Specific CaMKIIs mediate convergent extension cell movements in early zebrafish development

Noncanonical Wnts are morphogens that can elevate intracellular Ca2+, activate the Ca2+/calmodulin‐dependent protein kinase, CaMKII, and promote cell movements during vertebrate gastrulation.


| INTRODUCTION
Coordinated inductive and morphogenetic processes establish the vertebrate body plan during gastrulation.Three modes of cell migration enable these rearrangements and include epiboly, internalization of presumptive mesendoderm and convergent extension (CE). 1 CE movements narrow the germ layers mediolaterally (convergence) and elongate them anteroposteriorly (extension) to establish the embryonic axis.
2][3][4] During zebrafish development, suppression of the ncWnts, Wnt5, and Wnt11, 2,5-9 leads to a shortened anterior-posterior body axis, wider dorsal structures and defects in segmentation, which are stereotypical of CE defects seen in all vertebrates. 1,3Some investigators view Wnt5 and Wnt11 as having overlapping or even redundant functions, 10 but others observe distinct non-redundant functions of these morphogens during gastrulation. 7ntracellular Ca 2+ levels are elevated during gastrulation [11][12][13] and influence multiple developmental events. 14In particular, cells comprising the dorsal side of the blastoderm show elevated Ca 2+ which then propagates toward the embryonic anterior as a wave. 11,12During zebrafish gastrulation, activation of ncWnt (Wnt5 and Wnt11) signaling also results in an elevation of Ca 2+ , 10,15 whereas ncWnt mutants have altered or reduced Ca 2+ release. 10atural oscillations of Ca 2+ have also been reported during gastrulation. 16One potential mediator of Ca 2+ signals during gastrulation is CaMKII, a Ca 2+ /calmodulindependent serine/threonine protein kinase. 17,18n zebrafish, CaMKII is encoded by seven genes that give rise to at least two dozen splice variants in early embryos. 19,20CaMKII is evolutionarily conserved and is expressed throughout the life span and tissues of all animals. 21,22CaMKII is exquisitely sensitive to the amplitude and frequency of Ca 2+ elevations and can recall its activation even after Ca 2+ has subsided. 235][26][27] In mouse fibroblasts, both inhibition and hyperactivation of CaMKII block cell migration and focal adhesion turnover leading to the conclusion that cycles of CaMKII activation are necessary for cell migration. 279][30] This study identifies camk2b1 and camk2g1 as two CaMKII genes necessary for CE movements.Our findings not only link specific CaMKII gene products to CE cell movements but provide further evidence that CaMKII must be cyclically activated to enable cell migration during early development.

| Suppression of camk2b1 and camk2g1 expression cause CE defects
Seven transcriptionally active genes encoding full-length CaMKII have been identified and described in early zebrafish embryos. 19,20Gene-specific, translation-blocking, anti-sense MOs have been designed against all eight CaM-KII open reading frames 30 and used to demonstrate the importance of specific CaMKII genes in organ asymmetry, cardiac, ear, and kidney development. 19,28,30These MOs have all been validated by measuring their impact on CaMKII protein levels and by their unique effects on development.Further validation comes from the ability of CRISPR-based camk2g1 G0 mutants (crispants) to phenocopy morphants and to specifically repress expression. 22amk2b1 crispants have also been generated and phenocopy camk2b1 morphants (Figures S1 and S2).MO-based approaches remain valuable since stable CaMKII mutants can be compensated for by CaMKII paralogs. 22,31Os also remain advantageous for early developmental studies as they can target both maternal and zygotic transcripts.Only two of the eight CaMKII ORFs, camk2b1 and camk2g1, showed morphant phenotypes consistent with roles in CE.Interestingly, these two genes were the most conspicuously expressed prior to and during gastrulation and somitogenesis. 20The ultimate morphant phenotypes at 72hpf observed with these two genes has been noted in previous studies 19,[28][29][30] and CaM-KII activity levels decreased in a MO dose-dependent fashion. 19,22,30ocusing on the first 24hpf, both camk2b1 and camk2g1 morphants exhibited a truncated yolk extension (asterisk) with deformations of the otherwise clearly delineated somites (arrow) (Figure 1A-C).Camk2b1 morphants also exhibited compressed somites and curvature of the tail (Figure 1B).Camk2g1 morphants displayed severe somite and tail compression (Figure 1C).Camk2b1 G0 crispants also exhibit somite compression and notochord undulation (Figure S1).These phenotypes were absent in morphants of the corresponding paralogs (camk2b2 and camk2g2), whose amino acid sequences are 96%-98% identical. 19,20he early developmental phenotypes observed with these CaMKII morphants mirror those of genes associated with the CE cell movements associated with gastrulation. 32In particular, camk2b1 morphants showed similarities to Wnt5/ppt mutants 32 or morphants 33 whose phenotypes include a shortened axis, condensed somites and undulated notochord.Camk2g1 morphants resembled Wnt11/slb mutants/morphants whose phenotypes include a shortened and broadened axis or notochord. 9,33inking specific CaMKII genes to specific ncWnts would help resolve whether Wnts have distinct 7,32 or redundant 3,10 roles.
Body axis extension defects can be quantified by measuring gap angles 34,35 at the 10 somite stage (ss) (Figure 1D-F).The gap angle is defined as the angle formed between the head, mid-yolk, and tail of embryos (denoted by dashed lines in Figure 1D).Whereas control embryos had an average gap angle of 47 , both camk2b1 and camk2g1 morphants exhibited statistically significant dose-dependent increases in the average gap angle to >79 (Figure 1G).
Convergence and extension during zebrafish gastrulation primarily occurs in the anterior and posterior mesendoderm and neuroectoderm cell layers. 1 Mesendoderm tissue is defined as the hypoblast during zebrafish gastrulation and is composed of axial and paraxial cells. 36To determine if CaMKII suppression altered movement of the cells of the hypoblast, simultaneous in situ localization of the axial and paraxial transcripts tbxta and myoD was conducted on morphant embryos at 90% epiboly and early somitogenesis (6-7ss).While expression levels of both myoD and tbxta remained high, their patterns changed.At 90% epiboly, the width of axial tissue (denoted by arrowheads in Figure 2A) was greater in both camk2b1 and camk2g1 morphants compared to controls (Figure 2A-C).In these same embryos, myoD expressing cells remained dispersed at the base of the elongating notochord in both camk2b1 morphants (Figure 2B) and camk2g1 morphants (Figure 2C).These results indicate that camk2b1 and camk2g1 transcripts influence mesendoderm migration at the end of gastrulation.
At the 6-7ss, the normal shape of somites (Figure 2D) is altered in camk2g1 morphants, with somites extending less posteriorly and more horizontally with a decrease in their spacing (Figure 2F).Camk2b1 morphants revealed an intermediate somite phenotype (Figure 2E).We quantitated this phenomenon by determining the ratio of the length between the first and sixth somites and the width at the sixth somite (denoted by solid white lines, Figure 2D).We took measurements at the 6-7ss since we could be certain of comparing embryos at the same developmental stage.Camk2b1 and camk2g1 morphants both yielded statistically significant decreases in length:width ratios from 1.37 ± 0.15 (control) to 1.17 ± 0.08 for camk2b1 morphants and 0.90 ± 0.17 for camkg1 morphants (Figure 2G).
CE movements of neuroectodermal cells can also be revealed using simultaneous in situ hybridization with probes for the neural plate (dlx3), hatching gland/ prechordal plate (ctslb), and notochord (tbxta) of 1-3 somite embryos. 7An anterior view reveals the neural plate boundary (dlx3) surrounding the developing zebrafish embryo with a group of cells comprising the hatching gland/prechordal plate (ctslb) positioned directly anterior to the tip of the converged neural plate boundary (Figure 3A).From a posterior perspective, the neural boundaries (arrows) run parallel to the notochord (Figure 3D).Both camk2b1 and cam2g1 morphants contain neural plate boundary cells (dlx3) but these do not converge and remain out of view from either anterior (Figure 3B,C) or posterior (Figure 3E,F) perspectives.
In addition to the widened neural plate boundaries, CaMKII morphants show defective anterior migration of the prechordal plate (ctslb) relative to the neural plate boundary (dlx3) (Figure 3B,C).Camk2b1 and camk2g1 morphants also show that notochord cells expressing tbxta (white bar, Figure 3D) remain dispersed (Figure 3E,F).Taken together, these results demonstrate that a knockdown of specific CaMKIIs results in defective CE.

| Cell fate specification is unaltered in CaMKII morphants
The reduced convergence and extension of axial, paraxial and neuroectoderm tissue observed in CaMKII morphants suggests impairment of cell movements during gastrulation.However, these results could also be explained by alterations in cell fate specification which occur at the same time during development.To test whether cell fate specification was altered, whole-mount in situ hybridization was performed using the dorsal marker, goosecoid ( gsc) and the ventral marker bone morphogenetic protein 4 (bmp4) at the onset of gastrulation (shield stage) and during mid-gastrulation (75% epiboly).The expression levels of bmp4 and gsc remained unchanged in both camk2b1 and camk2g1 morphants at both stages when compared to control embryos (Figure 4).As shown above (Figures 2 and 3), the levels of other developmental markers (dlx3, ctslb, myoD, and tbxta) are also unchanged.These results indicate that defective CE is not the result of altered cell fate.

| Cell proliferation and survival are unaltered in CaMKII morphants
Morphological changes could also be explained by alterations in cell proliferation and cell survival/death.Therefore, proliferation was assessed by immunostaining mitotic nuclei (phosphorylated histone H3) and total nuclei (propidium iodide) at the 75% epiboly and shield stage (Figure 5).The average mitotic index in control embryos ($7%) was consistent with previous findings 37 and revealed no quantitative differences with morphant embryos.Apoptosis is indicated by cleaved caspase-3 immunostaining and occurs in only a handful of cells (<5/embryo) at the shield and 75% epiboly stages in wild-type embryos.This level was also unchanged in both camk2b1 and camk2g1 morphants (Figure 5).CE defects are therefore not the result of changes in cell proliferation or apoptosis.

| Dominant-negative CaMKII induces CE defects including midline bifurcation
The full-length CaMKII point mutant K 43 A is a kinaseinactive mutant that has previously been shown to act in a dominant-negative fashion when targeted to dorsal forerunner cells with the sox17 promoter 30 or to pronephric cells using the kidney specific Na + /K + -ATPase promoter. 28The use of full-length CaMKII enables it to be incorporated into endogenous CaMKII and act in a dominant-negative fashion.In fact, the kidney studies suggested a role for CaMKII in directed pronephric cell migration. 28For the present study, the expression of full-length green fluorescent protein (GFP)-tagged K 43 A CaMKII cDNA was driven by the carp β-actin promoter of the pFRM2.1 vector which yields an early and widespread, but mosaic expression pattern throughout the embryo. 19Three concentrations of GFP K 43 A CaMKII were injected and embryonic phenotypes were assessed at 24hpf for all injected embryos.35%-60% of all GFP K 43 A CaMKII expressing embryos showed relatively normal development (Figure 6A).The compressed phenotype, previously noted as a CE defect and observed in camk2b1 and camk2g1 morphants, was seen in 25%-35% of K 43 A expressing embryos (Figure 6B).Midline bifurcation (MB), never seen in CaMKII morphants, was observed in approximately one-fourth of embryos injected with the K 43 A CaMKII construct (Figure 6C).MB is never accompanied by duplicated heads or tails, is restricted to the trunk and is referred to as "split."In fact, clusters of cells expressing GFP K 43 A CaMKII (in the cytosol) were typically seen in the region between the two trunks (Figure 6C,F, asterisk).As development progressed, the duplicated trunks of some of the MB embryos could twitch and contort independently like the single trunk of a wild-type embryo, indicating duplicated, but functional tissue.The MB phenotype was dose dependent as it was seen in 16% of embryos injected with 80 ng, 24% of 100 ng, and 29% of 120 ng (Figure 6J).
In situ hybridization using the axial and paraxial markers tbxta and myoD, respectively, was used to further assess the MB phenotype of GFP K 43 A CaMKII embryos (Figure 6G-I).Representative images show that in each of these MB embryos, myoD expression in somites occurs in each half trunk, but then rejoins distally (Figure 6).MB is rare in other zebrafish mutants, but has been observed in embryos that have lost the expression of the nodal-related protein Squint. 38Like CaMKII, Squint disruption leads to a low (<20%) MB frequency 38 and also shows that tissue on both sides of the MB retains its identity.This finding is consistent with the morphant results that CaMKII suppression interferes with cell migration necessary for convergence but does not disrupt tissue specification.

| Hyperelevation of intracellular Ca 2+ increases CaMKII autonomy and causes CE defects
Overexpression of ncWnts during vertebrate development has been shown to cause CE defects that phenocopy those caused by ncWnt suppression. 39,40Similarly, hyperactivation of CaMKII causes CE defects 17,41 that resemble those caused by CaMKII suppression or inhibition (Figures 1-6).Therefore, we postulated that excessive elevation of Ca 2+ and the subsequent activation of CaMKII could cause a similar phenotype to that observed in CaM-KII morphants.In fact, the elevation of Ca 2+ caused by inhibition of the endoplasmic reticulum (ER) Ca 2+ pump, via thapsigargin (TG) prior to and during zebrafish gastrulation, has previously been shown to induce CE phenotypes. 42We therefore transiently exposed zebrafish embryos to increasing concentrations of TG just prior to (0-5hpf) and during gastrulation (5-7hpf) and not only evaluated CE, but measured CaMKII activation.The results indicate that TG-treated embryos exhibit dose-dependent axis compression defects at 24hpf (Figure 7E-H), much like those seen in CaMKII morphants and as previously reported. 42In addition, embryos at the 10ss showed a dose-dependent increase in gap angle from 34 in control embryos to as high as 150 in embryos treated with 0.75 μM TG (Figure 7A-D,I).This transient treatment with TG from 5 to 7hpf also caused a dose-dependent increase in whole embryo CaMKII activation (autonomy), when measured at the 12-16ss, but not at 24hpf (Figure 7J), when Ca 2+ levels in TG-treated embryos have returned to baseline. 42,43Embryos treated with TG prior to the 5-7hpf window did not exhibit these defects (data not shown).TG embryos probed for myoD and tbxta expression, as in Figure 2, show that expression of these markers is not blocked, but that somites do not converge and extend normally after transient TG treatment (Figure 7K-N).We can therefore verify that hyperactivation of CaMKII interferes with CE, but not specification.

| Partially activated CaMKII rescues CE defects of Wnt11 morphants
Since CaMKII is known to be activated in response to ncWnts 17 and Wnt11 may play a greater role in CE than Wnt5 in the anterior region, 7 we sought to rescue Wnt11 morphants with CaMKII.Constitutively active but not wild-type CaMKII has been shown to rescue the ppt/wnt5 truncated tail phenotype 10 and so we sought to test whether Wnt11 morphants could also be rescued.Fulllength CaMKII was used for these studies, as it was for previous rescues, 19,28,30 since it can hetero-oligomerize with endogenous CaMKII and influence the entire oligomeric complex.The Wnt 11 morphant phenotype observed in this study was similar to that previously described 9,33 and phenocopies the camk2b1 and camk2g1 morphants (Figure 8).When wnt11 MOs were co-injected with wildtype CaMKII mRNA alone or constitutively active CaMKII (T 287 D) mRNA alone, CE defects were not reversed (data not shown).However, by mixing mRNAs encoding wild-type and constitutively active CaMKII at a 3:2 ratio, the CE defects of wnt11 morphants could be partially rescued (Figure 8).Embryos at the 1-3ss were assessed by in situ hybridization using the dlx3/ctslb/tbxta probe combination, as in Figure 3. Fully defective CE embryos were defined as those in which expression of ctslb is posterior to dlx3, partially defective CE are those in which ctslb is on or within dlx3 expression, and normal embryos expressed ctslb completely anterior to dlx3 (Figure 8).These results reveal that untreated embryos and those in which only this CaMKII mixture is overexpressed are mostly normal (Figure 8A,B,E).Wnt11 morphants resulted in over 60% of embryos having fully defective CE (Figure 8C,E).Wnt11 morphants that co-expressed CaMKII had a reduction in defective CE to less than 20% (Figure 8D,E).These results show that mildly active CaMKII partially rescues the CE defects of Wnt11 morphants and is consistent with previous studies in which constitutively active CaMKII can rescue developmental defects at an intermediate level, but can be inhibitory at higher levels. 19ntracellular Ca 2+ elevation. 15,17,43Our findings provide genetic evidence that specific CaMKIIs are activated by Ca 2+ to enable CE.We have determined that two of the seven CaMKII genes responsible for these stereotypical cell migrations do so without impacting cell determination, proliferation, or survival.

| DISCUSSION
CaMKII morphants and crispants exhibit a similar phenotype to morphants and mutants of Wnt5 and Wnt11.In fact, the camk2b1 morphant and crispant exhibits an undulated notochord like that seen in both the Wnt5/ppt mutant 32 and the Wnt5 morphant, 33 while the camk2g1 morphant and crispant has shorter and broader axis defects like the Wnt11/slb mutant. 9,33The defects associated with ncWnt suppression in zebrafish have been characterized as gastrulation CE defects 7 thus we sought to determine whether the defects seen in CaM-KII morphants could also be attributed to disrupted CE.In general, we conclude that the roles of the proteins encoded by the camk2b1 and camk2g1 genes during gastrulation are both necessary for CE and cannot be fully distinguished, much like Wnt5 and Wnt11.
The expression patterns of camk2b1 and camk2g1 are consistent with tissues that exhibit CE cell movements.Previous studies have shown that camk2b1 and camk2g1 are expressed during gastrulation in both the mesendoderm and neuroectoderm layers where CE cell movements occur. 20uring gastrulation, it is possible that these CaMKII gene products collaborate.CaMKIIs expressed from different genes can freely hetero-oligomerize 44 to form their stereotypical dodecamers.Over two dozen splice variants are expressed from the seven zebrafish CaMKII genes during early development. 19,20All three of the camk2g1 splice variants encode putative cytosolic targeting domains while two of the three camk2b1 variants have nuclear targeting domains.Consequently, it is possible that camk2b1 or camk2g1 products could hetero-oligomerize and target the entire CaMKII complex to one location or another.Neither morphant altered the expression levels of tbxta, dlx3, myoD, gsc, ctslb, or bmp4, suggesting that CaMKII does not have a direct influence on the transcription of relevant genes.
0][51] We saw no evidence that CaMKII disruption in early zebrafish embryos affected cell cycle progression or apoptosis.
Our findings demonstrate that both camk2b1 and camk2g1 transcripts are important during CE, even though their splice variants are distinctive, and they have different expression patterns.9][30] Mutations in the camk2g1 gene show early (G0) phenotypes that phenocopy morphants, but over generations its paralog, camk2g2, is upregulated, compensating for the loss of camk2g1. 22e believe that CaMKIIs encoded by the camk2b1 and camk2g1 genes during gastrulation are transiently activated by ncWnt-mediated Ca 2+ elevations to directly enable cell migration.Oscillations of Ca 2+ are known to occur during gastrulation. 16Such Ca 2+ oscillations could transiently activate CaMKII to enable focal adhesion turnover and thus cell migration, as previously shown in mammalian fibroblasts. 27FAK and paxillin are conserved in zebrafish and morphants show evidence of a role for FAK in CE. 52 Cell migration is compromised when CaM-KII is either hyperactivated or inhibited. 27Such a model in which ncWnts cause transient elevations of Ca 2+ to transiently activate CaMKII and enable focal adhesion turnover is also consistent with findings in which both Wnt5 and Wnt11 gain of function and loss of function interventions inhibit CE. 41 The cyclical activation and deactivation of CaMKII in response to transient Ca 2+ elevations provides a logical explanation for these otherwise contradictory results.
Further evidence of a role for CaMKII in migration, not tissue specification, comes from the striking phenotype of MB.MBs were not seen when CaMKII was uniformly inhibited by MOs or uniformly activated by TG.Rather, the expression of a cytosolically targeted dominant-negative CaMKII is mosaic and random and ectopic CaMKII is known to hetero-oligomerize with endogenous CaMKII.This not only supports a primary role for CaMKII in embryonic cell migration but suggests that mosaic CaMKII mutations can exert a dominant effect on the surrounding tissue in preventing complete dorsal midline closure.
In conclusion, we believe that ncWnts initiate cycles of transient Ca 2+ elevations which enable cycles of CaM-KII activation and inactivation allowing cycles of focal adhesion assembly and disassembly which are required for cell migration.Since human birth defects are sometimes associated with incomplete tissue tube closure or midline assembly of organ precursors, CaMKII may be a responsible molecule worth further examining.

| Zebrafish strains and microinjections
Wild-type zebrafish embryos (AB, WIK, and AB/Tübingen Strains) were obtained through natural matings, raised at 28.5 C and staged as described. 53Embryos were injected at the one-to four-cell stage with $1 nL using freshly pulled micropipettes attached to a precision pressurized injector. 53Injection volume was calculated from micrometer measurements after injection into mineral oil.

| Morpholino oligonucleotides
Morpholino antisense oligonucleotides used in this study were designed to disrupt translation by complementarity to known translational start sites of zebrafish CaMKII and Wnt11 mRNAs.MOs were purchased from Gene Tools (Philomath, OR) and are shown for each MO in the 5 0 to 3 0 direction with complementary mRNA nucleotide positions indicated relative to the underlined start codon.
Morpholino stocks (1 mM) were stored at À80 C. Prior to injection, MO aliquots were heated to 65 C for 5 min, cooled to room temperature and then diluted in Danieau buffer. 36In this study, translation-blocking antisense MOs were used at 3 ng (camk2b1), 1.5 ng (camk2g1), and 3 ng (wnt11).The control mismatch MO was used at 3 ng.

| Validation of MOs
The Wnt11 MO was identical to that published. 33The CaMKII MOs have been used in previous studies of the zebrafish CaMKII gene family. 19,20,29,30,54In zebrafish, there are eight open reading frames that have been targeted by unique MOs.Morphants of camk2b1 and its paralog, camk2b2 showed a dose-dependent decrease in whole embryo CaMKII enzymatic activity, but only camk2b1 morphants displayed notochord undulation and axis shortening while camk2b2 morphants showed bradycardia, heart elongation and absent pectoral fins. 19Subsequent studies explored camk2g1 and camk2aKAP morphants, 30 each of which displayed unique phenotypes.Many of these traits have also been observed using the reversible CaMKII inhibitor, KN-93, 29 or dominantnegative CaMKII constructs expressed universally or targeted using tissue-specific promoters.CaMKII morphants have been rescued by overexpressing WT human CaMKII constructs, 19 which are resistant to translation-blocking MOs.CRISPR-based G0 mutants (crispants) of camk2g1 22 and camk2b1 (Figure S1) phenocopy morphants.

| CaMKII constructs
Full-length zebrafish wild-type and constitutively active (T 287 D) CaMKII δ1 C coding sequences were inserted into the pCS2P + vector and mRNAs were synthesized using SP6 RNA polymerase.RNA was purified, quantitated and either 250 pg of WT or T 287 D or a 150:100 pg mixture of wild-type and T 287 D CaMKII was injected.Full-length dominant-negative (K 43 A) GFP-tagged human CaMKII δ1 C was inserted into the pFRM2.1 vector under the control of the carp β-actin promoter 56 and injected at the one-cell stage.This GFP-CaMKII construct was identical to that previously used in zebrafish studies 19,30 other than the point mutation.
4.6 | Whole-mount in situ hybridization, probes, mRNA, and cDNA clones Embryos were fixed and hybridized with $0.5 kb digoxigenin-labeled antisense riboprobes and then developed as previously described. 19The myoD and gsc probes were prepared from cDNA clones provided by Dr. J. A. Lister, the tbxta and bmp4 probes from clones provided by Dr. D. M. Garrity and the ctslb and the dlx3 probes from cDNA provided by Drs.B. and C. Thisse.

| Gap angle measurements
Embryos were viewed laterally, anterior to the top, at the 10ss.The angle between the anterior and posterior end of 10ss embryos was measured by drawing lines from the most anterior part of the head and the most posterior part of the tail, to a center point of the embryo.Angles were quantified using Nikon Elements software.This protocol is a modified version of head-trunk angle assays used during developmental staging 53 and has previously been used to describe CE defects. 34,358 | Length:width measurements Length measurements were collected starting at the most anterior region of the first somite and extending to the posterior region of the sixth somite, with width measurements taken from the sixth somite (denoted by solid lines; Figure 3D).

| TG treatment
TG (Sigma, St Louis, MO) was dissolved in dimethyl sulfoxide and diluted in Danieau Buffer at the indicated concentrations.Embryos were treated for 2-h periods as indicated and then washed out by transferring embryos into fresh 1Â Danieau Buffer.

| CaMKII activity assays
Whole embryos were lysed and total CaMKII activity was assessed by measuring phosphate incorporation into the substrate, autocamtide-2, as previously described. 20he level of activity in the absence of Ca 2+ /CaM as a percentage of that in the presence of Ca 2+ /CaM yields percent autonomy.

| Immunolocalization
Whole-mount immunostaining in zebrafish embryos was used to determine mitotic indices as previously described. 37Embryos were fixed in 4% formaldehyde/ PBS, followed by 100% methanol and then immunostained as previously described. 29Rabbit anti-phospho Serine 10 Histone H3 (Life Technologies, Carlsbad, CA) and rabbit anti-cleaved caspase 3 (Cell Signaling Technologies, Beverly, MA) were used at 10 μg/mL.Goat antirabbit Alexa Fluor 488 (Life Technologies, Carlsbad, CA) was used at 1 μg/mL.Embryos were counter-stained for 1 h at room temperature with 1 μg/mL propidium iodide (Sigma, St Louis, MO).Washed embryos were mounted in low melting point agarose and imaged by confocal microscopy (Nikon C1 Plus two-laser) on a Nikon Eclipse Ni compound microscope using a 20Â dry objective or on a Nikon AZ-100 macro zoom fluorescent stereo microscope using NIS-Elements AR 4.50 software (Nikon Instruments, Melville, NY).

| Statistical analysis
Each figure represents conclusions from at least five replicate experiments.Quantitation reflected average values from at least three of these separate experiments that had matched control and experimental embryos at the same developmental stage.The sum of individual embryos per condition is the "n" value.Gap angle and length measurements were determined using NIS-Elements software (Nikon).Proliferation was determined by averaging the percentage of mitotic cells from multiple regions of interest per embryo across multiple embryos and replicate experiments.Statistical analysis was performed using Excel (Microsoft, Redmond, WA) t-test functions which were two-tailed, unpaired and assumed equal variance.An asterisk denotes statistically significant differences with p-values indicated in figure legends.

F
I G U R E 1 camk2b1 and camk2g1 morphant phenotypes.(A-C).Lateral view of 24hpf zebrafish embryos injected with 3 ng mismatch morpholino oligonucleotide (MO), 3 ng camk2b1 MO, or 1.5 ng camk2g1 MO.Control embryos show evenly spaced somites (arrow) with full yolk extension (asterisk).Camk2b1 morphants exhibit compression of the somites, curvature of the tail and reduced yolk extension.Camk2g1 morphants lack yolk and tail extension and exhibit compression of the somites.Scale bar 0.5 mm.(D-F) Lateral view images at the 10 somite stage (ss) of representative morphant embryos.Body gap angles were derived from the head-yolk-tail angle (dashed lines, D).Scale bar 0.25 mm.(G) Body gap angles of embryos at the 10-11ss were averaged from 34 to 80 embryos per condition across three replicated matched studies.****p < .0001compared to control.

F
I G U R E 2 camk2b1 and camk2g1 morphants display hypoblast movement defects.Dorsal views of representative morphant embryos injected with 3 ng mismatch morpholino oligonucleotide (MO), 3 ng camk2b1 MO, or 1.5 ng camk2g1 MO and then probed for myoD and tbxta at 90% epiboly (A-C) and at the 6-7 somite stage (ss) (D-F).(G) length:width (L:W) ratios at the 6-7ss were averaged from 24 to 29 embryos per condition across three replicated matched studies.Length was measured from the anterior region of somite 1 to the posterior region of somite 6 (denoted by vertical white line) and width of somite 6 (denoted by horizontal white line).Scale bar 0.05 mm.***p < .001.

F
I G U R E 3 camk2b1 and camk2g1 morphants display convergent extension defects at the tailbud stage.Representative morphant embryos injected with 3 ng mismatch morpholino oligonucleotide (MO), 3 ng camk2b1 MO, or 1.5 ng camk2g1 MO were probed at the 1-3 somite stage simultaneously for the expression of ctslb, dlx3, and tbxta.Dorsal views of representative embryos at their (A-C) anterior and (D-F) posterior ends.Arrows indicate dlx3 expression.White bar denotes band of tbxta expression across developing notochord.

F
I G U R E 4 Cell fate specification is unaltered in CaMKII morphants.Representative morphant embryos injected with 3 ng mismatch morpholino oligonucleotide (MO), 3 ng camk2b1 MO or 1.5 ng camk2g1 MO are shown at shield (A-F) and 75% epiboly (G-I) stages with dorsal to the right.(A-C) Animal pole view probed for the dorsal side organizer gsc.(D-I) Lateral view probed for bmp4 expression.

F I G U R E 5
Proliferation and Apoptosis are unaltered in CaMKII morphants.Lateral views of representative morphant embryos injected with 3 ng mismatch morpholino oligonucleotide (MO), 3 ng camk2b1 MO. or 1.5 ng camk2g1 MO after immunostaining for phospho-histone H3 (green, left panel) or cleaved caspase-3 (green, right panel) and propidium iodide (red) at shield and 75% epiboly stages.Proliferation (center panel) was the percentage of cells in mitosis at both shield and epiboly averaged from three experiments, at least 3 embryos per condition and five regions of interest per embryo.There were no statistically significant differences between conditions.Cleaved caspase-3 positive cells are marked by an asterisk.F I G U R E 6 Ectopic expression of dominant-negative CaMKII induces convergent extension (CE) defects.Embryos were injected with GFP-tagged dominant-negative (K 43 A) CaMKII cDNA and observed at 24hpf in differential interference contrast light (A-C) and fluorescent (D-F) optics.Embryos with normal, compressed, and split morphologies are shown and were also fixed and probed for tbxta and myoD expression (G-I).Individual fluorescent cells can be seen between the split axes and are marked by an asterisk (C, F).Bar chart represents the accumulated categorization of 58-93 embryos for each level of cDNA injected which spanned seven separate experiments.

F I G U R E 7
Persistent Ca 2+ elevations during gastrulation activate CaMKII and cause cell migration defects.Zebrafish embryos were treated with thapsigargin (TG) between 5 and 7hpf, washed free of TG and then harvested or imaged later.(A-H) Images of 10ss and 24hpf embryos after transient treatment with the indicated concentration of TG-treated embryos.Scale bar 0.25 mm (A) and 0.5 mm (E).(I) Body gap angles of control dimethyl sulfoxide and TG-treated embryos were measured at the 10 ± 1 somite stage (ss) and averaged from 17 to 54 embryos, ***p < .001.(J) CaMKII-specific activity was measured in whole embryo lysates of 12-16ss and 24hpf embryos after transient TG treatment using the autocamtide-2 peptide-based assay.CaMKII autonomy represents the percentage of whole embryonic CaMKII that is activated and thus Ca 2+ /CaM-independent.(K-N) Dorsal and lateral views of representative 15ss embryos either untreated or transiently treated with TG as described above and then probed for myoD and tbxta.Anterior is at top.Scale bar 0.5 mm.
CE cell movements during embryonic development require ncWnts such as Wnt5 and Wnt11 and lead to F I G U R E 8 Overexpression of CaMKII partially rescues the convergent extension (CE) defect of Wnt11 morphants.(A-D) Morphant embryos injected with 3 ng wnt11 morpholino oligonucleotide with or without CaMKII mRNA are shown at the 1-3 somite stage after in situ hybridization with a combination of the ctslb, dlx3, and tbxta probes as in Figure 3. (E) Individual examples of normal, partial, and defective CE with ctslb denoted by a white arrow and dlx3 by black arrows.(F) The percentage of embryos for each condition were scored for normal CE, partial CE defect or CE defect.Three separate experiments yielded 55-107 embryos for each condition.