-
Vol. 4, 90 1-909, November 1993 Cell Growth &
Differentiation 901
Identification of a Thyroid Hormone Response Elementin the Mouse
Myogenin Gene: Characterization ofthe Thyroid Hormone and Retinoid
X Receptor
. 8 #{149} #{149} #{149} 1Heterodimeric Binding Site
Michael Downes, Russell Griggs, Annette Atkins,Eric N. Olson,
and George E. 0. Muscat2
Centre for Molecular Biology and Biotechnology, Ritchie
ResearchLaboratories, University of Queensland, St. Lucia, 4072
Queensland,Australia FM. D., R. G., A. A., G. E. 0. Ml, and
Department of
Biochemistry and Molecular Biology, M. D. Anderson Cancer
Center,
University of Texas, Houston, Texas 77030 IE. N. 0.1
Abstract
Thyroid hormones are positive regulators of muscledevelopment in
vivo. Triiodo-t-thyronine (T3) treatmentof myogenic cell lines
results in the precociousexpression of myogenin, a muscle
specific,helix-loop-helix fador that can trans-adivate
musclespecific gene expression (G. Carnac et a!., Mol.Endocrinol.,
6: 1 1 85-i 1 94, 1 992). We have identified aT3 response element
(TRE) in the mouse myogenin (MM)promoter between nucleotide
positions -526 and -494(5’ GTGGTAGGTC1TTAGGGGTCTCATGGGACTGACA3’).
This sequence conferred appropriate hormonalregulation to an
enhancerless SV4O promoter.Eledrophoretic mobility shift analysis
experimentsshowed that thyroid hormone receptor a (TRa) andretinoid
X receptor a (RXRa) formed a heterodimericcomplex on the MM TRE
that was specifically competedby classical TREs and not by other
response elements.Analyses of this heterodimer with a battery of
steroidhormone response elements indicated that the complexwas
efficiently competed by a dired repeat of theAGGTCA motif separated
by 4 nucleotides, as predictedby the 3-4-5 rule. Eledrophoretic
mobility shift analysisexperiments showed that the myogenin,
growthhormone, and myosin heavy chain TREs interaded withan
identical nuclear fador(s) in muscle cells that wasconstitutively
expressed during myogenesis. Mutagenesisof the MM TRE indicated
that the sequence of thedired repeats (AGGTCA) and the 4-nucleotide
gap werenecessary for efficient binding to the
TRa/RXRaheterodimeric complex. In conclusion, our data suggestthat
the MM TRE is a target for dired cross-talk betweentwo different
hormonal signals (T3 and 9-cis-retinoicacid) at the receptor level.
To our knowledge, this is thefirst identification and
charaderization of a hormonalresponse element in the muscle
specific helix-loop-helixfamily that directly interads with TRa and
RXRa.These results indicate that the thyroid hormones, which
Received 5/19/93; revised 8/i6/93; accepted 9/2/93.
1 This work was supported by the University of Queensland Mayne
Bequest
Fund and the National Health and Medical Research Council
(NHMRC).G. E. 0. M. is an R. D. Wright Fellow of the NHMRC.2 To
whom requests for reprints should be addressed.
are positive regulators of myogenic terminaldifferentiation,
diredly target the hierarchical regulatorsof myogenesis.
Introdudion
During embryogenesis, myogenic progenitor cells arisewithin the
somites. These determined cells are committed toa myogenic fate;
however, they do not express a contractilephenotype until they
receive the appropriate environmentalsignals. In tissue culture,
proliferating mononucleated myo-blasts exit the cell cycle and
differentiate into postmitoticmultinucleated myofibers that express
muscle specific phe-notypic markers. Proliferation and
differentiation are mutu-ally exclusive events regulated by a
balance of antagonizing
cellular signals. The recent characterization of the MyoDgene
family (MyoD, myogenin, Myf-5, MRF-4!myf-6!herculin) of HLH3
proteins, which inhibit cell proliferationand trans-activate
directly/indirectly a multitude of contrac-tile genes, has provided
insight into this system. This genefamily functions at the nexus of
command circuits that con-trol the mutually exclusive events of
division and differen-tiation. The expression/function of the HLH
proteins andsubsequent contractile protein transcription are
suppressedby growth factors and oncogenic signals that promote
celldivision (1-3).
TheMyoDfamilyofhierarchical regulatory proteins direct
the fate of pluripotential mesodermal embryonic cells
andindirectly/directly activate muscle specific genes involved
in
terminal differentiation and contraction. The members of theMyoD
family have the ability to autoregulate their own ex-pression and
cross-activate one another’s expression in cul-tured cells. This
mechanism amplifies the expression of thesefactors above an
activation threshold and provides stabilityto the myogenic program
(i-3). These proteins contain a68-amino acid conserved
basic!myc-like region that is nec-essary and sufficient for
myogenic conversion. The basicregion and adjacent HLH motif mediate
DNA binding anddimerization. The MyoD family forms heterodimers
with
3 The abbreviations used are: HLH, helix-loop-helix; T�,
3,5,3’-triiodo-�-thy-
ronine; TR, thyroid hormone receptor; DR, direct repeat; PAL,
palindromicrepeat; RXR, retinoid x receptor; RA, retinoic acid;
COUP-IF, chicken ov-albumin upstream promoter transcription factor;
MM, mouse myogenin; IRE,thyroid hormone response element; a-MHC,
a-myosin heavy chain; EMSA,
electrophoretic mobility shift analysis; GH, growth hormone; r-,
rodent; h-,human; c-, chicken; CRBP I, cellular retinol binding
protein I; RARE, RAresponse element; MEF, myocyte enhancer factor;
SRE, serum response el-
ement; PPAR, peroxisome proliferator activated receptor; VDR,
vitamin Dreceptor; RAR, RA receptor; MI, myotube; CAT,
chloramphenicol acetyltrans-
ferase; CMV, cytomegalovirus; DMEM, Dulbecco’s modified Eagle’s
medium;FCS, fetal calf serum; DOTAP, N-Il
-(2,3-dioleoyloxy)propyll-N,N,N-
trimethyl ammonium-methylsulfate; HEPES, 4-(2-hydroxyethyl)-i
-pipera-zineethanesulfonic acid; PMSF, phenylmethylsulfonyl
fluoride; EGTA, ethyl-
eneglycol bis(f3-aminoethyl ether)-N,N,N’,N’-tetraacetic
acid.
-
902 Identification of IRE in Mouse Myogenin Gene
ubiquitously expressed members ofthe HLH protein family,
such as El 2 and E47 (the alternatively spliced products of
theE2A gene). The MyoD-E2A heterodimers bind to a consen-sus DNA
binding sequence, the E-box motif which includesa CANNTG motif,
present in most muscle specific enhanc-ers. The individual muscle
specific HLH proteins show dis-tinct half-site preferences for
binding that depend on thevariable nucleotides within and
surrounding the invariantdyad symmetry of the E-box motif
(1-3).
The expression and function of these hierarchical regu-lators is
repressed by growth factors and oncogene products[fibroblast growth
factor, transforming growth factor j3, pro-liferin, Ela, Src, Fps,
Ras, Myc, Fos, and Jun (summarizedand discussed in Ref. 2)]. These
agents act by (a) repressionof MyoD mRNA accumulation and protein
synthesis, (b)regulation of MyoD activity/function, or (c)
suppression ofmyogenic specific trans-acting factors that
positively regu-late the expression of myogenin and MRF-4, which
are ex-pressed only after terminal differentiation (2).
Very little is known about the growth factors and hor-mones that
act as positive regulators of the MyoD family(with respect to their
transcription and function) and asstimulators of differentiation.
Thyroid hormones are amongthe known activators of in vivo muscle
development. Thy-roid hormones exert marked effects on cardiac and
skeletalmuscle. T3 increases myocyte size, total RNA, and
protein
synthesis. This results in a substantial reorganization of
themyocyte, which, in turn, alters contractile protein perfor-
mance and ion transport. Hyper- and hypothyroidism resultin
increases and decreases, respectively, in heart rates, car-diac
output, and the velocity of shortening of unloadedmuscle. In
skeletal muscle, hyper- and hypothyroidism, re-spectively, result
in a precocious and retarded developmentof fast contractile protein
gene expression. Hypothyroidismresults in a decrease in muscle mass
and a reduced numberof myofibers (2, 4-1 1 ). Thyroid hormone
regulates the tran-scription of a number of contractile protein
genes [skeletaland cardiac ct-actin, cardiac and skeletal myosin
heavychain genes, atrial natriuretic factor, sarcolemmal
calciumATPase, fast and slow sarcoplasmic calcium ATPases,
car-bonic anhydrase Ill, the neural cell adhesion molecule, andthe
adrenergic receptors in ventricular cardiocytes (1 0-i
7)].Furthermore, T3 promotes terminal muscle
differentiation,positively regulates MyoD, and induces the
premature ex-pression of myogenin (6).
The effects of thyroid hormones are mediated by the
in-tracellular TRs, which are encoded by two distinct
genes,c-erbA-a and c-erbA-�3. The c-erbA-ct gene is
alternativelyspliced into a1 and a2 (hormone and non-hormone
binding)isoforms, respectively. The a-isoform of the receptor is
mostabundant in heart and skeletal muscle, central nervous sys-tem,
and kidney. Intriguingly, the c-erbA-a locus has alsobeen
demonstrated to contain an overlapping transcriptionunit utilizing
coding information on the opposite strand (rev-erb) (1 8, 1 9). The
a2 and rev-erb isoforms seem to regulatethefunction ofc-erbA-a1 .
The c-erbA-a1 and a2 isoforms areconstitutively expressed during
mouse myogenesis (6). TRsbind to response elements containing
tandem direct (anddegenerate) repeats of the AGGTCA N,, AGGTCA
motif,with spacings of 3 or 4 nucleotides, and activate gene
tran-scription (20-25). TR requires accessory/auxiliary
factorspresent in nuclear extracts for high affinity binding to
their
cognate sequences (26-28). The RXR family is one of
theseaccessory proteins (20, 24-31 ). RXR is activated by
9-cis-RA.The RXRs heterodimerize with TRs and function to
selec-
tively target the high affinity binding of these receptors
to
their cognate elements (20, 24, 25). These effects requiretwo
distinct hormones and suggest direct cross-talk be-tween
9-cis-retinoic acid (32, 33) and many different hor-monal signals,
implicating RXR as having a central role inmodulating thyroid
hormone signaling pathways. Othernuclear proteins, such as the T3
receptor auxiliary protein,stabilize the binding of TR to DNA
(26-28, 31-33). Fur-thermore, the stability of these homodimeric
and heterodi-meric receptor complexes is regulated by ligand
availabil-ity (34, 35). Recently, more elaborate control of
thispathway has been suggested by the observation thatCOUP-TF, a
dopamine regulated receptor, can regulatethe binding of these
heterodimers by competitive bindingto their cognate sequences
and/or formation of nonfunc-tional heterodimers (36).
The majority of the studies investigating T3 effects onmuscle
did not distinguish between the direct and indirecteffects of the
hormone. Direct effects in skeletal muscle areclouded byT3
regulation ofgrowth hormone and insulin-likegrowth factor
production and central nervous system matu-ration and innervation,
which regulate muscle maturation(4, 5, 36). Whether T3 directly
regulates the overall growth
of skeletal muscle at the level of transcription has not
beendemonstrated conclusively. To further understand the role ofT3
in growth, we investigated whether TR and RXR directlytargeted
myogenin or whether they modulated trans-actingfactors that
regulated the precocious transcription after T3treatment.
Results
The Retinoid X and Thyroid Hormone Receptors Form Het-erodimers
That Bind with High Affinity to the Mouse Myo-genin TRE. It has
been previously demonstrated that T3treatment of myogenic cell
lines resulted in elevated expres-sion of MyoD mRNA and the
precocious expression of myo-genin mRNA, which encode tissue
specific helix-loop-helixregulatory proteins (6).
We scanned the MM promoter for a putative TRE. Usingthe core
binding motif, PuGGT,’ACA!G, we identified a pu-tative MM TRE
between nucleotide positions -526 and -494(5 I
GTGGTAGGTC1TrAGGGGTCTCATGGGACTGACA3’). This putative TRE sequence
was arranged as a directrepeat ofthe core binding motifwith a
4-nucleotide gap andwas accommodated by the tandem direct repeat
model (3-4-5 rule) proposed by Evans and Rosenfeld (22, 23).
Wecharacterized the MM TRE between nucleotide positions-526 and
-494 by demonstrating that it bound heterodimersof the thyroid
hormone and retinoid X receptor with highaffinity.
A large body ofevidence (17, 20-25, 31 , 37, 38) indicates
that the heterodimer rather than the homodimer is the entitythat
recognizes functional TREs. We investigated the abilityof
bacterially expressed thyroid hormone and retinoid X re-ceptors
(TRa and RXRa) to interact with the putative myo-genin TRE. We used
the well characterized a-MHC TRE asa heterodimerization control;
three groups (1 7, 24, 38) havedemonstrated that RXR
heterodimerizes with TR and mark-edly enhances the binding to the
a-MHC TRE. EMSA wasused to evaluate the homodimeric and
heterodimeric bind-ing of RXRa and TRce to the MM TRE. Increasing
amounts(1 .5, 4.5, and 1 3.5 pmol) of RXRa and TRa were
incubatedwith fixed quantities ofthe MM TRE and the defined
a-MHCTRE (Fig. 1 ). These data demonstrate that TRa in pmol
quan-tities did not interact with the MM TRE; however, addition
-
Probe a-MHC ThE MM .5261-494 TRE
I 11pMolTRa . . . 1.54.513.54.5 . . . 1.54.513.54.5
pMoI RXRa ii 4.5 13.5 . . . 4.5 1.5 4.5 13.5 - - - 4.5
H_._�.. ==� ID-* a.
� .4W, 1
SRE GAAGGGGACCAAATAAGGCAAGGTG
�AAMAMA
Fig. 2. A, the TREs in the MM promoter and from a number of
other char-acterized genes. The synthetic direct repeat response
elements based on the3-4-5 rule are also depicted. The sequences of
one strand of each doublestranded oligonucleotide probe are shown.
Solid arrows, direction and lo-
cation ofthe TREs; dashedarrow, location ofa weak IRE in the MM
oromoter.These sequences match the core sequence binding motif for
the thyroid her-
mone receptor: c/AcGT/AcA/(�. PAL-O is a palindromic/inverted
repeat ar-
rangement of motifs that confers a transcriptional response to
RXR, TR, and
RAR. CRBPIis a RARE in the cellular retinol binding protein I.
MEF-!, niyocyte
enhancer factor 1 ; MEF-2, myocyte enhancer factor 2; SRE, serum
responseelement. The nucleotide positions are indicated with
respect to the transcrip-tional start site at + 1 . B, the MM IRE
interacts with TRCS/RXRO heterodimerin a sequence specific fashion.
The effect of competition 10-60-fold molar
excesses), by the DR-2, GH, MHC, PAL-0, DR-4, and RARE
characterizedresponse elements and various other nonhormonal DNA
binding sites (MEF-l,MEF-2, SRE) on the complex formed between the
probe MM -526/-494 IREand the TRa/RXRa heterodimer. The molar
excess of each DNA competitor
is indicated. C, the control binding reaction in the absence of
any unlabeledcompetitor.
Cell Growth & Differentiation 903
A� -8� -8’ --�I MM -524/-496
gatcGTGGTAGGTCTTTAGGGGTCTCATGGGAC
rGH -190/-166
rMHC-127/-159
-� -� �-
-� -� �-
-� �-PAL-0 gatcTCAGGTCATGACCTGA
-� -�
mCRBPI -1016/-989 TTAGTAGGTCA.AAAGGTCAGACACTGAA
-� -�DR-2 gatcAGG?CAggAGGTCA
-� -�DR-4 gatcAGG?CAcaggAGGTCk
MEF-1
MEF-2
CCCCCCCAA�ACc.T�CTGCCTGAGCC
GCCCCATATATCAGTGATATAAATAGAACCTGCAG
k��kk ,LAg�ALA= = LA.1g�M=���� .,g
� -
Fig. 1. The thyroid hormone and retinoid X receptors form
heteromericcomplexes on the cs-MHC and MM TREs. F. coli expressed
and affinity pu-
rified cTRo and hRXRo were incubated with the a-MHC and MM IRE.
M,D, and H denote the monomeric, dimeric, and heterodimeric
complexes,respectively.
of equivalent amounts of RXRt induced heterodimerization
and markedly enhanced binding of the receptor complexesto their
cognate sequences similar to the a-MHC TRE het-
erodimerization control (Fig. 1 ). The data here suggest thatRXR
may play an important role in modulating TR interac-
tion with this sequence. These findings raise the
possibilitythat the MM TRE (-526/-494) may be a site for direct
cross-talk between two different hormonal signals (9-cis-RA and
T3) at the receptor level. Homodimeric complexes of eachreceptor
were only observed on the MM TRE with greater
than nmol amounts of receptor (data not shown).
The Interaction of the TR and RXR Heterodimer with theMouse
Myogenin TRE (-5261-494) Is Specifically Com-
peted by the rGH, a-MHC, and PAL-O TREs. We conductedEMSA
competition studies using the classical and well char-
acterized natural rGH and cr-MHC and the synthetic PAL-0and DR-4
TRE sequences to assess the sequence specific
binding of the TRY/RXRO heterodimer to the MM (-526/-494)
sequences. The sequences of the oligonucleotides
used in this study are listed in Fig. 2A, with the TRE
orien-
tations defined by arrows. The competition studies were car-ned
out at 1 0- and 60-fold molar excesses of oligonucleotidewith
respect to the myogenin TRE -526/-494 probe (Fig.
28). These studies demonstrate that the complex formed be-tween
the -526/-494 sequences and TRZ/RXRa het-
erodimer could be specifically competed by a wide variety
of established characterized wild-type and synthetic TREs(which
have been demonstrated to interact with heterodi-
meric receptor complexes). However, the CRBP I RARE,
MEF-l and -2 (binding sequences which are integral com-ponents
of all muscle specific enhancers and interact withmuscle specific
helix-loop-helix and MADS box transcrip-tion factors), and SRE did
not compete for the formation of
the heterodimeric complex on the TRE. These results mdi-
B) Probe: MM .52�/494 TREReceptor Thyroid Hormone and Rettnoid x
Receptor HeterodimerCompetitor .sas4a� DR.2 rn_H MHC PAL.0 DR.4
RARE MEF.1 MEF-2 SREMolarExcess cioeoio�io�ioeoioooio�c
cio�io�ioaoio�c
. . . . .
�. �*w� � � �
cate that the -526/-494:TRa:RXRa interaction was highlyspecific
and involved in the triiodo-F-thyronine signalingpathway.
Direct Repeats of the Sequence PuGGTCA N,, PuGGTCAwith Spacings
of 4 and 5 Specifically Competed the For-mation of the TRa/RXRa
Heterodimer on the MM IRE. Afunctional relationship among the RXR,
PPAR, COUP-TF,VDR, TR, and RAR has recently been described in
whichthese receptors bind and activate through direct repeats ofthe
AGGTCA motif separated by 1 , 2, 3, 4, and 5 nucleo-tides,
respectively (x = 1 and 2, also binds RXR and RAR) (23,37). This
rule applies to most of the natural hormone re-sponse elements. In
an effort to further characterize the MM
-
A)
-� -�
DR- 1 ga t cAGGTCAgAGGTCA
-� -�
DR-2 gatcAGGTCAggAGGTCA
-� -�
DR- 3 gatcAGGTCkaggAGGTCk
-� -�
DR- 4 gatcAGGTCAcaggAGGTCA
-�
DR- 5 gatcAGGTCAccaggAGGTCA
B) Probe:Receptor:Competitor:Molar Excess:
Fig. 3. A, The synthetic direct repeat response elements based
on the 3-4-5rule are depicted. The sequences of one stran(l of each
double stranded oh-gonucleotide probe are shown. Solid arrows,
direction and location of the
TREs. DR.1, DR-2, DR-3, DR-4, and DR-5 are arranged as direct
repeats ofthe AGGTCA motif with spacings of 1 , 2, .3, 4, and S
nucleotides. A functionalrelationship among the RXR, VDR, TR, and
RAR has recently been described
inwhich these receptors bind and activate through tandem direct
repeatsAGGTCA N AGGTCAwiIh spacingof 1, 3, 4, and S nucleotides,
respectively)x - 2 mediates a positive response to RA and negative
response to I,). 8,the MM IRE interacts with TRa/RXRa heterodimer
and was preferentially
(onipeted by direct repeats with a 4- or 5-nucleotide gap. The
effects ofdornpetition by a battery of hormone response elements.
designated DR-i,DR-2, DR-3, DR-4, and DR-S 01) the complex formed
between the MM IRE
probe and TRa/RXRo. The molar excess ofeach DNA competitor is
indicated.F, the control binding reaction in the absence 01 any
unlabeled competitor.
904 Identitication 01 IRE in Mouse Myogeiiin Gene
MM IRE
Thyroid Hormon#{149}and Ritlnold X Hetorodimar
DR-I DR.2 DR-3 DR.4 DR.5
do eoio eoio eoio eoioeo C
��WWWa.�.a4 � w
TRE:TRa/RXRct complex, we synthesized an array of syn-thetic
hormone response elements as predicted by the 3-4-5rule, arranged
as direct repeats of the AGGTCA motif withspacings of 1 , 2, 3, 4,
and 5 nucleoticles (designated DR-i,DR-2, DR-3, DR-4, and DR-5,
respectively, as described inRef. 35; Fig. 3A).
We tested these hormone response elements in EMSAcompetition
assays to assess the specificity and gap/spacingpreference ofthe
TRct:RXRs heteromeric binding to the MM-526/-494 element. DR-i,
DR-2, DR-3, DR-4, and DR-S(37) were used in the binding reactions
at 10- and 60-foldmolar excesses with respect to the MM -526/-494
probe(Fig. 38). These data indicated that DR-4 > DR-S >
DR-3preferentially competed for the formation of the
TRa/RXRtheterodimeric complex on the MM TRE. DR-i and DR-2
sequences did not compete. These experiments
rigorouslydemonstrated that the TRct and RXRa heterodimer complexon
the MM TRE was selectively competed by direct repeatsof the AGGTCA
half-site motif separated by 4 nucleotidesand provided strong
evidence thatthe -526/-494 sequencesfunctioned as a TRE site
according to the 3-4-S rule. Thisanalysis was analogous to the work
of Umesono et al. (23),which showed that DR-4 > DR-5 > DR-3
preferentially
competed for the formation of the TRI3 heteromeric complexon a
sequence that functioned as a TRE in transfection stud-es.
Nuclear Extracts from Differentiated Myogenic CellsForm a
Complex with the MM TRE That Is Specifically Com-peted by the
Wild-Type rGH and a-MHC TREs. EMSA (38)was used to determine that
the MM TRE interacted with anuclear factor in vitro. Nuclear
extracts were prepared frommouse myogenic C2C1 2 myoblasts and
myotubes in differ-ent stages ofdifferentiation to assay the
developmental regu-lation ofTRE bound proteins in muscle.
Specifically, extractswere isolated from proliferating myoblasts,
confluent myo-blasts, and myotubes (MTi and MT4;l and 4 days after
mi-togen withdrawal, respectively). These extracts were as-sayed
for the levels of Oct-i , which is a constitutively
expressed factor used to standardize the amount of
nuclearextracts used in the experiments (Fig. 4). The amount
ofMEF-2 protein (39, 40), which is a differentially
expressedfactor, was also measured in the nuclear extracts. This
dem-onstrated the pattern of expression of a nuclear factor knownto
be differentially regulated in these cells and verified
thedevelopmental stage of these extracts (data not shown).
In the EMSA experiment, the MM TRE oligonucleotidespecifically
interacted with a factor(s) in mouse myogeniccells. The protein(s)
which interacted with the MM -526/-494 sequence is a constitutively
expressed factor (Fig. 4).The rGH and a-MHC TREs and the CRBP I
RARE and MEF-iwere utilized in EMSA competition experiments to
analyzethe myogenic factors interacting with this sequence. TheMM
TRE:myogenic factor complex was shown to be spe-cifically competed
by self-competition and by the rGH anda-MHC TREs. However, the
MEF-l and RARE oligonucleo-tides did not compete for the binding
ofthis complex. Theseresults were compatible with the interaction
of the MM TREwith purified TRa.
Mutagenesis of the MM TRE Identifies the PuGGTCA Mo-tifs as
Essential TRa/RXRa and Myogenic Fador BindingSites. We
characterized the nucleotides in the MM TRE thatinteracted with
cTRa and RXRa heterodimer and the myo-genic factors by mutating the
MM TRE sequentially from the5’ to 3’ direction by nine triplet
changes that spanned thethree half-site motifs. These mutant TREs
were designatedMi-M9 (see Fig. SA) and used in EMSA competition
analy-ses to ascertain their ability to disrupt the complexes
formedbetween the MM TRE and TRct/RXRa heterodimer. We
in-dependently incubated wild-type MM TRE (-526/-494)probe with
cTRa/RXRa heterodimer (Fig. SB) and competedwith 1 0- and 60-fold
molar excesses ofthe Mi , M2, M3, M4,MS, M6, M7, M8, and M9 mutant
TREs. Fig. SB depicts theability ofthe mutant MM TRE
oligonucleotides, Ml-M9, tocompete for binding to the heterodimer
relative to the wild-
type sequence. The mutant MM TREs Ml , M4, M7, and M8
competed efficiently for binding to the TRZ/RXRa het-erodimer.
This demonstrated that the sequences flanking thedirect repeats and
the sequence ofthe 4-nucleotide gap werenot important for binding
to the complex. In contrast, M2,M3, MS, and M6 TREs did not compete
for binding to the
-
.5261494 (;H TRE �1II(’ � KARl;
C 10 20 60 10 20 60 10 20 60 10 20 60 (‘
SI..’
� � � . �. L
#{149}:
. . � � - ________
w_w w
Cell Growth & Differentiation 905
Fig. -I. The MM IRE sequence
interacts with constitutively cx-
pressed tactors in myogenic cx-tracts that are specifically
com-
peted by thyroid hormoneresponse elements. PMB andCMB, extracts
derived from prolif-crating and confluent mynhlasts.MT-t and MT-4,
extracts derivedfroni myotubes, 1 and 4 days after
mitogen withdrawal. Oct-i, a fac.
tor that is constitutively expressed
and) unaffected by the mitotic stateof the cell. The eftect of
competi-
tion 10-60-bid molar excesses),))y the GH and MHC character-ized
TREs, and the CREW I RARE
on the complex formed betweenthe MM IRE probe and
myogenictactors. The molar excess of eachDNA compelitor is
inr)icated. C,the control binding reaction inthe absence ol any
unlal)eledCOI11l)etitor. MEF- 1, myocyte en-
hancer factor 1.
l’robc �-.Oct-I,
(ompetitor
�1oIar Eitcess
Extract �
-
�S26I-494 MEF-I
10 20 60 10 20 60
heterodimeric complex. These data strongly reinforced therole of
the direct repeats and the importance of the half-sitesequence in
complex formation. The mutant M9, which hadan 8-nucleotide gap
between the wild-type direct repeats,did not compete for binding,
demonstrating that the correctgap spacing was crucial to efficient
complex formation (Fig.SB). Similar results were observed when the
mutant oligo-nucleotides were used in EMSA competition with the
myo-genic extracts (data not shown).
Sequences between Nucleotide Positions -526 and -494in the Mouse
Myogenin Promoter Confer T3 Regulation toan Enhancerless Simian
Virus 40 Promoter. In general, for-mation of TR/RXR heterodimers
that strongly and selectivelyinteract with a target sequence
correlate with a functional T5dependent trans-activation in vivo.
Transfection studieswere carried out to investigate whether the
native mousemyogenin promoter was responsive to
triiodo-L-thyronine.For this purpose, we used the construct
pMYO7O4CAT,which contained the native mouse myogenin promoter
(41)linked to the CAT gene. This plasmid contained the cis-acting
sequences between nucleotide position -704 and+41 in the myogenin
promoter. When transfected into
COS-1 cells grown in thyroid hormone and retinoid
deficientmedium in the presence of cotransfected TRct
expressionvector (pCMV-TRs) or RXRa expression vector (pCMX-RXRs),
the CAT activity from pMYO7O4CAT was increased4-6-fold after T1
treatment in the presence of TRa. In con-trast, 9-cis-RA in the
presence of RXRs did not result in anysignificant induction of CAT
activity. The activity frompMYO7O4CAT was low and was not increased
after T4 treat-ment in the absence of cotransfected TRa (data not
shown).These transfections demonstrated that the sequences be-tween
nucleotide positions -704 and +4i in the mousemyogenin promoter are
capable of mediating T5 dependenttrans-activation of CAT expression
in a receptor dependentmanner.
An important question to address is whether the
synergismobserved in strong and specific heterodimeric binding
be-tween TR and RXR on the MM TRE can be directly correlatedwith
the trans-activation activity ofthe heterodimers on theirfunctional
response element in vivo. We cloned the -526/-494 region into an
enhancerless SV4O promoter (pCATpromoter-I); Promega) and conducted
experiments to see
whether the putative MM TRE conferred appropriate hor-monal
regulation to a heterologous promoter. The oligo-nucleotide was
cloned as a single copy in the correct on-entation upstream of the
SV4O promoter in the pCATpromoter plasmid (pMM-TRE-CAT-p). When
transfectedinto CV-i cells grown in thyroid hormone and retinoid
de-ficient medium in the presence of cotransfected TRct alone,the
MM TRE sequences conferred a -4-fold induction afterT3 treatment to
the enhancenless SV4O promoter in a recep-tor dependent manner
(Fig. 6, Lanes 5 and 6). The MM TREsequences did not show any
significant induction in thepresence of RXRt alone and RA above
that mediated by thebasal vector, pCAT-p, alone (Fig. 6, Lanes 3
and 4 versusLanes 7 and 8). The T3 response was minimally
increased,and the absolute CAT activity was increased by
cotransfec-tion of RXRs. These results confirmed that the
cis-actingregion between nucleotide positions -526 and -494
func-tions as a positively acting TRE in the muscle specific
mousemyogenin promoter (Fig. 6, Lanes 9 and 10).
This scenario is analogous to the malic enzyme TRE stud-ies
demonstrating that it was a very efficient binder of TRa/RXRa
heterodimers, conferred a 3.5-fold trans-activation toa
heterologous promoter after T3 treatment in the presenceof TR, but
did not function as a more effective functionalresponse element for
the TR/RXR heterodimen in the contextof TR/RXR cotransfection
experiments (38). These data werealso consistent with previous
observations that demon-strated that some response elements (e.g.,
PAL-0) requireparticular receptor isoform combinations for
efficient tran-scriptional activation in cell culture. We are
currently ex-amining these options since all RXR isoforms (co-,
f3-, and -y-)are expressed in muscle tissue (TRI3 is not expressed)
(30).
Discussion
T3 has been shown to promote terminal muscle differentia-tion in
vivo, positively regulates MyoD, and induces the pre-mature
expression of myogenin mRNA in myogenic culture(6). In the present
study, we have shown that the mousemyogenin promoter is regulated
by � and that the cis-actingsequences between nucleotide positions
-704 and +41were required for T jFR mediated -=5-fold
trans-activation ofCAT expression. These data indicated that the �
influenced
-
A)-*. N� -a-
WI LD TYPE �C - 52 6 1 - 4 94 )
CACCATCCAGAAATCCCCAGAGTACCCTGACTGT
Ml GTAAAAGGTCTTTAGGGGTCTCATGGGACTGACA
CATTTTCCAGAAATCCCCAGAGTACCCTGACTGT
M2 GTGGTTTTECTTTAGGGGTCTCATGGGACTGACA
CACCAAAMGAAATCCCCAGAGTACCCTGACTGT
M3 GTGGTAGGAAATTAGGGGTCTCATGGGACTGACA
CACCATCCT?TAATCCCCAGAGTACCCTGACTGT
M4 GTGGThGGTCTCA000GGTCTCATGGGACTGACA
CACCATCCAGAGTCCCCCAGAGTACCCTGACTGT
MS
GTGGTAGGTCTTTAGTTTTCTCATGGGACTGACACACCATCCAGAAATCA�AAGAGTACCCTGACTGT
M6
GTGGTP�GGTCTTTAGGGGAAACATGGGACTGACACACCATCCAGAAATCCCC???GTACCCTGACTGT
Ml GTGGTAGGTCTTTA0000TCTTIAGGGACTGACA
CACCATCCAGAAATCCCCAGSAATCCCTGACTGT
MO GTGGTAGGTCTTThG000TCTCATTTTACTGACA
M9 GTGGTAGGTCTTTAGCAOGGGGTCTCAT000ACTG
CACCATCCAGAAATCG?�CCCAGAGTACCCTGAC
B) Probe: MM IREReceptor Thyroid Hormone and Retinold X Receptor
Heterodimer
Competitor .5215494 Ml � i� i&� t�s us �i �s yeMolarExcess
c1s�i1o�o1s#{248}1oo1o�o1o�o c C 1I�I1C�I1O6O iO6OC
.4 ...4�.. . S. � �
...� � #{149}-�j - �
CACCATCCAGAAATCCCCAGAGTAAAATGACTGT
V IV ‘ ‘ � VI)
Fig. 5. A, pictorial representation ofthe various site specific
mutations in the
mouse myogenin IRE. The wild-type IRE sequence is depicted, and
the mu-
tations in the Mi , M2, M3, M4, M5, M6, M7, M8, and M9 TREs are
denotedby bold text. B, mutational analyses of the mouse myogenin
TRE. The corereceptor binding motifs in the MM -5261-494 IRE are
important for binding.The effects of competition by a battery of
mutations in the MM IRE, designated
Mi , M2, M3, M4, M5, M6, M7, M8, and M9 on the complex formed
betweenthe MM IRE probe and TRa/RXRa heterodimer. The molar excess
of each
DNA competitor is indicated. C, the control binding reaction in
the absenceof any unlabeled competitor.
induction of myogenin mRNA in myogenic cell lines in-volves
transcriptional mechanisms. Transfection expeni-ments and EMSA were
used to identify a functional TRE inthe mouse myogenin gene. This
TRE is located betweennucleotide positions -526 and -494 (5’
GTGGTAGGTCTT-TAGGGGTCTCATGGGACTGACA 3’) and is accommo-dated by the
cone receptor binding motif, AGGTCA. ThisTRE sequence conferred a
-=4-6-fold induction to an en-hancenless SV4O promoter, after T9
treatment in a TR de-pendent manner. This level of inducibility
correlated withthe -4-6-fold trans-activation of the human skeletal
a-actinand the rodent sancoplasmic reticulum Ca2� ATPase pro-motens
after T3 treatment, which are primarily expressed in
skeletal and cardiac muscle, respectively.EMSA experiments
showed that Escherichia coli
expressed/affinity purified TRcs and RXRa formed a hetenodi-
906 Identification of IRE in Mouse Myogenin Gene
meric complex that bound very efficiently to the MM TRE.This is
in agreement with all ofthe recentdata thatdefine theheterodimer
rather than the homodimer as the entity thatrecognizes thyroid
hormone response elements. Further-more, the TR/RXR complex on the
MM TRE was specificallyand selectively competed by other natural
and syntheticTREs; in contrast, a natural RARE and the DR-i, DR-2,
andDR-3 sequences that interacted with a wide variety ofother
steroid receptors (RXR/PPAR, COUP-TF, and VDR)and the MEF-1, MEF-2,
and SRE sites did not compete forthe formation of the TR/RXR
complex on the MM TRE.These in vitro binding data correlated with
the ability ofthese receptors after hormone treatment to
trans-activatethe myogenin promoter in vivo, in agreement with a
van-ety of other studies.
Mutagenesis of the MM TRE indicated that direct repeatsof the
cone AGGTCA binding motifs and the 4-nucleotidegap were necessary
and sufficient for efficient TRZ/RXRahetenodimer formation.
Furthermore, the ability of the MMTRE sequences to function in vivo
and confer hormonalregulation to a heterologous promoter correlated
with theheterodimenic interaction of purified TRa and RXRct
withthese sequences in vitro. Heterodimer formation on hor-monal
response elements has been previously shown to con-relate with
effective transcriptional activation (38).
The MM TRE interacted with a nuclear factor(s) from myo-genic
nuclean extnacts that were specifically competed bythe
characterized a-MHC and GH TREs but not by otherDNA elements.
Furthermore, these sequences interactedwith a nuclear factor(s) in
vitro that was constitutively ex-pressed during the differentiation
of mouse myogenic celllines. This correlates with the results
ofCannac etal. (6), whoobserved the constitutive expression
ofc-enbA-a1 during thedifferentiation of mouse myogenic cell lines.
These bindingand competition expeniments with myogenic extracts
con-related well with the intenaction ofthe MM TRE with
purifiedreceptors in vitro.
In summary, our experiments have demonstrated that the-526/-494
cis-acting sequence interacts with the hetenodi-menic TR/RXR
receptor complex, with high affinity and se-
quence specificity similar to the well characterized a-MHCTRE.
Furthermore, we demonstrated that the direct repeatcone sequences
and the 4-nucleotide gap spacing in thiselement wene necessany and
sufficient for the formation ofthe heterodimenic complex and that
this sequence confennedan appropriate T3 response to a hetenologous
promoter. Thissatisfies the criteria that have evolved to define
thyroid hon-mone nesponse elements.
Our experiments suggested that RXRs may modulate theinteraction
of TRa to the MM TRE and that the -526/-494sequence in the mouse
myogenin gene is a target for dinectcross-talk between two
different hormonal signals (T3 and9-cis-RA) at the receptor level.
To our knowledge, this is thefirst demonstration of hetenodimenic
binding of TRcr andRXRa to a TRE in a helix-loop-helix
transcription factor, ex-pressed specifically in skeletal muscle,
that trans-activatescontractile protein gene transcription. Our
results provide amolecular explanation for the precocious
expression ofmyogenin observed after T3 treatment of C2 myogenic
cul-tunes and the induction of myogenin transcription by RA
inrhabdomyosancoma cell lines (6, 42). Our study, which pro-vides a
target for T3 and 9-cis-RA on the myogenin gene,correlates with the
appearance of myogen in in the myotomalcompartment of the somites
8.5 days post coitum and RXRin day 1 0 mouse embryos. Furthermore,
we have observed
-
I+
+ +
+
+ +
+
+ +
+ +
+ + + +
0 20000 40000
�1”I . �1�’
6000080000100000120000
C.P.M.
Cell Growth & Differentiation 907
4 M. Downes, unpublished observation.
Fig. 6. Ihe MM IRE confers hor-monal regulation to an
enhancer-less SV4O promoter. The mean val-
ues are shown for the plasmid,pMM-TRE-CAI-p, containing 1
copy of the MM IRE cloned intothe pCAT promoter construct in
the correct orientation. Results areshown
fortransienttransfections in
CV-1 cells. Mean values and SDs(bars) were derived from a
tripli-cate experiment repeated twice.
Plasmid
1. pCAT-p
2. pCAT.p
3. pCAT-p
4. pCAT-p
5. pMM-TRE-CATp
6. pMM-TRE-CAT.p
7. pMM-TRE-CAT-p
8. pMM.TRE-CAT.p
9. pMM-TRE-CAT-p
10. pMM.TRE.CAT-p
TR RXR 13 RA
+
+ +
that myogenin, RXRa, and RXRy mRNA are coordinatelyinduced
during C2 differentiation after mitogen with-dnawal.4
Thyroid hormone levels play a major role in
regulatingcontractile protein isoform expression in vivo during
adultmuscle development. The effects of thyroid hormone levelsin
adult animals are only apparent in long term
intoxication.Hypothyroid rat neonates have a reduced number of
myo-fibers and a reduced muscle mass. In soleus muscle, T3
treat-ment results in an increase in type II fast fibers,
whereashypothyroidism reduces the number of these fibers. In
hy-pothyroid rats, expression of adult fast myosin is
drasticallyreduced, and neonatal myosin is abnormally high. The
de-velopmental program is slowed in hypothyroidism (4, 5).Although
c-enbA-a is abundantly expressed in muscle, it wasnot possible to
decide whether this is a direct effect of hor-mone action on fiber
development on an indirect conse-quence of retardation in the
maturation of the central ner-vous system. Recently, it was
demonstrated that MyoD andmyogenin mRNAs were stimulated and
precociously in-duced by thyroid hormone (6). This provided a
molecularbasis for the positive regulation of muscle development,
be-cause these muscle specific helix-loop-helix factors
trans-activate contractile protein gene transcription. The
ques-tions remaining were: (a) whether thyroid hormones viac-erbA-a
directly regulate skeletal muscle specific targets;(b) whether they
act by regulating transcription factors,which in turn regulate
thyroid hormone responsive genes inmuscle; and (c) whether T3
directly targets the MyoD family,which subsequently trans-activates
the contractile proteingenes; or whether T3 targets the
hierarchical HLH regulatorsand the contractile protein genes
simultaneously.
The skeletal a-actin mRNA, which is part of the thin fila-
ment in the sarcomene, increases in the adult heart
duringcardiac hypertrophy after the imposition of
hemodynamicoverload/’aortic restriction. T3 has been shown to cause
arapid increase in the amount of skeletal a-actin mRNA inhearts
from normal and hypophysectomized animals (Ref.43 and references
therein). We recently conducted experi-ments that indicated that T3
induced increases in a-actinmRNA in animals were mediated by direct
transcriptionalmechanisms which involved the cis-acting sequences
be-tween nucleotides -432/-i 53, c-erbA-a, and interactionswith
ubiquitous proteins (SpI, SRF, and NF-i!CTF) (43). Wesubsequently
identified a TRE in the human skeletal cx-actingene between
nucleotide positions -273 and -249, whichconferred appropriate
hormonal regulation to an enhancer-less SV4O promoter. TRa and RXRa
interacted with thea-actin TRE in a sequence specific fashion and
formed het-eromenic complexes on the a-actin TRE. These data
sug-gested that the human skeletal a-actin TRE is a target
fordirect cross-talk between two different hormonal signals (T3and
9-cis-RA) at the receptor level (1 7).
These resutts and the present data with the mouse myo-genin
promoter, as well as our previous observations dem-onstrating that
skeletal cr-actin was trans-activated by MyoD,myogenin, MRF-4 (44),
and MEF-2 (39, 40), firmly suggestthat 13 targets the hierarchical
MyoD family of regulatorsand the contractile protein genes
simultaneously, and thatfine tuned expression is achieved by the
MyoD family andother intermediate regulators such as MEF-2, which
arepart of a positive autoregulatory loop that maintains the
re-
quired threshold levels of these muscle specific transcrip-tion
factors.
Materials and MethodsCell Culture and Transfedion. Mouse
myogenic C2 cellswere grown in DMEM supplemented with 20% FCS in 1
0%
-
908 Identification of IRE in Mouse Myogenin Gene
CO2 as described previously (45). This cell line was inducedto
biochemically and morphologically differentiate intomultinucleate
myotubes by mitogen withdrawal (DMEMsupplemented with 2% FCS in 10%
CO2). Differentiationwas essentially complete within 72-96 h with
respect toisoform switching in the actin multigene family (22).
How-ever, these cells will spontaneously differentiate at a
veryhigh confluence (1 00%) in the presence of mitogens. COS-1cells
were grown in DMEM supplemented with 1 0% FCS iniO% CO2.
Each 60-mm dish of cells was transiently tnansfected with
S pg of reporter plasmid DNA expressing CAT, mixed with2-3 pg of
receptor expression vectors and pUCi 8 plasmidto a total of 1 2 pg
in each transfection experiment. Prior totnansfection, the cells
were cultured for 24 h in T3 and T4deficient medium containing S%
charcoal stripped FCS inDMEM. The DNA mixtures were contransfected
into CV-ifibroblasts by the liposome mediated procedure. We usedthe
cationic lipid DOTAP. Unilamellan vesicles were formedby mixing the
appropriate DNAs with 30-40 p1 of DOTAPand 1 x HEPES buffered
saline to a total volume of 200 ml.After a 1 0-mm incubation at
room temperature, this mixturewas added to 6 ml of fresh culture
(13 and T4 deficient)medium and added to the cells, which were at
S0-70% con-fluence. After a period of 20-24 h, fresh medium with
onwithout T3 (i0� M) was added to the cells. The cells
wereharvested for the assay ofCAT enzyme activity 48 h after
thetransfection period. Each transfection experiment was per-formed
three times using at least two different plasmid prepa-rations in
order to overcome the variability inherent in trans-fections.
CAT Assays. The cells were harvested, and the CAT ac-tivity was
measured as previously described (33). Aliquots ofthe cell extracts
were incubated at 37#{176}C,with 0.1-0.4 mCiof [t4Clchloramphenicol
(Amensham) in the presence of SmM acetyl CoA and 0.25 M Tnis-HCI,
pH 7.8. After a 20-mmto 3-h incubation period, the reaction was
stopped by theaddition of 1 ml ethyl acetate, which was used to
extract thechloramphenicol and its acetylated forms. The extracted
ma-tenials were analyzed on silica gel thin layer
chromatographyplates as described previously (33). Quantitation of
CAT as-says was performed by an AMBIS �3-scanner.
Plasmids. The plasmid pCAT promoter (an enhancerlessSV4O
promoter linked to CAT in a pUC1 9 backbone) waspurchased from
Promega. The plasmid, pCMV-TRa, ex-pressing the rodent c-erbAa gene
in the eukaryotic expres-sion vector CMV4, containing the
cytomegalovirus pro-moten and SV4O origin of replication, was
described by Zilzetal. (46). The plasmid pCMX-hRXRa, expressing the
humanRXRa gene in the eukaryotic expression vector pCMX, apCDM 8
derivative containing the cytomegalovmrus pro-moter and SV4O origin
of replication, was described byKliewer et al. (20). We constructed
pGEX-1 -cTRa by excis-ing the chicken TRa complementary DNA from
the pSG5expression vector by EcoRl digestion. This complementaryDNA
insert was then cloned in frame into EcoRl cleavedpGEX-1 (1 7). The
pGEX-2T-hRXRa was described by Man-gelsdorf et al. (21 ). We
constructed pMYO7O4CAT, whichcontained the sequences between
nucleotide positions-704 and +41 of the native myogenin promoter,
by po-lymerase chain reaction amplification of mouse EMBL 5genomic
DNA with the following primers: GMUQ68,
GCGAAGCTTCCCAACATCATGAGACCTGGTCA, andGMUQ,69,
GCGAAGCTTGTCGGAAAAGGCTTGTrCCT-GCCACTGGCCC. We utilized a
94#{176}C/i-mm denaturation,
59#{176}C/i-mm annealing, followed by a 65#{176}C!2-min
extensionfor 35 cycles. The product was cleaved with Hindlll
andcloned into the Hindlll site of pCAT-basic in the
senseorientation.
Oligonucleotides. The sequences of the oligonucleotideprobes
used in the EMSA experiments were described in thelegends to Figs.
3A, 4A, and 6A. The sense and antisensestrands of the MM TRE
sequences with gate ends were an-nealed, phosphorylated with T4
polynucleotide kinase, andself-ligated with T4 DNA ligase. These
products were thencloned into the BgIll site in the pCAT promoter
vector (Pro-mega) and initially screened by EcoRl digestion. Clones
con-taming putative inserts were sequenced by double
strandedsequencing to determine the orientation and number of
cop-ies cloned.
Expression and Purification of Receptors. Human RXRaand chicken
TRa were expressed as fusions withglutathione-5-transferase using
the pGEX-2T and pGEX-1bacterial expression vectors. BL21(DE3)pLysS
cells or DH5acells containing these expression vectors were induced
for1-2 h with 0.4 mM isopropyl thiogalactoside after the cellshad
grown to an A600 of -0.6. Pelleted cells were lysed; theclarified
lysates containing the fusion proteins were loadedonto
glutathione-agarose columns in Dignam buffer C (con-taming 0.5 m�
PMSF and 2 mg/mI of leupeptin and apro-tinin). After extensive
column washing, the fusion proteinwas eluted with Dignam buffer C
containing 5 miss reducedglutathione.
Nuclear Extrads and Gel Mobility Shift Assays. Nuclearextracts
were prepared as described previously (45). Briefly,the cells were
lysed using 10% Nonidet P-40 following in-cubation in iO m� HEPES
(pH 7.9), iO mivt KCI, 0.1 m�iEDTA, 0.i mM EGTA, 1 mt�i
dithiothreitol, 0.5 ms,i PMSF, and2 mg/mI ofleupeptin and aprotinin
(Boehringer-Mannheim).Nuclear proteins were extracted with 0.4 M
NaCl, 20 ms�iHEPES (pH 7.9), i mM EDTA, i mt�i EGTA, 1 mt�i
dithio-threitol, 1 mM PMSF, and 2 mg/mI aprotinin and
Ieupeptin.
Each binding mixture (25-30 pI) contained 1-2 ng ofa
T4polynucleotide kinase labeled DNA fragment, 1-30 pmol ofpurified
receptors or 5-i 0 pg of crude nuclear extract, and1-2 pg of
polydeoxymnosmnic-deoxycytidylic acid as a non-specific competitor
(only when crude nuclear extract wasused) in Dignam buffer C. The
assays were incubated atroom temperature for 20 mm and
electrophoresed througha 6% (20:i polyacrylamide:bisacrylamide) gel
in 80 m�i Trisborate and 2 m� EDTA. Gels were briefly soaked in
10%acetic acid, dried, and autoradiographed.
Synthesis of 9-cis-Retinoic Acid. 9-cis-Retinoic acid
wasprepared from the aldehyde by reacting with sodium cya-nide over
manganese dioxide to produce methyl ester,which was subsequently
hydrolyzed to the acid in alcoholicpotassium hydroxide. The product
was purified using highperformance liquid chromatography on a C18
column (33).
AcknowledgmentsWe sincerely thank Dr. Howard lowle for
generously providing the rat
c-erbA-a in the CMV4 expression vector and Dr. Ron Evans for the
plasmid,pGEx-2T-hRXRa, which expresses the retinoid x receptor in
Escherichia coli.
References1 . Olson, E. N. MyoD family: a paradigm for
development. Genes & 0ev.,4: 1454-1456, 1990.
2. Olson, E. N. Interplay between proliferation and
differentiation within the
myogenic lineages. Dev. Biol., 154: 261-272, 1992.
3. Weintraub, H., et a!. The MyoD gene family: nodal point
during speci-
-
Cell Growth & Differentiation 909
fication ofthe musclecell lineage. Science(Washington DC),
251:761-768,1 991.
4, Lompre, A. M., Mercadier, J. j., and Schwartz, K. Changes in
gene ex-pression during cardiac growth. Int. Rev. Cytol., 124:
137-186, 1991.
5, Swynghedauw, B. Developmental and functional adaption of
contractile
proteins in cardiac and skeletal muscle. Physiol. Rev., 66:
710-771, 1986.
6. Carnac, G., eta!. 3,5,3’-Triiodothyronine positively
regulates both MyoDigene transcription and terminal differentiation
in C2 myoblasts. Mol. Endo-
crinol., 6: 1 185-1 194, 1992.
7. Gambke, B., Lyons, G. E., Haselgrove, j., Kelly, A. M., and
Rubinstein, N.A. Thyroidal and neural control of myosin transitions
during development of
rat fast and slow muscles. FEBS Len., 156: 335-339, 1983.
8. lanuzzo, D., Pate), P., Chen, V., O’Brien, P., and Williams,
C. Thyroidal
trophic influence on skeletal muscle myosin. Nature (Lond.),
270: 74-76,1977.
9, Sugie, H., and Verity, M. A. Postnatal histochemical fiber
type differen-tiation in normal and hypothyroid rat soleus muscle.
Muscle & Nerve, 8:
654-660, 1985.
10, Butler-Browne, G. 5,, Herlicoviez, D., and Whaler, R. G.
Effects of hy-
pothyroidism on myosin isozyme transitions in developing rat
muscle. FEBSLeg,, 166:71-75, 1984.
1 1 , Whalen, R. G., Toutant, M., Butler-Browne, G. S., and
Watkins, S. C.
Hereditary pituitary dwarfism in mice affects skeletal and
cardiac myosinisozyme transitions differently. J. Cell Biol., 101:
603-609, 1985.
1 2. Fremont, P., Lazure, C., Tremblay, R. R., Chretien, M., and
Rogers, P. A.
Regulation of carbonic anhdrase Ill by thyroid hormone opposite
modulationin slow and fast-twitch skeletal muscle. Biochem, Cell
Biol., 65: 790-796,
1987.
1 3. Izumo, S., Nadal-Ginard, B., and Mahdavi, V. All members
ofthe MHC
multigene family respond to thyroid hormone in highly
tissue-specific man-ner. Science (Washington DC), 23 1 : 597-bOO,
1986.
14, Simonides, W. S., d. Linden, G. C. V., and Hardeved, C. V.
Thyroid
hormone differentially affects mRNA levels of Ca-ATPase
isoenzymes of sar-
coplasmic reticulum in fast and slow skeletal muscle. FEBS
Lett., 274: 73-76,
1990,
1 5. Thompson, 1,, Moore, S. E., and Walsh, F. S. Thyroid
hormones regulate
expression ofthe neural cell adhesion molecule in adultskeletal
muscle. FEBSLett., 219: 135-138, 1987.
16, Rohrer, D, K,, Hartong, R., and Dillmann, W. H. Influence of
thyroidhormone and retinoic acid on slow sarcoplasmic reticulum
Ca2’ ATPase and
myosin heavy chain a gene expression in cardiac myocytes. j.
Biol. Chem.,266: 8638-8646, 1991,
1 7. Muscat, G. E. 0., Griggs, R., Downes, M., and Emery, J.
Characterizationofthe thyroid hormone response element in the
skeletal a-actin gene: negativeregulation of I, receptor binding by
the retinoid X receptor. Cell Growth &
Differ,, 4:269-279, 1993,
1 8. Glass, C. K., and Holloway, j. M. Regulation of gene
expression by the
thyroid hormone receptor. Biochim. Biophys. Acta, 1032: 157-176,
1990.
1 9. Forman, B. M., and Samuels, H. H. Interactions amoung a
subfamily of
nuclear hormone receptors: the regulatory zipper model. Mol.
Endocrinol.,4: 1293-1301, 1990,
20. Kliewer, S. A., Umesono, K., Mangelsdorf, 0. j., and Evans,
R. Retinoidx receptor interacts with nuclear receptors in retinoic
acid, thyroid hormoneand vitamin D� signalling. Nature (Lond.),
355:446-449, 1992.
21 . Mangelsdorf, D, j., Umesono, K., Kliewer, S. A., Borgmeyer,
U., Ong, E.S., and Evans, R. M. Adirect repeat in the cellular
retinol-binding protein type
II gene confers differential regulation by RXR and RAR. Cell,
66: 555-561,1991.
22. Naar, A. M., Boutin, J-M., Lipkin, S. M., Yu, V. C.,
Holloway, j. M., Glass,
C, K., and Rosenfeld, M. G. The orientation and spacing ofcore
DNA-bindingmotifs dictate selective transcriptional responses to
three nuclear receptors.
Cell, 65: 1267-1279, 1991.
23. Umesono, K., Murakami, K. K., Thompson, C. C., and Evans, R.
M. Directrepeats as selective response elements for the thyroid
hormone, retinoic acid,
and vitamin 0, receptors. Cell, 65: 1255-1266, 1991.
24. Yu, ‘I. C., Dolsert, C., Anderson, B., Holloway, J. M.,
Devary, 0. V., Naar,A. M., Kim, S. V., Boutin, J. M., Glass, C. K.,
and Rosenfeld, M. G. RXRf3: acoregulator that enhances binding of
retinoic acid, thyroid hormone, and
vitamin D receptors to their cognate response elements. Cell,
67:1251-1266,
1 991.
25. Zhang, X-k., Hoffmann, B., Iran, P. B. V., Graupner, G., and
Pfahl, M.Retinoid X receptor is an auxiliary protein for thyroid
hormone and retinoic
acid receptors. Nature (Lond.), 355: 441 -449, 1992.
26. Darling, D. S., Beebe, J. S., Burnside, J., Winslow, E. R.,
and Chin, W. W.
3,5,3’-Triiodothyronine (13) receptor-auxiliary protein (TRAP)
binds DNAand forms heterodimers with the 13 receptor. Mo).
Endocrinol., 5: 73-84,
1991.
27. Glass, C. K., Devary, 0. V., and Rosenfeld, M. G. Multiple
cell type-specific proteins differentially regulate target sequence
recognition by the aretinoic acid receptor. Cell, 63: 729-738,
1990.
28. Murray, M. B., and Towle, H. C. Identification of nuclear
factors that
enhance binding of the thyroid hormone receptor to a thyroid
hormone re-
sponse element. Mol. Endocrinol., 3: 1434-1442, 1989.
29. Mangelsdorf, D. J., Ong, E. S., Dyck, J. A., and Evans, R.
M. Nuclearreceptor that identifies a novel retinoic acid response
pathway. Nature
(Lond.), 345:224-233, 1990.
30. Mangelsdorf, D. j., Borgmeyer, U., Heyman, R. A., Yang Zhou,
j., Ong,E. S., Oro, A. E., Kakizuka, A., and Evans, R. M.
Characterization ofthree RXR
genes that mediate the action of 9-cis retinoic acid. Genes Dev.
6: 329-344,1992.
31 . Leid, M. Purification, cloning, and RXR identity of the
HeLa cell factor
with which RAR or TR heterodimerize to bind target sequences
efficiently.
Cell, 68:377-395, 1992.
32. Levin, A. A., Sterzonbecker, L. J., Kazmer, S., Bosakowski,
I., Huselton,C., Rosenburger, M., Lovey, A., and Grippo, J. F.
9-cis-Retinoic acid stereo-isomer binds and activates the nuclear
receptor RXR a. Nature (Lond.), 355:
359-361, 1992.
33. Heyman, R. A., Mangelsdorf, D. J., Dyck, J. A., Stein, R.
B., Eichele, G.,Evans, R. M., and Thaller, C. 9-cis-Retinoic acid
is a high affinity ligand for
the retinoid x receptor. Cell, 68: 397-406, 1992.
34. Zhang, X-K., et a!. Homodimer formation of retinoid x
receptor inducedby 9-cis retinoic acid. Nature (Lond.), 358:
587-590, 1992.
35. Kliewer, S., Umesono, K., Heyman, R., Mangelsdorf, D., Dyck,
1., andEvans, R. Retinoid X receptor-COUP-IF interactions modulate
retinoic acid
signaling. Proc. NatI. Acad. Sci. USA, 89: 1448-1452, 1992.
36. Cooney, A. J., Tsai, S. Y., O’Malley, B. W., and Tsai, M-J.
Chicken ov-albumin upstream promoter transcription factor (COUP-IF)
dimers bind to
different GGTCA response elements, allowing COUP-TF to repress
hormonalinduction of the vitamin D3, thyroid hormone and retinoic
acid receptors.
Mo). Cell. Biol., 12:4153-4163, 1992.
37. Green, S. Promiscuous liaisons. Nature (Lond.), 361:
590-591, 1993.
38. Hermann, I., Hoffmann, B., Zhang, x-k., Iran, P., and Phaf),
M. Het-rodimeric receptor complexes determine triiodothyronine and
retinoid sig-naIling specificities. Mol. Endocrinol., 6: 1 1 53-1 1
62, 1992.
39. Muscat, G. E. 0., Perry, S., Prentice, H., and Kedes, L. The
human skeletala-actin gene is regulated by a muscle specific
enhancer that binds threenuclear factors. Gene Expression, 2: 1 1 1
-1 26, 1992.
40. Gosset, L. A., Kelvin, D. j., Sternberg, E. A., and Olson,
E. A new myocytespecific enhancer binding factor that recognises a
conserved element asso-
ciated with multiple muscle specific genes. Mol. Cell. Biol., 9:
5022-5033,1989.
41 . Edmondson, D. G., Cheng, I-C., Cserjesi, P., Chakraborty,
1., and Olson,E. N. Analysis of the myogenin promoter reveals an
indirect pathway forpositive autoregulation mediated by the muscle
specific enhancer factor-2.
Mo). Cell. Biol., 12: 3665-3677, 1992.
42. Arnold, H. H., Gerharz, C. D., Gabbert, H. E., and Salminen,
A. Retinoic
acid induces myogenin synthesis and myogenic differentiation in
the rat rhab-
domyosarcoma cell line BA-Han-iC. J. Cell Biol., 118:877-887,
1992.
43. Collie, E. S. R., and Muscat, G. E. 0. The human skeletal
a-actin promoter
is regulated by thyroid hormone: identification ofa thyroid
hormone responseelement. Cell Growth & Differ., 3: 31-42,
1992.
44. Muscat, G. E. 0., and Emery, 1.. and Collie, E. S. R. Tissue
specific ex-pression of the skeletal a-actin gene involves
sequences that can function
independently of MyoD and Id. Gene Expression, 2: 241-258,
1992.
45. Muscat, G. E. 0., Gobius, K., and Emery, I. Proliferin, a
prolactin/growthhormone-like peptide, represses myogenic specific
transcription by the sup-
pression ofan essential serum response factor-like DNA binding
activity. Mol.Endocrinol., 5:802-814, 1991.
46. Zilz, N. D., Murray, M. B., and lowle, M. Identification of
multiple thy-roid hormone response elements located far upstream
from the rat 51 4 pro-
moter. J. Biol. Chem., 265:8136-8143, 1990.47. Brent, G. A.,
Williams, G. R., Harney, J. W., Forman, B. M., Samuels, H.H.,
Moore, D. D., and Larsen, P. R. Capacity for co-operative binding
of the
thyroid hormone receptor dimers defines wild-type T� response
elements.Mol. Endocrinol., 6: 502-514, 1992.
