www.elsevier.com/locate/ydbio

Developmental Biology

The zebrafish down syndrome cell adhesion molecule is involved in

cell movement during embryogenesis

Dean Yimlamai, Liza Konnikova, Larry G. Moss, Daniel G. Jay*

Cellular and Molecular Physiology, Tufts University School of Medicine, 136 Harrison Avenue, M and V 709, Boston, MA 02111, USA

Received for publication 8 June 2004, revised 30 November 2004, accepted 1 December 2004

Available online 20 December 2004

Abstract

The Down syndrome cell adhesion molecule (Dscam) is a protein overexpressed in the brains of Down syndrome patients and

implicated in mental retardation. Dscam is involved in axon guidance and branching in Drosophila, but cellular roles in vertebrates

have yet to be elucidated. To understand its role in vertebrate development, we cloned the zebrafish homolog of Dscam and showed

that it shares high amino acid identity and structure with the mammalian homologs. Zebrafish dscam is highly expressed in developing

neurons, similar to what has been described in Drosophila and mouse. When dscam expression is diminished by morpholino injection,

embryos display few neurons and their axons do not enter stereotyped pathways. Zebrafish dscam is also present at early embryonic

stages including blastulation and gastrulation. Its loss results in early morphogenetic defects. dscam knockdown results in impaired cell

movement during epiboly as well as in subsequent stages. We propose that migrating cells utilize dscam to remodel the developing

embryo.

D 2004 Elsevier Inc. All rights reserved.

Keywords: Danio rerio; Zebrafish; Down syndrome; Dscam; Cell movement; Epiboly; Gastrulation

Introduction

Down syndrome (DS) is the most common form of

genetic birth defect, occurring in approximately one in

every thousand live births. This syndrome is defined by

mental retardation and a set of phenotypic features that is

caused by a trisomy of Chromosome 21 (Lejeune et al.,

1959). Due to this aneuploidy, multiple genes are involved

and it is presumed that overexpression of one or more of

these genes is the cause of the phenotypes seen in DS.

Over three hundred genes have been identified on

Chromosome 21 through genomic sequencing, but it is a

significant challenge to narrow the list of candidate genes

that are involved in DS. One gene that has been linked to

0012-1606/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.ydbio.2004.12.001

Abbreviations: Dscam, Down syndrome cell adhesion molecule; DS,

Down syndrome; hpf, hours post fertilization; IgCAM, immunoglobulin

cell adhesion molecule; CE, convergent-extension; epi, epiboly.

* Corresponding author. Fax: +1 617 636 0445.

E-mail address: daniel.jay@tufts.edu (D.G. Jay).

DS is the Down syndrome cell adhesion molecule

(Dscam).

Dscam was identified through a positional cloning

strategy for common Down syndrome (DS) phenotypes

(Yamakawa et al., 1998). Dscam is highly expressed

during brain development and is expressed at 20% higher

levels in DS subjects (Saito et al., 2000). Functional

experiments to define the role that Dscam plays in

development have primarily been undertaken in Droso-

phila. Drosophila dscam is also highly expressed in

embryonic and adult CNS. Dscam mutants show disrupted

development of neuronal tracts and perturbed axon

guidance (Hummel et al., 2003; Schmucker et al., 2000;

Wang et al., 2002, 2004; Wojtowicz et al., 2004; Zhan

et al., 2004).

Of particular interest is the finding that Drosophila

Dscam exhibits alternative splicing at multiple sites during

development resulting in over 36,000 potential splice

variants (Schmucker et al., 2000). Expression of these

Dscam variants appears to be regulated in a stage and tissue

279 (2005) 44–57

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 45

specific fashion (Celotto and Graveley, 2001; Neves et al.,

2004) with individual cells expressing distinct subsets of

these Dscam isoforms (Neves et al., 2004). Distinct

combinations of Dscam isoforms may act as molecular

labels differentiating cells within a local environment.

Dscams interact with each other through homophilic

adhesion (Agarwala et al., 2000, 2001c; Wojtowicz et al.,

2004), a property that is particularly specific in Drosophila.

Drosophila Dscam isoforms that are identical throughout

their ectodomain, but differ at one exon where identity is

shared up to 90% do not bind in vitro (Wojtowicz et al.,

2004).

In humans and rodents, there are two Dscam paralogs,

Dscam and Dscam-L1 (Agarwala et al., 2001c; Barlow et

al., 2002), neither of which have been described to display

large numbers of alternative splice variants. The two

mammalian Dscams are expressed in complementary

patterns in the CNS (Barlow et al., 2002) and their products

can be found on axons and dendrites (Agarwala et al.,

2001b,c). In the growing embryo, cells in vivo may

segregate or make other preferred cellular contacts depend-

ing on the type of Dscam that is expressed. DS brains show

developmental defects including delayed neuronal lamina-

tion and disorganized arborization (Golden and Hyman,

1994). The nature of these defects suggests that Dscam has

a role in development of cortical malformations, but thus

far cellular mechanisms for vertebrate Dscam are not

known.

The zebrafish (Danio rerio) is an excellent model

system for understanding vertebrate development. The

optical clarity of zebrafish facilitates detailed microscopic

analysis of their embryos. Embryonic development pro-

ceeds rapidly after fertilization with the first cell division

happening within the first hour, and subsequent divisions

proceeding at an increasingly rapid pace. Within 24 h,

zebrafish possess a rudimentary nervous system that is

beginning to form functional connections. In addition,

advances in forward and reverse genetics have made the

zebrafish a powerful system for understanding the biology

of new molecules.

To better understand Dscam’s role in vertebrate devel-

opment, we cloned the zebrafish homolog. We show that

dscam is present in developing neurons and is expressed

throughout the embryo during early embryogenesis. Loss

of dscam through morpholino knockdown results in a

severe morphologic disruption of the embryo. Treated

embryos show fewer neurons and their axons do not enter

stereotyped pathways, but this phenotype is also associated

with growth defects. These results suggest that dscam

plays an unexpected role earlier in development. We show

that dscam is broadly expressed in the early embryo and

its loss results in impaired morphogenetic movements

during epiboly and gastrulation. At the end of gastrulation,

Dscam knockdown embryos are significantly shorter than

controls and exhibit disorganized mesoderm and prechor-

dal plate. Here we describe a new developmental role for

dscam: dscam facilitates cell migrations that shape the

growing embryo.

Materials and methods

Fish

Wild-type Tqbingen AB zebrafish embryos were col-

lected from our breeding colony and maintained according

to standard procedures. Embryos were maintained at 28.58Cand staged in hours post fertilization (hpf) and by

morphology (Westerfield, 1995).

Cloning and analysis of Dscam orthologs

3V RACE primers were designed against the aligned

human, mouse, and Drosophila Dscam sequences. cDNA

pools were generated from total embryonic RNA from

48 hpf zebrafish using the SMART kit (BD Clontech)

according to the manufacturer’s directions. A 3V 4.7 kb

RACE product was isolated with the following primer, 5V-GGGAACGTGGTCAGCTACCTGAA-3V and the 3VSMART kit primer. Nested 5VRACE to generate the full-

length gene was performed on the same pools with the

following primers used in succession, 5V-CCGCCGAGT-TATTGCATGTGCAG-3V and 5V-CCTTCTGCACGTT-

CAGCAGCTTCAG-3Vwith the SMART kit primers.

DNA and protein sequences were aligned and analyzed

using tools available at the Biologist’s Workbench (http://

workbench.sdsc.edu). ClustalW alignments were prepared

to assess the identity of the cloned sequence against

previously identified homologs. Domain analysis was

performed using HMMPFAM. Rooted phylogenetic trees

were generated using DRAWGRAM. Amino acid sequence

annotation was produced using The Eukaryotic Linear Motif

Resource (http://elm.eu.org).

RT-PCR analysis

RT-PCR was performed on total RNA extracted from

embryos at the desired time point using Trizol Reagent

(Invitrogen, Carlsbad, CA) according to the manufacturer’s

directions. RNA (1 Ag) was subsequently treated with

DNase I (Amplification Grade, Invitrogen) and 1st strand

cDNA synthesized with Superscript III (Invitrogen). A

200 bp dscam fragment was amplified with the following

primers, DSRT1L (5V-AGGAGCTGTTCACGAGCATT-3V)and DSRT1R (5V-TCATCTGCAGCT CGTACCAC-3V).Primers for the transcription factor MAX were used as

control (Schreiber-Agus et al., 1993) on the same samples.

Whole-mount in situ hybridization (WISH)

Embryos were grown to the desired age, fixed over-

night in 4% PFA in PBS, and manually dechorionated.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–5746

Embryos were stored in 100% methanol until use.

Digoxigenin-labeled sense and anti-sense riboprobes were

synthesized from a 1.2 kb dscam fragment cloned into the

EcoRI/HindIII sites of Bluescript SK+ with the following

primers, DC95L (5V-AGTCAGAATTCCGAGGAGAGCTQACAACGTCC) and DC95R (5V-ATATGAAGCTTCTQ

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 47

CTGCCGTTATGCTGTGAA-3V). Hybridization and detec-

tion on whole-mount zebrafish embryos were performed

as described by Broadbent and Read (1999). Wnt11

(cb748), Bmp2b (cb670), Six3b (cb347) were obtained

from the Zebrafish Information Resource Center. A

Dharma probe was generated by RT-PCR with the

following primers, DhaL (5V-AGCAGGCAAACAGCA-GAATC-3V) and DhaR (5V-TCCTCCAGAGCTCGAT-

TAGC-3V). The PCR product was cloned into pCRII for

use. Squint (Feldman et al., 1998), floating head (Talbot et

al., 1995), and axial/foxA2 (Strahle et al., 1993) were

used as previously described. Images were taken on a

Leica MZLIII dissecting microscope (Leica Microsystems,

Bannockburn, IL) equipped with a Hamamatsu CC4742-

95 camera (Hamamatsu Corp., Bridgewater, NJ) with a

Varispec RGB color slider (CRI Inc, Woburn, MA).

Images were recorded on a Macintosh with Openlab

software (Improvision, Lexington, MA).

Dscam antibody production and use

We cloned a 3.5 kb extracellular portion of Dscam into

pFastBAC (Invitrogen) with a 6xHis tag. This vector was

used to synthesize baculovirus that was subsequently used

to infect HiFive cells for antigen production. Secreted

Dscam was collected from the media and purified using Ni-

NTA Agarose (Qiagen, Valencia, CA) according to the

manufacturer’s directions. Mice were immunized with this

antigen using a standard protocol. Eventually one mouse

was sacrificed for hybridoma fusion and screened for

positives by ELISA and immunocytochemistry. The ascites

produced from this is DS#1 (produced by Tufts GRASP).

The monoclonal antibody used in this study is DS1-11.

Western blots were performed with DS#1 (1:1000) and

DS1-11 (1:100).

Whole-mount immunohistochemistry

Embryos were blocked for 1 h at RT in 20% FBS in PBS.

Antibody incubations were done overnight at 48C in

blocking buffer. Five 1 h washes were performed with

PBS-T before incubation with secondary antibody. Anti-

mouse Alexa 488 was purchased from Molecular Probes

(Invitrogen), anti-mouse HRP was purchased from Dako-

Cytomation (Carpinteria, CA) and diluted according to the

manufacturer’s directions. Peroxidase substrate kit (Vector

Laboratories, Burlingame, CA) was used to develop HRP

Fig. 1. Sequence analysis of Dscams. (A) Amino acid sequence of zebrafish dsc

sequence. Black boxed characters represent hydrophobic sequence. Gray boxe

domains. Intracellular signaling motifs are indicated by underlined sequence and

fibronectin domain; SP, signal peptide; TM, transmembrane domain; WW, WW

binding motif; PDZ, PDZ class I binding motif. (B) Amino acid sequence identit

immunoglobulin domains, shaded rectangles represent fibronectin domains, and

analysis of Dscam family members. Horizontal length is proportional to estimated t

m, mouse; r, rat; zf, zebrafish.

staining according to the manufacturer’s directions. The

following primary antibody concentrations were used: anti-

acetylated tubulin (Sigma, clone 6-11B-1, 1:1000), anti-

DS1-11 (1:100). Confocal images were taken with a Leica

SP2 (Leica Microsystems).

Morpholino injection

Morpholinos were purchased from Gene Tools, LLC

(Philomath, OR). Three morpholinos were designed, DS2M

(5V-CGCTCCTTTCAATCTCCAAACTAAG-3V), against

the 5V UTR of Dscam (�29 to �5), DS2Mmis (5V-CGATCATTTCAAGCTCCAATCTCAG-3V), a correspond-

ing five base mismatch oligo, and DS3M (5V-AAAAGAT-GATGGCCAATATCCACAT-3V), an oligo overlapping with

the ATG start site (+1 to +25). Morpholinos were

resuspended in 1� Danieau Buffer with 0.1% phenol red

and injected into the yolk of one to two cell stage embryos.

10 pl of morpholino was injected using an Eppendorf 5242

Microinjector. Concentrations from 1 to 40 ng/embryos

were tested and assessed for non-specific phenotypes

appearing between DS2M and DS2Mmis oligos.

Labeling cells of early embryos

One or two cell embryos were injected with either a

solution of 0.025% DMNB-FITC dextran (10 kDa, Molec-

ular Probes) or a similar solution also containing DS2M

morpholino in 1� Danieau Buffer. Embryos were allowed

to develop in the dark until the high cell stage. These

embryos were then mounted in 1.2% agarose viewing

chambers and uncaging was performed by exposing a

central area of the embryo to UV light through a DAPI filter

set for 10 s. Embryos were briefly imaged to record the spot

size using a FITC filter set on a Nikon Diaphot 200 and

returned to the incubator to continue development. Dispersal

of the spot was recorded 2 h later for each treated embryo.

Results

Cloning a zebrafish Dscam homolog

To address the role of Dscam in vertebrate development,

we cloned a zebrafish homolog by RT-PCR. We performed

3V and 5V RACE using degenerate primers recognizing

conserved sequences in the human, mouse, and Drosophila

am. Amino acid motifs predicted by ELM are indicated above the shaded

s represent immunoglobulin domains. White boxes represent fibronectin

are conserved in all vertebrate dscams. Ig, immunoglobulin domain; FN

class IV binding motif; T2, Traf2 binding motif; SH2/3, src homology 2/3

y between zebrafish dscam and human Dscam by domain. Ovals represen

the dark rectangle represents a transmembrane domain. (C) Evolutionary

ime from divergence of the gene from the related family member; h, human

,

t

;

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–5748

sequences using the SMART kit. A 6.6 kb cDNA encoding

a 2024 amino acid protein was isolated (Fig. 1A). Several

previously isolated partial ESTs confirm the sequence of this

gene (AL928362, AL919997, CA496783). Zebrafish dscam

contains ten immunoglobulin and six fibronectin domains in

the same order and sequence as other family members. The

amino acid sequence is highly conserved with the human

gene with 59–94% amino acid identity by domain (Fig. 1B).

The cytoplasmic domain of dscam is highly related to

other vertebrate family members. It is predicted to be 401

amino acids in length with several types of signaling motifs

including three WW Class IV binding domains, two Traf2

binding sites, an SH2 and SH3 binding site, and a PDZ

binding domain at the C terminus. Each of these motifs was

highly conserved from human through zebrafish and

suggests likely functionality (Fig. 1A). The more distantly

related gene, Dscam-L1, has conserved several of these

motifs including the WW binding domains, an SH2 binding

site, and the PDZ recognition motif. Common cytoplasmic

molecules probably bind to these sites, although, Dscam and

Dscam-L1 also contains several types of non-overlapping

signal sequences likely conferring specificity.

It has been reported that Drosophila dscam is a gene

product that displays significant diversity through alter-

native exon splicing (Schmucker et al., 2000), although

neither vertebrate Dscam paralogs have been described to

display a significant number of splice variants. It is of

particular note that Drosophila contains four Dscams, only

one of which displays an extensive number of splice

variants (Dscam1). We closely examined our dscam gene

product to determine if extensive alternative splicing could

exist. The cDNAwas constructed from 4.7 kb, 3Vand 2.1 kb,5V overlapping RACE products. We sequenced several of

these isolated clones in order to determine if there were any

significant splice variants. We did not identify any

alternative splicing in our clones, although PCR or stage-

specific expression may have biased the dscam isoform that

was cloned. Of the two Dscams isolated from other

vertebrate species, the cDNAwe cloned shows significantly

more homology to Dscam than to Dscam-L1 (Fig. 1C).

Zebrafish dscam is expressed in regions of developing

neurons

We examined dscam expression during embryogenesis

by whole-mount in situ hybridization (WISH) analysis using

a dscam probe in wild-type embryos every 2 h following

fertilization. Discrete dscam expression appeared at approx-

imately 22 hpf. We detected strong expression from brain

regions including the midbrain, telencephalon, and dien-

cephalon (Fig. 2E). By 24 hpf, dscam expression spreads

widely throughout the CNS into the hindbrain and spinal

cord (Fig. 2F). At this stage, neuronal precursors in these

regions of the CNS are differentiating and beginning the

process of extending axons and dendrites (Chitnis and

Kuwada, 1990; Wilson et al., 1990). These results agree

with previous work demonstrating that dscam is highly

expressed in regions of developing mouse neurons (Agar-

wala et al., 2001a; Barlow et al., 2002).

dscam expression is maintained in the brain through at

least 5 dpf (Figs. 2I and J), while in the spinal cord

expression is greatly diminished between 2 and 5 dpf (Figs.

2G and I). As the neuropil segregates into more discrete

structures, dscam is localized to several synaptic layers of

the eye (Fig. 2H) as well as brain (Fig. 2J). The dscam

expression suggests that it plays a role in axon tract

development and maintenance.

Dscam knockdown results in growth and axon defects

We investigated the in vivo role of dscam by micro-

injecting morpholino oligos to interfere with expression of

dscam (Nasevicius and Ekker, 2000). Oligos against the 5VUTR (DS2M) and overlapping the ATG start site (DS3M)

were targeted with this strategy. We initially examined

embryos at 28 hpf and found that they exhibited marked

morphological defects generally characterized by overall

shortening of the embryo and disorganization of many

tissues. These embryos were sorted by a phenotypic grading

system that we developed: G0-wild-type appearing; G1-

mild defects such as small brain and/or crooked tail; G2-

moderate defects including moderate anterior–posterior

(AP) axis shortening and/or extremely small eyes and brain;

and G3-severe defect including severe AP axis shortening

and/or complete lack of anterior structures (Figs. 4A–D).

Both morpholinos exhibited similar defects and proportions

(data not shown and Table 1, respectively) that were

indistinguishable, suggesting that they targeted the same

gene. The DS2M oligo was used in the rest of this study as

the 5VUTR is less likely to share sequence conservation

with other genes.

Due to reports in Drosophila that perturbation of dscam

results in disruption of axonal tract development, we

examined early axonal tracts of the brain. It has been

previously described that by 30 hpf, several axon tracts

connecting the two hemispheres as well as various portions

of the brain are well developed and observable with

neuronal markers (Chitnis and Kuwada, 1990; Wilson et

al., 1990). We used an antibody against anti-acetylated

tubulin to label neurons and their processes. DS2M embryos

with a wild-type morphology (G0) displayed normal axon

tracts (Fig. 5A). All of the major tracts were present

including the supraoptic tract (SOT), tract of the posterior

commissure (TPC), and the tract of the post-optic commis-

sure (TPOC). The anterior and post-optic commissures were

also well developed (data not shown). Axon guidance errors

were not observed and these embryos appear equivalent to

non-injected embryos (data not shown). Embryos displaying

mild growth defects (G1) had visibly smaller brains,

displayed few neurons and disorganized axons (Fig. 5B).

Tracts and commissures were not discernable from these

embryos, and the few axons projecting from neuronal

Fig. 2. Expression of zebrafish dscam during early embryogenesis. (A) RT-PCR of dscam at various developmental stages within the first 24 h. Max is used as

loading control. Whole mount in situ hybridization of wild-type embryos at (B) 1 cell, (C) high, (D) 8 somite, (E) 22 hpf, (F) 24 hpf, (G) 2 dpf, (H) high

magnification of 3 dpf eye, (I) 5 dpf, and (J) anterior–dorsal view at 5 dpf. B–D show expression throughout the early embryo. Arrowhead in D indicates the

presumptive eye field. (E) 22 hpf embryo shows initial dscam discrete expression within midbrain, diencephalon, and telencephalon. (F) 24 hpf embryo shows

expanded dscam expression and new expression in hindbrain and spinal cord. (G) 2 dpf embryo shows expanding brain and spinal cord expression. (H)

Multiple layers of the laminated eye show strong dscam expression, particularly IPL and INL. (I) 5 dpf embryo shows expression only in the brain. (J) High

magnification dorsal view of the brain shows dscam expression is expressed discretely throughout. PE, pigmented epithelium; PRL, photoreceptor layer; OPL,

outer plexiform layer; INL, inner nerve layer; IPL, inner plexiform layer; GCL, ganglion cell layer; Ep, epiphysis; DC, diencephalon; HB, hindbrain; MB,

midbrain; TC, telencephalon; OSE, olfactory sensory epithelium.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 49

regions did not appear to project in the direction of known

tracts or commissures. The axons that were present appeared

to take tortuous courses through the brain. Moderate to

severe phenotypes showed no neurons by anti-acetylated

tubulin staining (data not shown).

The spectrum of dscam phenotypes is likely due to

differential uptake of the morpholino resulting in various

levels of dscam expression. We developed polyclonal and

monoclonal antibodies against dscam and confirmed their

specificity by Western blotting and immunohistochemistry

on dscam-transfected Cos-7 cells (Fig. 3A and data not

shown). Western blots of microinjected zebrafish embryos

show variable amounts of knockdown as compared to

controls (Fig. 4A). Although the axonal tract defects

observed in Grade 1 embryos may reflect a role that dscam

plays in axon guidance, it is also possible that generalized

Table 1

Zebrafish dscam knockdown results in growth defects

dscam ng/embryo Phenotypic grade Total

0 1 2 3

DS2Mmis 10 68

(100%)

68

DS2Mmis 20 65

(58%)

48

(42%)

113

DS2M 10 112

(49%)

77

(34%)

28

(12%)

11

(5%)

228

DS2M 20 33

(22%)

26

(17%)

62

(42%)

28

(19%)

149

DS3M 10 61

(41%)

48

(32%)

32

(21%)

8

(5%)

149

DS3M 20 39

(26%)

43

(29%)

28

(19%)

40

(27%)

150

Equivalent amounts of DS2M, DS3M, or DS2Mmis morpholinos were

injected into 1 to 2 cell embryos as described in Materials and methods and

assessed for growth defects at 24 hpf (see Fig. 4).

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–5750

growth defects cause axons to take different paths to their

targets because the local environment is significantly

different from the wild-type scenario. Also, because Grade

2 and 3 embryos show no observable neurons and they

displayed severe growth defects, these results suggested that

dscam may play an earlier developmental role to pattern the

embryo.

Dscam is a gene expressed throughout early development

We performed RT-PCR for dscam from embryo extracts

at various stages of embryonic development to determine if

dscam is present at embryonic stages that were not initially

detected by WISH. We were able to detect dscam at all

Fig. 3. dscam protein is expressed early in development. (A) Immunoblot of an

expressing dscam. Arrowhead indicates dscam at 220 kDa. (B) Secondary antibody

antibody alone, 90% epiboly. (E) a-dscam monoclonal, 90% epiboly.

embryonic stages examined (Fig. 2A). dscam is a mater-

nally expressed transcript that diminishes until near the

mid-blastula transition when zygotic transcription begins.

dscam is present throughout several developmental stages

including blastulation, gastrulation, neurulation, somito-

genesis, and organogenesis. Its presence suggests that

dscam could participate in development at any of these

stages. We repeated dscam WISH at stages earlier than 22

hpf and found staining could be detected throughout the

embryo at 1 cell (Fig. 2B), high (Fig. 2C), and 8 somite

(Fig. 2D) stages.

We used dscam monoclonal antibodies to determine if

the dscam protein is localized to particular cell populations

in the early embryo. Immunostaining of early embryos

showed strong staining at the 32 cell stage as well as at 90%

epiboly (Figs. 3B–E). These results suggest that dscam

could be utilized at the earliest stages of cell division. dscam

expression is broadly expressed superficially at 90% epiboly

(Fig. 3E) with diffuse staining appearing throughout

sectioned embryos (data not shown). These data suggest it

may play a role in multiple cell types. Since dscam is

present during epiboly, we asked if its knockdown inhibits

developmental mechanisms at this time.

dscam knockdown results in impaired cell movement during

epiboly

During epiboly and gastrulation, developmental pro-

grams regulating cell fate and movement cooperate in

proper embryo formation (Heisenberg and Tada, 2002;

Keller et al., 2003; Schier, 2001; Wallingford et al., 2002). If

dscam plays a role at this time, embryos should display

ti-dscam polyclonal and monoclonal antibodies against Cos-7 cell lysates

alone, 32 cell stage. (C) a-dscam monoclonal, 32 cell stage. (D) Secondary

Fig. 4. Zebrafish dscam knockdown results in growth defects. (A) Grade 0 (G0), wild-type appearance. (B) Grade 1 (G1), mild growth defects. (C) Grade 2

(G2), moderate defects. (D) Grade 3 (G3), severe defects. (E) Western blot analysis of individual zebrafish injected with DS2Mmis (10 ng) or DS2M (10 ng).

Arrowhead indicates dscam expression at 220 kDa.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 51

observable differences. We examined control and dscam

morphants at the completion of gastrulation and noted

significant shortening of the anterior–posterior axis (Figs.

5C and D). Gastrulation begins at approximately 50%

epiboly, when the dorsal shield appears displaying the first

morphologic sign of anterior–posterior polarity. We fol-

lowed control and dscam morphants from 50% epiboly

through the tailbud stage to determine if there were

observable morphologic differences.

Control and dscam morphants demonstrated several

differences at 50% epiboly: dscam morphants do not form

a prominent dorsal shield, their blastoderm is significantly

thickened, and the blastoderm of dscam morphants appears

to be bilayered (Fig. 6). This phenotype continues to worsen

through gastrulation. At 60% epiboly, DS2M embryos have

an excess number of cells at the animal pole with respect to

controls. Also, appearance of the polster is not obvious until

much later (see arrow, Fig. 6). The polster although visible

in DS2M embryos at 85% epiboly is extremely thickened,

possibly due to a lack of cell migration from the animal pole

where the polster is visible. By the completion of

gastrulation, DS2M embryos are shorter, thinner, and have

an excess number of cells near the anterior margin of the

embryo.

Fig. 5. dscam morphants show axon tract defects. (A) G0 embryo shows normal axon tracts. (B) G1 embryo shows few neurons and disorganized axons. Ep,

epiphysis; DC, diencephalon; SOT, supraoptic tract; TC, telencephalon; TPOC, tract of the post-optic commissure; TPC, tract of the posterior commissure. (C)

Control and (D) dscam knockdown embryos at tailbud stage. Arrowhead represents tailbud. Arrow represents leading edge of the polster.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–5752

Previous work suggested that dscam is a molecule

involved in cell adhesion and/or guidance, but the morpho-

logic deficits observed could be due to dscam playing a role

in cellular differentiation. dscam knockdown at early stages

may lead to loss or expansion of tissues resulting in

morphologic disruption. We used marker genes in situ to

determine if dscam knockdown disrupts the presence or

organization of early cell populations.

Patterning of the zebrafish embryo is governed by

expression of maternally derived genes that convey posi-

tional information. Squint (sqt) is initially expressed along

the yolk syncytial layer and induces ventral cells into

endoderm and mesoderm (Feldman et al., 1998). This gene

Fig. 6. Time lapse images of control and dscam morphants during gastrulation. Ima

represents the tailbud. Arrow represents leading edge of the polster.

was present in control and knockdown embryos, but in

knockdown embryos it appeared more diffusely distributed

along the border of the yolk syncytial layer at 30% epiboly

(Fig. 7C). Cells expressing the signaling molecule Wnt11

(Heisenberg et al., 2000) demonstrated a similar diffuse

pattern of expression. These results appear to suggest that

cellular induction in Dscam morphants occurs over longer

distances or cells are unable to reorganize themselves

properly. We observed significant differences in the

expression of other signaling molecules as well. At 50%

epiboly, Bmp2b appeared to be significantly diminished in

the area of the shield (Fig. 7D), while the Bmp antagonist

Dharma (Fig. 7B) (Fekany et al., 1999) was absent in

ges are of a single embryo photographed at the indicated stages. Arrowhead

Fig. 7. Marker gene in situ of dscam knockdown embryos. (A and C) Animal pole view. (B and D) Lateral view, dorsal is right, animal pole is up. (E and F)

Dorsal view, tailbud is at the bottom. (G, H, and J) Lateral view, tailbud is at the bottom. (I) Rostral view. (A and C) Wnt11 and Squint/Nodal at 30% epiboly,

respectively. (B and D) Dharma and Bmp2b at 50% epiboly, respectively. (E and F) Axial and floating head at 90% epiboly, respectively. (E and F) Axial and

floating head at 1 somite, respectively. (I and J) Six3b at 8 somites. All controls show representative samples. DS2M injections show severe phenotype in at

least five embryos in samples of twenty.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 53

Dscam morphants implying a requirement for Dscam in

normal induction of these molecules.

As gastrulation begins (50% epiboly), cells of the

developing zebrafish embryo participate in a coordinated

set of movements termed convergent-extension (CE) (Hei-

senberg and Tada, 2002; Wallingford et al., 2002). We used

the pattern of in situ markers for particular cell populations

to assess the overall development of the embryo and if CE

movements are occurring normally. The mesoderm markers

axial/foxA2 (axl, Strahle et al., 1993) and floating head (flh,

Talbot et al., 1995) demonstrate that dscam knockdown

results in less mesoderm migration to the midline at 90%

epiboly, producing staining patterns that are wider than

controls (Figs. 7E and F, respectively). These embryos are

also significantly shorter than controls and the phenotype

progressively worsens through somitogenesis (Figs. 7G and

H). The effects of dscam knockdown are not limited to

mesoderm. A marker of the prechordal plate, six3b (Seo et

al., 1998), shows that dscam knockdown does not inhibit its

specification. However, similar to the effects seen in

mesoderm, neural plate formation is more disorganized

than controls and is much thinner (Figs. 7I and J).

In order to directly address if Dscam may be an important

component of cell movement during epiboly, we labeled

small groups of cells centrally at the high cell stage in

control and Dscam morphant embryos (Figs. 8A and B).

Fig. 8. Dscam is required for cell dispersal during epiboly. (A) Wild-type embryo at high cell stage showing bright spot of FITC-dextran labeled cells after

uncaging. (B) DS2M injected embryo at high cell stage showing a similar size spot of FITC-dextran labeled cells. (C) Awild-type embryo imaged after 2 h of

development illustrating how cells within the original spot widely disperse across the embryo. This embryo is an example of Grade IV dispersal. (D) A DS2M

embryo imaged 2 h after uncaging shows most cells contained within the original spot. Limited migration from the spot has occurred as the border of the

original spot is irregular. This is an example of Grade I dispersal. (E) Microinjection of DS2M (n = 36) shifts population of embryos exhibiting less dispersal of

cells within the uncaged spot toward Grade I as compared to wild-type embryos (n = 28). This shift is statistically significant by v2, P = 0.01. Grade I, uncaged

spot is condensed with slight cell migration. Grade II, spot is slightly dispersed. Grade III, spot is moderately dispersed. Grade IV, spot is widely dispersed.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–5754

This method labeled both superficial and deep cells of the

early embryo. Two hours later, we examined the placement

of these cells out of the initial spot to determine if significant

cell movement had occurred. As epiboly begins following

the high cell stage, cells begin to migrate to the periphery

through radial cell intercalation (Warga and Kimmel, 1990).

The result of this behavior is that cells in close proximity at

the beginning of epiboly become uniformly distributed

throughout the embryo. An example of normal cellular

behavior can be seen in Fig. 8C where cells of the original

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 55

spot have disseminated across the embryo. Labeled cells in

Dscam morphants after 2 h showed much less dispersal

than controls (Fig. 8D) with a significant shift in the

population of embryos displaying less migration (Fig. 8E,

v2, P = 0.01).

Our results suggest that most dscam knockdown defects

are the results of aberrant or delayed cell movement. Normal

organization of the embryo is disturbed as observed in time

lapse images and marker gene WISH. Cells in dscam

morphants exhibit less movement from their initial position

at the start of epiboly than controls. An additional defect we

observed is that the induction of some genes is influenced

by dscam expression. This may be primarily due to their

activation through dscam or a secondary defect due to

improper formation of an inductive tissue.

Discussion

Understanding the role that Dscam plays in development

of the vertebrate embryo may lead to valuable insights into

its associated disease, Down syndrome (DS). Previous

studies have primarily focused upon potential roles of

Dscam in the nervous system due to its strong expression in

the mouse and human CNS (Agarwala et al., 2001a;

Yamakawa et al., 1998). Examinations of brains from

normal and DS patients have shown that Dscam is

overexpressed and is associated with certain types of

CNS pathology (Saito et al., 2000). Drosophila mutants

have been useful in providing insight into cellular roles for

Dscam in neural development (Hummel et al., 2003;

Schmucker et al., 2000; Wang et al., 2002, 2004;

Wojtowicz et al., 2004; Zhan et al., 2004), but descriptions

for the molecular role of vertebrate Dscams have been

limited.

Here we report the sequence of the zebrafish homolog

that is highly related to the gene found on human

Chromosome 21. It shares high structural and sequence

homology to all other Dscams in their extracellular

domains. Two orthologs of Dscam exist in humans and

rodents and it is a distinct possibility that four dscam genes

may exist in the zebrafish due to genomic duplication in the

ray fin fish lineage. Few splice variants for vertebrates have

been reported in the literature in contrast to reports in

Drosophila of several thousand possible splice variants

(Celotto and Graveley, 2001; Schmucker et al., 2000). We

have not observed extensive splice variants for zebrafish

dscam.

The vertebrate Dscams differ greatly from their Droso-

phila counterparts with respect to their intracellular

domains. The intracellular portion of zebrafish dscam shares

83% amino acid identity to the human homolog. The

intracellular portions of Drosophila and vertebrate Dscams

do not share significant homology, but there are some

signaling motifs that exist in both groups and may be

functionally conserved between Drosophila and zebrafish.

SH2/SH3 binding motifs can be found in both species. In

Drosophila, the SH3 adaptor protein Dock binds to dscam

in a cooperative fashion through its SH2/SH3 binding

motifs (Schmucker et al., 2000). Vertebrate dscams may

bind nck (vertebrate homolog of Dock) through SH2/SH3

motifs, although the structure appears to have changed over

time. Vertebrate Dscams may have evolved new modifica-

tions to their signaling cascades as human Dscam directly

binds and activates one of nck’s downstream targets, pak1

(Li and Guan, 2004). All vertebrate Dscams contain a C-

terminal consensus PDZ binding motif (YTLV) that is

similar to the C-terminus of Drosophila dscam (TMAV).

Dscams likely utilize PDZ binding proteins to localize itself

to particular microdomains. It has been previously described

that Dscam can be found on the axons and dendrites of

Purkinje cells (Agarwala et al., 2001b). PDZ domains may

localize Dscam to synapses, allowing it to stabilize synaptic

connections in the CNS.

Zebrafish dscam is highly expressed in the developing

brain. Discrete expression begins in the forebrain and

midbrain, expanding in the following hours into the hind-

brain and spinal cord (Figs. 2E–J). These results are

strikingly similar to previous work identifying areas of

neurogenesis in the zebrafish brain (Chitnis and Kuwada,

1990; Wilson et al., 1990). The expression patterns are

similar to those in other systems (Agarwala et al., 2001a,c;

Barlow et al., 2002; Hummel et al., 2003; Schmucker et al.,

2000; Wang et al., 2002; Yamakawa et al., 1998),

suggesting dscam is used in shaping the nervous system.

Unfortunately, we were unable to produce in vitro tran-

scribed mRNA and were therefore unable to directly address

the role overexpressed Dscam has on embryonic growth as

a model for DS. Until this point, the primary interest in

Dscam has focused on its roles in neural and heart develop-

ment due to its strong and specific expression in these

organs. In this study, we reveal that Dscam is broadly

expressed during blastulation. Future work should consider

the impact excess Dscam expression has upon overall growth

and development beginning shortly after fertilization.

We have used morpholino oligos in this study to

knockdown dscam levels to discern its in vivo role. Mild

growth defects are associated with zebrafish embryos

displaying fewer neurons and misguided axons. Coupled

with strong dscam expression from neuronal regions

involved in axonogenesis, these data suggest zebrafish

dscam plays a role in nervous system development.

However, our experiments do not eliminate the possibility

that general growth defects affect axon guidance.

Our work demonstrates that dscam plays a role in early

morphogenesis of the embryo. We show that dscam is a

maternally expressed gene that is upregulated before and

throughout gastrulation. Zebrafish dscam is broadly

expressed through at least somitogenesis. dscam morphants

begin to display morphologic defects shortly after the start

of epiboly. By the beginning of gastrulation, many dscam

morphants appear thickened and bilayered as seen in Fig. 6.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–5756

Consistent with this phenotype, we show that dscam

morphants exhibit cell movement defects after the onset of

epiboly (Fig. 8), suggesting that cells are unable to move to

their proper spatial position. Since Dscams interact homo-

philically in vitro (Agarwala et al., 2000, 2001c) and appear

to mediate specific cell–cell contact in vivo (Wojtowicz et

al., 2004), Dscam may be utilized in a manner that provides

adhesion and traction against other cells. Therefore, a

bilayered Dscam morphant may be a sign of inappropriate

adhesion to neighboring cells. Since our data show that

many cells in early embryo express dscam, any cell could

move with respect to another cell by regulating adhesion to

its neighbors. A number of immunoglobulin cell adhesion

molecules such as Dscam have been demonstrated to

mediate migration in vivo through homophilic adhesion

(Ishida et al., 2003; Izumoto et al., 1996; Nakada et al.,

2000; Wright et al., 1999).

It appears that a primary consequence of dscam knock-

down is a lack of convergent-extension in mesodermal cells

and disorganization in the prechordal plate. This indicates

that dscam is involved with different types of cell move-

ment through various developmental stages. Marker gene in

situ of knockdown embryos demonstrates that dscam does

not appear to interfere with cell fate, since knockdown

embryos have the correct cell layers in the approximately

correct position. In contrast, expression of Bmp2b and

Dharma in the shield was downregulated in Dscam

morphants. Dharma antagonizes Bmp2b expression (Leung

et al., 2003) and would be expected to upregulate Bmp2b

expression in this study as Dharma appeared to be absent.

This may indicate that morphologic deficits caused by

Dscam knockdown have subsequent consequences upon

normal gene expression.

Here, we show Dscam knockdown is required in zebra-

fish during epiboly when cells begin migrating to envelope

the yolk and take their eventual positions in the vertebrate

embryo. This situation is likely to be analogous to the fetal

development of rare cases of monosomy Chromosome 21.

Case reports of full, molecularly confirmed monosomy 21

describe severe intrauterine growth retardation, microce-

phaly, brain abnormalities, and congenital heart defects as

common traits (Mori et al., 2004; Pellissier et al., 1987;

Wisniewski et al., 1983). Specifically, partial deletion of

21q22.2, the chromosomal region containing Dscam, is

associated with microcephaly and intrauterine growth

retardation (Matsumoto et al., 1997). The work here

supports the notion that Dscam may be an important

component of embryonic and cephalic growth.

Recently, it has been suggested that Dscam is involved

in Drosophila dorsal–ventral patterning (Stathopoulos et

al., 2002). Perturbing the maternal Dorsal gradient results in

significant changes in dscam expression raising the

possibility that it functions in early Drosophila develop-

ment. In wild-type Drosophila, it is primarily expressed in

neuroectoderm although its expression can be found else-

where to a lesser degree. Our results show that dscam plays

a role during early embryonic vertebrate development. In

Grade 3/4 zebrafish dscam knockdown embryos, the

neuroectodermal-derived tissues are severely affected (Figs.

4C and D), and although mesoderm derived tissues are

present, they often appear disorganized. The use of Dscams

in cell movement during early development thus appears to

be an evolutionarily conserved mechanism for embryonic

morphogenesis.

Acknowledgments

The authors would like to thank members of the Jay

laboratory and Dr. Jennifer Moss (Tufts University) for

helpful discussions; Punita Koustabhan for excellent fish

care and technical assistance; the Tufts GRASP center for

antibody production and cell culture, NIH program grant

(DK34928); the Zebrafish International Resource Center for

providing cDNA clones, NIH Grant (RR15402); and the

useful comments of the anonymous reviewers of the

manuscript. This work was supported by NIH grants

(EY11992, NS34699) to D.G.J.

References

Agarwala, K.L., Nakamura, S., Tsutsumi, Y., Yamakawa, K., 2000. Down

syndrome cell adhesion molecule DSCAM mediates homophilic

intercellular adhesion. Brain Res., Mol. Brain Res. 79, 118–126.

Agarwala, K.L., Ganesh, S., Amano, K., Suzuki, T., Yamakawa, K.,

2001a. DSCAM, a highly conserved gene in mammals, expressed in

differentiating mouse brain. Biochem. Biophys. Res. Commun. 281,

697–705.

Agarwala, K.L., Ganesh, S., Suzuki, T., Akagi, T., Kaneko, K., Amano, K.,

Tsutsumi, Y., Yamaguchi, K., Hashikawa, T., Yamakawa, K., 2001b.

Dscam is associated with axonal and dendritic features of neuronal

cells. J. Neurosci. Res. 66, 337–346.

Agarwala, K.L., Ganesh, S., Tsutsumi, Y., Suzuki, T., Amano, K.,

Yamakawa, K., 2001c. Cloning and functional characterization of

DSCAML1, a novel DSCAM-like cell adhesion molecule that

mediates homophilic intercellular adhesion. Biochem. Biophys. Res.

Commun. 285, 760–772.

Barlow, G.M., Micales, B., Chen, X.N., Lyons, G.E., Korenberg, J.R.,

2002. Mammalian DSCAMs: roles in the development of the spinal

cord, cortex, and cerebellum? Biochem. Biophys. Res. Commun. 293,

881–891.

Broadbent, J., Read, E.M., 1999. Wholemount in situ hybridization of

Xenopus and zebrafish embryos. Methods Mol. Biol. 127, 57–67.

Celotto, A.M., Graveley, B.R., 2001. Alternative splicing of the Drosophila

dscam pre-mrna is both temporally and spatially regulated. Genetics

159, 599–608.

Celotto, A.M., Graveley, B.R., 2001. Alternative splicing of the Drosophila

dscam pre-mrna is both temporally and spatially regulated. Genetics

159, 599–608.

Chitnis, A.B., Kuwada, J.Y., 1990. Axonogenesis in the brain of zebrafish

embryos. J. Neurosci. 10, 1892–1905.

Fekany, K., Yamanaka, Y., Leung, T., Sirotkin, H.I., Topczewski, J., Gates,

M.A., Hibi, M., Renucci, A., Stemple, D., Radbill, A., Schier, A.F.,

Driever, W., Hirano, T., Talbot, W.S., Solnica-Krezel, L., 1999. The

zebrafish bozozok locus encodes Dharma, a homeodomain protein

essential for induction of gastrula organizer and dorsoanterior embry-

onic structures. Development 126, 1427–1438.

D. Yimlamai et al. / Developmental Biology 279 (2005) 44–57 57

Feldman, B., Gates, M.A., Egan, E.S., Dougan, S.T., Rennebeck, G.,

Sirotkin, H.I., Schier, A.F., Talbot, W.S., 1998. Zebrafish organizer

development and germ-layer formation require nodal-related signals.

Nature 395, 181–185.

Golden, J.A., Hyman, B.T., 1994. Development of the superior temporal

neocortex is anomalous in trisomy 21. J. Neuropathol. Exp. Neurol. 53,

513–520.

Heisenberg, C.P., Tada, M., 2002. Zebrafish gastrulation movements:

bridging cell and developmental biology. Semin. Cell Dev. Biol. 13,

471–479.

Heisenberg, C.P., Tada, M., Rauch, G.J., Saude, L., Concha, M.L., Geisler,

R., Stemple, D.L., Smith, J.C., Wilson, S.W., 2000. Silberblick/Wnt11

mediates convergent extension movements during zebrafish gastrula-

tion. Nature 405, 76–81.

Hummel, T., Vasconcelos, M.L., Clemens, J.C., Fishilevich, Y., Vosshall,

L.B., Zipursky, S.L., 2003. Axonal targeting of olfactory receptor

neurons in Drosophila is controlled by Dscam. Neuron 37, 221–231.

Ishida, T., Kundu, R.K., Yang, E., Hirata, K., Ho, Y.D., Quertermous, T.,

2003. Targeted disruption of endothelial cell-selective adhesion

molecule inhibits angiogenic processes in vitro and in vivo. J. Biol.

Chem. 278, 34598–34604.

Izumoto, S., Ohnishi, T., Arita, N., Hiraga, S., Taki, T., Hayakawa, T., 1996.

Gene expression of neural cell adhesion molecule L1 in malignant

gliomas and biological significance of L1 in glioma invasion. Cancer

Res. 56, 1440–1444.

Keller, R., Davidson, L.A., Shook, D.R., 2003. How we are shaped: the

biomechanics of gastrulation. Differentiation 71, 171–205.

Lejeune, J., Gautier, M., Turpin, R., 1959. Etude des chromosomes

somatiques de neuf enfants mongoliens. C. R. Acad. Sci. 248,

1721–1722.

Leung, T., Bischof, J., Soll, I., Niessing, D., Zhang, D., Ma, J., Jackle, H.,

Driever, W., 2003. Bozozok directly represses bmp2b transcription and

mediates the earliest dorsoventral asymmetry of bmp2b expression in

zebrafish. Development 130, 3639–3649.

Li, W., Guan, K.L., 2004. The down syndrome cell adhesion molecule

DSCAM interacts with and activates PAK. J. Biol. Chem. 279,

32824–32831.

Matsumoto, N., Ohashi, H., Tsukahara, M., Kim, K.C., Soeda, E., Niikawa,

N., 1997. Possible narrowed assignment of the loci of monosomy 21-

associated microcephaly and intrauterine growth retardation to a 1.2-Mb

segment at 21q22.2. Am. J. Hum. Genet. 60, 997–999.

Mori, M.A., Lapunzina, P., Delicado, A., Nunez, G., Rodriguez, J.I., de

Torres, M.L., Herrero, F., Valverde, E., Lopez-Pajares, I., 2004. A

prenatally diagnosed patient with full monosomy 21: ultrasound,

cytogenetic, clinical, molecular, and necropsy findings. Am. J. Med.

Genet. 127A, 69–73.

Nakada, M.T., Amin, K., Christofidou-Solomidou, M., O’Brien, C.D., Sun,

J., Gurubhagavatula, I., Heavner, G.A., Taylor, A.H., Paddock, C., Sun,

Q.H., Zehnder, J.L., Newman, P.J., Albelda, S.M., DeLisser, H.M.,

2000. Antibodies against the first Ig-like domain of human platelet

endothelial cell adhesion molecule-1 (PECAM-1) that inhibit PECAM-

1-dependent homophilic adhesion block in vivo neutrophil recruitment.

J. Immunol. 164, 452–462.

Nasevicius, A., Ekker, S.C., 2000. Effective targeted gene dknockdownT inzebrafish. Nat. Genet. 26, 216–220.

Neves, G., Zucker, J., Daly, M., Chess, A., 2004. Stochastic yet biased

expression of multiple Dscam splice variants by individual cells. Nat.

Genet. 36, 240–246.

Pellissier, M.C., Philip, N., Voelckel-Baeteman, M.A., Mattei, M.G.,

Mattei, J.F., 1987. Monosomy 21: a new case confirmed by in situ

hybridization. Hum. Genet. 75, 95–96.

Saito, Y., Oka, A., Mizuguchi, M., Motonaga, K., Mori, Y., Becker, L.E.,

Arima, K., Miyauchi, J., Takashima, S., 2000. The developmental and

aging changes of Down’s syndrome cell adhesion molecule expression

in normal and Down’s syndrome brains. Acta Neuropathol. (Berlin)

100, 654–664.

Schier, A.F., 2001. Axis formation and patterning in zebrafish. Curr. Opin.

Genet. Dev. 11, 393–404.

Schmucker, D., Clemens, J.C., Shu, H., Worby, C.A., Xiao, J., Muda, M.,

Dixon, J.E., Zipursky, S.L., 2000. Drosophila Dscam is an axon

guidance receptor exhibiting extraordinary molecular diversity. Cell

101, 671–684.

Schreiber-Agus, N., Horner, J., Torres, R., Chiu, F.C., DePinho, R.A., 1993.

Zebra fish myc family and max genes: differential expression and

oncogenic activity throughout vertebrate evolution. Mol. Cell. Biol. 13,

2765–2775.

Seo, H.C., Drivenes, O., Ellingsen, S., Fjose, A., 1998. Expression of two

zebrafish homologues of the murine Six3 gene demarcates the initial

eye primordia. Mech. Dev. 73, 45–57.

Stathopoulos, A., Van Drenth, M., Erives, A., Markstein, M., Levine, M.,

2002. Whole-genome analysis of dorsal–ventral patterning in the

Drosophila embryo. Cell 111, 687–701.

Strahle, U., Blader, P., Henrique, D., Ingham, P.W., 1993. Axial, a

zebrafish gene expressed along the developing body axis, shows

altered expression in cyclops mutant embryos. Genes Dev. 7,

1436–1446.

Talbot, W.S., Trevarrow, B., Halpern, M.E., Melby, A.E., Farr, G.,

Postlethwait, J.H., Jowett, T., Kimmel, C.B., Kimelman, D., 1995. A

homeobox gene essential for zebrafish notochord development. Nature

378, 150–157.

Wallingford, J.B., Fraser, S.E., Harland, R.M., 2002. Convergent extension:

the molecular control of polarized cell movement during embryonic

development. Dev. Cell 2, 695–706.

Wang, J., Zugates, C.T., Liang, I.H., Lee, C.H., Lee, T., 2002. Dro-

sophila Dscam is required for divergent segregation of sister

branches and suppresses ectopic bifurcation of axons. Neuron 33,

559–571.

Wang, J., Ma, X., Yang, J.S., Zheng, X., Zugates, C.T., Lee, C.H., Lee, T.,

2004. Transmembrane/juxtamembrane domain-dependent Dscam dis-

tribution and function during mushroom body neuronal morphogenesis.

Neuron 43, 663–672.

Warga, R.M., Kimmel, C.B., 1990. Cell movements during epiboly and

gastrulation in zebrafish. Development 108, 569–580.

Westerfield, M.E., 1995. The zebrafish book. University of Oregon Press,

Eugene, OR.

Wilson, S.W., Ross, L.S., Parrett, T., Easter Jr., S.S., 1990. The develop-

ment of a simple scaffold of axon tracts in the brain of the embryonic

zebrafish, Brachydanio rerio. Development 108, 121–145.

Wisniewski, K., Dambska, M., Jenkins, E.C., Sklower, S., Brown, W.T.,

1983. Monosomy 21 syndrome: further delineation including clinical,

neuropathological, cytogenetic and biochemical studies. Clin. Genet.

23, 102–110.

Wojtowicz, W.M., Flanagan, J.J., Millard, S.S., Zipursky, S.L., Clemens,

J.C., 2004. Alternative splicing of Drosophila Dscam generates axon

guidance receptors that exhibit isoform-specific homophilic binding.

Cell 118, 619–633.

Wright, J.W., Snyder, M.A., Schwinof, K.M., Combes, S., Copenhaver,

P.F., 1999. A role for fasciclin II in the guidance of neuronal migration.

Development 126, 3217–3228.

Yamakawa, K., Huo, Y.K., Haendelt, M.A., Hubert, R., Chen, X.N., Lyons,

G.E., Korenberg, J.R., 1998. DSCAM: a novel member of the

immunoglobulin superfamily maps in a Down syndrome region and

is involved in the development of the nervous system. Hum. Mol.

Genet. 7, 227–237.

Zhan, X.L., Clemens, J.C., Neves, G., Hattori, D., Flanagan, J.J., Hummel,

T., Vasconcelos, M.L., Chess, A., Zipursky, S.L., 2004. Analysis of

Dscam diversity in regulating axon guidance in Drosophila mushroom

bodies. Neuron 43, 673–686.