at SciVerse ScienceDirect

Journal of Cereal Science 56 (2012) 134e141

Contents lists available

Journal of Cereal Science

journal homepage: www.elsevier .com/locate/ jcs

Spectroscopic analysis of diversity in the spatial distribution of arabinoxylanstructures in endosperm cell walls of cereal species in the HEALTHGRAINdiversity collection

G.A. Toole a,*, G. Le Gall a, I.J. Colquhoun a, S. Drea d, M. Opanowicz e, Z. Bed}o c, P.R. Shewry b, E.N.C. Mills a

a Institute of Food Research, Norwich Research Park, Norwich NR4 7UA, UKbRothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UKcAgricultural Research Institute of the Hungarian Academy of Sciences, Brunszvik u. 2, Martonvásár HU-2462, HungarydDepartment of Biology, University of Leicester, Leicester LE1 7RH, UKeDepartment of Cell and Developmental Biology, John Innes Centre, NR4 7UH, UK

a r t i c l e i n f o

Article history:Received 7 April 2011Received in revised form28 October 2011Accepted 17 February 2012

Keywords:ArabinoxylanCerealEndosperm cell wallsFT-IRHEALTHGRAIN1H NMR

Abbreviations: Araf, arabinofuranosyl units; AX,se:xylose ratio; EtOH, ethanol; FPA, focal plane arinfra-red; HS-AX, highly substituted AX; LS-AX, low sAX; WE-AX, water-extractable AX; NMR, nuclear macipal Components Analysis; TSP, sodium 3-(trimethxylopyroanosyl residues.* Corresponding author. Tel.: þ44 (0)1603 255000;

E-mail address: geraldine.toole@ifr.ac.uk (G.A. Too

0733-5210 � 2012 Elsevier Ltd.doi:10.1016/j.jcs.2012.02.016

Open access under CC BY

a b s t r a c t

Two varieties each of spelt (Triticum aestivum var. spelta), durum wheat (Triticum turgidum var. durum),rye (Secale cereale), barley (Hordeum vulgare), oats (Avena sativa), einkorn (Triticum monococcum var.monococcum) and emmer (Triticum turgidum var. dicoccum) (all members of the Pooideae sub-family ofgrasses) were selected according to variation in their contents of soluble and/or total arabinoxylan (AX)determined during the HEALTHGRAIN diversity screen, together with one genotype of the related“model” grass species Brachypodium distachyon. The spatial distribution of low substituted (LS-AX) andhighly substituted arabinoxylan (HS-AX) was determined using FT-IR spectroscopic mapping of trans-verse thin cross-sections consisting of cell walls only. Variation in cell wall AX composition was observedbetween the cereals, and compared with that observed for wheat (Triticum aestivum var. aestivum). Oneline of each cereal type was analysed in more detail using 1H NMR spectroscopy. The results of the twoanalyses were consistent, showing variation in the composition and structure of the endosperm cell wallAX that is consistent with the genetic relationships of the cereals studied.

� 2012 Elsevier Ltd. Open access under CC BY license.

1. Introduction

The composition of the cell walls of cereal grains differs fromthat of most other agricultural crops with the endosperm cell wallsconsisting mainly of arabinoxylan (AX) and (1,3;1,4)-b-glucan, withlow levels of cellulose, glucomannan and structural proteins(Fincher and Stone, 1986). The relative proportions of (1,3;1,4)-b-glucan and AX vary significantly between cereal species, with theendosperm cell walls of rye andwheat being rich in AX, while thoseof barley and oats have high levels of (1,3;1,4)-b-glucan (Fincherand Stone, 1986). In addition to determining the physical proper-ties and structural characteristics of the endosperm cell walls, these

arabinoxylan; A/X, arabino-ray; FT-IR, Fourier-transformubstituted AX; TOT-AX, totalgnetic resonance; PCA, Prin-ylsilyl)-propionate-d4; Xylp,

fax: þ44 (0)1603507723.le).

license.

non-starch polysaccharides make an important contribution to theintake of dietary fibre and its associated health benefits.

In commonwith many other polysaccharides, cereal AX exhibitsa high degree of endogenous heterogeneity. Gruppen et al. (1993)suggested that AX contains highly substituted regions, mostly con-sisting of tetrameric repeating units of an un-substituted anda double arabinofuranosylated (Araf) xylose (Xylp) residue. Theseunits are interspersed with less substituted regions, which includesub-regions containingup to sevencontiguousun-substitutedxyloseresidues. The highly substituted regions are enriched in bothO-2 andO-3 disubstituted xylose residues. Wheat AX contains very smallamounts ofO-2monosubstituted xylose residues,whereas barley AXare characterised by a larger amount of O-2 monosubstitution,however, for both O-2 monosubstituted xylose residues are absentfrom the less substituted regions. Variation in the arabinose:xylose(A:X) ratio between different AX polymers therefore results fromvariation in the relative proportions of these regions, as well as fromvariation in the composition of the less substituted regions.

(1,3;1,4)-b-glucan consists of linear unbranched polysaccharidesof linked b-(1,3)- and b-(1,4)-D-glucopyranose units, which form

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141 135

‘worm-like’ cylindrical molecules containing up to 2500 glucoseresidues. The 1,3-linkages break up the uniform structure of the b-D-glucan molecule and make it soluble and flexible. b-glucan iscomposed primarily of b-1,3 linked cellotriosyl and cellotetraosylunits, with regions that are more cellulose-like in character withmore than four consecutive b-1,4 linked glucose units (Wood,1993).The molar ratio of cellotriose to cellotetraose units (DP3:DP4) istypically 1.5e2.3 for oat b-glucan (Miller and Fulcher, 1995).

In addition to variation in gross composition, it is clear that thereis significant spatial heterogeneity inAX composition in cereal grains(Barron et al., 2005; Toole et al., 2007). FT-IR microspectroscopyshowed that the structure of the AX in wheat endosperm cell wallschanged during the later stages of grain development from a highlysubstituted form (HS-AX) to a low substituted form (LS-AX), thetransition being initiated in the outer (prismatic) endosperm cellsand then spreading into the central (round) endosperm cells (Tooleet al., 2010, 2007). The rate of the transition was also faster inmaterial grown under hot/dry growing conditions and differedbetween cultivars. Toole et al. (2009, 2010) also used Raman micro-spectroscopy and 1H NMR to provide qualitative information on theextent of arabinose substitution of the AX in intact cell walls.

We have been studying the variation in phytochemicals anddietary fibre available to cereal breederswithin the context of the EUFP6 HEALTHGRAIN project, which sought to improve thewell-being,and reduce the risk ofmetabolic diseases, in European consumers byincreasing their intake of protective components present in whole-grain (Poutanen et al., 2008). This programme included a “diversityscreen” in which 150 bread wheat lines and 50 lines of other smallgrain cereals (spelt, durum wheat, Triticum monococcum (einkorn),Triticum dicoccum (emmer), oats, rye and barley) (Rakszegi et al.,2008) were grown on a single site in 2005 and analysed for a rangeof dietary fibre components (Andersson et al., 2008; Gebruers et al.,2008; Nystrom et al., 2008; Shewry et al., 2008).

In the present study, we use FT-IR spectroscopic mapping and1H NMR analysis to compare the composition, structure and spatialdistribution of AX in the endosperm cell walls of small grain cereals(barley, oats rye and wheat species), utilising material and infor-mation from the HEALTHGRAIN project, and compare them withlines of bread wheat (Toole et al., 2011) and the wild grass speciesBrachypodium distachyon. The latter has been proposed as a modelsystem for temperate cereals because of its ease of cultivation andtransformation, small genome size and rapid life cycle, and is thefirst member of the Pooideae sub-family to have a fully sequencedgenome (Vogel et al., 2010).

2. Materials and methods

2.1. Growth of cereal lines

Ten lines each of durum wheat (Triticum turgidum var. durum),barley (Hordeum vulgare) and rye (Secale cereale), together with fivecultivars each of oats (Avena sativa), spelt (Triticum aestivum var.spelta), einkorn (Triticum monococcum var. monococcum) andemmer (Triticum turgidum var. dicoccum) were selected from theHEALTHGRAIN diversity collection described by Ward et al. (2008).All lines were grown at the Agricultural Research Institute of theHungarian Academy of Sciences, Martonvásár, Hungary (latitude,47�210N; longitude, 18�490E; altitude, 150 m) in 2004/2005 in twoplots as described by Rakszegi et al. (2008). One genotype (Bd21) ofB. distachyon was grown at the John Innes Centre (UK).

2.2. Preparation of cell wall-only thin sections

Ten transverse cell wall-only thin sections were cut from thecentral region of a grain from each of the selected lines. The

sections were prepared using a Vibratome (Intracel, Royston, UK),and their cell contents washed out according to the methoddescribed by Barron et al. (2005) and Toole et al. (2007).

Images were collected for ten sections from each grain, usinglight microscopy as described by Toole et al. (2007). From theseimages, the best sections (those with intact cell wall networks withthe least remaining starch and protein) were selected and allowedto dry at room temperature onto a barium fluoride disc (13 mmdiameter � 1 mm thick, Crystran Ltd., Dorset, UK), for FT-IR spec-troscopic imaging.

2.3. FT-IR spectroscopic imaging

Four regions of each transverse section were selected to bemapped spectroscopically using FT-IR imaging as follows: region a)the top of a section, between the pericarp and the cavity, b) theupper corner of a section, between the pericarp and the cavity, c)the side of a section, from the pericarp at the side to the centre ofthe cheek region, and d) the centre of a section, from the cavity tothe centre of the cheek. The selected regions of the endosperm cellwall network in each transverse section were imaged using a Dig-ilab Stingray imaging spectrometer (Digilab, Cambridge, MA, US)according to the method described in Toole et al. (2007).

2.4. FT-IR data analysis

All data analysis was performed using the image analysis soft-ware ENVI 4.0 (Research Systems Inc, Boulder, CO, USA). In order toavoid effects of variation in sample thickness, and therefore spec-tral intensity, all spectra were baseline corrected and normalised.This was done by converting all spectra to transmittance units,correcting the baseline to 1 (using the continuum-removal func-tion), returning the spectra to absorbance units, and finallyapplying an area-normalisation to themaximumpeak at 1041 cm�1

(dividing by the height of the band then multiplying by a factor).Because of the size of the grain it was not possible to scan

a whole section using FT-IR imaging. Images were thereforecollected for four areas, 640 � 640 mm, selected to represent all celltypes within the endosperm (subaleurone, prismatic and centralround cells). Using ENVI 4.0, an average spectrum was determinedfor each cell wall-only thin section, by combining all spectra fromall four regions within a section (i.e., the average of4 � 16,384 ¼ 65,536 spectra). Principal Components Analysis (PCA)was then applied to various combinations of the cereals analysed.

Because of the preponderance of AX in the endosperm cell wallsof wheat and rye (approximately 70% of the cell wall poly-saccharides), the features of these components dominate thespectra, with little detectable apparent contributions from othercell wall polysaccharides. Clear differences were identified in theheight of the shoulder at 1075 cm�1, which may be assigned to thelevel of substitution (Barron et al., 2005). In order to compare thespectral images from different thin sections, a classificationmethodwas devised by creating an arbitrary boundary between tworegions (Toole et al., 2007). This criterion was transferable betweenall data sets for all sections. A false-colour-coded image wasproduced from the individual spectra by assigning a colour to eachpixel, with each of the four regions in each section comprising16,384 pixels. If the shoulder at 1075 cm�1 was above 66% of theheight of the peak at 1041 cm�1 (representing HS-AX), then thepixel corresponding to the spectrum was colour-coded blue; if theshoulder was below 66% (representing LS-AX), it was colouredgreen. It was coloured black if the absorbance of the spectrum wastoo low or too noisy, which was indicative of holes in the cell wallsection, and was coloured white if the spectral features weresimilar to those of starch (rather than AX). The 66% ratiowas chosen

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141136

because it was half way between the two average shoulder heightsin a PCA plot produced for the spectral image of a typical endo-sperm cell wall section (Toole et al., 2007). It should be noted thatthe difference between the ‘blue’ and ‘green’ cell wall types issignified by a lowering of the spectral shoulder at 1075 cm�1, whichis indicative of a change in the substitution pattern of the AX, bylowering the level of substitution present in the polymer.

For the spectral images for thin sections of barley and oats,a different classification was used. Because the cell walls in barleyand oats consist of 70% (1,3;1,4)-b-glucan and only 20% AX, the AXshoulder at 1075 cm�1 is overlapped by the spectral peak for(1,3;1,4)-b-glucan at 1065 cm�1. An ENVI classification was devisedin order to visualise the (1,3;1,4)-b-glucan. Any holes in the cell wallnetwork were coloured black, any remaining starch was colouredwhite, AX was defined as any spectra having a shoulder at1075 cm�1, rather than a peak, and these pixels were coloured blue,(1,3;1,4)-b-glucan was divided into two classes: the 1st class wasdefined by the emergence of the peak at 1065 cm�1, and signifieda lower level of (1,3;1,4)-b-glucan in the cell wall network, thesepixels were coloured in purple, and the 2nd class was coloured red,and signified that the (1,3;1,4)-b-glucan peak was higher than themain AX peak at 1041 cm�1, denoting a large percentage of(1,3;1,4)-b-glucan in the cell wall network.

2.5. 1H NMR analysis

Endosperm cell wall network samples were prepared by handdissection from approximately 200 transverse cell wall-only thinsections, and allowed to dry. These were then mixed with 700 mL ofa solution of D2O containing TSP (sodium 3-(trimethylsilyl)-propionate-d4) as a reference, and stirred using a spatula at roomtemperature for w10 s. 550 mL of the whole mixture was thentransferred to a 5 mm NMR tube for analysis. 1H NMR spectra wererecorded at 70 �C on a 600 MHz Bruker Avance spectrometer(Bruker Biospin, Rheinstetten, Germany) according to the methoddescribed by Toole et al. (2009, 2010).

Whilst only about 25e35 % of the cereal endosperm AX areextractable in water (Mares and Stone, 1973), hydration of thesamples was sufficient for the signals from the arabinose side-chainand xylan backbone sugars to be observed at high temperature(70 �C) by high resolution 1H NMR. However, the low solubility andthe preparation conditions used meant that only semi-quantitativedata could be collected. Despite this shortcoming, the NMR spectragive a unique insight into the structure of intact AX in cerealendosperm cell walls.

The signal assignments were based on detailed 1H NMR data forAX oligosaccharides (Gruppen et al., 1992; Hoffman et al., 1992).

3. Results and discussion

Two varieties each of spelt (T. aestivum var. spelta), durumwheat(T. turgidum var. durum), rye (S. cereale), barley (H. vulgare), oats (A.sativa), einkorn (T. monococcum var. monococcum) and emmer (T.turgidum var. dicoccum) were selected from the HEALTHGRAIN linesgrown together in 2005, based on analyses of white flour fractionswhich showed either high or low contents of soluble/total AX(Gebruers et al., 2008). A single genotype of B. distachyon wasgrown in a glasshouse.

3.1. FT-IR microspectroscopy

In order to determine the structure and distribution of AX andother non-starch polysaccharides across the grain, FT-IR spectro-scopic mapping was carried out on thin cell wall-only sectionsprepared from mature grain. It was not possible to scan the whole

sections so FT-IR spectral images were determined for four selectedareas of the starchy endosperm, as described in the Materials andmethods section. These images are overlaid onto the visualimages in Fig. 1 and discussed in detail below.

The spectra for these four areas were initially integrated to giveaveraged spectra for each cereal (Fig. 2). A similar averaged spec-trumwas also calculated from the analyses of 50 bread wheat linesgrown in the same field trial and reported by Toole et al. (2011). Theaveraged spectrum for the wheat lines is dominated by featuresrelated to AX, with a major carbohydrate peak at 1047 cm�1 anda shoulder at 1075 cm�1, which varies in height according to thelevel of arabinose substitution (Fig. 2). The spectra for barley andoat endosperm cell walls are dominated by features related to(1,3;1,4)-b-glucan, namely a peak at 1065 cm�1, which is lower forbarley than for oats. The (1,3;1,4)-b-glucans of barley and oats aresimilar in structure but differ in that oat (1,3;1,4)-b-glucan hasa higher proportion of b-(1-4) linkages (0.7) than b-(1-3) linkages(0.3) (Wood and Beer, 1998) and is therefore less soluble thanbarley (1,3;1,4)-b-glucan. The molecular weight of oat (1,3;1,4)-b-glucan is also higher than that of barley (1,3;1,4)-b-glucan (Brennanand Cleary, 2005). These differences may account for the differ-ences seen in the FT-IR spectra reported here. The spectrum forBrachypodium is similar to those of barley and oats, indicatinga high proportion of (1,3;1,4)-b-glucan, but the shoulder at980 cm�1 is higher which may result from the presence of otherinsoluble cell wall polysaccharides such as cellulose and xylan(Guillon et al., 2011).

Comparisons of the spatial distributions of the different spectraltypes (Fig. 1) show that tetraploid durum and hexaploid breadwheats (Toole et al., 2011) had similar patterns of AX, with most ofthe cell walls comprising LS-AX, with only a small amount of HS-AXpresent near the central region of the grain (Fig. 1). By contrast, theimages for spelt (a variety of hexaploid bread wheat) showed thepresence of more HS-AX, with only a small amount of LS-AX in theouter regions of the grain. However, significant heterogeneity wasobserved between the different samples of spelt, with some havingmore HS-AX. The images for the thin sections of einkorn (diploidwheat) and emmer (tetraploid wheat) were very similar to thosefor durum wheat, showing a mixture of HS-AX and LS-AX, withmore HS-AX towards the centre of the grain and more LS-AXtowards the outer regions. On the other hand, the spectral imagesof rye grain were different to those of the other cereals, with moreHS-AX being present in the outer regions of the grain and smallamounts of LS-AX being present in the centre. It is known that theAX present in the endosperm cell wall of rye contains a lowerproportion of disubstituted xylose (i.e., HS-AX) and a higherproportion of monosubstituted (i.e., LS-AX) xylose residues thanthat in wheat (Höije et al., 2008; Rantanen et al., 2007), which mayaccount for the high proportion of HS-AX shown in the spectralimages. The structure of rye AX is similar to that of wheat AX, butdifferences in the abundances of the O-2 monosubstituted and O-2and O-3 disubstituted xylose residues have been reported (Vinkxand Delcour, 1996). The spectral images for the thin sections ofbarley and oats show that themajor cell wall polysaccharide in bothspecies is (1,3;1,4)-b-glucan (shown in red) which is consistentwith their known composition (Fincher and Stone, 1986).

The spectral images of Brachypodium showed that the endo-sperm cell walls were similar in composition to those of barley andoats, but indicated that they contained a lower proportion of(1,3;1,4)-b-glucan. However, Guillon et al. (2011) recently reportedthat the starchy endosperm cell walls of Brachypodium contained atleast 80% b-glucan, compared with about 75% in barley (Fincher,1975). The spectroscopic imaging may therefore have under-estimated the content of b-glucan in Brachypodium due to loss fromthe cell wall sections during preparation because of its solubility in

Fig. 1. Spectroscopic images overlaid onto the corresponding visible microscope images, for transverse cell wall-only sections of durum wheat (T. turgidum var. durum), spelt(T. aestivum var. spelta), einkorn (T. monococcum var. monococcum), emmer (T. turgidum var. dicoccum), rye (Secale cereale), barley (Hordeum vulgare), oats (Avena sativa) andBrachypodium distachyon. Blue ¼ HS-AX, green ¼ LS-AX, red ¼ high levels of b-glucan, purple ¼ low b-glucan, black ¼ holes and white ¼ starch.

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141 137

water. It was also clear from the visual images of the sections thatthe endosperm cell walls of Brachypodium were considerablythicker than those of the other cereals. This is consistent withmeasurements by Scanning Electron Microscopy which showedthat the cell walls may be up to 4 mm thick (Opanowicz et al., 2011).

3.2. 1H NMR spectroscopy

Complementary analyses of the endosperm cell walls were alsomade using 1H NMR spectroscopy (Fig. 3). The signals of anomeric

protons of the arabinofuranosyl units (Araf) arise in thed 5.2e5.4 ppm region while those from the xylopyranosyl (Xylp)residues are at d 4.4e4.7 ppm (Hoffman et al., 1992). Single Arafunits linked to the 3-Oe of a Xylp residue (d 5.39) are distinguish-able from those attached to 3-Oe and 2-Oe of the same Xylpresidue (d 5.28 and 5.22, respectively). The protons of Araf unitsattached to the 2-Oe of amonosubstituted Xylp residuewould ariseat around d 5.29 but this type of substitution appears to be under-represented in the wheat AX model of Gruppen et al. (1993)compared with barley AX where a larger amount is present. A

Fig. 3. 1H NMR spectra for the endosperm cell walls of

Fig. 2. Average FT-IR spectra for the endosperm cell walls of bread wheat, durumwheat (T. turgidum var. durum), einkorn (T. monococcum var. monococcum), emmer(T. turgidum var. dicoccum), spelt (T. aestivum var. spelta), rye (Secale cereale), Brachy-podium distachyon, barley (Hordeum vulgare) and oats (Avena sativa).

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141138

weak doublet at d 4.66 corresponds to H1 of the disubstituted Xylpresidues. The weakness of this signal relative to the H1 signals ofthe two Araf anomeric signals (d 5.28 and 5.22) suggests that thedisubstituted Xylp backbone unit is motionally constrained at 70 �Cwhereas the two Araf units are relatively free. The anomeric signalsof the un-substituted and monosubstituted Xylp residues atd 4.4e4.5 are obscured in Fig. 3 by the suppressed water peakwhich is at d w4.4 at 70 �C. However, the H2 signal of un-substituted Xylp is seen at d 3.32.

Our analyses showed that the AX of durum wheat, spelt,einkorn and emmer varied in their proportions of mono-, di-, andun-substituted xylose residues, although the spectra were allsimilar to those obtained for bread wheat cultivars (Toole et al.,2010, 2011). However, the 1H NMR spectra of rye cell wallswere clearly different from those of the wheat species. Althoughthere was a broad underlying peak of unknown origin it was clearfrom the intensity of the sharp peak at d 5.39 relative to those atd 5.28 and d 5.22 that the ratio of mono- to di-substituted xyloseunits was considerably higher in rye than in the wheat species.This is consistent with the data of Rantanen et al. (2007), basedon HPAEC-PAD chromatography of enzymatically hydrolysed cellwall samples. The peaks corresponding to AX were almostundetectable in the spectrum for oat endosperm cell walls,although a large doublet in the d 4.75e4.79 ppm region wasobserved, which is consistent with the high (1,3;1,4)-b-glucancontent. It was not possible to remove all of the starch from thebarley cell wall sections which obscured the peaks correspondingto cell wall polysaccharides in the 1H NMR spectrum of thisspecies.

3.3. Principal Components Analysis (PCA)

PCA was carried out to compare the cell wall compositions ofthe cereals analysed here (coloured spots) with those of the 50bread wheat lines analysed previously (Toole et al., 2011) (blackspots). Analysis of all of the species (Fig. 4) showed that thedurum, einkorn, emmer and rye lines (shown in purple, red,

durum wheat, spelt, einkorn, emmer, rye and oats.

Fig. 4. PCA plot, plus loadings, for the averaged spectra for the spectral images from cell wall-only sections of all of the cereal species that were analysed, plus 50 bread wheatvarieties analysed previously (Toole et al., 2011).

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141 139

orange and yellow, respectively) were grouped together, close tothe 50 bread wheat lines, indicating similar, but distinct, cell wallcompositions. The two spelt lines (light blue spots) were distrib-uted randomly across the plot in the same region as the 50 breadwheat lines, suggesting greater variation and heterogeneitybetween the individual grains analysed, although they all hada similar composition to bread wheat. The barley and oat lines(pink and green spots, respectively) were separated only slightlyfrom the other cereals in loading 1 (47.2% of the variance), whichindicated differences in bulk composition in both the carbohydrateand protein regions of the spectra. They were, however, clearlyseparated in loading 2 (36.5% of the variance), as was Brachypo-dium, which indicated differences in the relative contents of(1,3;1,4)-b-glucan. The PCA separation is therefore consistent withthe differences in the endosperm cell wall compositions of thecereals, with those of barley and oats comprising about 30% AXand 70% (1,3;1,4)-b-glucan compared with about 70% AX and 20%(1,3;1,4)-b-glucan in wheat species and rye (Bacic and Stone,1980). The PCA analysis showed that the cell walls of oatsdiffered most from those of wheat, being furthest away in both PCscores 1 and 2. The spectra for the cell walls of Brachypodium againdiffered from those of all of the cultivated cereals, the greatestdifference being identified by a þve score in loading 1. This wasidentified as corresponding to the shoulder at 980 cm�1 whichprobably corresponded to insoluble cell wall polysaccharides such

as cellulose and xylan (Guillon et al., 2011). A difference in(1,3;1,4)-b-glucan content was also identified in loading 2, wherethe cell wall composition lay between those of the wheat speciesand those of barley and oats. This is consistent with B. distachyonbeing related to a similar extent to both Avena (oats) and Triticum(wheat) (Kellogg, 2001).

PCA analysis was also carried out on the spectra of the wheatand rye samples only (Fig. 5). This provided a clearer separationbetween the wheat lines than when all species were included inthe analysis, showing that the diploid einkorn (T. monococcumvar. monococcum) (genomes AbAb) and tetraploid durum (T. tur-gidum var. durum) and emmer (T. turgidum var. dicoccum) (bothAuAuBB) wheats grouped together, above, and distinctly separatefrom, the 50 hexaploid wheat lines (T. aestivum var. aestivum)(AuAuBBDD). The hexaploid spelt lines (T. aestivum var. spelta)(AuAuBBDD) are distributed randomly within the same area asthe bread wheat lines, which is consistent with the fact thatbread wheat and spelt are different varieties of the same species(T. aestivum), and that several spelt varieties contain a breadwheat background as a result of spelt � bread wheat crosses. Therye samples are clearly separated from the wheats. The loadingsfor these principal components (also given in Fig. 5) showdifferences in the shoulder at 1075 cm�1 which indicatesdifferences in arabinose substitution, and in the (1,3;1,4)-b-glucan peak at 1060 cm�1.

-10

-8

-6

-4

-2

0

2

4

6

8

10

-15 -10 -5 0 5 10 15 20

PC Score 1 (31.6%)

PC

S

co

re 2

(1

8.2

%)

50 wheat

lines

Durum wheat

Rye

Spelt

Monococcum

Dicoccum

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

900 950 1000 1050 1100 1150 1200

Loading 1

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

900 950 1000 1050 1100 1150 1200

Loading 2

Fig. 5. PCA plot, plus loadings, for the averaged spectra for the spectral images from cell wall-only sections of the wheat species plus rye, plus 50 bread wheat varieties analysedpreviously (Toole et al., 2011).

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141140

4. Concluding remarks

The differences and similarities in endosperm cell wallcomposition identified by the FT-IR and 1H NMR analyses reportedhere correlate with the genetic relationships of the species ana-lysed. Within the wheat species, the hexaploid bread wheats con-tained, on average, the highest levels of total dietary fibre in wholegrain (Gebruers et al., 2008), ranging from 11.5 to 18.3% of drymatter (dm), whereas diploid einkorn (T. monococcum L.) andtetraploid emmer wheats (T. dicoccum Schübler) contained thelowest levels (7.2e12.8% of dm) and tetraploid durum (Triticumdurum Desf.) and hexaploid spelt wheats contained intermediatelevels (10.7e15.5% of dm). The total AX (TOT-AX) levels in the flourswere comparable between the different types of wheat and variedbetween 1.35 and 2.75% of dm (Gebruers et al., 2008). For the barleyflours, the TOT-AX content was in the range determined for wheatand varied between 1.4 and 2.2%, whereas the rye flours had highercontents (3.11e4.31%) (Nystrom et al., 2008) and the oat flours hadlower contents (0.97e1.26%) (Andersson et al., 2008). Much highercontents of cell wall polysaccharides have been reported for theendosperm of Brachypodium, accounting for over half of the dryweight (Guillon et al., 2011).

The hexaploid bread wheat (T. aestivum) lines had similar cellwall compositions but differed from the tetraploid durum andemmer (T. turgidum) lines, which were similar to each other. The

barley (H. vulgare), rye (S. cereale), oats (A. sativa) and Brachypodiumlines have more distinctly different cell wall compositions. Thespectra of the Brachypodium cell walls also showed significantdifferences to those of the cultivated cereals which could relate tothe reported presence of significant amounts of cellulose, or theobservation that the b-glucan in Brachypodium cell walls havea much higher BG3/BG4 ratio than the other cereals (Guillon et al.,2011).

The diversity of cell wall composition between different cerealsis consistent with the observation that fibre content and compo-sition of cereals is genetically determined (Quraishi et al., 2011;Shewry et al., 2010). This suggests that it should be possible todevelop new cereal lines with contents and compositions (e.g.,different levels of substitution) of AX and proportions of poly-saccharides that are optimised for different end use requirements.For example, manipulating AX may be used to either increase theviscosity of whole grain products for human nutrition, thuspotentially reducing the glycaemic response, or reduce the viscosityof grain for animal feed to prevent sticky faeces in poultry andscouring in pigs.

Acknowledgements

The Institute of Food Research and Rothamsted Research receivegrant-aided support from the Biotechnology and Biological

G.A. Toole et al. / Journal of Cereal Science 56 (2012) 134e141 141

Sciences Research Council (BBSRC) in the UK. This publication isfinancially supported by the European Commission in theCommunities 6th Framework Programme, Project HEALTHGRAIN(Food-CT-2005-514008). It reflects the authors’ views and thecommunity is not liable for any use that may be made of theinformation contained in the publication.

References

Andersson, A.A.M., Lampi, A.M., Nyström, L., Piironen, V., Li, L., Ward, J.L.,Gebruers, K., Courtin, C.M., Delcour, J.A., Boros, D., Fra�s, A., Dynkowska, W.,Rakszegi, M., Bed}o, Z., Shewry, P.R., Åman, P., 2008. Phytochemicals and dietaryfiber components in barley varieties in the HEALTHGRAIN diversity screen.Journal of Agricultural and Food Chemistry 56, 9767e9776.

Bacic, A., Stone, B., 1980. (1/3)-and (1/4)-linked b-D-glucan in the endospermcell walls of wheat. Carbohydrate Research 82, 372e377.

Barron, C., Parker, M.L., Mills, E.N.C., Rouau, X., Wilson, R.H., 2005. FT-IR imaging ofwheat endosperm cell walls in situ reveals compositional and architecturalheterogeneity related to grain hardness. Planta 220, 667e677.

Brennan, C.S., Cleary, L.J., 2005. The potential use of cereal (1-3, 1-4)-b-D-glucans asfunctional food ingredients. Journal of Cereal Science 42, 1e13.

Fincher, G.B., 1975. Morphology and chemical composition of barley endosperm cellwalls. Journal of the Institute of Brewing 81, 116e122.

Fincher, G.B., Stone, B.A., 1986. Cell walls and their components in cereal graintechnology. In: Pomeranz, Y. (Ed.), Advances in Cereal Science and Technology.American Association of Cereal Chemists Inc., St. Paul, MN, pp. 207e295.

Gebruers, K., Dornez, E., Boros, D., Fra�s, A., Dynkowska, W., Bed}o, Z., Rakszegi, M.,Delcour, J.A., Courtin, C.M., 2008. Variation in the content of dietary fiber andcomponents thereof in wheats in the HEALTHGRAIN diversity screen. Journal ofAgricultural and Food Chemistry 56, 9740e9749.

Gruppen, H., Hoffman, R.A., Kormelink, F.J.M., Voragen, A.G.J., Kamerling, J.P.,Vliegenhart, J.F.G., 1992. Characterisation by 1H NMR spectroscopy of enzymi-cally derived oligosaccharides from alkali-extractable wheat-flour arabinox-ylan. Carbohydrate Research 233, 45e64.

Gruppen, H., Kormelink, F.J.M., Voragen, A.G.J., 1993. Water-unextractable cell wallmaterial from wheat flour. 3. A structural model for arabinoxylans. Journal ofCereal Science 18, 111e128.

Guillon, F., Bouchet, B., Jamme, F., Robert, P., Quemeneri, B., Barron, C., Larre, C.,Dumas, P., Saulnier, L., 2011. Brachypodium distachyon grain: characterisation ofendosperm cell walls. Journal of Experimental Botany 62, 1001e1015.

Hoffman, R.A., Geijtenbeek, T., Kamerling, J.P., Vliegenhart, J.F.G., 1992. 1H NMRstudy of enzymically generated wheat-endosperm arabinoxylan oligosaccha-rides: structures of hepta- to tetradeca-saccharides containing two or threebranched xylose residues. Carbohydrate Research 223, 19e44.

Höije, A., Sternemalm, E., Heikkinen, S., Tenkanen, M., Gatenholm, P., 2008. Materialproperties of films from enzymatically tailored arabinoxylans. Bio-macromolecules 9, 2042e2047.

Kellogg, E.A., 2001. Evolutionary history of the grasses. Plant Physiology 125,1198e1205.

Mares, D.J., Stone, B.A.,1973. Studies onwheat endosperm.1. Chemical composition andultrastructureof the cellwalls. Australian Journal ofBiological Science26, 793e812.

Miller, S.S., Fulcher, R.G., 1995. Oat endosperm cell walls: II. Hot-water solubilizationand enzymatic digestion of the wall. Cereal Chemistry 72, 428e432.

Nyström, L., Lampi, A.M., Andersson, A.A.M., Kamal-Eldin, A., Gebruers, K.,Courtin, C.M., Delcour, J.A., Li, L., Ward, J.L., Fra�s, A., Boros, D., Rakszegi, M.,Bed}o, Z., Shewry, P.R., Piironen, V., 2008. Phytochemicals and dietary fibercomponents in rye varieties in the HEALTHGRAIN diversity screen. Journal ofAgricultural and Food Chemistry 56, 9758e9766.

Opanowicz, M., Hands, P., Betts, D., Parker, M.L., Toole, G.A., Mills, E.N.C.,Doonan, J.H., Drea, S., 2011. Endosperm development in Brachypodium dis-tachyon. Journal of Experimental Botany 62, 735e748.

Poutanen, K., Shepherd, B., Shewry, P.R., Delcour, J.A., Bjorck, I., van der Kamp, J.W.,2008. Beyond whole grain: the European HEALTHGRAIN project aims athealthier cereal foods. Cereal Foods World 53, 32e35.

Quraishi, U.M., Murat, F., Abrouk, M., Pont, C., Confolent, C., Oury, X., Ward, J.,Boros, D., Gebruers, K., Delcour, J.A., Courtin, C.M., Bedo, Z., Saulnier, L.,Guillon, F., Balzergue, S., Shewry, P.R., Feuillet, C., Charmet, G., Salse, J., 2011.Combined meta-genomics analyses unravel candidate genes for the grain die-tary fiber content in bread wheat (Triticum aestivum L.). Functional IntegratedGenomics 11, 71e83.

Rakszegi, M., Boros, D., Kuti, C., Láng, L., Bed}o, Z., Shewry, P.R., 2008. Compositionand end-use quality of 150 wheat lines selected for the HEALTHGRAIN diversityscreen. Journal of Agricultural and Food Chemistry 56, 9750e9757.

Rantanen, H., Virkki, L., Tuomainen, P., Kabel, M., Schols, H., Tenkanen, M., 2007.Preparation of arabinoxylobiose from rye xylan using family 10 Aspergillusaculeatus endo-1,4-b-D-xylanase. Carbohydrate Polymers 68, 350e359.

Shewry, P.R., Piironen, V., Lampi, A.M., Edelmann, M., Kariluotos, S., Nurmi, T., Fer-nandez-Orozco, R., Ravel, C., Charmet, G., Andersson, A.A., Åman, P., Boros, D.,Gebruers, K., Dornez, E., Courtin, C.M., Delcour, J.A., Rakszegi, M., Bed}o, Z.,Ward, J.L., 2010. The HEALTHGRAIN wheat diversity screen: effects of genotypeand environment on phytochemicals and dietary fiber components. Journal ofAgricultural and Food Chemistry 58, 9291e9298.

Shewry, P.R., Piironen, V., Lampi, A.M., Nyström, L., Li, L., Rakszegi, M., Fra�s, A.,Boros, D., Gebruers, K., Courtin, C.M., Delcour, J.A., Andersson, A.A.M.,Dimberg, L., Bed}o, Z., Ward, J.L., 2008. Phytochemicals and fiber components inoat varieties in the HEALTHGRAIN diversity screen. Journal of Agricultural andFood Chemistry 56, 9777e9784.

Toole, G.A., Barron, C., Le Gall, G., Colquhoun, I.J., Shewry, P.R., Mills, E.N.C., 2009.Remodelling of arabinoxylan in wheat (Triticum aestivum) endosperm cell wallsduring grain filling. Planta 229, 667e680.

Toole, G.A., Le Gall, G., Colquhoun, I.J., Johnson, P., Bed}o, Z., Saulnier, L., Shewry, P.R.,Mills, E.N.C., 2011. Spectroscopic analysis of diversity of arabinoxylan structuresin endosperm cell walls of wheat varieties (Triticum aestivum) in theHEALTHGRAIN diversity collection. Journal of Agricultural and Food Chemistry59, 7075e7082.

Toole, G.A., Le Gall, G., Colquhoun, I.J., Nemeth, C., Saulnier, L., Lovegrove, A.,Pellny, T., Freeman, J., Mitchell, R.A.C., Mills, E.N.C., Shewry, P.R., 2010. Temporaland spatial changes in cell wall composition in developing grains of wheat cv.Hereward. Planta 232, 677e689.

Toole, G.A., Wilson, R.H., Parker, M.L., Wellner, N.K., Wheeler, T.R., Shewry, P.R.,Mills, E.N.C., 2007. The effect of environment on endosperm cell-wall devel-opment in Triticum aestivum during grain filling: an infrared spectroscopicimaging study. Planta 225, 1393e1403.

Vinkx, C.J.A., Delcour, J.A., 1996. Rye (Secale cereale L) arabinoxylans: a criticalreview. Journal of Cereal Science 24, 1e14.

Vogel, J.P., Garvin, D.F., Mockler, T.C., Schmutz, J., Rokhsar, D., Bevan, M.W., et al.,2010. Genome sequencing and analysis of the model grass Brachypodium dis-tachyon. Nature 463, 763e768.

Ward, J.L., Poutanen, K., Gebruers, K., Pironen, V., Lampi, A.M., Nyström, L.,Andersson, A.A.M., Åman, P., Boros, D., Rakszegi, M., Bed}o, Z., Shewry, P.R., 2008.The HEALTHGRAIN cereal diversity screen: concept, results, and prospects.Journal of Agricultural and Food Chemistry 56, 9699e9709.

Wood, P.J., 1993. Physicochemical characteristics and physiological properties of oat(1/3), (1/4)-b-D-glucan. In: Wood, P.J. (Ed.), Oat Bran. American Associationof Cereal Chemists, St. Paul, MN, USA, pp. 83e112.

Wood, P.J., Beer, M.U., 1998. Functional oat products. In: Mazza, G., Shi, J., LeMayuer, M. (Eds.), 1998. Functional Foods: Biochemical and Processing Aspects,vol. 2. Technomic Publishing Co, Lancaster, PA, USA, pp. 1e37.