The ability of positive transcription elongation factor B to transactivate human immunodeficiency virus transcription depends on a functional kinase domain, cyclin T1, and Tat

Abstract

By binding to the transactivation response element (TAR) RNA, the transcriptional transactivator (Tat) from the human immunodeficiency virus increases rates of elongation rather than initiation of viral transcription. Two cyclin-dependent serine/threonine kinases, CDK7 and CDK9, which phosphorylate the C-terminal domain of RNA polymerase II, have been implicated in Tat transactivation in vivo and in vitro. In this report, we demonstrate that CDK9, which is the kinase component of the positive transcription elongation factor b (P-TEFb) complex, can activate viral transcription when tethered to the heterologous Rev response element RNA via the regulator of expression of virion proteins (Rev). The kinase activity of CDK9 and cyclin T1 is essential for these effects. Moreover, P-TEFb binds to TAR only in the presence of Tat. We conclude that Tat-P-TEFb complexes bind to TAR, where CDK9 modifies RNA polymerase II for the efficient copying of the viral genome.

FIG. 1

Diagrammatic representation of plasmid effectors and reporters used in this study. (A) Plasmid effectors consisted of pRev, pRevTat, pRevCDK9, and pRevCDK9(D167N), which directed the synthesis of Rev, hybrid RevTat, hybrid RevCDK9, and hybrid Rev-kinase-negative-CDK9 proteins from the simian virus 40 early promoter, respectively. (B) Plasmid reporters consisted of pHIVSCAT, and pRRESCAT, which contained the wild-type HIV-1 LTR from HIV-1_SF2 and substituted HIV-1 LTRs, where TAR was replaced by SLIIB from the RRE linked to the CAT reporter gene, respectively. Ovals before arrows denote types of promoters, followed by start sites of transcription (lines with arrows), TAR and SLIIB structures, and reporters (CAT), followed by the polyadenylation site (pA). D167N stands for a mutation of aspartic acid to asparagine at position 167 of CDK9, which destroys the kinase activity of CDK9.

FIG. 2

RNA tethering is required for the activity of CDK9 on the HIV-1 LTR. HeLa cells were cotransfected with pHIVSCAT (A) or pRRESCAT (B) and indicated plasmid effectors. (A) Only the wild-type Tat and hybrid RevTat proteins transactivated pHIVSCAT, which contains TAR (99- and 86-fold, respectively [black bars]). Neither the hybrid RevCDK9 nor RevCDK9(D167N) protein transactivated the wild-type HIV-1 LTR (striped bars). (B) Both the hybrid RevTat and RevCDK9 but not Tat and RevCDK9(D167N) proteins increased transcription from pRRESCAT, which contains the binding site for Rev (SLIIB) in place of TAR. The hybrid RevCDK9 protein had about one-half the activity of the hybrid RevTat protein (46-fold versus 26-fold transactivation [black versus striped bars]). A total of 0.1 μg of each plasmid target and effector was used. The table below the bar graphs contains the mean slopes from liquid CAT data and absolute values for the fold transactivation. Experiments are representative of three independent transfections performed in duplicate where the standard errors of the means are given.

FIG. 3

Squelching assays revealed functional interactions between CDK9 and cyclin T1 in cells. The expression of free CDK9 reduced effects of the hybrid RevCDK9 protein on SLIIB (15- to 3-fold transactivation [lanes 2 and 3, white and cross-hatched bars]). However, the coexpression of increasing amounts of cyclin T1 together with CDK9 and the chimera restored levels of transactivation from 8- to 15-fold (lanes 4 to 6, cross-hatched bars). The coexpression of Tat had limited effects on the hybrid RevCDK9 protein (16-fold to 10-fold transactivation [lanes 7 and 8, black bars]). A total of 0.1 μg of the plasmid effector was used. Amounts of plasmid effectors are given below the bar graph. The introduction of various proteins had minimal effects on the expression of LacZ from the cotransfected pCMVβGal, which served as our internal control plasmid (table; expressed as absorbance units). Experiments are representative of two independent transfections performed in duplicate where the standard errors of the means are given. OD450, optical density at 450 nm.

FIG. 4

CDK9 as part of P-TEFb binds to TAR RNA in the presence of Tat. (A) Different TAR sequences were used in our RNA binding studies. The wild-type TAR contained the entire TAR (60 nucleotides) and an additional 20 nucleotides to the HindIII site at its 3′ end. Δbulge lacked the 5′ bulge (circled boldface letters). Δloop lacked 4 nucleotides of the central loop (boldface letters). Only relevant nucleotides from the upper stem are presented; the rest of the loop is schematized by the railroad track. (B) All three TAR sequences were labeled equivalently, and Δloop migrated faster on these 10% polyacrylamide gels. Immunoprecipitated P-TEFb complex (from HeLa cells expressing HA-tagged CDK9) bound to the wild-type but not mutant TAR sequences only in the presence of Tat (compare lanes 1 and 2 to lanes 3, 4, 5, and 6). In the absence of HA-tagged CDK9 (wild-type HeLa cells), no binding to TAR was observed even in the presence of Tat (lanes 5 and 6). wt, wild type; nt, nucleotide(s).

FIG. 5

A model for interactions among Tat, P-TEFb, and TAR. CDK9 exists in the P-TEFb complex in cells. This complex minimally contains CDK9 and cyclin T1. Tat binds to P-TEFb in the absence of TAR. This complex has also been referred to as the Tat-TAK (Tat-associated kinase) complex. The assembly of the Tat–P-TEFb complex results in the formation of a high-affinity surface for the binding to TAR RNA. Both the 5′ bulge and central loop are required for these protein-RNA interactions. Upon binding to TAR, CDK9 can then phosphorylate the CTD of RNAPIIo and possibly other targets, which results in efficient elongation of viral transcription. Importantly, the kinase activity of CDK9 and the presence of cyclin T1 are required for these effects.