR/RRE (SLIIB) sequence to give the reporter plasmid pG6RRE used in this study.

Synergistic Activation of Transcription by VP16 Tethered
Simultaneously to Downstream RNA and Upstream DNA-VP16 has been shown to activate transcription from the HIV-1 LTR when tethered downstream of the transcription start site to a promoter-proximal RNA element (Tiley et al., 1992). In order to determine if VP16 tethered downstream to RNA can synergize with VP16 tethered to DNA upstream of a promoter, we have constructed a modified HIV-1 LTR reporter plasmid, pG6RRE, which allowed us to target the 78-amino acid Cterminal VP16 acidic activation domain (amino acids 413-490) simultaneously to RNA and DNA (Fig. 1A). In the pG6RRE reporter, HIV-1 LTR (Sanchez-Pescador et al., 1985) is linked to the reporter gene CAT and has six synthetic GAL4-binding sites introduced upstream of NF-B sequences (Ϫ119 relative to the transcription start site), and downstream the apical region of the TAR element (ϩ30 to ϩ35) is replaced by the SLIIB subdomain of the HIV-1 RRE. The TAR element is rendered nonfunctional, while the RRE (SLIIB) sequence forms an in vivo RNA target for the HIV-1 Rev RNA-binding protein (Tiley et al., 1992). The effector plasmid Rev-VP16 (previously described as pcRev/VP16) (Tiley et al., 1992) directs the synthesis of a hybrid protein containing the full-length C-terminal activation domain of VP16 (amino acids 413-490) fused to HIV-1 RNA-binding protein Rev which allows targeting to the RRE (SLIIB) RNA element, while in the effector plasmid GAL4-VP16  the same full-length activation domain of VP16 (amino acids 413-490) is fused to the GAL4 DNA-binding domain (amino acids 1-147) for targeting to GAL4-binding sites. Both Rev-VP16 (Tiley et al., 1992) and GAL4-VP16 (Lin and Green, 1991) have been shown previously to activate transcription by increasing levels of RNA expres-sion. Jurkat cells were transfected with pG6RRE reporter and either GAL4-VP16 or Rev-VP16 or both. Fig. 1B and Table IA shows that cotransfection of Rev-VP16 and GAL4-VP16 along with pG6RRE resulted in high levels of synergistic activation: activation level was ϳ6-fold higher than the sum of the levels obtained with either fusion derivative alone (307-fold activation with both effectors versus 14-fold with GAL4-VP16 alone and 38-fold with Rev-VP16 alone). Thus RNA-bound VP16 acted synergistically with DNA-bound GAL4-VP16. The GAL4 DNA-binding domain (GAL4) and the Rev RNA-binding domain (Rev) did not have any effects on activation either individually or when coexpressed, indicating that the synergism obtained with the hybrid fusions is specifically due to the activation domains of VP16. The Rev-VP16 effector had no effect, when cotransfected with GAL4-VP16 on an HIV-1 LTR reporter containing upstream GAL4-binding sites but lacking downstream RRE sequences and vice versa GAL4-VP16 had no effect, when cotransfected with Rev-VP16, on an HIV-1 LTR reporter containing downstream RRE sequence but lacking upstream GAL4-binding sites. These results confirm that the synergism obtained is a result of targeting VP16 to the promoter via both GAL4 DNA-binding site and to the RRE RNA target (data not shown). It has been previously shown that the Rev-VP16 fusion activates transcription from an HIV-1 LTR reporter plasmid containing the RRE (SLIIB) sequence but was inactive when the critical nucleotides within the core Revbinding site of the RRE (SLIIB) sequence was mutated and rendered inactive for interaction with the Rev protein (Tiley et al., 1992). Synergistic activation was also obtained using a Lex-VP16 fusion where the C-terminal 78-amino acid activation domain of VP16 is fused to the heterologous DNA binding domain (amino acids 1-87) of the bacterial LexA protein and the pLRRE reporter, a G6RRE derivative containing LexA DNA-binding sites (one site in pL1RRE and four in pL4RRE) instead of GAL4. Lex-VP16 and Rev-VP16 fusions activated transcription from the pLRRE reporters synergistically when cotransfected into Jurkat cells. Greater than 3-fold synergistic levels of activation were obtained (data not shown). Furthermore, synergistic trans-activation of the pG6RRE reporter by GAL4-VP16 and Rev-VP16 were also obtained in another human cell line, HeLa, and in the monkey cell line, CV1, thereby indicating that the synergism seen is not cell type-dependent.
The Minimal Activation Domain (Amino Acids 413-453) of VP16 Is Sufficient for Synergistic Activation of Transcription when Tethered Simultaneously to RNA and DNA-The activation domain of VP16 (amino acids 413-490) has been subdivided into two independent activation domains, VP16N (proximal amino acids 413-456) and VP16C (distal amino acids 457-490) (Regier at al., 1993, Goodrich et al., 1993. O'Hare and colleagues refer to the proximal and distal portion of the VP16 activation domain as H1 (residues 410 -452) and H2 (residues 453-490), respectively (Walker et al., 1993). To determine if the residues involved in synergistic activation are located in the proximal N region (amino acids 413-453) or distal C region (amino acids 454 -490) of VP16 activation domain, we used truncated GAL4 and Rev derivatives of VP16 ( Fig. 2A). In GAL4-VP16N (previously described as SW3) (Walker et al., 1993) and Rev-VP16N effectors, the proximal minimal activation domain of VP16 (N region) is fused to GAL4 and Rev, respectively. Rev-VP16C expresses the distal C region of VP16 fused to Rev while in GAL4-VP16C the C region is fused to GAL4 DNA-binding domain. Fig. 2B and Table IB show that the N region of VP16 is sufficient for synergism. Rev-VP16N together with GAL4-VP16N resulted in ϳ8-fold synergistic levels of activation (120-fold activation with both effectors versus 10-fold with GAL4-VP16N and 5.5-fold with FIG. 1. VP16 tethered simultaneously to RNA and DNA act synergistically in trans-activating the HIV-1 LTR. A, schematic drawing of the reporter and effector plasmids used in cotransfections. In the pG6RRE reporter, HIV-1 LTR (Ϫ119 to ϩ80 relative to transcription start site) is linked to the reporter gene CAT and has six synthetic GAL4-binding sites introduced upstream of NF-B sequences and downstream the apical region of TAR element (ϩ30 to ϩ35) is replaced by the SLIIB subdomain of the HIV-1 RRE. The TAR element is rendered nonfunctional while the SLIIB RRE sequence forms an in vivo RNA target for the HIV-1 Rev RNA-binding protein (Tiley et al., 1992). GAL4-VP16 directs the synthesis of a hybrid protein containing the DNA-binding domain of GAL4 (amino acids 1-147) fused to the Cterminal acidic activation domain of VP16 (amino acids 413-490), while Rev-VP16 expresses the 113-amino acid Rev protein fused in-frame with the C-terminal acidic activation domain of VP16 (amino acids 413-490). B, Jurkat cells were transfected with pG6RRE and either GAL4-VP16 or Rev-VP16, or both. GAL4-VP16 fusion targets VP16 to GAL4 DNA and Rev-VP16 targets VP16 to RRE (SLIIB) RNA. At 48 h after transfection, cells were harvested, cell lysates were prepared, and the level of CAT enzyme activity was determined by the solvent partition method of Neumann et al. (1987) where CAT activity is expressed as the rate of formation of 3 H-labeled acetylated chloramphenicol (cpm/ min) and is normalized to the amount of protein in the lysate. Transactivation of the HIV-1 LTR reporter was monitored by CAT enzyme activity, and the data are reported as fold-trans-activation, where levels of CAT activity in the presence of the effector are divided by the basal activity level of the reporter in the absence of the effector and in the presence of carrier DNA (pGEM-7Zfϩ vector DNA, Promega). When Rev-VP16 and GAL4-VP16 were cotransfected alone, GAL4 and Rev, respectively, or vector DNA were used as carrier. The basal activity of the pG6RRE reporter plasmid as measured by CAT activity is 1.85 cpm/min. Synergistic activation was obtained when GAL4-VP16 and Rev-VP16 were cotransfected. GAL4 and Rev did not have any effects on activation indicating that the synergism obtained with the hybrid fusions is specifically due to the activation domains of VP16. The data presented are representative of three transfections, each done in duplicate. Transfection efficiency was examined by cotransfection of RSVluciferase reporter plasmid. Less than 10% variation in luciferase activity from the mean was detected among the different transfections.
Rev-VP16N alone) when cotransfected into Jurkat cells. The C region tethered to RNA (Rev-VP16C) did not synergize with a GAL4-VP16N or a GAL4-VP16C fusion. Thus the determinants for synergistic activation by VP16 are located within the 41 amino acid proximal minimal activation domain (N region).
Transcription-defective VP16 Minimal Activation Domain Mutated at Critical Residue 442 Maintains Synergism when Tethered to RNA with Wild-type VP16 Activation Domain Tethered to DNA-To define the nature of the synergism observed with VP16, we used truncated VP16N fusion derivatives where the critical aromatic residue phenylalanine at position 442 is mutated to proline (Fig. 3, A and B). GAL4-VP16Nm442 and Rev-VP16Nm442 have the mutated N region of VP16 activation domain fused to GAL4 and Rev, respectively. In context of the truncated minimal activation domain of VP16, mutation of the aromatic residue at 442 to proline abolished activation when this region was targeted to promoter DNA in both in vivo and in vitro studies (Regier et al., 1993;Jiang et al., 1994). Consistent with previous results, no activation was obtained with the GAL4-VP16Nm442 fusion even when targeted via multiple GAL4-binding sites. Neither did Rev-VP16Nm442 activate transcription via RNA. However, when Rev-VP16Nm442 was cotransfected with GAL4-VP16N, once again synergistic levels of activation were obtained ( Fig. 3C and Table IIB). The full-length activation domain of VP16 (amino acids 413-490) mutated at residue 442 (Phe 3 Pro) and targeted to RNA as a Rev-VP16m442 fusion and which did not significantly activate transcription by itself, also maintained synergism with the full-length intact GAL4-VP16 fusion. ϳFivefold synergistic levels of activation were obtained when Rev-VP16m442 was coexpressed with GAL4-VP16 ( Fig. 3C and Table IIA). Thus the "functionally defective" VP16 activation domain when targeted via RNA can cooperate with a functionally intact DNA bound VP16 activation domain leading to higher levels of activation. No significant synergism was detected between a wild-type Rev-VP16N fusion targeted to RNA and a mutated VP16N region (GAL4-VP16Nm442) targeted to upstream DNA. Also mutated VP16N activation domains did not synergize with each other when simultaneously targeted to DNA and RNA via GAL4-VP16Nm442 and Rev-VP16Nm442 fusions, respectively. These results indicate that for synergism to occur the aromatic residue at position 442 in the DNA-bound VP16 activation domain is critical, while this same residue is dispensable for the RNA-bound VP16. The data taken together suggest that synergistic activation by VP16 targeted simultaneously to DNA and RNA occurs by a sequence of at least two events. First, VP16 acts to stimulate rates of initiation upon interaction with basal components of the transcription initiation machinery. Thus, the requirement for the critical aromatic residue at position 442 within the minimal activation domain which has been extensively documented to play a crucial role in initiation and has been shown in in vitro studies to contact directly components of the basal transcription complex. Next, residues within the same 41-amino acid minimal activation domain required for initiation, but not including the critical aromatic residue at position 442, interact with the transcription complex to enhance overall rates of transcription. The mutated minimal activation domain is incapable of stimulating activation from the promoter by itself, and its activity is dependent and follows the stimulation of initiation by this domain mediated by the aromatic residue at 442. This second step activity of VP16 is most likely a postinitiation event, since targeting the mutated VP16 activation domain via DNA upstream of the promoter did not synergize with a functional VP16 activation domain targeted downstream via RNA. Cotransfection of GAL4-VP16Nm442 and Rev-VP16N did not lead to any further increase in activation levels over that of Rev-VP16N alone ( Fig.  3C and Table IIB). If this second pathway for VP16 functioned to activate the initiation step rather than act on a postinitiation stage, then one would have expected GAL4-VP16Nm442 to have synergized with Rev-VP16, similar to that observed upon cotransfecting GAL4-VP16N and Rev-VP16Nm442. Our studies, therefore, suggest that the 41-amino acid minimal activation domain of VP16, which plays an important role in transcription initiation, also acts at a postinitiation step to increase overall rates of transcription.

The Most C-terminal 39-Amino Acid VP16 Activation Subdomain Tethered to DNA Cooperates with the RNA Tethered 41-Amino Acid Minimal Activation Domain Mutated at Critical
Residue 442-It has been hypothesized that the C region of VP16 activation domain (amino acids 454 -490) may have an independent ability to activate transcription (Regier et al., 1993;Walker et al., 1993), and in vitro this region has been shown to bind dTAF II 40, a subunit of Drosophila TFIID b Basal CAT activity of target reporter plasmid in the absence of effector plasmids is expressed as the rate of formation of 3 H-labeled chloramphenicol acetyltransferase (cpm/min). Carrier DNA (pGEM-7Zf(ϩ), Promega) was used to keep the total amount of DNA transfected constant for all transfections.
c CAT activities of target plasmids in the presence of effector plasmids. d Fold-trans-activation was calculated by dividing CAT activities from cotransfections with target and effector plasmids over those obtained with target plasmids alone (column 4 over column 2). Values provided are from one set of transfections performed in duplicate. Standard errors of the mean for each transfection are given. (Goodrich et al., 1993) and also recently to hTAF II 32, the human homologue of TAF II 40 (Klemm et al., 1995). An expression plasmid containing two tandem copies of the VP16C region fused to GAL4 DNA binding domain (SW22) activated transcription weakly from a multisite target promoter (Walker et al., 1993). We predicted that if this C region is indeed capable of stimulating initiation from the promoter, then it would synergize when bound to DNA with the Rev-VP16Nm442 fusion. As shown in Fig. 3C, enhanced levels of activation are obtained when SW22 is cotransfected with the Rev-VP16Nm442 fusion, compared to each of the fusions alone. SW22 did not synergize with the Rev-VP16C fusion, further indicating that the determinants for the synergism observed above are located within the N region (data not shown). We conclude that the C region indeed has an independent activation domain. However, the GAL4-VP16C fusion containing a single copy of the C region of VP16 failed to synergize with both Rev-VP16N and Rev-VP16Nm442 (data not shown). It may be possible that our assay is not sensitive enough to detect synergism using a single copy of the C region. SW22 containing two copies of C region may have stimulated higher levels of initiation complexes from the promoter, sufficient for synergism to be detected in our assay. DISCUSSION Our analysis indicates that VP16 when tethered to RNA can synergize with upstream DNA-bound VP16 to activate transcription potently. Tethering the "functionally defective" minimal activation domain of VP16 mutated at critical residue 442 (F442P) via RNA and a wild-type minimal activation domain via DNA allowed us to uncover a second pathway for VP16 function, mediated by this minimal activation domain, hitherto unknown. This feature is distinct from its known function in initiation in that it does not require the participation of the critical residue Phe 442 . Our studies suggest that VP16 activates transcription by a sequence of at least two events. As shown in the model in Fig. 4, the first event involves both the proximal (N) and distal (C) regions of VP16 activation domain interacting with components of the basal transcription initiation machinery at the promoter to stimulate rates of initiation. The N region may stimulate initiation by contacts with TFIIB and TFIID, and for which the participation of the aromatic residue at 442 is critical Green, 1991, Ingles et al., 1991), while the C region may interact with TAF(s) in the TFIID complex (Goodrich et al., 1993;Klemm et al., 1995), leading to a stable initiation complex. VP16 has also been reported to bind in vitro to the human coactivator PC4 and the critical residue at 442 was found to be important for this interaction. Since PC4 directly binds TFIIA (or a TBP⅐TFIIA complex), it is suggested that PC4 mediates functional interactions between VP16 and TFIIA (or a TBP⅐TFIIA complex) (Ge and Roeder, 1991). Next, we hypothesize, once a stable initiation complex is assembled, the N region of VP16 activation domain may interact either directly or indirectly with target(s) in the transcription complex, modifying the complex to increase its processivity and for this function the aromatic residue at 442, found to be critical for initiation, is dispensable. The dispensability of this critical residue may suggest that the targets mediating this additional activity of VP16 may not include TFIID, TFIIB, or PC4, since interaction of these components with VP16 is disrupted by this mutation. The ability of VP16 to contact both general transcription factors involved in stimulating initiation and factor(s) modulating the processivity of chain elongation by RNA polymerase II would result in the potent activation characteristic of VP16. We suggest that the reason this postinitiation activity could not be detected in DNA targeting studies is that mutation of Phe 442 abolishes VP16's initiation function and consequently assembly of stable initiation complexes upon which VP16 could act to increase processivity. Tethering VP16 to RNA resulted in a large induction of transcription from the HIV-1 promoter even though multiple VP16 activation domains were tethered to DNA upstream of the promoter. We suggest that tethering VP16 to the 5Ј end of nascent RNA in close spatial proximity to the transcription start site may have allowed for a more efficient targeting of VP16 to the nascent elongating polymerase complexes initiated from the promoter and consequently for VP16 to exert its postinitiation processive activity leading to high overall rates of FIG. 2. The minimal activation domain of VP16 targeted simultaneously to RNA and DNA is sufficient for synergistic activation of transcription. A, schematic drawing of the effector plasmids coding for truncated hybrid fusions of VP16 with either GAL4 (1-147) or Rev used in cotransfections with the reporter, pG6RRE. The activation domain of VP16 (amino acids 413-490) has been divided into two independent activation domains, VP16N (proximal amino acids 413-452) and VP16C (distal amino acids 453-490). GAL4-VP16N and GAL4-VP16C are fusion's expressing the GAL4 (1-147) DNA-binding domain fused to VP16N and VP16C regions of VP16, respectively, while Rev-VP16N and Rev-VP16C fusion's express the Rev RNA-binding protein fused to VP16N and VP16C, respectively. B, trans-activation resulting from targeting truncated VP16 fusion derivatives simultaneously to RNA and DNA in pG6RRE reporter in Jurkat cells. The basal activity of pG6RRE as measured by CAT activity is 2.55 cpm/min. Reporter CAT activity data expressed as fold-trans-activation indicates that RNA-bound VP16N (Rev-VP16N) can synergize with DNA-bound VP16N (GAL4-VP16N) and that the N region contains all the determinants for synergism. RNA and DNA-bound VP16C (GAL4-VP16C and Rev-VP16C) did not synergize indicating that the synergism seen for VP16N is a specific event. Similar results were also obtained with CV1 and Hela cell lines. The data presented above are representative of three transfections, each done in duplicate.
transcription. Interestingly, the HIV-1 Tat protein which is an extremely potent viral trans-activator activates transcription from the HIV-1 promoter via interaction with the nascent TAR RNA element located at the 5Ј end of all HIV transcripts, and it has been shown by both in vivo and in vitro studies that Tat acts as an elongation factor and can also stimulate transcription initiation (see Zhou and Sharp (1995) and references therein). However, the mutated VP16 activation domain tethered to RNA did not have any effects on the basal activity of pG6RRE. Rev-VP16m442 or Rev-VP16Nm442 fusions did not alter the basal expression of pG6RRE, individually but synergized with an intact VP16 activation domain targeted to upstream DNA (Fig. 3C). The low levels of basal activity obtained with our modified HIV-1 LTR reporter, pG6RRE, is consistent with previous findings from several laboratories that the HIV-1 LTR promoter, in the absence of activator, is characterized by the production of low levels of productive full-length RNA. Phillip Sharp and colleagues have proposed that these RNA species are generated from the more processive class of polymerase elongation complexes that initiate from the HIV promoter (Marciniak and Sharp, 1991). We suggest that these more processive elongating polymerase complexes derived from the HIV promoter in the absence of activator may not be responsive any further to the processive activity of VP16. The processive function of VP16 specifically acts on the transcription initiation complexes stimulated from the promoter by the intact VP16 activation domain requiring critical residue 442, modifying it to increase its processivity resulting in increased rates of transcription.
It is generally believed that transcriptional synergy exhibited by VP16 on DNA may be a result of it interacting with more than one target affecting multiple steps involved in assembly of preinitiation complex. This hypothesis is supported by in vitro studies demonstrating GAL4-VP16 interactions with multiple basal transcription factors. However, transcriptional activation by RNA polymerase II is a multistep process, where each step may serve as a potential control point for activators and for transcriptional synergism. Transcriptional activation begins with assembly of a preinitiation complex followed by ATP-dependent events including melting of the DNA at the transcription start site to give an open complex and phosphorylation of the CTD (C-terminal domain) of RNA polymerase II. Next is the promoter clearance step resulting in the transition of the initiation complex to elongation complex and synthesis of phosphodiester bonds of the nascent RNA transcript followed by chain elongation (reviewed in Herschlag FIG. 3. Transcription-defective VP16 activation domain mutated at critical residue 442 (phenylalanine 3 proline) maintains synergism when tethered to RNA with wild-type VP16 activation domain tethered to DNA. A, amino acid sequence of the minimal activation domain of VP16 (VP16N; amino acids 413-453). The asterisk at position 442 highlights the aromatic residue phenylalanine shown to be critical for transcription activation in vivo and for interaction with basal transcription factors TFIIB, TFIID, TFIIH, and human coactivator PC4, in vitro. B, structure of the hybrid VP16 fusion plasmids and mutated derivatives used in cotransfections with the pG6RRE reporter. The phenylalanine residue at position 442 is mutated to proline in the GAL4-VP16N, Rev-VP16, and Rev-VP16N fusions to give GAL4-VP16Nm442, Rev-VP16m442, and Rev-VP16Nm442, respectively. C, CAT activity data from transfections in Jurkat cells show that the N region of VP16 when mutated at the critical residue 442 fails to activate transcription from the promoter when tethered to DNA, yet can maintain synergism when bound to RNA with a functional GAL4-VP16N fusion but not with the mutated GAL4-VP16Nm442 fusion tethered to upstream DNA. Basal activity of the pG6RRE reporter is 3.69 cpm/min for transfections shown in the top-most panel, and 2.55 cpm/min for the rest of the transfections shown. Levels of synergistic activation were always lower when the mutated Rev-VP16Nm442 fusion was cotransfected as compared to the unmutated Rev-VP16N. Also, the full-length VP16 activation domain mutated at Phe 442 and fused to Rev (Rev-VP16m442) synergized more effectively than the truncated and mutated Rev-VP16Nm442. We suggest that the Phe 3 Pro mutation at 442 within region N of VP16 may potentially disturb the structure of this region, resulting in lowered levels of activity. Higher levels of synergistic activation shown by Rev-VP16m442, containing both the mutated N and C region of VP16, may be due to the stabilization effect of the C region on the N domain that contains the potentially disrupting mutation at 442. and Johnson (1993)). It has been shown using HeLa cell nuclear extracts that binding of GAL4-VP16 to DNA was required for efficient open complex formation at the adenovirus E4 promoter, i.e. melting of the DNA around the transcription start site using energy derived from ATP hydrolysis to expose the template strand for subsequent events leading to the synthesis of RNA (Wang et al., 1992). Gralla and colleagues have shown that mutation of critical residue 442 within the VP16 activation domain which drastically affects transcriptional activity in vivo led to a defect in open complex formation in in vitro studies (Jiang et al., 1994). This finding is consistent with the previous studies showing that mutation at residue 442 (F442P) disrupts VP16's interaction with general transcription factors TBP, TFIIB, and PC4, all of which act at steps preceding open complex formation. Since the additional transcriptional activity of VP16 which we have uncovered in our synergism assay does not require the participation of the critical residue at 442, we suggest that this new activity of VP16 acts at a step that follows open complex formation. For example, VP16 could facilitate the steps leading to CTD phosphorylation which allows uncoupling of polymerase II from the promoter (reviewed in Eick et al. (1994)). Alternatively, this postinitiation activity of VP16 might act at the promoter clearance stage. It has been shown that TFIIE and TFIIH are not required for initiation, but are necessary for promoter clearance (Maxon et al., 1994;Goodrich and Tijan, 1994) and may be potential targets for the postinitiation activity of VP16. Of the two, TFIIE, may be a more likely candidate since interaction of VP16 with TFIIH has been shown to require residue 442 (Xiao et al., 1994). We suggest that the transcriptional synergism obtained when VP16 is tethered to multiple DNA-binding sites is a result of both the action of VP16 on multiple steps leading to preinitiation complex formation, via interactions with multiple targets, as well as on a postinitiation step(s) leading to overall potent levels of mRNA synthesis. However, our experimental approach has limitations and cannot determine the exact stage of the transcription process at which the proposed postinitiation activity of VP16 functions. We are currently examining the residue(s) within the N region of the VP16 activation domain which is(are) responsible for the postinitiation activity detected in our in vivo analysis of VP16 transcriptional activation, the stage in transcription at which it acts, and the cellular targets with which this region interacts using biochemical approaches. Our studies may lead to an understanding of the mechanism by which the herpes simplex virus VP16 protein triggers activation of the viral immediate early genes upon interaction with host factors, Oct-1, and host cell factor protein (reviewed in Thompson and McKnight (1992)). Finally, a dual role in both a HIV-1 LTR reporter plasmid which contains 6 GAL4 DNA-binding sites placed upstream of the HIV-1 promoter and RRE (SLIIB) RNA sequence downstream of the transcription start site.
b Basal CAT activity of target reporter plasmid in the absence of effector plasmids is expressed as the rate of formation of 3 H-labeled acetyl chloramphenicol transferase (cpm/min). Carrier DNA (pGEM-7Zf(ϩ), Promega) was used to keep the total amount of DNA transfected constant for all transfections.
c CAT activities of target plasmids in the presence of effector plasmids. d Fold-trans-activation was calculated by dividing CAT activities from cotransfections with target and effector plasmids over those obtained with target plasmids alone (column 4 over column 2). Values provided are from one set of transfections performed in duplicate. Standard errors of the mean for each transfection are given.
FIG. 4. Model for transcriptional activation by VP16. Analysis of synergistic activation features of VP16 acidic activation domain when targeted simultaneously to RNA downstream of the start site and to DNA upstream of the promoter suggest that VP16 in addition to stimulating rates of transcription initiation also plays a role in postinitiation control of transcriptional activation. We hypothesize that VP16 activates transcription by a sequence of at least two events. First, the activation domain of VP16 interacts with basal components of the transcription initiation machinery via both its proximal N region (amino acids 413-453) and the distal C region (amino acids 454 -490) to stimulate rates of transcription initiation from the promoter. This function of VP16 involves the participation of the critical aromatic residue at position 442 within the N region of its activation domain. Mutation of Phe 442 to proline (m442) drastically affects the initiation step and in context of the minimal activation domain (N region) abolishes transcription. Next, distinct determinant(s) within the same 41-amino acid N region of VP16 activation domain required for initiation, may interact with target(s) in the transcription complex at a postinitiation step to increase rates of transcription resulting in potent activation characteristic of VP16. Phe 442 critical for initiation is dispensable for this postinitiation function. transcription initiation and postinitiation may be a general property of other potent viral and cellular activators.