ABSTRACT
DOT1L is essential for early hematopoiesis but the precise mechanisms remain largely unclear. The only known function of DOT1L is histone H3 lysine 79 (H3K79) methylation. We generated two mouse models; a Dot1L-knockout (Dot1L-KO), and another possessing a point mutation in its methyltransferase domain (Dot1L-MM) to determine the role of its catalytic activity during early hematopoiesis. We observed that Dot1L-KO embryos suffered from severe anemia, while Dot1L-MM embryos showed minimal to no anemia. However, ex vivo culture of Dot1L-MM hematopoietic progenitors (HPCs) exhibited defective development of myeloid and mixed progenitors. DOT1L is a well-recognized, cell-type specific epigenetic regulator of gene expression. To elucidate the mechanisms underlying such diverse hematopoietic properties of Dot1L-KO and Dot1L-MM HPCs, we examined their whole transcriptomes. Extensively self-renewing erythroblast (ESRE) cultures were established using yolk sac (YS) cells collected on embryonic day 10.5 (E10.5). Dot1l-KO and Dot1l-MM cells expanded significantly less than the wildtype cells and showed slower progression through the cell cycle. Total RNA extracted from the wildtype and Dot1l-mutant ESRE cells were subjected to RNA-seq analyses. We observed that the majority (~82%) of the differentially expressed genes (DEGs) were upregulated in both of the Dot1L-mutants, which suggests that DOT1L predominantly acts as a transcriptional repressor in HPCs. We also observed that about ~40% of the DEGs were unique to either of the mutant group, suggesting that DOT1L possesses both methyltransferase domain-dependent and -independent functions. We further analyzed Gene Ontology and signaling pathways relevant to the DEGs common to both mutant groups and those that were unique to either group. Among the common DEGs, we observed upregulation of CDK inhibitors, which explains the cell cycle arrest in both of the Dot1L-mutant progenitors.
1. INTRODUCTION
DOT1L histone methyltransferase (DOT1L) is essential regulator of vital tissue and organ development during embryonic life, including hematopoiesis1. We observed that loss of DOT1L in mice (Dot1L-KO) results in lethal anemia during mid-gestation1. DOT1L is the only known methyltransferase in eukaryotic cells to methylate lysine 79 of histone H3 (H3K79)2. We generated a mouse line carrying a point mutation (Asn241Ala) in mouse Dot1L gene (Dot1L-MM) that renders its catalytic domain inactive3,4. Dot1L-MM mice contained an intact DOT1L protein that lacked only the H3K79 methyltransferase activity3. The methyltransferase mutant, Dot1L-MM mice, were also embryonic lethal-died around mid-gestation3. The mice also displayed defects in embryonic hematopoiesis, including a decreased ability to form definitive myeloid, and oligopotent (mixed) blood progenitors in ex vivo cultures3. However, unlike the Dot1L knockout (Dot1L-KO)1, HPCs from the Dot1L-MM YS were able to produce erythroid colonies in numbers similar to the wildtype3.
Histone methylation is important for permissive or repressive chromatin conformation and can have a profound effect on regulation of gene expression5. DOT1L is responsible for the mono, di- and tri-methyl marks on lysine 79 of histone H3 (H3K79)2. These histone modifications as well as the DOT1L protein have been strongly associated with actively transcribed chromatin regions6. Thus, it has been suggested that DOT1L is involved in epigenetic regulation of transcriptional activation of genes in a tissue specific manner.
In this study, we examined the expression of DOT1L-regulated genes on embryonic day 10.5 (E10.5) yolk sac (YS) derived hematopoietic progenitor cells (HPCs). We observed that more than 82% of the differentially expressed genes (DEGs) in Dot1L-KO or Dot1L-MM HPCs cultured ex vivo were upregulated, which suggests that DOT1L primarily acts as a transcriptional repressor in HPCs.
2. METHOLODOLOGY
2.1. Dot1L mutant mouse models
The Dot1L-KO mice were generated and maintained as described previously1. To produce the Dot1L-MM mouse, we generated mutant mESC as described3. Dot1L-KO and Dot1L-MM heterozygous mice were maintained by continuous backcrossing to 129 stocks. Genotyping was performed on tail clips by using RED extract-N-Amp Tissue PCR Kit, Sigma-Aldrich as previously described7,8. All animal experiments were performed in accordance with the protocols approved by the University of Kansas Medical Center Animal Care and Use Committee.
2.2. Extensively Self-Renewing Erythroblasts (ESRE) assays
Dot1L-KO or Dot1L-MM heterozygous mutant males and females were set up for timed mating to collect the conceptuses on E10.5. Pregnant females were sacrificed, and uteri were dissected to separate embryos and YS. Embryos were treated with RED extract-N-Amp Tissue PCR reagents (Millipore Sigma, Saint-Louis, MO) to purify genomic DNA and perform the genotyping PCR7,8. Digested E10.5 YSs were washed in IMDM alone, resuspended in 0.5ml expansion media for ESRE according to England et al.,9, and plated into gelatin-coated 24 well plates. Expansion media consisted of StemPro34 supplemented with nutrient supplement (Gibco/BRL), 2 U/ml human recombinant EPO (University of Kansas Hospital Pharmacy), 100 ng/ml SCF (PeproTech), 10−6 M dexamethasone (Sigma), 40 ng/ml insulin-like growth factor-1 (PeproTech) and penicillin-streptomycin (Invitrogen). After 1 day of culture, the nonadherent cells were aspirated, spun down, resuspended in fresh ESRE media, and transferred to a new gelatin coated well. After 3 days in culture, RNA was extracted from wildtype, Dot1L-KO or Dot1L-MM HPCs.
2.3. Assessment of cell proliferation, cell cycle analyses and apoptosis assays
Single-cell suspensions from E10.5 YS were cultured in MethoCult™ GF M3434 (StemCell Technologies, Vancouver, BC, Canada) for 4 days. The mix of cytokines in this methylcellulose medium promotes definitive erythroid, myeloid, and mixed progenitor differentiation. Cells were collected on day 4 and stained with Annexin V to assess apoptosis. Some cells were fixed by adding cold 70% ethanol slowly to single cell suspensions, and then stained with propidium iodide10. Flow cytometry was performed by the use of a FACSCalibur (BD Biosciences, San Jose, CA)11. Analyses of the cytometric data were carried out using CellQuest Pro software (BD Biosciences)12–14.
2.4. Sample collection, library preparation and RNA-sequencing
RNA quality was assessed by a Bioanalyzer at the KUMC Genomics Core, and samples with RIN values over 9 were selected for RNA-sequencing library preparation. RNA samples were extracted from multiple, expanded YS cells obtained from embryos of the same genotype. These samples were pooled to prepare each RNA-seq library. Approximately 500 ng of total RNA was used to prepare a RNA-seq library using the True-Seq mRNA kit (Illumina, San Diego, CA) as described previously15–17. The quality of RNA-seq libraries was evaluated by Agilent Analysis at the KUMC Genomics Core and the sequencing was performed on an Illumina NovaSeq 6000 sequencer (KUMC Genomics Core).
2.5. RNA-seq data analyses
RNA-sequencing data were demultiplexed, trimmed, aligned, and analyzed using CLC Genomics Workbench 12.2 (Qiagen Bioinformatics, Germantown, MD) as described previously15–17. Through trimming, low-quality reads were removed, and good-quality reads were aligned with the Mus musculus genome (mm10) using default guidelines: (a) maximum number of allowable mismatches = 2, (b) minimum length and similarity fraction = 0.8, and (c) minimum number of hits per read = 10. Gene expression values were measured in transcripts per million (TPM). DEGs were identified that had an absolute fold change of TPM ≥ 2 and a false discovery rate (FDR) p-value of ≤0.05.
2.6. Gene Ontology (GO) and disease pathway analyses for the RNA-sequencing data
DEGs were subjected to Gene Ontology (GO) analysis (http://www.pantherdb.org) and categorized in biological, cellular and molecular function. DEGs in Dot1L-mutant HPCs were further analyzed by Ingenuity Pathway Analysis (IPA; Qiagen Bioinformatics, Germantown, MD) to build gene networks related to placental development. Functional analyses were performed towards understanding the biological pathways and functions altered in either of the Dot1L-mutant progenitor cells.
2.7. Validation of RNA-sequencing data
DEGs were validated by RT-qPCR. RT-qPCR validation included cDNA samples prepared with wildtype, Dot1L-MM and Dot1L-KO ESRE cell derived total RNAs. The genes were selected from the IPA analyses and MGI data that impacted the proliferation and differentiation of HPCs.
2.8. Statistical analysis
Each RNA-seq library or cDNA was prepared from pooled RNA samples extracted from at least 3 different ESRE cultures of the same genotype. Each group for RNA sequencing consisted of three independent libraries and the DEGs were identified by CLC Genomics workbench as described previously15–17. RT-qPCR validation included at least six cDNA samples prepared from wildtype, Dot1L-MM and Dot1L-KO ESRE cell total RNA. The experimental results are expressed as mean ± standard error (SE). The RT-qPCR results were analyzed by one-way ANOVA, and the significance of mean differences was determined by Duncan’s post hoc test, with p ≤ 0.05. All the statistical calculations were done using SPSS 22 (IBM, Armonk, NY).
3. RESULTS
3.1. Dot1L-KO and Dot1L-MM embryos exhibit distinct hematopoietic phenotypes
We observed that Dot1L-KO embryos develop slower than WT embryos and suffer from lethal anemia 1 (Fig.1A-E). Dot1L-KO embryos die between embryonic day 11.5 (E11.5) and E13.5. Ex vivo culture of HPCs from E10.5 Dot1L-KO YS showed that erythroid differentiation was severely affected compared to myeloid lineage 1. We generated another mouse model that carries a point mutation (Asn241Ala) in endogenous DOT1L, rendering the catalytic domain inactive18 (Fig.1F-J). Although the Dot1L-methyl mutant (Dot1L-MM) embryos also died at midgestation, we observed remarkable differences in the hematopoietic phenotype between the Dot1L-KO and Dot1L-MM mice18 (Fig.1B-E and G-J); in particular, erythropoiesis was minimally affected in Dot1L-MM YS and embryo, suggesting that hematopoietic activity of DOT1L may not be limited to its MT domain. However, ex vivo culture of YS cells exhibited that formation of myeloid and mixed colonies were dramatically reduced in either Dot1L-KO or Dot1L-MM18. Culture of Dot1L-KO HPCs showed decreased cell proliferation (Fig. 2A), accumulation of cells in G0/G1 stage (Fig. 2B), and a greater percentage of Dot1L-KO or Dot1L-MM HPCs in ESRE culture were Annexin V-positive 1 (Fig. 2C). In addition, Alkaline Comet assays showed a greater DNA damage in Dot1L-MM compared to WT cells and the DNA damage was still higher in Dot1L-KO ESREs compared to MM (Fig. 2 D).
3.2. DEGs in Dot1L-KO or Dot1L-MM ESRE cells
Transcriptome data-sets were generated by sequencing of mRNA purified from ESRE cultures using E10.5 wildtype, Dot1L-KO or Dot1L-MM YS cells. The raw data have been deposited to NCBI SRA under PRJNA666736. Analyzed data include the DEGs are shown in Fig. 3(A-F). Of the total 25,749 reference genes in the mm10 genome, 16,806 genes were detected in wildtype and 16,678 in Dot1L-KO and 17,053 in Dot1L-MM ESRE cells. Analyses of the detected genes for level of gene expression revealed that ~40% had a very low abundance (<1 TPM), ~20% had low abundance (1-5 TPM), ~12% had modertate abundance (>5-10 TPM), ~25% had high abundance (>10-100 TPM), and only ~3% of the genes had a very high abundance (> 100 TPM). Among these genes, 2238 were differentially expressed (absolute fold change ≥2, p-value ≤ 0.05) in Dot1L-KO, with 358 downregulated and 1698 upregulated. In contrast, 2752 genes were differentially expressed (absolute fold change ≥2, p-value ≤ 0.05) in Dot1L-MM cells, with 328 downregulated and 1649 upregulated. The DEGs were evident in the hierarchical clustering (Fig. 3A, B) and Volcano plots (Fig. 3C, D), which demonstrate that most of the DEGs were upregulated in Dot1L-KO or Dot1L-MM HPCs (Fig. 3E, F).
3.4. Gene Ontology (GO) analyses of the DEGs
GO analysis classified the DEGs into three categories: Biological process (Fig. 4A, B), Molecular function (Fig. 4C, D) and Cellular component (Fig. 4E, F). GO analysis revealed that the majority of the genes in the biological process group were involved in biological process, cellular processes, or cell signaling (Fig. 4A, B). The genes in molecular function were involved in binding, protein-protein interactions, catalytic activity, and molecular and transcriptional regulation (Fig. 4C, D). The genes in cellular component were predominantly involved in cell parts, membranes, organelles, and protein-protein complexes (Fig. 4E, F).
3.5. Ingenuity Pathway Analysis (IPA) of the DEGs
IPA of the DEGs in Dot1L-KO or Dot1L-MM HPCs in ESRE culture identified upregulation of genes related to regulation of hematopoiesis. Among the hematopoietic pathways involved, we were particularly interested in proliferation and differentiation of hematopoietic progenitor cells (Fig. 5A-D). The DEGs included upregulation of CDK inhibitors and downregulation of Flt1, Flt3, Hoxa9 and Mpl. Interestingly, among the 165 known genes involved in differentiation of HPCs, only 9 were affected in Dot1L-KO cells and 21 in Dot1L-MM ESRE cells (Fig. 5 E, F).
3.6 RT-qPCR analyses validated DEGs involved in HPC proliferation and differentiation
Differentially expressed genes that were identified to be involved in proliferation and differentiation of HPCs were validated by RT-qPCR analyses. We observed that while genes involved in proliferation of HPCs were significantly downregulated in both Dot1L-KO and Dot1L-MM ESRE cells (Fig. 6 A-F), those involved in induction of differentiation were markedly upregulated (Fig. 6 G-I). Marked increase in CDK inhibitors (H, I) can explain the accumulation of cells in G0/G1 phase and increased proportion of the Dot1L-mutant ESRE cells undergoing apoptosis.
4. DISCUSSION
DOT1L is expressed at high levels in mouse HPCs (Supplemental Fig.1), suggesting a potential role for this chromatin organizer and transcriptional regulator in early blood development. This study analyzed the DEGs in both Dot1l-KO and Dot1L-MM mouse HPCs derived from E10.5 YS to examine early blood development. RNA-seq datasets were used to identify DOT1L-regulated genes in mouse HPCs and understand their potential role in early blood development.
We previously reported that hematopoietic transcription factor (TF) Gata2 was significantly reduced in Dot1L-KO HPCs, whereas Pu.1, an erythropoiesis inhibiting TF was upregulated1. We also observed that KIT-positive HPCs from Dot1L-KO YS expressed low levels of Trpc619. Our RNA-seq data recapitulated the previous observations regarding expression of Gata2, Pu.1 and Trpc6 (PRJNA666736).
DOTL1L is responsible for methylation of H3K792. Histone methylation is integral to permissive or repressive chromatin conformation, and regulation of gene expression20. Expression of Dot1/Dot1L is conserved across species2,6,21–23 and enrichment of H3K79 methyl marks is associated with actively transcribed chromatin regions2,6,21–23. However, DOT1L has also been associated with repression of gene transcription24. Thus far, the precise molecular mechanisms of DOT1L regulation of gene expression in HPCs remain undetermined.
We have detected a large number of DEGs in both Dot1L-KO or Dot1L-MM ESRE cells. Although ~60% of DEGs were common to both Dot1L-mutant groups, the remaining ~40% DEGs were unique to either mutant group (Fig. 3), which suggest that DOT1L regulates HPCs genes via methyltransferase-dependent and -independent ways. Remarkably, >82% of DEGs in either Dot1L-KO or Dot1L-MM ESRE cells were found to be upregulated, which indicate that DOT1L primarily acts as a transcriptional repressor in hematopoietic progenitors cells.
Gene ontology and IPA analyses indicated that DOT1L regulates genes, which are responsible for cell signaling and protein-protein interaction. Our previous studies have demonstrated that either loss of DOT1L expression1,19 or loss of its methyltransferase activity3 leads to G0/G1 cell cycle arrest and increased apoptosis. IPA analyses showed that the DEGs are linked proliferation and differentiation of HPCs, which were further validated by RT-qPCR analyses. We observed that while genes involved in proliferation of HPCs were significantly downregulated, those involved in induction of differentiation were markedly upregulated (Fig. 6). Marked increase in CDK inhibitors positively correlates with the accumulation of cells in G0/G1 phase and increased proportion of the Dot1L-mutant ESRE cells undergoing apoptosis. The gene expression profile also shows a positive correlation with mechanisms involved in incraesed DNA damage.
The most intriguing result of these RNA-seq analyses is the upregulation of >82% genes in Dot1L-KO and the Dot1L-MM ESRE cells, which proves that DOT1L acts as a transcriptional repressor in HPCs. The other striking result of these analyses is the apparent differences between the Dot1L-KO and the Dot1L-MM transcriptome profile, suggesting that DOT1L can act in both methyltransferase-dependent and -independent ways.
Disclosure
The authors do not have any conflicts of interest.
Acknowledgements
This work was supported by the National Institutes of Health (grant R01DK091277). The mouse model was generated in the Transgenic and Gene Targeting Institutional Facility of the University of Kansas Medical Center, supported in part by the Center of Biomedical Research Excellence (COBRE) Program Project in Molecular Regulation of Cell Development and Differentiation (NIH P30 GM122731) and the University of Kansas Cancer Center (NIH P30 CA168524).