Cabin1 (Calcineurin binding protein 1, or cain) is a gene that is essential for mammalian development (Esau et al., 2001). In mature animals, Cabin1 acts as a calcium-dependent repressor of calcineurin (protein phosphatase 3) in the CNS (Lai et al., 1998, 2000) and a repressor of both calcineurin and MEF2 (MADS/myocyte enhancer factor 2) in activated T-cells (Youn et al., 1999; Jang et al., 2007a). The MEF2 proteins were originally described as transcription factors required for muscle differentiation, but have since been found to play diverse roles in the CNS (Shalizi and Bonni, 2005 review). Calcineurin phosphatase activity is required for NFAT (nuclear factor of activated T cells) -mediated transcription. While the ability of Cabin1 to repress these transcriptional pathways has not been tested in neurons, calcineurin-NFAT and MEF2 transcriptional pathways have been implicated in differentiation (Okamoto et al., 2000), neuronal survival (Mao et al., 1999; Groth and Mermelstein, 2003; Shalizi et al., 2003; Pazyra-Murphy et al., 2009), and synaptogenesis (Yoshida and Mishina, 2005; Flavell et al., 2006; Shalizi et al., 2006).
Based on the requirements of calcineurin-mediated and MEF2 transcription for normal nervous system development, including such calcium-dependent processes as synaptic remodeling (Yoshida and Mishina, 2005) and activity-dependent survival (Mao et al., 1999), we predict Cabin1 may play important roles in the development of the nervous system by repressing the transcriptional activity of calcineurin-NFAT and MEF2. Here, we establish that Cabin1 is prominently expressed in the developing nervous system. Furthermore, in particular CNS tissues, Cabin1 expression is coincident with MEF2 and calcineurin. While the fatal developmental defects in Cabin1 null mice (Esau et al., 2001) have not been characterized, the developmental expression of Cabin1 suggests a major contribution could be attributed to CNS defects.
In the present study we isolated Cabin1, which we identified in a screen for genes expressed in neurons undergoing axon growth and synaptogenesis. Therefore, we predicted that Cabin1 functions in neuronal development and maturation. We mapped the spatial and temporal profile for Cabin1 during CNS development. Cabin1 expression is strong during early stages of neuronal development and is down-regulated as the CNS matures. We also examined the expression patterns of known Cabin1 binding partners, calcineurin and MEF2, in the developing nervous system to identify areas of potential interaction. We specifically identified two tissues where Cabin1, calcineurin, and MEF2 overlapped: the olfactory system and the cerebellum. We expect further analysis of these regions to provide more insight into the role of Cabin1 during development.
Isolation of Cabin1 cDNA
We identified Cabin1 in the developing nervous system from a cDNA library created to identify sequences enriched in neurons undergoing axon growth. The 283 base pair sequence isolated from the library represents part of a hypothetical protein-coding sequence in the zebrafish genome (Zebrafish Ensembl database, v8.0). The hypothetical zebrafish protein shows strong sequence similarity to mouse and human Cabin1. Full-length zebrafish Cabin1 mRNA sequence was obtained by rapid amplification of cDNA ends (RACE). The 5′ RACE did not return new sequence, but confirmed the predicted 85 bp of the 5′ noncoding region (XM_694965). The 3′ RACE returned 1.1 kb of new sequence (GenBank: GU733827). The combined sequences yielded a full-length zebrafish cDNA sequence of 7.8 kb with 6.6 kb of protein-coding sequence and 1.2 kb of 5′ and 3′ untranslated region (UTR). Molecular phylogenetic analysis demonstrates that zebrafish Cabin1 is very similar to its mammalian homologs (accession numbers are listed in Supp. Table S1, which is available online) and is well conserved among vertebrates (Fig. 1).
Cabin1 Amino Acid Sequence Is Conserved Across Species
To identify regions of Cabin1 protein sequence conservation across species, amino acid sequence alignment was performed using Multalin (Corpet, 1988). High sequence conservation (70% amino acid similarity) exists across mammals (human, mouse, rat), chicken, and zebrafish (Fig. 2). In addition to regions of identical amino acid sequence, many nonidentical amino acids are considered conservative substitutions, so protein structure should remain similar across species. Protein-binding sites for MEF2, calcineurin, and calmodulin are highly conserved. In addition, there are regions of unknown function throughout the protein sequence that are also highly conserved across all species examined. Thus, much of Cabin1 function should be conserved across species. Its function as a calcineurin and MEF2 repressor in activated T-cells has been well-documented (Sun et al., 1998; Youn et al., 1999; Jang et al., 2007a, b), but its role in developing systems has not been examined.
Cabin1 Is Expressed in the Developing Nervous System
To determine the spatial and temporal pattern of Cabin1 expression during development, we performed in situ hybridization (ISH). The ISH probe sequence is listed in supplementary materials (Supp. Table S2). We examined five time points during the first 3 days of development (16, 18, 24, 48, 72 hours postfertilization [hpf]) when neurons in different regions of the brain are known to differentiate, extend axons, and form connections. In the forebrain, midbrain, and hindbrain, the very first differentiated neurons appear around 15–18 hpf (Hanneman et al., 1988; Ross et al., 1992). We did not detect Cabin1 expression at the two earliest time points (data not shown), indicating that Cabin1 may be involved in later stages of differentiation. We first detected Cabin1 expression in the nervous system at 24 hpf in the midbrain and midbrain/hindbrain boundary, with weaker expression in the diencephalon, cerebellum, and hindbrain (Fig. 3A). Sense controls are shown in supplementary materials (Supp. Fig. S1). Taken together with previous studies of differentiation and axon growth in the developing zebrafish brain, these results suggest that Cabin1 expression in the brain occurs during a period when differentiated neurons are growing axons (Hanneman et al., 1988; Ross et al., 1992).
The number of neurons undergoing differentiation and axonogenesis in the brain has been previously shown to increase between 24 and 48 hpf (Kuwada et al., 1990; Ross et al., 1992). Cabin1 expression also expands during this period. At 48 hpf, Cabin1 expression is observed in various areas within the brain as well as outside the nervous system in the pharyngeal arches (Fig. 3B). Within the nervous system, expression is strongest in the olfactory epithelium and cerebellum (Fig. 3B). Expression is also observed for the first time in the retina, and continues to be expressed in other regions of the nervous system including the diencephalon, midbrain, and hindbrain (Fig. 3B). The increase in Cabin1 expression in the nervous system correlates with an increase in the number of differentiated neurons undergoing axon growth.
Cabin1 expression remains widespread in the brain at 72 hpf, but is diminished in intensity. Expression in the nervous system is still most evident in the olfactory epithelium and cerebellum (Fig. 3C), where neurons are undergoing activity-dependent survival and maturation (Yoshida and Mishina, 2005; Volkmann et al., 2008). In contrast, expression appears diminished in the midbrain and hindbrain (Fig. 3C). In sectioned embryos, we can also detect expression in the thalamus and preoptic area (Fig. 3D). In the retina, expression is no longer uniform across the retinal ganglion cell layer as at 48 hpf (Fig. 3B), but expression is still detected in the marginal zone of the retina (Fig. 3E) where new retinal neurons are added over the life of the animal (Easter and Nicola, 1996). Outside the nervous system, Cabin1 expression persists in the pharyngeal arches (Fig. 3C), but also appears in the fin buds, hatching gland, and skin of the head (data not shown). Thus, as Cabin1 expression within the maturing nervous system begins to diminish, its expression outside the nervous system expands.
In summary, Cabin1 expression appears to correlate with the maturation of the neurons in the brain. Specifically, Cabin1 expression is apparent in the forebrain, midbrain, hindbrain, and retina at times previously correlated with the presence of differentiated neurons in the process of forming mature connections (Table 1). In the olfactory and retinotectal systems in particular, Cabin1 expression correlates with times when neurons in these regions have previously been determined to undergo synaptogenesis and synaptic refinement (Stuermer, 1988; Yoshida and Mishina, 2005). To identify possible molecular pathways with which Cabin1 might interact during these periods of CNS development, we compared the developmental gene expression patterns of known interacting partners calcineurin and MEF2.
Relative expression is indicated by pluses: strong expression (+++), moderate expression (++), or weak expression (+).
Post Optic Commissure
Ganglion Cell Layer
Inner Nuclear Layer
Peripheral Nerve Ganglia
Cabin1 and Calcineurin Are Expressed in Overlapping Regions in the Developing CNS
To identify where Cabin1 and calcineurin could potentially interact, we performed in situ hybridization for genes encoding each of the calcineurin subunits at stages when Cabin1 mRNA was present (24, 48, and 72 hpf). The calcineurin catalytic subunits are encoded by the genes PPP3CA, PPP3CB, and PPP3CC, while the regulatory subunits are encoded by the genes PPP3R1 and PPP3R2. Probe sequences are given in Supp. Table S2. Overlapping expression of Cabin1 and the calcineurin genes is most prominent in three regions at all ages examined: the olfactory system, cerebellum, and hindbrain. Comparisons of 72 hpf animals illustrate overlapping expression in these structures (Fig. 4). Cabin1 expression overlaps with that of PPP3CA and PPP3CB in the olfactory epithelium (Fig. 4A–C) and all of the calcineurin genes overlap with Cabin1 in the cerebellum and hindbrain (Fig. 4). Regions with overlapping gene expression are regions of potential functional interaction between Cabin1 and calcineurin.
Each calcineurin gene is expressed in a distinct pattern that overlaps with Cabin1 in the nervous system and in non-neuronal tissues (Supp. Fig. S2). Of interest, PPP3R2 is robustly expressed throughout the developing nervous system (Supp. Fig. S2) and this is the first report of its neuronal expression in any species. The spatial and temporal gene expression patterns for all five calcineurin genes show overlap with that of Cabin1 in specific locations in the CNS (Fig. 4, Supp. Fig. S2, and Supp Table S3).
Cabin1 and MEF2 Are Expressed in Overlapping Regions in the Developing CNS
To determine the likelihood of a Cabin1-MEF2 interaction during development, we performed in situ hybridization for all five MEF2 genes, and compared their expression patterns with that of Cabin1. Probe sequences are listed in Supp. Table S2. Overlapping expression of Cabin1 and the MEF2 genes is most prominent in the cerebellum and hindbrain (Fig. 5). Cabin1, MEF2A, MEF2B, and MEF2Cb are also expressed in overlapping patterns in the diencephalon (Fig. 5). The expression patterns of Cabin1 and MEF2 overlap in specific regions in the CNS representing areas where Cabin1 could potentially regulate MEF2 activity in the developing CNS.
Each MEF2 gene is expressed in a different pattern in the developing nervous system and in non-neuronal tissues and overlaps with Cabin1 expression (Fig. 5, Supp. Fig. S3). Unlike mammals, which have a single MEF2C gene, zebrafish possess two MEF2 genes: MEF2Ca and MEF2Cb, for which we probed individually (Fig. 5, Supp. Fig. S3). Expression of all five MEF2 genes is summarized in supplementary materials (Supp. Table S4). The five MEF2 genes have a broader spatial distribution than Cabin1, but each partially overlaps spatially and temporally with Cabin1 during embryonic and early larval development during periods of neuronal differentiation.
Cabin1 has an essential role during normal development, because mice with a null mutation in the gene die at E12.5 (Esau et al., 2001). While several developmental defects could cause premature death, the developmental expression of Cabin1 suggests CNS defects could provide a major contribution. The repressive action of Cabin1 on calcineurin in both the mature CNS (Lai et al., 2000) and immune system (Jang et al., 2007a) and on MEF2 in the immune system (Sun et al., 1998; Youn et al., 1999; Han et al., 2003; Pan et al., 2005; Jang et al., 2007b) leave open the possibility that Cabin1 repression of these pathways also plays a broad role during the development of both tissues. In T-cells, Cabin1 acts as a repressor of calcineurin and MEF2 under normal conditions, but repression is alleviated in response to elevated levels of intracellular calcium (Youn et al., 1999). In the present study, we identified and cloned the zebrafish homolog of Cabin1 and found that it is expressed in the CNS during periods of neuronal differentiation and synaptogenesis. We also compared the spatial and temporal expression patterns of Cabin1 with those of the calcineurin and MEF2 families of genes. Because these proteins have roles in differentiation (Okamoto et al., 2000; Shalizi et al., 2003) and synaptogenesis (Yoshida and Mishina, 2005; Flavell et al., 2006), Cabin1 may function in the CNS through regulation of calcineurin and MEF2.
To determine the likelihood of Cabin1–calcineurin interactions during development, we compared Cabin1 gene expression with that of the calcineurin gene family. All five genes encoding calcineurin subunits are expressed in the developing CNS. Previously, only limited expression analysis was available for the calcineurin genes during development (Gerber et al., 2003; Eastwood et al., 2005). While calcineurin has many roles in the CNS, calcineurin-NFAT activity is required for synaptic maturation in the olfactory system (Yoshida and Mishina, 2005). Cabin1 expression occurs in the olfactory epithelium before and during synaptic maturation, which peaks between 60 and 84 hpf (Yoshida and Mishina, 2005). By inhibiting calcineurin, Cabin1 may regulate synapse maturation through transcriptional repression in cells that have not yet established robust connections with their targets. Once active synapses have been made, they would presumably be able to overcome Cabin1's calcium-sensitive repression. Thus the spatial and temporal pattern of Cabin1 expression indicates it could have a functional interaction with calcineurin.
Calcineurin–Cabin1 interactions occur in the cytoplasm, but cellular fractionation studies have reported that Cabin1 is also present in the nucleus (Sun et al., 1998). Therefore, Cabin1 functions as both a cytoplasmic and a nuclear protein. In the nucleus, Cabin1 represses MEF2 transcriptional activity in a calcium-dependent manner. The MEF2 proteins were originally identified as a family of transcription factors necessary for cardiac, skeletal, and smooth muscle differentiation, but have since been isolated from nervous tissue where they are involved in neuronal differentiation, growth, survival, and apoptosis (Mao et al., 1999; Okamoto et al., 2000; Flavell et al., 2006; Pazyra-Murphy et al., 2009; Shalizi and Bonni, 2005, for review).
To identify potential sites of interaction between Cabin1 and MEF2, we compared their spatial and temporal patterns of expression. Cabin1 and MEF2 are expressed in the developing cerebellum when granule cells are differentiating, migrating, and establishing circuits (48–72 hpf, Volkmann et al., 2008). MEF2 promotes survival of newly differentiated granule cells (Okamoto et al., 2000; Shalizi et al., 2003). Cabin1 may be involved in neuronal differentiation and survival by means of its repressive interactions with MEF2. Activity-dependent cerebellar granule cell survival is also dependent on MEF2 transcription (Mao et al., 1999). By repressing MEF2 when intracellular calcium is low, Cabin1 may function to prevent the survival of cells that are inactive due to limited or inappropriate connections. Only cells that are forming active connections will alleviate the calcium-sensitive Cabin1-mediated repression and survive. Cabin1 may also regulate synapse formation in the cerebellum by acting as a dual repressor of both MEF2 and calcineurin. Cabin1 may directly repress MEF2, which is also required for postsynaptic differentiation of granule cells. Cabin1 may also repress calcineurin-mediated dephosphorylation of MEF2, which increases the affinity of MEF2 for DNA (Mao and Weidmann, 1999) and promotes MEF2 acetylation in granule cells (Shalizi et al., 2006). Given the variety of roles for MEF2 in cerebellar development and the presence of Cabin1 in the developing cerebellum, we predict Cabin1 may have multiple roles in cerebellar development.
The studies examining the effects of Cabin1 on transcription have focused on the immune system, but the same interactions may occur in the nervous system since the same transcriptional pathways—calcineurin-NFAT and MEF2 pathways—are also functional in the CNS (Mao et al., 1999; Okamoto et al., 2000; Groth and Mermelstein, 2003; Cano et al., 2005; Flavell et al., 2006; Nguyen and Di Giovanni, 2008). Calcineurin activity can also directly influence the transcriptional activity of MEF2. Calcineurin activity regulates MEF2-dependent and MEF2-independent processes in the nervous system (Beg and Scheiffele, 2006). Cabin1 could potentially serve as a synergistic repressor of calcineurin-MEF2 activities in the CNS in addition to repressing each pathway individually.
This is the first report of Cabin1 expression during CNS development. Cabin1 is expressed in many regions of the developing CNS and expression decreases with CNS maturation. Cabin1 expression is most notable in the olfactory system and cerebellum, where the timing of expression appears to coincide with activity-dependent processes like synaptogenesis and activity-dependent survival. As a calcium-responsive inhibitor, Cabin1 may regulate these processes through interactions with known interacting partners calcineurin and MEF2. Calcineurin regulates synaptic remodeling in olfactory sensory neurons and MEF2 regulates survival of cerebellar granule cells. Both of these processes are calcium-dependent. Cabin1 expression patterns, known repressive actions on calcineurin and MEF2, and responsiveness to calcium signaling indicate it may be an important regulator of activity-dependent processes in CNS development.
Maintenance of Fish
Zebrafish husbandry and all experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC). Zebrafish colonies were maintained as previously described (Westerfield, 2000). Adult fish were housed in recirculating rack systems (Aquatic Habitats, Apopka, FL) at 28.5°C on a 14-hr light, 10-hr dark cycle, and fed twice daily with Artemia and/or Zeigler Adult Zebrafish Complete Diet (VWR, West Chester, PA). Fertilized eggs were generated by natural spawning. Embryos from the EkkWill strain (Ruskin, FL) were used for in situ hybridization studies.
BLAST and RACE
The Clontech polymerase chain reaction (PCR) -select cDNA subtraction kit was used to create a subtraction library using cells isolated from a GAP43/green fluorescent protein (GFP) transgenic reporter line (Udvadia, 2008; manuscript in preparation). A 283 base pair sequence isolated from this library was compared by means of nucleotide BLAST with the National Center for Biotechnology Information nucleotide database (NCBI; http://www.ncbi.nlm.nih.gov/) and Ensembl zebrafish Genome Browser (http://Ensembl.org/Danio_rerio/), and cross-species BLAT (BLAST-like alignment tool, Kent, 2002). These overlapping fragments were combined and verified using standard PCR techniques to assemble the zebrafish Cabin1 gene. 5′ and 3′ RACE (rapid amplification of cDNA ends) were performed on cDNA from 48 hpf embryos to isolate the full-length zebrafish Cabin1 cDNA. RACE data was used to supplement the sequences obtained from BLAST results. To identify regions of amino acid sequence conservation across species, protein sequence alignment was performed using a Blosum62 symbol weight comparison table in Multalin (Corpet, 1988; http://bioinfo.genotoul.fr/multalin/).
Sequences corresponding to the zebrafish calcineurin and MEF2 genes were obtained from GenBank sequences (Supp. Table S1). Zebrafish calcineurin catalytic subunit homologs were acquired by using BLink (NCBI precomputed BLAST Link) for the mouse and human protein sequences and confirmed in BLAST searches between both NCBI and Ensembl databases. PPP3CA and PPP3CB homologs were designated as two hypothetical zebrafish genes that are very similar at the protein level, but reside on different chromosomes. PPP3R1 and PPP3R2 were derived from hypothetical or predicted coding sequences returned as a result of searching translated nucleotide (tBLASTn) databases using the mouse and human protein sequences. Accession numbers are listed in supplementary materials (Supp. Table S1).
Sequence alignment of Cabin1 protein sequences obtained from GenBank were created using ClustalW2 (Larkin et al., 2007; http://www.ebi.ac.uk/Tools/clustalw2/index.html). The alignment was assembled into a phylogenetic tree by the neighbor joining method (Saitou and Nei, 1987) using NJ Plot (Perriére and Gouy, 1996; http://pbil.univ-lyon1.fr/software/njplot.html). Cabin1 sequences from zebrafish (Danio rerio), human (Homo sapiens), mouse (Mus musculus), rat (Rattus norvegicus), frog (Xenopus tropicalis), and chicken (Gallus gallus) were compared with the Florida lancelet (Branchiostoma floridae) protein obtained using BLink (NCBI) with the mouse Cabin1 sequence. Accession numbers for all homologs are included in the Supplementary Material (Supp. Table S1).
In Situ Hybridization
The Cabin1 probe was designed to include the region isolated from the subtraction library and surrounding sequences. Primers were designed to isolate fragments for probe synthesis corresponding to each of the five MEF2 genes and each of the five calcineurin genes. Plasmids used for in situ probe synthesis were generated by PCR under standard conditions from cDNA made from 48 hpf fish and sequenced. Probe sequences are given in supplementary materials (Supp. Table S2). Amplicons were cloned into pGEM-T Easy vectors (Promega). Digoxigenin (DIG) -labeled riboprobes were generated from PCR amplified templates containing T7 or SP6 RNA polymerase binding sites as previously described (Quiring et al., 2004). Template DNA was subsequently removed with RNase-free DNase treatment. DIG incorporation was determined using the dot blot method on positively charged nylon membrane as recommended by the manufacturer (Roche).
In situ hybridization experiments were carried out as previously described (Thisse et al., 1993) on whole mount embryos/larvae fixed in 4% paraformaldehyde at 16 hpf, 18 hpf, prim 5 (24 hpf), long pec (48 hpf), and protruding mouth (72 hpf) stages. Probe amounts (∼ 1 ng/μl) and hybridization temperatures (60° or 55°) were empirically determined. GAP-43 antisense probes were used as positive controls, while sense probes served as negative controls. Stained embryos were fixed for 30 min in 4% paraformaldehyde, rinsed briefly, and immersed in glycerol in preparation for digital imaging. Some embryos were manually de-yolked following glycerol immersion. Images were captured on a stereoscope (Olympus SZX12, Center Valley, PA) equipped with a cooled CCD color camera (Olympus DP70). Images were processed, cropped, and assembled into montages using Photoshop CS (Adobe, San Jose, CA).
Representative whole-mount stained embryos were cryoprotected in 20% sucrose in phosphate buffered saline (PBS) and embedded in flexible molds in embedding medium (1:2 20% sucrose and Histoprep) for cryosectioning. Twenty-micrometer sections were mounted on slides and allowed to dry overnight. Sections were washed in 1× PBS to remove embedding medium, and mounted immediately in Vectashield mounting medium with DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride; Vector Laboratories, Burlingame, CA) and coverslipped.