Role of ALKBH1 in the Core Transcriptional Network of Embryonic Stem Cells

In E. coli, the AlkB protein is involved in repair of methylations-induced DNA damage by oxidative demethylation [1,2]. However, the nine mammalian AlkB homologous appear to have diverse functions, with substrates including DNA, RNA and protein modifications [3,4,5,6,7,8], (reviewed in [9]).

In 2012, we demonstrated that the mammalian AlkB homologue 1 (ALKBH1) is a histone dioxygenase that removes methyl groups from histone H2A [10]. Various histone modifications have been extensively studied in pluripotent cells, and to our knowledge, this is the first dioxygenase that has been shown to remove methyl groups from histone H2A. Moreover, we demonstrated that homologous disruption of Alkbh1 in embryonic stem cells (ESCs) leads to delayed downregulation of pluripotency markers upon differentiation, and a delayed induction of neuroectodermal genes. In line with this, whole-genome expression analysis and chromatin immunoprecipitation sequencing (ChIP-seq) revealed that ALKBH1 binds to a subset of genes required for early neural development.

Pluripotency of human ESCs is maintained by a complex system that includes a core transcriptional network where OCT4, SOX2 and NANOG are among the key regulators. Precise levels of these transcription factors must be sustained for the maintenance of pluripotency, implying that the protein levels are tightly regulated by the cell. Null mutation of any of these genes results in early embryonic lethality [11,12,13]. Deletion of murine Nanog is driving the inner cell mass and ESCs into primitive endoderm, while forced expression enables autonomous self-renewal of ESCs in culture [12,14].

SOX2 is expressed not only in pluripotent stem cells, but also in multipotent cells of both embryonic- and extraembryonic lineages [11]. It can act synergistically with OCT4 to activate Oct-Sox enhancer elements and silencing of SOX2 in ESCs induces differentiation into a variety of cell lineages including trophoectoderm [15]. Interestingly, over-expression of SOX2 results in trophoectoderm differentiation as well [16].

Forced expression of the transcription factor OCT4 counteracts the effect of SOX2 deficiency [17]. This latter result suggests that the main function of SOX2 is to keep ESCs pluripotent by regulating the level of OCT4. In fact; the levels of these master pluripotency factors are crucial. Only subtle changes promote exit from the pluripotent state; illustrated by the fact that only a twofold increase of OCT4 induces differentiation toward primitive endoderm and mesoderm while a 50% decrease cause differentiation into trophoectoderm [18].

Genome-wide profiling of pluripotent- and differentiated cells suggests global chromatin remodelling during differentiation, which results in a progressive transition from an open chromatin configuration to a more compact state. This transition is formed by a close interaction by the core transcription factors, chromatin modifying enzymes, including histone demethylases, and microRNA (miRNA). miRNAs are short non-coding RNAs that function through the suppression of target genes [18,19,20,21]. A few years back, a genome-wide search was done in human ESCs to identify target genes of the core transcription factors, and OCT4 and NANOG was shown to co-occupy the promoter of the ALKBH1 gene [22]. Since ALKBH1 is a target gene of these core pluripotency transcription factors it is tempting to suggest that ALKBH1 may be involved in the regulation of self-renewal or differentiation of stem cells.

In this study we further investigate the role of ALKBH1 in pluripotency and early development. We demonstrate that NANOG and SOX2 bind to the ALKBH1 promoter, and we identify protein-protein interactions between ALKBH1 and these core transcription factors of the pluripotency network. Furthermore, lack of ALKBH1 affects the expression of developmentally important miRNAs. A better understanding of the molecular mechanisms that regulate pluripotency is critical for the utilization of pluripotent stem cells in disease modelling, drug development and regenerative medicine.

The assay was performed as described previously [23]. Briefly, radio labelled DNA was incubated with recombinant human proteins e.g. NANOG, OCT4, SOX2 (PeproTech NANOG 120-21; PeproTech SOX2 110-03; Primorigen OCT4 S2105) and filtered through a nitrocellulose membrane (Millipore MF-Millipore™ Membrane Filters VSWP01300). The nitrocellulose membrane retains proteins and DNA-protein complexes while the unbound DNA flows through. To increase the accuracy of the assay, a DEAE (diethylaminoethyl) filter was placed immediately below the nitrocellulose membrane to trap any DNA not bound by the nitrocellulose (Sigma Whatman, DE81 Z286605). Both filters were dried and subject to scintillation counting. DNA binding reaction mixtures (20μl) contained 25 mM Tris-HCl (pH = 7.5), 3 mM MgCl₂, 20 mM NaCl, 5 mM β-mercaptoethanol, 0.5 nM ³²P-labeled DNA (353 base pair NANOG/OCT4 binding sequence from Boyer et al., 2005 or 322 base pair negative control sequence (M13)), 2.9 mM AMP-PNP, and the indicated amount of either NANOG, OCT4, SOX2. The reactions were incubated for 10 min. at 37˚C, diluted to a final volume of 1ml with reaction buffer containing 50 μg/ml bovine serum albumin, and filtered onto nitrocellulose and DEAE filters as described. The background values representing DNA retention on the nitrocellulose filter in the absence of protein were typically less than 3% and were subtracted from the binding values reported.

0.5 µg of primary target proteins, NANOG, OCT4, SOX2, histones H1 (negative control) and H2A (positive control) were spotted onto a nitrocellulose membrane together with purified ALKBH1 and the negative control proteins heat-denatured lysozyme and BSA (fraction V) (PeproTech NANOG 120-21; PeproTech SOX2 110-03; Primorigen OCT4 S2105; New Englands Biolabs H1 M2501S, H2A M2502S). Membranes were then dried before blocking in 3% BSA/PBST for 2 h at RT. The membrane was subsequently incubated overnight at 4°C with purified ALKBH1 in 3% BSA/PBST (3.325 µg/ml) or with 3% BSA/PBST alone in the negative control. Membranes were then washed and incubated with anti-ALKBH1 primary antibody (Abcam ab18525, 1:1000) in 3% BSA/PBST for 2 h at 4°C. Subsequently the membranes were washed and incubated with a secondary antibody conjugated with alkaline phosphatase (Sigma A3812, 1:5000) in 3% BSA/PBST at 4˚C for 30 min. The membrane was then washed and protein-protein interactions were visualized by adding the Western Blue®Stabilized Substrate for alkaline phosphatise.

62.5 fmol of each primary target protein (NANOG, OCT4, SOX2, and histones H1, H2A, H2B, H3, H4 and BSA as a negative control) in binding buffer (25 mM HEPES (pH 7.5) and 50 mM NaCl) containing 1.6 mM Sulfo-NHS were added to a CovaLink plate. Proteins were bound to the plate by adding 6.5 mM EDC in binding buffer and incubating for 2 h at RT. Wells were washed with washing buffer (HEPES (pH 7.5) 25 mM, NaCl 50 mM, Tween20 0.1% (v/v), BSA 2% (w/v)) and then blocked with blocking buffer (HEPES (pH 7.5) 25 mM, NaCl 50 mM, Tween20 0.1% (v/v), BSA 3% (w/v)) overnight at 4°C. Wells were washed and a titration of the indicated amounts of purified ALKBH1 in washing buffer was added to the wells and incubated for 2 h at RT. Wells were washed and incubated for 2 h at RT with anti-ALKBH1 antibody (Abcam ab18525, 1:10 000) in washing buffer. Wells were then washed and incubated with alkaline phosphatise-conjugated secondary antibody (Sigma A3812, 1:10 000) in washing buffer for 30 min at RT. Wells were then incubated for 30 min with p-nitrophenyl phosphate (PNPP) at RT and 2M NaOH was then added to stop the reaction. The color reaction was read at 405 nm using a VICTOR21420 Multilabel counter.

Cultures of human ESCs were trypsinized, and the single cell suspension was cytospun onto Superfrost Plus slides (Thermo Scientific 4951PLUS). The cells were immediately fixed in 4% (w/v) paraformaldehyde and blocked in blocking buffer (0.1% Tween-20, 1% BSA and 10% goat serum in PBS) for 1 h. Cells were incubated overnight at 4°C with primary antibodies in blocking buffer. The following antibodies were used: mouse anti-ALKBH1 (Sigma, A8103, 1:200) and rabbit anti-NANOG (Abcam, ab21603, 1:200). Following incubation with primary antibodies, the cells were washed 3 times in PBST (PBS with 0.1% (v/v) Tween-20) and incubated with secondary antibodies in blocking buffer for 1 h at RT; Alexa488-Goat-anti-mouse or Alexa594-Goat-anti-rabbit (Life Technologies, A-11017 or A-11012, 1:500). 5 min incubation with 4'-6-diamidino-2-phenylindole (DAPI) (Life Technologies, D1306) was followed by mounting in Mowiol anti-fade mounting media (Merck Biosciences Ltd, 475904). All microscopy was done using an Axio Observer.Z1 microscope equipped plan-apochromat objectives and two cameras; one AxioCam MRm and one AxioCam ICc1. Picture analysis was done with AxioVision 4.8.1 Software (all from Carl Zeiss).

Cells were harvested at 24 h post-transfection and lysed in 50 mM Tris, pH 7.4, 150 mM NaCl, 1mM EDTA, 1% Triton X-100 supplemented with 1mM PMSF and protease inhibitor cocktail (Complete Mini EDTA free Protease Inhibitor Cocktail Tablets, Roche, 04693159001). Cell debris were removed by centrifugation at 20;000 × g for 10 min. Typically, 20 µl of anti-myc Mouse mAb Magnetic conjugate (Cell Signaling Technology, 9B11) or 40 µl of anti-flag antibody (Sigma, F7425) pre-adsorbed to Dynabeads® Protein G (Invitrogen, 100-03D), were incubated with 100 µg total MEF lysate at 4°C overnight, washed with the lysis buffer and subjected to immunoblotting.

Total MEF lysates (4 µg) or immunoprecipitated proteins were denatured at 70°C for 10 min in NuPAGE® LDS Sample Buffer (Invitrogen, NP007) (1 × final), proteins were separated in NuPAGE® Novex® 12% Bis-Tris Gels (Invitrogen, NP NP0344BOX), and blotted to nitrocellulose membranes (Trans-Blot® Turbo Mini Nitrocellulose Transfer Packs, Biorad, 170-4158). The membranes were incubated with the indicated primary antibodies at 4°C, overnight. Immunoreactivity was detected using goat-anti-rabbit (Abcam, Ab6721) or goat-anti-mouse (Abcam, Ab6789) antibodies conjugated to HRP (diluted 1:50,000 and 1:20,000, respectively) followed by chemiluminescence reaction using the SuperSignal West Dura Extended Duration Substrate (Pierce, 34075). Molecular weight of the proteins was estimated using the SuperSignal Molecular Weight Protein Ladder (Pierce, 84785). Primary antibodies were anti-myc antibody (diluted 1:5000, Invitrogen 46-0603) and anti-flag antibody (diluted 1:1500, Sigma F7425).

Mouse ESCs were cultured in KnockOut™ Dulbecco's modified Eagle's media (KO-DMEM) (Invitrogen 10829-018) supplemented with 20 % Knockout™ Serum Replacement (Knockout™ SR) (Invitrogen 10828-028), 100 U/ml Penicillin-Streptomycin (Invitrogen 15140-122), 0.1 mM nonessential amino acids (Invitrogen 11140-035), 2 mM GlutaMAX™ (Invitrogen 35050-038), 0.1 mM 2-mercaptoethanol (Sigma M7522) and 1000 U/ml leukaemia inhibitory factor (LIF) (Millipore ESG1107). All ESC cultures were grown on a layer of irradiated CF-1 mouse embryonic fibroblasts (Globalstem GSC-6001G) on gelatin coated plates. Differentiation was induced by adding 1 µM of all-trans retinoic acid (Stemgent 04-0021) and removing LIF. Proliferation and viability was assessed using the Countess® Automated Cell Counter (Invitrogen C10227) with trypan blue.

Total RNA was extracted from mouse ESCs with miRNeasy mini kit (Qiagen 217004) according to the manufacturer's protocol. Any DNA remnants were removed using DNase I, amplification grade (Invitrogen 18068-015). Quality was verified on Agilent Bioanalyzer 2100 (RIN value between 9.5 and 10.0). Analysis of differentially expressed microRNAs between wild type (WT) and Alkbh1^-/- mouse ESCs was done using a mouse microRNA microarray detecting transcripts listed in the Sanger miRBase Release 13.0. Hybridization was performed by LC Sciences (LC Science, TX, USA). Analysis of the array identified 8 miRNAs differentially expressed (p < 0.1, * = p < 0.05). TargetScan Release 5.1 (Whitehead Institute, Cambridge, USA) predicted potential biological targets of these transcripts and GO classification according to biological process (TopGO, Bioconductor) were done using the list of predicted targets. Real-time PCR using TaqMan probes was done to verify the results from the array (Applied Biosystems, miR-27b - 000409, miR-361 - 000554, miR-134 - 001186, miR-615-3p - 001960). The bar charts are presented as the mean ± SEM. For the multiple comparisons a two-way Student's t-test was performed, and the Bonferroni method was used to adjust the p-value threshold for significance (* = p < 0.1, ** = p < 0.05).

In 2006, a whole genome chromatin immunoprecipitation study, followed by microarray analysis identified that the ALKBH1 promoter, including part of exon 1, was bound by OCT4 and NANOG [22]. To investigate this binding in more detail, we assessed the in vitro binding of these core transcription factors to the 5' region of ALKBH1. Increasing amounts of recombinant NANOG, SOX2, and OCT4 were incubated with the ALKBH1-relevant 5' end-labeled fragment of template DNA. We included a 322 base pair sequence (M13) as a negative control. The amount of filter-bound radioactivity was measured to determine the fraction of bound DNA, which was plotted against the protein concentration (Fig. 1A). NANOG and SOX2 bind firmly to the ALKBH1 promoter. On the contrary, we observed no significant binding with OCT4 (Fig. 1B). This is in agreement with a recent report showing that NANOG, but not OCT4, binds to the promoter region of ALKBH1 [24]. Consistent with these data, the 5'-region of human/mouse ALKBH1/Alkbh1 contains a consensus NANOG binding site that is well conserved between mouse and human (Fig. 1C). A less conserved SOX2 binding site was identified in the human ALKBH1 5' region but not in the related mouse Alkbh1 sequence. In agreement with the data presented in Fig. 1A and B, we could not identify a conserved OCT4 binding site in the ALKBH1 gene. However, OCT4 functions as a heterodimer with SOX2 in ESCs [25], explaining why OCT4 might pull down the ALKBH1 promoter as in the study by Boyer et al., without direct binding of OCT4 to the promoter itself. According to Boyer et al., 492 human genes are co-occupied by NANOG and SOX2. These genes are either induced or repressed to maintain the pluripotent state [22]. Our data suggests that ALKBH1 is involved in the regulatory network that maintains the stem cell state.

SOX2 and NANOG bind to the promoter of ALKBH1, indicating that ALKBH1 might have an important role in ESC self-renewal and pluripotency, possibly through direct interaction with the core pluripotency factors themselves. We employed a dot-blot immunobinding assay to investigate protein-protein interactions between ALKBH1 and the core pluripotency factors. OCT4, SOX2, and NANOG was spotted onto nitrocellulose membranes and incubated with purified recombinant ALKBH1 protein. Histone H2A served as a positive control while histone H1 was used as a negative control. Subsequently, the membranes were probed with a primary antibody specific for ALKBH1 and an AP-conjugated secondary antibody. Protein-protein interactions were then visualized through Western Blue alkaline phosphatase staining. The data indicate that ALKBH1 interacts with all three core pluripotency factors in vitro (Fig. 2A). In order to substantiate this finding, we performed an adapted ELISA with increasing amounts of ALKBH1 protein. We used histone H2A as a positive control and histone H1 as a negative control. Once more we find that ALKBH1 interacts with OCT4, SOX2, and NANOG (Fig. 2B). Taken together, our in vitro data suggests a physical interaction between ALKBH1 and the core regulators of the pluripotent state.

We have previously shown that transcription of Oct4, SOX2, and Nanog is upregulated in Alkbh1^-/- mouse ESCs, particularly the Nanog transcript that was fourfold higher [10]. Furthermore, since NANOG previously was found to bind to the ALKBH1 promoter, we decided to study the interaction between ALKBH1 and NANOG in more detail [22,24]. In line with the in vitro data (Fig. 2A and B), immunofluorescent staining for ALKBH1 and NANOG revealed a highly similar intracellular localization, and we found that the two proteins co-localized in the nuclei of human ESCs (Fig. 2C). Moreover, we confirmed the interaction between ALKBH1 and NANOG in mouse embryonic fibroblast (MEF) cells using co-immunoprecipitation (co-IP) experiments (Fig. 2D). MEF cells were transiently transfected with a plasmid encoding myc-tagged ALKBH1 together with a plasmid encoding flag-tagged NANOG, followed by co-IP assays with an antibody against myc or flag (Fig. 2D, left panel). Indeed, western blot analysis of the anti-flag immunoprecipitates revealed the presence of the tagged ALKBH1 protein (Fig. 2D). Unfortunately, in our hands the anti-flag antibody did perform better in western blot analysis than in immunoprecipitation experiments hence a weak unspecific signal is visible in the flag precipitates. In the reciprocal experiment, however, a myc antibody co-immunoprecipitated NANOG (Fig. 2D). For the negative control, immunoprecipitates of non-transfected MEF cells were probed with myc-tagged ALKBH1 and/or flag-tagged NANOG (Fig. 2D). The transfection of flag-tagged NANOG and/or myc-tagged ALKBH1 or both was confirmed by western blot (Fig. 2D). The fact that NANOG co-immunoprecipitated with ALKBH1 and vice-versa combined with the co-localization seen upon co-staining of ESCs indicates that these proteins also interact in vivo.

Recently, miRNAs have emerged as important players in the regulatory network that governs pluripotency in stem cells and miRNA regulation has been integrated into the transcriptional circuitry model of ESC maintenance [26,27]. If ALKBH1 is an integral component of the pluripotency circuitry in part through interaction with the core pluripotency factors OCT4, SOX2, and NANOG, it is reasonable to assume that it could also be involved in the regulation of pluripotency-associated miRNAs. To test this hypothesis we examined the miRNA profile of WT and Alkbh1 deficient mouse ESCs cultured under self-renewal conditions by using a mouse microRNA microarray (LC Science, TX, USA). We retrieved 8 differentially expressed miRNAs in the Alkbh1^-/- mouse ESCs when compared to WT ESCs (Fig. 3A, left panel). A bioinformatics approach was used to map potential binding sites of these miRNAs. The list of genes was searched for overrepresented GO terms related to “biological process”. Interestingly, we found that the cluster of genes potentially regulated by the differentially expressed miRNAs were associated with regulation of transcription, chromatin modification and development of the nervous system (Fig. 3A, right panel). Increased expression of miR-361 and miR-134 and downregulation of miR-27b and miR-615 upon ALKBH1 depletion was confirmed by real-time PCR (Fig. 3B). miR-134 and miR-27b are known to be associated with neural differentiation [28,29], and we therefore looked into their expression upon differentiation of WT and Alkbh1^-/- ESCs (Fig. 3C). We found that whereas miR-134 was stably overexpressed in the Alkbh1^-/- cells throughout the differentiation course, downregulation of miR-27b in Alkbh1^-/- cells was observed until day 9 of the differentiation course when expression of miR-27b was restored and in fact surpassed that of the WT cells (Fig. 3C).

In this study we investigated a possible role for ALKBH1 in the pluripotency network. We have previously shown that ALKBH1 has an important function in early development in mice by regulating genes important for pluripotency and neural development in ESCs [10,30]. We identified a role for ALKBH1 in epigenetic regulation, specifically through histone H2A demethylation, suggesting that ALKBH1 exerts its effect on gene expression through regulation of H2A methylation status.

Here we show that SOX2 and NANOG bind the promoter of ALKBH1, indicating that these transcription factors regulate ALKBH1 expression. It appears that the expression of proteins interacting with the core pluripotency factors in ESCs is also controlled by key ESC transcription factors. Indeed, key ESC transcription factors control the expression of 56% of the genes in the NANOG interactome and 51% of the genes in the OCT4 interactome (reviewed in [31]). Moreover, ALKBH1 regulates the expression of the core pluripotency factors themselves, pointing towards a bi-directional regulation. A similar link between the core pluripotency factors and chromatin modifiers has previously been shown for JMJD1a and JMJD2c, two histone demethylases closely related to ALKBH1. OCT4 was found to regulate the expression of JMJD1a and JMJD2c which in turn regulates the expression of NANOG and TCL1 [32]. Depletion of JMJD1a and JMJD2c in ESCs induced differentiation, presumably caused by the reduction in TCL1 and NANOG protein levels. In our ALKBH1 depleted ESCs however, the level of the core pluripotency factors is increased and the cells displayed delayed differentiation in contrast to the spontaneous differentiation seen for the JHDM deficient cells.

We find that ALKBH1 interacts with the core pluripotency factors, OCT4, SOX2, and NANOG. All three core pluripotency factors showed interaction with ALKBH1 in vitro and we also observed co-localization of ALKBH1 and NANOG in the nuclei of ESCs. The interaction between ALKBH1 and NANOG was confirmed using co-immunoprecipitation experiments in MEF cells. There is a growing body of evidence pointing towards direct interaction of the core pluripotency factors with chromatin modifying enzymes, both on active and inactive promoters in ESCs (reviewed in [33]). We suggest that ALKBH1 could be involved in the transcriptional circuitry of pluripotent cells, regulating gene expression in concert with the core regulators of a subset of genes. Comparing published ChIP-seq data sets for ALKBH1, SOX2, OCT4 and NANOG we found overlap between their target genes, in support of this hypothesis (data not shown). Moreover, mining of our ALKBH1 ChIP-seq data revealed a range of processes important in regulation of stem cell self-renewal and pluripotency [10]. Notably, stem cell differentiation, epigenetic regulation and stem cell maintenance were identified as targets of ALKBH1 in ESCs. Interestingly, previously published iPSC datasets identified an increased ALKBH1 expression upon reprogramming of somatic cells into iPSCs [34,35,36,37]. This effect was seen regardless of somatic cell type used for reprogramming. Furthermore, ALKBH1 expression was higher in ESCs than in somatic cells and downregulated upon differentiation. Taken together, this suggests a specific role for ALKBH1 in stem cells, where it seems to be involved in fine tuning the levels of the core pluripotency factors and proteins important in the decision between pluripotency and differentiation.

Several miRNAs crucial for the maintenance of ESC pluripotency were differentially expressed in WT versus Alkbh1^-/- mouse ESCs. The important role of miRNAs in pluripotency and early development has recently become evident through a number of studies [27,29,38,39]. The expression of miRNA genes involved in regulation of pluripotency is controlled by the core pluripotency factors and recently it was shown that somatic cells could be reprogrammed into iPSCs using only miRNAs [40,41]. Of the miRNAs that were differentially regulated upon ALKBH1 depletion, we found miR-134, miR-183, miR-27b and miR-615 to be the most interesting as they are known to be involved in regulation of the ESC core transcription factors, differentiation and early development [38,42,43]. miR-361 has a proximal NANOG binding site indicating that its dysregulation might be due to the observed up-regulation of NANOG in Alkbh1^-/- mouse ESCs. miR-134 is involved in regulation of SOX2 and NANOG, while miR-27b has been implicated in a range of cell types and processes, but seems to have a role in differentiation [38,44,45]. Less is known about miR-615, however, it has been reported to reside in the Hoxc5 intron in mammals, suggesting a possible role in development as the Hox genes control patterning along the anterior-posterior axis in the developing embryo [43,46].

Previously published ChIP-seq data did not identify OCT4, SOX2 or NANOG as ALKBH1 target genes. Thus, the effect of ALKBH1 deficiency on the expression of OCT4, SOX2, and NANOG may be a result of downstream targets of ALKBH1. Such targets could be miRNAs which in turn regulate one or more of the core pluripotency factors. However, it is also possible that the repressive effect of ALKBH1 on the core pluripotency factors is due to their interaction in vivo. OCT4, SOX2, and NANOG are involved in a positive feedback loop involving their own and each other's expression. An interaction with ALKBH1 could therefore sequester the factors, interfering with the feedback loop. Furthermore, in addition to the miRNAs that directly controls OCT4, SOX2, and NANOG expression, the microarray identified miRNAs regulating processes of transcription and chromatin modification further implying a role for ALKBH1 in transcriptional regulation.

Taken together, our results point towards a role for ALKBH1 in regulation of the core transcriptional network of embryonic stem cells. Thus; aberrant expression of pluripotency markers in mice lacking ALKBH1 may explain the multitude of incompletely penetrated developmental phenotypes observed in these animals [30,47]. It has been suggested that the core pluripotency factors regulate transcription of downstream targets through the fine-tuning of chromatin states and ALKBH1 could have an important role governing the methylation status of histone H2A in concert with the core transcription factors in pluripotent stem cells.

ALKBH (AlkB homolog); ChIP-seq (Chromatin immunoprecipitation-sequencing); co-IP (co-immunoprecipitation); ELISA (enzyme-linked immunosorbent assay); ESCs (embryonic stem cells); miRNA (microRNA); MEF (mouse embryonic fibroblast); RT (room temperature); wild type (WT).

This work was funded by the Norwegian Research Council. We would like to thank Linda Tveterås and Torbjørn Rognes for technical assistance.

The authors declare that there is no conflict of interest regarding the publication of this paper.