The Mammalian DM Domain Transcription Factor Dmrta2 Is Required for Early Embryonic Development of the Cerebral Cortex

Development of the mammalian telencephalon is precisely organized by a combination of extracellular signaling events derived from signaling centers and transcription factor networks. Using gene expression profiling of the developing mouse dorsal telencephalon, we found that the DM domain transcription factor Dmrta2 (doublesex and mab-3-related transcription factor a2) is involved in the development of the dorsal telencephalon. Consistent with its medial-high/lateral-low expression pattern in the dorsal telencephalon, Dmrta2 null mutants demonstrated a dramatic reduction in medial cortical structures such as the cortical hem and the choroid plexus, and a complete loss of the hippocampus. In this mutant, the dorsal telencephalon also showed a remarkable size reduction, in addition to abnormal cell cycle kinetics and defective patterning. In contrast, a conditional Dmrta2 deletion in the telencephalon, which was accomplished after entry into the neurogenic phase, resulted in only a slight reduction in telencephalon size and normal patterning. We also found that Dmrta2 expression was decreased by a dominant-negative Tcf and was increased by a stabilized β-catenin form. These data suggest that Dmrta2 plays pivotal roles in the early development of the telencephalon via the formation of the cortical hem, a source of Wnts, and also in the maintenance of neural progenitors as a downstream of the Wnt pathway.


Introduction
During the development of the mammalian telencephalon, the regulation of temporal and spatial changes in characteristics of neural progenitors is fundamental for the growth control, regionalization, and layer formation of the cerebral cortex [1][2][3]. Signaling molecules secreted by signaling centers, such as fibroblast growth factors (FGFs) and Wingless-Int (Wnts) molecules, confer regional and temporal specificity to neural progenitors during early cortical development [4]. Intriguingly, these signals also modulate proliferation and differentiation of the neural progenitors in the telencephalon [5][6][7][8]; for this reason, the control of proliferation and regional specification appear to be tightly linked. Thus, to understand corticogenesis at a molecular level, it is essential to reveal the temporal and spatial regulation of the transcriptional network and its upstream signaling pathways controlled by signaling centers.
The anterior neural ridge (ANR) is a critical signaling center positioned at the anterior midline of the telencephalon. The ANR controls telencephalon formation by modulating rostrocaudal patterning through the secretion of FGFs during the early stages of cortical development [6,7,[9][10][11]. Furthermore, the cortical hem, another signaling center located in the medial edge of the dorsal telencephalon, regulates mediolateral patterning by the expression of multiple BMPs and Wnts [12,13] and functions as the organizing center for the development of the hippocampus [5,14,15]. Cortical hem-derived Wnt3a signaling regulates the neural progenitor proliferation in the medial part of the dorsal telencephalon [5]. This is presumably accomplished by changing the activity of its downstream nuclear Tcf/Lef effectors from transcriptional repressors to activators through the stabilization of b-catenin.
A number of studies using genetic models have shown that the spatially discrete expression of multiple TFs, including Coup-TFI, Pax6, Emx2, and Sp8, in the developing telencephalon appears to be important for the regional specification and proliferation of neural progenitors [1][2][3]. A remarkable feature of these TFs is that their expression pattern follows a distinct rostrocaudal and/or mediolateral gradient, suggesting that the expression of these molecules is tightly regulated by extracellular signaling from the ANR and cortical hem [1][2][3]. These studies have greatly increased our understanding of molecular mechanisms controlling cerebral cortical development; however, given the complexity of the temporal and regional regulation of corticogenesis, there seems to be several missing links in the TF network regulating cortical development.
In this study, we aimed to identify the molecules expressed in the developing telencephalon in a temporally restricted manner using gene expression profiling of neural progenitors from the dorsal telencephalon. We then focused on the function of the DM domain containing transcription factor Dmrta2 based on its unique expression pattern during cortical development. We thus investigated its role in the development of the dorsal telencephalon using a gene targeting strategy.

Animals
Embryonic stages were calculated by defining noon on the day of vaginal plug as embryonic day 0.5 (E0.5); the day of birth was defined as postnatal day 0 (P0). All animal manipulations were performed according to the guidelines for animal experiments at the RIKEN Center for Developmental Biology.

Microarray Analysis
Total RNA was extracted using the RNeasy MinElute kit (Qiagen, Tokyo, Japan). The cDNA synthesis and cRNA labeling reactions were performed according to the two-cycle protocol provided by Affymetrix. Affymetrix high-density oligonucleotide arrays for Mus musculus (Mouse Genome 430 2.0 Array) were hybridized, stained and washed, as described in the expression analysis technical manual (Affymetrix). The expression values were summarized using the RMA method. The resulting expression values were used in all subsequent analyses.

Plasmids
All mammalian expression vectors were based on pCAG [18]. The expression plasmid for NLS-tagged EGFP, pCAG-EGFP3NLS, was described previously [18]. The expression plasmids for dominant-negative tcf3 and constitutive-active b-Catenin, pCAG-DN-Tcf3 and pCAG-CA-bCat, were gifts from T. Iwano (CDB, RIKEN, Kobe, Japan). To construct the bacterial expression plasmids, cDNAs encoding the C-terminus of Dmrt3 and Dmrta2, as described below, were amplified and cloned into pGEX-6X (GE Healthcare, UK). To construct the expression plasmids for Flag-or HA-tagged Dmrta2, a cDNAs encoding the full-length Dmrta2 without the stop codon were amplified by PCR and then cloned into the pCAG-Flag-N1 plasmid [19]. All plasmids were confirmed using DNA sequencing.

Generation of Polyclonal and Monoclonal Antibodies
Polyclonal antibodies for Dmrt3 and Dmrta2 were generated by immunizing rabbits with glutathione S-transferase (GST)-fused Cterminal mouse Dmrt3 (aa 147-476) and mouse Dmrta2 (aa 129-531). The antisera were preabsorbed with GST protein, followed by purification using antigen coupled to a Sepharose 4B gel matrix.
A rat monoclonal antibody that specifically recognized Dmrta2 was generated based on the rat lymph node method [20]. A WKY/izm rat was immunized with GST-fused recombinant Dmrta2. After three weeks, rat lymphocytes were fused with mouse myeloma SP2/W cells. Hybridoma clones producing Dmrta2-specific antibody were screened by ELISA and immuno-staining. Finally, a hybridoma clone producing a monoclonal antibody (1D1) was selected.
Antibody specificity was confirmed by Western blotting ( Figure  S1).

Mutant Mice
To generate conditional mutants for Dmrta2, we constructed a targeting vector harboring the loxP-flanked exon 2 of Dmrta2 and a Frt-flanked neomycin resistant cassette (neo-cassette). The targeting vector was introduced into the TT2-derived mouse ES cells by electroporation as described previously [21]. After isolation of three homologous recombinants in the ES cells, the clones were introduced into 8-cell embryos of CD1 mice. To delete the neocassette, chimeric mice were mated with ACTB-FLPe mice (Stock No. 003800, The Jackson Laboratory, Maine, USA), which ubiquitously express Cre recombinase under the control of the human ACTB promoter. Two independent mutant mouse lines were analyzed to confirm the phenotype (Accession No. CDB1049K: http://www.cdb.riken.jp/arg/mutant%20mice% 20list.html). For experiments using Dmrta2 null mutants, we crossed Dmrta2 conditional mutants to EIIa-Cre transgenic mice (Stock No. 003724, The Jackson Laboratory). For the conditional deletion of Dmrta2 in the central nervous system, we crossed Dmrta2 conditional mutants with Nestin-Cre transgenic mice [22] and analyzed the phenotype of heterozygotic and homozygotic Dmrta2 mutants containing the hemizygous Nestin-Cre transgene. All mutant phenotypes were analyzed on a C57BL/6 mouse background. The genotypes of the resulting mutant mice were determined according to the size of their PCR products (wild-type: 319 bp, floxed: 516, deleted: 200 bp) using the described primers (Primer-1: GATCCTAGTGAACCTCTTCGAGGGAC, Primer-2: GAGTTGCAATTGCTGTACGGCACTG, Primer-3: GAG-CCACAGTTAAGTAGTTGGAG).
Dmrta2-heterozygous mice exhibited normal fertility and CNS development, and we therefore used the heterozygous mice as a control in this study.

Histology
Hematoxylin and Eosin (HE) staining was performed according to standard methods. Briefly, Mayer's Hematoxylin solution and 1% Eosin Y solution (Wako Pure Chemical Industries, Osaka, Japan) were sequentially used for nuclear and cytoplasmic staining, respectively.

DNA Labeling with EdU
Pregnant mice were injected intraperitoneally with 100 ml of EdU (10 mg/ml) (Invitrogen, Tokyo, Japan). Animals were euthanized at 30 min after injection to visualize the S-phase cells. The cell cycle exit rate was estimated as the ratio of Ki67negative/EdU-positive cells to all EdU-positive cells at E12.5 (18 h after EdU injection) or E15.5 (24 h after EdU injection).

Quantifications
To quantify the expression levels of Dmrt3 and Dmrta2, the immunostaining fluorescence intensity was evaluated using Olympus FV10-ASW v1.7 software. All statistical analyses were performed with the Student's t-test using Prism 5 or Excel software. The data were expressed as the mean6SD.

Identification of Young Neural Progenitor-enriched Genes
To identify the genes predominantly expressed in the neural progenitors of the dorsal telencephalon during the early stages of cortical development, we first used FACS to isolate progenitor cells from the dorsal telencephalon of Hes1-reporter mice at four developmental stages: the proliferative stage (E10.5), the early neurogenic stage (E12.5) and the later neurogenic stages (E15.5 and E18.5) ( Figure 1A). Hes1-reporter mice were shown to express destabilized EGFP in their neural progenitors under the control of the Hes1 promoter (pHes1-d2EGFP) [16,17]. Thus, the neural progenitors were easily purified using this method, with low levels of contamination from other cell types.
The gene expression profiles of the purified neural progenitors analyzed with DNA microarray demonstrated a steady expression level of Sox2, a neural progenitor marker, at all examined stages, as expected (Table S1). These data indicate that a pure population of neural progenitors was efficiently isolated using our strategy. We next found that 483 probe sets were abundantly expressed in the brain at E10.5 and E12.5, and that their expression gradually decreased in the later stages (Table S1 and data not shown).
Of these genes, we were most interested in the expression of three related genes, Dmrt3, Dmrta1 (Dmrt4), and Dmrta2 (Dmrt5), because of their putative functions as DNA-binding factors ( Figure 1B) and their unique expression pattern in the developing telencephalon. In situ hybridization revealed that these genes were expressed as early as E9.5 and were robustly expressed in the dorsal telencephalon at E10.5 and E12.5 ( Figure 1C). We found that Dmrt3 and Dmrta2 had medial-high to lateral-low gradient patterns of expression, while Dmrta1 did not show a graded pattern of expression ( Figure 1C). Taken together, these data suggest that Dmrt3 and Dmrta2 are involved in the spatiotemporal pattern of specification and/or in the differentiation of telencephalic neural progenitors.

Dmrt3 and Dmrta2 are Robustly Expressed in the Medial Telencephalon
We next investigated the expression of Dmrt proteins by immunostaining with specific antibodies ( Figure S1). Consistent with our in situ hybridization analysis, immunofluorescence staining for Dmrt3 and Dmrta2 revealed a graded expression pattern in the dorsal telencephalon at E10.5 and E12.5 ( Figure 2A). Both Dmrt3 and Dmrta2 proteins were detected in all Sox2positive neural progenitors in the midial high-lateral low manner in the dorsal telencephalon ( Figure 2B, top). Dmrt3 was also detected in Map2-positive cortical plate neurons, although its expression level was much lower in neurons than in neural progenitors ( Figure 2B, middle). In the medial cortex, Dmrt3 was strongly expressed in the cortical hem and its derivative, the choroid plexus, both of which are marked by Otx2 expression [23] ( Figure 2B, bottom). Dmrta2 was also strongly expressed in the cortical hem while its expression in the choroid plexus was faint.

Reduction of the Telencephalon in Dmrta2 Mutant Embryos
To examine the function of Dmrt3 and Dmrta2 in vivo, we first generated Dmrt3-knockout mice using KOMP (Knockout Mouse Project)-derived Dmrt3-knockout ES cells. Dmrt3 mutants showed a 16% decrease in the size of the neocortical area at postnatal day 0 (P0) compared with that of control brains (data not shown). However, we could not find abnormalities of proliferation and differentiation in neocortical neural progenitors at embryonic stages. We then focused on the function of Dmrta2 using Dmrta2knockout mice. Dmrta2-mutant mice were generated by flanking Dmrta2 exon 2, which includes the DM domain critical for DNAbinding, with loxP sequences ( Figure S2). LoxP-flanked (floxed)-Dmrta2 mutants (Dmrta2 flox/flox ) were crossed with an EIIa-cre transgenic mouse line, which expresses Cre recombinase ubiquitously, to generate Dmrta2-null mutants (Dmrta2 2/2 ).
To examine the effect of Dmrta2 deletion on the telencephalon, we analyzed the size and gross morphology of the brain at P0. In Dmrta2 2/2 mice, the size of the telencephalon appeared to decrease compared to that of the control brains, although the midbrain showed no changes in either morphology or size ( Figure 3A). Consistent with this observation, the size of the cerebral cortical area defined by HE staining in Dmrta2 2/2 mutants was reduced by approximately 40% compared with the control brains, as assessed in three different planes of the sections ( Figure 3B, C). Most evidently, the hippocampal structure almost disappeared in Dmrta2 2/2 mutants, and the formation of the entire medial structure was disorganized ( Figure 3B, D, E, F). The size of the olfactory bulb was also reduced in the mutant, while the layer formation of the major olfactory bulb appeared to be normal ( Figure 3F). Immunostaining for Reelin, which is a marker for mitral cells of the olfactory bulb in adulthood, confirmed that these cells were correctly organized in a single layer (data not shown). To more closely examine the neocortical development of Dmrta2 mutants, we performed immunofluorescence staining on P0 coronal brain sections for Tbr1, which is strongly expressed in the cortical layer VI and in the subplate neurons [24], Ctip2, which is highly expressed in a subset of corticospinal motor neurons [25], and Satb2, which is predominantly expressed in the callosal projection neurons [26,27]. Dmrta2 2/2 mice were found to have correct positioning of their Tbr1-, Ctip2-, and Satb2-positive cortical neurons in an inside-out manner, similar to that of the control brains ( Figure 3D). We note that the loss of Dmrta2 slightly affected the positioning of neurons as observed for Ctip2-positive neurons in the layer II-IV ( Figure 3D). This might be due to a massive loss of Cajal-Retzius cells, which play important roles in radial migration of cortical neurons, in the mutant telencephalon ( Figure S3E). In contrast to the nearly normal layering of cortical neurons, the number of such  (Fig. 3E). Remarkably, the number and density of the early born neurons per unit area of neocortex were significantly decreased in Dmrta2 2/2 brains. In contrast, the number of Satb2-positive late-born neurons was only slightly decreased while their density in the cortical layer increased ( Figure 3E). The production of cortical layer specific neurons is tightly controlled by temporal identities and/or cell cycle kinetics in neocortical neural progenitors, which in turn depends on the developmental stages of the brain [28]. For this reason, our observations suggest that Dmrta2 predominantly functions to maintain the early stagespecific features of the progenitors during cerebral cortical development. This finding is consistent with the observed temporal pattern of Dmrta2 expression. Next, we investigated the details of Dmrta2 function in the early embryonic development of the cerebral cortex.

Severe Disorganization of the Medial Cortex in Dmrta2 Mutants
One of the early critical steps during telencephalic development is the formation of signaling centers, such as the cortical hem and the anterior neural ridge [4].
The nearly complete loss of the hippocampus in Dmrta2 2/2 brains raises the possibility that Dmrta2 is involved in the formation of the cortical hem, a signaling center located in the medial edge of the telencephalon. Indeed, the cortical hemderived signal is known to regulate the growth of the hippocampal primordium [5]. Therefore, we investigated whether Dmrta2 deletion affects the formation of the hem at E10.5 and E12.5 by in situ hybridization for Wnt3a, a marker for the cortical hem. Although we did not observe any changes in body size and gross morphology in Dmrta2 2/2 embryos ( Figure 4A), Wnt3a-positive cortical hem region was dramatically reduced by E10.5, and only small portion was detected at E12.5 ( Figure 4B). Consistent with a previous report describing the involvement of the cortical hem in choroid plexus formation [14,29], the expression of transthyretin (Ttr), a marker for the choroid plexus, was also significantly reduced in Dmrta2 2/2 mutants ( Figure 4B). To examine the role of Dmrta2 in the formation of signaling centers other than the cortical hem, we checked the expression of Fgf8, a mediator of brain patterning expressed in the anterior midline of the telencephalon, but found no change in its expression (data not shown). We further confirmed the sustained disorganization of the medial structures of the telencephalon by immunostaining for regional markers; the expression of Otx2, which is predominantly expressed in the cortical hem and choroid plexus, was observed only in a very  small population of cells compared to the control brain ( Figure  S3). Prox1, which is expressed in the hippocampal primordium under the control of the Wnt signaling during embryonic brain development [30], almost disappeared in Dmrta2 2/2 mutants at E15.5 ( Figure S3). These data strongly suggest that cortical hem function is severely compromised in Dmrta2 2/2 mutants. In addition, Dmrta2 2/2 brains lost their lateral-high to medial-low gradient of Pax6 expression, which plays a critical role in cell fate determination and cell cycle progression in the mammalian CNS ( Figure 4C). Control brains showed an expression of Pax6 in the dorsal telencephalon with the gradient, whereas Dmrta2 2/ 2 mutant brains showed a uniform expression of Pax6 expression in the entire dorsal neuroepithelium ( Figure 4C). Taken together, these observations indicate that the genetic ablation of Dmrta2 causes the hypoplasia or loss of the medial structures in the telencephalic cortex, thus implying that Dmrta2 plays a pivotal role in the regulation of neocortical patterning and in the formation of the signaling center.

Altered Cell Cycle Kinetics in Dmrta2 Mutants
As shown above, Dmrta2 mutants showed severe defects in the formation of the medial cortex in addition to dramatic changes in the size of the telencephalon at P0 and P14 (Figure 3). These results raise the possibility that the cell cycle kinetics of neural progenitors is altered in Dmrta2 2/2 mutants, such that a smaller number of neural progenitors is formed or maintained during early cortical development. Consistent with this, Dmrta2 2/2 mutants showed decreased size of the telencephalon at E12.5 ( Figure 5A). The number of progenitor cells was smaller in the dorsal telencephalon of Dmrta2 2/2 mutants than that of controls based on the dimension of the dorsal telencephalon (the length of the ventricular zone) ( Figure 5B) and the density of Sox2-positive progenitors (the number of these cells in a unit area) ( Figure 5C). This difference in the number of the progenitors appeared to occur between E10.5 and E12.5 because no obvious difference was detected in the size of the telencephalon between Dmrta2 homozygotes and heterozygotes at E10.5 (data not shown). We next examined the cell cycle kinetics at E12.5 by checking the mitotic index, which we defined as the percentage of phosphorylated histone H3 (pH3)-positive cells in the Sox2-positive progenitors, and the number of S-phase cells in the dorsolateral area of the mutant brains. We found that the mitotic index was slightly but significantly decreased in Dmrta2 2/2 brains {4.9160.67 (hetero) vs. 4.0060.47 (homo)} ( Figure 5D). The number of S-phase cells, labeled by 30 min in vivo incorporation of EdU, also decreased by 68.5% compared to the control ( Figure 5E). We also examined the number of apoptotic cells and the cell cycle exit rate, and we found that the number of TUNEL-positive apoptotic cells and the cell cycle exit rate were not significantly different between control and mutant tissues ( Figure 5F, G, and data not shown), suggesting that the defective increase in the progenitor population in Dmrta2 mutants was unlikely to originate from the apoptosis of progenitors or from premature neurogenesis (see discussion). Taken together, these data suggest that alteration of cell cycle kinetics in progenitors during early brain development is at least partly involved in the dramatic size reduction observed in the dorsal telencephalon of Dmrta2 2/2 brains.

Conditional Ablation of Dmrta2 in Neurogenic Stages Causes a Moderate Reduction in Brain Size
To further clarify the predominant role of Dmrta2 in the early development of the cerebral cortex, we conditionally ablated Dmrta2 using the Nestin-Cre transgenic line. This mouse line is useful to investigate the function of genes after the initiation of the neurogenic phase of brain development because Cre recombinase in this line is not fully active in the dorsal telencephalon until E12.5 [22,31]. In Dmrta2-conditional mutants (cKO) at E12.5, Dmrta2 expression was nearly undetectable in the lateral cortex. However, many cells continued to express Dmrta2 at relatively low levels compared to those of Dmrta2 expression in the medial cortex of the control brains. This suggests that Dmrta2 ablation is completed around E12.5 ( Figure 6A). Dmrta2-cKO brains exhibited normal formation of the cortical hem ( Figure 6B) and the hippocampal primordium in the medial cortex ( Figure 6C), indicating that development of the cortical hem and the subsequent specification of the hippocampal primordium were normal. Consistent with the correct formation of the medial structures, the layering of cortical neurons was normal as revealed by immunostaining for Pou3f2 (Brn2), which is predominantly expressed in the upper layer neurons, and Tbr1 at E15.5 ( Figure 6D). In contrast to the significant size reduction observed in Dmrta2 2/2 telencephalon, no size reduction was observed at E12.5, and only a slight reduction of the ventricular surface dimension was observed at E15.5 in brains with the conditional ablation of Dmrta2 ( Figure 6E, data not shown). We also examined the cell cycle exit rate in Dmrta2-cKO brains, however, no significant change was observed (data not shown). Taken together, these results confirm that Dmrta2 predominantly contributes to early events of telencephalic development, including the formation of the cortical hem.

Canonical Wnt Signaling Regulates the Expression of Dmrt3 and Dmrta2
The requirement of a Dmrta2 gradient for the development of medial structures raised the question of what regulates the expression of Dmrta2. The canonical Wnt signaling pathway is one possible candidate for this role, given its involvement in the maintenance of neural progenitors in the medial cortex [5]. To assess the possible role of Wnt signaling in Dmrta2 expression, we examined the overexpression of a dominant-negative form of Tcf3 (DN-Tcf3), which lacks the b-catenin-interacting domain and thus inhibits Tcf/Lef-dependent transcription. Here, we also examined the expression of Dmrt3 to investigate whether its expression is under the control of the Wnt signaling pathway, given that it shares a similar expression profile with Dmrta2 during brain development. At 24 h after in utero electroporation, the percentage of Dmrta2-positive cells of the total EGFP-positive cells did not change significantly {84.562.3 (Control) vs. 87.865.3 (DN-Tcf3), n = 3, p = 0.38}. However, their expression in EGFP-positive electroporated cells significantly decreased compared to that of the controls ( Figure 7A, B). Furthermore, the overexpression of a constitutive-active form of b-catenin, which lacks the N-terminal phosphorylation domain essential for its degradation, elevated the expression of Dmrt3 and Dmrta2 in the lateral region of the ventricular zone ( Figure 7C, D). Thus, the expression of Dmrt3 and Dmrta2 in the medial cortex, including the cortical hem, is at least partly regulated through canonical Wnt signaling.

Discussion
Here, we report a critical role for the DM domain-containing protein Dmrta2 in the early embryonic development of the telencephalon. Dmrta2 mutants showed a dramatic reduction in the size of the dorsal telencephalon and a reduction in the total number of Sox2-positive neural progenitors. These defects suggest malfunctions of neural progenitors in Dmrta2 mutant telencephalon, which might include slow cell cycle progression, precocious neurogenesis, or increased apoptosis. However, the number of TUNEL-positive cells and cell cycle exit rate did not change significantly during the period of the phenotypic emergence in Dmrta2 mutants, suggesting that the latter two possibilities are unlikely. In contrast, the cell cycle kinetics of the neural progenitors in Dmrta2 mutants appeared to change, suggesting aberrant cell cycle progression. Thus, the regulation of cell cycle progression is most likely a predominant function of Dmrta2 in the neural progenitors.
As illustrated by the complete loss of the hippocampus and an extreme shrinkage of the cortical hem, the medial structure defects in Dmrta2-mutant brains were considerably more severe than those in the dorsolateral telencephalon. These malformations might be explained by the loss of the cortical hem, which acts as a signaling center by secreting Wnts [32]. The contribution of Dmrta2 to corticogenesis thus appears to depend on medio-lateral patterning of the cortex. Furthermore, in the absence of Dmrta2 function, the medio-lateral gradient of Pax6, which is a key patterning factor in cerebral cortical development, was lost in the dorsal cortex. These observations raise the possibility that Dmrta2 is involved in the patterning during telencephalon development. In this study, Wnt signaling was shown to partly control Dmrta2 expression, suggesting that Dmrta2 functions as a downstream of Wnts in coupling of cell proliferation with cortical patterning. Further studies on the functional relationship between Dmrta2 and the other patterning factors might help to reveal a transcriptional network that will tightly link cell proliferation and patterning in the neural progenitor.
The Wnt-dependent expression of Dmrt3 and Dmrta2 is partly consistent with the results of a recent report showing that a Tcfbinding site in a Dmrt3 enhancer element is necessary for the expression of Dmrt3, although Dmrta2 expression was not shown to have a clear graded expression in the telencephalon [33]. Our in situ hybridization data showed that Dmrta2 is expressed in a medial-high to lateral-low gradient during cortical development, although its gradient is less prominent than that of Dmrt3. However, expression of both proteins clearly showed very similar expression pattern of medial-high to lateral-low gradients during cortical development, suggesting that the expression of these two factors are regulated by same signaling pathway. In addition, we observed the residual uniform expression of both Dmrt3 and Dmrta2 in the Emx1/2 double mutant brains, in which the formation of the cortical hem and Wnt signaling are severely disorganized [23] (D.K. and F.M., unpublished data). These data suggest that the expression of Dmrt3 and Dmrta2 might be The vertebrate Dmrt gene family proteins were first identified by their homology with Drosophila melanogaster Dsx (Double sex) and Caenorhabditis elegans mab-3 (male abnormal-3) [34][35][36]. Of the seven mammalian Dmrt proteins, three (Dmrt3, Dmrta1, and Dmrta2) share a highly conserved protein domain of unknown function, called the DMA domain (Dmrta subfamily). Previous reports describe that the Dmrta gene subfamily functions in the nervous system of vertebrates and invertebrates [37][38][39][40]. In particular, the Dmrta proteins promote neurogenesis by regulating the expression of bHLH transcription factors in Xenopus and Zebrafish CNS development [37,40]. However, we observed no significant increases in the expression of Neurogenin2, a proneural factor that promotes neurogenesis, nor did we identify any changes in the neurogenic properties of Dmrta2 mutants (Figure 4, data not shown). This discrepancy might be explained by the different roles of Dmrt proteins in different biological contexts, and/or by the redundant roles of Dmrt with other proteins.
It is unclear if Dmrta2 maintains the cerebral cortical neural progenitors in a cell-autonomous manner. In this study, we could not address whether Dmrta2 is directly involved in the maintenance of the neocortical progenitors beyond its indirect contributions to cortical hem formation. However, the defects in Dmrta2 mutant brains appeared to be much more severe in terms of size and morphology than those observed in the cortical hem-ablated mutants generated through the conditional expression of diphtheria toxin-A in the Wnt3a-expressing cells [14], suggesting that the Dmrta2 mutant phenotype reflects the combination of cellautonomous and non-cell-autonomous Dmrta2 functions in the Table S1 Microarray analysis of the developing mouse telencephalon. The expression values of twenty genes that are highly expressed in the E10.5 brains and Sox2. (XLS)