Introduction

Capsosiphon fulvescens, a green alga and commonly called Maesaengi in Korea, grows plentifully along the south-western coast of Korea and also widely throughout the world, including European and North American coasts. Due to its unique flavour and soft texture, this alga has been traditionally consumed in Korea for centuries (Sohn 1998). Maesaengi has a great nutritive potential to be used as human food and as an ingredient in formulated food (Hwang et al. 2008a, b). Moreover, it is rich in vitamins A and C and in minerals such as iron, potassium and iodine and is often used to treat stomach disorders and hangovers (Go et al. 2011). Some biomedical activities like immunostimulatory, anti-cancer, anti-inflammatory, anti-gastritis and anti-atherosclerotic ones were observed for the hot-water extract and crude polysaccharides (natural and chemically modified) isolated from C. fulvescens (Go et al. 2011; Hwang et al. 2008a, b; Kwon and Nam 2007; Karnjanapratum and You 2011; Nam and Kwon 2007a, b, c; Nam et al. 2008). However, potential values and applications of structurally defined compounds originated from this alga and responsible for some biological activities have not been well studied.

Algal polysaccharides have emerged in recent years as a rich and valuable source of bioactive natural compounds. Among those, sulphated polysaccharides (SPS), i.e. those containing hemi-ester sulphate groups in their sugar residues, are a group of most attracting and thus extensively studied algal polysaccharides due to their wide range of pharmacological activities (Berteau and Mulloy 2003; Li et al. 2008). For example, fucoidans, the major SPS of brown algae and one of the most extensively studied SPS, have shown various activities, such as anti-inflammatory (Choi et al. 2010), anti-viral (Lee et al. 2004), anti-coagulant (Chevolot et al. 1999; Boisson-Vidal et al. 2000; Shanmugam and Mody 2000; Kim et al. 2010), anti-tumour and anti-metastatic (Alekseyenko et al. 2007; Synytsya et al. 2010), anti-oxidant (Jing et al. 2008), immunostimulating (Yoo et al. 2007) and anti-angiogenesis (Hahnenberger and Jakobson 1991) activities.

Anti-coagulant and anti-thrombotic activities are among the most widely studied properties of algal SPS for alternative sources of heparinoid-active compounds. Although heparins are commonly used to reduce thrombin formation in patients with arterial thrombotic diseases, serious side effects of these polysaccharides, such as hemorrhage, thrombocytopenia and osteoporosis, have been frequently reported (Fischer 2007; Boisson-Vidal et al. 2000). Furthermore, since heparin is obtained from animal sources, it has potential risk of contamination by animal pathogens such as viruses (Boisson-Vidal et al. 2000). These reasons have led to consideration of other options for the development of effective anti-coagulant with fewer side effects (Mourao and Pereira 1999).

Previously, we reported the isolation and immunomodulating activity of a water-soluble sulphated polysaccharide SPS-CF isolated from a Korean green alga C. fulvescens, Maesaengi (Na et al. 2010). In that paper, it was shown based on the high-performance liquid chromatography (HPLC) analysis of acidic hydrolyzates after 1-phenyl-3-methyl-5-pyrazolone (PMP) labelling of released sugars that the SPS-CF consists mainly of mannose (55.4 mol %), galactose (25.3 mol %), glucuronic acid (16.3 mol %) and arabinose (0.9 mol %), and thus it was suggested that the SPS-CF is a glucuronogalactomannan. However, subsequent experiments performed for elucidating a more detailed structure of SPS-CF using mainly various nuclear magnetic resonance (NMR) techniques including 1H-, 13C- and 2D correlation NMR experiments such as correlation spectroscopy (COSY), rotating-frame overhauser spectroscoPY (ROESY) and heteronuclear single-quantum correlation spectroscopy (HSQC), showed significant signals of deoxyhexose and pentose, which were further identified by GLC as rhamnopyranose and xylopyranose. However, no significant signals of hexoses were found. Thus, the results of NMR analysis lead us to perform a more detailed investigation for the monosaccharide composition of SPS-CF.

The present work is devoted to structural characterisation of this algal polysaccharide mainly by various NMR techniques, mass spectroscopy and monosaccharide composition analyses using several different chromatographic methods and evaluation of its anti-coagulant activity.

Experimental

Isolation and Purification of SPS-CF

Alga Maesaengi (C. fulvescens) was collected from the southern coastal area of Wando, Korea. The extraction and purification of polysaccharide SPS-CF was performed as described earlier (Kim et al. 2007; Na et al. 2010). Briefly, wet algal material (160 g) was immersed to 0.1 mol l−1 HCl (1 l) for overnight at room temperature. The extract was then filtered, neutralised with 1 mol l−1 NaOH and polysaccharides were recovered by 75 % ethanol precipitation. After centrifugation (30 min at 6000×g), the precipitate was dissolved in distilled water and the pH was adjusted to 2.0 with 1 mol l−1 HCl. To remove ballast polysaccharides, solid CaCl2 was added to the final concentration of 2 mol l−1. The resulting precipitate was removed by centrifugation, and the supernatant was collected. The ethanol precipitation and re-dissolving in distilled water was repeated three times, the final solution was dialyzed (molecular weight cutoff (MWCO), 14 kDa) at 4 °C against distilled water for 48 h, and the crude polysaccharide preparation was obtained by freeze-drying as a white material, with the yield of 2.49 % (w/w, in dry mass) (Na et al. 2010).

The crude polysaccharide (50 mg) was dissolved in distilled water (10 ml) and purified on a column (3 × 45 cm) of DEAE cellulose. The column was pre-equilibrated with distilled water and eluted in NaCl gradient (0–3 mol l−1). Fractions were collected and analysed for carbohydrate content by phenol–sulphuric acid method (Dubois et al. 1956). Individual tubes of the major carbohydrate fraction were pooled, exhaustively dialyzed (MWCO, 14 kDa) against distilled water, freeze-dried and stored at −20 °C until use. The major polysaccharide fraction was eluted from the column at approximately 1.3 mol l−1 NaCl in water (Na et al. 2010). Finally, the purified polysaccharide, designated as SPS-CF, was obtained with the yield of 1.32 % of the dry mass of C. fulvescens. The purity of isolated polysaccharide was confirmed by agarose gel electrophoretic analysis as described previously (Lee and Cowman 1994). The isolated polysaccharide migrated on the gel as a single band, showing it was purified to an apparent homogeneity (data not shown).

Monosaccharide Composition and Linkage Analyses

Total and Partial Acidic Hydrolysis

The SPS-CF polysaccharide was dissolved in 400 μl of 2 mol l−1 of trifluoroacetic acid (TFA) and incubated at 100 °C for 4 h in heating block (Fan et al. 1994). After cooling to room temperature and drying on a Speed-Vac Module spin 40 (Biotron, Korea), the hydrolyzates were dissolved in 0.5 ml of distilled water and repeatedly evaporated to dryness to remove residual TFA. The monosaccharide composition of the dried hydrolyzates was performed by (a) HPLC after labelling of sugars with PMP (Honda et al. 1989), (b) high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) (Park et al. 1999), (c) gas chromatography with flame ionising detection (GC-FID) and (d) thin-layer chromatography (TLC). For GC-FID analysis, the SPS-CF polysaccharide was hydrolyzed with 2 mol l−1 TFA but at different conditions (120 °C, 1 h). Partial hydrolysis of SPS-CF with TFA was made according to Lahaye and Ray (1996) with some modifications. Obtained hydrolysates were used for electrospray ionization mass spectrometry (ESI-MS) and matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) analyses (see details below).

PMP Labelling/HPLC Analysis

The sugars in dried hydrolyzates were labelled with PMP according to the method by Honda et al. (1989), and the resulting PMP-labelled monosaccharides were analysed as described earlier (Na et al. 2010) on a Watchers 120 ODS-BP column (250 × 4.6 mm, 20 μm, Watchers, Korea) using an HPLC system P680 equipped with a UV detector UVD170U (Dionex Inc., Sunnyvale, CA, USA) at a flow rate of 0.8 ml/min in a gradient mode with buffers A (0.1 mol l−1 sodium phosphate buffer, pH 7.0, containing 10 % CH3CN) and B (same buffer with 25 % CH3CN). Monosaccharides were identified and quantified by comparing with PMP-labelled authentic ones used as standard.

HPAEC-PAD Analysis

Samples of dried hydrolyzates were dissolved in distilled water and filtered through a 0.45-μm nylon microspin filter before injection onto the column (Park et al. 1999). For HPAEC-PAD analysis, a Dionex DX600 chromatography system (Dionex Inc., Sunnyvale, CA, USA) was used with a pulsed amperometric detector (PAD50) and a CarboPac PA-10 column which was pre-equilibrated with 100 mM NaOH. The monosaccharides in acid hydrolyzates were eluted at 0.8 ml/min with a concentration gradient of 200 mmol l−1 NaOH and DW. A mixture of authentic l-fucose (l-Fuc), l-rhamnose (l-Rha), l-arabinose (l-Ara), d-galactose (d-Gal), d-glucose (d-Glc), d-mannose (d-Man) and d-xylose (d-Xyl) (Sigma, USA) was used as the external standard.

Uronic Acid Analysis

The uronic acids of the hydrolyzate were also identified and quantified by HPAEC-PAD on a CarboPac PA-100 column (Dionex) coupled to a CarboPac guard column. The mobile phase consisted of 100 mM sodium hydroxide (eluent A), 600 mM sodium acetate in 100 mM sodium hydroxide (eluent B) and 500 mM sodium hydroxide (eluent C). Elution conditions were as follows: 95 % A and 5 % B over 0–5 min, 50 % A and 50 % B at 10 min, 20 % A and 80 % B over 15–35 min, 50 % B and 50 % C over 36–43 min and re-equilibrium over 44–50 min (Combo et al. 2012). The flow rate was 0.6 ml/min. A mixture of authentic d-glucuronic acid (d-GlcA) and d-galacturonic acid (d-GalA) (Sigma, USA) was used as the standard.

GC-FID Analysis

Polysaccharide SPS-CF was hydrolyzed with 2 mol l−1 TFA but at different conditions (at 120 °C for 1 h). The quantitative determination of the neutral sugars was carried out in the form of their trifluoroacetates by gas chromatography on a Hewlett-Packard 5890 Series II chromatograph using a PAS-1701 column (0.32 mm × 25 m) at a temperature programme of 80 °C (4 min), 80–140 °C (10 °C min−1), 140–160 °C (4 °C min−1), 160–200 °C (15 °C min−1) and 200 °C (3 min), hydrogen flow rate of 20 ml min−1 and FID detector of 300 °C (Shapira 1969).

TLC Analysis

The TLC analysis of the hydrolyzates was conducted on silica gel plate 10 × 10 cm (Merck, Germany) that was activated in 100 °C dry oven for 5 min before spotting. Samples were developed with mobile phase consisting of 1-propanol, ethyl acetate and water (8:1:1, v/v/v). The chromatogram was visualised by spraying the mixture of H2SO4 and ethanol (4:6, v/v) and drying at 100 °C for 10 min.

ESI and MALDI Mass Spectrometry

The ESI-MS of SPS-CF hydrolysate (0.1 mol l−1 TFA, 100 °C, 75 min) was carried out on a LTQ Orbitrap Velos (Thermo Scientific, USA) Q-TOF2 hybrid mass spectrometer in the positive ion mode. Prior analysis shows the hydrolysate was neutralised with 0.5 mol l−1 NaOH. Sample was introduced at a flow rate of 150 μl min−1 into the electrospray source. In FTMS experiments, the resolution was set to 30,000, scan 50–2000 Da. The spray voltage was maintained at 3 kV, source temperature was 250 °C and capillary temperature 300 °C. MALDI-MS analysis of native SPS-CF and its hydrolysate was performed on the UltrafleXtreme™ (Bruker) operated in the positive or negative ion mode. A solid-state laser (Nd/YAG, 532 nm; SmartBeam, Germany) with a frequency of 50 Hz was used for ionisation. The spectra were acquired using the FlexControl software. Prior analysis polysaccharide sample (1 mg) underwent to the partial hydrolysis (50 μl 1 mol l−1 TFA, 95 °C, 1 h) and fractionated on XC-cartridge (Sigma). Polysaccharide fractions were mixed with a matrix of 2,5-DHB (1.5 μl), loaded by a “dried-droplet” method on the surface of a stainless steel MALDI target and left air-dried.

NMR Spectroscopy

Proton NMR and 13C APT NMR spectra were recorded on Bruker Avance 600 and Bruker Avance 500 NMR spectrometers (Bruker, Germany) in D2O solution at 20 °C. Working frequencies were 600.1 and 499.8 MHz for 1H and 150.9 and 125.7 MHz for 13C, respectively. The spectra were referenced to dioxane as an external standard (δ(1H) = 3.75 ppm, δ(13C) = 69.3 ppm). Correlation spectroscopy such as 1H,1H-COSY, 1H,1H-total correlation spectroscopy (TOCSY), 1H,1H-ROESY, 1H,13C-HSQC and 1H,13C-heteronuclear multiple bond correlation (HMBC) were applied for signal assignment. Processing and analysis of NMR spectra and preparation of the graphs were made using MestReNov. 6.0 (Mestrelab Research, Spain) and Origin 6.0 (Microcal Origin, USA) software.

Other Chemical Analyses

The carbohydrate content was determined by the phenol–sulphuric acid assay (Dubois et al. 1956). The uronic acid content was determined with the 3-hydroxybiphenyl reagent (Blumenkrantz and Asboe-Hansen 1973). The amount of sulphate residues was determined by the Silvestri’s method (Silvestri et al. 1982), using Na2SO4 as a reference. Protein content was estimated according to Bradford (1976).

Anti-coagulation Activity

Human plasma provided from the Blood Bank, Seoul National University Hospital, was stored in a deep freezer (−70 °C). A portion of plasma sample was thawed at room temperature and directly assayed using the activated partial thromboplastin time (aPTT), prothrombin time (PT) and thrombin time (TT) assay kits. The aPTT clotting assay was carried out by the method of Andersson (1976). Normal human platelet-poor plasma (100 μl) was mixed with 3 μl of SPS-CF sample solution (final concentrations of 10–200 μg ml−1) and incubated for 2 min at 37 °C. For the control, equal volume of distilled water was added instead of polysaccharide solution. For comparison, commercial heparin (202 units mg−1, Sigma-Aldrich) and dextran (200 μg ml−1, Sigma, USA, the value of M w round down to 71,000 Da) were examined in the same way and compared with that of SPS-CF. To the reaction mixture, 100 μl of aPTT reagent (Sigma, STA®-PTT A) was added, incubated for 3 min at 37 °C, and then the clotting time was recorded using a Fibrometer (BBL Fibrosystem, Fisher Scientific) following the addition of 100 μl of 0.025 mol l−1 CaCl2. The TT assay was conducted using the method of Maraganore et al. (1989). Normal human plasma (100 μl) was mixed with 3 μl of SPS-CF sample solution (final concentrations of 10–200 μg ml−1) and incubated for 3 min at 37 °C. Then, 200 μl of TT reagent (STA®-Thrombin) was added and clotting time was recorded. The PT assays were carried out using the method of Matsubara et al. (2001). Normal human plasma (100 μl) was mixed with 3 μl of SPS-CF sample solution and incubated for 3 min at 37 °C. The clotting times were recorded after the addition of 200 μl of PT reagent (Neoplastine® CI plus). For the control, equal volume of distilled water was added instead of polysaccharide solution. Values were expressed as the mean of minimum triplicate tests.

Results and Discussion

Monosaccharide Composition of SPS-CF

As it was mentioned above, the results of NMR analyses of SPS-CF described below lead us to perform more extensive experiments for determination of its monosaccharide composition. We performed again the PMP labelling/HPLC analysis of sugars in the TFA hydrolyzates of SPS-CF following exactly the same protocols and the same HPLC systems that we routinely use in our laboratory as described in our previous paper (Na et al. 2010), and, in addition, newly performed other well-known methods for the determination of monosaccharide composition, i.e. HPAEC-PAD (Park et al. 1999), GC-FID (Shapira 1969) and TLC analyses. The resulting monosaccharide composition of SPS-CF is summarised in Table 1.

Table 1 Chemical composition of Maesaengi SPS-CF
Full size table

The PMP labelling/HPLC analysis showed that the SPS-CF consists of mainly l-Rha (45.0 mol %), d-Xyl (44.1 mol %), d-Man (10.2 mol %) and some minor sugars; the molar ratio between these major neutral monosaccharides was 1:1:0.2 (Table 1). The chromatogram of SPS-CF hydrolyzate (Fig. 1) showed three main peaks, but it was not clear whether peak 3 corresponds to d-Gal or d-Xyl because these sugars had similar retention times. To solve this problem, the assignment was confirmed by spiking the sample with the PMP-labelled, authentic d-Gal or d-Xyl and the result showed that this peak represents d-Xyl, so the SPS-CF should be a rhamnoxylan. This result was completely different from what we presented before in our previous paper (Na et al. 2010) where the SPS-CF was suggested to be a galactomannan. Thus, we tried to further confirm obtained result by using other chromatographic methods, i.e. HPAEC-PAD, GC-FID and TLC analyses.

Fig. 1
figure 1

HPLC chromatogram of (a) the mixture of PMP-labelled standard monosaccharides and (b) the PMP-labelled sugars of TFA hydrolyzate of Maesaengi SPS-CF polysaccharide, analysed on a C18-ODS column

Full size image

The HPAEC-PAD analysis for neutral monosaccharides composition of SPS-CF hydrolyzate (Fig. 2) showed two major (l-Rha, d-Xyl) and one minor (d-Man) peaks of neutral sugars. The HPAEC-PAD analysis for uronic acids showed the presence of only d-GlcA; d-GalA was not detected (Fig. 3). The content of these major sugars, l-Rha, d-Xyl, d-Man and d-GlcA, was shown to be approximately 4.0:3.9:1.0:1.0 with the molar ratios of 39.6, 38.6, 9.7 and 9.9 mol %, respectively, when setting the sum of each mole number of all the sugar species detected by HPAEC-PAD analysis as shown in Figs. 2 and 3 at 100 % (Table 1); d-Gal was not found at a significant level at all, confirming that this polysaccharide cannot be a galactomannan. These results were in agreement with those of PMP labelling/HPLC analysis described above, although there was some variation in molar ratios of each sugar between these two different methods (Figs. 1, 2, 3 and 4; Table 1). In addition, as shown in Fig. 4, the GC-FID analysis of SPS-CF also revealed l-Rha (33.0 mol %), d-Xyl (40.8 mol %) and, in a lesser extent, d-Man (4.2 mol %) as the major neutral monosaccharide components (1:1.2:0.1); other sugars, i.e. d-Gal (2.3 mol %), d-Glc (1.8 mol %), l-Fuc (1.1 mol %) and L-Ara (1.2 mol %), were found in smaller amounts (Table 1), when the values were calculated by setting the sum of each mole number of all the sugar species at 100 %. Possibly, these minor sugars originated from certain hexosan-type polysaccharide as a contaminant. Finally, the TLC analysis also showed only d-Xyl, l-Rha and d-Man as major spots on the chromatogram (Fig. 5), but there was no evidence for the presence of d-Gal as a major spot, confirming that the SPS-CF is a rhamnoxylan and not a galactomannan.

Fig. 2
figure 2

HPAEC-PAD chromatogram of (a) the mixture of standard monosaccharides (1 nmol of each) and (b) the TFA hydrolyzate of Maesaengi SPS-CF, analysed on a CarboPac PA1 column

Full size image
Fig. 3
figure 3

Uronic acid analysis of SPS-CF by HPAEC-PAD. (a) Authentic standard uronic acids: d-GalA and d-GlcA (10 nmol of each) and (b) the uronic acids in the TFA hydrolyzate of Maesaengi SPS-CF were analysed by HPAEC-PAD on a CarboPac PA 100 column

Full size image
Fig. 4
figure 4

GC-FID chromatogram of the TFA hydrolyzate of Maesaengi SPS-CF

Full size image
Fig. 5
figure 5

TLC chromatogram of TFA hydrolyzate of Maesaengi SPS-CF. Standard monosaccharides (25 μg each): Lanes 1, d-Gal; 2, l-Rha; 3, d-Man; 4, d-Xyl; 5, mixture of standard monosaccharides (lanes 1~4) dissolved in water; 6, mixture of standard monosaccharides treated with TFA in the same way used for preparation of TFA hydrolysis of Maesaengi SPS-CF; Lane 7, TFA hydrolyzate of Maesaengi SPS-CF (500 μg)

Full size image

Re-examination for chemical composition of SPS-CF polysaccharide revealed the presence of carbohydrates (~64 % m/m), uronic acids (13.2 % m/m), sulphate esters (15.4 % m/m) and proteins (0.5 % m/m) (Table 1). Previously, we reported that SPS-CF contains carbohydrates (49.8 % m/m), uronic acids (4.8 % m/m), sulphate (5.7 % m/m) and no proteins (Na et al. 2010). Current results showed somewhat higher contents of carbohydrates, uronic acids and sulphate than before, although protein content was shown to be almost the same, showing SPS-CF contains no or negligible level of protein.

Therefore, taken collectively, all these results of monosaccharide composition analyses performed by several methods mentioned above, clearly demonstrated that the most prominent neutral monosaccharides of SPS-CF are d-Xyl and l-Rha, and that d-Gal presents in negligible amount. These results strongly suggested that SPS-CF is a sulphated glucuronorhamnoxylan, which is an ulvan-type polysaccharide typical for green algae (Shanmugam and Mody 2000). It is completely different from what we presented before in our previous paper, where the SPS-CF was suggested as a glucuronogalactomannan based on the monosaccharide composition analysis performed by only PMP labelling/HPLC (Na et al. 2010).

As shown in Fig. 1 in the current study, even the same HPLC method with exactly the same protocols and the same HPLC systems resulted in completely different monosaccharide composition. It is now, however, obvious that from the results obtained by other additional methods, HPAEC-PAD, GC-FID and TLC, the SPS-CF is a glucuronorhamnoxylan and not a glucuronogalactomannan as it was mentioned in our previous paper (Na et al. 2010). Most probably, this discrepancy was a result of incomplete hydrolysis of SPS-CF owing to relatively high amount of dissolved polysaccharide in the reaction medium. Ulvans are known to be partially resistant to TFA hydrolysis (Quemener et al. 1997), so ballast galactomannan present in the SPS-CF may significantly contribute to the hydrolyzates. Underestimation of sugar contents in ulvan by TFA hydrolysis methods is possible due to the stability of glycosidic linkages in aldobiuronic acid fragments (BeMiller 1967). On the other hand, d-GlcA, d-Xyl and l-Rha, which are the main sugars of ulvans, are instable at stronger reaction conditions like 4 mol l−1 HCl in methanol (Chambers and Clamp 1971). Therefore, several independent analytical approaches should be used to be sure about true carbohydrate composition of complex polysaccharides like ulvans. In the case of SPS-CF, the presence of d-Xyl and l-Rha as major sugars, but no significant level of d-Gal or d-Man, was further confirmed by NMR analysis described below. Recently, Karnjanapratum et al. (2012) isolated sulphated polysaccharides (F1, F2, F3) from the same Korean C. fulvescens species mainly by thorough washing of raw materials, crushing into small particles, then extraction with 85 % ethanol. These polysaccharides showed significantly different chemical composition, i.e. contents of sugars, sulphate and protein, as well as structural features including linkage types and sulphation patterns and also molecular weight, from those of SPS-CF extracted with 0.1 mol l−1 HCl. For example, the proposed structure of F2 was a mannan-type polysaccharide. These compositional and thus structural differences may have arisen mainly from the distinct extraction and isolation procedures. As a result, we isolated polysaccharide (SPS-CF) totally different from those (F1, F2, F3) reported by Karnjanapratum et al. (2012); all of them are certain structural components of the same alga (C. fulvescens).

NMR Analysis

Proton NMR and 13C APT NMR spectra of the SPS-CF are shown in Fig. 6a, b, and resonance signal assignments are summarised in Table 2 according to the literature (Lahaye 1998; Lahaye and Ray 1996; Lahaye et al. 1998, 1999; Sugahara et al. 1996). Integral areas of the main proton signals in the anomeric region (δ 4.4–5.5) are shown in Fig. 7. Correlation 1H,1H-COSY and 1H,13C-HSQC spectra are shown in Figs. 8 and 9.

Fig. 6
figure 6

1H NMR (a) and 13C APT NMR (b) spectra of Maesaengi SPS-CF; integral areas of the proton signals in the region of δ 4.4–5.5 (c)

Full size image
Table 2 NMR signal assignments for the major and some minor sugar units of Maesaengi SPS-CF
Full size table
Fig. 7
figure 7

Correlation 1H,1H-COSY (a) and 1H,13C-HSQC (b) NMR spectra of Maesaengi SPS-CF with assignment of the main sugar units

Full size image
Fig. 8
figure 8

Correlation 1H,1H-ROESY NMR map of Maesaengi SPS-CF: the H-1 (a) and CH3 (b) regions; inter-unit NOE in the di- and trisaccharide fragments (c)

Full size image
Fig. 9
figure 9

NMR evidences of some minor sugar units in Maesaengi SPS-CF on the 1H NMR spectrum (a) and 1H,13C-HSQC (b), 1H,1H-TOCSY (c) and 1H,1H-ROESY (d) maps

Full size image

A strong proton doublet signal at δ ~1.21 (3 J H5, H6 = 4.6), corresponding carbon resonances at δ 16.9 and 17.0 ppm and HSQC cross peak at δ 16.9/1.21 were assigned to methyl groups at C-6 of Rhap (Fig. 6). Three COSY cross peaks at δ 3.96/1.21, 3.96/3.71 and 3.96/3.83 were assigned to H-5/H-6 and H-5/H-4 interactions in these units (Fig. 7a), three HSQC cross peak at δ 3.96/67.8, 3.71/77.3 and 3.83/77.0 indicate H-5/C-5 and H-4/C-4 interactions (Fig. 7b) and two HMBC cross peaks at δ 1.21/67.8 and 1.21/77.3 arose from H-6/C-5 and H-6/C-4 interactions (not shown). Carbon signals centred at δ 62.5 and 65.2 were assigned to CH2 at C-5 in Xylp. The HSQC cross peaks at δ 3.99/62.5 (3.85/65.2) and 3.24/62.5 (3.20/65.2) indicate the H-5 eq/C-5 and H-5ax/C-5 interactions in these units. Intense COSY cross peaks at δ 3.99/3.24 and 3.84/3.20 corresponded to strong geminal interaction between the methylene protons H-5 eq and H-5ax in Xylp. Therefore, all the NMR features mentioned above confirmed the presence of major sugars Rhap and Xylp in SPS-CF. In addition, two weak carbon signals at δ 175.9 and 175.4 were assigned to C-6 of uronic acids (Fig. 6b).

The anomeric proton (carbon) resonances at δ ~5.3–4.8 and ~4.7–4.3 (~93–97 and ~103–104) indicated the presence of both α- and β-anomeric forms. Two intense signals at δ 4.98 and 4.79 were assigned to H-1 of α-l-Rhap; next four resonances at δ 4.66, 4.58, 4.54 and 4.30 arose from H-1 of β-d-Xylp and β-d-GlcpA (Fig. 6a). The corresponding carbon signals were found at δ 96.8 and 97.2 (α-l-Rhap), 103.0 and 103.2 (β-d-GlcpA) and 103.2 and 103.7 (β-d-Xylp). The cross peaks H-1/C-1 HSQC found at δ 4.98/96.8, 4.79/97.2, 4.66/103.0, 4.58/103.2, 4.54/103.7 and 4.30/103.1 confirmed this assignment (Fig. 7b). Two HMBC cross peaks at δ 3.20/103.7 and 3.20/73.6 were assigned to H-2/C-1 and H-2/C-3 interactions in β-d-Xylp (not shown).

The correlation COSY, TOCSY, HSQC and HMBC spectra of SPS-CF confirmed the assignment of protons and carbons of the main five units (Figs. 7 and 8). Resonance signals of these units predominate in the spectra. Complete spin systems of two types of α-Rhap units (R, R′), one type of β-d-Xylp units (X) and two types of β-d-GlcAp units (G, G′) were found based on the analysis of COSY, TOCSY, HSQC and HMBC data. The results were in agreement with reports of Lahaye (1998) and Lahaye et al. (1997). Comparing the integral areas of the proton signals in the region of δ 4.4–5.5 (Fig. 6c), the ratio between the major units R/R′/X/G/G′ was deduced to be approximately 1:1:2.25:0.5:0.5. The R, R′ and G′ units are sulphated at O-3. This suggestion is confirmed by characteristic downfield shifts of H-3 and C-3 signals in comparison with those of the corresponding non-substituted units. Assignment of the rest less-pronounced C/H signals of other (minor) sugar units was complicated in many cases because of low intensity and overlapping of these signals. Nevertheless, some of these minor units were identified or their signals were partially described (Fig. 9; Table 2). Among them, only terminal β-d-Xylp (tX) exhibited pronounced and well-resolved signals of all H/C (Figs. 6, 7, 8 and 9).

The NOE inter-unit interactions observed for SPS-CF are illustrated in Fig. 8. Several ROESY inter-unit cross peaks were found confirming the proximity of α-l-Rhap3S (R and R′) and β-d-Xylp (X) units: δ 3.20/1.21 (X2/R6, R′6), 3.53/1.21 (X3, X4/R6, R′6), 4.79/3.19 (R1/X2), 4.79/3.52 (R1/X3, X4), 4.97/3.18 (R′1/X2) and 4.98/3.52 (R′1/X3, X4) (Fig. 8a, b). The H-1 resonances of X and G units are partially overlapped by the H-3 signals of R and R′ residues, respectively, which complicate significantly the analysis of ROESY signals in this region. Nevertheless, two ROESY inter-unit cross peaks of X were found at δ 4.54/3.71 (X1/R4) and 4.54/3.83 (X1/R′4). In addition, one inter-unit cross peak of tX was found at δ 4.30/3.72 (tX1/R4). Therefore, ROESY experiments confirmed the (1 → 4) linkages between β-d-Xylp and α-l-Rhap3S in the alternating sequence of the SPS-CF backbone (Fig. 8c). For the β-d-GlcpA (G) and β-d-GlcpA3S (G′) units, four ROESY inter-unit cross peaks were found at δ 4.58/4.27 (G1/R′2) and 4.66/4.28 (G′1/R′2) confirming that these units are bound as the side chains to O-2 of branching α-Rhap3S units (R′) of the polysaccharide backbone (Fig. 8a, c).

It is thus evident that α-l-Rhap3S (R, R′) and the β-d-Xylp (X) residues are alternating units of (1 → 4)-linked rhamnoxylan backbone. Its disaccharide building unit [4)-β-d-Xylp(1 → 4)-α-l-Rhap3S-(1→] is known as ulvanobiose U3s and, together with two ulvanobioronic acids A3s and B3s having β-d-GlcpA (A3s) or α-l-IdopA (B3s) instead of β-d-Xylp, represent a main building block of ulvan polysaccharides (Robic et al. 2009; Lahaye and Robic 2007). It has also been found that this rhamnoxylan sequence is typical for green algae (Ulvales, Chlorophyta) ulvans (Lahaye 1998). According to our results, in SPS-CF the terminal β-d-GlcpA (G) or β-d-GlcpA3S (G′) units are bound as side chains to O-2 of the α-l-Rhap3S (R′) ones. Branched ulvans containing β-d-GlcpA have been described previously, and this uronic acid can be attached as the side chain to β-d-GlcpA or α-l-Rhap3S in the backbone (Lahaye and Robic 2007; Lahaye and Ray 1996). Lahaye and Ray (1996) described linear and branched oligosaccharides obtained by partial acid hydrolysis of native and desulphated ulvan. Some of these ulvan fragments contained β-d-GlcpA bound to O-2 of α-l-Rha. Later, Lahaye and Robic (2007) noted about the branched type A ulvanobiouronic acid 3-sulphate substituted on O-2 of Rha3S by a single GlcA residue (A2g, 3s). Tabarsa et al. (2012) described ulvan-like sulphated polysaccharides extracted from Ulva pertusa. These polysaccharides had a backbone composed of (1 → 4)-linked α-l-Rhap, β-d-GlcpA and β-d-Xylp units; some α-l-Rhap were (1 → 2)-linked or branched at O-2.

Until recently sulphated β-d-GlcpA units were not assigned as typical for green algae ulvans. Most of the sulphate in ulvan from Ulva spp. was found on O-3 of α-l-Rhap, but the presence of 2-O-sulphated β-d-Xylp was also reported (Lahaye and Robic 2007; Lahaye et al. 1998, 1999; Ray and Lahaye 1995b). However, according to Domozych et al. (2012), sulphation of ulvans also appears on O-2 of the d-GlcpA side chains. Tabarsa et al. (2012) described ulvan-like U. pertusa polysaccharides sulphated mostly at O-3 position of GlcpA. Despite these findings, sulphated uronic acids are much more widespread in animal polysaccharides, especially glycosaminoglycans (GAGs). For example, 2-O- and/or 3-O-sulphated β-d-GlcpA were found in heparan sulphate of bivalve mollusk Nodipecten nodosus (Gomes et al. 2010), king crab cartilage chondroitin sulphate K (Sugahara et al. 1996) and squid cartilage chondroitin sulphate E (Kinoshita-Toyoda et al. 2004). The O-sulphated d-GlcpA units in heparin are formed by sulphation of glucuronic acid residues and not by “back epimerization” of sulphated l-iduronic acid units; O-sulphation of both types of uronic acid units appear simultaneously and before the 6-O-sulphation of glucosamine residues (Kusche and Lindahl 1990). Specific sulphotransferases catalyse in vivo sulphation at O-2 position of uronic acid residues in heparin that is blocked in some animal cell mutants (Bai and Esko 1996). Razi et al. (1995) found that human heparin-related GAGs containing d-GlcpA units sulphated at O-2 but not at O-3 position, whereas the analysis of chemically sulphated, bacterial capsular polysaccharide (generated by Escherichia coli K5) confirmed 3-sulphation of d-GlcpA units.

The NMR evidences for minor sugar units are illustrated in Fig. 9. The terminal β-d-Xylp (tX) are non-reducing end units of the rhamnoxylan chains. Reduced l-Rhap3S of both α- and β- configuration (Rα, Rβ) and reduced β-d-Xylp (Xβ) units were also found; R″ is probably internal l-Rhap3S substituted at O-2 (Fig. 9a, b). Chemical shifts of proton and carbon signals corresponding to these minor units were in agreement with those reported for ulvan oligosaccharides (Lahaye and Ray 1996; Lahaye et al. 1997; Lahaye 1998). An unusual 4,5-anhydro-β-GlcpA residue assigned as aG (Table 2; Fig. 9) was detected by COSY, TOCSY, ROESY, HSQC and HMBC experiments based on the literature data (Ragazzi et al. 1993; Dantas et al. 1994; Lahaye et al. 1998). The TOCSY map indicated four signals corresponding to four-proton spin system of aG (Fig. 9c), whereas the ROESY map had one more inter-unit cross peaks at δ 5.25/4.32 (Fig. 9d). The latter signal confirmed that aG, like G and G′ residues, is bound to another sugar unit (assigned “?” in Fig. 9a, d), which is probably non-identified minor α-l-Rhap3S of the backbone. The appearance of unsaturated derivative of β-d-GlcpA residues in SPS-CF is probably caused by the activity of specific algal lyases catalysing β-elimination of 4-O-linked glycosidic linkages. Similarly, microbial heparinases from Flavobacterium heparinum or other sources cleavage heparin or heparan sulphate molecule by elimination reaction and thus produce the mixture of oligosaccharides containing 4,5-unsaturated uronic acids (Yang et al. 1985). Unfortunately, algal enzymes of similar activity on ulvan are not studied yet. By contrast, bacterial ulvan lyases from Persicivirga ulvanivorans are described in recent articles (Collén et al. 2011). These endolytic enzymes cleave the glycosidic bond in ulvan between the sulphated rhamnose and a glucuronic or iduronic acid residues. Thus, β-GlcpA units can be linked not only in the side chains of SPS-CF but can be as well components of the backbone. Otherwise, 4,5-anhydro-β-GlcpA residues may originate from ballast alkali soluble (1 → 4)-β-d-glucuronan (Ray 2006), which has been co-extracted together with ulvans from the cell wall of Ulva sp. using hot water with dissolved calcium chelating agent like sodium oxalate (Ray and Lahaye 1995a). The presence of these glucuronans causes some difficulties in structural characterisation of ulvans because both polysaccharides contain β-d-GlcpA units (Ray and Lahaye 1995a, b).

It is still unclear about the structure of ballast arabinan (arabinohexosan) and hexosan because of insufficient NMR data. Two incomplete spin systems (A, A′) of minor sugar in furanose configuration were found that could be 1,2-linked α-l-Araf (Table 2; Fig. 9a, b); other smaller signals of unknown minor sugars were too weak to be assigned. Very low proportions of galactose, glucose, mannose and protein are generally found in ulvan preparations (Robic et al. 2009), and in the case of the SPS-CF, such impurities could be present as well. Arabinose containing polysaccharides (sulphated arabinan, sulphated arabinogalactan etc.) exhibiting anti-coagulant activity have been found in green algae (Qi et al. 2012; Shanmugam et al. 2001; Siddhanta et al. 1999; Uehara et al. 1992), and thus could be present in SPS-CF as impurities.

ESI and MALDI-MS Analyses

The positive electrospray mass spectrum of SPS-CF hydrolysate obtained by partial acid hydrolysis is represented in Fig. 10a. The mass spectrum showed the presence of ions between m/z 50 and 1000; the peaks above m/z 500 were of very low intensity. The major ion observed at m/z 159 was identified as sodium adduct of CF3COONa; most of the other pronounced ions were sodium adducts [M + Na]+ of mono- and oligosaccharides, which are the products of ulvan hydrolysis. Three ions at m/z 173, 187, 217 and 289 were identified as Xyl, Rha, GlcA and RhaSO3Na, respectively. By contrast, the Hex and HexSO3Na ions at m/z 203 and 305 were much less abundant. Ion at m/z 421 was identified as ulvanobiose U3S (Xyl-RhaSO3Na). The intense peak at m/z 319 was assigned mainly to Xyl-Rha, a product of U3S desulphation, with a contribution of GlcASO3Na. Several ions originated from di- and trisaccharides GlcA-Rha, (GlcA-Rha)(SO3Na)2, Rha-Xyl-Rha, (Rha-Xyl-Rha)SO3Na, Xyl-Rha-Xyl, Xyl-RhaSO3Na-Xyl etc. were also identified. Presence of these ions confirmed alternating sequence of Xyl and Rha in the rhamnoxylan backbone, possible sulphation of Rha and GlcA as well as the linkage between these two latter units.

Fig. 10
figure 10

a ESI-MS spectrum (LTQ Orbitrap Velos, Thermo Scientific) recorded in the reflectron-positive ion mode from Maesaengi SPS-CF (NaOH neutralised) hydrolysate; b, c MALDI-MS spectra (UltrafleXtreme, Bruker) recorded in the reflectron positive ion mode from native Maesaengi SPS-CF (b) and in the reflectron negative ion mode from fractionated SPS-CF (XC-cartridge) after partial hydrolysis (c); peaks with shown m/z values are detected as [M + Na]+ for the cationic mode (a, b) and as [M–H]− or [M–Na]− for the anionic mode (c); the structural assignment is based on monosaccharide composition found by GLC, HPAEC and knowledge from NMR study

Full size image

Although sulphated polysaccharides are generally difficult to analyse by MALDI-MS because of labile nature of the sulphate group in the positive ion mode (Fukuyama et al. 2002), it was possible to obtain a series of peaks with alternating saccharide units as suggested by the NMR. The peak with the highest observed m/z value was detected at 5890 ([M + Na]+), thus indicating polymer with molecular mass 5.8 kDa. The composition of this peak can be supported by the presence of others more pronounced accompanied peaks detected, for example, at m/z 5534 or 4163 (Fig. 10b). Like in the case of ESI-MS, the structural outputs of NMR can also be supported by MALDI-MS analysis of SPS-CF after partial acidic hydrolysis. In this case, the sample provided major peaks in the negative ion mode in the range of m/z 400–1000 (Fig. 10c) clearly indicating the presence of alternating sugar units mentioned above. The peak at m/z 499.1 corresponded to disulphated disaccharide (GlcA-Rha)(SO3H)2. The major peak detected at m/z 733.1 could be assigned to tetramer consisting of two repeating RhaSO3H-Xyl units and that one at m/z 925.3 to hexamer with composition 2 × (GlcA-Rha-Xyl). Although obtained NMR data were not supportive for the presence of hexosans probably because of concentration effect, nonetheless their incidence cannot be excluded from results obtained by mass spectrometry. Minor peaks detected at m/z 2087, 1925, 1763, 1601 and 1439 were differing by mass 162 units and indicating oligomer with composition Hex n SO3H. Because of low intensities, these parent ions under MS/MS conditions did not produce enough fragment ions to determine the exact structure for these peaks. However, a comparison of calculated masses with observed m/z values of peaks with consequent increase in mass of 162 Da was supportive for the presence of mentioned hexosan oligomers.

Anti-coagulant Activity

Anti-coagulant activity of SPS-CF was measured by aPTT, PT and TT assays. Heparin was used as a positive control, dextran and distilled water as negative controls instead of sample solution. It was found that SPS-CF significantly delayed the blood clotting time in a dose-dependent manner in both TT and aPTT assays (Fig. 11). TT assay is a simple screening technique for the fibrin polymerisation process, and it is a function of fibrinogen concentration and sensitive to heparin and direct thrombin inhibitors. It has been reported that sulphate rich polysaccharide showed relatively high anti-coagulant activity compared with those of poor sulphate contents when tested with TT and aPTT assays (Shanmugam and Mody 2000). As shown in Fig. 11a, although its effect was shown to be somewhat weaker than heparin, the SPS-CF at the concentrations tested (10–200 μg ml−1) significantly prolonged the clotting time in TT assay in a dose-dependent manner; TT of dextran and heparin were 14.6 ± 0.2 and 46.4 ± 2.2 s, respectively, whereas that of SPS-CF (200 μg ml−1) reached up to 31.7 ± 1.3 s (Fig. 11a). Prolongation of TT by the SPS-CF indicates the inhibitory effect of this polysaccharide on thrombin activity or fibrin polymerisation in common pathway of the coagulation cascade of human body (Matsubara et al. 2001). On the other hand, SPS-CF also dose-dependently prolonged the clotting time in APTT assay although its effect was still weaker than heparin; the aPTT values of dextran (200 μg ml−1) and heparin (0.3 μg ml−1) were 43.1 ± 1.2 and 76.3 ± 2.0 s, respectively, whereas that with 200 μg ml−1 of SPS-CF was 60.9 ± 1.7 s (Fig. 11b). The prolongation of aPTT suggested that the SPS-CF is able to inhibit the intrinsic pathway of the coagulation cascade of human body. However, no significant prolongation of PT was observed by SPS-CF at the concentration up to 200 μg ml−1; the PT value of SPS-CF (200 μg ml−1) was 17.1 ± 0.5 s, while those of dextran (200 μg ml−1), heparin (10 μg ml−1) and control (distilled water) were 15.1 ± 0.3, 37.7 ± 1.5 and 15.3 s, respectively, suggesting that this polysaccharide did not inhibit extrinsic pathway of blood coagulation (data not shown). Taken collectively, these results suggest that the SPS-CF inhibits intrinsic and common pathways, but not extrinsic pathway, of the blood coagulation cascade and the thrombin activity or conversion of fibrinogen to fibrin. Further detailed molecular mechanism and in vivo activities of anti-coagulant action of this polysaccharide remain to be resolved for its future application as therapeutic agents.

Fig. 11
figure 11

Anti-coagulant activity of Maesaengi SPS-CF: (a) thrombin time (TT) and (b) activated partial thromboplastin time (aPTT) assays. Control negative control (distilled water), dextran negative control (200 μg ml−1), heparin positive control (0.5 μg ml−1 for TT, 0.3 μg ml−1 for aPTT)

Full size image

Conclusions

The monosaccharide composition of SPS-CF was re-examined using different separation methods and approaches. Obtained results clearly demonstrated that, in contrast to our previous investigation (Na et al. 2010), xylose and rhamnose are the most prominent neutral monosaccharides of SPS-CF, while mannose and galactose are present at much lesser extent or even in negligible amounts. These results strongly suggested that the SPS-CF is glucuronorhamnoxylan, an ulvan type polysaccharide typical for green algae. Extensive monosaccharide composition, MS and NMR analyses led to following conclusions: (i) SPS-CF is composed of alternating sequences of ulvobiose U3s in the rhamnoxylan backbone with single d-glucuronic acid and 3-sulphato-d-glucuronic acid units linked to O-2 of l-rhamnose-3-sulphate as the side chains; (ii) other polysaccharides like hexosan, arabinan, arabinohexozan and/or glucuronan may be present in the SPS-CF preparations but only as impurity. Therefore, according to the correlation NMR and MS results mentioned above, the structure of algal glucuronorhamnoxylan (SPS-CF) is deduced as following:

The results from the present study demonstrated that the Maesaengi ulvan SPS-CF isolated from Korean seaweed C. fulvescens, in addition to previously reported immunomodulation property (Na et al. 2010), exhibited as well significant anti-coagulant activity by which the aPTT and TT were significantly prolonged. Although somewhat weaker activity was observed than that of heparin, Maesaengi ulvan may be considered as a potential anti-coagulant agent. Besides, additional O-3-sulphation of some glucuronic acid units increased the density of negative charge in this polysaccharide and thus supported as well its anti-coagulant activity similarly like in heparin and heparan sulphates.