Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

MIR sequences recruit zinc finger protein ZNF768 to expressed genes

MIR sequences recruit zinc finger protein ZNF768 to expressed genes Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 700–715 Nucleic Acids Research, 2019, Vol. 47, No. 2 Published online 22 November 2018 doi: 10.1093/nar/gky1148 MIR sequences recruit zinc finger protein ZNF768 to expressed genes 1 2 3 1 Michaela Rohrmoser , Michael Kluge , Yousra Yahia , Anita Gruber-Eber , Muhammad 3 4 5 5 6 Ahmad Maqbool , Ignasi Forne´ , Stefan Krebs ,HelmutBlum , Ann Katrin Greifenberg , 6 7,8 4 3 Matthias Geyer , Nicolas Descostes ,AxelImhof , Jean-Christophe Andrau , Caroline 2 1,* C. Friedel and Dirk Eick Department of Molecular Epigenetics, Helmholtz Center Munich and Center for Integrated Protein Science Munich (CIPSM), Marchioninistrasse 25, 81377 Munich, Germany, Institute for Informatics, Ludwig-Maximilians-Universitat Munchen, ¨ Amalienstrasse 17, 80333 Munich, Germany, Institut de Gen ´ etique ´ Moleculaire ´ de Montpellier (IGMM), Univ Montpellier, CNRS-UMR5535, Montpellier, France, Biomedical Center Munich, ZFP, Großhadener Strasse 9, 82152 Planegg-Martinsried, Germany, Laboratory for Functional Genome Analysis (LAFUGA) at the Gene Center, Ludwig-Maximilians-Universitat ¨ Munchen, ¨ Feodor-Lynen-Strasse 25, 81377 Munich, Germany, Institute of Structural Biology, University of Bonn, Sigmund-Freud-Str. 25, 53127 Bonn, Germany, Department of Biochemistry and Molecular Pharmacology, New York University Langone School of Medicine, New York, NY 10016, USA and Howard Hughes Medical Institute, New York University Langone School of Medicine, New York, NY 10016, USA Received July 06, 2018; Revised October 25, 2018; Editorial Decision October 29, 2018; Accepted October 29, 2018 ABSTRACT terspersed sequences (SINE) (1,2). Mammalian-wide in- terspersed repeats (MIRs) are an ancient family of retro- Mammalian-wide interspersed repeats (MIRs) are transposed SINEs that spread genome-wide before and dur- retrotransposed elements of mammalian genomes. ing mammalian radiation (3,4). MIRs are ∼240 bp long Here, we report the specific binding of zinc finger pro- and consist of tRNA-derived sequences, a 70 bp MIR- tein ZNF768 to the sequence motif GCTGTGTG (N ) 20 specific core region, and sequences similar to the 3 ends CCTCTCTG in the core region of MIRs. ZNF768 bind- of LINEs. MIRs are enriched at gene loci in euchromatin, harbor putative transcription-factor binding sites, provide ing is preferentially associated with euchromatin and insulator and enhancer function (5–8), encode microRNAs, promoter regions of genes. Binding was observed are transcribed by RNA polymerase III (9,10), are associ- for genes expressed in a cell type-specific manner ated with tissue-specific gene expression ( 5,11), and some- in human B cell line Raji and osteosarcoma U2OS times provide splicing signals and contribute to exonization cells. Mass spectrometric analysis revealed binding (12). MIRs constitute 5–16% of the genome in marsupials of ZNF768 to Elongator components Elp1, Elp2 and and monotremes and 0.5-3% in placentalia (13). Like other Elp3 and other nuclear factors. The N-terminus of transposable elements, MIRs have shaped gene regulatory ZNF768 contains a heptad repeat array structurally networks in vertebrates (14–17), but our understanding how related to the C-terminal domain (CTD) of RNA poly- MIRs regulate gene activity is still elusive. merase II. This array evolved in placental animals Similarly to MIRs, the family of zinc finger pro- but not marsupials and monotreme species, displays teins (ZNFs) strongly expanded in mammals (18,19). species-specific length variations, and possibly ful- Widespread binding of ZNFs to regulatory regions indi- cates that mammalian genomes contain an extensive ZNF fills CTD related functions in gene regulation. We pro- regulatory network that targets a diverse range of genes and pose that the evolution of MIRs and ZNF768 has ex- pathways (20,21). Zinc finger protein 768 (ZNF768) evolved tended the repertoire of gene regulatory mechanisms in mammals and is defined by a domain of ten zinc fingers in mammals and that ZNF768 binding is associated with >96% (Figure 1) identity in placentals and marsupials, with cell type-specific gene expression. but is less conserved in monotremes (Supplementary Fig- ure S1). Placentalia additionally evolved an array of 10–20 INTRODUCTION heptad repeats in the amino-terminus of ZNF768, which is absent in marsupials and monotremes. This array has a Approximately half of mammalian genomes is of repeti- striking similarity to the carboxy-terminal domain (CTD) tive nature and composed of long (LINE) and short in- To whom correspondence should be addressed. Tel: +49 89 3187 1512; Email: [email protected] C The Author(s) 2018. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected] Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 701 Figure 1. Domain structure of ZNF768 in placentalia and marsupials and comparison with the CTD of RNA polymerase II. (A) Human ZNF768 is composed of domains box A (red box) and box B (green box) at the N-terminus interrupted by an array of 15 heptad repeats (yellow box) and a domain of 10 zinc fingers at the C-terminus (blue box). ( B) Mouse ZNF768 evolved an array of 19 heptad repeats. (C) ZNF768 of the marsupial Tammar Wallaby contains conserved A, B, and zinc finger domains, while the array of heptad repeats is absent. ( D) Number of heptad repeats in RNA polymerase II in vertebrates and ZNF768 in placentalia (see also Supplementary Figure S1). of the large subunit (Rpb1) of RNA polymerase II (Pol II), 3 processing of the nascent transcript. Interestingly, CTD which is composed of 52 heptad repeats with the consensus can function as transcriptional activator after fusion to a sequence Y S P T S P S . GAL4 DNA binding domain (27). Furthermore, transition 1 2 3 4 5 6 7 The CTD functions as a platform for recruitment and of Pol II through the transcription cycle is also observed if dissociation of cellular factors to the transcription machin- CTD is fused to other subunits of Pol II (28). Recent re- ery and is mainly regulated during the transcription cycle ports further provide evidence that CTD of Pol II can ag- by phosphorylation of heptad repeats by various kinases gregate reversibly alone, or with low complexity domains of (22–26). It is required for initiation, elongation, and termi- other transcription factors, like FUS, and that the ability for nation of transcription, but also for capping, splicing, and Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 702 Nucleic Acids Research, 2019, Vol. 47, No. 2 phase separation in liquid droplets is an important feature Immunofluorescence microscopy for the regulation of transcriptional activity (29–32). U2OS cells were seeded on a coverslip and grown for Due to the striking similarity of the heptad repeat array in 24 h. Cells were washed with PBS and fixed with 2% ZNF768 with the array of heptad repeats in CTD of Pol II paraformaldehyde (PFA) at RT for 5 min. After perme- we investigated if ZNF768 can act as a transcription factor abilization with 0.15% TritonX-100, samples were blocked and fulfill gene regulatory functions in cells. with 1% BSA and incubated with 7D6 or HA-specific mAbs over night at 4 C. Samples were washed with PBS for 5 min MATERIALS AND METHODS at RT, 0.15% Triton X-100 for 10 min at RT, blocked with 1% BSA for 7 min and incubated with Cy5-conjugated don- Tissue culture and recombinant gene expression key anti-rat immunoglobulin (Dianova) in the dark for 45 U2OS osteosarcoma cells were cultured in Dulbecco’s mod- min. Cells were washed again, stained with 4 ,6-diamidino- ified Eagle’s medium (DMEM, Gibco) and Raji B-cells in 2-phenylindole (DAPI) (Sigma) and mounted on slides us- RPMI 1640 medium (Gibco) supplemented with 10% fe- ing fluorescent mounting medium (Dako). Confocal mi- tal calf serum (FCS, Bio&Sell), 2 mM L-glutamine (Gibco), croscopy was performed on a Leica LSCM SP2 fluorescence 100 U/ml penicillin (Gibco), and 100 g/ml streptomycin microscope using the objective HCX PL APO 63× 1.4. Im- (Gibco) at 37 Cat8%or5%CO , respectively. Stably trans- 2 ages were processed using ImageJ 1.37 V and Fuji software fected U2OS cell lines were generated with the expression and the plug-in RGB profiler. Scale bars were calculated as vector pRTS-1 (33) using Polyfect (QIAGEN) followed by follows: hygromycin B (200 g/ml) selection. Conditional gene ex- B × 5 m/P (B = picture length in m, P = (512 pixel × pression was induced with 1 g/ml doxycycline. Recombi- voxel size) in m) nant ZNF768 and mutants are tagged C-terminally by a hemagglutinin (HA) tag and synthesized with an optimized siRNA transfection codon usage (Gene Art, Regensburg). Details for cloning in pRTS has been described elsewhere (34). All plasmids were siRNA transfection was performed according to the manu- confirmed by DNA sequencing prior to expression. facturer’s protocol using HS ZNF768 1 FlexiTube siRNA (Qiagen) and the HiPerFect Transfection Reagent (Qiagen) Monoclonal antibody with the exception that transfection was repeated after 24 h. Negative (non-silencing) siRNA (Qiagen) was used as con- The generation of monoclonal antibodies has been de- trol. scribed previously (34). The ZNF768-specific peptide RSPESDSQSPEFESQSPRYEPQSPGYEPRSPG (synthe- sized by PSL GmbH, Heidelberg) was coupled to ovalbu- Cell proliferation assay min for immunization. The rat monoclonal antibody 7D6 Cell proliferation was determined using the Real-time used in this study (IgG2c) specifically recognizes human xCELLigence System (Roche). U2OS cells were seeded at a ZNF768. density of 3.000 cells per 100 l in equilibrated 96-well mi- crotiter xCELLigence assay plates (E-plates). Conditional Immunoprecipitation (IP) and SDS-PAGE gene expression was induced with 1 g/ml doxycycline at the indicated time points. Alternatively, siRNA transfection Cells were washed twice with cold phosphate-buffered was performed according to the manufacturer’s protocol. saline (PBS) and lysis was performed in 100 llysis buffer per 2 × 10 cells (50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% NP-40 (Roche), 1x PhosSTOP (Roche), 1× protease in- Purification of ZNF768 hibitor cocktail (Roche)) at 4 C for 30 min, followed by son- ication on ice using a BRANSON Sonifier 250 (15 s on, 15 s Expression plasmids of human ZNF768 (UniProt acces- off, 50% duty) and centrifugation at 16 400 rpm (FA-45-24- sion number Q9H5H4) were cloned from a synthetic gene 11 rotor) for 10 min at 4 C. Immunoprecipitation was per- that was codon optimized for expression in Escherichia coli formed using Dynabeads Protein A und G (1:1) (Invitro- cells (Gene Art, Regensburg). A full length ZNF768 (1– gen). Lysates were incubated with antibody-coupled beads 540) construct and a construct consisting of the N-terminal (2.5 g of antibodies for 4 h at 4 C, followed by three washes heptad-repeats only (1–197) were cloned by PCR with re- with 1 ml lysis buffer) overnight. Beads were washed three strictions sites NcoI/EcoRI and ligated into a pGEX-4T1 times with 1 ml lysis buffer and boiled in laemmli buffer vector modified with a TEV protease cleavage site. All plas- (2% SDS, 10% glycerol, 60 mM Tris–HCl, pH 6.8, 10 mM mids were confirmed by DNA sequencing prior to expres- EDTA, 1 mM PMSF, 100 mM DTT, 0.01% bromophenol sion. blue) for SDS-PAGE. Whole cell lysates or IP samples were Plasmids were transformed into E. coli BL21(DE3) cells resolved by SDS-PAGE (10% or 15%) and transferred onto and induced at an OD of 0.6 to 1.0 with 0.3 mM IPTG nitrocellulose transfer membranes (GE Healthcare). The for 4 h growth. Cells were harvested in lysis buffer (20 membrane was blocked with 5% milk/TBS-T for 1 h. Incu- mM Tris–HCl pH 8.0, 500 mM NaCl, 10% glycerol, 1 mM bation with primary antibodies was performed over night DTE) and lysed by ultrasound. Fusion proteins were iso- at 4 C, followed by incubation with HRP-conjugated sec- lated with GSH Sepharose FastFlow (GE Healthcare) affin- ondary antibodies for 1 h and chemiluminescence detection ity chromatography methods. Cleavage of the GST-tag was with ECL (GE Healthcare). achieved by adding TEV protease in a 1/50 ratio and was Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 703 performed for 20 h at 4 C. Protein solution was concen- lauroylsarcosine, 1× protease inhibitor cocktail). Sonica- trated and loaded on a preparative HiLoad 16/60 Superdex tion was carried out in a Bioruptor Pico ultrasounds wa- 200 prep grade gel filtration column (GE Healthcare) for ter bath (Diagenode B01060001) for 30 cycles of 30 s ON full length ZNF768 or on a HiLoad 16/60 Superdex 75 and 30 s OFF pulses in 4 C water. Sonicated extracts were column for the truncated version of ZNF768 (1–197), re- centrifuged at high speed in the presence of 0.1% of Triton spectively, and equilibrated in gel filtration buffer (50 mM X-100 and snap frozen in liquid nitrogen and then stored at HEPES pH 7.5, 150 mM NaCl and 1 mM TCEP). Frac- −80 C until subsequent use. tions of the peak containing ZNF768 proteins determined Prior to ChIP, the ZNF768 mAb was coupled to protein- by SDS PAGE analysis were pooled and concentrated to 5 G coated magnetic beads (Dynabeads, life technologies) by mg/ml. The protein was aliquoted, snap frozen in liquid ni- incubation in 0.5% BSA PBS overnight at 4 C. Pre-coated trogen and stored at −80 C. beads were then washed and incubated with the sonicated chromatin extracts. ChIP was carried out overnight at 4 C on a rotating wheel. The equivalent of 10 × 10 cells soni- Gel shift assay cated extract was used for each ChIP experiment for both Gel shift assay was performed according to the manu- cell lines. After incubation, the beads were washed 7× with wash buffer (50 mM HEPES pH 7.6, 500 mM LiCl, 1 mM facturer’s protocol using the DIG Gel Shift Kit, second generation (Sigma-Aldrich). Briefly, oligonucleotides EDTA pH 8.0, 1% NP-40, 0.7% Na-Deoxycholate, 1× pro- were annealed to equimolar amounts of their com- tease inhibitor cocktail) followed by one wash with TE- plementary strands (M1: 5 - CAGTGCTGTGTGAC NaCl buffer (10 mM Tris pH 8.0, 1 mM EDTA pH 8.0, CTTGGGCAAGTCACTTAACCTCTCTGCAGT-3 , 50 mM NaCl). Immunoprecipitated chromatin was eluted M2: 5 - CAGTGCTGTGTGCAGTCAGTCAGT by two sequential incubations with 100 l elution buffer (50 CAGTCAGTCCTCTCTGCAGT-3 and M3: 5 - mM Tris pH 8.0, 10 mM EDTA pH 8.0, 1% SDS) at 65 C CAGTCAGTTGTGACCTTGGGCAAGTCACTT for 15 min. The two eluates were pooled and incubated at ◦ ◦ AACCTCCAGTCAGT-3 ) by heating to 95 C for 5 min 65 C for 12 h to reverse-crosslink of chromatin, followed and cooling slowly to room temperature. Double-stranded by treatment with RNase A (0.2 g/ml) at 37 C for 2 h and oligonucleotide probes were labelled at the 3 end using proteinase K (0.2 g/ml) at 55 C for 2 h. The DNA was DIG-11-dUTP and terminal transferase. Binding reactions isolated by phenol:chloroform:isoamylalcohol (25:24:1 pH were performed in 20 l volumes containing binding buffer 8.0) extraction followed by Qiaquick PCR Purification (Qi- [20 mM HEPES, pH 7.6, 1 mM EDTA, 10 mM (NH4) SO , agen, Germany) and quantified with Qubit DS DNA HS 2 4 5 mM DTT, 0.2% (w/v) Tween 20, 30 mM KCl], 50 ng/l Assay (ThermoFisher Scientific, USA). Poly [d(I-C)] and 5 ng/l Poly L-lysine at room temperature At least 1 ng of ChIP DNA was used to prepare sequenc- for 15 min. 0.6 ng of DIG-labelled DNA and extract of ing library with Illumina ChIP Sample Library Prep Kit 1.5–15 g purified ZNF768-WT or ZNF768 1–197 was (Illumina, USA) with a few optimizations to the protocol. used. For competition experiments, unlabeled competitor The ChIP DNA was size selected using Ampure beads (Life DNA was added in excess. Protein-DNA-complexes were technologies) to enrich for fragments <400 bp prior to end- separated by a native 6% (w/v) polyacrylamide 0.5× TBE repair, 3 end adenylation and adapter ligation. Library frag- gel, transferred onto a positively charged Nylon membrane ments were then directly amplified by 10 cycles of PCR. Bar- (GE Healthcare), fixed by Stratagene cross-linker and coded libraries from different samples were pooled together detected by chemiluminescent substrate CSPD (Roche). and sequenced on Illumina HiSeq2000 platform in paired- end sequencing runs. Chromatin immunoprecipitation for ChIP-seq RNA-seq libraries Cells were crosslinked using a formaldehyde containing so- lution (10 mM NaCl, 0.1 mM EDTA pH 8.0, 0.05 mM For preparation of total RNA cells (0.9 Mio/ml) were EGTA pH 8.0, 5 mM HEPES pH 7.8 and 1% formalde- harvested and resuspended in TRIzol reagent (Life Tech- hyde) for 10 min at 20 C, the reaction was quenched by the nologies) and snap-frozen in liquid nitrogen. After thaw- addition of glycin to a final concentration of 250 M for 5 ing RNA was extracted from 0.4 ml of TriZol lysate us- min. Crosslinked cells were collected and washed twice with ing the direct-zol RNA Miniprep (Zymo Research, Irvine PBS before snap freezing in liquid nitrogen and storage at CA, USA) as described in the manufacturer’s protocol. −80 C until subsequent use. RNA was assessed for purity by UV–vis spectrometry (Nan- Prior to sonication, the crosslinked cells were resus- odrop) and for integrity by Bioanalyzer (Agilent Bioana- pended in lysis buffer (50 mM HEPES pH 7.5, 140 mM lyzer 2100, Agilent, Santa Clara USA). RNA was of high NaCl, 1 mM EDTA pH 8.0, 10% glycerol, 0.75% NP-40, purity (abs. 260/280 > 1.9, abs. 269/239 > 2.1) and integrity 0.25% Triton X-100, 1× protease inhibitor cocktail) at 4 C (Bioanalyzer RIN > 9) and thus used for further process- for 20 min. Nuclei were collected by centrifugation and ing. For production of RNA-seq libraries total RNA was washed in a second buffer (200 mM NaCl, 1 mM EDTA DNAse treated (dsDNAse, Fermentas) and 100 ng of this pH 8.0, 0.5 mM EGTA pH 8.0, 10 mM Tris pH 8.0, 1× RNA was processed with a strand-specific protocol (RNA- protease inhibitor cocktail) for 10 min at 4 C then collected seq complete kit, NuGEN, San Carlos, USA). In brief the by centrifugation and resuspended in the shearing buffer (1 RNA was reverse transcribed to cDNA with a reduced mM EDTA pH 8.0, 0.5 mM EGTA pH 8.0, 10 mM Tris set of hexamer primers, avoiding excessive representation pH 8.0, 100 mM NaCl, 0.1% Na-Deoxycholate, 0.5% N- of rRNA in the cDNA. Second strand cDNA synthesis Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 704 Nucleic Acids Research, 2019, Vol. 47, No. 2 was done in presence of dUTP. After ultrasonic fragmen- Purification of ZNF768 for mass spectrometric analysis tation of the cDNA and end repair, Illumina-compatible For purification of ZNF768, Raji or U2OS cells (3 × 10 ) adapter were ligated. Adapters contained uracil in one were collected and IP was performed in 3 biological repli- strand, allowing complete digestion of the second-strand cates as described in the respective paragraph of immuno- derived DNA. After strand selection the libraries were am- precipitation. Simultaneously, ZNF768 antibody (7D6) plified, assessed for correct insert size on the Agilent Bio- and Pes1 (8E9) antibody, respectively, was coupled to analyser and diluted to 10 nM. Barcoded libraries were Sepharose A and G beads for 4 h at 4 C. ZNF768 anti- mixed in equimolar amounts and sequenced on an Illumina body (7D6) was used to identify the interactome of ZNF768 HiSeq1500 in single-read mode with a read length of 100 bp. whereas Pes1 (8E9) antibody served as a subclass control (52,53). Deep sequencing ChIP-seq and RNA-seq analysis was performed as previ- On beads digestion ously described (35). Four biological replicates of U2OS After the last washing step with lysis buffer, beads were and Raji cells were used for RNA-seq library construction. washed three times by adding 100 lof50mMNH HCO . 4 3 For trypsin digest, beads were transferred to a clean tube ChIP-seq data processing and incubated with 100 lof10ng/l trypsin-solution in 1M urea and 50 mM NH HCO for 30 min at 25 C. Sam- Raw sequencing reads were aligned to the human genome 4 3 ples were centrifuged at 800 rpm and supernatant was trans- (hg38) using BWA (36). Sequence reads with an alignment ferred into a fresh tube. Beads were washed twice with 50 l score <30 for paired-end reads and <20 for single-end reads of 50 mM NH HCO . The supernatants were pooled into were discarded as well as all reads that aligned equally well 4 3 the corresponding tube and incubated overnight at 25 C to different positions in the genome. Peak calling was per- after addition of 1mM DTT. Iodoacetamide (IAA) 10 l formed using GEM (37) in GPS mode and with a q-value (5mg/ml) was added and incubated for 30 min in the dark cutoff of 0.01. Overlapping peaks (peak centers: ±100 bp) at 25 C. To quench the IAA, 1 lof1MDTT wasadded were merged both within samples and across all four sam- and samples were incubated for 10 min at 25 C, followed ples to obtain the final list of unique peaks. Motif discov- by addition of 2.5 l of trifluoroacetic acid (TFA) and de- ery was performed using MEME-ChIP (38) and sequence salting using 2× C18 Stagetips (54). Stagetips were washed logos of the binding motif and surrounding regions were three times with 20 l of 100% ACN (1000 rpm, 1 min) and created using weblogo (39). Annotation of peaks relative to three times by adding 20 l of 0.1% TFA (1800 rpm, 1 min). gene features was performed using the ChipSeeker package Subsequently, samples were added (800 rpm, 30 min) and in R (40). Gene annotations were taken from GENCODE washed 3 times with 20 l of 0.1% TFA, followed by elu- version 25 (41). Repeat annotations by RepeatMasker and tion into a clean tube by washing three times with 20 lof phyloP100 conservation scores (42) for hg38 were down- 80% ACN/25% TFA solution. Finally, samples were evap- loaded from the UCSC genome browser. Visualization of orated to dryness, resuspended in 20 l formic acid solution ZNF768 binding on the genome and corresponding peaks and stored at −20 C until LC–MS analysis. was performed using Gviz (43). For the analysis of binding frequencies of ZNF768 in promoter, UTR, exon and intron regions, the same number of 200 bp regions were randomly Protein quantification by liquid chromatography coupled to selected using BEDtools (44). tandem mass spectrometry (LC–MS/MS) Purified peptides (5 l) were automatically injected into in RNA-seq data processing an Ultimate 3000 RSLC HPLC system (Dionex Thermo), Quality check of sequencing reads was performed using separated on an analytical column C18 micro column (75 FastQC (available at: http://www.bioinformatics.babraham. m i.d. × 15 cm, packed in-house with Reprosil Pur C18 ac.uk/projects/fastqc). Sequencing reads were mapped AQ 2.4 m, Doctor Maisch) using a 50-min gradient from against the human genome (hg38) and human rRNA se- 5 to 60% acetonitrile in 0.1% formic acid. The effluent quences using ContextMap version 2.7.9 (45) (using BWA from the HPLC was subsequently electrosprayed into a as short read aligner and default parameters). Number of LTQ Orbitrap XL mass spectrometer (Thermo). The MS read fragments per gene were determined from mapped instrument was operated in a data dependent mode to RNA-seq reads in a strand-specific manner using feature- automatically switch between full scan MS and MS/MS Counts (46) and GENCODE version 25 gene annotations. acquisition. Survey full scan MS spectra (from m/z 300 to RPKM values were calculated using edgeR and averaged 1800) were acquired in the Orbitrap with a resolution of R between replicates (47). Differential gene expression anal- = 60 000 at m/z 400 (after accumulation to a ‘target value’ ysis was performed using limma (48). Functional enrich- of 500,000 in the linear ion trap). The six most intense ment analysis for UniProt keywords and Gene Ontology peptide ions with charge state between 2 and 4 were sequen- terms was performed with the DAVID webserver (49). Sig- tially isolated to a target value of 10,000 and fragmented in nificantly enriched terms were determined using a cutoff of the linear ion trap by collision induced dissociation (CID). 0.05 on the P-value adjusted for multiple testing using the For all measurements with the Orbitrap mass analyzer, method by Benjamini and Hochberg (50). Analysis work- three lock-mass ions from ambient air (m/z = 371.10123, flows were implemented and run using the workflow man- 445.12002, 519.13882) were used for internal. Usual MS agement system Watchdog (51). conditions were: spray voltage, 1.5 kV; no sheath and Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 705 auxiliary gas flow; heated capillary temperature, 200 C; U2OS cells (Figure 2A and Supplementary Figure S3A). normalized collision energy 35% for CID in LTQ. The Expression of the zinc finger-containing C-terminal domain threshold for ion selection was 10 000 counts for MS2. of ZNF768 (Figure 2E) caused a similar staining pattern, The used activation was 0.25 and activation time 30 ms. while expression of the N-terminus containing the array of MaxQuant 1.5.2.8 was used to identify proteins and quan- heptad repeats resulted in a more diffuse staining of the nu- tify by iBAQ with the following parameters: Database, cleus (Supplementary Figure S3B), suggesting that the zinc Uniprot Hsapiens 3AUP000005640 170526; MS tol, finger domain is responsible for the association of ZNF768 10ppm; MS/MS tol, 0.5 Da; Peptide FDR, 0.1; protein with euchromatin. Antibody 7D6 immunoprecipitated en- FDR, 0.01 Min. peptide length, 5; variable modifications, dogenous ZNF768 protein quantitatively from extracts of oxidation (M); fixed modifications, carbamidomethyl (C); osteosarcoma cell line U2OS and B-lymphoid cell line Raji peptides for protein quantitation, razor and unique; min. (Figure 2B), proving its high specificity and suitability to peptides, 1; min. ratio count, 2. Identified proteins were study the binding of ZNF768 to DNA in chromatin im- considered as interaction partners if their MaxQuant munoprecipitation (ChIP) experiments. Knockdown exper- iBAQ values displayed a greater value than log 5-fold iments of ZNF768 confirmed the specificity of mAb 7D6 enrichment (FC) and P-value 0.05 (t-test adjusted for (Figure 2C) and showed further that ZNF768 is required multiple comparisons) when compared to the control. for viability and proliferation of U2OS cells (Figure 2D). In line with its essential function, expression of mutants with deletions of either the N- or C-terminal domain of ZNF768 RESULTS have a dominant-negative phenotype and inhibit cell pro- ZNF768 domain structure and conservation liferation (Figure 2E–G). Finally, ZNF768 is a phospho- protein and can be phosphorylated at almost all heptad re- An array of ten zinc fingers at the C-terminus of ZNF768 peat serine-5 residues (www.cellsignal.com, Supplementary shows high conservation in placentalia and marsupials Figure S2A). Treatment of cellular extracts of U2OS cells (>96%) but is less conserved in monotremes (blue boxes in with alkaline phosphatase causes a shift of the hyperphos- Figure 1 and Supplementary Figure S1). At the N-terminus, phorylated form of ZNF768 (Supplementary Figure S2C) two sequence blocks (box A and box B, red and green and reveals that a large fraction of ZNF768 is hyperphos- boxes in Figure 1) are conserved in placentalia and marsu- phorylated in U2OS cells. pials, but replaced by unrelated sequences in monotremes. In addition, ZNF768 of placentalia has evolved an array of heptad repeats that is positioned between box A and box B. The number of repeats varies between 20 repeats Identification of the ZNF768 binding motif in cellular DNA in mouse lemur and 10 repeats in pika and malayan pan- To investigate if ZNF768 can bind to specific DNA se- golin. Mouse ZNF768 contains 19 repeats, chimpanzee 16 quences, we performed ChIP experiments with mAb 7D6 and human 15 repeats (Supplementary Figure S1). Similar using extracts of U2OS and Raji cells. DNA libraries of to heptad repeats in CTD of Pol II, the heptad repeats in two biological replicates were prepared for each cell line and ZNF768 show no length variation (except a single extended analyzed by next generation sequencing. Peak calling iden- repeat in pika). However, the composition of amino acids tified a total of 21 012 unique peaks and 13.1% of these in the heptad repeats shows higher variation in ZNF768, peaks were consistently identified in all four samples and both, within and between species (Supplementary Figures an additional 28.8% at least in both replicates for the same S1 and S2B). Serine-5 and proline-6 residues show the high- cell type (Supplementary Figure S4). Generally, ZNF768 est conservation between ZNF768 and Pol II, followed by binding sites distributed over all chromosomes (Supplemen- the residues corresponding to tyrosine-1 and proline-3 in tary Figure S5). Motif discovery identified several potential the CTD, the position of threonine-4 and serine-7 show binding motifs for ZNF768 (Supplementary Figure S6A). only little or almost no conservation. The position corre- The top two identified motifs were found in 46% and 37% sponding to serine-2 in the CTD is particularly remarkable of peaks, respectively, and for both motifs the other motif in ZNF768. It is replaced in almost all repeats by an acidic was often found as a secondary motif at a distance of ∼20 amino acid (mostly glutamic acid). The phosphorylation of bp. We thus hypothesized that the ZNF768 binding motif serine-2 residues in CTD by P-TEFb is a hallmark in RNA consists of anchor regions connected by a linker region of elongation control (55) and a replacement of serine-2 may ∼20 bp. In fact, 58.1% of identified peaks contained this mimic its phoyphorylation. It is thus tempting to speculate consensus motif with at most three mismatches in the an- that the array of heptad repeats in ZNF768 potentially can chor regions and a linker region of 20 ± 3bp(Figure 3A, mimic a CTD phosphorylated at serine-2 residues. Supplementary Figures S4 and S6B). For peaks identified in all 4 samples, this number was as high as 98.3% and the ZNF768 is associated with euchromatin and required for vast majority of peaks with motif hits (83.5%) had a linker growth and cell viability length of 20 bp (Supplementary Figure S6B). Gelshift ex- To study the cellular function of ZNF768 we first raised a periments with recombinant ZNF768 protein confirmed the monoclonal antibody (7D6) towards human ZNF768 us- motif, GCTGTGTG (N ) CCTCTCTG, and revealed that ing a peptide containing heptad repeats 8–12 as epitope the nucleotide sequence of the spacer between the two an- (Figure 1A, materials and methods). This antibody prefer- chor regions is likely not critical for binding (Supplemen- entially stains euchromatic regions in the nucleus of fixed tary Figure S7). Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 706 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 2. ZNF768 is associated with euchromatin and essential for cell viability and proliferation. (A) Confocal image of U2OS cells stained with DAPI and with the ZNF768-specific mAb 7D6; merge of images on the right hand site. White line marks the area of the RGB profiler, profiles of DAPI and ZNF768 at the bottom. (B) mAb 7D6 immunoprecipitates a 70 kD protein from extracts of Raji and U2OS cells. SN: supernatant, *: Ig heavy chain. (C)siRNA mediated knockdown of ZNF768 in U2OS cells. (D) Growth kinetics of U2OS cells after knockdown of ZNF768 measured by xCelligence (Roche). Arrows indicate consecutive addition of siRNA. (E) Expression constructs of HA-tagged ZNF768 wild-type and deletion mutants and (F) expression control in U2OS cells. (G) Growth kinetics of U2OS cells after expression of ZNF768-WT and ZNF768 mutants measured by xCelligence (Roche). Arrows indicate addition of doxycycline. ZNF768 binds to MIR sequences tained in a MIR sequence and the consensus sequence for all MIR types actually contains the ZNF768 binding motif A systematic comparison of ZNF768 binding sites to re- (Supplementary Figure S8A). Despite this fact, only a small peats in the human genome showed an enrichment of fraction (12.2%) of MIRs in the human genome contains binding sites within all four types of MIRs (Figure 3B). the ZNF768 binding motif, which is not surprising giving 12,488/21,012 peaks overlapped with MIRs. Furthermore, a per-base identity <80% for human MIR sequences. Al- almost all peaks (92%) with the binding motif were con- though only 15.8% of MIRs containing the binding mo- Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 707 Figure 3. DNA binding motif and genomic binding of ZNF768 to MIRs. (A) Consensus ZNF768 binding motif identified by ChIP-seq experiments in Raji and U2OS cells and determined from peaks containing the motif (±3 mismatches in the anchors and linker length of 20 ± 3 bp). (B) Number of ZNF768 binding sites overlapping with particular type of repetitive sequences (top; in case of multiple overlaps the largest overlap is used) compared to the genomic length covered by the corresponding type of repetitive sequence (bottom). The right-most bar shows the number of ZNF768 binding sites with no overlap to repetitive sequences. tif were found to be bound by ZNF768, this fraction in- bound MIR sequences showed no particular conservation creased to 54.2% when considering MIRs with a more strin- (Figure 4B) indicating that ZNF768 binding represents a gent and strict version of the motif (linker length: 19/20bp, conserved function of a subset of MIR sequences in mam- 1 mismatch in anchors). Thus, most MIRs diverged so far mals. from the consensus that the binding motif was lost and no general conservation of the binding motif in human MIRs ZNF768 binding is associated with transcribed genes was observed (Supplementary Figure S8B). This divergence also allowed reliably aligning reads to MIR sequences de- We next asked if ZNF768 binding sites were enriched in reg- spite their repetitive origin. Only reads that could be aligned ulatory elements of genes. We found a strong enrichment of uniquely to the genome were used for peak calling. Al- ZNF768 binding in promoters and a slight enrichment of though 8524 detected peaks (40.6%) were not within MIRs, binding in exons and introns, while the binding frequency in 13.5% of these peaks contained the binding motif. The re- intergenic sequences was reduced (Figure 5A). Interestingly, maining peaks may contain a weaker version of the binding the 1061 ZNF768 binding sites outside of MIRs that con- motif, recruit ZNF768 to chromatin by other mechanisms tained the binding motif showed an even higher enrichment (e.g. looping), or represent spurious binding. at promoters. To investigate whether genes with ZNF768 Interestingly, MIRs with ZNF768 binding show a clear binding tended to be more highly expressed, we analyzed conservation of the two anchor motifs in the human RNA-seq data of four replicates of total RNA of Raji and genome. Sequences of the linker in the binding motif and U2OS cells (Supplementary Table S1). In both cell lines, outside of the binding motif were not particularly con- protein-coding genes with ZNF768 binding in the promoter served, similar to MIRs without binding of ZNF768 (Sup- or 5 UTR were more highly expressed on average than the plementary Figure S8B). We further investigated whether remaining protein-coding genes (Figure 5B). In contrast, ZNF768 binding sites in MIRs were also conserved across binding in intronic regions showed only a small but signifi- species by analyzing phyloP100 conservation scores deter- cant effect (Figure 5B). This provides evidence that ZNF768 mined from a multiple alignment of 99 vertebrate genomes regulates transcription by binding in or near promoter re- against the human genome. Positive PhyloP scores indicate gions of active genes. slower than expected evolution. The analysis of phyloP100 scores within and around the binding motif (±25 bp) in MIR sequences bound by ZNF768 showed increased con- ZNF768 binds to genes with cell type-specific expression servation for most positions within the anchor regions (Fig- Raji and U2OS cells revealed common and cell type-specific ure 4A). Sequences outside of the binding motif or within binding sites of ZNF768. Common sites were for instance the linker region, however, were mostly not conserved. Un- associated with genes for RNA polymerase II subunit E Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 708 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 4. Conservation of the ZNF768 motif (+25 bp on either side) in MIRs either (A) bound or (B) not bound by ZNF768. Only MIRs were considered that align without gaps to the MIR consensus sequence in the region of the ZNF768 binding motif +25 bp on either side. Distribution of PhyloP100 scores are indicated by boxplots for each position (green = median PhyloP100 score > 0.2) and regions within the motif region are indicated in more intensive colors. (POLR2E) (Figure 6A) or Solute Carrier Family 1 Mem- the fold-changes in gene expression were lower. This may be ber 5 (SLC1A5) and Mitochondrial Ribosomal Protein S5 due to the higher number of peaks identified in Raji cells, in- (MRPS5) (Supplementary Figure S9A,B). These genes are dicating a higher sensitivity but lower specificity compared expressed in Raji and U2OS cells and show similar peaks to peaks in U2OS cells. Thus, a significant fraction of seem- in both cell lines. Thus, many of the 2747 identified com- ingly Raji-specific peaks may simply have been missed in mon binding sites in Raji and U2OS may be associated with U2OS. We conclude that binding of ZNF768 occurs prefer- commonly expressed genes. We also identified a large num- entially at expressed genes and at least in part in a cell type- ber of peaks that were present either in Raji or U2OS cells. specific manner. The underlying mechanisms regulating the In U2OS cells, strong peaks for ZNF768 were associated cell type-specific binding of ZNF768 in Raji and U2OS cells with the promoter region of the GAS2L1 gene (Figure 6B), are currently unclear, but may involve, e.g. DNA methyla- the ID1 and SNPH genes (Supplementary Figure S9C,D) tion or other epigenetic marks. and the gene body of the ANXA2 and ALDH7A1 genes (Supplementary Figure S9E, F), but were absent or only faintly detectable in Raji cells. These genes are expressed ZNF768-regulated genes in U2OS cells in U2OS but not in Raji cells. Inversely, Raji cells showed a strong ZNF768 binding site in the promoter region of To study the gene regulatory potential of ZNF768, the B-Lymphocyte Surface Antigene (CD19) gene (Figure we induced expression of the dominant-negative mutant 6C), which is a B cell-specific non-receptor tyrosine kinase ZNF768-N(Figure 2E) in U2OS cells and analyzed required for B cell receptor signaling. Strong Raji-specific changes in the transcriptome after 12 h. A >2-fold change peaks were further detected for the genes CD86, ATP2A3, in RNA levels was detected for 500 downregulated and 155 RHOH, PLCG2, LYN, and ARHGDIB (Supplementary upregulated genes (Supplementary Table S2). Functional Figure S9G–L). These genes are expressed in Raji but not enrichment analysis of repressed genes revealed several sig- U2OS cells. nificantly enriched gene sets including two gene sets con- A global analysis of differential gene expression between taining DNA binding proteins (105 genes) and zinc finger U2OS and Raji cells showed significant differences in fold- proteins (103 genes) (Figure 7A and Supplementary Table changes for genes with peaks specific to either cell line (Fig- S3). Repressed genes in both gene sets show a large over- ure 6D and Supplementary Table S1). In particular, genes lap (63 genes) with repressed transcription-associated genes with U2OS-specific peaks were on average 30-fold higher (Figure 7B). We conclude that ZNF768 can act as transcrip- expressed in U2OS cells. For genes with Raji-specific peaks, tional regulator and is required particularly for the expres- sion of other transcription factors. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 709 Figure 5. Genomic distribution of the ZNF768 binding motif in Raji and U2OS cells. (A) Frequency of ZNF768 binding sites in promoters (–1 kb to transcription start site) and other genomic regions compared to randomly selected binding sites with the same peak length distribution. This shows an enrichment of genomic binding of ZNF768 in promoters, in particular for motif-containing peaks outside of MIRs. (B) Boxplots illustrating the distribution of expression levels in total RNA (quantified as RPKM = reads per kilobase per Million mapped reads) in Raji or U2OS cells for genes without (wo) and with (w) peaks in the respective cells. A pseudocount of 1 was added to all RPKM values before plotting. P-values for a Wilcoxon rank sum test comparing −3 −5 −10 RPKM levels between the two groups are indicated as: *P < 10 ,**P < 10 , ***P < 10 Mass spectrometric analysis of ZNF768 associated factors (Elp1, Elp2 and Elp3), SR rich splicing factor (SUGP2), centromere protein E (CENPE), several E3 ligases (USP13, We used the mAb 7D6 for a combined immunoprecipitation Trim33, and HERC2), proteins with centrosomal functions (IP) and mass spectrometric (MS) assay to identify ZNF768 (CEP170-1, Cep170-2 and NIN), and other factors. The associated factors. The ZNF768 interactome of Raji and binding of Elongator subunit Elp3 to ZNF768 was con- U2OS cells showed a large overlap and twenty of the best firmed in IP experiments with an Elp3-specific antibody thirty interactors were found in both cell lines (Figure 8 (Figure 8C). mAb 7D6 could immunoprecipitate a signifi- A,B, Supplementary Figure S10). Among the common fac- cant fraction of Elp3 protein of cellular extracts of Raji cells. tors we identified three subunits of the Elongator complex Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 710 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 6. Common and cell type-specific peaks of ZNF768 in Raji and U2OS cells. ChIP-seq (replicates shown separately) and RNA-seq (mean of four replicates) read coverage (in counts per million) for example genes. Identified peaks are shown as rectangles below the corresponding ChIP-seq sampl e. Genomic coordinates and gene annotation (boxes = exons, lines = introns, strand indicated by arrowheads) are shown in the bottom row. (A) ZNF768 binding in the promoter upstream region of the RNA Polymerase II Subunit E (POLR2E) gene, which is expressed in both Raji and U2OS cells. (B) Binding of ZNF768 in the promoter region of the Growth Arrest Specific 2-Like (GAS2L1) gene, which is expressed in U2OS but not Raji cells. ( C) Binding of ZNF768 to the promoter region of the B-Lymphocyte Surface Antigene (CD19) gene, which is expressed in Raji but not U2OS cells. (D) Genes in Raji and U2OS cells with cell-specific peaks differ in gene expression. Boxplots illustrate the distribution of fold-changes in gene expression between b oth cell lines (determined with limma) for genes with cell-specific peaks ( = peaks identified in gene body or 1 kb upstream in both replicates of the corresponding cell line but not for the other cell line; to account for the differences in sensitivity between Raji and U2OS ChIP-seq, for Raji only the 103 genes with the top-scoring Raji-specific peaks were evaluated, i.e. the same number of genes as with U2OS-specific peaks). Significance of the difference in median va lues −10 was determined using the Wilcoxon rank sum test (***P ≤ 10 ). The results suggests that ZNF768 can recruit Elongator and near promoters, suggesting that binding of ZNF768 is asso- other factors to expressed genes in Raji and U2OS cells. ciated with gene expression. In agreement with this assump- tion we observed ZNF768 binding preferentially in euchro- matic regions of the nucleus. Likewise, MIR sequences have DISCUSSION been reported to be associated with transcriptional active ZNF768 binds to MIR sequences euchromatin but not heterochromatin (5,6). Strikingly, the number of MIR sequences in mammals varies considerably ZNF768 proteins in mammals contain an array of ten zinc from about 20% of the total genome in monotremes to 1% fingers that allow the specific binding to DNA. ChIP-seq ex- or 3% of the genome in mice and humans, respectively. Fur- periments revealed approximately ten to twenty thousand thermore, the ZNF768 binding motif, although part of the ZNF768 binding sites in the genome of Raji and U2OS MIR consensus sequence, is not conserved in all MIRs, but cells. The majority of these sites is contained within MIR only in those displaying a peak in ZNF768 ChIP-seq exper- sequences and shares a common binding motif that is part iments. Notably, we also detected ∼1000 peaks containing of the MIR consensus sequence. The motif of the binding the ZNF768 binding motif outside of MIRs. This category site is 36 bp long and consists of two anchor sequences of 8 of peaks showed the highest association with promoters. bp separated by a linker of 20 bp, which probably does not Given the length of the detected DNA binding motif and contribute to the binding specificity of ZNF768 as revealed the position of the two anchor sequences at its flanks it is by gel shift experiments. ZNF768 binds preferentially at or Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 711 Figure 7. Functional enrichment analysis for UniProt keywords for genes downregulated upon expression of the dominant-negative mutant ZNF768-N in U2OS cells. (A) Significantly enriched UniProt keywords (identified with DAVID at an adjusted P-value < 0.05) for downregulated genes (>2-fold down-regulated, adjusted P-value <0.01, for full details see also Supplementary Table S3). (B) Venn diagram of downregulated genes annotated with the keywords Transcription, DNA-binding, and Zinc finger. The data indicates that inhibition of ZNF768 downregulates other transcription factors with zinc finger domains. likely that the proximal and distal, but not the central zinc Cell type-specific binding of ZNF768 to gene loci fingers, of the array of 10 zinc fingers contribute to DNA ChIP-seq analysis of ZNF768 revealed common but also a binding. The potential function of the central zinc fingers is large number of differential binding sites in Raji and U2OS currently unknown. We have currently no evidence for other cells. Furthermore, many putative binding sites containing conserved motifs upstream or downstream of the ZNF768 the binding motif were not occupied by ZNF768 in either binding motif. Notably, we also observed ZNF768 peaks at Raji or U2OS cells. This observation suggests that binding gene loci that do not contain the DNA binding motif. It is of ZNF768 to DNA is regulated and that not all binding currently unclear if binding to these loci requires the zinc sites are equally accessible in Raji and U2OS. The mecha- finger domain and /or other parts of the protein. nism(s) regulating the different accessibility is currently un- known but may include DNA methylation, histone compo- sition at binding motifs or specific histone marks. Addition- ZNF768 is an essential gene for cell proliferation ally, other cellular factors may block or permit binding of Knockdown experiments as well as the expression of ZNF768 to the binding motif. In this context it will be im- dominant-negative mutants revealed the functional require- portant to determine at which stage of cell differentiation ment of ZNF768 for cell viability and proliferation. Expres- the access of ZNF768 to its binding motif is regulated. sion of a mutated form of ZNF768 containing only the C- The observed differential binding of ZNF768 in Raji and terminal or N-terminal domain, respectively, led to a de- U2OS cells further prompted us to ask whether binding of cline of the cell index in cell proliferation assays. A decline of ZNF768 can mark differentially expressed genes in both this index was also seen after siRNA-mediated knockdown cell lines. In fact, we found a general correlation between of ZNF768 expression. This indicates that the functional ZNF768 binding and the activity of adjacent genes. In par- loss of ZNF768 cannot be compensated by other cellular ticular, we found a correlation between ZNF768 binding factors. Our results suggest that ZNF768, despite being an and gene expression for those genes that are active only in evolutionary young gene, gained essential function(s) for Raji or U2OS cells. From these data we conclude that bind- the expression of growth related genes. A detailed genetic ing of ZNF768 can mark commonly as well as cell type- analysis combined with mass spectrometry experiments will specifically expressed genes in Raji and U2OS cells. be required in the future to analyze the function of ZNF768 in the context of growth control in more detail. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 712 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 8. ZNF768 interactome. ZNF768 was immunoprecipitated from cellular extracts of (A) Raji and (B) U2OS cells. Thirty interaction factors with the highest enrichment are shown. Common factors in both cell lines are depicted in red. The list of all interaction factors is shown in Supplementary Table S4. (C) ZNF768 mAb 7D6 specifically co-immunoprecipitates Elp3 from cellular extracts of Raji cells. ZNF768 functions as transcription factor ulated indirectly. The gene ontology analysis of repressed genes revealed several gene classes related to transcriptional Finally, we asked if binding of ZNF768 is required for ex- regulation suggesting that ZNF768 is hierarchically located pression of specific genes. To demonstrate this we stud- upstream of a network of transcription factor genes and ied the transcriptome of U2OS cells 12 h after overexpres- may function as a regulatory master gene for this network. sion of a ZNF768 mutant lacking the N-terminal domain. The notion that ZNF68 may act as a transcription factor We found several hundred genes that were significantly re- was further supported by mass spectrometric analysis of the pressed after expression of this dominant-negative mutant. ZNF768 interactome in Raji and U2OS cells. In both cell We also found a few induced genes, which may be upreg- Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 713 Figure 9. Known functions and regulation of Pol II CTD via its heptad repeat array and the implication with regards to possible functions and regulation of the heptad array in ZNF768 of placentalia (represented by human ZNF768). lines ZNF768 interacts with subunits Elp1, Elp2 and Elp3 and MIR sequences has contributed to speciation of pla- of the Elongator complex. The complex is conserved from centalia. yeast to mammals, consists of six subunits, Elp1-6, and has been proposed to function in the control of RNA elonga- DATA AVAILABILITY tion (56). The Elongator was found associated to the hy- perphoshorylated form of Pol II, but the mode of interac- GEO submissions: ChIP-Seq (GSE111879), RNA-seq Raji tion and the involved Elongator subunits are still elusive. cells (GSE111880), RNA-seq U2OS cells (GSE111881). Our data suggest that recruitment of Elongator to active The mass spectrometry proteomics data have been de- genes may also occur by ZNF768. ZNF768 binds first a posited to the ProteomeXchange Consortium via the subcomplex of Elongator consisting of Elp1-3 that subse- PRIDE (57) partner repository with the dataset identifier quently may assemble with subunits Elp4-6. In the future it PXD010831. will be interesting to study if heptad repeats of ZNF768 are involved in the recruitment of Elongator, as suggested for SUPPLEMENTARY DATA the CTD of Pol II, and if ZNF768 of marsupials lacks the Supplementary Data are available at NAR Online. ability of Elongator recruitment. Originally, the array of heptad repeats in ZNF768 at- tracted our attention to study the function of ZNF768 as ACKNOWLEDGEMENTS transcriptional activator due to its similarity to the array We thank Elisabeth Kremmer for help with the generation of heptad repeats in CTD of Pol II. This raises a couple of of ZNF768 mAb. intriguing questions. First, can this array fulfill similar or related functions as the array of heptad repeats in CTD? If so, can the acidic amino acids that are present at many po- FUNDING sitions in heptad repeats of ZNF768 mimic a hyperphos- D.E. and A.I. were supported by the Deutsche Forschungs- phorylated form of Pol ll? Such a mimicry is most likely gemeinschaft (DFG), SFB1064, Chromatin Dynamics and for position 2 of heptad repeats in ZNF768, which contains DFG excellence cluster CIPSM. In D.E. and J.C.A. labs, glutamic acid in almost all repeats across all species. It is the work was supported by a German-French BMBF- tempting to speculate that binding of ZNF768 can recruit ANR grant ‘EpiGlyco’. CNRS; ‘Agence Nationale de la cellular factors to genomic loci that otherwise are recruited Recherche’ (ANR); ‘amorcage jeunes equipes’ ´ Fondation only if serine-2 of CTD is phosphorylated, e.g. by Cdk9, or pour la Recherche Medicale FRM [AJE20130728183 other kinases (see model in Figure 9). In contrast, serine- to J.C.A.]; Deutsche Forschungsgemeinschaft (DFG) 5 residues are conserved between ZNF768 and the CTD [FR2938/7-1 and CRC 1123 (Z2) to C.C.F. and M.K.]; and may depend on phosphorylation in ZNF768, similar Deutsche Forschungsgemeinschaft (DFG) [GE 976/9-2 as in the CTD, to allow interaction with other factors. Fu- to M.G.] and is a member of the DFG excellence clus- ture work will address these and other questions and illu- ter ImmunoSensation. Funding for open access charge: minate if and how the new regulatory network of ZNF768 Helmholtz Center Munich. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 714 Nucleic Acids Research, 2019, Vol. 47, No. 2 Conflict of interest statement. None declared. 23. Eick,D. and Geyer,M. (2013) The RNA polymerase II carboxy-terminal domain (CTD) code. Chem. Rev., 113, 8456–8490. 24. Harlen,K.M. and Churchman,L.S. (2017) The code and beyond: transcription regulation by the RNA polymerase II carboxy-terminal REFERENCES domain. Nat. Rev. Mol. Cell Biol., 18, 263–273. 25. Jeronimo,C., Collin,P. and Robert,F. (2016) The RNA polymerase II 1. Kramerov,D.A. and Vassetzky,N.S. (2011) Origin and evolution of CTD: the increasing complexity of a low-complexity protein domain. SINEs in eukaryotic genomes. Heredity (Edinb.), 107, 487–495. J. Mol. Biol., 428, 2607–2622. 2. Redi,C.A. and Capanna,E. (2012) Genome size evolution: sizing 26. Zaborowska,J., Egloff,S. and Murphy,S. (2016) The pol II CTD: new mammalian genomes. Cytogenet Genome Res., 137, 97–112. twists in the tail. Nat. Struct. Mol. Biol., 23, 771–777. 3. Jurka,J., Zietkiewicz,E. and Labuda,D. (1995) Ubiquitous 27. Seipel,K., Georgiev,O., Gerber,H.P. and Schaffner,W. (1994) Basal mammalian-wide interspersed repeats (MIRs) are molecular fossils components of the transcription apparatus (RNA polymerase II, from the mesozoic era. Nucleic Acids Res., 23, 170–175. TATA-binding protein) contain activation domains: is the repetitive 4. Smit,A.F. and Riggs,A.D. (1995) MIRs are classic, tRNA-derived C-terminal domain (CTD) of RNA polymerase II a “portable SINEs that amplified before the mammalian radiation. Nucleic Acids enhancer domain”? Mol. Reprod. Dev., 39, 215–225. Res., 23, 98–102. 28. Suh,H., Hazelbaker,D.Z., Soares,L.M. and Buratowski,S. (2013) The 5. Jjingo,D., Conley,A.B., Wang,J., Marino-Ramirez,L., Lunyak,V.V. C-terminal domain of Rpb1 functions on other RNA polymerase II and Jordan,I.K. (2014) Mammalian-wide interspersed repeat subunits. Mol. Cell, 51, 850–858. (MIR)-derived enhancers and the regulation of human gene 29. Burke,K.A., Janke,A.M., Rhine,C.L. and Fawzi,N.L. (2015) expression. Mob. DNA, 5, 14. Residue-by-residue view of in vitro FUS granules that bind the 6. Jjingo,D., Huda,A., Gundapuneni,M., Marino-Ramirez,L. and C-terminal domain of RNA polymerase II. Mol. Cell, 60, 231–241. Jordan,I.K. (2011) Effect of the transposable element environment of 30. Hnisz,D., Shrinivas,K., Young,R.A., Chakraborty,A.K. and human genes on gene length and expression. Genome Biol. Evol., 3, Sharp,P.A. (2017) A phase separation model for transcriptional 259–271. control. Cell, 169, 13–23. 7. Smith,A.M., Sanchez,M.J., Follows,G.A., Kinston,S., 31. Kwon,I., Kato,M., Xiang,S., Wu,L., Theodoropoulos,P., Mirzaei,H., Donaldson,I.J., Green,A.R. and Gottgens,B. (2008) A novel mode of Han,T., Xie,S., Corden,J.L. and McKnight,S.L. (2013) enhancer evolution: the Tal1 stem cell enhancer recruited a MIR Phosphorylation-regulated binding of RNA polymerase II to fibrous element to specifically boost its activity. Genome Res., 18, 1422–1432. polymers of low-complexity domains. Cell, 155, 1049–1060. 8. Wang,J., Vicente-Garcia,C., Seruggia,D., Molto,E., 32. Schwartz,J.C., Ebmeier,C.C., Podell,E.R., Heimiller,J., Taatjes,D.J. Fernandez-Minan,A., Neto,A., Lee,E., Gomez-Skarmeta,J.L., and Cech,T.R. (2012) FUS binds the CTD of RNA polymerase II Montoliu,L., Lunyak,V.V. et al. (2015) MIR retrotransposon and regulates its phosphorylation at Ser2. Genes Dev., 26, 2690–2695. sequences provide insulators to the human genome. Proc. Natl. Acad. 33. Bornkamm,G.W., Berens,C., Kuklik-Roos,C., Bechet,J.M., Laux,G., Sci. U.S.A., 112, E4428–E4437. Bachl,J., Korndoerfer,M., Schlee,M., Holzel,M., Malamoussi,A. 9. Varshney,D., Vavrova-Anderson,J., Oler,A.J., Cowling,V.H., et al. (2005) Stringent doxycycline-dependent control of gene Cairns,B.R. and White,R.J. (2015) SINE transcription by RNA activities using an episomal one-vector system. Nucleic Acids Res., 33, polymerase III is suppressed by histone methylation but not by DNA e137. methylation. Nat. Commun.. 6, 6569. 34. Rohrmoser,M., Holzel,M., Grimm,T., Malamoussi,A., Harasim,T., 10. Yeganeh,M., Praz,V., Cousin,P. and Hernandez,N. (2017) Orban,M., Pfisterer,I., Gruber-Eber,A., Kremmer,E. and Eick,D. Transcriptional interference by RNA polymerase III affects (2007) Interdependence of Pes1, Bop1, and WDR12 controls expression of the Polr3e gene. Genes Dev., 31, 413–421. nucleolar localization and assembly of the PeBoW complex required 11. Carnevali,D., Conti,A., Pellegrini,M. and Dieci,G. (2017) for maturation of the 60S ribosomal subunit. Mol. Cell Biol., 27, Whole-genome expression analysis of mammalian-wide interspersed 3682–3694. repeat elements in human cell lines. DNA Res., 24, 59–69. 35. Shah,N., Maqbool,M.A., Yahia,Y., El Aabidine,A.Z., Esnault,C., 12. Krull,M., Petrusma,M., Makalowski,W., Brosius,J. and Schmitz,J. Forne,I., Decker,T.M., Martin,D., Schuller,R., Krebs,S. et al. (2018) (2007) Functional persistence of exonized mammalian-wide Tyrosine-1 of RNA polymerase II CTD controls global termination interspersed repeat elements (MIRs). Genome Res., 17, 1139–1145. of gene transcription in mammals. Mol. Cell, 69, 48–61. 13. Smit,A.F., Hubley,R. and Green,P. (2015). Repeatmasker Open 4.0. 36. Li,H. and Durbin,R. (2009) Fast and accurate short read alignment http://www.repeatmasker.org. with Burrows-Wheeler transform. Bioinformatics., 25, 1754–1760. 14. Cournac,A., Koszul,R. and Mozziconacci,J. (2016) The 3D folding of 37. Guo,Y., Mahony,S. and Gifford,D.K. (2012) High resolution genome metazoan genomes correlates with the association of similar wide binding event finding and motif discovery reveals transcription repetitive elements. Nucleic Acids Res., 44, 245–255. factor spatial binding constraints. PLoS Comput. Biol., 8, e1002638. 15. Feschotte,C. (2008) Transposable elements and the evolution of 38. Machanick,P. and Bailey,T.L. (2011) MEME-ChIP: motif analysis of regulatory networks. Nat. Rev. Genet., 9, 397–405. large DNA datasets. Bioinformatics, 27, 1696–1697. 16. Medstrand,P., van de Lagemaat,L.N. and Mager,D.L. (2002) 39. Crooks,G.E., Hon,G., Chandonia,J.M. and Brenner,S.E. (2004) Retroelement distributions in the human genome: variations WebLogo: a sequence logo generator. Genome Res., 14, 1188–1190. associated with age and proximity to genes. Genome Res., 12, 40. Yu,G., Wang,L.G. and He,Q.Y. (2015) ChIPseeker: an 1483–1495. R/Bioconductor package for ChIP peak annotation, comparison and 17. Testori,A., Caizzi,L., Cutrupi,S., Friard,O., De Bortoli,M., Cora,D. visualization. Bioinformatics, 31, 2382–2383. and Caselle,M. (2012) The role of Transposable Elements in shaping 41. Harrow,J., Frankish,A., Gonzalez,J.M., Tapanari,E., Diekhans,M., the combinatorial interaction of Transcription Factors. BMC Kokocinski,F., Aken,B.L., Barrell,D., Zadissa,A., Searle,S. et al. Genomics, 13, 400. (2012) GENCODE: the reference human genome annotation for The 18. Matthews,J.M. and Sunde,M. (2002) Zinc fingers–folds for many ENCODE Project. Genome Res., 22, 1760–1774. occasions. IUBMB Life, 54, 351–355. 42. Pollard,K.S., Hubisz,M.J., Rosenbloom,K.R. and Siepel,A. (2010) 19. Yang,P., Wang,Y. and Macfarlan,T.S. (2017) The role of Detection of nonneutral substitution rates on mammalian KRAB-ZFPs in transposable element repression and Mammalian phylogenies. Genome Res., 20, 110–121. evolution. Trends Genet.. 33, 871–881. 43. Hahne,F. and Ivanek,R. (2016) Visualizing genomic data using gviz 20. Imbeault,M., Helleboid,P.Y. and Trono,D. (2017) KRAB zinc-finger and bioconductor. Methods Mol. Biol., 1418, 335–351. proteins contribute to the evolution of gene regulatory networks. 44. Quinlan,A.R. and Hall,I.M. (2010) BEDTools: a flexible suite of Nature, 543, 550–554. utilities for comparing genomic features. Bioinformatics, 26, 841–842. 21. Najafabadi,H.S., Mnaimneh,S., Schmitges,F.W., Garton,M., 45. Bonfert,T., Kirner,E., Csaba,G., Zimmer,R. and Friedel,C.C. (2015) Lam,K.N., Yang,A., Albu,M., Weirauch,M.T., Radovani,E., ContextMap 2: fast and accurate context-based RNA-seq mapping. Kim,P.M. et al. (2015) C2H2 zinc finger proteins greatly expand the BMC Bioinformatics, 16, 122. human regulatory lexicon. Nat. Biotechnol., 33, 555–562. 22. Bentley,D.L. (2014) Coupling mRNA processing with transcription in time and space. Nat. Rev. Genet., 15, 163–175. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 715 46. Liao,Y., Smyth,G.K. and Shi,W. (2014) featureCounts: an efficient member of the Pes1-Bop1 complex and is required for ribosome general purpose program for assigning sequence reads to genomic biogenesis and cell proliferation. J. Cell Biol., 170, 367–378. features. Bioinformatics, 30, 923–930. 53. Kellner,M., Rohrmoser,M., Forne,I., Voss,K., Burger,K., Muhl,B., 47. Robinson,M.D., McCarthy,D.J. and Smyth,G.K. (2010) edgeR: a Gruber-Eber,A., Kremmer,E., Imhof,A. and Eick,D. (2015) Bioconductor package for differential expression analysis of digital DEAD-box helicase DDX27 regulates 3 end formation of ribosomal gene expression data. Bioinformatics, 26, 139–140. 47S RNA and stably associates with the PeBoW-complex. Exp. Cell 48. Ritchie,M.E., Phipson,B., Wu,D., Hu,Y., Law,C.W., Shi,W. and Res., 334, 146–159. Smyth,G.K. (2015) limma powers differential expression analyses for 54. Ishihama,Y., Rappsilber,J. and Mann,M. (2006) Modular stop and go RNA-sequencing and microarray studies. Nucleic Acids Res., 43, e47. extraction tips with stacked disks for parallel and multidimensional 49. Huang da,W., Sherman,B.T. and Lempicki,R.A. (2009) Systematic Peptide fractionation in proteomics. J. Proteome Res., 5, 988–994. and integrative analysis of large gene lists using DAVID 55. Guo,J. and Price,D.H. (2013) RNA polymerase II transcription bioinformatics resources. Nat. Protoc., 4, 44–57. elongation control. Chem. Rev., 113, 8583–8603. 50. Benjamini,Y. and Hochberg,Y. (1995) Controlling the false discovery 56. Otero,G., Fellows,J., Li,Y., de Bizemont,T., Dirac,A.M., rate: a practical and powerful approach to multiple testing. J. R. Stat. Gustafsson,C.M., Erdjument-Bromage,H., Tempst,P. and Soc. B Nr., 57, 289–300. Svejstrup,J.Q. (1999) Elongator, a multisubunit component of a novel 51. Kluge,M. and Friedel,C.C. (2018) Watchdog - a workflow RNA polymerase II holoenzyme for transcriptional elongation. Mol. management system for the distributed analysis of large-scale Cell, 3, 109–118. experimental data. BMC Bioinformatics., 19, 97. 57. Vizcaino,J.A., Csordas,A., del-Toro,N., Dianes,J.A., Griss,J., 52. Holzel,M., Rohrmoser,M., Schlee,M., Grimm,T., Harasim,T., Lavidas,I., Mayer,G., Perez-Riverol,Y., Reisinger,F., Ternent,T. et al. Malamoussi,A., Gruber-Eber,A., Kremmer,E., Hiddemann,W., (2016) 2016 update of the PRIDE database and its related tools. Bornkamm,G.W. et al. (2005) Mammalian WDR12 is a novel Nucleic Acids Res., 44, 447–456. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

MIR sequences recruit zinc finger protein ZNF768 to expressed genes

Download PDF
16 pages

Loading next page...
 
/lp/oxford-university-press/mir-sequences-recruit-zinc-finger-protein-znf768-to-expressed-genes-JnVLRZ50am

References (60)

Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of Nucleic Acids Research.
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/gky1148
Publisher site
See Article on Publisher Site

Abstract

Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 700–715 Nucleic Acids Research, 2019, Vol. 47, No. 2 Published online 22 November 2018 doi: 10.1093/nar/gky1148 MIR sequences recruit zinc finger protein ZNF768 to expressed genes 1 2 3 1 Michaela Rohrmoser , Michael Kluge , Yousra Yahia , Anita Gruber-Eber , Muhammad 3 4 5 5 6 Ahmad Maqbool , Ignasi Forne´ , Stefan Krebs ,HelmutBlum , Ann Katrin Greifenberg , 6 7,8 4 3 Matthias Geyer , Nicolas Descostes ,AxelImhof , Jean-Christophe Andrau , Caroline 2 1,* C. Friedel and Dirk Eick Department of Molecular Epigenetics, Helmholtz Center Munich and Center for Integrated Protein Science Munich (CIPSM), Marchioninistrasse 25, 81377 Munich, Germany, Institute for Informatics, Ludwig-Maximilians-Universitat Munchen, ¨ Amalienstrasse 17, 80333 Munich, Germany, Institut de Gen ´ etique ´ Moleculaire ´ de Montpellier (IGMM), Univ Montpellier, CNRS-UMR5535, Montpellier, France, Biomedical Center Munich, ZFP, Großhadener Strasse 9, 82152 Planegg-Martinsried, Germany, Laboratory for Functional Genome Analysis (LAFUGA) at the Gene Center, Ludwig-Maximilians-Universitat ¨ Munchen, ¨ Feodor-Lynen-Strasse 25, 81377 Munich, Germany, Institute of Structural Biology, University of Bonn, Sigmund-Freud-Str. 25, 53127 Bonn, Germany, Department of Biochemistry and Molecular Pharmacology, New York University Langone School of Medicine, New York, NY 10016, USA and Howard Hughes Medical Institute, New York University Langone School of Medicine, New York, NY 10016, USA Received July 06, 2018; Revised October 25, 2018; Editorial Decision October 29, 2018; Accepted October 29, 2018 ABSTRACT terspersed sequences (SINE) (1,2). Mammalian-wide in- terspersed repeats (MIRs) are an ancient family of retro- Mammalian-wide interspersed repeats (MIRs) are transposed SINEs that spread genome-wide before and dur- retrotransposed elements of mammalian genomes. ing mammalian radiation (3,4). MIRs are ∼240 bp long Here, we report the specific binding of zinc finger pro- and consist of tRNA-derived sequences, a 70 bp MIR- tein ZNF768 to the sequence motif GCTGTGTG (N ) 20 specific core region, and sequences similar to the 3 ends CCTCTCTG in the core region of MIRs. ZNF768 bind- of LINEs. MIRs are enriched at gene loci in euchromatin, harbor putative transcription-factor binding sites, provide ing is preferentially associated with euchromatin and insulator and enhancer function (5–8), encode microRNAs, promoter regions of genes. Binding was observed are transcribed by RNA polymerase III (9,10), are associ- for genes expressed in a cell type-specific manner ated with tissue-specific gene expression ( 5,11), and some- in human B cell line Raji and osteosarcoma U2OS times provide splicing signals and contribute to exonization cells. Mass spectrometric analysis revealed binding (12). MIRs constitute 5–16% of the genome in marsupials of ZNF768 to Elongator components Elp1, Elp2 and and monotremes and 0.5-3% in placentalia (13). Like other Elp3 and other nuclear factors. The N-terminus of transposable elements, MIRs have shaped gene regulatory ZNF768 contains a heptad repeat array structurally networks in vertebrates (14–17), but our understanding how related to the C-terminal domain (CTD) of RNA poly- MIRs regulate gene activity is still elusive. merase II. This array evolved in placental animals Similarly to MIRs, the family of zinc finger pro- but not marsupials and monotreme species, displays teins (ZNFs) strongly expanded in mammals (18,19). species-specific length variations, and possibly ful- Widespread binding of ZNFs to regulatory regions indi- cates that mammalian genomes contain an extensive ZNF fills CTD related functions in gene regulation. We pro- regulatory network that targets a diverse range of genes and pose that the evolution of MIRs and ZNF768 has ex- pathways (20,21). Zinc finger protein 768 (ZNF768) evolved tended the repertoire of gene regulatory mechanisms in mammals and is defined by a domain of ten zinc fingers in mammals and that ZNF768 binding is associated with >96% (Figure 1) identity in placentals and marsupials, with cell type-specific gene expression. but is less conserved in monotremes (Supplementary Fig- ure S1). Placentalia additionally evolved an array of 10–20 INTRODUCTION heptad repeats in the amino-terminus of ZNF768, which is absent in marsupials and monotremes. This array has a Approximately half of mammalian genomes is of repeti- striking similarity to the carboxy-terminal domain (CTD) tive nature and composed of long (LINE) and short in- To whom correspondence should be addressed. Tel: +49 89 3187 1512; Email: [email protected] C The Author(s) 2018. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected] Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 701 Figure 1. Domain structure of ZNF768 in placentalia and marsupials and comparison with the CTD of RNA polymerase II. (A) Human ZNF768 is composed of domains box A (red box) and box B (green box) at the N-terminus interrupted by an array of 15 heptad repeats (yellow box) and a domain of 10 zinc fingers at the C-terminus (blue box). ( B) Mouse ZNF768 evolved an array of 19 heptad repeats. (C) ZNF768 of the marsupial Tammar Wallaby contains conserved A, B, and zinc finger domains, while the array of heptad repeats is absent. ( D) Number of heptad repeats in RNA polymerase II in vertebrates and ZNF768 in placentalia (see also Supplementary Figure S1). of the large subunit (Rpb1) of RNA polymerase II (Pol II), 3 processing of the nascent transcript. Interestingly, CTD which is composed of 52 heptad repeats with the consensus can function as transcriptional activator after fusion to a sequence Y S P T S P S . GAL4 DNA binding domain (27). Furthermore, transition 1 2 3 4 5 6 7 The CTD functions as a platform for recruitment and of Pol II through the transcription cycle is also observed if dissociation of cellular factors to the transcription machin- CTD is fused to other subunits of Pol II (28). Recent re- ery and is mainly regulated during the transcription cycle ports further provide evidence that CTD of Pol II can ag- by phosphorylation of heptad repeats by various kinases gregate reversibly alone, or with low complexity domains of (22–26). It is required for initiation, elongation, and termi- other transcription factors, like FUS, and that the ability for nation of transcription, but also for capping, splicing, and Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 702 Nucleic Acids Research, 2019, Vol. 47, No. 2 phase separation in liquid droplets is an important feature Immunofluorescence microscopy for the regulation of transcriptional activity (29–32). U2OS cells were seeded on a coverslip and grown for Due to the striking similarity of the heptad repeat array in 24 h. Cells were washed with PBS and fixed with 2% ZNF768 with the array of heptad repeats in CTD of Pol II paraformaldehyde (PFA) at RT for 5 min. After perme- we investigated if ZNF768 can act as a transcription factor abilization with 0.15% TritonX-100, samples were blocked and fulfill gene regulatory functions in cells. with 1% BSA and incubated with 7D6 or HA-specific mAbs over night at 4 C. Samples were washed with PBS for 5 min MATERIALS AND METHODS at RT, 0.15% Triton X-100 for 10 min at RT, blocked with 1% BSA for 7 min and incubated with Cy5-conjugated don- Tissue culture and recombinant gene expression key anti-rat immunoglobulin (Dianova) in the dark for 45 U2OS osteosarcoma cells were cultured in Dulbecco’s mod- min. Cells were washed again, stained with 4 ,6-diamidino- ified Eagle’s medium (DMEM, Gibco) and Raji B-cells in 2-phenylindole (DAPI) (Sigma) and mounted on slides us- RPMI 1640 medium (Gibco) supplemented with 10% fe- ing fluorescent mounting medium (Dako). Confocal mi- tal calf serum (FCS, Bio&Sell), 2 mM L-glutamine (Gibco), croscopy was performed on a Leica LSCM SP2 fluorescence 100 U/ml penicillin (Gibco), and 100 g/ml streptomycin microscope using the objective HCX PL APO 63× 1.4. Im- (Gibco) at 37 Cat8%or5%CO , respectively. Stably trans- 2 ages were processed using ImageJ 1.37 V and Fuji software fected U2OS cell lines were generated with the expression and the plug-in RGB profiler. Scale bars were calculated as vector pRTS-1 (33) using Polyfect (QIAGEN) followed by follows: hygromycin B (200 g/ml) selection. Conditional gene ex- B × 5 m/P (B = picture length in m, P = (512 pixel × pression was induced with 1 g/ml doxycycline. Recombi- voxel size) in m) nant ZNF768 and mutants are tagged C-terminally by a hemagglutinin (HA) tag and synthesized with an optimized siRNA transfection codon usage (Gene Art, Regensburg). Details for cloning in pRTS has been described elsewhere (34). All plasmids were siRNA transfection was performed according to the manu- confirmed by DNA sequencing prior to expression. facturer’s protocol using HS ZNF768 1 FlexiTube siRNA (Qiagen) and the HiPerFect Transfection Reagent (Qiagen) Monoclonal antibody with the exception that transfection was repeated after 24 h. Negative (non-silencing) siRNA (Qiagen) was used as con- The generation of monoclonal antibodies has been de- trol. scribed previously (34). The ZNF768-specific peptide RSPESDSQSPEFESQSPRYEPQSPGYEPRSPG (synthe- sized by PSL GmbH, Heidelberg) was coupled to ovalbu- Cell proliferation assay min for immunization. The rat monoclonal antibody 7D6 Cell proliferation was determined using the Real-time used in this study (IgG2c) specifically recognizes human xCELLigence System (Roche). U2OS cells were seeded at a ZNF768. density of 3.000 cells per 100 l in equilibrated 96-well mi- crotiter xCELLigence assay plates (E-plates). Conditional Immunoprecipitation (IP) and SDS-PAGE gene expression was induced with 1 g/ml doxycycline at the indicated time points. Alternatively, siRNA transfection Cells were washed twice with cold phosphate-buffered was performed according to the manufacturer’s protocol. saline (PBS) and lysis was performed in 100 llysis buffer per 2 × 10 cells (50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% NP-40 (Roche), 1x PhosSTOP (Roche), 1× protease in- Purification of ZNF768 hibitor cocktail (Roche)) at 4 C for 30 min, followed by son- ication on ice using a BRANSON Sonifier 250 (15 s on, 15 s Expression plasmids of human ZNF768 (UniProt acces- off, 50% duty) and centrifugation at 16 400 rpm (FA-45-24- sion number Q9H5H4) were cloned from a synthetic gene 11 rotor) for 10 min at 4 C. Immunoprecipitation was per- that was codon optimized for expression in Escherichia coli formed using Dynabeads Protein A und G (1:1) (Invitro- cells (Gene Art, Regensburg). A full length ZNF768 (1– gen). Lysates were incubated with antibody-coupled beads 540) construct and a construct consisting of the N-terminal (2.5 g of antibodies for 4 h at 4 C, followed by three washes heptad-repeats only (1–197) were cloned by PCR with re- with 1 ml lysis buffer) overnight. Beads were washed three strictions sites NcoI/EcoRI and ligated into a pGEX-4T1 times with 1 ml lysis buffer and boiled in laemmli buffer vector modified with a TEV protease cleavage site. All plas- (2% SDS, 10% glycerol, 60 mM Tris–HCl, pH 6.8, 10 mM mids were confirmed by DNA sequencing prior to expres- EDTA, 1 mM PMSF, 100 mM DTT, 0.01% bromophenol sion. blue) for SDS-PAGE. Whole cell lysates or IP samples were Plasmids were transformed into E. coli BL21(DE3) cells resolved by SDS-PAGE (10% or 15%) and transferred onto and induced at an OD of 0.6 to 1.0 with 0.3 mM IPTG nitrocellulose transfer membranes (GE Healthcare). The for 4 h growth. Cells were harvested in lysis buffer (20 membrane was blocked with 5% milk/TBS-T for 1 h. Incu- mM Tris–HCl pH 8.0, 500 mM NaCl, 10% glycerol, 1 mM bation with primary antibodies was performed over night DTE) and lysed by ultrasound. Fusion proteins were iso- at 4 C, followed by incubation with HRP-conjugated sec- lated with GSH Sepharose FastFlow (GE Healthcare) affin- ondary antibodies for 1 h and chemiluminescence detection ity chromatography methods. Cleavage of the GST-tag was with ECL (GE Healthcare). achieved by adding TEV protease in a 1/50 ratio and was Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 703 performed for 20 h at 4 C. Protein solution was concen- lauroylsarcosine, 1× protease inhibitor cocktail). Sonica- trated and loaded on a preparative HiLoad 16/60 Superdex tion was carried out in a Bioruptor Pico ultrasounds wa- 200 prep grade gel filtration column (GE Healthcare) for ter bath (Diagenode B01060001) for 30 cycles of 30 s ON full length ZNF768 or on a HiLoad 16/60 Superdex 75 and 30 s OFF pulses in 4 C water. Sonicated extracts were column for the truncated version of ZNF768 (1–197), re- centrifuged at high speed in the presence of 0.1% of Triton spectively, and equilibrated in gel filtration buffer (50 mM X-100 and snap frozen in liquid nitrogen and then stored at HEPES pH 7.5, 150 mM NaCl and 1 mM TCEP). Frac- −80 C until subsequent use. tions of the peak containing ZNF768 proteins determined Prior to ChIP, the ZNF768 mAb was coupled to protein- by SDS PAGE analysis were pooled and concentrated to 5 G coated magnetic beads (Dynabeads, life technologies) by mg/ml. The protein was aliquoted, snap frozen in liquid ni- incubation in 0.5% BSA PBS overnight at 4 C. Pre-coated trogen and stored at −80 C. beads were then washed and incubated with the sonicated chromatin extracts. ChIP was carried out overnight at 4 C on a rotating wheel. The equivalent of 10 × 10 cells soni- Gel shift assay cated extract was used for each ChIP experiment for both Gel shift assay was performed according to the manu- cell lines. After incubation, the beads were washed 7× with wash buffer (50 mM HEPES pH 7.6, 500 mM LiCl, 1 mM facturer’s protocol using the DIG Gel Shift Kit, second generation (Sigma-Aldrich). Briefly, oligonucleotides EDTA pH 8.0, 1% NP-40, 0.7% Na-Deoxycholate, 1× pro- were annealed to equimolar amounts of their com- tease inhibitor cocktail) followed by one wash with TE- plementary strands (M1: 5 - CAGTGCTGTGTGAC NaCl buffer (10 mM Tris pH 8.0, 1 mM EDTA pH 8.0, CTTGGGCAAGTCACTTAACCTCTCTGCAGT-3 , 50 mM NaCl). Immunoprecipitated chromatin was eluted M2: 5 - CAGTGCTGTGTGCAGTCAGTCAGT by two sequential incubations with 100 l elution buffer (50 CAGTCAGTCCTCTCTGCAGT-3 and M3: 5 - mM Tris pH 8.0, 10 mM EDTA pH 8.0, 1% SDS) at 65 C CAGTCAGTTGTGACCTTGGGCAAGTCACTT for 15 min. The two eluates were pooled and incubated at ◦ ◦ AACCTCCAGTCAGT-3 ) by heating to 95 C for 5 min 65 C for 12 h to reverse-crosslink of chromatin, followed and cooling slowly to room temperature. Double-stranded by treatment with RNase A (0.2 g/ml) at 37 C for 2 h and oligonucleotide probes were labelled at the 3 end using proteinase K (0.2 g/ml) at 55 C for 2 h. The DNA was DIG-11-dUTP and terminal transferase. Binding reactions isolated by phenol:chloroform:isoamylalcohol (25:24:1 pH were performed in 20 l volumes containing binding buffer 8.0) extraction followed by Qiaquick PCR Purification (Qi- [20 mM HEPES, pH 7.6, 1 mM EDTA, 10 mM (NH4) SO , agen, Germany) and quantified with Qubit DS DNA HS 2 4 5 mM DTT, 0.2% (w/v) Tween 20, 30 mM KCl], 50 ng/l Assay (ThermoFisher Scientific, USA). Poly [d(I-C)] and 5 ng/l Poly L-lysine at room temperature At least 1 ng of ChIP DNA was used to prepare sequenc- for 15 min. 0.6 ng of DIG-labelled DNA and extract of ing library with Illumina ChIP Sample Library Prep Kit 1.5–15 g purified ZNF768-WT or ZNF768 1–197 was (Illumina, USA) with a few optimizations to the protocol. used. For competition experiments, unlabeled competitor The ChIP DNA was size selected using Ampure beads (Life DNA was added in excess. Protein-DNA-complexes were technologies) to enrich for fragments <400 bp prior to end- separated by a native 6% (w/v) polyacrylamide 0.5× TBE repair, 3 end adenylation and adapter ligation. Library frag- gel, transferred onto a positively charged Nylon membrane ments were then directly amplified by 10 cycles of PCR. Bar- (GE Healthcare), fixed by Stratagene cross-linker and coded libraries from different samples were pooled together detected by chemiluminescent substrate CSPD (Roche). and sequenced on Illumina HiSeq2000 platform in paired- end sequencing runs. Chromatin immunoprecipitation for ChIP-seq RNA-seq libraries Cells were crosslinked using a formaldehyde containing so- lution (10 mM NaCl, 0.1 mM EDTA pH 8.0, 0.05 mM For preparation of total RNA cells (0.9 Mio/ml) were EGTA pH 8.0, 5 mM HEPES pH 7.8 and 1% formalde- harvested and resuspended in TRIzol reagent (Life Tech- hyde) for 10 min at 20 C, the reaction was quenched by the nologies) and snap-frozen in liquid nitrogen. After thaw- addition of glycin to a final concentration of 250 M for 5 ing RNA was extracted from 0.4 ml of TriZol lysate us- min. Crosslinked cells were collected and washed twice with ing the direct-zol RNA Miniprep (Zymo Research, Irvine PBS before snap freezing in liquid nitrogen and storage at CA, USA) as described in the manufacturer’s protocol. −80 C until subsequent use. RNA was assessed for purity by UV–vis spectrometry (Nan- Prior to sonication, the crosslinked cells were resus- odrop) and for integrity by Bioanalyzer (Agilent Bioana- pended in lysis buffer (50 mM HEPES pH 7.5, 140 mM lyzer 2100, Agilent, Santa Clara USA). RNA was of high NaCl, 1 mM EDTA pH 8.0, 10% glycerol, 0.75% NP-40, purity (abs. 260/280 > 1.9, abs. 269/239 > 2.1) and integrity 0.25% Triton X-100, 1× protease inhibitor cocktail) at 4 C (Bioanalyzer RIN > 9) and thus used for further process- for 20 min. Nuclei were collected by centrifugation and ing. For production of RNA-seq libraries total RNA was washed in a second buffer (200 mM NaCl, 1 mM EDTA DNAse treated (dsDNAse, Fermentas) and 100 ng of this pH 8.0, 0.5 mM EGTA pH 8.0, 10 mM Tris pH 8.0, 1× RNA was processed with a strand-specific protocol (RNA- protease inhibitor cocktail) for 10 min at 4 C then collected seq complete kit, NuGEN, San Carlos, USA). In brief the by centrifugation and resuspended in the shearing buffer (1 RNA was reverse transcribed to cDNA with a reduced mM EDTA pH 8.0, 0.5 mM EGTA pH 8.0, 10 mM Tris set of hexamer primers, avoiding excessive representation pH 8.0, 100 mM NaCl, 0.1% Na-Deoxycholate, 0.5% N- of rRNA in the cDNA. Second strand cDNA synthesis Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 704 Nucleic Acids Research, 2019, Vol. 47, No. 2 was done in presence of dUTP. After ultrasonic fragmen- Purification of ZNF768 for mass spectrometric analysis tation of the cDNA and end repair, Illumina-compatible For purification of ZNF768, Raji or U2OS cells (3 × 10 ) adapter were ligated. Adapters contained uracil in one were collected and IP was performed in 3 biological repli- strand, allowing complete digestion of the second-strand cates as described in the respective paragraph of immuno- derived DNA. After strand selection the libraries were am- precipitation. Simultaneously, ZNF768 antibody (7D6) plified, assessed for correct insert size on the Agilent Bio- and Pes1 (8E9) antibody, respectively, was coupled to analyser and diluted to 10 nM. Barcoded libraries were Sepharose A and G beads for 4 h at 4 C. ZNF768 anti- mixed in equimolar amounts and sequenced on an Illumina body (7D6) was used to identify the interactome of ZNF768 HiSeq1500 in single-read mode with a read length of 100 bp. whereas Pes1 (8E9) antibody served as a subclass control (52,53). Deep sequencing ChIP-seq and RNA-seq analysis was performed as previ- On beads digestion ously described (35). Four biological replicates of U2OS After the last washing step with lysis buffer, beads were and Raji cells were used for RNA-seq library construction. washed three times by adding 100 lof50mMNH HCO . 4 3 For trypsin digest, beads were transferred to a clean tube ChIP-seq data processing and incubated with 100 lof10ng/l trypsin-solution in 1M urea and 50 mM NH HCO for 30 min at 25 C. Sam- Raw sequencing reads were aligned to the human genome 4 3 ples were centrifuged at 800 rpm and supernatant was trans- (hg38) using BWA (36). Sequence reads with an alignment ferred into a fresh tube. Beads were washed twice with 50 l score <30 for paired-end reads and <20 for single-end reads of 50 mM NH HCO . The supernatants were pooled into were discarded as well as all reads that aligned equally well 4 3 the corresponding tube and incubated overnight at 25 C to different positions in the genome. Peak calling was per- after addition of 1mM DTT. Iodoacetamide (IAA) 10 l formed using GEM (37) in GPS mode and with a q-value (5mg/ml) was added and incubated for 30 min in the dark cutoff of 0.01. Overlapping peaks (peak centers: ±100 bp) at 25 C. To quench the IAA, 1 lof1MDTT wasadded were merged both within samples and across all four sam- and samples were incubated for 10 min at 25 C, followed ples to obtain the final list of unique peaks. Motif discov- by addition of 2.5 l of trifluoroacetic acid (TFA) and de- ery was performed using MEME-ChIP (38) and sequence salting using 2× C18 Stagetips (54). Stagetips were washed logos of the binding motif and surrounding regions were three times with 20 l of 100% ACN (1000 rpm, 1 min) and created using weblogo (39). Annotation of peaks relative to three times by adding 20 l of 0.1% TFA (1800 rpm, 1 min). gene features was performed using the ChipSeeker package Subsequently, samples were added (800 rpm, 30 min) and in R (40). Gene annotations were taken from GENCODE washed 3 times with 20 l of 0.1% TFA, followed by elu- version 25 (41). Repeat annotations by RepeatMasker and tion into a clean tube by washing three times with 20 lof phyloP100 conservation scores (42) for hg38 were down- 80% ACN/25% TFA solution. Finally, samples were evap- loaded from the UCSC genome browser. Visualization of orated to dryness, resuspended in 20 l formic acid solution ZNF768 binding on the genome and corresponding peaks and stored at −20 C until LC–MS analysis. was performed using Gviz (43). For the analysis of binding frequencies of ZNF768 in promoter, UTR, exon and intron regions, the same number of 200 bp regions were randomly Protein quantification by liquid chromatography coupled to selected using BEDtools (44). tandem mass spectrometry (LC–MS/MS) Purified peptides (5 l) were automatically injected into in RNA-seq data processing an Ultimate 3000 RSLC HPLC system (Dionex Thermo), Quality check of sequencing reads was performed using separated on an analytical column C18 micro column (75 FastQC (available at: http://www.bioinformatics.babraham. m i.d. × 15 cm, packed in-house with Reprosil Pur C18 ac.uk/projects/fastqc). Sequencing reads were mapped AQ 2.4 m, Doctor Maisch) using a 50-min gradient from against the human genome (hg38) and human rRNA se- 5 to 60% acetonitrile in 0.1% formic acid. The effluent quences using ContextMap version 2.7.9 (45) (using BWA from the HPLC was subsequently electrosprayed into a as short read aligner and default parameters). Number of LTQ Orbitrap XL mass spectrometer (Thermo). The MS read fragments per gene were determined from mapped instrument was operated in a data dependent mode to RNA-seq reads in a strand-specific manner using feature- automatically switch between full scan MS and MS/MS Counts (46) and GENCODE version 25 gene annotations. acquisition. Survey full scan MS spectra (from m/z 300 to RPKM values were calculated using edgeR and averaged 1800) were acquired in the Orbitrap with a resolution of R between replicates (47). Differential gene expression anal- = 60 000 at m/z 400 (after accumulation to a ‘target value’ ysis was performed using limma (48). Functional enrich- of 500,000 in the linear ion trap). The six most intense ment analysis for UniProt keywords and Gene Ontology peptide ions with charge state between 2 and 4 were sequen- terms was performed with the DAVID webserver (49). Sig- tially isolated to a target value of 10,000 and fragmented in nificantly enriched terms were determined using a cutoff of the linear ion trap by collision induced dissociation (CID). 0.05 on the P-value adjusted for multiple testing using the For all measurements with the Orbitrap mass analyzer, method by Benjamini and Hochberg (50). Analysis work- three lock-mass ions from ambient air (m/z = 371.10123, flows were implemented and run using the workflow man- 445.12002, 519.13882) were used for internal. Usual MS agement system Watchdog (51). conditions were: spray voltage, 1.5 kV; no sheath and Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 705 auxiliary gas flow; heated capillary temperature, 200 C; U2OS cells (Figure 2A and Supplementary Figure S3A). normalized collision energy 35% for CID in LTQ. The Expression of the zinc finger-containing C-terminal domain threshold for ion selection was 10 000 counts for MS2. of ZNF768 (Figure 2E) caused a similar staining pattern, The used activation was 0.25 and activation time 30 ms. while expression of the N-terminus containing the array of MaxQuant 1.5.2.8 was used to identify proteins and quan- heptad repeats resulted in a more diffuse staining of the nu- tify by iBAQ with the following parameters: Database, cleus (Supplementary Figure S3B), suggesting that the zinc Uniprot Hsapiens 3AUP000005640 170526; MS tol, finger domain is responsible for the association of ZNF768 10ppm; MS/MS tol, 0.5 Da; Peptide FDR, 0.1; protein with euchromatin. Antibody 7D6 immunoprecipitated en- FDR, 0.01 Min. peptide length, 5; variable modifications, dogenous ZNF768 protein quantitatively from extracts of oxidation (M); fixed modifications, carbamidomethyl (C); osteosarcoma cell line U2OS and B-lymphoid cell line Raji peptides for protein quantitation, razor and unique; min. (Figure 2B), proving its high specificity and suitability to peptides, 1; min. ratio count, 2. Identified proteins were study the binding of ZNF768 to DNA in chromatin im- considered as interaction partners if their MaxQuant munoprecipitation (ChIP) experiments. Knockdown exper- iBAQ values displayed a greater value than log 5-fold iments of ZNF768 confirmed the specificity of mAb 7D6 enrichment (FC) and P-value 0.05 (t-test adjusted for (Figure 2C) and showed further that ZNF768 is required multiple comparisons) when compared to the control. for viability and proliferation of U2OS cells (Figure 2D). In line with its essential function, expression of mutants with deletions of either the N- or C-terminal domain of ZNF768 RESULTS have a dominant-negative phenotype and inhibit cell pro- ZNF768 domain structure and conservation liferation (Figure 2E–G). Finally, ZNF768 is a phospho- protein and can be phosphorylated at almost all heptad re- An array of ten zinc fingers at the C-terminus of ZNF768 peat serine-5 residues (www.cellsignal.com, Supplementary shows high conservation in placentalia and marsupials Figure S2A). Treatment of cellular extracts of U2OS cells (>96%) but is less conserved in monotremes (blue boxes in with alkaline phosphatase causes a shift of the hyperphos- Figure 1 and Supplementary Figure S1). At the N-terminus, phorylated form of ZNF768 (Supplementary Figure S2C) two sequence blocks (box A and box B, red and green and reveals that a large fraction of ZNF768 is hyperphos- boxes in Figure 1) are conserved in placentalia and marsu- phorylated in U2OS cells. pials, but replaced by unrelated sequences in monotremes. In addition, ZNF768 of placentalia has evolved an array of heptad repeats that is positioned between box A and box B. The number of repeats varies between 20 repeats Identification of the ZNF768 binding motif in cellular DNA in mouse lemur and 10 repeats in pika and malayan pan- To investigate if ZNF768 can bind to specific DNA se- golin. Mouse ZNF768 contains 19 repeats, chimpanzee 16 quences, we performed ChIP experiments with mAb 7D6 and human 15 repeats (Supplementary Figure S1). Similar using extracts of U2OS and Raji cells. DNA libraries of to heptad repeats in CTD of Pol II, the heptad repeats in two biological replicates were prepared for each cell line and ZNF768 show no length variation (except a single extended analyzed by next generation sequencing. Peak calling iden- repeat in pika). However, the composition of amino acids tified a total of 21 012 unique peaks and 13.1% of these in the heptad repeats shows higher variation in ZNF768, peaks were consistently identified in all four samples and both, within and between species (Supplementary Figures an additional 28.8% at least in both replicates for the same S1 and S2B). Serine-5 and proline-6 residues show the high- cell type (Supplementary Figure S4). Generally, ZNF768 est conservation between ZNF768 and Pol II, followed by binding sites distributed over all chromosomes (Supplemen- the residues corresponding to tyrosine-1 and proline-3 in tary Figure S5). Motif discovery identified several potential the CTD, the position of threonine-4 and serine-7 show binding motifs for ZNF768 (Supplementary Figure S6A). only little or almost no conservation. The position corre- The top two identified motifs were found in 46% and 37% sponding to serine-2 in the CTD is particularly remarkable of peaks, respectively, and for both motifs the other motif in ZNF768. It is replaced in almost all repeats by an acidic was often found as a secondary motif at a distance of ∼20 amino acid (mostly glutamic acid). The phosphorylation of bp. We thus hypothesized that the ZNF768 binding motif serine-2 residues in CTD by P-TEFb is a hallmark in RNA consists of anchor regions connected by a linker region of elongation control (55) and a replacement of serine-2 may ∼20 bp. In fact, 58.1% of identified peaks contained this mimic its phoyphorylation. It is thus tempting to speculate consensus motif with at most three mismatches in the an- that the array of heptad repeats in ZNF768 potentially can chor regions and a linker region of 20 ± 3bp(Figure 3A, mimic a CTD phosphorylated at serine-2 residues. Supplementary Figures S4 and S6B). For peaks identified in all 4 samples, this number was as high as 98.3% and the ZNF768 is associated with euchromatin and required for vast majority of peaks with motif hits (83.5%) had a linker growth and cell viability length of 20 bp (Supplementary Figure S6B). Gelshift ex- To study the cellular function of ZNF768 we first raised a periments with recombinant ZNF768 protein confirmed the monoclonal antibody (7D6) towards human ZNF768 us- motif, GCTGTGTG (N ) CCTCTCTG, and revealed that ing a peptide containing heptad repeats 8–12 as epitope the nucleotide sequence of the spacer between the two an- (Figure 1A, materials and methods). This antibody prefer- chor regions is likely not critical for binding (Supplemen- entially stains euchromatic regions in the nucleus of fixed tary Figure S7). Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 706 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 2. ZNF768 is associated with euchromatin and essential for cell viability and proliferation. (A) Confocal image of U2OS cells stained with DAPI and with the ZNF768-specific mAb 7D6; merge of images on the right hand site. White line marks the area of the RGB profiler, profiles of DAPI and ZNF768 at the bottom. (B) mAb 7D6 immunoprecipitates a 70 kD protein from extracts of Raji and U2OS cells. SN: supernatant, *: Ig heavy chain. (C)siRNA mediated knockdown of ZNF768 in U2OS cells. (D) Growth kinetics of U2OS cells after knockdown of ZNF768 measured by xCelligence (Roche). Arrows indicate consecutive addition of siRNA. (E) Expression constructs of HA-tagged ZNF768 wild-type and deletion mutants and (F) expression control in U2OS cells. (G) Growth kinetics of U2OS cells after expression of ZNF768-WT and ZNF768 mutants measured by xCelligence (Roche). Arrows indicate addition of doxycycline. ZNF768 binds to MIR sequences tained in a MIR sequence and the consensus sequence for all MIR types actually contains the ZNF768 binding motif A systematic comparison of ZNF768 binding sites to re- (Supplementary Figure S8A). Despite this fact, only a small peats in the human genome showed an enrichment of fraction (12.2%) of MIRs in the human genome contains binding sites within all four types of MIRs (Figure 3B). the ZNF768 binding motif, which is not surprising giving 12,488/21,012 peaks overlapped with MIRs. Furthermore, a per-base identity <80% for human MIR sequences. Al- almost all peaks (92%) with the binding motif were con- though only 15.8% of MIRs containing the binding mo- Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 707 Figure 3. DNA binding motif and genomic binding of ZNF768 to MIRs. (A) Consensus ZNF768 binding motif identified by ChIP-seq experiments in Raji and U2OS cells and determined from peaks containing the motif (±3 mismatches in the anchors and linker length of 20 ± 3 bp). (B) Number of ZNF768 binding sites overlapping with particular type of repetitive sequences (top; in case of multiple overlaps the largest overlap is used) compared to the genomic length covered by the corresponding type of repetitive sequence (bottom). The right-most bar shows the number of ZNF768 binding sites with no overlap to repetitive sequences. tif were found to be bound by ZNF768, this fraction in- bound MIR sequences showed no particular conservation creased to 54.2% when considering MIRs with a more strin- (Figure 4B) indicating that ZNF768 binding represents a gent and strict version of the motif (linker length: 19/20bp, conserved function of a subset of MIR sequences in mam- 1 mismatch in anchors). Thus, most MIRs diverged so far mals. from the consensus that the binding motif was lost and no general conservation of the binding motif in human MIRs ZNF768 binding is associated with transcribed genes was observed (Supplementary Figure S8B). This divergence also allowed reliably aligning reads to MIR sequences de- We next asked if ZNF768 binding sites were enriched in reg- spite their repetitive origin. Only reads that could be aligned ulatory elements of genes. We found a strong enrichment of uniquely to the genome were used for peak calling. Al- ZNF768 binding in promoters and a slight enrichment of though 8524 detected peaks (40.6%) were not within MIRs, binding in exons and introns, while the binding frequency in 13.5% of these peaks contained the binding motif. The re- intergenic sequences was reduced (Figure 5A). Interestingly, maining peaks may contain a weaker version of the binding the 1061 ZNF768 binding sites outside of MIRs that con- motif, recruit ZNF768 to chromatin by other mechanisms tained the binding motif showed an even higher enrichment (e.g. looping), or represent spurious binding. at promoters. To investigate whether genes with ZNF768 Interestingly, MIRs with ZNF768 binding show a clear binding tended to be more highly expressed, we analyzed conservation of the two anchor motifs in the human RNA-seq data of four replicates of total RNA of Raji and genome. Sequences of the linker in the binding motif and U2OS cells (Supplementary Table S1). In both cell lines, outside of the binding motif were not particularly con- protein-coding genes with ZNF768 binding in the promoter served, similar to MIRs without binding of ZNF768 (Sup- or 5 UTR were more highly expressed on average than the plementary Figure S8B). We further investigated whether remaining protein-coding genes (Figure 5B). In contrast, ZNF768 binding sites in MIRs were also conserved across binding in intronic regions showed only a small but signifi- species by analyzing phyloP100 conservation scores deter- cant effect (Figure 5B). This provides evidence that ZNF768 mined from a multiple alignment of 99 vertebrate genomes regulates transcription by binding in or near promoter re- against the human genome. Positive PhyloP scores indicate gions of active genes. slower than expected evolution. The analysis of phyloP100 scores within and around the binding motif (±25 bp) in MIR sequences bound by ZNF768 showed increased con- ZNF768 binds to genes with cell type-specific expression servation for most positions within the anchor regions (Fig- Raji and U2OS cells revealed common and cell type-specific ure 4A). Sequences outside of the binding motif or within binding sites of ZNF768. Common sites were for instance the linker region, however, were mostly not conserved. Un- associated with genes for RNA polymerase II subunit E Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 708 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 4. Conservation of the ZNF768 motif (+25 bp on either side) in MIRs either (A) bound or (B) not bound by ZNF768. Only MIRs were considered that align without gaps to the MIR consensus sequence in the region of the ZNF768 binding motif +25 bp on either side. Distribution of PhyloP100 scores are indicated by boxplots for each position (green = median PhyloP100 score > 0.2) and regions within the motif region are indicated in more intensive colors. (POLR2E) (Figure 6A) or Solute Carrier Family 1 Mem- the fold-changes in gene expression were lower. This may be ber 5 (SLC1A5) and Mitochondrial Ribosomal Protein S5 due to the higher number of peaks identified in Raji cells, in- (MRPS5) (Supplementary Figure S9A,B). These genes are dicating a higher sensitivity but lower specificity compared expressed in Raji and U2OS cells and show similar peaks to peaks in U2OS cells. Thus, a significant fraction of seem- in both cell lines. Thus, many of the 2747 identified com- ingly Raji-specific peaks may simply have been missed in mon binding sites in Raji and U2OS may be associated with U2OS. We conclude that binding of ZNF768 occurs prefer- commonly expressed genes. We also identified a large num- entially at expressed genes and at least in part in a cell type- ber of peaks that were present either in Raji or U2OS cells. specific manner. The underlying mechanisms regulating the In U2OS cells, strong peaks for ZNF768 were associated cell type-specific binding of ZNF768 in Raji and U2OS cells with the promoter region of the GAS2L1 gene (Figure 6B), are currently unclear, but may involve, e.g. DNA methyla- the ID1 and SNPH genes (Supplementary Figure S9C,D) tion or other epigenetic marks. and the gene body of the ANXA2 and ALDH7A1 genes (Supplementary Figure S9E, F), but were absent or only faintly detectable in Raji cells. These genes are expressed ZNF768-regulated genes in U2OS cells in U2OS but not in Raji cells. Inversely, Raji cells showed a strong ZNF768 binding site in the promoter region of To study the gene regulatory potential of ZNF768, the B-Lymphocyte Surface Antigene (CD19) gene (Figure we induced expression of the dominant-negative mutant 6C), which is a B cell-specific non-receptor tyrosine kinase ZNF768-N(Figure 2E) in U2OS cells and analyzed required for B cell receptor signaling. Strong Raji-specific changes in the transcriptome after 12 h. A >2-fold change peaks were further detected for the genes CD86, ATP2A3, in RNA levels was detected for 500 downregulated and 155 RHOH, PLCG2, LYN, and ARHGDIB (Supplementary upregulated genes (Supplementary Table S2). Functional Figure S9G–L). These genes are expressed in Raji but not enrichment analysis of repressed genes revealed several sig- U2OS cells. nificantly enriched gene sets including two gene sets con- A global analysis of differential gene expression between taining DNA binding proteins (105 genes) and zinc finger U2OS and Raji cells showed significant differences in fold- proteins (103 genes) (Figure 7A and Supplementary Table changes for genes with peaks specific to either cell line (Fig- S3). Repressed genes in both gene sets show a large over- ure 6D and Supplementary Table S1). In particular, genes lap (63 genes) with repressed transcription-associated genes with U2OS-specific peaks were on average 30-fold higher (Figure 7B). We conclude that ZNF768 can act as transcrip- expressed in U2OS cells. For genes with Raji-specific peaks, tional regulator and is required particularly for the expres- sion of other transcription factors. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 709 Figure 5. Genomic distribution of the ZNF768 binding motif in Raji and U2OS cells. (A) Frequency of ZNF768 binding sites in promoters (–1 kb to transcription start site) and other genomic regions compared to randomly selected binding sites with the same peak length distribution. This shows an enrichment of genomic binding of ZNF768 in promoters, in particular for motif-containing peaks outside of MIRs. (B) Boxplots illustrating the distribution of expression levels in total RNA (quantified as RPKM = reads per kilobase per Million mapped reads) in Raji or U2OS cells for genes without (wo) and with (w) peaks in the respective cells. A pseudocount of 1 was added to all RPKM values before plotting. P-values for a Wilcoxon rank sum test comparing −3 −5 −10 RPKM levels between the two groups are indicated as: *P < 10 ,**P < 10 , ***P < 10 Mass spectrometric analysis of ZNF768 associated factors (Elp1, Elp2 and Elp3), SR rich splicing factor (SUGP2), centromere protein E (CENPE), several E3 ligases (USP13, We used the mAb 7D6 for a combined immunoprecipitation Trim33, and HERC2), proteins with centrosomal functions (IP) and mass spectrometric (MS) assay to identify ZNF768 (CEP170-1, Cep170-2 and NIN), and other factors. The associated factors. The ZNF768 interactome of Raji and binding of Elongator subunit Elp3 to ZNF768 was con- U2OS cells showed a large overlap and twenty of the best firmed in IP experiments with an Elp3-specific antibody thirty interactors were found in both cell lines (Figure 8 (Figure 8C). mAb 7D6 could immunoprecipitate a signifi- A,B, Supplementary Figure S10). Among the common fac- cant fraction of Elp3 protein of cellular extracts of Raji cells. tors we identified three subunits of the Elongator complex Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 710 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 6. Common and cell type-specific peaks of ZNF768 in Raji and U2OS cells. ChIP-seq (replicates shown separately) and RNA-seq (mean of four replicates) read coverage (in counts per million) for example genes. Identified peaks are shown as rectangles below the corresponding ChIP-seq sampl e. Genomic coordinates and gene annotation (boxes = exons, lines = introns, strand indicated by arrowheads) are shown in the bottom row. (A) ZNF768 binding in the promoter upstream region of the RNA Polymerase II Subunit E (POLR2E) gene, which is expressed in both Raji and U2OS cells. (B) Binding of ZNF768 in the promoter region of the Growth Arrest Specific 2-Like (GAS2L1) gene, which is expressed in U2OS but not Raji cells. ( C) Binding of ZNF768 to the promoter region of the B-Lymphocyte Surface Antigene (CD19) gene, which is expressed in Raji but not U2OS cells. (D) Genes in Raji and U2OS cells with cell-specific peaks differ in gene expression. Boxplots illustrate the distribution of fold-changes in gene expression between b oth cell lines (determined with limma) for genes with cell-specific peaks ( = peaks identified in gene body or 1 kb upstream in both replicates of the corresponding cell line but not for the other cell line; to account for the differences in sensitivity between Raji and U2OS ChIP-seq, for Raji only the 103 genes with the top-scoring Raji-specific peaks were evaluated, i.e. the same number of genes as with U2OS-specific peaks). Significance of the difference in median va lues −10 was determined using the Wilcoxon rank sum test (***P ≤ 10 ). The results suggests that ZNF768 can recruit Elongator and near promoters, suggesting that binding of ZNF768 is asso- other factors to expressed genes in Raji and U2OS cells. ciated with gene expression. In agreement with this assump- tion we observed ZNF768 binding preferentially in euchro- matic regions of the nucleus. Likewise, MIR sequences have DISCUSSION been reported to be associated with transcriptional active ZNF768 binds to MIR sequences euchromatin but not heterochromatin (5,6). Strikingly, the number of MIR sequences in mammals varies considerably ZNF768 proteins in mammals contain an array of ten zinc from about 20% of the total genome in monotremes to 1% fingers that allow the specific binding to DNA. ChIP-seq ex- or 3% of the genome in mice and humans, respectively. Fur- periments revealed approximately ten to twenty thousand thermore, the ZNF768 binding motif, although part of the ZNF768 binding sites in the genome of Raji and U2OS MIR consensus sequence, is not conserved in all MIRs, but cells. The majority of these sites is contained within MIR only in those displaying a peak in ZNF768 ChIP-seq exper- sequences and shares a common binding motif that is part iments. Notably, we also detected ∼1000 peaks containing of the MIR consensus sequence. The motif of the binding the ZNF768 binding motif outside of MIRs. This category site is 36 bp long and consists of two anchor sequences of 8 of peaks showed the highest association with promoters. bp separated by a linker of 20 bp, which probably does not Given the length of the detected DNA binding motif and contribute to the binding specificity of ZNF768 as revealed the position of the two anchor sequences at its flanks it is by gel shift experiments. ZNF768 binds preferentially at or Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 711 Figure 7. Functional enrichment analysis for UniProt keywords for genes downregulated upon expression of the dominant-negative mutant ZNF768-N in U2OS cells. (A) Significantly enriched UniProt keywords (identified with DAVID at an adjusted P-value < 0.05) for downregulated genes (>2-fold down-regulated, adjusted P-value <0.01, for full details see also Supplementary Table S3). (B) Venn diagram of downregulated genes annotated with the keywords Transcription, DNA-binding, and Zinc finger. The data indicates that inhibition of ZNF768 downregulates other transcription factors with zinc finger domains. likely that the proximal and distal, but not the central zinc Cell type-specific binding of ZNF768 to gene loci fingers, of the array of 10 zinc fingers contribute to DNA ChIP-seq analysis of ZNF768 revealed common but also a binding. The potential function of the central zinc fingers is large number of differential binding sites in Raji and U2OS currently unknown. We have currently no evidence for other cells. Furthermore, many putative binding sites containing conserved motifs upstream or downstream of the ZNF768 the binding motif were not occupied by ZNF768 in either binding motif. Notably, we also observed ZNF768 peaks at Raji or U2OS cells. This observation suggests that binding gene loci that do not contain the DNA binding motif. It is of ZNF768 to DNA is regulated and that not all binding currently unclear if binding to these loci requires the zinc sites are equally accessible in Raji and U2OS. The mecha- finger domain and /or other parts of the protein. nism(s) regulating the different accessibility is currently un- known but may include DNA methylation, histone compo- sition at binding motifs or specific histone marks. Addition- ZNF768 is an essential gene for cell proliferation ally, other cellular factors may block or permit binding of Knockdown experiments as well as the expression of ZNF768 to the binding motif. In this context it will be im- dominant-negative mutants revealed the functional require- portant to determine at which stage of cell differentiation ment of ZNF768 for cell viability and proliferation. Expres- the access of ZNF768 to its binding motif is regulated. sion of a mutated form of ZNF768 containing only the C- The observed differential binding of ZNF768 in Raji and terminal or N-terminal domain, respectively, led to a de- U2OS cells further prompted us to ask whether binding of cline of the cell index in cell proliferation assays. A decline of ZNF768 can mark differentially expressed genes in both this index was also seen after siRNA-mediated knockdown cell lines. In fact, we found a general correlation between of ZNF768 expression. This indicates that the functional ZNF768 binding and the activity of adjacent genes. In par- loss of ZNF768 cannot be compensated by other cellular ticular, we found a correlation between ZNF768 binding factors. Our results suggest that ZNF768, despite being an and gene expression for those genes that are active only in evolutionary young gene, gained essential function(s) for Raji or U2OS cells. From these data we conclude that bind- the expression of growth related genes. A detailed genetic ing of ZNF768 can mark commonly as well as cell type- analysis combined with mass spectrometry experiments will specifically expressed genes in Raji and U2OS cells. be required in the future to analyze the function of ZNF768 in the context of growth control in more detail. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 712 Nucleic Acids Research, 2019, Vol. 47, No. 2 Figure 8. ZNF768 interactome. ZNF768 was immunoprecipitated from cellular extracts of (A) Raji and (B) U2OS cells. Thirty interaction factors with the highest enrichment are shown. Common factors in both cell lines are depicted in red. The list of all interaction factors is shown in Supplementary Table S4. (C) ZNF768 mAb 7D6 specifically co-immunoprecipitates Elp3 from cellular extracts of Raji cells. ZNF768 functions as transcription factor ulated indirectly. The gene ontology analysis of repressed genes revealed several gene classes related to transcriptional Finally, we asked if binding of ZNF768 is required for ex- regulation suggesting that ZNF768 is hierarchically located pression of specific genes. To demonstrate this we stud- upstream of a network of transcription factor genes and ied the transcriptome of U2OS cells 12 h after overexpres- may function as a regulatory master gene for this network. sion of a ZNF768 mutant lacking the N-terminal domain. The notion that ZNF68 may act as a transcription factor We found several hundred genes that were significantly re- was further supported by mass spectrometric analysis of the pressed after expression of this dominant-negative mutant. ZNF768 interactome in Raji and U2OS cells. In both cell We also found a few induced genes, which may be upreg- Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 713 Figure 9. Known functions and regulation of Pol II CTD via its heptad repeat array and the implication with regards to possible functions and regulation of the heptad array in ZNF768 of placentalia (represented by human ZNF768). lines ZNF768 interacts with subunits Elp1, Elp2 and Elp3 and MIR sequences has contributed to speciation of pla- of the Elongator complex. The complex is conserved from centalia. yeast to mammals, consists of six subunits, Elp1-6, and has been proposed to function in the control of RNA elonga- DATA AVAILABILITY tion (56). The Elongator was found associated to the hy- perphoshorylated form of Pol II, but the mode of interac- GEO submissions: ChIP-Seq (GSE111879), RNA-seq Raji tion and the involved Elongator subunits are still elusive. cells (GSE111880), RNA-seq U2OS cells (GSE111881). Our data suggest that recruitment of Elongator to active The mass spectrometry proteomics data have been de- genes may also occur by ZNF768. ZNF768 binds first a posited to the ProteomeXchange Consortium via the subcomplex of Elongator consisting of Elp1-3 that subse- PRIDE (57) partner repository with the dataset identifier quently may assemble with subunits Elp4-6. In the future it PXD010831. will be interesting to study if heptad repeats of ZNF768 are involved in the recruitment of Elongator, as suggested for SUPPLEMENTARY DATA the CTD of Pol II, and if ZNF768 of marsupials lacks the Supplementary Data are available at NAR Online. ability of Elongator recruitment. Originally, the array of heptad repeats in ZNF768 at- tracted our attention to study the function of ZNF768 as ACKNOWLEDGEMENTS transcriptional activator due to its similarity to the array We thank Elisabeth Kremmer for help with the generation of heptad repeats in CTD of Pol II. This raises a couple of of ZNF768 mAb. intriguing questions. First, can this array fulfill similar or related functions as the array of heptad repeats in CTD? If so, can the acidic amino acids that are present at many po- FUNDING sitions in heptad repeats of ZNF768 mimic a hyperphos- D.E. and A.I. were supported by the Deutsche Forschungs- phorylated form of Pol ll? Such a mimicry is most likely gemeinschaft (DFG), SFB1064, Chromatin Dynamics and for position 2 of heptad repeats in ZNF768, which contains DFG excellence cluster CIPSM. In D.E. and J.C.A. labs, glutamic acid in almost all repeats across all species. It is the work was supported by a German-French BMBF- tempting to speculate that binding of ZNF768 can recruit ANR grant ‘EpiGlyco’. CNRS; ‘Agence Nationale de la cellular factors to genomic loci that otherwise are recruited Recherche’ (ANR); ‘amorcage jeunes equipes’ ´ Fondation only if serine-2 of CTD is phosphorylated, e.g. by Cdk9, or pour la Recherche Medicale FRM [AJE20130728183 other kinases (see model in Figure 9). In contrast, serine- to J.C.A.]; Deutsche Forschungsgemeinschaft (DFG) 5 residues are conserved between ZNF768 and the CTD [FR2938/7-1 and CRC 1123 (Z2) to C.C.F. and M.K.]; and may depend on phosphorylation in ZNF768, similar Deutsche Forschungsgemeinschaft (DFG) [GE 976/9-2 as in the CTD, to allow interaction with other factors. Fu- to M.G.] and is a member of the DFG excellence clus- ture work will address these and other questions and illu- ter ImmunoSensation. Funding for open access charge: minate if and how the new regulatory network of ZNF768 Helmholtz Center Munich. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 714 Nucleic Acids Research, 2019, Vol. 47, No. 2 Conflict of interest statement. None declared. 23. Eick,D. and Geyer,M. (2013) The RNA polymerase II carboxy-terminal domain (CTD) code. Chem. Rev., 113, 8456–8490. 24. Harlen,K.M. and Churchman,L.S. (2017) The code and beyond: transcription regulation by the RNA polymerase II carboxy-terminal REFERENCES domain. Nat. Rev. Mol. Cell Biol., 18, 263–273. 25. Jeronimo,C., Collin,P. and Robert,F. (2016) The RNA polymerase II 1. Kramerov,D.A. and Vassetzky,N.S. (2011) Origin and evolution of CTD: the increasing complexity of a low-complexity protein domain. SINEs in eukaryotic genomes. Heredity (Edinb.), 107, 487–495. J. Mol. Biol., 428, 2607–2622. 2. Redi,C.A. and Capanna,E. (2012) Genome size evolution: sizing 26. Zaborowska,J., Egloff,S. and Murphy,S. (2016) The pol II CTD: new mammalian genomes. Cytogenet Genome Res., 137, 97–112. twists in the tail. Nat. Struct. Mol. Biol., 23, 771–777. 3. Jurka,J., Zietkiewicz,E. and Labuda,D. (1995) Ubiquitous 27. Seipel,K., Georgiev,O., Gerber,H.P. and Schaffner,W. (1994) Basal mammalian-wide interspersed repeats (MIRs) are molecular fossils components of the transcription apparatus (RNA polymerase II, from the mesozoic era. Nucleic Acids Res., 23, 170–175. TATA-binding protein) contain activation domains: is the repetitive 4. Smit,A.F. and Riggs,A.D. (1995) MIRs are classic, tRNA-derived C-terminal domain (CTD) of RNA polymerase II a “portable SINEs that amplified before the mammalian radiation. Nucleic Acids enhancer domain”? Mol. Reprod. Dev., 39, 215–225. Res., 23, 98–102. 28. Suh,H., Hazelbaker,D.Z., Soares,L.M. and Buratowski,S. (2013) The 5. Jjingo,D., Conley,A.B., Wang,J., Marino-Ramirez,L., Lunyak,V.V. C-terminal domain of Rpb1 functions on other RNA polymerase II and Jordan,I.K. (2014) Mammalian-wide interspersed repeat subunits. Mol. Cell, 51, 850–858. (MIR)-derived enhancers and the regulation of human gene 29. Burke,K.A., Janke,A.M., Rhine,C.L. and Fawzi,N.L. (2015) expression. Mob. DNA, 5, 14. Residue-by-residue view of in vitro FUS granules that bind the 6. Jjingo,D., Huda,A., Gundapuneni,M., Marino-Ramirez,L. and C-terminal domain of RNA polymerase II. Mol. Cell, 60, 231–241. Jordan,I.K. (2011) Effect of the transposable element environment of 30. Hnisz,D., Shrinivas,K., Young,R.A., Chakraborty,A.K. and human genes on gene length and expression. Genome Biol. Evol., 3, Sharp,P.A. (2017) A phase separation model for transcriptional 259–271. control. Cell, 169, 13–23. 7. Smith,A.M., Sanchez,M.J., Follows,G.A., Kinston,S., 31. Kwon,I., Kato,M., Xiang,S., Wu,L., Theodoropoulos,P., Mirzaei,H., Donaldson,I.J., Green,A.R. and Gottgens,B. (2008) A novel mode of Han,T., Xie,S., Corden,J.L. and McKnight,S.L. (2013) enhancer evolution: the Tal1 stem cell enhancer recruited a MIR Phosphorylation-regulated binding of RNA polymerase II to fibrous element to specifically boost its activity. Genome Res., 18, 1422–1432. polymers of low-complexity domains. Cell, 155, 1049–1060. 8. Wang,J., Vicente-Garcia,C., Seruggia,D., Molto,E., 32. Schwartz,J.C., Ebmeier,C.C., Podell,E.R., Heimiller,J., Taatjes,D.J. Fernandez-Minan,A., Neto,A., Lee,E., Gomez-Skarmeta,J.L., and Cech,T.R. (2012) FUS binds the CTD of RNA polymerase II Montoliu,L., Lunyak,V.V. et al. (2015) MIR retrotransposon and regulates its phosphorylation at Ser2. Genes Dev., 26, 2690–2695. sequences provide insulators to the human genome. Proc. Natl. Acad. 33. Bornkamm,G.W., Berens,C., Kuklik-Roos,C., Bechet,J.M., Laux,G., Sci. U.S.A., 112, E4428–E4437. Bachl,J., Korndoerfer,M., Schlee,M., Holzel,M., Malamoussi,A. 9. Varshney,D., Vavrova-Anderson,J., Oler,A.J., Cowling,V.H., et al. (2005) Stringent doxycycline-dependent control of gene Cairns,B.R. and White,R.J. (2015) SINE transcription by RNA activities using an episomal one-vector system. Nucleic Acids Res., 33, polymerase III is suppressed by histone methylation but not by DNA e137. methylation. Nat. Commun.. 6, 6569. 34. Rohrmoser,M., Holzel,M., Grimm,T., Malamoussi,A., Harasim,T., 10. Yeganeh,M., Praz,V., Cousin,P. and Hernandez,N. (2017) Orban,M., Pfisterer,I., Gruber-Eber,A., Kremmer,E. and Eick,D. Transcriptional interference by RNA polymerase III affects (2007) Interdependence of Pes1, Bop1, and WDR12 controls expression of the Polr3e gene. Genes Dev., 31, 413–421. nucleolar localization and assembly of the PeBoW complex required 11. Carnevali,D., Conti,A., Pellegrini,M. and Dieci,G. (2017) for maturation of the 60S ribosomal subunit. Mol. Cell Biol., 27, Whole-genome expression analysis of mammalian-wide interspersed 3682–3694. repeat elements in human cell lines. DNA Res., 24, 59–69. 35. Shah,N., Maqbool,M.A., Yahia,Y., El Aabidine,A.Z., Esnault,C., 12. Krull,M., Petrusma,M., Makalowski,W., Brosius,J. and Schmitz,J. Forne,I., Decker,T.M., Martin,D., Schuller,R., Krebs,S. et al. (2018) (2007) Functional persistence of exonized mammalian-wide Tyrosine-1 of RNA polymerase II CTD controls global termination interspersed repeat elements (MIRs). Genome Res., 17, 1139–1145. of gene transcription in mammals. Mol. Cell, 69, 48–61. 13. Smit,A.F., Hubley,R. and Green,P. (2015). Repeatmasker Open 4.0. 36. Li,H. and Durbin,R. (2009) Fast and accurate short read alignment http://www.repeatmasker.org. with Burrows-Wheeler transform. Bioinformatics., 25, 1754–1760. 14. Cournac,A., Koszul,R. and Mozziconacci,J. (2016) The 3D folding of 37. Guo,Y., Mahony,S. and Gifford,D.K. (2012) High resolution genome metazoan genomes correlates with the association of similar wide binding event finding and motif discovery reveals transcription repetitive elements. Nucleic Acids Res., 44, 245–255. factor spatial binding constraints. PLoS Comput. Biol., 8, e1002638. 15. Feschotte,C. (2008) Transposable elements and the evolution of 38. Machanick,P. and Bailey,T.L. (2011) MEME-ChIP: motif analysis of regulatory networks. Nat. Rev. Genet., 9, 397–405. large DNA datasets. Bioinformatics, 27, 1696–1697. 16. Medstrand,P., van de Lagemaat,L.N. and Mager,D.L. (2002) 39. Crooks,G.E., Hon,G., Chandonia,J.M. and Brenner,S.E. (2004) Retroelement distributions in the human genome: variations WebLogo: a sequence logo generator. Genome Res., 14, 1188–1190. associated with age and proximity to genes. Genome Res., 12, 40. Yu,G., Wang,L.G. and He,Q.Y. (2015) ChIPseeker: an 1483–1495. R/Bioconductor package for ChIP peak annotation, comparison and 17. Testori,A., Caizzi,L., Cutrupi,S., Friard,O., De Bortoli,M., Cora,D. visualization. Bioinformatics, 31, 2382–2383. and Caselle,M. (2012) The role of Transposable Elements in shaping 41. Harrow,J., Frankish,A., Gonzalez,J.M., Tapanari,E., Diekhans,M., the combinatorial interaction of Transcription Factors. BMC Kokocinski,F., Aken,B.L., Barrell,D., Zadissa,A., Searle,S. et al. Genomics, 13, 400. (2012) GENCODE: the reference human genome annotation for The 18. Matthews,J.M. and Sunde,M. (2002) Zinc fingers–folds for many ENCODE Project. Genome Res., 22, 1760–1774. occasions. IUBMB Life, 54, 351–355. 42. Pollard,K.S., Hubisz,M.J., Rosenbloom,K.R. and Siepel,A. (2010) 19. Yang,P., Wang,Y. and Macfarlan,T.S. (2017) The role of Detection of nonneutral substitution rates on mammalian KRAB-ZFPs in transposable element repression and Mammalian phylogenies. Genome Res., 20, 110–121. evolution. Trends Genet.. 33, 871–881. 43. Hahne,F. and Ivanek,R. (2016) Visualizing genomic data using gviz 20. Imbeault,M., Helleboid,P.Y. and Trono,D. (2017) KRAB zinc-finger and bioconductor. Methods Mol. Biol., 1418, 335–351. proteins contribute to the evolution of gene regulatory networks. 44. Quinlan,A.R. and Hall,I.M. (2010) BEDTools: a flexible suite of Nature, 543, 550–554. utilities for comparing genomic features. Bioinformatics, 26, 841–842. 21. Najafabadi,H.S., Mnaimneh,S., Schmitges,F.W., Garton,M., 45. Bonfert,T., Kirner,E., Csaba,G., Zimmer,R. and Friedel,C.C. (2015) Lam,K.N., Yang,A., Albu,M., Weirauch,M.T., Radovani,E., ContextMap 2: fast and accurate context-based RNA-seq mapping. Kim,P.M. et al. (2015) C2H2 zinc finger proteins greatly expand the BMC Bioinformatics, 16, 122. human regulatory lexicon. Nat. Biotechnol., 33, 555–562. 22. Bentley,D.L. (2014) Coupling mRNA processing with transcription in time and space. Nat. Rev. Genet., 15, 163–175. Downloaded from https://academic.oup.com/nar/article-abstract/47/2/700/5198530 by Ed 'DeepDyve' Gillespie user on 06 February 2019 Nucleic Acids Research, 2019, Vol. 47, No. 2 715 46. Liao,Y., Smyth,G.K. and Shi,W. (2014) featureCounts: an efficient member of the Pes1-Bop1 complex and is required for ribosome general purpose program for assigning sequence reads to genomic biogenesis and cell proliferation. J. Cell Biol., 170, 367–378. features. Bioinformatics, 30, 923–930. 53. Kellner,M., Rohrmoser,M., Forne,I., Voss,K., Burger,K., Muhl,B., 47. Robinson,M.D., McCarthy,D.J. and Smyth,G.K. (2010) edgeR: a Gruber-Eber,A., Kremmer,E., Imhof,A. and Eick,D. (2015) Bioconductor package for differential expression analysis of digital DEAD-box helicase DDX27 regulates 3 end formation of ribosomal gene expression data. Bioinformatics, 26, 139–140. 47S RNA and stably associates with the PeBoW-complex. Exp. Cell 48. Ritchie,M.E., Phipson,B., Wu,D., Hu,Y., Law,C.W., Shi,W. and Res., 334, 146–159. Smyth,G.K. (2015) limma powers differential expression analyses for 54. Ishihama,Y., Rappsilber,J. and Mann,M. (2006) Modular stop and go RNA-sequencing and microarray studies. Nucleic Acids Res., 43, e47. extraction tips with stacked disks for parallel and multidimensional 49. Huang da,W., Sherman,B.T. and Lempicki,R.A. (2009) Systematic Peptide fractionation in proteomics. J. Proteome Res., 5, 988–994. and integrative analysis of large gene lists using DAVID 55. Guo,J. and Price,D.H. (2013) RNA polymerase II transcription bioinformatics resources. Nat. Protoc., 4, 44–57. elongation control. Chem. Rev., 113, 8583–8603. 50. Benjamini,Y. and Hochberg,Y. (1995) Controlling the false discovery 56. Otero,G., Fellows,J., Li,Y., de Bizemont,T., Dirac,A.M., rate: a practical and powerful approach to multiple testing. J. R. Stat. Gustafsson,C.M., Erdjument-Bromage,H., Tempst,P. and Soc. B Nr., 57, 289–300. Svejstrup,J.Q. (1999) Elongator, a multisubunit component of a novel 51. Kluge,M. and Friedel,C.C. (2018) Watchdog - a workflow RNA polymerase II holoenzyme for transcriptional elongation. Mol. management system for the distributed analysis of large-scale Cell, 3, 109–118. experimental data. BMC Bioinformatics., 19, 97. 57. Vizcaino,J.A., Csordas,A., del-Toro,N., Dianes,J.A., Griss,J., 52. Holzel,M., Rohrmoser,M., Schlee,M., Grimm,T., Harasim,T., Lavidas,I., Mayer,G., Perez-Riverol,Y., Reisinger,F., Ternent,T. et al. Malamoussi,A., Gruber-Eber,A., Kremmer,E., Hiddemann,W., (2016) 2016 update of the PRIDE database and its related tools. Bornkamm,G.W. et al. (2005) Mammalian WDR12 is a novel Nucleic Acids Res., 44, 447–456.

Journal

Nucleic Acids Research – Oxford University Press

Published: Jan 25, 2019

There are no references for this article.