Nov 25, 2015 - cially initiated onto exogeneous &D-xylosides also ac- quire the 2-phosphoxylose ..... lose on one out of 3-4 chondroitin sulfate chains. A linkage ... droitin AC-I1 lyase degradation of chains released by alkaline borohydride ...
THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1985 by The American Society of Biological Chemists, Inc.
Vol. 260, No. 27, Issue of November 25, pp. 14722-14726,1986 Printed in U.S.A.
Structure of the Heparan Sulfate-ProteinLinkage Region DEMONSTRATION OF THE SEQUENCE GALACTOSYL-GALACTOSYL-XYLOSE-2-PHOSPHATE* (Received for publication, January 28, 1985, and in revised form, July 16,1985)
Lars- Ake Fransson, Ingrid Silverberg, and Ingemar Carlstedt From the Department of Physiological Chemistry 2, University of Lund, S-22100 Lund, Sweden
We have treated bovine lung heparan sulfate with the polymer chain and the protein is assumed to be -(1+4)alkaline [3H]borohydrideto end label the chains with P - D - G l c p A - ( 1 ~ 3 ) - P - D - G a l p - ( 1 + 3 ) - P - ~ - G a l p - ( l ~ 4 ) - P - ~ [3H]xylitol. After subsequent periodate oxidation-alXylp-(l*2)-Ser(peptide), in common with heparin, chonkaline elimination products were separated by gel per- droitin sulfate, and dermatan sulfate asshown previously by meation and ion exchange chromatography. The link- Rod& and co-workers (5). Although the linkage between age region fragment expected to have 2 galactoses and heparan sulfate and protein has notbeen studied in detail, it 1 [‘HJxylitol residue appeared in the tetra-/trisaccha- has been known for a long time that serine is the linkage ride region after gel filtration and was bound to the amino acid (6, 7) and that serine, xylose, and galactose are anion exchange resin. A similar negatively charged present in a molar ratio of 1:1:2 (8). In the course of studies fragment, expected to have2 galactoses, 1 xylose and 1 serine, was isolated after periodate oxidation-alka- on end-labeled [ [3H]xylitol]heparan sulfate we observed that line elimination of unlabeled heparan sulfate. The neg- the putative linkage region trisaccharide, Gal-Gal-XylOH, ative charge was due to the presence of alkaline phos- was negatively charged. The demonstration of 2-phosphoxyphatase-labile phosphateester. The molar ratio of ga- lose in theproteoglycan from Swarm rat chondrosarcoma by 1actose:phosphate:xylose was2.17:1.19:1.00. The Oegema and co-workers (9) led us to investigate whether or phosphate ester was associated with the ~ylose/[~H]not the negative charge of the above trisaccharide was due to xylitol moiety as indicated by the formation of phos- the presence of phosphate ester groups. phoxylose/-xylitol by &galactosidase digestion of the EXPERIMENTAL PROCEDURES phosphorylatedtrisaccharide.Furthermore,orcinol reactivity disappeared after periodate oxidation of the Materials-Heparan sulfate was prepared from the heparin bydephosphorylated trisaccharide. The phosphate ester products of bovine lung (10). The material was first fractionated must be located to C-2 of xylose/xylitol as the l-’H according to charge density into HS1, HS2, HS3, HS4, and HS5 (11) radioactivity could be released by periodate oxidation and then from subfractions HS2, HS3, and HS4, association-prone when it was preceded by alkaline phosphatase treat- variants (HS2-A, HS3-A, and HS4-A) were prepared by gel chromament. Itis estimated that almost every chain of heparan tography (10,12). Oligosaccharides of the general structure GlcNAc/ were obtained from heparan sulfate carries 2-phosphoxylose. It would ofbeinterest GlcNSO~-(GlyA-GlcNAc/GlcNSO~)n-R to know if glycosaminoglycan chains that are artifi- sulfate after periodate oxidation-alkaline elimination (11). The following enzymes were products of Sigma: alkaline phosphatase from cially initiated onto exogeneous &D-xylosides also ac-Escherichia coli (type III), @-galactosidasefrom the same source quire the 2-phosphoxylose moiety. (grade IX), ribonuclease A from bovine pancreas (type IA), and
Proteoheparan sulfateis a common cell surface constituent of adherent mammalian cells (for review see Ref. 1).The core protein of the proteoglycan is probably intercalated into the plasma membrane via a nonglycosylated hydrophobic peptide portion (1-4), whereas the peripheral part of the molecule carries the heparan sulfate side chains (1).Although these chainscan vary considerably in composition they have a common repetitive carbohydrate backbone, ((-(1+4)-a-~/p~-GlypA-(1+4)-a-~-Gl~pN-)),.~ The linkage region between
* This work was supported by grants from the Swedish Medical Research Council (567), “Kocks Stiftelser,” “Gustaf V s 80-Hsfond,” “Osterlunds Stiftelser,” and theMedical Faculty, University of Lund. The costs of publication of this article were defrayed in part by the payment of page charges. This articlemusttherefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‘The abbreviations used are: GlyA, glycuronic acid; GlcA, Dglucuronic acid; Gal, D-galaCtOSe; IdoA, L-iduronic acid; GlcN, GlcNAc, and GlcNS03,2-amino-2-deoxy, -2-acetamido-2-deoxy-,and 2-sulfamato-2-deoxy-D-glucose, respectively; Po4,phosphateester; SO4, sulfate ester; Xyl, D-XylOSe; XylOH, xylitol; R, remnant of oxidized and degraded GlcA. For further details see Document of the IUB-IUPAC Joint Commission on Biochemical Nomenclature, 1980.
phosphodiesterase I from Crotalus atrox venom. Ultrogel AcA202 and DE52 DEAE-cellulose were from LKB and Whatman, respectively. NaB3H4(366 mCi/mg) was purchased from the Radiochemical Centre, and Insta-Gel was from Packard Instrument Co. Preparation of Heparan Sulfates with ~3HlXylitol-Solutions of heparan sulfate (50 mg/ml) in 0.5 M NaOH, 0.05 M Na(3H)BHa (36 mCi/mmol) were kept at room temperature for 24 h. Excess borohydride wasdecomposed with glacial acetic acid, and [3H]heparan sulfate was recovered after extensive dialysis, freeze drying, and gel chromatography (10). Degradatiue Methods-Periodate oxidation ofGlcA in (GlcAGlcNAc),-block regions of heparan sulfate was carried out with 2 mg of polysaccharide/ml of 0.02 M NaI04, 0.05 M sodium formate, pH 3.0, a t 4 “C for 24 h (13). Reactions were stopped by the addition of mannitol, and oxyglycans were cleavedin alkali (pH 12,30min). The following conditions were used for the enzymatic degradations (using 100 wg or less of substrate/500 pl): alkaline phosphatase, 3 units in 50 mM Tris, pH 8.0, containing 1 mMMgC12; P-galactosidase, 200 units in 50 mM Tris, pH 7.3, containing 1 mM MgClz; ribonuclease, 100 pg in phosphate-buffered saline, pH 7.0; phosphodiesterase I, 0.1 unit in 500 mM Tris, pH 8.8; all digestions were carried out at37 “C overnight. Analytical Methods-Thefollowing colorimetric methods were used orcinol for hexuronate and pentose (14), anthrone for hexose (15), and Park-Johnson (16) for reducing power. Individual neutral sugars were determined by gas-liquid chromatography as the corresponding alditol acetates, with L-arabinose as internal standard (17). Samples were hydrolyzed in 2 M trifluoroacetic acid at 100 “C for 4 h in sealed tubes underN,. Phosphate was quantified by the molybdate-
14722
2-Phosphoxylose in Heparan Sulfate malachite green method (18) after ashing inHzS04.Radioactivity was determined by liquid scintillation.
n = 4 3 2
The nature of the low molecular weight 3H-labeled material (K, 0.7-1.0 in Fig. l b ) has not been determined. These components are also seen after chromatography of periodate-oxidized butnot alkalitreated [3H]heparansulfate. 3H radioactivity is also associated with the oligosaccharide degradation products of [3H]heparan sulfateobtained after treatments withHN02 or heparan sulfate lysase (L.-A. Fransson and I. Silverberg, unpublished observations). The 3Hmay be associated with the acetamido group of GlcNAc.
0
t t t t t
RESULTS
Isolation of Linkage Region Fragments-The bovine lung heparan sulfate subfractions HS1, HS2-A, HS3-A, and HS4A were treated with alkaline-[3H]borohydride to cleave the Xyl-Ser linkage and to generate [l-3H]Xy10H-end labeled heparan sulfate chains. The reducing terminal trisaccharide sequence, Gal-Gal-[1-3H]Xy10H,should be preceded by GlcA which can be cleaved by periodate oxidation-alkaline elimination. (13).Also, the terminal [3H]XyIOH should be periodate sensitive affording 3H-labeled C1fragments. The endlabeled heparan sulfate HS4-A gave the gel filtration profile shown in Fig. la after periodate oxidation followed byscission in alkali. The 3H radioactivity emerged largely with the V,, in trisaccharide position (Kav0.65) and in two peaks that eluted behindthe position of GlcNAc-R (n = O)? The labeled trisaccharide-like component which eluted slightly more retarded than did the standard saccharide GlcN-GlyA-GlcN-R (n = 1) was not seen when the heparan sulfate had been pretreated with alkaline phosphatase (Fig. lb). Similar results were obtained with the other heparan sulfate subfractions. End-labeled heparan sulfate (with or without phosphatase treatment) that had not been oxidized with periodate was completely eluted in the V, or closely thereafter. End-labeled heparan sulfatethat had been oxidized with periodate but not cleaved in alkaliemerged largely in theVo,but approximately 10-20% of the 3H radioactivity appeared in the two most retarded positions (Kav0.8 and 1.0, respectively; results not shown). For alarge-scale preparation of the linkage region fragment 500 mg of heparan sulfate HS3-A was subjected to periodate oxidation-alkaline elimination followed by gel chromatography on Ultrogel AcA202 as shown in Fig. 2. The fractions were analyzed for hexuronate plus pentose(orcinol), for hexose (anthrone), and for reducing power. The linkage region fragment which should consist of the sequence Gal-Gal-XylSer would be anthrone- andorcinol-positive but nonreducing. The material eluting in the position indicated by a bar displayed these properties. There was very little anthrone-positive material in more retarded positions. The trisaccharidecontaining fractions in Fig. 2 were further purified by ion exchange chromatography as shown in Fig. 3a. Most of the material was anionic affording two major components, at fractions 6-11 and 14-20, respectively. The latter contained more hexose relative to hexuronate/pentose than the former one. The hexuronate-rich component is presumably a low sulfated version of GlcN-GlyA-GlcN-R (see Ref. 11). The hexose- and pentose-containing material (see bar in Fig. 3a) was pooled, freeze-dried, and analyzed for neutral sugars by gas-liquid chromatography after acid hydrolysis. Only Gal and Xyl were obtained; the molar ratio was 2.1T1.00 when the sample had been pretreated with alkaline phosphatase. In the absence of phosphatasetreatment the ratio became 6.47:l.OO. The phosphate content of the original trisaccharide fraction was 1.19 mol/mol of Xyl. As shown in Fig. 3b the trisaccharidefraction became completely uncharged upon
1
(a)
T
v
n
5
0.4
4
0.:
3
0.2
2
0.1
1
0 IC
W
a
3
$
3
2 I
0
0.5
1
Kav
FIG. 1. Gel chromatography on Ultrogel AcA 202 of periodate oxidation-alkaline eliminationproducts of heparan sulfate HS4-A containing 3H-labeledxylitol before (a)and after treatment with alkaline phosphatase (b).Alkaline [3H]borohydride treatment, periodate oxidation-alkaline elimination and digestion with alkaline phosphatase were carried out as described under “Experimental Procedures.” Between each step the samples (2 mg each) were dialyzed against the appropriate buffer or solvent and finally chromatographed on a column (1.6 X 85 cm) eluted with 0.5 M NH4HC03at a rate of 7 ml/h. In a, 2 mg of unlabeled heparan sulfate HS4-Awas co-chromatographed with the 3H-labeledsample. Fractions (2 ml) were collected,andaliquots were analyzedfor uronic acid by the orcinal method 01.670, -) and for radioactivity (3H, - - -). Periodateoxidation was performedunderconditions which selectively cleave GlcA in -(GlcA-GlcNAc), segments and in -GlcA-Gal-Gal-Xyl-of the linkage region. The oligomers obtained after subsequent alkaline scission have the general structure GlcNAc/ G~~NS~~-(I~~A/G~~A-G~CN The A ~positions / G ~ ~ NofSsacO~)~-R. charides with n = 0, 1, 2, etc. are indicated above the top panel. A nonsubstituted Xyl (or XylOH) of the linkage region should also be periodate sensitive giving rise t o the fragmentGal-Gal-(3-carbon K, of 0.65 fragment). The 3H-labeled component which eluteda with (see arrow in b ) was pooled as indicated by the bar in a and freezedried. See also Footnote 2.
treatment with alkaline phosphatase; bothanthrone-and orcinol-positive material emerged in theearly fractions. When the corresponding linkage region trisaccharide obtained from end-labeled heparan sulfate HS4-A (Fig. l a ) was subjected to ion exchange chromatography the result shown in Fig. 5a was obtained. The 3H radioactivity emerged in essentially the same position as did the nonradioactive fragment (Fig. 3a).
14724
2-Phosphoxylose in Heparan Sulfate
h
I
n=
I
I I U
3 2 1
0
1 1 1
1
0 v
co 6
c’
U
I . 0
0
m
CD
6
T
v
0 IC
CD
6
0
0.5
1 o0.05 ? L -
10
20
K,V
FIG.2. Gel chromatography on Ultrogel AcA 202 of periodate oxidation-alkaline eliminationproducts of heparan sulfate HS3-A. The sample (500mg) was treated with periodate, dialyzed, and cleaved in alkali described under “Experimental Procedures.” Chromatography was performed on a column (2.5 X 160 cm) eluted as described in Fig. 1 a t a rate of 25 ml/h. Fractions (10 ml) were collected, and aliquots were analyzed for uronic acid by the for hexose by theanthrone method orcinol method (A670, -), (Aslo, - - -1, and for reducing power by the Park-Johnson method (Asso, . . ..).The positions indicated aboue the top panel are thesame as those in Fig. 1. The fractions marked by a bur were pooled and freeze-dried.
FRACTION NUMBER
10
20
30
FRACTION NUMBER
FIG. 3 (left).Ion exchange chromatography on DEAE-cellulose of trisaccharide-like componentfromtheperiodate oxidation-alkaline elimination productsofheparan sulfate HS3-A before treatment (a)and after treatments with alkaline phosphatase (b),&galactosidase (e), and &galactosidase plus alkaline phosphatase (d). The fractions corresponding to trisaccharide-like material and indicated by a bar in Fig. 2 were pooled, freeze-dried, and chromatographed on a column (0.6 X 12 cm) of DE52 DEAE-cellulosethat was equilibrated with 0.01 M NHJICO3, pH 8.2. Elution was performed with a linear gradient of 0.01-0.5 M NHIHC03, pH 8.2 (total volume, 100 ml) a t a rate of 3 ml/h. Fractions -) and for hexose were collected and analyzed for uronic acid (&O, (&lo, - - -). Material in a was pooled as indicated by the bur and Structure of Linkuge Region Fragments-The two linkage freeze-dried.Portions (0.1 mg) weretreated with alkaline phosphatase region fragments should have the structure Gal-Gal-Xyl-Ser or B-galactosidase as described under “Experimental Procedures” and (plus phosphate) and Gal-Gal-[1-3H]Xy10H(plus phosphate) rechromatographed (b and c, respectively). The orcinol-positive main the case of the unlabeled and labeled materials, respec- terial in c was pooled (see bar), freeze-dried, digested with alkaline tively. To demonstrate that the phosphate ester was located phosphatase, and rechromatographed (d). 4 (right).Periodate oxidation of nonphosphorylated(a, to theXyl/XylOH residue the components were treated with b)FIG. and phosphorylated (c, d ) trisaccharide-like linkage-region @-galactosidaseand rechromatographed on DEAE-cellulose. components from heparan sulfate HS3-A. The trisaccharide-like In the former case (Fig. 3c) anthrone-positive material (pre- and phosphorylated linkage region component was isolated by gel sumably Gal) emerged at low ionic strength, whereas the filtration (Fig. 2) and ion exchange chromatography (Fig. 3a) after orcinol-positive (presumably Xyl-containing) material still periodate oxidation-alkaline elimination of heparan sulfate HS3-A. appeared negatively charged. When this material (see bar in A portion of the material was treated with alkaline phosphatase and rechromatographed (Fig. 3b). The nonphosphorylated component was c) was subsequently digested with alkaline phosphatase, the recovered by freeze-drying. Portions (0.5 mg) of the phosphorylated elution position was shifted to the early fractions (Fig. 3d). and nonphosphorylated linkage region fragments were dissolved in 2 To confirm that thephosphate ester was on the Xyl moiety ml of0.05 M sodium formate, pH 3.0. One ml of each sample was of the trisaccharide-like linkage-region fragment, the phos- treated with NaI04 as described under “Experimental Procedures.” phorylated (Fig. 3a) and dephosphorylated (Fig. 3b) com- Untreated (aand c) as well as oxidized (b and d ) samples were then on a column (1 X 50 cm) of Bio-Gel P-2 eluted with pounds were oxidized with periodate. A phosphorylated Xyl chromatographed 10% (v/v) ethanol a t a rate of 4 ml/h. Aliquots of the fractions were moiety would beresistant to oxidation, whereas nonphospho- analyzed for pentose by the orcinol method (-4.570). The large peaks rylated Xyl would be cleaved betweenC-2 and C-3. To deter- around fractions 25-30 in b and d represent degradation products of mine the effect, treated and untreatedsamples were chromat- mannitol which is used to terminate the periodate oxidation. The ographed on Bio-Gel P-2 (Fig. 4). The nonphosphorylated void volume of the column is a t fraction 14,and the totalvolume is Gal-Gal-Xyl-Ser lost its orcinol reactivity upon periodate at fraction 35.
oxidation (Fig. 4,a and b). The phosphorylated Gal-Gal-XylSer retained all its orcinol reactivity (Fig. 4, c and d ) but gave To show that the3 H label was in thephosphorylatedXylOH a slightly broader peak. This is probably due to cleavage of moiety of the labeled trisaccharide (isolated as shown in Fig. the C-3 to C-4 bond of the nonreducing terminal Gal residue 5a), it was digestedwith@-galactosidase and rechromatoresulting in a slightly eroded trisaccharide. graphed. As seen from Fig. 5b the 3H emerged in the same
14725
2-Phosphoxylosein Heparan Sulfate
I
I
0
In
CCI
FRACTION NUMBER FIG. 5. Ion exchange chromatography on DEAE-cellulose of 'H-labeled trisaccharide-like component from the periodate oxidation-alkaline eliminationproducts of heparan sulfate HS4-A before treatment(u)and after treatments with8galactosidase (b) and alkaline phosphatase (c). The 3H-labeled material in the fractions indicated by a bur in Fig. l a was pooled, freeze-dried, and chromatographed as described in thelegend to Fig. 3. Material in u was pooled (see bar), freeze-dried, and portions were treated with the indicated enzymes and rechromatographed. The fractions indicated by burs in b and c were pooled, freeze-dried, and subjected to gel chromatography (see Fig. 6).
0.5
1
Kav
FIG. 6. Gel chromatography on Ultrogel AeA 202 of 'Hlabeled linkage-region trisaccharidefrom the periodate oxidation-alkaline elimination products of heparan sulfate HS4A after reoxidation with periodate(a),after treatment with &galactosidase (b),after subsequent treatment with alkaline phosphatase (c), and after treatment with alkaline phosphatase followed by periodate oxidation (d).The 3H-labeledmaterial in the fractions indicated by a bur in Fig. 5u was pooled, freeze-dried, and chromatographed after periodate oxidation in a. The fractions indicated by a bur in Fig. 5b @-galactosidase treated) were pooled, freeze-dried, and a portion chromatographed in b. Another portionof the same material was digested with alkaline phosphatase and chromatographed in c. The fractions indicated by a bur in Fig. 5e were pooled, freeze-dried, oxidized with periodate, and chromatographed in d. The column was the same as that used in Fig. 1. The arrow indicates the elution position of untreated trisaccharide and @ denotes a phosphate-ester group.
trisaccharide (isolated as shown in Fig. 5a) was oxidized with periodate and chromatographed on Ultrogel AcA 202 (Fig. 6a). Since the nonreducing terminal Gal residueis eroded by periodate the elution profile showeda trailing toward material of lower molecular weight. However, most of the radioactivity was still in thepositions of the untreatedtrisaccharide (arrow position as did the orcinol-positive material in Fig. 3c. After in Fig. 6a) and of the 6-galactosidase-treated trisaccharide alkaline phosphatase treatment of the labeled trisaccharide (Fig. 6b). After digestion of the labeled trisaccharide with both /3-galactosidase and alkaline phosphatase, the product, all of the 3H emergedwith the uncharged material (Fig. 5c). The position of the phosphate ester on the XylOH moiety [3H]Xy10H,emerged at K,,approximately0.9 (Fig. 6c). When was determined by the experiments shown in Fig. 6. The the original trisaccharide (Fig. 512) was dephosphorylated, XylOH of the trisaccharide Gal-Gal-[1-3H]Xy10Hhas only recovered (Fig.5c), andsubsequently oxidized with periodate, two available OH groups for PO, substitution (Fig. 7). Sub- 3H-labeled fragments smaller in size than [3H]Xy10Hwere stitution at C-2 renders the XylOH moiety resistant to per- obtained (Fig. 64.The phosphate, thus linked to C-2 of Xyl, is probably a monophosphate ester as it was also sensitive to iodate oxidation whereas substitution at C-3 wouldallow scission of the C-1 to C-2 bond. Therefore, the 3H-labeled ribonuclease and phosphodiesterase I (results not shown).
2-Phosphoxylose in Heparan Sulfate
14726 COOH (a3
CH20H
00
0
0
CH2OH O
I
G
O
-
C
H
I
OH
OH
2
-
C
I
0 I HO - P
I H I
- OH
II
0
(b3
“0 0 CH20H
CH20H
0
OH
0e
OH
2
0
H
0 I
HO - P
OH
II
0
FIG. 7. Proposed structures for linkage region trisaccharides. a, the proposed linkage region sequence, -Gal-Gal-Xyl(2-P04)Ser; b, the trisaccharide Gal-Gal-XylOH-2-POr generated by periodate oxidation-alkaline elimination of heparan sulfate with 3H-labeled xylitol residues.
25). It would be ofinterest to know if Xyl-initiated chondroitin sulfate chainsacquire a 2-phosphoxylose substructure. As many proteoglycans could be synthesized simultaneously by a given cell the need for a “sorting mechanism” is obvious. There arelarge, hyaluronate- or nonhyaluronate-binding proteochondroitin sulfates, small ubiquitous proteochondroitin sulfates (26), large and small proteodermatan sulfates (27), and, possibly, several types of proteoheparan sulfates (1).All have the same linkage region, i.e. the primers for elongation are indistinguishable. The phosphoxylose residue could be a “recognition signal” for primers that are going $0 acquire complex side chains. Alternatively or in addition the phosphoxylose moiety of a proteoheparan sulfate may serve as a “recognition signal” for interaction with other molecules or structures in the extracellular environment. In such a case, deshielding by removal of the side chains by carbohydratedegrading enzymes may be a necessary prerequisite. Acknowledgments-We thank Birgitta Havsmark for technical assistance, Birgitta Jonsson for artwork, Roger Sundler for the phosphate analyses, and Rose-Marie Akselsson and Inger Carlsson for secretarial aid. REFERENCES
1. Carlstedt, I., Coster, L., Malmstrom, A., and Fransson, L.-A. (1983) J. Bwl. Chem. 258, 11629-11635 DISCUSSION 2. KjellBn, L., Pettersson, I., and Hook, M. (1981) Proc. Natl. Acud. sci. U. S. A. 78,5371-5375 The datapresented demonstrate that heparan sulfate from 3. Norling, B., Glimelius, B., and Wasteson, A. (1981) Biochem. bovine lung contains a monophosphate ester on C-2 of the Biophys. Res. Commun. 103, 1265-1272 xylose moiety of the linkage region (Fig. 7). The results further 4. Raproeger, A.C., and Bernfield, M. (1983) J. Bwl. Chem. 258, suggest that most, if not all, of the heparan sulfate chains 3632-3636 carry 2-phosphoxylose. This conclusion is based on the obser5. Fransson, L.-A. (1984) in T h e Polysaccharides (Aspinall, G. O., vation (see Fig. 2) that very little material corresponding to ed) Vol. 111, pp. 337415, Academic Press, New York 6. Muir, H. (1958) Biochern. J. 69, 195-204 Gal-Gal-(3-carbon remnant) was seen when periodate-oxi7. Jacobs, S., and Muir, H. (1963) Biochem. J. 87,38P dized and cleaved heparan sulfate was chromatographed on 8. Knecht, J., Cifonelli, J. A., and Dorfman, A. (1967)J. Bwl. Chem. Ultrogel AcA 202. A nonphosphorylated Xyl moiety should 242,4652-4661 give rise to the3-carbon remnant provided that theXyl moiety 9. Oegema, T. R., Jr., Kraft, E.L., Jourdian, G. W., and Van Valen, was cleaved by periodate. However, further purification of the T. R. (1984)-J. Biol. Chem. 259,1720-1726 trisaccharide pool in Fig. 2 by ion exchange chromatography 10. Fransson, L.-A. Havsmark, B., and Sheehan, J. K. (1981) J. Biol. Chem. 256,013039-13043 (Fig. 3a) indicated that uncharged (nonphosphorylated) trisaccharide was largely absent. The 2-phosphoxylose has re- 11. Fransson, L.-A., Sjoberg, I., and Havsmark, B. (1980) Eur. J. Biochem. 106, 59-69 cently been demonstrated in the proteoglycan of the Swarm 12. Fransson, L.-A., Nieduszynski, I. A., and Sheehan, J. K. (1980) rat chondrosarcoma (9). This proteoglycan, which also conBiochim. Bwphys. Acta630,287-300 tains phosphorylated hydroxyamino acids, has 2-phosphoxy- 13. Fransson, L.-A., Malmstrom, A., Sjoberg, I., and Huckerby, T. N. (1980) Carbohydr. Res. 80,131-145 lose on one out of 3-4 chondroitin sulfate chains. A linkage region fragment of the proposed structure p-~-Glyp4enA-(l+ 14. Brown, A. H. (1946) Arch. Biochem. Biophys. 11,269-275 3)-/3-~-GalpNAc-(l -+ 4)-/3-~-GlcpA(l+ 3)-P-~-Galp-(l+ 15. Goa, J. (1955) Scand. J. Clin. Lab. Inuest. 7, Suppl. 22,19-25 16. Park, J. T., and Johnson, M. L. (1949) J. B i d . Chem. 181, 1493)-P-~-Galp-(14)-XylOH-(2PO~) was isolated after chon156 droitin AC-I1 lyase degradation of chains released by alkaline 17. Lindahl, U. (1970) Biochem. J. 116,27-34 borohydride treatment of the proteoglycan. 18. Buss, J. E., and Stull, J. T. (1983) Methods Enzymol. 99, 7-14 The phosphoxylose residue in proteochondroitin sulfate 19. Campbell, P., Jacobsson, I., Benzing-Purdie, L., Rodkn, L., and Fessler, J. H. (1984) Anal. Bwchem. 137,505-516 and proteoheparan sulfate could either be introduced in conH.-P., Schwartz, N. B., Rod& L., and Prockop, 0.J. junction with xylosylation of the nascent core protein (see 20. Hoffmann, (1984) Connect. Tissue Res. 12,151-163 Ref. 19) or during/after secretion of the completed proteogly- 21. Rod&, L. (1980) in The Biochemistry of Glycoproteins and Procan. Although thisremains to be determined the former teoglycans (Lennarz, W . J., ed) pp. 267-355, PlenumPress, New York alternative appears more reasonable. The Xyl moiety of a complete proteoglycan should be very inaccessible to a phos- 22. Schwartz, N. B., Galligani, L., Ho, P.-L., and Dorfman, A. (1974) Proc. Natl. Acad. Sci. U. S. A. 71,4047-4051 phorylating enzyme. During biosynthesis of a proteoglycan 23. Galligani, L., Hopwood, J., Schwartz, N. B., and Dorfman, A. the xylosylated precursor protein appearsto be formed in the (1975) J. Biol. Chern. 250,5400-5406 rough endoplasmic reticulum (20) whereas “multienzyme 24. Robinson, H. C., and Lindahl,, U. (1981) Biochem. J. 194, 575complexes” for chain elongation, modification, and sulfation 586 are located in the Golgi membranes (21). The elongation and 25. Spooncer, E., Gallagher, J. T., Krizsa, F., and Dexter, T. M. (1983) J. Cell Bwl. 96, 510-514 sulfation can be artificially initiated onto exogeneous p-DPaulson, M. (1984) in Extracellular Matrix xylosides to form single-chain glycosaminoglycans in consid- 26.Heinegiurd,D., and Biochemistry (Piez, K. A., and Hari Reddi, A., eds) pp. 277erable amounts (21, 22). In most cases, initiation of chon328, Elsevier, New York droitin sulfate synthesisis easily achieved, whereas synthesis 27. Carlstedt, I., Coster, L., and Malmstrom, A. (1981) Biochem. J. 197,217-225 of heparan sulfateand heparin is hardly stimulated at all (22-
