Arabinoxylan Gels: Impact of the Feruloylation Degree on Their

Alkaline extraction conditions determine gelling properties of corn bran arabinoxylans. Madhuvanti S. Kale , Bruce R. Hamaker , Osvaldo H. Campanella...
3 downloads 0 Views 198KB Size
Biomacromolecules 2005, 6, 309-317

309

Arabinoxylan Gels: Impact of the Feruloylation Degree on Their Structure and Properties Elizabeth Carvajal-Millan,† Virginie Landillon,† Marie-He´ le` ne Morel,† Xavier Rouau,† Jean-Louis Doublier,‡ and Vale´ rie Micard†* U.M.R. Inge´ nierie des Agropolyme` res et des Technologies Emergentes, ENSAM/INRA, UMII/CIRAD, 2 Place Pierre Viala 34060, Montpellier Cedex 1, France, and UPCM, INRA, rue de la Ge´ raudie` re, 44316, Nantes Cedex 3, France Received June 29, 2004; Revised Manuscript Received September 27, 2004

Arabinoxylan (AX) samples of decreasing ferulic acid (FA) contents were chemically prepared from waterextractable wheat arabinoxylans without affecting their other structural properties. Gels were obtained from these partially feruloylated WEAX (PF-WEAX) by enzymatic covalent cross-linking of FA leading to the formation of diferulic (di-FA) and tri-ferulic acid (tri-FA). WEAX gelling ability was found related to the WEAX FA content whereas the gel structure and properties depended on the density of newly formed covalent cross-links. FA content of WEAX ranging from 1.4 to 2.3 µg/mg AX gave gels with di-FA crosslinks contents from 0.20 to 0.43 µg/mg AX and G′ values from 5 to 44 Pa. For WEAX gels with initial FA contents from 1.6 to 2.3 µg/mg AX, average mesh size ranging from 331 to 263 nm were calculated from swelling experiments. Cross-linking densities of gels, determined from swelling experiments, were higher than those that could be theoretically estimated from the di-FA and tri-FA content of WEAX gels. This result suggests that, in addition to di-FA and tri-FA, higher ferulate cross-linking and physical entanglements would contribute to the final WEAX gel structure. Introduction Polysaccharide properties such as solubility, viscosity, and gelling capacity are closely related to their chemical structure, conformation, and molecular interaction. The close structureproperties relationship has increased the interest to design polysaccharides in order to develop and/or improve their functional properties.1 Arabinoxylans (AX), the main nonstarchy polysaccharides of cereal grains,2 are constituted of a linear backbone of β-(1f4)-linked D-xylopyranosyl units to which R-L-arabinofuranosyl substituents are attached through O-2 and/or O-3.3 Some of the arabinose residues are ester linked on (O)-5 to ferulic acid (FA) (3-methoxy, 4 hydroxy cinnamic acid)4 (Scheme 1a). Arabinoxylans from endosperm are partly water-extractable (WEAX). Once water extracted they form highly viscous solutions5 with gelling capacity by covalent cross-linking of AX chains through dimerization of ferulic acid substituents under oxidative conditions (e.g., use of enzymatic free radical generating agents as laccase and peroxidase-H2O2)6-9 (Scheme 1b). Five main dimers of ferulic acid (di-FA) (5-5′, 8-5′ benzo, 8-O4′, 8-5′, and 8-8′ forms) have been identified in gelled AX, the 8-5′ and 8-O-4′ forms being always preponderant10-12 (Scheme 1b). The presence of a trimer of ferulic acid (triFA) (4-O-8′, 5′-5′′-dehydrotriferulic acid) in laccase crosslinked wheat WEAX has been recently reported.13 * To whom correspondence should be addressed. Telephone: +33-499612889. Fax: +33-4-67522094. E-mail address: [email protected]. † U.M.R. ‡ UPCM.

Unlike most polysaccharide gels, arabinoxylan gelation process and gel properties are governed by the establishment of both covalent (di-FA, tri-FA bridges) linkages and weak (hydrogen) interactions,12 which depend on AX structural characteristics such as molecular weight (Mw), xylan backbone substitution (A/X ratio), as well as FA content and location.3,14,15 Up to now, the relative importance of each structural parameter on AX gelation and gel properties has been tentatively studied by varying AX sources.14 Even if ferulic acid has been considered in that way to be of major importance in the WEAX gelation, its real contribution to gelation and gel properties remained unknown since the Mw and A/X ratio differed with AX sources.14 The production of WEAX differing in its ferulic acid content but with identical mean chain structure was achieved in the present study by chemical deesterification of a single WEAX source. Their enzymatic gelation allowed us to investigate the effect of FA content and therefore di-FA and tri-FA bridges formation on the WEAX gelation process and gel properties. WEAX gels were characterized in terms of cross-links structures content, rheological properties, and equilibrium swelling degree. Knowledge of the equilibrium swelling degree allowed the calculation of structural gel parameters like molecular weight between cross-links (Mc), mesh size (ξ), and cross-linking density (Fc). To better understand the real contribution of covalent and physical bonds to the WEAX gel structure, the cross-linking density (Fc) issue from swelling experiments was compared to two theoretical crosslinking densities estimated from the FA content in WEAX

10.1021/bm049629a CCC: $30.25 © 2005 American Chemical Society Published on Web 11/13/2004

310

Biomacromolecules, Vol. 6, No. 1, 2005

Carvajal-Millan et al.

Scheme 1. Structure of Feruloylated Arabinoxylans (a) in Solution and (b) in Gel. Xylose (O), Arabinose (0), Ferulic Acid (4), Di-FA (22)

before gelation or from the di-FA and tri-FA content of WEAX gels. Experimental Section Materials. Water extractable arabinoxylans (WEAX) were isolated from the wheat milling fraction 1LC3 (endosperm of wheat kernel) provided by the Grands Moulins de Paris (Gennevilliers, France). Laccase (benzenediol:oxygen oxidoreductase, E. C.1.10.3.2) from Pycnoporus cinnabarinus was supplied by the Unite´ de Biotechnologie des Champignons Filamenteux (UMR 1163 INRA-ESIL, Luminy, France). Citric acid, sodium phosphate dibasic, calcium hydroxide (lime), TFA (trifluoroacetic acid), syringaldazine, ferulic acid, and TMCA (3,4,5-trimethoxy-trans-cinnamic acid) were purchased from Sigma Chemical Co. (St Louis, MO). Lithium nitrate and sodium azide were purchased from Fluka Chemical Co. (Buchs, Switzerland). Termamyl 120L, amyloglucosidase, and Pronase were from Novozyme (Bagsvaerd, Denmark), Fluka (Steinheim, Switzerland) and Boehringer (Manheim, Germany), respectively. Isolation of WEAX. 1LC3 wheat milling fraction (1 kg) was treated with water (3 L) for 15 min at 25 °C. After centrifugation (12 096g, 20 °C, 15 min), the supernatant was treated with Termamyl 120L (100 °C, 30 min, 2800 U/g of flour), then amyloglucosidase (50 °C, pH 5, 3 h, 24 U/g of flour), and finally Pronase (20 °C, pH 7.5, 16 h, 0.4 U/g of flour). The pH in the extract was adjusted by addition of 1 N HCl or 1 N NaOH. After heating at 100 °C for 10 min, the de-starched and de-proteinized extract was centrifuged (12 096g, 20 °C, 15 min). The supernatant was allowed to precipitate in 65% (v/v) ethanol for 4 h at 4 °C. Precipitate

was recovered and dried by solvent exchange (80% (v/v) ethanol, absolute ethanol and acetone) to give WEAX. WEAX De-esterification. A WEAX solution (2% w/v in AX) was diluted with a saturated 1% (w/v) Ca(OH)2 solution and incubated at 20 °C in darkness for 0, 10, 20, 40, 80, 160, 320, and 480 min. Lime was then eliminated by centrifugation (5000g, 20 °C, 10 min), and the supernatant was acidified to pH 2 with 4 N HCl. Solutions were then precipitated in 65% (v/v) ethanol and centrifuged (1000g, 4 °C, 5 min). Precipitates were dried by solvent exchange (80% (v/v) ethanol, absolute ethanol and acetone) to give partially feruloylated WEAX (PF-WEAX). Chemical and Physicochemical Analyses. Laccase ActiVity. Laccase activity was measured at 25 °C from a laccase solution at 0.125 mg/mL dissolved in 0.1 M citrate-phosphate buffer pH 5.5 as previously reported.9,13 Neutral Sugars. Neutral sugar content in native WEAX and PF-WEAX samples were determined by hydrolysis of the polysaccharides with 2 N trifluoroacetic acid at 120 °C for 2 h. The reaction was stopped in ice and the extracts were evaporated under air at 40 °C, rinsed twice with 200 µL of water. The evaporated extract was solubilized in 500 µL of water. Inositol was used as internal standard. Samples were filtered through 0.45 µm (Whatman) and analyzed by HPLC using a Polyspher RT CH Pb column (7.8 × 300 mm; Merck, Darmstadt Germany) eluted with water (filtered 0.2 µm, Whatman) at 0.4 mL/min and 80 °C. A refractive index detector (Waters 410) was used. Phenolic Acids. Ferulic acid (FA), dimers of ferulic acid (di-FA), and trimers of ferulic acid (tri-FA) contents were determined in native WEAX and PF-WEAX and in WEAX gels after gelation after saponification by RP-HPLC.12,16,17

Arabinoxylan Gels

An Alltima C18 column (250 × 4.6 mm) (Alltech associates, Inc. Deerfield, IL) and a photodiode array detector Waters 996 (Millipore Co., Milford, MA) were used. Detection was followed by UV absorbance at 320 nm. Proteins. Protein content of native WEAX was determined according to the Dumas method,18 using a NA 2000 nitrogen and protein analyzer (Fisons Instruments, Arcueil, France) (N x 5.7). Ash. The content of ash was determined according to the AACC methods (76-31).19 Molecular Weight. Molecular weight distribution of native WEAX and PF-WEAX was determined by SE-HPLC at 38 °C using a Waters (Millipore Co, Milford, MA) Ultrahydrogel 1000 column (7.8 × 300 mm). Isocratic elution was done at 0.6 mL/min with 0.2 M LiNO3 filtered through 0.2 µm (Whatman). A total of 20 µL of WEAX solution (0.2 or 0.4% w/v in AX) filtered through 2.7 µm (Whatman) was injected and a Waters 600 differential refractometer was used for detection. The molecular weight values of WEAX and PF-WEAX samples were calculated from their intrinsic viscosity ([η]) by using the Mark-Houwink equation [η] ) K Mwa as reported by Dervilly-Pinel et al.20 The values of the constants K and a published by Dervilly-Pinel et al.20 for arabinoxylans in water were used (-1.2 and +0.74, respectively). Intrinsic Viscosity [η]. Viscosity measurements were made by determination of the flow times of 2 mL of native WEAX and PF-WEAX solutions in water (from 0.1 to 0.6% w/v in AX). Measurements were performed at 25 °C using an AVS 400 capillary viscosimeter (Schott Gera¨te, Hofheim, Germany), equipped with an Oswald capillary tube (water flow time 73.14 s). Intrinsic viscosity [η] of native and PFWEAX samples was obtained using the Mead, Kraemer and Fouss method.21,22 Partial Specific Volume (υ). Partial specific volume of native WEAX in water was determined using aqueous WEAX solutions from 0.5 to 1.5% (w/v) in AX.23 WEAX Gel Preparation. Native and PF-WEAX solutions (1% w/v in AX) were prepared in 0.05 M citrate phosphate buffer pH 5.5 filtered through 0.2 µm (Whatman). Laccase (1.675 nkat per mg AX) was added to native and PF-WEAX solutions. Gels were allowed to develop for 2 h at 25 °C. Rheological Measurements. Small amplitude oscillatory shear was used to follow the gelation process of native WEAX and PF-WEAX solutions. Cold (4 °C) WEAX solutions (1% w/v in AX) in 0.05 M citrate phosphate buffer pH 5.5 were mixed with laccase and immediately poured on cone-plate geometry (5.0 cm in diameter and 0.04 rad in cone angle) of a strain controlled rheometer (ARES 2000, Rheometric Expansion System, Rheometric Scientific, Champ sur Marne, France) maintained at 4 °C. Exposed edges were recovered with silicone to prevent evaporation. WEAX gelation was started by a sudden increase of temperature from 4 to 25 °C and monitored at 25 °C for 2 h by recording the storage (G′) and loss (G′′) moduli. Measurements were carried out at 6.3 rad/s. From strain sweep tests, WEAX gels showed a linear behavior from 0.02 to 100% strain. 10% strain was used in all of the rheological measurements. The

Biomacromolecules, Vol. 6, No. 1, 2005 311

mechanical spectra of gels were obtained by frequency sweep from 1 to 100 rad/s at 25 °C. Swelling Experiments. A total of 2 mL of native WEAX and PF-WEAX solutions (1% w/v in AX)/laccase (1.675 nkat per mg AX) mixture was quickly transferred to a 5 mL tip-cutoff syringe (diameter 1.5 cm) and allowed to gel for 2 h at 25 °C. After gelation, the gels were removed from the syringes, placed in glass vials, and weighed. The gels were allowed to swell in 20 mL of 0.02% (w/v) sodium azide solution to prevent microbial contamination. During 36 h, the samples were taken out, blotted, and weighed. The equilibrium swelling was reached when the weight of the samples changed by no more than 3% (( 0.03 g). After weighing, a new aliquot of sodium azide solution was added to the gels. Gels were maintained at 25 °C during the test. The swelling ratio (q) was calculated as q ) (Ws - WAX)/WAX

(1)

where Ws is the weight of swollen gels and WAX is the weight of AX in the gel. The AX weight in the gel was calculated by taking into account the AX concentration in the WEAX solution (1% w/v) and the fresh weight of the gel after removing from the syringe. WEAX Gel Structure. From swelling measurements, the molecular weight between two cross-links (Mc) was calculated using the classic Flory-Rehner24 modified by Peppas and Merrill analysis.25 Mc was also calculated by using the Treloar equation.26 These two methods were chosen as they were found most appropriate for gel structure studies in related literature.27-30 From the Mc values, the average mesh size (ξ) and the cross-linking density (Fc) in the WEAX gels were calculated according to Peppas et al.28,31 Theoretical Mc values were also calculated by taking into account the following hypotheses: (a) All FA residues esterifying AX chain before laccase exposure participate to covalent linkage (di-FA, tri-FA, non identifiable higher FA oligomer crosslinking structures) after oxidative gelation (McFA); (b) the di-FA and tri-FA quantified in the WEAX gel are the only covalent cross-linking points in the gel (Mcd+t). From McFA and Mcd+t values, theoretical cross-linking density values were also estimated (FFA, Fd+t). Results and Discussion Extraction and Characterization of WEAX. 71.2% of the WEAX initially present in the milling fraction have been recovered. Their composition is presented in Table 1. Pure AX represented 63% db of the WEAX. This value was estimated from the sum of xylose + arabinose after correction from the arabinose provided by arabinogalactan-proteins (A/G ) 0.7).32 The ratio arabinose-to-xylose (A/X ) 0.59) was typical of wheat endosperm arabinoxylans.33 The molecular weight (Mw) and intrinsic viscosity ([η]) values were 438 kDa and 5.68 dL/g, respectively, which are similar values to those reported for other wheat WEAX.12,15 The ferulic acid (FA) content (1.5 µg/mg WEAX or 2.3 µg/mg AX) was similar to that obtained with other wheat WEAX.9,11 Small amounts of di-FA were detected in WEAX (0.09 µg/mg WEAX or 0.14 µg/mg AX) in agreement with earlier

312

Biomacromolecules, Vol. 6, No. 1, 2005

Carvajal-Millan et al.

Table 1. Composition of WEAXa arabinoseb xyloseb glucoseb galactoseb mannoseb proteinb ashb ferulic acidc diferulic acidsc triferulic acidc

Table 2. Effect of Lime Desesterification of WEAX on the AX Intrinsic Viscosity Valuesa

26.0 ( 0.5 39.4 ( 1.2 1.40 ( 0.02 4.00 ( 0.04 0.60 ( 0.03 3.00 ( 0.01 5.12 ( 0.11 1.50 ( 0.01 0.09 ( 0.01 0.006 ( 0.001

a All results are obtained from duplicates. b Results are expressed in g/100 g WEAX dry matter. c Phenolics are expressed in µg/mg WEAX dry matter.

time exposure to lime (min)

FA esterifying AX (µg/mg AX)

[η] (dL/g)

A/X

0 10 20 40 80 160 320 480

2.30 ( 0.22 1.96 ( 0.10 1.83 ( 0.16 1.43 ( 0.25 1.27 ( 0.16 0.61 ( 0.17 0.20 ( 0.10 0.10 ( 0.01

5.68 ( 0.30 5.87 ( 0.28 6.42 ( 0.16 5.75 ( 0.81 5.93 ( 0.06 5.66 ( 0.11 5.71( 0.05 5.75 ( 0.05

0.59 0.60 0.60 0.61 0.58 0.60 0.59 0.59

a

Results are obtained from triplicates.

Figure 1. Kinetics of the ferulic acid release from feruloylated native WEAX by 1% (w/v) lime at 20 °C as a function of time exposure.

studies,9,34 suggesting that some arabinoxylan chains might be cross-linked.35 The relative percentages of each di-FA were 45, 42, and 13% for the 8-5′(mainly in the benzofuran form), 8-O-4′, and 5-5′ structures, respectively. The 8-8′ di-FA was not detected in this study. The predominance of 8-5′ and 8-O-4′ dimer structures has been previously reported in cereal arabinoxylans.12,15 A trimer of FA (tri-FA 4-O-8′, 5′-5′′) was detected in very low amounts (0.006 µg/ mg WEAX or 0.01 µg/mg AX), as recently reported in other wheat WEAX.13 WEAX Deferuloylation. The kinetics of deferuloylation of native WEAX in 1% (w/v) lime at 20 °C is presented in Figure 1. Results are expressed as FA (µg/mg AX) remaining linked to the arabinoxylan backbone at different times along the lime action. Deferuloylation was performed in triplicates for each reaction time. Deferuloylation followed a first-order kinetics according to the following relation: [FA] ) 2.216e-0.0103t + 0.114

(2)

This equation, predicting the feruloylation degree of WEAX as a function of time exposure to lime, allowed the preparation of a range of WEAX samples with FA content varying from 0.1 to 2.3 µg/mg AX (PF-WEAX), corresponding to an average FA residues content from 0.1 to 2.5 per 1000 xyloses, respectively. Dervilly-Pinel et al.15 reported 2-6 FA per 1000 xylose residues in WEAX from wheat, triticale, rye, and barley. The PF-WEAX were characterized in terms of [η], Mw, A/X ratio and FA, di-FA and tri-FA contents. As showed in Table 2, the [η] and A/X values of the PF-WEAX were not significantly affected by the lime treatment whatever the exposition time. The gel permeation

Figure 2. Size-exclusion HPLC elution profiles of native (thin line) and 1% (w/v) lime (480 min, 20 °C) treated WEAX (bold line).

chromatography profile of PF-WEAX obtained after 480 min of lime action was almost not different to that of native WEAX, showing that lime treatment did not depolymerized the arabinoxylan backbone (Figure 2). WEAX Gelation. The kinetics of gelation of native WEAX and PF-WEAX was rheologically monitored by small amplitude oscillatory shear (Figure 3a). The gelation process of WEAX with initial FA contents higher than 1.4 µg/mg AX (60% of the initial FA content) followed a characteristic profile with a lag time (∼10 min), due to the high viscosity of the WEAX solutions slowing down the enzyme/substrate reaction, followed by a rapid G′ rise then plateauing. At WEAX FA content of 0.1 and 0.4 µg/mg AX (4 and 18% of initial FA content) no G′ increase was observed suggesting that no gelation took place (final G′