923RESEARCH ARTICLE

INTRODUCTIONDuring early vertebrate embryogenesis, the dorsal organizer playsan important role in the formation of the dorsoventral (DV) andanteroposterior (AP) body axes. The molecular mechanisms thatcontrol the zygotic gene cascades that induce the organizer havebeen elucidated (De Robertis, 2006; De Robertis et al., 2000;Gerhart et al., 1989; Hibi et al., 2002; Schier and Talbot, 2005).However, the mechanisms that initially specify the dorsal axis andinduce zygotic genes that are required for dorsal organizer formationare not fully understood. In amphibian and teleost embryos,specification of the dorsal axis begins soon after fertilization. InXenopus, the egg cortex rotates relative to the sperm entry pointduring the first cell cycle (cortical rotation). During this process,small particles and organelles are transported along a microtubulearray that is oriented with the plus ends towards the prospectivedorsal side (Elinson and Rowning, 1988; Houliston and Elinson,1991). In zebrafish, the sperm enters at the animal pole. A parallelmicrotubule array forms at the vegetal pole of the yolk ~20 minutespost-fertilization (mpf) (Jesuthasan and Stahle, 1997). Disruption ofthe microtubule array by treatment with nocodazol, cold or UVirradiation prior to the 32-cell stage leads to loss of the dorsalorganizer and to the ventralization of embryos (Jesuthasan andStahle, 1997). Removal of the vegetal yolk mass at the early 1-cellstage also results in severely ventralized embryos (Mizuno et al.,1999; Ober and Schulte-Merker, 1999). These studies establishedthe hypothesis that the dorsal determinants (DDs) are initially

localized to the vegetal pole and then transported along microtubulesto the prospective dorsal side, where they are incorporated by dorsalblastomeres and promote dorsal organizer formation (see Fig. S1 inthe supplementary material). However, the molecular identity of theDDs, and the mechanisms that localize them to the vegetal pole andmediate their subsequent transport to the prospective dorsalblastomeres, remain unknown.

Although the molecular nature of the DDs is not clear, they areknown to activate the canonical Wnt pathway, which leads tonuclear accumulation of b-catenin. Activation of the canonical Wntpathway in Xenopus and zebrafish embryos results in ectopicformation or expansion of the dorsal organizer, whereas itsinhibition impairs dorsal axis formation and reduces dorsal-specificgene expression (reviewed by Hibi et al., 2002). These data placecanonical Wnt pathway function downstream of the DDs. Inzebrafish, the b-catenin accumulates in the nuclei of dorsalblastomeres by the 128-cell stage and in the dorsal blastoderm anddorsal yolk syncytial layer of mid-blastula stage embryos (Douganet al., 2003; Schneider et al., 1996). Thus, the DDs activate canonicalWnt signaling in the dorsal blastomeres by the 128-cell stage. A b-catenin and Tcf/Lef complex is thought to regulate zygoticexpression of dorsal-specific genes, including the homeobox genedharma (dha, also known as bozozok) and the nodal-related genendr1 (also known as squint) (Dougan et al., 2003; Leung et al., 2003;Ryu et al., 2001; Shimizu et al., 2000a), which function in dorsalorganizer formation (Shimizu et al., 2000a; Sirotkin et al., 2000).Thus, the DDs induce the expression of these dorsal-specific genesthrough activation of the canonical Wnt pathway and therebypromote dorsal organizer formation.

Genetic studies in zebrafish have revealed maternal factors thatare involved in dorsal determination. The recessive maternal-effectmutant ichabod shows impaired maternal expression of b-catenin 2(ctnnb2; one of two b-catenin genes in zebrafish), defective dorsalorganizer formation and severe ventralization (Bellipanni et al.,2006; Kelly et al., 2000). The maternal-effect mutant brom bones,which has a nonsense mutation disrupting hnRNP I (ptbp1a –

Development 137, 923-933 (2010) doi:10.1242/dev.046425© 2010. Published by The Company of Biologists Ltd

1Laboratory for Vertebrate Axis Formation, RIKEN Center for Developmental Biology,Hyogo 650-0047, Japan. 2Department of Developmental and Molecular Biology,Albert Einstein College of Medicine, New York, NY 10461, USA. 3Department ofBiological Sciences and GRAST, Chungnam National University, Daejeion 305-764,Korea. 4Electron Microscope Laboratory, RIKEN Center for Developmental Biology,Hyogo 650-0047, Japan. 5Bioscience and Biotechnology Center, Nagoya University,Nagoya 464-8601, Japan.

*Author for correspondence (hibi@bio.nagoya-u.ac.jp)

Accepted 12 January 2010

SUMMARYIn amphibian and teleost embryos, the dorsal determinants (DDs) are believed to be initially localized to the vegetal pole and thentransported to the prospective dorsal side of the embryo along a microtubule array. The DDs are known to activate the canonicalWnt pathway and thereby promote the expression of genes that induce the dorsal organizer. Here, by identifying the locus of thematernal-effect ventralized mutant tokkaebi, we show that Syntabulin, a linker of the kinesin I motor protein, is essential for dorsaldetermination in zebrafish. We found that syntabulin mRNA is transported to the vegetal pole during oogenesis through the Buckyball (Buc)-mediated Balbiani body-dependent pathway, which is necessary for establishment of animal-vegetal (AV) oocyte polarity.We demonstrate that Syntabulin is translocated from the vegetal pole in a microtubule-dependent manner. Our findings suggestthat Syntabulin regulates the microtubule-dependent transport of the DDs, and provide evidence for the link between AV anddorsoventral axis formation.

KEY WORDS: Dorsal determination, Syntabulin, Kinesin, Microtubules, Balbiani body, Bucky ball, Zebrafish

Syntabulin, a motor protein linker, controls dorsaldeterminationHideaki Nojima1, Sophie Rothhämel2, Takashi Shimizu1, Cheol-Hee Kim3, Shigenobu Yonemura4,Florence L. Marlow2 and Masahiko Hibi1,5,*

DEVELO

PMENT

924

Zebrafish Information Network), shows egg activation defects,disorganized vegetal microtubule array formation, and subsequentlydisplays ventralization (Mei et al., 2009). Furthermore, a ubiquitinligase, tripartite motif-containing 36 (trim36), is localized to thevegetal cortex of Xenopus oocytes, and when depleted causesdefective vegetal microtubule array formation and ventralization(Cuykendall and Houston, 2009). Cumulatively, these data supportthe hypothesis that microtubule-dependent transport of DDs playsan essential role in canonical Wnt pathway activation and dorsal axisdetermination in teleost and amphibian embryos.

The DDs are present at the vegetal pole of eggs prior tofertilization and therefore achieve their vegetal pole localizationduring oogenesis. In Xenopus and zebrafish oocytes, mRNAslocalize to the oocyte vegetal pole via overlapping transportpathways. The early pathway involves recruitment of mRNAs,including germline-specific transcripts to the Balbiani body (Klocet al., 2001; Kloc and Etkin, 1995; Kloc et al., 1996; Wilk et al.,2005), an evolutionarily conserved aggregate of organelles andmolecules that is present in early oocytes (Guraya, 1979; Kloc et al.,2004). Balbiani body formation is controlled by Bucky ball (Buc).Oocytes lacking Buc function fail to localize mRNAs to the vegetalpole in early and late stage oocytes, and show defective animal-vegetal (AV) polarity (Bontems et al., 2009; Marlow and Mullins,2008). However, the relationship between the Balbiani body-dependent RNA transport system and vegetal pole localization of theDDs has not been fully elucidated.

The zebrafish maternal-effect recessive mutation tokkaebi (tkk)affects the formation of dorsal structures (Nojima et al., 2004).Embryos obtained from tkk homozygous females (tkk embryos)show severely ventralized phenotypes. Accumulation of b-cateninin the dorsal nuclei at mid-blastula stages is impaired in tkk embryos,indicating that tkk functions upstream of activation or stabilizationof b-catenin in the maternal Wnt signaling pathway. In this report,we show that the tkk locus encodes a zebrafish ortholog ofSyntabulin, a linker molecule that attaches cargo to the motor proteinkinesin I in neuronal axons (Cai et al., 2005; Cai et al., 2007; Su etal., 2004). We demonstrate that syntabulin RNA is transported to thevegetal pole during oogenesis through the Buc-mediated Balbianibody-dependent pathway. We further show that Syntabulin proteintranslocates from the vegetal pole in a microtubule-dependentmanner during early embryogenesis. Our findings suggest thatSyntabulin is involved in the vegetal pole localization andmicrotubule-dependent transport of the DDs, and that Syntabulinlinks oocyte AV polarity and embryonic DV polarity.

MATERIALS AND METHODSPositional cloningtkkrk4 homozygous fish were crossed with wild-type India fish (Knapik etal., 1996) to generate F1 families. Homozygous tkkrk4 female fish wereraised from the F1 cross and identified based on the ventralized phenotypesof their offspring. Since embryos from tkkrk4/rk4 female fish with Indiagenetic background show relatively mild phenotypes and low penetrance(Nojima et al., 2004), phenotypic analysis was performed at least twice.Genomic DNA from mutant females was used to carry out segregationanalyses with SSLP markers. z1543 on chromosome 16 was found to beclose to the tkk locus (one recombination event out of 280 meiosis). Usingthe PCR primers for z1543 as a probe, BAC DKEY-97P23 was obtained.Through BAC end sequencing and rescreening of BAC clones, seven BACclones (DKEY-22J10, 42F24, 9G17, 96G15, 71M14, 106H17, 97P23) werefound to cover the tkk locus (see Fig. 2A). Further linkage analyses revealedthat four BAC clones (DKEY-42F24, 9G17, 96G15, 71M14) contained thetkk locus. The inserts in these BAC clones were sequenced by a shotgunsequencing procedure and 12 genes were found in this region. The sequence

and gene information in the region were also confirmed by the currentzebrafish genome data (http://www.sanger.ac.uk/Projects/D_rerio/).Expression of the 12 candidate genes in ovary or in 1-cell stage embryos wasanalyzed by RT-PCR (see Table S1 in the supplementary material). ThecDNAs of syntabulin and nine other maternally expressed candidates wereisolated from wild-type and tkk 1-cell stage embryos by RT-PCR and clonedinto pBluescript II SK+ (Stratagene). A 3� terminal fragment of the insertionin the syntabulin promoter was isolated by inverse PCR with primers 5�-AGTATTTAATTCTGAACAAACAGAAC-3� and 5�-GTGTG TT ATGTA -ATATGCATTTTTAC-3�, as described (Kawakami et al., 2004). Theinsertion was detected by PCR with primer1, 5�-GATC CATTCC CAA -CTTTATTTTCTCG-3� and primer2, 5�-AGTAT TTAATTCTGAAC AA -ACAGAAC-3� (see Fig. 2G). The cDNA, genomic and insert sequences ofzebrafish syntabulin were deposited in the DDBJ databank under theaccession numbers AB517946, AB517947 and AB517948, respectively. Forphenotypic analysis, tkk embryos were obtained from crosses of young(3 month old) homozygous male and female fish in the TL geneticbackground (Nojima et al., 2004). For time course analysis, in vitrofertilization was performed (Westerfield, 1995). For studies of bucky ballmutants, bucp106re (Dosch et al., 2004; Marlow and Mullins, 2008) andbucp43bmtb (Bontems et al., 2009) alleles were used.

RT-PCRQuantitative (q) RT-PCR analysis was conducted to examine the expressionof syntabulin, the other candidate genes and Wnt genes in wild-type and tkkembryos, using an ABI Prism 7000 (Applied Biosystems) or LightCycler480 (Roche) with Power SYBR Green PCR Master Mix (AppliedBiosystems) or SYBR Green I Master (Roche), respectively. Primersused were: syntabulin, 5�-ATGGAGTTGGCGCAAAATG-3� and 5�-TCTTCGCCTCGGTAGAACTTTC-3�; zdhhc3, 5�-ATGAGCGTCAG -AACAGCATGTG-3� and 5�-AAGACGAGAAACCAGGTGATGATC-3�;e(y)2, 5�-GTTGCTGGAGTTACTCCCAAAGG-3� and 5�-GCGAGAA -AGGCTCTTATTCTCTGA-3�; gapdh, 5�-GGATCTGACAGTCC GTCTT -GA GA-3� and 5�-TGCAGCCTTGACGACTTTCTT-3�; for primerinformation for the other candidate genes, see Table S2 in the supplementarymaterial. All PCR assays were performed in triplicate. The relative mRNAlevel of the averaged CT (cycle threshold, ABI Prism 7000) or CP (crossingpoint, LightCycler 480) was quantified and normalized to gapdh. Fordetection of syntabulin in wild-type and buc mutant oocytes (see Fig. 6),ovaries were dissected from AB, bucp106re and bucp43bmtb mutant female fish.Fifty stage I/II or III oocytes were manually sorted for RNA isolation andcDNA preparation. qRT-PCR was carried out using the Realplex2 System(Eppendorf) with Power SYBR Green PCR Master Mix. Primers used were:syntabulin, 5�-CACACTGATGCTGTGCTGAA-3� and 5�-TCCTGCGA -GTGAATGAGAGA-3�; ef1a, 5�- AGCCTGGTATGGTTGTGACCTTCG-3� and 5�-CCAAGTTGTTTTCCTTTCCTGCG-3�. The relative mRNAlevel of the averaged CT was quantified and normalized to ef1a.

In situ hybridizationWhole-mount in situ hybridization was performed as described previously(Thisse and Thisse, 1998) except that hybridization was at 65°C. BM purpleAP substrate (Roche) was used as the substrate for alkaline phosphatase.dharma (bozozok) (Yamanaka et al., 1998) and dusp6 (mkp3) detection wasas described previously (Tsang et al., 2004). For the syntabulin probe, the3�UTR of syntabulin was amplified by PCR using the primers 5¢-GCTCTAGATAGCGGCCCTCACCTACTTCTACAAC-3¢ and 5¢-ACG -CGTCGACTTTTCCATGCTTTAATAATAAATATA-3¢. The BM purplesignals were acquired using an AxioPlan2 or AxioSkop2 microscopeequipped with an AxioCam CCD camera (Zeiss), or using an Olympus SZ16fluorescence dissecting microscope with a Microfire digital camera(Olympus). Images were processed in Photoshop (Adobe), Illustrator(Adobe) and Canvas (ACD systems). In situ hybridization on slides wasessentially as described previously (Marlow and Mullins, 2008). To obtaineggs, wild-type and mutant females were anesthetized in Tricaine asdescribed (Westerfield, 1995). Eggs were activated by adding embryomedium, and then fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) 2 or 5 minutes after activation.

RESEARCH ARTICLE Development 137 (6)

DEVELO

PMENT

ImmunostainingFor b-catenin and Syntabulin staining, embryos were fixed with 4% PFA inPBS for 24 hours at 4°C. For microtubule staining, embryos were fixed witha microtubule-stabilizing buffer: 80 mM potassium Pipes pH 6.8, 5 mMEGTA, 1 mM MgCl2, 3.7% formaldehyde, 0.25% glutaraldehyde, 0.2%Triton X-100 (Schroeder and Gard, 1992). The dechorionated embryos werewashed with PBS containing 0.1% Triton X-100 (PBS-T) and blocked with1% bovine serum albumin (BSA) in PBS-T for 1 hour. Detection of b-catenin was as described previously (Nojima et al., 2004), and the immunecomplexes were visualized with diaminobenzidine (DAB, Sigma). Fordetection of Syntabulin, the samples were incubated with anti-Syntabulinantibody solution (anti-Syn1, 1/50 dilution of hybridoma supernatant; anti-Syn2, 1/500 of hybridoma supernatant; anti-SynP, 1/500 of affinity-purifiedantibody) at 4°C overnight. For detection of microtubules, the samples wereincubated with mouse monoclonal anti-b-tubulin antibody solution (1/1000,Chemicon KMX-1). After four washes with PBS-T, the samples wereincubated with secondary antibodies [1/500, Alexa Fluor 555 goat anti-mouse or goat anti-rabbit IgG (H+L), Molecular Probes]. The DAB signals(b-catenin) and fluorescent images (Syntabulin) were obtained byAxioPlan2 imaging. Microtubule images were taken with an LSM5 Pascallaser-scanning inverted microscope (Carl Zeiss) and constructed fromz-stack sections by a 3D projection program associated with the microscope.

Plasmid construction and transgenesisThe Tol2 rescue plasmids were constructed using Gateway technology(Invitrogen) in pT2KDest-RfaF, which is derived from pT2KXIG(Kawakami et al., 2004). Further information about the plasmids is availableon request. The promoter of zona pellucida protein C (zpc), which is derivedfrom zp0.5GFP, was described previously (Onichtchouk et al., 2003). Thepolyadenylation signal (pAS) is derived from pCS2+. GFP refers tommGFP5 derived from zp0.5GFP. Venus is a modified YFP and is derivedfrom pCS2+venus (Nagai et al., 2002). The 2A peptide sequence is derivedfrom porcine teschovirus-1 (PTV1) (Provost et al., 2007) and the DNAfragment is generated with a synthesized oligonucleotide: 5�-GTCGACGGATCAGGCGCTACGAATTTCTCGCTACTGAAGCAGGC -TGGAGATGTAGAGGAAAACCCGGGACCGGCCATGG-3�(linkers areunderlined), which corresponds to GSGATNFSLLKQAGDVEENPGP. Toconnect syntabulin and Venus with the 2A peptide sequence, the stop codonof syntabulin was mutated to introduce a SalI site. The 2A peptide sequenceand Venus were linked at an NcoI site. To make transgenic fish, 25 pg of Tol2transgene plasmid DNA and 25 pg of transposase RNA were injected into1-cell stage embryos obtained from crosses of wild-type or relatively old tkkpairs. Expression plasmids for 6�Myc-tagged Syntabulin and 3�HA-tagged Kif5b were constructed in pCS2+.

AntibodiesTo raise monoclonal and polyclonal antibodies against Syntabulin,glutathione S-transferase (GST) fusion proteins containing amino acids 14-98 (Syn1 antigen) or amino acids 102-271 (Syn2 antigen) of Syntabulin, ora His-tagged protein containing amino acids 1-120 of Syntabulin weregenerated in E. coli. The GST fusion and His-tagged proteins were purifiedand used for immunization. For generating monoclonal antibodies, BALB/cmice were immunized four or five times with ~50 mg of the proteins. Spleencells of the immunized mice were fused with the mouse myeloma lineAg8.563, and hybridomas were obtained by conventional HAT selection.Supernatants of growth-positive cells were subjected to enzyme-likedimmunosorbent assays (ELISAs) to identify hybridomas producing specificantibodies. A polyclonal antibody (anti-SynP) was generated by immunizingrabbits with the His-tagged protein six times. The antibody was purifiedusing an affinity column that was covalently conjugated to the antigen.

Transfection, immunoprecipitation and immunoblottingHEK293T cells in a 6-cm dish were transfected using HilyMax (DOJINDOLaboratories). The cells were lysed at 24 hours post-transfection in 1 ml oflysis buffer: 10 mM HEPES-HCl pH 7.4, 150 mM NaCl, 1% NP40, 1 mMEDTA, protease inhibitor cocktail (Nacalai). The lysates were precipitatedwith anti-HA antibody (3F10, Roche), resolved by 5-20% SDS-PAGE(SuperSep, Wako, Japan), and blotted with either anti-HA (3F10, Roche) oranti-Myc (9E10, SantaCruz) antibodies. The proteins were detected with an

HRP-conjugated rabbit anti-rat IgG (H+L) (Zymed) or HRP-conjugatedanti-mouse IgG (Mouse TrueBlot, eBioscience) antibody, using achemiluminescence system (Western Lightning, Perkin Elmer Life Science).For detection of Syntabulin in embryos, 200 embryos were lysed in 700 mlof lysis buffer by grinding on ice. The lysates were cleared by centrifugationand immunoprecipitated with anti-Syntabulin antibodies (1 ml of ascites ofmonoclonal antibodies).

Treatment with microtubule inhibitorsStock solutions of nocodazole (Sigma; 2 mg/ml in DMSO) and colchicine(Sigma; 100 mg/ml in ethanol) were diluted to 1 mg/ml or 0.4 mg/ml,respectively, in embryo medium (Westerfield, 1995). Embryos were placedin inhibitor solutions with their chorion intact, just after in vitro fertilization,for the times indicated.

Electron microscopyElectron microscopy (see Fig. S2 in the supplementary material) was carriedout essentially as described previously (Shimizu et al., 2005; Yonemura etal., 2002).

StatisticsStatistical analyses were performed by two-tailed Student’s t-test usingGraphPad Prism 5. P<0.01 was considered significant.

RESULTSPositional cloning of the tokkaebi gene reveals anovel role for Syntabulin in dorsal axis formationtokkaebi (tkk) is a maternal-effect recessive mutation that disrupts agene required for dorsal axis formation in zebrafish. The progeny oftkk homozygous females (tkk embryos) display severely ventralizedphenotypes, which are more severe in the TL genetic background(Nojima et al., 2004). We reassessed the expression of the earliestknown zygotic dorsal-specific genes and nuclear accumulation of b-catenin in tkk embryos obtained after backcrossing with TL fish.Approximately 80% of tkk embryos from young females showedcomplete ventralization at the pharyngula stage [24 hours post-fertilization (hpf); class C1, Fig. 1A]. In tkk embryos, expression ofthe earliest known dorsal-specific genes dharma (dha) and dual-specificity phosphatase 6 [dusp6, also known as mitogen-activatedprotein kinase phosphatase 3 (mkp3)], two direct targets of thecanonical Wnt pathway (Ryu et al., 2001; Tsang et al., 2004;Yamanaka et al., 1998), was absent or strongly reduced at mid-to-late blastula stage (4 hpf, Fig. 1H-K). Dorsal accumulation of b-catenin at the 256-cell stage was observed in wild-type, but not intkk, embryos (Fig. 1F,G). A parallel microtubule array that is thoughtto play an important role in dorsal determination (Jesuthasan andStahle, 1997) was present at the vegetal pole at 20 mpf in wild-type(Fig. 1L) and in tkk (Fig. 1M) embryos. Electron microscopyrevealed a similar composition of organelles, including ribosomesat the vegetal pole, of wild-type and tkk embryos at 20 mpf (see Fig.S2 in the supplementary material), indicating that the vegetal yolkis not significantly affected in tkk embryos. Together, these datasuggest that the tkk gene product is dispensable for vegetal polemicrotubule formation, but functions in early embryogenesis oroogenesis to promote nuclear accumulation of b-catenin andexpression of zygotic genes required for dorsal organizer formation,including dha.

To reveal how the tkk gene product contributes to dorsaldetermination, we carried out positional cloning. We mapped the tkklocus to a 600 kb region of chromosome 16 that contains 12 potentialopen reading frames (Fig. 2A). Although ten of the 12 genes werematernally expressed, none had a nonsense or missense mutation intheir coding region (see Table S1 in the supplementary material forsequence information; data not shown). However, using qRT-PCR

925RESEARCH ARTICLEDorsal termination by a motor protein linker

DEVELO

PMENT

926

analysis, we found that the maternal expression of one candidategene was strongly reduced in tkk embryos (Fig. 2B and see Table S1in the supplementary material). The gene encodes the zebrafishortholog of Syntabulin (Su et al., 2004), based on strong sequencesimilarity to human syntabulin (see Fig. S3 in the supplementarymaterial) and on synteny between the tkk chromosomal region inzebrafish and the human syntabulin chromosomal region(chromosome 8, data not shown).

Expression of zebrafish syntabulin remained strongly reduced at24 hpf in tkk as compared with wild-type embryos (see Fig. S4 in thesupplementary material). Whole-mount in situ hybridizationrevealed vegetal pole localization of syntabulin mRNA in wild-typeeggs and embryos through the 16-cell stage (Fig. 2C-E see Fig. S4in the supplementary material), and its absence in tkk eggs andembryos (Fig. 2F). By genomic PCR and Southern blot analyses, wefound an insertion in the syntabulin promoter region (232 bpupstream of the transcription initiation site) in the tkk mutant

genome (Fig. 2G and see Fig. S5 in the supplementary material).The 3� sequence of the insertion shows similarities to variousgenome regions and to the pol-like protein of snail and Drosophilaretrotransposons (accession numbers ABN58714, AAA70222) atthe amino acid level, implying that the insertion is a transposableelement. Our expression data suggest that this insertion in thepromoter region strongly suppresses transcription of the syntabulingene.

Maternal Syntabulin is sufficient to rescueventralization of tkk mutant progenyTo address whether the strong reduction of syntabulin expressioncauses the ventralization of tkk embryos, we attempted to rescuethe mutant phenotype by injecting syntabulin mRNA into 1-cellstage embryos. Supplying syntabulin after egg activation was notsufficient to rescue ventralization in tkk mutant progeny (data notshown). The failure of these rescue attempts, and the localizationof syntabulin transcripts at the vegetal pole of early embryos (Fig.2C-E), suggest that Syntabulin functions in the zygote duringearly cleavage stages or earlier during oogenesis. With this inmind, we generated transgenic fish that express syntabulin in tkkoocytes (Fig. 3). We constructed the rescue plasmid in the Tol2vector (Kawakami et al., 2004) to contain the syntabulin cDNA,the first intron and 3� UTR, GFP, and the zona pellucida proteinC (zpc; zpcx) promoter, which drives maternal expression(Onichtchouk et al., 2003) (Fig. 3A). The rescue plasmid andtransposase RNA were injected into embryos obtained fromcrosses of relatively old tkk homozygotes, the offspring of whichshow only mild, or no, ventralized phenotypes (Nojima et al.,2004). These fish were raised to adulthood (see Fig. S6 in thesupplementary material). Maternal expression of the transgenewas observed in a subset of progeny from these tkk mutantfemales based on GFP visualization at the 1-cell stage, prior toactivation of zygotic gene expression. In total, from three crosses80.7% of GFP-negative (non-transgenic) tkk embryos werecompletely ventralized, and only 9.3% showed morphologicallynormal DV patterning (Fig. 3B) at 24 hpf. By contrast, in theGFP-positive (transgenic) class, only 43.7% of the embryos werecompletely ventralized, whereas 40.3% displayed normal DVpatterning (Fig. 3B). dha expression was also restored intransgenic tkk embryos (Fig. 3C,D). These data indicate thatmaternal syntabulin rescues the tkk axis defects, and provideevidence that reduced syntabulin expression in oocytes causesventralization of tkk progeny.

Exon-intron structure of syntabulin is required forvegetal pole localization and proper dorsaldeterminationThe 3� UTR of germline-specific mRNAs, such as vasa, nanos anddazl, is necessary and sufficient for their vegetal pole localization(Kosaka et al., 2007). In tkk embryos rescued by maternalexpression of syntabulin mRNA containing the first intron and 3�UTR, the mRNA was not vegetally restricted, but was insteaddistributed throughout the embryos (Fig. 3E,F). Furthermore,EGFP-3� UTR RNA was not localized to the vegetal pole intransgenic wild-type or tkk zygote-stage embryos (Fig. 4A-E).These data suggest that the 3� UTR of syntabulin is not sufficientfor its vegetal pole localization. In Drosophila oocytes, oskarmRNA localization at the posterior pole, which is required foroogenesis, is controlled in part by splicing (Hachet and Ephrussi,2004). Therefore, we examined whether the exon-intron structureof the syntabulin gene is required for the vegetal localization of

RESEARCH ARTICLE Development 137 (6)

Fig. 1. Maternal-effect tokkaebi (tkk) mutant zebrafish embryosshow defects in dorsal determination. (A-E) Phenotypes of tkkembryos (A-D) versus wild type (WT; E) at 24 hpf. Lateral views withdorsal to the right (A-D) or to the top (E). Embryos were classified intofour classes (C1-4) based on their ventralized phenotypes. C1 embryosdisplayed a completely ventralized phenotype. C2 embryos were similarto C1 embryos but had somites in the trunk region. C3 embryos hadsomites and spinal cord in the trunk. C4 embryos had somites, spinalcord and hindbrain, but lacked notochord or neuroectoderm anterior tothe midbrain. The percentage of tkk embryos from typical youngfemales that showed each phenotype is indicated.(F,G) Immunostaining of b-catenin at the 256-cell stage in wild-type(F; 88%, n=26) and tkk (G; 11%, n=35) embryos. Dorsal views. Arrowsindicate nuclear accumulation of b-catenin in dorsal blastomeres.(H,I) Expression of dha at sphere stage (4 hpf) in wild-type (H; 100%,n=36) and tkk (I; 8%, n=40) embryos. (J,K) Expression of dusp6 atsphere stage in wild-type (J; 100%, n=29) and tkk (K; 6%, n=36)embryos. Animal pole views. (L,M) Immunostaining with anti-b-tubulinantibody. Wild-type (L; 86%, n=14) and tkk (M; 81%, n=16) embryosat 20 mpf. Vegetal pole views. Scale bars: 200mm in A-E,H-K; 140mmin F,G; 20mm in L,M.

DEVELO

PMENT

syntabulin mRNA in zebrafish. The syntabulin gene is composedof five exons and four introns (Fig. 4F). We constructed a rescueplasmid that contains all the exons and introns, and inserted the 2Apeptide sequence of porcine teschovirus-1 (PTV1) and Venus (aYFP variant) cDNA between the coding region and the 3� UTR(Fig. 4G). Transgenic embryos were generated by co-injectionwith transposase RNA (see Fig. S6 in the supplementary material).Of Venus-negative (non-transgenic) tkk embryos, 78.3% showedcomplete ventralization and 13% showed normal DV patterning,whereas only 30.1% of Venus-positive (transgenic) tkk embryosdisplayed complete ventralization and 59.4% displayed normalDV patterning (Fig. 4H). This result indicates that this exon/intron-containing plasmid rescued the tkk phenotypes more efficientlythan the plasmid containing the cDNA, first intron and 3� UTR.This rescue was also more efficient than that by a plasmidcontaining the cDNA and 3�UTR (data not shown). dha expressionwas also more efficiently restored by this plasmid, compared withthe cDNA plus first intron and 3� UTR construct (Fig. 4I,J).Furthermore, more syntabulin mRNA was localized to the vegetalpole in tkk embryos rescued by this plasmid (asterisk in Fig. 4Kversus 4L as a control). These findings indicate that the exon-intron structure is required for vegetal pole localization and properfunction of syntabulin.

Microtubule-dependent transport of SyntabulinSyntabulin is a linker protein that attaches cargo to kinesin I, and inneurons Syntabulin functions in cargo transport to the synapticterminal along microtubules (Cai et al., 2005; Cai et al., 2007; Su etal., 2004) (see Fig. S3 in the supplementary material). We found thatkif5b, which encodes the heavy chain of kinesin I, is expressed inzebrafish oocytes (see Fig. S7 in the supplementary material). WhenHA-tagged Kif5b and Myc-tagged Syntabulin were co-expressed inhuman embryonic kidney (HEK) 293T cells, Syntabulin co-immunoprecipitated with Kif5b (Fig. 5A). These data indicate thatzebrafish Syntabulin can interact with the motor protein kinesin Ithat functions to transport cargo along microtubules.

We next examined the localization of Syntabulin protein. Weraised one polyclonal and two monoclonal antibodies againstSyntabulin (see Fig. S8 in the supplementary material). All theantibodies detected Syntabulin at the vegetal pole in the wild typebut not in tkk embryos from just after fertilization through 20 mpf(Fig. 5B-G and see Fig. S8C,D in the supplementary material). After20 mpf, when microtubule array formation occurs at the vegetal pole(Jesuthasan and Stahle, 1997), Syntabulin redistributed to one sideof the vegetal pole (Fig. 5H-J). This biased localization ofSyntabulin was detected through 50 mpf (the 2-cell stage, Fig. 5J).Furthermore, Syntabulin remained at the vegetal pole in embryos

927RESEARCH ARTICLEDorsal termination by a motor protein linker

Fig. 2. syntabulin is responsible for the tkk mutant phenotype. (A) Linkage mapping. The zebrafish tkk locus was initially mapped near theSSLP marker z1543. Seven BAC clones were used to construct a physical map. Linkage analysis (numbers indicate recombinants per meiosis),sequence analysis of BAC clones and database searches revealed that there were 12 candidate genes in the chromosomal region. (B) Quantificationof syntabulin (gene 6 in A) expression in wild-type and tkk embryos (expression in wild type taken as 100; *, P<0.01; the average ±s.d. is shown).zdhhc3 (zinc-finger, DHHC-type containing 3, gene 3) and e(y)2 (enhancer of yellow gene, gene 8) are located in the same chromosomal region assyntabulin and were used as controls. The expression of other genes in this chromosomal region is shown in Table S1 in the supplementary material.(C-F) Expression and localization of syntabulin RNA in an unfertilized wild-type egg (5 minutes after activation, C), in a 1-cell stage wild-type embryo(20 mpf, D), an 8-cell stage wild-type embryo (E), and in a 1-cell stage tkk embryo (F). Lateral views; A, animal pole; Vg, vegetal pole. (G) Molecularnature of the syntabulintkk4 allele. PCR detection of the insertion in wild-type and tkk embryos (n=3 each). Scale bars: 200mm.

DEVELO

PMENT

928

treated with nocodazole or colchicine, which are inhibitors ofmicrotubule formation (Fig. 5M-Q), indicating that Syntabulintranslocation is microtubule dependent. After 60 mpf, Syntabulinwas not detected by whole-mount immunohistochemistry (Fig. 5K),nor in whole embryonic lysates by immunoprecipitation andimmunoblotting (Fig. 5L), although syntabulin mRNA persisted atthe vegetal pole through the 16-cell stage (90 mpf, Fig. 2E and seeFig. S4A in the supplementary material). These data indicate that theinability to detect Syntabulin after 60 mpf was most likely due to itsdegradation, rather than to protein diffusion or a decline insyntabulin mRNA abundance. Together, our findings reveal that thelocalization and stability of Syntabulin are tightly controlled afterfertilization until the protein is no longer detected around the 2-cellstage.

syntabulin RNA is transported to the vegetal polethrough the Buc-mediated Balbiani body-dependent pathwaySome vegetal RNAs, including germline-specific gene products, aretransported to the vegetal pole through the Balbiani body (Kloc etal., 2001; Kloc and Etkin, 1995; Kloc et al., 1996; Wilk et al., 2005),which requires maternal Buc function for its formation (Bontems etal., 2009; Marlow and Mullins, 2008). As syntabulin mRNAlocalizes to the vegetal pole of eggs and early embryos, weinvestigated when syntabulin achieves its vegetal pole localizationand whether this involves the Buc-mediated Balbiani body-dependent pathway in oocytes. During oogenesis, syntabulin mRNAwas detected in the Balbiani body of wild-type stage Ib oocytes(compare Fig. 6A with 6B,E), but was not localized in buc mutantoocytes (Fig. 6C,D,F). This was not due to reduced expression ordegradation of syntabulin mRNA, as similar levels were detected byqRT-PCR in wild-type and buc mutant stage I-III oocytes (Fig. 6G).In unfertilized buc mutant eggs, syntabulin mRNA showedcircumferential localization, indicating that the later localization of

syntabulin to the vegetal pole depends on its proper localization inthe Balbiani body of early oocytes (Fig. 6H-J). These data suggestthat vegetal pole localization of syntabulin mRNA is mediatedthrough the Buc-mediated Balbiani body-dependent pathway, whichis essential for AV polarity in oocytes.

DISCUSSIONBuc-mediated Balbiani body-dependent transportof syntabulin RNAsyntabulin mRNA localization to the vegetal pole of unfertilizedeggs and early stage embryos is established during oogenesis (Fig.2C-E and Fig. 6). In early stage wild-type oocytes, syntabulintranscripts localize to the Balbiani body (Fig. 6B,E). Balbiani bodylocalization of syntabulin mRNA was disrupted in buc mutantoocytes (Fig. 6I,J), which lack the Balbiani body and show aberrantlocalization of vegetal RNAs during establishment of oocyte AVpolarity (Bontems et al., 2009; Dosch et al., 2004; Marlow andMullins, 2008). These data indicate that, after syntabulin RNA istranscribed, it is transported to the vegetal pole through the Balbiani

RESEARCH ARTICLE Development 137 (6)

Fig. 3. Maternal expression of syntabulin rescues tkk mutants.(A) Structure of the rescue plasmid. Tol2, inverted repeats of Tol2transposon; zpc, promoter of zone pellucida protein C; 3� UTR, 3�untranslated region; pAS, polyadenylation signal. (B) Rescue bysyntabulin cDNA transgene. The numbers of transgenic (green) andnon-transgenic (orange) zebrafish embryos showing each phenotypeare tabulated, and the percentages are shown in the bar chart.(C,D) Expression of dha in tkk (D; 8%, n=25) and transgene-rescued(C, GFP+; 48%, n=33) embryos at the late blastula stage.(E,F) Localization of syntabulin at the 1-cell stage in a transgene-rescued (E; 67%, n=27) versus control (F; 100%, n=23) embryo. Scalebars: 200mm.

Fig. 4. Exon-intron structure of syntabulin is required for vegetallocalization and proper dorsal determination. (A-E) The 3� UTR isnot sufficient for vegetal localization of syntabulin. (A) Structure of thetransgene plasmid. (B-E) Localization of GFP mRNA in transgenic (B,D)or non-transgenic (C,E) tkk (D,E) or wild-type (B,C) zebrafish embryos atthe 1-cell stage (20 mpf). (F,G) Genomic structure of syntabulin (F) andthe rescue plasmid that contains the exon-intron structure ofsyntabulin. 2A-peptide, 2A peptide of PTV1. (H) Rescue by genomicDNA. The numbers (tabulated) and percentages (bar chart) oftransgenic (green) and non-transgenic (orange) embryos showing eachphenotype. (I,J) Expression of dha in tkk (J; 9%, n=32) and rescued (I;66%, n=33) embryos at the late blastula stage. (K,L) Localization ofsyntabulin at the 1-cell stage in a rescued (K; 86%, n=35) versus control(L; 100%, n=31) embryo. Shown are animal pole views (I,J) and lateralviews (A-E,K,L). Vegetal pole localization is marked by an asterisk (K).Scale bars: 200mm.

DEVELO

PMENT

body pathway during oogenesis (Fig. 7). The Balbiani body-dependent RNA transport system is shared by germline-specificRNAs (Kloc et al., 2001; Kloc and Etkin, 1995; Kloc et al., 1996;Kosaka et al., 2007; Wilk et al., 2005). Despite their commonlocalization in early oocytes, important differences exist betweensyntabulin and germline-specific mRNA localization in late stageoocytes and in embryos. First, after the initial localization ofgermline-specific mRNAs to the Balbiani body during oogenesis,some of these transcripts, such as dazl (Maegawa et al., 1999;Kosaka et al., 2007), remain at the vegetal pole, whereas others, such

as vasa (Knaut et al., 2000), redistribute around the circumferenceof the oocyte cortex or, such as nanos (Draper et al., 2007), are foundthroughout late stage oocytes. Despite their distinct positions in latestage oocytes, the germline-specific mRNAs are transported to theanimal pole during egg activation, and subsequently localize to thedistal ends of the blastomere cleavage furrows, where germ plasmis thought to be present during early cleavage stages (Hashimoto etal., 2004; Kosaka et al., 2007; Maegawa et al., 1999; Suzuki et al.,2000). By contrast, syntabulin mRNA remained at the vegetal poleuntil the 16-cell stage, and was not transported to the cleavagefurrows or to the primordial germ cells (Fig. 2E and see Fig. S4 inthe supplementary material). Second, whereas the 3� UTR issufficient for localization of germline-specific mRNAs to the vegetalpole (Kosaka et al., 2007), the 3� UTR of syntabulin was notsufficient for its vegetal localization (Figs 3 and 4). Therefore,although both syntabulin and the germline-specific mRNAs rely onthe Balbiani body for transport to the vegetal pole, they may utilizedistinct mechanisms for later aspects of their localization.

In Drosophila oocytes, posterior localization of oskar mRNArequires elements within its 3� UTR and proper splicing, and occursby a mechanism that is proposed to involve the exon-exon junctioncomplex (EJC), which includes the nuclear shuttling proteins

929RESEARCH ARTICLEDorsal termination by a motor protein linker

Fig. 5. Microtubule-dependent transportation of Syntabulin.(A) Interaction between Syntabulin and Kif5b. HEK293T cells weretransfected with expression vectors of Myc-tagged Syntabulin and HA-tagged Kif5b. IP, immunoprecipitation; IB, immunoblotting.(B-E) Detection of Syntabulin with anti-Syntabulin monoclonalantibodies (anti-Syn1, anti-Syn2) in wild-type and tkk zebrafish embryosat early zygote stage (10 mpf). Expression of Syntabulin at the animalpole (asterisks in D,E) is non-specific as it is detected in tkk embryos,which lack significant amounts of Syntabulin (see Fig. S8C in thesupplementary material). (F-K) Time course of Syntabulin localization inthe wild type. Immunostaining with anti-Syn1. (L) Immunoblotting ofanti-Syntabulin (anti-Syn2) immunoprecipitates. (M-O) Localization ofSyntabulin at 45 mpf in wild-type embryos treated with 1mg/mlnocodazole (N) or 0.4 mg/ml colchicine (O) just after fertilization, versusuntreated control (M). (P,Q) Quantification of Syntabulin localization at45 mpf showing lopsided or vegetal pole localization, or undetectableexpression in embryos treated with nocodazole (P) or colchicine (Q),versus control (P,Q). Shown are lateral views with animal pole to thetop. Scale bars: 200mm.

Fig. 6. Balbiani body-mediated localization of syntabulin at thevegetal pole. (A-D) Localization of syntabulin RNA in wild-type andbucky ball mutant (bucp106 and bucp43) stage Ib (STI/Ib) zebrafishoocytes. Staining by whole-mount in situ hybridization with sense (A,control) or antisense (B-D) riboprobes. In wild-type oocytes, syntabulinRNA localizes to the Balbiani body (A,B; 84%, n=104). In bucp106 (C;97%, n=87 oocytes from three females) and bucp43 (D; 100%, n=36from one female) mutant oocytes, syntabulin is not spatially restrictedor localized. (E,F) Localization of syntabulin RNA. In wild-type oocytes(E), syntabulin localizes to the Balbiani body, whereas syntabulin isdetected throughout the cytoplasm in bucp106 mutants (F; 100%, n=27from one female). (G) Quantification of syntabulin expression in wild-type and buc oocytes. Although syntabulin is not localized in bucmutants, the transcripts are present in early and late stage oocytes.(H-J) Localization of syntabulin RNA in wild-type and buc activatedeggs. In activated eggs from wild type, syntabulin shows polarizedlocalization (H; 100% n=37 from two females). In activated eggs frombucp106 (I; 100%, n=44 from three females) and bucp43 (J; 100%, n=8from one female) mutants, syntabulin is detected around thecircumference. Arrows indicate localization of syntabulin at the vegetalpole in wild type and its circumferential expansion in buc eggs. nuc,nucleus; Bb, Balbiani body. Scale bars: 50mm.

DEVELO

PMENT

930

Tsunagi (Y14) and Mago nashi (Hachet and Ephrussi, 2004).syntabulin mRNA produced from the transgene containing the exon-intron structure and 3�UTR was more enriched at the vegetal polethan RNA from the transgene containing the cDNA, first intron and3� UTR (Figs 3 and 4), suggesting that splicing of syntabulin RNA,and its Buc-mediated Balbiani body-dependent transport, alsoinvolve EJC components. syntabulin mRNA expressed from thetransgene containing the full genomic structure of syntabulin, exceptthe promoter and enhancer, was localized to the vegetal pole, butwas also detected within the blastoderm (Fig. 4K). This is possiblydue to the presence of the 2A peptide sequence and the Venus cDNAbetween the coding region and the 3�UTR, which might affect thestructure of nascent syntabulin mRNA and its transport.Alternatively, proper expression levels of syntabulin RNA might berequired for its vegetal localization; for example, if components thatlocalize or tether syntabulin to the vegetal pole are limiting. Futurestudies will reveal the precise molecular mechanisms that controlthe vegetal pole localization of syntabulin RNA.

The link between AV oocyte polarity and DV axisformationThe DDs initially localize to the vegetal pole of embryos. It isanticipated that AV oocyte polarity is linked to dorsal determination,although this possibility has not been formally tested. We show thatsyntabulin gene products, which are necessary for dorsaldetermination, localize to the vegetal pole in a Buc-mediatedBalbiani body-dependent manner. Our findings provide evidencethat links early AV polarity in oocytes to DV embryonic axisformation. Since Syntabulin functions as a cargo linker (Cai et al.,2005; Cai et al., 2007; Su et al., 2004), we hypothesize thatSyntabulin localizes the DDs to the vegetal pole during oogenesisand/or in early stage zygotes (Fig. 7).

When tkk embryos were rescued with the plasmid containing thecDNA, first intron and 3� UTR of syntabulin, the transgene-derivedRNA was not enriched at the vegetal pole (Fig. 3E), implying thatvegetal pole localization of syntabulin RNA is not strictly requiredfor dorsal determination. Alternatively, because syntabulin RNAwas present throughout the embryo, it is possible that the amount ofsyntabulin RNA at the vegetal pole was sufficient for dorsaldetermination. Moreover, the rescue data suggest that additionalfactors contribute to localizing the DDs to the vegetal pole.Importantly, RNA from the plasmid containing the syntabulin exon-intron structure was enriched at the vegetal pole and rescued the tkkphenotypes more efficiently (Fig. 4K). Therefore, vegetallocalization of syntabulin RNA might ensure that Syntabulin proteinis positioned to localize the DDs at the vegetal pole to facilitate theirrobust transport to the prospective dorsal side.

Recently, a Balbiani body-localized ubiquitin ligase, Trim36,was identified that regulates microtubule polymerization orstability in the vegetal cortex of Xenopus (Cuykendall andHouston, 2009). Depletion of Trim36 leads to ventralization inXenopus embryos (Cuykendall and Houston, 2009). Notably, likeother Balbiani-localized transcripts, trim36 associates with thegerm plasm in Xenopus embryos (Cuykendall and Houston,2009). Our data identify syntabulin as a novel Balbiani body-localized mRNA. We find that unlike other Balbiani body-localized mRNAs, syntabulin is not a component of germ plasmin embryos, but instead mediates localization and transport of theDDs.

Role of Syntabulin in dorsal determinationRemoval of the vegetal yolk or disruption of microtubule arrayformation during zygote or early cleavage stages perturbs thedevelopment of dorsal structures in zebrafish (Jesuthasan and Stahle,1997; Mizuno et al., 1999; Ober and Schulte-Merker, 1999). Thesestudies established the hypothesis that the DDs, initially localized tothe vegetal pole, are subsequently transported to the prospectivedorsal side along microtubules, where they are incorporated bydorsal blastomeres. We show that Syntabulin, an essential factor fordorsal determination in zebrafish, is initially localized to, and latertranslocates from, the vegetal pole (Fig. 5B-K). Syntabulin functionsas a kinesin I motor protein linker that transports cargo from thesoma to the presynaptic terminal of neuronal axons alongmicrotubules towards their plus ends (Cai et al., 2005; Cai et al.,2007; Su et al., 2004). We demonstrate that zebrafish Syntabulin caninteract with zebrafish Kif5b, a maternally expressed kinesin I heavychain (Fig. 5A and see Fig. S7 in the supplementary material).Translocation of Syntabulin from the vegetal pole was microtubuledependent (Fig. 5B-K). Together, our results strongly suggest thatSyntabulin links the DDs to kinesin I to mediate their initial vegetalpole localization and subsequent transport to the prospective dorsalside along the microtubule array that emanates from the vegetal pole(Fig. 7).

Our results revealed that although maternal syntabulin RNApersisted through the 16-cell stage, Syntabulin protein was absentfrom around the 2-cell stage (Fig. 5K,L). It is unlikely thattranslation of Syntabulin is suppressed in this limited period. Wefavor the hypothesis that Syntabulin protein is degraded around the2-cell stage. The rapid degradation of Syntabulin protein temporallycoincides with stages when dorsalizing activity vacates the vegetalyolk (Mizuno et al., 1999; Ober and Schulte-Merker, 1999),indicating a possible mechanism whereby Syntabulin first positionsthe DDs and subsequently releases them from the vegetalmicrotubules following its own degradation (Fig. 7).

RESEARCH ARTICLE Development 137 (6)

Fig. 7. The role of Syntabulin in dorsal determination inzebrafish. During oogenesis, syntabulin RNA is transported from thenucleus to the vegetal pole through the Buc-mediated Balbiani body-dependent pathway. During early embryogenesis, Syntabulin links thedorsal determinants (DDs) to Kif5B, the heavy chain of kinesin I. At 20mpf, a microtubule array forms and Syntabulin transports the DDs tothe plus end of microtubules. Around the 2-cell stage (60 mpf),Syntabulin is degraded and releases the DDs, which are eventuallyincorporated by dorsal blastomeres and activate zygotic geneexpression cascades required for the formation of the dorsal organizer.

DEVELO

PMENT

How are the DDs transported to the dorsal blastomeres aftertheir release from Syntabulin protein? Inhibition of microtubuleformation at 8- or 16-cell stages leads to defective organizerformation (Jesuthasan and Stahle, 1997). Vegetal microtubules areonly transiently detected in zygotes before the first cleavage, butare not detected at the 8- or 16-cell stage (Jesuthasan and Stahle,1997). At these later stages, long microtubule arrays emanatefrom the marginal blastomeres (Jesuthasan and Stahle, 1997;Solnica-Krezel and Driever, 1994). Whether these microtubulearrays contribute to DD transport to the dorsal blastomeresrequires further study. Our findings reveal two steps in DDtransport from the vegetal pole to the dorsal blastomeres. Initially,the DDs are localized within the vegetal yolk. The first phaseinvolves asymmetric translocation of the DDs via the vegetalmicrotubules, followed by their release from the vegetal portionof the yolk. Second, the DDs are transported to, and incorporatedby, the prospective dorsal blastomeres. Syntabulin is essential forthe first phase of DD transport (Fig. 7).

Dorsal determinantsIn zebrafish, transcripts of the nodal-related gene ndr1 localizeasymmetrically in a microtubule-dependent manner, which predictsthe dorsal side of the embryos, and antisense morpholino-mediatedknockdown of ndr1 affects the formation of dorsal structures (Goreet al., 2005). In contrast to syntabulin transcripts, which arelocalized to the vegetal pole of early oocytes and embryos, ndr1mRNA is uniformly distributed throughout oocytes and localizes tothe animal pole during egg activation before the first cleavage(Gore and Sampath, 2002). During this time in development,syntabulin mRNA and protein remain in the vegetal yolk (Fig. 2C,Eand Fig. 6). Later, during early cleavage stages, ndr1 mRNAlocalizes asymmetrically within the blastomeres of the 2- to 4-cellstage embryo (Gore et al., 2005). At this time, maternal Syntabulinprotein is no longer detected in the embryo (Fig. 5J,K). Moreover,asymmetric localization of ndr1 is b-catenin independent (Gore etal., 2005), and Nodal signaling is not required before the mid-blastula transition (Hagos and Dougan, 2007; Hagos et al., 2007),after b-catenin is detected in the nuclei of dorsal blastomeres (Fig.1) (Dougan et al., 2003). Although it is possible that Syntabulin isinvolved in ndr1 mRNA transport, the reports discussed above andour spatial-temporal analysis of Syntabulin localization andfunction suggest that the mechanisms that mediate the translocationof ndr1 and the Wnt pathway-activating DDs are distinct. Thenotion that distinct mechanisms contribute to translocation of theDDs and ndr1 is compatible to a model whereby Syntabulintransports the DDs to activate b-catenin, which then inducesessential zygotic regulators of axis formation, including ndr1(Bennett et al., 2007; Dougan et al., 2003; Shimizu et al., 2000b).

In Xenopus, maternal Wnt11 and Wnt5a function in dorsaldetermination (Cha et al., 2008; Tao et al., 2005). wnt11 mRNAlocalizes to the Balbiani body (Kloc and Etkin, 1995) and to thevegetal pole (Ku and Melton, 1993) of Xenopus oocytes. Wnt11protein is enriched dorsally in cleavage-stage Xenopus embryos(Schroeder et al., 1999). Depletion of maternal wnt11 from latestage Xenopus oocytes leads to dorsal axis defects (Tao et al.,2005). In zebrafish, expression of wnt11/wnt11r and wnt5a/bgenes is weak or below detection in 1-cell stage embryos, ascompared with gastrula and/or pharyngula stages (see Fig. S9 inthe supplementary material) when these Wnts function in cellmigration (Heisenberg et al., 2000; Matsui et al., 2005; Rauch etal., 1997). Maternal zygotic wnt11 (silberblick) mutant embryosdo not exhibit disrupted DV axis formation (Heisenberg et al.,

2000). Injection of wnt5a/b and/or wnt11/11r RNA just afterfertilization impairs gastrulation movements (data not shown), butneither affected dha expression in wild-type embryos nor rescueddha expression in tkk embryos (see Fig. S10 in the supplementarymaterial), suggesting negligible contributions of these Wnts toSyntabulin-mediated dorsal determination. Furthermore,overexpression of the Wnt signaling components Dishevelled 3(Dvl3), Glycogen synthase kinase (Gsk) binding protein (Gbp),dominant-negative Axin1 or dominant-negative Gsk3b rescuedand/or induced hyper-dorsalization in tkk embryos; by contrast,Wnt8, Wnt3a or Frizzled 8b did not efficiently dorsalize tkkembryos (Nojima et al., 2004). These data suggest that Syntabulinfunctions downstream of Wnt receptors and upstream of b-catenin, in agreement with the hypothesis that Syntabulin does nottransport Wnt RNAs or proteins. In Xenopus, Dishevelled (Dvl)and Gbp undergo microtubule- and/or motor protein-dependenttransport and contribute to dorsal axis determination (Miller et al.,1999; Weaver et al., 2003; Yost et al., 1998). In early zebrafishembryos, biased localization of Dvl2, Dvl3 and Gbp was notdetected (data not shown). Therefore, zebrafish and Xenopus mayutilize different DDs, or there are DDs that are shared amongvertebrate species that remain to be discovered. In either case, theanswer to this question requires the isolation and identification ofthe DDs. As a linker protein that attaches cargo, presumablyincluding the DDs, to kinesin I, Syntabulin might facilitateidentification of the elusive DDs.

In summary, the regulation of dorsal determination viamicrotubule dynamics and association with motor and adaptorproteins is shared among zebrafish and amphibians. Our findingsprovide a missing component of the microtubule-dependent DDtransport machinery. Moreover, our finding that syntabulin, anessential regulator of dorsal axis formation, relies on early AVoocyte polarity for its localization provides new insight into the linkbetween the first embryonic axis and dorsal determination inzebrafish.

AcknowledgementsWe thank K. Kawakami for the Tol2 transposon system; M. Tsang, M. Tada, D.Onichtchouk, T. Matsui and J. C. Belmonte for plasmids; M. Mullins forcomments on the manuscript; K. Bando, S. Fujii, A. Katsuyama, S. Onishi fortechnical assistance; and the members of the Hibi laboratory for helpfuldiscussions. This work was supported by Grants-in-Aid for Scientific Researchfrom the Ministry of Education, Science, Sports and Technology, Japan(21370103 to M.H.).

Competing interests statementThe authors declare no competing financial interests.

Supplementary materialSupplementary material for this article is available athttp://dev.biologists.org/lookup/suppl/doi:10.1242/dev.046425/-/DC1

ReferencesBellipanni, G., Varga, M., Maegawa, S., Imai, Y., Kelly, C., Myers, A. P., Chu,

F., Talbot, W. S. and Weinberg, E. S. (2006). Essential and opposing roles ofzebrafish beta-catenins in the formation of dorsal axial structures andneurectoderm. Development 133, 1299-1309.

Bennett, J. T., Stickney, H. L., Choi, W. Y., Ciruna, B., Talbot, W. S. and Schier,A. F. (2007). Maternal nodal and zebrafish embryogenesis. Nature 450, E1-E2;discussion E2-E4.

Bontems, F., Stein, A., Marlow, F., Lyautey, J., Gupta, T., Mullins, M. C. andDosch, R. (2009). Bucky ball organizes germ plasm assembly in zebrafish. Curr.Biol. 19, 414-422.

Cai, Q., Gerwin, C. and Sheng, Z. H. (2005). Syntabulin-mediated anterogradetransport of mitochondria along neuronal processes. J. Cell Biol. 170, 959-969.

Cai, Q., Pan, P. Y. and Sheng, Z. H. (2007). Syntabulin-kinesin-1 family member5B-mediated axonal transport contributes to activity-dependent presynapticassembly. J. Neurosci. 27, 7284-7296.

931RESEARCH ARTICLEDorsal termination by a motor protein linker

DEVELO

PMENT

932

Cha, S. W., Tadjuidje, E., Tao, Q., Wylie, C. and Heasman, J. (2008). Wnt5aand Wnt11 interact in a maternal Dkk1-regulated fashion to activate bothcanonical and non-canonical signaling in Xenopus axis formation. Development135, 3719-3729.

Cuykendall, T. N. and Houston, D. W. (2009). Vegetally localized Xenopustrim36 regulates cortical rotation and dorsal axis formation. Development 136,3057-3065.

De Robertis, E. M. (2006). Spemann’s organizer and self-regulation in amphibianembryos. Nat. Rev. Mol. Cell Biol. 7, 296-302.

De Robertis, E. M., Larrain, J., Oelgeschlager, M. and Wessely, O. (2000). Theestablishment of Spemann’s organizer and patterning of the vertebrate embryo.Nat. Rev. Genet. 1, 171-181.

Dosch, R., Wagner, D. S., Mintzer, K. A., Runke, G., Wiemelt, A. P. andMullins, M. C. (2004). Maternal control of vertebrate development before themidblastula transition: mutants from the zebrafish I. Dev. Cell 6, 771-780.

Dougan, S. T., Warga, R. M., Kane, D. A., Schier, A. F. and Talbot, W. S.(2003). The role of the zebrafish nodal-related genes squint and cyclops inpatterning of mesendoderm. Development 130, 1837-1851.

Draper, B. W., McCallum, C. M. and Moens, C. B. (2007). nanos1 is required tomaintain oocyte production in adult zebrafish. Dev. Biol. 305, 589-598.

Elinson, R. P. and Rowning, B. (1988). A transient array of parallel microtubulesin frog eggs: potential tracks for a cytoplasmic rotation that specifies the dorso-ventral axis. Dev. Biol. 128, 185-197.

Gerhart, J., Danilchik, M., Doniach, T., Roberts, S., Rowning, B. and Stewart,R. (1989). Cortical rotation of the Xenopus egg: consequences for theanteroposterior pattern of embryonic dorsal development. Development 107,37-51.

Gore, A. V. and Sampath, K. (2002). Localization of transcripts of the zebrafishmorphogen Squint is dependent on egg activation and the microtubulecytoskeleton. Mech. Dev. 112, 153-156.

Gore, A. V., Maegawa, S., Cheong, A., Gilligan, P. C., Weinberg, E. S. andSampath, K. (2005). The zebrafish dorsal axis is apparent at the four-cell stage.Nature 438, 1030-1035.

Guraya, S. S. (1979). Recent advances in the morphology, cytochemistry, andfunction of Balbiani’s vitelline body in animal oocytes. Int. Rev. Cytol. 59, 249-321.

Hachet, O. and Ephrussi, A. (2004). Splicing of oskar RNA in the nucleus iscoupled to its cytoplasmic localization. Nature 428, 959-963.

Hagos, E. G. and Dougan, S. T. (2007). Time-dependent patterning of themesoderm and endoderm by Nodal signals in zebrafish. BMC Dev. Biol. 7, 22.

Hagos, E. G., Fan, X. and Dougan, S. T. (2007). The role of maternal Activin-likesignals in zebrafish embryos. Dev. Biol. 309, 245-258.

Hashimoto, Y., Maegawa, S., Nagai, T., Yamaha, E., Suzuki, H., Yasuda, K.and Inoue, K. (2004). Localized maternal factors are required for zebrafishgerm cell formation. Dev. Biol. 268, 152-161.

Heisenberg, C. P., Tada, M., Rauch, G. J., Saude, L., Concha, M. L., Geisler, R.,Stemple, D. L., Smith, J. C. and Wilson, S. W. (2000). Silberblick/Wnt11mediates convergent extension movements during zebrafish gastrulation.Nature 405, 76-81.

Hibi, M., Hirano, T. and Dawid, I. B. (2002). Organizer formation and function.Results Probl. Cell Differ. 40, 48-4871.

Houliston, E. and Elinson, R. P. (1991). Evidence for the involvement ofmicrotubules, ER, and kinesin in the cortical rotation of fertilized frog eggs. J.Cell Biol. 114, 1017-1028.

Jesuthasan, S. and Stahle, U. (1997). Dynamic microtubules and specification ofthe zebrafish embryonic axis. Curr. Biol. 7, 31-42.

Kawakami, K., Takeda, H., Kawakami, N., Kobayashi, M., Matsuda, N. andMishina, M. (2004). A transposon-mediated gene trap approach identifiesdevelopmentally regulated genes in zebrafish. Dev. Cell 7, 133-144.

Kelly, C., Chin, A. J., Leatherman, J. L., Kozlowski, D. J. and Weinberg, E. S.(2000). Maternally controlled (beta)-catenin-mediated signaling is required fororganizer formation in the zebrafish. Development 127, 3899-3911.

Kloc, M. and Etkin, L. D. (1995). Two distinct pathways for the localization ofRNAs at the vegetal cortex in Xenopus oocytes. Development 121, 287-297.

Kloc, M., Larabell, C. and Etkin, L. D. (1996). Elaboration of the messengertransport organizer pathway for localization of RNA to the vegetal cortex ofXenopus oocytes. Dev. Biol. 180, 119-130.

Kloc, M., Bilinski, S., Chan, A. P., Allen, L. H., Zearfoss, N. R. and Etkin, L. D.(2001). RNA localization and germ cell determination in Xenopus. Int. Rev. Cytol.203, 63-91.

Kloc, M., Bilinski, S. and Etkin, L. D. (2004). The Balbiani body and germ celldeterminants: 150 years later. Curr. Top. Dev. Biol. 59, 1-36.

Knapik, E. W., Goodman, A., Atkinson, O. S., Roberts, C. T., Shiozawa, M.,Sim, C. U., Weksler-Zangen, S., Trolliet, M. R., Futrell, C., Innes, B. A. et al.(1996). A reference cross DNA panel for zebrafish (Danio rerio) anchored withsimple sequence length polymorphisms. Development 123, 451-460.

Knaut, H., Pelegri, F., Bohmann, K., Schwarz, H. and Nusslein-Volhard, C.(2000). Zebrafish vasa RNA but not its protein is a component of the germplasm and segregates asymmetrically before germline specification. J. Cell Biol.149, 875-888.

Kosaka, K., Kawakami, K., Sakamoto, H. and Inoue, K. (2007). Spatiotemporallocalization of germ plasm RNAs during zebrafish oogenesis. Mech. Dev. 124,279-289.

Ku, M. and Melton, D. A. (1993). Xwnt-11: a maternally expressed Xenopus wntgene. Development 119, 1161-1173.

Leung, T., Soll, I., Arnold, S. J., Kemler, R. and Driever, W. (2003). Directbinding of Lef1 to sites in the boz promoter may mediate pre-midblastula-transition activation of boz expression. Dev. Dyn. 228, 424-432.

Maegawa, S., Yasuda, K. and Inoue, K. (1999). Maternal mRNA localization ofzebrafish DAZ-like gene. Mech. Dev. 81, 223-226.

Marlow, F. L. and Mullins, M. C. (2008). Bucky ball functions in Balbiani bodyassembly and animal-vegetal polarity in the oocyte and follicle cell layer inzebrafish. Dev. Biol. 321, 40-50.

Matsui, T., Raya, A., Kawakami, Y., Callol-Massot, C., Capdevila, J.,Rodriguez-Esteban, C. and Izpisua Belmonte, J. C. (2005). NoncanonicalWnt signaling regulates midline convergence of organ primordia duringzebrafish development. Genes Dev. 19, 164-175.

Mei, W., Lee, K. W., Marlow, F. L., Miller, A. L. and Mullins, M. C. (2009).hnRNP I is required to generate the Ca2+ signal that causes egg activation inzebrafish. Development 136, 3007-3017.

Miller, J. R., Rowning, B. A., Larabell, C. A., Yang-Snyder, J. A., Bates, R. L.and Moon, R. T. (1999). Establishment of the dorsal-ventral axis in Xenopusembryos coincides with the dorsal enrichment of dishevelled that is dependenton cortical rotation. J. Cell Biol. 146, 427-437.

Mizuno, T., Yamaha, E., Kuroiwa, A. and Takeda, H. (1999). Removal ofvegetal yolk causes dorsal deficiencies and impairs dorsal-inducing ability of theyolk cell in zebrafish. Mech. Dev. 81, 51-63.

Nagai, T., Ibata, K., Park, E. S., Kubota, M., Mikoshiba, K. and Miyawaki, A.(2002). A variant of yellow fluorescent protein with fast and efficient maturationfor cell-biological applications. Nat. Biotechnol. 20, 87-90.

Nojima, H., Shimizu, T., Kim, C. H., Yabe, T., Bae, Y. K., Muraoka, O., Hirata,T., Chitnis, A., Hirano, T. and Hibi, M. (2004). Genetic evidence forinvolvement of maternally derived Wnt canonical signaling in dorsaldetermination in zebrafish. Mech. Dev. 121, 371-386.

Ober, E. A. and Schulte-Merker, S. (1999). Signals from the yolk cell inducemesoderm, neuroectoderm, the trunk organizer, and the notochord in zebrafish.Dev. Biol. 215, 167-181.

Onichtchouk, D., Aduroja, K., Belting, H. G., Gnugge, L. and Driever, W.(2003). Transgene driving GFP expression from the promoter of the zonapellucida gene zpc is expressed in oocytes and provides an early marker forgonad differentiation in zebrafish. Dev. Dyn. 228, 393-404.

Provost, E., Rhee, J. and Leach, S. D. (2007). Viral 2A peptides allow expressionof multiple proteins from a single ORF in transgenic zebrafish embryos. Genesis45, 625-629.

Rauch, G. J., Hammerschmidt, M., Blader, P., Schauerte, H. E., Strahle, U.,Ingham, P. W., McMahon, A. P. and Haffter, P. (1997). Wnt5 is required fortail formation in the zebrafish embryo. Cold Spring Harb. Symp. Quant. Biol. 62,227-234.

Ryu, S. L., Fujii, R., Yamanaka, Y., Shimizu, T., Yabe, T., Hirata, T., Hibi, M.and Hirano, T. (2001). Regulation of dharma/bozozok by the Wnt pathway.Dev. Biol. 231, 397-409.

Schier, A. F. and Talbot, W. S. (2005). Molecular genetics of axis formation inzebrafish. Annu. Rev. Genet. 39, 561-613.

Schneider, S., Steinbeisser, H., Warga, R. M. and Hausen, P. (1996). Beta-catenin translocation into nuclei demarcates the dorsalizing centers in frog andfish embryos. Mech. Dev. 57, 191-198.

Schroeder, K. E., Condic, M. L., Eisenberg, L. M. and Yost, H. J. (1999).Spatially regulated translation in embryos: asymmetric expression of maternalWnt-11 along the dorsal-ventral axis in Xenopus. Dev. Biol. 214, 288-297.

Schroeder, M. M. and Gard, D. L. (1992). Organization and regulation of corticalmicrotubules during the first cell cycle of Xenopus eggs. Development 114, 699-709.

Shimizu, T., Yamanaka, Y., Ryu, S., Hashimoto, H., Yabe, T., Hirata, T., Bae,Y., Hibi, M. and Hirano, T. (2000a). Cooperative roles of Bozozok/Dharma andnodal-related proteins in the formation of the dorsal organizer in zebrafish.Mech. Dev. 91, 293-303.

Shimizu, T., Yamanaka, Y., Ryu, S., Hashimoto, H., Yabe, T., Hirata, T., Bae,Y., Hibi, M. and Hirano, T. (2000b). Cooperative roles of Bozozok/Dharma andnodal-related proteins in the formation of the dorsal organizer in zebrafish.Mech. Dev. 91, 293-303.

Shimizu, T., Yabe, T., Muraoka, O., Yonemura, S., Aramaki, S., Hatta, K., Bae,Y. K., Nojima, H. and Hibi, M. (2005). E-cadherin is required for gastrulationcell movements in zebrafish. Mech. Dev. 122, 747-763.

Sirotkin, H. I., Dougan, S. T., Schier, A. F. and Talbot, W. S. (2000). bozozokand squint act in parallel to specify dorsal mesoderm and anteriorneuroectoderm in zebrafish. Development 127, 2583-2592.

Solnica-Krezel, L. and Driever, W. (1994). Microtubule arrays of the zebrafishyolk cell: organization and function during epiboly. Development 120, 2443-2455.

RESEARCH ARTICLE Development 137 (6)

DEVELO

PMENT

Su, Q., Cai, Q., Gerwin, C., Smith, C. L. and Sheng, Z. H. (2004). Syntabulin is amicrotubule-associated protein implicated in syntaxin transport in neurons. Nat.Cell Biol. 6, 941-953.

Suzuki, H., Maegawa, S., Nishibu, T., Sugiyama, T., Yasuda, K. and Inoue, K.(2000). Vegetal localization of the maternal mRNA encoding an EDEN-BP/Bruno-like protein in zebrafish. Mech. Dev. 93, 205-209.

Tao, Q., Yokota, C., Puck, H., Kofron, M., Birsoy, B., Yan, D., Asashima, M.,Wylie, C. C., Lin, X. and Heasman, J. (2005). Maternal wnt11 activates thecanonical wnt signaling pathway required for axis formation in Xenopusembryos. Cell 120, 857-871.

Thisse, C. and Thisse, B. (1998). High resolution whole-mount in situhybridization. In Zebrafish Science Monitor, vol. 5, issue 1. Eugene, OR:University of Oregon Press (http://zfin.org/zf_info/monitor/vol5.1/vol5.1.html).

Tsang, M., Maegawa, S., Kiang, A., Habas, R., Weinberg, E. and Dawid, I. B.(2004). A role for MKP3 in axial patterning of the zebrafish embryo.Development 131, 2769-2779.

Weaver, C., Farr, G. H., 3rd, Pan, W., Rowning, B. A., Wang, J., Mao, J., Wu,D., Li, L., Larabell, C. A. and Kimelman, D. (2003). GBP binds kinesin light

chain and translocates during cortical rotation in Xenopus eggs. Development130, 5425-5436.

Westerfield, M. (1995). The Zebrafish Book: a Guide for the Laboratory Use ofZebrafish. Eugene, OR: University of Oregon Press.

Wilk, K., Bilinski, S., Dougherty, M. T. and Kloc, M. (2005). Delivery of germinalgranules and localized RNAs via the messenger transport organizer pathway tothe vegetal cortex of Xenopus oocytes occurs through directional expansion ofthe mitochondrial cloud. Int. J. Dev. Biol. 49, 17-21.

Yamanaka, Y., Mizuno, T., Sasai, Y., Kishi, M., Takeda, H., Kim, C. H.,Hibi, M. and Hirano, T. (1998). A novel homeobox gene, dharma, caninduce the organizer in a non-cell-autonomous manner. Genes Dev. 12,2345-2353.

Yonemura, S., Matsui, T. and Tsukita, S. (2002). Rho-dependent and -independent activation mechanisms of ezrin/radixin/moesin proteins: anessential role for polyphosphoinositides in vivo. J. Cell Sci. 115, 2569-2580.

Yost, C., Farr, G. H., 3rd, Pierce, S. B., Ferkey, D. M., Chen, M. M. andKimelman, D. (1998). GBP, an inhibitor of GSK-3, is implicated in Xenopusdevelopment and oncogenesis. Cell 93, 1031-1041.

933RESEARCH ARTICLEDorsal termination by a motor protein linker

DEVELO

PMENT