Abstract

For genetically heterogeneous diseases a better understanding of how the underlying gene defects are functionally interconnected will be important for dissecting disease etiology. The Immunodeficiency, Centromeric instability, Facial anomalies (ICF) syndrome is a chromatin disorder characterized by mutations in DNMT3B, ZBTB24, CDCA7 or HELLS. Here, we generated a Zbtb24 BTB domain deletion mouse and found that loss of functional Zbtb24 leads to early embryonic lethality. Transcriptome analysis identified Cdca7 as the top down-regulated gene in Zbtb24 homozygous mutant mESCs, which can be restored by ectopic ZBTB24 expression. We further demonstrate enrichment of ZBTB24 at the CDCA7 promoter suggesting that ZBTB24 can function as a transcription factor directly controlling Cdca7 expression. Finally, we show that this regulation is conserved between species and that CDCA7 levels are reduced in patients carrying ZBTB24 nonsense mutations. Together, our findings demonstrate convergence of the two ICF genes ZBTB24 and CDCA7 at the level of transcription.

Introduction

The Immunodeficiency, Centromeric instability and Facial anomalies (ICF; OMIM 242860/614069) syndrome is an autosomal recessive disorder characterized by immunodeficiency, developmental delay and facial anomalies. Patients present with hypo- or a- gammaglobulinemia, in the presence of circulating B cells, causing recurrent and often fatal respiratory and gastrointestinal infections. All patients show DNA hypomethylation of the centromeric satellite II and III repeats, which together with chromosomal abnormalities in phytohemagglutinin-stimulated cells are the molecular hallmarks for the ICF syndrome (1,2). Based on the underlying genetic defects, ICF syndrome can be divided into at least five different clinically indistinguishable subgroups (ICF1, ICF2, ICF3, ICF4 and ICFX) (3). About half of the ICF patients carry mutations in the de novo DNA methyltransferase 3B gene (DNMT3B), which is referred to as ICF1 (1,4). ICF2 patients carry mutations in the zinc-finger- and BTB domain containing 24 gene ZBTB24 and account for ∼30% of the ICF cases (1,5,6). Recently, mutations in cell division cycle associated 7, CDCA7 and helicase, lymphoid-specific, HELLS/LSH, have been shown to be causative for the ICF3 and ICF4 subtypes respectively, and together with few unexplained cases (ICFX) account for the remaining ICF cases (3).

Explaining the molecular mechanisms underlying ICF syndrome has remained difficult. The majority of molecular studies have focused on the analysis of gene expression and DNA methylation patterns in cell lines of normal and ICF1 patients. DNA hypomethylation and aberrant expression of germ line genes and genes important for immune function have been reported (7–11). ICF1 mouse models carrying hypomorphic Dnmt3b mutations have been described and shown to recapitulate some of the disease aspects such as hypomethylation of repeat sequences, aberrant expression of germ line genes and craniofacial abnormalities (12–14). The most prominent difference between mouse and human is the absence of B cell defects in an ICF1 mouse model. In contrast, apoptosis of T-cells was reported after birth (12). Hells loss of function mice display genome wide DNA hypomethylation and perinatal lethality (15,16). In addition, B and T cell abnormalities have been reported in Rag2-/- mice that were injected with Hells-/- fetal liver cells (17). In mouse embryonic stem cells (mESCs), Hells has been proposed to facilitate Dnmt3b recruitment to its targets (18). This highlights a functional convergence between Dnmt3b and Hells and provides a possible explanation for the involvement of both genes in the ICF syndrome.

Little is known about the molecular function of Zbtb24 and Cdca7. Zbtb24 is a member of the BTB domain family of proteins, of which many are important for hematopoietic differentiation (19). Zbtb24 contains eight C2H2 zinc fingers and a function as a transcription factor has been suggested (20). Cdca7 has previously been shown to interact with Myc and is involved in neoplastic transformation and hematopoietic stem cell emergence (21,22). Mouse Zbtb24 can localize to heterochromatin (5) and it has been shown that knock down of Zbtb24 and Cdca7 in mouse embryonic fibroblasts leads to hypomethylation of minor satellite repeats, suggesting a role for the two factors in the maintenance of DNA methylation (3).

In this study, we show that mice homozygous for a Zbtb24 BTB-domain deletion are early embryonic lethal and that Cdca7 expression is lost in ES cells from these mice. We find that this regulation is conserved between species and that overexpression of human ZBTB24 can restore Cdca7 mRNA levels in homozygous mutant mESCs. We further demonstrate enrichment of ZBTB24 at the CDCA7 promoter suggesting a direct transcriptional regulation. Finally, we show that CDCA7 levels are affected in ICF2 patients carrying ZBTB24 nonsense mutations.

Results

Generation of Zbtb24 BTB-domain deletion mouse and derivation of mESC lines

To study the molecular function of Zbtb24, the Zbtb24 gene was mutated by partial replacement of the BTB-domain region with a PGK-Neomycin (PGK-NEO) cassette, in antisense orientation, using standard gene targeting techniques (Fig. 1A and

). This results in the deletion of the last 15 bps of the non-coding exon 1 and the first 430 bps of exon 2, which encodes the BTB and AT-hook domains, and the first zinc finger motif. Southern blot and long range PCR were used to show integration of the PGK- NEO cassette and no random insertion was found (). Based on previous biochemical studies on BTB-domain containing proteins, this deletion is predicted to disrupt protein-protein interactions and regulation of gene expression through altered chromatin conformation (23). Mice heterozygous for the Zbtb24 BTB-domain deletion, Zbtb24ΔBTB, appeared to be normal, fertile and indistinguishable from their wild-type littermates at weaning (∼3 weeks of age) and throughout adulthood. Following heterozygous intercrosses, no viable homozygous Zbtb24ΔBTB mutants were recovered from a total of 129 offspring. The proportions of wild-type (n= 45) and heterozygous (n= 84) offspring were as expected for homozygous lethality (Fig. 1B). Timed matings between heterozygous Zbtb24ΔBTB mice produced no viable homozygous Zbtb24ΔBTB embryos at E9.5 (Fig. 1B). The presence of implantation sites at E9.5 suggests that embryonic death of homozygotes occurs sometime between E4.5–9.5. To date, no knock out mouse model for Zbtb24 has been described but recently, mousephenotype.org reported complete homozygous lethality pre-weaning for the Zbtb24tm1b(EUCOMM)Hmgu deletion allele (https://www.mousephenotype.org/, date last accessed July 26, 2016). This is in agreement with our observations.
Figure 1.

Characterization of homozygous Zbtb24ΔBTB mutant mESCs. (A) Genotyping of Zbtb24ΔBTB heterozygous mice. The wild-type allele was detected as a 363-bp band (P2 + P1) and the mutant allele as a 321-bp band (P3 + P1). Black arrows indicate primers. (B) Timed matings and intercrosses of Zbtb24ΔBTB mice. Data shows the number of mice observed (and in brackets the percentage) at E9.5–11.5 and 3 weeks. (C) Cellular morphology of wild-type and homozygous Zbtb24ΔBTBmutant mESCs grown feeder free on 2i. Scale bar, 250 μm. (D) qRT-PCR analysis of pluripotent gene expression levels in wild-type and homozygous Zbtb24ΔBTB mESCs. Error bars, SEM from three biological replicates per genotype; NS, non-significant.

Figure 1.

Characterization of homozygous Zbtb24ΔBTB mutant mESCs. (A) Genotyping of Zbtb24ΔBTB heterozygous mice. The wild-type allele was detected as a 363-bp band (P2 + P1) and the mutant allele as a 321-bp band (P3 + P1). Black arrows indicate primers. (B) Timed matings and intercrosses of Zbtb24ΔBTB mice. Data shows the number of mice observed (and in brackets the percentage) at E9.5–11.5 and 3 weeks. (C) Cellular morphology of wild-type and homozygous Zbtb24ΔBTBmutant mESCs grown feeder free on 2i. Scale bar, 250 μm. (D) qRT-PCR analysis of pluripotent gene expression levels in wild-type and homozygous Zbtb24ΔBTB mESCs. Error bars, SEM from three biological replicates per genotype; NS, non-significant.

The early embryonic lethality prompted us to establish Zbtb24ΔBTB mutant mESC lines. In total, 24 blastocysts from two heterozygous intercrosses were put into culture, which resulted in the establishment of 19 independent mESC lines. Genotyping identified six wild-type, eight heterozygous and five homozygous Zbtb24ΔBTB mutant mESC lines. We decided to continue with three wild-type and three Zbtb24ΔBTB homozygous mutant mESC lines and maintained the mESCs in the presence of the two small-molecule inhibitors PD0325901 and CHIR99021, together known as 2i. mESCs grown in 2i resemble cells in the pre-implantation embryo (24) and have reduced DNA methylation levels genome wide (25). Wild-type and homozygous Zbtb24ΔBTB mutant mESCs were phenotypically normal, showed a homogeneous morphology characteristic for 2i and were positive for the expression of pluripotency marker genes (Fig. 1C and D).

Transcriptome-wide differences between wild-type and homozygous Zbtb24ΔBTB mutant mESCs

To characterize the effect of the Zbtb24 BTB-domain deletion mutation on the mESC transcriptome we performed RNA sequencing (RNA-seq) on total RNA isolated from independent wild-type and Zbtb24ΔBTB homozygous mutant mESC lines, established and cultured feeder free in 2i (n= 3 per genotype). Genome-wide we found a total of 139 significantly differentially expressed genes, using a cutoff that defined significant changes as at least 0.5 log 2-fold difference in gene expression and an adjusted P-value of <0.05 (Fig. 2A and

). A similar number of up- and down-regulated genes were observed relative to wild-type controls. Functional annotation clustering of differentially expressed genes using KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis did not show any enrichment for specific terms and chromosome graph showed homogeneous distribution of differentially expressed genes over the mouse genome (). Analysis of gene expression levels for known lineage specification genes showed that there was a good correlation with a published RNA-seq dataset from mESCs grown on 2i (26) and did not reveal any major differences between wild-type and homozygous Zbtb24ΔBTB mutant mESCs ().
Figure 2.

Transcriptome analysis of wild-type and homozygous Zbtb24ΔBTB mutant mESCs. (A) MA plot showing RNA-seq analysis results from Zbtb24 wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Three independent biological replicates were analyzed per genotype. Significantly differentially expressed genes are shown in red, adj.P <0.05. Black data points represent genes whose expression were not significantly altered in Zbtb24ΔBTBmutant mESCs. (B) IGV RNA-seq track showing a putative translation start site at the transgenic Zbtb24 locus in a homozygous Zbtb24ΔBTB mutant mESC line. Bidirectional transcription of the PGK promoter was detected. An ATG is present near the TSS region of the mutant Zbtb24 transcript that does not affect the reading frame of the protein suggesting that a truncated Zbtb24 protein can be produced. The green and blue lines represent a nucleotide mismatch in sequencing reads. (C) Polysome profiling in wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Error bars = SEM, two biological replicates for each genotype. (D) Western blot showing full-length Zbtb24 (∼80 kDa) in wild-type and severely reduced levels in homozygous Zbtb24ΔBTB mutant mESCs. Tubulin is shown as a loading control.

Figure 2.

Transcriptome analysis of wild-type and homozygous Zbtb24ΔBTB mutant mESCs. (A) MA plot showing RNA-seq analysis results from Zbtb24 wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Three independent biological replicates were analyzed per genotype. Significantly differentially expressed genes are shown in red, adj.P <0.05. Black data points represent genes whose expression were not significantly altered in Zbtb24ΔBTBmutant mESCs. (B) IGV RNA-seq track showing a putative translation start site at the transgenic Zbtb24 locus in a homozygous Zbtb24ΔBTB mutant mESC line. Bidirectional transcription of the PGK promoter was detected. An ATG is present near the TSS region of the mutant Zbtb24 transcript that does not affect the reading frame of the protein suggesting that a truncated Zbtb24 protein can be produced. The green and blue lines represent a nucleotide mismatch in sequencing reads. (C) Polysome profiling in wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Error bars = SEM, two biological replicates for each genotype. (D) Western blot showing full-length Zbtb24 (∼80 kDa) in wild-type and severely reduced levels in homozygous Zbtb24ΔBTB mutant mESCs. Tubulin is shown as a loading control.

By disrupting exon 1 and 2 of the Zbtb24 gene we intended to create a knock out mouse model. Unexpectedly, our RNA-seq analysis revealed that Zbtb24 mRNA levels were not affected in Zbtb24ΔBTB homozygous mutant mESCs (Fig. 2B and

) suggesting that an aberrant Zbtb24ΔBTB transcript can be produced in the mutants. To enable characterization of the transgenic Zbtb24 locus we mapped to an artificial mouse genome reference containing the sequence of the Zbtb24 locus with the inserted reverse PGK-NEO cassette, resembling our transgene insertion. This analysis suggested that an aberrant Zbtb24 transcript initiated in the PGK promoter flanking the NEO cassette (). No aberrant transcription from genes neighboring the Zbtb24 locus was observed in the Zbtb24ΔBTB homozygous mutant mESCs (). We used polysome profiling to assess whether the aberrant Zbtb24ΔBTB transcript can be engaged by ribosomes. We found no difference in the ratio of total RNA/polysome-bound RNA between wild-type and Zbtb24ΔBTB homozygotes suggesting that Zbtb24ΔBTB mRNA could be translated into an amino-terminal truncated Zbtb24 protein that lacks the BTB and AT hook domains (Fig. 2C). Finally, we used a C-terminal Zbtb24 antibody and Western-blot analysis to determine whether Zbtb24 protein could be detected in Zbtb24ΔBTBhomozygous mutant mESCs. Zbtb24 was detected at its predicted size of around 80 kDa in wild-type but severely reduced in the Zbtb24ΔBTBhomozygous mutant mESCs (Fig. 2D). Furthermore, we could not detect an additional band in the mutant mESCs that could be indicative of an amino-terminal truncated Zbtb24 protein. Together, our results suggest that the amino-terminal truncated Zbtb24 protein lacking the BTB and AT hook domains is likely to be unstable.

The top significantly down-regulated gene (2.57 log2 fold change and adjusted P-value 4,58E-131) in our RNA-seq dataset was Cdca7 (Fig. 3A), in which mutations have recently been shown to be causative for ICF3 (3). Western-blot analysis revealed that the decrease in mRNA levels correlates with reduced levels of Cdca7 protein in Zbtb24ΔBTB homozygous mutant mESCs (Fig. 3B). The finding that Cdca7 expression levels are down regulated in Zbtb24ΔBTB homozygous mutant mESCs was unexpected and to our knowledge, no functional relationship between these two genes has previously been reported.

Figure 3.

Cdca7 is down regulated in homozygous Zbtb24ΔBTB mutant mESCs and upon siRNA knock down. (A) RNA-seq normalized read count showing Cdca7 down regulation in homozygous Zbtb24ΔBTB mutant mESCs. Each black dot represents a biological replicate. P-value is corrected for multiple testing. (B) Western blot showing that altered Cdca7 transcript levels correlate with a decrease in Cdca7 protein levels (43 kDa) in homozygous Zbtb24ΔBTB mutant mESCs. Two independent biological replicates for each genotype are shown and Tubulin (∼50 kDa) was used as a loading control. (C) qRT-PCR analysis of gene expression levels of the four ICF genes Dnmt3b, Zbtb24, Cdca7 and Hells in wild-type mESCs upon siRNA knock down. Error bars represent the SEM from two independent experiments. t-test ***P <0.001. (D) Western blot showing Dnmt3b (top, ∼110 kDa, multiple variants), Zbtb24 (middle, ∼80 kDa), Cdca7 (middle) and Hells (bottom, ∼100 kDa) protein levels after siRNA knock down. Tubulin is shown as a loading control.

Figure 3.

Cdca7 is down regulated in homozygous Zbtb24ΔBTB mutant mESCs and upon siRNA knock down. (A) RNA-seq normalized read count showing Cdca7 down regulation in homozygous Zbtb24ΔBTB mutant mESCs. Each black dot represents a biological replicate. P-value is corrected for multiple testing. (B) Western blot showing that altered Cdca7 transcript levels correlate with a decrease in Cdca7 protein levels (43 kDa) in homozygous Zbtb24ΔBTB mutant mESCs. Two independent biological replicates for each genotype are shown and Tubulin (∼50 kDa) was used as a loading control. (C) qRT-PCR analysis of gene expression levels of the four ICF genes Dnmt3b, Zbtb24, Cdca7 and Hells in wild-type mESCs upon siRNA knock down. Error bars represent the SEM from two independent experiments. t-test ***P <0.001. (D) Western blot showing Dnmt3b (top, ∼110 kDa, multiple variants), Zbtb24 (middle, ∼80 kDa), Cdca7 (middle) and Hells (bottom, ∼100 kDa) protein levels after siRNA knock down. Tubulin is shown as a loading control.

Our observation of Cdca7 dysregulation in Zbtb24ΔBTB homozygous mutant mESCs prompted us to examine whether additional relationships between the four ICF genes exist. We conducted knock down experiments on wild-type mESCs, using two independent siRNAs against each of the four ICF genes, Dnmt3b, Zbtb24, Cdca7 and Hells. Consistent with the results from our Zbtb24ΔBTB mutant mESCs, we found significantly reduced Cdca7 mRNA levels (t-test P < 0.05) upon Zbtb24 depletion (Fig. 3C and D). Reversely, Cdca7 knock down did not affect Zbtb24 mRNA levels, suggesting that Zbtb24 acts upstream of Cdca7. No additional links between the four ICF genes at transcript levels were observed (Fig. 3C and D). Together our results show that Zbtb24 can directly or indirectly regulate transcription of Cdca7 in mESCs, suggesting convergence of the ICF2 and ICF3 genes.

Consistent with our data from mESCs, we found down regulation of CDCA7 mRNA (t-test P < 0.05) and protein levels upon siRNA-mediated ZBTB24 knock down in human fetal lung fibroblasts (WI38) and a human osteosarcoma cell line (U2OS) (Fig. 4). When we knocked down CDCA7 using three siRNAs in U2OS cells, we did not detect down regulation of ZBTB24, which is consistent with our result in mESCs (

). These results suggest that the effect of ZBTB24 on CDCA7 transcription is conserved between species and seen in pluripotent mESCs as well as somatic cells.
Figure 4.

CDCA7 is down regulated in siZBTB24 human cell lines. (top) qRT-PCR and (bottom) a representative Western-blot analysis showing down regulation of both ZBTB24 and CDCA7 levels in siZBTB24 U2OS (left) and WI38 (right) cells. Error bars = SEM from three independent knock down experiments. t-test *P < 0.05, **P <0.01, ***P <0.001. TUBULIN (∼50 kDa) is shown as a loading control. CDCA7 (∼40 kDa).

Figure 4.

CDCA7 is down regulated in siZBTB24 human cell lines. (top) qRT-PCR and (bottom) a representative Western-blot analysis showing down regulation of both ZBTB24 and CDCA7 levels in siZBTB24 U2OS (left) and WI38 (right) cells. Error bars = SEM from three independent knock down experiments. t-test *P < 0.05, **P <0.01, ***P <0.001. TUBULIN (∼50 kDa) is shown as a loading control. CDCA7 (∼40 kDa).

Zbtb24 regulates Cdca7 expression level by activating the promoter region of Cdca7

ZBTB24 protein is highly conserved between mouse and human and contains a BTB domain, an AT hook domain and eight zinc fingers of the classical Cys2-His2 type (

). Based on its zinc finger structure it has been suggested to function as a transcription factor (20). To investigate whether ZBTB24 has transcriptional capacity we overexpressed full-length human GFP-ZBTB24 and 3xTy1-ZBTB24 in homozygous Zbtb24ΔBTB mutant mESCs. First, we measured expression levels of Cdca7, Cdc40 and Ostc. These three genes were most significantly down regulated in our RNA-seq experiment and therefore considered to be putative direct targets of Zbtb24 (). We found that overexpression of ZBTB24 could restore transcription levels of Cdca7, Cdc40 and Ostc (Fig. 5A).
Figure 5.

Zbtb24 regulates Cdca7 expression level by modulating the Cdca7 promoter. (A) qRT-PCR showing Cdca7, Cdc40 and Ostc expression levels in wild-type and Zbtb24ΔBTB homozygous mutant mESCs. mRNA levels of all three genes could be rescued by overexpression of GFP- or 3xTy1 tagged ZBTB24. Error bars = SEM from three biological replicates. t-test *P < 0.05, **P <0.01. (B) Luciferase activity assay showing the interaction between endogenous Zbtb24 and overexpressed full-length ZBTB24 with the Cdca7 and Ostc promoters in wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Error bars = SEM from three biological replicates. t-test *P < 0.05, **P < 0.01, ***P < 0.001. (C) (top) TSS regions of CDCA7 and OSTC. Primers (black arrows) for ChIP-qPCR were designed in the predicted promoter regions (∼1 kb upstream of TSS) of CDCA7 and OSTC. (bottom) ChIP-qPCR analysis showing an enrichment of ZBTB24 at the CDCA7 promoter region when compared to a R28S control region in U2OS cells transfected with a 3xTy1_ZBTB24 construct. Error bars = SEM from two independent experiments. t-test *P < 0.05, **P < 0.01.

Figure 5.

Zbtb24 regulates Cdca7 expression level by modulating the Cdca7 promoter. (A) qRT-PCR showing Cdca7, Cdc40 and Ostc expression levels in wild-type and Zbtb24ΔBTB homozygous mutant mESCs. mRNA levels of all three genes could be rescued by overexpression of GFP- or 3xTy1 tagged ZBTB24. Error bars = SEM from three biological replicates. t-test *P < 0.05, **P <0.01. (B) Luciferase activity assay showing the interaction between endogenous Zbtb24 and overexpressed full-length ZBTB24 with the Cdca7 and Ostc promoters in wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Error bars = SEM from three biological replicates. t-test *P < 0.05, **P < 0.01, ***P < 0.001. (C) (top) TSS regions of CDCA7 and OSTC. Primers (black arrows) for ChIP-qPCR were designed in the predicted promoter regions (∼1 kb upstream of TSS) of CDCA7 and OSTC. (bottom) ChIP-qPCR analysis showing an enrichment of ZBTB24 at the CDCA7 promoter region when compared to a R28S control region in U2OS cells transfected with a 3xTy1_ZBTB24 construct. Error bars = SEM from two independent experiments. t-test *P < 0.05, **P < 0.01.

Next, we cloned the predicted mouse Cdca7, Cdc40 and Ostc promoters into a Luciferase reporter vector and transfected wild-type and homozygous Zbtb24ΔBTB mutant mESCs. Luciferase activity was detected in wild-type but not homozygous Zbtb24ΔBTB mutant mESCs for Cdca7 and Ostc, whereas we could not detect luciferase activity for Cdc40 (Fig. 5B). The Cdc40 gene is located adjacent to the Wasf1 gene and it could be that Cdc40 expression is regulated by elements distal to the Cdc40 gene (

). We then co-transfected homozygous Zbtb24ΔBTB mutant mESCs with the reporter constructs and a human GFP-ZBTB24 overexpression construct. Luciferase activity was detected for Cdca7 and Ostc, demonstrating that the GFP-ZBTB24 overexpression construct was able to rescue the Zbtb24ΔBTB phenotype (Fig. 5B). These results suggest that functional Zbtb24 is needed at the Cdca7 and Ostc promoters to activate transcription.

To test whether ZBTB24 directly or indirectly affects CDCA7 expression, we overexpressed both 3xTy1 and 3xTy1_ZBTB24 in U2OS cells separately and performed ChIP followed by qRT-PCR. We first performed immunoprecipitation to test specificity of the Ty1 antibodies (

). For the ChIP-qPCR we decided to focus on CDCA7 and OSTC since we failed to detect Luciferase activity in the predicted promoter region of Cdc40. Using primers in the predicted promoter regions of CDCA7 and OSTC and 28S ribosomal RNA as a control region we found ZBTB24 enrichment at the CDCA7 and OSTC promoter regions when compared to 28S ribosomal RNA (Fig. 5C). Together, these experiments suggest that ZBTB24 modulates CDCA7 expression levels by directly binding to the CDCA7 promoter.

CDCA7 deregulation upon ZBTB24 loss is reflected in ICF2 patient-derived cell lines

Next, we were interested to determine if CDCA7 levels were also affected in ICF2 patients carrying ZBTB24 mutations. To address this, we used qRT-PCR to measure ZBTB24 and CDCA7 mRNA levels in two ICF2 families. One previously unpublished patient was found to be homozygous for a 1 bp deletion in exon 2 that encodes the Zinc finger domain of ZBTB24 (c.917delA; p.[N306IfsX4]) by Sanger candidate gene sequencing (

). This mutation results in a frameshift and a premature stop codon three amino acids downstream and has previously been described in a different family (6). We found reduced levels of CDCA7 mRNA in the newly identified ICF2 patient when compared to the heterozygous unaffected parents in fibroblasts (Fig. 6). A less apparent result was obtained using RNA isolated from expanded T-cells from a different patient (patient 55 in Weemaes et al.) (1) carrying a homozygous nonsense ZBTB24 mutation (c.958C > T; p.[R320X]). One explanation for the observed discrepancy between fibroblasts and T-cells is that the regulation is not conserved between all tissues or cell types. Alternatively, T-cell stimulation could mask the ZBTB24-dependent regulation of CDCA7 since in zebrafish, cdca7 was shown to participate in a regulatory network controlling thymus development (22).
Figure 6.

ZBTB24 and CDCA7 mRNA levels in two ICF2 families. qRT-PCR analysis of ZBTB24 and CDCA7 mRNA levels in ICF2 patient-derived fibroblasts (Rf1461) and T-cells (Rf1449). Both patients are homozygous for ZBTB24 nonsense mutations. Error bars = SEM from three technical replicates.

Figure 6.

ZBTB24 and CDCA7 mRNA levels in two ICF2 families. qRT-PCR analysis of ZBTB24 and CDCA7 mRNA levels in ICF2 patient-derived fibroblasts (Rf1461) and T-cells (Rf1449). Both patients are homozygous for ZBTB24 nonsense mutations. Error bars = SEM from three technical replicates.

Discussions

In this study we have revealed that the two ICF genes ZBTB24 and CDCA7 converge at the level of transcriptional regulation. Absence or disruption of ZBTB24 protein results in the down regulation of CDCA7, most likely because of a direct interaction of the ZBTB24 transcription factor with the CDCA7 promoter, and this is found in mESCs and in somatic cells.

Mice homozygous for a BTB-domain deletion are early embryonic lethal suggesting an important function for Zbtb24 in development. The early embryonic lethality phenotype of the Zbtb24ΔBTBhomozygotes was surprising given that ICF is a recessive disorder and most ICF2 patients carry nonsense mutations in the ZBTB24 gene. One possibility for the more severe phenotype in the mouse is that Zbtb24 has acquired a different function during mouse early development. Alternatively, using a different inbred mouse strain or a mixed background could result in improved viability of the homozygous embryos. A third possibility could be that the aberrant PGK/NEO-driven Zbtb24ΔBTBtranscript contributes to the early embryonic lethality observed in the homozygotes. However, the observation that an independent allele produced by the IPMC shows a similar homozygous lethality phenotype makes this less likely.

So far, no clear phenotypic differences between the ICF subtypes have been identified. However, it appears that ICF2 patients show a higher incidence of intellectual disability (1). At a molecular level, up-regulation of germ-line genes including Mael and Syce1 has been reported in ICF1 mouse models and in whole blood from ICF1 patients (10,13). In contrast, ICF2 patients did not show this molecular signature (10). Consistent with this, expression levels of germ line genes were not affected in our homozygous Zbtb24ΔBTB mutant mESCs. Our findings in mouse and human cell lines indicate that disruption of ZBTB24 leads to down regulation of CDCA7, the gene causative for ICF3. If and to what extent this contributes to the ICF syndrome phenotype requires further investigation. Highly overlapping phenotypic features suggest convergence of all ICF genes in pathways involved in immunity, chromatin regulation and development. HELLS can interact with DNMT3B to mediate DNA methylation establishment (17,18,27,28). It is possible that ZBTB24 and CDCA7 act in common at the level of DNA methylation maintenance.

In conclusion, we have identified ZBTB24 as a direct transcriptional activator of CDCA7 and established a functional link between the two ICF genes. Our observations suggest that investigations into functional relationships between disease causing genes are likely to shed light on the mechanisms underlying other genetically heterogeneous diseases.

Materials and Methods

Ethics statement

All procedures involving animals were approved by the Animal Ethics Committee of the Leiden University Medical Centre and by the Commission Biotechnology in animals of the Dutch Ministry of Agriculture. The study was approved by the Medical Ethical Committee of the Leiden University Medical Center and patient material was obtained after obtaining informed consent.

Zbtb24 BTB deletion mouse

Zbtb24ΔBTB mice were generated by homologous recombination using standard gene targeting techniques in a C57BL/6J background. Transgenic mice were genotyped by PCR and the wild-type allele was detected as a 363-bp band and the mutant allele as a 321-bp band (genotyping primers are provided in

). For timed matings, the day of a vaginal plug was designated as E0.5. All experimental procedures and crosses were carried out on mice that were backcrossed to C57BL/6J for at least 10 generations.

Patient Material

Patient derived skin fibroblasts were generated by expanding primary cultures from skin biopsies and cultured in DMEM F12 (31331) supplemented with 20% FCS, 1% Pen-Strep, 1% sodium pyruvate and 1% HEPES (all Invitrogen Life Technologies, Bleiswijk, The Netherlands). T-cells were obtained from fresh blood samples as described previously (1).

Cell culture and siRNA transfection

mESCs were generated from E3.5 blastocysts after natural mating and were cultured without feeders in mESC medium [Knockout DMEM (10829-018; Gibco), 10% FBS (DE14-801F; BioWhittaker), NEAA (11140; Gibco), l-Glutamine (25030-123; Gibco), Sodium Pyruvate (11360; Gibco), 2-Mercaptoethanol (31350; Gibco) and Leukemia Inhibitory Factor (ESG1107; Millipore)] plus MEK inhibitor PD0325901 (1 µm) and GSK3 inhibitor CHIR99021 (3 µm) (Axon Medchem) on 0.1% gelatine. U2OS cells were maintained in DMEM (31966-021; Gibco) supplemented with 10% FCS (10270-106; Gibco) and 1% Pen-Strep (15140-122; Gibco). WI38 cells were maintained in the same medium used for patient-derived skin fibroblasts. For transfections, siRNAs with a final concentration of 40 nm and RNAiMAX (13778; Life Technologies) reagent were used. About 1 mg/ml (pH 7.4) PEI (23966; Polysciences) was used in the transfection assay for overexpression. Cell cultures tested negative for mycoplasma on a regular basis.

RNA isolation and qRT-PCR

Total RNA was extracted using QIAzol (5346994; Qiagen). About 1 µg of total RNA was used for reverse transcription with the RevertAid H Minus First Strand cDNA Synthesis Kit (K1632; Thermo). qRT-PCR was performed in triplicate on a C1000TM Thermal cycler (Bio-Rad) with SYBR Green (170-8887; Bio-Rad). Data was normalized to β-actin for mouse and GUSB for human. Primer sequences are provided in

.

RNA sequencing libraries

RNA-seq libraries were generated by BGI from total RNA using the TruSeq RNA-Seq library prep kit (Illumina), with an insert size of ∼160 bp. Three independent biological replicates for each group (wild-type or homozygous Zbtb24ΔBTB mutant ESCs) were sequenced paired-end (2 × 91 bp) on the Illumina HiSeq2000 platform (Accession number for RNA-seq data: European Nucleotide Archive ERP010761 http://www.ebi.ac.uk/ena/data/view/ERP010761, date last accessed July 26, 2016).

Pre-processing, alignment of RNA-seq reads and differential expression analysis

After sequencing, adapter sequences were removed, and raw reads were filtered for low quality reads (Q20> 97%). An average of 20 million paired-end reads were aligned to the mouse genome (mm9) using GSNAP (version 2013-11-27), with the following parameters: -N 1 –n 1. BAM files were converted into mpileup files using SAMtools, and subsequently converted into WIG format for visualization purposes, using a custom python script. The number of mapped reads was quantified at gene level using HTSeq (version 0.6.1p1) with default parameters (union mode), based on Ensembl annotation (version 64) (ftp://ftp.ensembl.org/pub/release-64/gtf/mus_musculus/Mus_musculus.NCBIM37.64.gtf.gz) (29). The R Bioconductor package DESeq2 (version 1.6.3 on R version 3.1.1) was used for analysis of differential gene expression analysis between wild-type mice and homozygous Zbtb24ΔBTB mutant mESCs (30).

Mapping of RNA-seq reads to the artificially engineered Zbtb24 locus

The mouse mm9 genome reference sequence was modified by inserting a reverse PGK-Neomycin cassette between exon1 and exon2 of the Zbtb24 locus to create an artificial genome reference resembling our transgenic mouse. Gmap_build.data (31) was used to build a reference genome for GSNAP and subsequently reads from one wild-type and one homozygous Zbtb24ΔBTB mutant mESC samples were aligned to this artificial reference genome. The Sashimi-plot option in IGV was used to visualize the alignment (32).

Statistical Analysis

A Student’s t test and Standard error of mean (SEM) were used for all the statistical analysis. Unless otherwise stated statistical analysis was done either on at least two biological replicates or at least two independent experiments. Values of P < 0.05 were considered significant.

Immunoprecipitation

Cells were washed twice with cold PBS and lysed in IP Buffer (50 mm Tris (pH 7.5), 150 mm NaCl, 1% NP-40, 1 mm EDTA) with Protease inhibitor cocktail (05056489001; Roche) and Phosphatase inhibitor cocktail (04906837001; Roche) on ice. A BCA kit (23225; Thermo) was used to determine protein concentration. Protein A/G plus-Agarose Beads (sc-2003; Santa Cruz) were first incubated with 1 µg Ty1 antibodies (SAB4800032; Sigma; C15200054; Diagenode) at 4 °C for at least 1 h. Total cell extracts were pre-cleared with beads and equal amounts of total protein were added into beads-antibody mix and incubated at 4 °C overnight. After immunoprecipitation, beads were washed with IP buffer 4 times and then boiled for 10 min to collect samples. IP samples were analyzed by Western blotting.

Chromatin immunoprecipitation

Cells were cross linked with 1% formaldehyde (344198; Calbiochem) for 10 min at RT and glycine (125 mm) was used to quench cross-linking for 5 min. Cells were washed twice with cold PBS and lysed with NP Buffer [150 mm NaCl, 50 mm Tris–HCl (pH 7.5), 5 mm EDTA, 0.5% NP-40, 1% Triton X-100, Protease inhibitor cocktail (05056489001; Roche)]. Nuclei were sheared by sonication (Diagenode Biorupter Pico). Protein A and G Beads (10002D, 10003D; Life Technologies) were first blocked with BSA (A7906; Sigma) and then incubated with antibodies at 4 °C for at least 4 h. About 2 µg Ty1 antibody (C15200054; Diagenode) coupled with beads were incubated with sheared chromatin at 4 °C overnight. After immunoprecipitation, beads were washed with low-salt washing buffer [0.1% SDS, 1% Triton X-100, 2 mm EDTA, 20 mm Tris–HCl (pH 8.1), 150 mm NaCl], high-salt washing buffer (0.1% SDS, 1% Triton X-100, 2 mm EDTA, 20 mm Tris–HCl (pH 8.1), 500 mm NaCl), LiCl washing buffer [0.25 m LiCl, 1% NP40, 1% deoxycholate, 1 mm EDTA, 10 mm Tris–HCl (pH 8.1)] and TE buffer (10 mm Tris–HCl (pH 8.0), 1 mm EDTA). DNA was extracted using phenol-chloroform-isoamylol (15593-049; Life Technologies) and used for analysis. Primers can be found in

.

Western blotting

Cells were lysed in Cell Lysis buffer (20 mm triethanolamine (T1377; Sigma), 0.14 m NaCl, 0.1% Sodium deoxycholate (D6750; Sigma), 0.1% SDS, 0.1% Triton X-100) with Protease inhibitor cocktail (05056489001; Roche) and Phosphatase inhibitor cocktail (04906837001; Roche) on ice. BCA kit was used to determine protein concentration. Equal amounts of total cell extracts were separated on a NuPAGE gel (4–12%; NP0321), and transferred to a Nitrocellulose Blotting Membrane (10600016; Life Sciences). The following primary antibodies were used: Zbtb24 (PM085; MBL Life Science, 1:1000), Cdca7 (15249-1-AP; Proteintech, 1:500), Dnmt3b (Ab16049; Abcam, 1:1000), Hells (11955-1-AP; Proteintech, 1:1000), Ty1 (SAB4800032; Sigma, 1:1000) and Tubulin (T6199; Sigma, 1:5000). Goat anti-Rabbit 800CW (926-32211; Westburg, 1:5000), Donkey anti-mouse 680RD (926-68072; Li-Cor, 1:5000), Goat anti-guinea pig A594 (A-11076; ThermoFisher, 1:5000) and Donkey anti-Goat 800CW (926-32214; LI-COR, 1:5000) were used as secondary antibodies. Membranes were analyzed on Odyssey (Westburg).

Luciferase reporter assay

Suspended Cells were transfected with Lipofectamine 2000 (11668-027; Life Technologies). Cells were harvested 2 days after transfection and treated with Dual-Luciferase Reporter Assay System kit (E1910; Promega). A Perkin Elmer precisely 1420 Multilabel counter victor 3 was used to measure Luciferase intensity.

Southern blot

Southern blotting was used to determine integration of the PGK-NEO cassette. Primers used to generate the probe can be found in

. A buffer containing 0.125 m Na2HPO4 (pH 7.2), 10% PEG6000, 0.25 m NaCl, 1 mm EDTA and 7% SDS was used for hybridization for 16–24 h at 65 °C. Final washing conditions were 2× SSC-0.1% SDS (p13E-11). The blot was exposed for 16–24 h to Phosphorimager screens and analyzed with the Image Quant software program (Molecular Dynamics).

Cloning

Full length human ZBTB24 was PCR amplified from a cDNA library obtained from HCT116 cells using Phusion high fidelity DNA polymerase (NEB) with the following primers: ZBTB24_F and ZBTB24_R. The obtained fragment was subcloned into the Zero Blunt Topo cloning vector. To introduce restriction sites (BglII/SalI), we amplified the coding sequence using Phusion with the following primers: ZBTB24_F2 and ZBTB24_R2. Subsequently, the ZBTB24 ORF and the pEGFP-c1 expression vector (Promega) were digested with BglII and SalI and ligated. To generate the 3xTy1_ZBTB24 construct, 3xTy1 tag was amplified from pTALYM9 vector (47876; Addgene) with flanking restriction sites (NheI/BglII) (3xTy1-NheI_F, 3xTy1-BglII_R), and GFP was replaced by 3xTy1 in pEGFP-c1 vector containing ZBTB24. For Luciferase assay experiments around 1000 bp upstream of the Cdca7 and Cdc40 transcription start sites were PCR amplified (Cdca7-pro-MluI_F, Cdca7-pro-HindIII_R, Cdc40-pro-MluI_F, Cdc40-pro-HindIII_R) with flanking restriction sites for cloning (MluI/HindIII), and the PCR product was inserted into the pGL3-Basic vector. Around 1000-bp upstream of the Ostc transcription start site was obtained by PCR (Ostc-pro-MluI_F, Ostc-pro-BglII_R) with flanking restriction sites for cloning (MluI/BglII), the PCR product was inserted into the pGL3-Basic Vector. All constructs were verified using Sanger sequencing. Primers used can be found in

.

Polysome profiling

Polysome RNA was extracted from mESCs cultured in 10-cm dishes. Translation elongation was blocked and cytoplasmic RNA was recovered as described previously (33). Lysate was layered on frozen sucrose gradients (7–46% sucrose) and separated by ultracentrifugation at 35 000 rpm in a SW 41 Ti rotor (210 000 g) for 3 h at 4 °C. About 15 fractions (750 μl each) were collected from the top and digested with proteinase K (0.15 mg/750 μl) for 30 min at 42 °C in the presence of 1% sodium dodecyl sulphate. RNA was extracted by acid phenol (Ambion) purification followed by ethanol precipitation. For each sucrose gradient separation, the polysome profile was determined on a Bioanalyzer (Agilent) with the RNA 6000 Nano kit. Fractions containing polysomes (corresponding to fractions 10–15) were combined.

Total RNA was extracted from mESCs using NucleoSpin RNA II kit (Macherey-Nagel) according to the manufacturer’s instructions. cDNA was synthesized from 1 μg of polysome or total RNA using BioScript MMLV Reverse Transcriptase (Bioline) with 40 ng of random hexamer and oligo(dT)18 primers. qRT-PCR was performed with Zbtb24-RT-PCR primers (

) on a LightCycler 480 (Roche) using 2× SensiMix reagent (Bioline). PCR efficiency was determined using the LinRegPCR program (34,35). Ratios between Zbtb24 relative expression levels in total RNA and polysome RNA were calculated and significance was tested performing a t-test.

Supplementary Material

is available at HMG online.

Acknowledgements

We thank the ICF patients and their families for their participation. We thank the members of the Transgenesis Facility Leiden for generation of the Zbtb24 mouse and assistance with ESC derivation.

Conflict of Interest statement. None declared.

Funding

LUMC (Fellowship to L.D.) and Dutch Cancer Society (UL2012-5460).

References

1
Weemaes
C.M.
van Tol
M.J.
Wang
J.
van Ostaijen-ten Dam
M.M.
van Eggermond
M.C.
Thijssen
P.E.
Aytekin
C.
Brunetti-Pierri
N.
van der Burg
M.
Graham Davies
E
. et al.  . (
2013
)
Heterogeneous clinical presentation in ICF syndrome: correlation with underlying gene defects
.
Eur. J. Hum. Genet
 .,
21
,
1219
1225
.
2
Hagleitner
M.M.
Lankester
A.
Maraschio
P.
Hulten
M.
Fryns
J.P.
Schuetz
C.
Gimelli
G.
Davies
E.G.
Gennery
A.
Belohradsky
B.H.
et al.  . (
2008
)
Clinical spectrum of immunodeficiency, centromeric instability and facial dysmorphism (ICF syndrome)
.
J. Med. Genet
 .,
45
,
93
99
.
3
Thijssen
P.E.
Ito
Y.
Grillo
G.
Wang
J.
Velasco
G.
Nitta
H.
Unoki
M.
Yoshihara
M.
Suyama
M.
Sun
Y.
et al.  . (
2015
)
Mutations in CDCA7 and HELLS cause immunodeficiency-centromeric instability-facial anomalies syndrome
.
Nat. Commun
 .,
6
,
7870.
4
Xu
G.L.
Bestor
T.H.
Bourc'his
D.
Hsieh
C.L.
Tommerup
N.
Bugge
M.
Hulten
M.
Qu
X.
Russo
J.J.
Viegas-Pequignot
E.
(
1999
)
Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene
.
Nature
 ,
402
,
187
191
.
5
Nitta
H.
Unoki
M.
Ichiyanagi
K.
Kosho
T.
Shigemura
T.
Takahashi
H.
Velasco
G.
Francastel
C.
Picard
C.
Kubota
T.
et al.  . (
2013
)
Three novel ZBTB24 mutations identified in Japanese and Cape Verdean type 2 ICF syndrome patients
.
J. Hum. Genet
 .,
58
,
455
460
.
6
de Greef
J.C.
Wang
J.
Balog
J.
den Dunnen
J.T.
Frants
R.R.
Straasheijm
K.R.
Aytekin
C.
van der Burg
M.
Duprez
L.
Ferster
A.
et al.  . (
2011
)
Mutations in ZBTB24 are associated with immunodeficiency, centromeric instability, and facial anomalies syndrome type 2
.
Am. J. Hum. Genet
 .,
88
,
796
804
.
7
Ehrlich
M.
Buchanan
K.L.
Tsien
F.
Jiang
G.
Sun
B.
Uicker
W.
Weemaes
C.M.
Smeets
D.
Sperling
K.
Belohradsky
B.H.
et al.  . (
2001
)
DNA methyltransferase 3B mutations linked to the ICF syndrome cause dysregulation of lymphogenesis genes
.
Hum. Mol. Genet
 .,
10
,
2917
2931
.
8
Heyn
H.
Vidal
E.
Sayols
S.
Sanchez-Mut
J.V.
Moran
S.
Medina
I.
Sandoval
J.
Simo-Riudalbas
L.
Szczesna
K.
Huertas
D.
et al.  . (
2012
)
Whole-genome bisulfite DNA sequencing of a DNMT3B mutant patient
.
Epigenetics
 ,
7
,
542
550
.
9
Jin
B.
Tao
Q.
Peng
J.
Soo
H.M.
Wu
W.
Ying
J.
Fields
C.R.
Delmas
A.L.
Liu
X.
Qiu
J.
et al.  . (
2008
)
DNA methyltransferase 3B (DNMT3B) mutations in ICF syndrome lead to altered epigenetic modifications and aberrant expression of genes regulating development, neurogenesis and immune function
.
Hum. Mol. Genet
 .,
17
,
690
709
.
10
Velasco
G.
Walton
E.L.
Sterlin
D.
Hedouin
S.
Nitta
H.
Ito
Y.
Fouyssac
F.
Megarbane
A.
Sasaki
H.
Picard
C.
et al.  . (
2014
)
Germline genes hypomethylation and expression define a molecular signature in peripheral blood of ICF patients: implications for diagnosis and etiology
.
Orphanet J. Rare Dis
 .,
9
,
56.
11
Huang
K.
Wu
Z.
Liu
Z.
Hu
G.
Yu
J.
Chang
K.H.
Kim
K.P.
Le
T.
Faull
K.F.
Rao
N.
et al.  . (
2014
)
Selective demethylation and altered gene expression are associated with ICF syndrome in human-induced pluripotent stem cells and mesenchymal stem cells
.
Hum. Mol. Genet
 .,
23
,
6448
6457
.
12
Ueda
Y.
Okano
M.
Williams
C.
Chen
T.
Georgopoulos
K.
Li
E.
(
2006
)
Roles for Dnmt3b in mammalian development: a mouse model for the ICF syndrome
.
Development
 ,
133
,
1183
1192
.
13
Velasco
G.
Hube
F.
Rollin
J.
Neuillet
D.
Philippe
C.
Bouzinba-Segard
H.
Galvani
A.
Viegas-Pequignot
E.
Francastel
C.
(
2010
)
Dnmt3b recruitment through E2F6 transcriptional repressor mediates germ-line gene silencing in murine somatic tissues
.
Proc. Natl. Acad. Sci. U.S.A
 .,
107
,
9281
9286
.
14
Youngson
N.A.
Epp
T.
Roberts
A.R.
Daxinger
L.
Ashe
A.
Huang
E.
Lester
K.L.
Harten
S.K.
Kay
G.F.
Cox
T.
et al.  . (
2013
)
No evidence for cumulative effects in a Dnmt3b hypomorph across multiple generations
.
Mamm. Genome
 ,
24
,
206
217
.
15
Dennis
K.
Fan
T.
Geiman
T.
Yan
Q.
Muegge
K.
(
2001
)
Lsh, a member of the SNF2 family, is required for genome-wide methylation
.
Genes Dev
 .,
15
,
2940
2944
.
16
Geiman
T.M.
Tessarollo
L.
Anver
M.R.
Kopp
J.B.
Ward
J.M.
Muegge
K.
(
2001
)
Lsh, a SNF2 family member, is required for normal murine development
.
Biochim. Biophys. Acta
 ,
1526
,
211
220
.
17
Geiman
T.M.
Muegge
K.
(
2000
)
Lsh, an SNF2/helicase family member, is required for proliferation of mature T lymphocytes
.
Proc. Natl. Acad. Sci. U.S.A
 .,
97
,
4772
4777
.
18
Ren
J.
Briones
V.
Barbour
S.
Yu
W.
Han
Y.
Terashima
M.
Muegge
K.
(
2015
)
The ATP binding site of the chromatin remodeling homolog Lsh is required for nucleosome density and de novo DNA methylation at repeat sequences
.
Nucleic Acids Res
 .,
43
,
1444
1455
.
19
Lee
S.U.
Maeda
T.
(
2012
)
POK/ZBTB proteins: an emerging family of proteins that regulate lymphoid development and function
.
Immunol. Rev
 .,
247
,
107
119
.
20
Edgar
A.J.
Dover
S.L.
Lodrick
M.N.
McKay
I.J.
Hughes
F.J.
Turner
W.
(
2005
)
Bone morphogenetic protein-2 induces expression of murine zinc finger transcription factor ZNF450
.
J. Cell. Biochem
 .,
94
,
202
215
.
21
Gill
R.M.
Gabor
T.V.
Couzens
A.L.
Scheid
M.P.
(
2013
)
The MYC-associated protein CDCA7 is phosphorylated by AKT to regulate MYC-dependent apoptosis and transformation
.
Mol. Cell. Biol
 .,
33
,
498
513
.
22
Guiu
J.
Bergen
D.J.
De Pater
E.
Islam
A.B.
Ayllon
V.
Gama-Norton
L.
Ruiz-Herguido
C.
Gonzalez
J.
Lopez-Bigas
N.
Menendez
P.
et al.  . (
2014
)
Identification of Cdca7 as a novel Notch transcriptional target involved in hematopoietic stem cell emergence
.
J. Exp. Med
 .,
211
,
2411
2423
.
23
Deweindt
C.
Albagli
O.
Bernardin
F.
Dhordain
P.
Quief
S.
Lantoine
D.
Kerckaert
J.P.
Leprince
D.
(
1995
)
The LAZ3/BCL6 oncogene encodes a sequence-specific transcriptional inhibitor: a novel function for the BTB/POZ domain as an autonomous repressing domain
.
Cell Growth Differ
 .,
6
,
1495
1503
.
24
Ying
Q.L.
Wray
J.
Nichols
J.
Batlle-Morera
L.
Doble
B.
Woodgett
J.
Cohen
P.
Smith
A.
(
2008
)
The ground state of embryonic stem cell self-renewal
.
Nature
 ,
453
,
519
523
.
25
Ficz
G.
Hore
T.A.
Santos
F.
Lee
H.J.
Dean
W.
Arand
J.
Krueger
F.
Oxley
D.
Paul
Y.L.
Walter
J.
et al.  . (
2013
)
FGF signaling inhibition in ESCs drives rapid genome-wide demethylation to the epigenetic ground state of pluripotency
.
Cell Stem Cell
 ,
13
,
351
359
.
26
Marks
H.
Kalkan
T.
Menafra
R.
Denissov
S.
Jones
K.
Hofemeister
H.
Nichols
J.
Kranz
A.
Stewart
A.F.
Smith
A.
et al.  . (
2012
)
The transcriptional and epigenomic foundations of ground state pluripotency
.
Cell
 ,
149
,
590
604
.
27
Yu
W.
McIntosh
C.
Lister
R.
Zhu
I.
Han
Y.
Ren
J.
Landsman
D.
Lee
E.
Briones
V.
Terashima
M.
et al.  . (
2014
)
Genome-wide DNA methylation patterns in LSH mutant reveals de-repression of repeat elements and redundant epigenetic silencing pathways
.
Genome Res
 .,
24
,
1613
1623
.
28
Zhu
H.
Geiman
T.M.
Xi
S.
Jiang
Q.
Schmidtmann
A.
Chen
T.
Li
E.
Muegge
K.
(
2006
)
Lsh is involved in de novo methylation of DNA
.
EMBO J
 .,
25
,
335
345
.
29
Anders
S.
Pyl
P.T.
Huber
W.
(
2015
)
HTSeq–a Python framework to work with high-throughput sequencing data
.
Bioinformatics
 ,
31
,
166
169
.
30
Love
M.I.
Huber
W.
Anders
S.
(
2014
)
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
 .,
15
,
550.
31
Wu
T.D.
Watanabe
C.K.
(
2005
)
GMAP: a genomic mapping and alignment program for mRNA and EST sequences
.
Bioinformatics
 ,
21
,
1859
1875
.
32
Robinson
J.T.
Thorvaldsdottir
H.
Winckler
W.
Guttman
M.
Lander
E.S.
Getz
G.
Mesirov
J.P.
(
2011
)
Integrative genomics viewer
.
Nat. Biotechnol
 .,
29
,
24
26
.
33
de Klerk
E.
Fokkema
I.F.
Thiadens
K.A.
Goeman
J.J.
Palmblad
M.
den Dunnen
J.T.
von Lindern
M t.
Hoen
P.A.
, (
2015
)
Assessing the translational landscape of myogenic differentiation by ribosome profiling
.
Nucleic Acids Res
 .,
43
,
4408
4428
.
34
Ramakers
C.
Ruijter
J.M.
Deprez
R.H.
Moorman
A.F.
(
2003
)
Assumption-free analysis of quantitative real-time polymerase chain reaction (PCR) data
.
Neurosci. Lett
 .,
339
,
62
66
.
35
Ruijter
J.M.
Ramakers
C.
Hoogaars
W.M.
Karlen
Y.
Bakker
O.
van den Hoff
M.J.
Moorman
A.F.
(
2009
)
Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR data
.
Nucleic Acids Res
 .,
37
,
e45.

Supplementary data