Quantification of Glucosinolates in Seeds by Solid-State ¹³C-Nuclear Magnetic Resonance (NMR)

Abstract

Solid-state ¹³C NMR spectroscopy using cross-polarization magic-angle spinning is a highly valuable technique for the semi-quantitative analysis of complex solid matrices. One of its key advantages is that it does not require any manipulation of the matrix, such as extractions or other treatments, which is particularly important for preserving the integrity of unstable secondary metabolites. Glucosinolates (β-thioglucoside-N-hydrosulphates) are crucial secondary metabolites specific to Brassica species, and many of them are known to be highly unstable. In this study, we evaluated solid-state nuclear magnetic resonance spectroscopy as an alternative method for the identification and quantification of total glucosinolates in the seeds of Sisymbrium officinale, Brassica napus, Sinapis alba, Brassica nigra, and Moringa oleifera. The results obtained with this method showed good agreement with those from conventional chemical analyses of the seed material. Although, based on a limited number of samples, this preliminary study suggests that the proposed approach could be a useful alternative for quantifying total glucosinolate content in seeds.

1. Introduction

Glucosinolates (GLSs) are a class of secondary metabolites predominantly found in plants of the family Brassicaceae, such as broccoli (Brassica oleracea var. italica), cauliflower (B. oleracea var. botrytis), and mustard (Sinapis alba). These compounds, along with their enzymatic degradation products, particularly isothiocyanates, have raised considerable scientific interest due to their bioactive properties, most notably their potential role in cancer chemoprevention. This interest is supported by epidemiological evidence indicating a correlation between the consumption of Brassica vegetables and a reduced incidence of various cancers [1,2,3].

Among the GLSs, sinigrin (2-propenyl-glucosinolate) is one of the most extensively studied. Found abundantly in various Brassicaceae, it is often used as a reference standard in analytical studies due to its commercial availability in potassium salt form. The bioactivity of sinigrin is primarily attributed to its hydrolysis product, allyl-isothiocyanate, which has demonstrated anticancer, antimicrobial, antioxidant, and anti-inflammatory effects [4]. Notably, sinigrin can also generate alternative products—such as nitriles, thiocyanates, and epithionitriles—depending on pH, iron ions, and the presence of specific proteins.

In addition to their relevance to human health, GLSs also play a key role in plant–insect interactions, acting as either feeding deterrents or attractants.

From a chemical standpoint, glucosinolates are β-D-thioglucoside-N-hydroxysulfates biosynthetically derived from amino acids. In their intact form, glucosinolates are biologically inert. However, mechanical damage to plant tissues leads to the rupture of cellular compartments, bringing glucosinolates into contact with the enzyme myrosinase (a thioglucoside glucohydrolase), which rapidly hydrolyzes the thioglucosidic bond. This enzymatic reaction releases glucose and a sulfated thiohydroximate intermediate, which spontaneously rearranges—via a Lossen-like mechanism—into isothiocyanates (ITCs) and other biologically active compounds [5,6].

GLSs can be classified according to the nature of their side chain (R-group) into aliphatic, indolic, or aromatic types [7]. Alternatively, they may be grouped based on their amino acid precursors, a classification that more accurately reflects their biosynthetic origin and structural relationships [8]. After degradation, three principal groups can be distinguished: stable isothiocyanates (ITCs), thiocyanates, and oxazolidine-2-thione compounds. A comprehensive review by Blažević et al. has recently outlined the structural diversity of GLSs across plant species, along with their associated biochemical properties and key analytical and synthetic methodologies [9].

Several analytical techniques have been developed for the identification and quantification of glucosinolates (GLSs) in plant matrices, including chromatographic, spectroscopic, enzymatic, and microchip-based approaches [10]. Early studies on GLS separation date back to the early 20th century, with initial attempts employing paper chromatography.

Currently, analytical methods can be broadly divided into two categories based on the nature of the analyte: destructive and non-destructive techniques. Destructive methods involve chemical or enzymatic hydrolysis of the thioglucosidic bond, followed by the quantification of the resulting degradation products [11]. Although widely used [12], these approaches are often labor-intensive and depend on extraction protocols that can affect GLS stability and yield. In particular, exposure to high temperatures promotes GLS degradation, making low-temperature procedures preferable for preserving their native structure [13,14]. Non-destructive approaches, on the other hand, enable the direct analysis of intact GLSs [15], offering a more streamlined workflow.

Among non-destructive techniques, spectroscopic methods such as near-infrared reflectance spectroscopy [16] and X-ray fluorescence spectroscopy [17] allow for rapid, extraction-free quantification. However, they may be less sensitive or specific than chromatographic methods.

Nuclear magnetic resonance (NMR) spectroscopy, first applied to GLS analysis in 1967 [18], represents a powerful, non-destructive tool capable of providing detailed structural information. It is particularly valuable for the qualitative and quantitative assessment of GLSs that are difficult to detect using conventional methods [19].

Solid-state NMR spectroscopy represents an advanced analytical technique for the quantitative analysis of chemical species in solid matrices. Although based on the same physical principles as solution-state NMR, its application to solids is complicated by the absence of molecular tumbling. This absence enhances anisotropic interactions such as chemical shift anisotropy and dipolar couplings. These effects historically limited the resolution and quantitative applicability of the method [20].

The advent of magic-angle spinning (MAS) and cross-polarization (CP) significantly improved spectral quality by attenuating anisotropic interactions and enhancing signal intensity, respectively. The combined CP-MAS technique enables the analysis of complex solid samples without the need for extraction, a major advantage when dealing with labile compounds such as certain glucosinolates [12].

In this study, we evaluated the use of solid-state CP-MAS ¹³C-NMR as a non-destructive method for the identification and quantification of total glucosinolates in food plant seeds. The choice to focus exclusively on seeds was based on their typically higher glucosinolate content compared to other plant tissues, such as leaves or roots. This makes them particularly suitable for evaluating the sensitivity and applicability of the solid-state NMR approach.

Results were compared with those obtained via ultra-performance liquid chromatography (UPLC). To our knowledge, this is the first report demonstrating the applicability of solid-state CP-MAS ¹³C-NMR for GLS quantification, supporting its potential as a robust and extraction-free analytical alternative.

2. Materials and Methods

2.1. Chemicals

Sinigrin monohydrate potassium salt, methanol-d₄ (isotopic 99.8%), DSS (3-(trimethylsilyl)-1-propanesulfonic acid sodium salt), cyclohexanone oxime, pure adamantane, and NaCl were purchased from Sigma-Aldrich (Milan, Italy).

2.2. Plant Material

Sisymbrium officinale (L.) Scop. was cultivated in a greenhouse at the Department of Agricultural and Environmental Sciences—Production, Landscape and Agroenergy of the University of Milan. Plants were harvested in 2018, and seeds were collected in September of the same year. Brassica napus L. seeds, used as a certified reference material (ERM-BC367 RAPESEED), were purchased from Sigma-Aldrich (Milan, Italy). Seeds of Sinapis alba L. and Brassica nigra (L.) W.D.J.Koch were commercially available products acquired from a local supermarket. Moringa oleifera Lam. seeds, originating from Haiti, were kindly provided as a gift by Professor Franco Sangiorgi.

2.3. CP-MAS ¹³C NMR

The CP-MAS ¹³C NMR spectra were recorded using a Bruker Avance 600 MHz spectrometer (Bruker GmbH, Mannheim, Germany) at 298 K. The spectra were acquired with a standard cross-polarization (CP) pulse sequence featuring a contact time of 1 ms and an acquisition time of 16 h. The rotor spinning rate was set to 10 kHz, and magic-angle spinning was conducted at 150.91 Hz. Cylindrical zirconium dioxide rotors with a 7 mm diameter were used as sample holders.

To optimize the signal-to-noise ratio, an appropriate number of scans were performed for each experiment. A line broadening (LB) of +50 Hz was applied to transform all free induction decays using a transform size of 4K. Baseline flattening and phase correction were employed to enhance the accuracy of the NMR integrals. All spectra were manually phased and referenced to the DSS peak at 0 ppm. Data processing, including phase correction, baseline correction, integration, and peak picking, was carried out using TopSpin v. 3.1 software (Billerica, MA, USA).

CP-MAS experiments were conducted using various contact times to determine the optimal value, minimizing quantification errors. Glucosinolate (GLS) quantification requires a standard reference that provides a known number of ¹³C nuclei for peak integration. The most used reference was 3-(trimethylsilyl)-1-propanesulfonic acid sodium salt (DSS). The experiments were performed in duplicate, and inter-day reproducibility tests were conducted on selected samples. These tests yielded consistent integral values, with deviations remaining within acceptable limits (<5%). The analyses were conducted on two biological replicates (n = 2), with each sample subjected to four technical replicates to ensure measurement reliability.

In order to calculate the GLS content, the integral of the characteristic GLS signal (SC = N, 156–160 ppm) was normalized vs. the internal standard DSS, according to the following Equation (1):

GLS content = mmol CH_{3 DSS} × (I SC = N/100) × 1000 µmol/mmol/g DW

(1)

where mmol CH_{3 DSS} values are the mmols of the C of three methyl groups of the DSS signal (0 ppm, normalized at 100), I SC = N is the integrated signal of GLS, and g DW is the g of the dried weight of the sample.

The NMR sample preparation did not require solvent extraction; instead, the dried sample was finely ground, mixed with the internal standard, and directly analyzed by NMR. The chemical shifts of the glucosinolate carbons are clearly distinguishable in the 157–160 ppm range, a region with minimal signal overlap. Signal assignments were initially confirmed using a pure reference compound, sinigrin, analyzed by both solid0 and liquid-state NMR.

Sinigrin was used as a standard for GLS quantification. The ¹³C liquid-state NMR spectrum of sinigrin was similar to that of solid-state NMR (Figure 1).

A good correlation was observed between the two spectra, with acceptable signal broadening in the solid-state spectrum due to the presence of a single molecule without matrix effects. Specifically, the quaternary carbon of sinigrin resonated at 160.5 ppm in solution and at 162.0 ppm in solid-state NMR, a chemical shift that is distinct and unambiguous for this compound. Solid-state NMR analyses were performed in duplicate, and the data are expressed as the mean of the two determinations.

For solid samples, accurate weighing was performed before adding approximately 10% DSS by weight, followed by thorough grinding in a mill (MM400, Retsch GmbH, Haan, Germany).

2.4. Extraction and Glucosinolates Quantification by UPLC

The extraction and quantification of glucosinolates by UPLC are described in detail in the Supporting Information.

The MS/MS spectra of glucosinolates showed the presence of typical product ions with (m/z)⁻ 97 Da corresponding to the sulfate moiety.

Neoglucobrassicin and 4-methoxyglucobrassicin showed a typical UV spectrum, identical parent (m/z 477), and product ions (m/z 97); thus, they were differentiated by comparison with a reported elution sequence during RP-LC. Alkyl-glucosinolates like glucoiberin, progoitrin, and sinigrin were not well separated in RP-LC due to their high polarity. On the other hand, the successful separation of these compounds was achieved by LC-MS/MS with MRM detection; thus, the partial peaks’ overlap did not affect the quantification of these compounds.

The following fragmentation transitions for the multiple reaction monitoring (MRM) were used, with a dwell time of 0.2 s per transition: (m/z) 358 to 97 (sinigrin), 388 to 97 (progoitrin), 422 to 97 (glucoiberin), 436 to 97 (glucoraphanin), 447 to 97 (glucobrassicin), 463 to 97 (4-hydroxy-glucobrassicin), 477 to 97 (4-methoxy-glucobrassicin and neoglucobrassicin), 570 to 97 (glucomoringin) and 390 to 97 (hydroxybutyl-glucosinolate). Calibration curves were obtained from sinigrin stock solutions prepared by dissolving 5 mg of standard powder in 50 mL methanol. The working solutions were prepared in 0.1% aqueous formic acid in the range of 0.02–10 μg/mL. GLSs were assayed using sinigrin calibration curves, and their amounts were normalized by the molecular mass ratios.

3. Results

CP-MAS NMR offers valuable insights into the molecular structure of complex matrices. Parameters such as peak shape, intensity, and chemical shift are instrumental for analyzing molecular structures and chemical compositions. Furthermore, peak areas—interpreted as relative intensities—can be used to quantify molecules in a mixture by integrating the signals and comparing them with those of NMR standards.

We used Brassica napus L. seeds, a certified reference material for GLS content, as a standard for GLS quantification in our samples. The total GLS content in this reference material corresponded to 99 mmol/kg (99 µmol/g dry weight).

The sinigrin content in the sinigrin/DSS mixture was verified using our protocol, and the results showed good agreement.

The NMR spectra of all the considered samples are reported in Figure 2 and Figures S1–S6.

Table 1. The total GLS content in seeds determined by ¹³C CP-MAS and UPLC. Concentrations are expressed per gram of dry weight. The standard deviation for CP-MAS quantifications is not reported due to the insufficient number of replicates for a meaningful estimation of variability.

Seeds	Chemical Shift 13C Thiocyanate (δ, ppm)	CP-MAS (µmol/g DW)	UPLC (µmol/g DW)
Brassica napus	157.3	116	113 ± 5	113
Sisymbrium off.	157.6	110	20 ± 10	20
Sinapis alba	157.2	156	152 ± 17	156
Brassica nigra	157.2	67	69 ± 10
Moringa oleifera	153.7	60	32 ±11

summarizes the results obtained from both solid-state NMR and UPLC analyses of the seed samples. Only two samples—Sisymbrium officinale and Moringa oleifera—showed discrepancies in GLS content between the two analytical techniques, warranting further investigation. In contrast, the results obtained for the other seed samples showed good consistency across the two independent methods.

The GLSs detected in the various samples are summarized below:

• Brassica napus: Progoitrin and gluconapin were the predominant compounds, along with smaller amounts of glucoalyssin, glucobrassicin, and napoleiferin;
• Sisymbrium officinale: glucoputranjivin, glucobrassicin, and 4-methoxyglucobrassicin were detected, along with lesser amounts of glucocochlearin and glucosinalbin;
• Sinapis alba: sinalbin, sinigrin, and glucobrassicin were identified, while neoglucobrassicin, gluconasturtiin, and 4-methoxyglucobrassicin were found in trace amounts;
• Brassica nigra: sinigrin, gluconapin, and gluconasturtiin were present, with 4-hydroxyglucobrassicin detected in trace amounts;
• Moringa oleifera: glucosinalbin-rhamnoside (glucomoringin), gluconaringin, and a lesser amount of acetylated glucomoringin were also detected.

4. Discussion

Solid-state NMR is applied in the biological and environmental field to study membrane proteins, fungi, algae, and plant biomass, including cellulose and lignin [21,22,23]. Recently Xue et al. [24] employed advanced solid-state nuclear magnetic resonance spectroscopy to analyze bark samples from various species. However, to the best of our knowledge, no recent scientific study has performed solid-state NMR for quantitative analysis on seeds.

Quantification data for GLSs indicated that only two samples—Sisymbrium officinale and Moringa oleifera—exhibited discrepancies in GLS content between the two analytical techniques. Further investigation is required to clarify the underlying causes. This observation raises the possibility that the higher GLS content detected by NMR may reflect either an actual overestimation or a loss of analytes during the extraction procedures required by conventional methods.

In contrast, data from the remaining samples demonstrated good consistency between the two independent analytical approaches, particularly for seed-derived materials.

Several challenges were encountered during the analysis, including factors related to the solid nature of the matrix, low GLS concentrations in certain plant tissues (such as leaves and roots), and intrinsic molecular characteristics. For instance, the quaternary carbon of the thiocyanate group yielded weak and broad signals, complicating accurate integration. Seeds, which typically contain high levels of glucosinolates, produced more intense and readily integrable NMR signals, potentially explaining the greater consistency observed for seed samples.

Nonetheless, the apparent overestimation in some cases may also result from interference by other chemical compounds resonating within the same spectral region, which have yet to be identified. One hypothesis is that these interfering substances may be carbamates. Additional studies will be necessary to confirm this.

5. Conclusions

This study provides the first evidence supporting the use of solid-state CP-MAS ¹³C-NMR as a non-destructive technique for the identification and quantification of total glucosinolates (GLSs) in food plant seeds. The method, which does not require extraction or derivatization steps, was evaluated through direct comparison with UPLC analysis. Overall, a good correlation was observed between the two techniques, particularly for seed samples characterized by high GLS content, where signal intensity and resolution were sufficient to ensure accurate quantification.

The discrepancies identified—limited to Sisymbrium officinale and Moringa oleifera—may arise either from actual NMR overestimation or from partial loss of analytes during the extraction procedures required for chromatographic analysis. Although further studies are needed to clarify these aspects, the present findings highlight the critical role of matrix composition, compound abundance, and signal quality in determining analytical reliability.

Despite these limitations, solid-state NMR shows significant potential as a complementary analytical tool for glucosinolate profiling, particularly in contexts where sample integrity must be preserved or rapid screening is desirable. Its application could be further extended to the characterization of other plant-derived matrices, provided that sensitivity and selectivity constraints are adequately addressed.