ABSTRACT

Sinapic acid esters (SAE) are important antinutritional compounds of the seeds of rapeseed and related Brassica species. The objective of the present study was to develop a fast and nondestructive screening method for these compounds by near-infrared reflectance spectroscopy (NIRS) and to apply it to a selection for reduced SAE content in a germplasm collection of rapeseed. Reference analyses for NIRS calibration were performed with a colorimetric method. A NIRS calibration equation for SAE content was developed from a calibration set containing 400 intact-seed samples from 17 Brassica species, ranging from 2.1 to 17.6 g kg^-1. The coefficient of determination was 0.91 and the standard error of crossvalidation was 1.1 g kg^-1 (mean ± SD of the calibration set = 10.0 ± 3.6 g kg^-1). This calibration equation was used for analysing a collection of 1361 entries of breeding material of rapeseed (Brassica napus L.), with SAE content ranging from 5.0 to 17.7 g kg^-1. The entries with the lowest SAE content (n=75, SAE<8.5 g kg^-1) were further analysed by the laboratory reference method, confirming the accuracy of NIRS predictions. Because NIRS is a multi-trait technique, information on SAE content is obtained simultaneously to other traits such as oil, protein, and glucosinolate content and the fatty acid composition of the seed oil.

INTRODUCTION

Esterified phenolic acids are the predominant forms of phenolic compounds in rapeseed meals. Among them, the choline ester of sinapic acid (sinapine) is the most abundant. The presence of phenolic compounds reduces the nutritive value of proteins, and therefore they are undesirable in rapeseed (Naczk et al., 1998). The total content of phenolic acids in rapeseed flour ranges from 6.2 to 12.8 g kg^-1 dry weight (Shahidi and Naczk, 1992). Breeding for reduced levels of phenolic compounds in this species has been rarely attempted, mainly because of the lack of adequate methods for large-scale screenings of breeding materials and germplasm accessions.

The development of near infrared reflectance spectroscopy (NIRS) revolutionised the evaluation of quality traits of agricultural commodities. In the analysis of rapeseed, NIRS provides a rapid, nondestructive, and simultaneous analysis of multiple constituents such as oil, protein and glucosinolate content (Daun and Williams, 1995), or the fatty acid composition of the seed oil (Velasco and Becker, 1998), without grinding or chemical alteration. The objective of this study was to study the potential of NIRS to analyse sinapic acid esters (SAE) content in intact rapeseed and to use this methodology to screen a core collection of breeding materials of this species.

MATERIALS AND METHODS

A set of 587 intact-seed samples from 20 Brassica species was used for the development and validation of NIRS calibration equation. The selection of these samples is described by Velasco et al. (1998). NIRS calibration equation was applied to selection for reduced SAE content in a core collection of 1361 breeding samples of rapeseed (Brassica napus L.).

The samples were scanned on a monochromator NIR Systems model 6500 (NIR Systems, Inc., Silver Springs, MD). The standard ring cup was implemented with a special adapter for small samples (300 mg intact seeds). Calibration equation was developed by using the spectral information from 400 to 2500 nm and modified partial least squares (MPLS) regression, after applying second derivative transformation (2,5,5,1), SNV, and De-trend scatter correction.

The SAE concentration was determined by using the colorimetric method of Thies (1991). It was expressed as g kg^-1 seeds. Quantification of SAE was performed with a calibration curve developed with pure sinapine thiocyanate (Matthäus, 1997). The standard error of the laboratory (SEL) was calculated from a flour sample analysed 13 times, each time at a different date. This sample had an average SAE content of 8.1 g kg^-1, and the SEL was 0.8 g kg^-1.

RESULTS AND DISCUSSION

NIRS calibration

The calibration equation for SAE content developed from a set containing 587 samples of 20 species of Brassica showed a very close relationship between the reference method and NIRS data, with coefficients of determination of 0.95 in calibration and 0.93 in cross validation (Table 1). The standard error of cross validation (SECV) was 1.0 g kg^-1, as compared with an standard error of the laboratory (SEL) of 0.8 g kg^-1.

TABLE 1. Calibration and cross validation statistics for analysis of sinapic acid esters (SAE) content by near infrared reflectance spectroscopy in a set containing 587 intact-seed samples of 20 species of Brassica.

† SEC=standard error of calibration.

‡ SECV=standard error of cross validation.

Selection for reduced SAE content in breeding materials of rapeseed

The SAE content of the breeding lines of B. napus ranged from 5.0 to 17.7 g kg^-1. This range was considerably greater than that found in germplasm material of this species (5.0 to 11.1 g kg^-1) (Velasco and Möllers, 1998). Such a difference could simply be caused by the higher number of samples analysed from breeding lines, 1361 vs. 90. The lowest levels of SAE found in the breeding material were similar to those found for B. napus in a germplasm collection (about 5.0 g kg^-1) (Velasco and Möllers, 1998), and not very different from those reported by Shahidi and Naczk (1992) (6.2 g kg^-1, dry weight basis).

Fig. 1a shows the SAE content (NIRS values) of the set of 1361 samples from breeding lines of B. napus. A total of 75 of them having SAE content below 8.5 g kg^-1 were selected and analysed by the reference method to confirm the accuracy of NIRS. The results of these analyses are shown in Fig. 1b. The SAE content in the selected samples ranged from 4.9 to 9.4 g kg^-1, demonstrating the high reliability of NIRS analyses to perform a selection for reduced levels of this trait.

Fig. 1. Selection for reduced sinapic acid esters (SAE) content based on (a) near infrared reflectance spectroscopy (NIRS) performed on 1361 samples of Brassica napus breeding material, and (b) evaluation of 75 selected samples by the reference method.

REFERENCES

1. Daun, J.K., and P.C. Williams, 1995. Use of NIR spectroscopy to determine quality factors in harvest surveys of canola, pp. 864-866. IN: Groupe Consultatif International de Recherche sur le Colza (ed.), Proceedings of the 9^th International Rapeseed Congress, Cambridge, U.K., 4-7 July 1995. Henry Ling, Dorchester, U.K.

2. Naczk, M., R. Amarowicz, A. Sullivan, and F. Shahidi, 1998. Current research developments on plyphenolics of rapeseed/canola: a review. Food Chemistry 62:489-502.

3. Matthäus, B. 1997. Antinutritive compounds in different oilseeds. Fett/Lipid 99: 170-174.

4. Shahidi, F., and M. Naczk, 1992. An overview of the phenolics of canola and rapeseed: Chemical, sensory and nutritional significance. Journal of the American Oil Chemists’ Society 69: 917-924.

5. Thies, W. 1991. Determination of the phytic acid and sinapic acid esters in seeds of rapeseed and selection of genotypes with reduced concentrations of these compounds. Fat Science Technology 93: 49-52.

6. Velasco, L., and H.C. Becker, 1998. Estimating the fatty acid composition of the oil in intact-seed rapeseed (Brassica napus L.) by near-infrared reflectance spectroscopy. Euphytica 101: 221-230.

7. Velasco, L., and C. Möllers. 1998. Nondestructive assessment of sinapic acid esters in species of Brassica: II. Evaluation of germplasm and identification of phenotypes with reduced levels. Crop Science 38:1650-1654.

8. Velasco, L., B. Matthäus, and C. Möllers. 1998. Nondestructive assessment of sinapic acid esters in species of Brassica: I. Analysis by near infrared reflectance spectroscopy. Crop Science 38:1645-1650.