Eur. J. Biochem. 166,431 -436 (2987) 0 FEBS 1987

Structural features of sulfated glycans from the tunic of StyeZa plicata (Chordata-Tunicata) A unique occurrence of L-galactose in sulfated polysaccharides

Paulo A. S. MOURAO' and Arthur S. PERLIN'

Departamento de Bioquimica, Centro de CiEncias da Saude, Universidade Federal do Rio de Janeiro Department of Chemistry, McGill University, Montreal

(Received February 6,1987) - EJB 87 0131

The sulfated polysaccharides in the tunic of Styela plicata occur as three fractions that differ markedly in molecular mass and chemical composition. The high-molecular-mass fraction has a high galactose content and a strong negative optical rotation while the low-molecular-mass fractions have a higher proportion of amino sugars and glucose. The galactose occurs in these polysaccharides entirely in the L-enantiomeric form. Although L-galactose is a constituent of several polysaccharides, this is the first report of sulfated polysaccharides that contain high amounts of L-galactose, and that lack the D enantiomorph of this sugar. Furthermore, the structure of the high-molecular-mass fraction, which is composed mainly of a core of a-L-galactopyranose residues, sulfated at position 3 , linked glycosidically though position 1 +4, and with non-sulfated L-galactopyranose non-reducing end-units, is unique among other previously described sulfated glycans. These data are of considerable interest as they show an unusual example of possible variants of polyanionic glycans with structure function in living tissues.

Connective tissues of invertebrates and of vertebrates have been compared in relation to function and chemical composi- tion. In our laboratory we have looked for structures in in- vertebrates that showed the presence of sulfated glycans with functions presumably similar to those of cartilaginous glycosaminoglycans. Interestingly, we isolated sulfated gly- cans from the tunic of various species of ascidians (Chordata- Tunicata), which contain galactose, glucose and amino sugar [l]. These polysaccharides are distinct from the glycosamino- glycans of other animal tissues and also from the sulfated polysaccharides of marine algae.

In the tunic of the species Styela plicata these polymers occur as three main fractions, which are markedly distinct in their molecular mass and chemical composition [2]. The high- molecular-mass fraction contains a high proportion of galactose, whereas the other two fractions of low molecular mass contain a higher proportion of amino sugars and glucose.

On more detailed examination it has now been found that the galactose occurs in these polysaccharides entirely in the L-enantiomeric form. Although L-galactose is a constituent of several polysaccharides [3 - 61, this is the first report of sulfated polysaccharides that contain high amounts of L-galactose and that lack the D enantiomorph of this sugar. In the high-molecular-mass fraction, which is primarily a galactan sulfate, the galactose undoubedly exists mainly as a-L-galactopyranosides, linked glycosidically though position 1 -t4 and sulfated at position 3.

Our data are of considerable interest as they show an unusual example of possible variants of polyanionic glycans with structural function in living tissues. From our results it

Correspondence to P. A. S. MourBo, Centro de CiEncias da Saude, Universidade Federal do Rio de Janeiro, Caixa postal 68041, BR-21910 Rio de Janeiro, Brazil

is possible to speculate that such biological function in animal tissues may be served by sulfated polymers showing consider- able variations as observed among keratan sulfate, other glycosaminoglycans and ascidian polysaccharides.

MATERIALS AND METHODS

Materials

Tunicates were collected from Guanabara Bay (Urca) in Rio de Janeiro. Bovine liver a-D-galactosidase, Aspergillus niger p-D-galactosidase, baker's yeast a-D-glucosidase, al- mond p-D-glucosidase, A . niger D-glucose oxidase, Dactylium dendroides D-galactose oxidase, porcine liver L-fucose dehy- drogenase, peroxidase and L-galactose were purchased from Sigma Chemical Company (St Louis, MO, USA); l-methyl- 1 -phenylhydrazine from Aldrich Chemical Company (Milwaukee, WI, USA) and o-tolidine from Baker Chemical Company (Phillipsburg, NJ, USA).

Extraction and fractionation of the sulfated glycans from the ascidian tunic

Styela plicata and fractionated, as previously described [2]. The sulfated glycans were extracted from the tunic of

NMR spectra

NMR spectra ('H and I3C) were recorded with D20 as the solvent (following prior H-D exchange) at 90°C, with a Bruker WH-400 spectrometer. The chemical shifts (6) were measured with respect to internal sodium 4,4-dimethyl-4- silapentane-1-sulfonate for 'H spectra and dioxane for l3C spectra.

432

Methylation of the ascidian polysaccharides Methylation of the ascidian polysaccharides was per-

formed by the Hakomori method [7], with the modifications introduced by Conrad [8]. The properties of the sulfated polysaccharides after four methylations indicated that the reaction was almost quantitative, and did not remove sulfate groups. This is supported by the fact that the products were chloroform-insoluble, and that the comparison of the infrared spectra of the original and methylated polysaccharides showed that the OH stretch band was diminished, with a concomitant increase in intensity of the CH stretch bands at 3000-2800 cm-l, with no apparent loss in the bands cor- responding to sulfate. The desulfated polysaccharides re- quired three methylations to obtain a product without OH stretch bands in the infrared spectrum, and was completely soluble in chloroform, as expected.

The methylated polysaccharides were hydrolysed with 4.0 M trifluoroacetic acid for 6 h at lOO"C, reduced with borohydride and the alditols were acetylated with 1 : 1 acetic anhydride/pyridine [9]. The alditol acetates from the methyl- ated sugars were dissolved in chloroform and analysed with a model 4000 Finnigan GLC-MS unit. GLC was performed with a column coated with OV-225 (Chrompack, Canada). Injections were made in the splitless mode at 50°C in order to obtain the 'Grob solvent effect' [lo], and the elution was quickly programmed (40"C/min) to 182°C. The carrier gas was helium with a linear velocity of 22 cm/s.

Isolation of galactose and glucose from the ascidian glycans A solution of the ascidian glycans (50 mg) in 5 ml 6.0 M

trifluoroacetic acid was heated at 100°C for 4 h. After evapo- ration of the acid in a rotary evaporator, the mixture was applied to Whatman no.1 paper and chromatographed in ethyl acetate/pyridine/water (8:2: 1, v/v/v) for 24 h. A strip- guide of the chromatogram was developed with silver nitrate, and the separate regions corresponding to galactose and glucose were eluted with distilled water. Each eluate was con- centrated to about 1 ml, and the concentration of hexose was estimated by the DuBois et al. reaction [ll].

Enzymatic degradation

a ) Oxidations with o-galactose oxidase. Varying amounts of the galactose isolated, as well as of D and L-galactose, were incubated with 2 units D-galactose oxidase [12], peroxidase (4 units) and the chromogen o-tolidine (0.1 mg) in 0.5 mlO.05 M sodium acetate buffer (pH 7.0). The reaction mixtures were incubated at 37 "C and the absorbance was measured periodi- cally at 425 nm. In some experiments 100-pl samples were removed from the incubation mixtures, boiled for 2 min and applied to Whatman no. 1 paper. After descending chroma- tography in ethyl acetate/pyridine/water (8 : 2: 1, v/v/v) for 24 h, the chromatograms were stained with silver nitrate.

b) Oxidation with o-glucose oxidase. Varying amounts of the glucose isolated, and of D-glucose, were incubated with A . niger D-glucose oxidase (2 units), peroxidase (4 units) and the chromogen o-tolidine (0.1 mg) in 0.5 ml 0.05 M sodium ace- tate buffer (pH 7.0). The reaction mixtures were incubated at 37°C and the absorbance was measured periodically at 620 nm.

c) Oxidation with L-jiucose dehydrogenase. Varying amounts of D or L-galactose were incubated with 0.2 unit porcine liver L-fucose dehydrogenase [13] and 2.5 pmol

NAD' in 2 ml 0.01 M glycine/NaOH buffer (pH 8.0). The reaction mixtures were incubated at 37 "C for different times, and the formation of NADH was followed by the measure- ment of absorbance at 340 nm.

d) Incubation with a-D and P-o-glucosidase. About 2 mg desulfated polysaccharides from S. plicata was incubated with 0.2 unit baker's yeast a-D-glucosidase, or with 0.2 unit almond /3-D-glucosidase in 0.5 ml 0.05 M sodium acetate buffer (pH 5.0). After incubation for 8 h at 37"C, the amount of glucose liberated was measured by oxidation with D-glucose oxidase.

The controls for all of the enzymatic reactions consisted of incubation mixtures lacking substrate, or containing heat- inactivated enzyme in place of active enzyme.

Desulfation of the ascidian glycans Desulfation of the sulfated glycans was performed as de-

scribed by Nagasawa et al. [14]. About 200mg sulfated glycans in 20 ml water was mixed with 2 g (dry weight) of Dowex 50-W (H', 200-400 mesh). After filtration and neutralization with pyridine, the solution was lyophilized. The pyridinium salt was dissolved in 20 ml dimethylsulfoxide/ methanol (9: 1, v/v), heated at 80°C for 6 h, and the desulfated product was dialyzed against 3 1 distilled water. The degree of sulfation was estimated from the molar ratio of sulfate/total sugar, as observed in the infrared spectrum.

Preparation of galactose-methylphenylhydrazone

The galactose-methylphenylhydrazone was prepared as described by Hirst et al. [15]. 1 ml 1-methyl-1-phenylhydrazine reagent (250 mg I-methyl-I-phenylhydrazine in 1 ml absolute ethanol and 30 pl glacial acetic acid) was mixed with the sugar sample (about 2 mg) contained in 1 ml water. The mixture was kept at 33 "C for 12 h with occasional shaking, then cooled to 0°C for 9 h. The crystals formed were recovered, washed three times with ice-cold ethanol (5 ml) and dried at 100°C for 30 min and their melting points were determined.

Chemical analysis Total hexose was measured by the phenol/sulfuric acid

method of DuBois et al. [Ill. After acid hydrolysis (6.0 M trifluoroacetic acid, at 100°C for 4 h), total hexosamine was measured by a modified Elson-Morgan reaction [16], and sulfate by the BaCl,/gelatin method [17]. The percentages of the different hexoses in the acid hydrolysates were estimated by gas-liquid chromatography of the acetylated, borohydride- reduced hexoses [9], and the relative proportions of glucos- amine and galactosamine by means of a Beckman amino acid analyzer. The percentages of the different hexoses were also estimated by densitometry, after separation by descending chromatography on Whatman no. 1 paper for 24 h in ethyl acetate/pyridine/water (8 : 2: 1, v/v/v), and visualization with silver nitrate.

Other methods The molecular masses of the sulfated glycans were deter-

mined by polyacrylamide gel electrophoresis, as described by Hilborn and Anastassiadis [18]. Densitometry was performed using a Quick Scan densitometer (Helena Laboratories, Beaumont, TX, USA). Infrared spectra were recorded with a Perkin-Elmer infrared spectrophotometer (model 298), and

433 U Table 1. Chemical composition, specific optical rotation and average A

molecular mass of the different sulfated glycans from the tunic of S. plicata

Frac- Molar ratio Sulfate/ [a];'""" Average tion total molecular

Gal Glc Man HexN sugar mass

mol/mol kDa B

Total 0.49 0.32 0.04 0.15 0.95 - 79 F-1 0.82 0.14 0 0.04 0.66 -119 >I000 F-2-A 0.20 0.57 0 0.23" 0.73 - 32 20 F-2-B 0.31 0.27 0.12 0.40" 0.67 - 13 8

a About 85% of the hexosamine is glucosamine. C

a-H-1 p-H-1

I , I , , I I I ,

5.0 4 .O 3.0 2.0 PPm

Fig. 1. Proton magnetic resonance spectra at 400 MHz of fraction F-1 (A),chemicallydesulfatedF-1 (B),F-2-A ( C ) andF-2-B ( D ) . Solvent, deuterium oxide; 90 "C

optical rotations were determined with a digital polarimeter (Japan Spectroscopic Co. Ltd, Tokyo, Japan).

RESULTS AND DISCUSSION Chemical analysis of the ascidian polysaccharides

The sulfated glycans extracted from the ascidian tunic were purified by DEAE-cellulose column and fractionated by Sepharose CL4B and Sephadex G-200 gel chromatography, as previously described [2]. Table 1 shows the chemical analy- sis, the optical rotation and average molecular mass of the three fractions of sulfated polysaccharide obtained from the tunic of S. plicata. F-1 has a high galactose content, an unusually high molecular mass and a strong negative optical rotation. The hexosamine content increases as the molecular

- . . . . . l , l , l , l , l , l , l , ~ . ~

w o w w 70 o a s a UI 1 to PPm

Fig. 2. 13C magnetic resonance spectra at 400 MHz of fraction F-1 ( A ) , CH subspectrum (B), CH2 subspectrum ( C ) and chemically desulfated F-1 ( D )

mass decreases. F-2-A has the highest content of glucose, whereas mannose is present only in the F-2-B fraction. All fractions have a high content of sulfate ester.

'H and 13C-NMR studies of the ascidian polysaccharides

Fig. 1 shows the 400-MHz 'H-NMR spectra of the various fractions of ascidian polysaccharides. Signals clearly attrib- utable to a-anomeric protons were identified in the 'H-NMR spectra of F-1 and chemically desulfated F-1 (Fig. 1 A, B), whereas the spectra of fractions F-2-A and F-2-B show a preponderance of signals attributable to p-anomeric protons (Fig. 1 C, D). The signals of acetyl groups, which resonate at approximately 6 = 2.0 ppm, increase in intensity from frac- tion F-1 to F-2-B, as expected from the data of chemical analysis (Table 1). The sharp signal at about 6 = 1.8 ppm in Fig. 1 A is attributable to the acetate anion [19].

The 13C-NMR spectra of fraction F-1 (Fig.2), before and after chemical desulfation, show that the 13C nuclei of F-I resonate at 6 = 72.7-69.6ppm, that two major signals, attributable to non-substituted carbon-6, resonate at 6 = 62.2 ppm and 61.4 ppm and that a large variety of signals attributable to glycosidically-linked or sulfated secondary carbons resonate at 6 = 80.1 - 76.5 ppm.

The main change that occurs after desulfation of F-I (Fig. 2 D) is an upfield displacement by approximately 8.0 ppm of signals 7, 9 and 10, with a concomitant integral increase mainly of signals 12 and 13. These results are in

434

Table 2. I3C chemical shgt ( 6 ) data for desulfated F-I and reference methyl galactosides

Fraction c-1 c-2 to c-5 C-6

PPm

Desulfated F-1 101.1 -98.9' 72.7-69.6'. 62.2-61.4' a-Galactopyrano-

/I-Galactofurano- side [23, 241 100.1 71.6-69.2 62.2

side 1251 109.9 84.7 -71.7 63.6

a Major signals. Not including the main signals attributable to glycosidically

linked secondary carbons, which resonate at 80.1 - 76.5 ppm.

accord with previous 13C-NMR studies of sulfated sugars, showing downfield displacements of approximately 8.0 ppm, with respect to the parent sugars, of the shifts of carbons atoms carrying a sulfate group [20-221. However, since the pattern of the displacements is obscured by overall complexity of the spectra, these results do not afford reliable information about the site of sulfation of this polysaccharide.

Similarly the complexity of the I3C-NMR spectra of F-1 in the region of substituted secondary carbons does not permit the identification of the main glycosidically substituted carbons. However, the spectra unequivocally demonstrate that because carbon-6 resonates only in the vicinity of 6 = 62-61 ppm, it is neither sulfated nor substituted by a glycosidic linkage. In addition, the results of Fig. 2 show that the 13C nuclei of F-1 are more strongly shielded than are the corresponding nuclei of a or P-galactofuranosides, which resonate at 84.7-71.7 ppm [23, 241, whereas the 13C nuclei of F-1 resonate in the range expected for an o! or P- galactopyranoside (71.6 - 69.2 ppm). These comparative data are summarized in Table 2.

The 13C-NMR spectra of fraction F-2-A (Fig. 3) show that 11 peaks have chemical shifts in common with the 13C-NMR spectra of F-1. The main differences are that about 1/4 of the carbon-6 signals correspond to substitution by a glycosidic linkage or a sulfate group (see CH2 spectrum in Fig.3C) in F-2-A, and that the peak of the acetyl group (peak 20) is more intense in F-2-A, as expected from the chemical analysis (Table 1).

Methylation studies

B I

4 & 1 1 1 1 1 1 1 1 1 t 1 1 1 1 1

100 80 80 70 60 50 40 30 20 ppm

Fig. 3. ' 3 C magnetic resonance spectra at 400 MHz of fraction F-2-A ( A ) , and subspectra of CH ( B ) , CH2 (C) and CH3 ( D )

Table 3. Methylation analysis of the ascidian polysaccharides Retention time, tR, referred to 2,3,4,6-tetra-O-rnethylglucose. n. d. = not detected

Methylated sugars tR Molar ratios (as alditol acetates)

F-1 desulfated F-I F-2-A

mol/mol

2,3,4,6-Gl~ 2,3,4,6-Gal

2,4,6-Gal 2,3,6-Gal 3,4,6-Gal

2,6-Gal

3,6-HexN 3,4-HexN 6-HexN

2,4,6-Gk

2,3,6-Glc

2,6431~

1.00 0.02 0.04 1.15 0.20 0.21 1.63 0.02 0.07 1.80 0.09 n.d. 1.91 0.07 0.48 1.93 0.02 n.d. 2.00 0.07 0.12 2.70 0.50 0.04 2.79 0.01 0.04 3.00 n.d. n.d. 3.10 n.d. n.d. 3.31 n.d. n.d.

0.04 0.10 0.05 0.02 0.12 n.d. 0.27 0.06 0.1 1 0.03 0.15 0.05

The methylation studies of the ascidian polysaccharides indicate that fraction F-1 is constituted mainly of a carbo- hydrate core of galactose linked glycosidically though position 1+4 and sulfated at position 3. That is, 2,3,6-tri-0- methylgalactose is the main methyl ether derivative obtained from desulfated F-1, whereas 2,6-di-O-methylgalactose is the predominant methyl ether derived from sulfated F-1 (Table 3). The formation of 2,3,4,6-tetra-O-methyl derivates of both galactose and glucose indicates the presence of non-sulfated end-groups in this ascidian polysaccharides. The detection of a small proportion of 2,4,6-tri-O-methyl and 3,4,6-tri-O- methyl derivates may indicate minor sites of sulfation and/or small proportions of other types of glycosidic linkages.

The results of the methylation studies on F-1 agree with the 13C-NMR spectra (Fig. 2) in showing the absence of sub- stitution at carbon-6. Furthermore, analysis of the products formed from the periodate-oxidized ascidian polysaccharides

[2], i. e. high yields of glycerol and, after chemical desulfation, an increased proportion of threitol, is in agreement with the above-proposed structure for the main part of F-1. However, it is not possible from these data to elucidate the position of branch points or the distribution of the minor glucose constituent among the galactose residues in the polysaccha- ride.

Fraction F-2-A produced different proportions of the 0-methyl derivatives from the neutral sugars, when compared with F-1 (Table 3), and also gave three types of amino sugar derivatives (3,6-di-O-methyl, 3,4-di-O-methyl and 6-mono-O- methylhexosamines). The formation of this 3,4-di-O-methyl derivative indicates that the substituted carbon-6, detected by the 13C-NMR spectrum (Fig.3), may be among the amino sugar residues. Moreover, few other conclusions can be

435

obtained from the methylation on NMR studies of F-2-A, because this fraction probably has a more heterogeneous and complex structure than F-1.

L-Galactose in the ascidian glycans

The strongly negative specific rotation (- 119") of F-1 is especially noteworthy. As this fraction consists largely of galactose, its optical activity is compatible with two possible types of residue, namely, P-D-galactofuranosyl and a - ~ - galactopyranosyl. That is, the specific rotations of methyl P-D-galactofuranoside and a-L-galactopyranoside are - 1 12" and - 179" respectively. However, the first of these pos- sibilities is inconsistent with the fact that the I3C nuclei of F-1 are far more strongly shielded than are the corresponding nuclei of P-galactofuranosides (Table 2), whereas they res- onate in the range expected for an a-L- (or D-)galacto- pyranoside. The latter structural type is favored unequivo- cally by the finding that F-1 (as well as the other fractions) contains L-galactose.

In the experiments performed to establish the identity of the sugar constituents of the ascidian polysaccharides, the original, unfractionated, material was used. Therefore, the results obtained apply equally well for each of the fractions isolated subsequently.

The major sugar component, isolated from ,the total hydrolysate, was found to cochromatograph with galactose in three solvents, and after borohydride reduction its hexaace- tate had the same GLC retention time as galactitol hexaace- tate. Its specific rotation of -79", as compared with that of - 80" recorded for a mutarotated solution of authentic L-galactose, showed that the sugar is the L enantiomer. More- over, it afforded a crystalline l -methyl-l -phenylhydrazone of m.p. 185"C, undepressed by mixing with the hydrazone of L-galactose (m.p. 186°C).

Experiments with D-galactose oxidase (Fig. 4A, B) served not only to show that the isolated sugar was resistent to this enzyme, but also to test for the possible presence of some D-galactose. The measurements by the chromagen method (Fig.4A) included a control in which the rate of oxidation of the D-galactose standard was compared with that in the presence of the ascidian galactose specimen, to ensure that the latter contained no inhibitor of the enzyme. Consequently the fact that no D-galactose was detected colorirnetrically (which should have been feasible at a level of 1 pg), clearly demonstrates that at least 99% of the galactose produced by acid hydrolysis of the ascidian polysaccharides is the L isomer. The resistance of the desulfated polysaccharides to the action of a- or P-D-galactosidase, measured by the D-galactose oxidase method, is another indication of the absence of the D-galaCtOSe in these polysaccharides (Table 4). In our previous paper [2] we reported a small release of reducing sugar after incubation with P-D-galactosidase (1 S%), which may be attributed to a small contamination of D-glucosidase in the P-D-galactosidase preparation.

Another observation worth noting is that the sugar was oxidized by L-fucose dehydrogenase. Its rate of reaction, although slower than that of L-fucose, was the same as the rate for authentic L-galactose (Fig. 4C), an additional indica- tion that the galactose isolated was exclusively of the L configu- ration.

The glucose present in the hydrolysate was the D isomer, as shown by the fact that it was oxidized completely by D-glucose oxidase. In the desulfated polysaccharides at least 20% of this constituent hexose occurs in the form of 8-D-glucopyranosyl

0 20 LO 60 120 Time lminl

0.81 fa

n C &.-xx-x-x-x-xx-x-x~+x

0 20 LO 60" 120 Time lminl

Fig. 4. Oxidation of the galactose, obtained from the hydrolysates of the ascidian polysaccharides by Dgalactose oxidase and L-fucose dehydrogenase. The incubation conditions are described in Materials and Methods. (A) The rate of oxidation of galactose by D-galactose oxidase. The incubation mixtures contained 50 pg and 100 pg stan- dard D-galactose (0-O), or 100 pg galactose obtained from S. plicata (0-0). The arrow indicates the time at which 50 pg standard D-galactose was added to the incubation mixture containing galactose from the ascidian polysaccharides. (B) The oxidation of galactose by D-galactose oxidase, as determined by paper chromatog- raphy. The incubation mixtures are the same as those described in Materials and Methods, except that they contain 100 pg standard D- galactose ( 3 , 2) or 100 pg galactose prepared from S. plicata (3, 4). At zero time (1 ,3) , or after 6 h incubation (2,4), 100-pl samples were removed from the incubation mixtures, boiled for 2 min and applied to Whatman no. 1 paper. After descending chromatography in ethyl acetate/pyridine/water (8 : 2: 1, v/v/v) for 24 h the chromatogram was stained with silver nitrate. (C) The oxidation of galactose by L-fucose dehydrogenase. The incubation mixtures contained 100 pg L- galactose (0-O), 100 pg D-galactose ( x- x ) or 100 pg galactose obtained from the ascidian polysaccharides (0-0)

Table 4. Degradation of the chemically desuuated ascidian polysac- charides with galactosidases and glucosidases The amounts of galactose or glucose liberated by the enzymes were measured by oxidation with D-galactose or D-glUCOSe oxidase (see Materials and Methods) ~~

Enzyme ~ ~ ~

D-Galactose or D-glucose liberated

% of total a-D-Galactosidase < I P-D-Galactosidase < I a-D-Glucosidase < I /?-D-Glucosidase 20

end-units, based on the amount of D-glucose liberated upon incubation with almond P-D-glucosidase (Table 4).

Overall, then, glycans of the ascidian tunic are unique among known sulfated polysaccharides, in that their constitu- ent galactose residues occur only in the L form In fraction

Two other polysaccharides that contain L-galactose are not sulfated. In snail galactan [26 - 281, the L-galactose is accompanied by at least three times as much D-galaCtOSe, whereas in the mucilage of flax seed [29] it co-occurs with D-galacturonic acid (as well as L-rhamnose and D-xylose).

436

OR OR

R = H oTSO4 w--- OR

(2b)

Fig. 5. Structure of the galactose-rich sulfatedpolysaccharides reported in living tissues in comparison with fraction F-I of the ascidian polysaccharides. The repeating disaccharide unit of keratan sulfate (l), the units of algal carrageenans (2) and the main structural features of fraction F-1 obtained from the tunic of S. plicata (3)

F-1, which is primarily a galactan sulfate, they undoubtedly exist as a-L-galactopyranosides.

CONCLUSION

The structural features of the ascidian polysaccharides reported in this paper may be relevant for a phylogenetic comparison among galactose-rich sulfated polysaccharides having a structural function in living tissue. The repeating disaccharide unit of keratan sulfate, a galactose-rich sulfated polysaccharide of mammalian tissues [30], is P-D-galaCtO- pyranose 1 -+blinked glycosidically to N-acetyl-D-glucos- amine 6-sulfate (1, Fig. 5), whereas the algal carrageenans (2, Fig.5) present a more heterogeneous structure [3]. The carrageenans are linear chains of P-D-galactopyranose resi- dues (A units) linked glycosidically through position 1 +3 to a-galactopyranose (B units). The B units can occur in either D or L form [3] or can be wholly or partly converted to 3,6- anhydro forms (2 b, Fig. 5). The sulfate esters may occur at position 2,4 or 6 of the galactose residues in the carrageenans.

Ascidian glycans appear to be 'intermediate' in nature with respect to these other polysaccharides. Like keratan sulfate they contain hexosamines (which are especially promi- nent in F-2-A and F-2-B), whereas they resemble the algal polysaccharides in having residues based on a-L-galacto- pyranose. Moreover, the ascidian polysaccharides are unique among known sulfated polysaccharides in living tissues. Frac- tion F-1 (3, Fig. 5), which is primarily a galactan sulfate, is a branched polysaccharide, composed mainly of a core of a-L- galactopyranose residues, sulfated at position 3, linked glycosidically through positions 1 and 4 (3 a, Fig. 5), and with non-sulfated L-galactopyranose non-reducing end-units (3 b, Fig. 5). Also, it is the first example of a sulfated polysaccharide that contains L-galactose, although not its D enantiomorph.

The authors wish to express their appreciation to Izaias G. Bastos for his help in the fractionation of the ascidian polysaccharides and to Dr Lawrence Hogge (Prairie Regional Laboratory, National Research Council, Saskatoon, Canada) for the GLC-MS analysis. This work was supported by grants from Natural Sciences and Engineering Research Council of Canada (NSERC), Financiadora de Estudos e Projetos (FINEP), Conselho Nacional de Desenvolvimento Cientifico e Tecnolbgico (CNPq) and Fundo de Incentivo a Pesquisa Tkcnico- Cientifico (FIPEC).

REFERENCES 1.

2.

3.

4.

5. 6.

7. 8.

9. 10. 11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26. 27.

28.

29. 30.

Albano, R. M. & MourHo, P. A. S. (1983) Biochim. Biophys. Acta

Albano, R. M. & Mouriio, P. A. S . (1986) 1. Biol. Chem. 261,

Painter, T. J. (1983) in The polysaccharides, vo1.2 (Aspinall, G. D., ed.) pp. 195-285, Academic Press, New York.

Percival, E. & McDowell, R. H. (1967) Chemistry andenzymology of marine algalpolysaccharides, pp. 127 - 175, Academic Press, New York.

760,192-196.

758 - 765.

Rees, D. A. & Conway, E. (1962) Biochem. J. 84,411 -416. Molchanova, V. I., Ovodova, R. G., Ovodov, Y. S., Elkin, Y.

N. & Santana, V. F. (1985) Carbohydr. Res. 141,289-293. Hakomori, S. (1964) Biochem. J . (Tokyo) 55,205-208. Conrad, H. E. (1972) in Methods in carbohydrate chemistry

(Whistler, R. L. & BeMiller, J. N., eds) vol. 6, pp. 361 -364, Academic Press, New York.

Kircher, H. W. (1960) Anal. Chem. 32,1103-1106. Grob, K. & Grob, K. Jr (1974) J . Chromatogr. 94,53-64. DuBois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A. &

Avigad, G., Amaral, D., Asensio, C. & Horecker, B. L. (1962) J .

Schachter, H., Sarney, J., McGuire, E. J. & Roseman, S. (1969)

Nagasawa, K., Inove, Y. & Yamagata, T. (1977) Carbohydr. Res.

Hirst, E. L., Jones, J. K. N. & Woods, E. A. (1947) J . Chem. SOC.

Rondle, C. J. & Morgan, W. T. J. (1955) Biochem. J. 61, 586-

Saito, H., Yamagata, T. & Suzuki, S. (1968) J. Biol. Chem. 243,

Hilborn, J. C. & Anastassiadis, P. A. (1969) Anal. Biochem. 31,

Perlin, A. S., Casu, B., Sanderson, G. R. &Johnson, L. F. (1970)

Archbald, P. J., Fenn, M. D. & Roy, A. B. (1981) Carbohydr.

Honda, S . , Yuki, H. & Takiura, K. (1973) Carbohydr. Res. 28,

Honda, S., Yuki, H. & Takiura, K. (1974) J . Biochem. (Tokyo)

Dorman, D. E. & Roberts, J. D. (1970) J . Am. Chem. SOC. 92,

Perlin, A. S., Casu, B. & Koch, H. J. (1970) Can. J. Chem. 48,

Ritchie, R. G. S., Cyr, N., Korsch, B., Koch, H. J. & Perlin, A.

Bell, D. J. & Baldwin, E. (1941) 1. Chem. SOC. 125-131. Correa, J. B. C., Dmytraczenko, A. & Duarte, J. H. (1967)

Bretting, H., Jacobs, G., Benecke, I., Konig, W. A. & Thiem, J.

Anderson, E. & Lowe, H. J. (1947) J. Biol. Chem. 168,284-297. Rod& L. (1981) in The biochemistry of glycoproteins and pro-

teoglycans (Lennarz, w . J., ed.) pp.267- 371, Plenum Press, New York.

Smith, F. (1956) Anal. Chem. 28, 350-354.

Biol. Chem. 237,2736-2743.

J. Biol. Chem. 244,4785-4792.

58,47- 55.

1048.

589.

1536- 1542.

51 -55.

Can. J . Chem. 48,2260-2268.

Res. 93, 177-190.

150 - 153.

76,209-211.

1355 - 1361.

2596 - 2606.

S . (1975) Can. J. Chem. 53,1424-1433.

Carbohydr. Res. 3,445 -452.

(1985) Carbohydr. Res. 139,225-236.