ABSTRACT

Indian mustard (Brassica juncea L. Czern.+ Coss.) is an important animal and human food source, mainly because its high quality oil and high protein seed meal. However, there is a need to eliminate antinutritive glucosinolates present in the meal. Near infrared reflectance spectroscopy (NIRS) is a rapid, reliable and non-destructive method of estimating quality components in many agricultural products. The purpose of this study was to test the accuracy of NIRS to estimate individual and total glucosinolate content in the intact seed. A total of 69 samples of Brassica juncea were analysed for sinigrin (SIN) and gluconapin (GLA), and 47 for total glucosinolate content (GSL) by high performance liquid chromatography (HPLC). Calibration equations for SIN, GLA and GSL were developed. Standard error of calibration (SEC) ranged values from 1.50 to 9.84 μmol·g^-1 and coefficients of determination (R²) from 0.97 to 0.99. These results show that NIRS could be used as a reliable technique for screening Brassica juncea germplasm for individual and total glucosinolate content.

INTRODUCTION

Modern cultivars and breeding lines of Indian mustard (Brassica juncea (L.) Czern.+ Coss.) yield high-quality oil, high-protein seed meal, and are relatively resistant to heat stress, water stress, pod-shattering and fungal diseases (Palmer et al., 1988), and could be used as an oilseed crop in the Mediterranean area (Fereres et al., 1983). However, the high glucosinolate content in seeds make the meal unacceptable for animal and human consumption. The high cost and time-consuming of high performance liquid chromatography (HPLC) analyses are serious handicaps to analyse a large set of samples. With an awareness of this problem, we test in this study the accuracy of near infrared reflectance spectroscopy (NIRS) to estimate these parameters in the intact seed. In this way NIRS has been shown to be a rapid, reliable, non-destructive method of testing the composition of Brassica juncea seed in a cost-effective manner (Font et al., 1998a, 1998b). To determine the concentration of sinigrin (SIN, allylglucosinolate) and gluconapin (GLA, butenylglucosinolate), which are the principal seed glucosinolates of brown-seeded Indian mustard lines (Gland et al., 1981; Palmer et al., 1987), and total glucosinolate content (GSL), calibration equations for both individual and total glucosinolate content were developed.

MATERIALS AND METHODS

Selected samples

One set of 69 seed samples of B. juncea, being part of a germplasm collection from different geographical origins was chosen to develop this study. These samples were selected to range all the variation present in the collection for sinigrin (SIN), gluconapin (GLA) and total glucosinolate content (GSL). Plant material was sown in November, 1995 in the Centro de Investigación y Formación Agraria (CIFA) (Córdoba) in a typic xerofluvent soil (pH 8.0). Individual plants were harvested in July, 1996.

High performance liquid chromatography

About 100 mg of seeds from the same samples previously scanned, were homogenised in a mill and then dried in a oven at 75ºC for 18-20 hours. For the first extraction step the homogenised seeds were heated for 15 min in 2.5 ml 70% aqueous methanol and an internal standard (200 μl 10 mM glucotropaeolin). After centrifugation (5 min, 5x10³ g) a second extraction was achieved using 2 ml 70% aqueous methanol. The temperature in the waterbath was held at 75ºC during the process. 1 ml of the centrifuged extraction of each sample was pippeted on the top of an ion-exchange column containing 1 ml DEAE Sephadex A25 in the formiate form. Desulfation was carried out by the addition of 75μl of purified sulfatase (Type H-1 from Helix pomatia). Desulfated glucosinolates were eluted with 2.5 ml (0.5 ml x 5 times) with an ultra-pure water system (Milli-Q from Millipore) and analysed in a Waters HPLC equipment with an UV Tunable Absorbance Detector at a wavelength of 229 nm. Separation was carried out by using a Lichrospher 100 RP-18 in Lichrocart 125-4 column (5 μm particle size).

NIRS analysis

To perform NIRS analysis, intact seed samples described above were scanned in a NIRS monochromator (NIR Systems mod. 6500, NIRSystems, Inc., Silversprings, MD, USA), and their spectra collected between 400-2500 nm., registering the absorbance values (log 1/R) at 2 nm intervals for each sample.

Prediction equations for SIN, GLA and GSL were developed using the ISI program CALIBRATE with the modified partial least squares regression option. Four mathematical treatments (0,0,1,1 (derivative, gap, smooth, second smooth); 1,4,4,1; 1,10,10,1 and 2,5,5,1) were tested on the calibration set. Standard normal variate and detrend transformations were used to correct scattering, and two passes was the option chosen to eliminate outliers. Wavelengths from 400 to 2500 nm every 4 nm, were used for calibration. The standard error of calibration (SEC), coefficient of determination (R²), standard error of cross-validation (SECV) and 1-VR (1 minus the ratio of unexplained variance to total variance) statistics were used to characterise the different equations obtained and to determine the best calibration equation (Windham et al., 1989). SECV was used as an estimate of the standard error of performance (SEP) (Mark and Workman, 1991). The best calibration equations were obtained with the following mathematical treatments: 1,10,10,1; 1,4,4,1 and 2,5,5,1 for SIN, GLA and GLS, respectively.

RESULTS AND DISCUSSION

High performance liquid chromatography.

Table 1 shows reference values for SIN, GLA and GSL in the sample set. GSL content (145.06 ± 33.08 μmol·g^-1) is similar to those reported by Palmer (Palmer et al., 1988) in reference to 3 different lines of Indian mustard plants which exhibited values between 105 and 242 μmol·g^-1 (dry wt., oil free basis). Nevertheless, the highest values reported by these authors could be related with the addition of sulphur to the field . In our entries, GLA and SIN contents varied from 0.00 to 162.19 μmol·g^-1 (mean= 52.82) and from 0.56 to 181.72 μmol·g^-1 (mean= 80.80), respectively. Profiles for individual glucosinolates showed large variation among samples with different combinations between them.

Table 1. High performance liquid chromatography (HPLC) laboratory reference value statistics for sinigrin, gluconapin and total glucosinolate content.

SIN= sinigrin; GLA= gluconapin; GSL= total glucosinolate content;

Std= standard deviation.

NIRS

The NIRS intact-seed equations for GSL, GLA and SIN showed low SEC (1.50, 4.49 and 9.84 μmol·g^-1) and high R² (0.99, 0.99 and 0.97) (Table 2) respectively, which indicates high correlation between the reference method values and the spectral data (Fig. 1a, 1b and 1c). 1-VR values were moderate in all the cases, being the GSL equation which showed the higher 1-VR value and the lower SECV. Predicting ability for these equations was similar for GSL and GLA (2.85 and 2.23, respectively) and lower for SIN (1.95) based on the standard deviation to SECV ratio.

Table 2. Calibration and cross-validation statistics for sinigrin, gluconapin and total glucosinolate content (μmol·g^-1 dry wt.).

Std= standard deviation; R²= coefficient of determination; SEC= standard error of calibration; 1-VR= percentage of variation in HPLC values explained by NIRS; SECV= standard error of cross-validation; RSC= ratio Std/SECV; SIN= sinigrin; GLA= gluconapin; GSL= total glucosinolate content.

GSL (LAB) GNA (LAB) SIN (LAB)

Fig. 1: Laboratory vs. predicted values calibration plots for a) total glucosinolate content b) gluconapin and c) sinigrin

CONCLUSIONS

The purpose of developing these NIRS equations was to have a rapid screening way for estimating individual and total glucosinolate contents in intact seed of Indian mustard. The results obtained showed that NIRS can be used as a screening technique to evaluate large number of samples in a short period of time. Once the screening process is done, HPLC could be used to accurately determine these parameters in samples previously selected by NIRS, avoiding the need of the analyses.

Future research should aim to improving the accuracy of the prediction by adding new samples to the calibration set, and by optimizing sample presentation.

ACKNOWLEDGEMENTS

The authors thank technician Gloria Fernández Marín, Instituto de Agricultura Sostenible (CSIC), Córdoba, Spain, for her excellent technical assistance in obtaining HPLC reference values.

This work has been supported by the C.I.C.Y.T. (Proj. AGF95-0222) of the Spanish Government.

REFERENCES

1. Fereres, E.,. Fernández-Martínez, J.M., Minguez, Y., Domínguez, J. 1983. Productivity of Brassica juncea and B. carinata in relation to rapeseed. Proc. 6th Int. Rapeseed Congr., Paris, France, 293-298.

2. Font, R., Del Río, M., Fernández-Martínez, J.M., De Haro, A. 1998a. Determining quality components in Indian mustard by NIRS. Cruciferae Newsletter. 20, 67-68.

3. Font, R., Del Río, M., Fernández-Martínez, J.M., De Haro, A. 1998b. Evaluation of Brassica juncea germplasm for quality components. Eucarpia. International Symposium on Breeding of Protein and Oil Crops. Pontevedra (Spain). European Association for Research on Plant Breeding. pp. 119-120.

4. Gland, A., Robbelen, G., Thies, W. 1981. Variation of alkenyl glucosinolates in seeds of Brassica species. Zeitschrift fuer Pflanzenzuechtung. 87, 96-110.

5. Mark, H., Workman, J. 1991. Statistics in Spectroscopy. Academic Press Inc.

6. Palmer, M.V., Yeung, S.P., Sang, J.P. 1987. Glucosinolate content of seedlings, tissue cultures and regenerant plants of Brassica juncea. Journal of Agricultural and Food Chemistry. 35, 262-265.

7. Palmer, M.V., Sang, J.P., Oram, R.N., Tran, D.A., Salisbury, P.A. 1988. Variation in seed glucosinolate concentrations of Indian mustard (Brassica juncea (L.) Czern. + Coss.). Australian Journal of Experimental Agriculture. 28, 779-782.

8. Windham, W.R., Mertens, D.R., Barton II, F.E. 1989. Protocol for NIRS calibration: Sample selection and equation development and validation. In Near infrared reflectance spectroscopy (NIRS): Analysis of forage quality. G.C. Marten et al., (ed.). Agric. Handb. 643. USDA-ARS, Washington, DC.