the journal of biological no. 10, 5, 654643555, … · · 2001-06-16using carbopac pa1* ... carry...

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol .267, No. 10, Issue of April 5, pp. 654643555, 1992 Printed cn U.S.A.

Structural Analysis of the Linkage Region Oligosaccharides and Unsaturated Disaccharides from Chondroitin Sulfate Using CarboPac PA1*

(Received for publication, October 11, 1991)

Shunichi ShibataS, Ronald J. Midurag, and Vincent C. Hascall From the Proteoglycan Chemistry Section, Bone Research Branch, National Institute of Dental Research, National Institutes of Health, Bethesda, MaTyhnd 20892

Swarm rat chondrosarcoma cell cultures were metabolically labeled with [3sS]sulfate, [3H]glucose, or r3H] glucosamine. Chondroitin sulfate chains were isolated from purified aggrecan using alkaline borohydride treatment and Superose 6 chromatography. Various linkage region oligosaccharide alditols were derived from these chains using sequential chondroitinase digestions (ABC lyase followed by ACII lyase). They were then further processed by mercuric acetate treatment, which removed the 4,5-unsaturated uronosyl residue from the nonreducing end of the linkage, and then &galactosidase digestion which liberated the 2 galactose residues from the xylitol reducing terminus. Alkaline phosphatase digestions were performed to verify the presence of phosphate esters. All linkage region structures were isolated and identified using a combination of Progel-TSK G2500 and CarboPac PA1 chromatography steps in conjunction with monosaccharide analyses. This study revealed that chondroitin sulfate chains from aggrecan synthesized by rat chondrosarcoma cells in vitro have the following properties: 1) three out of every four of their linkage regions carry a phosphate ester on xylose, 2) nearly three out of every five chains begin the repeating disaccharide region with an unsulfated first disaccharide unit, 3) nearly twice as many nonphosphorylated chains have a sulfated first disaccharide than their phosphorylated counterparts, and 4) the vast majority of these chains do not contain sulfated galactose in their linkage regions.

This report also describes a borohydride reduction procedure to confer alkali stability to the 3-substituted, unsaturated disaccharides derived from chondroitinase digests of chondroitin sulfate. Furthermore, a CarboPac P A 1 method is demonstrated that separates these reduced disaccharides with exceptional resolution.

The large, hyaluronan-binding proteoglycan in cartilage, refered to as aggrecan, provides resistance to compressive forces on this tissue. The chondroitin sulfate chains on aggre-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

3 Present address: Oral Anatomy Dept., School of Dentistry, Tokyo Medical and Dental University, 1-5-45, Yushima, Bunkyo-ku, 113 Tokyo, Japan.

To whom correspondence should be addressed National Institute of Dental Research, National Institutes of Health, Bldg. 30, Rm. 106, Bethesda, MD 20892. Tel.: 301-496-4913; Fax: 301-402-0824.

can determine many of its physical and chemical properties. These chains are attached to serine residues in aggrecan’s core protein through a tetrasaccharide refered to as the linkage region oligosaccharide. The carbohydrate structure of this linkage region, GlcAPl+ 3GalP1 + 3GalBl- 4XylP1+ 0- serine,’ is common among chondroitin, dermatan, and heparan sulfate (as well as heparin) chains (1, 2). Structural variation in this linkage oligosaccharide results from phosphorylation of xylose (3-5) and sulfation of the second galactose from the xylose residue (6). The amounts of these phosphate and sulfate groups in the linkage region appear to vary with glycosaminoglycan type as well as with source of tissue and animal species.

Chondroitin ABC lyase degrades the repeating unit structure of chondroitin sulfate to unsaturated disaccharides and yields the following linkage region oligosaccharide: A4,5GlyA/31 + 3GalNAc(+S04)P1 + 4GlcAP1 + 3GalP1 + 3GalP1 + 4Xyl (7, 8). Chondroitin ACII lyase can further digest this hexasaccharide by removing the A4,5GlyA/31 + 3GalNAc(f SO,) disaccharide, thus yielding a modified linkage region tetrasaccharide, i.e. as shown above but with a terminal A4,5GlyA (7, 8). Chemical reactions to degrade the linkage region have been reported (9), but they are nonselective and nonsequential in nature. Mercuric acetate has been demonstrated to remove the 4,5-unsaturated uronosyl from A,,,GlyAPl + 3GlcNAc disaccharides derived from hyaluronan (10). Furthermore, mercuric chloride has been reported to remove the terminal A4,5GlyA residue from the linkage region tetrasaccharide isolated from dermatan sulfate (11). We report here the use of this mercuric acetate reaction to remove selectively the terminal A4.sGlyA residue from the linkage region tetrasaccharide, followed by P-galactosidase treatment to liberate both galactose residues, thereby com- pletely degrading the linkage region in a sequential manner (11).

This study sought a detailed structural analysis of the linkage region oligosaccharides from chondroitin sulfate using recent carbohydrate separation technology in order to address some literature conflicts concerning the presence of galactose sulfate and xylose phosphate in these structures. To these ends, Swarm rat chondrosarcoma cell cultures were metabol-

~

The abbreviations used are: Gal, galactose; Xyl(-OH), xylose (xylitol); GlcA, glucuronic acid A,,,GlyA, 4,5-unsaturated uronosyl; NeuAc, N-acetylneuraminic acid ADi-OS, 2-acetamido-Z-deoxy-3-0- (@-D-gluco-4-enepyranosyluronic acid)-D-galactose; ADi-4S(-6S), 2-acetamido-2-deoxy-3-O-(~-~-gluco-4-enepyranosyluronic acid)-4 (or 6)-O-sulfo-~-galactose; CHAPS, 3-[(3-cholamidopropyl)dimethyl- ammoniol-1-propanesulfonic acid; HEPES, N-2-hydroxyethylpiper- azine-N”2-ethanesulfonic acid; BES, N,N-bis(2-hydroxyethyl)-2- aminoethanesulfonic acid; TES, N-tris(hydroxymethyl)methyl-2- aminoethanesulfonic acid.

6548

Chondroitin Sulfate Linkage Region Oligosaccharides 6549

ically labeled with [35S]sulfate, [3H]glucose, or [3H]glucosamine, and chondroitin sulfate chains were isolated from purified aggrecan. High resolution procedures using CarboPac PA1 have been developed for the isolation of N-linked (12, 13) and 0-linked (14) oligosaccharide alditols. These procedures were used to analyze the linkage region oligosaccharide alditols derived from the chondroitin sulfate chains. All of the major types of linkage region oligosaccharides, and their se- quentially degraded products, were identified and quantified. These procedures were also used to resolve the unsaturated disaccharides derived from chondroitinase digests of the chondroitin sulfate chains after their reduction to disaccharide alditols with borohydride.

EXPERIMENTAL PROCEDURES

Materials-Guanidine HC1, formamide, and cesium chloride were purchased from Bethesda Research Laboratories; Centricon 30 and 3 microconcentrators as well as Centrifree devices were from Amicon; ["s]sulfate (carrier free), ~-[ l -~H]glucose (15.5 Ci/mmol) and ~ - [ 6 - 3H]glucosamine (30 Ci/mmol) from Du Pont-New England Nuclear; Sephadex G-50 (fine), Q-Sepharose, and a Superose 6 (HR 10/30) column from Pharmacia LKB Biotechnology Inc.; a Progel-TSK G2500PWXL column (0.78 X 30 cm) from Toso Haas through Su- pelco; a CarboPac PA1 column (0.4 X 25 cm) from Dionex Corp.; chondroitin ABC lyase (Proteus uulgaris) and chondroitin ACII lyase (Arthrobacter aurescens) from Seikagaku America; mercuric acetate from Aldrich Chemical Co.; bovine serum albumin, insulin, and sodium pyruvate from Sigma. Dulbecco's modified Eagle's minimal essential medium was from the National Institutes of Health Media Unit and fetal bovine serum from GIBCO. Alkaline phosphatase (calf intestine) was from Boehringer Mannheim, and P-galactosidase (bovine testis) was kindly provided by Dr. G . w . Jourdian (University of Michigan, Ann Arbor, MI). AG 50W-X8 (200-400 mesh, H' form) was purchased from Bio-Rad Labs. Optiphase High Safe 111 scintil- lation mixture was obtained from Pharmacia LKB Nuclear, Inc. All other reagents were of the highest purity commercially available.

Cell Culture-Chondrocytes were isolated from the Swarm rat chondrosarcoma by sequential digestion with trypsin and collagenase as previously described (15). After isolation, cells were suspended in Dulbecco's modified Eagle's minimal essential medium containing 20% fetal bovine serum, 100 unit/ml penicillin, 100 pg/ml strepto- mycin, 4.5 mg/ml glucose, 15 mM HEPES, 10 mM BES, and 10 mM TES, pH 7.2. The cells were plated a t lo' cells/lOO-mm plastic culture dish and incubated in 20 ml of medium overnight a t 37 "C under an atmosphere of 5% COZ and 95% air.

Cultures were then labeled for 24 h in 10 ml of Dulbecco's modified Eagle's minimal essential medium containing 100 pCi/ml [3H]glucose, 1 pg/ml insulin (161, 0.5 mg/ml glucose, 0.5 mg/ml sodium pyruvate, 15 mM HEPES, 10 mM BES, and 10 mM TES, pH 7.2. Other cultures were labeled with either 100 pCi/ml [3H]glucosamine or 20 pCi/ml [~'lsS]sulfate in the same medium without pyruvate and with 4.5 mg/ ml glucose.

Isolation of Proteoglycans-Media samples were removed and solid guanidine HC1 added to a final concentration of -4 M and CHAPS to a final concentration of -0.5%. The solutions were eluted on Sephadex G-50 columns equilibrated with 10 M formamide, 50 mM sodium acetate, 0.5 M NaC1,0.5% CHAPS, pH 6.0, to separate labeled macromolecules from unincorporated precursors and to exchange the guanidine HCl for formamide (14, 17). The excluded fraction from each sample was mixed with slurried Q-Sepharose equilibrated in the 10 M formamide solution to give a settled volume -1/10 of the sample volume. The mixtures were gently vortexed several times over 15 min and then centrifuged a t -500 X g for 10 min. Each supernatant was removed and each pellet resuspended twice in 10 volumes of the formamide solution followed by centrifugation to wash the resin. Bound macromolecules were then extracted by resuspending the washed resin in 5 volumes of 4 M guanidine HCl, 50 mM sodium acetate, 0.5% CHAPS, pH 6.0. After centrifugation, the solution phase was removed and the pellet re-extracted with 5 volumes of the guanidine HC1 solvent. After centrifugation, the second extract was combined with the first. For the sample labeled with ["Slsulfate, -85% of the total incorporated radiactivity was recovered in the combined extract.

Cesium chloride was added to each extract to give an initial density of 1.40 g/ml. Isopycnic density gradients were established by centrif-

ugation in a SW-50.2 rotor (Beckman) a t 35,000 rpm, 10 "C, 48 h (18). The bottom 1/6 of each gradient (Dl fractions) was isolated. For the sample labeled with [35S]sulfate, -95% of the 3sS activity in the gradient was recovered in the Dl fraction. T h e D l fractions, containing purified aggrecan monomer, were first concentrated and then subsequently exchanged into water by repeated centrifugation on Centricon 30 microconcentrators pretreated with serum albumin (19).

Isolation of Chondroitin Sulfate Chains-Portions of the purified aggrecan monomer samples were treated with 50 mM NaOH, 1 M sodium borohydride for 24 h a t 45 "C in a final volume of 0.5 ml(20). The excess borohydride was destroyed by adding small aliquots of 5 M acetic acid to the samples on ice with rapid mixing to give a final pH of -5. Absolute methanol (1 ml) was added to each sample followed by drying in a Speed Vac (Savant). Samples were resuspended in 1 ml of methanol, acidified with acetic acid, and redried three times to remove methyl borate derivatives. Samples were redis- solved in 0.25 ml of water and eluted on Superose 6 in 0.5 M pyridinium acetate, pH 5.0, to separate the chondroitin sulfate chains from 0-linked oligosaccharide alditols and N-linked glycopeptides (19, 21). Aliquots of eluant fractions were analyzed for radioactivity, and the chondroitin sulfate peak pooled for each sample as shown in Fig. 2 below. The pooled fractions were dried in a Speed Vac. Recov- eries were 90-95% for the 3sSS-labeled sample.

Enzyme Digestions-Preparations were dissolved in 0.1 M Tris- acetate, pH 7.3 (chondroitinase buffer), and treated with either chondroitin ABC or ACII lyase (50 milliunits/50 pl) for 3 h a t 37 "C (8). For phosphatase digestions, preparations were dissolved in 0.1 M Tris, pH 8.5, with 0.1% bovine serum albumin and digested at 37 "C for 5 h with 5 units of alkaline phosphatase in 50 pl. Phosphatase digests were desalted by elution on Progel-TSK G2500 in 0.5 M pyridinium acetate, pH 5.0. For p-galactosidase digestion, preparations were dissolved in 0.05 M citrate-phosphate buffer, pH 4.3, and digested at 37 "C for 5 h with 10 milliunits of P-galactosidase in 10 pl (22).

Sugar Analyses-Samples were hydrolyzed in 100 pl of 4 M triflu- oroacetic acid a t 100 'C for 3 h (23,24). After removing acid by drying in a Speed Vac, sugars were separated on CarboPac PA1 using three successive, isocratic solvents: 16 mM NaOH for 25 min, 0.15 M sodium acetate in 0.1 M NaOH for the next 20 min, and 1 M sodium acetate in 0.1 M NaOH for an additional 15 min. Fractions of 0.5 ml were collected and analyzed for radioactivity. Internal sugar standards (-1 nmol each of xylitol, galactosamine, galactose, mannose, N-acetylneuraminic acid, and glucuronolactone) were added to each sample and detected by pulsed amperometry (24). Xylitol was prepared from xylose as previously described (21).

Reduction of Disaccharides for Analysis on CarboPac PAl-Disac- charides containing 3-substituted hexosamines can be reduced with borohydride, thereby stabilizing them to alkaline pH (25-28). Here, we report a borohydride reduction method compatible with the buffer conditions of chondroitin lyases which stabilizes the unsaturated disaccharides from chondroitin sulfate to alkali. A sample of chondroitin sulfate labeled with [3H]glucosamine as a precursor was digested with chondroitin ACII lyase. The digests were then centrifuged through a Centricon 3 followed by an equal volume of water to recover quantitatively the digestion products while removing macromolecules in the enzyme preparation. The ultrafiltrate was collected and a portion was borohydride-reduced for 60 min a t room temperature by adding an aliquot of 1 M NaBH4 in 100 p~ NaOH to make a final concentration of 200 mM NaBH,. The reduction reaction was terminated by adding 1 M acetic acid to the sample on ice to a final concentration of 300 mM. Aliquots of the untreated and of the reduced samples were eluted on CarboPac PA1 (see Fig. 2 below).

In a separate experiment, a portion of the chondroitin sulfate preparation labeled with [3H]glucosamine was first digested with chondroitin ABC lyase. This digest was then applied to Progel-TSK G2500 as described below to isolate linkage region oligosaccharides. These linkage region structures were then digested with alkaline phosphatase followed by chondroitin ACII lyase to liberate the remaining unsaturated disaccharide from the linkage region tetrasaccharide. These disaccharides were reduced with borohydride as above prior to CarboPac PA1 analysis.

Isolation of Chondroitin Sulfate Linkage Region Structures-Chon- droitinase ABC digests were eluted on Progel-TSK G2500 with 0.5 M pyridinium acetate, pH 5.0, a t a flow rate of 0.5 ml/min. Fractions of 0.1 ml were collected and aliquots counted for radioactivity. Peaks were pooled as indicated in Fig. 3 below (Linkage A ) and dried. The column was cleaned with 4 M guanidine HC1 after every two to three sample applications.

6550 Chondroitin Sulfate Linkage Region Oligosaccharides

Portions of the Linkage A samples were dissolved in water and analyzed on CarboPac PA1 either directly or after treatment with phosphatase. The column was equilibrated in 0.1 M sodium acetate, 0.1 M NaOH. After sample application and wash for 5 min with starting buffer, a linear gradient of sodium acetate from 0.1 to 1.0 M in 0.1 M NaOH was applied over 90 min followed by 10 min with the 1.0 M sodium acetate, 0.1 M NaOH solution. The flow rate was 1 ml/ min, and fractions of 1 ml were collected and analyzed for radioactivity.

Portions of Linkage A samples were dissolved in chondroitinase buffer and digested with chondroitin ACII lyase. These samples were eluted on Progel-TSK (32500 in 0.5 M pyridinium acetate, pH 5.0, to isolate ACII-generated linkages (Linkage B). The Linkage B peaks were pooled as indicated in Fig. 5 below and dried. Portions of Linkage B samples were dissolved in water and analyzed on CarboPac PA1 either directly or after treatment with phosphatase as described above.

Mercuric Acetate Treatment-Portions of the Linkage B structures with and without a phosphate ester (peaks 5 and 6, see Fig. 5) were recovered from CarboPac PAI, neutralized with acetic acid, concentrated by Speed Vac, and desalted by elution on Progel-TSK G2500

dissolved in 50 pl of 35 mM mercuric acetate, pH 5.0 (10). After 30 in 0.5 M ammonium acetate, pH 6.0. After drying, samples were

min at room temperature, the mercuric ion was removed by adding a 1/10 volume of AG 50W-X8 resin equilibrated in water with periodic stirring. After 5 min, the resin was pelleted by centrifugation. The supernatant fractions, containing more than 90% of the radioactivity, were recovered and eluted on Progel-TSK G2500 (Linkage C) prior to analysis on CarboPac PAL

0-Galactosidase Treatments-Portions of Linkage C structures without and with a phosphate ester (peaks 7 and 8, see Fig. 6) were recovered from Progel-TSK G2500 chromatography. After drying, samples were dissolved in galactosidase buffer and then digested with @galactosidase. A portion of each digest was eluted on CarboPac PA1 using monosaccharide analysis conditions. A portion of the peak 8 sample was treated with alkaline phosphatase prior to 6-galactosidase digestion and monosaccharide analysis (see Fig. 7).

RESULTS

Analyses of Reduced Disaccharides on CarboPac PAl- Chondroitin lyases are eliminases which release 81 + 3 disaccharides having a 4,5-unsaturated uronosyl residue on the nonreducing end of an N-acetylgalactosamine residue (Fig. 1) (29). When [3H]glucosamine is used as a labeling precursor, virtually all of the radioactivity is found in the galactosamine residues (see “Discussion”), and when chondroitin ACII lyase is used, all of the disaccharides are released from the linkage region (Linkage B, Fig. 1) (7). Accordingly, cultures of chon-

Chondroitin Sulfote 1 *oso3 *psos *?SO, iOPO,

C ~ I N A C - ( G ~ C A - C O ~ N A C ) ~ - G ~ C A - G O ~ N A C - G I C A - G O I - G O ~ - X ~ I - O H

1 Chondroitin ABC lyase

tOSOJ GolNAc

J. ipsos- + AClyA-GolNAc-GlcA-Gal-Gol-Xyl-OH

n[AClyA-dolNAc] I Chondroitin ACll lyose

i O S O , I

AGlyA-GolNAc + AGlyA-Gal-GOl-Xyl-OH

Mercuric ocetote

AGlyA keto acid + Gol-Gal-Xyl-OH

Goloctose (2) + Xyl-OH

FIG. I. Schematic representation of chondroitin sulfate linkage region structures resulting from various enzymatic and chemical treatments of chondroitin sulfate. AGlyA, 4,5- unsaturated uronosyl.

drosarcoma chondrocytes were metabolically labeled with [3H]glucosamine and radiolabeled aggrecan monomers purified from the media as described under “Experimental Pro- cedures.” A portion of this aggrecan sample was treated with alkaline borohydride to release chondroitin sulfate chains, 0- linked oligosaccharide alditols and N-linked glycopeptides (30, 31). This sample was then eluted on Superose 6 to separate the larger chondroitin sulfate chains from the smaller oligosaccharide fractions (Fig. 2, top panel inset). About 90% of the 3H activity was recovered in the chondroitin sulfate peak. The peak Kd value (0.44) corresponds to a molecular mass of -37 kDa when compared with chondroitin sulfate standards (21, 32). A chondroitin ACII lyase digest was prepared for this chondroitin sulfate sample to provide a total disaccharide analysis of the chains.

Chondroitin sulfate disaccharides are unstable in alkali because they undergo a “peeling” reaction (33, 34) releasing the uronosyl substituent and the sulfate group (35). Thus, when a portion of the above digest was eluted on CarboPac PA1 using the normal alkaline elution conditions, the disaccharides decomposed and the resulting degradation products eluted as a broad peak in the beginning of the sodium acetate gradient (Fig. 2, top). The sulfate group was quantitatively released from these chondroitin sulfate disaccharides during this decomposition (data not shown). These disaccharides are pH-stabilized when reduced to their hexosaminitol derivatives (25, 27). In this case, each disaccharide-hexosaminitol elutes in a narrow peak at a characteristic position in the gradient (Fig. 2, bottom). The positions of standards of reduced unsul-

CarboPac PA1 CarboPac PA1

0 0 10 20 30 40 50 60 70 80 90

Fraction number

FIG. 2. CarboPac PA1 analysis of all unsaturated disaccharides liberated from chondroitin sulfate by chondroitin ACII lyase without (top) and with (bottom) prior reduction with sodium borohydride. Chondroitin sulfate ( C S ) chains labeled from [3H]glucosamine were recovered from Superose 6 chromatography (bar In top panel inset) and then directly digested by chondroitin ACII lyase without prior ABC lyase treatment. The abbreviations for the unsaturated disaccharide standards are defined in footnote 1. Shown here are the peak elution positions for the reduced forms of these standards. The programmed sodium acetate gradient is shown as a dashed line in the top panel. The bottom panel inset depicts an expanded scale of the borohydride-reduced profile revealing some minor peaks. The asterisk in the top panel and arrow in the bottom panel inset indicate minor peaks discussed in the text. Note that the 3H activity scale for the Superose 6 profile is IO6 cpm.

Chondroitin Sulfate Linkage Region Oligosaccharides 6551 TSK 62500

; .?

n t 2

0 20 40 60 80 100 Fractlon number

FIG. 3. Analysis of the linkage region oligosaccharide alditols derived from chondroitin sulfate digested with chondroitin ABC lyase. The bar in the top panel (denoted Linkage A ) indicates the linkage region preparation recovered for further analysis on CarboPac PA1 before (middle panel) and after (bottom panel) alkaline phosphatase treatment. The dashed line in the middle panel denotes the programmed sodium acetate gradient as described in Fig. 2. The elution position of free sulfate (SO,') is indicated in the middle panel. Baseline values (<5 cpm) in the middle and bottom panels are suppressed to enhance the clarity of the profiles.

fated (ADi-OS,), 4-sulfated (ADi-4Sr), and 6-sulfated (ADi- 6S,) disaccharides are indicated. Radioactivity in these three disaccharides constituted -93% of the total 3H, and they were in the proportion of 4.5:95:0.5 for ADi-OSr:ADi-4S,:ADi-6S,. The identity of the peak in the ADi-6S, position remains uncertain because an unknown peak with the same amount of radioactivity (Fig. 2, top, asterisk) eluted near the same position in the unreduced sample, indicative of alkali stability which would not be the case for ADi-6s.

A minor peak (-4%) eluted between ADi-OS, and ADi-4Sr (Fig. 2, arrow in bottompanel inset) and is tentatively assigned as N-acetylgalactosaminitol-S04 for the following reasons: (a) it contains a sulfate ester, (b) it is monosaccharide in size, and ( c ) it contains galactosaminitol as its only sugar compo- nent (data not shown). If every chain (-37 kDa) terminated with a GalNAc-SO, (-300 Da), then the expected amount of 'H activity in N-acetylgalactosaminitol-S04 would be -1%. The observed amount of this product is -4-fold greater than expected. Thus, it appears to be derived primarily from GalNAc-S04 residues released from some of the unsaturated disaccharides by a contaminating uronidase activity in the chondroitin lyase preparation as documented previously (29).'

Isolation of Linkage A Structures from Chondroitin ABC Lyase Digests-Aliquots of chondroitin sulfate in which [35S] sulfate and [3H]glucose were used as precursors were mixed and digested with chondroitin ABC lyase. The digest was eluted on Progel-TSK G2500 (Fig. 3, top). The linkage region

Detectable amounts of uronidase activity were measured in both chondroitin lyase preparations, with the levels of this activity varying in each lot of enzyme tested (R. J. Midura and V. C. Hascall, unpublished observations).

oligosaccharide alditols (Linkage A) were separated from the disaccharide digestion products. The linkage oligosaccharide alditols contain one disaccharide from the chain (Fig. 1) since this enzyme will not remove the first disaccharide (7). The Linkage A peak contained -4% of the 3H and -0.4% of the 35S in the digest. Approximately 3% of the 3H and -4% of the 35S eluted near the column total volume after the disaccharide peak; molecules such as GalNAc-S04 would elute in this peak.

Aliquots of the Linkage A fraction were analyzed on CarboPac PA1 either directly (Fig. 3, middle) or after treatment with alkaline phosphatase (Fig. 3, bottom). Four peaks with 3H activity (peaks 1-4) were observed before and two (peaks 1 and 3) after phosphatase treatment. Some free [35S] sulfate was observed in the direct analysis (Fig. 3, middle) which resulted from the decomposition of small amounts of contaminating disaccharide recovered in the Linkage A sam- le.^ This sulfate peak is absent in the sample digested with phosphatase (Fig. 3, bottom) because the digest was desalted on Progel-TSK G2500 before analysis on CarboPac PA1, thus removing any residual disaccharides.

Peaks 1 and 2 contained no 35S activity. Sugar analyses for peak 1 (-1500 3H cpm) recovered after phosphatase treatment (Fig. 3, bottom) revealed 3H label in galactosamine (-195 cpm), galactose (-500 cpm), and xylitol (-260 cpm) yielding a ratio of 0.75:l.g:l.O. The less than stoichiometric amount of 3H label in galactosamine is consistent with previous results which showed that there is some loss of label when [3H] glucose is metabolically converted to hexosamine (5). Label in unsaturated uronosyl and glucuronate was not recovered because both readily degrade in aqueous acid at high temperature (36, 37). However, the amount of 3H label unaccounted for in these sugar analyses (36%) approximates the amount which should have been incorporated into 1 residue each of unsaturated uronosyl and glucuronate per xylitol. The results are consistent with the structures shown in Fig. 4 in which the first disaccharide in peak 1 is unsulfated. Peak 2 appears to be a phosphorylated version of peak 1 since it is converted to peak 1 by phosphatase treatment (also see the sugar analyses for Linkage B below).

Peaks 3 and 4 both contained 35S activity in addition to the 3H activity. Also, peak 4 was converted to peak 3 after phosphatase treatment. This suggests that the first disaccharide in peaks 3 and 4 contains a sulfate ester (see Linkage B analyses below) and that peak 4 additionally contains a phosphate ester, consistent with the structures shown in Fig. 4. The 3H recoveries of peaks 1-4 from CarboPac PA1 (Fig. 4) would argue that the first disaccharide is sulfated -36% of the time and that -77% of the linkage structures are phosphorylated. Additionally, phosphorylated linkage structures contain a much lower proportion of sulfate on the first disaccharide than do nonphosphorylated linkages (29 us. 59%).

Isolation of Linkage B Structures from Chondroitin ACII Lyase Digestion of Linkage A Fraction-An aliquot of the Linkage A fraction (Fig. 3, top) was treated with chondroitin ACII lyase to remove the residual disaccharide, and the digest was eluted on Progel-TSK G2500 (Fig. 5, top). The Linkage B oligosaccharide alditols (-67% of the 3H) eluted ahead of the disaccharide peaks (-90% of the 35S and -29% of the 3H). The 35S peak, representing the sulfated disaccharides, eluted behind the Linkage B peak but slightly ahead of the 3H disaccharide peak, which contains the predominant nonsul-

Chondroitin sulfate disaccharide analysis by the method of Ze- brower et al. (39) detected the presence of an %-labeled ADi-4S contaminant in the Linkage A sample that accounted for this amount of 35S activity (S. Shibata, R. J. Midura, and V. C. Hascall, unpublished observations).


Peak 1 Linkage A

g% AGlyA-GolNAc-GlcA-Gal-Gal-Xyl-OH

Peak 2 ?PO,

53% AGlyA-GoINAc-GlcA-Gal-Gol-Xyl-OH

Peak 3 oso, 1% A G I ~ A - ~ ~ I N A ~ - G I ~ A - G ~ I - G ~ ~ - x ~ ~ - o H

Peak 4 ?SO, y o ,

2256 AGlyA-GolNAc-GlcA-Gal-Gal-Xyl-OH

Peak 5 Linkage B

25% AGlyA-Gal-Gal-Xyl-OH

Peak 6 y o 3

6% AGlyA-Gal-Gal-Xyl-OH

Peak 7 Linkage C

Gal-Gal-Xyl-OH

Peak 6

Gal-Gal-Xyl-OH

FIG. 4. Data summary for linkage region oligosaccharide alditols A, B, and C isolated from chondroitin sulfate. Linkage region designations are shown in Fig. 1. Percentages shown reflect the amount of 3H label from glucose incorporated into each peak as a proportion of the total recovered sample (data from Fig. 3 for Linkage A. and Fig. 5 for Linkage B). These proposed structures are deduced from monosaccharide compositional analysis described in the text as well as data shown in Fig. 7. The abbreviations are as defined in Fig. 1.

fated disaccharides. Only a small amount (-8%) of the 35S activity in the profile eluted in the region of Linkage B structures.

The Linkage B peak was recovered and aliquots were analyzed on CarboPac PA1 directly (Fig. 5, middle) or after digestion with alkaline phosphatase (Fig. 5, bottom). Two 3H peaks (peaks 5 and 6) were observed before, and one (peak 5) after, phosphatase digestion. These peaks eluted in characteristic locations in the sodium acetate gradient well separated from peaks 1-4. Sugar analyses revealed ratios of galactose:xylitol of 2.O:l.O for peak 5 and 2.0:0.5 for peak 6.* The low recovery of xylitol in peak 6 after acid hydrolysis is characteristic of phosphorylated xylitol (3,5). The results are consistent with the structures shown in Fig. 4. The overall recovery for 3H label in these two peaks was -94%. About 73% of the Linkage B structures were phosphorylated, in good agreement with the results described for Linkage A above. Both CarboPac PA1 analyses contain free [35S]sulfate from the degradation of contaminating disaccharides. In this case, unlike the results for Linkage A analyses, the TSK desalting step after the phosphatase digestion failed to remove all of the residual disaccharide because of a partial overlap of the sulfated disaccharide peak with the unphosphorylated Link- age B alditol peak. A minor peak (-3% of the 3H label) in the CarboPac PA1 analysis of Linkage B structures was observed to contain some 35S activity (Fig. 5, middle, asterisk). It was not analyzed further because of its low abundance, but may represent a structural variant of the linkage region containing sulfate (6).

Sugar analysis of peak 5 (-1500 cpm) gave -670 cpm of galactose and -330 cpm of xylitol. Analysis of peak 6 (-1000 cpm) revealed -490 cpm of galactose and -120 cpm of xylitol, while after alkaline phosphatase digestion it yielded -475 cpm of galactose and -230 cpm of xylitol.

TSK 02500 30 Unkage B 15

0.0 0.5 K 1.0

1: o 20 40 60 a0 loo

Fractlon numtmr

FIG. 5. Analysis of the linkage region oligosaccharide alditols generated by chondroitin ACII lyase treatment of Linkage A structures. In the top panel, the arrow indicates the peak elution position of the Linkage A sample, while the bar (denoted Linkage B ) depicts the linkage region preparation recovered for further analysis on CarboPac PA1 before (middle panel) and after (bottom panel) alkaline phosphatase treatment. The dashed line in the middle panel represents the programmed sodium acetate gradient with the elution positions of the Linkage A structures (1-4) presented as references (V). The elution position of free sulfate (SO,=) is indicated in the middle and bottom panels. The asterisk in the middle panel denotes a minor peak discussed in the text. Baseline values ( 4 cpm) in the middle and bottom panels are suppressed to enhance the clarity of the profiles.

Chondroitin ABC lyase treatment and elution on Progel- TSK G2500 as described for Fig. 3 above was used to isolate a Linkage A fraction from chondroitin sulfate that had been labeled with [3H]glucosamine as a precursor (data not shown). A portion of this Linkage A sample was treated with alkaline phosphatase and then with chondroitin ACII lyase. The digestion products were then reduced with sodium borohydride and analyzed on CarboPac PA1 to determine directly the composition of the remaining disaccharide. The predominant peaks (-92% of the 3H)5 were ADi-OS, and ADi-4Sr in a ratio of about 3:2 (data not shown), close to that expected from the results described above (Fig. 3, bottom, and Fig. 4).

Isolation of Linkage C Structures from Mercuric Acetate Treatment of Peaks 5 and 6-Samples of peaks 5 and 6 were isolated from CarboPac PA1 analyses as shown in Fig. 5. Aliquots were treated with mercuric acetate (10) and eluted on Progel-TSK G2500 (Fig. 6, left panels). In each case the treated samples revealed a major peak eluting slightly later than their respective untreated controls (arrows) with the one derived from peak 6 eluting slightly earlier than that derived from peak 5. Each major peak was followed by a small, later eluting one, 23 and 27% of the total label in peaks 5 and 6, respectively, which represent the products from the mercuric

A minor peak (-5%) was observed in the position of peak 5 , the nonphosphorylated Linkage B structure. It is possible that a labeled aldose contaminant in the [3H]glucosamine preparation (-98% purity) could have resulted in this level of 3H associated with the Linkage B structure.


0.0 0.5 1.0 0 20 40 60 80 100

Kd Fraction number

FIG. 6. Analysis of the linkage region oligosaccharide alditols resulting from the mercuric acetate treatment of Linkage B structures. Samples of peaks 5 and 6 (Fig. 5, middle panel) were isolated from a separate CarboPac PA1 run and desalted on TSK G2500. These samples were treated with mercuric acetate and linkage region preparations (represented by bars in the two panels to the left) were recovered from TSK G2500 (denoted Linkage C (-PO,) from peak 5 and Linkage C (+PO,) from peak 6). Arrows in these panels indicate the peak elution positions of the respective Linkage B structures before treatment. CarboPac PA1 analyses of these Linkage C structures (peaks 7 and 8) are shown in the two panels to the right. Baseline values (<5 cpm) are suppressed to enhance the clarity of the profiles. The dashed lines in these panels represent the programmed sodium acetate gradient with the elution positions of Linkage A (V) and B (V) structures shown as references. Small amounts (-250 cpm each) of Linkage C (+PO,) were treated with alkaline phosphatase (0) and then applied to TSK G2500 and CarboPac PA1.

degradation of the 4,5-unsaturated uronosyl residue at the nonreducing end of the Linkage B oligosaccharide alditols (Fig. 1). This reaction exposes the nonreducing terminal galactoses thereby yielding Linkage C structures. After alkaline phosphatase treatment, the Linkage C structure derived from peak 6 then eluted from Progel-TSK G2500 in a position similar to that for the structure derived from peak 5 (Fig. 6, left panels).

The Linkage C peaks were recovered as indicated by the bars and analyzed on CarboPac PA1 (Fig. 6, right panels). The product derived from nonphosphorylated peak 5 did not bind under the conditions used and eluted as a peak (peak 7) in the breakthrough fractions, as would be expected for nonphosphorylated Linkage C which no longer contains a strong negative charge. The product derived from phosphorylated peak 6 bound to the column and eluted as a single peak (peak 8) at a characteristic position in the sodium acetate gradient distinct from the locations of peaks 1-4 and 6, but near to that for peak 5. After aIkaline phosphatase treatment, peak 8 no longer bound to CarboPac PA1, eluting instead at the position of peak 7 (Fig. 6, lower right panel). Sugar analyses of these linkage oligosaccharide alditols were performed as described below, and the results are consistent with the structures shown in Fig. 4.

Treatment of Linkage C Structures with @-Galactasidme- Portions of both Linkage C oligosaccharide alditols (Fig. 6, left panels) were treated with @-galactosidase and the digests analyzed on CarboPac PA1 using an elution protocol for monosaccharide analysis (Fig. 7). In each case, about two- thirds of the 3H eluted in the position of galactose. In the digest of the nonphosphorylated structure, the remaining one-

3

2

f 8 - = r '

E 3

-= 2

w & e c

N 0 u

5

0

1

CarboPac PA1 Linkage C (-PO4 )

I +phosphatase I

4 o 20 40 60 ao l oo

Fraction number FIG. 7 . CarboPac PA1 monosaccharide analyses of Linkage

C structures after 8-galactosidase digestion. The peak elution positions of various monosaccharide internal standards are shown. Baseline values (<5 cpm) are suppressed to enhance the clarity of the profiles. The dashed line in the bottom panel depicts the isocratic, step-elution program: I , 16 mM NaOH; I I , 100 rnM NaOH/150 mM sodium acetate; III , 100 mM NaOH/1 M sodium acetate. A portion of Linkage C (+PO,) was treated with alkaline phosphatase (0) prior to @-galactosidase digestion and represents about one-half the amount used for 8-galactosidase digestion alone.

third of the label eluted near the breakthrough of the column where xylitol elutes. In the digest of the phosphorylated form, the remaining one-third of the 3H eluted as a single, more tightly bound peak during the second eluant which contained 150 mM sodium acetate. When a portion of this phosphorylated linkage was treated with alkaline phosphatase prior to @-galactosidase digestion, this tightly bound peak disappeared with the appearance of the same proportion of label eluting in the position of xylitol. The results indicate that the more tightly bound peak contains phosphorylated xylitol. Thus, @- galactosidase effectively removed the 2 galactose residues from both forms of Linkage C.

DISCUSSION

This study provides structural analyses of the linkage region oligosaccharides and unsaturated disaccharides derived from the chondroitin sulfate chains on the aggrecan proteoglycan synthesized by Swarm rat chondrosarcoma cells in culture. High performance anion-exchange chromatography designed for carbohydrate analysis (CarboPac PA1) was used to isolate each type of linkage oligosaccharide with superior resolution and no apparent chromatographic artifacts. Procedures are described that removed the unsaturated uronosyl and the 2 galactose residues from the linkage oligosaccharides. Also described is a procedure to reduce the @l + 3-linked unsaturated disaccharides derived from chondroitinase digests of chondroitin sulfate. Such reduced disaccharides were base- stabilized, and preliminary results show that they can be separated by chromatography on CarboPac PA1 at alkaline pH with exceptional resolution.

This study revealed that about three out of every four


chondroitin sulfate chains are linked to the aggrecan core protein via xylose which carries a phosphate ester. Similar ratios for phosphoxylose-linked glycosaminoglycans have been reported for chondroitin sulfate isolated from chondrosarcoma cell cultures (5) and for heparan sulfate isolated from bovine lung tissue (4). This proportion of phosphorylated xylose appears to be greater than that reported for chondroitin sulfate chains derived from the intact chondrosarcoma tumor (3, 6). The apparent discrepancy may reflect a difference between the tumor-state in vivo uersus cultured cells, or may be related to phosphatase activities in enzyme preparations (6, 11) as well as phosphatases binding to chromatographic matrices in a highly active form.6 The functional significance of phosphorylated xylose in the oligosaccharide region that links glycosaminoglycans to their core proteins is as yet unknown.

Interestingly, there is a lower tendency for the first disaccharide to be sulfated in those chondroitin sulfate linkage oligosaccharides carrying a phosphate ester on xylose. In this study, only about one out of every three phosphorylated linkage regions had a sulfated first disaccharide, while nearly two out of every three nonphosphorylated linkages contained one. The significance of this observation is not clear because it is not known whether phosphorylation of xylose occurs in a different (and perhaps earlier) Golgi subcompartment than that for the sulfation of the first N-acetylgalactosamine residue. Thus, it seems that the presence of one anionic group in the linkage region appears to interfere with the addition of the other group to that same structure.

It is also apparent that nearly three out of every five chondroitin sulfate chains begin the repeating disaccharide region with an unsulfated disaccharide which yields ADi-OS after chondroitinase treatment. When intact chondroitin sulfate chains were digested with chondroitin ACII lyase, the resultant disaccharide composition indicated that -95% were sulfated (virtually all ADi-4s). Thus, there is a 25-30-fold- greater frequency that the first disaccharide in these chondroitin sulfate chains will be unsulfated as opposed to being ~ulfated.~ This observation is consistent with previous reports indicating a predominance of an unsulfated disaccharide in the linkage region (11, 38). Such a great preference for an unsulfated first disaccharide in these chains might be the result of (a) a lack of necessary recognition signals for the N-acetylgalactosamine sulfotransferase to act efficiently at that site, (6) the presence of 2-phosphoxylose in the linkage region which might precede and inhibit the sulfation of that site, or (c) steric hinderance which blocks access of the sulfotransferase near the linkage region.

Given the amount of 3H label from a [l-3H]glucose precursor detected in the linkage oligosaccharide released from chondroitin sulfate after digestion with chondroitin ABC lyase, we estimate that the average mass of the chondroitin sulfate chains, including its linkage, is nearly 41 kDa.’ This

Alkaline phosphatase-digested samples must be deproteinized prior to column chromatography (S. Shibata, R. J. Midura, and V. C. Hascall, unpublished observations).

Three out of five chondroitin sulfate chains have an unsulfated first disaccharide, and only 5% of the total disaccharides in the chain are unsulfated. Therefore, the relative frequency of an unsulfated disaccharide compared to that of a sulfated disaccharide in the first position would he 28.5 (= (3/5)/(2/95)).

* About 4% of the 3H label from [3HJglucose is incorporated into Linkage A structures which have 5.75 labeled aldose equivalents [3(3 aldoses) + 2(2 hexuronates) + 0.75(1 hexosamine)]. Accordingly, there would be -144 total aldose equivalents per chondroitin sulfate chain (= 5.75/0.04). Excluding the linkage region tetrasaccharide and the nonreducing end hexosamine, the total aldose equivalents in the disaccharide-repeating units per chondroitin sulfate chain would be

estimate agrees quite well with the calculated value for the average mass of these chains (-37 kDa) as determined from their elution position on a calibrated Superose 6 column. The above values are nearly twice the reported weight average of chondroitin sulfate chains isolated directly from the chondrosarcoma tumor (7, 32). Other investigators (5 , 15) have also observed larger sized chondroitin sulfate chains synthesized by chondrosarcoma cells in vitro. The exact reasons for this difference in chain size are not known, however differences in tumor source, nutrient conditions in vitro (15), or the amount of matrix surrounding the chondrocytes may be fac- tors which influence chondroitin sulfate metabolism and chain length.

Sugahara et al. (6) reported a novel sulfation on the 4- position of the second galactose residue from the xylose in the linkage region of chondroitin sulfate isolated from the Swarm rat chondrosarcoma tumor. However, these authors estimated that less than 10% of the linkages would contain 4-0-sulfated galactose. Further, Fransson et al. (11) reported that the linkage oligosaccharides from pig skin dermatan sulfate do not contain detectable amounts of sulfated galactose. In the present study, none of the chondroitin sulfate linkage oligosaccharides isolated in significant proportion contained sulfated galactose. After chondroitin ACII lyase digestion, we could account for nearly all of the linkage oligosaccharides as either of two structures (peaks 5 and 6), neither of which contains sulfated galactose. If a linkage structure containing galactose 4-sulfate existed in the sample and escaped detection, then it would constitute no more than 3-4% of the total linkage oligosaccharide population on the prote~glycan.~ Close inspection of Fig. 5 (middle panel, asterisk) reveals the presence of a minor peak (-3% of the 3H label) in the Linkage B preparation that also contains 35S activity. It eluted from the CarboPac PA1 column in a position consistent with it containing only one negatively charged ester group and may represent a minor species of linkage region oligosaccharide. Further structural analysis is needed to con- firm whether this peak could represent a linkage structure containing sulfated galactose.

In addition to analyzing chondroitin sulfate linkage oligosaccharides, we also successfully developed a method to analyze chondroitin sulfate disaccharides on a high resolution, carbohydrate analysis column (CarboPac PA1). The alkaline conditions of this chromatography provide superior resolution capabilities. However, /3l + 3-linked disaccharides in free- aldehyde form are unstable in alkali. Reducing these disaccharides to their respective alcohol forms using borohydride conferred a complete stabilization to alkaline pH. We have now optimized the CarboPac PA1 procedure to provide a complete separation of all the mono-, di-, and unsulfated disaccharides from chondroitin and dermatan sulfate as well as from hyaluronan.” Furthermore, the resolution capacity of this chromatography is such that we can now analyze some minor and as yet unknown components of chondroitin sulfate,

-139 (= 144 - 4.75). With an aldose equivalent per disaccharide of 1.75, there would be -79 disaccharides/chain (= 139/1.75). At a mass of -500 Da/disaccharide, -800 Da/linkage region tetrasaccharide,

these chondroitin sulfate chains would he -41,000 Da. and -300 Da/nonreducing end hexosamine, the total mass of one of

Two out of every five Linkage A structures have a sulfated first disaccharide. After further digestion with chondroitin ACII lyase, the

B) is only 8% of the total 35S label associated with the Linkage A 35S radioactivity left on the linkage region tetrasaccharide (Linkage

sample (-90% is incorporated in the first disaccharide). Thus, only -3.6% of the Linkage A structures could potentially carry a sulfate ester other than that on the first disaccharide (= (8190) X (2/5)).

lo R. J. Midura and V. C. Hascall, manuscript in preparation.


thus providing a more detailed structural analysis beyond what has been published to date.

Acknowledgments-We express our gratitude to Drs. Masaki Yan- agishita and Neal S. Fedarko for their input and critical review of this manuscript. We would also like to thank Dr. Anthony Calabro, Jr. for his advice on culturing the Swarm rat chondrosarcoma cells and on preparing the aggrecan proteoglycan.

REFERENCES 1. Lindahl, U., and Roden, L. (1972) in Glycoproteins (Gottschalk,

A., ed) pp. 491-517, Elsevier Scientific Publishing Co., New York

2. Roden, L. (1980) in The Biochemistry of Glycoproteins and Pro- teoglycans (Lennarz, w. J., ed) pp. 267-371, Plenum Publishing Co., New York

3. Oegema, T. R., Jr., Kraft, E. L., Jourdian, G. W., and Van Valen,

4. Fransson, L.-A., Silverberg, I., and Carlstedt, I. (1985) J. Biol. Chem. 260,14722-14726

5. Lohmander, L. S., Hascall, V. C., Yanagishita, M., Kuettner, K. E., and Kimura, J. H. (1986) Arch. Biochem. Biophys. 250 , 211-227

6. Sugahara, K., Yamashina, I., De Waard, P., Van Halbeek, H., and Vliegenthart, J. F. G. (1988) J. Biol. Chem. 2 6 3 , 10168- 10174

7. Hascall, V. C., Riolo, R. L., Hayward, J., Jr., and Reynolds, C. C. (1972) J. Bwl. Chem. 247,4521-4528

8. Oike, Y., Kimata, K., Shinomura, T., Nakazawa., K., and Suzuki, S. (1980) Biochem. J . 191, 193-207

9. Gunnarsson, A., Svensson, S., and Roden, L. (1984) Carbohydr. Res. 133,75-82

10. Ludwigs, U., Elgavish, A., Esko, J. D., Meezan, E., and Rodhn, L. (1987) B i o c k m . J. 245,795-804

11. Fransson, L.-A., Havsmark, B., and Silverberg, I. (1990) Biochem.

12. Hardy, M. R., and Townsend, R. R. (1988) Proc. Natl. Acud. Sci.

13. Townsend, R. R., Hardy, M. R., Hindsgaul, O., and Lee, Y. C. (1988) Anal. Biochem. 174,459-470

14. Midura, R. J., McQuillan, D. J., Benham, K. J., Fisher, L. W., and Hascall, V. C. (1990) J. Biol. Chem. 265, 5285-5291

15. Kimura, J. H., Hardingham, T. E., Hascall, V. C., and Solursh, M. (1979) J. Bwl. Chem. 254, 2600-2609

T. R. (1984) J. BWl. Chem. 2 5 9 , 1720-1726

J. 269,381-388

U. S. A. 85,3289-3293

16.

17.

18.

19.

20. 21.

22.

23. 24.

25. 26.

27.

28.

29.

30.

31.

32. 33. 34.

35.

36.

37.

38.

39.

Stevens, R. L., Nissley, S. P., Kimura, J. H., Rechler, M. M., Caplan, A. I., and Hascall, V. C. (1981) J. Biol. Chem. 256,

Fedarko, N. S., Termine, J. D., and Gehron Robey, P. (1990)

Hascall, V. C., and Kimura, J. H. (1982) Methods Enzymol. 82,

Midura, R. J., Toledo, 0. M. S., Yanagishita, M., and Hascall, V.

Carlson, D. M. (1968) J. Biol. Chem. 243, 616-626 Midura, R. J., and Hascall, V. C. (1989) J. Biol. Chem. 264,

Distler, J. J., and Jourdian, G. W. (1973) J. Biol. Chem. 248,

Lohmander, L. S. (1986) And. Biochem. 154,75-84 Hardy, M. R., Townsend, R. R., and Lee, Y. C. (1988) Anal.

Lloyd, K. O., and Kabat, E. A. (1969) Carbohydr. Res. 9,41-48 Weber, P., and Winder, R. J. (1969) Arch. Biochem. Biophys.

Mayo, J. W., and Carlson, D. M. (1970) Carbohydr. Res. 15,300-

Glaser, J . H., and Conrad, H. E. (1979) J. Biol. Chem. 254,658&

Yamagata, T., Saito, H., Habuchi, O., and Suzuki, S. (1968) J.

Yanagishita, M., and Hascall, V. C. (1979) J. Bwl. Chem. 264 ,

Lohmander, L. S., De Luca, S., Nilsson, B., Hascall, V. C., Caputo, C. B., Kimura, J. H., and Heinegird, D. (1980) J. Biol. Chem.

2045-2052

Anal. Biochem. 188,398-407

769-800

C. (1989) J. Biol. Chem. 2 6 4 , 1414-1422

1423-1430

6772-6780

Biochem. 170, 54-62

129,534-538

303

6597

Biol. Chem. 243,1523-1535

12355-12364

255,6084-6091 Wasteson, A. (1971) J. Chromutogr. 59,87-97 Ballou, C. E. (1954) Adv. Carbohydr. Chem. 9,59-95 Whistler, R. L., and BeMiller, J. N. (1958) Adu. Carbohydr. Chem.

Inoue, H., Nakayama, R., Otsu, K., Habuchi, H., Suzuki, S., and

Fransson, L.-A., Roden, L., and Spach, M. L. (1968) Anal.

Hjerpe, A., Antonopoulos, C. A., Classon, B., Engfeldt, B., and

Fransson, L.-A. (1985) in The Polysaccharides (Aspinall, G. O.,

Zebrower, M. E., Kieras, F. J., and Brown, W. T. (1986) Anal.

13,289-329

Nakanishi, x. (1986) Curbohydr. Res. 155 , 277-282

Biochem. 2 3 , 317-330

Nurminen, M. (1982) J. Chromatogr. 235,221-227

ed) Vol. 3, pp. 337-415, Academic Press, New York

Biochem. 157,93-99

the journal of biological no. 10, 5, 654643555, … · · 2001-06-16using carbopac pa1* ... carry...

Documents