Development of a selective CDK7 covalent inhibitor reveals predominant cell cycle phenotype

Summary

Cyclin-dependent kinase 7 (CDK7) regulates both cell cycle and transcription, but its precise role remains elusive. We previously described THZ1, a CDK7 inhibitor, which dramatically inhibits super-enhancer associated gene expression. However, potent CDK12/13 off-target activity obscured CDK7s contribution to this phenotype. Here, we describe the discovery of a highly selective, covalent CDK7 inhibitor. YKL-5-124 causes arrest at the G1/S transition and inhibition of E2F-driven gene expression; these effects are rescued by a CDK7 mutant unable to covalently engage YKL-5-124, demonstrating on-target specificity. Unlike THZ1, treatment with YKL-5-124 resulted in no change to Pol II CTD phosphorylation; however, inhibition could be reconstituted by combining YKL-5-124 and THZ531, a selective CDK12/13 inhibitor, revealing potential redundancies in CDK control of gene transcription. These findings highlight the importance of CDK7/12/13 polypharmacology for anti-cancer activity of THZ1 and posit that selective inhibition of CDK7 may be useful for treatment of cancers marked by E2F misregulation.

Graphical Abstract

eTOC blurb

Olson et al describe the development and characterization of YKL-5-124, a potent, selective and covalent CDK7 inhibitor. YKL-5-124 displays biochemical and cellular selectivity for CDK7 over CDK12/13, structurally related kinases. CDK7 inhibition by YKL-5-124 induces a strong cell cycle arrest and a surprisingly weak effect on RNA Pol II phosphorylation.

Introduction

Cyclin-dependent kinase 7 (CDK7) is a master regulator of cell cycle progression and gene transcription (Fisher and Morgan, 1994, Devault et al., 1995, Glover-Cutter et al., 2009, Kelso et al., 2014). In the cytoplasm, CDK7 comprises the enzymatic core of the CDK-activating kinase (CAK), a trimeric complex that also includes CDK7’s obligate binding partners, cyclin H and MAT1. CAK phosphorylates key cell cycle CDKs (1, 2, and 4) leading to their full activation (Larochelle et al., 2007, Schachter et al., 2013, Bisteau et al., 2013, Gerber et al., 2008, Larochelle et al., 2012). Activation of CDKs by CAK plays an essential role in control over cell cycle progression: (1) activation of CDK4 results in the phosphorylation of retinoblastoma (Rb) protein and consequent G1 progression (Sherr and Roberts, 1999), (2) CDK2 activation promotes progression from G1 into S phase (Furuno et al., 1999, Satyanarayana and Kaldis, 2009), and (3) phosphorylation of CDK1 positively regulates the G2/M transition. These results have been demonstrated both genetically and through the use of the ‘bumped hole’ chemical genetic approach which has enabled selective inhibition of a genetically engineered CDK7 mutant that is sensitive to a bulky ATP-like inhibitor (Larochelle et al., 2007). The predominant cell cycle phenotype caused by inhibition of CDK7 (and the kinases that depend on CDK7 for activity) is accumulation of G1 or G2/M cells and a concomitant reduction in the number of S phase cells (Larochelle et al., 2007, Ali et al., 2009, Wang et al., 2016, Kwiatkowski et al., 2014), showing that cell proliferation requires CDK7 kinase activity.

In the nucleus, the CAK module is a component of the general transcription factor, TFIIH, a multi-protein complex that is essential for RNA Polymerase II (Pol II)-mediated transcription (Adamczewski et al., 1996, Glover-Cutter et al., 2009, Serizawa et al., 1995, Shiekhattar et al., 1995, Chen et al., 2015, Feaver et al., 1994). Phosphorylation of the Pol II CTD at S2/S5/S7 residues is tightly regulated as a means to coordinate transcription initiation, elongation and termination (Buratowski, 2009, Egloff and Murphy, 2008). However, the conclusion that CDK7 kinase activity is required for Pol II CTD phosphorylation and thus global gene transcription has been controversial. Several reports find that CDK7 is non-essential in transcription both in vivo and in vitro (Kanin et al., 2007, Glover-Cutter et al., 2009, Serizawa et al., 1993). In contrast, other reports show that CDK7 inhibition reduces serine (S5 and S7) phosphorylation across the 52 heptad repeats comprising the Pol II C-terminal domain (CTD) (Kwiatkowski et al., 2014, Akhtar et al., 2009, Glover-Cutter et al., 2009) with a concurrent reduction in global basal transcription (Kwiatkowski et al., 2014, Kelso et al., 2014, Poss et al., 2016), however these experiments were mostly conducted using in vitro biochemical assays or with less selective chemical probes. Consistent with its role in transcription, CDK7 is also implicated in mRNA processing through the recruitment of capping enzymes, transcription activation through regulating the deposition of histone marks associated with active transcription, pause induction (Nilson et al., 2015, Ebmeier et al., Glover-Cutter et al., 2009), and pause release through phosphorylation of CDK9, the kinase subunit of the elongation complex pTEFb (Shiekhattar et al., 1995, Serizawa et al., 1995, Marshall et al., 1996). Together these results suggest a role for CDK7 in gene transcription; however it remains unclear how CDK7 kinase activity ultimately impact steady state mRNA levels (Marshall et al., 1996, Rodríguez-Molina et al., 2016, Larochelle et al., 2012).

Although the “bumped hole” approach to kinase inhibition has yielded key insights into CDK7’s biological role, this approach requires engineering mammalian cell lines and is difficult to use broadly; it also does not provide a route to developing new drugs. Selective small molecules capable of selective inhibition of endogenous CDK7 are therefore required to further delineate CDK7 activities in cell cycle progression and transcription. Previously, we reported the discovery of THZ1, a small molecule kinase inhibitor that targets CDK7 by covalently labeling a cysteine (C312) that lies in the CTD of CDK7 and is remote from the active site (Kwiatkowski et al., 2014). We also found that THZ1 inhibited CDK12/13 via the same covalent mechanism of action (Kwiatkowski et al., 2014, Zhang et al., 2016). To separate these activities, we sought chemical probe molecules with a narrower selectivity profile. This has yielded THZ531, a derivative of THZ1 with the same phenylaminopyrimidine core scaffold, that inhibits CDK12/13 approximately 20 times more potently than CDK7 (Zhang et al., 2016, Gao et al., 2018). However, efforts to generate a CDK7 selective inhibitor using the THZ1 scaffold have not been successful.

Here, we describe the development and mechanistic validation of a selective, covalent CDK7 inhibitor, YKL-5-124, based on a previously unexplored CDK-targeting scaffold that originates from the PAK4 inhibitor PF-3758309 (Murray et al., 2010a). We performed a series of structured biochemical and cell-based assays to validate a covalent mode of CDK7 inhibition by YKL-5-124 involving C312 modification and determined the compound’s selectivity profile. We also employed YKL-5-124 as a tool compound to investigate the functions of CDK7 in regulating cell cycle progression and gene transcription. Our data confirm the reported CAK activities of CDK7 in regulating cell cycle progression, but suggest a non-essential role for CDK7 in regulating Pol II CTD phosphorylation and global basal gene expression.

Results

YKL-5-124 covalently targets CDK7

Our initial efforts to develop a selective CDK7 inhibitor used THZ1 (Fig. 1) as a lead compound were not productive. We turned our attention to PF-3758309, a PAK4 inhibitor (Murray et al., 2010b) since molecules from the series of compounds that gave rise to PF-3758309 are reported to have strong CDK7 inhibitory activity (Rudolph et al., 2015). We envisioned that the hybridization of the covalent warhead from THZ1 and the pyrrolidinopyrazole core from PF-3758309 could lead to more selective CDK7 inhibitors. The first compound in the new series, YKL-1-116 (Fig. 1A) (Kalan et al., 2017), showed good selectivity for CDK7, but only moderate potency and had minimal anti-proliferative effects on cancer cell lines (Kalan et al., 2017). Further iterative optimization has focused on tuning the acidity of the aminopyrazole core, gaining additional hydrophobic interactions with the side chain residue, and optimizing the length and trajectory of the covalent warhead that targets C312. This focused medicinal chemistry campaign yielded YKL-5-124, a potent and selective covalent CDK7 inhibitor, and YKL-5-167, an inactive analog lacking the acrylamide reactive center and therefore incapable of forming a covalent bond with CDK7- C312 (Fig. 1A).

Figure 1. Biochemical characterization of YKL-5-124.

A. Chemical structures of THZ1, THZ531, YKL-1-116, YKL-5-124 and YKL-5-167. B. Analysis of time and dose-dependent inhibition of CDK7/CycH/MAT1 enzymatic activity by YKL-05-124 or THZ1 reveals a leftward shift in IC₅₀ values over time. Kinetics parameters of enzyme inactivation were obtained using nonlinear least-squares regression. Error bars represent the standard deviation for duplicate measurements. C. Docking model of YKL-5-124 in the ATP-binding site of CDK7. CDK7 is shown as blue ribbons with key residues highlighted in yellow and YKL-5-124 is modeled as a green structure, left. Expansion of the box in left is shown to the right and key residues that make contact with YKL-5-124 are labeled and highlighted. D. Efficiency of labeling was estimated to be approximately 99% gauged by the reduction in signal of triply and quadruply charged YFSNRPGPTPGCQLPRPNCPVETLK ions (residues 294–318). The peptides VPFLPGDSDLDQLTR (residues 180–194) and LDFLGEGQFATVYK (residues 15–28) were used for normalization. E. In vitro kinase assays of exogenous tagged-CDK7 protein immunoprecipitated with FLAG antibody after Hela cells were treated with YKL-5-124 or YKL-5-167 as indicated. WT, wildtype; C312S, C312 to serine mutation.

Biochemical evaluation of YKL-5-124 using a fixed time point, in vitro kinase assay indicated that YKL-5-124 inhibited CDK7/Mat1/CycH with an IC₅₀ of 9.7 nM, while the other CDKs tested biochemically, CDK2 and CDK9, had IC₅₀ values of 1300 nM and 3020 nM, respectively. To more precisely assay the contribution of covalent bond formation to compound inhibitory activity, k_inact/K_i was measured using a kinetic assay that monitors the shift in the electrophoretic mobility of a CDK7 peptide substrate following incubation with YKL-5-124 (Blackwell et al., 2009a, Tan et al., 2017). YKL-5-124 and THZ1 displayed similar k_i values (1.9 nM and 2.1 nM, respectively) showing that they achieved nearly equivalent inhibition of CDK7 (Fig. 1B, Sup Fig. 1A, B). However, YKL-5-124 exhibited a faster k_inact of 103 μs⁻¹nM⁻¹ as compared to k_inact of 9 μs⁻¹nM⁻¹, for of THZ1, demonstrating that YKL-5-124 covalently modifies CDK7 approximately 11 times faster than THZ1 (Fig. 1B, Sup Fig. 1A, 1B).

Molecular docking studies place the reactive acrylamide of YKL-5-124 near residue CDK7-C312, the same cysteine targeted by THZ1 (Fig. 1C). These docking results are confirmed by X-ray structures of kinases other than CDK7 complexed with inhibitors containing a tetrahydropyrrolopyrazole core such as maternal embryonic leucine zipper kinase (MELK, PDB accession codes 5MAG, 4BKY) and Serine/threonine-protein kinase PAK 4 (4APP) wherein the same binding mode is observed (Klaeger et al., 2017, Canevari et al., 2013, Guo et al., 2012).

The docking results additionally highlighted the possibility that YKL-5-124 can make two additional contacts with CDK7 at K41 and Q141, which may contribute to binding affinity and potency (Fig. 1C). Mass spectrometry analysis of recombinant CDK7-cyclin H-MAT1 trimeric complex incubated with YKL-5-124 demonstrated covalent labeling of CDK7 with an efficiency of ~99% (Fig. 1D, Sup. Fig. 1C, 1D) and capillary electrophoresis/mass spectrometry (CE-MS) confirmed the site of labeling to be C312 (Sup. Fig. 1E).

To determine whether YKL-5-124 is capable of inhibiting CDK7 in cells, Hela cells with wild-type or engineered CDK7 with the C312S mutation were immuno-precipitated from vehicle- or YKL-5-124-treated cells and precipitates subjected to in vitro kinase assays. YKL-5-124 blocked the in vitro phosphorylation of RNA polymerase II substrate by immuno-purified CDK7 at concentrations as low as 62.5 nM whereas the reversible YKL-5-167 analog demonstrated little or no inhibition of phosphorylation at 10μM (Fig. 1E). Mutation of the reactive cysteine to a less nucleophilic serine (C312S) abolished the inhibitory activity of YKL-5-124 in the in vitro kinase assay (Fig. 1E). Taken together, these data show that YKL-5-124 is a nanomolar covalent inhibitor that targets C312 on CDK7 with faster kinetics than THZ1 and 100-fold greater selectivity for CDK7 than CDK9 and CDK2.

YKL-5-124 specifically targets CDK7

To confirm that YKL-5-124 does not inhibit CDKs 12 and 13 we performed in vitro kinase reactions with CDK7, CDK12, and CDK13 at an ATP concentration of 1mM (Fig 2A). We found that YKL-5-124 had an IC₅₀ of 53.5 nM for CDK7, yet displayed no inhibition of CDK12 or CDK13 at the concentrations tested. In comparison, we found that THZ1 was equipotent on CDKs 7, 12, and 13. To determine the cellular selectivity of YKL-5-124 for CDK7 vs CDK12/13 we used HAP1 and Jurkat cells. HAP1 is a near-haploid human cell line derived from the male chronic myelogenous leukemia (CML) cell line KBM-7; Jurkat is a human T lymphocyte cell line previously used to characterize THZ1 (Essletzbichler et al., 2014, Andersson et al., 1996, Kwiatkowski et al., 2014). Lysates from cells treated with increasing concentrations of THZ1, YKL-5-124, or YKL-5-167 were treated with biotinylated-THZ1 (bioTHZ1) to affinity purify CDK complexes, which were then analyzed by Western blotting using antibodies against CDK7, CycH and CycK, the obligate binding partner of CDK12/13. THZ1 bound to both CDK7 and CDK12/13 based on its ability to block pull down of all three proteins (Fig. 2B, Sup Fig. 2); treatment of cells with YKL-5-124 as a competitor at a concentration of about 30 nM blocked pulldown of CDK7-cyclin H but had no effect on the pulldown of cyclin K- CDK12/13 (Fig. 2B, Sup Fig. 2A). YKL-5-124 displayed engagement of CDK7 three hours after compound was washed out of cells down to 125 nM, further confirming that YKL-5-124 functions in an irreversible manner (Fig. 2C, Sup Fig. 2B). In contrast, YKL-5-167, which cannot react with CDK7-C312, was 100-fold less effective as a competitor for CDK7-cyclin H in bioTHZ1 pulldown experiments and inactive following washout. To study the kinetics of target engagement, cells were treated with drug for 30 min to 6 hr prior to pulldown. Treatment with 100 nM YKL-5-124 reduced CDK7-cyclin H binding to bioTHZ1 by >50% at 30 min whereas THZ1 required twice as long, consistent with the faster k_inact of YKL-5-124 (Fig. 2D, Sup Fig. 2C). YKL-5-167 required 6 hours to achieve equivalent engagement of CDK7 in HAP1 cells and was even less active in Jurkat cells, requiring 10-fold higher concentrations and longer times to function as a competitor (Fig. 2D, Sup Fig. 2C).

Figure 2. Cellular engagement of CDK7 by YKL-5-124.

A. In vitro biochemical dose curves of CDK7, CDK12, and CDK13 with YKL-5-124 (left) and THZ1 (right) treatments. Error bars represent standard deviation of triplicate experiments B. Competitive pulldown assays in HAP1 cells treated with a concentration course of THZ1, YKL-5-124, or YKL-5-167 for 6 hours. Westerns show the pulldown or input of CDK7, Cyclin H, and Cyclin K. C. Competitive pulldown assays in HAP1 cells treated for 6 hours with a concentration course of YKL-5-124 or YKL-5-167 with or without washout after 3-hour treatment. D. Competitive pulldown assays in HAP1 cells treated with DMSO, THZ1, or YKL-5-124 for the time indicated before lysing and treating with bioTHZ1 to pulldown all available CDK7. PD, pulldown. E. Kinativ target engagement kinase tree showing kinases engaged over 65% with 1μM YKL-5-124 treatment (left) or YKL-5-167 treatment (right) in Jurkat cells. See also Table S1.

To assess target selectivity, we performed KiNativ profiling, a mass-spectrometry-based method that measures the ability of a test compound to block binding of kinases to a desthiobiotin-ATP probe. This showed that CDK7 is the only target bound by 1μM YKL-5-124 at more than 65% in Jurkat extracts pretreated with YKL-5-124 for 6 hours (Fig. 2E, Sup Table 1).Furthermore, YKL-5-167 exhibited a similar activity profile to YKL-5-124 except that CDK7 was only weakly engaged (Fig. 2E, Sup Table 1). To further demonstrate covalent binding of YKL-5-124 for CDK7, we used CRISPR to create an isogenic pair of cell lines expressing either wild type or CDK7-C312S (Sup Fig. 2D) and then used bio-THZ1 to perform pull-down experiments from cell extracts. We found that the C312S mutation prevented binding of CDK7-cyclin H to beads but had no effect on binding by CDK12/13-CycK (Sup Fig. 2E). Collectively, these data show that YKL-5-124 is a CDK7-selective agent, and provide additional support for the C312-dependent covalent mode of binding.

YKL-5-124 disrupts cell cycle progression through inhibition of CDK7 CAK activity

CDK7 functions as the CAK for several CDKs, including CDK1 and CDK2, activating them by phosphorylating key threonine residues in the T-loopT-loops of each kinase (Larochelle et al., 2007, Gerber et al., 2008). We, therefore examined the effect of exposing HAP1 cells 24 hours to 125 nM to 2 μM YKL-5-124 or 2 μM YKL-5-167 and then blotted cells extracts for the levels of total and phospho-CDK1(T161) and CDK2(T160). Blotting for PARP cleavage was used to monitor the extent of apoptosis. We found that YKL-5-124 inhibited CDK1 T-loop phosphorylation, and to a lesser extent CDK2 T-loop phosphorylation in (Fig. 3A) in a concentration-dependent fashion while YKL-5-167 had no detectable effect. Reduction in the phosphorylation of CDK1 and CDK2 T-loops was not observed in HAP1 cells expressing CDK7-C312S implying that the effect of YKL-5-124 on T-loop phosphorylation is mediated by CDK7 inhibition. (Fig. 3A). When HAP1 cells were exposed to 2μM YKL-5-124 for 6, 12, 24, and 48 hours, inhibition of T-loop phosphorylation was observed within 12 hr (Fig. 3B) The absence of PARP cleavage across all conditions showed that apoptosis was not induced by YKL-5-124 exposure. Similar data were obtained in Jurkat cells although the effect on CDK2 T-loop phosphorylation was greater than on the CDK1 T-loop (Sup Fig. 3A), suggesting variability among cell lines in CAK-substrate modification.

Figure 3. Cell cycle effects from YKL-5-124 treatment.

A. Western blot analysis of proteins shown after treatment of HAP1 WT and C312S cells with YKL-5-124 at concentrations indicated for 24 hours. (pCDK1- phospho T161 CDK1, tCDK1- total CDK1, pCDK2- phospho T160 CDK2, tCDK2- total CDK2) with % signal shows as a percentage compared to DMSO treatment in the relevant cell line. B. Western blot analysis of proteins shown after treatment of HAP1 WT and C312S cells with 2μM YKL-5-124 for time points indicated. (pCDK1- phospho T161 CDK1, tCDK1- total CDK1, pCDK2- phosphor T160 CDK2, tCDK2- total CDK2) C. Cell-cycle analysis using Hoechst, EdU, and pHH3 staining across concentrations of YKL-5-124 (left), YKL-5-167 (middle), or THZ1 (right) at 72 hours in WT cells. Error bars represent standard deviation of technical duplicates. D. Cell-cycle analysis using Hoechst, EdU, and pHH3 staining across concentrations of YKL-5-124 (left), YKL-5-167 (middle), or THZ1 (right) at 72 hours in CDK7 C312S cells. Error bars represent standard deviation of technical duplicates. E. GR value across concentrations in WT (red) and C312S (blue) HAP1 cells with YKL-5-124, left, YKL-5-167, middle, or THZ1, right. Error bars represent standard deviation of technical duplicates. F. Percent of LDR positive cells across concentrations in WT (red) and C312S (blue) HAP1 cells with YKL-5-124, left, YKL-5-167, middle, or THZ1, right. Error bars represent standard deviation of technical duplicates.

To monitor effects on cell cycle progression, HAP1 cells were exposed to YKL-5-124 or THZ1 for 72 hours and cell cycle analysis was performed by pulsing live cells 1 hour with EdU, fixing cells and staining with Hoechst and Phospho-histone H3 (pHH3) antibodies followed by flourescence microscopy. YKL-5-124 caused a dose-dependent increase in G1 and G2/M phase cells and a corresponding loss of S phase cells (Fig. 3C, left panel). Concurrent with S phase loss, we observed a notable increase in ‘S dropout’, or cells that have intermediate DNA content, but no EdU stain indicating inhibition of DNA synthesis. As a control we showed that cell cycle distribution was not affected by exposure of HAP1 cells expressing CDK7-C312S to YKL-5-124, further supporting the selectivity of YKL-5-124 and that CDK7 is mediating the effects of YKL-5-124 (Fig. 3C-D, left panel). The reversible compound, YKL-5-167, also showed no effect on cell cycle distribution.

To quantify drug-induced phenotypes we used growth rate inhibition (GR) metrics, which measure the effect of a drug on viable cell number corrected for the rate of cell division (Hafner et al., 2016, Hafner et al., 2018); GR₅₀ is a measure of potency and GR_max a measure of efficacy at high concentration. The GR_max value for THZ1 was negative, indicating cytotoxicity, whereas that of YKL-5-124 was zero indicating a cytostatic response (Fig. 3E) (Hafner et al., 2016). For THZ1, the GR₅₀ value was 3.6-fold higher in CDK7-C312S expressing than in wild-type cells but the GR_max value was unchanged. In contrast, for YKL-5-124 the GR₅₀ value was at least 150-fold higher in CDK7-C312S that in wild-type cells and GR_max was too high to be estimated accurately. Staining cells with LIVE/DEAD dye (Hafner et al., 2016) confirmed that THZ1 induced substantially more cell death than YKL-5-124 and killing was relatively unaffected by CDK7-C312S mutation (whereas cell killing by YKL-5-124 was substantially reduced in CDK7-C312S cells). From these data we conclude that YKL-5-124 primarily induces G1/S cell cycle and a low level of cell killing in a CDK7-dependent manner whereas THZ1 is cytotoxic in a largely CDK7 independent manner, as indicated by the potent effect in CDK7-C312S cells (Fig 3F). Cell Viability data based on ATP content were obtained in Jurkat cells where treatment with THZ1 was able to inhibit proliferation to zero, indicating total loss of cell viability at the highest concentrations, while treatment with YKL-5-124 could only inhibit proliferation to within 20% of the original cell count as measured by cell titer glo (Sup Fig. 3C). Together these data show that YKL-5-124 is primarily cytostatic at G1/S and that arrest in CDK7-dependent; in contrast, THZ1 is cytoxic in CDK7- independent manner.

YKL-5-124 treatment has no discernable effect on RNA Pol II CTD phosphorylation

CDK7 is reported to be the kinase responsible for phosphorylating S5 in the RNA polymerase II CTD heptapeptide repeat, and therefore we asked whether exposure of cells to YKL-5-124 would reduce Pol II CTD phosphorylation. Surprisingly, in both HAP1 (Fig. 4A) and Jurkat (Sup. Fig. 4A) cells we found that YKL-5-124 concentrations 100-fold above GR_max (up to 2 uM) had no effect on the phosphorylation of S5 and other Pol II CTD serine residues cells. However, inhibition of CTD phosphorylation by THZ1 has previously been demonstrated at nanomolar doses (Kwiatkowski et al., 2014). This inconsistency led us to ask whether loss of Pol II CTD phosphorylation requires inhibtion of CDK7, and CDK12/13 by THZ1. To test this hypothesis we combined YKL-5-124 with a previously described inhibitor of CDK12/13, THZ531 and found that the combination was active as a concentration dependent inhibitor of CTD S2/S5/S7 phosphoryation (Fig. 4A-B, Sup 4A-B). Inhibition of CDK7 was require for this effect, since it was not observed in CDK7-C312S cells. We conclude that the CTD of Pol II is modified by a combination of CDK7 and Cdk12/13 (Fig4. A,B).

Figure 4. Transcriptional effects from YKL-5-124 treatment.

A. Western blot analysis of proteins shown after treatment of HAP1 WT and C312S cells with YKL-5-124 at concentrations indicated for 24 hours and co-treatment with 250nM THZ531, right. B. Western blot analysis of proteins shown after treatment of HAP1 WT and C312S cells with 2μM YKL-5-124 for time points indicated and co-treatment with 250nM THZ531, right. C. Scatter plot of FKPM of genes comparing YKL-5-124 500nM treatment in wildtype and C312S cells. D. Jaccard plot comparing the genes downregulated by YKL-5-124 500nM, THZ1 100nM, THZ531 250nM, or YKL-5-124 and THZ531 co-treatment. E. Radar plot comparing the effects of different drug treatments on genes upregulated (left) or downregulated (right) by treatment with YKL-5-124 at 500nM on wildtype (red) and C312S (gray) cell lines.

Inhibition of CDK7 by YKL-5-124 elicits a transcriptional signature distinct from that of THZ1

To measure the effects of CDK7 inhibtion of YKL-5-124 on gene transcription we performed RNA sequencing. After 24 hr of exposure of WT HAP1 cells to 500nM YKL-5-124, differential expression of 1452 genes was observed (a log₂ fold change >1.5) as compared to 583 genes in CDK7-C312S cells (Fig. 4C, Sup Fig 4D). Under DMSO-only control conditions the transcript profiles of WT and CDK7-C312S were highly similar demonstrating that the mutation does not appreciably affect baseline transcription (Sup Fig. 4C). Next, we compared differentially expressed genes in cells exposed to YKL-5-124, THZ1, THZ531, or a combined of YKL-5-124 and THZ531 by Jaccard analysis, a 0-1 bounded measure of gene set overlap (Levandowsky and Winter, 1971). In the WT HAP1 cells, the effects of 500 nM YKL-5-124 were most similar to those of a combination of 500 nM YKL-5-124 plus 250 nM THZ531 (Jaccard similarity index = 0.45) somewhat less similar to 100 nM THZ1 (index = 0.23) and least similar to 250 nM THZ531 alone (index = 0.05; Fig. 4D). Genes that were upregulated or downregulated by YKL-5-124 are shown as radar plots in Fig. 4E and were largely similar to those differentially regulated by 100 nM THZ1. The number of genes differentially regulated by YKL-124 was substantially reduced to 583 in CDK7-C312S cells, showing that the effect was was mediated by inhibition of CDK7. Morever, the similarity between YKL-5-124 and the YKL-5-124/THZ531 combination fell in CDK7-C312S relative to WT cells (index = 0.29 v. 0.45) and rose relative to THZ531 alone (index = 0.23 v. 0.05; Sup Fig. 4D). These data suggest that CDK7 inhibition alone is not sufficient to affect global basal transcription and that simultaneous inhibition of CDK12/13 is required. Interestingly, our data show that the converse is also true: CDK12/13 inhibition without CDK7 inhibition is insufficient to block global basal transcription.

CDK7 inhibition leads to a strong cell cycle transcriptional profile

GSEA analysis of expression data from WT HAP1 cells treated with YKL-5-124 showed strong leading edge enrichment for downregulation of E2F expression programs (Fig. 5A)(Subramanian et al., 2005). For THZ531, we observed no significant enrichment for E2F gene expression and instead saw downregulation of genes normally repressed by EZH2 (Lu et al., 2010) (Sup Fig. 4E). When we focused specifically on a curated set cell cycle regulated genes (Chang et al., 2004) (which includes the E2F program) differentially expressed in WT and CDK7-C312S cells at 500 nM or 2 μM we observed a statistically significant decrease in their expression (Fig. 5B). Among genes in the E2F3 cell cycle program, PLK1 and CDCA3 both exhibited dose-dependent decreases in transcription with YKL-5-124 treatment (Fig. 5C, 5D).Taken together, we find a strong inhibition of cell cycle transcriptional program in cells treated with YKL-5-124 for 24 hours.

Figure 5. Cell Cycle Transcriptional effects from YKL-5-124 treatment.

A. Gene set enrichment analysis of YKL-5-124 500nM treatment. B. Bar plot quantitating the expression levels on cell cycle genes with DMSO, YKL-5-124 500nM, or YKL-5-124 2μM treatment in wildtype and C312S cell lines. C. Bar plot quantitating PLK1 expression with DMSO, YKL-5-124 500nM, or YKL-5-124 2μM treatment in wildtype and C312S cell lines. D. Bar plot quantitating CDCA3 expression with DMSO, YKL-5-124 500nM, or YKL-5-124 2μM treatment in wildtype and C312S cell lines with error bars representing standard deviation.

Discussion

Here we performed target deconvolution on the previously reported CDK7/12/13 inhibitor, THZ1, parsing the CDK7 and CDK12/13 activities to (1) to isolate and examine the relative contribution of CDK7 inhibition to the overall THZ1 phenotype, and (2) further explore the consequences of CDK7 pharmacological inhibition. To do this, we leveraged a covalent inhibitor strategy to target CDK7 and other kinases with similarly positioned cysteines to the exclusion of the broader kinome, followed by scaffold diversification to gain further selectivity for CDK7, an approach we previously deployed to selectively inhibit a diverse array of kinase targets (Zhang et al., 2012, Zhang et al., 2016, Kwiatkowski et al., 2014). This method culminated in the development of YKL-5-124, an irreversible CDK7 inhibitor derived from the previously reported pyrrolopyrazole-based PAK4 inhibitor, PF-3758309 (Murray et al., 2010b). Covalent targeting of CDK7 at C312 is obligatory for YKL-5-124 activity as mutation to the less nucleophilic serine completely rescued both CDK7 activity and cellular proliferation, highlighting both a clear mechanism of action and exceptional selectivity enabling YKL-5-124 use across cellular systems.

Unique among CDKs, CDK7 has been reported to play major roles in both the cell cycle and transcription regulation, through its CAK- and TFIIH-related functions, respectively (Gerber et al., 2008, Ebmeier et al., Feaver et al., 1994, Adamczewski et al., 1996, Glover-Cutter et al., 2009, Kelso et al., 2014, Rodríguez-Molina et al., 2016). However, many mechanistic details remain elusive and the range of CDK7 functions is incompletely understood. In this study, we employed our newly developed CDK7-selective chemical probe, YKL-5-124, to examine the effects of CDK7 pharmacological inhibition on both the cell cycle and global transcription. Our treatment of HAP1 CML and Jurkat T-ALL cells with YKL-5-124 reduced both the levels of activated CDK1/2 and the number of cells in S phase, consistent with previous findings (Larochelle et al., 2007). Interestingly, whereas reduced CDK1/2 activation occurred at concentrations exceeding 1 μM YKL-5-124, strong anti-proliferative effects were observed beginning at 100 nM, suggesting additional substrates of CDK7 that are important mediators of cell cycle progression remain to be discovered. RNA-seq analysis following treatment with YKL-5-124 revealed a strong cell cycle signature enriched for E2F3 target genes, indicating that CDK7 may indeed impinge on the cell cycle by various mechanisms. Expression of the inhibitor-refractory C312S mutation rescued all these phenotypes indicating these are on-target effects of CDK7 inhibition.

Surprisingly, YKL-5-124 treatment alone did not globally downregulate RNA Pol II CTD phosphorylation or Pol II-mediated gene expression. This stands in direct contrast to our previous inhibitor, THZ1, which elicited robust downregulation of RNA Pol II CTD phosphorylation and gene expression (Kwiatkowski et al., 2014) . However, we were able to recapitulate THZ1-mediated effects on gene expression and CTD phosphorylation when YKL-5-124 was combined with our previously characterized CDK12/13 inhibitor, THZ531 (Zhang et al., 2016). This indicates that combined inhibition of CDK7, CDK12, and CDK13 is central to the anti-transcriptional effects of THZ1. Similar results have been observed when the CDK7 bumped hole inhibitor, 3MB-PP1, was combined with the pan-CDK inhibitor, flavopiridol, which strongly inhibits CDK9, another Ser2 CTD kinase (Larochelle et al., 2007). This suggests that phosphorylation of the CTD by CDK7 may be dispensable for basal global gene expression and that CDKs 9/12/13 or other kinases such as ERK1/2 (Bonnet et al., 1999) may largely compensate for loss of CDK7. Alternatively, CDK7’s role may become essential when Ser2 CTD phosphorylation and corresponding Pol II elongation is first perturbed. This does not exclude the possibility that: (1) CDK7 inhibition leads to potent, but transient loss of Pol II CTD phosphorylation and gene expression, which is quickly reset following engagement of compensatory mechanisms; (2) CDK7 is solely responsible for Ser5 or Ser7 phosphorylation at specific genes whose expression changes following its inhibition; (3) CDK7 kinase activity is not required for basal expression of most genes, but may be involved in gene induction; and (4) CDK7 may regulate gene expression by other means including by transcription factor phosphorylation. These findings highlight the power of using selective kinase inhibitors in combination with complementary genetic tools and systems level analyses to investigate complex signaling pathways.

Significance

CDK7 has long been reported to act as both the CDK activating kinase (CAK) and an RNA Polymerase II (Pol II) C-terminal domain (CTD) kinase, thereby imparting top-down regulation of the cell cycle and gene transcription, respectively. However, results and interpretations thereof have been complicated by the use of non-selective CDK-targeting small molecules. Here we describe the development of YKL-5-124, a potent and selective covalent inhibitor of CDK7. Using YKL-5-124 we confirmed a strong cell cycle phenotype in cells following CDK7 inhibition with reduction in cycling cells, CAK activity, and cell cycle-regulated gene expression. However, cells exhibited little-to-no change in global Pol II CTD phosphorylation or general loss of transcription factor genes, which are hallmarks of THZ1 treatment. Furthermore, while YKL-5-124 treatment of HAP1 and Jurkat cells led to a strong cell cycle arrest, it did not induce cell death like THZ1. The ability to recapitulate THZ1 phenotypes using potent by separate inhibitors of CDK7 and CDK12/13 signal the importance of this polypharmacology to the transcriptional activity of THZ1 and derivative compounds. Therefore, YKL-5-124 represents a compound that enables rapid inactivation of CDK7 to study the consequences of selective CDK7 inhibition on both the cell cycle and gene transcription. As YKL-5-124 does not indiscriminately induce cell death, it also presents a unique therapeutic opportunity to find cancer cell states that are uniquely dependent on CDK7 inhibition for survival. As YKL-5-124 demonstrated suppression of the cell cycle at both the gene expression and post-translational levels, therapeutic strategies aimed at targeting cancers that are driven by direct misregulation of cell cycle factors may show promise.

Contact for Reagent and Resource Sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Nathanael Gray (nathanael_gray@dfci.harvard.edu).

Experimental Model and Subject Details

Cell lines

HAP1 (male) cells (Horizon Discovery) were cultured in IMDM media (Cat # 12440061, Thermo Fisher) supplemented with 10% FBS and 1% penicillin/streptomycin (Cat # 10378016, Thermo Fisher) and Jurkat (male) cells (T. Looke Lab, DFCI) were cultured in RPMI media (Cat# 11875093, Thermo Fisher) and incubated at 37 °C with 5% CO₂. All cell lines were routinely tested for mycoplasma infection (Lonza, LT07-318). Cell lines were not authenticated.

Constructs

For CRISPR studies, guide RNAs (sgRNA) targeting CDK7 were cloned into pX330 (Addgene: 42230) (Cong et al., 2013). pUC57-AMP was used as storage vector for CDK7 genome reference sequence.

Method Details

Genome Editing

The CRISPR/Cas9 system was used to mutate the endogenous CDK7 WT locus to encode for CDK7 C312S, a THZ1 and YKL-5-124 -refractory mutant of CDK7. Target-specific oligonucleotides were cloned into pX330 (Addgene: 42230), which carries a codon-optimized version of Cas9 and was further modified to express GFP for identifying transfectants. Cells were co-transfected (X-tremeGENE 9 (Roche)) with (1) pX330 expressing Cas9 and CDK7 targeting sgRNA and (2) a pUC57-AMP construct bearing 1500 bp of modified CDK7 reference genome that is centered around the CRISPR targeting site in CDK7. Two days after transfection, cells were sorted using GFP as a marker of transfected cells and cells were replated for five days. Cells were then re-plated at low density to facilitate the isolation of individual clones. Individual clones were isolated, expanded, and PCR genotyped using mutant specific PCR primers. Following initial PCR screening, individual clones were Sanger sequenced to confirm the presence of the desired mutation. Western blot confirmed the presence of intact CDK7 kinase. Subsequent experiments were conducted using a CDK7 C312S clone and a WT control clone that was carried through the entirety of the CRISPR protocol, but that was verified by Sanger sequencing to be WT CDK7.

The genomic sequence complementary to the CDK7 -directed guide RNA that was cloned into pX330 and used in the genome editing experiments is: GCCAAGACCAAACTGTCCAG.

The modified genomic sequence that was cloned into pUC57-AMP by Genewiz and used as the repair template for genome editing is:

GCCATGAGGAGTCGACTTAATCTTGGCTTCCATCCCCAAGGTACGTCATGTATATGCAAATATTCCACAGTCCAAAATCTGAAGCACTTCTGGTCCCAAGCATTTTAGATAAGATATATTCAGCCTGTATAGGCTATTTAATTATGGAAAAATTAGAAGGGAGTCTTTTTTTCTTTTCTTTTTTTGATACGGAGCCTTGCTCTGTCGCCCAGGCTGGAGTGCAGTGGTGCGATCTCGGCTCACTGCAACCTCTGCCTCTGGGTTCAAGCGATTCTCCTACCTCAGCCTCCCGAGTAGTGGGGACTACAGACATGCACCACCACGCCCAGCTTATTTTTGTATTTTTAGTAGAGATGGAGTTTCGCCACGTTGGCCAAGCTGATCTCGAACTTCTGACCTCATGTGATCCACCCGCCTCAGCCTCCCAAGGTGCTGGGATTACAGGCATGAGCTACTGTGCCCGGCCAGTCTTTTAATGGAATTATCAGCCAATCTTAGGATCCTAGGTATTTATTTGTAGTAATTTCCATTTGCAAATACTTGACATTAATTTTGTTTTGTTGGGAATGAGGGCATTCACATAGTATTTTTAATGCTCCCTATAATCTTGAAGGTAAGATTTTCCTTAACACATTTCAGATACTCACCTACCTTACTTTTGGTATCTTTTCTTTTAAAAGGCACTGAAAATGAAGTATTTCAGTAATCGGCCAGGGCCAACACCTGGATGTCAGCTGCCAAGACCAAACTCACCGGTAGAAACCTTAAAGGAGCAATCAAATCCAGCTTTAGCAATAAAAAGGAAAAGAACAGAGGCCTTAGAACAAGGTAAGATTCCCACTTTTAAAAGAAATTAAATGAATTTAGAAAACTCCAAATAGCTCGTGTATGGCTAGCGACTGAACAACAGCAAAGAAATGTAGCTTGCTAGCTATTTAATTTTTATAAATTTAATTGTATAAAGTAGAGAGACTGTGCCGTTTTAGGTATGTTTACTTTCATTGTGAATACTTCGGCAGCTTTTGTGTATACCTCAGAATTATGGGAACTATGTATGTGTACCTTGTTCCCTTCATAAATTCCTTCTGTAAACTACATGCTAAATGTTTAACAAGTGCTACTGGAAAAAATGTGTCAGTGAAATAATGTAGTTGGAATGCAGTGACTGGTTAAAATTGCTATGAGAATGTGAACTTTGTTAAATGTGAACAACTTTTTTTTTTCTTGTTCTTTTTGCTTCCTAGGAGGATTGCCCAAGAAACTAATTTTTTAAAGAGAACACTGGACAACATTTTACTACTGAGGGAAATAGCCAAAAAGGCAAATAATGGAAAAATAGTAAACATTAAGTAAATGCTGTAGAAGTGAGTTTGTAAATATTCTACACATGTAAAATATGTAAAACTATGGGTTATTTTTATTAAATGTATTTTAAAATAAAAATTTAATTCTGGTTTTTCTGATTAGAGTGCAAAAGTGAGAAAGAATTCTCGAGTACCG

1. Green highlighting indicates the introduced desired TCA mutation, which codes for serine (C312S), replacing TGT which codes for cysteine (C312, WT)

2. Yellow highlighting indicates wobble mutations introduced to remove Cas9 –targeting sites, to prevent cutting of repair template.

3. Pink highlighting indicates Wobble mutations introduced for PCR-based screening to permit WT vs. mutated allele discrimination.

4. Red highlighting indicates Sal I and EcoRI sites used for pUC57 cloning

Immunoblotting

HAP1 or Jurkat cells were lysed in RIPA buffer (50mM Tris-HCl, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, and 0.1% SDS, pH 7.4 ± 0.2) with Protease Inhibitor (Roche), Phosstop Phosphatase Inhibitor (Roche), and 2.5 U/ml Universal Nuclease for Cell Lysis (Pierce) by incubating on ice 30 min. The lysates were clarified by spinning at 21,000 g for 30 min at 4 °C and the concentration of the lysate was determined using BCA protocol (Pierce). Primary antibodies in this study include: CDK7 (Cell Signaling Technology, 2916, 1:1,000), Cyclin H (Bethyl Labs, A301-674A , 1:1000), Cyclin K (Bethyl Labs, A301-939A, 1:1000), CDK2 (Bethyl Labs, A301-812, 1:1,000), Phospho-CDK2 T160(Cell Signaling Technology, 2561, 1:1000) CDK1 (Bethyl Labs, A303-663, 1:1,000), Phospho-CDK1 T161 (Cell Signaling Technology, 9114, 1:1000)Tubulin (Cell Signaling Technology, 3873 1:5,000), PARP (Cell Signaling Technology, 9542, 1:2,000), Phospho-CTD Ser2 (Millipore, 04-1571, 1:2,000), Phospho-CTD Ser5 (Millipore, 04-1572, 1:5,000), Phospho-CTD Ser7 (Millipore, 04-1570-I, 1:1000), Total Pol II (Santa Cruz Biotechnology, sc-899, 1:200). Secondary antibodies were infra-red labeled antibodies (LI-COR, used at 1:10,000) and blots were imaged on an Odyssey CL_Ximager.

Commercial in vitro kinase assays

Adapta Eu kinase assays were conducted for CDK7/CycH/MNAT1 and CDK9/CycT1 at Life Technologies using Km ATP concentrations. Z'LYTE kinase assays was conducted for CDK2/CycA at Life Technologies using Km ATP concentrations.

In vitro kinase assays

Recombinant kinases GST-CDK12 (714-1063)/GST-Cyclin K (1-267) and GST-CDK13 (694-1093)/GST-Cyclin K (1-267) were co-expressed with Cdk-Activating Kinase (CAK) from S. cerevisiae in Sf9 insect cells. The tripartite complex GST-CDK7 (2-346)/Cyclin H (1-323)/MAT1 (1-309) was co-expressed from a single vector in baculo-virus infected Sf9 insect cells using the MultiBac^Turbo system (Bieniossek et al., 2012).

Cells expressing the respective kinase complexes were harvested by centrifugation and resuspended in lysis buffer (50 mM HEPES pH 7.6, 500 mM NaCl, 10% glycerol, 5 mM β-mercaptoethanol). Cells were lysed by sonication. All kinases were purified via GST-affinity purification using pre-packed GSTrap 4FF columns (GE Healthcare). The lysate was cleared by centrifugation in a Beckman Optima L-80 XP Ultracentrifuge with a Ti45 rotor (45,000 r.p.m. for 45 min at 4°C) and applied to GSTrap 4FF columns (G E Healthcare) equilibrated with lysis buffer using an ÄKTA PrimePlus chromatography system (GE Healthcare). Following extensive washing with 10 column volumes (CV) of lysis buffer and 5 CV of wash buffer (50 mM HEPES pH 7.6, 1000 mM NaCl, 10% glycerol, 5 mM β-mercaptoethanol), the protein was eluted in elution buffer (50 mM HEPES pH 7.6, 500 mM NaCl, 10% glycerol and 5 mM β-mercaptoethanol, 10 mM glutathione). GST-tags were removed by TEV-protease cleavage overnight at 4°C. Proteins were further purified by size exclusion chromatography on a Superdex 200 PG column (GE Healthcare) equilibrated in 20 mM HEPES pH 7.6, 150 mM NaCl, 1 mM TCEP). Proteins were eluted from the column with the same buffer used for equilibration in 2 mL fractions. Fractions were analyzed by Coomassie staining of SDS-PAGE and fractions containing stoichiometric complexes were pooled and concentrated by ultrafiltration in Amicon filters (Millipore). Kinases were aliquoted in single use aliquots, snap frozen, and stored at −80°C until used.

Radioactive kinase activity measurements were performed at a concentration of 0.2 μM CDK/Cyclin complex. Kinase was incubated with varying concentrations of inhibitory compounds in presence of 1 mM ATP containing 0.45 μCi ³²P /μL (Perkin Elmer) for 5 minutes at 30°C prior to starting the kinase reaction by ad dition of substrate. Reactions were incubated for 15 minutes at 30°C and stopped by adding EDTA t o a final concentration of 50 mM. Reaction mixture was spotted onto filter sheets of Amersham Protran nitrocellulose membrane (GE Healthcare). Filters were washed three times for 5 minutes with 0.75% (v/v) phosphoric acid. Radioactivity was counted in a Beckman Liquid Scintillation Counter (Beckman-Coulter) for 1 minute. Data were obtained from three independent experiments and normalized to DMSO control for comparison.

Radioactive in vitro kinase assays from cellular CDK7

For kinase assays Hela (female) cells with stable expression of wildtype or C312S flag-tagged CDK7 were first treated with DMSO, YKL-5-124, or YKL-5-167 for 6 hours. Cells were then harvested by lysis in 50 mM TrisHCl pH 8.0, 150 mM NaCl, 1% NP-40, 5 mM EDTA, Protease Inhibitor (Roche), and Phosstop Phosphatase Inhibitor (Roche). CDK7-FLAG was immunprecipitated from lysates using M2-agarose resin (Sigma, A2220). Precipitated proteins were washed with lysis buffer 4 times, followed by 2 washes with kinase buffer (40 mM Hepes pH 7.5, 150 mM NaCl, 10 mM MgCl2, 5% glycerol) and subjected to in vitro kinase assays at 30°C for 45 minutes using 1 μg of the large subunit of RNAPII (RPB1) as substrate and 25 μM ATP and 10 μCi of 32P ATP. Reactions were run on SDS-PAGE gel, fixed and stained with coomassie reagent, and exposed to phospho-imager screen before being scanned on a TYPHOON imager (GE Healthcare Life Sciences).

Pulldown assays

HAP1 or Jurkat cells were treated with DMSO, YKL-5-124, YKL-5-167, or THZ1 for time and concentration indicated. Cells were trypsinized (HAP1 only), pelleted, washed with 1× PBS and resuspended in 50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 5 mM EDTA, 1 mM DTT and protease/phosphatase cocktails. Lysates were clarified and input samples were taken before 1 μM biotinylated-THZ1 probe was added and samples were rotated overnight at 4 °C. Lysates with probe were then incubated wit h streptavidin resin for 3 h at 4 °C. Beads were washed with lysis buffer five times and resuspended in 2× LDS load dye. Samples were boiled at 95 °C for 5 min before they were loaded o nto SDS-PAGE gels and subjected to immunoblotting.

Covalent inhibition kinetics using Mobility Shift Assay

k_inact/K_i measurements were determined using kinetic CDK7 enzyme activity data obtained from a peptide substrate mobility shift assay (Blackwell et al., 2009b). Recombinant CDK7/CycH/Mat1 and fluorescent peptide substrate were obtained from Carna Biosciences (CDK7/CycH/Mat1, Cat. 04-108; CTD3 peptide, Cat. 04-108MS). Reactions contained 3-5 nM CDK7 enzyme, 1 μM CTD3 peptide and 1 mM ATP in the assay buffer (20 mM HEPES pH7.5, 0.01% TritonX-100, 5 mM MgCl₂, 0.5 mM DTT). Phosphorylated product and non-phosphorylated substrate were separated and detected using a LabChip EZ Reader® (PerkinElmer) using the following settings: pressure 1.8 psi and voltage differentail −2400V to −500V with separation buffer consisting of 100 mM Hepes pH 7.3, 0.015% Brij-35, 1 mM disodium EDTA, 0.1% coating reagent 3, 5% DMSO and 1X coating reagent 8, similar to previous reports (Tan et al., 2017) . Nonlinear least-squares regression was performed using the DynaFit software package (Kuzmic, 1996) using with default settings for a 2-step inactivation model. Figures were generated with GraphPad Prism software (version 6.07).

Mass Spectrometry Analysis

nanoLC/MS

Recombinant CAK complex (CDK7/Cyclin H/MNAT1, Thermo Fisher Scientific) was incubated with YKL-05-124 or DMSO for 4 hours at 37 °C, and reactions were resolved by SDS-PAGE. Bands corresponding to CDK7 were excised, digested in gel with trypsin according to standard protocols, and analyzed by nanoLC-MS (Ficarro et al., 2009). Extracted peptides were loaded via autosampler injection (NanoAcquity Sample Manager) onto a precolumn (4 cm POROS 10R2, Applied Biosystems, Framingham, MA) and eluted with an HPLC gradient (NanoAcquity Binary Solvent Manager, Waters; 2-35% B in 60 minutes; A=0.1 M acetic acid in water, B=0.1M acetic acid in acetonitrile). Peptides were resolved on a self-packed analytical column (50 cm Monitor C18, Column Engineering) and introduced to a Q-Exactive HF mass spectrometer (ThermoFisher Scientific) at a flow rate of ~30 nL/min (ESI spray voltage = 2.2 kV). The instrument was operated in data dependent mode such that the 10 most abundant precursors were subjected to MS/MS (HCD, NCE=30%, resolution 15K). mzStudio software (Ficarro et al., 2017) was used to perform peak integrations for the triply and quadruply charged precursors of the doubly carbamidomethylated C312 containing peptide. Two unlabeled, non-cysteine containing CDK7 peptides were used for normalization.

ZipChip CE-MS

Recombinant CAK complex was treated with YKL-05-124 as above, denatured with Rapigest (0.1% final concentration; Waters, Milford, MA), reduced (10 mM dithiothreitol), alkylated (22.5 mM iodoacetamide), and digested with trypsin (Promega) overnight at 37 °C. After cleaving Rapigest according to the manufacturer’s instructions, peptides were desalted using C18, dried by vacuum centrifugation, and reconstituted in 1% formic acid/50% acetonitrile with 100 mM ammonium acetate. Peptides were then analyzed by CE-MS using a ZipChip CE system and autosampler (908 Devices) interfaced to an Orbitrap Lumos mass spectrometer (ThermoFisher Scientific). Peptide solution was loaded for 30 seconds and separation performed at 500 V/cm on an HR chip for 10 minutes with a background electrolyte consisting of 1% formic acid in 50% acetonitrile. Pressure assist was utilized and started at 1 minute. The mass spectrometer was operated in data dependent mode and subjected the 5 most abundant ions in each MS scan (60k resolution, 1E6 target, lock mass enabled) to MS/MS (15k resolution, 2E5 target, 100 ms max inject time). Dynamic exclusion was enabled with a repeat count of 1 and an exclusion time of 6 seconds. MS/MS data was extracted to .mgf using multiplierz scripts (Alexander et al., 2017, Askenazi et al., 2009) and searched against a custom database containing the sequences of CDK7, Cyclin H, and MAT1 using Mascot version 2.2 (Perkins et al., 1999). Search parameters specified fixed carbamidomethylation of cysteine, and variable oxidation (methionine) and variable YKL-05-124 modification (cysteine). Precursor mass tolerance was set to 10 ppm and product ion tolerance was 25 mmu. To confirm the site of labeling, the mass spectrometer performed cycles of one MS scan (m/z 400-2000, electron multiplier detection, target = 5E4) followed by targeted ETD-MS/MS scans corresponding to the quadruply and quintuply charged precursor ions of the YKL-05-124 modified C312 containing CDK7 peptide (image current detection, 15k resolution, 5E5 target, 250 ms max inject time).

Kinativ sample preparation and analysis

Jurkat cells were treated with DMSO, YKL-5-124 (1μM), or YKL-5-167 (1μM) for 6 hours. Cells were washed with PBS 1X and frozen in liquid nitrogen. Samples were processed by Kinativ (ActivX Biosciences Inc.) as previously described (Patricelli et al., 2011). Briefly, cells were lysed using sonication in lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 0.1% Triton-X-100, phosphatase inhibitors (Cocktail II AG Scientific #P-1518)) before being cleared by centrifugation. The supernantant is labeled with 20 μM desthiobiotin-ATP acyl-nucleotide probe for 10 minutes to allow covalent engagement of the probe to proteins with available active sites to identify target engagement. Lysates are then denatured (6 M urea, 10 mM DTT, 65°C, 15 min), alkylated(40 mM iodoacetamide, 37°C, 30 mi n), and filtered into 10 mM ammonium bicarbonate, 2 M urea, 5 mM methionine before digestion with trypsin for 1 hour at 37°C and subjecting peptides to pulldown using high-capacity streptavidin resin (Thermo Scientific). Peptides were was in 50% CH₃CN/water mixture containing 0.1% TFA at room temperature then subjected to mass spectrometric analysis.

Proliferation assays.

Proliferation assays were performed in triplicate by plating Jurkat cells at 2 × 10⁴ cells/mL in 50μLin a 384-well plate. Cells were treated with 100nL of YKL-5-124 or THZ1 for final concentrations ranging from 20μM to 600pM for 72 h. Anti-proliferative effects of compounds were assessed by adding 25 μL of Cell Titer Glo (Promega) and luminescence was measured on an Envision plate reader (PerkinElmer). IC₅₀ values were determined using Graphpad Prism nonlinear regression curve fit.

Cell Cycle and Growth rate analysis

HAP1 WT and C312S mutant cells were plated at a density of ranging 500 cells per well in 384-well Cell Carrier plates (Perkin Elmer) using a Multidrop Combi Reagent Dispenser (Thermo Fisher Scientific) and allowed to adhere to for 24 hours prior to drug treatment. Cells were treated with a dilution series of the indicated drugs using a D300 Digital Dispenser (Hewlett-Packard). Cells were stained and fixed for analysis at the time of drug delivery and after 72 hours of incubation. Cells were pulsed for one hour with EdU (Lumiprobe) and stained with 1:2000 LIVE/DEAD Far Red Dead Cell Stain (LDR) (Thermo Fisher Scientific). Cells were then fixed with 3.7% formaldehyde (Sigma Aldrich) for 30 minutes and permeabilized with 0.5% Triton X-100 in PBS. The EdU was labeled with cy3-azide (Lumiprobe) for 30 min. The cells were then blocked for one hour with Odyssey blocking buffer (LI-COR), and stained overnight at 4°C with 2 μg/ml Hoechst 33342 (Sigma Aldrich) and a 1:1000 dilution of anti-phospho-histone H3 (pHH3)Alexa 488 (Ser10, clone D2C8) conjugated antibody (Cell Signaling Technologies). Fixed cells were imaged with a 10x objective using an Operetta microscope and analyzed using the Columbus image data storage and analysis system (Perkin Elmer). Nuclei were segmented using Columbus software (Perkin Elmer) based on their Hoechst signal. DNA content was defined by the total Hoechst intensity within the nuclear mask to identify cells in the G1 and G2 phases of the cell cycle. The average LDR, EdU and phospho-histone H3 intensities within the nuclear masks were determined. The LDR signal was used to classify cells as live or dead, the EdU and pHH3 signals to identify S and M phase cells respectively. Cells with intermediate DNA content and no EdU signal were classified as S phase dropout cells. Live cell counts were normalized to DMSO-treated controls on the same plates to yield normalized growth rate inhibition (GR) values as described previously (Hafner et al., 2016). Experiments were performed as technical duplicates in biological triplicate.

RNA Sequencing.

1×10⁶ HAP1 Cells were treated for 24 h with 500nM or 2μM YKL-5-124, 100nM THZ1, 250nM THZ531, or the combination of 250nM THZ531 and 500nM or 2μM YKL-5-124. Library preparation was performed using TruSeq Stranded mRNA Library Prep Kit.

RNA-seq Analysis

All RNA-seq analysis was performed using the human reference genome build HG19 with HG19/GRCh37 Refseq gene annotations. Data were aligned using Hisat2 with default parameters to the UCSC Refseq transcriptome as provided by the Illumina iGenomes project (https://support.illumina.com/sequencing/sequencing_software/igenome.html). Raw and processed data are deposited into the NCBI GEO repository (accession numberGSE 124607).

DESeq Analysis

Differential Expression analysis was performed using the “DESeq2” package in R. Counts were first generated from Hisat2 outputted single-end bam files in R via “Rsamtools”. These counts were then filtered based on minimum count number (>10). DESeq datasets were created using these filtered counts to compare counts between the treatment and DMSO.

Jaccard Heatmap Analysis

Heatmaps were generated based on gene set similarity scores to show the amount of overlap in genes downregulated by each treatment. Jaccard scores were calculated based on union and intersect of genes downregulated in the two treatments being compared. A table consisting of these scores was imported into python. The “seaborn” and “matplotlib.pyplot” packages were used to generate the heatmap.

Gene Set Enrichment Analysis

Gene set enrichment analysis was performed using the GSEA computational platform developed at the Broad Institute. This analysis required expression data, phenotype data, and a gene set database. Expression data was derived from FPKM data which was used to create a Gene Cluster Text file (.gct). This file is comprised of FPKM averages across the treatment and the control. The phenotype data was input in the form of a Categorical Class file (.cls). This file featured information regarding the number of classes and samples in our analysis as well as classification for each of the classes. Finally, we utilized the C2 all curated gene set directly from the GSEA website for the gene set database in the form of a Gene Matrix Transposed file (.gmt). All other parameters were kept at defaults with the exception of the metric for ranking genes which changed from “Signal2Noise” to “Ratio_of_Classes”.

Radar Plot Generation

Radar plots were created via the “fmsb” package in R to examine differences in gene expression changes across different treatments. This was performed both on a specific cell cycle related gene set derived from GSEA as well as specific gene examples. Log₂ fold change values were used to determine magnitudes of gene expression changes between the treatment and control conditions. Calculation of these values was done in R for both the wildtype and YKL-5-124 refractory cell lines. Following compilation of the log₂ fold change values into a separate data frame, the “fmsb” package was used to create radar plots with overlapping plots for both cell lines.

Single Gene Expression Analysis

Specific genes from the GSEA derived cell cycle gene set were further analyzed for changes in expression across different dosages of YKL-5-124 treatment in both the wildtype and mutant cell lines. To do this, FPKM averages were calculated across the triplicate for the DMSO control and the two dosages of YKL-5-124. These averages were combined with standard deviation values in a single data frame. Using “ggplot” in R, we generated a bar plot for expression of these specific genes across different dosages in both cell lines. Significance levels in changes of expression were calculated via the paired Student’s t-test.

Chemical Synthesis

Synthesis of (R)-3-(4-acrylamidobenzamido)-N-(2-(dimethylamino)-1-phenylethyl)-6,6-dimethyl-4,6-dihydropyrrolo[3,4-c]pyrazole-5(1H)-carboxamide (YKL-05-124)

5-(tert-butyl) 1-ethyl 3-amino-6,6-dimethyl-4,6-dihydropyrrolo[3,4-c]pyrazole-1,5-dicarboxylate (1- and 2-protected ethyl formate was obtained as an inseparable mixture, which was used together until the last step)

To a solution of tert-butyl 3-amino-6,6-dimethyl-4,6-dihydropyrrolo[3,4-c]pyrazole-5(1H)-carboxylate (1.0 g, 3.95 mmol) and TEA (0.6 g, 0.82 mL, 5.93 mmol) in THF (10 mL) was added ethyl chloroformate (0.43 g, 0.38 mL, 3.95 mmol) dropwise at 0 °C. The mixture was stirred at 0 °C for 1 h. The solvent was removed and the residue was partitioned with EtOAc and sat. NaHCO₃. The organic layer was washed with water, brine and dried with Na₂SO₄. The solvent was then removed under vacuum to provide 5-(tert-butyl) 1-ethyl 3-amino-6,6-dimethyl-4,6-dihydropyrrolo[3,4-c]pyrazole-1,5-dicarboxylate as an off white solid (1.28 g, 100 %). LC/MS (ESI) m/z = 325 (M + H)⁺.

5-(tert-butyl) 1-ethyl 6,6-dimethyl-3-(4-nitrobenzamido)-4,6-dihydropyrrolo[3,4-c]pyrazole-1,5-dicarboxylate

To a solution of 5-(tert-butyl) 1-ethyl 3-amino-6,6-dimethyl-4,6-dihydropyrrolo[3,4-c]pyrazole-1,5-dicarboxylate (162 mg, 0.5 mmol) and DIEA (98 μL, 1.0 mmol) in DCM (5 mL) was added 4-nitrobenzoyl chloride (110 mg, 0.6 mmol) at 0 °C. The reaction was stirred at room temperature for 2 h and concentrated. The crude was purified by flash column chromatography on silica gel (EtOAc/hexanes, 0-70%) to provide 5-(tert-butyl) 1-ethyl 6,6-dimethyl-3-(4-nitrobenzamido)-4,6-dihydropyrrolo[3,4-c]pyrazole-1,5-dicarboxylate (177 mg, 75%). LC/MS (ESI) m/z = 474 (M + H)⁺.

1-Ethyl 6,6-dimethyl-3-(4-nitrobenzamido)-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate

To a solution of 5-(tert-butyl) 1-ethyl 6,6-dimethyl-3-(4-nitrobenzamido)-4,6-dihydropyrrolo[3,4-c]pyrazole-1,5-dicarboxylate (177 mg, 0.23 mmol) in DCM (1 mL) was added TFA (1 mL). The reaction was stirred at room temperature for 1 h and concentrated to provide 1-ethyl 6,6-dimethyl-3-(4-nitrobenzamido)-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate as a TFA salt, which was used directly in next step. LC/MS (ESI) m/z = 374 (M + H) ⁺.

1-Ethyl (S)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-3-(4-nitrobenzamido)-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate

To a mixture of ethyl 6,6-dimethyl-3-(4-nitrobenzamido)-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate (TFA salt, 0.23 mmol) in DCM (2 mL) was added DIEA (148 mg, 0.2 mL, 1.15 mmol) at 0 °C, followed by (S)-2-isocyanato-N,N-dimethyl-2-phenylethan-1-amine (62 mg, 0.27 mmol). The solution was stirred at 0 °C for 1 h and diluted with CHCl₃/i-PrOH (v/v = 4:1) and washed with sat. NaHCO₃, brine and dried with Na₂SO₄. The solvent was then removed and the crude was purified by flash column chromatography on silica gel to provide 1-ethyl (S)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-3-(4-nitrobenzamido)-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate (113 mg, 85% over 2 steps). LC/MS (ESI) m/z = 564 (M + H)⁺.

1-Ethyl (S)-3-(4-aminobenzamido)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate

To a solution of 1-ethyl (S)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-3-(4- nitrobenzamido)-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate (113 mg, 0.20 mmol) in EtOAc (2 mL) was added SnCl₂2H₂O (226 mg, 1.0 mmol). The mixture was stirred at 70 °C for 2 h and diluted with CHCl₃/i-PrOH (v/v = 4:1) and washed with sat. NaHCO₃, brine and dried with Na₂SO₄. The crude was purified by flash column chromatography on silica gel to provide 1-ethyl (S)-3-(4-aminobenzamido)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate (193 mg, 88%). LC/MS (ESI) m/z = 534 (M +H)⁺.

(S)-3-(4-acrylamidobenzamido)-N-(2-(dimethylammo)-1-phenylethyl)-6,6-dimethyl-4,6-dihydropyrrolo[3,4-c]pyrazole-5(1H)-carboxamide (YKL-05-124)

To a mixture of 1-ethyl (S)-3-(4-aminobenzamido)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate (20 mg, 0.032 mmol) in THF (1 mL) and sat. NaHCO₃ (1 mL) was added acryloyl choride (5.8 mg, 0.064 mmol) dropwise at 0 °C. The mixture was stirred at 0 °C for 10 min and diluted with CHCl₃/i-PrOH (v/v 4:1) and washed with sat. NaHCO₃ and brine, dried with Na₂SO₄ and then concentrated. The crude was then dissolved in 1.0 mL of i-PrOH. The solution was then treated with LiOH (1 M, 1 mL) and was stirred at room temperature for 10 min. The mixture was diluted with CHCl₃/i-PrOH (v/v 4:1) and washed with sat. NaHCO₃ and brine, dried (Na₂SO₄), and concentrated. The crude was purified by HPLC (MeOH/H₂O, 0-100%) to provide YKL-05-124 as TFA salt (8.1 mg, 40%). LC/MS (ESI) m/z = 516 (M + H)⁺. ¹H NMR (500 MHz, DMSO-d₆) δ 10.89 (s, 1H), 10.72 (s, 1H), 9.49 (s, 1H), 8.03 (d, J = 8.8 Hz, 2H), 7.85 (d, J = 8.8 Hz, 2H), 7.46 (d, J = 7.4 Hz, 2H), 7.40 (dd, J = 8.5, 6.8 Hz, 2H), 7.30 (td, J = 7.0, 1.4 Hz, 1H), 6.79 (d, J = 9.1 Hz, 1H), 6.57 (dd, J = 16.9, 10.2 Hz, 1H), 6.31 (dd, J = 17.0, 1.9 Hz, 1H), 5.81 (dd, J = 10.1, 2.0 Hz, 1H), 5.37 (ddd, J = 12.4, 9.0, 3.8 Hz, 1H), 4.82 (d, J = 11.9 Hz, 1H), 4.56 (d, J = 11.9 Hz, 1H), 3.57 (td, J = 12.5, 2.8 Hz, 1H), 2.87 (d, J = 4.8 Hz, 3H), 2.83 (d, J = 4.8 Hz, 3H), 1.68 (s, 3H), 1.60 (s, 3H).

Synthesis of (S)-N-(2-(dimethylamino)-1-phenylethyl)-6,6-dimethyl-3-(4-propionamidobenzamido)-4,6-dihydropyrrolo[3,4-c]pyrazole-5(1H)-carboxamide (YKL-05-167)

To a mixture of 1-ethyl (S)-3-(4-aminobenzamido)-5-((2-(dimethylamino)-1-phenylethyl)carbamoyl)-6,6-dimethyl-5,6-dihydropyrrolo[3,4-c]pyrazole-1(4H)-carboxylate (20 mg, 0.032 mmol) in THF (1 mL) and sat. NaHCO₃ (1 mL) was added propionyl chloride (5.9 mg, 0.064 mmol) dropwise at 0 °C. The mixture was stirred at 0 °C for 10 min and diluted with CHCl₃/i-PrOH (v/v 4:1) and washed with sat. NaHCO₃ and brine, dried (Na₂SO₄), and concentrated.

The crude was then dissolved in 1.0 mL of i-PrOH. The solution was then treated with LiOH (1 M, 1 mL) and was stirred at room temperature for 10 min. The mixture was diluted with CHCl₃/i-PrOH (v/v 4:1) and washed with sat. NaHCO₃ and brine, dried (Na₂SO₄), and concentrated. The crude was purified by HPLC (MeOH/H₂O, 0-100%) to provide YKL-05-167 as TFA salt (9.0 mg, 45%). LC/MS (ESI) m/z = 518 (M + H)⁺. H NMR (600 MHz, DMSO-d₆) δ 12.38 (s, 1H), 10.71 (s, 1H), 10.12 (s, 1H), 7.95 (m, 2H), 7.69 (m, 2H), 7.35 (d, J = 7.0 Hz, 2H), 7.28 (t, J = 7.6 Hz, 2H), 7.18 (d, J = 7.6 Hz, 1H), 6.25 (m, 1H), 4.87 (m, 1H), 4.52 (m, 2H), 2.68 (m, 1H), 2.41 (m, 1H), 2.34 (q, J = 7.6 Hz, 2H), 2.20 (m, 6H), 1.62 (m, 3H), 1.55 (s, 3H), 1.08 (t, J = 7.6 Hz, 1H).

Quantification and Statistical Analysis

The method of determining error bars is indicated in the corresponding figure legend with the replicate number also indicated. Statistical tests for RNA-seq data is outlined in the methods section under the relevant analysis. Data met the assumptions for all tests used.

Data and Software Availability

The RNA-seq raw data files have been deposited in GEO under the accession number GSE124607. Original data can be found at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE124607.

Supplementary Material

2

Supplemental Table 1. Complete Kinativ Table of Kinases Engaged by YKL-5-124 or YKL-5-167, Related to Figure 2E.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-CDK7	Cell Signaling Technology	Cat#2916
Anti-Cyclin H	Bethyl Labs	Cat#A301-674A
Anti-Cyclin K	Bethyl Labs	Cat#A301-939A
Anti-CDK2	Bethyl Labs	Cat#A301-812
Anti-Phospho-CDK2 T160	Cell Signaling Technology	Cat#2561
Anti-CDK1	Bethyl Labs	Cat#A303-663
Anti-Phospho-CDK1 T161	Cell Signaling Technology	Cat#9114
Anti-Tubulin	Cell Signaling Technology	Cat#3873
Anti-PARP	Cell Signaling Technology	Cat#9542
Anti-Phospho-CTD Ser2	Millipore	Cat#04-1571
Anti-Phospho-CTD Ser5	Millipore	Cat#04-1572
Anti-Phospho-CTD Ser7	Millipore	Cat#04-1570-I
Anti-Total RNA Polymerase II	Santa Cruz Biotechnology	Cat#sc-899
Anti-phospho-histone H3 Alexa 488 S10	Cell Signaling Technology	Cat#3377
Goat IRDye 680RD anti-Rat IgG (H+L)	LI-COR Biosciences	Cat#926-68076
Goat IRDye 680RD anti-Mouse IgG	LI-COR Biosciences	Cat#926-68070
Goat IRDye 800CW anti-Rabbit IgG	LI-COR Biosciences	Cat#926-32211
Chemicals, Peptides, and Recombinant Proteins
Biotinylated-THZ1	Kwiatkowski et al 2014	N/A
YKL-5-124	This paper
YKL-5-167	This paper
THZ1	Kwiatkowski et al 2014	N/A
THZ531	Zhang et al 2016	N/A
CellTiter-Glo Luminescent Cell Viability Assay	Promega	Cat#G7571
M2-Agarose Resin	Sigma	Cat#A2220
cOmplete, Mini Protease Inhibitor Cocktail	Roche	Cat#11836153001
PhosSTOP Phosphotase Inhibitor Tablets	Roche	Cat#04906837001
Mycoalert	Lonza	Cat#LT07-318
CDK7/Cyclin H/ MAT1	Thermo Fisher Scientific	Cat#PR6749B
[ɣ-32P] ATP	Perkin-Elmer	Cat#NEG002A
CDK12/Cyclin K	Bösken et al 2014	N/A
CDK13/Cyclin K	Greifenberg et al 2016	N/A
CDK7/CycH/Mat1	Carna Biosciences	Cat#04-108
CTD3 peptide	Carna Biosciences	Cat#04-108MS
EdU	Lumiprobe	Cat#10540
LIVE/DEAD Far Red Dead Cell Stain	Thermo Fisher Scientific	Cat#L10120
Cy3-azide	Lumiprobe	Cat#11030
RNAPII (RPB1)	abcam	Cat#ab81888
Odyssey Blocking Buffer	LI-COR Biosciences	Cat#927-50100
Hoechst 33342	Sigma Aldrich	Cat#B2261
X-tremeGENE 9	Sigma Aldrich	Cat# 6365779001
Critical Commercial Assays
Selectscreen- Adapta Eu- CDK7/CycH/MNAT1	Invitrogen	N/A
Selectscreen- Adapta Eu- CDK9/CycT1	Invitrogen	N/A
Selectscreen- Z’LYTE- CDK2/CycA	Invitrogen	N/A
Deposited Data
RNA-Seq Data	This paper	Accession Number GSE 124607
Experimental Models: Cell Lines
Human: Jurkat cells (male)	Laboratory of Thomas Looke	N/A
Human: HAP1 cells (male)	Horizon Discovery	C859
Human: Hela cells (female)	Kwiatkowski et al, 2014	N/A
Oligonucleotides
GCCAAGACCAAACTGTCCAG- sgRNA against CDK7	This paper	N/A
CCAAGACCAAACTGTCCAGTGG- WT specific amplification	This paper	N/A
ACCAAACTCACCGGTAGAAAC- mutation specific amplification	This paper	N/A
TGGGTCAGATCTTCCAAATAACT- genomic CDK7 primer	This paper	N/A
Recombinant DNA
pX330	N/A	Addgene: 42230; Cong et al, 2013
pUC57-AMP	Laboratory of Richard Young	N/A
Software and Algorithms
DynaFit software	Kuzmic et al 1996	http://www.biokin.com/dynafit/
GraphPad Prism	GraphPad Software, Inc	http://www.graphpad.com/
mzStudio software	Ficarro et al, 2017	http://github.com/BlaisProteomics/mzStudio
Mascot v 2.2	Perkins et al, 1999	http://www.matrixscience.com
Multiplierz scripts	Askenazi et al, 2009	https://github.com/MaxAlex/multiplierz/wiki
R	R Core Team, 2016	https://www.r-project.org/
Columbus image storage and analysis	PerkinElmer Informatics.	http://www.perkinelmer.com/product/image-data-storage-and-analysis-system-columbus
R scripts for RNAseq Analysis	This Paper	https://github.com/linlabcode/olson_YKL5124

Highlights.

• Development of YKL-5-124 a potent, selective and covalent CDK7 inhibitor

• Selective CDK7 inhibition results in cell cycle inhibition rather than apoptosis

• YKL-5-124 exhibited little effect on RNA Polymerase II phosphorylation status