Antimicrobial isothiocyanates from Brassicaceae glucosinolates : Analysis, reactivity, and quantitative structure-activity relationships

The increased consumer awareness of health-associated risks of synthetic preservatives along with increasing antimicrobial resistance drive research towards exploring new natural antimicrobial compounds. The Brassicaceae plant family is a potential source of new antimicrobials, e.g. as food preservatives. These plants abundantly produce glucosinolates (GSLs), which serve as precursors of antimicrobial isothiocyanates (ITCs), formed upon contact of GSLs and myrosinase. GSLs as well as ITCs are structurally diverse, differing in side chain configuration. The main aim of this thesis research was to explore the potential of ITCs as natural antimicrobials. For this, the production, the analysis, the reactivity, and the (quantitative) structure-antimicrobial activity relationships ((Q)SAR) of ITCs were studied. In Chapter 1, the structural diversity, the occurrence, and the analysis of GSLs and ITCs were described. Furthermore, approaches to modulate structural diversity and content of GSLs in Brassicaceae seeds were explained. Lastly, the state-of-the-art of ITCs as new natural antimicrobial candidates was reviewed, also by considering their reactivity (i.e. electrophilicity). In Chapter 2, simultaneous fungal elicitation and germination of Brassicaceae seeds was studied as an attempt to increase content and structural diversity of GSLs. Compositional changes of aliphatic, benzenic, and indolic GSLs of Sinapis alba, Brassica napus, and B. juncea seeds by germination and fungal elicitation were determined. Rhizopus oryzae (non-pathogenic), Fusarium graminearum (non-pathogenic), and F. oxysporum (pathogenic) were employed. Thirty-one GSLs were detected by RP-UHPLC-PDA-ESI-MSn. Aromatic-acylated derivatives of 3-butenyl GSL, p-hydroxybenzyl GSL, and indol-3-ylmethyl GSL were for the first time tentatively annotated and confirmed not to be artefacts. Germination alone increased total GSL content in S. alba, mainly consisting of p-hydroxybenzyl GSL, by 2- to 3-fold. Meanwhile, germination alone did not significantly affect total GSL content in B. napus and B. juncea. Regardless of the pathogenicity of the fungi, fungal elicitation did not increase the content nor the structural diversity of GSLs further than what could be obtained by germination alone. In Chapter 3, a new method to simultaneously analyze various GSLs and ITCs by RP-UHPLC-ESI-MSn was described. The method was validated for 14 GSLs and 15 ITCs. It involved derivatization of ITCs with N-acetyl-L-cysteine (NAC). The limits of detection were 0.4−1.6 μM for GSLs and 0.9−2.6 μM for NAC-ITCs. The analysis of S. alba, B. napus, and B. juncea extracts spiked with 14 GSLs and 15 ITCs indicated that the method generally had good intraday (≤10% RSD) and interday precisions (≤16% RSD). Recovery of the method did not differ for each extract and was within 71−110% for GSLs and 66−122% for NAC−ITCs. The method was able to monitor the enzymatic hydrolysis of standard GSLs to ITCs in mixtures. Furthermore, GSLs and ITCs were simultaneously determined in Brassicaceae plant extracts before and after in vitro myrosinase treatment. In Chapter 4, the antimicrobial activity (minimum inhibitory concentration, MIC; minimum bactericidal/fungicidal concentration, MBC/MFC) of 10 ITCs against pathogenic and food spoilage Gram– bacteria, Gram+ bacteria, and fungi was determined. The activity of the long-chained (C9) 9-(methylthio)nonyl ITC (9-MTITC), 9-(methylsulfinyl)nonyl ITC (9-MSITC), and 9-(methylsulfonyl)nonyl ITC (9-MSoITC) was determined for the first time. Due to the electrophilicity of ITCs, the activity of ITCs was evaluated in nucleophile-rich and nucleophile-poor growth media. ITCs reacted with components in a nucleophile-rich growth medium at a rate of 39-141 µM/h, depending on their side chain configuration and temperature. The reaction rates were lowered by a factor of 2-21 when using nucleophile-poor growth media. Consequently, the activity of ITCs was generally improved, with MSITC and MSoITC being the most positively affected (activity increased by a factor of > 4). 9-MSITC and 9-MSoITC had good activity (MIC ≤ 25 µg/mL) against Gram+ bacteria and fungi. The short-chained (C3) analogues had good activity against Gram+ bacteria and Gram– bacteria. ITCs exhibit different antimicrobial activity, depending on their structure. Chapter 5 aimed at determining QSAR-based physicochemical properties of ITCs important for their antimicrobial activity and developing QSAR models to predict the activity of ITCs as individual and in mixtures. Twenty-six ITCs covering 9 subclasses were tested against Escherichia coli and Bacillus cereus. MIC (mM) and growth inhibition response (GIR, h/mM) were determined and used to develop QSAR models. The most active ITC subclasses were bifunctional ITC, MSITC, and MSoITC. MIC of the most active ITCs was 9.4-25 µg/mL against E. coli and 6.3-25 μg/mL against B. cereus. Multiple linear regression models were proposed with a good fit (R2adj 0.86–0.93) and high internal predictive power (Q2adj 0.80–0.89). Partial charge, polarity, shape, and reactivity were key physicochemical properties for antibacterial activity of ITCs. Furthermore, ITC compositions and antibacterial activity of Brassicaceae ITC-rich extracts were determined. B. oleracea ITC-rich extract, mainly comprising of short-chained (C3, C4) MSITCs, had MIC 750-1000 µg/mL against both bacteria, and Camelina sativa ITC-rich extract, mainly comprising of long-chained (C9, C10, C11) MSITCs, had MIC 188 µg/mL against B. cereus. Moreover, based on the ITC compositional analysis, the models successfully predicted the antibacterial activity of these extracts. In Chapter 6, the main findings of this thesis were discussed. The approaches in compositional analysis and antimicrobial assays used in this thesis were examined. Prospective external ITCs, to be tested for the external validation of the QSAR models, were given in detail with their predicted activity. Some hints on mode of action of ITCs based on important QSAR-based physicochemical properties and previous studies were elaborated for future studies. Finally, the prospects of ITCs as natural food preservatives, considering their electrophilicity, pungency and bitterness, toxicity, and potential synergisms with other antimicrobial compounds, were elaborated. In conclusion, the newly developed simultaneous RP-UHPLC-ESI-MSn analysis of various GSLs and ITCs is useful to monitor the in vitro enzymatic degradation of GSLs to ITCs in complex mixtures. Furthermore, this study confirms that ITCs from subclasses MSITC and MSoITC had potent antimicrobial activity and, therefore, might be potential new natural food preservatives, but their reactivity with food matrix components should be considered. Furthermore, the QSAR models developed in this thesis can be applied to predict antibacterial activity of new ITCs and (natural) mixtures of ITCs. Overall, ITCs are promising natural antimicrobial candidates worth further studies