EVOLUTION & DEVELOPMENT

2:2, 85–92 (2000)

©

BLACKWELL SCIENCE, INC.

85

Characterization of amphioxus

AmphiWnt8

: insights into the

evolution of patterning of the embryonic dorsoventral axis

Michael Schubert,

a,

* Linda Z. Holland,

a

Georgia D. Panopoulou,

b

Hans Lehrach,

b

and Nicholas D. Holland

a

a

Scripps Institution of Oceanography, University of California at San Diego, La Jolla, CA 92093-0202, USA;

b

Max-Planck-Institut für Molekulare Genetik, D-14195 Berlin (Dahlem), Germany

*Author for correspondence (email:mschubert@ucsd.edu)

SUMMARY

The full-length sequence and developmental

expression of an amphioxus

Wnt

gene (

AmphiWnt8

) are de-scribed. In amphioxus embryos, the expression patterns of

AmphiWnt8

suggest patterning roles in the forebrain, in thehindgut, and in the paraxial mesoderm that gives rise to themuscular somites. Phylogenetic analysis indicates that a sin-gle

Wnt8

subfamily gene in an ancestral chordate duplicatedearly in vertebrate evolution into a

Wnt8

clade and a

Wnt8b

clade. Coincident with this gene duplication, the functions ofthe ancestral

AmphiWnt8

-like gene appear to have been di-vided between vertebrate

Wnt8b

(exclusively neurogenic, es-pecially in the forebrain) and vertebrate

Wnt8

(miscellaneous,especially in early somitogenesis). Amphioxus

AmphiWnt8

and its vertebrate

Wnt8

homologs probably play comparableroles in the early dorsoventral patterning of the embryonicbody axis.

INTRODUCTION

For dorsoventral axis development in bilaterian embryos, itis convenient to recognize an initial establishment phase anda subsequent patterning phase. The key molecular mecha-nisms initially establishing the dorsoventral axis appear todiffer between protostomes and deuterostomes. Thus, in theprotostome

Drosophila

, a localized protease cascade gener-ates Toll ligand on the ventral side of the embryo (Le Mosyet al. 1998), whereas, in the deuterostome

Xenopus

, corticalrotation transports dishevelled protein dorsally where it trig-gers the downstream portions of a

Wnt

signalling pathwayfollowed by a patterning phase (Miller et al. 1999).

In comparison with the initial establishment phase, thedorsoventral patterning phase is more comparable betweenprotostomes and deuterostomes. In both groups, morpho-gens from the neural side and antineural side interact to sub-divide the dorsoventral axis into several tissue types (Nellenet al. 1996; Piccolo et al. 1996). Of the morphogens that helppattern the dorsoventral embryonic axis, the best studied arevertebrate BMP4 and

Drosophila

DPP (encoded by homolo-gous genes on the antineural side), which antagonize, respec-tively, chordin and short gastrulation (encoded by homolo-gous genes on the neural side).

It has been proposed that the molecular basis of dorsoven-tral axis patterning is phylogenetically ancient and was al-ready present in the last common ancestor of

Drosophila

andvertebrates (De Robertis and Sasai 1996), because BMP4and DPP are functionally interchangeable, as are chordin and

short gastrulation (Padgett et al. 1993; Holley et al. 1995,1996). While polarities of the morphogenetic systems pat-terning the dorsoventral axes correspond in

Drosophila

andvertebrates, the body plan topographies are reversed; thus,the neural side of the former is ventral, but the neural side ofthe latter is dorsal. It has thus been suggested (Arendt andNübler-Jung 1994; De Robertis and Sasai 1996) that the to-pology of animal bodies became dorsoventrally inverted withrespect to the substratum during deuterostome evolution.

In comparison with the much discussed inversion of thebody plan, there is another aspect of dorsoventral axis pat-terning that has received less attention. This concerns thequestion of which embryonic tissue is the earliest to be sub-divided dorsoventrally during development. In

Drosophila

,the first dorsoventral patterning appears in the ectoderm,which considerably later imposes a dorsoventral pattern onthe mesoderm (St. Johnston and Gelbart 1987; Ferguson andAnderson 1992; De Robertis and Sasai 1996). In contrast,the dorsoventral axis of vertebrate embryos is patterned ap-proximately concurrently in the ectoderm (Graff 1997; Hem-mati-Brivanlou and Melton 1997) and mesoderm (Christianand Moon 1993; Fainsod et al. 1994; De Robertis and Sasai1996; Graff 1997; Tonegawa et al. 1997; Hoppler and Moon1998; Marom et al. 1999). Several evolutionary scenarioscould explain this difference between protostomes and deu-terostomes in the initial patterning of the dorsoventral axis. Thesimplest scenario is that the site of this patterning was primi-tively ectodermal, but became concurrently ectodermal andmesodermal at some point in the deuterostome line of descent.

86 EVOLUTION & DEVELOPMENT

Vol. 2, No. 2, March–April 2000

To gain insights into the evolutionary origin of vertebrate-type dorsoventral patterning, we have been studying am-phioxus homologs of vertebrate genes known to influencedorsoventral patterning. Amphioxus is the closest living in-vertebrate relative of the vertebrates (Wada and Satoh 1994)and is thus the best available stand in for the proximate inver-tebrate ancestor of the vertebrates. The present paper con-cerns amphioxus

AmphiWnt8

, a close relative of vertebrate

Wnt8

subfamily genes known to play a key part in patterningthe early mesoderm dorsoventrally (Christian et al. 1991;Fainsod et al. 1994; Hoppler et al. 1996; Hoppler and Moon1998; Marom et al. 1999; Tian et al., 1999). The

Wnt8

genescomprise one of the more than a dozen subfamilies by nowrecognized for the

Wnt

gene family as a whole. Although

Wnt

genes are widely distributed in the animal kingdom, the full-length sequences described to date for this gene family arepredominantly from vertebrates. For amphioxus

AmphiWnt8

is the first full-length

Wnt

sequence determined, although par-tial sequences of

AmphiWnt4

and

AmphiWnt6

have previouslybeen described by Holland et al. (1994).

Genes of the

Wnt

family encode structurally related, cys-tein-rich glycoproteins involved in cell–cell signalling for awide variety of developmental processes (Nusse and Varmus1992; Wodarz and Nusse 1998; Dierick and Bejsovec 1999;Sokol 1999). Wnt ligands interact with receptors at the cellsurface, typically initiating a signaling cascade that shunts cy-toplasmic

b

-catenin away from degradation and toward com-bination with Lef1/T cell-specific factor to modulate genetranscription in the nucleus. There is also a less common

Wnt

signaling pathway acting via the phosphoinositol cycle to ele-vate intracellular calcium ion concentration and activate proteinkinase C (Sheldahl et al. 1999). Wnts can act either on the cellthat secreted them (autocrine feedback) or on neighboring cellsby paracrine signaling (Blader et al. 1996).

Wnt

expression be-fore the mid-blastula transition is maternal, while subsequentdevelopmental (and even adult)

Wnt

expression is zygotic.The present study shows that

AmphiWnt8

is expressedduring amphioxus development in the hindgut endoderm,forebrain, and early somitic mesoderm. Our phylogeneticanalysis of the

Wnt8

subfamily suggests that a single

Wnt8

gene in an ancestral invertebrate chordate was duplicatedearly in vertebrate evolution, giving rise to a

Wnt8b

cladewith purely neurogenic functions and a

Wnt8

clade with di-verse functions, including the dorsoventral patterning ofthe embryo. For example,

Xenopus

Wnt8

in the paraxialmesoderm acts as an upstream regulator of myogenesis inthe presomitic mesoderm and helps delimit it from the ven-tral mesoderm and nascent notochord (Fainsod et al. 1994;Hoppler and Moon 1998; Marom et al. 1999; Tian et al.1999). Similarly, our results suggest that one of the key de-velopmental functions of

AmphiWnt8

is to help pattern theamphioxus embryo dorsoventrally by specifying the pre-somitic mesoderm.

MATERIALS AND METHODS

Animal collection, library screening, and polymerase chain reaction

Ripe males and females of the Florida amphioxus (

Branchiostomafloridae

) were collected in Tampa Bay, Florida, and stimulated elec-trically to obtain gametes. After fertilization, the embryonic and larvalstages were cultured in the laboratory (Holland and Holland 1993).

Three cDNA libraries were used: one from mixed 8–18 h em-bryos in Lambda Zap II (Stratagene, La Jolla, CA) and two griddedlibraries in pSport1 (GIBCO-BRL, Gaithersburg, MD) (Zehetnerand Lehrach 1994; Lehrach et al. 1997). The gridded libraries, com-prising about 60,000 clones each, were from 5–6 h embryos andfrom 26 h embryos, respectively.

A 741-bp fragment (MPIMGBFLG 41F14) of amphioxus

AmphiWnt8

was isolated by oligonucleotide fingerprinting (G. D. Pa-nopoulou, unpublished data) from the gridded library of 5–6 h em-bryos. This fragment comprised the 3

9

end of the coding region plusthe 3

9

UTR. The remaining 5

9

portion of the

AmphiWnt8

clone wasobtained by polymerase chain reaction with the TaqPlus Precisionpolymerase chain reaction system (Stratagene) using the pBluescript-specific primer together with an

AmphiWnt8

-specific primer and 2

m

lof the library in Lambda Zap II as a template. The amplificationyielded two polymerase chain reaction products, each including the5

9

end of

AmphiWnt8

and overlapping the rest of the clone by 619 bp.

Southern blot analysis

Southern blots were prepared from genomic DNA from 20 speci-mens of

Branchiostoma floridae

according to Holland et al. (1996).Numbers at the top of the lanes in Fig. 2 refer to digestion in the fol-lowing restriction enzymes: 1,

Bam

H1; 2,

Bgl

II; 3,

Bst

II; 4,

Eco

0109;5,

Eco

RI; 6,

Bst

XI; 7,

Hind

III; 8,

Kpn

I; 9,

Pst

I; 10,

Not

I; 11,

Nco

I;12,

Pvu

I; 13,

Sal

I; 14,

Stu

I; 15,

Xba

I; 16,

Xho

I. To determine thenumber of amphioxus genes related to

Wnt8

and the gene copy num-ber, the blot was hybridized with a 540 bp fragment (from base 765to base 1315) that included the relatively conserved 3

9

part of the cod-ing region. The hybridization was at moderately low stringency in6

3

SSC, 0.2% SDS, 10

3

Denhardt’s, 0.1 mg/ml tRNA overnight at60

8

C. Washes were at 50

8

C in 2

3

SSC, 0.1% SDS.

Phylogenetic analysis of

Wnt8

subfamily proteins

The degree of conservation within the coding region of Wnt pro-teins, although greatest in the C-terminal half, is high enough in theN-terminal half to justify comparing full-length sequences. Full-length Wnt8 subfamily sequences were compared among nine dif-ferent species, and the

Caenorhabditis

Wnt protein Lin44 was usedas the outgroup. Accession numbers for the sequences were Lin44

Caenorhabditis

(2133466), AmphiWnt8 (AF190470), Wnt8 ze-brafish (U10869), Wnt8b zebrafish (U10870), Wnt8

Xenopus

(X57234), Wnt8b

Xenopus

(U22173), Wnt8 chicken (originallynamed Wnt8c) (U02097), Wnt8 mouse (Z68889), Wnt8b mouse(AF130349), and Wnt8b human (X91940). Trees were calculatedusing PAUP (version 3.1.1) in 100 rounds of heuristic random step-wise addition with 784 characters in the data matrix for the 10 taxa.Only one most parsimonious tree was retained (with a length of1027). Branch stability was assessed by bootstrap analysis with1000 cycles and 10 random stepwise additions per cycle.

Schubert et al.

Amphioxus

Wnt8

gene

87

In situ hybridization

In situ hybridization was performed according to the method ofHolland et al. (1996). A 741-bp 3

9

fragment including the 3

9

UTRand 390-bp of coding sequence was used as a template for an anti-sense riboprobe. The corresponding sense probe control producedno signal. Embryos and larvae were photographed as whole mountsand then prepared as histological sections (Holland et al. 1996).

RESULTS

Structure of amphioxus

AmphiWnt8

and Southern blot analysis

Figure 1 shows the nucleotide and deduced amino acid se-quences for the full-length cDNA of

AmphiWnt8.

The long-est open reading frame codes for 364 amino acids. The 70-bp

5

9

UTR includes one in-frame stop codon upstream from thepresumed start codon, and the 351-bp 3

9

UTR includes a ca-nonical polyadenylation signal (AATAAA) 15-bp upstreamfrom the poly(A) tail.

The genomic Southern blot analysis of

AmphiWnt8

(Fig.2) is based on genomic DNA pooled from 20 amphioxusadults and digested with 16 restriction enzymes. Most of thedigests yielded only a single band when hybridized at moder-ately low stringency against a probe that included the mostconserved part of the coding region of

AmphiWnt8.

The tripleand double bands, respectively, in the

Bst

EII (Fig. 2, lane 3)and

Eco

RI (lane 5) digests are presumably due to polymor-phism or to cutting within an intron in the probed region. Be-cause the probe recognized only a single band in most of thedigests, it is likely that amphioxus has only one

AmphiWnt8

gene that is present as a single copy in the genome.

Fig. 1. Nucleotide and deduced aminoacid sequence of AmphiWnt8 of Bran-chiostoma floridae (GenBank accessionnumber AF190470). In-frame stop codonspreceding the presumed translational startsite are underlined, as is the canonicalpolyadenylation signal.

88 EVOLUTION & DEVELOPMENT

Vol. 2, No. 2, March–April 2000

Phylogeny of the Wnt8 subfamily

Figure 3 shows phylogenetic analyses of the Wnt8 proteinsubfamily, based on full-length sequences with

Caenorhabdi-tis

Lin44 as the outgroup. In the tree, the branching order (in-vertebrates, invertebrate chordates, vertebrates) is congruentwith the commonly accepted phylogeny based on morpholog-ical characters, and AmphiWnt8 branches off at the base of thevertebrates. The tree topology and high bootstrap support val-ues indicate that a single, ancestral gene encoding Wnt8 dupli-

cated early in vertebrate evolution, giving rise to a Wnt8 cladeand a Wnt8b clade. In our analysis, the vertebrate Wnt8b cladeshows the expected branching order (fish, amphibian, mam-mal), which is supported by robust boostrap values. In con-trast, for reasons that are not obvious, the vertebrate Wnt8clade has an inverted branching order (mammal, bird, anam-niote) and is poorly supported by bootstrapping.

Developmental expression of

AmphiWnt8

Whole-mount in situ hybridization of oocytes, eggs, andcleavage stages revealed no consistent evidence for maternalexpression of

AmphiWnt8.

It is possible that transcripts werepresent in these early stages, but at levels detectable only bymore sensitive methods, like the reverse transcriptase poly-merase chain reaction used by Cui et al. (1995) to demonstratematernal transcripts of

Xenopus

Wnt8b. All the AmphiWnt8expression described in the present paper follows the mid-blastula transition and is, therefore, zygotic.

Expression of AmphiWnt8 is first detected in the cup-shaped gastrula. Transcripts are localized dorsolaterally (Fig.4A, arrows) on either side of the archenteron in regions ofthe invaginated mesendoderm that constitute the prospectiveparaxial mesoderm. By the very early neurula stage (Figs.4B–4D), the paraxial mesoderm is evaginating from the arch-enteron (as diagrammed in Fig. 5B) to form a presomiticgroove on either side of the midline; AmphiWnt8 is ex-pressed posteriorly in the paraxial mesoderm (Figs. 4C, ar-rowhead, 4D), but not anteriorly (Fig. 4C, arrow).

By the stage of the hatching neurula, several somites havepinched off from the presomitic grooves, and AmphiWnt8 tran-scripts are in the second, third, and sometimes fourth somites(somite numbering according to Holland et al. 1992) on eitherside of the midline (Figs. 4 E and 4F). In contrast, no expressionis detectable in the first (most anterior) or in somites posterior tothe third pair. Within each AmphiWnt8-expressing somite, tran-scripts are most conspicuous in the lateral (non-myotomal) cells(Fig. 4G). In the hatching neurula, there is also expression in theventral endoderm of the hindgut. (Figs. 4E and 4H).

By the midneurula stage, expression of AmphiWnt8 has al-most disappeared from the somites, but is still strong in thehindgut (Figs. 4I and 4J). In the 3-day larva (Figs. 4K and 4L)transcripts are no longer detectable in the hindgut, but have ap-peared near the anterior end of the neural tube in a few ventrallylocated cells of the cerebral vesicle (see Fig. 4L). In 1-week-oldlarvae, AmphiWnt8 expression can no longer be detected by insitu hybridization.

DISCUSSION

Evolution within the Wnt8 subfamilyWithin the Wnt8 subfamily, AmphiWnt8 occurs at the base ofthe vertebrate split into a Wnt8 clade and a Wnt8b clade. Thisbranching pattern suggests that a single precursor gene in the

Fig. 2. Genomic Southern blot analysis of DNA pooled from 20amphioxus adults. Numbers at the top of lanes refer to digestionwith restriction enzymes given in the Materials and Methods sec-tion. Blot probed at moderately low stringency with a 540-bpstretch of AmphiWnt8, including the relatively conserved 39 partof the coding region. Size markers at left.

Fig. 3. Phylogenetic tree of the Wnt8 subfamily based on full-length Wnt proteins (accession numbers are in the Materials andMethods section). The tree is constructed by random stepwise ad-dition with Caenorhabditis elegans Wnt protein Lin44 as an out-group, and the values at the branching points are bootstrappercentages (bootstrap values lower than 50% are not depicted).

Schubert et al. Amphioxus Wnt8 gene 89

Fig. 4. Embryonic and larval expression of AmphiWnt8 in whole mounts (anterior toward left; side views except for A, C, and F; scale lines50 mm) and cross-sections (counterstained pink; scale lines 25 mm). (A) Vegetal pole view of cup-shaped gastrula showing AmphiWnt8 expres-sion (arrows) dorsolaterally in prospective paraxial mesoderm on either side of archenteron. (B) Very early neurula with expression in paraxialmesoderm. (C) Dorsal view of preceding embryo showing expression posteriorly in paraxial mesoderm (arrowhead), but not anteriorly (ar-row). (D) Cross-section through level of arrowhead in (C), with expression in paraxial mesoderm. (E) Hatching neurula with expression in mus-cular somites posterior to the first and in the hindgut. (F) Dorsal view of preceding embryo with expression in somites posterior to the first andin the hindgut. (G) Section through level of arrowhead in (E) with expression in the lateral mesothelial cells of the somites. (H) Section throughlevel of arrow in (E) showing expression in the ventral endoderm of the hindgut. (I) Midneurula with conspicuous expression in the hindgut. (J)Dorsal view of preceding embryo. (K) Three-day larva with the primary pigment spot (arrow); expression is limited to ventral cells in the ce-rebral vesicle (arrowhead). (L) Enlargement of the anterior end of preceding larva showing expression in ventral cells of the cerebral vesicle (cv).

90 EVOLUTION & DEVELOPMENT Vol. 2, No. 2, March–April 2000

common ancestor of amphioxus and vertebrates continued asa single gene in the cephalochordate line of evolution, butduplicated during early vertebrate evolution. In the verte-brates, separate Wnt8 and Wnt8b genes are already present atthe level of the teleost fishes. A more exact timing of thisgene duplication could be established by isolating Wnt8 sub-family genes from jawless vertebrates (hagfishes and lam-preys). If the expression patterns for chordate genes in theWnt8 subfamily are mapped onto the phylogenetic tree inFig. 3, it appears that vertebrate members of this subfamilyhave split into a Wnt8b clade expressed exclusively in theneuroectoderm plus a Wnt8 clade expressed in the mesodermand several other tissues. In comparison, amphioxus AmphiWnt8is expressed in domains that resemble those of both these cladesof vertebrate genes.

AmphiWnt8 and its vertebrate homologs in the Wnt8b cladeIn amphioxus, neural expression of AmphiWnt8 is detected inearly larvae in the cerebral vesicle at a position that probably cor-responds to the posterior diencephalic forebrain, as judged frommicroanatomy and the expression domains of other develop-mental genes (Holland and Holland 1999). The neural expres-

sion of AmphiWnt8 appears most comparable to Wnt8b tran-scription in the diencephalon of zebrafish (Kelly et al. 1995),Xenopus (Cui et al. 1995), chicken (Hollyday et al. 1995), andmammals (Lako et al. 1998; Richardson et al. 1999), but not toadditional expression domains of these vertebrate genes, whichmay be found more anteriorly and/or more posteriorly in the de-veloping brain. The functions of the vertebrate Wnt8b geneshave not been studied experimentally, but it has been suggestedthat they play roles in the rostrocaudal patterning of the anteriorneural tube (Cui et al. 1995; Kelly et al. 1995; Lako et al. 1998),and a comparable neural function for amphioxus AmphiWnt8would seem likely.

AmphiWnt8 and its vertebrate homologs in the Wnt8 clade: nonmesodermal expressionBoth AmphiWnt8 and Xenopus Wnt8 are expressed tran-siently in the hindgut (Christian et al. 1991; Lemaire andGurdon 1994), although there have been no speculationsabout the possible functional importance of this endodermaltranscription. In higher vertebrates, Wnt8 genes are ex-pressed in the developing central nervous system, most con-spicuously at the level of hindbrain rhombomere 4 (Humeand Dodd 1993; Bouillet et al. 1996). This neural expression,which has not been observed in lower vertebrates, appears tobe distinct from that of vertebrate Wnt8b (discussed above).Thus, genes of the Wnt8 clade may have been co-opted forhindbrain functions relatively late in vertebrate evolution.

AmphiWnt8 and its vertebrate homologs in the Wnt8 clade: mesodermal expressionBoth amphioxus AmphiWnt8 and vertebrate Wnt8 are con-spicuously expressed in the early mesoderm. There is no uni-form terminology for the subdivisions of the early mesodermalong the dorsoventral (or, correspondingly in higher verte-brates, the mediolateral) axis. Here, we will use the termsmid-dorsal mesoderm (the source of the notochord), paraxial(alternatively called lateral) mesoderm (the source of thesomites), and ventral mesoderm. For this discussion, it is im-portant to understand that only the mid-dorsal and paraxialmesoderm are initially produced by amphioxus gastrulation,leaving the ventral mesoderm to be produced considerablylater in development (Figs. 5A–5C). In contrast, all threesubdivisions of the mesoderm are produced almost simulta-neously by vertebrate gastrulation (Figs. 5D and 5E).

In amphioxus and lower vertebrates, the earliest zygotictranscripts of Wnt8 genes are detected in newly induced me-soderm, except mid-dorsally (Christian et al. 1991; Smithand Harland 1991; Christian and Moon 1993; Kelly et al.1995; Hoppler et al. 1996; Hoppler and Moon 1998; Bang etal. 1999; Tian et al. 1999). In higher vertebrates, the earlymesodermal transcription of Wnt8 genes is preceded by aneven earlier phase of expression in the epiblast (Hume andDodd 1993; Bouillet et al. 1996).

Fig. 5. Diagrammatic cross sections of embryos. Amphioxus atthe stages of (A) early gastrula, (B) early neurula, and (C) lateneurula; generalized vertebrate at stages of (D) early gastrula and(E) early neurula. In amphioxus, the mid-dorsal mesoderm andthe paraxial mesoderm arise before the ventral mesoderm. Bycontrast, in vertebrates, the advent of these three subdivisions ofthe mesoderm is virtually simultaneous. Abbreviations: en, endo-derm; mdm, mid-dorsal mesoderm; nc, nerve cord, no, notochord,np, neural plate; pm, paraxial mesoderm; vm, ventral mesoderm.

Schubert et al. Amphioxus Wnt8 gene 91

Intermediate levels of BMP4 expression can up-regulateWnt8 transcription in vertebrates (Hoppler and Moon 1998;Marom et al. 1999), and AmphiBMP2/4, the amphioxus ho-molog of vertebrate BMP4 (Panopoulou et al. 1998), couldwell up-regulate expression of AmphiWnt8 in the paraxial me-soderm, where these two amphioxus genes are co-expressed.

Xenopus Wnt8 has been strongly implicated in the earlieststages of somitic myogenesis, because it maintains the ex-pression of MyoD (Hoppler et al. 1996). However, in bothXenopus and mouse, Wnt8 is rapidly downregulated in theparaxial mesoderm (Bouillet et al. 1996; Tian et al. 1999),evidently to prevent overcommitment of somitic cells to amyogenic fate (Tian et al. 1999). Amphioxus AmphiWnt8 issimilarly expressed in the paraxial mesoderm before thesomites become morphologically detectable, but, in contrastto its vertebrate homologs, continues to be transcribed in atleast some of the differentiated somites. It is unclear why thefirst somite expresses no detectable AmphiWnt8, while thesecond through fourth somites do; this inconsistency doessuggest that AmphiWnt8 may not be involved in intrasomiticpatterning (say, distinguishing myotome from sclerotome).

In vertebrates, there is evidence that mesodermally pro-duced Wnt8 ligands not only act within the mesoderm itself,but also influence the overlying neuroectoderm. For example,experimental ectopic expression and dominant-negative ap-proaches indicate that Xenopus Wnt8 ligands, presumably em-anating from the mesoderm, help pattern the overlying neuralplate and nascent neural crest both rostrocaudally and me-diolaterally (Fredieu et al. 1997; LaBonne and Bronner-Fraser1998; Bang et al. 1999; Niehrs 1999). For amphioxus, althoughthere is no experimental evidence that mesodermal transcrip-tion of AmphiWnt8 affects the developing neuroectoderm, thespatiotemporal relations between mesodermal AmphiWnt8 andneuroectodermal AmphiPax3/7 (Holland et al. 1999) are strik-ingly similar to those between mesodermal Wnt8 and the neu-ral marker Pax3, which have been shown experimentally to becausally related in Xenopus (Bang et al. 1999).

Wnt8 homologs in vertebrate gastrulae are expressed notonly in the paraxial mesoderm, which is somitogenic, butalso in the ventral mesoderm. Hoppler et al. (1996) sug-gested that Wnt8 transcription in the ventral mesoderm mightactivate downstream genes specifying that tissue. In con-trast, Salic et al. (1997) proposed that Xenopus Wnt8 ligandsin the ventral mesoderm are not effective, because they aretitrated there by the product of the sizzled gene, and Kelly etal. (1995) demonstrated a very rapid down-regulation of ze-brafish Wnt8 in the nascent ventral mesoderm.

In sum, the mesodermal expression pattern of AmphiWnt8in amphioxus suggests that Wnt8 genes in the protochordateancestors of the vertebrates probably functioned for the earlystages of somitogenesis in the paraxial mesoderm, but didnot pattern the ventral mesoderm. Marom et al. (1999) haverecently suggested that vertebrate BMP4, when expressed at

high levels, is involved in the specification of ventral meso-derm. In amphioxus, the homologous gene, AmphiBMP2/4,is expressed strongly in only a few cells of the ventral meso-derm (Panopoulou et al. 1998) and thus may not be involvedin the specification of the amphioxus ventral mesoderm as awhole. Thus, there are preliminary indications that the ge-netic program initiating ventral mesoderm formation in am-phioxus may not be vertebrate-like, and its further studyshould help elucidate the evolutionary roots of dorsoventralpatterning mechanisms in the vertebrates.

AcknowledgmentsWe are deeply indebted to John M. Lawrence and Ulrike Schönherrand Jim Langeland for their generous provision of facilities, libraries,and assistance. This research was supported in part by National Sci-ence Foundation research grant IBN 96-309938 (N. D. H. and L. Z.H.) and in part by the Max-Planck-Gesellschaft (G. D. P. and H. L.).

REFERENCES

Arendt, D., and Nübler-Jung, K. 1994. Inversion of dorsoventral axis? Na-ture 371: 26.

Bang, A. G., Papalopulu, N., Goulding, M. D., and Kintner, C. 1999. Expres-sion of Pax-3 in the lateral neural plate is dependent on a Wnt-mediatedsignal from posterior nonaxial mesoderm. Dev. Biol. 212: 366–380.

Blader, P., Strähle, U., and Ingham, P. W. 1996. Three Wnt genes expressed ina wide variety of tissues during development of the zebrafish, Danio rerio:developmental and evolutionary perspectives. Dev. Genes Evol. 206: 3–13.

Bouillet, P., Oulad-Abdelghani, M., Ward, S. J., Bronner, S., Chambon, P.,and Dollé, P. 1996. A new mouse member of the Wnt family, mWnt-8, isexpressed during early embryogenesis and is ectopically induced by re-tinoic acid. Mech. Dev. 58: 141–152.

Christian, J. L., and Moon, R. T. 1993. Interactions between wnt-8 and Spe-mann organizer signaling pathways generate dorsoventral pattern in theembryonic mesoderm of Xenopus. Genes Dev. 7: 13–28.

Christian, J. L., McMahon, J. A., McMahon, A. P., and Moon, R. T. 1991.Xwnt-8, a Xenopus Wnt-1/int-1-related gene responsive to mesoderm-inducing growth factors, may play a role in ventral mesoderm pattern-ing during embryogenesis. Development 111: 1045–1055.

Cui, Y. Z., Brown, J. D., Moon, R. T., and Christian, J. L. 1995. Xwnt-8: amaternally expressed Xenopus Wnt gene with a potential in establishingthe dorsoventral axis. Development 121: 2177–2186.

De Robertis, E. M., and Sasai, Y. 1996. A common plan for dorsoventralpatterning in Bilateria. Nature 380: 37–40.

Dierick, H., and Bejsovec, A. 1999. Cellular mechanisms of wingless/Wntsignal transduction. Curr. Top. Dev. Biol. 43: 153–190.

Fainsod, A., Steinbeisser, H., and De Robertis, E. M. 1994. On the functionof BMP-4 in patterning the marginal zone of the Xenopus embryo.EMBO J. B13: 5015–5025.

Ferguson, E. L., and Anderson, K. V. 1992. decapentaplegic acts as a mor-phogen to organize dorso-ventral pattern in the Drosophila embryo.Cell 71: 451–461.

Fredieu, J. R., Cui, Y. Z., Maier, D., Danilchik, M. V., and Christian, J. L.1997. Xwnt-8 and lithium can act upon either dorsal mesodermal orneurectodermal cells to cause a loss of forebrain in Xenopus embryos.Dev. Biol. 186: 100–114.

Graff, J. M. 1997. Embryonic patterning: to BMP or not to BMP, that is thequestion. Cell 89: 171–174.

Hemmati-Brivanlou, A., and Melton, D. 1997. Vertebrate embryonic cellswill become nerve cells unless told otherwise. Cell 88: 13–17.

Holland, L. Z., and Holland, N. D. 1999. Chordate origins of the vertebratecentral nervous system. Curr. Opin. Neurobiol. 9: 596–602.

Holland, L. Z., Holland, P. W. H., and Holland, N. D. 1996. Revealing ho-mologies between body parts of distantly related animals by in situ hy-bridization to developmental genes: amphioxus versus vertebrates. In

92 EVOLUTION & DEVELOPMENT Vol. 2, No. 2, March–April 2000

J. D. Ferraris and S. R. Palumbi (eds.). Molecular Zoology: Advances,Strategies, and Protocols. Wiley-Liss, New York, pp. 267-282, 473–483.

Holland, L. Z., Schubert, M., Kozmik, Z., and Holland, N. D. 1999.AmphiPax3/7, an amphioxus paired box gene: insights into chordatemyogenesis, neurogenesis, and the possible evolutionary precursor ofdefinitive vertebrate neural crest. Evol. Dev. 1: 153–165.

Holland, N. D., and Holland, L. Z. 1993. Embryos and larvae of inverte-brate deuterostomes. In C. D. Stern and P. W. H. Holland (eds.). Essen-tial Developmental Biology: A Practical Approach. IRL Press, Oxford,pp. 21–32.

Holland, P. W. H., Holland, L. Z., Williams, N. A., and Holland, N. D.1992. An amphioxus homeobox gene: sequence conservation, spatialexpression during development and insights into vertebrate evolution.Development 122: 653–661.

Holland, P. W. H., Garcia-Fernàndez, J., Williams, N. A., and Sidow, A.1994. Gene duplications and the origins of vertebrate development. De-velopment 1994 supplement: 125–133.

Holley, S. A., Jackson, P. D., Sasai, Y., Lu, B., De Robertis, E. M., Hoff-mann, F. M., and Ferguson, E. L. 1995. A conserved system for dorsal-ventral patterning in insects and vertebrates involving sog and chordin.Nature 376: 249–253.

Holley, S. A., Neul, J. L., Attisano, L., Wrana, J. L., Sasai, Y., O’Connor,M. B., De Robertis, E. M., and Ferguson, E. L. 1996. The Xenopus dor-salizing factor noggin ventralizes Drosophila embryos by preventingDPP from activating its receptor. Cell 86: 607–617.

Hollyday, M., McMahon, J. A., and McMahon, A. P. 1995. Wnt expressionpatterns in chick embryo nervous system. Mech. Dev. 52: 9–25.

Hoppler, S., and Moon, R. T. 1998. BMP-2/-4 and Wnt-8 cooperatively pat-tern the Xenopus mesoderm. Mech. Dev. 71: 119–129.

Hoppler, S., Brown, J. D., and Moon, R. T. 1996. Expression of a domi-nant-negative Wnt blocks induction of MyoD in Xenopus embryos.Genes Dev. 10: 2805–2817.

Hume, C. R., and Dodd, J. 1993. Cwnt-8c: a novel Wnt gene with a potentialrole in primitive streak formation and hindbrain formation. Develop-ment 119: 1147–1160.

Kelly, G. M., Greenstein, P., Erezyilmaz, D. F., and Moon, R. T. 1995. Ze-brafish wnt8 and wnt8b share a common activity but are involved in dis-tinct developmental pathways. Development 121: 1787–1799.

La Bonne, C., and Bronner-Fraser, M. 1998. Neural crest induction in Xe-nopus: evidence for a two-signal model. Development 125: 2403–2414.

Lako, M., Lindsay, S., Bullen, P., Wilson, D. I., Robson, S. C., and Strachan,T. 1998. A novel mammalian Wnt gene, WNT8B, shows brain-restrictedexpression in early development, with sharply delimited expressionboundaries in the developing forebrain. Hum. Mol. Genet. 7: 813–822.

Lehrach, H., Bancroft, D., and Maier, E. 1997. Robotics, computing, andbiology: an interdisciplinary approach to the analysis of complex ge-nomes. Interdisc. Sci. Rev. 22: 37–44.

Lemaire, P., and Gurdon, J. B. 1994. A role for cytoplasmic determinantsin mesoderm patterning: cell-autonomous activation of the goosecoidand Xwnt-8 genes along the dorsoventral axis of early Xenopus em-bryos. Development 120: 1191–1199.

Le Mosy, E. K., Kemler, D., and Hashimoto, C. 1998. Role of nudel pro-tease activation in triggering dorsoventral polarity of the Drosophilaembryo. Development 125: 4045–4053.

Marom, K., Fainsod, A., and Steinbeisser, H. 1999. Patterning of the meso-derm involves several threshold responses to BMP-4 and Xwnt-8. Mech.Dev. 87: 33–44.

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 inXenopus embryos coincides with the dorsal enrichment of dishevelledthat is dependent on cortical rotation. J. Cell Biol. 146: 427–437.

Nellen, D., Burke, R., Struhl, G., and Balser, K. 1996. Direct and longrange action of a DPP morphogen gradient. Cell 85: 357–368.

Niehrs, C. 1999. Head in the Wnt: the molecular nature of Spemann’s headorganizer. TiG 15: 314–319.

Nusse, R., and Varmus, H. E. 1992. Wnt genes. Cell 69: 1073–1987.Padgett, R. W., Wozney, J. M, and Gelbart, W. M. 1993. Human BMP se-

quences can confer normal dorsal-ventral patterning in the Drosophilaembryo. Proc. Natl. Acad. Sci. USA 90: 2905–2909.

Panopoulou, G. D., Clark, M. D., Holland, L. Z., Lehrach, H., and Holland,N. D. 1998. AmphiBMP2/4, an amphioxus bone morphogenetic proteinclosely related to Drosophila decapentaplegic and vertebrate BMP2and BMP4: insights into evolution of dorsoventral axis specification.Dev. Dynam. 213: 130–139.

Piccolo, A., Sasai, Y., Lu, B., and De Robertis, E. M. 1996. Dorsoventralpatterning in Xenopus: inhibition of ventral signals by direct binding ofchordin to BMP-4. Cell 86: 589–598.

Richardson, M., Redmond, D., Watson, C., and Mason, J. O. 1999. MouseWnt8b is expressed in the developing forebrain and maps to chromo-some 19. Mamm. Genome 10: 923–925.

Salic, A. N., Kroll, K. L., Evans, L. M., and Kirschner, M. W. 1997. Sizzled:a secreted Xwnt8 antagonist expressed in the ventral marginal zone ofXenopus embryos. Development 124: 4739–4748.

Sheldahl, L. C., Park, M., Malbon, C. C., and Moon, R. T. 1999. Protein ki-nase C is differentially stimulated by Wnt and Frizzled homologs in aG-protein-dependent manner. Curr. Biol. 9: 695–698.

Smith, W. C., and Harland, K. M. 1991. Injected Xwnt-8 RNA acts early inXenopus embryos to promote formation of a vegetal dorsalizing center.Cell 67: 753–765.

Sokol, S. Y. 1999. Wnt signaling and dorso-ventral axis specification in ver-tebrates. Curr. Opin. Genet. Dev. 9: 405–410.

St. Johnston, R. D., and Gelbart, W. M. 1987. Decapentaplegic transcriptsare localized along the dorsal-ventral axis of the Drosophila embryo.EMBO J. 6: 2785–2791.

Tian, Q., Nakayama, T., Dixon, M. P., and Christian, J. L. 1999. Post-tran-scriptional regulation of Xwnt-8 expression is required for normal myo-genesis during vertebrate embryonic development. Development 126:3371–3380.

Tonegawa, A., Funayama, N., Ueno, N., and Takahashi, Y. 1997. Mesoder-mal subdivision along the mediolateral axis in chicken controlled by dif-ferent concentrations of BMP-4. Development 124: 1975–1984.

Wada, H., and Satoh, N. 1994. Details of the evolutionary history from in-vertebrates to vertebrates, as deduced from the sequence of 18S rDNA.Proc. Natl. Acad. Sci. USA 91: 1801–1804.

Wodarz, A., and Nusse, R. 1998. Mechanisms of Wnt signaling in develop-ment. Ann. Rev. Cell Dev. Biol. 14: 59–88.

Zehetner, G., and Lehrach, H. 1994. The reference library system—sharingbiological material and experimental data. Nature 367: 489–491.