Vol. 5, 1357-1365, December 1994 Cell Growth & Differentiation 1357

Rev-erbg3, a New Member of the Nuclear ReceptorSuperfamily, Is Expressed in the NervousSystem during Chicken Development’

Edith Bonnelye, Jean-Marc Vanacker, Xavier Desbiens,Agnes Begue, Dominique Stehelin, and Vincent Laudet2

Centre National de Ia Recherche Scientifique URA 1 160, Institut Pasteur,1 Rue Calmette, 59019 Lille Cedex, France [E. B., A. B., D. S., V. Li;Angewandte Tumorvirologie, Abteilung 061 0, DeutschesKrebsforschungszentrum, P. 1 01 949, D-69009, Heidelberg, Germany

[J-M. V.]; and Laboratoire de Biologie Cellulaire et Molbculaire duDbveloppement, Universit#{233} des Sciences et Techniques de Lille,

Villeneuve d’Ascq, France [X. D.]

Abstract

We have identified and characterized a new orphanmember of the nuclear hormone receptor superfamily inthe chicken. This new gene, called Rev-erbfi, exhibitsstrong homologies with the Rev-erba/ear-1 orphanreceptor gene, which partially overlaps the thyroidhormone receptor a gene in opposite orientation. Wedemonstrate that both Rev-erlxx and Rev-erbf.3 genes areconserved in their C and E domains. Rev-erbfi binds toDNA as a monomer and recognizes the same bindingmotif as the a gene product. The Rev-erbg3 gene productdoes not interact with retinoid X receptors, as revealedby gel shift experiments. In situ hybridizationexperiments show that Rev-erbl3 is expressed in thecentral and peripheric nervous system, spleen, andmandibular and maxillar processes, as well as in bloodislands. During embryonic development, we noticed astriking specific distribution of Rev-erbfi transcripts inthe notochord at 24 h and later on, in the floor plate ofthe neural tube. We propose that Rev-erbl3 may play animportant role in the complex network of inductivesignals, which control neuron differentiation.

IntroductionNuclear receptors play an important part in cellular regu-lation, providing a direct link between extracellular hor-monal signals and transcriptional responses (1 , 2). Thesemolecules are ligand-activated transcription factors thatregulate the expression oftarget genes by binding to specificcis-acting sequences (3-5). This family includes nuclearreceptors for hydrophobic ligands, such as glucocorticoids,estrogens, vitamin D, ecdysone, thyroid hormones, andretinoic acid. In addition to these receptors for known Ii-gands, several authors have described “orphan receptors,”which are putative receptors for ligands still to be identified(6, 7). Alternatively, it is conceivable that some orphanreceptors act as constitutive transcriptional regulators (8).The unique action of nuclear receptors, such as thyroidhormone receptors, are due to their ability to bind tandem

repeats of the core sequence AGGTCA, which are orga-nized either as palindromes, direct repeats, or invertedpalindromes. These receptors avidly bind their target se-quences through heterodimerization with RXRs3 (5). A largenumber of heterodimers between RXRs and other nuclearreceptors can be formed, and these various combinatorialinteractions are thought to play a major role in the finetuning of hormonal gene regulation (4).

Recently, some nuclear receptors have been shown tobind as monomers to the core sequence AGGTCA, such asNGF1 B, FTZ-F1 , or RORa (8-1 0). In these cases, the pro-tein directly contacts a specific Air rich sequence, which islocated just 5’ to the AGGTCA motif. It has been suggestedthat these proteins make direct contacts with this A/I richregion through small regions of charged amino acids, the Tand A boxes, situated in the vicinity of the DNA binding Cdomain (8, 1 1 ). This has also been shown for Rev-erba, forwhich the AT� rich sequence is Air A ATF N T (1 1).

The Rev-erba gene exhibits an additional interesting fea-ture; both in humans and rats, it partially overlaps thec-erbA-1 gene, which encodes THRa (1 2-14). Thus, Rev-erba mRNA is complementary to the mRNA of a non-T3-binding splice variant of THRa, called a2, which acts as adominant negative inhibitor of normal THRa (1 5-1 7). Sucha structure probably plays an important role in the regula-tion of thyroid hormone action, although the mechanismsby which the two genes can possibly regulate their respec-tive expressions are not well understood (1 8-20). Rev-erbctis expressed in adipocytes, brown fat, and muscles (1 3, 20).This gene is also widely expressed in brain and, morespecifically, in the neocortex, as shown in rats by in situhybridization (21). These experiments have also suggestedthat, in the neocortex, a mutually exclusive expression ofRev-erba and THRa2 transcripts may occur.

In this paper, we present the cloning and characterizationof the chicken Rev-erbf3, a new orphan receptor whichexhibits strong sequence homologies with Rev-erbct. Wedemonstrate that two Rev-erb genes actually exist in hu-mans and chickens. The product of Rev-erbf3 is able to bindDNA as a monomer and recognizes the same target se-quence as Rev-erbct. The expression pattern studied by insitu hybridization suggests that Rev-erbf3 may be implicatedin some aspects of nervous system development.

Results

Cloning of Rev-erbfi. Since Rev-erbct is highly expressed inskeletal muscles in humans and rats (1 2, 1 3), we used totalmRNA from chicken muscle in order to isolate its avianhomologue by RT-PCR experiments. To obtain a specific

Received 7/19/94; revised 9/26/94; accepted 9/29/94.

1 This work was supported by Association pour Ia Recherche contre le

Cancer, Centre National de Ia Recherche Scientifique, and Institut Pasteur deLille. V. L. is funded by the Fondation pour Ia Recherche M#{233}dicale.2 To whom requests for reprints should be addressed.

3 The abbreviations used are: RXR, retinoid X receptor; THRa, thyroid hor-mone receptor a; RT-PCR, reverse transcription-polymerase chain reaction;bp, base pair(s); cDNA, complementary DNA; EMSA, electrophoretic mo-bility shift assay; HRE, hormone-responsive element; ROR, related orphanreceptor.

1358 Rev-erb�, a new orphan receptor

4 D. Moore, personal communication.

Rev-erb amplification product, we used primers (Rev-erbland Rev-erb2) corresponding to the most conserved regionsof the C domain and to the T-box specific of Rev-erb. Weobtained a 150-bp fragment which has 75% sequence iden-tity with mammalian Rev-erbcr sequences. This PCR frag-ment was used to screen a chicken fibroblast embryo (El 1)cDNA library. A 2.2-kilobase cDNA clone containing asingle open reading frame was isolated. The ATG of thisclone is in accordance with Kosak rules, with the onlyexception of a C in position -3, which is found only in rareoccasions in vertebrates mRNAs (data not shown). Transla-tion of this open reading frame predicts a protein of 549amino acids clearly related to human and rat Rev-erba (Fig.1 A). Nevertheless, comparison with sequences of domainsC, D, and E of mammalian Rev-erba revealed lower ho-mologies (54%) than expected. In the conserved domainsC, D, and E, the homology between mammalian and avianhomologues of other nuclear receptors range between 99%(e.g., COUP-TF II) to 84% (e.g., estrogen receptor). The lowdegree of similarity between human and chicken Rev-erbproteins may be explained by rapidly accumulating muta-tions over time specific for this orphan receptor sequence.Alternatively, this could indicate that we had not isolatedthe real chicken homologue of Rev-erbcr but rather anorthologous, related gene.

To test this possibility, we used RT-PCR to search for asecond Rev-erb gene in the human genome. Using the sameset of PCR primers, we amplified reverse-transcribed mRNAfrom human HeLa cells. PCR products were cloned, andsequences of individual recombinants were analyzed. Ofseven clones, we obtained one clone that is identical to thepreviously described human Rev-erba gene and six whichrepresent a new, Rev-erb-related sequence. Comparison ofthe predicted amino acid sequence deduced from these sixclones revealed a single amino acid change when com-pared to our chicken Rev-erb cDNA and three when com-pared to the mammalian Rev-erbct gene (Fig. 1 8). At thenucleotide level, the new human PCR clone contains 90%homology with the chicken clone and only 80% with mam-malian Rev-erba. These data suggest that we have isolateda chicken cDNA corresponding to a new gene related, butdistinct from, Rev-erbct that we have called Rev-erbf3. Dur-ing the course of this work, we learned that the group ofDavid D. Moore4 isolated a new human cDNA calledBD73, which is identical to our human PCR product at theamino acid level and can be considered as the humanhomologue ofour chicken cDNA clone (Fig. 1). Amino acidsequences of the domains C, D, and E of human andchicken Rev-erbj3 exhibit 85% of identity, a value which issimilar to those obtained between other chicken and hu-man homologous nuclear receptors. Recently, we obtaineda new chicken cDNA clone that encodes a chicken homo-logue of the mammalians Rev-erba (data not shown). Thus,as in humans, two Rev-erb genes exist in the chicken ge-nome.

The predicted chicken Rev-erb�3 product appears clearlyto be a member of the Rev-erb group since it exhibits thesame characteristics of the a gene: highly conserved spe-cific T and A boxes located immediately downstream of theC domain and a long D domain shorter in the 3 gene thanin the a gene (12, 13; Fig. 1A). The NH2-terminal A/B

domain is poorly conserved. When compared to othernuclear receptors, Rev-erbf3 exhibits the strongest homolo-gies with peroxisome proliferator-activated receptors, reti-noic acid receptors, and THRs (31 to 34%), confirming theplacement of Rev-erb genes inside the first subfamily ofnuclear receptors (7, 22, 23). This observation is also con-firmed by the phylogenetic reconstruction, which alwaysclusters together peroxisome proliferator-activated recep-tors and Rev-erb genes (Fig. 1 D).

To check whether the Rev-erbf3 gene is conserved inother vertebrate groups, we decided to search for Rev-erbf3-specific sequences in Xenopus /aevis. Using nested degen-erated primers Rev-erb3, Rev-erb4, and Rev-erb5, we per-formed PCR experiments searching for an exon of the Edomain. We amplified a PCR fragment which was cloned,sequenced, and compared to both Rev-erbct and Rev-erbf3

genes (Fig. 1 C). We obtained a clear Rev-erbf3 signature,since the amino acid homology is 84% between the Xeno-pus sequence and the chicken Rev-erbf3 and only 62% withthe human Rev-erbct. Thus, the Rev-erbf3 gene appearsconserved in amphibians.

Rev-erbg3 DNA Binding Activity. The Rev-erbct geneproduct was recently shown to bind, as a monomer, theRevRE target sequence (Fig. 2A; Ref. 1 1 ), which can beseparated into a AGGTCA core sequence recognized byalmost all types of nuclear receptors, and a 5’ AlT-richspecific sequence. We thus examined whether the Rev-erbfi product was also able to bind to the RevRE. In vitrotranslation of Rev-erbf3 yielded a peptide of the expectedsize (60 kilodaltons; data not shown) that was used in EMSAand shown to bind to the RevRE as strongly as the productof Rev-erba (1 1). This binding is specific since it was com-peted out by a 10- or 100-fold molar excess of unlabeledoligonucleotide, whereas a RevRE mutated in the three finalpositions ofthe AGGTCA motifwas unable to compete withthe probe (Fig. 2A, compare Lane 27 to Lanes 1 and 2 orLanes 25 and 26, respectively). Consistent with the relativesize of the proteins, the complex obtained with Rev-erbf3migrated slightly faster than the Rev-erba-induced complex(Fig. 2B).

The DNA binding features of Rev-erbct and Rev-erbf3gene products were compared (Fig. 2A) by studying thecompetition potential of several mutations of the RevRE. Allthe competitors behaved, irrespective of the Rev-erb prod-uct used, an observation consistent with the strong se-quence conservation ofthe DNA binding domain ofthe twoproteins. We obtained essentially the same results as didHarding and Lazar (1 1) for the binding of Rev-erba, exceptfor the T9C mutant, which clearly exhibited competitivebehavior in our laboratory, whereas these authors onlynoticed a weak effect. Of note, strong alterations (but nota single point mutation of the 5’ AlT-rich region, nude-otides 9 to 1 1 ) almost completely abolished competition,pointing to the importance of these bases. In contrast,mutations further upstream (e.g., 1 1-1 3C) or inside theAGGTCA core motif (C2G or Al C) only weakly affectedcompetition.

Since numerous nuclear receptors are able to interactwith the RXRs, we next examined whether Rev-erb�3 wasable to form a complex with RXRI3 or RXR’y. We showedthat the binding patterns observed with Rev-erba or Rev-erbj3 alone on RevRE probe were not altered by the pres-ence of RXR�3 or RXRy (Fig. 28). We also verified that noneof the Rev-erb products, either alone or in combination withRXRf3 or RXRy, was able to bind to a palindromic HRE or to

Cell Growth & Differentiation 1359

A10 20 30 40 50 60 70 80 90 100 110 120

�zHtaii --KS

uRat - - . G

�l4um EV .A....A..S..S.AS.PA.CH---.EG.EN.... .SS.V.S.PNSSN.DTNGNPKNGDLANIEGILKNDRIDCS.KT.KSSA.GMT..

�Ck .S QRELRQRF PVILFAC T.FSEQLLL.KQLNGR. . EG.----DGAPKSDR.EE.VKP.QSGVAGLT.G

ETh .G.AVVPADSRPQT . E.IKSY.V NDTTVAS.VKGE.ELN--

130 140 150 160 170 180 190 2 210 220 230 240

� �250 260 270 280 290 300 310 320 330 340 350 360

eHum

aRat . . . .AP. . .S.SV A.T T QC. .5. . . .ST I

�Hum D.L-VE.HEQTALPAQEQLR.K.Q LEQENIKSSSP. .SDFAK.E. .GP4.T. . .KDT.M.NQEQQEPI.AESMQPQRGERI.KNMEQYNLNHDH.GNGLSS-HFPCSESQQH-...

�Ck .AL-AE.QDQ---A.QEDLSSK.K .ERET.KS.SP. .SDI4AK.E. .GM.T.V.KDT.M.NQEQSQNPAEMMQPQSGERVSKN.EQYMLSSEH.VSGLSGPQYP--DSEQH-.G.

ETh -ALA.ELDDQPRL LAA.L. . .L.TCEFTKE.VSAMRQR ARD. .5

370 380 390 400 410 420 430 440 450 E460 470 48�xtlum LRQAPSS- -YPPTWPPGPAHHSCHQ-SNSNGHRLC- --PTHVYAAPEGKAPANSPRQGNSKNV LLACPMNMYPHGRSGRTVQEIWEDFSMSF4�AVREVVEFAKHIPGFRDLSQHDQVT

czRat . . .G. . .-- S -P ---. . . . .5 GL. . . .T

�l4um QFKGRNIMH. .NGHAICI.NGH.MNF. .AYTQ.V.DRV.IDGFSQN.N.NSYLCNTG.R---MH.V. . .SKS.YVDPHKSGH. . . .E K R N.

�Ck QYKGR.TMH. .SGHAICFTNGH.MNFT.GYTQ. . .DRI.ED.FSPNRNT-TYSCNTG.R---MH.V. . .SKT. .VDPNKSGH.V. .E. .L K R N.

ETh .514.1 L.PA.ELQ.E QE. .QR.A�iVI.G.ID. .GM. . . .QL.T.D.KF..

490 500 510 520 530 540 550 560 570 580 590 60eHum

uRat -N - S�IIum DA.ER. .T. . .GKK. .VDD.HS. .A. . . .-NS. .E A.Q.SD. .MS NRK.VN. . .A I. . . .T.IM. .II.N.A.I

�Ck DARER. .T. . .GKK. .VDD.HS. .A. . . .-NS. .E A.Q.SD. .MS I-. .VN. . .A I. . . .T.IM. .H.N.A.I

ETh

610 620aHum LKLPDLRTLNNMHSEKLLSFRVDAQ

uRat�Hum S E. .A.K.HP-

�Ck S E. .A.K.HP-ETh ETP4 STL.T. . .VV. .TE-H

B aHum KRCLKNENCSIVRINRNRCQQCRFKKCLSVGMSR D�Hum .K M.M

�Ck .K. . . .N. . . .M.ME75 RP.T. .QQ. . .L Y. .L. . .IA

C aHum SRTTYSLQELGAMGMGDLL-SAMFDFSEKLNSLABRat -NiHum .GKK. .VDD.HS. .A. . . .-NS. .E A.Q�Ck .GKK. .VDD.HS. .A. . . .-NS. .E A.Q

�Xen .GKK. .VD. .RS. .A. . . .-NSV.E SA.Q

E75 NGQVMRRDAIQNGANARF.VDST.N.A.RM. .MN

Fig. 1. A, alignment ofthe sequences of the chicken Rev-erb�3 (�Ck) with other members of the Rev-erb group: human Rev-erbcx (cxHum; Ref. 12), rat Rev-erbcr

(aRat; Ref. 1 3), human Rev-erbf3 (f3Hum)4 and Drosophila E75 (E75; Ref. 45). The C and E domains are boxed; the TA boxes are underlined. B, comparison ofthe translated sequence of the human Rev-erbf3 PCR clone with other Rev-eth in the DNA binding domain. C, comparison of the translated sequence of theXenopus Rev-erbf3 PCR clone (f3Xen) with other Rev-erb factors in the ligand-binding domain. Sequences were aligned with Clustal V; gaps are indicated byslashes, and amino acids identical to those of aHum are noted by points. D, phylogenetic tree obtained by the Neighbor-Joining method from the alignment of(A) showing the Rev-erb group inside the first subfamily of nuclear receptors.

a direct repeat of the core motif AGGTCA spaced by 4 products are able to interact with RXRs, at least for bindingnucleotides (DR4 element; data not shown; see also Fig. 2C to RevRE or HREpal.for a positive control of RXRI3 action). We conclude that, Expression of Rev-erbj3. The expression pattern of Rev-using these EMSA conditions (24, 25), none of the Rev-erb erbf3 was first investigated by Northern blot analysis. Two

A FoldColnpelilor Rev.crtkt Rcv-erb� Lan��sexcer �

14 lll2lll() 9 II 7 6 5 4 3 2 I I() I

RcvRE TAGAATGTAGGTCA - -v 100 2

It) 3

GIIA TAGAATATAGGTCA100 -

10g-

t4

5

AIC TAGAATGTAGGTC�100

C20 TAGAATGTAGGTQAlOt)

IOW

- -

6

7

II

10 #{149} 9

lI-DC T � A I G I A G G I C A1(X)

10 W �T10

II

lO-12C TA�...�...�TGTAGGTCA1(X)

l()

9-IIC T A C LC...� G T A G G T C A100

10

N.l(%� TAGA�...�...�TAGGTCA1(X)

10

7-9C TAGAA�...�....�AGGTCA1(X)

l()

Al()C TAGA�TGTAGGTCAI(S)

-

�

�

Z�

#{149}#{149}j�,

-

�

��

12

13

14

15

16

17

is

19

20

10 21

i’x: TAGAA�GTAGGTCA1(5)

-22

It) 23

Tic TAGAATG�AGGTCA101)

0

intiiRcs’Rl� T A G A A I G T A G G �I(S)

No coinpelilor

‘T”#{163}

�

#{149}#{149}#{149}j#{149}#{149}

�

24

25

26

27

C ,�4?ds*�sX

Fig. 2. A, EMSA realized with re-

ticulocyte lysate-produced Rev-erbe or Rev-erbf3 proteins and a Ia-

beled RevRE oligonucleotide asprobe. For each lane, a iO- or 100-

fold excess of the indicated coldcompetitor oligonucleotide was

added, except for Lane 27, forwhich no competitor was added.

The sequence of the RevRE and ofthe various mutants used for compe-tition are indicated 8). EMSA real-

ized with Rev-erbci or Rev-erbfi in

absence (Lane 1) or presence ofRXRJ3 (Lane 2) or RXRy (Lane 3).Lane R corresponds to an EMSA re-

alized with unprogrammed reticulo-cyte lysate. C, EMSA realized withhuman THRai in the presence or

absence of RXRf3 using a palm-dromic HRE. The position ofthe het-

erodimeric complex is indicated.

BRev-erbcz Rev-erb�

RI 2 3 1 23

1360 Rev-erbjl, a new orphan receptor

I� � Heterodin�er

�

4 Nl(,n()Iner

A

D �- :� � �� , � ...p. � � �

.. ...‘ : ‘� .“i:m, k; . � � ,..- ‘� 0. ‘�

� ,� ;. ��:�:‘�; .� . � �

.�.a �, �.. ‘: �, ��tr � �. .,

� .1 � �.. . . . ., �: : � .-� p�. �. . � � � ‘

� � �; � ��

Fig. 3. Expression of Rev-erbj3 in 1 -day-old chickens. A, Northern blotshowing Rev-erb)3 expression in various organs after hatching. B, expression

of Rev-erbg3 by in situ hybridization in the spleen of 1 -day-old-chickens. Anadjacent section of the spleen hybridized with sense probe displays no signal

(data not shown).

Cell Growth & Differentiation 1361

/1 u/si

. � 4.5kb

. .4- 3.0kb

mRNA transcripts of 4.5 and 3 kilobases were detected (Fig.3A). A very high expression level of these two transcriptswas detectable in the spleen after hatching. In situ hybrid-ization of spleen sections revealed a specific expression ingranulocyte-rich regions (Fig. 38). With a Northern blotcontaining mRNA extracted from mechanically dissociatedspleen cells (26), we detected Rev-erbf3 transcripts in thefraction containing granulocytes but not in other fractions(data not shown). Furthermore, we never observed a Rev-erbf3 signal in mRNA extracted from thymus or bursa ofFabricius from 1-day-old chickens (Fig. 3A). Interestingly,we also detected Rev-erbfi mRNA in the extraembryonic

blood islands of El-stage embryos (Fig. 4, A and B) sug-gesting an expression in early erythroid cells. Consistentwith this observation, Rev-erbf3 mRNAs are present in ery-throid cell lines such as 6C2 (data not shown). Thus, Rev-erbj3 seems to be expressed in a variety of chicken hema-topoietic cells ofthe erythrocytic and granulocytic lineages.

In situ hybridization was also performed on sagittal ortransversal sections of E4, E6, and E7 chicken embryos.Rev-erbf3 transcripts were distributed in the wall of all ner-vous system vesicles at E4, E6, and E7 (Fig. 4Cand data notshown). Analysis of transversal sections also demonstrated

Rev-erbf3 expression in the cephalic mesenchyme and neu-roepithelium (Fig. 4H). Moreover, Rev-erb/3 mRNAs weredetected in both cranial ganglions 3 and 5 (Fig. 4D) at E4and in E6 and E7 dorsal root ganglions (Fig. 4N). TheNorthern blot experiment showed a weak expression ofRev-erb/3 in the brain after hatching (Fig. 3A). Interestingly,

at E4 (Fig. 4, Kand L), E6 (Fig. 4M), and E7 (data not shown),Rev-erb� was strongly expressed in the ventral midline ofthe neural tube (i.e., in the floor plate). No specific expres-

sion was seen in the lateral or dorsal part of the spinal cord.This specific floor plate expression prompted us to examineyounger embryos, and we observed that, at El , Rev-erbf3wasexpressed in notochord and in all ofthethe neural plate(Fig. 41).

We also observed Rev-erb� transcripts in mesodermalcells. At E4, a strong signal for Rev-erbf3 was seen in dif-ferentiating somites (Fig. 4H). In order to precisely localizethis expression, we used the monoclonal antibody 13F4,which is a marker of early embryonic heart muscle (Fig. 4/,arrowhead) and somitic myotome (Fig. 41, arrow). Compar-ison between localization of sclerotome and myotome in-dicated that Rev-erb/3 was expressed in migrating sclero-tomal cells, which will give rise to vertebral bodies. Rev-erb�3 expression was also detected in cephalicmesenchyme, which participates in the cephalic skullcapstructure (Fig. 4H), as well as in maxillary components (datanot shown) and in mandibular components of the first bra-chial arch (Fig. 4, E and F). As shown by 1 3F4 immunohis-tochemical labeling, Rev-erb� was also detected in themyocardium (Fig. 4, H, arrowhead; Fig. 4/ for 1 3F4 label-ing). Finally, we noticed that the dorsal mesocardium (Fig.4H, doub/e-arrow; see an enlargement of this structure inFig. 4G) was strongly labeled by our probe.

Discussion

We have isolated and characterized Rev-erb�, a new or-phan receptor, isolated from a chicken cDNA library. TheRev-erb genes are new examples of nuclear receptors forwhich two different genes seem to encode proteins with

apparently overlapping functions. As in other cases, the twoRev-erb genes seem to be conserved in all vertebrates andcould be viewed as recently duplicated homologues of aunique ancestral gene which remains unique in Drosophi/a(E75; Ref. 45). However, this has to be experimentallyproved by the analysis of Rev-erb genes in lower organisms.The Rev-erbj3 gene seems not implicated in an overlappingstructure like Rev-erbct in mammals since it does not exhibitany open reading frame in its opposite strand. Furthermore,the rate of accumulation of mutations in the third position ofcodons in the region homologous to the overlapping regionof chicken Rev-erb�3 is not significantly different from thevalue in the more 5’ regions. This contrasts with what it isobserved in mammalian Rev-erbcs genes (data not shown).The discovery of Rev-erb/3 also sheds light on the origin ofthe c-erbA-1/Rev-erbcs overlapping locus. It has been pro-posed that this locus was created by translocation of one ofthe members inside the gene of the other (7, 27). Thistranslocation would have created the alternative c-erbA-1exon 10, which overlaps Rev-erba. This model implies thatRev-erba, and particularly its 3’ part, is conserved duringevolution and that the c-erbA- 1 exon 10 is not (27). The factthat the 3’ region of Rev-erb� shows a high level of homol-

ogy with Rev-erbcs gives a strong argument for this model,as does the absence of c-erbA-1 exon 10 in chicken (28).

CA � ..

:��\ #{149}‘

� bi

B �. : . #{149}��

.�. i �.. te

. . . :. � : � �‘ � . : . �. . . . � � � -. �

�..,.:..‘

� bi..

. � w hI I\�_�4�_ .J-

q*’.:;. b

� I ‘i

.o�

a

. si

SI � �

J .,........� y�:�’

I. �

.. , . ‘:: I�-

I.. � ‘.�... . .�. . ,�. :#{149}� � � � #{149} .r#{149}-

. I� . � : . : � � � � � #{149}� � � .�

.I�. � � #{149}.� . �; .

M

1362 Rev-erb[3, a new orphan receptor

Cell Growth & Differentiation 1363

Rev-erbf3 presents strong homologies with the Rev-erbagene product. In fact, the chicken Rev-erb�3 protein ap-peared identical in structure to the rat or human Rev-erba.Indeed, both domains C and E are well conserved (94 and73% identity, respectively). Interestingly, one of the mostconserved regions among Rev-erba, Rev-erb(3, and E75 liesoutside the C domain in the T and A boxes. These boxeswere described previously as necessary for the DNA bind-ing as monomers of the orphan receptors NGF1 B and SF1(8). These two boxes seem also to be important for the DNAbinding of RXRs, retinoic acid receptors, and THRs, prob-ably by establishing DNA contacts across the minor groovewith the phosphate backbone located 5’ to the AGGTCAcore motif (29-31). Thus, it seems likely that, in mostnuclear receptors, the T and A boxes are parts of the DNAbinding domain, which could strengthen the interactionbetween the receptor and its target sequence. For monomerbinding receptors (NGF1 B, SF1/FTZ-Fl , Rev-erb, RORs,and also THRa), this stabilization is sufficient to ensure astable monomer to bind to DNA, whereas in other cases,the receptors need homo- or heterodimeric interactions toachieve efficient DNA binding. Consistent with this model,known monomer binding receptors contain a high propor-tion of basic amino acids in the T and A boxes, which couldbe responsible for this increased interaction strength (8, 10).At least for NGF1 B and SF1 , some of the basic amino acidsof the T and A boxes have been directly implicated in DNAbinding. This model could explain why, although evolu-tionarily distant, the monomer binding receptors have allbeen acquired, by the same convergent way, the ability tobind DNA as monomers.

Consistent with its high homology with Rev-erba, Rev-erbf3, the newly discovered nuclear receptor we describedin this paper, binds to the RevRE as a monomer. In ouranalysis, we were unable to detect any binding of Rev-erbf3or Rev-erba in palindromic HRE or in direct repeat elementsspaced from 0 to 5 nucleotides (data not shown). Thisobservation contrasts with the recent study of the rat Rev-erba product, which was shown to bind to a DR4 element(32). Nevertheless, the DR4 element used by these authorscontained 5’ to the first repeat an AlT-rich sequence, whichcould be a perfect RevRE site with A or T nucleotides atcritical positions (nucleotides 9 to 12; Fig. 2A). Based onmutant RevRE analysis, the Rev-erbg3 product exhibits thesame DNA binding specificities when compared to Rev-erba. This observation suggests that both proteins, whenexpressed in the same cell, will be able to compete forthe occupation of the RevRE and, presumably, to regulateoverlapping sets of genes. Furthermore, it is possible thatother nuclear receptors capable of binding DNA as

monomers can bind to sites close to the RevRE. Thiscould be the case of the recently described RORs, whichrecognize a monomeric response element closely relatedto the RevRE (10).

In this study, we have analyzed the expression of Rev-

erbf3. We were particularly intrigued by the high expressionin notochord at El and in floor plate at later stages. Nu-merous studies have demonstrated that the notochord playsa crucial role in inducing cells of the ventral midline of theneural tube to become floor plate cells (recently reviewed inRefs. 33 and 34). The floor plate provides signals that reg-ulate the development of motor neurons (35) and releases adiffusible chemoattractant molecule that stimulates and di-rects the growth of commissural axons in vitro (36-39). Thedifferentiation of floor plate cells and motor neurons isassociated with expression of different classes of genes.Only one class of transcription factors is known to bespecifically expressed in floor plate: the members ofthe forkhead/HNF3 family originally identified as liver-specifictranscription factor (Ref. 40 and references therein). Indeed,HNF3� has been shown to control floor plate developmentsince ectopic expression of this gene leads to various ab-normalities of the brain and neural tube, including ectopicexpression of floor plate-specific genes (40). Contrary toHNF3-�3 expression, Rev-erbf3 expression in the neuralplate is intense at early stages (48 h) and later becomesrestricted to the floor plate. Therefore, we hypothesized thatRev-erbf3 is not particularly important for the induction offloor plate but rather in later aspects of neuronal develop-ment, such as axon guidance. The timing of expression ofRev-erbfi seems comparable with those of the recently de-scribed nestrins (38, 39), which are actively implicated in

this process. The precise delineation of the spatio-temporalexpression of chicken Rev-erbf3 in the neural tube is nowunder way. The fact that Rev-erbf3 encodes a member of thenuclear receptor superfamily, which could be regulated bya ligand still to be identified, is also very interesting. Such aligand may play a role in the complex signaling pathwaysimplicated in ventral neural tube differentiation.

In this study, we were also intrigued by Rev-erbf3 expres-sion in muscle. Whereas expression was visualized byNorthern blot and in dorsal muscles at E6 (Fig. 4N), theRev-erb�3 transcription in myotome was absent. This obser-vation would argue in favor of the possibility that Rev-erbf3occurs at later stages of muscle differentiation. This situa-tion has been observed for some myogenic factors likemyf-5, which is transiently expressed in myotome and ac-cumulates later in the muscles (41).

Fig. 4. In situ hybridization study of Rev-erbf3 expression in chicken embryos. A, the section of an El embryo blood island hybridized with sense Rev-erb� probeexhibit no signal. B, expression of Rev-erbg3 in an adjacent section with the antisense Rev-eth�3 probe. C, sagittal sections through E4 embryo hybridized withRev-erb�3 antisense probe; a conspicious expression is seen in neural vesicles, ganglions, and the mandibular component of the first branchial arch and, to a lesserextent, in heart and subsets of mesenchymal cells. 0, E, and F, higher magnifications. Rev-erb/3 is expressed in the cephalic ganglions 3 and 5 (0) and in themandibular component of the first branchial arch (F). Hybridization performed with the sense probe showed no labeling over the background signal in theadjacent section (F). C to I, transverse sections of E4 chicken embryo hybridized with Rev-erbB antisense probe or stained using a specific procedure. G, PicroIndigo Carmin and nuclear red staining of the section at the level of lungbud and the dorsal mesocardium. H, expression of Rev-erbfi in cephalic mesenchyme,heart (arrowhead), dorsal mesocardium (double arrow), and sclerotomal cells (arrow). I, adjacent section processed for immunohistochemistry using themonoclonal antibody 1 3F4, which recognizes a specific antigen expressed in the myogenic lineage in early embryonic heart (arrowhead) and in somitic myotome(arrow). I though N, expression of Rev-erbj3 in the nervous system. I, expression of Rev-erbf3 in El embryo in notochord and neural groove; K and L, expressionin the floor plate of the neural tube at E4 (K, antisense probe; L, sense probe) and at E6 (M); N, parasagittal section of a E6 embryo showing labeled dorsal rootganglions and dorsal muscles. a, aorta; bi, blood island; cm, cephalic mesenchyme; di, diencephalon; dm, dorsal muscle; ec, ectoderm; ed, endoderm; eo,

esophagus; fo, foregut; fp, floor plate; g3, ganglion 3; gS, ganglion 5; h, heart; lb. lung bud; m, mesocardium; me, mesencephalon; mp, mandibular process; my,myelencephalon; myo, myotome; n, neuroepithelium; nt, notochord; ng, neural grove; Sc, spinal cord; sI, sclerotome; sg, spinal ganglion; te, telencephalon; e,encephalon.

1364 Rev-erbf3, a new orphan receptor

Materials and Methods

PCR, Sequencing, and Northern Blots. Messenger RNAsfrom chicken muscles were extracted according to standardmethods (42). RT-PCR was performed using 1 5 pg of mRNAand denaturated at 90#{176}Cfor 2 mm with 3 units of Moloneymurine leukemia virus reverse transcriptase from GIBCO/BRL with the buffer [0.5 mt�i of each deoxynucleotidetriphosphate (Pharmacia), 1 unit of RNAsin (Promega), 40mg/mi bovine serum albumin (Promega), and 100 ng of

specific primer] furnished by the supplier. The reactionmixture was incubated at 37#{176}Cfor 1 h, and the reactionproducts were denaturated at 100#{176}Cfor 5 mm. PCR wasconducted with TAQ polymerase and buffer from Amer-sham, using standard conditions, with the following 40cycles: 94#{176}Cfor 1 mm, 45#{176}Cfor 2 mm, 72#{176}Cfor 1 mm, anda final elongation step of 7 mm at 72#{176}C.RT reaction wasperformed with degenerated primer Rev-erb2 (correspond-ing to positions 1229-1 251 in the human Rev-er/ia cDNA;Ref. 1 2), which is localized in domain D, whereas the PCRreaction was performed with primers Rev-erbl (positions1104-1124) and Rev-erb2. Oligonucleotides sequencesused were: Rev-erbl, 5’-CAT CCA GCA GAA CAT CCAGTA-3’ (1 1 04-1 1 24); Rev-erb2, 5’-AT(G/A) CG(T/C) CCAAAA CGC ACA GC(G/A) TC-3’ (1 229-1 251 ); Rev-erb3,

5’-GT(G/A) AAI A(A/G)(A/G) C(CTr)C A(Gfl)C TCC TC-3’(2230-2249); Rev-erb4, 5’-C(C/T)C A(G/T)C TCC TC(C/A)TC(G/A) (G/C)TA AG-3’ (2221-2241); Rev-erb5, 5’GA(C/G)C(A/G)(GTr) AC(Af�) GT(G/C) A(T/C)(G/C) TTC CT-3’(2100-2120).

PCR products were cloned in the PCR II vector from theTA cloning kit of Invitrogen Corporation with the conditionsof the supplier. Sequencing was done on an Applied Bio-system 373A automatic sequencer using the PRISM kit.Northern blot analyses were performed using 20 pg of totalRNA and standard procedures (42). The probe used encom-passed the 3’ part of the chicken cDNA (position 448 to1469). Spleen cells were fractionated as described (26).

Primers Rev-erb3, Rev-erb4, and Rev-erb5 were used toperform nested PCR to search for Xenopus Rev-erb se-quences. Rev-erb3 and Rev-erb5 allow the amplification ofa 1 50-bp fragment, which was reamplified with primersRev-erb4 and Rev-erb5 to obtain a 142-bp fragment.

Isolation of cDNA Clones. Approximately 800,000 re-combinant bacteriophages from El 1 chicken embryo fibro-blast cDNA library constructed in Agtl 1 (Clontech) werescreened with a 1 50-pb EcoRl fragment of the chickenRev-erbf3 clone obtained by RT-PCR, which had been Ia-beled with ks-32PIdCTP (NEN) by nick translation (Amer-sham). Positive clones were purified, and the EcoRl insertwas subcloned into pUC1 9 vector for sequencing and inPSG5 vector for in vitro translation.

Gel Shift Experiments. Full-length human Rev-erbct orchicken Rev-erbf3 was synthesized in reticulocyte lysatesusing the TNT kit from Promega. Synthesis yield waschecked with l35Slmethionine labeling of the proteins andmigration ofan aliquot in a sodium dodecyl sulfate-page gel(data not shown). Gel shift experiments (EMSA) were doneas described (24). For the study of Rev-erb-RXR interactions,two stringency conditions for EMSA were used: those ofHupp et a!. (Ref. 24; Fig. 2, B and C) and those of Graupneret a/. (25), with identical results.

In Situ Hybridization and Immunocytochemistry. El,E4, E6, and E7 chicken embryos and dissected organs werefixed at 4#{176}Cfor 1 6 h in 4% paraformaldehyde in phosphate-

buffered saline containing 5 mist MgCI2, dehydrated andembedded in paraffin. Five-pm-thick sections were trans-ferred to 3-aminopropyltriethoxysilane (TESPA; Aldrich)-coated slides and dried at 42#{176}Cfor 2 days. In situ hybrid-ization was performed as described (43) using a 35S-labeledantisense RNA probe synthesized from a full-length chickenRev-erbf3 cDNA cloned in Bluescript II KS (Stratagene). Toavoid any cross-hybridization with Rev-erba mRNAs, hy-bridization was performed with stringent conditions (60#{176}C).As a negative control, adjacent sections were hybridizedwith a 35S-labeled sense RNA probe synthesized from thesame template. At the end of the in situ hybridization pro-tocol, nuclei were stained with bisbenzimide and appearedblue under fluorescent light. 1 3F4 monoclonal antibodywas used as a specific marker for myogenic cells (44).Staining was performed using mouse antibody to the mono-clonal antibody 1 3F4 and the biotin extravidin kit (Sigma),according to the manufacturer’s specifications.

AcknowledgmentsWe thank David D. Moore for communicating results prior to publication.We are grateful to Christophe Queva for discussions about Rev-erb)3 expres-sion in floor plate, Violetta lotsova for help in gel shift experiments, Rachid

Safi for help in sequencing, and Catherine Ziller for the 1 3F4 antibody. Wethank Hinrich Gronemeyer, Fran#{231}oiseDieterlin, Simon Saule, Jean CoIl, JeanRommelaere, and Ji Hshiang Chen for stimulating discussions and critical

reading of the manuscript. We thank members of the Endocrin’os Group fordiscussions, support, and friendship.

References1 . Green, S., and Chambon, P. In: M. Parker (ed), Nuclear Hormone Re-ceptors, pp. 15-33. London: M. G. Academic Press, 1991.

2. Forman, B. M., and Samuels, H. H. Interactions among a subfamily ofnuclear hormone receptors: the regulatory zipper model. Mol. Endocrinol.,4: 1293-1301, 1990.

3. Evans, R. The steroid and hormone receptor superfamily. Science (Wash-ington DC), 240: 889-895, 1988.

4. Leid, M., Kastner, P., and Chambon, P. Multiplicity generates diversity inthe retinoic acid signalling pathways. Trends Biochem. Sci., 17: 427-433,1992.

5. Laudet, V., and St#{233}helin,D. Nuclear receptors: flexible friends. Curr.Biol., 2: 293-295, 1992.

6. Moore, D. D. Diversity and unity in the nuclear hormone receptors: aterpenoid receptor superfamily. New Biol., 2: 100-105, 1990.

7. Laudet, V., H#{228}nni,C., CoIl, J., Catzeflis, F., and St#{233}helin,D. Evolution ofthe nuclear receptor gene superfamily. EMBO J., 1 1: 1003-1 01 3, 1992.

8. Wilson, T. E., Fahrner, T. J., and Milbrandt, J. The orphan receptor NGFI-Band steroidogenic factor 1 establish monomer binding as a third paradigm ofnuclear receptor-DNA interaction. Mol. Cell. Biol., 13: 5794-5804, 1993.

9. Ellinger-Ziegelbauer, H., Hihi, A. K., Laudet, V., Keller, I-I., Wahli, W.,and Dreyer, C. FTZ-F1 -related orphan receptors in Xenopus laevis: transcrip-tional regulators differentially expressed during early embryogenesis. Mol.Cell. Biol., 14:2786-2797, 1994.

10. Gigu#{232}re,V., Tini, M., Flock, G., Ong, E., Evans, R. M., and Otulakowski,G. Isoform-specific amino-terminal domains dictate DNA-binding propertiesof RORcr, a novel family of orphan hormone nuclear receptors. Genes Dev.,8:538-553, 1994.

1 1 . Harding, H. P., and Lazar, M. A. The orphan receptor Rev-erbAa acti-yates transcription via a novel response element. Mol. Cell. Biol., 13: 31 13-3121, 1993.

12. Miyajima, N. N., Horiuchi, R., Shibuya, Y., Fukushige, S-I., Matsubara,K-I., Toyoshima, K., and Yamamoto, T. Two erbA homologs encoding pro-teins with different T3 binding capacities are transcribed from opposite DNAstrands of the same genetic locus. Cell, 57: 31-39, 1989.

13. Lazar, M. A., Hodin, R. A., Darling, D. S., and Chin, W. W. A novelmember of the thyroid/steroid hormone receptor family is encoded by theopposite strand of the rat c-erbAa transcriptional unit. Mol. Cell. Biol., 9:1128-1136, 1989.

Cell Growth & Differentiation 1365

14. Laudet, V., B#{232}gue,A., Henry-Duthoit, C., Joubel, A., Martin, P., Stfthelin,D., and Saule, S. Genomic organization of the human thyroid receptor a(c-erbA-1) gene. Nucleic Acids Res., 19: 1 105, 1991.

15. lzumo, S., and Mahdavi, V. Thyroid hormone receptor a isoforms gen-erated by alternative splicing differentially activate myosin HC gene tran-

scription. Nature (Lond.), 334: 539-543, 1988.

16. Koenig, R. J., Lazar, M. A., Hodin, R. A., Brent, G. A., Larsen, P. R., Chin,W. W., and Moore, D. D. Inhibition of thyroid hormone action by a non-hormone binding c-erbA protein generated by alternative mRNA splicing.Nature (Lond.), 337: 659-661 , 1989.

1 7. Katz, D., and Lazar, M. A. Dominant negative activity of an endogenousthyroid hormone receptor variant (a2) is due to competition for binding sites

on target genes. J. Biol. Chem., 268: 20904-20910, 1993.

18. Lazar, M. A., Hodin, R. A., Cardona, G., and Chin, W. W. Geneexpression from the c-erbAa/Rev-ErbAa genomic locus. J. Biol. Chem., 265:12859-1 2863, 1990.

19. Munroe, S. H., and Lazar, M. A. Inhibition of c-erbA mRNA splicing bya naturally occurring antisense RNA. J. Biol. Chem., 266: 22083-22086,1991.

20. Chawla, A., and Lazar, M. A. Induction of Rev-erba, an orphan receptorencoded on the opposite strand of the a-thyroid hormone receptor gene,during adipocyte differentiation. J. Biol. Chem., 268: 16265-16269, 1993.

21 . Bradley, 0. 1., Young, W. S., III, and Weinberger, C. Differential expres-sion of a and f3 thyroid hormone receptor genes in rat brain and pituitary.Proc. NatI. Acad. Sci. USA, 86: 7250-7254, 1989.

22. Amero, S. A., Kretsinger, R. H., Moncrief, N. D., Yamamoto, K. R., andPearson, W. R. The origin of nuclear receptor proteins: a single precursordistinct from other transcription factors. Mol. Endocrinol., 6: 3-7, 1992.

23. Dreyer, C., Keller, H., Mahfoudi, A., Laudet, V., Krey, G., and Wahli, W.Positive regulation of the peroxisomal b-oxidation pathway by fatty acidsthrough activation of peroxisome proliferator-activated receptors (PPAR).Biol. Cell, 77:67-76, 1993.

24. Hupp, T. R., Meek, D. W., Midgley, C. A., and Lane, D. P. Regulationof the specific DNA binding function of p53. Cell, 71: 875-886, 1992.

25. Graupner, R., Wills, K. N., Tzukerman, M., Zhang, x-K., and Pfahl, M.Dual regulatory role for thyroid hormone receptors allows control of retinoicacid-receptor activity. Nature (Lond.), 340: 653-656, 1989.

26. Leprince, D., Gesqui#{232}re,i-C., and St#{233}helin,D. The chicken cellularprogenitor of the v-eta oncogene, p68�’� , is a nuclear DNA-binding pro-tein not expressed in lymphoid cells of the spleen. Oncogene Res., 5:

255-265, 1990.

27. Keese, P. K., and Gibbs, A. Origins of genes: “Big Bang” or continuous

creation? Proc. NatI. Acad. Sci. USA, 89: 9489-9493, 1992.

28. Forrest, D., SjOberg, M., and VennstrOm, B. Contrasting developmentaland tissue-specific expression of a and f3 thyroid hormone receptor genes.EMBOJ., 9:1519-1528, 1990.

29. Lee, M. S., Kliewer, S. A., Provencal, J., Wright, P. E., and Evans, R. M.Structure ofthe retinoid x receptor a DNA binding domain: a helix required

for homodimeric DNA binding. Science (Washington DC), 260: 1 1 1 7-1 121,1993.

30. Zechel, C., Shen, �-Q., Chambon, P., and Gronemeyer, H. Dimerizationinterfaces formed between the DNA binding domains determine the coop-erative binding of RXRJRAR and RXR/TR heterodimers to DR and D elements.EMBO J., 13: 1414-1424, 1994.

31 . Zechel, C., Shen, �-Q., Chen, J-Y., Chen, Z-P., Chambon, P., andGronemeyer, H. The dimerization interfaces formed between the DNAbinding domains of RXR, RAR and TR determine the binding specificity andpolarity of the full-length receptors to direct repeats. EMBO J., 13: 1425-1433, 1994.

32. Spanjaard, R. A., Nguyen, V. P., and Chin, W. W. Rat Rev-erbAa, anorphan receptor related to thyroid hormone receptor, binds to specificthyroid response elements. Mol. Endocrinol., 8: 286-295, 1994.

33. Smith, J. C. Dorso-Ventral patterning in the neural tube. Curr. Biol., 3:

582-585, 1993.

34. Ingham, P. W. Hedgehog points the way. Curr. Biol., 4: 347-350, 1994.

35. Yamada, T., Placzek, M., Tanaka, H., Dodd, I., and Jessell, T. M. Controlof cell pattern in the developing nervous system: polarizing activity of thefloor plate and notochord. Cell, 64: 635-647, 1991.

36. Tessier-Lavigne, M., Placzek, M., Lumsden, A. G. S., Dodd, J., andJessell, T. M. Chemotropic guidance of developing axons in the mammaliancentral nervous system. Nature (Lond.), 336: 775-778, 1988.

37. Bovolenta, P., and Dodd, J. Perturbation of neuronal differentiation andaxon guidance in the spinal cord of mouse embryos lacking a floor plate:analysis of Danforth’s short-tail mutation. Development, 1 13: 625-639,1991.

38. Serafini, T., Kennedy, T. E., Galko, M., Mirzayan, C., Jessell, T. M., andTessier-Lavigne, M. The nestrins define a family of axon outgrowth-promot-ing proteins homologous to C. elegans UNC-6. Cell, 78: 409-424, 1994.

39. Kennedy, T. E., Serafini, T., de Ia Torre, J., and Tessier-Lavigne, M.Nestrins are diffusible chemotropic factors for commissural axons in the

embryonic spinal cord. Cell, 78: 425-435, 1994.

40. Sasaki, H., and Hogan, B. L. M. HNF-3� as a regulator for floor platedevelopment. Cell, 76: 103-1 15, 1994.

41 . Buckingham, M., and Tajbakhsh, S. C. R. Expression des facteurs myo-g#{233}niqueschez Ia souris: myf-5, premier membre de Ia famille des genesMyoD a #{234}tretranscrits au cours de Ia formation du muscle squelettique.Acad. Sci./Life Sci., 316: 1032-1039, 1993.

42. Maniatis, T., Fristch, E. F., and Sambrook, J. Molecular Cloning: ALaboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor LaboratoryPress, 1989.

43. Qu#{233}va,C., Ness, S. A., Gr#{228}sser,F. A., Graf, T., Vandenbunder, B., andSt#{233}helin,D. Expression pattern of c-myb and v-myb induced myeloid-l(mim-1) gene during the development of the chick embryo. Development,114: 125-133, 1992.

44. Rong, P., Ziller, C., Pena-Melian, A., and Le Douarin, N. M. A mono-clonal antibody specific for avian early myogenic cells and differentiatedmuscle. Dev. Biol., 122: 338-353, 1987.

45. Segraves, W. A., and Hogness, D. S. The E75 ecdysone-inducible generesponsible for the 75B early puff in Drosophila encodes two new membersof ther steroid receptor superfamily. Genes Dev., 4: 204-21 9, 1990.