Incorporation of N-Acetylmannosamine and N-Acetylglucosamine into Thyroglobulin in Rat Thyroid in Vitro

Abstract The incorporation of N-acetylmannosamine ([3H]ManNAc) and N-acetylglucosamine ([3H]GlcNAc and [14C]GlcNAc) has been studied in rat thyroid hemilobes and homogenates. The radioactive label is incorporated into thyroglobulin (TG), [3H]ManNAc showing a time course more rapid than [3H]GlcNAc. [3H]ManNAc is incorporated directly into 19 S thyroglobulin showing no lower molecular weight-labeled precursors of TG either in the soluble or in the solubilized proteins; [3H]GlcNAc is incorporated into 19 S TG and into precursors both in soluble and solubilized proteins. [3H]ManNAc is converted only into N-acetylneuraminic acid (sialic acid). [3H]GlcNAc and [14C]GlcNAc are incorporated mainly as GlcNAc, and an additional small proportion is converted into sialic acid. Both labeled GlcNAc and sialic acid are found in soluble TG but only GlcNAc is found in membrane-bound TG. [3H]ManNAc is particularly useful to study the last step of carbohydrate incorporation while [3H]GlcNAc is a better tool to study the initiation of carbohydrate incorporation. The presence of the label as sialic acid only in soluble TG suggests that sialic acid is incorporated at the time of TG release from endoplasmic membranes. It may, therefore, serve as an indicator of TG secretion.

[3H]GlcNAc and [W]GlcNAc are incorporated mainly as GlcNAc, and an additional small proportion is converted into sialic acid. Both labeled Glc-NAc and sialic acid are found in soluble TG but only GlcNAc is found in membrane-bound TG.
[3H]ManNAc is particularly useful to study the last step of carbohydrate incorporation while [3H]GlcNAc is a better tool to study the initiation of carbohydrate incorporation. The presence of the label as sialic acid only in soluble TG suggests that sialic acid is incorporated at the time of TG release from endoplasmic membranes. It may, therefore, serve as an indicator of TG secretion.
Although in recent years there have been many reports on glycoprotein biosynthesis, t.he process of incorporation of carbohydrates into protein is not yet completely understood.
Since the thyroid secretes a glycoprotein, thyroglobulin, this gland can provide a good model for the study of glycoprotein biosynthesis and secretion.
TG' contains about 1O70 carbohydrate which has been shown to be present in three distinct units; one has a molecular weight of 1050 and contains mannose and Wacetylglucosamine, the second has a molecular weight of 3200 and contains in addition to N-acctylglucosamine and mannosr, galactose, sialic acid (N-acetylneuraminic acid), and fucose (l-3). Recently a third unit from human TG has been partially characterized; it contains galactosamine and possibly sialic acid (4). It has been suggested that fucose and galactosc are incorporatrd directly into an 18 S protein whereas mannose incorporation has a time course similar to that observed for leucine (5, 6); more recently it has been suggested that mannose is pre>ent in two intermediates of TG synthesis, the 6 S and 7 S precursors (7). Since all the labeled carbohydrates used so far are present in the middle of the oligosaccharide chains and are subject to conversion, we wanted to study specifically the first and the last steps of carbohydrate incorporation into TG. A:-Acetylmannosamine is the precursor of sialic acid and believed not to be transformed into other sugars (8); N-acetylglucosamine is the first sugar to be incorporated into the polypeptidr chain through an N-glycosidic linkage to aspartic acid (8). Neither of these precursors have heretofore been used to study TG synthesis. Glucosamine, reported previously (9, 10) is a less specific precursor since it is incorporated into A-acetylglucosamine, N-acetylgalactosamine, COZ, and sialic acid (11,12).

MATERIALS AND METHODS
XIale Fisher rats, -250 g weight, were bred in the Sational Institutes of Health Laboratory Aids Branch and fed l'ul,ina laboratory chow.
Digitonin prepared as previously described (13), was from Fisher.
Aquasol, universal liquid scintillation counter mixture, was from SW-England Suclcar, NCS solubilizer from Amersham-Searle. Earle's solution was provided by the Media ITnit of t.he National Institutes of Health. Neuraminidase of Vibrio cholerae was from Schwarz-Mann.
The rats were lightly anesthetized with ether and then killtd by cssanguination; connective tissue n-n-: removed from the thyroids which were then immersed in cold (4") 0.1 M Tris-I-ICI buffer, pH 7.4. The glands from several animals were cut in quarters, pooled, and pieces totaling -20 mg were used in each incubation flask. The pooled thyroids were immersed in 2 ml of Earle's solution pregassed for 20 min with O&O2 (95:5) and preincubated for 20 min at 37" in a constant temperature shaking water bath at 120 oscillations per min. After the preincubation, 50 &i of the labeled carbohydrate was added and incubation was carried out for 1, 3, 6, and 12 hours, gassing the mixture every 15 min with O&OZ.
In some experiments the pooled glands (~20 mg) were directly homogenized in 2 ml of Earle's solution, and the homogenate was incubated as above.
After incubation the glands were homogenized in 2 ml of 0.1 M Tris-HCl, 0.25 M sucrose, pH 7.4, in a Potter-Elvehjem homogenizer with a Teflon pestle (clearance 0.003 to 0.005 inch) at 1,100 rpm, three strokes in ice. The homogenate was centrifuged at 105,000 x g for 1 hour at 4" in a Spinco L centrifuge rotor 40. The supernatant, S1, was dialyzed for 2 days at 4" against four to five changes of 0.1 M Tris-HCl, pH 7.4, buffer. The pellet was rehomogenized in 1 ml of 2% digitonin (13), 10 strokes at, 2,4CO rpm in ice, and recentrifuged at 105,000 X g for 1 hour at 4". Using I251 as label it was shown that up to 71 70 of t,he pellet radioactivity was solubilized by this procedure whereas using sodium deoxycholate the recovery was up to 83%; these values are the mean of six experiments for digitonin and of three experiment,s for deoxycholate.
The released labeled proteins examined by sucrose gradient ultracentrifugation showed that more than 90y0 of the total radioactivity was present in the 19 S TG peak. The amount of nonlabeled protein solubilized from the pellet was frequently low; the protein peak of 19 S TG in sucrose gradient patterns was variable and did not exceed -20% of the total protein applied on the gradient.
No estimation for the recovery of the protein content was carried out.
The supernatant after digitonin treatment, &, was dialyzed against 0.1 M Tris-HCI, pH 7.4, buffer for 2 days, and the pellet was dissolved in 0.5 ml of NCS.
Aliquots of X1, S2, and pellet were counted in a Nuclear-Chicago Mark I scintillation counter in 15 ml of Aquasol containing 1 ml of distilled water.
Aliquots of soluble (S1) and solubilized (X,) proteins were precipitated with 95y0 ethanol or 10% trichloroacetic acid; the mixture was kept for 30 min at 4", then centrifuged at 3000 X g for 10 min. The precipitate was dissolved in 1 ml of distilled water, reprecipitated, and centrifuged as above; the washed precipitate was then dissolved in 1 ml of distilled water and 15 ml of Aquasol and counted in a scintillation counter. Sialic acid was released from labeled 19 S thyroglobulin isolated by sucrose gradient ultracentrifugation (14) by hydrolyzing with 0.1 N HzS04 for 1 hour in a sealed tube at 80". After hydrolysis a drop of 0.5% phenolphthalein in ethanol was added, and the mixture was neutralized with 0.1 M 13a(OH)z; the precipitate was removed by centrifuging at 3000 x g for 10 min at 4" and the supernatant was dried under vacuum; the dried material was dissolved in 0.2 ml of distilled water and applied to a strip of Whatman No. 3MM filter paper, 5 X 47 cm. Descending chromatography was performed for 8 hours in an isoamylacetate-acetic acid-water (3 :3 : 1) system (15). After chromatography the strips were dried and cut into 2-cm pieces; each piece was extracted for 24 hours with 1 ml of distilled water, and the extract plus paper were put into 15 ml of Aquasol and counted in a scintillation counter. Unlabeled sialic acid, GlcNAc, A-acetylgalactosamine, and neutral sugars were run on parallel strips and were detected by the silver nitrate method 2073 ["H]ManNAc after hydrolysis and paper chromatography was 84 to 95% (mean 90%) of the amount in the protein.
TO further identify the labeled product in soluble or solubilized proteins after incubation with N-[%]acetylmannosamine, 19 S TG isolated by sucrose gradient centrifugation was incubated with neuraminidase from V. cholerae (5 mg of protein per 25 i.u. of enzyme) (17). After the incubation the mixture was precipitated between 1.4 and 1.8 M ammonium sulfate, redissolved, dialyzed, and then centrifuged in a sucrose gradient as described previously.
The pattern of the protein showed that the 19 S peak of unlabeled TG was unchanged after neuraminidase treatment although the label had been completely displaced from the 19 S zone. To further establish whether the neuraminidase used had any protease or other glycosidase activity a 19 S TG labeled with [l%]galactose in vitro was incubated with neuraminidase as above; after the incubation the mixture was precipitated with 10% trichloroacetic acid, centrifuged at 3000 x g for 10 min and the radioact,ivity in the precipitate and supernatant was measured in a liquid scintillation counter.
More than 92% of the 14C was trichloroacetic acid precipitable.
GlcNA2c was released by hydrolyzing labeled 19 S TG, isolated as above, with 4 N HCl for 6 hours in a sealed tube in boiling water.
The hydrolysat,e was then dried under vacuum, dissolved in water, and l)assed through a small column of Dowex 50-X4 (H+ form) (1X), eluted with 2 N HCl and t.he eluate dried at 50" under T-acuum. The material, dissolved in water, was then acctylated as described by O'Rrien. 2 To 0.4 ml of the sample, 60 ~1 of 1 s sodium carbonate and 2 ~1 of acetic anhydride were added at room temperature.
After 5 min Dower; 50 resin (hydrogen form) was added until the pH reached 2.5, then the mixture was centrifuged at 3000 x g for 10 min, and the resulting supernatant was applied on a borate-buffered Whatman No. 3MRI paper shcrt.
High voltage paper electrophoresis was performed with 0.2 hf borate buffer, pH 10, at 20 volts per cm for X$4 hours as described (19,20). Standards (GlcNAc, Gal'Ac, and ilIan;";Xc) were run in parallel on t,he same sheet. After electrophoresia the markers were located by the silver nitrate method.
The portion of the sheet containing the labeled samples was cut into 2.cm segments, extracted, and counted as described above.
Sialic acid was determined by the thiobarbituric acid assay of Warren (on samples hydrolyzed in 0.1 N H2S04 for 60 min at 80") (21) and by the resorcinol reaction (22) on the 1.4 to 1.8 M ammonium sulfate fractions. Density gradient ultracentrifugation in sucrose was performed as previously described (14). The double antibody t,echnique of immunoprecipitat'ion was done according to Salabi! and Robbins (23).

RESULTS
The rat,e of incorporation of label into thyroid hemilobes and homogenates is shown in Fig. 1. [3H]ManNAc incorporation increased with time reaching a plateau of -0.5% of the dose at 6 hours after incubat,ion both in hemilobes and in homogenates.
[3H]GlcNAc was incorporated at a slower rate, increasing up to 12 hours and reaching at this time about 0.8% of the dose in hemilobes and 0.5% in homogenates.
With [3H]Gl~NA~ but not [3H]ManNAc incorporation into hemilobes was greater than in homogenates at all times studied.
%I from both compounds was incorporated into the soluble protein (S1) of hemilobes, increasing to -40% of the total ra-  These findings, however, were inconstant and variable in amount; hence their significance is uncertain.
The 27 S peak was seen in about onc-third of the analyses, except for hemilobes incubated with [5H]ManNAc where it was seen in one-tenth of the analyses.
The 12 S peak was seen in one-half of the analyses of hemilobes incubated with [3H]l\lanNAc, in one-third of hemilobes incubated with [3H]GlcNAc, and in about one-tenth of analyses of homogenates.
In hemilobes incubated with [3H]GlcNAc the solubilized fraction (8,) contained 12 S, 19 S, and 27 S peaks of radioactivity similar to the 81 fraction, as shown in Fig. 4. In the case of [3H] ManNXc, however, no TG was detected in the Xp fraction at any time.
To investigate the proport,ion of the labeled protein immunologically related to TG, aliquots of the radioactive soluble and solubilized proteins were subjected to immunoprecipitation. As shown in Table II  The 19 S TG recovered from the sucrose gradient of the 81 fraction was pooled, conctntrated, and dialyzed. This was hydrolyzed with 0.1 K HZSOI, and also with 4 x I-ICI in t.he case of hemilobes incubated with [14C]GlcNAc. Chromatographic and electrophoretic analysis of the hydrolgsates is presented in Table III.
All of the label from [3H]ManNAc was found in sialic acid. Furthermore, treatment of the thyroglobulin fraction with neuraminidase resulted in complete disappearance of the labeled 19 S peak as determined by sucrose gradient ultra-protein.
The homogenate (Fig. 2)    The radioactivity present in the fractions of the thyroid extracts which do not sediment in the sucrose gradient is not related to TG since 85oj, of the radioactivity cannot be precipitated by trichloroacetic acid or ethanol, nor is there any precipitable radioactivity immunochemically related to TG. Presumably these fractions contain oligosaccharides related to the thyroid cell membrane.
[3H]ManNAc and [311]GlcNAc are known to be incorporated into glycolipid of mouse cell membrane (24). A considerable amount of radioactivity found in the particulate fraction after incubation with [$H]ManNAc is not extractable with digitonin and probably is also unrelated to TG.
Our data show that [3H]ManNAc is converted only into sialic acid in thyroid glycoprotein and is found in 19 S TG. Thus it is a specific detector of the incorporation of sialic acid in TG. Furthermore, the labeled 19 S is found only in the soluble fractions, indicating the addition occurs at, the time of TG release from endoplasmic membranes.
[3H]GlcNAc is also incorporated into soluble 19 S TG and its specific activity increases with time, the label showing a shift from a membrane-bound form and from slowly sedimenting fractions to the more mature molecule. This is in agreement with the finding of two glycoprotein intermediates in the biosynthesis of TG (7). A portion of [%-GlcNAc is converted to sialic acid and, consistent with the findings after [3H]ManN4c incorporation, this is found only in the soluble TG.
The labeling experiments are consistent with our finding of a low sialic acid content. in thyroglobulin solubilized from cell particles.
It has been reported that the carbohydrate content, including sialic acid, of the proteins solubilized from thyroid particles is similar to that from soluble proteins (25). Alt'hough these data were from a study on calf thyroid and the procedures used to release proteins from particles and to isolate and part)ially purify thyroglobulin were different from ours; we have no esplanation for the discrepancy between these results and our own on t,he sialic acid content of the membrane-bound thyroglobulin.
Several reports in the literature show that the recovery of solubilized labeled proteins is always high even when different procedures are used but the unlabeled protein recovery can be quite variable (26)(27)(28).
With our method it is possible to obtain good recovery of lz51-labeled protein, most of which is 19 S TG. The fact that a significant amount of radioactivity after incubation with [3H]ManNAc is still present in the pellet after treatment with detergent suggests that this label is probably in constituents of the cell membranes that are unrelated to TG synthesis.
Since sialic acid is present in a terminal position in the oligosaccharide chains of thyroglobulin and since we could not detect it in membrane-bound t.hyroglobulin, it is suggested that eialic acid incorporation may be used as an index for the secretion of this glycoprotein.
Its use in studying the secretion of other glycoproteins will be of interest.

Fabrizio Monaco and Jacob Robbins in Vitro
Thyroglobulin in Rat Thyroid -Acetylglucosamine into N -Acetylmannosamine and N Incorporation of