Detection of 2- O -Sulfated Iduronate and N -Acetylglucosamine Units in Heparan Sulfate by an Antibody Selected against Acharan Sulfate (IdoA2S-GlcNAc) n *

The snail glycosaminoglycan acharan sulfate (AS) is structurally related to heparan sulfates (HS) and has a repeating disaccharide structure of (cid:1) - D - N -acetylglu-cosaminyl-2- O -sulfo- (cid:1) - L -iduronic acid (GlcNAc-IdoA2S) residues. Using the phage display technology, a unique antibody (MW3G3) was selected against AS with a V H 3, DP 47, and a CDR3 amino acid sequence of QKKRPRF. Antibody MW3G3 did not react with desulfated, N- deacetylated or N- sulfated AS, indicating that reactivity depends on N- acetyl and 2- O -sulfate groups. Antibody MW3G3 also had a high preference for (modified) heparin oligosaccharides containing N- acetylated glucosamine and All performed at AS, Heparin, Heparin Oligosaccharides— evaluated the modified

The snail glycosaminoglycan acharan sulfate (AS) is structurally related to heparan sulfates (HS) and has a repeating disaccharide structure of ␣-D-N-acetylglucosaminyl-2-O-sulfo-␣-L-iduronic acid (GlcNAc-IdoA2S) residues. Using the phage display technology, a unique antibody (MW3G3) was selected against AS with a V H 3, DP 47, and a CDR3 amino acid sequence of QKKRPRF. Antibody MW3G3 did not react with desulfated, N-deacetylated or N-sulfated AS, indicating that reactivity depends on N-acetyl and 2-O-sulfate groups. Antibody MW3G3 also had a high preference for (modified) heparin oligosaccharides containing N-acetylated glucosamine and 2-O-sulfated iduronic acid residues. In tissues, antibody MW3G3 identified a HS oligosaccharide epitope containing N-acetylated glucosamine and 2-O-sulfated iduronic acid residues as enzymatic N-deacetylation of HS in situ prevented staining, and 2-O-sulfotransferase-deficient Chinese hamster ovary cells were not reactive. An immunohistochemical survey using various rat organs revealed a distinct distribution of the MW3G3 epitope, which was primarily present in the basal laminae of most (but not all) blood vessels and of some epithelia, including human skin. No staining was observed in the glycosaminoglycan-rich tumor matrix of metastatic melanoma. In conclusion, we have selected an antibody that identifies HS oligosaccharides containing N-acetylated glucosamine and 2-O-sulfated iduronic acid residues. This antibody may be instrumental in identifying structural alterations in HS in health and disease.
Acharan sulfate (AS) 1 is a glycosaminoglycan (GAG) abundantly present in the giant African snail Achatina fulica (1). It largely consists of the repeating disaccacharide structure of 34)-␣-D-2-acetamido-2-deoxyglucopyranose(134)-␣-L-idopyranosyluronic acid-2-sulfate(13 and is closely related to heparan sulfates. AS is located in large granules present in the outer surface of the snail and is secreted onto the surface as mucus. This mucus consists of 26% proteins, and the GAG moiety is entirely formed by AS. Whether AS is present as a proteoglycan is unclear (2). The pattern of adjacent N-acetylglucosamine and 2-sulfoiduronic acid residues is unusual and suggests interesting biological activities. Among the proposed biological functions of AS in snails are binding, uptake, and transport of divalent cations and functioning as an antidesiccant (1). AS inhibits the mitogenic activity induced by basic fibroblast growth factor (3), angiogenesis (4), and tumor growth (5).
To study the cell biological importance of specific HS epitopes and their location in tissue, single chain antibodies have been generated as described by our group (6 -9). These antibodies were selected against HS from various sources and stained differentially in tissues sections from various organs. Defined HS oligosaccharides were used to reveal details of the epitopes recognized by the antibodies (6). Although some chemical groups in HS essential for antibody reactivity could be determined, the oligosaccharide sequence recognized by these antibodies remains largely unknown. The homogenous structure of AS and its resemblance to HS, together with its biological activities, especially its anti-tumor potential, made us decide to select single chain variable fragment (scFv) antibodies directed against AS using the phage display technology. All chemicals used were purchased from Merck unless stated otherwise. Bacterial medium (2ϫ TY) was from Invitrogen. Cell culture media were from Invitrogen. Isopropyl-␤-D-thiogalactopyranoside and bovine serum albumin (fraction V) were from Sigma. Protease inhibitor mixture was from Roche Applied Science. Mowiol (4-88) was obtained from Calbiochem. The ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit was from PerkinElmer Life Sciences. Microlon 96-well microtiter plates were from Greiner (Frickenhausen, Germany). Polystyrene maxisorp immunotubes were from Nunc (Roskilde, Denmark). Multispot slides were from Nutacon (Leimuiden, The Netherlands). Protein A-agarose beads were from Kem-Tec (Copenhagen, Denmark).

Materials
Heparin and HS from porcine intestinal mucosa, HS from bovine kidney, chondroitin 4-sulfate from whale cartilage, chondroitin 6-sulfate from shark cartilage, dermatan sulfate from porcine intestinal mucosa, hyaluronate from human umbilical cord, and DNA from calf thymus were from Sigma. (GlcA-GlcNAc) n (K5) and K5-O-S were a kind gift from the Department of Nephrology, University Medical Center Nijmegen (11). Recombinant heparinase III (from Flavobacterium heparinum) was a kind gift from IBEX Technologies (Montreal, Canada).
Chemically Modified AS, Heparin, and Heparin Oligosaccharides-Isolation of AS was previously described (1) and was evaluated by agarose gel electrophoresis as described (13). AS was N-deacetylated by hydrazinolysis and N-sulfated using the sulfur trioxide-trimethylamine complex (14). 2-O-Desulfated AS was prepared under alkaline lyophilization conditions (15). Preparation of the modified heparin samples (N-desulfated (de-NS), re-N-acetylated heparin; 2-O-desulfated heparin; and de-NS, re-N-acetylated, 2-O-desulfated heparin) and the disaccharide analysis of these samples including heparin from bovine lung were described previously (16,17). The preparation of heparin dp6 (where dp represents degree of polymerization) and dp8 as well as of N-acetylated heparin dp6 and dp8 oligosaccharides was carried out as described by Goger et al. (18) and Ostrovsky et al. (19).
Selection of Anti-AS Antibodies-The human semisynthetic scFv library 1 (10) was used to select anti-acharan sulfate antibodies. This library contains 10 8 different clones with 50 different V H genes with synthetic random complementarity-determining region 3 regions and one V L gene. The selection of phages displaying scFv antibodies was performed as described (7)(8)(9). The scFv antibodies expressed contain a c-myc tag for detection. Phagemid DNA of clones expressing scFv antibodies reactive with AS were isolated and digested with restriction enzymes NotI and NcoI to obtain the DNA encoding the scFv. This fragment was subcloned into vector pUC119 His VSV (J. M. H. Raats, Department of Biochemistry, Faculty of Sciences, University of Nijmegen, Nijmegen, The Netherlands), which is similar to the original pHEN1 vector but does not contain the c-Myc tag and pIII gene. Instead, this vector contains a polyhistidine and VSV tag for detection and purification of the antibodies.
Large Scale Preparation of Antibodies-To produce scFv antibodies, periplasmic fractions of infected bacteria were isolated as described (7)(8)(9)20). Bacteria (expressing anti-AS scFv antibodies) were grown and induced by isopropyl-␤-D-thiogalactopyranoside to produce antibodies. The bacterial periplasmic fraction, containing the antibody, was isolated, dialyzed against PBS, and stored at Ϫ20°C.
Evaluation of Specificity by ELISA-The reactivity of the anti-AS antibody with a large number of GAG preparations was tested using ELISA, as described previously (7,8). Briefly, GAGs were immobilized to 96-well microtiter plates and incubated with periplasmic fractions containing antibodies, followed by incubation with anti-VSV tag antibody P5D4 and alkaline phosphatase conjugated rabbit anti-mouse IgGs, respectively. Enzyme activity was detected using 1 mg of pnitrophenyl phosphate per ml of 1 M diethanolamine, 0.5 mM MgCl 2 , pH 9.8, and absorbance was measured at 405 nm. All assays were performed at least three times, and representative results are shown.
Immunoprecipitation of Heparin Oligosaccharides with Anti-AS Antibodies-Heparin dp6 and dp8 and N-desulfated reacetylated heparin dp6 and dp8 oligosaccharides were subjected to immunoprecipitation using the anti-AS antibody (VSV-tagged). Oligosaccharides (10 g) were incubated with 0.5 ml of periplasmic fraction for 1 h while shaking. Protein A-agarose beads (100 l) were loaded with 500 g of anti-VSV tag antibody (P5D4). Next, 100 l of protein A-agarose beads loaded with P5D4 antibodies were added to the oligosaccharide/anti-AS antibody mixture and incubated for 1 h while shaking. Beads were spun down (10,000 ϫ g), washed three times with PBS, mixed with PAGE loading buffer (Tris acetate buffer, pH 7.0, with 0.5 M NaCl), and heated for 2 min. Samples were analyzed by PAGE using a 33% gel (21), using alcian blue fixation (0.8% (w/v) in 2% (v/v) acetic acid) and silver staining. Control incubations included omission of antibody and heparin oligosaccharides and the use of an irrelevant antibody.
Isolation of Metabolically Radiolabeled HS and Disaccharide Analysis-HS chains ( 3 H-labeled) were isolated from wild type (CHOK1) and 2-O-sulfotransferase mutant CHO cells (pgsF17) and digested with a combination of heparinases I, II, and III, and disaccharide compositional analysis was performed as previously described (22).
Evaluation of Specificity by Immunohistochemistry-Tissue specimens were snap frozen in liquid nitrogen and stored at Ϫ80°C. Tissue sections (5 m) were incubated with periplasmic fractions of anti-AS antibodies in 1% (w/v) bovine serum albumin in PBS with 0.05% (v/v) Tween 20. Bound antibodies were detected using anti-tag antibodies (anti-VSV, P5D4, and IgG1) followed by Alexa-labeled (488) anti-mouse antibodies. Finally, tissue sections were fixed in ethanol, air-dried, and embedded in mowiol (10% (w/v) in 0.1 M Tris-HCl, pH 8.5, 25% (v/v) glycerol, and 2.5% (w/v) NaN 3 ). As a control, primary antibodies were omitted. To evaluate the HS specificity of the antibodies, tissue sections were pretreated with heparinase III to remove all HS (0.02 IU/ml in 50 mM NaAc, 50 mM Ca(Ac) 2 , pH 7.0; overnight incubation at 37°C). As a control, tissue sections were incubated with reaction buffer without enzyme. After HS removal, tissue sections were washed in PBS and processed for immunofluorescence analysis as described above. Digestion of HS was analyzed by incubation with antibody 3G10 (12), which recognizes HS stubs generated after HS digestion. Tissue sections were also incubated with a mixture of anti-AS antibodies and AS or AS oligosaccharides (14-mers) and evaluated for inhibition of antibody staining. Mixtures of anti-AS antibodies and oligosaccharides (10 and 50 g/ml) were preincubated for 2 min before adding to the tissue section. Tissue sections were processed for immunofluorescence analysis as described above.
Wild type CHOK1 and mutant pgsF17 CHO (defective in HS 2-Osulfation) cells were cultured in Ham's F-12 culture medium supplemented with 10% fetal calf serum on multispot slides for 48 h to reach 80% confluence, washed with PBS, and fixed with 4% paraformaldehyde for 10 min and processed for immunofluorescence analysis as described above. Staining patterns were analyzed by fluorescence microscopy.
N-Deacetylase/N-Sulfotransferase 1 and 2 Cell Lysates and Deacetylation of Rat Kidney Sections-Stably transfected human kidney 293 cell lines clone 11 (NDST 1) and clone S5 (NDST 2) were grown in Ham's F-12 culture medium supplemented with 10% fetal calf serum and harvested at 80% confluence. Cell lysates were prepared by homogenizing cells on ice in 50 mM Tris, pH 7.4, containing 2 mM EDTA, 1% Triton X-100, and protease inhibitors. Homogenates were centrifuged for 15 min, at 10,000 ϫ g at 4°C, and the supernatant was collected, diluted 1:1 in glycerol, and stored at Ϫ20°C. Protein concentration varied between 10 and 20 mg/ml as determined by the method of Lowry. Cell lysates were stable for at least 1 month. For deacetylation of tissue sections, rat kidney sections were incubated with cell lysates containing 500 g of protein/ml for 16 h at 37°C. As a control, tissue sections were incubated with lysis buffer. Tissue sections were washed with PBS and processed for immunofluorescence analysis as described above.

RESULTS
Acharan Sulfate-AS was analyzed by agarose gel electrophoresis followed by silver staining of the gel (Fig. 1). AS migrates somewhat faster than HS from bovine kidney but slower than dermatan sulfate.
Selection of Antibodies against Acharan Sulfate-Four rounds of panning were performed against AS using the semisynthetic scFv library 1, which resulted in a total increase of phage titer from 6 ϫ 10 2 colony-forming units in the first round to 1 ϫ 10 8 colony-forming units in the fourth round. Supernatants containing antibodies of the third and fourth selection were tested for reactivity with AS by ELISA. Of the 188 clones screened, four reacted with AS, which were all from the third selection round. DNA sequence analysis revealed that all clones expressed identical antibodies. Clone MW3G3 was selected for further analysis. This antibody belongs to the V H 3 family, has a DP 47 germline gene segment, and contains the heavy chain complementarity-determining region 3 amino acid sequence QKKRPRF.
Evaluation of the Specificity of Anti-AS Antibody MW3G3-Antibody MW3G3 was analyzed for reactivity with various GAG species and control samples by ELISA (Fig. 2, Table I). Antibody MW3G3 reacted strongly with AS, moderately with heparin, and very weakly with HS from bovine kidney and porcine intestinal mucosa. No reactivity was seen with chondroitin sulfate-A, chondroitin sulfate-C, dermatan sulfate, and hyaluronate nor with DNA. To verify the necessity for N-acetylated glucosamine residues and 2-O-sulfated iduronic acid residues, as present in AS, various modified AS, heparin, and K5 preparations were analyzed (Table I). Antibody MW3G3 reacted strongly with AS, but not with N-sulfated AS, N-deacetylated AS, and de-O-sulfated AS. These results suggest that both types of modification (N-acetylation and 2-O-sulfation) are important for antibody recognition. Antibody MW3G3 reacted very strongly with N-desulfated, N-acetylated heparin, which is strongly in favor of the above suggested specificity of antibody MW3G3. Reactivity for heparin was dependent on the source used. Heparin from intestinal mucosa reacted quite strongly, whereas heparin from bovine lung, which is more sulfated and less N-acetylated, reacted much more weakly. Desulfated and re-N-sulfated heparin, desulfated and N-acetylated heparin, and 2-O-desulfated heparin showed no reactivity with antibody MW3G3, demonstrating the need for 2-O-sulfated iduronic acid residues. Antibody MW3G3 showed weak reactivity with a N-desulfated, N-acetylated, and partly 2-Odesulfated (15% (17)) heparin preparation (Table I).
Reactivity of antibody MW3G3 for N-desulfated reacetylated heparin is very strong and suggested a strong preference for N-acetylated glucosamine residues. To address this, antibody MW3G3 reactivity for heparin dp6 and dp8 and N-desulfated, N-acetylated heparin dp6 and dp8 oligosaccharides was analyzed. A mixture of all four classes of oligosaccharides was allowed to react with antibody MW3G3 using an immunoprecipitation approach (Fig. 3). The results clearly demonstrated that the antibody strongly precipitates N-acetylated heparin dp8 oligomers, suggesting that antibody MW3G3 gives high preference to N-acetylated glucosamine residues instead of Nsulfated glucosamine residues (which are abundant in heparin).

N-Acetylation Is Essential for Binding of Antibody MW3G3 in Situ as Determined by HS Deacetylation of Rat Kidney
Tissue Sections-The necessity of N-acetyl groups for antibody recognition was further analyzed by enzymatic removal of Nacetyl groups using cell lysates of human kidney 293 cells stably transfected with NDST 1 and NDST 2 cDNA (23, 24), followed by anti-HS antibody staining of the tissue section (25) (Fig. 4, Table I). After incubation, hardly any staining with antibody MW3G3 was observed anymore (Fig. 4, A and C). Staining with anti-HS antibody HS4C3, which recognizes highly sulfated HS structures (9), was not affected (Fig. 4, B and D). No differences were observed between NDST 1 or NDST 2 cell lysates. These data indicated that the presence of N-acetyl groups is essential for staining of antibody MW3G3.  Table II. Wild type CHO cells showed high expression of MW3G3 epitopes (Fig. 5A), whereas mutant CHO cells were negative (Fig. 5B). We used antibody AO4B08 as a control antibody for the presence of 2-O-sulfation (6) (Fig. 5, C and D).
Immunofluorescence Detection of the MW3G3 Epitope in the Giant Snail A. fulica and Rat Tissues-To ascertain that antibody MW3G3 recognizes HS, tissue sections of rat kidney and intestine and of human skin, atypical nevi, and metastatic melanoma (see below) were pretreated with heparinase III. A total loss of staining was observed, indicating that indeed HS was recognized by antibody MW3G3 (data not shown). Antibody staining was completely inhibited by incubation of the antibody with AS and AS oligosaccharides (14-mers) at a concentration of 50 g/ml (data not shown). At a concentration of 10 g/ml, staining was greatly reduced. In the snail A. fulica, large granules present in the outer surface of the snail, which contain AS, stained very intensely (Fig. 6A), whereas the control showed no staining at all (Fig. 6B). Rat tissue sections (intestine, liver, pancreas, kidney, testis, and tongue) were analyzed for reactivity with antibody MW3G3 (Fig. 7, Table  III). In general, blood vessels were stained (i.e. (larger) arteries, venules, and capillaries). However, staining of blood vessels was variable, and some did not stain at all. For example, the vena centralis in the liver was strongly stained (Fig. 7, Liver,  left), whereas the vena interlobularis did not stain (Fig. 7, Liver, right); a gradual staining was observed in the sinusoids. In addition, the basal lamina of some epithelia were stained, like that of the transitional epithelium in the renal calyx (Fig.  7, Kidney, right), the epithelium of the tongue (Fig. 7, Tongue,  left), and, weakly, the crypt epithelium in the intestine (Fig. 7,  Intestine, left). The basal lamina of the tubuli in the kidney (Fig. 7, Kidney, left), the testis (Fig. 7, Testis), and intestine villi (Fig. 7, Intestine, right) hardly stained with antibody MW3G3. Intracellularly, strong staining was observed in Paneth cells (Fig. 7, Intestine, left) and mast cells in the intestine (Fig. 7, Intestine, left) and cells at the periphery of the islets of Langerhans probably representing glucagon-producing A-cells (Fig.  7, Pancreas, left).
The Reactivity of Antibody MW3G3 with Normal Human Skin, Atypical Nevi, and Metastatic Melanoma-AS can act as an anti-tumor agent and demonstrates antiangiogenic activities (4, 5). Therefore, we analyzed the expression of the epitope of antibody MW3G3 in normal human skin, atypical nevi, and FIG. 3. Immunoprecipitation analysis of (modified) heparin oligosaccharides using antibody MW3G3. Heparin dp6 (hep dp6; 1) and dp8 (hep dp8; 3) and N-desulfated reacetylated heparin dp6 (hep dp6, N-acetylated; 2) and dp8 (hep dp8, N-acetylated; 4) oligosaccharides were combined and incubated with antibody MW3G3, followed by incubation with protein A-agarose beads loaded with anti-tag antibody. Immunoprecipitated fragments were analyzed by 33% polyacrylamide gel electrophoresis followed by alcian blue fixation and silver staining. Individual oligosaccharides (indicated by 1, 2, 3, and 4) and the immunoprecipitated fragments (indicated by 1ϩ2 ϩ 3ϩ4) were loaded on the gel. Note that there is a high preference for N-acetylated heparin oligosaccharides of 8 saccharides in length.

FIG. 4. Immunolocalization of the MW3G3 epitope in rat kidney section after treatment with N-deacetylase/N-sulfotransferase enzymes.
Cryosections of rat kidney were incubated without (A and B) and with (C and D) cell lysates of 293 cells stably transfected with NSDT enzymes to remove N-acetyl groups from GlcNAc residues (note that no sulfotransferase activity is present due to a lack of sulfate donor). After incubation for 16 h at 37°C with NDST 1 cell lysate (500 g of protein/ml), cryosections were washed and stained using antibody MW3G3 (A and C) and HS4C3 (B and D; control). Staining with antibody MW3G3, but not HS4C3, is almost completely abolished after NDST incubation. No differences were observed between cell lysates of NDST 1-or 2-transfected cells. Bar, 25 m. An Antibody against Acharan Sulfate (IdoA2S-GlcNAc) n metastatic melanoma (Fig. 8). In normal skin and atypical nevi, the basal lamina of the epidermis strongly stained; however, no staining of blood vessels in the dermis was observed (Fig. 8, Skin and AN, left). An anti-chondroitin sulfate scFv antibody IO3H10 (27) (Fig. 8, right) showed strong staining associated with the basal lamina of the epidermis and of the blood vessels. No staining was observed in metastatic melanoma with antibody MW3G3, whereas the anti-chondroitin sulfate antibody intensely stained the extracellular matrix surrounding nests of tumor cells and the blood vessels (Fig. 8, MM).

DISCUSSION
The disaccharide structure of AS is quite homogenous and largely consists of repeating GlcNAc-IdoA2S disaccharides (1). This uniform GAG structure seemed ideal to select antibodies against, since this might predict the epitope structure recognized by the antibody. Therefore, the antibody phage display technology was applied to obtain antibodies against AS. Antibody MW3G3 (DP47, V H 3, complementarity-determining region 3 QKKRPRF) was selected and reacted strongly with AS. Both the N-acetyl and 2-O-sulfate groups were found essential for antibody recognition. In general, anti-GAG single chain antibodies react well with oligosaccharides consisting of five or more monosaccharides (6), and therefore the most compatible epitope in AS may be (IdoA2S-GlcNAc) 3 .
The antibody also reacts with HS and heparin, and, as in AS, the essential groups are N-acetylated glucosamine and 2-Osulfated uronic acid residues, as was demonstrated by ELISA, immunoprecipitation, NDST enzyme treatment on tissue sections, and mutant cell lines (2-O-sulfotransferase-deficient cells). HS as well as heparin contain IdoA2S-GlcNAc disaccha-  7. Immunolocalization of the HS epitope recognized by antibody MW3G3 in rat tissues. Cryosection of various rat tissues were incubated with the MW3G3 antibody. Rat intestine, liver, pancreas, kidney, testis, and tongue were analyzed. Note that mainly blood vessels (indicated by arrows) were stained. In intestine cells of Paneth (curved arrow) and in the pancreas cells at the periphery of the islets of Langerhans, most likely A-cells (arrowhead), were stained. Basal lamina of the transitional epithelium of the calyx (Kidney, right, arrowhead) and of the epithelium of the tongue (arrowhead) stained intensely, whereas the basal lamina of the intestine crypts stained weakly (arrowhead). Bar, 25 m. rides, albeit in low amounts (22, 28 -33). Since heparin octasaccharides bind much better to the antibody than hexasaccharides, it is unlikely that one single IdoA2S-GlcNAc disaccharide forms the epitope. However, a repeating disaccharide structure as found in AS has not been demonstrated in HS/heparin and appears biosynthetically impossible, since N-and 2-O-sulfation reactions are interdependent such that N-sulfation on the nonreducing side is a prerequisite for 2-O-sulfation on the reducing side (34).
To create IdoA2S-GlcNAc units, epimerization has to occur. The epimerase acts on GlcA only when it is located at the reducing side of a GlcNS residue, and the enzyme does not react with GlcA that is O-sulfated or that is adjacent to an O-sulfated glucosamine (35)(36)(37). Therefore, epimerization starts after N-deacetylation and N-sulfation but before 6-Oand 3-O-sulfation (38). The iduronic acid residue may then be 2-O-sulfated by the 2-O-sulfotransferase, which acts on both IdoA and GlcA units but prefers the former (39,40). The percentage of 2-O-sulfation of the GlcA units is very low (32), and GlcA2S-GlcNAc disaccharide units have not been reported in HS/heparin. Taken together, biosynthetic constraints make it unlikely that in HS/heparin the epitope of antibody MW3G3 is (IdoA2S-GlcNAc) n (n Ͼ 1), although it cannot be excluded that such an epitope, be it very rare, occurs. The epitope may there-fore be an oligosaccharide with a core of IdoA2S-GlcNAc, flanked by other disaccharides, or an oligosaccharide in which IdoA2S and GlcNAc residues are scattered.
Given the notion that at least one IdoA2S and one GlcNAc residue are present in the epitope, one can speculate which sites in HS/heparin chains are most reactive with the antibody. An HS/heparin chain is made up of N-acetylated (NA), Nsulfated (NS), and N-acetylated/N-sulfated (NA/NS) or mixed domains. The total number of IdoA is almost evenly distributed over NS and mixed domains. However, almost all IdoA are 2-O-sulfated in the NS domains, whereas almost none are in the NA/NS domains. Since 2-O-sulfated IdoA are largely restricted to NS domains, it is predicted that the location of the MW3G3 epitope is at the interface of NS domains and the adjacent NA/NS regions (41) in sequences of the type -GlcNS-IdoA2S-GlcNAc-GlcA-.
Immunohistochemical staining using antibody MW3G3 revealed a very specific expression pattern in rat tissues. The MG3G3 epitope is almost exclusively found in basal lamina of most, but not all, blood vessels and of some epithelia, indicating expression of IdoA2S-and GlcNAc-containing oligosaccharides in these structures. Remarkable is the absence of the epitope in extracellular matrix structures present in melanoma metastases, which are very rich in GAGs (42). The absence of IdoA2S/ GlcNAc residues in melanoma metastasis is in line with the  8. Immunolocalization of the HS epitope recognized by antibody MW3G3 in human skin, atypical nevi and melanoma metastasis. Cryosections of human skin, atypical nevi, and melanoma metastasis were stained with antibody MW3G3. The origin of the sections is indicated in the lower right corner (Skin, normal skin; AN, atypical nevi; MM, metastatic melanoma). MW3G3 staining patterns are shown on the left, and IO3H10 (anti-chondroitin sulfate) staining patterns are shown on the right. Note that antibody MW3G3 stained the basal lamina of the epithelium in normal skin and atypical nevi (arrows). No staining was observed in the extracellular matrix of metastatic melanoma, which is rich in glycosaminoglycans (as demonstrated by antibody IO3H10). In addition, antibody MW3G3 does not stain with blood vessels (arrowheads), whereas antibody IO3H10 does (arrowheads). Bar, 25 m.
observation that AS can act as an anti-tumor agent (5), suggesting that AS-like residues are inhibitory for tumor growth.
In conclusion, we have developed an antibody that strongly reacts with AS and a rare oligosaccharide structure present in HS/heparin. IdoA2S and GlcNAc residues are essential for antibody recognition. The synthesis of this epitope is tightly regulated in adult tissues, and it may have important functions for controlling cell growth and migration. It will be very interesting to evaluate the expression of the MW3G3 epitope in embryonic development and in human cancer (in view of the finding with human metastatic melanoma). Reduced levels of IdoA2S-GlcNAc units may be predictive of metastatic potential in human tumors.