Mammalian Type II Iodothyronine Deiodinase
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 98, Number 2, July 1996, 405–417
Cloning of the Mammalian Type II Iodothyronine Deiodinase
A Selenoprotein Differentially Expressed and Regulated in Human and Rat Brain and Other Tissues
Walburga Croteau, Jennifer C. Davey, Valerie Anne Galton, and Donald L. St. Germain
Departments of Medicine and Physiology, Dartmouth Medical School, Lebanon, New Hampshire 03756
The deiodination of thyroid hormones in extrathyroidal tis-
sues plays an important role in modulating thyroid hor-
mone action. The type II deiodinase (DII) converts thyrox-
ine to the active hormone 3,5,3
rat is expressed in the brain, pituitary gland, and brown ad-
ipose tissue (BAT). Complementary DNAs (cDNAs) for the
types I and III deiodinases (DI and DIII, respectively) have
been isolated and shown to code for selenoproteins. How-
ever, information concerning the structure of the mamma-
lian DII remains limited, and the pattern of its expression in
human tissues is undefined. We report herein the identifica-
tion and characterization of rat and human DII cDNAs.
Both code for selenoproteins and exhibit limited regions of
homology with the DI and DIII. In the rat pituitary and
BAT, DII mRNA levels are altered more than 10-fold by
changes in the thyroid hormone status of the animal. North-
ern analysis of RNA derived from human tissues reveals ex-
pression of DII transcripts in heart, skeletal muscle, pla-
centa, fetal brain, and several regions of the adult brain.
These studies demonstrate that: (
are selenoproteins, (
) DII expression in the rat is regulated,
at least in part, at the pretranslational level in some tissues,
) DII is likely to be of considerable physiologic impor-
tance in thyroid hormone economy in the human fetus and
J. Clin. Invest.
1996. 98:405–417.) Key words: thy-
-triiodothyronine, and in the
) the rat and human DII
In some mammalian tissues, such as the developing brain, the
anterior pituitary gland, and brown adipose tissue (BAT)
where thyroid hormone appears to have particularly important
regulatory effects, a relatively high proportion of the receptor-
sue itself, rather than being derived from plasma (1–4). The ex-
pression in these tissues of the type II iodothyronine deiodi-
nase (DII), which catalyzes deiodination of thyroxine (T
exclusively on the outer ring (5
suggests that this enzyme is responsible for this “local” produc-
tion of T
and is thus important in influencing thyroid hormone
action in these tissues (5). In addition, DII activity is markedly
elevated in the hypothyroid state and appears to be responsi-
ble for catalyzing the production of a large proportion of the
under such conditions (6). For example, in neo-
natal rats, circulating T
levels are not reduced when the type I
deiodinase (DI) is inhibited by 6-
a specific inhibitor of this enzyme (2). Thus, DII is of critical
importance for both the local generation of T
sues and systemic thyroid hormone homeostasis.
Complementary DNAs for the DI and type III deiodinase
(DIII) have recently been isolated and demonstrated to con-
tain in-frame TGA codons that code for selenocysteine (7–10).
The catalytic properties and tissue patterns of expression of
these selenoproteins differ from those of the DII. Thus, DI ap-
pears to play an important role in converting T
the thyroid gland itself (11, 12), and, unlike DII, is expressed in
the liver and kidney and is capable of inner ring (3- or 5-posi-
tion) deiodination of sulfated thyroid hormone conjugates
(13). DIII functions as an inner ring deiodinase to convert T
to inactive metabolites (3,3
-diiodothyronine, respectively). Its expression in pla-
centa and several fetal tissues during early development sug-
gests that it plays a role in preventing premature exposure of
developing tissues to adult levels of thyroid hormones (14).
The DII also is present in several fetal and neonatal tissues
(15) and is essential for providing the brain with appropriate
levels of T
during the critical period of development (16).
Isolating cDNAs for DII has been problematic. Our at-
tempts at screening appropriate cDNA libraries under re-
duced stringency conditions using DI and DIII cDNAs as
probes has proved unsuccessful (Croteau, W., M. Schneider,
and D.L. St. Germain, unpublished data). Furthermore, unlike
the DI and DIII, the DII is poorly expressed in
oocytes after the injection of poly(A)
taining tissues (17), thus rendering this expression system un-
tenable as a tool for screening cDNA libraries. However, using
a PCR-based strategy, we have recently cloned a cDNA (des-
DII) for the DII of the amphibian species
(Davey et al., reference 18). In the present studies,
we used this amphibian DII cDNA to identify homologous rat
and human cDNAs. We have demonstrated that these cDNAs
code for selenoproteins with DII activity. The predominant
mRNA species for these mammalian enzymes are 6–8 kb in
size and are highly expressed in a number of tissues including
the fetal and adult brain of rats and humans, and in human
heart, placenta, and skeletal muscle.
) is produced within the tis-
-position) to yield T
in selective tis-
RNA from DII-con-
Address correspondence to Donald L. St. Germain, Dartmouth Med-
ical School, One Medical Center Drive, Lebanon, NH 03756. Phone:
603-650-7910; FAX: 603-650-6130; E-mail: email@example.com
Received for publication 13 February 1996 and accepted in revised
form 9 April 1996.
nase; BAT, brown adipose tissue; DI, type I deiodinase; DII, type II
deiodinase; DIII, type III deiodinase; PTU, 6-
-triiodothyronine; SECIS, selenocysteine insertion se-
Abbreviations used in this paper:
5-D, 5-deiodinase; 5
W. Croteau, J.C. Davey, V.A. Galton, and D.L. St. Germain
RNA preparation and Northern analysis.
viously described (19) from BAT of male Sprague-Dawley rats (Charles
River Laboratory, Wilmington, MA) (150–175 g) exposed to cold (4
for 24 h. Poly(A) RNA was isolated by one or two cycles of chroma-
tography over oligo(dT)-cellulose (Collaborative Biomedical Prod-
ucts, Bedford, MA). RNA from other rat tissues used for Northern
analysis was prepared by the same methods. RNA blots of human tis-
sues were purchased from Clontech (Palo Alto, CA) and contained
g of poly(A) RNA per lane according to the supplier’s specifi-
cations. Hybridization and washing of Northern blots were per-
formed as previously described for rat tissues (20), or according to the
supplier’s instructions for blots of human tissues, except that final
washes were performed at 42 or 60
iments, rats were rendered hypothyroid by the inclusion of 0.05%
methimazole in their drinking water for 25–30 d, or hyperthyroid by a
single daily subcutaneous injection of 50
before killing. After hybridization with the specific DII probes identi-
fied in these studies, blots were stripped and reprobed with a mouse,
rat, or human
-actin probe. Hybridization signals were quantified by
densitometric measurements of scanned computer images of the auto-
radiographs using the IPLab Gel program (Signal Analytics Corp.,
Vienna, VA) on a Macintosh computer. In all cases, signals obtained
using the DII probes were normalized using the signals generated
cDNA probe preparation.
cDNA probes for Northern analysis
and library screening were prepared by PCR using the appropriate
cDNAs as templates and gene-specific sense and antisense primers that
flanked the regions of interest. After amplification, the PCR reaction
products were separated on a low melt agarose gel stained with ethid-
ium bromide, and the appropriate band was then excised and purified
using a QIAquick Gel Extraction Kit (Qiagen Inc., Chatsworth, CA).
The PCR product was then labeled with
Kit from Pharmacia LKB Biotechnology Inc. (Piscataway, NJ).
cDNA library construction and screening.
BAT of cold-exposed rats was used to prepare a cDNA library in the
Uni-Zap XR Vector according to the kit manufacturer’s instructions
(Stratagene Inc., La Jolla, CA). First-strand cDNA synthesis was
primed by the oligo-(dT) linker provided in the kit. Screening of the
library was performed using plaque hybridization under low strin-
gency conditions according to the methods of Lees et al. (21). The
first 305 nucleotides of the coding region of the RC5
as a probe. (This region of the deiodinase proteins is the least con-
served among the three enzyme subtypes. Thus, the use of the 5
tion of the coding region made it less likely that DI and DIII cDNAs
would be identified during the screening process.) Positive plaques
were detected by autoradiography and purified by additional rounds
of screening using the hybridization conditions described above.
cDNA inserts were sequenced on both strands using vector and gene-
specific primers and an automated sequencing system with fluores-
cent dye terminators (Applied Biosystems Inc., Foster City, CA).
Expression studies in X. laevis oocytes and COS-7 cells.
oocytes were isolated and each microinjected as previously de-
scribed (19) with 50 ng of in vitro–synthesized, capped RNA tran-
scripts prepared using the MEGAscript kit (Ambion Inc., Austin,
TX). After injection, oocytes were incubated for 3–4 d in Barth’s me-
dium (for determination of 5-deiodinase [5D] activity) or L-15 me-
dium (for determination of 5
harvested. Membrane fractions were prepared as described previ-
For expression in COS-7 cells, rat and human DII cDNAs and the
G21 rat DI cDNA (kindly provided by Drs. M. Berry and P.R.
Larsen, Brigham and Women’s Hospital, Boston, MA) were sub-
cloned into the pcDNA3 mammalian expression vector (Invitrogen
Corp., San Diego, CA). Cells were grown and maintained in DME
supplemented with 10% iron-supplemented calf serum (Sigma
Chemical Co., St. Louis, MO). Cells (10
RNA was prepared as pre-
C instead of 50
C. In some exper-
/100 g body wt for 4 d
P using the Oligolabeling
DII were used
D] activity), and then
in Hepes-buffered saline)
were transfected with 22
(Bio-Rad Laboratories, Hercules, CA), and then maintained in cul-
ture medium for 48 h before harvesting. After aspiration of the me-
dium, cell monolayers were washed twice with phosphate-buffered
saline, and then scraped from the dish, pelleted, and sonicated in 0.25 M
sucrose, 0.02 M Tris/HCl, pH 7.4.
D and 5D activity were determined in oocyte membrane prepa-
rations or COS-7 cell sonicates according to published methods (17,
22). For the 5
D assay, the reaction buffer contained 1 mM EDTA. In
kinetic studies, 5
D activity was determined using either 0.5–16 nM
or 0.5–12 nM [ I]T
with 20 mM dithiothreitol as cofactor.
Kinetic constants were determined from double reciprocal plots.
labeled iodothyronines used as substrates were obtained from Du
Pont de Nemours (Boston, MA) and purified by chromatography us-
ing Sephadex LH-20 (Sigma Chemical Co.) before use. In other ex-
periments, the deiodinase activities in oocyte membrane preparations
or COS-7 cell sonicates were determined in the absence or presence
of PTU (10–900
M) or aurothioglucose (0.01–10
was measured using 1.5 nM [I]rT
threitol as cofactor.
Protein concentrations were determined by the method of Brad-
ford (23) with reagents obtained from Bio-Rad Laboratories.
Reverse transcriptase-PCR assay.
and PCR amplification was used to determine the presence of DII
transcripts in selected adult rat tissues including BAT from a hy-
pothyroid animal, and liver and white fat from a normal animal. Re-
actions used the Access RT-PCR System (Promega Corp., Madison,
WI) with poly(A) RNA (1
g) as a template. Reaction conditions
were as specified by the manufacturer except that 30 or 35 cycles
were used in the PCR. Specific oligonucleotide primers derived from
the coding region of the rat DII sequence (sense: ACTCGGTCAT-
TCTGCTCAAG; antisense: TTCAAAGGCTACCCCATAAG) were
used to prime first-strand cDNA synthesis, and then the amplification
of a predicted 590-bp PCR product. Reaction mixtures lacking re-
verse transcriptase or an RNA template were used as controls. Prod-
ucts were then separated on a 1.0% agarose gel, transferred to a ny-
lon membrane (Magna Charge Micron Separations, Inc., Westboro,
MA), and hybridized with a radiolabeled nested oligonucleotide rat
DII probe (AATGCCACCTTCTTGACTTT). After washing, the
blot was exposed to x-ray film for 12 h. The signals were quantified
using a phosphoimager (Molecular Dynamics, Inc., Sunnyvale, CA).
Preparation of chimeric DII/DIII cDNAs.
constructed by splicing part or all of the 3
full length rat DIII cDNA (rNS43-1), which contains an active seleno-
cysteine insertion sequence (SECIS element), to the 3
coding regions of the rat BAT 1-1 or human Z44085 DII cDNAs
(Fig. 1). The 3
-untranslated portion of the rNS43-1 was amplified by
PCR using sense primers that included, at their 5
cleotides of the rat or human DII cDNA, and the antisense pBlue-
Script M13 forward primer. Conditions for these PCR reactions var-
ied, but generally included 28 to 35 cycles of amplification involving
C for 1 min, 53 or 57
C for 45 sec, and 72
10-min extension period. PCR reaction products were purified as de-
scribed above for cDNA probe preparation.
For construction of the rBAT 1-1 coding region/rat DIII SECIS
chimera cDNA (Fig. 1
), the rNS43-1 3
amplified with primers 1 and 2, and the PCR product was then di-
gested with XbaI and ApaI. This was then subcloned into the rat
BAT 1-1 DII pcDNA3 construct that had been truncated in the
-untranslated region at a BamHI site and, after relegation, digested
with XbaI and ApaI to remove the 3
DII clone. This construct retained the entire coding region and stop
codon of the rat DII clone, although the stop codon was changed
from TAG to TAA. The 5
-truncated Bat 1-1 clone was used for this
construct because of the possibility that a shorter 5
gion would allow higher levels of expression in
For constructing the hZ44085 coding region/rat DIII SECIS chi-
mera cDNA, convenient restriction sites near the 3
g of plasmid DNA using an Electroporator
as substrate and 20 mM dithio-
Coupled reverse transcription
Chimeric cDNAs were
-untranslated region of the
end of the
ends, 19 or 20 nu-
C for 1.5 min, and a final
-untranslated region was
-untranslated region of the rat
end of the human
Mammalian Type II Iodothyronine Deiodinase
DII coding region were not available, hence an overlap extension
PCR method (24) was used (Fig. 1
man Z44085 DII cDNA was amplified using sense primer 3, whose
sequence was derived from the 5
(with a BamHI restriction site included near the 5
nucleotide), and an antisense primer 4, located in the 3
region at 70 nucleotides 3
to the TAA stop codon. The rNS43-1
-untranslated region for this construct was amplified with primer 5,
that contains hZ44085 sequences at its 5
purification, these PCR products were spliced together in an overlap
extension PCR reaction to yield the full length chimeric hZ44085/rat
DIII SECIS cDNA, which was then digested with BamHI and XhoI
and cloned into the pcDNA3 vector. Conditions of these PCR reac-
tions were the same as those noted above for amplification of the
-untranslated portion of the rNS43-1. Critical regions of the chi-
meric cDNAs were sequenced to insure the accuracy of the construc-
Kinetic constants were determined by linear re-
gression analysis of double reciprocal plots. Results are reported as
SEM. Statistical analysis used Student’s
Bonferroni correction applied for multiple comparisons (25). The
evolutionary tree diagram was produced with the GeneWorks pro-
gram (IntelliGenetics, Mountain View, CA) using the unweighted
pair group method with arithmetic mean.
). The coding region of the hu-
-untranslated region of the clone
end of the oligo-
end, and primer 2. After gel
test with the
Previous studies employing actinomycin D have provided indi-
rect evidence that the adrenergically mediated cold-induced
stimulation of DII activity in rat BAT results from transcrip-
tional activation and an increase in DII mRNA levels (26, 27).
Thus, as an initial step in isolating a cDNA for the rat DII, we
constructed a cDNA library from poly(A)
from the BAT of rats exposed to 4
brary by plaque filter hybridization at reduced stringency with
a 305-bp fragment from the coding region of the RC5
sulted in the isolation of 13 clones containing two different
cDNA inserts. The clone containing the largest cDNA (1.9 kb,
designated rBAT 1-1) was sequenced on both strands (Fig. 2
). rBAT 1-1 contains an open reading frame of 798 nucle-
otides that includes two in-frame TGA codons. The first
(cDNA nucleotide 951, codon 130) is located in a region that
exhibits high homology with the regions in RC5
DIII cDNAs that contain a TGA triplet that codes for seleno-
cysteine. The second TGA triplet (codon 262) is located four
codons upstream from an unambiguous TAG stop codon and
is not present in any other deiodinase cDNA previously iso-
lated. The other, smaller cDNA isolated from the BAT library
was also sequenced and found to be identical to portions of the
BAT 1-1 clone, including the presence of both TGA codons,
thus confirming their presence in the rat DII homolog. The
798-nucleotide open reading frame of the rBAT 1-1 cDNA is
predicted to code for a protein of 266 amino acids of 29.8 kD.
If translation stopped at the second TGA codon, rather than
the downstream TAG, a slightly smaller protein would result.
A search of GenBank revealed a DNA sequence (acces-
C for 24 h. Screening the li-
DII, DI, and
Figure 1. Schematic diagram outlining the strategy for construction of (A) the rBAT 1-1/rDIII SECIS and (B) the hZ44085/rDIII SECIS chi-
meric cDNAs. Sequences of the oligonucleotides used as PCR primers are shown in the lower part of each figure. Primer 2 was used in both
schemes. See text for detailed description of methodology.
W. Croteau, J.C. Davey, V.A. Galton, and D.L. St. Germain
Figure 2. (A) Nucleotide and predicted amino acid
sequence for the rBAT 1-1 DII cDNA. The first in-
frame TGA codon is designated as coding for seleno-
cysteine (SeC), whereas the second TGA codon is
shown by an asterisk since it is uncertain if this triplet
codes for selenocysteine or is read as a termination
signal. An unambiguous TAG stop triplet is located
four codons 3? of the second TGA. (B) Comparison
of the amino acid sequences deduced from the R.
catesbeiana (RC5?DII), rat (rBAT 1-1), and human
(hZ44085) DII cDNAs. X, selenocysteine; x, position
of second TGA codon in rat and human cDNAs.
Mammalian Type II Iodothyronine Deiodinase
sion Z44085) that is highly homologous (89% nucleotide iden-
tity) to a portion of the 5
-untranslated region and the begin-
ning of the coding region of the rBAT 1-1 cDNA. The 320
nucleotides of the hZ44085 sequence were derived from the 5
end of an expressed sequence tag isolated from a cDNA li-
brary prepared from the brain tissue of a 3-mo-old human in-
fant. The high homology suggests that the Z44085 cDNA rep-
resents the human DII homolog. The hZ44085 cDNA was
kindly provided by Genethon (Evry, France), and we se-
quenced it in its entirety. This cDNA is
an open reading frame of 819 nucleotides that codes for a pro-
tein of 30.0 kD (Fig. 2
) that is highly homologous to those
coded by the rBAT 1-1 and RC5
amino acid residue identity, respectively). Overall, 72% of the
amino acid residues have been conserved among these three
proteins. The hZ44085 cDNA contains two in-frame TGA
1.9 kb and contains
DII cDNAs (87 and 73%
codons in locations analogous to those found in rBAT 1-1. As
is the case for the DI and DIII proteins, the amino-terminal
portions of these three proteins contain a region of 42 amino
acid residues that is highly hydrophobic and may represent a
membrane-spanning domain (28). The complete sequences of
the rBAT 1-1 and hZ44085 cDNAs have been submitted to
GenBank/EMBL/DDBJ and have been assigned accession
numbers U53505 and U53506, respectively.
Attempts to demonstrate functional deiodinase activity of
the rBAT 1-1 and hZ44085 proteins by either (a) the injection
of in vitro–synthesized RNA transcripts into X. laevis oocytes,
or (b) transfection of the corresponding cDNAs into COS-7
cells were unsuccessful. Truncating the rBAT 1-1 cDNA of 405
nucleotides of the 5?-untranslated region at a BamHI site did
not result in a cDNA that codes for a functional deiodinase.
Examination of the 3?-untranslated region in both cDNAs sug-
gested that they lack a classic stem-loop SECIS element that is
necessary for the read through and translation of the TGA
codon(s) into selenocysteine. Lacking such elements, one
would predict that attempts at expression of these cDNAs
would result in the formation of truncated, inactive proteins, as
Figure 3. Expression of rat deiodinase cDNAs in X. laevis oocytes.
Individual oocytes were injected with 50 ng of in vitro–synthesized
RNA transcripts prepared using T3 or T7 RNA polymerase and, as
template, either the rat Bat 1-1 DII/rDIII SECIS chimera cDNA, the
rat G21 DI cDNA, or the rat NS43-1 DIII cDNA. (A) 5?-deiodinase
or (B) 5-deiodinase activity was measured in oocyte membrane prep-
arations. Because of high levels of activity, the G21 and NS43-1 mem-
brane preparations were diluted 20-fold before the 5?-deiodinase or
5-deiodinase assays, respectively. For each group of oocytes, both as-
says were conducted on aliquots of the same membrane preparations
using substrate concentrations of ? 1.5 nM rT3 (5?-deiodinase assay)
or 1.0 nM T3 (5-deiodinase assay).
Figure 4. Kinetic analysis using T4 or rT3 as substrates of the protein
products of the (A) rat Bat 1-1 DII/rDIII SECIS chimera and (B) hu-
man Z44085 DII/rDIII SECIS chimera as expressed in COS-7 cells.
Vmax values are expressed in units of activity where 1 U ? 1 pmol/hr
per mg protein.
Figure 5. Sensitivity of the G21 rDI and
the rBAT 1-1 DII/rDIII SECIS chimera
deiodinases to the inhibitory effects of (A)
PTU as determined in oocyte membrane
preparations and (B) aurothioglucose as
defined in COS-7 cell homogenates. Oo-
cytes previously injected with RNA synthe-
sized in vitro using the G21 rDI or rBAT
1-1 DII/rDIII SECIS chimera cDNAs as
templates were harvested, membranes pre-
pared, and then assayed for 5?-D activity in
the presence of PTV. Control incubations
were performed in aliquots of the same
membrane preparations in the absence of
inhibitors. For studies examining the ef-
fects of aurothioglucose, COS-7 cells were
transfected with either the G21 or rBAT 1-1
DII/rDIII SECIS chimera and assays per-
formed in the absence or presence of the
inhibitor in cell sonicates. Entirely analo-
gous results were obtained with the
hZ44085/DIII SECIS chimera.
W. Croteau, J.C. Davey, V.A. Galton, and D.L. St. Germain
has been demonstrated for other deiodinase cDNAs (7, 9).
Thus, chimeric constructs were prepared by replacing part or
all of the 3?-untranslated regions of the rBAT 1-1 and hZ44085
cDNAs, respectively, with the 3?-untranslated region of the
rNS43-1 that contains a potent SECIS element (Moyer, B., and
D.L. St. Germain, unpublished observations). In both chimeric
constructs, the coding regions of the rBAT 1-1 and hZ44085
cDNAs remained intact.
Figure 6. Expression of DII
transcripts in rat tissues (A–C)
by Northern analysis of
poly(A)? RNA (6 ?g/lane)
from several tissues of euthy-
roid rats. A portion of the cod-
ing region between nucleotides
615–1238 of the rBAT 1-1
cDNA was used as a probe. The
final wash step was for 1 h at
42?C in 0.1 ? SSC, 0.1% SDS.
Blots were then subjected to
autoradiography for 1 wk.
Identical patterns of hybridiza-
tion were noted after a 1-h
wash in the same solution at
60?C; however, the intensity of
the signal was generally some-
what diminished. When using
the actin probe, blots were
washed in the same solution at
60?C and autoradiographed
overnight. (D) Coupled re-
verse transcriptase/PCR assay
using poly(A)? RNA from
liver, white fat, and hypothy-
roid BAT as template. For
each sample, reactions were
performed with (?RT) and
without (?RT) reverse tran-
scriptase in the mixture. Also,
a mixture that did not include
RNA (no template) was run as
an additional negative control.
PCR products were separated
on an agarose gel, transferred
to a nylon membrane, and then
probed with a radiolabeled,
nested rat DII oligonucleotide.
Mammalian Type II Iodothyronine Deiodinase
As shown in Fig. 3, the injection into oocytes of RNA tran-
scripts derived from the rBAT 1-1/rDIII SECIS chimera
cDNA induced 5?-, but not 5-deiodinase activity. In the same
experiment, the G21 and the rNS43-1 cDNAs induced high
levels of 5?- or 5-deiodinase activity, respectively.
When examined in rat tissue homogenates, the DII exhib-
its several characteristic properties including: (a) Km values in
the nanomolar range for iodothyronine substrates; (b) the
ability to efficiently 5?-deiodinate T4 as well as rT3; and (c) rel-
ative insensitivity to the inhibitory effects of PTU and gold
compounds such as aurothioglucose (29). Using the oocyte
and COS-7 cell expression systems, the functional properties
of the proteins coded by the rat and human chimeric cDNA
constructs were investigated. Kinetic analyses performed in
sonicates of transfected COS-7 cells are shown in Fig. 4. Both
enzymes showed saturable reaction kinetics with Km values
for T4 and rT3 of ? 1 and 5 nM, respectively. Vmax values using
rT3 as substrate were 75 and 33% higher for the rat and hu-
Figure 7. Expression of
DII transcripts in human
tissues. (A–D) Northern
analysis of poly(A)?
RNA (2 ?g/lane) derived
from several human tis-
sues using a portion of
the coding region of the
hZ44084 cDNA (nucle-
otides 220–814) as probe.
Final wash conditions
were the same as those
described in Fig. 6, and
the blots were autoradio-
graphed for 5–14 d.
Washing at 60?C did not
affect the patterns of hy-
bridization in any of the
blots. After stripping,
blots were reprobed
with a human actin
W. Croteau, J.C. Davey, V.A. Galton, and D.L. St. Germain
man enzymes, respectively, when compared to values ob-
tained with T4.
The rBAT 1-1 deiodinase proved to be insensitive to inhi-
bition by PTU and aurothioglucose. PTU, 100 ?M, had no ef-
fect on 5?-deiodinase activity, whereas DI activity induced by
the expression of the G21 cDNA was inhibited by ? 90% (Fig.
5 A). The hZ44085 protein also proved completely insensitive
to PTU at concentrations as high as 900 ?M (data not shown).
Although aurothioglucose did inhibit both the rBAT 1-1
and hZ44085 enzyme activity, a much higher concentration
(? 10-fold) was required to achieve inhibition comparable to
that observed with the DI (Fig. 5 B, and data not shown for the
The expression of rBAT 1-1–associated mRNA was inves-
tigated in rat tissues by Northern analysis and coupled reverse
transcription PCR. A relatively weak hybridization band of 7.5
kb was noted in RNA from normal BAT (Fig. 6 A). The abun-
dance of this species was increased 17-fold in the sample of
RNA from BAT of cold-exposed animals. (In a second experi-
ment, a ninefold increase was noted, data not shown.) In addi-
tion, smaller and much less abundant species of ? 4.0, 2.5, and
2.0 kb were also noted in this and other RNA samples from
Figure 8. Effect of altered thyroid hormone
status on DII (A) activity and (B) mRNA lev-
els in the BAT, anterior pituitary gland, cere-
bral cortex, and cerebellum of rats. Rats were
rendered hypo- and hyperthyroid as described
in Methods. For determination of DII activity,
individual tissues were harvested from three
rats, homogenized, and assayed in the pres-
ence of 0.1 mM PTU using 1.5 nM rT3 as sub-
strate and 20 mM dithiothreitol as cofactor.
Results represent the mean?SEM. *P ? 0.05
vs euthyroid value. For Northern blot analysis,
RNA was prepared from BAT and cerebral
cortex of individual animals, from a pool of
three cerebellums, and from a pool of 10 ante-
rior pituitary glands. Approximately 10 ?g
(brown fat, cerebral cortex, and cerebellum) or
5 ?g (pituitary) of poly(A)? RNA were loaded
per lane on the gel.
Mammalian Type II Iodothyronine Deiodinase
this tissue. No hybridization signal was seen with RNA sam-
ples from rat heart, white fat, placenta, liver, kidney, neonatal
skin, or skeletal muscle (Fig. 6, B and C). In adult rat cerebral
cortex, two major hybridizing species of approximately equal
abundance at 7–8 kb were observed (Figs. 6 B and 8 B, and be-
low). A similar pattern was seen in the cerebellum, though the
smaller species of the doublet appears somewhat more abun-
dant in this tissue.
Although no hybridization signal was detected on North-
ern blots using RNA from rat liver and white fat (Fig. 6 C), the
lesser ?-actin control signal in these RNA samples suggested
that this analysis was less than definitive. Hence, a coupled re-
verse transcription-PCR technique was used to examine
whether DII transcripts were expressed in these tissues. As
shown in Fig. 6 D, an abundant PCR product of the expected
size (590 bp) was observed when RNA from BAT of a hy-
pothyroid rat was used in the assay (positive control). This sig-
nal was markedly diminished in the absence of reverse tran-
scriptase and was absent when the template was excluded from
the reaction mixture. A faint signal, only 1/100th the intensity
of that observed in BAT, was noted in white fat, whereas no
signal was observed in this experiment when liver RNA was
used as the template. In another experiment using a higher cy-
cle number, a faint band was detected in liver (data not
shown). Such a finding is likely of little physiologic significance
and may correspond to the phenomenon of “illegitimate tran-
scription” as defined by Chelly et al. (30).
The hZ44085 cDNA was used to probe RNA blots of hu-
man tissues. High levels of expression of a 6.6-kb transcript
were noted in brain, skeletal muscle, and heart, with lesser
amounts of a larger 7.5-kb species present in placenta and
heart (Fig. 7 A). No signal was detectable in pancreas, kidney,
liver, or lung. The distribution of hZ44085-associated tran-
scripts in human brain was investigated further (Fig. 7, B and
C). Again the predominant species observed was ? 6.6 kb with
transcripts of 4.4, 4.0, 3.2, and 1.9 kb also observed in some tis-
sues. Transcripts appeared abundant throughout the neocor-
tex as well as the putamen, amygdala, and hippocampus.
Lesser degrees of expression were noted in the cerebellum,
caudate nucleus, substantia nigra, thalamus, subthalamic nu-
cleus, corpus callosum, medulla, and spinal cord.
Because of the known importance of thyroid hormones for
brain development, an RNA blot of human fetal tissues (18–24
wk gestation according to the blot manufacturer, Clontech)
was also probed with hZ44085 (Fig. 7 D). High levels of ex-
pression of a 6.6-kb species were detected in fetal brain, but
not in fetal lung, liver, or kidney.
Previous studies have demonstrated that the regulation of
DII activity is complex and appears to involve both pre- and
posttranslational mechanisms in different tissues (29, 31). In
particular, thyroid hormones exert important regulatory ef-
fects on DII activity as well as on the other deiodinases. Thus,
an investigation of the effects of altered thyroid hormone sta-
tus on the BAT 1-1–associated transcripts was of interest. For
this study, DII activity was determined and RNA extracted
from several tissues of normal (euthyroid), hypothyroid, and
hyperthyroid adult rats. DII activity levels are shown in Fig. 8
A. As compared to the euthyroid state, hypothyroidism was
associated with a significant increase in activity of 6-, 9-, 11-,
and 118-fold in the cerebellum, cerebral cortex, anterior pitu-
itary, and BAT, respectively. In addition, hyperthyroidism in-
duced by T3 administration was associated with a trend toward
higher DII activity in all tissues, and this was statistically signif-
icant in BAT (11-fold increase in activity) and cerebral cortex
The Northern analysis of RNA samples from these rats is
shown in Fig. 8 B. Comparison of the hybridization signals in
the various tissues from the euthyroid animals demonstrates
that rBAT 1-1–associated transcripts are most abundant in the
anterior pituitary gland. Thus, the DII transcript levels in the
cerebellum, cerebral cortex, and pituitary, when normalized for
the actin mRNA signal in each tissue, are ? 7-, 30-, and 50-fold
higher, respectively, than those observed in euthyroid BAT. In
pituitary and BAT RNA samples, hypothyroidism was associ-
ated with a marked increase (3.3- and 12-fold, respectively) in
DII transcripts when compared to the levels in euthyroid sam-
ples. Furthermore, hyperthyroidism resulted in a 70% de-
crease in pituitary DII mRNA relative to the control samples.
In contrast, altered thyroid status had little or no effect on the
DII transcript levels in the cerebral cortex and cerebellum,
where again two species between 7 and 8 kb were observed.
The present studies demonstrate that the rBAT 1-1 and hu-
man hZ44085 code for mammalian DIIs. This conclusion is
based on both the functional activity of the expressed proteins
and the tissue patterns of expression of their associated mRNAs.
Both the rBAT 1-1 and hZ44085 cDNAs appear to lack
functional SECIS elements in their 3?-untranslated regions.
(Although it is theoretically possible that the SECIS element
could be located in a missing portion of the 5?-untranslated re-
gion of these cDNAs, stem loop structures function very ineffi-
ciently as SECIS elements in this location, reference 32, and no
native 5?-untranslated region SECIS elements have been de-
scribed.) This explains the inability of these cDNAs to code for
proteins with functional deiodinase activities; without an ac-
tive SECIS element, translation likely terminates at the first
TGA triplet located midway through the coding regions of
these genes. However, by fashioning chimeric cDNAs in which
the open reading frames of the rBAT 1-1 or the hZ44085 were
fused to the SECIS-containing 3?-untranslated region of the
rNS43-1 rat DIII cDNA, expression of functional deiodinases
The properties of these expressed enzymes are entirely
consistent with those of the endogenous DII previously de-
fined in mammalian tissue homogenates (33). Thus, both en-
zymes manifest Km values for T4 and rT3 in the low nanomolar
range when using 20 mM dithiothreitol as cofactor. Further-
more, using the Vmax/Km ratio as an approximate indicator of
catalytic efficiency, T4 is the preferred substrate for both en-
zymes; values of this calculated parameter for T4 are three- to
fourfold higher than for rT3. This finding contrasts sharply
with the catalytic properties of the DI where rT3 is the much
preferred substrate, and T4 is converted to T3 with relatively
poor efficiency (33). Also contrasting with the DI is the
marked insensitivity of these expressed deiodinases to the in-
hibitory effects of PTU and gold compounds.
Several studies have provided indirect evidence that the
mammalian DII is not a selenoprotein. Such evidence includes
(a) the observation that DII activity in rat astrocytes is not di-
minished when the cells are cultured in selenium deficient me-
dium (34); (b) the inability to label candidate DII proteins with
[75Se] (34); and (c) the insensitivity of the DII to PTU and gold
W. Croteau, J.C. Davey, V.A. Galton, and D.L. St. Germain
inhibitors as noted above (35). Furthermore, although DII ac-
tivity is modestly diminished in the brain and anterior pituitary
gland of selenium-deprived rats (36–38), Chanoine et al. (38)
have provided convincing evidence that this results from an in-
creased rate of posttranslational inactivation of the enzyme
secondary to elevations in serum, and presumably tissue, T4
levels rather than a selenium-induced decrease in enzyme syn-
The present study, however, demonstrates that the rat and
human DIIs are selenoproteins that contain a selenocysteine
within a region that is highly conserved with the other deiodi-
nases (Fig. 9 A). This observation, along with prior site-
directed mutagenesis studies (7, 9, 10, 18), demonstrates that
this rare amino acid is essential for efficient catalysis of iodo-
thyronines by all members of this enzyme family. The inability
of Safran et al. (34) to affect DII activity in cultured rat astro-
glial cells by the manipulation of selenium concentrations in
the medium may be related to the findings of others that the
brain and pituitary gland, like the thyroid gland and other en-
docrine organs, are resistant to selenium depletion (39). That
this property is maintained in vitro in primary cell culture was
recently demonstrated by Beech et al. (40), who were unable
to affect the activity of the DI in cultured human thyroid cells
by selenium starvation. Furthermore, in whole animal studies,
they observed that thyroidal DI activity was actually increased
by 42% in rats on a selenium-deficient diet in spite of a modest
(30%) decrease in thyroidal glutathione peroxidase activity,
another selenoenzyme. Such results confirm the findings of
others (41) indicating that there is a hierarchy of expression of
different selenoproteins in a cell in the face of selenium depri-
Alternatively, the DII in cultured rat glial cells could lack
selenocysteine and be coded by a mRNA that is unrelated to
the rBAT 1-1. However, our observations that rBAT 1-1– and
hZ44085-associated transcripts are highly expressed in the rat
anterior pituitary gland and in the central nervous system of
fetal and adult rats and humans strongly suggest that brain DII
is coded by these genes. Furthermore, the predicted size of
the rBAT 1-1 protein (29.8 kD) is virtually identical to a 29-kD
protein identified by affinity labeling techniques in cultured glial
cells and proposed by Safran and Leonard to be the DII (42).
In addition to the highly conserved regions surrounding the
selenocysteine residue, other shared structural features of the
deiodinases are apparent (Fig. 9 A). All three subclasses of
deiodinase contain regions of strong hydrophobicity near their
amino termini and this may represent membrane-spanning do-
mains (28). Furthermore, two histidine residues demonstrated
by Berry (43) to be essential for catalysis by the DI (residues
158 and 174 in the rat DI) are conserved in the DII and the
DIII enzymes. The second in-frame TGA triplet is unique to
the rat and human DII cDNAs. Whether these code for a sec-
ond selenocysteine in these proteins or function in their more
Figure 9. (A) A comparison of the amino acid se-
quence of the rat DI, DII, and DIII proteins as de-
duced from the cDNA nucleotide sequence. Asterisks
represent amino acid residues that have been con-
served identically between these three proteins. X, se-
lenocysteine; x, position of second TGA codon in rat
DII. (B) Evolutionary tree diagram based on the de-
duced amino acid sequences of 11 deiodinase cDNAs
reported to date (7–10, 18, 71–75).
Mammalian Type II Iodothyronine Deiodinase
traditional role as stop codons is uncertain. It appears likely,
however, that a second selenocysteine is not required for DII
activity since this second TGA codon is not present in the
RC5?DII cDNA (Fig. 2 B). A comparison of the amino acid
sequence of the 11 deiodinases for which cDNAs have been
isolated to date is shown as an evolutionary tree diagram in
Fig. 9 B. This analysis demonstrates that the DIs, DIIs, and
DIIIs form three separate subfamilies, each approximately
equally dissimilar to the other two. Furthermore, based on
these limited data, the members within the DII subfamily ap-
pear more highly conserved across species lines than the DI or
The tissue expression of rBAT 1-1–associated transcripts is
consistent with the known distribution of DII activity in the rat
(33); hybridizing RNA species were observed in the anterior
pituitary gland, cerebral cortex, cerebellum, and BAT. In BAT,
where cold exposure induces DII activity (26), a 9–17-fold in-
crease in abundance of a 7.5-kb band was noted in response to
this stimulus. The large size of the predominant DII transcript
observed in both the rat and human was unexpected given the
much smaller size (? 1.6–2.2 kb) of the DI (44) and DIII mRNAs
(9, 10), and our prior observation that in Rana catesbeiana the
predominant DII mRNA is 1.5 kb (18). Thus, the rBAT 1-1
and hZ44085 cDNAs, both ? 1.9 kb, are not full length with re-
spect to the 6–8-kb DII mRNAs. This likely explains the lack
of SECIS elements in their presumably truncated 3?-untrans-
lated regions. Both cDNAs, however, do contain the entire
coding region for the DII proteins. The finding that the rBAT
1-1 cDNA contains a long poly(A)? tail at its 3? end suggests
that it is derived from one of the smaller, less abundant RNA
species observed on Northern blots of RNA from BAT of
cold-exposed animals. Thus, either alternative RNA splicing
or the use of alternative polyadenylation sites may be a feature
of the processing of the mammalian DII mRNA.
Information on the distribution of the deiodinases in hu-
mans is limited (8, 45, 46). The present studies demonstrate
that DII transcripts are more widely distributed in this species
than in the rat. Thus, in addition to its expression in the central
nervous systems, DII mRNA was also observed in the human
placenta, heart, and skeletal muscle at levels comparable to
those observed in the brain. In contrast, no hybridizing species
could be detected with RNA from these nonneural tissues in
rats, consistent with earlier reports that rat heart (47, 48) and
skeletal muscle (49) contain only low levels of DI activity and
no detectable DII activity. Rat placenta, however, does con-
tain DII activity (46). Thus, our failure to detect DII mRNA
by Northern analysis in rat placenta likely results from the rel-
atively low abundance of this transcript.
The pattern of DII expression in the human fetus (Fig. 7 D)
is of considerable interest given the prior report by Bernal and
Pekonen (50) that T3 and nuclear T3 receptors are detectable
in human fetal brain as early as 10 wk of gestation, long before
T3 becomes detectable in the serum (51). Thus, DII expression
in the central nervous system early in gestation is likely of
great importance for the generation of the T3 needed by this
tissue. Of note, tissues such as the liver and lung, which do not
express DII at this early stage of development, do not contain
detectable T3 (50), thus reinforcing the importance of this en-
zyme in regulating local T3 tissue concentrations.
Northern analyses revealed that the expression of DII tran-
scripts in the human adult brain is widespread and shows a
general rostral–caudal distribution. This regional pattern of
DII expression correlates closely with the relative abundance
of thyroid hormone nuclear receptors in various regions of the
adult rat central nervous system as defined by both saturation
analysis (52) and T3 receptor mRNA expression (53). Al-
though the importance of thyroid hormone in the developing
brain is unquestioned (54, 55), there continues to be contro-
versy concerning the role of these hormones in adult brain
function (56). Thus, the apparent coexpression of the DII and
thyroid hormone receptors in several regions of the adult hu-
man brain provides strong evidence for a continuing important
role of T3 after the developmental period and reinforces clini-
cal observations of altered central nervous system function in
states of hypo- and hyperthyroidism (56–58). Of particular
note are the relatively high levels of expression of DII tran-
scripts in the hippocampus, amygdala, and basal ganglion, sug-
gesting an important influence of thyroid hormones on mem-
ory and learned patterns of movement (59). Such effects might
explain some of the cognitive dysfunction seen in adults with
altered thyroid function as well as the movement disorders ob-
served in neurologic cretinism (60).
To our knowledge, adult human heart and skeletal muscle
have not been examined for the presence of DII activity. How-
ever, the expression of DII in these tissues, as suggested by the
results of the Northern analysis, could be of considerable phys-
iologic importance. For example, LoPresti et al. (61) have
demonstrated that the serum T3 level in euthyroid subjects is
unaltered after 4 d by the administration of high doses of PTU,
suggesting an important role for the DII in maintaining circu-
lating T3 concentrations in healthy humans. DII expression in
skeletal muscle could provide a source for such T3 production.
In addition, several investigators have demonstrated that in
rats, DII-containing tissues (predominantly the cerebral cortex
and BAT) are able to maintain relatively normal tissue T3 lev-
els, and hence presumably a euthyroid state, over a wide range
of plasma T4 levels (62–64). Thus, human heart and skeletal
muscle may also be relatively protected from the effects of hy-
DII activity in rat tissues is markedly influenced by thyroid
hormone status. For example, hypothyroidism is associated
with marked elevations in DII activity (65, 66). In the central
nervous system and anterior pituitary gland, posttranslational
processes are important in mediating this effect (67, 68),
whereas in BAT, multiple pre- and posttranslational mecha-
nisms appear to influence DII activity in hypo- and hyperthy-
roid states (27, 31). The cloning of the rBAT 1-1 cDNA al-
lowed us to investigate further these complex processes by
focusing specifically on pretranslational mechanisms of regula-
tion (Fig. 8). As expected from previous studies, the induction
of hypothyroidism resulted in a marked increase in DII activity
in all tissues studied. This was associated with a significant in-
crease in DII mRNA expression in BAT and the pituitary, but
little if any change in the cerebral cortex and cerebellum. It is
thus apparent that in the brain, thyroid hormone regulates DII
activity predominantly, if not exclusively, via posttranslational
mechanisms as previously described (67). In BAT, Silva and
Larsen (31) have demonstrated that the growth hormone defi-
ciency that accompanies hypothyroidism stimulates DII activ-
ity, an effect that is blocked by actinomycin. Thus, the alterations
in DII transcript levels in this tissue are likely due at least in
part to a secondary effect of the thyroid hormone deficiency.
The significant increase in DII transcripts in the pituitary of
hypothyroid animals demonstrates that pretranslational fac-
W. Croteau, J.C. Davey, V.A. Galton, and D.L. St. Germain
tors, in addition to the previously cited posttranslational ef-
fects of thyroid hormones, are responsible in part for the
marked rise in DII activity in this tissue. Whether this increase
represents a direct effect of thyroid hormone on DII mRNA
levels is uncertain, but it seems plausible given that, unlike
BAT, catecholamines and growth hormone have no demon-
strable effect on pituitary DII activity (26, 31). A direct effect
is also consistent with a recent study by Halperin et al. (69)
who noted that T3 was more potent than rT3 in downregulating
DII activity in cultured rat pituitary GC cells, an effect that
they suggest is mediated by the nuclear thyroid hormone re-
ceptor. Alternatively, the increase in DII transcripts in hy-
pothyroidism could be secondary to a relative increase in
abundance of a pituitary cell type that expresses comparatively
high DII levels. Indeed, the fraction of thyrotrophs has been
observed to increase from 10.7% in the euthyroid rat pituitary
gland to 34.4% in chronic hypothyroidism (70). Thyrotrophs,
however, appear to contain less DII activity than other pitu-
itary cell types (71). Thus an increase in the proportion of thy-
rotrophs is an unlikely explanation for our findings.
In our experiment, hyperthyroidism, as induced by 4 d of
T3 injections, was also associated with an increase in DII activ-
ity, though to a considerably lesser extent than that seen with
hypothyroidism. Although not proven, this likely results pri-
marily from thyroid-stimulating hormone suppression, leading
to reduced serum and tissue T4 levels, and hence decreased
posttranslational inactivation of the DII protein (67). Such a
phenomenon has been noted previously in BAT (31), but has
not been investigated in other tissues expressing DII.
In summary, we have identified cDNAs for the rat and hu-
man DII and have demonstrated that they code for selenopro-
teins that manifest critical areas of homology with the other
deiodinase subtypes. The expression of the DII in mammalian
brain, pituitary, BAT, heart, skeletal muscle, and placenta
highlights the importance of thyroid hormone in regulating
metabolic activity in these tissues, and lends further support to
the concept that the DII is a critical modulator of thyroid hor-
The authors thank Pascal Soularve (Genethon) for generously pro-
viding the Z44085 cDNA.
This study was supported by the National Institutes of Health
grants DK-42271 (D.L. St. Germain), HD-27706 (V.A. Galton), and
HD-09020 (V.A. Galton) and the Norris Cotton Cancer Center Core
grant CA 23108.
1. Crantz, F.R., J.E. Silva, and P.R. Larsen. 1982. An analysis of the sources
and quantity of 3,5,3?-triiodothyronine specifically bound to nuclear receptors
in rat cerebral cortex and cerebellum. Endocrinology. 110:367–375.
2. Silva, J.E., and P. Matthews. 1984. Thyroid hormone metabolism and the
source of plasma triiodothyronine in 2-week-old rats: effects of thyroid status.
3. Bianco, A.C., and J.E. Silva. 1987. Nuclear 3,5,3?-triiodothyronine (T3) in
brown adipose tissue: receptor occupancy and sources of T3 as determined by
in vivo techniques. Endocrinology. 120:55–62.
4. van Doorn, J., F. Roelfsema, and D. van der Heide. 1985. Concentration
of thyroxine and 3,5,3?-triiodothyronine at 34 different sites in euthyroid rats as
determined by isotopic equilibrium techniques. Endocrinology. 117:1201–1208.
5. Silva, J.E., and P.R. Larsen. 1986. Regulation of thyroid hormone expres-
sion at the prereceptor and receptor levels. In Thyroid Hormone Metabolism.
G. Hennemann, editor. Marcel Dekker, Inc., New York. 441–500.
6. Silva, J.E., M.B. Gordon, F.R. Crantz, J.L. Leonard, and P.L. Larsen.
1984. Qualitative and quantitative differences in the pathways of extrathyroidal
triiodothyronine generation between euthyroid and hypothyroid rats. J. Clin.
7. Berry, M.J., L. Banu, and P.R. Larsen. 1991. Type I iodothyronine deio-
dinase is a selenocysteine-containing enzyme. Nature (Lond.). 349:438–440.
8. Mandel, S.J., M.J. Berry, J.D. Kieffer, J.W. Harney, R.L. Warne, and
P.R. Larsen. 1992. Cloning and in vitro expression of the human selenoprotein,
type I iodothyronine deiodinase. J. Clin. Endocrinol. & Metab. 75:1133–1139.
9. St. Germain, D.L., R. Schwartzman, W. Croteau, A. Kanamori, Z. Wang,
D.D. Brown, and V.A. Galton. 1994. A thyroid hormone regulated gene in Xe-
nopus laevis encodes a type III iodothyronine 5-deiodinase. Proc. Natl. Acad.
Sci. USA. 91:7767–7771. Correction. Proc. Natl. Acad. Sci. USA. 91:11282.
10. Croteau, W., S.L. Whittemore, M.J. Schneider, and D.L. St. Germain.
1995. Cloning and expression of a cDNA for a mammalian type III iodothyro-
nine deiodinase. J. Biol. Chem. 270:16569–16575.
11. Beech, S.G., S.W. Walker, A.M. Dorrance, J.R. Arthur, J.R. Nicol, D.
Lee, and G.J. Beckett. 1993. The role of type-I iodothyronine deiodinase in tri-
iodothyronine production by human and sheep thyrocytes in culture. J. Endo-
12. Chanoine, J., L.E. Braverman, A.P. Farwell, M. Safran, S. Alex, S.
Dubord, and J.L. Leonard. 1993. The thyroid gland is a major source of circu-
lating T3 in the rat. J. Clin. Invest. 91:2709–2713.
13. Moreno, M., M.J. Berry, C. Horst, R. Thoma, F. Goglia, J.W. Harney,
P.R. Larsen, and T.J. Visser. 1994. Activation and inactivation of thyroid hor-
mone by type I iodothyronine deiodinase. FEBS Lett. 344:143–146.
14. St. Germain, D.L. 1994. Biochemical study of type III iodothyronine
deiodinase. In Thyroid Hormone Metabolism: Molecular Biology and Alterna-
tive Pathways. S.-Y. Wu, and T. J. Visser, editors. CRC Press, Ann Arbor, MI.
15. McCann, U.D., E.A. Shaw, and M.M. Kaplan. 1984. Iodothyronine
deiodination reaction types in several rat tissues: effects of age, thyroid status,
and glucocorticoid treatment. Endocrinology. 114:1513–1521.
16. Calvo, R., M.J. Obregón, C. Ruiz de Ona, F. Escobar del Rey, and G.M.
de Escobar. 1990. Congenital hypothyroidism, as studied in rats: crucial role of
maternal thyroxine but not 3,5,3?-triiodothyronine in the protection of the fetal
brain. J. Clin. Invest. 86:889–899.
17. St. Germain, D.L., and W. Croteau. 1989. Expression of phenolic and
tyrosyl ring iodothyronine deiodinases in Xenopus laevis oocytes is dependent
on the tissue source of injected poly(A)? RNA. Mol. Endocrinol. 3:2049–2053.
18. Davey, J.C., K.B. Becker, M.J. Schneider, D.L. St. Germain, and V.A.
Galton. 1995. Cloning of a cDNA for the type II iodothyronine deiodinase. J.
Biol. Chem. 270:26786–26789.
19. St. Germain, D.L., and C.M. Morganelli. 1989. Expression of type I io-
dothyronine 5?-deiodinase in Xenopus laevis oocytes. J. Biol. Chem. 264:3054–
20. DePalo, D., W.B. Kinlaw, C. Zhao, H. Engelberg-Kulka, and D.L. St.
Germain. 1994. Effect of selenium deficiency on type I 5?-deiodinase. J. Biol.
21. Lees, J.A., M. Saito, M. Vidal, M. Valentine, T. Look, E. Harlow, N.
Dyson, and K. Helin. 1993. The retinoblastoma protein binds to a family of E2F
transcription factors. Mol. Cell. Biol. 13:7813–7825.
22. Sharifi, J., and D.L. St. Germain. 1992. The cDNA for the type I iodo-
thyronine 5?-deiodinase encodes an enzyme manifesting both high Km and low
Km activity. J. Biol. Chem. 267:12539–12544.
23. Bradford, M.M. 1976. A rapid and sensitive method for quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding.
Anal. Biochem. 72:248–254.
24. Horton, R.M., S.N. Ho, J.K. Pullen, H.D. Hunt, Z. Cai, and L.R. Pease.
1993. Gene splicing by overlap extension. Methods Enzymol. 217:270–278.
25. Godfrey, K. 1985. Comparing the means of several groups. N. Engl. J.
26. Silva, J.E., and P.R. Larsen. 1983. Adrenergic activation of triiodothyro-
nine production in brown adipose tissue. Nature (Lond.). 305:712–713.
27. Silva, J.E., and P.R. Larsen. 1986. Hormonal regulation of iodothyro-
nine 5?-deiodinase in rat brown adipose tissue. Am. J. Physiol. 251:639–643.
28. Toyoda, N., M.J. Berry, J.W. Harney, and P.R. Larsen. 1995. Topologi-
cal analysis of the integral membrane protein, type 1 iodothyronine deiodinase
(D1). J. Biol. Chem. 270:12310–12318.
29. St. Germain, D.L. 1994. Iodothyronine deiodinases. Trends Endocrinol.
30. Chelly, J., J.-P. Concordet, J.-C. Kaplan, and A. Kahn. 1989. Illegitimate
transcription: transcription of any gene in any cell type. Proc. Natl. Acad. Sci.
31. Silva, J.E., and P.R. Larsen. 1986. Interrelationships among thyroxine,
growth hormone, and the sympathetic nervous system in the regulation of 5?-iodo-
thyronine deiodinase in rat brown adipose tissue. J. Clin. Invest. 77:1214–1223.
32. Berry, M.J., L. Banu, J.W. Harney, and P.R. Larsen. 1993. Functional
characterization of the eukaryotic SECIS elements which direct selenocysteine
insertion at UGA codons. EMBO (Eur. Mol. Biol. Organ.) J. 12:3315–3322.
33. Leonard, J.L. 1991. Biochemical basis of thyroid hormone deiodination.
In Thyroid Hormone Metabolism: Regulation and Clinical Implications. S. Wu,
editor. Blackwell Scientific Publications, Boston, MA. 1–28.
Mammalian Type II Iodothyronine Deiodinase
34. Safran, M., A.P. Farwell, and J.L. Leonard. 1991. Evidence that type II
5?-deiodinase is not a selenoprotein. J. Biol. Chem. 266:13477–13480.
35. Berry, M.J., J.D. Kieffer, and P.R. Larsen. 1991. Evidence that cysteine,
not selenocysteine is in the catalytic site of type II iodothyronine deiodinase.
36. Beckett, G.J., D.A. MacDougal, F. Nicol, and J.R. Arthur. 1989. Inhibi-
tion of types I and II iodothyronine deiodinase activity in rat liver, kidney and
brain produced by selenium deficiency. Biochem. J. 259:887–892.
37. Meinhold, H., A. Campos-Barros, B. Walzog, R. Köhler, F. Müller, and
D. Behne. 1993. Effects of selenium and iodine deficiency on type I, type II, and
type III iodothyronine deiodinases and circulating thyroid hormones in the rat.
Exp. Clin. Endocrinol. 101:87–93.
38. Chanoine, J.P., M. Safran, A.P. Farwell, P. Tranter, D.M. Ekenbarger,
S. Dubord, S. Alex, J.R. Arthur, G.J. Becker, L.E. Braverman, and J.L. Le-
onard. 1992. Selenium deficiency and type II 5?deiodinase regulation in the eu-
thyroid and hypothyroid rat: evidence of a direct effect of thyroxine. Endocri-
39. Behne, D., H. Hilmert, S. Scheid, H. Gessner, and W. Elger. 1988. Evi-
dence for specific selenium target tissues and new biologically important sele-
noproteins. Biochim. Biophys. Acta. 966:12–21.
40. Beech, S.G., S.W. Walker, G.J. Beckett, J.R. Arthur, F. Nicol, and D.
Lee. 1995. Effect of selenium depletion on thyroidal type-I iodothyronine deio-
dinase activity in isolated human thyrocytes and rat thyroid and liver. Analyst.
41. Gross, M., M. Oertel, and J. Köhrle. 1995. Differential selenium-depen-
dent expression of type I 5?-deiodinase and glutathione peroxidase in the por-
cine epithelial kidney cell line LLC-PK1. Biochem. J. 306:851–856.
42. Safran, M., and J.L. Leonard. 1991. Comparison of the physiochemical
properties of type I and type II iodothyronine 5?-deiodinase. J. Biol. Chem. 266:
43. Berry, M. 1992. Identification of essential histidine residues in rat type I
iodothyronine deiodinase. J. Biol. Chem. 267:18055–18059.
44. St. Germain, D.L., W. Dittrich, C.M. Morganelli, and V. Cryns. 1990.
Molecular cloning by hybrid arrest of translation in Xenopus laevis oocytes:
identification of a cDNA encoding the type I iodothyronine 5?-deiodinase from
rat liver. J. Biol. Chem. 265:20087–20090.
45. Kaplan, M.M., C.Y. Pan, P.R. Gordon, and B.A. Gilchrest. 1987. Hu-
man epidermal keratinocytes convert thyroxine to triiodothyronine by type II
deiodination. Clin. Res. 35:388(Abstr.)
46. Kaplan, M.M., and E.A. Shaw. 1984. Type II iodothyronine 5?-deiodina-
tion in human and rat placenta in vitro. J. Clin. Endocrinol. & Metab. 59:253–
47. Mori, Y., M. Nishikawa, N. Toyoda, T. Yonemoto, H. Matsubara, and
M. Inada. 1991. Iodothyronine 5?-deiodinase activity in cultured rat myocardial
cells: characteristics and effects of triiodothyronine and angiotensin II. Endocri-
48. Nauman, A., J. Nauman, S. Porta, and W. Grzesiuk. 1994. Iodothyro-
nine-5?-deiodinase in rat heart, regulatory action of catecholamines. In Heart
and Thyroid. L. E. Braverman, O. Eber, and W. Langsteger, editors. Blackwell
Scientific Publications Inc., Cambridge, MA. 169–172.
49. Tsukahara, F., T. Nomoto, and M. Maeda. 1989. Properties of 5?-deiodi-
nase of 3,3?,5?-triiodothyronine in rat skeletal muscle. Acta Endocrinol. 120:69–
50. Bernal, J., and F. Pekonen. 1984. Ontogenesis of the nuclear 3,5,3?-tri-
iodothyronine receptor in the human fetal brain. Endocrinology. 114:677–679.
51. Abuid, J., A.H. Klein, T.P.J. Foley, and P.R. Larsen. 1974. Total and
free triiodothyronine and thyroxine in early infancy. J. Clin. Endocrin. & Me-
52. Gullo, D., A.K. Sinha, A. Bashir, M. Hubank, and R.P. Ekins. 1987. Dif-
ferences in nuclear triiodothyronine binding in rat brain cells suggest phyloge-
netic specialization of neuronal functions. Endocrinology. 120:2398–2403.
53. Mellström, B., J.R. Naranjo, A. Santos, A.M. Gonzalez, and J. Bernal.
1991. Independent expression of the ? and ? c-erbA genes in developing rat
brain. Mol. Endocrinol. 5:1339–1350.
54. Pasquini, J.M., and A.M. Adamo. 1994. Thyroid hormones and the cen-
tral nervous system. Dev. Neurosci. 16:1–8.
55. Porterfield, S.P., and C.E. Hendrich. 1993. The role of thyroid hor-
mones in prenatal and neonatal neurological development—current perspec-
tives. Endocr. Rev. 14:94–106.
56. Sinha, A.K., M.R. Pickard, K.D. Kim, M.T. Ahmed, F. Al Yatama, I.M.
Evans, and R.P. Ekins. 1994. Perturbations of thyroid hormone homeostasis in
the adult and brain function. Acta Med. Austriaca. 21:35–43.
57. Whybrow, P.C. 1991. Behavioral and psychiatric aspects of Hypothy-
roidism. In The Thyroid: a Fundamental and Clinical Text. L.E. Braverman and
R.D. Utiger, editors. J.B. Lippincott Co., New York. 1078–1083.
58. Whybrow, P.C. 1991. Behavioral and psychiatric aspects of thyrotoxico-
sis. In The Thyroid: a Fundamental and Clinical Text. L.*******E. Braverman
and R.D. Utiger, editors. J. B. Lippincott Co., New York. 863–870.
59. Bunsey, M., and H. Eichenbaum. 1996. Conservation of hippocampal
memory functions in rats and humans. Nature (Lond.). 379:255–257.
60. DeLong, G.R., and R.D. Adams. 1991. The neuromuscular system and
brain in hypothryoidism. In The Thyroid: a Fundamental and Clinical Text.
L.E. Braverman and R.D. Utiger, editors. J.B. Lippincott Co., New York.
61. LoPresti, J.S., A. Eigen, E. Kaptein, K.P. Anderson, C.A. Spencer, and
J.T. Nicoloff. 1989. Alterations in 3,3?,5?-triiodothyronine metabolism in re-
sponse to propylthiouracil, dexamethasone, and thyroxine administration in
man. J. Clin. Invest. 84:1650–1656.
62. Silva, J.E., and P.S. Matthews. 1984. Production rates and turnover of
triiodothyronine in rat-developing cerebral cortex and cerebellum: responses to
hypothyroidism. J. Clin. Invest. 74:1035–1049.
63. Carvalho, S.D., E.T. Kimura, A.C. Bianco, and J.E. Silva. 1991. Central
role of brown adipose tissue thyroxine 5?-deiodinase on thyroid hormone-
dependent thermogenic response to cold. Endocrinology. 128:2149–2159.
64. Escobar-Morreale, H., M.J. Obregón, F. Escobar del Rey, and G. Mor-
reale de Escobar. 1995. Replacement therapy for hypothyroidism with thyrox-
ine alone does not ensure euthyroidism in all tissues, as studied in thyroidecto-
mized rats. J. Clin. Invest. 96:2828–2838.
65. Leonard, J.L., S.A. Mellen, and P.R. Larsen. 1983. Thyroxine 5?-deiodi-
nase activity in brown adipose tissue. Endocrinology. 112:1153–1155.
66. Silva, J.E., and J.L. Leonard. 1985. Regulation of rat cerebrocortical and
adenohypophyseal type II 5?-deiodinase by thyroxine, triiodothyronine, and re-
verse triiodothyronine. Endocrinology. 116:1627–1635.
67. Leonard, J.L., J.E. Silva, M.M. Kaplan, S.A. Mellen, T.J. Visser, and
P.R. Larsen. 1984. Acute posttranscriptional regulation of cerebrocortical and
pituitary iodothyronine 5?-deiodinases by thyroid hormone. Endocrinology.
68. St. Germain, D.L. 1985. Metabolic effect of 3,3?,5?-triiodothyronine in
cultured growth hormone producing rat pituitary tumor cells: evidence for a
unique mechanism of thyroid hormone action. J. Clin. Invest. 76:890–893.
69. Halperin, Y., L.E. Shapiro, and M.I. Surks. 1994. Down-regulation of
type II L-thyroxine, 5?-monodeiodinase in cultured GC cells: different path-
ways of regulation by L-triiodothyronine and 3,3?,5?-triiodo-L-thyronine. En-
70. Surks, M.I., and C.R. DeFresi. 1977. Determination of the cell number
of each cell type in the anterior pituitary of euthyroid and hypothryoid rats. En-
71. Koenig, R.J., J.L. Leonard, D. Senator, N. Rappaport, A.Y. Watson,
and P.R. Larsen. 1984. Regulation of thyroxine 5?-deiodinase activity by 3,5,3?-
triiodothyronine in cultured rat anterior pituitary cells. Endocrinology. 115:
72. Maia, A.L., M.J. Berry, R. Sabbag, J. Harney, and P.R. Larsen. 1995.
Structural and functional differences in the dio1 gene in mice with inherited
type I deiodinase deficiency. Mol. Endocrinol. 9:969–980.
73. Salvatore, D., S.C. Low, M. Berry, A.L. Maia, J.W. Harney, W. Croteau,
D.L. St. Germain, and P.R. Larsen. 1995. Type 3 iodothyronine deiodinase:
cloning, in vitro expression, and functional analysis of the placental selenopro-
tein. J. Clin. Invest. 96:2421–2430.
74. Becker, K.B., M.J. Schneider, J.C. Davey, and V.A. Galton. 1995. The
type III 5-deiodinase in Rana catesbeiana tadpoles is encoded by a thyroid hor-
mone-responsive gene. Endocrinology. 136:4424–4431.
75. Toyoda, N., J.W. Harney, M.J. Berry, and P.R. Larsen. 1994. Identifica-
tion of critical amino acids for 3,5,3?-triiodothyronine deiodination by human
type I deiodinase based on comparative functional-structural analysis of the hu-
man, dog, and rat enzymes. J. Biol. Chem. 269:20329–20334.