Calmin expression in embryos and the adult brain, and its regulation by all-trans retinoic acid

Abstract

The vitamin A metabolite, all-trans retinoic acid (atRA), is a regulator of nervous system development. Using a subtracted cDNA library constructed from neuroblastoma cells, the atRA-responsive gene calmin (Clmn) was identified (Merrill et al. [2004] Biol Chem 385:605-614). The Clmn transcript is detected very early in rat embryonic development and is sensitive to retinoid status. In vitamin A-deficient embryos, Clmn mRNA is dramatically down-regulated in the neuroepithelium adjacent to the somites, and this expression can be rescued with the addition of atRA. In embryonic day 18.5 embryos, CLMN is detected in regions where newly differentiated neurons are found, including the neural retina and the cortical plate; and in the adult brain, CLMN is most highly expressed in the neuron cell bodies of the hippocampus, cerebellum, and olfactory bulb. Thus, Clmn is sensitive to retinoid status during early gestational stages, and its expression is relegated to postmitotic neuronal cells in the adult rat brain. Developmental Dynamics 239:610–619, 2010. © 2009 Wiley-Liss, Inc.

INTRODUCTION

Vitamin A plays an essential role in vertebrate embryogenesis, including development of the nervous system (McCaffery and Drager, 2000; Clagett-Dame and DeLuca, 2002; Duester, 2008). The vitamin A metabolite, all-trans retinoic acid (atRA), regulates the transcription of target genes by binding to the retinoic acid receptor (RAR) (Chambon, 1996). The RAR heterodimerizes with the retinoid X receptor, and this ligand-bound complex interacts with specific enhancer regions of DNA, called retinoic acid response elements (RARE), to modulate the expression of atRA-regulated genes (Clagett-Dame and Plum, 1997; Balmer and Blomhoff, 2002). In an effort to identify atRA-responsive genes that play a role in neuronal development, a subtractive cDNA library prepared from atRA-treated and untreated human SH-SY5Y cells was screened, and calmin (Clmn or calponin-like, transmembrane; also known as retinoic acid induced in neuroblastoma 12 or Rainb12) was identified (Merrill et al., 2004). Our group has shown that Clmn is widely distributed in cancer cell lines, and is up-regulated in human neuroblastoma, breast cancer, and myeloid leukemia cells within 4 to 24 hr after exposure to atRA.

CLMN is a carboxy-terminal transmembrane-containing protein of unknown function. Based on sequence homology databases, CLMN is proposed to contain two N-terminal calponin homology domains, which are known actin-binding interfaces found in several scaffolding proteins, including spectrin and dystrophin (Gimona et al., 2002). Additionally, studies reveal sequence homology with enaptin as well as NUANCE, a protein shown to connect the nucleus to the actin cytoskeleton (Zhen et al., 2002). In addition to an ORF encoding for a 1021 amino acid (aa) murine protein (NCBI accession no. AB047978), several Clmn splice variants have been described, including two isoforms predicted to result in truncated proteins both lacking the transmembrane domain, and a third with a 31 aa insertion that retains the transmembrane domain (Ishisaki et al., 2001).

Analysis of adult mouse tissues shows Clmn is expressed in the brain, liver, kidney, large intestine, and the testis (Ishisaki et al., 2001; Takaishi et al., 2003). Clmn expression increases in the mouse testis during maturation, and the CLMN protein is found specifically in the maturing spermatids. Other than skin, the expression of calmin mRNA or protein has not been studied in developing embryos.

In the present work, we have characterized the spatio-temporal expression of calmin using immunohistochemistry and in situ hybridization in rat embryos. We show that Clmn mRNA expression in the neural tube of early embryos is sensitive to retinoid status. Additionally, we report that the expression of the full-length CLMN protein is confined to neurons in the adult hippocampus, cerebellum, and olfactory bulb.

RESULTS

Immunoblot Analysis of CLMN Protein

To study the CLMN protein, an antibody was generated in rabbit against a C-terminal peptide (aa 947-961, Fig. 1A). This antibody was designed to detect the full-length protein. To test antibody specificity, the full-length CLMN open reading frame was expressed using both a mammalian and a bacterial expression system. Murine CLMN containing a C-terminal 3x-flag tag was detected at 160 kDa using the anti-flag or the anti-CLMN antibody, and was not detected when nonimmune serum was used (Fig. 1B). Similar results were obtained when the full-length human protein was overexpressed in mammalian cells (data not shown). A protein of approximately 160 kDa was also specifically detected by the anti-CLMN antibody in rat and mouse brain (Fig. 1C), and agrees with the molecular size reported by Takaishi et al. (2003) in mouse cerebrum and cerebellum using a polyclonal antibody generated using bacterially expressed CLMN. CLMN expressed as a N-terminal fusion protein with GST in bacteria appeared at approximately 190 kDa (27 kDa GST + 160 kDa CLMN) using both a GST monoclonal antibody and the anti-CLMN antibody (data not shown). Thus, the anti-peptide CLMN antibody generated here specifically detects human, mouse, and rat CLMN.

Figure 1. Analysis of CLMN protein by immunoblotting. A: Schematic of the full-length mouse CLMN protein and the putative functional domains including two calponin homology (CH) domains (hatched) and a transmembrane domain (solid); the location of the CLMN peptide used to generate the CLMN antibody is also shown (black bar). B: Immunodetection of overexpressed murine CLMN (arrowhead) containing a 3X flag tag (CLMN-3xFl) using either the CLMN antibody or a monoclonal antibody to the flag sequence. Immunoblotting with nonimmune serum is shown as a control. Lysates from N2A cells were prepared after transfection with CLMN-3xFl or empty vector (pIRES; negative control). C: Immunoblotting of endogenous CLMN protein in adult rat and mouse brain lysates using the CLMN antibody or nonimmune serum. Immunoblottting of lysates from cells transfected with CLMN-3xFl was performed as a positive control. αTubulin (55 kDA) was analyzed in B and C by immunoblotting as a control for protein loading.

Distribution of CLMN Protein in the Adult Rat Brain

CLMN immunostaining was observed in the hippocampus, olfactory bulb, cerebellum, cerebral cortex, piriform cortex of the rhinencephalon, medial habenular nucleus of the dorsal thalamus, and hypothalamic nuclei (Fig. 2, and data not shown). It was also observed in a subset of large cells lining the border of the periaqueductal gray; these cells may represent neurons of the mesencephalic nucleus of the trigeminal nerve (Supp. Fig. S1, which is available online). In the hippocampus, CLMN was found in the dentate gyrus and all CA regions as well as the hilus (Fig. 2A and inset). The CLMN staining was specific, as shown by the lack of staining in the control section in which nonimmune serum was used (Fig. 2B). Within the cortex, staining was most notable in layers II–V (Fig. 2C) and staining in the piriform cortex was also observed (Fig. 2D). In the olfactory bulb, CLMN was specifically found within cell bodies of neurons in the granule cell layer, as evidenced by expression in neuron-specific NeuN-positive cells (Fig. 2E,F). CLMN was highly expressed in Purkinje cells in the cerebellum (Fig. 2G,H) as confirmed by colocalization with calbindin D28K (Fig. 2I,J). While calbindin D28K was found both in the cell bodies and neuronal processes (red) of the Purkinje cells, CLMN staining (green) was confined to the Purkinje cell body (Fig. 2J). Very light CLMN staining was also noted in the internal granule cell layer, but was absent from the both the molecular layer and the white matter of the cerebellum. The expression of CLMN was verified in several regions of the adult brain by immunoblotting. A 160 kDa protein was specifically detected in the cortex, cerebellum, hippocampus, and olfactory bulb (Supp. Fig. S2).

Figure 2. CLMN expression in coronal sections of the adult rat brain. A: Specific CLMN immunostaining in the dentate gyrus (dg) and cornu ammonis (CA) regions of the hippocampus. A higher magnification showing CLMN labeling of cells in the hilus (h) region is shown in the inset. B: Note the absence of specific staining in the hippocampus when nonimmune serum is used (compare B with A). C,D: CLMN immunoreactivity is detected in the adult cortical laminae (II–V; C), and the piriform cortex (pc; D) in the rhinencephalon. E: CLMN immunostaining in the olfactory bulb. F: dotted box in E is the region of the granule cell layer (gcl) shown in higher magnification in F, where CLMN (green immunofluorescence) is found in the same cells stained with nuclear neuronal marker NeuN (red). G: CLMN is detected at high levels in the cell soma of Purkinje neurons (pur) as well as in the granule cell layer (gcl) in the cerebellum. H: The region shown in the dotted box in G is shown in higher magnification. I: Calbindin D28K is expressed in the cell soma and dendritic processes of the Purkinje neurons in the molecular layer (ml) while CLMN expression is evident in the cell soma of the Purkinje neurons and in the internal granule cell layer (compare H with I). J: Overlay of CLMN (green) and calbindin D28K (red) immunofluorescence shows colocalization (yellow) in the cell soma of Purkinje neurons. wm, white matter.

Because CLMN protein was highly expressed in the adult rodent hippocampus, we next examined whether staining was confined to neurons, or whether the protein was also expressed in non-neuronal cells. CLMN (green) staining was found in the same cells as the neuronal marker, NeuN (red) in the CA1–3 regions (Fig. 3A, and data not shown), as well as in the dentate gyrus (Fig. 3B) and hilus (data not shown). CLMN staining did not overlap that of the glial cell marker, glial fibrillary acidic protein (GFAP; red; Fig. 3C,D). CLMN staining was found in the cell bodies of neurons that also stained positive for MAP2, a brain specific microtubular marker, but did not colocalize with MAP2 in the dendritic compartment of cells (Fig. 3F,G). Thus, in the adult brain, CLMN immunostaining was found exclusively in postmitotic neuronal cell bodies.

Figure 3. CLMN immunostaining is confined to the cell body of hippocampal neurons. A,B: CLMN (green) and NeuN (red) staining is found in the same cells in the CA3 region and dentate gyrus (dg). C,D: Non-neuronal cells that stain positive for glial fibrillary acidic protein (GFAP; red) are not immunoreactive for CLMN (green). F,G: Within neurons, CLMN and MAP2 staining do not overlap; CLMN immunostaining is confined to the cell body whereas MAP2, a microtubular marker, specifically stains the neuronal processes. E: A schematic of the adult mammalian hippocampus. Regions within the CA3 and dg that were shown in A–G are indicated by the dotted boxes. CA, cornu ammonis; h, hilus.

CLMN Expression in the Embryo

Because little is known about CLMN expression in developing embryos, antibody staining was examined in frontal sections of the head of an embryo at 18.5 days (E18.5). In the forebrain region, light CLMN immunostaining was seen in the cortical plate, whereas staining was absent in the intermediate zone (Fig. 4A,A′). There was also very faint staining in regions corresponding to the ventricular and subventricular zones, identified by staining an adjacent section with Ki67 (Fig. 4B), which marks cells in interphase and undergoing mitosis (Scholzen and Gerdes, 2000). There was little CLMN expression in the midbrain, except in a small subset of cells that appeared to define the border of the periaqueductal gray (Fig. 4C,C′), and in the inferior tectal neuroepithelium (Fig. 4D). In the cerebellum, CLMN protein was most abundant in the external granule cell layer, a transient cell population that will ultimately give rise to the internal granule cell layer (Fig. 4D,D′). In all of these regions, the specificity of CLMN antibody staining was confirmed by the absence of staining in the nonimmune control sections (Fig. 4A″,C″,D″). Staining of the hypoglossal nerve (CN12) was also evident in the medulla (data not shown). In the eye, CLMN was prominently expressed in the differentiating cell layer, as well as in a few cells at the outer edge of the proliferating retina (Fig. 4E,G), and the staining was specific as shown by the nonimmune control section (Fig. 4F). The differentiating region of the retina was identified with ISL1 (Islet-1; Fig. 4H), a protein expressed in the early differentiating retinal ganglion cells (Gong et al., 1995; Galli-Resta et al., 1997) and the proliferating layer was observed by Ki67 staining (Fig. 4I). Finally, CLMN was expressed in several other head structures, including the epithelium of the nasal cavity, dorsal surface of the tongue, and the ocular muscles (Fig. 4G, and data not shown).

Figure 4. CLMN immunostaining in frontal (coronal) sections of the head at embryonic day (E) 18.5. A: CLMN immunostaining is present in the forebrain cortical plate (cp). A higher magnification of the cortical plate region stained with the CLMN antibody is shown in (A′), whereas specific staining is absent in the adjacent panel (A″) in which nonimmune serum was used. B: Adjacent section to (A) probed with the proliferating marker Ki67 show prominent staining in the ventricular (vz) and subventricular zones (svz) of the forebrain where faint CLMN staining is also detected (compare A with B). C,D: In the midbrain, CLMN is localized in a small subset of cells at the border of the dorsal and ventral periaqueductal gray (C, arrow) and is also found in the inferior tectal neuroepithelium (D, itn). In the cerebellum (cb), CLMN is localized to the external granule cell layer (D, egl). A higher magnification of specific CLMN staining in the cells lining the periaqueductal gray and in the cerebellum is shown in (C′) and (D′), respectively; and the lack of staining in these regions with nonimmune serum is shown in (C″) and (D″). E,F: CLMN is expressed in the differentiating cell layer of the neural retina (nr), and in a smaller subset of cells lining the outer edge of the proliferating retina (pr; E); F shows the lack of staining in the retina when nonimmune serum is used. G: A higher magnification of specific CLMN immunostaining in the retina and ocular muscle (om). H,I: ISL1, a protein expressed in the early differentiating retinal ganglion cells (H), and Ki67 (I) staining are shown for reference. iz, intermediate zone.

At very early stages of embryonic development, Clmn expression was examined by in situ hybridization. This approach was used because the Clmn signal to background staining ratio was better when using the riboprobe as compared to antibody staining in early embryos. Clmn mRNA was first observed in presomitic (late head fold stage) embryos, with strong staining observed in the caudal mesoderm (Fig. 5A). At the 1- to 2-somite (s) stage, light staining was also noted in the head region (data not shown), and at 3–4s, staining was found in association with the presumptive fore- and mid-brain regions of the embryo (Fig. 5B). At 3–4s, transcript was also present in the neuroepithelium adjacent to the developing somites (Fig. 5C) as well as in the tail region in the mesoderm but not the neuroepithelium of the tail region (Fig. 5D). At 7s, staining in the forebrain and midbrain neural plate was evident, and staining appeared in the neuroepithelium just caudal to the otic sulcus, with the strongest staining immediately adjacent to the entire somitic mesoderm (Fig. 5E). With development, the caudal domain of Clmn expression in the neuroepithelium/neural tube followed that of somitic development (Fig. 5F–H). By E12.5, Clmn mRNA was detected in the ependymal layer of the developing spinal cord, dorsal root ganglion and emerging motor axons, sympathetic ganglion, as well as the developing lung bud (Fig. 5I).

Figure 5. Clmn mRNA is expressed in early rat embryos. A: Detection of Clmn by whole-mount in situ hybridization in the caudal mesoderm (cm) at the late head fold (presomite) stage. B–D: At the 3- to 4-somite (s) stage, Clmn staining is seen in the presumptive forebrain (fb) and midbrain (mb) (B), the neuroepithelium adjacent to the developing somites (C, arrow), and in the caudal mesodermal layers of the tail region (D). E: At 7s, Clmn mRNA is expressed in the neuroepithelium caudal to the otic sulcus (os) and in the forebrain and midbrain regions. F–H: Clmn is found in the neuroepithelium of the developing neural tube at 12–19s. Note, Clmn expression in the neuroepithelium ends coincident with the most recently formed somite (G). I: Clmn mRNA is also expressed in the developing nervous system at E12.5, with expression in the dorsal spinal cord (sc), floor plate (fp), dorsal root ganglion (drg), motor neurons (mn) and sympathetic ganglion (sg); Clmn is also found at high levels in the developing lung bud (lb). LHF, late head fold; ov, otic vesicle; pos, preotic sulcus.

Clmn Expression Is Down-regulated in Early Vitamin A-Deficient Embryos

Because Clmn mRNA is regulated by atRA in some cell lines, we next determined whether Clmn expression was altered in early embryos made severely deficient in vitamin A. Embryos (4–5s) from vitamin A-sufficient (VAS) mothers showed the expected pattern of Clmn expression in the neuroepithelium in the head region and adjacent to the somites, as well as in the tail mesoderm. In contrast, as shown in transverse vibratome sections, vitamin A-deficient (VAD) embryos showed a dramatic loss of staining in the neuroepithelium adjacent to the developing somites (compare Fig. 6E to 6B), whereas expression in the head region was less affected and that in the tail region was unchanged (compare Fig. 6D to 6A; Fig. 6F to 6C). At 6–7s, VAS embryos continued to express Clmn in the neuroepithelium adjacent to the somites (Fig. 6G), whereas expression in this region was lost in VAD embryos (Fig. 6H). However, VAD embryos from mothers supplemented with atRA (250 μg/g diet) from the onset of pregnancy showed a rescue of Clmn expression in the trunk region (Fig. 6I). Thus, atRA acts to regulate Clmn expression in the neuroepithelium adjacent to the developing somites in the early embryo.

Figure 6. Clmn expression in the neuroepithelium of the developing neural tube is sensitive to retinoid status. A: At the 4- to 7-somite (s) stage, Clmn transcript is expressed in the neuroepithelium adjacent to the somites, as well as in the developing forebrain/midbrain and tail mesoderm in vitamin A-sufficient (VAS) embryos. D: Vitamin A-deficient (VAD) embryos show a specific loss of Clmn expression in the neuroepithelium adjacent to the somites. Transverse sections reveal a loss of Clmn due to VAD in the neuroepithelium (compare E with the VAS section, B) but not in the mesodermal layer of the developing tail (compare F with C). G–I: At 6–7s, Clmn expression in the neuroepithelium adjacent to the developing somites (s) is present in VAS (G) but not VAD (H) embryos, and is rescued by the addition of atRA (I). pos, preotic sulcus.

VAD embryos contain undetectable levels of atRA as assessed using a sensitive HPLC-F9-RARE-lacZ reporter assay, whereas VAS embryos at E10.5 from mothers supplemented with retinol contain approximately 60 fmol atRA (Supp. Table S1). The presence of Clmn expression in the head and tail regions in embryos devoid of measurable atRA suggests that ex-pression in these regions may be independent of atRA. Embryos from mothers fed a high level of atRA (250 μg/g diet) have detectable levels of atRA and also show rescued Clmn expression in the neuroepithelium adjacent to the somites supporting that, at least in this region, maintenance of Clmn expression requires atRA.

To determine whether the effect of atRA on Clmn expression is direct, SH-SY5Y cells were treated with vehicle or 10⁻⁶ M atRA in the absence and presence of the transcriptional inhibitor, actinomycin-D, or cycloheximide, an inhibitor of protein translation. The induction of Clmn mRNA by atRA at 4 hr was eliminated with the addition of actinomycin-D, but was not reduced by cycloheximide (Fig. 7). In fact, Clmn mRNA was increased 7.6-fold above vehicle when protein synthesis was inhibited, suggesting a stabilization of the mRNA. The rapid induction of Clmn mRNA and its inhibition by actinomycin-D, suggest that Clmn is directly regulated by atRA and its receptors.

Figure 7. The induction of Clmn mRNA by atRA does not require new protein synthesis, but is blocked by an inhibitor of transcription. Clmn is up-regulated by exposure of SH-SY5Y cells to atRA for 4 hr. Co-treatment with the transcriptional inhibitor actinomycin-D results in a total loss of Clmn induction by all-trans retinoic acid (atRA), whereas exposure of atRA-treated cells to the protein translation inhibitor cycloheximide does not block the induction of Clmn mRNA. Values are mean ± SEM.

DISCUSSION

In the present study, we show that calmin is widely distributed in both the embryonic and adult nervous system. In early embryos, at least a subset of Clmn expression is responsive to retinoid status. The vitamin A metabolite all-trans retinoic acid (atRA) is required for early embryogenesis, including the developing nervous system (Clagett-Dame and DeLuca, 2002; Duester, 2008). Embryos made deficient in retinoid, either through nutritional or genetic methods, die early in development (White et al., 1998, 2000; Niederreither et al., 1999; Mic et al., 2002). Provision of atRA in amounts sufficient for early embryogenesis, but limiting at later gestational times, yields fetuses with a host of developmental abnormalities (See et al., 2008). The first report of Clmn regulation by retinoid was in human neuroblastoma cells, where it is up-regulated by atRA (Merrill et al., 2004). Here, we report that Clmn expression is lost in the neuroepithelium adjacent to the developing somites in embryos that are severely deficient in vitamin A. Furthermore, the expression of Clmn can be rescued by the addition of atRA to these embryos. In the early embryo, atRA is produced in the paraxial mesoderm with a rostral boundary at the level of the first somite, and later it is generated by the somites. Several groups have shown that the atRA-synthesizing enzyme, Raldh2, is expressed in these regions, whereas it is not expressed by the neuroepithelium proper (Niederreither et al., 1997; Berggren et al., 1999; Swindell et al., 1999; Molotkova et al., 2005). Taken together, this supports the idea that Clmn expression in the neuroepitheum is maintained by retinoid signaling from the underlying somitic mesoderm. Because atRA induction of Clmn mRNA is eliminated by the transcriptional inhibitor, actinomycin-D, but is not altered by a protein synthesis inhibitor, Clmn appears to be directly regulated by atRA and its receptors. Thus, Clmn is found in developing embryos before the time that neurons differentiate, and vitamin A is required to maintain expression specifically in the neuroepithium adjacent to the developing somites.

Several regions in which CLMN immunostaining is found at a later stage (E18.5) in the rat embryo correspond to sites in which neuronal progenitors have recently exited the cell cycle and undergone differentiation and migration. In the developing eye, retinal neurogenesis occurs in a fixed histogenetic order during late embryonic and early postnatal life, and at E18.5 in the rat, both retinal progenitor cells and differentiated neurons are present. The nuclei of the retinal progenitor cells move within the outer neuroblastic layer according to the phase of the cell cycle, whereas cell differentiation occurs in the inner layer of the retina. CLMN expression is strong in the differentiating region of the developing retina, and is found in the same region that stains positively for ISL1, a homeodomain protein associated with several specific populations of early differentiating neurons in the central nervous system (CNS), including postmitotic retinal neurons (Galli-Resta et al., 1997). CLMN immunostaining is also found at the outermost edge of the retina, the site where cells would be completing mitosis and generating both cells that will continue to divide and postmitotic daughter cells.

At E18.5, CLMN is also expressed in the nondividing cell layer of the cortex (cortical plate). Neuronal differentiation in the mammalian neocortex peaks at about this time in gestation, whereas non-neuronal cells such as astrocytes and oligodendrocytes develop later (Ross et al., 2003). Neurogenesis in the cortex occurs in an inside out manner. Nuclear migration and division of cells in the ventricular zone is followed by cell cycle exit, generating a neuron that then migrates through the intermediate zone into the cortical plate (Guillemot, 2005). The formation of the CNS involves precise regulation of neural cell progenitor cell proliferation, cell fate specification, and migration and synaptogenesis (Donovan and Dyer, 2005). It will be important to determine whether CLMN plays a role in the regulation of cell cycle exit or cell fate specification, as it is most abundantly expressed in the late embryonic CNS in regions of the retina and cortex where these processes are actively taking place.

The strongest CLMN immunostaining in the adult brain is observed in the pyramidal neurons of the hippocampus (CA1–CA3) and the dentate gyrus, as well as the olfactory bulb and the cerebellum, where the Purkinje cell body and cells in the granular layer, but not the molecular layer, are stained. Co-staining experiments with glial and neuronal markers in regions of the adult brain in which CLMN is expressed show specific immunostaining that is restricted to postmitotic neurons, and more specifically to the cell bodies of these neurons. The detection of CLMN in the hippocampus agrees with a previous report in which Clmn mRNA was found in these same regions using a riboprobe designed to the 3′ end of the full-length transcript (Takaishi et al., 2003). Of interest, vitamin A deficiency in mice affects adult rodent synaptic plasticity between pyramidal neurons in the CA1 and CA3 neurons of the Schaffer Collateral pathway (Misner et al., 2001). Impairment of synaptic plasticity and a deficiency in spatial learning is also observed in RARβ knockout mice (Chiang et al., 1998). Furthermore, retinoid activity is present in the adult hippocampus as evidenced by several in vivo RA reporter systems (Misner et al., 2001; Luo et al., 2004). Whether CLMN plays a role in any of these events remains to be determined.

In summary, Clmn is expressed in embryos from the late headfold stage (presomite stage) onward, and is most abundant in tissues of the central and peripheral nervous systems. At the early somite stage, vitamin A is required to maintain the normal expression of Clmn in the neuroepithelium adjacent to the somitic mesoderm. In E18.5 embryos, CLMN is detected in regions where newly differentiated neurons are found, including the neural retina and the cortical plate; and in the adult brain, CLMN is most highly expressed in the neuron cell bodies of the hippocampus, cerebellum, and olfactory bulb.

EXPERIMENTAL PROCEDURES

Retinoids

All-trans retinoic acid (atRA) was obtained from Spectrum Chemical Co. (New Brunswick, NJ) and was deemed greater than 99% pure by reverse-phase high-performance liquid chromatography (Motto et al., 1989).

Cell Culture and Transfection

The murine neuroblastoma cell line Neuro2A (N2A, ATCC, Manassas, VA) was grown at 37°C with 5% CO₂ in MEM with 2 mM L-glutamine, 1.5 g/L sodium bicarbonate, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, and 10% fetal bovine serum. N2A cells were transfected with Fugene 6 (Roche, USA) according to the manufacturer's guidelines. SH-SY5Y human neuroblastoma cells were cultured as previously described (Clagett-Dame et al., 1993).

Actinomycin-D/Cycloheximide Experiments

SH-SY5Y cells were exposed to atRA (10⁻⁶ M) or vehicle (0.2% ethanol) for 4 h in the absence or presence of actinomycin-D (5 μg/ml) or cycloheximide (10 μg/ml) after which time poly(A)⁺ RNA was isolated and reverse transcribed using AMV enzyme (Promega, Madison, WI) and random hexamers. Quantitative PCR was performed using the LightCycler faststart DNA master SYBR green 1 kit (Roche, Indianapolis, IN) using primers for human Clmn (NCBI accession no. AB047979; upstream 5′-GTGAAAGACCAGAGGA AGGCTA-3′ and downstream 5′-TGA TGCGAACAAAAGTGGAT-3′) as well as human GAPDH (NCBI accession no. AB062273; upstream 5′-GCTTAG CACCCCTGGCCAAGGTCA-3′ and downstream 5′-CATGTGGGCCATG AGGTCCACCAC-3′) to normalize for input cDNA. The identity of polymerase chain reaction (PCR) products was verified by sequence analysis.

Generation of Embryo and Brain Samples

Female rats (Harlan Sprague-Dawley, Madison, WI) were maintained on laboratory chow diet (Harlan Teklad, USA) and mated with male rats of the same strain. Vitamin A-deficient (VAD) embryos were generated and atRA rescue studies were performed as described previously (White et al., 1998, 2000). Briefly, the VAD group was fed the vitamin A-free purified diet supplemented with 1.5 μg atRA/g diet starting at embryonic day (E) 0.5; this amount of atRA is close to the minimum needed to support embryonic survival (White et al., 2000). For atRA rescue studies, the 250 μg atRA/g diet was fed. The VAS group received a daily oral bolus of retinyl palmitate (500 IU/day) starting at E0.5. For adult brain studies, female rats (Harlan Sprague-Dawley) were maintained on a laboratory chow diet and euthanized at approximately 100 days of age. All animals were maintained according to conditions under a research protocol approved by the Institutional Animal Care and Use Committee at the University of Wisconsin-Madison.

Riboprobe Generation and In Situ Hybridization

The polymerase chain reaction (PCR) was used to amplify the rat Clmn cDNA using the following primers: upstream 5′-TTTTCGGTTGAACAA CATAGCG-3′ and downstream 5′-ACAGACGAAGACCTTGTCAGGC-3′ to generate a 788-bp product. The product was subcloned into the Zero Blunt TOPO PCR vector (Invitrogen), and was sequenced to verify the identity of the insert cDNA. For whole-mount in situ hybridization, embryos were fixed in 4% PFA overnight at 4°C and studies were performed as previously described (Merrill et al., 2004; See et al., 2008). A subset of embryos after staining were embedded in 3% agarose in phosphate buffered saline (PBS) and sectioned at a thickness of 40 μm using a Leica VT1000S vibrating microtome.

Antibody Production and Immunoblotting

A polyclonal antibody was generated in rabbit to a 15-mer CLMN peptide (VQLRNAADLDDRRNR; amino acids 947-961, NCBI accession no. AB047978) conjugated to keyhole limpet hemocyanin. To test antibody specificity, the full-length murine Clmn cDNA (3.0 kb) was amplified from reverse transcribed cDNA from the murine neuroblastoma Neuro2A (N2A) cell line with upstream 5′-TCCATGGCTGCTCAGGAGT-3′ and downstream 5′-GAGCCGGCTGACAT CAAGCT-3′ primers and cloned into the EcoRI-ScaI site of pIRES-hrGFP-1α (Stratagene) and sequenced, and encodes for a protein with a predicted molecular weight of 112 kDa containing a 3x C-terminal flag tag (CLMN-3xFl).

N2A cells were transfected with CLMN-3xFl or pIRES, lysed in buffer containing 0.05 M Hepes (pH 7.4), 0.15 M NaCl, and 1% NP40 with protease inhibitors (5 μg/ml leupeptin, 25 mM benzamidine, and 1 mM phenhylmethyl sulfonyl fluoride), and protein was analyzed by BCA assay (Pierce). The sample was mixed 1:1 in 2× Laemmli sample buffer (60 mM Tris HCl, pH 6.8, 5% sodium dodecyl sulfate (SDS), 5% β-mercaproethanol, 10% glycerol, and 0.05% bromophenol blue). Protein (5 μg) was resolved on a 10% SDS-polyacrylamide gel, and proteins were transferred to a nitrocellulose membrane (Bio-Rad). Membranes were blocked with 8% nonfat dry milk powder in PBS with 0.2% Tween-20 for CLMN antibody or nonimmune serum and 5% nonfat dry milk powder in PBS with 0.1% Tween-20 for the M2 anti-flag antibody (Sigma, 1:2,000, Catalog F1804) and the anti–α-tubulin antibody (Clone B-5-12, Sigma, 1:10,000) for 1 hr at room temperature, followed by primary antibody incubation in blocking buffer for 1 hr. Membranes were washed three times in PBS before incubating with secondary antibodies conjugated to horseradish peroxidase in blocking buffer for 1 hr. Antigen-antibody conjugates were visualized by SuperSignal West Pico Chemiluminescent kit (Pierce) and exposed to X-ray film. Rat and mouse brain tissues were lysed in homogenization buffer (1 mg of tissue: 4 μl of buffer; 20 mM Tris-HCl, pH 8.0, 2 mM ethylenediaminetetraacetic acid, 2 mM ethyleneglycoltetraacetic acid, 10 mM Na₂HPO₄, 25 μg/ml soybean trypsin inhibitor, 5 μg/ml leupeptin, 2 mM dithiothreitol, 25 mM benzamidine, and 1 mM phenylmethylsulfonyl fluoride) using the Tekmar Tissuemizer (Tekmar Co.) and lysed in 10% SDS Buffer (62.5 mM Tris-HCl, pH 6.8, 10% glycerol, 10% SDS, 0.03% bromophenol blue, and 10% β-mercaptoethanol).

Immunohistochemical and Immunofluorescence Studies

Embryos (E18.5) and adult brain tissues were incubated in Dent's fixative for 1 week. Embryos were taken through ethanol and xylene washes, and transferred to xylene: paraffin (1:1) overnight at 60°C. Samples were then incubated in 100% paraffin at 60°C for 7 hr with three exchanges, and, during the final incubation, were placed under 15 mmHg pressure for 1 hr at 60°C. Paraffin blocks were sectioned at 10 μm in series on a Microm HM325 microtome.

Tissue sections were deparaffinized through a series of xylene washes to PBS and incubated in 0.3% H₂O₂ /1× PBS for 30 min at room temperature to quench endogenous peroxidase activity. Sections were blocked with 2% goat serum, 1% bovine serum albumin (BSA), 0.1% cold fish skin gelatin, 0.05% Tween-20, and 0.05% NaN₃ in PBS for 30 min in a humidified chamber. Samples were then incubated with primary antibodies in 1% BSA, 0.1% fish skin gelatin, and 0.05% thiomerosal in PBS overnight at 4°C in a humidified chamber. Primary antibodies included anti-calbindin D28K (Clone CB-955, Sigma, mouse IgG, 1:1,000), anti-Ki67 (Clone B56, BD Pharmingen, 1:400), anti-islet 1 (clone 40.2D6 supernatant, Developmental Studies Hybridoma Bank), rabbit IgG (Sigma, 1:200), and anti-CLMN (rabbit polyclonal immune serum). The Vectastain ABC kit (mouse IgG or rabbit IgG) and 3,3′-diaminobenzadine tetrahydrochloride with metal enhancer was used to evaluate antibody binding.

For immunofluorescence studies, tissue sections were prepared as above except the hydrogen peroxide incubation was eliminated. Primary antibodies included anti-GFAP (Clone G-A-5, Sigma, 1:1,000), anti-NeuN (Clone A60, Millipore, 1:1,000), anti-MAP2 (Clone AP-20, Sigma, 1:100) and anti-CLMN (rabbit polyclonal immune serum). After overnight primary antibody incubation, sections were incubated in Alexa-conjugated secondary antibodies (Alexa Fluor 594 goat-anti-mouse IgG [Invitrogen, 1: 200]; Alexa Fluor 488 goat anti-rabbit IgG [Invitrogen, 1:200]).

Acknowledgements

We thank Jamie Ahrens and Vivian Poon for their contribution to the cell biological experiments. We also thank Mary Kaiser for assistance in the generation of embryo samples and in situ hybridization studies, and Jean Chung for embryonic retinoid analysis.