Development 112, 933-943 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

933

Developmental and regional expression of thyroid hormone receptor genes

during Xenopus metamorphosis

AKIRA KAWAHARA*, BETTY S. BAKER and JAMSHED R. TATAf

Laboratory of Developmental Biochemistry, National Institute for Medical Research, The Ridgeway, Mill Hill, London NW7 1AA, UK

•Present address: Laboratory of Cell & Developmental Biology, Faculty of Integrated Arts & Sciences, Hiroshima University,Higashisenda-machi 1-1-89, Naka-ku, Hiroshima 730, Japant Author for correspondence

Summary

A characteristic feature of the obligatory control ofamphibian metamorphosis by thyroid hormones is theearly acquisition of response of tadpole tissues to thesehormones well before the latter are secreted, with'exponentially' increasing hormonal sensitivity upon theonset of metamorphosis. We have therefore analyzed theexpression of the two thyroid hormone receptor genes(TRa and ff) before, during and after metamorphosis inXenopus tadpoles and froglets. Using non-cross-hybrid-izing cRNA probes for 5' and 3' sequences of XenopusTRa and fi transcripts for RNAase protection assays, thetwo mRNAs can be detected in tadpoles as early as stage39. Their concentration increases abruptly at stage 44and continues to increase differentially at the onset ofmetamorphosis (stage 55) and through metamorphicclimax at stages 58-62, after which they decline uponcompletion of metamorphosis at stage 66. Quantitativedensitometric scanning of autoradiograms showed that,although the concentration of TR/? transcripts is aboutl/30th of that of TRa mRNA at stages 44-48, dependingon the region, it accumulates 3-10 times more rapidlythan does the a isoform during further development. Asubstantial proportion of the increase in TR/J mRNA islocalized to the head region of tadpoles. Using thehormone-binding domain (HBD) and 3' end of XenopusTRa cRNA as probe for in situ hybridization, the highestconcentration of TR transcripts in stage 44 tadpoles isseen in the brain and spinal cord. High concentrations ofmRNA are also present in the intestinal epithelium andtail tip, tissues programmed for regression. At later

stages (55 onwards), strong hybridization signals arealso exhibited by hindlimb buds. This pattern persiststhrough metamorphic climax, after which TR mRNAsdecline in all tissues to low levels in froglets at stage 66.In developing froglets, TR transcripts were detected inlarge amounts in the cytoplasm of stage 1 and 2 oocytesbut the rate of their accumulation did not increase withfurther oocyte growth. This observation raises thepossibility that the response to thyroid hormones at earlystages of tadpoles (42-44) may be due to TR synthesizedon maternally derived mRNA. Exposure of tadpoles atpremetamorphic stages (48-52) to exogenous thyroidhormone (T3) substantially enhanced the accumulationof TR mRNA, especially that of TR0 message, whichcould explain the accelerated increase in sensitivity oftadpoles to thyroid hormones at the onset of naturalmetamorphosis. This autoinduction by T3 of the mRNAfor its own receptor was hormone-specific and could notbe induced by ligands of other members of the nuclearreceptor family, such as oestrogen (E2) or retinoic acid(RA). We propose that stored maternal mRNA forthyroid hormone receptor and the phenomenon ofupregulation of its own receptor may explain theprecocious hormonal response and the 'exponential'kinetics of metamorphic climax characteristic of thisamphibian postembryonic developmental process.

Key words: Xenopus tadpoles, metamorphosis, thyroidhormones, TRcr and /3 gene expression, in situhybridization, localization of mRNA.

Introduction

Metamorphosis in both vertebrates and invertebrates isunder strict hormonal control (see Gilbert and Frieden,1981). In amphibia this postembryonic developmentalprocess is obligatorily induced and sustained by thyroidhormones, the deprivation of which results in perma-nent arrest at the tadpole stage (Weber, 1967;

Beckingham Smith and Tata, 1976; Tata, 1984).Another major characteristic of metamorphosis is thatvery different types of developmental changes areinitiated in different tissues by the same hormonalsignal. Thus, thyroid hormones control such diverseprocesses as induction of morphogenesis in limb buds,cell death underlying the regression of tail, gills andintestine, and extensive morphological and biochemical

934 A. Kawahara, B. S. Baker and J. R. Tata

remodelling of the liver, pancreas, skin and centralnervous system. Tissue culture studies have demon-strated that the hormone acts directly on individualtissues to produce this diversity of developmentalswitching (Tata, 1966; Derby and Etkin, 1968; Whiteand Nicoll, 1981; Kikuyama et al. 1983; Kawahara et al.1989; Mathisen and Miller, 1989).

In Xenopus laevis, the larval thyroid gland is firstactivated during the prometamorphosis period, ataround stage 54, when the two thyroid hormones,L-thyroxine and 3,3',5-triiodo-L-thyronine, are firstsecreted. The exposure of larval cells to the hormonesleads to metamorphic climax between stages 59 and 62.It is followed by a drastic reduction in hormonalsecretion at the froglet stage 66 when metamorphosis iscompleted (Leloup and Buscaglia, 1977; White andNicoll, 1981). A similar surge of circulating thyroidhormones during metamorphosis has also been de-scribed in other amphibian tadpoles (Gilbert andFrieden, 1981; Piotrowski and Kaltenbach, 1985).Although endogenous thyroid hormones first appear atabout stage 54, it is known that several larval tissuesacquire sensitivity to exogenous thyroid hormones wellbefore that, i.e. as early as stage 44, as determined bybiochemical and morphological criteria (Tata, 1968;Mathisen and Miller, 1989; Moskaitis et al. 1989). Thisfinding implies that functional thyroid hormone recep-tors must be expressed well before the hormone firstappears during normal development.

Thyroid hormone receptors belong to the proto-oncogene c-erb-A-related superfamily of genes en-coding nuclear receptors for other developmentallyimportant signals such as steroid hormones and retinoicacid (Evans, 1988; Green and Chambon, 1988; Carl-stedt-Duke et al. 1988; Hollenberg et al. 1989). It istherefore significant that, in a recent study usingheterologous cDNA probes, we observed a rapid build-up of c-erb-A like transcripts between stages 44 and 55Xenopus tadpoles (Baker and Tata, 1990). However,definitive conclusions could not be drawn about species-specific thyroid hormone receptor gene expression.Work from many laboratories on mammals, chickenand Xenopus has shown the existence of severalisoforms of thyroid hormone receptors encoded by twogenes termed TRar and TR/S (Weinberger et al. 1986;Sap et al. 1986; Forrest et al. 1990; Brooks et al. 1989;Cook and Koenig, 1990; Yaoita et al. 1990). What is ofparticular importance is that the a and /S isoforms inrodents and chicken are expressed in a developmentaland tissue-specific manner (Hodin et al. 1989; Koenig etal. 1989; Forrest et al. 1990). In view of this fact and ofthe diversity of tissue responses to the hormone inamphibia, it became crucial to investigate the tissuedistribution of both the a and fi TR genes duringXenopus development, particularly before and duringmetamorphosis.

In this paper, using homologous sense and antisensecomplementary RNA (cRNA) as probes for RNAaseprotection assays and in situ molecular hybridization,we describe the tissue distribution of TRa- and TR/Jtranscripts in Xenopus larvae before, during and after

metamorphosis. In all tissues, TRormRNA is present ina considerably higher concentration than the /3 isoformtranscript, being particularly elevated in the CNS,spinal cord, limb-bud, tail and intestine, the latter twoorgans being programmed for extensive cell death.After metamorphosis, the concentration of TRa-mRNA declined rapidly in all tissues, except that itaccumulated to very high levels in stages 1 and 2oocytes. Treatment of tadpoles with T3 stronglyinduced TR/3 mRNA, especially in the head region.Transcripts of two members of the c-erb-A nuclearreceptor superfamily, closely related to TR, namelyretinoic acid (RARa) and oestrogen (ER) receptors(Evans, 1988; Carlstedt-Duke et al. 1988), were alsoexamined. RARa- mRNA was also distributed at highconcentration in all tissues in stages 44-66 Xenopus,particularly in nervous tissues, but remained elevatedafter metamorphosis. ER mRNA could only bedetected in liver, kidney and skin gland at relatively lowconcentrations, pretreatment of tadpoles or frogletswith oestrogen causing an upregulation of mRNA forits own receptor. These findings are discussed in thecontext of early acquisition of response of Xenopuslarvae to the metamorphic effects of thyroid hormones.

Materials and methods

AnimalsAdult male and female Xenopus laevis, purchased from BladeBiological, Cowden, Kent, UK, were maintained at 22±1°Cand fed once a week. Breeding pairs were injected twice at 8 hintervals with 100 and 500 units (females) and 50 or 150 units(males) of human chorionic gonadotrophin (Paines and ByrneLtd, London, UK). Fertilized eggs and embryos until stage 36were maintained in 10 % Barth-X solution, in 5 % Barth-Xuntil stages 42-45, and after that in distilled water. Tadpoleswere fed on Complan. Embryos and larvae were stagedaccording to Nieuwkoop and Faber (1967).

Materials[35S]UTP (1000-1500 CimmoF1) and placental ribonucleaseinhibitor were from Amersham International (Amersham,UK); SP6 and T7 RNA polymerases were from New EnglandBiolabs (Beverly, MA, USA) and Pharmacia (Uppsala,Sweden), respectively; RNAase-free DNAase I was pur-chased from Boehringer (Mannheim, Germany); all nucleo-tides and other reagents were purchased from Sigma (Poole,UK) or BDH (Poole, UK).

Plasmids and cRNA probesGoned cDNAs for Xenopus TRa and (3 mRNAs were kindlyprovided by Dr R. Old (University of Warwick, UK) and DrD. D. Brown (Carnegie Institution, Baltimore, USA),respectively. Professor P. Chambon (Universite Louis Pas-teur, Strasbourg) provided us with cloned cDNAs to humanRARo-, Professor B. Vennstrom (Karolinska Institute,Stockholm) gave us the chicken TRar and fi clones, while thecDNA to Xenopus ER (pXERO) was a gift of Professor D.Shapiro (University of Illinois, Urbana, USA). Descriptionsof these clones have been previously published (Brooks et al.1989; Yaoito et al. 1990). For RNAase protection assays,antisense cRNAs were synthesized and labelled with 32P andfor in situ hybridization the sense and antisense cRNAs were

Developmental localization of Xenopus TRa and /3 mRNA 935

342HhcII

ptKXERO

tk promoter

Fig. 1. Diagrammaticrepresentation of fragments ofcloned cDNAs to Xenopus TRaand P, and ER mRNAs togenerate cRNA probes forRNAase protection assay(fragments I and IV) and in situhybridization (fragments II, III,V). HBD: hormone-bindingdomain (•); DBD: DNA-binding domain (S); The lengthof the fragments and theirposition is given by nucleotidenumbers and restriction sites.Cloning sites ( • ) are alsoindicated. Details of these andother probes are given inMaterials and methods.

synthesized with [35S]UTP, as described by Melton et al.(1984). In most cases, only the ligand binding domain (HBD)with or without the 3' untranslated region were excised byrestriction digestion, but for some probes the full-lengthcDNAs were used. For these studies, the following plasmidswere cloned in pGEM3 (Promega Corp., Madison, WI,USA): human RARa- (HBD): 1185bp of Sacl-EcoRIfragment (Petkovich et al. 1987); Chicken TRa (HBD):1300 bp of EcoRl-BaK fragment of c-erb-A clone Fl chickenTR£ (Sap et al. 1986): ~1 kb of EcoRl-Bglll insert of cloneT10 Xenopus ER (HBD) (Forrest et al. 1990): 792 bp ofHindlU fragment of plasmid TkXERO (Chang and Shapiro,1990). The following plasmids were cloned in Bluescript(Stratagene, La Jolla, CA, USA). Xenopus TRa (HBD):549bp of HindlU-Sacl fragment of xenTRa^l plasmid(Brooks et al. 1989); Xenopus TRa (3' end): 1181 bp of//ircdm-EcoRI fragment of xenTRa--l plasmid (Brooks et al.1989);. Xenopus TRfi (305bp of HaeUl-EcoRI fragment ofXT3R£(A1) (Yaoita et al. 1990); Xenopus cardiac actin: the328bp RSAI fragment pSpol (Mohun et al. 1988). The HBD,3' and 5' end fragments of Xenopus TRa, TR/3 and ER usedfor generating cRNA probes for in situ hybridization areschematically depicted in Fig. 1.

RNA extractionTotal RNA from individual tissues or from the head, middle(containing liver, kidney, intestine and pancreas) and tailregions (for early stages of tadpoles) was prepared by theLiCl-urea procedure (Baker and Tata, 1990). Poly(A)+ RNAwas separated by column chromatography on oligo(dT)(Maniatis et al. 1982). The extent of degradation of RNA waschecked by gel electrophoresis, and northern blot hybridiz-ation with Xenopus cytoplasmic actin cDNA as probe(Mohun and Garrett, 1987).

RNAase protection assayThe presence and relative amounts of TRa- and p mRNAs wasestablished by RNAase protection assay, as described by Zinnet al. (1985) using the 261 nucleotide Hincll-Pstl and the 305nucleotide //aeIII-£coRI (protected fragment of 260 nt)fragments of Xenopus TRa and /S cDNAs, respectively, forprotection of the two mRNAs (Fig. 1). The procedure wasvalidated by protection with Xenopus cytoplasmic actincDNA and comparing the results with those obtainedpreviously (Mohun and Garrett, 1987). In order to quantifythe relative distribution of TRa-and TR/3 mRNAs, autoradio-grams were scanned in a Joyce-Loebl double-beam integrat-

ing densitometer, under conditions in which the intensity ofautoradiographic bands was linearly related to the amount ofradioactivity and restricting the comparison of differentsamples to those on the same autoradiogram. The reason forthe multiplicity of bands of higher mobility than the majorprotected band, a characteristic feature of RNAase protectionassays, is not known. However, with few exceptions, thesmaller protected RNA bands varied proportionately to themajor band. Therefore, in making quantitative comparisons,bands representing ~90 % of total radioactivity loaded on thegel for each RNA sample were grouped together fordensitometric scanning.

Filter hybridization analysisTotal RNA samples were used for. RNAase protection andslot-blot hybridization assays with cRNA and cDNA probesfor Xenopus TRa and /3 mRNAs, as well as the HBDs of ERand RARa mRNAs. Poly(A)+-RNA was used for northernblot hybridization analysis, both procedures followed asdescribed previously (Baker and Tata, 1990). Multiple bandswere observed for both TRa and /3 mRNAs from varioustissues or regions and at all developmental stages, in line withthe presence of several isoforms of TR transcripts in otherspecies (Hodin et al. 1989; Koenig et al. 1989; Forrest et al.1990). A transcript bigger than 10 kb was also detected forboth Xenopus TRa and fi transcripts (Yaoita et al. 1990).

In situ hybridizationThe procedure for fixing, embedding and sectioning wholetadpoles, individual segments or tissues was as for mouseembryos and, as with the procedure for in situ hybridization,was essentially the same as described by Wilkinson and Green(1990). High-stringency conditions for hybridization werefollowed for Xenopus-s^cdfic probes with post-hybridizationwashing in 2xSSC - 50 % formamide - 50 mM DTT at 65° for30min. With heterologous probes (chicken TRa-and /S humanER and human RARa-), the signal was greatly diminishedwith post-hybridization washing under these stringent con-ditions; washing at 60°C for 20min, however, allowed thedetection of specific signals with the antisense probe, with lownon-specific hybridization with the sense probe. Autoradi-ography was carried out with 40% Ilford K5 nuclear trackemulsion, 1 % glycerol - 60% H2O and developed in KodakD19 developer, as described elsewhere (Wilkinson andGreen, 1990). Autoradiographs were followed by nuclearstaining with toluidine blue.

936 A. Kawahara, B. S. Baker and J. R. Tata

Results

Biochemical analysis of TRa and B mRNAsSince virtually every tissue in the tadpole is known torespond to thyroid hormones and that response toexogenous hormone is elicited precociously, it wasexpected that thyroid hormone receptors would bewidely distributed in early larval tissues. Fig. 2 showssemi-quantitatively by RNAase protection assay theappearance of TRa- and TR/J mRNAs during Xenopusdevelopment using Xenopus-specific probes. Pre-viously, using heterologous (chicken) c-erb-a and Bprobes for filter hybridization, we had shown that thecorresponding Xenopus transcripts accumulated differ-entially and that the a-mRNA was present in at least 10-fold higher concentration than the /S homologue (Bakerand Tata, 1990). It is therefore significant, from theRNAase protection analysis depicted in Fig. 2, that theXenopus probe also revealed that the a isoform of TRmRNA in all three regions of developing and metamor-phosing tadpoles was present in considerable excessover the B form. Quantification by double-beamdensitometry, under conditions in which the autoradio-

St 28 39 44 48 p—52—, - 5 6 - r

h m t h m tr58ih m t

rh

•60- ]1 t

r62ih m t l i Pr

mnt

1 2 9 9

<261

B

h m t

S t 28 39 44 48 L-52—'

h m t h m t

6 0 - 1 L62-J

graphic intensities can be directly compared, showedthat up to stage 48 TRa-mRNA was about 30 times inexcess over TR/3 mRNA. It can also be seen thatsubstantial variations in regional concentrations werefound for TRa-and TR/? mRNAs. For example, at stage52 the ratio of a/B mRNAs was 24, 78 and 57 for head,middle and tail regions, respectively. During furtherdevelopment, the amount of both TRa- and TR/JmRNA increased but to different extents in differentregions. That of TRa- transcripts had accumulated tohigh levels well before metamorphosis, and thenincreased only slightly at metamorphic climax and atlate stages of the process. TR/? mRNA was present athigher concentrations in the head region at all stagesuntil the end of metamorphosis. Densitometric analysisshowed a continuous increase in the relative amounts ofTR/? mRNA during prometamorphosis and climax.Thus, the ratio of a/B mRNAs at stage 56 in head,middle and tail regions dropped to 8, 16 and 10,respectively, and even further to 1, 4 and 2 at stage 62.The concentration of both isoforms of mRNA wassharply reduced in all tissues upon completion ofmetamorphosis (data not shown).

Fig. 2. Detection by RNAaseprotection assay of TRa- and fitranscripts in different regionsand tissues of Xenopustadpoles and froglets atdifferent developmental stagesbefore, during and aftermetamorphosis. 10 (C) or 20(A,B) ng of total RNA fromwhole tadpoles (stages 28, 39,44, 48), the head (h), middle(m) and tail (t) portions oftadpoles, hindlimbs (li) offroglets and liver (1) of adultXenopus were hybridized tolxlCPctsmin"1 of 32P-labelledcRNA probes (Pr) for (A)TRa- cRNA, (B) TR/3 or (C)cardiac actin cRNA.Unhybridized RNA and probewere digested with RNAase Aand Tl and the protectedfragments resolved byelectrophoresis on sequencinggels with appropriate sizemarkers (Zinn et at. 1985).tRNA was used as a negativecontrol (not shown) and thecomplete digestion of theunprotected probe was verifiedfor each run. Autoradiogramswere exposed for 16 h for Aand B and for 30min for C.Arrow indicates the cRNAprobe lengths and the expectedsize of the protected probes(260, 261 nt, respectively) forTRa-, TR0 CRN As (see Fig. 1).The multiple protected bandsof actin cRNA probe areindicated by small arrows in C.

1325

1260

1250

li Pr

Developmental localization of Xenopus TRa and ft mRNA 937

The procedure of RNAase protection that we haveadopted would not allow us to distinguish between theindividual transcripts of the two TR/J genes that havebeen reported for Xenopus (Yaoita et al. 1990). Sincelarge amounts of RNA (usually 20jtg), which could beaccurately measured optically, were used for thisprocedure, it was not essential to normalize sampleswith a probe for another Xenopus-sp&cific RNA. Wehave failed to find a probe for Xenopus RNA whoseconcentration is invariable in different tissues atdifferent developmental stages. However, the validityof comparison of TR mRNA in different regions andtissues of Xenopus tadpoles is confirmed by theconverse pattern of distribution of actin mRNA in thesame RNA samples obtained at different developmen-tal stages (Fig. 2C). Actin transcripts, which are moreabundant in muscle-rich tissues, are found at higherconcentration in tail and limbs and this pattern does notchange with the onset of metamorphosis as seen for TRmRNA.

Localization of TR transcripts by in situ hybridizationThe above findings of relative TR mRNA distributionwere confirmed and extended by in situ hybridizationanalysis, in order to derive a more accurate picture oftissue distribution of the receptor transcripts.Fig. 3A-C shows the localization of TR mRNA in threebody regions of the early tadpole stage of 44, using anantisense cRNA probe of the 3' end incorporating thehormone-binding domain of Xenopus TRa transcript.Since the antisense HBD containing fragment ofXenopus TRa (fragment III, Fig. 1), which is highlyconserved between the different TR genes, was used asa probe, the hybridization signals would depict thelocalization of all isoforms of TRo- and /3 transcripts.However, the signals obtained, because of much lowerabundance of TR/3 transcripts and 3' end untranslatedregion which differs from that of TR/3, would largelyrepresent the distribution of TRa transcripts. In fact,both the Xenopus and the heterologous chicken TR/3cRNA probes did not produce any hybridization signalin Xenopus larval tissues. Although virtually everylarval tissue gave a positive hybridization signal(relative to the sense probe), the highest signal strengthwas detected in the central nervous system andepithelial tissues or organs (intestinal epithelium,pharyngeal floor epithelium, nasal epithelium, liver,kidney and skin), with moderate signal strengthdetected in heart, tail muscle, intestinal muscle andskeletal muscle, with the lowest hybridization in almostall connective tissues. In summary, TR transcriptsaccumulate to high levels in larval epithelial tissues andneural tissues, including eye retina (data not shown),while muscular and connective tissues contain arelatively lower amount of this transcript.

The hybridization signal strength of TR transcriptschanged with development. As summarized in thecomposite images in Fig. 4 for stages 44, 54/58(prometamorphosis) and 62 (postmetamorphic climax),the extent and intensity of hybridization signalincreased with development, except that the relative

tissue distribution was not greatly altered. However, atabout stage 55 when hindlimb buds could be well-discerned, there was a substantial accumulation of TRmRNA in this organ (Fig. 3D), and also the liver(Fig. 4). However, during metamorphic climax (stages59-64), there was a decline in the relative distributionof these transcripts in the central nervous system(Fig. 4).

As regards the possible differential localization ofmRNAs of the a and fl isoform of TR, several attemptswere made to localize'the possible differential distri-bution of these transcripts, but the signals from TR/3mRNA were too weak to draw firm conclusions abouttissue distribution or developmental variations. Furtherwork in improving the sensitivity of the technique willbe necessary to obtain definitive results in thisdirection.

In situ hybridization analysis was also extended tostages well beyond the completion of metamorphosis.In general, TR mRNA concentration decreased in alltissues that underwent substantial remodelling, such asthe brain, liver, heart, etc. (data not shown). Therewas, however, one exception. Rather surprisingly, thedeveloping gonads of froglets showed considerableaccumulation of TR transcripts, which increased withdevelopment. This pattern of expression was restrictedto the ovary (phenotypic sex in many anurans isestablished upon completion of metamorphosis(Witschi; 1956; Deuchar, 1975). Upon closer examin-ation, the accumulation of mRNA was restricted to thedeveloping oocytes, as shown in Fig. 5 for stages 1 and 2oocytes. The accumulation of TR transcripts in oocyteswas sustained and continued to increase through furtherdevelopment.

Other receptor transcripts and hormonal auto-upregulation of mRNAAmong the different members of the c-erb-A relatedsteroid/thyroid family, the mRNAs encoding oestrogenand retinoic acid receptors bear the closest homology toTR mRNA. Since the 3' hormone-binding domains ofreceptors for the different ligands are substantiallydifferent while the sequences for any given ligand areconserved across species (Evans, 1988; Green andChambon, 1988; Carlstedt-Duke et al. 1988; Weiler etal. 1987), we carried out hybridization analysis forXenopus development with probes covering only theHBDs of human ER and RARa cDNAs (data notshown). Compared with the results described above forTR mRNAs, transcripts for RAR (using a heterologousRARa- probe) are more abundant and more homo-geneously distributed in all regions of tadpoles andfroglets from stages 43 through 66. It was in thedeveloping oocytes that the quantitative differences inthe signal from the two transcripts were most obvious;the level of RAR mRNA was estimated to be about 10times that of TRa- mRNA, which in turn would be atleast another order of magnitude higher than that of theTR/J isoform. In contrast, the distribution of ERmRNA was quite different from that of the above two.In stages 57-60 tadpoles, treated with oestrogen to raise

938 A. Kawahara, B. S. Baker and J. R. Tata

the level of ER mRNA, its tissue distribution did notcorrespond to that of TR and RAR mRNAs, and wasnot detectable in target tissues for T3 such as the brain,tail and intestine. The difference in distribution of ERmRNA from that of TR transcripts confirms that theresults described above for TR mRNA are not due to anon-specific hybridization of the cRNA probes amongdifferent members of the c-erb-A related nuclearreceptor transcripts.

A well-known feature of amphibian metamorphosisis the rapidly increasing sensitivity of larval tissues tothyroid hormones at the onset of metamorphosis whenthese hormones begin to appear in the blood (Leloupand Buscaglia, 1977; Beckingham Smith and Tata, 1976;Gilbert and Frieden, 1981). It was therefore of someimportance to test the possibility that exposure ofXenopus tadpoles to exogenous thyroid hormones justbefore the secretion of endogenous hormones mightaccelerate the accumulation of thyroid hormone recep-tors. This possibility was reinforced by the RNAaseprotection assays shown in Fig. 6. Exposure of stage 52tadpoles to 2X1CT9M T3 for 6 days caused a substantialincrease in the accumulation of both TRa and TR/3transcripts, but that of the latter was more marked.Although all regions of the tadpole exhibited theelevated levels of TR/SmRNA, it was in the head regionthat the auto-upregulation of this transcript was mostnoticeable following T3 treatment. Thus, densitometricanalysis showed that in T3-treated stage 52 tadpole headregion, the ratio of TRa/TR/3 dropped to a value of 2from that of 29 in untreated controls.

Finally, it is important to emphasize the use ofheterologous probes for receptor mRNAs and thecomparison of such data with those obtained withhomologous probes. In general, at least qualitatively,similar results were obtained with Xenopus, human andchicken probes that were restricted to the hormone-binding domains or the 3' end of the mRNA forreceptors for T3, oestrogen and RA. Only in oneinstance did we notice an exception while comparingthe localization of TR mRNA in oocytes by in situhybridization with the hormone-binding domain ofXenopus and chicken TRarcRNAs as probes. With thehomologous probe, TRarmRNA was found to accumu-late, as would be expected, only in the cytoplasm of thedeveloping Xenopus oocytes (Fig. 5), but, surprisingly,in sections of the same ovaries the chicken TRarcRNArecognized transcripts in both the nuclear and cytoplas-mic compartments (data not shown).

Discussion

A major characteristic of metamorphosis, in vertebratesand invertebrates, is that each organ undergoes adifferent pattern of preprogrammed morphogeneticand functional changes in direct response to themetamorphic hormonal signal (Gilbert and Frieden,1981). Earlier studies on cultured tail, liver and skin ofXenopus confirmed the diversity and tissue specificity ofresponse to thyroid hormones (Tata, 1966; Beckingham

Fig. 3. Dark-field imaging of localization by in situhybridization of TR mRNA in different tissues of stage 44(premetamorphosis) and 55 (prometamorphosis) Xenopustadpoles. Sagittal sections were prepared from differentregions of tadpoles hybridized to a mixture of antisense35S-labelled cRNA probe 3 (lxK^ctsmin" V " 1 ) , asshown in Fig. 1, unhybridized probe digested with RNAaseand the distribution of transcripts visualized byautoradiography. Sense probes gave virtually nohybridization and the autoradiographs are not shown here.(A) Head region of stage 44 tadpole; (B) middle or trunkregion, stage 44; (C) tail, stage 44; (D) trunk and tail withhindlimb bud in stage 55 (prometamorphosis). Left andright sides of each image are the ventral and dorsal sides ofthe tadpoles, respectively. Bright-field micrographs werealso prepared (not shown) and used for identifying thedifferent organs and tissues for which the abbreviations areas follows: Br, brain; Ct, connective tissue; Gi, gill; He,heart; Hlb, hindlimb bud; In, intestinal epithelium; Ki,kidney; Li, liver; Mu, muscle; No, notochord; Pf,pharyngeal floor epithelium; Ph, pharyngeal cavity; Sc,spinal chord; Sk, skinj.Tf, tail fin; Tm, tail muscle. Notethat pigmented layers (Pi) and food particles (Fp) appearas bright zones, also seen with sense probe controls.Autoradiograms exposed for 9 days.

Smith and Tata, 1976; White and Nicoll, 1981;Kawahara et al. 1989; Mathisen and Miller, 1989).Other studies on precocious induction by T3 ofmetamorphosis in Xenopus tadpoles at the early stagesof 44-48, as opposed to normal metamorphosis which isnot initiated until several weeks later at stage 57,suggested an early developmental acquisition of thyroidhormone receptors (Tata, 1968; Mathisen and Miller,1989; Moskaitis et al. 1989). It is therefore significantthat using RNAase protection assay we could clearlyidentify the a and f3 forms of TR mRNA from stage 44tadpole onwards (Figs 2 and 6). Although RNAaseprotection can at best be semiquantitative, TRamRNA, at all developmental stages, accumulates to ahigher extent (up to 50 times) than the ft isoform. Also,in line with an earlier study based on filter hybridizationwith heterologous probes (Baker and Tata, 1990) andthat in chick embryonic development (Forrest et al.1990), it is the /S transcript that seems to be develop-mentally regulated with greater differences in tissue orregional distribution than noted for the a mRNA(Figs 2 and 6).

The high concentration of TR transcripts, in the headregion of tadpoles at different developmental stageswas confirmed by in situ hybridization analysis, as canbe judged from the high signal strength observed in thetadpole brain and spinal chord with Xenopus TRacRNA (Figs 3-5). Because of the relatively lowconcentration of the TR/3 mRNA in Xenopus tadpoleand adult tissue, it was difficult to visualize thedistribution of TR/3 mRNA by the same procedure. It isknown that the amphibian CNS undergoes extensivestructural and functional reorganization during naturaland exogenous thyroid hormone-induced metamorpho-sis (Fox, 1981; Kollros, 1981; Leone et al. 1976;Deuchar, 1975; Gilbert and Frieden, 1981; Hauser and

1

(

. K

u0

•

. ~-' • • • ~ n -

* •

<

•

^^^^HB^^V. I >

• f

)

'•̂ ^

o

EE

CO

0 1 2mm

Fig. 5. Localization of TR mRNA in stages 1 and 2 oocytes visualized in sectionsthrough Xenopus froglet ovary hybridized to Xenopus TRa cRNA. (A) Antisense probe;(B) sense probe. Autoradiograms exposed for 3 and 9 days for antisense and senseprobes, respectively.

Developmental localization of Xenopus TRa and fi mRNA 939

Gi

yl

Hlb

1

-Tm

Ph

He

Hlc

B

0 0 15mm.

Fig. 4. Composite assembly of sagittal sections of (A) bright- and (B) dark-field imaging of Xenopus tadpoles at differentdevelopmental stages to show the localization of TR mRNA using the labelled TRa cRNA as probe. (1) Stage 44; (2) stage56 (prometamorphosis); (3) stage 62 (late-metamorphic climax). The three stages shown are to the same scale.Abbreviations: Ne, nasal epithelium; Oe, oesophagus; Gb, gall bladder; Go, gonad; Hlc, hindlimb connection. Otherabbreviations as in Fig. 3.

940 A. Kawahara, B. S. Baker and J. R. Tata

T3- - _ + P r

E 2 - - - - - - + + +h m t h m t h m t

Pr nt[299

TRoc

T 3 - - - 325

h m t h m t h m t

Fig. 6. Preferential autoinduction of TR/3 mRNA by T3 indifferent regions of Xenopus tadpoles, as revealed byRNAase protection assays. 20 fig of RNA from head,middle and tail regions of stage 52 tadpoles, treated or notwith 2X10~9M T3 or 10~8M oestrogen (E2) were hybridizedto 32P-labelled TRo- or TR/3 cRNA probes and analyzed asdescribed in Fig. 2. Autoradiograms of TRar and TR/?probes were exposed for 16 and 40 h, respectively. AHabbreviations and sizes of probes and protected fragmentsalso as in Fig. 2.

Gona, 1984). Also, thyroid hormones play a crucial rolein the maturation of mammalian brain and nervoussystem during foetal and perinatal development (Mor-reale de Escobar et al. 1983; DeLong et al. 1989).Therefore, the finding of high concentrations of TR/SmRNA in a selective manner in rodent and avian brainis highly relevant to our findings (Hodin et al. 1989;Bradley et al. 1989; Koenig et al. 1989; Cook andKoenig, 1990; Forrest et al. 1990).

Besides the CNS, our in situ analysis has revealedhigh levels of accumulation of TR transcripts in theXenopus tadpole digestive tract, particularly the intes-tine, and tail (Figs 3 and 4). It is noteworthy that thesetwo tissues are programmed for cell death andregression which are a major feature of amphibian

metamorphosis, these tissues also exhibiting an extremesensitivity to thyroid hormones (Gilbert and Frieden,1981; Tata, 1966; Beckingham Smith and Tata, 1976;Derby and Etkin, 1968). Another characteristic featureof amphibian metamorphosis is the growth of limb budsduring prometamorphosis and the eruption of hind-limbs, followed by that of forelimbs, subsequent tothyroid hormone secretion at around stage 54. Hind-limb buds are easily distinguishable after this stage andit is therefore significant to note the accumulation of TRmRNAs in this tissue at stage 55 (Fig. 3D). Amongother tissues exhibiting significant accumulation of TRmRNAs are the retina and pancreas. These two tissuesare known to undergo considerable structural andfunctional reorganization, as, for example, the tran-sition from aquatic to terrestrial visual pigmentationand the acquisition of new pancreatic enzymes,respectively (Gilbert and Frieden, 1981; BeckinghamSmith and Tata, 1976). While considering such prefer-ential distribution, lower but reproducible concen-trations of TR transcripts were detected in most othertissues, which is in line with the fact that virtually alltadpole tissues undergo to some extent structural orfunctional alteration during metamorphosis (Deuchar,1975; Gilbert and Frieden, 1981).

In the developmental context, the finding of TRtranscripts at tadpole stages as early as 39 is alsosignificant. It is relevant to the observations madeseveral years ago, well before the advent of techniquesfor assaying nuclear receptors or their transcripts, thatexposure of stage 43 Xenopus tadpoles to exogenous T3provoked several biochemical changes characteristic ofmetamorphosis (Tata, 1968). Furthermore, this earlyexpression of TR genes in the tadpole, coupled with theprecocious acquisition of responsiveness to thyroidhormones, indicates the developmental importance ofreceptor formation in postembryonic development. Ofparticular interest is the substantial accumulation of TRmRNAs in the tadpole tail and gut between stages 44and 54 (Figs 3 and 4). Although these tissues wouldundergo cell death and regression under the control ofthyroid hormones during normal metamorphosis, TRtranscripts have already accumulated to substantiallevels while these tissues are still actively growing inpremetamorphosis stages.

Upon completion of metamorphosis at stage 66,thyroid hormone synthesis and secretion are drasticallyreduced (Leloup and Buscaglia, 1977; White andNicoll, 1981; Piotrowski and Kaltenbach, 1985), and nophysiological response of froglet or adult tissue to thesehormones has yet been described. Our observation thatTR mRNA concentrations also started to decline inpostmetamorphic animals is therefore relevant to thesefacts. The only exception to this observation was thedeveloping oocyte in Xenopus froglet ovary in whichhigh levels of TR transcripts were detected (Fig. 5). Itwill be interesting to follow up this finding with thedetection and measurement of stored receptor protein,not only in the oocyte but also in tadpole tissues prior tometamorphosis, once techniques of sufficiently highsensitivity for thyroid hormone reception are devel-

Developmental localization of Xenopus TRa and fi mRNA 941

oped. Thus, it can be argued that the primary responseto thyroid hormones, especially when administered atearly premetamorphic larval stages, is due to receptorsof maternal origin, whose interaction with the hormonewould lead to the activation of the tadpole TR genes.We realize that detection of a receptor transcript in agiven tissue cannot be automatically equated with thepresence in those cells of functional receptor. However,the fact that the accumulation of TR mRNA in manyorgans and tissues of Xenopus tadpoles at early stagescan be correlated with a biochemical response tothyroid hormones at the same early stages (Tata, 1968;Mathisen and Miller, 1989; Moskaitis et al. 1989)suggests a correlation, at least qualitatively if notquantitatively, between the presence of TR mRNA andfunctional TR protein. Also, many studies on thedevelopmental or pharmacological modulation of avariety of steroid/thyroid hormone nuclear receptorgene family have established a semiquantitative re-lationship between receptor mRNA and functionalreceptor in diverse types of cells and organisms(Carlstedt-Duke et al. 1988; Sheridan et al. 1988).

On the basis of nucleotide and amino acid sequence,the steroid/thyroid hormone receptor supergene familyhas been divided into two major subclasses (Evans,1988; Green and Chambon, 1988; Carlstedt-Duke et al.1988). One comprises receptors for gfucocorticoids,androgen and progesterone, while the other includesreceptors for oestrogen, thyroid hormone and retinoicacid. It is therefore of some interest that, in ourpreliminary studies, RAR transcripts were found to bedistributed in tadpole tissues, as well as in stage 1 and 2oocytes, in a manner analogous to TR mRNA (data notshown). Although the exact significance of this simi-larity of tissue distribution is not fully understood, it isworth recalling that TR and RAR can be co-expressedin some cells and that the two receptors or theirrespective HBDs, can replace one another in activationof target genes or interaction with the heterologousresponsive elements in cell transfection experiments(Graupner etal. 1989; Glass et al. 1989). The similarityin receptor distribution may also simply reflect the factthat both thyroid hormones and retinoic acid areimportant morphogenetic signals in amphibia (Rags-dale et al. 1989; Durston et al. 1989). Conversely,despite their close relationship the distribution of TRand ER mRNAs in Xenopus tadpole tissues and theirdevelopmental expression are very different. By in situhybridization and biochemical analysis (data notshown), it was possible to localize ER mRNA only instage 57 onwards tadpoles, in line with earlier obser-vations (Baker and Tata, 1990). Under these conditionsthe distribution of ER transcripts was restricted to a fewtissues known to be targets for oestrogen (liver, kidney,skin).

The requirement for pretreatment with oestrogen tovisualize ER mRNA was explained by the property ofthe ligand to induce strongly the mRNA for its ownreceptor (data not shown), a property related to theupregulation of ER protein or mRNA by the hormoneseen earlier in adult Xenopus and rat tissues that are

targets for oestrogen (Perlman et al. 1984; Barton andShapiro, 1988; Shapiro et al. 1989; Shupnik et al. 1989;Varriale and Tata, 1990). This property of autoinduc-tion of receptor is not unique to oestrogen. Indeed, wefound that T3 was also able to upregulate its ownreceptor mRNA before normal metamorphosis wouldbegin (Fig. 6). Previously, TR, or a T3-binding activity,has been found to increase in erythrocytes, but not inliver nuclei, during metamorphic climax in Ranacatesbeiana tadpoles (Moriya etal. 1984; Galton and St.Germain, 1985). It would be important to determine ifthe Xenopus ER and TR promoters contain oestrogenand thyroid hormone responsive elements (ERE andTRE), respectively, analogous to a functional RAresponsive element recently described for the RARaand /3 gene promoters (deThe et al. 1990; Clifford et al.1990; Leroy et al. 1991). In any event, our presentfindings and unpublished data show, firstly, that theproperty of autoinduction of TR and ER is establishedduring premetamorphosis and prometamorphosis, re-spectively, and, secondly, that small amounts offunctional TR and ER must already be present in sometissue by these developmental periods. The phenom-enon of autoinduction of receptor may be a feature ofdevelopmental systems that are hormone-regulated,since generally in adult tissues most hormones tend todownregulate their own receptor (Carlstedt-Duke et al.1988). Furthermore, the developmental modulation ofreceptors restricted to only some tissues merits furtherinvestigation. Since several tadpole tissues responddirectly in organ culture to thyroid hormones, we areextending our present studies in whole animals to tissueculture in order to derive more precise information onthe developmental autoactivation of the receptor genes.

After our present studies were completed, Yaoitaand Brown (1990) recently reported a correlationbetween the expression of Xenopus TRarand fi mRNAsand progression of metamorphosis. Using a procedureof cDNA synthesis and primer extension, they observeda higher amount of TRa than TR/3 transcripts at allstages up to completion of metamorphosis. They alsofound that TR/J gene expression followed more closelythe progression of metamorphosis and that T3 adminis-tration to tadpoles upregulated this transcript to agreater extent than the a mRNA. Unlike our findingswith in situ hybridization, they failed to detect TRmRNA in oocytes, for which we have no explanation.

We are most grateful to the following colleagues for givingus the cloned cDNAs from which the probes used in this studywere derived: Dr R. W. Old (University of Warwick) forXenopus TRar, Dr D. D. Brown (Carnegie Institution,Baltimore) for Xenopus TR/3; Professor B. Vennstrom(Karolinska Institute, Stockholm) for chicken c-erb-A andTR/J; Professor D. J. Shapiro (University of Illinois, Urbana)for Xenopus ER; Professor P. Chambon (Universite LouisPasteur, Strasbourg) for human RAR<r, Dr T. J. Mohun(National Institute for Medical Research, London) forXenopus cardiac actin. We would also like to thank Dr D. G.Wilkinson and Dr T. J. Mohun for advice and practical helpwith some of the experimental work in this study.

942 A. Kawahara, B. S. Baker and J. R. Tata

References

BAKER, B. S. AND TATA, J. R. (1990). Accumulation of proto-oncogene c-erb-A related transcripts during Xenopusdevelopment: association with early acquisition of response tothyroid hormone and estrogen. EMBO J. 9, 879-885.

BAJTTON, M. C. AND SHAPIRO, D. J. (1988). Transientadministration of estradiol-17/3 establishes an autoregulatoryloop permanently inducing estrogen receptor raRNA. Proc.natn. Acad. Sci. U.S.A. 85, 7119-7123.

BECKINGHAM SMITH, K. AND TATA, J. R. (1976). Amphibianmetamorphosis. In Developmental Biology of Plants andAnimals (ed. C. F. Graham and P. F. Wareing) pp. 232-245.Oxford: Blackwell.

BRADLEY, D. J., YOUNG III, W. S. AND WEINBERGER, C. (1989).Differential expression of a and /S thyroid hormone receptorgenes in rat brain and pituitary. Proc. natn. Acad. Set. U.S.A.86, 7250-7254.

BROOKS, A. R., SWEENEY, G. AND OLD, R. W. (1989). Structureand functional expression of a cloned Xenopus thyroid hormonereceptor. Nucl. Acids Res. 17, 9395-9405.

CARLSTEDT-DUKE, J., ERIKSSON, H. AND GUSTAFSSON, J.-A. (eds)(1988). The Steroid/ Thyroid Hormone Receptor Family andGene Regulation. Basel: Birkhauser.

CHANG, T. C. AND SHAPIRO, D. J. (1990). An NFl-relatedvitellogenin activator element mediates transcription from theestrogen-regulated Xenopus laevis vitellogenin promoter. J. biol.Chem. 265, 8176-8182.

CLIFFORD, J. L., PETKOVTTCH, M., CHAMBON, P. AND LOTAN, R.(1990). Modulation by retinoids of mRNA levels for nuclearretinoic acid receptors in murine melanoma cells. Molec.Endocrinol. 4, 1546-1555.

COOK, C. B. AND KOENIG, R. J. (1990). Expression of erbA a-and/3 mRNAs in regions of adult rat brain. Molec. cell. Endocnnol.70, 13-20.

DELONG, G. R., ROBBINS, J. AND CONDUFFE, P. G. (eds) (1989).Iodine and the Brain. New York: Plenum Press.

DERBY, A. AND ETKIN, W. (1968). Thyroxine induced tailresorption in vitro as affected by anterior pituitary hormones. /.exp. Zool. 169, 1-8.

DETHE, H., VrvANCo-Ruiz, M., TIOLLAIS, P., STUNNENBERG, H.AND DEJEAN, A. (1990). Identification of a retinoic acidresponsive element in the retinoic acid receptor beta gene.Nature 343, 177-180.

DEUCHAR, E. (1975). Xenopus. The South African Clawed Frog.London: John Wiley.

DURSTON, A . J . , TlMMERMANS, J . P . , H A G E , W. J . , HENDRICKS, H .F . , DE VR1ES, N . J . , HEIDEVELD, M. AND NlEUWKOOP, P . D .(1989). Retinoic acid causes an anteroposterior transformation inthe developing central nervous system. Nature 340, 140-144.

EVANS, R. M. (1988). The steroid and thyroid hormone receptorsuperfamily. Science 240, 889-895.

FORREST, D., SJOBERG, M. AND VENNSTROM, B. (1990). Contrastingdevelopmental and tissue specific expression of a and fl thyroidhormone receptor genes. EMBO J. 9, 1519-1528.

Fox, H. (1981). Cytological and morphological changes duringamphibian metamorphosis. In Metamorphosis. A Problem inDevelopmental Biology (ed. L. Gilbert and E. Frieden) pp.327-362. New York: Plenum Press.

GALTON, V. A. AND ST GERMAIN, D. L. (1985). Putative nucleartriiodothyronine receptors in tadpole liver during metamorphicclimax. Endocrinology 117, 912-916.

GILBERT, L. I. AND FRIEDEN, E. (eds.) (1981). Metamorphosis. AProblem in Developmental Biology. New York: Plenum Press.

GLASS, C. K., LIPHN, S. M., DEVARY, O. V. AND ROSENFELD, M.G. (1989). Positive and negative regulation of gene transcriptionby a retinoic acid-thyroid hormone receptor heterodimer. Cell59, 697-708.

GRAUPNER, G., WILLS, K. N., TZUKERMAN, M., ZHANG, X.-K.AND PFAHL, M. (1989). Dual regulatory role for thyroid-hormone receptors allows control of retinoic-acid receptoractivity. Nature 340, 653-656.

GREEN, S. D. AND CHAMBON, P. (1988). Nuclear receptors enhance

our understanding of transcription regulation. Trends Genet. 4,309-314.

HAUSER, K. F. AND GONA, A. G. (1984). Purkinje cell maturationin the frog cerebellum during thyroxine-induced metamorphosis.Neuroscience 11, 139-155.

HODIN, R. A., LAZAR, M. A., WINTMAN, B. I., DARLING, D. S.,KOENIG, R. J., LARSEN, P. R., MOORE, D. D. AND CHTN, W. W.(1989). Identification of a thyroid hormone receptor that ispituitary-specific. Science 244, 76-79.

HOLLENBERG, S. M. , THOMPSON, C. C. AND EVANS, R. M. (1989).Hormone-dependent transcriptional activation by thyroidhormone receptors: Functional homology with steroid hormonereceptors. In Molecular Mechanisms of Hormone Action (ed. U.Gehring, E. Helmreich and G. Schutz), pp. 35-43. Springer:Berlin.

KAWAHARA, A., KOHARA, S. AND AMANO, M. (1989). Thyroidhormone directly induces hepatocyte competence for estrogen-dependent vitellogenin synthesis during the metamorphosis ofXenopus laevis. Devi Biol. 132, 73-80.

KIKUYAMA, S., NIKI, K., MAYUMI, M., SHIBAYAMA, R., NISHIKAWA,M. AND SHINTAKE, N. (1983). Studies on corticoid action on thetoad tadpole tail in vitro. Gen. comp. Endocrinol. 52, 395-399.

KOENIG, R. J., WARNE, R. L., BRENT, G. A., HARNEY, J. W.,LARSEN, P. R. AND MOORE, D. D. (1989). Isolation of a cDNAclone encoding a biologically active thyroid hormone receptor.Proc. natn. Acad. Sci U.S.A. 85, 5031-5035.

KOLLROS, J. J. (1981). Transitions in the nervous system duringamphibian metamorphosis. In Metamorphosis. A Problem inDevelopmental Biology (ed. L. Gilbert and E. Frieden) pp.445-459. New York: Plenum Press.

LELOUP, J. AND BUSCAGLIA, M. (1977). La triiodothyronine,hormone de la metamorphose des amphibiens. C.r. hebd. StoneAcad. Sci. Paris 284, 2261-2263.

LEONE, F., LAMBERT-GARDINI, S., SARTORI, C. AND SCAPIN, S.(1976). Ultrastructural analysis of some functional aspects ofXenopus laevis pancreas during development andmetamorphosis. /. Embryol. exp. Morph. 36, 711-724.

LEROY, P., KRUST, A., ZELENT, A., MENDELSOHN, C , GARNIER,J.-M., KASTNER, P., DIERICH, A. AND CHAMBON, P. (1991).Multiple isoforms of the mouse retinoic acid receptor a aregenerated by alternative splicing and differential induction byretinoic acid. EMBO J. 10, 59-69.

MANIATIS, T., FRTTSCH, E. F. AND SAMBROOK, J. (1982). MolecularCloning. A Laboratory Manual. Cold Spring HarborLaboratory. Cold Spring Harbor, New York.

MATHISEN, P. M. AND MILLER, L. (1989). Thyroid hormoneinduces constitutive keratin gene expression during Xenopuslaevis development. Molec. cell. Biol. 9, 1823-1831.

MELTON, D. A., KRIEG, P. A., REBAGUATI, M. R., MANIATIS, T.,ZINN, K. AND GREEN, M. R. (1984). Efficient in vitro synthesisof biologically active RNA and RNA hybridization probes fromplasmids containing a bacteriophage SP6 promoter. Nucl. AcidsRes. 12, 7035-7056.

MOHUN, T., GARRETT, N., STUTZ, F. AND SPOHR, G. (1988). Athird striated muscle actin gene is expressed during earlydevelopment in the amphibian Xenopus laevis. J. molec. Biol.202, 67-76.

MOHUN, T. J. AND GARRETT, N. (1987). An amphibiancytoskeletal-type actin gene is expressed exclusively in muscletissue. Development 101, 393-402.

MORTYA, T., THOMAS, C. R. AND FRIEDEN, E. (1984). Increase in3,5,3'-triiodothyronine (T3)-binding sites in tadpole erythrocytenuclei during spontaneous and T3-induced metamorphosis.Endocrinology 114, 170-175.

MORREALE DE ESCOBAR, G . , R U I Z - M A R C O S , A . AND ESCOBAR DELREY, F. (1983). Thyroid hormones and the developing brain. InCongenital Hypothyroidism (ed. J. H. Dussault and P. Walker)pp. 85-126. New York: Marcel Dekker Inc.

MosKAms, J. E., SARGENT, T. D., SMITH, L. H., JR, PASTORI, R.L. AND SCHOENBERG, D. R. (1989). Xenopus laevis serumalbumin: sequence of the complementary deoxyribonucleic acidsencoding the 68- and 74-kilodalton peptides and the regulationof albumin gene expression by thyroid hormone duringdevelopment. Molec. Endocrinol. 3, 464-473.

Developmental localization o/Xenopus TRa and fi mRNA 943

NIEUWKOOP, P. O. AND FABER, J (1%7) Normal Table ofXenopus laevis (Daudin), 2nd Edition, Amsterdam: North-Holland.

PERLMAN, A. J., WOLFFE, A. P., CHAMPION, J. AND TATA, J. R.

(1984). Regulation by estrogen receptor of vitellogenin genetranscription in Xenopus hepatocyte cultures. Molec. cell.Endocrinol. 38, 151-161.

PETKOVICH, M., BRAND, N. J., KRUST, A. AND CHAMBON, P.

(1987). A human retinoic acid receptor which belongs to thefamily of nuclear receptors. Nature 330, 444-450.

PIOTROWSKI, D. C. AND KALTENBACH, J. C. (1985).

Immunofluorescent detection and localization of thyroxine inblood of Rana catesbeiana from early larval throughmetamorphic stages. Gen. comp. Endocrinol. 59, 82-90.

RAGSDALE, C. W., JR, PETKOVTTCH, M., GATES, P. B., CHAMBON,

P. AND BROCKES, J. P. (1989). Identification of a novel retinoicacid receptor in regenerative tissues of the newt. Nature 341,654-657.

SAP, J., MUNOZ, A., DAMM, K., GOLDBERG, Y., GHYSDAEL, J.,

LEUTZ, A., BEUG, H. AND VENNSTROM, B. (1986). The c-erb-A

protein is a high-affinity receptor for thyroid hormone. Nature(London) 324, 635-640.

SHAPIRO, D. J., BARTON, M. C , MCKEARIN, D. M., CHANG, T. -C,

LEW, D., BLUME, J., NIELSEN, D. A. AND GOULD, L. (1989).

Oestrogen regulation of gene transcription and mRNA stability.Recent Progr. Hormone Res. 45, 29-58.

SHERIDAN, P. J., BLUM, K. AND TRACHTENBERG, M. C. (eds.)

(1988). Steroid Receptors and Disease. New York: MarcelDekker.

SHUPNIK, M. A., GORDON, M. S. AND CHIN, W. W. (1989). Tissue-

specific regulation of rat estrogen receptor mRNAs. MolecEndocrinol 3, 660-665.

TATA, J. R. (1966). Requirement for RNA and protein synthesisfor induced regression of the tadpole tail in organ culture. DeviBiol. 13, 77-94.

TATA, J. R. (1968). Early metamorphic competence of Xenopuslarvae. Devi Biol. 18, 415-440.

TATA, J. R. (1984). The action of growth and developmentalhormones: evolutionary aspects. In Biological Regulation and

Development, Vol. 3B, (eds. R. F. Goldberger and K. R.Yamamoto) pp. 1-58. New York: Plenum Press.

VARRIALE, B. AND TATA, J. R. (1990). Autoinduction of estrogenreceptor is associated with FOSP-1 mRNA induction byestrogen in primary cultures of Xenopus oviduct cells. Molec.cell. Endocrinol. 71, R25-R31.

WEBER, R. (1967). Biochemistry of amphibian metamorphosis. InThe Biochemistry of Animal Development (ed. R. Weber) pp.227-301. New York: Academic Press.

WEILER, I. J., LEW, D. AND SHAPIRO, D. J. (1987). The Xenopuslaevis estrogen receptor: Sequence homology with human andavian receptors and identification of multiple estrogen receptormessenger ribonucleic acids. Molec. Endocrinol. 1, 355-362.

WEINBERGER, C , THOMPSON, C. C , ONG, E. S , LEBO, R., GRUOL,

D. J. AND EVANS, R. M. (1986) The c-erb-A gene encodes athyroid hormone receptor. Nature 324, 641-646.

WHITE, B A. AND NICOLL, C. S. (1981). Hormonal Control ofAmphibian Metamorphosis. In Metamorphosis. A Problem inDevelopmental Biology (eds. L. I. Gilbert and E. Frieden) pp.363-3%. New York: Plenum Press.

WILKINSON, D. AND GREEN, J. (1990). In situ hybridization andthree-dimensional reconstruction of serial sections. In ThePractical Approach Series: Postimplantation Mouse Embryos: APractical Approach (ed. D. Rickwood and D. L. Cockcroft), pp.155-171. Oxford: IRL Press.

WITSCHI, E. (1956). Development of Vertebrates. Philadelphia:Saunders.

YAOITA, Y. AND BROWN, D. D. (1990). A correlation of thyroidhormone receptor gene expression with amphibianmetamorphosis. Genes Dev. 4, 1917-1924.

YAOITA, Y., SHI, Y-B. AND BROWN, D. D. (1990). Xenopus laevisa and R thyroid hormone receptors. Proc. natn. Acad. Sa.U.S.A. 87, 7090-7094.

ZINN, K., Di MAIO, D AND MANIATIS, T (1985). Identification of

two distinct regulatory regions adjacent to the human /3-mterferon gene. Cell 34, 865-879.

(Accepted 23 April 1991)