To analyze transcription stimulation by Tat and other activators we performed ChIP experiments in transiently transfected mammalian tissue culture cells. Plasmids expressing a reporter construct, containing a core promoter and various combinations of activator-binding sites, were transiently transfected into HeLa or 293T cells. In some experiments, a second plasmid expressing an activator was co-transfected. Reporter activity or primer extension was used to quantitate transcription levels. TC assembly was analyzed by a ChIP assay using antibodies directed against components of four representative GTFs involved in distinct stages of TC assembly: TFIID (TBP and typically TAF1 and TAF5), TFIIB, mediator (CDK8 and hMed6), and RNA polymerase II (RBP1). In most experiments the analysis was performed using two sets of primers: one set encompassed the core promoter, transcription start site, and immediate downstream region, and a second set was located far downstream of the transcription start site, within the open reading frame (ORF).
Diverse Activators Function by Increasing TC Assembly
In yeast, it has been shown that activators stimulate TC assembly, which is evident at the earliest step of this process, interaction of TBP with the
TATA box [18
]. To confirm that this was also the case in mammalian cells, we used a ChIP assay to analyze TC assembly in three well-characterized model systems: transcription directed by Gal4-VP16, transcription directed by the SV40 enhancer, and transcription directed by the adenovirus (Ad) E1a protein.
The artificial activator Gal4-VP16 contains an unusually potent acidic activation domain and can stimulate transcription in many species and cell types including mammalian tissue culture cells [5
]. A shows, consistent with previous studies, that on a synthetic promoter containing four Gal4-binding sites upstream of the
TATA box, transcription was not detectable in the absence of Gal4-VP16, whereas addition of Gal4-VP16 led to a large transcriptional increase. The accompanying ChIP assay shows that in the absence of Gal4-VP16 there was no detectable association of GTFs with the promoter. However, addition of Gal4-VP16 led to a large increase in recruitment of all GTFs analyzed, explaining the transcriptional stimulation. Consistent with the transcription results, RNA polymerase II was associated with the ORF in the presence but not absence of Gal4-VP16. Also, as expected, the GTFs were associated with the core promoter and not the ORF.
Diverse Activators Function by Increasing TC Assembly
Transcription of certain genes requires a cis-
acting enhancer element that is often located at a distance from the transcription start site (reviewed in [20
]). The enhancer functions by providing binding sites for cellular activators. For example, a minimal β-globin
promoter is inefficiently expressed, but addition of an appropriate enhancer element dramatically increases transcription [21
]. B shows, as expected, that a minimal rabbit β-globin
promoter was virtually inactive, whereas transcriptional activity was greatly increased by addition of the SV40 enhancer. The ChIP analysis shows that in the absence of the SV40 enhancer there was no detectable association of GTFs with the β-globin
promoter. Upon addition of the SV40 enhancer, there was a large increase in GTF recruitment that correlated with the increased transcriptional activity.
Finally, we analyzed transcription directed by the Ad E1a protein. Efficient transcription from Ad early promoters requires the viral E1a protein as well as the participation of cellular transcription factors that also bind to the promoter (reviewed in [22
]). For example, the Ad E4 promoter contains multiple binding sites for cellular ATF proteins, which are required for efficient transcription activation by E1a [23
]. C shows, as expected, that in the absence of E1a there was only a background level of Ad E4 transcription, and that E1a increased transcription dramatically. The accompanying ChIP assay shows that in the absence of E1a there was no significant association of GTFs with the promoter, whereas transfection of E1a resulted in recruitment of all GTFs analyzed in a manner that paralleled the increased transcriptional activity.
In summary, the results of indicate, as in yeast, that in three well-studied higher eukaryotic examples, transcription activation involves promotion of TC assembly. Moreover, as in yeast, the stimulatory effect is evident at the earliest step of TC assembly, the TBP–
TATA box interaction.
The HIV-1 Tat Protein Stimulates TC Assembly through Recruitment of TBP in the Absence of TAFs
We next investigated Tat-mediated transcription activation and TC assembly on the HIV-1 LTR promoter. A shows, as expected, that in the absence of Tat there was no detectable transcription from the HIV-1 LTR, and that addition of Tat increased transcription dramatically. The accompanying ChIP experiment (A, left) shows that in the absence of Tat, association of GTFs with the core promoter was virtually undetectable. Sp1, a constitutive cellular activator that binds upstream of the HIV-1 core promoter was, as expected, associated with the promoter in the absence of Tat and unaffected by Tat addition (A, right). Significantly, in the absence of Tat there was no detectable association of RNA polymerase II near the transcription start site or within the ORF (A, left). However, following addition of Tat, there was a large increase in association of TBP, TFIIB, mediator, and RNA polymerase II with the promoter, which paralleled the transcriptional increase. Also, as expected, there was a large increase in association of RNA polymerase II with the ORF. Following recruitment of P-TEFb to the HIV-1 LTR by Tat, P-TEFb associates with and phosphorylates RNA polymerase II [12
]. Accordingly, the ChIP assay (A, right) shows that in the presence of Tat the two P-TEFb subunits, CycT1 and CDK9, were present at both the promoter and the ORF. Unexpectedly, although TBP and the other GTFs were efficiently recruited to the promoter in the presence of Tat, there was no significant recruitment of the two TAFs analyzed, TAF1 or TAF5 (left panel).
The HIV-1 Tat Protein Stimulates TC Assembly through Recruitment of TBP in the Absence of TAFs
We note that recruitment of Tat was not observed in the ChIP assay, which detects proteins bound either directly or indirectly to DNA (reviewed in [25
]). Unlike all the other transcription factors, which associate with the TC through DNA–protein or protein–protein interactions, Tat is bound to nascent RNA [11
To determine whether the lack of TAF recruitment was a general feature of Tat-directed transcription, we analyzed a chronically HIV-1-infected cell line, 8E5/LAV, which harbors an integrated provirus that is constitutively transcribed [26
]. B shows that TBP, TFIIB, mediator, Sp1, P-TEFb, and RNA polymerase II, but not TAF1 or TAF5, were associated with the integrated proviral promoter, consistent with the results of the transient transfection assay.
We also analyzed a second chronically HIV-1-infected cell line, U1, which has been used as a model to study viral latency. U1 cells harbor an integrated viral genome, but, unlike 8E5/LAV cells, the constitutive level of viral expression is extremely low. Viral expression can be induced by phorbol esters, which stimulate transcription through upstream NF-κB-binding sites in the HIV-1 LTR, leading to Tat synthesis and a subsequent substantial Tat-mediated transcriptional increase [27
]. Thus, U1 cells provide an experimental system to analyze TC assembly from an integrated HIV-1 LTR in the inactive (−PMA) or active (+PMA) state. C shows that following activation of the HIV-1 LTR by PMA addition, there was a large increase in recruitment of TBP and RNA polymerase II, whereas TAF1 and TAF5 were once again not detected.
To investigate further the composition of the TC formed on the HIV-1 LTR, we obtained a panel of antibodies directed against nine additional human TAFs (TAF2, TAF4, TAF6, TAF7, TAF8, TAF9, TAF11, TAF12, and TAF13), as well as antibodies against four subunits of the S
TAGA complex (PAF65β, hSpt3, hGCN5, and TRRAP) and the TBP-interacting protein Mot1. The results (A) show that when transcription was directed by Gal4-VP16, all 11 TAFs analyzed were bound to the promoter, as expected for recruitment of TFIID. By contrast, when transcription was directed by Tat, none of the 11 TAFs were recruited to the HIV-1 LTR. Gal4-VP16 also recruited all four S
TAGA subunits analyzed, none of which were associated with the HIV-1 LTR in the presence of Tat. Finally, Mot1 was not recruited when transcription was directed by either Gal4-VP16 or Tat. On the basis of these data we conclude that the TC formed on the HIV-1 LTR lacks at least 11, and probably all, of the 14 TFIID TAFs. In the experiments presented below, TAF1 and TAF5 were analyzed as representative TAFs.
TAFs Are Not Recruited to the HIV 1-LTR and Not Required by Tat for Transcription Activation
TAFs Are Not Required for Tat-Mediated Transcription Activation
The results of the ChIP experiments strongly suggested that transcription activation by Tat did not require TAFs. To confirm this supposition, we examined the ability of Tat to stimulate transcription following TAF inactivation. First, we examined the ability of Tat to activate transcription in ts13 cells, which harbor a temperature-sensitive mutation in TAF1 [29
]. B shows, as expected, that inactivation of TAF1 significantly diminished transcription activation by VP16 and E1a, whereas Tat-mediated transcription activation was unaffected. In a second approach, we used short hairpin RNAs (shRNAs) to knockdown expression of TAF5 or TAF12 in 293A cells. Immunoblot analysis confirmed that shRNA-mediated knockdown of TAF expression was efficient and specific (C, left). Transcriptional analysis revealed that knockdown of TAF5 or TAF12 substantially decreased transcription directed by VP16, whereas Tat-mediated transcription activation was unaffected (C, right). (Because 293A cells are transformed by and express E1a, activation by E1a could not be analyzed in this experiment.) Collectively, the results of demonstrate that TAFs are not involved in transcription activation directed by the HIV-1 Tat protein.
Tat and P-TEFb Direct Recruitment of TBP in the Absence of TAFs
The results described above indicate that on the HIV-1 LTR transcription activation involves assembly of an atypical TC that contains TBP but not TAFs. However, these experiments do not distinguish whether the critical determinant for recruitment of TBP and not TFIID is Tat or the promoter, the HIV-1 LTR. To address this issue we performed a series of artificial tethering experiments.
Previous studies have shown that Tat can activate transcription when directed to the HIV-1 LTR through a heterologous DNA-binding domain [31
]. We analyzed TC assembly using an HIV-1 LTR derivative that lacked the TAR element and contained upstream Gal4-binding sites. We first examined the TCs formed on this promoter by three Gal4 fusion proteins: Gal4-VP16, Gal4-E1a, and Gal4-Tat. Consistent with previous studies [32
], shows that all three Gal4 fusion proteins activated transcription from this modified HIV-1 LTR derivative. The accompanying ChIP experiments show that Gal4-VP16 and Gal4-E1a supported assembly of a TC that contained all of the GTFs, including TAF1 and TAF5. By contrast, Gal4-Tat directed assembly of a TC in which the TAFs were present at a level significantly below that of TBP and other GTFs. These results indicate that the activator Tat, and not the HIV-1 LTR promoter, directs the selective recruitment of TBP in the absence of TAFs.
Tat and P-TEFb Direct Recruitment of TBP in the Absence of TAFs When Tethered to DNA
As discussed above, Tat interacts directly with CycT1, a subunit of the transcription elongation factor P-TEFb. Thus, Tat is an adaptor whose function is to recruit P-TEFb. We therefore considered that P-TEFb was ultimately responsible for the selective recruitment of TBP in the absence of TAFs. To address this possibility, we analyzed transcription activation by the two P-TEFb subunits, CycT1 and CDK9. Consistent with previous results [34
], both Gal4-CycT1 and Gal4-CDK9 activated transcription. The ChIP analysis indicates that both Gal4-CycT1 and Gal4-CDK9 stimulated assembly of a TC that contained all of the GTFs but lacked TAF1 and TAF5. These results confirm that P-TEFb is responsible for directing selective recruitment of TBP in the absence of TAFs.
In the experiments shown in , P-TEFb was tethered to DNA, whereas on the HIV-1 LTR, P-TEFb is normally bound through Tat to nascent RNA. We therefore sought to verify that P-TEFb, when bound to nascent RNA, would also selectively recruit TBP in the absence of TAFs. Previous studies have shown that recruitment of CycT1/P-TEFb to the HIV-1 LTR through a heterologous RNA-binding domain can activate transcription in the absence of Tat [36
]. We therefore asked whether such an artificially recruited CycT1/P-TEFb protein would stimulate assembly of a TC that contained TBP but not TAFs. We used a previously characterized HIV-1 LTR derivative [36
] in which the TAR element was replaced by a minimal binding site for the HIV-1 Rev protein, and examined the ability of a Rev-Tat or Rev-CycT1 fusion protein to stimulate TC assembly.
shows that in the absence of a Rev fusion protein, both transcription and GTF recruitment were virtually undetectable. Addition of Rev-Tat or Rev-CycT1 resulted in a large transcriptional increase, as expected from previous studies. Significantly, addition of Rev-Tat or Rev-CycT1 also led to a large increase in recruitment of TBP, TFIIB, mediator, and RNA polymerase II, but not TAF1 or TAF5. Thus, in the absence of Tat, Rev-CycT1 stimulated assembly of a TC that contained TBP but not TAFs.
Tat and P-TEFb Direct Recruitment of TBP in the Absence of TAFs When Tethered to RNA
Identification of Cellular Promoters That Recruit TBP in the Absence of TAFs
P-TEFb is also a cofactor for several cellular activators, the best studied of which is CIITA, a transcription factor involved in the expression of major histocompatibility complex (MHC) class II genes (reviewed in [37
]). In fact, overexpression of Tat can inhibit transcription of MHC class II genes by competing with CIITA for P-TEFb binding [38
We therefore tested whether the promoters of MHC class II genes also recruited an atypical TC that contained TBP but not TAFs. We analyzed two MHC class II genes known to use CIITA, HLA-DM and HLA-DR, and, as a negative control, GAPDH. RT-PCR analysis confirmed, as expected, that both HLA-DM and HLA-DR were expressed in 293T cells (data not shown). The associated ChIP experiment () shows that the TCs formed on both HLA-DM and HLA-DR contained all of the GTFs but lacked significant levels of TAF1 and TAF5. By contrast, the TC formed on GAPDH contained all of the GTFs, including TAF1 and TAF5. These results indicate that TBP, and not TFIID, is also selectively recruited to specific cellular promoters and strongly support the conclusion that P-TEFb directs this recruitment.
The Cellular Activator CIITA Directs Recruitment of TBP in the Absence of TAFs