Colinear and Segmental Expression of Amphioxus Hox Genes

C

Developmental Biology 213, 131–141 (1999)Article ID dbio.1999.9369, available online at http://www.idealibrary.com on

Colinear and Segmental Expressionof Amphioxus Hox Genes

Hiroshi Wada,*,1 Jordi Garcia-Fernandez,† and Peter W. H. Holland**School of Animal & Microbial Sciences, University of Reading, Whiteknights, PO Box 228,Reading RG6 6AJ, United Kingdom; and †Departament de Genetica, Facultat de Biologia,Universitat de Barcelona, Av. Diagonal 645, 08071 Barcelona, Spain

The cephalochordate amphioxus has a single Hox gene cluster. Here we describe the genomic organization of four adjacentamphioxus genes, AmphiHox-1 to AmphiHox-4, together with analysis of their spatiotemporal expression patterns. Wedemonstrate that these genes obey temporal colinearity and that three of the genes also obey spatial colinearity in thedeveloping neural tube. AmphiHox-1, AmphiHox-3, and AmphiHox-4 show segmental modulation of their expressionlevels, a two-segment phasing of spatial colinearity, and, at least for AmphiHox-4, asymmetrical expression. AmphiHox-2is unlike other amphioxus Hox genes: it does not obey spatial colinearity and it has no positional expression in the neuraltube. AmphiHox-2 is expressed in the preoral pit of larvae, from which the homologue of the anterior pituitary develops. Wesuggest that the ancestral role of chordate Hox genes was primarily in the neural tube and that chordate Hox genes canfunctionally diverge in a manner analogous to that of Drosophila ftz or zen. © 1999 Academic Press

Key Words: amphioxus; Hox genes; colinearity; segmentation.

INTRODUCTION

The conservation of Hox gene clustering is remarkable,having been described from animals as diverse as nema-todes, arthropods, nemerteans, echinoderms, cephalochor-dates, and vertebrates (Kenyon and Wang, 1991; Garcia-Fernandez and Holland, 1994; Popodi et al., 1996; Kmita-

unisse et al., 1998; Martinez et al., 1999). Nonetheless,several examples of modification have been described. Thephysical integrity of the Hox gene cluster has been second-arily broken in two Drosophila species (von Allmen et al.,1996) and interrupted by a probable inversion in Caeno-rhabditis (Burglin and Ruvkun, 1993). Examples of diver-gence of individual genes include Drosophila ftz and zen:Hox genes that have escaped the constraints of spatialcolinearity and diverged structurally and functionally fromtheir neighbors within a Hox cluster (Dawes et al., 1994;Falciani et al., 1996).

Another means by which Hox genes have been modifiedin evolution is by elaboration of their expression patterns.

Sequence data from this article have been deposited with theGenBank Data Library under Accession Nos. AB028206–AB028208.

1 Present address: Seto Marine Biological Laboratory, Kyoto Uni-versity, Shirahama-cho, Nishimuro-gun, Wakayama 649-22, Japan.

0012-1606/99 $30.00Copyright © 1999 by Academic PressAll rights of reproduction in any form reserved.

For example, mouse Hox genes are spatially expressed in amultitude of tissues including the neural tube, derivativesof neural crest cells, somitic mesoderm, intermediate me-soderm, and lateral plate mesoderm. Several lines of evi-dence imply that this complexity reflects addition of regu-latory controls during evolution. For example, the sameHox gene usually has different spatial expression limits indifferent tissues, and enhancers specific for different tissueshave been identified in Hox gene clusters (for example,Morrison et al., 1997; Zhang et al., 1997). Furthermore,gain-of-function and loss-of-function mutations in micedemonstrate that Hox genes play distinct spatial patterningroles within each tissue type, implying that the elaborationof Hox gene regulation during evolution has been of func-tional relevance (reviewed by Krumlauf, 1994).

It is unclear how and when these functional changesoccurred during Hox gene evolution. For example, in whichtissues were Hox genes deployed before the origin of verte-brates? Which additional expression sites were added beforeand after duplication of the Hox gene cluster? In addition tothe evolutionary implications, resolution of these questionsis relevant to analysis of vertebrate Hox gene functions,since it may reveal the extent to which roles are shared
between the gene clusters. To distinguish ancestral andderived roles, it is essential to deduce the primitive expres-
131

HaAsmidpisptpoa

mwDAcot

f1tsus

ntraor all

132 Wada, Garcia-Fernandez, and Holland

sion patterns of Hox genes in the ancestors of vertebrates.Analysis of Hox genes in the cephalochordate amphioxuscan give clues to ancestral patterns when considered in aphylogenetic context. Cephalochordates are the closestliving invertebrates to the vertebrates (Wada and Satoh,1994) so comparisons to vertebrates are less likely to becomplicated by excessive divergence in the two lineages.Any differences observed could have arisen in either lin-eage; use of an outgroup (such as urochordates) is requiredto confidently determine the ancestral state. From thepresent perspective, it is also helpful that amphioxus pos-sesses a single Hox gene cluster; hence, any complicationsrelating to the independent divergence of paralogous geneclusters is avoided. In addition, most of the amphioxus Hoxgenes are related to the vertebrate Hox paralogy groups in asimple 1:1 correspondence (Garcia-Fernandez and Holland,1994). There have been no complicating tandem gene du-plications in the lineage leading to amphioxus (at least inthe “anterior” and “middle” parts of the cluster).

Expression patterns have been described previously fortwo amphioxus Hox genes, AmphiHox-3 and AmphiHox-1(Holland et al., 1992; Holland and Garcia-Fernandez, 1996;

olland and Holland, 1996). Each gene was shown to haven anterior expression limit in the neural tube, withmphiHox-1 also having epidermal expression. Here we

ignificantly refine these descriptions using more accurateethodology, including higher sensitivity protocols, double

n situs, and examination from multiple angles. We alsoescribe expression of AmphiHox-2 and AmphiHox-4 andresent full gene sequences for the genes studied. We detectntrasegmental periodicity of expression levels, double-egment periodicity of spatial colinearity, asymmetric ex-ression of AmphiHox-4, and deviation from colinearity byhe AmphiHox-2 gene. Expression of the latter gene only inosterior mesoderm (early) and preoral pit (late) is unlikether Hox genes; we suggest it has lost its original role innteroposterior patterning.

MATERIALS AND METHODS

Cloning and Sequencing

FIG. 1. Genomic organization of AmphiHox-1 to AmphiHox-4. Ulines, open boxes, and closed boxes, respectively. 59 is to the left f

Gene sequences for AmphiHox-1, AmphiHox-2, andAmphiHox-4 were determined from Branchiostoma floridae

(c

Copyright © 1999 by Academic Press. All right

genomic clones (Garcia-Fernandez and Holland, 1994) after sub-cloning into pUC plasmid vectors. To determine intron positionsand polyadenylation sites, cDNA clones of AmphiHox-1,AmphiHox-2, AmphiHox-3, and AmphiHox-4 were isolated fromcDNA libraries using genomic clones as probes and partiallysequenced. The cDNA libraries were a B. floridae 5- to 24-h embryolibrary from J. Langeland (University of Kalamazoo, MI) and a B.floridae larval library from L. Z. Holland (Scripps Institution ofOceanography, San Diego, CA).

Whole-mount in situ hybridization

B. floridae embryos were collected, fixed, and stored as describedby Holland et al. (1992). Whole-mount in situ hybridization wasperformed as described for ascidian embryos by Yasuo and Satoh(1994), except that treatment with acetic anhydride/triethanolamine and RNase was omitted, and proteinase K treat-ment was 2 mg/ml for 10 min (embryos younger than 36 h) or 20–30

in (36- and 60-h embryos). Double-staining in situ hybridizationas performed as described in Wada et al. (1998). Clones used forIG-labeled probe preparation were the cDNA clones ofmphiHox-1, AmphiHox-2, and AmphiHox-4 and the genomic

lone of AmphiHox-3 used by Holland et al. (1992). A cDNA clonef AmphiMLC-alk described by Holland et al. (1995) was used as aemplate for the fluorescein-labeled probe.

RESULTS

Genomic Structure and Sequences of AmphioxusHox Genes

Garcia-Fernandez and Holland (1994) demonstrated thatamphioxus Hox genes are arranged in a single cluster in thegenome. Here we investigated the transcription units of thefour most 39 genes in the cluster: AmphiHox-1 toAmphiHox-4. Comparison of cDNA and genomic clonesrevealed the presence of a single intron 59 of the homeoboxin AmphiHox-1 and AmphiHox-4, as previously predictedrom genomic sequence for AmphiHox-3 (Holland et al.,992). We also confirmed this AmphiHox-3 predicted in-ron position using a cDNA clone. AmphiHox-2 shares theame intron position, plus an additional intron in the 39ntranslated region (Fig. 1). The complete deduced proteinequences of AmphiHox-1, -2, and -4 were determined

nslated regions, coding regions, and homeoboxes shown as thickgenes.

AmphiHox-3 was reported previously). All four genes en-ode a characteristic hexapeptide N-terminal to the homeo-

s of reproduction in any form reserved.

F

sA

apneAtoupmai

133Expression of Amphioxus Hox Genes

domain, predicted to interact with products of the pbx/exdclass of homeobox genes. The DNA sequences ofAmphiHox-1, AmphiHox-2, and AmphiHox-4 are given inig. 2 and are available from the GenBank database.

Spatiotemporal Expression of AmphiHox Genes

Partial descriptions of AmphiHox-1 and AmphiHox-3expression have been reported previously (Holland et al.,1992; Holland and Garcia-Fernandez, 1996; Holland andHolland, 1996). Here we examine these genes in more detailand compare to the previously unstudied genesAmphiHox-2 and AmphiHox-4. Refinement of the in situhybridization protocol and examination of a greater agerange of embryos and larvae allowed us to detect details ofspatiotemporal expression pattern that were not revealed inprevious studies. Furthermore, because the somites of am-phioxus are slightly out of phase between the left and theright side (right somites are displaced posteriorly relative totheir left counterparts), we observed that the relative posi-tion of neural tube expression against somites can bemisinterpreted if observed from the lateral side. In order toavoid this problem, we examined boundaries of expressionfrom the dorsal side of embryos.

AmphiHox-1. The first sign of AmphiHox-1 expressionwas observed in late-gastrula embryos (8–9 h) (Fig. 3A).Expression at this stage is weak and restricted to cellssurrounding the blastopore. We conclude that this is theearliest stage and site of expression, since no signal wasdetected in younger embryos, including 7-h embryos (cup-shaped gastrula). At a slightly late stage (10 h), a strongsignal is also detected in the neural plate (Fig. 3B); we notethat the anterior border of this neural plate expressioncoincides with that in the mesoderm. By 12.5 h develop-ment, when four or five somite pairs are visible, thestrongest staining in the neural plate is at the level ofsomite 4 (s4; Fig. 3C). In 14-h embryos (when 7 somites arevisible), the neural plate expression clearly resolves intofour bilateral groups of cells, at intervals along the antero-posterior axis. The most anterior patch of expression isweak and adjacent to the posterior part of s2; each of theother patches of expression are stronger and located adja-cent to the posterior part of each subsequent somite (Fig.3D). Thus AmphiHox1 is expressed segmentally in theneural plate from the level of mid-s2 to the boundarybetween s5 and s6 (s5/6) with a periodicity matching thesomitic segmentation. There is no comparable repeatingexpression in mesoderm, although the most posterior me-soderm is expressing AmphiHox-1 at this stage (Figs. 3Dand 3E). The anterior boundary of AmphiHox-1 neuralexpression is one somite more anterior than reported byHolland and Holland (1996); this reflects the weak expres-sion detected here at the level of posterior s2. In 18-hembryos (when 10 somites are visible) expression is ob-served in the neural tube from the same anterior limit
(mid-s2), but now a posterior limit to expression is evident(level of s6/7). This is confirmed in embryos in which the

somites are made clearer by double-staining in situ hybrid-ization with AmphiMLC-alk (Holland et al., 1995; Figs.3F–3H). Expression is still detectable in posterior meso-derm, although the level of expression is decreased com-pared to that of 14-h embryos (compare Figs. 3E and 3F). Asnoted by Holland and Garcia-Fernandez (1996),AmphiHox-1 is also expressed in epidermis with clearanterior and posterior boundaries; we determined that theseboundaries lie adjacent to somite boundaries s3/4 and s6/7(Fig. 3H). Epidermal expression has not been observed inother AmphiHox genes; however, the ascidian group 1 Hoxgene HrHox1 shows a similar epidermal expression domain(Katsuyama et al., 1994).

AmphiHox-3. The earliest expression of AmphiHox-3 isobserved in 10-h embryos, slightly later than the earliestAmphiHox-1 expression. At this stage, expression is de-tected in posterior mesoderm and more strongly in neuralplate cells. The two expressing tissues share a commonanterior boundary (Fig. 4A). Between 10 and 12.5 h ofdevelopment the intensity of AmphiHox-3 expressiongreatly increases, as noted for AmphiHox-1 between 8 and10 h. In 12.5-h embryos, AmphiHox-3 expression in theneural tube has an anterior boundary at the level of s4/5(Fig. 4B); by 14 h this expression limit is opposite mid-s4(Fig. 4C). As with AmphiHox-1, the neural expressionshows intrasegmental periodicity in its intensity (strongerat the level of the posterior halves of somites), and the mostanterior expression is weakest. In 18-h embryos neural tubeexpression continues to be observed from the level ofmid-s4, as confirmed by double-staining in situ hybridiza-tion with AmphiMLC-alk (Figs. 4D–4F). This neural tubeexpression reaches to the posterior end of the embryo.Unlike AmphiHox-1, AmphiHox-3 does not have posteriorlimit to expression in the neural tube. Strong expression inposterior mesoderm is also observed in this stage. This doesnot retain a fixed anterior limit relative to any particularsomite, but remains posterior during development. Theintensity of AmphiHox-3 expression in posterior mesodermis as strong as in neural tube, in contrast to the mesodermalexpression of AmphiHox-1, which becomes weaker at thistage. No epidermal expression is observed formphiHox-3.AmphiHox-4. Expression of AmphiHox-4 is not detect-

ble until 14 h, when a weak but specific signal is seen inosterior mesoderm and neural tube (Figs. 5A and 5B). Theeural tube expression has a clear anterior boundary in 18-hmbryo, at the level of mid-s6 (Figs. 5C–5E). As withmphiHox-1 and AmphiHox-3, there is a segmental pattern

o the neural expression, with stronger signals at the levelf the posterior half of somites. Like AmphiHox-3, butnlike AmphiHox-1, the neural tube expression reaches theosterior end without a boundary. Expression in posterioresoderm is also observed; this does not respect a fixed

nterior boundary relative to the somites (Fig. 5C). Annteresting feature of AmphiHox-4 expression is the asym-

metrical nature of its anterior expression limit in the neuraltube; expression extends more anteriorly on the left side of



FIG. 2. DNA and deduced protein sequences of AmphiHox-1, AmphiHox-2, and AmphiHox-4. Intron sequence shown in lowercasecharacters.

Copyright © 1999 by Academic Press. All rights of reproduction in any form reserved.

AbpAeAeesr6t

(e1(aepcs

Con


the embryo (Fig. 5D). This confirms a close relationshipbetween neural expression patterns and segmentation insomitic mesoderm, since left somites are displayed slightlyanterior to right somites in amphioxus, particularly towardthe posterior of the embryo.

AmphiHox-2. Although the expression patterns ofmphiHox-1, AmphiHox-3, and AmphiHox-4 genes areroadly similar, AmphiHox-2 shows a completely differentattern of expression. The earliest expression ofmphiHox-2 is observed in the posterior part of 10-hmbryos (Fig. 6A). However, unlike the expression ofmphiHox-1 and AmphiHox-3 at the same stage, no clearxpression is observed within the neural plate. In 12.5-hmbryos, expression of AmphiHox-2 is observed withinomite 5, and weaker expression is detected in more poste-ior mesoderm; signal is not detected in the neural tube (Fig.

FIG. 2—

B). Optical sectioning of 12.5-day embryos confirms thathe expression of AmphiHox-2 is predominantly in somites


Fig. 6C), quite unlike the expression of the other genesxamined. The mesodermal expression is still detectable in4-h embryos, but no expression is detected in neural plateFig. 6D). This unusual pattern of expression is transient,nd no expression of AmphiHox-2 is detected in 18- to 36-hmbryos. In 60-h larvae, a strong new signal emerges in thereoral pit, a large organ located on the left side of the oralavity (Fig. 6E). A dorsal view clearly indicates that theignal is on the left side of the body (Fig. 6F).

DISCUSSION

Spatial and Temporal ColinearityAnalysis of the genomic organization of AmphiHox-1 to

AmphiHox-4 confirmed that these genes are organized in a

tinued

manner similar to that of their homologues in vertebrates.All genes encode a hexapeptide motif, separated from the


edspoAFb


FIG. 3. Spatial expression pattern of AmphiHox-1. (A) 8- to 9-h amphioxus embryo. (B) Dorsal view of 10-h embryo showing in-phasexpression in mesoderm and neurectoderm and first segmental stripe (arrowheads). (C) Dorsal view of 12.5-h embryo; somite positionsenoted by arrows. (D) Dorsal view of 14-h embryo, showing four segmental stripes of neural expression adjacent to posterior halves ofomites 2, 3, 4, and 5 but not somite 1 (arrows). (E) Lateral view of 14-h embryo showing domain of neural expression with anterior andosterior boundaries (arrowheads), and strong expression in posterior mesoderm (arrow). (F) Lateral view of 18-h embryo showing domainf neural expression, epidermal expression, and reduced expression in posterior mesoderm. (G) Same as F, but double-stained formphiMLC-alk expression (red) to reveal somite positions. (H) Dorsal view of 18-h embryo stained as in G. Anterior to the left in all.
IG. 4. Spatial expression pattern of AmphiHox-3. (A) Oblique dorsal view of 10-h amphioxus embryo showing same anterior expressionoundary in mesoderm and neurectoderm (arrowhead). (B) Dorsal view of 12.5-h embryo; somite boundaries shown by arrows. (C) Dorsal


FIG. 5. Spatial expression pattern of AmphiHox-4. (A) Lateral and (B) dorsal views of 14-h amphioxus embryo showing weak expressionin posterior mesoderm and neurectoderm. (C) Lateral and (D) dorsal view of 18-h embryo showing neural expression. Note asymmetry inanterior boundary (arrowheads in D), lack of posterior boundary, and segmental modulation. (E) 18-h embryo double-stained forAmphiMLC-alk showing the characteristic out-of-phase position of left and right somites; the arrows mark the intersomitic cleft betweensomites 5 and 6. The anterior neural expression limit for AmphiHox-4 is opposite the posterior half of s6. Anterior to the left in A–D;anterior to the right in E. To facilitate comparison, the image in A is reversed horizontally.FIG. 6. Expression pattern of AmphiHox-2 (A) 10-h amphioxus embryo. (B) Dorsal view and (C) optical transverse section of 12.5-h
embryo, showing expression in somites (arrows). (D) Dorsal view of 14-h embryo. (E and F) 60-h larva showing expression in the preoral pit on the left side of the body (arrow). Anterior to the left in A, B, D, E, and F.
view of 14-h embryo showing segmental modulation of neural expression with the most anterior stripe adjacent to posterior somite 4;somites indicated by arrows. (D) Dorsal view of 18-h embryo showing segmental modulation of neural expression and persistent strongexpression in posterior mesoderm (arrow). (E) Lateral view of 18-h embryo, double-stained by AmphiMLC-alk expression (red), showing the
neural expression limit of AmphiHox-3 to be adjacent to the posterior half of s4. (F) Same as D, but double-stained for AmphiMLC-alkexpression. Anterior to the left in all.

abaodHammcfp

eAAAer

e8anAa

teespcttsda

pbmn(HtrAsiNotpF


homeobox by an intron. The hexapeptide is expected tointeract with the product of a pbx/exd gene (Mann andChan, 1996); this has not yet been cloned from amphioxus.Genomic analysis revealed no particular surprises, exceptfor the presence of an additional intron in AmphiHox-2.

Vertebrate Hox genes display a colinear relationshipbetween position within a gene cluster, anterior expressionlimit, and temporal profile of expression (Dolle et al., 1989;McGinnis and Krumlauf, 1992; Duboule, 1994). The rela-tion between chromosomal position and expression limit(spatial colinearity) is also displayed by AmphiHox genes,with the clear exception of AmphiHox-2. A differencebetween spatial colinearity in amphioxus and in vertebratesrelates to tissue specificity. In amphioxus, spatial colinear-ity is confined to the developing neural tube, while invertebrates it has been described in neural tube, neural crestderivatives, pharyngeal ectoderm, somitic mesoderm, andlateral plate mesoderm. Comparison to urochordates indi-cates clearly that the amphioxus condition is most likely tobe primitive for chordates and the vertebrate conditionderived. Urochordates (which include the ascidia) representthe outgroup to amphioxus plus vertebrates. Both ascidianHox genes that have been examined to date show spatialexpression only in the central nervous system (Katsuyamaet al.,1994; Gionti et al., 1998), as also described here foramphioxus. We argue, therefore, that spatial colinearity ofHox genes was confined to the neural tube in the ancestralchordates. Vertebrates elaborated on this condition, byextending spatial colinearity to other tissues. We suggestthis modification of Hox gene regulation was instrumentalin permitting the evolution of increased body plan complex-ity in vertebrates.

It is intriguing that in echinoderms, also members of theDeuterostomia, Hox genes do not have colinear expressionin the larval nervous system (Arenas-Mena et al., 1998).This may indicate that the hypothesized ancestral role forHox genes in patterning the nervous system does notextend beyond the chordates. Alternatively (or in addition),the echinoderm condition may be derived, particularlysince some other genes have modified expression patternsin echinoderm development (Lowe and Wray, 1997).

It is intriguing that it is the paralogy group 2 gene thatbreaks the spatial colinearity rule in amphioxus, while it isparalogy group 1 genes in mouse (Wilkinson et al., 1989;Hunt et al., 1991). These differences imply that neither themphioxus nor the mouse variant of spatial colinearity cane ancestral; both must have derived features. Deducing thencestral pattern for chordates requires information fromther taxa, including ascidians. Wada et al. (1998) recentlyemonstrated that the ascidian paralogy group 1 Hox gene,rHox1, has an anterior expression limit that precisely

buts the stripe of expression of Pax-2/5/8 (a putativearker of the midbrain–hindbrain junction). If this is theost anterior neural region that Hox genes can pattern in

hordates, then the primitive pattern of spatial colinearity
or chordates may have commenced at this point androceeded posteriorly. The amphioxus condition could have

volved from this state either by posterior displacement ofmphiHox-1 expression or by anterior displacement ofmphiHox-3, AmphiHox-4, etc., when the ancestralmphiHox-2 domain was “vacated.” Determination of the

xpression pattern of ascidian Hox-2 will be useful inesolving between these alternatives.

Temporal colinearity of amphioxus Hox genes is alsovident. Expression of AmphiHox-1 was first detected at–9 h, followed by AmphiHox-2 and AmphiHox-3 at 10 hnd AmphiHox-4 from 14 h. Considering just expression ineural tube, where spatial colinearity is evident,mphiHox-1 and AmphiHox-3 are both detected at 10 h

nd AmphiHox-4 at 14 h.

Escape from Colinearity

The expression pattern of AmphiHox-2 is different fromhat of the other AmphiHox genes examined. AmphiHox-2xpression does not seem to respect a precise boundary, andxpression is never detected in the neural tube at least intages examined here: 7 to 60 h of development. This isarticularly significant in light of the fact that spatialolinearity of AmphiHox genes is restricted to the neuralube. Thus, AmphiHox-2 has lost the anteroposterior pat-erning role typical for Hox genes. Similar cases are ob-erved in Drosophila in which zen and ftz are deduced to beerived from Hox genes that have lost their function innteroposterior patterning (reviewed by Averof et al., 1996).

These genes have other roles in Drosophila embryogenesis;ftz is a pair-rule segmentation genes, while zen has a role inatterning dorsally derived structures, including extraem-ryonic membranes. Similarly, although AmphiHox-2ust have lost the typical role of Hox genes, it has gained a

ew specific expression site in the developing preoral pitprecursor of Hatschek’s pit). In adult amphioxus,atschek’s pit is a secretory organ putatively homologous

o the adenohypophysis, thought to produce hormoneselated to gonadotropins (Nozaki and Gorbman, 1992).cquisition of a new role may have imposed the necessary

elective constraint to prevent AmphiHox-2 degeneratingnto a pseudogene after it had lost its typical Hox function.onetheless, it is noteworthy that the AmphiHox-2 home-

domain is more divergent from its vertebrate homologueshan is seen for any other amphioxus homeodomain fromaralogy group 1 to 10 (77% versus 80 to 97%; Garcia-ernandez and Holland, 1994).

Segmental Expression

Two aspects of amphioxus Hox gene expression could bedescribed as segmental. First, the anterior expression limitsof AmphiHox-1, AmphiHox-3, and AmphiHox-4 in theneural tube are spaced at regular two-somite intervals.These limits lie at the levels of mid-s2, mid-s4, and mid-s6,respectively (Fig. 7). Double-segment phasing is reminis-
cent of the two-rhombomere spacing for the anterior ex-pression limits of mouse Hoxb-2, Hoxb-3, Hoxb-4, and

mtse

o shA tion,


Hoxb-5 (Wilkinson et al., 1989). This similarity may be noore than coincidental, since there is currently no reason

o suppose that rhombomeric segmentation is related toomitic segmentation by a 1:1 correspondence. It is inter-sting, however, that several Drosophila homeotic genes

also show a two-metamere phasing of their anterior expres-sion boundaries in the ventral nerve cord (Holland, 1990;Hirth et al., 1998).

The second aspect of expression that is segmental relatesto the intensities of RNA detection. AmphiHox-1,

FIG. 7. Schematic diagram of 18-h or later amphioxus embrymphiHox-4. Note the spatial colinearity, intrasegmental modula

AmphiHox-3, and AmphiHox-4 each have a higher inten-sity of expression opposite the posterior half of each somite


(Fig. 7). This intrasegmental modulation of AmphiHox geneexpression in the neural plate commences early in develop-ment; indeed the first intense stripe of AmphiHox-1 expres-sion is detected at 10 h. This is earlier than the segmentalexpression of AmphiEn (the amphioxus homologue of en-grailed) in somites, reported by Holland et al. (1997).Modulation of expression level in the neural tube has alsobeen described in vertebrates, although the patterns haveless regularity. For example Hoxa-2 shows strong expres-sion in both rhombomere 3 and rhombomere 5, with

owing expression domains of AmphiHox-1, AmphiHox-3, andand weaker anterior expression (black shading).

weaker expression in other regions of the hindbrain (Huntet al., 1991). If we assume that this aspect of Hox gene


a(amwHd

D

D

D

F

G

G

H

H

H

H

H

H

H

H

K

K

K

K

K

L

M

M


regulation is homologous between amphioxus and verte-brates, it is simplest to conclude that the vertebrate condi-tion evolved by secondary reduction of more extreme intra-segmental modulation, the latter being retained byamphioxus. In this regard, it is noteworthy that the prod-ucts of several arthropod Hox genes also show intrasegmen-tal modulation in intensity (striping), including DrosophilabdA (Karch et al., 1990; Macias et al., 1990) and Ubx

Akam and Martinez-Arias, 1985; Struhl and White, 1985)nd Tribolium Abdominal (Shippy et al., 1998) and Ther-obius Antp (Peterson et al., 1989). It is unclear at presenthether intrasegmental modulation is an ancient feature ofox genes or a derived feature acquired independently inifferent taxa.

ACKNOWLEDGMENTS

H.W. held a postdoctoral fellowship from the Human FrontiersScience Program Organisation. J.G.F. acknowledges support fromGrant CICYT PB95-0579; P.W.H.H. acknowledges support fromthe BBSRC. Collaboration between P.W.H.H. and J.G.F. was facili-tated by a grant from the Acciones Integradas of the BritishCouncil/Ministerio de Educacion y Ciencia. We thank Jim Lange-land and Linda Holland for cDNA libraries.

REFERENCES

Akam, M. E., and Martinez-Aras, A. (1985). The distribution ofUltrabithrax transcripts in Drosophila embryos. EMBO J. 4,1689–1700.

Arenas-Mena, C., Martinez, P., Cameron, R. A., and Davidson,E. H. (1998). Expression of the Hox gene complex in the indirectdevelopment of a sea urchin. Proc. Natl. Acad. Sci. USA 95,13062–13067.

Averof, M., Dawes, R., and Ferrier, D. (1996). Diversification ofarthropod Hox genes as a paradigm for the evolution of genefunctions. Semin. Cell Dev. Biol. 7, 539–551.

Burglin, T. R., and Ruvkun, G. (1993). The Caenorhabditiselegans homeobox gene cluster. Curr. Opin. Genet. Dev. 3,615– 620.

Dawes, D., Dawson, I, Falciani, F., Tear, G., and Akam, M. (1994).Dax, a locust Hox gene related to fushi tarazu but showing nopair-rule expression. Development 120, 1561–1572.olle, P., Izpisua-Belmonte, J.-C., Falkenstein, H., Renucci, A., andDuboule, D. (1989). Coordinate expression of the murine Hox-5complex homeobox-containing genes during limb pattern forma-tion. Nature 342, 767–772.uboule, D. (1994). Temporal colinearity and the phylotypic pro-gression: A basis for stability of the vertebrate Bauplan and theevolution of morphologies through heterochrony. DevelopmentSuppl. 135–142.uboule, D., and Dolle, P. (1989). The structural and functionalorganization of the murine Hox gene family resembles that ofDrosophila. EMBO J. 8, 1497–1505.

alciani, F., Hausdorf, B., Schroder, R., Akam, M., Tautz, D.,
Denell, R., and Brown, S. (1996). Class 3 Hox genes in insects andthe origin of zen. Proc. Natl. Acad. Sci. USA 16, 8479–8484.

arcia-Fernandez, J., and Holland, P. W. H. (1994). Archetypalorganization of the amphioxus Hox gene cluster. Nature 370,563–566.ionti, M., Ristoratore, F., Di Gregorio, A., Aniello, F., Branno, M.,and Di Lauro, R. (1998). Cihox5, a new Ciona intestinalisHox-related gene is involved in regionalization of the spinal cord.Dev. Genes Evol. 207, 515–523.irth, F., Hartmann, B., and Reichert, H. (1998). Homeotic geneaction in embryonic brain development of Drosophila. Develop-ment 125, 1579–1589.olland, L. Z., and Holland, N. D. (1996). Expression ofAmphihHox-1 and AmphiPax-1 in amphioxus embryos treatedwith retinoic acid—Insights into evolution and patterning ofthe chordate nerve cord and pharynx. Development 122,1829 –1838.olland, L. Z., Pace, D. A., Blink, M. L., Kene, M., and Holland,N. D. (1995). Sequence and expression of amphioxus alkalimyosin light-chain (AmphiMLC-alk) throughout development—Implications for vertebrate myogenesis. Dev. Biol. 171, 665–676.olland, L. Z., Kene, M., Williams, N. A., and Holland, N. D.(1997). Sequence and embryonic expression of the amphioxusengrailed gene (AmphiEn): The metameric pattern of transcrip-tion resembles that of its segment polarity homolog in Drosoph-ila. Development 124, 1723–1732.olland, P. W. H. (1990). Homeobox genes and segmentation:Co-option, co-evolution, and convergence. Semin. Dev. Biol. 1,135–145.olland, P. W. H., and Garcia-Fernandez, J. (1996). Hox genes andchordate evolution. Dev. Biol. 173, 382–395.olland, P. W. H., Holland, L. Z., Williams, N. A., and Holland,N. D. (1992). An amphioxus homeobox gene: Sequence conser-vation, spatial expression during development and insights intovertebrate evolution. Development 116, 653–661.unt, P., Gulisano, M., Cook, M., Sham, M. H., Faiella, A.,Wilkinson, D., Boncinelli, E., and Krumlauf, R. (1991). A distinctHox code for the branchial region of the vertebrate head. Nature353, 861–864.arch, F., Bender, W., and Weiffenbach, B. (1990). abdA expressionin Drosophila embryos. Genes Dev. 4, 1573–1587.

atsuyama, Y., Wada, S., Yasugi, S., and Saiga, H. (1995). Expres-sion of the labial group gene HrHox-1 and its alteration byretinoic acid in development of the ascidian Halocynthia roretzi.Development 121, 3197–3205.

enyon, C., and Wang, B. (1991). A cluster of Antennapedia classhomeobox genes in a nonsegmented animal. Science 253, 516–517.mita-Cunisse, M., Loosli, F., Bierne, J., and Gehring, W. J. (1998).Homeobox genes in the ribbonworm Lineus sanguineus: Evolu-tionary implications. Proc. Natl. Acad. Sci. USA 95, 3030–3035.rumlauf, R. (1994). Hox genes in vertebrate development. Cell 78,191–201.

owe, C. J., and Wray, G. A. (1997). Radical alterations in the rolesof homeobox genes during echinoderm evolution. Nature 389,718–721.acias, A., Casanova, J., and Morata, G. (1990). Expression andregulation of the abd-A gene of Drosophila. Development 110,1197–1207.ann, R. S., and Chan, S.-K. (1996). Extra specificity from extra-
denticle: The partnership between Hox and PBX/EXD homeodo-main proteins. Trends Genet. 12, 258–262.

M

M

N

P


Martinez, P., Rast, J. P., Arenas-Mena, C., and Davidson, E. H.(1999). Organization of an echinoderm Hox gene cluster. Proc.Natl. Acad. Sci. USA 96, 1469–1474.cGinnis, W., and Krumlauf, R. (1992). Homeobox genes and axialpatterning. Cell 68, 283–302.orrison, A., Ariza-McNaughton, L., Gould, A., Featherstone, M.,and Krumlauf, R. (1997). Hoxd-4 and regulation of the group 4paralog genes. Development 124, 3135–3146.ozaki, M., and Gorbman, A. (1992). The question of functionalhomology of Hatcheck’s pit of amphioxus (Branchiostomabelcheri) and the vertebrate adenohypophysis. Zool. Sci. 9, 387–395.

etersen, M. D., Rogers, B. T., Popadic, A., and Kaufman, T. C.(1999). The embryonic expression pattern of labial, posteriorhomeotic complex genes and the teashirt homologue in anapterygote insect. Dev. Genes Evol. 209, 77–90.

Popodi, E., Kissinger, J. C., Andrews, M. E., and Raff, R. A. (1996).Sea urchin Hox genes—Insights into the ancestral Hox cluster.Mol. Biol. Evol. 13, 1078–1086.

Shippy, T. D., Brown, S. J., and Denell, R. E. (1998). Molecularcharacterization of the Tribolium abdominal-A ortholog andimplications for the products of the Drosophila gene. Dev. GenesEvol. 207, 446–452.

Struhl, G., and White, R. A. H. (1985). Regulation of the Ultra-
bithorax gene of Drosophila by other Bithorax complex genes.Cell 43, 507–519.

Von Allmen, G., Hogga, I., Spierer, A., Karch, F., Bender, W.,Gyurkovics, H., and Lewis, E. (1996). Splits in fruitfly Hox genecomplexes. Nature 380, 116.

Wada, H., and Satoh, N. (1994). Details of the evolutionary historyfrom invertebrates, as deduced from the sequences of 18S rDNA.Proc. Natl. Acad. Sci. USA 91, 1801–1804.

Wada, H., Saiga, H., Satoh, N., and Holland, P. W. H. (1998).Tripartite organisation of the ancestral chordate brain and theantiquity of placodes: Insights from ascidian Pax-2/5/8/, Hox andOtx genes. Development 125, 1113–1122.

Wilkinson, D. G., Bhatt, S., Cook, M., Boncinelli, E., and Krumlauf,R. (1989). Segmental expression of Hox 2 homeobox genes in thedeveloping hindbrain. Nature 341, 405–409.

Yasuo, H., and Satoh, N. (1994). An ascidian homolog of themouse Brachyury (T) gene is expressed exclusively in noto-chord cells at the fate restricted stage. Dev. Growth Differ. 36,9 –18.

Zhang, F., Popperl, H., Morrison, A., Kovacs, E. N., Prideaux, V.,Schwarz, L., Krumlauf, R., Rossant, J., and Featherstone, M. S.(1997). Elements both 59 and 39 to the murine Hoxd-4 geneestablish anterior borders of expression in mesoderm and neurec-toderm. Mech. Dev. 67, 49–58.

Received for publication May 22, 1999Accepted May 24, 1999


Colinear and Segmental Expression of Amphioxus Hox Genes

Documents

Transcript of Colinear and Segmental Expression of Amphioxus Hox Genes