Post-translational histone modifications have essential roles in controlling nuclear processes; however, the specific mechanisms regulating these modifications and their combinatorial activities remain elusive. Cyclin-dependent kinase 9 (CDK9) regulates gene expression by phosphorylating transcriptional regulatory proteins, including the RNA polymerase II carboxy-terminal domain. Here, we show that CDK9 activity is essential for maintaining global and gene-associated levels of histone H2B monoubiquitination (H2Bub1). Furthermore, CDK9 activity and H2Bub1 help to maintain correct replication-dependent histone messenger RNA (mRNA) 3 0 -end processing. CDK9 knockdown consistently resulted in inefficient recognition of the correct mRNA 3 0 -end cleavage site and led to increased read-through of RNA polymerase II to an alternative downstream polyadenylation signal. Thus, CDK9 acts to integrate phosphorylation during transcription with chromatin modifications to control co-transcriptional histone mRNA processing.

Extensive changes in posttranslational histone modifications accompany the rewiring of the transcriptional program during stem cell differentiation. However, the mechanisms controlling the changes in specific chromatin modifications and their function during differentiation remain only poorly understood. We show that histone H2B monoubiquitination (H2Bub1) significantly increases during differentiation of human mesenchymal stem cells (hMSCs) and various lineage-committed precursor cells and in diverse organisms. Furthermore, the H2B ubiquitin ligase RNF40 is required for the induction of differentiation markers and transcriptional reprogramming of hMSCs. This function is dependent upon CDK9 and the WAC adaptor protein, which are required for H2B monoubiquitination. Finally, we show that RNF40 is required for the resolution of the H3K4me3/H3K27me3 bivalent poised state on lineage-specific genes during the transition from an inactive to an active chromatin conformation. Thus, these data indicate that H2Bub1 is required for maintaining multipotency of hMSCs and plays a central role in controlling stem cell differentiation.

Many anticancer therapies function largely by inducing DNA double-strand breaks (DSBs) or altering the ability of cancer cells to repair them. Proper and timely DNA repair requires dynamic changes in chromatin assembly and disassembly characterized by histone H3 lysine 56 acetylation (H3K56ac) and phosphorylation of the variant histone H2AX (γH2AX). Similarly, histone H2B monoubiquitination (H2Bub1) functions in DNA repair, but its role in controlling dynamic changes in chromatin structure following DSBs and the histone chaperone complexes involved remain unknown. Therefore, we investigated the role of the H2B ubiquitin ligase RNF40 in the DSB response. We show that RNF40 depletion results in sustained H2AX phosphorylation and a decrease in rapid cell cycle checkpoint activation. Furthermore, RNF40 knockdown resulted in decreased H3K56ac and decreased recruitment of the facilitates chromatin transcription (FACT) complex to chromatin following DSB. Knockdown of the FACT component suppressor of Ty homolog-16 (SUPT16H) phenocopied the effects of RNF40 knockdown on both γH2AX and H3K56ac following DSB induction. Consistently, both RNF40 and SUPT16H were required for proper DNA end resection and timely DNA repair, suggesting that H2Bub1 and FACT cooperate to increase chromatin dynamics, which facilitates proper checkpoint activation and timely DNA repair. These results provide important mechanistic insights into the tumor suppressor function of H2Bub1 and provide a rational basis for pursuing H2Bub1-based therapies in conjunction with traditional chemo- and radiotherapy.

C yclin-dependent kinase-9 (CDK9) was originally characterized as a transcription elongation factor which regulates RNA Polymerase II (RNAPII) activity following transcriptional initiation. However, recent evidence from a number of studies have shown that CDK9 plays an important role in regulating not only RNAPII activity but also cotranscriptional histone modification and mRNA processing events such as splicing and 3' end processing. Importantly, our previous work and the work presented here demonstrate that CDK9 functions to guide a complex network of chromatin modifications including histone H2B monoubiquitination (H2Bub1), H3 lysine 4 trimethylation (H3K4me3) and H3K36me3. This function appears to be dependent upon not only the phosphorylation of the RNA Polymerase II C-terminal domain but also upon other CDK9 targets such as the Suppressor of Ty Homolog-5 (SUPT5H), Negative Elongation Factor-E (NELF-E) and probably the human Rad6 homolog UBE2A. We provide a working model by which CDK9 may control co-transcriptional replication-dependent histone mRNA 3' end processing in an H2Bub1 and H3K4me3-dependent manner and uncover new and important differences between the functions of human CDK9 and its yeast counterparts Ctk1 and Bur1.

Cyclin-dependent kinase-9 (CDK9) plays a central role in transcriptional elongation and controls multiple cotranscriptional histone modifications, including histone H2B monoubiquitination (H2Bub1). Like other CDK9-dependent histone modifications, the role of CDK9 in maintaining H2Bub1 was shown to be partially dependent upon the phosphorylation status of Ser2 of the RNA polymerase II (RNAPII) C-terminal domain (CTD). Since mutation of Ser2 within the RNAPII CTD resulted in a milder effect on H2Bub1 compared with CDK9 knockdown, we explored whether another CDK9 target may also influence H2Bub1. Based on its homology to yeast Bur1, we hypothesized that CDK9 may directly phosphorylate and activate the ubiquitin-conjugating enzyme utilized for H2B monoubiquitination. Indeed, we demonstrate that UBE2A specifically interacts with CDK9, but not CDK2. Furthermore, UBE2A is phosphorylated by CDK9 in vitro and increases UBE2A activity. Interestingly, CDK9 knockdown not only decreases UBE2A phosphorylation and H2Bub1, but also significantly impairs the induction of UBE2A-dependent monoubiquitination of proliferating cell nuclear antigen (PCNA). Thus, we provide the first evidence that CDK9 is required for the activity of UBE2A in humans, and that its activity is not only required for maintaining H2Bub1, but also for the monoubiquitination of PCNA. The common involvement of these two ubiquitinations in distinct DNA repair pathways may provide a mechanistic rationale for further exploring CDK9 as a combinatorial target for increasing the efficacy of existing cancer therapies based on the induction of DNA damage and are repaired by mechanisms which require H2Bub1 and/or PCNA ubiquitination.