Calcium (Ca2+) is essential for the regulation of a wide variety of biological processes such as contraction, secretion, fertilization, cell proliferation, apoptosis, learning, and memory (Berridge et al.,1998). Calmodulin (CaM) is primarily known as a Ca2+ binding protein in all eukaryotic cells and plays a critical role as a Ca2+ mediator (Cheung, 1980; Klee et al.,1980). CaM is highly abundant in the vertebrate central nervous system (CNS) and spatially enriched in postsynaptic membranes, postsynaptic densities, and synaptic vesicles (Sola et al.,2001). Cellular Ca2+ signaling induced by CaM requires an interaction between CaM and specific proteins. Among the several CaM binding proteins, the Ca2+/CaM-dependent protein kinase II (CaMK-II) is important in numerous neural functions. CaMK-II is highly conserved from Drosophila to human. This well-conserved serine/threonine protein kinase CaMK-II is encoded by at least four genes (α, β, γ, and δ) in vertebrates (Tombes et al.,2003). In contrast to that in vertebrates, Drosophila CaMK-II is encoded by a single gene (Griffith and Greenspan,1993). In the brain, the α and β isoforms of CaMK-II constitute between 0.25% and 2% of the total protein and is especially abundant in the hippocampus (Klee,1991; Braun and Schulman,1995; Sola et al.,2001). In contrast to the α and β isoforms, γ and δ are widely expressed in various tissues (Tobimatsu and Fujisawa,1989). Other CaMKs abundant in neurons include members of the CaMK cascade: CaMKI, CaMKIV, and CaMKK (CaMK kinase; Means,2000). A previous report showed the exclusive expression of CaMK-II in the CNS in Drosophila embryos (Rachidi et al.,1999). Another report also demonstrated that a CaMK-II–like gene, 1G5, was predominantly expressed in the rat brain, including olfactory bulb, cerebral cortex, caudate nucleus, hippocampus, hypothalamus, and midbrain, and mild expression was observed in pons, medulla, and spinal cord but was not detected in the non-neural tissues (Godbout et al.,1994). Recently, Thisse et al. reported the zebrafish CaMK-IID expression pattern to the ZFIN (unpublished data; http://zfin.org/cgi-bin/webdriver?MIval=aa-xpatview.apg&OID=ZDB-XPAT-041008-109). Of interest, the expression of CaMK-IID began at the late segmentation stage and became confined in brain tissues throughout development. However, detailed studies of the expression pattern of CaMK or CaMK-like genes during vertebrate embryonic development have not been conducted.
Here, we report the isolation and expression of a CaMK-like gene, the zebrafish homologue of mammalian 1G5 (1G5z). 1G5z expression starts long after primary neurogenesis begins and is restricted to the particular neuronal cells of the developing CNS. At 3 days postfertilization, 1G5z expression is further restricted in the brain hierarchy, particularly in the eye. The number of 1G5z-positive sensory neurons and interneurons are greatly enhanced in the mindbomb mutant embryos at 24 hpf. However, 1G5z transcripts in the trigeminal ganglion and trunk sensory neurons are completely abolished in neurogenin1 morphants. By using the bromodeoxyuridine (BrdU) incorporation assay, we further identified that 1G5z-expressing cells were terminally differentiated sensory and interneurons. The application of 1G5z as a marker gene for neurons and brain is discussed.
RESULTS AND DISCUSSION
Molecular Cloning and Characterization of 1G5z
We isolated a partial cDNA through the in situ screening for genes with tissue-specific expression patterns during zebrafish embryonic development. The isolated cDNA fragment encodes a novel member of vertebrate 1G5 homologue. In an attempt to obtain a complete open reading frame (ORF), we initially searched the zebrafish homologue of 1G5 gene by using Ensemble Zebrafish Genome Server Genescan (http://www.ensembl.org/Danio_rerio) and NCBI Blast the Zebrafish Genome EST Blast Search (http://www.ncbi.nlm.nih.gov/genome/seq/DrBlast.html) programs. As a result, we found several DNA fragments from the database enough to encompass the full ORF of zebrafish 1G5 homologue. After connecting the DNA fragments, we could deduce putative ORF. The expected ORF was obtained by reverse transcriptase-polymerase chain reaction (RT-PCR) and verified by sequencing analysis. The zebrafish homologue of 1G5 shared high amino acids sequence homology with the proteins from several species, including rat (Godbout et al.,1994), mouse (Strausberg et al.,2002), and human (Fig. 1A), and was designated as 1G5z (1G5 of zebrafish; accession no. AY450361). 1G5z encoding 436 amino acids contains major three domains: (1) a serine/threonine kinase domain (catalytic domain) between residues 24 and 286, (2) a calmodulin-binding domain between residues 309 and 325, and (3) a region between residues 334 and 398 that contains an alanine-rich domain (Fig. 1B). A phylogenetic tree generated from a CLUSTALW multiple amino acid sequence alignment (Thompson et al.,1994) of 1G5z with other mammalian homologues shows that 1G5z shares highest sequence homology with rat 1G5 (Fig. 1C). A previous report suggested that 1G5 could be classified as a subfamily of CaMK-II (Godbout et al.,1994). However, based upon the accumulated sequence data, we identified that 1G5 shares most sequence homology with CaMK-I (Fig. 1C). Rat 1G5 shows 67% sequence homology with rat CaMK-I but shares only 57% of sequence similarity with rat CaMK-II. But the serine/threonine kinase domain and the calmodulin-binding domain of all CaMK members are relatively well matched to each other.
Whereas 1G5z protein shares 83–84% amino acid sequence homology with other vertebrate homologues, the C-terminal alanine-rich domain of 1G5z does not show significant homology, which makes it the 1G5z-specific domain. As we mentioned above, the amino terminal region of 1G5z encompassing the serine/threonine kinase motif and calmodulin-binding domain shows high sequence similarity with CaMK-II, like as mouse 1G5 does (Fig. 1B,C; Godbout et al.,1994). This sequence information suggests that 1G5z might have calmodulin-binding and serine/threonin kinase activities. However, 1G5z lacks the consensus autophosphorylation site (RXXS/T) that is commonly found in other vertebrates and located between the catalytic and calmodulin-binding domains in CaMK-II (Miller et al.,1988; Godbout et al.,1994). Previous studies showed that the autophosphorylation of CaMK-II is important for maintenance of its kinase activity, without prolonged calcium signal stimulation (Hanson and Schulman,1992). This difference between CaMK-II and 1G5/1G5z suggests that 1G5/1G5z is possibly involved in short-term calcium-dependent signaling. In addition, the high sequence homology between zebrafish and its mammalian homologues suggests that they are evolutionarily well conserved and play important roles in vertebrate development.
Spatial and Temporal Expression Patterns of 1G5z
In Drosophila early embryonic stage, CaMK-II gene was expressed throughout the CNS at the first instar larvae (Rachidi et al.,1999). As development progresses, CaMK-II signals were restricted in several ganglia tissues and eye imaginal disc (Rachidi et al.,1999). In mice, Godbout et al. (1994) demonstrated that 1G5 was exclusively detected in brain and eye and not in the non-neural tissues using Northern blot analysis. However, the detailed expression pattern of CaMK and/or CaMK-like genes throughout vertebrate development has not been investigated.
To determine the spatiotemporal expression profile of 1G5z, we performed both RT-PCR and whole-mount in situ hybridization (WISH). RT-PCR analysis showed that the 1G5z transcripts are initially detectable at 24 hpf, and the transcriptional level is constantly maintained until 48 hpf (Fig. 2A) but not detectable before the 18-somite stage (Fig. 2B). In the embryos at 19.5 hpf (20 somites), cells expressing 1G5z are located in the developing neurons along the trigeminal ganglion, hindbrain, and spinal cord (Fig. 2C). Thus, these data suggested that 1G5z is not transcribed until the onset of primary neurogenesis.
At 24 hpf, 1G5z is highly expressed in the neuronal cells at the trunk neural tube (Fig. 2D). Transverse section through the spinal cord resolved a more detailed expression pattern of 1G5z in the interneuron and sensory neuron (Fig. 2G–J). It is of interest to note that 1G5z is differentially expressed through the anterior to posterior trunk spinal cord. 1G5z is expressed in the interneuron and sensory neuron from anterior to medial trunk (Fig. 2G–J). Beyond the yolk sac extension, the expression domain of 1G5z is restricted in the sensory neurons. At this stage, 1G5z is newly transcribed in the brain region (Fig. 2E). Horizontal sections of the 24-hr-old zebrafish brain revealed the 1G5z expression in the diencephalon and trigeminal ganglion (Fig. 2F). We further determined the precise location of 1G5z expression domain with two-color WISH using isl1 (islet1), isl2 (islet2), th (tyrosine hydroxylase), lim1, and krox20 marker genes. is1 family (isl1, isl2), and LIM/homeodomain-type transcription factors are the markers for the cranial and trunk motor neurons as well as caudal Rohon–Beard (RB) neurons (Inoue et al.,1994; Appel et al.,1995; Tokumoto et al.,1995; Segawa et al.,2001). th and krox20 are the markers for dopaminergic neurons (Holzschuh et al.,2001) and the rhombomere (r) 3 and 5 (Oxtoby and Jowett,1993), respectively. In 24 hpf, lim1 mRNA is expressed in the brain and spinal cord (Toyama and Dawid,1997). 1G5z is partly coexpressed with isl-1 in the telencephalon, anterior–ventral diencephalons and trigeminal ganglion (Fig. 2J,K). Of interest, 1G5z-expressing cells in the diencephalon form three different neuronal cell clusters (Fig. 2K–M). To clarify the identity of these cell clusters, we carried out two-color WISH using 1G5z and th as in situ probes. At 24 hpf, th is specifically expressed in the diencephalic catecholaminergic cluster (Holzschuh et al.,2001). Bicolor WISH experiments found that 1G5z and th are expressed in the similar regions of diencephalon but not overlapping with each other (Fig. 2L). It suggests that 1G5z may not be expressed in the diencephalic dopaminergic neurons. 1G5z expression domains in the diencephalon and tegmentum were further analyzed by bicolor WISH with the lim1 marker gene. As shown in Figure 2M, lim1 expression domain completely encompasses 1G5z-expressing cells in the diencephalon and tegmentum. In the hindbrain, 1G5z appeared in the lateral margin of r2, 3, and 4 (Fig. 2N–P). 1G5z expression pattern in the trunk region is also confirmed by two-color WISH using isl1, isl2, and lim1 markers. In the trunk spinal cord, lim1 demonstrates restricted expression pattern in the interneurons, whereas isl1 and isl2 are expressed in the sensory (RB) and motor neurons. As shown in Figure 2Q,R, 1G5z transcripts in the anterior trunk region are specifically detected in the RB neurons and internuerons but not in the ventral motor neurons where isl2 is exclusively expressed. In the posterior trunk and tail region, 1G5z is also colocalized with isl1 in the RB neurons, but the expression in the interneuron gradually fades as it reaches the end of tail (Fig. 2U). Transverse section of spinal cord after staining with 1G5z and isl1 further clarified RB and interneuron-specific expression of 1G5z (Fig. 2J).
By 35 hpf, 1G5z transcripts appear in the anterior neural tube as well as in the hindbrain, tegmentum, hypothalamus, and telencephalon (Fig. 3A,B). It is worth noting that 1G5z transcripts no longer exist in the trunk (data not shown) at 48 hpf, whereas the overall expression pattern in the brain hierarchy has not been altered after 35 hpf (Fig. 3C,D). At 72 hpf, new transcripts of 1G5z appear in the tectum, are more distinct in the eyes, but gradually vanished in the rhombencephalon (Fig. 3E,F). Transverse sections resolved a more detailed expression pattern of 1G5z in the tectum, lateral margin of tegmentum, hypothalamus, proximal region of inner nuclear layer, and some ganglion cells of retina (Fig. 3G–I). In particular, 1G5z appears to be restricted in the amacrine cells of the inner nuclear layer (Fig. 3H). Of interest, the distinct expression pattern of 1G5z in the trigeminal ganglia and eye is reminiscent of the expression profile of Drosophila CaMK-II at the late embryonic stage (Rachidi et al.,1999). In Drosophila, the CaMK-II expression is restricted to the thoracic and cephalic ganglia of the CNS and the eye imaginal disc at the white pupae stage (Rachidi et al.,1999). Drosophila CaMK-II is encoded by a single gene, and its expression is initiated from very early in CNS development (Griffith and Greenspan,1993; Rachidi et al.,1999). On the other hand, 1G5z is not expressed at the onset of primary neurogenesis. These observations suggest that CaMK or 1G5-like genes other than 1G5z could be expressed in the early primary neurons of zebrafish embryos.
Because 1G5z expression is restricted to the specific neuronal cells in the developing embryos, we analyzed its expression pattern in the mindbomb mutant and neurogenin 1 morphant. The mindbomb mutant is characterized by a severe neurogenic phenotype, with overproduction of differentiating neurons that results from deficits in Notch signaling (Jiang et al.,1996; Itoh et al.,2003). The number of 1G5z-expressing cells is remarkably increased in the mindbomb mutant at 24 hpf when compared with wild-type embryos (Fig. 4A,B). Ectopic expression of 1G5z in the neuronal tissues in mindbomb mutant embryos is obvious at the 20-somite stage, when 1G5z transcription begins (data not shown). However, lack of 1G5z expression in the ventral motor neurons in mindbomb further supports the distinct expression pattern of 1G5z in the developing sensory and interneurons of the spinal cord. This result suggests that 1G5z expression is directly or indirectly regulated by mindbomb during the zebrafish embryonic neurogenesis. The correlation between the function of 1G5z with Delta-Notch signaling remains to be investigated further.
Previous reports showed that neurogenin1 (ngn1) depletion caused complete loss of sensory ganglia and trunk sensory neurons without affecting interneurons and motor neurons in zebrafish embryos (Andermann et al.,2002; Cornell and Eisen,2002). To further specify the 1G5z expression domains in developing neuronal tissues, we injected ngn1 morpholino (MO) and then analyzed the 1G5z expression pattern. As shown in Figure 4C,D, ngn1 injection completely abolished 1G5z expression in the trigeminal ganglion. Of interest, the lateral marginal expression of 1G5z in the r2 is also abolished in the ngn1 MO-injected embryos (Fig. 4C,D). However, hindbrain-specific expression of 1G5z was relatively unaffected, except in r2, in the ngn1 morphants (Fig. 4C,D). ngn1 MO injection also completely eliminated 1G5z expression in the caudal spinal cord, demonstrating restricted expression of 1G5z in the posterior sensory neuron at the 24 hpf (Fig. 4E,F). In addition, 1G5z-expressing interneuronal cells from the anterior to medial trunk were not sensitive to ngn1 MO injection, further supporting that 1G5z is exclusively expressed in the interneuron of anterior and coexpressed in both interneuron and sensory neuron of medial trunk (data not shown; Fig. 4F, arrow).
We further confirmed that the 1G5z-expressing cells were postmitotic sensory and interneurons by incubating them with BrdU, which marks S-phase cells at 24 hpf (Fig. 5). As shown in Figure 5B–G, BrdU failed to incorporate into the cells where 1G5z was strongly expressed. So, we verified that 1G5z is exclusively expressed in the postmitotic sensory and interneurons but not in the neural precursors.
In conclusion, a novel zebrafish 1G5z encodes a CaMK-like protein, and its expression pattern is dynamic and specified in the developing CNS throughout embryonic development. This article is the first report of dynamic expression of a CaMK-like gene during the early neuronal differentiation among the vertebrates.
Zebrafish and embryos were maintained essentially as described in the zebrafish book (Westerfield,1995). Developmental stages were classified as hour postfertilization (hpf). Embryos were obtained after natural spawning and cultured at 28.5°C in fishwater containing 0.2 mM (PTU) to prevent pigmentation and 0.001% methylene blue to prevent fungi contamination.
cDNA Cloning and DNA Sequencing
Randomly selected cDNA clones (Clontech, catalog no. QL 40000AB) were analyzed by sequencing and in situ hybridization. Among the tissue specifically expressing cDNA clones, one encoded a novel member of the 1G5 (designated as 1G5z). The full-length 1G5z was subcloned using RT-PCR from 24 hpf embryos and then verified by sequencing analysis.
Total RNA was isolated from the nine different stages of embryos using RNAsol B (Tel-Test. Inc.), and 3 μg of total RNA was used for the RT reaction. The first strand of cDNA was synthesized using MMLV reverse transcriptase (Promega). The full length of 1G5z was amplified with forward (5′-GATGAATTCGCCATGCCATTCGGCTGTTTAAAACCTGG-3′) and reverse (5′-GATCCTCGAGTTAGCCGCTCTGCTCCTCCACTGTGTCC-3′) primers containing EcoRI and XhoI sites using Pfu DNA polymerase. The RT-PCR products were digested with EcoRI and XhoI restriction enzymes and then subcloned into the EcoRI and XhoI sites of pcDNA3 vector. This vector construct was used for further assay. For the analysis of temporal expression patterns of 1G5z, 0.5-kb fragments were obtained by RT-PCR amplification of the first-strand cDNA (predenaturation, 94°C, 3 min; denaturation, 94°C, 30 sec; annealing, 58°C, 30 sec; elongation, 72°C, 30sec; 30 cycles). For the loading control, zebrafish β-actin–specific primers were used for the RT-PCR under the same condition. Two sets of primers were used to amplify 1G5z and β-actin products: 1G5z, 5′-CGTCCACAGGAACCTTAAGCTTGAG-3′ (forward) and 5′-AATTCTTCTCAATTTGAGCGCACAC-3′ (reverse); β-actin, 5′-GAGGAGCACCCCGTCCTGCTCAC-3′ (forward) and 5′-GATGGCTGGAACAGGGCCTCTGG-3′ (reverse). The amplified PCR products were separated on 1.5% agarose gel, and the products were visualized by ultraviolet light after ethidium bromide (0.4 μg/ml) staining.
The full-length 1G5z was PCR amplified then cloned into the multiple cloning sites of pcDNA3 (Invitrogen). Antisense digoxigenin- or fluorescence-labeled riboprobes were generated according to the instructions provided from the labeling kit (Roche). Single or double WISH was performed as previously described (Hauptmann and Gerster,1994), with minor modification. Proteinase K treatment (10 μg/ml) was performed for 3 to 20 min, depending on the stage of the embryos. The hybridized probes were detected using preabsorbed antibodies (Roche) diluted (1/2,000) in blocking solution (PBS, 0.1% Tween 20, 5% sheep serum). After staining, embryos were mounted in a 2:1 mixture of benzylbenzoate:benzylalcohol or 80% glycerol in PBS, then examined under the microscope (MZ16, Leica). Whole-mount stained embryos were subject to cryostat sectioning for detailed analysis following the protocol of Westerfield (1995).
Antisense MO Injection
Antisense MO complementary to the cDNA of neurogenin1 (ngn1) was designed and synthesized by GeneTools (Corvallis, OR). The ngn1 MO sequence is 5′-ACGGATCTCCATTGGATAACCTGG-3′. The MO was injected into the yolk of one- to four-cell stage embryos by using a pressure injection apparatus. Approximately 2 ng of MO per embryo was injected into more than 50 embryos for each experiment.
BrdU Labeling, In Situ RNA Hybridization, and Immunocytochemistry
Manually dechorionated embryos were labeled with BrdU by incubating them for 20 min on ice in a solution of 10 mM BrdU and 15% dimethyl sulfoxide in embryo medium (EM). The embryos were then placed in EM and incubated 20 min at 28.5°C and fixed using 4% paraformaldehyde in PBS. Embryos were processed for in situ RNA hybridization first to detect 1G5z message. The stained embryos were immersed in 2 M HCl for 1 hr, and then processed for anti-BrdU immunocytochemistry. For immunocytochemistry, we used mouse anti-BrdU antibody (G3G4; 1:1,000; Developmental Studies Hybridoma Bank, Iowa City, IA), and for fluorescent detection of antibody labeling, we used Alexa Fluor 568 goat anti-mouse conjugates (1:500; Molecular Probes). Embryos for sectioning were embedded in 1.5% agar/5% sucrose and frozen in 2 methyl-butane chilled by immersion in liquid nitrogen. Sections (10 μm) were obtained by using a cryostat microtome. In situ hybridization and fluorescence images were collected separately and combined using Adobe Photoshop (Adobe Systems, San Jose, CA).
We thank Dr. Feng Liu in Vanderbilt University for critical reading of the manuscript and the Korea Zebrafish Organogenesis Mutant Bank for providing zebrafish lines. We also thank Hye-Won Lee and Yeon-Ju Kim for their technical support of cryostat section.