Negative Regulation of p21Waf1/Cip1 by Human INO80 Chromatin Remodeling Complex Is Implicated in Cell Cycle Phase G2/M Arrest and Abnormal Chromosome Stability

We previously identified an ATP-dependent human Ino80 (INO80) chromatin remodeling complex which shares a set of core subunits with yeast Ino80 complex. Although research evidence has suggested that INO80 complex functions in gene transcription and genome stability, the precise mechanism remains unclear. Herein, based on gene expression profiles from the INO80 complex-knockdown in HeLa cells, we first demonstrate that INO80 complex negatively regulates the p21Waf1/Cip1 (p21) expression in a p53-mediated mechanism. In chromatin immunoprecipitation (ChIP) and a sequential ChIP (Re-ChIP) assays, we determined that the INO80 complex and p53 can bind to the same promoter region of p21 gene (-2.2kb and -1.0kb upstream of the p21 promoter region), and p53 is required for the recruitment of the INO80 complex to the p21 promoter. RNAi knockdown strategies of INO80 not only led to prolonged progression of cell cycle phase G2/M to G1, but it also resulted in abnormal chromosome stability. Interestingly, high expression of p21 was observed in most morphologically-changed cells, suggesting that negative regulation of p21 by INO80 complex might be implicated in maintaining the cell cycle process and chromosome stability. Together, our findings will provide a theoretical basis to further elucidate the cellular mechanisms of the INO80 complex.

Like the other chromatin remodelers, INO80 complex possesses ATPase and DNA nucleosomal sliding activities. Using the energy of ATP hydrolysis, INO80 can alter histone-DNA interactions, causing nucleosomes to move in cis (sliding) [8]. Thus, INO80 plays a role in concert to organize chromatin structure by depositing, moving, evicting, or selectively altering nucleosomes in an ATP-dependent manner [9]. Research into the functions of INO80 in replication, DNA damage repair and gene transcription in higher eukaryotes has been rather incomplete, but increasing data reveals that INO80 complex plays an important role in those processes. Depletion of Ino80 not only resulted in slower growth and reduced S-phase progression, but also led to defective telomere replication, impaired DNA repair and hydroxyurea (HU) sensitivity [4,5,10]. Recent electron microscopy (EM) studies showed that the Ino80 complex formed an elongated embryo-like shape with head-neck-body-foot topology in yeast [11,12]. Although the complete crystal structure of INO80 complex is still unknown, three modules that assemble on three distinct domains of the INO80 protein have been reported [13]. All shared core subunits composed of two modules and assembled on the conserved helicase-SANT-associated/post-HSA (HSA/PTH) and ATPase domains of INO80 protein. Both HSA/PTH and ATPase domains are essential for catalyzing the ATP-dependent nucleosome remodeling activity of the INO80 complex. Lacking Arp4 and Arp8 in yeast, Ino80 complex is defective in DNA binding, ATPase, and nucleosome remodeling activities [14]. In human cells, INO80 complex is recruited to DNA damage sites in an Arp8-dependent manner [15].
Chromatin remodelers play key roles in maintaining normal dynamic changes of chromatin structure in cells [16]. However, dynamic changes of chromatin formation are often affected by many factors during the cell cycle, of which DNA damage and its repair processes are the most important factors. Evidence from yeast and human cell research suggests that the Ino80 complex participates in DNA damage repair [4,6], and the recruitment of Ino80 complex to DNA double strand breaks is in a cell cycle-dependent manner [17]. Further studies revealed that checkpoint kinases and the Ino80 complex enhance global chromatin mobility in response to DNA damage [18], suggesting the roles of Ino80 complex in DNA damage repair and in cell cycle checkpoint control.
CDKN1A (p21 Waf1/Cip1 ) is a universal inhibitor of cyclin kinases which controls cell cycle by activating and/or inactivating the cyclin-dependent kinases (CDKs) [19,20]. It is well known that p21 is both a classical tumor suppressor and an oncogene, and is involved in many cellular processes [21]. Accumulation of cytoplasmic p21 has been reported in many cancers and is associated with tumor aggressiveness, metastasis, and prognosis [22]. Activated tumor suppressor p53 can bind directly to two highly conserved p53 response elements (PRE) in the p21 promoter and induce transcriptional activation of the p21 gene [23]. Over the past few decades, research in the mechanisms regulating p21 transcription has been investigated mostly in cells that have defective cellular senescence pathways [24]. Interestingly, recent findings indicate that epigenetic regulators might be involved in transcriptional regulation of p21. For instance, SWI2/SNF2-related chromatin remodeling complexes including p400 and SRCAP are localized to the p21 promoter region near the p53 binding sites, and negatively regulate p21 transcription in a p53-dependent manner [25,26]. Moreover, p400 chromatin remodeler plays a key role in the deposition of the histone variant H2AZ within the p21 promoter, therefore repressing p21 gene expression [27].
Although it has been known that INO80 complex functions in DNA damage repair and therefore leads to delayed or arrested cell cycle, the precise mechanisms, especially its target genes, remains unclear. In an effort to resolve which genes are responsible for INO80 involved cellular functions including cell cycle regulation, we carried out the gene expression profiles in INO80 core subunits knockdown in HeLa cells. With combined methods of ChIP and Re-ChIP assays, RNAi knockdown, flow cytometry, immunofluorescence staining, RT-PCR and Western blots, we first demonstrate: 1) the p21, an inhibitor of cyclin kinases, is negatively regulated by the INO80 complex in a p53-mediated mechanism; 2) INO80 complex and p53 cooccupy the p21 promoter (-2.2 kb and -1.0 kb upstream of the p21 promoter region), the existence of p53 is required for recruiting the INO80 complex to the p21 promoter; and 3) INO80 complex might be implicated in maintaining the normal cell cycle and genome stability. Our results shed light on how INO80 complex is involved in cell cycle and in the maintenance of chromosome stability.

Chemical treatment
Camptothecin (CPT) (Cat. C991, Sigma) was dissolved in dimethyl sulfoxide (DMSO) and prepared into 5 mg/ml concentration for storage. The final concentration of CPT in cell culture medium is 5μM. 1M/L hydroxyurea (HU) (Cat. H8267, Sigma) was prepared with double distilled water (ddH 2 O) as a stock solution. Doxorubicin hydrochloride (D1515) from Sigma was dissolved in DMSO. The final concentration of doxorubicin in cell culture medium is 0.5μM.

Constructions of plasmids and transient transfection
Flag-tagged full length (FL) of p53 (GI:371502114) were constructed by inserting cDNAs into the XhoI/BamHI sites of pcDNA3.1(-). The pBS/U6 vector encoding INO80 shRNA (pBluescript-shINO80) was a kindly gift from Dr. Yang Shi in Harvard Medical School [25,17]. For transfection of either Flag-tagged p53FL proteins or INO80-shRNA plasmids, cells were seeded in 6-well tissue culture plates (~3 × 10 5 cells/well). Next day, 2μg plasmids were then transiently transfected using 6μL polyethylenimine (PEI) (Cat. 23966, Polysciences) according to the manufacturer's recommendation. After 48 hours post transfection, whole cell lysates were prepared and subjected to 12% SDS-PAGE gel electrophoresis. The proteins were analyzed by Western blot with appropriate antibodies.

DNA microarray and statistical analysis
HeLa cells were cultured in 6-well tissue culture plates (~2 × 10 5 cells/well) in DMEM medium containing 10% fetal bovine serum. Arp8 siRNA (sc-60072) was obtained from Santa Cruz, while all specific siRNAs including non-targeting siRNA (D-001206), INO80 siRNA (D-004176), Ies6 siRNA (D-019327) and Ies2 siRNA (D-009848) SMART pool were from Dharmacon (U.S.A.). The cells were transiently transfected with 10~20pmol specific siRNAs using Lipofectamine RNAiMAX transfection kit (Invitrogen, Cat.No-864425) following the manufacturer's instruction. 48 hours after siRNA transfection, cells were then stored in an RNA hold solution (ER501-01, Beijing Transgen Biotech Co., Ltd.) and sent to EMTD Science and Technology Development Co., Ltd. (Beijing, China) for DNA microarray analysis. The Illumina microarray datasets were accessible from the National Center for Biotechnology Information Gene Expression Omnibus (GEO) data repository (http://www.ncbi.nlm.nih.gov/geo/) using the series accession number GSE68655. In order to obtain functional annotation of INO80-regulated genes in our study, the online biological classification tool DAVID (Database for Annotation, Visualization and Integrated Discovery) was used to perform enrichment analysis [28,29]. Furthermore, DAVID was utilized to conduct Gene Ontology (GO) [30] function and Kyoto Encyclopedia of Genes and Genomes (KEGG) [31] pathway enrichment analysis of differentially expressed genes (DEGs). During the analysis, p < 0.05 and FDR < 0.02 genes used in the annotation were defined as statistically significant.

Flow cytometry analysis
HEK293T cells were grown to 50-60% confluence in DMEM medium containing 10% FBS. In order to determine the effect of INO80 on cell cycle, cells were transfected with pBluescriptvector (pBS-Vector) or pBlueScript-shINO80 (pBS-shINO80) plasmids. Cells were harvested 48 hours after transfection for flow cytometry analysis. In the cell cycle arrest experiment, cells were synchronized by treatment with 1mM hydroxyurea (HU). Cells were harvested by trypsinization 0, 2, 4, 6h, 8h and 12 hours after removal of HU. 10 6 cells with or without synchronization were suspended as single cell dispersions in 70% ethanol at -20°C at least 4 hours. After ethanol fixation, cells were centrifuged at 300 × g for 5 minutes and ethanol removed. Cells were washed twice with PBS buffer and then re-suspended in 300μl PBS containing 0.1% (v/v) Triton X-100 (Sigma, Cat. T8787), 0.3 mg/ml DNase-free, RNase A (Sigma, Cat. R5500), 50μg/ ml propidium iodide (Beijing Dingguo Cat. CF0031), then were incubated at 37°C for 1 hour. Data collection was performed using EPICSXL TM Cytometers (Beckman Coulter). Acquired data was analyzed using ModFit LT software (Verity Software House, U.S.A.). Table 1. RT-qPCR primer sets used in paper.

Primers
Directions Sequences Size (bp)

Gene expression profiles of INO80-knockdown HeLa cells reveal potential functional versatility and complexity
We previously identified a human ATP-dependent chromatin remodeling complex which functions in gene transcription [2,3]. Although increasing evidence suggests that INO80 chromatin remodeling complex plays a key role in gene transcriptional regulation [33,34], the exact regulatory mechanisms are still unclear. In order to identify which potential target genes were specifically regulated via knocking down the core components of the INO80 complexes, HeLa cells with specific siRNA (siINO80, siArp8, siIes2 and siIes6) transfections were sent to EMTD Science and Technology Development Co., Ltd. (Beijing, China) for gene expression profile analyses. The efficiency of siRNA knockdown was confirmed by RT-qPCR ( Fig 1A) and Western blots (Fig 1B). To analyze the difference between differently expressed genes (DEGs) in each siRNA knockdown cells, a Venn diagram was utilized to identify overlapping genes ( Fig 1C). Interestingly, a total of 251 genes including 149 down-and 102 up-regulated genes were co-regulated by INO80, hIes2, hIes6, and hArp8. To gain insight into whether INO80 complex was involved in important cell functional pathways by regulating pathwayrelated genes, a total of 2144 DEGs in at least two knocked down subunits were used for KEGG annotation by using DAVID web annotation tool. After KEGG pathway enrichment analysis, the DEGs were mainly enriched in 12 KEGG pathways including p53 signaling pathway, pathways in cancer, adherens junction, and cell cycle ( Table 2).

INO80-knockdown prolongs progression of G2/M to G1 phase in 293T cells
Since the DEGs were linked to cell cycle according to KEGG pathway enrichment analysis, we speculated that the INO80 complex was involved in cell cycle progress. Thus, cell flow     while the gene expression levels of cyclinG1 (CCNG1), cyclin-dependent kinase 2 (CDK2), cyclin-dependent kinase 6 (CDK6) and p21 were increased in both INO80-and hArp8-knockdown HeLa cells. Table 3 shows a comparison of Illumina fold change and RT-qPCR fold change. Note that changes in mRNA levels of CDK inhibitors, in particular p21, emerged after microarray studies. To clarify this result, Western blot and immunofluorescence analyses were carried out in INO80-and Arp8-knockdown HeLa cells. In both protein expression assays, we observed an apparent high protein expression of p21 in INO80-or Arp8-knockdown cells ( Fig  3B and 3C). Obvious p21 increased cells accounted for 37.8-43.3% and 19.2-22.7% of the total cell numbers in INO80-and Arp8-knockdown cells, respectively ( Fig 3D).
The above experimental results clearly suggested that p21 is regulated by INO80 chromatin remodeling complex. In order to determine whether the effect of INO80 on cell cycle progress was due to its regulation of the p21 gene transcription, the recruitment of INO80 complex to p21 gene was studied. Six primer sets in the p21 locus were used for amplifying ChIP'd DNA (Fig 4A). The specificity of each primer sets was confirmed by qPCR and 2.5% DNA agarose ( Fig 4B). ChIP assays were performed using INO80, YY1 (a subunit of INO80 complex) and p53 antibodies in 293T cells. Interestingly, INO80, and YY1 in conjunction with p53, co-occupied the -2.2kb and -1.0kb upstream of the p21 transcriptional start site (Fig 4C). To further clarify whether INO80 complex indeed enriched at the -2.2kb and -1.0kb upstream of the p21 transcriptional start site, ChIP assays using INO80 and YY1 antibodies in 293T cells with or without pBS-shINO80 transfection were carried out. As expected, silencing INO80 significantly reduced both of INO80 and YY1 binding around the p21 transcriptional start site (-2.2 kb and -1.0 kb). (Fig 4D).

INO80-mediated upregulation of p21 may be in a p53-dependent manner
According to published reports, the tumor suppressor gene p53, a regulator of p21, can be enriched in the specific sites of the p21 gene. Although a slightly different distribution of p53 in the p21 gene was reported by different research groups, two regions including -2.5-2.2kb and -1.3-1.0kb upstream of the p21 transcriptional start site are considered as the main binding sites for p53 [23,35]. Distribution of INO80 complex on the p21 gene -2.2kb and -1.0kb upstream of the transcriptional start site suggested that the regulation of INO80 on p21 transcription may be related to p53-mediated mechanism. To address this hypothesis, p53 +/+ and p53-/-HCT116 colon carcinoma cell lines were used in this experiment. The cell lines were confirmed by RT-PCR ( Fig 5A) and Western blot analysis. The induction of total p53, phosphorylated p53-S15p, p53-S33p, and p53-S46p proteins was only observed in 5μM CPT treated p53+/+ HCT116 cells (Fig 5B). Next, to evaluate the correlation between regulation of p21 gene by INO80 and p53 at the transcriptional level, three consecutive (48 hours interval) INO80 knockdown with siRNA was performed in both p53+/+ and p53-/-HCT116 cells. 48 hours after each transfection, cells were harvested for RT-qPCR to assess indicated gene expression. As shown in Fig 5C, high expression of p53 target genes, in particular p21, remained continuously higher in three consecutive INO80 knockdown p53+/+ HCT116 cells. In contrast, there was no significant increase of gene expression including p21 detected in p53-/-HCT116 cells after INO80 knockdown. RNAi approaches in two kinds of HCT116 cells further argued that a role for INO80 in regulation of p21 may be associated with p53-mediated mechanism. Silencing of INO80 by pBS-INO80 shRNA (2 μg/well/6-well plate) led to an obvious increase in p21 protein level in  Table 1). BCCIP is the PCR control, and p53 plasmid is the positive control of the PCR product. No template in the reaction is the negative control. (B) Western blot analysis. Both cell lines were cultured in IMDM medium containing 5μM CPT for 8 hours. WCE was then prepared, and the proteins were detected with indicated antibodies. (C) High expression of p21 gene in three consecutive INO80-knockdown cells. Both HCT116 cells were transfected with 20pmol INO80 and NT (as control) siRNAs three times every 48 hours. Then, 48 hours after each transfection, cells were harvested for RT-qPCR to assess indicated gene expression. Bar graphs show ratios of RT-qPCR signals to GAPDH (all signals normalized to siNT). (n = 3) (D) Obvious up-regulation of p21 in INO80-knockdown p53+/+ HCT116 cells. HCT116 cells were transfected with 2μg pBS-shINO80 (well/6-well plate) and pBS-Vector (as control). 48 hours later, indicated proteins were detected by WB with specific antibodies. Representative results from three independent experiments are shown in the upper panel. Western blot images (n = 3) were quantified with densitometry using Quantity One Basic software (Bio-Rad). *p < 0.05, **p < 0.01 in comparison with pBS-Vector control (Student t test). (lower panel). (E) p53-dependent up-regulation of p21 in INO80-knockdown cells. p53+/+ HCT116 cells were transfected with pBS-shINO80 or pBS-Vector. 24 hrs after transfection, cells were treated with 0.5μM doxorubicin (Dox) for 24 hours. WCE was then prepared and indicated proteins were measured by WB with specific antibodies. Representative results from three independent experiments are shown. (F) Quantified proteins. Western blot images (n = 3) were quantified with densitometry using Quantity One Basic software. **p < 0.05, p < 0.01 in comparison with pBS-Vector control (Student t test). p53+/+ HCT116 cells(p<0.01), but not in p53-/-HCT116 cells(p>0.05) (Fig 5D). To further understand the p53-mediated functions of INO80 in regulation of p21, pBS-shINO80 or pBS-Vector transfected p53+/+ HCT116 cells were treated with 0.5μM doxorubicin (Dox) for 24 hours, and Western blot analysis were performed with the indicated antibodies (Fig 5E). In agreement with other reports [36], Dox induced high expression of p53 including phosphomodified p53 and p21 proteins (right panel). Note that the expression of p21 protein was higher in INO80-knockdown cells than those in pBS-control cells regardless of treatment (p<0.01) or non-treatment with Dox (p<0.05), suggesting that up-regulation of p21 by knocking down INO80 may be regulated in a p53-mediated mechanism (Fig 5E and 5F).
To further confirm this hypothesis, pBS-Vector-or pBS-shINO80-treated p53-/-HCT116 cells were transiently transfected with p53 or pcDNA3.1 plasmid. As shown in Fig 6A, although a slight heightened expression of p21 appeared in transfection of pcDNA3.1 in INO80 knocked down p53-/-HCT116 cells (lane 1-3) (p>0.05 compared to NT control, data not shown), transient transfection of p53 plasmids (lane 4-6) caused evident increase of p21 in a dose-dependent manner (lane 5-6), In order to clarify whether the p53 was required for the localization of INO80 to p53 binding sites in p21, another ChIP assay was performed using INO80 antibody in p53+/+ and p53-/-HCT116 cells. As shown in Fig 6B, once again the distribution of INO80 was confined to-2.2 kb and -1.0 kb upstream of the p21 in p53+/+ HCT116 cells, but the peaks of distribution of INO80 on p21 disappeared in p53-/-cells, suggesting that the existence of p53 is necessary for the recruitment of the INO80 complex to the p21 promoter. To further determine the co-occupancy of the INO80 and p53 in the p21 promoter region, a sequential ChIP (Re-ChIP) was performed based on the previous report [32]. The experiment process is shown in Fig 6C. As shown in Fig 6D,

Roles of INO80 in chromosome stability and cytokinesis
It has been reported that p21, a potent cyclin-dependent kinase inhibitor, plays an important role in regulating cell cycle phase G2/M [37,38]. Silencing of INO80 resulted in both up-regulation of p21 and cell cycle G2/M phase arrest, suggesting that INO80 complex might block the progression of G2/M to G1 by regulating p21. To gain a clear idea of the roles of INO80 complex in G2/M phase, the impact of INO80 on chromosome stability in p53+/+ HCT116 and HeLa cells were studied. Knockdown of INO80 with siRNA in p53+/+ HCT116 cells led to an increase of large nuclei cells (Fig 7A). Significant elevation of large nuclei cells was found in INO80-knockdown cells (Fig 7B, p < 0.01). Again, increased p21-stained cells were observed in INO80-silenced cells (Fig 7C, p < 0.01). Interestingly, high-expressed p21 was observed in most of the large nuclei cells, indicating the roles of INO80 in chromosome stability might be mediated by regulating p21. In addition, significant increases of cells with more than two centrosomes per nucleus were observed in INO80-knockdown cells (Fig 7D). Moreover, INO80-knockdown also resulted in cytokinesis failure ( Fig 7E) and multipolar spindle formations ( Fig 7F).

Discussion
In this report we first present evidence that the INO80 chromatin remodeling complex negatively regulates the p21 expression in a p53-mediated mechanism using in vitro biological experiments combined knockdown/over-expression approaches. In addition, ChIP and Re-ChIP assays indicated that the recruitment of INO80 complex in the p21 gene relied on the p21 Is Negatively Regulated by hIno80 Chromatin Remodeling Complex existence of p53. By combining methods of flow cytometry and immunofluorescence staining, we postulate that the role of the INO80 complex to maintain the normal cell cycle and chromosome stability is at least partly related to the regulation of p21.
Ino80 complex, a member of the SWI/SNF2 superfamily of chromatin remodeling complexes, is highly conserved from yeast to human [39]. Except for a set of highly evolutionarily orthodox core subunits, six metazoan-specific subunits and YY1 are identified in the INO80 complex [2,3]. Interestingly, all of the metazoan-specific subunits are composed of an N-terminal regulatory module. The other subunits form two modules assembled on the conserved HSA/PTH and ATPase domains of the INO80 protein, therefore catalyzing ATP-dependent nucleosome remodeling activity of the INO80 complex [12,13]. Although the exact functions of Ino80 HSA/PTH domains remain unclear, increasing evidence argues that HSA/PTH domains play an important role in cellular processes in various species. Several nuclear Arps such as Arp4 and Arp8 have been shown to bind to DNA and histones, suggesting a possibility that Ino80 HSA/PTH-containing modules may contribute to recognition of DNA and/or nucleosome substrates [14,40], thereby regulating gene expression through chromatin structure alteration. Our experimental results also support this view. Based on the analysis of performed gene expression profiles from INO80-, Arp8-, hIes2-or hIes6-knockdown HeLa cells, we clarified that the INO80 complex was involved in multiple signal pathways including cell cycle ( Table 2), suggesting the importance of the INO80 complex in cellular biological processes.
It has been reported that the Ino80 chromatin remodeling complex contributes to a wide variety of chromatin-dependent nuclear transactions, including gene transcription; DNA replication and DNA repair [33]. Data from DNA microarray and ChIP experiments previously revealed that as many as two-thirds of the yeast genes in which transcription were affected by the Ino80 complex [41,42]. In contrast, studies of the INO80 complex on gene transcription, especially on genes related to cell cycle in human cells, are rare. Based on gene expression profiles and in vitro biological experiments, we observed the up-regulation of p21 by knocking down either INO80 or Arp8 in human cells (Fig 3). ChIP assay with available INO80 and YY1 antibodies revealed that the two binding sites of the INO80 complex in p21 gene were similar to those of p53 in p21 [23,35]. Furthermore, ChIP and Re-ChIP assays confirmed that the INO80 complex and p53 co-occupy the p21 gene promoter (-2.2 and -1.0kb upstream of the p21 transcriptional start site), suggesting the collaborative mechanism of INO80 complex and p53 ( Fig 4C and Fig 6D). As expected, further investigation verified that the up-regulation of p21 by INO80 was in a p53-dependent manner (Figs 5 and 6): 1) Significant increase of p21 expression was observed in INO80 knockdown p53+/+ HCT116 cells; 2) Recruitment of INO80 in p21 gene only occurred in p53+/+ HCT116 cells, and not in p53-/-cells; 3) Obvious dose-dependent increases of p21 were produced by over-expression full length of p53 in INO80 knockdown p53-/-HCT116 cells. On the other hand, research evidence demonstrated that high-expression of p21 in response to various stimuli and stress signal can be regulated in a p53-dependent and/or p53-independent manner. For example, p21 expression is up-regulated by forkhead box A1/2 in p53-null H1299 lung carcinoma or by SP1 in p53-null Caco-2 colon carcinoma cells [43,44]. In our study, note that p21 protein also had a slight increase in INO80 knockdown p53-/-HCT116 cells (Fig 5D and Fig 6A), suggesting that p53-mediated regulation of p21 may not be the only mechanism of the INO80 complex.
According to a report by Hur SK et al. [4], INO80 is critical for normal cell cycle progression, particularly in S-phase progression and mitosis. In this study, although the cell cycle arrest was exhibited in both S and G2/M phases, the significant prolonged progression of cell cycle phase G2/M to G1 was found in INO80-knockdown 293T cells (Fig 2), indicating that the INO80 complex might be implicated in the progression from DNA replication to cell division. Consistent with this idea, we carried out cell staining with primary antibodies in INO80knockdown p53+/+ HCT116 and HeLa cells to know the effects of INO80 on chromosome stability. Similar to that previous report [4], siINO80 treatment not only increased the large nuclei cells, but also caused multipolar spindle formations, cytokinesis failure and centrosome amplification in p53+/+ HCT116 or HeLa cells (Fig 7). Interestingly, hyper-expression of p21 was also observed in most of morphologically changed cells, suggesting p21-mediated function of INO80 in genome stability. It is worth noting that p21 shows bidirectional functions as both a tumor suppressor and an oncogene. In the nucleus, p21 as a negative cell cycle regulator, functions as a tumor suppressor. While in the cytoplasm, p21 acts as an oncogene by inhibiting apoptosis and facilitating cell proliferation [21,45]. In addition, phosphorylation of p53 at p53-S15p and p53-S20p can activate downstream target genes such as p21 which play an important role in G2/M checkpoint through inhibition of Cdk1/Cyclin B [46]. In this study, the presence of Dox in INO80 silencing p53+/+ HCT 116 cells also increased the protein level of p53 including p53-S15p and p21 (Fig 5E and 5F), suggesting the up-regulation of p21 by INO80 chromatin remodeling complex might be associated with cell cycle phase G2/M checkpoint and abnormal chromosome stability.
Although our ChIP and Re-ChIP experimental data suggest that INO80 complex inhibits the gene expression of p21 by acting as a transcriptional co-repressor of p53 in cells, the question of how to switch the transcriptional co-activator of p53 remains yet unclear. There are several possibilities we can think of at the moment: 1) Knockdown INO80 destroyed the joint action between the INO80 and p53; 2) p53 may switch to co-activator after knockdown INO80 and activate the p21 gene transcription; 3) non-p53 mediated mechanism may be involved in the transcription of p21 gene. Therefore, more experimental data is required to address those issues.

Conclusions
In summary, given that INO80 complex performs an important function in maintaining the normal cell cycle and genome stability, it is necessary to investigate the functions of INO80 complex in tumorigenesis and cancer therapy in the future.