Isolation, Fractionation, and Structural ... - Cereal Chemistry

Isolation, Fractionation, and Structural Characteristics of Alkali-Extractable β-Glucan from Rye Whole Meal S. M. Ragaee,1,2 P. J. Wood,1 Q. Wang,1 S. M. Tosh,1 Y. Brummer,1 and X. Huang1 ABSTRACT

Cereal Chem. 85(3):289–294

The main nonstarch polysaccharide of rye is arabinoxylan (AX), but rye contains significant levels of (1→3)(1→4)-β-D-glucan, which unlike oat and barley β-glucan, is not readily extracted by water, possibly because of entrapment within a matrix of AX cross-linked by phenolics. This study continues objectives to improve understanding of factors controlling the physicochemical behavior of the cereal β-glucans. Rye βglucan was extracted by 1.0N NaOH and increasing concentrations of ammonium sulfate were used to separate the β-glucan from AX and prepare a series of eight narrow molecular weight (MW) distribution frac-

tions. Composition and structural characteristics of the isolated β-glucan and the eight fractions were determined. High-performance size-exclusion chromatography (HPSEC) with both specific calcofluor binding and a triple detection (light scattering, viscometry, and refractive index) system was used for MW determination. Lichenase digestion followed by high-performance anion exchange chromatography of released oligosaccharides, was used for structural evaluation. The overall structure of all fractions was similar to that of barley β-glucan.

Rye is a hardy crop, its production is environmentally friendly and, compared with wheat, it can be grown advantageously by organic methods because of reduced requirements for fertilizer and pesticides. Yields of rye grain, particularly those of the newer hybrid cultivars, are comparable to wheat yields on good soils and superior on poor soils where wheat performs badly (Seibel and Weipert 2001). Consumption of whole grain rye, a good source of dietary fiber, has potential health benefits such as reduced risk of heart disease and improved glucose metabolism (Salmeron et al 1997; Liu et al 1999). Clinical studies indicate potentially favorable modification of lipid and carbohydrate metabolism (Leinonen et al 1999, 2000). The major nonstarch polysaccharide of rye is arabinoxylan (AX 7–12%) (Saastamoinen et al 1998; Saini and Henry 1998) but (1→3)(1→4)-β-D-glucan is also present. Interest in cereal βglucan has arisen because of its physiological effects and potential health benefits. Oat β-glucan has been linked to lower serum cholesterol levels (Braaten et al 1994) and attenuated postprandial blood glucose and insulin (Wood et al 1994a). The latter effect is viscosity related. Lower insulin levels are associated with lower serum cholesterol levels (Jenkins et al 1989) and, in general, increased luminal viscosity is believed to be of importance in lowering serum cholesterol levels (Jenkins et al 2001). Viscosity depends, among other things, on molecular weight (MW), structure, and concentration of β-glucan in the intestinal fluid. In the United States, both oat and barley have been allowed a health claim based on the β-glucan content, which states that products containing these cereals might lower risk of heart disease (FDA 1997, 2005). The amounts of β-glucan found in rye are lower than in oats and barley, where values are reported generally at ≈3–8% (Peterson 1991; Lee et al 1997; Colleoni-Sirghie et al 2003; Yao et al 2007). The amounts in rye (2.1–3.1%) are at or below the lower end of that range (Henry 1987; Ragaee et al 2001; Ragaee et al, unpublished) but levels are enriched in bran (Nilsson et al 1996). Rye β-glucan is mostly water-insoluble, which might render it biologically unavailable for functions dependent on viscosity. To render rye β-glucan water-soluble and to facilitate separation from AX, previous studies have used barium hydroxide to extract AX before hot water extraction of the β-glucan (Roubreks et al 2000), but barium hydroxide also results in considerably

depolymerized β-glucan (Brummer et al 2008), and water soluble yields were low. Ragaee et al (unpublished) observed that good yields (62–74% of total) of high molecular weight (chromatographic peak MW ≈1,200K) rye β-glucan were obtained in 1.0N NaOH. The major structural feature of cereal β-glucan is β-(1→3)linked cellotriosyl and cellotetraosyl units (Parrish et al 1960). The ratio between these major building blocks distinguishes the β-glucans from various sources. Wood et al (1991a, 1994b) demonstrated that there is a higher ratio of tri- to tetrasaccharide in βglucan from whole barley (2.8–3.3) and rye (3.0–3.2) compared with that of oat (2.1–2.4). Li et al (2006) reported a tri- to tetrasaccharide ratio of 4.4 for β-glucan from wheat bran. The use of ammonium sulfate for purification of β-glucan, first developed by Preece and Mackenzie (1952) has been applied by several authors (Wood et al 1991a; Izydorczyk et al 1998; Wang et al 2003; Li et al 2006). In our laboratory, Wang et al (2003) demonstrated that molecular size was the dominant mechanism of fractionation by ammonium sulfate and careful application of this method gave structurally homogeneous fractions. In the present study, we used the gradient ammonium sulfate fractionation procedure of Wang et al (2003) to separate β-glucan from AX, as well as to determine whether the rye β-glucan, unlike oat, barley and wheat, might exhibit structural heterogeneity.

1 Agriculture & Agri-Food Canada, Food 2 Corresponding

Research Program, Guelph, ON N1G 5C9. author. Phone: (519) 829-2649. E-mail: [email protected]

doi:10.1094 / CCHEM-85-3-0289 © 2008 Department of Agriculture and Agri-Food, Government of Canada.

MATERIALS AND METHODS Rye (Secale cereale) cultivar AC Rifle, grown in western Canada (Lethbridge) in 2004 was kindly provided by J. G McLeod (Agriculture and Agri-Food Canada, Swift Current, Saskatchewan). Thermostable-α-amylase (Bacillus lichenformis, E.C. 3.2.1.1, Lot 50901), β-xylanase M6 (from rumen microorganism, E.C. 3.2.1.8, Lot 51206), and (1→3)(1→4)-β-D-glucan-4-glucanohydrolase (lichenase; Bacillus subtilis, E.C. 3.2.1.73, Lot 60502) were purchased from Megazyme International Ireland Ltd (Bray, Ireland). All chemicals were of reagent grade unless otherwise specified. Rye kernels were ground to pass through a 0.3-mm screen to obtain rye whole meal. Endogenous enzymes were deactivated by suspending the rye whole meal in aqueous ethanol (70% v/v) and stirring under reflux for 2 hr at a ratio of (1:5 w/v), followed by refluxing for 1 hr with 80% aqueous ethanol (v/v). The supernatant was removed by centrifugation at 5,000 × g for 15 min, and the residue was washed twice with two volumes of ethanol (95%) and dried at 45°C overnight. The enzyme-deactivated dry rye sample was reground to pass through a 0.3-mm screen using a ball mill. Vol. 85, No. 3, 2008

289

Analytical Methods β-glucan content in the rye whole meal, alkali-extractable βglucan, and the fractions isolated by ammonium sulfate precipitation were determined by the method of McCleary and GlennieHolmes (1985) using a β-glucan assay kit (Megazyme International), in an automated glucose analysis (Wood et al 1991a), and in solubilized fractions by flow-injection analysis (FIA) (Beer et al 1997). Monosaccharide composition of the isolated rye β-glucan was determined after hydrolysis with 1.0M H2SO4 at 100°C for 3 hr. The hydrolyzate was diluted as appropriate, filtered (0.45 μm) and analyzed by high-performance anion-exchange chromatography (HPAEC) (Dionex, Sunnyvale, CA) using a Carbopac PA1 column (4 × 250 mm) and guard column (3 × 25 mm) with detection by pulsed amperometry with a gold electrode as described by Wood et al (1994b). Protein was determined based on the combustion method using a FP 2000 Leco nitrogen analyzer . The sample (20–30 mg) was combusted at 1150°C in a sealed furnace. Measurements were validated by analyzing four standard compounds: atropina (4.84% N), DL-metionina (9.39% N), acetanilide (10.36% N), and nicotinamide (22.94% N), and by running blank and standard samples before actual sample analysis. Protein content was expressed as nitrogen × 5.83. Structural Characterization Oligosaccharides released by lichenase (Megazyme β-glucan analysis kit) were analyzed using HPAEC (Dionex) according to Tosh et al (2004b). Relative amounts of oligosaccharides were calculated without correction for different response factors as relative normalized area with the total area of oligosaccharides of DP 3–9 set as 100%. Molecular Weight (MW) Determination Peak molecular weight (Mpc) of crude alkali extracted β-glucan and the eight fractions were determined by HPSEC using a Shodex (Showa Denko K.K., Tokyo, Japan) OHpak Kb-806M column (with OHpak guard) and a Waters ultrahydrogel linear column, with postcolumn addition of calcofluor enabling fluorescence detection (Wood et al 1991a). The system consisted of a Perkin Elmer ISS-100 auto sampler and injector, a Shimadzu model ADvt HPLC pump, a Shimadzu RF-10Axl fluorescence detector, a Waters (Milford, CT) model 510 HPLC pump for postcolumn addition of calcofluor, and the Viscotek DM 400 data manager. Data integration was performed using TriSEC 3.0 (Viscotek, Houston, TX) software. Five β-glucan molecular weight standards, both prepared in-house (Wang et al 2003) and obtained commercially

Rye Whole Meal ↓ Ethanol Reflux ↓ Amylase Treatment ↓ Water Wash, Dry, Grind ↓ Extract with 1N NaOH ↓ Neutralize and Extract with Phenol ↓ Collect Top Layer, Precipitate with Ethanol (80%) ↓ Wash with IPA (100%), Dry, Grind ↓ F0 Fig. 1. Scheme for preparing alkali-extractable rye β-glucan. 290

CEREAL CHEMISTRY

(Megazyme), were used to construct a calibration curve for βglucan by plotting retention time versus log Mpc. Calcofluor binds selectively to β-glucan, which allows analysis of unpurified, crude extracts, but binding declines below MW≈10–20K and so rather than report weight (Mw) or number (Mn) average MW, with this system, we use the value at the peak of the chromatogram, which for normal distributions of purified material corresponds to the refractive index peak as used in the Viscotek system described below because both respond to concentration. The HPSEC with triple detection also used a Shodex OHpak KB-806M and Ultrahydrogel linear column with a Shimadzu SCL-10Avp pump and auto injector (Shimadzu Scientific Instruments, Columbia, MD). Elution was with 0.1M NaNO3 containing 0.03% (w/w) NaN3 at a flow rate of 0.6 mL/min. Peaks were detected with the Viscotek triple detector system consisting of a refractive index (RI), differential pressure viscometer (DP), and right-angle laser light scattering (RALLS) detector. The Viscotek DM 400 data manager and TriSEC 3.0 software were used to determine the molecular weight distribution of the five β-glucan molecular weight standards and the obtained β-glucan fractions. Pullulan (J.M. Science, New York, NY) was used to calibrate the detectors. A refractive index increment (dn/dc) of 0.146 mL/g was used (Beer et al 1997; Wang et al 2003). Isolation of Rye Alkali-Extractable β-Glucan The isolation procedure is outlined in Fig. 1. Water (2,000 mL) was added to the deactivated rye whole meal (100 g) and the mixture was stirred and heated to 80°C. Thermostable α-amylase (22,500 unit; preheated to 80°C) was added and heating continued for 2 hr at 80°C. After cooling to room temperature, the mixture was centrifuged at 5,000 × g for 15 min, the residue was washed twice for 30 min with 500 mL of water and centrifuged as before, then dried overnight at 45°C and ground as described for the deactivated meal. The residue was then extracted for 90 min with 1.0N NaOH containing 0.1% KBH4, using a magnetic stirrer (90 min, 25°C). The mixture was neutralized with 2.0N HCl and centrifuged for 20 min at 5,000 × g and 25°C. The extract was treated with phenol (Westphal et al 1952) at a ratio of 1:1 (v/v) for 35 min at room temperature to remove protein. The mixture was allowed to cool in an ice bath and centrifuged (10,000 × g, 20 min, 10°C). The top layer, which contains the polysaccharides, was recovered and 20 mL of water was added to the phenol-protein mixture, shaken for a few minutes, then centrifuged and the top layer again was recovered and added to the previous collection. Ethanol was slowly added to the extract to bring it to 80% (v/v) ethanol, the mixture was centrifuged, and the precipitate was suspended in 100% 2-propanol (iso-propanol, IPA) and kept at 4°C overnight. After removing the IPA, the crude alkali extract was dried in a vacuum oven (80°C) for 4 hr and ground using a ball mill to give fraction F0. Fractionation of Alkali Extracted Rye β-Glucan A portion of F0 (1,300 mg) was dispersed in 500 mL of deionized water and heated at 90°C for 2 hr under constant stirring, cooled to 25°C, and then centrifuged at 5,000 × g for 10 min. The fractionation procedure (Fig. 2) was according to Wang et al (2003). Ammonium sulfate (NH4)2SO4 (60 g) was added slowly to the centrifuged solution under constant stirring in a water bath at 30°C to bring the (NH4)2SO4 concentration to 16.4% (w/w). The precipitate was collected by centrifuging (8,000 × g, 10 min, 25°C) and redissolved in distilled water by heating at 80°C for 1 hr and 60°C for 2 hr with constant stirring, then it was dialyzed against deionized water (48 hr, 4°C) using membrane tubing (MW cutoff of 6,000–8,000; Spectra Medical Industries, Los Angeles, CA) to eliminate LMW components. β-Glucan was precipitated by adding an equal volume of 100% IPA and the mixture was centrifuged (10,000 × g, 20 min, 25°C). The pellet was dispersed in 100% IPA, left overnight at 4°C, then recovered by filtration on a

glass filter and dried in vacuum oven at 80°C for 4 hr. This fraction was designated as F1. In the supernatant from the initial ammonium sulfate precipitation, the concentration of (NH4)2SO4 was increased stepwise to concentrations as indicated in Fig. 2 to a final concentration of 33.5%. The corresponding fractions at each step, obtained by centrifugation as described above, were designated F2, F3, F4, F5, F6, F7, and F8 The composition, MW, and structural characteristics of F0 and the eight ammonium sulfate precipitated fractions were determined as described above. RESULTS Alkali Extraction Treatment of 100 g of rye whole meal with refluxing aqueous ethanol and hot aqueous amylase followed by removal of the solubilized material yielded 18.9 g of dry matter, which is mainly nonstarch polysaccharide and protein. A portion (5.0 g) of the insoluble residue was then extracted with 1.0N NaOH containing 0.1% KBH4.. To lower the protein content, the described process of neutralization and phenol treatment, followed by precipitation with IPA, was used and reduced the protein content to 1.6% in the extract (F0), which was used for further purification and fractionation. F0 contained similar amounts of β-glucan and AX (32.0 and 29.3%, respectively). The Mpc of β-glucan was 372K as determined by HPSEC with calcofluor detection. Other studies (Ragaee et al, unpublished) found the Mpc of β-glucan extracted by 1.0N NaOH to be of the order of ≈1,200K, which was similar to the value in the original amylase-treated meal. The MW was not affected by phenol treatment. Fractionation and Molecular Weight Distribution of Alkali-Extracted Rye β-Glucan The alkali-extracted β-glucan isolate (F0) was fractionated into eight fractions (F1–F8) at ammonium sulfate concentrations of 16.4, 16.8, 17.5, 18.5, 19.5, 23.2, 28.5, and 33.5% (w/w), respectively. Table I summarizes the yield of each fraction as a percent-

age of the alkali-extracted isolate, F0, and provides the β-glucan content, AX content, A/X ratio, and molecular characteristics of each fraction. Most of the protein in F0 was found in the first fraction, F1. Phenol was used to remove the protein, increasing the β-glucan content in the fraction (designated F1p) from 31.5 to 75%. The concentration of β-glucan in the eight fractions ranged from 75.0% for F1p to 88.0% for F5, while for the later fractions (F6, F7, and F8), β-glucan content declined to 79.0, 23.5, and 9.4%, respectively. The amount of AX in each fraction determined by HPAEC and the A/X ratio (Table I) indicated that the lowest concentration of AX (3–3.8 %) was found in F2–F5, and the highest concentration (18.3%) was in F8. Values for the A/X ratio ranged between 0.25 for F1p to 0.89 in F8. Phenol treatment of fraction F1 increased AX content from 6.5 to 12.6%, proportionately slightly less than the increase in β-glucan (31.5–75%), and reduced the A/X ratio from 0.45 to 0.25. The phenol treatment did not greatly modify the MW distribution of F1 (Table I) determined using HPSEC with the triple detector system. Mw of F1 was 770K; Mw of F1p, after the phenol treatment, was 716K. Molecular Characteristics of Fractions from Alkali-Extracted β-Glucan Precipitated by Ammonium Sulfate HPSEC with calcofluor detection (Fig. 3) showed a broad MW distribution with Mpc of 372K for β-glucan in F0, whereas the βglucan in fractions obtained by gradient ammonium sulfate concentrations were clearly of narrower MW distributions. As the ammonium sulfate concentration was increased, the fractions F1 and F2 had Mpc higher than F0 (785 and 524K, respectively) and that of F3, 367K, was similar. The subsequent fractions had lower Mpc than the original (264, 185, 112, 64, and 47K) for F4, F5, F6, F7, and F8, respectively. Molecular characteristics (Mw, Mp, [η] and polydispersity index [Pd = Mw/Mn]) of fractions F1–F8 as determined by HPSEC with the triple detector system (Table I) confirmed narrow MW distributions (Pd ≈1.1) for F1–F6, but F7 and F8 had a very broad MW distribution. (Pd 6.0 and 5.0, respectively). Mp values were similar to those found using calcofluor detection and followed the same decreasing trend with increasing ammonium sulfate concentration, up to F6, but then increased in fractions F7 and F8 as polydispersity increased. The increase was particularly apparent in Mw, reflecting the very broad MW distribution. Structure of F0 and Eight Fractions Precipitated by Ammonium Sulfate Essentially no differences were observed in the profiles of lichenase released oligosaccharides from F0 and fractions F1–F8 (Table II), although there was a slightly greater relative amount of

Fig. 2. Scheme for fractionation of alkali-extractable rye β-glucan through stepwise precipitation by ammonium sulfate. Numbers given at each step are the (NH4)2SO4 concentrations used for fractionation.

Fig. 3. High-performance size-exclusion chromatography using calcofluor detection of fraction F0 (heavy line) of alkali-extractable rye βglucan and fractions obtained by stepwise precipitation with ammonium sulfate (F1–8, consecutively, at increasing retention times). Vol. 85, No. 3, 2008

291

DP 5–9 in F0 and F7 and F8. The molar ratio of 3-O-β-cellobiosyl-D-glucose to 3-O-β-cellotriosyl-D-glucose ranged from 3.44 to 3.50. The ratio for oat β-glucan was 2.4–2.5 and for barley βglucan it was 3.3–3.6 using the same HPAEC system. The elution system used can modify relative response factors as discussed by Brummer et al (2008). DISCUSSION The β-glucan of oats and barley are relatively easily extracted in hot water or dilute alkali (Wood et al 1989, 1991b; Bhatty 1993; Edney et al 1991; Beer et al 1997; Johansson et al 2000). Purification to ≈80% β-glucan by precipitation with ethanol or IPA is straightforward (Wood et al 1989; Zhang et al 1998) and >90% purity is achieved by ammonium sulfate precipitation (Preece and Mackenze 1952). Wheat and rye β-glucan are less readily extracted and purified (Nilsson et al 1996; Cui et al 1999; Brummer et al 2008). Because we found that the usual method for extraction of rye β-glucan, which uses a prior treatment with barium hydroxide to preferentially extract AX (Nilsson et al 1996; Roubreks et al 2000) results in depolymerization of the β-glucan (Brummer et al 2008), we explored alternate methods for extraction and purification and on the small scale (≈1g) found that 1.0N NaOH provided a good yield (62%) of product of high MW (≈1,200K) (Ragaee et al, unpublished), but to reduce content of AX and protein, xylanase, and phenol treatment were required. Cyran et al (2004) and Cyran and Saulnier (2005, 2007) reported that protein is associated with rye cell wall, as indicated by the effect of protease treatment in releasing cell wall materials from rye grains and totally eliminating all the protein present in the sample. In consecutive extraction (using barium hydroxide, water, 1.0N NaOH,

and 4.0N NaOH) of rye whole meal after treatment with amylase and protease, the population of both polysaccharides (β-glucans and AX) extracted with 1.0N NaOH was associated with protein and lignin-like components. The polysaccharides extracted by 1.0N NaOH were 66–68% AX and 28–32% β-glucan. In this study, direct addition of alcohol to the neutralized alkali extract from rye whole meal resulted in a product containing (in addition to β-glucan and AX) ≈16.0% protein, and therefore a prior step using phenol to remove protein was used. Based on earlier observations, the low Mpc of the β-glucan in F0 was unexpected. Although a higher MW would have been preferable, we felt that the product F0 was adequate for purifying and assessing structural uniformity and molecular characteristics of rye β-glucan using ammonium sulfate fractionation. Precise application of gradient ammonium sulfate precipitation (Wang et al 2003) was used to effect purification of β-glucan from the alkali extracted rye β-glucan fraction F0. In the study of Wang et al (2003), this fractionation was applied to already partially purified oat and barley β-glucan. In the present study, an additional purpose was to simultaneously separate the β-glucan and AX. Thus fractions F2–F6 contained mostly β-glucan, whereas F7 and F8 contained significant levels of AX; the A/X ratio was higher in these latter two fractions. Somewhat unusually for ammonium sulfate fractionation (Preece and Mackenzie 1952; Izydorczyk et al 1998), the fraction F1 obtained at the lowest concentration of ammonium sulfate had a low concentration of β-glucan, contained a significant amount of AX, and was dark in color, indicating the presence of significant amount of protein. Additional phenol treatment gave fraction F1p in which most of the protein was removed with a resultant increase in the β-glucan content from 31.5 to 75%. The treatment

TABLE I Details of Fractionation of Alkali-Extractable Nonstarch Polysaccharides from Rye and Physicochemical Characteristics of Arabinoxylan and β-Glucan Fractions Fraction No.

Yielda (%)

β-Glucanb (%)

naf

F0 F1 F1pg F2 F3 F4 F5 F6 F7 F8

AXc (%)

32.0 31.5 75.0 80.0 81.0 83.0 88.0 79.0 23.5 9.4

7.7 na 6.9 4.6 5.4 6.2 5.4 9.2 12.3

A/X Ratio

29.3 6.5 12.6 3.4 3.0 3.8 3.0 5.1 16.3 18.3

Mpcd × 10–3

··· 0.45 0.25 0.52 0.56 0.53 0.58 0.58 0.75 0.89

Mwe × 10–3

Mpe × 10–3

Pde

[η]e

··· 770 716 519 537 300 229 161 1,516 1,781

··· 778 722 518 507 286 218 150 592 430

··· 1.15 1.11 1.08 1.08 1.01 1.01 1.06 5.99 4.95

··· 6.8 6.6 5.5 5.8 4.3 3.6 2.7 5.9 6.8

372 785 778 524 367 264 185 112 64 47

% of β-glucan from F0. β-Glucan content in each fraction. c (Arabinose + Xylose) × 0.88. d Peak MW determined by HPSEC-calcofluor detector. e Weight average (M ) and peak (M ) MW, polydispersity index (P )and intrinsic viscosity [η] determined by HPSEC with the Viscotek triple detector system. w p d f Not applicable. g Fraction F1 after phenol treatment. a

b

TABLE II Molar Ratio of 3-O-Β-cellobiosyl-D-Glucose and 3-O-β-Cellotriosyl-D-Glucose and Relative Proportions (%) of Oligosaccharides for Alkali-Extractable Rye β-Glucan (F0) and Ammonium Sulfate Precipitated Fractions (F1–F8). Fraction F1 F2 F3 F4 F5 F6 F7 F8 F0 a

Molar Ratioa

Tri

Tetra

Penta

Hexa

Hepta

Octa

Nona

DP 5–9

3.44 3.48 3.50 3.45 3.45 3.48 3.46 3.42 3.46

66.2 66.4 66.6 66.3 66.3 66.5 66.0 65.5 65.8

25.4 25.2 25.2 25.4 25.4 25.2 25.2 25.3 25.1

3.1 3.1 3.0 3.0 3.1 3.0 3.1 3.3 3.3

1.8 1.9 1.8 1.8 1.8 1.8 1.9 2.0 2.0

1.7 1.7 1.6 1.7 1.6 1.7 1.8 1.9 1.9

0.7 0.6 0.6 0.6 0.6 0.6 0.7 0.6 0.6

1.2 1.2 1.2 1.2 1.2 1.2 1.3 1.4 1.4

8.4 8.4 8.2 8.3 8.3 8.3 8.8 9.2 9.1

Molar ratio = [(peak area of the trisaccharide/peak area of tetrasaccharide) ×1.321].

292

CEREAL CHEMISTRY

also increased AX content from 6.5 to 12.6%, slightly less proportionately than the increase in β-glucan, and reduced the A/X ratio from 0.45 to 0.25, suggesting either a lability of the arabinofuranose residues or removal, with protein, of a more branched AX fraction, leaving a less branched fraction, which Izydorczyk and MacGregor (2000) suggested tightly associates with β-glucan. The usual range quoted for A/X ratio in whole rye products is ≈0.5– 0.8 and the lowest for the water-soluble fraction is ≈0.5 (Vinkx and Delcour 1996). Different fractions of rye have a wider range (0.1–1.1) of A/X ratio for AX extracted by alkali (Vinkx and Delcour 1996), but the lower values (0.1–0.3) usually lead to water insolubility following neutralization so we would not expect such a fraction to be present in F0 or F1. This suggests that the lower value (0.25) in F1p is at least in part due to acid lability of the arabinofuranose substituents, although the water solubility is anomalous and instead indicates association with the more water-soluble β-glucan, as suggested by Izydorczyk and MacGregor (2000). The slight reduction (Table I) in the Mw (770K) of F1 to 716K in F1p reflects one or both of these behaviors in the AX because phenol treatment did not affect MW of β-glucan (Ragaee et al, unpublished), which was our primary interest. HPSEC with calcofluor detection showed the expected decline in Mpc of β-glucan in fractions obtained with increasing concentrations of ammonium sulfate for F1–F8. The HPSEC triple detector system revealed more details about the characteristics of these fractions. There was a gradual reduction in Mp as the fractionation procedure proceeded from F1 to F6, and these fractions had a narrow molecular size distribution. The results are similar to those reported by Wang et al (2003) for oat and barley β-glucan. Mp values obtained by HPSEC calcofluor (Mpc) and HPSEC triple detector (Mp) were similar when the fraction was mostly β-glucan (F1–F6), but when AX became the predominant polysaccharide (F7 and F8), the results differed. Mp, values obtained by the triple detector were higher than those with calcofluor detection, indicating that the AX that co-precipitated with the lower MW β-glucan had significantly higher MW. This became very apparent in 1,516 and 1,781K values for Mw (Table I), which reflects the marked increase in Pd for F7 and F8. This indicated the heterogeneity of the AX component of these fractions, which continued to show narrow distributions with calcofluor detection because this is selective for β-glucan. Fractions F2–F6 were sufficiently pure to apply the Mark-Houwink relationship, [η] = K(MW)α. The value of the exponent α was 0.60, close to the values 0.65 and 0.62 reported for oat and barley β-glucan in water (Wang et al 2003), indicating similar conformations for all three β-glucans in aqueous solution. The oligosaccharides released by lichenase, a (1→3)(1→4)-βD-glucan-4-glucanohydrolase that specifically cleaves the (1→4) linkage of the 3-O-substituted glucose units in β-glucan, provide a fingerprint of the original structure (Wood et al 1991a, 1994b). The major products (≈90%) 3-O-β-cellobiosyl-D-glucose and 3O-β-cellotriosyl-D-glucose arise from the β-(1→3)-linked cellotriosyl and cellotetraosyl units in the original chain. The relative amounts of these are characteristic of different cereal β-glucans. Our results indicate that the different fractions obtained from rye, like oat and barley β-glucan (Wang et al 2003), essentially have a homogeneous structure, although there was a slight trend toward a greater amount of more cellulose-like regions in the fractions more soluble in ammonium sulfate. The structures were similar to the whole original unfractionated rye whole meal (Ragaee et al, unpublished). Thus, the dominant mechanism of fractionation by ammonium sulfate for rye is MW, as for oat and barley β-glucans (Wang et al 2003). The greater the proportion of β-(1→3)-linked cellotriosyl units in the polysaccharide (i.e., the greater the molar ratio between 3O-β-cellobiosyl-D-glucose and 3-O-β-cellotriosyl-D-glucose), the greater the frequency of consecutive occurrence of this structure in the chain, which leads to a greater tendency to gel (Böhme and

Kulicke 1999; Lazaridou et al 2004; Tosh et al 2004a). The molar ratio for F0 and the eight fractions was 3.4–3.5. This ratio is intermediate between oat and wheat and similar to barley using the HPAEC fractionation described. LITERATURE CITED Beer, M. U., Wood, P. J., and Weisz, J. 1997. Molecular weight distribution and (1→3),(1→4)-β-D-glucan content of consecutive extracts of various oat and barley cultivars. Cereal Chem. 74:476-80. Bhatty, R. S. 1993. Extraction and enrichment of (1→3)(1→4)-β-Dglucan from barley and oat brans. Cereal Chem. 70:73-77. Böhm, N., and Kulicke, W. M. 1999. Rheological studies of barley (1→3),(1→4)-β-glucan in concentrated solution: Mechanistic and kinetic investigation of the gel formation. Carbohydr. Res. 315:302-311. Braaten, J. T., Wood, P. J., Scott, F. W., Wolynetz, M. S., Lowe, M. K., Bradley-White, P., and Collins M. W. 1994. Oat β-glucan reduces blood cholesterol concentration in hyperchole-sterolemic subjects. Eur. J. Clin. Nutr. 48:465-474. Brummer, Y., Jones, S., Tosh, S. M., and Wood, P. J. 2008. Extraction and physicochemical characterization of rye β-glucan and effects of barium on polysaccharide molecular weight. Cereal Chem. 85:174-181. Colleoni-Sirghie, M., Kovalenko, I. V., Briggs, J. L., Fulton, B., and White, B. J. 2003. Rheological and molecular properties of water soluble (1,3)(1,4)-β-D-glucan from high β-glucan and traditional oat lines. Carbohydr. Polym. 52:439-447. Cui, W., Wood, P. J., Weisz, J., and Beer, M. U. 1999. Nonstarch polysaccharides from preprocessed wheat bran: Carbohydrate analysis and novel rheological properties. Cereal Chem. 76:129-133. Cyran, M. R., Courtin, C. M., and Delcour, J. A. 2004. Heterogeneity in the fine structure of alkali-extractable arabinoxylans isolated from two rye flours with high and low breadmaking quality and their coexistence with other cell wall components. J. Agric. Food Chem. 52:2671-2680. Cyran, M. R., and Saulnier, L. 2005. Cell wall fractions isolated from outer layers of rye grain by sequential treatment with α-amylase and proteinase: Structural investigation of polymers in two ryes with contrasting breadmaking quality. J. Agric. Food Chem. 53:9213-9224 Cyran, M. R., and Saulnier, L. 2007. Association and structural diversity of hemicelluloses in the cell walls of rye outer layers: Comparison between two ryes with opposite breadmaking quality. J. Agric. Food Chem. 55:2329-2341. Edney, M. J., Marchylo, B. A., and MacGregor, A. W. 1991. Structure of total barley β-glucan. J. Inst. Brew. 97:39-44. FDA. 1997. 21 CFR Part 101. Food Labeling, Health Claims; Soluble Dietary Fiber From Certain Foods and Coronary Heart Disease. Federal Register 62(15):3584-3601. FDA. 2005. 21 CFR Part 101. Food Labeling, Health Claims; Soluble Dietary Fiber From Certain Foods and Coronary Heart Disease. Federal Register, 70(246):76150-76162. Henry, R. J. 1987. Pentosan and (1→3),(1→4)-β-glucan concentrations in endosperm and whole grain of wheat, barley, oats and rye. J Cereal Sci. 6:253-258. Izydorczyk, M. S., Macri, L. J., and MacGregor, A. W. 1998. Structure and physicochemical properties of barley non-starch polysaccharides. II. Alkali-extractable β-glucans and arabinoxylans. Carbohydr. Polym. 35:259-269. Izydorczyk, M. S., and MacGregor, A. W. 2000. Evidence of intermolecular interactions of β-glucans and arabinoxylans. Carbohydr. Polym. 41:417-420. Jenkins, D. J. A., Wolever, T. M. S., Vuksan, V., Brighenti, F., Cunnane, S., Venketeshwer, V. R., Jenkins, A. L., Buckley, G., Patten, R., Singer, W., Corey P., and Josse, R. G. 1989. Nibbling versus gorging: Metabolic advantages of increased meal frequency. New Engl. J. Med. 321:929-934. Jenkins, D. J. A., Jenkins, A. L., Kendall, C. W. C., Augustine, L., and Vuksan, V. 2001. Dietary fibre, carbohydrate metabolism and chronic disease. Pages 162-167 in: Advanced Dietary Fibre Technology. B. V. McCleary and L. Prosky, eds. Blackwell Science: Oxford, UK. Johansson, L., Virkki, L., Maunu, S., Lehto, M., Ekholm, P., and Varo, P. 2000. Structural characterization of water soluble β-glucan of oat bran. Carbohydr. Polym. 42:143-148. Lazaridou, A., Biliaderis, C. G., Micha-Screttas, M., and Steele, B. R. 2004. A comparative study on structure-function relations of mixedlinkage (1→3),(1→4)-β-D-glucans. Food Hydrocolloids 18:837-855. Vol. 85, No. 3, 2008

293

Lee, C. J., Horsley, R. D., Manthey, F. A., and Schwarz, P. B. 1997. Comparisons of β-glucan content of barley and oat. Cereal Chem. 74:571-575. Leinonen, K., Liukkonen, K., Poutanaen, K., Uusitupa, M., and Mykkänen, H. 1999. Rye bread decreases postprandial insulin reponse but does not alter glucose reponse in healthy subjects. Eur. J. Clin. Nutr. 53:262-267. Leinonen, K., Poutanaen, K., and Mykkänen, H. 2000. Rye bread decreases serum total and LDL cholesterol in Finnish men with moderately elevated serum cholesterol. J. Nutr. 130:164-170. Li, W., Cui, S. W., and Kakuda, Y. 2006. Extraction, fractionation, structural and physical characterization of wheat β-D-glucans. Carbohydr. Polym. 63:408-416. Liu, S., Manson, J. E., Stampfer, M. J., Hu, F. B., Giovannucci, E., Colditz, G. A., Hennekens, C. H., and Willett, W. C. 1999. Whole-grain consumption and risk of coronary heart disease, results from the Nurses Health Study. Am. J. Clin. Nutr. 70:412-419. McCleary, B. V., and Glennie-Holmes, M. 1985. Enzymic quantification of (1→3)(1→4)-β-D-glucan in barley and malt. J. Inst. Brew. 91:285295. Nilsson, M., Saulnier, L., Andersson, R., and Åman, P. 1996. Water unextractable polysaccharides from three milling fractions of rye grain. Carbohydr. Polym. 30:229-237. Parrish, F. W., Perlin, A. S., and Reese, E. T. 1960. Selective enzymolysis of poly-β-D-glucans, and the structure of the polymers. Can. J. Chem. 38:2094-2104. Peterson, D. M. 1991. Genotype and environmental effects on oat βglucan concentration. Crop Sci. 31:1517-1520. Preece, I. A., and MacKenzie, K. G. 1952. Non-starch polysaccharides of cereal grains. II. Distribution of water soluble gum like material in cereals. J. Inst. Brew. 58:457-464. Ragaee, S. M., Campbell, G. L., Scoles, G. J., McLeod, J. G., and Tyler, R. T. 2001. Studies on rye (Secale cereale L.) lines exhibiting a range of extract viscosities. I. Composition, molecular wight distribution of water extracts, and biochemical characteristics of purified waterextractable arabinoxylan. J. Agric. Food Chem. 49:2437-2445. Roubroeks, J. P., Andersson, R., and Åman, P. 2000. Structural features of (1→3),(1→4)-β-D-glucan and arabinoxylan fractions isolated from rye bran. Carbohydr. Polym. 42:3-11. Saastamoninen, M., Plaami, S., and Kumpulainen, J. 1989. Pentosan and β-glucan content of Finnish winter rye varieties as compared with rye of six other countries. J. Cereal Sci. 10:199-207. Saini, H. S., and Henry, R. J. 1989. Fractionation and evaluation of triticale pentosans: Comparison with wheat and rye. Cereal Chem. 66:11-14. Salmerón, J., Manson, J. E., Stampfer, M. J., Colditz, G. A., Wing, A. L., and Willett, W. C. 1997. Dietary fiber, glycemic load, and risk of noninsulin-dependent diabetes mellitus in women. J. Am. Med. Assoc. 277:472-477. Seibel, W., and Weipert, D. 2001. Bread baking and other food uses

around the world. In: Rye Production Chemistry and Technology. W. Bushuk, ed. AACC International: St. Paul, MN. Tosh, S. M., Brummer, Y., Wood, P. J., Wang, Q., and Weisz, J. 2004a. Evaluation of structure in the formation of gels by structurally diverse(1→3),(1→4)-β-D-glucans from four cereal and one lichen species. Carbohydr. Polym. 57:249-259. Tosh, S. M., Wood, P. J., Wang, Q., and Weisz, J. 2004b, Structural characteristics and rheological properties of partially hydrolyzed oat betaglucan: The effects of molecular weight and hydrolysis method. Carbohydr. Polym. 55:425-436. Vinkx, C. J. A., and Delcour, J. A. 1996. Rye (Secale cereal L.) arabinoxylan: A critical review. J. Cereal Sci. 24:1-14. Wang, Q., Wood, P. J., Huang, X., and Cui, W. 2003. Preparation and characterization of molecular weight standards of low polydispersity from oat and barley (1→3),(1→4)-β-D-glucan. Food Hydrocolloids 17:845-853. Wood, P. J., Weisz, J., Fedec, P., and Burrows, V. D. 1989. Large scale preparation and properties of oat fractions enriched in (1→3)(1→4)-Dglucan. Cereal Chem. 66:97-103. Wood, P. J., Weisz, J., and Blackwell, B. A. 1991a. Molecular characterization of cereal β-D-glucans, Structural analysis of oat β-glucan and rapid structural evaluation of β-glucans from different sources by highperformance liquid chromatography of oligosaccharides released by lichenase. Cereal Chem. 68:31-39. Wood, P. J., Weisz, J., and Mahn, W. 1991b. Molecular characterization of cereal β-D-glucans. II. Size-exclusion chromatography for comparison of molecular weight. Cereal Chem. 68:530-536. Wood, P. J., Braaten, J. T., Scott, F. W., Riedel, K. D., Wolynetz M. S., and Collins, M. W. 1994a. Effect of dose and modification of viscous properties of oat gum on plasma glucose and insulin following an oral glucose load. Brit. J. Nutr. 72:731-743. Wood, P. J., Weisz, J., and Blackwell, B. A. 1994b. Structural studies of (1→3),(1→4)-β-D-glucans by 13C-nuclear magnetic resonance spectroscopy and by rapid analysis of cellulose-like regions using highperformance anion exchange chromatography of oligosaccharides released by lichenase. Cereal Chem. 71:301-307. Wood , P. J., Weisz, J., Beer, M. U., Newman, C. W., and Newman, R. K. 2003. Structure of (1→3)(1→4)-β-glucan in waxy and nonwaxy barley. Cereal Chem. 80:329-332. Westphal, V. O., Ludebuz, O., and Bister, F. 1952. Uber die extraction von bakterien mit phenol/wasser. Zeitschr. Naturforschung 7b:148-155. Yao, N., Jannink, J.–L., and White, P. J. 2007. Molecular weight distribution of (1→3)(1→4)-β-glucan affects pasting properties of flour from oat lines with high and typical amounts of β-glucan. Cereal Chem. 84:471-479. Zhang, D., Doehlert, D. C., and Moore, W. R. 1998. Rheological properties of (1→3)(1→4)-β-D-glucans from raw, roasted, and steamed oat groats. Cereal Chem. 75:433-438.

[Received July 24, 2007. Accepted December 12, 2007.]

294

CEREAL CHEMISTRY