| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Molecular Endocrinology Laboratories, Departments of Physiology, Nuclear Medicine, Pediatrics (R.W.), and Medicine, and the Program in Molecular Medicine (M.L.Z.), University of Massachusetts Medical School, Worcester, Massachusetts 01655
Address all correspondence and requests for reprints to: Jack L. Leonard, Ph.D., Molecular Endocrinology Laboratories, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655. E-mail: jack.leonard{at}banyan.ummed.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2500 U/mg protein) and contained abundant
levels of the 7.5-kb SeD2 mRNA. However, no immunoreactive 30-kDa SeD2
protein was identified by Western analysis, immunoprecipitation, or
immunocytochemistry, and the specific C-terminus antiserum failed to
immunoprecipitate deiodinase activity from
(Bu)2cAMP-stimulated astrocytes, brown adipose tissue or
brain. Instead, the native 7.5-kb SeD2 mRNA encoded a 15-kDa protein
that terminated at the first UGA codon and contained the catalytically
inactive, N-terminal 129 amino acids of SeD2. These data show that the
native 7.5-kb SeD2 mRNA in stimulated astrocytes does not encode D2. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The 29-kDa substrate-binding subunit, p29, is the best characterized subunit of mammalian D2. This subunit was identified by affinity-labeling techniques (4, 22) and was characterized by showing that the rate of bromoacetyl-T4 (BrAcT4) label incorporation equaled the rate of BrAcT4-dependent inhibition of D2 activity, and that the quantity of affinity-labeled p29 paralleled D2 activity levels under conditions of enzyme over- and underexpression (22). Competition studies revealed that substrates, but not products, selectively blocked the affinity labeling of p29 and preserved D2 activity (22). Gel filtration of the detergent-soluble enzyme showed that D2 had a Mr of approximately 200,000 (4, 23) and was composed of one or more p29 subunits, an approximately 60-kDa cAMP-induced activation protein (CAP) (23), and one or more catalytic subunits.
One major difference between D2 and the other selenodeiodinases (D1 and D3) is the failure of Se deficiency to eliminate catalytic activity (12, 24, 25). In Se-depleted astrocytes, steady state levels of D2 activity are modestly, if at all, affected, and the synthesis of its p29 subunit is not altered (12), even though two other selenoenzymes, glutathione peroxidase (GPx) and D3, are almost completely lost (12, 15). Similarly, in vivo, cerebrocortical D2 activity is unaffected by Se deficiency, whereas brain GPx levels show the expected fall (24). Differential preservation of the different classes of selenoproteins has been proposed to account for the lack of effect of Se on D2 activity during periods of Se deprivation; however, the short biological half-life (t1/2, <20 min) of D2 and its low abundance in the cell make this protein especially susceptible to an impaired translational machinery. In fact, direct analysis of expressed selenoproteins in astrocytes failed to identify any 30-kDa translation product(s), suggesting that D2 is not Se dependent and that the p29 subunit of D2 is not a selenoprotein (12).
Despite characterization of the biology and biochemistry of D2 and its p29 subunit, a selenoprotein with the some of the catalytic properties of D2 from frog skin was recently cloned (26). Homology cloning led to the discovery of a similar gene product in rat BAT and human brain (27), challenging the widely held view that D2 was not a selenodeiodinase. Unlike mammals, frogs lack D1 and rely on a D2-like enzyme to catalyze T3 production during metamorphosis (28). Galton and co-workers isolated a full-length, frog type II selenodeiodinase complementary DNA (cDNA), SeD2, containing both the selenocysteine (SeC) codon (UGA) and the obligatory selenocysteine insertion sequence (SECIS) located in the 3'-untranslated region (3'UTR) (26). In all eukaryotic transcripts, the UGA triplet functions as a stop codon unless it is reinterpreted to code for SeC insertion by a specific stem loop element (SECIS) located in the 3'UTR (29). Transient expression of the frog SeD2 cDNA in COS cells produced a deiodinase that catalyzed the PTU-resistant 5'-deiodination of T4 (26), consistent with the D2 operational classification (1, 2, 3).
A homologous, partial gene product (1.5-kb cDNA) was isolated from a
cold stimulated rat BAT cDNA library that coded for a putative 30-kDa
selenoprotein with approximately 80% amino acid identity to that of
frog SeD2 (27). However, attempts to express a functional enzyme failed
(27), presumably due to the lack of the SECIS in the BAT SeD2 cDNA. As
we showed that attaching a SECIS to the 3'UTR of any eukaryotic
transcript containing a UGA in its open reading frame reinterprets this
triplet, programs SeC incorporation, and produces a selenoprotein (30),
it was not surprisingly to find that fusion of an exogenous SECIS to
the BAT SeD2 cDNA produced an artificial construct that yielded a
functional deiodinase when transfected into COS cells (27). Northern
analysis showed that the putative mammalian SeD2 was encoded by a
7.5-kb transcript that was abundant in the BAT, pituitary, and brain of
rats. In humans, the SeD2 messenger RNA (mRNA) is also found in
placenta, skeletal muscle, skin, and thyroid (27, 31), organs that show
little, if any, D2 activity in other mammals; only human placental
membranes contain a putative D2-like deiodinase (32). To date, the
mammalian homolog of SeD2 remains a virtual enzyme, because expression
of the native SeD2 polypeptide has not been identified, and the
required SeD2 SECIS has not been found.
To determine whether the native homolog of virtual SeD2 was the missing subunit of the 200-kDa brain D2, we set out to identify the native SeD2 protein in cultured astrocytes, BAT, and brain. In this report we show that this virtual 30-kDa selenoprotein is not expressed in mammals and is unrelated to native D2 activity, and that the native 7.5-kb SeD2 transcript codes for a 15-kDa protein lacking both deiodinating activity and SeC residue(s) in mammals.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-35S]Deoxy (d)-ATP (3000
Ci/mmol), [-32P]dCTP (800 Ci/mmol), and
[35S]Met (1000 Ci/mmol) were purchased from New England
Nuclear (Billerica, MA). Synthetic oligonucleotides were prepared
in-house or purchased from Life Technologies (Grand
Island, NY). BrAc-[125I]T4 and
[125I]rT3 was prepared by radioiodination of
BrAcT3 and 3,3'-diiodothyronine, respectively, using
methods described previously (33). DMEM, antibiotics, Hanks’ Balanced
Salt Solution, glucose, and trypsin were obtained from Life Technologies; supplemented bovine calf serum was obtained from
HyClone Laboratories, Inc. (Boulder, CO; selenium content,
42 ng/ml); (Bu)2cAMP and hydrocortisone were purchased from
Sigma Chemical Co. (St. Louis, MO). All iodothyronines
were of the L-configuration and were purchased from
Henning Berlin GmbH & Co.
Culture conditions
Rat type I astrocyte cultures were prepared by enzymatic
dispersion of neonatal rat brains as described previously (34). Cells
were grown in a humidified atmosphere of 5% CO2 and 95%
air at 37 C in complete growth medium, composed of DMEM supplemented
with 15 mM sodium bicarbonate, 33 mM glucose, 1
mM sodium pyruvate, and 15 mM HEPES, pH 7.4
(supplemented DMEM), and containing 10% calf serum, 50 U/ml
penicillin, and 90 µg/ml streptomycin. The culture medium was changed
twice weekly, and the cells were subcultured (3 x
104 cells/cm2) when they reached confluence
(5–7 days). Confluent cell monolayers between passages 1–4 were
used and contained more than 95% astrocytes, as determined by staining
for the astrocyte-specific protein, glial fibrillary acidic protein.
Unless otherwise noted, maximal D2 activity was induced in confluent
monolayers of astrocytes by growth in serum-free medium for 24 h,
followed by an additional 16 h in serum-free medium containing 1
mM (Bu)2cAMP and 100 nM
hydrocortisone.
Rat C6 astrocytoma cells were obtained from the American Tissue Culture Collection (Manassas, VA) and grown in supplemented DMEM containing 10% calf serum and antibiotics.
Functional SeD2 constructs
The 5'-overhangs of an approximately 1-kb
BamHI-XhoI restriction fragment containing the
coding region of SeD2 were filled in using the Klenow fragment of DNA
PolI, and the blunt-ended SeD2 was ligated into the
EcoRV site of pcDNA3. The eukaryotic expression plasmid
(pOPAL) was completed by inserting the AvrII-XbaI
restriction fragment of the GPX1 gene containing the SECIS (nucleotides
902-1132) into the XbaI site of the pcDNA3-SeD2 construct.
SeD2 orientation and construct integrity were confirmed by DNA
sequencing.
Production of SeD2 cells
C6 astrocytoma cells (200,000 cells/25-cm2
flasks) were transfected with 5 µg pOPAL-SeD2 construct using
DOTAP (Promega Corp., Madison, Wi) according to
the manufacturer’s instructions. After 24 h, the growth medium
was supplemented with 200 µg/ml G418 (Life Technologies), and the cells were grown until individual
G418-resistant clonal growth was visible, usually within 2–3 weeks.
Individual colonies of G418-resistant cells were isolated by limiting
dilution in the presence of 200 µg/ml G418, and the SeD2 cells used
were chosen based on functional SeD2 expression. SeD2 cells were
maintained in G418-supplemented medium to prevent loss of the SeD2
phenotype.
RNA isolation and Northern analysis
RNA was prepared from cell pellets by the method of Chomczynski
and Sacchi (35). Polyadenylated RNA was isolated by two cycles of
chromatography over oligo(deoxythymidine)-cellulose
(Stratagene, La Jolla, CA). RNA was separated on 1.2%
agarose-formaldehyde gels and transferred to nitrocellulose (Duralose,
Stratagene, La Jolla, CA) by diffusion blotting.
Hybridization probes were prepared by random primer labeling
(Promega Corp.), and blots were probed at 42 C for 16
h with 10 ng/ml 32P-labeled SeD2 cDNA (nucleotides 1–1368)
and 10 ng/ml 32P-labeled human glyceraldehyde-3-phosphate
dehydrogenase (G3PDH) cDNA control (CLONTECH Laboratories, Inc., Palo Alto, CA) in a solution composed of 50% formamide
(vol/vol), 1 M NaCl, 10 mg/ml SDS, 0.1 g/ml dextran
sulfate, and 100 µg/ml salmon sperm DNA. All blots were washed to
high stringency (30 mM NaCl, 3 mM sodium
citrate, and 1 mg/ml SDS) at 65 C for 15 min. Hybridization signals
were visualized and quantified by PhosphorImager analysis using
ImageQuant software (Molecular Dynamics, Inc., Sunnyvale,
CA). RNA loading differences were normalized using the hybridization
signals simultaneously generated with the control G3PDH probe.
Anti-SeD2 antibodies
Antibodies were raised in rabbits against synthetic peptides
corresponding to the C-terminus of full-length (anti-SeD2-long) and
truncated (anti-SeD2-short) SeD2. Two synthetic peptides to the
full-length SeD2 protein were prepared: one corresponding to amino
acids 247–266 [NH2-YNLQEVRSWLEKNFSKRCILD-COOH,
SeD2(266)] contained a conservative substitution of a Cys for SeC at
position 263 and one corresponding to amino acids 247–262 lacking the
terminal four amino acid residues (-CILD-COOH) of SeD2(263). One
synthetic peptide corresponding to the C-terminus of the truncated,
15-kDa polypeptide programmed when the first UGA codon of SeD2 mRNA
signals translational arrest, amino acids 115–129
[NH2-YASAERPLVVNFGSAT-COOH, SeD-(2–129)], was also
synthesized. All peptides contained an N-terminal tyrosine to
facilitate diaminobenzidine coupling to keyhole limpet hemocyanin and
for radioiodination.
Domain-specific SeD2 antibodies were purified by affinity chromatography before use on peptide-Affigel 10 (Bio-Rad Laboratories, Inc., Hercules, CA) affinity matrices. Primary antiserum was adsorbed to the cognate peptide-Affigel 10 matrix at 25 C, and unbound proteins were removed by repeated washes with 150 mM NaCl and 20 mM sodium phosphate buffer (pH 7.4; PBS). The bound, peptide-specific antibodies were eluted with 100 mM acetic acid and monitored by absorbance at 280 nm. Fractions containing antibody were neutralized by adding 0.1 vol 1 M Tris buffer (pH 8.6), pooled, and stored frozen at -70 C until use.
Immunocytochemistry
Cells were seeded onto glass coverslips (22 x 22 mm)
coated with poly-D-lysine (10 µg/ml) and grown for 1–4
days as indicated in individual experiments.
(Bu)2cAMP-stimulated astrocytes were treated with 10
mM colchicine for 60 min to relax the cell border before
fixation; SeD2 cells were fixed directly. All cells were fixed with 4%
paraformaldehyde and permeabilized with 0.1% Triton X-100 (vol/vol) in
PBS. SeD2-related proteins were identified by incubation with
affinity-purified, domain-specific anti-SeD2 IgG (0.1–1 µg/ml) for
2 h at 4 C. Where indicated, peptide-blocked IgG was prepared by
preincubation of the anti-SeD2 IgG with a 100-fold molar excess of
cognate peptide for 60 min at room temperature before incubation with
the fixed cells. Immune complexes were identified with Texas
Red-conjugated, antirabbit IgG, and the coverslips were mounted and
examined with epifluorescence illumination using a Zeiss Axioskop
microscope (Carl Zeiss, New York, NY) equipped with an
Olympus Corp. OM4 camera (New Hyde Park, NY). The
micrographs shown are representative of 30–40 independent fields.
Antisense deoxyoligonucleotide effects on D2 activity in SeD2 and
stimulated glial cells
SeD2 and astrocytes were seeded at 5 x 104
cells/well in six-well cluster plates and grown in complete growth
medium for 24 h before the experiment. Increasing concentrations
of a 24-mer thiophosphinate ester oligonucleotide complementary to the
Kozak start site of SeD2 (5'-CTGAGGAGTCCCATGTTCTTTACC-3', SeD2
antisense) were added to the growth medium, and the cells were grown
for an additional 72 h. During the final 24-h period, D2 activity
was induced in the euthyroid glial cell by the addition of 1
mM (Bu)2cAMP and 100 nM
hydrocortisone. Cells were then washed free of the treatment medium,
scraped from the plate, and collected by centrifugation. D2 activity
was determined as detailed below; protein disulfide isomerase
(PDI) expression was determined by Western blot analysis.
Hybrid selection of SeD2-encoding mRNA
Hybrid selection of SeD2-encoding mRNAs was performed using
established procedures (36). The prBAT 1–1 plasmid was denatured,
filtered through 0.45-mm pore size nitrocellulose (10 µg
plasmid/cm2), and immobilized by UV cross-linking.
Approximately 50 µg heat-denatured, polyadenylated RNA from
approximately 5 x 108 cells of each cell type were
hybridized to the nitrocellulose-immobilized SeD2 cDNA for 2 h at
50 C in a hybridization solution composed of 65% formamide (vol/vol),
400 mM NaCl, 0.2% (wt/vol) SDS, 30 mM
1,4-piperazine-diethanesulfonic acid buffer (pH 6.5), and 50 µg/ml
yeast transfer RNA. Unbound mRNA was eluted from the nitrocellulose
filters by repeated washes at 65 C with 150 mM NaCl, 10
mM Tris buffer (pH 7.6), and 1 mM EDTA, and the
specifically absorbed mRNA was eluted with diethylpyrocarbonate-treated
H2O containing 75 µg/ml yeast transfer RNA. Eluted mRNAs
were extracted once with phenol-chloroform-isoamyl alcohol (50:48:2,
vol/vol/vol), precipitated by ethanol, resuspended in
diethylpyrocarbonate-treated H2O, and stored at -70 C
until use.
Xenopus laevis oocyte expression of functional SeD2
Stage 5–6 X. laevis oocytes were defolliculated and
microinjected by established procedures (14) with 2–5 ng hybrid
selected mRNA from C6-, SeD2-, and cAMP-stimulated astrocytes. After
injection, oocytes were incubated for 4–5 days in modified Barth’s
medium. Pools of two oocytes were lysed in 100 µl 10 mM
HEPES (pH 7.0), 1 mM EDTA, and 10 mM
dithiothreitol and used for enzyme analysis.
Analytical procedures
D2 activity was determined in cell lysates by measuring the
release of radioiodide from 2 nM
[125I]rT3 (5 x 105
cpm/pmol), unless otherwise indicated, at 20 mM
dithiothreitol and 1 mM PTU in a total volume of 100 µl.
Deiodination reactions were performed in triplicate in 100
mM potassium phosphate buffer (pH 7.0) and 1 mM
EDTA with 50–100 µg cell protein and incubated at 37 C for 60–120
min. Product formation was measured as described previously (5), and
the data are expressed as units per mg protein, where 1 U = 1 fmol
I- released/h.
All experiments were performed at least three times, and statistical analysis was performed using Student’s t test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
are the SeD2 mRNA levels (A) and the D2
activity (B) in astrocytes treated with (Bu)2cAMP in the
absence or presence of hydrocortisone. Unstimulated hypothyroid
astrocytes lacked both D2 activity and the SeD2 transcript, whereas
hydrocortisone treatment induced the 7.5-kb SeD2 mRNA without a
corresponding appearance of D2 activity. (Bu)2cAMP
treatment induced both SeD2 mRNA and D2 activity, and maximal
expression of both SeD2 mRNA and D2 activity was observed in
hypothyroid astrocytes treated with both (Bu)2cAMP and
hydrocortisone. In euthyroid astrocytes, the abundance of SeD2 mRNA
reached levels identical to those in the hypothyroid astrocytes, but D2
activity reached only 20% of that in the hypothyroid cells.
|
, anti-SeD2-long
specifically recognized an epitope(s) in the C-terminus of SeD2 that is
upstream of the last four deduced amino acids SeD2(266) and SeD2(263),
but did not recognize peptides derived from the C-termini of rat D1,
rat D3, or the p29 subunit of D2. Anti-SeD2-short (Fig. 2B) recognized
only its cognate peptide.
|
We first examined the ability of the C-terminus-specific, anti-SeD2
antibody (SeD2-long) to immunoprecipitate catalytically active D2 using
detergent extracts of microsomes isolated from hypothyroid rat BAT and
cerebral cortex and detergent extracts of SeD2 cells and stimulated
astrocytes. D2-containing detergent extracts (BAT, 68 ± 7 U/mg
protein; cerebral cortex, 37 ± 3 U/mg protein; SeD2 cells,
49 ± 5 U/mg protein; (Bu)2cAMP-stimulated astrocytes,
89 ± 10 U/mg protein; n = 3) were incubated with
anti-SeD2-long antiserum (1:100 final dilution) in the absence or
presence of a 100-fold molar excess of blocking peptide. The immune
complexes were then removed using protein A-Sepharose, and the unbound
D2 activity remaining in the clarified supernatant was determined. As
shown in Fig. 3, the anti-SeD2-long
antiserum selectively immunoprecipitated 60% of D2 activity
solubilized from SeD2 cells, but failed to precipitate D2 activity in
detergent extracts from (Bu)2cAMP-stimulated rat BAT
or cerebral cortex. These data show that the anti-SeD2 long recognizes
the heterologous SeD2 translation product, but does not recognize the
native D2 in astrocytes or other rat tissues.
|
are immunoblots of the SeD2-related
proteins in the C6 null cell, the SeD2 cell, and unstimulated and
(Bu)2cAMP-stimulated astrocytes. Compared with the C6 null
cell, which lacked D2 activity, SeD2 cell lysates (390 U/mg protein;
D2) contained two new immunoreactive bands with
Mr of about 30 and 15 kDa, and both bands were
specifically blocked by excess peptide, indicating that both the
full-length (30 kDa) and the truncated (15 kDa) SeD2 proteins were
present in the transfected SeD2 cells. In contrast to the SeD2 cell,
(Bu)2cAMP-stimulated astrocytes (Fig. 4B) expressing
approximately 10-fold greater levels of D2 activity (2960 U/mg protein)
and maximal levels of SeD2 mRNA contained only the 15-kDa
immunoreactive protein, indicating that the first UGA of the SeD2 mRNA
terminated translation. No peptide-blocked 30-kDa protein was observed
in (Bu)2cAMP-stimulated astrocytes. Both unstimulated
astrocytes and C6 null cells lacked any specific 30- or 15-kDa
immunoreactive signal. Thus, the anti-SeD2-short antibody recognizes
both full-length (30-kDa) and truncated (15-kDa) SeD2 translation
products on immunoblots.
|
, the SeD2 activity was decreased by
about 80% after 3 days of selenium deficiency. These data confirm that
SeC is incorporated at the first UGA codon in transcripts derived from
the SeD2:SECIS cDNA chimera and that read-through is dependent on Se in
SeD2 cells.
|
are representative photomicrographs of
SeD2 cells, hypothyroid BAT, hypothyroid cerebral cortex, and
cAMP-stimulated astrocytes stained for full-length (30-kDa) SeD2 with
anti-SeD2-long IgG. Abundant, peptide-blocked immunostaining was
observed in the transfected SeD2 cells, consistent with the ability of
anti-SeD2 long to immunoprecipitate deiodinase activity. In contrast,
no specific staining was observed in (Bu)2cAMP-stimulated
astrocytes, hypothyroid BAT, or hypothyroid cerebral cortex. In control
studies, the p29 subunit of D2 was detected in all cells and tissues
containing D2 activity, such as brain, BAT, and stimulated astrocytes,
but not in the SeD2 cells expressing the virtual SeD2 protein (Fig. 6).
Although no immunoreactive full-length (30-kDa) SeD2 protein was found
in rat tissues, all cells and tissues expressing the native SeD2 mRNA
showed abundant immunoreactive, peptide-blocked signals that were
uniformly distributed throughout the cell when the 15-kDa SeD2
translation product was visualized with anti-SeD2-short antiserum (Fig. 7). These data show that a full-length
(30-kDa) native SeD2 protein is not translated from the 7.5-kb SeD2
mRNA in rat cells and organs, whereas the truncated 15-kDa product is
readily observed. Thus, the native SeD2 transcript programs the
translation of a catalytically inactive protein in rat tissues.
|
|
, addition of increasing concentrations
of SeD2 antisense led to a progressive loss of D2 activity in SeD2
cells compared with the D2 activity in cells treated with the control
antisense oligonucleotides for an unrelated protein (PDI). In contrast,
the D2 activity in euthyroid, (Bu)2cAMP-stimulated
astrocytes was unaffected by the SeD2 antisense oligonucleotide,
indicating that enzyme activity in astrocytes does not depend on any
translation product encoded by the SeD2 mRNA.
|
, inset)
confirming that an intact SeD2 mRNA was isolated by this enrichment
strategy. As shown in Fig. 9A, the hybrid-selected SeD2 transcripts
from SeD2 cells and (Bu)2cAMP-stimulated astrocytes
programmed the synthesis of the expected immunoreactive, 15-kDa SeD2
protein in reticulocyte lysates, illustrating that the enriched SeD2
mRNAs was intact and functional. As reticulocyte lysates cannot
incorporate SeC into the elongating polypeptide chain, only the 15-kDa
translation product from the enriched SeD2 transcripts was expected. As
shown in Fig. 9B, oocytes injected with hybrid-selected, artificial
SeD2 mRNA isolated from SeD2 cells programmed the synthesis of
functional D2 activity, whereas the enriched SeD2 mRNA from
(Bu)2cAMP-stimulated astrocytes produced no D2 activity.
RNA from control C6 cells also yielded no D2 activity. These data show
that SeD2 transcripts containing a heterologous SECIS encode a
functional D2, whereas the native 7.5-kb SeD2 mRNA programs the
synthesis of the N-terminal 15 kDa of SeD2, but not a full-length,
catalytically active enzyme, suggesting that the native transcript lack
the required SECIS.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The abundant D2 activity produced in stimulated astrocytes was exploited in the search for the native SeD2 translation product. Despite a rough correlation between the expression of the 7.5-kb SeD2 transcript and D2 activity, Western blot, immunoprecipitation, and immunohistochemistry failed to find any immunoreactive 30-kDa SeD2 protein in mammalian tissues. In contrast, the 15-kDa translation product encoded by the first 390 nucleotides of the open reading frame of the native SeD2 transcript was readily expressed. On the other hand, the catalytically active, virtual SeD2 was readily immunoprecipitated from cells transfected with the SeD2-SECIS chimera, but native D2 activity was unaffected. Direct analysis of translation products programmed by native SeD2 mRNA in situ using antisense oligonucleotides or in vitro using oocytes injected with hybrid selected SeD2 mRNA also failed to identify any native functional SeD2. Taken together, these data show that the catalytically active D2 found in mammalian tissues is not encoded by the native 7.5-kb SeD2 transcript, as the mature mRNA appears to lack the SECIS required to incorporate the SeC residue essential for full-length translation of this gene product.
The relative abundance of the 7.5-kb SeD2 transcript in the stimulated astrocyte was unexpected and was 35% of that for the glycolytic enzyme, G3PDH, indicating that this putative SeD2 transcript is not a rare message. In fact, relative to G3PDH mRNA, the quantity of the 7.5-kb SeD2 transcript is about 5-fold greater than that of the D1 mRNA in PK1 cells that express approximately 10,000-fold more deiodinating activity that astrocytes. The unexpected abundance of this mRNA suggests that the transcript either codes for another gene product, presumably the catalytically inactive 15-kDa protein, or is a remnant of ontogeny.
The large size of the SeD2 mRNA relative to that of other members of the mammalian selenodeiodinase family, is also an unusual feature of this transcript. From fish to man the mRNAs encoding D1 and D3 range in size from about 1.5–2.4 kb (13, 15, 17, 26, 40). Similarly, the full-length SeD2 transcript in Rana catesbeiana is only 1.5 kb in length (26). As the SECIS element in all known selenoproteins is located no more than 1.5 kb downstream from the UGA triplet, St. Germain and co-workers assumed that this essential element was located in the approximately 5 kb of the 3'UTR presumably lost during the cloning of rBAT 1–1 and its human homolog, even though both clones were polyadenylated (27), and a naturally occurring polyadenylation signal variant, AAUACA (41), is located 13 bases upstream from the polyadenylase tail present in the rBAT 1–1 clone. Recently, Toyoda and colleagues3 cloned an approximately 5.1-kb SeD2 cDNA (accession no. AB011068) from a rat brain cDNA library, adding more than 4 kb of 3'UTR to the sequence provided by rBAT 1–1. A complete search of the new 3'UTR failed to identify any SECIS element(s), indicating that this essential element, if present, would be located more than 5 kb downstream of the UGA codon, where it would function poorly, if at all. However, direct experimental analysis of the translation products encoded by the native 7.5-kb SeD2 mRNA clearly indicates that the mature transcript lacks a functional SECIS element.
Alternative splicing has also been proposed to lead to the loss of the 3'UTR and the presumed SECIS, from rBAT 1–1. Although only a single SeD2 transcript was observed in astrocytes, there is one putative splice site in the open reading frame of SeD2 (starting at nucleotide 1095) that could exchange the C-terminal 85 amino acids of the mammalian processed transcript. However, this hypothesis is very unlikely, because the C-terminus of the SeD2 protein shares more than 80% amino acid identity from amphibian to man. Although it seems likely that both alternate splicing and/or gene rearrangement of the primordial SeD2 transcript led to the large increase in the size of the mature transcript in mammals, this also eliminated the essential SECIS and disabled this potential selenodeiodinase transcript. Thus, the mammalian SeD2 transcript belongs to the pool of primordial gene products that have undergone repeated recombination and whose original translation product(s) has been altered. The only translation product of the native 7.5-kb SeD2 transcript found in rats is the 15-kDa truncated protein.
Unlike the mammalian virtual SeD2, the amphibian SeD2 fulfills all of the criteria for an authentic selenodeiodinase. The full-length cDNA contains the UGA triplet in the open reading frame and has the required SECIS in the 3'UTR of the transcript (26), and expression studies using unadulterated cDNA produced a deiodinase with the catalytic properties of the native frog enzyme. As frogs lack a PTU-sensitive D1, they rely on SeD2 to produce T3 at critical times during metamorphosis. Thus, the amphibian SeD2 most likely represents one of the earliest known members of the selenodeiodinase family with at least some of the characteristics of all of the mammalian deiodinase isozymes.
Many of the problems surrounding deiodinase isozyme identification have occurred (42, 43, 44) because the properties of these enzymes in most lower vertebrates deviate from the classical operational definitions of the mammalian isozymes. For example, neither the frog SeD2 nor a D1 variant in the fish kidney is inhibited by the classical D1 inhibitors, PTU, or thiogold glucose (40). Cloning of Tilapia kidney enzyme (TN-12) revealed that it was a true selenodeiodinase containing both a UGA triplet in the open reading frame and two SECIS elements in the 3'UTR (40), and the coding region of this cDNA shared approximately 50% identity with mammalian D1 (40). As the lack of PTU inhibition is a hallmark of the mammalian D2 enzyme (1, 2, 3), by functional analysis alone both the fish kidney and the amphibian skin D2 would have been mistakenly classified as D2 isozymes. In fact, closer analysis of the amphibian SeD2 reveals that the frog enzyme is inhibited by PTU, albeit at 10- to 50-fold higher concentrations than those required to inhibit D1 (28). Thus, it appears that the deiodinase(s) found in lower vertebrates represents intermediate forms in the evolving selenodeiodinase family and that PTU inhibition is a poor discriminator between isozymes when evaluating nonmammalian deio-dinases.
The influence of selenium on the expression of D2 catalytic activity has recently been revisited by Courtin and co-workers (45). They found that at early times after cAMP-dependent induction, D2 activity was temporally correlated with the appearance of the 7.5-kb SeD2 mRNA and that about 50% of the T4-deiodinating activity in acutely stimulated (6-h) primary brain cell cultures was lost after prolonged withdrawal of selenium. In light of our findings that the presumed SeD2 mRNA is unrelated to D2 activity in astrocytes and encodes a 15-kDa catalytically inactive polypeptide, the effects of selenium on T4 deiodination are probably related to the presence of the selenium-dependent D1 and D3 enzymes in primary brain cell cultures employed by these workers (34, 46) and by the absence of the D1 inhibitor, PTU, from their T4 deiodination assay. Alternatively, prolonged selenium deprivation may render the cell more susceptible to oxidative stress, and this would be especially damaging to enzymes found in the plasma membrane, such as D2, although we were unable to observe any effect on cAMP-induced, steady state, D2 activity after 5 days of selenium deprivation (12).
Although the identity of the remaining subunits of mammalian D2 is unknown, clearly SeD2 is not a component of the functional enzyme in vivo, and the p29 substrate-binding D2 subunit remains the only known subunit.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Safran, M., and J. Leonard, unpublished
observation.
3 Gondou, A., N. Toyoda, M. Niskikawa, S. Tabata,
T. Yonemoto, Y. Ogawa, T. Tokoro, N. Sakaguchi, F. Wang, and M. Inada,
submission to GenBank, February 12, 1998.
Received May 5, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G. I. C. Simpson, D. M. Leonard, and J. L. Leonard Identification of the Key Residues Responsible for the Assembly of Selenodeiodinases J. Biol. Chem., May 26, 2006; 281(21): 14615 - 14621. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Alkemade, E. C Friesema, G. G Kuiper, W. M Wiersinga, D. F Swaab, T. J Visser, and E. Fliers Novel neuroanatomical pathways for thyroid hormone action in the human anterior pituitary. Eur. J. Endocrinol., March 1, 2006; 154(3): 491 - 500. [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]
|
|
|
|
 |
|
 A. Alkemade, E. C. Friesema, U. A. Unmehopa, B. O. Fabriek, G. G. Kuiper, J. L. Leonard, W. M. Wiersinga, D. F. Swaab, T. J. Visser, and E. Fliers Neuroanatomical Pathways for Thyroid Hormone Feedback in the Human Hypothalamus J. Clin. Endocrinol. Metab., July 1, 2005; 90(7): 4322 - 4334. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Leonard, G. Simpson, and D. M. Leonard Characterization of the Protein Dimerization Domain Responsible for Assembly of Functional Selenodeiodinases J. Biol. Chem., March 25, 2005; 280(12): 11093 - 11100. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Curcio-Morelli, B. Gereben, A. M. Zavacki, B. W. Kim, S. Huang, J. W. Harney, P. R. Larsen, and A. C. Bianco In Vivo Dimerization of Types 1, 2, and 3 Iodothyronine Selenodeiodinases Endocrinology, March 1, 2003; 144(3): 937 - 946. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Montero-Pedrazuela, J. Bernal, and A. Guadano-Ferraz Divergent Expression of Type 2 Deiodinase and the Putative Thyroxine-Binding Protein p29, in Rat Brain, Suggests that They Are Functionally Unrelated Proteins Endocrinology, March 1, 2003; 144(3): 1045 - 1052. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. G. J. M. Kuiper, W. Klootwijk, and T. J. Visser Substitution of Cysteine for a Conserved Alanine Residue in the Catalytic Center of Type II Iodothyronine Deiodinase Alters Interaction with Reducing Cofactor Endocrinology, April 1, 2002; 143(4): 1190 - 1198. [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]
|
|
|
|
 |
|
-counter. C-Terminus D3 peptide, -RTWLERYDEQLHGTRPRRL;
C-terminus D1 peptide, YNPEEVRAVLEKLCIPPGHMPQF; C-terminus p29
peptide, LAQEMAFEEATPVDSLGGEKI.
