Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 218 (9), 2896-2918

CDK2 Regulates the NRF1/ Ehmt1 Axis During Meiotic Prophase I

Affiliations

CDK2 Regulates the NRF1/ Ehmt1 Axis During Meiotic Prophase I

Nathan Palmer et al. J Cell Biol.

Abstract

Meiosis generates four genetically distinct haploid gametes over the course of two reductional cell divisions. Meiotic divisions are characterized by the coordinated deposition and removal of various epigenetic marks. Here we propose that nuclear respiratory factor 1 (NRF1) regulates transcription of euchromatic histone methyltransferase 1 (EHMT1) to ensure normal patterns of H3K9 methylation during meiotic prophase I. We demonstrate that cyclin-dependent kinase (CDK2) can bind to the promoters of a number of genes in male germ cells including that of Ehmt1 through interaction with the NRF1 transcription factor. Our data indicate that CDK2-mediated phosphorylation of NRF1 can occur at two distinct serine residues and negatively regulates NRF1 DNA binding activity in vitro. Furthermore, induced deletion of Cdk2 in spermatocytes results in increased expression of many NRF1 target genes including Ehmt1 We hypothesize that the regulation of NRF1 transcriptional activity by CDK2 may allow the modulation of Ehmt1 expression, therefore controlling the dynamic methylation of H3K9 during meiotic prophase.

Figures

Figure 1.
Figure 1.
H3K9me2 expression is abnormally retained in pachytene stage mutant spermatocytes. Immunostaining of P18 meiotic chromosome spreads from wild-type, Prdm9KO, Sun1KO, Cdk2T160A, Cdk2KO, Cdk2D145N, and Speedy AKO mice. Representative images of leptotene, zygotene, and pachytene stages (or the closest possible stage) of prophase I of the first meiotic division according to wild type are shown as indicated (top to bottom). For Prdm9KO, Sun1KO, SpeedyAKO, Cdk2KO, and Cdk2D145N, meiotic arrest of spermatocytes occurs before the completion of pachytene, whereas Cdk2T160A show normal synaptonemal complex dynamics during early meiotic prophase I, but arrest later due to an unrelated meiotic defect (unpublished data). In A–C, costaining of SYCP3 (red) and H3K9me2 (green) is shown, whereas in D–F, costaining of SYCP3 (red) and SYCP1 (green) is shown. At least 50 images of distinct spermatocytes were analyzed for each genotype and stage for each costaining. Original images were individually pseudo-colored and combined as indicated using Adobe Photoshop CC 2018. Scale bars, 5 µm.
Figure 2.
Figure 2.
Immunofluorescence analysis of P16 wild-type and Cdk2KO testis. Immunostaining was performed on P16 testis sections from wild-type (left) and Cdk2KO mice (right). Double immunostaining of EHMT1 (white) or H3K9me2 (green) with either SYCP3, a marker of spermatocytes for all stages of meiotic prophase I (red; A and B), or CDK2 (purple; C and D). For each image, areas of interest are highlighted (white boxes) and displayed as magnified images toward the upper right-hand side. Additional single channel images for either EHMT1 or H3K9me2 are also shown to the lower right-hand side. This is to indicate where costaining has occurred in the merged image. For wild-type images, yellow arrowheads indicate early prophase I (preleptotene, leptotene, and zygotene) spermatocytes, and white arrowheads indicate pachytene stage spermatocytes. For Cdk2KO images, potential pachytene-arrest stage spermatocytes are indicated by white asterisks. At least 25 images of distinct areas of seminiferous tubules were analyzed for each genotype and for each costaining. Original images were individually pseudo-colored and combined as indicated using Adobe Photoshop CC 2018. Scale bars, 50 µm.
Figure 3.
Figure 3.
EHMT1 expression and H3K9me2 levels are increased in Cdk2KO testis. (A) Western blotting of EHMT1, EHMT2, H3K9me2, pan-histone H3, γ-tubulin, and CDK2 in P56 wild-type and Cdk2KO testis (three biological replicates are shown for each). The expression of EHMT1 and EMHT2 at the protein level was increased in Cdk2KO testis compared with wild-type. This was also associated with an increase in H3K9me2 levels. Pan-histone H3 is displayed as a loading control. (B) qPCR analysis of P56 wild-type and Cdk2KO testis. Expression is displayed as fold change normalized to the expression of the eEF2 housekeeping gene. Error bars are representative of the SD of normalized fold change values from at least three biological replicates as compared to wild type. Gene expression between biological replicates was assumed to follow a normal distribution but this was not formally tested. The levels of Ehmt1 transcript are significantly increased as compared with wild-type (***, P < 0.001, as determined by unpaired t test).
Figure 4.
Figure 4.
CDK2 binds to chromatin at promoter regions enriched for NRF1. (A) Pie chart indicating the percentages of CDK2 bound genomic loci found within intergenic, intragenic, or promoter regions. Promoter regions here are defined as within ±5 kb of a transcription start site. (B) Density plot of mapped reads centered on genomic loci positive for CDK2 binding. Density plots of histone mark occupation from isolated spermatocyte DNA are shown for comparison (Hammoud et al., 2014). The majority of CDK2-bound regions are occupied by the epigenetic marks of active transcription (H3K4me3, H3K4me1, and H3K27ac) but not by the transcriptionally repressive H3K27me3 mark. Scale shows ±5 kb relative to the center of CDK2-bound sites. (C) Motif analysis of genomic loci positive for CDK2 binding: The over-represented motif from CDK2-bound loci as generated by MEME suite (top) and NRF1-binding motif extracted from JASPAR database of transcription factor binding motifs (bottom). The significance of the identified CDK2 binding motif to the NRF1 binding motif as determined by GREAT INPUT (McLean et al., 2010) is P = 6.7444 × 10−6. (D) Venn diagram illustrating the overlap of CDK2-bound and NRF1-bound genomic regions from CDK2 and NRF1 ChIPseq analysis in testis. (E) Representative NRF1/CDK2 ChIP-reChIP analysis of the Ehmt1 promoter via RT-qPCR. Two ChIP-reChIPs were performed using either CDK2 or NRF1 antibodies followed by the reciprocal antibody as indicated. Control immunoprecipitations were also performed using nonspecific rabbit IgG followed by CDK2 or NRF1 antibodies. Immunoprecipitated genomic DNA was used for RT-qPCR directed against the region of the Ehmt1 promoter identified as bound by both CDK2 and NRF1 by ChIPseq. (F) Representative ChIP-qPCR for multiple genomic loci using antibodies against NRF1 or nonspecific rabbit IgG for immunoprecipitation. Enrichment of specific NRF1 binding to genomic loci is shown relative to input. Each ChIP was performed on chromatin extracted from corn oil–treated (black bars) and tamoxifen-treated (white bars) Cdk2-iKO testis at P68, 12 d after corn oil/tamoxifen treatment (started at P56). CDK2 and NRF1 ChIPseq experiments were performed using pooled spermatogenic cells from distinct groups of animals (biological replicates) at least three times with similar results. One representative dataset was analyzed to create these figures, as shown in Tables S1 and S3, respectively. ChIP-reChIP and ChIP-qPCR experiments were performed using pooled spermatogenic cells from distinct groups of animals (biological replicates). These experiments were repeated at least four times for each genomic loci analyzed, with similar results. Error bars for ChIP-reChIP data are representative of the SD of at least three technical replicates of the same experiment.
Figure 5.
Figure 5.
NRF1 binds to CDK2 and is a novel CDK2 substrate in vitro. (A) HEK-293T cells expressing either empty vector control, NRF1-MYC, or CDK2-HA, or coexpressing NRF1-MYC and CDK2-HA were lysed as described. NRF1-MYC could be coimmunoprecipitated with CDK2-HA and vice versa (lane 4). Bottom panels show Western blots of 10% of total protein lysate used for immunoprecipitation (Input). HSP90 was used as a loading control. (B) Western blotting of immunoprecipitates from whole testis lysate. Lane 1: 1% of input (16 µg of whole testis lysate); lane 2: IP CDK2; lane 3: IP NRF1; lane 4: IP nonspecific IgG. Coimmunoprecipitation can be seen in lanes 2 and 3 of both of the upper and lower panels. (C) Evolutionary conservation of the amino acid sequence surrounding serines 102 and 136 of NRF1 in multiple species. Potential CDK phosphorylation sites are highlighted in pink. (D) Representative kinase assay using 1 µg of wild-type (lane 1), S136A mutant (lane 2), or S102A/S136A (lane 3) full-length NRF1–GST fusion proteins as a substrate of CDK2/cyclin A in the presence of radiolabeled ATP (upper panel). Quantification of the corresponding phospho-signal was done by PhosphoImager using the Multi Gauge software (Ver3.X) from three replicate experiments as displayed below. Significance was determined by one-way ANOVA. Data were assumed to be normally distributed, but this was not formally tested (****, P < 0.0001). Error bars are representative of the SD of phospho-signal values from at least three technical replicates of the same experiment. Equal protein loading is shown by Coomassie staining of the dried gel (lower panel). (E) EMSA using a probe of the Ehmt1 promoter. A DIG-labeled Ehmt1 promoter probe was incubated with increasing concentrations of either CDK2/cyclin A2 phosphorylated NRF1 (pNRF1; lanes 6–8; 1, 2, or 3 µg, respectively) or nonphosphorylated NRF1 (lanes 3–5; 1, 2, or 3 µg, respectively). “Probe only” and “Probe + Poly DIDC only” conditions are shown in lanes 1 and 2, respectively. Shifted probe (complexes) or free probe (probe) were detected via anti-DIG-alkaline phosphatase (AP) antibody treatment followed by chemiluminescent detection of alkaline phosphatase. Protein loading is shown by both Western blotting and Coomassie staining of NRF1 and pNRF1 on separate SDS-PAGE gels. All experiments were repeated at least three times with similar results. One representative image is shown in each figure. AU, arbitrary units.
Figure 6.
Figure 6.
NRF1 target mRNA expression is increased upon Cdk2 deletion. (A) qPCR analysis of spermatocyte cDNA isolated from corn oil treated (black triangles) and tamoxifen treated (white circles) Cdk2-iKO mice—referred to as noninduced and induced Cdk2-iKO, respectively. Expression is displayed as fold change normalized to the expression of the eEF2 housekeeping gene. STA-PUT spermatocyte isolation (as described in Materials and methods) for this experiment was performed at P68, 12 d after corn oil/tamoxifen treatment (started at P56). 17/25 NRF1 target genes showed significant increases in expression. Three biological replicates were tested for each group for each gene. Gene expression of each biological replicate was assumed to be normally distributed, but this was not formally tested. Error bars are representative of the SD of normalized fold change values from at least three biological replicates as compared to noninduced Cdk2-iKO controls. Significance was determined via unpaired t test (****, P ≤ 0.0001; ***, P < 0.001; **, P < 0.005; *, P < 0.05). (B) Western blotting of NRF1, EHMT1, and H3K9me2 in corn oil–treated Cdk2-iKO (lanes 1–3) or tamoxifen-treated Cdk2-iKO (lanes 4–6) whole testis lysates. For Cdk2-iKO animals, testis isolation for this experiment was performed at P68, 12 d after corn oil/tamoxifen treatment (started at P56). γ-Tubulin and pan-histone H3 (PAN H3) are shown as loading controls.
Figure 7.
Figure 7.
Pharmacological inhibition of EHMT1 partially rescues the phenotype of early Cdk2KO spermatocytes. (A) Time course experiment to determine the effectiveness of UNC0642 in reducing EHMT1 activity against H3K9 in Cdk2-iKO mice. Cdk2-iKO mice at P56 were injected with corn oil (lane 1) or tamoxifen (lanes 2–7), and 1 wk later were injected with UNC0642 (lanes 2–7). Whole testis lysates extracted from mice culled 1–6 d after the initial UNC0642 injection were blotted using the indicated antibodies. H3K9me2 levels were reduced by UNC0642 treatment for up to 3 d as compared with control. After this time, H3K9me2 levels were seen to rise again. Cdk2 deletion was confirmed by the loss of CDK2 protein in all mice receiving tamoxifen (lanes 1–7). Upon Cdk2 deletion, the expression of EHMT1 rose steadily over time in agreement that Cdk2 is a potential negative regulator of this protein acting through NRF1. GAPDH was used as a loading control. (B) Apoptotic counts in induced (+ tamoxifen) Cdk2-iKO mice treated with UNC0642 (blue bars) or control solution (CNTRL; red bars). Apoptotic counts are measured as mean apoptotic (TUNEL-positive) cells counted per tubule. Tissues collected here follow the schedule shown in Fig. S3 A with the addition of a single microinjection of UNC0642 or vehicle control (CNTRL) solution 7 d after the first treatment of tamoxifen or corn oil. Tissue collection was then performed 8, 10, 12, 16 or 30 d after either tamoxifen or corn oil treatment. At least 50 images were counted for each biological replicate. At least three biological replicates were taken for each time point and for each condition. Values shown are presented as the mean of at least three biological replicates ± SD. (C) Representative TUNEL-stained images from uninduced (+ Corn oil) or induced (+ Tamoxifen) Cdk2-iKO mice microinjected with either UNC0642 or vehicle control solution 16 d after either corn oil or tamoxifen treatment. Hematoxylin is used as a nuclear counterstain. Scale bars, 62.7 µm. (D) Staging of seminiferous tubules showing apoptotic counts, in induced Cdk2-IKO mice treated with UNC0642 (blue bars) or without UNC0642 (red bars) 16 d after tamoxifen treatment. Tubules were classified as stages I–V, VI–VIII, or IX–XII based on the morphology of cells. Apoptotic counts are measured as mean apoptotic (TUNEL-positive) cells counted per tubule. At least 20 images were counted for each stage, for each biological replicate. Values shown are presented as mean of at least three biological replicates ± SD. For B and D, mean apoptotic cell numbers/tubule were found to have a nonnormal distribution. Significance when comparing UNC0642-treated (blue) and control-treated (red) conditions in these panels was calculated by Kruskal–Wallis one-way ANOVA (**, P ≤ 0.01; ****, P < 0.0001). (E) Proposed mechanism of NRF1-CDK2 interaction and how this pathway could potentially be altered in the absence of Cdk2 to increase the expression levels of many NRF1 target genes. (1) In wild-type germ cells, chromatin-associated NRF1 is bound and phosphorylated by CDK2; (2) upon phosphorylation, NRF1 DNA-binding activity is reduced and NRF1 dissociates from chromatin. Phosphorylation here is indicated by “P” in yellow circles. (3) Upon dissociation, the transcription of NRF1 target genes such as Msh4, Asz1, and Ehmt1 is decreased; (4) reduced Ehmt1 expression results in a concurrent reduction in the levels of H3K9me2 on chromatin, resulting in widespread demethylation. Methylation here is indicated by “CH3” in red circles. (5) In the absence of Cdk2 expression, NRF1 is able to bind target promoters and transcription is activated; (6) NRF1 transcriptional activity remains high, and Ehmt1 is inappropriately transcribed; and (7) the continued expression of EHMT1 allows H3K9me2 levels to remain high. This occurs in parallel with the observation of meiotic arrest.

Similar articles

See all similar articles

Cited by 1 article

References

    1. Ahmed E.A., and de Rooij D.G. 2009. Staging of mouse seminiferous tubule cross-sections. Methods Mol. Biol. 558:263–277. 10.1007/978-1-60761-103-5_16 - DOI - PubMed
    1. Bailey T.L., and Elkan C. 1994. Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol. Biol. 2:28–36. - PubMed
    1. Bean C.J., Schaner C.E., and Kelly W.G. 2004. Meiotic pairing and imprinted X chromatin assembly in Caenorhabditis elegans. Nat. Genet. 36:100–105. 10.1038/ng1283 - DOI - PMC - PubMed
    1. Bellvé A.R., Cavicchia J.C., Millette C.F., O’Brien D.A., Bhatnagar Y.M., and Dym M. 1977. Spermatogenic cells of the prepuberal mouse. Isolation and morphological characterization. J. Cell Biol. 74:68–85. 10.1083/jcb.74.1.68 - DOI - PMC - PubMed
    1. Bergerat A., de Massy B., Gadelle D., Varoutas P.C., Nicolas A., and Forterre P. 1997. An atypical topoisomerase II from Archaea with implications for meiotic recombination. Nature. 386:414–417. 10.1038/386414a0 - DOI - PubMed

Associated data

Feedback