The doublesex-related Dmrta2 safeguards neural progenitor maintenance involving transcriptional regulation of Hes1

Significance

Maintaining an intricate balance between continued progenitor proliferation and cell cycle exit/differentiation is pivotal for proper brain development. Disruption of this delicate process can lead to brain malformations, such as microlissencephaly. In this paper, we identify Dmrta2 (doublesex- and mab-3–related transcription factor a2, also known as Dmrt5) as an important transcription factor that helps regulate the fine tuning between cell cycle progression and neuronal differentiation. Mechanistically, this function of Dmrta2 involves direct transcriptional regulation of a known repressor of neurogenesis Hes1. Our findings thus add Dmrta2 to the complex regulatory machinery controlling cortical NPC maintenance, and provide an explanation for the microlissencephaly caused by Dmrta2 deficiency in model organisms and humans.

Abstract

The mechanisms that determine whether a neural progenitor cell (NPC) reenters the cell cycle or exits and differentiates are pivotal for generating cells in the correct numbers and diverse types, and thus dictate proper brain development. Combining gain-of-function and loss-of-function approaches in an embryonic stem cell-derived cortical differentiation model, we report that doublesex- and mab-3–related transcription factor a2 (Dmrta2, also known as Dmrt5) plays an important role in maintaining NPCs in the cell cycle. Temporally controlled expression of transgenic Dmrta2 in NPCs suppresses differentiation without affecting their neurogenic competence. In contrast, Dmrta2 knockout accelerates the cell cycle exit and differentiation into postmitotic neurons of NPCs derived from embryonic stem cells and in Emx1-cre conditional mutant mice. Dmrta2 function is linked to the regulation of Hes1 and other proneural genes, as demonstrated by genome-wide RNA-seq and direct binding of Dmrta2 to the Hes1 genomic locus. Moreover, transient Hes1 expression rescues precocious neurogenesis in Dmrta2 knockout NPCs. Our study thus establishes a link between Dmrta2 modulation of Hes1 expression and the maintenance of NPCs during cortical development.

Balancing neural progenitor cell (NPC) self-renewal and neuronal differentiation is essential for generating cells in correct numbers and diverse types during brain development (1, 2). As such, cortical neurogenesis is tightly regulated by a complex array of transcription factors that work in concert to coordinate NPC maintenance and differentiation. Proneural transcription factors, such as neurogenin (Neurog) and NeuroD, act as the primary initiators of differentiation through their direct regulation of target genes associated with cytoskeletal reorganization, migration, and other critical differentiation processes (3, 4). Proneural transcription factors are themselves subject to transcriptional regulation by other cortical transcription factors, such as Pax6 and Hes1. Pax6 acts upstream to promote neuronal differentiation through its direct activation of proneural genes (5). On the other hand, the basic helix–loop–helix transcription factor Hes1 promotes NPC proliferation and self-renewal through its repressive actions on proneural gene expression, thereby restricting spontaneous differentiation (6).

Significant disruptions to this delicate regulatory network can result in severe developmental defects due to altered neuronal production (1, 2). One such disorder is microlissencephaly, a rare genetic-linked group of neurodevelopmental malformations characterized by the absence of sulci and gyri of the cerebral cortex and an accompanying reduction in cortical size and volume. Recently, a loss-of-function mutation in the doublesex- and mab-3–related transcription factor a2 (DMRTA2, also known as DMRT5) gene has been reported in a case of microlissencephaly, implicating DMRTA2 as a critical regulator of cortical NPC dynamics (7).

Dmrta2 belongs to the highly conserved family of Dmrt transcription factors, whose roles in the developing reproductive system have been extensively characterized (8). Another site of expression and function of Dmrta2 has been found in the embryonic brain, however (9, 10). Dmrta2 loss of function in zebrafish leads to significant reductions in cortical size, coupled with reduced neuronal numbers (10, 11). Likewise, a smaller neocortex, particularly the dorsomedial neocortex, has been observed in mice carrying null deletions of Dmrta2 (12–14). Together with the association of DMRTA2 mutation and microlissencephaly in humans, these findings implicate Dmrta2 as an important regulator for cortical neurogenesis.

Dmrta2-null mice also exhibit agenesis of the embryonic cortical hem, however. The cortical hem is the embryonic organizer for the hippocampus and a major regulator of cortical patterning outside the hippocampus. It provides a source of Wingless-related (WNT) and bone morphogenetic protein (BMP) signaling in the dorsomedial telencephalon to control proper cortical regionalization and NPC expansion in a paracrine fashion (15, 16). Thus, the severe patterning and arealization defects in Dmrta2-null model organisms prohibit a clear dissection of a direct role of Dmrta2 in NPC behavior from the secondary effect of an overall reduction in extrinsic hem-derived signals. More recently, conditional Dmrta2 mutant mice (Dmrta2^fl/fl;Emx1-cre), which delete Dmrta2 in cortical progenitors after cortical hem formation, also have been found to have reduced cortical hemisphere size, suggesting a direct role of Dmrta2 in the control of NPC behavior that remains to be defined (14).

Embryonic stem cells (ESCs) are capable of giving rise to all somatic cell types with easy access during in vitro differentiation. Mouse and human ESCs can efficiently generate cortical NPCs in culture without any added morphogens and subsequently differentiate into layer-specific neurons in a temporally regulated fashion, recapitulating major steps of normal cortical development (17–19). In this study, we analyzed the behavior of mouse ESC-derived cortical progenitors either lacking Dmrta2 or conditionally expressing transgenic Dmrta2 (9). We report that enforced expression of Dmrta2 in cortical NPCs suppresses neuronal differentiation without affecting neurogenic competence, whereas in its absence cortical NPCs undergo precocious cell cycle exit and neuronal differentiation in vitro and in vivo. We provide evidence that Dmrta2 maintains NPC status via transcriptional regulation of Hes1. Thus, this study identifies an additional layer of genetic control by Dmrta2 in the fine-tuning of cortical NPC proliferation and terminal differentiation.

Results

Expression of Dmrta2 by ESC-Derived Cortical NPCs.

To achieve efficient induction of cortical fate from mouse ESCs, we incorporated in our protocol several measures previously shown to promote a dorsal telencephalic fate (Fig. 1A) (17). These included dual SMAD inhibition with SB431542 and LDN193189 to accelerate neural induction (20); the addition of a Wnt inhibitor, XAV, to suppress caudalization (21, 22); and the addition of cyclopamine to antagonize ventralization of NPCs by endogenous sonic hedgehog (SHH) signaling (18). (Cultures generated by this paradigm are referred to hereinafter as cortical cultures.) As negative controls for cortical identity, we induced ESCs in parallel toward a ventral telencephalic fate with SHH, a caudal fate with retinoic acid, and a ventral mesencephalic (dopaminergic) fate using a combination treatment with ERK inhibitor and SHH (Fig. S1A) (23).

Fig. 1.

Dmrta2 expression in ESC-derived cortical NPCs. (A) Schematic representation of the ESC cortical differentiation protocol. (B) Dmrta2 immunostaining on day 6 of differentiation showing colabeling with other cortical markers. (C) Colocalized immunostaining of Dmrta2 and Pax6 in neural rosettes on day 8 of differentiation, with Eomes⁺ basal progenitor extending distally. (D) Quantification of the proportion of Dmrta2⁺ cells between days 4 and 10 of differentiation. Data are presented as mean ± SEM of three independent experiments. (E) Immunocytochemistry for deep [Bcl11b (Ctip2) and Tbr1] and superficial (Satb2) layer cortical neuronal markers at day 15. (F) High-magnification images of vGlut1 and Bcl11b staining in dorsal telencephalic neurons. (Scale bars: 100 µm.)

Fig. S1.

ESC differentiation produces highly homogenous cultures of cortical NPCs and neurons. (A) Schematic representation of alternative differentiation protocols used to derive NPCs and neurons of ventral telencephalic, ventral midbrain, and more caudal identities. (B) Additional characterization data confirming that Dmrta2 expression is restricted to telencephalic NPCs of dorsal, but not ventral, identity, as demonstrated by colocalization with Coup-TF1 but not with Nkx2.1. (C) Additional characterization data demonstrating the absence of FoxA2 staining in cortical NPC cultures. (D) Additional immunostaining data demonstrating the colocalization of Dmrta2 with Otx2 in cortical NPC cultures. (E–G) Additional characterization data confirming a glutamatergic phenotype of neurons derived from cortical NPC cultures demonstrated by positive immunostaining for Bcl11b and vGlut1, and an absence of staining for GAD65/67 (E), Th and Nurr1 (F), and Isl1 and Olig2 (G). (Scale bars: 100 µm.)

The generation of cortical NPCs was verified by immunostaining at day 6 to day 8 for cortical-specific or cortical-enriched markers Pax6, Lmx1a, Eomes (Tbr2), Otx2, Coup-TF1, and FORSE-1 (also known as LeX) along with Nestin as a pan-NPC marker (Fig. 1 B and C and Fig. S1 B and D). The vast majority of cells in the cortical cultures stained positive for Pax6, FORSE-1, Otx2, Lmx1a, Coup-TF1, and Nestin, whereas a proportion of cells also expressed Eomes. The transcription factor Nkx2.1 is specifically expressed by medial ganglionic eminence progenitors. Although abundant Nkx2.1⁺ cells were detected in SHH-treated ventral telencephalic cultures, negligible numbers of Nkx2.1⁺ cells were found in cortical cultures (Fig. S1B). Foxa2 is a marker for the ventral midbrain and spinal cord. Few Foxa2⁺ cells were observed in the cortical cultures, whereas they constituted the major population in dopaminergic differentiated cultures (Fig. S1C).

Double immunocytochemistry showed that Dmrta2⁺ cells were confined to Nestin⁺ and FORSE-1⁺ NPCs in cortical cultures, representing 30–54% of the total cell population between day 4 and day 10 (Fig. 1 B and D). During development, Dmrta2 expression is restricted to the dorsal telencephalon, where it is coexpressed with Pax6 but in an opposite gradient (12, 13, 24). Consistent with its expression in vivo, we found that Dmrta2 and Pax6 staining largely overlapped in ESC-derived NPCs localized in neural rosettes, from which Eomes⁺ basal progenitor cells could be seen extending distally (Fig. 1C). In contrast, no Dmrta2⁺ cells were found in SHH- or retinoic acid-treated NPC cultures (Fig. S1 B and D).

Dmrta2⁺ cells were no longer detectable by day 15 of differentiation. At this stage, the presence of postmitotic cortical neurons was confirmed by immunostaining for cortical layer-specific neuronal markers Tbr1 (layer VI), Bcl11b (Ctip2; layers V and VI), and Satb2 (layer II/III) and a pan-glutamatergic neuronal marker, vGlut1 (Fig. 1 E and F and Fig. S1E). Very few GABAergic neurons, identified by GAD65/67 immunostaining, were observed in the cortical cultures (Fig. S1E). Moreover, we did not observe TH⁺/Nurr1⁺ dopaminergic neurons or Isl1⁺/Olig2⁺ spinal motor neurons, confirming an enrichment of cortical neurons in our cultures (Fig. S1 F and G). Taken together, our data demonstrate the ability to reproduce in vitro Dmrta2⁺ dorsal telencephalic NPCs and their neuronal progeny.

Enforced Expression of Dmrta2 Suppresses NPC Neuronal Differentiation.

To investigate a role for Dmrta2 in telencephalic NPC behavior, we first examined the effect of Dmrta2 gain of function in neuronal differentiation of ESC-derived NPCs using a tetracycline-inducible Dmrta2 transgenic mESC model (Dmrta2-ESCs) reported previously (9). These cells harbor the reverse tetracycline-controlled transactivator (rtTA) and produce significant levels of Dmrta2 protein in response to the addition of doxycycline to the culture medium. The Dmrta2 transgene was induced at the peak of NPC production for 7 d starting at day 5, and the expression of several neurogenic genes was examined by quantitative PCR (qPCR) (Fig. 2 A and B). In the control condition, without doxycycline, the level of proneural gene transcripts (Neurog2, Neurod1, and Neurod4) increased gradually from day 6, along with the immature neuronal marker gene Tubb3 (β3-tubulin); however, lower levels of all these transcripts were detected in parallel sister cultures treated with doxycycline at all time points analyzed. In contrast, the transcript levels of Hes1, a repressor of cortical neurogenesis, was robustly up-regulated (Fig. 2B). These gene expression changes were concurrent with the induced transgenic Dmrta2 from day 6, which remained at a higher level than in control cultures throughout.

Fig. 2.

Enforced expression of Dmrta2 in NPCs suppresses neuronal differentiation. (A) Experimental scheme. Monolayer cultures of Dmrta2-ESCs were exposed to doxycycline or vehicle control from day 5 to day 12. Cultures were harvested every day from day 6 to day 12, and samples were processed for qPCR. (B) qPCR analysis of the genes indicated from day 6 to day 12. Levels of mRNA expression were normalized to day 5. Error bars indicate mean ± SEM of three biological replicates. (C) Phase-contrast view of day 10 cultures treated with doxycycline (Bottom) or vehicle (Top) from day 5. (D) Sister cultures as in C double-stained with antibodies against Nestin (green) and Tubb3 (red). (E) Day 10 cultures as in C double-stained for Tubb3 (green) and Dmrta2 (red). (Scale bars: 100 µm.)

Consistent with the qPCR observations, cells exposed to doxycycline for 5 d maintained a largely NPC morphology, whereas the sister control cells progressed to terminal differentiation into neurons (Fig. 2C). Immunostaining confirmed a marked reduction of Tubb3⁺ cells in doxycycline-treated cultures compared with controls. In contrast, doxycycline-treated cultures contained more Nestin⁺ NPCs (Fig. 2D). Moreover, double immunocytochemistry for Dmrta2 and Tubb3 revealed mutually exclusive staining in doxycycline-treated cultures, providing direct evidence that a high level of Dmrta2 suppresses the neuronal differentiation of NPCs (Fig. 2E).

Interestingly, on removal of doxycycline after 4 d of treatment, the NPCs readily gave rise to Tubb3⁺ neurons (Fig. S2). This finding suggests that a high level of Dmrta2 favors NPC maintenance over neuronal differentiation without affecting their neurogenic competence.

Fig. S2.

Transgenic expression of Dmrta2 does not compromise NPC neurogenic competence. (A) Experimental scheme. Monolayer cultures of Dmrta2-ESCs were exposed to doxycycline or vehicle control from day 5 to day 9. Cultures were harvested at days 7, 9, and 14 for qPCR and at day 14 for immunocytochemistry. (B) Phase-contrast images of day 14 cultures treated with vehicle or doxycycline between day 5 and day 9, showing similar neuronal content. (Magnification: 20×.) (C) Tubb3 antibody staining of day 14 sister cultures. (Magnification: 20×.) (D) qPCR for Tubb3 demonstrating the recovery of Tubb3 expression levels at day 14 in doxycycline-treated cultures compared with control cultures.

Loss of Dmrta2 Accelerates Neuronal Differentiation of Cortical NPCs.

To gain further insight into the physiological function of Dmrta2 in neurogenesis and the cellular mechanisms that might underpin the microcephaly caused by Dmrta2 loss-of-function mutation, we generated lines of mESCs with homozygous deletion of Dmrta2 (Dmrta2^−/−) by gene targeting and directed these cells toward a cortical fate (Fig. 3 and Fig. S3). We closely monitored neural induction and neuronal differentiation in Dmrta2^−/− and isogenic control (Dmrta2^flox/flox) cultures by immunocytochemistry and qPCR (Fig. 3 A and B). Rapid neuroepithelial fate conversion was observed in both genotypes, as demonstrated by the generation of a similar proportion of Nestin⁺ NPCs at days 4 and 6 (Fig. 3A). However, following the onset of Dmrta2 expression at day 4, we observed marked differences in the temporal expression profile of the intermediate progenitor marker gene Eomes, neuronal marker Map2, and proneural transcription factors Neurog2 and Neurod1 (Fig. 3B). Map2 and Eomes levels were increased in the Dmrta2^−/− cultures, whereas both Neurog2 and Neurod1 RNA reached their highest levels sooner in Dmrta2^−/− cells compared with control cells, suggesting early initiation of a neurogenesis program. Consistent with this observation, we detected an increase in the production of Tubb3⁺ neurons in Dmrta2^−/− cultures by immunostaining at days 4, 6, and 8 compared with the isogenic control cultures (Fig. 3C). Similarly, Dmrta2^−/− cultures also contained significantly more NeuN⁺ cells (mature neurons) at days 10–14 (Fig. 3D).

Fig. 3.

Loss of Dmrta2 in cortical NPCs accelerate neurogenesis in vitro. (A) Day 4 and day 6 cultures were immunostained for Nestin, revealing comparable generation of NPCs in the control and Dmrta2^−/− cultures. (B) qPCR analysis of neuronal differentiation markers. Data are representative of three independent differentiation experiments. (C) Immunostaining and quantification of cells expressing an immature neuronal marker, Tubb3. Two-way ANOVA identified a significant increase in the overall production of Tubb3⁺ neurons by Dmrta2^−/− NPCs (F_1,16 = 8.005; P = 0.012). (D) Immunostaining and quantification of cells expressing mature neuronal marker NeuN. Two-way ANOVA revealed a significant increase in the overall maturation of neurons derived from Dmrta2^−/− NPCs (F_1,16 = 11.991; P = 0.003). (Scale bars: 100 µm.)

Fig. S3.

Generation and validation of Dmrta2^flox/flox and Dmrta2^−/− mESC lines. (A) The Dmrta2 gene was targeted by homologous recombination using a Dmrta2 targeting construct containing 5′ and 3′ arms for homologous recombination. (B) Targeting was screened by Southern blot analysis for 5′ recombination. Scal restriction enzyme digestion generated an 8.1-kb fragment in WT and an 11.29-kb fragment in Dmrta2^mut ESC cells detected with the 5′ probe. (C) Dmrta2^flox alleles were distinguished from WTDmrta2 by PCR to generate 479-bp and 365-bp fragments, respectively. (D) Corresponding images of Pax6 and Dmrta2 immunostaining of Dmrta2^flox/flox and Dmrta2^−/− NPCs on day 8 of differentiation, confirming knockout of Dmrta2 protein. (Scale bar: 100 µm.)

In the reduced cortex of Dmrta2^−/− embryos, a transient increase in neuronal production also has been observed during early corticogenesis. This excess neuronal production during early neurogenesis may be a secondary consequence of the reduction of Wnt cortical hem signals or a direct consequence of the loss of Dmrta2 in cortical progenitors (12). To test this latter possibility, we used immunostaining to assay the amount of Tubb3⁺ and Eomes⁺ cells in the cortical plate of E11 conditional Dmrta2 mutant mice (Dmrta2^fl/fl;Emx1-cre), in which the Wnt signaling pathway appears to be unaffected (14). We found increased numbers of Tubb3⁺ and Eomes⁺ cells in the conditional Dmrta2 mutant mice relative to controls (Fig. 4). Thus, premature neuronal differentiation is also a feature in vivo of conditional Dmrta2 mutant mice, corroborating our in vitro observations and suggesting a direct role for Dmrta2 in cortical NPC neurogenesis.

Fig. 4.

Loss of Dmrta2 in cortical NPCs accelerate neurogenesis in vivo. Immunostaining of Tubb3 (Left) and Eomes (Right) on coronal brain sections of E11 embryos. Note the increase of Tubb3⁺ and Tbr2⁺ cells (arrows in Right) in the Dmrta2^fl/fl;Emx1-Cre mutants compared with controls. The graph is a representation of the number of Tubb3⁺ cells in Dmrta2^fl/fl;Emx1-Cre and control embryos. Data are mean ± SEM of three independent experiments; *P < 0.05, two-tailed Student’s t test. (Scale bars: 100 µm.)

Altered Dmrta2 Levels Lead to Cell Cycle Dysregulation in Cortical NPCs.

Disruptions of cortical NPC proliferation and cell cycle progression have been implicated as underlying mechanisms for microlissencephaly (25). To determine whether Dmrta2 plays a role in cell cycle regulation, we performed a flow cytometry-based cell cycle analysis to reveal the distribution of Nestin⁺ NPCs in three major phases of the cell cycle: G0/1, S, and G2/M (Fig. 5A). We found significantly more Dmrta2^−/− cortical NPCs than control cells in the G0/1 phase (P < 0.01) at day 6 of differentiation (Fig. 5B). Accordingly, the number of cells in the S phase was reduced (P < 0.05) in day 6 Dmrta2^−/− NPCs compared with controls, although no differences were seen on days 8 and 10.

Fig. 5.

Disruption of cell cycle progression in Dmrta2^−/− NPCs. (A) Cell cycle analysis of Dmrta2^flox/flox and Dmrta2^−/− NPCs by flow cytometry. NPCs were first immunostained with antibodies against Nestin (Top), and DNA content was measured by DAPI labeling (Bottom). (B) Quantification of cell distribution in G0/1, S and G2/M phases of the cell cycle. Data are presented as mean ± SEM of three independent experiments. *P < 0.05, **P < 0.01, two-way ANOVA (F_1,12 = 7.109; P = 0.021), followed by Sidak’s post hoc test. (C) Quantification of EdU uptake, and the number of cells expressing cdkn1b (p27kip1) and cdkn1c (p57kip2). Two-way ANOVA identified overall reduced proliferation in Dmrta2^−/− NPCs, as indicated by EdU uptake (F_1,12 = 11.336; P = 0.006), and an increase in the proportion of cells staining positive for Cdkn1b (F_1,12 = 20.804; P < 0.001) and Cdk1nc (F_1,12 = 10.477; P = 0.007). Data are presented as mean ± SEM of three independent experiments. (Scale bar: 100 µm.)

To gain further insight into Dmrta2-regulated cell cycle progression, we carried out EdU incorporation assays at days 6, 8, and 10 of differentiation. These assays revealed a reduced number of EdU-labeled cells in Dmrta2^−/− cultures compared with control cultures at all three time points, providing independent evidence of altered S phase in Dmrta2^−/− NPCs (Fig. 5C). Moreover, an immunocytochemical analysis showed that the numbers of cells expressing Cdkn1b (p27kip1) and Cdkn1c (p57Kip2) were increased in Dmrta2^−/− cultures compared with controls (Fig. 5C). Cdkn1b and Cdkn1c are cell cycle regulators with a major function in halting or slowing the G1-S phase transition, and thus their up-regulation is closely associated with cell cycle exit and neuronal differentiation. In contrast, we observed a reduction at the transcript level of these two cell cycle regulators and an increase in the proportion of Ki67⁺ cells when NPCs were forced to express Dmrta2 (Fig. S4). Taken together, these findings identify an unexpected role for Dmrta2 in NPC cell cycle regulation, and suggest that the disrupted regulation of cell cycle progression in Dmrta2^−/− NPCs may be a significant contributor to the precocious neurogenesis described above.

Fig. S4.

Enforced expression of Dmrta2 transgene keeps cortical NPCs in the cell cycle. (A) Experimental scheme. Monolayer cultures of Dmrta2-ESCs were exposed to doxycycline or vehicle control from day 5 to day 12. Cultures were harvested each day from day 6 to day 12, and samples were processed for qPCR. (B) qPCR analysis of Cdkn1b and Cdkn1c from day 6 to day 12. mRNA expression levels were normalized to day 5. Error bars indicate mean ± SEM of three biological replicates. (C) Day 12 cultures treated with doxycycline (Middle) or vehicle (Left) from day 5 were stained with antibodies against Ki67 (red) and counterstained with DAPI (blue), followed by quantification of Ki67⁺ cells. (Scale bar: 100 µm.)

Genome-Wide Transcriptome Profiling Supports a Role for Dmrta2 in NPC Neurogenesis.

To gain insight into the molecular mechanisms underlying the altered neurogenesis in Dmrta2-deficient NPCs, we carried out a transcriptome analysis by RNA sequencing (RNA-seq) using day 8 cultures, a timepoint associated with the highest number of Dmrta2⁺ cells and when both neurogenic and proliferative defects were apparent. Analysis of this RNA-seq dataset identified 7,343 differentially expressed transcripts at a significance level of P < 0.05 (Fig. 6A, Fig. S5, and Dataset S1). Among the Dmrta2-regulated genes were transcription factors involved in cortical development and patterning, including Foxg1, Pax6, Emx1, Emx2, Nr2f1 (Coup-TF1), and Sp8 (Fig. 6 B and C). Similarly, Lmx1a and Msx1, genes associated with development of the cortical hem, were significantly down-regulated in Dmrta2^−/− cultures. Interestingly, Dmrta2 itself was identified as one of the most significantly up-regulated transcripts on loss of Dmrta2 (Fig. 6B and Fig. S5C). The closely related Dmrt family member Dmrta1 was also up-regulated in Dmrta2^−/− NPCs, whereas no significant change was found for Dmrt3. In the reduced cortex of Dmrta2 conditional mutants, Dmrta2 and Dmrta1 also have been found to be up-regulated (14). Taken together, these observations further support a negative autoregulatory function for Dmrta2, as well as regulatory interactions with other Dmrt family members.

Fig. 6.

Genome wide transcriptome analysis supports a role for Dmrta2 in neurogenesis. (A) Heatmap depicting 7343 differentially expressed mRNA transcripts (P < 0.05) identified by RNA-seq. (B–D) Examples of differentially expressed genes associated with telencephalic patterning and development (B), transcription factors known to regulate neuronal differentiation (C), and markers of different stages of neuronal maturation (D).

Fig. S5.

Extended RNA-seq data. (A and B) Top-10 enriched GO terms for biological processes associated with transcripts identified as up-regulated (A) or down-regulated (B) in Dmrta2^−/− NPCs relative to Dmrta2^flox/flox NPCs. Differentially regulated transcripts used for this analysis were restricted to those with an adjusted P value < 0.01 and a greater than twofold fold change. Modified Fisher’s exact P value and Benjamini–Hochberg correction for multiple comparisons. (C) Top-15 transcripts identified as up-regulated or down-regulated in by Dmrta2^−/− NPCs relative to Dmrta2^flox/flox NPCs.

We are particularly interested in genes and gene sets that have a functional role in NPC proliferation and/or neuronal differentiation. Our data reveal a significant down-regulation of Hes1 (Fig. 6C). Down-regulated expression also was identified for other transcription factors known to complex with Hes1 to repress proneural gene expression and promote NPC proliferation, including Id1, Id3, and Tcf3 (6, 26, 27). In contrast, we found an up-regulation in the expression of Hes1 target proneural genes Neurog1, Neurog2, and Ascl1, together with their downstream target genes Neurod1, Neurod4, and Nhlh1 (3, 4). Furthermore, genes known to perform opposing actions to Hes1 by promoting the expression of proneural genes, including Pax6 and Btg2, were up-regulated (5, 28). Other up-regulated transcription factors with proneuronal functions included Insm1, Myt1l, and Brn2 (29, 30). Consistent with these findings, molecular markers for intermediate progenitor (Eomes), immature (Dcx), and mature (Mapt, Nefh, and Rbfox3) neurons also were up-regulated in Dmrta2^−/− cells (Fig. 6D).

To provide a broader overview of the cellular functions of Dmrta2-regulated genes/gene sets, we performed a Gene Ontology (GO) functional enrichment analysis using gene lists meeting the stringent criteria of P < 0.01 and an absolute fold change value >2. This analysis revealed that the 650 up-regulated genes meeting these criteria are enriched in transcripts associated with biological processes including neuronal differentiation, neurogenesis, and nervous system development in general (Fig. S5A). Similarly, enrichments for genes linked to the regulation of cell proliferation, organ morphogenesis, and locomotion were identified in 936 transcripts down-regulated in Dmrta2^−/− cortical NPCs (Fig. S5B). Overall, our global gene expression analysis provides strong independent support for the role of Dmrta2 in balancing NPC proliferation and neurogenesis.

Dmrta2-Controlled Neuronal Differentiation Involves Direct Regulation of Hes1.

The significant down-regulation of Hes1 revealed by the RNA-seq analysis, along with its robust induction in response to Dmrta2 transgene expression in NPCs, suggest that Dmrta2 may maintain cortical NPC status via regulation of Hes1 transcription. To determine whether this is mediated by direct binding of Dmrta2, we performed chromatin immunoprecipitation (ChIP) on day 8 Dmrta2^flox/flox NPCs, with Dmrta2^−/− NPCs serving as a negative control. Based on the published consensus binding sequence for Dmrta2, we identified three potential binding sites (Bs1–3) at the Hes1 locus (Fig. 7A) (31). Dmrta2-immunoprecipitation of DNA fragments at each of these sites was quantified relative to a nonbound control region (NBCR) by qPCR (Fig. 7B). An enrichment in Dmrta2-bound fragments was identified at binding sites Bs1 and Bs2 (P < 0.05), but not at Bs3, in Dmrta2^flox/flox cells. In contrast, no enrichment was detected at any of the three potential binding sites in Dmrta2^−/− control cells. Thus, Dmrta2 binds to the Hes1 gene in NPCs.

Fig. 7.

Hes1 is a direct target for Dmrta2 transcriptional regulation in NPCs. (A) Schematic representation of the Hes1 genomic locus showing the relative positions of predicted Dmrta2 binding sites (Bs1–Bs3) and nonbinding control region (NCBR), and the primer pairs used to amplify each region after ChIP. (B) ChIP-qPCR for each of the regions depicted in A using chromatin prepared from Dmrta2^flox/flox and Dmrta2^−/− NPCs on day 8 of differentiation. Data are presented as mean ± SEM fold enrichment relative to the NBCR of three immunoprecipitations, each prepared from an independent differentiation experiment. *P < 0.05, one-tailed Student’s t test. (C) Reporter assay performed in Dmrta2^flox/flox and Dmrta2^−/− NPCs on day 8 of differentiation using WT or mutant Hes1 promoter-luciferase vectors carrying a point mutation at Bs1 or Bs2, respectively. Data are presented as mean ± SEM of three independent transfections with reporter plasmids. **P < 0.01, one-way ANOVA with Tukey’s HSD post hoc test. (D) Dmrta2^flox/flox and Dmrta2^−/− NPCs were cotransfected with GFP and Hes1 expression vectors on day 6 of differentiation. Cultures were immunostained for Tubb3 at 48 h later. Successfully transfected cells overexpressing Hes1 were identified based on GFP expression. Data are presented as mean ± SEM of three transfections, each from an independent differentiation experiment. *P < 0.05, two-tailed Student’s t test. (Magnification: 10×.) (E) Monolayer cultures of Dmrta2-ESCs were exposed to doxycycline with or without nontargeting control or Hes1 siRNA from day 5. The proportions of Nestin⁺ NPCs and Tubb3⁺ neurons were quantified at day 12. Data are presented as mean ± SEM of >20 individual fields of view. **P < 0.05, one-way ANOVA with Tukey‘s HSD post hoc test.

To provide evidence for Dmrta2 regulatory activity at Bs1 and Bs2, we performed reporter assays using a Hes1 promoter-driven luciferase vector (32). A twofold higher basal Hes1 promoter activity level was recorded in the isogenic Dmrta2^flox/flox cultures relative to the Dmrta2^−/− NPCs on day 8 of differentiation (P < 0.001) (Fig. 7C). An A > C point mutation at position −3 of the binding motif has been shown to significantly impair the ability of Dmrta2 to bind to its target sequence (31); thus, we introduced A > C point mutations at Bs1 and Bs2 via site-directed mutagenesis to yield two mutated reporter constructs, pHes1-luc-Bs1g.2007A > C and pHes1-luc-Bs2g.2365A > C, respectively. The mutation at Bs1, but not that at Bs2, resulted in a significant reduction of elevated Hes1 promoter activity level in Dmrta2^flox/flox NPCs relative to the parental luciferase construct (P < 0.01), suggesting reduced binding of Dmrta2 to the mutated Bs1 (Fig. 7C). Taken together, these data strongly support the ability of Dmrta2 to bind to and regulate the transcriptional activity at Bs1 on the Hes1 genomic locus.

We next sought to determine the extent to which reduced Hes1 expression in Dmrta2^−/− NPCs may contribute to their altered neurogenesis. To this end, we transfected Dmrta2^flox/flox and Dmrta2^−/− NPCs with a Hes1 expression vector on day 6 of differentiation together with a GFP-coding plasmid to distinguish between Hes1-transfected and nontransfected cells (Fig. 7D). By quantifying Tubb3 staining at 48 h posttransfection, we identified significantly increased numbers of neurons in the nontransfected population of Dmrta2^−/− cells relative to nontransfected isogenic controls (P < 0.01). As predicted based on known Hes1 function, Hes1 transgene expression led to reduced neuronal production for each cell line. Interestingly, we found no significant differences between cell lines in the Hes1-transfected populations. Thus, Hes1 transgene expression leads to a rescue of the precocious neurogenesis associated with the loss of Dmrta2.

We then asked whether siRNA-mediated knockdown of Hes1 expression could attenuate the antineurogenic effect of Dmrta2 transgenic expression in cortical NPCs. Dmrta2 ESCs were treated with doxycycline and Hes1 siRNA or a control nontargeting siRNA from day 5 to day 12 of differentiation. Hes1 knockdown resulted in a significant reduction of Nestin⁺ NPCs (P < 0.05) and a concurrent increase in the proportion of Tubb3⁺ neuronal cells (Fig. 7E). Thus, Hes1 knockdown partially reverses Dmrta2-mediated suppression of neuronal differentiation.

Taken together, our data identify Hes1 as a downstream target of Dmrta2 transcriptional regulation and a mechanism through which Dmrta2 safeguards NPCs from premature differentiation.

Discussion

Dmrta2 in NPC Cell Cycle Regulation.

Loss-of-function mutations in Dmrta2 have been linked with microcephaly in zebrafish, mice, and humans (7, 10–13); however, the role and mechanism of action of Dmrta2 in the control of NPC maintenance and expansion have remained unknown until now. Recently, conditional Dmrta2 mutant mice (Dmrta2^fl/fl;Emx1-cre) were created that delete Dmrta2 in cortical progenitors after cortical hem formation without impacting Wnt signaling. The Dmrta2 cKO embryos also show reduced cortical hemisphere size, suggesting a direct role of Dmrta2 in the control of NPC behavior (14). Cells in a monolayer ESC-neural differentiation system are generally exposed to the same extracellular environment and do not form “signaling centers,” unlike those found in the developing brain. Moreover, daily changes of culture medium will reduce the impact of any secreted molecules that may elicit a secondary effect. Our pathway analysis of the RNA-seq data did not reveal any significant changes in Wnt signaling; thus, the observed effect of Dmrta2 on cell cycle changes is likely cell-autonomous.

In line with the accumulation of Dmrta2^−/− NPCs in the G0/G1 phase of the cell cycle, we found altered expression of various cell cycle regulatory genes, particularly those acting on the G1-to-S phase transition (Cdkn1b, Cdkn1c, and Btg2). Considered together, these data suggest a disruption in cell cycle progression and potential lengthening of the G1 phase in Dmrta2^−/− NPCs. During normal corticogenesis, the duration of the G1 phase is linked to neuronal differentiation and is always longer in cells committed to undergoing differentiative rather than proliferative division (33, 34). Furthermore, experimental lengthening of the G1 phase pharmacologically or by the induction of Cdkn1b or Cdkn1c expression promotes neuronal differentiation and depletion of the NPC pool, resulting in microcephaly (1, 33, 35). The observed increases of Cdkn1b and Cdkn1c in Dmrta2^−/− NPCs strongly implicate delayed cell cycle progression of Dmrta2^−/− NPCs as a cellular mechanism contributing to precocious neurogenesis. Whether this is achieved through the direct regulation of cell cycle progression genes by Dmrta2 is unclear, however. Although various G1-to-S transition regulatory molecules are known to act as downstream targets for Hes1 repression, Cdkn2c (p18ink4c) also has been identified as a candidate gene for direct Dmrta2-mediated regulation in the zebrafish testes (11, 36, 37). Our findings demonstrate that through a direct or indirect effect, Dmrta2 is intricately linked to the control of cell cycle progression, a feature conserved across species and tissues.

Along with their function in the regulation of cell cycle progression, cyclin-dependent kinase inhibitors, including Cdkn1b, directly influence and promote NPC differentiation by stabilizing protein levels of Neurog2 via direct binding and the regulation of both interkinetic and radial migration (35, 38, 39). Our transcriptome analysis of Dmrta2^−/− NPCs identified an enrichment of down-regulated transcripts associated with GO terms for cell adhesion (GO:0007155) and locomotion (GO:004001). Thus, it is possible that defective migration is a cellular phenotype associated with the loss of Dmrta2 that could potentially lead to increased neuronal differentiation. Similarly, our RNA-seq data allude to a potential switch in the mode of cell division of Dmrta2^−/− NPCs. The transcription factors Emx2 and Pax6 were found to be down-regulated and up-regulated, respectively, by Dmrta2^−/− NPCs, as well as in the brains of Dmrta2-null mice (12). Regarding its role in telencephalic patterning, Emx2 is known to promote symmetric proliferative division of NPCs and Pax6 asymmetric differentiative division (40, 41). Thus, a switch in the mode of proliferation to neurogenic divisions may be a further cellular factor contributing to increased differentiation in the absence of Dmrta2. This idea is supported by the strong up-regulation of mRNA transcripts for Btg2, which is expressed exclusively in cortical NPCs committed to undergoing neurogenic, but not proliferative, division (42). Thus, although we provide evidence that Dmrta2 safeguards NPCs from precocious neurogenesis via regulation of Hes1, other targets also may contribute to the fine control of neurogenesis by Dmrta2.

Dmrta2 Targets and Neurogenesis.

To date, Hes1 and Cdkn2c are the only two transcriptional targets that have been proposed for Dmrta2. Owing to high levels of conservation in DNA-binding motifs between Dmrt proteins, further insight may be provided by examining DNA-binding sites of related Dmrt family members in other tissues (31). Close to 1,400 direct binding sites for Dmrt1 in the mouse testis have been identified using ChIP-chip techniques (43). Many of these genes also were identified as dysregulated by our Dmrta2^−/− NPC transcriptome analysis, including Cdkn2c, Igf2r, Meis1, Hox family members, and other Dmrt genes. Although Dmrt1 is not expressed by cortical NPCs, our data suggest significant overlaps in the regulatory targets of different Dmrt proteins. This is of particular interest because of the similar expression patterns of Dmrt3, Dmrta1, and Dmrta2 in the dorsal telencephalon, suggesting a potential for functional redundancy (13, 24). A similar but less severe phenotype as that seen in Dmrta2 null mutants has been observed in mice with a Dmrt3 null mutation, further supporting the idea that the two factors have overlapping functions in cortical development (12-14). In contrast, Dmrta1-null mice produce viable offspring with no overt anatomic defects in the brain (13, 44). This implies a hierarchical structure of importance of the Dmrt proteins to cortical development. Similar to our transgenic Dmrta2 findings, forced expression of Dmrt3 or Dmrta1 in the rodent telencephalon is linked to the regulation of Neurog2 expression (24). Further studies may reveal the extent of this functional overlap among Dmrt family members in the dorsal telencephalon.

The functions of the Notch target gene Hes1 in maintaining NPC self-renewal have been well characterized (6). Notch ligands produced by newborn neurons activate notch signaling in neighboring cells, which in turn induces expression of Hes1 to repress the transcription of proneural factors and cell cycle progression regulators, thereby inhibiting neuronal differentiation (6, 36, 37). This lateral inhibition of spontaneous neuronal differentiation by neighboring cells favors NPC proliferation and self-renewal; however, Hes1 expression is dynamically regulated by various mechanisms, including a strong negative autoregulatory function, which is in turn inhibited by interactions with Id proteins (45, 46); the activity of other transcription factors, such as Lhx2 (47); and the activation of signaling pathways by growth factors and mitogens, including Fgf2 and Notch (6, 48). Under healthy conditions, these homeostatic mechanisms help maintain a NPC pool through development by regulating the expression of Hes1 and thus the balance between progenitor cell self-renewal and differentiation. This study has thus identified a previously unrecognized role for Dmrta2 in the dynamic regulation of Hes1 expression in cortical NPCs. By promoting Hes1, and thereby suppressing downstream proneural gene expression, Dmrta2 contributes to the maintenance of NPC self-renewal (Fig. 8A). In the absence of Dmrta2, Hes1 levels are reduced, leading to the up-regulation of proneural genes and increased neuronal differentiation (Fig. 8B).

Fig. 8.

Dmrta2 modulates neurogenesis through regulation of Hes1. Schematic model depicting the function of Dmrta2 in the modulation of NPC maintenance and neuronal differentiation. (A) High Dmrta2 expression in cortical NPCs ensures high expression of Hes1 and low levels of proneural genes, thereby promoting NPC maintenance and cortical expansion. (B) On differentiation, Dmrta2 expression declines, resulting in reduced levels of Hes1 and increased expression of proneural genes, thereby enhancing the differentiation of NPCs into postmitotic neurons.

In summary, we have identified Dmrta2 as a modulator controlling neuronal differentiation of cortical NPCs and have provided evidence that Dmrta2 exerts this function, at least in part, by direct transcriptional regulation of the neurogenesis inhibitor Hes1. Thus, this work points to another layer of control mechanisms coordinating NPC maintenance and neurogenesis, and begins to elucidate how Dmrta2 loss-of-function mutations may lead to microcephaly.

Materials and Methods

Cell Culture.

Six mouse ESC lines were used: a Dmrta2^flox/flox control line and two Dmrta2^−/− lines derived from the control line (SI Materials and Methods), two independent doxycycline-inducible Dmrta2-overexpressing ESC lines, and their parental lines harboring rtTA, as reported previously (9). All ESCs were maintained under standard conditions as described previously (17). For monolayer-based cortical differentiation, ESCs were seeded at 10,000 cells/cm² in gelatin-coated six-well plates and cultured in N2B27 medium. The differentiation medium was supplemented with 100 nM LDN193189 (Tocris) and 10 µM SB431542 (Tocris) from day 0 to day 4, 1 µM XAV939 (Tocris) from day 0 to day 6, and 1 µM cyclopamine (Sigma-Aldrich) between days 2 and 10. On day 5 or 6 of differentiation, NPCs were dissociated using trypsin/EDTA and then replated onto a poly-d-lysine/laminin–coated surface at a density of 50,000 cells/cm² for neuronal differentiation and maturation.

Transient transfections were performed using Lipofectamine 3000 reagent (Thermo Fisher Scientific). The following vectors were used: pHes1(2.5k)-luc (Addgene) (32), pGL4.73[hRluc/SV40] (Promega), pCAG-Hes1-IP (Lonza), and pmaxGFP (Lonza). Accell Hes1 and nontargeting control siRNA (GE Dharmacon) were used at a concentration of 1 µM.

Mouse Lines.

Emx1-cre and Dmrta2^fl/fl mouse lines were generated and maintained as described previously (14). The experiments were performed in compliance with the relevant laws and institutional guidelines and were approved by the local ethics committee (Universite Libre de Bruxelles).

qPCR.

Total RNA was extracted using TRI reagent treated with TURBO DNase (Thermo Fisher Scientific). cDNA was generated using the qScript cDNA Synthesis Kit (Thermo Fisher Scientific). qPCR was performed with MESA GREEN qPCR Master Mix (Eurogentec) with specific primers listed in Table S1, and dissociation curves were recorded to check for amplification specificity. C_q values were normalized to a minimum of two housekeeping reference genes, and changes in expression were calculated using the 2^-∆∆CT method (49). Three independent experiments were performed (n = 3), and each sample was assayed in duplicate on a CFX Connect Real Time PCR machine (BioRad).

Table S1.

Primers used in PCR and qPCR

Gene	Forward primer (5′→3′)	Reverse primer (5′→3′)	Size, bp
Dmrta2 genotyping - p1r	GGACCTGCCCCCTAACAAGAGTA	—	—
Dmrta2 genotyping - p3f	CGCGGCCTGCCTACGAAGTCTTTG	—	—
Dmrta2 genotyping - p4f	GGGGTGGGGGCGTACTTGTTTACAG	—	—
Cyclophillin	GGCAAATGCTGGACCAAACAC	TTCCTGGACCCAAAACGCTC	147
Dmrta2	GCCTGCCTACGAAGTCTTTGGCTCGGTTT	CGTCTTGGGAAACAGATCAAACTTCTGCAATTT	136
GAPDH	AGGTCGGTGTGAACGGATTTG	ACTGTGCCGTTGAATTTGCC	166
Hes1	TCGGTGGGTCCTAACGCAGT	ACGGGTAGCAGTGGCCTGAG	105
HMBS	ACTGGTGGAGTCTGGAGTCTAGATGGC	GCCAGGCTGATGCCCAGGTT	181
Map2	CTGGACATCAGCCTCACTCA	AATAGGTGCCCTGTGACCTG	163
Neurod1	CATGAGCGAGTCATGAGTGC	GCACAGTGGATTCGTTTCCC	92
Neurod4	AGGCCAATGCTAGAGAACGG	TCCTTGCCAGTCGAAGAGTC	129
Neurog2	GCTGTGGGAATTTCACCTGT	AAATTTCCACGCTTGCATTC	236
Cdkn1b	ACGCCAGACGTAAACAGCTC	GAGGCAGATGGTTTAAGAGTGC	192
Cdkn1c	GAGGAGCAGGACGAGAATCA	CACGTTTGGAGAGGGACACC	198
Eomes	GAGAAAGCGCCTGTCTCCCA	CCCATGCCTTTGGAGGTGTC	108
Tubb3	ATCAGCAAGGTGCGTGAGGAG	ATGGACAGGGTGGCGTTGTAG	113
ChIP - NBCR	AGGACATCAGGTTCTGTGCC	AGATTCCCCACAAGATTCCC	240
ChIP - Bs1	ACATACAGAGTTCGAGCGGG	TTCTCTGGGCTTTGCTTAG	235
ChIP - Bs2	CGTGTCTCTTCCTCCCATTG	ATTCCGCTGTTATCAGCACC	155
ChIP - Bs3	CCACCTCTCTCTTCTGACGG	AGGCGCAATCCAATATGAAC	184

Immunocytochemistry and EdU Labeling.

Cultures were fixed with 4% (wt/vol) paraformaldehyde and permeabilized with 0.1% (vol/vol) Triton X-100. Following blocking with 2% (wt/vol) BSA and 5% (vol/vol) donkey serum, cells were incubated with primary antibodies overnight at 4 °C, followed by incubation with complementary Alexa Fluor-conjugated antibodies and counterstaining with DAPI. All antibodies used are listed in SI Materials and Methods. To quantify proliferation, differentiating cultures were incubated with 5 µM EdU for 30 min before fixation. EdU detection was then carried out using the Click-iT EdU Alexa Fluor 488 Imaging Kit (Life Technologies). Images were subsequently acquired using an inverted microscope (DMI600b; Leica Microsystems). Manual cell counts were performed on a minimum of 10 randomly placed fields of view per stain, and the mean of three separate differentiations was calculated (n = 3).

Embryo sections (6–8 µm) were fixed overnight in 4% paraformaldehyde/PBS, dehydrated, paraffin-embedded, and then processed as described previously (14). For quantification of cells expressing Tubb3, cells of the entire dorsal telencephalon at the medial level were counted; at least two embryos of each genotype were analyzed, with a quantification of three to six sections per embryo.

Cell Cycle Analysis by Flow Cytometry.

NPCs were dissociated with EDTA, washed in PBS, and fixed with ice-cold 70% ethanol. After washing with 1% BSA, cell samples were incubated with mouse anti-Nestin (BD Biosciences; 611659, 1 µg/mL) or mouse IgG isotype (Sigma-Aldrich; I5381, 1 µg/mL) antibodies overnight at 4 °C. After incubation with donkey anti-mouse Alexa Fluor 647 secondary antibody (Thermo Fisher Scientific; A-31571, 1:1,000), DNA content was labeled by incubating cells with 1 µg/mL DAPI. Stained cells were analyzed on an Amnis Flowsight (Merck Millipore) under excitation from 405-nm and 642-nm lasers. IgG isotype control samples were used to set gating parameters for Nestin⁺ NPC and DAPI staining to identify cells at different stages of the cell cycle. Samples from three individual differentiation experiments were analyzed for each time point (n = 3).

RNA-seq.

RNA was extracted and purified using the PureLink RNA Mini Kit (Thermo Fisher Scientific). The TruSeq Stranded mRNA kit (Illumina) was used to prepare libraries from 1 µg of RNA from three independent differentiations (n = 3). The 75-bp paired-end sequencing was performed with the Illumina HiSeq 4000 sequencing system, yielding 30–45 million reads per sample. Reads were mapped to the mouse genome (mm10, GRCm38) using Burrows–Wheeler aligner algorithms (50), and individual gene read counts were calculated using featureCounts (51). DeSeq2 was used to calculate differential gene expression (52). GO functional enrichment for biological processes was performed using DAVID version 6.8 with the Mus musculus genome set as background (53). Calculated P values were adjusted for multiple testing using Benjamini–Hochberg correction. Raw sequence data files are publicly available from the National Center for Biotechnology Information’s Gene Expression Omnibus (accession no. GSE90827).

Luciferase Reporter Assays.

Point mutations were introduced into the pHes1(2.5k)-luc firefly luciferase vector using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies) and the following mutagenic primers (5′→3′): CAAGGTAAAGAGGATGTGTTCCTCTAATGTCTTCCGGAATT and AATTCCGGAAGACATTAGAGGAACACATCCTCTTTACCTTG for Bs1; and GAAAGTTCCTGTGGGAAAGAAAGTTTGGGAAGTTTCAC and CAAACTTTCTTTCCCACAGGAACTTTCAGCCAATGG for Bs2. Generation of mutagenized plasmids were confirmed by Sanger sequencing.

For luciferase reporter assays, cells were cultured in 24-well plates and cotransfected with 290 ng/well of firefly luciferase [pHes1(2.5k)-luc and its derivatives]. Renilla luciferase vector (pGL4.73; 10 ng/well) served as an internal control to normalize for transfection efficiency, and 50 ng of notch intracellular domain (NICD; pCAG-NotchIC-IP) served as a positive control. Cells were harvested at 24 h posttransfection and processed using the Dual-Glo Luciferase Assay System (Promega). Luciferase activity was measured with GloMax 96 Microplate Luminometer (Promega). Triplicate readings were taken for each sample, and all experiments were repeated with three biological replicates (n = 3).

ChIP.

Approximately 10⁷ cells on day 8 of differentiation were used for each immunoprecipitation. Protein and DNA were cross-linked with 1% formaldehyde before cell lysis. The extracted chromatin was subsequently sonicated at high power for 20 cycles of 30 s on/30 s off with a Bioruptor (Diagenode). Immunoprecipitation was performed by incubating chromatin with custom rabbit anti-Dmrta2 or rabbit IgG isotype control antibodies and salmon sperm DNA/Protein A agarose beads (Merck Millipore). Following denaturation of cross-links, Dmrta2-bound DNA fragments were purified using the Wizard SV Gel and PCR Clean-Up System (Promega). Immunoprecipitated DNA was subsequently amplified in qPCR reactions using the primers specified in Table S1. Three immunoprecipitations, each from a separate differentiation experiment, were performed (n = 3).

Statistical Analysis.

Statistical analyses were performed using IBM SPSS 20 software. Where specified, two-way ANOVA tests were performed, using Dmrta2 genotype status and day of differentiation as independent variables. Simple effects analysis by the post hoc Sidak test was used to correct for separate orthogonal comparisons between groups at each time point and to identify statistical significance. For luciferase assays and siRNA knockdown experiments, one-way ANOVA with Tukey’s honest significant difference (HSD) post hoc test were performed. One-tailed and two-tailed Student’s t tests were used to analyze ChIP-qPCR and Hes1 transfection rescue experiments, respectively.

SI Materials and Methods

Generation of Dmrta2^−/− and Dmrta2^flox/flox mESCs.

The Dmrta2 gene was targeted by homologous recombination using a Dmrta2 targeting construct containing 5′ and 3′ arms for homologous recombination, a CMVhygrotkpA cassette flanked by two flippase (Flp) recombination target (FRT) sites for selection and subsequent removal of the selection cassette by Flp, exon 2 (E2) flanked by two loxP sites for Cre-mediated conditional knockout, and a Lox2272 site for facilitating exchange of reporters to the Dmrta2 locus (Fig. S3A). Targeting was screened by Southern blot analysis for homologous recombination (Fig. S3B). To target the second Dmrta2 allele, the CMVhygrotkpA cassette was removed from the first targeted Dmrta2^mut allele by transfection with the Flp construct to result in Dmrta2^flox. The Dmrta2^flox allele was distinguished from WT Dmrta2 by PCR genotyping using forward primer Dmrta2-KO7-p3f and reverse primer Dmrta2-KO7-p1r (Fig. S3C and Table S1).

After targeting of the second Dmrta2 allele, Dmrta2^flox/mut mESCs were transfected with the Flp construct to remove CMVhygrotkpA cassette from the second allele to generate control Dmrta2^flox/flox mESCs. To generate Dmrta2 null ES cells (Dmrta2^−/−), Dmrta2^flox/mut cells were cotransfected with both Flp and Cre constructs to simultaneously remove the CMVhygrotkpA cassette.

ESC Differentiation.

To differentiate ESCs toward a ventral telencephalic fate, cells were seeded at 10,000 cells/cm² in gelatin-coated six-well plates and cultured in N2B27 medium. Differentiation medium was supplemented with 100 nM LDN193189, 10 µM SB431542 at day 0 to day 4, 1 µM XAV939 at day 0 to day 6, and 1 µg/mL SHH and 10 nM SAG (all Tocris) between days 2 and 10. To obtain a more caudal phenotype, ESCs were seeded at 10,000 cells/cm² in gelatin-coated six-well plates and cultured in N2B27 supplemented with 100 nM LDN193189 and 10 µM SB431542 at day 0 to day 4. Between days 2 and 6 of differentiation, N2B27 prepared with B27 supplement containing vitamin A was used.

ESCs were differentiated toward a ventral midbrain fate as described previously (23). In brief, ESCs were seeded at 2,500 cells/cm² in gelatin-coated six-well plates and cultured in N2B27 medium. Differentiation medium was supplemented with PD0325901 from day 2 to day 4 and 100 ng/mL SHH, 1 µM purmorphamine (all Tocris), and 100 ng/mL Fgf-8 between days 7 and 11. On day 5 of differentiation, NPCs were dissociated using trypsin/EDTA and then replated onto a poly-d-lysine/laminin–coated surface at a density of 50,000 cells/cm² for neuronal differentiation and maturation.

Primary Antibodies.

The following primary antibodies were used: rat anti-Bcl11b (Abcam; ab18465, 1:500), mouse anti-Coup-TF1 (Perseus Proteomics; PP-H8132-00, 1:200), custom rabbit anti-Dmrta2 (1:2,000), custom guinea pig anti-Dmrta2 (1:2,000), rabbit anti-Eomes (Abcam; ab23345, 1:500), mouse anti-FORSE-1 (Developmental Studies Hybridoma Bank; 1:500), goat anti-FoxA2 (Santa Cruz Biotechnology; sc6554, 1:100), mouse anti-Gad65/67 (Merck Millipore; Mab5406, 1:500), mouse anti-GFP (Thermo Fisher Scientific; A11120, 1:200), rabbit anti-Ki67 (Novocastro; NCL-Ki67p, 1:500), mouse anti-Lim1 (DSHB, 1:10), custom guinea pig anti-Lmx1a (1:1,000), rabbit anti-Musashi1 (Merck Millipore; AB5977, 1:200), chicken anti-Nestin (Neuromics; CH23001, 1:500), mouse anti-NeuN (Merck Millipore; MAB377, 1:400), rabbit anti-Nkx2.1 (Abcam; ab40880, 1:1,000), rabbit anti-Nurr1 (Santa Cruz Biotechnology; sc990, 1:200), goat anti-Olig2 (R&D Systems; AF2418, 1:200), goat anti-Otx2 (Santa Cruz Biotechnology; sc30659, 1:100), mouse anti-Pax6 (DSHB, 1:1,000), rabbit anti-p27^kip1 (Santa Cruz, sc528, 1:200), rabbit anti-p57^kip2 (Abcam; ab4058, 1:250), mouse anti-SatB2 (Abcam; ab51502, 1:10), rabbit anti-Tbr1 (Abcam; ab31940, 1:1,000), sheep anti-tyrosine hydroxylase (Pel-Freez; P60101/1, 1:500), rabbit anti-TUBB3 (Covance; PRB-435p, 1:1,000), rabbit anti-Tubb3 (Sigma-Aldrich; T2200, 1:200), and mouse anti-vGlut1 (Merck Millipore; MAB5502, 1:500).