A Major Common Trisulfated Hexasaccharide Core Sequence, Hexuronic Acid(2-Sulfate)-Glucosamine(N-Sulfate)-Iduronic Acid-N-Acetylglucosamine-Glucuronic Acid-Glucosamine(N-Sulfate), Isolated from the Low Sulfated Irregular Region of Porcine Intestinal Heparin*

The major structure of the low sulfated irregular region of porcine intestinal heparin was investigated by characterizing the hexasaccharide fraction prepared by extensive digestion of the highly sulfated region with Flavobacterium heparinase and subsequent size fractionation by gel chromatography. Structures of a tetrasaccharide, a pentasaccharide, and eight hexasaccharide components in this fraction, which accounted for approximately 19% (w/w) of the starting heparin representing the major oligosaccharide fraction derived from the irregular region, were determined by chemical and enzymatic analyses as well as 1H NMR spectroscopy. Five compounds including one penta- and four hexasaccharides had hitherto unreported structures. The structure of the pentasaccharide with a glucuronic acid at the reducing terminus was assumed to be derived from the reducing terminus of a heparin glycosaminoglycan chain and may represent the reducing terminus exposed by a tissue endo-β-glucuronidase involved in the intracellular post-synthetic fragmentation of macromolecular heparin. Eight out of the 10 isolated oligosaccharides shared the trisaccharide sequence, -4IdceAα1–4GlcNAcα1–4GlcAβ1-, and its reverse sequence, -4GlcAβ1–4GlcNAcα1–4IdceAα1-, was not found. The latter has not been reported to date for heparin/heparan sulfate, indicating the substrate specificity of the d-glucuronyl C-5 epimerase. Furthermore, seven hexasaccharides shared the common trisulfated hexasaccharide core sequence ΔHexA(2-sulfate)α1–4GlcN(N-sulfate)α1–4IdceAα1–4GlcNAcα1–4GlcAβ1–4GlcN(N-sulfate) which contained the above trisaccharide sequence (ΔHexA, IdceA, GlcN, and GlcA represent 4-deoxy-α-l-threo-hex-4-enepyranosyluronic acid, l-iduronic acid, d-glucosamine, andd-glucuronic acid, respectively) and additional sulfate groups. The specificity of the heparinase used for preparation of the oligosaccharides indicates the occurrence of the common pentasulfated octasaccharide core sequence, -4GlcN(N-sulfate)α1–4HexA(2-sulfate)1–4GlcN(N-sulfate)α1–4IdceAα1–4GlcNAcα1–4GlcAβ1–4GlcN(N-sulfate)α1–4HexA(2-sulfate)1-, where the central hexasaccharide is flanked by GlcN(N-sulfate) and HexA(2-sulfate) on the nonreducing and reducing sides, respectively. The revealed common sequence constituted a low sulfated trisaccharide representing the irregular region sandwiched by highly sulfated regions and should reflect the control mechanism of heparin biosynthesis.

Heparin is a highly sulfated co-polymer of glucosamine and uronic acid residues that are alternatively 134-linked. Most of the heparin molecule is accounted for by the major trisulfated disaccharide repeating unit, 34IdceA(2-sulfate)␣134GlcN (N,6-disulfate)␣13. This repeating sequence forms highly sulfated regions and represents at least 75% heparin from porcine intestine (1). Undersulfation and substantial structural variability are observed in the rest of the region which is called the irregular region and distributed along the chain flanked by the fully sulfated region composed of the trisulfated disaccharide units, accounting for approximately one-quarter of the heparin polysaccharide chain (for reviews see Refs. [2][3][4]. Heparin exhibits a wide range of biological activities such as inhibition of blood coagulation (5), modulation of cellular proliferation (6,7), potentiation of angiogenesis (8), and interactions with various growth factors (9 -12). These activities result from the ability of heparin to interact with various proteins causing their activation, deactivation, or stabilization. Interactions between heparin and proteins generally depend on the presence of sulfate groups. Some proteins such as lipoprotein lipase (13), thrombin (14), and platelet factor 4 (15) bind to the highly sulfated region consisting exclusively of the trisulfated disaccharide unit in a seemingly nonspecific fashion. Still, many other proteins are thought to require specific sequences for binding, and the precise requirement for individual sulfate groups may vary from one protein to another. However, it should be noted that heparin is oversulfated and contains not only the essential sulfate groups but also additional nonessential ones for protein binding. The contribution of the irregular region to the biological activities of heparin is not well understood mainly because of the difficulty in analyzing the variably sulfated structure. However, it is well known that the antithrombin-binding minimum pentasaccharide sequence GlcN(6S) 1 -␣1-4GlcA␤1-4GlcN(NS,3S)␣1-4IdceA(2S)␣1-4GlcN(NS,6S) (16,17) is embedded over the connecting border of the high and low sulfated regions and contains both low and high sulfated disaccharide units. Hence, it is conceivable that the low sulfated irregular region is involved in some other active domains as well.
Biosynthetic reactions required to generate heparin sequences include the formation of an initial, simple polysaccharide structure, 34GlcA␤134GlcNAc␣13, that is subsequently modified through N-deacetylation/N-sulfation of GlcNAc units, C-5 epimerization of GlcA to IdceA units, and O-sulfation at C-2 of IdceA and C-6 of GlcN units (for review see Ref. 4). The undersulfated region of the final product heparin is thought to have escaped such modifications. However, the mechanism by which certain residues are selected for modification is not understood, partly because of a lack of sequence information on relatively large fragments although certain rules about the sequential arrangement of various disaccharide units have been noted (4).
In this study we isolated and systematically characterized 10 oligosaccharide structures from the undersulfated region of porcine intestinal heparin after extensive digestion with heparinase, which acts only on the highly sulfated repeating region, to investigate the variability and/or regularity, if any, of the low sulfated region. A majority of these oligosaccharides were revealed to share a common trisulfated hexasaccharide core sequence.
Matrix-assisted Laser Desorption Ionization (MALDI) Time-of-flight (TOF) Mass Spectrometry-MALDI TOF mass spectra of a sulfated heparin oligosaccharide were recorded on a Konpact MALDI 1 (Shi-madzu/Kratos) linear instrument in the positive ion mode at Toray Research Center, Kanagawa, Japan. Ultraviolet MALDI experiments were carried out using an N 2 laser (Laser Science, Newton, MA; 337-nm wavelength, 3-ns pulse width). The ions were accelerated to 20 keV energy. Caffeic acid was used as a matrix at a concentration of 10 mg/ml in 1:1 water/MeCN mixture. A synthetic peptide (Arg-Gly) 15 was used as a complexing agent to shield the negatively charged groups of a sulfated oligosaccharide according to Juhasz and Biemann (29). Aqueous solutions of a heparin oligosaccharide (10 pmol/l) and the peptide (10 pmol/l) were mixed in advance and then diluted with an equimolar proportion of the matrix solution. Of this sample/matrix solution, 0.5-1.0 l was placed on the probe surface and dried under a stream of air.
Other Analytical Methods-Uronic acid was determined by the carbazole method (30). Unsaturated uronic acid was spectrophotometrically quantified based upon an average millimolar absorption coefficient of 5.5 at 232 nm (31). Amino acids and amino sugars were quantified after acid hydrolysis in 6 M HCl at 110°C for 20 h or 3 M HCl at 100°C for 16 h, respectively, using a Beckman 6300E amino acid analyzer (32). Capillary electrophoresis was carried out to examine the purity of each isolated fraction in a Waters capillary ion analyzer as reported (33).
Enzymatic Characterization of the Isolated Oligosaccharides-Disaccharide compositions of the isolated oligosaccharides were determined by digestion with heparitinases I and/or II, followed by HPLC analysis on an amine-bound silica column. Substrate specificities of the three heparin lyases are shown in Fig. 2. Most of the fractions except for fractions 6-32 and 6-34 were degraded into disaccharides by the enzymes, although fractions 6-32 and 6-34 were degraded into approximately 1 mol each of a di-and a tetrasaccharide unit. The results are summarized in Table I. Recoveries of the oligosaccharides in Table I were calculated taking the absorbance of the parent oligosaccharide in each fraction as 100%. Excess or insufficient recoveries of the products were observed in some cases partly due to the products derived from possible impurities and partly due to the use of the average millimolar absorption coefficient of 5.5 at 232 nm obtained from unsaturated disaccharides (31) for quantification of oligosaccharides. Millimolar absorption coefficients of individual oligosaccharides deviate from the average value to various degrees. As representative chromatograms, those obtained with fractions 6-23 and -27 are shown in Fig. 3.  (Table I) and 500-MHz 1 H NMR analysis. 500-MHz 1 H NMR Analysis-All the individual oligosaccharides were also analyzed by 500-MHz 1 H NMR, and individual monosaccharide units were identified based on the chemical shifts of the proton signals and the coupling constants J 1,2 . Chemical shifts were assigned by two-dimensional homonuclear Hartmann-Hahn and correlation spectroscopy analyses (data not shown) as reported for the sulfated oligosaccharides isolated previously from heparin (26) and heparan sulfate (33). Although some fractions were still mixtures as revealed by capillary electrophoresis, it was possible to extract sequence information about the major compound in these fractions by taking advantage of 1 H NMR spectroscopy. Since peak heights of resonances reflect molar ratios of the components, signals for the major compound in the mixtures could be easily distinguished from those of the minor impurities. The internal uronic acid residues of each isolated oligosaccharide were unambiguously identified based upon the chemical shifts of the anomeric proton signals and the coupling constants J 1,2 . Anomeric proton signals of an ␣IdceA and a ␤GlcA residue in heparin/ heparan sulfate oligosaccharides are observed around ␦ 5.2-5.0 and 4.7-4.5, respectively (22,35). The coupling constants J 1,2 of ␣IdceA and ␤GlcA in heparin/heparan sulfate oligosaccharides are approximately 3.0 and 8.0 Hz, respectively (22,36). The NMR data obtained in this study for the oligosaccharides are summarized in Table II. In the spectrum of fraction 6-25, six individual saccharide residues were readily identified. Two internal uronic acid residues were determined as IdceA-4 and GlcA-2 based on the chemical shifts of the anomeric proton signals, at ␦ 5.192 and 4.574, respectively, as well as the coupling constants J 1,2 of 2.0 and 8.0Hz, respectively. The chemical shifts of H-1 and H-2 of the IdceA residue of the compound in fraction 6-25 were shifted downfield by approximately 0.2 and 0.6 ppm, respectively, when compared with those of the nonsulfated IdceA residue of the oligosaccharides isolated from bovine kidney heparan sulfate or porcine intestinal heparin (22,33). In contrast, the proton chemical shifts of the GlcA residue of the compound in this fraction were very similar to those of the nonsulfated GlcA residue. These results indicated the 2-sulfation of IdceA-4 and the nonsulfation of GlcA-2 of this compound. Based upon these NMR data and the sequential arrangement of the disaccharide units determined by enzymatic analysis, the structure of the major compound in fraction 6-25 has been determined as the following.   (Table I). Taking the UV absorbance of the parent hexasaccharide in fraction 6-30 as 100%, the total recovery of these disaccharide products was greater than 300%, suggesting that the millimolar absorption coefficient of the parent hexasaccharide may be smaller than those of the produced disaccha- The hexasaccharide fraction obtained from gel filtration on Bio-Gel P-10 was separated into subfractions 6-1 to 6-37, on an amine-bound silica column using an NaH 2 PO 4 gradient (indicated by the dashed line). Fraction 6-28 was the right shoulder of fraction 6-27 and was removed by rechromatography. For experimental details, see "Experimental Procedures." rides. Upon successive digestion with ⌬hexuronate-2-sulfatase and then heparitinase II, it yielded ⌬HexA-GlcN(NS,6S) and ⌬HexA(2S)-GlcNAc with the recoveries of 183 and 112%, respectively (Table I). These results indicated that ⌬HexA(2S)-GlcN(NS,6S) was located on the nonreducing terminus. When digested with heparitinase I, fraction 6-30 gave rise to equimolar amounts of ⌬HexA-GlcN(NS,6S) and a component that eluted near the elution position of a tetrasulfated tetrasaccharide (Table I), indicating that ⌬HexA-GlcN(NS,6S) was derived from the reducing end. Therefore, the compound in fraction 6-30 is a hexasulfated hexasaccharide with a sequence of ⌬HexA(2S)-GlcN(NS,6S)-HexA(2S)-GlcNAc-HexA-GlcN(NS,6S).
In the spectrum of fraction 6-30, two internal uronic acid residues were identified as IdceA-4 and GlcA-2 based on the chemical shifts of the anomeric proton signals, at ␦ 5.203 and 4.572, respectively, as well as the coupling constants J 1,2 of 1.5 and 7.5 Hz, respectively. The chemical shifts of H-1 and H-2 of the IdceA residue of the compound in fraction 6-30 were shifted downfield by approximately 0.2 and 0.6 ppm, respectively, when compared with those of the nonsulfated IdceA residue of the oligosaccharides isolated from bovine kidney heparan sulfate or porcine intestinal heparin (22,33). In contrast, the proton chemical shifts of the GlcA residue of the compound in this fraction were very similar to those of the nonsulfated GlcA residue. These results indicated the 2-sulfation of IdceA-4 and the nonsulfation of GlcA-2 of this compound. Based upon these NMR data and the sequential arrangement of the disaccharide units determined by enzymatic analysis, the structure of the major compound in fraction 6-30 has been determined as the following. Fraction 6-22-Fraction 6-22 was resolved into several subcomponents by capillary electrophoresis (data not shown), the major component accounting for only 57% of the UV absorbing materials in this fraction. However, it was not possible to fractionate it preparatively into its subcomponents. Therefore, it was first digested with ⌬hexuronate-2-sulfatase and then the digest was analyzed by HPLC. The ⌬hexuronate-2-sulfatase digest gave four peaks in a molar ratio of 7:24:54:15, all of which eluted 11-16 min earlier than the parent fraction (data not shown), indicating that fraction 6-22 was a mixture of at least four different compounds. The major product, designated as fraction 6-22S-1, was isolated and subjected to structural analysis. The yield of fraction 6-22S-1 was 42 nmol/100 mg starting heparin. When examined by capillary electrophoresis, fraction 6-22S-1 was 69% pure (results not shown). Upon heparitinase II digestion, fraction 6-22S-1 yielded ⌬HexA-Glc-N(NS) and ⌬HexA(2S)-GlcNAc with the recoveries of 220 and 66%, respectively, taking the UV absorbance of the total parent oligosaccharides in fraction 6-22S-1 as 100% (Table I). Amino sugar and uronic acid analyses showed that fraction 6-22S-1 contained a hexasaccharide as a major compound. These results altogether indicate that the major compound in this fraction was composed of 2 mol of ⌬HexA-GlcN(NS) and 1 mol of ⌬HexA(2S)-GlcNAc. The excess recovery of ⌬HexA-GlcN(NS) upon heparitinase II digestion over the other product was probably due to the disaccharide produced by degradation of contaminating oligosaccharide(s). Since this fraction was isolated after ⌬hexuronate-2-sulfatase digestion, the disaccharide unit on the nonreducing terminus of the parent hexasaccharide was not ⌬HexA(2S)-GlcNAc but ⌬HexA-GlcN(NS). Heparitinase I digestion of both fractions 6-22 and 6-22S-1 resulted in mainly two unsaturated components, the monosulfated disaccharide ⌬HexA-GlcN(NS) and a presumable tetrasaccharide component that eluted near the elution position of the tri-or disulfated tetrasaccharide. Recoveries of the di-and tetrasaccharide components from fraction 6-22 or 6-22S-1 were 56 and 49% or 91 and 83%, respectively (Table I). The low recoveries of the components from fraction 6-22 were due to the corresponding low content of the major component in the fraction. Since the presumable tetrasaccharide peak from fraction 6-22 was shifted by the prior ⌬hexuronate-2-sulfatase digestion to the position corresponding to the loss of one sulfate group on HPLC, the tetrasaccharide component was derived from the nonreducing terminus. Therefore, the structure of the compound in fraction 6-22S-1 is ⌬HexA-GlcN(NS)-HexA(2S)-Glc-NAc-HexA-GlcN(NS). Consequently, the major compound in the parent fraction 6-22 is a tetrasulfated hexasaccharide with a sequence of ⌬HexA(2S)-GlcN(NS)-HexA(2S)-GlcNAc-HexA-GlcN(NS).
In the spectrum of fraction 6-22S-1, two internal uronic acid residues were identified as IdceA-4 and GlcA-2 based on the chemical shifts of the anomeric proton signals, at ␦ 5.217 and 4.542, respectively, as well as the coupling constants J 1,2 of 2.  (22,23,34). The C-3 OH of the glucosamine residue (marked by an asterisk) adjacent to the disaccharide cleavage site has to be free for sensitivity to heparitinases I and II (26). due of the oligosaccharides isolated from bovine kidney heparan sulfate or porcine intestinal heparin (22,33). In contrast, the proton chemical shifts of the GlcA residue of the compound in this fraction were very similar to those of the nonsulfated GlcA residue. These results indicated the 2-sulfation of IdceA-4 and the nonsulfation of GlcA-2 of this compound. Based upon these NMR data and the sequential arrangement of the disaccharide units determined by enzymatic analysis, the structure of the major compound in fraction 6-22S-1 has been determined as the following: fraction 6-22S-1, ⌬HexA␣1-4GlcN(NS)␣1-4IdceA(2S)␣1-4GlcNAc␣1-4GlcA␤1-4GlcN(NS). Since fraction 6-22S-1 was isolated after ⌬hexuronate-2-sulfatase digestion as described above, the major compound in the parent fraction 6-22 contained the following structure.   (Table II) were indistinguishable from those of fraction b-15 obtained from porcine intestinal heparin reported previously (24). Hence, the major compound in fraction 6-27 was identical with ⌬HexA(2S)␣1-4GlcN(NS,6S)␣1-4IdceA␣1-4GlcNAc(6S)␣1-4GlcA␤1-4GlcN(NS,6S). Likewise, 1 H NMR data of fractions 6-17, -32, and -34 (Table II) were also indistinguishable from those of fractions VII, b-20, and b-22, respectively, that were obtained from porcine intestinal heparin as reported previously (22,24), indicating that the major compounds in fractions 6-17, -32, and -34 were identical with those in the compounds in the above fractions, respectively. These structures are in good agreement with the results obtained from enzymatic characterization (Table I). Therefore, the struc-  GlcN(NS,6S). The peak marked by an asterisk around 35 min is often observed upon high sensitivity analysis and is due to an unknown substance eluted from the column resin.  Fraction 6-31-The spectral data of fraction 6-31 were compared with those of fraction b-19S, which were obtained from porcine intestinal heparin after ⌬hexuronate-2-sulfatase digestion as reported previously (24). No significant differences were observed except for the downfield shifts (⌬0.05-0.84 ppm) of the proton signals belonging to the nonreducing terminal ⌬HexA. It should be noted that in our previous report (24), the resonances of H-2 and -3 of the compound in fraction b-19S had been interchanged by mistake. The results indicated that the major compound in fraction 6-31 has an additional sulfate group on C-2 of the nonreducing terminal ⌬HexA in the structure for the compound in fraction b-19S. Therefore, the structure of the compound in fraction 6-31 is the following, and this structure is in good agreement with the results obtained from the enzymatic analysis (Table I).   GlcN(NS,6S) with the recoveries of 53, 67, and 67%, respectively, taking the UV absorbance of the total parent oligosaccharides in fraction 6-26 as 100% (Table I). The low recoveries of the major unsaturated disaccharides were probably due to the corresponding low content of the major component in fraction 6-26. The major compound in this fraction was H chemical shifts of the constituent monosaccharides of the isolated oligosaccharides derived from heparin Chemical shifts are given in ppm downfield from internal sodium 4,4-dimethyl-4-silapentane-1-sulfonate but were actually measured indirectly relative to acetone (␦ 2.225 ppm) in 2 H 2 O at 26°C. The estimated error for the values to two decimal places was only Ϯ0.01 ppm because of partial overlap of signals. That for the values to three decimal places was Ϯ0.002 ppm. Coupling constants J 1,2 (in Hz) are given in parentheses. a pentasulfated hexasaccharide composed of equimolar amounts of the above three components and accounted for only 53%, as judged by the recovery of one of the three major disaccharides. The sensitivity of this fraction to ⌬hexuronate-2-sulfatase indicated that the disaccharide unit on the nonreducing terminus was ⌬HexA(2S)-GlcN(NS,6S) (Table I) The two internal uronic acid residues of the hexasaccharide in fraction 6-26 were identified as IdceA and GlcA, based on the chemical shifts (␦ 4.997 and 4.538) of the anomeric proton signals and the coupling constants J 1,2 (1.5 and 8.0 Hz), respectively. Analysis of the deamination products identified the location of internal uronic acid residues. A tetrasaccharide component derived from the reducing side of the parent hexasaccharide was obtained by nitrous acid degradation. It was sensitive to ␣-iduronidase but resistant to ␤-glucuronidase (results not shown), indicating that the uronic acid HexA-4 at the nonreducing end was IdceA, and the GlcA residue was in turn localized at position 2 of the hexasaccharide. Thus, the sequence of the major compound in fraction 6-26 is ⌬HexA(2S)-GlcN(NS,6S)-IdceA-GlcNAc(6S)-GlcA-GlcN(NS), which was confirmed by 500-MHz 1 H NMR spectroscopy (Table II). The proton chemical shifts of ⌬HexA-6, GlcN-5, IdceA, GlcNAc-3, and GlcA of the compound in fraction 6-26 were very similar to those of ⌬HexA-6, GlcN-5, IdceA-4, GlcNAc-3, and GlcA-2 of the compound in fraction 6-27. Those of GlcN-1 were also analogous to those of GlcN-1 of the compound in fraction 6-27 except for the upfield shifts of H-5, H-6, and H-6Ј. These results indicated that the compound in fraction 6-26 lacks a sulfate group on C-6 of the GlcN-1 in the structure of the compound in fraction 6-27. Therefore, the structure of the major compound in fraction 6-26 is the following. Fraction 6-23-Amino sugar and uronic acid analyses showed that fraction 6-23 contained 1.8 mol of HexA and 1.9 mol of GlcN/mol of ⌬HexA, where the GlcN value has been corrected for the degradation (16%) during acid hydrolysis, indicating that the major compound in this fraction is a pentasaccharide. Upon exhaustive heparitinase I digestion of fraction 6-23, the absorption at 232 nm doubled, and ⌬HexA-GlcNAc(6S) and ⌬HexA(2S)-GlcN(NS,6S) were observed with recoveries of 88 and 121%, respectively, taking the UV absorbance of the total parent oligosaccharides in fraction 6-23 as 100% ( Fig. 3A and Table I). The expected unsaturated uronic acid residue to be derived from the reducing end was not detected probably since it was labile and decomposed into an ␣-keto acid as reported previously for the enzymatic digestions of unsaturated di-and trisaccharide from chondroitin sulfate (37,38). The excess recovery of ⌬HexA(2S)-GlcN(NS,6S) over the other products was probably due to the disaccharide derived from contaminating oligosaccharide(s). When fraction 6-23 was digested successively with ⌬hexuronate-2-sulfatase and then heparitinase II, it yielded ⌬HexA-GlcN(6S) and ⌬HexA-GlcN(NS,6S) (Table I). These results indicated the presence of the nonreducing end tetrasaccharide sequence of ⌬HexA(2S)-GlcN(NS,6S)-HexA-GlcNAc(6S)-. However, it remained to be determined whether the reducing terminal uronic acid was sulfated or not.
To define the pentasaccharide structure, fraction 6-23 was analyzed by mass spectrometry. However, it was not possible to obtain spectra of good quality by fast atom bombardmentionization mass spectrometry unlike for heparin tetrasaccharides (23) probably due to the high negative charge. The fraction was successfully analyzed by MALDI TOF mass spectrometry, where the negatively charged groups of the oligosaccharide were shielded with a synthetic peptide as a complexing agent (Arg-Gly) 15 (29). Internal calibration by the peptide yielded a molecular ion signal of the protonated 1:1 complex at m/z 4434 (data not shown). After subtracting the contribution of the protonated peptide (m/z 3217), the molecular mass of the oligosaccharide was calculated to be 1217 Da, in reasonable agreement with the theoretical value (1213 Da) for an unsat- urated tetrasulfated pentasaccharide ⌬HexA 1 HexA 2 HexNAc 1 -HexN 1 (OSO 3 H) 3 . The mass accuracy of the present method could be lower (Ϯ0.2-0.3%) (29) than that obtained using a spectrometer equipped with the recently developed delayed ion extraction device (39) but still sufficed to determine the number of saccharide units and sulfate groups present. Hence, the major compound in this fraction is a pentasaccharide with a sequence of ⌬HexA(2S)-GlcN(NS,6S)-HexA-GlcNAc(6S)-HexA.
The spectrum of fraction 6-23 ( Fig. 4) contained additional H-1 signals in the anomeric region when compared with fraction 6-17 containing tetrasaccharides. Characteristic H-1 resonances at ␦ 5.2 and 4.6 led to identification of ␣GlcA and ␤GlcA residues at the reducing end, respectively (38), confirming that the reducing terminal sugar residue was GlcA. The signals of the nonreducing terminal trisaccharide region were very similar to those of fraction 6-27, showing the presence of the structural element, ⌬HexA(2S)-GlcN(NS,6S)-IdceA-. Compared with the chemical shifts of the protons belonging to the reducing terminal GlcA residue of the reference compound, ⌬HexA␣1-3GalNAc(4-sulfate)␤1-4GlcA, isolated from commercial pig skin dermatan sulfate (38), no significant differences were observed, confirming the presence of a GlcA residue at the reducing terminus of the compound in fraction 6-23. The NMR data indicated that the compound in fraction 6-23 has the pentasaccharide structure which contains a GlcA residue extension on the reducing side. Based upon the NMR data, the MALDI TOF mass data, and the sequential arrangement of the disaccharide units determined by the enzymatic analysis, the structure of the compound in fraction 6-23 was deduced as the following tetrasulfated pentasaccharide structure.

DISCUSSION
In this study, we determined the structures of a tetra-, a penta-, and eight hexasaccharides isolated from the heparin hexasaccharide fraction, which was prepared by the extensive digestion of porcine intestinal heparin with Flavobacterium heparinase. Five subfractions, 6-22, -23, -25, -26, and -30, were isolated for the first time as discrete structures. Since heparinase cleaves most glucosaminidic linkages in the highly sulfated region (22,40), which accounts for three-quarters of a heparin polysaccharide chain, but does not cleave those in the less sulfated irregular region scattered along the polysaccharide chain being flanked by the highly sulfated region, the isolated oligosaccharides are derived from the irregular region. , which is an HNO 2 degradation product of porcine mucosal heparin, has also been isolated (44). However, all these are considered to be derived from the highly sulfated region. A few larger oligosaccharides have also been isolated from the highly sulfated region (41,(45)(46)(47). In contrast, the oligosaccharides isolated in this study are derived from the irregular region of heparin. The isolation of hexa-and larger oligosaccharides from the irregular region has been limited to those derived from the antithrombin III-binding site, and the structural variability of several such oligosaccharides has been summarized and discussed (24,26,48).
The isolated pentasaccharide appears to be derived from the reducing end of a parent heparin glycosaminoglycan chain. During biosynthetic processing, heparin glycosaminoglycan chains are synthesized on the core protein as part of a proteoglycan, released from macromolecular proteoglycans by an endo-␤-glucuronidase, and stored in mast cell granules that may be discharged from cells suitably stimulated (49 -51). The pentasaccharide in fraction 6-23 could be derived from the newly formed reducing end exposed by the specific endo-␤glucuronidase. Endoglycosidase activities toward heparin/ heparan sulfate have also been found in various other cells and tissues (for a review, see Ref. 52). Some heparanases are secreted from cells to play a role in remodeling basement membranes after injury or at inflammation sites. Other heparanases are intracellular and important for degrading cell surface heparan sulfate proteoglycans once they have been internalized. One glucuronidic linkage cleavable by the endoglucuronidase of human platelets is in the sequence -GlcNAc-GlcA-Glc-N(NS)- (53). However, the substrate specificities of most of the endoglycosidases for heparin/heparan sulfate have been only partially characterized. One approach to investigate substrate specificity of endo-␤-glucuronidases is to analyze the reducing and nonreducing terminal structures of heparin glycosaminoglycan chains, which would reflect the specific sites cleaved by endoglycosidases. The nonreducing terminus of a heparin glycosaminoglycan chain will be isolated as a saturated oligosaccharide after digestions with heparin lyases (23,54). The oligosaccharides in fractions 6-26, -27, -32, and -34, which contain the pentasaccharide sequence of the compound in fraction 6-23, may be useful for elucidating the structural requirement for recognition by an endo-␤-glucuronidase. Bame and Robson (55) have recently characterized the reducing terminal structures of heparanase-derived short heparan sulfate chains isolated from