expression of estrogen-receptor related receptors in ... · have been identiﬁed (holland and...

Expression of estrogen-receptor related receptors in amphioxus and

zebrafish: implications for the evolution of posterior brain segmentation

at the invertebrate-to-vertebrate transition

Pierre-Luc Bardet,a,1 Michael Schubert,a Beatrice Horard,a Linda Z. Holland,b Vincent Laudet,a,�

Nicholas D. Holland,b and Jean-Marc Vanackera,2,�

aLaboratoire de Biologie Moleculaire de la Cellule, CNRS UMR 5161, IFR128 BioSciences Lyon-Gerland, Ecole Normale

Superieure de Lyon, 46 Allee d’Italie, 69007 Lyon, FrancebMarine Biology Research Division, Scripps Institution of Oceanography, La Jolla, CA 92093-0202, USA�Authors for correspondence (emails: [email protected] and [email protected])

1 Present address: NIMR, Mammalian development, The Ridgeway-Mill Hill, London NW7 1AA, UK.

2Present address: Laboratoire Genotypes et Phenotypes Tumoraux, INSERM EMI229, CRLC Val d’Aurelle-Paul Lamarque, 34298 Montpellier cedex 5, France.

SUMMARY The evolutionary origin of vertebrate hindbrainsegmentation is unclear since the amphioxus, the closestliving invertebrate relative to the vertebrates, possesses ahindbrain homolog that displays no gross morphologicalsegmentation. Three of the estrogen-receptor related (ERR)receptors are segmentally expressed in the zebrafish hindbrain,suggesting that their common ancestor was expressed in asimilar, reiterated manner. We have also cloned and deter-mined the developmental expression of the single homolog ofthe vertebrate ERR genes in the amphioxus (AmphiERR). Thisgene is also expressed in a segmented manner in a regionconsidered homologous to the vertebrate hindbrain. In contrastto the expression of amphioxus islet (a LIM-homeobox genethat also labels motoneurons), AmphiERR expression persistslonger in the hindbrain homolog and does not later extend

to additional posterior cells. In addition, AmphiERR and oneof its vertebrate homologs (ERRa) are expressed in thedeveloping somitic musculature of amphioxus and zebrafish,respectively. Altogether, our results are consistent with finestructural evidence suggesting that the amphioxus hindbrainis segmented, and indicate that chordate ERR gene expressionis a marker for both hindbrain and muscle segmentation.Furthermore, our data support an evolution model of chordatebrain segmentation: originally, the program for anteriorsegmentation in the protochordate ancestors of the vertebr-ates resided in the developing axial mesoderm which imposedreiterated patterning on the adjacent neural tube; during earlyvertebrate evolution, this segmentation program was trans-ferred to and controlled by the neural tube.

INTRODUCTION

Segmentation is the process that subdivides tissues into re-

peated, identical subunits that will later acquire particular

characteristics according to their position. In adult verte-

brates, this phenomenon is particularly obvious in the spinal

cord where vertebrae are iterated elements sharing identical

features, such as the presence of dorsal sensory- and ventral

motor roots although the territories they serve vary along the

axis. This importance of segmentation is also reflected by the

complexity of the vertebrate head, the formation of which

depends on brain segmentation. In addition to cryptic seg-

mentation in the diencephalon, the hindbrain displays an

overt segmentation, which is required for proper craniofacial

morphogenesis (Trainor and Krumlauf 2000). Hindbrain is

transiently subdivided into repeated elements called rhombo-

meres. Neural crest cells emerging from the hindbrain are

specified according to their antero-posterior position and

migrate through the whole cephalic region to help pattern

definite territories. In addition to this morphological segmen-

tation, each rhombomere displays a specific gene expression

code, as exemplified by Hox genes (Prince et al. 1998).

The evolution of the vertebrate complex head, as a dis-

tinctive trait of their phylum, has been extensively discussed

(Guthrie 1995; Holland 2000; Holland and Chen 2001). Re-

garding this issue, the cephalochordate amphioxus (Branch-

iostoma floridae), being the closest living invertebrate relative

to the vertebrates, represents a powerful tool. The general

vertebrate brain organization is indeed conserved in the amp-

hioxus since structures homologous to fore- and hindbrain

have been identified (Holland and Holland 1999), but simpler

than in the vertebrates, as the amphioxus hindbrain homolog

EVOLUTION & DEVELOPMENT 7:3, 223–233 (2005)

& BLACKWELL PUBLISHING, INC. 223

is not segmented at the gross anatomical level and does not

give rise to neural crest cells. However, amphioxus genes have

been cloned that display an iterated expression pattern in the

hindbrain. For instance, the LIM homeobox gene islet is ex-

pressed in a repeated manner throughout the posterior part of

the amphioxus nervous system (Jackman et al. 2000; Jackman

and Kimmel 2002). This suggests that even though morpho-

logical segmentation is not obvious, molecular segmentation

already exists in the amphioxus. However, expression do-

mains for amphioxus genes previously proposed as hindbrain

markers (Jackman and Kimmel 2002), although suggestive of

a hindbrain homology with vertebrates, are transiently ex-

pressed and their expression pattern is not exactly homolo-

gous to those of their vertebrate orthologs. For instance,

whereas amphioxus islet has been reported to be only ex-

pressed in the hindbrain homolog, its vertebrate islet 1 coun-

terpart not only labels the hindbrain but is also expressed in

the whole spinal cord (Appel et al. 1995). Furthermore, the

developmental mechanisms directing amphioxus and verte-

brate hindbrain segmentation could be different. Indeed, it

has been proposed that in amphioxus signals originating from

the segmented head mesoderm trigger neural segmentation

(Mazet and Shimeld 2002). In contrast, the vertebrate hind-

brain segmentation program is initiated entirely within the

neural tube.

Nuclear receptors form a family of ligand-dependent tran-

scription factors implicated in a variety of developmental and

physiological processes (Laudet and Gronemeyer 2002).

These receptors can be recognized through their typical do-

main organization, with a highly conserved DNA-binding

domain localized in the middle of the molecule and a ligand

binding, transactivation domain in the C-terminus. However,

for some members of the family, referred to as ‘‘orphan re-

ceptors’’, no ligand has been identified to date (Giguere 1999).

This is the case for estrogen-receptor related (ERR) a, b, andg receptors (NR3B1, 2 and 3, respectively) which may activate

transcription in a constitutive manner (Xie et al. 1999), al-

though the existence of an unidentified ligand has also been

suggested (Vanacker et al. 1999a; reviewed in Horard and

Vanacker 2003). While ERR receptors have been suggested to

positively or negatively interfere with the signaling driven by

estrogens (Vanacker et al. 1999b; Lu et al. 2001; Kraus et al.

2002; reviewed in Giguere 2002), little is known of their de-

velopmental roles. During mouse embryonic development

ERRa is expressed in a variety of tissues (Bonnelye et al.

1997) whereas ERRb expression is restricted to the developing

placenta (Pettersson et al. 1996) to the formation of which it

contributes (Luo et al. 1997). Besides its complex pattern of

expression in the brain (Hermans-Borgmeyer et al. 2000), the

territories in which ERRg is expressed during mouse devel-

opment have not been thoroughly characterized.

To gain insight into the developmental roles of the ERR

genes, we have cloned their zebrafish orthologs and deter-

mined their expression pattern during development (Bardet

et al. 2004). Among other expression sites, we found that three

zebrafish ERR receptors are expressed in the hindbrain in

an iterated manner. Furthermore zebrafish ERRa is also ex-

pressed in a segmented manner in the somites. We next cloned

the single ERR ortholog in the amphioxus (AmphiERR) and

determined its expression pattern. AmphiERR is expressed in

the hindbrain homolog where it most likely labels dorsal

compartment (DC) motoneurons, and also in somitic meso-

derm. Strikingly AmphiERR is expressed in a segmented

manner in both tissues. Our results show that a structure

similar to the segmented hindbrain was already present before

the separation between invertebrates and vertebrates. Our

data also support a model in which neural segmentation in

invertebrate chordates requires signaling originating from the

segmented mesoderm. During early vertebrate evolution this

signaling was evidently transferred to the hindbrain and the

neural tube.

MATERIALS AND METHODS

Molecular cloningOne microgram of total RNA extracted from pooled amphioxus

larvae (24-h post fertilization) was retrotranscribed using MMLV

reverse transcriptase (Gibco BRL, Cergy, France) and submitted to

nested polymerase chain reaction (PCR) using degenerated primers

located in the conserved DNA binding and ligand-binding do-

mains. PCR products were cloned into pCR2.1 vector (Invitrogen,

Cergy, France) and individual clones were sequenced. B. floridae (2

to 4 days of development) cDNA library was screened using

standard methods. RACE experiments were performed using the

50/30 RACE kit (Roche, Meylan, France) following the manufac-

turer’s instructions. Cloning of zebrafish ERRs is described else-

where (Bardet et al. 2004).

Sequences of the degenerated oligonucleotides used:

50: GA(T/C)TA(T/C)GCITCIGGITA(T/C)CA

50 nested: TT(T/C)AA(G/A)AG(G/A)AG(T/C)ATICAAGG

30: GIIGTIG(T/C)CA(C/G)IA(G/A)CATGTC

30nested:CAITTICC(T/C)T(G/C)(A/G)(T/C)(T/C)CCTGTCCA

GenBank accession number for ERRs:

AmphiERR: zebrafish ERRa: AY556395; zebrafish ERRb:AY556396; zebrafish ERRg: AY556397.

Phylogenetic analysisProtein sequences were aligned automatically by Clustalw (Thom-

son et al. 1994) with manual correction in Seaview (Galtier et al.

1996). Phylogenetic reconstruction was done using amino-acid

alignments. Only complete sites (no gap) were used. Tree was con-

structed by Neighbor-Joining (Saitou and Nei 1987) with distances

corrected for rate heterogeneity between sites, using a gamma law

of parameter alpha estimated in Tree-Puzzle (Schmidt et al. 2002),

with eight categories. Reliability of nodes was estimated by 1000

bootstrap replicates (Felsenstein 1985).

224 EVOLUTION & DEVELOPMENT Vol. 7, No. 3, May^June 2005

Expression analysisIn situ hybridizations on amphioxus embryos were performed ac-

cording to Holland et al. (1996). Anti-sense probes were synthe-

sized from complete cDNA cloned in pCS21 plasmid using T7

polymerase after linearization (DIG RNA labeling kit, Roche). In

situ hybridizations of zebrafish embryos were performed according

to Thisse et al. (1993). All zebrafish probes were synthesized from

partial cDNA (around 1kb from the DNA- up to the ligand-

binding domains thus including the highly divergent hinge domain)

cloned in pBluescript, using the same DIG RNA labeling kit (T3

polymerase for ERRa, and ERRg, T7 polymerase for ERRb. Afterphotography of the whole mounts, specimens were counterstained

in Ponceau S, embedded in Spurr’s resin, and sectioned (Holland

et al. 1996).

RESULTS

ERR genes are expressed in a segmented mannerin zebrafish hindbrain

We have previously reported the cloning and expression of

five ERR species in the zebrafish (Bardet et al. 2004). The

present report will mainly extend the data concerning the

segmented expression in the hindbrain of the three zebrafish

ERRs (a, b, and g) genes that are expressed in the hindbrain.

There were differences as well as striking similarities among

the hindbrain expression domains of the three related genes

(Fig. 1). At 36 hpf, ERRa was detected in the hindbrain

region in two patches of cells in r5 and 6 (Fig. 1A; and not in

r4 as mistakingly stated in Bardet et al. 2004). Transverse

section revealed that ERRa was expressed in a ventro-lateral

position (Fig. 1B). ERRb also displayed a segmented expres-

sion in the hindbrain at 36 hpf (Fig. 1C). As for ERRa,ventrolateral positions of the labeling was also revealed by

cross-sections (Fig. 1, D and E). More laterally, ERRb was

also detected in structures whose position corresponds to the

trigeminal (tg), the anterior and posterior lateral line ganglia

(allg and pllg, respectively; Fig. 1, C and D; Andermann et al.

2002). At 36 hpf ERRg was detected in all three posterior

cranial ganglia (tg, allg and pllg; Fig. 1, F and G) and dis-

played a segmented expression in the rhombomeres in ventro-

lateral position (Fig. 1, F–H). The rhombomeric expression of

ERR genes appeared in a sequential order (data not shown)

as summarized in Fig. 1I. In addition, cohybridization indi-

cated that at least a subset of rhombomeric cells co-expressed

ERRb and ERRg at 24 hpf and 36 hpf (data not shown). To

confirm the segmented nature of ERR genes expression, we

used valentinomutant, where the r5 and r6 fail to individualize

(giving rise to an undetermined rX rhombomere; Moens

et al. 1996). In this context, ERRg expression is lacking

(or at least deeply attenuated) at the level of the affected

rhombomere at 24 hpf (compare Fig 1, J and K). Thus ERRs

are robust markers of the hindbrain segmentation in the

zebrafish.

Identification of a unique ERR receptor inamphioxus

To analyze the evolutionary conservation of the segmented

expression of ERRs, we used the amphioxus (Branchiostoma

floridae), a cephalochordate that displays no overt hindbrain

segmentation at the gross anatomical level. Furthermore, a

unique ERR gene may be predicted in this species given that

(i) its genome is not duplicated (Panopoulou et al. 2003) and

(ii) unique copies of ERR were found in other invertebrates

such as C. intestinalis (Dehal et al. 2002) and D. melanogaster

(Ostberg et al. 2003). We first isolated an ERR cDNA frag-

ment by RT-PCR using total RNA extracted from amp-

hioxus larvae (24h post fertilization) and degenerate primers

located in the two conserved regions of nuclear receptors: the

DNA binding domain (DBD) and the ligand binding domain

(LBD). This experiment gave rise to a fragment that was used

to screen an amphioxus cDNA library. Partial cDNAs were

obtained and 50 RACE was then performed to produce com-

plete cDNAs that were subsequently sequenced. A single

ORF in the cDNA was predicted to produce a putative pro-

tein of 455 amino acids, hereafter referred to as AmphiERR

(NR3B4). This sequence was aligned with zebrafish ERRs

(Fig. 2A). AmphiERR bears the typical DNA binding do-

main of nuclear receptors organized in two zinc finger mod-

ules with cysteines in fixed positions and a putative ligand-

binding domain located in the C-terminal end of the protein

with 12 conserved a-helices (Laudet and Gronemeyer 2002).

As expected, there is a high level of sequence identity within

these domains between species (over 90% for the DBD, close

to 60% for the LBD when comparing amphioxus to zebrafish

sequences).

Using the Neighbor-Joining method, we then analyzed the

phylogenetic relationships of AmphiERR to others ERRs

(Fig. 2B). Low bootstrap values were obtained for the posi-

tion of zebrafish ERRd, probably because of a high mutation

rate. AmphiERR sequence branched at the root of vertebrate

(human and zebrafish) ERRs and is most likely the unique

amphioxus homologue of the ancestral ERR.

Expression of ERR during amphioxusdevelopment

We determined the expression pattern of AmphiERR using in

situ hybridization of amphioxus embryos (Fig. 3). Expression

was first detected at 14h of development (neurula stage) in

somites 2 to 6 (Fig. 3, A and B) and then extended to newly

formed posterior somites (Fig. 3, E–G). Transverse sections

indicate that AmphiERR is expressed in the dorsal part of the

muscular masses of the somites (Fig. 3, C and D). These will

later give rise to superficially-located slow muscle types (re-

viewed in Lacalli 2001), a population that is labelled by Am-

phiERR until at least 3 days of development. In contrast,

AmphiERR expression was always excluded from somite 1,

ERR genes as markers of posterior brain segmentation 225Bardet et al.

Fig. 1. Expression of estrogen-receptor related (ERR) genes in the zebrafish hindbrain. (A and B): ERRa (A) Dorsal view of the head(anterior to the left) of 36 hpf embryos, showing ERRa expression in the diencephalon (arrow) and in cell patches in rhombomeres (r) 5 and6 (arrowheads). Double arrowhead: muscle expression in the trunk. Position of the otic vesicle (ot.) is indicated. (B) Cross section throughlevel x in (A), showing ventro-lateral ERRa expression corresponding to motoneurons (arrowhead). (C–E): ERRb (C) Dorsal view of thehead of 36 hpf embryos (anterior to the left) showing ERRb expression in the diencephalon (arrow), in cell patches in all rhombomeres.ERRb is also expressed in trigeminal ganglia (double open arrowhead), anterior and posterior lateral line ganglia (open arrowhead andopen arrow, respectively). (D) Cross section through level x in (C) showing ERRb expression in motoneurons (arrowhead) and anteriorlateral line ganglia (open arrowhead). (E) Cross section through level y in (C) showing ERRb expression in motoneurons (arrowhead).(F–H): ERRg (F) Dorsal view of the head of 36 hpf embryos (anterior to the left) showing ERRg expression in the epiphysis (asterisk), thediencephalon (arrow), in cell patches in the rhombomeres. ERRg is also expressed in the trigeminal ganglia (open double arrowhead), in theanterior and posterior lateral line ganglia (open arrowhead and open arrow, respectively). Position of the ot. is indicated. (G) Cross sectionthrough level x in (F) showing ERRg expression in motoneurons (black arrowhead) and in anterior lateral line ganglia (open arrowhead).(H) Cross section through level y in (F) showing ERRg expression in motoneurons (arrowhead). (I) Summary of the spatio-temporalexpression of ERR genes. Hatched bars indicate weaker gene expression. Position and numbering of the rhombomeres, as well as of the ot.,are schematized on the left. (J and K) Loss of ERRg expression in valentino mutants. Dorsal views of 24 hpf wild type (J) or val mutant (K)embryos (anterior to the left) hybridized with an ERRg probe. Rhombomeres are numbered, rX corresponds to the undetermined r5–r6equivalent in val. ERRg expression is absent in rX in val mutant as compared with r5 and r6 in wild type (wt) animal.


which only contains deep, fast fibres (Fig. 3, A and B). In

addition to somites, individual mesodermal and epidermal

cells also began expressing AmphiERR, starting at 24h of

development (Fig. 3, K–M). By three days of development,

expression in these cells tended to disappear or at least to be

strongly reduced (Fig. 3Q). At these later stages, AmphiERR

was also weakly transcribed in pharyngeal endoderm cells

(Fig. 3Q).

We also detected AmphiERR in neural derivatives, such as

the presumptive frontal eye and the presumptive photoreceptor

Fig. 2. A unique estrogen-receptorrelated (ERR) gene in the amp-hioxus. (A) Alignment of ERRsequences. Zebrafish and amp-hioxus sequences were aligned us-ing Clustalw with Seaview manualcorrections. A dot represents iden-tical amino acid whereas a dashsymbolizes a gap, using Amphi-ERR sequence as a reference.Conserved Cys residues of theDNA binding domains are boxed.The 12 a-helices of the putativeligand-binding domain are underbrackets. (B) Phylogenetic tree ofERR proteins. Tree was built withDanio rerio (dr� ), Homo sapiens(hs� ) and AmphiERR usingNeighbor-Joining method, using305 sites with global gap removal.One thousand bootstrap replicateswere performed. Bootstrap valuesare indicated. Drosophila melano-gaster (dm� ) and Ciona intestinalis(ci� ) ERR sequences were used asoutgroups.


Fig. 3. AmphiERR expression demonstrated by in situ hybridization of developing Branchiostoma floridae. All side and dorsal views withanterior toward left; all cross sections viewed from tail end of animal. Scale lines for whole mounts550mm; scale lines for cross and frontalsections (counterstained pink)525mm. (A) Side view of 14-h neurula showing strong expression in the second through sixth somite oneither side and weak expression in some cells of the presumptive frontal eye (arrow). (B) Dorsal view of (A) showing expression in frontaleye cells (single arrow), presumptive photoreceptor cells of Hess (arrowhead) and muscular somites, excluding the first (tandem arrow). (C)Cross section through level x in (B) showing strong expression in the muscle cells of the somites, but no detectable transcripts in the neuralplate (arrow). (D) Cross section through level y in (B) showing strong expression in the photoreceptor cells of Hess in the neural plate andweaker expression in somitic muscles. (E) Side view of 19-h neurula showing conspicuous expression in the muscular somites. (F) Dorsalview of (E) showing expression in the somites and in the photoreceptor cells of Hess (arrowhead) as well as in some neurons (one of which isindicated by the arrow). (G) Frontal section through (F) showing expression in somites and, more medially, in neurons. (H) Cross sectionthrough level x in (G) showing expression in neurons (arrow). (I) Cross section through level y in (G) showing expression in neurons (arrow)and somitic muscle cells (arrowhead). (J) Cross section through level z in (G) showing expression in somitic muscle cells (arrowhead). (K)Twenty-four-hour embryo with focus on upper surface showing transcripts in individual epidermal/mesodermal cells. (L) Twenty-four-hourembryo with focus on lower surface showing transcripts in individual epidermal/mesodermal cells. (M) Cross section through level x in (K)and (L) showing expression in individual mesodermal cell (arrow) and individual epidermal cell (arrowhead). (N) Dorsal view of 24-hembryo showing transcripts in somitic muscle cells and neurons. (O) Frontal section through (N) showing transcripts in muscle cells andneurons; pigment cup cells have appeared medial to the photoreceptor cells of Hess (arrow and enlarged in inset). (P) Cross section throughlevel x in (O) showing expression in somitic muscle cells and in photoreceptor cells of Hess; the arrow indicates the pigment cup cells. (Q)Side view of a 3-day larva with weak expression in somitic muscle cells and endodermal cells of the pharynx and strong expression inneurons. The inset is an enlarged dorsal view of the expressing neurons; the most anterior (arrowhead) are at the posterior end of thecerebral vesicle, and the most posterior are the photoreceptor cells of Hess (arrow). (R) Cross section through the level of the arrowhead in(Q) showing expression in several neurons located ventrally in the neural tube.


of Hesse (Fig. 3B) and other neuronal derivatives starting at

19h of development. Indeed, AmphiERR was expressed in

paired cells from the border between somites 1 and 2 and up

to somite 4 (Fig. 3, F–H). At 24 hpf, this expression was

expanded to six pairs of cells from the level of somites 2 to 5

(Fig. 3, N and O). This clearly segmented expression persisted

until at least 3 days of development with an additional ex-

pression domain in the posterior part of the cerebral vesicle

(Fig. 3Q). Based on their wedge-shaped morphology and on

their ventro-lateral position within the nerve cord (Fig. 3I),

the paired cells expressing AmphiERR are likely to represent

DC motoneurons that innervate superficial muscular fibers

(Lacalli and Kelly 1999).

As pointed out above and in contrast to vertebrates, the

amphioxus hindbrain homolog is not segmented at the gross

anatomical level (Holland and Holland 1998). However, the

fine structural data (Lacalli and Kelly 1999) and the iterated

expression of AmphiERR suggests that the hindbrain homo-

log is fundamentally segmented. It has been previously sug-

gested that the amphioxus islet gene could be a similar marker

for hindbrain segmentation since it is expressed as a series of

seven paired spots spread throughout the nerve cord (Jack-

man et al. 2000; Jackman and Kimmel 2002). Given that

AmphiERR also presented this type of pattern, we compared

the expression domains of both genes (Fig. 4). As expected,

islet was expressed at 20 hpf in the nerve cord in a segmented

manner. Sections of embryos at this stage of development

revealed a heterogeneous expression of this gene within the

nerve cord. Indeed, in some of the sections, ventro-lateral

expression was observed (Fig. 4B) which most likely corre-

sponded to DC motoneurons, while on others islet was ex-

pressed in laterally positioned cells (Fig. 4C). These cells could

correspond to sensory neurons by analogy with zebrafish islet

1 and islet 2 that label Rohon-Beard (sensory) neurons in the

dorsal neural tube (Appel et al. 1995). Notably, islet expres-

sion was also observed in the floor plate. Double in situ hy-

bridizations (Fig. 4, D–F) revealed that islet and AmphiERR

share common expression regions in the anterior half of

the embryo which corresponds to the amphioxus hindbrain

homolog (Holland and Holland 1998). However, the

sequentially repeated expression of islet corresponded only

partially to that of AmphiERR and does not always occur in

DC motoneurons. Furthermore, islet was also detected more

posteriorly suggesting that this gene may be involved in seg-

mentation of the whole nerve cord. In contrast, AmphiERR

expression was limited to motoneurons of the hindbrain ho-

molog. Altogether our results allow us to propose ERR genes

as conserved markers of posterior brain segmentation.

Fig. 4. Comparison of AmphiERR and islet expression in the amphioxus hindbrain. (A–C) Islet expression. (A) Dorsal view of a 20-hembryo showing segmented islet expression. (B) Cross section through level w in (A) showing islet expression in the endoderm and in thenervous system. Islet is expressed in the floor plate (arrow) and in ventro-lateral position (arrowhead). (C) Cross section through level x in(A) showing islet expression in the endoderm and in the nervous system. At this level, islet is again expressed in the floor plate (arrow) andalso in medio-lateral position (arrowhead), indicating heterogeneity between cross-sections. (D–F) AmphiERR and islet co-hybridization.(D) Dorsal view of a 24-h embryo showing AmphiERR expression in somites and in the anterior neural system (compare with Fig. 3N) in asegmented manner. Islet expression extends more posteriorly (arrow). (E) Cross section through level y in (D). Neural labeling in ventro-lateral position might be attributed to both AmphiERR and islet, whereas medio-lateral signal (arrowhead) originate from islet alone(compare with Fig. 3I and Fig. 4, B and C). Note AmphiERR expression in the muscle masses (double arrowhead). (F) Cross sectionthrough level z in (D) showing islet expression in the floor plate (arrow). AmphiERR is only expressed in muscular masses but not in theneural process.


DISCUSSION

Hindbrain segmentation is conserved betweencephalochordates and vertebrates

The hindbrain is a typical vertebrate structure transiently

characterized during development by a repetition of subunits

called rhombomeres. Individual rhombomeres can be identi-

fied morphologically through their clear boundaries and also

through expression of a given combination of molecular

markers. For instance vertebrate Krox20 is only expressed in

rhombomere 3 (r3) and r5 (Wilkinson et al. 1989). Within

rhombomeres, individual cranial nerves will differentiate and

innervate distinct muscular territories. The evolutionary or-

igin of the segmented hindbrain of vertebrates is unclear since

no gross morphological segmentation can be observed in

cephalochordates, the sister group of vertebrates.

In the amphioxus we have found ERR labeling in neu-

roectodermal structures, the frontal eye, the two initial pho-

toreceptor cells of Hesse, and six pairs of additional cells at

the level of somite 2–6. These cells have the typical wedge-

shape of DC somatic motoneurons with axons extending

posteriorly (Lacalli and Kelly 1999). This view is further sup-

ported by the fact that no such labeling was observed at the

level of somite 1, whose anterior neural territory does not

contain DC motoneurons. Furthermore, the distance between

the cells expressing AmphiERR is compatible with the inter-

vals predicted to exist between DC motoneurons (Lacalli

2002). Nevertheless only five pairs of DC motoneurons have

been reported (Lacalli and Kelly 1999). However, these cells

were identified by tracing back axons on serial sections of the

anterior nerve cord. The sampling of the sections used, to-

gether with the asymmetry that can be observed within a pair

of neurons renders it possible that an additional DC moto-

neuron pair might have been missed (T. Lacalli, personal

communication). It is thus very likely that AmphiERR is an

exclusive marker of all DC motoneurons.

A number of genes, such as AmphiNk2-1 (Venkatesh et al.

1999), AmphiFoxB (Mazet and Shimeld 2002), AmphiKrox,

and AmphiShox (Jackman and Kimmel 2002), have been

identified whose expression demonstrated segmentation of the

amphioxus neural tube, including the hindbrain homolog.

Vertebrate hindbrain segmentation can also be observed

through the LIM-homeodomain transcription factor Islet1

that patterns cranial nerves (reviewed in Hobert andWestphal

2000). An amphioxus islet gene has been cloned that is ex-

pressed as an iterated series of cells identified as motoneurons

(Jackman et al. 2000; Jackman and Kimmel 2002).

However, several differences can be noted between the ex-

pression of AmphiERR and islet. From a temporal point of

view, islet expression appears much earlier (at 12h when so-

mite 4 is formed) than that of ERR (whose first neuronal

labeling appears at 19h). The latter distinctly persists at least

up to 3 days of development whereas islet expression has

faded at 30h. The difference in regional expression between

the two genes is even more striking. The caudal-most site

of AmphiERR expression (the photoreceptor of Hesse)

is at the level of the boundary between somites 4 and 5,

whereas islet expression extends back to the boundary be-

tween somites 6 and 7. Islet has been suggested to label the

whole amphioxus hindbrain homolog. However, it could be

argued that the posterior boundary of the amphioxus hind-

brain is not clearly defined because of the lack of markers.

For instance, whereas the anterior expression limit of verte-

brate Hox5 paralogous genes abuts on the boundary between

the hindbrain and the spinal cord, AmphiHox5 expression

pattern has not been determined to date. In vertebrates islet

genes are expressed not only in the hindbrain, but also more

caudally in the neural tube (Appel et al. 1995). By analogy,

the possibility exists that, in amphioxus as well, islet labels

posterior neural structures and not only the hindbrain ho-

molog. In contrast, expression of ERR genes in the zebrafish

is posteriorly limited to the hindbrain in which each rhombo-

mere expresses a combination of the ERR genes. Alterna-

tively it is possible that islet labels the whole amphioxus

hindbrain homolog and that the posterior domain of expres-

sion is a vertebrate innovation. In this case, AmphiERR

would mark a subdomain of the hindbrain, displaying a more

refined segmentation pattern, in which DC motoneurons

are found.

Whatever the case, it is worth noting that all markers

found to date are transiently expressed and define embryonic

fields in which functional cells will differentiate. In contrast,

AmphiERR is expressed during a longer period of time and

defines functional structures (likely DC motoneurons). The

fact that islet is a marker of DC motoneurons (Jackman et al.

2000) is challenged by our experiments. Indeed, these cells

have only been described in the anterior part of the nerve cord

and project caudally to innervate the whole set of somites

(Lacalli 2002). Coexpression of AmphiERR and islet has been

detected in the anterior part of the nerve cord, but evidently

neural cells expressing islet extend much more posteriorly. It is

therefore possible that islet labeling was detected in some, but

not all, DC motoneurons. This is supported by the ventro-

lateral islet expression observed on cross sections. In contrast

the medio-lateral expression found on certain cross sections

suggests that islet is also expressed in ventral compartment

(VC) motoneurons or in sensory neurons. In agreement with

the latter hypothesis, it should be remembered that zebrafish

LIM homeobox genes including islet1 also labels the sensory

Rohon–Beard neurons (Appel et al. 1995).

In summary, our results lead us to propose ERR genes

as markers of hindbrain segmentation. Since this is true

for amphioxus and zebrafish, it is likely that segmented ERR

expression was present in the hindbrain equivalent

of the last common ancestor of cephalochordates and

vertebrates.


ERR in motoneuron-muscle complex

Slow and fast muscles can be clearly identified during

zebrafish development. Fast fibers originate from the lateral

presomitic cells present in the segmental plate, whereas

slow fibers originate from adaxial cells, adjacent to the noto-

chord. The latter cells migrate radially across the somitic mass

and differentiate in muscle pioneer cells and in superficial

muscle layer (Devoto et al. 1996), two structures that express

zebrafish ERRa (Bardet et al. 2004 and data not shown).

ERRa is expressed in muscular masses during mouse embry-

onic development (Bonnelye et al. 1997). Furthermore, pre-

liminary results indicate that this receptor is preferentially

expressed in slow fibers versus fast ones in rat muscles

(P. L. Bardet and J. M. Vanacker, unpublished). The slow

muscle tropism is therefore a conserved feature of ERRaexpression.

Trunk musculature in amphioxus is composed of (i) deep,

fast fibers used in escape, emergency swimming; (ii) interme-

diate fibers which are thought to constitute a reserve for fur-

ther differentiation; and (iii) superficial slow fibers which are

dedicated to slow regular swimming (Lacalli and Kelly 1999).

In the amphioxus, these structures can be identified during

somitogenesis. The different types of muscle fibers are inner-

vated by specific motoneurons: fast fibers are innervated by

VC motoneurons whereas slow fibers receive influx from DC

motoneurons. Axons from the latter neurons lie entirely in the

nerve cord, and muscle cells send processes dorsally where

they contact neurons at the level of the DC (reviewed in La-

calli 2001). We found that AmphiERR is expressed in each

newly formed somite, at a superficial location, suggesting that

this receptor labels slow fibers but not fast ones. This view is

further supported by the fact that AmphiERR is not ex-

pressed in somite 1 which only comprises deep, fast fibers. The

transcription of AmphiERR in the mesoderm as well as in

DC motoneurons (but not in other types of muscle-innervat-

ing nerves), is consistent with the expression of this receptor in

both the nervous and the muscular compartments of a given

nerve-muscle complex.

Gene expression in nerve and muscle is tightly regulated

with information originating from one compartment acting

on the other. Examples of such cross-talk involve neuron-

secreted ligands and receptors expressed on the post-synaptic

membrane (reviewed in Schaeffer et al. 2001). On the other

hand, transcription factors belonging to the same gene family

have been reported to be expressed in both compartments,

but in this instance, different members of the family are found

in nerve and muscle. This is the case for the ETS gene family,

where combinations of ER81 and/or PEA3 are expressed in

motor and sensory neurons (Lin et al. 1998), whereas GABP

is the predominant ETS member in muscle cells (Schaeffer

et al. 1998; Khurana et al. 1999). To our knowledge, expres-

sion of an identical gene (AmphiERR) in both the nerve and

muscle cells that establish functional contacts with each other

is a unique feature of ERR genes.

The identity of the ERR-expressing cells in the zebrafish

hindbrain remains to be determined. Indeed, we performed

double in situ hybridizations with ERRg and either islet 1 or

evx 1 (data not shown). As in the amphioxus, expression of

islet1 does not overlap that of ERRg, thus excluding the latteras labeling cranial nerves. In the zebrafish hindbrain Evx1 is

expressed in a segmented manner in the primary and second-

ary interneurons (Thaeron et al. 2000). Although ERR-pos-

itive cells have striking similarities with the evx11 secondary

interneurons, technical limits prevented us from demon-

strating a coexpression of both genes.

Evolution of the mechanisms of hindbrainsegmentation

Several lines of evidence suggest that retinoic acid (RA) is a

key determinant in the regulation of vertebrate hindbrain

segmentation and rhombomere identity (reviewed in Gavalas

and Krumlauf 2000; see also Dupe and Lumsden 2001). RA is

produced by the somitic mesoderm adjacent to the posterior

brain and establishes an antero-posterior (AP) gradient of

concentration with the help of Cyp26 gene (an RA-degrading

enzyme) expressed in the midbrain (Swindell et al. 1999). In

Fig. 5. Evolution of head segmentation. (A) In ancestral chordates,neural (gray) segmentation is triggered by signals (vertical arrows)originating from already segmented mesoderm (hatched) with con-tributions (dashed arrow) from the posterior neuroectoderm. Noovert segmentation is present in the hindbrain. This situation stillexists in the amphioxus. (B) Neuroectoderm becomes autonomous:its segmentation program (solid arrow) lies in the neural plate itselfand uncouples from mesoderm segmentation. (C) Mesoderm seg-mentation is lost in the head. Overt segmentation appears in thehindbrain triggered by signals (solid arrow) originating from theneural tube. (D) Segmented hindbrain contributes to the patterningof the mesoderm through migration and differentiation of neuralcrest cells (vertical arrows). This situation prevails in modern ver-tebrates. Note that the intermediate stages have not been formallyidentified and that they do not necessarily appear stepwise duringthe course of evolution. Estrogen-receptor related (ERR) expres-sion are indicated as markers of segmentation (including in neuralcells), and do not imply any functional implication in segmentation.


the amphioxus RA has also been shown to exhibit AP pat-

terning activities (Holland and Holland 1996; Escriva et al.

2002). However this is not concomitant with the establish-

ment of morphologically conspicuous rhombomeres. Based

on these and other observations, including gene expression

pattern, Mazet and Shimeld (2002) have recently proposed

that posterior brain segmentation in ancestral chordates was

regulated by vertical signals, originating from the adjacent

mesoderm, which is also segmented in the head, in contrast to

the situation in vertebrates.

The data presented here support this hypothesis. Indeed,

we suggest that a segmentation control program, first active in

the mesoderm, is revealed by AmphiERR segmented expres-

sion in somites. In some way, over evolutionary time, mes-

oderm then instructs the neuroectoderm to undergo the

segmentation program. This might also be recapitulated dur-

ing development by AmphiERR expression in motoneurons,

which appears later as compared with that in the somites. It

should also be stressed that the compartments that express

AmphiERR (DC motoneurons and slow fiber muscles) are

intimately connected. To our knowledge, AmphiERR is the

only marker to date to display segmented expression both in

the mesoderm and in the neuroectoderm.

This hierarchy of control (summarized in Fig. 5A) could

have been present in primitive protochordates, a remnant of

which can be observed in the amphioxus. As a first step

towards the vertebrate situation, the instructive program

triggered by the segmented mesoderm might have been lost

(Fig. 5B). This could have been accompanied by a rein-

forcement of the influence of the posterior signaling (Fig.

5C). Ultimately the instructive program has been transferred

to the neural tube leading to morphologically segmented

hindbrain in which RA plays a prominent role (Fig. 5D). At

some point head mesoderm segmentation, now dispensable

for neural segmentation, was lost (Guthrie 1995; Kuratani

1997). As a final step, the segmented hindbrain controls the

patterning of the head mesoderm through the migration and

differentiation of neural crest cells (a vertebrate innovation),

to give rise to the complex vertebrate head (Fig. 5D).

Whether an already present segmentation program (as meas-

urable at the level of gene expression) triggered the iterated

appearance of given cell type, or whether the ancestral ex-

istence of segmented neurons led to repeated gene expression

is not clear. Whatever the case, it has been suggested (ap-

ropos of islet expression; Jackman et al. 2000) that this phe-

nomenon eventually led to morphological segmentation.

Expression of ERR in segmented neurons in the amphioxus

posterior brain is in favor of the latter hypothesis.

AcknowledgmentsWe thank the Centre National de la Recherche Scientifique (CNRS),Ecole Normale Superieure de Lyon and Ligue contre le Cancer,comites Ardeche, Drome and Loire for financial support. B. H. isfunded by Proskelia Pharmaceuticals (Romainville).

We are indebted to Thurston Lacalli for his help with interpretationof the amphioxus brain expression data and for communicating dataprior to publication. We thank Philippe Vernier for comments onzebrafish brain, Hector Escriva-Garcia and Frederic Flamant forcomments on the manuscript, and Marc Robinson-Rechavi forphylogenetical expertise. We thank William Jackman for providingus with amphioxus islet probe, Patrick Laurenti for the gift ofzebrafish evx1 probe, and Cecilia Moens for the gift of valentinozebrafish mutants.

REFERENCES

Andermann, P., Ungos, J., and Raible, D. W. 2002. Neurogenin1 defineszebrafish cranial sensory ganglia precursors. Dev. Biol. 251: 45–58.

Appel, B., et al. 1995. Motoneuron fate specification revealed by patternedLIM homeobox gene expression in embryonic zebrafish. Development121: 4117–4125.

Bardet, P.-L., Obrecht-Pflumio, S., Thisse, C., Laudet, V., Thisse, B., andVanacker, J.-M. 2004. Cloning and developmental expression of fiveestrogen-receptor related genes in the zebrafish. Dev. Genes Evol. 214:240–249.

Bonnelye, E., et al. 1997. Expression of the estrogen-related receptor 1(ERR-1) orphan receptor during mouse development. Mech. Dev. 65:71–85.

Dehal, P., et al. 2002. The draft genome of Ciona intestinalis: insights intochordate and vertebrate origins. Science 298: 2157–2167.

Devoto, S. H., Melancon, E., Eisen, J. S., and Westerfeld, M. 1996. Iden-tification of separate slow and fast muscle precursor cells, in vivo, priorto somite formation. Development 122: 3371–3380.

Dupe, V., and Lumsden, A. 2001. Hindbrain patterning involves gradedresponses to retinoic acid signaling. Development 128: 2199–2208.

Escriva, H., Holland, N. D., Gronemeyer, H., Laudet, V., and Holland,L. Z. 2002. The retinoic acid signaling pathway regulates anterior/pos-terior patterning in the nerve cord and pharynx of amphioxus, a chordatelacking neural crest. Development 129: 2905–2916.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approach usingthe bootstrap. Evolution 39: 783–791.

Galtier, N., Gouy, M., and Gautier, C. 1996. SEAVIEW and PHYLO_WIN: two graphic tools for sequence alignment and molecular phylo-geny. Comput. Appl. Biosci. 12: 543–548.

Gavalas, A., and Krumlauf, R. 2000. Retinoid signaling and hindbrainpatterning. Curr. Opin. Genet. Dev. 10: 380–386.

Giguere, V. 1999. Orphan nuclear receptors: from gene to function.Endocrine Rev. 20: 689–725.

Giguere, V. 2002. To ERR in the estrogen pathway. Trends Endocrinol.Metab. 13: 220–225.

Guthrie, S. 1995. The status of the neural segment. Trends Neurosci. 18:74–79.

Hermans-Borgmeyer, I., Susens, U., and Borgmeyer, U. 2000. Develop-mental expression of the estrogen-receptor related receptor g in thenervous system during mouse embryogenesis. Mech. Dev. 97: 197–199.

Hobert, O., and Westphal, H. 2000. Functions of the LIM-homeoboxgenes. Trends Genet. 16: 75–83.

Holland, L. Z., and Holland, N. D. 1996. Expression of AmphiHox-1 andAmphiPax-1 in amphioxus embryos treated with retinoic acid: insightsinto evolution and patterning of the vertebrate nerve cord and pharynx.Development 122: 1829–1838.

Holland, L. Z., and Holland, N. D. 1998. Developmental gene expression inamphioxus: new insights into the evolutionary origin of vertebrate brainregions, neural crest, and rostrocaudal segmentation. Am. Zool. 38:647–658.

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

Holland, L. Z., Holland, P. W. H., and Holland, N. D. 1996. Revealinghomologies between body parts of distantly related animals by in situhybridization to developmental genes: amphioxus versus vertebrates.


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

Holland, N. D., and Chen, J. 2001. Origin and early evolution of the ver-tebrates: new insights from advances in molecular biology, anatomy andpalaeontology. BioEssays 23: 142–151.

Holland, P. W. H. (2000). Embryonic development of heads, skeletons andamphioxus: Edwin S. Goodrich revisited. Int. J. Dev. Biol. 44: 29–34.

Horard, B., and Vanacker, J.-M. 2003. Estrogen receptor-related receptors:orphan receptors desperately seeking a ligand. J. Mol. Endo. 31: 349–357.

Jackman, W. R., and Kimmel, C. B. 2002. Coincident iterated gene ex-pression in the amphioxus neural tube. Evol. Dev. 4: 366–374.

Jackman, W. R., Langeland, J. A., and Kimmel, C. B. 2000. Islet revealssegmentation in the amphioxus hindbrain homolog. Dev. Biol. 220:16–26.

Khurana, T. S., Rosmarin, A. G., Shang, J., Krag, T. O., Das, S., andGammeltoft, S. 1999. Activation of utrophin promoter by heregulin viathe ets-related transcription factor complex GA-binding protein alpha/beta. Mol. Cell. Biol. 10: 2075–2086.

Kraus, R. J., Ariazi, E. A., Farrell, M. L., and Mertz, J. E. 2002. Estrogenrelated receptor a1 actively antagonizes estrogen receptor-regulatedtranscription in MCF-7 mammary cells. J. Biol. Chem. 277: 24826–24834.

Kuratani, S. 1997. Spatial distribution of postotic crest cells defines thehead/trunk interface of the vertebrate body: embryological interpretationof peripheral nerve morphology and evolution of the vertebrate head.Anat. Embryol. 195: 1–13.

Lacalli, T. C. 2001. New perspectives on the evolution of protochordatesensory and locomotory systems, and the origin of brains and heads.Philos. Trans. R. Soc. London B 356: 1565–1572.

Lacalli, T. C. 2002. The dorsal compartment locomotory control system inamphioxus larvae. J. Morphol. 252: 227–237.

Lacalli, T. C., and Kelly, S. J. 1999. Somatic motoneurones in amphioxuslarvae: cell types, cell position and innervation patterns. Acta Zool.(Stockh.) 80: 113–124.

Laudet, V., and Gronemeyer, H. 2002. The Nuclear Receptor Factbooks.Academic Press, San Diego.

Lin, J. H., Saito, T., Anderson, D. J., Lance-Jones, C., Jessell, T. M., andArber, S. 1998. Functionally related motor neuron pool and musclesensory afferent subtypes defined by coordinate ETS gene expression.Cell 95: 393–407.

Lu, D., Kiriyama, Y., Lee, K. Y., and Giguere, V. 2001. Transcriptionalregulation of the estrogen-inducible pS2 breast cancer marker geneby the ERR family of orphan nuclear receptors. Cancer Res. 61:6755–6761.

Luo, J., Sladek, R., Bader, J. A., Matthyssen, A., Rossant, J., and Giguere,V. 1997. Placental abnormalities in mouse embryos lacking the orphannuclear receptor ERR-beta. Nature 388: 778–782.

Mazet, F., and Shimeld, S. M. 2002. The evolution of chordate neuralsegmentation. Dev. Biol. 251: 258–270.

Moens, C. B., Yan, Y.-L., Appel, B., Force, A. G., and Kimmel, C. B. 1996.valentino: a zebrafish gene required for normal hindbrain segmentation.Development 122: 3981–3990.

Ostberg, T., Jacobsson, M., Attersand, A., Mata de Urquiza, A., andJendeberg, L. 2003. A triple mutant of the Drosophila ERR confersligand-induced suppression of activity. Biochemistry 42: 6427–6435.

Panopoulou, G., et al. 2003. New evidence for genome-wide duplications atthe origin of vertebrates using an amphioxus gene set and completedanimal genomes. Genome Res. 13: 1056–1066.

Pettersson, K., Svensson, K., Mattsson, R., Carlsson, B., Ohlsson, R., andBerkenstam, A. 1996. Expression of a novel member of estrogenresponse element-binding nuclear receptors is restricted to the earlystages of chorion formation during mouse embryogenesis. Mech. Dev.54: 211–223.

Prince, V. E., Moens, C. B., Kimmel, C. B., and Ho, R. K. 1998. ZebrafishHox genes: expression in the hindbrain region of wild-type and mutantsof the segmentation gene, valentino. Development 125: 393–406.

Saitou, N., and Nei, M. 1987. The neighbor-joining method: a new methodfor reconstructing phylogenetic trees. Mol. Biol. Evol. 4: 406–425.

Schaeffer, L., Duclert, N., Huchet-Dymanus, M., and Changeux, J.-P. 1998.Implication of a multisubunit Ets-related transcription factor in synapticexpression of the nicotinic acetylcholine receptor. EMBO J. 17:3078–3090.

Schaeffer, L., de Kerchove d’Exaerde, A., and Changeux, J.-P. 2001.Targeting transcription to the neuromuscular synapse. Neuron 31: 15–22.

Schmidt, H. A., Strimmer, K., Vingron, M., and von Haeseler, A. 2002.TREE-PUZZLE: maximum likelihood analysis using quartets and par-allel computing. Bioinformatics 18: 502–504.

Swindell, E. C., Thaller, C., Sockanathan, S., Petkovich, M., Jessell, T. M.,and Eichele, G. 1999. Complementary domains of retinoic acid produc-tion and degradation in the early chick embryo. Dev. Biol. 216: 282–296.

Thaeron, C., et al. 2000. Zebrafish evx1 is dynamically expressed duringembryogenesis in subsets of interneurons, posterior gut and urogenitalsystem. Mech. Dev. 99: 167–172.

Thisse, C., Thisse, B., Schilling, T. F., and Postlethwait, J. H. 1993. Struc-ture of the zebrafish snail1 gene and its expression in wild-type, spadetailand no tail mutant embryos. Development 119: 1203–1215.

Thomson, J. D., Higgins, D. G., and Gibson, T. J. 1994. CLUSTAL W:improving the sensitivity of progressive multiple sequence alignmentthrough sequence weighting, position-specific gap penalties and weightmatrix choice. Nucleic Acids Res. 22: 4673–4680.

Trainor, P. A., and Krumlauf, R. 2000. Patterning the cranial neural crest:hindbrain segmentation and Hox gene plasticity. Nat. Rev. Neurosci. 1:116–124.

Vanacker, J.-M., Bonnelye, E., Chopin-Delannoy, S., Delmarre, C., Cava-illes, V., and Laudet, V. 1999a. Transcriptional activities of the orphannuclear receptor ERRa. Mol. Endocrinol. 13: 764–773.

Vanacker, J.-M., Pettersson, K., Gustafsson, J.-A., and Laudet, V. 1999b.Transcriptional targets shared by estrogen receptorrelated receptors(ERRs) and estrogen receptor (ER) alpha, but not by ERbeta. EMBO J.18: 4270–4279.

Venkatesh, T. V., Holland, N. D., Holland, L. Z., Su, M. T., and Bodmer,R. 1999. Sequence and developmental expression of amphioxus Amph-iNk2-1: insights into the evolutionary origin of the vertebrate thyroidgland and forebrain. Dev. Genes Evol. 209: 254–259.

Wilkinson, D. G., Bhatt, S., Chavrier, P., Bravo, R., and Charnay, P. 1989.Segment-specific expression of a zinc-finger gene in the developing nerv-ous system of the mouse. Nature 337: 461–464.

Xie, W., et al. 1999. Constitutive activation of transcription and binding ofcoactivator by Estrogen-related receptors 1 and 2. Mol. Endocrinol. 13:2151–2162.


expression of estrogen-receptor related receptors in ... · have been identiﬁed (holland and...

Documents