ABSTRACT

Dietary fibres (DF) are the qualitatively most dominating part of cruciferous oilseed next to the oil. These DF constituents have binding properties quite different from legume and cereal DF. Studies performed on cruciferous DF have been directed both at their β-glucane constituents and their associated compounds based on use of supercritical fluid techniques (SFT), micellar electrokinetic capillary chromatography (MECC) and chemometric methods. By use of the developed Biorefining technique based on separation of the seed components at gentle conditions in aqueous solutions or emulsions, hulls are removed and different protein concentrate and isolates are produced as added-value products containing various amounts of DF. SFT comprising extraction (SFE) and chromatography (SFC) are together with MECC proven to be effective for extraction and analysis of DF associated lipophilic and amphiphilic compounds. These compounds constitute an appreciable part of DF, and they are considered to be important for the properties of DF, and thereby the plants cell walls. Ongoing research is focused on use of chemometric methods as a fast and simple tool to differentiate between DF from different sources.

Introduction

DF includes a complex group of mainly plant cell wall derived compounds, which are non-digestible in the human small intestine. These compounds include non-starch polysaccharides (NSP), proteins, and lignins, and to a smaller extent various non-carbohydrate compounds such as various lipids and phenolics (Andersen et al., 1997, 1998). Certain physiological responses have been associated with the consumption of dietary fibres and the physical and chemical properties of individual DF compounds appear to be important in determining the physiological response to sources of DF in the diet (Schneeman, 1987).

The aim of the present work has been concentrated on investigations concerning variations in the content of dietary fibre and compounds associated to these fractions. By use of SFE, UV-VIS-spectroscopy, MECC and principal component analysis (PCA) (Martens and Næs, 1989) the content and composition of lipophilic and amphiphilic compounds associated to total dietary fibre (TDF) from rapeseed, rapeseed fractions, four cereal grains, and seeds of three legumes were compared.

Material and methods

TDF of rapeseed (RS), rapeseed protein rich meal (PRM), rapeseed lipoprotein (LIPRO), rapeseed hulls (RpH), lupine seed (LS), dehulled lupine seed (LM), lupine hulls (LH), round pea (RP), wrinkled pea (WP), wheat (W), barley (B), rye (R), and oat (O) were isolated by a modification of the enzymatic gravimetric procedure originally developed by Asp et al. (1983) (Andersen et al., 1998). Protein and ash analysis were performed on the weighted residuals by Kjeldahl analysis and a standard AOAC method, respectively (Andersen et al., 1998). SFE on 1 g samples of TDF were performed using pure CO₂ (SFE A) followed by 15 % methanol in CO₂ (SFE B), and the extracts were afterwards analysed by UV-VIS-spectroscopy and MECC (Andersen et al., 1998). Finally chemometric analysis of the UV-VIS-spectra and the normalised area of the compounds analysed by MECC were performed (Andersen et al., 1999).

Results and discussion

Results of the TDF analysis is shown in together with the content of protein and ash. Contents of soluble DF (SDF) and insoluble DF (IDF) depends upon the source from which the fibres are isolated. Levels vary from 8.3 to 31.4 g/100 g for IDF and 11.3 to 22.6 g/100 g for SDF regarding whole grain sources.

The plant cell walls consist of a complex matrix of various types of molecules responsible for maintaining the shape and rigidity of the cell as well as transportation of metabolites and signal molecules across the cell wall membranes. The major part of the cell walls consist of polysaccharides but for transportation of lipophilic as well as hydrophilic molecules across the cell wall, various types of lipophilic, amphiphilic, and hydrophilic molecules are found in close connection to the polysaccharide fraction. As the major part of DF consists of polysaccharides from the cell walls it is expected that minor parts of non-carbohydrate cell wall constituents are bound to the DF carbohydrates in a way that makes hydrolysis impossible. This type of molecules (e.g. proteins, lignins, and phospholipids) are thus included in the AOAC working definition of enzymatic gravimetric determined DF, which states that DF are: “the remnants of plant cells, polysaccharides, lignin and associated substances resistant to hydrolysis (digestion) by the alimentary enzymes of human” (Cho et al., 1997).

Table. Uncorrected levels of IDF, SDF and TDF along with the values of ash, protein and TDF corrected for ash and protein (TDF_C). Levels (g /100 g) are given as mean values and relative standard deviation (%) in parenthesis where available (Andersen et al., 1998).

SFE of DF from rapeseed (RS) and lupine (LS) revealed pronounced differences in the amount and binding strength of SFE extractable compounds). Results from determination of total extractable lipophilic materials including fat (oil) from RS and LS revealed almost complete extraction within the first 30 min by pure CO₂ (SFE A) regardless of the DF type used. Changing the polarity of the fluid by addition of MeOH resulted in further extraction of lipid-amphiphilic material with the efficiency of the modifier depending on the amount of modifier used as well as the DF type. The results (Figure 1) show that the RS fibres bind the extractable compounds more tightly than the LS fibres. This results in continuous release of materials from RS whereas extraction of LS seemed to end within 60 min from modifier addition (150 mL / L). The DF fractions from RS, PRM, and RH have a much stronger binding of SF-extractable material than DF

Figure. Mass of DF-associated compounds extractable by supercritical CO₂ as function of time and modifier concentration (left). Gravimetric determination of TDF associated SF extractable compounds (right). The results are given as mg/g. SFE A refers to extraction with pure CO₂ whereas SFE B refers to extraction using 15 % MeOH. Maximal level of extractable enzyme derived material (reference samples) equals 18.6 and 8.6 mg/g TDF for SFE A and B respectively. (− ⋅ − ⋅) 0 % MeOH, () 15 % MeOH, (- - -) 25 % MeOH (Andersen et al., 1998).

from cereals and legumes except DF from LM. The high level of SF-extractable material in LM may be more or less a result of the alkaloids and other lipophilic compounds present in lupine. It is noteworthy that the conditions used to obtain SFE A are identical to the degreasing extractions carried out on rapeseed, lipoprotein and protein rich meal prior to TDF isolation. The lipids extracted from TDF within the first 30 min have thus been made available for SFE extraction by the DF isolation procedure. The results could therefore express the ability of the TDF fractions to prevent digestion of lipids by pancreatic lipase during the DF preparation.

Examinations of the extracts using UV-VIS-spectroscopy revealed great variations in the composition of the SFE extractable DF associated compounds (Table). Especially the variations in the amounts of cinnamic acid derivatives seems to differentiate the DF materials from the three plant families represented in this study. Analysis by MECC also showed pronounced variations in the compositions of the SFE extracts. As seen in Figure, various compounds are analysed, but the

Table. SF-extractable compounds in TDF analysed by UV-VIS-spectroscopy and quantified using Lambert Beer’s law. The wavelength areas used were 278-280 nm (proteins), 325-330 nm (cinnamic acid derivatives; CAD’s), 470-474 nm (carotenoides) and 667-669 nm (chlorophyll) respectively. Relative standard deviations (%) are given in brackets when available (Andersen et al., 1998).

Figure. MECC of LIPRO DF, SFE B. Analysis was performed using an ABI Model 270A-HT capillary electrophoresis system with a 760 x 0,05 mm I.D. fused-silica capillary. Detection (235 nm) was performed on-column 530 mm from inlet. The separation was performed at a temperature and voltage of 30^OC and +20 kV, respectively. Buffer system consisted of 400 mM taurine, 15 mM Na₂HPO₄, 140 mM cholic acid and 2 % 1-PrOH (Buskov et al., 1997). 0.1 mM trigonellinamide was used as external standard (Std) (Andersen et al., 1998).

final identification of these compounds are yet to be performed. Chemometric analysis of the UV-VIS-spectra and MECC data (Andersen et al., 1999) have though shown, that the compounds 1-6, 9-12, and 14 have spectra which are closely related to the UV-VIS-spectra of cinnamic acid derivatives. In the same manner are the compounds 17, 18, and 20 related to a high content of DF associated chlorophyll, and might thus be chlorophyll derivatives.

Principal component analysis of the UV-VIS-spectra is shown in Figure . The three groups marked Legumes, Cereals, and Cruciferous are primarily separated in the direction of the first principal component which describes the variations in the amount of cinnamic acid derivatives (data not shown). Chemometric methods of near-infrared spectra of dietary fibres have resulted in fast methods for determination of the DF content in various types of foodstuffs (Williams et al., 1991; Baker, 1995; Kays et al., 1996, Windham et al., 1997; Kays et al., 1998). Another possibility is to determine small impurities, or “indicator substances” for evaluation of the quality parameters and process efficiency as provided by Nørgaard (1995). Further studies of the SFE B extractable compounds of TDF might, in the same manner, result in chemometric models that could supplement the methods for DF determination by estimating the fibre composition and perhaps the plant materials from which the DF materials were derived.

Figure. Score plot of principal component analysis on SFE of TDF analysed by UV-VIS-spectroscopy. Unmarked: SFE A; Bold + *: SFE B. The three crops are marked Legume, Cereal and Cruciferous (Andersen et al., 1999).

CONCLUSION

The physico-chemical properties and the physiological effects of dietary fibre depend on the composition of the DF fraction. This initial study using chemometric methods on UV-VIS-spectra and peaks in the MECC chromatograms have indicated that further investigations of amphiphilic TDF associated compounds could result in models for prediction of the composition of the TDF fraction in relation to the plant materials from which the DF are derived.

ACKNOWLEDGEMENTS

This work was financially supported by the Commission of European Union (Contract Nos. FAIR CT 95-0260 and FAIR CT 98-3778).

REFERENCES

1. Andersen K.E., Buskov S., Bjergegaard C., Sørensen H., Sørensen J.C. and Sørensen S. (1997). Analyses of dietary fibre associated non-carbohydrate constituents based on supercritical fluid extraction. In: Proceedings of Euro Food Chem IX, September 24-26, Interlaken Switzerland (eds. R Amadó and R. Battaglia), Swiss Society of Food and Environmental Chemistry (SGLUC), vol. 1, 209-214.

2. Andersen K.E., Bjergegaard C. and Sørensen H. (1998). Characterisation of dietary fibre associated constituents based on supercritical fluid extraction. Polish Journal of Food and Nutrition Science, 7/48 No 4, 617-630.

3. Andersen K.E., Bjergegaard C., Nørgaard L., and Sørensen H. (1999). Chemometric analysis of dietary fibre associated compounds. GCIRC Bulletin, 16, in press.

4. Asp N.-G., Johansson C.-G., Hallmer H., and Siljeström M. (1983). Rapid enzymatic assay of insoluble and soluble dietary fiber. Journal of Agricultural and Food Chemistry, 31: 476-482.

5. Baker D. (1995). The determination of fiber, starch, and total carbohydrate in snack foods by near-infrared reflectance spectroscopy. Cereal Foods World 30, 6, 389-392.

6. Buskov, S., Mortensen, K., Møller, P., Sørensen, H., Sørensen, J.C. and Sørensen, S. (1997). Comparison of micellar electrokinetic capillary chromatography and high performance liquid chromatography for determination of different hydroxycinnamic acid esters of malic acid occurring in cruciferous seedlings. In: Proceedings of Euro Food Chem IX, September 24-26, Interlaken Switzerland (eds. R Amadó and R. Battaglia), Swiss Society of Food and Environmental Chemistry (SGLUC), pp. 632-637.

7. Cho S., Devreis J.W. and Prosky L. (1997). Dietary fiber analysis and applications. AOAC International, Gaithersburt, Maryland, USA, pp. 1-10.

8. Kays S.E., Windham W.R., and Barton F.E. II. (1998). Prediction of total dietary fiber in cereal products using near-infrared reflectance spectroscopy. Journal of Agricultural and Food Chemistry, 44, 2266-2271.

9. Kays S.E., Windham W.R., and Barton F.E. II. (1996). Prediction of total dietary fiber by near-infrared reflectance spectroscopy in high-fat- and high-sugar-containing cereal products. Journal of Agricultural and Food Chemistry, 46, 854-861.

10. Martens H., and Næs T. (1989). Multivariate calibration. Wiley, New York.

11. Nørgaard L. (1995). Classification and prediction of quality and process parameters of thick juice and beet sugar by fluorescence spectroscopy and chemometrics. Zuckerindustrie, 120, 11 970-981.

12. Schneeman B.O. (1987). Soluble vs. insoluble fiber – different physiological responses. Food Technology, 2, 81-82.

13. Windham W.R., Kays S.E., and Barton F.E. II. (1997). Effect of cereal product residual content on total dietary fibre determined by near-infrared reflectance spectroscopy. Journal of Agricultural and Food Science, 45, 140-144.

14. Williams P.C., Cordeiro H.M., and Harnden M.F.T. (1991). Analysis of oat bran products by near-infrared reflectance spectroscopy. Cereal Foods World, 36, 571-574.