Loss of TIP60 (KAT5) abolishes H2AZ lysine 7 acetylation and causes p53, INK4A, and ARF-independent cell cycle arrest

Histone acetylation is essential for initiating and maintaining a permissive chromatin conformation and gene transcription. Dysregulation of histone acetylation can contribute to tumorigenesis and metastasis. Using inducible cre-recombinase and CRISPR/Cas9-mediated deletion, we investigated the roles of the histone lysine acetyltransferase TIP60 (KAT5/HTATIP) in human cells, mouse cells, and mouse embryos. We found that loss of TIP60 caused complete cell growth arrest. In the absence of TIP60, chromosomes failed to align in a metaphase plate during mitosis. In some TIP60 deleted cells, endoreplication occurred instead. In contrast, cell survival was not affected. Remarkably, the cell growth arrest caused by loss of TIP60 was independent of the tumor suppressors p53, INK4A and ARF. TIP60 was found to be essential for the acetylation of H2AZ, specifically at lysine 7. The mRNA levels of 6236 human and 8238 mouse genes, including many metabolism genes, were dependent on TIP60. Among the top 50 differentially expressed genes, over 90% were downregulated in cells lacking TIP60, supporting a role for TIP60 as a key co-activator of transcription. We propose a primary role of TIP60 in H2AZ lysine 7 acetylation and transcriptional activation, and that this fundamental role is essential for cell proliferation. Growth arrest independent of major tumor suppressors suggests TIP60 as a potential anti-cancer drug target.


INTRODUCTION
Chromatin modifications affect DNA accessibility and are essential for the activation of gene expression [1][2][3][4][5][6]. Histone lysine acetylation mediated by lysine acetyltransferases (KATs) is one of the most abundant histone modifications and is generally associated with actively transcribed gene loci [7][8][9]. Genetic alterations of genes encoding histone acetyltransferases are found in a wide range of cancers [10]. Targeting the histone acetyltransferases KAT6A and KAT7, members of the MYST family of histone acetyltransferases has emerged as a possible avenue for the treatment of cancer [11,12].
In the context of HIV infection, TIP60 appears to promote cell death, and HIV-1 tat binding to TIP60 facilitates its degradation resulting in impaired TIP60-mediated apoptosis [33,34]. Overexpression of catalytically inactive TIP60 leads to defects in DNA repair and decreased apoptosis upon γ-irradiation [14]. Similarly, siRNA knockdown of TIP60 impairs the UV-induced DNA damage response and apoptosis [35]. Relevant to the regulation of apoptosis, TIP60 has been reported to acetylate p53 at lysine 120, which is thought to be required for p53-induced apoptosis [20,21]. In contrast, germline deletion of Tip60 in mice causes developmental arrest and apoptotic cell death in peri-implantation mouse embryos [36], indicating that in this instance TIP60 is required for cell survival. The conflicting reports on the role of TIP60 in promoting apoptosis or cell survival have not been resolved.
TIP60 appears to either promote or retard cancer progression. Heterozygous loss of Tip60 accelerates tumor progression in mouse models of breast cancer [37] and lymphoma [38]. Reduced levels of TIP60 are associated with metastatic colorectal cancer [39], metastatic gastric cancer [40], and with poor outcomes in melanoma patients [41]. In contrast, in prostate cancer, high levels of nuclear TIP60 correlate with hormone treatment-refractory disease [42]. The seemingly contradictory reports of TIP60 in various cancers may reflect context-specific functions, affected by parameters such as cancer type and stage, or they may result from changes in TIP60 levels/function that are a consequence rather than a cause of certain oncogenic processes.
Here we used inducible genetic deletion of the TIP60 gene in human and mouse cells to investigate its role in cell proliferation, cell death, histone acetylation, and transcriptional regulation.

MATERIALS AND METHODS Mice
Tip60 fl/+ mice were generated by CRISPR/Cas9 on a C57BL/6 J background, using techniques previously described [43,44]. Briefly, two sgRNAs of the sequence GTACGGAGATGATCCGGGCG and TGGAAGCTACGCCTGCAACT were used to insert loxP sites flanking exons 3 and 4 of the Tip60 gene ( Figure S1). The deletion of exon 3 and 4 is predicted to result in a frameshift and disruption of proper TIP60 protein synthesis. Successful targeting was confirmed by sequencing. Tip60 fl/+ mice were crossed to Rosa26-CreERT2 transgenic mice [45] on a C57BL/6 background and used to isolate Tip60 fl/fl ;ERT2 MEFs.
Male and female young adult mice were used. Mice were genotyped by PCR; details are described in the supplement. Genotyping primers are listed in the supplement Methods Table 1.
Experiments were performed with the approval of the Walter and Eliza Hall Institute for Medical Research Animal Ethics Committee and according to the Australian code of practice for the care and use of animals for scientific purposes.

MEF isolation
Briefly, E13.5 embryos were dissected and washed in PBS. Head, extremities, and liver were removed, and the torso was collected in MEF medium, supplemented with 1% (v/v) penicillin and streptomycin (Gibco, 15140), and minced by pipetting several times consecutively through 18 G, 21 G, and 24 G needles. The cells were centrifuged with low speed (500 × g) for 5 min and resuspended in warm MEF medium supplemented with penicillin and streptomycin and plated onto gelatin-coated 10 cm Petri dishes.
Inducible Tip60 CRISPR knockout cell lines Dox-inducible CRISPR knockout cell lines were generated as described [48,49]. Two single guide RNAs, sgRNA1 GCCGGCACCGCCTCAAGCCG and sgRNA2 CTGCCTCCCTACCAGCGCCG, were used to target exons 8 and 11 of the human TIP60 gene and murine Tip60 gene. Cells were treated with 1 μg/ml dox for the indicated times. CRISPR indels were analyzed using the primers displayed in the supplement Methods Table 2. Indel frequencies were determined by NGS and calculated as a proportion of sequences with indels of all sequences obtained. Only sequences with at least 25 reads were included in this calculation.

SDS-PAGE and immunoblotting
Detailed protocols for acid histone extraction and quantification, cell fractionation protein extraction, and SDS-PAGE and immunoblotting can be found in the supplement. Histones were acid extracted, as described [50] with slight modifications. Cells were fractionated using a subcellular protein fractionation kit for cultured cells (Thermo Scientific, 78840) according to the manufacturer's instructions. SDS-PAGE and immunoblotting were performed using the Li-Cor system (LI-COR Biosciences), details are described in the supplement. Antibodies are listed in the supplement Methods Table 3.
SA-β-galactosidase staining by X-gal SA-β-galactosidase staining was performed with a senescence β-galactosidase staining kit (CST, 9860) according to the manufacturer's instructions. SA-β-galactosidase positive and negative cells were counted in triplicate in three images with at least 300 nuclei per replicate using ImageJ. Cells were irradiated with 32 Gy and cultured for 5 days prior to staining as positive controls.

RT-qPCR
RT-qPCRs were performed in triplicate. RNA was isolated using RNeasy Kit (Qiagen, 74106) according to the manufacturer's instructions. Details of the RT-qPCR procedure can be found in the supplement. RT-qPCR primers are listed in the supplement Methods Table 4.

RNA-seq
RNA-seq was performed on Tip60 fl/fl ;ERT2 and Tip60 +/+ ;ERT2 MEFs induced with 4-OHT for 3 days and 5 days, as well as HEK293g1/C9, HEK293g2/C9, and guide-only HEK293g1 and HEK293g2 controls induced with dox for 3 days. RNA-seq was performed on 2-3 replicates each of U2OSg1/C9, U2OSg2/C9, and guide-only U2OSg1 and U2OSg2 control cells, in total 6 replicates of TIP60 deleted cells and 5 replicates of control cells, induced with dox for 4 days and with D. melanogaster S2 cell spike-in. Details of the RNA-seq procedure can be found in the supplement.

CUT&Tag sequencing
CUT&Tag was performed on replicate cultures and 100,000 cells per sample, 43,000 human cells each of U2OSg1/C9 and U2OSg2/C9 either treated with doxycycline for 4 days (KO) or untreated (CTL) and 57,000 D. melanogaster cells as a spike-in, in total two replicates of TIP60 deleted cells and two replicates of control cells, as described by the Henikoff laboratory [51] with slight modifications and described in detail in the supplement. The CUT&Tag method produces results similar to ChIP-seq. Briefly, cells are bound by magnetic beads, permeabilized, and subjected with intervening wash steps to incubation with a primary antibody against the target of interest (e.g., H2AZ), followed by a secondary antibody incubation and then pAG-conjugated transposase enzyme (pAG-Tn5). pAG-Tn5 binds to the antibody (target) and is activated. Tn5 cuts the DNA and inserts DNA oligomers (previously loaded on the Tn5) site-specifically into the genome and thereby produces genomic fragments near the position of the target protein. Oligomer-tagged ("tagmented) fragments are recovered, amplified, and sequenced. D. melanogaster S2 cells as a spike-in was used to enable normalization and quantification of potential global changes in read coverage between control and TIP60 deleted cells.

Flow cytometry analysis
For flow cytometry analysis, including BrdU incorporation, live/cell death cell assay, and γH2AX staining, cells were collected, stained, and analyzed as described in the supplement.

Live-cell imaging
Cells were imaged using a live-cell imaging system (IncuCyte S3) at 37°C and 5% CO 2 with a CMOS camera and ×10 objective for up to 4 days. For 3D confocal live-cell time-lapse imaging, cells were grown in DKSH Ibidi 8-well chamber (DKSH, 80826) and stained with 1 μM SiR-DNA stain and 1 μM verapamil (Spirochem, SC-007). Cells were imaged every 5 min for a total of 24 h using a Leica SP8 and images were analyzed using ImageJ.

Statistical analysis
Statistical analysis was carried out using the Prism v8.0a software. Statistical methods are stated in the figure legends. Data are presented as mean ± SEM, unless indicated otherwise.

RESULTS
To assess the role of TIP60 we selected cell types to capture human and mouse cells, normal cells (mouse embryonic fibroblasts, MEFs), immortalized cells (HEK293), and cancer cells (U2OS). U2OS cells are osteosarcoma cells, that can be considered as a cancerous equivalent of primary isolates of fibroblasts [52]. Like MEFs, HEK293 and U2OS cells are of mesodermal origin.
Treatment of Tip60 fl/fl ;ERT2 and Tip60 fl/+ ;ERT2 MEFs with 4-OHT for 3 days followed by passaging in culture for 2 weeks in the absence of 4-OHT resulted in complete growth arrest in Tip60 iKO/iKO ;ERT2 MEFs, whereas heterozygous Tip60 +/iKO ;ERT2 control MEFs continued to proliferate normally (Fig. 1B). Tip60 iKO/iKO ;ERT2 MEFs retained the ability to adhere to the cell culture plastic after passaging and no excessive floating cells were observed, suggesting an absence of cell death. Tip60 iKO/iKO ;ERT2 MEFs displayed a large and flat morphology typical of senescent MEFs ( Fig. 1C; compare to [53,54]. In parallel, we deleted Tip60 by inducible CRISPR/Cas9 genome editing, involving doxycycline (dox) inducible sgRNA expression marked with GFP, constitutively expressed mCherry-tagged Cas9 [48], using two sgRNAs targeting exons 8 and 11 of TIP60 in human and mouse cells (Figure S1C, D; referred to as g1/C9 and g2/C9 or g/ C9 for merged data). Cell lines containing only sgRNA, lacking Cas9 (referred to as g1 and g2 or g-only for merged data) and either treated with dox or not treated (g-only dox or g-only CTL), and cell MEFs. Knockout of Tip60 was induced (Tip60 iKO ) by treatment for 3 days with 4-OHT, or cells were treated with EtOH vehicle control before culture in normal growth medium (means ± SEM, n = 3, unpaired two-tailed t test on slopes of log-transformed values). C Representative phase-contrast images of Tip60 fl/fl ;ERT2 MEFs 5 days after 4-OHT induction vs. heterozygous and EtOH controls. D Quantification of indel frequencies after doxycycline (dox) induction of sgRNA expression in iC-Tip60 MEFs for sgRNA#1 (g1/C9) and sgRNA#2 (g2/C9) and sgRNA-only controls (g-only CTL) (means ± SEM, n = 3, one-way ANOVA). E Proliferation of iC-Tip60 MEFs induced to deleted Tip60 with dox treatment vs. untreated cells (expressing Cas9 only) and g-only CTL controls (means ± SEM, n = 3, unpaired two-tailed t test on slopes of log-transformed values). F Representative phasecontrast images of iC-Tip60 MEFs 5 days after dox induction vs. untreated cells (expressing Cas9 only) and g-only CTL controls. Circles represent individual data points of cell isolates from individual embryos (A, D).
lines containing the sgRNA and Cas9 constructs, but not induced to express the sgRNA (referred to as g/C9 CTL), were used as controls. After dox treatment, insertion and deletion (indels) mutations of Tip60 were analyzed in the bulk population via high-throughput sequencing ( Figure S1E). After 48 h of doxycycline induction, the indel frequency for both sgRNAs plateaued at around 85% in MEFs (Fig. 1D). sgRNA-only controls showed no indel activity. Assessment of the top four predicted off-target loci for each sgRNA, identified by WTSI Genome Editing (WGE) [55], showed no indel activity ( Figure  S1F). RT-qPCR analysis showed a reduction of Tip60 mRNA in bulk MEFgC/9 cultures after 3 days of induction ( Figure S1G); however, small indels are expected to cause a loss of function even without a reduction in mRNA. Currently commercially available TIP60 antibodies (Invitrogen (PA5-23290), Sigma (SAB4500118), Abcam (ab135490), SantaCruz (sc-166323), CST (12058 S)) did not detect endogenous TIP60 protein in our experiments, preventing us from assessing TIP60 protein levels.
The slight outgrowth of cells observed after 9 days of dox treatment was consistent with an approximately 85% efficiency of the CRISPR/Cas9 genome editing system (Fig. 1D). Analysis of cells grown for an extended period in dox supplemented medium revealed a partial loss of sgRNA or Cas9 expression ( Figure S2). Cells that either did not integrate or lost or silenced the sgRNA or Cas9 constructs would be expected to grow similarly fast as controls.
Although TIP60 has been reported to play a central role in the DNA damage response [14,15], we did not observe an increase in γH2AX staining, used as an indication of DNA damage, in Tip60 iKO/iKO ;ERT2 MEFs 3 days after 4-OHT induction ( Figure S1H).
TIP60 deletion in human cells causes cell cycle arrest and failure to progress to metaphase TIP60 was mutated in U2OS and HEK293 cells using dox-inducible CRISPR/Cas9 genome editing. Continuously culturing the cells in dox resulted in a strong reduction in TIP60 mRNA after 3 days ( Fig. 2A) and a reduction in cell growth (Fig. 2B).
Live-cell imaging and tracking of individual cell divisions on iC-TIP60 U2OS and control cells revealed that most dox-treated iC-TIP60 U2OS cells underwent a final cell division before complete proliferation arrest ( Figure S3A and Movies 1A-C). Some cells maintained normal division rates, consistent with <100% efficiency of CRISPR/Cas9 genome editing (Fig. 1D, S2A).
Brightfield microscopy of iC-TIP60 U2OS cells revealed a senescent cell morphology. In contrast, iC-TIP60 HEK293 cells displayed reduced adherence to the tissue culture surface and largely lost contact after 4-5 days of dox treatment (Fig. 2C). TIP60 deletion led to a 2.4-fold (TIP60 iKO/iKO ;ERT2 MEFs) to fivefold (iC-TIP60 U2OS cells) increase in senescence-associated-β-galactosidase (β-gal) positive cells ( Figure  S3B). Further analysis of senescence markers revealed only a partial senescence footprint, with increased levels of p21 and p53 in the nucleus of iC-TIP60 U2OS cells, while other previously reported senescence-associated markers, such as TNF-α, lamin B1, MMP-3, RB, γ-H2AX, and macroH2A remained unchanged and IL-6 was not expressed in U2OS cells ( Figure S4A-D). CDKN2A (coding for p16 and p19) is deleted in U2OS cells [56] and therefore could not be used as a senescence marker in these cells.
Three different tests assessing cell death (live/dead cell assay, TMRE, and annexin V flow cytometric analysis), showed that cell survival was not affected in U2OS or adherent and floating HEK293 cells after 3 days of doxycycline treatment ( Figure S5A-C).
Loss of TIP60 in iC-TIP60 U2OS cells led to a twofold decrease in S-phase, a 47% increase in G2/M-phase cells, a 34% increase in G1 phase cells, and an eightfold increase in cells with a DNA content >2 N, indicating endoreplication as assessed by DNA content and BrdU incorporation (Fig. 2D). High-resolution 3D confocal live-cell imaging of iC-TIP60 U2OS cells revealed severely disturbed mitoses with chromosome condensation prolonged from the 25 min (control) to 4 h 45 min during mitosis, failure to arrange chromosomes in a metaphase plate and ultimate failure to divide ( Fig. 2E and Movies 2A, B).
iC-Tip60 MEFs showed a progressive decrease in H2AZK7ac over 3 days to sevenfold and a twofold decline in H2AZK4ac (Fig. 3D). The rapid onset of reduction in H2AZ acetylation after genomic deletion of Tip60 supports a direct dependence on TIP60. We further observed a~29% reduction in H4K12ac, but no other changes (Fig. 3D, S6B).
Western blotting of histones and modified histones on the cytosolic, soluble nuclear, and chromatin-bound fractions revealed that chromatin-bound total H2AZ levels remained unchanged in iC-TIP60 U2OS cells 3 days after dox treatment, while H2AZac levels were reduced by 85% ( Figure S8A, B).
TIP60 is required for normal gene expression in mouse and human cells Since H2AZ acetylation is reported to play an important role in gene expression in yeast [59], we examined the effect of TIP60 mutation on transcription. RNA-seq analysis of Tip60 +/+ ;ERT2 control and Tip60 fl/ fl ;ERT2 MEFs 3 days and 5 days after 4-OHT treatment revealed 8238 differentially expressed genes in Tip60 iKO/iKO ;ERT2 cells compared to control cells (Table S1, Fig. 5A). Differentially expressed genes after 3 days and 5 days were highly correlated (p = 5×10 -5 ; Figure S9A). For the subsequent analysis, we focused on genes differentially expressed 3 days after induction of Tip60 deletion to avoid secondary effects of loss of TIP60, at which time a total of 6311 genes were differentially expressed (Table S1). Among the top 50 differentially expressed genes ranked by FDR, 92% were downregulated (Fig. 5B) and among the top 500 differentially expressed genes 77.6% were downregulated, suggesting a bias towards transcriptional repression in the absence of TIP60 (Table S2).
No compensatory upregulation of another KAT gene was seen in the absence of TIP60 ( Figure S9B). Only minor changes in mRNA levels of the TIP60 protein complex subunits were seen ( Figure S9B). p53 and its target genes, including Cdkn1a (encoding p21 CIP1/ WAF1 ), were generally downregulated in the absence of TIP60 in MEFs (Fig. 5C), suggesting that the p53 pathway was not involved in the cell cycle arrest mechanism caused by the absence of TIP60. Interestingly, Cdkn2b and Cdkn2c (encoding p15 INK4B and p18 INK4C respectively) were upregulated (Fig. 5C). Cell adhesion genes and collagen genes were up to 25-fold upregulated ( Figure S9C), consistent with a senescent cell phenotype. This notably included cell-substrate adhesion genes, such as integrin and collagen genes, with 30% of the genes within this biological process upregulated (p = 0.0002; Table S3).
Among the top 10 enriched biological processes, 8 were downregulated, including gene expression, protein translation, and associated processes ( Fig. 5D; Table S3), consistent with cell cycle arrest and reduced requirement for metabolism and protein synthesis after the loss of TIP60. Transcription and gene expression processes overall were downregulated in the absence of TIP60 (Fig. 5D; Table S3). Downregulated pathways included p53-signaling, cell cycle regulation, apoptosis, and necroptosis (Table S4), congruent with the observation that TIP60 knockout MEFs neither proliferate, nor undergo apoptosis or other forms of cell death, but rather enter cell cycle arrest.
In HEK293 cells, the use of sgRNA#1 or sgRNA#2 resulted in highly correlated expression changes 3 days after induction of TIP60 deletion ( Figure S10A) and therefore the data from the two different sgRNAs were combined for all subsequent analyses.
Similar to MEFs, TIP60 complex subunit and KAT genes were not greatly dysregulated in HEK293 cells lacking TIP60 ( Figure S10B). Cell cycle and mitosis processes and genes were predominantly downregulated in the absence of TIP60 ( Figure S10C; Table S6). Similar to MEFs, cyclin-dependent kinase genes CDKN2B and CDKN2C were upregulated in HEK293 cells lacking TIP60 ( Figure S10D).  Figure S6 and full-size Western blots are displayed in the supplemental information.
TIP60 deleted HEK293 displayed a large proportion of cells that lifted off the cell culture plastic. However, the proportion of dead cells was not increased (Figure S5B), suggesting a primary failure in cell-substrate adhesion. Genes encoding cell-cell adhesion proteins, in particular of the protocadherin, cadherin, and ephrin receptor families were upregulated in the absence of TIP60 ( Fig. 6B and S10E, Table S5), suggesting cell-cell adhesion increased as cell-substrate adhesion decreased. Overall, 33% of genes with the annotation "cell-cell adhesion" were upregulated in the absence of TIP60 (Table S6).
Six of the top 10 biological processes were downregulated in HEK293 cells lacking TIP60. These included metabolic processes, whereas organ development processes were upregulated (Fig. 6C, Table S6). Similarly, among the pathways affected, metabolic pathways and the cell cycle were downregulated (Table S7). Gene expression and transcription processes were comprehensively downregulated (Fig. 6D).
We performed RNA-seq on iC-TIP60 U2OS cells 4 days after induced TIP60 mutation adding in a D. melanogaster spike-in cells to enable detection of global changes in mRNA, as previously described [70]. We found a total of 12716 expressed genes, 2707 genes were downregulated, and 1555 genes were upregulated (Table S8; Figure  S11A, S11B), representing a shift towards downregulated genes compared to RNA-seq results from HEK293 cells and MEFs without D. melanogaster spike-in normalization, while total mRNA levels remained similar. The directional change and the amplitude of change in gene expression in iC-TIP60 U2OS cells correlated strongly positively with gene expression changes in iC-TIP60 HEK293 cells ( Figure S11C) and Tip60 iKO/iKO MEFs ( Figure S11D).
We integrated the iC-TIP60 U2OS RNA-seq dataset and CUT&Tag dataset to examine correlations between gene expression and H2AZ occupancy and H2AZ acetylation. H2AZ and H2AZac occupancy was detected in wild-type cells at the start of transcription of more than 99% of expressed genes (Table S9). Of the genes decorated with H2AZ and H2AZac, 35% were differentially expressed in the absence of TIP60, whereas the RNA levels of 65% were not affected by the loss of TIP60. Of the differentially expressed genes, 64% were downregulated in RNA levels (Table S9). In TIP60 wild-type U2OS cells, RNA levels correlated significantly with H2AZ occupancy (p < 10 −6 ) and H2AZK4-K7 acetylation levels (p < 10 −6 ; Figure S12A). Furthermore, downregulation of genes in iC-TIP60 vs. control U2OS cells correlated significantly with H2AZ (p < 10 −6 ) occupancy and H2AZK4-K7 acetylation levels in genes that were differentially expressed in iC-TIP60 vs. control U2OS cells (p < 10 −6 ; Figure S12B).
Loss of TIP60 causes cell cycle arrest even in the absence of the cell regulators p53, INK4A, and ARF Our RNA-sequencing results suggested that the ARF-p53 tumor suppressor pathway (Fig. 7A) was not engaged in the cell cycle arrest induced by loss of TIP60 in MEFs. This finding could potentially have bearing on the suitability of TIP60 as a target for cancer therapeutics, as the TP53 gene is mutated in 42% of cancers [71], and CDKN2A, encoding INK4A and ARF, is mutated in 3.6% in a pan-cancer study [71] and in up to 78% in specific types of pediatric leukemia [72]. To formally test if p53 or INK4A-ARF were required for cell cycle arrest induced by loss of TIP60, we deleted the Tip60 gene in MEFs lacking p53 (Trp53 -/-MEFs) or INK4A-ARF (Cdkn2a -/-MEFs). In Trp53 -/-MEFg1/C9 cells indel frequency reached 80% after only 24 h of doxycycline induction (Fig. 7B). TIP60 deletion in Trp53 -/-MEFs or in Cdkn2a -/-MEFs resulted in growth arrest, indicating a p53 and INK4A-ARF-independent mechanism of cell cycle arrest (Fig. 7C). The morphology was indistinguishable from Tip60 only deleted MEFs (Figs. 7D vs. 1C, F).

DISCUSSION
In this study, we induced TIP60 deletion in normal mouse cells, immortalized human cells, and human cancer cells, as well as cells lacking the tumor suppressors p53, INK4A, and ARF to investigate the role of TIP60. TIP60 deletion caused cell growth arrest, but cell survival was not affected.
In our study loss of TIP60 caused a p53-independent cell cycle arrest. In contrast, TIP60 has previously been reported to induce cell cycle arrest, either as a co-activator of p53 at target genes such as CDKN1A (p21 CIP1/WAF1 ) [73], or in a p53-independent manner [74] after genotoxic stress in the context of DNA repair. The active role of TIP60 in inducing cell cycle arrest in the context of genotoxic stress appears to involve specific molecular mechanisms, including interaction with DNA damage repair proteins [15]. Our data, showing that the loss of TIP60 causes cell cycle arrest in a range of cell cycle phases, suggest that the observed cell cycle arrest is not due to effects on a single cell cycle regulatory mechanism, but is rather mediated through a broad Our data identify the specific lysine residues that rely on TIP60 for acetylation. Loss of TIP60 results in a marked decline in H2AZK7 acetylation and a reduction in H2AZK4 acetylation, consistent with previous reports proposing that TIP60 is required for the acetylation of H2A.Z and gene activation in human and mouse cells [75,76]. Apart from the profound reduction in H2AZ acetylation, we observed only a modest reduction in acetylation levels at some H4 lysine residues in the absence of TIP60. However, the effects on H4 acetylation were not present in all cell types and the specific H4 residue affected was not consistent between cell types. Various other HATs are known to acetylate H4 lysines [77]. Redundancy between those HATs might conceal a potential effect of loss of TIP60 on global H4 acetylation.
Global chromatin-bound H2AZ levels were not reduced upon TIP60 deletion in U2OS cells, indicating either no effect of TIP60 deletion on H2AZ dynamics or a complete halt of H2AZ exchange.
The mitotic defect and increased endoreplication indicate that Tip60 deleted cells were able to synthesize DNA, but failed to arrange chromosomes in a metaphase plate, and a failure of   5 TIP60 is required for normal gene expression in mouse cells. A-D RNA-seq results of Tip60 iKO/iKO ;ERT2 vs. Tip60 +/+ ;ERT2 control MEFs after 3 days and 5 days (A, combined analysis) or after 3 days of 4-OHT treatment B-D to induced Tip60 deletion (n = 3 cell isolates from individual embryos per genotype). A Mean-difference plot showing the log2 fold change vs. average log2 counts-per-million of each mRNA in Tip60 iKO/iKO ;ERT2 MEFs vs. Tip60 +/+ ;ERT2 control MEFs after treatment for 3 and 5 days with 4-OHT. Differentially expressed genes (FDR < 0.05) are indicated. B Heatmap of the top 50 differentially expressed (DE) genes by FDR in Tip60 iKO/iKO ;ERT2 vs. Tip60 +/+ ;ERT2 control MEFs after 3 days of 4-OHT treatment. C Expression level changes of the Trp53 (p53) gene and p53 target genes and cyclin-dependent kinase inhibitor genes Cdkn2b and Cdkn2c in the absence of TIP60. D Top 10 biological processes (GO terms) and gene expression and transcription processes (GO terms) enriched for genes differentially expressed (DE) in the absence of TIP60 in MEFs. RNA-seq results of Tip60 iKO/iKO and control MEFs, including genes encoding other histone acetyltransferases, TIP60 complex proteins, cell adhesion proteins, extracellular matrix proteins, and gene expression regulators, are displayed in Figure S9.
chromosome segregation. These observations may relate to the knockout phenotype of the yeast homolog of TIP60, esa1. Esa1 null mutant or catalytic mutant Saccharomyces cerevisiae display a budding phenotype characterized by incomplete DNA segregation and aberrant spindle orientation [78]. In this context, it is interesting that TIP60-mediated acetylation of aurora B kinase, the catalytic subunit of the chromosome passenger complex, promotes accurate chromosome segregation in mitosis [79].
H2AZ has been shown to play an important role in centromere structure formation [80] and H2AZ depletion has been linked to abnormal nuclear chromatin structure, as well as lagging chromosomes and chromosome bridges during cell division in mammalian cells [81]. The mitosis defect upon loss of TIP60 (and 80% of H2AZac) suggests that TIP60-dependent acetylation of H2AZ may be important in centromere and spindle formation.
Our RNA-seq data on TIP60-deficient cells support a broad role of TIP60 in gene transcription, congruent with the role of H2AZ acetylation. Thousands of genes relied on TIP60 for normal expression levels in MEFs, including genes related to protein translation, ribosome biogenesis, and general gene expression. The TIP60 mutant cells phenotype is likely to be due to widespread failure of transcription rather than a small subset of genes being responsible for the phenotype. While more genes are downregulated than upregulated in response to the loss of TIP60, the amplitude of change is generally larger in the upregulated genes, suggesting that we also observed gene expression changes secondary to cellular events such as cell cycle arrest. H2AZ acetylation has been correlated with active transcription, consistent with a role of TIP60 in H2AZ acetylation and gene activation, albeit less locus-specific than previously reported [82][83][84]. H2AZ is found broadly at the transcription start site (TSS) of active and inactive genes as well as in gene bodies of inactive genes [63]. In contrast, acetylated H2AZ is shown to exclusively accumulate at the TSS of active genes [63]. Our data suggest that TIP60 is responsible for the acetylation of H2AZK7 at active genes. A   Fig. 7 Loss of TIP60 causes cell cycle arrest even in the absence of the cell cycle regulators p53, INK4A and ARF. A Simplified schematic displaying the regulatory interplay between ARF, MDM2, and p53. B Quantification of indel frequencies in iC-Tip60 Trp53 -/-MEFs. (Means ± SEM, n = 3, one-way ANOVA). C Proliferation of iC-Tip60 Trp53 -/-MEFs and iC-Tip60 Cdkn2a -/-MEFs vs. untreated controls (expressing Cas9 only) and g-only controls (means ± SEM, n = 6, unpaired two-tailed t test on slopes of log-transformed values). D Representative phasecontrast images iC-Tip60 Trp53 -/-MEFs and iC-Tip60 Cdkn2a -/-MEFs 5 days after dox treatment vs. untreated controls. E Fold change of total H2AZ and histone acetylation levels normalized to H3 in iC-Tip60 Trp53 -/and iC-Tip60 Cdkn2a -/vs. g-only control MEFs after 3 days of dox treatment (means ± SEM, n = 2, unpaired two-tailed t test on log-transformed values). Experiments assessing histones H2A, H3, and H4 lysine acetylation levels are displayed in Figure S6D.
comparison of the RNA-seq and the Cut&Tag data revealed a positive correlation of RNA levels with H2AZ occupancy and H2AZ acetylation. High H2AZ acetylation and H2AZ occupancy correlated with stronger downregulation in gene expression upon TIP60 mutation. However, RNA levels of 65% of genes displaying decreased H2AZ acetylation at their TSS were not significantly changed in their expression levels after TIP60 mutation. Residual H2AZ acetylation, i.e., at lysine 4, might be sufficient to keep genes active in the absence of TIP60. Furthermore, high expression levels and thus more open chromatin structure combined with active RNAPII transcription through the locus might facilitate continued transcription even upon loss of H2AZ acetylation. Lastly, H2AZ might only be more important for the onset of gene transcription but not for maintaining open chromatin structure and active gene expression.
In conclusion, we propose a mechanism by which TIP60 acts as a co-activator of gene transcription through H2AZ acetylation. Identifying H2AZK7 and H2AZK4 as endogenous histone acetylation targets of TIP60 in a number of human and mouse cell types is an important step in understanding the role of TIP60 and its histone acetylation function. Furthermore, the finding that loss of TIP60 induces cell cycle arrest independent of the tumor suppressors p53, INK4A, and ARF, and identification of H2AZK7 acetylation as a potential biomarker are essential steps for the assessment of TIP60 as a potential therapeutic target.

DATA AVAILABILITY
Complete results from our RNA-seq studies, including complete lists of differentially expressed genes, differentially enriched GO terms and KEGG pathways, are included in the in the supplementary tables. Raw data NCBI Short Read Archive accession number GSE208383.