Synthesis of 3-O-Sulfated Disaccharide and Tetrasaccharide Standards for Compositional Analysis of Heparan Sulfate

Abstract

3-O-Sulfation on the glucosamine sugar unit in heparan sulfate (HS) is linked to various biological functions, including the anticoagulant activity to treat thrombotic disorders in hospitals. The 3-O-sulfated glucosamine is biosynthesized by heparan sulfate glucosamine 3-sulfotransferases. Because of its biological significance, there is a need for 3-O-sulfated oligosaccharide standards to facilitate the compositional analysis of HS. These oligosaccharides must contain a Δ^4,5-unsaturated uronic acid (ΔUA) residue at the nonreducing end, which is due to the depolymerization reaction catalyzed by heparin lyases used during the compositional analysis procedure. Here, we describe a protocol for the preparation of one 3-O-sulfated disaccharide (compound 4) and three 3-O-sulfated tetrasaccharides (compound 1–3) in a milligram scale. The synthesis of 3-O-sulfated disaccharide and tetrasaccharide standards was completed by degrading synthetic octasaccharides using heparin lyases. Further analysis revealed that 3-O-sulfated oligosaccharide standards are labile under basic conditions, confirming the findings from a previous study. The unwanted degradation was reduced by decreasing the pH in the presence of phosphate buffer. The 3-O-sulfated oligosaccharide standards are reagents to characterize 3-O-sulfation in HS derived from biological sources.

Graphical Abstract

Heparan sulfate (HS) is a linear sulfated polysaccharide containing disaccharide repeating units of alternate 1→4 linked β-d-glucuronic acid (GlcA) or α-l-iduronic acid (IdoA) and α-d-glucosamine residues with various sulfation patterns. Biosynthesis of HS mainly occurs in the endoplasmic reticulum (ER) and Golgi apparatus including the formation of a tetrasaccharide attached to serine side chains of a core protein followed by a long precursor chain of GlcNAc-GlcA repeating unit.^1,2 HS is ubiquitously present in the extracellular matrix and on the cell surface. HS is involved with various biological functions, including infection by bacteria and viruses, blood coagulation, angiogenesis, inflammatory responses, and embryonic development.^3–6

In the past, intense efforts have been devoted to the chemical synthesis of HS oligosaccharides. However, it generally entails multistep synthesis of monosaccharides building blocks, protection, deprotection sequences, low stereoselectivity, and poor yields. This makes the chemical synthesis of HS oligosaccharides a challenging task.^7–10 Chemoenzymatic synthesis is an alternative and straightforward approach for the efficient synthesis of HS oligosaccharides. This protocol comprises the use of glycosyl-transferases, sulfotransferases, and C₅-epimerase to afford a wide range of oligosaccharides with good yields.^11,12 A number of HS oligosaccharides that are synthesized via the chemoenzymatic approach are commercially available from Glycan Therapeutics (www.glycantherapeutics.com).

The sulfation pattern in HS determines its biological function.¹³ One essential aspect of HS-related research is to investigate the relationship between the structure of HS and their biological functions. The IdoA or GlcA and glucosamine residues in HS carry sulfo groups at different OH positions, including 2-O-sulfation of IdoA or GlcA residue and 6-O-sulfation, N-sulfation, and 3-O-sulfation of the glucosamine residues. A 3-O-sulfate group on the glucosamine sugar unit is linked to biological functions, including anticoagulant activity, binding to fibroblast growth factor (FGF) receptors to regulate cellular proliferation and binding to viral glycoprotein D of herpes simplex virus to establish the infection.^14–17 Heparin, an FDA-approved anticoagulant, consists of an antithrombin-binding pentasaccharide sequence and contains a 3-O-sulfated glucosamine residue critical for the anticoagulant activity.

Structural analysis of pharmaceutical heparin and HS isolated from tissues primarily remains at the stage of the disaccharide or oligosaccharide compositional analysis. Mass spectrometer-based methods for high-level sequence analysis are emerging,^18,19 but the technique requires access to sophisticated mass spectrometers and highly trained operators. During the compositional analysis procedure, the polysaccharides are exhaustively degraded by a mixture of lyases to yield disaccharides and tetrasaccharides.²⁰ The degraded products are then subjected to chromatographic analysis. The identities of each different disaccharide and tetrasaccharide are confirmed by the chromatographic analysis by coeluting with authentic standards. A number of disaccharide standards are commercially available, but none of these standards contain 3-O-sulfated glucosamine residues. The lack of reference standards impedes the structural analysis of HS containing 3-O-sulfated glucosamine residues. Therefore, well-characterized, highly pure standards that contain the 3-O-sulfated glucosamine residues are desirable.²¹

In the present study, we report the synthesis of four 3-O-sulfated oligosaccharide standards (compounds 1–4) that contain a Δ^4,5-unsaturated uronic acid at the nonreducing end (Figure 1). These oligosaccharides resemble, in part, those released from heparin lyases degraded pharmaceutical heparin or 3-O-sulfated HS isolated from biological tissues.²² The preparation of 3-O-sulfated oligosaccharide standards was completed from synthetic octasaccharide precursors after the degradation with a mixture of heparin lyases. Although some 3-O-sulfated oligosaccharides have been isolated from heparin or HS,^20,23,24 the purification steps involved in multistep chromatographic purification, typically resulting in only a limited amount of products. The method reported here is scalable because the synthetic octasaccharide precursors are made in the gram scale.²⁵ The availability of 3-O-sulfated oligosaccharide standards provides key research reagents for investigating the complex composition and biological roles of heparin and HS.

Figure 1.

Synthesis of 3-O-sulfated tetrasaccharides (compounds 1–3) and disaccharides (compound 4) from octasaccharide precursors (compounds 5–8). The octasaccharides (5–8) were digested by H. lyases to yield the 3-O-sulfated tetrasaccahrides (1–3) and disaccharide (4). The cleavage sites by H. lyases are indicated.

EXPERIMENTAL PROCEDURES

Preparation of Compounds 1–4.

Octasaccharides were synthesized using a chemoenzymatic approach and are available from Glycan Therapeutics (www.glycantherapeutics.com). Octasaccharides (5 mg/mL) were treated exhaustively with H. lyase I (100 μL, 3.5 mg/mL), H. lyase II (30 μL, 5.5 mg/mL), and H. lyase III (30 μL, 5 mg/mL) in the presence of Na₂HPO₄ buffer (2 mL, 50 mM, pH 7.2) overnight at 37 °C. Recombinant H. lyase I, II, and III are expressed in E. coli and purified by a Ni-agarose column (GE Healthcare).³¹ The reaction was monitored by injecting a small amount of reaction mixture to anion-exchange HPLC (TSKgel DNA-NPR-column (4.6 mm × 7.5 cm, 2.5 μm, from Tosoh Bioscience)) to monitor the formation of products using UV absorbance at 232 and 310 nm. The purification of the product was performed with a Q-Sepharose column using a linear gradient of NaCl from 0 to 1000 mM in 25 mM Na₂HPO₄, pH 4. For compound 3, the digested octasaccharide 7 was purified by using Biogel P2 size exclusive chromatography (BioRad) eluted with 0.1 M ammonium bicarbonate. To prevent degradation in ammonium bicarbonate buffer after P-2 column purification, compound 3 was further purified by Q-Sepharose column. Isolated tetrasaccharide was further redigested using H. lyases, I, II, and III with the same protocol; however, no further digestion was observed.

NMR Analysis of Compounds 1–4.

The full NMR assignment of each compound is shown in Supporting Information. NMR experiments were performed at 298 K on Bruker Avance 600 and 850 MHz spectrometers with Topsin 3.2 software. Samples (0.5–3.0 mg) were each dissolved in 0.5 mL of D₂O (99.996%, Sigma-Aldrich) and lyophilized three times to remove the exchangeable protons. The samples were redissolved in 0.5 mL of D₂O and transferred to NMR microtubes (O.D. 5 mm, Norrell). Compounds 1, 2, 3, and 4 were prepared by adding D₂O in H₂O (5:1) and lock with D₂O and water. The 1D ¹H NMR experiment “zg” pulse sequence was performed with 64 scans and an acquisition time of 3.8 s. The 1D ¹³C NMR experiment “zgdc30” pulse sequence was performed with 10 000 scans and an acquisition time of 1.0 s. The 2D ¹H–¹³C HSQC experiment “hsqcgpph” pulse sequence was performed with 48 scans, 512 increments, 1.5 s relaxation delay, and 120 ms acquisition time. 2D spectra were recorded with GARP carbon decoupling, and 48 dummy scans were used prior to the start of acquisition. 2048 total points were collected in f2. ¹³C transmitter offset was set at 90.0 ppm. 1D ¹H NMR, ¹³C NMR, and 2D experiments for compounds 1–3 were performed using water suppression NMR.

Electrospray Ionization Mass Spectrometry (ESI-MS) Analysis.

Molecular weight conformation of synthesized compounds 1–4 (tetrasaccharides and disaccharides) was determined by the ESI-MS (Thermo LCQ-Deca). ESI-MS analysis was performed in the negative ion mode and with the following parameters: spray voltage at 3.0 kV, curved desolvation line temperature at 120 °C. The mass range was set at 300–1000.

HPLC Analysis.

YMC-Pack polyamine II column (from YMC) was applied to detect the degree of completion of the reaction and the purity of synthesized compounds 1–4 after purification. All compounds were purified using Q-Sepharose chromatography. Mobile phase A was 20 mM NaOAc, pH 5. Mobile phase B was 20 mM NaOAc and 1 M NaCl, pH 5. The gradient step was 0–100% B in 120 min with a flow rate of 2 mL/min. Mobile phase A was 25 mM Na₂HPO₄, pH 4. Mobile phase B was 25 mM Na₂HPO₄ and 1 M NaCl, pH 4. The gradient step was 0–100% B in 120 min with a flow rate of 2 mL/min. Purification of compounds 1–4 was performed in 25 mM phosphate buffer at pH 4. Formic acid (0.1 M) was used to lower the pH to 3.0. The absorption at 310, 232, and 260 nm was used to monitor the eluent.

RESULTS AND DISCUSSION

Three 3-O-sulfated tetrasaccharides, including compounds 1, 2, and 3, and one 3-O-sulfated disaccharide (compound 4) were generated by the depolymerization of synthetic octasaccharide precursors (Figure 1 and Table 1). Octasaccharide precursors (compounds 5–8, Figure 1) were synthesized via the chemoenzymatic method as reported previously.¹² The octasaccharide precursors were digested exhaustively with heparin lyases (H. lyase) I, II, and III in a phosphate buffer to yield the products as illustrated in Figure 1A,B. H. lyases reportedly display different reactivity toward different linkages in HS. H. lyase I specifically cleaves the glycosidic bond between N-sulfated glucosamine (GlcNSO₃) and 2-O-sulfated iduronic acid (IdoA2S). H. lyase II also cleaves the bond between N-sulfated or N-acetylated glucosamine and glucuronic or iduronic acid, whereas H. lyases III cleaves the bond between N-sulfated or N-acetylated glucosamine (GlcNSO₃ or GlcNAc) and glucuronic acid (GlcA).²³ During the experiment, we used a mixture of H. lyase I, II, and III to obtain maximum depolymerization of octasaccharide precursors. The conditions employed were exhaustive digestions; namely, the products should be disaccharides after the reaction. However, the presence of 3-O-sulfated glucosamine residue renders the resistance to the mixture of H. lyases digestion for the glycosidic bond at the nonreducing end resistance,^20,26,27 resulting in 3-O-sulfated tetrasaccharide products ΔUA-GlcNS/Ac6S-GlcA-GlcNS3S6S.²² Indeed, we obtained compound 1, ΔUA-GlcNAc6S-GlcA-GlcNS3S6S, and compound 2, ΔUA-GlcNS6S-GlcA-GlcNS3S6S, from degradation of appropriate octasaccharide precursors. It should be noted that the GlcNS3S6S residue in compounds 1 and 2 is adjacent by a GlcA residue at its nonreducing end with the sequences of -GlcA-GlcNS3S6S-.

Table 1.

Structures and Molecular Mass Properties of Compounds 1–4

cmpd	sequence of the oligosaccharides	measured MWa	calcd MW	amount (mg)
1	ΔUA-GlcNAc6S-GlcA-GlcNS3S6S	1036.2	1036.8	6
2	ΔUA-GlcNS6S-GlcA-GlcNS3S6S	1074.2	1074.9	8
3	ΔUA-GlcNS-IdoA2S-GlcNS3S	994.4	994.8	4
4	ΔUA2S-GlcNS3S6S	657.2	657.5	5

^aMeasured MW was based on the value of m/z of the molecular ion of [M −2H]² using ESI-MS.

The impact on the resistance to H. lyases digestion appears to be different if the 3-O-sulfated glucosamine residue is under different contexts of saccharide sequence as demonstrated in the preparation of compounds 3 and 4. To prepare compound 4, compound 8 was subjected to the degradation to H. lyases (Figure 1C). The 3-O-desulfated glucosamine residue in compound 4 contains a 3-O-sulfated glucosamine residue linked to an IdoA2S residue at the nonreducing end. We obtained a disaccharide product, instead of a tetrasaccharide. This disaccharide was previously identified from 3-O-sulfotransferase isoform 5 modified HS and HS from Chinese hamster ovary cells expressing 3-O-sulfotransferase isoform 3.^28,29 It should be noted that we obtained a tetrasaccharide compound 3 from compound 7 using heparin lyases (Figure 1B). The structural difference between octasaccharide precursor 7 and 8 is that 8 contains 6-O-sulfation, whereas 7 does not. Our results demonstrate that the 6-O-sulfation in 8 diminishes the resistance effect from the 3-O-sulfated glucosamine residue toward H. lyases digestion; however, such effect only occurs when the 3-O-sulfated glucosamine residue is in the context of -IdoA2S-GlcNS3S6S-.

The structural characterization of compounds 1–4 was completed using electrospray ionization mass spectrometry (ESI-MS) and NMR. Measured molecular weights (MWs) of compounds 1–4 are very close to calculated MWs (Table 1). NMR analyses of compounds 1 and 2 have been reported previously.^20,24 Our analyses further confirmed the structures of these two compounds. Briefly, compounds 1 and 2 were identified with the presence of four anomeric protons in ¹H NMR and corresponding carbon in ¹³C NMR. In the 2D COSY experiment, the peak at δ 5.85 ppm for H-4 in Δ^4,5UA, showing the connectivity to H-3 and further correlation with H-2 and H-1, indicates the presence of hexenepyranosyluronic acid in compound 1. The presence of the N-acetyl group in ¹H NMR indicates the presence of the N-acetylated glucosamine residue. The anomeric proton in glucuronic acid was almost undetectable due to water suppression. However, its anomeric carbon was clearly indicating presence of GlcA in tetrasaccharide. The peak at 5.38 ppm in C-1 with J = 4.0 Hz indicates α (1→4) connectivity with GlcA in the tetrasaccharide. Similarly, compound 2 showing a peak at 5.64 ppm in glucosamine with J = 3.8 Hz indicates α (1→4) linkage with GlcA. The H-3 proton with a SO₃ group in the glucosamine residue was more downfield and appeared as a doublet of the doublet with J = 10.2, 9.2 Hz at 4.50 ppm, whereas H-3 without the SO₃ group appears at 3.74 ppm. Moreover, the peak at 5.81 and 4.66 ppm indicates the presence of Δ^4,5UA and GlcA in compound 2. HSQC and the DEPT experiment indicate the presence of methylene protons in the tetrasaccharide. The remaining protons were assigned by HSQC-TOSCY and COSY experiment. Compound 2 was isolated as a single anomer (α/β = 1:0). The NMR signal assignments are consistent with previous reports (Figures S1–S4).

The structural analysis using NMR was primarily focused on compounds 3 and 4 in the present study as it was not reported previously. Compound 3 was identified by the presence of 4 anomeric carbons in ¹H NMR at 5.42, 5.40, 5.05, and 4.67 ppm, indicative of a tetrasaccharide (Figure 2). A downfield peak at δ 5.75 for H-4 in Δ^4,5-uronic acid, showing the connectivity to H-3 in the COSY experiment, was observed. Further correlation with H-2 and H-1 indicates the presence of Δ^4,5-uronic acid. The unique H-1 and H-5 protons in IdoA2S (residue B) appear as a multiplet at 5.19–5.18 and 4.99–4.85 ppm, respectively. The peak at 5.40 ppm in C-1 shows doublet with J = 3.3 Hz indicates α (1→4) linkage with IdoA2S and peak at 5.04 with J = 6.8 Hz in D-1 confirmed the presence of β(1→4) linkage with glucosamine residue. In the glucosamine residue, both H-6 protons 3.86–3.85 ppm were observed at the upfield region due to a lack of 6-O-sulfation. Typically, O-sulfation at C-6 appeared at 4.20–4.50 ppm in compounds 1 and 2. The presence of the additional 3-O-sulfo group in ring A indicates downfield H-6 protons as compared to ring C, which is clearly observed in the HSQC spectrum. Additional NMR analysis and purity analysis data for compound 3 are shown in Figures S5 and S6.

Figure 2.

¹H NMR spectrum of compound 3. Signals from anomeric protons of each residue are indicated. The signal from the characteristic proton (H-4) from the nonreducing end is indicated.

NMR spectroscopy analysis was performed to solve the structure of 4 (Figure 3). The presence of two anomeric protons at 5.50 and 5.34 ppm with the corresponding two anomeric carbons in the HSQC experiment indicates a disaccharide. The presence of a C-6 sulfo group at 4.26–4.20 ppm appears as a multiplet, which was clearly observed in the HSQC experiment (Figure S8). The downfield position of H-4 at δ 6.06–6.05 ppm in ring B indicates the presence of Δ^4,5-uronic acid. The signal of the A-4 proton at 4.54 ppm was suppressed because of the water in D₂O and sometimes disappeared in proton NMR; however, its corresponding carbon was clearly observed in HSQC (Figure S8). In addition, the correlation of A4-B1 was observed in HSQC-TOCSY. The downfield proton of the 3-sulfo group at 4.35–4.32 appeared as a multiplet in proton NMR (Supporting Information). Additional structural and purity analysis data are shown in Figure S7

Figure 3.

¹H NMR spectrum of compound 4. Signals from anomeric protons and the characteristic proton from the nonreducing residue (H-4) are indicated.

We next examined the stability of compounds 1–4 under different conditions. Although HS disaccharides are generally stable, Huang et al. reported that a trisaccharide with a 3-O-sulfated glucosamine residue at the reducing end is susceptible to degradation under mildly basic conditions.³⁰ In the study, authors only tested the 3-O-sulfated glucosamine in a trisaccharide sequence of GlcNS6S-GlcA-GlcNS3S6S. They also discovered that lowering the pH reduced the degradation rate. Their report prompted us to test if the presence of a 3-O-sulfated glucosamine residue at the reducing end leads to the instability among all four 3-O-sulfated oligosaccharide standards prepared in the present study. To this end, we incubated these compounds in 0.1 M NH₄HCO₃ buffer at pH 8 at room temperature over a period of 48 h. The samples were then subjected to the high-resolution anion-exchange HPLC analysis under different incubation periods. The data for the HPLC analysis for 1 is shown in Figure 4. A steady decrease in the peak intensity for compound 1 and an increase in the degradation products were observed in the first 24 h, and a nearly complete degradation occurred after 36 h. The stability tests for compounds 2–4 were also conducted, and all three compounds were degraded in 0.1 M NH₄HCO₃ buffer (Figures S9 and S11). In a control experiment, a disaccharide without 3-O-sulfation showed only 6% degradation in 0.1 M NH₄HCO₃ buffer after 48 h incubation, confirming that the presence of 3-O-sulfo group at the reducing end accelerates the degradation (Figure S12). We discovered that the stability of 3-O-sulfated oligosaccharides is significantly improved by being stored in 25 mM phosphate buffer at pH 3 at −80 °C. Under this condition, the compounds are stable for 4 months as analyzed by HPLC (Figure S13)

Figure 4.

HPLC chromatograms of compound 1 over time in 0.1 M NH₄HCO₃ buffer. Compound 1 (50 μg/mL) was incubated at room temperature in 0.1 M NH₄HCO₃. The sample was withdrawn periodically and analyzed by HPLC.

Finally, we analyzed the structures of the degraded products by ESI-MS using compound 1 as a model. The mechanism for the degradation reaction was proposed by Huang and colleagues as illustrated in Figure S14.³⁰ Two products were anticipated from 1, including a trisaccharide and a desulfated tetrasaccharide (Figure 5). Indeed, we observed two products after incubating 1 in NH₄HCO₃ buffer over 48 h, confirming the degradation mechanism reported previously.

Figure 5.

ESI-MS analysis of degradation products from compound 1. The left panel shows the ESI-MS spectrum of intact compound 1. The right panel shows the ESI-MS spectrum of degraded compound 1 after it was incubated at room temperature in 0.1 M NH₄HCO₃ for 48 h. Two molecular ions were detected. The molecular ion of 476.7 is associated with the desulfated tetrasaccharide, and the molecular ion of 634.1 is associated with a trisaccharide resulted from the pealing reaction.

In conclusion, we report the preparation of 3-O-sulfated oligosaccharide standards that may be used for the HS disaccharide/oligosaccharide compositional analysis. The unique structural features of these standards contain a Δ^4,5UA residue at the nonreducing end and a 3-O-sulfated glucosamine residue at the reducing end. During our studies, we discovered that a 3-O-sulfated glucosamine residue at the reducing end is highly labile under basic conditions. One way to avoid the degradation is to store the oligosaccharides in a phosphate buffer at pH 3. Other buffer conditions were also tested, such as sodium acetate buffer (pH 5.0) and Tris buffer (pH 7.0). However, none of these buffer conditions were superior to the phosphate buffer to stabilize the 3-O-sulfated oligosaccharides (data not shown). The 3-O-sulfation has been viewed as the rarest modification in HS isolated from biological sources; however, this perception does not consider the unstable nature of 3-O-sulfated oligosaccharides during the compositional analysis. Our results raise a question about the conditions to use H. lyases to conduct the structural analysis of pharmaceutical low-molecular weight heparin recommended by US pharmacopeia and HS to accurately measure the 3-O-sulfated oligosaccharides after H. lyases digestion. The availability of highly pure 3-O-sulfated oligosaccharides will help researchers to characterize the complex and biologically important sulfation patterns in heparin and HS.