DNA AND CELL BIOLOGY Volume 10, Number 3, 1991 Mary Ann Liebert, Inc., Publishers Pp. 211-221

Isolation and Characterization of a cDNA Encoding a Chicken ß Thyroid Hormone Receptor MARK O.

SHOWERS, DOUGLAS S. DARLING, GORDON D. KIEFFER, and WILLIAM W. CHIN

ABSTRACT We have isolated and characterized a cDNA encoding a chicken ß homolog of c-erbA, or thyroid hormone receptor (TR). Chicken liver cDNA libraries were screened with a rat TR/3-1 cDNA probe, and several cDNA inserts were isolated and characterized. The sequence of one cDNA predicts a 369-amino-acid open reading frame (ORF), with a protein sequence that possesses 96% identity with that of rat TR/3-1, but only 88% identity with chicken TRa. These data indicate that the cDNA likely encodes a ß form of TR that has the expected putative DNA and T., binding domains. The chicken TR/3 (chTR/3) in vitro translated protein binds T3 with high affinity, and binds both the thyroid hormone response element (TRE) from the rat growth hormone gene and the Xenopus vitellogenin A2 gene estrogen response element (ERE), similarly to that of the rat TR/3-1. Northern blot analysis revealed the expression of a 7.0-kb RNA in several tissues including cerebellum, pituitary, kidney, and liver. This chicken liver TR/3 cDNA sequence varies in both the 5' and 3' untranslated regions from the chicken kidney TR/3 cDNA sequence recently reported (Forrest et al., 1990). The 5' untranslated cDNA sequence divergence occurs near a potential splice site junction of the human TR/3 gene, suggesting that this chicken liver cDNA may represent an alternatively spliced RNA product of the chicken TR/3 gene.

bryogenesis (Dasmahapatra et ai, 1987; Hentzen et ai, 1987). Conceivably, multiple TRs or modified TRs mediate these T3 effects during chick ontogenesis and throughout

INTRODUCTION and

T3 (3,5,3'-triiodothyronine) TheT4 (3,5,3',5'-tetraiodothyronine) profound effects metabolism, thyroid hormones

via thyroid hormone receptors (TRs) velopment, and differentiation in all vertebrates (Gorbman et ai, 1983; Ichikawa er a/., 1989). These effects have been well characterized in chicken, in which thyroid hormones regulate chick skin development, and the differentiation, pigmentation, and molting of feathers (Gorbman er ai, 1983). T3 increases malic enzyme mRNA 100-fold and malic enzyme synthesis in chick embryo hepatocytes in culture, suggesting that T3 regulation of the enzyme is largely pretranslational (Winberry et ai, 1983). TRs undergo changes in binding and number during ontogenesis in the liver, lung, brain, and pectoral muscle of chick embryos (Goldberg et ai, 1989), as well as in chick skeletal muscle (Dainat et ai, 1984). A decrease in nuclear T3 binding sites in chick embryo erythrocytes during development has been reported, and the plasma level of T3 increases during emon

Division of Genetics, Department of Medicine, Medical School, Boston, MA 02115.

de-

Brigham and

life, as do their mammalian homologs (Bellabara et ai, 1983; Dainat et ai, 1984; Bigler and Eisenman, 1988; Freake et ai, 1988). adult

exert

Considerable evidence suggests that these physiologic TRs are encoded by c-erbA, the cellular homolog of the viral erbA oncogene, \-erbA (Sap et ai, 1986; Weinberger et ai, 1986). In turn, c-erbAs comprise a small subfamily of the thyroid steroid receptor superfamily (Evans, 1988). Multiple TR cDNAs have been identified in mammals and have been classified into a and ß homologs based on their sequence similarity, and by mapping to human chromosomes 17 and 3, respectively (Spurr et ai, 1984; Wein-

berger et ai, 1986; Thompson et ai, 1987; Evans, 1988; Murray et ai, 1988; Nakai et ai, 1988; Hodin et ai, 1989; Lazar et ai, 1989; Mitsuhashi et ai, 1988; Sakurai et ai, 1989). Two forms of TRj3 have been identified in rat, rTR0-l and rTR/3-2 (Koenig er ai, 1988; Hodin et ai,

Women's

211

Hospital, Howard Hughes Medical Institute, and Harvard

SHOWERS ET AL.

212

1989). Rat TR/3-2 is expressed only in pituitary, while rTR/3-1 is expressed in most tissues. The two rat TR/3s share identical amino acid sequences at the central and

boxy-terminal portions of the protein, but differ

car-

com-

pletely at the amino terminus. This divergence represents tissue-specific splicing of a single rTR/3 gene transcript (Hodin et ai, 1989). Alternative splicing of the rat TRa gene also generates multiple forms of rTRa mRNAs (Izumo and Mahdavi, 1988; Mitsushashi and Nikodem, 1988; Lazar et ai, 1989). The functional significance of this rTR diversity has not yet been established but may involve tissue-specific and/or gene-specific regulation by T3. A single avian TRa cDNA has been isolated and characterized (Sap et ai, 1986; Forrest et ai, 1990). The chTRa cDNA encodes a 47-kD protein that binds T3 with high affinity (Kd 2 x 10-'° M) and specificity (Sap et ai, 1986; Goldberg et ai, 1989). Chicken TRa confers T3 inducibility to a reporter construct containing the 5' end of the rat prolactin gene, when transfected into TR-deficient cells (235-1) (Horowitz et ai, 1989). To study the structural and functional preservation of this TR diversity in a different class of vertebrate, a chicken liver TR/3 cDNA was isolated and characterized. The =

encoded amino acid sequence is 90% identical to the mammalian TR/3 homologs, containing nearly identical DNA and T3 binding domains. The chTR/3 binds T3 with high affinity, as well as a thyroid hormone response element (TRE) oligomer, suggesting possible functional preservation of TR/3 in aves. While this work was in progress, a chicken kidney TR/3 was reported (Forrest et ai, 1990), which differs in both the 5' and 3' untranslated regions from the chTR/3 described herein, suggesting possible alternative splicing of chTR/3 RNAs. The presence of both the a and ß TRs in chickens, and of the alternatively spliced TR/3 mRNAs, suggests, by analogy with mammalian TRs, that TR diversity is an important aspect of T3 responses in vertebrates.

MATERIALS AND METHODS

Genomic Southern

analysis

Chicken and rat genomic DNAs (Clontech Labs, Inc., Palo Alto, CA) were analyzed by Southern blot analysis (Ausubel et ai, 1989). The blots were hybridized with either the chicken TRa cDNA (Hind III fragment) (Vennstrom et ai, 1982), or the rTR/3 "common" cDNA fragment (the 0.55-kb Xba l-Pst I portion common to both the rTR/3-1 and rTR/3-2 cDNAs) (Hodin et ai, 1989), for at least 24 hr at 55°C in 5 x SSC. Cloned cDNA fragments were radiolabeled by the random-primer method, typically to a specific activity of at least 9 x 10" cpm/fig DNA (Feinberg and Vogelstein, 1983). RNA probes (riboprobes) were synthesized with T7 or T3 polymerase from linearized Bluescript SK II +/— plasmids containing the cloned TR cDNA inserts, as previ-

ously described (Lazar et ai, 1988). Oligonucleotide primers were phosphorylated with T4 DNA kinase and [7-32P]ATP, and plasmids were prepared by standard techniques (Ausubel et ai, 1989). DNA sequencing was per-

formed using the dideoxy method (Sanger et ai, 1977) with T4 DNA polymerase (Sequenase, USB Biochemicals). Nucleotide sequence analyses were performed using the Microgenie (Beckman Instruments, Palo Alto, CA) and the Wisconsin Genetics Computer Group sequence analysis programs.

cDNA

cloning

Two independent chicken liver cDNA libraries were screened for the presence of chicken TR/3 cDNAs. The first, a XZAP I library (Stratagene), was prepared from total RNA and random primers. The second, a Xgtll library, was prepared from poly(A)*RNA and oligo-dT primers (Sambrook et ai, 1989). The XZAP or Xgtll cDNA libraries were screened by standard methods (Ausubel et ai, 1989; Sambrook et ai, 1989), using the radiolabeled rTR/3 "common" cDNA or the chTRa cDNA as hybridization probes. Approximately 1 x 106 recombinant phage were screened for the presence of chTR|3 cDNAs, as previously described (Sambrook et ai, 1989). Positive plaques yielding coincident hybridization signals for the TRa and TR/3 cDNA probes were rescreened, and the phage were plaque-purified. The isolated cDNAs from the purified XZAP phage were subcloned into Bluescript plasmids by the excision process (Stratagene, La Jolla, CA), or by standard cloning techniques for the Xgtll clones, and the nucleotide sequence of each cDNA insert was determined

(Fig. 1). Northern blot

analysis

RNA was prepared by the guanidinium isothiocyanate method (Chirgwin et ai, 1979) from quick-frozen chicken tissues. Ten to twenty micrograms of total RNA were subjected to Northern blot analysis (Sambrook et ai, 1989). The blots were hybridized with either the chTR/3 (1 1.8-kb Eco RI fragment) cDNA probe or the chTRa (1.6-kb Hind III fragment) cDNA probe in 5x SSC (with no formamide) at 42° for 18 hours. Sequence-specific riboprobes were made by subcloning the probe sequence downstream from the T7 or T3 promoter (Stratagene). The relative amount of RNA on Northern blot autoradiographs was quantified by laser scanning densitometry (Molecular Dynamics, Sunnyvale, CA). The data were normalized to 28S and 18S RNA hybridization signals to correct for unequal gel loading of RNA.

In vitro translation In vitro translation of the chTR/3 was performed by subcloning chTR/3 cDNAs into Bluescript plasmids, and linearizing them with Bam HI such that the sense RNA strand was synthesized from the T7 promoter, as previously described (Burnside et ai, 1989). Protein was synthesized by translation of the chTR/3 RNA in rabbit reticulocyte (Bethesda Research Labs, BRL) lysates in the presence of [35S]methionine, (800 Ci/mmole, New England Nuclear, NEN) according to the manufacturer's instructions. 35SLabeled receptor was analyzed by NaDodS04/polyacryl-

213

CHICKEN ß THYROID HORMONE RECEPTOR cDNA

Eco RI

0.8 kb-

Eco RI 2130

1277

INSERT 1 XZAP I

Eco RI

-1.8 kb-

1

Eco RI 1797

INSERT 2 \gt11

Eco RI 1

Eco RI

Eco RI

1797

2130

COMPOSITE cDNA

COMPOSITE cDNA

FIG. 1. Sequence strategy and overlap of phage cDNA inserts. The overlapping nucleotide sequences of the two cDNA inserts are shown. Insert 1 is a 0.8-kb isolate from a chicken liver XZAP cDNA library. Insert 2 is a 1.8-kb isolate from a different chicken liver Xgtll cDNA library. The insert 1 cDNA (nucleotides 1,277-2,130) contains 60 bp of protein coding sequence and 793 bp of 3' untranslated region. The insert 2 cDNA (nucleotides 1-1,797) contains the entire protein coding sequence of the chicken TR/3 plus 227 bp of 5' untranslated region and 460 bp of 3' untranslated region. The two cDNA inserts are then joined at their overlapping sites to generate a composite cDNA sequence, as the cDNA inserts are identical over their 520 overlapping nucleotides at the 3' portion of the composite chicken TR/3 cDNA. The 2,130-bp composite cDNA sequence encodes a chicken TR/3 protein. A polyadenylation signal is not apparent in the 3' untranslated region. The sequence strategy used to determine the nucleotide sequence of the inserts is indicated by arrows above and below the cDNA insert. These arrows indicate the strand, direction, and extent of each of the sequences determined. The numbers above the bars refer to the nucleotide sequence. The larger shaded bar depicts the region of the cDNA encoding the chicken TR/3.

amide gel electrophoresis followed by fluorography. Incorporation of [35S]methionine into protein was determined by TCA precipitation, as previously described (Burnside et ai, 1989).

T3 binding The translated proteins were analyzed for T3 binding by Scatchard analyses. Binding of 0.1 nM 125I-T3 to in vitrotranslated protein products was analyzed with 10 pi of programmed rabbit reticulocyte lysate, as previously described (Burnside et ai, 1989). Nonspecific binding was measured in the presence of 400-fold excess of unlabeled ligand. Rat /3-actin cDNA was used as a negative control, and rTR/3-1 cDNA was used as a positive control for these studies

(Darling DNA

et

ai, 1989).

binding

Binding of the products to several response elements was assessed by the ABCD binding assay of Glass et ai (1987). Complementary oligonucleotides were designed to have 15-20 bases of single stranded 5' overhanging regions (tails) when annealed (Burnside et ai, 1989). Doublestranded oligonucleotides encompassing the TRE (bases -209 to -146) of the rat GH gene or the Xenopus vitellogenin A2 gene estrogen response element (ERE) were constructed as described (Glass et ai, 1988). An oligonucleo-

tide corresponding to the third exon of the chicken actin gene (+2,077 to +2,146) was included as a negative control (Darling et ai, 1989). Binding of the c-erbA protein to biotinylated oligomers was measured by incubation of 3-5 ¡il of 35S-labeled protein with biotinylated DNA in 40 ¡il of binding buffer containing 250 fig/ml poly(dl-dC), 0.1% (vol/vol) Nonidet P-40, and nonradioactive T3 (10 nM). After incubation at room temperature for 40 min, protein-DNA complexes were precipitated by the addition of streptavidin-agarose as described (Glass et ai, 1987) and

quantitated by scintillation counting. RESULTS

Isolation and characterization of chTRß The existence of a chTR/3 gene was suggested by genomic Southern blot analysis using a rTR/3 cDNA probe containing sequences common to both the rTR/3-1 and the rTR/3-2 cDNAs. The hybridization pattern generated with the rat TR/3 "common" cDNA probe (using less stringent conditions) was distinct from that obtained using the chicken TRa cDNA probe for each of the restriction enzyme digests of chicken genomic DNA (data not shown). These differences in hybridization pattern suggested the existence of a TR/3 gene in the chicken.

SHOWERS ET AL.

214

99

gaattccgccgggccgcgcttccccccgccccgcgccccctcccggcgaccctctccgcggcggaggggccgttgcgtaacgcgtccgcacggcgccgg ggccggcagcgggcgaggcgcgcggtggccgcggccggagtgagagaacagaaaatgaagctctagtaaggtggcagtggtgacagaagaagaacaaga atgattacgtatctgtaactctcagcagt

227

ATG TCA GGG TAT ATA CCC AGT TAC TTA GAC AAG GAT GAG CTA TGT GTA GTA TGT GGG GAC AAA

290

M

S

G

Y

I

P

S

Y

L

D

K

D

E

L

C

V

V

C

6

D

198

K

+1

GCC ACC GGA TAT CAT TAT CGC TGC ATC ACT TGT GAA GGT TGC AAG GGA TTT TTC AGA AGA ACC ATT CÄG AAA AAC

365

ATGYHYRCITCËGCKGFfRRTÏQRN +26

CTC CAT CCA ACC TAT TCC TGT AAA TAT GAA GGA AAA TGT GTG ATA GAC AAA GTA ACA AGA AAT CAG TGC CAG GAA

440

LHPTYSCKYEGKCVIDKVTKNQCQE +51 TGT CGC TTC AAA AAA TGT ATC TTT GTT GGC ATG GCA ACA GAT TTG GTG TTG GAT GAC AGC AAG AGG CTG GCA AAG F K C î F V 6 H A T D R K LVLD D SKRLAK C +86

515

AGG AAG CTG ATA GAA GAA AAT CGA GAG AAG AGG CGT CGG GAA GAG CTG CAG AAA ACG ATG GTT CAC AAA CCA GAA

RKLIEENREKRRREELQKTMGHKPE +101 CCA ACA GAT GAG GAA TGG GAG CTG ATC AAA ATT GTC ACT GAA GCA CAT GTG GCC ACC AAT GCA CAA GTA AGC CAC

PTDEEWELIKIVTEAHVATNAQGSH + 126

TGG AAG CAG AAA AGG AAA TTT CTG CCA GAA GAC ATT GGG CAA GCA CCA ATA GTT AAT GCC CCA GAA GGG GGG AAA

WKQKRKFLPEDIGCjAPIVNAPEGGK + 151

GTG GAT TTA GAA GCC TTC AGC CAG TTT ACA AAA ATT ATC ACA CCA GCG ATT ACA AGA GTG GTG GAT TTT GCC AAA

VDLEAFSQFTKI

815

ITPAITRVVDFAK

+ 176

AAG TTA CCI ATG TTT TTG GAG CTG CCA TGT GAA GAC CAG ATC ATC CTT CTG AAA GTC TGC TGT ATG GAG ATA ATG

890

KLPMFCELPCEDQIILLKGCCMEIM +201 TCC CTC CGA GCA GCA GTT CGC TAT GAC CCC GAG AGT GAG ACT TTA ACG CTA AAT GGG GAG ATG GCG GTG ACA AGG

965

SLRAAVRYDPESETLTLNGEMAVTR +226

GGC CAG CTG AAA AAT GGG GGT CTT GGC GTA GTT TCT GAT GCC ATT TTT GAC CTG GGC ATG TCT CTT TCT TCA TTT 1040

GQLKNGGLGVVSDAIFDLGMSLSSF +251 AAC CTG GAT GAC ACC GAG GTT GCC CTT CTC CAG GCT GTC CTG CTC ATG TCA TCA GAT CGC CCA GGC CTT GTT TGC 1115

NLDDTEVALLQAVLLMSSDRPGLVC +276 GTC GAG AGA ATA GAA AAG TGT CAA GAG GGT TTC CTC CTG GCA TTT GAA CAC TAC ATT AAT TAC AGA AAA CAC CAT 1190

VERIEKCQEGFLLAFEHYINYRKHH 1-301 GTT V

GC^ A

CAT TTT TGG CCA AAA CTG CTG ATG AAA GTG ACA GAT CTG CSA ATS ATT GGA GCC TGC CAT GCC AGC CGC 1265 S 8 c'a A t» M I G A fî F W. P K L L M K V T D

+326 '

TTC CTG CAC ATG AAG GTG GAG TGC CCC ACA