DDX6 transfers P-TEFb kinase to the AF4/AF4N (AFF1) super elongation complex

Abstract

AF4/AFF1 and AF5/AFF4 are both backbones for the assembly of "super elongation complexes" (SECs) that exert 2 distinct functions after the recruitment of P-TEFb from the 7SK snRNP: (1) initiation and elongation of RNA polymerase II gene transcription, and (2) modification of transcribed gene regions by distinct histone methylation patterns. In this study we aimed to investigate one of the initial steps, namely how P-TEFb is transferred from 7SK snRNPs to the SECs. In particular, we were interested in the role of DDX6 that we have recently identified as part of the AF4 complex. DDX6 is an evolutionarily conserved member of the DEAD-box RNA helicase family that is known to control miRNA and mRNA biology (translation, storage and degradation). Overexpressed DDX6 is associated with different cancer types and with c-Myc protein overexpression. We could demonstrate that DDX6 binds to 7SK snRNA and causes the release and transfer of P-TEFb to the AF4/AF4N SEC. DDX6 also binds stably to AF4 and AF4N as demonstrated by GST pull-down and co-immunoprecipitation experiments. As a consequence, overexpression of either AF4/AF4N or DDX6 resulted in a strong increase of mRNA production (5-6 fold), while their simultaneous expression increased the cellular mRNA production by 11-fold. Conversely, the corresponding knockdown of DDX6 decreased mRNA production by 70%. In conclusion, AF4/AF4N and DDX6 represent key molecules for the elongation process of gene transcription and a model will be proposed for the hand-over process of P-TEFb to SECs.

Keywords: 7SK snRNP; AF4/AFF1; AF4N; DDX6; P-TEFb; RNA polymerase II; SEC; elongation control.

Figure 1

Cell lines expressing AF4 and AF4N form SEC complexes. A. Induction kinetics of the AF4 and AF4N protein. Both proteins are strongly induced 24-72 h after Dox treatment. B. Affinity-purified (AP) AF4 or AF4N assemble into identical super elongation complexes (SECs). CCNT1: Cyclin T1; L: lysate; FT: flow through; E: eluate. Lower panel: AF4N purifications under different stringency conditions. C. Immunoprecipitation (IP) experiments reveal that DDX6 is binding to either AF4 and AF4N proteins or complexes thereof (SECs). L: lysate; W: washing fraction; P: precipitate. D. GST Pull-down experiment to confirm a direct protein-protein interaction between AF4N and DDX6. The interaction between both recombinant proteins was not dependent on the addition of 7SK snRNA. E. Dox-induced overexpression (~3-fold) or shRNA-mediated knockdown of DDX6 (~80%) is shown.

Figure 2

Quantification of mRNA levels in DDX6-/AF4N-overexpressing and DDX6 knockdown cells. A. Identical cell numbers (2 x 10⁵) of stable transfected HEK 293T cells were used to isolate total RNA with a standardized method. Total amount of RNA is displayed. B. Amount of mRNA in the same cells after subtraction of rRNA and RNAPIII transcripts. C. RNA gel with 28S and 18S rRNA isolated from the different cell lines used in this study; below: real-time PCR data of 18S rRNA. Relative quantification was carried out by using two housekeeping genes (GAPDH and RPL13A). Both displayed no changes in Ct values during qRT-PCR. These experiments revealed that the quantitative amount of ribosomal RNA was not significantly affected. Small differences 1-4% are in the range of experimental variations. D. qRT-PCR data of HEXIM1 and Cyclin T1 mRNA.

Figure 3

Association and binding of 7SK snRNA to components of the 7SK snRNP and DDX6/AF4N. A. In vitro transcript of 7SK snRNA (IVT 7SK snRNA). B. Immunoprecipitations performed with antibodies against DDX6, HEXIM1, LARP7, AF4N and control IgG. The amount of bound 7SK snRNA was visualized by RT-PCR of the available 7SK snRNA in these precipitates. Adding 25 ng of IVT 7SK snRNA to the precipitates revealed that DDX6 and AF4N are both able to bind to this particular snRNA. C. In vitro 7SK snRNA binding experiment. Recombinant proteins expressed in E. coli were used to demonstrate a direct binding capacity of LARP7, HEXIM1 (both positive controls), AF4N and DDX6 towards the 7SK snRNA. Neither GST protein alone nor GST-beads were able to bind to the 7SK snRNA under these conditions. PC and NC: ± reverse transcribed 7SK snRNA to validate the PCR assay. D. Alignment of the SIAH1/2 interacting motif of AF4 protein family members.

Figure 4

DDX6 is a recruiting factor for P-TEFb. A. Affinity purified AF4N SECs are asscoiated with endogenous 7SK snRNA, while the addition of RNase A destroy the 7SK snRNA. B. Addition of RNase A destroys the binding of DDX6 to the AF4 SEC, but not binding of P-TEFb. Adding 25 ng 7SK snRNA restores DDX6 binding to the AF4N SEC. L: lysate; E: eluate. C. Overexpressed or repressed DDX6 changed the amount of P-TEFb transferred to the AF4N SEC. L: lysate; E: eluate. D. RNase A treatment destroys completely the 7SK snRNP.

Figure 5

Proposed model for DDX6 under physiological conditions. P-TEFb is stored as inactive kinase in the 7SK snRNP. Currently known releasing factors are the HIV Tat protein, BRD4 and certain MLL fusion proteins. DDX6 is a new P-TEFb releasing factor. RNase A treatment destroys 7SK snRNPs and generates a large pool of freely available P-TEFb. The amount of DDX6, as well as the amount of AF4/AF4N is critical for the assembly of P-TEFb containing AF4/AF4N SECs that drive transcript initiation and elongation. P-TEFb kinase activity destroys the AF4/AF4N SECs by enhancing the turn-over of AF4/AF4N. This allows free P-TEFb to be recruited via HEXIM1 back to 7SK snRNPs.