Interaction of thyroid peroxidase with concanavalin A covalently coupled to agarose.

We have investigated the interaction between concanavalin A-agarose (Con A-agarose) and thyroid peroxidase, an integral membrane protein found in the 105,000 X g, 1-h particulate fraction of thyroid tissue. An intact form of porcine thyroid peroxidase was obtained by solubilization with the nonionic detergent Triton X-100 and two fragmented, hydrophilic forms of the enzyme were prepared by trypsin treatment of the membrane. The three types of thyroid peroxidase bind to Con A-agarose and can be eluted with alpha-methyl-D-mannoside. The alpha-methyl-D-mannoside eluate of the most purified thyroid peroxidase preparation has been analyzed by polyacrylamide gel electrophoresis. Peroxidase activity corresponds with a glycoprotein band. The binding of thyroid peroxidase to Con A-agarose can be inhibited by sugars in the following order: alpha-methyl-D-mannoside greater than D-mannose greater than alpha-methyl-D-glucoside greater than D-glucose greater than D-galactose. This order of specificity is typical of Con A-sugar interactions. Furthermore, inactivation of the carbohydrate binding site of Con A by demetallization greatly reduces the extent of thyroid peroxidase binding. Reactivation of the carbohydrate binding site by the addition of Ca2+ and Mn2+ to demetallized Con A-agarose restores thyroid peroxidase binding. These and other experiments suggest that htyroid peroxidase is, like several other peroxidases, a glycoprotein. In addition, the interaction between thyroid peroxidase and Con A-agarose may provide a new purification tool for thyroid peroxidase.

concanavalin A-agarose (Con A-agarose) and thyroid peroxidase, an integral membrane protein found in the 105,000 x g, l-h particulate fraction of thyroid tissue. An intact form of porcine thyroid peroxidase was obtained by solubilization with the nonionic detergent Triton X-100 and two fragmented, hydrophilic forms of the enzyme were prepared by trypsin treatment of the membrane. The three types of thyroid peroxidase bind to Con A-agarose and can be eluted with cu-methyl-D-mannoside.
The a-methyl-D-mannoside eluate of the most purified thyroid peroxidase preparation has been analyzed by polyacrylamide gel electrophoresis. Peroxidase activity corresponds with a glycoprotein band. The binding of thyroid peroxidase to Con A-agarose can be inhibited by sugars in the following order: a-methyl-D-mannoside > D-mannose > cu-methyl-D-glucoside > D-glucose > D-galactose.
This order of specificity is typical of Con Asugar interactions.
Furthermore, inactivation of the carbohydrate binding site of Con A by demetallization greatly reduces the extent of thyroid peroxidase binding. Reactivation of the carbohydrate binding site by the addition of Ca2+ and Mn2+ to demetallized Con A-agarose restores thyroid peroxidase binding.
These and other experiments suggest that thyroid peroxidase is, like several other peroxidases, a glycoprotein. In addition, the interaction between thyroid peroxidase and Con A-agarose may provide a new purification tool for thyroid peroxidase. Thyroid peroxidase is an integral membrane enzyme (1) which plays a key role in the biosynthesis of thyroid hormones because of its ability to catalyze the iodination of tyrosine residues in thyroglobulin. The properties of thyroid peroxidase and the mechanism of iodination have not been clearly elucidated.
The most widely used procedure for solubilizing thyroid peroxidase has been proteolysis (2)(3)(4)(5)(6)(7)(8) EXPERIMENTAL PROCEDURES Preparation of Triton X-100~solubilized Thyroid Peroxidase -The 105,000 x g, l-h particulate fraction from frozen porcine thyroids (Swift & Co., Chicago, 111.) was prepared as previously described (12) and stored at -20". In order to release occluded cytoplasmic proteins, the particulate fraction was suspended in 0.1 rnM KI, mixed for 3 to 4 h on a Dubnoff shaker at 5", and centrifuged at 105,000 x g, 1 hr, 5". Solubilization of thyroid peroxidase by 0.1% Triton X-100 and at a particulate protein concentration of 1.5 mglml was conducted as previously described (1). The supernatant, which contained about 65% of the thyroid peroxidase activity, was concentrated by ultrafiltration.
The specific activity of the Triton X-100 extract ranged between 1.2 and 3.6 pmol of oxidized guaiacol/min/mg of protein. Remaining active groups were blocked by treatment with 1 M glycine, pH 9, for 2 h (22). The Con A-agarose was finally washed with 0.15 M NaCl and stored at 5". The efficiency of the coupling reaction, as judged from the extinction of the washings at 280 nm, ranged from 62 to 80% for six batches of Con Aagarose. For a binding control, "activated" agarose was prepared by removing a portion of the cyanogen bromide-treated agarose before addition of Con A. This sample was then washed in exactly the same manner as the Con A-agarose. Demetallized Con A-agarose was prepared by treating Con A-agarose with 0.1 N HCl for 30 min followed by exhaustive dialysis against H,O. Sumner and Howell (23) first described the demetallization of Con A in 1936, and more recently Kalb and Levitzki (24) reported the details of the procedure that was followed here.
Conditions of Binding and EL&ion-These are described in the legends to the tables and figures.
Preparation of Immobilized Anti-porcine Thyroglobulin -Thyroglobulin, the major glycoprotein in thyroid tissue, is found in many thyroid peroxidase preparations.
In order to remove thyroglobulin from the peroxidase preparations, anti-porcine thyroglobulin was prepared and immobilized on glass beads. To determine the amount of anti-thyroglobulin-CPG necessary to remove thyroglobulin from the Triton X-100 peroxidase preparation, a trace amount of 1251-thyroglobulin was added to the Triton X-100 extract and then various amounts of antithyroglobulin-CPG were added to the tubes. The mixtures were shaken intermittently over a period of 3 h at room temperature and after centrifugation, the pellets were counted in a Packard y spectrometer.
These experiments showed that 4 mg of anti-thyroglobulin-CPG in a 0.5% (w/v) suspension absorbed a maximum of 85% of the thyroglobulin from a Triton X-100 extract containing 0.1 mg of protein.
The extent of thyroglobulin removal was qualitatively determined by double immunodiffusion against anti-porcine thyroglobulin.
No precipitin line was observed between the immunoadsorbed Triton X-100 extract and anti-porcine thyroglobulin, whereas a strong line developed between an untreated Triton X-100 extract and the antiserum.
Thyroglobulin was quantitatively measured by radioimmunoassay for porcine thyroglobulin, using a procedure similar to the radioimmunoassay described by Van Herle et al. (26), except that porcine antigen and anti-porcine thyroglobulin were substituted for human thyroglobulin and anti-thyroglobulin. Assays of the Triton X-100 extract before and after immunoadsorption indicated that 85% of the thyroglobulin was adsorbed from the Triton X-100 extract by anti-thyroglobulin-CPG.
(Use of the radioimmunoassay also showed that no detectable anti-thyroglobulin leaked off the glass beads since the supernatant from the anti-thyroglobulin-CPGimmunoadsorbed sample did not bind ""I-thyroglobulin.) The immunoadsorbed Triton X-100 extract is referred to as the "thyroglobulindepleted thyroid peroxidase." Analytical Methods -Thyroid peroxidase activity was routinely measured by the guaiacol assay (27, in the absence of Na dodecyl SO, was conducted essentially as described by Davis (33). The separating gel was 10% acrylamide; the pH of the separating gel was 10 and the pH of the reservoir buffer was 9.0. Electrophoresis was initiated at 1 mA/gel; when the tracking dye was concentrated in the stacking gel, the current was increased to 3 mA/gel. Electrophoresis was terminated when the tracking dye was within a few millimeters of the bottom of the gel. Peroxidase activity was detected by transferring the gel to the standard guaiacol/H,O, solution used in the spectrophotometric assays. An orange-colored band developed within 10 min.

Binding
of Thyroid Peroxidase to Con A-Agarose-As shown in Table I, detergent-solubilized, intact thyroid peroxidase binds to Con A-agarose. After incubation with Con Aagarose, only 3% of the thyroid peroxidase activity was recovered in the supernatant. In contrast to this, full recovery of thyroid peroxidase activity was obtained from "activated" agarose, thereby indicating that thyroid peroxidase binds to Con A and not to an agarose product(s) resulting from the cyanogen bromide activation reaction or the procedure used to block unreacted groups.
It is also of interest that, even though essentially all of the peroxidase binds to Con A-agarose, more than 50% of the protein in the Triton X-100 extract does not bind. Goldstein et al. (15) have shown that a-methyl-n-mannoside is a potent inhibitor of the Con A-dextran precipitation reaction.
When a-methyl-n-mannoside (100 mg/ml) and thyroid peroxidase are mixed and then added to Con A-agarose, none of the thyroid peroxidase binds to Con A-agarose (  One milliliter (settled volume) of the indicated agarose preparation was washed three times with 10 ml of 10 rnM phosphate buffer (pH 6.01, 0.1% Triton X-100, 1 mM CaCl,, 1 rnr,r MnCl,, 0.1 rnM KI (equilibrating buffer). One milliliter of Triton X-100 extract, which contained 2.4 units of thyroid peroxidase activity, 1.9 mg of protein, and 1.5 ml of the equilibrating buffer (or, in Experiment 4, 1.5 ml of the equilibrating buffer containing 100 mg/ml of a-methyl-n-mannoside) were added to each of the washed agarose gels. The suspensions were mixed at 4" for 2 h and centrifuged at 2000 rpm in a Sorvall HL-8 swinging bucket rotor at 5" for 10 min, and the packed gels were washed with 2 ml of equilibrating buffer (or, in Experiment 4, with 2 ml of equilibrating buffer containing 100 mg/ml of e-methyl-n-mannoside  Table II, the binding of thyroid peroxidase to Con A-agarose can be specifically inhibited by sugars. a-Methyl-nmannoside (10 mglml) inhibits binding by 50%; n-mannose and cY-methyl-n-glucoside are also effective inhibitors, n-glucose is less effective, and n-galactose has little effect on the binding of thyroid peroxidase. Goldstein and coworkers (15) studied the specificity of the carbohydrate binding site of Con A by investigating the ability of various sugars to inhibit the dextran-Con A precipitation reaction. The sugars that were tested in our study inhibit the interaction between thyroid peroxidase and Con A in the same order of specificity as in the dextran-Con A interaction investigated by Goldstein et al. (15). We interpret this high degree of specificity as further evidence that thyroid peroxidase binds to the carbohydrate binding site of Con A. Katzen and Soderman (34) found a similar order of inhibition by sugars for the interaction of adipocyte receptors with the carbohydrate binding sites of Con A-agarose.
Studies with Demetallized Con A-Agarose -In 1936, Sumner and Howell (23) reported on the role of divalent metal ions in the reversible inactivation of jack bean hemagglutinin. Subsequent studies (24, 35) have confirmed the observation that removal of divalent cations from Con A by treatment with dilute acid destroys the ability of Con A to interact with polysaccharides and that the addition of Ca'+ and Mn"+ can restore carbohydrate binding. However, demetallization does not destroy the ability of Con A to bind hydrophobic molecules such as the dye rose bengal." As shown in Fig. 1, when increasing amounts of thyroid peroxidase are added to a prepa-' R. H. Clew, personal communication. Con A-agarose was washed with 10 mM phosphate buffer (pH 6.01, 0.1% Triton X-100, 1 mM CaCl,, 1 mM MnCl,, 0.1 mM KI (equilibrating buffer) as described in Table I. One-half  milliliter  of Triton X-100  extract, which contained 1.47 units of thyroid peroxidase activity, was mixed with 2 ml of a 62.5 mM sugar solution in equilibrating buffer (or, in Sample 1, with 2 ml of equilibrating buffer) and the resulting solution was added to 1 ml (settled volume) of the washed Con A-agarose.
The suspensions were mixed at 4" for 1.5 h and centrifuged at 2000 rpm in a Sorvall HL-8 swinging bucket rotor at 5" for 10 min. The supernatants were assayed for thyroid peroxidase activity.
The amount of peroxidase bound to Con A-agarose was determined by subtracting the peroxidase recovered in the supernatant from the applied peroxidase.
Sugar (10 mglml binds is probably due to incomplete demetallization of the Con A-agarose.) However, when the carbohydrate binding sites are restored by the addition of CaZ+ and Mn2+ to demetallized Con A-agarose, thyroid peroxidase binding increases dramatically and reaches a saturating level, much greater than that obtained in the absence of Ca2+ and Mn2+.
Elution of Thyroid Peroxidase from Con A-Agarose with CY-Methyl-D-mannoside - Table III shows the ability of a-methyln-mannoside to release detergent-solubilized thyroid peroxidase from Con A-agarose. Approximately 60% of the thyroid peroxidase activity was recovered in the combined supernatants and most of the recovered activity was found in the first a-methyl-nmannoside supernatant.
A control experiment showed that a-methyl-n-mannoside is essential for thyroid peroxidase release since only 4% of the applied activity was eluted in the absence of a-methyl-nmannoside.
Experiments conducted over a period of 1 year indicate that the amount of thyroid peroxidase released from Con A-agarose by a-methyl-n-mannoside can vary from 30 to 95%, depending on the batch of Con A-agarose and on the thyroids used as starting material for the preparation of the thyroid particulate fraction. From the same batches of Con A-agarose, we found that the elution of a known glycoprotein, porcine thyroglobulin (Sigma, type II), varied from 30 to 60%. The consistency of glycoprotein elution from lectin-agarose systems has not been discussed frequently, although a variable elution of a-glutamyltransferase by mannose has been reported (36), and yields ranging from 30% (37) to 75% (38) of electric eel acetylcholinesterase have been described. The recovery of thyroid peroxidase from Con A-agarose compares favorably with the yields obtained for other proteins from lectin-agarose systems. For   TABLE III Release of thyroidperoxidase activity, protein, and sugar from con Aagarose by a-methyl-D-mannoside Two milliliters (settled volume) of Con A-agarose was washed three times with 5-ml portions of 10 mM phosphate buffer (pH 6.01, 0.1% Triton X-100, 1 rnM CaCl,, 1 m&i MnCl,, 0.1 mM KI (equilibrating buffer).
One milliliter of Triton X-100 extract was mixed with 4 ml of equilibrating buffer and added to the washed Con A-agarose. The suspension was mixed for 1.5 h at 4", centrifuged at 2000 rpm in a Sorvall HL-8 swinging bucket rotor at 5" for 10 min, and the supernatant was decanted. The gel was washed twice with 5-ml portions of 10 mM phosphate buffer (pH 6.01, 0.1% Triton X-100, 0.1 mM KI. The first supernatant and the two supernatants from the washes were combined and were labeled "not bound fraction." Five milliliters of or-methyl-n-mannoside (100 mg/ml) in 0.1 M Tris/HCl (pH 8.91, 0.1% Triton X-100, 0.1 mM KI were added to the gel and the suspension was mixed overnight at 4". A supernatant was obtained by centrifugation as described above and the gel was washed twice with 5 ml of the a-methyl-n-mannoside solution, as described above. The three (Ymethyl-n-mannoside-eluted fractions were analyzed separately and the figures in the  (20) have been recovered from Con A-agarose.
The fractions described in Table III were subjected to polyacrylamide gel electrophoresis in the presence of Na dodecyl S04, and the results are shown in Fig. 2. It is not possible to identify thyroid peroxidase in these electrophoresis gels because the peroxidase is inactivated by dithiothreitol and Na dodecyl SO,. The fraction eluted with a-methyl-n-mannoside contains three readily visible glycoprotein bands and several faint bands which are not visible in the photograph.
The protein bands in the a-methyl-n-mannoside eluted fractions that stain with Coomassie blue but do not stain with the PAS reagent may represent glycoproteins which are low in sialic acid content. It is generally accepted that sialic acid-rich glycoproteins readily stain with PAS (42,43). However, not all glycoproteins react with the PAS reagent. Adair and Kornfeld (44) have reported on a glycoprotein receptor from human erythrocytes for Ricinus communis agglutinin which does not stain with PAS reagent. This glycoprotein is rich in N-acetylglucosamine but almost devoid of sialic acid.
In the fraction which does not bind to Con A, a strong PAS band is present close to the tracking dye. The position of this fast moving band is similar to that reported by Lenard (45) and Fairbanks et al. (32) and may be a glycolipid. This component may account for most of the sugars in the fraction that does not bind to Con A (Table III). Some glycoproteins, which do not have an affinity for Con A-agarose and which do not react with the PAS stain, may also be present in this fraction.
Effect ofpH on a-Methyl-D-mannoside-dependent Release of Thyroid Peroxidase -As shown in Fig. 3, the optimum pH for the release of thyroid peroxidase with a-methyl-n-mannoside is about 9. Under these conditions, 60% or more of the peroxidase is released. In the absence of a-methyl-n-mannoside only 6% or less of the peroxidase is eluted at pH 8.9, thereby indicating that the increase in pH is not sufficient to release FIG The decrease in the activity curve above pH 9.5 may be due in part to the slight decrease in activity of peroxidase in alkaline medium (10) rather than a decrease in amount of peroxidase eluted above pH 9. Znteraction of Con A-Agarose with Thyroid Peroxidase Preparation Treated with Zmmobilized Anti-thyroglobulin-Thyroglobulin, the major glycoprotein in thyroid tissue, is found in many peroxidase preparations. As shown in Table IV, when 85% of the thyroglobulin is removed from the detergentsolubilized peroxidase preparation by immunoprecipitation, none of the peroxidase activity is lost. Furthermore, the thyroglobulin-depleted peroxidase binds to and elutes from Con Aagarose as efficiently as the untreated peroxidase. Znteraction of Two Highly Purified Thyroid Peroxidase Preparations with Con A-Agarose -A fragmented, hydrophilic form of thyroid peroxidase (6,7,19) can be obtained in a state of greater purity than the detergent-solubilized, intact peroxidase.
For the experiments shown in Table  V, fragmented  thyroid  peroxidase  was prepared  by the method  of  Taurog and associates (6, 7), with the modifications given under "Experimental Procedures." This preparation is at least 25% pure, as judged by a peroxidase activity stain and a protein stain of polyacrylamide electrophoresis gels (6). As shown in Table V, this fragmented form of thyroid peroxidase also binds to Con A-agarose and can be eluted with cy-methyl-Dmannoside (Experiments 2 and 3). Furthermore, when amethyl-D-mannoside is mixed with this form of thyroid peroxidase prior to addition to Con A-agarose, peroxidase binding to Con A is greatly reduced (Experiment 4). The interaction between Con A and thyroid peroxidase pre- of immunoadsorbed preparation on Con A-agarose For Sample 1, 270 al of a Triton X-100 extract were added to 3.0 ml of 0.1% Triton X-100, 10 mM Tris (pH 7.4), 0.1 rnM KI. For Sample 2, 270 ~1 of the same Triton X-100 extract were added to 2.75 ml of a 0.5% (w/v) immobilized anti-porcine thyroglobulin mixture that had previously been exhaustively washed with the 0.1% Triton X-100, 10 mM Tris (pH 7.4), 0.1 rnM KI buffer to remove bovine serum albumin and sodium azide. After intermittent mixing for 3 h at room temperature, both samples were centrifuged at 2200 x g for 10 min, and 2 ml of each supernatant (initial sample) were added to separate tubes containing 1 ml (settled volume) of Con A-agarose and 1 ml of 10 mM phosphate (pH 6.0), 0.1% Triton X-100, 0.1 rnM KI, 1 mM CaCl,, and 1 mM MnCl,.
The Con A-agarose had previously been equilibrated as described in Table III. The suspensions were mixed for 1.5 h at 4" and supernatants were obtained as described in Table III. The gels were washed with 2.5-ml portions of 10 rnM phosphate buffer (pH 6.01, 0.1% Triton X-100, 0.1 mM KI. The supernatants were assayed separately for peroxidase activity and the totals were subtracted from the initial fraction to determine the amount bound. Three milliliters of o-methyl-n-mannoside (100 mg/ml) in 0.1 M Tris/HCl (pH 8.91, 0.1% Triton X-100, 0.1 rnM KI were added to each gel and the suspensions were mixed overnight at 4". A supernatant was obtained by centrifugation as described in Table III. The gels were washed with 3 ml of the cu-methyl-n-mannoside solution but no further peroxidase activity was obtained in the subsequent supernatant.  Thyroid peroxidase as prepared by the method of Nunez and co-workers (19) was applied to Con A-agarose and eluted with cu-methyl-n-mannoside as described in Table V. The eluted fraction was concentrated and the sugar removed by vacuum dialysis against 5 mM phosphate buffer (pH 7.4) containing 0.1 mM KI. Samples (35 pg of protein/gel) were then applied to polyacrylamide gels and electrophoresis was conducted as described under "Experimental Procedures." Individual gels were stained for (a) thyroid peroxidase (Z'PO) activity by means of the guaiacol/H,Oz assay reagent; (5) for carbohydrate by means of the PAS reagent; and (c) for protein with Coomassie blue (CB). The black lines near the bottom of the gels indicate the position of the tracking dye. The UFFOW denotes the correspondence between the peroxidase activity band, the PAS band, and the major protein band. The peroxidase activity band developed within 10 min of H,O, addition. pared by the method of Nunez and co-workers (19) has also been studied. This preparation is also about 25% pure as judged by electrophoretic techniques (19). Results similar to those presented in Table V have been obtained with this preparation.
In addition, the ol-methyl-n-mannoside-eluted peroxidase fraction has been analyzed by polyacrylamide gel electrophoresis. The gels were stained for peroxidase activity, PAS-reactive material, and proteins. As shown in Fig. 4, the cY-methyl-n-mannoside eluate contains a PAS component which corresponds with the peroxidase band. The peroxidase, PAS-positive region is associated with the major protein band.

DISCUSSION
The data presented here can be summarized as follows: (a) thyroid peroxidase binds to Con A-agarose and the binding is inhibited by monosaccharides in the same general order of specificity as observed in the competition between monosaccharides and dextran for the carbohydrate binding site of Con A; (b) the binding of thyroid peroxidase to Con A-agarose can be reversed by the addition of cu-methyl-n-mannoside to the thyroid peroxidase . Con A . agarose complex; (cl inactivation of the carbohydrate binding site of Con A by demetallization greatly reduces the binding of thyroid peroxidase,,and reactivation of the carbohydrate binding site by addition of Ca2+ and Mn*+ to demetallized Con A-agarose restores thyroid peroxidase binding; (d) catalytically active-trypsin prepared forms of thyroid peroxidase which do not exhibit hydrophobic properties bind to and elute from Con A-agarose as efficiently as the intact form of peroxidase obtained by nonionic detergent solubilization; and (e) on polyacrylamide electrophoresis gels, the thyroid peroxidase activity band from the cu-methyl+mannoside-eluted fraction corresponds with a glycoprotein band.
This evidence strongly suggests that thyroid peroxidase is a glycoprotein which binds to the carbohydrate binding site of Con A. However, alternative explanations for the interaction of thyroid peroxidase with Con A have been considered since the binding of membrane components and glycoproteins to Con A may involve more than sugar-Con A interactions (18,(46)(47)(48)(49) and since the thyroid peroxidase is not available in a homogeneous state. One type of noncarbohydrate interaction with Con A that has been reported involves hydrophobic binding (47,(50)(51)(52). However, the following observations indicate that hydrophobic interactions do not play a major role in the binding of thyroid peroxidase to Con A-agarose. First, demetallization of Con A greatly reduces the binding of thyroid peroxidase (Fig. 1). Demetallization destroys the carbohydrate binding site of Con A (23, 24, 351, but does not alter the ability of Con A to bind hydrophobic molecules.2 Addition of Ca2+ and Mn2+ reactivates the carbohydrate site of Con A and restores the binding of thyroid peroxidase (Fig. 1). Second, the fragmented form of thyroid peroxidase, which does not exhibit hydrophobic properties, binds to and elutes from Con A-agarose as effectively as the intact form of thyroid peroxidase. This indicates that the hydrophobic portion of the intact peroxidase is not responsible for the binding of thyroid peroxidase to Con A. Third, the Triton X-100 micelles present in the experiments described here for intact thyroid peroxidase would be expected to block the hydrophobic binding site of Con A. Fourth, we were not able to improve the recovery of intact thyroid peroxidase from Con A-agarose by use of ethylene glycol, a polarity reducing solvent which was effective in increasing the yield of human interferon from Con'A-agarose (52). This protein binds to Con A-agarose both by hydrophobic interaction and carbohydrate recognition.
It has also been suggested that protein-protein interactions, i .e . between Con A and adsorbed proteins, which are mediated by electrostatic forces, may stabilize the glycoprotein * Con A complex (18,53). The combined requirement of sugar and elevated pH for maximum release of thyroid peroxidase from Con A-agarose may reflect such interactions (Fig. 3). Bishayee and Bachhawat (54) also reported a pH-dependent release of a lysosomal glycoprotein, arylsulfatase A, from a Con A-glycoprotein precipitate. However, this type of protein-protein interaction cannot solely account for the interaction of thyroid peroxidase and Con A-agarose because of the evidence summarized in the first paragraph under "Discussion," and the fact that, in the absence of sugar, only about 5% of the peroxidase can be eluted at pH 9.
Finally, since the peroxidase is not pure, we have considered the possibility that the interaction between Con A and thyroid peroxidase could be mediated by a glycoprotein contaminant in the peroxidase preparations.
Thyroglobulin is the major glycoprotein contaminant of crude thyroid peroxidase preparations. A thyroglobulin . thyroid peroxidase complex, if pres- and can be eluted with a-methyl-D-mannoside. However, because the peroxidase is not pure, we have considered the possibility that the binding of thyroid peroxidase to Con A could be mediated by a glycoprotein contaminant in the preparation.
This possibility seems unlikely, but it cannot be completely excluded until a homogeneous preparation of thyroid peroxidase is available. We hope that these studies will help to achieve this goal by providing a new tool for the purification of this and other peroxidases.