Polysaccharides of Salsola passerina: Extraction, Structural Characterization and Antioxidant Activity

The above-ground part of the Salsola passerine was found to contain ~13% (w/w) of polysaccharides extractable with water and aqueous solutions of ammonium oxalate and sodium carbonate. The fractions extracted with aqueous sodium carbonate solutions had the highest yield. The polysaccharides of majority fractions are characterized by similar monosaccharide composition; namely, galacturonic acid and arabinose residues are the principal components of their carbohydrate chains. The present study focused on the determination of antioxidant activity of the extracted polysaccharide fractions and elucidation of the structure of polysaccharides using nuclear magnetic resonance (NMR) spectroscopy. Homogalacturonan (HG), consisting of 1,4-linked residues of α-D-galactopyranosyluronic acid (GalpA), rhamnogalacturonan-I (RG-I), which contains a diglycosyl repeating unit with a strictly alternating sequence of 1,4-linked D-GalpA and 1,2-linked L-rhamnopyranose (Rhap) residues in the backbone, and arabinan, were identified as the structural units of the obtained polysaccharides. HMBC spectra showed that arabinan consisted of alternating regions formed by 3,5-substituted and 1,5-linked arabinofuranose residues, but there was no alternation of these residues in the arabinan structure. Polysaccharide fractions scavenged the 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical at 0.2–1.8 mg/mL. The correlation analysis showed that the DPPH scavenging activity of polysaccharide fractions was associated with the content of phenolic compounds (PCs).


Introduction
Plants of the Amaranthaceae family are associated with noxious garden weeds and ruderal plants. Perennial or annual herbaceous flowering plants of various species of the Chenopodium genus, known as the goosefoots, grow almost everywhere in the world and are among the most common cosmopolitan weeds. However, this family also contains valuable, useful plants. The genus of halophyte plant Salsola L. is one of the largest in the family Amaranthaceae. Plants of this genus are characterized by rapid regeneration, the ability to grow large biomass, resistance to high environmental temperature, tolerance to soil salinity and to extended drought conditions. Therefore, the role of plants of this genus is great in saline, arid regions of various countries with developed distant pastures. Over 150 species of the genus Salsola L., including annual semi-dwarf and dwarf shrubs and woody trees, are distributed in arid and semi-arid regions of the Middle East, Asia, Europe and Africa [1].
Extracts and decoctions of plants of this genus are used in world folk medicine to treat bacterial and viral, cardiovascular, skin diseases, coughs and flu, and in cosmetics [2]. Previously, several biologically active compounds were isolated from different types of Salsola: flavonoids, phenolic acids, saponins, triterpenes, lignans, sterols, fatty acids,

Isolation and Characterization of Polysaccharides
Five extractants-cold water, hot water, water acidified to pH 2.0, 0.7% aqueous solutions of ammonium oxalate and 0.5% aqueous solutions of sodium carbonate-were successively used to extract polysaccharides from S. passerine (Figure 1). We performed the extraction with each extractant out until there were no sugars in the corresponding extract. As a result, eleven polysaccharide fractions were obtained. Fractions extracted with the sodium carbonate solution had the highest yield and those extracted with cold water-the lowest. Polysaccharides isolated with cold water containing a significant amount of Man, GalA and Ara residues were the principal components of polysaccharides extracted by other extractants.
HW1, AC, OK1, SO1 were fractionated using a DEAE-cellulose (OH − ) column. As a result, three polysaccharide fractions were obtained from each fraction by elution with 0.01, 0.1 and 0.2 M NaCl. The polysaccharides of the obtained fractions had a similar monosaccharide composition. The GalA and Ara residues were the principal components of their carbohydrate chains. The exception was the HW1-1 fraction whose polysaccharides, similar to the polysaccharides of fractions extracted by cold water, were characterized by a significant content of Man and Glc residues (Table 1). This indicates that the extraction with cold water was incomplete, and a small part of the polysaccharides was extracted with hot water in the next step.
All parent fractions included protein components. Fractions extracted with sodium carbonate included the largest amount of protein (up to 39%). The largest part of the proteins was not connected to polysaccharides because it was removed during separation on DEAE-cellulose. However, a small part of the co-eluted protein seemed to be connected to polysaccharides. HW1, AC, OK1, SO1 were fractionated using a DEAE-cellulose (OH -) column. As a result, three polysaccharide fractions were obtained from each fraction by elution with 0.01, 0.1 and 0.2 M NaCl. The polysaccharides of the obtained fractions had a similar monosaccharide composition. The GalA and Ara residues were the principal components of their carbohydrate chains. The exception was the HW1-1 fraction whose polysaccharides, similar to the polysaccharides of fractions extracted by cold water, were characterized by a significant content of Man and Glc residues (Table 1). This indicates that the extraction with cold water was incomplete, and a small part of the polysaccharides was extracted with hot water in the next step.
All parent fractions included protein components. Fractions extracted with sodium carbonate included the largest amount of protein (up to 39%). The largest part of the proteins was not connected to polysaccharides because it was removed during separation on DEAE-cellulose. However, a small part of the co-eluted protein seemed to be connected to polysaccharides.

NMR Spectroscopic Study
Information about the structure of the main polysaccharides from S. passerines was obtained by a combined analysis of the NMR spectra of SO1-1, SO1-2 and SO1-3. The NMR spectra of the three samples were similar (Figures 2-5). The 13 C NMR spectra of the samples ( Figure 2) were assigned using 1 H, 13 C heteronuclear single quantum coherence spectroscopy (HSQC) spectra. Analysis of the 1 H, 13 C HSQC spectra (Figures 3-5, Table 2) revealed substitutions in the monosaccharide residues based on the comparison of their 13 C chemical shifts with those of the parent pyranoses and furanoses [22] and considering the glycosylation effects in the 13 C NMR spectra of the carbohydrates [23,24], as well as data from our previous NMR studies of pectins [25]. The correlated spectroscopy (COSY), total correlation spectroscopy (TOCSY), rotating frame Overhauser effect spectroscopy (ROESY) and heteronuclear multiple bond correlation (HMBC) spectra revealed residues of α-D-galactopiranoside uronic acid (GA in Table 2), α-L-rhamnopyranose (R) and α-arabinofuranose (A) in all three samples. Conclusions regarding monosaccharide composition, ring size and anomeric configuration were drawn based on the comparison of visible coupling constants and chemical shifts of the sugar residues and corresponding pyranoses [26,27] and furanoses [28,29].      13 C HSQC spectrum of the SO1-1. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2.  13 C HSQC spectrum of the SO1-1. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2.   13 C HSQC spectrum of the SO1-2. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2.  13 C HSQC spectrum of the SO1-2. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2.  13 C HSQC spectrum of the SO1-2. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2.

Figure 5.
Parts of 1 H, 13 C HSQC spectrum of the SO1-3. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2.  13 C HSQC spectrum of the SO1-3. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals refer to the carbon atoms in the residues, as designated in Table 2. Table 2. Chemical shifts of the signals in the 1 H and 13 C NMR spectra of the SO1-1, SO1-2 and SO1-3 (323 K, D 2 O, TSP, δ H 0.0, δ C −1.6). The occurrence of 1,2-linked (label R) and 2,4-substituted (label R ) α-L-rhamnose residues in polysaccharides was confirmed by cross peaks at 1.25/17.9 ppm and 1.31/18.1 ppm in the 1 H, 13   The occurrence of 1,2-linked (label R) and 2,4-substituted (label R′) α-L-rhamnose residues in polysaccharides was confirmed by cross peaks at 1.25/17.9 ppm and 1.31/18.1 ppm in the 1 H, 13 C HSQC spectra (Figures 3-5) and HMBC spectra (Figures 6 and S2). The ROESY spectrum of SO1-3 ( Figure 7) included an inter-residue correlation peak of the anomeric proton of Rha residues and H-4 of GalA residues at δH/H 5.26/4.44 ppm, confirming the RG-I regions in polysaccharides.   Table 2.  13 C HMBC spectrum of the SO1-1. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals before slash refer to the protons, and those after slash refer to carbon atoms in the corresponding residues.

C NMR Chemical Shifts (δ C ) and 1 H (δ H , Italic), ppm
Three intense signals at δH 5.16, 5.12 and 5.09 ppm belonging to terminal nonreduction arabinose residues (label A T ), 3,5-substituted arabinose residues (label A S ) and 1,5linked arabinose residues (label A L ), respectively, were found in the anomeric region of 1 H NMR spectra of SO1-1 and SO1-2 ( Figure S1). In the anomeric region, the 1 H NMR Figure 7. Part of 1 H, 13 C HMBC spectrum of the SO1-1. The corresponding parts of the 1 H and 13 C NMR spectra are shown along the X and Y axes, respectively. Arabic numerals before slash refer to the protons, and those after slash refer to carbon atoms in the corresponding residues.
The resonance of C-6 at δ C 176.0 ppm indicated the predominance of non-methylesterified α-1,4-linked D-GalA residues in the structure of polysaccharides SO1-3 (Figure 2c), but the signal of low intensity at δ H/C 3.86/54.4 ppm confirmed that some GalA residues were methyl esterified.
The correlation peak at δ H/C 2.09/21.56 ppm in the HSQC spectrum of SO1-3 ( Figure 5) confirmed the O-acetylated residues in the structure of polysaccharides from S. passerina. No clear evidence was obtained for the attachment of the O-acetyl group to specific residues, since the intensity of their signals was low. Rha and GalA residues may be acetyl esterified [30]. The signals of O-acetyl groups are present only in the spectrum of sample SO1-3, which included polysaccharides with a high content of GalA, which may indirectly indicate the O-acetylation of GalA residues.
The resonance of C-6 at δ C 176.0 ppm indicated the predominance of non-methylesterified α-1,4-linked D-GalA residues in the structure of polysaccharides ( Figure 2).
The sequence of Ara residues in the repeating units was determined using the H/C correlations in the HMBC spectra and the H/H correlations in the ROESY spectra.
The average length of branches in the arabinan side chains, derived from the relative amounts of terminal and branched arabinose residues, was equal to one, confirming that the branches in the arabinan side chains comprised a single arabinose residue.
The ratio of A T , A S and A L was approximately 1:1:4 in the 1 H NMR spectrum of SO1-1 and indicated that the lengths of the linear regions were four times the lengths of the branched regions.
The ratio of Ara residues in the spectra of SO1-2 and SO1-3 was not determined because of the overlap of the signal of the anomeric proton A L with the signal of the anomeric proton of GA in the 1 H NMR spectrum ( Figure S1).
A possible structure of the repeating unit of the arabinan chain of polysaccharides from S. passerines is proposed below (Scheme 1), where the lengths of the structural regions are arbitrary. The average length of branches in the arabinan side chains, derived from the r amounts of terminal and branched arabinose residues, was equal to one, confirmin the branches in the arabinan side chains comprised a single arabinose residue.
The ratio of A T , A S and A L was approximately 1:1:4 in the 1 H NMR spectrum o 1 and indicated that the lengths of the linear regions were four times the lengths branched regions.
The ratio of Ara residues in the spectra of SO1-2 and SO1-3 was not determin cause of the overlap of the signal of the anomeric proton A L with the signal of the ano proton of GA in the 1 H NMR spectrum ( Figure S1).
A possible structure of the repeating unit of the arabinan chain of polysacch from S. passerines is proposed below (Scheme), where the lengths of the structural r are arbitrary.

Scheme.
A possible structure of the repeating unit of the arabinan from S. passerines.
In addition, the following low-intensity peaks were found in the anomeric reg the 1 H, 13 (Figures 3 and 4). In addition, the following low-intensity peaks were found in the anomeric region of the 1 H, 13 (Figures 3 and 4).
The occurrence of 1,2-linked (label R) and 2,4-substituted (label R') α-L-rhamnose residues in polysaccharides was confirmed by the cross peaks at 1.25/17.9 ppm and 1.31/18.1 ppm in the 1 H, 13 C HSQC spectra (Figures 3-5) and the HMBC spectra (Figures 7 and S2). The ROESY spectrum of SO1-3 ( Figure 6) included an inter-residue correlation peak of the anomeric proton of Rha residues, and H-4 of the GalA residues at δH/H 5.26/4.44 ppm confirmed the RG-I regions in polysaccharides.
Thus, three structural domains were identified in the polysaccharides isolated from S. passerine: arabinan, HG and RG-I. Considering the intensity of signals in the NMR spectrum, SO1-1 is dominated by the arabinan units, while SO1-3 is dominated by the galacturonan units. In the present study, no links between them were established. Nonetheless, it is possible that they represent domains of a complex pectin macromolecule.
Arabinans have been found in the cell wall of several plants and are believed to form RG-I side chains [31]. However, most of the evidence is based on co-extraction and/or coelution of RG-I and arabinans [32][33][34]. Only a few studies found that the L-Ara residues are covalently attached to rhamnose residues at the O-4 position of the RG-I backbone [35,36].
1,5-linked residues of α-L-arabinofuranose form both the backbone and the side chains of most of the arabinans studied [37]. Backbone residues are usually substituted at O-2 and/or O-3 and/or at both positions, with O-3 substitutions predominating [32]. However, several arabinans with a high percentage of substitution at the O-2 position have also been found [38,39]. Other structures of arabinans have also been described. For example, in arabinans, both the furanose and pyranose forms of arabinose were found [40]. Terminal β-arabinofuranose residues may glycosylate 1,5-linked α-arabinofuranose residues of the backbone at position O-5 [41]. Various degrees of branching have been found, including single, linear and branched oligomeric and polymeric chains, with different linkage types. The almost linear 1,5-arabinan associated with protein was isolated from red wine [42]. The arabinans in pectins often have single substituted side chains [32,43]. Arabinans from soybean [44], apple [45], the inner bark of Norway spruce [46] were found to have a highly branched structure. Arabinan-rich pectins, which constituted 50% of the total pectic polysaccharides, have been obtained from pea Pisum sativum L. [47].
The roles of arabinans in plant cell walls remain unclear. It was established that arabinans can be substituted by terminal phenolic esters, particularly feruloyl or coumaroyl esters. Ferulic acid groups may be ester linked to O-2 of the arabinose residues [48,49]. Feruloyl esters may determine guard cell wall flexibility by providing the cross-links between arabinans and other wall polymers; this testifies a unique role for arabinans in determining the physical and functional properties of guard cell walls [50].

DPPH Radical-Scavenging Activity
Polysaccharide fractions from S. passerine scavenged the DPPH radical at concentrations of 0.2-1.8 mg/mL. The half-maximal DPPH inhibitory concentration (IC 50 ) of them is given in Table 3. CW2, CW3, SO1 and SO2 demonstrated the highest activity, which exceeded 2.51-2.96 times that of commercial apple pectin (AP) activity. CW2, CW3, SO1 and SO2 scavenged 67, 69, 55 and 67% of DPPH radicals at a concentration of 1 mg/mL. Other fractions were less effective and scavenged only 31-48% of DPPH radicals at a concentration of 1 mg/mL.
The DPPH radical scavenging assay is widely used to evaluate the antioxidant property of plant polysaccharides. The activity of CW2, CW3, SO1 and SO2 seems to be comparable to that of polysaccharides from cantaloupe rinds [51], hawthorn wine pomace [52], fruit bodies of Tremella fuciformis [53] and apple pomace [54]. It should be noted that some other polysaccharides demonstrated the same level of DPPH scavenging activity at lower concentrations. These include pectins from Chaenomeles sinensis fruits [55], Epilobium angustifolium L. [56], Thymus quinquecostatus Celak. leaves [57], Gardenia jasminoides J. Ellis flowers [58] and Ziziphus jujuba cv. Muzao [59]. The DPPH radical scavenging activity of the fractions obtained by DEAE-cellulose elution was compared with the activity of the parent fraction SO1 (Figure 8). The polysaccharides SO1-1, SO1-2 and SO1-3 obtained were less effective (p < 0.05) than the parent fraction SO1, exhibiting IC 50 equal to 3.64 ± 0. 18  On the basis of the yield and content of PCs in SO1-1, SO1-2, SO1-3 and SO1, mo of the PCs providing antioxidant activity were removed by anion exchange chromato raphy, assuming that they were not bound to the polysaccharide chains. The sum conten of PCs in polysaccharides SO1-1, SO1-2, SO1-3 included about 18% from the content PCs in the parent fraction SO1. It was detected that three fractions obtained on DEAE cellulose provided only 24% of the DPPH radical scavenging activity of SO1, althoug they represented about 70% of the parent pectin (Table 1). This suggests that the antiox dant activity of SO1 was mainly provided by the associated PCs but not by polysacch rides. On the basis of the yield and content of PCs in SO1-1, SO1-2, SO1-3 and SO1, most of the PCs providing antioxidant activity were removed by anion exchange chromatography, assuming that they were not bound to the polysaccharide chains. The sum contents of PCs in polysaccharides SO1-1, SO1-2, SO1-3 included about 18% from the content of PCs in the parent fraction SO1. It was detected that three fractions obtained on DEAEcellulose provided only 24% of the DPPH radical scavenging activity of SO1, although they represented about 70% of the parent pectin (Table 1). This suggests that the antioxidant activity of SO1 was mainly provided by the associated PCs but not by polysaccharides.
The relationship between the chemical characteristics of polysaccharides and DPPH scavenging ability was further investigated using correlation analysis. The total content of sugars and the (Ara + Gal)/Xyl ratio correlated negatively, whereas the content of PCs, Gal and Man, as well as PDI correlated positively with DPPH scavenging activity (Table 4). We tested the five regression models, subsequently removing the less significant factors (according to the p-value). The linear regression, including the contents of PCs and Man as independent variables, resulted in the best model for prediction (adj. R 2 = 0.82, p = 0.000) ( Table 5). The content of PCs was the only factor contributing significantly to DPPH scavenging activity (p = 0.000, β = 0.79). Thus, the correlation analysis showed that the DPPH scavenging activity of the sample from S. passerine is associated with the content of PCs. This is consistent with the results of Ref [54], whose authors evaluated the activity of apple pectins, and our previous study on fireweed pectins [56]. It is known that PCs may bind covalently to the side chains of RG I through the Ara and Gal residues and may be involved in the cross-linking of macromolecules [60].
It was shown that the removal of PCs from polysaccharides reduces the antioxidant activity but does not completely abolish it [60]. Several authors suggest that the antioxidant activity of pectins may be due to the hydroxyl and carboxyl groups of GalA residues [52,61]. Previously, we showed that the antioxidant activity of fireweed pectins is partly related to the xylogalacturonan chains [56]. However, in the present study, we failed to find the polysaccharide chains responsible for the DPPH radical scavenging activity of Salsola pectins. The small sample size (n = 14), which determines the statistical power of multiple regression [62], may be the reason we failed to identify the polysaccharide chains that contribute to the antioxidant activity of Salsola pectins.

Materials
Biological material: plant material, consisting of yellow-green annual branches with spherical dwarf leaves, was collected in August 2019 from the semi-shrubs of Salsola gemmascens ssp. passerina

Isolation of Polysaccharides of S. passerina
Polysaccharides from the plant material were sequentially extracted, as described below; the extraction scheme is shown in Figure 1. At each stage, an exhaustive extraction of polysaccharides was carried out until the absence of reaction of the extract to the carbohydrate; the extraction mixtures were mixed in a mechanical stirrer.
Freshly picked plant material (234 g) was milled in a blender, distilled water (1 L) was added, and the resulting mixture was stirred in a mechanical mixer at 20 • C for 3 h. The mixture was centrifuged, and the residue of the plant material was treated again; the treatment was repeated three times. In the next stage, polysaccharides from the residues of plant materials were extracted with hot water at 80 • C for 3 h. The extraction was repeated twice (each time, the volume of added water was 1 L). Finally, the five aqueous extracts (three obtained with cold water (CW1, CW2, CW3) and two with hot water (HW1, HW2)) were obtained. Next, polysaccharides were extracted with acidified water (pH 2.0, 1 L) at 50 • C for 3 h. As a result, one extract (AC) was obtained. Next, polysaccharides were extracted with aqueous solutions of ammonium oxalate (0.7% w/v) at 70 • C for 6 h. The extraction was repeated three times (the first volume of salt solution added was 2 L; the second and third volumes were 1 L). Finally, the three extracts (OK1, OK2, OK3) were obtained. Next, polysaccharides were extracted with aqueous solutions of Na 2 CO 3 (0.5% w/v) containing NaBH 4 at 70 • C for 3 h. The extraction was repeated twice (the first volume of soda solution added was 3 L, the second-2 L). The two extracts (SO1, SO2) were obtained.
The carbohydrate content of each extract was detected using a phenol-sulfuric acid assay [63]. All extracts were dialyzed against distilled water for 48 h at 10 • C. Extracts SO1, SO2 were previously acidified with a diluted solution of acetic acid to pH 5.6. The dialyzed extracts were concentrated on a Heidolph 4002 rotary evaporator (Germany) under reduced pressure at 40 • C.
Polysaccharides were precipitated from the extracts with a four-fold volume of 95% ethanol, centrifuged, washed twice with 95% ethanol, dissolved in distilled water, frozen and lyophilized. The yields of the polysaccharide fractions obtained are expressed in % (w/w) of mass of dry plant material and are presented in Table 1.

Ion Exchange Chromatography of Polysaccharide Fractions
The major polysaccharide fractions HW1, AC, OK1, SO1 were separated on a DEAEcellulose (OH-) column (2.5 cm × 40 cm). Each polysaccharide fraction (100 mg) was dissolved in 5 mL of 0.01 M NaCl, and the solution was applied to the column. The column was stepwise eluted with 0.01, 0.1, 0.2, 0.3, 0.5 and 1.0 M NaCl solution (400 mL of each eluent) at a flow rate of 0.9 mL/min. The fractions were collected at 12 min intervals using a low-pressure system Pharmacia Biotech (Sweden) with a FRAC-100 fraction collector, P-50 pump. The carbohydrate content in each tube was determined by the phenol-sulfuric acid method [63]. When separating each of the HW1, OK1, SO1, three major polysaccharide fractions were obtained (eluted with 0.01, 0.1 and 0.2 M NaCl). When separating AC, the fraction eluted 0.2 M NaCl was obtained as minor. In addition, minor fractions were obtained from all fractions by elution with 0.3, 0.5 and 1.0 M NaCl.
The separation procedure was repeated twice for HW1, OK1 and four times for AC, SO1. Data on the monosaccharide composition and the yield of the fractions are presented in Table 1 as a mean of these experiments.

General Analytical Methods
The content of uronic acids was determined as described earlier [64,65]. The quantitative determination of protein was calculated using the Bradford method [66]. The quantitative determination of phenolics was performed with the Folin-Ciocalteu reagent using gallic acid as a standard [67]. The content of neutral monosaccharides was determined by gas-liquid chromatography (GLC), as described earlier in detail [68]. The sugar concentration was determined at 490 nm using the phenol-sulfuric acid assay [63].
The relative molar mass distributions (RMM) (including Mn, Mw and PDI) of the polysaccharide samples were determined by size exclusion chromatography with high-performance liquid chromatography (HPSEC); the procedure was described in detail earlier [69].

Nuclear Magnetic Resonance Spectroscopy
All homo-and heteronuclear NMR experiments of the samples were carried out on a Bruker Avance 600 spectrometer (Germany) at a probe temperature of 303, 313 and 318 K, which provided a minimum overlap of the signal of deuterated water with the polymer signals. The procedures for preparing the polysaccharide samples and the conditions of the NMR experiments were described earlier [69].

Antioxidant Activity
The DPPH solution (0.2 mM, in ethanol) was added to the pectin solution (0.4-3.6 mg/mL water) in equal proportions (v/v) and mixed. After incubation at 25 • C for 1 h, the absorbance of the sample was measured at 517 nm. The scavenging activity of the pectins was measured at four different concentrations, and the half-maximal inhibitory concentration (IC 50 , mg/mL) values were calculated based on a polynomial regression curve [70].

Statistical Analysis
The significance of the difference among the means in determining the antioxidant activity was estimated with one-way analysis of variance (ANOVA) and Fisher's least significant difference (LSD) post hoc test at p < 0.05. The relationship between the chemical characteristics and activity of polysaccharide fractions was evaluated by the calculation of the Pearson correlation coefficients and multiple linear regression analysis. All calculations were performed using the statistical package Statistica 10.0 (StatSoft, Inc., USA). The data were expressed as the means ± s.d. of three independent experiments.

Conclusions
Polysaccharide fractions isolated from S. passerine with water and aqueous solutions of ammonium oxalate and sodium carbonate were characterized by a similar composition, including polysaccharides, protein and PCs. HG, RG-I and arabinan with regions formed by 3,5-substituted and by 1,5-linked arabinose residues were identified as the principal units of the polysaccharides obtained. Polysaccharide fractions of S. passerine demonstrated a moderate antioxidant potential. Fractions isolated with cold water and sodium carbonate scavenged the DPPH radical in vitro to a much greater extent than commercial apple pectin. The correlation analysis of the composition and activity of polysaccharide fractions obtained by anionic exchange chromatography revealed that the antioxidant capacity of polysaccharides of S. passerine is mainly due to the associated PCs.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.