Induction of Larval Tissue Resorption in Xenopus laevis Tadpoles by the Thyroid Hormone Receptor Agonist GC-1* J. David Furlow‡**, Hayung Yang‡, Mei Hsu‡§, Wayland Lim‡║, Davy J. Ermio‡, Grazia Chiellini¶, and Thomas S. Scanlan¶ From the ‡ Section of Neurobiology, Physiology, and Behavior, University of California, Davis, CA 95616-8519 and the ¶Departments of Pharmaceutical Chemistry and Cellular and Molecular Pharmacology, University of California, San Francisco, CA 94143-0446 Present address: §Department of Biochemistry, Dartmouth College, Hanover, New Hampshire 03755; ║St. Louis University School of Medicine, St. Louis, MO 63104 **To whom correspondence should be addressed: Section of Neurobiology, Physiology, and Behavior, University of California, One Shields Avenue, Davis, CA 95616-8519. e-mail: jdfurlow@ucdavis.edu. Running title: Receptor isotype regulation of Xenopus metamorphosis * This work was supported by grants from the National Institutes of Health (DK-52798 to T.S.S. and DK-55511 to J.D.F.)

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

2

A major challenge in understanding nuclear hormone receptor function is to determine

how the same ligand can cause very different tissue specific responses. Tissue specificity

may result from the presence of more than one receptor subtype arising from multiple

receptor genes or by alternative splicing. Recently, high affinity analogs of nuclear receptor

ligands have been synthesized that show subtype selectivity. These analogs can greatly

facilitate the study of receptor subtype-specific function in organisms where mutational

analysis is problematic or where it is desirable for receptors to be expressed in their normal

physiological contexts. We describe here the effects of the synthetic thyroid hormone

analog GC-1 on metamorphosis of the frog Xenopus laevis. The most potent natural thyroid

hormone, 3’, 3, 5- triiodothyronine or T3, shows similar binding affinity and

transactivation dose response curves for both thyroid hormone receptor isotypes,

designated TR α and TR β. GC-1, however, binds to and activates TR β at least an order of

magnitude better than TRα. GC-1 efficiently induces the larval tissue death and resorption

in premetamorphic tadpoles such as the gills and tail, two tissues that strongly induce

thyroid hormone receptor β during metamorphosis. GC-1 has less effect on the growth of

adult tissues such as the hindlimbs that express high TRα levels. The effectiveness of GC-1

in inducing tail resorption and tail gene expression correlates with increasing TRβ levels.

These results illustrate the utility of subtype selective ligands as probes of nuclear receptor

function in vivo.

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

3

Nuclear hormone receptors and their ligands are essential regulators of metazoan development,

reproduction, and homeostasis. One of the most dramatic nuclear receptor mediated events in

nature is amphibian metamorphosis (1,2). In the frog, metamorphosis occurs as a result of gene

expression cascades induced by thyroid hormone (TH) secreted from the developing tadpole’s

thyroid gland (3). These cascades result in dramatically different morphological responses such

as growth and differentiation of the limbs, death and resorption of the larval gills and tail, and

remodeling of a large number of larval organs for new adult functions (1). As in all vertebrates,

TH exerts its effects via a pair of TH receptor subtypes encoded by separate genes, TRα and TRβ

(4). TRs are ligand-regulated transcription factors that are thought to repress target gene

transcription in the absence of hormone and activate transcription upon hormone binding (5).

The most detailed molecular studies on amphibian metamorphosis have been done on the frog

Xenopus laevis. Xenopus TRα is expressed early in development (6,7), long before the embryo

and larval tadpole have a functional thyroid gland. Just prior to metamorphosis, xTRα is

ubiquitously expressed but particularly high levels are detected in the brain, limb buds, skin, and

other tissues that are destined to respond to TH by proliferating and differentiating into adult

organs (8,9). In contrast, much lower levels of Xenopus TRβ mRNA are detectable prior to

metamorphosis (6,10). At metamorphosis, TRβ is strongly induced by TH in larval tissues that

will die and resorb such as the tail. TRβ levels remain very low in the growing limbs, although

expression in this tissue may be highly localized (11). In remodeling tissues, such as the intestine

and cartilage of the head, TRβ mRNA is up-regulated in both dying and proliferating cells

suggesting a possible role for TRβ in both processes (8,12,13).

Xenopus TRs are extremely well conserved compared to their avian and mammalian

counterparts (4). xTRα is over 95% and 91% conserved in its DNA and ligand binding domains,

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4

respectively, when compared to rat TRα. xTRβ is over 96% and 94% conserved in those

domains compared to rat TRβ. The major difference between Xenopus TR subtypes is the lack

of an amino terminal domain in xTRβ that may participate in cell specific transcriptional

activation or altered DNA binding properties in xTRα. A recently developed synthetic thyroid

hormone analog, GC-1, preferentially binds and transactivates mammalian TRβ versus TRα (14).

In human TRs, a single amino acid difference in the ligand binding pocket is responsible for this

selectivity (15), and this amino acid difference is conserved in Xenopus TRs. In vivo, GC-1

selectively decreases plasma lipids and thyroid stimulating hormone secretion without increasing

heart rate in hypothyroid rats (16). The effects of GC-1 are consistent with the known relative

expression patterns of rat TR subtypes in liver, pituitary, and heart. Recently, we have

established that Xenopus laevis metamorphosis provides a simple screening assay for the

biological activity of thyroid hormone receptor antagonists(17). Since the Xenopus and

mammalian TRs are strongly conserved, and the tissue specific responses to TH are very

dramatic, we used GC-1 to probe TR subtype function in living Xenopus laevis tadpoles during

induced metamorphosis.

MATERIALS AND METHODS

Protease protection assays -35S-methionine labeled (Amersham Pharmacia) TRα and

TRβ proteins were synthesized using TnT SP6 Coupled Wheat Germ Extract in vitro

transcription and translation system (Promega). 2µl of translated 35S labeled TR was mixed with

the indicated concentrations of T3, GC-1, NH-3, or vehicle and incubated for 30 minutes at room

temperature. 2µg of type IV Elastase (Sigma) was mixed with the TR and hormone in 96-well

plates (Costar 9017) bringing the total volume to 20µl. After a 20 minute room temperature

incubation SDS-PAGE sample buffer was added followed by four 1 minute freeze-thaw cycles.

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

5

Samples were boiled for 10 minutes before SDS-PAGE analysis. The SDS-PAGE gel was fixed

with 25% isopropanol and 10% acetic acid, and dried for 2 hours. The gel was exposed to a

phosphoimager screen overnight and scanned with Storm 860 (Molecular Dynamics). The bands

were quantitated using ImageQuant software.

Transient transfection assays- DNA encoding the Xenopus TRα ligand-binding domain

(amino acids 132-418) or Xenopus TRβ ligand binding domain (amino acids 85-369) was

amplified by PCR using proof reading (Pfu) DNA polymerase from cDNA encoding Xenopus

laevis TRα or TRβ using specific primers flanked by BamHI sites (5' and 3’). These fragments

were subcloned in frame with the Gal4 DNA binding domain into pSG5 Gal4 DBD vector (a gift

from Marty Privalsky, University of California, Davis) between the BamHI and BglII sites of the

dephosphorylated vector. All clones were verified by DNA sequencing (UC Davis DNA

Sequencing Facility). All cell culture media and transfection reagents were obtained from

Gibco/BRL unless otherwise indicated. XLA kidney cells or XTC fibroblast cells were seeded in

6 well tissue culture plates at a density of 1X 105 cells per well in 70% Leibovitz L-15 medium

plus 10% fetal bovine serum and gentamycin. The cells were maintained at room temperature

without CO2 supplementation. XLA cells were transfected with 50 ng pSG5 Gal4/TRa or pSG5

Gal4/TRb, 100 ng MH100X4 (a luciferase reporter gene with four UAS or Gal4 binding sites,

kindly provided by David Mangelsdorf, UT Southwestern), and 350 ng pCS2-βgalactosidase per

well(18). XTC cells were transfected with 100 ng Xenopus laevis TR expression vector

(miwTRα, miwTRβ, or miwCAT as control (19)), 100 ng TH/bZIP TRE-luciferase reporter

vector (19) and 300 ng pCS2-βgalactosidase (18) . DNA for XLA or XTC cell transfections

were mixed with 2 µl Lipofectamine 2000 (Invitrogen) per well, and added to the cells for 24

hours. Cells were allowed to recover for 12-16 hours in the culture medium, then treated for 48

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6

hours with the indicated doses of T3, GC-1, or vehicle. Cells were harvested by scraping in

10mM TrisCl, pH 7.8, 150 mM NaCl, 1 mM EDTA and cell pellets were resuspended in 125 µl

reporter lysis buffer (Promega) followed by a single freeze thaw cycle. Cell extracts were

assayed for luciferase and βgalactosidase activity as described elsewhere (19).

Animal studies-Xenopus laevis tadpoles were purchased from Nasco, Inc. (Fort

Atkinson, WI) and staged according to Nieuwkoop and Faber (20). All chemicals were

purchased from Sigma unless indicated otherwise. For hormone treatments, a 1 mM stock of T3

was dissolved in 4 mM NaOH, and a 10 mM GC-1 stock was made in dimethylsulfoxide. GC-1

was synthesized as described elsewhere (14). For whole animal treatments, stage 52/53 tadpoles

were placed in 1X Steinberg’s solution with 1 mM methimazole and the indicated hormone

concentrations or appropriate vehicle and observed daily. After one week, animals were fixed in

phosphate buffered saline containing 10% formalin. Animals were photographed with a Nikon F-

5 camera and slides were scanned into Adobe Photoshop 5.0 using a Kodak RFS 3570 Plus Film

Scanner. Images were then assembled in Adobe Pagemaker 6.5.

Tail organ culture- Tadpole tails were cultured in vitro as described previously (21), with

some modifications. Staged tadpoles were pre-treated in 1X Steinberg's solution containing

gentamycin (70 µg/ml) and streptomycin (200 µg/ml) 24 hours before removal of the tail tips for

culture. The tadpoles were briefly anesthetized in water containing 0.01% MS222

(aminoesterbenzoic acid, Sigma) after which 6 mm of the tail tip was cut and placed in 1 ml of

Steinberg's solution with antibiotics and 5 µg/ml insulin in 12 well Falcon tissue culture dishes

(Becton-Dickinson). Tail tips were left in the culture media for 24 hours before the first hormone

treatment. T3 (Sigma) and GC-1 were diluted to the appropriate final concentration. The same

final dilution of the vehicle was added to control cultures. Afterwards, the culture media with

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

7

hormones were changed every 48 hours. The length of each tail was measured every 24 hours,

and each experimental point consisted of the mean change of 4 to 6 tails. Each experiment was

repeated twice.

RNA analysis- Tadpoles of the indicated stage were treated with T3 or GC-1 in 1X

Steinberg’s solution for 48 hours. Tails were harvested and frozen in liquid nitrogen. Total RNA

was isolated and analyzed by Northern hybridization using cDNA probes as described previously

(22). The amount of hybridized probe was determined using a Molecular Dynamics Storm

Phosphorimager and ImageQuant software and normalized to expression of the ribosomal

protein rpL8 gene (23).

RESULTS

GC-1 preferentially binds to Xenopus laevis TRβ over TRα- GC-1 is an iodine-free

thyromimetic that preferentially binds mammalian TRβ over TRα (14). In order to determine

both the relative binding affinity of GC-1 for the Xenopus TRs, protease protection assays were

performed. Protease protection assays have been used previously to both estimate binding

affinity and to reveal different conformations induced by various natural and synthetic nuclear

receptor ligands (24). In addition, the assay is performed at equilibrium and bound and free

ligands are not separated, so results tend to accurately reflect relative transcriptional activation

potency of ligands in question (25). Increasing amounts of T3, the most potent form of TH,

induce a major protected band of about 35 kDa in both xTRα and xTRβ (Fig. 2A). The binding

curves generated for xTRα and xTRβ were virtually identical (Fig. 2B). Using GC-1, a major

protected band of the same size was produced; however, GC-1 bound to xTRβ at a concentration

at least ten times lower than observed for xTRα (Fig. 2A and B). Our protease protection assay

revealed a preference of GC-1 for TRβ over TRα remarkably consistent with competitive ligand

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

8

binding assays used for mammalian TRs. Therefore, GC-1 is a TRβ selective ligand for Xenopus

TRs.

Activation of Xenopus laevis Thyroid Hormone Receptors by GC-1- We next tested the

ability of GC-1 to activate TRs in Xenopus laevis cells using a transient transfection assay.

Using Xenopus laevis XLA kidney cells, which express functional TRs, we examined the T3 and

GC-1 regulated transactivation properties of the TRα and TRβ ligand binding domains as Gal4

fusion proteins. The dose response curves for T3 activation of TRα and TRβ on a UAS-based

reporter gene were virtually identical, whereas GC-1 activated TRβ at lower concentrations than

TRα (Fig. 3A and 3B). The leftward shift in the GC-1 dose response curve was approximately

ten- to twenty-fold at 50% maximal induction, which essentially mirror the protease protection

assay results shown in Figure 2. We next tested the relative ability of GC-1 to activate full-

length Xenopus TRα and TRβ by transfecting expression vectors for each subtype individually

into the Xenopus XTC fibroblast cell line, which in our hands expresses a low level of functional

TRs relative to XLA cells. For example, in the absence of transfected TR, T3 or GC-1 only

activated a TRE based reporter gene by less than two-fold (see insets, Figure 3C and 3D). In

response to T3, both TRα and TRβ strongly activated reporter gene transcription. In this case,

the maximum level activation is different for TRα versus TRβ but the EC50 (about 10nM) does

not change, as has been previously reported for mammalian TR (26). When GC-1 is used, the

maximum activation is similar but the EC50 value (approximately 20 nM for TRβ vs. 200 nM

for TRα) is shifted in favor of TRα. (Fig. 3C and 3D). Therefore, in both ligand binding and

transactivation assays, GC-1 is a TRβ selective ligand for Xenopus as well as mammalian TRs.

Effect of GC-1 on Xenopus laevis tadpoles-An important advantage of metamorphosis as

an assay system is that thyroid hormone can be added directly to the rearing water, causing

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

9

accelerated effects that mimic what occurs during spontaneous metamorphosis. We treated

premetamorphic tadpoles (stages 52-53) with increasing doses of both T3 and GC-1. These

developmental stages are just prior to when the thyroid gland becomes active, but after the

animals are fully competent to respond to exogenous hormones. We observed the first clear

response to GC-1 at 50 nM, where obvious gill resorption and modest tail resorption are induced

(Fig. 4A.). As the dose of GC-1 is increased, progressively more gill and tail resorption occurs.

Very little change is observed in the limbs except for a loss of tissue around the forelimb buds

allowing them to extend away from the body. At high doses of GC-1 (> 200 nM), the animals

develop a beak like appearance due to the expansion of Meckel's cartilage. Finally, at 500 nM

GC-1, the treated tadpoles are considerably smaller than controls with an almost complete loss of

the tail and gills. Internally, we noted extensive resorption of the intestine beginning above 100

nM GC-1, similar to what is observed with T3 treatments above 10 nM. Like T3, the effects of

GC-1 on intact tadpoles are first apparent about two days after treatment begins, and most

animals survive the one-week treatments even at the highest doses of GC-1. In some batches of

animals, we observed limited limb growth and some widening of the brain at concentrations of

GC-1 greater than 200 nM, but these occurred infrequently and at higher concentration than

those required for tail and gill resorption that were always observed.

Our results with GC-1 contrast significantly with the results obtained with the

endogenous hormone T3. Low doses of T3 (2 nM) induced hind- and forelimbs growth and

differentiation without significant loss of tail tissue (Fig. 4A). In dorsal views (Fig. 4B),

expansion of the brain is readily observed at low T3 doses; we observe a consistent 30 to 40 %

expansion of the brain in response to 2 nM T3. Higher T3 concentrations (10 nM) induced limb

growth to a greater extent, and gill loss becomes clearly noticeable. Both growth of adult

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

10

structures such as the limbs and resorption of larval tissues such as the tail and gills are all

apparent at the highest doses of T3 used in these studies (50 nM). A "hunchbacked" appearance

is also apparent at these higher T3 doses that is not seen in GC-1 treated animals. This

presumably results from the accelerated development of adult dorsal muscle [Nishikawa, 1994

#480]. Toxicity was observed when T3 was used in excess of 50 nM, with many animals dying

within a few days after treatment. Thus, when GC-1 and T3 dose responses are compared, GC-1

induces death and resorption of larval tissues at lower concentrations than are required to induce

limb growth. This is completely opposite to the developmental program induced by the natural

hormone T3 where growth responses generally precede larval tissue death and resorption.

The sensitivity of tadpole tails to GC-1 increases with increasing TRβ expression-Since

higher levels of GC-1 than T3 were required to cause tissue responses in premetamorphic

tadpoles, we next tested whether the efficiency of GC-1 in inducing larval cell death would

improve in later developmental stages when TRβ levels are significantly elevated. TRβ protein

expression increases eight- to ten-fold as tadpoles develop from premetamorphosis (prior to

thyroid gland maturation) to metamorphic climax when tail resorption commences in earnest

(27). On the other hand, despite a modest increase in TRα message, TRα protein levels remain

constant throughout metamorphosis. Therefore we predicted a modest increase in sensitivity to

T3 but a large increase in sensitivity to the TRβ selective ligand GC-1 at two developmental

stages: premetamorphic stages 52-54 with low TRβ levels and prometamorphic stages 57-58

when TRβ levels were nearly maximal but tail resorption had not yet begun.

We first established that isolated tail tips from premetamorphic tadpoles would resorb in a T3

dose responsive and reproducible manner (Fig. 5A). Maximal resorption rates occurred at

concentrations above 32 nM T3. We then measured the efficiency of tail resorption in response

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

11

to increasing T3 and GC-1 at stages 53/54 and 57/58. Tails were measured after six days

treatment, when the control tails were still healthy and there were clear differences in the

response to various doses of hormones (Fig. 5B). Stage 57-58 tails were modestly more sensitive

to T3 (about two fold at 50% resorption) than stage 52-54 tails. Strikingly, the stage 57-58 tails

were over twelve-fold more sensitive to GC-1 than the stage 52-54 tails when the doses required

to induce 50% resorption are compared. This increased sensitivity to GC-1 is especially

apparent at low doses of GC-1 (50 nM) where there is no response in young tadpoles but a robust

response is observed in older tadpoles.

Induction of TH response genes by T3 and GC-1-A large collection of TH response genes

has been cloned from Xenopus laevis tadpole tails (28), as well as from several other tissues (29-

31). Tail TH response genes include rapidly induced early genes that generally encode

transcription factors and delayed genes that mostly encode membrane bound, secreted, or

intracellular proteases and extracellular matrix components and receptors (28,32). Delayed gene

expression is more tissue specific than early gene expression and is especially strong in

fibroblasts under the epidermis and between muscle fibers of the tail and in the gills (13). We

have analyzed the induction of tail TH response genes by increasing doses of T3 and GC-1,

again using stages of development with different levels of TRβ expression.

xBTEB (33) is a zinc finger transcription factor whose expression is typical of early gene

responses to both T3 and GC-1 (Fig. 6 A and B). xBTEB is quite sensitive to T3, since it is

induced by as low as 1 nM T3 in stage 54 tadpoles. The induction of xBTEB by T3 at stage 57

is similar. GC-1 induces xBTEB above 100nM in stage 54 tadpoles; however, at stage 57/58,

GC-1 is unable to induce xBTEB expression above baseline. Gene 12, TH/bZIP, and TRβ itself

are other early genes that show a similar dose response profile to T3 and GC-1 as xBTEB (data

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

12

not shown). We next looked at the delayed class of tail response genes, exemplified by the

membrane bound protease fibroblast activation protein α (FAPα) (32) (Fig. 6C and D). First, a

five-to -ten fold higher level of T3 is required to activate delayed genes like FAPα than early

genes like xBTEB. In response to GC-1, FAPα mRNA is first induced above baseline at 100 nM,

with a similar dose response curve to xBTEB. Strikingly, FAPα becomes more, not less,

sensitive to GC-1 in older animals that are expressing more TRβ. A clear induction of FAPα is

seen in response to as low as 20 nM GC-1 in stage 57-58 animals. Other genes in this delayed

class that show the same pattern of responses are the secreted matrix metalloproteinase

collagenase-3 and the intracellular protease pepE (data not shown). We also noted that the dose

response curves for delayed gene expression closely parallels both the T3 and GC-1 dose

response curves in tail culture assays, further implicating these genes as key players in the

process of tail resorption.

DISCUSSION

A recently developed TRβ selective agonist, GC-1, preferentially induces a subset of TH

regulated developmental programs in Xenopus laevis tadpoles, namely the death and resorption

of larval tissues. Thus, we provide compelling evidence for different functional roles of the two

TR subtypes during metamorphosis, extending earlier expression studies that were strictly

correlative (8,13,28). Our results agree well with known TR isotype gene expression in that GC-

1 poorly induces TRα rich tissues but is most effective in tissues that have low TRα levels but

strongly induce TRβ. TRβ expression in the tail is particularly high in fibroblasts under the skin

and around the notochord that also strongly express delayed tail response genes like FAPα. The

pronounced effects of GC-1 on gill and tail resorption are intriguing in light of recent

experiments on the expression of TRα and TRβ in various salamander species. Obligatorily

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

13

neotenic species like Necturus are unable to complete metamorphosis such that the external gills

and tail fins of these animals are resistant to TH even though they have fully developed limbs.

Interestingly, they express TRα in those tissues but there is no detectable baseline or inducible

TRβ message, even though an apparently normal TRβ gene exists in the genome (34).

Most tail response genes are induced by similarly high doses of GC-1 in premetamorphic

animals (>100 nM) but show very different responses at later developmental stages when TRβ is

the predominant receptor isotype. Based on these results, we propose that there are two classes

of genes in terms of TR subtype control. In young animals, sufficient TRβ must be induced via

TRα to activate delayed genes, but other early genes like xBTEB and gene 12 are also being

induced in parallel. At later stages, only the delayed genes become more sensitive to both T3 and

GC-1 indicating that this gene class, at least in the tail, is preferentially controlled by TRβ. Our

tail culture results are also most consistent with at least a partial requirement for activation for

xTRα to induce sufficient xTRβ levels for a response in larval tissues like the tail. Indeed, the

xTRβ gene contains a strong thyroid hormone response element near its transcription start site

(35,36), and TRβ can be induced in tadpoles even in the presence of protein synthesis inhibitors

(10). However, there are other possibilities to account for the differences in tissue sensitivity to

T3 and GC-1. There may be less efficient uptake of GC-1 than T3 from the rearing media in

young tadpoles or stage differences in the metabolism of the ligands. In rats, GC-1 accumulates

in the liver to a higher degree than in the heart, potentiating its TR subtype selective properties

(16). In this study, we deliberately focused on limb versus tail responses rather than internal

organs due to equal access of each tissue to the media.

Although our results are consistent with a dominant role for TRβ in larval tissue death

and resorption, we cannot completely rule out a role for TRβ in formation of adult structures

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

14

during metamorphosis. For instance, TRβ activation may be important in the differentiation of

adult cells after several rounds of TRα mediated proliferation have occurred. TRβ mRNA is

detectable in certain remodeling tissues undergoing both resorption and growth such as head

cartilage and the intestine (8,12). In addition, the effects of GC-1 on premetamorphic tadpoles

are perhaps more dramatic than expected based on the dose response profiles of transient

transfection assays alone. It is possible that inappropriately timed TRβ activation may antagonize

TRα function. The estrogen inducible progesterone receptor antagonizes estrogen receptor

function in the mammalian uterus (37). Another interpretation of our results is that the effects of

GC-1 on premetatmorphic tadpoles may simply reflect where each isotype is most abundant as

the tissue responds. In this model, the two TR isotypes differ quantitatively but not qualitatively.

In order to test this possibility, transgenic approaches are now available in Xenopus where it

should be possible to express xTRβ in xTRα dominated tissues and vice versa depending on the

availability of appropriate tissue-specific promoters.

There is growing evidence in other animals for isotype specific functions of nuclear

receptors. In insects, ecdysone receptor (EcR) diversity is generated by alternate promoter usage

of the same gene to create different amino terminal transactivation domains (38). Three EcR

subtypes in Drosophila melanogaster are expressed in different ratios in different tissues, which

correlate with different tissue fates at metamorphosis (38). Specific EcR-B1 mutations prevent

salivary gland histolysis and development of a subset of imaginal disks which cannot be rescued

by overexpression of the EcR-A subtype (39). In vertebrates, knockout studies of each TR gene

alone and in combination in mice have demonstrated both overlapping and distinct functions for

the two TR genes (40). For instance, proper functional development of the inner ear appears to

require TRβ (41) and TRα appears to play a more dominant role in TH control of resting heart

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

15

rate and body temperature (42,43). However, other processes such as negative feedback control

of the thyroid stimulating hormone gene were at least partially affected in both TR knockouts

(40). It is possible that TR subtype specific function is more distinct in taxa other than mammals.

Certainly, the study of receptor signaling during development is easier in model systems with

externally developing embryos. Our findings illustrate the utility of subtype-selective ligands as

probes of nuclear receptor function in vivo, and demonstrate the potential of natural systems such

as amphibian metamorphosis as simple bioassays for the development of selective agonists and

antagonists of thyroid hormone receptors.

Acknowledgements-The authors would like to thank Jock Hamilton for photography and

the Privalsky laboratory for providing CV-1 cells and for advice on their transient transfection.

We also thank Phuoc Le, Sherman Lau, and Yvette Chu for excellent technical assistance.

REFERENCES 1. Dodd, M. H. I., and Dodd, J. M. (1976) in Physiology of the Amphibia (Lofts, B., ed), pp.

467–599, Academic Press, New York 2. Tata, J. R. (1998) Cell Res 8, 259-272 3. Brown, D. D., Wang, Z., Kanamori, A., Eliceiri, B., Furlow, J. D., and Schwartzman, R.

(1995) Recent Prog Horm Res 50, 309-315 4. Yaoita, Y., Shi, Y. B., and Brown, D. D. (1990) Proc Natl Acad Sci U S A 87, 7090-7094 5. Zhang, J., and Lazar, M. A. (2000) Annu Rev Physiol 62, 439-466 6. Yaoita, Y., and Brown, D. D. (1990) Genes Dev 4, 1917-1924 7. Banker, D. E., Bigler, J., and Eisenman, R. N. (1991) Mol Cell Biol 11, 5079-5089 8. Berry, D. L., Rose, C. S., Remo, B. F., and Brown, D. D. (1998) Dev Biol 203, 24-35 9. Kawahara, A., Baker, B. S., and Tata, J. R. (1991) Development 112, 933-943 10. Kanamori, A., and Brown, D. D. (1992) J Biol Chem 267, 739-745 11. Fairclough, L., and Tata, J. R. (1997) Dev Growth Differ 39, 273-283 12. Shi, Y. B., Su, Y., Li, Q., and Damjanovski, S. (1998) Int J Dev Biol 42, 107-116 13. Berry, D. L., Schwartzman, R. A., and Brown, D. D. (1998) Dev Biol 203, 12-23 14. Chiellini, G., Apriletti, J. W., Yoshihara, H. A., Baxter, J. D., Ribeiro, R. C., and Scanlan,

T. S. (1998) Chem Biol 5, 299-306 15. Wagner, R. L., Huber, B. R., Shiau, A. K., Kelly, A., Cunha Lima, S. T., Scanlan, T. S.,

Apriletti, J. W., Baxter, J. D., West, B. L., and Fletterick, R. J. (2001) Mol Endocrinol 15, 398-410

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

16

16. Trost, S. U., Swanson, E., Gloss, B., Wang-Iverson, D. B., Zhang, H., Volodarsky, T., Grover, G. J., Baxter, J. D., Chiellini, G., Scanlan, T. S., and Dillmann, W. H. (2000) Endocrinology 141, 3057-3064

17. Lim, W., Nguyen, N. H., Yang, H. Y., Scanlan, T. S., and Furlow, J. D. (2002) J Biol Chem

18. Turner, D. L., and Weintraub, H. (1994) Genes Dev 8, 1434-1447 19. Furlow, J. D., and Brown, D. D. (1999) Mol Endocrinol 13, 2076-2089 20. Nieuwkoop, P. D., and Faber, J. (1994) Normal Tables of Xenopus laevis (Daudin),

Garland Publishing, New York 21. Tata, J. R., Kawahara, A., and Baker, B. S. (1991) Dev Biol 146, 72-80 22. Furlow, J. D., Berry, D. L., Wang, Z., and Brown, D. D. (1997) Dev Biol 182, 284-298 23. Shi, Y. B., and Liang, V. C. (1994) Biochim Biophys Acta 1217, 227-228 24. Yoh, S. M., and Privalsky, M. L. (2002) Methods Mol Biol 202, 129-152 25. Blumberg, B., Bolado, J., Jr., Derguini, F., Craig, A. G., Moreno, T. A., Chakravarti, D.,

Heyman, R. A., Buck, J., and Evans, R. M. (1996) Proc Natl Acad Sci U S A 93, 4873-4878

26. Hollenberg, A. N., Monden, T., and Wondisford, F. E. (1995) J Biol Chem 270, 14274-14280

27. Eliceiri, B. P., and Brown, D. D. (1994) J Biol Chem 269, 24459-24465 28. Wang, Z., and Brown, D. D. (1993) J Biol Chem 268, 16270-16278 29. Shi, Y. B., and Brown, D. D. (1993) J Biol Chem 268, 20312-20317 30. Buckbinder, L., and Brown, D. D. (1992) J Biol Chem 267, 25786-25791 31. Denver, R. J., Pavgi, S., and Shi, Y. B. (1997) J Biol Chem 272, 8179-8188 32. Brown, D. D., Wang, Z., Furlow, J. D., Kanamori, A., Schwartzman, R. A., Remo, B. F.,

and Pinder, A. (1996) Proc Natl Acad Sci U S A 93, 1924-1929 33. Furlow, J. D., and Kanamori, A. (2002) Endocrinology 143, 3295-3305 34. Safi, R., Begue, A., Hanni, C., Stehelin, D., Tata, J. R., and Laudet, V. (1997) J Mol Evol

44, 595-604 35. Ranjan, M., Wong, J., and Shi, Y. B. (1994) J Biol Chem 269, 24699-24705 36. Machuca, I., Esslemont, G., Fairclough, L., and Tata, J. R. (1995) Mol Endocrinol 9, 96-

107 37. Tong, W., and Pollard, J. W. (1999) Mol Cell Biol 19, 2251-2264 38. Talbot, W. S., Swyryd, E. A., and Hogness, D. S. (1993) Cell 73, 1323-1337 39. Bender, M., Imam, F. B., Talbot, W. S., Ganetzky, B., and Hogness, D. S. (1997) Cell 91,

777-788 40. Forrest, D., and Vennstrom, B. (2000) Thyroid 10, 41-52 41. Forrest, D., Hanebuth, E., Smeyne, R. J., Everds, N., Stewart, C. L., Wehner, J. M., and

Curran, T. (1996) Embo J 15, 3006-3015 42. Johansson, C., Gothe, S., Forrest, D., Vennstrom, B., and Thoren, P. (1999) Am J Physiol

276, H2006-2012 43. Wikstrom, L., Johansson, C., Salto, C., Barlow, C., Campos Barros, A., Baas, F., Forrest,

D., Thoren, P., and Vennstrom, B. (1998) Embo J 17, 455-461

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

17

FIGURE LEGENDS

FIG. 1: Structures of T3 and GC-1. Chemical structures of 3, 5, 3’-triidothyronine, or T3, and

the isotype selective agonist GC-1.

FIG. 2: GC-1 binds preferentially to Xenopus laevis thyroid hormone receptor β. A. In vitro

translated, 35S-labeled xTRα or xTRβ were incubated with the indicated nM concentrations of T3

or GC-1, then treated with 4 µg of elastase. The resulting bands were resolved by 12% SDS-

PAGE. I, input. B. T3 binding curves for xTRα (circles) and xTRβ (squares) were generated by

quantitation of the protected bands shown in A, and expressed as a percentage of the maximally

protected bands. C. GC-1 binding curves for xTRα (circles) and xTRβ (squares) were generated

by quantitation of the protected bands shown in A, and expressed as a percentage of the

maximally protected bands. Error bars in B and C represent standard deviation.

FIG. 3. GC-1 preferentially activates Xenopus laevis TRβ versus TRα relative in transient

transfection assays. A and B. A UAS-containing promoter driving luciferase reporter gene

expression was co-transfected into Xenopus laevis kidney epithelial cells (XLA) along with

expression vectors for Xenopus laevis TRα/Gal4 (circles) or TRβ/Gal4 (squares) fusion

proteins, and the cells were treated with increasing concentrations of T3 (A) or GC-1 (B).

Luciferase activity was normalized to constitutive β-galactosidase activity. C and D. A TRE-

containing minimal promoter driving luciferase reporter gene expression was co-transfected in

Xenopus laevis XTC cells with expression vectors for full-length Xenopus TRα (circles) or TRβ

(squares), and cells were treated with in creasing concentrations of T3 (C) or GC-1 (D).

Induction of the TRE-based reporter gene by T3 or GC-1 without co-transfection of either TR

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

18

expression vector is shown in the insets of C and D. Luciferase activity was normalized to

constitutive β-galactosidase activity. Each data point was determined in triplicate, and error bars

represent SD.

FIG. 4. GC-1 strongly induces the loss of larval tissues in premetamorphic (stage 52/53)

tadpoles, whereas T3 induces both the death of larval tissues and growth of adult tissues

like the hind and forelimbs. A, top panels: Tadpoles treated for one week with the indicated

GC-1 concentrations, ventral views. G, gills; HL, Hindlimbs. Bottom panels, Tadpoles treated

for one week with the indicated T3 concentrations, also ventral views. B. Same treated animals

as in A, dorsal views. Top panels, GC-1 treated animals. Br, brain; G, Gills. Bottom panels, T3

treated animals.

FIG. 5. The efficiency of GC-1 induced tail resorption in vitro correlates with TRβ but not

TRα levels. A. Tail tips from stages 53/54 cultured in vitro show a dose dependent resorption in

response the indicated concentration of T3 compared to untreated tail tips. The extent of tail

resorption is expressed as the % of tail length remaining in hormone treated cultures compared to

cultures treated with vehicle alone. B. Tails were obtained from stage 53/54 tadpoles (open

symbols) or stage 57-58 tadpoles (filled symbols). Squares, T3; Circles, GC-1. Error is

expressed as SEM.

FIG. 6. Northern blot analysis of thyroid hormone response genes induced by T3 and GC-

1. Two thyroid hormone response genes, BTEB (A and B) and FAPα (C and D), were assayed

for induction by increasing doses of T3 (A and C, squares) or GC-1 (B and D, circles). RNA

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

19

was analyzed from tails of tadpoles at stages 53/54 (open symbols) or stages 57/58 (closed

symbols). Northern data was normalized by stripping the blots and reprobing for constitutive

rpL8 expression, and shown as a % of the maximum expression for each set of dose response

curves on a given blot.

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

HO

I

OCO2H

NH2

H

I

I

T3

HO

H3C

CH3

O CO2H

GC-1

Figure 1

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

0

20

40

60

80

100

120

-11 -10 -9 -8 -7 -6

Log [T3]

% M

ax

. B

ind

ing

B.

0

0

20

40

60

80

100

120

-10 -9 -8 -7

Log [GC-1]

% M

ax. B

ind

ing

C.

0

I 0 0.1 1 10 100 1000 nM T3

I 0 1 2 5 10 20 50 100 200 nM GC-1

TRα

TRβ

TRα

TRβ

A.

Figure 2 by guest on A

pril 9, 2018http://w

ww

.jbc.org/D

ownloaded from

0

2000

4000

6000

8000

-10 -9 -8 -7 -6

0

2000

4000

6000

8000

-10 -9 -8 -7 -6

A.

B.

Figure 3

[Log T3, M]

[Log GC-1, M]

Rel

ativ

e Lu

c A

ctiv

ityR

elat

ive

Luc

Act

ivity

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

0

40

80

120

160

-11 -10 -9 -8 -7 -6

0

40

80

120

160

0 nM 100 nM

No TR

0

100

200

300

400

500

0 nM 100 nM

No TR

C.

D.

[Log T3, M]

[Log GC-1, M]

Figure 3, cont.R

elat

ive

Luc

Act

ivity

Rel

ativ

e Lu

c A

ctiv

ity

0

100

200

300

400

500

-10 -9 -8 -7 -6 -5

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

nM GC-1: 0 50 100 200 500

nM T3: 0 2 10 50

A.G

HL

B.

nM T3: 0 2 10

nM GC-1: 0 500

G Br

Tail

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8

2

nM T3

Days after first T3 treatment

% T

ail R

emai

ning

4

8

16

3264

128

A.

0

25

50

75

100

% T

ailr

eman

ing

rela

tive

to c

ontr

ols

0 -8 -7 -6

log [Hormone]

-8.7

B.

Figure 5

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

0

25

50

75

100

Perc

ent m

ax. i

nduc

tion

-8

-7.5 -7

-6.5 -6

Log [GC-1]

FAPα0

25

50

75

100

Perc

ent m

ax. i

nduc

tion

-9

-8.5 -8

-7.5 -7

-6.5

Log [T3]

FAPα

0 00

25

50

75

100

Perc

ent m

ax. i

nduc

tion

-9 -8.5 -8 -7.5 -7 -6.5

Log [T3]

0

0

25

50

75

100

Perc

ent m

ax. i

nduc

tion

-8

-7.5 -7

-6.5 -6

Log [GC-1]

0BTEB BTEB

A. B.

C. D.

Figure 6

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

and Thomas S. ScanlanJ. David Furlow, Hayung Yang, Mei Hsu, Wayland Lim, Davy J. Ermio, Grazia Chiellini

hormone receptor agonist GC-1Induction of larval tissue resorption in Xenopus laevis tadpoles by the thyroid

published online March 30, 2004J. Biol. Chem. 

  10.1074/jbc.M402847200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from