| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLE |
Thyroid Section, Division of Endocrinology (C.C.-M., A.M.Z., B.W.K., S.H., J.W.H., P.R.L., A.C.B.), Diabetes and Hypertension, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts 02115; and Institute of Experimental Medicine (B.G.), Department of Neurobiology, Budapest and University of Pécs, Faculty of Sciences, Institute of Biology, Budapest H-1083, Hungary
Address all correspondence and requests for reprints to: Antonio C. Bianco, M.D., Ph.D., Brigham and Women’s Hospital, 77 Avenue Louis Pasteur, Harvard Institutes of Medicine Building, Room 566, Boston, Massachusetts 02115. E-mail: abianco{at}rics.bwh.harvard.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All three deiodinase enzymes are integral membrane proteins. Topology software predicts and experimental data for D1, D2, and D3 indicate that they have a single transmembrane domain that lies within the first 30–40 NH2-terminal residues (2, 3, 4, 5). D1 is a plasma membrane protein with a relative molecular mass (Mr) of approximately 29,000, of which the approximately Mr 25,000 COOH-terminal domain is in the cytosol (6, 7, 8, 9). Unlike D1, D2 is an approximately Mr 31,000 endoplasmic reticulum (ER)-resident protein with its NH2 terminus located in the ER lumen and the COOH terminus in the cytosol (4). D3 is an approximately Mr 33,000 protein that inactivates thyroid hormones. It is located in the plasma membrane and recycles rapidly to the early endosomal compartment. Most of its molecule is extracellular, including the active center, and can be biotinylated with a cell impermeant probe (5).
Early attempts to purify D1 and D2 using biochemical approaches identified activity in higher molecular weight forms than would be predicted retrospectively from their respective deduced amino acid sequences (Mr 29,000–33,000) (10, 11). For example, solubilization of kidney or liver microsomal membranes with deoxycholate or Brij56 followed by gel filtration or gradient centrifugation identified D1 activity in proteins of Mr 44,000–65,000 (10, 11). Based on sedimentation coefficients and the Stokes radius, the calculated molecular weight for detergent-solubilized D2 was estimated to be approximately Mr 200,000 (11). Although these higher molecular weight complexes could be due to residual tightly bound membrane fragments, some have suggested that D1 and D2 could be present in multimeric forms (12, 13, 14, 15). This raises the possibility that homodimerization or association with other proteins is necessary for enzymatic activity. On the other hand, these higher molecular weight forms could reflect associations with other cellular protein(s) not primarily involved in their catalytic function, but which could, for example, regulate half-life, transport, or subcellular localization. A recent study showed that when an inactive rat D1 (ratD1) mutant protein (Sec126Ser) was expressed in porcine LLC-PK1 cells, all the endogenous porcine D1 (pigD1) activity was eliminated from the cell lysate after immunoprecipitation (IP) with a specific anti-ratD1 antiserum although D1 activity was not demonstrated in the IP pellets (16). This suggested that a large fraction of the endogenous porcine and mutant ratD1 were present as heterodimers (pigD1-ratD1). By analogy, dimers might also be formed with D2 or D3. On the other hand, no such higher molecular weight complexes structures have been identified in cells overexpressing wild-type (wt) deiodinases (17, 18, 19, 20, 21).
In the present studies, we used site-directed mutagenesis, epitope tagging, 75Se labeling, and dithiothreitol (DTT) sensitivity to determine whether recombinant D1, D2, and D3 protein can form homodimers as well as to determine whether such forms are required for their function. We identified small amounts of dimer forms of all three proteins and show that a dimer of D1 or D2 with only one active monomer is functionally active.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-iodoacetamide (IodAc), N-ethylmaleimide (NEM), and protease inhibitors were obtained from Calbiochem (La Jolla, CA). T3, T4, and rT3 were dissolved in 40 mM NaOH. Protein A plus G agarose beads suspension was obtained from Oncogene Research Products (Boston, MA). BM chemiluminescence Western blotting kit (mouse/rabbit) was from Roche (Indianapolis, IN). The polyvinylidene fluoride (Immobilon) membrane was from Millipore Corp. (Bedford, MA). Restriction enzymes and Vent DNA polymerase were from New England Biolabs, Inc. (Beverly, MA). TnT Quick Coupled transcription/translation system and canine pancreatic microsomal membranes were from Promega Corp. (Madison, WI). Glutathione Sepharose 4B resin was obtained from Amersham Pharmacia Biotech (Piscataway, NJ). L-[35S]-Methionine Easytag was obtained from Perkin-Elmer (Boston, MA). The Expand High Fidelity PCR System was purchased from Roche. Custom oligos were synthesized by Invitrogen (Carlsbad, CA), and the Gateway cloning, expression, and PCR systems were also obtained from Invitrogen. Autofluor was obtained from National Diagnostics (Atlanta, GA). Outer ring-labeled [125I]-T3, rT3 or T4 (specific activity: 4400 Ci/mmol) were from NEN Life Science Products (Boston, MA). Na2[75Se]O3 was kindly provided by the University of Missouri Research Reactor, courtesy of Drs. Marla Berry and Dolph L. Hatfield. All other reagents were of analytical grade.
Constructs employed in experiments involving Western blot analysis, IP of 75Se-labeled deiodinases, or IP followed by activity measurement
The plasmids encoding wt deiodinases were previously described: wtD1 is a human wt D1 (G21-D10), wtD2 is a human wt D2 (hD2-SelP), and wtD3 is a human wt D3 (hD3-CDM8) (19, 22, 23). The reference to the FLAG fusion with each deiodinase protein is either FLAG-[deiodinase], when the FLAG peptide is fused to the NH2 terminus, or [deiodinase]-FLAG, when it is fused to the COOH terminus (Fig. 1A
). FLAG-CysD1 is an NH2 terminus FLAG-tagged rat D1, and CysD1-FLAG is a COOH terminus FLAG-tagged rat D1. In both constructs, the Sec in position 126 was replaced by cysteine (Cys) (24).
|
FLAG-CysD3 is an NH2 terminus FLAG-tagged human D3 in which Cys replaced Sec145. The plasmid was constructed using overlap extension PCR. Briefly, Bp78-Bp81 and Bp79-Bp80 (Table 1) were used as oligos for the wt human D3 template. The resulting fragments were amplified by PCR using the outer oligos sense Bp78 and antisense Bp79 and subcloned in the EcoRI/NotI sites of a modified T7 pGem-T vector, containing an EcoRI site in its 5' cloning site. The insert was reamplified using oligos Bp86-Bp87. The resulting fragment was digested by NheI/BamHI and subcloned in a FLAG containing D10, fusing the FLAG to the NH2 terminus of CysD3. CysD3-FLAG is a COOH terminus FLAG-tagged human D3 in which the Sec145 was replaced by Cys. Oligos Bp122 and 123 were used to generate a CysD3 with COOH terminus FLAG using FLAG-CysD3 as template. The product was digested with EcoRI/XbaI and cloned in the same sites of a D10 vector.
|
AlaD3-FLAG is a COOH terminus FLAG-tagged human D3 with Ala replacing Sec145 and was generated by overlap extension PCR using oligos Bp86-C8 and on FLAG-CysD3 as template. The resulting PCR fragments were amplified by PCR together using the outer oligos Bp86 and Bp87. The fragment was subcloned in the NheI/BamHI sites of FLAG-CysD3.
In some experiments, we used FLAG-tagged wt enzymes, designated as FLAG-wtD1, wtD2-FLAG, and FLAG-wtD3, prepared as follows. FLAG-wtD1 is an NH2 terminus FLAG-tagged wt human D1. Briefly, a PCR fragment was obtained using C52-C53 on wt hD1 template (22). The resulting fragment was subcloned in the EcoRI/HindIII sites of a D10 vector containing minimal Sec insertion sequence (SECIS) element. wt D2-FLAG (wtD2-FLAG) is a COOH terminus FLAG-tagged human D2 derived from the wtD2 template (23) with minimal SelP SECIS element. The PCR fragment was obtained on wt hD2 template, digested with EcoR1/XbaI and subcloned into a D10 vector containing minimal SECIS element. The same cloning strategy used for FLAG-wtD1 was applied for FLAG-wtD3, which is an NH2 terminus FLAG-tagged human D3. C50-C51 PCR fragment was obtained on wt hD3cdm template (19), digested with EcoRI/HindIII and subcloned into sites of a D10 vector containing minimal SECIS element.
Constructs used in experiments involving bacterial expression of glutathione-S-transferase (GST)-fused deiodinases or in vitro transcription-translation
FLAG-CysD2-pGEM is an NH2 terminus FLAG-CysD2 generated by RI/SalI digestion of FLAG-CysD2 (4) and subcloned into the same sites of a modified T7 pGem-T vector, containing an EcoRI site in its 5' cloning site. FLAG-CysD1-pGEM and FLAG-CysD3-pGEM are T7-driven pGEM-T based plasmids containing an NH2 terminus FLAG-CysD1 and NH2 terminus FLAG-CysD3, respectively, and were generated following the same strategy used for FLAG-CysD2-pGEM-T, but using FLAG-CysD1 or FLAG-CysD3 as templates.
GST fusions of D1, D2, and D3 for an Escherichia coli overexpression vector were constructed as follows. PCR product containing CysD2 was created by amplifying CysD2-FLAG with the Expand High Fidelity PCR System using the oligos A and B. The resulting product was used to create pDONR201-CysD2 using the BP reaction with Gateway PCR system following the manufacturer’s instructions. This vector was then used to generate an E. coli expression vector containing a NH2 terminus GST fusion of D2, denominated GST-CysD2, via the ligation reaction (LR) of the Gateway E. coli Expression System. The same cloning strategy was used to generate GST-CysD1 and GST-CysD3. Briefly, a PCR fragment containing CysD1 (22) and another containing CysD3, were amplified with the oligos C and D; E and F, respectively, to generate PCR fragments. These fragments were cloned into pDONR201 via the BP reaction of the Gateway cloning system as above and then transferred into pDest15 via the LR to create GST-CysD1 and GST-CysD3. The GST-ß-glucoronidase (GUS) gene construct was generated as a positive control for the LR and contains the Arabidopsis thaliana GUS gene.
Cell culture and transfections
Human embryonic kidney epithelial cells (HEK-293) were plated in 60-mm dishes and grown until confluence in DMEM supplemented with 10% fetal bovine serum. HEK-293 cells were transfected using the calcium phosphate method as described earlier (26). In all experiments, human GH (thymidine kinase GH) was used as a control for the transfections, as described before (27). The cotransfection ratios of Cys mutant constructs and wt enzymes varied from 1:1 to 9:1 depending on the experiment. For cotransfections, the following plasmid combinations were used:
For activity measurements:
D1: a) 9 µg FLAG-AlaD1 and 1 µg wtD1; b) 9 µg CysD1-FLAG and 1 µg wtD1; c) 9 µg of CysD1-FLAG and 1 µg of D10; or d) 1 µg wtD1 and 9 µg of D10.
D2: a) 9 µg AlaD2-FLAG and 1 µg wtD2; b) 9 µg CysD2-FLAG and 1 µg wtD2; c) 9 µg of CysD2-FLAG and 1 µg of D10; or d) 1 µg wtD2 and 9 µg of D10.
D3: a) 9 µg FLAG-AlaD3 and 1 µg wtD3; b) 9 µg CysD3-FLAG and 1 µg wtD3; c) 9 µg of CysD3-FLAG and 1 µg of D10; or d) 1 µg wtD3 and 9 µg of D10.
For 75Se labeling experiments:
D1: a) 5 µg FLAG-CysD1 and 5 µg wtD1; c) 5 µg FLAG-CysD1 and 5 µg of D10; d) 5 µg wtD1 and 5 µg of D10; e) 5 µg FLAG-CysD3 and 5 µg wtD1.
D2: a) 5 µg FLAG-CysD2 and 5 µg wtD2; c) 5 µg FLAG-CysD2 and 5 µg of D10; d) 5 µg wtD2 and 5 µg of D10; e) 5 µg FLAG-CysD3 and 5 µg wtD2.
D3: a) 5 µg FLAG-CysD3 and 5 µg wtD3; c) 5 µg FLAG-CysD3 and 5 µg of D10; d) 5 µg wtD3 and 5 µg of D10.
For all other experiments, when only one type of plasmid encoding for deiodinases was used for transfection, the DNA amount used per plate was 10 µg. Cells transfected with D10 were used as negative control in all experiments. Each experiment was performed two to four times for each condition. At the appropriate times, cells were harvested and processed for Western blot analysis or IP followed by activity assay or SDS-PAGE (see below).
Western blot analysis
This was performed as previously described (24). Briefly, cells were scraped, washed in PBS, and sonicated in lysis buffer containing 50 mM Tris (pH 7.4), 1 mM EDTA, 140 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, 1 µg/ml leupeptin, and 2% sodium dodecyl sulfate. Some cell samples were processed in lysis buffer containing 1 mM NEM and 1 mM IodAc as indicated. DTT was added to final concentrations of 0–200 mM. Samples containing 20–40 µg protein were boiled for 5 min, resolved by 12% SDS-PAGE and electrotransferred to a polyvinylidene fluoride membrane. The blots were probed with anti-FLAG antibody at a final dilution of 1:2000. The signal was detected using the BM chemiluminescence Western blotting kit according the instructions of the manufacturer (Roche).
IP of deiodinases
HEK-293 cells were transfected with different plasmid combinations as indicated above and on d 2 after transfection were labeled with 4–6 µCi Na2(75Se)O3/dish, in the presence of DMEM supplemented with 10% fetal bovine serum. On d 3, the cells were washed with PBS and sonicated in 500 µl lysis buffer containing 0.5%, 1% Triton X-100 or 5 mM tDOC, 1% bovine hemoglobin, 0.2 U aprotinin/ml, 0.2 U leupeptin/ml, 1 mM phenymethylsulfonyl fluoride in TSA buffer (0.01 M Tris-HCl, pH 8.0; 0.14 M NaCl) 0.5 ml/dish. Ten microliters of total sonicate were saved from each cell lysate. Lysis was completed under slow agitation at 4 C for 30 min. The supernatant of 1200 x g for 15 min was incubated under slow agitation at 4 C for 1 h with preimmune mouse serum to a final dilution of 1:100 and 20 µl protein A plus G agarose suspension followed by centrifugation at 200 x g for 15 min. The supernatant was incubated with 2 µl of anti-FLAG antibody and 20 µl of protein A plus G agarose beads suspension for 12–18 h under slow agitation at 4 C. After centrifugation at 450 x g for 15 min, pellets were washed three times in 1:10 diluted lysis buffer, twice in 0.3 M TSA buffer (0.3 M Tris-HCl, pH 8.0; 0.14 M NaCl; 0.025% NaN3) and twice in 0.14 M TSA (0.14 M Tris-HCl, pH 8.0; 0.14 M NaCl; 0.025% NaN3). All pellets were resolved by 10% SDS-PAGE. To obtain IP pellets for activity assays a similar IP method was used, with minor modifications (18). Figure 2 presents all steps involved in this procedure and all experiments with controls done for its optimizations are described in the results section.
|
E. coli overexpression of GST-deiodinase fusion proteins
GST-D1, GST-D2, GST-D3, and GST-GUS were expressed in E. coli BL21 DE3 pLysS. Cultures were grown up until log phase and then induced with 0.2 mM isopropyl-ß-D-thiogalactopyranoside for 3 h at 37 C. Cells were pelleted and frozen. Pellets were thawed on ice and ressuspended in cold extraction buffer (25 mM Tris, pH 7.5; 1 mM EDTA; 20 mM NaCl; 20% glycerol, plus protease inhibitors) and sonicated. This lysate was spun at 16,000 x g for 15 min at 4 C, and the supernatant was saved (crude bacterial extract). To quantitate the amount of protein expressed, 50 µl of this crude bacterial extract were incubated for 1 h at room temperature with 30 µl of GST resin previously washed twice with 1 ml PBS. After incubation, the resin was washed with 1 ml of 25 mM Tris buffer (pH 7.0) containing 0.5% Triton X-100, 300 mM NaCl, 1 mM CaCl2 (wash buffer 1) followed by 1 ml of 25 mM Tris buffer (pH 7.0) containing 140 mM NaCl, 1 mM CaCl2 (wash buffer 2). Pellets were ressuspended in sample buffer, run on a 10% SDS-PAGE gel and visualized with coomassie blue staining.
In vitro studies of deiodinase dimerization
FLAG-CysD1-pGEM, FLAG-CysD2-pGEM, and FLAG-CysD3-pGEM were translated in vitro. One microgram of plasmid DNA was used per 50-µl reaction containing 2 µl L-[35S]-methionine. For in vitro dimerization studies, 50 µl of GST-D1, D2, and D3 bacterial crude lysate and 25 µl of GST-GUS bacterial extract were bound to 30 µl of GST resin for 1 h at room temperature under gentle agitation. All samples were centrifuged at 10,000 x g for 1 min and the supernatant discarded. Ten microliters of the indicated in vitro translation was added and incubated overnight at 4 C under gentle agitation. Total volume was centrifuged at 10,000 x g and washed with 1 ml of cold wash buffer 1 and by 1 ml of cold wash buffer 2. Pellets were boiled in sample buffer for 3 min, and the amount of 35S-methionine-labeled protein precipitated was resolved in a 12.5% SDS-PAGE gel. After it was fixed, the gel was enhanced using autoradiographic image enhancer Autofluor, following the manufacturer’s instructions, and dried before autoradiography.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Addition of Na2[75Se]O3 to the media permitted labeling and visualization of the wt selenodeiodinases using a strategy for specific immunoprecipitation of FLAG-tagged deiodinases. In other experiments, the lysates and immunoprecipitates were also assayed for deiodinase activity. The fusion of the FLAG to either the NH2 or COOH terminus sequence does not alter the kinetics the respective deiodinases. FLAG-CysD1 and CysD1-FLAG have similar Michaelis constant (Km)[rT3] (
7 µM) to a CysD1 mutant without the FLAG (3 µM) (24). FLAG-wtD2 and wtD2-FLAG (5–8 nM) have similar Km[T4] as wtD2 (0.4–1.3 nM). The same is true for D3 as wtD3 or FLAG-wtD3 have Km[T3] values of the same magnitude (1 nM T3) (data not shown).
Identification of 75Se-labeled wt-mutant deiodinase dimers in transfected HEK-293 cells
After transient expression of wt and Cys-mutant deiodinase constructs and exposure to Na2(75Se)O3, a number of 75Se-proteins were present in the SDS-PAGE of total cell lysate (Fig. 3, A–C). 75Se-labeled proteins of the characteristic size of D1, D2, and D3 appeared only in cells transfected with a plasmid encoding the wt construct. Dimerization of a fraction of the wt deiodinase with the FLAG-tagged mutant protein is confirmed by the presence of labeled protein in the SDS-PAGE gels of the stringently washed anti-FLAG IP complexes. For each deiodinase, an intense 75Se-labeled monomeric band of the appropriate size is found only when both the FLAG-Cys mutant and wt deiodinase encoding plasmid are transfected together, although trace amounts of nonspecific 75Se-labeled deiodinase are also present in other lanes. There is minimal cross-hybridization of FLAG-CysD3 with wt D1 or D2, indicating the dimerization between deiodinase monomers is relatively specific (Fig. 3, A and B). Quantification of these 75Se-labeled bands indicated that the wt 75Se-monomer in the IP pellets of each deiodinase is 2–5% of the total 75Se-deiodinase present in the cell lysate (Fig. 3, A–C). However, this is an underestimate of the fraction of wt dimerized to the FLAG-Cys deiodinases due to the high stringency of the washing procedure (see below).
|
, A–D). The lysis buffer was supplemented with 1 mM NEM and 1 mM IodAc to prevent the formation of disulfide bonds during lysate processing. In all samples, the predominant band was the predicted size of the full-length deiodinase monomer. When isolated under nonreducing conditions, an additional high molecular weight FLAG-containing band is present in all samples, Mr 55,000 (D1), Mr 62,000 (D2), and Mr 65,000 (D3), which are the predicted sizes of the respective dimers. The relative proportion of these higher molecular weight bands varied for the different deiodinases, being approximately 1–5% of the total FLAG signal for D1 and D2, and approximately 15% for D3. Exposure to 100 C for 5 min in the presence of increasing DTT concentrations before SDS-PAGE progressively reduced these dimeric forms, indicating they are due to disulfide bonds.
|
Studies of functional activity in IP pellets
Having determined that one can immunoprecipitate 75Se-labeled wt monomers dimerized with FLAG-tagged Cys mutant monomers, we wished to determine whether activity could be detected in the dimeric IP pellets. HEK-293 cells were cotransfected with FLAG-tagged Ala or Cys deiodinase mutants and plasmids encoding the corresponding wt molecules (wtD1, wtD2, or wtD3, respectively). After lysis, FLAG-tagged proteins were immunoprecipitated with anti-FLAG antibody. The IP efficiency was determined by comparing the deiodinase activities of the Cys-mutant deiodinase in the pre- and post-IP supernatants.
For optimization of the deiodinase imunoprecipitation, a series of experiments was performed to allow detection of deiodinase activity in the IP pellet. D1 activity was followed during the five steps (I–V) of cell lysis and IP (Fig. 2). Addition of anti-FLAG antibody or proteins A plus G agarose beads suspension to total cell lysate, as well as overnight incubation, did not inhibit D1 activity. The specificity of the procedure is monitored by assay for wt deiodinase activity in the IP pellet in the absence of cotransfected FLAG-tagged deiodinase derivative.
Because of the substantial loss of deiodinase activity in the IP pellet during washing, we attempted to optimize the conditions for the IP by varying the detergent type and concentration in the cell lysis buffer. Triton X-100 inhibited D1 activity in a concentration-dependent fashion as compared with the same samples processed with 5 mM tDOC, as previously described (29). However, the IP efficiency was decreased using tDOC resulting in similar overall D1 activity in the IP pellet. Similar results were obtained for D2 and D3, and therefore 0.5% Triton X-100 was used in all subsequent experiments (data not shown).
Because in some experiments we wished to compare the efficiencies of Cys vs. Ala deiodinase mutants to co-IP wt monomers, we also developed assay conditions in which wt and Cys mutant deiodinase activities could be discriminated. For D1, this was accomplished using GTG as an inhibitor of wtD1 (30). The inhibition constant (Ki) for GTG is 100-fold lower for inhibition of wtD1 than for CysD1 (30). Therefore, we determined the GTG concentration that inhibited more than 90% wtD1 but had little effect on CysD1-FLAG (960 nM; Fig. 5A). This approach permitted specific assays for wtD1 activity in the presence of the respective Cys mutant.
|
).
Because there are no data on the functional effects of replacing Sec with Cys in D3, we assayed the CysD3-FLAG and determined it has an approximately 150-fold higher Km[T3] (0.15 µM T3) than wt (Fig. 5C). Accordingly, a similar strategy to that used for D2 was applied to distinguish between wtD3 and CysD3-FLAG activities by comparing 125I T3 deiodination at 2 and 200 nM T3 (Fig. 5D).
Wt D1 and D2 activities require only one active dimer partner
In HEK-293 cells transfected with CysD1-FLAG, IP with anti-FLAG antibody precipitated 52% of CysD1 activity in the lysate but the amount of CysD1 activity in the washed IP pellet was only 1.5% of the total lysate (Table 2). This low yield is due to the high stringency of the wash procedure, which was designed to reduce the nonspecific binding of wt deiodinases in the IP pellet to negligible levels as explained above. In cells expressing only wtD1, only 6% of the activity was present in the crude precipitate but insignificant activity remained in the washed IP pellet (Table 2, column 2). In contrast, in cells coexpressing FLAG-AlaD1 and wtD1, about 11% of the D1 activity was lost from the lysate during the IP process, 5% more than when no FLAG-tagged mutant D1 was present (Table 2, column 3 vs. column 2). Thus, approximately 10% of the D1 activity was specifically immunodepleted from the lysate using anti-FLAG antibody when corrected for the approximately 50% IP efficiency. As expected from the stringency of the wash conditions, only a small fraction of this wtD1 activity was present in the pellet whereas no D1 activity was found in the pellet when no FLAG D1 bait was coexpressed (Table 2, compare columns 2 and 3). These results establish that wtD1 monomers complexed to an inactive FLAG-AlaD1 are active. The low yield of wtD1 might be explained if an enzymatically active bait (e.g. CysD1-FLAG) was superior to the inactive AlaD1. Therefore, we repeated these experiments replacing the inactive FLAG-Ala deiodinase bait with FLAG-Cys D1, but the results were similar (Table 2, column 4).
|
, column 1) with the recovery of 4% of the D2 activity in the IP pellet. Under these conditions, there was negligible nonspecific depletion of wtD2 (Table 3, column 2). However, approximately 38% of the wtD2 activity was immunoprecipitated by anti-FLAG antibody from cells transiently coexpressing AlaD2-FLAG and wtD2 (Table 3, column 3). Although the amount of wtD2 in the washed pellet was much lower than expected, it was 15-fold higher than in absence of AlaD2-FLAG (0.3 vs. 0.02%). The results indicate that the wtD2 monomer in the IP pellets is also active when complexed with an inactive AlaD2-FLAG. We also employed an active CysD2-FLAG bait to pull down wtD2 activity. However, the CysD2-FLAG appeared to trap less wtD2 activity than did AlaD2-FLAG. The accuracy of this estimate is compromised due to the necessity for assay of small amounts of wtD2 in the presence of large quantities of CysD2-FLAG.
|
In vitro studies of deiodinase dimerization
To analyze for possible posttranslation protein-protein interaction we used an in vitro system consisting of a GST-tagged Cys-deiodinases bait and FLAG-tagged radiolabeled deiodinases target. GST-deiodinase fusion proteins were generated by expression in bacteria. D1 activity is present in bacterial lysates expressing CysD1 mutant (Berry, M. J., and P. R. Larsen, personal communication). The 35S-radiolabeled target FLAG-CysD1-pGEM, FLAG-CysD2-pGEM and FLAG-CysD3-pGEM were synthesized in reticulocyte lysates. Samples from both: FLAG-tagged radiolabeled deiodinases target (10 µl of a 50-µl reticulocyte lysate translation reaction) and bait GST-deiodinase (1 µg protein) were incubated for 60 min at 30 C. The samples were processed for pull down using glutathione-sepharose resin. No significant amounts of the corresponding monomers were pulled down as assessed by the SDS-PAGE analysis of the resin pellets (data not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Furthermore, homodimers of Cys mutants are directly seen in Western blots of cell lysates after when DTT is omitted from the lysis buffer (Fig. 4).
Based on three different experimental approaches, i.e. co-IP of 75Se-labeled deiodinase (Fig. 3), Western analysis (Fig. 4), homodimers constitute a relatively modest fraction of the total enzyme expressed, varying from 1–15% of the total pool. Dimers identified by Western analysis require DTT at 100 C for 5 min for disruption, suggesting that S-S bonds are involved (Fig. 4). Our data, however, do not allow us to exclude the possibility that a substantial fraction of deiodinase protein is present in loosely complexed dimers not stabilized by S-S bonds. It is likely that the stringency of the wash conditions would dissociate such complexes. For example, we were able to specifically immunodeplete 10% of wtD1 and 38% of wtD2 using the respective Ala mutant baits but found only low amounts of activity in the washed pellets (Tables 2 and 3). In vitro formation of dimers was not observed during incubation of GST-tagged bacterially expressed deiodinases with the respective 35S-labeled counterparts in pull-down experiments, suggesting that dimer assembly may only occur cotranslationally (data not shown).
In the case of D1 and D2, we could also confirm the presence of small quantities of specifically complexed wt monomers and FLAG-tagged inactive (Ala) mutants by IP (Tables 2 and 3). The most significant result of these studies, however, was that wt enzyme activity was attributable specifically to a wt monomer by virtue of the technique used to isolate it (Tables 2 and 3). This constitutes the first direct evidence that wt D1 or D2 deiodinase monomers are active, at least when dimerized with an inactive partners. Our data do not allow a conclusion as to whether an uncomplexed monomer would be active. Surprisingly, even though wtD3-CysD3 dimers were visualized by 75Se-labeling (Fig. 3C) and in nondenaturing SDS-PAGE (Fig. 4, C and D), we could not immunoprecipitate active CysD3-FLAG. Our data thus do not allow a determination of the functional state of the D3 monomer complexed to an inactive protein.
The present results for D1 agree qualitatively with a previous study with respect to the fact that dimeric complexes of D1 can be formed in vivo (16). In that study, however, the major evidence for dimer formation was that all the endogenous pig D1 activity could be immunodepleted by an antibody directed against a unique epitope of a catalytically inactive ratD1 coexpressed in LLC-PK1 cells. It was inferred from these results that all the endogenous pig D1 was in a heterodimer configuration with the inactive ratD1 (ratD1-pigD1). Because D1 activity was still present in such cells, it was argued that the pigD1 monomer was sufficient to confer D1 activity to the heterodimer. Our results confirm that supposition by direct assay of the obligate heterodimer isolated in the IP pellets. Our results show in addition that small amounts (1–5% of the deiodinase protein) of covalently linked dimers due to S-S bonds are formed as mentioned. It is possible that higher estimates, which would be inferred from immunodepletion studies in our evaluation and in reference (16), are due to the IP of noncovalently associated monomers, perhaps dependent on stabilization by coassociation in small membrane fragments. A similar phenomenon could explain the identification of D1 and D2 activities in higher molecular weight complexes by gel filtration techniques.
Although our results established the existence of deiodinase dimerization for the three deiodinases, each of three approaches used in the present investigation indicated that the amount of dimer present for all deiodinases is small. In fact, to be able to detect the dimerization by activity assays, the amount of Cys or Ala mutant DNA (bait) used was 10-fold higher than their respective wt deiodinase. Considering also that the translation efficiency of Cys and Ala proteins is about 100-fold times higher that of wt (17, 25), it seems likely that the bait:target ratio is at least 1000:1. Similar stable overexpression of Sec126Ser-rat D1 bait was used to trap pig D1 in the earlier study (16). This is further supported by our failure to detect deiodinase dimers when FLAG-wt D1, D2, or D3 were transiently expressed and 75Se-labeled, even using sensitive techniques such as cell lysis in DTT-free buffer, mild wash of the IP pellet and film-detection for 2–3 wk. Further studies are needed to determine if deiodinase dimers occur in cells expressing deiodinases at endogenous levels.
In the course of these studies, we prepared several novel deiodinase mutants in which the Sec contained in the active center was replaced by Ala or Cys, namely FLAG-AlaD1, AlaD3-FLAG, FLAG-CysD3 and CysD3-FLAG. Site-directed mutagenesis has been used in the past to delineate the critical role played by Sec in the active center of D1 and D2. For D1 and D2, the replacement of Sec by Cys increased the Km[rT3] or Km[T4] by 10- and 500-fold, respectively (17, 25). A similar finding was obtained in the present investigation with D3. Independently of NH2 or COOH FLAG added to CysD3, the replacement of Sec by Cys increased the Km[T3] of D3 approximately 150-fold suggesting a critical structural role for Sec in substrate interaction with the active center, similar to its effect on D2. Substitutions of other amino acids for Sec are known to cause inactivation of D1 (Sec110Ser) and D2 (Sec133Ala) (22, 25). We can now extend this generalization to D3 because the Sec to Ala substitution inactivated the enzyme.
In summary, our results indicate that D1, D2, and D3 are present as dimers after transient expression and that S-S bonds stabilize a small fraction of these. The functional significance of this dimer form is not yet clear because at least for D1 and D2 a single wt monomer complexed with an inactive partner retains catalytic activity at least under in vitro conditions.
|
Footnotes |
|---|
Abbreviations: Ala, Alanine; Cys, cysteine; D1–D3, types 1, 2, and 3 iodothyronine selenodeiodinases; DTT, dithiothreitol; GST, glutathione-S-transferase; GTG, gold thioglucose; GUS, ß-glucoronidase; HEK, human embryonic kidney epithelial cells; IodAc, -iodoacetamide; IP, immunoprecipitation; LR, ligation reaction; Mr, relative molecular mass; NEM, N-ethylmaleimide; rT3, reverse T3; Sec, selenocysteine; SECis, Sec insertion sequence; tDOC, taurodeoxycholic acid; wt, wild-type.
Received September 11, 2002.
Accepted for publication December 3, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Gereben, A. M. Zavacki, S. Ribich, B. W. Kim, S. A. Huang, W. S. Simonides, A. Zeold, and A. C. Bianco Cellular and Molecular Basis of Deiodinase-Regulated Thyroid Hormone Signaling Endocr. Rev., December 1, 2008; 29(7): 898 - 938. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D. V. Sagar, B. Gereben, I. Callebaut, J.-P. Mornon, A. Zeold, C. Curcio-Morelli, J. W. Harney, C. Luongo, M. A. Mulcahey, P. R. Larsen, et al. The Thyroid Hormone-Inactivating Deiodinase Functions as a Homodimer Mol. Endocrinol., June 1, 2008; 22(6): 1382 - 1393. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. P Brouwer, B. C Appelhof, R. P Peeters, W. J G Hoogendijk, J. Huyser, A. H Schene, J. G P Tijssen, R. Van Dyck, T. J Visser, W. M Wiersinga, et al. Thyrotropin, but not a polymorphism in type II deiodinase, predicts response to paroxetine in major depression. Eur. J. Endocrinol., June 1, 2006; 154(6): 819 - 825. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 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]
|
|
|
|
|
|
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. H. A. Gouveia, M. A. Christoffolete, C. R. Zaitune, J. M. Dora, J. W. Harney, A. L. Maia, and A. C. Bianco Type 2 Iodothyronine Selenodeiodinase Is Expressed throughout the Mouse Skeleton and in the MC3T3-E1 Mouse Osteoblastic Cell Line during Differentiation Endocrinology, January 1, 2005; 146(1): 195 - 200. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. G. J. M. Kuiper, F. Wassen, W. Klootwijk, H. van Toor, E. Kaptein, and T. J. Visser Molecular Basis for the Substrate Selectivity of Cat Type I Iodothyronine Deiodinase Endocrinology, December 1, 2003; 144(12): 5411 - 5421. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE |
