| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Thyroid Division (B.G., J.W.H., H.M.T., P.R.L.) Department of
Medicine Brigham and Women’s Hospital Harvard Medical
School Boston, Massachusetts 02115
Dipartimento di
Endocrinologia ed Oncologia Molecolare e Clinica (D.S.)
Università degli Studi di Napoli "Federico II" Napoli,
Italy 80131
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The transcription factors governing thyroid-specific protein expression include thyroid transcription factors 1 and 2 (TTF-1, TTF-2) and Pax-8. TTF-1 (also called T/ebp or Nkx-2.1) cDNA has been cloned and encodes a homeodomain (HD)-containing protein (10). Its expression is restricted to the thyroid, lung, and certain regions of the fetal central nervous system (11). The expression of thyroglobulin (Tg) and thyroperoxidase (TPO) are stimulated by TTF-1 and Pax-8 (12). TTF-1 also affects the expression of the TSH receptor and the lung-specific surfactant B protein (12, 13). An important role for TTF-1 in organogenesis was shown in the TTF-1 knock-out mouse (14). The homozygous animals are born dead, without a thyroid or pituitary gland and with atrophic lung parenchyma. Defects in the ventral forebrain are also present. An infant with heterozygous deletion of the TTF-1 gene was found to have thyroid dysfunction, respiratory failure, and delayed mental and motor development, indicating it has similar importance in humans (15).
On the other hand, the Pax-8 gene encodes a paired domain-containing protein that is expressed in the thyroid, kidney, and pituitary (16, 17). It binds the "C" site in the 5'-flanking region (FR) of the Tg gene with TTF-1 in a mutually exclusive fashion (17). Pax-8 is more important in the transcriptional activation of TPO expression than of the Tg promoter (12, 18). The core binding site contains a TGCCC motif, but the binding is also affected by the adjacent 3'-region that contains an A(G/C)TC sequence in the rat Tg and TPO promoter. However, mutational studies suggest that important determinants of Pax-8 DNA binding are located outside of this consensus sequence (17).
TTF-2 has been only recently cloned (19). It is a forkhead-containing protein involved in thyroid- specific gene expression and necessary for thyroid morphogenesis. Its specific function, as either a positive or a negative regulator of thyroid-gene expression, is still under investigation (12, 19, 20).
There is no information as to the potential role of any of these transcription factors in the expression of the deiodinase genes. The present studies were performed to explore the molecular basis for thyroidal deiodinase expression in human and rat thyroid. We addressed the role of the two established transcriptional activators of thyroid-specific gene transcription, TTF-1 and Pax-8. Also, since prior studies have shown that D2 activity in rat is increased by cAMP, we also defined the molecular basis for the cAMP responsiveness of rdio2 (21, 22).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). In contrast, the hD2 mRNA is easily
detected after approximately 16 h in human thyroid under similar
conditions (7). D1 mRNA was abundant in the rat thyroid, requiring only
48 h exposure of the same blot (Fig. 1). The hypothyroid rat
thyroid expressed lower levels of D2 mRNA than did that of normal
animals (Fig. 1). The D2/ß-actin band density ratios from euthyroid
and hypothyroid rat thyroids were 0.8 and 0.1, respectively, while that
for D1 was 9.6 and 3.2.
|
2 nM and 2
µM, respectively). We have shown that the release of
125I- from 2 nM
125I-T4 by D2 can be
completely blocked by 100 nM T4 while
that from D1 catalyzed 5'-T4 deiodination is
unaffected (7). Such an approach is required because 6-n
propylthiouracil (PTU) cannot completely inhibit D1 activity when
substrate concentration is low (7). Using this approach, we found no
difference in the 125I released at 2 and at 100
nM T4 in the presence of rat thyroid
sonicate, indicating the absence of low Km 5'- T4
deiodinase activity in the adult rat thyroid. This is in contrast to
the 41% inhibition of 125I release by 100
nM T4 using human thyroid homogenate
(Table 1).
125Iodide release from 100 nM
[125I]T4 is catalyzed by
D1, as shown by its suppression in the presence of 10 µM
rT3 in the human and rat thyroid samples. A
direct D1 activity assay using 1 µM
rT3 as substrate confirmed that D1 activity was
much higher in rat thyroid than in this sample of human thyroid (Table 1). Thus, there is little D2 mRNA and no detectable D2 activity in rat
thyroid.
|
). One motif in poor context, GCAAGA
(C2), was also included because it is located only 6 bp 3' to the C1
site, and we did wish to overlook the possibility of cooperative
effects. The hdio2 5'-FR fragment contains numerous TGCCC
motifs, which are the core of the Pax-8 binding site (17). However,
none of these had the preferred A(G/C)TC sequence in the adjacent
3'-region, such as are found in the Pax-8 binding site of the rat Tg
and TPO promoter (17).
|
). Surprisingly, this response was not
significantly reduced by deletion of approximately 5.9 kb of the 5'-FR,
indicating that sites A, B, E, F, and G were not functional. However,
deletion of either the C1 and C2 (hdio2–633
C) or the D binding
sites (hdio2–633D) reduced but did not abolish the
hdio2CAT response (Fig. 3). The responsiveness to TTF-1 was
reduced to only 3-fold by eliminating all three sites
(hdio2–633CD), suggesting that the functionally critical TTF-1
binding sites were the C and D motifs. The most 3' 83 bp of the
hdio2 promoter (hdio2–83) showed an approximately 1.8-fold
TTF-1 response (Fig. 3). With respect to the specificity of the
promoter response to TTF-1 under these conditions, none of the
following promoter CAT constructs were induced more than about
1.7-fold: the 3.7-kb hdio1 5'-FR (24), the
7.5-kb hdio3 gene, the 109-bp thymidine kinase promoter,
or the empty pOCAT2 vector (25). There was a 2.4-fold induction of a
137-bp rGH promoter CAT construct (26), which is of unknown
significance.
|
). In
contrast to hdio2, the 3.8-kb rdio2 CAT showed
only a 3-fold response to coexpression of TTF-1.
Comparison of the Human and Rat dio2 5'-FR
Sequences
Since the TTF-1 response of hdio2 was localized to the
most 3' 633 bp of the 5'-FR, we compared these sequences to the
corresponding region of rdio2 (21). These were 73%
identical, including complete conservation of the cAMP response element
(CRE) at about -90 (Fig. 4). However,
the C1 and C2 TTF-1 binding sites are absent in the rat sequence, and
the D site is eliminated by a G-to-A exchange in the CAAG core of the
motif (Fig. 4). The most 3' 120-bp region is virtually identical to
that of rdio2 and contains the CRE, the TATA box, and the
most 5'-hdio2 TSS.
|
. The binding affinity of a given
hdio2 site was estimated by comparing its capacity to
inhibit binding of labeled rat Tg oligo "C" to the in
vitro expressed rat TTF-1 HD protein (Fig. 6). The densitometric analysis of that
experiment is depicted graphically in Fig. 7. Only two of the four consensus
sequences, C1/C2 and D, were effective inhibitors of the binding of the
Tg site "C" sequence to the rat TTF-1 HD. The D site was the most
effective inhibitor, approximately equivalent to the rat Tg oligo
"C." Mutation of the two adenine nucleotides in the core of the D
site abolished its potency (Figs. 6 and 7). Mutation of the C1 site
reduced the competitive potency of the C1, C2 oligo to that of the
negative control while mutation of the C2 site had no effect. However,
the C1/C2 fragment was approximately 10-fold less potent than the
positive control or the D oligo (Fig. 7). To confirm the validity of
the results of these competition studies, we demonstrated that the
binding of the labeled D site to the rat TTF-1 HD is saturable (Fig. 8).
|
|
|
|
). A potential explanation for this
paradoxical result was that the BglII deletion moved the C1
TTF-1 binding site to a more proximal position in the promoter, thereby
increasing its effectiveness (see Fig. 4). Therefore, we introduced the
specific C1 and D site mutations already shown to eliminate the TTF-1
binding by EMSA into hdio2–633 CAT and transfected these
constructs into a human glioblastoma cell line U87. This line was
selected since, in transient transfection analyses with TTF-1, we found
it to have reduced nonspecific TTF-1-stimulated activity when compared
with COS cells. Furthermore, D2 is normally expressed in astroglial
cells (27) although it is not present in the U87 cells (data not
shown). A 10-fold response of the 633-bp hdio2 promoter
occurred with 60 ng cotransfected TTF-1 plasmid but the
rdio2 5'-FR construct was again only 3-fold stimulated.
Surprisingly, given the results of the EMSA studies, mutations of C1 or
D were equally effective in reducing the transcriptional response of
the hdio2–633, and mutation in both were required to reduce
the response to TTF-1 to that of the 3.8-kb rat dio2 (Fig. 9). Qualitatively similar results were
obtained in COS-7 cells (data not shown).
|
) (21). As expected, the response of rat dio2
to the catalytic subunit of protein kinase A (PKA) was abolished only
when the region containing this CRE is removed (rdio2#4). This confirms
that rdio2 responds to cAMP via a single canonical CRE (Fig. 10).
|
3.8 kb rdio2#1 CAT construct, and a
promoter reporter plasmid containing the 3.7 kb long
hdio1 5'-FR were also transfected into COS-7 to evaluate a
functional response along with 0.5 µg rat Pax-8-expressing plasmid
(17, 24). None of these were induced more than approximately 2.3-fold
while a 5- to 10-fold induction was found with the positive control,
CP5-CAT (28).
The hdio2 5'-FR contains a consensus AP-1 site at -524 to
-518 but its function has not been evaluated (Fig. 4). This site was
mutated by introducing an AGACCTC replacement
for TGACTCA in the 633-bp hdio2 5'-FR (hdio2-AP1Mut). The
mutant and the wild-type hdio2–633 constructs were transfected into
COS-7 cells in three separate experiments along with an hGH plasmid to
monitor transfection efficiency. Inactivation of the AP-1 site
increased the hdio2 basal promoter activity approximately
2-fold, suggesting that the hdio2 promoter is down-regulated
via the AP-1 site (data not shown). There was no response of either the
wild-type hdio2 or the AP-1 mutant to phorbol ester
[12-O-tetradecanoylphorbol 13-acetate (TPA)].
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We found extremely low concentrations of dio2 mRNA in rat
thyroid, and D2 activity was undetectable consistent with earlier
results in FRTL-5 cells (7, 9). The absence of significant D2 activity
differs from results reported by Bates et al. (3–4
pmol/h/mg protein) (29). However, these authors found only 14%
inhibition of 1 nM
[125I]rT3 deiodination by
100 nM unlabeled T4 in rat
thyroid homogenate. This is similar to our results (Table 1) and
indicates that rT3 deiodination by D2 is minimal
in rat as opposed to human thyroid. The methodological description
provided is not sufficiently detailed to resolve this apparent internal
discrepancy (29).
A functional 5'-FR-CAT transient expression assay showed that the
hdio2 gene was highly responsive (10 fold) to coexpressed
rat TTF-1 HD. We isolated and subcloned a 3.8-kb rat dio2
5'-FR/promoter clone to compare with the human gene. Consistent with
our hypothesis, the rat dio2 gene responded only
approximately 3- fold to coexpressed TTF-1. Neither human nor rat
dio2 was induced by Pax-8, and both genes contain one
functional CRE about 90 bp 5' to the transcriptional start site (Fig. 4
and Ref. 21). These results suggest that the lack of response to TTF-1
could explain the much higher expression of D2 mRNA in human than in
rat thyroid, although the alternative possibility of a rat
thyroid-specific transcriptional repressor cannot be completely ruled
out. The human dio1 gene is unresponsive to Pax-8 or
TTF-1.
A number of studies have evaluated the DNA sequence requirements for
TTF-1 binding. Detailed in vitro analysis of binding of
randomly generated Tg "C" site sequences to the rat TTF-1 HD
protein showed that the binding site consensus was CCCAGTCAAGTGTTCTT
(23). The consensus derived from analysis of the TTF-1 binding sites of
the rat, human, bovine, and dog Tg and TPO promoters (ACTCAAGTNNNN)
(12) or only from the rat Tg promoter (...
NNNAC(G)TCAAGTA(G)NNNN ... ) (30) were more flexible. These studies
indicate that a CAAG core is usually present but not sufficient for
efficient TTF-1 HD binding. Seven TCAAG(G/T) motifs were found in the
3' 4 kb hdio2 5'-FR. However, the functional studies of
the hdio2 gene indicated that only the most proximal two
sites, C1 (620 bp 5' to the major TSS) and D (230 bp 5' to this
site), were required for a complete response to TTF-1. In comparison,
footprinting and sequence homology search-based approaches identified
three TTF-1 binding sites in the rat and human Tg promoter, three in
the rat and five in the human TPO gene, and one site in the rat TSH
receptor promoter (12).
The order of magnitude of the response of hdio2 to TTF-1
(10 fold) in both COS-7 and U87 cells is similar to that conferred
by the rat Tg "A" and "C" TTF-1 binding sites in FRTL-5 cells
(30). This transcriptional response is quite large compared with the
less than 2-fold response of the sodium iodine symporter promoter in
FRT cells (31) or to the approximately 2.5-fold response of the TSH
receptor in COS-7 cells (32).
In vitro determination of the TTF-1 binding affinity of the
hdio2 sites by EMSA showed that the C1 and D
hdio2 TTF-1 binding sites that were important in the
functional studies showed the highest HD binding affinity. However,
despite the fact that the D site binding affinity was approximately
10-fold higher than that of the C1, C1 was as active as the D site
(Figs. 6, 7, and 9) even though it is about 390 bp farther away from
the TSS.
TTF-1 binds its recognition sequence via a HD, that is highly similar to the Drosophila NK2 HD (33). The rat TTF-1 HD can reproduce the binding of the entire protein (10). It is well known that the DNA binding specificity of HDs is promiscuous since the same HD can recognize different sequences without a clear consensus (34, 35, 36). For example, the functionally important TTF-1 binding sites in the bovine Tg gene upstream enhancer element do not contain a CAAG motif (37). However, the TTF-1 binding motifs in most of the known functional TTF-1 sites do contain the CAAG core as in hdio2 in contrast to the typical TAAT core recognition sequence of the known HDs (12, 34, 38).
The TTF-1 HD binding to the core is significantly affected by the flanking nucleotides. On the basis of the comparison of the characterized TTF-1 binding sites with the available in vitro TTF-1 HD binding data, the presence of a TCAAG(G/T) motif is generally required for efficient TTF-1-governed transcription (12, 23). However, this rule must be tempered by functional studies on the rat Tg promoter. Three TTF-1-protected regions (A–C) were found by DNAse footprinting each containing a TCAAGT motif. However, only two are functional since studies in FRTL-5 cells show that mutation of the B motif has a limited influence on promoter activity (12, 30, 39). This discrepancy between in vitro TTF-1 binding data and functional relevance was also demonstrated in the bovine Tg upstream enhancer element, where one of the three sites footprinted by TTF-1 is not functional (37).
In the case of hdio2, the flanking nucleotides might explain the difference between the TTF-1 HD binding of the C1 and D sites. The strong in vitro binding of the D site may be due to the G in position 7 (1 T CAAG G/T G 7), which increases the TTF-1 HD binding in vitro (23) since only the D site fulfills this criterion among the A, B, C, and D hdio2 promoter TTF-1 binding sites. However, the discrepancy between the in vitro binding and in vivo function of the C1 and D sites is in accordance with the data showing that the TTF1-HD affinity of a site does not necessarily parallel its function. Clearly, the latter criterion is the most specific requirement.
It is of interest that despite the 73% identity between the 3'-portion of the rat and human dio2 promoters, the C1 and D TTF-1 binding sites of the hdio2 are not present in rdio2. We conclude that this is responsible for the very different TTF-1 responsiveness of these two genes and is likely to explain the striking difference in the expression levels of the D2 mRNA in the human and rat thyroid. It may also be noted that, in the recently published mouse dio2 promoter region (40), both the C1 and D TTF-1 binding sites are not present, suggesting this is characteristic of the rodent dio2 gene. These TTF-1 binding sites may also be relevant to expression of D2 in other human TTF-1-expressing tissues such as the developing brain.
It should also be noted that D2 activity is lower in human thyroid than might be expected from the high level of D2 mRNA. There are several potential reasons for this. The 5'-untranslated region (UTR) of the D2 mRNA contains three short open reading frames, which may reduce translation efficiency (21). It is also known that the proteasomal destruction of D2 is markedly accelerated by T4 deiodination (41, 42). If the T4 in human thyroid cells is deiodinated by D2, this would also reduce the ratio of D2 protein to D2 mRNA.
TTF-1 is present in a variety of thyroid and lung tumors (43). TTF-1 expression is reduced in certain human thyroid carcinoma cell lines, and its expression can be restored by demethylating agents (44). On the other hand, the nuclear extract from toxic human thyroid adenomas contains elevated levels of TTF-1 while phosphorylated CREB is reduced (45). Thyroid adenomas also express high D2 mRNA levels (7). These results suggest that it is TTF-1, not cAMP, that accounts for the high levels of D2 mRNA in human thyroid cells. Since TTF-1 is reduced in many thyroid carcinomas, one might expect that D2 as well would be reduced and could serve as an additional marker for differentiation.
The expression of TTF-1 in the human central nervous system has not
been localized, and there is no direct comparison of D2 expression in
the human and rat brain and pituitary. However, D2, but not D1,
activity is found in the adult human central nervous system (46). We
also found no D1, but high D2, activity in two 
-glycoprotein
pituitary tumors (7). Further studies are required to investigate
whether TTF-1 expression increases D2 expression in the human pituitary
and brain.
With respect to other potential influences on D2 expression, we have shown that the hdio2 gene responds to PKA via a single canonical CRE located 90 bp 5' to the TSS (21). In vivo D2 activity and mRNA are increased in rat brown fat by a cAMP-mediated pathway linked to the sympathetic nervous system, and D2 activity is induced in cultured astrocytes by (Bu)2cAMP (4, 22, 27, 47). This response can be explained by the single canonical CRE in the rdio2 promoter in the same position as that in hdio2.
While the hdio2 contains a consensus AP-1 site, we could not induce the hdio2 promoter in HEK-293 cells by a combination of the PKC activator, phorbol ester, and A23187 (21). On the other hand, in primary cultures of human thyroid cells, TPA causes a 50% decrease in D2 activity (48). While neither the wild-type nor AP-1 mutant hdio2 constructs responded to phorbol ester in our transient expression system, mutation of the AP-1 site caused a 50% increase in basal promoter activity, suggesting a constitutive negative effect due to this site.
In conclusion, the present findings indicate that D2 expression in the human thyroid is positively controlled by TTF-1 via two DNA binding sites, C1 and D, which are not present in the rdio2 gene. The lack of the C1 and D sites in the rat D2 promoter can explain the very low D2 mRNA levels in the rat thyroid. The human dio1 gene does not respond to TTF-1 or Pax-8, but both human and rat dio2 genes respond to cAMP. These results, together with the presence of D2 mRNA in the human, but not rodent, myocardium and skeletal muscle, indicate that species-specific differences in dio2 genes may cause significant phenotypic differences in tissue D2 expression.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
) generated from a rat D2 cDNA clone kindly
provided by Drs. V. Galton and D. St. Germain and labeled by
[32P]dCTP. Four positive clones were
isolated and two (12–15 kb each) were subcloned into Bluescript KS.
One of these was digested 5' by NotI and 3' by
BamHI; the latter was in the 5'-UTR 160 bp 5' to the
initiator ATG of the rat D2 cDNA previously reported (GenBank U53505,
position 404). The approximately 4.4-kb fragment, which hybridized to
the labeled rD2/d1s, was subcloned into the pOCAT2 vector to form
rdio2#1. The 5'-transcriptional start site was identified by comparison
to the hdio2 gene. The rdio2 5'-FR has been
entered in GenBank (accession no. AF249274).
|
6.5 kb hdio2 5'-FR (hdio2–6.5) or its
most 3' 633 bp (hdio2–633) has been described (21). The latter
contains an artificial SacI site just 5' to -633 so that a
SacI/HindIII deletion was used to remove the
fragment between -633 and -610, which contained the two putative
TTF-1 binding motifs, C1 and C2 (hdio2–633?C) (Fig. 4). A
BglII deletion removed the region between -492 and -84,
including the D site and the CRE (hdio2–633?D). The hdio2–633?CD
construct was obtained by digesting hdio2–633?C by BglII
combining the two internal deletions in the same plasmid. A
SacI-BglII deletion removed the region between
-633 and -84 (hdio2–83).
PCR-based mutagenesis was used to introduce point mutations into the C1
and D TTF-1 binding sites, changing the CAAG core to CgtG in both. To
prepare the C1 mutant (mut), D wild-type (wt) construct (hdio2C1Mut), a
mut sense (Bp56, Table 2) and a wt antisense (hD2g12r) oligonucleotide
were used. The SacI/Pst (+7; blunt) digested fragment was
inserted between the SacI and HincII sites of
pOCAT2. For the C1 wt, D mut construct (hdio2DMut) overlap-extension
Vent PCR was used. In brief, Tp12 (wt sense) and Bp57 (mut antisense,
Table 2) fragment and the hD2DS1Mut (mut sense) and hD2g12r (wt
antisense) fragment were combined by PCR using Tp12 and hD2g12r as the
outside oligonucleotides, and the 3'-blunt-ended fragment inserted
into pOCAT2 at a SacI and blunt-ended PstI site.
The C1, D double mutant (hdio2C1DMut) was constructed by exchanging the
BglII fragment of the hdio2DMut with that in the
hdio2C1Mut.
To construct the AP-1 mutated 633-bp hdio2 construct (hdio2-AP1Mut), the Tp12-Bp110 Vent fragment was cut by HindIII and AvrII and inserted between the corresponding sites of the wild-type hdio2–633 CAT, mutating the TGACTCA site to aGACctc.
The rdio2#1 constructs contained a NotI and a
BamHI site, the latter located in the 5'-UTR. The fragment
was blunted and subcloned into the HincII site of pOCAT2 and
contained 3.8 kb rdio2 5'-FR and 600 bp rD2 5'-UTR.
This 3'- end was 160 bp 5' to the initiator ATG of the rD2 coding
region (GenBank U53505). Further 5'-truncations were performed starting
with the rdio2#1, removing an EcorI-HindIII
(resulting in rdio2#3) or EcorI-BglII fragment
(resulting in rdio2#4) (Fig. 10). The two truncated rdio2
constructs contain 658 or 83 bp 5' to the putative rdio2
TSS, respectively, as well as a portion of the 5'-UTR sequence. All
constructs were sequenced to confirm the accuracy of the constructions
and the mutations.
DNA Transfection and CAT Expression Assays
The reporter CAT plasmids were transfected into U87 or COS-7
cells as previously described using calcium phosphate precipitation
(26). For each transfection, 10 µg pOCAT2-based vector were
cotransfected with 0.06 or 0.1 µg rat TTF-1 for U87 or COS-7 cells,
respectively (10). The Pax-8 responsiveness was studied in COS-7 cells
using 0.5 µg Pax-8 coding expression vector (17). The CP5-CAT was
used as positive control (28). The TTF-1, Pax-8, and CP5-CAT clones
were kindly provided by Dr. R. Di Lauro. The rat dio2
response to the -catalytic subunit of PKA was tested in HEK-293
cells using 10 µg of rdio2CAT plasmid and 100 ng PKA -catalytic
subunit expressing plasmids. The PKA- expressing plasmid was a gift
of Dr. R. Maurer (50). For the uninduced control, a cytomegalovirus
(CMV)-ßGal or an empty CDM-8 vector was transfected in the same
quantity as the TTF-1- or Pax-8-expressing vector. The cells were
sonicated and assayed for CAT activity as described by Seed and
Sheen (51). The transfection efficiency was monitored by the
cotransfection of 3 µg TKhGH, as described previously (26). In the
Pax-8 experiments, the pXGH5 plasmid was used (52). The results are
expressed as the mean of CAT activity/hGH ±
SEM. Each construct was studied in duplicate in
at least three separate transfections.
Semiquantitative EMSA
Sense and antisense oligonucleotides containing hdio2
putative TTF-1 binding sites were annealed to produce double-stranded
DNA. As a positive control, we used the high-affinity TTF-1 binding
site (site "C") of the rat Tg promoter sequence (see Fig. 5). For
probe generation, fragment D and the positive control were labeled by
T4 polynucleotide kinase using 
[32P]dATP.
The binding buffer contained 20 mM Tris-HCl (pH
7.5), 200 mM KCl, 1 mM
dithiothreitol, 10% glycerol, 1 mg BSA per ml, 1 mg poly(dI-dC) per
ml, and varying concentrations of the unlabeled DNA fragment. The
protein was purified rat TTF-1 HD protein expressed in bacteria, kindly
provided by Dr. R. Di Lauro (34). The unlabeled competitor was
incubated with HD for 10 min at room temperature. Then 30,000 cpm
probe was added, and the mixture was incubated for an additional 30 min
at room temperature. Bound and free probe were separated in a
nondenaturing PAGE (7%) at 4 C in 1xTBE running buffer. After drying,
autoradiography was performed at -80 C for approximately 3 h. For
the semiquantitative analysis of competitive potency, samples were run
at least twice. The density was determined using a Computing
Densitometer (software v.3.22, Molecular Dynamics, Inc.
Sunnyvale, CA). The density of the shifted band was expressed as a
fraction of the sum of the densities of the upper and lower bands for
each lane. This was then expressed as a fraction of that found without
competitor.
Northern Blots
Male Sprague Dawley rats were made hypothyroid using 0.05%
methimazole for 3 weeks under a protocol approved by the Animal Use
Committee. Hypothyroidism was verified by the absence of
T4 in serum by T4 RIA. All
human tissues were obtained under protocols approved by the
Institutional Review Board at the Brigham and Women’s Hospital
(Boston, MA). Thyroids were collected at the operating room immediately
after removal from the patients and placed on ice. After the
pathologist’s inspection and removal of relevant samples, the
remaining tissues were immediately frozen in liquid nitrogen. This was
within approximately 15 min of removal from the patient. After this,
they were transferred to -80 C and kept at that temperature until
aliquots were processed for either RNA extraction or deiodinase assays.
RNA was isolated according to the single-step RNA isolation method of
Chomczynski (53). Various amounts were precipitated with 0.3
M sodium acetate, washed in 75% ethanol, resuspended in
1x RNA loading dye, and electrophoresed at room temperature according
to the procedure of Lehrach et al. (54). Thirty micrograms
of rat thyroid gland RNA of hypothyroid and euthyroid rats were loaded
on a 1.2% agarose/1x3-(N-morpholino)propanesulfonic
acid/1.85% formaldehyde gel containing 0.5 µg/ml ethidium bromide.
The RNA was transferred to Genescreen Plus (NEN Life Science Products, Boston, MA) by the capillary method against 10xSSC
and probed first with the 1.1-kb EcoRI fragment from rD2
in pBS-KS (4). The cDNA probe was labeled by the random primer method
using Prime-It kit (Stratagene, La Jolla, CA) and
[32P]dCTP. One million counts/min per ml of
hybridization solution was applied. Blots were hybridized at 42 C for
16 h in 40% formamide/0.64 M NaCl/0.8
mM EDTA, pH 7.4/0.04 M
PO4 buffer, pH 5.6/2xDenhardt’s solution/1.6%
SDS/7% dextran sulfate/80 µg/ml sonicated denatured thymus DNA.
After hybridization, blots were sequentially washed three times with
2xSSC/0.1% SDS at room temperature, 20 min; once at 42 C, 20 min;
0.5xSSC/0.1% SDS at 42 C for 20 min, 0.2xSSC/0.1% SDS at 42 C for
20 min; and then at 50 C, 55 C, 60 C. Blots were exposed to film for 2
weeks. They were stripped of residual signal, and then probed with a
mouse ß-actin probe (450 bp, KpnI/SacI),
random-primed, and washed as above. Blots were exposed to film for 1
day. Blots were then stripped and checked to ensure that there was no
residual signal, and then probed with rD1 probe [G21-KS
(XhoI/HindIII), 775 bp fragment]. Hybridization
was as for rD2 except that 100 µg/ml of denatured, sonicated thymus
DNA was used. The blots were washed up to 0.2xSSC/0.1% SDS at 45 C
for 30 min and exposed for 2 days. Densitometry was performed using a
computing densitometer (see above). The rD2 and rD1 signals were
normalized to that of ß-actin.
D2 Activity
The thyroids of three adult (12–14 weeks) male Sprague Dawley
rats and a patient with TSH-induced hyperthyroidism were sonicated in
0.1 M potassium phosphate buffer (pH 6.9), 1 mM
EDTA (PE) buffer containing 0.25 M sucrose and 10
mM dithiothreitol (DTT). The animal and the human samples
were obtained under approved of animal and IRB protocols.
Sonicated protein (100 µg) was assayed in duplicate for 5'-deiodinase
activity for 120 min at 37 C in a final volume of 300 µl PE buffer
containing 20 mM DTT and 2 or 100 nM
T4 or 10 µM
rT3 in the presence of approximately 90,000 cpm
of [125I]T4 purified by
chromatography on an LH-20 column (7). Five and 20 µg of sonicate
protein were assayed in duplicate for D1 activity for 60 min at 37 C in
300 µl PE buffer containing 1 µM rT3, 120,000 cpm of
[125I]rT3, and 10 mM DTT. The
125I-
and substrate were separated by
trichloroacetic acid (TCA) after the addition of horse serum as
described previously (55). Quantities of sonicate were adjusted to
consume less than 30% of the substrate. The background for the assays
was 1.8% of total T4 and 1.5% of total
rT3.
Sequencing
The rdio2 5'-FR fragment and all constructs were
sequenced in the ABI Prism 377 automated sequencer using dye
terminators.
Reagents
[125I]T4 or
rT3 and or [32P]
were purchased from NEN Life Science Products
(Boston, MA). Other chemicals were of molecular biology or reagent
grade. All primers were synthesized by Life Technologies, Inc. (Gaithersburg, MD).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This work was supported by NIH Grants DK-36256 and T-32-DK-07529.
1 These authors contributed equally to this paper. 
2 The sequence reported in this paper has been
deposited in the GenBank under accession no. AF249274.
Received for publication August 2, 2000. Revision received September 27, 2000. Accepted for publication October 3, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Gereben, A. M. Zavacki, S. Ribich, B. W. Kim, S. A. Huang, W. S. Simonides, A. Zeold, and A. C. Bianco Cellular and Molecular Basis of Deiodinase-Regulated Thyroid Hormone Signaling Endocr. Rev., December 1, 2008; 29(7): 898 - 938. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Canettieri, A. Franchi, M. D. Guardia, I. Morantte, M. G. Santaguida, J. W. Harney, P. R. Larsen, and M. Centanni Activation of Thyroid Hormone Is Transcriptionally Regulated by Epidermal Growth Factor in Human Placenta-Derived JEG3 Cells Endocrinology, February 1, 2008; 149(2): 695 - 702. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. L. Chik, M. T. Wloka, D. M. Price, and A. K. Ho The Role of Repressor Proteins in the Adrenergic Induction of Type II Iodothyronine Deiodinase in Rat Pinealocytes Endocrinology, July 1, 2007; 148(7): 3523 - 3531. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zeold, M. Doleschall, M. C. Haffner, L. P. Capelo, J. Menyhert, Z. Liposits, W. S. da Silva, A. C. Bianco, I. Kacskovics, C. Fekete, et al. Characterization of the Nuclear Factor-{kappa}B Responsiveness of the Human dio2 Gene Endocrinology, September 1, 2006; 147(9): 4419 - 4429. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Dentice, V. Cordeddu, A. Rosica, A. M. Ferrara, L. Santarpia, D. Salvatore, L. Chiovato, A. Perri, L. Moschini, C. Fazzini, et al. Missense Mutation in the Transcription Factor NKX2-5: A Novel Molecular Event in the Pathogenesis of Thyroid Dysgenesis J. Clin. Endocrinol. Metab., April 1, 2006; 91(4): 1428 - 1433. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Kohrle, F. Jakob, B. Contempre, and J. E. Dumont Selenium, the Thyroid, and the Endocrine System Endocr. Rev., December 1, 2005; 26(7): 944 - 984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Dentice, C. Luongo, A. Elefante, R. Ambrosio, S. Salzano, M. Zannini, R. Nitsch, R. Di Lauro, G. Rossi, G. Fenzi, et al. Pendrin Is a Novel In Vivo Downstream Target Gene of the TTF-1/Nkx-2.1 Homeodomain Transcription Factor in Differentiated Thyroid Cells Mol. Cell. Biol., November 15, 2005; 25(22): 10171 - 10182. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. W. J. S. Wassen, W. Klootwijk, E. Kaptein, D. J. Duncker, T. J. Visser, and G. G. J. M. Kuiper Characteristics and Thyroid State-Dependent Regulation of Iodothyronine Deiodinases in Pigs Endocrinology, September 1, 2004; 145(9): 4251 - 4263. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Fekete, B. Gereben, M. Doleschall, J. W. Harney, J. M. Dora, A. C. Bianco, S. Sarkar, Z. Liposits, W. Rand, C. Emerson, et al. Lipopolysaccharide Induces Type 2 Iodothyronine Deiodinase in the Mediobasal Hypothalamus: Implications for the Nonthyroidal Illness Syndrome Endocrinology, April 1, 2004; 145(4): 1649 - 1655. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. C. Moeller, S. Kimura, T. Kusakabe, X.-H. Liao, J. Van Sande, and S. Refetoff Hypothyroidism in Thyroid Transcription Factor 1 Haploinsufficiency Is Caused by Reduced Expression of the Thyroid-Stimulating Hormone Receptor Mol. Endocrinol., November 1, 2003; 17(11): 2295 - 2302. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Dentice, C. Morisco, M. Vitale, G. Rossi, G. Fenzi, and D. Salvatore The Different Cardiac Expression of the Type 2 Iodothyronine Deiodinase Gene between Human and Rat Is Related to the Differential Response of the dio2 Genes to Nkx-2.5 and GATA-4 Transcription Factors Mol. Endocrinol., August 1, 2003; 17(8): 1508 - 1521. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L.-H. Wang, R. Chmelik, and M. Nirenberg Sequence-specific DNA binding by the vnd/NK-2 homeodomain of Drosophila PNAS, October 1, 2002; 99(20): 12721 - 12726. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Gereben, A. Kollar, J. W. Harney, and P. R. Larsen The mRNA Structure Has Potent Regulatory Effects on Type 2 Iodothyronine Deiodinase Expression Mol. Endocrinol., July 1, 2002; 16(7): 1667 - 1679. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. C. Bianco, D. Salvatore, B. Gereben, M. J. Berry, and P. R. Larsen Biochemistry, Cellular and Molecular Biology, and Physiological Roles of the Iodothyronine Selenodeiodinases Endocr. Rev., February 1, 2002; 23(1): 38 - 89. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
