Differential expression of CaMK-II genes during early zebrafish embryogenesis

Authors


Abstract

CaMK-II is a highly conserved Ca2+/calmodulin-dependent protein kinase expressed throughout the lifespan of all vertebrates. During early development, CaMK-II regulates cell cycle progression and “non-canonical” Wnt-dependent convergent extension. In the zebrafish, Danio rerio, CaMK-II activity rises within 2 hr after fertilization. At the time of somite formation, zygotic expression from six genes (camk2a1, camk2b1, camk2g1, camk2g2, camk2d1, camk2d2) results in a second phase of increased activity. Zebrafish CaMK-II genes are 92–95% identical to their human counterparts in the non-variable regions. During the first three days of development, alternative splicing yields at least 20 splice variants, many of which are unique. Whole-mount in situ hybridization reveals that camk2g1 comprises the majority of maternal expression. All six genes are expressed strongly in ventral regions at the 18-somite stage. Later, camk2a1 is expressed in anterior somites, heart, and then forebrain. Camk2b1 is expressed in somites, mid- and forebrain, gut, retina, and pectoral fins. Camk2g1 appears strongly along the midline and then in brain, gut, and pectoral fins. Camk2g2 is expressed early in the midbrain and trunk and exhibits the earliest retinal expression. Camk2d1 is elevated early at somite boundaries, then epidermal tissue, while camk2d2 is expressed in discrete anterior locations, steadily increasing along either side of the dorsal midline and then throughout the brain, including the retina. These findings reveal a complex pattern of CaMK-II gene expression consistent with pleiotropic roles during development. Developmental Dynamics 236:295–305, 2007. © 2006 Wiley-Liss, Inc.

INTRODUCTION

CaMK-II, also known as the multifunctional Ca2+/calmodulin-dependent protein kinase, is encoded by four genes in mammals (α, β, γ, and δ) and has widespread roles throughout the lifespan of metazoan organisms (Tombes et al., 2003). Among a variety of serine/threonine CaM kinases, CaMK-II is perhaps best known for its extremely high concentration in the adult central nervous system (Hanson and Schulman, 1992) where it enables spatial memory (Silva et al., 1992; Mayford et al., 1996). However, CaMK-II also plays roles during gametogenesis and early development, where its activity promotes meiotic resumption in frog and mouse oocytes (Lorca et al., 1993; Winston and Maro, 1995; Johnson et al., 1998) and cell cycle progression in sea urchin embryos (Baitinger et al., 1990; Tombes and Peppers, 1995). CaMK-II enables embryonic neurite extension, arborization, and dynamics (Zou and Cline, 1996; Johnson et al., 2000; Fink et al., 2003; Wen et al., 2004; Easley et al., 2006). CaMK-II is activated by members of the “non-canonical” Wnt family, such as Wnt5, which lead to widespread cell movements during and after gastrulation (Kühl et al., 2000a,2000b; Sheldahl et al., 2003). Cells activated in this fashion acquire planar polarity, undergo shape changes, and migrate along stereotyped pathways (Tada et al., 2002).

Among Ca2+-dependent targets, CaMK-II is uniquely capable of distinguishing Ca2+ signals by their frequency and amplitude. This is accomplished through oligomerization and autophosphorylation, which allows this enzyme to maintain different activity states even after a Ca2+ stimulus has subsided (Bayer et al., 2002). These common properties coupled with variations in targeting and expression enable CaMK-IIs to distinguish spatially and temporally limited Ca2+ stimuli, which occur during early stages of development (Reinhard et al., 1995; Gilland et al., 1999; Whitaker, 2006).

CaMK-II variability is primarily achieved through alternative exon usage in the central variable domain, yielding over 40 splice variants from the four CaMK-II genes (Hudmon and Schulman, 2002; Tombes et al., 2003). Some of these variants are specifically targeted to the nucleus, the plasma membrane, the actin cytoskeleton, and post-synaptic densities (Srinivasan et al., 1994; Urquidi and Ashcroft, 1995; Heist et al., 1998; Shen and Meyer, 1999; Takeuchi et al., 2000; Caran et al., 2001). Differential expression of CaMK-II genes occurs throughout the adult vertebrate body (Tobimatsu and Fujisawa, 1989). During embryogenesis, γ and δ are the most commonly expressed in the developing central nervous system (Sakagami and Kondo, 1993; Bayer et al., 1999) and in the developing cardiac system (Schworer et al., 1993; Edman and Schulman, 1994; Baltas et al., 1995; Mayer et al., 1995; Ramirez et al., 1997; Singer et al., 1997; Hoch et al., 1998, 2000; Hagemann et al., 1999). β CaMK-II is required for mouse embryonic development (Karls et al., 1992), but α CaMK-II is not expressed until after birth (Brocke et al., 1995).

In this study, we evaluated CaMK-II expression during the first three days of development in the zebrafish, Danio rerio. Three-day-old zebrafish embryos are known to express a δ CaMK-II splice variant (Strausberg et al., 2002), which resembles a cytosolic CaMK-II originally discovered in a human neuroblastoma cell line (Tombes and Krystal, 1997). Other than the “1G5” CaM kinase (Won et al., 2006), no other zebrafish CaM kinases have been characterized. Our study identified 20 splice variants from six CaMK-II genes, which exist in locations consistent with the multifunctional nature of CaMK-II in transducing Ca2+ signals even during the earliest phases of vertebrate development.

RESULTS

CaMK-II Activity Increases in Two Phases During Early Development

CaMK-II activity can be measured in species as diverse as sea urchins (Baitinger et al., 1990; Tombes and Peppers, 1995) and humans (Tombes and Krystal, 1997) using a simple peptide-based assay on cell lysates. CaMK-II activity was undetectable in zebrafish embryos at the 1–2 cell stage (0.5hpf), but began increasing immediately (Fig. 1). At all developmental stages, activity was Ca2+/CaM-dependent and could be inhibited by 20 μM AIP, a peptide that mimics an autoinhibitory domain found in all CaMK-IIs and that has previously been used to inhibit sea urchin CaMK-II (Baitinger et al., 1990; Tombes and Peppers, 1995). The activity increase occurred at the same time that Ca2+ mobilization has been observed (Reinhard et al., 1995; Gilland et al., 1999). The absence of activity in newly fertilized eggs was surprising, since CaMK-II was easily detected in echinoderm eggs (Tombes and Peppers, 1995) using the same assay. After the initial increase, zebrafish specific activity levels remained relatively stable between 4 and 12hpf, but then increased again, rising 100-fold by 3 days of development before leveling off after approximately 6 days. Specific activity of 6-day embryos was comparable to that found in mouse embryonic cells (Tombes et al., 1995; Johnson et al., 2000).

Figure 1.

Developmental kinetics of CaMK-II enzymatic activity. CaMK-II-specific activity was measured in whole embryo lysates using the autocamtide-2 peptide–based assay and plotted during the first 24 hr (left) and the first 10 days (right). Values shown here were completely dependent on Ca2+/CaM and were also sensitive to 20 μM AIP, a CaMK-II specific autoinhibitory peptide.

Identification of Zebrafish CaMK-II Genes

The identification of zebrafish CaMK-II genes was determined using best fit BLAST searches corroborated by rules established in a phylogenetic study of all CaMK-II genes (Tombes et al., 2003). Although there are four known CaMK-II genes in mammals and birds, and a single gene in Drosophila and C. elegans (Tombes et al., 2003), we identified six zebrafish CaMK-II genes. The known genome duplication in teleost fish is likely responsible for the extra genes (Postlethwait et al., 2004). The six genes are shown with their predicted and aligned protein sequences (Fig. 2). Alternative sequences from the central variable region are omitted from this alignment as described previously (Tombes et al., 2003). Based on similarity and gene-specific residues, we conclude that there is one α gene (camk2a1), one β gene (camk2b1), two γ genes (camk2g1 and camk2g2), and two δ genes (camk2d1 and camk2d2). Camk2d2 has been previously reported as camk2d (Thisse and Thisse, 2004). Amongst all six genes, pairwise identity of amino acid sequences ranged from 80–96% and from 90–96% when compared to the corresponding human CaMK-II.

Figure 2.

Zebrafish CaMK-II genes. Six zebrafish CaMK-II genes are shown as their predicted amino acid sequence lacking alternative variable domains. The catalytic domain comprises the first 315 and the oligomerization domain the last 135 amino acids. The variable region extends from amino acids 317 to 345 and corresponds to exons encoding conserved linker domains II and VII (Tombes et al., 2003). Gene-specific residues are underlined. Dots indicate identity and vertical lines indicate conservative replacements among all 6 genes.

Identification of CaMK-II Splice Variants in Zebrafish

The presence of splice variants, if any, from the six zebrafish CaMK-II genes was next evaluated by RT-PCR as previously described (Tombes and Krystal, 1997). Redundant and gene-specific PCR primers, which flank the variable region of each gene, were used to amplify oligo dT-primed cDNA, synthesized from RNA isolated from zebrafish embryos between 1 and 72hpf (Lister et al., 2001). PCR products were cloned and screened by sequence.

All six of the CaMK-II genes expressed at least one mRNA during this time period indicating their transcriptional activity. Twenty different splice variants from these six genes were identified (Fig. 3). According to zebrafish genome conventions and those previously established for this gene family (Hudmon and Schulman, 2002; Tombes et al., 2003), each splice variant is identified by a Greek letter and number referring to the paralog followed by a letter subscript. The predicted amino acid sequence of each of these variants predicts alternative exon utilization consistent with that previously determined for human CaMK-II genes (Tombes et al., 2003). As determined in that study, all splice variants contain sequences encoding conserved linker domains II and VII and alternative domains I, III, IV, V, VI, and VIII/IX.

Figure 3.

CaMK-II splice variant exon utilization. Twenty different CaMK-II variants were identified by sequencing and are shown as their amino acid sequences in the central variable region aligned to demonstrate alternative exon utilization (A). Variants fall into eleven categories of exon utilization (B). Exons are named and boundaries set as previously determined (Tombes et al., 2003). Variants are named with dashes rather than subscripts in this figure for visibility.

Between one and eight variants were identified from each gene. Ten of the twenty variants are novel. Six splice variants contain domain I, which is believed to influence CaM binding (Brocke et al., 1999; Bayer et al., 2002). Five of the variants contain the nuclear targeting domain III, including one maternally expressed variant. The functions of domains IV and V remain unknown while domain VI is believed to regulate nuclear targeting (Takeuchi et al., 2000). Three variants contain domain VIII/IX, which may influence cell surface binding (Urquidi and Ashcroft, 1995). The 20 variants fall into 11 categories of alternative exon utilization (Fig. 3B). For example, category 4 encodes domains II, VI, and VII and was the most common combination with 4 variants represented in this category, including γ1C, which is the major variant identified from 4-cell stage (maternal) mRNA. Other maternal transcripts included α, β1K, γ2L, γ2O, and γ2P variants. The predicted full-length molecular weight of all 20 CaMK-II variants is between 54 and 64 kDa.

Embryonic Localization of CaMK-IIs During Early Development

The timing and location of CaMK-II expression was examined by in situ hybridization for all six genes (Figs. 4–6). Zebrafish embryos were fixed and stained at various times up to 3 days post-fertilization (72hpf). Expression at the 8-cell stage (1hpf) and the 512-cell stage (2.5hpf) reflects maternal mRNA since zygotic expression does not commence until time points represented by the dome stage (4hpf) and the shield stage (6hpf). Somite stage embryos were evaluated at the 3-somite (∼11hpf) and 18-somite stage (∼18hpf). Each gene revealed unique spatial and temporal expression, as summarized in Table 1. Although there are some maternally encoded CaMK-II transcripts, the bulk of CaMK-II expression was zygotic, consistent with the dramatic increase in activity starting at 10hpf (Fig. 1).

Figure 4.

In situ localization of α and β CaMK-II mRNAs. CaMK-II expression was assessed by in situ hybridization with a probe for camk2a1 and camk2b1 at the indicated stages. Stages analyzed in Figures 4–6 include the 8-cell stage (8; 1hpf), the 512-cell stage (512; 2.5hpf), the dome stage (Do; 4hpf), the shield stage (Sh; 6hpf), the 3-somite stage (3s; ∼11hpf), and the 18-somite stage (18s; ∼18hpf). The 8-cell stage is an animal poll view; others are lateral except for dorsal views at 24, 48, and 72hpf. Arrow in camk2a1 indicates heart. Arrows in camk2b1 at 24hpf indicate somites and at 72hpf gut. Letters locate cross-sections for Figure 7. Scale bar = 1 mm.

Figure 5.

In situ localization of γ CaMK-II mRNAs. CaMK-II expression was assessed by in situ hybridization with a probe for camk2g1 and camk2g2 at the indicated stages as in Figure 4. Arrows in camk2g1 at 24hpf indicate dorsal cell bodies and at 72hpf gut. Letters locate cross-sections for Figure 7. Scale bar = 1 mm.

Figure 6.

In situ localization of δ CaMK-II mRNAs. CaMK-II expression was assessed by in situ hybridization with a probe for camk2d1 and camk2d2 at the indicated stages as in Figure 4. The 8-cell stage is an animal poll view; others are lateral except for dorsal views at 24, 48, and 72hpf. Arrows in camk2d1 at 24hpf indicates hatching gland. Letters locate cross-sections for Figure 7. Scale bar = 1 mm.

Table 1. Relative CaMK-II Expression Levelsa
 camk2a1camk2b1camk2g1camk2g2camk2d1camk2d2
  • a

    Relative levels of expression at different times and places were assessed and are summarized for all 6 genes during the first 3 days of zebrafish development.

8-cell++++++
512-cell++++
Dome+++
Shield+++
3-somites++++
18-somites+++++++
24hpf trunk++++++++
24hpf Epidermis+
24hpf Brain+++++
24hpf Hatching Gland++
48hpf Heart++
48hpf Forebrain++++++
48hpf Midbrain+++++++
48hpf Retinal epithelia++
72hpf Retinal epithelia++++++
72hpf Forebrain+++++++++++
72hpf Midbrain+++++++++++
72hpf Hindbrain+++++++++
72hpf Pectoral fin++
72hpf Gut+++
72hpf Trunk++

α1 (camk2a1) transcripts appear diffusely in the 8-cell embryo, but are undetectable at the 512-cell, the dome, and the shield stages (Fig. 4). Expression then increases throughout the 3-somite embryo, along the ventral edge of 18-somite embryos and on either side of the midline in anterior somites at 24hpf. Expression is strong in the heart at 48hpf (arrow). At 72hpf, α1 is expressed at low levels in the forebrain and persists in heart tissue (Fig. 4). Only one α1 CaMK-II catalytically active splice variant was cloned by RT-PCR throughout this time.

β1 (camk2b1) mRNA was detected at low levels at all time points from the 8-cell stage through gastrulation and into the somite stage (Fig. 4). Expression localized more ventrally at the 18-somite stage. At 24hpf, expression was in somites (arrows) along the entire trunk and in the hatching gland. Expression diminished along the trunk at 48hpf, but appeared in the forebrain in discrete anterior structures. By 72hpf, expression was strong in mid- and forebrain and lighter in gut (arrow), retina, trunk, and pectoral fins (Fig. 4). Two of the three β1 splice variants identified have nuclear targeting domains.

The strongest mRNA signal detected during early cleavages and through gastrulation was encoded by the γ1 (camk2g1) gene (Fig. 5). As determined above, this most likely represents the γ1C variant, which was the only γ1 variant identified in mRNA obtained from 1hpf embryos. mRNA staining was high throughout all blastomeres from the 8-cell through the 512-cell stage, and then diminished, but did not disappear at the dome, shield, and 3-somite stages. Expression re-appeared throughout the entire embryo, not just ventrally, at the 18-somite stage. γ1 expression was strong in the brain and in cell bodies (arrows) on either side of the dorsal midline at 24hpf, but then, like β1, diminished at 48hpf, reappearing strongly in the mid- and hindbrain, gut (arrow) and pectoral fins at 72hpf. One of the three γ1 variants detected during development contains domain I; none are nuclear.

Weak, but significant, levels of γ2 (camk2g2) transcripts were detected from the 8-cell to the 3-somite stage (Fig. 5). γ2 expression increased ventrally at the 18-somite stage. At 24hpf, γ2 expression remained diffusely ventral in the developing brain and along the trunk. At 48hpf, expression increased in the midbrain and along the trunk and was the first CaMK-II gene to be expressed in the retina. The overall expression of γ2 diminished slightly at 72hpf but remained significant in the mid- and hindbrain and retina. This gene encoded 8 diverse variants during this time period; its diffuse localization may be a reflection of its complex expression.

Expression from the δ1 (camk2d1) and δ2 (camk2d2) genes was undetectable until the 18-somite stage (Fig. 6), which was consistent with the lack of any δ transcripts identified from cleavage-stage mRNA. δ1 appeared restricted to somite boundaries at the 18-somite stage and at 24hpf is elevated in epidermal tissue and in the hatching gland (arrow). After 24hpf, δ1 expression diminished, but persisted at low levels in discrete locations along the dorsal trunk. At 48hpf, expression was almost undetectable, except for light forebrain staining, while at 72hpf, expression re-appeared very weakly along the entire dorsal trunk in discrete cell bodies. Like the one variant expressed from the α1 gene, the two δ1 splice variants were simple, utilizing a minimal number of alternative exons.

δ2 (camk2d2) transcripts are absent until the 18-somite stage. At 24hpf, expression is observed in discrete anterior locations and along either side of the midline. At 48hpf, expression is primarily in the forebrain and then accumulates in the forebrain, hindbrain, and retinal epithelium at 72hpf. Three non-nuclear splice variants were identified from this gene.

Frozen thin sectioning was used to further locate CaMK-IIs in the central nervous system at 72hpf (Fig. 7). Four of the six CaMK-II genes, including camk2b1 (β1), camk2g1 (γ1), camk2g2 (γ2) and camk2d2 (δ2) exhibited strong staining in the central nervous system at this time point and were comparatively imaged in cross-section. Cross-sectional locations are indicated (A–F) in Figures 4–6. Staining corroborated whole mount images and provided additional insight into the gene-specific expression patterns (note arrows). For example, all four genes were expressed in retinal epithelia (B); β1, γ1, and γ2 were preferentially found in an inner cell layer closest to the lens, which most likely represents the ganglion cell layer. In contrast, δ2 preferentially labeled an outer retinal cell layer (arrow). This cross-sectional analysis revealed that β1 and γ1 were intensely co-expressed in a cortical rim in the midbrain (C, D: arrows), whereas δ2 was expressed in a thin intense layer in the hindbrain (E: arrow). At the base of the anterior trunk (F), γ1 was expressed in discrete ventral spots while δ2 was expressed in more dorsal foci (F: arrows).

Figure 7.

Cross-sectional localization of CaMK-II mRNAs. CaMK-II expression was localized at 72hpf by 30-μm frozen thin sections of pre-stained embryos. Sections shown were acquired at the locations (A–F) indicated in Figures 4–6 along the anterior-posterior axis. Scale bar = 0.1 mm.

DISCUSSION

In this study, CaMK-II expression was comprehensively monitored from fertilization through early development in the zebrafish. Our findings are consistent with other studies, which have implicated CaMK-II during events around the time of gastrulation and then in the developing circulatory and central nervous systems. This study also raises possibilities for the involvement of specific CaMK-IIs in the formation of the anteroposterior axis, somites, retinal epithelia, gut, and other tissues. The numerous CaMK-II variants encoded during this time period in a spatially and temporally regulated manner supports complex transcriptional and post-transcriptional gene regulation and implies multiple roles for members of this protein kinase family.

Like other genes in teleost fish, CaMK-II genes have been duplicated. At least half of the duplicated CaMK-II genes have been retained and are transcriptionally active. In contrast, only 20–30% of all duplicated zebrafish genes are retained (Postlethwait et al., 2004). The six CaMK-II genes encode between one (α1) and eight (γ2) splice variants. By both RT-PCR and in situ hybridization, several of these genes are expressed maternally, including mRNAs from the α1, β1, γ1, and γ2 genes. Maternal CaMK-IIs are not just the simple variants, but utilize alternative domains implicated in CaM binding or in targeting to the nucleus or membrane.

Surprisingly, no CaMK-II enzymatic activity was detected at fertilization, in contrast to the significant activity found in other species (Baitinger et al., 1990; Tombes and Peppers, 1995; Winston and Maro, 1995; Johnson et al., 1998; Stevens et al., 1999). This is surprising since zebrafish eggs are fertilized at metaphase II (Selman et al., 1993), where CaMK-II has been implicated in enabling the release from meiosis II arrest (Lorca et al., 1993; Winston and Maro, 1995; Johnson et al., 1998). Within the first two hr of development, however, active CaMK-II is detected at the time when developmentally important Ca2+ transients are known to occur (Gilland et al., 1999; Creton, 2004). Activity assays and in situ localization support complex CaMK-II expression patterns as a result of maternal mRNA and zygotic gene expression.

At least five of the six zebrafish CaMK-II genes are expressed in the developing nervous system, supporting the importance of this gene family in central nervous system function (Hudmon and Schulman, 2002). In the mouse, the α CaMK-II gene is expressed exclusively post-natally in the hippocampus and frontal cortex (Burgin et al., 1990; Bayer et al., 1999). Only one α CaMK-II gene has been uncovered so far in the zebrafish, so it is possible that another α CaMK-II gene may be expressed in the adult zebrafish brain. For this reason, we have numbered each zebrafish CaMK-II gene with a numerical postscript even though, in the case of α and β, only one paralog has so far been identified from either the searches of zebrafish genome databases or from splice variant sequences.

Cardiac expression was restricted to the α1 CaMK-II gene. Only one catalytic variant was cloned from this gene. A non-catalytic splice variant of the α gene, αKAP, has been found in mammalian cardiac muscle (Sugai et al., 1996; Singh et al., 2005), where it oligomerizes with and targets active CaMK-II to membranes, such as the sarcoplasmic reticulum (Bayer et al., 1996, 1998; Sugai et al., 1996; Takeuchi and Fujisawa, 1997). In situ α locations could reflect this non-catalytic variant; however, it would have to hetero-oligomerize with the catalytic α CaMK-II variant cloned here since no other CaMK-II gene products were detected in the developing heart. These findings are consistent with the expression of catalytically active α CaMK-II in the heart.

Biphasic expression was most apparent with β1 and γ1 CaMK-II, as described above and summarized in Table I. β1 was expressed within the first day in all somites, was lost along the trunk at 48hpf, but re-appeared strongly in the brain, gut, retina, and pectoral fins at 72hpf. γ1 mRNA was prevalent in cleavage-stage embryos and then along either side of the midline at 24hpf, was significantly decreased at 48hpf, and then reappeared in brain, gut, and pectoral fins. These patterns of expression suggest that these two CaMK-II genes are controlled by one set of transcriptional regulatory influences at 24hpf and another set at 72hpf. The additional influences of mechanisms that control alternative splicing during development is not known for CaMK-II genes, but could be initially assessed by quantitative analysis of the relative levels of each splice variant through development.

Retinal expression of CaMK-II has previously been reported in the synaptic and pigment epithelial layers (Bronstein et al., 1988), but these and other studies have primarily implicated α CaMK-II (Liu and Cooper, 1996; Laabich et al., 2000). Our studies indicate that four CaMK-II genes are expressed in the retina; the only CaMK-IIs not expressed in the developing retina are encoded from the α1 and the δ1 genes. Among the retinal layers, our results indicate that each CaMK-II gene encodes its own unique pattern of expression.

Zebrafish CaMK-II localization is consistent with its involvement in “non-canonical” Wnt pathways (Kühl et al., 2000a,2000b; Sheldahl et al., 2003). Non-canonical Wnts induce convergent extension movements during and after gastrulation (Tada et al., 2002). In zebrafish, defects in convergent extension movements are observed with the Wnt5 pipetail mutant (ppt) (Kilian et al., 2003) and its receptor, frizzled 2 (Sumanas et al., 2001). Pharmacological disruption of Ca2+ dynamics during the discrete developmental window (6–8hpf) associated with gastrulation also leads to defects in axis formation (Creton, 2004). Wnt5 has been implicated in the specification of myoblasts (Anakwe et al., 2003) and retinal cells (Yu et al., 2004) and activates CaMK-II (Kühl et al., 2000a). Wnts 4, 5a, and 11 also influence the midline assembly of organ precursors, including liver, gut, pancreas, and heart, all of which require endoderm migration (Pandur et al., 2002; Kim et al., 2005; Matsui et al., 2005). The reported elevation of CaMK-II activity on the ventral side of the embryo (Kühl et al., 2000a) is consistent with our observation that four of the six zebrafish genes express ventrally at early stages of development. In general, the co-incidence of CaMK-II expression with members of the Wnt family supports the possibility that CaMK-II diversity reflects variations in its responsiveness to members of the Wnt family.

Convergent extension is typically accomplished through either directed migration, cell shape changes, or cell rearrangements (Wallingford et al., 2002). These post-transcriptional mechanisms are likely the result of alterations in the cytoskeleton through discrete targets. Transient Ca2+ elevations have been proposed to promote focal complex disassembly and detachment from the extracellular matrix at the periphery of motile cells (Marks and Maxfield, 1990; Conklin et al., 2005). CaMK-II has been implicated in cell motility in mammalian cells (Lundberg et al., 1998; Pfleiderer et al., 2004) and δ CaMK-II has been detected in extracts of isolated pseudopods by mass spectrometry (Lin et al., 2004). A role for CaMK-II in convergent extension is, therefore, likely through its direct action on the cytoskeleton. In support of this, of the five maternal CaMK-II transcripts detected during early development, four are extranuclear.

Why are there so many CaMK-II splice variants expressed during embryogenesis? One fourth of the zebrafish embryo CaMK-IIs identified in this study have nuclear targeting sequences. Nuclear targeted CaMK-IIs are known to act directly on transcription factors, such as CREB (Matthews et al., 1994; Shimomura et al., 1996; Sun et al., 1996; Ramirez et al., 1997). Other CaMK-IIs that we have identified have variable spacers, which may enable them to respond differently to exposure to Ca2+ and CaM. Others have newly discovered domains, which may enable them to interact with binding partners, substrates, or organelles in a manner that supports their specific functions. This study has laid the groundwork for assessing the role of specific CaMK-IIs in discrete developmental functions.

EXPERIMENTAL PROCEDURES

Zebrafish Care and General Reagents

Wild type fish embryos were obtained through natural crosses (AB Strain), raised at 28°C, and staged as described (Kimmel et al., 1995).

Whole Cell Lysate Preparation

Embryos were dechorionated and then lysed in 30 mM Hepes, pH 7.4, 20 mM MgCl2, 80 mM β-glycerol phosphate, 5 mM EGTA, 0.1 μM okadaic acid (Life Technologies Invitrogen, Carlsbad, CA), 0.01 mg/ml each chymostatin, leupeptin, aprotinin, pepstatin, and soybean trypsin inhibitor (Sigma Chemical Co., St. Louis, MO), sonicated for 5 sec on ice, and centrifuged at 12,000g for 15 min at 4°C. This buffer has been optimized for maximal CaMK-II recovery into this supernatant (Tombes et al., 1995, 1999). The addition of the detergent NP-40 did not increase the level of activity recovery. Protein concentrations were determined using the BCA assay (Pierce Chemicals, Rockford, IL).

CaMK-II Activity Assay

Total CaMK-II activity was assessed by measuring phosphate incorporation into autocamtide-2, a peptide modeled after the autophosphorylation site of CaMK-II (KKALRRQETVDAL). Reactions were carried out on 1–2 μg protein from cell lysates in a total volume of 25 μl containing final concentrations of 20 mM HEPES (pH 7.4), 0.1 mM dithiothreitol, 15 mM magnesium acetate, 20 mM β-glycerophosphate, 10 mM NaF, 0.5 μM PKA inhibitor peptide, 0.1 μM okadaic acid, 10 μM [γ-32P]-ATP (0.5 mCi per assay), 35 μM autocamtide-2, and either 1 mM EGTA (-Ca2+) or 1 μM bovine calmodulin plus 2.0 mM CaCl2 (+Ca2+). After 10 min at 32°C, 20 μl was pipetted onto P81 phosphocellulose paper squares (Whatman, Florham Park, NJ) that were air dried for 1 min and washed five times in 500 ml 1% phosphoric acid. Dried paper squares were quantitated by Cerenkhov counting. These assay conditions were optimized for compatibility with the buffer in which cell lysates were prepared (Tombes et al., 1995, 1999). Activity detected in this assay could be inhibited by AIP, the autoinhibitory peptide (Biomol, Plymouth Meeting, PA) whose sequence is KKALRRQEAVDAL. Using partially purified CaMK-II from 6-day zebrafish embryos, we determined that this peptide could half-maximally inhibit at 10 μM and was 90% effective at 25 μM, which are values very similar to that seen with partially purified mouse CaMK-II.

Genomic Identification

Sequences were obtained by searching GenBank and Sanger Institute genomic databases using a BLAST (Basic Local Alignment Search Tool) search for zebrafish CaMK-II homologs. Six CaMK-II genes were identified and are listed with accession number and GenInfo sequence identification numbers (gi). They included camk2a1 (chromosome 21, NM_001017741, gi: 82524370); camk2b1 (chromosome 5, XM_685461, gi: 68360079); camk2g1 (chromosome 12, BC096785, gi: 66911241); camk2g2 (chromosome 13, XM_686010, gi: 68437880); camk2d1 (chromosome 10, BC077143, gi: 50417146) and camk2d2 (chromosome 1, NM_001002542, gi: 50540149).

RT-PCR and Sequencing

Total RNA was prepared from dechorionated embryos at 1hpf, 24hpf, and 72hpf and cDNA was prepared as described (Lister et al., 2001). PCR primers which bracketed the variable region included two sense primers (TGGATCTGCCAACGCTCCACTGT- GGC, TGGATATCACATCGCTCCACCGTCGC), which encode WICQRSTVA and WISHRSTV, respectively, and three antisense primers (CCTCATGTGCACCTGATTGGAGA, CACAGTTTGTGGATGGCCAGGGCC, GTAACCGAGCAGCTGATTGAAGCC), which encode PHVHLIGE, QFVDGQG, and VTEQLIEA. These primers were used in various combinations in order to achieve the greatest coverage of potential CaMK-II sequences. CaMK-II PCR products were cloned into the TOPO/TA vector (Invitrogen) and clones were screened by PCR, purified, and then sequenced. Over 100 clones were screened in this fashion.

Whole Mount In Situ Hybridization

Digoxigenin-labeled anti-sense riboprobes (0.5–1.5 kb) were synthesized using T3 or T7 RNA polymerase (depending on the orientation) from TOPO/TA vector-based cDNAs and then incubated with fixed, staged embryos as described (Dutton et al., 2001). Individual probes derived from cDNAs encoding the splice variants αKAP, β1e, γ1G, γ2L, δ1G, and δ2E were used to localize all transcripts from each gene. The αKAP and the δ2E probes were synthesized from purchased clones (Open Biosystems, Huntsville, AL) whereas other probes were synthesized from TOPO/TA clones. Anti-digoxigenin and alkaline phosphatase conjugated secondary antibodies were used and developed with NBT/BCIP as substrate. Sense probes showed no signal. In some cases, stained embryos were embedded in NEG-50 frozen sectioning medium (Richard-Allan Scientific, Kalamazoo, MI), frozen to −20°C and 30-μm sections obtained using an HM-550 cryostat (Richard-Allan Sci). Sections were re-hydrated in 50% glycerol and photographed with a Nikon Cool-Pix 990 color camera on an Olympus IX-70 bright-field microscope. Whole-mount specimens were photographed with an Olympus DP70 digital camera on an Olympus SZX12 stereo microscope.

Acknowledgements

The authors are grateful to Cassandra Barrett, Sabrina Cline, Charles A. Easley, Ludmilla Francescatto, and AnhThu Nguyen for assistance with this study. Some of the fish used in this study were obtained from the Zebrafish International Research Center, supported by NIH-NCRR grant # RR12546. This work was supported by a Massey Cancer Center Pilot Project, National Science Foundation grant IBN-0238821 (R.M.T.), and American Cancer Society Institutional Research Grant IRG-99-225-0 (J.A.L.).

Ancillary