Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2007 Sep 18;104(39):15382–15387. doi: 10.1073/pnas.0707456104

The E-protein Tcf4 interacts with Math1 to regulate differentiation of a specific subset of neuronal progenitors

Adriano Flora *,†,‡, Jesus J Garcia *,†, Christina Thaller §, Huda Y Zoghbi *,†,¶,‖,**,‡
PMCID: PMC1978485  PMID: 17878293

Abstract

Proneural factors represent <10 transcriptional regulators required for specifying all of the different neurons of the mammalian nervous system. The mechanisms by which such a small number of factors creates this diversity are still unknown. We propose that proteins interacting with proneural factors confer such specificity. To test this hypothesis we isolated proteins that interact with Math1, a proneural transcription factor essential for the establishment of a neural progenitor population (rhombic lip) that gives rise to multiple hindbrain structures and identified the E-protein Tcf4. Interestingly, haploinsufficiency of TCF4 causes the Pitt–Hopkins mental retardation syndrome, underscoring the important role for this protein in neural development. To investigate the functional relevance of the Math1/Tcf4 interaction in vivo, we studied Tcf4−/− mice and found that they have disrupted pontine nucleus development. Surprisingly, this selective deficit occurs without affecting other rhombic lip-derived nuclei, despite expression of Math1 and Tcf4 throughout the rhombic lip. Importantly, deletion of any of the other E-protein-encoding genes does not have detectable effects on Math1-dependent neurons, suggesting a specialized role for Tcf4 in distinct neural progenitors. Our findings provide the first in vivo evidence for an exclusive function of dimers formed between a proneural basic helix–loop–helix factor and a specific E-protein, offering insight about the mechanisms underlying transcriptional programs that regulate development of the mammalian nervous system.

Keywords: basic helix–loop–helix, mental retardation, neural development, proneural

The establishment of neural progenitor territories in mammals relies on the expression of <10 basic helix–loop–helix (bHLH) transcription factors, collectively called proneural proteins (1–3). The mechanism by which a few transcriptional regulators control the differentiation of hundreds of distinct neuronal subtypes remains unclear (4). Distinct neuronal populations can originate from cells that express a single proneural gene, suggesting that progenitors with similar expression profiles are highly plastic and can adopt different fates (5–7). The heterogeneity of neurons derived from the same progenitor population begs the question of how a single proneural factor activates diverse genetic programs to regulate the differentiation of distinct neurons. Transcription factors typically interact with different cofactors to form distinct functional complexes (8). However, only a few cofactors interacting with neural bHLHs have been identified to date (9–11).

Math1 (also called Atoh1) is a bHLH transcription factor that is highly expressed in proliferating progenitors of the rhombic lip (RL), a specialized neuroepithelium in the developing dorsal hindbrain (12, 13). Math1 is required for the differentiation of most cerebellar and precerebellar structures, including the pontine, lateral reticular, and external cuneate nuclei (7, 14). Moreover, Math1−/− mice lack most of the cochlear nucleus and all cerebellar granule neurons. Like other proneural proteins, Math1 dimerizes with the widely expressed bHLH E-proteins to bind DNA and activate transcription (3). There are four E-proteins in mammals: Tcf4 (also called E2.2 or ITF2), E12, E47, and HEB, encoded by three genes, E12 and E47 being different splice isoforms of E2a (3, 15–17). Because their bHLH domains are highly homologous, it has been proposed that E-proteins have redundant functions, despite obvious phenotypes in animals lacking any one of them (18, 19). Although all E-proteins are highly expressed in neural progenitors (20, 21), their role in nervous system development has not been investigated in depth (22). Recently, genetic studies demonstrated that loss of one copy of TCF4 causes Pitt–Hopkins syndrome (PHS) (23–25), a neurodevelopmental disease characterized by mental retardation, seizures, and hyperventilation (26, 27), suggesting that TCF4 is critical for human nervous system development.

In this study we report the identification of Tcf4 as an interactor of Math1, and we demonstrate through genetic studies that Math1/Tcf4 heterodimers have an exclusive role in pontine neuron differentiation, thus providing the first evidence that a specific proneural bHLH/E-protein heterodimer plays a unique function in the developing nervous system.

Results

Math1 and Tcf4 Form Transcriptionally Active Heterodimers.

To identify Math1 interactors, we designed a yeast two-hybrid strategy using the bHLH domain because of its fundamental role (28, 29). However, the Math1 bHLH-Gal4 DNA-binding domain fusion protein proved to be unsuitable as bait because it was a strong transactivator in yeast (data not shown). To overcome this problem, we used the highly conserved bHLH domain of Drosophila atonal [supporting information (SI) Fig. 7], the orthologue known to rescue phenotypes of Math1-null mice (30), to screen a mouse embryonic cDNA library.

After two rounds of selection, we isolated 73 colonies that represented six different proteins. Seven colonies corresponded to three independent clones of Tcf4; one of these, comprising the entire coding sequence, was used to confirm the interaction (Fig. 1A). Reciprocal coimmunoprecipitation of the two proteins transiently expressed in NIH 3T3 cells confirmed the interaction in mammalian cells (Fig. 1B).

Fig. 1.

Fig. 1.

Open in a new tab

Identification of Tcf4 as a functional partner of Math1. (A) A β-galactosidase assay confirms the interaction between atonal and Tcf4 in yeast. Positive (p53+ T antigen) and negative (p53 or the atonal bait, Ato) controls are shown. (B) Coimmunoprecipitation of Math1 and Tcf4 in NIH 3T3 cells. Samples subjected to immunoprecipitation are shown on top of the gel; antibodies are indicated at the bottom. Anti-Flag and anti-V5 antibody detected tagged Math1 and Tcf4 proteins, respectively (arrows). Molecular markers are shown. (C) Autoradiography of a nondenaturing acrylamide gel used for EMSA; transfected constructs are indicated at the bottom. The top indicates which samples were incubated with antibodies against tagged Math1 and Tcf4 proteins or with unlabeled oligonucleotide. The arrows indicate the complexes formed between nuclear protein extracts and labeled oligonucleotide. Lanes are numbered at the bottom. (D) Bar diagram shows luciferase data. The expression constructs transfected in NIH 3T3 cells are indicated at the bottom. Bars show the fold induction over luciferase activity from cells transfected with empty vector pCDNA3. The standard deviation is shown.

To determine whether Tcf4 can heterodimerize with Math1 to form a functional DNA-binding complex, we performed EMSA using an oligonucleotide that contains two variants of the canonical E-box consensus sequence, previously demonstrated to interact with Math1 (12). Incubation of nuclear extracts from GFP-transfected NIH 3T3 cells with the radioactive probe revealed a very faint band, probably due to binding of endogenous bHLH proteins to the probe (Fig. 1C, lanes 2–4, band a). Transfection of a vector expressing Tcf4 alone did not alter the mobility of the oligonucleotide (Fig. 1C, lanes 6–8). Incubation of the probe with nuclear extracts of Math1-transfected cells (Fig. 1C, lanes 9–11) caused the formation of a new specific DNA–protein complex (band c), confirmed by competition with cold oligonucleotide and antibody supershifting (Fig. 1C, lanes 10 and 11, band d). Because Math1 homodimers do not bind to DNA (12), the complex observed in lane 9 likely results from the dimerization of Math1 with endogenous E-proteins. When we incubated the probe with nuclear extracts from cells transfected with both Tcf4 and Math1, a strong band appeared that was supershifted by antibodies against either Math1- or Tcf4-tagged proteins (Fig. 1C, lanes 12, 14, and 15, bands c and d). These data show that Math1/Tcf4 heterodimers bind the DNA target sequence. We also observed a faster-migrating complex (Fig. 1C, lane 12, band b), probably due to binding of a Math1/Tcf4 dimer on one of the two E-box sequences of the probe. The specificity of these interactions was verified by competition with unlabeled oligonucleotide (Fig. 1C, lane 13).

To determine whether the Math1/Tcf4 heterodimers are transcriptionally active, we used a reporter plasmid bearing a minimal promoter flanked by seven E-box sequences cloned in front of the luciferase reporter gene (12). As expected, transfection of Math1 in NIH 3T3 cells induced the luciferase activity (Fig. 1D), probably by interacting with endogenous E-proteins. Transfection of Tcf4 by itself did not alter the expression of the reporter gene. However, coexpression of Tcf4 and Math1 increased the Math1-dependent activation of the reporter gene. Together, these data indicate that Math1 and Tcf4 interact and form transcriptionally competent heterodimers.

Math1 and Tcf4 Are Coexpressed in the RL.

To investigate the physiological relevance of the Math1/Tcf4 interaction, we sought to determine whether the two genes are coexpressed during mouse development. Because the expression of Math1 in the embryonic brain is restricted to the RL (a population of neuronal progenitors located in the dorsal aspect of the developing hindbrain) we verified that Tcf4 is expressed in the RL by in situ hybridization of sections from mouse embryos at embryonic day 14.5 (RL in Fig. 2B and SI Fig. 8 and data not shown). The expression profile of Tcf4 overlaps with and extends beyond that of Math1 (SI Fig. 8 A, A′, and Inset). Both genes are expressed in the rostral, or upper, RL (uRL in Fig. 2) and in the caudal, or lower, RL (lRL in Fig. 2). Sense probes did not show any signal (data not shown). The temporal and spatial coexpression of Math1 and Tcf4 suggests that Tcf4 may play a role in the differentiation of RL neural progenitors.

Fig. 2.

Fig. 2.

Open in a new tab

In situ hybridization of sequential sagittal sections of a mouse embryo at embryonic day 14.5. The antisense probes are shown on the right. The section shown in A has been hybridized with the Math1 antisense probe, and the section shown in B has been hybridized with the Tcf4 antisense probe. uRL, upper RL; lRL, lower RL; Cb, cerebellum.

Tcf4 Is Required for the Differentiation of a Specific Subset of Math1-Expressing Progenitors.

To analyze the hindbrain development in Tcf4 mutant mice, we injected embryonic stem cells bearing a null Tcf4 allele (19) into mouse blastocysts. Tcf4 heterozygous animals are viable and fertile and show no obvious phenotype. Tcf4-null mice die in the first 24 h after birth, as previously reported. However, in contrast to previous observations (19), we did not observe embryonic lethality of null animals, and we recovered the expected Mendelian ratio of Tcf4−/− mice at birth. The brains of Tcf4-null animals were grossly normal and did not show any evident morphological defect (data not shown). We crossed Tcf4+/− mice to Math1+/lacZ mice to study the differentiation of RL progenitors in Tcf4 mutant animals. In Math1+/lacZ mice, the coding region of one allele of Math1 has been replaced with the lacZ gene, leading to the expression of β-galactosidase in cells that would normally express Math1, allowing the visualization of Math1-expressing progenitors and structures derived from them (7, 31).

Two main groups of neuronal structures are derived from the RL: dorsal structures, such as the granule cell layer of the cerebellum (Fig. 3A, Cb) and the external cuneate nucleus (Fig. 3A, ECN), and ventral structures, such as the pontine nucleus and the lateral reticulate nucleus (Fig. 3C, PN and LRt) (7, 32). Surprisingly, despite the expression of Tcf4 in the entire RL, its deletion caused a differentiation defect only in the pontine nucleus, leaving the other Math1-dependent structures intact (Figs. 3 and 4). This defect resulted in a substantial reduction in the neurons of the pontine nucleus (compare Fig. 3 C and C″, arrowheads) as well as ectopic accumulation of β-galactosidase-expressing cells in a dorsal–lateral region of the hindbrain (compare Fig. 3 B and B″ with C and C″, arrows), likely because of aberrant migration of Math1-dependent neuronal progenitors. Math1/Tcf4 doubly heterozygous mice showed a persistent anterior extramural migratory stream (AES), suggesting a delay or interruption in migration (compare Fig. 3 B and B′ with C and C′, arrows).

Fig. 3.

Fig. 3.

Open in a new tab

β-Galactosidase staining of structures derived from the RL of Math1+/lacZ mice. The pictures show dorsal (A–A″), lateral (B–B″), and ventral (C–C″) views of P0 hindbrains. The arrows point to the AES, and the arrowhead points to the pontine nucleus (PN). ECN, external cuneate nucleus; Cb, cerebellum; LRt, lateral reticulate nucleus. The genotype of the animals is shown at the top.

Fig. 4.

Fig. 4.

Open in a new tab

Comparison of the phenotypes of Math1−/− and Tcf4−/− mice. The outline of the sections is shown on the left of each row of images, and the framed portions represent the regions shown in the pictures. Arrowheads in A and B indicate the lateral reticulate (LRt) and external cuneate (ECN) nuclei, respectively. The arrowheads in C indicate the external granule layer cells (EGL), and the arrow points to the lower RL (lRL). In D, the arrowheads indicate the pontine nucleus (PN), and the arrows show the location of the AES. Pax6 staining is revealed as a red signal over the blue color of the nuclear stain TOTO3. The structures missing in Math1−/− animals are indicated between parentheses. The genotypes of the animals are shown at the top.

To confirm that the differentiation defect observed in Tcf4−/− mice is limited to pontine neurons, we compared the phenotype of Tcf4−/− animals to that of Math1−/− mice, which lack all of the structures derived from Math1-expressing progenitors. To evaluate the differentiation of the RL progenitors, we analyzed the expression pattern of Pax6, a homeodomain transcription factor expressed by all Math1-expressing progenitors in the hindbrain, and whose expression is maintained in the migratory streams that originate in the RL (32). Analysis of Pax6 staining confirmed that the pontine nucleus is the only structure disrupted by deletion of Tcf4, whereas the other Math1-dependent structures appear to develop normally (Fig. 4).

Math1 and Tcf4 Interact Genetically.

To verify that the differentiation defects in the Tcf4−/− animals are due to the absence of Math1/Tcf4 heterodimers rather than the lack of a Math1-independent function of Tcf4 in the RL, we tested the genetic interaction between Math1 and Tcf4. Nissl staining of coronal sections from brains at postnatal day 0 (P0) revealed very few cells in the region immediately lateral to the caudal pontine nucleus in wild-type animals, demonstrating completion of migration of pontine nucleus neurons (Fig. 5A). As expected, sections from Math1+/lacZ animals did not differ from those of wild-type mice. Similarly, Tcf4+/− mice had very few cells in this region, suggesting that a single copy of Tcf4 in an otherwise wild-type background is sufficient for the normal migration of the AES cells (Fig. 5D). On the other hand, Math1+/lacZ;Tcf4+/− animals showed clear accumulation of cells in the region lateral to the pontine nucleus, a phenotype consistent with the β-galactosidase staining data (compare Fig. 5 E and B, arrows). Analysis of Tcf4-null animals revealed a massive accumulation of neuronal precursors that were unable to reach the pontine nucleus region as well as a drastic reduction in the size of the pontine nucleus itself (compare Fig. 5 F and B, arrows and PN). Immunofluorescence analysis of the caudal portion of the pontine nucleus of P0 animals showed that the ectopic cluster of cells present in Math1+/lacZ;Tcf4+/− mice are positive for Pax6, indicating that these cells are part of the AES that normally gives rise to the pontine nucleus (Fig. 5J).

Fig. 5.

Fig. 5.

Open in a new tab

Genetic interaction between Math1 and Tcf4. A schematic of the coronal sections is shown in A. The framed portion represents the region analyzed by Nissl and Pax6 staining. (B–F) Nissl staining of coronal sections at the level of the caudal pontine nucleus of hindbrains collected from P0 mice. The brackets show the region corresponding to the location of the AES. The arrows indicate the cells of the AES. Magnifications of the region between brackets are shown in the Insets. PN, pontine nucleus. At least six animals per genotype were analyzed; genotypes are shown in each panel. (G–J) Pax6 staining of the same region analyzed by Nissl staining. Pax6 expression is revealed by the red color superimposed on the blue nuclear staining. Arrowheads indicate the position of the AES.

Other E-Proteins Are Expressed in RL and Can Interact with Math1.

Tcf4 is a member of the E-protein family, a group of highly conserved bHLH proteins with similar biochemical and functional characteristics. The presence of a phenotype in Tcf4-null animals raises the question as to why other members of the family are not able to compensate for the lack of Tcf4 in the progenitors of the pontine neurons. The simplest explanation is that the other E-protein coding genes are not expressed in the lower RL during mouse development. Alternatively, the E-proteins encoded by these genes might not be able to interact with Math1 to form functional DNA-binding dimers. To test these hypotheses, we performed in situ hybridization for the E2a and HEB genes on sagittal sections of embryos at embryonic day 14.5. Both E2a and HEB are expressed in the RL, suggesting that lack of compensation is not caused by different patterns of expression between E-proteins (SI Fig. 9A). To confirm that Math1 can form dimers with E-proteins other than Tcf4 (12, 33), we analyzed the ability of E47, a splice isoform of E2a, to interact with Math1. We transfected NIH 3T3 cells with a constant amount of a Math1-expressing plasmid and increasing concentrations of plasmids encoding the V5-tagged version of Tcf4 or E47, and immunoprecipitated with an antibody recognizing tagged Math1. Comparable amounts of Tcf4 and E47 coimmunoprecipitated (SI Fig. 9B), suggesting that the two E-proteins have a similar affinity for Math1. To verify that the Math1/E47 heterodimers bind to DNA, we performed EMSAs using the E-box-containing DNA sequence. The incubation of nuclear extracts from NIH 3T3 cells cotransfected with Math1 and E47 with the labeled oligonucleotide resulted in the formation of a DNA–protein complex (SI Fig. 9C, lane 11, band a), indicating that the Math1/E47 heterodimer binds the DNA target sequence. The ability of the Math1/E47 dimers to bind E-boxes was comparable to that of the Math1/Tcf4 dimers (SI Fig. 9C, lanes 5 and 11).

Tcf4/Math1 Dimers Play a Unique Role in Pontine Nucleus Development.

Having established that the lack of compensation by other E-proteins in Tcf4−/− mice is not due to absence of E2a and HEB in the RL, or to their inability to interact with Math1, we pondered whether the pontine nucleus phenotype of Tcf4−/− mice is due to a reduction in the dosage of genes coding for E-proteins. In the immune system, deletion of either Tcf4 or E2a causes a similar differentiation defect in B cells, an effect also present in Tcf4+/−;E2a+/− doubly heterozygous animals (19). These data suggested that the E-proteins have identical functions and that a reduction in the total E-protein level is the cause of the observed differentiation phenotype. To test for the role of dosage, we crossed Tcf4+/− mice with E2a+/− and HEB+/− mice and analyzed the doubly heterozygous animals by Pax6 staining (Fig. 6). Remarkably, Tcf4+/−;E2a+/− and Tcf4+/−;HEB+/− animals displayed normal development of the pontine nucleus (compare Fig. 6 C and D with A), which is distinctly different from what is seen in Tcf4−/− mice (Fig. 6B). To rule out compensatory mechanisms due to transcriptional up-regulation of the remaining allele of E2a or HEB in doubly heterozygous animals, we analyzed coronal sections obtained from E2a−/− or HEB−/− mice and observed normal development of the pontine nucleus (Fig. 6 E and F). Analysis of the other Math1-dependent structures in E2a- and HEB-null animals did not reveal any differentiation defect (data not shown). These findings strongly suggest that the defects in Tcf4−/− pontine nucleus cannot be attributed to a simple “gene dosage” effect.

Fig. 6.

Fig. 6.

Open in a new tab

Analysis of pontine nuclear neuron differentiation in mice lacking different E-proteins. The cells of the anterior migratory stream (AES) are visualized by immunofluorescence staining using the anti-Pax6 antibody. Pax6 expression is represented by the red staining in the pictures. Cell nuclei are stained in blue. The arrowheads are pointing to the AES. The genotype of the animals is shown in the left bottom part of each panel. PN, pontine nucleus.

Math1 Heterodimers Have Different Cell-Specific Transcriptional Activities.

To better understand the molecular basis of the differentiation phenotype seen in Tcf4-null mice, we compared the transcriptional activity of the different Math1 heterodimers in several cell types using the E-box–luciferase reporter construct. We cotransfected a Math1-expressing vector along with plasmids encoding any one of the four E-proteins in NIH 3T3 fibroblasts, neuroblastoma line Neuro2a cells, the teratoma cell line P19, and primary neural precursors purified from postnatal day 5 cerebella. In all cell types, the transcriptional activity of Math1 was highly dependent on the E-protein partner. The ratio between the activities of the different heterodimers varied dramatically between different cell lines (compare NIH 3T3 and P19 profiles in SI Fig. 10), suggesting that that the activity of each Math1 heterodimer can be differentially modulated in a cell context-specific manner.

Deletion of Tcf4 Does Not Affect the Locus Ceruleus (LC).

To further evaluate the role of Tcf4 in hindbrain development, we investigated the effect of its deletion on Mash1-dependent neural progenitors, which give rise to the main noradrenergic structure of the brainstem, the LC. We focused our attention on LC neurons because Tcf4 was reported to interact in vitro with Mash1 (34), and because the respiratory defects observed in patients with PHS have been attributed to defective function of the LC (24, 25). Staining of coronal sections of Tcf4−/− mice with an antibody raised against tyrosine hydroxylase, an enzyme expressed in the noradrenergic cells that form the LC, did not show any obvious morphological defect of the LC, suggesting that the Mash1-expressing progenitors of the hindbrain are able to differentiate properly in the absence of Tcf4 (SI Fig. 11).

Discussion

In this study we identified the interaction between the proneural protein Math1 and the E-protein Tcf4 and investigated the role of their heterodimers in development of the mouse hindbrain. We show that lack of Math1/Tcf4 heterodimers results in the loss of a restricted neuronal population in the hindbrain. We also demonstrate that the E-proteins E47, E12, and HEB do not compensate for the absence of Tcf4, despite their overlapping expression pattern and similar biochemical properties. Our data provide the first in vivo evidence for a specialized function of dimers formed between a proneural bHLH factor and a specific E-protein, revealing a new layer of complexity to the transcriptional events regulating the development of the mammalian nervous system.

Our data suggest that two different classes of neural progenitors exist in the RL. The first class of progenitors, to which most Math1-dependent progenitors belong, can differentiate normally when any one of the three E-protein genes is deleted, presumably because of functional compensation by the remaining E-proteins. The differentiation behavior of this class of neural progenitors is consistent with the current model of bHLH heterodimer function in the development of the nervous system, which suggests that proneural proteins mediate the transcriptional specificity of the heterodimers, whereas E-proteins are interchangeable (3, 35). The second class of progenitors, to which the pontine nuclear progenitors belong, requires the presence of Tcf4/Math1 dimers to activate the normal differentiation program. The reasons for the failure of other members of the E-protein family to compensate for the deletion of Tcf4 in this class of progenitors are not entirely clear. Our analysis of E2a and HEB mutant mice indicates that the inability of other E-proteins to compensate for the absence of Tcf4 in pontine nucleus progenitors is due to intrinsic molecular differences between these proteins and not to a generic gene dosage effect caused by the reduction of the number of alleles coding for E-proteins. The observation that the transcriptional activity of various Math1/E-protein heterodimers is differentially modulated by the cellular context supports the hypothesis that different Math1/E-protein dimers might interact with distinct cell-specific transcriptional cofactors to differentially activate target genes. In the developing embryo, factors expressed by pontine nuclear neuron progenitors might interact selectively with Math1/Tcf4 dimers to activate a specific differentiation program. The ability of a single proneural bHLH to interact with E-proteins to form heterodimers with different transcriptional activity would increase the number of specific genetic programs regulated by this class of transcription factors and possibly clarify how <10 proneural proteins coordinate the differentiation of all of the different progenitor populations of the mammalian nervous system.

Our data suggest a possible developmental model to explain how mutations in TCF4 cause the complex features of PHS, characterized by widespread neurological symptoms without obvious defects in the anatomical structure of the nervous system. If TCF4 plays limited and specialized roles in different neural progenitor domains, its deletion would be expected to affect the differentiation of restricted neuronal populations in several portions of the nervous system, causing a wide variety of neurological symptoms. The mental retardation and epilepsy, for example, could be due to aberrant migration of specific groups of cortical neurons, and the intestinal phenotype could be caused by a partially defective differentiation of neural crest cells. Some of the symptoms, like the hyperventilation phenotype, could also be caused by subtle deficits in different systems that are part of the same functional network. The breathing defects of PHS patients have been previously suggested to arise from a deficit in the function of the proneural bHLH Mash1 leading to aberrant development of the noradrenergic neurons of the LC (24, 25). However, our data show that mice lacking Tcf4 have a normal LC, suggesting that the function of the proneural gene Mash1 in the progenitors of these neurons is maintained in the absence of Tcf4. If this holds true in humans, it would imply that the hyperventilation phenotype of PHS is not caused by the absence of the LC, but most likely by the concomitant lack of small subsets of Mash1- and Math1-dependent medullary neurons involved in breathing control.

In sum, the interaction between Math1 and Tcf4 is necessary for the normal activation of the differentiation program of a subset of neural progenitors. We propose that a similar requirement exists in specific progenitor populations in different regions of the developing neural tube. Therefore, detailed analyses of the developmental function of the heterodimers formed between Tcf4 and the various proneural bHLH proteins will be important to define the role of this E-protein in development of the mammalian nervous system and to understand the basis of PHS.

Materials and Methods

Yeast Two-Hybrid Screen.

We screened a pretransformed mouse embryonic 12.5 library using the Matchmaker yeast two-hybrid system following the supplier's instructions (Clontech, Mountain View, CA). The bait consisted of a portion of Drosophila atonal from amino acid 248 to amino acid 312. After the first round of selection, we isolated and sequenced DNA from yeast. Independent clones were retested for growth on selective medium and for β-galactosidase activity.

Coimmunoprecipitation, EMSA, and Luciferase Assays.

For coimmunoprecipitation, cells were harvested 48 h after transfection and resuspended in lysis buffer (10 mM Tris·HCl, pH 8/140 mM NaCl/0.5% Triton X-100/1 mM iodoacetamide/1 mM PMSF/1% hemoglobin/1× protease inhibitors; all from Sigma, St. Louis, MO). After 4 h of incubation with agarose-conjugated mouse monoclonal antibodies (Sigma), proteins associated to the beads were subjected to Western blot by using rabbit polyclonal anti-Flag and anti-V5 antibodies (Sigma). EMSAs were performed as previously described (36). Detailed information about constructs and oligonucleotides used can be found in SI Text.

Nonradioactive in Situ Hybridization.

Tissue preparation, automated in situ hybridization, and digital imaging were performed as previously described (37). Probe details can be found in SI Text.

Histochemistry.

β-Galactosidase staining was performed as previously described (7). For Nissl staining, P0 brain paraffin sections were incubated in histoclear (Baxter, Deerfield, IL) and gradually rehydrated through a series of ethanol–water solutions. The slides were then incubated in cresyl violet solution (0.5% cresyl violet acetate from Sigma in 0.3% glacial acetic acid in water) and washed in increasing concentrations of ethanol in water. After complete dehydration in histoclear, slides were mounted with Cytoseal (Richard-Allan Scientific, Kalamazoo, MI) and analyzed with an Axiocam2 (Zeiss, Berlin, Germany).

Immunofluorescence.

P0 brains were collected, fixed in 4% paraformaldehyde, and frozen in OCT (Sakura, Torrance, CA). Sections (25 μm thick) were incubated overnight at 4°C with a 1:500 dilution of a rabbit polyclonal anti-Pax6 antibody (Covance, Princeton, NJ) or a dilution 1:1,000 of a mouse monoclonal anti-tyrosine hydroxylase antibody (Immunostar). Secondary antibodies conjugated with Alexa Fluor 488 or Cy3 (Molecular Probes, Carlsbad, CA) were then added for 1 h at room temperature. Cell nuclei were visualized by using TOTO3 (Molecular Probes). Images were acquired with a confocal Axiocam microscope (Zeiss). All of the experiments involving animals were performed following the Institutional Animal Care and Use Committee guidelines.

Supplementary Material

Supporting Information
pnas_0707456104_index.html (6.4KB, html)

Acknowledgments

We thank Yuan Zhuang at Duke University (Durham, NC) for the generous gift of embryonic stem cells with a targeted Tcf4 allele and for the HEB- and E2a-null mice; Gabriele Schuster, Sukeshi Vaishnav, Barbara Antalffy, and Richard Atkinson for technical support; members of the H.Y.Z. laboratory for critical reading of the manuscript and advice; and Mental Retardation and Developmental Disabilities Research Center Gene Expression and Neuropathology Core Grant 5 P30 HD024064. H.Y.Z. is an investigator and A.F. is a postdoctoral research associate with the Howard Hughes Medical Institute.

Abbreviations

RL

rhombic lip

LC

locus ceruleus

AES

anterior extramural migratory stream

bHLH

basic helix–loop–helix

PHS

Pitt–Hopkins syndrome

P0

postnatal day 0.

Footnotes

The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/cgi/content/full/0707456104/DC1.

References

  • 1.Guillemot F. Curr Opin Cell Biol. 2005;17:639–647. doi: 10.1016/j.ceb.2005.09.006. [DOI] [PubMed] [Google Scholar]
  • 2.Ross SE, Greenberg ME, Stiles CD. Neuron. 2003;39:13–25. doi: 10.1016/s0896-6273(03)00365-9. [DOI] [PubMed] [Google Scholar]
  • 3.Bertrand N, Castro DS, Guillemot F. Nat Rev Neurosci. 2002;3:517–530. doi: 10.1038/nrn874. [DOI] [PubMed] [Google Scholar]
  • 4.Parras CM, Schuurmans C, Scardigli R, Kim J, Anderson DJ, Guillemot F. Genes Dev. 2002;16:324–338. doi: 10.1101/gad.940902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Guillemot F, Lo LC, Johnson JE, Auerbach A, Anderson DJ, Joyner AL. Cell. 1993;75:463–476. doi: 10.1016/0092-8674(93)90381-y. [DOI] [PubMed] [Google Scholar]
  • 6.Casarosa S, Fode C, Guillemot F. Development (Cambridge, UK) 1999;126:525–534. doi: 10.1242/dev.126.3.525. [DOI] [PubMed] [Google Scholar]
  • 7.Wang VY, Rose MF, Zoghbi HY. Neuron. 2005;48:31–43. doi: 10.1016/j.neuron.2005.08.024. [DOI] [PubMed] [Google Scholar]
  • 8.Remenyi A, Scholer HR, Wilmanns M. Nat Struct Mol Biol. 2004;11:812–815. doi: 10.1038/nsmb820. [DOI] [PubMed] [Google Scholar]
  • 9.Castro DS, Skowronska-Krawczyk D, Armant O, Donaldson IJ, Parras C, Hunt C, Critchley JA, Nguyen L, Gossler A, Gottgens B, et al. Dev Cell. 2006;11:831–844. doi: 10.1016/j.devcel.2006.10.006. [DOI] [PubMed] [Google Scholar]
  • 10.Seo S, Richardson GA, Kroll KL. Development (Cambridge, UK) 2005;132:105–115. doi: 10.1242/dev.01548. [DOI] [PubMed] [Google Scholar]
  • 11.Lee SK, Pfaff SL. Neuron. 2003;38:731–745. doi: 10.1016/s0896-6273(03)00296-4. [DOI] [PubMed] [Google Scholar]
  • 12.Akazawa C, Ishibashi M, Shimizu C, Nakanishi S, Kageyama R. J Biol Chem. 1995;270:8730–8738. doi: 10.1074/jbc.270.15.8730. [DOI] [PubMed] [Google Scholar]
  • 13.Ben-Arie N, McCall AE, Berkman S, Eichele G, Bellen HJ, Zoghbi HY. Hum Mol Genet. 1996;5:1207–1216. doi: 10.1093/hmg/5.9.1207. [DOI] [PubMed] [Google Scholar]
  • 14.Ben-Arie N, Bellen HJ, Armstrong DL, McCall AE, Gordadze PR, Guo Q, Matzuk MM, Zoghbi HY. Nature. 1997;390:169–172. doi: 10.1038/36579. [DOI] [PubMed] [Google Scholar]
  • 15.Johnson JE, Birren SJ, Saito T, Anderson DJ. Proc Natl Acad Sci USA. 1992;89:3596–3600. doi: 10.1073/pnas.89.8.3596. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hu JS, Olson EN, Kingston RE. Mol Cell Biol. 1992;12:1031–1042. doi: 10.1128/mcb.12.3.1031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Murre C, McCaw PS, Vaessin H, Caudy M, Jan LY, Jan YN, Cabrera CV, Buskin JN, Hauschka SD, Lassar AB, et al. Cell. 1989;58:537–544. doi: 10.1016/0092-8674(89)90434-0. [DOI] [PubMed] [Google Scholar]
  • 18.Zhuang Y, Soriano P, Weintraub H. Cell. 1994;79:875–884. doi: 10.1016/0092-8674(94)90076-0. [DOI] [PubMed] [Google Scholar]
  • 19.Zhuang Y, Cheng P, Weintraub H. Mol Cell Biol. 1996;16:2898–2905. doi: 10.1128/mcb.16.6.2898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Uittenbogaard M, Chiaramello A. Brain Res Gene Expression Patterns. 2002;1:115–121. doi: 10.1016/s1567-133x(01)00022-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Soosaar A, Chiaramello A, Zuber MX, Neuman T. Brain Res Mol Brain Res. 1994;25:176–180. doi: 10.1016/0169-328x(94)90297-6. [DOI] [PubMed] [Google Scholar]
  • 22.Ik Tsen Heng J, Tan SS. BioEssays. 2003;25:709–716. doi: 10.1002/bies.10299. [DOI] [PubMed] [Google Scholar]
  • 23.Brockschmidt A, Todt U, Ryu S, Hoischen A, Landwehr C, Birnbaum S, Frenck W, Radlwimmer B, Lichter P, Engels H, et al. Hum Mol Genet. 2007;16:1488–1494. doi: 10.1093/hmg/ddm099. [DOI] [PubMed] [Google Scholar]
  • 24.Amiel J, Rio M, Pontual LD, Redon R, Malan V, Boddaert N, Plouin P, Carter NP, Lyonnet S, Munnich A, Colleaux L. Am J Hum Genet. 2007;80:988–993. doi: 10.1086/515582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Zweier C, Peippo MM, Hoyer J, Sousa S, Bottani A, Clayton-Smith J, Reardon W, Saraiva J, Cabral A, Gohring I, et al. Am J Hum Genet. 2007;80:994–1001. doi: 10.1086/515583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Pitt D, Hopkins I. Aust Paediatr J. 1978;14:182–184. doi: 10.1111/jpc.1978.14.3.182. [DOI] [PubMed] [Google Scholar]
  • 27.Peippo MM, Simola KO, Valanne LK, Larsen AT, Kahkonen M, Auranen MP, Ignatius J. Clin Dysmorphol. 2006;15:47–54. doi: 10.1097/01.mcd.0000184973.14775.32. [DOI] [PubMed] [Google Scholar]
  • 28.Nakada Y, Hunsaker TL, Henke RM, Johnson JE. Development (Cambridge, UK) 2004;131:1319–1330. doi: 10.1242/dev.01008. [DOI] [PubMed] [Google Scholar]
  • 29.Quan XJ, Denayer T, Yan J, Jafar-Nejad H, Philippi A, Lichtarge O, Vleminckx K, Hassan BA. Development (Cambridge, UK) 2004;131:1679–1689. doi: 10.1242/dev.01055. [DOI] [PubMed] [Google Scholar]
  • 30.Wang VY, Hassan BA, Bellen HJ, Zoghbi HY. Curr Biol. 2002;12:1611–1616. doi: 10.1016/s0960-9822(02)01144-2. [DOI] [PubMed] [Google Scholar]
  • 31.Ben-Arie N, Hassan BA, Bermingham NA, Malicki DM, Armstrong D, Matzuk M, Bellen HJ, Zoghbi HY. Development (Cambridge, UK) 2000;127:1039–1048. doi: 10.1242/dev.127.5.1039. [DOI] [PubMed] [Google Scholar]
  • 32.Landsberg RL, Awatramani RB, Hunter NL, Farago AF, DiPietrantonio HJ, Rodriguez CI, Dymecki SM. Neuron. 2005;48:933–947. doi: 10.1016/j.neuron.2005.11.031. [DOI] [PubMed] [Google Scholar]
  • 33.Helms AW, Abney AL, Ben-Arie N, Zoghbi HY, Johnson JE. Development (Cambridge, UK) 2000;127:1185–1196. doi: 10.1242/dev.127.6.1185. [DOI] [PubMed] [Google Scholar]
  • 34.Persson P, Jogi A, Grynfeld A, Pahlman S, Axelson H. Biochem Biophys Res Commun. 2000;274:22–31. doi: 10.1006/bbrc.2000.3090. [DOI] [PubMed] [Google Scholar]
  • 35.Zhuang Y, Barndt RJ, Pan L, Kelley R, Dai M. Mol Cell Biol. 1998;18:3340–3349. doi: 10.1128/mcb.18.6.3340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Flora A, Lucchetti H, Benfante R, Goridis C, Clementi F, Fornasari D. J Neurosci. 2001;21:7037–7045. doi: 10.1523/JNEUROSCI.21-18-07037.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Carson JP, Thaller C, Eichele G. Curr Opin Neurobiol. 2002;12:562–565. doi: 10.1016/s0959-4388(02)00356-2. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information
pnas_0707456104_index.html (6.4KB, html)
pnas_0707456104_07456Fig7.jpg (70.8KB, jpg)
pnas_0707456104_07456Fig8.jpg (117.3KB, jpg)
pnas_0707456104_07456Fig9.jpg (65KB, jpg)
pnas_0707456104_07456Fig10.jpg (46KB, jpg)
pnas_0707456104_07456Fig11.jpg (26.1KB, jpg)

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES