Pontin Functions as A Transcriptional Co-activator for Retinoic Acid-induced HOX Gene Expression

Abstract

Pontin is a AAA+ ATPase protein that has functions in various biological contexts including gene transcription regulation, chromatin remodeling, DNA damage sensing and repair, as well as assembly of protein and ribonucleoprotein complexes. Pontin is known to regulate the transcription of several important signaling pathways, including Wnt signaling. However, its role in early embryonic signaling regulation remains unclear. Retinoic acid (RA) signaling plays a central role in vertebrate development. Using an in vivo biotin tagging technology, we mapped the genome-wide binding pattern of Pontin before and after RA-induced differentiation in the pluripotent embryo carcinoma cell line NTERA-2. Biotin ChIP-seq revealed significant changes in genome-wide Pontin binding sites upon RA stimulation. We also identified a substantial amount of overlapping binding peaks between Pontin and RARα, especially on all of the HOX gene loci (A-D clusters). Pontin knockdown experiments showed that its chromatin binding at the HOX gene clusters is required for RA-induced HOX gene expression. Furthermore, we performed Global Run-On sequencing (GRO-seq) to map de novo transcripts genome-wide and found that Pontin knockdown significantly diminished nascent HOX gene transcripts, indicating that Pontin regulates HOX gene expression at the transcriptional level. Finally, proteomic analysis demonstrated that Pontin associates with chromatin organization/remodeling complexes and various other functional complexes. Altogether, we have demonstrated that Pontin is a critical transcriptional co-activator for RA-induced HOX gene activation.

Graphical Abstract

Introduction

Chromatin remodeling plays a critical role during embryonic development. Pontin is a chromatin remodeling factor with both ATPase and DNA helicase activities and is evolutionarily conserved [1–3]. Pontin is found to be an integral subunit of different chromatin-modifying complexes, included in the Ino80 complex, the Swr1 complex, and Tip60 complexes [4–6]. It interacts with numerous other target proteins to regulate transcription and protein complex assembly [7–9]. Pontin has also been implicated in multiple processes encompassing the regulation of gene transcription, DNA damage sensing and repair, assembly of protein and ribonucleoprotein, and tumor biology [3, 8, 10–12]. Pontin has been reported to function as a co-activator for various transcription factors including androgen receptor (AR) in prostate cancer, T-cell factor (TCF) in the Wnt signaling pathway, and regulation of p53 and JNK activity [12–18]. Additionally, Pontin is implicated in proper embryonic development in multiple organisms by complexing with c-Myc and regulating proper organ development [19–21]. However, much is still unknown about the function of Pontin and the mechanisms it uses, as it interacts with many molecular complexes that regulate vastly different downstream effectors.

Retinoic acid (RA) is a universal differentiation agent that regulates transcription by interacting with nuclear RA receptors (RARs) bound to RA response elements (RAREs) near target genes [22, 23]. RA signaling is mediated through one of the three retinoic acid receptors RARα, β, and γ that form heterodimerized complexes with the rexinoid receptors RXR α, β, and γ [24], as well as fibroblast growth factor (FGF) signaling [25]. In the absence of RA, RARs and RXRs associate with repressive complexes on RAREs to inhibit transcription [26]. The interaction between RA and RARs results in conformational changes of RARs, allowing recruitment of co-activators and transcriptional activation [27, 28]. It has also been reported that, for some target genes of RA signaling, RARs and RXRs are only recruited to RARE in the presence of RA ligand [29]. RA-RAR signaling is indispensable for proper early stage embryonic development and RA deletion results in impaired development of muscles, neurons, and heart [23, 30, 31].

Homeobox (HOX) genes are a highly conserved family of homeodomain containing transcription factors that are under the direct control of RA in embryonic development [32]. HOX genes are essential for patterning the anterior-posterior axis of developing embryos. In addition to their roles in specifying positional identity during development, they have been implicated in cell proliferation, inflammation, epithelial-mesenchymal transition, and cancer progression [33–35]. Early studies using embryonic stem cells (ESCs) demonstrated that RA treatment results in rapid HOX gene response and induces cell differentiation [36–38]. Further analyses in this system revealed direct regulation of HOX gene expression by RA signaling through RAREs located in HOX gene loci. Extensive studies using vertebrate models also demonstrated that RA is a morphogen regulating anterior-posterior patterning through HOX gene regulation [39–42]. More recently, functional enhancers have been identified to mediate RA-induced HOX gene expression and ESCs differentiation [43]. Interestingly, Pontin has been reported to act as a co-activator within the TrxG complex to maintain HOX gene transcription in Drosophila [44]. However, whether and how Pontin is involved in HOX gene regulation in mammalian cells has not been explored.

In this study, we utilized pluripotent NTERA-2 cells to study the function of Pontin in RA-induced gene expression. Our ChIP-seq identified genome-wide co-binding of Pontin and RARα in response to RA signaling, especially on all of the HOX gene loci (A-D clusters). We show that Pontin is required for RA-induced HOX gene expression. We further mapped the global nascent transcripts with GRO-seq and demonstrated that Pontin regulates HOX gene expression at the transcriptional level. Finally, our interactome network analysis identified the association of Pontin with multiple functional complexes. Together, our work revealed a novel role of Pontin as a critical transcriptional co-activator in regulating RA-induced HOX gene activation.

Results

Identification of genome-wide binding sites of Pontin in response to RA signaling

To investigate the function of Pontin, we first generated NTERA-2 cell lines stably expressing a biotin ligase recognition peptide (BLRP) tagged full length Pontin protein, as well as bacterial biotin ligase (BirA) that specifically biotinylates BLRP (Supplemental Figure 1A). This in vivo Biotin tagging system has been successfully applied to study protein-protein and protein-DNA interactions [45–47]. We confirmed that the expression of BLRP-Pontin was at a similar level of endogenous Pontin in these cells with Western blots (Supplemental Figure 1B). The expression of biotin-labeled Pontin was controlled by a Tet-on promoter and could be eluted out from streptavidin magnet beads with TEV protease digestion (Supplemental Figure 1C). We next performed ChIP-seq using NTERA-2 cells, with or without RA treatment, to map genome-wide binding of Pontin. We found that Pontin binding sites were predominantly mapped to promoters and intergenic regions, regardless of RA treatment (Figure 1A), consistent with the known function of Pontin in transcription regulation and chromatin remodeling. Intriguingly, RA treatment triggered a global change in Pontin binding, with 24,977 sites identified specifically from RA-treated NTERA-2 cells (Figure 1B). We then organized Pontin binding sites into three groups: Veh-specific, Common and RA-specific (Figure 1C and Supplemental Table 1). Representative binding peaks from each group were provided to show the changes of Pontin binding in response to RA signaling (Figure 1D–F). Therefore, these data indicate that RA signaling induces a genome-wide change in Pontin binding sites on chromatin.

Figure 1.

ChIP-seq identified changes in genome wide binding of Pontin in response to RA signaling.

A. Genomic annotation of the total peaks of Pontin ChIP-seq in NTERA-2 cells with and without RA treatment. Peak number distributions at distinct regions are presented. Pontin binding sites are predominantly located in promoter and distal intergenic regions.

B. Venn diagram of Pontin ChIP-seq showing the total peak numbers and overlapping peaks between conditions with or without RA treatment.

C. Heatmap representation of Pontin ChIP-seq tag density for the three groups of Pontin binding peaks: vehicle-specific, common, and RA-specific. The binding strength of Pontin on these sites were significantly altered following RA treatment.

D-F. Genome browser view of Pontin ChIP-seq signals in the absence or presence of RA signaling on one representative genomic locus from each group: RA specific (D), Common (E), and Veh-specific (F).

RA signaling triggers global binding of Pontin to RARα-bound sites

RA signaling is predominantly mediated by RARs, which form heterodimers with RXRs and bind to RAREs in RA-responsive genes [48, 49]. The induction of Pontin binding by the RA signaling prompted us to test whether Pontin was recruited to RAR sites in response to RA treatment. We performed ChIP-seq for RARα and compared binding sites of RARα and Pontin in NTERA-2 cells with or without RA stimulation. In the absence of RA signaling, Pontin and RARα shared 861 common binding sites. In cells treated with RA, the number of common binding sites between Pontin and RARα increased to 3,661 (Figure 2A and Supplemental Table 2). On these 3,661 common sites, there was strong RARα binding in the absence of RA treatment, and the binding strength of RARα was slightly but significantly enhanced upon RA stimulation (Figure 2B, 2C). In contrast, binding of Pontin on these 3, 661 sites was relatively weak in the vehicle control condition, and was dramatically augmented in response to RA treatment. Within these 3,661 common sites, RXR and RAR motifs were significantly enriched (Figure 2D), consistent with the strong binding of RARα on these sites. Furthermore, we found that the level of H3K36me3, a marker of transcriptionally active regions [50], was elevated following RA treatment (Figure 2E). Therefore, these data suggest that RA signaling promotes the binding of Pontin to RARα-bound sites to activate transcription.

Figure 2.

RA signaling induces co-binding of Pontin and RARα.

A. Venn diagrams showing overlapping ChIP-seq binding sites between Pontin and RARα in the absence (top panel) or presence (bottom panel). The number of overlapping sites increased from 861 to 3,661 after RA treatment.

B. Heatmap representation of Pontin/RARα ChIP-seq tag density on the 3,661 Pontin/RARα co-binding peaks under the indicated conditions. The binding of Pontin on these sites was greatly augmented, while the binding of RARα was only slightly elevated.

C. Aggregation plots of RARα ChIP-seq (left) and Pontin ChIP-seq (right) on the 3,661 Pontin/RARα co-binding peaks. Pontin binding showing a more dramatic enhancement following RA treatment.

D. RXR and RAR motifs were significantly enriched in the 3,661 Pontin/RARα co-binding sites.

E. Aggregation plots of H3K36me3 ChIP-seq signals (+-RA) along the gene bodies of genes in the closest proximity to the 3,661 Pontin/RARα co-binding peaks.

Pontin binds to RARα-occupied HOX gene clusters upon RA stimulation

Having shown that RA signaling regulates global Pontin binding, we next examined how RA treatment affects Pontin binding on specific sites. HOX genes are direct targets of RA signaling during the development of different tissues [51]. Indeed, we detected RARα binding on HOXA, B, C, and D gene clusters regardless of the absence or presence of RA (Figure 3A–D). Consistent with the effect of RA signaling on the genome-wide binding pattern of Pontin, Pontin was recruited to RARα sites on HOX gene clusters upon RA stimulation (Figure 3A–D). Previous studies have suggested that RARs co-bind with repressors to some target genes to inhibit target gene expression in the absence of ligand, and that the presence of RA will allow the recruitment of co-activators and the relief of co-repressors, subsequently activating gene expression [51]. We found that the recruitment of Pontin to HOX gene clusters was associated with increased H3K36me3 signals on these loci, suggesting that Pontin functions as a co-activator of RARα to regulate HOX gene expression.

Figure 3.

Pontin is recruited to RARα-occupied HOX gene clusters to activate HOX gene expression upon RA stimulation.

A-D. Genome browser views of Pontin, RARα and H3K36me3 ChIP-seq signals on HOXA (A), HOXB (B), HOXC (C) and HOXD (D) gene clusters in the absence or presence of RA treatment. Pontin binding signals were increased at the sites of RARα binding upon RA stimulation. This increasement was associated with elevated H3K36me3 signals. The shaded regions highlight the changes in ChIP-seq signals following RA treatment.

Pontin is required for RA-induced HOX gene expression

After finding that Pontin was recruited to HOX gene clusters in response to RA stimulation, we next sought to determine whether Pontin is required for RA-induced HOX gene expression. We generated NTERA-2 cells expressing Tet-on inducible Pontin shRNA and induced Pontin knockdown using doxycycline treatment. We observed efficient protein and mRNA knockdown of Pontin following doxycycline induction of two different Tet-on shRNAs for Pontin (Figure 4A). We next performed qRT-PCR to determine how Pontin knockdown affects HOX gene expression. We observed that gene expression of HOXA1, HOXA2, HOXB1, HOXB2, HOXC4, and HOXD1 were all significantly upregulated by RA in control cells, consistent with the known roles of HOX genes as direct effectors of RA signaling [51]. However, when Pontin was depleted, the RA-induced HOX gene expression was significantly diminished (Figure 4B–E). These data demonstrate that Pontin is required for RA-induced HOX gene activation, further supporting that Pontin can act as a co-activator of RARα to activate gene expression.

Figure 4.

Pontin knockdown diminished RA-induced expression of HOX genes.

A. Doxycycline-inducible Pontin knockdown in two NTERA-2 stable cell lines established with two independent Tet-on Pontin shRNAs. Doxycycline-induced Pontin knockdown was confirmed with Western blot and RT-PCR. Both Tet-on shRNAs were able to effectively knock down Pontin at mRNA and protein levels.

B-E. Pontin knockdown inhibits RA-induced expression of selected HOX genes. mRNA levels of HOXA1 and HOXA2 (B), HOXB1 and HOXB2 (C), HOXC4 (D), HOXD1 (E) measured by qRT-PCR were greatly augmented in response to RA signaling. Upon Pontin knockdown, this RA-induced augment in HOX gene mRNA levels was significantly dampened. Data represented as means ± SEM. n = 3, * p < 0.05, ** p < 0.01.

Pontin is required for de novo HOX mRNA transcription in response to RA signaling

Our RT-PCR assays described above demonstrate the important role of Pontin in RA-induced HOX gene expression. However, RT-PCR detects steady-state level of total RNA, which does not accurately reflect the dynamic transcriptional level. In fact, RNA stability plays an important role in controlling steady-state level of RNA [52]. To examine the role of Pontin in regulating RA-dependent transcriptional activity, we performed Global Run-On sequencing (GRO-seq) to assess nascent transcription from engaged RNA polymerase [53]. As expected, we detected low level of GRO-seq signals at the HOX gene loci in cells without RA treatment, indicating low transcriptional activity in the absence of RA signaling (Figure 5A–E). Upon RA stimulation, we observed significantly elevated de novo transcription from HOX gene loci. This elevation of GRO-seq signals was associated with the recruitment of Pontin to HOX gene clusters detected by ChIP-seq (Figure 5A–E). However, in cells with DOX-induced Pontin knockdown, this RA-induced transcriptional elevation was compromised (Figure 5A–E and Supplemental Figure 2A–E), indicating that Pontin is required for the transcriptional activation of HOX genes in response to RA signaling.

Figure 5.

Pontin plays its role at the transcriptional level to regulate RA-dependent HOX gene expression.

A-D. Genome browser views of GRO-seq signals at HOXA1 and HOXA2 (A), HOXB1 and HOXB2 (B), HOXC4 (C), HOXD1 (D) gene loci. RA signaling activated nascent RNA transcripts, which were measured by GRO-seq, from HOX gene loci. This RA-dependent transcriptional activation was greatly abolished upon doxycycline-induced Pontin knockdown. ChIP-seq signals are also presented to show that nascent transcription was associated with RA-triggered recruitment of Pontin to these sites.

E. Quantification of GRO-seq signals for HOXA1, HOXA2, HOXB1, HOXB1, HOXC4 and HOXD1 genes, showing that RA-dependent transcriptional activation of these genes was greatly abolished upon doxycycline-induced Pontin knockdown. To measure the nascent transcriptional level of each HOX gene, we counted the reads mapped to the whole gene body of each gene from transcription start site to 1kb downstream of the transcription stop site, and further normalized them to 1 million reads of the mapped library size. For each condition, the expression levels of the HOX genes were log2 transformed for comparison.

Pontin interacts with various functional protein complexes

A basic aspect of the mechanism by which Pontin regulates gene transcription is its interaction with other molecular complexes. We next utilized NTERA-2 cells expressing BLRP-tagged Pontin to identify Pontin interacting proteins. Using SDS-PAGE, we confirmed the isolation of tagged Pontin protein complexes (Figure 6A), and identified a number of known or novel Pontin-interacting protein partners using mass spectrometry (Figure 6B and Supplemental Table 3). Along with the Pontin bait and its partner protein Reptin, we identified components of various chromatin remodeling complexes that are known to associate with Pontin, including INO80A, INO80K, P400, SRCAP, DNAPK, PARP1, and TIP60 (Figure 6B). When we performed network analysis on the Pontin interactome, we found that proteins involved in chromatin/chromosome organization and proteins related to ATP-dependent chromatin remodeling were highly enriched (Figure 6C). Proteins involved in histone acetylation, ribosome assembly, ribonucleoprotein complex biogenesis and mRNA processing were also enriched in the protein-protein interaction network of Pontin. Therefore, these findings are consistent with the known role of Pontin as a chromatin modifier, and highlight that Pontin interacts with various cofactors to fulfill its context-specific roles. Although under the regular NTERA-2 cell growing condition, we did not identify RAR protein in the complexes pulled down by BLRP-tagged Pontin (Figure 6A–C and Supplemental Table 3), we detected a strong interaction between RARα and Pontin when we treated NTERA-2 cells with RA (Figure 6D). This is consistent with the RA-induced binding of Pontin at HOX gene clusters detected by ChIP-seq experiments (Figure 3A–D). Altogether, our data suggest that Pontin binds to HOX gene clusters and functions as a transcriptional co-activator for RAR to activate HOX Gene Expression upon RA signal stimulation.

Figure 6.

Pontin interactions with various protein functional complexes.

A. BirA-BLRP in vivo biotinylation system was used to identify Pontin-interacting proteins through mass spectrometry. Doxycycline-inducible BLRP-tagged Pontin was expressed in NTERA-2 cells growing under regular medium and biotinylated Pontin was pulled-down along with its interacting proteins using streptavidin beads. Pontin protein is then eluted using TEV protease. The SDS-PAGE gel image confirmed the isolation of BLRP-tagged Pontin protein complexes.

B. Identification of Pontin interacting co-factors by mass spectrometry analyses on the protein complexes pulled down from nuclear fractions of biotinylated Pontin. Peptide numbers detected from mass spectrometry are listed.

C. Network visualization of statistically over-represented biological processes for the Pontin interactome. Nodes represent enriched processes that are grouped according to network connectivity indicated by Kappa Score. Node size is proportional to the significance level.

D. Co-IP assays in NTERA-2 cells treated with 5uM retinoic acid for 24 hours confirming protein-protein interactions of endogenous RARα and Pontin under RA stimulation condition. Nuclear fractions were used for immunoprecipitation with antibody against RARα and blotted with RARα and Pontin antibodies respectively.

Discussion

For the first time, we have demonstrated that Pontin is a RA responsive protein that interacts with RARα on chromatin. We have shown that Pontin is recruited to HOX gene clusters following RA treatment, and that RA-induced HOX gene transcription activation is abrogated following Pontin knockdown, providing the first evidence that Pontin acts as a transcriptional co-activator of RARα to control the expression of target genes of RA signaling.

Pontin is a versatile protein with proposed functions ranging from chromatin remodeling, transcriptional regulation, modulating mitotic spindle function, DNA repair, and mRNA processing to processing small nucleolar RNAs. It plays its roles by interacting with distinct protein complexes and interacting with distinct co-factors in different cellular processes. Indeed, we found that Pontin is associated with chromatin remodeling complexes and various other functional complexes (Figure 6). Pontin is an AAA+ superfamily member and has conserved Walker A and Walker B motifs that are important for ATP binding and hydrolysis. It has been previously reported that the ATPase activity of Drosophila Pontin is critical for its role in regulating Hox gene transcription [44]. Pontin also shares homology with the bacterial RuvB helicase and is a putative DNA helicase [54]. However, its helicase activity is not well characterized and remains controversial. We speculate that, after its binding to RAR sites in response to RA signaling, Pontin might stabilize RAR complex or facilitate recruitment of other components independent of the helicase activity. Further studies are required to test this notion.

Pontin and its partner protein Reptin are normally co-expressed and often co-bind to common sites on chromatin. They form multimeric ring systems and associate with various chromatin-modifying and/or transcription-regulating multiprotein complexes [55]. Mutation of either gene leads to a lethal phenotype, indicating that they play non-redundant essential roles during development. They are also found to have opposite activities in various contexts of transcriptional control, in which Pontin functions as a transcriptional co-activator and Reptin acts as a co-repressor [55]. Their opposite activities are likely mediated by different protein complexes, as so far, there is no evidence to show that complexes containing both Pontin and Reptin can play both active and repressive roles in the same target genes. For instance, Drosopila Pontin and Reptin have been shown to associate with the Brm-C TrxG complex and PRC1 PcG complex respectively to function antagonistically in controlling Hox gene expression [44]. As expected, we have identified Reptin as the most abundant protein in Pontin complexes (Figure 6), but whether Reptin can respond to RA signaling and interact with RAR to control HOX gene transcription remains to be investigated.

Our ChIP-seq data showed that the common binding sites of Pontin and RARα on HOX clusters were pre-occupied by RARα in the absence of RA signaling. Early studies have demonstrated that RAR/RXR and nuclear receptor co-repressor (NCoR) co-occupy the RARE of Hoxa1 gene in murine ES cells in the absence of RA ligand, resulting in Hoxa1 gene silencing [56]. In response to RA signaling, RARs undergo conformational changes that allow for recruitment of co-activators and transcriptional activation [27, 28]. In contrast, on some other sites, the recruitment of RARs and RXRs is dependent on the presence of RA [29]. In this study, we observed that RARα pre-binds to HOX gene clusters before RA stimulation, with a HOX gene expression level measured by H3K36me3 (Figure 3). In the presence of RA, Pontin was recruited to RARα-bound sites, activating HOX gene transcription. This process might be associated with a dynamic change in the components of RAR/Pontin protein complex. Future studies will be needed to map the protein components that are released or recruited to the complex on the RAR-bound RARE sites.

Materials and Methods

Cell culture

NTERA-2 cell line was obtained from ATCC. NTERA-2 cells were cultured in Minimum Essential Medium α (MEMα) (Gibco) supplemented with 10% fetal bovine serum and penicillin/streptomycin. Cells were grown at 37°C in a humidified incubator with 5% CO₂. For differentiation induction, NTERA-2 cells were treated with 5 μM all-trans retinoic acid (RA) (Sigma) or dimethyl sulfoxide (DMSO) (as vehicle control) for 24 hours before collection for experiments.

Generation of BLRP-tagged stable cell lines for Pontin

All commercially available Pontin antibodies tested were not effective enough for proteomic complex or genome-wide binding studies in NTERA-2 cell line. To solve this issue, we used an in vivo biotinylation tagging BirA-BLRP system as previously described [45–47]. The BLRP-tagged Pontin cDNA was cloned into a retrovirus-based Tet-On expression vector pRetroX-Tight-Pur (Clontech #632104). The finished construct was then co-transfected with pCL-Ampho packaging plasmid into 293T cells for retrovirus production. Then the retroviruses were transduced into a parental NTERA-2 stable cell line that has been engineered to stably express the BirA enzyme and Tet Repressor. G418 (250 μg/ml), hygromycin (100 μg/ml), and puromycin (0.3 μg/ml) were used for selection and stable cell line maintenance. To induce BLRP-tagged protein expression, stable cell lines were treated with 2 μg/ml doxycycline for approximately 24 hours and were collected to check for BLRP-tagged protein expression levels by Western blots with specific antibodies before biotin ChIP or proteomic complex pull-down experiments. When picking cell colonies for stable lines, we made sure the BLRP-tagged Pontin expression levels were not too high. Ideally, the exogenous expression level should be similar to the endogenous level.

Western blot analysis

Cells were lysed in RIPA lysis buffer (50 mM Tris-Cl pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with 1 mM DTT, 1 mM PMSF, and 1x protease inhibitor cocktail (Roche). Protein concentrations were quantified with the Bio-Rad protein assay kit. Western blotting was performed as previously described [46]. Briefly, 30 μg of protein extracts were loaded and separated by SDS–PAGE gels. Blotting was performed with standard protocols using PVDF membrane (Bio-Rad). Membranes were blocked for 1 hour in blocking buffer (5% Non-fat milk in PBST) and probed with primary antibodies at 4°C overnight. After three washes with PBST, the membranes were incubated with HRP-conjugated secondary antibody. Signals were visualized with Clarity Western ECL Substrate (Bio-Rad) as described by the manufacturer. The antibodies used in this assay were: anti-Pontin (Ab51500, abcam), streptavidin-HRP (016–030-084, Jackson ImmunoResearch), anti-Flag (F1804, Sigma) and anti-GAPDH (sc-25778, Santa Cruz).

Coimmunoprecipitation (co-IP)

To study the interaction between RARα and Pontin under RA stimulation condition, we used NTERA-2 cells treated with 5uM retinoic acid for 24 hours. Cytoplasmic/nuclear fractionation was performed as previously described [45–47]. Nuclear pellets were collected and lysed with NP-40 lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% NP-40, 0.1% Triton X-100. 1 mM PMSF and 1x protease inhibitor were freshly added.). Sonication was performed for 3 minutes with Q800R sonicator (QSonica) at 4°C to improve lysis before centrifuging to collect nuclear lysate. The nuclear lysate was precleared using 10 μl Dynabeads Protein G (Life Technologies) and 2 μg anti-RARα rabbit polyclonal antibody (sc-551, Santa Cruz) was added to lysate and incubated at 4°C overnight. The protein complex was collected using 20 μl Dynabeads Protein G and washed 6 times with NP-40 lysis buffer before elution with 2x SDS loading buffer. The antibodies used for Western Blot detection were: anti-RARα mouse monoclonal antibody (sc-515796, Santa Cruz) and anti-Pontin mouse monoclonal antibody (Ab51500, abcam).

Chromatin immunoprecipitation (ChIP) assays

Regular ChIP assays for RARα were performed as previously described [46], with some modifications. Briefly, NTERA-2 cells were cross-linked with 1% formaldehyde for 10 minutes at room temperature. Cross-linking was quenched with 0.125 M glycine for 5 minutes. Cells were successively lysed in lysis buffer LB1 (50 mM HEPES-KOH, pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.25% Triton X-100, 1x PI), LB2 (10 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 1x PI), and LB3 (10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% Na-Deoxycholate, 0.5% N-lauroylsarcosine, 1x PI). Chromatin was sonicated to an average size of ~200–500 bp using QSonica’s Q800R sonicator system (20% amplitude, 10s on and 20s off for 10 minutes). A total of 5 μg of RARα antibody (sc-551, Santa Cruz) was added to the sonicated chromatin and incubated overnight at 4°C. Subsequently, 50 μl of Dynabeads Protein G (Invitrogen) were added to each ChIP reaction and incubated for 4 hours at 4°C. Dynabeads were washed with RIPA buffer (50 mM HEPES pH 7.6, 1 mM EDTA, 0.7% Na-Deoxycholate, 1% NP-40, 0.5 M LiCl) 6 times, and once with TE. The chromatin was eluted, followed by reverse cross-linking and DNA purification. ChIP DNA was resuspended in 10 mM Tris-HCl pH 8.5. The purified DNA was subjected to qPCR to confirm target region enrichment before moving on to deep sequencing library preparation. Each experiment was performed for at least twice to make sure similar results are reproducible.

Biotin ChIP-seq experiments for BLRP-tagged Pontin were performed with our previous protocol [46]. Because Pontin is a transcriptional cofactor without DNA binding domain, we modified the fixation step for better pull-down efficiency. BLRP-tagged Pontin stable cells were double cross-linked with 2 mM DSG (CovaChem) for 45 minutes followed by secondary fixation with 1% formaldehyde for 10 minutes. Briefly, cross-linked protein-DNA complexes were pulled down by MyOne Streptavidin T1 Dynabeads (Thermo-Fisher Scientific) and the washing was performed under much more stringent conditions that included four washes with 1% SDS in TE (20 minutes each) and two washes with 1% Triton X-100 in TE. The washed streptavidin beads were then subjected to TEV protease (Life Technologies) digestion twice for tagged protein-DNA complex elution before de-crosslinking at 65°C overnight. The following day, the final ChIP DNA was purified and resuspended in 10 mM Tris-HCl pH 8.5. The purified DNA was subjected to qPCR directly to confirm target region enrichment. BLRP-tagged Pontin stable cell lines without doxycycline induction were used as controls for background detection by ChIP-qPCR tests before moving on to deep sequencing library preparation. Each experiment was performed for at least twice to make sure similar results are reproducible.

For both RARα ChIP and Pontin biotin ChIP, the extracted DNA samples were used to construct the ChIP-seq library using KAPA Hyper Prep kit (KK8504), followed by deep sequencing with the Illumina sequencing systems (Illumina Genome Analyzer IIX for Pontin biotin ChIP-seq and HiSeq 2000 for RARα ChIP-seq) according to the manufacturer’s instructions.

Generation of doxycycline-inducible knockdown stable cell lines for Pontin in NTERA-2

To perform inducible knockdown experiments for Pontin in NTERA-2, we established two Tet-On inducible stable cell lines for Pontin shRNAs (Tet-on shPontin#1 and Tet-on shPontin#2). We have chosen two validated shRNA sequences for Pontin gene from Sigma Mission shRNA library (shRNA #1: TRCN0000018914; shRNA #2: TRCN0000018912). Both shRNAs have very good knockdown efficiency and these two shRNA sequences were cloned into Tet-pLKO-puro (addgene #21915) following the cloning protocol associated with this plasmid from addgene [57]. The lentiviral Tet-on shPontin vectors were co-transfected with packaging plasmids (psPAX2 and pMD2.G from addgene) into 293T cells. Culture medium containing lentivirus particles were harvested, filtered, and used to infect cells in media containing polybrene (8 μg/ml). Cells were selected by 0.3 μg/ml Puromycin (Invitrogen) after infection to set up doxycycline-inducible stable cell lines. To induce Pontin expression knockdown, 2μg/ml doxycycline was added into culture media for about 5 days before the RT-qPCR or GRO-seq sample collection (the RA/DMSO treatment was performed on Day 4).

RNA isolation and quantitative RT-PCR

Total RNA was isolated with RNeasy Mini Kit (Qiagen) according to the manufacture’s protocol and 1 μg RNA was used to convert to cDNA using iScript Select cDNA Synthesis Kit (Bio-Rad) in the presence of both oligo (dT) and random primers. qPCR was conducted with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) using CFX384 Real-Time PCR Detection System (Bio-Rad) according to the manufacturer’s instructions. Relative expression of RNA was determined by the ΔΔCT method using GAPDH as an internal control for quantification analyses of gene targets. The primers used for qPCR are as follows: GAPDH: F: AACATCATCCCTGCCTCTACTGG, R: GTTTTTCTAGACGGCAGGTCAGG; PONTIN: F: GGTGCTGATGGAGAACTTCC, R: GGTTCCTTTGGCTGTTTTGA; HOXA1: F: CTTCTCCAGCGCAGACTTTT, R: CGCGTCAGGTACTTGTTGAA; HOXA2: F: ACCCCTGGATGAAGGAGAAG, R: GCTGTGTGTTGGTGTAAGCAG; HOXB1: F: CTCCGAGGACAAGGAAACAC, R: AGCTGCCTTGTGGTGAAGTT; HOXB2: F: AAGAAATCCGCCAAGAAACC, R: CAGCTGCGTGTTGGTGTAAG; HOXC4: F: CCAGCAAGCAACCCATAGTC, R: GGGTCAGGTAGCGGTTGTAA; HOXD1: F: ACCTACCCCAAGTCCGTCTC, R: GTCAGTTGCTTGGTGCTGAA.

RT-qPCR experiments were performed with at least three independent biological replicates and three technical replicates for each reaction. Results are reported as mean ± SEM for three independent experiments. Data were analyzed and statistics were performed using unpaired two-tailed Student’s t-test or one-way ANOVA (Graphpad Prism 8). Significant differences between two groups were noted by * P<0.05, ** P<0.01, *** P<0.001.

Global run-on sequencing (GRO-seq)

To investigate the effect of Pontin knockdown on HOX gene expression, we performed GRO-seq in NTERA-2 cells. Two Tet-On inducible stable cell lines for Pontin shRNAs (Tet-on shPontin#1 and Tet-on shPontin#2) were treated with 2μg/ml doxycycline for 5 days before collection for GRO-seq (the RA/DMSO treatment was performed on Day 4). GRO-seq experiments were performed as previously reported [46]. Briefly, ~5–10 million NTERA-2 cells treated with vehicle or RA for 24 hours were washed 3 times with cold PBS and then sequentially swelled in swelling buffer (10 mM Tris-HCl pH7.5, 2 mM MgCl₂, 3 mM CaCl₂) for 10 minutes on ice, harvested, and lysed in lysis buffer (swelling buffer plus 0.5% NP-40, 20 units of SUPERase-In, and 10% glycerol). The resultant nuclei were washed two more times with 10 ml lysis buffer, and finally re-suspended in 100 μl of freezing buffer (50 mM Tris-HCl pH8.3, 40% glycerol, 5 mM MgCl₂, 0.1 mM EDTA). For the run-on assay, re-suspended nuclei were mixed with an equal volume of reaction buffer (10 mM Tris-HCl pH 8.0, 5 mM MgCl₂, 1 mM DTT, 300 mM KCl, 20 units of SUPERase-In, 1% sarkosyl, 500 μM ATP, GTP, and Br-UTP, 2 μM CTP) and incubated for 5 minutes at 30°C. The resultant nuclear-run-on RNA (NRO-RNA) was then extracted with TRIzol LS reagent (Life Technologies) following manufacturer’s instructions. NRO-RNA was fragmented to ~300–500 nt by alkaline base hydrolysis on ice for 30 minutes followed by treatment with DNase I and Antarctic Phosphatase. At this step, only a small portion of total RNA species are Br-UTP-labeled. To purify the Br-UTP labeled nascent RNA, the fragmented NRO-RNA was immunoprecipitated with anti-BrdU agarose beads (Santa Cruz Biotechnology) in binding buffer (0.5xSSPE, 1 mM EDTA, 0.05% tween) for 1–3 hours at 4°C with rotation. Subsequently, T4 PNK was used to repair the ends of the immunoprecipitated Br-UTP labeled nascent RNA at 37°C for 1hour. The RNA was extracted and precipitated using acidic phenol-chloroform.

The RNA fragments were subjected to poly-A tailing reaction by poly-A polymerase (NEB) for 30 minutes at 37°C. Subsequently, reverse transcription was performed using oNTI223 primer and SuperScript III RT kit (Life Technologies). The cDNA products were separated on a 10% polyacrylamide TBE-urea gel and only those fragments migrating between 100–500bp were excised and recovered by gel extraction. Next, the first-strand cDNA was circularized by CircLigase (Epicentre) and re-linearized by APE1 (NEB). Re-linearized single strand cDNA (sscDNA) was separated on a 10% polyacrylamide TBE gel and the appropriately sized product (~120–320 bp) was excised and gel-extracted. Finally, sscDNA template was amplified by PCR using the Phusion High-Fidelity enzyme (NEB) according to the manufacturer’s instructions. The oligonucleotide primers oNTI200 and oNTI201 were used to generate DNA libraries for deep sequencing. The sequences for primers oNTI223, oNTI200 and oNTI201 are as follows: oNTI223-TruSeqLT:/5Phos/AGATCGGAAGAGCGTCGTGTA/idSp/CAGACGTGTGCTCTTCCGATC TTTT TTT TTT TTT TTT TTT TVN oNTI201-TruseqLT: 5’AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT 3’ oNTI200-TruSeqID1: 5’CAAGCAGAAGACGGCATACGAGAT CGTGAT GTGACTGGAGTT CAGACGTGTG CTCTTCCGATC 3’.

The purification of Pontin protein complexes using BLRP-tagged stable line

BLRP-tagged stable cells for Pontin were washed for three times with cold PBS and nuclei were prepared with a similar method as GRO-seq protocol (see above). The nuclear pellet was lysed with NP40 lysis buffer (50mM Tris-HCl pH 7.4, 150mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% TritonX-100, protease inhibitor cocktail (PIC) freshly added). To improve lysis, sonication was performed several times (10 seconds each) on ice before centrifuging to obtain the supernatant fraction. The nuclear lysate was precleared using 20μl Dynabeads Protein G (Life Technologies) per 1ml before IP. NanoLink™ Streptavidin Magnetic Beads (Solulink) (blocked by 1%BSA/PBS for 1hr before use) were added into the diluted nuclear lysate and incubated overnight at 4°C. The next day, the bead-protein complexes were washed four times with NP40 lysis buffer. The tagged protein complex could be eluted with TEV protease (Life Technologies) digestion following manufacturer’s instructions for cleaner results and then used for Western blot analysis or Mass-Spectrometry analysis.

Mass-Spectrometry was performed following a published method (Guttman et al., 2009). Briefly, protein samples were diluted in TNE (50mM Tris-HCl pH 8.0, 100mM NaCl, 1mM EDTA) buffer. RapiGest SF reagent (Waters Corp.) was added to the mix to a final concentration of 0.1% and samples were boiled for 5 minutes. TCEP (Tris (2-carboxyethyl) phosphine) was added to 1mM (final concentration) and the samples were incubated at 37°C for 30 minutes. Subsequently, the samples were carboxymethylated with 0.5 mg/ml of iodoacetamide for 30 minutes at 37°C followed by neutralization with 2mM TCEP (final concentration). Protein samples were then digested with trypsin (trypsin:protein ratio-1:50) overnight at 37°C. RapiGest was degraded and removed by treating the samples with 250mM HCl at 37°C for 1 hour followed by centrifugation at 14,000 rpm for 30 minutes at 4°C. The soluble fraction was then added to a new tube and the peptides were extracted and desalted using C18 desalting columns (Thermo Scientific) for LC-MS/MS analysis. For LC-MS/MS, Trypsin-digested peptides were analyzed by ultra-high pressure liquid chromatography (UPLC) coupled with tandem mass spectroscopy (LC-MS/MS) using nano-spray ionization, which was performed by a Triple TOF 5600 hybrid mass spectrometer (AB SCIEX) interfaced with nano-scale reversed-phase UPLC (Waters corporation, nano ACQUITY) using a 20cm-75 micron ID glass capillary packed with 2.5-μm C18 (130) CSHTM beads (Waters corporation). Peptides were eluted from the C18 column into the mass spectrometer using a linear gradient (5–80%) of ACN (Acetonitrile) at a flow rate of 250μl/minute for 1 hour. The buffers used to create the ACN gradient were: Buffer A (98% H₂O, 2% ACN, 0.1% formic acid, and 0.005% TFA) and Buffer B (100% ACN, 0.1% formic acid, and 0.005% TFA). MS/MS data were acquired in a data-dependent manner, in which the MS1 data was acquired for 250 ms at m/z of 400 to 1,250 Da and the MS/MS data was acquired from m/z of 50 to 2,000 Da. The independent data acquisition (IDA) parameters were as follows: MS1-TOF acquisition time of 250 milliseconds, followed by 50 MS2 events of 48 milliseconds acquisition time for each event. The threshold to trigger MS2 event was set to 150 counts when the ion had the charge state +2, +3 and +4. The ion exclusion time was set to 4 seconds. Finally, the collected data were analyzed using Protein Pilot 4.5 (AB SCIEX) for peptide identification.

ChIP-seq data analysis

Reads were aligned to human genome (hg19) using bowtie with “--best --strata –m 1” parameters [58]. Only uniquely mapped and non-duplicated reads were selected for subsequent analysis. MACS2 was employed to call peaks with default parameters and a q-value cutoff of 1e-5 [59]. For Pontin ChIP-seq data, the broad mode of MACS2 was switched on. The peaks overlapping with the blacklist regions from UCSC were removed. For all ChIP-seq analyses, the peaks within ± 1000bp of RefSeq gene TSSs were considered as promoter-bound peaks. Annotated positions for promoters, exons, introns and other features were based on RefSeq transcripts and repeat annotations from the University of California, Santa Cruz. Heatmap matrices were created by counting tags using a 4kb window (±2kb of the peak center) and 100bp bin size.

GRO-seq data analysis

GRO-seq reads were aligned to human genome (hg19) using bowtie with “--best --strata –m 1 –v 2” parameters [60]. Duplicated reads were eliminated for subsequent analysis. To balance the clonal amplification bias and total useful reads, only three reads at most were allowed for each unique genomic position. When measuring the expression level of genes, mapped reads from the first 30 kb of gene body were counted, excluding promoter-proximal region (transcription start site (TSS) to 1000 bp downstream of TSS; if the length of a gene is shorter than 10 kb, then the reads mapped to the first 10%*length regions were excluded). If the length of a gene is shorter than 30 kb, then the mapped reads from the whole gene were counted, excluding promoter-proximal region and gene end (500 bp upstream of transcription termination site (TTS) to TTS).

Other data analysis

For Pontin proteomic complex studies, ClueGO was used to calculate the significance of the enriched biological processes for Pontin interactome [61]. Only those biological processes with adjusted p values less than 0.05 were shown and Kappa Score was set to 0.4 to group the connected biological processes.

Data visualization

All the genome browser viewers were visualized in Integrative Genomics Viewer (IGV). ChIP-seq and GRO-seq samples were normalized to 10 million mapped reads per experiment.

Data availability

The ChIP-seq data of H3K36me3 were retrieved from GSE42602 [62]. All the other ChIP-seq data and GRO-seq data have been deposited in the Gene Expression Omnibus (GEO) under accession code GSE162372.

Supplementary Material

1

Supplemental Figure 1. In vivo biotinylation of Pontin for ChIP-seq or proteomic complex studies in NTERA-2 cells.

A. BirA-BLRP method diagram for the setup of in vivo Pontin biotinylation. BirA biotinylates the BLRP-tagged Pontin protein, which is pulled-down along with its interacting DNA or proteins by magnetic streptavidin beads. Pontin protein is then isolated using TEV protease.

B. Western blot analysis demonstrating expression of biotinylated BLRP-tagged Pontin at the same level as endogenous Pontin.

C. Western blot analysis showing the TEV-mediated elution of biotin-tagged Pontin protein from magnetic streptavidin beads.

2

Supplemental Figure 2. Genome browser views of ChIP-seq signals and GRO-seq signals derived from another Pontin knockdown experiment by a different Tet-on shRNA for genomic regions covering HOXA1 and HOXA2 (A), HOXB1 and HOXB2 (B), HOXC4 (C), HOXD1 (D) gene loci.

E. Quantification of GRO-seq signals for HOXA1, HOXA2, HOXB1, HOXB1, HOXC4 and HOXD1 genes, showing that RA-dependent transcriptional activation of these genes was greatly abolished upon doxycycline-induced Pontin knockdown. To measure the nascent transcriptional level of each HOX gene, we counted the reads mapped to the whole gene body of each gene from transcription start site to 1kb downstream of the transcription stop site, and further normalized them to 1 million reads of the mapped library size. For each condition, the expression levels of the HOX genes were log2 transformed for comparison.

3

Supplementary Table 1. Lists of Pontin binding peaks (vehicle-specific, common, and RA-specific peaks shown in Figure 1C).

4

Supplementary Table 2. List of the overlapping 3,661 peaks between Pontin ChIP-seq and RARα ChIP-seq data (+RA condition) shown in Figure 2A.

5

Supplementary Table 3. List of Pontin-interacting proteins from BirA-BLRP IP mass spectrometry data.

Research Highlights.

1. Pontin is an essential gene but its function is not fully understood.

2. ChIP-seq identified genome-wide binding sites of Pontin in response to RA signaling.

3. Pontin was recruited to RARα-bound sites upon RA stimulation.

4. GRO-seq demonstrated the requirement of Pontin in RA-activated HOX gene transcription.

5. For the first time we revealed the role of Pontin as a cofactor of RARα.