Determination of glucosinolates and isothiocyanates in glucosinolate-rich vegetables and oilseeds using infrared spectroscopy: A systematic review

Figures & data

Figure 1. General structure of glucosinolates and isothiocyanates and their conversion. The R group is a varying aglycone amino acid-deriving side chain. Glucosinolates can be classified as aliphatic, indolic or aromatic glucosinolates based on the R group.

Figure 2. Development of infrared spectroscopy-based calibration models for determination of compounds as glucosinolates and isothiocyanates. The images used in this figure are copyright free or have been downloaded from BioRender.com.

Table 1. Summary of the NIR calibration models used to determine glucosinolate content in vegetables and oilseeds.

Source	Method	Wavelength range (nm)	Reference technique	Compounds	Number of samples used for calibration (without outliers)	Spectra preprocessing method	Regression method	Min (μmol/g DW)	Max (μmol/g DW)	Average (μmol/g DW)	SD Ref (μmol/g DW)	Calibration	Validation					Reference
													Internal validation			External validation
												R^2	R^2	SECV	RPD	R^2	RPD
Vegetables
Brassica oleracea (broccoli)	NIR	400 -2500	HPLC	Glucoiberin	62	–	MPLS	0.02	0.36	0.12	0.13	0.38	0.31	0.20	0.67			(Sahamishirazi et al. 2017)
				Glucosinigrin	64			0.36	0.37	0.37	0.004	0.44	0.05	0.01	0.62
				Glucoraphanin	67			0.03	2.87	1.10	0.70	0.71	0.66	0.54	1.63
				Glucobrassicin	64			0.21	0.73	0.55	0.24	0.24	0.11	0.30	0.81
				4-methoxyglucpbrassicin	67			0.41	0.45	0.43	0.01	0.34	0.21	0.01	1.11
				1-methoxyglucobrassicin	63			0.02	0.72	0.40	0.39	0.25	0.03	0.68	0.56
				Indolic glucosinolates	66			0.19	3.16	1.56	0.80	0.50	0.26	0.85	0.95
				Aliphatic glucosinolates	67			0.21	4.00	1.59	0.90	0.76	0.64	0.58	1.65
				Total glucosinolates	66			0.43	6.00	3.27	1.50	0.69	0.55	1.17	1.36
Brassica oleracea (broccoli)	VIS/NIR	2000 − 2500	MECC	Glucobrassicin	46	De-trending 1, 4, 4, 1	MPLS	1.60	6.83	3.93	1.09	0.89						(Hernández-Hierro et al. 2012)
				Glucoraphanin	47	De-trending 1, 4, 4, 1		1.09	4.82	2.86	0.87	0.40
				4-methoxyglucobrassicin	48	SNV 1, 4, 4, 1		0.09	0.55	0.28	0.11	0.69
				Neoglucobrassicin	48	Multiplicative scatter correction 2, 8, 6, 1		0.74	4.45	1.93	0.86	0.68
				Total glucosinolates	45	None 2, 10, 10, 1		4.74	15.10	9.40	1.87	0.73
Brassica albograbra Baiely (Chinese kale)	NIR	400-2500	HPLC	Total glucosinolates	141	SNV, de-trending, and algorithms (2,5,5,2)	MPLS			1.42	0.84	0.91	0.92	2.96	3.59			(Chen et al. 2014)
				Progoitrin	137					1.89	1.36	0.34	0.69	0.36	2.32
				Sinigrin	141					2.30	1.97	0.86	0.83	0.55	2.46
				Glucoraphanin	142					10.88	5.19	0.67	0.84	0.62	3.18
				Gluconapin	142					0.12	0.12	0.82	0.87	1.87	2.78
				4-hydroxyglucobrassicin	142					2.30	4.68	0.64	0.82	0.05	2.70
				Glucobrassicin	145					0.14	0.08	0.93	0.93	1.22	3.84
				4-methoxyglucobrassicin	142					0.80	1.62	0.58	0.71	0.04	1.96
				Neoglucobrassicin	141					19.83	10.63	0.90	0.86	0.44	3.70
Eruca vesicaria (rocket) leaves	VIS/NIR	400-2500	LC-UV	Total glucosinolates	30	Second derivative treatment (2,5,5,2) with SNV-DT	MPLS	4.86	44.65	23.56	7.32	0.79	0.70	4.02	1.83	0.61	1.59	(Toledo-Martin et al. 2017)
				Glucoraphanin	30			0.13	27.85	9.11	6.41	0.94	0.82	2.72	2.37	0.79	2.48
				Glucosativin	30			0.25	18.61	9.66	3.92	0.86	0.64	2.42	1.62	0.60	1.92
				Glucoerucin	30			0.14	8.01	1.45	1.60	0.97	0.93	0.41	3.99	0.59	1.56
Cabbage and Brussels sprout leaf tissues (fresh)	NIR/REF	950-1650	HPLC	Glucobrassicin	92	First derivative treatment	PLS	3.69	379.16	65.16	80.01	0.89	0.76		2.6			(Renner and Fritz 2020)
Cabbage and Brussels sprout leaf tissues (dry)	NIR/REF	950-1650	HPLC	Glucobrassicin	92	First derivative treatment	PLS	0.41	22.25	5.33	4.95	0.90	0.79		2.4			(Renner and Fritz 2020)
Oilseeds
Brassica juncea (brown mustard)	VIS/NIR	400-2500	HPLC	Singrin	377	Norris method, derivative method (2,5,5,2), and SNV-DT	MPLS	0.00	132.44	30.90	33.85	0.99	0.97	6.33	5.35			(Gohain et al. 2021)
				Gluconapin	379			0.00	171.47	70.73	33.58	0.95	0.91	9.96	3.37
				Glucobrassicanapin	378			0.00	2.49	0.76	0.58	0.65	0.91	0.47	1.24
				Glucoiberin	390			0.00	13.18	3.30	3.29	0.98	0.34	1.36	2.42
				Glucoraphanin	377			0.00	36.54	3.87	10.89	0.88	0.84	5.49	1.98
				Glucoalyssin	390			0.00	2.87	1.04	0.61	0.92	0.74	0.35	1.74
				Glucobrassicin	380			0.00	5.17	0.71	1.49	0.75	0.67	0.98	1.51
				4-hydroxy glucobrassicin	387			0.00	23.58	5.85	5.91	0.96	0.56	1.81	3.27
				4-methoxy-glucobrassicin	386			0.00	0.93	0.19	0.25	0.67	0.91	0.18	1.38
				Total glucosinolates	385			15.85	219.00	117.43	33.86	0.87	0.47	13.12	2.58
				Aliphatic glucosinolates	374			0.00	220.94	109.89	37.02	0.90	0.85	13.02	2.84
				Total indolic glucosinolates	383			0.00	30.83	6.90	7.97	0.97	0.88	2.30	3.47
Brassica juncea (brown mustard)	NIR/REF	400-2500	Chemical assay followed by ELISA	Total glucosinolates	240	Derivatives (1,4,4,1), multiple scatter correction, and smoothing	MPLS	28.10	114.04	73.31	6.55	0.939	0.865			0.951		(Sen et al. 2018)
Brassica napus (rapeseed)	NIR	900-1680	HPLC	Total glucosinolates	77	SNV and SG second derivative transformation	KPLS	4.9	23.7	11.3	3.6	0.948	0.533	2.4				(Barthet, Petryk, and Siemens 2020)
Brassica napus (rapeseed)	NIR/REF	400-2500	Chemical assay followed by ELISA	Total glucosinolates	240	Derivatives (1,4,4,1), multiple scatter correction, and smoothing	MPLS	15.49	139.09	70.26	18.86	0.865	0.866			0.986		(Sen et al. 2018)
Brassicaceae seeds	NIR	400-2500	HPLC	Total glucosinolates	124	Derivatives (1,4,4,1) and SNV/de-trend scatter correction	MPLS	5.5	117.4	44.2	29.2	0.92			2.3			(Oblath et al. 2016)

Abbreviations: Infrared techniques—NIR, near-infrared spectroscopy; VIS/NIR, visible/near-infrared spectroscopy; NIR/REF, near-infrared reflectance spectroscopy. Reference techniques—HPLC, high performance liquid chromatography; MECC, micellar electrokinetic capillary chromatography; LC-UV, liquid chromatography with ultraviolet absorbance detection; ELISA, enzyme-linked immunoassay reader. Pretreatments: SNV, standard normal variate; SNV-DT, standard normal variate-detrend; SG, Savitzky-Golay. Regression methods: PLS, partial least square; MPLS, modified partial least square; KPLS, Kernel partial least square. Statistical analysis: R², coefficient of determination; SECV, ratio of the standard error of cross-validation; RPD, residual prediction deviation; RER, range error ratio, SD Ref: standard deviation reference, Max; maximum value; Min; minimum value; WV; wavelength.

Table 2. Specific wavelengths in the NIR region reported to measure total glucosinolates in leaves and plants as reported by different authors.

Bond	Wavelengths (nm)
	Rocket leaves (Eruca vesicaria) (Toledo-Martin et al. 2017)	Leaf rape (Brassica napus ssp. pabularia) (Font et al. 2005)	Cabbage (Brassica oleracea var. capitata f. alba) and brussel sprouts (Brassica oleracea var. gemmifera) (Renner and Fritz 2020)
O–H stretching	1432	1434	1440-1470
N–H asymmetric stretching of indole			1450
S–H stretch first overtone or C–H stretch first overtone of CH3 groups	1696	1694
C–H stretching by methylene groups	1730	1728, 1764
O-H stretch plus O-H deformation of water	1920	1922
N-H stretch of amides	2054	2056
O-H plus C-C stretch groups of cellulose	2270	2270

Table 3. Specific wavenumbers in the MIR region used to measure isothiocyanate (i.e., NCS group) as reported by different authors.

Bond		Wavenumbers (cm^-1)
		Methyl isothiocyanate (Zheng et al. 2007)	Germyl isothioc yanate (Zheng et al. 2007)	2-Methoxyphenyl isothiocyanate (Yenagi, Nandurkar, and Tonannavar 2012)
Asymmetric stretching of -N=C=S group	1990 − 2150	2083	2047	2034, 2112
Symmetric N-C-S stretch	925 - 1250	680	959	931
Deformation nodes perpendicular to -NCS plane (out-of-plane)	520 - 570	417	464	539
Deformation nodes parallel to -NCS plane (in-plane)	425 - 440	535	464	472

Table 4. Characteristic bands in the MIR region reported to measure pure glucosinolates and isothiocyanates.

Compound	Molecular formula	Wavenumbers of key absorbance bands (cm^−1)	Reference
Glucosinolates
(R,S)S-Glucoraphanin	C12H23NO10S3	3316 (OH), 2976, 2868, 1651 (C=N), 1495, 1265, and 1063	(Vo et al. 2013)
α-Neoglucobrassicin	C17H22N2O10S2	3125, 2936, 1714, 1575, 1453, 1242, and 1060	(Vo et al. 2018)
α-4-Methoxyglucobrassicin	C17H22N2O10S2	3454, 3267, 2908, 2842, 1260, and 1060	(Vo et al. 2018)
Glucobretschneiderin	C15H20NO11S2	3373 (OH), 2915 (CH), 1658 (CN), 1059 (C–O), 801 (sulphate), and 668 (C–S)	(Montaut et al. 2015)
Isothiocyanates
Sulforaphane	C6H11NOS2	3426 (O-H from H2O adsorbed), 2923, 2867 (C-H), 2179, 2100 (N = C=S), 1451, 1349 (C-H), 1260 (C-N), 1021 (S = O), 739 (C-H), and 688 (C-S)	(De Nicola et al. 2014)
Iberverin	C5H9NS2	3054, 2925, 2106, 1264, and 736	(Ernst et al. 2013)
Iberin	C5H9NOS2	2932, 2179, 2087, 1444, 1348, 1013, and 690	(Ernst et al. 2013)
Cheirolin	C5H9NO2S2	3002, 2923, 2180, 2110, 1442, 1298, 1264, 1131, 1016, and 764	(Ernst et al. 2013)

Sahamishirazi, S., S. Zikeli, M. Fleck, W. Claupein, and S. Graeff-Hoenninger. 2017. Development of a near-infrared ­spectroscopy method (NIRS) for fast analysis of total, indolic, ­aliphatic and individual glucosinolates in new bred open pollinating genotypes of broccoli (Brassica oleracea convar. botrytis var. italica). Food Chemistry 232:272–7. doi: 10.1016/j.foodchem.2017.04.025.

 PubMedGoogle Scholar

Hernández-Hierro, J. M., J. Valverde, S. Villacreces, K. Reilly, M. Gaffney, M. L. Gonzalez-Miret, F. J. Heredia, and G. Downey. 2012. Feasibility study on the use of visible-near-infrared spectroscopy for the screening of individual and total glucosinolate contents in broccoli. Journal of Agricultural and Food Chemistry 60 (30):7352–8. doi: 10.1021/jf3018113.

 PubMed Web of Science ®Google Scholar

Chen, J., L. Li, S. Wang, X. Tao, Y. Wang, A. Sun, and H. He. 2014. Assessment of glucosinolates in Chinese kale by near-infrared spectroscopy. International Journal of Food Properties 17 (8):1668–79. doi: 10.1080/10942912.2012.678535.

 Web of Science ®Google Scholar

Toledo-Martin, E. M., R. Font, S. Obregon-Cano, A. D. Haro-Bailon, M. Villatoro-Pulido, and M. Del Rio-Celestino. 2017. Rapid and cost-effective quantification of glucosinolates and total phenolic content in rocket leaves by visible/near-infrared spectroscopy. Molecules 22 (5):851. doi: 10.3390/molecules22050851.

 PubMedGoogle Scholar

Renner, I. E., and V. A. Fritz. 2020. Using Near-infrared reflectance spectroscopy (NIRS) to predict glucobrassicin concentrations in cabbage and brussels sprout leaf tissue. Plant Methods 16:136. doi: 10.1186/s13007-020-00681-7.

 PubMed Web of Science ®Google Scholar

Gohain, B., P. Kumar, B. Malhotra, R. Augustine, A. K. Pradhan, and N. C. Bisht. 2021. A comprehensive Vis-NIRS equation for rapid quantification of seed glucosinolate content and composition across diverse Brassica oilseed chemotypes. Food Chemistry 354:129527. doi: 10.1016/j.foodchem.2021.129527.

 PubMedGoogle Scholar

Sen, R., S. Sharma, G. Kaur, and S. S. Banga. 2018. Near-infrared reflectance spectroscopy calibrations for assessment of oil, phenols, glucosinolates and fatty acid content in the intact seeds of oilseed Brassica species. Journal of the Science of Food and Agriculture 98 (11):4050–7. doi: 10.1002/jsfa.8919.

 PubMedGoogle Scholar

Barthet, V. J., M. W. P. Petryk, and B. Siemens. 2020. Rapid nondestructive analysis of intact canola seeds using a handheld near‐­infrared spectrometer. Journal of the American Oil Chemists’ Society 97 (6):577–89. doi: 10.1002/aocs.12335.

 Google Scholar

Oblath, E. A., T. A. Isbell, M. A. Berhow, B. Allen, D. Archer, J. Brown, R. W. Gesch, J. L. Hatfield, J. D. Jabro, J. R. Kiniry, et al. 2016. Development of near-infrared spectroscopy calibrations to measure quality characteristics in intact Brassicaceae germplasm. Industrial Crops and Products 89:52–8. doi: 10.1016/j.indcrop.2016.03.022.

 Google Scholar

Font, R., M. del Rio-Celestino, E. Cartea, and A. de Haro-Bailon. 2005. Quantification of glucosinolates in leaves of leaf rape (Brassica napus ssp. pabularia) by near-infrared spectroscopy. Phytochemistry 66 (2):175–85. doi: 10.1016/j.phytochem.2004.11.011.

 PubMed Web of Science ®Google Scholar

Zheng, C., G. A. Guirgis, H. Deeb, and J. R. Durig. 2007. On the structural parameters and vibrational spectra of CH3NCS, SiH3NCS and GeH3NCS. Journal of Molecular Structure 829 (1–3):88–110. doi: 10.1016/j.molstruc.2006.06.011.

 Google Scholar

Yenagi, J., A. R. Nandurkar, and J. Tonannavar. 2012. 2-Methoxyphenyl isocyanate and 2-methoxyphenyl isothiocyanate: Conformers, vibration structure and multiplet fermi resonance. Spectrochimica Acta, Part A, Molecular and Biomolecular Spectroscopy 91:261–8. doi: 10.1016/j.saa.2012.01.024.

 PubMedGoogle Scholar

Vo, Q. V., C. Trenerry, S. Rochfort, and A. B. Hughes. 2013. A total synthesis of (R, S) S -glucoraphanin. Tetrahedron 69 (41):8731–7. doi: 10.1016/j.tet.2013.07.097.

 Google Scholar

Vo, Q. V., S. Rochfort, P. C. Nam, T. L. Nguyen, T. T. Nguyen, and A. Mechler. 2018. Synthesis of aromatic and indole alpha-glucosinolates. Carbohydrate Research 455:45–53. doi: 10.1016/j.carres.2017.11.004.

 PubMed Web of Science ®Google Scholar

Montaut, S., W. D. Zhang, J. M. Nuzillard, G. R. De Nicola, and P. Rollin. 2015. Glucosinolate diversity in bretschneidera sinensis of Chinese origin. Journal of Natural Products 78 (8):2001–6. doi: 10.1021/acs.jnatprod.5b00338.

 PubMed Web of Science ®Google Scholar

De Nicola, G. R., P. Rollin, E. Mazzon, and R. Iori. 2014. Novel gram-scale production of enantiopure R-sulforaphane from Tuscan black kale seeds. Molecules (Basel, Switzerland) 19 (6):6975–86. doi: 10.3390/molecules19066975.

 PubMed Web of Science ®Google Scholar

Ernst, I. M., K. Palani, T. Esatbeyoglu, K. Schwarz, and G. Rimbach. 2013. Synthesis and Nrf2-inducing activity of the isothiocyanates iberverin, iberin and cheirolin. Pharmacological Research 70 (1):155–62. doi: 10.1016/j.phrs.2013.01.011.

 PubMed Web of Science ®Google Scholar