CHONDROITIN SULFATE MODIFICATIONS

DISSERTATION

Presented in Partial Fulfillment of the Requirements

for the Degree Doctor of Philosophy in the

Graduate School of the Ohio State

University

By

BIENVENIDO OCHOA JULIANO, B. S., M. Sc.

The Ohio State University

1959

Approved by;

Ju.

Adviser Department of Chemistry ACKNOWLEDGMENTS

The candidate expresses appreciation to Professor M. L. Wolfrom for his advice, encouragement and interest in this work. The research group has been very accomodating and its pleasant company provided stimulating discussions and criticisms which were invaluable. This investigation was made possible through a National Science Foundation Predoctoral Fellowship from November, 1957, to June, 1958, under Grant NSF-GVf9^ to The Ohio State University, and through a C. F. Kettering Research Foundation Fellowship from July, 1958, to August, 1959* TABLE OF CONTENTS

Page 1. INTRODUCTION AND STATEMENT OF PROBLEM 1

II. HISTORICAL Chemistry of Chondroitin Sulfates Early developments 2 Methods of isolation and purification 5 Isomeric chondroitin sulfates, distribution and properties 7 Nature of chondrosine l^f Sulfate ester group attachment 18 Hexuronidic linkage 21 Galactosaminidic linkage 21 Chondroitin Zb Keratosulfate Zb If-Deacetylation and Sulfation of Polysaccharides N-Deacetylation of mucopolysaccharides 25 Sulfated polysaccharides 27

III. DISCUSSION OF RESULTS Purification and Characterization of Chondroitin Sulfate A Sodium chondroitin sulfate A 33 Paper chromatographic analysis 35 Calcium chondroitin sulfate A 36 Keratosulfate isolation and desulfation 36 Barium chondroitin sulfate A 38 Potassium chondroitin sulfate A 39 Carboxyl-reduced Chondroitin______Polymeric reduced chondroitin 39 Carboxyl-reduced chondrosine bZ Degraded, reduced chondroitin 50

iii xv Page N-Deacetylation Studies N-Deacetylation of Barium Chondroitin Sulfate A Alkaline reagents 52 Acidic reagents 56 Hydrolytic studies on N-deacetylated chondroitin sulfate A 59 Hydrazinolysis of Chondroitin Sulfate Modifications 6l Sulfated Chondroitin Sulfate Modifications 63 IV. EXPERIMENTAL

i Purification and Characterization of Chondroitin Sulfate A Sodium chondroitin sulfate A 66 Paper chromatographic analysis 67 Calcium chondroitin sulfate A 68 Keratosulfate isolation and desulfation 69 Barium chondroitin sulfate A 70 Potassium chondroitin sulfate A 72 Carboxyl-reduced Chondroitin 90% Reduced Polymeric Chondroitin 72 96% Reduced Polymeric Chondroitin 75 Carboxyl-reduced Chondrosine Fractionation of hydrolyzate 79 P-D-Glucose pentaacetate 80 N-Acetyl-oc-D-galactosamine monohydrate 8l p-D-Galactosamine pentaacetate 8l Carboxyl-reduced N-acetylchondrosine dihydrate 82 Carboxyl-reduced N-acetylchondrosinol 83 83% Reduced, Degraded Chondroitin 85 N-Deacetylation Studies N-Deacetylation of Barium Chondroitin Sulfate A Hydrazinolysis 89 V Page Hydrolytic studies on N-deacetylated chondroitin sulfate A 90 Deacetylation with 5 N sodium hydroxide 91 Deacetylation with 5 N hydrochloric acid 91 Deacetylation with methanolic hydrogen chloride 91 Deacetylation with benzyl alcohol- hydrogen chloride 93 Hydrazinolysis of Chondroitin Sulfate Modifications N-Deacetylated carbonyl-reduced sodium chondroitin sulfate A 9k IJ-Deace tylated sodium chondroitin sulfate A 9^ N-Deacetylated Sodium chondroitin 9*+ N-Deacetylated carboxyl-reduced chondroitin 95 Sulfated Chondroitin Sulfate Modifications Sulfated Jtf-deacetylated chondroitin sulfate, sodium salt 95 Sulfated N-deacetylated carbonyl-reduced chondroitin sulfate, sodium salt 98 Sulfated chondroitin sulfate, sodium salt 98 Sulfated N-deacetylated chondroitin, sodium salt 99 Sulfated N-deacetylated carboxyl-reduced chondroitin, sodium salt 99 V. SUGGESTIONS FOR FURTHER RESEARCH 108 VI. SUMMARY AND CONCLUSIONS 109 CHRONOLOGICAL BIBLIOGRAPHY 111 AUTOBIOGRAPHY 120 LIST OF TABLES

Comparison of Properties of Chondroitin Sulfates A, B and C Deacetylation of Sodium Chondroitin Sulfate A Characterization of Silicate-purified Sodium Chondroitin Sulfate A rf-Deacetylation of Barium Chondroitin Sulfate A Hydrazinolysis of Chondroitin Sulfate A and Its Modifications

S t Values of Component Sugars in Three Solvent glucose ° Systems LIST OF FIGURES

Figure Page 1. Reaction of 5 W Hydrochloric Acid with Polymeric Chondroitin Sulfate A at 37-^0° 57 2. Infrared Spectrum of Sodium Chondroitin Sulfate A 100 3. Infrared Spectrum of Crude Keratosulfate 100 h. Infrared Spectrum of Partially Desulfated Crude Keratosulfate 101 5* Infrared Spectrum of Chondroitin Acetate 101 6. Infrared Spectrum of Chondroitin Methyl Ester 102 7. Infrared Spectrum of Carboxyl-reduced Chondroitin 102 8. Infrared Spectrum of Carboxyl-reduced N- Acetylchondrosine Dihydrate 103 9. Infrared Spectrum of N-Acetyl-D-galactosamine Monohydrate 103 10. Infrared Spectrum of P-D-Galactosamine Pentaacetate 104 11. Infrared Spectrum of p-D-Glucose Pentaacetate 104 12. Infrared Spectrum of Sodium Chondroitin 105 13. Infrared Spectrum of Partially N-Deacetylated Sodium Chondroitin 105 1^+. Infrared Spectrum of Partially N-Deacetylated Chondroitin Sulfate A, Sodium Salt 106 15* Infrared Spectrum of Sulfated Partially N-Deacetylated Chondroitin Sulfate, Sodium Salt 106 16. Infrared Spectrum of Sulfated Chondroitin Sulfate, Sodium Salt 107 17. Infrared Spectrum of Heparin, Sodium Salt 107

vii I. INTRODUCTION AND STATEMENT OF PROBLEM

Chondroitin sulfate A, an acidic mucopolysaccharide readily available from cartilage, is closely related structurally to heparin, a known blood anticoagulant. An important difference is that the amino group of their anhydrohexosamine units is acetylated in chondroitin sulfate and sulfated in heparin. In attempts in this laboratory to prepare polymeric sulfated N-deacetylated chondroitin sulfate, the serious drawbacks had been the absence of a suitable jN-deacetylating reagent and the sensitivity of chondroitin sulfate A toward alkaline degradation, alkaline reagents having been used for deacetylation. This work concerns itself with the preparation and characterization of chondroitin sulfate modifications and the search for a fairly mild and efficient N-deacetylation technique.

1 II. HISTORICAL Chemistry of Chondroitin Sulfates Early developments. Although Krukenburg (1) was not the first

(1) C. F. W. Krukenburg, Z. Biol., 20, 30? (1884). to isolate chondroitin sulfuric acid from cartilage, his preparation was of relatively pure form, whereas previous workers (2) obtained

(2) G. Fischer and C. Boedeker, Ann., 117, 111 (l86l). mixtures. Among the nineteenth century researchers, Schmiedeberg (3)

(3) 0. Schmiedeberg, Arch, exptl. Pathol. Pharmakol., 2£3, 335 (1891). contributed most toward the elucidation of the structure of chondroitin sulfate. He was the first to isolate the disaccharide, chondrosine, on acid hydrolysis of the polymer. To him was also given the credit of first noting the presence of sulfate and acetic acid in the hydrolyzate. He claimed to have obtained, by the action of dilute hydrochloric acid on chondroitin sulfate, the desulfated polymer which he termed chondroitin. However, his proposed formulas for these compounds were later proven to be inadequate (*0.

(*0 P. A. Levene, ’'Hexosamines and Mucoproteins," Longmans, Green and Co., London, 1925*

The preparation by Hebting (5) of a crystalline derivative

(5) J* Hebting, Biochem. Z., 63, 353 (191*0 • of chondrosine was of great importance since hitherto all 3 preparations and all degradation products characterized were amorphous and thus conclusions deduced from these materials were not very reliable. Hebting obtained the crystalline ethyl ester hydrochloride of chondrosine by treating oxalic acid-hydrolyzed chondroitin sulfate with 3% ethanolic hydrogen chloride. Oxidative hydrolysis of chondroitin sulfate from cartilage with acidic bromine gave D-glucaric acid as a principal product (6).

(6) M. L. SiVolfrom and W. B. Neely, J. Am. Chera. Soc., 7 31 2778 (1953); P* A. Levene and W. A. Jacobs, J. Exptl. Med., 10, 537 (1908).

This acid was identified as the potassium acid salt by Wolfrom and Neely (6) and as the silver salt, by analysis, by Levene and Jacobs (6). The glycaric acid thus obtained may be accounted for by the oxidation of the anomeric group of either D-glucuronic or L-guluronic acid. The isolation of the first crystalline derivative of the uronic acid from chondroitin sulfate was made by Stacey and coworkers (7). They isolated the crystalline

(7) H. G. Bray, J. E. Gregory and M. Stacey, Biochem. J., 38, 142 (1944). methyl glycoside of 2,3»4-tri-0-methyl-a-D-glucuronamide from the methanolysis of the methylated, degraded polysaccharide. This ether was identical to that derived from authentic D-glucuronic acid, thus proving that the uronic acid of chondroitin sulfate from cartilage is D-glucuronic acid. Levene and LaForge (8) isolated the hexosamine component of

(8) P. A. Levene and F. B. LaForge, J. Biol. Chem., 18, 123 (1914). k chondroitin sulfate, which they termed chondrosamine. They showed that it was different from D-glucosamine (2-amino-2- deoxy-D-glucose or chitosamine). Preparation was effected by hydrolysis of the barium salt of chondroitin sulfate in refluxing hydrochloric acid in the presence of stannous chloride. This amino sugar was observed to give the same phenylosazone as D-galactose, signifying that the amino group was attached at C2, and that chondrosamine was either of the epimeric hexosamines, D-galactosamine (2-amino-2-deoxy-D-galactose) or D-talosamine (2-amino-2-deoxy-D-talose) (9)- These findings were confirmed

(9) P. A. Levene, J. Biol. Chem., 26, 1^3 (1916).

by the synthesis of these two hexosamines from D-lyxose by a Strecker type reaction (10). The addition of ammonia and

(10) P. A. Levene, J. Biol. Chem., 31* 609 (1917)* hydrogen cyanide to the pentose yielded two l.yxo-hexosaminic acids, whose lactones, on reduction, gave the two epimeric amino sugars. The absolute identity of chondrosamine was solved unequivocably by the Birmingham group of Stacey and coworkers (11), who synthesized

(11) Sybil P. James, F. Smith, M. Stacey and L. F. Wiggins, Nature, 156, 308 (19^5)* J* Chem. Soc., 625 (19^6).

D-galactosamine identical in all respects to the naturally occurring chondrosamine. The synthesis involved the epoxide-opening action of ammonia on 1,6:2,3-dianhydro-P-D-talose yielding, on acidification, the hydrochlorides of both 3-amino-3-deoxy-l,6-anhydro-P-D-idose and 2-amino-2-deoxy-l,6-anhydro-P-D-galactose. The anhydro-D-galactosamine, on hydrolysis with strong acid, gave chondrosamine. These early investigations have shown that chondroitin sulfate from cartilage consisted of N-acetyl-D-galactosamine, D-glucuronic acid and sulfate ester group in nearly equivalent amounts. The last decade elucidated most of the remaining details of the structure and the chemistry of this heteropolysaccharide and these investigations are discussed below. Methods of isolation and purification. The present means of isolation of chondroitin sulfate from cartilage are not much different from that used by Krukenberg (1), who extracted rib cartilage with cold 5-10% sodium hydroxide for a period of two to three days, neutralized the extract with hydrochloric acid, concentrated the neutral solution at 80° and precipitated the material with alcohol, A modification of the procedure included precipitation with glacial acetic acid. Various techniques have been proposed and used for the isolation and purification of chondroitin sulfates. They occur naturally in protein combination, probably as salts. Dilute alkaline solutions have been commonly employed for the extraction, to bring about separation from protein aggregates. Although high yields of products have been extracted by this means, some degradation must certainly occur in the alkaline media, mainly changes in macromolecular properties (12) and partial desulfation.

(12) R. L. Whistler and J. N. BeMiller, Advances in Carbohydrate Chem., 13, 289 (1958).

To minimize degradation, neutral extractants have been tried. Protein-free and relatively undegraded fractions have been obtained through the use of 10% calcium chloride (13) and 30% potassium

(13) K. Meyer and Elizabeth M. Smyth, J. Biol. Chem., 119, 507 (1937). chloride (14), while Partridge (15) employed 30% formamide for the 6

(1*0 Julia Einbinder and M. Schubert, J. Biol. Chem., 185, 725 (1950). (19) S. M. Partridge, Biochem. J. , 4£, 38? (1948).

same purpose, Karl Meyer and coworkers (16) isolated chondroitin

(16) K. Meyer, A. Linker, E. A. Davidson and B. Weissman, J. Biol. Chem., 209, 611 (1953). sulfate from ground bovine cornea by digesting the protein aggregates with the proteolytic enzymes, pepsin and trypsin, followed by a tedious alcohol precipitation technique from calcium acetate and sodium acetate buffers and treatment with Lloyd's reagent (hydrated aluminum silicate). Although neutral extractants avoided degradation, only a fraction of the total chondroitin sulfate in the tissue was extractable. In this laboratory, silicate adsorption has been used exclusively to purify commercial sodium chondroitin sulfate (17).

(17) M. L. Wolfrom and K. Onodera, J. Am. Chem. Soc. , 79, 4737 (1957).

The complexing ability of acidic polysaccharides has been utilized in the purification of chondroitin sulfate preparations. Mathews and Dorfman (18) made principal use of precipitation of the

(18) M. B. Mathews and A. Dorfman, Arch. Biochem. Biophys., 42, 41 (1953). luteocobaltic complex of chondroitin sulfate from 0.1 M acetic acid at pH 3 .6-3 .7* Various quaternary ammonium salts, such as Cetavlon (cetyltrimethylammonium bromide) and cetylpyridinium halides, have also been employed for this purpose (19).

(19) B. C. Bera, A. B. Foster and M. Stacey, J. Chem. Soc., 3788 (1955); J- E. Scott, Biochem. J., 62, 3l£ (1956).

Isomeric chondroitin sulfates, distribution and properties. Three isomeric chondroitin sulfates, A, B and C, have been characterized, after the classification of Meyer and coworkers (20).

(20) K. Meyer, E. Davidson, A. Linker and P. Hoffman, Biochim. et Biophys. Acta, 21, 506 (1956); P. Hoffman, A. Linker and K. Meyer, Federation Proc., 17» IO78 (1958).

They are widely distributed mucopolysaccharides of connective tissues. Chondroitin sulfate A is the major component of cartilage mucopolysaccharides and has also been isolated from cornea and from human and bovine aorta. Chondroitin sulfate B is the major polysaccharide of adult skin and is widely distributed in most connective tissues, except in cartilage and in cornea. The chondroitin sulfate extract from cartilage contain a very small amount of chondroitin sulfate C. Chondroitin sulfate C is also present in other connective tissues, admixed with other mucopolysaccharides. The chondroitin sulfates have been differentiated by optical rotation, anticoagulant activity, relative ease of enzymic and- acid hydrolysis, color yield by the carbazole reaction, solubility of salts and infrared spectral analysis. Karl Meyer and coworkers (20) summarized their properties and a modified resume is provided in Table 1. Although the elementary analyses of the isomers were identical, chondroitin sulfate A and C have no anticoagulant activity, whereas the B polysaccharide (P-heparin) has 30% of the activity of heparin (21). Chondroitin sulfate B 8

Table 1

Comparison of Properties of Chondroitin Sulfates A, B and C (20)

Chondroitin Sulfate

A B C

ri20-25° , . , , ,„o \aj ^ (water) -28 to -32 -55 to -63° -16 to -22°

Ethanol fraction precipitating the calcium salt, % 30 to 40 18 to 25 40 to 50

Hydrolysis by testicular enzyme (hyaluronidase) + +

Anticoagulant activity - + -

Electrophoretic mobility ca. -13 ca. -13 ca. -13

Hexosamine D-galactos- D-galactos- D-galactos­ amine amihe .. amine

Uronic acid (molar ratio to hexosamine) Carbon dioxide assay equimolar equimolar equimolar Carbaaole assay equimolar half mole equimolar per mole Identity D-glucuronic L-iduronic D-glucuronic 9

(21) R. Marbet and A. Winterstein, Helv. Chim. Acta, 3^£* 2311 (1951); Experientia, 8, kl (1952).

has not been isolated from cartilage and differs markedly in properties from A and C in being more acid-labile. All gave equivalent amounts of hexosamine, hexuronic acid, sulfate and acetic acid. Although their hexosamine residue has been characterised as D-galactosamine, colorimetric assays for hexuronic acid, the carbazole (22) and the orcinol (23) tests,

(22) 2. Dische, J. Biol. Chem., 167, 189 (19^7). (23) J. X. Khym and D. G. Doherty, J. Am. Chem. Soc., 7^. 3199 (1952).

of chondroitin sulfate B were inconsistent with the decarboxylation assay (2k), This led K. Meyer and coworkers to

(2k) M. V. Tracey, Biochem. J., ^3, 185 (19^S). deduce that chondroitin sulfate B has a different hexuronic acid residue from D-glucuronic acid, present in chondroitin sulfates A and C . In 1956, Hoffman, Linker and Meyer (25) tentatively proposed

(25) P. Hoffman, A. Linker and K. Meyer, Science, 12^-, 1252 (1956). iduronic acid as the hexuronic acid moiety of the B polysaccharide. From the acid hydrolysate of chondroitin sulfate B, they obtained two disaccharides, the major disaccharide of which yielded a similar carbazole-to-orcinol ratio (0.2) to that obtained for iduronic acid and on hydrolysis, produced a spot on a paper 10 chromatogram with value identical to that of iduronic acid. It was similarly shown that the second disaccharide contained glucuronic acid as its uronic acid moiety. Both disaccharides contained only galactosamine as the demonstrable hexosamine. The isolated uronic acid and the synthetic iduronic acid (26) had

(26) F. Shafizadeh and M. L. Wolfrom, J. Am. Chem. Soc., 77, 2568 (1955). identical R^. values for both acid and lactone in four different solvent systems. 3-Heparin, first isolated from beef lung by Marbet and Winterstein (21), has been shown to be identical to chondroitin sulfate B, first isolated from hog skin (27). This was

(27) K. Meyer and Eleonor Chaffee, J. Biol. Chem,, 138, 491 (1941). demonstrated by Cifonelli, Ludowieg and Dorfman (28) and by Jeanloz

(28) J. A. Cifonelli, J. Ludowieg and A. Dorfman, Federation Proc., 16, 165 (1957); J. Biol. Chem., 233, 541 (195S). and Stoffyn (29), These workers simultaneously proved that the

(29) R. W. Jeanloz and P. J. Stoffyn, Federation Proc., 17, 249 (1958). iduronic acid of P-heparin is the L-form. Jeanloz and Stoffyn hydrolyzed P-heparin, previously desulfated by the method of Kantor and Schubert (30), and further reduced with sodium borohydride (31-33),

(30) T. G. Kantor and M. Schubert, J. Am. Chem. Soc., 79, 152 (1957). 11 (31) M. L. Wolfrom and H. B. Wood, J. Am. Chem. Soc., 73, 2933 (1951). (32) M. L. Wolfrom and Kimiko Anno, J. Am. Chem. Soc., 7^, 5583 (1952). (33) Harriet L. Frush and H. S. Isbell, J. Am. Chem. Soc., 78, 28*f*f (1956).

to obtain L-idose, subsequently transformed to L-idosan (1,6-anhydro-L-idopyranose). This compound was characterized as its crystalline 2,3,*<— tri-O-acetyl derivative, identical with a sample prepared by synthesis from 1,2-isopropylidene-L-idofuranose (3*0. Dorfman and coworkers isolated the amorphous hexuronic acid

(3*0 I* Vargha, Chem. Ber. , 87, 1351 (195*0*

lactone from P-heparin,[a^ +30° (water), which behaved chromatographically like the authentic L-iduronic acid (26). L- Iduronic acid has +33° (water, final) (26). They reduced the methyl ester of this acid to a mixture of L-idose and L-iditol with sodium borohydride. The products were chromatographically identical with authentic samples. Chondroitin sulfate B (p-heparin) is the first acidic mucopolysaccharide reported to contain a uronic acid other than D-glucuronic acid and the work represents the first isolation of iduronic acid from a biological source. Most structural investigations have been confined to chondroitin sulfate (mostly A) from cartilage. Periodate oxidation data supports a linear polymer, with the consumption of one mole of oxidant per anhydrodisaccharide unit with the destruction of the hexuronic acid moiety. Methylation procedures after the method of Haworth (35) have been employed on highly degraded chondroitin

(35) W. N. Haworth, J. Chem. Soc., 107, 8 (1915).

sulfate in attempts to determine the linkages involved— K. H. Meyer and coworkers (36) proposed a linear polymer exclusively 12

(36) K. H. Meyer and M. E. Odier, Experientia, 2, 311 (19^6); K. H. Meyer, M. E. Odier and A. E. Siegrist, Helv. Chim. Acta, 31, 1400 (19^8).

(l-*3)~linked, whereas Stacey and coworkers (?) envisaged a repeating trisaccharide unit of the type — D-glucuronic acid------rJ-acetyl-D-galactosamine— D-glucuronic acid denoting a highly branched structure, not substantiated by established analytical results of equimolar amount of hexosamine and hexuronic acid. These structures were inconclusive since they were based on highly degraded material and because of the incomplete methylation achieved by these workers. The light-scattering studies of Mathews (37) showed that the

(37) M. B. Mathews, Arch. Biochem. Biophys., _6l, 367 (1956).

sodium chondroitin sulfate derived from bovine hyaline cartilage has a molecular weight of 50,000. Chondroitin sulfates have been found to be linear polymers with similar chain configuration- charge relationships. Whereas Mathews and Dorfman (18) proposed a spherical structure for the molecule, Bernard! (38 )

(38) G. Bernard!, Compt. rend,, BMt, 1918 (1957)*

demonstrated that a rod-like structure was equally plausible from the experimental data. Shatton and Schubert (39) were the first to isolate

(39) Jennie Shatton and M. Schubert, J. Biol. Chem., 211, 365 (195*0. 13 chondromucoprotein, consisting of 30% protein, 60% potassium chondroitin sulfate and 10% water, by aqueous extraction from slices of fresh bovine nasal cartilage. This material behaved as a compound of protein and polysaccharide when precipitated as the potassium, barium or luteocobaltic salt. Free protein was not detected by protein precipitants, nor was free chondroitin sulfate detected by ultrafiltration in the first products extracted. Later studies supported these observations, which were also supported by Partridge and Davis (40). It has been

(40) S. M. Partridge and H. F. Davis, "Chemistry and Biology of Mucopolysaccharides," G. E. W. Wolstenholme and Maeve O'Connor, Ed., Little, Brown and Co., Boston, Mass., 1958, p. 93* proposed that the complex consisted of polysaccharide chains of variable but moderate length, cemented together by polypeptide units attached to carbohydrate by a bond as sensitive to alkali as an ester or lactone (4l). Mathews and Lozaityte (42), and also

(41) Helen Muir, Biochem. J, , 69., 195 (1958). (42) M. B. Mathews and Irene Lozaityte, Arch. Biochem. Biophys., 74, 158 (1958).

Partridge and Davis (40), reported the protein component to be markedly different in amino acid composition from that of collagen. Their light-scattering and viscosity data supported the existence of a rod-like basic unit of molecular weight 4.0 x 10 and 3,700 A long, with the sodium chondroitin sulfate molecules in the complex of molecular weight 5 0 ,000. A new and more efficient extraction method was described for chondromucoprotein by Malawista and Schubert (43) in 1958,

(43) Ina Malawista and M. Schubert, J. Biol. Chem., 230, 535 (1958). 14 through the use of a high-speed (45,000 r.p.m,) homogenizer at low temperature, a 75% yield being obtained from cartilage. Alkaline hydrolysis of chondromucoprotein so prepared was recommended as a convenient source of undegraded chondroitin sulfate. The electrophoretic behavior of the chondromucoprotein has been extensively studied by Warner and Schubert (44),

(44) B. C. Warner and M. Schubert, J. Am. Chem. Soc., 80, 5166 (1958).

Chondromucoprotein and chondroitin sulfate migrated essentially as single components over pH 5 to 11 with mobility constants of 13*6 —5 2 —1 —1 and 14.4 x 10 cm. volt sec. , respectively, a difference of only about 6%. They showed that the ionic groups of chondroitin sulfate are free in the complex and thus are not involved in bonding with the protein. Chondromucoprotein appeared completely dissociated at pH 12.5 into two components, one of which had the same mobility as chondroitin sulfate, while the other was slower. As the pH fell from 5 to 2, the drop of mobility of both substances was only 30% (at pH 2, the authors considered the carboxylic acid groups bound and only the sulfate groups ionic), indicating that at pH 7 the carboxylic acid groups of the hexuronic acid moiety are not completely ionized, but might be extensively associated in solution. Mature of chondrosine. Since the preparation of the crystalline ethyl ester of chondrosine by Hebting (5)» extensive investigations have concentrated on chondrosine in attempts to elucidate the structure of chondroitin sulfate. Levene (45)

(45) P. A. Levene, J. Biol. Chem., 140, 267 (1941). 15 prepared a number of crystalline derivatives of this disaccharide but did not succeed in establishing its structure. He prepared the crystalline hexaacetate and hexamethyl ether of N-acetylchondrosine methyl ester, and the crystalline heptaacetate and heptamethyl ether of N-acetylchondrosinol methyl ester. Catalytic reduction of hepta-()-methyl-N-acetylchondrosinol methyl ester effected the reduction of the ester group, giving a crystalline product. In this laboratory, Levene's (45) work was reevaluated by Wolfrom, Madison and Cron (46). They also prepared the crystalline

(46) M. L. Wolfrom, R. K. Madison and M. J. Cron, J. Am. Chem. Soc., 74, 1491 (1952). hexa-O-benzoyl-N-benzoylchondrosine methyl ester, Hepta-()- acetyl-N-acetylchondrosinol methyl ester was converted, by the action of ethanolic ammonia, to the diamide glycitol which on periodate oxidation underwent formaldehyde and formic acid scission in the reduced portion to yield a crystalline oxidation product. Further oxidation of this product showed the consumption of one more mole of oxidant. This periodate oxidation data is definitive for only one structure for either sugar sequence for the disaccharide. Influenced by the results of Levene and LaForge (47), Wolfrom and coworkers (46) proposed the structure of

(47) P. A. Levene and F. B. LaForge, J. Biol. Chem., 15, 69 (1913). chondrosine as 2-amino-2-deoxy-0-(3 (? )-D-galactopyranosyl- (l-Vf) -D-glucuronic acid. 16 For the diamide glycitol corresponding to this structure, the periodate oxidation data is consistent with the immediate oxidation of the glycitol group, consuming two moles of oxidant and liberating formic acid and formaldehyde, and the slow oxidation step involving the scission of the 5,4-cis-diol group of the N-acetylgalactosaminide portion. However, the sequence chosen by Wolfrom and coworkers (46) was shown to be otherwise. This sequence was inconsistent with the observation of Japanese workers (48) that their disaccharide

(48) H. Masamune, Z. Yosizawa and M. Maki, Tohoku J. Exptl. Med., 5 5 , 47 (1951). preparation gave a positive Elson-Morgan (49) test for free amino

(49) I. A. Elson and W. T. J. Morgan, Biochem. J., 27, 1824 (1933). sugar, implying that the hexosamine, and not the uronic acid, was the reducing end. Davidson and K. Meyer (50) confirmed this

(50) E. A. Davidson and K. Meyer, J. Am. Chem. Soc., 76, 5686 (1954). alternative sequence for chondrosine of C)-(3-D-glucopyranosyluronic acid-(1— >3)-2-amino-2-deoxy-D-galactopyranose. This is the only structure for this sequence consistent with the periodate oxidation data of Wolfrom and coworkers (46). The initial consumption of two moles of oxidant giving the crystalline oxidation product involves the glycitol portion and the slow further oxidation must be the oxidation of the glucuronamide moiety. C02 Me CHgOAc I)[h], IMi OAc 2,MeOH.H® CH20Ac OH H lOAc H, 3) Ac2 0, C5 H5 N AcO\l -/H H^

H OH NH® Cl0 NHAc

NH 3, EtOH *

CONHj, CONH H20H 0 HC02H + HCHO 2 IO4 CH20H H,OH ,OH ( fa s t) HO

H NHAc H OH H NHAc

C 0N H 2 CONH2 Jk1 i o r NH Ac NHAC (slow)

HP H K H^CHgOH 18 Davidson and Meyer (50) reduced the methyl ester hydrochloride of chondrosine with sodium borohydride and the Elson-Morgan (49) negative product gave, on hydrolysis, D-glucose as the only reducing sugar isolated, in support of their proposed sequence. Their N-acetylchondrosinol and N-acetylchondrosine acetate derivative methyl esters were identical to those prepared by Levene (45) and Wolfrom and coworkers (46). Since only P-glucosidase hydrolyzed the N-acetyl derivative of the glycitol, the P-D-configuration was assigned to the glucuronidic linkage. The attachment of the glucuronidic linkage was confirmed by Davidson and Meyer (51)• They oxidized chondrosine methyl ester

(51) E. A. Davidson and K. Meyer, J. Am. Chem. Soc., 77, 4796 (1955). hydrochloride with ninhydrin and reduced the product with sodium borohydride to give P-D-glucopyranosyl-(1—>4)-D-arabinitol (2-D-lyxitol) as the final product, characterized as the crystalline octaacetate and by periodate oxidation. This confirmed the C3 linkage for chondrosine. Sulfated ester group attachment. Infrared spectral analyses by Orr (52), refined recently by Mathews (53) and Hoffman, Linker

(52) S. F. D. Orr, Biochim. et Biophys. Acta, 14, 173 (1954). (53) M. B. Mathews, Nature, l8l, 421 (1958). and Meyer (54), have shown that in the carbohydrate fingerprint

(54) P. Hoffman, A. Linker and K. Meyer, Biochim. et Biophys. Acta, j?0, 184 (1958). region of 700 to 1,000 cmT^", chondroitin sulfate A and B have similar absorption spectra quite distinct from that of ohondroitin sulfate C. Mathews proposed that the difference in infrared spectra in this region was consistent with the sulfate group having an axial configuration in chondroitin sulfate A and B and an equatorial configuration in the C isomer. Structural studies along this line had been hampered by the lack of a suitable means of desulfation of chondroitin sulfate. The acetylative desulfation, using absolute sulfuric acid-acetic anhydride mixture (55)» was accompanied by considerable

(55) L. Wolfrom and R. Montgomery, J. Am. Chem. Soc., 72, 2859 (1950). acetolysis. Recently, the Kantor-Schubert (28) desulfation procedure, using 0.06 N methanolic hydrogen chloride, provided undegraded chondroitin (desulfated chondroitin sulfate) methyl ester in high yield. This desulfation technique was used with advantage by Karl Meyer and coworkers. Hoffman, Linker and Meyer (5^) showed that the sulfated tetrasaccharides produced by the action of testicular hyaluronidase oh chondroitin sulfates A and C maintained the parent polymers' characteristic infrared spectra in the region 700 to 1,000 cm."*" On hydrolyses to the disaccharides (removal of sulfate and acetyl groups and hydrolysis of the hexosaminidic linkage), the products were identical in spectra and in gross properties. Desulfated chondroitin sulfates A, B and C were shown to lose their characteristic absorption bands, again signifying that the difference in infrared absorption was due to the difference in sulfate ester attachment in the molecule. Meanwhile, chondroitin isolated from cornea has an infrared spectrum similar to the desulfated chondroitin sulfates. Action of 3-glucuronidase on the sulfated tetrasaccharide from chondroitin sulfate A gave free D-glucuronic acid and the sulfated trisaccharide, indicating that the sulfated entity was the hexosamine residue. 20 Meanwhile, Jeanioz and coworkers (29,56) employed methylation

(56) R. W. Jeanioz, P. J. Stoffyn and Monique Tremege, Federation Proc., 16, 201 (1957); "Chemistry and Biology of Mucopolysaccharides," 0. E. W. Wolstenholme and Maeve O'Connor, Ed., Little, Brown and Co., Boston, Mass., 1958, P» 86.

techniques with the Kantor-Schubert (30) desulfation to determine the positions of the sulfate ester group and the glycosidic linkage on galactosamine in chondroitin sulfate B. They succeeded in attaining total methylation of the polymer after six successive methylations with dimethyl sulfate and sodium hydroxide at 0 to 5°, and the nondialyzable, etherified product was recovered in 75% yield. Methanolysis yielded crystalline methyl 2-acetamido-2- deoxy-6-()-methyl-a~D-galactopyranoside as the principal product. Traces of a methylated hexuronic acid was isolated. Thus, the C3 and C*f of galactosamine are blocked in the polymer. Applying the same procedure on 90% desulfated chondroitin sulfate B, they obtained crystalline methyl 2-acetamido-2-deoxy-4,6-di-0?-methyl- P-D-galactopyranoside. These results indicate that the location of the sulfate ester has to be at Ck of the galactosamine moiety and the L-iduronidic linkage, at C3. Due to the similar infrared spectra of chondroitin sulfate A and B, Hoffman and coworkers (5^) inferred that these isomers must have the same sulfate ester attachments (at C*f) at the galactosamine moiety, which has been unequivocably proven by Jeanioz and coworkers (29,56) to be the case for the B isomer. It is interesting that A and B differ only in the configuration at C5 of the hexuronic acid moiety, this being D in chondroitin sulfate A and L in B. Since chondroitin sulfate C had a different infrared spectrum from A and B, implying that the sulfate is not at of galactosamine, the only possible attachment is at C6 of the hexosamine, since C3 is the point of glucuronidic linkage and C2 has the acetamido group, assuming the 21 pyranose form. All these conclusions confirmed the empirical deductions of Mathews (53)• Hexuronidic linkage. Davidson and Meyer (51) elucidated the attachment of the glucuronidic linkage in chondrosine, and by inference, that of chondroitin sulfate A and C, much earlier than the elucidation by methylation and desulfation techniques of the iduronidic linkage in chondroitin sulfate B by Jeanioz and coworkers (29,56), who by default showed also a C3 attachment to galactosamine. They made use of ninhydrin oxidation, first applied by Gardell and coworkers (57) and refined by Jeanioz and Stoffyn (57)

(57) S. Gardell, F. Heijkenskjold and A. Roch-Norlund, Acta Chem. Scand., 4, 970 (1950); P. J. Stoffyn and R. W. Jeanioz, Arch. Biochem. Biophys., 52, 373 (1954).

to distinguish between glucosamine and galactosamine by paper chromatography of the resulting pentoses. Meanwhile, Dorfman and Cifonelli (58) described a simple

(58) J* A. Cifonelli and A. Dorfman, J, Biol. Chem., 231. 11 (1958). colorimetric method for identifying C3-linked hexosamines by a 510 mu absorption maximum in the Elson-Morgan (49) reaction. On this basis, they (28,58) proposed the C3 glucuronidic linkage for the galactosamine moiety of chondroitin sulfate B, consistent with the results of Jeanioz and coworkers (29,56). Galactosaminidic linkage. The attachment of the hexosaminidic linkage on the hexuronic acid moiety of chondroitin sulfates was solved by the Columbia University group with Korn (59), through

(59) P. Hoffman, A. Linker, Phyllis Sampson, K. Meyer and E. Korn, Biochim. et Biophys. Acta, 2]5, 658 (1957). 22 enzymic studies using Flavobacterium heparinum enzymes, the bacterial source having been isolated from soil by Korn (60).

(60) E. D. Korn, J. Biol. Chem., 226, 8^1 (1957).

Unsaturated oligosaccharides were obtained from the action of unadapted bacterial enzyme on chondroitin sulfate A and C. The sulfated tetrasaccharide from A, on enzymic hydrolysis,gave two N-acetylated disaccharides, one saturated and the other unsaturated, showing that unsaturation resulted only from enzymic (hexosaminidic) hydrolysis. They isolated also sulfated fractions, hence, glycosidic cleavage preceded desulfation. By using adapted bacterial enzymes, they were able to partially hydrolyze chondroitin sulfate B to an unsaturated sirupy disaccharide, identical to that derived from chondroitin sulfate A and C. This disaccharide was characterized by ultraviolet spectral analysis to be the a,p- unsaturated acid,Ak-5 glucuronidoacetylgalactosamine.

Since the formation of an a ,3-unsaturated acid was consistent with uj3-dealkoxylation," the action of the hydrolytic enzyme consisted of the removal of hydrogen from C5 and of the hexo- sarainidic linkage from of the hexuronic acid moiety, leading to unsaturation and destruction of the asymmetry at C*f and C5 of the uronic acid. Hence, the same unsaturated disaccharide was derived from all the isomeric sulfates. These results were consistent with the assignment of the hexosaminidic linkage at Ok of the uronic acid moiety in all the chondroitin sulfates. In 23 general, results from enzymic hydrolysis should be interpreted with caution because of the possibility of transglycosylation (6l).

(6l) P. Hoffman, K. Meyer and A. Linker, J. Biol. Chem., 219, 653 (1956).

The presently accepted structures of the isomeric chondroitin sulfates are Chondroitin Sulfate A CO'

Chondroitin Sulfate B

s CO

H

Chondroitin Sulfate C

9 CO*

rt nhcoch3 24 Chondroitin. In 1954, Davidson and Meyer (62) isolated from

(62) E. A, Davidson and K. Meyer, J. Biol. Chem., 211, 605 (1954). bovine cornea a polysaccharide fraction containing equivalent amounts of hexosamine and uronic acid, but low (2%) in sulfate. By fractional elution from ion-exchange resin, it was separated into a high and a low sulfate fraction. The latter on hydrolysis produced chondrosamine. The disaccharide unit was prepared in high yield and had identical infrared spectrum with that of chondrosine prepared from chondroitin sulfate of hyaline cartilage. The name chondroitin was given to this polysaccharide which may be considered as the precursor of chondroitin sulfate A and C. Testicular and pneumococcal hyaluronidases hydrolyzed chondroitin at a rate comparable to that of hyaluronate. Keratosulfate. A mucopolysaccharide of still unknown structure, keratosulfate, - 1 0 to +6°, composed of equimolar amounts of N-acetylglucosamine, galactose and sulfate, has been isolated from bovine cornea (16) and nucleus pulposus (63) where

(63) S. Gardell, Acta Chem. Scand., 9» 1035 (1955)* it constituted about one-half of the total mucopolysaccharides and from the end pieces of the long bones of calf where it was only 1% of the total mucopolysaccharides (20). The other half of the mucopolysaccharides in cornea was made up of chondroitin sulfate A and chondroitin, whereas chondroitin sulfate C appeared to make up the rest in nucleus pulposus. Meyer, Hoffman and Linker (64)

(64) K. Meyer, P. Hoffman and A. Linker, Science, 128, 896 (1958). contended that keratosulfate may be more widely occurring than 25 presently believed and that it had been missed during the isolation of chondroitin sulfates because of the greater alcohol solubility of its salts* The occurrence of glucosamine in articular cartilage, reported by Kuhn and Leppelmann (65), may be attributed to

(65 ) R. Kuhn and H. J. Leppelmann, Ann., 611, 25^ (1958).

keratosulfate (66). Cifonelli and Dorfman (58) suggested a

(66) A. Hallen, Acta Chem. Scand., 12, I869 (1958).

C3-substitution of the hexosamine in keratosulfate based on the Elson-Morgan (^9) ultraviolet absorption spectrum. It was resistant to the action of hyaluronidases. Meanwhile, a new amino sugar, probably talosamine, has been detected in the acid hydrolyzates of chondroitin sulfate from hyaline cartilage (67). However, Crumpton (68) isolated N-

(67) Helen Muir, Biochem. J., ^5, 33p (1957)* (68) M. J. Crumpton, Nature, 180, 605 (1957)*

acetyltalosamine chromatographically in a sample of N- acetylgalactosamine derived from galactosamine isolated from sheep tracheal cartilage. He could not detect talosamine in the starting sample of galactosamine from the same source, indicating that the former probably does not occur in nature. Future work will surely shed light on these anomalies. N-Deacetylation and Sulfation of Polysaccharides N-Deacetylation of mucopolysaccharides. N-Deacetylation methods so far tried on mucopolysaccharides have been accompanied by alteration of the remainder of the polymer. Both basic and acidic reagents have been applied but with no marked selectivity noted. In 1953» Meyer and Baldin (69) prepared N-deacetylated 26

(69) K. H. Meyer and G. Baldin, Helv. Chim. Acta, 3jS, 597 (1953).

desulfated chondroitin sulfate by heating chondroitin sulfuric acid with N hydrochloric acid at 65° for 80 hours, but the product was very degraded. Aqueous mineral acids generally catalyze deacetylation, desulfation and also glycosidic cleavage. Alkaline reagents have been used successfully to N-deacetylate chitin to chitosan. Fusion with solid potassium hydroxide at a high temperature for over an hour has been employed (70,71)* which (70) C. R. Ricketts, Research (London), £i, 17jS (1953). (71) S* A. Barker, A. B. Foster, M. Stacey and J. M. Webber, J. Chem. Soc., 2218 (1958).

exemplifies the drastic conditions required for N-deacetylation. In fact, acetamido groups are unaffected by methanolic ammonia, the reagent commonly used for ^-deacetylation. Jeanioz and Forchielli (72) obtained 50% deacetylated hyaluronate by saponifying with 5 N

(72) R. W. Jeanioz and E. Forchielli, J. Biol. Chem., 190, 537 (1951).

sodium hydroxide at 80° for an hour. Less than 20% recovery was reported for the treatment. Hydrazinolysis was attempted by Matsushima and Fujii (73) on barium chondroitin sulfate for a

(73) Y* Matsushima and N. Fujii, Bull. Chem. Soc. Japan, J50, 48 (1957).

10 hour period at 100°. From the Van Slyke analysis of the product, they calculated 90% deacetylation, but they did not characterize the partially deacetylated material further. Although partial 2? desulfation (saponification) would be expected to have occurred, no quantitative study of this side-reaction was made. In this laboratory, various methods of deacetylation have been applied to chondroitin sulfate A with the preparation of sulfated N-deacetylated chondroitin sulfate in mind (7^— -76)

(7k) M. L. Wolfrom, (Miss) T. M. Shen and C. G. Summers, J. Am. Chem. Soc., 75, 1519 (1953). (75) C. G. Summers, Ph. D. Dissertation, The Ohio State University, (1955). (76) E. D, Toro-Feliciano, M. Sc. Thesis, The Ohio State University, (1957).

(Table 2). No notable procedure was evolved since the better and promising methods were not reproducible. Deacetylation data were calculated from acetyl analyses after Chaney and Wolfrom (77)•

(77) A. Chaney and M. L. Wolfrom, Anal. Chem., 28, l6l4 (1956).

To reduce the extent of stepwise degradation during alkaline deacetylation of chondroitin sulfate, Toro-Feliciano (76) essentially eliminated the terminal (free) potential aldehyde group by reduction with sodium borohydride, but treatment of reduced chondroitin sulfate with 20% sodium hydroxide for 100 hours resulted only in 18% deacetylation. A series of other experiments with reduced chondroitin sulfate and strongalkali were also of little value for improving the efficiency of deacetylation. Sulfated polysaccharides. Sulfated polymers, natural or synthetic, have the general property of retarding in vitro the clotting of freshly shed blood. Among the naturally occurring sulfated polysaccharides, notably the algal polysaccharides and the animal polysaccharides, chondroitin sulfates, heparin and 28

Table 2 Deacetylation of Sodium Chondroitin Sulfate A (75)

Method Deacetylation %

Sodium methoxide in absolute methanol 13*5 Sodium benzyloxide in absolute benzyl alcohol 16.0 Cuprammonium hydroxide 0 4% Sodium hydroxide 8.0 509^ Sodium hydroxide 60.0 — Amide exchange 77.0 - Methanolic hydrogen chloride (trace) 0

— Nonreproducible 29 keratosulfate, only heparin is highly sulfated. Since Demole and Reinert (78) and Fischer (79) noted that blood anticoagulants

(78) V. Demole and M. Reinert, Arch, exptl. Pathol. Pharmakol., 158, 211 (1930). (79) A. Fischer, Biochem. Z., 2k0, 36** (1931). such as hirudine (a crude extract from the leech) and heparin contained sulfate ester functions, much work has been done in attempts to develop sulfated polysaccharides of high anticoagulant activity comparable to that of heparin, but without success despite the fact that preparations of molecular weights and degrees of sulfation similar to heparin have been achieved. Sulfuric acid esters of polyvinyl alcohol (80,8l), cellulose (80,82), xylan (83)*

(80) E. Ghargaff, F. W. Bancroft and M. Stanley Brown, J. Biol. Chem., 11£, 155 (1936). (81) P. Karrer, E. Usteri and B. Camerino, Helv. Chim. Acta, 27, l*f22 (19*+*0. (82) P. Karrer, H. Koenig and E. Usteri, Helv. Chim. Acta, 26, 1296 (19^3). (83) K. H. Meyer, R. P. Piroue and M. E. Odier, Helv. Chim. Acta, 35, 57^ (1952). chitin (8 2 ,8 3 ), chondroitin sulfuric acid (82,83), dextran (8*f,85)

(8*0 A. Gronwall, B. Ingelman and H. Mosiman, Upsala Lakareforen. Forh., 50, 397 (19^5)* (8 5 ) C. R. Ricketts and K. W. Walton, Chem. & Ind. (London), 869 (1952). and synthetic polyglucose (86,87), have been prepared. The 30

(86) E. London, R. S. Theobald and G. D. Twigg, Chem. & Ind. (London), 1060 (1955). (87) J. W. Wood and P. T. Mora, J. Am. Chem. Soc., 80, 3700 (1958). importance of the N-sulfate group of heparin has been discussed frequently. /'-Heparin (N-desulfated heparin) has virtually no activity (88). However, the presence of the sulfoamino group (89)

(88) M. L. Wolfrom, D. I. Weisblat, J. V. Karabinos, W. H. McNeely and J. McLean, J. Am. Chem. Soc., 65, 2077 (19^5), M. L. Wolfrom and W. H. McNeely, J. Am. Chem. Soc., 6£, 7^8 (19^+5), A. B. Foster and A. J. Huggard, Advances in Carbohydrate Chem., 10, 555 (1955). it (89) J* E. Jorpes, H. Bostrom and V. Mutt, J. Biol Chem., 183, 607 (1950). alone cannot account for the high anticoagulant activity of this mucopolysaccharide. In the case of dextran sulfate, the influence of molecular weight and degree of sulfation on anticoagulant activity and toxicity has been studied (83) and a relatively low molecular weight product with two to three sulfate groups per anhydro-D-glucose unit, has been developed in England for use as an anticoagulant, having one fifth the activity of heparin (90).

(90) G. D. Forwell and G. I. C. Ingram, J. Pharm. Pharmacol,, 8, 550 (1956).

The two most commonly used sulfating agents for polysaccharides are chlorosulfonic acid and sulfur trioxide (sulfuric anhydride) in various solvents and in the form of complexes. Pyridine- chlorosulfonic acid has been the standard O-sulfating agent for most polysaccharide preparations (80-87)*. Wood and Mora (87) reported 31 obtaining polyglucose sulfate about one-third as active as heparin. In a critical study of the sulfation of chondroitin sulfuric acid, Meyer, Pirou/ and Odier (83) reported that the use of sulfur dioxide-sulfur trioxide or sulfur dioxide-chlorosulfonic acid mixtures gave the best results. They found their sulfated (12 to 17 per cent) preparation, by in vivo and in vitro assays, to have anticoagulant activity about 50% that of heparin. Since heparin is a mucopolysaccharide, related polymers with potential N- and ^-sulfate groups have been exploited. Chitosan (N-deacetylated chitin) was homogeneously sulfated in this laboratory, using the sulfur trioxide-N,N-dimethylformamide complex in N,N-dimethylformamide (91*92). The preparation had about one half

(91) T, M. Shen Han, Ph. D. Dissertation, The Ohio State University, (195*0* (92) M. L. Wolfrom and T. M. Shen Han, J. Am. Chem. Soc., 81, 1764 (1959). of the anticoagulant action of heparin. Shen Han (92) also prepared fully N- and partially O^-sulfated chitosan with pyridine- chlorosulfonic acid, half as active although twice as toxic as heparin. Summers (75) similarly esterified partially N- deacetylated chondroitin sulfate and his product had approximately 40% of the activity of heparin. Since the sulfoamino function is present in the D-glucosamine unit of heparin, derivatives of D-glucose containing the sulfoamino group were similarly prepared employing pyridine-chlorosulfonic acid (93)*

(93) M. L. Wolfrom, R. A. Gibbons and A, J. Huggard, J. Am. Chem. Soc., 79, 5043 (1957).

Chlorosulfonic acid in dichloroethane had also been successfully tried for esterifying chitosan (94). Recently, selective sulfation 32

(9*0 I. B. Cushing, R. V. Davis, E. J. Kratovil and D. W. MacCorquodale, J. Am. Chem. Soc., 76, ^590 (195^)•

of chitosan and representative amino alcohols was achieved by Warner and Coleman (95)* N-Sulfation of chitosan was accomplished

(95) D. T. Warner and L. L. Coleman, J. Org. Chem., 235., 1133 (1958). with pyridine-sulfur trioxide at pH 9 to 10, and the product was found to have no activity (95)* This N-sulfated chitosan, on ^-sulfation to the extent of 75% with sulfur dioxide-sulfur trioxide, acquired high activity. However this sulfated preparation exhibited a delayed toxicity which makes it unusable (96).

(96) L. L. Coleman, personal communication.

The authors (95) also claimed selective O-sulfation of amino alcohols with chlorosulfonic acid in carbon tetrachloride. III. DISCUSSION OF RESULTS Purification and Characterization of Chondroitin Sulfate A Sodium chondroitin sulfate A. All investigations here reported were made on the same lot of tan-colored crude sodium chondroitin sulfate. The material was rid of impurities by silicate absorption, which had been used extensively in this laboratory for purifying chondroitin sulfate (17,75*76). The product, & Jjj -21.4° (water), was characterized by infrared absorption spectral analysis (Fig.2) to be chondroitin sulfate A. Mathews (53) and Karl Meyer and coworkers (54) elaborated on the differentiation of the isomeric chondroitin sulfates by their infrared spectra in the region 700 to 1,000 cm7^: chondroitin sulfate C has unique bands at 1,000, 820 and 775 CHIT'S whereas A has bands at 928, 855 and 725 cmT^. (B, absent in cartilage, resembles A with bands at 928, 855, 840 and 712 cmT’*'). The purified chondroitin sulfate sample had sulfate bands at 920, 848 and 720 cm?1 , denoting that the material was essentially chondroitin sulfate A. Hitherto, this material had been assigned as the A isomer, only by analogy to the findings of other workers. In the light of the reported presence in cartilage of two chondroitin sulfates, A and C, and of the related keratosulfate (20), the purified sample was subjected to the ethanol fractionation analysis of Meyer and coworkers (16) from calcium acetate buffer, to verify the homogeneity of this preparation. The fractions were characterized for component sugars by paper chromatography, the detection of uronic acid being supplemented by the Dische (22) carbazole assay. However, although the results (Table 5) indicated that the fractions precipitating below 50% ethanol were chondroitin sulfate and the more soluble fractions were keratosulfate, attempts to refractionate the combined former fractions gave anomalous results for isomer distribution. Meyer and coworkers (20) had encountered fractions, especially in bone, which could not be classified as either chondroitin sulfate A or C since their optical

33 Table 3

Characterization of Silicate-purified Sodium Chondroitin Sulfate A

Chondroitin Sulfate Keratosulfate

A A' —

Fraction from calcium acetate buffer, % ethanol 30 to 40 40 to 50 50 to 80

{a]20^, (c_ 1, water) — -25.0° -14.7° -4.0°; -2.1° (-28 to -32°X-l6 to -22°)(-10 to +6°)

Uronic acid (carbazole) test + + -

Wt. % of precipitated sample 92 6 2

Galactose (chromatography) -- +

Galactosamine (chromatography) + 4* -

Glucosamine (chromatography) - - +

Color white white light tan

— Identical in infrared absorption spectrum to A, denoting partially degraded (desulfated) A polysaccharide.—Reported values in parentheses are for chondroitin sulfates A and C and keratosulfate, respectively (20). 35 rotations and solubilities did not correspond. In addition, the hydrolyzates of these polysaccharides (A and C) are identical, such that supplementary data from infrared analysis was sought to verify the assignments from ethanol fractionation. The two fractions collected below ^Ofo alcohol both gave infrared spectra identical with that of the starting material, hence were chondroitin sulfate A. They probably differed only in molecular weight and degree of sulfation, since the sample was extracted from bovine cartilage with dilute alkali, conducive to alkaline degradation (12). This would explain the anomalous "isomer distribution" on ethanol refractionation of the combined chondroitin sulfate fractions, since this procedure was purposely for component distribution. This same behavior was observed from the larger scale preparation of calcium chondroitin sulfate A and for calcium chondroitin sulfate A derived from the barium salt by cation exchange. Since the silicate-treated sample contained traces of keratosulfate, alcohol fractionation procedures were employed to prepare keratosulfate-free chondroitin sulfate A. The calcium, potassium and barium salts were accordingly prepared and are discussed elsewhere. The high (less negative) specific rotation of this preparation as compared to those reported (20) may be attributed to partial desulfation during alkaline degradation and to traces of keratosulfate, with specific rotation close to zero, in the sample. Paper chromatographic analysis. Descending paper chromatography on Whatman No. 1 filter paper using 1-butanol, pyridine and water (3:2:1.5 by vol.) readily separated the components of hydrolyzates, except the hexosamines, into distinct zones. The acid hydrolyzate of polysaccharides, after neutralizing with barium carbonate, was passed through a weakly acidic cation exchange resin to make the effluent acidic, since hexosamines are unstable in neutral and basic solutions (97)* Strongly acidic resins adsorb hexosamines (98)* 36

(97) P- W. Kent and M. W. Whitehouse, "Biochemistry of the Aminosugars," Academic Press Inc., New York, 1953* P* 211. (98) K. Freudenberg, H. Walch and H. Molter, Naturwissenschaften, 3 0 , 87 (19^2).

Multiple development reduced the amount of tailing, especially of hexosamines. Since incomplete separation between galactosamine and glucosamine was attained, prior oxidation the the hexosamine with ninhydrin, after Stoffyn and Jeanioz (57), served to identify the amino sugar from the pentose produced, lyxose or arabinose, which were readily distinguishable by paper chromatography. Calcium chondroitin sulfate A . Since the ethanolic fractionation of sodium chondroitin sulfate from calcium acetate •buffer, after Karl Meyer and coworkers (l6), removed keratosulfate contamination, this method was implemented on a larger scale on silicate-treated sodium salt to yield pure calcium chondroitin sulfate A. The fraction collected, in 72% yield, was that precipitating between ethanol concentrations of 30 to 50%. Most of the fraction precipitated between 30 to 40% ethanol. (The mother liquor was processed for the isolation of keratosulfate.) The product, -2^.3° (water), had identical spectrum to chondroitin sulfate A, This value was close to the of -25° for the chondroitin sulfate A fraction prepared above (Table 3)* Attempts to ethanol-refractionate this sample from calcium acetate buffer gave an anomalous "isomer" distribution of 1 to 1 in contrast to the first fractionation, which confirmed the contention already made that these two fractions (precipitating at 30 to 40 and 40 to 50% ethanol) were chondroitin sulfate A differing only in degree of desulfation and alkaline degradation. Keratosulfate isolation and desulfation. Since a less that 2% keratosulfate content in silicate-purified sodium chondroitin sulfate A was shown in the preliminary investigation (p. 3^), 37 attempt was made to isolate this mucopolysaccharide contamination from the mother liquor of the precipitation of calcium chondroitin sulfate A, after Meyer and coworkers (16). Refractionation was made because of the presence of chondroitin sulfate A impurities, shown by the presence of galactosamine and also by infrared analysis of the fraction separating at 50% ethanol, which verified the absence of substantial amounts of the C polysaccharide. The crude keratosulfate preparation, still contaminated with chondroitin sulfate A, [aJ^+O.40 (water), constituted 0.5% of the commercial chondroitin sulfate. Comparison of the infrared spectrum of keratosulfate (Fig.5 ) with that of chondroitin sulfate A (Fig.2) revealed interesting differences. In the region 700 to 1,000 cm.\ keratosulfate peaks at 998, 820 and 775 cm. resembled those of chondroitin sulfate C reported at 1,000, 820 and 775 cm. considered by Mathews (53) to be characteristic for an equatorial sulfate group, whereas those of chondroitin sulfate A were characteristic for an axial group. Since in keratosulfate, D-glucosamine is the hexosamine residue, both the C k and C6 hydroxyl groups are equatorial such that sulfation at either carbon should result in similar infrared absorption spectra. Since keratosulfate is devoid of uronic acid, the 1,612 cm7^" carboxylate band was absent and the acetamido bands at 1 ,6^8 and 1,585 cm.^ were more distinct than in chondroitin sulfate A. The infrared spectrum of keratosulfate has not been hitherto reported. Desulfation of keratosulfate (5»59% S) was attempted to determine the change in the infrared spectrum of such structural change. However, methanolic hydrogen chloride treatment (30) gave incomplete desulfation, as evidenced by infrared spectral analysis and sulfur content (3. 2*f% S). No marked change in the spectrum was noted (Fig. 3 and 4). However, the presence of chondroitin sulfate A in this preparation was confirmed by the presence of weak methyl glucuronate absorption at 1,730 cm. 38 The observed resistance to desulfation of keratosulfate is noteworthy in view of the reported similar behavior of heparin. Danishefsky and Eiber (99) subjected heparin (N:S ratio of 2:5)

(99) I. Danishefsky and H. B. Eiber, Federation Proc., 18, 210 (1959). to methanolic hydrogen chloride (3 0 ) for 4 days and the product had the N:S:0Me ratio of 1:1:1, indicating a difference in lability to acid of the sulfate groups in heparin. The lability of the sulfate group in chondroitin sulfate A and C and the stability of that in keratosulfate is hard to explain since the structure of the latter is still unknown. However, chondroitin sulfate B had not been completely desulfated by this method (29), although differing only from the A polysaccharide in the hexuronic acid moiety. Although heparin and keratosulfate have the same hexosamine moiety, they differ in that the former contains D- glucuronic acid, whereas the latter has D-galactose. Barium chondroitin sulfate A . Malawista and Schubert (*f3) had applied ethanol fractionation from barium chloride solution of chondroitin sulfate, derived from the alkaline hydrolysis of chondromucoprotein, to prepare the pure barium salt. This procedure was applied on silicate-purified sodium chondroitin sulfate A to determine the merits of the method for removing keratosulfate impurities. The chondroitin sulfate fraction collected in 53% yield at ethanol concentrations from 20 to kCP/o was devoid of keratosulfate as shown by paper chromatographic analysis. The infrared spectrum of this salt,Calj) -20.0° (water), was identical to that of chondroitin sulfate A. The formula, Cl'tH19.l8Ba0.50N0ll(S03Ba0.5o'5Ha0 )0.82’ “ dicatea the polysaccharide had been partially desulfated during isolation, besides being slightly degraded as noted from anomalous ethanol fractionation from calcium acetate buffer. Malawista and Schubert 39 reported C^H^^BaNO^S* 5*^0 for the barium salt, [aj^ -20.1° (water), and observed about 10% liberation of inorganic sulfate (as the barium salt) in the chondroitin sulfate isolated from the hydrolysis of chondromucoprotein with dilute sodium hydroxide. Potassium chondroitin sulfate A. The procedure of Malawista and Schubert (^3) was applied to silicate-purified sodium chondroitin sulfate, using potassium chloride in place of barium chloride, and provided potassium chondroitin sulfate A, Ca3j) -19-7° (water) in 66% yield, devoid of keratosulfate. This indicated the potentiality of purification of chondroitin sulfate A from aqueous solutions containing the desired cation by ethanol fractionation. Again, infrared spectral analysis showed the potassium salt to be the A polysaccharide. Carboxyl-reduced Chondroitin The main drawback in the use of alkaline deacetylating agents on chondroitin sulfate has been alkaline degradation (^-elimination). P-Elimination occurs in alkaline solution, especially with (1— >4)- and (1— >3 )_linked polysaccharides, proceeding exclusively from the terminal reducing sugar and forming isosaccharinates and metasaccharinates, respectively (12). To suppress this degradation, Toro-Feliciano (76) reduced the terminal carbonyl group with sodium borohydride, but even then, the material resisted N- deacetylation and the recovery also was low from sodium hydroxide treatment. It is probable that the carboxylate function of the uronic acid exerts some effect in this reactivity. The approach here adopted to further minimize the extent of degradation during N-deacetylation was to reduce the carboxyl group of the anhydroglucuronic acid units of chondroitin sulfate A with sodium borohydride through the methyl ester, simultaneously reducing the terminal carbonyl function of the polymeric chain. Polymeric reduced chondroitin. Chondroitin methyl ester derived from potassium chondroitin sulfate A by the action of methanolic hydrogen chloride after Kantor and Schubert (30), was 4o subjected to exhaustive sodium borohydride reduction (32,33)• The best yield and reduction was attained by the use of borate buffer and sodium borohydride. The first reduction was 66% efficient, the second raised the total ester reduction to 85% and a third raised the reduction to 36%, the final product being recovered in 71% yield from the starting chondroitin methyl ester. Reduction was followed with a modified hexuronic acid assay of Dische (22) to allow correction for the presence of D-glucose. These data were consistent with those of Jeanioz (100) on desulfated chondroitin sulfate B methyl ester; the

(100) R. W. Jeanioz, personal communication.

first reduction attained 66% reaction and a repetition increased the reduction of the L-iduronic acid moiety to 85%. Potassium chondroitin sulfate A was derived from the sodium salt either by ethanol fractionation from dilute potassium chloride solution or by cation exchange. The former method was preferred since it removed keratosulfate contamination beforehand, although all the reduced products showed absence of keratosulfate. Chondroitin methyl ester was recovered from 60 to 82% and gave a methoxyl analysis slightly lower than the theoretical. Rotations were comparable to that of Kantor and Schubert (30)* Infrared spectral analysis (Fig. 6) showed the absence of the sulfate band at 1,240 cm?1 and the presence of a strong band at 1,740 cm7^ characteristic for carboxylate esters. The methyl ester was reduced after the general sodium borohydride procedure of Frush and Isbell (33), in. the presence either of borate buffer or cation exchange resin, to make the medium initially acidic. Reduction was started in acid medium since esters were more reactive to hydride attack in such medium and also to avoid saponification of the ester function prior to reduction. The reduction of carbonyl to alcohol groups proceeded *+1 in slightly alkaline medium. Considerable reduction of the terminal carbonyl group of the polymer was observed; Toro- Feliciano (76) attained as much as 97% reduction of this function for chondroitin sulfate for a 72-hour period with sodium borohydride. That the aldehyde form of sugars is involved in such reduction has been supported by the observation that the susceptibility to hydride reduction of sugars and sugar derivatives paralleled their rate of mutarotation (101). Of the

(101) H. Endres and M. Oppelt, Chem. Ber., 91, ^78 (1958). two buffering agents, borate buffer was easier to process since inorganic material was readily removed from the polymeric suspension by dialysis against distilled water. Dialysis followed by lyophilizing was preferred to ethanol precipitation for purifying the product. Although the material became increasingly water-insoluble with reduction, the reductions were effected even on the sparingly soluble, esterified substances. The use of sodium borohydride-alurainum chloride complex (102)

(102) H. C. Brown and B. C. Subba Eao, J. Am. Chem. Soc., 78, 2582 (1956). produced good yields of the alditols for D-glucosamine, D-glucurone and D-glucuronic acid methyl ester. However, extension to the polymeric material was hampered by the insolubility of the material in diglyme (diethylene glycol dimethyl ether), even in the presence of tetrahydrofuran and also by the anhydrous conditions required. Thus, the choice of reduction in aqueous medium was deserving since carbohydrate (uronate and aldonate) esters are notable exceptions to the general unreactivity of carboxylic acid esters toward sodium borohydride (31-53)* Esterification with methanolic hydrogen chloride was employed with high recovery of the ester. Diazomethane was objectionable, k2 since prior decationizing of the partially reduced material was impractical because the substance is sparingly water-soluble and will be hard to separate from cation exchange resins. Besides, Neukom and Deuel (103) had shown that pectin (partial methyl

(103) H. Neukom and H. Deuel, Chem. & Ind. (London), 683 (1958). ester of polygalacturonic acid) behaved like an oxidized poly­ saccharide in the presence of diazomethane, even at -20°, whereby glycosidic linkages within the chain were readily split by"p-dealkoxylation" (involving the cleavage of the glycosidic linkage P to the carboxylic ester group of pectin following the removal of the activated a-hydrogen at C5), which results in double bond formation between C*f and C5» However, this reaction was significant with pectate, since the carboxylate function was not very activating. This phenomenon may be extended to chondroitin sulfate. The reduced chondroitin samples were fully characterized. Paper chromatographic analysis detected only glucose and galactosamine, plus traces of glucuronic acid. Infrared spectral analysis (Fig. 7) verified the absence of sulfate bands at l,2*f0, 928, 832, and 723 cm7^ and that of the carboxylate band at 2,930 -1 cm. was detected, which may be attributed to the Cl-H of the P-D-glucopyranose residues (10*0.

(10*0 S. A. Barker, E. J. Bourne and D. H. Whiffen, Methods of Biochem, Anal., 3t 213 (1956).

Carboxyl-reduced chondrosine. Chondrosine, j3-(P-D- glucopyranosyluronic acid) — (1—>3)-2-amino-2-deoxy-D-galactose, has been the only isolable disaccharide from the acid hydrolysis of barium chondroitin sulfate (50). Possible explanations for 43 this selective hydrolysis, in which only one of two theoretically possible disaccharides has been isolated, have been the stability of the P-D-glycopyranosyluronidic bond, the instability of the 2-acetamido-2-deoxy-P-D-glycopyranosidic bond, or both. Reduced chondroitin (desulfated chondroitin sulfate), whose anhydro-D-glucose units were derived by sodium borohydride reduction of D-glucuronic acid methyl ester residues, converted D-glucopyranosyluronidic linkages into D-glucopyranosidic ones. Hydrolytic studies on this modification should determine the (selective) influence of the hexuronidic linkage in determining the disaccharide composition. Only the Elson-Morgan (49) disaccharide, carlboxyl-reduced chondrosine, 0_-P — D-glucopyranosyl-(1— -^3)-2-araino-2-deoxy — D-galactose, was detected on following the hydrolysis of carboxyl-reduced chondroitin paper chromatographically and polarimetrically, Constant specific rotation was noted within 3 hours of refluxing, with a final reading of +50°. Paper chromatography employed the 1-butanol, pyridine and water solvent mixture and the Elson Morgan spray (105) and alkaline (105) S. Ffi Partridge, Biochem. J. , 42j 238 (1948). silver nitrate reagent (106). Zones for N-acetyl-D-galactosamine

(106) W. E. Trevelyan, D. P. Proctor and J. S. Harrison, Nature, 166, 444 (1950)* and carboxyl-reduced N-acetylchondrosine were present during the first two hours of the reaction, but were subsequently absent, denoting that with this modification, hydrolysis was faster than N-deacetylation. This same phenomenon had been noted for chondroitin sulfate A (50)• After 13 hours, only traces of the disaccharide were detected. Since only carboxyl-reduced chondrosine was detected, aside 44 from traces of chondrosine (due to incomplete reduction), this dramatically showed that N-acetyl-|3-D-galactosaminidic linkages were more labile to acid hydrolysis than P-B-glucosidic linkages. This is in accord with the comparison of kinetic data (107,108),

(10?) R. C. G. Moggridge and A. Neuberger, J. Chem. Soc., 745 (1938). (108) A. B. Foster, D. Horton and M. Stacey, J. Chem. Soc., 81 (1957).

that the rate of hydrolysis of methyl N-acetyl-p-D-glucosaminide was 9-18 times that for methyl (3-D-glucopyranoside (109). Thus,

(109) E. A. Moelwyn-Hughes, Trans. Faraday Soc., 25* 503 (1929). the stabilization of the (3-D-glucuronidic linkage alone, for which the inductive effect of the carboxyl group had been advanced as an explanation (110), cannot account for the selective

(110) R. L. Whistler and G. N. Richards, J. Am. Chem. Soc., 80 , 4880 (1958). nature of the hydrolysis of chondroitin sulfate A, evidenced by the also selective hydrolysis of carboxyl-reduced chondroitin. (Desulfation of the latter molecule was not considered critical for the hydrolysis since the sulfate group was cleaved faster than the glycosidic centers.) This is in contrast with the results on nonnitrogenous heteroglycans containing uronic acid units, where hydrolysis occurs more extensively elsewhere in the molecule than at the glycopyranosyluronidic linkages, consequently aldobiouronic acids accumulate as resistant disaccharides in the hydrolyzates. Here, the inductive effect of the carboxyl group 45 of uronic acids is the major factor involved for this selective hydrolysis. To complete the hydrolysis of the aldobiouronic units, acid concentrations of 4% at 120° for 10 to 12 hours may be necessary (ill). However, liberation of glycuronic acid is

(ill) W. A. G. Nelson and E. G. V. Percival, J. Chem. Soc., 58 (1942).

accompanied by its considerable degradation (112); a plausible

(112) R, L. Whistler, A. R. Martin and M. Harris, J. Research Nat. Bur. Standards, 24, 13 (1940).

mechanism has been postulated for this acidic decarboxylation of hexuronic acids (113).

(113) E. Stutz and H. Deuel, Helv. Chim. Acta, 4l, 1722 (1958).

Isolation of the carboxyl-reduced chondrosine and the non- carboxylated monosaccharides, D-glucose and D-galactosamine, was undertaken to verify the assignments made above. From the preliminary hydrolytic studies, a hydrolysis period of 234 to 3 hours was found optimum for the recovery of the disaccharide. N-Acetylation (114) of the hydrolyzate. and adsorption of D-

(114) S. Roseman and J. Ludowieg, J. Am. Chem. Soc. 76, 301 (1954).

glucuronic acid and N-acetylchondrosine on anion exchange resin and subsequent fractionation on a carbon (115) column with

(115) R. L. Whistler and D. F. Durso, J. Am. Chem. Soc., 46 72, 677 (1950).

aqueous ethanol facilitated the isolation of the component monosaccharides and the disaccharide. D-Glucose was characterized as the (3-pentaacetate, whereas the N-acetyl-oc- D-galactosamine monohydrate, which was Morgan-Elson (116) positive,

(116) W, T. J. Morgan and L. A. Elson, Biochem. J., 28, 988 (1934).

was characterized also as the (3-pentaacetate, by reaction with acetic anhydride-zinc chloride (117). Chromatographically pure,

(117) M. Stacey, J. Chem. Soc., 272 (1944).

Morgan-Elson reactive carboxyl-reduced N-acetylchondrosine, ,0-P- D~glucopyranosyl~(l— >3 )-2-acetamido-2-deoxyl-a-D-galactose, was recovered from the 5% and 6% ethanol eluate as the dihydrate, m.p. 155-157°* JaJD+47°+19.2° (water). This disaccharide was readily hydrolyzed with acid to D-glucose and D-galactosamine. It is interesting that the repeating disaccharide unit of carboxyl-reduced chondroitin is also a dihydrate. In addition, their infrared spectra (Fig. 7 and 8 ) were very similar, except for the presence of an 807 cmT'*' peak for the disaccharide, absent in the polymer. Sharp bands for the acetamido function — 1 —1 at 1,620 and 1,560 cm. were noted, together with an 886 cm. peak characteristic for P-D-glucopyranose (104). This technique of reduction of the uronic acid of mucopolysaccharides and subsequent hydrolysis and identification of the hydrolyzates has potentialities in the study of mucopolysaccharide structures. This avoids the loss of the free uronic acid in the hydrolyzates, since the corresponding acid- stable hexose is formed. 4? Carbon column chromatography of neutral N-acetylated hydrolyzates was found better than fractionation on paper chromatograms using the basic solvent, 1-butanol, pyridine and water (3 :2 :1.5 by vol.), and better than cation-exchange column chromatography using aqueous acetic acid developer. In the latter cases, the carboxyl-reduced chondrosine fractions were isolated mixed with their component sugars, D-glucose and D-galactosamine. Degradation of amino sugars in the presence of organic bases has also been reported (118). The futile

(118) D. Aminoff, W. T. J. Morgan and W. M. Watkins, Biochem, J., 46, 426 (1950).

purification of the free disaccharide by carbon (115) column chromatography is consistent with the reported (71) unsuccessful fractionation of the homologous chitosaccharide hydrochloride salts by this method. Barker and coworkers (71) also solved the problem by prior N-acetylation (114) of the hydrolyzates before carbon column chromatography. The alternative use of cation exchange column chromatography, employed by Davidson and Meyer (50) for the isolation of chondrosine, was impractical. This is because carboxyl-reduced chondrosine was firmly adsorbed by the resin, requiring high acid concentrations for elution, since it is an amine base, unlike chondrosine which, although ionic (containing carboxyl and amino groups), acts as a uzwitterionM. Considerable hydrolysis of the disaccharide occured during purification. Carboxyl-reduced N-acetylchondrosine was readily degraded in aqueous alkaline solution to D-glucose and a Morgan-Elson (116) reactive sugar (®giucose nanhydro-N-acetyl-D- galactosamine," different from N-acetyl-D-galactosamine (R _ 1.2). This result is analogous to those reported by ^ 1 U C 0 S 6 Kuhn and coworkers (119) on 3-0-derivatives of N-acetyl-D- 48

(119) H. Kuhn, Adeline Gauhe and H. H. Baer, Chem. Ber., 87, 289, 1138 (1954). glucosamine. Alkaline degradation of 0-[3-D-galactopyranosyl- (1— *>3)-2-acetamido-2-deoxy-D-glucopyranose gave D-galactose and "anhydro-Itf-acetyl-D-glucosamine" ^ g i ucose 1*70) * different from N-acetyl-D-glucosamine (R , 1.24). Since the 4-O-p- — J 0 glucose — D-galactopyranosyl analogue did not yield "anhydro-N- acetylhexosamine and the 6-0-p-D-galactopyranosyl-N-acetyl-D- glucosamine gave instead D-galactopyranosyl-(l...>6 )- "anhydro-N-acetyl-D-glucosamine" (R , 1.04), alkaline — glucose degradation should be a valuable tool for determining 0- substitution of N-acetylhexosamines. 3-0-Methyl-IJ-acetyl-D- glucosamine also provided "anhydro-N-acetyl-D-glucosamine." These data support the C3 glycosidic linkage reported for chondrosine (51)* To further confirm the assigned structure of carboxyl- reduced N-acetylchondrosine for the isolated disaccharide, the substance was converted to the alditol by sodium borohydride reduction and the alditol was subsequently hydrolyzed. Acid hydrolysis gave (by paper chromatography) glucose as the only reducing monosaccharide and D-galactosaminol (2-amino-2-deoxy~ D-galactitol), which confirmed the fact that the hexose was the non-reducing end of the disaccharide. In addition, the alditol was chromatographically identical (®giuCOSe 0 *76) in three solvent systems to that derived from chondrosine methyl ester hydrochloride by sodium borohydride reduction and subsequent N-acetylation (50). Since partial degradation of the disaccharide to D-glucose and Manhydro-N-acetyl-D-galactosamine" occurred during alkaline borohydride reduction, the product consisted of the corresponding alditols. That the alkaline degradation occurred prior to reduction was confirmed by the alkali-stability of the carboxyl- reduced N-acetylchondrosinol purified by carbon column ^9 chromatography. This stability of the alditol is in keeping with the [3-elimination mechanism (12) for alkaline degradation, which requires a carbonyl group. Paper chromatographic analysis of the alditols showed them to have similar R^ values to the corresponding sugars. Crimmin (120) similarly noted related R^ values for N-acetyl-D-

(120) W. R. C. Crimmin, J. Chem. Soc., 2838 (1957).

galactosamine and N-acetyl-D-galactosaminol. Although the N-acetylhexosaminols were unreactive to aniline hydrogen phthalate (121), they were reactive to the Morgan-Elson (116) test.

(121) S. M. Partridge, Nature, 16^, hk3 (19^9).

The positive Morgan-Elson (116) reaction of carboxyl-reduced N-acetylchondrosinol was very interesting in view of the results of Kuhn and coworkers (119,122) and those of Jeanloz and associate

(122) R. Kuhn, Adeline Gauhe and H. H. Baer, Chem. Ber., 86 , 827 (1953).

(123). Although methyl and benzyl N-acetyl-D-glucosaminides were

(123) R. W. Jeanloz and Monique Tremege, Federation Proc., 15, 282 (1956). unreactive to the reagent, 6-0-(2-acetamido-2-deoxy-|3-D- glucopyranosyl)-D-glucose and-D-galactose and the oligosaccharide, lacto-N-tetraose, containing an N-acetyl-D-glucosamine residue linked at C3» gave positive Morgan-Elson tests despite the fact that the amino sugar is not an end group (122). In keeping with this seemingly characteristic reaction of C3-linked N- acetylhexosaminides, chondroitin sulfate A and its terminal 50

carbonyl-reduced modification gave the characteristic, but faint purple color for the test. N-Acetyl-D-glucosaminol (124) also

(124) Prepared by K. Onodera by borohydride reduction.

gave a positive color test. Since some N-acetylhexosaminides were reported (119,122) to be reactive, together now with N- acetylhexosaminols, and 4-0-substituted N-acetylhexosamines are unreactive without exception (119,122,123), the Morgan-Elson test must be interpreted with care, because of its nonspecific nature. These results eliminate a C4 glucosidic linkage for the disaccharide in question, because of its positive Morgan-Elson reaction. Degraded, reduced chondroitin. Degraded chondroitin methyl ester was derived from the deacetylation of chondroitin acetate (55) an^ subsequent esterification of decationized chondroitin with diazomethane. A 67% yield of 85% carboxyl-reduced, degraded chondroitin was attained by repeated sodium borohydride reductions (32,33). The product was less soluble in methanol and in water than the starting chondroitin, which can be attributed to the decreased ionic character of the compound because of the conversion of a carboxyl group to a primary alcohol group. Paper chromatography showed glucose and galactosamine with traces of glucuronic acid, glucosamine and galactose. The anomalously high nitrogen content may be partly explained by the degradation of the uronic acid moiety during acetylative desulfation (55)1 since its uronic acid assays (22) were consistently low. Reaction with ammonia during _0-deacetylation may account for the rest of the increase in nitrogen. Chondroitin acetate, obtained from sodium chondroitin sulfate A (silicate treated) by the acetylative desulfation technique of Wolfrom and Montgomery (55) had properties consistent with those reported by Madison (125); it was degraded and dialyzable. The C02CH

HO No B H 4

(7 1 % )

OH NHCOCH H OH H NHCOCH3

I) H ® , A

(1 0 % ) 2 ) A czO, Dowex I (CO

3) Carbon column

CH2OH CH 2 0H C H 2 OH

c h 2 o h 1) N 0 BH 4 2 ) C a rb o n column HO OH (6 0 % ) NHCOCH: OH 52

(125) R» K. Madison, Ph. D. Dissertation, The Ohio State University, 1950.

infrared spectrum (Fig. 5) revealed a sharp band at 1,470 cm?’*' for esters. The high acetyl and low hexuronic acid assays indicated marked degradation. The nitrogen analysis was also low. Deacetylation of the chondroitin acetate was done with methanolic ammonia at 0° and the product was decationiaed prior to esterification with etheral diazomethane. Madison reported that ammonolysis gave only partial saponification, since his product still contained 0.8 0-acetyl per anhydrodisaccharide unit. Diazomethane was readily prepared from "Diazald" (N-methyl- N-nitroso-ji-toluenesulfonamide) (126). The generator was simple

(126) T. J. De Boer and H. J. Backer, Rec. trav. chim., 731, 229 (1954). and yields of 90% were attained. Reduction of the methyl ester utilized sodium borohydride in the presence of cation exchange resin-borate buffer after Frush and Isbell (33)* This was convenient since the product was rid of inorganic matter by passing through cation exchange resin column and evaporating the boric acid as the volatile methyl borate before reesterifying with diazomethane. Hydrolytic studies on degraded, reduced chondroitin were hampered by the degraded state of the material and the presence of keratosulfate contamination. No defined zone for carboxyl- reduced chondrosine was observed, in comparison to the results with the polymeric preparation above. N-Deacetylation Studies

N-Deacetylation of Barium Chondroitin Sulfate A Alkaline reagents. Summers (75) and Toro-Feliciano (76) 53

had extensively employed alkaline reagents in the absence of oxygen in attempts to N-deacetylate sodium chondroitin sulfate A, without notable success. In the preparation of N- deacetylated chondroitin sulfate A, the major drawback had been the competing alkaline degradation, P-elimination, proceeding from the terminal reducing group of (1—»3 )-and (1 — ^ - l i n k e d polysaccharides (12). Even chondroitin sulfate A, whose terminal carbonyl had been reduced, resisted deacetylation and ivas recovered in low yield (76). Since these degradations occurred readily in aqueous alkaline media, amide exchange reaction under anhydrous condition was attempted by a modification of the hydrazinolysis of chondroitin sulfate by Matsushima and Fujii (73). Consistently good (^3%) yield of cream colored', highly (59-68%) N-deacetylated polymer was attained (Table k ), as compared to the 11% (weight) yield of 90% deacetylated product reported by these authors. The modifications consisted of the use of sealed tube as reaction vessel to exclude atmospheric moisture and oxygen from the reaction mixture during the 10-hr. 100° treatment and a slightly larger excess (100 to 1 ) of hydrazine than in the original (50 to 1) was used. Scaling up the amount of reactants three­ fold increased the recovery to 66%. The quality of the hydrazine reagent was also found to be critical. Although the selective nature of the reaction made the reagent useful for the determination of carboxyl-terminal amino acids of polypeptides (127) because of the resistance of these

(127) S. Akabori, K. Ohno and K. Narita, Bull. Chem. Soc. Japan, 25, 214 (1952).

amino acids to hydrazide formation to which peptide-linked amino acids were very susceptible, hydrazinolysis of barium chondroitin t sulfate A was found to be more complex than simply Ih—*K-acetyl migration. Partial saponification of the sulfate group was noted. 5^

Table k

N-Deacetylation of Barium Chondroitin Sulfate A

Polymeric product

Deacetyl- Yield, Desulfation,— Method ation, — % %%

Alkaline Reagents Hydrazine, 10 hr., 100° 59-68 39-53

Sodium hydroxide (5 N)» 1 hr., 80° 62 25 29

Acidic Reagents Hydrochloric acid (5 N), 3 hr., 37-^0° 21 ~100 36 6 hr., 37-^0° 25 67 52 12 hr., 37-^0° 32 3h 60

Methanolic hydrogen chloride, (IN), 72 hr., ambient 19 32 ~100

Benzyl alcohol-hydrogen

chloride (IN), 72 hr. 9 ambient low high high

— Acetyl analysis after Chaney and Wolfrom (77)* - The starting barium chondroitin sulfate A was already 18% desulfated. 55 Alkaline hydrolysis of sulfate esters sometimes leads to the formation of anhydrosugars (128); however, the sulfate group

(128) W. G. Overend and C. R. Ricketts, Chem. & Ind. (London) 632 (1957). at C^f of the N-acetyl-D-galactosamine residues has no vicinal hydroxyl group. More disturbing was the anomalous nitrogen analysis of the product; N-deacetylated chondroitin sulfate A had 1.60 and 1.72 nitrogen per anhydrodisaccharide unit; N-deacetylated chondroitin (desulfated chondroitin sulfate) had 1.22; and N-deacetylated carboxyl-reduced chondroitin had 0.89. These results indicate that the increase in nitrogen content arose from the reaction of hydrazine with both carboxylate-and sulfate-containing modifications. Both these functions were absent in the last sample, where no increase in nitrogen was noted. This sample also has negligible terminal carbonyl. It is probable that hydrazone, osazone, hydrazide and sulfohydrazino group, ROSC^NHNH^, formation occurred in the other modifications. No enlightenment from infrared data was possible for hydrazide detection since the spectrum was identical to that of the starting material. Studies of hydrazinolysis of model compounds would give more insight into the nature of the reaction than further study of the materials at hand. The normal hexuronic acid assay (22) precludes extensive degradation of this moiety on hydrazine action. Comparison of various deacetylation methods studied (Table *0 showed that hydrazinolysis gave the most consistent reaction, without total desulfation, in spite of the side reactions. This method was thus applied for the N-deacetylation of chondroitin sulfate modification discussed elsewhere. Although Summers (75) and Toro-Feliciano (76) had done extensive studies on the action of sodium hydroxide solutions of various concentrations on sodium chondroitin sulfate A at room temperature for various' reaction times, no high-temperature 56 reaction was tried to decrease the reaction period. Jeanloz and Forchielli (72) reported 50% N-deacetylation of hyaluronate by heating in 5 N sodium hydroxide at 80° for an hour. However, they used acetic acid to neutralize the solution, and, even then, a l6% yield of highly deacetylated product was obtained. A higher deacetylation would have resulted if mineral acid were used instead for neutralizing. Thus, to investigate the feasibility of extending this drastic method to chondroitin sulfate A, the barium salt was heated in 5 N sodium hydroxide for 1 hr. at 80° and neutralized with dilute sulfuric acid. Dialyzing conveniently retained the polymeric fraction, 62% N-deacetylated tan chondroitin sulfate A, in 25% yield. Alkaline degradation and oxidation of the polysaccharide accounted for the low recovery of degraded product, making the method impractical. Acidic reagents. Although acidic reagents have been extensively employed for the N-deacylation of monomeric amides, these have seldom been used on polymeric materials because acids also catalyze hydrolysis. The drawbacks against acidic N-deacetylation of chondroitin sulfate A are that simultaneous desulfation occurs together with considerable hydrolysis of the polysaccharide (Table 4), The facts that a polymeric deacetylated product was desired complicated the matter further, since even with the disaccharide, chondrosine, both N-acetylated and free amino forms were obtained on hydrolysis of the polymer, indicative that hydrolysis was faster than N-deacetylation (50). Meyer, Odier and Siegrist (56) reported that sodium chondroitin sulfate can be desulfated 51*4% by reacting for 3 hours with 5 N hydrochloric acid at 37°, but with only 3% glycosidic hydrolysis. However, since their data were those for the reaction medium and not for the isolable polymeric product, an attempt was made to isolate and characterize this product (Table 4 and Fig. 1). From the results, time reaction with acidic reagents may be deduced as very inefficient for the Efficiency, % 100 0 4 0 6 20 0 8 0 gur . i 5 sulurc ai ih oyei chondr itin re d n o h c polymeric with acid ric lfu u s , 5N f o n tio c a e R I. re u ig F f e A t . ° 0 4 - 7 3 at A te lfa u s 5 me, hr. , e im T Deocet oton tio lo ty e c o e -D N A fc rotation ific c e p S • ecovery R □ n tio lfo esu D O 10

16 -1 - 15 4 1 - -20 18 -1 -10' -12

C«c.] q (w a te r) 58 purpose of attaining a high recovery of highly deacetylated chondroitin sulfate A. A steady increase in specific rotation was noted for the reaction products with time. Intramolecular acetyl migration has been reported for some IT-acetyl amino alcohols to proceed N— >0 to some extent in acid media (129), under conditions relatively drastic for hydrolyzable

(129) G. E. McCasland, J. Am. Chem. Soc., 73>, 2295 (1951); V. Bruckner, G. Fodor, J. Kiss and J. Kovacs, J. Chem. Soc;, 885 (19^8) •, G. Fodor and J. Kiss, J. Chem. Soc., 1589 (1952)*, G. Fodor and K. Nador, J. Chem. Soc., 721 (1953)* functions. In chondroitin sulfate A, the possibility of intramolecular N-acetyl migration is practically nil since the C3 hydroxyl is involved in glycosidic linkage and the anomeric group of the D-galactosamine residue is similarly blocked. The addition of reactive hydroxyl compounds, such as benzyl or methyl alcohol, may alleviate intermolecular N—>0 acetyl transfer under acid conditions. Barium chondroitin sulfate A was suspended in alcoholic hydrogen chloride, made anhydrous to suppress hydrolysis and N (b%) for effective acid catalysis. Since esterification of the uronic acid residues occur in the media, careful saponification of the recovered product was performed, Methanolic hydrogen chloride treatment of barium chondroitin sulfate A followed that of Kantor and Schubert, except the use of N instead of 0.06 N acid. Considerable methanolysis occured since only a 32% recovery of 19% deacetylated chondroitin resulted (Table *+). Complete desulfation was assumed since even with the less acidic solution, complete desulfation was observed. This showed that inefficient N-deacetylation of chondroitin sulfate A resulted from the action of acidic methanol. The use of benzyl alcohol instead of methanol as the alcohol medium gave anomalous results. Based on sulfur and acetyl analyses, high deacetylations were calculated, but the corresponding 59 recoveries were higher than the theoretical. The higher specific gravity of benzyl alcohol and its water insolubility may have hampered the reaction as compared to methanol. The fact that benzyl chloride, a lachrymator, was also detected in the reaction, poses the possibility of benzyl glycosidation of chondroitin sulfate A, which can account for the increase in weight, and thus for the low acetyl and sulfate assays. This is indirectly supported by infrared analysis, which showed a still marked sulfate absorption in the region 700 to 1,000 cmT^" characteristic for chondroitin sulfate A, in spite of the low sulfate content (0.31%). However, no defined characteristic aromatic ring absorption was noted due to the polyfunctional nature of the compound. Hydrolytic studies on N-deacetylated chondroitin sulfate A,

One possible explanation for the selective hydrolysis of chondroitin sulfate A to chondrosine'is the lability of the N- acetyl-P-D-galactosaminidic linkages. Partially N-deacetylated chondroitin sulfate A, with a major fraction of the amino function free, was subjected to hydrolytic studies to determine the effect of deacetylation on the nature of the disaccharides obtained. The hydrolysis was followed polarimetrically and paper chromatographically. An almost linear increase in polarimetric reading was noted which became constant within 13 hours, indicating a slower rate of hydrolysis than chondroitin sulfate A. Paper chromatography showed only chondrosine as the detectable disaccharide. 1£-Acetylated fractions (purple to Elson-Morgan spray (IO5 ) ) were noted during the first hour of refluxing, but were only in traces since the starting material was already partly deacetylated. No glucuronic acid was detected after 2k hours of hydrolysis, again confirming its acid-instability (113). The results may be considered tentative until further evidence is found in support of this selective hydrolysis. The slow change in specific rotation is indicative of the 60 resistance to hydrolysis of P-D-galactosaminidic linkage compared to N-acetyl-p-D-galactosaminidic linkage. The detection only of chondrosine reflects the relative stability of p-D-glucuronidic linkages to P-D-galactosaminidic linkages. The fact that heparin yields on hydrolysis only heparosin sulfate with the uronic acid as the reducing sugar (88), which involved preferential cleavage of a-D-glucuronidic linkage to N-sulfated-a-D-glucosaminidic linkage, may be rationalized on the basis of differences in anomeric configuration. In line with the fact that a-D- glycopyranosides are generally more resistant to hydrolysis than are the P-D-anomers, Moggridge and Neuberger (107) reported that methyl P-D-glucosaminide was hydrolyzed much faster than methyl a-D-glucosaminide, the a-D-anomer being three times more resistant (108). On the basis of an ionic repulsion mechanism by the amino center (at C2) of catalytic hydronium ions, the effect should be more important if the glycosidic group is nearest (cis) to the amino group, which was the case for methyl a-D-glucosaminide and also for heparin. Since P-D-galactosaminidic linkages are involved in N-deacetylated chondroitin sulfate, the analogy from glucosaminides may apply, since they have the same configuration at C2, but being p-D-linked, the protection afforded by the free amino group (NH^ ) to the trans glycosidic linkage should be less. Considering the stability of P-D- glucuronidic linkages, the detection only of chondrosine is in accord with these considerations. However, the definitive proof for the absence in the hydrolyzate of N-deacetylated polymer of the other disaccharide, C)-(2-amino-2-deoxy— P— D- galactopyranosyl)-(l—>4)-D-glucuronic acid, proposed by Wolfrom and coworkers (*f6 ) for chondrosine, has yet to be made. It is interesting to note that Linker, Hoffman and Meyer (130)

(130) A. Linker, P. Hoffman and K. Meyer, Nature, l80, 810 (1957). 61 recently extracted from the medicinal leech a hyaluronidase which hydrolyzes the endoglucuronidic linkages of hyaluronic acid. On exhaustive digestion of hyaluronate by this extract, the main product obtained was a tetrasaccharide with D-glucuronic acid as the reducing end. Until then, hydrolyses of hyaluronate with either acid or hyaluronidase had preferentially cleaved the endo-N-acetyl-hexosaminidic linkage. However, this hyaluronidase from leech was unreactive to chondroitin sulfate and chondroitin, preventing the exploiting of this enzyme to prepare the other disaccharide from chondroitin sulfate, different from chondrosine (46). Hydrazinolysis of Chondroitin Sulfate Modifications Since hydrazinolysis gave the most consistent N-deacetylation results for barium chondroitin sulfate A, the relative influence of the various substituents of chondroitin sulfate on the efficiency of the reaction was studied. Hydrazinolysis of various chondroitin sulfate modifications was undertaken and compared with the data for barium chondroitin sulfate A (Table 5). Carbonyl-reduced sodium chondroitin sulfate A was prepared by sodium borohydride reduction of chondroitin sulfate A after the general method of Head (131) as employed by Toro-Feliciano

(131) F. S. H. Head, J. Textile Inst., 46, T584 (1955).

(76). The 50% increase in yield of N-deacetylated product was consistent with the suppression of P-elimination in the alkaline reaction medium, proceeding from the terminal reducing sugar of the polymer chain (12). The use of the sodium instead of the barium salt of the compound did not affect the validity of the results. Unexpected results were obtained for the desulfated modifications, chondroitin and reduced chondroitin. Chondroitin (desulfated chondroitin sulfate), on hydrazinolysis, gave about twice as much highly N-deacetylated material as the parent Table 5

Hydrazinolysis of Chondroitin Sulfate A and Its Modifications

Nondialyzable product

N-Deacetyl- Yield, Desulfation, ation,4. * —a % % %

Chondroitin Sulfate A — Barium salt 59-68 42-44 59-55- Sodium salt 59 48 50 -

Chondroitin Sulfate Modifications Sodium chondroitin sulfate A, 92% carbonyl-reduced 60-64 66-67 28-55 Chondroitin, sodium salt 62-75 80-82 (100) - Chondroitin, 96% carboxyl- reduced 50 82 (100) — Chondroitin, 90% carboxyl- reduced 28 68 (100) —

Si ][) — Acetyl analysis after Chaney and Wolfrom (77). — Starting Q material was 18% desulfated. — Desulfated starting material. 63 polysaccharide, whereas carboxyl-reduced chondroitin, while giving similar recoveries, hampered the N-deacetylation efficiency by one-half. This implies that the carboxylate function of the uronic acid moiety of chondroitin sulfate A is critical to the reaction. A study of the molecular model for chondroitin sulfate A showed the capacity of the carboxylic acid carbonyl group to form hydrogen bonds with the amino hydrogen of the acetamido group of the hexosamine residues. However, the presence of a limiting amount of anionic groups in the molecule as essential to effective N—+N*acetyl transfer, whether sulfate or carboxylate anions, cannot be discounted as an alternative explanation. The high recovery of material from reduced chondroitin may also stem from the suppression of alkaline degradation ((3-elimination). However, the best results were obtained from chondroitin, as far as deacetylation was concerned, but the serious objection is the prior desulfation of starting material, especially when the product in mind was a sulfated N-deacetylated polymer. The sulfate on C4 of the hexosamine residue of chondroitin sulfate A is too far removed from the acetamido group at C2 to influence the reaction. The sodium salt of chondroitin was employed instead of the methyl ester since the latter is more prone to hydrazide formation than the former. Similarly, partial desulfation accompanied the hydrazine treatment and, with the sulfate and carboxylate containing modifications, an increase in the nitrogen content was noted. No significant change in the infrared absorption spectra (Fig. 2 and 12-14) was noted from the hydrazine treatment. All the N-deacetylated products had high specific rotations than the starting materials. Sulfated Chondroitin Sulfate Modifications Sulfation of various chondroitin sulfate modifications was undertaken to determine their relative anticoagulant potencies as compared to heparin. Sulfation with chlorosulfonic acid- pyridine (75) or with sulfur trioxide-N,N-dimethylforraamide (9 1 ,92) increased the anticoagulant activity (91*92,132) of the 64

(132) 0. F. Swoap and M. H. Kuizenga, J. Am. Pharm. Assoc., 38, 563 (1949).

preparations to only 15% that of heparin, but was comparable to that of sulfated chondroitin sulfate. This indicates that prior N-deacetylation by hydrazine treatment of chondroitin sulfate A did not enhance the activity of the sulfated product. The high activity (40% that of heparin) reported by Summers (7^,75) for chlorosulfonic acid-pyridine sulfated sodium hydroxide N- deacetylated chondroitin sulfate was not duplicated. However, the alteration of the rest of the molecule by the hydrazine treatment, as denoted by the increase in nitrogen content, prevented comparison with the sodium hydroxide-deacetylated material of Summers. Of the two sulfating methods employed, the sulfur trioxide- N,N-dimethylformamide method (91,92), although using a larger excess of reagent, had the advantage of being essentially homogeneous, since the partially sulfated polymer became progressively soluble in N,JM-dimethylformamide. Since both of these reactions were heterogeneous initially, activation of the polymer was critical for effective sulfation. A highly sulfated product with 3 sulfate groups per anhydrodisaccharide unit was obtained by the sulfur trioxide sulfation, whereas 2.5 sulfate groups per anhydrodisaccharide unit was attained by the chlorosulfonic acid method, although the recovery was higher for the latter method. The higher sulfation by the former method may be explained by the fact that only sulfated fractions dissolved in the amide solvent and the improperly activated fraction remained in suspension and was readily removed by decantation. Although these degrees of sulfation are comparable to that of heparin (5 sulfates per anhydrotetrasaccharide unit), the sulfated chondroitin sulfate modifications had lower 65 anticoagulant activity. N-Deacetylated chondroitin gave the least sulfation and the lowest anticoagulant activity of the preparations. Sulfation enhanced the water solubility of carboxyl reduced N-deacetylated chondroitin. Increase in specific rotation was noted upon sulfation. The presence of sulfoamino groups was verified by the negative ninhydrin reaction and from infrared absorption spectral analysis. The N-H deformation band at 1,560 cm. , prominent in sulfated chondroitin sulfate (Fig. 16), is absent in heparin (Fig. 17), and has been proposed (lO^f) to distinguish between acetamido and sulfoamino groups in mucopolysaccharides. Examination of the spectrum of sulfated N-deacetylated chondroitin sulfate (Fig. 15) showed only a weak absorption in this region, indicative of N-sulfation of the molecule, but also the presence of acetamido functions since the material was only partially deacetylated. No deacetylation was noted during the sulfation as evidenced by the consistent acetyl content (77) of the products. The appearance of very strong sulfate bands at 1,2^0 cm7^ and in the region 700 to 1,000 cm.’*' characteristic for equatorial sulfate (55) and the high sulfur analysis supported O-sulfation. Despite the unsuitability of hydrazine in the preparation of sulfated N-deacetylated chondroitin sulfate of high activity (7^* 75)* it should find more use especially in deamination studies (73*133) on N-acetylated mucopolysaccharides, which could be of

(133) M. Stacey, "Chemistry and Biology of Mucopolysaccharides," G. E. W. Wolstenholme and Maeve O'Connor, Ed., Little, Brown and Co., Boston, Mass., 1958* p. 8 . potential value for sequence determination of the component monosaccharides. IV. EXPERIMENTAL Chondroitin sulfate A was obtained from Wilson Laboratories, Inc., Chicago, 111., as a crude preparation. Magnesol is a synthetic magnesium acid silicate produced by the Westvaco Division of the Food Machinery and Chemical Corp., South Charleston, W. Va. Celite (No. 535) is a siliceous filter-aid produced by the Johns-Manville Co., New York, N. Y. The West Virginia Pulp and Paper Co., Chicago, 111., produces the carbon, Nuchar C unground. Dowex 1 and 50 are products of the Dow Chemical Co., Midland, Mich., whereas Rohm and Haas Co., Resinous Products Division, Philadelphia, Pa., produces Amberlite IRC-50 and MB-3. All elementary analyses were performed in part by the Galbraith Microanalytical Laboratories, Knoxville, Tenn., and by the Schwarzkopf Microanalytical Laboratory, Woodside, N. Y. Infrared absorption spectra of samples (potassium bromide pellet, Fig. 2-17) were obtained with the Baird-Atomic infrared recording spectrophotometer (model B). Purification and Characterization of Chondroitin Sulfate A Sodium chondroitin sulfate A . The procedure of Summers (75) as simplified by Wolfrom and Onodera (17) was employed on commercial sodium chondroitin sulfate. From the tan powder (100.0 g.), a white Magnesol-Celite (5*1 by weight) treated product was obtained; yield 60 to 80 g. 25n-21.4° (c 1 .32,

m *i25 o water); Summers (75) cites [aj v -25*1 (_c 1.46, water) for this preparation. Anal. Calcd. for C ^ H ^ N N a g O ^ S : uronic acid, 38.6 .Found: uronic acid (22), 32. Paper chromatography of acid hydrolyzate on Whatman No. 1 filter paper, using the 1-butanol, pyridine and water (3 :2 :1.5 by vol.) solvent and aniline hydrogen phthalate spray (121) revealed the presence of a trace of galactose contamination. Infrared spectral analysis (Fig. 2) showed the characteristic —X sulfate bands at 920, 848 and 720 cm. for chondroitin Sulfate A (53).

66 To determine the homogeneity of the sample the alcohol fractionation procedure of Meyer and coworkers (l6 ) was adapted. An amount of 1.50 g. of Magnesol-Celite treated material was dissolved in 25 ml. of acetate buffer (5% in calcium acetate and 0.5 N in acetic acid). Upon dilution with ethanol to 50% alcohol, no precipitation was observed. A voluminous precipitate formed when the alcohol concentration was increased to *f0%, the insoluble fraction being collected after standing overnight at 5°. Similarly, some precipitation occurred on dilution to 50% ethanol and this fraction was collected as above by centrifugation. Fractions precipitating at ethanol concentrations 50 to 70% and 70 to 80% were also isolated. The first two fractions were each dissolved in water, treated with Lloyd's reagent, filtered and precipitated from ethanol, washed thoroughly with ethanol, ether, and stored over phosphorus pentoxide. Physical properties and composition of these various fractions were determined (Table 5). Infrared spectral analyses of these fractions showed that in the region 700 to 1,000 cm.'*', the two fractions collected below 50% alcohol absorbed differently from those collected above 50% alcohol. The former fractions gave identical spectra, which were also identical with that of the purified sodium chondroitin sulfate, with bands at 720, 8^-8 and 920 cm."*', reported for chondroitin sulfate A (53)* Paper chromatographic analysis. Descending paper chromatography on Whatman No. 1 filter paper, usually using a mixture of 1-butanol, pyridine and water (5 :2 :1.5 by vol.) as developer , was applied on acid hydrolyzates. An amount of 10 mg. of polysaccharide was heated in a sealed tube with 5 ml* of N sulfuric acid in a boiling water bath, cooled, neutralized with solid barium carbonate and the inorganic residue filtered. The filtrate and washing were passed through a column (60 x 15 mm., diam.) of Amberlite IBC—50 H+ form) and the effluent collected till neutral, and concentrated to less than 1 ml. Sprays used were aniline hydrogen phthalate (121), alkaline silver nitrate (106), Elson-Morgan 68 reagent (105) and ninhydrin (134). Phthalate was suitable for

(134) J. J. Pratt, Jr. and J. L. Auclair, Science, 108, 213 (1948). reducing sugars, especially aldoses, but was not very sensitive to amino sugars and uronic acids. Alkaline silver nitrate was less selective and alditols also reacted. The latter two sprays were for amino sugars. Elson-Morgan reagent reacted also with N- acetylhexosamines, however, giving purple instead of rose red color characteristic for free amino sugars. All hydrolyzates, except the hexosamines, were separated into distinct zones. The ninhydrin-oxidation technique of Stoffyn and Jeanloz (57) was adopted on the hydrolyzate, selectively reducing the hexosamine to the pentose, which was readily identified by paper chromatography on Whatman No. 1 filter paper, using the 1-butanol, pyridine and water solvent. Heating with a 2% solution of ninhydrin in 4% aqueous pyridine for 30 min. in a sealed capillary tube at 100°, sufficed for the reaction. Complete separation of arabinose and lyxose was obtained on chromatography. Aniline hydrogen phthalate (121) gave red zones for pentoses and brown zones for hexoses. Multiple development of the paper chromatogram was usually resorted to in attempts to decrease the "tailing" of hexosamine zones. Calcium chondroitin sulfate A . Alcohol fractionation of the polysaccharide from calcium acetate buffer according to Meyer and coworkers (l6 ) was applied on a larger scale to the Magnesol- Celite-purified sodium chondroitin sulfate A. A solution of the purified sample from 60 g. of commercial sodium salt in 1,500 ml. of calcium acetate buffer (5% in calcium acetate and 0.5 N in acetic acid) was diluted with ethanol to 20% by vol. of the latter, without any precipitation noted. Only a slight turbidity resulted on increasing the alcohol concentration to 30%, which was removed by centrifugation after keeping overnight at 5°. Since no separation of isomeric chondroitin sulfates was needed, the chondroitin sulfate fraction precipitating between 30 to 50% ethanol was collected, also after standing overnight at 5°* The voluminous precipitate was washed thoroughly with 80% ethanol and redissolved in 1500 ml. of the calcium acetate buffer. (Meanwhile, the supernatant above was processed for the isolation of keratosulfate described below.) This time no precipitation occurred at 30% ethanol. The fraction precipitating between 30 to 50% ethanol was again collected by centrifugation after keeping overnight at 5° and washed well with 80% ethanol, before dissolving in water, dialyzing the solution against running distilled water for 3 days, filtering through asbestos and lyophilizing the concentrated filtrate; yield 42.9 g. (72%) of 22 o white powder, [a} -24.3 (_c 2.18, water). Karl Meyer and coworkers (20) reported -28 to -32° for their preparation. Infrared absorption spectral analysis in the region 700 to 1,000 cm7^ showed identity with that reported for chondroitin sulfate A (53) and that obtained from sodium chondroitin sulfate A above. No galactose was detected on paper chromatographic examination. Keratosulfate isolation and desulfation. The mother liquor (supernatant), from the scale ethanol fractionation (16) of calcium chondroitin sulfate A described above, was processed for the isolation of keratosulfate. The solution was concentrated under reduced pressure below 40° to a thin sirup and dialyzed against running distilled water for 4 days. The dialyzate was passed through asbestos to remove turbidity and the filtrate concentrated into an orange brown thin sirup and lyophilized; yield 1.40 g. of tan brittle solid. Paper chromatographic analysis revealed the presence of chondroitin sulfate A - 22 contamination, also supported by the specific rotation ofQxJ

-10.4° (jc 1.44, water). A refractionation from calcium acetate buffer after Meyer and coworkers (16) was done to further purify the preparation. The impurity precipitating below 50% ethanol 70 was reprecipitated from aqueous ethanol and washed with ethanol and ether and dried over phosphorus pentoxide; yield O .85 g. of light tan powder, with infrared spectrum identical to that of chondroitin sulfate A (53)* The mother liquor from the above fractionation was dialyzed against running distilled water for 4 days, filtered through asbestos, concentrated and lyophilized; yield 0.32 g. of light cream powder, + 0.4° (c^ 0.7» water); reported (20) value, “10 ho +6° (water).

Anal. Calcd. for Cl4H22.28N010^S03Ca0 . 5 0 ^ 0 . 7 7 ^ : S ’ 5*59. Found: S, 5*59* On paper chromatographic analysis, the principal sugars detected were D-galactose and D-glucosamine with traces of D-galactosamine. Infrared spectrum (Fig. 3) showed absence of the characteristic absorption bands for chondroitin sulfate A (Fig. 2) in the region 700 to 1,000 cm.'*" Instead, bands at 998, 820 and 775 cmT were observed, which were similar to those reported for chondroitin sulfate C (53)• An amount of 0.05 g- of keratosulfate preparation was ,subjected to methanolic hydrogen chloride (0.06. M) after Kantor and Schubert (30). The orange residue was washed thoroughly with methanol and dried over phosphorus pentoxide; yield 0,02 g. of light tan powder. Anal. Calcd. for C2.h^22 60^10 ^°3®^0 40 1 S ’ 3*23. Found: S, 3.24. Infrared spectral analysis (Fig. 4) showed no marked change in spectrum except for the appearance of a weak carboxylic ester band at 1, 749 cmT1, indicating methyl ester formation from chondroitin methyl ester contamination. Barium chondroitin sulfate A . Sodium chondroitin sulfate A, purified with Magensol-Celite treatment above from 100 g. of crude material, was dissolved in barium chloride solution and fractionated after Malawista and Schubert (43). A solution of the polymer in 1,500 ml. water made up to pH 5*7 with glacial acetic acid and containing 100 g. of barium chloride was diluted with one-fourth its volume of absolute ethanol and the solution shaken. The small amount of precipitation was centrifuged and 71 discarded. Then, ethanol was again added slowly to the supernatant with vigorous stirring till the solution was 50% by vol. in ethanol and the mixture kept overnight at 5°* The precipitate was recovered by centrifugation and washed with 80?6 ethanol, before redissolving in 1,500 ml. of water containing 60 g. of barium chloride. The fraction precipitating between 20 to 4-0% ethanol was collected this time. A third precipitation was done, again collecting the 20 to 4-0% ethanol fraction. This material was then dissolved in 500 ml. of water, filtered through asbestos, concentrated to a thick sirup and lyophilized; yield 68 g. (53% from commercial source) of white powder. An amount of 0.5 g* of this material was dialyzed against distilled water for 3 days, concentrated and lyophilized and used for analyses, (a] ^ -20.0° (c^ 2 .51 , water); reported (4-3) value, [aj -20.1° (water). Anal. Calcd. for C ^ H ^ ^ g B a ^ ^ N O ^ C S O j B a ^ ^ 5H20>0ig2 : C, 26.19; H, 4.28; N, 2.18; S, 4-.10; COCH^, 6.70; uronic acid, 32.1; ash, as sulfate, 33*10. Found: C,25.02; H, 4-. 91; N, 2.26; S, 4-.25; COCH^(77), 6.43; uronic acid (22), 27; ash, as sulfate, 32.10. Paper chromatographic analysis showed the absence of D-galactose and D-glucosamine, and thus of keratosulfate, originally present in the starting material. Its infrared spectrum was identical to that of sodium chondroitin sulfate A above. An amount of 0.70 g. of barium chondroitin sulfate A in water was passed through a column (30 x 13 mm., diam.) of Dowex 50 ■4* form) and the effluent was carefully neutralized with calcium acetate solution, the slight turbidity removed by filtration through asbestos. The filtrate was concentrated under reduced pressure below 40° and precipitated by pouring with stirring into twice the volume of ethanol. The precipitate was washed with ethanol and ether, and dried over phosphorus pentoxide to produce calcium chondroitin sulfate,[a]^ -19.2° Cc 2.08, water). The calcium chondroitin sulfate A, prepared by the ethanol fractionation from calcium buffer above, had [a]22D -2^.3° (c 2 .18, water). 72 Potassium chondroitin sulfate A, Sodium chondroitin sulfate A, purified with Magnesol-Celite from 50 g. of crude material, was dissolved in 800 ml. of water made acidic (pH 5*7) by the addition of acetic acid, similar to the procedure for barium salt (43), only changing the cation involved. An amount of 50 g. of potassium chloride was dissolved in the solution and ethanol (200 ml.) was added, also with stirring. A small amount of residue which formed was centrifuged and discarded. More ethanol was added to the supernatant with stirring until the ethanol concentration was 50% by vol. On standing overnight at 5°i the precipitated fraction was collected by centrifugation and redissolved in 500 ml, of water containing 30 g. of potassium chloride. The fraction precipitating between 20 to 4-5% ethanol was collected as above. This fraction was dissolved in 500 ml. of water, filtered through asbestos and lyophilized; yield 34.9 g. (66%) of white potassium chondroitin sulfate -19.7° (c^ 2 .3 6 , water), identical in infrared spectrum to sodium chondroitin sulfate A above. Kantor and Schubert (30) cite -25*0° (water) for this substance. Paper chromatographic identification of the component sugars of this potassium chondroitin sulfate A preparation revealed only glucuronic acid and galactosamine, and no galactose nor glucosamine present, thus showing absence of any keratosulfate contamination. Carboxyl-reduced Chondroitin 90% Reduced Polymeric Chondroitin Chondroitin methyl ester. Chondroitin methyl ester was derived from potassium chondroitin sulfate A according to Kantor and Schubert (30)* Potassium chondroitin sulfate A was obtained from the Magnesol-Celite treated sodium 6alt by cation exchange; yield 78.6 g. (74% from commercial chondroitin sulfate) of the potassium salt, fa ] ^ -22.8° (_c 2.32, water). A stock solution of 0.06 N methanolic hydrogen chloride was prepared by carefully adding (hood) 5 ml. of acetyl chloride to every liter of absolute methanol and stood for a day before use. Dried powdered potassium 73 chondroitin sulfate A (10.0 g.) was shaken with 1,500 ml. of 0.06 N methanolic hydrogen chloride for 24 hr. The acidic alcohol was replaced daily for a total reaction time of 3 days. The suspension collected by centrifugation was washed well with methanol, air-dried, dissolved in 400 ml. of water, and dialyzed against running distilled water for 3 days. Insoluble material was removed by filtration and the filtrate was concentrated under reduced pressure and lyophilized; yield 4.1 g. (56%) of white fluffy ester, (a] 2** -14.0° (£ 1.2 8 , water); reported (30) value, -Ik. 6° (water). Anal. Calcd. for C ^ E ^ q N0^(.C0^CE^) : S, none; OCH^, 7.89; uronic acid, 49.4. Found: S, absent; OCH^, 7*05; uronic acid (22), 39*0. Infrared absorption spectrum (Fig. 6 ) showed — 1 characteristic ester absorption at 1,740 cm. and the absence of strong sulfate bands at 1,250 cmT'1' and in the region 700 to 1,000 cm7^ A pronounced peak at 894 cm.'*' was noted. Sodium borohydride reductions. The procedure of Wolfrom and Anno (32) as employed by Frush and Isbell (for aldonate esters) (33) was utilized. Chondroitin methyl ester (15.22 g.) prepared above in 0,05 M boric acid solution (430 ml.) was cooled to 0 . Dowex 50 (H form, 90 ml.) was then added and with stirring, the mixture was treated dropwise with 250 ml. of a freshly prepared solution of 0,5 H sodium borohydride for 45 min. The reaction mixture was stirred for another hour before another lot of reducing agent was added again dropwise, with the stirring continued for 1 hr. more. Then dilute sodium hydroxide solution was added to adjust the pH to 9 and the solution was kept overnight at 5 °» The solution was rid of the resin, neutralized with acetic acid, dialyzed against running distilled water for 3 days, filtered through washed asbestos, concentrated under reduced pressure to a milky sirup and lyophilized; yield 12.30 g. of white powder. Anal. Calcd. for C ^ H ^ N O ^ C H ^ H ) ^ (CO^a)^ uronic acid, 17.6. Found: uronic acid (22), 10.7 . 7^ This partially reduced chondroitin was esterified with methanolic hydrogen chloride as above. Attempts to use ethereal diazomethane on a methanolic suspension of partially reduced chondroitin were unsuccessful; only trace amounts of OCH-, were —1 detected, consistent with the absence of the 1,740 era. ester band in the infrared spectrum. However, some loss of material was noted. The methanolic hydrogen chloride treated material was washed with methanol, air dried and directly subjected to sodium borohydride reduction, using borate buffer after Frush and Isbell (33)* The material was dissolved in 0.4 M boric acid (100 ml.) at 0° and to the cool solution was added dropwise, with stirring, 200 ml. of a freshly prepared 0.3 M sodium borohydride solution over a period of 30 min. Stirring was continued for 30 min. more and the solution was made alkaline (pH 8.5) with dilute sodium hydroxide solution before storing overnight at 5°» it was then neutralized with formic acid and dialyzed against running distilled water for 3 days. The dialyzate was filtered through asbestos, concentrated to a thick sirup and lyophilized; yield 11.0 g. of sparingly water-soluble white powder. Anal. Calcd. for C - ^ H ^ N O ^ C H ^ H ^ g,_(C02Na)0 uronic acid, 7 *85. Found: uronic acid (22), 5*5. A final reduction sequence was tried on this reduced material. This was esterified with 400 ml. of 0.06 N methanolic hydrogen chloride for 3 days, the solvent being replaced every 1.5 days. Reduction was with sodium borohydride-borate buffer and the product was isolated as described above. A white water-insoluble powder was obtained on lyophilizing; yield 8.5 g. (5^% from chondroitin methyl ester), +5*7° (c, 0.44, dimethyl sulfoxide). Anal. Calcd. for C^H^NO^CH^H)^^(CO^a)^ 1Q’ 2H20: C, 41.55; H, 6.15; N, 3*46; COCH^, 10.62; ash, as oxide, 0 .76; uronic acid, 4.80, Found: C, 40.97; H, 7.09; N, 3*37? COCH^ 75

(77)* 10.5? ash, as oxide, 0.39; uronic acid (22), 3»5* Paper chromatography of the acid hydrolyzate of this reduced material showed glucose and no galactose, although the starting material was contaminated with traces of keratosulfate. A modified hexuronic acid assay of Dische (22) was employed for the partially reduced chondroitin, since hexose (glucose) interfered with the colorimetric reaction. From a series of determinations with D-glucose and D-glucuronic acid solutions of various concentrations and ratios, a correction factor was calculated which fairly accounted for the presence of glucose. Reproducible results were obtained even with a glucose to glucuronic acid ratio of 24 to 1. The presence of acetylated alditols did not interfere with the assay. Colorimetry involved the use of a Klett-Summerson photoelectric colorimeter with Filter 54. 96% Reduced Polymeric Chondroitin Chondroitin methyl ester. Potassium chondroitin sulfate A was treated with acidic methanol after Kantor and Schubert (30) to prepare chondroitin methyl ester. Potassium chondroitin sulfate, QxJ -19.7° (c 2 .36, water), was prepared by the alcohol fractionation of Magnesol-Celite treated sodium chondroitin sulfate A in potassium chloride solution. Dried powdered potassium chondroitin sulfate A (lO.OOg.) was treated for 3 days with 1600 ml. of 0.06 M methanolic hydrogen chloride, derived by passing dry hydrogen chloride through absolute methanol and diluting to the desired concentration. The collected suspension was washed twice with methanol and stored overnight over phophorus pentoxide; yield 6.0 g. (82%) of white ester, -15.2° (jc 1.21, water); reported (30) value, [aj^p -14.6° (water). Anal. Calcd. for C ^ H „ N 0 o (C0_CH,): S, absent; OCH,, 7*89; 13 20 9 2 5 3 uronic acid, 49.4. Found: S, none; OCH^, 7-64; uronic acid (22) 40.1. 76 Sodium borohydride reductions. Reduction of the chondroitin methyl ester followed the general procedure of Frush and Isbell (33) in borate buffer. An amount of 17.6 g. of chondroitin methyl ester in 300 ml. of 0.4 M boric acid solution was treated dropwise at 0°, under stirring, with a fresh solution of 6.4 g. of sodium borohydride in 500 ml. of water during a period of50 min. Stirring was continued 45 min. more and processed as above, except that dialysis was replaced with ethanol precipitation and washing several times with ethanol. The washed powder was dried over phosphorus pentoxide. Anal. Calcd. for C ^ H ^ N O ^ (CH^0H)„ (C0„Na)„ : uronic 13 20 9 2 0.66 2 0.34 acid, 17.5. Found: uronic acid (22), 13.7» This partially reduced chondroitin was esterified with methanolic hydrogen chloride and reduced with sodium borohydride in borate buffer as above, processed and rid of inorganic impurities by dialysis; yield 14.3 g. of white powder, sparingly soluble in water.

Anal. Calcd. for C ^ H ^ N O ^ (CH^OIOq gg (CO^a)^ uronic acid, 7.34. Found. uronic acid (22), 5*6. A third esterification and reduction was made on this material and the product processed and dialyzed, concentrated to a thick white paste and lyophilized; yield 13.04 g. (71% from chondroitin methyl ester) of very sparingly water-soluble powder, +11° (jc 0.46, dimethyl sulfoxide). Anal. Calcd. for C-^H^NOg (CH20H)Q gg(C02Na)0 ^<21^0: C, 41.74; H, 6.73; N, 3.48; COCH^, 10.69; ash, as oxide, O.3 8 ; uronic acid, 1.93* Found: C,41.75; H, 6 .65; N, 3*77; COCH, 3 (77) » 11.2; ash, as oxide, 0.60; uronic acid (22), 1.3* Paper chromatographic analysis confirmed the absence of D-galactose and D-glucosamine and the presence of D-glucose and D- galactosamine. Infrared absorption spectral examination (Fig.7 ) of the reduced chondroitin showed the absence of sulfate absorption at 1,240, 928, 852 and 725 cm.^ and also of carboxylate absorption at 1,612 cm.^ Compared with sodijtm 77 chondroitin sulfate A, sharper peaks at 2,930? 1,570, 1 ,380, -1 -1 783 and 704 cm. and the appearance of a new band at 889 cm. were noted. Acid hydrolytic studies. Ah amount of 100 mg. of 96% reduced chondroitin was refluxed in a 230 ml. two-necked round-bottomed flask with 43 ml. of N sulfuric acid. The sparingly water-soluble polymer dissolved in the medium within 1 hr. and samples (4 ml.) were withdrawn from the colorless clear solution intermittently, starting immediately after the dissolution of the substrate. The hydrolysis was followed polarimetricaliy and paper chromatographically. Readings were made immediately on the samples on cooling. The solution was then neutralized with solid barium carbonate, the inorganic residue centrifuged and the supernatant treated with Amberlite IRC-30 (H+form, 8 ml.). The deionized samples were concentrated under reduced pressure into thin sirups for paper chromatography. Multiple development of the samples on Whatman No. 1 filter paper with 1-butanol, pyridine and water (3:2:1.'3 by vol.) solvent was employed. Constant polarimetric reading was noted within 3 hr. of refluxing. Paper chromatographic analysis showed, aside from glucose, galactosamine, glucuronic acid, chondrosine, a distinct zone of Blson-Morgan (49) positive material, later identified as carboxyl reduced chondrosine, which was present in all samples except after 13 hr. hydrolysis, where only traces of it were detectable. The disaccharide moved faster than chondrosine but slower than galactosamine (Table 6). N-Acetyl derivatives, which gave purple spots with the Elson-Morgan reagent (105), were present within the first 2 hr. of hydrolysis. N- Acetylhexosamine derivatives moved much faster on paper than the free hexosamines. Of the three solvent systems used, the basic solvent (1-butanol, pyridine and water, 3*2:1.5 by vol.) gave the best separation with the least overlapping of component substances (Table 6). 78

Table 6

^glucose ^a-*-ues Component Sugars in Three Solvent Systems

H: t Value — glucose

1-Butanol: Ethyl acetate: 1-Butanol: pyridine: acetic acid: ethanol: Sugar water formic acid: water (3:2:1.5) water (18:3: (40:11:19) 1:4) Chondrosine 0.06 O.32 0.13 N-Acetylchondrosine 0.20 0.?4 0.16 D-Glucuronic acid 0.25 1.0 0.88 Carboxyl-reduced chondrosine 0.32 0.37 0.38 D-Galactosamine hydrochloride 0.56 0.64 0.71 Carboxyl-reduced N-acetylchondrosine 0.7^ 0.?4 0.70 N-Acetyl-D-galactosamine 1.2 1.7 1.3

— Average value — 0.08 at room temperature. R^ values for D-glucose in the three solvents were 0 .32, 0.18 and 0.24 in the order above. Solvent composition in parts by vol. 79 Carboxyl-reduced Chondrosine Fractionation of hydrolyzate. An amount of 4.00 g. of 90% carboxyl-reduced chondroitin was refluxed for 2.25 hr. in 125 of N sulfuric acid and the cooled hydrolyzate was neutralized with solid barium carbonate. After the filtration of the inorganic residue, the filtrate was made acidic with 10 ml. of N hydrochloric acid prior to N-acetylation, after the general procedure of Hoseman and Ludowieg (114). A solution of the concentrated yellow hydrolyzate in water (75 ml.) was treated at 0° with 7*5 ml. of methanol, 90 ml. of Dowex 1 (carbonate form) and 2 ml. of acetic anhydride, and stirred for 90 min. at 0-5°. The reaction mixture was filtered and the filtrate and washings passed through a column (180 x 12 mm., diam.) of Dowex 50 (H+ form) to remove free uronic acid and amino sugar. Paper chromatographic analysis at this stage showed the absence of glucuronic acid and N-acetylchondrosine and the presence of distinct zones for N-acetyl-D-galactosamine (R glucose D-glucose and carboxyl-reduced N-acetylchondrosine (R , 0.74). glucose Fractionation of the N-acetylated hydrolyzate followed the general procedure of Whistler and Durso (115). A carbon (Nuchar C unground) column (210 x 44 mm,, diam.) was prepared by pouring a slurry of carbon into the tube under reduced pressure and the column was well packed and washed with 2 liters of water. The sample was placed on the column and developed with water (9 liters), 2% ethanol (4.5 liters), 3% ethanol (1.8 liters), 3% ethanol (3.2 liters) and 6% ethanol (4 liters). Paper chromatographic analysis showed that D-glucose was present pure in the first liter of the water effluent, but was mixed with N- acetyl-D-galactosamine in the rest of the water effluent; pure N-acetyl-D-galactosamine was eluted with 2% ethanol, but mixed with traces of carboxyl-reduced N-acetylchondrosine in the 3% ethanol eluate; and pure carboxyl-reduced N-acetylchondrosine was recovered from the 3% and 6% ethanol fractions. The fractions corresponding to pure substance were concentrated to a small 8o volume, filtered through a fhitted glass filter and deionized by­ passing through a column (60 x 13 mm., diam.) of mixed bed resin (Amberlite MB-3)» Effluents were then concentrated to dryness under reduced pressure. Attempts to isolate free carboxyl-reduced chondrosine as the neutral sulfate salt, both by cation exchange column chromatography and by filter paper chromatography, failed. In the slightly alkaline and acid media, carboxyl-reduced chondrosine was eluted with traces of its component sugars, glucose and galactosamine (salt). Subjecting this crude preparation to carbon column chromatography (115) did not separate the disaccharide from its component sugars. With filter paper chromatography, all zones were contaminated with traces of glucose, which may have been extracted from the cellulosic material. The solvent used was 1-butanol, pyridine and water (3:2:1.5 by vol.) on Whatman No. 3 filter paper. The sample was eluted, on the other hand, from the cation exchange column (Dowex 50, H+ form) with increasing concentrations of acetic acid. Fractions were analyzed by paper chromatography. (3-D-Glucose pentaacetate. The glucose fraction (first liter of water effluent) was acetylated with acetic anhydride-sodium acetate. The sirup (0.*+l g.) was heated with 1 g. of anhydrous sodium acetate and 7 ml. of acetic anhydride until the reaction proceeded spontaneously. The reaction mixture became homogeneous and was reboiled twice, cooled slightly and poured into 5 times as much ice and water and stirred for 5 hr. at room temperature. The solution was extracted with b 15 ml. portions of chloroform and the combined extracts were evaporated to dryness. The residue was dissolved in anhydrous diethyl ether, filtered through a fritted glass filter and crystallized by the addition of petroleum ether (b.p. 30-60°) till incipient cloudiness; yield 0.31 g. (35%) of white microneedles, m.p. 133*5-13^°, Ca3 +^° Ql chloroform), x-ray powder diffraction data identical to authentic 81

sample: 12.4 (135) m (136), 9 .38vs (1 ), 5 -6ls (3), 5 .21vw, 4.91m,

(135) Interplanar spacing, $, CuK radiation. oc (136) Relative intensity, estimated visually; s, strong; m, medium; w, weak; v, very. First three strongest lines are numbered (1 , strongest); double numbers indicate approximately equal intensities.

4.66w, 4,47s (2 ), 4.30m, 3 »75vw, 3 *53m, 3*4lvw, 3 *25vw, 3 *10vw,

2 .56vw, 2 .44 v w , 2 .35 v w , 2 .20 v w , 2 .12vw, 1 .82v w .

Anal. Calcd. for ^i 5^22^11! ^ ’ ^9*23; 5*68. Found: C, 49.22; H, 5*78. Infrared spectral analysis (Fig. 11) showed a band at 906 and a shoulder at 888 cm. indicative of [3-D anomeric configuration (104). TJ-Acetyl-a-D-galactosamine (2-acetamido-2-deoxy-a-D-galactose) monohydrate. The deionized 2% ethanol fraction yielded a dry sirup (0,24 g.) which was crystallized by the addition of a small amount of absolute ethanol and standing for a few hours at 5°* Recrystallization was effected in the same manner yield 0.18 g. of white crystals; m.p. 118-120° (with preliminary softening), C«] 25d +84.4° (_c 1.04, water, final downward mutorotation), x-ray powder diffraction data: 10.5 (135) s (136) (1 ,1 ), 7 »80m, 7 *l8w, 5 *l6m (2 ), 4.64m, 4.40m, 4.19s (1 ,1 ), 3 *90w, 3 *63vw, 3 *19vw, 2 .17vw. Stacey (117) cites m.p. 120-122°, [aj +80° (water, final) for this substance. Anal. Calcd. for CgH^NOg'HgO: C, 4o.l6; H, 7-16; N, 5 .86. Found: C, 40.30; H, 7-37? N, 5*81. Infrared spectral analysis (Fig. 9) showed strong peaks at 832, 8l8 and 780 cm. but one at — 1 887 cm. , indicative of a predominantly a-D-configuration (104). (3-D-Galactosamine pentaacetate (2-acetamido-tetra^0-acetyl- 2-deoxy-P-D-galactopyranose). The procedure was essentially that of Stacey (117). An amount of 90 mg. of N-acetyl-D-galactosamine monohydrate v/as suspended in 1.2 ml. of acetic anhydride and shaken with powdered fused zinc chloride for 24 hr. The reaction 82 mixture was poured into 4 vol. of ice-water and the suspension was carefully neutralized with solid sodium carbonate. The mixture was made slightly alkaline with dilute sodium hydroxide and extracted six times with chloroform (10-ml. portions). The combined extracts were dried overnight over sodium sulfate, filtered and the filtrate and washings concentrated under reduced pressure until crystallization commenced. Then the solution was diluted with ethanol and kept at 5°; yield 40 mg. (37%) of white crystals, m.p. 235° i L°0 +8° (£ 0.4, chloroform), x-ray powder diffraction data: 8.10 (135) w (136),

7.44s (2,2), 6.28m, 5.10VW, 4.00s (1), 3«84w, 3 *55 v w , 3 .31s (2,2), 3*03vw, 2.34m. Stacey (117) cites m.p. 235°, +7° (c_ 0.6, chloroform) for this compound. Anal. Calcd. for G > ^9-35; H, 5*95; N, 3.60. Found: C, 49.25; H, 5.85; N, 3*70. Bands at 895 and 865 cm. were noted in the infrared spectrum (Fig, 10), characteristic for the p-D-configuration (104). Carboxyl reduced N-acetylchondrosine (()-p-D-glucopyranosyl- (1— 3)-‘2-acetamido-2-deoxy-a-D-galactose) dihydgdte . The dry sirup (1.0 g.) from deionized 5% and 6% ethanol eluates was crystallized by the addition of a small volume of absolute ethanol and keeping for a few hours at 5°. Recrystallization was effected in the same manner; yield 0.40 g. (10% from the polymer) of white microneedles, m.p. 155-157° (with preliminary ^ ^ Q softening), (a] + 47 (extrapolated)— > +19 (c 1.07, water), x-ray powder diffraction data: 14.3 (135) w (136), I3«0vs (1), 10.4s (2,2), 7 .36vw, 6 .56vw, 4.91vw, 4.59s, 4 .36m, 4.l4s (2,2),

4.01m, 3 .88m, 3 ,60m, 3«33w, 2.89‘W, 2.79vw, 2.5IVW, 2. 28 v w ,

2 .17vvw, 2.l4vvw, 1 .90vw, 1 .87v v w . Anal. Calcd. for C ^ H ^ N O ^ * 2H20: C, 40.09; H, 6-97; N, 3*34; H20, 8.59. Found: C, 40.09; H, 7.09; N, 3.27; H20, 8 .3 2 ; positive Morgan-Elson (ll6) test. Its infrared spectrum (Fig. 8) was very similar to that of carboxyl-reduced chondroitin except for the presence of an 807 cm.'*' peak absent in the latter. Paper chromatographic analysis of the acid hydrolyzate of this substance showed only D-glucose and D-galactosamine. To determine the alkaline lability of the disaccharide, an amount of 5 mg. of carboxyl-reduced N-acetylchondrosine was treated with 1 ml. of 0.04 N sodium carbonate for 2 hr. at room temperature and the mixture was analyzed paper chromatographically with the 1-butanol, pyridine and water mixture as solvent. Aside from the starting material, D-glucose and a Morgan-Elson (116) reactive sugar ^ g p ucose 1-SJ, "anhydro-N-acetyl-D-galactosamine," different from N-acetyl-D-galactosamine (R n 1.2), were — glucose ’ detected. Carboxyl-reduced N-acetylchondrosinol (C)-3-D-glucopyranosyl - 2-ace tamido-2-deoxy-D-galactitol) . Carboxyl reduced N- acetylchondrosine was reduced to the alditol with sodium borohydride after the general procedure of Davidson and Meyer (50). An amount of 100 mg. of carboxyl-reduced N-acetylchondrosine was dissolved in 5 ml. of cold 5G?£ methanol and was added in portions, with occasional stirring, to a solution of 40 mg. of sodium borohydride in 5 ml. of borate buffer (0.1 M, pH 8) at 0°. The mixture was stirred at 0° for 2 hr., then for an additional hour at room temperature and was acidified to pH 5 with acetic acid and passed through a column (100 x 13 mm., diam.) of mixed bed resin (Amberlite MB-3). Traces of boric acid were removed by evaporation of the volatile methyl borate. Paper chromatographic analysis using the 1-butanol, pyridine and water solvent mixture showed carboxyl-reduced N-acetylchondrosinol (R n 0.76) as the main product, unreactive to aniline glucose hydrogen phthalate (121) and reactive to Morgan-Elson (49) spray (105). Traces of D-glucitol f®gQ_ucose 1*0) and Manhydro-N-acetyl- D-galactosaminol,t (R _ 1.8) were also noted, the latter glucose being Morgan-Elson reactive. Purification of the carboxyl-reduced N-acetylchondrosinol was facilitated by carbon (115) (Nuchar C unground) column 84 (130 x 34 mm., diam.) chromatography, collecting water (1.0 liter), 2% ethanol (0.45 liter) and 20% ethanol (1.1 liters) eluates. The 20% ethanol fraction was concentrated under reduced pressure, filtered through a fine fritted-glass filter, deionized by passing through a column (100 x 13mm., diam.) of Amberlite MB-3 mixed bed resin and evaporated to dryness. Paper chromatographic analysis showed trace contamination of the same impurities to the carboxyl-reduced N-acetylchondrosinol. Attempts to crystallize the hygroscopic glass failed. On lyophilizing, 0.06 g. of white powder was obtained. Carboxyl-reduced N-acetylchondrosinol was unaffected by treatment with 0.04 N sodium carbonate for 2 hr. at room temperature; no glucose was detected on paper chromatography of the reaction mixture. To correlate the disaccharide from carboxyl-reduced chondroitin to that derived from chondrosine by Davidson and Meyer(50), crude carboxyl-reduced N-acetylchondrosinol was prepared. An amount of 0.20 g. of chondrosine was esterified to the methyl ester hydrochloride by the action of 0.03 N methanolic hydrogen chloride (8 ml.) at 5° for 4 days (30)* The clear solution was evaporated to dryness, washed several times with absolute methanol and crystallized from methanol; yield 0.10 g. of crystalline chondrosine methyl ester hydrochloride, m.p. 155-157 » (“J -q (c_ 2, methanol). This ester was reduced to carboxyl-reduced chondrosinol, according to the general procedure of Frush and Isbell (33) i in borate buffer. The ester (0.10 g.) was dissolved in 10 ml. of 0.4 M boric acid at 0°, to which 15 ml. of 0.6 M sodium borohydride solution was added dropwise, at 0°, over a period of 30 min. with stirring, stirred 1 hr. more at 0° before warming to room temperature for 2 hr. The reaction mixture was then acidified to pH 5 with acetic acid, passed through a column (90 x 13 mm., diam.) of Amberlite IRC-50 (H+ form) and evaporated to dryness. Repeated washing with methanol removed the boric acid, but the 85 product was converted directly to the N-acetyl derivative, without characterizing, due to inorganic contamination. The general procedure of Roseman and Ludowieg (114) was followed. The alditol above in 15 ml. of water and 1.5 ml. of methanol was stirred for 90 min, at 0.5° with 20 ml. of Dowex 1 (carbonate form) and 0.30 ml. of acetic anhydride. Filtration of the solution and passing through a column of mixed bed resin (Amberlite MB-3) gave a clear effluent which on concentration to dryness yielded a glass (0.06 g.). Paper chromatographic analysis showed a Morgan-Elson (ll6) reactive non-reducing zone as the major component . (■^a.]_ucose O.76), similar, in three solvent systems, to that for carboxyl-reduced N_-acetylchondrosinol prepared above. Impurities also present were "anhydro-N-acetyl-D-galactosaminolM and D-glucitol, which prevented further characterization of the product. 85% Reduced, Degraded Chondroitin Preparation of chondroitin acetate. Magnesol-Celite treated sodium chondroitin sulfate A was converted to degraded chondroitin acetate by the acetylative-desulfation procedure of iVolfrom and Montgomery (55)> modified such that 3-0 g. insteat of 2.0 g. of the polysaccharide was used with each run. This involved the reaction of sodium chondroitin sulfate with a mixture of absolute sulfuric acid (m.p. 5°) and acetic anhydride at subzero temperature; yield 1.3 to 2.0 g. of light tan powder; total yield (from five replications) 9.10 g. (48%) , [a]^ +18° to +31° (.£, l*9i chloroform); reported (125) values, +17.8° (chloroform). Anal. Calcd. for Cn_H_,NNa01^(COCH,)c : N, 2.46; S, none; — — 12 13 10 3 5 COCH^, 37*8; uronic acid, 34.1. Found: N, 1.98; S,<0.03; C0CH„, 42.0; uronic acid (22), 21.8. Infrared spectral analysis (Fig. 5) showed ester absorption at the regions 1,740-1,755 cm. and 1,220-1,240 cm7^ and the disappearance of the characteristic sulfate bands in the region 700 to 1,000 cm.'*’ 86 Degraded chondroitin methyl ester. The chondroitin acetate (9.10 g.) prepared by the sulfuric acid-acetic anhydride method above was dissolved in ^00 ml. of absolute methanol saturated with anhydrous ammonia at 0° to give a brownish orange solution. The solution was maintained at room temperature for 6 hr. with occasional stirring and the resultant turbid solution was concentrated under reduced pressure at J>0°. The brown sirupy residue was washed twice with 50 ml. of absolute methanol and the solvent was evaporated under reduced pressure. Acetamide was removed by sublimation under reduced pressure at room temperature. This degraded chondroitin was passed through a column of cation- exchange resin (Dowex 50, H+ form, 220 x 50 mm., diam.) and the turbid light-yellow effluent was concentrated under reduced pressure at 30° to a thin sirup and finally lyophilized; yield 6.2 g. (.98%) of a tan glass. Esterification of the above dried chondroitin was facilitated by the use of an ethereal solution of diazomethane, prepared from "Diazald" (N-methyl-N-nitroso-p-toluenesulfonamide, available from Aldrich Chemical Co., Milwaukee, Wise.) according to De Boer and Backer (126). An amount of 8.3 ml. of 95% ethanol was added to a solution of 1.7 g. of potassium hydroxide in 2.7 ml* of water in a 125 ml. distilling flask fitted with a dropping funnel and an inclined Liebig condenser, the adapter of which was connected tightly in series to two receivers, such that the inlet of the second receiver from the first dipped into the 15 ml. of ether in it. This prevented the escape of any diazomethane vapors from the set-up. Both receivers were maintained at 0° in an ice-water bath. The flask was then heated to and kept at 65° and a solution of 7*2 g. of "Diazald11 in about ^3 ml. of diethyl ether was added dropwise through the dropping funnel in about 25 min. or such that the rate of distillation should about equal the rate of addition. The dropping funnel was then washed with 7 ml. of ether and distillation stopped when the distillate was no longer yellow. 87 The distillate plus the ether in the second receiver contained about 1 g. of diazomethane, which was immediately used. (Ground glass joints and rough glass surfaces should not get in contact with diazomethane as a precautionary measure.) To the dried chondroitin dissolved in absolute methanol (1,000 ml.) at 0° was added dropwise the ethereal solution of diazomethane with vigorous stirring, causing complete precipitation of the degraded polymer. The reaction mixture was then placed at room temperature for ^ hr. during which time the characteristic yellow color of diazomethane disappeared. The ether-methanol was removed by evaporation under reduced pressure below ^0°, and the orange- brown residue was stored over phosphorus pentoxide; yield 6. A- g. (99%) of a very hygroscopic methyl ester. Sodium borohydride reductions. Reduction with sodium borohydride followed essentially the modification of Frush and Isbell (33) of the procedure of Wolfrom and Anno (32). Chondroitin methyl ester (6.3 ) was dissolved in 130 ml. of water made 0.03 M in boric acid and to which 33 ml. of Dowex 50(H ) had been added, and maintained at 0°. With efficient stirring, 150 ml. of a freshly prepared 0.3 M sodium borohydride solution was added dropwise in about 35 min. Stirring was continued for an additional 30 min. after which a second 150 ml. of reducing solution was added as above. After stirring for an additional 30 min. after all the reducing agent had been added, the pH of the reaction solution was adjusted to about 9 with aqueous sodium hydroxide solution (seldom required), and this basic solution was cooled overnight at 5°* The turbid light yellow solution was passed through a column (220 x 25 mm., diam.) of Dowex 50 (H+ form) and the turbid light yellow effluent was collected until the washings were neutral. The effluent was filtered through asbestos to remove the turbidity, and the cream-colored solution was concentrated under reduced pressure at 35° to a thin sirup, or until the boric acid started crystallizing out of solution. Then the solution was diluted with methanol until all the boric acid was dissolved and was 88 concentrated to a thick sirup. Several washings 'with methanol were required to completely remove the volatile methyl borate, after which the residual sirup was dissolved in the least amount of water and lyophilized; yield 5*00 g. of light brown hygroscopic material. The partially reduced material was reesterified in methanolic solution with ethereal diazomethane (1 g.); yield 5.0 g., [o J28d -4. 7° (c 1.5, water).

Anal. Calcd. for C ^ H ^ N O g (CH^OH^ ( C O ^ K ^ q ?9= OCH^, 3*44; uronic acid, 20.5. Found: OCH^, 3*44; uronic acid (22), 1 1 .8 . This partially reduced chondroitin methyl ester (5»0 g.) was again reduced with sodium borohydride in the presence of cation exchange resin and boric acid as above. Anal. Calcd. for C,-.H^NO^ (CH^0H)rt nr (CO^H)^ „, : uronic 13 20 9 2 0.7b 2 0.24 acid, 12.6. Found: uronic acid (22), 7*24. A third diazomethane esterification and subsequent sodium borohydride reduction was attempted using the same procedure; yield 4.0 g. (67% from the starting chondroitin methyl ester) of light tan, brittle, very hygroscopic glass, -3.1° (_c 0 .82, water). Anal. Calcd. for C ^ H ^ N O g (CH20H)Q g,_ (CO^^ ^ : C, 45.35; H, 6.13; N, 3 .78; uronic acid, 7*92. Found: C, 42.75; H, 7*24; N, 8 .71; uronic acid (22) 4.57. Paper chromatographic analysis showed the presence of glucose with a trace of galactose. Trial paper chromatographic characterization of the 85% reduced degraded chondroitin was made to determine its extent of degradation. No spots were detected with aniline hydrogen phthalate spray (121), indicating the absence of reducing sugars and the complete reduction of terminal carbonyl functions. Alkaline silver nitrate (106) revealed three fast moving zones, one with mobility comparable to that of D-glucitol (sorbitol), whereas the other faster moving zones were unidentified. 89 Acid hydrolytic studies. An amount of 96 mg. of degraded 85% reduced chondroitin was dissolved in 45 ml. of N sulfuric acid and refluxed in a two-necked flask. Samples (5 ml.) were pipetted out intermittently and analyzed polarimetrically and paper chromatographically. Readings were made immediately on cooling and the samples were neutralized with solid barium carbonate and the supernatant was filtered and concentrated under reduced pressure to sirups for chromatography. The specific rotation was essentially constant in 2 hr., with a final reading of +50°. Paper chromatographic analysis using the 1-butanol, pyridine and water (3:2:1.5 by vol.) solvent mixture and Elson-Morgan (49) reagent (105) and alkaline silver nitrate (106) showed no defined zone for carboxyl-reduced chondrosine, although spots corresponding to glucose, galactosamine, chondrosine and traces of galactose were distinctly resolved. The degraded state of the substrate may explain these observations in contrast to the polymeric material above from which carboxyl-reduced chondrosine was detected as a distinct zone. Keratosulfate contamination complicated the analysis further. JJ-Deacetylation Studies N-Deacetylation of Barium Chondroitin Sulfate A Hydrazinolysis. Barium chondroitin sulfate A, prepared by alcohol fractionation from barium chloride solution of Magnesol- Celite treated sodium salt, was treated with hydrazine by a modification of the procedure of Matsushima and Fujii (73)* An amount of 2.28 g. of barium chondroitin sulfate A, [a"] ^ -20° (water), was heated with 11 ml. (11.1 g.) of anhydrous hydrazine (1:100) in a sealed tube for 10 hr. in a boiling water bath. The turbid reaction mixture was concentrated to dryness under reduced pressure to remove the excess hydrazine, dissolved in 30 ml. of water and dialyzed against running distilled water for 3 days. The inorganic residue was then separated by filtration through 90 asbestos and the filtrate was passed through a column (100 x 13 mm., diam.) of Dowex 50 (H+ form). Effluent and washings were carefully neutralized with dilute sodium hydroxide, concentrated under reduced pressure and lyophilized; yield 0.66 g. (42%, 2^ triplicate) of light cream powder, f o r P -13.7° _(_c 1.16, water). Anal. Calcd. for 012Hl8>02H1>60Ma0>7009#70(C0CH3 )0 o 2 (S0_Na) N, 5.24; S, 4.20; COCH,, 3-22. Found: N, 5.23; 3 O.35 3 S, 4,22; COCH^ (77)* 3.22; positive ninhydrin test. Infrared spectrum (Fig. 14) of the hydrazine-treated material was very similar to that of barium chondroitin sulfate A, the starting substance. Scaling down the amount of reactants to half gave a light cream product; yield O .33 g. (44%, duplicate). Anal. Calcd. for .^N Na 0^Q (COCH^)^ ^gCSO^Na)^ ^ : S, 3.57* COCH^, 3.67. Found: S, 3.59; COCH^ (77), 3.68; positive ninhydrin test. Scaling up the amount of reactants three-fold gave a slightly cream-colored powder; yield 2.57 g. (66%, triplicate), §t]22D -13.9° (£ 1.19, water). Anal. Calcd. for 0iNi. 72NaO. 64°9.64 ^GOCH3^0.38 (S0_Wa)_ N, 5*44; S, 5*00; COCH,, 3*70; ash, as sulfate, 3 O . 0 9 3 21.33, uronic acid, 44.2. Found: N, 5*46; S, 5*07; COCH, (77), 3 3 .72; ash, as sulfate, 14.31; uronic acid (22), 40; positive ninhydrin test. Hydrolytic studies on N-deacetylated chondroitin sulfate A. An amount of 100 mg. of 62% N-deacetylated sodium chondroitin sulfate A in 45 ml. of N sulfuric acid was hydrolyzed by refluxing and the hydrolysis followed polarimetrically and paper chromatographically for a 24 hr. period from samples drawn from the reaction mixture at appropriate intervals. Readings were made on the samples immediately upon cooling and the solutions neutralized with barium carbonate, and the supernatant passed through a column (60 x 13 mm., diam.) of Amberlite IRC-50 (H+ form). The effluent and washings were concentrated to a sirup 91 for chromatography on Whatman No. 1 filter paper, using 1-butanol, pyridine and water (3:2:1.5 by vol.) developer. Only chondrosine was detected as a distinct zone for disaccharide, using Elson- Morgan spray (105) and alkaline silver nitrate (106) reagent. D-Glucuronic acid was detectable in the hydrolyzate^ but was absent after a 2b hr. reaction. Both the free acid and lactone forms were present. tl-Acetyl derivatives of chondrosine and galactosamine were present during the first hr. of the reaction. The specific rotation increased steadily but slowly for 13 hr., and then remained constant at +50°. Deacetylation with 5 N sodium hydroxide. The procedure of Jeanloz and Forchielli (72) for hyaluronate was applied to chondroitin sulfate A. An amount of 2.00 g. of barium chondroitin sulfate A was dissolved in 100 ml. of a 5 N solution of sodium hydroxide by mixing a solution of the compound in 33 nil. of water with 67 ml. of 30% sodium hydroxide, all at 80°. The reaction mixture was maintained at 80° with occasional stirring for 1 hr. On cooling, the yellow-orange turbid solution was neutralized to pH 8 with dilute sulfuric acid, with cooling, and the solution was dialyzed against running distilled water for 5 days. The inorganic.residue was removed by filtration through asbestos, the filtrate concentrated under reduced pressure below ^t0° to a thin sirup and lyophilized; yield 0.35 g* (25%) of dark tan powder, -11.6° (jd 1 .03, water). Anal. Calcd. for ^..NNaO.. _(C0CH, )„. ,Q(S0,Na)„ : 111 X<~ Xo • 7J. XU (j • po J O • f X S, 5.07; COCH-,, 3-66. Found: S, 5.08; COCH, (77), 3*70. 5 5 Deacetylation with 5 W hydrochloric acid. The procedure of K. H. Meyer and coworkers (36) was adapted. An amount of 0.50 g. of barium chondroitin sulfate A was dissolved in 17.5 ml* of water in a glass-stoppered flask and 12.5 ml- of concentrated hydrochloric acid was added with stirring to bring the acid concentration to 5 H.* The reaction solution was maintained at 37-^0°, and runs were made for 3-, 6- and 12-hr. periods. The product was collected by pouring the turbid reaction mixture into 100 ml. of ethanol, 92 centrifuging and washing the residue thoroughly with ethanol. The washed material was then dissolved in water, passed through a Dowex 50 (H^ form) column (75 x 13 mm., diam.) and the effluent was treated with dilute sodium hydroxide until mildly alkaline and dialyzed against running distilled water for 3 days. The dialyzate was rid of turbidity by filtration through asbestos, concentrated and lyophilized; yield 0.32 g. (100)6), 0.23 g* (67%) and 0.18 g. (54%) of white powders,[a] ^ -17.9°i -15.8° and -13.8° (_c 1.04, 1.07 and 1.02, water) for the 3~» 6- and 12-hr. runs, respectively. Anal. (a) Three-hr. run. Calcd. for Cn„H..-NNaO..- Xci -Lo• J_0

(COCH^Jq (SO Na)Q 6Zf: S, 4.48; COCH^, 7-43. Found: S, 4.50; COCH, (77), 7.41. (b) Six-hr. run. Calcd. for C.^H.,/; ^JINaO, „ j l£ . JlO • / / i.U (C0CH3 )0<75(S03Na)0100%) of white powder, [a] -19.0° (c_ 1 .26, water). Anal. Calcd. for ^ 0010(C0CE^) (S0^BaQ 50)Q (?): S, 0.31; COCH^, 6.54; ash, as sulfate, 23.60. Found: S, 0.31; COCH^, (77) 6 .51; ash, as sulfate, 30.55* The infrared spectrum of this preparation was very similar to that of chondroitin sulfate A, despite the 96% desulfation calculated above. A faint purple coloration was obtained on ninhydrin treatment. Yields were verified by further repetition of the work. The benzyl alcohol was not replaced for the 3 day treatment, and the product passed through a Dowex 50 (H+ form) column and neutralized with sodium hydroxide. A higher than theoretical recovery was again noted of a white powder, M 25d -19.7°

(137) M. Somogyi, J. Biol. Chera., l60, 6l (1945). treated slightly cream-colored powder, W 23d -10-7° (_c I.03, water), was obtained in 66% yield. Anal. Calcd. for C12Hl6 >93MNa01 0 (C0CH )0^ 0 (SO Na )Q>6 : S, 4.81; C0CH-,, 3 .86. Found: S, 4.83; COCH,(77), 3*90; positive 3 3 ninhydrin test. Repetition of this reaction gave 67% yield of product. Anal. Calcd. for 92NNa010(C0CH3 )0< (SO N a ) ^ ?2: S, 5.13; COCH,, 3.44. Found: S, 5.14; COCH.. (77), 3.47; positive 3 3 ninhydrin test. tl-Deacetylated sodium chondroitin sulfate A. Sodium chondroiting sulfate A, C«r ^ -19.3 (.£, 1.42, water), with anticoagulant activity (91 ,92) determined at less than 3*3 I.U. per mg., was obtained from the barium salt by cation exchange. White hydrazine-treated powder, M -16.2° (£ 1.05, water), was obtained in 48% yield. Anal. Calcd. for c12Hl6.89HNa0lo(c“ V o A l (s05Ma)o.70: s- 5.01; OOCH , 5 .9^. Found: S, 5-02; COOH (77), 3.91*; positive ninhydrin test. N-Deacetylated sodium chondroitin. Sodium chondroitin (desulfated chondroitin sulfate) was derived from calcium chondroitin sulfate A by the action of 0.06 M methanolic hydrogen chloride after Kantor and Schubert (30) and the saponification of 95 the resulting methyl ester with dilute sodium hydroxide and lyophilizing; 8l% yield of white powder, -23.4° (c 1.09 water). Hydrazine treated product was obtained in 82% yield as a white powder, fa] -7 -5° (c 0.66, water).

Anal. Calcd. for 06N1. 22Na0.89°10. 89^G0CH3^ 0.27: N, 4.62; COCH^, 3*14. Found: N, 4.60; COCH^ (77), 3.14; positive ninhydrin test. A duplicate run gave 80 % yield of white powder. Anal. Calcd. for 62NWa°10^C0CIV 0 3 8 : C0CH3 » Found: COCH, (77), 4.31. 5 N-Deacetylated carboxyl-reduced chondroitin. A yield of 82 % of white powder, +13° (c. 0.47, dimethyl sulfoxide), was obtained from 96% carboxyl-reduced chondroitin on hydrazine treatmeht. The product was still water-insoluble. Anal. Calcd. for 30NOg(CH2OH)Q > 6 (C0CH )Q 70(C02Na) 0>Qif: N, 3.95; C0CH3 , 8.49. Found: N, 3.52; COCH^ (77), 8.51. White fluffy hydrazine-treated powder was recovered from 90% carboxyl reduced chondroitin in 68% yield. Anal. Calcd. for 2gN0g(CH20H)0> (COCH )Q 72(C02Na) COCH,, 8 .69. Found: COCH, (77), 8.72. U.1U 3 3 Sulfated Chondroitin Sulfate Modifications Sulfated N-deacetylated chondroitin sulfate, sodium salt. The sulfation procedure initiated in this laboratory by Wolfrom and Shen Han (91,92) for chitosan using sulfur trioxide-N,N- dimethylformamide in excess N,N-dimethyIformamide was adapted with some modifications. The amide solvent was redistilled at 149-150°• Sulfur trioxide was generated by heating fuming sulfuric acid (30% oleum)over phosphorus pentoxide and the fumes absorbed in the amide till a saturated solution (ca 3*38 N) was formed. The saturated complex was stored at 5° with a drying tube outlet. Prior activation of the polysaccharide involved precipitation from, aqueous solution with 3 vol. of ethanol, accelerated by the addition of a few ml. of satd. sodium chloride solution. The voluminous precipitate was collected in a 96 fritted-glass filter (medium pore size) and, without exposure to air, washed successively with 80% ethanol, grain alcohol, diethyl ether and N,N-dimethylformamide and stored over phosphorus pentoxide. The activated suspension of 62% N-deacetylated sodium chondroitin sulfate A (6.00 g.) in N,N-dimethylformamide was transferred into a one liter 2-necked round-bottomed flask fitted with a dropping funnel, drying tube and containing a Teflon-covered magnetic stirrer. Sulfur trioxide in N,N- dimethyIformamide (3*36 N, 200 ml.) was added to the activated material, with stirring at room temperature, in portions over a period of 2 hr. All the suspension dissolved within 2 hr., forming a clear brownish-red solution. After stirring for 10 hr. more, the reaction mixture was neutralized with solid sodium bicarbonate and the inorganic residue was removed by filtration. The filtrate and N,N-dimethylformamide washings were diluted with 2 vol. of ethanol and a few ml. of satd. sodium chloride solution and the tan precipitate was collected and dissolved in water. To this was added the water washings of the inorganic residue above. This solution was made slightly alkaline with sodium hydroxide, dialyzed against running distilled water for 4 days; yield 5*95 S» (64%) of tan powder, -1° (£ 1* water). Anal. Calcd. for Q (SO^Na)^ Qg : N, 3-52; S, 14.40; COCH^, 2 .38. Found: N, 3 .28; S, 13.58; COCH^ (77)« 2.46; negative ninhydrin test; anticoagulant activity (91,92), 8.2 I.U. per mg. Infrared absorption spectral analysis (Fig. 15) revealed very strong sulfate absorption at 1,220 to 1,260 cm."*" and bands absent in the starting material (Fig. 14) at about 990, 808 and 780 cmT^ besides the less prominent characteristic bands at 918, 848 and 720 cm.'*' Much sharper peaks were noted at 1,630 cm7^ and 3,400 to 3,500 cm7^, whereas the N-H deformation band at 1,560 cm7^ was much weaker than in sulfated chondroitin sulfate (Fig. 16). 97 The sheep blood plasma for the anticoagulant bioassays (91,92) was supplied by the Wilson Laboratories, Inc., Chicago, 111., in 50 ml* bottles containing 0,4% sodium citrate. Standard heparin was a complimentary sample from The Upjohn Co., Kalamazoo, Mich. The optimum volume of 2% calcium chloride solution required for recalcification of the particular lot of blood plasma was 0.12 ml. A concentration of 12>'of standard heparin in 4 ml. of 0.9% sodium chloride solution was found minimal to maintain the fluidity of 1 ml. of plasma. Sulfated preparations were made for the assay at a concentration of 1.00 mg. per ml. in 0 .9% sodium chloride solution. The chlorosulfonic acid-pyridine method was also applied to N-deacetylated sodium chondroitin sulfate A, essentially after Summers (75)* Technical pyridine was distilled over barium oxide at 112-113°. Technical chlorosulfonic acid was used without further purification. Prior activation of the polysaccharide involved precipitation from aqueous solution by the addition of three volumes of methanol with the addition of a few drops of satd. sodium chloride solution, the precipitate being collected by centrifuging and washed five times with dry pyridine before storing over phosphorus pentoxide. Pyridine (5 ml*) was cooled (salt-ice-water bath) in a 100 ml. three-necked round-bottomed flask fitted with a mercury seal, mechanical stirrer, dropping funnel and reflux condenser with a drying tube outlet. To this was added dropwise, with stirring, 0.6 ml. of chlorosulfonic acid, forming the white solid complex. An amount of 0.30 g. of the activated 62% N-deacetylated chondroitin sulfate (sodium salt) in pyridine was added quickly and the mixture was heated over a boiling water-bath for 1 hr. with stirring. During this heating the sulfated product formed a viscous mass at the bottom of the flask. On cooling, the yellow supernatant was carefully decanted and the brown residue was dissolved and neutralized with dilute sodium hydroxide. The solution was then dialyzed against running distilled water for 98 three days, filtered, concentrated under reduced pressure and lyophilized; yield 0.34 g. (88%) of brown powder, [a]^ +10° (c^ 0.4, water) .

Anal. Calcd. for c12Hl6.90W1.?2Na0.64°9.64(COCH3 )0.38^S°3 Na)i g^: S, 10.40. Found: S, 10.3 8 ; anticoagulant activity (91.92), 5.5 I.U. per mg.; negative ninhydrin test. Repetition of the sulfation on 59% N-deacetylated chondroitin sulfate A gave an 89% yield of brown powder. Anal. Calcd. for C ^ H ^ ^N N a O ^ C C O C H ^ ) ^ 4l^S03Wa ^l. 69 : S ’ 9-88. Found: S, 9*86; anticoagulant activity (91*92), 8.2 I.U. per mg. A modification of the procedure using twice the amount of chlorosulfonic acid (1.2 ml.) in pyridine (10 ml.) was employed on 0.25 g. of 62% N-deacetylated chondroitin sulfate A; yield 0.34 g. (95%) of light tan powder, +10° (jc 0.4, water). Anal. Calcd. for ^ 640 _ ^ 3 Na)2 S, 13.0. Found: S, 13*01; anticoagulant activity (91.92), 9.4 I.U. per mg.; negative ninhydrin test. Sulfated N-deacetylated carbonyl-reduced chondroitin sulfate* sodium salt. Sulfation of 60% N-deacetylated, 92% carbonyl- reduced sodium chondroitin sulfate A (0.35 g.) hy the chlorosulfonic acid-pyridine method described above gave a light tan powder: yield 0.39 g* (75%); [a] +5° (.c 1 * water). Anal. Calcd. for ^ N N a O ^ C O C H ^ ) ^ ( S O ^ N a ) ^ ?9: S, 13.50* Found: S, 13.50; anticoagulant activity (91*92), 16 I.U. per mg.; negative ninhydrin test, Sulfated chondroitin sulfate, sodium salt. Chondroitin sulfuric acid A (0.50 g.), derived from barium chondroitin sulfate A by deionizing with Dowex 50 (H form), upon treatment with chlorosulfonic acid-pyridine as above, gave light tan powder; yield O .63 g. (83%); [aj2^ -11*3° (c 1.04, water). Anal. Calcd. for ^^NNaO,„ (SO^Na)„ : S, 10.63. 1H lb.99 11 3 2.01 Found: S, 10.63; anticoagulant activity (91,92), 16 I.U. per mg.; ninhydrin test, negative. Infrared spectral analysis 99 (Fig. 16) showed strong and broad sulfate bands in the region 1,220 to 1,270 cm7^ and at 985i 810 and ?80 cm.'*' A comparison- with the spectrum of heparin (Fig, 17) showed that the N-H deformation band at 1,560 cmT'*' of the acetamido group in sulfated chondroitin sulfate was absent in heparin, which had the sulfoamino group instead. Sulfated N-deacetylated chondroitin, sodium salt. Chlorosulfonic acid-pyridine sulfation of 73% N-deacetylated sodium chondroitin (0.22 g.) provided a tan powder; yield 0 .2*+ g. (8l%); +8° (_c 0 .*+, water). Anal. Calcd. for C 12H 17i21N 1 _22N a 0i890 1 0 _8 9 (C0CH3.)0 _27,(S0J Na)© : S, 5*97- Found: S, 5*95; anticoagulant activity (91» 92), 3.3 I.U. per mg. Sulfated N-deacetylated carboxyl-reduced chondroitin, sodium salt. Water-insoluble 3>0% N-deacetylated 96% carboxyl-reduced chondroitin (0.35 g.) was treated with chlorosulfonic acid- pyridine as described above; yield 0.60 g, (98%) of a white water- I* *1 23 o soluble powder; LfxJ ^ +7 (c_ 1 .0 , water). Anal. Calcd. for 71N0g(CH20H)0 ^(COCH^)^?0 (C02Na) q Q^(S0^Na)2 S, 13-*+2. Found: S, 13.*+2; anticoagulant activity (91i92), 11 I.U. per mg.; negative ninhydrin test. rtANCE 100 100 00 4000 5000 00 4000 5000 1000 3000 2500 AEUKS N M 1 CM IN WAVENUMKRS AE UBR I CM-1 IN NUMBERS WAVE AE EGH N IRN AE EGH N MICRONS IN LENGTH WAVE MICRONS IN LENGTH WAVE AEEGH N MICRONS IN WAVELENGTH 2000 00 ISC 2000

50 40 1100 1400 1500 1200 1100 11,00 1000 1000 OIM CHONDROITIN SODIUM ULAT A TE LFA SU RD KERATOGULFATE CRUDE 3 FIGURE AENMES N CM-1 IN NUMBERSWAVE AE UBR I CM-' IN NUMBERS WAVE AEEGH N MICRONS IN WAVELENGTH •00

700 425 100 100 \ WAVE HUMMUS IN CM' WAVE NUMOEAS IN CM-' sooo n o o 1100 1000 MO too — -U_L 100

FMURE 4

PARTIALLY DC SULFATED

CRUDE KERATOBULFATE H O H WAVE LEN6TH M MKRONS WAVE LEN6TH IN MKOONS

WAVE HUMM US IN C M ' WAVE NUMMRS IN CM' sooo 3000 2000 1000 ISOO 1400 1100 1100 1000 100 100 100

iPMURC 8

'oNONDRorroi a c e t a t e

WAVE WAVE LMRIH M WAVE WAVE M CM-' SOOO 1400 I MO I I N MD 100 100

CMONOROITIN METHYL ESTER

2 1 4 i 0 7 * 10 II 12 13 14 II I*I WAVE LENGTH M MCRONS WAVE LENGTH IN MCSONS

WAVE NUMHRS IN CM' WAVE HUMMUS M C M ' 1000 100

FIGURE 7

CARBOXYL-REDUCED

CHONDROITIN

WAVE LENGTH M MCRONS WAVE LENGTH M MCRONS PERCENT TRANSMITTANCE PERCENT TRANSMITTANCE 100 100 00 4000 5000 00 00 00 50 00 50 40 30 20 10 00 0 10 0 425 700 100 900 1000 1100 1200 1300 1400 1500 2000 2500 3000 4000 5000 2 WVNMR I CM1 AEUKS N ' M C IN WAVENUMKRS MICRONS IN WAVELENGTH 1 M C IN WAVENUMKRS MICRONS IN WAVELENGTH I | 1 SOOO 2500 4 WAVENUMKRS IN C M ' ' M C IN WAVENUMKRS AEEGH N MICRONS IN WAVELENGTH 2000 5 6 50 40 1100 I 1400 1500 7 I 1200 1100 4 1000 10 ABXLRDCD N-ACETYL CARBOXYL-REDUCED MONOHYDRATE N-ACETYL-O-GALACTOSAMINE IUE 9 FIGURE 400 WAVENUMKRS IN CM-* CM-* IN WAVENUMKRS AEEGH N MICRONS IN WAVELENGTH 12 DMYDRATE •00 13 14 700 IS li 100 100 40 C ~r PERCENT PERCENT TRANSMITTANCE © w *701 0 m M 'EM PEKCS4T PEKCS4T TRANSMITTANCE PERCENT TRANSMITTANCE TRANSMITTANCE PERCENT TRANSMITTANCE PERCENT m

WAVE UMTH M MICRONS WAVE LVM1H M MCRONS WAVE NUMBERS IN CM * WAVE NUMIERS IN CM-' 5000 4000 3000 2500 MOO IS40 1400 1100 1100 1000 *25 ,00 i— I— i-L 100

•O N U M OHONDROITIN

WAVE LENGTH IN MICRONS WAVE LENGTH IN MICRONS

WAVE NUMIERS IN CM-1 WAVE NUMIERS IN CM-' SOOO 4000 3000 25002000 1500 1400 1300 1100 1000 100 *25 100 100

PARTIALLY N-DEACETYLATED

•ODIUM CHONDROITIN

WAVE LENGTH IN MICRONS WAVE LENGTH IN MICRONS £ PERCENT TRANSMITTANCE 100 100 O O 4000 SOOO 00 4000 5000 2 4 3 3000 002500 3000 2SOO AE UIR I CM-' IN NUMIERS WAVE AE UIR I CM-' IN NUMIERS WAVE AE EGH ORN WV LNT M MCRONS M LENGTH WAVE WOCRONS M LENGTH WAVE AE EGH N MICRONS IN LENGTH WAVE 2000 2000 I S

50 40 1300 1400 1500 50 40 1300 1400 1500 7 1100 1100 * 1000 1000 10 ATAL NDAEYAE OHONORCHTIN N-DEACETYLATED PARTIALLY 14 E R U I8 F UFT A, OIM SALT SODIUM , A SULFATE UFT, ONM SALT SOONJM SULFATE, M T O M N O H C N-OEAOCTYLATED PARTIALLY SULFATED AE UIR I CM' M C IN NUMIERS WAVE AE UIR I CM-' IN NUMIERS WAVE AE EGH N MICRONS IN LENGTH WAVE II II 4IS 14 S U S U IS 100 100 40 at PERCENT TRANSMITTANCE look 20 40 40 O O 00 00 50 00 50 40 10 10 1100 1100 1100 1400 1500 1000 2500 3000 4000 SOOO WVNMES N ' M C IN WAVENUMOERS I . J ..... : 1 1 1 1 t 1 1 1 1 ...... h .. . I_L 1 I 1 AEUKS M'WVNMR I CM' M C IN WAVENUMKRS CM-' M WAVENUMKRS AEEGH N MICRONS IN WAVELENGTH AEEGH N MCRONS IN WAVELENGTH 4 7 4 S 4 7 4 S , 1 . 1 11 , 1 1 1 1 , . II 1 111 1 1 1 f I ■ 1 : 1: 1 I ;.:T T O 40 10 10 1100 1100 1100 1400 I HO ; : ; : .... ■: i ■ : i ; ■ i '! : I i i : : : \ . V ! V V i I 1 V i \ \ - f - -1 ■ : - r - / : / : i i i : ' ; i ' : • : : : ' i i l - - ill ; : ■£.r: it H ■ 1 : : - : 13 1 1 , I , !1 , : : :: . ; . ; : : i : ; : ; : i ■ ;; : : : , ‘ : .I.;.;. : : : ' I ’ ! ..... ■ : .... :i : 1 !

! i H i : ' . , , i ' i . :■ i • I , 10 ; . ■ ' 1 i 1 : t i r ■ i j | : , i . i . ii i -f i • ' 1 r-r- \ / A V '• ; :yr ...... : ■ ■ * 14 I I I I I I 14 I I I I I I V 1 V , 1 AEUKS N ' M C IN WAVENUMKRS t GURE 17 E R U IG F AEEGH N MICRONS IN WAVELENGTH AEEGH N MICRONS IN WAVELENGTH r .,i,i, ONUH U N IO . i ■ALT i.i.i , . i.i.i, i 700 IS 14 I

£ 107 V. SUGGESTIONS FOR FURTHER RESEARCH

1. Hydrolytic studies should be made on partially N- deacetylated carboxyl-reduced chondroitin with emphasis on the disaccharide composition. 2. Isolation of the disaccharide fraction from partially N-deacetylated chondroitin sulfate A should be pursued and the fraction characterized to establish the homogeneity of the chondrosine thus obtained. 3* Attempts should be made to isolate and characterize "anhydro-N-acetyl-D-galactosamine" (the p-elimination product) by the larger scale alkaline degradation of carboxyl-reduced chondroitin or of carboxyl-reduced Itf-acetylchondrosine. 4. Deamination studies may be pursued on the partially N-deacetylated polysaccharide. Hydrazine treatment of model compounds may help elucidate the nature of the increase in nitrogen content noted for the hydrazine action on chondroitin sulfate A. 3. Selective degradation of chondrosine to 2-C>- (P-D-glucopyranosyl)-glycerol should establish the p-D-glycoside assignment for the disaccharide.

108 VI. SUMMARY AND CONCLUSIONS 1. Carboxyl-reduced chondroitin was prepared by exhaustive sodium borohydride reduction in borate buffer of chondroitin methyl ester. A degraded product was obtained by reduction of the degraded chondroitin methyl ester, derived by diazomethane esterification of _0-deacetylated chondroitin acetate. Partial acid hydrolysis of the carboxyl-reduced polymer, N-acetylation of the hydrolyzate and carbon column fractionation facilitated the isolation and characterization of D-glucose, N-acetyl-oc-D- galactosamine monohydrate and the sole disaccharide, O-p-D- glucopyranosyl-(1—»3)-2-acetamido-2-deoxy-a-D-galactose dihydate, a new compound. This demonstrated that p-D- glucopyranosidic linkages were more resistant to acid cleavage than N-acetyl-P-D-galactosaminidic linkages in the polymer. The disaccharide ^ g ^ ucose 0*7*0 was readily degraded in alkaline solution to D-glucose and a Morgan-Elson (*f9) reactive sugar (^gpucose 1*3), "anhydro-N-acetyl-D-galactosamine,11 different from N-acetyl-D-galactosamine (R n 1.2). — glucose 2. Sodium borohydride treatment of the disaccharide gave the alditol (O-P-D-glucopyranosyl-(1—£5)-2-acetamido-2-deoxy- D-galactitol), which was chromatographically identical (R g-LUC OS © O.76) in. three solvent systems with that derived from chondrosine methyl ester hydrochloride (50)* The alditol was resistant to alkaline degradation and together with N-acetylhexosaminols, gave positive Morgan-Elson (116) test. They are new exceptions to this test for N-acetylhexosamines. 3. Of the various N-deacetylating agents reevaluated for barium chondroitin sulfate A, a modification of the hydrazine treatment of Matsushima and Fujii (73) gave consistently partially desulfated, highly (39-68%) N-deacetylated polymer, whereas acidic deacetylating agents effected too much hydrolysis and desulfation to be practical; sodium hydroxide gave a highly degraded product. The nature of the inorganic cation and desulfation of the polymer had negligible effect on the

- 109 110 efficiency of the hydrazine reaction. However, carboxyl-reduced chondroitin showed low (29%) N-deacetylation, denoting that the absence of anionic groups, specifically carboxylate, adversely reduced the efficiency of the reaction. Increase in nitrogen content was noted also with uronate and sulfate containing modifications of chondroitin sulfate A. 4. Sulfur trioxide-N,N-dimethylformamide and chlorosulfonic acid-pyridine sulfation of chondroitin sulfate A and its N- deacetylated modifications increased their anticoagulant activity to only 15% that of heparin. However, prior W- deacetylation of chondroitin sulfate A with hydrazine did not improve the activity of the sulfated preparation (sodium salt). No direct relationship was evident between the degree of sulfation and anticoagulant activity. The absence of the 1,560 crnT^" N-H deformation band in the infrared was confirmed as characteristic of the sulfoamino function as against the acetamido group in mucopolysaccharides. 5* Crude keratosulfate isolated from commercial chondroitin sulfate A showed bands at 775» 820 and 998 cm7'L in the infrared, consistent for an equatorial sulfate group (53)* It was only partially desulfated with methanolic hydrogen chloride. CHRONOLOGICAL BIBLIOGRAPHY

Reference Page 2 G. Fischer and C. Boedeker, Ann., 117, 111 (l86l). 2 1 C. F. W. Krukenburg, Z. Biol., PO, 307 (1884). 2 3 0. Schmiedeberg, Arch, exptl. Pathol. Pharmakol., 281 335 (1891). 2 6 P. A. Levene and W. A. Jacobs, J. Exptl. Med., 10, 557 (1908). 3 lf7 P. A. Levene and F. B. LaForge, J. Biol. Chem., 15, 69 (1913). 15 5 J. Hebting, Biochem. Z., 63, 353 (1914). 2 8 P. A. Levene and F. B. LaForge, J. Biol. Chem., 18, 123 (191*0. 3 35 W. N. Haworth, J. Chem. Hoc., 107, 8 (1915). 11 9 P. A. Levene, J. Biol.Chem., 2£>, l*f3 (1916). 4 10 P. A. Levene, J. Biol. Chem., 31, 609 (1917). 4 4 P. A. Levene, "Hexosamines and Mucoproteins,11 Longmans, Green and Co., London, 1925 2 109 E. A. Moelwyn-Hughes, Trans. Faraday Soc., 25, 503 (1929). *0 78 V. Demole and M. Reinert, Arch, exptl. Pathol. Pharmakol., 158, 211 (1930). 29 79 A. Fischer, Biochem. Z., 240, 364 (1931). 29 49 L. A. Elson and W. T. J. Morgan, Biochem. J., 27, 1824 (1933). 16 116 W. T. J. Morgan and L. A. Elson, Biochem. J., 28, 988 (193*0. A-6 80 E. Chargaff, F. W. Bancroft and M. Stanley Brown, J. Biol. Chem., 115, 155 (1936). 29 13 K. Meyer and Elizabeth M. Smyth, J. Biol, Chem., 119, 507 (1937)-. 5 107 R. C. G. Moggridge and A. Neuberger, J. Chem. Soc., 7^5 (1938). *0 111 112 Reference Page 112 R. L. Whistler, A. R. Martin and M. Harris, J. Research Nat. Bur. Standards, 2k, 13 (1940). 45 45 P. A. Levene, J. Biol. Chem., l40, 267 (1941). 14 27 K. Meyer and Eleonor Chaffee, J. Biol. Chem., 138, 491 (1941). 10 98 K. Freudenberg, H. Walch and H. Molter, Naturwissenschaften, 30, 87 (1942). 36 111 W.A.G. Nelson and E. G. V. Percival, J. Chem. Soc., 58 (1942). 45 82 P. Karrer, H. Koenig and E. Usteri, Helv. Chim. Acta, 26, 1296 (1943). 29 88 M. L. Wolfrom, D. I. Weisblat, J. V. Karabinos, vV. H. McNeely and J. McLean, J. Am.. Chem. Soc., 65, 2077 (19^3). 30 7 H. G. Bray, J. E. Gregory and M. Stacey, Biochem. J., 38, 142 (1944). 3 81 P. Karrer, E. Usteri and B. Camerino, Helv. Chim. Acta, 27, 1422 (1944). 29 117 M. Stacey, J. Chem, Soc., 272 (1944). 46 11 84 A. Gronwall, B. Ingelman and H. Mosiman, Upsala Lakareforen. Forh., 30,397 (19^5)* 29 11 Sybil P. James, F. Smith, M. Stacey and L. F. Wiggins, Nature, 136, 308 (1945). 4 137 M. Somogyi, J. Biol. Chem., 160, 6l (1945). 94 88 M. L. Wolfrom and W. H. McNeely, J. Am. Chem. Soc. 67, ?48 (1945). 30 11 Sybil P. James, F. Smith, M. Stacey and L. F. Wiggins, J. Chem. Soc., 625 (1946). 4 36' K. H. Meyer and M. E. Odier, Experientia, 2 , 311 (19k6). 12 22 Z. Dische, J. Biol. Chem., 167, 189 (1947). 9 129 V. Bruckner, G. Fodor, J. Kiss and J. Kovacs, J. Chem. Soc., 885 (1948). 38 115 Reference Page 5.6 K. H. Meyer, M. E. Odier and A. E. Siegrist, Helv. Chim. Acta, 51, 1400 (1948). 12 105 S. M. Partridge, Biochem. J. , 42, 258 (1948). 45 15 S. M. Partridge, Biochem. J., 45, 587 (1948). 6 154 J. J. Pratt, Jr. and J. L. Auciair, Science, 108, 215 (1948). 68 24 M. V. Tracey, Biochem. J., 45, 185 (1948). 9 121 S. M. Partridge, Nature, 164, 445 (1949). 49 152 0. F. Swoap and M. H. Kuizenga, J. Am. Pharm. Assoc., 5 8 , 565 (1949). 64 118 D. Aminoff, W. T. J. Morgan and W. M. Watkins, Biochem. J. , _46, 426 (1950). 47 14 Julia Einbinder and M. Schubert, J. Biol. Chem., 185, 725 (1950). 6 it 57 S. Gardell, F. Heijkenskjold and A. Roch-Norlund, Acta Chem. Scand. , 4^, 970 (1950). 21 11 89 J- E. Jorpes, H. Bostrom and V. Mutt, J. Biol. Chem., 185, 607 (1950). 50 125 R. K. Madison, Ph. D. Dissertation, The Ohio State University, 1950. 52 106 W. E. Trevelyan, D. P. Proctor and J. S. Harrison, Nature, 166, 444 (1950). 45 115 R. L. Whistler and D. F. Durso, J. Am, Chera. Soc., 72, 677 (1950). 45 55 M. L. Wolfrom and R. Montgomery, J. Am. Chem. Soc., 72, 2859 (1950). 19 72 R. W. Jeanloz and E. Forchielli, J. Biol. Chem., 190, 557 (1951). 26 21 R. Marbet and A. Winterstein, Helv. Chim. Acta, 54, 2511 (1951). 9 48 H. Masamune, 2. Yosizawa and M. Maki, Tohoku J. Exptl. Med., 55, 47 (1951). 16 129 G. E. McCasland, J. Am. Chem. Soc., 75, 2295 (1951). 58 Reference 31 M. L. Wolfrom and H. B. Wood, J. Am. Chem. Soc., 73, 2933 (1951). 127 S. Akabori, K. Ohno and K. Narita, Bull. Chem. Soc. Japan, 25, 214 (1952). 129 G. Fodor and J. Kiss, J. Chem. Soc., 1589 (1952). 23 J. X. Khym and D. G. Doherty, J. Am. Chem. Soc., 74, 3199 (1952). 21 R. Marbet and A. Winterstein, Experientia, 8_, 41 (1952). 83 K. H. Meyer, R. P. Piroue and M. E. Odier, Helv. Chim. Acta, 35, 574 (1952). 85 C. R. Ricketts and K. IV. Walton, Chem. & Ind. (London), 869 (1952). 32 M. L. Wolfrom and Kimiko Anno, J. Am. Chem. Soc. 74, 5583 (1952). 46 M. L. Wolfrom, R. K. Madison and M. J. Cron, J. Am. Chem. Soc., 74, 1491 (1952). 129 G. Fodor and K. Nador, J. Chem. Soci., 721 (1953) 122 R. Kuhn, Adeline Gauhe and H. H. Baer, Chem. Ber., 86, 827 (1955). 18 M, B. Mathews and A, Dorfman, Arch. Biochem. Biophys., 42, 4l (1953). 16 K. Meyer, A. Linker, E. A. Davidson and B. Weissmann, J. Biol. Chem., 205, 6ll (1953)* 69 K. H. Meyer and G. Baldin, Helv. Chim. Acta, 36, 597 (1953). 70 C. R. Ricketts, Research (London), £>, 17S_ (1953) 6 M. L. Wolfrom and W. B. Neely, J. Am. Chem. Soc. 75, 2778 (1953). 74 M. L. Wolfrom, (Miss) T. M. Shen and C. G. Summers, J. Am. Chem. Soc., 75, 1519 (1953). 115 Reference Page 94 I. B. Cushing, R. V. Davis, E. J. Kratovil and D. W. MacCorquodale, J. Am. Chem. Soc,, 76, 4590 (195*0. 32 62 E. A. Davidson and K. Meyer, J. Biol. Chem., 211, 605 (195*0. 24 50 E. A. Davidson and K. Meyer, J. Am. Chem.Soc., 76, 5686 (195*0. 16 126 T. J. De Boer and H. J. Backer, Rec. trav. chim., 73, 229 (195*0. 52 119 R. Kuhn, Adeline Gauhe and H. H. Baer, Chem. Ber. , 87, 289, 1183 (195*0- 48 52 S. F. D. Orr, Biochim. et Biophys. Acta, 14, 173 (195*0. 18 114 S. Roseman and J. Ludowieg, J. Am. Chem. Soc., 76, 301 (195*+). 45 39 Jennie Shatton and M. Schubert, J. Biol. Chem., 211, 565 (195*0. 12 91 T. M. Shen Han, Ph. D. Disseration, The Ohio State University, 1954. 31 57 P- J* Stoffyn and R. W. Jeanloz, Arch. Biochem. Biophys., 52, 373 (195*0. 21 34 L. Vargha, Chem. Ber., 87, 1351 (195*0» 11 19 B. C. Bera, A. B. Foster and M. Stacey, J. Chem. Soc., 3788 (1955). 7 51 E. A. Davidson and K. Meyer, J. Am. Chem. Soc., 77, 4796 (1955). 18 88 A. B. Foster and A. J. Huggard, Advances in Carbohydrate Chem., 10, 335 (1955). 30 63 S. Gardell, Acta Chem. Scand. , j?, 1035 (1955). 24 131 F. S. H. Head, J. Textile Inst., 46^, T584 (1955). 61 97 P* W. Kent and M. W. Whitehouse, "Biochemistry of the Aminosugars," Academic Press Inc., New York, 1955. 36 ll6 Reference Page 86 E. London, R. S. Theobald and G. D. Twigg, Chem. 8t Ind. (London), 1060 (1955)* 30 26 F. Shafizadeh and M. L. Wolfrom, J. Am. Chem. Soc., 77, 2568 (1955). 10 75 C. G. Summers, Ph. D. Dissertation, The Ohio State University, 1955* 27 104 S. A. Barker, E. J. Bourne and D. H. Whiffen, Methods of Biochem, Anal., J5, 213 (1956). 42 102 H. C. Brown and B. C. Subba Rao, J. Am. Chem. Soc., 78, 2582 (1956). 4l 77 A. Chaney and M. L. Wolfrom, Anal. Chem., 28, lbl4 (1956). 27 90 G. D. Forwell and G. I. C. Ingram, J. Pharm. Pharmacol., 8_, 530 (1956). 30 33 Harriet L. Frush and H. S. Isbell, J. Am. Chem, Soc., 78, 2844 (1956). 11 25 P. Hoffman, A. Linker and K. Meyer, Science, 124, 1252 (1956). 9 61 P. Hoffman, K. Meyer and A. Linker, J. Biol. Chem., 219, 653 (1956). 23 123 R. W. Jeanloz and Monique Tremege, Federation Proc., 15, 282 (1956). 49 37 M. B. Mathews, Arch. Biochem. Biophys., 6l, (1956). 12 20 K. Meyer, E. Davidson, A. Linker and P. Hoffman, Biochim. et Biophys. Acta, 21, 506 (1956). 7 19 J. E. Scott, Biochem. J,, 62, 3l£ (1956). 7 38 G. Bernard!, Compt, rend., 244, 1918 (1957)* 12 28 J. A. Cifonelli, J. Ludowieg and A. Dorfman, Federation Proc., 16, 165 (1957). 10 120 W. R. C. Crimmin, J. Chem. Soc., 2838 (1957). 49 68 M. J. Crumpton, Nature, l80, 605 (1957)* 25 108 A. B. Foster, D. Horton and M. Stacey, J. Chem. Soc., 81 (1957). 44 117 Reference Page 59 P. Hoffman, A. Linker, Phyllis Sampson, K. Meyer and E. Korn, Biochim. et Biophys. Acta, 25, 658 (1957). 21 56 R. W. Jeanloz, P. J. Stoffyn and Monique Tremege, Federation Proc., 16, 201 (1957)* 20 30 T. G. Kantor and M. Schubert, J. Am. Chem. Soc., 79, 152 (1957). 10 60 E. D. Korn, J. Biol. Chem., 226, 84l (1957). 22 130 A. Linker, P. Hoffman and K. Meyer, Nature, 180, 810 (1957). 60 73 Y- Matsushima and N. Fujii, Bull. Chem. Soc. Japan, 30, 48 (1957). 26 67 Helen Muir, Biochem. J. , 65, 33j2 (1957). , 25 128 W. G. Overend and C. R. Ricketts, Chem. & Ind. (London), 632 (1957). 55 76 E. D. Toro-Feliciano, M. Sc. Thesis, The Ohio State University, 1957. 27 93 M. L. Wolfrom, R. A. Gibbons and A. J. Huggard, J. Am. Chem. Soc., 79, 5043 (1957). 31 17 M. L. 'Wolfrom and K. Onodera, J. Am. Chem. Soc., 79, 4737 (1957). 6 71 S. A. Barker, A. B. Foster, M. Stacey and J. M. Webber, J. Chem. Soc., 2218 (1958). 26 58 J. A. Cifonelli and A. Dorfman, J. Biol. Chem., 231, 11 (1958). 21 28 J. A. Cifonelli, J. Ludowieg and A. Dorfman, J. Biol. Chem., 233, 541 (1958). 10 101 H. Endres and M. Oppelt, Chem. Ber., 91* 478 (1958). 41 66 A. Hallen, Acta Chem. Scand., 12, 1869 (1958). 25 54 P. Hoffman, A. Linker and K. Meyer, Biochim. et Biophys. Acta, 30, 184 (1958). 18 118 Reference Page 20 P. Hoffman, A. Linker and K. Meyer, Federation Proc., 17, 1078 (1958). 7 29 R. IV. Jeanloz and P. J. Stoffyn, Federation Proc., 17, 249 (1958). 10 56 R. W. Jeanloz, P. J. Stoffyn and Monique Tremege, "Chemistry and Biology of Mucopolysaccharides," G. E. W. Wolstenholme and Maeve O'Connor, Ed., Little, Brown and Co., Boston, Mass., 1958. 20 65 R. Kuhn and H. J. Leppelmann, Ann., 6ll, 254 (1958). 25 43 Ina Malawista and M. Schubert, J. Biol. Chem., 230, 535 (1958). 13 53 M. B. Mathews, Nature, l8l, 421 (1958). 18 42 M. B, Mathews and Irene Lozaityte, Arch. Biochem. Biophys., 74, 158 (1958). 13 64 K. Meyer, P. Hoffman and A. Linker, Science, 128, 896 (1958). 24 4l Helen Muir, Biochem. J. , 69., 195 (1958). 13 103 H. Neukom and H. Deuel, Chem. & Ind. (London), 683 (1958). 42 40 S. M. Partridge and H. F. Davis, "Chemistry and Biology of Mucopolysaccharides," G. E. W. Wolstenholme and Maeve O'Connor, Ed., Little, Brown and Co., Boston, Mass., 1958. 13 133 M. Stacey, "Chemistry and Biology of Mucopolysaccharides," G. E. W. Wolstenholme and Maeve O'Connor, Ed., Little, Brown and Co., Boston, Mass., 1958. 65 113 E. Stutz and H. Deuel, Helv. Chim. Acta, 4l, 1722 (1958). 45 95 D. T. Warner and L. L. Coleman, J. Org. Chem., 23, 1133 (1958). 32 Reference 44 R. C. Warner and M. Schubert, J. Am. Chem. Soc 80, 5166 (1958). 12 R. L. Whistler and J. N. BeMiller, Advances in Carbohydrate Chem., 131 289 (1958). 110 R. L. Whistler and G. N. Richards, J. Am. Chem Soc. , 80, *+880 (1958). 87 J. W. Wood and P. T. Mora, J. Am. Chem. Soc., 80, 3700 (1958). 99 I. Danishefsky and H. B. Eiber, Federation Proc., 18, 210 (1959). 92 M. L. Wolfrom and T. M. Shen Han, J. Am. Chem. Soc., 81, 1764 (1959). AUTOBIOGRAPHY I, Bienvenido Ochoa Juliano, was born in Los Banos, Laguna, Philippines, August 15, 1936. I completed my elementary and secondary education at my home town of Calamba, Laguna. At the University of the Philippines, I received the degree Bachelor of Science in Agriculture, magna cum laude, majoring in sugar chemistry, in October, 1955* I was assistant instructor in agricultural chemistry at the university until July, 1956, when I resigned to accept an assistantship in the Department of Chemistry of The Ohio State University. In March, 1958, I was conferred the degree Master of Science by the University. While completing the requirements for the degree Doctor of Philosophy, I held the position of Assistant, October, 1956, to March, 1957; Research Assistant, April to October, 1957; National Science Foundation Predoctoral Fellow, November, 1957, to June, 1958, under Grant NSF-GW^^ to The Ohio State University; and C. F. Kettering Research Foundation Fellow, July, 1958, to date.

120