Biosynthesis of the acetylgalactosamine 4,6-disulfate unit of squid chondroitin sulfate by transsulfation from 3'-phosphoadenosine 5'-phosphosulfate.

Abstract A mucopolysaccharide sulfotransferase utilizing 3'-phosphoadenosine 5'-phosphosulfate (sulfate donor) and endogenous proteinpolysaccharides (sulfate acceptor) was characterized from squid cartilage. This enzyme had the unique property of catalyzing the introduction of sulfate into position 6 of an acetylgalactosamine moiety already bearing a sulfate residue at position 4. The sulfotransferase was separated from endogenous proteinpolysaccharides and purified some 9-fold on DEAE-Sephadex A-50. The enzyme thus purified was effective only when exogenous acceptors were added. Although various mono-, di-, and polysaccharides were active as sulfate acceptors, the presence of a sulfate at position 4 of their acetylgalactosamine residue was essential for the acceptor activity. These observations suggest that an interpretation of the known change in type of chondroitin sulfates with phylogenic and ontogenic development must take into consideration the development of a specific family of sulfotransferases with respect to the site of sulfation.

This enzyme had the unique property of catalyzing the introduction of sulfate into position 6 of an acetylgalactosamine moiety already bearing a sulfate residue at position 4.
The sulfotransferase was separated from endogenous proteinpolysaccharides and purified some g-fold on DEAE-Sephadex A-50. The enzyme thus purified was effective only when exogenous acceptors were added. Although various mono-, di-, and polysaccharides were active as sulfate acceptors, the presence of a sulfate at position 4 of their acetylgalactosamine residue was essential for the acceptor activity.
These observations suggest that an interpretation of the known change in type of chondroitin sulfates with phylogenic and ontogenic development must take into consideration the development of a specific family of sulfotransferases with respect to the site of sulfation.
In view of the cxistc11w of various different, c~holldroitirr sulfato chains, we are q"itc: ill(c~wstrd in osploring their tliffcrcrrl ial synthesis. 'It, is a reasonable assutnl)l,iolr that the enzyme s;ysl,cw~ neccs-;ary for chondroitin sulfalc synlhasis exerts a primary control over tire differential synt~liwis.
II seems possible, for example, tllat 1 here is a family of sulfol ransfcrascs with more or less sharply tlwoloped specificity and I,hey arc responsible for determining * This research was supported in part; by a research grant from 111~: RZinistry of Education of Japan and by a grant from the 'I'akc~la Scierlw Foundation. the site of sull'atc iticorl~or:~tio~~, Cotwitlcrable support \~a8 give11 to this hgwthesis N+IPII Suzuki, 'l'lw~~n, and St,reminger (4) oMained partial scparai iorl or suII'o~r:msfcrase activit,ics, specific for ChS-A', (1hS-(:, and hcparitirr sulfate, from an enzyme preparation solubilized from trcri oviduc~l. IIowever, an alternat.ive hypothesis has been l)rol)osctl by l\lcc~al~ alld J>avidson (5). The authors have reported that a solul~lc sulfotransferasc preparation l'rom chick cmbrpo cartilage synthesized ChS-A in the presence of crrdogcnous arrcl)t,or wl~crc:rs in the presence of added protein-free :iccel)t,or it, s)-nthesieed Cl&C, a result which led to their proposal that the site of sulfa1 ion depends on whet,her t,he awcptor polysaccharide exists as :i 1)rotcin cwnples or :I prot,eilL-free pol~s:lc~c,liaridc.
BcCorc going into the details, it may 1~: i~l~l~rol~riale t,o show in Fig. 1 the structure of tile major rclwating unil, of squid chondroitin sulfate (termed "tyl)c E"). The two sulfates are situated 011 the gulacl,osamine moiety.
'l'hc difference between t8ypc Ii: unit rind the well known A and (' units is in the number and po:'it,ion of the sulfate residues. 111 both A and C, only nne sulfate is sitllatcd on the a~~lac,tosallline moiety; in A, the sulfate is nt position 4, while in C: it, is at Iwsilion 6. As will become evident, the squid cartilage wrries a \-aricty of polysacchnride chains with different proportioir h of tlisullatcd (type E), monohulfntcd (lype I1 and C,), :uld tlonaulfalctl rclEY~ting unils.  For the enzyme preparation, the head cartilage (found in the cephalic region behind the tentacles) was dissected free of soft tissues, cut into slices with a razor, placed in four volumes of ice-cold 0.02 M Tris-HCl, pH 7.0, containing 0.5% Triton X-100, 10% glycerol, and 0.01 M mercaptoethanol, and ground with sea sand in a chilled mortar.
The homogenate was centrifuged at 12,000 x g for 20 min. The clear supernatant fluid was collected, dialyzed for 24 hours against three 10.volume changes of 0.02 M Tris-HCl, pH 7.0, containing 10% glycerol and 0.01 31 mercaptoethanol, and stored at -20" ("crude extract").
Enzyme Assay-Previous studies of Suzuki et al. (2) revealed t'he presence of four types of sulfate residues in chondroitin sulfate prepared from sqrrid cartilage, i.e. sulfate linked to position 4 of acetylgalactosamine moiety (type A sulfate), sulfate linked to Position 6 of acetylgalactosamine moiety (type C sulfate), and sulfates located at position 4 and position 6, respectively, of acetylgalactosamine 4,6-disulfate moiety (hereinafter, the two sulfates will be referred to as t,ype Eq sulfate and type Es sulfate, respectively).
The assay described below is so designed as to permit the quantitative measurement of the incorporation of [3%]sulfate into the four different sites and involves (a) precipitation of labeled polysaccharides with ethanol, (b) degradation of the polysaccharides to disaccharides (ADi-4S, ADi-BS, and ADi-diSn) with chondroitinase-ABC, (c) determination of radioactivity in each of the disaccharide fractions after paper-chromatographic separation, (d) further hydrolysis of the ADi-diSn product to ADi-4S and inorganic sulfate with chondro 6-sulfatase, and (e) determination of radioactivity in each of the sulfatase products after paper electrophoretic separation. The reaction used for the enzymatic determination are summarized in Fig. 1 Controls contained heat-inactivated enzyme. After incubation at 25", the reaction was stopped by immersing the reaction tubes in a boiling water bath for 1 min.
Step a: Pronase-P, 0.4 mg, was added to each solution and the mixture was incubated for 1 hour at 37". To the solution were added 200 ~1 of water and 700 ~1 of a solution of 1 g of potassium acetate and 19 mg of EDTA in 100 ml of 95a/, ethanol to precipitate mucopolysaccharides.
The mixture was left for 30 min in an ice bath. The resultant precipitate was collected by centrifugation at 11,500 x g for 30 min. The precipitate was dissolved in 300 ~1 of water, 700 ~1 of the ethanol solution (see above) were added, and the centrifugation was repeated. The supernatant solution was discarded, and the washing procedure was repeated three times. The precipitate was dissolved in 100 ~1 of water and, in order to measure the total radioactivity incorporated into the mucopolysaccharide fraction, a 20+1 portion of the solution was applied on a disc (2.4 cm in diameter) of filter paper. The disc was dried in an oven at 60" and the radioactivity of the disc was determined in a Horiba liquid scintillation spectrometer, model LS-500 (Horiba Seisakusho, Kyoto), with the solvent system recommended by the manufacturer.
Step b, the remainder (80 ~1) of the mucopolysaccharide solution was dried in a vacuum over PnOs and was then dissolved in 80 ~1 of a solution containing 4 pmoles of Tris-acetate, pH 8.0, 80 pg of bovine serum albumin, and 0.1 unit of chondroitinase-ABC.
The mixture was incubated at 37" for 2 hours.
Step c, a 20-~1 portion of the resulting mixture was applied to filter paper together with 0.1 pmole each of unlabeled ADi-4S, ADi-GS, and ADi-diSn as internal markers.2 Chromatography in Solvent A (see below) for 48 hours separated the disaccharides from one another.
After drying, the disaccharides were located by viewing under ultraviolet light, cut out, placed in scintillation vials, and counted as described above.
Step cl, another 20+1 portion of the chondroitinase digest was subjected to paper electrophoresis (50 min, see below) together with 0.05 pmole each of unlabeled ADi-6S, and ADi-di& as internal markers2 After drying, the region containing AD-diSn was located by viewing under ultraviolet light, cut out, and eluted with water.
The eluate was dried in a vacuum over PZOs and was then dissolved in 40 ~1 of a solution containing 2 pmoles of Tris-acetate, pH 8.0, 40 fig of bovine serum albumin, and 0.012 unit of chondro 6-sulfatase.
The mixture was incubated at 37" for 20 min.
Xtep e, the result.ing mixture was subjected to paper electrophoresis (30 min) to separate inorganic sulfate from ADi-4s. After drying, inorganic sulfate and ADi-4S were visualized, respectively, by spraying with BaC$rhodizonate (14) and by viewing under ultraviolet light, cut out, and counted.
In some experiments (e.g. purification of sulfotransferase by DEAE-Sephadex chromatography, see below), the supernatant solution from boiled (loo', 10 min), centrifuged crude extract (for the preparation, see above) was used as sulfate acceptor. In the experiments with mono-and disaccharides as acceptor, the reaction products were located, after separation by paper chromatography in Solvent A, by placing the chromatogram on x-ray film for 2 days. The radioactive spots were cut out and counted.
Analytical Methods-The following compounds were determined by the indicated methods: protein by the method of Lowry et al. (15) ; uranic acid by the method of Dische (16); hexosamine by the modified Elson-Morgan method (17) ; and sulfate by the method of Dodgson (18) .&-:I--::" 2. Effect of pH on the incorporation of sulfate into four different, sit'es by the crude extract.
The conditions of the experiment were those described under "Enzyme Assay," except for the pH of the bufYers. The buffers used were 5 pmoles of Tris-HCl (pH 7.25 to 9.0) and Tris-maleate (pH 5.75 to 8.6) per incubation mixture.
Incubation was carried out for 1 hour with 0.2 mg (as protein) of crude extract.
No exogenous acceptor was added.
sine 5'-phosphosulfate was transferred to endogenous acceptors within 5 hours. However, some extracts prepared from different squid showed higher incorporation.
The variability of enzyme preparation resulted in the different specific activities given in the individual experiments described below. The labeled product was recovered from the reaction mixture by deproteinization followed by alcohol precipitation (see "Enzyme Assay") and carefully examined for the location of labeled sulfate groups as shown in Table I. The preparation appears to catalyze five types of sulfate transfer to endogenous acceptor, i.e. introduction of type A, type C, type Ed, and type Es sulfate, and of the sulfate which is located in a chondroitinase-resistant polymer fraction. 3 It is evident that apparent velocities are different among the five reactions, the velocity for type Es sulfate being highest.
It should be noted that the assays for these activities contain at least two variables.
The first is the concentration of enzyme and the second the availability of endogenous acceptor.
Furthermore, various other factors such as ionic composition and strength, 3'-phosphoadenosine 5'-phosphosulfate concentration, and pH may also be responsible for the difference in activities observed with the crude extract (see below).
Regardless of these remaining uncertainties, it is clear that the squid preparation has a unique property of catalyzing the introduction of two sulfates (type Ed and type E6) per acetylgalactosamine unit. When a crude extract of cartilage from Gday-old chick embryo was incubated under the same condition, the disulfated unit was not formed in more than trace amounts.4 3 The radioactive sulfate remaining at, the origin of paper chromatogram of chondroitinase digest. It is not clear at present whether this sulfate ester is derived from the same polysaccharide acceptor as that yielding the chondroitin sulfate products or derived from a different acceptor. 4 In a typical experiment, the Y+labeled products obtained with the crude extract gave, upon digestion with chondroitinase-ABC, radioactive ADi-4S, ADi-6S, and ADi-diSE in the radio- The conditions of the experiment were those described under "Enzyme Assay," except that CaClz was omitted and KC1 or NaCl was added to give the indicated concentrations.
Incubation was carried out for 1 hour with 0.18 mg (as protein) of crude extract.
No exogenous acceptor was added. Fig. 2 gives the apparent sulfotransferase activities measured at various pH. Based on experiments with 0.05 M Tris-HCl, the optimal pH for type Ed, type Es, and type C activities was found to be approximately pH 8.2 while that for type A, pH 7.2, or lower.
In contrast, when the buffer was replaced by 0.05 M Tris-maleate, the optimal pH for all these activities shifted to higher pH. In both cases, the activity for type Eg at the optimal pH is higher than the other activities.
A study was made to determine if certain cations elicited an activation of some of the activities.
As shown in Fig. 3, KC1 and NaCl were found to activate the rate of introduction of type Es sulfate without any significant activation of the other three reactions.
From the fact that MgClz and MnCle did not produce any measurable effect on this reaction as well as on the other three reactions, it can be surmised that the chloride ion of itself does not account for the observed activations with sodium and potassium salts. At the optimal concentration of K+ ion (0.2 M), activity ratio of 9:16:1, respectively.
A minor product which was depolymerized by Bacillus keratanus "keratanase" (a hydrolase acting on cornea1 keratosulfate) was also obtained (K. Kimata and S. Suzuki, unpublished observation). the introduction of type Es sulfate proceeded approximately 18 times faster than the other reactions.
CaClz and BaC12, on the other hand, activated all the four reactions, although the concentration at which maximum activation occurred and the extent of activation varied considerably (Fig. 4). Again, the activity for type Es sulfate was highest at the optimal Ca++ or Ba++ concentration.
At present it is not clear whether these activations by K+, Na+, Caf*, and Ba++ are effects on the enzymes or on the endogenous mucopolysaccharide acceptors. B selective activation was also observed when mercaptoethanol or glutathione was added to the reaction mixture (Table II). In this experiment, the crude extract which had been dialyzed against 0.02 M Tris-HCl, pH 7.0, was used. These sulfhydryl agents increased most of the reaction velocities, but the activation of type EJ and type A sulfate incorporation was relatively higher compared with type Eg and type C sulfate incorporation.
When heat labilities of the activities for type Ed and Es sulfate were compared at 47" (Fig. 5), the inactivation of the activity for type Ee sulfate proceeded much more slowly than the activity for type EJ sulfate.
After incubation for 30 min, the former activity had decreased by 40%, while the latter activity had decreased by 90%.
These inactivat,ion phenomena might reflect a direct effect on the enzyme proteins or, alternatively, an effect on the stability of endogenous acceptors. The experiments described below indicate that the latter situation is unlikely.
PuriJication of E&?uljotransjerase-The experiments with the The adsorbent was eluted stepwise with 2 liters each of 0.2 M NaCl and 0.5 M NaCl, and 1.6 liters each of 0.8 M NaCl, 1.0 NaCl, 1.5 M NaCl, and 2 M NaCl in the same Tris buffer. The flow rate was 60 ml per hour, and 20-ml fractions were collected.
Each fraction was checked for the sulfateincorporating ability, with the boiled crude extract (see "Enzyme Assay") as exogenous sulfate acceptor.
The elution pattern is shown in Fig. 6, together with the patterns of ultraviolet (280 m&absorbing material and uranic acid-reacting material. It can be seen that incorporation of sulfate was found in only one peak (Fraction II).
The other fractions had appreciable quantities of protein (as measured by absorption at 280 mp), but showed no sulfotransferase activity. Fractions III through VI contained proteinpolysaccharides, with the protein content progressively decreasing and the uranic acid content increasing (for more detailed analysis, see Table III).
Fraction II was pooled and concentrated to 20 ml by pressure r -L 2.0 z 2 1.6 N dialysis against 2 liters of the Tris-glycerol-mercaptoethanol buffer. If the preparation was tested for sulfate incorporation into the individual sites, it was apparent that about Q-fold purification of EG-sulfotransferase was realized and a large proportion of the other activities was eliminated (Table I). Since the three activities were not recovered from any of the other fractions of chromatography, the apparent separation of Es-sulfotransferase may be interpreted as reflecting the difference in stabihties of sulfotransferase;, i.e. the enzymes responsible for incorporation of type A, type C, and type Ed sulfate seem to be more labile than E8-sulfotransferase and they are inactivated during the purification.
Properties of EG-Xuljotransjerase-The formation of type Es sulfate by the purified preparation had an absolute dependence on added acceptor.
In addition to the endogenous acceptors in the crude extract, several authentic mucopolysaccharide preparations served as acceptors in the reaction.
The relationship   I  I  I   I  II  III TV v VT  a See Fig. 6. b Sum of the disaccharide products bmole)/total glucuronic acid of proteinpolysaccharide sample bmole) X 100. c The sum of the four disaccharides &mole) is assigned an arbitrary value of 100%. d The acceptor concentrations used were 0.1 mM (as glucuronic acid) for Fraction IV, V, and VI, and 0.5 mM for Fraction III; PAPS, 3'-phosphoadenosine 5'-phosphosulfate. between mucopolysaccharide concentration and reaction velocity is shown in Fig. 7. In these experiments, an [%]3'-phosphoadenosine 5'-phosphosulfate preparation with relatively low specific activity (about 3.7 x lo7 cpm per pmole) was added to give a final concentration of 1.33 X lop4 M. It is evident that only mucopolysaccharides containing type A sulfate, i.e. ChS-A and ChS-B, can serve as acceptors.
A commercial preparation of ChS-C exhibited a significant activity.
However, this activity is of questionable significance because the ChS-C preparation has been shown to contain about 10% ChS-A. Fig. 7 also shows that analogous reactions (i.e. the introduction of a second sulfate resulting in formation of an acetylgalactosamine 4,6-disulfate moiety) occurred when acetylgalactosamine 4-s& fate and ADiIS were added as model acceptors.
The products, isolated from the reaction mixtures containing acetylgalactosamine 4-sulfate and ADi-4S, were acetylgalactosamine 4,6-disulfate and ADi-diSn, respectively, as judged by paper chromatography and paper electrophoresis (for the mobilities of authentic samples, see Reference 2). Previously, Suzuki and Strominger (21) reported that 35S could be transferred from [35S]3'-phosphoadenosine 5'-phosphosulfate to acetylgalactosamine-containing oligosaccharides and monosaccharides in the presence of crude extracts from hen oviduct.
The Es-sulfotransferase preparation of squid is different, however, from the oviduct preparation in that it was inactive with such nonsulfated derivatives as acetylgalactosamine and ADi-OS (Fig. 7). Also to be noted is the fact that neither acetylgalactosamine 6-sulfate nor ADi-6S was active as acceptor.
It is clear from these results that the presence of a sulfate residue at Position 4 of hexosamine moiety is essential for the acceptor ability.
In the above studies, the sulfate donor, 3'-phosphoadenosine 5'-phosphosulfate, was added to give a final concentration of Es-sulfotransferase as described under "Enzyme Assay" except that varying amounts of ADi-QS were used as acceptor and the indicated amounts of ADi-di& were added. The velocity, 0, is expressed as radioactivity of [WlADi-diSn formed in 1 hour.
above were catalyzed at almost maximum rates (see below). The K, values for acetylgalactosamine 4-sulfate and ADi-4S were 0.04 rnM and 0.14 mM, respectively.
The reaction with ADilS was inhibited by one of its products, ADi-diS,.
The inhibition by ADi-diSn was strictly competitive with respect to ADi-4S (Fig. 8). The Ki value calculated from Fig. 8 is 0.46 mM for ADi-diSn.
One would speculate, therefore, that there is a regulatory interaction of precursor and product on the sulfotransferase. Properties of Endogerwus Acceptors-As shown in Fig. 6, the uranic acid-containing material in the crude cartilage extract could be separated into several subfractions on DEAE-Sephadex. Experiments were therefore designed to examine these subfractions for their chemical compositions and their behaviors toward Es-sulfotransferase in the hope that a structure-activity relationship may be determined.
The uranic acid-reacting fractions from the DEAE-Sephadex column corresponding to Fractions III through VI were separately pooled and concentrated by pressure dialysis. The apparent yields of the fractions from 670 pmoles (as glucuronic acid) of the crude extract were 25.2 pmoles (Fraction III), 25.4 pmoles (Fraction IV), 99.5 pmoles (Fraction V), and 190 pmoles (Fraction VI).
The chemical analyses (Table III) indicated that Fraction III through VI contain uranic acid, galactosamine, sulfate, and protein, with the protein content progressively decreasing and the ratio of sulfate to uranic acid or galactosamine increasing.
The difference in polysaccharide structure among these fractions was further illustrated by the comparison of disaccharides produced by chondroitinase digestion.
As can be seen in Table III  All the four fractions served as acceptors in the reaction with E,+ulfotransferase.
The results of a typical assay with varying amounts of each acceptor showed that maximum velocities do not differ among the four fractions (Table III).s In the assay conditions used here, an excess of 3'-phosphoadenosine 5'-phosphosulfate (0.133 111~) was used so that the incorporation of [Wlsulfate was enzyme and accept,or-dependent.
The relationship between 3'-phosphoadenosine 5'-phosphosulfate concentration and reaction velocity was also measured under the conditions saturated with the acceptors. The K, values for 3'.phosphoadenosine 5'-phosphosulfate thus obtained are presented in Table III and show t'hat the lower the sulfate content of the acceptor, the greater the affinity of enzyme and 3'.phosphoadenosine 5'-phosphosulfate. Since the acceptor preparations used in these experiments appear to be mixtures of highly hybridized molecules (see "Discussion") the final interpretation of the data in Table III will have to await further work in the purification of endogenous acceptors.

DISCUSSION
The EG-sulfotransferase system described in this work shows a st,rict requirement for acceptors, which is typical of many sulfotransferases.
There is a correlation between the variation of v Tnax and K, for 3'.phosphoadenosine 5'-phosphosulfate and the differences in the molecular form of acceptors, as pointed out in the experiments with the four fractions of endogenous acceptors (Table III).
However, the site of sulfation on hexosamine residues never changes when different acceptors are used. The purified enzyme specifically catalyzes the sulfation of acetygalactosamine 4-sulfate residues, leading to acetylgalactosamine 4,6-disulfate formation, whether acetylgalactosamine 4-sulfate in a proteinpolysaccharide form or in a protein-free form is the acceptor.
It is remarkable that, even when acetylgalactosamine 4.sulfate in a disaccharide form (ADi-4s) or in the monosaccharide form is the acceptor, the enzyme introduces an "extra" sulfate to position 6 of the 4-sulfate residue.
It is most likely therefore that the specificity of this enzyme does not involve recognition of a particular size or a particular monosaccharide sequence of acceptor molecule.
It should be noted that, although the present study has been restricted to the soluble preparation from the cartilage, a large proportion of the squid sulfotransferases appears to be present in the cell in a membrane-bound form. This is indicated by our control experiments in which a disruption of the squid cartilage without the addition of Triton X-100 has resulted to solubilize only 509ib' or less of the sulfotransfe?ase activities solubilized with the aid of the detergent.
At any rate, the detection of such an enzyme as Es-sulfotransferase in an organism that makes an oversulfated chondroitin sulfate (type E) is highly suggestive.
One could assume that the enzyme enables the squid cartilage to synthesize an acetylgalactosamine 4,6-disulfate residue from an acetylgalactosamine 4-sulfate residue in a precursor molecule.
According to our com-6 From the cartilage of 12-day-old chick embryo, a sulfotrsnsferase has been purified, which catalyzes the transfer of sulfate from phosphoadenosine 5'-phosphosulfate to position 6 of an unsubstituted acetylgalactosamine residue of endogenous proteinpolysaccharides.
The activity for sulfation at position 4 of an acetylgalactosamine residue, which is present in the crude enzyme preparation, was completely removed during purification (K. Kimata and S. Suzuki, unpublished observation). 7  is not known at the present time.