INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

The type 1, 2, and 3 iodothyronine deiodinases (D1, D2, and D3, respectively)¹ constitute a family of oxidoreductases that catalyze the removal of iodide from the outer (5'-iodine) and/or inner (5-iodine) ring of thyroid hormones (1). These enzymes have been highly conserved during vertebrate evolution, and of particular note, all deiodinases identified to date contain the uncommon amino acid selenocysteine in the midregion of the polypeptide chain. Using site-directed mutagenesis, this residue has been demonstrated to be essential for efficient catalysis (2-5).

In addition to the selenocysteine, each of the three deiodinase isoforms contains 6 or 7 cysteine residues (Fig. 1). Of particular interest is the cysteine at position 124 of the D1. This residue is vicinal to the selenocysteine at position 126 and also present in the analogous position in all D3 isoforms but not in the D2 where the corresponding amino acid is an alanine. Also of interest is Cys¹⁹⁴ in the D1, which has been conserved in all three deiodinase isoforms. The conservation of these cysteine residues suggests that they may play an important role in the catalytic properties of these enzymes. In particular, the presence or absence of the cysteine at position 124 could be responsible in part for the unique catalytic properties of these isoforms. For example, the D1 and D2 both catalyze 5'-deiodination, yet they differ markedly in terms of reaction kinetics, substrate and cofactor utilization, and sensitivity to inhibitors (6).

View larger version (6K): [in this window] [in a new window]   Fig. 1.   Location of cysteine and selenocysteine residues in the rat D1, D2, and D3. These proteins have predicted molecular masses of 29, 30, and 32 kDa, respectively.  and , cysteine residues; , selenocysteine.

In addition to acting as active site nucleophiles, a role that appears to be subserved by selenocysteine in the deiodinases, cysteine residues in other enzymes have a number of important functions. These include optimal positioning of substrates within the catalytic cleft (7), as well as the transfer of reducing equivalents from in vivo reducing systems to the active site (8). For example, the presence of a cysteine vicinal to a selenocysteine in selenoprotein P_A of the bacterial glycine reductase complex has been implicated in the activation of that enzyme by thioredoxin-like cofactors (9). Finally, cysteines have been demonstrated to act as important structural components in certain enzymes through the formation of disulfide bridges (10) or as a result of the hydrophobic properties of the cysteine side chain (11). Such actions serve to stabilize protein confirmation and/or promote homodimeric or multimeric assemblies necessary for catalytic activity. The present studies were thus conducted to evaluate systematically the roles of Cys¹²⁴ and Cys¹⁹⁴ in D1 activity.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Mutagenesis-- The G21 full-length rat D1 cDNA (kindly provided by Drs. M. Berry and P. R. Larsen) was subcloned into the pcDNA3 mammalian expression vector (Invitrogen, San Diego, CA). Site-directed mutagenesis was then performed in this vector using the Transformer Mutagenesis kit (CLONTECH, Palo Alto, CA) according to the manufacturer's instructions. The Cys¹²⁴ and Cys¹⁹⁴ codons were individually changed to code for alanine. A double C124A/C194A mutant was also prepared. Confirmation of the mutated sequence, as well as the integrity of the remainder of the coding region, was performed in all cases using an automated sequencing system with fluorescent dye terminators (Applied Biosystems, Foster City, CA).

Transfection and Expression Studies in COS-7 Cells-- COS-7 cells were grown and maintained in Dulbecco's modified Eagle's medium supplemented with 10% iron-supplemented calf serum (Sigma). Cells (10⁷ in 425 µl of HEPES-buffered saline) were transfected by electroporation using a Bio-Rad (Hercules, CA) Electroporator at settings of 280 volts and 960 microfarads with 20 µg of DNA consisting of the pcDNA3 plasmid containing either the G21 (D1), the C124A mutant (A/C), the C194A mutant (C/A), or the C124A/C194A double mutant (A/A). Cells transfected with the pcDNA3 plasmid that contained no insert served as controls. In some experiments, the amount of D1 or A/A plasmid transfected was varied from 0.5 to 20 µg. In all such cases, sufficient control pcDNA3 plasmid was included such that the total amount of plasmid transfected equaled 20 µg. As noted previously, COS-7 cells do not express endogenous 5'- or 5-deiodinase activity (12). After transfection, cells were plated onto 150-mm culture dishes.

Experimental Protocols-- At 48 h after transfection, cell monolayers were washed twice with sterile phosphate-buffered saline and then cultured for an additional 24 h in serum-free Dulbecco's modified Eagle's medium to which was added [¹²⁵I]rT₃ or [¹²⁵I]T₄, radiolabeled at the 5'-position, in a final concentration of 0.2-0.9 nM. For the determination of the rate of in vivo 5'-deiodination, 50-µl aliquots of medium were sampled in duplicate immediately after the addition of radiolabeled compounds (basal samples) and at additional time points during the next 24 h. 10 µl of a 4% bovine serum albumin solution plus 100 µl of 20% trichloroacetic acid were immediately added to each aliquot of medium. This mixture was then centrifuged at 12,000 × g for 5 min, and the supernatant was applied to an AG50W-X8 ion-exchange column as described previously (13). Radiolabeled iodide was eluted from the column by washing with 2.5 bed volumes of 10% acetic acid and counted. The iodide formed was then calculated as the percentage of the total counts present in the 50-µl aliquot of medium minus the percentage of iodide present in the basal (time 0) sample.

Determination of in Vitro 5'-Deiodinase Activity-- 5'-Deiodinase activity in COS-7 cell sonicates was determined as described previously (12) using 1-3 nM [¹²⁵I]rT₃ as substrate, and dithiothreitol (DTT), glutathione (GSH), or a reconstituted thioredoxin system (which included thioredoxin, NADPH, and thioredoxin reductase as described previously (14)) as cofactors. In other experiments, deiodinase activity in COS-7 cell sonicates was determined in the absence or presence of 6-n-propyl-2-thiouracil (PTU; 0.3-100 µM) or aurothioglucose (0.01-10µM). In kinetic experiments, rT₃ concentrations of 0.3-1000 nM were used as substrate with varying concentrations of DTT or GSH or the reconstituted thioredoxin system as cofactors.

Immunoblotting-- SDS-polyacrylamide gel electrophoresis and electrotransfer of proteins to Immobilon membranes (Millipore Corporation, Bedford, MA) was performed as described previously using 50 or 75 µg of COS-7 cell sonicate protein (15). On all blots, 50 µg of liver homogenates derived from hyperthyroid and hypothyroid rats were included as controls. A polyclonal antiserum prepared against the carboxyl-terminal half of the rat D1 protein was used for immunoblotting. The preparation and characterization of this antiserum have previously been described (15). Blotting procedures were carried out essentially as described previously (15), except that the final detection step made use of Vistra^TM ECF Western blotting detection reagents from Amersham Pharmacia Biotech. The fluorescent signal was detected using a Molecular Dynamics Fluorimager 575, and the signal was quantitated with the ImageQuant program (Molecular Dynamics, Sunnyvale, CA) on a Macintosh computer. Quantitation units are arbitrary. Sonicate protein content was determined by the method of Bradford (16) with reagents obtained from Bio-Rad.

Calculations and Statistical Analysis-- In vitro and in vivo 5'-deiodinase activities were corrected where indicated using the expressed levels of the D1 and mutant proteins as quantified by Western analysis. These protein expression data were normalized within and between experiments by comparison with the level of D1 observed in a hyperthyroid rat liver control homogenate sample that was included on all blots.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

The levels of expression of the native and mutant D1 enzymes were determined by immunoblot analysis utilizing a polyclonal antiserum directed against the carboxyl-terminal half of the D1 protein (15). Representative results are shown in Fig. 2. A 27-kDa species was identified in sonicates from cells transfected with the native D1 plasmid or with the A/C, C/A, or A/A mutant constructs. These proteins corresponded in both size and approximate abundance to that observed in a control liver homogenate derived from a hyperthyroid rat. A second liver homogenate from a methimazole-treated rat was also included as a control on all the blots, and as previously demonstrated (15), the D1 protein content is significantly decreased under hypothyroid conditions. No signal was detected in sonicates of cells transfected with the empty pcDNA3 plasmid, consistent with the lack of endogenous deiodinase activity in this cell line (12). The relative amounts of native or mutant deiodinase protein expressed in COS-7 cells transfected with 20 µg of recombinant plasmid were determined in six separate experiments. A trend toward lower levels of immunoreactive protein were noted in cells transfected with the C194A constructs (i.e. the C/A and A/A). In the case of the A/A construct, expressed protein levels were 38% lower (p < 0.02) than in cells expressing the native D1.

View larger version (80K): [in this window] [in a new window]   Fig. 2.   Western blot analysis of COS-7 cell sonicates probed with a polyclonal antiserum directed against the carboxyl-terminal half of the native rat D1. Results of a representative experiment where 50 µg of sonicate protein or of liver homogenates derived from a hyperthyroid or a hypothyroid rat were loaded per lane.

Effects of Cysteine Mutations on in Vitro Determined 5'-Deiodinase Activity-- The D1 demonstrates ping-pong reaction kinetics whereby both an iodothyronine and a thiol cofactor serve as reactants with the enzyme (21). The effects of the C124A and C194A mutations on deiodinase activity were determined by kinetic analysis using rT₃ and DTT over a broad range of concentrations (15-1000 nM and 0.2-30 mM, respectively). Secondary plots yielded the limiting V_max and K_m values shown in Table I. Mean values and the individual determinations from two experiments are given. Velocity data were corrected based on the amount of D1 or mutant protein present in the cell sonicates as determined by immunoblots. The native D1 manifested limiting K_m values for DTT and rT₃ of 0.62 mM and 0.25 µM, respectively. The most striking observation was a 7-11-fold increase in the K_m of DTT for both of the C124A mutants (A/C and A/A) but not with the C194A protein (C/A). The C/A mutant, however, did manifest a 3-fold higher K_m of rT₃, a finding not noted with the A/A double mutant. Limiting V_max values for the native and mutant constructs differed by up to 2.5-fold with the A/C mutant showing the lowest maximal activity.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Comparison of GSH- and thioredoxin-stimulated 5'-deiodinase activities determined in COS-7 cell sonicates transfected with the native D1 or the various cysteine mutants. A, kinetic analysis using rT3 as substrate and 20 mM GSH. Data are plotted on an Eadie-Hofstee plot where Km values are represented by the negative value of the slope, and Vmax is determined by the y intercept. Specific velocity is expressed per unit of D1 or mutant protein as determined by immunoblot, and the rT3 concentration is in nanomolar. B, pooled results of four experiments quantitating 5'-deiodinase activity utilizing rT3 (1.5 nM) and GSH (20 mM) as substrate and cofactor, respectively. For each experiment, activities were first corrected for the levels of D1 or mutant protein as determined by immunoblot and then normalized to the activity level determined in the native D1-expressing cell sonicate, which was assigned a value of 1. C, pooled results of four experiments quantitating 5'-deiodinase activity utilizing rT3 (1.5 nM) and the thioredoxin system as substrate and cofactor, respectively. Data are expressed and analyzed as indicated in panel B. *, p < 0.02 versus D1; , p < 0.001 versus D1.

Effects of Cysteine Mutations on 5'-Deiodination in Intact Cells-- The results of the above studies performed in broken cell preparations suggested that mutations of the D1 at the Cys¹²⁴ and Cys¹⁹⁴ positions alter the properties of the D1 and impair its activity, particularly when GSH or thioredoxin are utilized as cofactors. To test the effects of these mutations on enzyme activity in vivo, transfected COS-7 cells were incubated in serum-free medium with [¹²⁵I]rT₃ for 24 h before harvesting. At various times after the addition of substrate, the medium was sampled, and the amount of radioiodine generated was quantitated.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Analysis of 5'-deiodination in intact transfected COS-7 cells. A, the fraction of [125I]rT3 substrate deiodinated as a function of time during the 24-h incubation period. Cells were incubated in serum-free medium with an initial rT3 concentration of 0.3 nM. Values are expressed as the fraction of rT3 deiodinated per dish and are not corrected for levels of enzyme expression. The results of a representative experiment are shown. B, in vivo 5'-deiodinase activity calculated as the rate of iodide generation from [125I]T4 during a 24-h incubation period. Values are corrected for nonspecific iodide generation observed in the cells transfected with the empty pcDNA3 plasmid and for the levels of expression of the D1 or mutant proteins as determined by immunoblot analysis in cell sonicates harvested at the end of the incubation period.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Comparison of in vitro versus in vivo determined 5'-deiodinase activity expressed per mg of cellular protein in COS-7 cells transfected with 2, 8, or 20 µg of the native D1 expression plasmid. 48 h following transfection, cells were incubated in serum-free medium with 0.3 nM [125I]rT3, and the rate of iodide generation was quantitated by sampling the medium at 0, 2, 4, 6, 10, and 24 h. Data obtained at the 6- and 10-h sampling times are shown plotted against the 5'-deiodinase activity determined in sonicates of the same cells harvested at the end of the 24-h incubation period. In vitro activity was determined using 1 nM rT3 and 20 mM GSH as the substrate and cofactor.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Comparison of in vitro versus in vivo determined 5'-deiodinase activity in COS-7 cells transfected with 0.5, 1, 2, or 8 µg of the native D1 or the A/A expression plasmids. A, levels of expressed D1 or A/A protein, as determined by immunoblot, are plotted against the amount of plasmid transfected. B, the rate of in vivo deiodination expressed per mg of cellular protein as a function of the amount of the native D1 or the A/A plasmid transfected. In vivo deiodination was determined as in the experiment described in the legend to Fig. 5. Data from the 6-h time point are shown. This pattern is representative of the results obtained at other sampling times. C, comparison of the rate of GSH-stimulated 5'-deiodinase activity, determined as in Fig. 5, in sonicate preparations from cells transfected with 0.5 µg of the native D1 plasmid or 2 µg of the A/A construct. Cells were harvested at the end of the 24-h incubation period used to determine the in vivo deiodinase activity. Activity is expressed per mg of cellular protein. Levels of expression of the native D1 and A/A mutant enzyme were equivalent under these conditions in this experiment (see panel A and text). D, comparison during a 24-h incubation period of the extent of in vivo deiodination, determined as in Fig. 5, in cells transfected with 0.5 µg of the native D1 plasmid or 2 µg of the A/A construct. Deiodination is expressed per mg of cellular protein.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In the present studies, we have utilized site-directed mutagenesis to explore the function of the Cys¹²⁴ and Cys¹⁹⁴ in the D1 by changing these residues to alanine. We chose to substitute alanine for the cysteines because the native D2 contains alanine in the position corresponding to Cys¹²⁴ in the D1 (12). We demonstrate that mutagenesis of the Cys¹²⁴ and Cys¹⁹⁴ alter the kinetic characteristics and impair the 5'-deiodinase activity of the D1 as determined using in vitro assays but do not alter the function of this enzyme in intact cells.

The effects of C124A and C194A on 5'-deiodinase activity in cell sonicates are complex and dependent on the cofactor system utilized in the assay. A striking finding was a decrease in the efficiency by which the C124A mutants utilize DTT as a cofactor. This was manifest by a 7-11-fold increase in the K_m of DTT for the A/C and A/A mutants. This finding confirms the recent report of Sun et al. (23) and is consistent with their thesis that the Cys¹²⁴ forms a sulfo-seleno adduct that can be efficiently reduced by a small dithiol compound such as DTT. However, the presence of the Cys¹²⁴ appears to confer no such advantage when the monothiol GSH is used as cofactor; the concentration of GSH resulting in half-maximal velocity was 5 mM for the native D1 and all of the mutants tested. Other alterations in enzyme function observed with the cysteine mutants were minimal or inconsistent when DTT was used as cofactor. For example, although the V_max values for the A/C and C/A mutants were reduced 51 and 31%, respectively, no decrease in this parameter was noted for the A/A mutant, suggesting that the maximal deiodinating capacity of the D1 in the presence of DTT is dependent to only a minimal extent on the presence of these cysteines.

The activity of the cysteine mutants in the presence of PTU and aurothioglucose over a broad range of concentrations was unchanged from that of the native D1. Because we did not perform a formal kinetic analysis of the effects of these inhibitors, subtle changes in K_i values, such as the 2-fold increase in the K_i of PTU described by Sun et al. (23) for the C124A mutant, could have been missed. However, it is apparent that the mutant constructs, like the native D1, are very susceptible to inhibition by PTU and aurothioglucose

The cysteine mutations had more dramatic effects on 5'-deiodinase activity when GSH or a reconstituted thioredoxin cofactor system were used in the broken cell assays. K_m values were largely unchanged; however, marked decreases in V_max were noted with both the C124A and C194A substitutions, and these effects appeared to be additive in that the A/A mutant showed the greatest impairment in activity. This dichotomy in the effects of cysteine mutations on enzyme reactivity with different thiols has previously been noted. For example, Mao et al. (8) observed that mutagenesis of specific cysteines in ribonucleotide reductase resulted in marked impairment of enzyme activity in the presence of thioredoxin but did not alter maximal reactivity with DTT. These investigators speculated that these cysteines were necessary for the transfer of reducing equivalents from thioredoxin to the active site and that the unimpaired reactivity with DTT was secondary to this small compound being able to directly access the catalytic site. Similar circumstances may hold for the D1 with regard to the Cys¹²⁴ and/or Cys¹⁹⁴. Alternatively, mutagenesis of these cysteines may alter other structural features of the enzyme such that GSH and thioredoxin are limited in their ability to effect reduction of the active site selenocysteine.

Intact COS-7 cells transfected with the various D1 constructs actively deiodinated both rT₃ and T₄. For all constructs, the appearance of iodide in the incubation medium was a linear function of time, indicating that the rate of deiodination was constant throughout the 24-h incubation period. This is somewhat surprising given that typically up to 95% of the rT₃ was deiodinated during the period of observation. The maintenance of a constant reaction rate in the face of such a large decrease in substrate concentration suggests that the enzyme was operating at or near its V_max during the entire incubation period. A comparison of the K_m values determined in vitro with the medium substrate concentration used in these experiments suggests that this was likely the case. Thus, in these and prior studies (14), we have demonstrated that the D1 manifests a K_m of rT₃ of approximately 0.2-1.5 nM when assayed in broken cell preparations using either no added cofactors or potential native cofactors such as GSH or thioredoxin. Such K_m values, however, are based on total rT₃ concentrations. K_m values calculated as a function of free substrate concentrations in the sonicate mixture would be significantly lower due to protein binding of the rT₃. In the present studies, rT₃ at concentrations of 0.3-0.9 nM was added to COS-7 cells cultured in serum-free medium. These concentrations thus reflect the free rT₃ levels. Such concentrations, if also indicative of intracellular free hormone levels, as suggested by the studies of Mendel et al. (24), should be saturating to the D1 in the presence of endogenous cofactors.

A novel and unexpected observation during the course of these studies was that the amount of deiodinase expressed in COS-7 cells was not always the limiting factor in determining the rate of in vivo rT₃ metabolism. Carrier-mediated transport of iodothyronines has been described in several cell systems (25-27) and could limit substrate availability for deiodination (28). However, even in systems where evidence for specific transporters has been found, the total uptake of various iodothyronines is usually rapid and largely nonsaturable over a wide range of concentrations (29, 30). Thus, it is possible that other factors (e.g. cofactor availability) limit the rate of deiodination in intact COS-7 cells.

A second striking finding in these studies was that cysteine mutations that dramatically impair GSH- and thioredoxin-stimulated 5'-deiodinase activity in broken cell preparations have no effect on reaction rates in intact cells. Limitations in cellular hormone uptake, if present, are unlikely to explain these discrepant results; given that the impairment in in vitro determined activity of the mutants involves a decrease in V_max, iodide production from intact cells expressing these less active constructs should have been less than control cells expressing the native D1 at any given substrate level. Thus, activities determined in vitro do not necessarily predict the behavior of the deiodinases in vivo, where additional protein-protein or enzyme-cofactor interactions may take place. In this regard, the present results suggest that GSH and thioredoxin are not the major native thiol cofactors for the D1. A similar conclusion regarding GSH was reached by Sato and Robbins (31). It is noteworthy that the D2 and D3 also appear to utilize other cofactors; in in vitro assay systems the D2 and D3, like the D1 cysteine mutants, are poorly reactive with GSH and thioredoxin.² It is plausible that native dithiols such as dihydrolipoic acid or dihydrolipoamide may serve as cofactors in vivo for these enzymes (32).

An important technical issue relates to the finding that the levels of immunodetectable protein for the Cys¹⁹⁴ mutants were somewhat less than for the native D1 and the A/C. The polyclonal antiserum used in these studies was produced using as antigen the carboxyl-terminal half of the native rat D1 protein beginning immediately distal to the selenocysteine residue (15). If the C194A mutation alters an important antigenic epitope on the protein, then the C/A and A/A mutants might manifest reduced reactivity on immunoblotting. This would result in an underestimation of the amount of C/A and A/A protein being expressed. Alternatively, the C/A and A/A mutants may undergo more rapid turnover in COS-7 cells due to this structural alteration, in which case the decreased immunoreactive signals accurately reflect protein levels in the cell. However, neither scenario negates the basic dichotomy between the in vivo and the in vitro activity of these mutant constructs.

In summary, we have demonstrated that mutation of the Cys¹²⁴ and/or Cys¹⁹⁴ alters the reactivity of the D1 in vitro, especially when GSH and thioredoxin are utilized as cofactors. However, these residues do not appear to be determinants of the unique properties of the D1, namely its sensitivity to inhibition by propylthiouracil or thioredoxin or its high K_m kinetic behavior in the presence of DTT. Furthermore, mutagenesis of Cys¹²⁴ and Cys¹⁹⁴ does not impair the activity of the D1 in intact cells, suggesting that these residues have neither critical catalytic or structural roles in determining deiodinase activity.