Dissecting the Pol II transcription cycle and derailing cancer with CDK inhibitors

Abstract

Largely non-overlapping sets of cyclin-dependent kinases (CDKs) regulate cell division and RNA polymerase II (Pol II)-dependent transcription. Here we review the molecular mechanisms by which specific CDKs are thought to act at discrete steps in the transcription cycle and describe the recent emergence of transcriptional CDKs as promising drug targets in cancer. We emphasize recent advances in understanding the transcriptional CDK network that were facilitated by development and deployment of small-molecule inhibitors with increased selectivity for individual CDKs. Unexpectedly, several of these compounds have also shown selectivity in killing cancer cells, despite the seemingly universal involvement of their target CDKs during transcription in all cells. Finally, we describe remaining and emerging challenges in defining functions of individual CDKs in transcription and co-transcriptional processes and in leveraging CDK inhibition for therapeutic purposes.

Fig. 1 |. Multiple CDK-cyclin complexes act in sequence in the Pol II transcription cycle.

During progression through transcription, multiple CDK-cyclin complexes act sequentially, approximately in the intervals depicted by the curved arrows, to promote transitions between and progression through the different stages of transcription: initiation, elongation, termination (which comprises primary transcript cleavage and polyadenylation followed by termination, per se) and a hypothesized recycling of unphosphorylated Pol II to re-enter the cycle.

Fig. 2 |. Modes of Cdk8-mediated transcription regulation.

a, The Cdk8 module and Pol II bind in a, mutually exclusive fashion to the core Mediator, while CTD phosphorylation (P) blocks Pol II recruitment to promoters. b, Direct phosphorylation of transcription factors regulates their activity (both positively and negatively). c, Cdk8 stabilization of the scaffold complex favors Pol II recruitment and reinitiation. d, Cdk8 facilitates recruitment of P-TEFb and Brd4 to promote elongation at some genes. TF, transcription factor; GTF, general transcription factor; SEC, super elongation complex.

Fig. 3 |. CDK-dependent transitions in the Pol II transcription cycle.

After initiation, Cdk7 phosphorylates the Pol II CTD at Ser5 and Ser7 positions and promotes the exchange of an initiation factor, TFIIE, for the elongation factor Spt5 (subunit of DSIF), which recruits NELF to establish a promoter-proximal pause ~50–100 bp downstream of the TSS. This coincides with 5′-end capping of the nascent transcript (indicated by a purple dot). Pausing is relieved by Cdk9, which is activated by Cdk7 and phosphorylates multiple components of the elongation complex, including the CTRs of Spt5. As Pol II traverses the CPS, Spt5 undergoes protein phosphatase 1 (PP1)-dependent dephosphorylation, the elongation complex pauses, and Cdk12 and/or Cdk13 phosphorylate Pol II CTD Ser2 to promote cleavage, polyadenylation of the nascent pre-mRNA and Xrn2-dependent termination. Green represents active CDK; grey represents inactive CDK. Dashed straight arrow indicates a requirement for Cdk7 activity in pause establishment for which the direct targets are unknown Dashed curved arrow indicates eviction of NELF from the elongation complex.

Fig. 4 |. Strategies for selective CDK inhibition by chemical genetics.

a, Mutation of the conserved gatekeeper to a less bulky amino acid residue (Phe to Gly) renders a kinase analog sensitive (AS). b, Structures of bulky adenine analogs that inhibit AS kinases, but not wild-type kinases. c, An additional Cys substitution within the active site renders an AS kinase irreversibly sensitized (IS) to bulky adenines conjugated to CMK.

Fig. 5 |. Strategies for selective CDK inhibition by covalent inhibitors.

a, Covalent kinase inhibitors have dual selectivity for the active site and a naturally occurring Cys residue situated in close proximity to the ATP-binding pocket of some kinases. Different colors are used to distinguish different chemical moieties within the inhibitor or different regions within the target kinase. Curved arrows represent the flow of electrons in formation of the covalent kinase-inhibitor adduct. b, Structure of THZ1, which targets Cdk7, Cdk12 and Cdk13. c, Structure of THZ531, which targets Cdk12 and Cdk13.

Fig. 6 |. Strategies for selective CDK inhibition by degraders.

a, Conjugation of an E3-binding moiety such as thalidomide (arrowhead) to a moderately selective kinase inhibitor (ball) enables selective ubiquitylation (Ub) by an E3 ubiquitin ligase and kinase destruction by the proteasome. b, Structure of THAL-SNS-032 (thalidomide moiety is boxed), which targets Cdk9 for CRBN-dependent, proteasome-mediated degradation.