- Email: [email protected]
Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles
Article
Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles Graphical Abstract
Authors Anat Bahat, Or Lahav, Alexander Plotnikov, Dena Leshkowitz, Rivka Dikstein
Correspondence [email protected]
In Brief Spt5 promotes promoter-proximal pausing, promoter escape, mRNA processing, and selective transcription of the mutant huntingtin gene. Bahat et al. identified Spt5-Pol II small-molecule inhibitors that mimic Spt5 knockdown effects and uncouple its different functions. The inhibitors uncovered Spt5 regulatory roles in metabolism and 30 end processing of histone genes.
Highlights d
Discovery of Spt5-Pol II small-molecule inhibitors (SPIs) that mimic Spt5 knockdown
d
SPIs relieve promoter-proximal pausing and inhibit inflammatory gene activation
d
d
SPIs inhibit mutant huntingtin gene transcription and 30 end processing of histone genes Selective targeting of Spt5-Pol II activities by SPIs uncouples its different functions
Bahat et al., 2019, Molecular Cell 76, 1–15 November 21, 2019 ª 2019 Elsevier Inc. https://doi.org/10.1016/j.molcel.2019.08.024
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Molecular Cell
Article Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles Anat Bahat,1 Or Lahav,1 Alexander Plotnikov,2 Dena Leshkowitz,3 and Rivka Dikstein1,4,* 1Department
of Biomolecular Sciences, The Weizmann Institute of Science, Rehovot 76100, Israel Nancy and Stephen Grand Israel National Center for Personalized Medicine, The Weizmann Institute of Science, Rehovot 76100, Israel 3Bioinformatics Unit, Department of Life Sciences Core Facilities, The Weizmann Institute of Science, Rehovot 76100, Israel 4Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.molcel.2019.08.024 2The
SUMMARY
Spt5 is a conserved and essential transcription elongation factor that promotes promoter-proximal pausing, promoter escape, elongation, and mRNA processing. Spt5 plays specific roles in the transcription of inflammation and stress-induced genes and tri-nucleotide expanded-repeat genes involved in inherited neurological pathologies. Here, we report the identification of Spt5-Pol II small-molecule inhibitors (SPIs). SPIs faithfully reproduced Spt5 knockdown effects on promoter-proximal pausing, NF-kB activation, and expanded-repeat huntingtin gene transcription. Using SPIs, we identified Spt5 target genes that responded with profoundly diverse kinetics. SPIs uncovered the regulatory role of Spt5 in metabolism via GDF15, a food intake- and body weight-inhibitory hormone. SPIs further unveiled a role for Spt5 in promoting the 30 end processing of histone genes. While several SPIs affect all Spt5 functions, a few inhibit a single one, implying uncoupling and selective targeting of Spt5 activities. SPIs expand the understanding of Spt5-Pol II functions and are potential drugs against metabolic and neurodegenerative diseases.
INTRODUCTION The regulation of transcription elongation by RNA polymerase II (Pol II) is central to many gene expression programs, and the deregulation of this stage is implicated in various pathologies. Elongating Pol II is associated with several transcription elongation regulatory factors, of which Spt5 is the most highly conserved, with homologs from bacteria to human (Werner, 2012). Spt5 interacts with Spt4 to form the DRB sensitivityinducing factor (DSIF) complex (Wada et al., 1998a; Yamaguchi et al., 1999b), which is mostly known for its function in promoter-proximal pausing in which Pol II stalls after transcribing 30–120 nt downstream of the transcription start site
(TSS). DSIF-mediated promoter-proximal pausing involves interaction with the negative elongation factor (NELF) (Yamaguchi et al., 1999a; Vos et al., 2018). The release of paused Pol II into productive elongation is mediated by the positive transcription elongation factor b (P-TEFb), a kinase that phosphorylates serine 2 of the Pol II C-terminal repeat domain (CTD) and facilitates this transition (Ni et al., 2008; Peng et al., 1998; Wada et al., 1998b). Genes known to be highly susceptible to inhibition via promoter-proximal pausing are those that are immediately activated in response to stress and extracellular signals such as heat shock and proinflammatory genes (Boehm et al., 2003; Diamant and Dikstein, 2013; Missra and Gilmour, 2010; Rougvie and Lis, 1990; Wu et al., 2003). During elongation, Spt5 has the ability to stabilize Pol II on the DNA template by interacting with upstream sequences and with RNA (Bernecky et al., 2016; Blythe et al., 2016; Crickard et al., 2016; Hirtreiter et al., 2010; Klein et al., 2011; Martinez-Rucobo et al., 2011). This activity was recently implicated in maintaining high transcriptional speed on long mRNAs (Fitz et al., 2018). Consistent with this, Spt5 was shown to accompany Pol II along the transcribed region until termination (Mayer et al., 2010; Pavri et al., 2010; Rahl et al., 2010). While in yeast and Drosophila, Spt5 depletion causes genome-wide transcription defects during the early elongation of sense and antisense transcripts (Henriques et al., 2018; Shetty et al., 2017), the depletion of Spt5 in mammalian cells does not result in broad transcriptional effects (Diamant et al., 2012, 2016b; Fitz et al., 2018; Komori et al., 2009; Pavri et al., 2010; Rahl et al., 2010; Stanlie et al., 2012), suggesting greater redundancy among mammalian transcription elongation factors. Our previous studies established Spt5 as a central player in the cellular response to nuclear factor kB (NF-kB) activation by controlling several steps during transcription. Induction of a subset of NF-kB target genes is remarkably fast and involves the presence of the transcription machinery on the promoters before activation and promoter-proximal pausing mediated by DSIF (Ainbinder et al., 2002, 2004). At the NF-kB target A20 gene, promoter-proximal pausing by DSIF-Pol II is mediated by an enhancer box (E-box) at the promoter (Ainbinder et al., 2004; Amir-Zilberstein et al., 2007; Amir-Zilberstein and Dikstein, 2008). In addition, in a subset of NF-kB target genes, DSIF coordinates elongation with mRNA splicing and export (Diamant Molecular Cell 76, 1–15, November 21, 2019 ª 2019 Elsevier Inc. 1
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A Spt5
Rpb1
Rpb1
Spt5
C-RL
N-RL
B Establishment of split-RL assay for Spt5-Pol II interaction Screening ~100,000 compounds (>30% inhibition)
587 Screening with full-length RL and removal of RL enzyme inhibitors
309 IC50 dose response assay at 5 concentrations: 0.55, 1.6, 5, 15, 45μM
148 Live cell assay
140 Filtering out overlapping hits
96
Molecules with IC50<40μM were selected for further analysis
41 Biologically active
18
C H
N
100
O
Cl
N
Cl
50
50
75
O
IC50 = 8.9μM
O
SPI-31
25
Ac t i v i t y ( % )
H3 C
N
60
SPI-29
N
IC50 = 36.5μM
70
Ac t i v i t y ( % )
N
75
N
80
N
SPI-17
Ac t i v i t y ( % )
90
N O
N
IC50 <5μM
50
CH 3 H N
25
S
O
O
40 O
100
1
H
80 O
O
50
Ac t i v i t y ( % )
Ac t i v i t y ( % )
H
IC50 = 25.2μM
75
SPI-57 CH 3
40
1
S
S S
SPI-95
O
10 Concent rat ion (μM)
O
50
N
IC50 = 25.4μM
75
50
CH 3
O
+
H
N H S
S
25
100
N
1 10 Concent rat ion (μM)
100
100
N
90
IC50 = 32.4μM
80 70 60
H3C
N
SPI-39
N
O N
80
IC50 = 16.9μM
70 60 50 40
50
S
H 3C
CH 3
90
H
N
60
N
SPI-85
100 H
IC50 = 33.1μM
70
Ac t i v i t y ( % )
Ac t i v i t y ( % )
50
20
100
80
100
100
IC50 = 36.1μM
60
100
90
10 Concent rat ion (μM)
N
70
30
O
25 1 10 Concent rat ion (μM)
1
Ac t i v i t y ( % )
100
SPI-18
H
100
90
O
SPI-21
10 Concent rat ion (μM)
Ac t i v i t y ( % )
1 10 Concent rat ion (μM)
30 1 10 Concent rat ion (μM)
100
1 10 Concent rat ion (μM)
100
1 10 Concent rat ion (μM)
H
85 80
SPI-157
90
75
IC50 = 5.8μM
50
S
H
80
100
1 10 Concent rat ion (μM)
–
100
1 10 Concent rat ion (μM)
40
SPI-42
O F
30
IC50 = 26.1μM
70 60 50
H 3C N
10 Concent rat ion (μM)
+
H
100
1 10 Concent rat ion (μM)
Cl N
IC50 = 21.6μM
80 70 60
S
F F
100
F
1 10 Concent rat ion (μM)
100
H 3C
110
70 60
H
SPI-46
N
H N
N
50
H
N
–
IC50 = 12μM
60 50
SPI-86
O
F
100
60
IC50 = 5.5μM
50 40
S N O
N
S
40
40 1 10 Concent rat ion (μM)
O
O
70
70
Ac t i v i t y ( % )
IC50 = 18.2μM
80
O
80
O
90
Ac t i v i t y ( % )
SPI-74
Ac t i v i t y ( % )
100
N
Cl
50
H 3C
30 1
SPI-16
O
40
S
20
N
90
80
Ac t i v i t y ( % )
IC50 = 26.5μM
50
100
100
Cl
90
60
40 20
100
O
Ac t i v i t y ( % )
Ac t i v i t y ( % )
SPI-68
O
50
O
1 10 Concent rat ion (μM)
60
30
70
70
IC50 = 14μM
70
S
O
O
25
75
SPI-09
O
Ac t i v i t y ( % )
IC50 = 40μM
90
Ac t i v i t y ( % )
Ac t i v i t y ( % )
95
O
100
O
100
SPI-06
100
H
100
105
1
10 Concent rat ion (μM)
100
N
30
O
1
10 Concent rat ion (μM)
+
O
–
100
(legend on next page)
2 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
et al., 2012). In addressing the activity of Spt5 in transcription elongation dynamics, we unexpectedly discovered that it also plays an important role in the early stages of transcription, supporting promoter escape and H3K4me3 and H4K5Ac chromatin modifications. These, in turn, facilitate TFIID maintenance on the promoter and re-initiation for the rapid induction of NF-kB response genes (Diamant et al., 2016a). Recently, Spt4/Spt5 has been shown to play a specific role in transcribing genes with lengthy stretches of multiple repeats that are the hallmarks of several inherited degenerative neurological disorders, including Huntington disease (HD), amyotrophic lateral sclerosis (ALS), and frontotemporal dementia (FTD) (Cheng et al., 2015; Kramer et al., 2016; Liu et al., 2012). Interference with Spt4 or Spt5 by small interfering RNA (siRNA) was reported to decrease the transcription of mutant huntingtin (Htt) gene alleles containing long CAG expansions, but had little effect on the expression of genes containing short CAG stretches (Cheng et al., 2015; Liu et al., 2012). The underlying basis for the requirement of Spt4/Spt5 in transcribing these genes is presently unknown. Because Spt5 is essential for cell viability, most studies aiming to interfere with its activity in Drosophila and mammalian cells use knockdown (KD) methods. However, these methods are lengthy, as treatment with siRNA requires at least 72 h, making distinguishing between direct and indirect effects challenging. In addition, it is labor-intensive and expensive. To expand the available tools for the study of Spt5, we report here the identification and characterization of small-molecule inhibitors (SPIs). SPIs precisely mimic the effect of Spt5 depletion on basal and activated NF-kB target genes and on the expanded repeat Htt gene. Using these inhibitors, we identified heat shock genes as transient target genes and uncovered novel regulatory roles of Spt5 that include the facilitation of the unique mode of histone gene termination and the regulation of an important metabolic pathway mediated by GDF15, a food intake- and body weightregulatory hormone. A subset of SPIs is selective to a single activity of Spt5, suggesting that the various activities of this factor can be uncoupled. RESULTS Identification of Spt5-Pol II Inhibitors Using HighThroughput Drug Screen The genetic manipulation of many proteins in mammals is limited by their importance for cell viability and the presence of multifunctional domains. Spt5 is an essential transcription elongation factor that has been mostly studied in metazoan cells using KD, as presently no inhibitor is available that can interfere with its activity. A major limitation of the KD approach against Spt5 is the lengthy treatment of at least 72 h. As the activities of Spt5 are critically dependent on its ability to bind Pol II (see, for example, Diamant et al., 2012, 2016a, 2016b), we envisaged that targeting this complex may enable the identification of Spt5 inhibitors,
which would be useful not only for Spt5 research but also for developing therapies against HD and inherited ALS and FTD. Spt5 is a large protein that contains several conserved functional domains (Werner, 2012), including NusG N-terminal homology domain (NGN), which mediates the interaction with Pol II; 5 KOW domains, which are also involved in the interaction with Pol II; a repetitive heptapeptide motif in the C-terminal region called CTR, implicated in the positive elongation activity of Spt5 (Chen et al., 2009; Ivanov et al., 2000; Yamada et al., 2006); and a C-terminal domain with an unknown function. We applied the split Renilla luciferase (RL) complementation assay to detect the Spt5 interaction with Pol II, as previously described (Ashkenazi et al., 2016, 2017). In this assay, RL is split into 2 inactive N- and C-terminal fragments and fused to target proteins. Interaction of the target proteins brings the N and C termini into close proximity, which can restore the RL enzymatic activity. In this system, the spatial arrangement of the 2 RL parts is critical for the enzymatic activity; therefore, it is sensitive not only to direct interference with the interacting proteins but also to changes in the conformation of the fused proteins (Ashkenazi et al., 2016; Haimov et al., 2018). For this purpose, the region spanning NGN and KOW 1 and 2 of Spt5 was fused to N-terminal Renilla luciferase (N-RL), and the coiled-coil (CC) domain of Rpb1 (the large subunit of Pol II) was fused to C-terminal Renilla luciferase (C-RL) (Figure 1A). These constructs were tested for specific interactions in mammalian cells and then co-expressed in Escherichia coli from a single plasmid, as described by Ashkenazi et al. (2017). A major advantage of the use of recombinant protein fragments of large proteins such as Spt5 and Rpb1 is the potential to identify specific compounds that directly bind to these protein domains. Bacterial cell lysates were used in a 1,536-well plate format to screen several small-molecule libraries with diverse chemical characteristics, consisting of 100,000 compounds (Figure 1B). A total of 587 inhibitors of the RL activity were identified and tested with the full-length RL enzyme to eliminate inhibitors of the RL enzymatic activity. The resulting 309 compounds were further validated in a doseresponse assay using Spt5 and Pol II split-RL pair and the fulllength RL. Approximately half of the compounds also inhibited the full-length RL at the high concentration, leaving 148 compounds that specifically inhibited the activity of the Spt5-Pol II split-RL pair. To select compounds that can enter mammalian cells, they were further tested in a live-cell split-RL assay, as previously described (Ashkenazi et al., 2017), and 140 retained their inhibitory activity. After filtering out 44 potentially non-specific compounds, as they also appeared in other screens of these libraries, we selected 41 Spt5-Pol II inhibitors (SPIs) displaying a half-maximal inhibitory concentration (IC50) % 40 mM for further analysis. Of the 41 compounds analyzed, 18 were found to exert at least one biological effect (see below), whereas the rest exerted none. The chemical structures and the IC50 measurements of the biologically active compounds are shown in Figure 1C. Notably, several of these compounds appear to be chemically
Figure 1. A High-Throughput Drug Screen (HTS) against Spt5-Pol II (A) An illustration of Spt5 (NGN + KOW1/2 domains) and Rpb1 (coiled-coil domain) split-RL assay used for the HTS. (B) A flowchart describing the steps carried out during the HTS in search of direct inhibitors of Spt5-Pol II. (C) The chemical structure and the IC50 of the SPIs found to be biologically active.
Molecular Cell 76, 1–15, November 21, 2019 3
IEX1
Relative mRNA level
IL6
JunB
B Relative A20 mRNA level
Relative mRNA level
CXCL2
CCL20
Relative mRNA level
Relative mRNA level
IL8
Relative mRNA level
A20
Relative mRNA level
A
Relative mRNA level
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
E
D
C
Average of A20, IL6, IL8, CXCL2 Relative mRNA level
W/O Promoterproximal pausing
F
With Promoterproximal pausing 73%
G
HSPA6
EGR1
FOS
A20
IL8
(legend on next page)
4 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
related—for example, one class includes SPI-21, SPI-29, and SPI-57 and another SPI-68, SPI-46, and SPI-157. SPIs Relieve Promoter-Proximal Pausing In vitro transcription assays, along with Spt5 KD studies, established the NF-kB-responsive A20 gene as a target of Spt5 for promoter-proximal pausing under basal conditions (Ainbinder et al., 2004; Amir-Zilberstein and Dikstein, 2008). Following the activation of NF-kB by tumor necrosis factor a (TNF-a), Spt5 turns into a positive regulatory factor and facilitates promoter escape and mRNA splicing (Diamant et al., 2012, 2016a). Genome-wide analysis of transcripts from Spt5 KD cells identified other Spt5-regulated genes under resting and TNFa-induced conditions (Diamant et al., 2016b). To examine the ability of SPIs to interfere with Spt5 activity, we selected a subset of TNF-a-responsive genes that are also targeted by Spt5 (A20, C-C motif chemokine ligand 20 [CCL20], interleukin [IL]-6, IL-8, C-X-C motif chemokine ligand 2 [CXCL2], JUNB, and IEX1) from the aforementioned studies and from others (Aida et al., 2006; Fujita et al., 2009) and tested their effect under basal conditions. Cells were treated with vehicle (DMSO) or SPI (%50 mM) for 2 h, and then RNA was extracted and analyzed by qRT-PCR. Most of the examined SPIs enhanced the basal activity of the selected NF-kB-responsive genes (Figure 2A), akin to the effect of Spt5 KD. To determine whether the observed effect is linked to activated transcription or an increase in mRNA stability, we treated cells with actinomycin D in the presence or absence of one randomly selected compound, SPI-21, and then measured the mRNA levels at different time points. The stability of A20 mRNA was not enhanced but actually was moderately reduced in SPI-21 samples (Figure 2B). To obtain a genome-wide view of the impact of SPIs on transcription, we performed Global run-on sequencing (GRO-seq) using the nuclei of HeLa cells treated with either DMSO (control) or SPI-21. The GRO-seq experiment measures the distribution of transcriptionally active Pol II along the length of the gene (Core et al., 2008). This analysis revealed that short-term SPI-21 treatment caused a transcription upregulation (R2-fold) of 1,371 genes, among them A20 (TNFAIP3), IL-6, IL-8 (CXCL8), and CCL20 (analyzed above), while CXCL2, IEX1 (IER3), and JUNB were unchanged. The reduction of promoter-proximal pausing
is expected to release the paused Pol II into the gene body, and thus upregulate transcription. We therefore calculated the ratio between Pol II density around the TSS and the gene body (pausing index [PI]) of the SPI-21 upregulated gene set and found that the majority display promoter-proximal pausing (PI >1.2) (Figure 2D). Furthermore, treatment with SPI-21 resulted in a general reduction in their PI (Figure 2E). The sequencing tracks of a few examples of upregulated genes with reduced PI values following SPI-21 treatment are shown in Figure 2F. These findings together confirm that elevated mRNA levels following short-term SPI treatment are, at least in part, a consequence of promoter-proximal pausing relief, which is consistent with a known function of Spt5. We examined the SPI-21 upregulated genes from the GROseq data for the enrichment of regulatory elements in their control regions (2,000 to +2,000 around the TSS) using a gene set enrichment analysis (GSEA) database and software (Subramanian et al., 2005). One of the highest-scoring motifs identified has a CAGGTG/CACGTG sequence (Figure 2G) that is remarkably similar to the E-box motif of the A20 promoter CACGTG, which was found to control promoter-proximal pausing through Spt5 (Ainbinder et al., 2004; Amir-Zilberstein and Dikstein, 2008). Another enriched motif is the binding site of heat shock factor 1 (HSF1), a transcription activator of heat shock genes, which we found to be regulated by Spt5 via promoter-proximal pausing in mammalian cells (see Figures 2F and 4D). SPIs Mimic the Effects of Spt5 KD on Pro-inflammatory NF-kB Target Genes and on Cell Survival Next, we analyzed the effect of SPIs following NF-kB activation, a state in which Spt5 turns into a positive regulatory factor. Cells were pre-incubated with DMSO or SPIs (%50 mM) for 1 h and then treated with TNF-a for an additional 1 h. The selected genes were either moderately (CXCL2, JUNB, IEX1) or strongly (A20, CCL20, IL-6, IL-8) activated by TNF-a (Figure 3A, DMSO+). The induced levels of all of these genes were diminished to a variable degree by the SPIs, similar to the effect seen with Spt5 KD (Diamant et al., 2016a). The potent inhibitors decreased the induction rate of all of the analyzed genes, while the weaker inhibitors such as SPI-68, -74, -95, -42 and -06 affected mainly
Figure 2. Effects of SPIs on Promoter-Proximal Pausing and Pol II Recruitment (A) HeLa cells were treated with DMSO or 50 mM of the indicated SPIs (except SPI-57 and SPI-42, which were at 25 mM, and SPI-31, which was at 7.5 mM) for 2 h. Then, total RNA was extracted, and the levels of the indicated mRNAs were determined by quantitative real-time PCR and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Bars represent the means ± SEMs of at least 3 independent experiments. (B) The effect of SPI-21 on the basal levels of A20 mRNA stability. Cells were treated with SPI-21 and actinomycin D for the indicated time points, and the levels of A20 and GAPDH mRNA were determined by qRT-PCR. The data present the means ± SEMs of 3 independent experiments. (C) A graph summarizing the average effect of SPIs on the indicated genes organized in descending order of potency. (D) HeLa cells, pretreated with either DMSO or SPI-21 for 90 min, were subjected to GRO-seq analysis as described in the STAR Methods. A pie chart demonstrating the percentage of SPI-21 upregulated genes that display promoter-proximal pausing (PI R1.2). The PI was calculated by dividing Pol II density (reads per 1,000 bp) around the TSS (300 to +700 bp) by Pol II density in the gene body (+700 to +1,700). (E) A boxplot presenting the change in the pausing index (PI) of the upregulated genes that display promoter-proximal pausing (PI R1.2 in DMSO) upon SPI-21 treatment. The significance of the difference in the median values between control and treatment was calculated by the Wilcoxon signed-rank test. (F) Several examples showing GRO-seq reads aligned to the genome. DMSO and SPI-21 samples are presented in blue and red, respectively. The PIs, calculated for each gene, are shown at right. (G) Enrichment of regulatory sequences in SPI-21 upregulated genes (R2-fold change). Shown are the top 11 output sequences, along with their enrichment score and statistical significance. The analysis was done using a GSEA database and software (Subramanian et al., 2005). The asterisks denote statistical significance differences according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
Molecular Cell 76, 1–15, November 21, 2019 5
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A
A20 Induction fold
Induction fold
CCL20
IL6 Induction fold
Induction fold
IL8
JunB
B
Induction fold
IEX1
C
Relative A20 mRNA level
Induction fold
Induction fold
CXCL2
D
Induction fold
Average of A20, IL6, IL8, CXCL2
Relative occupancy
Relative occupancy
E
A
B
C
A
B
C
(legend on next page)
6 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
the strongly induced genes (compare A20, CCL20, IL-6, and IL-8 to CXCL2, JUNB, and IEX1 in Figure 3A). The effect of the weak inhibitors is remarkably similar to that seen with Spt5 depletion, which primarily diminished the induction of strongly induced genes (Diamant et al., 2016b). Using actinomycin D on SPI-21 and TNF-a-treated cells, we ruled out the possibility that the reduced mRNA levels are a consequence of diminished mRNA stability (Figure 3B). In fact, the half-life of the induced mRNA levels of A20 in control samples is shorter than that in the SPI21-treated cells. We confirmed that the observed effects are not a consequence of changes in Rpb1 and Spt5 protein levels caused by SPI-21 (Figure S1A). In addition, we found no significant alterations in the occupancy of the NF-kB protein p65 of its target genes upon SPI-21 treatment under basal and TNF-a conditions (Figure S1B). We determined the IC50 of a few SPIs on the TNF-a-induced transcription and found them to be mostly in the low micromolar range (Figure S1C). By summarizing the effect of the different SPI compounds on the selected genes, we obtained a ranking of their potency in inhibiting the negative and positive regulatory activities of Spt5 (Figures 2C and 3C). This analysis revealed that several compounds, including SPI-68, -31, -18, and -39 inhibited mainly the positive activity of Spt5, with little or no effect on promoter-proximal pausing, while SPI-06 predominantly affected promoter-proximal pausing. These results suggest that the positive and negative activities of Spt5 can be uncoupled. As Spt5 KD also reduced the splicing efficiency of the TNFa-induced A20 mRNA (Diamant et al., 2012, 2016b), we examined the effect of SPI-21 on splicing by determining the ratio between mature and immature mRNAs of several genes following TNF-a induction. The results revealed a clear decrease in the mature:immature ratio following treatment with SPI-21 (Figure 3D), which is consistent with the effect seen upon Spt5 KD. IkBa, the major inhibitor of NF-kB, is also a primary target gene of NF-kB itself, forming a negative feedback loop. Spt5 was shown to be central in maintaining this regulatory circuit (Diamant et al., 2012, 2016b). We therefore examined the effect of SPI-21 and SPI-18 on IkBa protein. In control cells, the initial levels of IkBa are diminished after 30 min of TNF-a induction as a consequence of its degradation and then recover after 150 min due to its induction by NF-kB (Figure S1D, DMSO).
Upon SPI-21 and SPI-18 treatment, the recovery of IkBa at 150 min is markedly reduced (Figure S1D, lanes 6 and 9 and corresponding columns), as observed in Spt5 KD cells (Diamant et al., 2012, 2016a). To examine further the impact of SPIs on the chromatin occupancy of Spt5 and Pol II, we performed chromatin immunoprecipitation (ChIP) in the presence of DMSO, SPI-21, and SPI-18, with or without 30 min of TNF-a treatment. Analysis of 3 loci of the A20 gene under basal conditions revealed a moderate but significant enhancement of Spt5 and Pol II occupancy at the 3 loci by SPI-21, while their levels were unchanged with SPI-18 (Figure 3E, basal). This finding is consistent with the increase in basal transcription exerted by SPI-21 but not by SPI-18 (Figure 2A). Upon TNF-a induction, the levels of Pol II and Spt5 are dramatically elevated, and this induction is diminished by both SPI-21 and SPI-18 at all 3 loci (Figure 3E, TNF-a), which correlates well with their effect on transcription (Figure 3A). The effects of SPIs on NF-kB target genes under basal and induced conditions are in full agreement with the previously reported negative and positive effects of Spt5 on pro-inflammatory genes, as revealed by the KD approach. We selected SPI-21 and SPI-18 to validate their direct interaction with either the Spt5 or the Rpb1 CC domain. His-Rpb1 and His-Spt5 (fused to Spt4 to increase its stability) and His-eIF4E, used as a negative control, were each expressed in E. coli and purified by nickel agarose beads. Using the purified proteins, we performed fluorescence intensity measurements in the presence of increasing concentrations of SPI-21 and SPI-18. This method analyzes binding by monitoring the changes in the intrinsic fluorescence of the proteins as a consequence of conformational changes. Spt5, Rpb1, and eIF4E display significant intrinsic fluorescence, as expected from the presence of aromatic residues in all of them. SPI-21 decreased the intrinsic fluorescence of Rpb1 in a dose-dependent manner with an IC50 of 23 mM, while it had no significant effect on Spt5 and eIF4E (Figure S2A). This IC50 is comparable to the 25 mM IC50 value seen in the split-RL assay with SP-21 (Figure 1C) and suggests that SPI-21 primarily binds the Rpb1 CC domain. SPI-18, however, seems to affect the fluorescence of both Spt5 and Rpb1, and the measured IC50 is 40 mM for Spt5 and 76 mM for Rpb1. The IC50 of Spt5 binding by SPI-18 is close to
Figure 3. Effect of SPIs on TNF-a-Induced Inflammatory Response Genes (A) Cells were incubated with DMSO or 50 mM indicated SPIs (or less, as indicated in Figure 2) for 1 h and then treated with TNF-a (20 ng/mL) for an additional 1 h. Then, RNA was extracted and analyzed for the indicated mRNAs by qRT-PCR and normalized to GAPDH. The results are presented as fold of induction relative to non-treated DMSO, which was set to 1. Bars represent the means ± SEMs of at least 3 independent experiments. The asterisks (as described for Figure 2A) denote statistically significant differences relative to DMSO + TNF-a. (B) The effect of SPI-21 on the stability of the TNF-a-induced A20 mRNA. Cells, pre-treated with SPI-21 and TNF-a, were incubated with actinomycin D for the indicated times, and A20 and GAPDH mRNA levels were determined by qRT-PCR. The data points represent the means ± SEMs of 3 independent experiments. (C) A graph summarizing the average effects of SPIs on the indicated genes organized in descending order of potency. (D) A dot plot presenting the effect of SPI-21 on the splicing efficiency of TNF-a-induced genes. Cells were incubated with DMSO or 50 mM SPI-21, as in (A). Following RNA extraction, cDNA was prepared using random hexamers, and levels of mature and immature transcripts were determined by qRT-PCR using different sets of primers (see scheme on top: red and green arrows for mature and immature transcripts, respectively). Each dot presents the ratio between mature and immature transcripts in each sample. The data represent the means ± SEMs of 3 independent experiments, and the asterisks denote statistically significant differences relative to DMSO + TNF-a. (E) SPIs effect on Pol II and Spt5 occupancies of the A20 gene. Cells, pre-treated with DMSO or SPI-21 and SPI-18 (50 mM) for 1 h, were induced with TNF-a for 30 min and then subjected to ChIP using anti-Rpb1 and anti-Spt5 antibodies. The occupancy of the A20 gene was analyzed by qRT-PCR using primers from the first (A), third (B), and last (C) exon-intron junctions. Bars represent the levels of Pol II and Spt5 relative to the input, and are the means ± SEMs of 3 independent experiments. The asterisks denote statistically significant difference compared to the DMSO counterpart with or without TNF-a; according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
Molecular Cell 76, 1–15, November 21, 2019 7
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A
B
TNF - induced genes
induced genes
11
31
9
SPI-21 up Genes (basal)
C
SPI-21 down genes TNF vs ctrl
D
SPI-21 vs ctrl
SPI-21 +TNF vs TNF
SPI-21 +TNF vs ctrl
E
HSPA1A
SPI-21 up-regulated genes under basal conditions
Relative mRNA levels
HSPA6
528
315
G Relative GDF15 level
Relative GDF15 level
F
93
Relative GDF15 level
24h
I Relative GDF15 mRNA level
24h
H
(legend on next page)
8 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
the values obtained with the split-RL (33 mM; Figure 1C) and with the inhibition of inflammatory genes (39 mM; Figure S1C). These findings are consistent with the direct interaction of the identified SPI compounds with the Spt5 and Rpb1 proteins. Next, we performed co-immunoprecipitation in the presence of SPI-21 and SPI-18 to examine their effect on the endogenous Spt4/Spt5-Pol II complex and found that these compounds did not disrupt their binding (Figure S2B). These findings are not surprising, as the binding of Spt5 to the multi-subunit Pol II involves multiple Spt5 domains, several subunits, and many contact points; thus, interference with one domain may not be sufficient to detach it from Pol II. In addition, SPI-21 and SPI-18 binding to Rpb1 and Spt5 are more likely to induce conformational changes that diminish Spt5-Pol II function rather than their binding. Spt5 was reported to be essential for cell growth and survival in mammalian cells (Komori et al., 2009). We therefore examined the effect of increasing concentrations of SPI-21 and SPI-18 on cell growth and found that both compounds dramatically diminish cellular viability in a dose-dependent manner (Figure S3). SPIs Unveil Kinetically Distinct Spt5 Target Genes and a Metabolic Function A major limitation of KD and knockout (KO) studies is the prolonged deficiency of the target protein. Small molecule inhibitors, on the other hand, inactivate their target soon after their application. To examine the global effect of a short-term Spt5 inhibition, cells were treated with SPI-21 for 2 h or with SPI-21 for 1 h and TNF-a for an additional 1 h and then subjected to RNA sequencing (RNA-seq). Reads were aligned to the human genome, and the normalized expression of each gene was determined. Subsequently, the following ratios were calculated: TNFa versus control, which describes the response to TNF-a; SPI-21 versus control, which describes the effect of SPI-21 on basal gene expression; and SPI-21 + TNF-a versus TNF-a, which describes the effect of SPI-21 on TNF-a induction rate and SPI21 + TNF-a versus control. A total of 506 genes displayed a statistically significant fold change in at least 1 of the ratios, and these were grouped into 5 distinct clusters (Figure 4A). The
vast majority of SPI-21-affected genes were upregulated as also seen with the GRO-seq. This is similar to the global effect seen upon Spt5 KD in human (Diamant et al., 2016b) and mouse (Fitz et al., 2018) cells. Clusters 1–4 contain genes that are upregulated by SPI-21 under basal conditions and cluster 5 contains SPI-21 downregulated genes (see SPI-21 versus control). Cluster 1 also consists of TNF-a-induced genes that are inhibited by SPI-21, while the TNF-a-induced genes in cluster 2 are not affected by it. We selected several genes from the different clusters to examine the effect of a few other SPIs on their expression by qRT-PCR (excluding cluster 1, which is represented in Figures 2 and 3). The changes in their levels are highly similar to those of SPI-21 (Figure S4). Analysis of the RNA-seq data provided several insights. There is a substantial overlap between genes prone to induction by TNF-a and those that are upregulated by SPI-21 under basal conditions, suggesting that promoter-proximal pausing is central to this signaling pathway (see clusters 1 and 2). Comparing the effect of SPI-21 with that of Spt5 KD (Diamant et al., 2016b) on the TNF-a-induced genes, we found that most of the downregulated genes are common to the two treatments (Figure 4B). In addition, biological pathway analysis of the SPI-21 upregulated genes revealed remarkable enrichment of inducible genes in response to a variety of extracellular signals other than TNF-a, including unfolded protein response, heat, hypoxia, cyclic AMP (cAMP), and mechanical stimulus (Figure 4C). Of particular interest are heat shock genes, which were very strongly upregulated. While heat shock genes are known Spt5 targets for promoter-proximal pausing in Drosophila (Missra and Gilmour, 2010; Wu et al., 2003), these genes were not identified as such by the Spt5 KD studies in mammalian cells (Diamant et al., 2016b). Activation of these genes is also apparent in the GRO-seq data (see Figures 2F and 2G). We raised the possibility that the effect of Spt5 inhibition on heat shock genes may be transient and therefore is not detected by the prolonged KD approach. To test this possibility, cells were treated with SPI-21 for 2 and 24 h and levels of heat shock genes HSPA6 and HSPA1A were determined by qRT-PCR. Both genes were strongly upregulated after 2 h,
Figure 4. Identification of Kinetically Distinct Spt5 Target Genes (A) The global effect of SPI-21 and TNF-a on mRNA levels. Cells were pre-treated with SPI-21 (50 mM) or DMSO for 1 h and then with TNF-a for an additional 1 h. Sequencing libraries from mRNA extracted from these cells were prepared and subjected to deep sequencing. Reads were aligned to the human genome, and the ratios indicated at the bottom were calculated between the samples. The resulting gene list (average of 2 experiments) was clustered into 5 distinct groups and is shown as a heatmap of the log fold change of differentially expressed genes. (B) A Venn diagram showing the overlap between SPI-21 downregulated genes from the TNF-a-induced set (presented in cluster 1) and the TNF-a-induced genes that were downregulated upon Spt5 KD retrieved from the previously published RNA-seq data (clusters 2 and 3 in Diamant et al., 2016b). (C) Gene set enrichment analysis (using the GeneAnalytics Gene Ontology [GO] function) of the biological processes associated with SPI-21 upregulated genes. (D) Analysis of the effects of short- and long-term treatments (2 and 24 h, respectively) of SPI-21 on the expression of heat shock genes HSPA6 and HSP1A1, using qRT-PCR (normalized to GAPDH). (E) Comparison of the global effects of long- and short-term SPI-21 treatment. Cells were treated with SPI-21 (50 mM) or DMSO for 24 h and subjected to RNAseq. The extent of overlap between the upregulated gene sets (R2-fold change) from 2 h (described in A) and 24 h was determined using a Venn diagram. (F) The change in GDF15 levels following 72-h Spt5 KD and 24-h SPI-21 treatment. The Spt5 KD data were retrieved from the previously published RNA-seq analysis (Diamant et al., 2016b). (G) Effect of short- and long-term treatments of SPI-21 on the expression of GDF15 mRNA analyzed by qRT-PCR and normalized to GAPDH. (H) The effect of 24-h treatment of the indicated SPIs on GDF15 levels analyzed by qRT-PCR and normalized to GAPDH (final concentrations: SPI-18, 50 mM; SPI06, -17, -29, and -95, 30 mM; SPI-42, 10 mM; SPI-86, 5 mM). (I) Analysis of the effect of 24-h SPI-21 and SPI-18 treatments on GDF15 mRNA levels in mouse embryonic fibroblasts (MEFs), MCF7, and Jurkat cells. The graphs in (D) and (G)–(I) represent the means ± SEMs of at least 3 independent experiments. The asterisks denote statistically significant differences relative to DMSO; according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
Molecular Cell 76, 1–15, November 21, 2019 9
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
but after 24 h their activation had considerably decreased (Figure 4D), indicating that the effect of Spt5 depletion on their expression is temporary, and thus explaining the absence of these genes from the Spt5 KD studies. To further test the idea of kinetically distinct Spt5 target genes, we determined the global effect of long-term SPI-21 treatment (24 h) by RNA-seq and compared the affected genes between the shortand long-term treatments. We found that 80% of the affected genes differed between the short- and long-term treatments (Figure 4E). Examination of the RNA-seq data of Spt5 KD cells (Diamant et al., 2016b) and the long-term SPI-21 treatment revealed GDF15 as one of the top upregulated genes, elevated by 27fold in the 2 experiments (Figure 4F). GDF15 is a member of the transforming growth factor b (TGF-b) superfamily and is associated with the regulation of body weight and food intake in humans and rodents (Emmerson et al., 2017; Hsu et al., 2017; Mullican et al., 2017; Tsai et al., 2018; Yang et al., 2017). GDF15 levels were almost unchanged by the 2-h SPI-21 treatment. To gain further support that the strong upregulation of GDF15 is associated with the prolonged inhibition of Spt5, we analyzed the GDF15 expression following 2 and 24 h of treatment with SPI-21 by using qRT-PCR. Contrary to the heat shock genes, GDF15 was 100-fold upregulated after 24 h, while only moderately upregulated after 2 h of SPI-21 treatment (Figure 4G). These findings support the idea that the strong upregulation of GDF15 requires persistent Spt5 inhibition and may be partially indirect. The dramatic elevation of GDF15 is also seen with other SPIs (Figure 4H), and it is not limited to HeLa cells but other human and mouse cell lines (Figure 4I). Thus, temporary and constitutive Spt5-regulated genes can be distinguished by using SPIs, which also unveiled a previously unknown but significant role for Spt5 in the control of an important mammalian metabolic pathway via GDF15. To validate that the effects of SPIs are at the transcriptional level, cellular RNA was metabolically labeled with 4-thiouridine (4sU) for 2 h in the presence of DMSO or SPI-21 in the presence or absence of TNF-a. Newly synthesized labeled RNA was then purified and subjected to RNA-seq. We determined the overlap between the SPI-21 and Spt5 KD upregulated genes of the TNF-a-responsive set, and found that 67% of Spt5 KD upregulated genes are common to the newly synthesized SPI-21 upregulated genes from this set (Figure S5A). Analysis of the effect of TNF-a induction in these data showed a remarkably high number of TNF-a-induced genes (467) compared to the conventional RNA-seq (<100; see Figure 4A and Diamant et al., 2016b). A substantial fraction of these genes (285 genes, 61%) is downregulated by SPI-21 (Figure S5B), while only 301 of 13,890 expressed genes (2.2%) are downregulated under basal conditions. Thus, the observed reduction in TNF-a activation by SPI-21 is highly specific and is a consequence of diminished transcription. Selective Inhibition of Mutant but Not Wild-Type Htt Gene Transcription by SPIs HD is an inherited genetic disorder caused by the expansion of trinucleotide CAG repeats (R36) encoding glutamine (Q) in the Htt gene. Expression of the mutant gene in neurons of the stria-
10 Molecular Cell 76, 1–15, November 21, 2019
tum and the frontal cortex leads to the degeneration of neuronal cells that control body movements, emotions, and intellect. Previous studies identified Spt4 and Spt5 as important players in promoting the transcription of mutant Htt gene containing long CAG repeats, but with little effect on short CAG stretches (Cheng et al., 2015; Liu et al., 2012). To examine the potential effect of SPIs in this context, we used a striatal cell line established from wild-type (WT) (Q7) or mutant (Q111) Htt knockin mice. In these cell lines, the coding region of the first exon of the mouse Htt gene was replaced with the same region of human WT (Q7) or mutant (Q111) Htt, as shown schematically in Figure 5A. We treated the 2 cell lines with SPIs for 48 h and then determined the levels of Htt mRNA. The results revealed that several SPIs selectively diminished the Q111 but not the Q7 Htt mRNA (Figure 5B). Notably, the effective concentration of SPI-86 and SPI-31 is 15-fold lower than the concentration needed for these drugs to diminish the proinflammatory gene expression (Figures 3 and S1C). To examine whether the selective effect on mutant Htt is at the transcriptional level, we performed metabolic labeling of cellular RNA with 4sU for 2 h in the presence of DMSO or SPI-21. Analysis of the RNA levels using primers from different regions of the Htt gene (shown in Figure 5A) revealed that levels of the newly synthesized Q111 but not Q7 Htt were diminished by SPI-21, and the extent of inhibition is greater compared to its effect on steady-state levels (Figure 5C), indicating that this inhibitor acts at the level of mRNA transcription. In addition, these data show that in controls (DMSO), the relative levels of the first intron are much lower in Q7 as compared to Q111 (first exon:intron ratio differs significantly in Q7 versus Q111, p = 0.05), suggesting that the splicing efficiency of the longer repeats is reduced. Moreover, the levels measured from the second exon relative to the first is somewhat higher in Q7 compared to Q111, and this small difference is observed all along the gene, a finding that is consistent with a modest reduction in transcription elongation. The reduction in the transcription elongation rate and in splicing are associated with reduced expression levels of the Q111 Htt mRNA relative to the Q7 (Figure 5D) and may be the underlying basis for the sensitivity of Q111 to Spt5 inactivation. We were intrigued by the observation that only a subset of SPIs affects mutant Htt expression. We therefore summarized the influence of the different SPIs on the various activities of Spt5 (Figure 5E). From this combined data, it is apparent that SPIs can be classified according to their effect on the various Spt5 activities. The first class, which includes SPI-21, -86, -29, -74, and -68, affect all of the examined Spt5 activities and can be considered general inhibitors. The second class affects a single activity: SPI-85 and SPI-09 diminished mutant Htt expression but have little or no effect on basal and induced inflammatory gene expression; SPI-39 inhibits only the induced inflammatory gene transcription, while SPI-06 predominantly affects promoter-proximal pausing. The largest class, which consists of SPI-17, -157, -46, -57, -16, -42, and -95, does not affect mutant Htt expression, but it does affect the basal and the induced inflammatory gene transcription (Figure 5E). These findings strongly suggest that there may be discrete mechanisms underlying the requirement of Spt5 for expanded repeats
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Figure 5. SPIs Selectively Inhibit the Transcription of the Mutant Htt Gene
A
Relative mRNA levels
B
C
D Relative Htt levels
Relative mRNA levels
4sU labeling
Q7
Q111
(A) A scheme showing the structure of WT (Q7) and mutant (Q111) Htt gene of the transgenic striatal cell lines. (B) Q7 and Q111 expressing cells were treated with SPIs at the indicated concentrations for 48 h. Then, RNA was extracted, and levels of Htt mRNA were determined by qRT-PCR and normalized to b-actin. (C) Q7 and Q111 cells were metabolically labeled with 5-thiouridine for 2 h in the presence or absence of SPI-21. Newly synthesized Htt transcript levels were determined by using primers from the indicated locations. (D) The Htt levels in Q7 and Q111 cells normalized to b-actin. Levels of Q7 Htt were set to 1. The data shown in (B)–(D) represent the means ± SEMs of at least 3 independent experiments. The asterisks denote statistically significant differences relative to DMSO (B and C) or to Q7 (D). (E) A table summarizing the effects of the indicated SPIs on the various Spt5 biological activities. The compounds are clustered according to their effects on the 3 different Spt5 activities indicated at top and are color-coded and numbered. For example, the compounds in cluster 1 inhibit all activities, while those in cluster 4 inhibit primarily mutant Htt expression. PPP, promoter-proximal pausing. Plus sign (+) indicates statistically significant effect. The asterisks denote statistical significance differences according to Student’s t tests (typically onetailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
E SPI
Inflammatory genes Basal (PPP)
Induced
Mutant Htt
transcription, for induction of inflammatory genes, and for promoter-proximal pausing. Spt5 Regulation of Histone Genes Is Linked to 30 End Processing Spt5 KD resulted in a relatively small number of downregulated genes (43 genes), of which 20 are replication-dependent histone
genes (Figure 6A), suggesting that Spt5 plays an important role in their expression. Histone genes are coordinately synthesized with DNA during S phase and do not end with a polyA tail but with a conserved stem-loop structure. We analyzed the 24h RNA-seq data for the effect of SPI-21 on histone gene expression and found them to be similarly downregulated (Figure 6A). Likewise, 24 h of SPI-18 treatment resulted in a decreased expression of selected histone genes (Figure S6). The requirement of Spt5 for the expression of histone genes cannot be explained by a positive effect on elongation or mRNA processing since histone genes are very short and are intronless. Moreover, analysis of the RNA-seq data of the 4sU metabolic labeling experiment actually revealed a moderate increase in histone gene transcription upon SPI-21 treatment (Figure 6A), suggesting that their synthesis during S phase is not affected. To resolve the discrepancy between the effect of Spt5 inhibition on steady-state levels and synthesis rate, we considered the possibility that Spt5 is required for the processing of histone mRNAs. Besides the conserved stemloop, histone genes have a canonical polyadenylation signal
Molecular Cell 76, 1–15, November 21, 2019 11
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A
B PolyA
HIST1H1B
Hist1H1B
Hist1H2AI
Hist1H3A
Hist1H4H
Total/Poly-A
HIST1H1E HIST1H2AG HIST1H2AH HIST1H2AI HIST1H2AK HIST1H2AL HIST1H2BF HIST1H3A HIST1H3B HIST1H3D Total/Poly-A
HIST1H3F HIST1H3G HIST1H4A HIST1H4B HIST1H4C HIST1H4D HIST1H4H HIST2H2AC HIST2H3D
C Hist1H2AI
Hist1H3A
Hist1H4H
Total/Poly-A
Total/Poly-A
Hist1H1B
Spt5 Tub
Figure 6. Regulation of 30 End Processing of Histone Genes by Spt5 (A) The levels of 20 replication-dependent histone genes following 72-h Spt5 KD, 24-h SPI-21 treatment, and 4sU metabolic labeling. All of the values were retrieved from the RNA-seq data carried out in this study and by Diamant et al. (2016b) and are shown as the mean of 2 experiments. In all of the experiments, cDNA was generated with random hexamers. (B) The ratio between total and poly(A) containing histone genes. A schematic illustration of the histone gene structure is shown at top. For the total and poly(A) mRNA analyses, cDNA was prepared with random hexamers or poly-deoxythymine (dT) oligonucleotide, respectively. The levels were normalized to GAPDH. (C) An analysis that is similar to that in (B) using RNA from control and Spt5 KD cells. The bottom panel shows a western blot of control and Spt5 KD cells using the indicated antibodies. The data in (B) and (C) represent the means ± SEMs of 3 independent experiments. The asterisks denote statistically significant differences relative to DMSO; according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005.
downstream of the stem-loop structure (Figure 6B). We therefore examined the possibility that Spt5 controls the choice between 30 end processing via the stem-loop pathway or the polyadenylation pathway of histone pre-mRNAs. By using cDNA prepared by random hexamer and Poly-T we determined the ratio between total and poly-A tailed histone genes. We found that relative to control cells, this ratio is significantly reduced, which is
12 Molecular Cell 76, 1–15, November 21, 2019
consistent with an increase in the relative amount of poly-A tailed histone genes upon Spt5 inhibition by the drugs (Figure 6B). A similar analysis with RNA extracted from Spt5 KD cells shows the same effect (Figure 6C). Thus, the reduction in the histone gene expression upon Spt5 inhibition either by the SPIs or KD may be associated, at least in part, with a defect in 30 end processing.
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
DISCUSSION In the present study, we report the development of a pharmacological approach to elucidate the function of Spt5, a central transcription elongation factor in mammalian cells. By targeting the functional Spt5-Rpb1 complex in a high-throughput drug screen, we identified several SPIs that faithfully reproduced the effect of Spt5 KD on the expression of basal and activated pro-inflammatory genes. Our findings with SPI-21 revealed regulatory sequences that potentially control the negative activity of Spt5. A few of these sequences are similar to sequences previously implicated in pausing in mammalian cells (Ainbinder et al., 2004; Amir-Zilberstein and Dikstein, 2008; Gromak et al., 2006; Yonaha and Proudfoot, 1999), but are distinct from those identified in Drosophila, one of which binds M1BP, a Drosophila exclusive factor (Lee et al., 2008; Li and Gilmour, 2013). Thus, the control of promoter-proximal pausing by Spt5 involves diverse mechanisms. Consistent with this notion, these inhibitors identified target genes that respond to Spt5 inhibition in profoundly diverse kinetics. While the relief of Spt5 inhibition from heat shock genes by SPIs resulted in their transient upregulation, the opposite is seen with GDF15, in which its upregulation was apparent only upon prolonged Spt5 inhibition. As heat shock genes were absent from Spt5 KD studies, this observation highlights the advantage of small molecules in discriminating between temporary (direct) and possibly indirect target genes. The dramatic upregulation of GDF15 upon Spt5 inhibition by siRNA or by SPIs points to the involvement of Spt5 in metabolic functions related to food intake and control of body weight. Furthermore, the use of these SPIs revealed a previously unknown role of Spt5 in promoting the unique 30 end processing of histone genes over use of the polyadenylation pathway. As earlier work implicated NELF in the 30 end processing of histone genes (Narita et al., 2007), it is plausible that DSIF and NELF cooperate in controlling this process. The examination of the effects of SPIs on the expression of WT and mutant Htt genes provided further support for the reported importance of Spt4/Spt5 for the transcription of the mutated expanded repeat Htt gene (Cheng et al., 2015; Liu et al., 2012). Notably, some of the identified compounds inhibited mutant Htt expression, with no apparent effect on the expression of WT Htt and pro-inflammatory genes. In addition, the effective concentration of 2 other compounds, SPI-86 and SPI-31, for inhibiting Htt expression was substantially lower than that required to diminish the expression of pro-inflammatory genes. These findings imply that the Htt-specific compounds may be useful as leads for therapy against HD and also suggest that the underlying basis of the Spt4/Spt5 requirement for Htt gene expression is distinct from that of inflammatory gene expression. This notion is supported by the finding that Spt4 is dispensable for pro-inflammatory gene expression (Diamant et al., 2016a), while it is crucial for mutant Htt expression (Cheng et al., 2015; Liu et al., 2012). Thus, the development of these compounds is expected to show therapeutic potential. A challenging aspect with SPIs relates to their specificity. While the issue could not be fully resolved in this study, our experiments provide several arguments on the subject. First, a major advantage of the screening approach used here is the use of
recombinant proteins, which led to the identification of compounds that directly interact with the target proteins. Second, the transcriptional effects of SPIs reported here are highly specific and in full agreement with the expected effects of Spt5 KD. Third, the transcriptomic data revealed that only a limited number of genes were affected and most of them were expected from Spt5 inhibition. Fourth, the effect of the compounds on Htt gene expression is highly specific to long but not to short expanded repeats. Nevertheless, the effect of these compounds on cellular processes other than transcription cannot be ruled out. In summary, we have discovered the first specific Spt5-Pol II inhibitors by applying a conformation-sensitive protein-protein interaction drug screen. The use of SPIs validated the known activities exerted by Spt5 in mammalian cells and helped to reveal novel functions. The identified compounds provide an excellent approach for assessing the function of Spt5 in complex in vivo settings. Further development of SPIs is expected to lead to their preclinical and clinical evaluations as a therapeutic approach against metabolic and neurodegenerative disorders. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d d
d d
KEY RESOURCES TABLE LEAD CONTACT AND MATERIALS AVAILABILITY EXPERIMENTAL MODEL AND SUBJECT DETAILS METHOD DETAILS B Manipulations in cultured cells B Plasmid construction B High-throughput drug screening (HTS) B RNA preparation, RNA metabolic labeling, and Global Run-On B High-throughput sequencing of the total, labeled and Run-On transcripts (RNA-Seq and Gro-seq) B RNA-Seq and GRO-seq Bioinformatics analysis B Chromatin immunoprecipitation (ChIP) B Cells extract, western blotting B Co-immunoprecipitation (CO-IP) B Binding assay B Cells viability assay QUANTIFICATION AND STATISTICAL ANALYSIS DATA AND CODE AVAILABILITY
SUPPLEMENTAL INFORMATION Supplemental Information can be found online at https://doi.org/10.1016/j. molcel.2019.08.024. ACKNOWLEDGMENTS This work was supported in part by grants from the Nella and Leon Benoziyo Center for Neurological Diseases, Yeda Research and Development Co. Ltd., the Estate of David Levinson, the Estate of Fannie Sherr, and the Israel Innovation Authority (KAMIN). We would like to thank Galit Cohen from the Maurice and Vivienne Wohl Institute for Drug Discovery of the Nancy and Stephen Grand Israel National Center for Personalized Medicine (G-INCPM) (Weizmann Institute of Science) for contributions to the drug screen, Dr. Sima Benjamin
Molecular Cell 76, 1–15, November 21, 2019 13
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
and Dr. Shlomit Gilad from the Crown Institute for Genomics of the G-INCPM for libraries preparation and RNA-seq services, Michael Gershovits from the Ilana and Pascal Mantoux Institute for Bioinformatics of the G-INCPM, Dr. Noa Wigoda from the Weizmann core facilities for the RNA-seq data analysis, and Mr. Arsenii Hordeichyk for technical assistance in the binding assays. R.D. is the incumbent of the Ruth and Leonard Simon Chair of Cancer Research. AUTHOR CONTRIBUTIONS R.D. and A.B. conceived and designed the study, analyzed the data, and wrote the paper. A.B. carried out most of the experiments. O.L. provided technical assistance. A.P. and A.B. carried out the high-throughput drug screen. D.L. performed part of the bioinformatics analysis. DECLARATION OF INTERESTS R.D. and A.B. declare a patent application (pending) related to this work. Received: September 13, 2018 Revised: June 12, 2019 Accepted: August 26, 2019 Published: September 26, 2019 REFERENCES Aida, M., Chen, Y., Nakajima, K., Yamaguchi, Y., Wada, T., and Handa, H. (2006). Transcriptional pausing caused by NELF plays a dual role in regulating immediate-early expression of the junB gene. Mol. Cell. Biol. 26, 6094–6104. Ainbinder, E., Revach, M., Wolstein, O., Moshonov, S., Diamant, N., and Dikstein, R. (2002). Mechanism of rapid transcriptional induction of tumor necrosis factor alpha-responsive genes by NF-kappaB. Mol. Cell. Biol. 22, 6354–6362. Ainbinder, E., Amir-Zilberstein, L., Yamaguchi, Y., Handa, H., and Dikstein, R. (2004). Elongation inhibition by DRB sensitivity-inducing factor is regulated by the A20 promoter via a novel negative element and NF-kappaB. Mol. Cell. Biol. 24, 2444–2454. Amir-Zilberstein, L., and Dikstein, R. (2008). Interplay between E-box and NF-kappaB in regulation of A20 gene by DRB sensitivity-inducing factor (DSIF). J. Biol. Chem. 283, 1317–1323. Amir-Zilberstein, L., Ainbinder, E., Toube, L., Yamaguchi, Y., Handa, H., and Dikstein, R. (2007). Differential regulation of NF-kappaB by elongation factors is determined by core promoter type. Mol. Cell. Biol. 27, 5246–5259. Anders, S., Pyl, P.T., and Huber, W. (2015). HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169. Ashkenazi, S., Plotnikov, A., Bahat, A., Ben-Zeev, E., Warszawski, S., and Dikstein, R. (2016). A Novel Allosteric Mechanism of NF-kB Dimerization and DNA Binding Targeted by an Anti-Inflammatory Drug. Mol. Cell. Biol. 36, 1237–1247. Ashkenazi, S., Plotnikov, A., Bahat, A., and Dikstein, R. (2017). Effective cellfree drug screening protocol for protein-protein interaction. Anal. Biochem. 532, 53–59. Bernecky, C., Herzog, F., Baumeister, W., Plitzko, J.M., and Cramer, P. (2016). Structure of transcribing mammalian RNA polymerase II. Nature 529, 551–554. Blythe, A.J., Yazar-Klosinski, B., Webster, M.W., Chen, E., Vandevenne, M., Bendak, K., Mackay, J.P., Hartzog, G.A., and Vrielink, A. (2016). The yeast transcription elongation factor Spt4/5 is a sequence-specific RNA binding protein. Protein Sci. 25, 1710–1721. Boehm, A.K., Saunders, A., Werner, J., and Lis, J.T. (2003). Transcription factor and polymerase recruitment, modification, and movement on dhsp70 in vivo in the minutes following heat shock. Mol. Cell. Biol. 23, 7628–7637. Chen, H., Contreras, X., Yamaguchi, Y., Handa, H., Peterlin, B.M., and Guo, S. (2009). Repression of RNA polymerase II elongation in vivo is critically dependent on the C-terminus of Spt5. PLoS One 4, e6918. Cheng, H.M., Chern, Y., Chen, I.H., Liu, C.R., Li, S.H., Chun, S.J., Rigo, F., Bennett, C.F., Deng, N., Feng, Y., et al. (2015). Effects on murine behavior
14 Molecular Cell 76, 1–15, November 21, 2019
and lifespan of selectively decreasing expression of mutant huntingtin allele by supt4h knockdown. PLoS Genet. 11, e1005043. Core, L.J., Waterfall, J.J., and Lis, J.T. (2008). Nascent RNA sequencing reveals widespread pausing and divergent initiation at human promoters. Science 322, 1845–1848. Crickard, J.B., Fu, J., and Reese, J.C. (2016). Biochemical Analysis of Yeast Suppressor of Ty 4/5 (Spt4/5) Reveals the Importance of Nucleic Acid Interactions in the Prevention of RNA Polymerase II Arrest. J. Biol. Chem. 291, 9853–9870. Diamant, G., and Dikstein, R. (2013). Transcriptional control by NF-kB: elongation in focus. Biochim. Biophys. Acta 1829, 937–945. Diamant, G., Amir-Zilberstein, L., Yamaguchi, Y., Handa, H., and Dikstein, R. (2012). DSIF restricts NF-kB signaling by coordinating elongation with mRNA processing of negative feedback genes. Cell Rep. 2, 722–731. Diamant, G., Bahat, A., and Dikstein, R. (2016a). The elongation factor Spt5 facilitates transcription initiation for rapid induction of inflammatory-response genes. Nat. Commun. 7, 11547. Diamant, G., Eisenbaum, T., Leshkowitz, D., and Dikstein, R. (2016b). Analysis of Subcellular RNA Fractions Revealed a Transcription-Independent Effect of Tumor Necrosis Factor Alpha on Splicing, Mediated by Spt5. Mol. Cell. Biol. 36, 1342–1353. Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., Chaisson, M., and Gingeras, T.R. (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. Edgar, R., Domrachev, M., and Lash, A.E. (2002). Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 30, 207–210. Emmerson, P.J., Wang, F., Du, Y., Liu, Q., Pickard, R.T., Gonciarz, M.D., Coskun, T., Hamang, M.J., Sindelar, D.K., Ballman, K.K., et al. (2017). The metabolic effects of GDF15 are mediated by the orphan receptor GFRAL. Nat. Med. 23, 1215–1219. Fitz, J., Neumann, T., and Pavri, R. (2018). Regulation of RNA polymerase II processivity by Spt5 is restricted to a narrow window during elongation. EMBO J. 37, e97965. Fujita, T., Piuz, I., and Schlegel, W. (2009). Negative elongation factor NELF controls transcription of immediate early genes in a stimulus-specific manner. Exp. Cell Res. 315, 274–284. Gromak, N., West, S., and Proudfoot, N.J. (2006). Pause sites promote transcriptional termination of mammalian RNA polymerase II. Mol. Cell. Biol. 26, 3986–3996. Haimov, O., Sehrawat, U., Tamarkin-Ben Harush, A., Bahat, A., Uzonyi, A., Will, A., Hiraishi, H., Asano, K., and Dikstein, R. (2018). Dynamic interactions of eIF4G1 with eIF4E and eIF1 underlie scanning dependent and independent translation. Mol. Cell Biol. 28, e00139–18. Henriques, T., Scruggs, B.S., Inouye, M.O., Muse, G.W., Williams, L.H., Burkholder, A.B., Lavender, C.A., Fargo, D.C., and Adelman, K. (2018). Widespread transcriptional pausing and elongation control at enhancers. Genes Dev. 32, 26–41. Hirtreiter, A., Damsma, G.E., Cheung, A.C., Klose, D., Grohmann, D., Vojnic, E., Martin, A.C., Cramer, P., and Werner, F. (2010). Spt4/5 stimulates transcription elongation through the RNA polymerase clamp coiled-coil motif. Nucleic Acids Res. 38, 4040–4051. Hsu, J.Y., Crawley, S., Chen, M., Ayupova, D.A., Lindhout, D.A., Higbee, J., Kutach, A., Joo, W., Gao, Z., Fu, D., et al. (2017). Non-homeostatic body weight regulation through a brainstem-restricted receptor for GDF15. Nature 550, 255–259. Ivanov, D., Kwak, Y.T., Guo, J., and Gaynor, R.B. (2000). Domains in the SPT5 protein that modulate its transcriptional regulatory properties. Mol. Cell. Biol. 20, 2970–2983. Kim, D.K., Inukai, N., Yamada, T., Furuya, A., Sato, H., Yamaguchi, Y., Wada, T., and Handa, H. (2003). Structure-function analysis of human Spt4: evidence that hSpt4 and hSpt5 exert their roles in transcriptional elongation as parts of the DSIF complex. Genes Cells 8, 371–378.
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Klein, B.J., Bose, D., Baker, K.J., Yusoff, Z.M., Zhang, X., and Murakami, K.S. (2011). RNA polymerase and transcription elongation factor Spt4/5 complex structure. Proc. Natl. Acad. Sci. USA 108, 546–550. Kohen, R., Barlev, J., Hornung, G., Stelzer, G., Feldmesser, E., Kogan, K., Safran, M., and Leshkowitz, D. (2019). UTAP: User-friendly Transcriptome Analysis Pipeline. BMC Bioinformatics 20, 154. Komori, T., Inukai, N., Yamada, T., Yamaguchi, Y., and Handa, H. (2009). Role of human transcription elongation factor DSIF in the suppression of senescence and apoptosis. Genes Cells 14, 343–354. Kramer, N.J., Carlomagno, Y., Zhang, Y.J., Almeida, S., Cook, C.N., Gendron, T.F., Prudencio, M., Van Blitterswijk, M., Belzil, V., Couthouis, J., et al. (2016). Spt4 selectively regulates the expression of C9orf72 sense and antisense mutant transcripts. Science 353, 708–712. Lee, C., Li, X., Hechmer, A., Eisen, M., Biggin, M.D., Venters, B.J., Jiang, C., Li, J., Pugh, B.F., and Gilmour, D.S. (2008). NELF and GAGA factor are linked to promoter-proximal pausing at many genes in Drosophila. Mol. Cell. Biol. 28, 3290–3300. Li, J., and Gilmour, D.S. (2013). Distinct mechanisms of transcriptional pausing orchestrated by GAGA factor and M1BP, a novel transcription factor. EMBO J. 32, 1829–1841. Liu, C.R., Chang, C.R., Chern, Y., Wang, T.H., Hsieh, W.C., Shen, W.C., Chang, C.Y., Chu, I.C., Deng, N., Cohen, S.N., and Cheng, T.H. (2012). Spt4 is selectively required for transcription of extended trinucleotide repeats. Cell 148, 690–701. Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550.
Quinlan, A.R., and Hall, I.M. (2010). BEDTools: a flexible suite of utilities for 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. Rougvie, A.E., and Lis, J.T. (1990). Postinitiation transcriptional control in Drosophila melanogaster. Mol. Cell. Biol. 10, 6041–6045. Shetty, A., Kallgren, S.P., Demel, C., Maier, K.C., Spatt, D., Alver, B.H., Cramer, P., Park, P.J., and Winston, F. (2017). Spt5 Plays Vital Roles in the Control of Sense and Antisense Transcription Elongation. Mol. Cell 66, 77–88.e5. Stanlie, A., Begum, N.A., Akiyama, H., and Honjo, T. (2012). The DSIF subunits Spt4 and Spt5 have distinct roles at various phases of immunoglobulin class switch recombination. PLoS Genet. 8, e1002675. Subramanian, A., Tamayo, P., Mootha, V.K., Mukherjee, S., Ebert, B.L., Gillette, M.A., Paulovich, A., Pomeroy, S.L., Golub, T.R., Lander, E.S., and Mesirov, J.P. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102, 15545–15550. Tsai, V.W., Zhang, H.P., Manandhar, R., Lee-Ng, K.K.M., Lebhar, H., Marquis, C.P., Husaini, Y., Sainsbury, A., Brown, D.A., and Breit, S.N. (2018). Treatment with the TGF-b superfamily cytokine MIC-1/GDF15 reduces the adiposity and corrects the metabolic dysfunction of mice with diet-induced obesity. Int. J. Obes. 42, 561–571. Vos, S.M., Farnung, L., Urlaub, H., and Cramer, P. (2018). Structure of paused transcription complex Pol II-DSIF-NELF. Nature 560, 601–606.
Martinez-Rucobo, F.W., Sainsbury, S., Cheung, A.C., and Cramer, P. (2011). Architecture of the RNA polymerase-Spt4/5 complex and basis of universal transcription processivity. EMBO J. 30, 1302–1310.
Wada, T., Takagi, T., Yamaguchi, Y., Ferdous, A., Imai, T., Hirose, S., Sugimoto, S., Yano, K., Hartzog, G.A., Winston, F., et al. (1998a). DSIF, a novel transcription elongation factor that regulates RNA polymerase II processivity, is composed of human Spt4 and Spt5 homologs. Genes Dev. 12, 343–356.
Mayer, A., Lidschreiber, M., Siebert, M., Leike, K., So¨ding, J., and Cramer, P. (2010). Uniform transitions of the general RNA polymerase II transcription complex. Nat. Struct. Mol. Biol. 17, 1272–1278.
Wada, T., Takagi, T., Yamaguchi, Y., Watanabe, D., and Handa, H. (1998b). Evidence that P-TEFb alleviates the negative effect of DSIF on RNA polymerase II-dependent transcription in vitro. EMBO J. 17, 7395–7403.
Missra, A., and Gilmour, D.S. (2010). Interactions between DSIF (DRB sensitivity inducing factor), NELF (negative elongation factor), and the Drosophila RNA polymerase II transcription elongation complex. Proc. Natl. Acad. Sci. USA 107, 11301–11306.
Werner, F. (2012). A nexus for gene expression-molecular mechanisms of Spt5 and NusG in the three domains of life. J. Mol. Biol. 417, 13–27.
Mullican, S.E., Lin-Schmidt, X., Chin, C.N., Chavez, J.A., Furman, J.L., Armstrong, A.A., Beck, S.C., South, V.J., Dinh, T.Q., Cash-Mason, T.D., et al. (2017). GFRAL is the receptor for GDF15 and the ligand promotes weight loss in mice and nonhuman primates. Nat. Med. 23, 1150–1157. Narita, T., Yung, T.M., Yamamoto, J., Tsuboi, Y., Tanabe, H., Tanaka, K., Yamaguchi, Y., and Handa, H. (2007). NELF interacts with CBC and participates in 30 end processing of replication-dependent histone mRNAs. Mol. Cell 26, 349–365. Ni, Z., Saunders, A., Fuda, N.J., Yao, J., Suarez, J.R., Webb, W.W., and Lis, J.T. (2008). P-TEFb is critical for the maturation of RNA polymerase II into productive elongation in vivo. Mol. Cell. Biol. 28, 1161–1170. Pavri, R., Gazumyan, A., Jankovic, M., Di Virgilio, M., Klein, I., AnsarahSobrinho, C., Resch, W., Yamane, A., Reina San-Martin, B., Barreto, V., et al. (2010). Activation-induced cytidine deaminase targets DNA at sites of RNA polymerase II stalling by interaction with Spt5. Cell 143, 122–133. Peng, J., Marshall, N.F., and Price, D.H. (1998). Identification of a cyclin subunit required for the function of Drosophila P-TEFb. J. Biol. Chem. 273, 13855–13860.
Wu, C.H., Yamaguchi, Y., Benjamin, L.R., Horvat-Gordon, M., Washinsky, J., Enerly, E., Larsson, J., Lambertsson, A., Handa, H., and Gilmour, D. (2003). NELF and DSIF cause promoter proximal pausing on the hsp70 promoter in Drosophila. Genes Dev. 17, 1402–1414. Yamada, T., Yamaguchi, Y., Inukai, N., Okamoto, S., Mura, T., and Handa, H. (2006). P-TEFb-mediated phosphorylation of hSpt5 C-terminal repeats is critical for processive transcription elongation. Mol. Cell 21, 227–237. Yamaguchi, Y., Takagi, T., Wada, T., Yano, K., Furuya, A., Sugimoto, S., Hasegawa, J., and Handa, H. (1999a). NELF, a multisubunit complex containing RD, cooperates with DSIF to repress RNA polymerase II elongation. Cell 97, 41–51. Yamaguchi, Y., Wada, T., Watanabe, D., Takagi, T., Hasegawa, J., and Handa, H. (1999b). Structure and function of the human transcription elongation factor DSIF. J. Biol. Chem. 274, 8085–8092. Yang, L., Chang, C.C., Sun, Z., Madsen, D., Zhu, H., Padkjær, S.B., Wu, X., Huang, T., Hultman, K., Paulsen, S.J., et al. (2017). GFRAL is the receptor for GDF15 and is required for the anti-obesity effects of the ligand. Nat. Med. 23, 1158–1166. Yonaha, M., and Proudfoot, N.J. (1999). Specific transcriptional pausing activates polyadenylation in a coupled in vitro system. Mol. Cell 3, 593–600.
Molecular Cell 76, 1–15, November 21, 2019 15
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
STAR+METHODS KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
Antibodies Purified Mouse Anti-Human IkBa
BD-transductions
Cat#610690; RRID: AB_398013
Mouse monoclonal BrdU antibody (IIB5) conjugated to agarose
Santa Cruz
Cat#sc-32323AC; RRID: AB_626766
Mouse monoclonal c-Rel Antibody (B-6)
Santa Cruz
Cat#sc-6955; RRID: AB_670194
Mouse monoclonal NFkB p65 Antibody (F-6)
Santa Cruz
Cat#sc-8008; RRID: AB_628017
Normal mouse IgG
Santa Cruz
Cat#sc-2025; RRID: AB_737182
Mouse monoclonal [8WG16] to RNA polymerase II CTD repeat YSPTSPS - ChIP Grade
Abcam
Cat#ab817; RRID: AB_306327
Rat monoclonal antibody against DSIF p160
Wada et al., 1998a
N/A
Mouse monoclonal antibody against hSpt4
Kim et al., 2003
N/A
Recombinant Human TNF-a
Peprotech
Cat#300-01A-50
Coelenterazine, native
Gold Biotechnology
Cat#CZ5; CAS: 55779-48-1
BioTri (RNA/DNA/protein extraction using Trireagent)
Bio-Lab
Cat#959758027100
4-Thiouridine
Cayman Chemical
Cat#16373-25
Biotin-HPDP
APExBIO
Cat#A8008; CAS: 129179-83-5
Streptavidin-coated magnetic beads
NEB
Cat#S1420S
IGEPAL CA-630
Sigma
Cat#I8896; CAS: 9002-93-1
5-Bromouridine 5-triphosphate sodium salt
Santa Cruz
Cat#SC-214314A; CAS: 161848-60-8
RNA Fragmentation reagent
Thermo Fisher Scientific
Cat#AM8740
Actinomycin D
APExBIO
Cat#A4448; CAS: 50-76-0
Chemicals, Peptides, and Recombinant Proteins
D-Luciferin Firefly, Potassium Salt
Gold Bioscience
Cat#LUCK-100; CAS: 115144-35-9
nProtein A Sepharose 4 Fast Flow
GE Healthcare
Cat#17-5280-01
Protein G Sepharose 4 Fast Flow
GE Healthcare
Cat#17-0618-01
Reporter lysis buffer
Promega
Cat#E4030
Salmon sperm DNA
R&D Systems
Cat#9610-5-D
Ni-NTA His,Bind Resin
Merck Millipore
Cat#70666
tRNA
Sigma
Cat#R8508
Critical Commercial Assays High-Capacity cDNA Reverse Transcription kit
Thermo Fisher Scientific
Cat#4368814
Fast qPCRBIO SYBR Green Mix
PCR Biosystems
Cat#PB20.16-50
CellTiter-Glo Lum Cell Viability kit
Promega
Cat#G7571
Micro Bio-Spin P-30 Gel Columns
Bio-Rad
Cat#7326250
QIAquick PCR Purification Kit
QIAGEN
Cat#28106
RNA Clean & Concentrator kit
Zymo Research
Cat#ZR-R1017
Direct-zol RNA MiniPrep kit
Zymo Research
Cat#R2052
Raw and analyzed data of RNA-seq 2h
This paper
GEO:GSE136022
Raw and analyzed data of RNA-seq 24h
This paper
GEO:GSE136023
Raw and analyzed data of GRO-Seq
This paper
GEO:GSE136024
Raw and analyzed data of 4sU-seq
This paper
GEO:GSE136025
Transgenic mice: STHdhQ7/Q7
Coriell Institute for Medical Research
Cat# CH00097
Transgenic mice: STHdhQ111/Q111
Coriell Institute for Medical Research
Cat# CH00095
Deposited Data
Experimental Models: Cell Lines
(Continued on next page)
e1 Molecular Cell 76, 1–15.e1–e4, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Continued REAGENT or RESOURCE
SOURCE
IDENTIFIER
MCF7
ATCC
N/A
HeLa
ATCC
N/A
HEK293T
ATCC
N/A
Mouse embryonic fibroblasts (MEFs)
ATCC
N/A
Jurkat
ATCC
N/A
This paper
Table S1
Plasmid: RSV-DC/DN Renilla luciferase (pRL-DC/DN)
Ashkenazi et al., 2017
N/A
Plasmid: pRSFDuet-1 DNA
Merck-Novagen
Cat# 71341
Plasmid: DC RL-spt5
This paper
N/A
Oligonucleotides Primers for RF-cloning, ChIP and real-time see Table S1 Recombinant DNA
Plasmid: Pol II-DN RL
This paper
N/A
Plasmid: His-DC RL-spt5/Flag-Pol II-DN RL in pRSFDuet
This paper
N/A
Software and Algorithms GeneData software GSEA database and software
https://www.genedata.com/products/ profiler/software/ Subramanian et al., 2005
http://software.broadinstitute.org/gsea/ msigdb/index.jsp
LI-COR Image Studio Software
https://www.licor.com/bio/image-studio/
OriginLab for IC50 calculation
https://www.originlab.com/
Others RNA-Seq data from TNFa and Spt5 KD HeLa cells
Diamant et al., 2016b
GEO:GSE70268
LEAD CONTACT AND MATERIALS AVAILABILITY Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contacts (Rivka Dikstein). EXPERIMENTAL MODEL AND SUBJECT DETAILS HEK293T, HeLa, MCF7, Jurkat and mouse embryonic fibroblasts (MEFs) cells were grown and maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum (Invitrogen) and 1% penicillin-streptomycin. To avoid basal NF-lB activity, HeLa cells were kept from reaching confluence and re-plated after initial thawing no more than 10 times. To induce NF-lB activity, cells were treated with 20 ng/ml recombinant human TNF-a (Peprotech) for the indicated times. STHdh Q111 and STHdh Q7 striatal cell lines (Coriell Institute for Medical Research) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, 1% penicillin-streptomycin and 0.4 mg/ml G418 at 33C. The cells were re-plated no more than 5-6 times, as recommended. METHOD DETAILS Manipulations in cultured cells Spt5 knockdown (KD) was carried out as previously described (Diamant et al., 2012). Briefly, 1,250,000 cells were plated in 10cm dish and transfected 24h later using ICAFectin441 (In-Cell-Art) and 12 mg of either pSUPER or a mix of two distinct pSUPER-spt5-RNAi, together with 0.5 mg CMV-GFP-Puro plasmid. Transfected cells were selected for 24h with puromycin (1 mg/ml). Plasmid construction The mammalian expression plasmids RSV-DC/DN Renilla luciferase (pRL-DC/DN), encoding the N-terminal amino acid residues 1-229 or the C terminus 230-311 of the Renilla luciferase were previously described (Ashkenazi et al., 2016, 2017). The split-Renilla luciferase fusion plasmids were constructed by two-steps PCR using the pRL-DC or pRL-DN as backbones and plasmids encoding
Molecular Cell 76, 1–15.e1–e4, November 21, 2019 e2
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
human Rpb1, Spt5 and Spt4 as insert sources. For bacterial expression of the split-RL fusion proteins, the double expression plasmid pRSFDuet was used. All constructs were verified by sequencing. The primer sequences are shown in Table S1. High-throughput drug screening (HTS) The HTS was carried out in 1536 well plate format as previously described (Ashkenazi et al., 2017), with the following modifications: the final concentration of each drug was 15 mM and the incubation time with the Coelenterazine (GOLDBIO, CZ5) RL substrate was 30 min at RT. Compounds that diminished the luminescence activity by > 30% were subjected for an additional screen with bacterial lysate expressing recombinant full-length RL (diluted 1:500 in phosphate buffer) that served as a control for false-positive compounds that inhibit the RL enzymatic activity itself. Compounds that did not affect RL activity were further analyzed in a doseresponse assay (0.55 mM, 1.6 mM, 5 mM, 15 mM and 45 mM) in duplicate. The selected compounds were further subjected to livecell split-RL assay as described (Ashkenazi et al., 2017). The analysis of the luminescence data for all the above-described assays was performed using GeneData software (Basel, Switzerland). RNA preparation, RNA metabolic labeling, and Global Run-On Total RNA was extracted using Tri-reagent (Bio-Lab Chemicals) and Direct-zol RNA MiniPrep kit (Zymo Research). For gene-specific analysis cDNA was synthesized from 0.5-1 mg of total RNA using the High-Capacity cDNA Reverse Transcription Kit (ABI, Thermo Fisher Scientific) with a poly-dT primer or random hexamers as indicated for each experiment. cDNA samples were analyzed by quantitative PCR (qPCR) in a ViiA 7 Real-Time PCR System using Fast qPCRBIO SYBR Green Mix (PCR Biosystems). The primer sequences are shown in Table S1. For metabolic labeling of newly synthesized RNA with 4-thiouridine (4sU), cells were treated with SPI-21 or DMSO for 30 min. Then 4sU (150 mM final concentration) was added for an additional 2h. The labeling was terminated by removing the 4sU-containing medium and washing the cells 3 times with phosphate-buffered saline (PBS). Total RNA was extracted as described above. 10 mg of RNA were biotinylated using biotin-HPDP (1:5; A8008, APExBIO) for 1.5 h at room temperature. The biotinylated RNA was purified using chloroform and isopropanol extraction and dissolved in DEPC water. Thereafter the biotinylated RNA was captured using Streptavidin-coated magnetic beads (NEB, S1420S) for 30 min at room temperature in constant rotating. Bound RNA was eluted with 100 mL of 0.1 M fresh DTT. RNA isolation was performed using the RNA Clean & Concentrator kit (Zymo Research). cDNA was prepared as described above using random hexamers. For the Global Run-On experiment, HeLa cells (15 cm plate x 3) were treated with SPI-21 (50 mM) for 1.5h, washed 3 times with cold PBS, collected and pelleted at 1000 X g for 5 min at 4 C. Cells were swollen by resuspension in 2 mL swelling buffer [10 mM Tris-Cl pH 7.5, 10% glycerol, 3 mM CaCl2, 3 mM MgCl2, protease inhibitor cocktail (EDTA free), and 4 units/ml of RNase inhibitor], incubated at 4 C for 20 min and pelleted again at 1000 X g for 5 min at 4 C. Cells were resuspended in 1 mL of ice-cold lysis buffer (10 mM Tris-Cl pH7.5, 300 mM sucrose, 10 mM NaCl, 3 mM CaCl2, 2 mM MgCl2, 0.5% Igepal, 0.5 mM DTT, protease inhibitors and RNase inhibitor) and then homogenized and lysed by transferring them 5 times through 27G needle. Nuclei were pelleted at 1000 X g for 5 min at 4 C, washed once with 10 mL of lysis buffer and once with 1 mL storage buffer (50 mM Tris-Cl pH 8.0, 25% glycerol, 5 mM MgCl2, 0.1 mM EDTA, 5 mM DTT). For the Run-on reaction nuclei were resuspended at 5 3 10^6 nuclei/100 ml of storage buffer and mixed with equal volume of reaction buffer (10 mM Tris-Cl pH 8.0, 5 mM MgCl2, 1 mM DTT, 300 mM KCl, 20 units of RNase Inhibitor, 1% sarkosyl, 500 mM ATP, GTP, CTP and Br-UTP). The reaction was allowed to proceed for 5 min at 30 C, followed by the addition of 1 mL Trizol. RNA was further extracted with Direct-zol RNA MiniPrep kit and DNase I treatment was performed. Fragmentation of NRO-RNA was performed using fragmentation reagents (AM8740) for 10 min at +70 C and then terminated by addition of stop solution (1:1). The fragments were purified by Micro Bio-Spin Columns with Bio-Gel P-30 (BioRad) according to the manufacturer’s instructions. BrU-labeled RNA was purified with pre-blocked anti-BrdU beads and eluted 4 times (10 min each) with elution buffer (20 mM DTT, 300 mM NaCl, 5 mM Tris-cl pH 7.5, 1 mM EDTA, and 0.1% SDS) at +42 C while shaking. Eluted RNA was extracted using Trizol and Direct-zol RNA MiniPrep. High-throughput sequencing of the total, labeled and Run-On transcripts (RNA-Seq and Gro-seq) The RNA-seq libraries were prepared by the Genomics unit in The Nancy and Stephen Grand Israel National Center for Personalized Medicine (G-INCPM, Weizmann Institute of Science). Samples from 2h SPI-21 treatment were prepared using in-house mRNA Seq protocol and polyA capturing and those from 24h and 4sU treatments were prepared using the same protocol w/o PolyA capture (with random hexamers), and sequenced using an Illumina HiSeq 2500 system yielding 30M SE reads of 61 bases. The GRO-seq RNA samples were reverse transcribed and amplified (15 cycles) using in-house mRNA Seq protocol w/o PolyA capture (with random hexamers) and sequenced in one lane using an Illumina HiSeq High Output (480 million reads, yielding more than 100M SE reads of 47 bases per sample). Each RNA-seq analysis was carried out using two independent replicates. RNA-Seq and GRO-seq Bioinformatics analysis For the analysis of the 24h, 4sU and GRO-seq data, we used the RNA-seq pipeline of UTAP (Kohen et al., 2019) to determine the differentially expressed genes. For the 2h data, reads were mapped to the H. sapiens GRCh38 reference genome using STAR (Dobin et al., 2013) and expression levels for each gene were quantified using htseq-count (Anders et al., 2015). Differentially expressed genes were identified using DESeq2 (Love et al., 2014). The pausing index (PI) analysis was determined for the upregulated genes
e3 Molecular Cell 76, 1–15.e1–e4, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
from the GRO-seq experiment with at least 50 reads in the DMSO sample. The reads coverage was calculated using bed files containing the desired genomic regions and orientation (awk command) and the bedtools coverage tool (Quinlan and Hall, 2010). Chromatin immunoprecipitation (ChIP) HeLa cells were treated with different SPIs (50 mM; 60 min) or DMSO as a control, followed by TNF-a induction (30 min) in 100-mm plates. The cells were then cross-linked with 1% formaldehyde for 10 min at room temperature, and fixing was terminated by adding a 1/20 volume of 2.5 M glycine. Chromatin extraction and immunoprecipitations were carried out as previously described (Diamant et al., 2012). For immunoprecipitation, 3 mg of anti-RNA polymerase II, anti-Spt5, anti-p65 or control IgG was added to 1 mL of the soluble chromatin DNA. IP and input samples were analyzed by qPCR. The primer sequences are shown in Table S1. Cells extract, western blotting The whole-cell extract was prepared using commercial Reporter lysis buffer (E4030, Promega). Samples were then separated by SDS-PAGE and subjected to western blotting. Western blots were quantitated using Image Studio Software. The source of the antibodies used throughout the study is shown below. Co-immunoprecipitation (CO-IP) HeLa cells were detached from 100-mm plate with ice-cold PBS and a cell scraper, centrifuged for 5 min at 2,000 rpm and resuspended in 200 mL of lysis buffer (20 mM Tris-HCl pH 8, 400 mM NaCl, 2 mM EDTA, 0.5% NP-40, 10% glycerol) to which protease inhibitor cocktail (1:100) and PMSF (200 mM) were freshly added. The lysates were rotated for 10 min at 4C, and centrifuged for 15 min at 13,000 rpm. The supernatants were diluted with lysis buffer without NaCl to reach a final concentration of 100 mM NaCl. The cell lysate was then incubated overnight with anti-Spt5 antibody or rabbit IgG at 4C, followed by incubation with Protein A/G Sepharose beads (GE Healthcare) for 2h at 4C. The lysate was washed 3 times and then split and incubated with different SPIs or DMSO for 2h at 4C. The immunoprecipitated proteins were detected by SDS-PAGE followed by western blot using anti-Spt5, anti-Pol II and anti-Spt4 proteins antibodies. Binding assay To measure binding of SPIs to Spt5, Pol II and control protein (eIF4E), these proteins were expressed in E. coli as a fusion with histidine tag and purified on nickel agarose beads as described previously (Ashkenazi et al., 2016). The proteins were incubated with increasing concentrations of SPIs in black microplate 384 wells in a total volume of 20 mL for 10 min. The binding was assessed by measuring the changes in the intrinsic fluorescence (280 nm) using Cytation 5 (BioTek). IC50 was calculated and graphed using AAT Bioquest online calculator (https://www.aatbio.com/tools/ic50-calculator). Cells viability assay HeLa cells were plated in an opaque-walled 96 well plate (50,000 cells per well) and treated with increasing concentrations of SPI-21 or SPI-18 for 48h. After equilibrating the plate and its contents at room temperature for 30 min, 100 mL (1:1) of CellTiter-Glo Reagent (G7571, Promega) was added to each well and shook for 2 min to induce cell lysis. Following 10 min incubation at room temperature, the luminescent signal was detected using the Cytation 5 instrument. QUANTIFICATION AND STATISTICAL ANALYSIS To calculate p values, Student’s t tests (typically one-tail, paired) were performed. The significance of the difference in the median values was calculated by the Wilcoxon Signed Rank test. Significance symbols in all experiments are: * = p < 0.05; ** = p < 0.01; *** = p < 0.005; **** = p < 0.001. DATA AND CODE AVAILABILITY The RNA-seq datasets generated during this study have been deposited in NCBI’s Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE136026.
Molecular Cell 76, 1–15.e1–e4, November 21, 2019 e4
Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles Graphical Abstract
Authors Anat Bahat, Or Lahav, Alexander Plotnikov, Dena Leshkowitz, Rivka Dikstein
Correspondence [email protected]
In Brief Spt5 promotes promoter-proximal pausing, promoter escape, mRNA processing, and selective transcription of the mutant huntingtin gene. Bahat et al. identified Spt5-Pol II small-molecule inhibitors that mimic Spt5 knockdown effects and uncouple its different functions. The inhibitors uncovered Spt5 regulatory roles in metabolism and 30 end processing of histone genes.
Highlights d
Discovery of Spt5-Pol II small-molecule inhibitors (SPIs) that mimic Spt5 knockdown
d
SPIs relieve promoter-proximal pausing and inhibit inflammatory gene activation
d
d
SPIs inhibit mutant huntingtin gene transcription and 30 end processing of histone genes Selective targeting of Spt5-Pol II activities by SPIs uncouples its different functions
Bahat et al., 2019, Molecular Cell 76, 1–15 November 21, 2019 ª 2019 Elsevier Inc. https://doi.org/10.1016/j.molcel.2019.08.024
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Molecular Cell
Article Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles Anat Bahat,1 Or Lahav,1 Alexander Plotnikov,2 Dena Leshkowitz,3 and Rivka Dikstein1,4,* 1Department
of Biomolecular Sciences, The Weizmann Institute of Science, Rehovot 76100, Israel Nancy and Stephen Grand Israel National Center for Personalized Medicine, The Weizmann Institute of Science, Rehovot 76100, Israel 3Bioinformatics Unit, Department of Life Sciences Core Facilities, The Weizmann Institute of Science, Rehovot 76100, Israel 4Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.molcel.2019.08.024 2The
SUMMARY
Spt5 is a conserved and essential transcription elongation factor that promotes promoter-proximal pausing, promoter escape, elongation, and mRNA processing. Spt5 plays specific roles in the transcription of inflammation and stress-induced genes and tri-nucleotide expanded-repeat genes involved in inherited neurological pathologies. Here, we report the identification of Spt5-Pol II small-molecule inhibitors (SPIs). SPIs faithfully reproduced Spt5 knockdown effects on promoter-proximal pausing, NF-kB activation, and expanded-repeat huntingtin gene transcription. Using SPIs, we identified Spt5 target genes that responded with profoundly diverse kinetics. SPIs uncovered the regulatory role of Spt5 in metabolism via GDF15, a food intake- and body weight-inhibitory hormone. SPIs further unveiled a role for Spt5 in promoting the 30 end processing of histone genes. While several SPIs affect all Spt5 functions, a few inhibit a single one, implying uncoupling and selective targeting of Spt5 activities. SPIs expand the understanding of Spt5-Pol II functions and are potential drugs against metabolic and neurodegenerative diseases.
INTRODUCTION The regulation of transcription elongation by RNA polymerase II (Pol II) is central to many gene expression programs, and the deregulation of this stage is implicated in various pathologies. Elongating Pol II is associated with several transcription elongation regulatory factors, of which Spt5 is the most highly conserved, with homologs from bacteria to human (Werner, 2012). Spt5 interacts with Spt4 to form the DRB sensitivityinducing factor (DSIF) complex (Wada et al., 1998a; Yamaguchi et al., 1999b), which is mostly known for its function in promoter-proximal pausing in which Pol II stalls after transcribing 30–120 nt downstream of the transcription start site
(TSS). DSIF-mediated promoter-proximal pausing involves interaction with the negative elongation factor (NELF) (Yamaguchi et al., 1999a; Vos et al., 2018). The release of paused Pol II into productive elongation is mediated by the positive transcription elongation factor b (P-TEFb), a kinase that phosphorylates serine 2 of the Pol II C-terminal repeat domain (CTD) and facilitates this transition (Ni et al., 2008; Peng et al., 1998; Wada et al., 1998b). Genes known to be highly susceptible to inhibition via promoter-proximal pausing are those that are immediately activated in response to stress and extracellular signals such as heat shock and proinflammatory genes (Boehm et al., 2003; Diamant and Dikstein, 2013; Missra and Gilmour, 2010; Rougvie and Lis, 1990; Wu et al., 2003). During elongation, Spt5 has the ability to stabilize Pol II on the DNA template by interacting with upstream sequences and with RNA (Bernecky et al., 2016; Blythe et al., 2016; Crickard et al., 2016; Hirtreiter et al., 2010; Klein et al., 2011; Martinez-Rucobo et al., 2011). This activity was recently implicated in maintaining high transcriptional speed on long mRNAs (Fitz et al., 2018). Consistent with this, Spt5 was shown to accompany Pol II along the transcribed region until termination (Mayer et al., 2010; Pavri et al., 2010; Rahl et al., 2010). While in yeast and Drosophila, Spt5 depletion causes genome-wide transcription defects during the early elongation of sense and antisense transcripts (Henriques et al., 2018; Shetty et al., 2017), the depletion of Spt5 in mammalian cells does not result in broad transcriptional effects (Diamant et al., 2012, 2016b; Fitz et al., 2018; Komori et al., 2009; Pavri et al., 2010; Rahl et al., 2010; Stanlie et al., 2012), suggesting greater redundancy among mammalian transcription elongation factors. Our previous studies established Spt5 as a central player in the cellular response to nuclear factor kB (NF-kB) activation by controlling several steps during transcription. Induction of a subset of NF-kB target genes is remarkably fast and involves the presence of the transcription machinery on the promoters before activation and promoter-proximal pausing mediated by DSIF (Ainbinder et al., 2002, 2004). At the NF-kB target A20 gene, promoter-proximal pausing by DSIF-Pol II is mediated by an enhancer box (E-box) at the promoter (Ainbinder et al., 2004; Amir-Zilberstein et al., 2007; Amir-Zilberstein and Dikstein, 2008). In addition, in a subset of NF-kB target genes, DSIF coordinates elongation with mRNA splicing and export (Diamant Molecular Cell 76, 1–15, November 21, 2019 ª 2019 Elsevier Inc. 1
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A Spt5
Rpb1
Rpb1
Spt5
C-RL
N-RL
B Establishment of split-RL assay for Spt5-Pol II interaction Screening ~100,000 compounds (>30% inhibition)
587 Screening with full-length RL and removal of RL enzyme inhibitors
309 IC50 dose response assay at 5 concentrations: 0.55, 1.6, 5, 15, 45μM
148 Live cell assay
140 Filtering out overlapping hits
96
Molecules with IC50<40μM were selected for further analysis
41 Biologically active
18
C H
N
100
O
Cl
N
Cl
50
50
75
O
IC50 = 8.9μM
O
SPI-31
25
Ac t i v i t y ( % )
H3 C
N
60
SPI-29
N
IC50 = 36.5μM
70
Ac t i v i t y ( % )
N
75
N
80
N
SPI-17
Ac t i v i t y ( % )
90
N O
N
IC50 <5μM
50
CH 3 H N
25
S
O
O
40 O
100
1
H
80 O
O
50
Ac t i v i t y ( % )
Ac t i v i t y ( % )
H
IC50 = 25.2μM
75
SPI-57 CH 3
40
1
S
S S
SPI-95
O
10 Concent rat ion (μM)
O
50
N
IC50 = 25.4μM
75
50
CH 3
O
+
H
N H S
S
25
100
N
1 10 Concent rat ion (μM)
100
100
N
90
IC50 = 32.4μM
80 70 60
H3C
N
SPI-39
N
O N
80
IC50 = 16.9μM
70 60 50 40
50
S
H 3C
CH 3
90
H
N
60
N
SPI-85
100 H
IC50 = 33.1μM
70
Ac t i v i t y ( % )
Ac t i v i t y ( % )
50
20
100
80
100
100
IC50 = 36.1μM
60
100
90
10 Concent rat ion (μM)
N
70
30
O
25 1 10 Concent rat ion (μM)
1
Ac t i v i t y ( % )
100
SPI-18
H
100
90
O
SPI-21
10 Concent rat ion (μM)
Ac t i v i t y ( % )
1 10 Concent rat ion (μM)
30 1 10 Concent rat ion (μM)
100
1 10 Concent rat ion (μM)
100
1 10 Concent rat ion (μM)
H
85 80
SPI-157
90
75
IC50 = 5.8μM
50
S
H
80
100
1 10 Concent rat ion (μM)
–
100
1 10 Concent rat ion (μM)
40
SPI-42
O F
30
IC50 = 26.1μM
70 60 50
H 3C N
10 Concent rat ion (μM)
+
H
100
1 10 Concent rat ion (μM)
Cl N
IC50 = 21.6μM
80 70 60
S
F F
100
F
1 10 Concent rat ion (μM)
100
H 3C
110
70 60
H
SPI-46
N
H N
N
50
H
N
–
IC50 = 12μM
60 50
SPI-86
O
F
100
60
IC50 = 5.5μM
50 40
S N O
N
S
40
40 1 10 Concent rat ion (μM)
O
O
70
70
Ac t i v i t y ( % )
IC50 = 18.2μM
80
O
80
O
90
Ac t i v i t y ( % )
SPI-74
Ac t i v i t y ( % )
100
N
Cl
50
H 3C
30 1
SPI-16
O
40
S
20
N
90
80
Ac t i v i t y ( % )
IC50 = 26.5μM
50
100
100
Cl
90
60
40 20
100
O
Ac t i v i t y ( % )
Ac t i v i t y ( % )
SPI-68
O
50
O
1 10 Concent rat ion (μM)
60
30
70
70
IC50 = 14μM
70
S
O
O
25
75
SPI-09
O
Ac t i v i t y ( % )
IC50 = 40μM
90
Ac t i v i t y ( % )
Ac t i v i t y ( % )
95
O
100
O
100
SPI-06
100
H
100
105
1
10 Concent rat ion (μM)
100
N
30
O
1
10 Concent rat ion (μM)
+
O
–
100
(legend on next page)
2 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
et al., 2012). In addressing the activity of Spt5 in transcription elongation dynamics, we unexpectedly discovered that it also plays an important role in the early stages of transcription, supporting promoter escape and H3K4me3 and H4K5Ac chromatin modifications. These, in turn, facilitate TFIID maintenance on the promoter and re-initiation for the rapid induction of NF-kB response genes (Diamant et al., 2016a). Recently, Spt4/Spt5 has been shown to play a specific role in transcribing genes with lengthy stretches of multiple repeats that are the hallmarks of several inherited degenerative neurological disorders, including Huntington disease (HD), amyotrophic lateral sclerosis (ALS), and frontotemporal dementia (FTD) (Cheng et al., 2015; Kramer et al., 2016; Liu et al., 2012). Interference with Spt4 or Spt5 by small interfering RNA (siRNA) was reported to decrease the transcription of mutant huntingtin (Htt) gene alleles containing long CAG expansions, but had little effect on the expression of genes containing short CAG stretches (Cheng et al., 2015; Liu et al., 2012). The underlying basis for the requirement of Spt4/Spt5 in transcribing these genes is presently unknown. Because Spt5 is essential for cell viability, most studies aiming to interfere with its activity in Drosophila and mammalian cells use knockdown (KD) methods. However, these methods are lengthy, as treatment with siRNA requires at least 72 h, making distinguishing between direct and indirect effects challenging. In addition, it is labor-intensive and expensive. To expand the available tools for the study of Spt5, we report here the identification and characterization of small-molecule inhibitors (SPIs). SPIs precisely mimic the effect of Spt5 depletion on basal and activated NF-kB target genes and on the expanded repeat Htt gene. Using these inhibitors, we identified heat shock genes as transient target genes and uncovered novel regulatory roles of Spt5 that include the facilitation of the unique mode of histone gene termination and the regulation of an important metabolic pathway mediated by GDF15, a food intake- and body weightregulatory hormone. A subset of SPIs is selective to a single activity of Spt5, suggesting that the various activities of this factor can be uncoupled. RESULTS Identification of Spt5-Pol II Inhibitors Using HighThroughput Drug Screen The genetic manipulation of many proteins in mammals is limited by their importance for cell viability and the presence of multifunctional domains. Spt5 is an essential transcription elongation factor that has been mostly studied in metazoan cells using KD, as presently no inhibitor is available that can interfere with its activity. A major limitation of the KD approach against Spt5 is the lengthy treatment of at least 72 h. As the activities of Spt5 are critically dependent on its ability to bind Pol II (see, for example, Diamant et al., 2012, 2016a, 2016b), we envisaged that targeting this complex may enable the identification of Spt5 inhibitors,
which would be useful not only for Spt5 research but also for developing therapies against HD and inherited ALS and FTD. Spt5 is a large protein that contains several conserved functional domains (Werner, 2012), including NusG N-terminal homology domain (NGN), which mediates the interaction with Pol II; 5 KOW domains, which are also involved in the interaction with Pol II; a repetitive heptapeptide motif in the C-terminal region called CTR, implicated in the positive elongation activity of Spt5 (Chen et al., 2009; Ivanov et al., 2000; Yamada et al., 2006); and a C-terminal domain with an unknown function. We applied the split Renilla luciferase (RL) complementation assay to detect the Spt5 interaction with Pol II, as previously described (Ashkenazi et al., 2016, 2017). In this assay, RL is split into 2 inactive N- and C-terminal fragments and fused to target proteins. Interaction of the target proteins brings the N and C termini into close proximity, which can restore the RL enzymatic activity. In this system, the spatial arrangement of the 2 RL parts is critical for the enzymatic activity; therefore, it is sensitive not only to direct interference with the interacting proteins but also to changes in the conformation of the fused proteins (Ashkenazi et al., 2016; Haimov et al., 2018). For this purpose, the region spanning NGN and KOW 1 and 2 of Spt5 was fused to N-terminal Renilla luciferase (N-RL), and the coiled-coil (CC) domain of Rpb1 (the large subunit of Pol II) was fused to C-terminal Renilla luciferase (C-RL) (Figure 1A). These constructs were tested for specific interactions in mammalian cells and then co-expressed in Escherichia coli from a single plasmid, as described by Ashkenazi et al. (2017). A major advantage of the use of recombinant protein fragments of large proteins such as Spt5 and Rpb1 is the potential to identify specific compounds that directly bind to these protein domains. Bacterial cell lysates were used in a 1,536-well plate format to screen several small-molecule libraries with diverse chemical characteristics, consisting of 100,000 compounds (Figure 1B). A total of 587 inhibitors of the RL activity were identified and tested with the full-length RL enzyme to eliminate inhibitors of the RL enzymatic activity. The resulting 309 compounds were further validated in a doseresponse assay using Spt5 and Pol II split-RL pair and the fulllength RL. Approximately half of the compounds also inhibited the full-length RL at the high concentration, leaving 148 compounds that specifically inhibited the activity of the Spt5-Pol II split-RL pair. To select compounds that can enter mammalian cells, they were further tested in a live-cell split-RL assay, as previously described (Ashkenazi et al., 2017), and 140 retained their inhibitory activity. After filtering out 44 potentially non-specific compounds, as they also appeared in other screens of these libraries, we selected 41 Spt5-Pol II inhibitors (SPIs) displaying a half-maximal inhibitory concentration (IC50) % 40 mM for further analysis. Of the 41 compounds analyzed, 18 were found to exert at least one biological effect (see below), whereas the rest exerted none. The chemical structures and the IC50 measurements of the biologically active compounds are shown in Figure 1C. Notably, several of these compounds appear to be chemically
Figure 1. A High-Throughput Drug Screen (HTS) against Spt5-Pol II (A) An illustration of Spt5 (NGN + KOW1/2 domains) and Rpb1 (coiled-coil domain) split-RL assay used for the HTS. (B) A flowchart describing the steps carried out during the HTS in search of direct inhibitors of Spt5-Pol II. (C) The chemical structure and the IC50 of the SPIs found to be biologically active.
Molecular Cell 76, 1–15, November 21, 2019 3
IEX1
Relative mRNA level
IL6
JunB
B Relative A20 mRNA level
Relative mRNA level
CXCL2
CCL20
Relative mRNA level
Relative mRNA level
IL8
Relative mRNA level
A20
Relative mRNA level
A
Relative mRNA level
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
E
D
C
Average of A20, IL6, IL8, CXCL2 Relative mRNA level
W/O Promoterproximal pausing
F
With Promoterproximal pausing 73%
G
HSPA6
EGR1
FOS
A20
IL8
(legend on next page)
4 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
related—for example, one class includes SPI-21, SPI-29, and SPI-57 and another SPI-68, SPI-46, and SPI-157. SPIs Relieve Promoter-Proximal Pausing In vitro transcription assays, along with Spt5 KD studies, established the NF-kB-responsive A20 gene as a target of Spt5 for promoter-proximal pausing under basal conditions (Ainbinder et al., 2004; Amir-Zilberstein and Dikstein, 2008). Following the activation of NF-kB by tumor necrosis factor a (TNF-a), Spt5 turns into a positive regulatory factor and facilitates promoter escape and mRNA splicing (Diamant et al., 2012, 2016a). Genome-wide analysis of transcripts from Spt5 KD cells identified other Spt5-regulated genes under resting and TNFa-induced conditions (Diamant et al., 2016b). To examine the ability of SPIs to interfere with Spt5 activity, we selected a subset of TNF-a-responsive genes that are also targeted by Spt5 (A20, C-C motif chemokine ligand 20 [CCL20], interleukin [IL]-6, IL-8, C-X-C motif chemokine ligand 2 [CXCL2], JUNB, and IEX1) from the aforementioned studies and from others (Aida et al., 2006; Fujita et al., 2009) and tested their effect under basal conditions. Cells were treated with vehicle (DMSO) or SPI (%50 mM) for 2 h, and then RNA was extracted and analyzed by qRT-PCR. Most of the examined SPIs enhanced the basal activity of the selected NF-kB-responsive genes (Figure 2A), akin to the effect of Spt5 KD. To determine whether the observed effect is linked to activated transcription or an increase in mRNA stability, we treated cells with actinomycin D in the presence or absence of one randomly selected compound, SPI-21, and then measured the mRNA levels at different time points. The stability of A20 mRNA was not enhanced but actually was moderately reduced in SPI-21 samples (Figure 2B). To obtain a genome-wide view of the impact of SPIs on transcription, we performed Global run-on sequencing (GRO-seq) using the nuclei of HeLa cells treated with either DMSO (control) or SPI-21. The GRO-seq experiment measures the distribution of transcriptionally active Pol II along the length of the gene (Core et al., 2008). This analysis revealed that short-term SPI-21 treatment caused a transcription upregulation (R2-fold) of 1,371 genes, among them A20 (TNFAIP3), IL-6, IL-8 (CXCL8), and CCL20 (analyzed above), while CXCL2, IEX1 (IER3), and JUNB were unchanged. The reduction of promoter-proximal pausing
is expected to release the paused Pol II into the gene body, and thus upregulate transcription. We therefore calculated the ratio between Pol II density around the TSS and the gene body (pausing index [PI]) of the SPI-21 upregulated gene set and found that the majority display promoter-proximal pausing (PI >1.2) (Figure 2D). Furthermore, treatment with SPI-21 resulted in a general reduction in their PI (Figure 2E). The sequencing tracks of a few examples of upregulated genes with reduced PI values following SPI-21 treatment are shown in Figure 2F. These findings together confirm that elevated mRNA levels following short-term SPI treatment are, at least in part, a consequence of promoter-proximal pausing relief, which is consistent with a known function of Spt5. We examined the SPI-21 upregulated genes from the GROseq data for the enrichment of regulatory elements in their control regions (2,000 to +2,000 around the TSS) using a gene set enrichment analysis (GSEA) database and software (Subramanian et al., 2005). One of the highest-scoring motifs identified has a CAGGTG/CACGTG sequence (Figure 2G) that is remarkably similar to the E-box motif of the A20 promoter CACGTG, which was found to control promoter-proximal pausing through Spt5 (Ainbinder et al., 2004; Amir-Zilberstein and Dikstein, 2008). Another enriched motif is the binding site of heat shock factor 1 (HSF1), a transcription activator of heat shock genes, which we found to be regulated by Spt5 via promoter-proximal pausing in mammalian cells (see Figures 2F and 4D). SPIs Mimic the Effects of Spt5 KD on Pro-inflammatory NF-kB Target Genes and on Cell Survival Next, we analyzed the effect of SPIs following NF-kB activation, a state in which Spt5 turns into a positive regulatory factor. Cells were pre-incubated with DMSO or SPIs (%50 mM) for 1 h and then treated with TNF-a for an additional 1 h. The selected genes were either moderately (CXCL2, JUNB, IEX1) or strongly (A20, CCL20, IL-6, IL-8) activated by TNF-a (Figure 3A, DMSO+). The induced levels of all of these genes were diminished to a variable degree by the SPIs, similar to the effect seen with Spt5 KD (Diamant et al., 2016a). The potent inhibitors decreased the induction rate of all of the analyzed genes, while the weaker inhibitors such as SPI-68, -74, -95, -42 and -06 affected mainly
Figure 2. Effects of SPIs on Promoter-Proximal Pausing and Pol II Recruitment (A) HeLa cells were treated with DMSO or 50 mM of the indicated SPIs (except SPI-57 and SPI-42, which were at 25 mM, and SPI-31, which was at 7.5 mM) for 2 h. Then, total RNA was extracted, and the levels of the indicated mRNAs were determined by quantitative real-time PCR and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Bars represent the means ± SEMs of at least 3 independent experiments. (B) The effect of SPI-21 on the basal levels of A20 mRNA stability. Cells were treated with SPI-21 and actinomycin D for the indicated time points, and the levels of A20 and GAPDH mRNA were determined by qRT-PCR. The data present the means ± SEMs of 3 independent experiments. (C) A graph summarizing the average effect of SPIs on the indicated genes organized in descending order of potency. (D) HeLa cells, pretreated with either DMSO or SPI-21 for 90 min, were subjected to GRO-seq analysis as described in the STAR Methods. A pie chart demonstrating the percentage of SPI-21 upregulated genes that display promoter-proximal pausing (PI R1.2). The PI was calculated by dividing Pol II density (reads per 1,000 bp) around the TSS (300 to +700 bp) by Pol II density in the gene body (+700 to +1,700). (E) A boxplot presenting the change in the pausing index (PI) of the upregulated genes that display promoter-proximal pausing (PI R1.2 in DMSO) upon SPI-21 treatment. The significance of the difference in the median values between control and treatment was calculated by the Wilcoxon signed-rank test. (F) Several examples showing GRO-seq reads aligned to the genome. DMSO and SPI-21 samples are presented in blue and red, respectively. The PIs, calculated for each gene, are shown at right. (G) Enrichment of regulatory sequences in SPI-21 upregulated genes (R2-fold change). Shown are the top 11 output sequences, along with their enrichment score and statistical significance. The analysis was done using a GSEA database and software (Subramanian et al., 2005). The asterisks denote statistical significance differences according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
Molecular Cell 76, 1–15, November 21, 2019 5
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A
A20 Induction fold
Induction fold
CCL20
IL6 Induction fold
Induction fold
IL8
JunB
B
Induction fold
IEX1
C
Relative A20 mRNA level
Induction fold
Induction fold
CXCL2
D
Induction fold
Average of A20, IL6, IL8, CXCL2
Relative occupancy
Relative occupancy
E
A
B
C
A
B
C
(legend on next page)
6 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
the strongly induced genes (compare A20, CCL20, IL-6, and IL-8 to CXCL2, JUNB, and IEX1 in Figure 3A). The effect of the weak inhibitors is remarkably similar to that seen with Spt5 depletion, which primarily diminished the induction of strongly induced genes (Diamant et al., 2016b). Using actinomycin D on SPI-21 and TNF-a-treated cells, we ruled out the possibility that the reduced mRNA levels are a consequence of diminished mRNA stability (Figure 3B). In fact, the half-life of the induced mRNA levels of A20 in control samples is shorter than that in the SPI21-treated cells. We confirmed that the observed effects are not a consequence of changes in Rpb1 and Spt5 protein levels caused by SPI-21 (Figure S1A). In addition, we found no significant alterations in the occupancy of the NF-kB protein p65 of its target genes upon SPI-21 treatment under basal and TNF-a conditions (Figure S1B). We determined the IC50 of a few SPIs on the TNF-a-induced transcription and found them to be mostly in the low micromolar range (Figure S1C). By summarizing the effect of the different SPI compounds on the selected genes, we obtained a ranking of their potency in inhibiting the negative and positive regulatory activities of Spt5 (Figures 2C and 3C). This analysis revealed that several compounds, including SPI-68, -31, -18, and -39 inhibited mainly the positive activity of Spt5, with little or no effect on promoter-proximal pausing, while SPI-06 predominantly affected promoter-proximal pausing. These results suggest that the positive and negative activities of Spt5 can be uncoupled. As Spt5 KD also reduced the splicing efficiency of the TNFa-induced A20 mRNA (Diamant et al., 2012, 2016b), we examined the effect of SPI-21 on splicing by determining the ratio between mature and immature mRNAs of several genes following TNF-a induction. The results revealed a clear decrease in the mature:immature ratio following treatment with SPI-21 (Figure 3D), which is consistent with the effect seen upon Spt5 KD. IkBa, the major inhibitor of NF-kB, is also a primary target gene of NF-kB itself, forming a negative feedback loop. Spt5 was shown to be central in maintaining this regulatory circuit (Diamant et al., 2012, 2016b). We therefore examined the effect of SPI-21 and SPI-18 on IkBa protein. In control cells, the initial levels of IkBa are diminished after 30 min of TNF-a induction as a consequence of its degradation and then recover after 150 min due to its induction by NF-kB (Figure S1D, DMSO).
Upon SPI-21 and SPI-18 treatment, the recovery of IkBa at 150 min is markedly reduced (Figure S1D, lanes 6 and 9 and corresponding columns), as observed in Spt5 KD cells (Diamant et al., 2012, 2016a). To examine further the impact of SPIs on the chromatin occupancy of Spt5 and Pol II, we performed chromatin immunoprecipitation (ChIP) in the presence of DMSO, SPI-21, and SPI-18, with or without 30 min of TNF-a treatment. Analysis of 3 loci of the A20 gene under basal conditions revealed a moderate but significant enhancement of Spt5 and Pol II occupancy at the 3 loci by SPI-21, while their levels were unchanged with SPI-18 (Figure 3E, basal). This finding is consistent with the increase in basal transcription exerted by SPI-21 but not by SPI-18 (Figure 2A). Upon TNF-a induction, the levels of Pol II and Spt5 are dramatically elevated, and this induction is diminished by both SPI-21 and SPI-18 at all 3 loci (Figure 3E, TNF-a), which correlates well with their effect on transcription (Figure 3A). The effects of SPIs on NF-kB target genes under basal and induced conditions are in full agreement with the previously reported negative and positive effects of Spt5 on pro-inflammatory genes, as revealed by the KD approach. We selected SPI-21 and SPI-18 to validate their direct interaction with either the Spt5 or the Rpb1 CC domain. His-Rpb1 and His-Spt5 (fused to Spt4 to increase its stability) and His-eIF4E, used as a negative control, were each expressed in E. coli and purified by nickel agarose beads. Using the purified proteins, we performed fluorescence intensity measurements in the presence of increasing concentrations of SPI-21 and SPI-18. This method analyzes binding by monitoring the changes in the intrinsic fluorescence of the proteins as a consequence of conformational changes. Spt5, Rpb1, and eIF4E display significant intrinsic fluorescence, as expected from the presence of aromatic residues in all of them. SPI-21 decreased the intrinsic fluorescence of Rpb1 in a dose-dependent manner with an IC50 of 23 mM, while it had no significant effect on Spt5 and eIF4E (Figure S2A). This IC50 is comparable to the 25 mM IC50 value seen in the split-RL assay with SP-21 (Figure 1C) and suggests that SPI-21 primarily binds the Rpb1 CC domain. SPI-18, however, seems to affect the fluorescence of both Spt5 and Rpb1, and the measured IC50 is 40 mM for Spt5 and 76 mM for Rpb1. The IC50 of Spt5 binding by SPI-18 is close to
Figure 3. Effect of SPIs on TNF-a-Induced Inflammatory Response Genes (A) Cells were incubated with DMSO or 50 mM indicated SPIs (or less, as indicated in Figure 2) for 1 h and then treated with TNF-a (20 ng/mL) for an additional 1 h. Then, RNA was extracted and analyzed for the indicated mRNAs by qRT-PCR and normalized to GAPDH. The results are presented as fold of induction relative to non-treated DMSO, which was set to 1. Bars represent the means ± SEMs of at least 3 independent experiments. The asterisks (as described for Figure 2A) denote statistically significant differences relative to DMSO + TNF-a. (B) The effect of SPI-21 on the stability of the TNF-a-induced A20 mRNA. Cells, pre-treated with SPI-21 and TNF-a, were incubated with actinomycin D for the indicated times, and A20 and GAPDH mRNA levels were determined by qRT-PCR. The data points represent the means ± SEMs of 3 independent experiments. (C) A graph summarizing the average effects of SPIs on the indicated genes organized in descending order of potency. (D) A dot plot presenting the effect of SPI-21 on the splicing efficiency of TNF-a-induced genes. Cells were incubated with DMSO or 50 mM SPI-21, as in (A). Following RNA extraction, cDNA was prepared using random hexamers, and levels of mature and immature transcripts were determined by qRT-PCR using different sets of primers (see scheme on top: red and green arrows for mature and immature transcripts, respectively). Each dot presents the ratio between mature and immature transcripts in each sample. The data represent the means ± SEMs of 3 independent experiments, and the asterisks denote statistically significant differences relative to DMSO + TNF-a. (E) SPIs effect on Pol II and Spt5 occupancies of the A20 gene. Cells, pre-treated with DMSO or SPI-21 and SPI-18 (50 mM) for 1 h, were induced with TNF-a for 30 min and then subjected to ChIP using anti-Rpb1 and anti-Spt5 antibodies. The occupancy of the A20 gene was analyzed by qRT-PCR using primers from the first (A), third (B), and last (C) exon-intron junctions. Bars represent the levels of Pol II and Spt5 relative to the input, and are the means ± SEMs of 3 independent experiments. The asterisks denote statistically significant difference compared to the DMSO counterpart with or without TNF-a; according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
Molecular Cell 76, 1–15, November 21, 2019 7
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A
B
TNF - induced genes
induced genes
11
31
9
SPI-21 up Genes (basal)
C
SPI-21 down genes TNF vs ctrl
D
SPI-21 vs ctrl
SPI-21 +TNF vs TNF
SPI-21 +TNF vs ctrl
E
HSPA1A
SPI-21 up-regulated genes under basal conditions
Relative mRNA levels
HSPA6
528
315
G Relative GDF15 level
Relative GDF15 level
F
93
Relative GDF15 level
24h
I Relative GDF15 mRNA level
24h
H
(legend on next page)
8 Molecular Cell 76, 1–15, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
the values obtained with the split-RL (33 mM; Figure 1C) and with the inhibition of inflammatory genes (39 mM; Figure S1C). These findings are consistent with the direct interaction of the identified SPI compounds with the Spt5 and Rpb1 proteins. Next, we performed co-immunoprecipitation in the presence of SPI-21 and SPI-18 to examine their effect on the endogenous Spt4/Spt5-Pol II complex and found that these compounds did not disrupt their binding (Figure S2B). These findings are not surprising, as the binding of Spt5 to the multi-subunit Pol II involves multiple Spt5 domains, several subunits, and many contact points; thus, interference with one domain may not be sufficient to detach it from Pol II. In addition, SPI-21 and SPI-18 binding to Rpb1 and Spt5 are more likely to induce conformational changes that diminish Spt5-Pol II function rather than their binding. Spt5 was reported to be essential for cell growth and survival in mammalian cells (Komori et al., 2009). We therefore examined the effect of increasing concentrations of SPI-21 and SPI-18 on cell growth and found that both compounds dramatically diminish cellular viability in a dose-dependent manner (Figure S3). SPIs Unveil Kinetically Distinct Spt5 Target Genes and a Metabolic Function A major limitation of KD and knockout (KO) studies is the prolonged deficiency of the target protein. Small molecule inhibitors, on the other hand, inactivate their target soon after their application. To examine the global effect of a short-term Spt5 inhibition, cells were treated with SPI-21 for 2 h or with SPI-21 for 1 h and TNF-a for an additional 1 h and then subjected to RNA sequencing (RNA-seq). Reads were aligned to the human genome, and the normalized expression of each gene was determined. Subsequently, the following ratios were calculated: TNFa versus control, which describes the response to TNF-a; SPI-21 versus control, which describes the effect of SPI-21 on basal gene expression; and SPI-21 + TNF-a versus TNF-a, which describes the effect of SPI-21 on TNF-a induction rate and SPI21 + TNF-a versus control. A total of 506 genes displayed a statistically significant fold change in at least 1 of the ratios, and these were grouped into 5 distinct clusters (Figure 4A). The
vast majority of SPI-21-affected genes were upregulated as also seen with the GRO-seq. This is similar to the global effect seen upon Spt5 KD in human (Diamant et al., 2016b) and mouse (Fitz et al., 2018) cells. Clusters 1–4 contain genes that are upregulated by SPI-21 under basal conditions and cluster 5 contains SPI-21 downregulated genes (see SPI-21 versus control). Cluster 1 also consists of TNF-a-induced genes that are inhibited by SPI-21, while the TNF-a-induced genes in cluster 2 are not affected by it. We selected several genes from the different clusters to examine the effect of a few other SPIs on their expression by qRT-PCR (excluding cluster 1, which is represented in Figures 2 and 3). The changes in their levels are highly similar to those of SPI-21 (Figure S4). Analysis of the RNA-seq data provided several insights. There is a substantial overlap between genes prone to induction by TNF-a and those that are upregulated by SPI-21 under basal conditions, suggesting that promoter-proximal pausing is central to this signaling pathway (see clusters 1 and 2). Comparing the effect of SPI-21 with that of Spt5 KD (Diamant et al., 2016b) on the TNF-a-induced genes, we found that most of the downregulated genes are common to the two treatments (Figure 4B). In addition, biological pathway analysis of the SPI-21 upregulated genes revealed remarkable enrichment of inducible genes in response to a variety of extracellular signals other than TNF-a, including unfolded protein response, heat, hypoxia, cyclic AMP (cAMP), and mechanical stimulus (Figure 4C). Of particular interest are heat shock genes, which were very strongly upregulated. While heat shock genes are known Spt5 targets for promoter-proximal pausing in Drosophila (Missra and Gilmour, 2010; Wu et al., 2003), these genes were not identified as such by the Spt5 KD studies in mammalian cells (Diamant et al., 2016b). Activation of these genes is also apparent in the GRO-seq data (see Figures 2F and 2G). We raised the possibility that the effect of Spt5 inhibition on heat shock genes may be transient and therefore is not detected by the prolonged KD approach. To test this possibility, cells were treated with SPI-21 for 2 and 24 h and levels of heat shock genes HSPA6 and HSPA1A were determined by qRT-PCR. Both genes were strongly upregulated after 2 h,
Figure 4. Identification of Kinetically Distinct Spt5 Target Genes (A) The global effect of SPI-21 and TNF-a on mRNA levels. Cells were pre-treated with SPI-21 (50 mM) or DMSO for 1 h and then with TNF-a for an additional 1 h. Sequencing libraries from mRNA extracted from these cells were prepared and subjected to deep sequencing. Reads were aligned to the human genome, and the ratios indicated at the bottom were calculated between the samples. The resulting gene list (average of 2 experiments) was clustered into 5 distinct groups and is shown as a heatmap of the log fold change of differentially expressed genes. (B) A Venn diagram showing the overlap between SPI-21 downregulated genes from the TNF-a-induced set (presented in cluster 1) and the TNF-a-induced genes that were downregulated upon Spt5 KD retrieved from the previously published RNA-seq data (clusters 2 and 3 in Diamant et al., 2016b). (C) Gene set enrichment analysis (using the GeneAnalytics Gene Ontology [GO] function) of the biological processes associated with SPI-21 upregulated genes. (D) Analysis of the effects of short- and long-term treatments (2 and 24 h, respectively) of SPI-21 on the expression of heat shock genes HSPA6 and HSP1A1, using qRT-PCR (normalized to GAPDH). (E) Comparison of the global effects of long- and short-term SPI-21 treatment. Cells were treated with SPI-21 (50 mM) or DMSO for 24 h and subjected to RNAseq. The extent of overlap between the upregulated gene sets (R2-fold change) from 2 h (described in A) and 24 h was determined using a Venn diagram. (F) The change in GDF15 levels following 72-h Spt5 KD and 24-h SPI-21 treatment. The Spt5 KD data were retrieved from the previously published RNA-seq analysis (Diamant et al., 2016b). (G) Effect of short- and long-term treatments of SPI-21 on the expression of GDF15 mRNA analyzed by qRT-PCR and normalized to GAPDH. (H) The effect of 24-h treatment of the indicated SPIs on GDF15 levels analyzed by qRT-PCR and normalized to GAPDH (final concentrations: SPI-18, 50 mM; SPI06, -17, -29, and -95, 30 mM; SPI-42, 10 mM; SPI-86, 5 mM). (I) Analysis of the effect of 24-h SPI-21 and SPI-18 treatments on GDF15 mRNA levels in mouse embryonic fibroblasts (MEFs), MCF7, and Jurkat cells. The graphs in (D) and (G)–(I) represent the means ± SEMs of at least 3 independent experiments. The asterisks denote statistically significant differences relative to DMSO; according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
Molecular Cell 76, 1–15, November 21, 2019 9
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
but after 24 h their activation had considerably decreased (Figure 4D), indicating that the effect of Spt5 depletion on their expression is temporary, and thus explaining the absence of these genes from the Spt5 KD studies. To further test the idea of kinetically distinct Spt5 target genes, we determined the global effect of long-term SPI-21 treatment (24 h) by RNA-seq and compared the affected genes between the shortand long-term treatments. We found that 80% of the affected genes differed between the short- and long-term treatments (Figure 4E). Examination of the RNA-seq data of Spt5 KD cells (Diamant et al., 2016b) and the long-term SPI-21 treatment revealed GDF15 as one of the top upregulated genes, elevated by 27fold in the 2 experiments (Figure 4F). GDF15 is a member of the transforming growth factor b (TGF-b) superfamily and is associated with the regulation of body weight and food intake in humans and rodents (Emmerson et al., 2017; Hsu et al., 2017; Mullican et al., 2017; Tsai et al., 2018; Yang et al., 2017). GDF15 levels were almost unchanged by the 2-h SPI-21 treatment. To gain further support that the strong upregulation of GDF15 is associated with the prolonged inhibition of Spt5, we analyzed the GDF15 expression following 2 and 24 h of treatment with SPI-21 by using qRT-PCR. Contrary to the heat shock genes, GDF15 was 100-fold upregulated after 24 h, while only moderately upregulated after 2 h of SPI-21 treatment (Figure 4G). These findings support the idea that the strong upregulation of GDF15 requires persistent Spt5 inhibition and may be partially indirect. The dramatic elevation of GDF15 is also seen with other SPIs (Figure 4H), and it is not limited to HeLa cells but other human and mouse cell lines (Figure 4I). Thus, temporary and constitutive Spt5-regulated genes can be distinguished by using SPIs, which also unveiled a previously unknown but significant role for Spt5 in the control of an important mammalian metabolic pathway via GDF15. To validate that the effects of SPIs are at the transcriptional level, cellular RNA was metabolically labeled with 4-thiouridine (4sU) for 2 h in the presence of DMSO or SPI-21 in the presence or absence of TNF-a. Newly synthesized labeled RNA was then purified and subjected to RNA-seq. We determined the overlap between the SPI-21 and Spt5 KD upregulated genes of the TNF-a-responsive set, and found that 67% of Spt5 KD upregulated genes are common to the newly synthesized SPI-21 upregulated genes from this set (Figure S5A). Analysis of the effect of TNF-a induction in these data showed a remarkably high number of TNF-a-induced genes (467) compared to the conventional RNA-seq (<100; see Figure 4A and Diamant et al., 2016b). A substantial fraction of these genes (285 genes, 61%) is downregulated by SPI-21 (Figure S5B), while only 301 of 13,890 expressed genes (2.2%) are downregulated under basal conditions. Thus, the observed reduction in TNF-a activation by SPI-21 is highly specific and is a consequence of diminished transcription. Selective Inhibition of Mutant but Not Wild-Type Htt Gene Transcription by SPIs HD is an inherited genetic disorder caused by the expansion of trinucleotide CAG repeats (R36) encoding glutamine (Q) in the Htt gene. Expression of the mutant gene in neurons of the stria-
10 Molecular Cell 76, 1–15, November 21, 2019
tum and the frontal cortex leads to the degeneration of neuronal cells that control body movements, emotions, and intellect. Previous studies identified Spt4 and Spt5 as important players in promoting the transcription of mutant Htt gene containing long CAG repeats, but with little effect on short CAG stretches (Cheng et al., 2015; Liu et al., 2012). To examine the potential effect of SPIs in this context, we used a striatal cell line established from wild-type (WT) (Q7) or mutant (Q111) Htt knockin mice. In these cell lines, the coding region of the first exon of the mouse Htt gene was replaced with the same region of human WT (Q7) or mutant (Q111) Htt, as shown schematically in Figure 5A. We treated the 2 cell lines with SPIs for 48 h and then determined the levels of Htt mRNA. The results revealed that several SPIs selectively diminished the Q111 but not the Q7 Htt mRNA (Figure 5B). Notably, the effective concentration of SPI-86 and SPI-31 is 15-fold lower than the concentration needed for these drugs to diminish the proinflammatory gene expression (Figures 3 and S1C). To examine whether the selective effect on mutant Htt is at the transcriptional level, we performed metabolic labeling of cellular RNA with 4sU for 2 h in the presence of DMSO or SPI-21. Analysis of the RNA levels using primers from different regions of the Htt gene (shown in Figure 5A) revealed that levels of the newly synthesized Q111 but not Q7 Htt were diminished by SPI-21, and the extent of inhibition is greater compared to its effect on steady-state levels (Figure 5C), indicating that this inhibitor acts at the level of mRNA transcription. In addition, these data show that in controls (DMSO), the relative levels of the first intron are much lower in Q7 as compared to Q111 (first exon:intron ratio differs significantly in Q7 versus Q111, p = 0.05), suggesting that the splicing efficiency of the longer repeats is reduced. Moreover, the levels measured from the second exon relative to the first is somewhat higher in Q7 compared to Q111, and this small difference is observed all along the gene, a finding that is consistent with a modest reduction in transcription elongation. The reduction in the transcription elongation rate and in splicing are associated with reduced expression levels of the Q111 Htt mRNA relative to the Q7 (Figure 5D) and may be the underlying basis for the sensitivity of Q111 to Spt5 inactivation. We were intrigued by the observation that only a subset of SPIs affects mutant Htt expression. We therefore summarized the influence of the different SPIs on the various activities of Spt5 (Figure 5E). From this combined data, it is apparent that SPIs can be classified according to their effect on the various Spt5 activities. The first class, which includes SPI-21, -86, -29, -74, and -68, affect all of the examined Spt5 activities and can be considered general inhibitors. The second class affects a single activity: SPI-85 and SPI-09 diminished mutant Htt expression but have little or no effect on basal and induced inflammatory gene expression; SPI-39 inhibits only the induced inflammatory gene transcription, while SPI-06 predominantly affects promoter-proximal pausing. The largest class, which consists of SPI-17, -157, -46, -57, -16, -42, and -95, does not affect mutant Htt expression, but it does affect the basal and the induced inflammatory gene transcription (Figure 5E). These findings strongly suggest that there may be discrete mechanisms underlying the requirement of Spt5 for expanded repeats
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Figure 5. SPIs Selectively Inhibit the Transcription of the Mutant Htt Gene
A
Relative mRNA levels
B
C
D Relative Htt levels
Relative mRNA levels
4sU labeling
Q7
Q111
(A) A scheme showing the structure of WT (Q7) and mutant (Q111) Htt gene of the transgenic striatal cell lines. (B) Q7 and Q111 expressing cells were treated with SPIs at the indicated concentrations for 48 h. Then, RNA was extracted, and levels of Htt mRNA were determined by qRT-PCR and normalized to b-actin. (C) Q7 and Q111 cells were metabolically labeled with 5-thiouridine for 2 h in the presence or absence of SPI-21. Newly synthesized Htt transcript levels were determined by using primers from the indicated locations. (D) The Htt levels in Q7 and Q111 cells normalized to b-actin. Levels of Q7 Htt were set to 1. The data shown in (B)–(D) represent the means ± SEMs of at least 3 independent experiments. The asterisks denote statistically significant differences relative to DMSO (B and C) or to Q7 (D). (E) A table summarizing the effects of the indicated SPIs on the various Spt5 biological activities. The compounds are clustered according to their effects on the 3 different Spt5 activities indicated at top and are color-coded and numbered. For example, the compounds in cluster 1 inhibit all activities, while those in cluster 4 inhibit primarily mutant Htt expression. PPP, promoter-proximal pausing. Plus sign (+) indicates statistically significant effect. The asterisks denote statistical significance differences according to Student’s t tests (typically onetailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.001.
E SPI
Inflammatory genes Basal (PPP)
Induced
Mutant Htt
transcription, for induction of inflammatory genes, and for promoter-proximal pausing. Spt5 Regulation of Histone Genes Is Linked to 30 End Processing Spt5 KD resulted in a relatively small number of downregulated genes (43 genes), of which 20 are replication-dependent histone
genes (Figure 6A), suggesting that Spt5 plays an important role in their expression. Histone genes are coordinately synthesized with DNA during S phase and do not end with a polyA tail but with a conserved stem-loop structure. We analyzed the 24h RNA-seq data for the effect of SPI-21 on histone gene expression and found them to be similarly downregulated (Figure 6A). Likewise, 24 h of SPI-18 treatment resulted in a decreased expression of selected histone genes (Figure S6). The requirement of Spt5 for the expression of histone genes cannot be explained by a positive effect on elongation or mRNA processing since histone genes are very short and are intronless. Moreover, analysis of the RNA-seq data of the 4sU metabolic labeling experiment actually revealed a moderate increase in histone gene transcription upon SPI-21 treatment (Figure 6A), suggesting that their synthesis during S phase is not affected. To resolve the discrepancy between the effect of Spt5 inhibition on steady-state levels and synthesis rate, we considered the possibility that Spt5 is required for the processing of histone mRNAs. Besides the conserved stemloop, histone genes have a canonical polyadenylation signal
Molecular Cell 76, 1–15, November 21, 2019 11
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
A
B PolyA
HIST1H1B
Hist1H1B
Hist1H2AI
Hist1H3A
Hist1H4H
Total/Poly-A
HIST1H1E HIST1H2AG HIST1H2AH HIST1H2AI HIST1H2AK HIST1H2AL HIST1H2BF HIST1H3A HIST1H3B HIST1H3D Total/Poly-A
HIST1H3F HIST1H3G HIST1H4A HIST1H4B HIST1H4C HIST1H4D HIST1H4H HIST2H2AC HIST2H3D
C Hist1H2AI
Hist1H3A
Hist1H4H
Total/Poly-A
Total/Poly-A
Hist1H1B
Spt5 Tub
Figure 6. Regulation of 30 End Processing of Histone Genes by Spt5 (A) The levels of 20 replication-dependent histone genes following 72-h Spt5 KD, 24-h SPI-21 treatment, and 4sU metabolic labeling. All of the values were retrieved from the RNA-seq data carried out in this study and by Diamant et al. (2016b) and are shown as the mean of 2 experiments. In all of the experiments, cDNA was generated with random hexamers. (B) The ratio between total and poly(A) containing histone genes. A schematic illustration of the histone gene structure is shown at top. For the total and poly(A) mRNA analyses, cDNA was prepared with random hexamers or poly-deoxythymine (dT) oligonucleotide, respectively. The levels were normalized to GAPDH. (C) An analysis that is similar to that in (B) using RNA from control and Spt5 KD cells. The bottom panel shows a western blot of control and Spt5 KD cells using the indicated antibodies. The data in (B) and (C) represent the means ± SEMs of 3 independent experiments. The asterisks denote statistically significant differences relative to DMSO; according to Student’s t tests (typically one-tailed, paired): *p < 0.05; **p < 0.01; ***p < 0.005.
downstream of the stem-loop structure (Figure 6B). We therefore examined the possibility that Spt5 controls the choice between 30 end processing via the stem-loop pathway or the polyadenylation pathway of histone pre-mRNAs. By using cDNA prepared by random hexamer and Poly-T we determined the ratio between total and poly-A tailed histone genes. We found that relative to control cells, this ratio is significantly reduced, which is
12 Molecular Cell 76, 1–15, November 21, 2019
consistent with an increase in the relative amount of poly-A tailed histone genes upon Spt5 inhibition by the drugs (Figure 6B). A similar analysis with RNA extracted from Spt5 KD cells shows the same effect (Figure 6C). Thus, the reduction in the histone gene expression upon Spt5 inhibition either by the SPIs or KD may be associated, at least in part, with a defect in 30 end processing.
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
DISCUSSION In the present study, we report the development of a pharmacological approach to elucidate the function of Spt5, a central transcription elongation factor in mammalian cells. By targeting the functional Spt5-Rpb1 complex in a high-throughput drug screen, we identified several SPIs that faithfully reproduced the effect of Spt5 KD on the expression of basal and activated pro-inflammatory genes. Our findings with SPI-21 revealed regulatory sequences that potentially control the negative activity of Spt5. A few of these sequences are similar to sequences previously implicated in pausing in mammalian cells (Ainbinder et al., 2004; Amir-Zilberstein and Dikstein, 2008; Gromak et al., 2006; Yonaha and Proudfoot, 1999), but are distinct from those identified in Drosophila, one of which binds M1BP, a Drosophila exclusive factor (Lee et al., 2008; Li and Gilmour, 2013). Thus, the control of promoter-proximal pausing by Spt5 involves diverse mechanisms. Consistent with this notion, these inhibitors identified target genes that respond to Spt5 inhibition in profoundly diverse kinetics. While the relief of Spt5 inhibition from heat shock genes by SPIs resulted in their transient upregulation, the opposite is seen with GDF15, in which its upregulation was apparent only upon prolonged Spt5 inhibition. As heat shock genes were absent from Spt5 KD studies, this observation highlights the advantage of small molecules in discriminating between temporary (direct) and possibly indirect target genes. The dramatic upregulation of GDF15 upon Spt5 inhibition by siRNA or by SPIs points to the involvement of Spt5 in metabolic functions related to food intake and control of body weight. Furthermore, the use of these SPIs revealed a previously unknown role of Spt5 in promoting the unique 30 end processing of histone genes over use of the polyadenylation pathway. As earlier work implicated NELF in the 30 end processing of histone genes (Narita et al., 2007), it is plausible that DSIF and NELF cooperate in controlling this process. The examination of the effects of SPIs on the expression of WT and mutant Htt genes provided further support for the reported importance of Spt4/Spt5 for the transcription of the mutated expanded repeat Htt gene (Cheng et al., 2015; Liu et al., 2012). Notably, some of the identified compounds inhibited mutant Htt expression, with no apparent effect on the expression of WT Htt and pro-inflammatory genes. In addition, the effective concentration of 2 other compounds, SPI-86 and SPI-31, for inhibiting Htt expression was substantially lower than that required to diminish the expression of pro-inflammatory genes. These findings imply that the Htt-specific compounds may be useful as leads for therapy against HD and also suggest that the underlying basis of the Spt4/Spt5 requirement for Htt gene expression is distinct from that of inflammatory gene expression. This notion is supported by the finding that Spt4 is dispensable for pro-inflammatory gene expression (Diamant et al., 2016a), while it is crucial for mutant Htt expression (Cheng et al., 2015; Liu et al., 2012). Thus, the development of these compounds is expected to show therapeutic potential. A challenging aspect with SPIs relates to their specificity. While the issue could not be fully resolved in this study, our experiments provide several arguments on the subject. First, a major advantage of the screening approach used here is the use of
recombinant proteins, which led to the identification of compounds that directly interact with the target proteins. Second, the transcriptional effects of SPIs reported here are highly specific and in full agreement with the expected effects of Spt5 KD. Third, the transcriptomic data revealed that only a limited number of genes were affected and most of them were expected from Spt5 inhibition. Fourth, the effect of the compounds on Htt gene expression is highly specific to long but not to short expanded repeats. Nevertheless, the effect of these compounds on cellular processes other than transcription cannot be ruled out. In summary, we have discovered the first specific Spt5-Pol II inhibitors by applying a conformation-sensitive protein-protein interaction drug screen. The use of SPIs validated the known activities exerted by Spt5 in mammalian cells and helped to reveal novel functions. The identified compounds provide an excellent approach for assessing the function of Spt5 in complex in vivo settings. Further development of SPIs is expected to lead to their preclinical and clinical evaluations as a therapeutic approach against metabolic and neurodegenerative disorders. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d d
d d
KEY RESOURCES TABLE LEAD CONTACT AND MATERIALS AVAILABILITY EXPERIMENTAL MODEL AND SUBJECT DETAILS METHOD DETAILS B Manipulations in cultured cells B Plasmid construction B High-throughput drug screening (HTS) B RNA preparation, RNA metabolic labeling, and Global Run-On B High-throughput sequencing of the total, labeled and Run-On transcripts (RNA-Seq and Gro-seq) B RNA-Seq and GRO-seq Bioinformatics analysis B Chromatin immunoprecipitation (ChIP) B Cells extract, western blotting B Co-immunoprecipitation (CO-IP) B Binding assay B Cells viability assay QUANTIFICATION AND STATISTICAL ANALYSIS DATA AND CODE AVAILABILITY
SUPPLEMENTAL INFORMATION Supplemental Information can be found online at https://doi.org/10.1016/j. molcel.2019.08.024. ACKNOWLEDGMENTS This work was supported in part by grants from the Nella and Leon Benoziyo Center for Neurological Diseases, Yeda Research and Development Co. Ltd., the Estate of David Levinson, the Estate of Fannie Sherr, and the Israel Innovation Authority (KAMIN). We would like to thank Galit Cohen from the Maurice and Vivienne Wohl Institute for Drug Discovery of the Nancy and Stephen Grand Israel National Center for Personalized Medicine (G-INCPM) (Weizmann Institute of Science) for contributions to the drug screen, Dr. Sima Benjamin
Molecular Cell 76, 1–15, November 21, 2019 13
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
and Dr. Shlomit Gilad from the Crown Institute for Genomics of the G-INCPM for libraries preparation and RNA-seq services, Michael Gershovits from the Ilana and Pascal Mantoux Institute for Bioinformatics of the G-INCPM, Dr. Noa Wigoda from the Weizmann core facilities for the RNA-seq data analysis, and Mr. Arsenii Hordeichyk for technical assistance in the binding assays. R.D. is the incumbent of the Ruth and Leonard Simon Chair of Cancer Research. AUTHOR CONTRIBUTIONS R.D. and A.B. conceived and designed the study, analyzed the data, and wrote the paper. A.B. carried out most of the experiments. O.L. provided technical assistance. A.P. and A.B. carried out the high-throughput drug screen. D.L. performed part of the bioinformatics analysis. DECLARATION OF INTERESTS R.D. and A.B. declare a patent application (pending) related to this work. Received: September 13, 2018 Revised: June 12, 2019 Accepted: August 26, 2019 Published: September 26, 2019 REFERENCES Aida, M., Chen, Y., Nakajima, K., Yamaguchi, Y., Wada, T., and Handa, H. (2006). Transcriptional pausing caused by NELF plays a dual role in regulating immediate-early expression of the junB gene. Mol. Cell. Biol. 26, 6094–6104. Ainbinder, E., Revach, M., Wolstein, O., Moshonov, S., Diamant, N., and Dikstein, R. (2002). Mechanism of rapid transcriptional induction of tumor necrosis factor alpha-responsive genes by NF-kappaB. Mol. Cell. Biol. 22, 6354–6362. Ainbinder, E., Amir-Zilberstein, L., Yamaguchi, Y., Handa, H., and Dikstein, R. (2004). Elongation inhibition by DRB sensitivity-inducing factor is regulated by the A20 promoter via a novel negative element and NF-kappaB. Mol. Cell. Biol. 24, 2444–2454. Amir-Zilberstein, L., and Dikstein, R. (2008). Interplay between E-box and NF-kappaB in regulation of A20 gene by DRB sensitivity-inducing factor (DSIF). J. Biol. Chem. 283, 1317–1323. Amir-Zilberstein, L., Ainbinder, E., Toube, L., Yamaguchi, Y., Handa, H., and Dikstein, R. (2007). Differential regulation of NF-kappaB by elongation factors is determined by core promoter type. Mol. Cell. Biol. 27, 5246–5259. Anders, S., Pyl, P.T., and Huber, W. (2015). HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169. Ashkenazi, S., Plotnikov, A., Bahat, A., Ben-Zeev, E., Warszawski, S., and Dikstein, R. (2016). A Novel Allosteric Mechanism of NF-kB Dimerization and DNA Binding Targeted by an Anti-Inflammatory Drug. Mol. Cell. Biol. 36, 1237–1247. Ashkenazi, S., Plotnikov, A., Bahat, A., and Dikstein, R. (2017). Effective cellfree drug screening protocol for protein-protein interaction. Anal. Biochem. 532, 53–59. Bernecky, C., Herzog, F., Baumeister, W., Plitzko, J.M., and Cramer, P. (2016). Structure of transcribing mammalian RNA polymerase II. Nature 529, 551–554. Blythe, A.J., Yazar-Klosinski, B., Webster, M.W., Chen, E., Vandevenne, M., Bendak, K., Mackay, J.P., Hartzog, G.A., and Vrielink, A. (2016). The yeast transcription elongation factor Spt4/5 is a sequence-specific RNA binding protein. Protein Sci. 25, 1710–1721. Boehm, A.K., Saunders, A., Werner, J., and Lis, J.T. (2003). Transcription factor and polymerase recruitment, modification, and movement on dhsp70 in vivo in the minutes following heat shock. Mol. Cell. Biol. 23, 7628–7637. Chen, H., Contreras, X., Yamaguchi, Y., Handa, H., Peterlin, B.M., and Guo, S. (2009). Repression of RNA polymerase II elongation in vivo is critically dependent on the C-terminus of Spt5. PLoS One 4, e6918. Cheng, H.M., Chern, Y., Chen, I.H., Liu, C.R., Li, S.H., Chun, S.J., Rigo, F., Bennett, C.F., Deng, N., Feng, Y., et al. (2015). Effects on murine behavior
14 Molecular Cell 76, 1–15, November 21, 2019
and lifespan of selectively decreasing expression of mutant huntingtin allele by supt4h knockdown. PLoS Genet. 11, e1005043. Core, L.J., Waterfall, J.J., and Lis, J.T. (2008). Nascent RNA sequencing reveals widespread pausing and divergent initiation at human promoters. Science 322, 1845–1848. Crickard, J.B., Fu, J., and Reese, J.C. (2016). Biochemical Analysis of Yeast Suppressor of Ty 4/5 (Spt4/5) Reveals the Importance of Nucleic Acid Interactions in the Prevention of RNA Polymerase II Arrest. J. Biol. Chem. 291, 9853–9870. Diamant, G., and Dikstein, R. (2013). Transcriptional control by NF-kB: elongation in focus. Biochim. Biophys. Acta 1829, 937–945. Diamant, G., Amir-Zilberstein, L., Yamaguchi, Y., Handa, H., and Dikstein, R. (2012). DSIF restricts NF-kB signaling by coordinating elongation with mRNA processing of negative feedback genes. Cell Rep. 2, 722–731. Diamant, G., Bahat, A., and Dikstein, R. (2016a). The elongation factor Spt5 facilitates transcription initiation for rapid induction of inflammatory-response genes. Nat. Commun. 7, 11547. Diamant, G., Eisenbaum, T., Leshkowitz, D., and Dikstein, R. (2016b). Analysis of Subcellular RNA Fractions Revealed a Transcription-Independent Effect of Tumor Necrosis Factor Alpha on Splicing, Mediated by Spt5. Mol. Cell. Biol. 36, 1342–1353. Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., Chaisson, M., and Gingeras, T.R. (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. Edgar, R., Domrachev, M., and Lash, A.E. (2002). Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 30, 207–210. Emmerson, P.J., Wang, F., Du, Y., Liu, Q., Pickard, R.T., Gonciarz, M.D., Coskun, T., Hamang, M.J., Sindelar, D.K., Ballman, K.K., et al. (2017). The metabolic effects of GDF15 are mediated by the orphan receptor GFRAL. Nat. Med. 23, 1215–1219. Fitz, J., Neumann, T., and Pavri, R. (2018). Regulation of RNA polymerase II processivity by Spt5 is restricted to a narrow window during elongation. EMBO J. 37, e97965. Fujita, T., Piuz, I., and Schlegel, W. (2009). Negative elongation factor NELF controls transcription of immediate early genes in a stimulus-specific manner. Exp. Cell Res. 315, 274–284. Gromak, N., West, S., and Proudfoot, N.J. (2006). Pause sites promote transcriptional termination of mammalian RNA polymerase II. Mol. Cell. Biol. 26, 3986–3996. Haimov, O., Sehrawat, U., Tamarkin-Ben Harush, A., Bahat, A., Uzonyi, A., Will, A., Hiraishi, H., Asano, K., and Dikstein, R. (2018). Dynamic interactions of eIF4G1 with eIF4E and eIF1 underlie scanning dependent and independent translation. Mol. Cell Biol. 28, e00139–18. Henriques, T., Scruggs, B.S., Inouye, M.O., Muse, G.W., Williams, L.H., Burkholder, A.B., Lavender, C.A., Fargo, D.C., and Adelman, K. (2018). Widespread transcriptional pausing and elongation control at enhancers. Genes Dev. 32, 26–41. Hirtreiter, A., Damsma, G.E., Cheung, A.C., Klose, D., Grohmann, D., Vojnic, E., Martin, A.C., Cramer, P., and Werner, F. (2010). Spt4/5 stimulates transcription elongation through the RNA polymerase clamp coiled-coil motif. Nucleic Acids Res. 38, 4040–4051. Hsu, J.Y., Crawley, S., Chen, M., Ayupova, D.A., Lindhout, D.A., Higbee, J., Kutach, A., Joo, W., Gao, Z., Fu, D., et al. (2017). Non-homeostatic body weight regulation through a brainstem-restricted receptor for GDF15. Nature 550, 255–259. Ivanov, D., Kwak, Y.T., Guo, J., and Gaynor, R.B. (2000). Domains in the SPT5 protein that modulate its transcriptional regulatory properties. Mol. Cell. Biol. 20, 2970–2983. Kim, D.K., Inukai, N., Yamada, T., Furuya, A., Sato, H., Yamaguchi, Y., Wada, T., and Handa, H. (2003). Structure-function analysis of human Spt4: evidence that hSpt4 and hSpt5 exert their roles in transcriptional elongation as parts of the DSIF complex. Genes Cells 8, 371–378.
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Klein, B.J., Bose, D., Baker, K.J., Yusoff, Z.M., Zhang, X., and Murakami, K.S. (2011). RNA polymerase and transcription elongation factor Spt4/5 complex structure. Proc. Natl. Acad. Sci. USA 108, 546–550. Kohen, R., Barlev, J., Hornung, G., Stelzer, G., Feldmesser, E., Kogan, K., Safran, M., and Leshkowitz, D. (2019). UTAP: User-friendly Transcriptome Analysis Pipeline. BMC Bioinformatics 20, 154. Komori, T., Inukai, N., Yamada, T., Yamaguchi, Y., and Handa, H. (2009). Role of human transcription elongation factor DSIF in the suppression of senescence and apoptosis. Genes Cells 14, 343–354. Kramer, N.J., Carlomagno, Y., Zhang, Y.J., Almeida, S., Cook, C.N., Gendron, T.F., Prudencio, M., Van Blitterswijk, M., Belzil, V., Couthouis, J., et al. (2016). Spt4 selectively regulates the expression of C9orf72 sense and antisense mutant transcripts. Science 353, 708–712. Lee, C., Li, X., Hechmer, A., Eisen, M., Biggin, M.D., Venters, B.J., Jiang, C., Li, J., Pugh, B.F., and Gilmour, D.S. (2008). NELF and GAGA factor are linked to promoter-proximal pausing at many genes in Drosophila. Mol. Cell. Biol. 28, 3290–3300. Li, J., and Gilmour, D.S. (2013). Distinct mechanisms of transcriptional pausing orchestrated by GAGA factor and M1BP, a novel transcription factor. EMBO J. 32, 1829–1841. Liu, C.R., Chang, C.R., Chern, Y., Wang, T.H., Hsieh, W.C., Shen, W.C., Chang, C.Y., Chu, I.C., Deng, N., Cohen, S.N., and Cheng, T.H. (2012). Spt4 is selectively required for transcription of extended trinucleotide repeats. Cell 148, 690–701. Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550.
Quinlan, A.R., and Hall, I.M. (2010). BEDTools: a flexible suite of utilities for 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. Rougvie, A.E., and Lis, J.T. (1990). Postinitiation transcriptional control in Drosophila melanogaster. Mol. Cell. Biol. 10, 6041–6045. Shetty, A., Kallgren, S.P., Demel, C., Maier, K.C., Spatt, D., Alver, B.H., Cramer, P., Park, P.J., and Winston, F. (2017). Spt5 Plays Vital Roles in the Control of Sense and Antisense Transcription Elongation. Mol. Cell 66, 77–88.e5. Stanlie, A., Begum, N.A., Akiyama, H., and Honjo, T. (2012). The DSIF subunits Spt4 and Spt5 have distinct roles at various phases of immunoglobulin class switch recombination. PLoS Genet. 8, e1002675. Subramanian, A., Tamayo, P., Mootha, V.K., Mukherjee, S., Ebert, B.L., Gillette, M.A., Paulovich, A., Pomeroy, S.L., Golub, T.R., Lander, E.S., and Mesirov, J.P. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102, 15545–15550. Tsai, V.W., Zhang, H.P., Manandhar, R., Lee-Ng, K.K.M., Lebhar, H., Marquis, C.P., Husaini, Y., Sainsbury, A., Brown, D.A., and Breit, S.N. (2018). Treatment with the TGF-b superfamily cytokine MIC-1/GDF15 reduces the adiposity and corrects the metabolic dysfunction of mice with diet-induced obesity. Int. J. Obes. 42, 561–571. Vos, S.M., Farnung, L., Urlaub, H., and Cramer, P. (2018). Structure of paused transcription complex Pol II-DSIF-NELF. Nature 560, 601–606.
Martinez-Rucobo, F.W., Sainsbury, S., Cheung, A.C., and Cramer, P. (2011). Architecture of the RNA polymerase-Spt4/5 complex and basis of universal transcription processivity. EMBO J. 30, 1302–1310.
Wada, T., Takagi, T., Yamaguchi, Y., Ferdous, A., Imai, T., Hirose, S., Sugimoto, S., Yano, K., Hartzog, G.A., Winston, F., et al. (1998a). DSIF, a novel transcription elongation factor that regulates RNA polymerase II processivity, is composed of human Spt4 and Spt5 homologs. Genes Dev. 12, 343–356.
Mayer, A., Lidschreiber, M., Siebert, M., Leike, K., So¨ding, J., and Cramer, P. (2010). Uniform transitions of the general RNA polymerase II transcription complex. Nat. Struct. Mol. Biol. 17, 1272–1278.
Wada, T., Takagi, T., Yamaguchi, Y., Watanabe, D., and Handa, H. (1998b). Evidence that P-TEFb alleviates the negative effect of DSIF on RNA polymerase II-dependent transcription in vitro. EMBO J. 17, 7395–7403.
Missra, A., and Gilmour, D.S. (2010). Interactions between DSIF (DRB sensitivity inducing factor), NELF (negative elongation factor), and the Drosophila RNA polymerase II transcription elongation complex. Proc. Natl. Acad. Sci. USA 107, 11301–11306.
Werner, F. (2012). A nexus for gene expression-molecular mechanisms of Spt5 and NusG in the three domains of life. J. Mol. Biol. 417, 13–27.
Mullican, S.E., Lin-Schmidt, X., Chin, C.N., Chavez, J.A., Furman, J.L., Armstrong, A.A., Beck, S.C., South, V.J., Dinh, T.Q., Cash-Mason, T.D., et al. (2017). GFRAL is the receptor for GDF15 and the ligand promotes weight loss in mice and nonhuman primates. Nat. Med. 23, 1150–1157. Narita, T., Yung, T.M., Yamamoto, J., Tsuboi, Y., Tanabe, H., Tanaka, K., Yamaguchi, Y., and Handa, H. (2007). NELF interacts with CBC and participates in 30 end processing of replication-dependent histone mRNAs. Mol. Cell 26, 349–365. Ni, Z., Saunders, A., Fuda, N.J., Yao, J., Suarez, J.R., Webb, W.W., and Lis, J.T. (2008). P-TEFb is critical for the maturation of RNA polymerase II into productive elongation in vivo. Mol. Cell. Biol. 28, 1161–1170. Pavri, R., Gazumyan, A., Jankovic, M., Di Virgilio, M., Klein, I., AnsarahSobrinho, C., Resch, W., Yamane, A., Reina San-Martin, B., Barreto, V., et al. (2010). Activation-induced cytidine deaminase targets DNA at sites of RNA polymerase II stalling by interaction with Spt5. Cell 143, 122–133. Peng, J., Marshall, N.F., and Price, D.H. (1998). Identification of a cyclin subunit required for the function of Drosophila P-TEFb. J. Biol. Chem. 273, 13855–13860.
Wu, C.H., Yamaguchi, Y., Benjamin, L.R., Horvat-Gordon, M., Washinsky, J., Enerly, E., Larsson, J., Lambertsson, A., Handa, H., and Gilmour, D. (2003). NELF and DSIF cause promoter proximal pausing on the hsp70 promoter in Drosophila. Genes Dev. 17, 1402–1414. Yamada, T., Yamaguchi, Y., Inukai, N., Okamoto, S., Mura, T., and Handa, H. (2006). P-TEFb-mediated phosphorylation of hSpt5 C-terminal repeats is critical for processive transcription elongation. Mol. Cell 21, 227–237. Yamaguchi, Y., Takagi, T., Wada, T., Yano, K., Furuya, A., Sugimoto, S., Hasegawa, J., and Handa, H. (1999a). NELF, a multisubunit complex containing RD, cooperates with DSIF to repress RNA polymerase II elongation. Cell 97, 41–51. Yamaguchi, Y., Wada, T., Watanabe, D., Takagi, T., Hasegawa, J., and Handa, H. (1999b). Structure and function of the human transcription elongation factor DSIF. J. Biol. Chem. 274, 8085–8092. Yang, L., Chang, C.C., Sun, Z., Madsen, D., Zhu, H., Padkjær, S.B., Wu, X., Huang, T., Hultman, K., Paulsen, S.J., et al. (2017). GFRAL is the receptor for GDF15 and is required for the anti-obesity effects of the ligand. Nat. Med. 23, 1158–1166. Yonaha, M., and Proudfoot, N.J. (1999). Specific transcriptional pausing activates polyadenylation in a coupled in vitro system. Mol. Cell 3, 593–600.
Molecular Cell 76, 1–15, November 21, 2019 15
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
STAR+METHODS KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
Antibodies Purified Mouse Anti-Human IkBa
BD-transductions
Cat#610690; RRID: AB_398013
Mouse monoclonal BrdU antibody (IIB5) conjugated to agarose
Santa Cruz
Cat#sc-32323AC; RRID: AB_626766
Mouse monoclonal c-Rel Antibody (B-6)
Santa Cruz
Cat#sc-6955; RRID: AB_670194
Mouse monoclonal NFkB p65 Antibody (F-6)
Santa Cruz
Cat#sc-8008; RRID: AB_628017
Normal mouse IgG
Santa Cruz
Cat#sc-2025; RRID: AB_737182
Mouse monoclonal [8WG16] to RNA polymerase II CTD repeat YSPTSPS - ChIP Grade
Abcam
Cat#ab817; RRID: AB_306327
Rat monoclonal antibody against DSIF p160
Wada et al., 1998a
N/A
Mouse monoclonal antibody against hSpt4
Kim et al., 2003
N/A
Recombinant Human TNF-a
Peprotech
Cat#300-01A-50
Coelenterazine, native
Gold Biotechnology
Cat#CZ5; CAS: 55779-48-1
BioTri (RNA/DNA/protein extraction using Trireagent)
Bio-Lab
Cat#959758027100
4-Thiouridine
Cayman Chemical
Cat#16373-25
Biotin-HPDP
APExBIO
Cat#A8008; CAS: 129179-83-5
Streptavidin-coated magnetic beads
NEB
Cat#S1420S
IGEPAL CA-630
Sigma
Cat#I8896; CAS: 9002-93-1
5-Bromouridine 5-triphosphate sodium salt
Santa Cruz
Cat#SC-214314A; CAS: 161848-60-8
RNA Fragmentation reagent
Thermo Fisher Scientific
Cat#AM8740
Actinomycin D
APExBIO
Cat#A4448; CAS: 50-76-0
Chemicals, Peptides, and Recombinant Proteins
D-Luciferin Firefly, Potassium Salt
Gold Bioscience
Cat#LUCK-100; CAS: 115144-35-9
nProtein A Sepharose 4 Fast Flow
GE Healthcare
Cat#17-5280-01
Protein G Sepharose 4 Fast Flow
GE Healthcare
Cat#17-0618-01
Reporter lysis buffer
Promega
Cat#E4030
Salmon sperm DNA
R&D Systems
Cat#9610-5-D
Ni-NTA His,Bind Resin
Merck Millipore
Cat#70666
tRNA
Sigma
Cat#R8508
Critical Commercial Assays High-Capacity cDNA Reverse Transcription kit
Thermo Fisher Scientific
Cat#4368814
Fast qPCRBIO SYBR Green Mix
PCR Biosystems
Cat#PB20.16-50
CellTiter-Glo Lum Cell Viability kit
Promega
Cat#G7571
Micro Bio-Spin P-30 Gel Columns
Bio-Rad
Cat#7326250
QIAquick PCR Purification Kit
QIAGEN
Cat#28106
RNA Clean & Concentrator kit
Zymo Research
Cat#ZR-R1017
Direct-zol RNA MiniPrep kit
Zymo Research
Cat#R2052
Raw and analyzed data of RNA-seq 2h
This paper
GEO:GSE136022
Raw and analyzed data of RNA-seq 24h
This paper
GEO:GSE136023
Raw and analyzed data of GRO-Seq
This paper
GEO:GSE136024
Raw and analyzed data of 4sU-seq
This paper
GEO:GSE136025
Transgenic mice: STHdhQ7/Q7
Coriell Institute for Medical Research
Cat# CH00097
Transgenic mice: STHdhQ111/Q111
Coriell Institute for Medical Research
Cat# CH00095
Deposited Data
Experimental Models: Cell Lines
(Continued on next page)
e1 Molecular Cell 76, 1–15.e1–e4, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
Continued REAGENT or RESOURCE
SOURCE
IDENTIFIER
MCF7
ATCC
N/A
HeLa
ATCC
N/A
HEK293T
ATCC
N/A
Mouse embryonic fibroblasts (MEFs)
ATCC
N/A
Jurkat
ATCC
N/A
This paper
Table S1
Plasmid: RSV-DC/DN Renilla luciferase (pRL-DC/DN)
Ashkenazi et al., 2017
N/A
Plasmid: pRSFDuet-1 DNA
Merck-Novagen
Cat# 71341
Plasmid: DC RL-spt5
This paper
N/A
Oligonucleotides Primers for RF-cloning, ChIP and real-time see Table S1 Recombinant DNA
Plasmid: Pol II-DN RL
This paper
N/A
Plasmid: His-DC RL-spt5/Flag-Pol II-DN RL in pRSFDuet
This paper
N/A
Software and Algorithms GeneData software GSEA database and software
https://www.genedata.com/products/ profiler/software/ Subramanian et al., 2005
http://software.broadinstitute.org/gsea/ msigdb/index.jsp
LI-COR Image Studio Software
https://www.licor.com/bio/image-studio/
OriginLab for IC50 calculation
https://www.originlab.com/
Others RNA-Seq data from TNFa and Spt5 KD HeLa cells
Diamant et al., 2016b
GEO:GSE70268
LEAD CONTACT AND MATERIALS AVAILABILITY Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contacts (Rivka Dikstein). EXPERIMENTAL MODEL AND SUBJECT DETAILS HEK293T, HeLa, MCF7, Jurkat and mouse embryonic fibroblasts (MEFs) cells were grown and maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum (Invitrogen) and 1% penicillin-streptomycin. To avoid basal NF-lB activity, HeLa cells were kept from reaching confluence and re-plated after initial thawing no more than 10 times. To induce NF-lB activity, cells were treated with 20 ng/ml recombinant human TNF-a (Peprotech) for the indicated times. STHdh Q111 and STHdh Q7 striatal cell lines (Coriell Institute for Medical Research) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, 1% penicillin-streptomycin and 0.4 mg/ml G418 at 33C. The cells were re-plated no more than 5-6 times, as recommended. METHOD DETAILS Manipulations in cultured cells Spt5 knockdown (KD) was carried out as previously described (Diamant et al., 2012). Briefly, 1,250,000 cells were plated in 10cm dish and transfected 24h later using ICAFectin441 (In-Cell-Art) and 12 mg of either pSUPER or a mix of two distinct pSUPER-spt5-RNAi, together with 0.5 mg CMV-GFP-Puro plasmid. Transfected cells were selected for 24h with puromycin (1 mg/ml). Plasmid construction The mammalian expression plasmids RSV-DC/DN Renilla luciferase (pRL-DC/DN), encoding the N-terminal amino acid residues 1-229 or the C terminus 230-311 of the Renilla luciferase were previously described (Ashkenazi et al., 2016, 2017). The split-Renilla luciferase fusion plasmids were constructed by two-steps PCR using the pRL-DC or pRL-DN as backbones and plasmids encoding
Molecular Cell 76, 1–15.e1–e4, November 21, 2019 e2
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
human Rpb1, Spt5 and Spt4 as insert sources. For bacterial expression of the split-RL fusion proteins, the double expression plasmid pRSFDuet was used. All constructs were verified by sequencing. The primer sequences are shown in Table S1. High-throughput drug screening (HTS) The HTS was carried out in 1536 well plate format as previously described (Ashkenazi et al., 2017), with the following modifications: the final concentration of each drug was 15 mM and the incubation time with the Coelenterazine (GOLDBIO, CZ5) RL substrate was 30 min at RT. Compounds that diminished the luminescence activity by > 30% were subjected for an additional screen with bacterial lysate expressing recombinant full-length RL (diluted 1:500 in phosphate buffer) that served as a control for false-positive compounds that inhibit the RL enzymatic activity itself. Compounds that did not affect RL activity were further analyzed in a doseresponse assay (0.55 mM, 1.6 mM, 5 mM, 15 mM and 45 mM) in duplicate. The selected compounds were further subjected to livecell split-RL assay as described (Ashkenazi et al., 2017). The analysis of the luminescence data for all the above-described assays was performed using GeneData software (Basel, Switzerland). RNA preparation, RNA metabolic labeling, and Global Run-On Total RNA was extracted using Tri-reagent (Bio-Lab Chemicals) and Direct-zol RNA MiniPrep kit (Zymo Research). For gene-specific analysis cDNA was synthesized from 0.5-1 mg of total RNA using the High-Capacity cDNA Reverse Transcription Kit (ABI, Thermo Fisher Scientific) with a poly-dT primer or random hexamers as indicated for each experiment. cDNA samples were analyzed by quantitative PCR (qPCR) in a ViiA 7 Real-Time PCR System using Fast qPCRBIO SYBR Green Mix (PCR Biosystems). The primer sequences are shown in Table S1. For metabolic labeling of newly synthesized RNA with 4-thiouridine (4sU), cells were treated with SPI-21 or DMSO for 30 min. Then 4sU (150 mM final concentration) was added for an additional 2h. The labeling was terminated by removing the 4sU-containing medium and washing the cells 3 times with phosphate-buffered saline (PBS). Total RNA was extracted as described above. 10 mg of RNA were biotinylated using biotin-HPDP (1:5; A8008, APExBIO) for 1.5 h at room temperature. The biotinylated RNA was purified using chloroform and isopropanol extraction and dissolved in DEPC water. Thereafter the biotinylated RNA was captured using Streptavidin-coated magnetic beads (NEB, S1420S) for 30 min at room temperature in constant rotating. Bound RNA was eluted with 100 mL of 0.1 M fresh DTT. RNA isolation was performed using the RNA Clean & Concentrator kit (Zymo Research). cDNA was prepared as described above using random hexamers. For the Global Run-On experiment, HeLa cells (15 cm plate x 3) were treated with SPI-21 (50 mM) for 1.5h, washed 3 times with cold PBS, collected and pelleted at 1000 X g for 5 min at 4 C. Cells were swollen by resuspension in 2 mL swelling buffer [10 mM Tris-Cl pH 7.5, 10% glycerol, 3 mM CaCl2, 3 mM MgCl2, protease inhibitor cocktail (EDTA free), and 4 units/ml of RNase inhibitor], incubated at 4 C for 20 min and pelleted again at 1000 X g for 5 min at 4 C. Cells were resuspended in 1 mL of ice-cold lysis buffer (10 mM Tris-Cl pH7.5, 300 mM sucrose, 10 mM NaCl, 3 mM CaCl2, 2 mM MgCl2, 0.5% Igepal, 0.5 mM DTT, protease inhibitors and RNase inhibitor) and then homogenized and lysed by transferring them 5 times through 27G needle. Nuclei were pelleted at 1000 X g for 5 min at 4 C, washed once with 10 mL of lysis buffer and once with 1 mL storage buffer (50 mM Tris-Cl pH 8.0, 25% glycerol, 5 mM MgCl2, 0.1 mM EDTA, 5 mM DTT). For the Run-on reaction nuclei were resuspended at 5 3 10^6 nuclei/100 ml of storage buffer and mixed with equal volume of reaction buffer (10 mM Tris-Cl pH 8.0, 5 mM MgCl2, 1 mM DTT, 300 mM KCl, 20 units of RNase Inhibitor, 1% sarkosyl, 500 mM ATP, GTP, CTP and Br-UTP). The reaction was allowed to proceed for 5 min at 30 C, followed by the addition of 1 mL Trizol. RNA was further extracted with Direct-zol RNA MiniPrep kit and DNase I treatment was performed. Fragmentation of NRO-RNA was performed using fragmentation reagents (AM8740) for 10 min at +70 C and then terminated by addition of stop solution (1:1). The fragments were purified by Micro Bio-Spin Columns with Bio-Gel P-30 (BioRad) according to the manufacturer’s instructions. BrU-labeled RNA was purified with pre-blocked anti-BrdU beads and eluted 4 times (10 min each) with elution buffer (20 mM DTT, 300 mM NaCl, 5 mM Tris-cl pH 7.5, 1 mM EDTA, and 0.1% SDS) at +42 C while shaking. Eluted RNA was extracted using Trizol and Direct-zol RNA MiniPrep. High-throughput sequencing of the total, labeled and Run-On transcripts (RNA-Seq and Gro-seq) The RNA-seq libraries were prepared by the Genomics unit in The Nancy and Stephen Grand Israel National Center for Personalized Medicine (G-INCPM, Weizmann Institute of Science). Samples from 2h SPI-21 treatment were prepared using in-house mRNA Seq protocol and polyA capturing and those from 24h and 4sU treatments were prepared using the same protocol w/o PolyA capture (with random hexamers), and sequenced using an Illumina HiSeq 2500 system yielding 30M SE reads of 61 bases. The GRO-seq RNA samples were reverse transcribed and amplified (15 cycles) using in-house mRNA Seq protocol w/o PolyA capture (with random hexamers) and sequenced in one lane using an Illumina HiSeq High Output (480 million reads, yielding more than 100M SE reads of 47 bases per sample). Each RNA-seq analysis was carried out using two independent replicates. RNA-Seq and GRO-seq Bioinformatics analysis For the analysis of the 24h, 4sU and GRO-seq data, we used the RNA-seq pipeline of UTAP (Kohen et al., 2019) to determine the differentially expressed genes. For the 2h data, reads were mapped to the H. sapiens GRCh38 reference genome using STAR (Dobin et al., 2013) and expression levels for each gene were quantified using htseq-count (Anders et al., 2015). Differentially expressed genes were identified using DESeq2 (Love et al., 2014). The pausing index (PI) analysis was determined for the upregulated genes
e3 Molecular Cell 76, 1–15.e1–e4, November 21, 2019
Please cite this article in press as: Bahat et al., Targeting Spt5-Pol II by Small-Molecule Inhibitors Uncouples Distinct Activities and Reveals Additional Regulatory Roles, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.08.024
from the GRO-seq experiment with at least 50 reads in the DMSO sample. The reads coverage was calculated using bed files containing the desired genomic regions and orientation (awk command) and the bedtools coverage tool (Quinlan and Hall, 2010). Chromatin immunoprecipitation (ChIP) HeLa cells were treated with different SPIs (50 mM; 60 min) or DMSO as a control, followed by TNF-a induction (30 min) in 100-mm plates. The cells were then cross-linked with 1% formaldehyde for 10 min at room temperature, and fixing was terminated by adding a 1/20 volume of 2.5 M glycine. Chromatin extraction and immunoprecipitations were carried out as previously described (Diamant et al., 2012). For immunoprecipitation, 3 mg of anti-RNA polymerase II, anti-Spt5, anti-p65 or control IgG was added to 1 mL of the soluble chromatin DNA. IP and input samples were analyzed by qPCR. The primer sequences are shown in Table S1. Cells extract, western blotting The whole-cell extract was prepared using commercial Reporter lysis buffer (E4030, Promega). Samples were then separated by SDS-PAGE and subjected to western blotting. Western blots were quantitated using Image Studio Software. The source of the antibodies used throughout the study is shown below. Co-immunoprecipitation (CO-IP) HeLa cells were detached from 100-mm plate with ice-cold PBS and a cell scraper, centrifuged for 5 min at 2,000 rpm and resuspended in 200 mL of lysis buffer (20 mM Tris-HCl pH 8, 400 mM NaCl, 2 mM EDTA, 0.5% NP-40, 10% glycerol) to which protease inhibitor cocktail (1:100) and PMSF (200 mM) were freshly added. The lysates were rotated for 10 min at 4C, and centrifuged for 15 min at 13,000 rpm. The supernatants were diluted with lysis buffer without NaCl to reach a final concentration of 100 mM NaCl. The cell lysate was then incubated overnight with anti-Spt5 antibody or rabbit IgG at 4C, followed by incubation with Protein A/G Sepharose beads (GE Healthcare) for 2h at 4C. The lysate was washed 3 times and then split and incubated with different SPIs or DMSO for 2h at 4C. The immunoprecipitated proteins were detected by SDS-PAGE followed by western blot using anti-Spt5, anti-Pol II and anti-Spt4 proteins antibodies. Binding assay To measure binding of SPIs to Spt5, Pol II and control protein (eIF4E), these proteins were expressed in E. coli as a fusion with histidine tag and purified on nickel agarose beads as described previously (Ashkenazi et al., 2016). The proteins were incubated with increasing concentrations of SPIs in black microplate 384 wells in a total volume of 20 mL for 10 min. The binding was assessed by measuring the changes in the intrinsic fluorescence (280 nm) using Cytation 5 (BioTek). IC50 was calculated and graphed using AAT Bioquest online calculator (https://www.aatbio.com/tools/ic50-calculator). Cells viability assay HeLa cells were plated in an opaque-walled 96 well plate (50,000 cells per well) and treated with increasing concentrations of SPI-21 or SPI-18 for 48h. After equilibrating the plate and its contents at room temperature for 30 min, 100 mL (1:1) of CellTiter-Glo Reagent (G7571, Promega) was added to each well and shook for 2 min to induce cell lysis. Following 10 min incubation at room temperature, the luminescent signal was detected using the Cytation 5 instrument. QUANTIFICATION AND STATISTICAL ANALYSIS To calculate p values, Student’s t tests (typically one-tail, paired) were performed. The significance of the difference in the median values was calculated by the Wilcoxon Signed Rank test. Significance symbols in all experiments are: * = p < 0.05; ** = p < 0.01; *** = p < 0.005; **** = p < 0.001. DATA AND CODE AVAILABILITY The RNA-seq datasets generated during this study have been deposited in NCBI’s Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE136026.
Molecular Cell 76, 1–15.e1–e4, November 21, 2019 e4