Controlling cellular P-TEFb activity by the HIV-1 transcriptional transactivator Tat

Abstract

The human immunodeficiency virus 1 (HIV-1) transcriptional transactivator (Tat) is essential for synthesis of full-length transcripts from the integrated viral genome by RNA polymerase II (Pol II). Tat recruits the host positive transcription elongation factor b (P-TEFb) to the HIV-1 promoter through binding to the transactivator RNA (TAR) at the 5'-end of the nascent HIV transcript. P-TEFb is a general Pol II transcription factor; its cellular activity is controlled by the 7SK small nuclear RNA (snRNA) and the HEXIM1 protein, which sequester P-TEFb into transcriptionally inactive 7SK/HEXIM/P-TEFb snRNP. Besides targeting P-TEFb to HIV transcription, Tat also increases the nuclear level of active P-TEFb through promoting its dissociation from the 7SK/HEXIM/P-TEFb RNP by an unclear mechanism. In this study, by using in vitro and in vivo RNA-protein binding assays, we demonstrate that HIV-1 Tat binds with high specificity and efficiency to an evolutionarily highly conserved stem-bulge-stem motif of the 5'-hairpin of human 7SK snRNA. The newly discovered Tat-binding motif of 7SK is structurally and functionally indistinguishable from the extensively characterized Tat-binding site of HIV TAR and importantly, it is imbedded in the HEXIM-binding elements of 7SK snRNA. We show that Tat efficiently replaces HEXIM1 on the 7SK snRNA in vivo and therefore, it promotes the disassembly of the 7SK/HEXIM/P-TEFb negative transcriptional regulatory snRNP to augment the nuclear level of active P-TEFb. This is the first demonstration that HIV-1 specifically targets an important cellular regulatory RNA, most probably to promote viral transcription and replication. Demonstration that the human 7SK snRNA carries a TAR RNA-like Tat-binding element that is essential for the normal transcriptional regulatory function of 7SK questions the viability of HIV therapeutic approaches based on small drugs blocking the Tat-binding site of HIV TAR.

Figure 1. In vivo association of HIV Tat with 7SK.

A. Transient expression of Tat-FL in HeLa cells. Schematic structure of the pTat-FL expression construct is shown. The cytomegalovirus promoter (CMV) and the polyadenylation region (PA) are indicated. Tat-FL was immunoprecipitated (α-Flag) from extracts (E) prepared from transfected or non-transfected (NT) cells. Distribution of Tat-FL and Paf1 was monitored by Western blot analysis. B. Detection of Tat-associated HeLa RNAs. RNAs co-precipitated with Tat-FL (T) were labeled in vitro and separated on a 6% sequencing gel. Lanes NT and M, control IP from non-transfected cells and molecular size markers. C. RNA G-tracking. The Tat-associated RNA was partially digested with RNase T1 or moderately hydrolyzed with formamide (OH⁻) and analyzed on a 6% gel. G residues and their positions in the human 7SK snRNA sequence are indicated. Asterisk indicates a fragile U residue. D. In vivo cross-linking of Tat-FL and 7SK. HeLa cells expressing Tat-FL were treated (x link) or not treated (cont) with formaldehyde before extract (E) preparation. Tat-FL was immunoprecipitated (α-Flag) or mock-precipitated (no Ab) under stringent conditions. Distributions of Tat-FL and 7SK snRNA were monitored by Western blot analysis and RNase A/T1 mapping, respectively.

Figure 2. Identification of putative Tat-binding motifs in 7SK snRNA.

A. Tat binds to the 5′ hairpin of 7SK. Schematic structures of the p7SK expression construct and the expressed truncated 7SK RNAs are shown. Dashed boxes indicate deletions. Pol III transcription of the 7SK gene terminates within four consecutive T residues. HeLa cells were transfected with the indicated expression plasmids. After extract (E) preparation, Tat-FL was immunoprecipitated (IP). Distribution of the endogenous and transiently expressed 7SK RNAs and Tat-FL was monitored by Western and Northern blotting. B. Consensus structure of the minimal Tat-binding motif of HIV TAR. Nucleotides with essential and moderate contribution to Tat binding are in red and green, respectively. C. Phylogenetic comparison of the 5′ hairpins of 7SK snRNAs. The sequences of lancelet, snail, fruit fly and tick 7SK snRNAs have been published , . The potential Tat-binding motifs of human 7SK snRNA are shaded. The evolutionarily invariant nucleotides are in red. Nucleotides common to the putative proximal Tat-binding motif of 7SK and the Tat-binding site of HIV TAR are in orange.

Figure 3. Characterization of the Tat-binding element of human 7SK snRNA.

A. Tat binds to the distal part of the 5′ hairpin of 7SK. About 2 fmol of ³²P-labeled RNA representing the distal (Dist) or proximal (Prox) part of the 7SK 5′ hairpin was incubated with the indicated amount (fmol) of Tat(38–72) oligopeptide and analyzed on a 4% native gel. B. The distal Tat-binding motif of 7SK directs in vivo binding of Tat. The 5′ hairpin of 7SK (5′HP) carrying the pm and/or dm mutations was co-expressed with Tat-FL and their interaction was tested by co-IP and Northern blotting. Structure of the p5′HP expression plasmid with the pm and dm mutations and the expected length of the 5′HP RNA is shown. C. In vivo association of Tat with mutant 7SK 5′ hairpin RNAs. Nucleotides indicated by red (essential) or green (dispensable) lines were replaced with complementary nucleotides in the p5′HP expression plasmid. Tat-FL and the mutant 5′HP RNAs were co-expressed in HeLa cells and their interactions were tested.

Figure 4. The 5′ hairpin of human 7SK snRNA contains two HEXIM-binding sites.

A. Detection of HEXIM-binding sites by mobility shift assays. About 2 fmol of in vitro synthesized probe RNAs representing the entire (5′HP) or the distal (Dist) and proximal (Prox) parts of the 5′ hairpin of human 7SK were incubated with increasing amounts (fmol) of recombinant HEXIM1 and analyzed on a 5% native gel. Appropriate cold RNAs were used as specific competitors. B. In vitro interaction of mutant 7SK 5′ hairpin RNAs (5′HPdm, 5′HPpm, 5′HPdm+pm) with HEXIM1. Complexes were analyzed on a 4% gel. C. The minimal HEXIM1-binding elements of 7SK. The in vitro HEXIM-binding capacity of a distal (DHBS) and proximal (PHBS) fragment of the 7SK 5′ hairpin was tested by mobility shift assay on a 4% gel. Sequences derived from wild-type 7SK are boxed. D. Gelshift analysis of an artificial hairpin RNA (5′HPsyn) carrying the distal and proximal HEXIM-binding motifs of 7SK. Sequences originated from the human 7SK snRNA are boxed.

Figure 5. In vivo binding of HEXIM1 to 7SK RNA.

A. Interaction of HEXIM1 with mutant 7SK 5′ hairpin RNAs. HA-HEXIM1 was immunoprecipitated (IP) from extracts (E) prepared from HeLa cells also expressing 5′HP, 5′HPpm or 5′HPdm RNAs. Interaction of HA-HEXIM1 with the endogenous 7SK and the ectopically expressed 5′HP, 5′HPpm, 5′HPdm RNAs was monitored by Northern blot analyses. B. Interaction of HEXIM1 and P-TEFb with mutant 7SK snRNAs. 7SKpm and 7SKdm RNAs were expressed in HeLa cells together with HA-HEXIM1 (lanes 1–10) or in G3H cells accumulating HA-CycT1 (lanes 11–18). After IP, recovery of HA-HEXIM1 and HA-CycT1 was confirmed by Western blot analysis and co-precipitation of the endogenous and ectopically expressed 7SK RNAs was determined by RNase A/T1 mapping. Lane C, control mapping with E. coli tRNA. Lane B, control IP with beads alone.

Figure 6. In vivo disruption of the 7SK/HEXIM/P-TEFb snRNP by HIV Tat.

A. Tat disrupts the interaction of 7SK and HEXIM1. About 5×10⁶ HeLa cells were transfected with 0.5, 1.5, 2.5 or 3.5 µg of pTat-FL. After 48h of incubation, cell extracts were prepared, HEXIM1 was immunoprecipitated and association of 7SK snRNA was measured by Northern blotting. NT, control IP from non-transfected cells. B. The TAR RNA-binding capacity of Tat is essential for 7SK binding and for disruption of the 7SK/HEXIM/P-TEFb snRNP. Transiently expressed wild-type and mutant (K50Q and K50A+K51A) Tat-FL proteins as well as endogenous HEXIM1 were immunoprecipitated and co-purification of endogenous 7SK snRNA and CycT1 was monitored. C. Transiently expressed mutant C22G and K41A Tat proteins lacking CycT1-binding ability can disrupt the in vivo interaction of HEXIM1 and 7SK. (For other details, see the captures to panel A) D. Tat disrupts 7SK/HEXIM/P-TEFb independently of its CycT1-binding capacity. Wild-type and mutant (C22G and K41A) Tat-FL proteins were expressed in HeLa G3H cells stably expressing HA-CycT1. Association of HA-CycT1 with 7SK, HEXIM1 and Tat-FL proteins was monitored by co-IP. E. Tat and HEXIM1 bind to 7SK in a mutually exclusive manner. From extracts (E) prepared from HeLa cells non-transfected (NT) or transfected with the pTat-FL, pTat-FL(C22G) or pTat-FL(K41A) expression plasmids the accumulating Tat-FL proteins were immunoprecipitated (IP). Distribution of 7SK snRNA and Tat-FL, HEXIM1 and CycT1 proteins was monitored with RNase A/T1 mapping and Western blot analysis. F. Expression of Tat-FL, Tat-FL(C22G) and Tat-FL(K41A) increases the cellular level of active P-TEFb. From extracts of G3H cells expressing Tat-FL proteins, the HA-tagged P-TEFb was immobilized on beads saturated with anti-HA antibody and incubated with a recombinant GST-CTD protein and [γ-³²P]ATP. Distribution of HA-CycT1 and Tat-FL and phosphorylation of GST-CTD at serine 2 were monitored by Western blot analysis. CTD phosphorylation was quantified by PhosphorImager.

Figure 7. The 7SK/Tat snRNP does not interact with hnRNP proteins.

A. Expression of Tat has no effect on the interaction of 7SK and hnRNP proteins. HnRNP A1 and A2/B1 were immunoprecipitated from extracts prepared from HeLa cells expressing or not expressing (NT) Tat-FL. Co-purification of 7SK was assayed by Northern blotting. Lanes no Ab, control IPs without antibody. B. The 7SK/Tat snRNP lacks hnRNP proteins, but associates with Larp7. Transiently expressed Tat-FL was immunoprecipitated from extracts (E) prepared from transfected (T) or non-transfected (NT) HeLa cells. Co-IP of 7SK snRNA and hnRNP and Larp7 proteins was assayed.