Coumarins of Lovage Roots (Levisticum officinale W.D.J.Koch): LC-MS Profile, Quantification, and Stability during Postharvest Storage

Lovage (Levisticum officinale W.D.J. Koch) is a known aromatic apiaceous species that is widely used as a culinary and medicinal plant. Traditionally, more scientific attention has been paid to lovage volatiles, while other groups of compounds have been underutilized. In this study, metabolites of fresh lovage roots were investigated by liquid chromatography–mass spectrometry, and 25 compounds were identified, including coumarins as basic components and minor hydroxycinnamates; most were detected for the first time in the plant. Four major coumarins (including apterin, xanthotoxin, isopimpinellin, and pimpinellin) were successfully separated by a validated HPLC–PDA method, and the fresh roots of seven lovage cultivars as well as the dry roots of commercial lovage were quantified. The coumarin content deviation was 1.7–2.9 mg/g in the fresh roots and 15–24 mg/g in the dry roots. A variation in the coumarin level was found during storage of the fresh lovage roots at chill and room temperatures, while storage of the dried roots at room temperature showed the lowest loss of target compounds. This new information about the metabolites of lovage indicates the prospects of the plant roots as a source of dietary coumarins.

Apiaceous vegetables (e.g., carrot, celery, and parsley) have a culinary application as fresh roots; therefore, owing to the limited postharvest shelf life, the study of chemical changes deserves special attention to reduce losses. In particular, it is known that the process of storing reduces the content of carotenes, phenolics, and ascorbic acid in fresh carrot roots [46,47], and chill storage allows to slow down destructive processes [48]. Twenty-days of storage of fresh celery roots resulted in the decrease or increase in the content of chlorogenic acids, depending on the variety [49], while the ascorbic acid content reduced after six days of postharvest refrigerated storage [50]. There is no information about postharvest stability of coumarins in lovage and other apiaceous species despite their obligate presence in root products.
As part of the ongoing study of Apiaceae coumarins [36][37][38][39], high-performance liquid chromatography with photodiode array detection with electrospray ionization triple quadrupole mass spectrometric detection (HPLC-PDA-ESI-TQ-MS) was applied for phenolic metabolite profiling of the fresh roots of lovage (L. officinale), followed by the quantification of the principal components by rapid HPLC-PDA of the fresh roots of seven lovage cultivars and dry commercial products, and the postharvest changes of coumarins in the lovage roots were studied.

Plant Extracts Preparation
Fresh roots of 3 y.o. lovage plants (25-30 cm long) were homogenized by X-1740 homogenizer (Goldleaf Scientific, Riverside, CA, USA), and a portion of homogenate (5 g) was treated by 45 mL of methanol and sonicated twice (ultrasonic bath, 20 min, 50 • C, ultrasound power 100 W, frequency 35 kHz). Dry lovage roots were ground in laboratory grinder KM-100 (MRC group, Harlow, Essex, UK) till particle size 0.125 µm, and 1-g sample was extracted by 50 mL of methanol with double sonication (ultrasonic bath, 40 min, 50 • C, ultrasound power 100 W, frequency 35 kHz). Methanolic extract (after fresh or dry tissue extraction) was filtered through 0.22-µm syringe filters into a measuring flask (100 mL) and the final volume was filled up to 100 mL by methanol. The resultant extract was stored at 2 • C before analysis. Lovage roots metabolite profiling was performed by HPLC-PDA-ESI-TQ-MS assay on the liquid chromatograph LC-20 Prominence coupled with photodiode array detector SPD-M30A (wavelength range 200-600 nm), triple-quadrupole mass spectrometer LCMS 8050 (all Shimadzu, Columbia, MD, USA) and ProntoSIL 120-5 C18 AQ column (1 mm × 50 mm, 1 µm; Knauer, Berlin, Germany). The gradient elution used eluents A (1% acetic acid in water) and B (1% acetic acid in acetonitrile) and the gradient program (%B): 0-4 min 5-100%, 4-5 min 100%, 5-6 min 100-5%, and 6-7 min 5%. The injection volume was 0.5 µL and the flow rate was 500 µL/min. Ultraviolet spectra were recorded in a spectral range 200-600 nm. Electrospray ionization triple quadrupole mass spectrometric detection used temperature 300 • C in the ESI interface, 250 • C in the desolvation line, and 400 • C in the heat block. The nebulizing gas (N 2 ) flow value was 3 L/min, heating gas (air)-10 L/min, and collision-induced dissociation gas (Ar)-0.3 mL/min. The source voltage of mass spectra was 3 kV, collision energy was +10-+25 eV (positive ionization), and the scanning range was m/z 80-1900. LabSolution's workstation software managed the LC-MS system. The integrated analysis of retention time, ultraviolet and mass spectra data after comparison with the inner LC-MS library, reference standards and the literature data were used for the identification of metabolites.

HPLC-PDA-MS Metabolite Quantification
To quantify four coumarins (apterin, xanthotoxin, isopimpinellin, pimpinellin) in plant extracts, the HPLC-PDA-ESI-TQ-MS separation and detection conditions were used (Section 2.3). Reference standards were separately weighed (10 mg), dissolved in the methanol in volumetric flasks (10 mL), and the stock solution (1000 µg/mL) was used to prepare the calibration solutions (1-100 µg/mL). After the separation, PDA data were used to create 'concentration-PDA peak area' correlation. Correlation coefficient (r 2 ), standard deviation (S YX ), limit of detection (LOD), limit of quantification (LOQ), and linear range were calculated in Advanced Grapher 2.2 (Alentum Software Inc., Ramat-Gan, Israel) using calibration curve data. Values of intra-day and inter-day precisions and recovery of spiked sample were determined as described early [38]. Three HPLC runs were sufficient for the quantitative analyses, and the results were expressed as mean values ± standard deviation (S.D.).

Lovage Roots Storage Experiment
Six and five portions of the fresh lovage samples (10 roots, approx. equal; cv. Lider) were deposited into the individual polystyrene bags (2 L) and incubated at (1) 1 • C (6 months) or (2) at 20 • C (14 days), respectively, in a ventilated MK 53 thermostat (BINDER GmbH, Tuttlingen, Germany). Five roots of stored samples were taken out for analysis (1) every month or (2) at 1, 3, 7, 11 and 14 days, and extraction/analysis procedure was applied (Sections 2.2-2.4). The samples of dry lovage roots (manufacturer Evalar, CJSC; production year 2016; 1 kg) were deposited into the individual polystyrene bags (2 L) and incubated at 10 • C in a ventilated MK 53 thermostat (BINDER GmbH, Tuttlingen, Germany) for 6 years. Two-hundred portions of stored sample were taken out for analysis every year and extraction/analysis procedure was applied (Sections 2.2-2.4).

Statistical Analysis
Statistical analyses were performed by one-way analysis of variance, and the significance of the mean difference was determined by Duncan's multiple range test. Differences at p < 0.05 were considered statistically significant. The results are presented as the mean ± S.D. The linear regression analysis and generation of calibration graphs were conducted using Advanced Grapher 2.2 (Alentum Software, Inc., Ramat-Gan, Israel).
Coumarins have specific absorbance in the UV region [60] and gave a typical triplet in the positive ionization mass spectra, featuring signals of protonated ion   12, x FOR PEER REVIEW (a)   Table 2.  Table 2.

Quantification of Four Principal Coumarins in Lovage Roots
Chromatographic conditions applied for metabolite profiling of fresh lovage roots gave appropriate separation of four principal coumarins with more abundant peak areas, such as for apterin (peak 9), xanthotoxin (peak 18), isopimpinellin (peak 19), and pimpinellin (peak 20), enabling their use for quantification of the mentioned coumarins in plant samples. To simplify and lower the cost of the assay, in this study, PDA detection was used, resulting in a fast and easy method of analysis. The validation procedure demonstrated the good linearity of the calibration equations built for four coumarins with correlation coefficients (r 2 ) of 0.9925-0.9981 and standard deviations of (SYX) 9.76-11.52 × 10 −2 ( Table 3).  Non-coumarin metabolites of fresh lovage roots are derivatives of caffeic acid as mono-(4) and di-O-caffeoylquinic acids (7, 11-13) as well as 5-O-feruloylquinic acid [53]. Acid 4 has been previously found in the herbal part of lovage [12,14,15,31], and the remaining phenolics have been discovered for the first time in L. officinale. A previous report [29] showed that some flavonoid compounds can be detected in lovage roots; however, in our case, no member of this group was found.

Quantification of Four Principal Coumarins in Lovage Roots
Chromatographic conditions applied for metabolite profiling of fresh lovage roots gave appropriate separation of four principal coumarins with more abundant peak areas, such as for apterin (peak 9), xanthotoxin (peak 18), isopimpinellin (peak 19), and pimpinellin (peak 20), enabling their use for quantification of the mentioned coumarins in plant samples. To simplify and lower the cost of the assay, in this study, PDA detection was used, resulting in a fast and easy method of analysis. The validation procedure demonstrated the good linearity of the calibration equations built for four coumarins with correlation coefficients (r 2 ) of 0.9925-0.9981 and standard deviations of (S YX ) 9.76-11.52 × 10 −2 (Table 3). The limits of detection and limits of quantifications were 0.18-0.26 µg/mL and 0.56-0.78 µg/mL, respectively, and the linear range was 0-1000 µg/mL. The intra-day and inter-day precisions were high and showed relative standard deviations (RSDs) of 0.96-1.20% and 1.40-1.93%, respectively, and spiked samples demonstrated high recovery levels from 98.51% to 101.70%. All these results showed the appropriateness of the method for quantification of the principal coumarins in lovage roots.

Post-Harvest Changes in Four Principal Coumarins in Lovage Roots
Traditionally, the methods of lovage root storage have been similar to those of other apiaceous root (carrot, parsley, celery, and fennel). The best preservation has been observed for chilled storage when the temperature is close to zero; however, room temperature storage is popular for fresh roots. Therefore, postharvest changes in fresh lovage roots were studied under two conditions: one group of samples was stored at 1 • C for 6 months, and second group was conditioned at 20 • C for two weeks. These periods were chosen taking into consideration the satisfactory appearance of vegetables; as a rule, after these dates, roots became flabby (lost firmness) and were no longer stored. Additional study was focused on the changes in dry lovage roots over long-term storage for 6 years at 10 • C (the temperature of a dry plant repository). Studies of the two different types of samples were due to the widespread use of both fresh and dried lovage roots for which it is necessary to determine the composition of coumarins before and after storage (Table 5 and Figure 3).  Chilled storage of fresh lovage roots negatively affected the total coumarin content. Storing roots for 6 months resulted in a loss of 25% of total coumarins, mostly because glycoside apterin losses resulted in 67% damage. Reduction in the content of non-glycosidic xanthotoxin, isopimpinellin, and pimpinellin was no more than 30% of the initial level. Postharvest changes occurred much more rapidly when fresh lovage roots were stored at room temperature. After 2 weeks of storage, a 32% loss of apterin was observed with almost full preservation of other coumarins. Dried lovage roots demonstrated good stability of coumarin content upon long-term storage. Non-glycosidic compounds were resistant and demonstrated approximately a 5% loss after 6 years of storage, and the decrease in glycoside apterin was more than 25%. Despite the loss of compounds during all types of storage, the lovage roots remained a source of coumarins even at the end of the expiration date.
The general trend of postharvest changes in both fresh and dried lovage roots is the significant loss of the glycosidic coumarin apterin. The same changes were observed for other storing plants. The roots of Hansenia forbesii (H.Boissieu) Pimenov and Kljuykov (syn. Notopterygium forbesii H.Boissieu) can lose up to 60% of the coumarin glycoside Chilled storage of fresh lovage roots negatively affected the total coumarin content. Storing roots for 6 months resulted in a loss of 25% of total coumarins, mostly because glycoside apterin losses resulted in 67% damage. Reduction in the content of non-glycosidic xanthotoxin, isopimpinellin, and pimpinellin was no more than 30% of the initial level. Postharvest changes occurred much more rapidly when fresh lovage roots were stored at room temperature. After 2 weeks of storage, a 32% loss of apterin was observed with almost full preservation of other coumarins. Dried lovage roots demonstrated good stability of coumarin content upon long-term storage. Non-glycosidic compounds were resistant and demonstrated approximately a 5% loss after 6 years of storage, and the decrease in glycoside apterin was more than 25%. Despite the loss of compounds during all types of storage, the lovage roots remained a source of coumarins even at the end of the expiration date.
The general trend of postharvest changes in both fresh and dried lovage roots is the significant loss of the glycosidic coumarin apterin. The same changes were observed for other storing plants. The roots of Hansenia forbesii (H.Boissieu) Pimenov and Kljuykov (syn. Notopterygium forbesii H.Boissieu) can lose up to 60% of the coumarin glycoside nodakenin (nodakenetin O-glucoside) during storage while maintaining non-glycosidic compounds [75]. Scopolin (scopoletin O-glucoside) and scopoletin reduction was observed in cassava roots (Manihot esculenta Crantz) after 7 days of dark storage at 29 • C [76]. Instability of flavonoid glycosides during short-and long-term storage was found for apple [77] and strawberry fruits [78,79]. The possible reasons are increasing cleavage processes that involved the influence of water, acids, and enzymes, resulting in hydrolysis of storage compounds [80]. However, in the case of lovage roots, most coumarins are found in the non-glycosidic form, thus preserving the valuable potential of the plant.

Conclusions
This study for the first time elucidated the phenolic profile of fresh lovage roots, a traditional food product that is still scarcely investigated. The basic components were simple coumarins and furanocoumarins with various structures with or without glycosidic fragments. Even though some metabolites were in the lovage roots, most identified compounds were new for the Levisticum officinale species. Successful chromatographic separation of the principal compounds resulted in creation of a convenient assay for quantification of four coumarins (i.e., apterin, xanthotoxin, isopimpinellin, and pimpinellin), which were found at high levels in both the various cultivars of the fresh lovage roots and in the dry commercial roots. These findings suggest for the first time that lovage roots are a good source of furanocoumarins with proven bioactivity, making lovage a functional food product. The results of the postharvest stability study of lovage coumarins demonstrated a gradual decrease in target compounds, especially the glycoside apterin in fresh and dried roots. However, the final losses accounted for less than one quarter of the total coumarin content, which confirmed satisfactory retention of furanocoumarins in lovage during postharvest storage. Therefore, the known spicy-aromatic vegetable lovage roots accumulate not only phthalides and volatile compounds but also coumarins, making it one of the most valuable apiaceous plants used as food and medicine

Data Availability Statement:
The data presented in this study are available in the article.