Vol. 85, No. 4, 2008 571 Bulk Carbohydrate Grain Filling of Barley E-Glucan Mutants Studied by 1 H HR MAS NMR Helene Fast Seefeldt, 1,2 Flemming Hofmann Larsen, 2 Nanna Viereck, 2 Bernd Wollenweber, 1 and Søren Balling Engelsen 2,3 ABSTRACT Cereal Chem. 85(4):571–577 Temporal and genotypic differences in bulk carbohydrate accumulation in three barley genotypes differing in the content of mixed linkage E- (1o3),(1o4)-D-glucan (E-glucan) and starch were investigated using proton high-resolution, magic angle spinning, nuclear magnetic resonance ( 1 H HR MAS NMR) during grain filling. For the first time, 1 H HR MAS NMR spectra of flour from immature barley seeds are analyzed. Spectral assignments are made using two-dimensional (2D) NMR methods. Both D- and E-glucan biosynthesis were characterized by inspection of the spectra as well as by calibration to the reference methods for starch and E-glucan content. Starch was quantified with very good calibrations to the D-(1o4) peak (5.29–5.40 ppm) and the region 3.67–3.83 ppm covering starch glycopyranosidic protons from H5 and H6. In contrast, the spectral inspection of the E-anomeric region 4.45–4.85 ppm showed unexpected lack of intensity in the high E-glucan mutant lys5f at seed maturity, result- ing in poor calibration to reference E-glucan content. We hypothesize that the lack of E-glucan signal in lys5f indicates partial immobilization of the E-glucan that appears to be either genotypic dependent or water/E-glucan ratio dependent. The functionality of cereals for food and industrial purposes is mainly determined by the composition of starch and fibers, the architecture of starch (Tester et al 2004), and the content and structure of cell wall fibers (Lazaridou and Biliaderis 2007). In particular, barley seeds contain high amounts of the cell wall fiber E-glucan (mixed linkage E-(1o3),(1o4)-D-glucan) that has at- tracted much attention due to its physical and biological proper- ties (Storsley et al 2003; Tsuchiya et al 2005; Lazaridou and Biliaderis 2007; Queenan et al 2007). E-Glucan has two aspects in production: on one hand causing problems in brewing and animal feed industries, and on the other hand having beneficial influence on human health. For example, E-glucan has been able to reduce serum cholesterol in hypercholesterolemic individuals (Kalra and Joad 2000) and to modulate gluco-regulation in diabetics (Léon et al 2000). In contrast, E-glucan is unwanted in the brewing indus- try because it forms a viscous gel that leads to hazing (Fincher and Stone 1986). Furthermore, barley is considered a less valuable food source for chickens and pigs, as they gain less energy due to the viscous properties of E-glucan, which reduce colon emptying (Knudsen 2001). A renewed interest in E-glucan arises from its functionality in food processing due to its water-binding capacity (Holtekjolen et al 2006) and stabilizing and thickening ability. For example, it can be used as a fat-replacer (Burkus and Temelli 2000). The barley seed endosperm cell walls consist of |75% of E- glucan (Fincher and Stone 1986). E-Glucan acts partly as a struc- tural element and partly as a flexible storage material hydrolyzed during grain filling and germination (Buckerigde et al 2004). E- Glucan is not a strictly defined polysaccharide and structure dif- ferences occur between cereals and within cereal grains (Zhang et al 2002; Johansson et al 2004). The water-soluble ȕ-glucan from barley contains |70% E-(1o4)-glycosyl linkages and 30% E- (1o3)-glycosyl linkages. Often repetitions of two or more suc- cessive E-(1o4)-linkages are separated by single E-(1o3)- linkages (Fincher and Stone 1986). Insoluble barley E-glucan contains a higher ratio of E-(1o4) to E-(1o3)-linkages (Johans- son et al 2004). The insoluble fiber fraction appears to be nonco- valently bound to arabinoxylans and therefore remains insoluble, even in small molar masses (Johansson et al 2004). The content of E-glucan in barley seeds varies from 3% to 18– 20% (dry matter) E-glucan depending on genotype (Aastrup and Munck 1985; Munck et al 2004) and environmental factors (Aas- trup 1979; Fincher and Stone 1986). Besides the cell wall fibers, three major constituents are accu- mulated in barley during grain filling: starch, lipids, and proteins. Starch constitutes |61% of the mature grain dry weight in barley (MacGregor and Fincher 1993). Cereals contain different kinds of lipids: membrane-bound oil droplets in the aleurone layer, scutel- lum and embryo, and lipids found in the endosperm (Morrison 1978). Lipids constitute 1–3% of the cereal grain depending on genetic constitution (Jacobsen et al 2005). The lipids in the en- dosperm are lysophospholipids complexed with amylose (Morri- son 1993). In barley, proteins account for 8–13% dry weight, the majority being storage proteins surrounding the starch. The pro- tein content and composition are also genetically dependent and mutants in an isogenic background; for example, the lys3a mutant has an extreme reduction in alcohol soluble proteins and hordeins compared with the parental line (Jacobsen et al 2005). It would be valuable to correlate spectral fingerprints (unique identification of genotypes) with starch or ȕ-glucan in plant breed- ing screening programs. Nuclear magnetic resonance (NMR) spec- troscopy is a versatile technique that provides chemical as well as quantitative information (Kuchenbrod et al 1995) and has histori- cally attracted much attention in plant biology studies (Fan 1996; Bardet et al 2001; Baker et al 2006; Glidewell 2006) and it can provide detailed information on the plant metabolome (Krishnan et al 2005). Cereals consist mainly of semicrystalline compounds and the standard NMR technique to detect signals from rigid, solid-state material is by 13 C cross-polarization (CP) magic angle spinning (MAS) NMR (Pines et al 1972), by which resonances from the protonated carbons belonging to the rigid domains are enhanced. 13 C CP-MAS NMR has been used to study protein and starch in various types of seeds (O’Donnell et al 1981), triacylglycerols (Bardet et al 2001), and cell wall properties (Jarvis and McCann 2000; Tang et al 2000). Mobile-phase 1 H high-resolution (HR) MAS NMR is faster and more sensitive than 13 C CPMAS solid- phase NMR because of the higher sensitivity of 1 H. 1 H HR MAS NMR spectroscopy in combination with chemometrics has previ- ously been used for analysis of durum wheat flour for the dis- crimination of varietal and geographical origin (Brescia et al 2002). Similarly, liquid-state 1 H NMR and chemometrics have been used 1 University of Aarhus, Faculty of Agricultural Sciences, Dept. of Genetics and Biotechnology, Forsøgsvej 1, 4200 Slagelse, Denmark. 2 University of Copenhagen, Faculty of Life Sciences, Dept. of Food Science, Quality & Technology, Rolighedsvej 30, 1958 Frederiksberg C, Denmark. 3 Corresponding author. Phone: +45 3533 3205. Fax: +45 3533 3245. E-mail: se@life.ku.dk doi:10.1094/ CCHEM-85-4-0571 © 2008 AACC International, Inc.