Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Mar 10.
Published in final edited form as: Neuron. 2011 Mar 10;69(5):930–944. doi: 10.1016/j.neuron.2011.02.008

An isoform-specific SnoN1-FOXO1 repressor complex controls neuronal morphogenesis and positioning in the mammalian brain

Mai Anh Huynh 1,2,3, Yoshiho Ikeuchi 1,†, Stuart Netherton 4,†, Luis de la Torre-Ubieta 1,5,†, Rahul Kanadia 6, Judith Stegmüller 1,7, Constance Cepko 6, Shirin Bonni 4, Azad Bonni 1,2,5,*
PMCID: PMC3073069  NIHMSID: NIHMS272227  PMID: 21382553

SUMMARY

Control of neuronal positioning is fundamental to normal brain development. However, the cell-intrinsic mechanisms that govern neuronal positioning remain to be elucidated. Here, we report that the spliced protein products of the transcriptional regulator SnoN, SnoN1 and SnoN2, harbor opposing functions in the coordinate regulation of neuronal branching and positioning. Knockdown of SnoN2 stimulates axon branching in primary neurons, and impairs migration of granule neurons in the rat cerebellar cortex in vivo. By contrast, SnoN1 knockdown suppresses SnoN2 knockdown-induced neuronal branching, and strikingly triggers excessive migration of granule neurons in the cerebellar cortex. We also find that SnoN1 forms a complex with the transcription factor FOXO1 that represses the X-linked lissencephaly gene encoding doublecortin (DCX). Accordingly, repression of DCX mediates the ability of SnoN1 to regulate branching in primary neurons and granule neuron migration in vivo. These data define an isoform-specific SnoN1-FOXO1 transcriptional complex that orchestrates neuronal branching and positioning in the brain, with important implications for the study of developmental disorders of cognition and epilepsy.

INTRODUCTION

The correct positioning of neurons is crucial for the establishment of neuronal circuitry and hence normal brain function (Ayala et al., 2007; Marin and Rubenstein, 2003). Defective migration and positioning of neurons is thought to form the cellular basis of inherited mental retardation and epilepsy syndromes (Gleeson, 2001; McManus and Golden, 2005; Schwartzkroin and Walsh, 2000; Sisodiya, 2004). Therefore, elucidation of the mechanisms governing neuronal migration and positioning will advance our understanding of both brain development and disease.

Regulation of the cytoskeleton plays a key role in the control of neuronal migration and positioning in the brain. The microtubule-associated protein doublecortin (DCX) has been implicated as a critical player in neuronal migration and morphology (des Portes et al., 1998; Gleeson et al., 1998). Mutations of DCX cause X-linked lissencephaly in males and the milder phenotype subcortical band heterotopia, also known as “double cortex”, in females. Inhibition of DCX function impairs neuronal migration and concomitantly stimulates branching of processes in neurons (Bai et al., 2003; Bielas et al., 2007; Kappeler et al., 2006; Koizumi et al., 2006). Notably, the association of impaired neuronal migration and increased neuronal branching has been observed upon inhibition of other migration genes (Guerrier et al., 2009; Heng et al., 2008; Nagano et al., 2004). These observations raise the question of whether cell-intrinsic transcriptional mechanisms might coordinately regulate neuronal migration and branching in neurons.

Granule neurons of the rodent cerebellar cortex provide a robust model system for studies of neuronal development in the brain (Ramon y Cajal, 1911). Granule neurons are generated in the external granule layer (EGL) of the cerebellar cortex. As the postmitotic granule neurons extend parallel fiber axons, their somas migrate radially in the molecular layer (Hatten, 1999). Upon arrival in the internal granule layer (IGL), granule neurons migrate farther to adopt their final position in a temporally defined manner, with older neurons residing deeper inside the IGL and younger neurons taking up residence in more superficial positions within the IGL (Altman and Bayer, 1997; Komuro and Rakic, 1998). However, the mechanisms that control granule neuron positioning within the IGL have remained unexplored.

The transcriptional regulator SnoN plays a critical role in axon morphogenesis, including the development of parallel fibers in the cerebellar cortex (Stegmuller et al., 2008; Stegmuller et al., 2006). Outside the nervous system, SnoN operates as a versatile transcriptional modulator that can either repress or activate transcription and thereby promotes or suppresses tumorigenesis (Luo, 2004; Pot and Bonni, 2008). As a transcriptional corepressor in proliferating cells, SnoN forms a complex with the transcription factor Smad2 and thereby inhibits Smad-dependent transcription (He et al., 2003; Stroschein et al., 1999). Intriguingly, SnoN’s transcriptional activating function mediates its ability to promote the growth of axons in neurons (Ikeuchi et al., 2009). These observations raise the question of whether SnoN’s transcriptionally repressive functions might regulate other features of neuronal development besides axon growth. Importantly, SnoN is found in two isoforms, SnoN1 and SnoN2, that are generated from alternative splicing of the Sno gene (Pelzer et al., 1996). However, the isoform-specific functions of SnoN1 and SnoN2 have remained unknown.

In this study, we identify unique functions for SnoN1 and SnoN2 in the control of neuronal branching and positioning. SnoN2 knockdown induces axon branching in primary granule neurons and inhibits their migration in the cerebellar cortex in vivo. In contrast, SnoN1 knockdown suppresses SnoN2 knockdown-induced branching in primary neurons, and induces migration of granule neurons to the deepest regions within the IGL in vivo. We also uncover a mechanism that underlies SnoN isoform-specific regulation of neuronal branching and migration. SnoN1 forms a complex with the transcription factor FOXO1 that represses DCX transcription in neurons. Accordingly, FOXO knockdown phenocopies the SnoN1 knockdown migration phenotype in the cerebellar cortex in vivo. In addition, DCX RNAi overrides the ability of SnoN1 RNAi to stimulate migration to the deepest regions of the IGL. Collectively, our data define the SnoN1-FOXO1 transcriptional repressor complex as a novel cell-intrinsic transcriptional mechanism that controls neuronal branching and positioning in the mammalian brain.

RESULTS

Isoform-specific functions of SnoN1 and SnoN2 in axon branching

SnoN1 and SnoN2 are the products of alternative splicing of the Sno gene. SnoN2 is generated by the use of a different 5’ splice site within exon 3, which results in a 46 amino acid deletion (Figure 1A) (Pearson-White and Crittenden, 1997; Pelzer et al., 1996). Both SnoN1 and SnoN2 are highly expressed in primary granule neurons and in the rat cerebellar cortex (Stegmuller et al., 2006). To characterize the isoform-specific functions of SnoN1 and SnoN2 in neurons, we employed a plasmid-based RNAi approach to induce acute knockdown of SnoN1 or SnoN2 specifically. Expression of short hairpin RNAs (shRNAs) targeting SnoN1 and SnoN2 robustly and specifically reduced the levels of endogenous SnoN1 and SnoN2 protein, respectively, in primary granule neurons (Figure 1B). We also confirmed that shRNAs targeting a region in both SnoN1 and SnoN2 (pan-SnoN shRNA) reduced both SnoN1 and SnoN2 protein in primary granule neurons (Figure 1B).

Figure 1. Isoform-specific functions of SnoN1 and SnoN2 in axon branching.

Figure 1

Open in a new tab

(A) Schematic of SnoN1 and SnoN2, the alternatively spliced isoforms of the Sno gene. The dach homology domain (DHD), SAND domain, Smad2-interacting motif, and coiled-coil domains (CC) are indicated. (B) Lysates of granule neurons transfected with the isoform-specific SnoN1, isoform-specific SnoN2, pan-SnoN (SnoN) RNAi, or control U6 RNAi plasmid were immunoblotted with the SnoN or Erk antibody, the latter to serve as a loading control. Asterisk denotes non-specific band. (C) Granule neurons transfected with the SnoN1, SnoN2, or control U6 RNAi plasmid together with a GFP expression plasmid were subjected to immunocytochemistry with an antibody to GFP after 1, 2, or 3 days in vitro (DIV). Representative neurons at DIV3 are shown along with magnified images of the axon shaft. Arrowheads, arrows, and asterisks indicate axons, dendrites, and cell bodies, respectively. Scale bars represent 50 µm. For additional images, see Figure S1A. (D) The percentage of neurons bearing exuberant axon branching was significantly higher in SnoN2 knockdown neurons compared to control and SnoN1 knockdown neurons at DIV2 and DIV3 (p<0.0001; ANOVA, n=3). (E) SnoN2 knockdown, but not SnoN1 knockdown, increased the number of secondary and tertiary protrusions per neuron at DIV2 and DIV3 (p<0.005, ANOVA, n=3). A total of 540 neurons were measured. (F) Lysates of 293T cells transfected with the SnoN2-WT or SnoN2-RES expression plasmid along with the SnoN2 RNAi or control U6 RNAi plasmid and the GFP expression plasmid were immunoblotted with the SnoN, GFP, or Erk antibody. (G) Granule neurons transfected with the SnoN2 RNAi or control U6 RNAi plasmid together with GFP and the expression plasmid encoding SnoN2-WT, SnoN2-RES, or the control vector pcDNA3 were analyzed as in (C). Scale bars represent 50 µm. (H) SnoN2-RES, but not SnoN2-WT, significantly reduced the percentage of neurons bearing exuberant axon branching (left panel: p<0.0001; ANOVA, n=3) and the number of secondary axon branches per neuron (right panel: p<0.005; ANOVA) in the background of SnoN2 RNAi. A total of 380 neurons were measured. (I) Lysates of neurons transfected with both SnoN2 RNAi and SnoN1 RNAi plasmids or the control U6 RNAi plasmid were immunoblotted with the SnoN or Erk antibody. Asterisk denotes non-specific band. (J) Neurons were transfected with the SnoN2 RNAi plasmid, SnoN2 RNAi and SnoN1 RNAi plasmids together, or the control U6 RNAi plasmid and analyzed as in (C). Scale bars represent 50 µm. For additional images, see Figure S1C. (K) Neurons treated as in (J) were analyzed as in (D) and (E). SnoN1 knockdown in the background of SnoN2 RNAi restored both the percentage of neurons bearing exuberant axon branching (left panel: p<0.0001; ANOVA, n=3) at DIV2 and DIV3 and the number of secondary and tertiary branches per neuron to control U6 baseline levels at DIV2 and DIV3 (right panel: p<0.0005; ANOVA, n=3). A total of 546 neurons were measured.

SnoN2 knockdown unexpectedly led to a striking branching phenotype in granule neurons, characterized by numerous protrusions emanating from the axon shaft (Figures 1C and S1A). In time course analyses, the percentage of cells with exuberant axon branching increased over time (Figure 1D). Quantification of the number of axon branches per neuron revealed that SnoN2 knockdown increased the number of both secondary and tertiary axon branches (Figure 1E). By contrast to the robust axon branching phenotype in SnoN2 knockdown neurons, SnoN1 knockdown failed to increase axon branching (Figures 1C, 1D, 1E, and S1A). Interestingly, neither SnoN1 RNAi nor SnoN2 RNAi reduced axon length (Figure S1B). Because pan-SnoN RNAi reduces axon length in granule neurons (Stegmuller et al., 2006), these results suggest that SnoN1 and SnoN2 have redundant functions in axon growth. In agreement with this conclusion, the combination of SnoN1 RNAi and SnoN2 RNAi reduced axon length, thus phenocopying the effect of pan-SnoN RNAi on axon growth (Figures 1J, S1C, and S1D) (Stegmuller et al., 2006). In addition, although pan-SnoN RNAi induced robust downregulation of the axon growth promoting signaling molecule Ccd1, a transcriptional target of SnoN (Ikeuchi et al., 2009), SnoN1 RNAi or SnoN2 RNAi alone failed to reduce Ccd1 mRNA levels in neurons (Figure S1E). In other experiments, SnoN1 RNAi and SnoN2 RNAi had little or no effect on neuron survival, suggesting that the morphological phenotypes were not due to impaired cell health (Figure S1F). SnoN1 RNAi and SnoN2 RNAi failed to alter the expression of the granule neuron marker MEF2A (data not shown), suggesting that the morphology phenotypes were not secondary to a change in the general differentiation state of granule neurons. Taken together, these results suggest that SnoN2 RNAi specifically impairs the restriction of axon branching in neurons.

To determine whether the SnoN2 RNAi-induced effect on neuronal morphology is the result of specific knockdown of SnoN2, we performed a rescue experiment. We generated an expression plasmid encoding SnoN2 using a cDNA containing silent mutations. SnoN2 RNAi induced knockdown of SnoN2 encoded by wild type cDNA (SnoN2-WT) but not the RNAi-resistant cDNA (SnoN2-RES) (Figure 1F). Importantly, expression of SnoN2-RES but not SnoN2-WT in the background of SnoN2 RNAi in granule neurons restored axon branching to levels similar to that of control-transfected neurons (Figures 1G and 1H). Expression of SnoN2 in the absence of SnoN2 RNAi in granule neurons had little or no effect on axon branching (data not shown). These results support the conclusion that the SnoN2 RNAi-induced axon branching phenotype is the result of specific knockdown of SnoN2.

Because exuberant branching is not observed in neurons expressing pan-SnoN shRNAs (Stegmuller et al., 2006), we asked whether SnoN1 and SnoN2 might harbor antagonistic functions, such that SnoN2 knockdown unmasks a function of SnoN1 in promoting axon branching. To test this possibility, we assessed the ability of SnoN1 RNAi to reverse the SnoN2 RNAi-induced branching phenotype in neurons. Simultaneous expression of SnoN1 shRNAs and SnoN2 shRNAs induced knockdown of both SnoN1 and SnoN2 isoforms in neurons (Figure 1I). SnoN1 knockdown in the background of SnoN2 RNAi restored both the percentage of branched neurons and the number of axon branches per neuron to baseline levels (Figures 1J, 1K, and S1C), suggesting that SnoN1 RNAi suppresses the SnoN2 RNAi-induced branching phenotype. Although the combined knockdown of SnoN1 and SnoN2 also reduced axon length (Figures S1C and S1D), suppression of axon branching occurred at a faster pace than the reduction of axon length (see right panel in Figure 1K and S1D). In addition, branching was suppressed in the subpopulation of SnoN1, SnoN2 double knockdown neurons that harbor short axons as effectively as in those with long axons (Figure S1G). These data suggest that the ability of SnoN1 RNAi to suppress SnoN2 RNAi-induced axon branching is not due to the reduction in axon length. SnoN2 knockdown but not SnoN1 knockdown also stimulated branching of dendrites without changing dendrite length (Figures S1J–S1L), and SnoN1 RNAi suppressed the SnoN2 RNAi-induced dendrite branching phenotype without reducing dendrite length (Figures S1M and S1N). These data further support the conclusion that SnoN1 RNAi suppresses SnoN2 knockdown-induced neuronal branching independently of reducing process length. Collectively, our findings suggest that SnoN1 and SnoN2 exert opposing effects on neuronal branching.

SnoN2 promotes neuronal migration in the cerebellar cortex in vivo

Growing evidence suggests that impaired neuronal migration in vivo is often associated with increased branching in primary neurons (Bielas et al., 2007; Guerrier et al., 2009; Kappeler et al., 2006; Nagano et al., 2004). We therefore explored whether SnoN1 and SnoN2 might have isoform-specific functions in the control of granule neuron migration and positioning in the cerebellar cortex. We used an in vivo electroporation method in postnatal rat pups to characterize neuronal migration and positioning within the developing rat cerebellar cortex (Konishi et al., 2004). Because the electroporation procedure targets cells in the EGL (data not shown), all transfected neurons are granule neurons. We injected rat pups at postnatal day 3 (P3) with a plasmid encoding the U6 promoter and cmv-driven green fluorescent protein (U6-cmvGFP), and returned pups to moms (Figure 2A). Animals were then sacrificed 3, 5, or 7 days after electroporation and coronal sections of the cerebellar cortex were subjected to immunohistochemistry using the GFP antibody. These analyses revealed transfected granule neurons at distinct stages of development, beginning with their appearance in the EGL at P6 and their subsequent positioning in the IGL observed at P8 and P10 (Figures 2A and 2B).

Figure 2. SnoN2 impairs neuronal migration in the cerebellar cortex in vivo.

Figure 2

Open in a new tab

(A) Schematic of in vivo electroporation and development of the cerebellar cortex. Rat pups were electroporated at postnatal day 3 (P3) with a U6 control plasmid that also expressed GFP (U6-cmvGFP). Pups were returned to moms and sacrificed on various days after electroporation. Inset contains an illustration of granule neuron development (Ramon y Cajal, 1911). (B) Rat pups electroporated at P3 with the control U6-cmvGFP plasmid were returned to moms and sacrificed 3, 5, and 7 days following electroporation at P6, P8, and P10, respectively. Representative images of coronal sections subjected to immunohistochemistry with the GFP antibody. Scale bar represents 50 µm. (C) Rat pups electroporated with a SnoN2 RNAi plasmid that also encodes GFP (U6/snoN2-cmvGFP) or the corresponding control U6-cmvGFP RNAi plasmid were sacrificed 5 days after electroporation, and cerebella were subjected to immunohistochemistry with the GFP antibody. Scale bar represents 50 µm. (D) SnoN2 RNAi significantly increased the number of granule neurons in the EGL/ML (p<0.005; ANOVA), and concomitantly reduced the proportion of granule neurons in the IGL (p<0.0005; ANOVA). The number of animals in each condition is indicated. A total of 18,465 neurons were counted. (E) Rat pups electroporated at P3 with the U6/snoN2-cmvGFP or corresponding control U6-cmvGFP RNAi plasmid together with a plasmid encoding RNAi-resistant SnoN2 (SnoN2-RES) or its control vector were sacrificed at P8, and cerebella analyzed as in (C). Scale bar represents 50 µm. (F) SnoN2 knockdown significantly increased the proportion of neurons in the EGL/ML (p<0.0005; ANOVA), and expression of SnoN2-RES significantly reduced the proportion of neurons in the EGL/ML in the presence of SnoN2 knockdown (p<0.005; ANOVA). The number of animals in each condition is indicated. A total of 2,586 neurons were counted. (G) Rat pups were electroporated with plasmids encoding the SnoN2, SnoN1, pan-SnoN shRNA SnoN RNAi, or corresponding control U6-cmvGFP RNAi plasmid. Scale bar represents 50 µm. (H) Migration of granule neurons from the EGL/ML to the IGL was significantly impaired in SnoN2 knockdown animals (p<0.0005; ANOVA), whereas animals with knockdown of SnoN1 alone (SnoN1 RNAi) or knockdown of both SnoN1 and SnoN2 (SnoN RNAi) did not exhibit a robust migration defect. The number of animals in each condition is indicated. A total of 7,171 neurons were counted.

We next electroporated P3 rat pups with a SnoN2 RNAi plasmid that also expressed GFP or the corresponding control U6-cmvGFP RNAi plasmid (Figure 2C). We quantified the effect of SnoN2 RNAi on neuronal migration by counting the number of GFP-positive granule neurons in the different layers of the cerebellar cortex. SnoN2 knockdown substantially increased the proportion of GFP-positive granule neurons in the EGL and molecular layer and reduced the number of neurons that reach the IGL in P8 rat pups (Figure 2D). SnoN2 knockdown also induced the formation of ectopic protrusions in parallel fibers and within somatic processes of granule neurons in the molecular and Purkinje cell layers (Figure S2A). Although the branching phenotype was more subtle in SnoN2 knockdown animals than in primary neurons, the in vivo phenotype was consistent and reproducible. Importantly, expression of the RNAi-resistant rescue form of SnoN2 (SnoN2-RES) in rat pups reversed the SnoN2 RNAi-induced phenotypes of impaired migration and ectopic protrusions in the cerebellar cortex (Figures 2E, 2F, S2B, and S2C). The SnoN2 knockdown-induced impairment of granule neuron migration was sustained in rat pups at P12 (Figures S2D and S2E). These results suggest that SnoN2 plays a critical role in promoting the migration of granule neurons to the IGL in the cerebellar cortex in vivo.

In contrast to the inhibition of granule neuron migration in SnoN2 knockdown animals, knockdown of SnoN1 or the combined knockdown of SnoN1 and SnoN2, using pan-SnoN RNAi, had little inhibitory effect on the migration of granule neurons from the EGL to the IGL (Figures 2G and 2H). These results suggest that SnoN1 knockdown suppresses the SnoN2 knockdown-induced phenotype. Notably, parallel fiber axons were significantly impaired upon pan-SnoN knockdown, but knockdown of SnoN1 or SnoN2 had a reduced or little effect, respectively, on parallel fiber formation (Figure S2F; Stegmuller et al, 2006), consistent with redundant roles of SnoN1 and SnoN2 in axon growth in primary neurons. In control experiments in which the bromodeoxyuridine derivative EdU was injected in rat pups 24 hours after electroporation, SnoN1 knockdown and SnoN2 knockdown had little or no effect on the proliferation of granule cell precursors in the cerebellar cortex in vivo (Figures S2G and S2H). SnoN knockdown does not affect expression of the granule marker MEF2A in vivo (Stegmuller et al., 2006). Together, these data suggest that SnoN1 and SnoN2 have antagonistic functions in the control of neuronal branching and granule neuron migration.

SnoN1 is required for neuronal positioning in the cerebellar cortex in vivo

In view of the opposing roles of SnoN1 and SnoN2 in granule neuron migration in vivo, we reasoned that inhibition of SnoN1 on its own might trigger excessive migration of granule neurons in the cerebellar cortex. Analyses of neuronal positioning in rat pups 9 days after electroporation at P12 uncovered a dramatic difference in neuronal positioning in SnoN1 knockdown animals (Figure 3A). In control-transfected rat pups, granule neurons were distributed throughout the IGL (Figure 3A). By contrast, granule neurons in SnoN1 knockdown animals were predominantly aligned at the bottom of the IGL, forming a seam-like collection of neurons at the junction of the IGL with the white matter (Figure 3A). To quantify the effect of SnoN1 knockdown on positioning, we stratified the IGL into three domains—upper, middle, and lower—and measured the number of GFP-positive cells in each domain. In control animals, more than two thirds of the granule neurons were in the upper and middle domains of the IGL and nearly a third were in the lower IGL domain (Figure 3B). However, nearly two thirds of granule neurons in SnoN1 knockdown animals were in the lower domain of the IGL and the remainder in the middle and upper IGL domains (Figure 3B). Thus, SnoN1 knockdown induced excessive migration of granule neurons within the IGL, increasing the proportion of neurons in the lower IGL by more than two-fold (Figure 3B). These findings suggest that SnoN1 is required for proper granule neuron positioning in the cerebellar cortex.

Figure 3. SnoN1 is required for neuronal positioning in the cerebellar cortex in vivo.

Figure 3

Open in a new tab

(A) Rat pups electroporated with the U6/snoN1-cmvGFP or corresponding control U6-cmvGFP RNAi plasmid were sacrificed 9 days later at P12, and cerebella were subjected to immunohistochemistry with the GFP antibody (green), the calbindin antibody to label the Purkinje-cell defined PCL and ML (red), and the DNA dye bisbenzimide (Hoechst 33258; blue). Representative coronal images of each condition are shown. Scale bar represents 50 µm. (B) SnoN1 knockdown significantly increased the proportion of neurons in the lower IGL (p<0.0005; ANOVA). The number of animals in each condition is indicated. A total of 9,289 neurons were counted. (C) Rat pups electroporated at P3 with the U6/snoN1-cmvGFP or corresponding control U6-cmvGFP RNAi plasmid together with a plasmid encoding RNAi-resistant human SnoN1 (SnoN1-RES) or its control vector were sacrificed at P12, and cerebella analyzed as in (A). Representative coronal images of SnoN1 knockdown together with the control vector or the SnoN1-RES expression plasmid are shown. Scale bar represents 50 µm. (D) SnoN1 knockdown significantly increased the proportion of neurons in the lower IGL (p<0.0001; ANOVA), and expression of SnoN1-RES significantly reduced the proportion of neurons in the lower IGL in the presence of SnoN1 knockdown (p<0.0001; ANOVA). The number of animals in each condition is indicated. A total of 25,591 neurons were counted.

We next determined whether the SnoN1 RNAi-induced effect on neuronal positioning in the cerebellar cortex is the result of specific knockdown of SnoN1. To rescue the SnoN1 RNAi-induced phenotype, we used an expression plasmid encoding human SnoN1 (SnoN1-RES), which contains five nucleotide mismatches in the region targeted by SnoN1 shRNAs. We confirmed that SnoN1 RNAi induced knockdown of SnoN1 encoded by wild type cDNA but not human cDNA (SnoN1-RES) (Figure S3A). Importantly, expression of SnoN1-RES in the background of SnoN1 RNAi in postnatal rat pups almost completely reversed the effect of SnoN1 RNAi on the positioning of granule neurons within the IGL in vivo (Figures 3C and 3D). Expression of SnoN1-RES on its own in the absence of SnoN1 RNAi had little or no effect on granule neuron positioning in the IGL in vivo (Figures S3B). Together, these results indicate that the SnoN1 RNAi-induced neuronal positioning phenotype is the result of specific knockdown of SnoN1 in the cerebellar cortex in vivo.

The X-linked lissencephaly gene DCX is a repressed target of SnoN1

The identification of opposing functions of the SnoN isoforms in neuronal branching and positioning led us to the question of the mechanism underlying SnoN isoform-specific functions in neurons. Since SnoN1 and SnoN2 are transcriptional regulators, we reasoned that a target gene may mediate biological responses in an isoform-specific manner. Because the X-linked lissencephaly protein doublecortin (DCX) controls both neuronal migration and branching (Bielas et al., 2007; Kappeler et al., 2006), we asked whether DCX might operate downstream of the SnoN isoforms in neurons.

DCX levels declined with neuronal maturation both in primary granule neurons and in the cerebellum (Figure 4A), suggesting that expression of DCX is developmentally regulated. Upon knockdown of SnoN isoforms in granule neurons, DCX protein and mRNA levels were elevated in SnoN1 knockdown but not SnoN2 knockdown neurons (Figures 4B, 4C, and data not shown). Chromatin immunoprecipitation (ChIP) assays revealed that endogenous SnoN occupied the endogenous DCX gene in granule neurons (Figure 4D). Together, these results suggest that DCX represents a novel direct repressed target gene of SnoN1 in neurons.

Figure 4. The X-linked lissencephaly gene DCX is a direct repressed target gene of SnoN1.

Figure 4

Open in a new tab

(A) Lysates of granule neurons (upper panel) or cerebellar tissue (lower panel) were immunoblotted with the DCX, 14-3-3β, or ERK antibody, the latter to serve as loading controls. DIV, days in vitro; P, postnatal day. (B) Lysates of granule neurons transfected with the SnoN1 RNAi plasmid or the control U6 RNAi plasmid were immunoblotted with the DCX or Erk antibody, the latter to serve as a loading control. (C) Quantitative RT-PCR on RNA extracted from granule neurons transfected with the SnoN1 RNAi plasmid or the control U6 RNAi plasmid. The levels of DCX mRNA were normalized relative to GAPDH mRNA levels in each sample and the fold change in DCX mRNA levels quantified. SnoN1 knockdown led to a 2.5-fold increase in DCX mRNA levels (p<0.05; t-test, n=3). (D) ChIP analysis of granule neuron lysates using the SnoN or IgG antibody. Quantitative PCR (qPCR) was performed using primers designed to encompass the putative FOXO binding site on the intronic DCX silencing region or to the control β-actin gene. SnoN occupancy at the DCX gene was significant relative to the control gene (p<0.05; t-test, n=3). For additional controls for the SnoN antibody, see Figures S4C–S4E. (E) Rat pups electroporated in vivo with expression plasmids encoding GFP and DCX RNAi, SnoN1 RNAi, the combination of DCX RNAi and SnoN1 RNAi, or the control U6-cmvGFP and psiSTRIKE-scramble-cmvGFP plasmids were sacrificed 9 days after electroporation, and cerebella were subjected to immunohistochemistry with the GFP (green) and calbindin antibody (red). Scale bar represents 50 µm. (F) DCX knockdown significantly increased the number of granule neurons in the upper IGL (p<0.05; ANOVA), and reduced the number of granule neurons in the middle and lower domains of the IGL (p<0.0005; ANOVA). SnoN1 knockdown alone increased the proportion of neurons in the lower IGL (p<0.0005; ANOVA). The combination of DCX knockdown and SnoN1 knockdown significantly increased the proportion of neurons in the upper IGL (p<0.0005; ANOVA), and decreased the number of neurons in the middle (p<0.005; ANOVA) and lower domains of the IGL (p<0.05; ANOVA). The number of animals in each condition is indicated. A total of 23,386 neurons were counted. (G) Granule neurons were transfected with the SnoN2 RNAi plasmid or control U6 RNAi plasmid, together with DsRed and the DCX expression plasmid pCAG-DCX-IRES-GFP or its control vector pCAG-IRES-GFP and subjected to immunocytochemistry with the DsRed antibody. Representative neurons are shown along with magnified images of the axon shaft. Arrowheads, arrows, and asterisks indicate axons, dendrites, and cell bodies, respectively. Scale bar represents 50 µm. For additional images, see Figure S4A. (H) DCX expression restored the SnoN2 knockdown-induced increase in axon branching (upper panel: p<0.0005; ANOVA, n=3) and number of axon branches per neuron to control vector baseline levels (lower panel: p<0.005; ANOVA, n=3). A total of 359 neurons were measured. Increased levels of DCX suppressed the SnoN2 RNAi-induced increase in axon branching.

Because DCX promotes neuronal migration and SnoN1 represses DCX expression, we asked whether inhibition of DCX might suppress the SnoN1 knockdown-induced neuronal positioning phenotype in the cerebellar cortex. DCX knockdown on its own in rat pups led to the accumulation of granule neurons in the upper IGL and reduced the proportion of granule neurons in the lower IGL (Figures 4E and 4F), suggesting that DCX plays a critical role in promoting granule neuron migration within the IGL. In epistasis analyses, we found that while SnoN1 knockdown increased the proportion of granule neurons in the lower domain of the IGL, the phenotype in animals in which DCX knockdown was induced in the background of SnoN1 knockdown was nearly indistinguishable from the positioning phenotype induced by DCX knockdown alone (Figures 4E and 4F). These results suggest that DCX knockdown suppresses the SnoN1 knockdown-induced neuronal positioning phenotype in vivo. In other experiments, DCX overexpression mimicked the ability of SnoN1 knockdown in completely suppressing the SnoN2 knockdown-induced branching phenotype in primary granule neurons (Figures 4G, 4H, and S4A). Collectively, these data suggest that repression of DCX expression mediates SnoN1’s function to coordinately regulate neuronal branching and migration.

FOXO1 and SnoN1 form a complex that represses DCX expression and thereby controls neuronal positioning

As a transcriptional corepressor, SnoN function is contingent upon its association with DNA-binding transcription factors. SnoN forms a complex with the transcription factor Smad2 and thereby represses Smad-dependent transcription in proliferating cells (He et al., 2003; Stroschein et al., 1999). However, knockdown of Smad2 surprisingly failed to alter levels of endogenous DCX expression in granule neurons (Figure S5A), suggesting that SnoN1 might repress DCX in a Smad-independent manner. Interrogation of the regulatory sequences within the DCX gene revealed an evolutionarily conserved FOXO binding site within a reported DCX gene silencing region in the first intron of the DCX gene (Karl et al., 2005). We asked whether SnoN1 might operate in concert with a FOXO family protein and thereby repress DCX transcription.

We found that exogenous FOXO1 associated with endogenous SnoN1 in transfected 293T cells (Figure 5A). In addition, endogenous FOXO1 interacted with endogenous SnoN1 in primary granule neurons (Figure 5B). These results suggest that SnoN1 forms a physical complex with FOXO1. Expression of SnoN1, but not SnoN2, significantly reduced the ability of FOXO1 to induce the expression of a FOXO-responsive luciferase reporter gene in cells (Figure S5B). These data suggest that SnoN1 represses FOXO1-dependent transcription.

Figure 5. FOXO1 and SnoN1 form a complex that represses DCX expression and thereby controls neuronal positioning.

Figure 5

Open in a new tab

(A) Lysates of 293T cells transfected with Flag-FOXO1 or control vector pcDNA3 were immunoprecipitated with the Flag antibody and immunoblotted with the SnoN antibody. Input lysates were also immunoblotted with the Flag or SnoN antibody. (B) Lysates of granule neurons were immunoprecipitated with the FOXO1 antibody and immunoblotted with the SnoN or FOXO1 antibody. Asterisk denotes non-specific band. (C) Lysates of granule neurons were immunoblotted with the FOXO1, DCX, or 14-3-3β antibody, the latter to serve as a loading control. DIV, days in vitro. (D) Lysates of granule neurons transfected with the FOXO RNAi or control U6 RNAi plasmid were immunoblotted with the FOXO1 or 14-3-3β antibody, the latter to serve as a loading control. FOXO RNAi reduced FOXO1 protein levels. (E) Neurons treated as in (D) were immunoblotted for DCX or 14-3-3β antibody. Endogenous DCX protein levels increased upon knockdown of FOXO1. (F) qRT-PCR on RNA extracted from granule neurons transfected with the FOXO RNAi or control U6 RNAi plasmid. The levels of DCX mRNA were normalized relative to GAPDH mRNA levels in each sample and the fold change in DCX mRNA levels quantified. FOXO knockdown led to a 6-fold increase in DCX mRNA levels (p<0.05; t-test, n=3). (G) ChIP analysis of granule neuron lysates using the FOXO1 or IgG antibody, followed by qPCR using primers for the DCX gene as in Figure 4D or the control synaptotagmin gene. FOXO1 occupancy at the DCX gene was significant relative to the control gene (p<0.05; t-test, n=3). (H) Granule neurons were transfected with the SnoN2 RNAi, FOXO RNAi, both SnoN2 RNAi and FOXO RNAi, or control U6 RNAi plasmid, together with GFP, and analyzed as in Figure 1C. Representative images of axons in control, SnoN2 knockdown, FOXO knockdown, or double SnoN2 and FOXO knockdown neurons are shown. Scale bar represents 50 µm. (I) Neurons transfected as in (H) were analyzed as in Figures 1D and 1E. FOXO knockdown in the background of SnoN2 RNAi restored both the percentage of neurons bearing exuberant branching (left panel: p<0.0005; ANOVA, n=3) and the number of axon branches per neuron (right panel: p<0.005; ANOVA, n=3) to control U6 baseline levels. A total of 371 neurons were measured. (J) Rat pups were electroporated in vivo with plasmids encoding FOXO RNAi and GFP (U6/foxo-cmvGFP) or the corresponding control U6-cmvGFP, sacrificed 9 days after electroporation, and cerebella were subjected to immunohistochemistry with the GFP antibody. Scale bar represents 50 µm. (K) Animals electroporated as in (J) were analyzed. FOXO knockdown significantly increased the number of granule neurons in the lower IGL (p<0.0005; ANOVA), and reduced the number of neurons in the upper (p<0.0005; ANOVA) and middle domains (p<0.005; ANOVA) of the IGL. The number of animals in each condition is indicated. A total of 10,834 neurons were counted. (L) Rat pups electroporated at P3 with the U6/foxo-cmvGFP or corresponding control U6-cmvGFP RNAi plasmid together with a plasmid encoding RNAi-resistant FOXO1 (FOXO1-RES) or its control vector were sacrificed at P12, and cerebella analyzed as in (J). Scale bar represents 50 µm. (M) Although FOXO RNAi stimulated granule neuron migration to the deepest region of the IGL in vivo (p<0.005; ANOVA), expression of FOXO1-RES significantly reversed the FOXO RNAi-induced effects on neuronal positioning (p<0.05; ANOVA). The number of animals in each condition is indicated. A total of 5,357 neurons were counted.

Consistent with the hypothesis that FOXO1 might repress DCX transcription in neurons, DCX protein levels declined in granule neurons over time with maturation, which correlated with a corresponding increase in FOXO1 protein levels (Figure 5C). We next determined the role of endogenous FOXO1 in the control of endogenous DCX expression. FOXO RNAi reduced the levels of endogenous FOXO1 in neurons (Figure 5D). Importantly, FOXO RNAi triggered a marked increase in endogenous DCX protein and mRNA levels (Figures 5E and 5F), suggesting that FOXO RNAi leads to derepression of DCX gene expression. ChIP analyses revealed that, like SnoN1, FOXO1 also occupied the endogenous DCX promoter in granule neurons (Figure 5G). Electrophoretic mobility shift assays revealed that recombinant FOXO1 robustly binds the putative FOXO binding sequence within the DCX promoter, and mutation of key consensus nucleotides of the FOXO binding motif within the DCX promoter abrogated binding to FOXO1 (Figure S5C). Together, these results suggest that FOXO1 directly binds the DCX promoter and represses DCX transcription in neurons.

We next determined the role of FOXO1 in mediating isoform-specific functions of SnoN1 in neuronal morphology and positioning. We first assessed whether FOXO1 mimics SnoN1 in antagonizing SnoN2 function in the control of branching in primary granule neurons. FOXO RNAi completely reversed the SnoN2 knockdown-induced increase in axon branching to baseline levels, suggesting that FOXO RNAi phenocopies the effect of SnoN1 RNAi in the control of neuronal branching (Figures 5H and 5I). We next asked whether FOXO1 controls neuronal positioning within the IGL in the cerebellar cortex in vivo. Remarkably, FOXO RNAi induced excessive migration of granule neurons within the IGL in rat pups analyzed at P12, increasing the proportion of granule neurons within the lower domain of the IGL to over 70% as compared to 30% in control animals (Figures 5J and 5K). Thus, FOXO RNAi phenocopies the effect of SnoN1 RNAi on neuronal positioning within the IGL. Importantly, the expression of an RNAi-resistant form of FOXO1 (FOXO1-RES) in the background of FOXO RNAi in rat pups reversed the FOXO RNAi-induced neuronal positioning phenotype in the cerebellar cortex (Figures 5L and 5M), supporting the conclusion that the FOXO RNAi-induced neuronal positioning phenotype is the result of specific knockdown of FOXO1 in vivo. The combination of SnoN1 RNAi and FOXO RNAi in rat pups did not additively increase the proportion of granule neurons in the deepest region of the IGL (Figure S5D), suggesting that SnoN1 and FOXO1 operate in a shared pathway to regulate neuronal positioning in the cerebellar cortex in vivo.

To determine the role of the SnoN1-FOXO1 interaction in the regulation of neuronal positioning in the cerebellar cortex, we performed structure-function analyses. Deletion of the C-terminal domain of SnoN1, which is dispensable for SnoN1’s ability to interact with Smad2 (He et al., 2003; Stroschein et al., 1999), impaired the ability of SnoN1 to associate with FOXO1 (Figures S5E and S5F). The SnoN1 mutant protein lacking the C-terminal domain (SnoN1 1–366) failed to repress FOXO1-dependent transcription (Figure S5G). Importantly, by contrast to SnoN1-RES, expression of SnoN1 1–366, which is not targeted by SnoN1 RNAi, failed to reverse the SnoN1 RNAi-induced phenotype of excess granule neurons in the deepest region of the IGL in vivo (Figure S5H). These results suggest that the C-terminal domain of SnoN1 is required for the formation of a transcriptional repressor complex with FOXO1 and hence for the proper positioning of granule neurons in the developing cerebellar cortex. Collectively, our findings support a model in which SnoN1 and FOXO1 function as components of a transcriptional complex that represses DCX transcription and thereby controls neuronal branching and positioning in the mammalian brain.

SnoN2 interacts with SnoN1 and thereby antagonizes SnoN1 function

We next determined the molecular basis underlying the antagonism of the SnoN isoforms in the regulation of neuronal branching and migration. We first asked whether SnoN2 and SnoN1 interact with each other. SnoN2 robustly associated with SnoN1 in co-immunoprecipitation analyses (Figures 6A–6C). Structure-function analyses revealed the C-terminal region containing the coiled-coil domains in both SnoN1 and SnoN2 are required for the SnoN2-SnoN1 interaction (Figures 6A–6C). Accordingly, the SnoN1 mutants SnoN1 1–539 and SnoN1 1–477 failed to effectively associate with SnoN2 (Figure 6B). Conversely, the SnoN2 mutant SnoN2 1–493 failed to effectively associate with SnoN1 (Figure 6C).

Figure 6. SnoN2 interacts with SnoN1 and thereby antagonizes SnoN1 function in branching and migration.

Figure 6

Open in a new tab

(A) Schematic of SnoN1 and SnoN2, along with deletion mutants. (B) Lysates of 293T cells transfected with expression plasmids encoding GFP-SnoN2 and wildtype Flag-HA-SnoN1 (F-HA-SnoN1), Flag-HA-SnoN1 1–539 (F-HA-SnoN1 1–539), Flag-HA-SnoN1 1–477 (F-HA-SnoN1 1–477), or the control vector pcDNA3 were immunoprecipitated with the Flag antibody and immunoblotted with the SnoN antibody. Input lysates were also immunoblotted with the SnoN or Flag antibody. (C) Lysates of 293T cells transfected with expression plasmids encoding GFP-SnoN1 and wildtype Flag-HA-SnoN2 (F-HA-SnoN2), Flag-HA-SnoN2 1–493 (F-HA-SnoN2 1–493), or the control vector pcDNA3 were immunoprecipitated with the Flag antibody and immunoblotted with the SnoN antibody. Input lysates were also immunoblotted with the SnoN or Flag antibody. The interaction of exogenous SnoN1 and SnoN2 requires the coiled-coil domains within the C-terminus of SnoN2. Asterisk denotes non-specific band. (D) C2C12 cells were transfected with a FOXO-responsive reporter gene that contains three copies of the FOXO responsive element from the IGFBP1 promoter (Tang et al., 1999), and a control renilla reporter gene together with expression plasmids encoding FOXO1, SnoN2, wildtype SnoN1 (WT), the mutant SnoN1 1–539 or SnoN1 1–477, or their control vectors. Firefly luciferase activity was normalized relative to renilla luciferase activity. SnoN2 significantly reduced the ability of wildtype SnoN1 but not the SnoN1 mutants to repress FOXO1-dependent transcription (p<0.01; ANOVA, n=6). (E) C2C12 cells were transfected with the FOXO-responsive reporter gene and control renilla reporter gene together with expression plasmids encoding FOXO1, SnoN1, wildtype SnoN2 (WT), the mutant SnoN2 1–493, or their control vectors. Firefly luciferase activity was normalized relative to renilla luciferase activity. Wildtype SnoN2 but not the SnoN2 mutant significantly reduced the ability of SnoN1 to repress FOXO1-dependent transcription (p<0.01; ANOVA, n=5). (F) Axon branching was quantified in granule neurons transfected with the SnoN2 RNAi or control U6 RNAi plasmid together with GFP and the expression plasmid encoding SnoN2-RES, SnoN2-RES 1–493, or the control vector pcDNA3 and were analyzed as in Figures 1C–1E. SnoN2-RES, but not SnoN2-RES 1–493, significantly reduced the percentage of neurons bearing exuberant axon branching (left panel: p<0.0005; ANOVA, n=3) and the number of secondary and tertiary axon branches per neuron (right panel: p<0.0001; ANOVA, n=3) in the background of SnoN2 RNAi. A total of 358 neurons were measured. (G) Rat pups electroporated at P3 with the U6/snoN2-cmvGFP or corresponding control U6-cmvGFP RNAi plasmid together with a plasmid encoding RNAi-resistant SnoN2 (SnoN2-RES), the SnoN2-RES 1–493 mutant, or control vector were sacrificed at P8, and cerebella subjected to immunohistochemistry with the GFP antibody. Scale bar represents 50 µm. (H) SnoN2 knockdown significantly increased the proportion of neurons in the EGL/ML (p<0.0001; ANOVA) and concomitantly reduced the number of neurons in the IGL (p<0.0001; ANOVA). Expression of SnoN2-RES, but not SnoN2-RES 1–493, significantly reduced the proportion of neurons in the EGL/ML (p<0.0001; ANOVA), and significantly increased the number of neurons in the IGL in the presence of SnoN2 knockdown (p<0.0001; ANOVA). The number of animals in each condition is indicated. A total of 9,239 neurons were counted. (I) A model of the isoform-specific SnoN1-FOXO1 transcriptional repressor complex and SnoN2 inhibition of the SnoN1-FOXO1 repressor complex in the coordinate regulation of neuronal branching and positioning.

We next determined the impact of the SnoN2-SnoN1 interaction on SnoN1 repression of FOXO1-dependent transcription. Expression of SnoN2 antagonized the ability of SnoN1 to repress FOXO1-dependent transcription (Figure S6A). In structure-function analyses, SnoN1 1–539 and SnoN1 1–477, which failed to effectively associate with SnoN2, repressed FOXO1-dependent transcription but were refractory to derepression by SnoN2 (Figure 6D). Conversely, in contrast to wildtype SnoN2, SnoN2 1–493, which failed to effectively interact with SnoN1, also failed to inhibit the ability of SnoN1 to repress FOXO1-dependent transcription (Figure 6E). These results suggest that SnoN2 interacts via its coiled-coil domains with SnoN1 and thereby derepresses the SnoN1-FOXO1 transcriptional repressor complex.

We next assessed the functional relevance of the SnoN2 interaction with SnoN1 on the antagonistic, isoform-specific functions of SnoN2 in the control of neuronal morphology and migration in primary neurons and the cerebellar cortex in vivo. Remarkably, in structure-function analyses, in contrast to SnoN2-RES, the SnoN2-RES 1–493 mutant failed to rescue the branching phenotype induced by SnoN2 knockdown in primary granule neurons (Figure 6F). In addition, in contrast to SnoN2-RES, the SnoN2-RES 1–493 mutant failed to rescue the SnoN2 RNAi-induced migration and branching phenotypes in the cerebellar cortex in rat pups (Figures 6G, 6H, and S6B). These data suggest that the SnoN2-SnoN1 interaction via their coiled-coil domains plays a critical role in the regulation of neuronal branching and migration. Collectively, our findings suggest SnoN2 interacts with SnoN1 and thereby derepresses the SnoN1-FOXO1 transcriptional repressor complex, providing a model whereby the opposing activities of SnoN1 and SnoN2 on neuronal morphology and positioning are mediated via the interaction of the two SnoN isoforms (Figure 6I).

DISCUSSION

In this study, we have discovered an isoform-specific SnoN1-FOXO1 transcriptional repressor complex that plays a fundamental role in neuronal positioning in the brain. Specific depletion of the transcriptional regulator SnoN1 or SnoN2 in primary granule neurons and in the rat cerebellar cortex in vivo reveals that the two SnoN isoforms have opposing functions in the control of neuronal branching and migration. Whereas SnoN2 restricts neuronal branching and promotes migration of granule neurons to the IGL in the cerebellar cortex, SnoN1 promotes branching and inhibits the migration of granule neurons within the IGL. We have also uncovered the molecular basis of SnoN isoform-specific functions in neurons. SnoN1 interacts with the transcription factor FOXO1, forming a complex that directly inhibits expression of the lissencephaly gene DCX in neurons. Accordingly, repression of DCX mediates the ability of SnoN1 to control granule neuron position within the IGL. Finally, we have uncovered a mechanism by which SnoN2 antagonizes the functions of SnoN1 in neurons. SnoN2 associates with SnoN1 via a coiled-coil domain interaction and thereby inhibits the ability of SnoN1 to repress FOXO1-dependent transcription. Importantly, the SnoN2-SnoN1 interaction plays a critical role in the regulation of neuronal branching and migration. Collectively, these findings define SnoN1 and FOXO1 as components of a transcriptional complex that directly represses DCX expression and thereby orchestrates neuronal morphology and positioning in the mammalian brain.

The identification of the transcriptional regulators SnoN1 and SnoN2 as novel cell-intrinsic regulators of both neuronal branching and positioning supports the concept that neuronal migration and branching are intimately linked mechanistically. Besides the lissencephaly protein DCX, which associates with microtubules and promotes their stabilization (Gleeson et al., 1999), the Elongator complex, the slit-robo GTPase activating protein srGAP2, and the small GTP-binding protein Rnd2 represent regulators of cytoskeletal and membrane dynamics that have been implicated in the coordinate control of branching and cortical migration (Creppe et al., 2009; Guerrier et al., 2009; Heng et al., 2008). These observations raise the question of whether in addition to controlling DCX transcription, the SnoN isoforms might also regulate the expression of other local effectors of neuronal morphology and migration.

Elucidation of SnoN1’s isoform-specific function in the control of neuronal position in different domains within the IGL suggests that this major cerebellar cortical layer harbors functionally distinct sublayers. Prior to our study, the molecular basis of granule neuron migration within the IGL remained unknown. Identification of a transcriptional mechanism that is required for proper neuronal positioning within the IGL may provide the basis in future studies for characterization of programs of gene expression that define the distinct domains of the IGL within the cerebellar cortex.

The isoform-specific function of SnoN1 and SnoN2 in neurons raises the intriguing question of whether expression of the SnoN isoforms is developmentally regulated. In situ analyses utilizing fluorescent probes specific for SnoN1 and SnoN2 in the developing cerebellar cortex revealed differences in their pattern of expression. SnoN1 is expressed in both the EGL and IGL and at relatively low levels in the molecular layer. By contrast, SnoN2 is expressed in the EGL and molecular layer and is found at modestly lower levels in the IGL (Figure S2I). The apparent enrichment of SnoN2 in the molecular layer and SnoN1 in the IGL are consistent with the isoform-specific requirement for SnoN2 in granule neuron migration from the EGL to the IGL and for the isoform-specific requirement for SnoN1 in granule neuron positioning in the IGL. Because the antagonism of the two SnoN isoforms requires their physical interaction, lower levels of SnoN1 in the molecular layer may enhance the ability of SnoN2 to antagonize SnoN1 and hence enable the isoform-specific function of SnoN2 in promoting granule neuron migration to become manifest within the molecular layer. Therefore, the protein-protein interaction dependent mechanism of SnoN2 antagonism of SnoN1 may work hand-in-hand with the differential expression pattern of the SnoN isoforms to allow for isoform-specific functions of SnoN to operate at distinct points in neuronal development. Notably, FOXO1 levels increase with neuronal maturation (Figure 5C), suggesting that FOXO1 expression is also regulated during brain development. Together, these observations suggest that after granule neurons differentiate and begin arriving in the IGL, the abundance of the SnoN1-FOXO1 repressor complex may increase, correlating with the role of this complex in the control of positioning in maturing neurons.

The identification of an intimate link between SnoN1 and FOXO1 bears significant ramifications for our understanding of the biology of both major families of SnoN and FOXO transcriptional proteins. The FOXO proteins activate or repress transcription (Paik et al., 2007; Ramaswamy et al., 2002; van der Vos and Coffer, 2008). However, although the mechanisms by which FOXO proteins induce transcription have been intensely studied (Van Der Heide et al., 2004; van der Vos and Coffer, 2008), the molecular basis of FOXO-dependent repression remained unknown. The finding that SnoN1 serves as a corepressor for FOXO1-regulated transcription illuminates a mechanism by which FOXOs directly repress transcription. Since the FOXO proteins regulate diverse biological processes from cell survival to metabolism to longevity (Accili and Arden, 2004; Salih and Brunet, 2008), our findings raise the possibility that SnoN1 may play a role in these fundamental biological processes.

Characterization of DCX as a novel direct target gene of the SnoN1-FOXO1 transcriptional repressor complex highlights the importance of regulation of DCX gene expression in the control of neuronal positioning in brain development and disease. In light of the dramatic consequence of DCX loss-of-function mutations in mental retardation and epilepsy, it will be important to determine whether deregulation of SnoN1 and FOXO1 function might contribute to the pathogenesis of neurodevelopmental disorders of cognition and epilepsy. Interestingly, forced expression of DCX in the early postnatal period reduces subcortical band heterotopia and seizure threshold in an animal model of human double cortex syndrome (Manent et al., 2009). Therefore, identification of a SnoN1-FOXO1 repressor complex as a regulator of DCX gene expression raises the prospect that manipulation of SnoN1 or FOXO1 function may provide a potential avenue of treatment for developmental disorders of cognition and epilepsy.

EXPERIMENTAL PROCEDURES

Plasmids

shRNA plasmids were produced by cloning the following oligonucleotides into pBS/U6 or pBS/U6-cmvGFP (targeted sequence is underlined): SnoN1 RNAi: 5’-AACCAGTAGAGAATTATACAGTTGTTAACTATAACTGTATAATTCTCTACTGGTTCTTTTTTG-3’ and SnoN2 RNAi: 5’-AAGGCAGAGACAAATTCATCAATCCGTTAACAATTGATGAATTTGTCTCTGCCTTCTTTTTTG-3’. The pan-SnoN RNAi, FOXO RNAi, and FOXO1-RES expression plasmids have been described (Bernard, 2004; Daitoku et al., 2004; Lehtinen et al., 2006; Sarker et al., 2005; Yuan et al., 2008). The RNAi-resistant rescue construct (SnoN2-RES) was generated using Quik-Change Site Directed Mutagenesis (Stratagene) and verified by sequencing. The cDNAs encoding the mutants SnoN1 1–539, SnoN1 1–477, SnoN1 1–366, and SnoN2 1–493 were generated by PCR, subcloned into pcDNA3 or pEGFP-C2 (Clontech), and verified by sequencing.

Primary neuron cultures and transfection

Granule neurons were prepared from postnatal day 6 (P6) Long-Evans rat pups and transfected either 8 hours, 2 days, or 4 days in vitro after plating using a modified calcium phosphate method as described (Konishi et al., 2004) with indicated plasmids together with either GFP, DsRed, or β-galactosidase expression plasmid to visualize transfected neurons. To rule out the possibility that the effects of RNAi or protein expression on morphology were due to any effect of these manipulations on cell survival, the anti-apoptotic protein Bcl-xL was co-expressed in all neuronal transfections except those in which survival was assessed. The expression of Bcl-xL has little or no effect on axon or dendrite morphology (Gaudilliere et al., 2004; Konishi et al., 2004). In additional control experiments, granule neurons were transfected with the SnoN1 RNAi, SnoN2 RNAi, or control U6 RNAi plasmid along with the GFP expression plasmid in the absence of Bcl-xL (Figures S1H and S1I). In other control experiments, granule neurons were transfected with the SnoN2 RNAi plasmid or control U6 RNAi plasmid, together with DsRed and the DCX expression plasmid pCAG-DCX-IRES-GFP or its control vector pCAG-IRES-GFP, in the absence of the Bcl-xL expression plasmid and subjected to immunocytochemistry with the DsRed antibody (Figure S4B). For all these RNAi plasmids, the branching phenotypes observed in the presence of Bcl-xL paralleled exactly the branching phenotypes observed in the absence of Bcl-xL (Figures S1H, S1I, and S4B). For high-efficiency transfection of granule neurons, we employed a nucleofection method.

Immunoprecipitation and immunoblotting analyses

Immunoprecipitation and immunoblotting analyses were performed as described (Kim et al., 2009).

In vivo electroporation and immunohistochemistry

In vivo electroporation was performed as described (Konishi et al., 2004). For more details, please see supplemental information.

Chromatin immunoprecipitation (ChIP)

ChIP was perfomed as described (Yuan et al., 2009). For more details, please see supplemental information.

Analysis of neuronal morphology

Axon and dendrite length morphometry was done as described (Gaudilliere et al., 2004; Konishi et al., 2004). Briefly, images of transfected GFP or DsRed-positive neurons were captured in a blinded manner using a Nikon Eclipse TE2000 epifluorescence microscope. Length was measured using SPOT imaging software. The percentage of neurons bearing exuberant axon branching was qualitatively assessed in a blinded manner and is approximated by a cut-off of 8 or more secondary axon branches. Axon branching was measured by quantifying the number of secondary, tertiary, and quaternary protrusions emanating from the axon shaft in images of GFP or DsRed-positive neurons.

Statistics

Statistical analyses were done using Statview 5.0.1 software. Bar graphs are presented as the mean + SEM except for analyses of neuronal branching in Figures 1, 4H, 5I, 6F, S1G–S1I, S1L, S1N, S2C, S4, and S6B where mean + SD is shown. For experiments in which only two groups were analyzed, the t-test was used. Pairwise comparisons within multiple groups were done by analysis of variance (ANOVA) followed by the Fischer’s PLSD post-hoc test.

Supplementary Material

01

ACKNOWLEDGEMENTS

We thank John Blenis and Randy King for helpful discussions, John Parnavelas for the DCX plasmids, Dan Bernard for the Smad2 RNAi plasmid, Akiyoshi Fukamizu for the Flag-FOXO1 plasmid, and members of the Bonni laboratory for helpful discussions and critical reading of the manuscript. This work was supported by NIH grants NS041021 and NS051255 (A.B.), the Canadian Institutes of Health Research (S.B.), the Alberta Cancer Board (S.B. and S.N.), an NIH Training Grant GM077226 (M.A.H.), the Albert J. Ryan Foundation (M.A.H. and L.T.U.), the Human Frontier Science Program Long-Term Fellowship (Y.I), the National Science Foundation (L.T.U.), and the Deutsche Forschungsgemeinschaft (J.S.).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  1. Accili D, Arden KC. FoxOs at the crossroads of cellular metabolism, differentiation, and transformation. Cell. 2004;117:421–426. doi: 10.1016/s0092-8674(04)00452-0. [DOI] [PubMed] [Google Scholar]
  2. Altman J, Bayer S. Development of the Cerebellar System: In Relation to Its Evolution, Structure, and Functions. Boca Raton: CRC Press; 1997. [Google Scholar]
  3. Ayala R, Shu T, Tsai LH. Trekking across the brain: the journey of neuronal migration. Cell. 2007;128:29–43. doi: 10.1016/j.cell.2006.12.021. [DOI] [PubMed] [Google Scholar]
  4. Bai J, Ramos RL, Ackman JB, Thomas AM, Lee RV, LoTurco JJ. RNAi reveals doublecortin is required for radial migration in rat neocortex. Nature neuroscience. 2003;6:1277–1283. doi: 10.1038/nn1153. [DOI] [PubMed] [Google Scholar]
  5. Bernard DJ. Both SMAD2 and SMAD3 mediate activin-stimulated expression of the follicle-stimulating hormone beta subunit in mouse gonadotrope cells. Molecular endocrinology (Baltimore, Md. 2004;18:606–623. doi: 10.1210/me.2003-0264. [DOI] [PubMed] [Google Scholar]
  6. Bielas SL, Serneo FF, Chechlacz M, Deerinck TJ, Perkins GA, Allen PB, Ellisman MH, Gleeson JG. Spinophilin facilitates dephosphorylation of doublecortin by PP1 to mediate microtubule bundling at the axonal wrist. Cell. 2007;129:579–591. doi: 10.1016/j.cell.2007.03.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Creppe C, Malinouskaya L, Volvert ML, Gillard M, Close P, Malaise O, Laguesse S, Cornez I, Rahmouni S, Ormenese S, et al. Elongator controls the migration and differentiation of cortical neurons through acetylation of alpha-tubulin. Cell. 2009;136:551–564. doi: 10.1016/j.cell.2008.11.043. [DOI] [PubMed] [Google Scholar]
  8. Daitoku H, Hatta M, Matsuzaki H, Aratani S, Ohshima T, Miyagishi M, Nakajima T, Fukamizu A. Silent information regulator 2 potentiates Foxo1-mediated transcription through its deacetylase activity. Proceedings of the National Academy of Sciences of the United States of America. 2004;101:10042–10047. doi: 10.1073/pnas.0400593101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. des Portes V, Francis F, Pinard JM, Desguerre I, Moutard ML, Snoeck I, Meiners LC, Capron F, Cusmai R, Ricci S, et al. doublecortin is the major gene causing X-linked subcortical laminar heterotopia (SCLH) Human molecular genetics. 1998;7:1063–1070. doi: 10.1093/hmg/7.7.1063. [DOI] [PubMed] [Google Scholar]
  10. Gaudilliere B, Konishi Y, de la Iglesia N, Yao G, Bonni A. A CaMKII-NeuroD signaling pathway specifies dendritic morphogenesis. Neuron. 2004;41:229–241. doi: 10.1016/s0896-6273(03)00841-9. [DOI] [PubMed] [Google Scholar]
  11. Gleeson JG. Neuronal migration disorders. Mental retardation and developmental disabilities research reviews. 2001;7:167–171. doi: 10.1002/mrdd.1024. [DOI] [PubMed] [Google Scholar]
  12. Gleeson JG, Allen KM, Fox JW, Lamperti ED, Berkovic S, Scheffer I, Cooper EC, Dobyns WB, Minnerath SR, Ross ME, et al. Doublecortin, a brain-specific gene mutated in human X-linked lissencephaly and double cortex syndrome, encodes a putative signaling protein. Cell. 1998;92:63–72. doi: 10.1016/s0092-8674(00)80899-5. [DOI] [PubMed] [Google Scholar]
  13. Gleeson JG, Lin PT, Flanagan LA, Walsh CA. Doublecortin is a microtubule-associated protein and is expressed widely by migrating neurons. Neuron. 1999;23:257–271. doi: 10.1016/s0896-6273(00)80778-3. [DOI] [PubMed] [Google Scholar]
  14. Guerrier S, Coutinho-Budd J, Sassa T, Gresset A, Jordan NV, Chen K, Jin WL, Frost A, Polleux F. The F-BAR domain of srGAP2 induces membrane protrusions required for neuronal migration and morphogenesis. Cell. 2009;138:990–1004. doi: 10.1016/j.cell.2009.06.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hatten ME. Central nervous system neuronal migration. Annual review of neuroscience. 1999;22:511–539. doi: 10.1146/annurev.neuro.22.1.511. [DOI] [PubMed] [Google Scholar]
  16. He J, Tegen SB, Krawitz AR, Martin GS, Luo K. The transforming activity of Ski and SnoN is dependent on their ability to repress the activity of Smad proteins. The Journal of biological chemistry. 2003;278:30540–30547. doi: 10.1074/jbc.M304016200. [DOI] [PubMed] [Google Scholar]
  17. Heng JI, Nguyen L, Castro DS, Zimmer C, Wildner H, Armant O, Skowronska-Krawczyk D, Bedogni F, Matter JM, Hevner R, et al. Neurogenin 2 controls cortical neuron migration through regulation of Rnd2. Nature. 2008;455:114–118. doi: 10.1038/nature07198. [DOI] [PubMed] [Google Scholar]
  18. Ikeuchi Y, Stegmuller J, Netherton S, Huynh MA, Masu M, Frank D, Bonni S, Bonni A. A SnoN-Ccd1 pathway promotes axonal morphogenesis in the mammalian brain. J Neurosci. 2009;29:4312–4321. doi: 10.1523/JNEUROSCI.0126-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kappeler C, Saillour Y, Baudoin JP, Tuy FP, Alvarez C, Houbron C, Gaspar P, Hamard G, Chelly J, Metin C, et al. Branching and nucleokinesis defects in migrating interneurons derived from doublecortin knockout mice. Human molecular genetics. 2006;15:1387–1400. doi: 10.1093/hmg/ddl062. [DOI] [PubMed] [Google Scholar]
  20. Karl C, Couillard-Despres S, Prang P, Munding M, Kilb W, Brigadski T, Plotz S, Mages W, Luhmann H, Winkler J, et al. Neuronal precursor-specific activity of a human doublecortin regulatory sequence. Journal of neurochemistry. 2005;92:264–282. doi: 10.1111/j.1471-4159.2004.02879.x. [DOI] [PubMed] [Google Scholar]
  21. Kim AH, Puram SV, Bilimoria PM, Ikeuchi Y, Keough S, Wong M, Rowitch D, Bonni A. A centrosomal Cdc20-APC pathway controls dendrite morphogenesis in postmitotic neurons. Cell. 2009;136:322–336. doi: 10.1016/j.cell.2008.11.050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Koizumi H, Higginbotham H, Poon T, Tanaka T, Brinkman BC, Gleeson JG. Doublecortin maintains bipolar shape and nuclear translocation during migration in the adult forebrain. Nature neuroscience. 2006;9:779–786. doi: 10.1038/nn1704. [DOI] [PubMed] [Google Scholar]
  23. Komuro H, Rakic P. Distinct modes of neuronal migration in different domains of developing cerebellar cortex. J Neurosci. 1998;18:1478–1490. doi: 10.1523/JNEUROSCI.18-04-01478.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Konishi Y, Stegmuller J, Matsuda T, Bonni S, Bonni A. Cdh1-APC controls axonal growth and patterning in the mammalian brain. Science (New York, NY. 2004;303:1026–1030. doi: 10.1126/science.1093712. [DOI] [PubMed] [Google Scholar]
  25. Lehtinen MK, Yuan Z, Boag PR, Yang Y, Villen J, Becker EB, DiBacco S, de la Iglesia N, Gygi S, Blackwell TK, et al. A conserved MST-FOXO signaling pathway mediates oxidative-stress responses and extends life span. Cell. 2006;125:987–1001. doi: 10.1016/j.cell.2006.03.046. [DOI] [PubMed] [Google Scholar]
  26. Luo K. Ski and SnoN: negative regulators of TGF-beta signaling. Current opinion in genetics & development. 2004;14:65–70. doi: 10.1016/j.gde.2003.11.003. [DOI] [PubMed] [Google Scholar]
  27. Manent JB, Wang Y, Chang Y, Paramasivam M, LoTurco JJ. Dcx reexpression reduces subcortical band heterotopia and seizure threshold in an animal model of neuronal migration disorder. Nature medicine. 2009;15:84–90. doi: 10.1038/nm.1897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Marin O, Rubenstein JL. Cell migration in the forebrain. Annual review of neuroscience. 2003;26:441–483. doi: 10.1146/annurev.neuro.26.041002.131058. [DOI] [PubMed] [Google Scholar]
  29. McManus MF, Golden JA. Neuronal migration in developmental disorders. Journal of child neurology. 2005;20:280–286. doi: 10.1177/08830738050200040301. [DOI] [PubMed] [Google Scholar]
  30. Nagano T, Morikubo S, Sato M. Filamin A and FILIP (Filamin A-Interacting Protein) regulate cell polarity and motility in neocortical subventricular and intermediate zones during radial migration. J Neurosci. 2004;24:9648–9657. doi: 10.1523/JNEUROSCI.2363-04.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Paik JH, Kollipara R, Chu G, Ji H, Xiao Y, Ding Z, Miao L, Tothova Z, Horner JW, Carrasco DR, et al. FoxOs are lineage-restricted redundant tumor suppressors and regulate endothelial cell homeostasis. Cell. 2007;128:309–323. doi: 10.1016/j.cell.2006.12.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pearson-White S, Crittenden R. Proto-oncogene Sno expression, alternative isoforms and immediate early serum response. Nucleic acids research. 1997;25:2930–2937. doi: 10.1093/nar/25.14.2930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pelzer T, Lyons GE, Kim S, Moreadith RW. Cloning and characterization of the murine homolog of the sno proto-oncogene reveals a novel splice variant. Dev Dyn. 1996;205:114–125. doi: 10.1002/(SICI)1097-0177(199602)205:2<114::AID-AJA3>3.0.CO;2-L. [DOI] [PubMed] [Google Scholar]
  34. Pot I, Bonni S. SnoN in TGF-beta signaling and cancer biology. Current molecular medicine. 2008;8:319–328. doi: 10.2174/156652408784533797. [DOI] [PubMed] [Google Scholar]
  35. Ramaswamy S, Nakamura N, Sansal I, Bergeron L, Sellers WR. A novel mechanism of gene regulation and tumor suppression by the transcription factor FKHR. Cancer cell. 2002;2:81–91. doi: 10.1016/s1535-6108(02)00086-7. [DOI] [PubMed] [Google Scholar]
  36. Ramon y Cajal S. Histology of the Nervous System of Man and Vertebrates. Oxford, UK: Oxford University Press; 1911. [Google Scholar]
  37. Salih DA, Brunet A. FoxO transcription factors in the maintenance of cellular homeostasis during aging. Current opinion in cell biology. 2008;20:126–136. doi: 10.1016/j.ceb.2008.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sarker KP, Wilson SM, Bonni S. SnoN is a cell type-specific mediator of transforming growth factor-beta responses. The Journal of biological chemistry. 2005;280:13037–13046. doi: 10.1074/jbc.M409367200. [DOI] [PubMed] [Google Scholar]
  39. Schwartzkroin PA, Walsh CA. Cortical malformations and epilepsy. Mental retardation and developmental disabilities research reviews. 2000;6:268–280. doi: 10.1002/1098-2779(2000)6:4<268::AID-MRDD6>3.0.CO;2-B. [DOI] [PubMed] [Google Scholar]
  40. Sisodiya SM. Malformations of cortical development: burdens and insights from important causes of human epilepsy. Lancet neurology. 2004;3:29–38. doi: 10.1016/s1474-4422(03)00620-3. [DOI] [PubMed] [Google Scholar]
  41. Stegmuller J, Huynh MA, Yuan Z, Konishi Y, Bonni A. TGFbeta-Smad2 signaling regulates the Cdh1-APC/SnoN pathway of axonal morphogenesis. J Neurosci. 2008;28:1961–1969. doi: 10.1523/JNEUROSCI.3061-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Stegmuller J, Konishi Y, Huynh MA, Yuan Z, Dibacco S, Bonni A. Cell-intrinsic regulation of axonal morphogenesis by the Cdh1-APC target SnoN. Neuron. 2006;50:389–400. doi: 10.1016/j.neuron.2006.03.034. [DOI] [PubMed] [Google Scholar]
  43. Stroschein SL, Wang W, Zhou S, Zhou Q, Luo K. Negative feedback regulation of TGF-beta signaling by the SnoN oncoprotein. Science (New York, NY. 1999;286:771–774. doi: 10.1126/science.286.5440.771. [DOI] [PubMed] [Google Scholar]
  44. Tang ED, Nunez G, Barr FG, Guan KL. Negative regulation of the forkhead transcription factor FKHR by Akt. The Journal of biological chemistry. 1999;274:16741–16746. doi: 10.1074/jbc.274.24.16741. [DOI] [PubMed] [Google Scholar]
  45. Van Der Heide LP, Hoekman MF, Smidt MP. The ins and outs of FoxO shuttling: mechanisms of FoxO translocation and transcriptional regulation. The Biochemical journal. 2004;380:297–309. doi: 10.1042/BJ20040167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. van der Vos KE, Coffer PJ. FOXO-binding partners: it takes two to tango. Oncogene. 2008;27:2289–2299. doi: 10.1038/onc.2008.22. [DOI] [PubMed] [Google Scholar]
  47. Yuan Z, Becker EB, Merlo P, Yamada T, DiBacco S, Konishi Y, Schaefer EM, Bonni A. Activation of FOXO1 by Cdk1 in cycling cells and postmitotic neurons. Science (New York, NY. 2008;319:1665–1668. doi: 10.1126/science.1152337. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

01
NIHMS272227-supplement-01.doc (4.9MB, doc)

RESOURCES