Identification of Physiologically Relevant Substrates for Cloned Gal: 3- O -Sulfotransferases (Gal3STs) DISTINCT HIGH AFFINITY OF Gal3ST-2 and LS180 SULFOTRANSFERASE FOR THE GLOBO H BACKBONE, Gal3ST-3 FOR N -GLYCAN MULTITERMINAL Gal (cid:1) 1,4GlcNAc (cid:1) UNITS AND 6-SULFOGal (cid:1) 1,4GlcNAc (cid:1) , AND Gal3ST-4 FOR THE MUCIN CORE-2 TRISACCHARIDE*

Sulfated glycoconjugates regulate biological pro-cesses such as cell adhesion and cancer metastasis. We examined the acceptor specificities and kinetic properties of three cloned Gal:3- O -sulfotransferases (Gal3STs) ST-2, ST-3, and ST-4 along with a purified Gal3ST from colon carcinoma LS180 cells. Gal3ST-2 was the dominant Gal3ST in LS180. While the mucin core-2 structure Gal (cid:1) 1,4GlcNAc (cid:1) 1,6(3- O -MeGal (cid:1) 1,3)GalNAc (cid:2) O -Bn (where Bn is benzyl) and the disaccharide Gal (cid:1) 1,4GlcNAc served as high affinity acceptors for Gal3ST-2 and Gal3ST-3, 3- O -MeGal (cid:1) 1,4GlcNAc acceptor quantitated TLC method CHCl as the Identification of the 35 S-Sulfated Compound Arising from the Accep- tor Gal (cid:1) 1,3GalNAc (cid:1) 1,3Gal (cid:2) -O-Me by the Action of LS180 Sulfotransferase— A 10-fold standard reaction mixture (200 (cid:3) l) containing Gal (cid:1) 1,3GalNAc (cid:1) (cid:2) - O -Me was incubated for 2 h at 37 °C with purified enzyme. After dilution with 1.0 ml of water, it was fractionated by the Dowex-1-Cl method. The 0.2 M NaCl eluate was lyophilized to a small volume (1 ml) and desalted on a Biogel P 2 column (1.0 (cid:3) 116.0 cm) using 0.1 M pyridine acetate, pH 5.4, as the eluting buffer. The radioactive fractions emerging first as a major peak were pooled, lyophilized to dryness, and dissolved in 200 (cid:3) l of water. A small aliquot was subjected to TLC (silica gel GHLF, 250 (cid:3) m scored 20 (cid:3) 20 cm, Analtech) using 1-propanol:NH along with authentic Gal (cid:1) 1,3GalNAc (cid:1) 1,3Gal (cid:2) - O -Me and 3- O -sulfoGal (cid:1) 1,-3GalNAc (cid:1) 1,3Gal (cid:2) - O -Me as reference compounds. The reference com- pounds were located on TLC plates by spraying with sulfuric acid in ethanol and heating at 100 °C. The radioactive compound was located by scraping 0.5-cm-wide segments of the silica and soaking in 2.0 ml of water followed by liquid scintillation counting. Metal Ions on Gal:3-O-Sulfotransferases— For studying the effect of divalent metal ions, the incubation mixture con- tained varying concentrations (0–50 m M ) of magnesium acetate, man-ganese acetate, or calcium acetate under the standard incubation conditions. Testing for Competitive Inhibition— For studying the competitive inhibition by the acceptors on Gal:3- O -sulfotransferases, we took ad-vantage of the fact that the radioactive product arising from the mono- sulfated compounds 3- O -sulfoGal (cid:1) 1,4GlcNAc (cid:1) 1,6(Gal (cid:1) 1,3)GalNAc (cid:2) O -Bn and Gal (cid:1) 1,4GlcNAc (cid:1) 1,6(3- O -sulfoGal (cid:1) 1,3)GalNAc (cid:2) - O -Bn cannot be eluted, whereas the product from the neutral acceptors can be eluted from the Dowex-1-Cl column by 0.2 M NaCl. For these runs, the con- centration of the neutral acceptor was left constant (6 m M ) and that of the sulfated compound was varied from 0 to 7.5 m M under the standard conditions. Thin layer chro- matography of Gal (cid:1) 1,3GalNAc (cid:1) 1,3Gal (cid:2) O -Me ( A ), synthetic 3- O sulfoGal (cid:1) 1,3GalNAc (cid:1) 1,3Gal (cid:2) - O -Me ( lane B ), and the 35 S-sulfated com- pound arising from the action of the purified LS180 Gal3ST enzyme 1,3GalNAc (cid:1) 1,3Gal (cid:2) O

Sulfate groups located at various defined positions in glycoconjugates are thought to play crucial roles in biological pro-cesses. For example, a sulfated Lewis x determinant has been identified to be a major structural motif in mucins isolated from a nude mice xenograft tumor produced by the human colon carcinoma LS174T-HM7 cell (1). It is suggested that this sulfated determinant may contribute to the high metastatic potential of this cell. A monoclonal antibody evoked to high molecular weight salivary mucins recognizes the epitope 3-O-sulfoGal␤1,3GlcNAc in mucinous epithelia of salivary glands, colon, and uterine cervix, but this epitope is not detectable in healthy stomach, breast, and small intestine (2). Thus, a tissuespecific distribution of sulfated glycans is noted. A novel cell substrate recognition phenomenon was demonstrated in the interaction between the lectin domains of chondroitin sulfate proteoglycans and cells expressing sulfated glycolipids (3). It is suggested that such molecular recognition may contribute to cell adhesion and migration.
Carbohydrate sulfation has been shown to be important for the formation of ligands that bind adhesion molecules belonging to the selectin family (4,5). Studies that examine the ability of sulfated carbohydrates to inhibit selectin-ligand recognition also highlight the importance of sulfation to human health. Polymers displaying the selectin recognition epitopes 3Ј,6-disulfo Lewis x and 3Ј,6Ј-disulfo Lewis x inhibited L-selectin binding to heparin under static cell-free binding conditions with similar efficacies; however, under the conditions of shear flow, only the polymer displaying 3Ј,6-disulfo Lewis x inhibited the rolling of L-selectin-transfected cells on the glycoprotein ligand glycosylation-dependent cell adhesion molecule-1 (6). Binding inhibition assays utilizing paucivalent L-selectin also identified 3Ј,6-sulfo Lewis x and 3Ј-sulfo Lewis a as potent inhibitors of L-selectin binding (7). In our studies, we have also found that substitution of a sialyl group with a sulfate group in GalNAc␤1,4(Fuc␣1,3)GlcNAc␤1,6(NeuAc␣2,3Gal␤1,3)-GalNAc␣-O-Me reduced considerably its inhibitory potential of L-and P-selectin binding (8).
In previous studies (9, 10), we identified two distinct Gal:3-O-sulfotransferases (Gal3STs) 1 in tumor tissues and cancer cells that exhibited distinct acceptor preferences. Specifically, while enzymes from colon cell lines and colon tissue prefer to sulfate the C-3 position of Gal in the Gal␤1,4GlcNAc␤-moiety of the mucin core-2 structure (Fig. 1), breast cancer cells prefer the Gal␤1,3GalNAc␣-moiety. In similar biochemical studies, we also observed that prostate carcinoma cell LNCaP has ␣1,2-L-fucosyltransferase activity that exhibits 4-fold higher activity toward Gal␤1,4GlcNAc␤-compared with the Gal residue on the mucin core-2 ␤1,3 branch (11). This enzyme also acted on the cancer antigen Globo H backbone Gal␤1,3GalNAc␤1,3Gal␣very efficiently (Fig. 1). Similar to the activity of the fucosyltransferase enzyme, we have observed that a cloned ␣2,3-sialyltransferase, ST-3GalIV, utilized both the Globo H backbone and the Gal␤1,4GlcNAc␤-structure in the core-2 tetrasaccharide. 2 The possibility that the Gal:3-O-sulfotransferase from colon tissue, similar to the ␣1,2-L-fucosyltransferase and ␣2,3sialyltransferase, also acts on the Globo H antigen remains undetermined, and we tested this here. Further the cloning of three distinct Gal:3-O-sulfotransferases recently, Gal3ST-2 (12), Gal3ST-3 (13), and , provides us with an opportunity to determine the acceptor specificities and kinetic properties of these proteins with emphasis on their action on core-2-based acceptors. These studies have led us to identify novel, high affinity, and specific acceptors for each of these enzymes, suggesting that each enzyme has a distinct physiological role.

EXPERIMENTAL PROCEDURES
Cell Culture-Human colon carcinoma LS180 and Chinese hamster ovary CHO-S cells were obtained from American Type Culture Collection (ATCC, Manassas, VA). LS180 was grown in 2-liter roller bottles in Leibovitz's L-15 medium supplemented with 10% fetal bovine serum (Hyclone, Logan, UT) and antibiotics (penicillin, streptomycin, and amphotericin B). Harvested cells were stored frozen at Ϫ20°C prior to Gal3ST purification. CHO-S was cultured in Dulbecco's modified Eagle's medium (Invitrogen) with 10% heat-inactivated fetal bovine serum in tissue culture incubators.
Expression of Cloned Gal3ST-The following plasmids containing genes that encode various Gal3STs were used in the current work: pSV-GP3ST for Gal3ST-2 (12), pCXNGal3ST for Gal3ST-3 (13), and pCMV-SPORT for . These plasmids were kindly provided by Drs. K. Honke (Osaka University Medical School, Osaka, Japan), K. Uchimura (Nagoya University School of Medicine, Nagoya, Japan) and A. Seko (Sasaki Institute, Tokyo, Japan), respectively. For protein expression, CHO cells in 6-well plates at subconfluency were transiently transfected with 6 l of LipofectAMINE 2000 (Invitrogen)/well along with 6 g/ml of one of the above plasmids. The cells were passed into T75 flasks 1 day after transfection, cultured for 3 days, then harvested, and kept frozen at Ϫ20°C until use.
Purification of Gal:3-O-Sulfotransferase from LS180 Cells-5.5 ϫ 10 9 LS180 cells were homogenized in a Dounce all-glass hand-operated grinder with 60 ml of 0.1 M Tris maleate buffer (pH 7.2) containing 10 mM magnesium acetate, 2% Triton X-100, 20% glycerol, 30 M phenylmethylsulfonyl fluoride, and 0.1% NaN 3 and then stirred for 2 h at 4°C. The homogenate was centrifuged at 10,000 ϫ g for 1 h at 4°C. The supernatant was subjected to chromatography on a 25-ml bed volume of Aleuria aurantia lectin-agarose (Vector Laboratories, Burlingame, CA) column, which had been washed and equilibrated with the above buffer. After entry of the sample into the column bed, the column was washed with 60 ml of the buffer. The bound proteins were eluted sequentially with 100 ml each of 0.5 M fucose and 2.0 M NaCl in the same buffer. Both eluates were concentrated separately by Amicon ultrafiltration using PM10 membranes to a small volume and dialyzed against 2 liters of the buffer with five changes in the cold room for 48 h. The concentrated and dialyzed fucose-eluted fraction (10 ml) was then applied to a 10-ml bed volume affinity gel-GDP (Calbiochem) column equilibrated with the same buffer. The affinity column was washed with 20 ml of the buffer and then eluted with 30 ml of 2 M NaCl in the same buffer. The NaCl eluate was concentrated and dialyzed as above. This preparation (3 ml) was further purified on a Sephacryl S-100 HR column (2.5 ϫ 118.0 cm) at 4°C, equilibrated, and eluted with 0.1 M Tris maleate, pH 7.2, containing 0.1% Triton X-100 and 0.02% NaN 3 . Fractions of 2 ml at a flow rate of 6 ml/h were collected, and 10 l of alternate fractions were assayed for sulfotransferase activity using Gal␤1,4GlcNAc␤1,6(3-O-MeGal␤1,3)GalNAc␣-O-Bn as the acceptor. The enzyme activity emerged as a single peak from the column soon after the void volume. The fractions under this peak were pooled and concentrated by ultrafiltration and dialyzed against the extraction buffer. This fraction (0.8 ml) was stored frozen at Ϫ20°C and used for enzymology studies.
RT-PCR to Identify Sulfotransferase Genes in LS180 -Total RNA extracted from LS180 cells using TRIzol reagent (Invitrogen) was subjected to RT-PCR using Superscript one-step RT-PCR with platinum Taq (Invitrogen). Total RNA was reverse-transcribed with oligo(dT) primers at 42°C for 1 h and then subjected to PCR using Taq DNA polymerase. Two sets of primers were designed for PCR amplification. The first set corresponded to the amino acid sequences FLKTHKT, FLKTHKS, DESLVLLR, and DESLVLLA, which are conserved among all cloned Gal3STs. The sense primers corresponding to these sequences were 5Ј-TTCCTGAAGACTCACAAGACG-3Ј and 5Ј-TTCCTGA-AGACACATAAATCC-3Ј, and the antisense primers used were 5Ј-CC-GCAGCAGCACTAGCGACTCGTC-3Ј and 5Ј-TGCCAGCAGAACCAAT-GACTCATC-3Ј. The second set of primers were designed against the PAPS binding domain of the cloned Gal3STs corresponding to the amino acid sequence, either HKTASSTV and LRNPVFQLESSFI or HKSGSSSV and RDPAALARSAF. One set of primers for this was 5Ј-CACAAGACGGCCAGCAGCACGGTG-3Ј (sense) and 5Ј-GTAGATGAA-GGAGGACTCCAGCTGGAACACGGGGTTCCT-3Ј (antisense), and the other set was 5Ј-ATAAATCCGGGAGCAGCTCTGTGCTGAGCC-3Ј (sense) and 5Ј-CAGAGCGAGCCAGAGCCGCTGGGTCTCG-3Ј (antisense). PCR was performed in all cases using the denaturation temperature of 94°C, annealing temperature of 58°C, and extension temperature of 72°C. RT-PCR products were sequenced by dideoxy chain termination method to determine the sulfotransferase mRNA in LS180.
Assay of Sulfotransferase-The incubation mixture run in duplicate contained 100 mM Tris maleate, pH 7.2, 5 mM magnesium acetate, 5 mM ATP, 10 mM NaF, 10 mM British anti-Lewisite, 7.5 mM acceptor, 0.5 Ci of [ 35 S]PAPS (2.4 Ci/mmol), and the enzyme in a total volume of 20 l unless otherwise stated. The control incubation mixtures contained everything except the acceptor. Incubation was carried out for 2 h at 37°C. Under our assay conditions, for all acceptors and enzymes tested, less than 25% of the donor [ 35 S]PAPS was utilized. 35 S enzymatic transfer was in the linear range for the first 2 h, and it reached maximum incorporation at 5-6 h. Similar observations were made by us in our earlier studies (9,10). The Dowex-1-Cl method was used to measure the radioactive product from neutral acceptors as follows. The incubation mixture was diluted with 1.0 ml of water and passed through Dowex-1-X8 (200 -400 mesh, chloride form) of 1-ml bed volume in a Pasteur pipette. The column was washed twice with 1.0 ml of water. A quantitative elution of the radioactive product (the 35 S-sulfated compound) was achieved by eluting the column with 3.0 ml of 0.2 M NaCl. The radioactivity present in the NaCl eluate was measured using 3a70 scintillation mixture (Research Products International, Mount Prospect, IL) and a Beckman LS6500 scintillation counter. The 0.2 M NaCl from the control reaction mixture (containing no acceptor) contained a negligible amount of radioactivity (Ͻ50 cpm) in all cases. The values for the duplicate runs did not vary by more than 5%. The 35 S-sulfated compound resulting from sialylated or sulfated synthetic FIG. 1. Core-2 tetrasaccharide and Globo H precursor. Enzymes that prefer to act on the Gal residue of the Gal␤134GlcNAc␤ moiety on the core-2 structure also typically act on the Gal residue on the Globo H backbone.
Identification of the 35 S-Sulfated Compound Arising from the Acceptor Gal␤1,3GalNAc␤1,3Gal␣-O-Me by the Action of LS180 Sulfotransferase-A 10-fold standard reaction mixture (200 l) containing Gal␤1,3GalNAc␤1,3Gal␣-O-Me was incubated for 2 h at 37°C with purified enzyme. After dilution with 1.0 ml of water, it was fractionated by the Dowex-1-Cl method. The 0.2 M NaCl eluate was lyophilized to a small volume (1 ml) and desalted on a Biogel P 2 column (1.0 ϫ 116.0 cm) using 0.1 M pyridine acetate, pH 5.4, as the eluting buffer. The radioactive fractions emerging first as a major peak were pooled, lyophilized to dryness, and dissolved in 200 l of water. A small aliquot was subjected to TLC (silica gel GHLF, 250 m scored 20 ϫ 20 cm, Analtech) using 1-propanol:NH 4 OH (25%):water (60:10:25, v/v) along with authentic Gal␤1,3GalNAc␤1,3Gal␣-O-Me and 3-O-sulfoGal␤1,-3GalNAc␤1,3Gal␣-O-Me as reference compounds. The reference compounds were located on TLC plates by spraying with sulfuric acid in ethanol and heating at 100°C. The radioactive compound was located by scraping 0.5-cm-wide segments of the silica and soaking in 2.0 ml of water followed by liquid scintillation counting.

Effect of Divalent Metal Ions on Gal:3-O-Sulfotransferases-
For studying the effect of divalent metal ions, the incubation mixture contained varying concentrations (0 -50 mM) of magnesium acetate, manganese acetate, or calcium acetate under the standard incubation conditions.
Testing for Competitive Inhibition-For studying the competitive inhibition by the acceptors on Gal:3-O-sulfotransferases, we took advantage of the fact that the radioactive product arising from the monosulfated compounds 3-O-sulfoGal␤1,4GlcNAc␤1,6(Gal␤1,3)GalNAc␣-O-Bn and Gal␤1,4GlcNAc␤1,6(3-O-sulfoGal␤1,3)GalNAc␣-O-Bn cannot be eluted, whereas the product from the neutral acceptors can be eluted from the Dowex-1-Cl column by 0.2 M NaCl. For these runs, the concentration of the neutral acceptor was left constant (6 mM) and that of the sulfated compound was varied from 0 to 7.5 mM under the standard conditions.

Gal3ST Purified from LS180 Acts Both on Gal␤1,4GlcNAc␤ in Mucin Core-2 as Well as Terminal Gal␤ in Globo H Precursor
Gal3ST was purified from LS180 cells using a series of chromatography steps ( Table I). The enzyme was purified 750-fold with a recovery of 40% using the three steps of Triton X-100 solubilization, A. aurantia lectin-agarose chromatography, and fractionation on affinity gel-GDP. When this purified preparation was further subjected to chromatography on a Sephacryl S-100 HR column, we obtained a single peak of activity (data not shown) emerging from the column soon after the void volume as measured with blue dextran 2000 but prior to our bovine serum albumin standard (molecular mass, 66,000 Da). The enzyme at this stage was purified 3000-fold with 30% recovery of activity. The enzyme exhibited a relatively low specific activity of ϳ700 milliunits/mg suggesting that it was only partially purified, and this was confirmed by silver staining of the SDS-polyacrylamide gel of the partially purified fraction (data not shown). GlcNAc:6-O-sulfotransferase activity is not exhibited by sulfotransferases in LS180 cells (10), and this was the case for our fraction as well. Further, as shown below, the partially purified enzyme exhibited activity that resembled only one of the three cloned Gal3STs studied in the current work suggesting that it contains only one Gal3ST.
TLC experiments were conducted to determine that the purified enzyme transferred sulfate to the C-3 hydroxyl of the terminal Gal moiety in Gal␤1,3GalNAc␤1,3Gal␣ (Fig. 2). A monosulfated standard compound, 3-O-sulfoGal␤1,3-GalNAc␤1,3Gal␣-O-Me, was synthesized for this purpose. When the 35 S-sulfated product isolated from the action of LS180 sulfotransferase on Gal␤1,3GalNAc␤1,3Gal␣-O-Me was subjected to TLC (Fig. 2) along with the synthetic sulfated standard, it was found that the mobility of the radioactive product coincided with that of the synthetic sulfated compound. Thus, based on the combined results from Table II and Fig. 2, it has been tentatively identified that sulfation by LS180 Gal3ST takes place at the C-3 OH of the terminal Gal moiety in the Globo H precursor.

TABLE II
The Globo H structure and its 6-deoxy analog are better acceptors than the mucin core-2 structure for colon carcinoma LS180 Gal3ST a Value in parentheses is the activity in cpm of the reference acceptor (100% activity) against which all other acceptors were compared. Blank samples (without any acceptor) were of less than 50 cpm, and this corresponds to 0% activity.

A Comparison of LS180 Gal3ST with the Cloned Gal3STs
Comparison of the enzyme activities in Tables II-IV suggests that the identity of the Gal3ST isolated from LS180 cells is Gal3ST-2. Experiments were conducted to verify this finding.
Effect of Divalent Metal Ions-The enzymatic transfer of sulfate was measured separately with each Gal3ST using specific acceptors in the presence of varying concentrations of Mg 2ϩ , Mn 2ϩ , and Ca 2ϩ (Fig. 3). In the case of LS180 sulfotransferase, both Mg 2ϩ and Ca 2ϩ showed a similar profile of activity when Gal␤1,4GlcNAc␤1,6(3-O-MeGal␤1,3)GalNAc␣-O-Bn (Fig. 3A) and Gal␤1,3GalNAc␤1,3Gal␣-O-Al (Fig. 3B) were used as acceptors. In both cases, Mn 2ϩ stimulated the enzyme activity up to a concentration of 10 mM, and the activity gradually decreased with further increase in Mn 2ϩ concentration. The stimulation of Gal3ST-2 activity by Mg 2ϩ , Mn 2ϩ , and Ca 2ϩ was similar to that of LS180 Gal3ST (Fig. 3, compare C with A  and B).  Table II.  Table II. b AA-CP, acrylamide copolymer.  Table II.

Distinct Specificities of Gal:3-O-Sulfotransferases
Among the Gal3STs tested, Mg 2ϩ exhibited a stimulating effect only on Gal3ST-3 when fetuin triantennary asialo-GP was used as an acceptor (Fig. 3D). None of the cations, Mg 2ϩ , Mn 2ϩ , or Ca 2ϩ , had a stimulating effect on Gal3ST-4. In fact, a gradual decrease in the enzyme activity of Gal3ST-4 was noticed upon increasing the concentration of Mn 2ϩ in the reaction mixture (Fig. 3E).
Effect of pH on Gal3ST Activities-The activities of LS180 Gal3ST and Gal3ST-2 were measured over a pH range from 5.2 to 8.4 using the Globo H precursor Gal␤1,3GalNAc␤1,3Gal␣-O-Me as the acceptor. Enzyme activity in both cases had an optimum at pH 6.8 (data not shown). The near identical pH-dependent activity profiles of the two enzymes support the proposition that they have the same identity.

DISCUSSION
Studies of cellular glycosyltransferase expression at the mRNA level can be performed using microarrays and Northern blot analysis, and such analysis can also be performed at the protein level using enzymology methods. Both strategies when combined with knowledge of glycan biosynthesis pathways can be useful in predicting oligosaccharide structures on cell glycolipids and glycoproteins. Enzymatic studies using a range of well defined acceptors can quantitatively compare enzyme activities, and they can be used to study the competition of various enzymes for a single substrate. Identification of unique substrates for each enzyme can also allow development of rapid assay strategies for the identification of a particular glycosyltransferase in a complex mixture. We discuss here results of our studies with three cloned Gal:3-O-sulfotransferases: Gal3ST-2, Gal3ST-3, and Gal3ST-4. Emphasis is placed on acceptor studies with the mucin core-2 structure and the Globo H precursor due to the physiological importance of these molecules and because such studies have not been carried out previously.
Gal3ST in Colon and Breast Cancer Cells and Tissue-Various Gal:3-O-sulfotransferases have been noted to act on glycoproteins and glycolipids. In one study, it was observed that the human renal cancer cell line SMKT-R3 sulfotransferase utilized both GalCer and LacCer as acceptors but did not act on the terminal ␤-Gal moiety of oligosaccharides that are not associated with lipids (24). This enzyme was cloned and designated human cerebroside 3Ј-sulfotransferase (24). Honke et al. (12) further used the cerebroside 3Ј-sulfotransferase cDNA sequence as a probe and cloned a human ␤-Gal:3-O-sulfotransferase (GP3ST or Gal3ST-2) that acts on LacNAc types 1 and 2 as well as mucin core 1 structure. The amino acid sequence of Gal3ST-2 indicated 33% identity to the cerebroside 3Ј-sulfotransferase sequence, and this enzyme acted on the terminal ␤-Gal moiety of oligosaccharide chains in glycoproteins only (12). Others have noted the existence of a Gal:3-O-sulfotransferase from human respiratory mucosa that acts on terminal LacNAc unit in mucins, but this enzyme did not utilize GalCer (25). The activity of this enzyme does not match any of the Gal3STs tested in the current study. Chance and Mawhinney (26) reported the occurrence of the 3-O-sulfoGal␤1,4(6-O-sulfo) GlcNAc␤-sequence in tracheobronchial mucin. The present study identifies Gal␤1,4(6-O-sulfo)GlcNAc␤-O-Ac as a novel acceptor for Gal3ST-3 and suggests that this enzyme may be responsible for the biosynthesis of this bronchial mucin. The 3-O-sulfoGal␤1,3GalNAc␤1,3Gal␣1-sequence has also been suggested to be part of sulfated glycolipids in kidney (27,28).
Our studies suggest that the sulfotransferase involved in this process is either Gal3ST-2 or Gal3ST-4. We have also reported earlier (9, 10) the existence of two distinct Gal:3-O-sulfotransferases showing acceptor preference to either LacNAc type 2 unit (Gal␤1,4GlcNAc␤) or the T-hapten unit (Gal␤1,3GalNAc␣) of the mucin core-2 structure. The enzymes that acted on Gal␤1,4GlcNAc␤ were observed in colon tumor tissue (9) and colon cancer cell lines (10), while the Gal3STs acting on Gal␤1,3GalNAc␣ were observed in breast tumor tissues and breast cancer cell lines (9, 10). These two enzymes exhibited significant difference in their kinetic properties such as pH optima and divalent metal ion activation.
In the present study, we further purified the Gal3ST from LS180 cells using a series of chromatography steps. In these studies, we noted that the LS180 Gal3ST can bind to a GDP affinity column. The reason for this binding may be due to the fact that the biological sulfate donor, PAPS, contains adenine (a purine base), and Gal:3-O-sulfotransferase is able to bind to another purine base guanine. Further, even after significant purification, the Gal3ST exhibited the same kind of acceptor specificity and kinetics toward the LacNAc type 2 structure of mucin core-2 as the enzyme from crude lysate. Our enzyme preparation from LS180 was free of GlcNAc:6-O-sulfotransferase as Me-O-Gal␤1,3(GlcNAc␤1,6)GalNAc␣-O-Bn did not act as an acceptor. Upon comparison with the three cloned Gal3STs, the activity of LS180 sulfotransferase most closely resembled Gal3ST-2. Drawing analogies from the current work, based on the acceptor specificity studies, it appears very likely that the predominant Gal3ST in breast cancer cells and tissue is Gal3ST-4.
Identification of Unique Acceptors for Cloned Gal3STs-We performed studies on various cloned Gal:3-O-sulfotransferases (Gal3ST-2, Gal3ST-3, and Gal3ST-4) with the objective of determining both unique and overlapping substrates for each of these enzymes. Identification of this acceptor specificity can reveal the glycan structure facilitated by these enzymes and the biological role of these enzymes.
The current study defines the acceptor specificity of carbohydrate-specific Gal3STs cloned thus far, and it poses new questions for the future. In this regard, the characterization of specific substrates in the current work will likely facilitate assessment of the functional importance of these enzymes in the future. For example, based on the current work, it appears that the generation of monoclonal antibodies against the 3-Osulfated mucin core-2 tetrasaccharide structure and the 3-Osulfated Globo H trisaccharide may be useful reagents for future immunohistochemical studies that examine the tissue distribution of sulfated glycans. Our enzymatic studies suggest the existence of novel disulfated structures in vivo. They also suggest that it will be important to compare the competitive action of various ␣2,3-sialyltransferases and Gal3STs on the core-2 mucin structures along with structural analysis of cell surface carbohydrates. Finally we showed here that a 3-fluoro or 4-O-methyl substituent on either Gal moiety of the mucin core-2 tetrasaccharide reduced the function of various Gal3STs. It will be interesting to determine whether such molecules are poor acceptors for these enzymes or whether they act as competitive inhibitors that may find in vivo application as sulfo-and sialyltransferase inhibitors.