Essential Dosage-Dependent Functions of the Transcription Factor Yin Yang 1 in Late Embryonic Development and Cell Cycle Progression^†

Histological analysis of hypomorphic mice.

Newborn mice were fixed in Bouin solution, dehydrated in ethanol, incubated in xylene, and embedded in paraffin. Lung sections stained with eosin and hematoxylin were observed by light microscopy.

Isolation of MEFs.

Mouse embryonic fibroblasts (MEFs) were derived from embryos 13.5 days postcoitum (dpc). Briefly, after removal of the head and internal organs, embryo carcasses were digested in trypsin solution for 30 min at 37°C. After addition of Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% of fetal bovine serum (FBS), the remaining tissues were disrupted by pipetting. The resulting cell suspension was plated and cultured in DMEM supplemented with 10% FBS and antibiotics. Infection with adenovirus-Cre (Gene Transfer Vector Core, University of Iowa) was performed 2 days postisolation. Cells were maintained in the presence of viruses in DMEM containing 5% of FBS for 16 h before being plated for various assays.

RNAi.

HeLa cells were transfected with a control or yy1 RNA interference (RNAi) plasmid (65) using Lipofectamine 2000 (Invitrogen Life Technologies). In some cases (see below), RNAi vectors were mixed with a puromycin resistance-encoding vector, and transfected cells were selected by adding puromycin (1.5 μg/ml; Sigma-Aldrich) to the culture medium for 2 days (65). Cells were then trypsinized and plated for various assays.

Analysis of growth properties.

For the colony formation assay, equal numbers (5,000 to 20,000) of MEFs were seeded in six-well plates and cultured for 10 to 12 days. Cell colonies were then stained with a solution containing 30 mg/ml crystal violet, 9 mg/ml ammonium oxalate, and 20% ethanol. The same protocol was used for HeLa cells, but the colony formation assay was performed 14 days after puromycin selection of the transfected cells.

For the MTT [3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide; Sigma-Aldrich] assay, equal numbers (10,000 to 30,000) of MEFs were seeded in six-well plates. After 2 to 10 days of culture, cells were incubated in the presence of 400 μg/ml of MTT in DMEM for 30 min at 37°C, allowing metabolic transformation of the MTT by the mitochondria and subsequent formation of an intracellular purple formazan (41). After solubilization of this precipitate in dimethyl sulfoxide, the absorbance was measured at 490 nm.

Analysis of cell viability and apoptotic cell death.

Puromycin-selected HeLa cells were treated with various apoptotic agents 4 days after transfection of the RNAi vectors. Briefly, cells were treated with 200 nM of staurosporine (Sigma-Aldrich), 100 μM of etoposide (Sigma-Aldrich), 100 ng/ml of Fas ligand (Alexis Corporation), or 10 μg/ml of tumor necrosis factor alpha (R&D Systems) and harvested at various time points posttreatment. While viable cells were trypsinized and counted, apoptotic cell death was quantified by Western blot analysis of poly(ADP-ribose) polymerase (PARP) cleavage.

Flow cytometry analysis.

The DNA content of MEF cells and puromycin-selected HeLa cells was analyzed essentially according to Shah et al. (58). Briefly, cells were harvested by trypsinization 4 days postinfection by adenovirus-Cre or posttransfection of the RNAi vectors and fixed with 50% ethanol. After one wash with phosphate-buffered saline (PBS), cells were treated with RNase A (100 μg/ml; Sigma-Aldrich) for 30 min at 37°C, stained with 50 μg/ml propidium iodide (Sigma-Aldrich), and analyzed with a FACSCalibur machine and Cellquest software (Becton Dickinson).

Bromodeoxyuridine incorporation and indirect immunofluorescence.

MEFs and unselected RNAi-treated HeLa cells were cultured on coverslips and incubated with 10 μM bromodeoxyuridine (5-bromo-2′-deoxyuridine; Sigma-Aldrich) for 4 h prior to fixation in a 3% paraformaldehyde solution for 20 min. Cells were then permeabilized with 0.5% of NP-40 in PBS for 20 min and washed with PBS containing 0.1% NP-40. The cells were further treated with 4 M HCl for 30 min, followed by one wash with PBS and another wash with 100 mM borax solution. After an additional wash in PBS containing 0.1% NP-40, cells were incubated in blocking solution (PBS containing 0.1% NP-40 and 10% of FBS) and costained with a mouse monoclonal antibromodeoxyuridine (BU33; Sigma-Aldrich) and a rabbit polyclonal anti-YY1 (H414; Santa Cruz Biotechnology, Inc.).

Expression of various cell cycle markers was analyzed by immunostaining of HeLa cells, using antibodies directed against PCNA (PC10; Santa Cruz Biotechnology, Inc.), cyclin B1 (GNS1; Santa Cruz Biotechnology, Inc.), cyclin E (HE12; Santa Cruz Biotechnology, Inc.), phosphorylated (Ser10) histone H3 (06-570; Upstate Cell Signaling Solutions), α-tubulin (B512; Sigma-Aldrich), or YY1 (H10; Santa Cruz Biotechnology, Inc.). With the exception of cyclin E immunodetection, which required methanol fixation, all immunofluorescence experiments were conducted on paraformaldehyde-fixed HeLa cells, as described previously (66).

Western blot analysis.

Total cell extracts were prepared in lysis buffer (50 mM Tris-HCl, pH 7.3; 5 mM EDTA; 50 mM KCl; 0.1% NP-40; 1 mM phenylmethylsulfonyl fluoride; 1 mM dithiothreitol) supplemented with a complete antiprotease cocktail (Roche Diagnostics), and protein concentration was assessed by Bradford assay. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, Western blotting, and peroxidase-based chemiluminescence detection were conducted according to standard laboratory procedures. Anti-YY1 (H414 and H10) and anti-PARP (F2) were from Santa Cruz (Santa Cruz Biotechnology, Inc.), the anti-lamin A and C antibody (2032) was from Cell Signaling (Cell Signaling Technology, Inc.), and the antiactin antibody (MAB1501) was from Chemicon International Inc. p53 induction was analyzed using an anti-p53 antibody (PAB240; Santa Cruz Biotechnology, Inc.), adriamycin treatment (2 μg/ml for 10 h) being used as a positive control. Densitometry analysis of YY1 protein levels in tissues and MEFs was performed using a digital camera coupled with band quantification software (Quantity One; Bio-Rad).

Microarray.

MEFs were isolated from three wild-type (yy1^+/+) embryos and compound heterozygous mutant (yy1^flox/−) littermates, and RNA samples were prepared 4 days postisolation. Briefly, RNAs were prepared using Trizol reagent (Life Technologies) and the RNeasy kit (QIAGEN). cDNAs were generated from total RNAs with the Superscript Choice system (Gibco BRL Life Technologies) and T7-oligo(dT)₂₄ primers. Subsequently, biotin-labeled cRNAs were synthesized using a BioArray high-yield RNA transcript labeling kit (Enzo Diagnostic, Inc.). After cleanup, fragmentation, and quality controls, the fragmented cRNAs were subjected to hybridization with a DNA-Chip mouse genome 430A array (Affymetrix) containing over 22,600 probe sets corresponding to transcripts and variants from approximately 15,000 well-characterized mouse genes. Data analysis was performed using DNA-Chip expression analysis software. Briefly, after determining the presence call and average differentiation parameters of each gene, gene expression levels from yy1^+/+ and yy1^flox/− MEFs were subjected to comparative analysis.

Dosage-dependent functions of YY1 during development.

Mice heterozygous for the yy1^flox conditional allele were born at the expected frequency and developed into fertile adults phenotypically indistinguishable from their wild-type littermates. Genotype analysis of the offspring from heterozygous (yy1^flox/+) intercrosses revealed that the number of yy1^flox/flox newborn pups represented approximately 74% of that expected according to the Mendelian ratio, suggesting that a small subset of homozygous mice died during embryonic development. The remaining yy1^flox/flox mice were viable and appeared normal despite a significant reduction in body size and weight (Fig. ​(Fig.2A).2A). This developmental delay is reminiscent of the phenotype displayed by mice heterozygous for the constitutive yy1 knockout allele (20), raising the possibility that the yy1^flox allele was hypomorphic. Consistent with this hypothesis, Western blot analysis demonstrated that YY1 expression in various tissues of the yy1^flox/flox mice was reduced to approximately 50% of that expressed by their wild-type littermates (Fig. ​(Fig.2B2B).

Open in a separate window

FIG. 2.

Hypomorphism of the yy1^flox allele. (A) Weight of the conditional knockout mice. Heterozygous yy1^flox/+ mice were intercrossed, and the weight of their offspring was determined 3 to 6 weeks after birth. The weight of mice heterozygous for the constitutive yy1 knockout allele (yy1^+/−) generated from independent crosses is shown as a comparison. Within each litter, the weight of the mutant (yy1^flox/+, yy1^flox/flox, and yy1^+/−) mice was expressed as the percentage of the average weight of their wild-type (yy1^+/+) littermates of the same sex. Each bar corresponds to the mean ± standard deviation of n mice, with values obtained from several (>5) independent crosses. *, P < 0.001, yy1^flox/flox and yy1^+/− versus yy1^+/+ littermates (Student's t test). (B) YY1 expression in the yy1^flox/flox conditional knockout mice. Western blot analysis of YY1 expression level in various organs harvested from 3-week-old homozygous yy1^flox/flox mice and their wild-type (yy1^+/+) littermates. YY1 was detected with an anti-YY1 (H414) antibody, and an antiactin (MAB1501) antibody was used to verify equal loading of the protein extracts. Densitometry analysis revealed that the YY1 protein level in the liver, lung, kidney, and thymus of the yy1^flox/flox mice corresponded to 52%, 46%, 54%, and 46%, respectively, of that detected in tissues isolated from their wild-type littermates.

Since the presence of selection markers in the targeted locus is the most common cause of hypomorphism, the neomycin resistance cassette was removed by intercrossing the yy1^flox/+ mice with “flipper” mice ubiquitously expressing Flp site-specific recombinase (22). Flp-mediated recombination targeting the Frt sites (see Fig. S1 in the supplemental material) allowed us to generate homozygous yy1^f/f mice, which were born at the expected frequency, were indistinguishable in size and weight from their wild-type littermates, and did not display any other detectable defects (data not shown). As anticipated based on the rescue of the growth retardation phenotype observed after deletion of the PGK-Neo cassette, Western blot analysis revealed that wild-type and yy1^f/f MEFs expressed identical levels of YY1 protein (see Fig. ​Fig.4C;4C; see also Fig. S1 in the supplemental material). These data demonstrate that removal of the selection cassette restores normal expression of YY1 and strongly suggest that the yy1^flox allele is a hypomorphic allele.

Open in a separate window

FIG. 4.

Dose-dependent effect of YY1 on cell proliferation in MEFs. (A to C) Western blot analysis of YY1 expression level in MEFs isolated from yy1^flox/+, yy1^+/−, yy1^flox/−, and yy1^f/f littermates and their wild-type (yy1^+/+) littermates. In panels B and C, MEFs were infected with adenovirus-Cre and harvested at various time points postinfection. In all panels, YY1 was detected with an anti-YY1 antibody (H414), and equal loading of the protein extracts was assessed using an antiactin (MAB1501) antibody. Densitometry analysis revealed that the YY1 protein level in yy1^flox/+, yy1^+/− and yy1^flox/− MEFs corresponded to 71%, 54%, and 23%, respectively, of that detected in MEFs isolated from their wild-type littermates. (D and E) Dosage-dependent growth defects of yy1-deficient cells. yy1^+/+, yy1^flox/+, yy1^flox/−, and yy1^f/f MEFs were infected with adenovirus-Cre, and their growth properties were analyzed by the colony-forming assay. (F) Growth properties of MEFs expressing different levels of YY1. yy1^+/+, yy1^flox/+, yy1^+/−, and yy1^flox/− MEFs were subjected to the MTT assay 6 days postisolation. *, P < 0.014, yy1^+/− versus yy1^+/+ MEFs; **, P < 0.001, yy1^flox/− versus yy1^+/+ and yy1^+/− MEFs (Student's t test, n = 6). (G) Growth kinetics of MEFs after complete depletion of YY1. yy1^+/+ and yy1^flox/− MEFs were infected with adenovirus-Cre, and cell proliferation was assessed by the MTT assay performed at various time points (5, 8, and 10 days) postinfection. In both panels F and G, the results are expressed in relative units (RU) and correspond to the mean ± standard deviation of three values obtained in a representative experiment. *, P < 0.001, untreated yy1^flox/− versus untreated yy1^+/+ MEFs; **, P < 0.001, Cre-treated yy1^flox/− versus Cre-treated yy1^+/+ and untreated yy1^flox/− littermates (Student's t test). (H) Cell cycle profile resulting from complete depletion of YY1. MEFs isolated from yy1^f/f and yy1^+/+ littermates were submitted to FACS analysis 4 days postinfection by adenovirus-Cre. (I) Cytokinesis defects induced by complete depletion of YY1. MEFs isolated from yy1^+/+ and yy1^f/f littermates were infected with adenovirus-Cre, and the number of cells with more than one nucleus was counted at 2, 3, and 6 days postinfection. Each value corresponds to the mean ± standard deviation of three values obtained in independent experiments in which at least 500 cells per sample were scored. *, P < 0.001, Cre-treated yy1^f/f versus Cre-treated yy1^+/+ MEFs (Student's t test).

Intercrossing of the hypomorphic yy1^flox/flox mice with yy1^+/− heterozygotes carrying a null yy1 allele (20) allowed us to generate compound heterozygous mutant embryos (yy1^flox/−) expressing only approximately 25% of the normal amount of yy1 mRNA (see Fig. ​Fig.5;5; see also Fig. S1 in the supplemental material) and protein (see Fig. 4A and B; see also Fig. S1 in the supplemental material). As shown in Table ​Table1,1, genotype analysis of the progeny obtained from such crosses demonstrated that the majority of the compound heterozygous mutants (yy1^flox/−) developed beyond the time of implantation (4.5 to 6 dpc), the stage at which embryonic lethality induced by a complete loss of yy1 function was previously shown to occur (20). Indeed, comparison of the number of yy1^flox/− embryos obtained from our crosses to that predicted by Mendel's law revealed that only ∼25% of the yy1^flox/− animals died before day 14.5 dpc (Table ​(Table1).1). The discrepancies in the frequency of yy1^flox/− embryos expected and obtained became even more pronounced at later stages, indicating that a fraction of embryos died between 13.5 and 14.5 dpc and 16.5 and 19.5 dpc, while another subpopulation died between 16.5 and 19.5 dpc and birth (Table ​(Table1).1). Examination of the morphological appearance of the yy1^flox/− mice before birth (17.5 dpc) revealed that a significant subset of yy1^flox/− embryos were severely growth retarded and displayed other developmental defects with various degrees of severity (Fig. ​(Fig.3A3A).

Open in a separate window

FIG. 3.

Phenotype of the hypomorphic yy1^flox/− mice. (A) Morphological appearance of the growth-retarded subset of hypomorphic embryos. Heterozygous yy1^flox/+ mice were crossed with yy1^+/− mice, and the morphology of their offspring was observed at 17.5 dpc. While the yy1^flox/− embryo shown in the middle is representative of most animals within the growth-retarded subset, the rightmost embryo displays more-severe developmental defects observed in only a few embryos. A wild-type (yy1^+/+) littermate at the same developmental stage is shown for comparison. (B) Weight of the hypomorphic knockout mice at birth. Homozygous yy1^flox/flox mice were crossed with yy1^+/− mice, and the weight of their newborn pups was measured. Within each litter, the weight of the yy1^flox/− mice was expressed as a percentage of the average weight of their yy1^flox/+ littermates, whose weights are indistinguishable from those of their wild-type littermates. Each bar corresponds to the mean ± standard deviation of n embryos, and values were obtained from several (>5) independent crosses. *, P < 0.001, yy1^flox/− versus yy1^flox/+ littermates (Student's t test). (C) Gross morphology of the yy1^flox/− pups at birth. Homozygous yy1^flox/flox mice were crossed with yy1^+/− mice, and the morphological appearance of their newborn pups was analyzed. Two yy1^flox/+ neonates, whose size and appearance are indistinguishable from those of wild-type pups, are shown for comparison. (D). Histological analysis of lungs from yy1^flox/− and yy1^flox/+ newborn littermates stained with hematoxylin and eosin. Bar, 100 μm.

Open in a separate window

FIG. 5.

Confirmation of potential targets of YY1 by semiquantitative RT-PCR. (A) Validation of the microarray data in yy1^flox/− MEFs. Eleven genes whose expression was higher in the yy1^flox/− MEFs than in their wild-type (yy1^+/+) counterparts (Table ​(Table22 and Table ​Table11 in the supplemental material) were selected for RT-PCR analysis using RNAs extracted from untreated yy1^+/+ and yy1^flox/− MEFs 4 days postisolation (left panel). Eleven genes expected to be down-regulated in the mutant cells were also analyzed (middle panel). (B) Deregulation of putative YY1 target genes upon complete depletion of YY1 in MEFs. Nine genes whose expression was induced by partial depletion of YY1 (i.e., untreated yy1^flox/− MEFs) were selected for RT-PCR analysis performed on RNAs isolated from Cre-treated yy1^+/+ and yy1^f/f MEFs at 6 days posttreatment (left panel). The expression of 12 genes expected to be down-regulated upon partial depletion of YY1 was also analyzed (middle panel). In both panels A and B, the YY1 and actin control RT-PCRs are shown in the right panels.

While half of the yy1^flox/− animals died during late embryogenesis, the other half survived until birth (Table ​(Table1).1). This subpopulation of yy1^flox/− animals displayed only a marginal delay in overall growth, the weight of the newborn yy1^flox/− pups representing approximately 80% of that of their wild-type littermates (Fig. ​(Fig.3B).3B). Morphological examination of these embryos did not reveal any obvious developmental defects (Fig. ​(Fig.3C).3C). While the majority of the newborn yy1^flox/− animals were alive at birth, they appeared pale and slightly cyanotic compared to their wild-type littermates, suggesting possible abnormalities in hematopoiesis, vascularization, circulation, and/or respiratory function. Importantly, most yy1^flox/− animals died within the first day after birth, with the remainder dying within the next 24 h.

To gain further insights into the cause of perinatal lethality of the yy1^flox/− mice, embryos were isolated and examined at 19.5 dpc. As expected, most of the wild-type embryos isolated at this stage were able to breathe autonomously. In contrast, the yy1^flox/− animals that were alive and reacted to mechanical stimulation at the time of isolation appeared unable to breathe despite frequent gasping and died within an hour postisolation. Postmortem analysis of lungs isolated from newborn yy1^flox/− pups indicated that they were not filled with air. Consistently, histological examination revealed collapsed alveoli (Fig. ​(Fig.3D),3D), confirming that the lungs failed to inflate at birth. Together, these results demonstrate an essential, dosage-dependent requirement for YY1 in late embryonic development.

Inhibition of YY1 function impairs cell proliferation in primary cells.

As mentioned earlier, generation of the yy1^flox hypomorphic allele allowed us to produce mouse strains expressing different levels of YY1. As shown in Fig. 4A and B, Western blotting and densitometry analysis of YY1 expression in yy^flox/+, yy1^+/−, and yy1^flox/− MEFs demonstrated that these cells expressed ∼75%, 50%, and ∼25% of the normal YY1 level, respectively. In contrast, MEFs isolated from embryos carrying two copies of the conditional yy1^f allele (which lacks the PGK-Neo cassette [see above and Fig. S1 in the supplemental material]) expressed YY1 levels indistinguishable from that detected in their wild-type counterparts (Fig. ​(Fig.4C4C and Fig. S1H in the supplemental material). Time course analysis revealed that infection of yy1^flox/− and yy1^f/f MEFs with adenovirus-Cre led to a gradual decrease in YY1 protein level, YY1 expression becoming undetectable 2 to 3 days postinfection (Fig. 4B and C; see also Fig. S1 in the supplemental material). To gain insights into YY1 cellular functions, we took advantage of these conditional MEFs and examined the phenotype induced by partial and complete inhibition of YY1.

We first analyzed the viability and growth properties of MEFs of various genotypes in the absence and presence of adenovirus-Cre by a colony-forming assay. As shown in Fig. ​Fig.4D,4D, untreated yy1^flox/+ MEFs expressing ∼75% of the normal YY1 complement displayed growth properties very similar to those of wild-type cells. While infection of the wild-type cells by adenovirus-Cre did not significantly affect their growth properties, reduction of the YY1 level to ∼50% in Cre-treated yy1^flox/+ MEFs led to a marked decrease in cell proliferation (Fig. ​(Fig.4D).4D). The growth-suppressive effect induced by depletion of YY1 became even more pronounced in untreated yy1^flox/− cells expressing ∼25% of normal YY1, with complete depletion of YY1 in Cre-treated yy1^flox/− MEFs further increasing the severity of this proliferation defect (Fig. ​(Fig.4D).4D). Interestingly, while reduction of YY1 expression to ∼25% of the normal complement in untreated yy1^flox/− MEFs affected mostly the size of the colonies formed by these cells, a marked decrease in both colony size and colony number was observed after complete depletion of YY1 in yy1^flox/− MEFs infected with adenovirus-Cre (Fig. ​(Fig.4D4D).

Consistent with their ability to express wild-type levels of YY1 (Fig. ​(Fig.4C),4C), no significant differences in cell colony formation were observed when comparing the untreated and yy1^f/f MEFs with their wild-type counterparts (Fig. ​(Fig.4E).4E). As expected based on our previous results (Fig. ​(Fig.4D),4D), infection of the yy1^f/f MEFs with adenovirus-Cre led to a dramatic decrease in cell colony size and cell colony number (Fig. ​(Fig.4E).4E). Immunofluorescence experiments revealed that the few cell colonies formed by the Cre-treated yy1^f/f MEFs were mainly attributable to the presence of YY1-positive cells (i.e., cells that had not been infected by adenovirus-Cre and therefore displayed normal YY1 expression levels and proliferation rates; data not shown). Taken together, these results demonstrate that inhibition of YY1 function in primary cells impairs cell viability and/or proliferation in a dose-dependent manner.

The dosage-dependent effect of YY1 on cell viability and/or proliferation was confirmed by the MTT assay, performed on yy1^+/+, yy1^flox/+, yy1^+/−, and yy1^flox/− MEFs expressing 100%, ∼75%, 50%, and ∼25%, respectively, of the normal amount of YY1 (Fig. ​(Fig.4F).4F). Using a similar approach, we analyzed the growth kinetics of untreated and Cre-treated yy1^+/+, yy1^flox/−, and yy1^f/f MEFs (Fig. ​(Fig.4G4G and data not shown). Consistent with our previous observations (Fig. ​(Fig.4D),4D), the MTT assay confirmed that yy1^flox/− MEFs expressing ∼25% of normal YY1 proliferate at a much lower rate than wild-type cells (Fig. ​(Fig.4G).4G). Despite their impaired growth properties, untreated yy1^flox/− MEFs were found to proliferate steadily throughout the course of the experiment (Fig. ​(Fig.4G).4G). In contrast, complete depletion of YY1 in both Cre-treated yy1^flox/− MEFs (Fig. ​(Fig.4G)4G) and Cre-treated yy1^f/f MEFs (data not shown) abolished cell proliferation.

The decrease in viable cell number caused by depletion of YY1 could result from an increase in cell death, a decrease in cell proliferation, or both. Microscopic observation of the untreated and Cre-treated yy1^flox/− and yy1^f/f MEFs demonstrated that these cells did not detach and/or die (data not shown). Consistently, these cells did not display any characteristics of apoptotic cells, including nuclear fragmentation or cleavage of lamins A and C, and no significant accumulation of apoptotic (sub-G₀) cells was detected by fluorescence-activated cell sorting (FACS) analysis in either untreated or Cre-treated yy1^flox/− and yy1^f/f cell populations (Fig. ​(Fig.4H4H and data not shown).

Besides the appearance of a small subpopulation of cells having an 8N DNA content (see below), the major alteration of cell cycle profile observed upon treatment of the yy1^f/f cells with Cre was an accumulation of cells with a 4N DNA content (Fig. ​(Fig.4H).4H). Indeed, 4 days posttreatment, the tetraploid cell population in the yy1^f/f MEFs treated with Cre was ∼62%, while the percentage of cells with a 4N DNA content in the wild-type (treated or untreated) and untreated yy1^f/f MEFs ranged from 26 to 39% (Fig. ​(Fig.4H).4H). Based on these observations, we conclude that inhibition of YY1 function in primary cells does not trigger apoptosis but impairs cellular proliferation in a dosage-dependent manner. Strikingly, complete depletion of YY1 leads to cell cycle arrest and accumulation of tetraploid cells.

YY1 is required for cytokinesis in primary cells.

The accumulation of tetraploid cells observed after inhibition of YY1 function could be due to the inability of the cells to progress from G₂ phase to mitosis (blockage at the G₂/M transition), to successfully complete mitosis (mitotic arrest, failure of chromosome segregation), to initiate or complete cytokinesis, or to enter G₁ phase. To characterize the proliferative arrest caused by depletion of YY1 in more detail, we examined the cellular and nuclear morphology of the YY1-depleted cells by phase-contrast light microscopy and fluorescence microscopy after 4′,6′-diamidino-2-phenylindole (DAPI) staining. Most cells had nuclei with uncondensed chromosomes, arguing against mitotic arrest (data not shown). However, depletion of YY1 resulted in the gradual accumulation of binucleated cells (Fig. ​(Fig.4I).4I). No sign of DNA condensation or cell cleavage was observed in these binucleated cells (data not shown), strongly suggesting that inhibition of YY1 function results in cytokinesis failure.

Identification of YY1 target genes in primary cells.

While a substantial body of studies established a role for YY1 in transcriptional regulation, most YY1-responsive genes identified so far have yet to be validated as bona fide YY1 target genes in vivo. To gain mechanistic insights into YY1 developmental and cellular functions, we used a genomewide approach to identify YY1 target genes in mouse embryonic fibroblasts. Microarray analysis performed on yy1^+/+ and yy1^flox/− MEFs (expressing ∼25% of the normal complement of YY1) allowed us to identify 524 genes whose expression was affected by a reduction of the YY1 level. The complete list of YY1-responsive genes is provided in Table S1 in the supplemental material, and the target genes whose functions are the most relevant to the phenotypes described in this study are presented in Table ​Table2.2. It is worth mentioning that approximately 90% of the genes affected by partial depletion of YY1 were found to display similar (i.e., comparable or more pronounced) changes in their expression profiles after complete depletion of YY1 in Cre-treated yy1^flox/− MEFs (data not shown).

Consistent with the ability of YY1 to act as a transcriptional activator and a transcriptional repressor, inhibition of YY1 function led to up-regulation of 231 genes and down-regulation of 293 genes. In addition to key regulators of cell cycle progression, various genes involved in cytokinesis, mitosis, DNA replication, and apoptosis were also affected. YY1 depletion also led to the deregulation of a plethora of genes involved in various aspects of cellular functions such as cell survival, growth, and proliferation; oncogenic transformation; cell adhesion/migration; cytoskeleton biogenesis and organization; DNA repair; transcription; and epigenetic regulation of gene expression, metabolism, and signal transduction. Consistent with its essential role during development, YY1 depletion affected the expression of various genes required for cellular differentiation, embryonic patterning, and morphogenesis. Finally, genes required for homeostasis, DNA repair, and stress and immune responses were found among the most-represented functional groups, suggesting that YY1 might influence a broad spectrum of biological processes through a diverse and complex network of transcriptional changes.

To confirm the microarray data, the expression of 22 putative YY1 target genes was analyzed by RT-PCR in untreated yy1^+/+ and yy1^flox/− MEFs (Fig. ​(Fig.5A).5A). A similar approach was used to investigate the effects of complete depletion of YY1 on the expression of 21 of these putative targets in Cre-treated yy1^+/+ and yy1^f/f MEFs (Fig. ​(Fig.5B).5B). Among the candidates analyzed were the genes for cell cycle regulators p21, p57, cdc46/mcm5, polα1, cyclin B1 and cyclin B2, known regulators of mitosis and cytokinesis (Incenp, CenpA, Plk1, Birc-5/Survivin, and Cap-d2), and epigenetic regulators of mammalian development (Suv39h2 and Dnmt3a). The p53 target genes Mdm2, Sip27, Gadd45γ, Dda3, and Perp were selected due to their involvement in p53-dependent regulation of the cell cycle and apoptosis. Several genes involved in the regulation of cell survival, growth, and differentiation during development (e.g., Bdnf, Ntf3, Nsg1, Tgfβ2, and Btg2), as well as a few other genes (BM88, Hmgb3, Nisnap1, and Sfxn1) regulating diverse biological processes, were also studied. The expression of all candidate target genes tested was found to be affected by partial (Fig. ​(Fig.5A)5A) and/or complete (Fig. ​(Fig.5B)5B) depletion of YY1, as predicted based on the microarray results (Table ​(Table2;2; see also Table S1 in the supplemental material).

Depletion of YY1 does not lead to p53 induction in primary cells.

We and others have previously shown that, in some cellular contexts, ablation of YY1 results in elevated levels of the tumor suppressor p53 and subsequent activation of p53-dependent pathways controlling cell growth and/or apoptosis (5, 26, 65). Strikingly, analysis of the microarray data revealed that approximately 40 YY1-responsive genes have previously been identified as p53 target genes (18, 27, 31, 40, 42, 74), their deregulation upon depletion of YY1 mimicking the effects of p53 induction (see Table ​Table2;2; see also Table S1 in the supplemental material). Taken together, these results raise the possibility that the cellular phenotype (i.e., cell cycle arrest) and changes in gene expression profiles induced by depletion of YY1 in mouse embryonic fibroblasts might be mediated, at least in part, via an increase in p53 expression and/or stability.

To investigate this hypothesis, we analyzed the effects of YY1 depletion on the cellular levels of p53 mRNA and protein (Fig. ​(Fig.6).6). Consistent with our microarray data (see Table S1 in the supplemental material, and data not shown), we did not observe any significant effect of partial (i.e., untreated yy1^flox/− MEFs) or complete (i.e., Cre-treated yy1^flox/− MEFs) depletion of YY1 on p53 mRNA levels (Fig. ​(Fig.6A).6A). While a strong increase in p53 protein level was observed in adriamycin-treated wild-type MEFs (used as a positive control for p53 induction), Western blot analysis showed that YY1 depletion did not affect the steady-state level of p53 protein in yy1^flox/− MEFs (Fig. ​(Fig.6B).6B). Consistently, no significant fluctuations of p53 mRNA and protein levels were detected after complete depletion of YY1 in yy1^f/f MEFs (data not shown).Together, these results strongly suggest that the cellular phenotype and molecular targets induced by depletion of YY1 in MEFs are unlikely to result from induction of p53.

Open in a separate window

FIG. 6.

Depletion of YY1 does not affect p53 expression in MEFs. (A) RT-PCR analysis of p53 mRNA level in MEFs isolated from yy1^flox/− mice and their wild-type (yy1^+/+) littermates 2 days postinfection with adenovirus-Cre (i.e., 4 days postisolation). The YY1 and actin RT-PCRs are shown as controls. (B) Western blot analysis of p53 protein levels in yy1^flox/− and yy1^+/+ MEFs was performed at various time points postinfection with adenovirus-Cre, using an anti-p53 antibody (PAB240). Adriamycin treatment (2 μg/ml for 10 h) of wild-type (p53^+/+) and p53 null (p53^−/−) MEFs was used as a control. Equal loading of the protein extracts and Cre-mediated depletion of YY1 were ensured by using antiactin (MAB1501) and anti-YY1 (H414) antibodies, respectively.

Depletion of YY1 inhibits cell proliferation and cytokinesis in HeLa cells.

In order to circumvent some of the limitations associated with mouse embryonic fibroblasts and confirm our results in a different cellular context, we investigated the effect of RNAi-mediated inhibition of YY1 in HeLa cells, in which p53 is inactivated by the human papillomavirus oncoprotein E6.

The effects of YY1 depletion on cell viability and proliferation were assessed by a colony-forming assay (Fig. ​(Fig.7B)7B) and FACS analysis (Fig. ​(Fig.7C),7C), performed after transfection with a control or YY1 RNAi vector and subsequent selection of the transfected cells. Under these conditions, the YY1 protein level was dramatically reduced upon transfection of the YY1 RNAi vector (Fig. ​(Fig.7A).7A). We found that the number and size of cell colonies formed by the RNAi-depleted cells were greatly reduced compared to those of mock-transfected cells (Fig. ​(Fig.7B).7B). This is consistent with the phenotype resulting from complete depletion of YY1 in mouse primary cells (Fig. 4D and E, Cre-treated yy1^flox/− and yy1^f/f MEFs) and reinforces the idea that YY1 is required for cell survival and/or cell proliferation.

Open in a separate window

FIG. 7.

Depletion of YY1 induces proliferation and cytokinesis defects in HeLa cells. (A) Efficient depletion of YY1 after RNAi in HeLa cells. After cotransfection of a control or YY1 RNAi plasmid with a puromycin resistance-encoding vector, the transfected cells were selected by puromycin treatment, and the expression level of YY1 was assessed by Western blot using an anti-YY1 antibody (H414). Equal loading of the protein extracts was confirmed using a monoclonal antiactin (MAB1501) antibody. (B) Effect of YY1 depletion on cell proliferation. HeLa cells were cotransfected and selected as in panel A, and their growth properties were analyzed by the colony-forming assay. (C) Cell cycle profile induced by depletion of YY1. FACS analysis was performed on HeLa cells cotransfected and selected (as described for panel A) 4 days after transfection with the RNAi plasmids. (D to F) Cytokinesis defects and nuclear abnormalities induced by depletion of YY1. HeLa cells transfected with a control or a YY1 RNAi plasmid were plated on coverslips and analyzed by immunofluorescence using an anti-YY1 antibody (H414; red), DAPI to stain the DNA (blue), and an antibody directed against α-tubulin (B512; green). Representative pictures are shown in panel E, and cells displaying more than one nucleus are indicated by solid arrows. Examples of the nuclear abnormalities induced by inhibition of YY1 (scored as other nuclear abnormalities in panel D) are shown in panel F. These nuclear defects include micronuclei (open arrowheads), DNA bridges (solid arrowheads), bilobed and dumbbell-shaped nuclei (open arrows), and other alterations of the nuclear morphology (solid arrows). Note that this panel is a montage generated by selecting only YY1-negative cells displaying nuclear anomalies. In panels E and F, the bar corresponds to 10 μm. The percentages of cells with more than one nucleus or with nuclear abnormalities were determined 4 and 5 days posttransfection. The results are shown in panel D and correspond to the mean ± standard deviation of three values obtained in independent experiments. Within each experiment, at least 500 cells per sample were scored, with only YY1-negative cells being analyzed in the YY1 RNAi samples. *, P < 0.001, YY1 RNAi versus the control (Student's t test).

Microscopic observation of the cells (see below), Western blot analysis of PARP cleavage (a marker of apoptotic cell death; see Fig. ​Fig.8),8), as well as FACS analysis (Fig. ​(Fig.7C)7C) did not reveal any sign of apoptosis in the RNAi-treated cells. Although inhibition of YY1 function also resulted in the appearance of a small fraction of polyploid cells (i.e., cells having a DNA content higher than 4N), the most dramatic effect of YY1 RNAi on the cell cycle profile was a pronounced accumulation of cells having a 4N DNA content (Fig. ​(Fig.7C).7C). Based on FACS analysis of three independent experiments, the percentage of tetraploid cells in the control and YY1 RNAi cell populations was estimated to be 24.7% ± 1.8% and 43.2% ± 1%, respectively.

Open in a separate window

FIG. 8.

RNAi-mediated depletion of YY1 sensitizes HeLa cells to various apoptotic agents. Puromycin-selected HeLa cells were treated with various apoptotic agents (200 nM of staurosporine, 100 μM of etoposide, 100 ng/ml of Fas ligand, and 10 ng/ml of tumor necrosis factor [TNF] alpha) 4 days after transfection of the control and YY1 RNAi vectors. Cell viability (A) and cell death (B) were then quantified at various time points posttreatment. The results shown in panel A correspond to the mean ± standard deviation of three values obtained in independent transfections, the number of viable cells in the treated control and YY1 RNAi cell populations being expressed as a percentage of that observed in the corresponding untreated samples harvested at the same time points. *, P < 0.01, YY1 RNAi versus control cells; **, P < 0.05, YY1 RNAi versus control cells (Student's t test). Western blot analysis of PARP cleavage was used to determine apoptotic cell death (panel B); full-length PARP and cleavage fragment are indicated by a solid and an open arrowhead, respectively. Efficient depletion of YY1 upon transfection of the RNAi plasmid and equal loading of the total cell extracts were ensured by using anti-YY1 (H10) and antiactin (MAB1501) antibodies.

To characterize the proliferative arrest caused by depletion of YY1 in more detail, we examined the cellular and nuclear morphology of the YY1-depleted cells by phase-contrast light microscopy (data not shown) and fluorescence microscopy after DAPI staining (Fig. 7D to F). Costaining of the control and YY1 RNAi-treated cells with anti-YY1 and antitubulin antibodies revealed that most cells affected by the YY1 RNAi had nuclei with uncondensed chromosomes, excluding mitotic arrest as the cause of the accumulation of the tetraploid cells (data not shown; see Fig. ​Fig.7E).7E). The most striking effect induced by inhibition of YY1 function was the gradual accumulation of cells containing more than one nucleus and/or displaying various other types of nuclear abnormalities (Fig. ​(Fig.7D7D).

As shown in Fig. ​Fig.7E,7E, the majority of YY1-negative cells (i.e., in which YY1 nuclear staining was extremely low or undetectable) containing more than one nucleus were binucleated and showed no detectable sign of DNA condensation or cell cleavage. Only a very small percentage of YY1-negative cells accumulated more than two nuclei (Fig. ​(Fig.7E7E and data not shown). The YY1-RNAi treated cells also displayed other nuclear anomalies, including deformed or bilobed nuclei, micronuclei, and DNA bridges, with a high frequency (Fig. 7D and F). Besides these alterations of their nuclear morphology, we found that most cells affected by the YY1 RNAi were abnormally flat and large compared to the control cells, and the size of their nucleus was also increased (Fig. 7E and F, and data not shown). Thus, the phenotype induced by depletion of YY1 in HeLa cells is essentially identical to that observed in mouse primary cells (Fig. ​(Fig.4)4) and strongly suggests that inhibition of YY1 induces cytokinesis failure.

To gain further insights into the molecular mechanisms underlying the role of YY1 in the regulation of cell cycle progression and cell ploidy, we examined the expression of various markers specific to the different phases of the cell cycle (Fig. S2 and S3 in the supplemental material). Based on these stainings, the percentages of cells in G₁, S, G₂/early M (i.e., prophase-end of metaphase), and late M/cytokinesis (i.e., anaphase-telophase) phases in the control and YY1 RNAi cell populations were determined (Table ​(Table33).

YY1 RNAi did not significantly affect the percentage of G₂ cells (data not shown) or G₂/early M cells (Table ​(Table3).3). In contrast, inhibition of YY1 function led to a pronounced diminution of the percentage of cells undergoing cytokinesis during late M phase (Table ​(Table3).3). We did not observe any significant increase in the percentage of G₁ cells in the YY1 RNAi samples, arguing against a G₁ cell cycle arrest. Similarly, depletion of YY1 did not significantly affect the percentage of cells in S phase, strongly suggesting that a significant number of the binucleated cells (representing ∼35% of the YY1-depleted cells) were undergoing DNA replication.

Considering exclusively YY1-negative cells containing more than one nucleus, we determined that 44% ± 3% of such cells contained PCNA-positive DNA replication foci. Consistently, a significant number of binucleated YY1-negative cells (24.3% ± 1%) were found to be in G₂ phase (see also Fig. S2 in the supplemental material). Based on these immunostaining experiments, we conclude that (i) the proliferation defects and accumulation of tetraploid cells induced by depletion of YY1 do not result from a blockage at the G₂/M transition, (ii) depletion of YY1 impairs cytokinesis initiation, leading to the accumulation of cells containing two nuclei, and (iii) binucleated cells do not undergo cell cycle arrest immediately after cytokinesis failure, but instead progress through the next G₁ phase and subsequently undergo at least one round of illegitimate DNA replication leading to polyploidization. The small percentage of YY1-negative cells displaying more than two nuclei strongly suggests that most binucleated cells cannot proceed to the next round of mitosis, but rather undergo cell cycle arrest. Supporting this view, the percentage of YY1-negative cells initiating mitosis (i.e., cells in early M phase) decreased slowly but gradually throughout the course of the experiment, while the percentage of control cells undergoing mitosis was not significantly affected (data not shown).

Taken together, these results demonstrate that complete inhibition of YY1 function results in cytokinesis failure and induces severe cell proliferation defects, uncovering new p53-independent functions of YY1-mediated regulation of cell cycle progression.

We thank Grace Gill (Harvard Medical School, Boston) for critical reading of the manuscript and helpful comments. We are extremely grateful to Shidong Jia and Thomas Roberts (Dana Farber Cancer Institute, Boston) for assistance during the genome-wide gene expression profiling analysis, which was performed at the Dana-Farber Microarray Core Facility. We also thank Roderick Bronson (Harvard Medical School, Boston) for histological analysis of the compound mutant mice.

El Bachir Affar and Frédérique Gay were fellowship recipients from the Taplin Funds for Discovery. Yujiang Shi and Maite Huarte were supported by fellowships from the NIH (F32GM070690) and from the Fundación Ramón Areces, respectively. This work was supported by a grant from the National Institutes of Health (GM53874) to Yang Shi.