Thyroid-stimulating Hormone Binding to Beef Thyroid Membranes ROLE OF N-ACETYLNEURAMINIC ACID*

The effect of sugars on ‘*“I-thyroid-stimulating hormone binding to beef thyroid membranes was studied to determine their role in thyroid-stimulating hormone (TSH) binding. At 0.1 M concentration, N-acetylneuraminic acid produced a 3- to 7-fold increase in TSH binding, was the only sugar to enhance TSH binding, and did so whether binding was determined in the cyclase medium or under conditions of optimum binding. The enhanced TSH binding remained after the membranes were removed from the high NeuAc concentration and an effect was observed at concentrations of 10 mM NeuAc. NeuAc did not alter the kinetics of TSH binding but the pH optimum for TSH binding shifted from pH 5.5 to 7.5 in the presence of NeuAc. Incubation of the membranes with increasing concentrations of NeuAc resulted in increased sialic acid content of the membranes. The NeuAc concentration curve of membrane sialic acid and TSH binding were roughly parallel. The capacity of the low affinity site increased from 0.74 to 2.5 nmol/mg of protein in the presence of NeuAc. The apparent affinity (0.88 x 10s M-‘) of this site was unaffected by NeuAc. With the high affinity site, NeuAc increased both the apparent affinity and capacity from 2.2 x lOa M-’ to 5.5 x lOa M-’ and 1.6 to 3.1 pmol/mg of protein, respectively.

The effect of sugars on '*"I-thyroid-stimulating hormone binding to beef thyroid membranes was studied to determine their role in thyroid-stimulating hormone (TSH) binding. At 0.1 M concentration, N-acetylneuraminic acid produced a 3-to 7-fold increase in TSH binding, was the only sugar to enhance TSH binding, and did so whether binding was determined in the cyclase medium or under conditions of optimum binding. The enhanced TSH binding remained after the membranes were removed from the high NeuAc concentration and an effect was observed at concentrations of 10 mM NeuAc. NeuAc did not alter the kinetics of TSH binding but the pH optimum for TSH binding shifted from pH 5.5 to 7.5 in the presence of NeuAc.
Incubation of the membranes with increasing concentrations of NeuAc resulted in increased sialic acid content of the membranes. The NeuAc concentration curve of membrane sialic acid and TSH binding were roughly parallel.
The capacity of the low affinity site increased from 0.74 to 2.5 nmol/mg of protein in the presence of NeuAc. The apparent affinity (0.88 x 10s M-') of this site was unaffected by NeuAc. With the high affinity site, NeuAc increased both the apparent affinity and capacity from 2.2 x lOa M-' to 5.5 x lOa M-' and 1.6 to 3.1 pmol/mg of protein, respectively. Neuraminidase or neuraminidase plus /3-galactosidase incubation of the membranes removed approximately 60% of the sialic acid from the membranes within 15 to 30 min but did not affect TSH binding. Large quantities of sialic acid were detected in the soluble fractions during isolation of the membranes, 4 to 5% of which was ultrafilterable and not associated with high molecular weight proteins. It is concluded that among the sugars tested, NeuAc exhibits an unique effect on TSH binding th& may have physiological significance. The inability to alter TSH binding by enzymatic removal of endogenous sialic acid suggests that either NeuAc resistant to hydrolysis is sufficient to maintain TSH binding or that NeuAc important in TSH binding is removed during membrane preparation but is replaced by incubation with exogenous NeuAc.
Bovine thyroid-stimulating hormone is a glycoprotein with rine-stimulated adenylate cyclase activity in fat cell and fat carbohydrate residues of N-acetylglucosamine, mannose, Ncell homogenates, respectively. The effects of the lectins could acetylgalactosamine, fucose, and galactose (1, 2) listed in order be reversed by simple sugars which bind specifically to the of abundance. These sugars probably exist as oligosaccharides lectins. From the insulin-like effects of these lectins and their in the intact molecule and are distributed between the (Y and /3 binding properties, he concluded that sugars on the memsubunits. The role of these sugars in TSH' binding has not been explored. Cuatrecasas and Tell (3) explored the role of carbohydrates in insulin binding and adenylate cyclase activation. They observed that, like insulin, concanavalin A and wheat germ agglutinin enhanced glucose transport and inhibited epineph-  In line with common usage, the nondisplaceable 'Y-TSH was subtracted from total '*"I-TSH bound to yield specific binding which is reported unless otherwise specified.

Assay of Sialic Acid
Assay of NeuAc by this assay is reported as sialic acid since the assay is not entirely specific for NeuAc; however, NeuAc is suggested to be the major component assayed (10). The sialic acid content of the membranes was determined by the thiobarbituric acid assay of Warren (10). Since glycerol and sucrose produced interfering colors with the assay, it was necessary to do one of the following: (a) remove glycerol and sucrose from the membrane suspension as explained under "Preparation of Membranes," (b) perform a spectral analysis of the colored solution and compensate for increased or decreased absorbance at 549 nm resulting from the contaminants, or (c) measure the absorbance at 549 nm before and after treating the colored solution with strong base. The color produced by NeuAc is destroyed by the strong base (10). The residual absorbance at 549 nm due to glycerol or sucrose was 0.055 + 0.013 which accounted for approximately 30 to 50% of the total absorbance of the solution. The residual absorbance of the NaHCO,-isolated membranes after strong base was equivalent to that of solvent controls. The approaches give roughly equivalent results. For convenience, the first approach was used routinely. Prior to the assay, the membranes or solutions were incubated at 80" with 0.1 N H,SO, for at least 60 min with intermittent shaking to remove essentially all the sialic acid from the membranes.
Then the membranes or solutions were centrifuged at 40,000 x g for 10 min and an aliquot of the supernatant was assayed for sialic acid.

RESULTS
Under binding and cyclase conditions 0.1 M a-methyl mannoside, mannose, N-acetylmannosamine, galactose, N-acetylgalactosamine (not shown), fucose, glucose, and N-acetylglucosamine did not alter TSH binding (Figs. 1 and 2). This was true at low TSH concentrations used in the binding assay and high TSH concentrations used in the cyclase assay. As Moore and Wolff (4) demonstrated with other agents that inhibit or enhance TSH binding to thyroid membranes, there was no qualitative difference between the effect of the sugars on TSH binding under cyclase and binding conditions. Among the sugars tested, N-acetylneuraminic acid was unique in its ability to enhance V-TSH binding (Figs. 1 and 2). NeuAc concentrations of 100 mM caused a 3-to 7-fold increase in TSH binding. A significant increase in TSH binding was observed at NeuAc concentrations as low as 10 mM (Fig. 3 NeuAc occurred within 2 to 3 min and was stable for at least 90 to 120 min. Under binding conditions, the pH optimum for TSH binding to thyroid membranes was approximately pH 5.5 (Fig. 4). In the presence of 30 IIIM NeuAc, the pH optimum for TSH binding was approximately pH 7..5 (Fig. 4). At pH 7.5, the amount of TSH bound to the NeuAc-treated membrane was approximately 3-fold higher than to control membranes at the same pH. The peak of TSH binding with NeuAc at pH 7.5 was The aggregation of the membranes as measured by light scattering (absorbance at 700 nm) was determined as a approximately 30% higher than the peak of TSH binding at pH function of pH. Aggregation of the membranes at pH less than 6.5 in the presence of 30 mM NeuAc was half-maximal at pH 5.7 5.5 without NeuAc. The two pH curves on TSH binding were to 5.9. This was not significantly different from aggregation in the absence of NeuAc.
qualitatively similar with a fairly narrow pH optimum for binding.
The membranes were exposed briefly to high NeuAc concentrations to determine if high NeuAc concentrations of the medium are necessary for enhanced binding. Membranes were incubated for 30 min at 22O with 100 mM NeuAc followed by 15-fold dilution, sedimentation, and resuspension of the membranes in 50 mM Hepes (pH 7.5). When binding was determined after each dilution and resuspension, the membranes continued to exhibit TSH binding severalfold above control membranes (Table II)  after pretreatment of the membranes with NeuAc to determine if high concentrations of NeuAc result in increased sialic acid content of the membranes. Membranes were incubated with increasing concentrations of NeuAc at 22' for 30 min and sedimented by centrifuging at 40,000 x g for 10 min. The surface of the pellet was washed and the membranes were Analysis of the '*?-TSH binding by Scatchard plots revealed two orders of binding sites with affinities of 0.88 x lo8 M-' and resuspended in 50 mM Hepes. An increase in membrane sialic 2.2 x 10' M-l. In the presence of 30 mM NeuAc, the apparent affinity of the low affinity site was unchanged at 0.88 x lo8 M-' acid content was detected at NeuAc concentrations above 10 while the capacity of this site increased from 0.74 nmol/mg of protein to 2.5 nmol/mg of protein (Fig. 6). The NeuAc mM in the incubation medium (Fig. 5). This NeuAc concentraconcentration was established emperically from the NeuAc concentration curve as one which produced significant en-tion curve of membrane sialic acid content is roughly parallel hancement of TSH binding but did not require expenditures of large quantities of NeuAc.
to the NeuAc concentration curve of TSH binding.
Both the apparent affinity and capacity of the high affinity site increased in the presence of NeuAc with 9 mM NeuAc in the incubation medium. The apparent affinity of the high affinity site increased from 2.2 x 10' M-' to 5.5 x lo* M-' and the capacity increased from 1.6 to 3.1 pmol/mg of protein (Fig.  7)  The incubation solution was then diluted 15.fold with 50 mM Hepes and followed by sedimentation of the membranes by centrifuging at 40,000 x g for 10 min. Complete recovery of membrane protein is assumed so that the increase in NeuAc is not due to selective loss of membrane protein during sedimentation and resuspension, but approximately 10% of the membrane protein was lost during the procedure.
The surface of the pellet was washed twice with 50 rnM Hepes and then resuspended in 0.1 N HBO,.
The membranes were incubated at 80" for 60 min in H,SO, prior to assay of the sialic acid content by the thiobarbituric acid assay. was not studied further.
membrane degradation since a similar release was observed from control membranes. Incubation with 50 wg/ml of fi-galactosidase did not cause release of membrane sialic acid until 3 to 4 hours of incubation.
Incubation of the membranes with &galactosidase did cause the release of a substance with peak absorbance at 532 nm in the thiobarbituric acid assay. This compound was not identified further.
In contrast to their effect on membrane sialic acid content, incubation of the membranes with 50 rg/ml of neuraminidase or neuraminidase plus B-galactosidase at 22" for 30 to 60 min did not alter TSH binding. Decrease in binding was observed at 3 to 4 hours of incubation but was attributed to membrane degradation (Table IV).
The sialic acid content of the fractions obtained during membrane isolation was assayed to determined if significant NeuAc is removed from the membranes during isolation. Analysis of the sialic acid content of the soluble and membrane fractions during preparation of the membranes revealed that a large quantity of sialic acid was present in the soluble fractions (Table V). Approximately 4 to 5% of the soluble sialic acid was ultrafilterable through an Amicon membrane with exclusion molecular weight of greater than 1000. Consequently, this portion of the soluble sialic acid was not associated with high molecular weight proteins such as thyroglobulin which is a rich source of NeuAc in the thyroid. This free sialic acid may represent endogenous NeuAc that is removed from the membrane during preparative procedures. If the sialic acid content of the filtrate is expressed in terms of nanomoles of NeuAc/mg of protein, there was a go-fold increase between the filtrate of the initial homogenate and the filtrate of the final membrane suspension while the nanomoles of NeuAc/mg of protein in the Top, the incubation solution contained 0.099 mg/ml of membrane protein, 50 mM Hepes (pH 7.51, 1% glycerol, 1.25% albumin, and the appropriate concentration of 'Y-TSH. Binding was started by adding the membranes and was performed at room temperature for 10 min. The amount of TSH bound (picomoles/mg of membrane protein) is plotted versus the ratio of the amount bound to the concentration of free hormone (BOUND/FREE). Bottom, the conditions of incubation are the same as above except the incubation solution contained 9 mM NeuAc. A lower concentration of NeuAc was necessary to study the binding to the high affinity site since higher NeuAc concentrations produced maximum TSH binding at low TSH concentrations.
The relation between the low affinity and high affinity site has been demonstrated previously (5). The affinity of the high affinity site was 2.2 x lOa M-' and 5.5 x 10' M-' in the absence and presence of NeuAc, respectively.
The capacity of the site was 1.6 and 3.1 pmol/mg of protein in the absence and presence of NeuAc, respectively. soluble fraction remains fairly constant throughout the isolation procedure.

DISCUSSION
Previous studies by Moore and Wolff (4) have described the binding of TSH to bovine thyroid membranes and the relation of binding to adenylate cyclase activation. TSH binding to cells (ll-13), slices (14), and membranes (8, 9, 12, 13) have been described by other investigators. The studies of TSH binding have been predominantly descriptive and the mechanism of TSH binding remains unknown. Several points concerning the binding process are known. Through a study of the effect of pH, cations, and anions,* Moore and Wolff (4) suggested that electrostatic interactions may be important in the binding process. The effect of phospholipase A and filipin on TSH binding indicated that membrane lipids are involved in the binding process (4). Wolff et al. (15) demonstrated that the binding capacity of TSH resides mainly in the /3 subunit.
The effect of selected lectins on TSH binding indicated that the carbohydrate moieties of the hormone and/or membrane binding in the presence of NeuAc results from an increase in the capacity of the low affinity receptors rather than from an increase in their affinity and from an increase in both the affinity and capacity of the high affinity sites. Amir et al. (8) failed to observe an increase in TSH binding with 1 and 30 mM NeuAc concentrations.
The difference between their observations and the present data is difficult to explain. The effect of NeuAc on binding was consistently observed throughout these experiments from 10 to 100 mM NeuAc concentration.
The pH optimum for TSH binding shifts from approximately pH 5.5 to pH 7.5 in the presence of NeuAc. The decrease in binding at pH below 5.5 in the absence of NeuAc had previously been attributed to degradation of either TSH or the receptor (4). However, TSH binding to the NeuAc-treated membranes decreases rapidly at pH less than 7.5 so that the contours of the pH curve of the control and NeuAc-treated membranes are qualitatively similar. The significance of this similarity is not known at the present time. Moore and Wolff (4) demonstrated that thyroid membranes aggregate at pH less than approximately 6.5, being half-maximal at pH 5.6. Aggregation when measured by light scattering (absorbance at 700 nm) in the presence of 30 mM NeuAc was half-maximal at pH 5.8 to 5.9 and did not differ significantly from aggregation of control membranes.
Concentrations of NeuAc in the incubation solution which promote TSH binding also result in increased sialic acid content of the membranes which is retained after several sedimentations and resuspensions in solutions free of NeuAc. These membranes continue to exhibit increased TSH binding even after the membranes are removed from the high NeuAc concentration which indirectly suggests that NeuAc binding to the membranes is important in the enhancement of TSH binding.
The biological significance of these observations is open to conjecture. Neuraminidase and neuraminidase plus p-galactosidase treatment of the membranes effectively removed approximately one-half of the NeuAc from the membranes without affecting the TSH binding. This suggests the following possibilities: (a) endogenous membrane NeuAc is not involved in TSH binding, (5) the NeuAc resistant to hydrolysis is sufficient to maintain TSH binding, or (c) membrane NeuAc that is physiologically important in TSH binding is removed during membrane isolation but is replaced by incubation with high concentrations of NeuAc.
This research and other recent reports indicate that carbohydrates and more specifically NeuAc are involved in the binding and activation process so that it is impossible to eliminate a role of NeuAc in TSH binding. Winand and Kohn (16) have recently reported that the soluble TSH receptor is rich in carbohydrate. Their 15,000 to 30,000 molecular weight receptor fragment contains approximately 30% carbohydrate and 10% sialic acid. Using this soluble receptor system, Tate et al. (17) observed that TSH binding to the soluble receptor is inhibited by preincubation with Sepharose-linked concanavalin A or neuraminidase.
When compared to the present data, this indicates the presence of NeuAc residues on the intact membrane which are resistant to neuraminidase hydrolysis and thereby maintain TSH binding following neuraminidase incubation. These same NeuAc residues might become available to neuraminidase through conformational changes in the receptor occurring during solubilization and account for the observations in the soluble receptor system. One might expect NeuAc residues resistant to neuraminidase in the intact membrane to be unavailable to TSH binding, but this assumes that NeuAc is directly rather then indirectly involved in the binding process such as by electrostatic attraction. The appropriateness of the third possibility is difficult to determine from the present data. Even though considerable amounts of sialic acid become soluble during isolation of the membranes, its exact source is not known. In most receptor systems, the measured affinities of the hormones for their membrane receptor are rather low compared to the affinity that might be expected from concentrations of the hormone that produce a biological response. This discrepancy varies from 1 to several orders of magnitude (4,18,19) and may result from alterations of the receptor during preparative procedures or storage (20). Exposure of the thyroid membranes to NeuAc results in a significant increase in both the affinity and capacity of the high affinity site and may represent partial "repair" of a damaged receptor or replacement of a component lost during membrane preparation. Whether this repair results in an increased sensitivity of TSH stimulation of sdenylate cyclase is presently under investigation.
In conclusion, treatment of bovine thyroid membranes with NeuAc results in an increase in TSH binding which appears to result from an increase in the capacity of the low affinity receptors and also from an increase in the affinity and capacity of the high affinity receptors. The enhancement of TSH binding by NeuAc seems related to NeuAc binding to the membranes but removal of endogenous NeuAc from the membranes by neuraminidase digestion does not affect TSH binding.
The biological significance of these observations is obscure, but the unique qualities of NeuAc and its role in membrane receptor systems suggests that its role in TSH binding may be physiologically meaningful.