5016–5037 Nucleic Acids Research, 2019, Vol. 47, No. 10 Published online 29 March 2019 doi: 10.1093/nar/gkz195 Chromatin regulation by Histone H4 acetylation at Lysine 16 during cell death and differentiation in the myeloid compartment Rocio G. Urdinguio1,2,† , Virginia Lopez1,† , Gustavo F. Bayón2 , Rafael Diaz de la Guardia3,4 , Marta I. Sierra2 , Estela Garcı́a-Toraño2 , Raúl F. Perez1,2 , Marı́a G. Garcı́a1,2 , Antonella Carella1,2 , Patricia C. Pruneda2 , Cristina Prieto2 , Marija Dmitrijeva2 , Pablo Santamarina1,2 , Thalı́a Belmonte1,2 , Cristina Mangas2 , Elena Diaconu2 , Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Cecilia Ferrero2 , Juan Ramón Tejedor2 , Juan Luis Fernandez-Morera2 , Cristina Bravo2 , Clara Bueno3,4 , Alejandra Sanjuan-Pla5 , Ramon M. Rodriguez6 , Beatriz Suarez-Alvarez6 , Carlos López-Larrea6 , Teresa Bernal7 , Enrique Colado7 , Milagros Balbı́n8 , Olivia Garcı́a-Suarez9 , Marı́a Dolores Chiara10 , Inés Sáenz-de-Santa-Marı́a10 , Francisco Rodrı́guez11 , Ana Pando-Sandoval12 , Luis Rodrigo13 , Laura Santos14 , Ana Salas15 , Jesús Vallejo-Dı́az16 , Ana C. Carrera16 , Daniel Rico 17 , Inmaculada Hernández-López17 , Amparo Vayá18 , José M. Ricart19 , Edward Seto20 , Núria Sima-Teruel21 , Alejandro Vaquero21 , Luis Valledor22 , Maria Jesus Cañal22 , David Pisano23 , Osvaldo Graña-Castro23 , Tim Thomas24 , Anne K. Voss24 , Pablo Menéndez3,4,25 , Ana Villar-Garea26 , Rainer Deutzmann26 , Agustı́n F. Fernandez 2,* and Mario F. Fraga1,* 1 Nanomaterials and Nanotechnology Research Center (CINN-CSIC), Universidad de Oviedo-Principado de Asturias, Spain, 2 Cancer Epigenetics Laboratory, Institute of Oncology of Asturias (IUOPA), ISPA-Hospital Universitario Central de Asturias HUCA, Universidad de Oviedo, Oviedo, Spain, 3 Josep Carreras Leukemia Research Institute and Department of Biomedicine, School of Medicine, University of Barcelona, Barcelona, Spain, 4 Centro de Investigación Biomédica en Red en Cáncer (CIBER-ONC), Barcelona, Spain, 5 Hematology Research Group, Instituto de Investigación Sanitaria La Fe (IIS La Fe), Valencia, 46026, Spain, 6 Translational Immunology Laboratory, Instituto de Investigación Sanitarias del Principado de Asturias (ISPA), Immunology Department, Hospital Universitario Central de Asturias (HUCA), Oviedo, Spain, 7 Servicio de Hematologı́a, Hospital Universitario Central de Asturias (HUCA), Oviedo, Spain, 8 Service of Molecular Oncology, Hospital Universitario Central de Asturias, Instituto Universitario de Oncologı́a del Principado de Asturias, Universidad de Oviedo, Oviedo, Spain, 9 Department of Morphology and Cellular Biology, Faculty of Medicine, University of Oviedo, Oviedo, Spain, 10 Otorhinolaryngology Service, Hospital Universitario Central de Asturias, Instituto Universitario de Oncologı́a del Principado de Asturias, Universidad de Oviedo, CIBERONC, Oviedo, Spain, 11 Departamento de Bioquı́mica y Biologı́a Molecular, Facultad de Medicina, Instituto Universitario de Oncologı́a (IUOPA), Universidad de Oviedo, Oviedo, Spain, 12 Hospital Universitario Central de Asturias (HUCA), Instituto Nacional de Silicosis (INS), Área del Pulmón, Facultad de Medicina, Universidad de Oviedo, Avenida Roma s/n, Oviedo, Asturias 33011, Spain, 13 Hospital Universitario Central de Asturias (HUCA), Gastroenterology Service, Facultad de Medicina, Universidad de Oviedo, Avenida de Roma s/n, Oviedo, Asturias 33011, Spain, 14 Fundación para la Investigación Biosanitaria de Asturias (FINBA). Instituto de Investigación Sanitaria * To whom correspondence should be addressed. Tel: +34 985652411; Fax: +34 985109905; Email: mffraga@cinn.es Correspondence may also be addressed to Agustin F. Fernandez. Email: affernandez@hca.es † The authors wish it to be known that, in their opinion, the first two authors should be regarded as Joint First Authors. Present address: Ana Villar-Garea, R&D, Sanofi-Aventis Deutschland GmbH, Industriepark Höchst Building G878, D-65926 Frankfurt am Main, Germany. C The Author(s) 2019. 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 journals.permissions@oup.com Nucleic Acids Research, 2019, Vol. 47, No. 10 5017 del Principado de Asturias (ISPA). Avenida de Roma s/n, 33011 Oviedo. Asturias. España, 15 Cytometry Service, Servicios Cientı́fico-Técnicos (SCTs). Universidad de Oviedo, Oviedo, Spain, 16 Department of Immunology and Oncology, National Center for Biotechnology, CNB-CSIC, Cantoblanco, 28049 Madrid, Spain, 17 Institute of Cellular Medicine, Newcastle University, UK, 18 Hemorheology and Haemostasis Unit, Service of Clinical Pathology, La Fe University Hospital, Valencia, Spain, 19 Dermatology Service, Quirón Clinic, Valencia, Spain, 20 George Washington University Cancer Center, Department of Biochemistry and Molecular Medicine, George Washington University, Washington, DC 20037, USA, 21 Chromatin Biology Laboratory, Cancer Epigenetics and Biology Program (PEBC), Bellvitge Biomedical Research Institute (IDIBELL), Av. Gran Via de l’Hospitalet, 199–203, 08907- L’Hospitalet de Llobregat, Barcelona, Spain, 22 Plant Physiology Lab, Department of Organisms and Systems Biology, Faculty of Biology, University of Oviedo, Oviedo, Asturias, Spain, 23 Bioinformatics Unit, Structural Biology and Biocomputing Program, Spanish National Cancer Research Center (CNIO), C/ Melchor Fernández Almagro, 3. 28029 Madrid, Spain, 24 The Walter and Eliza Hall Institute of Medical Research, Melbourne, Victoria, Australia; Department of Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Medical Biology, University of Melbourne, Melbourne, Victoria, Australia, 25 Instituciò Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain and 26 Institute of Biochemistry, Genetics and Microbiology, University of Regensburg, 93053 Regensburg, Germany Received June 29, 2018; Revised February 26, 2019; Editorial Decision March 11, 2019; Accepted March 15, 2019 ABSTRACT which are capable of giving rise to all blood cell lineages whereby multipotent progenitors transit multiple decision Histone H4 acetylation at Lysine 16 (H4K16ac) is a points while becoming increasingly lineage-restricted. Loss key epigenetic mark involved in gene regulation, DNA of such homeostasis underlies hematopoietic defects af- repair and chromatin remodeling, and though it is fecting erythro-myeloid and lymphoid lineages, resulting in known to be essential for embryonic development, its life-threating malignancies, infections, anemia and bleed- role during adult life is still poorly understood. Here ing. The activation and repression of specific gene pro- we show that this lysine is massively hyperacety- grams during hematopoiesis is highly regulated by epige- lated in peripheral neutrophils. Genome-wide map- netic mechanisms (2), PTMs playing a central role (4). One ping of H4K16ac in terminally differentiated blood of these histone PTMs, histone H4 acetylation at Lysine 16 cells, along with functional experiments, supported (H4K16ac), can regulate the expression of HOX9 (5,6), a a role for this histone post-translational modification gene involved in the hematopoietic commitment of embry- onic stem cells (7). In line with this, it has been proposed in the regulation of cell differentiation and apoptosis that the major H4K16 histone acetyltransferase KAT8 (also in the hematopoietic system. Furthermore, in neu- known as males absent on the first (MOF), or MYST1) trophils, H4K16ac was enriched at specific DNA re- (8) is central to the maintenance of hematopoietic stem peats. These DNA regions presented an accessible cell (HSC) target genes (9). The H4K16ac-specific NADH- chromatin conformation and were associated with dependent histone deacetylase Sirt1 also plays an important the cleavage sites that generate the 50 kb DNA frag- role in hematopoiesis (6,10), and, in addition, HDAC1 and ments during the first stages of programmed cell HDAC2, which can deacetylate H4K16ac, are highly regu- death. Our results thus suggest that H4K16ac plays lated during hematopoietic differentiation (11) and may be a dual role in myeloid cells as it not only regulates essential for HSC homeostasis (12). differentiation and apoptosis, but it also exhibits a Mice carrying null mutations of Kat8 revealed that non-canonical structural role in poising chromatin H4K16ac is necessary for normal development and the maintenance of normal distribution of euchromatin in for cleavage at an early stage of neutrophil cell death. mammals (13), although accompanying molecular events were not studied due to the very early death of the preim- INTRODUCTION plantation embryos. At a physicochemical level, acetylation Histone post-translational modifications (PTMs) play an of H4K16 was shown to regulate both higher order chro- important role in the regulation of chromatin structure and matin structure and functional interactions between non- have been associated with the regulation of gene expres- histone proteins and the chromatin fiber (14), which could sion (1–3). Consequently, histone PTMs are one of the cen- explain the biological role of this histone PTM in transcrip- tral epigenetic modifications involved in embryonic devel- tional activation (14–16). opment and cell differentiation, as well as in pathological In addition to the regulation of hematopoietic genes, conditions including cancer. H4K16ac has been associated with apoptosis and cell death, Hematopoiesis is a very dynamic spatial-temporal- and, specifically, to have an important role in the repression regulated process controlling the homeostasis of the of apoptotic genes in human cancer (17,18), which is in line hematopoietic system throughout the entire lifespan of with the capacity of HDAC inhibitors to induce apoptosis an individual. It is maintained by the activity of rare (19,20). Interestingly global levels of H4K16 acetylation are self-renewing, long-lived hematopoietic stem cells (HSCs) reduced in various types of tumor (21–24), and decreased 5018 Nucleic Acids Research, 2019, Vol. 47, No. 10 levels of KAT8 (23,25–29) and increased levels of HDACs gradient centrifugation using Ficoll-Hypaque from 20 um- (19,30–33) are often found in cancer cells. In accordance bilical cord samples collected, with Clinical Research Ethics with this, it has been demonstrated that H4K16 acetylation Committee approval, by Banc de Sang i Teixits (Barcelona). controls the sensitization of multidrug-resistant cancer cells After lysing red cells, the CD34+ cells were purified by mag- to topoisomerase inhibitors (34,35), supporting the notion netic bead separation using the human CD34 MicroBead that this histone modification might play a general role in kit (130-046-703, Miltenyi Biotec) and the AutoMACS Pro- apoptosis, although the underlying molecular mechanisms separator (Miltenyi Biotec) as per the manufacter’s instruc- involved remain elusive. tions and the cells collected for subsequent analyses. The role of H4K16ac in apoptosis might have implica- tions in the regulation of the outcome of autophagy since High-performance capillary electrophoresis (HPCE) quan- high levels of this histone PTM have been shown to promote tification of global histone acetylation cell death responses upon autophagy induction (36). In the same vein both KAT8 and H4K16ac have been shown to Histone acid extraction (sulfuric acid + acetone precipita- Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 be essential in the DNA damage response and for double- tion) and global histone acetylation was performed as pre- strand break repair (37), a role that may be mediated, at viously described (41,42) with slight modifications. In brief, least in part, by the modulation of the recruitment of DNA individual fractions of histones H4, H3, H2A and H2B damage repair proteins (38). were purified by reversed-phase high-performance liquid In this study, the initial discovery that human periph- chromatography (HPLC) on a Jupiter C18 column (Phe- eral blood (PB) neutrophils are massively hyperacetylated nomenex, Inc.) and eluted with an acetonitrile gradient (20– at H4K16 as compared to other blood cell types, including 60%) in 0.3% trifluoroacetic acid using an HPLC gradient CD34+ progenitors, CD19+ B cells, CD3+ T cells, CD15+ system (Beckman Coulter). Acetylated histone derivatives CD16− eosinophils and CD15+ CD16+ neutrophils, led us were resolved by high-performance capillary electrophore- to investigate in more detail, both in vitro and in vivo, the sis (HPCE) using an uncoated fused-silica capillary tubing role of this histone PTM in hematopoiesis. Our results in- (Beckman-Coulter; 60.2 cm × 75 ␮m, effective length 50 dicate that H4K16ac plays a complex role in this process, cm) in a capillary electrophoresis system (P/ACE MDQ, which has implications in terms of gene regulation and, Beckman-Coulter) with 100 mM phosphate buffer (pH 2.0) more importantly, for the regulation of neutrophil chro- 0.02% (w/v) HPM-cellulose as running buffer and an oper- matin structure and apoptosis. Neutrophil homeostasis is ating voltage of 12 kV. tightly regulated by a process known as neutrophil constitu- tive apoptosis, a type of programmed cell death that shares Western blotting many features of classical apoptosis (39,40). Although the underlying molecular mechanisms are still not fully under- Histones were separated on a 17% sodium dodecyl sulfate- stood, our results suggest that nuclear chromatin and, in polyacrylamide gel electrophoresis (SDS-PAGE) and trans- particular H4K16ac, may play an important role in the pro- ferred to a 45-␮m-pore polyvinylidene difluoride (PVDF) cess. membrane (Immobilon PSQ, Millipore). The membrane was blocked in 5% BSA/TBS-T (Tris-buffered saline with 0.1% Tween-20) and immunoprobed with antibodies to MATERIALS AND METHODS acetylation at lysine 16 of histone H4 (H4K16ac) (Milli- pore, 07–329), and using histone H3 (Abcam, ab1791) as Purification of peripheral white blood cells a loading control. The secondary antibody used was anti- Blood samples from 26 healthy donors and from two pa- rabbit IgG conjugated to horseradish peroxidase (Amer- tients with eosinophilia resulting from asthma exacerbation sham Biosciences, NA934) at 1:3000 dilution in TBS-T. Pri- were collected from the Hospital Universitario Central de mary antibodies were prepared for incubation in 5% BSA Asturias (HUCA). The study was approved by the Clini- in TBS-T while secondary antibody was blocked with 5% cal Research Ethics Committee of HUCA and all the indi- milk in TBS-T. Signals were detected with ECL detection viduals provided written informed consent. Mononuclear kit (Amersham Biosciences). Protein bands were quanti- cells (MNC) and granulocytes were separated from whole fied using ImageJ software (Rasband, W.S., ImageJ, U.S. blood by density gradient centrifugation using a standard National Institutes of Health, Bethesda, MD, USA, https: Ficoll protocol (Histopaque-1077, Sigma-Aldrich, 10771). //imagej.nih.gov/ij/, 1997–2018). After lysing red cells (BioLegend, 420301), we performed the isolation from the MNC fraction of CD3+ T cells, and Mass spectrometry assays CD15+ CD16+ neutrophils or CD15+ CD16− eosinophils from the granulocyte fraction (eosinophils from patients), Histone H4 preparation for mass spectrometry. Bands cor- using the corresponding selection of the AutoMACS sys- responding to histone H4 were excised from SDS-PAGE tem (130-050-101 for CD3, 130-046-601 for CD15, 130-092- gels and processed as previously described (43) with mi- 010 for CD16) (Miltenyi Biotec). Cells were collected for nor modifications. Briefly, after destaining the gel pieces subsequent analyses. Blood samples from 18 Pelger-Hüet with 0.1 M ammonium bicarbonate in 30% HPLC-grade individuals were collected at Hospital Universitario La Fe acetonitrile, samples were washed twice with 70% HPLC- (Valencia) with the corresponding informed consent. His- grade methanol and acylated at RT with propanoic anhy- tones were isolated from those samples for mass spectrom- dride. To this end, 15 ␮l 0.71 M propanoic anhydride in etry analysis. Another set of MNCs was isolated by density 70% methanol was added to each sample and, after 2.0 min Nucleic Acids Research, 2019, Vol. 47, No. 10 5019 incubation at RT, 40 ␮l 1M ammonium bicarbonate was by oral gavage at a dose of 6 mg tamoxifen/40 g body added to each crude. Reactions were kept at RT for a fur- weight for a total of 7 days. For hematological determi- ther 15 min, then the gel pieces were washed again with 70% nations, blood was extracted from the mandibular sinus of methanol and the acylation was repeated once more. Next, mice anesthetized with isoflurane. Blood samples were an- the gel pieces were washed sequentially with 0.1 M ammo- alyzed using Abacus junior vet equipment (Diatron Labs). nium bicarbonate, water, 50% acetonitrile, and then shrunk Details of the in vitro murine neutrophil differentiation with 100% acetonitrile. Proteins were then digested over assays and the subsequent analyses are available in Supple- night at 37◦ C with 0.26 ␮g trypsin (Trypsin Gold, Promega) mentary Materials and Methods. per sample, according the manufacturer’s instructions. Analysis of H4 acetylation by MALDI-TOF/TOF mass ChIP-seq spectrometry. Three 1 ␮l and three 0.5 ␮l aliquots of For high-throughput sequencing of H4K16ac-enriched ge- each crude were spotted onto a stainless steel MALDI tar- nomic regions (ChIP-seq), CD3+ T cells and neutrophils Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 get plate, allowed to dry, overlaid with 0.5 ␮l freshly pre- isolated from human peripheral blood of 5 donors (Sup- pared, ice-cold 5 mg/ml alpha-cyano-4-hydroxy cinnamic plementary Table S2) were freshly processed using the Sim- acid (ultra-pure, for mass spectrometry, Fluka) in 50% pleChIP Enzymatic Chromatin IP Kit (Magnetic Beads) acetonitrile/0.3% TFA, and then dried again. Spectra were (Cell Signaling, 9003). The amount of Microccocal Nu- recorded in a MALDI-TOF/TOF instrument (4800, Ap- clease required to digest DNA to the optimal length was plied Biosystems) operated according to the manufacturer’s determined as 2.5 ␮l per 2 × 107 cells. Library prepara- instructions. MS1 spectra were acquired in the reflector tion and sequencing were performed at Centre de Regulació modus, setting the m/z window between 700 and 2500 and Genòmica (CRG) in Barcelona on an Illumina HiSeq 2000 the focus on 1500, with 20 sub-spectra (50 shots each) being system in 50-single end mode. collected per spot. Tandem spectra (MS2) of the unmod- A detailed description of the bioinformatic analysis of ified and monoacetylated forms of the peptide 4–17 were Chip-seq data is available in Supplementary Materials and recorded with 1 kV collision energy without the use of col- Methods. lision gas (CID off), with the isolation window width set to 2.0 Da, and 55 sub-spectra (50 shots of each) were collected per spot. For all MS and MS/MS spectra, the laser intensity Cell culture and in vitro treatments of HL60 cell line was manually adjusted for a good S/N. MS1 spectra were processed with mMass 5.5.0 (44,45). The human HL-60 cell line was grown in Gibco RPMI Each spectrum was recalibrated by using the signals of pep- 1640 medium supplemented with 10% FBS (Sigma-Aldrich, tides 46–55 and 79–92 and deisotoped. Next, the relative F6178), 100 U/ml penicillin along with 100 ␮g/ml strep- proportion of a given form of the peptide 4–17 was calcu- tomycin (Gibco, 15070) and 2.5 ␮g/ml amphotericin B lated by dividing the absolute intensity of its monoisotopic (Gibco, 15290) at 37◦ C in a humidified atmosphere with 5% peak by the sum of the intensities of the monoisotopic peaks CO2 . corresponding to the unmodified, mono- and diacetylated Differentiation was induced by treatment with 1 ␮M peptide 4–17. Finally, the results for all the spectra acquired ATRA (Sigma-Aldrich, R2625) and/or 1.25% DMSO from each sample were averaged and minimum and maxi- (Sigma-Aldrich, D5879) for 72 h. Etoposide (Sigma- mum values were extracted. Aldrich, E1383) treatment was carried out at a concentra- tion of 5 and 10 ␮M and cells were collected after 1, 2, 3, 4 and 5 h for the lower or 1, 2 and 3 h for the higher con- Generation of a conditional Kat8-deficient mouse model centration. Treatment with the HDAC inhibitor SAHA was A targeting construct for the Kat8 locus was created carried out at 1 ␮M for 48 h. The KAT8 overexpression through a previously described recombineering-specific vector was generated as previously described (48). Opti- methodology (46). Briefly, loxP sites were introduced be- MEM (Gibco, 31985) and Lipofectamine 3000 transfec- tween exons 1 and 2, and between exons 5 and 6. This tion reagent (Invitrogen Corp., L3000015) were used for the resulted in a mouse strain in which bases 2373–8550 (in- transient transfection of plasmid DNA. shKAT8 lentiviral corporating exons 2–5) 3 from the start point of trans- expression vectors were obtained from OriGene (Rockville, lation, could be deleted leaving a single loxP site be- MD). The cells were collected 72 h after transfection. tween exons 1 and 6 (13). These mice were crossed with a Apoptosis related to Annexin V detection was deter- CreERT2 strain which allowed the removal of exons 2–5 in mined using the FITC Annexin V Apoptosis Detection a tamoxifen-dependent manner (47). Mice were maintained Kit with 7-AAD (BioLegend, 640922) or the antibody Dye on a C57BL/6 background, fed ad libitum, and housed un- 634 Annexin V (Immunostep, ANXVDY) depending on der a 12-h light-dark cycle. PCR genotyping was performed the experiment, using a Facs Aria IIu flow cytometer and using the oligonucleotides in Supplementary Table S1. The FACSDiva v8.02 software (Becton Dickinson) Cell cycle PCR products consisted of fragments of 422 bp (WT allele), phases and debris percentage were analyzed using propid- 509 bp (floxed allele) and 277 bp (excised allele). ium iodide (Sigma-Aldrich, P4170) as previously described Animal experiments were conducted in accordance with (49,50). Data were acquired using a Cytomics FC500 Flow the guidelines of the Committee for Animal Experimenta- Cytometer (Beckman Coulter) and analyzed using CXP tion of the Universidad de Oviedo. To induce Kat8 dele- Software (Beckman Coulter 2006) and ModFit LT 3.0 (Ver- tion, tamoxifen (Sigma-Aldrich, T5648) was administrated ity Software House). 5020 Nucleic Acids Research, 2019, Vol. 47, No. 10 Quantitative real-time reverse transcriptase polymerase Pulsed field gel electrophoresis chain reaction (qRT-PCR) Temperature-dependent induction of synchronous neutrophil Cells for RNA extraction were pelleted and immediately apoptosis. For the arrest of neutrophils and the subse- frozen in dry ice. Storage was performed at -80◦ C. To- quent induction of synchronous apoptosis, we followed tal RNA was isolated using TRIzol Reagent (Life Tech- the protocol described by Pryde et al. (55) with minor nologies, Ref. 15596) following manufacturer´s instruc- modifications. Briefly, neutrophils were cultured in T75 tions. Chloroform (Ref. 1.02445.100) was purchased from flasks (Corning) at 1 × 106 cells/ml in Gibco RPMI Merck; and isopropanol (Ref. 1.09634.100) and ethanol 1640 growth medium supplemented with 100 U/ml peni- (Ref. 100983) from Merck-Millipore. To remove any pos- cillin and 100 ␮g/ml streptomycin (Gibco), 250 ␮g/ml sible contamination by DNA, 10 ug of total RNA were amphotericin B (Gibco), 10% (v/v) autologous serum, treated with DNase I (Turbo DNA-free kit, Ambion-Life 25 mM HEPES (Merck), 0.2% (w/v) albumin from bovine Technologies, Ref. AM1906) according to supplier’s proto- serum (BSA, Sigma-Aldrich), and 20 ␮g/ml cycloheximide col. (Sigma-Aldrich). After incubation at 15◦ C for 20 h, neu- Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 RNA was quantified both before and after DNase treat- trophils underwent apoptosis synchronously when the tem- ment with Nanodrop (ThermoScientific) checking purity perature was raised to 37◦ C. The cells for the preparation of as A260/280 ratio. One microgram of total DNA-free agarose plugs were collected 2 h after the temperature shift. RNA was reverse-transcribed in 20 ul of final reaction vol- ume using Oligo(dT)12–18 Primer (Invitrogen, Ref. 18418) Preparation of agarose plugs and Pulsed Field Gel Elec- and dNTP Mix (Invitrogen, Ref. 18427) in combination trophoresis (PFGE). CD15+ CD16+ neutrophil DNA with SuperScript™ III Reverse Transcriptase (Invitrogen, embedded in agarose plugs was prepared as follows: about 5 Ref. 18080) and RNaseOUT™ Recombinant RNase In- × 106 cells for each 100 ␮l of agarose plug to be made were hibitor (Invitrogen, Ref. 10777). After cDNA synthesis, 1 collected, washed once with 1× PBS (Oxoid), and resus- ␮l (2 units) of E. coli RNase H (Invitrogen, Ref.18021) was pended in 35 ␮l of 50 mM EDTA (Merk), 10 mM Tris (pH added and each sample was incubated at 37◦ C for 20 min to 7.5) (Roche). The suspension was quickly warmed at 42◦ C finally remove RNA complementary to the cDNA. All the and mixed with 65 ␮l of prewarmed 1% low melt agarose process was performed following the instructions detailed (NuSieve GTG Low Melting Temperature Agarose, Lonza) for SuperScript™ III RT (Invitrogen, Ref. 18080). and 125 mM EDTA. The suspension was distributed into Real-time PCR was carried out for each sample of cDNA well plug molds placed on a cool surface. The plugs were using 2× SYBR Green PCR Master Mix (Applied Biosys- extruded and incubated overnight at 53◦ C without agitation tems, 4309155). All measurements were performed in trip- in 600 ␮l of 500 mM EDTA, 10 mM Tris (pH 8.0), 1% (w/v) licate, non-template controls were included, and a calibra- N-Laurylsarcosine (Sigma-Aldrich), and 0.4 mg/ml Pro- tion curve was determined for each primer set. PCR reac- teinase K (Roche). Plugs were soaked for 1 h three times in 1 tions were run and analyzed using the StepOnePlus™ Real- ml of 50 mM EDTA, and 10 mM Tris (pH 7.5). As positive Time PCR System (Applied Biosystems). Gene expression controls, 50kb random fragments from a pool of samples was normalized using GAPDH as endogenous control for were prepared as follows: concentrations of genomic DNA human samples (51) and beta Actin for mouse samples and from neutrophils were determined with a Qubit 2.0 fluorom- analyzed by the comparative threshold (Ct) method. eter (Life Technologies). About 4 ␮g of DNA from each All primers were acquired from Sigma-Aldrich. Oligonu- sample was sonicated using a Bioruptor UCD-200 sonica- cleotide sequences are listed in Supplementary Table S1. tion device (Diagenode) performing for 2–4 min bursts in low power mode in order to shear the DNA into 50 kb frag- CB collection and CD34+ HSPC isolation for KAT8 down- ments. Sonicated samples were subjected to PFGE in liquid regulation on human primary cells form under the same running conditions described above. Fresh umbilical cord blood (CB) units from healthy Pulsed field electrophoresis was carried out in a 0.9% neonates were obtained from the Banc de Sang i Teixits, agarose gel (D5 agarose, Conda) in 0.5× TBE at 14◦ C with Barcelona, upon approval by the Ethics and Biohazard a CHEF DR III from Bio-Rad with a switch time ramping Board Committee. CB samples were pooled to reduce vari- from 5 to 30 s over a 22 h period, to resolve fragments from ability among individual CB units. Mononuclear cells were 20 to 400 kb. Lambda ladder was used as size marker for isolated using Ficoll-Hypaque and after lysing the red cells PFGE (New England Biolabs). For DNA extraction, the (Cytognos, Salamanca, Spain), CD34+ cells were purified portion of the gel containing the 50 kb DNA fragments of by magnetic bead separation using the human CD34 Mi- interest was excised. The agarose was melted at 55◦ C and croBead kit and the AutoMACS Pro separator (Miltenyi the DNA extracted using the QIAEX II Gel Extraction Kit Biotec, Bergish Gladbach, Germany) as per manufacturer’s (QIAGEN). instructions (52–54). Purity of the CD34+ fraction was as- sessed by flow cytometry using anti-CD34-PE (BD Bio- Bisulfite pyrosequencing science, San Jose, CA, USA) and only CD34+ fractions showing purity >95% were used (52,53). DNA methylation patterns of Alu repetitive sequences were Details of the KAT8 down-regulation (shKAT8) assays analyzed by bisulfite pyrosequencing in neutrophils and in human CD34+ progenitors and the subsequent analyses CD3+ T cells. Bisulfite modification of DNA was per- following induction of neutrophil differentiation are avail- formed with the EZ DNA methylation-gold kit (Zymo Re- able in Supplementary Materials and Methods. search) following the manufacturer’s instructions. The set Nucleic Acids Research, 2019, Vol. 47, No. 10 5021 of primers for PCR amplification and sequencing were de- Large Ct shifts between digested and undigested samples, signed using the specific software PyroMark assay design rendering high FE%, indicated that the target region was (version 2.0.01.15) (Supplementary Table S1). After PCR in the open chromatin, while minimal Ct shifts, resulting in amplification, pyrosequencing was performed using Pyro- low FE%, indicated that the target region was in the closed Mark Q24 reagents and a vacuum prep workstation, equip- chromatin. ment, and software (Qiagen). Formaldehyde-Assisted isolation of regulatory elements Immunofluorescence (FAIRE) quantitative PCR (FAIRE-qPCR) For immunofluorescence experiments 2 × 105 cells were ad- For the FAIRE analyses we followed the protocol for sus- hered to slides by Cytospin centrifugation (Thermo Scien- pension cells described by Simon et al. (56,57), using 1 × 107 tific). Samples were fixed with 4% formaldehyde for 15 min CD15+ CD16+ neutrophils and 1 × 107 CD3+ lymphocytes at room temperature, and permeabilized with 1× PBS/0.1% from four different healthy donors as starting material. The Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Triton X-100 (X100 Sigma-Aldrich) for 20 min at room enriched fragments were detected by real time qPCR using temperature. Blocking was performed with 1× PBS/10% primers for different repeats (see Supplementary Table S1). BSA (A7906 Sigma-Aldrich) for one hour at room tempera- ture, followed by incubation with the primary antibody rab- bit anti-H4K16ac (07-329, Merck Millipore) at 1:500 dilu- Histological analysis tion in antibody diluent (EnVision FLEX DM830 Dako), O/N at 4◦ C. Secondary antibody chicken anti-rabbit IgG Mouse tissues were fixed in 4% buffered paraformaldehyde Alexa Fluor 488 (A-21441 Invitrogen) at 1:500 dilution, was solution. Paraffin sections were stained with hematoxylin incubated for 1 h at room temperature and protected from and eosin. The May-Grünwald-Giemsa staining procedure light. Finally, slides were mounted using EverBrite mount- was followed to study human blood smears. ing medium with 4 ,6-diamidino-2-phenylindole (DAPI) (23002 Biotium). Immunofluorescence images were obtained using a Zeiss Chromatin immunoprecipitation assays and real-time PCR microscope equipped with a 63×/1.4 NA immersion lens, quantification (qChIP) and an AxioCam MRm camera (Carl Zeiss). DAPI and For the preparation of crosslinked chromatin, fresh samples H4K16ac images were acquired sequentially for a total were incubated in PBS with 1% formaldehyde for 10 min of 10 z-sections (Z step of 5 ␮m) with at the appropri- at RT. Crosslinking was stopped by the addition of glycine ate emission wavelengths. Fluorescence intensity measure- to 125 mM, and cells were washed in PBS. All subsequent ments were performed using the ZEN lite software (ZEN procedures were performed on ice, with buffers contain- lite 2.3 SP1, Carl Zeiss). ing the protease inhibitors (Complete EDTA-free) as pre- For immunofluorescence analysis of HL60 confluence viously described (58). Sonication yielded chromatin frag- experiment, cells were fixed in acetone at −20◦ C for ments of 150–500 bp in length. Immunoprecipitations were 20 min, washed with PBS, and incubated with primary performed using antibodies against H4K16ac (Active Mo- antibodies overnight at 4◦ C at 1:100 dilutions. Next, tif, 39167), total histone H3 (Abcam, ab1791) as positive fluorescent-conjugate secondary antibodies (Alexa fluor control, or IgG antiserum (Abcam, ab46540) as negative 488–conjugated goat anti-rabbit IgG from Serotec and control. CyTM3-conjugated donkey anti-mouse antibody from After testing the PCR amplification sensitivity of Jackson-ImmunoResearch) at 1:100 dilution were added primers, quantitative real-time PCR analysis was per- to the fixed cells, which were then incubated for 1 h at formed for each specific region using SYBR® Green of the room temperature in a darkened humidity chamber. Finally, StepOnePlus™ Real-Time PCR System (Applied Biosys- samples were counterstained and mounted with DAPI (10 tems). All measurements were performed in triplicate, non- ng/ml; Sigma-Aldrich) diluted 1:1000 in fluoromount G template controls were included, and a calibration curve (SouthernBiotech). Confocal immunofluorescence micro- was determined for each primer set. Three independent ex- graphs were captured with a TCS-SP2-AOBS confocal mi- periments were tested for each sequence. Oligonucleotide croscope (Leica Microsystems). sequences are listed in Supplementary Table S1. Chromatin accessibility assay Heat density scatterplots For the analysis of chromatin accessibility in multiple repet- itive elements, 1 × 106 CD15+ CD16+ neutrophils and 1 × Two-dimensional kernel density estimations were gener- 106 CD3+ lymphocytes from four donors were subjected to ated with the Bioconductor R-package MASS (7.3–47). chromatin extraction and nuclease digestion according to H4K16ac and DAPI signals were normalized by the maxi- manufacturer’s instructions (EpiQuik Chromatin Accessi- mum intensities obtained for each of the experiments prior bility Assay kit, P-1047 Epigentek). Purified DNA was ana- to the computation step. The resulting data comprised 9198 lyzed via real time qPCR using primers for different repeats and 5925 pairwise data points (H4K16ac and DAPI nor- (see Supplementary Table S1). Fold enrichment (FE) was malized intensities) across longitudinal sections of the cell calculated as the ratio of amplification efficiency (Ct) of the corresponding to observations from neutrophils (n = 72) or digested DNA sample over that of the non-digested sample. CD3+ cells (n = 62), respectively. 5022 Nucleic Acids Research, 2019, Vol. 47, No. 10 RESULTS CD3+ cells and 164 229 168 (73.46%) in neutrophils were Global H4K16 hyperacetylation in PB neutrophils aligned to one or the other strand of the reference genome (GRCh37/hg19) (Supplementary Table S2). In addition to To study the possible role of H4K16ac in the hematopoi- the unique alignments, other reads aligned in multiple posi- etic system we first used HPLC and HPCE to analyze the tions (multi-reads or ambiguous reads) (Supplementary Ta- global levels of H4 acetylation in different human blood ble S2). Similar counts were obtained in four additional in- cell types. Intriguingly, we found a 1.7-fold relative increase dependent biological replicates (Supplementary Table S2). in the monoacetylation of H4K16 in neutrophils (CD15+ To identify H4K16ac enriched regions (peaks) we used CD16+ ) compared with CD3+ T cells, which, in absolute the Model-based Analysis of ChIP-seq (MACS2) algorithm terms, implies an increase in H4K16ac in over 20% of the (62), and the corresponding inputs as control samples (see neutrophil genome (Figure 1A). Analysis of other acety- Materials and Methods). We identified a total of 229 474 lated forms of H4 (59), showed that the relative percentage peaks in CD3+ cells, and 266 860 peaks in neutrophils (Fig- of diacetylated isoforms (where Lysine 12 and 16 are acety- ure 2A; Supplementary Tables S3 and S4). The peak size Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 lated) was slightly lower in neutrophils (Supplementary Fig- differed significantly between the two cell types (Wilcoxon ure S1A), and tri- and tetracetylated isoforms (where, re- rank sum test, P < 2.2e–16): CD3+ showed a median peak spectively, Lysines 8, 12 and 16 or Lysines 5, 8, 12 and 16 size of 615 bp, while that of neutrophils was 2.29 kb (Fig- are acetylated) were not detected in blood cells with this ure 2A). Consequently, the total length of neutrophil peaks technology. Histone H3 and histone H2B did not show any was 611 116 535 bp, covering 19.48 % of the genome, and substantial changes in global acetylation between CD15+ for CD3+ T cells it was 150 618 607 bp, accounting for CD16+ and CD3+ T cells (Supplementary Figure S1B). only 4.8 % of the genome (Figure 2A). This higher cover- As it had been previously reported that the majority of age by peaks in neutrophils than in T cells was evident in histone H4 monoacetylation occurs at H4K16 (60,61), we all chromosomes (Supplementary Figure S2A), confirming hypothesized that the global H4 acetylation changes we ob- the genome-wide level of the global differences between spe- served primarily occur at this position. To test this, we used cific blood cell types observed by HPLC/HPCE and mass two different approaches. First, we used Western Blotting spectrometry. to analyze the levels of H4K16ac and H4K12ac in sev- To study the associations between H4K16ac ChIP-seq eral blood cell types, including those previously analyzed peaks and genomic features we determined their distribu- by HPLC/HPCE, using antibodies against these histone tion throughout the genome. We found that peaks of neu- PTMs. K12 is another lysine residue that has been reported trophils were located in regions with higher CpG densi- to be acetylated in histone H4 (60,61). The results showed ties than those corresponding to the peaks of CD3+ T cells an enrichment of H4K16ac (Figure 1A, and Supplemen- (Wilcoxon rank sum test, P < 2.2e–16) (Figure 2B), irre- tary Figure S1C) but not of H4K12ac (Supplementary Fig- spective of the chromosome analyzed (Supplementary Fig- ure S1d) in neutrophils (CD15+ CD16+ ), suggesting that, ure S2A). In line with these results, neutrophil peaks were the observed global histone H4 acetylation changes pri- located preferentially on CpG islands and GpG shores (Per- marily did occur at K16. To corroborate this further, we mutation tests, P < 0.01, FC CI = [1.32, 1.43] and P < used mass spectrometry to analyze the levels of H4K16ac 0.01, FC CI = [1.92, 2.06] respectively) (Supplementary Fig- in five types of blood cells (CD34+ hematopoietic progeni- ure S2B). Interestingly, a detailed analysis of the relative tors, CD19+ , CD3+ , CD15+ CD16− and CD15+ CD16+ ). genomic distribution of ChIP-seq peaks showed a signifi- The integration of the signals corresponding to the pep- cant overlap between neutrophil peaks and promoter and tide 4–17 confirmed that the global levels of monoacety- intragenic regions (Permutation tests, P < 0.01, FC CI = lated histone H4 were much higher in myeloid cells than in [1.50, 1.61], and P < 0.01, FC CI = [1.28, 1.35] respectively). CD34+ hematopoietic progenitors and B and T cells (Fig- However, the overlap at intergenic regions was lower than ure 1B). In addition, tandem mass spectrometry revealed expected in both cell types, and this effect was more pro- that, in all samples analyzed, the great majority of his- nounced in CD3+ T cells (Permutation tests, P < 0.01, FC tone H4 monoacetylation occurred at lysine 16 (Figure 1C), CI = [0.69, 0.72] and P < 0.01, FC CI = [0.75, 0.77] re- thus confirming that H4K16ac is highly regulated in the spectively) (Figure 2c). Similar results were obtained in four hematopoietic system. additional independet biological replicates (Supplementary Figure S3). Genome-wide differential distribution of H4K16ac in termi- Further analysis of H4K16ac peaks identified high num- nally differentiated blood cells bers of associated genes in both neutrophils and T cells (Figure 2D; Supplementary Figure S4; Supplementary Ta- To study the role of H4K16ac in blood cells in more de- bles S3 and S4). Although the number of common peak- tail, we determined the genome-wide distribution of this hi- associated genes was high, the genic sequences (in bp) en- stone mark in human CD3+ T cells and neutrophils (CD15+ riched in H4K16ac were cell-type specific, and only 4.2% CD16+ ) combining chromatin immunoprecipitation, tar- of the total were common to both CD3+ T cells and neu- geting this histone mark with high-throughput sequencing trophils (Figure 2E). (ChIP-seq) (see Materials and Methods). To find the relationship between H4K16ac and other hi- We generated over 200 million 50-base paired-end reads stone marks in neutrophils, we trained a 12 state model from H4K16ac enriched and input (control) libraries of with ChromHMM (63). We identified histone mark com- CD3+ T cells and neutrophils (Methods, and Supplemen- binations reflecting different functional chromatin states tary Table S2). Of these raw reads, 191 640 570 (77.33%) in Nucleic Acids Research, 2019, Vol. 47, No. 10 5023 A B Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 C Figure 1. Global H4K16ac level in different types of blood cell. (A) Global H4 monoacetylation analyzed by HPLC and HPCE in neutrophils (CD15+ CD16+ ) and CD3+ T cells. Error bars represent standard deviation ***P-value <0.001. On the right, western blot comparing H4K16ac in CD3+ T cells versus neutrophils. H3 was used as loading control. The intensity (relative units) of H4K16ac bands were normalized to H3 and quantified using ImageJ software (numbers below the bands). (B) Levels of monoacetylated histone H4 analyzed by MALDI-TOF in the hematopoietic cells. The error bars indicate the maximum and minimum values obtained from six spectra per sample (technical replicates). (C) Fragment spectra showing that acetylation of the peptide 4–17 primarily occurs on K16 in the samples indicated. The left panel is a zoom of the region encompassing the y4 fragment ion and its acetylated form, while the right panel plots the region corresponding to b9 and its acetylated form. In both panels, the upper spectrum corresponds to the unmodified peptide 4–17 (positive control for y4), the others are the fragment spectra of the monoacetylated peptide 4–17 of the samples indicated. The detection of acetylated y4 but not its unmodified form, as well as the detection of only unmodified b9 indicate that it is mostly K16 that is acetylated. *: impurity. 5024 Nucleic Acids Research, 2019, Vol. 47, No. 10 A B C D Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 E F Figure 2. Genomic distribution of H4K16 acetylation in neutrophils and CD3+ T cells. (A) Bar plot showing the number of H4K16ac-enriched peaks identified by ChIP-seq in neutrophils and CD3+ T cells. On the right, a violin plot shows the distribution of the peak size in both cell types, and a bar plot shows the total coverage of the genome in base pairs (bp). (B) Box plot showing the genomic distribution of CpG density for the H4K16ac-enriched peaks in both cell types. Dotted red line represents average CpG density in the genome. (C) Bar plots showing the overlap between H4K16ac-enriched peaks and different genomic regions in both neutrophils and T cells. Violin plots represent a permutation-based null distribution for the specific genomic region overlap ratio for each peak data set. (D) Representative ChIP-seq tracks for H4K16ac in neutrophils and CD3+ T cells. Peaks showing fold change above 1.5 are underlined below the profiles. The RefSeq genes track is shown below. (E) Euler diagrams showing the overlap between peaks of neutrophils and CD3+ T cells, for both the whole genome and the different genomic regions. (F) ChromHMM training of a 12-chromatin states model (left panel) in neutrophils. The central panel shows the enrichment of each of the 12 chromatin states in different genomic features. The right panels show the percentage of neutrophil-specific segments for two neutrophil Blueprint samples (C0010KH2 and C0011lH2). Nucleic Acids Research, 2019, Vol. 47, No. 10 5025 (Figure 2F). Interestingly, despite the fact that H4K16ac ferences in proliferation potential between the scramble had a higher correlation with enhancers and actively tran- and shKAT8 cells (Supplementary Figure S7C). How- scribed states, we identified a specific chromatin state char- ever shKAT8 cells showed a decresase in the potential of acterized by the presence of H4K16ac alone (state 10). This their hematopoietic progenitor to generate CFUs, as well was one of the most extensive chromatin states (Supple- as changes in CFU distribution (Supplementary Figure mentary Figure S5) and it was enriched in CpG hyper- S7D). Specifically, we observed a decrease in the percent- methylated regions (Figure 2F). After segmentation with age of CFUs of erythrocytes, granulocytes and mixed cells, ChromHMM, 7.71 % of the bins (segment of 200 bp con- and an increase in CFUs of granulocyte–macrophage and taining H4K16ac) were unique to neutrophils, in concor- macrophages (Supplementary Figure S7E). dance with results shown in Figure 2A, and the majority of To confirm the role of H4K16ac in the regulation of which overlapped with states 9 and 10, characterized by the myeloid differentiation we analyzed the levels of this epige- presence and absence of H4K16ac respectively (Figure 2F). netic mark in the human acute promyelocytic leukemia cell line HL60 incubated with all-trans retinoic acid (ATRA) Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 and dimethylsulfoxide (DMSO), two drugs that can induce H4K16ac and KAT8 play a role in hematopoietic differenti- neutrophil differentiation in both hematopoietic progeni- ation tors (64) and myeloid cancer cells (65,66). In line with pre- To determine whether H4K16ac changes between different viously published data (21–24) the cancer cells showed a types of blood cells reflects a role of this histone PTM in lower H4K16 acetylation level than their non-tumorigenic hematopoietic cell differentiation, we first generated an in- terminally differentiated perihperal blood cell counterparts ducible Kat8 knockout mouse model by crossing “floxed” (Supplementary Figure S8A). The analysis of the expres- Kat8 mice with tamoxifen-inducible ubiquitous Cre mice sion levels of neutrophil cytosol factor 1 (NCF1) indicated (Kat8fl/fl ;ER-Cre mice), KAT8 being the principal H4K16 that combined treatment with ATRA and DMSO induced acetyltransferase in mammals (8). Tamoxifen administra- the greatest degree of neutrophilic differentiation in HL60 tion (Figure 3A) induced the excission of LoxP-flanked re- cells (Supplementary Figure S8B). Supporting this, im- gion of Kat8 (Supplementary Figure S6A), and 1 week munoflurorescent staining revealed that most of the HL60 after the start of the treatment these mice began to lose cells treated with ATRA and DMSO acquired the lob- weight (Figure 3B), and died after three weeks (Figure 3C). ulated nuclear shape typical of neutrophils (Supplemen- The precise cause of death was unclear. However, post- tary Figure S8C). However, the HPLC/HPCE analysis of mortem pathological examination of the Kat8−/− ;ER-Cre global histone H4 monoacetylation showed no significant mice identified an evident decrease in the size of the spleen, differences of cells after differentiation with ATRA and thymus, and liver (Figure 3D). At the microscopic level, DMSO (Supplementary Figure S8D), and this was con- all the samples analyzed showed major tissular alterations firmed by mass spectrometry (Supplementary Figure S8E). (Supplementary Figure S6B). To determine the effects of In addition we performed a ChIP-seq analysis to deter- Kat8 deletion in the blood compartiment, we analyzed the mine the locus-specific changes in H4K16ac during HL60 levels of peripheral blood cell types using a hematology an- neutrophilic differentiation induced by ATRA and DMSO alyzer (Abacus). This showed that tamoxifen administra- (Supplementary Table S2), and confirmed that only minor tion markedly reduced the number of white blood cells of global changes occurred (overlapping test; P < 0.005; see Kat8−/− ;ER-Cre mice (Figure 3E), affecting both myeloid Methods) (Supplementary Figures S8F and G, and Sup- and lymphoid lineages. Moreover, Kat8 deletion inverted plementary Tables S5 and S6). We found that H4K16ac the percentage of myeloid and lymphoid cells present in the was enriched in many GO terms of biological processes blood compared to controls (Figure 3F), suggesting a pos- related to development and differentiation in both HL60 sible role of Kat8 in blood cells survival. and in vitro HL60 differentiated cells (FDR < 0.1) (Sup- To study in more detail the possible role of KAT8 and plementary Table S7). Interestingly, the small differences H4K16ac in the regulation of myeloid differentiation we (genes associated with H4K16ac specific peaks) between also compared in vitro differentiation to neutrophils of control and differentiated cells were related to, among oth- murine cKit+ progenitors sorted from bone marrow from ers, ontology terms such us cell development or regulation Kat8fl/fl ; ER-Cre and Kat8+/+ ;ER-Cre mice. Kat8 excision of signaling (FDR < 0.1) (Supplementary Tables S8 and was induced in vitro by incubating with 4-OH-tamoxifen re- S9), which indicates that the absence of global H4K16ac ducing Kat8 expression and H4K16ac (Figure 3G and H). changes measured by mass spectrometry is compatible with Differentiation was confirmed by May Grünwald-Giemsa minor locus-specific H4K16ac changes, and also that locus- staining (Figure 3I), and the levels of the two neutrophil ter- specific H4K16ac may play a role in the regulation of minally differentiated markers were quantified and showed hematopoietic genes during in vitro differentiation of HL60 lower levels of expression in KO cells (Figure 3J). cells. Collectively, these results suggest that, in HL60 cells, To determine whether KAT8 and H4K16ac were also neutrophilic differentiation can be achieved, at least in part, involved in hematopoietic differentiation in humans, we without the global hyperacetylation of H4K16, and that knocked down KAT8 (shKAT8) in CD34+ progenitors ob- locus-specific changes in H4K16ac during hematopoietic tained from umbilical cord blood (Supplementary Figure differentiation do not contribute significantly to the global S7A) and analyzed clonogenic potential as measured by changes of this histone PTM that are observed by mass the ability to form colony formation units (CFUs). KAT8 spectrometry in terminally differentiated neutrophils (Fig- downregulation induced a global decrease of H4K16ac ure 1). (Supplementary Figure S7B). We did not find any dif- 5026 Nucleic Acids Research, 2019, Vol. 47, No. 10 A B C D Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 E F G H I J Figure 3. Role of H4K16ac in peripheral white blood cells. (A) Diagram showing the timeline for tamoxifen administration in Kat8fl/fl mice with and without the inducible ubiquitous Cre (Kat8fl/fl (n: 5) and Kat8fl/fl ;ER-Cre (n: 8)). (B) Weight evolution of control and Kat8−/− ;ER-Cre mice after tamoxifen treatment. (C) Kaplan–Meier survival curve showing that Kat8−/− ; ER-Cre mice (red line) differs significantly from Kat8fl/fl mice (blue line) (P-value = 0.0011, Kaplan–Meier log-rank test). (D) Representative images of spleen, thymus, and liver from Kat8fl/fl and Kat8−/− ;ER-Cre mice showing differences in the size of the organs. Line plots showing (E) differences in total number of white blood cells and (F) percentage of lymphoid and myeloid cells between Kat8fl/fl ; ER-Cre and Kat8−/− mice after tamoxifen treatment. (G). Genotype analysis of Kat8 locus using PCR specific primers on genomic DNA from Kat8+/+ ;ER-Cre (pool, n: 3) as well as Kat8−/− ;ER-Cre (pool, n: 3) mice cells after tamoxifen treatment. The double band indicating approximately 30% excision of Kat8 was only detected in mice containing the floxed alleles after treatment with tamoxifen. On the right, RT-qPCR showing downregulation of Kat8 in the cKit+ progenitors. Expression levels determined by real time-qPCR are an average of triplicated measurements (+SD), normalized to transcript levels of beta Actin, and represented as a fold-change relative to the Kat8+/+ ;ER-Cre control (a t-test showed significant differences; ***P < 0.001). (H) Representative immunofluorescence images showing lower levels of H4K16ac in Kat8−/− ;ER-Cre. Average H4K16ac fluorescence intensity was quantified using ZEN lite software (a t-test showed significant differences; **P < 0.01). (I) May Grünwald Giemsa stain (40×) after in vitro differentiation to neutrophils (day 7). (J) Expression of two markers (Collagenase and Gelatinase) of neutrophilic differentiation (Pham, C., Nature Reviews Immunology, 2006) (day 7). Expression levels determined by real time-qPCR are an average of triplicated measurements (+SD), normalized to transcript levels of beta Actin, and represented as a fold-change relative to the Kat8+/+ ;ER-Cre control (a t-test showed significant differences; **P < 0.01, ***P < 0.001). Nucleic Acids Research, 2019, Vol. 47, No. 10 5027 H4K16ac level is functionally associated with apoptosis in umbilical cord blood and then we induced neutrophil differ- myeloid cells entiation (Figure 4D). Six days after the induction of neu- trophil differentiation, between 50 and 60% of scramble and As locus-specific H4K16ac changes in hematopoietic genes shKAT8 cells were positive for CD33 and CD15 (data not during HL60 cell differentiation did not seem to have a shown). shKAT8 reduced H4K16 acetylation (Figure 4D), substantial impact on global H4K16ac level, the regulation and Annexin V-7AAD staining at days 8–11 revealed a de- of hematopoietic cell differentiation was considered an un- crease in apoptosis in cells with lower KAT8/H4K16ac lev- likely candidate to explain the massive H4K16 hyperacety- els (Figure 4E). lation observed in peripheral neutrophils. We therefore pos- Next, we used the in vitro model of neutrophil differen- tulated the hypothesis that this epigenetic change might play tiation on murine cKit+ progenitors from the Kat8 con- a role in neutrophil apoptosis. Indeed, it has been previously ditional knockout mouse (Figure 3) to analyze neutrophil shown that H4K16ac is involved in the regulation of apop- apoptosis with and without staurosporine treatment, which totic genes in leukemia cancer cells (18). is known to be an apoptosis inductor. Analysis of the To study the role of H4K16ac in apoptosis in myeloid Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Annexin V-7AAD staining of differentiated neutrophils at cells in more detail, we first analyzed H4K16ac changes day 7 revealed that apoptosis levels were similar in con- during HL60 cell culture. These cells have previously been trol and Kat8 knockout-derived cells, both with (85.98% shown to undergo spontaneous apoptosis in a cell-density and 87.76%, respectively) and without (26.81% and 25.96%, dependent fashion (67). HPLC/HPCE and tandem mass respectively) staurosporine treatment (Supplementary Fig- spectrometry experiments carried out here showed that ure S10). Within each treatment there were, however, dif- global H4K16ac increased with cell density until the cells ferences between control and Kat8 knockout-derived cells reached confluence (day 5), after which it stabilized at ∼35% in their distribution between the early and late phases of (Figure 4A and B, Supplementary Figure S9a). Interest- apoptosis. Without staurosporine, late apoptosis was de- ingly, the highest H4K16ac levels coincided with the in- tected in 4.61% of control cells but in only 2.26% of Kat8 duction of spontaneous apoptosis (day 5) (Figure 4A). knockout-derived cells. Similarly, after treatment with stau- H4K16ac has been shown to fluctuate during the cell cy- rosporine we found 8.23% of control cells to be in the late cle, peaking during the S phase, and reaching its lowest lev- apoptosis phase and only 5.56% of Kat8 knockout-derived els in the G2/M phase (68). To determine any possible ef- cells (Supplementary Figure S10). Thus, these results indi- fect of the cell cycle on changes in H4K16ac during the cate that Kat8 downregulation in murine cKit+ progenitor- increase in cell confluence we analyzed the percentage of derived neutrophils induces a slight retardation of entry into cells in each phase of the cycle in the assay previously de- the late apoptosis phase, both in the presence and absence scribed. The results showed a negative correlation between of staurosporine (Supplementary Figure S10). global H4K16ac and the percentage of cells in either the S To further study the relationship between global H4K16 or the G2/M phase during exponential cell growth (Pearson acetylation and apoptosis, we treated the cancer cell line Correlation coefficient = –0.770, P-value = 0.025; Pearson HL60 with the topoisomerase inhibitor etoposide (Supple- Correlation coefficient = –0.714, P-value = 0.047, respec- mentary Figure S11A), which has been shown to induce tively), but no clear associations following the induction of apoptosis in myeloid tumors (70). Additionally, H4K16 apoptosis (Supplementary Figure S9B). Taken together our acetylation was modulated by treating HL60 cells with ei- findings lead us to discard the notion that the cell cycle has ther SAHA or KAT8 interference or upregulation (Supple- any marked effect on the cell-density dependent increase in mentary Figure S11B and C). Etoposide has been demon- H4K16ac. strated to induce an apoptosis that cannot be measured by The association between H4K16ac and apoptosis led Annexin V staining in HL60 due to the compound’s lipid us to study the possible role of this epigenetic mark in anti-oxidant action, which inhibits phosphatidylserine ex- programmed cell death in more detail. We first analyzed ternalization and, therefore, its staining (71). For this rea- the levels of apoptosis in undifferentiated vs. differentiated son, although we measured both Annexin V+ cells and de- HL60 cells which were either untreated or treated with bris percentages (from Propidium Iodide staining), only de- suberoylanilide hydroxamic acid (SAHA) (Supplementary bris values account for etoposide-induced apoptosis. We Figure S9C), an inhibitor of the H4K16 histone deacety- observed that, when incubation was with etoposide alone, lases HDAC1 and HDAC2 (69). Treatment with SAHA in- there was induction of apoptosis which was independent creased monoacetylated histone H4 from 33% to 56% in of any increase in H4K16ac (Supplementary Figure S11A). control HL60, and to 60% in HL60 cells differentiated with Indeed, after 4 h of incubation with the topoisomerase in- ATRA+DMSO (Figure 4C; Supplementary Figure S9D). hibitor there was even a small decrease in acetylation, which Interestingly, incubation with SAHA also increased apop- is in line with previously published data (72). When cells tosis in HL60 cells, with a more pronounced effect in differ- were treated with SAHA, there were initial (i.e. prior to entiated cells (Figure 4C). etoposide incubation) high levels of Annexin V+ apopto- Histone deacetylase inhibitors can induce acetylation sis, but these decreased over time during etoposide treat- changes at positions other than H4K16 (69). ment. As the accompanying increase in debris percentages In order to establish the particular role of global could be explained by the fragmentation of Annexin V+ H4K16ac in apoptosis, we ectopically regulated the levels of dying cells, it can be seen that etoposide-induced apopto- the H4K16-specific histone acetyltransferase KAT8 in both sis does not seem to change with SAHA treatment (Sup- human primary cells and mouse cells. Firstly, we knocked plementary Figure S11c). Taken together, these data sug- down KAT8 (shKAT8) in human CD34+ progenitors from gest that H4K16ac is not essential for the induction of pro- 5028 Nucleic Acids Research, 2019, Vol. 47, No. 10 A B Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 C D E Figure 4. Association between global levels of H4K16ac and apoptosis in myeloid cells. (A) Global H4 monoacetylation analyzed by HPLC and HPCE (blue bars in the upper panel), and percentage of Annexin+ cells (lower panel) during HL60 cell increasing confluence (black line). (B) Representative immunofluorescence images showing the changes in H4K16ac during the exponential growth of the cells at days 2 and 8. (C) Bar plots showing the effects of the treatment with SAHA in control and in vitro differentiated HL60 cells. Upper panel depicts the effects of the treatment on H4 monoacetylation levels, and lower panels, the effect on apoptosis measured by Annexin V-7AAD staining and percentage of debris from Propidium Iodide staining. A: ATRA, D: DMSO, S: SAHA. (D) RT-qPCR showing downregulation of KAT8 in transduced CD34+ progenitors. Expression levels determined by real time-qPCR are an average of triplicated measurements (+SD), normalized to transcript levels of GAPDH, and represented as a fold-change relative to the scramble. Below, a western blot comparing H4K16ac in shKAT8 cells versus the scramble. H3 was used as loading control. The intensity (relative units) of H4K16ac bands were normalized to H3 and quantified using ImageJ software (numbers below the bands). On the right, representative immunofluorescence images showing lower levels of H4K16ac in shKAT8 cells at day 11. Average H4K16ac fluorescence intensity was quantified using ZEN lite software (a t-test showed significant differences; **P < 0.001). (E) Bar plots showing the lower percentage of apoptosis of shKAT8 cells. Apoptosis levels were assessed by Annexin V-7AAD staining. Nucleic Acids Research, 2019, Vol. 47, No. 10 5029 grammed cell death, and that basal H4K16 acetylation level Figure S13). Enrichment analysis of different repeat fami- is sufficient for programmed cell death to occur. In line with lies across the chromatin states (Figure 5B) confirmed that this, KAT8 over expression did not affect the apoptosis in- H4K16ac peaks (emission order 10) specific for neutrophils duced by etoposide. However, ectopic downregulation of (Figure 2F) are enriched in Alu repeats, MIR and LINE 2. KAT8 in HL60 cells was associated with reduced etoposide- Interestingly, although in the analyses of the ‘unique induced cell death at 3h of treatment (Supplementary Fig- mapped reads’ we identified a significant overlap of ure S11C). H4K16ac peaks with several types of repetitive sequences In order to explore the possible connection between the (Figure 5A and Supplementary Figure S13A), we also iden- cell death detected in HL60 cells in our study and the gen- tified that the regions flanking the peaks in both cell types eration of neutrophil extracellular traps (NETs) (73), we were highly enriched in repetitive elements (Figure 5C), sug- tested the NETosis of HL60 cells under different condi- gesting that H4K16ac could be even more enriched in repet- tions: treated or not with SAHA, and after the induction itive elements than initially revealed, though this would be of ectopic up- or downregulation of KAT8. In all the con- masked by the limitation of the computational method em- Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 ditions analyzed, the levels of elastase activity, as a measure- ployed to deal with ambiguous reads and detect enrichment ment of NETs formation, were similar to background signal of peaks in those regions. We also found similar results in 4 (Supplementary Figure S12). It would thus seem that, under additional independet biological replicates (Supplementary these conditions, NETosis is not induced. Figure S13B). To investigate this further we carried out a new anal- ysis of the ‘ambiguous reads’, that is, those that align in Enrichment of H4K16ac at specific repeated DNA in neu- multiple parts of the genome and could in theory be as- trophils sociated with repetitive regions (Supplementary Table S2) As the H4K16ac changes in differentiation and apopto- (see Materials and Methods). In line with the results ob- sis genes between lymphoid and myeloid cells were un- tained with the unambiguosly mapped reads, this approach likely to explain the global differences observed by mass showed that neutrophil H4K16ac peaks were enriched in spectrometry (>20% of the genome), we reanalyzed our SINEs (particularly Alu sequences) (Homer Genome test, ChIP-seq data of CD3+ T cells and neutrophils to iden- P < 0.001, logRE = 0.1) and LINE-2 (Homer Genome test, tify the DNA regions where these changes occur. To this P < 0.001, logRE = 0.03), but depleted in LINE-1 (Homer end we focused our attention on repetitive DNA, a type of Genome test, P < 0.001, logRE = –0.12) and in centromeric DNA sequence that is not usually given much attention in repeats (Homer Genome test, P < 0.001, logRE = –0.15) ChIP-seq experiments because they are difficult to map in (Figure 5D). Similar results were obtained in 4 additional the genome. Interestingly, the HOMER enrichment anal- independent biological replicates, although in the case of yses showed that H4K16ac peaks, overlapped with specific CD3+ T cells we observed some interindividual variability repetitive sequences in the genome in both CD3+ T cells and (Supplementary Figure S14). Collectively these results in- neutrophils (Supplementary Tables S3 and S4). To further dicate that, in neutrophils, H4K16ac is enriched at specific investigate these relationships we carried out a comparative repeated DNA regions. statistical analysis of H4K16ac enrichment in three types of repetitive sequence identified in the ChIP-seq: Short In- H4K16ac-rich DNA repeats mark genomic regions frag- terspersed Nuclear Elements (SINEs), Long Interspersed mented early on in neutrophil apoptosis Nuclear Elements (LINEs) and Satellites. These analyses showed that the H4K16ac peaks of neutrophils were signif- As H4K16ac has been proposed to play an important role icantly enriched in SINEs (Permutation test, P < 0.01, FC in making chromatin accessible in DNA repair (37), we con- CI = [1.34, 1.42]), while being depleted in LINEs (Permuta- sidered it reasonable to assume that it could also play a tion test, P < 0.01, FC CI = [0.61, 0.63]) and satellites (per- similar role in DNA cleavage during neutrophil apoptosis. mutation test, P < 0.01, FC CI = [0.26, 0.38]). Conversely, In the first stages of the apoptotic process, genomic DNA H4K16ac peaks of CD3+ T cells were significantly depleted is cleaved into 300 and 50 kb fragments (74). Therefore, in SINEs (Permutation test, P < 0.01, FC CI = [0.70, 0.72]) to study the possible role of global H4K16 hyperacetyla- (Figure 5A, upper panels). The analyses of specific families tion in early DNA fragmentation during neutrophil apop- of repetitive elements showed that H4K16ac peaks of neu- tosis, we determined whether the genomic regions specifi- trophils were significantly enriched in Alu elements (Homer cally enriched in H4K16ac in neutrophils were fragmented Genome test, P < 0.001, logRE = 0.50), mammalian-wide during the early stages of neutrophil apoptosis. Our ChIP- interspersed repeat (MIR) sequences (Homer Genome test, seq data suggested that the global H4K16 hyperacetyla- P < 0.001, logRE = 0.45), and LINE-2 (Homer Genome tion observed in neutrophils by mass spectrometry is, at test, P < 0.001, logRE = 0.28), but depleted in LINE- least in part, due to the increase of this histone modifica- 1 and satellites (Homer Genome test, P < 0.001, logRE tion at specific repeated DNA regions (Figure 5). We thus = –0.85 and P < 0.001, logRE = –1.11, respectively). In designed PCR oligonucleotides against 6 of these regions contrast, H4K16ac peaks of CD3+ T cells were signifi- (Supplementary Table S1) and used them to amplify the cantly depleted in Alu sequences (Homer Genome test, P 50 kb DNA fragments isolated by pulsed field gel elec- < 0.001, logRE = –0.57) (Figure 5A, lower panels). Simi- trophoresis (Figure 6A; Supplementary Figure S15A, B) lar results were obtained in four additional independet bi- from apoptotic neutrophils obtained from two independent ological replicates, although in the case of CD3+ T cells we series of healthy individuals (Figure 6B; Supplementary observed some interindividual variability (Supplementary Figure S15C). Bisulfite pyrosequencing analysis of these ge- 5030 Nucleic Acids Research, 2019, Vol. 47, No. 10 Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Figure 5. Distribution of H4K16 acetylation at specific repetitive DNA elements in neutrophils and CD3+ T cells. (A) Bar plots showing the overlap between H4K16-enriched peaks and different types of DNA repeats, obtained in the analyses of unique mapped reads. Violin plots represent a permutation-based null distribution for each type of repeated DNA overlap ratio for each peak data set. Lower panels show associations between different types of repetitive DNA elements with H4K16ac peaks in neutrophils and CD3+ T cells, ranked by Q-value, and enrichment score (relative risk). (B) Enrichment of the 12 chromatin states described in Figure 2F in different types of DNA repeats. (C) Plot showing the coverage of repetitive elements in the regions flanking the H4K16ac enriched peaks obtained in the analyses of unique mapped reads in both cell types. (D) Stacked bar charts describing the classification of ambiguous reads for neutrophil and CD3+ T cell signals relative to the corresponding input, and according to the enrichment in different types of DNA repeats. Lower panels show associations between different types of repetitive DNA elements with probable H4K16ac-enriched regions obtained from ambiguous mapping in neutrophils and CD3+ T cells, and ranked by Q-value, and enrichment score (relative risk). Nucleic Acids Research, 2019, Vol. 47, No. 10 5031 A B Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Figure 6. Associations between H4K16ac enriched regions and DNA breakage during neutrophil apoptosis. (A) Schematic representation of the pulsed- field gel electrophoresis (PFGE) experiment to analyze the 50 kb DNA fragments from apoptotic neutrophils. (B) qPCR amplification of specific H4K16ac- enriched repeats using the 50kb DNA fragments from the PFGE. Data are an average of triplicated measurements, normalized to a H4K16ac-poor region of the albumin gene, and represented as a fold-change relative to the 50kb sonicated DNA. Intact DNA and 50kb random fragments from a pool of samples were used as positive controls. H4K16ac peaks identified by MACS2, and genome diagrams showing the relative enrichment of H4K16ac in neutrophils (blue), and CD3+ T cells (green) are shown on the left. Significant H4K16ac peaks showing fold change above 1.5 are indicated by a colored bar below the profile. nomic regions showed that they are highly hypermethylated ments generated by sonication (Figure 6B; Supplementary in both neutrophils and CD3+ T cells (Supplementary Fig- Figure S15C). In contrast, the oligonucleotides designed ure S16). As a negative control, we used oligonucleotides against the H4K16ac-rich DNA repeats amplified the un- designed against DNA regions lacking H4K16ac in neu- specific 50 kb DNA fragments to a far greater extent than trophils (Supplementary Table S1). The oligonucleotides the early DNA apoptotic 50 kb DNA fragments isolated by designed against H4K16ac-poor DNA regions amplified pulsed field gel electrophoresis (Figure 6B; Supplementary both the early apoptotic 50 kb DNA fragments isolated by Figure S15C), which indicates that H4K16ac is enriched in pulsed field gel electrophoresis and unspecific 50 kb frag- 5032 Nucleic Acids Research, 2019, Vol. 47, No. 10 the DNA fragments generated during the early stages of showing the critical importance of KAT8 and H4K16ac in neutrophil apoptosis. these processes (77,78). Finally, to study the possible effect of H4K16 hyperacety- Although our results supported a role for H4K16ac in lation at these DNA repeats on the chromatin conforma- gene regulation during cell differentiation, such changes in tion state we took two different approaches. First, we used the small number of genes involved in cell differentiation an H4K16ac-specific antibody to determine by immunoflu- are insufficient to explain the massive global differences be- orescence the enrichment of this histone PTM in the nu- tween neutrophils and CD3+ T cells observed by mass spec- clear regions of human neutrophils and CD3+ T cells (Fig- trometry that affects >20% of the genome. It has been re- ure 7A). As expected, DAPI staining was very different be- ported that mice lacking a processing protease of the nu- tween lymphoid and myeloid cells. Neutrophils displayed clear envelope protein lamin A show a low level of H4K16ac large blocks of highly compacted heterochromatin in the (79,80). However, Pelger-Huët patients, who have lamin B nuclear periphery. These blocks were separated by regions mutations, and individuals with myelodysplastic syndrome, of mildly compact chromatin and the least compacted chro- who present with an altered neutrophil nuclear morphology, Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 matin was in the center of the nucleus (Figure 7A). In con- have a normal distribution of H4K16ac (data not shown), trast, CD3+ cells showed a more uniform chromatin com- suggesting that this histone PTM does not play a central paction pattern, with only a few small blocks of heterochro- role in these disease processes. Future research is needed to matin attached to the nuclear periphery. H4K16ac stain- determine the biological consequences, if any, of the alter- ing was almost completely absent from the large blocks of ations of H4K16ac in Zmpste24 mutant mice, which are de- heterochromatin of the neutrophil nuclear envelop (Figure ficient in the proteolytic processing of prelamin A. 7A, B) and, interestingly, although the low compacted chro- Previous findings have also suggested that H4K16ac also matin areas of the center of the nucleus were stained, the plays an important role in gene regulation during pro- highest H4K16ac signal was observed in the mildly compact grammed cell death (17,18). Interestingly, we found a corre- chromatin regions interspersed between the large blocks of lation and a functional association between H4K16ac and heterochromatin of the nuclear periphery (Figure 7A, B). apoptosis in HL60 cells and in neutrophils obtained by in In CD3+ cells, H4K16ac staining was similar to that re- vitro myeloid differentiation of CD34+ mouse and human ported in other cell types and was far more uniform than progenitors, providing further support for this histone mod- in neutrophils (Figure 7A, B). Generally, H4K16ac stain- ification having a role in the process. However, as previ- ing and DAPI were less correlated in neutrophils than in in ously stated with respect to the cell differentiation genes, CD3+ (Pearson correlation = 0.03 and 0.68, respectively) differences in H4K16ac in a few apoptotic genes cannot (Figure 7B; Supplementary Table S10). In the second ap- explain the massive genome-wide H4K16 hyperacetylation proach, we used a chromatin accessibility assay, based on observed in peripheral blood neutrophils. Indeed, in vitro how accessible the DNA is to nucleases, to determine the differentiated neutrophils do not display the same global level of chromatin compaction in six of the previously iden- H4K16 hyperacetylation and staining pattern as that ob- tified H4K16ac-rich repeats which are fragmented at early served in peripheral blood neutrophils, something which stages of neutrophil apoptosis (Alu 257398, Alu 257394, should be taken into account in future studies using such L2 12327, L2 100294, Sat 101966 and Sat 28530) in neu- in vitro models. trophils and CD3+ cells obtained from four healthy individ- We next therefore considered the possibility that, in ad- uals. We further validated these results by FAIRE quanti- dition to gene regulation, H4K16ac plays another role in tative PCR (FAIRE-qPCR) analysis (Supplementary Fig- neutrophil apoptosis. The magnitude of the differences and ure S17). In general, there was considerable interindivid- the enrichment of H4K16ac in specific repeated DNA se- ual variability for most of the sequences analyzed in both quences in neutrophils pointed to a structural role for this myeloid and CD3+ cells. However, chromatin was always histone modification in the process. In line with this, our more accessible in neutrophils than in CD3+ in all the re- results revealed an enrichment of this histone mark in re- peated DNA regions analyzed (Figure 7C; Supplementary gions associated with early DNA fragmentation during neu- Figure S17). trophil apoptosis. We also showed that these DNA regions are more accessible in this cell type than in CD3+ T cells, intimating that H4K16ac plays a role in the establishment of a specific chromatin status that facilitates DNA frag- DISCUSSION mentation early during apoptosis. In support of this no- Neutrophils are defined by their multilobed nuclear archi- tion, in vitro experiments with nucleosomal arrays and com- tecture (75), and they display a characteristic pattern of putational modeling have shown that acetylation at lysine histone PTMs (76). Our initial observation that human 16 of histone H4 plays a central role in the regulation of PB neutrophils are hyperacetylated at H4K16 as compared higher-order chromatin structure and the interactions be- with other terminally differentiated types of blood cells led tween non-histone proteins and chromatin fiber (14,81,82). us to investigate the role of this histone PTM in myeloid However, the relationship between chromatin structure and cells in more detail. Genome-wide mapping of H4K16ac H4K16ac is far more complex, because, in vivo, this his- in peripheral neutrophils and CD3+ T cells confirmed its tone modification was also found to be associated with com- enrichment in the myeloid lineage. The results obtained in pacted chromatin and repression (83), and loss of H4K16ac vitro and in vivo in this work suggest that this PTM plays an does not alter higher-order chromatin compaction in mouse important role in blood cell survival and hematopoietic dif- embryonic stem cells (84). In line with this, although we ob- ferentiation, and they are in line with other previous results served the clear exclusion of H4K16ac from the large blocks Nucleic Acids Research, 2019, Vol. 47, No. 10 5033 A Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 B C Figure 7. Effect of H4K16ac enrichment at the repeated DNA fragmented early in neutrophil apoptosis. (A) Representative immunofluorescence images of nuclear distribution of H4K16ac in neutrophils and CD3+ T cells. On the right, line plots comparing DAPI and H4K16ac staining signals in several nuclear longitudinal positions. The red lines in the merged images indicates the position and direction of measurements. (B) Heat density scatterplots depicting the different distribution of the H4K16ac signal with respect to DAPI intensity in neutrophils (n = 72 cells from five donors) (upper panel) and CD3+ T cells (n = 62 cells from one donor) (lower panel) performed with 9198 and 5925 pairwise datapoints respectively. Colors represent the relative density of the number of measurements (data points) obtained across the different axes (blue – low density areas, red – high density areas). H4K16ac and DAPI intensities were positively correlated in lymphocyte cells (pcorr = 0.68), whereas neutrophils displayed a more sparse, non-correlated distribution (pcorr = 0.03) of H4K16ac-DAPI signals across the nucleus. (C) Chromatin accessibility assay in four donors showing that H4K16ac-enriched repetitive sequences in neutrophils are more associated with open chromatin than in CD3+ T cells. %FE: fold enrichment percentage. of heterochromatin of the nuclear periphery of peripheral acteristic chromatin structure/state prone to DNA frag- blood neutrophils, we did not find any global correlation mentation during apoptosis in neutrophils specifically. In- between this epigenetic modification and nuclear chromatin terestingly, H4K16ac-mediated chromatin relaxation might compaction in this type of cells. These data thus suggest that facilitate DNA repair of double-strand breaks (79). Taken H4K16ac function in chromatin conformation may well be together, these results could imply that chromatin relaxation highly dependent on the type of chromatin region and, con- by H4K16 acetylation plays a general role in DNA cleav- sequently, this should be explored further. age during programmed cell death, a possibility that should Although the genomic regions targeted during the apop- be studied in more detail in the future. Moreover, both the totic early DNA fragmentation are still poorly understood, H4K16ac-rich repeats cleaved during the first stages of neu- it has been proposed that the cleavage might occur prefer- trophil apoptosis and the chromatin state specifically en- entially at repeated DNA (85) and in regions of accessible riched in this histone PTM in neutrophils (state 10) are chromatin (86). Thus, it is possible that hyperacetylation of enriched in hypermethylated DNA. Thus, future research H4K16 at specific repeated DNA regions establishes a char- 5034 Nucleic Acids Research, 2019, Vol. 47, No. 10 should also determine the role of this latter chromatin state tion is still poorly understood. Our data suggest that, in ad- in the DNA cleavage occurring early on during apoptosis. dition to gene regulation, HDAC inhibitors might also pro- In summary, our study is in line with previous liter- mote programmed cell death through the establishment of a ature showing a positive correlation between high lev- higher-order chromatin structure prone to DNA fragmen- els of H4K16ac and apoptosis (17,18,34,35), which sug- tation during the first stages of apoptosis. In support of this gests that H4K16 hyperacetylation is directly involved in notion, we observed that the treatment of the HL60 cancer programmed cell death. However, our experiments which cell line with SAHA increased H4K16ac levels at specific showed a lack of global H4K16 hyperacetylation in apop- DNA repetitive regions identified in our ChIP-seq (Supple- tosis cell death induced by etoposide in myeloid cells indi- mentary Figure S18). As this possibility could have impor- cate either that the non-canonical structural role of this hi- tant clinical implications, future research is needed to con- stone PTM that is described in this manuscript is not essen- firm the role of H4K16 hyperacetylation of repeated DNA tial to induce programmed cell death, or that basal levels of on the mechanism of HDAC inhibitor treatment of cancer. H4K16ac might be sufficient. Making the story even more Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 complex, our data shows that ectopic downregulation of DATA AVAILABILITY KAT8 reduced etoposide-induced cell death in HL60 cells, Supplementary Tables generated during this study are in- which suggests that this histone modification could still play cluded in the online version of this manuscript and, with a role in the process. For example, downregulation of KAT8 the source code, are also available in the Zenodo public might impair the H4K16ac-dependent upregulation of spe- repository, DOI: 10.5281/zenodo.2425053. Raw data cor- cific proapoptotic genes, which is compatible with the fact responding to ChIP-seq experiments is uploaded in SRA that the downregulation of KAT8 in this work only induced database BioProject: PRJNA480011 with the accession a relatively small decrease in H4K16ac. A possible canon- numbers SUB4245394 and SUB5124343. ical role for H4K16ac in the regulation of apoptotic genes in HL60 cells exposed to etoposide is also compatible with SUPPLEMENTARY DATA the lack of global changes of H4K16ac in HL60 exposed to etoposide because, as previously reasoned, changes in Supplementary Data are available at NAR Online. this histone modification at a small number of specific genes would not be expected to be detected at a global level. Fi- ACKNOWLEDGEMENTS nally, it could also be possible that the effect of KAT8 down- We thank Ronnie Lendrum for editorial assistance. We also regulation in HL60 cells treated with etoposide in fact im- thank the Blood Bank of HUCA, and M. Rosario Rodicio pacts on the role of this histone acetyltransferase in the and Vanesa Garcı́a for their support. modulation of other regulators of apoptosis (87). Further research is thus still needed to determine the precise roles FUNDING of KAT8 and H4K16ac in etoposide-induced cell death in HL60 cells. Plan Nacional de I+D+I co-funding FEDER [PI15/00892 Further research will also be needed to determine and PI18/01527 to M.F.F. and A.F.F.; PI16/01318 and whether H4K16 hyperacetylation is required for apopto- PI14/01244 to C.L.]; ISCIII-Subdirección General de sis in other cell types. However, the fact that H4K16 hy- Evaluacicción y Fomento de la Investigación, and Plan poacetylation underpins resistance to the topoisomerase I Nacional de I+D+I 2008–2011/FEDER [CP11/00131 inhibitor irinotecan in colon cancer cells lends support to to A.F.F.]; IUOPA (to G.F.B. and M.I.S.); Fundación the notion that this histone modification might play a gen- Cientı́fica de la AECC (to R.G.U.); Ministry of Econ- eral role in apoptosis (35). Interestingly, our results showed omy and Competitiveness Juan de la Cierva postdoc- that tumoral HL60 cells have lower H4K16 acetylation toral fellowships [FJCI-2015-26965 to J.R.T., IJCI-2015- levels than their non-tumorigenic terminally differentiated 23316 to V.L.]; Fundación Ramón Areces (to M.F.F); peripheral blood cell counterparts (Supplementary Figure FICYT (to E.G.T., M.G.G., A.C.); Asturias Regional S8A), which is in line with previously published data show- Government [GRUPIN14-052 to M.F.F.]; Gobierno del ing that H4K16ac is frequently down-regulated in cancer Principado de Asturias, PCTI-Plan de Ciencia, Tec- (21–24), although the underlying molecular mechanisms nologı́a e Innovación co-funding Fondos FEDER (grant were largely unclear. Our results suggest that myeloid tu- number IDI/2018/146 to M.F.F. and IDI/2018/144 to mors display low H4K16ac levels, because they do not un- C.L.); Deutsche Forschungsgemeinschaft (DFG) [SFB960 dergo the massive hyperacetylation that occurs during the to A.V.G., R.D.]; European Research Council [CoG-2014- last stages of myeloid differentiation. Moreover, the lack 646903]; Spanish Ministry of Economy and Competitive- of H4K16 hyperacetylation could even favor the resistance ness [SAF-SAF2013-43065 to P.M.]; Asociación Española of these tumor cells to apoptosis, one of the hallmarks of Contra el Cáncer [AECC-CI-2015]; FERO Foundation, cancer (88). In line with this, our results might help to un- and the ISCIII [PI14-01191 to C.B.]; P.M. acknowl- derstand the mechanism of action of HDAC inhibitors, a edges financial support from The Obra Social La Caixa- family of small molecules with antitumor activity that can Fundaciò Josep Carreras and The Generalitat de Catalunya induce histone hyperacetylation and apoptosis in myeloid (SGR330). P.M. an investigator from the Spanish Cell cells (88). Although it was proposed that their mechanism Therapy cooperative network (TERCEL). The IUOPA is of action is mediated, at least in part through the reac- supported by the Obra Social Liberbank-Cajastur, Spain. tivation of proapoptotic genes (18,89), these compounds Funding for open access charge: Plan Nacional de I+D+I have many pleiotropic effects and their mechanism of ac- co-funding FEDER [PI18/01527]. Nucleic Acids Research, 2019, Vol. 47, No. 10 5035 Conflict of interest statement. None declared. deacetylase 2 increases apoptosis and p21Cip1/WAF1 expression, independent of histone deacetylase 1. Cell Death Differ., 12, 395–404. 20. Zhang,J. and Zhong,Q. (2014) Histone deacetylase inhibitors and cell death. Cell Mol. Life Sci., 71, 3885–3901. REFERENCES 21. Elsheikh,S.E., Green,A.R., Rakha,E.A., Powe,D.G., Ahmed,R.A., 1. Bannister,A.J. and Kouzarides,T. (2011) Regulation of chromatin by Collins,H.M., Soria,D., Garibaldi,J.M., Paish,C.E., Ammar,A.A. histone modifications. Cell Res., 21, 381–395. et al. (2009) Global histone modifications in breast cancer correlate 2. Alvarez-Errico,D., Vento-Tormo,R., Sieweke,M. and Ballestar,E. with tumor phenotypes, prognostic factors, and patient outcome. (2015) Epigenetic control of myeloid cell differentiation, identity and Cancer Res., 69, 3802–3809. function. Nat. Rev. Immunol., 15, 7–17. 22. Fraga,M.F., Ballestar,E., Villar-Garea,A., Boix-Chornet,M., 3. Kouzarides,T. (2007) Chromatin modifications and their function. Espada,J., Schotta,G., Bonaldi,T., Haydon,C., Ropero,S., Petrie,K. Cell, 128, 693–705. et al. (2005) Loss of acetylation at Lys16 and trimethylation at Lys20 4. Rice,K.L., Hormaeche,I. and Licht,J.D. (2007) Epigenetic regulation of histone H4 is a common hallmark of human cancer. Nat. Genet., of normal and malignant hematopoiesis. Oncogene, 26, 6697–6714. 37, 391–400. 5. Dou,Y., Milne,T.A., Tackett,A.J., Smith,E.R., Fukuda,A., 23. Pfister,S., Rea,S., Taipale,M., Mendrzyk,F., Straub,B., Ittrich,C., Thuerigen,O., Sinn,H.P., Akhtar,A. and Lichter,P. (2008) The histone Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Wysocka,J., Allis,C.D., Chait,B.T., Hess,J.L. and Roeder,R.G. (2005) Physical association and coordinate function of the H3 K4 acetyltransferase hMOF is frequently downregulated in primary methyltransferase MLL1 and the H4 K16 acetyltransferase MOF. breast carcinoma and medulloblastoma and constitutes a biomarker Cell, 121, 873–885. for clinical outcome in medulloblastoma. Int. J. Cancer, 122, 6. Singh,S.K., Williams,C.A., Klarmann,K., Burkett,S.S., Keller,J.R. 1207–1213. and Oberdoerffer,P. (2013) Sirt1 ablation promotes stress-induced 24. Van Den Broeck,A., Brambilla,E., Moro-Sibilot,D., Lantuejoul,S., loss of epigenetic and genomic hematopoietic stem and progenitor Brambilla,C., Eymin,B., Khochbin,S. and Gazzeri,S. (2008) Loss of cell maintenance. J. Exp. Med., 210, 987–1001. histone H4K20 trimethylation occurs in preneoplasia and influences 7. Ramos-Mejia,V., Navarro-Montero,O., Ayllon,V., Bueno,C., prognosis of non-small cell lung cancer. Clin. Cancer Res., 14, Romero,T., Real,P.J. and Menendez,P. (2014) HOXA9 promotes 7237–7245. hematopoietic commitment of human embryonic stem cells. Blood, 25. Cai,M., Hu,Z., Liu,J., Gao,J., Tan,M., Zhang,D., Zhu,L., Liu,S., 124, 3065–3075. Hou,R. and Lin,B. (2015) Expression of hMOF in different ovarian 8. Taipale,M., Rea,S., Richter,K., Vilar,A., Lichter,P., Imhof,A. and tissues and its effects on ovarian cancer prognosis. Oncol Rep., 33, Akhtar,A. (2005) hMOF histone acetyltransferase is required for 685–692. histone H4 lysine 16 acetylation in mammalian cells. Mol. Cell Biol., 26. Cao,L., Zhu,L., Yang,J., Su,J., Ni,J., Du,Y., Liu,D., Wang,Y., 25, 6798–6810. Wang,F., Jin,J. et al. (2014) Correlation of low expression of hMOF 9. Mishra,B.P., Zaffuto,K.M., Artinger,E.L., Org,T., Mikkola,H.K., with clinicopathological features of colorectal carcinoma, gastric Cheng,C., Djabali,M. and Ernst,P. (2014) The histone cancer and renal cell carcinoma. Int. J. Oncol., 44, 1207–1214. methyltransferase activity of MLL1 is dispensable for hematopoiesis 27. Liu,N., Zhang,R., Zhao,X., Su,J., Bian,X., Ni,J., Yue,Y., Cai,Y. and and leukemogenesis. Cell Rep., 7, 1239–1247. Jin,J. (2013) A potential diagnostic marker for ovarian cancer: 10. Ou,X., Chae,H.D., Wang,R.H., Shelley,W.C., Cooper,S., Taylor,T., involvement of the histone acetyltransferase, human males absent on Kim,Y.J., Deng,C.X., Yoder,M.C. and Broxmeyer,H.E. (2011) SIRT1 the first. Oncol Lett., 6, 393–400. deficiency compromises mouse embryonic stem cell hematopoietic 28. Wang,Y., Zhang,R., Wu,D., Lu,Z., Sun,W., Cai,Y., Wang,C. and differentiation, and embryonic and adult hematopoiesis in the mouse. Jin,J. (2013) Epigenetic change in kidney tumor: downregulation of Blood, 117, 440–450. histone acetyltransferase MYST1 in human renal cell carcinoma. J. 11. Wada,T., Kikuchi,J., Nishimura,N., Shimizu,R., Kitamura,T. and Exp. Clin. Cancer Res., 32, 8. Furukawa,Y. (2009) Expression levels of histone deacetylases 29. Zhu,L., Yang,J., Zhao,L., Yu,X., Wang,L., Wang,F., Cai,Y. and Jin,J. determine the cell fate of hematopoietic progenitors. J. Biol. Chem., (2015) Expression of hMOF, but not HDAC4, is responsible for the 284, 30673–30683. global histone H4K16 acetylation in gastric carcinoma. Int. J. Oncol., 12. Heideman,M.R., Lancini,C., Proost,N., Yanover,E., Jacobs,H. and 46, 2535–2545. Dannenberg,J.H. (2014) Sin3a-associated Hdac1 and Hdac2 are 30. Choi,J.H., Kwon,H.J., Yoon,B.I., Kim,J.H., Han,S.U., Joo,H.J. and essential for hematopoietic stem cell homeostasis and contribute Kim,D.Y. (2001) Expression profile of histone deacetylase 1 in gastric differentially to hematopoiesis. Haematologica, 99, 1292–1303. cancer tissues. Jpn J. Cancer Res., 92, 1300–1304. 13. Thomas,T., Dixon,M.P., Kueh,A.J. and Voss,A.K. (2008) Mof 31. Halkidou,K., Gaughan,L., Cook,S., Leung,H.Y., Neal,D.E. and (MYST1 or KAT8) is essential for progression of embryonic Robson,C.N. (2004) Upregulation and nuclear recruitment of development past the blastocyst stage and required for normal HDAC1 in hormone refractory prostate cancer. Prostate, 59, chromatin architecture. Mol. Cell Biol., 28, 5093–5105. 177–189. 14. Shogren-Knaak,M., Ishii,H., Sun,J.M., Pazin,M.J., Davie,J.R. and 32. Song,J., Noh,J.H., Lee,J.H., Eun,J.W., Ahn,Y.M., Kim,S.Y., Lee,S.H., Peterson,C.L. (2006) Histone H4-K16 acetylation controls chromatin Park,W.S., Yoo,N.J., Lee,J.Y. et al. (2005) Increased expression of structure and protein interactions. Science, 311, 844–847. histone deacetylase 2 is found in human gastric cancer. APMIS, 113, 15. Horikoshi,N., Kumar,P., Sharma,G.G., Chen,M., Hunt,C.R., 264–268. Westover,K., Chowdhury,S. and Pandita,T.K. (2013) Genome-wide 33. Zhang,Z., Yamashita,H., Toyama,T., Sugiura,H., Ando,Y., Mita,K., distribution of histone H4 Lysine 16 acetylation sites and their Hamaguchi,M., Hara,Y., Kobayashi,S. and Iwase,H. (2005) relationship to gene expression. Genome Integr., 4, 3. Quantitation of HDAC1 mRNA expression in invasive carcinoma of 16. Katoh,H., Qin,Z.S., Liu,R., Wang,L., Li,W., Li,X., Wu,L., Du,Z., the breast*. Breast Cancer Res. Treat., 94, 11–16. Lyons,R., Liu,C.G. et al. (2011) FOXP3 orchestrates H4K16 34. Hajji,N., Wallenborg,K., Vlachos,P., Fullgrabe,J., Hermanson,O. and acetylation and H3K4 trimethylation for activation of multiple genes Joseph,B. (2010) Opposing effects of hMOF and SIRT1 on H4K16 by recruiting MOF and causing displacement of PLU-1. Mol. Cell, acetylation and the sensitivity to the topoisomerase II inhibitor 44, 770–784. etoposide. Oncogene, 29, 2192–2204. 17. Kapoor-Vazirani,P., Kagey,J.D., Powell,D.R. and Vertino,P.M. (2008) 35. Meisenberg,C., Ashour,M.E., El-Shafie,L., Liao,C., Hodgson,A., Role of hMOF-dependent histone H4 lysine 16 acetylation in the Pilborough,A., Khurram,S.A., Downs,J.A., Ward,S.E. and maintenance of TMS1/ASC gene activity. Cancer Res., 68, El-Khamisy,S.F. (2016) Epigenetic changes in histone acetylation 6810–6821. underpin resistance to the topoisomerase I inhibitor irinotecan. 18. Lara,E., Mai,A., Calvanese,V., Altucci,L., Lopez-Nieva,P., Nucleic Acids Res., 45, 1159–1176. Martinez-Chantar,M.L., Varela-Rey,M., Rotili,D., Nebbioso,A., 36. Fullgrabe,J., Lynch-Day,M.A., Heldring,N., Li,W., Struijk,R.B., Ropero,S. et al. (2009) Salermide, a Sirtuin inhibitor with a strong Ma,Q., Hermanson,O., Rosenfeld,M.G., Klionsky,D.J. and Joseph,B. cancer-specific proapoptotic effect. Oncogene, 28, 781–791. (2013) The histone H4 lysine 16 acetyltransferase hMOF regulates 19. Huang,B.H., Laban,M., Leung,C.H., Lee,L., Lee,C.K., the outcome of autophagy. Nature, 500, 468–471. Salto-Tellez,M., Raju,G.C. and Hooi,S.C. (2005) Inhibition of histone 5036 Nucleic Acids Research, 2019, Vol. 47, No. 10 37. Sharma,G.G., So,S., Gupta,A., Kumar,R., Cayrou,C., 57. Simon,J.M., Giresi,P.G., Davis,I.J. and Lieb,J.D. (2013) A detailed Avvakumov,N., Bhadra,U., Pandita,R.K., Porteus,M.H., Chen,D.J. protocol for formaldehyde-assisted isolation of regulatory elements et al. (2010) MOF and histone H4 acetylation at lysine 16 are critical (FAIRE). Curr. Protoc. Mol. Biol., for DNA damage response and double-strand break repair. Mol. Cell doi:10.1002/0471142727.mb2126s102. Biol., 30, 3582–3595. 58. Urdinguio,R.G., Lopez-Serra,L., Lopez-Nieva,P., Alaminos,M., 38. Li,X., Corsa,C.A., Pan,P.W., Wu,L., Ferguson,D., Yu,X., Min,J. and Diaz-Uriarte,R., Fernandez,A.F. and Esteller,M. (2008) Mecp2-null Dou,Y. (2010) MOF and H4 K16 acetylation play important roles in mice provide new neuronal targets for Rett syndrome. PLoS One, 3, DNA damage repair by modulating recruitment of DNA damage e3669. repair protein Mdc1. Mol. Cell Biol., 30, 5335–5347. 59. Zhang,K., Williams,K.E., Huang,L., Yau,P., Siino,J.S., 39. Luo,H.R. and Loison,F. (2008) Constitutive neutrophil apoptosis: Bradbury,E.M., Jones,P.R., Minch,M.J. and Burlingame,A.L. (2002) mechanisms and regulation. Am. J. Hematol., 83, 288–295. Histone acetylation and deacetylation: identification of acetylation 40. Witko-Sarsat,V., Pederzoli-Ribeil,M., Hirsch,E., Sozzani,S. and and methylation sites of HeLa histone H4 by mass spectrometry. Cassatella,M.A. (2011) Regulating neutrophil apoptosis: new players Mol. Cell Proteomics, 1, 500–508. enter the game. Trends Immunol., 32, 117–124. 60. Galasinski,S.C., Resing,K.A. and Ahn,N.G. (2003) Protein mass 41. Lindner,H., Helliger,W., Dirschlmayer,A., Jaquemar,M. and analysis of histones. Methods, 31, 3–11. Puschendorf,B. (1992) High-performance capillary electrophoresis of 61. Turner,B.M., O’Neill,L.P. and Allan,I.M. (1989) Histone H4 Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 core histones and their acetylated modified derivatives. Biochem. J., acetylation in human cells. Frequency of acetylation at different sites 283, 467–471. defined by immunolabeling with site-specific antibodies. FEBS Lett., 42. Sarg,B., Koutzamani,E., Helliger,W., Rundquist,I. and Lindner,H.H. 253, 141–145. (2002) Postsynthetic trimethylation of histone H4 at lysine 20 in 62. Feng,J., Liu,T. and Zhang,Y. (2011) Using MACS to identify peaks mammalian tissues is associated with aging. J. Biol. Chem., 277, from ChIP-Seq data. Curr. Protoc. Bioinformatics, 39195–39201. doi:10.1002/0471250953.bi0214s34. 43. Villar-Garea,A., Israel,L. and Imhof,A. (2008) Analysis of histone 63. Ernst,J. and Kellis,M. (2012) ChromHMM: automating modifications by mass spectrometry. Curr. Protoc. Protein Sci., chromatin-state discovery and characterization. Nat. Methods, 9, doi:10.1002/0471140864.ps1410s51. 215–216. 44. Strohalm,M., Hassman,M., Kosata,B. and Kodicek,M. (2008) 64. Purton,L.E., Bernstein,I.D. and Collins,S.J. (1999) All-trans retinoic mMass data miner: an open source alternative for mass spectrometric acid delays the differentiation of primitive hematopoietic precursors data analysis. Rapid Commun. Mass Spectrom., 22, 905–908. (lin-c-kit+Sca-1(+)) while enhancing the terminal maturation of 45. Strohalm,M., Kavan,D., Novak,P., Volny,M. and Havlicek,V. (2010) committed granulocyte/monocyte progenitors. Blood, 94, 483–495. mMass 3: a cross-platform software environment for precise analysis 65. Breitman,T.R., Selonick,S.E. and Collins,S.J. (1980) Induction of of mass spectrometric data. Anal. Chem., 82, 4648–4651. differentiation of the human promyelocytic leukemia cell line (HL-60) 46. Liu,P., Jenkins,N.A. and Copeland,N.G. (2003) A highly efficient by retinoic acid. Proc. Natl. Acad. Sci. U.S.A., 77, 2936–2940. recombineering-based method for generating conditional knockout 66. Nishizawa,Y., Saeki,K., Hirai,H., Yazaki,Y., Takaku,F. and Yuo,A. mutations. Genome Res., 13, 476–484. (1998) Potent inhibition of cell density-dependent apoptosis and 47. Sheikh,B.N., Bechtel-Walz,W., Lucci,J., Karpiuk,O., Hild,I., enhancement of survival by dimethyl sulfoxide in human myeloblastic Hartleben,B., Vornweg,J., Helmstadter,M., Sahyoun,A.H., HL-60 cells. J. Cell Physiol., 174, 135–143. Bhardwaj,V. et al. (2016) MOF maintains transcriptional programs 67. Saeki,K., Yuo,A., Kato,M., Miyazono,K., Yazaki,Y. and Takaku,F. regulating cellular stress response. Oncogene, 35, 2698–2710. (1997) Cell density-dependent apoptosis in HL-60 cells, which is 48. Peng,L., Ling,H., Yuan,Z., Fang,B., Bloom,G., Fukasawa,K., mediated by an unknown soluble factor, is inhibited by transforming Koomen,J., Chen,J., Lane,W.S. and Seto,E. (2012) SIRT1 negatively growth factor beta1 and overexpression of Bcl-2. J. Biol. Chem., 272, regulates the activities, functions, and protein levels of hMOF and 20003–20010. TIP60. Mol. Cell Biol., 32, 2823–2836. 68. Vaquero,A., Scher,M.B., Lee,D.H., Sutton,A., Cheng,H.L., Alt,F.W., 49. Crissman,H.A. and Steinkamp,J.A. (1973) Rapid, simultaneous Serrano,L., Sternglanz,R. and Reinberg,D. (2006) SirT2 is a histone measurement of DNA, protein, and cell volume in single cells from deacetylase with preference for histone H4 Lys 16 during mitosis. large mammalian cell populations. J. Cell Biol., 59, 766–771. Genes Dev., 20, 1256–1261. 50. Krishan,A. (1975) Rapid flow cytofluorometric analysis of 69. Marks,P.A. and Breslow,R. (2007) Dimethyl sulfoxide to vorinostat: mammalian cell cycle by propidium iodide staining. J. Cell Biol., 66, development of this histone deacetylase inhibitor as an anticancer 188–193. drug. Nat. Biotechnol., 25, 84–90. 51. Stegmaier,K., Ross,K.N., Colavito,S.A., O’Malley,S., Stockwell,B.R. 70. Dubrez,L., Goldwasser,F., Genne,P., Pommier,Y. and Solary,E. and Golub,T.R. (2004) Gene expression-based high-throughput (1995) The role of cell cycle regulation and apoptosis triggering in screening(GE-HTS) and application to leukemia differentiation. Nat. determining the sensitivity of leukemic cells to topoisomerase I and II Genet., 36, 257–263. inhibitors. Leukemia, 9, 1013–1024. 52. Bueno,C., Catalina,P., Melen,G.J., Montes,R., Sanchez,L., Ligero,G., 71. Tyurina,Y.Y., Serinkan,F.B., Tyurin,V.A., Kini,V., Yalowich,J.C., Garcia-Perez,J.L. and Menendez,P. (2009) Etoposide induces MLL Schroit,A.J., Fadeel,B. and Kagan,V.E. (2004) Lipid antioxidant, rearrangements and other chromosomal abnormalities in human etoposide, inhibits phosphatidylserine externalization and embryonic stem cells. Carcinogenesis, 30, 1628–1637. macrophage clearance of apoptotic cells by preventing 53. Bueno,C., Montes,R., Martin,L., Prat,I., Hernandez,M.C., Orfao,A. phosphatidylserine oxidation. J. Biol. Chem., 279, 6056–6064. and Menendez,P. (2008) NG2 antigen is expressed in CD34+ HPCs 72. Boix-Chornet,M., Fraga,M.F., Villar-Garea,A., Caballero,R., and plasmacytoid dendritic cell precursors: is NG2 expression in Espada,J., Nunez,A., Casado,J., Largo,C., Casal,J.I., Cigudosa,J.C. leukemia dependent on the target cell where leukemogenesis is et al. (2006) Release of hypoacetylated and trimethylated histone H4 triggered? Leukemia, 22, 1475–1478. is an epigenetic marker of early apoptosis. J. Biol. Chem., 281, 54. Prieto,C., Lopez-Millan,B., Roca-Ho,H., Stam,R.W., 13540–13547. Romero-Moya,D., Rodriguez-Baena,F.J., Sanjuan-Pla,A., Ayllon,V., 73. Brinkmann,V., Reichard,U., Goosmann,C., Fauler,B., Uhlemann,Y., Ramirez,M., Bardini,M. et al. (2018) NG2 antigen is involved in Weiss,D.S., Weinrauch,Y. and Zychlinsky,A. (2004) Neutrophil leukemia invasiveness and central nervous system infiltration in extracellular traps kill bacteria. Science, 303, 1532–1535. MLL-rearranged infant B-ALL. Leukemia, 32, 633–644. 74. Oberhammer,F., Wilson,J.W., Dive,C., Morris,I.D., Hickman,J.A., 55. Pryde,J.G., Walker,A., Rossi,A.G., Hannah,S. and Haslett,C. (2000) Wakeling,A.E., Walker,P.R. and Sikorska,M. (1993) Apoptotic death Temperature-dependent arrest of neutrophil apoptosis. Failure of Bax in epithelial cells: cleavage of DNA to 300 and/or 50 kb fragments insertion into mitochondria at 15 degrees C prevents the release of prior to or in the absence of internucleosomal fragmentation. EMBO cytochrome c. J. Biol. Chem., 275, 33574–33584. J., 12, 3679–3684. 56. Simon,J.M., Giresi,P.G., Davis,I.J. and Lieb,J.D. (2012) Using 75. Zhu,Y., Gong,K., Denholtz,M., Chandra,V., Kamps,M.P., Alber,F. formaldehyde-assisted isolation of regulatory elements (FAIRE) to and Murre,C. (2017) Comprehensive characterization of neutrophil isolate active regulatory DNA. Nat. Protoc., 7, 256–267. genome topology. Genes Dev, 31, 141–153. Nucleic Acids Research, 2019, Vol. 47, No. 10 5037 76. Olins,D.E. and Olins,A.L. (2005) Granulocyte heterochromatin: 83. Zhou,Y. and Grummt,I. (2005) The PHD finger/bromodomain of defining the epigenome. BMC Cell Biol., 6, 39. NoRC interacts with acetylated histone H4K16 and is sufficient for 77. Gupta,A., Hunt,C.R., Pandita,R.K., Pae,J., Komal,K., Singh,M., rDNA silencing. Curr. Biol., 15, 1434–1438. Shay,J.W., Kumar,R., Ariizumi,K., Horikoshi,N. et al. (2013) 84. Taylor,G.C., Eskeland,R., Hekimoglu-Balkan,B., Pradeepa,M.M. T-cell-specific deletion of Mof blocks their differentiation and results and Bickmore,W.A. (2013) H4K16 acetylation marks active genes in genomic instability in mice. Mutagenesis, 28, 263–270. and enhancers of embryonic stem cells, but does not alter chromatin 78. Valerio,D.G., Xu,H., Eisold,M.E., Woolthuis,C.M., Pandita,T.K. compaction. Genome Res., 23, 2053–2065. and Armstrong,S.A. (2016) Histone acetyltransferase activity of 85. Khodarev,N.N., Bennett,T., Shearing,N., Sokolova,I., Koudelik,J., MOF is required for adult but not early fetal hematopoiesis in mice. Walter,S., Villalobos,M. and Vaughan,A.T. (2000) LINE L1 Blood, 129, 48–59. retrotransposable element is targeted during the initial stages of 79. Krishnan,V., Chow,M.Z., Wang,Z., Zhang,L., Liu,B., Liu,X. and apoptotic DNA fragmentation. J. Cell Biochem., 79, 486–495. Zhou,Z. (2011) Histone H4 lysine 16 hypoacetylation is associated 86. Di Filippo,M. and Bernardi,G. (2009) The early apoptotic DNA with defective DNA repair and premature senescence in fragmentation targets a small number of specific open chromatin Zmpste24-deficient mice. Proc. Natl. Acad. Sci. U.S.A., 108, regions. PLoS One, 4, e5010. 12325–12330. 87. Rea,S., Xouri,G. and Akhtar,A. (2007) Males absent on the first 80. Osorio,F.G., Varela,I., Lara,E., Puente,X.S., Espada,J., Santoro,R., (MOF): from flies to humans. Oncogene, 26, 5385–5394. Downloaded from https://academic.oup.com/nar/article/47/10/5016/5421813 by guest on 20 February 2021 Freije,J.M., Fraga,M.F. and Lopez-Otin,C. (2010) Nuclear envelope 88. Cotter,T.G. (2009) Apoptosis and cancer: the genesis of a research alterations generate an aging-like epigenetic pattern in mice deficient field. Nat. Rev. Cancer, 9, 501–507. in Zmpste24 metalloprotease. Aging Cell, 9, 947–957. 89. Conway,K.E., McConnell,B.B., Bowring,C.E., Donald,C.D., 81. Yang,D. and Arya,G. (2011) Structure and binding of the H4 histone Warren,S.T. and Vertino,P.M. (2000) TMS1, a novel proapoptotic tail and the effects of lysine 16 acetylation. Phys. Chem. Chem. Phys., caspase recruitment domain protein, is a target of 13, 2911–2921. methylation-induced gene silencing in human breast cancers. Cancer 82. Zhang,B. and Wolynes,P.G. (2017) Genomic energy landscapes. Res., 60, 6236–6242. Biophys. J, 112, 427–433.