Development of an Ion-Exchange Chromatography Method for

Mar 29, 2013 - Development of an Ion-Exchange Chromatography Method for Monitoring the Degradation of Prebiotic Arabinoxylan-Oligosaccharides in a ...
0 downloads 0 Views 427KB Size
Article pubs.acs.org/ac

Development of an Ion-Exchange Chromatography Method for Monitoring the Degradation of Prebiotic ArabinoxylanOligosaccharides in a Complex Fermentation Medium Audrey Rivière,† Sebastiaan Eeltink,‡ Christophe Pierlot,† Tom Balzarini,† Frédéric Moens,† Marija Selak,† and Luc De Vuyst*,† †

Research Group of Industrial Microbiology and Food Biotechnology and ‡Department of Chemical Engineering, Vrije Universiteit Brussel, Pleinlaan 2, B-1050 Brussels, Belgium ABSTRACT: Arabinoxylan-oligosaccharides (AXOS) are a new class of prebiotics with promising health-promoting characteristics. However, the mechanism by which bacteria break down these compounds in the colon is still uncharacterized, due to their structural complexity. A new analytical method that offers structural information was developed to characterize AXOS degradation during fermentation. The method was based on the simultaneous determination of arabinose, xylose, xylo-oligosaccharides (XOS), and AXOS by applying high-performance anion-exchange chromatography with pulsed amperometric detection. To study the structural features of AXOS in solution without the use of spectroscopic techniques or standards, enzymatic-based reference degradation chromatograms were generated based on enzymes with known specificity. The new method for fingerprinting showed to be a powerful and fast tool to study AXOS degradation with high repeatability with respect to peak area, peak width at half height, and retention time (respective relative standard deviations of ≤3.1%, 2.8%, and 0.8%). This method was successfully applied to study the degradation kinetics of AXOS in a complex fermentation medium by Bifidobacterium longum LMG 11047. The results showed that this strain could use both the arabinose side chains and xylose backbones up to xylotetraose. The characterization of the degradation abilities of AXOS by colon bacteria will allow a better understanding of the beneficial effects of these prebiotics. Furthermore, if the appropriate enzymes are available to design the reference degradation chromatograms, this new method for the qualitative fingerprinting of AXOS breakdown can also be applied for the breakdown of other complex oligosaccharides and polysaccharides.

T

preventing colon cancer.4−6 In this way, AXOS and AX contribute to the bifidogenic effect, that is, the selective stimulation of indigenous bifidobacteria; also, they stimulate the production of the health-promoting metabolites propionate and butyrate.7−9 Although several studies highlight the beneficial effects of AXOS and AX, their mechanistic breakdown by colon bacteria during fermentation is still unknown.10,11 This is due to the high variability in structural features, encompassing different degrees of polymerization and substitution, which makes these substrates difficult to study during the colon fermentation process. AX consist of a linear backbone of 1500−15 000 β(1,4)-linked D-xylopyranosyl (Xylp) units, which can be substituted with α-L-arabinofuranosyl (Araf) residues positioned on C-(O)-2 or C-(O)-3 (monosubstituted) or on both C-(O)-2 and C-(O)-3 (disubstituted) (Figure 1).12 The xylose residues can be esterified with glucuronic acid and to a lesser

he human colon can be seen as an anaerobic bioreactor, containing up to 1011−1012 bacterial cells per mL of luminal content that break down the undigested food compounds leaving the small intestine.1 The fermentation activity of colon bacteria on undigested food compounds leads to different metabolite profiles. In the proximal colon, carbohydrate fermentation produces beneficial short-chain fatty acids (SCFA), such as acetate, propionate, and butyrate; in the distal colon, protein fermentation generates mainly unwanted and potentially toxic end-products, such as amines and phenolic compounds.2 This has encouraged the development of approaches to manage growth and/or activity of desirable colon bacteria to improve nutrition and health by consuming, among others, nondigestible food components, termed prebiotics.3 The consumption of complex oligo- and polysaccharides, such as arabinoxylan-oligosaccharides (AXOS) and arabinoxylans (AX), is of growing interest. These carbohydrates are slower fermented by the colon bacteria than linear carbohydrates. Hence, they shift carbohydrate fermentation from proximal toward more distal parts of the colon, possibly © XXXX American Chemical Society

Received: January 18, 2013 Accepted: March 29, 2013

A

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

AXOS and AX because of its many advantages. These encompass a fast analysis time, no need for time-consuming pretreatments such as derivatization, a high resolution (separation of anomeric and positional isomers), and detection down to the picomolar level.28−30 However, no straightforward elution rules are applicable for branched oligosaccharides in contrast with the monosaccharides that they are composed of and their linear oligosaccharide homologues.28 Hence, standards for all possible oligosaccharides are necessary to perform a quantitative analysis, which is a big disadvantage, as no commercial AXOS and AX standards are available.28 The aim of this paper was to describe a simple, fast, and highresolution analytical method based on high-performance anionexchange chromatography (HPAEC) coupled to PAD to characterize AXOS degradation by colon bacteria. Whereas standards or advanced spectroscopic techniques are needed to identify complex carbohydrates in HPAEC-PAD chromatograms in general, a new method for identification was applied during the present study.

Figure 1. Example of arabinoxylan-oligosaccharide with its main structural units: unsubstituted D-xylopyranosyl (Xylp), monosubstituted Xylp at C-(O)-2 or C-(O)-3 with α-L-arabinofuranosyl (Araf), and disubstituted Xylp at C-(O)-2,3 with Araf. The cleavage sites of arabinofuranosidases GH51 (Meripilus giganteus) and GH43 (Humicola insolens) are shown.

extent with 4-O-methyl glucuronic acid at the C-(O)-2 positions,13 while acetyl groups can be linked to the C-(O)-2 and/or C-(O)-3 positions.12 Additionally, arabinose residues can be esterified with phenolic acids, such as ferulic acid and pcoumaric acid, at the C-(O)-5 positions.12,14 AX can be enzymatically cleaved with endoxylanase into AXOS and xylooligosaccharides (XOS; backbone of Xylp units).11,15 Consequently, there is a growing need for simple, fast, and highresolution analytical methods to characterize AXOS and AX routinely. Both separation and spectroscopic techniques have been used in the past for the analysis of complex oligosaccharides. Examples of separation techniques include gas chromatography, high-performance liquid chromatography, capillary electrophoresis, and size-exclusion chromatography (SEC). However, several disadvantages are associated with these techniques, including a slow and tedious nature of sample preparation, a limited resolution and high detection limit, a low sensitivity for low concentrations, and the disability to separate linkage isomers, respectively.16 Often, these separation techniques are used together with spectroscopic ones, such as nuclear magnetic resonance (NMR) and mass spectrometry (MS), to obtain additional structural information.16,17 The main disadvantages of NMR are the requirement of relatively high amounts of sample and the poor resolution of epimers.18 MS has the disadvantage that isomers with the same m/z values cannot be distinguished.16 As a result, anion-exchange chromatography (AEC) coupled with pulsed amperometric detection (PAD) is often used for AXOS and AX analysis. Detection with PAD is based on the oxidation of the eluted carbohydrates as oxyanions at a gold electrode that generates an electrical current. To avoid that oxidation products poison the electrode surface and hence cause signal loss, the electrode surface is cleaned by a series of potentials, defined as a waveform, once the detection potential has been applied.19 AEC is often used in combination with NMR,20 NMR and matrix-assisted laser desorption/ionization with time-of-flight MS,21,22 or NMR and SEC.23 Also, the online coupling of AEC and MS is possible by means of an anion micromembrane suppressor to remove the excess of sodium ions to be compatible with electrospray ionization.24 Likewise, AEC/MS in combination with microdialysis sampling is effective for the in situ profiling of oligo- and polysaccharides during enzymatic degradation.25 Finally, complex AXOS and AX can be hydrolyzed with H2SO4 or HCl, followed by separation of the monomers through AEC-PAD.26,27 However, the latter methodology is associated with the need to remove sulfate ions from the hydrolysates and a loss of detailed structural information, including side-chain substitution patterns.27 Yet, AEC-PAD is a promising technique for routine analysis of



MATERIALS AND METHODS Materials. Wheat AXOS, with an arabinose/xylose ratio of 0.27, average degree of polymerization (DP) of 5, 1% (m m−1) ferulic acid, and no glucuronic acid, were kindly provided by Fugeia (Leuven, Belgium). Aqueous solutions of 5 g L−1 were used throughout this study. Arabinose (Sigma−Aldrich, Stenheim, Germany), xylose (Sigma−Aldrich), and XOS (xylobiose, xylotriose, xylotetraose, xylopentaose, and xylohexaose; Megazyme International, Bray, Ireland) were used as standards for identification during HPAEC-PAD analysis. The arabinofuranosidases GH43 and GH51 isolated from Humicola insolens and Meripilus giganteus, respectively, were kindly provided by Novozymes (Bagsvaerd, Denmark). According to the supplier’s instructions, solutions of 3.36 and 18.86 mg mL−1, respectively, were used throughout this study. NaOH solution (50%) and anhydrous NaCH 3 COOH (≥99.0%) were purchased from Boom (Meppel, The Netherlands) and Sigma−Aldrich, respectively. For the mobile phase of the HPAEC runs, a ternary gradient was used, consisting of ultrapure water (eluent A), 200 mM NaOH (eluent B), and 100 mM NaOH with 400 mM NaCH3COOH (eluent C), modified after Pollet et al.31 The mobile phase C was filtered with a 0.45 μm nylon filter (Nalgene, Langenselbold, Germany). During analysis, the eluents were stored under an inert gas atmosphere of helium (Praxair, Schoten, Belgium) to prevent carbonate accumulation. Medium for colon bacteria (MCB), a complex fermentation medium, which supports the growth of various human colon bacteria when supplemented with an appropriate energy source, was used to perform the laboratory fermentations.32 This medium is composed of the following components (in g L−1): bacteriological peptone (Oxoid, Basingstoke, Hampshire, UK), 6.5; soy peptone (Oxoid), 5.0; yeast extract (VWR International, Darmstadt, Germany), 3.0; tryptone (Oxoid), 2.5; NaCl (VWR International), 4.5; KCl (Merck, Darmstadt, Germany), 2.0; MgSO4·7H2O (Merck), 0.5; CaCl2·2H2O (Merck), 0.45; cysteine−HCl (Merck), 0.4; NaHCO3 (VWR International), 0.2; MnSO4·H2O (VWR International), 0.2; FeSO4·7H2O (Merck), 0.005; ZnSO4·7H2O (VWR International), 0.005; hemin (Sigma−Aldrich), 0.005; menadione (Sigma−Aldrich), 0.005; H3PO4 (Merck), 0.5 mL L−1; and Tween 80 (Merck), 2 mL L−1. All components were dissolved in ultrapure water, and the medium was adjusted to pH 6.3 by means of a 37% HCl B

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

and a 10 M NaOH solution (VWR International). MCB that did not contain AXOS was used as the blank to evaluate possible matrix interferences. Bifidobacterium longum LMG 11047 was purchased from the Belgian Co-Ordinated Collections of Microorganisms/Laboratory for Microbiology Ghent (BCCM/LMG; Ghent, Belgium). This strain grows well on inulin-type fructans.33 Ion-Exchange Chromatography Instrumentation and Conditions. Separation experiments were performed on an ICS-3000 chromatograph at controlled room temperature, equipped with an ED40 PAD, an AS50 autosampler, and a SP-1 single pump, all from Thermo Scientific (Waltham, MA). Two different types of columns were tested: a CarboPac PA100 analytical column (250 mm × 4 mm) with a CarboPac PA100 guard column (50 mm × 4 mm) and a CarboPac PA200 analytical column (250 mm × 3 mm) with a CarboPac PA200 guard column (50 mm × 3 mm), all from Thermo Scientific. The respective stationary phases of the CarboPac PA100 and PA200 columns consisted of an 8.5 μm (55% cross-linked) and 5.5 μm diameter (6% cross-linked) ethylvinylbenzene/ divinylbenzene substrate, agglomerated with a 275 or 34 nm MicroBead quaternary amine functionalized latex. Chromeleon software 6.7 (Thermo Scientific) was used for system control and data analysis. Before injection (10 μL) into the column, samples were deproteinized, to avoid interference and signal diminution through trailing of amino acids and peptides on the gold electrode,34 by adding 300 μL of Carrez A reagent (36 g L−1 of K4Fe(CN)6·3H2O) and 300 μL of Carrez B reagent (72 g L−1 of ZnSO4·7H2O) to 600 μL of a two times diluted AXOS sample. After microcentrifugation (14 000 rpm for 15 min), the supernatant was filtered (pore size of 0.2 μm; UnifloRC filters, Whatman, Dassel, Germany). The mobile phase consisted of a linear segmented ternary gradient (Table 1). As the separation performance of two columns differing in internal diameter (i.d.) was compared, using the same eluent concentrations and injection volumes, the flow rate (usually 1 mL min−1) and gradient conditions for the CarboPac PA100 column were adapted (Table 1). Settings for Pulsed Amperometric Detection. The optimal settings for the PAD ED40 detector were determined by testing three different waveforms, consisting of a detection potential, cleaning potential(s) and reactivation potential. The first waveform (waveform A) had the following quadruple potential: E1 = 0.10 V (t1 = 0.00−0.40 s) with integration from 0.20 to 0.40 s; E2 = −2.00 V (t2 = 0.41−0.42 s); E3 = 0.60 V (t3 = 0.43 s); and E4 = −0.10 V (t4 = 0.44−0.50 s). The second triple waveform (waveform B) tested was as follows: E1 = 0.05 V (t1 = 0.00−0.40 s) with integration from 0.20 to 0.40 s; E2 = 0.75 V (t2 = 0.41−0.60 s); and E3 = −0.15 V (t3 = 0.61−1.00 s). Finally, the triple waveform C was tested as follows: E1 = 0.05 V (t1 = 0.00−0.48 s) with integration from 0.28 to 0.48 s; E2 = 0.60 V (t2 = 0.49−0.61 s); and E3 = −0.60 V (t3 = 0.62−0.69 s). Qualitative Identification Method for Uncharacterized AXOS. To elucidate the structures of the AXOS in solution using the HPAEC-PAD chromatogram profiles, the AXOS chromatograms were compared with reference degradation chromatograms that were generated through degradation of AXOS in solution by enzymes with known specificities, namely, arabinofuranosidase GH43 releasing Araf residues from doubly substituted Xylp residues at position C-(O)-3 in AX and AXOS and arabinofuranosidase GH51 liberating Araf residues from singly substituted Xylp in AX and AXOS at positions C-(O)-2

Table 1. Gradient Elution Conditions Applied for the Separation of Arabinose, Xylose, XOS, and AXOS through HPAEC-PAD with CarboPac PA200 or CarboPac PA100 Columns, Using a Ternary Gradient of Eluent A, Ultrapure Water; Eluent B, 200 mM NaOH; and Eluent C, 100 mM NaOH with 400 mM NaCH3COOH ternary gradient −1

flow rate (mL min ) CarboPac PA200 0.500

CarboPac PA100 0.899

time (min)

A (%)

B (%)

C (%)

0.0 10.0 10.1 15.0 55.0 80.0 85.0 85.1 90.0

90.0 90.0 50.0 50.0 34.0 0.0 0.0 90.0 90.0

10.0 10.0 50.0 50.0 34.0 0.0 0.0 10.0 10.0

0.0 0.0 0.0 0.0 32.0 100.0 100.0 0.0 0.0

0.0 10.0 10.1 15.0 55.0 80.0 85.0 85.1 90.0

94.4 94.4 71.9 71.9 62.9 43.8 43.8 94.4 94.4

5.6 5.6 28.1 28.1 19.1 0.0 0.0 5.6 5.6

0.0 0.0 0.0 0.0 18.0 56.2 56.2 0.0 0.0

and C-(O)-3 (Figure 1).35 Therefore, the AXOS solutions (5 g L−1) were incubated with 10 ppm of GH43 solution at pH 6.7 and 53 °C for 24 h or with 10 ppm of GH51 solution at pH 6.0 and 40 °C for 24 h.35 After 24 h of incubation, the enzyme reaction was stopped, and proteins were removed by adding 300 μL of Carrez A reagent and 300 μL of Carrez B reagent to 600 μL of a two times diluted AXOS sample. Next, the samples were microcentrifuged at 14 000 rpm for 15 min and filtered (UnifloRC filters, Whatman). The final qualitative analysis of the AXOS breakdown by the GH43 or GH51 enzymes was performed through HPAEC-PAD with a CarboPac PA200 column and a ternary gradient profile (Table 1). Statistics. A signal in the HPAEC-PAD chromatogram was considered as relevant when the peak height exceeded three times the standard deviation (SD) of the noise.36 The SD of the noise was determined in a region without peaks. Decreases in peak height after enzymatic incubation with GH43 or GH51 were confirmed by a t-test on independent samples of untreated and enzymatically treated AXOS. The decreases in peak height were considered significant when p < 0.05. The data were analyzed using SPSS version 13.0 for Windows (SPSS Inc., Chicago, IL). The run-to-run repeatability of the method was determined by repetitive injection (30 times) of the same homogeneous AXOS sample. All data were subjected to a statistical Grubbs’ outlier test37 at a significance level of α = 0.05, followed by averaging of the peak area, peak width at half height, peak retention time of nine different peaks (arabinose, xylose, five XOS peaks, and two AXOS peaks), and determination of the relative standard deviation (RSD) of all injections. Fermentation Experiment with a Bifidobacterial Strain. Fermentations with B. longum LMG 11047 were carried out in duplicate in Schott flasks containing 100 mL of C

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 2. Optimization of HPAEC-PAD chromatography fingerprint of pure AXOS solution: (A) fingerprint obtained with CarboPac PA100 column; (B) fingerprint obtained with CarboPac PA200 column. Gradient conditions were applied as described in Table 1. A = arabinose, X = xylose, X2 = xylobiose, X3 = xylotriose, X4 = xylotetraose, X5 = xylopentaose, and X6 = xylohexaose. Double-headed arrow represents the scale of the PAD response.



MCB. The flasks were sterilized in an autoclave at 121 °C and 1.1 bar overpressure for 20 min. After autoclaving, they were put immediately under anaerobic conditions in a modular atmosphere-controlled system (MG anaerobic workstation; Don Whitley Scientific, West Yorkshire, UK) that was continuously sparged with a mixture of 80% nitrogen, 10% carbon dioxide, and 10% hydrogen (Air Liquide, Paris, France). Afterward, AXOS were added as a sterile aqueous solution of 5 g L−1 to the sterile medium aseptically. The AXOS solution was not autoclaved because of the marked increase in monosaccharides (in particular arabinose) and XOS concentrations during heat treatment. Therefore, the AXOS solutions were filtersterilized using a polycarbonate (hydrophilic) highperformance MediaKap-50 Plus hollow fiber membrane filter (0.2 μm pores; Microgon, Rancho Dominguez, CA). There was no partial loss of the sample during filtration, as the HPAECPAD profiles were identical to the reference AXOS solution profiles. The inoculum for the 100 mL fermentations was prepared through two subcultures under anaerobic conditions. The bifidobacterial strain was first inoculated into 10 mL of reinforced clostridial medium (RCM; Oxoid) and incubated anaerobically at 37 °C for 12 h. Next, 5% (v v−1) of the first subculture was grown in 10 mL of MCB (pH 6.3), containing 1 g L−1 arabinose and 4 g L−1 xylose as the added energy sources. After a final anaerobic incubation of 12 h at 37 °C, 5% (v v−1) was transferred into 100 mL of MCB containing 5 g L−1 AXOS. The fermentations were sampled at different time points (0, 4, 8, 12, 24, and 48 h). Sample preparation for HPAEC-PAD analysis involved microcentrifugation (14 000 rpm for 15 min) to remove the cells, followed by deproteinization as described above.

RESULTS AND DISCUSSION Separation of Arabinose, Xylose, XOS, and AXOS. As the present work aimed at developing a single technique for structural characterization of arabinose, xylose, XOS, and AXOS in a solution of AXOS, the HPAEC parameters, including the ternary gradient of the mobile phase and the stationary phase, were optimized as follows. As a high-resolution separation of the complex AXOS was desirable (Figure 2), strong alkaline eluent conditions were used (≥100 mM [OH−]).38 The composition of the mobile phase has an important influence on the selectivity and retention time of the oligosaccharides to be separated, since it affects their oxyanion formation and simultaneous binding interaction with the stationary phase.28 As separation of complex mixtures of carbohydrates often requires gradient elution chromatography, a NaOH- and acetate-segmented linear gradient was applied.39−42 To get a good simultaneous separation of arabinose and xylose, mild starting conditions (90% A and 10% B during 10.0 min) were applied to separate them not only from each other but also from the monosaccharides present in the complex MCB. To obtain a high resolution of the AXOS peaks, a slowly increasing NaCH3COOH gradient, with a positive slope of 3.2 mM NaCH3COOH min−1 between 15.0 and 55.0 min, was applied. The baseline was stable as a consequence of the NaOH concentration that was kept constant during the NaCH3COOH gradient. To compare the separation performance of the CarboPac PA100 and PA200 columns with different dimensions under the same conditions, the flow rate and ternary gradient were adjusted to maintain the gradient volume constant, as the volume of the column is proportional to (i.d./2)2 . The available standards were used to identify the peaks correspondD

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

with waveforms A and C compared with waveform B. Indeed, for waveform A, a baseline drift effect (+7 nC after 65.0 min compared with waveform B) and a dip in the baseline around 25.0 min, which did not always appear, were detected, while for waveform C, a slightly higher baseline drift was obtained (+10 nC after 65.0 min compared with waveform B). A similar baseline dip has been found for separations with a CarboPac PA1 column in combination with waveform A, due to the reduction of dissolved oxygen at the gold electrode.43,44 Waveforms B and C differ from the waveform A in the use of a positive potential to clean the gold electrode in the case of a triple-potential waveform compared with the negative potential that is used to clean the gold electrode in the case of a quadruple-potential waveform.19,43,44 The positive cleaning potential causes electrode wearing and a gradual decrease in carbohydrate response. As this method will be used to study changes in peak heights and/or peak areas of the carbohydrates as a function of time during laboratory fermentation experiments with colon bacteria, a long-term repeatability is needed, and thus, the quadruple-potential waveform was chosen for further analysis. Qualitative Identification Method for Uncharacterized AXOS. The elution order of branched oligosaccharides in the HPAEC-PAD chromatograms cannot be predicted. This is because their retention times depend on the relative pKa values of the hydroxyl groups of the individual monosaccharides, their glycosidic linkages, the molecular size of the oligomers, and the accessibility of the oxyanions to the stationary phase.28,45,46 To identify the branched oligosaccharide peaks in the chromatograms, complex and time-consuming spectroscopic techniques or standards for all possible oligosaccharides are needed.28 During the present study, the complex branched oligosaccharides, eluting between 30.0 and 80.0 min under the conditions tested, were identified using a new method based on enzymaticbased reference degradation chromatograms generated with arabinofuranosidases (GH43 and GH51) with known specificities acting on AXOS. The degradation profile of AXOS by GH51, as compared with a pure AXOS solution, led to the statistically significant (independent samples t-test; p < 0.05) identification of 11 AXOS peaks that contained at least 1 monosubstituted Xylp residue (A2/3Xn; Figure 4B). The peaks could be clustered into two separate groups, further referred to as the early monosubstituted AXOS (A2/3,earlyXn) and the late monosubstituted AXOS (A2/3,lateXn). The A2/3,earlyXn group consisted of five peaks with retention times of approximately 31.0, 32.4, 33.8, 35.7, and 37.5 min. The A2/3,lateXn group was characterized by AXOS peaks with retention times of approximately 44.0, 45.8, 47.5, 51.4, 52.3, and 63.8 min. Also, a statistically significant (independent samples t-test; p < 0.05) increase of eight intermediate AXOS peaks (AinterXn) could be noticed with retention times between 39.2 and 43.6 min. Next, the enzymatic reaction of a pure AXOS solution with GH43 led to the statistically significant (independent samples t-test; p < 0.05) identification of four AXOS peaks containing at least one doubly substituted Xylp residue (A2+3Xn). The retention times of these peaks were 36.4, 37.9, 39.4, and 40.8 min (Figure 4D). Finally, two of the AinterXn peaks corresponded with the A2+3Xn peaks (39.4 and 40.8 min) that were identified based on the degradation of AXOS by arabinofuranosidase GH43. These two independent enzymatic incubations proved to be sufficient for the identification of a large part of the AXOS peaks between 30.0 and 65.0 min of elution. Peaks that could not be identified

ing with arabinose, xylose, xylobiose, xylotriose, xylotetraose, xylopentaose, and xylohexaose (Figure 2). Arabinose, xylose, XOS, and AXOS were well separated in 90.0 min using the CarboPac PA200 column (Figure 2B) in contrast with the CarboPac PA100 column, in particular with respect to the AXOS (Figure 2A), under the conditions tested. The elution order of the monosaccharides corresponded with their pKa values, as arabinose (pKa = 12.43) eluted before xylose (pKa = 12.29) (Figure 2). When the monosaccharides were eluted, the linear oligosaccharides xylobiose up to xylohexaose eluted according to their DP (Figure 2). The arabinose, xylose, xylobiose, xylotriose, xylotetraose, xylopentaose, and xylohexaose peaks eluted 0.5, 2.1, 11.4, 14.6, 12.3, 11.7, and 12.2 min earlier, respectively, from the CarboPac PA200 column (Figure 2B) compared with the same peaks from the CarboPac PA100 column (Figure 2A). This can be explained by the lower anionexchange capacity of the CarboPac PA200 column (35 μequiv column−1) compared to the CarboPac PA100 column (90 μequiv column−1). Moreover, the higher efficiency of the CarboPac PA200 profile allowed the tentative identification of 10 additional oligosaccharides, eluting between 30.0 and 80.0 min (Figure 2B), under the conditions tested. Owing to the smaller particle size, an increase in resolution was obtained (Rs ∼ √N, with N defined as the number of theoretical plates that increases with decreasing particle size), which resulted in lower detection limits.41 Optimization of Settings for Pulsed Amperometric Detection. To find out whether the sensitivity of the method for the complex AXOS could be improved by changing the waveform of the PAD detector under the conditions tested, one quadruple-potential waveform (A) and two different triplepotential waveforms (B and C) were compared using the CarboPac PA200 column (Figure 3). Waveforms A−C showed the same sensitivity for the AXOS peaks, as the peaks did not change in height or width. However, minor deviations occurred

Figure 3. Optimization of the PAD settings: (red line) waveform A, (black line) waveform B, and (dotted line) waveform C. An enlargement of the complex carbohydrates of a pure AXOS solution, eluting between 20.0 and 65.0 min, is shown. Similar separation conditions were applied as described in Figure 2B. X4 = xylotetraose, X5 = xylopentaose, and X6 = xylohexaose. Double-headed arrow represents the scale of the PAD response. E

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 4. Identification of AXOS peaks by analysis of an overlay of (black line) enzymatically cleaved AXOS solutions and (red line) its corresponding original AXOS solution. The enzymes used were (A + B) GH51 and (C + D) GH43. Figures B and D are enlargements of the AXOS chromatograms between 30.0 and 65.0 min. Similar separation conditions were applied as described in Figure 2B with waveform A. A = arabinose, X = xylose, X2 = xylobiose, X3 = xylotriose, X4 = xylotetraose, X5 = xylopentaose, X6 = xylohexaose, A2+3Xn = AXOS peak containing at least one doubly substituted Xylp residue, and A2/3Xn = AXOS peak containing at least one monosubstituted Xylp residue. Double-headed arrow represents the scale of the PAD response.

using this method may represent AXOS with arabinose residues that were esterified with ferulic acid or may present longer XOS (>X6) for which no commercial standards were available. Apart from substituent linkage-type information, the GH43 and GH51 incubations revealed the DP of the AXOS-derived XOS backbones (Figure 4A,C). Xylobiose and xylotriose peak areas hardly changed upon enzymatic hydrolysis with GH51. The peak areas of the xylotetraose, xylopentaose, and xylohexaose peaks doubled, tripled, and quadrupled, respectively, after GH51 hydrolysis. Except for an increase of the arabinose peak area, the XOS peaks did not change after GH43 hydrolysis. This indicates that the AXOS with disubstituted xylose residues were possibly less present in the AXOS substrate or contained monosubstituted xylose residues. Run-to-Run Repeatability. The average peak area, average peak width at half height, and average peak retention time with corresponding RSD were determined for 9 peaks eluting across the chromatogram of 30 repetitive injections of a pure AXOS solution into a CarboPac PA200 column under the conditions tested (Table 2). One statistical outlier was excluded to determine the average peak area and RSD, no outliers were excluded to determine the average peak width at half height and RSD, and one outlier was excluded to calculate the average retention time and RSD. The data showed that this method provided excellent run-to-run repeatability of the chromatograms with respect to peak area (RSD ≤ 3.1%), peak width at

Table 2. Repeatability of the HPAEC-PAD Method Was Determined for 9 Peaks by 30 Repetitive Injections of a Pure AXOS Solution into a CarboPac PA200 Columna peak area

peak width at half height

retention time

carbohydrate

averageb (nC·min)

RSDb (%)

average (min)

RSD (%)

averageb (min)

RSDb (%)

arabinose xylose xylobiose xylotriose xylotetraose xylopentaose xylohexaose AXOS 1 AXOS 2

35.46 35.52 289.53 155.67 51.60 17.49 6.50 69.05 15.26

2.21 1.75 3.09 1.72 1.42 1.54 2.83 2.11 1.85

0.15 0.15 0.56 0.23 0.21 0.15 0.14 0.20 0.22

1.71 1.13 2.39 1.73 1.33 2.78 1.20 1.12 0.95

3.98 5.51 12.09 16.50 22.79 26.63 29.06 32.60 36.25

0.28 0.43 0.82 0.36 0.36 0.19 0.16 0.20 0.19

a

Gradient conditions were applied as described in Table 1. The average peak area, average peak width at half height, and average peak retention time with the corresponding RSDs were calculated. A Grubb’s test was used to search for and to eliminate possible outliers. b One result eliminated as an outlier using the Grubb’s test.

half height (RSD ≤ 2.8%), and retention time (RSD ≤ 0.8%).41 This good repeatability was essential for the qualitative identification of uncharacterized AXOS peaks, whereas the information of the enzymatic-based reference degradation F

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 5. AXOS degradation by B. longum LMG 11047 in medium for colon bacteria supplemented with 5 g L−1 AXOS as a function of time (0, 4, 8, 12, 24, and 48 h). The chromatograms represent (A) monosaccharide and XOS consumption and (B) enlargements of the AXOS consumption profiles between 30.0 and 65.0 min. Similar separation conditions were applied as described in Figure 2B with waveform A. Peaks were named as in Figure 4. Double-headed arrow represents the scale of the PAD response.

chromatograms was used to identify the peaks of the chromatograms representing the mechanistic breakdown of AXOS during fermentation. Fermentation Experiment with a Bifidobacterial Strain. Despite the increasing amount of studies indicating an AXOS-related bifidogenic effect,8,11,47 little is known about the mechanistic AXOS degradation strategies of bifidobacterial species. The newly developed method for qualitative fingerprinting of prebiotic AXOS of the present paper was applied to study their mechanistic breakdown during fermentation by a selected member of the colon microbiota, namely, B. longum LMG 11047. This bifidobacterial strain represents a cluster of bifidobacteria that shares both high fructose consumption and oligofructose degradation rates and is able to degrade short fractions of inulin in a very competitive way compared to other inulin-type fructan degraders.33,48 As MCB is a complex fermentation medium, it led to an overlap of a few peaks in the HPAEC-PAD chromatograms under the conditions tested, especially an overlap of the AinterXn peaks with retention times ranging from 39.0 to 43.0 min and an overlap of A2/3,lateXn peaks between 50.0 and 55.0 min (Figure 5). B. longum LMG 11047 was able to break down the AXOS in different steps using a combination of two mechanisms, according to the data obtained with the method outlined above. First, it cleaved both monosubstituted (A2/3,earlyXn and A2/3,lateXn) and disubstituted arabinose side chains, which resulted in an increase of the AXOS-derived backbones (xylobiose, xylotriose, xylotetraose, xylopentaose, and xylohexaose) in the medium after 4 h of fermentation, indicating an extracellular breakdown. Afterward, when the arabinose side chains were consumed, B. longum LMG 11047 consumed the AXOS-derived backbones together with the XOS initially present in the AXOS substrate, albeit only up to xylotetraose, without xylose accumulation in the medium, indicating an intracellular breakdown. The AXOS-derived xylopentaose and xylohexaose were left unaffected. This suggests a combination of the two mechanisms that have been proposed for the degradation of fructo-oligosaccharides by bifidobacteria, namely, transport of small oligosaccharides into the cell and

subsequent intracellular degradation as well as external degradation by membrane-bound and/or extracellular enzymes followed by uptake of the monosaccharides.32,47,49 AXOS with doubly substituted xylose units were equally efficiently fermented as monosubstituted AXOS. This is in contrast with the results of Pastell et al.50 that showed that doubly substituted xylose residues could not be fermented by the bifidobacterial strains tested. However, several genes encoding putative arabinofuranosidases are present in the genome of B. longum NCC2705; for example, a gene (BL1544) was found showing amino acid similarity with an AX arabinofuranosidase of Bifidobacterium adolescentis DSM 20083 that hydrolyzes Araf residues from doubly substituted Xylp residues at position C(O)-3.51,52 These results suggest that the new method for HPAEC-PAD fingerprinting of the present study, based on comparison with a reference degradation chromatogram generated with enzymes with known specificities, is an interesting tool to study the fermentation capabilities of AXOS in complex growth media by colon bacteria. Until now, studies of the degradation of AXOS through fermentation were restricted to the monitoring of bacterial growth, pH, and SCFA production; also, fermentations of self-made short-chain AXOS standards were performed, which only reveal a minor aspect of the true fermentation capacity of the microorganisms under study.50,53,54 Furthermore, Crittenden et al.,53 Van Laere et al.54 and Pastell et al.50 only assessed changes in AXOS consumption and/or structures after fermentation and may therefore have missed some interesting kinetic changes during fermentation. The establishment of a simple and fast qualitative analytical method to follow the degradation kinetics of AXOS during fermentation is of great importance. The application of this method will allow the study of mechanistic variations in AXOS degradation by colon bacteria to understand how prebiotics can exert their health-beneficial properties. Although the developed method may be restricted to the eluents used and the elution program, it forms a powerful basis for the chromatographic separation of various AXOS substrates. G

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry



Article

(10) Grootaert, C.; Delcour, J. A.; Courtin, C. M.; Broekaert, W. F.; Verstraete, W.; Van de Wiele, T. Trends Food Sci. Technol. 2007, 18, 64−71. (11) Broekaert, W. F.; Courtin, C. M.; Verbeke, K.; Van de Wiele, T.; Verstraete, W.; Delcour, J. A. Crit. Rev. Food Sci. Nutr. 2011, 51, 178− 194. (12) Izydorczyk, M. S.; Biliaderis, C. G. Carbohydr. Polym. 1995, 28, 33−48. (13) Schooneveld-Bergmans, M. E. F.; Beldman, G.; Voragen, A. G. J. J. Cereal Sci. 1999, 29, 63−75. (14) Lequart, C.; Nuzillard, J. M.; Kurek, B.; Debeire, P. Carbohydr. Res. 1999, 319, 102−111. (15) Broekaert, W. F.; Courtin, C.; Delcour, J. PCT International Publication Number. WO 2009/117790 A2, October 1, 2009. (16) Sanz, M. L.; Ruiz-Matute, A. I.; Corzo, N.; Martínez-Castro, I. Prebiotics and Probiotics Science and Technology; Charalampopoulos, D., Rastall, R. A., Eds.; Springer: London, 2009; Vol. 1, pp 465−534. (17) Dell, A.; Reason, A. J. Curr. Opin. Biotechnol. 1993, 4, 52−56. (18) Williams, D. H.; Fleming, I. Spectroscopic Methods in Organic Chemistry; McGraw-Hill Book Company: Berkshire, UK, 1995; pp 1− 329. (19) LaCourse, W. R.; Johnson, D. C. Anal. Chem. 1993, 65, 50−55. (20) Swennen, K.; Courtin, C. M.; Lindemans, G. C. J. E.; Delcour, J. A. J. Sci. Food Agric. 2006, 86, 1722−1731. (21) Broberg, A.; Thomsen, K. K.; Duus, J. O. Carbohydr. Res. 2000, 328, 375−382. (22) Rantanen, H.; Virkki, L.; Tuomainen, P.; Kabel, M.; Schols, H.; Tenkanen, M. Carbohydr. Polym. 2007, 68, 350−359. (23) Van Laere, K. M. J.; Voragen, C. H. L.; Kroef, T.; Van den Broek, L. A. M.; Beldman, G.; Voragen, A. G. J. Appl. Microbiol. Biotechnol. 1999, 51, 606−613. (24) Vanderhoeven, R. A. M.; Niessen, W. M. A.; Schols, H. A.; Bruggink, C.; Voragen, A. G. J.; Vandergreef, J. J. Chromatogr. 1992, 627, 63−73. (25) Okatch, H.; Torto, N. Afr. J. Biotechnol. 2003, 2, 636−644. (26) Houben, R.; de Ruijter, C. F.; Brunt, K. J. Cereal Sci. 1997, 26, 37−46. (27) Lebet, V.; Arrigoni, E.; Amado, R. Food Res. Technol. 1997, 205, 257−261. (28) Lee, Y. C. Anal. Biochem. 1990, 189, 151−162. (29) Lee, Y. C. J. Chromatogr., A 1996, 720, 137−149. (30) Cataldi, T. R. I.; Campa, C.; Benedetto, G. E. Anal. Bioanal. Chem. 2000, 368, 739−758. (31) Pollet, A.; Beliën, T.; Fierens, K.; Delcour, J. A.; Courtin, C. M. Enzyme Microb. Technol. 2009, 44, 189−195. (32) Van der Meulen, R.; Makras, L.; Verbrugghe, K.; Adriany, T.; De Vuyst, L. Appl. Environ. Microbiol. 2006, 72, 1006−1012. (33) Falony, G.; Lazidou, K.; Verschaeren, A.; Weckx, S.; Maes, D.; De Vuyst, L. Appl. Environ. Microbiol. 2009, 75, 454−461. (34) Weitzhandler, M.; Pohl, C.; Rohrer, J.; Narayanan, L.; Slingsby, R.; Avdalovic, N. Anal. Biochem. 1996, 241, 128−134. (35) Sørensen, H. R.; Jørgensen, C. T.; Hansen, C. H.; Jørgensen, C. I.; Pedersen, S.; Meyer, A. S. Appl. Microbiol. Biotechnol. 2006, 73, 850−861. (36) Felinger, A. Data Analysis and Signal Processing in Chromatography; Elsevier: The Netherlands, 1998; pp 125−141. (37) Grubbs, F. E. Technometrics 1969, 11, 1−21. (38) Dionex. Application Note 67; Dionex: Sunnyvale, CA, 2003. (39) Basa, L. J.; Spellman, M. W. J. Chromatogr. 1990, 499, 205−220. (40) Ammeraal, R. N.; Delgado, G. A.; Tenbarge, F. L.; Robert, B. F. Carbohydr. Res. 1991, 215, 179−192. (41) Snyder, L. R.; Kirkland, J. J.; Glajch, J. L. Practical HPLC Method Development; Wiley Interscience: New York, 1997; p 542. (42) Pouvreau, L.; Joosten, R.; Hinz, S. W. A.; Gruppen, H.; Schols, H. A. Enzyme Microb. Technol. 2011, 48, 397−403. (43) Rocklin, R. D.; Clarke, A. P.; Weitzhandler, M. Anal. Chem. 1998, 70, 1496−1501. (44) Dionex. Technical Note 21; Dionex: Sunnyvale, CA, 1998.

CONCLUSION A new method for the qualitative simultaneous analysis of arabinose, xylose, XOS, and AXOS was developed. The advantages of this method were its simplicity, speed, ease of sample preparation, high resolution and high repeatability, as well as the structural information that is obtained concomitantly. The method was successful for the analysis of AXOS in samples from complex fermentation media and could provide interesting information on the fermentation capacities of colon bacteria. Also, it could help to understand the beneficial properties of this prebiotic compound. Although this new method depends on the availability of appropriate enzymes, the emerging interest in various fields of bioconversion and fermentation of complex oligo- and polysaccharides ensures that more enzymes will become commercially available and that other complex heterooligo- and polysaccharides will enable studies with the same approach.



AUTHOR INFORMATION

Corresponding Author

*Phone: +32 2 6293245; fax: +32 2 6292720; e-mail: ldvuyst@ vub.ac.be. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS



REFERENCES

The authors acknowledge their financial support of the Research Council of the Vrije Universiteit Brussel and the Research Foundation-Flanders (FWO-Vlaanderen). A.R. is the recipient of a Ph.D. fellowship of the FWO-Vlaanderen. We thank Prof. Dr. ir. Christophe Courtin (KU Leuven) for providing us with AXOS. We also thank Filip Timmermans (Thermo Scientific) for kindly providing the CarboPac PA200 column. Eva Tyteca and Axel Vaast are gratefully acknowledged for their technical advice.

(1) Bäckhed, F.; Ley, R. E.; Sonnenburg, J. L.; Peterson, D. A.; Gordon, J. I. Science 2005, 307, 1915−1920. (2) Falony, G.; De Vuyst, L. Prebiotics and Probiotics Science and Technology; Charalampopoulos, D., Rastall, R. A., Eds.; Springer: London, 2009; Vol. 2, pp 640−679. (3) Gibson, G. R.; Probert, H. M.; Van Loo, J.; Rastall, R. A.; Roberfroid, M. B. Nutr. Res. Rev. 2004, 17, 259−275. (4) Van Craeyveld, V.; Swennen, K.; Dornez, E.; Van de Wiele, T.; Marzorati, M.; Verstraete, W.; Delaedt, Y.; Onagbesan, O.; Decuypere, E.; Buyse, J.; De Ketelaere, B.; Broekaert, W. F.; Delcour, J. A.; Courtin, C. M. J. Nutr. 2008, 138, 2348−2355. (5) Grootaert, C.; Van den Abbeele, P.; Marzorati, M.; Broekaert, W. F.; Courtin, C. M.; Delcour, J. A.; Verstraete, W.; Van de Wiele, T. FEMS Microbiol. Ecol. 2009, 69, 231−242. (6) Sanchez, J. I.; Marzorati, M.; Grootaert, C.; Baran, M.; Van Craeyveld, V.; Courtin, C. M.; Broekaert, W. F.; Delcour, J. A.; Verstraete, W.; Van de Wiele, T. Microb. Biotechnol. 2009, 2, 101−113. (7) Damen, B.; Verspreet, J.; Pollet, A.; Broekaert, W. F.; Delcour, J. A.; Courtin, C. M. Mol. Nutr. Food Res. 2011, 55, 1862−1874. (8) Neyrinck, A. M.; Possemiers, S.; Druart, C.; van de Wiele, T.; De Backer, F.; Cani, P. D.; Larondelle, Y.; Delzenne, N. M. PLoS One 2011, 6, 1−12. (9) Van den Abbeele, P.; Gerard, P.; Rabot, S.; Bruneau, A.; El Aidy, S.; Derrien, M.; Kleerebezem, M.; Zoetendal, E. G.; Smidt, H.; Verstraete, W.; Van de Wiele, T.; Possemiers, S. Environ. Microbiol. 2011, 13, 2667−2680. H

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(45) Hardy, M. R.; Townsend, R. R. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 3289−3293. (46) Koizumi, K.; Kubota, Y.; Tanimoto, T.; Okada, Y. J. Chromatogr. 1989, 464, 365−373. (47) Cloetens, L.; Broekaert, W. F.; Delaedt, Y.; Ollevier, F.; Courtin, C. M.; Delcour, J. A.; Rutgeerts, P.; Verbeke, K. Br. J. Nutr. 2010, 103, 703−713. (48) Falony, G.; Calmeyn, T.; Leroy, F.; De Vuyst, L. Appl. Environ. Microbiol. 2009b, 75, 2312−2319. (49) Amaretti, A.; Tamburini, E.; Bernardi, T.; Pompei, A.; Zanoni, S.; Vaccari, G.; Matteuzzi, D.; Rossi, M. Appl. Microbiol. Biotechnol. 2006, 73, 654−662. (50) Pastell, H.; Westermann, P.; Meyer, A. S.; Tuomainen, P.; Tenkanen, M. J. Agric. Food Chem. 2009, 57, 8598−8606. (51) Schell, M. A.; Karmirantzou, M.; Snel, B.; Vilanova, D.; Berger, B.; Pessi, G.; Zwahlen, M. C.; Desiere, F.; Bork, P.; Delley, M.; Pridmore, R. D.; Arigoni, F. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 14422−14427. (52) van den Broek, L. A. M.; Lloyd, R. M.; Beldman, G.; Verdoes, J. C.; McCleary, B. V.; Voragen, A. G. J. Appl. Environ. Microbiol. 2005, 67, 641−647. (53) Crittenden, R.; Karppinen, S.; Ojanen, S.; Tenkanen, M.; Fagerstrom, R.; Matto, J.; Saarela, M.; Mattila-Sandholm, T.; Poutanen, K. J. Sci. Food Agric. 2002, 82, 1−9. (54) Van Laere, K. M. J.; Hartemink, R.; Bosveld, M.; Schols, H. A.; Voragen, A. G. J. J. Agric. Food Chem. 2000, 48, 1644−1652.

I

dx.doi.org/10.1021/ac400187f | Anal. Chem. XXXX, XXX, XXX−XXX