Structure and anticoagulant activity of a fucosylated chondroitin sulfate from echinoderm. Sulfated fucose branches on the polysaccharide account for its high anticoagulant action.

A polysaccharide isolated from the body wall of the sea cucumber Ludwigothurea grisea has a backbone like that of mammalian chondroitin sulfate: [4-beta-D-GlcA-1-->3-beta-D-GalNAc-1]n but substituted at the 3-position of the beta--glucuronic acid residues with sulfated alpha--fucopyranosyl branches (Vieira, R. P., Mulloy, B., and Mourão, P. A. S. (1991) J. Biol. Chem. 266, 13530-13536). Mild acid hydrolysis removes the sulfated alpha--fucose branches, and cleaved residues have been characterized by 1H NMR spectroscopy; the most abundant species is fucose 4-O-monosulfate, but 2,4- and 3, 4-di-O-sulfated residues are also present. Degradation of the remaining polysaccharide with chondroitin ABC lyase shows that the sulfated alpha-L-fucose residues released by mild acid hydrolysis are concentrated toward the non-reducing end of the polysaccharide chains; enzyme-resistant polysaccharide material includes the reducing terminal and carries acid-resistant -fucose substitution. The sulfated alpha-L-fucose branches confer anticoagulant activity on the polysaccharide. The specific activity of fucosylated chondroitin sulfate in the activated partial thromboplastin time assay is greater than that of a linear homopolymeric alpha-L-fucan with about the same level of sulfation; this activity is lost on defucosylation or desulfation but not on carboxyl-reduction of the polymer. Assays with purified reagents show that the fucosylated chondroitin sulfate can potentiate the thrombin inhibition activity of both antithrombin and heparin cofactor II.

Anticoagulant and antithrombotic activities are among the most widely studied properties of sulfated polysaccharides. The anticoagulant glycosaminoglycan heparin is an important therapeutic agent used in the prophylaxis and treatment of thrombosis (14); dermatan sulfate is also an anticoagulant, although of lower potency than heparin (15)(16)(17). A chemically sulfated xylan from beechwood, pentosan polysulfate, has been available for many years as an anticoagulant polysaccharide (18 -20). Sulfated fucans from brown seaweed have anticoagulant activity due to the ability to potentiate inhibition of thrombin by antithrombin or heparin cofactor II (21,22).
Recently, we isolated novel sulfated polysaccharides from the body wall of a sea cucumber (5, 7, 9 -11). We found that the main fraction has a chondroitin sulfate-like structure, containing large numbers of sulfated ␣-L-fucopyranose branches linked to position 3 of the ␤-D-glucuronic acid residues (5,7). We now present both revision and further refinement of our previous structure.
The analogy in structure between the fucosylated chondroitin sulfate from sea cucumber, heparin, or dermatan sulfate from mammalian tissues and sulfated fucans from brown algae led us to investigate the possible anticoagulant activity of the echinoderm polysaccharide. We observed a high anticoagulant activity in the fucosylated chondroitin sulfate due to its ability to potentiate inhibition of thrombin and factor Xa by antithrombin or heparin cofactor II. Measurements of anticoagulant activities of chemically modified polymers show that the high anticoagulant activity of the sea cucumber polysaccharide can be assigned mainly to sulfated fucose branches linked to the chondroitin sulfate core.

EXPERIMENTAL PROCEDURES
Native and Chemically Modified Polysaccharides-Fucosylated chondroitin sulfate was extracted from the body wall of the sea cucumber Ludwigothurea grisea by papain digestion, and purified by procedures previously described (5,7). Desulfation of this polysaccharide by solvolysis in dimethyl sulfoxide/methanol (9:1, v/v) at 80°C for 6 h (23) and reduction of the hexuronic acid carboxyl groups in the polysaccharide by 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide-NaBH 4 (24) were performed as described previously. Yield and integrity of the polysaccharides obtained after these two procedures were the same as in our previous study (7,11). Partial removal of sulfated fucose branches from the fucosylated chondroitin sulfate was performed by mild acid hydrolysis. In these experiments, the fucosylated chondroitin sulfate (50 mg) was dissolved in 1.0 ml of 150 mM H 2 SO 4 , maintained at 100°C for 30 min, and the pH of the solution was adjusted to 7.0 with 0.3 ml of ice-cold 1.0 M NaOH. Chondroitin 4-sulfate from whale cartilage, chondroitin 6-sulfate from shark cartilage, dermatan sulfate from bovine mucosa, and heparin from porcine intestinal mucosa were from Sigma. Dermatan sulfate was treated with nitrous acid (25) to remove contaminating heparin. Heparan sulfate from human aorta was extracted and purified as described previously (26). The heparin for the APTT 1 assay was the 4th International Standard (85/502), obtained from the National Institute for Biological Standards and Control, Potters Bar, UK.
3 H Labeling of the Fucosylated Chondroitin Sulfate-Alkaline-catalyzed ␤-elimination of the linkage region of the fucosylated chondroitin sulfate (50 mg) was performed in 1.0 ml of 0.1 M NaOH in the presence of ϳ2 mCi of [ 3 H]NaBH 4 at room temperature. After 8 h non-labeled NaBH 4 was added to a final concentration of 0.1 M and the solution was maintained at room temperature for an additional 8-h period. The solution was then neutralized with acetic acid, and the products were fractionated on a Bio-Gel P-4 column (see below). Fractions were collected, assayed by metachromasia using 1,9-dimethylmethylene blue (27) and the radioactivity was counted on a scintillation counter. The fractions containing the 3 H-labeled fucosylated chondroitin sulfate were pooled and lyophilized.
Analysis of the Products Formed by Mild Acid Hydrolysis of the Fucosylated Chondroitin Sulfate-3 H-Labeled fucosylated chondroitin sulfate (50 mg) was submitted to mild hydrolysis with acid (see above). After neutralization with NaOH, the solution was applied to a Bio-Gel P-4 column (88 ϫ 1.5 cm) and eluted with 50 mM pyridine/acetate buffer (pH 6.0) at a flow rate of 6 ml/h. Fractions of 1.5 ml were collected and assayed by metachromasia using 1,9-dimethylmethylene blue (27), the carbazole (28), and Dubois (29) reactions, and the radioactivity was counted on a scintillation counter. The fractions containing the "acidresistant fragments," hereafter designated as partial defucosylated chondroitin sulfate, were identified by the carbazole and metachromatic positive tests. The released "sulfated fucose" was identified by positive Dubois test and the absence of carbazole and metachromatic reactions. Both fractions were pooled and lyophilized.
In order to remove small amounts of unsulfated fucose (and also of inorganic sulfate) from the sulfated fucose sample, the fraction was re-applied to a Bio-Gel P-2 column (90 ϫ 0.8 cm) and eluted with distilled water at a flow rate of 6 ml/h. Fractions of 1.0 ml were collected and assayed by the Dubois reaction (29). The fractions of this second column containing sulfated fucose were pooled and lyophilized. This product (ϳ15 g) was applied to Whatman No. 1 paper and separated by descending chromatography in isobutyric acid, 1.0 M NH 4 OH (5: 3,v/v) for 24 h. In addition, the purified sulfated fucose (ϳ15 g) was applied to Whatman 3MM chromatographic paper and submitted to electrophoresis in 0.3 M pyridine/acetate buffer (pH 5.0), run for 4 h at 500 V. The chromatogram and paper electrophoresis were stained with silver nitrate.
Analysis of the Products Formed by Digestion of the Partial Defucosylated Chondroitin Sulfate with Chondroitin ABC Lyase-The partial defucosylated chondroitin sulfate (ϳ20 mg) was incubated with 1 unit of chondroitin ABC lyase (EC 4.2.2.4) from Proteus vulgaris (Seikagaku American Inc., Rockville, MD) in 1 ml of 50 mM Tris/HCl buffer (pH 8.0), containing 5 mM EDTA and 15 mM sodium acetate. After incubation at 37°C for 12 h, the reaction mixture was applied to a Bio-Gel P-4 column and chromatographed as described above. The fractions containing the "chondroitin lyase-resistant fragments" were identified by the carbazole and metachromatic positive tests. The released disaccharides were identified by the positive carbazole reaction. Both groups of fractions were pooled and lyophilized. The disaccharides fraction (ϳ100 g) was In Panel A, the 3 Hlabeled fucosylated chondroitin sulfate (50 mg) was submitted to mild acid hydrolysis (see "Experimental Procedures"). After neutralization with NaOH, the solution was applied to a Bio-Gel P-4 column (88 ϫ 15 cm) and eluted with 50 mM pyridine/acetate buffer (pH 6.0) at a flow rate of 6 ml/h. Fractions of 1.5 ml were collected and assayed by the carbazole (q) and Dubois (E) reactions, for metachromasia (Ç), and counted in a liquid scintillation counter (å). The fractions containing the acid-resistant fragments and sulfated fucose (horizontal bars) were pooled and lyophilized. The sulfated fucose was re-purified on a Bio-Gel P-2 column (90 ϫ 0.8 cm), eluted with distilled water at a flow rate of 6 ml/h (inset in Panel A). Fractions of 1.0 ml were collected and assayed by the Dubois reaction (E). Fractions 72, 73, and 74 (horizontal bar) were pooled and lyophilized. In Panel B, the purified sulfated fucose (ϳ15 g), before (Ϫ) and after (ϩ) strong acid hydrolysis and a mixture of standard sugars containing 10 g each of galactose, mannose, fucose, and galactosamine were spotted on Whatman No. 1 paper and subjected to chromatography in isobutyric acid, 1 N NH 4 OH (5:3, v/v) for 24 h. The products were located on the chromatogram by silver nitrate staining. In Panel C, the sulfated fucose (ϳ15 g) before (Ϫ) and after (ϩ) strong acid hydrolysis and a mixture of standard sugars containing 10 g each of glucuronic acid, fucose, and galactosamine were spotted on Whatman 3MM chromatographic paper and submitted to electrophoresis in 0.3 M pyridine/acetate buffer (pH 5.0), run for 4 h at 500 V. The electrophoresis was stained with silver nitrate. fractions of 0.5 ml were collected and assayed by metachromasia using 1,9-dimethylmethylene blue (27) and by the carbazole (28) and Dubois (29) reactions.
Chemical Analysis-After strong acid hydrolysis (4.0 M HCl, 100°C for 6 h) of the polysaccharide total hexosamine and sulfate were estimated by a modified Elson-Morgan reaction (30) and by the BaCl 2gelatin method (31), respectively. Standard curves for hexosamine and sulfate were constructed from glucosamine and Na 2 SO 4 . The hexuronic acid content was estimated by the carbazole reaction (28). The percentages of fucose and galactosamine in the acid hydrolysates were estimated by paper chromatography in 1-butanol/pyridine/water (3:2:1, v/v) for 48 h or in isobutyric acid, 1.0 M NH 4 OH (5:3, v/v) for 24 h and by gas-liquid chromatography of the corresponding alditol acetates (32).
NMR Methods-1 H spectra were recorded at 500 MHz and 13 C spectra at 125 MHz using a Varian Unity 500 spectrometer in the FT mode. Polysaccharide samples were converted to their sodium salts by passage through a column 1 ϫ 10 cm of Dowex 50 ϫ 8 Na ϩ form. About 15 mg of each polysaccharide sample was dissolved in approximately 0.7 ml of 99.8% D 2 O (Goss Scientific, Ingatestone, United Kingdom) for NMR spectroscopy. The polysaccharide spectra were recorded at 60°C, with suppression of the HOD signal by presaturation. 13 C spectra were recorded with full proton decoupling using the WALTZ sequence (33). Two-dimensional double-quantum filtered COSY (34), TOCSY (35), and NOESY (36) spectra were recorded in the phase-sensitive mode using the pulse programs supplied by the manufacturer. TOCSY spectra were run with a spin-lock field of about 10 kHz and a mixing time of 120 ms; the NOESY spectrum was run with a mixing time of 100 ms. All chemical shifts are relative to internal or external trimethylsilylpropionic acid.
Anticoagulant Action Measured by APTT (Activated Partial Thromboplastin Time)-APTT clotting assays were carried out as described previously (37,38). Normal human plasma (90 l) was incubated with 10 l of a solution of polysaccharide (0 -100 g) and 100 l of kaolin ϩ bovine phospholipid reagent (National Institute for Biological Stand- Effect of Sulfated Polysaccharides on the Inactivation of Thrombin or Factor Xa by Antithrombin or Heparin Cofactor II-These experiments were based on the assay of amidolytic activity of thrombin or factor Xa using chromogenic substrates, as described previously (39,40).
Effect on the Inactivation of Thrombin by Antithrombin-Fifty l of the sulfated polysaccharide solution was mixed with 50 l of 1 unit/ml of purified human antithrombin from Chromogenix AB (Molndal, Sweden) in 0.02 M Tris-HCl buffer (pH 7.4) containing 0.15 M NaCl, and the mixture was preincubated at 37°C for 2 min. Then, 50 l of a 4 NIH units/ml solution of purified human thrombin (Sigma) in the same Tris buffer was added to initiate the reaction. After a 1-min incubation period (inhibition period), 50 l of 1.48 mM chromogenic substrate S-2238 from Chromogenix AB (Molndal, Sweden) was added and the remaining thrombin activity recorded by the absorbance at 405 nm. In the short incubation periods, used to measure activity versus concentration of sulfated polysaccharide, amidolysis was stopped by the addition of 100 l of 50% acetic acid, and the absorbance was measured at 405 nm.
Effect on the Inactivation of Thrombin by Heparin Cofactor II-This assay was essentially as described above except that heparin cofactor II (25 g/ml) from Diagnostica Stago (Asnières, France) instead of antithrombin was added to the incubation mixtures.
Effect on the Inactivation of Factor Xa by Antithrombin-This assay was performed as described for the inhibition of thrombin by antithrombin (see above) except that 4 units/ml purified bovine factor Xa (Chromogenix, Molndal, Sweden) instead of thrombin was added to the mixtures. In addition, the chromogenic substrate employed in this experiment was 2.73 mM S-2222 Chromogenix AB (Mölndal, Sweden).
In the short incubation periods used to measure activity versus concentration of the sulfated polysaccharide no inhibition occurred when thrombin or factor Xa was incubated with antithrombin or heparin cofactor II in the absence of sulfated polysaccharide, nor did inhibition occur when thrombin or factor Xa was incubated with sulfated polysaccharide alone over the range of concentrations tested.

Fucose 4-O-Monosulfate, Fucose 2,4-O-Disulfate, and Fucose 3,4-O-Disulfate Are the Major Species Released by Partial Acid
Hydrolysis of the Fucosylated Chondroitin Sulfate-The fucosylated chondroitin sulfate from the body wall of sea cucumber has a chondroitin sulfate-like core, containing side chains of sulfated ␣-L-fucose linked at the C-3 position of the ␤-D-glucuronic acid. The intact polysaccharide is totally resistant to chondroitin lyase digestion (5,7).
Since fucose forms a glycosidic linkage that is more sensitive to acid than that formed by glucuronic acid or by hexosamine (41), we attempted to defucosylate the polysaccharide using partial hydrolysis with acid. Indeed, this chemical treatment releases sulfated fucose and leaves "acid resistant-fragments," eluted at V o and V t of Bio-Gel P-4 column, respectively (Fig.  1A). The 3 H radioactivity (closed triangles in Fig. 1A) introduced to label the reducing terminal of the polysaccharide still eluted at the V o of the column.
Sulfated fucose was further purified on a Bio-Gel P-2 column (inset to Fig. 1A), and on paper chromatography shows a major component (Fig. 1B) which migrated as glucuronic acid standard on paper electrophoresis at pH 5.0 (Fig. 1C). Upon strong acid hydrolysis it releases almost exclusively fucose (Fig. 1, B and C). Another component, which migrated as disulfated fucose, was also observed on paper electrophoresis (Fig. 1C). Overall, these experiments indicate that the major products released from the fucosylated chondroitin sulfate by partial acid hydrolysis are fucose monosulfate and fucose disulfate.
Characterization of Released Sulfated Fucose by 1 H NMR-The 1 H spectrum of fucose released from the fucosylated chondroitin sulfate by mild acid treatment is shown in Fig. 2.
Eight spin systems consistent with assignment to fucose residues could be identified using the TOCSY and DQCOSY 1 H spectra of the released fucose, and a further five minor anomeric doublets were visible which may also come from ␣-fucose. The DQCOSY and TOCSY spectra gave connectivities for H1 through to H4, and H5 to H6. The coupling constant between H4 and H5 is small for fucose residues, so H4-H5 cross-peaks cannot be seen, but the connectivity can be established by cross-peaks in the NOESY spectrum resulting from the close spatial proximity of H4 and H5. Comparison of the chemical   (28), amino sugar by a modified Elson-Morgan reaction (30), and sulfate by the BaCl 2 /gelatin method (31). Fucose/galactosamine molar ratios were estimated from the integrals of deoxy CH 3 and acetamidomethyl (COCH 3 ) at 18.5 and 25.5 ppm, respectively, in the 13 C NMR spectra (Fig. 7). shifts for each of these fucose residues with shifts for standard, unsulfated fucose (Table I) showed strong downfield shifts of some signals consistent with sulfation at those positions, and less strong downfield shifts attributable to sulfation at the adjoining position. These signals are marked in bold type in Table I.
Three ␣and ␤-fucose systems correspond: those for fucose 4-O-monosulfate, fucose 2,4-O-disulfate, and fucose 3,4-O-disulfate. Approximate integration of the anomeric signals gives the proportions of these three as 49, 20, and 17% of the sample, respectively. A further ␣-fucose 4-O-monosulfate makes up at least 4% of the sample, and a ␤-fucose 3-O-monosulfate at least 6%. It was not possible to identify the partner anomeric forms of these last two species; either they are the non-reducing residues of disaccharides or the anomeric equilibrium strongly favors the form seen over the form not seen.  Fig. 1A and Table II). Thus, this chemical treatment produces only a partially defucosylated polysaccharide.
If the sulfated L-fucose branches released by partial acid hydrolysis were distributed randomly through the chondroitin sulfate core, we would expect that chondroitin ABC lyase digestion of the acid-resistant fragments would produce a wide variety of oligosaccharides with different molecular weights. Instead, mainly disaccharides and small amounts of tetrasaccharides were released, as well a proportion of polymeric chon-droitin lyase-resistant material (Fig. 3). 2 Of the disaccharides produced, 15% are saturated and therefore originated from the non-reducing ends of polysaccharide molecules (Fig. 3, A and  C). The average molecular mass of the fucosylated chondroitin sulfate decreases from ϳ30 kDa in the native polysaccharide to ϳ15 kDa after partial acid hydrolysis and to ϳ10 kDa after subsequent incubation with chondroitin ABC lyase (Fig. 3B). In addition, the 3 H label at the reducing terminal of the polysaccharide (closed triangles in Fig. 3A) is found in the 10-kDa chondroitin lyase-resistant fragments.
Our interpretation of these experiments is that sulfated Lfucose branches susceptible to release by mild acid treatment are located as a cluster at the non-reducing terminal of the polysaccharide. 3 Overall, the combination of partial acid hydrolysis and incubation with chondroitin lyase allows a sequential degradation of the fucosylated chondroitin sulfate. The fucose:glucuronic acid molar ratio decreases in the polysaccharide after partial acid hydrolysis and increases again after digestion with chondroitin ABC lyase (Table II).
Analysis of the fucosylated chondroitin sulfate by anion exchange chromatography on a Mono Q-FPLC column (Fig. 4) shows a homogeneous compound and confirms the high negative charge density of this polysaccharide. Thus, the sea cucumber chondroitin sulfate (Fig. 4B) was eluted from the column at a higher NaCl concentration than mammalian glycosaminoglycans (Fig. 4A). After sequential degradation by partial acid hydrolysis (Fig. 4C) and incubation with chondroitin lyase (Fig. 4D), the resistant fragments showed increasingly wider chromatographic fractions and eluted at lower NaCl concentrations than the native polysaccharide.
Study of the Sequential Degradation of the Fucosylated Chondroitin Sulfate Using 1 H NMR Spectra-The 1 H NMR spectra of the fucosylated chondroitin sulfate, before and after partial acid hydrolysis and degradation with chondroitin ABC lyase, are shown in Fig. 5.
The two signals at 3.39 and 3.59 ppm, attributable to H-2 and H-3, respectively, of non-substituted glucuronic acid residues (42,43) are almost absent in the 1 H NMR spectrum of the fucosylated chondroitin sulfate (Fig. 5A). After mild acid hydrolysis, which partially releases the sulfated fucose branches, the intensity of these two signals increases markedly (Fig. 5B),

FIG. 5. 1 H NMR spectra at 500 MHz of the fucosylated chondroitin sulfate (Panel A), the acid-resistant (Panel B), and the chondroitin lyase-resistant (Panel C) fragments.
All spectra were recorded at 60°C for samples in D 2 O solution. Chemical shifts are relative to internal or external trimethylsilylpropionic acid at 0 ppm. The HOD signal has been suppressed by presaturation. Signals designated by U refer to those produced by ␤-D-glucuronic acid residues. Expansions of the 5.0 -5.9-ppm regions of the spectra are shown in the insets. and decreases after incubation with chondroitin lyase (Fig. 5C). Some L-fucose residues remain after partial acid hydrolysis of the sea cucumber chondroitin sulfate (Fig. 5B), and are concentrated in the polymeric remainder after chondroitin lyase degradation (Fig. 5C). A signal at 5.22 ppm in the spectrum of these preparations is particularly noticeable, but there are several possible fucose anomeric signals between 5.1 and 5.7 ppm. The TOCSY spectrum (not shown) displays crosspeaks for two of these signals, the signal at 5.22 ppm and a further less intense signal at 5.44 ppm. Two H5-H6 cross-peaks can also be identified in the TOCSY spectrum, one more intense than the other. The chemical shifts for fucose residues which can be inferred from these cross-peaks may be tabulated as in Table III. There is no evidence other than relative intensity to connect the spin system from H1 with the H5,H6 pairs.
Comparison of these chemical shifts with those for the acidliberated fucose residues and the standard fucose listed in Table I indicates that the remaining fucose residues after partial acid hydrolysis and chondroitin lyase treatment are unlikely to be sulfated at the 2 or 3 positions, but we have no evidence as to the state of sulfation at the 4-position.
Comparison of the various 1 H NMR spectra in Fig. 5 shows clearly that the fucose residues removed by partial acid hydrolysis are distinguished from those in the resistant polysaccharide. Thus, the signal at 5.60 ppm consistent with the anomeric proton of sulfated ␣-fucose is intense in the fucosylated chondroitin sulfate but almost absent in the acid-resistant and chondroitin lyase-resistant fragments. In contrast, the signal at 5.22 ppm remains in the polymer after both treatments.
This proposition is also supported by the time course experiment of partial acid hydrolysis of the fucosylated chondroitin sulfate at 60°C, followed by 1 H NMR spectroscopy (Fig. 6). The integrals of the signal at ϳ5.60 ppm decrease (Fig. 6C) concomitantly with an increase in the integrals of the signal at ϳ3.39 ppm (Fig. 6D). These two signals are ascribed to anomeric protons of ␣-fucose residues and to H-2 of unsubstituted ␤-Dglucuronic acid in the chondroitin sulfate (42,43), respectively. However, the integral of the peak at ϳ5.22 ppm (Fig. 6E) also attributed to anomeric protons of ␣-fucose is only slightly affected during the time course experiment of acid hydrolysis. This peak is slightly modified during the partial acid hydrolysis (Fig. 6A), possibly as a consequence of the other chemical modifications that occur within the polysaccharide. Therefore, we believe that not all L-fucose residues in the fucosylated chondroitin sulfate are equally susceptible to removal by mild acid hydrolysis; the resistant fucose residues are in some way structurally distinct from those removed. 13 C NMR Spectra-The 13 C-spectrum of the partially defucosylated chondroitin sulfate from sea cucumber (Fig. 7B) resembles for the most part that of standard chondroitin 4-or 6-sulfate (42,43). Signals attributable to fucose residues in the spectrum of the original fucosylated chondroitin sulfate (Fig.  7A) are much reduced after mild acid treatment (for example, the fucose CH 3 at 18.6 ppm). The spectrum of chondroitin backbone residues is also simplified (for example, the C2 signal from N-acetyl-␤-D-galactosamine, split into two separate reso-  Fig. 8 shows our proposition for the structure of the major components found in the fucosylated chondroitin sulfate from sea cucumber. The results presented in this study and in our previous publications (5, 7) suggest a highly heterogeneous polysaccharide. But, some structural features are now very clear.

Summary of the Structural Features of the Fucosylated Chondroitin Sulfate-
Most of the ␤-D-glucuronic acid units from the chondroitin sulfate core are substituted at the O-3 position. This conclusion is based on the near absence of signals at 3.59 and 3.39 ppm in the 1 H NMR spectrum (Fig. 5A), which correspond to H-2 and H-3, respectively of non-substituted glucuronic acid residues (42,43) and in the formation of 2,6-di-O-methyl-glucitol after methylation of the carboxyl-reduced chondroitin sulfate from sea cucumber (7).
Our earlier conclusion that only half of the ␤-D-glucuronic acid residues would be fucosylated (5, 7) was based on the formation of approximately equimolar proportions of 2,3,6-tri-O-methyl-and 2,6-di-O-methylglucitol after methylation of the desulfated and carboxyl-reduced polysaccharide. However, sulfated fucose is highly sensitive to chemical treatment. The desulfation reaction, which requires heating of the polysaccharide solution in dimethyl sulfoxide/methanol (9:1, v/v) at 80°C for several hours, partially defucosylates the molecule (Table  II). 4 In addition, we observed that it is difficult to obtain reliable proportions of methylated derivatives from polysaccharides rich in sulfated fucose residues (9, 10). The current study uses less severe conditions to approach the structure of the fucosylated chondroitin sulfate.
We have also previously suggested the presence of a high proportion of sulfate esterification at position O-3 of the ␤-Dglucuronic acid residues (7). This suggestion was based on a mistaken assumption that fucose was not removed during the desulfation reaction by solvolysis in dimethyl sulfoxide. We have now shown that fucose branches are in fact partially removed from the polysaccharide during this reaction (Table  II). Therefore, high amounts of 3-sulfo-␤-D-glucuronosyl residues do not occur in the sea cucumber chondroitin sulfate. However, the immunoreaction of this polysaccharide with anti-Leu-7 monoclonal antibody (7), which specifically recognizes 3-sulfoglucuronic acid residues, suggests that these residues do occur in the sea cucumber chondroitin sulfate, although in a smaller proportion of disaccharide units than was previously proposed (7). It may be that these residues are located in the chondroitin lyase-resistant region, which is apparently a more heterogeneous portion of the molecule.
Mild acid hydrolysis produces a partially defucosylated chondroitin sulfate, releasing a mixture of mono-and disulfated fucose. The non-reducing portion of the partially defucosylated polysaccharide is totally degraded by chondroitin AC or ABC lyase (Fig. 3). These experiments suggest a cluster of sulfated ␣-L-fucose branches susceptible to release by mild acid treatment at the non-reducing portion of the polysaccharide (Fig. 8).
The fucose branches which resist mild acid hydrolysis are in some way structurally distinct, as suggested by the 1 H NMR spectra (Figs. 5 and 6). These residues are not sulfated at positions 2 and 3 (Table III), but 4-O-sulfation and the presence of disaccharides formed by fucosyl residues cannot be excluded. They are clustered toward the reducing end of the polysaccharide. The precise distinction between acid-releasable and nonreleasable fucose branches was not determined by the methods  Table IV). Comparison between native and chemically modified (desulfated, carboxyl-reduced, and partial defucosylated) polysaccharides suggests that sulfated fucose branches, which are released in the course of mild acid hydrolysis, are responsible for the high anticoagulant activity of the fucosylated chondroitin sulfate. This higher anticoagulant activity is not a specific property of polymers composed of sulfated fucose units, since a sulfated ␣-L-fucan from sea urchin has a low anticoagulant activity (b in Table IV) in spite of its high sulfate/fucose molar ratio (9). This higher activity also cannot be attributed exclusively to the higher anionic nature of the fucosylated chondroitin sulfate, since the partially defucosylated polysaccharide has no discernible anticoagulant activity (a in Table IV) in spite of its elution from a Mono Q-FPLC column at a higher NaCl concentration than heparin (Fig. 4, A and C). Carboxyl reduction of the sea cucumber chondroitin sulfate does not affect its anticoagulant action. 5 Fucosylated Chondroitin Sulfate Accelerates Thrombin Inhibition by Antithrombin and Heparin Cofactor II-Native and carboxyl-reduced fucosylated chondroitin sulfates have inhibitory effect on thrombin amidolytic activity of normal human plasma (Fig. 9A), whereas the desulfated and partially defucosylated polysaccharides have no effect. These results and the APTT assays (a in Table IV) indicate the requirement of sulfated fucose branches for the anticoagulant activity of the fucosylated chondroitin sulfate. Fucosylated chondroitin sulfate is more effective than mammalian dermatan sulfate on inhibition of thrombin activity of normal human plasma (Fig. 9B).  Heparin and some other sulfated polysaccharides have anticoagulant action mediated mainly by plasma protease inhibitors (14,39,40,44,45). Thus, heparin inhibits thrombin, factor Xa, and other coagulation enzymes in the presence of antithrombin. Dermatan sulfate and heparin have an additional inhibitory effect on coagulation through heparin cofactor II (46).
In order to trace a parallel between the anticoagulant actions of mammalian glycosaminoglycans and that of the fucosylated chondroitin sulfate we compared the influence of these sulfated polysaccharides on thrombin and factor Xa inactivation by antithrombin and heparin cofactor II.
Native and carboxyl-reduced fucosylated chondroitin sulfate inhibits the thrombin amidolytic activity in the presence of antithrombin (Fig. 10A), but higher concentrations are required to obtain the same effect as with heparin (Fig. 10B). Mammalian dermatan sulfate has no effect on this assay, as expected.
The effect of the sea cucumber chondroitin sulfate is essentially the same if factor Xa instead of thrombin is the target protein for antithrombin inactivation (Fig. 11A) although slight differences are observed between the native and the carboxylreduced polysaccharide (Fig. 11B). In addition, a marked difference is observed in the concentration of fucosylated chondroitin sulfate for thrombin and factor Xa inhibition in the presence of antithrombin. The IC 50 for fucosylated chondroitin sulfate inhibition of thrombin and factor Xa are ϳ10 and ϳ500 times greater when compared with IC 50 for heparin inhibition, respectively (see Figs. 10B and 11B).
Finally, the sea cucumber chondroitin sulfate also inactivates thrombin in the presence of heparin cofactor II, and again, carboxyl reduction of the polysaccharide does not abolish the inhibitory effect (Fig. 12A). In this case, the inhibitory effect occurs in approximately the same range of concentrations required for mammalian dermatan sulfate or heparin and slight differences were observed between the IC 50 for native and the carboxyl-reduced polysaccharide (Fig. 12B).
The sulfated fucose branches are apparently essential for the anticoagulant action of fucosylated chondroitin sulfate (a in Table IV, Figs. 9A and 12B) and thus these branches could constitute the structural requirement for the binding of the polysaccharide to heparin cofactor II and antithrombin. Nevertheless the IC 50 for fucosylated chondroitin sulfate inactivation of heparin cofactor II was unchanged in the presence of 100 g/ml sulfated fucose released by partial acid hydrolysis of the sea cucumber chondroitin sulfate (Fig. 13). In addition, a sulfated ␣-L-fucan from sea urchin, composed of fucose units sulfated at O-2 and/or O-4 positions (9) has a low anticoagulant activity (b in Table IV). Therefore we believe that the specific spatial array of the sulfated fucose branches in the fucosylated chondroitin sulfate is essential for its anticoagulant action.
There are few compounds which provide a suitable comparison for the anticoagulant activity of the fucosylated chondroitin sulfate from the sea cucumber L. grisea. An anticoagu- FIG. 9. Thrombin inhibition in normal human plasma by different sulfated polysaccharides. Panel A shows the time course of thrombin inhibition. A 50-l solution at 25 g/ml of fucosylated chondroitin sulfate (fucCS) before (q) and after carboxyl-reduction (CR) (E), partial defucosylation (defuc) (Ç), or desulfation (deSO 4 ) (å), mammalian dermatan sulfate (DS) (Ⅺ), or without sulfated polysaccharide (buffer) (f) was mixed with 50 l of normal citrated human plasma and the mixture was preincubated at 37°C for 2 min. Then, 50 l of 4 NIH units/ml purified human thrombin was added to initiate the reaction. After a 1-min incubation period (inhibition period), 50 l of 1.48 mM chromogenic substrate S-2238 was added and remaining thrombin activity was recorded by absorbance at 405 nm. Panel B shows the dependence on the sulfated polysaccharide concentration for thrombin inactivation by normal human plasma. The reaction mixtures were as described in Panel A, except that different concentrations of sulfated polysaccharides were used and amidolysis was stopped at 4 min by the addition of 100 l of 50% acetic acid. NaCl, and the mixture was preincubated at 37°C for 2 min. Then, 50 l of a 4 NIH unit/ml of purified human thrombin in the same Tris buffer was added to initiate the reaction. After a 1-min incubation period (inhibition period), 50 l of 1.48 mM chromogenic substrate S-2238 was added and the remaining thrombin activity was recorded by the absorbance at 405 nm. Panel B shows the dependence on the sulfated polysaccharide concentration for thrombin inactivation. The reaction mixtures were as described in Panel A, except that different concentrations of sulfated polysaccharides were used and the reaction was stopped at 4 min by the addition of 100 l of 50% acetic acid. lant sulfated glycosaminoglycan containing galactosamine, glucuronic acid, and fucose has been extracted from the sea cucumber Stichopus japonicus (47,48). However, the absence of a more detailed description on the structure of the polysaccharide used in these studies makes difficult any comparison with our results. Polymers with sulfated fucose as their primary constituent, the fucoidans of brown algae, also have anticoagulant properties (49). A study of the fucoidan from Fucus vesiculosus (21) established that the major antithrombin activity of this preparation was mediated by heparin cofactor II, with low ability to potentiate antithrombin; however, another species, Ascophyllum nodosum, has yielded a fucoidan in which the balance of antithrombin and heparin cofactor II mediated activities is more in favor of the former (50). Fucoidans from other species also vary in their antithrombin-mediated activity (22,51).
The structures of fucoidans vary from species to species (9), and must give rise to variation in the detailed mechanisms of anticoagulant action; it is also the case that differences in experimental techniques between laboratories make direct comparisons between the results of different studies difficult. CONCLUSION A fucosylated chondroitin sulfate extracted from the sea cucumber body wall exhibits a potent anticoagulant action due to its ability to potentiate inhibition of thrombin by both heparin cofactor II and antithrombin. Comparison between native and chemically modified (desulfated, partially defucosylated, or carboxyl reduced) polysaccharides suggests that the sulfated fucose branches are responsible for the high anticoagulant activity of the fucosylated chondroitin sulfate.
This activity is mediated mainly through heparin cofactor II, but the fucosylated chondroitin sulfate can also potentiate antithrombin; this pattern of anticoagulant activity is similar to that described for some algal fucoidans (22,50), but is in contrast to that of other fucoidan preparations (21,51) and of dermatan sulfate (46) which act only through heparin cofactor II. The potent anticoagulant action of the fucosylated chondroitin sulfate and the possible absence of bleeding side effect make this polysaccharide a promising molecule for testing in experimental thrombosis.
Acknowledgments-We are grateful to Dr. Elaine Gray for practical advice and help with anticoagulant assays, to Dr. Helena B. Nader for the bleeding time assays on scarified rat tail, and Adriana A. Eira for technical assistance.  Fig. 10, except that heparin cofactor II was used instead of antithrombin. The sulfated polysaccharide solutions (25 g/ ml) were either fucosylated chondroitin sulfate (fucCS) before (q) and after carboxyl reduction (CR) (E), mammalian dermatan sulfate (DS) (Ⅺ), or without sulfated polysaccharide (buffer) (f). Panel B shows the dependence on the sulfated polysaccharide concentration for thrombin inactivation. The reaction mixtures were as described in Panel A, except that different concentrations of sulfated polysaccharides were used and amidolysis was stopped at 4 min by the addition of 100 l of 50% acetic acid.
FIG. 13. Effect of sulfated fucose on the activity of fucosylated chondroitin sulfate with heparin cofactor II. The reaction mixtures were as described in the legend of Fig. 12, except that various concentrations of fucosylated chondroitin sulfate (fucCS) (q), sulfated fucose released by partial acid hydrolysis (see Fig. 1) of the fucosylated chondroitin sulfate (Ⅺ), or fucosylated chondroitin sulfate plus a fixed concentration of sulfated fucose (100 g/ml) (E) were used. The incubation time was 100 s.