Deglycosylation of chondroitin sulfate proteoglycan by hydrogen fluoride in pyridine

ANALYTICAL

BIOCHEMISTRY

146,232-237

(1985)

Deglycosylation of Chondroitin Sulfate Proteoglycan by Hydrogen Fluoride in Pyridine’ CAROL A. OLSON, RICHARD KRUEGER,

AND NANCY

Department of Pediatrics, University of Chicago, Box 413, 5841 South Maryland

B. SCHWARTZ~ Avenue, Chicago, Illinois 60637

Received October 1, 1984 The original deglycosylation procedure using HF/pyridine has been modified for maximal removal of carbohydrate from chondroitin sulfate proteoglycan, with minimal alteration of the core protein. Gas-liquid chromatography analysis after treatment for various times showed that 95% of xylose and mannose and 7085% of other sugars were removed within 30 min, indicating that almost all chondroitin sulfate chains and about 80% of N- and U-linked oligosaccharides were removed. In contrast to the loss of carbohydrate, no change in amino acid composition or loss of immunoreactivity occurred. Longer treatment of up to 16 h resulted in little additional removal of carbohydrate, but did cause a significant decrease in solubility and recovery of the deglycosylated product. Optimal removal of xylose residues after about 1 h was also shown by maximal acceptor activity of the product in a xylosyltransferase assay. Rapid removal of the HF reagent by vacuum evacuation and ion-exchange chromatography, coupled with the reduced time of treatment allowed recovery of an intact, homogenous protein core that is amenable to structural and sequence studies. 0 1985 Academic press, Inc. KEY WORDS: deglycosylation; hydrogen fluoride; proteoglycan; carbohydrate; chondroitin sulfate; core protein.

We previously reported on a procedure for deglycosylating proteoglycans that removed greater than 95% of chondroitin sulfate chains by treatment with HF3-pyridine (1,2). While the product was a good xylosyltransferase acceptor, immunoreactivity with an antibody to proteoglycan core protein was reduced about 50-fold, indicating some structural al’ This research was supported by USPHS Grants AM19622, HD-04583, HD-09402, and HD-17332 and Grant l-737 from the March of Dimes. * Recipient of Research Career Development Award AM-00603 from the National Institutes of Health. 3 Abbreviations used: CSPG, chondroitin sulfate proteoglycan; PGHF, HF-deglycosylated CSPG; HF. hydrogen fluoride; MES, (2-N-morpholino)ethanesulfonic acid; GuHCl, guanidine hydrochloride; SDS, sodium dodecyl sulfate; AID,, proteoglycan collected from the bottom fractions following associative and dissociative gradient centrifugation; AID,-ZB, A,D, material that has been fractionated by chromatography on Sepharose CL-2B; AID,-ZB-H, hyaluronidase-digested AID,-2B; GalNAc, N-acetylgalactosamine; GluNAc, N-acetylglucosamine; NeuNAc, N-acetylneuraminic acid. 0003-2697185 $3.00 Copyright 0 1985 by Academic Press. Inc. All rights of reproduction in any form reserved.

teration of the deglycosylated product. Furthermore, a high-molecular-weight homogenous product was not consistently obtained, leading to the suspicion that residual HF was causing peptide cleavage during extraction and purification of the product. Therefore, the procedure has been substantially modified to more rapidly and completely remove HF after the HF-pyridine treatment. A further improvement involved reducing the reaction time from 8 or 16 h to 30-60 min. Carbohydrate and amino acid analysis and xylosyltransferase acceptor studies all indicate that maximum carbohydrate removal and minimal protein destruction occur within the first 60 min of treatment. MATERIALS

AND

METHODS

Materials. UDP-[‘4C]xylose (278 mCi/ mmol) was purchased from New England Nuclear and ‘251-labeled sodium iodide was obtained from ICN. Other materials were 232

DEGLYCOSYLATION

obtained from the following sources: anisole from Aldrich Chemical Company; MES from Calbiochem; polyhydrogen fluoride (70%) in pyridine from Columbia Organic Chemicals Company, Inc., Columbia, South Carolina; cesium chloride (special biochemical grade) from Gallard Schlesinger: testicular hyaluronidase (20,000 units/mg) from Leo Helsingborg Laboratories, Sweden; Sephadex and Sepharose gels from Pharmacia Fine Chemicals; ethyl acetate and guanidine hydrochloride (both sequanal grade) from Pierce Chemical Company; deoxyribose, glucuronolactone, ribose, and xylitol from P-L Biochemicals: 6-aminohexanoic acid, benzamidine hydrochloride, dithiothreitol, iodoacetamide, pepstatin A, and UDP-xylose from Sigma; anhydrous methanolic HCl (0.5 N), fucose, galactose, galactosamine, glucose, glucosamine, inositol, mannose, N-acetylgalactosamine, N-acetylglucosamine, N-acetylneuraminic acid, xylose, and 5% SP-2401 on 1OO/ 120 Supelcoport prepacked gas-chromatography columns from Supelco. Ion-retardation resin AGl lA8 was from Bio-Rad. Degl-vcosylation of proteoglycan with HF in pyridine. Chondroitin sulfate proteoglycan

(CSPG) from rat chondrosarcoma was extracted, purified, and partially deglycosylated by treatment with testicular hyaluronidase as described (2,3). Briefly, AID, proteoglycan purified by chromatography on Sepharose CL-2B (designated AID,-2B) was digested with testicular hyaluronidase in 0.05 M sodium acetate buffer, pH 5.0, containing 0.15 M NaCl, 0.0 1 M EDTA, 0.1 M 6-aminohexanoic acid, and 0.005 M benzamidine hydrochloride (4). The mixture was chromatographed on Sephadex G-100 equilibrated with 0.05 M sodium acetate buffer, pH 5.8, and 0.15 M NaCl. The hyaluronidase-digested proteoglycan (designated AID,-2B-H) eluted in the void volume: appropriate fractions were pooled, dialyzed extensively against H20, and lyophilized. This starting material for HF-deglycosylation, AID,-ZB-H (15-20 mg per reaction), was dried in a polyethylene container for 48 h at 7°C under vacuum,

OF

PROTEOGLYCAN

233

prior to chemical deglycosylation. The dried proteoglycan was then treated with polyhydrogen fluoride in pyridine with anisole as scavenger (0.1 ml anisole/l ml of HF in pyridine/lO mg AID,-2B-H) in a desiccator flushed with dry Nz (5). The vessel was quickly capped and the reaction was carried out with stirring at 23°C. At the end of the reaction the vessel was rapidly evacuated at 4°C by water aspiration for 30 min and then held under vacuum at 4°C for 3 h (6). The residue was resuspended in 2 ml of HZ0 at 4°C and quickly loaded onto a Bio-Rad AG 1 lA8 column (1 X 30 cm) to remove the residual acid and eluted with H20. The deglycosylated product was localized with the Bio-Rad reagent (7). The positive fractions (5-10) of 1.3 ml each were pooled, equilibrated with 0.05 M sodium acetate buffer, pH 5.8, containing 0.15 M NaCl, and loaded onto a Sephadex G-100 column (1.5 X 100 cm), equilibrated and eluted with the same buffer. Deglycosylated proteoglycan (PGHF) was detected at 230 nm and noncovalently bound carbohydrate by the carbazole method (8). PGHF (void volume fractions) was pooled, dialyzed, lyophilized, and used for analysis of the core protein of CSPG. Xylosyltransferase assay. The 100,OOOg supernatant, prepared from freshly harvested rat chondrosarcoma, was used as the enzyme source in experiments presented here (9). Enzyme activity was assayed according to Method 2A ( 10). The assay mixture contained 2.5 pmol MES buffer (pH 6.5) 12.5 kmol KCI, 0.6 pmol MgCl*, 0.15 pmol MnC&, 1.5 pmol RF, 3.0 nmol UDP-[‘4C]xylose (specific activity 35 nCi/nmol), 45 pugof xylosyltransferase, and varying amounts of xylose acceptor (PGHF) in a total volume of 80 ~1. Radioimmune assa)‘. Hyaluronidasetreated rat chondrosarcoma proteoglycan was used as antigen to elicit antibodies in rabbits (2). Proteoglycan core proteins were iodinated by the chloramine-T method and repurified by Sephadex G- 100 chromatography. The radioimmune assay and calculations of inhibition were performed as described (2).

234

OLSON, KRUEGER,

AND SCHWARTZ

Analytical procedures. All samples for protein or carbohydrate analysis were dried over phosphorous pentoxide for 48 h at 65°C under vacuum before weighing. Uranic acid concentration was estimated by the carbazole method (8). Protein concentrations were determined by the method of Lowry et al. (1 I) or by quanti~tive amino acid analysis, using norleucine as an internal standard. Amino acid analyses were performed using a Durrum D502 analyzer ( 12). Samples (50-75 pg protein) were hydrolyzed in distilled 6 N HCl containing 0.1% phenol under vacuum at 110°C for 22 h. Corrections were made for loss of serine (15%) and threonine (10%) during hydrolysis. Sugars were quantitated as the trifluoroacetate derivatives of the O-methyl glycosides (13). Inositol was added to each sample as an internal standard. The derivatized samples were analyzed by gas-liquid chromatography (Perkin-Elmer Sigma 2 chromatograph and Sigma 10 data processor) on an SP-240 1 column (6 ft. X $ in. id.). The column temperature was pro~amm~ from 110 to 200°C at a rate of l”C/min.

a substrate for xylosyltransferase (2). In the present experiments, purified CSPG was digested with testicular hyaluronidase, which removed about 80% of the chondroitin sulfate, prior to deglycosylation with HF in pyridine. Proteoglycan could be deglycosylated without hyaluronidase digestion, but recovery of soluble PGHF decreased about twofold, most likely because of increased aggregation and loss of the aggregated material during chromatography. Therefore, enzymatic removal of the majority of chondroitin sulfate chains is recommended before HF deglycosylation. The optimal time of treatment with HF in pyridine was determined by varying the length of the reaction from 0.5 to 8 h. The sugar compositions of the starting material and the resulting PGHF preparations are shown in Table 1. Treatment for 0.5 h removed more than 95% of the xylose and mannose, about 85% of the galactose, fucose, and NeuNAc, 75% of the remaining GalNAc, and 70% of the GlcNAc from the core protein. Longer treatment of up to 8 h removed some additional amounts of sugar. In contrast, no change in the amino acid composition of the core protein was detected following treatment with HF in pyridine from 0.5 to 8 h (Table 2). The recovery of PGHF following ethyl acetate precipitation and Sepharose CL&B chromatography was an indication of the

RESULTS

Preliminary studies have indicated that polyhydrogen fluoride in pyridine may be used to effectively deglycosylate CSPG and that the deglycosylated product, PGHF, was TABLE SUGAR ANALYSES

OF CSPG

1

BEFORE AND AFER

HF DEGL~COSYLA~O~

Sugar (PGHF),O &mg AID,-2B-H Xyfose Fucose Galactose Mannose

GlcNAc GalNAc NeuNAc

85 4.4 210 28 51 100 61

protein

0.5 h

lh

2h

4h

8h

3.3 0.55 35 1.2 15 25 IO

2.9 0.53 30 1.0 13 21 II

2.2 0.37 28 0.75 12 20 13

1.7 Trace 20 Trace 4.0 22 10

0.8 Trace 22 Trace 4.2 24 12

* All values are the average of two dete~inations.

DEGLYCOSYLATION

235

OF PROTEOGLYCAN TABLE 2

AMINO ACID ANALYSES OF CSPG BEFORE AND AFTER HF DEGLYCOSYLATION

(REsIDuEs/~OOO)~

PGHF AID,-2B-H 71.9 93.7 141.5 135.7 69.8 150.2 66.8 64.8 3.5 29.7 82.0 10.0 22.2 10.5 16.7 27.8

Asx Thr Ser Glx Pro GUY Ala Val Met Ile Leu Tyr Phe His LYS A%

0.5 h

lh

2h

4h

8h

71.6 91.0 138.8 133.1 71.9 147.4 74.9 69.0 3.5 30.3 19.2 9.4 21.0 9.3 16.3 29.0

71.6 95.0 134.4 135.4 73.6 142.3 68.8 65.0 4.0 30.6 78.4 12.7 23.1 10.7 17.0 32.2

73.9 93.3 133.8 137.8 73. I 148.1 70.0 66.1 3.8 24.6 75.6 11.9 24.5 10.6 19.0 30.8

72.4 94.4 135.6 137.6 71.0 149.0 69.5 64.5 3.4 31.1 79.2 8.9 23.6 10.5 19.6 28.8

73.4 94.4 139.9 137.7 72.5 145.3 62.3 62.3 4.0 28.0 76.2 11.9 23.1 10.1 19.0 30.5

’ All values are the average of two determinations.

effect of HF on the core protein of CSPG. Since AID,-2B-H was approximately 65% protein and PGHF was about 92% protein, the maximum yield expected after deglycosylation would be approximately 70%. As shown in Table 3, the recovery by this method after 0.5 or 1 h of HF treatment was about

65% of the predicted value. However, recovery decreased with longer treatments, and the yield was only 8% after 16 h. The recovery of protein was also verified by deglycosylation of ‘251-labeledAIDI-2B-H. After 1 h in HF, 98% of the radioactivity was recovered, while only 64% remained following the 8-h treat-

TABLE 3 CHARACTERISTICS

Time in HF/pyridine (h) 0

0.5 1 2 4 8

OF HF DEGLYCOSYLATED

CSPG

Recovery a (%I

Xylose incorporation” (k-W50 PCS) acceptor)

Xylosefserine’ Wmg)

Radioimmune response’ (rg/50% inhibition)

100 45 41 34 25 20

350 10,240 12.530 11.360 10,500 8.500

550 22 19 15 11 5.3

7 11 15 NDb ND ND

’ All values are the average of two determinations. Proteincontentwasdetermined by the method of Lowry et al. (11). Values were also determined by quantitative amino acid analysis, and were within 3% of those obtained calorimetrically. ’ Not determined.

236

OLSON,

KRUEGER,

ment. The decrease in recovery is probably due to decreased solubility of PGHF, which was apparent in both water and 4 M GuHCl. PGHF deglycosylated for 1 h was quite soluble, while the 8-h sample was difficult to solubilize in water and precipitated after freezing and thawing several times. The sample treated with HF for 16 h was sparingly soluble in water, and required several hours to dissolve in 4 M GuHCl. It should also be mentioned that anhydrous HF-deglycosylation of AID,-2B-H was attempted without success. The product was a brown pellet that was insoluble in several solvents, including pyridine and 4 M GuHCl. The ability to accept xylose from UDPxylose in the presence of xylosyltransferase is a rapid means of estimating the effectiveness of the deglycosylation of the core protein. As shown in Table 3, the acceptor prepared after 1 h of HF treatment had the highest xylose incorporation, while longer treatment led to decreased incorporation and a larger K, value for PGHF. However, the ratios of xylose to serine of these preparations (Table 3) indicated that more serine residues were liberated following longer deglycosylation. Most importantly, rapid removal of HF following the deglycosylation reaction resulted in minimal alteration of the core protein, as determined by immunoreactivity by quantitative radioimmunoassay (Table 3). DISCUSSION

Deglycosylation is essential for structural analysis of the protein moieties of complex macromolecules. Thus, several procedures have been described for removal of carbohydrate from glycoproteins and only a few for proteoglycans. Proteoglycans present special problems because of the large amount and diversity of sugar substituents. A portion of carbohydrate can be removed by enzymes that specifically degrade certain of the glycosaminoglycan substituents. Although these methods are gentle and probably do not cause degradation of the protein backbone,

AND

SCHWARTZ

the product is very heterogeneous and still covalently bonded with various oligosaccharide stubs, and thus is not amenable to structural analysis. Of the chemical methods that have been successful for deglycosylating glycoproteins, anhydrous HF at 23°C for 3 h has been reported to cleave all linkages between sugars as well as 0-glycosidic linkage to amino acids, while leaving N-glycosidic linkages and peptide bonds uncleaved (5). Variations on the HF procedure have been used to deglycosylate peptides from immunoglobulin D and ceruloplasmin ( 14) and carcinoembryonic antigen ( 15). Walton et al. (16) have used HF in pyridine and Edge et al. ( 17) have used trifluoromethanesulfonic acid to remove carbohydrate from bovine nasal septum proteoglycan. In neither study was removal quantitated nor the products characterized. We previously reported on the use of HF in pyridine to remove greater than 95% of the chondroitin sulfate chains from rat chondrosarcoma proteoglycan (2). Unfortunately, significant loss of immunoreactivity and inability to obtain homogenous high-molecularweight deglycosylated products suggested structural alteration of the protein core during the deglycosylation treatment. Determination of mimimum reaction time, maintenance of a reducing atmosphere during the reaction, and rapid removal of the HF reagent have now optimized the deglycosylation reaction while minimizing nonspecific degradation. Removal of almost all of chondroitin sulfate chains and most of the N-linked oligosaccharides, except the chitobiose linkage region, was achieved within 30 min of treatment of hyaluronidase-digested proteoglycan. Some of the remaining galactose, GalNAc, and NeuNAc, are probably components of the O-linked oligosaccharides. Removal of sugars following 1 h of treatment was comparable to that reported for glycoproteins treated for 1 h with anhydrous HF (5). Anhydrous HF has been reported to remove GalNAc, galactose, or NeuNAc after 3 h (5), but no further reduction in the amounts of these sugars was

DEGLYCOSYLATION

observed in the present studies when samples of CSPG were treated for more than 1 h with HF in pyridine. Longer times at elevated temperatures would probably remove these components, but this would most certainly lead to increased structural alteration of the peptide as well. We also found that CSPG prepared by deglycosylation with anhydrous HF could not be analyzed because of problems encountered with the solubility of CSPG after the anhydrous HF treatment. The deglycosylated product produced after 1 h of HF treatment was an excellent xylose acceptor and completely soluble in low ionicstrength buffers. Furthermore, it was immunologically comparable to the glycosylated starting material, indicating that our antibody recognizes a protein determinant that is not altered by the HF treatment. The product of deglycosylation procedure is an intact, homogenous polypeptide of molecular weight 2 10,000 by SDS-polyacrylamide gel electrophoresis (18). It is also in agreement with a molecular weight of 210,000, which was calculated for the core protein from the amino acid composition, assuming the presence of seven methionine residues per molecule. Although all results reported here were obtained for rat chondrosarcoma, the HF/pyridine procedure has been used successfully to deglycosylate proteoglycan from embryonic chick and bovine nasal septum cartilage. The procedure should also be useful for deglycosylating other proteoglycans with different types of glycosaminoglycan substituents. ACKNOWLEDGMENTS We are grateful to Dr. Robert Heinrikson for performing the amino acid analysis, Dr. Eugene Goldwasser for the use of the gas-liquid chromatograph, and Dr. John Westley for the computer program used to determine

OF PROTEOGLYCAN

237

kinetic values. The excellent technical assistance of Mr. Daniel Frank is also acknowledged. REFERENCES Coudron, C.. Philipson, L., Ellis, K., and Schwartz. N. B. (1979) Fed. Proc. 38, 651. 2. Coudron. C.. Ellis, K., Philipson, L.. and Schwartz, N. B. (1980a) Biochem. Biophgs. Res. Commun. 92, 618-623. 3. Oegema, T. R., Jr., Hascall, V. C., and Dziewiatkowski, D. D. (1975) J. Biol. Chem. 250, 615 I6159. 4. Keiser, H. D., and Hatcher, V. B. (1979) Connect. 1.

Tissue Res. 6, 229-233.

5. Mot-t, A. J., and Lamport, D. T. A. (1977) Anal. Biochem.

82, 289-309.

6. Sakakibara, S., Kishida, Y ,. Nishizawa, R.. and Shimonishi. S. (1968) Bull. Chem. Sot. Jpn. 41, 438-440. 7. Bradford, M. (1976) Anal. Biochem. 72, 248-252. 8. Bitter, T., and Muir, H. (1962) Anal. Biochem. 4, 330-334. 9. Schwartz, N. B., and Dorfman. A. (1975) Arch. Biochem. Biophvs. 171, 136- 144. 10. Roden. L., Baker, J. R.. Helting, T., Schwartz, N. B., Stoolmiller, A. C.. Yamagata, S., and Yamagata, T. (1972) Methods in Enzymology (V. Ginsburg, ed.), Vol. 28, pp. 638-684. Academic Press. New York. Il. Lowry, 0. H., Rosebrough, N. J., Farr, A. L.. and Randall. R. J. (1951) J. Biol. Cllem. 193, 265275. 12. Spackman, D. H., Stein, W. H., and Moore, S. (1958) Anal. Chem. 238, 622-627. 13. Zanetta, J. P.. Breckenridge, W. C., and Vincendon, G. (1972) J. Chromatogr. 69, 291-304. 14. Tataert, D.. Takahasi, N., and Putnum. F. (1982) Anal. Biochem. 123,430-437. 15. Glassman. J. N. S.. Todd, C. W., and Shively, J. E. (1978) Biochem. Biophys. Res. Commun. 85, 209-216. 16. Walton. A. G., Volger. H. G., and Jaynes, E. N. (1979) Int. J. Biol. Macromol. 1, 89-92. 17. Edge, A. S. B., Faltynek, C. R., Hof, L., Reichert. L. E., and Weber, P. (1981) Anal. Biochem. 118, 131-137. 18. Schwartz, N. B., Habib, G., Campbell. S., D’Elvlyn, D.. Gartner, M., Krueger, R., Olson, C.. and Philipson, L. Fed. Proc.. in press.