Targeting Processive Transcription Elongation via SEC Disruption for MYC-Induced Cancer Therapy

Article

Targeting Processive Transcription Elongation via

SEC Disruption for MYC-Induced Cancer Therapy

Graphical Abstract

Authors

Kaiwei Liang, Edwin R. Smith, Yuki Aoi, ...,

Rintaro Hashizume, Gary E. Schiltz,

Ali Shilatifard

Correspondence

ash@northwestern.edu

In Brief

Targeting transcriptional elongation with

small-molecule inhibitors of the super

elongation complex blocks

transcriptional programs driven by the

oncogene MYC

Highlights

d

Discovery of small-molecule inhibitors of SEC and

transcription elongation by Pol II

KL-1 and KL-2 disrupt the cyclin T1-AFF4 interaction

within SEC

SEC inhibitors attenuate SEC-dependent rapid

transcriptional responses

MYC transcriptional programs are inhibited by SEC chemical

disruptors KL-1/KL-2

Liang et al., 2018, Cell 175, 766–779

October 18, 2018 ª 2018 Elsevier Inc.

https://doi.org/10.1016/j.cell.2018.09.027

Article

Targeting Processive Transcription Elongation

via SEC Disruption for MYC-Induced Cancer Therapy

Kaiwei Liang,^1,2,7Edwin R. Smith,^1,2,6Yuki Aoi,^1,2Kristen L. Stoltz,^1,2,3Hiroaki Katagi,⁴Ashley R. Woodﬁn,^1,2

Emily J. Rendleman,^1,2Stacy A. Marshall,^1,2David C. Murray,^1,2Lu Wang,^1,2Patrick A. Ozark,^1,2Rama K. Mishra,^3,5

Rintaro Hashizume,^2,4,6Gary E. Schiltz,^3,5,6and Ali Shilatifard^1,2,6,8,

*

¹Simpson Querrey Center for Epigenetics, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611, USA

²Department of Biochemistry and Molecular Genetics, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611, USA

³Center for Molecular Innovation and Drug Discovery, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208, USA

⁴Department of Neurosurgery, Northwestern University Feinberg School of Medicine, Chicago, IL 60611, USA

⁵Department of Pharmacology, Northwestern University Feinberg School of Medicine, 303 E. Superior Street, Chicago, IL 60611, USA

⁶Robert H. Lurie Comprehensive Cancer Center, Northwestern University Feinberg, School of Medicine, 303 E. Superior Street, Chicago,

IL 60611, USA

⁷Present address: School of Basic Medical Sciences, Wuhan University, Wuhan 430071, China

⁸Lead Contact

*Correspondence: ash@northwestern.edu

https://doi.org/10.1016/j.cell.2018.09.027

SUMMARY

et al., 2010; Bradner et al., 2017; Shilatifard et al., 1996; Smith

et al., 2011; Takahashi et al., 2011).

The super elongation complex (SEC) is required for

robust and productive transcription through release

of RNA polymerase II (Pol II) with its P-TEFb module

and promoting transcriptional processivity with its

ELL2 subunit. Malfunction of SEC contributes to mul-

tiple human diseases including cancer. Here, we

identify peptidomimetic lead compounds, KL-1 and

its structural homolog KL-2, which disrupt the inter-

action between the SEC scaffolding protein AFF4

and P-TEFb, resulting in impaired release of Pol II

from promoter-proximal pause sites and a reduced

The majority of P-TEFb, comprising CDK9 and its cyclin part-

ner Cyclin T1 (CCNT1), is sequestered in an inactive form by RNA

binding proteins HEXIM1 or HEXIM2 associating with 7SK small

nuclear ribonucleoprotein particle (snRNP) (Luo et al., 2012a;

Zhou et al., 2012). The inactive complex occupies promoter-

proximal regions on chromatin (Ji et al., 2013; McNamara

et al., 2016). P-TEFb can be released from the 7SK/HEXIM com-

plex into active complexes such as BRD4/P-TEFb (Yang et al.,

2005) and the super elongation complex (SEC) (Lin et al., 2010;

Luo et al., 2012b; Zhou et al., 2012). Inhibition of all P-TEFb ac-

tivity can be achieved with inhibitors such as ﬂavopiridol (Chao

and Price, 2001). The activity of BRD4/P-TEFb can be inhibited

average rate of processive transcription elongation. with small molecules blocking its recruitment to chromatin (Daw-

son et al., 2011; Filippakopoulos et al., 2010) and phthalimide-

conjugated analogs can induce rapid degradation of BRD4

SEC is required for induction of heat-shock genes

and treating cells with KL-1 and KL-2 attenuates

the heat-shock response from Drosophila to human.

SEC inhibition downregulates MYC and MYC-depen-

dent transcriptional programs in mammalian cells

and delays tumor progression in a mouse xenograft

model of MYC-driven cancer, indicating that small-

molecule disruptors of SEC could be used for tar-

(Winter et al., 2015, 2017). Blocking BRD4/P-TEFb with these

compounds inhibits release of paused Pol II into productive elon-

gation with profound effects on MYC targets (Bradner et al.,

2017; Delmore et al., 2011). However, studies of SEC in cells

are more limited due to a lack of small molecular inhibitors spe-

ciﬁcally targeting this form of P-TEFb.

In addition to P-TEFb, SEC contains AF4/FMR2 (AFF) family

geted therapy of MYC-induced cancer.

proteins (AFF1–4), the YEATS domain-containing proteins ENL

or AF9 (encoded by the MLLT1 and MLLT3 genes), the Pol II elon-

gation factors eleven-nineteen lysine-rich leukemia (ELL) pro-

teins, and ELL-associated factor 1 (EAF1) or EAF2 (Lin et al.,

2010; Chen et al., 2018; Luo et al., 2012b). Within SEC, the AFF

proteins function as scaffolds for binding the other subunits.

P-TEFb is required for phosphorylation of Pol II CTD on serine 2

(Ser2P) andtranscription elongation factor SPT5 on its C-terminal

region to promote release from the promoter-proximal pausing,

while ELL proteins have been demonstrated to enhance proces-

sivity of elongation by RNA Pol II using in vitro transcription as-

says (Shilatifard et al., 1996, 1997) and in cells (Hu et al., 2013).

SEC is required for rapid induction of transcription in response

to cellular signals (Galbraith et al., 2013; Lin et al., 2010, 2011;

INTRODUCTION

In most metazoans, the majority of polymerase II (Pol II)-tran-

scribed genes are regulated at a step called promoter-proximal

pausing (Chen et al., 2018; Jonkers and Lis, 2015). Release from

pausing requires phosphorylation of the Pol II C-terminal domain

(CTD) by cyclin-dependent kinase 9 (CDK9)-containing positive

transcription elongation factor b (P-TEFb) (Peterlin and Price,

2006; Zhou et al., 2012). Proper regulation of this transcriptional

checkpoint is vital for physiological responses in metazoan

development, and misregulation of this checkpoint has been

found to contribute to human diseases, including cancer (Lin

766 Cell 175, 766–779, October 18, 2018 ª 2018 Elsevier Inc.

Figure 1. Peptidomimetic Identiﬁcation of

Disruptors of the AFF4-CCNT1 Interaction

within SEC

A

B

(A) Schematic for identifying small-molecule dis-

ruptors of the AFF4-CCNT1 interaction using

in silico screening. Compounds from the ZINC

database were docked with the AFF4-CCNT1

structure (4IMY) using a three-tier glide-docking

algorithm, which led to 40 candidates. Residues in

the binding pocket of CCNT1 are labeled orange,

and residues of the AFF4 peptide are labeled red.

(B) Validation screening identiﬁed KL-1 as a po-

tential SEC disruptor using AlphaLISA screening.

GST-CCNT1 (aa 1–300) and biotin-labeled AFF4

peptide (aa 32–67) were used to measure the

AFF4-CCNT1 interaction with the AlphaLISA

assay. Data are represented as mean SD.

(C) Similarity search for KL-1-like molecules

identiﬁed KL-2. The peptidomimetic potential of

KL-1 and KL-2 can be seen by overlaying them

with the AFF4 peptide LFAEP structure.

C

D

E

(D) Dose-dependent inhibition of KL-1 and KL-2 on

AFF4-CCNT1 interaction. The K_iconstants of both

compounds were measured with the AlphaLISA

assays. Data are represented as mean SD.

(E) KL-1 and KL-2 treatment results in reduced

protein levels of SEC components AFF1 and AFF4

but not CDK9 or CCNT1. HEK293T cells were

treated with 20 mM of SEC inhibitors KL-1 or KL-2

for 6 hr. 5-, 10-, and 20-mL cell lysates were

loaded, and the protein levels of AFF1, AFF4,

CCNT1, CDK9, and Tubulin (load control) were

determined by western blotting.

G

H

F

(F–H) ChIP-seq analysis demonstrates that KL-1

and KL-2 treatment results in decreased occu-

pancy of AFF1 and AFF4 on chromatin as seen at

the HSPA8 gene (F) or by metaplot analysis at

AFF1 (G) and AFF4 peak regions (H).

See also Figure S6.

Cell 175, 766–779, October 18, 2018 775

(Figure S7B). We assessed the toxicity of KL-1 and KL-2 in mice SEC complex is involved in MYC-dependent transcription

with increasing doses and found that injection of 5 doses of through promotion of transcription elongation rates and these

50 mg/kg KL-1 or 10 mg/kg KL-2 for 5 days does not result in sig- SEC disruptors can be potentially used in vivo.

niﬁcant weight loss in mice after monitoring for 35 days, and no

obvious sign of sickness was observed during this period and increases messenger RNA synthesis, which leads to an

(Figure S7C). increased burden on the core spliceosome to properly process

MYC hyperactivation induces transcriptional ampliﬁcation

To measure the potential of SEC inhibition in the MDA231-LM2 mRNA, suggesting that RNA splicing is a therapeutic vulnera-

tumor model, we initiated injection of mice with SEC inhibitors bility in MYC-driven cancer (Hsu et al., 2015; Lee and Abdel-Wa-

on day 17 after inoculation, when the average tumor size reached hab, 2016). RNA expression proﬁling analysis shows that KL-1

100 mm³(Figure 7A). After once-daily administration for 15 days, and KL-2 treatment leads to a signiﬁcantly decreased output of

we further monitored tumor weights and mice were euthanized the MYC transcriptional program, including RNA splicing-related

when the tumor size reached 1,000 mm³. Both KL-1 and KL-2 genes including the PRMT5 gene, which is a key regulator

delayed tumor progression as monitored by tumor sizes (Figures among the MYC-upregulated genes (Bezzi et al., 2013; Koh

7B–7D). Our study demonstrated that both SEC inhibitors signif- et al., 2015), suggesting that KL-1 and KL-2 could directly target

icantly extended survival of recipient mice (Figure 7E). Together, MYC to lead to an impaired downstream MYC-PRMT5 axis.

these data suggest that SEC disruptors could potentially be used KL-1 and KL-2 share the same scaffold and have similar activ-

to delay progression and improve the survival of MYC-depen- ities toward SEC disruption and Pol II processivity, suggesting

dent cancer.

that this scaffold could function as a lead for future optimization.

These leads already exhibit efﬁcacy in impairing SEC function in

rapid response models and delaying the progression of a MYC-

dependent tumor. Therefore, we anticipate that development of

DISCUSSION

SEC functions as a transcriptional cofactor that is required for small-molecule inhibitors targeting SEC or otherwise slowing

driving high rates of transcription for the classic immediate-early RNA Pol II processivity will be useful both for understanding

genes, the heat shock genes under stress, and for production of the regulation of transcription elongation in cells and as thera-

the HIV provirus, and additionally contributes to oncogenesis by peutic tools for the treatment of human disease including cancer.

driving high rates of coordinated transcriptional programs such

as occurs in MYC-ampliﬁed cancers (Lin et al., 2010; Smith

et al., 2011; Luo et al., 2012b). Treatment of mammalian cells

STAR+METHODS

with KL-1 and KL-2 result in reduced levels of subunits of SEC

AFF1, AFF4, and ELL2. Germline mutations that stabilize AFF4

cause the human developmental disorder CHOPS syndrome

(Izumi et al., 2015). Both AFF4 and ELL2 are targeted by the

SIAH1 E3 ligase (Liu et al., 2012). The SEC destabilizing property

of KL-1 and KL-2 likely enhances the efﬁcacy of our lead com-

pounds in the transcription elongation assays and in vivo animal

tumor model. Proteolysis-targeting chimera (PROTAC) methods

that allow targeted degradation of proteins with small molecules

have been shown to be much more efﬁcacious than the small-

molecule inhibitor alone (Neklesa et al., 2017; Winter et al.,

2015, 2017). For example, JQ1 and related compound IBET-

151 bind to the bromodomains of BRD4 and block its interac-

tions with acetylated histones on chromatin (Dawson et al.,

2011; Filippakopoulos et al., 2010). When JQ1-like molecules

are fused to phthalimides to target BRD4 degradation by the

endogenous cellular ubiquitin ligase cereblon, the loss of

BRD4 protein obviates the need for constant interaction of JQ1

with BRD4 (Lu et al., 2015; Winter et al., 2015, 2017).

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d CONTACT FOR REAGENT AND RESOURCE SHARING

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

B Cell Lines

B Plasmids, Peptides and Chemicals

B MDA231-LM2 Tumor Model

d METHOD DETAILS

B Chemical Synthesis

B In silico High-throughput Screening

B AlphaLISA Assay

B Heat Shock Induction

B Induction of J-Lat 6.3 Cells

B Chromatin Immunoprecipitation Sequencing

B Precision Nuclear Run-on and Sequencing

B 4sU-FP-seq

B RNA-seq Analysis

d QUANTIFICATION AND STATISTICAL ANALYSIS

d DATA AND SOFTWARE AVAILABILITY

In this study, we found that SEC disruptors KL-1 and KL-2

could block transcription elongation in multiple SEC-dependent

transcriptional models, demonstrating these compounds can be

used a convenient chemical perturbation tool for the mechanistic

and functional studies of SEC in other SEC-related cellular and

developmental processes. Interestingly, we found that the

MYC-dependent cancer cells are sensitive to the SEC inhibitors,

providing the mechanistic ﬁnding that SEC is co-localized with

MYC and exhibit increased occupancy in the MYC highly ex-

pressing cells, suggesting a dependency of transcription elonga-

tion for MYC-dependent cancers. Indeed, we also showed that

SUPPLEMENTAL INFORMATION

Supplemental Information includes seven ﬁgures and can be found with this

article online at https://doi.org/10.1016/j.cell.2018.09.027.

ACKNOWLEDGMENTS

We thank Dr. David Bentley for the slow and fast Pol II mutant cells and Dr. Yib-

ing Kang for the MDA231-LM2 cells. We thank Dr. Lihua Zou for help with the

776 Cell 175, 766–779, October 18, 2018

Hidden Markov analysis. The following reagent was obtained through the NIH

AIDS Reagent Program, Division of AIDS, NIAID, NIH: J-Lat full-length clone

(6.3) from Dr. Eric Verdin. A part of this work was performed by the North-

western University ChemCore, which is funded by Cancer Center Support

Grant P30CA060553 from the National Cancer Institute awarded to the Robert

H. Lurie Comprehensive Cancer Center, and the Chicago Biomedical Con-

sortium with support from the Searle Funds at the Chicago Community Trust.

L.W. was supported by the Training Program in Signal Transduction and

Cancer (T32CA070085). Y.A. was supported by a JSPS Research Fellowship

for Young Scientists. E.R.S. was supported by the National Cancer Institute

grant R50CA211428. Studies on transcription elongation control in the Shilati-

fard laboratory were supported by the National Cancer Institute grant

R01CA214035 to A.S.

domain inhibition as

904–917.

a therapeutic strategy to target c-Myc. Cell 146 ,

Erb, M.A., Scott, T.G., Li, B.E., Xie, H., Paulk, J., Seo, H.S., Souza, A., Roberts,

J.M., Dastjerdi, S., Buckley, D.L., et al. (2017). Transcription control by the ENL

YEATS domain in acute leukaemia. Nature 543 , 270–274.

Filippakopoulos, P., Qi, J., Picaud, S., Shen, Y., Smith, W.B., Fedorov, O.,

Morse, E.M., Keates, T., Hickman, T.T., Felletar, I., et al. (2010). Selective inhi-

bition of BET bromodomains. Nature 468 , 1067–1073.

Fong, N., Kim, H., Zhou, Y., Ji, X., Qiu, J., Saldi, T., Diener, K., Jones, K., Fu,

X.D., and Bentley, D.L. (2014). Pre-mRNA splicing is facilitated by an optimal

RNA polymerase II elongation rate. Genes Dev. 28 , 2663–2676.

Fong, N., Brannan, K., Erickson, B., Kim, H., Cortazar, M.A., Sheridan, R.M.,

Nguyen, T., Karp, S., and Bentley, D.L. (2015). Effects of transcription elonga-

tion rate and Xrn2 Exonuclease activity on RNA polymerase II termination sug-

gest widespread kinetic competition. Mol. Cell 60 , 256–267.

AUTHOR CONTRIBUTIONS

K.L. and A.S. conceived and designed the experiments. K.L. conducted most

of the cell and biochemical experiments. G.E.S. and R.K.M. performed the vir-

tual high-throughput screening. K.L.S. synthesized the compounds KL-1 and

KL-2. G.E.S. supervised the chemical aspects of these studies. Y.A. per-

formed the PRO-seq experiments, and D.C.M. performed western blotting ex-

periments. E.J.R. performed the ﬂow cytometry analysis. H.K., L.W., and R.H.

performed the MDA231-LM2 mice studies. S.A.M. and E.J.R. performed NGS

sequencing. K.L. and A.R.W. performed bioinformatics analyses, except

P.A.O. performed the HMM analyses. K.L., E.R.S., and A.S. analyzed and in-

terpreted results and wrote the manuscript with input from all authors.

Fong, N., Saldi, T., Sheridan, R.M., Cortazar, M.A., and Bentley, D.L.

(2017). RNA Pol II dynamics modulate co-transcriptional chromatin modiﬁ-

cation, CTD phosphorylation, and transcriptional direction. Mol. Cell 66 ,

546–557.

Fuchs, G., Voichek, Y., Benjamin, S., Gilad, S., Amit, I., and Oren, M. (2014).

4sUDRB-seq: Measuring genomewide transcriptional elongation rates and

initiation frequencies within cells. Genome Biol. 15 , R69.

Galbraith, M.D., Allen, M.A., Bensard, C.L., Wang, X., Schwinn, M.K., Qin, B.,

Long, H.W., Daniels, D.L., Hahn, W.C., Dowell, R.D., and Espinosa, J.M.

(2013). HIF1A employs CDK8-mediator to stimulate RNAPII elongation in

response to hypoxia. Cell 153 , 1327–1339.

DECLARATION OF INTERESTS

Gu, J., Babayeva, N.D., Suwa, Y., Baranovskiy, A.G., Price, D.H., and Tahirov,

T.H. (2014). Crystal structure of HIV-1 Tat complexed with human P-TEFb and

AFF4. Cell Cycle 13 , 1788–1797.

The authors declare no competing interests.

He, N., Liu, M., Hsu, J., Xue, Y., Chou, S., Burlingame, A., Krogan, N.J., Alber,

T., and Zhou, Q. (2010). HIV-1 Tat and host AFF4 recruit two transcription elon-

gation factors into a bifunctional complex for coordinated activation of HIV-1

transcription. Mol. Cell 38 , 428–438.

Received: April 2, 2018

Revised: July 2, 2018

Accepted: September 13, 2018

Published: October 18, 2018

Hsu, T.Y., Simon, L.M., Neill, N.J., Marcotte, R., Sayad, A., Bland, C.S.,

Echeverria, G.V., Sun, T., Kurley, S.J., Tyagi, S., et al. (2015). The spliceo-

REFERENCES

some is

384–388.

a therapeutic vulnerability in MYC-driven cancer. Nature 525 ,

Baell, J.B., and Holloway, G.A. (2010). New substructure ﬁlters for removal of

pan assay interference compounds (PAINS) from screening libraries and for

their exclusion in bioassays. J. Med. Chem. 53 , 2719–2740.

Hu, D., Smith, E.R., Garruss, A.S., Mohaghegh, N., Varberg, J.M., Lin, C.,

Jackson, J., Gao, X., Saraf, A., Florens, L., et al. (2013). The little elongation

complex functions at initiation and elongation phases of snRNA gene tran-

scription. Mol. Cell 51 , 493–505.

Bezzi, M., Teo, S.X., Muller, J., Mok, W.C., Sahu, S.K., Vardy, L.A., Bonday,

Z.Q., and Guccione, E. (2013). Regulation of constitutive and alternative

splicing by PRMT5 reveals a role for Mdm4 pre-mRNA in sensing defects in

the spliceosomal machinery. Genes Dev. 27 , 1903–1916.

Irwin, J.J., Sterling, T., Mysinger, M.M., Bolstad, E.S., and Coleman, R.G.

(2012). ZINC: A free tool to discover chemistry for biology. J. Chem. Inf. Model.

52 , 1757–1768.

Bolger, A.M., Lohse, M., and Usadel, B. (2014). Trimmomatic: A ﬂexible

trimmer for Illumina sequence data. Bioinformatics 30 , 2114–2120.

Izumi, K., Nakato, R., Zhang, Z., Edmondson, A.C., Noon, S., Dulik, M.C., Ra-

jagopalan, R., Venditti, C.P., Gripp, K., Samanich, J., et al. (2015). Germline

gain-of-function mutations in AFF4 cause a developmental syndrome func-

tionally linking the super elongation complex and cohesin. Nat. Genet. 47 ,

338–344.

Bradner, J.E., Hnisz, D., and Young, R.A. (2017). Transcriptional addiction in

cancer. Cell 168 , 629–643.

Chao, S.H., and Price, D.H. (2001). Flavopiridol inactivates P-TEFb and blocks

most RNA polymerase II transcription in vivo. J. Biol. Chem. 276 ,

31793–31799.

Ji, X., Zhou, Y., Pandit, S., Huang, J., Li, H., Lin, C.Y., Xiao, R., Burge, C.B., and

Fu, X.D. (2013). SR proteins collaborate with 7SK and promoter-associated

nascent RNA to release paused polymerase. Cell 153 , 855–868.

Chen, F.X., Smith, E.R., and Shilatifard, A. (2018). Born to run: Control of tran-

scription elongation by RNA polymerase II. Nat. Rev. Mol. Cell Biol. 19 ,

464–478.

Jonkers, I., and Lis, J.T. (2015). Getting up to speed with transcription elonga-

tion by RNA polymerase II. Nat. Rev. Mol. Cell Biol. 16 , 167–177.

Danko, C.G., Hah, N., Luo, X., Martins, A.L., Core, L., Lis, J.T., Siepel, A., and

Kraus, W.L. (2013). Signaling pathways differentially affect RNA polymerase II

initiation, pausing, and elongation rate in cells. Mol. Cell 50 , 212–222.

Jordan, A., Bisgrove, D., and Verdin, E. (2003). HIV reproducibly establishes

a latent infection after acute infection of T cells in vitro. EMBO J. 22 ,

1868–1877.

Dawson, M.A., Prinjha, R.K., Dittmann, A., Giotopoulos, G., Bantscheff, M.,

Chan, W.I., Robson, S.C., Chung, C.W., Hopf, C., Savitski, M.M., et al.

(2011). Inhibition of BET recruitment to chromatin as an effective treatment

for MLL-fusion leukaemia. Nature 478 , 529–533.

Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., and Salzberg, S.L.

(2013). TopHat2: Accurate alignment of transcriptomes in the presence of in-

sertions, deletions and gene fusions. Genome Biol. 14 , R36.

Delmore, J.E., Issa, G.C., Lemieux, M.E., Rahl, P.B., Shi, J., Jacobs, H.M.,

Kastritis, E., Gilpatrick, T., Paranal, R.M., Qi, J., et al. (2011). BET bromo-

Koh, C.M., Bezzi, M., Low, D.H., Ang, W.X., Teo, S.X., Gay, F.P., Al-Haddawi,

`

M., Tan, S.Y., Osato, M., Sabo, A., et al. (2015). MYC regulates the core

Cell 175, 766–779, October 18, 2018 777

pre-mRNA splicing machinery as an essential step in lymphomagenesis.

Nie, Z., Hu, G., Wei, G., Cui, K., Yamane, A., Resch, W., Wang, R., Green, D.R.,

Tessarollo, L., Casellas, R., et al. (2012). c-Myc is a universal ampliﬁer of ex-

pressed genes in lymphocytes and embryonic stem cells. Cell 151 , 68–79.

Nature 523 , 96–100.

Kwak, H., Fuda, N.J., Core, L.J., and Lis, J.T. (2013). Precise maps of RNA po-

lymerase reveal how promoters direct initiation and pausing. Science 339 ,

950–953.

Peterlin, B.M., and Price, D.H. (2006). Controlling the elongation phase of tran-

scription with P-TEFb. Mol. Cell 23 , 297–305.

Quinlan, A.R., and Hall, I.M. (2010). BEDTools: A ﬂexible suite of utilities for

Langmead, B., Trapnell, C., Pop, M., and Salzberg, S.L. (2009). Ultrafast and

memory-efﬁcient alignment of short DNA sequences to the human genome.

Genome Biol. 10 , R25.

comparing genomic features. Bioinformatics 26 , 841–842.

Rahl, P.B., Lin, C.Y., Seila, A.C., Flynn, R.A., McCuine, S., Burge, C.B., Sharp,

P.A., and Young, R.A. (2010). c-Myc regulates transcriptional pause release.

Cell 141 , 432–445.

Lee, S.C., and Abdel-Wahab, O. (2016). Therapeutic targeting of splicing in

cancer. Nat. Med. 22 , 976–986.

Robinson, M.D., McCarthy, D.J., and Smyth, G.K. (2010). edgeR: A Bio-

conductor package for differential expression analysis of digital gene expres-

sion data. Bioinformatics 26 , 139–140.

Liang, K., Woodﬁn, A.R., Slaughter, B.D., Unruh, J.R., Box, A.C., Rickels, R.A.,

Gao, X., Haug, J.S., Jaspersen, S.L., and Shilatifard, A. (2015). Mitotic tran-

scriptional activation: Clearance of actively engaged Pol II via transcriptional

elongation control in mitosis. Mol. Cell 60 , 435–445.

Sab o ` , A., Kress, T.R., Pelizzola, M., de Pretis, S., Gorski, M.M., Tesi, A.,

Morelli, M.J., Bora, P., Doni, M., Verrecchia, A., et al. (2014). Selective tran-

scriptional regulation by Myc in cellular growth control and lymphomagenesis.

Nature 511 , 488–492.

Liang, K., Volk, A.G., Haug, J.S., Marshall, S.A., Woodﬁn, A.R., Bartom, E.T.,

Gilmore, J.M., Florens, L., Washburn, M.P., Sullivan, K.D., et al. (2017). Ther-

apeutic targeting of MLL degradation pathways in MLL-rearranged leukemia.

Cell 168 , 59–72.

Saldanha, A.J. (2004). Java Treeview—extensible visualization of microarray

data. Bioinformatics 20 , 3246–3248.

Lin, C., Smith, E.R., Takahashi, H., Lai, K.C., Martin-Brown, S., Florens, L.,

Washburn, M.P., Conaway, J.W., Conaway, R.C., and Shilatifard, A. (2010).

AFF4, a component of the ELL/P-TEFb elongation complex and a shared sub-

unit of MLL chimeras, can link transcription elongation to leukemia. Mol. Cell

37 , 429–437.

Schulze-Gahmen, U., Upton, H., Birnberg, A., Bao, K., Chou, S., Krogan, N.J.,

Zhou, Q., and Alber, T. (2013). The AFF4 scaffold binds human P-TEFb adja-

cent to HIV Tat. eLife 2 , e00327.

Shilatifard, A., Lane, W.S., Jackson, K.W., Conaway, R.C., and Conaway, J.W.

(1996). An RNA polymerase II elongation factor encoded by the human ELL

gene. Science 271 , 1873–1876.

Lin, C., Garrett, A.S., De Kumar, B., Smith, E.R., Gogol, M., Seidel, C., Krum-

lauf, R., and Shilatifard, A. (2011). Dynamic transcriptional events in embryonic

stem cells mediated by the super elongation complex (SEC). Genes Dev. 25 ,

1486–1498.

Shilatifard, A., Duan, D.R., Haque, D., Florence, C., Schubach, W.H., Con-

away, J.W., and Conaway, R.C. (1997). ELL2, a new member of an ELL family

of RNA polymerase II elongation factors. Proc. Natl. Acad. Sci. USA 94 ,

3639–3643.

Lin, C.Y., Lov e ´ n, J., Rahl, P.B., Paranal, R.M., Burge, C.B., Bradner, J.E., Lee,

T.I., and Young, R.A. (2012). Transcriptional ampliﬁcation in tumor cells with

elevated c-Myc. Cell 151 , 56–67.

Smith, E., Lin, C., and Shilatifard, A. (2011). The super elongation complex

(SEC) and MLL in development and disease. Genes Dev. 25 , 661–672.

Liu, M., Hsu, J., Chan, C., Li, Z., and Zhou, Q. (2012). The ubiquitin ligase Siah1

controls ELL2 stability and formation of super elongation complexes to modu-

late gene transcription. Mol. Cell 46 , 325–334.

Sobhian, B., Laguette, N., Yatim, A., Nakamura, M., Levy, Y., Kiernan, R., and

Benkirane, M. (2010). HIV-1 Tat assembles a multifunctional transcription elon-

gation complex and stably associates with the 7SK snRNP. Mol. Cell 38 ,

439–451.

Lu, J., Qian, Y., Altieri, M., Dong, H., Wang, J., Raina, K., Hines, J., Winkler,

J.D., Crew, A.P., Coleman, K., and Crews, C.M. (2015). Hijacking the E3

ubiquitin ligase cereblon to efﬁciently target BRD4. Chem. Biol. 22 ,

755–763.

Takahashi, H., Parmely, T.J., Sato, S., Tomomori-Sato, C., Banks, C.A., Kong,

S.E., Szutorisz, H., Swanson, S.K., Martin-Brown, S., Washburn, M.P., et al.

(2011). Human mediator subunit MED26 functions as a docking site for tran-

scription elongation factors. Cell 146 , 92–104.

Luo, Z., Lin, C., Guest, E., Garrett, A.S., Mohaghegh, N., Swanson, S.,

Marshall, S., Florens, L., Washburn, M.P., and Shilatifard, A. (2012a). The

super elongation complex family of RNA polymerase II elongation factors:

Gene target speciﬁcity and transcriptional output. Mol. Cell. Biol. 32 ,

2608–2617.

Tripathi, S., Pohl, M.O., Zhou, Y., Rodriguez-Frandsen, A., Wang, G., Stein,

D.A., Moulton, H.M., DeJesus, P., Che, J., Mulder, L.C., et al. (2015). Meta-

and orthogonal integration of inﬂuenza ‘‘OMICs’’ data deﬁnes a role for

UBR4 in virus budding. Cell Host Microbe 18 , 723–735.

Luo, Z., Lin, C., and Shilatifard, A. (2012b). The super elongation complex

Walz, S., Lorenzin, F., Morton, J., Wiese, K.E., von Eyss, B., Herold, S., Rycak,

L., Dumay-Odelot, H., Karim, S., Bartkuhn, M., et al. (2014). Activation and

repression by oncogenic MYC shape tumour-speciﬁc gene expression pro-

ﬁles. Nature 511 , 483–487.

(SEC) family in transcriptional control. Nat. Rev. Mol. Cell Biol. 13 , 543–547.

Mahat, D.B., Kwak, H., Booth, G.T., Jonkers, I.H., Danko, C.G., Patel, R.K.,

Waters, C.T., Munson, K., Core, L.J., and Lis, J.T. (2016a). Base-pair-resolu-

tion genome-wide mapping of active RNA polymerases using precision nu-

clear run-on (PRO-seq). Nat. Protoc. 11 , 1455–1476.

Wan, L., Wen, H., Li, Y., Lyu, J., Xi, Y., Hoshii, T., Joseph, J.K., Wang, X., Loh,

Y.E., Erb, M.A., et al. (2017). ENL links histone acetylation to oncogenic gene

expression in acute myeloid leukaemia. Nature 543 , 265–269.

Mahat, D.B., Salamanca, H.H., Duarte, F.M., Danko, C.G., and Lis, J.T.

(2016b). Mammalian heat shock response and mechanisms underlying its

genome-wide transcriptional regulation. Mol. Cell 62 , 63–78.

Wang, L., Collings, C.K., Zhao, Z., Cozzolino, K.A., Ma, Q., Liang, K.,

Marshall, S.A., Sze, C.C., Hashizume, R., Savas, J.N., and Shilatifard, A.

(2017). A cytoplasmic COMPASS is necessary for cell survival and triple-

negative breast cancer pathogenesis by regulating metabolism. Genes

Dev. 31 , 2056–2066.

Martin, M. (2011). Cutadapt removes adapter sequences from high-

throughput sequencing reads. EMBnetjournal 17 , 10–12.

McNamara, R.P., Bacon, C.W., and D’Orso, I. (2016). Transcription elongation

control by the 7SK snRNP complex: Releasing the pause. Cell Cycle 15 ,

2115–2123.

Winter, G.E., Buckley, D.L., Paulk, J., Roberts, J.M., Souza, A., Dhe-Paganon,

S., and Bradner, J.E. (2015). Drug Development. Phthalimide conjugation as a

strategy for in vivo target protein degradation. Science 348 , 1376–1381.

Mohan, M., Lin, C., Guest, E., and Shilatifard, A. (2010). Licensed to elongate:

A molecular mechanism for MLL-based leukaemogenesis. Nat. Rev. Cancer

10 , 721–728.

Winter, G.E., Mayer, A., Buckley, D.L., Erb, M.A., Roderick, J.E., Vittori, S.,

Reyes, J.M., di Iulio, J., Souza, A., Ott, C.J., et al. (2017). BET bromodomain

proteins function as master transcription elongation factors independent of

CDK9 recruitment. Mol Cell 67 , 5–18.

Neklesa, T.K., Winkler, J.D., and Crews, C.M. (2017). Targeted protein degra-

dation by PROTACs. Pharmacol. Ther. 174 , 138–144.

778 Cell 175, 766–779, October 18, 2018

Yang, Z., Yik, J.H., Chen, R., He, N., Jang, M.K., Ozato, K., and Zhou, Q.

(2005). Recruitment of P-TEFb for stimulation of transcriptional elongation

by the bromodomain protein Brd4. Mol. Cell 19 , 535–545.

Zeller, K.I., Jegga, A.G., Aronow, B.J., O’Donnell, K.A., and Dang, C.V.

(2003). An integrated database of genes responsive to the Myc oncogenic

transcription factor: Identiﬁcation of direct genomic targets. Genome Biol.

4 , R69.

Yokoyama, A., Lin, M., Naresh, A., Kitabayashi, I., and Cleary, M.L. (2010). A

higher-order complex containing AF4 and ENL family proteins with P-TEFb fa-

cilitates oncogenic and physiologic MLL-dependent transcription. Cancer Cell

17 , 198–212.

Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E.,

Nusbaum, C., Myers, R.M., Brown, M., Li, W., and Liu, X.S. (2008). Model-

based analysis of ChIP-Seq (MACS). Genome Biol. 9 , R137.

Zang, C., Schones, D.E., Zeng, C., Cui, K., Zhao, K., and Peng, W. (2009).

A clustering approach for identiﬁcation of enriched domains from histone

modiﬁcation ChIP-Seq data. Bioinformatics 25 , 1952–1958.

Zhou, Q., Li, T., and Price, D.H. (2012). RNA polymerase II elongation control.

Annu. Rev. Biochem. 81 , 119–143.

Cell 175, 766–779, October 18, 2018 779

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies

Rabbit anti-AFF1 Antibody, Afﬁnity Puriﬁed

Rabbit anti-ELL2 Antibody, Afﬁnity Puriﬁed

Pol II Rpb1 NTD (D8L4Y) Rabbit mAb

Bethyl Laboratories

Bethyl Cat# A302-344A; RRID:AB_1850255

Bethyl Cat# A302-505A; RRID:AB_1966087

Bethyl Laboratories

Cell Signaling Technology

Cell Signaling Technology Cat# 14958;

RRID:AB_2687876

BRD2 (D89B4) Rabbit mAb

BRD4 (E2A7X) Rabbit mAb

Cell Signaling Technology

Cell Signaling Technology Cat# 5848;

RRID: AB_10835146

Cell Signaling Technology Cat# 13440;

RRID: AB_2687578

MED26 (D4B1X) Rabbit mAb

c-Myc (D3N8F) Rabbit mAb

Cell Cell Signaling Technology

Cell Signaling Technology

Cell Signaling Technology Cat# 14950

Cell Signaling Technology Cat# 13987;

RRID:AB_2631168

Tubulin beta antibody

AFF4 Antibody

DSHB

DSHB Cat# E7; RRID:AB_528499

Proteintech

Proteintech Group Cat# 14662-1-AP;

RRID:AB_2242609

Cyclin T1 Antibody (H-245)

HSP 90a/b Antibody

Santa Cruz

Active Motif

Santa Cruz Biotechnology Cat# sc-10750;

RRID:AB_2073888

Santa Cruz Biotechnology Cat# sc-7947;

RRID:AB_2121235

CDK9 Antibody (C-20)

Santa Cruz Biotechnology Cat# sc-484;

RRID:AB_2275986

Anti-RNA polymerase II subunit B1 (phospho

CTD Ser-2) Antibody, clone 3E10

Active Motif Cat# 61083; RRID:AB_2687450

FLAG-synthetic antibody

ANTI-FLAG M2 Afﬁnity Gel

Anti-SPT5 Antibody, clone 6F1

Rabbit anti-AFF1 serum

Sigma-Aldrich

Millipore

Sigma-Aldrich Cat# F3165; RRID:AB_259529

Sigma-Aldrich Cat# A2220; RRID:AB_10063035

Millipore Cat# MABE1803

Lin et al., 2011

Lin et al., 2010

N/A

Rabbit anti-ELL2 serum

Rabbit anti-AFF4 serum

Rabbit anti-Drosophila Rpb1 antibody

Chemicals, Peptides, and Recombinant Proteins

Flavopiridol

Cayman

Cat# 10009197

Cat# T4509

N/A

4-Thiouridine

Sigma-Aldrich

Screening compounds

ChemDiv, ChemBridge and

Enamine

EZ-Link HPDP-Biotin

KL-1

Thermo

Cat# 21341

This study

VCPBIO

N/A

KL-2

N/A

Biotin-AFF4 (32-67) peptides

Biotin-AFF4 mutant peptides

Glycogen

N/A

VCPBIO

N/A

Sigma-Aldrich

This study

Santa Cruz

Perkin Elmer

Cat# 10901393001

N/A

GST-CCNT1 (1-300)

a-Amanitin

CAS 23109-05-9

Cat# NEL544001EA

Cat# NEL542001EA

(Continued on next page)

Biotin-11-ATP

Biotin-11-CTP

e1 Cell 175, 766–779.e1–e9, October 18, 2018

Continued

REAGENT or RESOURCE

Biotin-11-GTP

SOURCE

IDENTIFIER

Perkin Elmer

NEB

Cat# NEL545001EA

Cat# NEL543001EA

Cat# M0356S

Biotin-11-UTP

RNA 5⁰Pyrophosphohydrolase (RppH)

2% Agarose, PippinHT, 100-600 bp.10/pkg.

BD Matrigel Matrix

Sage science

BD Biosciences

Cat# HTC2010

Cat# 354234

Critical Commercial Assays

TruSeq Stranded Total RNA LT - (with Ribo-

ZeroTM Human/Mouse/ Rat) - Set A

Illumina

RS-123-2201

RS-123-2202

KK8234

TruSeq Stranded Total RNA LT - (with Ribo-

ZeroTM TM Human/ Mouse/Rat) - Set B

Illumina

High-Throughput Library Preparation Kit

Standard PCR Amp Module - 96 rxn

KAPA Biosystems

RNeasy Mini Kit

QIAGEN

Cat# 74104

Vi-CELL Reagent Quad Pak

Beckman Coulter

Perkin Elmer

Cat# 383198

Cat# 6760603C

AlphaScreen GST Detection Kit, 500 assay

points

AlphaScreen Streptavidin Donor beads

RNeasy MinElute Cleanup Kit

Glutathione Superﬂow Agarose

Dynabeads MyOne Streptavidin C1

Streptavidin M280 beads

Dynabeads Protein G

Perkin Elmer

QIAGEN

Cat# 6760002S

Cat# 74204

Thermo

Cat# 25236

Thermo

Cat# 65001

Invitrogen

VWR

Cat# 11206D

Cat# 10003D

Cat# 10847-802

Cat# 15596-018

Cat# HT90132-1L

Cat# V13245

Phase Lock Gel Heavy

Trizol Reagent

Invitrogen

Sigma-Aldrich

Thermo

Crystal violet solution

Dead Cell Apoptosis Kit with Annexin V Alexa

Fluor 488 & Propidium Iodide (PI)

CellTiter-Glo 2.0 Assay

Promega

Cat# G9242

Deposited Data

Raw and analyzed data

This paper

GEO: GSE112608

Human reference genome GRCh37/hg19

Genome Reference Consortium

https://www.ncbi.nlm.nih.gov/assembly/

GCF_000001405.13/

Drosophila reference genome BDGP

Genome Reference Consortium

http://www.fruitﬂy.org/sequence/

release3genomic.shtml

Release 5/dm3

Experimental Models: Cell Lines

HCT-116, human, male

ATCC

ATCC Cat# CCL-247, RRID:CVCL_0291

NIH-ARP Cat# 9846-446, RRID:CVCL_8280

ATCC Cat# CRL-3216, RRID:CVCL_0063

N/A

J-Lat 6.3 clone (Jurkat cells), human, male

HEK293T, human, male

NIH AIDS Program

ATCC

HEK293T Flag-AFF1

Lin et al., 2010

ATCC

HEK293T Flag-AFF4

N/A

NCI-H2171 [H2171] (ATCC CRL-5929),

human, male

ATCC Cat# CRL-5929, RRID:CVCL_1536

SW 1271 [SW1271] (ATCC CRL-2177),

human, male

ATCC

ATCC Cat# CRL-2177, RRID:CVCL_1716

N/A

FAST Pol II Mutant E1126G HEK293 cells,

human, female

Fong et al., 2014

WT Pol II Mutant HEK293 cells, human, female

Fong et al., 2014

N/A

Slow Pol II Mutant R749H HEK293 cells,

human, female

(Continued on next page)

Cell 175, 766–779.e1–e9, October 18, 2018 e2

Continued

REAGENT or RESOURCE

MDA231-LM2, human, female

S2 cells, Drosophila, male

SOURCE

IDENTIFIER

N/A

Wang et al., 2017

Invitrogen

Cat# R690-07

Experimental Models: Organisms/Strains

athymic nude mice, nu/nu

Envigo

Hsd:Athymic Nude-Foxn1^nu

Oligonucleotides

Recombinant DNA

shRNA targeting sequence: AFF1 #1

GCCTCAAGTGAAGTTTGACAA

Sigma-Aldrich

This paper

Clone ID: TRCN0000021975

Clone ID: TRCN0000330908

Clone ID: TRCN0000426769

Clone ID: TRCN0000015825

N/A

shRNA targeting sequence: AFF1 #2

TAGGTTGGGAAAGCCGAAATA

shRNA targeting sequence: AFF4 #1

GCACGACCGTGAGTCATATAA

shRNA targeting sequence: AFF4 #2

GCACCAGTCTAAATCTATGTT

shRNA targeting sequence: ELL2 #1

AACGCCAGAATTATAAGGATG

shRNA targeting sequence: ELL2 #2

AAATGATCCCCTCAATGAAGT

This paper

N/A

Lenti-sh1368 knockdown c-myc shRNA

targeting sequence: GACGAGAACAGT

TGAAACA

Addgene

Addgene plasmid # 29435

pGEX-2TK cyclin T1 (1-300)

Addgene

Addgene plasmid # P432

Software and Algorithms

TopHat 2.1.0

Kim et al., 2013

https://ccb.jhu.edu/software/tophat/index.shtml

http://liulab.dfci.harvard.edu/MACS/

MACS 1.4.2

Zhang et al., 2008

Robinson et al., 2010

EdgeR 3.12.0

http://bioconductor.statistik.tu-dortmund.

de/packages/2.11/bioc/html/edgeR.html

Bowtie version 1.1.2

Trimmomatic 0.33

Langmead et al., 2009

Bolger et al., 2014

http://bowtie-bio.sourceforge.net/index.shtml

http://www.usadellab.org/cms/index.php?

page=trimmomatic

Cutadapt 1.14

Bedtools 2.17

Martin, 2011

https://cutadapt.readthedocs.io/en/stable/

guide.html

Quinlan and Hall, 2010

https://launchpad.net/ubuntu/+source/

bedtools/2.17.0-1

Metascape

Tripathi et al., 2015

N/A

http://metascape.org/gp/index.html

https://www.r-project.org/

http://zinc.docking.org

R 3.3.3

ZINC database

Irwin et al., 2012

¨

Schrodinger

Small-Molecule Drug Discovery Suite 2017-2

https://www.schrodinger.com/suites/

small-molecule-drug-discovery-suite

Prism 7

FlowJo

GraphPad Software

FlowJo

https://www.graphpad.com

https://www.ﬂowjo.com

`

Dassault Systemes BIOVIA

`

Dassault Systemes

https://www.3ds.com/products-services/biovia/

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for reagents may be directed to and will be fulﬁlled by the Lead Contact, Ali Shilatifard (ash@

northwestern.edu).

e3 Cell 175, 766–779.e1–e9, October 18, 2018

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell Lines

HEK293T (female cell line, ATCC CRL-3216), HCT-116 (male cell line, ATCC CCL-247), MDA231-LM2 (female cell line), Flag-AFF1

and Flag-AFF4 HEK293T stable cell lines were cultured in Dulbecco’s Modiﬁed Eagle Medium (DMEM) supplemented with 10% fetal

bovine serum (FBS, catalog No. F6178, Sigma). NCI-H2171 [H2171] (male cell line, ATCC CRL-5929) and SW1271 (male cell line,

ATCC CRL-2177) small cell lung cancer cells were maintained in RPMI-1640 and DMEM/F12 medium supplemented with 20%

FBS. The Jurkat J-Lat full-length cells (6.3) (male cell line, NIH-ARP Cat# 9846-446) were provided by the NIH AIDS Reagent Program

and cultured in RPMI-1640 medium with 10% FBS. Drosophila melanogaster S2 cells, a male cell line, were maintained in Schneider’s

medium. The wild-type Pol II, slow Pol II (R749H) and fast Pol II (E1126G) mutant HEK293T cell lines (Fong et al., 2014) were provided

by Dr. David Bentley (University of Colorado School of Medicine) and cultured in DMEM with 10% FBS. After induction with doxy-

cycline (2.0 mg/mL) for 16 hr, speed mutant cells were treated with a-amanitin (2.5 mg/mL, Santa Cruz) for 42 hr prior to ChIP-seq

analysis.

Plasmids, Peptides and Chemicals

pGEX-2TK cyclin T1 (1-300) (Addgene P#432) was purchased from Addgene and used to express recombinant GST-CCNT1(1-300)

in Rosetta cells. GST-CCNT1 (AA1-300) recombinant protein was puriﬁed with glutathione superﬂow agarose (Thermo, Cat# 25236).

shRNAs for human AFF1 (TRCN0000021975 and TRCN0000330908), and AFF4 (TRCN0000426769 and TRCN0000015825) were

obtained from Sigma. ELL2 was also depleted with shRNAs targeting the sequences AAC GCC AGA ATT ATA AGG ATG and

AAA TGA TCC CCT CAA TGA AGT.

Biotin labeled AFF4 peptide (AA32-67) and mutant AFF4 peptide abolishing the binding with CCNT1 were synthesized and puriﬁed

(purity > 96%) by VCPBIO with further Triﬂuoroacetic acid removal. The sequence for wild-type AFF4 peptide is Biotin-GABA-SPL

FAE PYK VTS KED KLS SRI QSM LGN YDE MKD FIG-amide and the mutant AFF4 peptide sequence is Biotin-GABA-SAA AAE PYK

VTS KAA KLSS RIQ SAA GNY DEM KDF IG-amide where Biotin indicates N-terminal biotin labeling and GABA indicates a g-amino-

butyric acid spacer. The candidate chemicals from the in silico screening were purchased from the vendors ChemDiv, ChemBridge

and Enamine.

MDA231-LM2 Tumor Model

MDA231-LM2 tumor model was established as previously reported (Wang et al., 2017). Brieﬂy, Six-week-old female athymic mice

(nu/nu genotype, BALB/c background) were purchased from Envigo (Indianapolis, IN) and housed under aseptic conditions. All pro-

tocols, described below, were approved by the Northwestern University Institutional Animal Care and Use Committee. 4 3 10⁶

MDA231-LM2 cells, in 0.4 mL of cell culture media with matrigel (BD Bioscience) were injected in the right mammary pad of mice

under anesthetization by isoﬂurane. For the in vivo therapy-response study, mice were randomly assigned to vehicle (DMSO,) KL-

1, and KL-2 treatment groups when the size of tumor reached 100 mm³. Mice were treated with drug administration by intraperitoneal

injection at 50 mg/kg of KL1 and 10 mg/kg of KL2, with once daily administration for 15 days for 3 weeks. The tumor sizes were

measured twice a week and the mice were euthanized when the tumor size reached 1000 mm³.

METHOD DETAILS

Chemical Synthesis

All chemical reagents were obtained from commercial suppliers and used without further puriﬁcation unless otherwise stated. Anhy-

˚

drous solvents were purchased from Sigma-Aldrich and dried over 3 A molecular sieves when necessary. Normal phase ﬂash column

chromatography was performed using Biotage KP-Sil 50 mm silica gel columns and ACS grade solvents on a Biotage Isolera ﬂash

puriﬁcation system. Analytical thin layer chromatography (TLC) was performed on EM Reagent 0.25 mm silica gel 60 F254 plates

and visualized by UV light. Proton (¹H), and carbon (¹³C) NMR spectra were recorded on a 500 MHz Bruker Avance III with direct

cryoprobe spectrometer. Chemical shifts were reported in ppm (d) and were referenced using residual non-deuterated solvent as

an internal standard. The chemical shifts for ¹H NMR and ¹³C NMR are reported to the second decimal place. Proton coupling

constants are expressed in hertz (Hz). The following abbreviations were used to denote spin multiplicity for proton NMR: s = singlet,

d = doublet, t = triplet, q = quartet, m = multiplet, brs = broad singlet, dd = doublet of doublets, dt = doublet of triplets, quin = quintet,

tt = triplet of triplets. Low resolution liquid chromatography/mass spectrometry (LCMS) was performed on a Waters Acquity-H UPLC/

MS system with a 2.1 mm 3 50 mm, 1.7 mm, reversed phase BEH C18 column and LCMS grade solvents. A gradient elution from

95% water +0.1% formic acid/5% acetonitrile +0.1% formic acid to 95% acetonitrile +0.1% formic acid/5% water +0.1% formic

acid over 2 min plus a further minute continuing this mixture at a ﬂow rate of 0.85 mL/min was used as the eluent. Total ion

current traces were obtained for electrospray positive and negative ionization (ESI+/ESI-). High-resolution mass spectra were

obtained using an Agilent 6210 LC-TOF spectrometer in the positive ion mode using electrospray ionization with an Agilent

Cell 175, 766–779.e1–e9, October 18, 2018 e4

G1312A HPLC pump and an Agilent G1367B autoinjector at the Integrated Molecular Structure Education and Research Center

(IMSERC), Northwestern University.

Ethyl 4-(3-methoxyphenyl)-2,4-dioxobutanoate (2a): To a solution of diisopropylamine (1.4 mL, 10 mmol) in THF (33 mL) at ꢁ78^ꢀC

was added n-BuLi (4.0 mL, 10 mmol). 3⁰-methoxyacetophenone (0.91 mL, 6.7 mmol) was added slowly, and the reaction was stirred

at ꢁ78^ꢀC for 15 min. Diethyl oxalate (1.4 mL, 10 mmol) was added slowly, and the reaction stirred at ꢁ78^ꢀC for 1.5 hr. The reaction

was slowly warmed to room temperature, then was quenched by the addition of 1M HCl (10 mL). The aqueous layer was extracted

with EtOAc (3 3 75 mL) and then combined organic layers were washed with 1M HCl (10 mL), saturated aqueous NaHCO₃(10 mL),

water (10 mL), and brine (10 mL). The organic phase was dried over Na₂SO₄, decanted into a round bottom ﬂask and concentrated by

rotary evaporation. The crude material was recrystallized from EtOH to obtain 2a (0.764 g, 46% yield) as an off-white solid. ¹H NMR

(500 MHz, CDCl₃) d 15.47 (s, 1H), 7.78 (dt, J = 7.7, 1.3 Hz, 1H), 7.73 (t, J = 2.1 Hz, 1H), 7.62 (t, J = 7.9 Hz, 1H), 7.47 (s, 1H), 7.36 (dd,

J = 8.3, 2.6 Hz, 1H), 4.61 (q, J = 7.2 Hz, 2H), 4.09 (s, 3H), 1.63 (t, J = 7.2 Hz, 3H). ¹³C NMR (125 MHz, CDCl₃) d 190.98, 169.47, 162.36,

160.17, 136.53, 130.03, 120.64, 120.38, 112.38, 98.38, 62.79, 55.67, 14.26.

Ethyl 4-(4-ﬂuorophenyl)-2,4-dioxobutanoate (2b): To a solution of diisopropyl amine (6.2 mL, 44 mmol) in THF (44 mL) at 0^ꢀC was

added n-BuLi (16.2 mL, 40.5 mmol). The cloudy yellow solution was stirred at 0^ꢀC for 30 min., then cooled to ꢁ78^ꢀC. 4’-ﬂuoroace-

tophenone (3.2 mL, 26 mmol) was added slowly along the sides of the ﬂask and was stirred for 15 min. Diethyl oxalate (7.9 mL,

58 mmol) was added and the reaction stirred at ꢁ78^ꢀC for 1 hour. The mixture was warmed to room temperature and stirred for

20 min and the reaction was quenched by the addition of 1M HCl. The organic solvent was removed by rotary evaporation. The

aqueous phase was extracted with EtOAc (3 3 75 mL) and the combined organic layers were washed with 1M HCl (25 mL), saturated

aqueuos NaHCO₃(25 mL), and brine (25 mL). The organic phase was dried over Na₂SO₄, ﬁltered, and concentrated. The crude ma-

terial was puriﬁed by ﬂash column chromatography and recrystallized from EtOH to obtain 2b (3.38 g, 54% yield) as a yellow solid.

¹H NMR (500 MHz, CDCl₃) d 15.83 – 15.03 (m, 1H), 8.42 – 8.09 (m, 2H), 7.50 (s, 1H), 7.42 (t, J = 8.5 Hz, 2H), 4.64 (q, J = 7.1 Hz, 2H),

1.65 (t, J = 7.2 Hz, 3H). ¹³C NMR (125 MHz, CDCl₃) d 189.86, 169.33, 166.36 (d, J = 256.5 Hz), 162.27, 131.50 (d, J = 2.5 Hz), 130.72

(d, J = 9.5 Hz)*, 116.3 (d, J = 22.0 Hz)*, 97.96, 62.83, 14.25.

4-(3-methoxyphenyl)-2,4-dioxobutanoic acid (3a): To a solution of 2a (0.764 g, 3.05 mmol) in THF (15 mL) was added a solution of

NaOH (1.22 g, 30.5 mmol) in 15 mL of water. The reaction stirred at room temperature for 15 min. The organic solvent was removed by

rotary evaporation. The aqueous phase was extracted with Et₂O (3 3 30 mL), then acidiﬁed with 1M HCl. The aqueous layer was

extracted wtih EtOAc (3 3 50 mL), and the combined organic layers were washed with brine (50 mL), dried over Na₂SO₄, ﬁltered,

and concentrated by rotary evaporation to obtain 3a (0.450 g, 66% yield) as an off-white solid. ¹H NMR (500 MHz, CDCl₃) d 15.32

(s, 1H), 7.74 – 7.66 (m, 1H), 7.66 – 7.59 (m, 1H), 7.54 (t, J = 8.0 Hz, 1H), 7.37 (s, 1H), 7.29 (dd, J = 8.7, 3.0 Hz, 2H), 4.00 (s, 3H).

¹³C NMR (125 MHz, CDCl₃) d 186.77, 174.75, 161.49, 160.07, 134.35, 130.09, 120.69, 120.49, 112.27, 95.27, 55.56.

4-(4-ﬂuorophenyl)-2,4-dioxobutanoic acid (3b): To a solution of 2b (3.38 g, 14.2 mmol) in THF (47 mL) was added a solution of

NaOH (5.68 g, 142 mmol) in 45 mL of water. The reaction stirred at room temperature for 15 min., then the organic solvent was

removed by rotary evaporation. The aqueous phase was extracted with Et₂O (3 3 50 mL), then acidiﬁed with conc. HCl. The aqueous

layer was extracted with EtOAc (3 3 100 mL), and the combined organic layers were washed with brine (50 mL), dried over Na₂SO₄,

ﬁltered, and concentrated to obtain 3b (1.92 g, 64% yield) as an off-white solid. ¹H NMR (500 MHz, CDCl₃) d 15.21 (s, 1H), 8.25 – 8.01

(m, 2H), 7.27 (s, 1H), 7.22 (t, J = 8.6 Hz, 2H), 7.14 (s, 1H). ¹H NMR (500 MHz, DMSO-D6) d 8.30 – 8.02 (m, 2H), 7.41 (t, J = 8.8 Hz, 2H),

e5 Cell 175, 766–779.e1–e9, October 18, 2018

7.10 (s, 1H), (Carboxylic acid -OH and enol-OH not observed). ¹³C NMR (125 MHz, CDCl₃) d 185.90, 174.19, 166.49 (d, J = 257.7 Hz),

161.63, 130.62 (d, J = 9.5 Hz)*, 129.42 (d, J = 3.1 Hz), 116.44 (d, J = 22.1 Hz)*, 95.08. * Indicates two equivalent carbons with the same

chemical shift that couple with ¹⁹F.

N-(5-chloro-2-methylphenyl)-4-(3-methoxyphenyl)-2,4-dioxobutanamide (4a): Acid 3a (0.400 g, 1.80 mmol) was dissolved in THF

(9.00 mL) and 5-chloro-2-methylaniline (0.33 mL, 2.7 mmol) was added, followed by EEDQ (0.467 g, 1.89 mmol). The reaction stirred

at room temperature for 18 hr then was diluted with EtOAc. The organic phase was washed with 1M HCl (2 3 20 mL), saturated

aqueous NaHCO₃(2 3 20 mL), water (20 mL) and brine (20 mL). The organic phase was dried over Na₂SO₄, ﬁltered, and concen-

trated. The crude material was recrystallized from MeOH to obtain 4a (0.412 g, 66% yield) as a yellow powder. ¹H NMR

(500 MHz, CDCl₃) d 15.65 (s, 1H), 9.02 (s, 1H), 8.28 (d, J = 2.2 Hz, 1H), 7.63 (d, J = 7.7 Hz, 1H), 7.59 – 7.49 (m, 1H), 7.43 (t, J =

8.0 Hz, 1H), 7.31 (s, 1H), 7.25 – 7.15 (m, 2H), 7.10 (dd, J = 8.1, 2.2 Hz, 1H), 3.90 (s, 3H), 2.35 (s, 3H). ¹³C NMR (125 MHz, CDCl₃)

d 185.77, 179.73, 160.15, 159.08, 135.81, 134.96, 132.59, 131.55, 130.12, 126.23, 125.43, 121.28, 120.48, 120.38, 112.19, 94.18,

55.68, 17.20.

N-(5-chloro-2-methylphenyl)-4-(4-ﬂuorophenyl)-2,4-dioxobutanamide (4b): To a solution of 3b (0.208 g, 0.990 mmol) in THF (5 mL)

was added 5-chloro-2-methylaniline (0.18 mL, 1.5 mmol), followed by EEDQ (0.257 g, 1.04 mmol). The reaction stirred at room tem-

perature for 18 hr, then was diluted with EtOAc. The organic phase was washed wtih 1M HCl (3 3 10 mL), saturated aqueous NaHCO₃

(3 3 10 mL), water (10 mL), and brine (10 mL). The organic layer was dried over Na₂SO₄, ﬁltered, and concentrated by rotary evap-

oration. The crude material was recrystallized from MeOH to obtain 4b (0.061 g, 59% yield) as a yellow powder. ¹H NMR (500 MHz,

CDCl₃): d 15.65 (s, 1H), 9.00 (s, 1H), 8.27 (d, J = 2.3 Hz, 1H), 8.07 (dd, J = 8.5, 5.3 Hz, 2H), 7.21 (t, J = 8.4 Hz, 2H), 7.16 (d, J = 8.1 Hz,

1H), 7.10 (dd, J = 8.3, 2.2 Hz, 1H), 2.35 (s, 3H). ¹H NMR (500 MHz, DMSO-D6) d 10.24 (s, 1H), 8.17 (dd, J = 8.5, 5.3 Hz, 2H), 7.58 (d, J =

2.5 Hz, 1H), 7.46 – 7.35 (m, 3H), 7.31 (d, J = 8.3 Hz, 1H), 7.24 (dd, J = 8.3, 2.3 Hz, 1H), 7.17 (s, 1H), 2.22 (s, 3H). 13C NMR (125 MHz,

CDCl₃) d 184.81, 179.44, 166.34 (d, J = 256.6 Hz), 158.98, 135.76, 132.60, 131.57, 130.50 (d, J = 9.4 Hz)*, 129.91 (d, J = 2.85 Hz),

126.26, 125.48, 121.30, 116.42 (d, J = 22.2 Hz)*, 93.80, 17.19. * Indicates two equivalent carbons with the same chemical shift

that couple with ¹⁹F.

In silico High-throughput Screening

In silico Filtering of The Small Molecule Database for Ligand Preparation

The ZINC database (Irwin et al., 2012), which contains approximately 41 million commercially available compounds, was used for

virtual high-throughput screening (vHTS). All compounds in the ZINC library were subjected to a panel of PAINS substructures ﬁlters

with Smiles ARbitrary Target Speciﬁcations (SMARTS) strings (Baell and Holloway, 2010) to eliminate promiscuous and non-drug-like

molecules that interfere with functionality of the target proteins. Filtering generated a list of approximately 10 million commercially

available compounds for further screening. The 10 million compound dataset was then subjected to the LigPrep module of

¨

Schrodinger (Small-Molecule Drug Discovery Suite 2017-2, Schrodinger, LLC, New York) in OPLS2005 force ﬁeld at pH 7.4 1 (phys-

iological pH) retaining the speciﬁc chirality. A low energetic 3D structure for each molecule was generated in this ligand prepara-

tion panel.

Protein Preparation for Small Molecule Screening and Grid Generation

¨

The protein preparation (prot-prep) engine implemented in the Schrodinger software suite was utilized to prepare the protein for small

˚

molecule docking simulations. Analysis of the tripartite complex crystal structure (4IMY.pdb) having a resolution 2.94 A reveals the

binding of the AFF4 protein to CCNT1, a subunit of P-TEFb. We observed that ﬁve terminal residues of AFF4 (L34, F35, A36, E37 and

P38) are having good interactions with the binding groove of CCNT1containing the residues W221, Y224, L163, V164, R165, Y175,

F176, D169, W207, W210 and E211. Furthermore, the mutation data of Y175, E211, D169, F176, R165, W210 and W207 of CCNT1

3

˚

reported in the literature (Schulze-Gahmen et al., 2013) guided us to select the small molecule ligand-binding site. A 12 A grid was

generated considering the centroid of the above mention critical residues in the CCNT1 groove.

Virtual Screening Workﬂow

For vHTS, we began with the curated library of approximately 10 million drug-like compounds described above and the OPLS 2005

˚

force ﬁeld was set. The ligand van der Waals radii was scaled to 0.80 A with partial atomic charges < 0.15 esu. A three-tier Glide

¨

docking algorithm (Small-Molecule Drug Discovery Suite 2017-2, Schrodinger, LLC) was employed that incorporates vHTS followed

by Standard Precision (SP) and Extra Precision (XP) docking protocols. The output of this three-tier docking engine was analyzed

using the XP-visualization tools by considering the interactions of the compounds with the critical residues reported above. Based

Cell 175, 766–779.e1–e9, October 18, 2018 e6

on the docking scores, a list of 122 compounds was selected for cross validation using a 5-point structure focus pharmacophore

generated by the pharmacophore module implemented in BIOVIA software considering the interactions of residues of AFF4 and

Cyclin T1. Using the 5-point pharmacophore as the query, the glide hits were scored. Based on ﬁtting scores and low energetics

conformers, 67 hits were selected. We selected 40 available compounds having a Glide docking score < ꢁ6.0. The Glide score is

¨

a function of the binding energy (Small-Molecule Drug Discovery Suite 2017-2, Schrodinger, LLC)

AlphaLISA Assay

The interaction of CCNT1-AFF4 was measured by Perkin Elmer’s bead-based AlphaLISA assays. Recombinant GST-CCNT1 (1-300),

which was puriﬁed by Glutathione Superﬂow Agarose from Rosetta cells, and AFF4 peptides were diluted in incubation buffer (25 mM

HEPES, PH 7.4, 100 mM NaCl, 0.1% NP-40). GST-CCNT1 (AA1-300) and AFF4 peptides at indicated concentrations were mixed

together with 0.5 mg of AlphaScreen Streptavidin Donor beads, and 0.5 mg of Glutathione AlphaLISA Acceptor Beads. For inhibition

assays, inhibitors were added right after the mixture of CCNT1 and AFF4 peptide. Reactions were subsequently incubated for 2 hr

with agitation in the dark. Plates were read with a Tecan INFINITE M1000 PRO. The dissociation constant Kd of CCNT1-AFF4 inter-

action was calculated based on the hyperbolic binding equation in Prism 7 (Graphpad). The IC₅₀values of the KL-1 and KL-2 were

calculated with a four-parameter sigmoid ﬁtting equation in Prism 7 and converted to the inhibitory Constants (K_i) with the Cheng and

Prusoff equation.

Heat Shock Induction

Heat shock of mammalian cells was performed using ꢂ70%–80% conﬂuent HCT-116 cells by adding pre-heated (42^ꢀC) conditioned

media collected from identically growing cells (Mahat et al., 2016b). The heat shock cells were incubated at 42^ꢀC for 1 hour. After

washing with PBS, the heat shock and non-heat shock HCT-116 cells were ﬁxed with 1% formaldehyde in PBS for downstream

ChIP-seq analysis. For heat shock induction of Drosophila S2 cells, the S2 cells were mixed with pre-heated medium to instantly in-

crease the medium temperature from 24^ꢀC to 37^ꢀC and maintained in a water bath at 37^ꢀC for 10 min before ﬁxation for ChIP-seq.

Induction of J-Lat 6.3 Cells

The J-Lat 6.3 cell line was derived from human Jurkat cells with the integration of a full-length green ﬂuorescent protein (GFP)-

encoding HIV-1 vector (HIV-R7/E^ꢁ/GFP) under the control of the viral 5⁰-LTR (Jordan et al., 2003). To measure the effects on Tat-

mediated HIV inducibility with ﬂow cytometry, J-Lat 6.3 cells were incubated with 10 nM PMA (Phorbol 12-myristate 13-acetate)

in the presence of vehicle or SEC inhibitors at the indicated concentrations for 17 hr. Cells were washed in PBS and GFP ﬂuorescence

was measured with a FACSVantage instrument (Becton Dickinson, San Jose, CA). Analysis was gated on live cells according to for-

ward and side scatter. A two-parameter analysis to distinguish GFP-derived ﬂuorescence from background ﬂuorescence was used:

GFP was measured in FL1 and cellular autoﬂuorescence was monitored in FL2. The percentage of GFP-positive cells was calculated

based on live cells (Jordan et al., 2003). The J-Lat 6.3 cells were also induced with 10 nM PMA for 11 hr and then treated with Vehicle

or SEC inhibitors for 6 hr prior to ChIP-seq analysis.

Chromatin Immunoprecipitation Sequencing

Chromatin Immunoprecipitation Sequencing (ChIP-seq) was performed according to a previously published protocol (Liang et al.,

2015). Brieﬂy, cells were crosslinked with 1% paraformaldehyde for 10 min and were quenched with glycine for 5 min at room tem-

perature. Fixed chromatin was sonicated with a Covaris Focused-ultrasonicator for 6 min and immunoprecipitated with the indicated

antibody and Dynabeads Protein G. Libraries were prepared with the HTP Library Preparation Kit for Illumina (KAPA Biosystems) and

sequenced on a NextSeq 500. ChIP-seq reads were aligned to the Drosophila genome (UCSC dm3) or human genome (UCSC hg19).

Alignments were processed with Bowtie version 1.1.2, allowing only uniquely mapping reads with up to two mismatches within the

50 bp read. The resulting reads were extended to 150 bp toward the interior of the sequenced fragment and normalized to total reads

aligned (reads per million, rpm). Peaks were called using MACS (model based analysis of ChIP-Seq) (Zhang et al., 2008) version 1.4.2

using default parameters. Ensembl version 75 transcripts were chosen with the highest total coverage from the annotated TSS to 200

nt downstream for protein coding genes that also have a RefSeq identiﬁer, were at least 2 kb long, and 2 kb away from the nearest

gene. The genes with SEC and Pol II occupancy were deﬁned by the overlapping of peaks with Pol II and SEC peaks by MACS 1.4.2

using default parameters. For pausing indexes, the promoter region was deﬁned as ꢁ200 bp upstream to 400 bp downstream, and

the body region was the remainder of the entire gene body. The ratio of the average coverage (rpm) of the promoter over the average

coverage of the gene body was then taken to be the pausing index. ECDF plots were made in R version 3.3.3 using the ecdf function.

P-values were calculated with a two-sided Kolmogorov-Smirnov test. Heatmap tables were made for the indicated windows around

the TSS or TES using the average coverage (rpm) in 25 bp & 50 bp bins (50 bp bins for 50 kb downstream of the TSS). Metagene tables

were made by approximating the coverage across all genes to the same length. All ChIP-seq heatmaps were sorted by the

decreasing coverage in indicated windows by the control samples and visualized using JavaTreeView version 1.6.4 (Saldanha,

2004). Average proﬁle plots were made by averaging the coverage for all genes using colMeans in R.

e7 Cell 175, 766–779.e1–e9, October 18, 2018

Precision Nuclear Run-on and Sequencing

Precision Nuclear Run-on and Sequencing (PRO-seq) was performed according to the previously published protocol (Mahat et al.,

2016a) with minor modiﬁcations. All 4 biotinylated nucleotides were used at 25 mM each ﬁnal concentration for the run-on reaction.

RPPH (NEB) was used to remove the 5⁰RNA cap. Libraries were size selected using a 2% agarose gel on a Pippin HT programmed to

elute 140–350 bp. After sequencing, adaptors were removed with cutadapt version 1.14 (Martin, 2011). Reads were trimmed from the

3⁰end to 36 bp with removing low quality bases using Trimmomatic version 0.33 (Bolger et al., 2014) requiring a minimal read length of

16nt. Reads were then mapped to the human genome (UCSC hg19) using Bowtie version 1.1.2 (Langmead et al., 2009). Only uniquely

mapped reads with up to 2 mismatches in the entire read were used for further analysis. Read where then converted to single nucle-

otide 3⁰BigWig strand speciﬁc tracks by taking 5⁰positions of the read (using bedtools genomecov version 2.17 (Quinlan and Hall,

2010) with options -strand -bg ꢁ5. Strands were then swapped to give the correct orientation with the 5⁰end now becoming the

3⁰end of the read (Mahat et al., 2016a). PRO-seq genome browser track examples show coverage of the entire length of the read

for easier visualization. The single nucleotide 3⁰BigWig strand speciﬁc tracks were used to generate all other ﬁgures. For the

Pol II-selected genes described above, we found the site of maximum coverage, in the region from the annotated TSS to 500bp

downstream, to which we assign the pausing site. Heatmap tables were made as described above but instead centering at this calcu-

lated pausing site.

4sU-FP-seq

Cell Labeling with 4sU, RNA Extraction and Fragmentation

20-50 million cells were treated with ﬂavopiridol for 1-2 hr to pause Pol II near the TSS sites. For the release of Pol II and measurement

of elongation rates, the cells were labeled with 4-thiouridine (4sU, Sigma-Aldrich, St. Louis, MO, USA) either in water bath or in plates.

For water bath labeling, the cells were harvested through centrifugation for 3 min at 350 g and washed with PBS twice. Then, the cells

were released with prewarmed medium containing 500 mM 4-thiouridine for 15 min, and harvested by centrifuge at 1800 g for 3 min.

For labeling in plates, the cells were washed with PBS twice after ﬂavopiridol treatment and released with prewarmed medium con-

taining 500 mM 4-thiouridine in CO₂incubator at 37^ꢀC for 15 min. RNA was extracted with 4 mL Trizol (Invitrogen) and 5 mL 20 mg/mL

glycogen. The extracted RNA was further fragmented by base hydrolysis in 0.2 M NaOH on ice for 18 min, neutralized by adding 1x

volume of 1 M Tris-HCl pH 6.8 and precipitated with isopropanol.

Biotinylation of RNA

Biotinylation of 4sU-labeled RNA was performed using EZ-Link Biotin-HPDP (Pierce) dissolved in dimethylformamide (DMF, Sigma)

at a concentration of 1 mg/mL and stored at 4^ꢀC. Biotinylation was carried out in 10 mM Tris (pH 7.4), 1 mM EDTA, and 0.2 mg/mL

Biotin-HPDP at a ﬁnal RNA concentration of 200 ng/mL for 1.5 hr. at room temperature. After biotinylation, unbound Biotin-HPDP was

removed by extracting twice with chloroform and phase lock gel. Afterward, a 1/10 volume of 5 M NaCl and an equal volume of iso-

propanol was added to precipitate RNA. RNA was collected by centrifugation at 20,000 g for 20 min and the pellet was washed with

an equal volume of 80% ethanol. The pellet was resuspended in 200 mL RNase-free water.

Puriﬁcation of 4sU-Labeled RNA, Library Preparation and Alignment

After denaturation of RNA samples at 65^ꢀC for 5 min followed by rapid cooling on ice for 5 min, biotinylated RNA was captured using

streptavidin beads. Up to 200 mg of biotinylated RNA were incubated with 50 mL of Dynabeads MyOne Streptavidin C1 with rotation

for 15 min at room temperature. Beads were washed two times with 65^ꢀC wash buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA,

1 M NaCl, 0.1% Tween20) followed by four washes with room temperature wash buffer. Labeled RNA was eluted twice for 5 min

with 100 mL of freshly prepared 100 mM dithiothreitol (DTT). RNA was recovered from the elute fractions and puriﬁed using the

RNeasy MinElute Spin columns (QIAGEN). Libraries were made with the TruSeq RNA Sample Prep kit (Illumina) and subjected to

Illumina sequencing. 4sU-FP-seq reads were aligned to the human genome (UCSC hg19). Alignments were processed with Bowtie

version 1.1.2, allowing only uniquely mapping reads with up to two mismatches within the 50 bp read. The resulting reads were

extended to 150 bp toward the interior of the sequenced fragment and normalized to total reads aligned (reads per million, rpm)

for each strand.

Elongation Rate Analysis

Genes used for calculating elongation rates were required to have an observed transcription start site with 4sU RNA signals, a min-

imum gene length of 50 kb and must be at least 2 kb away from the transcription start site of another gene. Genes were further ﬁltered

for activity/coverage by ﬁltering on the reads-per-million count within each gene body (+400 to TES) in our untreated, wild-type data.

Thus, the read coverage must be present at levels above background (rpm > 1).

We ﬁrst used the Hidden Markov Model (HMM) to calculate elongation rates. Advancing waves were identiﬁed using a three state

Hidden Markov Model (HMM) that was previously developed and implemented on GRO-seq data from a human cell line (Danko et al.,

2013). We also used SICER peak calling to determine elongation rates. Peaks were called with SICER version 1.1 (Zang et al., 2009)

with the following options–windowSize 150–fragSize 150–gapSize 3 with strand separated reads over input. These strand-speciﬁc

peaks were ﬁltered for an FDR < 0.01. Peaks were merged if there was a gap less than 2 kb. Peaks were then overlapped with TSS’s of

genes on the same strand that were greater than 50 kb. The distance traveled was calculated from the TSS to the 3⁰end of the merged

peak and the elongation rate in kb/min was calculated using the time after release.

Cell 175, 766–779.e1–e9, October 18, 2018 e8

RNA-seq Analysis

Total RNA-seq reads were trimmed from the 3⁰end until the ﬁnal base had a quality score > 30, using Trimmomatic version 0.33

(Bolger et al., 2014) and then aligned to the human genome (UCSC hg19, using Tophat version 2.1.0 (Kim et al., 2013)) with the

following options -no-novel-juncs–read-mismatches 2–read-edit-dist 2–num-threads 10–max-multihits 20 then post ﬁltering for

uniquely mapped reads using the NH ﬂag. Protein coding genes from Ensembl version 75 that also had a RefSeq identiﬁer were

only considered for analysis. Raw read counts were normalized to rpm per sample and then displayed in the UCSC genome browser

as bigWig-formatted coverage tracks. Exonic reads were assigned to speciﬁc genes from Ensembl release 75 using Bioconductor

package GenomicRanges countOverlaps. The R Bioconductor package edgeR (Robinson et al., 2010), version 3.12.0 was used to ﬁt

the data to a negative binomial generalized log–linear model and estimate a dispersion parameter. To ﬁlter out lowly expressed

genes, genes had to have at least 1 count per million (c.p.m) in at least 2 samples in each comparison. The total number of

uniquely mapped reads was provided to edgeR for the calcNormFactors normalization rather than the default column sums. An

adjusted-p value threshold of 0.01 and a log₂r.p.m cut off of 3 was used to identify genes signiﬁcantly differentially expressed in

one experimental condition relative to another. GO term analysis was done using Metascape (Tripathi et al., 2015).

QUANTIFICATION AND STATISTICAL ANALYSIS

Data are presented as Mean SD. The sample sizes (n) in the ﬁgure legends indicate the number of replicates in each experiment and

is provided in the corresponding ﬁgure legends. The peak or gene size (N) in the heatmaps indicates the number of peaks or genes

included. For Figures S1I and S5E, One-Way ANOVA tests were performed with Prism 7 (GraphPad Software, La Jolla, CA) to deter-

mine the statistical signiﬁcance. P value < 0.005 (**) was considered as highly signiﬁcantly different, p value < 0.05 was considered as

signiﬁcantly different, n.s, not signiﬁcantly different, p R 0.05. For Figures 2F, 2H, S2G, and S2H, the two-sided Kolmogorov-Smirnov

test was performed for the ECDF curves and the p values were provided in each ﬁgure. For Figures S2E, S2F, 4G, S4E, 5E, and 5G,

the statistical signiﬁcance was determined by a two-sided Wilcoxon signed-rank test using R 3.3.3 package with the p values

provided in each ﬁgure. For Figure 7D, a 2-tailed unpaired t test was used for comparison the tumor size between each treatment

group. For Figure 7E, the Kaplan-Meier survival curves were plotted with GraphPad Prism 7 and the p values were calculated using

the log-rank test.

DATA AND SOFTWARE AVAILABILITY

The accession number for the raw and processed ChIP-seq, RNA-seq, PRO-seq and 4sU-FP-seq data reported in this paper is GEO:

GSE112608.

e9 Cell 175, 766–779.e1–e9, October 18, 2018

Supplemental Figures

A

B

C

5

8

6

4

2

10

CCNT1 (1-300) 5 nM

Kd = 23.46 nM

WT AFF4 peptide (10 nM)

Mutant AFF4 peptide

6

1.0 10

No CCNT1

O2

Excitation

5

7.5 10

Emission

5

5.0 10

5

Biotin-AFF4

2.5 10

GST-CCNT1

peptide

0

0.1

0

0.01

1

10

100

0.1

1

10

100

Biotin-AFF4 peptide Conc. nM

CCNT1(1-300) Conc. nM

a. LDA,

diethyl oxalate

O

O O

D

b. NaOH

O

OH

R

46-54%

64-66%

O

1a R = 3-OMe

1b R = 4-FI

2a R = 3-OMe

2b R = 4-FI

3a R = 3-OMe

3b R = 4-FI

c. EEDQ,

5-chloro-2-methylaniline

O

H

N

Cl

R

59-66%

O

4a R = 3-OMe

4b R = 4-FI

DMSO

KL-2

KL-1

E

293T lysate

loading (μl):

F

DMSO

KL-2

KL-1

BRD2

100KD

250KD

HCT116 lysate

loading (μl):

BRD4

250KD

AFF1

150KD

250KD

150KD

Pol II

(RPB1)

AFF4

CCNT1

CDK9

SPT5

MED26

MYC

75KD

50KD

75KD

50KD

Tubulin

50KD

100KD

HSP90

Tubulin

SEC mRNA expression levels

after KL-1/KL-2 treatment

50KD

I

10

8

Flag-AFF4 stable cells

Flag-AFF1 stable cells

G

H

***

n.s.

**

***

n.s.

6 hrs

treatment:

6

250KD

150KD

Flag-AFF1

n.s.

Flag-AFF4

150KD

75KD

50KD

4

DMSO

KL-1

KL-2

75KD

50KD

CCNT1

Tubulin

CCNT1

Tubulin

2

0

Figure S1. Peptidomimetic Identiﬁcation of Disruptors of the AFF4-CCNT1 Interaction within SEC, Related to Figure 1

(A) Schematic of AlphaLISA assay of the AFF4-CCNT1 interaction. Anti-GST AlphaLISA acceptor beads and AlphaScreen streptavidin donor beads (Perkin Elmer)

were used to detect the interaction of GST-CCNT1 (AA1-300) and biotin-AFF4 (AA32-67).

(B and C) Optimization of the AlphaLISA assay with various biotin-AFF4 peptide and GST-CCNT1 concentrations. No CCNT1 (B) and mutant AFF4 (C) were used

as negative controls. Data are represented as mean SD.

(D) Scheme for chemical synthesis of SEC inhibitors KL-1 and KL-2. Reagents and conditions: a. LDA, THF, ꢁ78^ꢀC, 15 min., then diethyl oxalate, ꢁ78^ꢀC to room

temperature. b. NaOH, THF/H₂O, room temperature, 15 min. c. EEDQ, 5-chloro-2-methylaniline, THF, room temperature, 24 hr.

(E) 6 hr of treatment with KL-1 or KL-2 does not signiﬁcantly reduce the protein levels of BRD2, BRD4, Pol II (RPB1) and MYC in 293T cells. 293T cells were treated

with 20 mM of KL-1 or KL-2 for 6 hr, and the protein levels of BRD2, BRD4, Pol II (RPB1), SPT5, MED26, MYC, HSP90 (load control) and Tubulin (load control) were

measured by western blotting.

(F) Treatment of HCT-116 cells with KL-1 and KL-2 reduced protein levels of SEC components AFF1 and AFF4 but not CDK9 or CCNT1. HCT-116 cells were

treated with 20 mM of KL-1 or KL-2 for 6 hr, and the protein levels of AFF1, AFF4, CCNT1, CDK9 and Tubulin (load control) were determined by western blotting.

(G and H) Treating cells with KL-1 and KL-2 results in reduced levels of AFF1 and AFF4. FLAG-AFF1 (G) and FLAG-AFF4 (H) were expressed under the Tet-

inducible promoter in HEK293T cells. Cells were treated with 20 mM of the indicated inhibitors for 6 hr for western blotting.

(I) KL1 and KL-2 do not lead to reduced mRNA expression of SEC subunits AFF1, AFF4, ELL2, CCNT1 and CDK9 in HEK293T cells. RNA-seq was performed in

HEK293T cells treated for 12 hr with 20 mM of the indicated inhibitors. The log₂counts per million (CPM) SEC genes were calculated with HTseq. KL-1 and KL-2

had no signiﬁcant effects on CCNT1, CDK9 and AFF1 expression, but modest increases in the mRNA levels of AFF4 and ELL2 were observed (n = 3). Unpaired

two-way ANOVA was used for the statistical testes. **p < 0.05, ***p < 0.001.

SRSF4

HCT-116 cells

Pol II occupancy at TSS

Jurkat cells

Pol II occupancy at TSS

Pol II ChIP-Seq

A

B

C

40 -

15

DMSO

KL-1

KL-2

KL-1

KL-2

0 _

40 -

15

10

5

KL-1

KL-2

10

5

0 _

40 -

0 _

60 -

DMSO

0 _

60 -

0

KL-1

KL-2

−2000

TSS

+2000

−2000

TSS

+2000

0 _

60 -

HCT-116 cells

G

ECDF plot of Pausing Index

0 _

1.00

0.75

0.50

0.25

0.00

DMSO

KL-1

KL-2

AFF1

D

E

Log2 Pol II FC vs DMSO

KL-1 KL-2

p<2.2E-16

N=6681

Group I

N=3062

more pausing

2

4

6

8

Log2(Pausing Index)

Groups:

I

II

III

Jurkat cells

ECDF plot of Pausing Index

H

AFF4

F

1.00

0.75

0.50

0.25

0.00

DMSO

KL-1

KL-2

Group II

N=2493

p<2.2E-16

N=6073

Group III

N=1285

more pausing

2

4

6

8

- 1kb TSS +1kb

N=6840

Log2(Pausing Index)

Groups:

HSPA8

I

II

III

I

J

Pol II ChIP-Seq

SRSF4

TES

5 -

DMSO

0 _

5 -

0 _

5 -

5’ shift

KL-1

KL-2

KL-1

0 _

5 -

0 _

5 -

KL-2

0 _

Figure S2. Small-Molecule Disruption of SEC Increases Promoter-Proximal Pausing, Related to Figure 2

(A) Genome browser tracks of Pol II occupancy at the SRSF4 gene in vehicle or SEC inhibitor-treated cells gene in HCT-116 (top panel) and Jurkat cells (bot-

tom panel).

(B and C) Metaplots of Pol II occupancy for all expressing genes with Pol II occupancy in HCT-116 (B) and Jurkat cells (C). SEC inhibitor treatments result in

increased Pol II occupancy around the TSS. Pol II density is plotted in a ꢁ2 kb and +2 kb window around the TSS.

(D) K-means 3 clustered heatmap of Pol II fold change of the 6,840 genes after SEC inhibitor treatments. The region around the TSS is shown. Group I genes are

preferentially affected by the SEC inhibitors.

(E and F) Boxplot analysis of AFF1 (E) and AFF4 (F) occupancy at the three clusters from (D).

(G and H) ECDF plots of Pol II pausing indexes for all expressing genes in HCT-116 (G) and Jurkat cells (H) in the presence of vehicle or SEC inhibitors.

(I and J) SEC inhibition results in a 5⁰shift of Pol II density near the Transcription End Site (TES) at the HSPA8 and SRSF4 genes. Pol II coverage in HEK293T cells is

displayed in the UCSC genome browser.

Figure S3. Disruption of SEC Phenocopies Slow Pol II Mutants and Reduces Pol II Processivity, Related to Figure 3

(A) Identiﬁcation of the 1,057 genes with typical Pol II termination signals around the TES. Pol II signals from WT Pol II mutant expressing cells were K-means

clustered into 5 groups with a window of 15 kb around the TES. The expressing genes from clusters with a typical Pol II termination signals were selected and

plotted.

(B) SEC inhibitor treatments of the WT Pol II mutant expressing cells lead to a Pol II phenotype similar to slower Pol II mutant expressing cells as viewed at the

3⁰end of genes.

(C and D) Time-dependent and dose-dependent shift of Pol II signals around the TES sites by KL-2 treatment (N = 1,057).

(E) Genome browser tracks of Serine 2 phosphorylated (Ser2P) Pol II at the SRSF1 gene in HEK293T cells treated with vehicle or SEC inhibitors. SEC inhibitor

treatments result in increased Pol II occupancy at the promoter-proximal region and increased occupancy of Ser2P Pol II in the gene body.

(F and G) Metagene plot of Pol II Ser2P occupancy (F) and log₂Ser2P/total Pol II ratio (G) for 6,840 well-expressed genes after SEC inhibitor treatments, indicating

that disruption of SEC leads to altered Pol II dynamics, with increased CTD Ser2 phosphorylation ratio near TSS sites and decreased Ser2P occupancy after the

annotated TES sites.

(H) SEC inhibition results in decreased protein levels of the SEC subunit ELL2. 293T cells were treated with 20 mM of the indicated inhibitors for 6 hr before

harvesting cells for western blotting.

(legend continued on next page)

(I–K) Depletion of ELL2 in HEK293T cells (I) results in apparent early termination of Pol II with reduced Pol II occupancy downstream of the PIM3 (J) and ACTB (K)

annotated TES.

(L) Metaplots of Pol II occupancy at TES regions for all of the 1,057 genes in Figure S3A. ELL2 knockdown results in a 5⁰shift of Pol II in these regions.

(M and N) Metaplot and heatmap analysis of PRO-seq signal in HEK293T cells treated with vehicle or SEC inhibitors for the region 500 bp upstream and 500 bp

downstream of the empirical promoter-proximal pausing site identiﬁed in the control condition. Rows represent genes and are sorted according to total PRO-seq

signal in the control condition (N). The right two panels display log₂fold changes in PRO-seq signal with the indicated inhibitors. Increased Pol II occupancy in

promoter-proximal regions is observed along with decreased occupancy in the region downstream.

Figure S4. Small-Molecule Disruption of SEC Slows Pol II Elongation Rates, Related to Figure 4

(A) Hidden Markov Model (HMM) for elongation rates analysis. Raw changes in 4sU-FP-seq read counts in non-overlapping 50 bp windows used to infer

elongation rates for the ACTN2 gene. The boxes show the span of advancing wave inferred by a 3-state HMM analysis, demonstrating that SEC inhibitors

decrease the elongation rates of Pol II elongation at the ACTN2 gene.

(B) Calculation of Pol II elongation rate using the length of SICER-called peaks of 4sU-FP-seq data to measure distance traveled. Genome browser tracks of 4sU-

FP-seq signals at the MTR gene in vehicle and SEC inhibitor-treated HEK293T cells. Black bars beneath the tracks indicate SICER-called peaks. Note that in the

DMSO condition, Pol II has traveled further during the 15 minutes of release from ﬂavopiridol resulting in a longer distance traveled.

(C and D) Histograms of elongation rates (N = 1,484) calculated form SICER data for DMSO (C-D) and SEC inhibitors KL-1 (C) and KL-2 (D).

(E) Boxplot analysis of SICER-based elongation rates (N = 1,484, left panel) and log₂fold change of elongation rates (right panel) for the indicated treatments

Wilcoxon test was used for the statistical analysis.

Figure S5. SEC Inhibitors Block Transcription Elongation in SEC-Dependent Rapid Response Models, Related to Figure 5

(A) Genome browser tracks of Pol II occupancy demonstrating that SEC inhibitors attenuate the induction of heat shock-induced gene EGR1.

(B) Gene Ontology analysis of the 136 identiﬁed heat shock-induced genes in HCT-116 cells.

(C) Workﬂow of Tat-induced HIV genome activation in J-Lat 6.3 cells. J-Lat 6.3 cells contain an integrated copy of the HIV genome that has GFP replacing HIV nef

and could be activated by Phorbol-12-Myristate-13-Acetate (PMA). J-Lat 6.3 cells were treated with vehicle or 20 mM SEC inhibitor along with PMA to activate the

HIV genome transcription, which could be monitored by ﬂow cytometry analysis.

(D and E) SEC inhibitors block Tat-induced HIV genome activation in J-Lat 6.3 cells. (D) FACS analysis of J-Lat 6.3 cells treated with 10 nM PMA and SEC in-

hibitors for 17 hr. Treatment with SEC inhibitors attenuates HIV genome activation in a dose-dependent manner (E) as revealed by ﬂow cytometry analysis with

GFP (n = 3-5). Unpaired two-way ANOVA was used to compare the groups between DMSO and SEC inhibitors. ***p < 0.001. Data are represented as mean SD.

(F) Genome browser views demonstrating that SEC inhibitors inhibit the Tat-dependent induction of the integrated HIV genome. J-Lat 6.3 cells were ﬁrst induced

with 10 nM PMA for 11 hr to induce the Tat expression and then treated with 20 mM SEC inhibitors for 6 hr prior to Pol II ChIP-seq.

Article Doi

Targeting Processive Transcription Elongation via SEC Disruption for MYC-Induced Cancer Therapy

DOI: 10.1016/j.cell.2018.09.027

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.cell.2018.09.027

Full text of DOI:10.1016/j.cell.2018.09.027

Products guided by the article

R&D Labs maybe for 64393-83-5

Relevant to this article

Hot Product