|Home | About | Journals | Submit | Contact Us | Français|
Metastasis-associated protein 3 (MTA3) is a constituent of the Mi-2/nucleosome remodeling and deacetylase (NuRD) protein complex that regulates gene expression by altering chromatin structure and can facilitate cohesin loading onto DNA. The biological function of MTA3 within the NuRD complex is unknown. Herein, we show that MTA3 was expressed highly in granulosa cell nuclei of all ovarian follicle stages and at lower levels in corpora lutea. We tested the hypothesis that MTA3-NuRD complex function is required for granulosa cell proliferation. In the ovary, MTA3 interacted with NuRD proteins CHD4 and HDAC1 and the core cohesin complex protein RAD21. In cultured mouse primary granulosa cells, depletion of endogenous MTA3 using RNA interference slowed cell proliferation; this effect was rescued by coexpression of exogenous MTA3. Slowing of cell proliferation correlated with a significant decrease in cyclin B1 and cyclin B2 expression. Granulosa cell populations lacking MTA3 contained a significantly higher percentage of cells in G2/M phase and a lower percentage in S phase compared with control cells. Furthermore, MTA3 depletion slowed entry into M phase as indicated by reduced phosphorylation of histone H3 at serine 10. These findings provide the first evidence to date that MTA3 interacts with NuRD and cohesin complex proteins in the ovary in vivo and regulates G2/M progression in proliferating granulosa cells.
Ovarian granulosa cells surround an oocyte and together with theca-interstitial cells comprise an endocrine syncytium that provides morphological and biochemical support for ovarian follicle development. Follicle cell proliferation and steroid hormone biosynthesis, rigorously controlled by pituitary hormones and autocrine/paracrine growth factors, are prerequisites for development of a healthy follicle and ovulation of a developmentally competent oocyte [1–3]. Follicle growth is a carefully orchestrated developmental process controlled by a tightly regulated program of gene expression driven by multiple transcription factors .
Dynamic alterations in chromatin structure affect DNA accessibility to the transcriptional apparatus and serve as a mechanism for regulating gene expression . Chromatin structure is altered by repositioning nucleosomes along DNA and by covalent histone or DNA modifications, such as acetylation or methylation. Several chromatin-modifying multiprotein complexes have been identified that are composed of different combinations of subunits that have DNA-binding, DNA helicase, and histone-modifying activities. Chromatin-modifying complexes containing histone acetyltransferases are typically associated with open transcriptionally active chromatin, whereas complexes containing histone deacetylases (HDACs) are associated with densely packed transcriptionally inactive chromatin. In addition, some transcription factors recruit HDACs to chromatin and can act as histone modifiers themselves [6–9].
Chromatin remodeling is important in regulating gene expression responsible for controlling the steroidogenic activity of luteinized granulosa cells [10, 11]. SNF2L, official symbol SMARCA1, the ATP-dependent helicase subunit of the imitation switch (ISWI) chromatin-remodeling complex, is highly upregulated in granulosa cells in response to the luteinizing hormone surge . SMARCA1 becomes associated with the steroidogenic acute regulatory protein (STAR) promoter and induces expression of this critical gene during luteinization. In addition, association of specific transcription factors with the STAR promoter has been linked with alterations in histone modifications . These findings suggest that chromatin modifications are necessary for direct action of transcription factors on gene promoters regulating granulosa cell luteinization.
The Mi-2/nucleosome remodeling and deacetylase (NuRD) protein complex is a chromatin remodeling complex reported to function as a transcription repressor [13, 14]. The NuRD complex contains an ATP-dependent chromodomain helicase DNA-binding protein (CHD3 or CHD4) that provides nucleosome-remodeling activity and either HDAC1 or HDAC2 that provides HDAC activity [15–17]. Additional NuRD subunits include methyl-CpG-binding domain (MBD) proteins, retinoblastoma-associated binding proteins, and metastasis-associated (MTA) proteins. Combinatorial assembly of these subunits into the NuRD complex occurs variably depending on species and cell type and can result in distinct functions [5, 13]. In addition to its chromatin-modifying activities, the NuRD complex was shown in HeLa cells to facilitate association of cohesin with chromosomes . This function could be important in gene regulation because of recent reports that cohesin binding regulates three-dimensional chromosome architecture to bring enhancer and promoter sequences together . The NuRD complex also participates with cohesin in the cellular response to DNA damage, perhaps by affecting cell cycle checkpoints [20–24].
In breast cancer cells, MTA3 expression is positively regulated by estradiol, and the MTA3-NuRD complex supports maintenance of epithelial differentiation [25, 26]. Although these findings connect MTA3 expression with hormonal signaling and cell differentiation in breast tumors, little is known about whether this relationship is conserved in normal female reproductive tissues that are tightly regulated by hormonal signaling. In the present study, we show that MTA3 is highly expressed in granulosa cells of ovarian follicles at all stages of follicular development and that it interacts with NuRD and cohesin complex components in the ovary. Depletion of MTA3 from granulosa cells results in reduction of cyclin B expression and slowing of cell proliferation at the G2/M transition. These findings suggest a functional role for MTA3 in supporting granulosa cell proliferation, possibly via interactions with NuRD and cohesin complex proteins.
CD1 mice were obtained from the in-house breeding colony at the National Institute of Environmental Health Sciences (NIEHS). All animal procedures were performed under an approved NIEHS protocol and were consistent with the Guide for Care and Use of Laboratory Animals. For immunohistochemical analysis, ovarian tissue was collected from mouse pups on Postnatal Days 1–19. To obtain granulosa cells for MTA3 splice variant analysis, adult (6- to 8-wk-old) mice were superovulated using a single injection of 5 IU of equine chorionic gonadotropin (Calbiochem), and pure populations of granulosa cells were collected by mechanical separation of cumulus-oocyte complexes. For granulosa cell culture, ovaries were collected from unstimulated mice on Postnatal Day 20, and cells were obtained by puncturing whole ovaries with a 30G needle in Dulbecco modified Eagle medium-Ham F-12 medium (DMEM/F-12) (1:1; Invitrogen). Cells were then collected by brief centrifugation at 200 × g, and viability was determined by trypan blue exclusion. Granulosa cells were seeded (105 cells/well) into 12-well plates (Corning) and cultured in DMEM/F-12 (1:1) in the presence of 10% fetal bovine serum (Hyclone) and media supplements. Cells were washed twice with culture medium 6 h after plating, and then treatments (viral infection) were applied.
Reproductive tract tissues were fixed in cold formalin for 24 h, embedded in paraffin, and cut at 6 μm. For antigen retrieval, slides were treated with citrate buffer (pH 6.0; Biocare Medical) and then placed in decloaker solution (Biocare Medical). Endogenous peroxidase was quenched with 3% H2O2. Tissues were blocked with 10% bovine serum albumin (BSA) and incubated with affinity-purified rabbit polyclonal MTA3 antibody (3.75 μg/ml)  in 1× automation buffer (AB) (Biomeda Corp.) containing 1% BSA and 1% milk. Slides were rinsed and incubated with 3 μg/ml of biotinylated anti-rabbit IgG (Vector Laboratories) in AB for 30 min, followed by incubation for 30 min in ExtrAvidin peroxidase (Sigma) diluted 1:50 in AB. Protein was visualized using NOVARed (Vector Laboratories).
For immunoprecipitation of MTA3 and CHD4, ovaries from 20-day-old mice were homogenized in immunoprecipitation (IP) lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM edetic acid [EDTA], 1% NP-40, and 5% glycerol). For RAD21 immunoprecipitation, ovarian nuclear extracts were obtained in a two-step isolation process using cytosolic (50 mM Tris-HCl, pH 7.4, 1% NP-40, 0.25% deoxycholic acid, and 1 mM EDTA) and nuclear (1% sodium dodecyl sulfate, 50 mM Tris-HCl, pH 8.1, and 10 mM EDTA) extraction buffers. Protein concentration was determined using the BCA protein assay (Pierce). A total of 1000 μg of total protein (MTA3 and CHD4) or 50 μg of nuclear protein (RAD21) was brought to a 600-μl volume with IP lysis buffer and precleared with 40 μl of a 50% slurry of agarose beads (Pierce). Samples were incubated overnight at 4°C with either rabbit IgG (Santa Cruz), CHD4 (Abcam), MTA3, or RAD21 (Cell Signaling) antibodies at 1.5 mg of antibody per milligram of input protein. Immune complexes were captured by incubation with protein A/protein G agarose resin and eluted using an immunoprecipitation kit (Pierce) following the manufacturer's instructions.
Total protein was isolated from cultured granulosa cells or from postnatal ovaries using tissue protein extraction reagent (T-PER; Pierce). For cellular fractionation, we used nuclear, cytoplasmic, and membrane protein extraction reagents (Pierce) following the manufacturer's instructions. Protein content was determined using the BCA protein assay. Protein extracts (50 μg/lane for γH2AX and 10 μg/lane for all other tested proteins) were separated on 4%–12% Tris-glycine minigels and transferred to polyvinylidene fluoride membrane (Invitrogen). Blots were incubated with antibodies at the following concentrations: MTA3 (0.375 μg/ml), cyclin B1 (0.025 μg/ml; Cell Signaling), γH2AX (0.025 μg/ml; Cell Signaling), Trp53 (0.2 μg/ml; Santa Cruz), CHD4 (0.375 μg/ml; Abcam), HDAC1 (0.2 μg/ml; Santa Cruz), or RAD21 (0.075 μg/ml; Cell Signaling). To check protein loading, immunoblots were stripped and reprobed using either actin (Santa Cruz) or glyceraldehyde phosphate dehydrogenase (GAPDH) (Abcam) antibodies. Following extensive washes with Tris-buffered saline with Tween 20, either a horseradish peroxidase-conjugated secondary anti-rabbit IgG (Amersham), anti-mouse IgG (Abcam), or anti-goat IgG (Jackson Immunoresearch) was applied. Signals were visualized using Pierce Femto or Pico reagents and exposure to film.
Total RNA was isolated from granulosa cells using either a PicoPure RNA isolation kit (Arcturus) or TRIzol reagent (Invitrogen). Reverse transcription was performed using SuperScript first-strand synthesis system (Invitrogen). To determine which MTA3 splice forms were present in granulosa cells, primers were designed based on sequences of the four splice main forms found in the University of California Santa Cruz (UCSC) genome browser (http://genome.ucsc.edu/). The PCR was carried out as follows: 5 min at 95°C, then 30 cycles of 95°C for 30 sec, 58°C for 30 sec, and 72°C for 60 sec. The PCR products were resolved using a 1% agarose gel and visualized using a BioSpectrum imaging system (UVP). Real-time PCR with detection by differentially labeled TaqMan minor groove binding probes (TaqMan PCR; Applied Biosystems) was used to determine the ratio of MTA3 sequences containing a three-nucleotide insertion in exon 12. Standard real-time PCR to test expression of various genes relative to ribosomal protein L19 (Rpl19) was performed using gene-specific primers and Power SYBR Green Master Mix (ABI). Relative expression was calculated using the delta-delta cycle threshold method . Real-time and TaqMan PCR was carried out using an ABI Prism 7200 sequence detection system. The primers and probes used for all analyses are listed in Supplemental Table S1 (available online at www.biolreprod.org).
Plasmid pLKO.1 carrying either a scrambled or Mta3-specific short hairpin RNA (shRNA) sequence was purchased from Sigma. The sequence of Mta3-specific shRNA was as follows: 5′-CCGGCGGCAAAGATTTCAACGACATCTCGAGATGTCGTTGAAATCTTTGCCGTTTTTG-3′. The lentivirus was packaged and titered as previously described . Multiplicity of infection was expressed in transducible units per cell (TU/cell) and calculated based on the number of viable granulosa cells plated. Viral particles were added to the cell culture 6 h after seeding.
Granulosa cell-derived MTA3 variants 1A and 1B cDNA was tagged at the N-terminus by PCR amplification using a primer containing a 2× FLAG sequence. The FLAG-tagged sequences were cloned into the BamHI and XhoI sites of pEYFP-N1 (Clontech) to add enhanced yellow fluorescent protein (EYFP) at the C-terminus. The FLAG-tagged and FLAG/EYFP-tagged sequences were inserted by recombination into destination viral vector pDest-673 (generously provided by Dr. Dominic Esposito of the National Cancer Institute). An RNA interference-resistant mutant of the 2× FLAG/MTA3 sequence was generated using the QuikChange Lightning Site-Directed mutagenesis kit (Stratagene). Primers were designed to create a mismatch in the region of Mta3 mRNA targeted by the Mta3-specific shRNA, while at the same time keeping the amino acid sequence unchanged. The sequences of all primers used for cloning are provided in Supplemental Table S1.
Cell proliferation was measured 24, 48, 72, and 96 h after lentiviral infection using an MTS cell proliferation assay (Promega). Briefly, 20 μl of staining solution was added to cells at the end of treatment. After incubation for 1–4 h, absorbance at 490 nm was measured.
Granulosa cells were seeded (105 cells/well) in glass-bottom 12-well plates (MatTek) and infected with lentivirus carrying MTA3-EYFP. After 96 h, 5 μM DRAQ5 (Biostatus Limited) was added to stain DNA. Confocal images were recorded using a Zeiss LSM 510-UV META microscope and processed using Photoshop software (Adobe).
Granulosa cells were collected 96 h after lentiviral infection, fixed in 70% ethanol, permeabilized in 0.2% Triton X-100, and incubated with mouse anti-phospho-histone H3 antibody (10 μg/ml; Cell Signaling) for 2 h. The secondary antibody was fluorescein isothiocyanate-conjugated anti-mouse IgG (12 μg/ml; Jackson Immunoresearch). DNA was stained with 3 μg/ml of propidium iodide (PI). Flow cytometric analysis was carried out using an FACSort flow cytometer equipped with CELLQuest software (Becton Dickinson Immunocytometry Systems). Individual cells (7500 per experimental sample) were selected by gating on a PI area vs. width dot plot to exclude cell aggregates and debris. Cell cycle analysis was performed using ModFit software (Verity Software House).
Statistical analyses were performed using Prism software (GraphPad Software, Inc.). The statistical tests used are indicated in the corresponding figure legends.
To gain new insights into the role of chromatin remodeling during ovarian follicle development, expression of the NuRD component MTA3 was examined in mouse ovaries. MTA3 was expressed in maturing follicles at all stages of postnatal development as indicated by immunohistochemical analysis (Fig. 1A). In the somatic cell compartment, intense positive staining was observed in granulosa cells of all follicle stages (primordial through preovulatory). There was some staining in theca-interstitial and corpus luteum cells, but the signal was much weaker than in granulosa cells. MTA3 staining in all somatic cell types was most intense in the nuclei, although some cytoplasmic signal was also observed. MTA3 was also expressed in developing oocytes at all stages, from primordial to fully grown. Expression of MTA3 in the ovary increased with postnatal age as indicated by immunoblotting (Fig. 1B). To determine if MTA3 expression was induced by gonadotropin stimulation, real-time PCR was performed on ovary cDNA from unstimulated adults and adults treated 48 h previously with equine chorionic gonadotropin. There was no difference in the relative expression of MTA3 after gonadotropin stimulation (data not shown).
To confirm the apparent nuclear localization, granulosa cell extracts were separated into membrane, cytoplasm, and nuclear fractions. Immunoblotting of these fractions indicated that the vast majority of MTA3 localizes to the nuclear compartment (Fig. 1C). Of note, a second immunoreactive band that migrated slightly slower than the predominant MTA3 band was consistently observed in immunoblots of ovary protein, suggesting the possibility of a posttranslational modification. This second band remained even after the protein extracts were treated with phosphatase, indicating that it was not a result of MTA3 phosphorylation (data not shown).
Initial computational analysis using the UCSC genome browser revealed four major MTA3 splice variants that encode proteins composed of 513–591 amino acids (Fig. 2A). In addition, one reported MTA3 transcript differed from variant 1 by insertion of three nucleotides encoding a serine residue. This variant, which we referred to as variant 1A (accession number uc008dsf.1), can be explained by usage of an alternative splice site three nucleotides upstream of the exon 12 splice junction used by variant 1B (accession number uc008dsg.1) (Fig. 2A). Of note, all four MTA3 isoforms retain the highly conserved protein-protein interaction domains typical of the MTA3 family, including histone interaction, GATA-like zinc finger, and DNA-binding domains, suggesting that all are capable of serving within protein complexes interacting with chromatin.
To determine which of these MTA3 forms were expressed in granulosa cells, RT-PCR was used. Granulosa cell cDNA encoding three regions of MTA3 was amplified: exons 2–7, exons 7–12, and exon 12 through a 3′ untranslated region (3′-UTR) that was unique for each variant (Fig. 2A). All variants were present in mouse granulosa cells (Fig. 2B), although variant 1 appeared to be the most abundantly expressed based on the intensity of the amplified band. The PCR products were sequenced to confirm their identity, revealing that all four variants sometimes included the three-nucleotide insertion. To determine the relative expression of the two subtypes (A and B) in granulosa cells, we designed fluorescent TaqMan PCR probes to differentiate between the variants. In granulosa cells isolated from Postnatal Day 20 ovaries, expression of subtype B (not containing the serine residue) was approximately twice that of subtype A (Fig. 2C). This subtype B:subtype A ratio was consistently observed in cDNA isolated from ovaries at all stages of postnatal development and from adult ovaries after gonadotropin stimulation (data not shown). To determine if the subtype A serine residue affected protein localization, EYFP-tagged forms of MTA3 variants 1A and 1B were expressed in granulosa cells. Both variants localized only to the nuclei; localization of variant 1A is shown in Figure 2D.
It has been suggested that members of the MTA family are mutually exclusive within the NuRD complex [5, 13]. To date, all three members of the MTA family have been found to be NuRD subunits in different cell types [25, 29–31]. To determine whether MTA3 is a component of the NuRD complex in ovaries in vivo, we performed coimmunoprecipitation experiments using antibodies specific for two major subunits of the NuRD complex, CHD4 and HDAC1.
Immunoprecipitation of CHD4 from whole-ovary extracts resulted in coprecipitation of both MTA3 and HDAC1 (Fig. 3A). The reciprocal was also true (i.e., immunoprecipitation of MTA3 resulted in coprecipitation of CHD4 and HDAC1) (Fig. 3B). These findings indicate that MTA3 interacts with NuRD complex proteins in vivo. Because the NuRD complex participates in cohesin loading onto chromosomes in HeLa cells , we tested whether MTA3 interacts with RAD21, a major cohesin complex protein. RAD21 coimmunoprecipitated with MTA3 (Fig. 3C), indicating that these two proteins also interact in ovarian cells in vivo.
To determine the function of MTA3, we knocked down MTA3 in primary mouse granulosa cells proliferating in vitro. Granulosa cells were infected with a lentiviral construct encoding an shRNA targeting a region of Mta3 conserved in all variant forms. Uninfected granulosa cells or cells infected with a scrambled shRNA sequence were used as controls. A dose-dependent decrease in Mta3 mRNA was observed 96 h after infection only in Mta3 shRNA-treated cells (Fig. 4A). Twenty TU/cell of Mta3 shRNA resulted in a 95% decrease in Mta3 mRNA 96 h after infection compared with control cells infected with the scrambled shRNA sequence. MTA3 protein was partially reduced 48 h after infection with both 10 and 20 TU/cell of Mta3 shRNA and was not detected at all 96 h after infection with 20 TU/cell (Fig. 4B). No change in MTA3 protein was observed in control cells at either time point.
Treatment of proliferating granulosa cells with Mta3 shRNA resulted in an apparent loss of confluence compared with controls, suggesting that the cells were proliferating more slowly or that cell death was occurring. Cell proliferation over time after infection was measured using a metabolic assay. Control cells treated with scrambled shRNA proliferated at a rate similar to uninfected cells, increasing their number almost eight-fold over 96 h (Fig. 5A). MTA3-depleted cells proliferated but not as rapidly as controls, such that by 96 h after infection their number was 40% lower than the scrambled shRNA-treated cells. To determine whether the inhibitory effects of Mta3 shRNA on cell proliferation were due specifically to loss of MTA3 protein, we coinfected the cells with a lentiviral expression vector encoding MTA3 variant 1A mutated in the region targeted by Mta3 shRNA but encoding an amino acid sequence identical to the wild-type variant. Granulosa cells treated concomitantly with Mta3 shRNA and lentivirus encoding exogenous MTA3 variant 1A proliferated normally, suggesting that inhibition of cell proliferation was a specific result of MTA3 protein depletion (Fig. 5B). Similar results were observed when MTA3 variant 1B was coexpressed with the Mta3 shRNA (data not shown). These rescue experiments suggested that cells lacking MTA3 do not undergo apoptosis but rather that the inhibitory effect on proliferation was caused by slowing of the cell cycle or induction of earlier quiescence in culture.
To determine whether loss of MTA3 was associated with specific alterations in the cell cycle, we compared the expression of various cyclins in the treated cells vs. controls. There was a decrease in mRNA encoding both cyclin B1 and cyclin B2 by 48 h after Mta3 shRNA treatment was initiated (Fig. 6A). This finding was confirmed at the protein level for cyclin B1 (Fig. 6B). Expression of cyclins A1, D1, D2, E1, and G1 was no different between treatment groups (data not shown). Interestingly, downregulation of cyclin B1 and cyclin B2 expression preceded the inhibitory effects on cell proliferation, suggesting that joint action of MTA3 and cyclin B is required during preparation for mitosis.
It was previously shown that inhibitory effects on cell proliferation similar to those we induced by MTA3 depletion can also be induced by depletion of CHD4, the major NuRD complex subunit, in immortalized human osteosarcoma U2OS cells . In that study, slowing of proliferation was apparently a result of accumulation of spontaneous DNA damage characterized by increased phosphorylation of histone γH2AFX, changes in CDKN1A (previously p21) and TP53 (previously p53 [human homolog of mouse Trp53]) expression, and activation of the G0/S checkpoint. To determine if a similar mechanism explained our findings, we examined expression of Cdkn1a and Trp53 in MTA3-depleted granulosa cells. Expression of Cdkn1a and Trp53 mRNA was no different in MTA3-depleted cells compared with controls (Fig. 6C). Because TRP53 protein is typically regulated by ubiquitin-mediated protein degradation, the level of TRP53 was also examined by immunoblotting. We found that TRP53 protein was unchanged; moreover, phosphorylation of γH2AFX was unaffected by MTA3 knockdown (Fig. 6D).
We next examined whether there was any alteration in the percentage of cells in the various stages of the cell cycle in MTA3-depleted cells compared with controls. Flow cytometry experiments revealed that cell populations lacking MTA3 contained a significantly higher percentage of cells in G2/M phase and a lower percentage of cells in S phase compared with controls (Fig. 7A). There was no measurable sub-G1 fraction in either treatment group, indicating a lack of apoptotic cells. Because the G2/M transition appeared to be slowed, we examined phosphorylation of histone H3 at serine 10 (H3Ser10), a marker of a successful G2/M shift. Phosphorylation of histone H3 at this epitope was significantly reduced (Fig. 7B), suggesting that reduction in MTA3 affects the G2/M transition of cells and results in delayed entry into mitosis.
In the present study, we characterized MTA3 in the mouse ovary and examined its function in primary mouse granulosa cells. MTA3 not only interacts with NuRD complex proteins, as expected, but also coprecipitates with the major cohesin complex subunit RAD21. MTA3 protein knockdown in primary granulosa cells results in slowing of cell proliferation associated with reductions in cyclin B1 and cyclin B2 expression and a delay in the G2/M cell cycle transition. These findings demonstrate that MTA3 is a NuRD complex constituent that can interact with the cohesin complex in the ovary and suggest that the MTA3-NuRD complex functions with the cohesin complex to support cell cycle progression and growth of the granulosa cell compartment during follicle development in vivo.
Based on coprecipitation with CHD4 and HDAC1, at least some of the NuRD complex in the ovary contains MTA3. Although this does not exclude the possibility that MTA3 could act in association with other complexes or alone, it is likely that the slowing of cell proliferation is caused by alterations in NuRD complex function that result from a lack of MTA3. Indeed, knockdown of MBD3, a subunit required for NuRD integrity and stability, delays the G2/M transition and is associated with downregulation of cyclin B1 in several cancer cell lines . Similar delays in the G2/M transition associated with reduced cyclin B1 expression are observed following chemical and functional inhibition of HDAC1 . These observations suggest that MTA3 is required to support effective chromatin-remodeling activity of the NuRD complex in granulosa cells.
Cells duplicate their DNA content during S phase, and toward the end of this phase, newly synthesized sister chromatids are held together by the cohesin complex, particularly in pericentromeric regions, to facilitate faithful chromosome segregation during M phase . In human cell lines, association of the cohesin complex with chromatin is facilitated by both ISWI and NuRD chromatin-remodeling complexes through their SMARCA5 and CHD4 subunits, respectively, which interact directly with the core cohesin component RAD21 . Interestingly, NuRD associates with pericentromeric heterochromatin during late S phase in rapidly proliferating lymphoid cell lines and in primary germinal center B cells, suggesting that it facilitates cohesin loading in lymphoid cells . In the present study, we show that, similar to the ISWI complex subunit SMARCA5 , MTA3 expression is significantly lower in terminally differentiated nondividing corpus luteum cells than in ovarian granulosa cells. These findings, combined with the observation that MTA3 depletion slows granulosa cell proliferation by affecting the G2/M transition, suggest that both the ISWI and NuRD complexes function to facilitate cohesin loading and sister chromatid stabilization during mitosis in granulosa cells.
The NuRD components MTA1, MTA2, and CHD4 facilitate DNA double-strand break repair , and this effect could be mediated by modulation of cohesin association with repair foci. It is unlikely that MTA3 depletion in granulosa cells resulted in slowing of cell proliferation due to accumulation of DNA double-strand breaks. This statement is based on our findings that MTA3 depletion did not affect Cdkn1a or Trp53 expression, TRP53 protein levels, or γH2AFX phosphorylation. Instead, slowing of the transition through the G2/M checkpoint due to inhibition of cyclin B expression or slower cohesin complex loading in the absence of MTA3 is a more likely explanation.
The concerted activity of several kinases is required for mitotic entry to occur with normal timing. One such kinase, aurora B, phosphorylates H3Ser10 in pericentromeric chromatin regions beginning in late G2 phase . This phosphorylation mark spreads along the chromosomes as mitosis proceeds and is complete in prophase; thus, it can serve as a mark of G2/M progression . In the rat, H3Ser10 phosphorylation is most prominent during proestrus in granulosa cells of growing follicles , confirming an association of this mark with estrogen-mediated proliferation of primary granulosa cells. In mouse primary granulosa cell cultures, we found that MTA3 depletion caused more than a two-fold decrease in cells positive for H3Ser10 phosphorylation; however, less than 1% of cells in either treatment group exhibited this mark. The low incidence of H3Ser10 phosphorylation in both groups is consistent with the relatively slow cell proliferation we observed in these primary cells cultured in the absence of mitogenic stimuli like follicle-stimulating hormone or estradiol.
Later stages of chromosome condensation rely on the master mitotic complex cyclin-B1-CDK1, which triggers mitotic entry by initiating nuclear envelope breakdown . At this point, cells become committed to mitosis and proceed into prometaphase, even if stressed . Expression of type B cyclins in normally proliferating animal cells begins to rise in late S phase, reaches a maximum in the middle of G2/M phase, and drops immediately upon completion of mitosis . Transcription of cyclin B1 and cyclin B2 is regulated by transcriptional activators and repressors that bind promoter elements common to many cell cycle genes . We found that expression of cyclin B1 and cyclin B2 is significantly reduced following MTA3 knockdown. One explanation for this finding is that MTA3 serves directly as a transcriptional activator or coactivator. There are precedents for MTA proteins serving in this role because MTA1 stimulates PAX5 transcription in B cells by a direct interaction with its promoter and interacts with the transactivator HBx to promote nitric oxide synthase 2 transcription in hepatocellular carcinoma cells [44, 45]. A direct role for MTA3 in regulating cyclin B expression via promoter interactions is possible even though there is no evidence published to date that MTA3 serves as a transcriptional activator. However, it was recently shown that the cohesin complex can function to modulate chromatin architecture in ways that stabilize enhancer-promoter interactions important for gene transcription . This observation raises the possibility that MTA3 and/or the MTA3-NuRD complex supports cohesin-chromatin interactions required to promote sufficient cyclin B expression to transit the G2/M checkpoint.
Taken together, our studies indicate that MTA3 is a NuRD complex subunit in granulosa cells and support the idea that the MTA3-NuRD complex functions to support chromatin remodeling required for cohesin loading and G2/M progression. These findings highlight the critical role of the NuRD chromatin-remodeling complex in cell cycle control, in addition to its well-documented functions in modulating epigenetic marks that control gene transcription and in DNA double-strand break repair.
We thank Gary Powell, Negin Martin, Carl Bortner, and Maria Sifre (NIEHS) for technical support; Grace Kissling (NIEHS) for help with statistical analysis; and Dominic Esposito (National Cancer Institute, National Institutes of Health) for providing the viral expression vector pDest-673.
1Supported by grants Z01-ES102985, and Z01-ES101965 from the National Institute of Environmental Health Sciences, Intramural Research Program of the National Institutes of Health.