(Received for publication, May 9, 1997, and in revised form, June 2, 1997)
From INSERM, Unité 369, Faculté de Médecine Lyon-RTH Laënnec, 69372 Lyon Cédex 08, France
Anti-peptide antibodies directed against the C-terminal portion (amino acids 603-618) of the rat thyroid iodide transporter (rTIT) have been produced to characterize the molecular forms of rTIT in the rat thyroid and in the functional rat thyroid cell line, FRTL-5. rTIT is located on the basolateral membrane of rat thyroid follicular cells and randomly distributed on the plasma membrane of FRTL-5 cells that do not exhibit cell polarity. The major rTIT component corresponds to an 80-90-kDa glycosylated protein. After treatment of cell membrane fractions with N-glycosidase F or incubation of FRTL-5 cells with tunicamycin, rTIT has an apparent molecular mass of about 55 kDa. FRTL-5 cells cultured in the presence of TSH exhibit a high rTIT content and a high iodide uptake activity (IUA). Upon either removal of TSH or addition of cycloheximide, IUA declines more rapidly than rTIT. The half-life of rTIT was about 4 days. Re-exposure of 7-day TSH-deprived FRTL-5 cells to TSH causes a rapid synthesis of the glycosylated rTIT but a delayed re-induction of IUA. Tunicamycin totally prevents the TSH-dependent re-expression and activity of rTIT. Our data bring basic information on the location, structure, and turnover of rTIT and suggest that its activity is subjected to diverse control mechanisms including regulatory proteins.
Iodide trapping by epithelial thyroid cells is the first step of thyroid hormone synthesis within thyroid follicles. Thyroid iodide uptake brings into play a membrane transporter (1) defined as a Na+/iodide symporter (2). After the pioneering studies of Wolff (1), a definite step in the knowledge of the thyroid iodide transporter (TIT),1 has been made by Dai et al. (3) who cloned the cDNA of the rat TIT (rTIT) allowing the design of tools for further molecular analyses. Secondary structure prediction and hydropathic profile analyses indicated that rTIT is a membrane protein with 12 membrane-spanning domains thus resembling the other Na+-dependent cotransporters. The availability of the rTIT primary sequence prompted us to produce antibodies to characterize the TIT protein of the rat and other animal species. We choose to generate polyclonal antibodies against a peptide corresponding to the last 16 amino acids (amino acids 603-618) of the rTIT. Antibodies that were raised exhibited a very high titer and a high specificity for a rat thyroid membrane glycoprotein fulfilling the criteria for being the rTIT. These antibodies have been used (a) to visualize the rTIT on rat thyroid tissue sections, (b) to characterize the molecular form(s) of rTIT in the rat thyroid and FRTL-5 cells, and (c) to determine the turnover of rTIT in FRTL-5 cells in relation with changes in their capacity to concentrate iodide.
Coon's modified Ham's F-12 medium was obtained from Seromed (Biochrom KG, Berlin). Bovine TSH (2 units/mg), insulin, cortisol, transferrin, glycyl-L-histidyl-L-lysine acetate, tunicamycin, cycloheximide, and anti-rabbit IgG antibody conjugated to alkaline phosphatase were from Sigma. N-glycosidase F was purchased from Boehringer Mannheim and Immobilon P membrane from Millipore Corp. Na125I was obtained from ICN.
Preparation and Test of AntibodiesA peptide corresponding to the published sequence (3) of the C-terminal segment of the rTIT (amino acids 603-618) was synthesized by a solid-phase procedure using an automated peptide synthesizer (Neosystem Laboratories, Strasbourg, France). The peptide was conjugated to keyhole limpet hemocyanin using glutaraldehyde as a coupling reagent. Antisera were raised in rabbits by multipoint injections of 200 µg of the conjugate in complete Freund's adjuvant. After 2, 4, 8, and 12 weeks, rabbits were given a booster injection. Blood was collected 10 days after the last injection. Two immune sera pAb 716 and pAb 792 were obtained. Sera were tested for their antibody titer against the peptide by enzyme-linked immunosorbent assay.
Multiwell plates from Corning Costar Corp. (Cambridge, MA) were coated with the unconjugated peptide (10 µg/ml) in carbonate buffer, pH 9.6, for 16 h at 4 °C. After saturation of residual binding sites with bovine serum albumin (5 mg/ml) in 10 mM Tris, 0.15 M NaCl, pH 7.4 (TN buffer), serial dilutions of antisera in TN buffer were added, and incubation was performed for 2 h at 37 °C. After washings in TN buffer containing 0.05% Tween 20, immune complexes were detected using an anti-rabbit IgG antibody conjugated to alkaline phosphatase. The enzyme activity was assayed using p-nitrophenyl phosphate as substrate. Absorbance measurements were made at 405 nm. Controls for autohydrolysis of the substrate and nonspecific binding of the conjugate were run in parallel.
Cell CultureFRTL-5 cells, a continuous line of functional
epithelial cells from Fisher rat thyroid (4), remain differentiated
growing in 5% serum and require the presence of TSH for growth. Cells were grown in Coon's modified Ham's F-12 medium supplemented with 5%
calf serum, insulin (10 µg/ml), cortisol (108
M), transferrin (5 µg/ml),
glycyl-L-histidyl-L-lysine acetate (10 ng/ml),
and TSH (1 milliunit/ml). They were subcultured in either 24-well
plates or 60-mm culture dishes, and culture medium was changed every
other day. Cells were allowed to reach confluence before any treatment
was applied. FRT cells, a continuous line of non-functional but
polarized thyroid cells, were grown in the same medium as FRTL-5 with
two exceptions, normal calf serum was replaced by fetal calf serum and
TSH was omitted.
FRTL-5 cells,
grown in 24-well plates, were washed twice with Earle's balanced salt
solution, pH 6.8, supplemented with 1 mM methimazole and 10 µM NaI and incubated in the same medium containing
0.5-1.0 µCi of carrier-free Na125I for 60 min at
37 °C. In each experimental condition, incubations were performed in
the absence or in the presence of sodium perchlorate (0.1 mM) to measure the active iodide transport. At the end of the incubation period, the medium was aspirated, and cells were quickly
washed twice with ice-cold Earle's medium. The complete procedure was
achieved within 40 s. Radiolabeled iodide taken up by the cells
was released by the addition of 10% trichloracetic acid. Radioactivity
of the trichloracetic acid-soluble fraction was measured in a counter from Packard Instruments. Incubations were made in triplicate.
IUA was expressed in picomoles of iodide taken up per 106
cells. Cells detached from culture wells by a 15-min incubation in PBS
containing 0.25% trypsin and 0.02% EDTA were counted using an
hemocytometer.
Thyroid glands from Harlan Sprague Dawley rats (200-300 g body weight) rapidly removed after the animal sacrifice were fixed by immersion in the Bouin-Hollande solution. After paraffin embedding, 10-µm tissue sections were prepared and processed to allow immunofluorescence labeling. FRTL-5 cells were fixed in 4% paraformaldehyde for 30 min at 20 °C, then permeabilized using 1% Triton X-100 for 30 min at 20 °C and pre-incubated for 1 h at 20 °C in PBS/bovine serum albumin (1 mg/ml). Fixed and permeabilized FRTL-5 cells and rat thyroid tissue sections were incubated with the pAb 716 immune serum at a 1:1000 dilution overnight at 4 °C. After washings with 0.05% Triton X-100 in PBS, immune complexes were detected using a fluorescein-labeled secondary antibody. Images were prepared using the videomicroscope equipment previously described (5). Control incubations included the rabbit pre-immune serum, omission of the rabbit serum, and saturation of antibodies with the synthetic peptide.
SDS-PAGE and Western BlotFRTL-5 cells, collected by scraping in PBS containing protease inhibitors aprotinin, leupeptin, and pepstatin at a concentration of 1 µg/ml, were lysed by freezing-thawing. Cell homogenates were centrifuged at 100,000 × g for 60 min at 4 °C to obtain membrane fractions. Rat thyroid and adrenal homogenates were prepared in PBS supplemented with protease inhibitors in a glass-glass homogenizer. Particulate material was collected by centrifugation at 10,000 × g for 30 min at 4 °C. Protein was assayed by the Lowry method (16) after solubilization in 0.1% deoxycholate. Protein samples (2-15 µg) from total cell extracts or membrane fractions were separated by SDS-PAGE on 7-8% acrylamide slab minigels and then transferred onto an Immobilon P membrane. Incubations with the immune serum and the secondary antibody conjugated to alkaline phosphatase were performed as described previously (5). Immunolabeled protein spots were quantified by videoimage analysis (5).
The peptide (amino acids 603-618) selected by
antigenicity prediction analyses from the rTIT primary sequence (3) and
coupled to keyhole limpet hemocyanin has proved to be highly
immunogenic. The anti-peptide antibody titer of the two immune sera,
pAb 716 and pAb 792, was higher than 107, the titer being
defined as the highest dilution that gives a signal significantly
higher than that of the controls in the enzyme-linked immunosorbent
assay. This is illustrated for pAb 716 in Fig. 1. The
pre-immune sera were devoid of activity. When tested on FRTL-5 cell
extracts by Western blot, the two immune sera detected a main protein
migrating as an 80-90-kDa component and some other bands of lower
intensity with an apparent molecular mass ranging from 60 to 80 kDa.
The 80-90-kDa protein was still detected on Western blot by using a
10
6 dilution of the pAb 716 immune serum (Fig. 1). The
reactivity of anti-peptide antibodies toward the synthetic peptide
analyzed by solid-phase immunoassay and the ability of the antibodies
to label FRTL-5 protein(s) on Western blot were parallel on a large scale of dilution (up to 10
6) of the immune serum. The
binding of anti-peptide antibodies to the immobilized peptide was
inhibited by a prior incubation with the free peptide. The peptide
concentration that produced a 50% inhibition was about
10
7 M (data not shown). Anti-peptide
antibodies (from pAb 716 immune serum) labeled the same bands in cell
extracts from TSH-treated FRTL-5 cells and in rat thyroid particulate
fractions (lanes 2 and 4 and lane 6 of
Fig. 2A); an 80-90-kDa component was the
most abundant in both cases. In contrast, FRTL-5 cells deprived of TSH
for 7 days exhibited a lower amount of the 80-90-kDa protein and were
devoid of the component of lower apparent molecular mass (60-65 kDa)
(lines 3 and 5 of Fig. 2A). The
absence of protein labeling in cells (FRT cells) and tissue (rat
adrenals) known to be devoid of iodide uptake activity gives evidence
for the specificity of the reaction obtained with cells and tissue
known to express TIT. Treatment of the TSH-treated FRTL-5 cell membrane fraction with N-glycosidase F led to a large shift in the
distribution of the immunoreactive material; a main band with an
apparent molecular mass of about 55 kDa was formed. The deglycosylation
reaction yielded the same product when performed on membranes or
Nonidet P-40-solubilized membrane proteins (Fig. 2B). These
results indicate that rTIT is a glycoprotein. That anti-peptide
antibodies properly react with rTIT is further documented in Fig.
3. Anti-peptide antibodies intensely labeled the
periphery of FRTL-5 cells cultured in the presence of TSH (Fig.
3A). The labeling was concentrated in the regions of
cell-cell contacts but not limited to these parts of the plasma
membrane. Looking apart from cell boundaries, there was a diffuse
fluorescence becoming faint toward the middle of the cells. Numerous
fluorescent dots also mainly located at the cell periphery were found
over most cells. There was no labeling when the pre-immune serum was
used instead of the immune serum (Fig. 3B) or when the
immune serum was omitted. FRTL-5 cells deprived of TSH for 7 days were
hardly labeled (Fig. 3C). Given the extreme low background,
the low level of fluorescence of TSH-deprived cells probably
corresponded to a specific labeling, the distribution of which did not
appear different from that of TSH-treated cells. The immunofluorescence
staining of rat thyroid sections was more easily definable. As shown at
low (Fig. 3D) or high (Fig. 3E) magnification,
the anti-peptide antibodies clearly labeled the basolateral plasma
membrane domain which was remarkably delineated. On the opposite, the
apical plasma membrane domain facing the follicle lumen was devoid of
any fluorescence staining. The labeling intensity of thyrocytes
appeared rather homogeneous within individual follicles and between
follicles. As mentioned for FRTL-5 cells, the fluorescence background
(nonspecific labeling) of tissue sections was very low, indicating that
the anti-peptide antibodies react very selectively with the rTIT.
Relationship between rTIT Content and Iodide Uptake Activity of FRTL-5 Cells
FRTL-5 cells cultured in a TSH-free medium are known
to progressively lose the IUA within 5-8 days (6) and to re-acquire rapidly this activity upon addition of TSH (7). The turnover of rTIT in
FRTL-5 cells in relation to the loss and the re-establishment of IUA
(associated with the withdrawal and re-addition of TSH, respectively)
is reported in Figs. 4 and 5.
Quantitative densitometric measurements on Western blot were carried
out on the 80-90-kDa band which was considered as the mature,
potentially active, rTIT protein. IUA and rTIT protein content of
FRTL-5 cells maintained in the presence of TSH remained rather stable
for up to 7 days. Removal of TSH from the culture medium led to a
parallel 30% reduction of IUA and rTIT content within 1 day. From day
1 up to day 7 of TSH deprivation, there was a dissociation between the
disappearance rate of rTIT and the decrease of IUA (Fig.
4B). At day 5, IUA was abolished whereas cells exhibited a
rTIT content equal to about 30% of its initial value. We compared the
time-dependent decreases of rTIT and IUA obtained after TSH
withdrawal to the disappearance rate of rTIT and IUA resulting from the
inhibition of protein synthesis in FRTL-5 continuously treated with
TSH. Under cycloheximide (3 µM) treatment, rTIT slowly
decreased (Fig. 4C). The half-life of the 80-90-kDa rTIT
protein was about 4 days. It is worth noticing that the disappearance
rate of rTIT was very similar, if not identical, in cells treated with
cycloheximide and TSH and in TSH-deprived cells. More surprising was
the rapid cycloheximide-induced decrease of IUA. One day after blocking of protein synthesis, IUA was reduced by 55% whereas the rTIT cell
content was depleted by only 10-15%. By day 5, IUA and rTIT were
reduced to the same extent and declined in parallel thereafter (result
not shown). The treatment of FRTL-5 cells by 3 µM
cycloheximide for up to 7 days altered neither the cell number nor the
cell viability. Both TSH-deprived cells and cells cultured in the
presence of TSH and cycloheximide still contained rTIT after 7 days.
Only the 80-90-kDa rTIT form persisted; it amounted to about 15-20% of the rTIT content of TSH-treated cells. Addition of TSH to 7-day TSH-deprived FRTL-5 cells (Fig. 5) re-induced the synthesis of rTIT.
After 8 h (7.3 days from the outset of the experiments) the
60-65-kDa form reappeared (Fig. 5A), and there was a
significant increase in the 80-90-kDa rTIT. Within 24 h, the
80-90-kDa rTIT protein reached a value equal to or higher than that
observed in FRTL-5 cells at day 0 or FRTL-5 cells continuously cultured in the presence of TSH. The same results were obtained when FRTL-5 were
treated with forskolin instead of TSH (result not shown). As described
previously (7), we found that the restoration of IUA was delayed.
Twenty four h after TSH addition (day 8 from the beginning of the
experiments), IUA reached only 15-20% of its initial (day 0) level
whereas rTIT expression was maximum. Thereafter, IUA continued to
increase and plateaued at a value about 50% lower than that obtained
with FRTL-5 cells continuously cultured with TSH. Cycloheximide (3 µM) prevented the TSH-induced synthesis of rTIT. This
indicates that at the concentration of 3 µM,
cycloheximide very efficiently blocked protein synthesis in FRTL-5
cells. Tunicamycin, an inhibitor of the synthesis of N-linked oligosaccharides totally blocked the synthesis of
the 80-90-kDa rTIT. Accordingly, tunicamycin prevented the
TSH-dependent re-induction of IUA. It was verified that
tunicamycin did not alter cell viability. An immunoreactive protein
with an apparent molecular mass of 55 kDa (distinct from the 60-65-kDa
protein normally found in both FRTL-5 cells and rat thyroid) was
transiently detected at days 8 and 9 (1 and 2 days after TSH addition)
(Fig. 5A). This 55-kDa protein could correspond to the
non-glycosylated rTIT polypeptide; it has the same mobility in SDS-PAGE
as the deglycosylated protein shown in Fig. 2B.
Antibodies directed against the synthetic peptide corresponding to the last 16 amino acids of the rTIT sequence (3) allowed the identification of rTIT protein(s) (a) on membrane fractions by Western blot and (b) on fixed cells by indirect immunofluorescence in both FRTL-5 cells and rat thyroid gland. The major immunoreactive form of rTIT migrates as a broad band in SDS-PAGE; its apparent molecular mass of 80-90 kDa is substantially higher than that expected from the amino acid sequence, i.e. 65 kDa (3). This difference in size and the width of the electrophoretic band is probably attributable to the presence of N-linked oligosaccharide chains. Two potential N-linked glycosylation sites have been identified in the sixth and last extracellular loop of the predicted rTIT structure (3). On the opposite, the deglycosylated rTIT form obtained by N-glycosidase F treatment of FRTL-5 cell membranes has an apparent molecular mass (about 55 kDa) lower than the predicted molecular mass of the polypeptide chain. Abnormal SDS binding, contributions of the intrinsic charge, and conformation of the protein as well as proteolytic cleavage are possible explanations for the abnormal electrophoretic mobility of the deglycosylated rTIT (8). The occurrence of a proteolytic cleavage during the deglycosylation reaction is unlikely. Indeed, an immunoreactive form of the rTIT of the same apparent size has been identified in tunicamycin-treated FRTL-5 cells. The 55-kDa protein might thus correspond to the rTIT polypeptide chain. In addition to the 80-90-kDa component, both FRTL-5 cells and rat thyroid gland contain a second immunoreactive component with an apparent molecular mass of 60-65 kDa that we interpret as a glycosylation intermediate. This is supported by the fact that this component (a) disappeared in TSH-deprived FRTL-5 cells, (b) re-appeared (without accumulation) after re-addition of TSH, and (c) was not formed in response to TSH when cells were treated with tunicamycin. The 80-90-kDa protein, being the most abundant immunoreactive species in both rat thyroid gland and FRTL-5 cells, probably corresponds to the mature form of rTIT. This is strengthened by the fact that the anti-peptide antibodies that predominantly detected the 80-90-kDa protein, unequivocally labeled the basolateral membrane of thyrocytes where the functional form of TIT or sodium/iodide symporter should be located (9). It was recently reported that sera of patients with autoimmune thyroid diseases contain antibodies against TIT, some sera labeled a single band with an apparent molecular mass of about 80 kDa on FRTL-5 cell membranes (10), and some other sera inhibited the iodide uptake activity on CHO cells stably transfected with the rTIT cDNA but did not recognize rTIT on Western blot (11). These anti-TIT auto-antibodies probably interact with extracellular loops of rTIT. The 80-90-kDa protein detected by anti-peptide antibodies in the present study and the 80-kDa species labeled by human autoantibodies likely correspond to the same molecular entity.
Regulation of iodide transport by thyroid cells had been the subject of numerous studies (reviewed in Ref. 12) using different cell systems including FRTL-5 cells (2, 6, 7, 13). It has generally been accepted that TSH regulates IUA by acting on the synthesis of the TIT and/or the formation of an activating factor (2, 7, 14). Three conclusions can be drawn from the present study.
First, TSH is required for the maintenance of a steady-state level of rTIT in FRTL-5 cells, and the synthesis of rTIT is rapidly switched off or on upon suppression or re-addition of TSH. Indeed, the rate of disappearance of rTIT after TSH withdrawal was not different from the rTIT decay obtained by inhibiting protein synthesis. This is at variance with the model proposed by Kaminsky et al. (2) in which rTIT was supposed to constitutively reside in the plasma membrane. Addition of TSH to TSH-deprived FRTL-5 cells re-activated the expression of rTIT, and within 24 h, the rTIT cell content reached a steady-state level corresponding to the rTIT level of cells continuously cultured in the presence of TSH.
Second, the activity of rTIT appears to be subjected to regulatory processes. There was a marked dissociation in the time-dependent decrease of IUA and rTIT in cells treated with cycloheximide. The rapid initial loss of IUA after inhibition of protein synthesis in TSH-treated cells suggests the existence of a rTIT stimulatory protein with a turnover rate markedly higher than that of the rTIT. The involvement of such a stimulatory protein in iodide transport had been proposed long time ago by Knopp et al. (14) using dispersed thyroid cells and more recently further analyzed in FRTL-5 cells (2). However, the dissociation between IUA and rTIT content under cycloheximide treatment cannot be solely explained by the disappearance of an activating factor. Indeed, after the initial decrease, IUA remained rather stable for several days despite the progressive disappearance of rTIT. One has to postulate the implication of another regulatory event. The prolonged cycloheximide treatment could deplete cells in a long-lived protein endowed with an inhibitory action on rTIT. The dissociation between IUA and rTIT cell content after both TSH suppression and TSH re-addition also suggests the existence of an activating factor, the synthesis or the activity of which slowly decreases in TSH-deprived cells and slowly increases in cells re-exposed to TSH. Modulation of the activity of rTIT might possibly involve post-translational modifications of rTIT including phosphorylation or changes in the oligomerization state. Indeed, rTIT possesses a putative phosphorylation site for protein kinase A in its C-terminal part, and functional rTIT is possibly oligomeric.
Third, N-glycosylation of rTIT is essential for its synthesis and stability. In the presence of tunicamycin, the response of FRTL-5 cells (depleted in rTIT by a prolonged TSH deprivation) to TSH was limited to the transient formation of a low amount of a 55-kDa protein, presumably the rTIT polypeptide chain. As demonstrated for many other glycoproteins (15), the non-glycosylated form of rTIT likely accumulates in the endoplasmic reticulum where it is degraded. These data suggest that N-glycosylation is essential for correct folding and stabilization of the conformation of rTIT.
We are grateful to Prof. L. Nitsch (Naples, Italy) for providing us with the FRT and FRTL-5 cell lines. We thank Christine Audebet for her contribution to the management of the FRTL-5 cell line.