Chemical and bioactive characterization of the aromatic plant Levisticum o ﬃ cinale W.D.J. Koch: a comprehensive study

,


Introduction
Since ancient times, several aromatic plants and spices have been used worldwide for culinary purposes aiming to modify, enhance or improve the flavour of foods.Besides their significance in gastronomy, several of them are also recognized for their empirical use in folk medicine due to their beneficial health effects.In fact, beyond their nutritional value, several of these plants are known to have many phytochemicals, including polyphenols and terpenoid compounds, which can significantly contribute to their biological activity. 1,2Currently, aromatic plants are also considered as a valuable source of natural products, many of them being secondary metabolites that have diverse applications in the chemical, pharmaceutical, cosmetic and food industries. 3Among such products, essential oils have long been used by the food industry primar-ily as flavouring agents. 4However, because several essential oils have a Generally Recognized as Safe (GRAS) status attributed by the Food and Drug Administration (FDA), presently, the possibility of their utilization as natural preservatives to extend the shelf-life of foods has been attracting the attention of the food industry. 5In fact, the addition of synthetic additives in foods is increasingly raising concerns in consumers about eventual harmful effects.In this regard, several studies focusing on different aromatic plants have suggested the use of essential oils from these plants, or their constituents, as promising alternatives to synthetic preservatives. 4,5mong the several species of aromatic plants used in culinary products, Levisticum officinale W.J.D. Koch, also known as lovage, was once much recognized, being considerably used by the condiment's industry 6 and by households in soups, stews, meat dishes, etc.Nevertheless, with the exception of quite a few culinary chefs, nowadays, the use of lovage in gastronomy is most probably unknown to the vast majority of people.L. officinale is a plant native to south-western Asia and Southern Europe, being classified as a perennial dicotyledonous plant belonging to the Apiaceae family.As several other plants of this family, lovage is mainly used for its aromatic properties.Due to its strong taste like celery combined with parsley with a scent of aniseed and curry, it reminds us of some commercial condiments, thus being also commonly designated as "Maggi plant" in several countries. 7Consumers who know this plant frequently grow it in private gardens and use it as a natural substitute of commercial condiments or stocks.Besides their significance in culinary products, L. officinale has also been used as a medicinal plant due to its carminative, spasmolytic and diuretic effects, being approved by the German Commission E for lower urinary tract infections. 8Moreover, the decoction extract prepared from the aerial parts of the plant has been described as an antiseptic for wounds. 9Despite the significance of this aromatic plant, so far only some studies concerning the volatile composition of the essential oil extracted from lovage grown in different countries are available in the literature.To the best of our knowledge, we are unaware of any previous studies regarding the nutritional value, composition in bioactive compounds and biological properties of lovage.Therefore, this study aims at addressing this gap, contributing towards expanding the knowledge on this plant species.

Samples
Edible aerial parts (leaves and stems) of L. officinale were bought in October 2018 from a local supermarket in Porto, Portugal.The leaves and stems were weighed, lyophilized (FreeZone 4.5, Labconco, MO, USA), powdered (20 mesh) and stored at −20 °C for further analysis.

Preparation of extracts
Two types of extracts were prepared, namely a decoction and a hydroethanolic extract.To prepare the decoction, 3 g of lyophilized leaves and stems were extracted with 300 mL of deionized water by boiling for 5 min.After standing at room temperature for another 5 min, the mixture was filtered through Whatman filter paper no. 4 and then subjected to freeze-drying (FreeZone 4.5, Labconco, USA).
For the hydroethanolic extract, 3 g of the lyophilized leaves and stems were extracted with 90 mL of ethanol : water (80 : 20, v/v) under stirring at room temperature, for 1 hour.The mixture was filtered through Whatman paper filter no. 4 and the residue was re-extracted by repeating the procedure.After gathering the two filtrates, ethanol was eliminated under vacuum at 40 °C by using a rotary evaporator (Buchi R-2010).The obtained solution was frozen and further lyophilized.

Chemical parameters
2.3.1.Macronutrients and energetic value.The lyophilized leaves and stems were analysed for moisture, ash and macronutrients according to the AOAC methods. 10Briefly, the crude protein was evaluated by the macro-Kjeldahl method (N × 6.25) using an automatic distillation and titration unit (model Pro-Nitro-A, JP Selecta, Barcelona), the crude fat was determined by extraction with petroleum ether using Soxhlet apparatus, the ash content was determined by incineration at 550 ± 15 °C, and total carbohydrates were determined by calculating the difference.Energy was calculated according to the Atwater system: energy (kcal per 100 g) = 4 × (g proteins + g carbohydrates ) + 9 × (g fat ).
2.3.2.Free sugars.The content of free sugars was determined in the lyophilized sample and in both extracts.For the lyophilized sample, the procedure was performed as previously described 11 while for the extracts 30 mg of each extract was dissolved in 2 mL of distilled water, filtered through a 0.22 μm disposable LC filter disk and directly injected on a chromatographic system consisting of a high performance liquid chromatograph coupled with a refraction index detector (Knauer, Smartline system 1000, Berlin, Germany) operating under the conditions described by Barros et al. 11 Compounds were identified by comparison with standards and quantification was achieved using melizitose (25 mg mL −1 ) as an internal standard.The obtained data were handled using the Clarity 2.4 software (DataApex, Prague, Czech Republic) and the results were expressed in g per 100 g of fresh weight (fw) or in mg g −1 extract, for the lyophilized aerial parts or the extracts, respectively.
2.3.3.Organic acids.Organic acids were determined in the lyophilized sample and in both extracts.The lyophilized sample was extracted using a methodology previously described and optimized, 11 while the extracts (10 mg) were redissolved in 1 mL of metaphosphoric acid (4.5%), filtered (0.22 μm) and directly analysed in the chromatographic system.The analysis was performed using an ultra-fast liquid chromatography system coupled with a diode-array detector (Shimadzu Corporation, Japan) as previously described. 11ompounds were identified and quantified by comparison of the retention time, spectra and peak area recorded at 245 nm for ascorbic acid and at 215 nm for the remaining compounds with those of commercial standards (oxalic, quinic, malic, ascorbic, citric and fumaric acids; Sigma-Aldrich, St Louis, USA).The results were processed using the software LabSolutions Multi LC-PDA (Shimadzu Corporation, Kyoto, Japan), and expressed in g per 100 g fw or in mg g −1 extract, for the lyophilized aerial parts or the extracts, respectively.
2.3.4.Fatty acids.Fatty acid methyl esters (FAME) were prepared as previously described, 12 and determined by gas-liquid chromatography with flame ionization detection, using a DANI model GC 1000 instrument.Separation was achieved on a Zebron-Kame column (30 m × 0.25 mm ID × 0.20 µm, Phenomenex, Lisbon, Portugal) operating under the following oven temperature program: initial temperature of 100 °C, held for 2 min, increase at 10 °C min −1 to 140 °C, followed by a 3 °C min −1 ramp to 190 °C, a 30 °C min −1 ramp to 260 °C and held for 2 min.The carrier gas (hydrogen) flow-rate was 1.1 mL min −1 , measured at 100 °C.The injection split ratio was 1 : 50, with injector and detector temperatures being set at 250 °C and 260 °C, respectively.Fatty acid identification and quantification (Clarity DataApex 4.0 Software, Prague, Czech Republic) were performed by comparing the relative retention times of FAME peaks from samples with standards (standard mixture 47885-U, Sigma, St Louis, USA) and the results were expressed as the relative percentage of each fatty acid.
2.3.5.Tocopherols.Tocopherols were determined in the lyophilized sample using a HPLC system coupled to a fluorescence detector (FP-2020, Jasco, Easton, USA) as previously described. 13Compounds were identified by comparison with tocopherol standards (α-, β-, γ-, and δ-isoforms, Sigma, St Louis, MO, USA) and quantification was performed by the internal standard (IS) method using tocol (Matreya, Pleasant Gap, USA) as the IS.The results were processed using the Clarity 2.4 software and expressed in mg per 100 g fw.
2.3.6.Volatile compounds.Leaves and stems were subjected to essential oil extraction by hydrodistillation in Clevenger apparatus.The essential oil was separated from water and recovered directly without adding any solvent.After drying by adding anhydrous sodium sulphate, the oil was diluted in HPLC grade n-hexane and analysed in a GC-2010 Plus (Shimadzu) gas chromatography system with an AOC-20iPlus automatic injector and a mass spectrometry detector.Separation was achieved on an SH-RXi-5 ms column (30 m × 0.25 mm × 0.25 μm; Shimadzu, USA) operating under the following conditions: initial oven temperature of 40 °C for 4 min, raised at 3 °C min −1 to 175 °C, then at 15 °C min −1 to 300 °C and held for 10 min.Helium was used as the carrier gas adjusted to a linear velocity of 30 cm s −1 .The injected sample was 1 µL, with a split ratio of 1 : 10.The injector temperature was set at 260 °C, with a transfer line at 280 °C and an ion source at 220 °C.The ionization energy was 70 eV and a scan range of 35-500 u with a scan time of 0.3 s was used.
Identification of compounds was based on the NIST17 mass spectral library and by determining the linear retention index (LRI) based on the retention times obtained for a mixture of n-alkanes (C8-C40, ref. 40147-U, Supelco) analyzed under identical conditions.When possible, comparisons were also performed with commercial standard compounds and with published data.
Quantification was performed and the results were expressed as the relative percentage of total volatiles using relative peak area values obtained directly from the total ion current (TIC) values.
2.3.7.Phenolic compounds.Phenolic compounds were analysed in both hydroethanolic and decoction extracts after they were re-dissolved in ethanol/water (80 : 20, v/v) and water, respectively, to a concentration of 5 mg mL −1 and filtered (0.22 μm).The compounds were evaluated using a Dionex Ultimate 3000 UPLC system (Thermo Scientific, San Jose, USA) equipped with a quaternary pump and a diode array coupled in-series to an electrospray ionization mass spectrometry detector (LC-DAD-ESI/MS n ) operating under the conditions described by Bessada et al. 14 The identification of compounds was performed by comparison of data regarding retention time, UV-VIS spectra, mass spectra (full scan mode from m/z 100 to 1500) and fragmentation patterns of the sample compounds with those obtained from the available standards and/ or reported data from the literature.For quantification, calibration curves constructed based on the UV-vis signal of the following standards, chlorogenic acid, apigenine-6-C-glucoside and quercetin-3-O-glucoside (Extrasynthese, Genay, France), were used.The results were expressed in mg g −1 extract.

Bioactive properties
2.4.1.Antioxidant activity.Antioxidant properties were evaluated following different in vitro assays, namely 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical-scavenging activity, reducing power, β-carotene bleaching inhibition, lipid peroxidation inhibition by evaluating thiobarbituric acid reactive substances (TBARS), and antihaemolytic activity evaluating the inhibition of oxidative hemolysis (OxHLIA).The hydroethanolic and decoction crude extracts were re-dissolved in ethanol : water (80 : 20, v/v) and water, respectively, at a final concentration of 5 mg mL −1 and further diluted at different concentrations (in the range of 5-0.0391 mg mL −1 ) to perform the different in vitro assays, as previously described. 15he results of the assays were expressed as EC 50 , corresponding to the extract concentrations providing 50% of antioxidant activity, with the exception for the reducing power assay for which EC 50 corresponds to 0.5 of absorbance at 690 nm.Trolox (Sigma-Aldrich, St Louis, USA) was used as a standard.
The antihaemolytic activity was evaluated by the oxidative haemolysis inhibition assay (OxHLIA) as previously described by Takebayashi et al., 16 with some modifications.Blood was collected from healthy sheep and centrifuged at 1000g for 5 min at 10 °C.After discarding the plasma and buffy coats, the erythrocytes were first washed with NaCl (150 mM) and then three times with phosphate-buffered saline (PBS, pH 7.4).The erythrocyte pellet was then resuspended in PBS to obtain a concentration of 2.8% (v/v).Using a flat bottom 48-well microplate, 200 μL of the erythrocyte solution were mixed with 400 μL of PBS solution (control), antioxidant extracts dissolved in PBS, or water (for complete haemolysis).Trolox was used as a positive control.After pre-incubation at 37 °C for 10 min with shaking, 2,2′-azobis(2-methylpropionamidine) dihydrochloride (AAPH, 160 mM in PBS, 200 μL) was added to each well and the optical density was measured at 690 nm every 10 min over 400 min while the microplate was re-incubated. 16he percentage of the erythrocyte population that remained intact (P) was calculated according to the equation: P(%) = (S t − CH 0 /S 0 − CH 0 ) × 100, where S t and S 0 correspond to the optical density of the sample at t and 0 min, respectively, and CH 0 is the optical density of the complete haemolysis at 0 min.The results were expressed as delayed time of haemolysis (Δt ), which was calculated according to the equation: Δt (min) = Ht 50 (sample) − Ht 50 (control), where Ht 50 is the time (min) corresponding to 50% haemolysis, graphically obtained from the haemolysis curve of each antioxidant sample concentration.The Δt values were then correlated with the extract concentrations, 16 and from the correlation obtained, the extract concentration able to promote a Δt haemolysis delay was calculated.The results were given as IC 50 values (μg mL −1 ) at Δt 60 and 120 min, i.e., the extract concen-tration required to keep 50% of the erythrocyte population intact for 60 and 120 min.
2.4.2.Hepatotoxicity and cytotoxic activity.Hepatotoxicity was evaluated by the Sulforhodamine B (Sigma-Aldrich, St Louis, USA) assay as described by Abreu et al., 17 using a primary cell culture (PLP2) prepared from a porcine liver sample and different concentrations of the hydroethanolic or decoction extracts, ranging from 400 µg mL −1 to 6.5 µg mL −1 .The anti-proliferative capacity of the extracts was also evaluated by the same method but using four human tumour cell lines (acquired from Leibniz-Institut DSMZ): MCF-7 (breast adenocarcinoma), NCI-H460 (non-small cell lung cancer), HeLa (cervical carcinoma), and HepG2 (hepatocellular carcinoma).In both the cases, ellipticine (Sigma-Aldrich, St Louis, USA) was used as the positive control and the results were expressed as GI 50 values (μg mL −1 ), corresponding to the extract concentration that provides 50% of cell growth inhibition.
2.4.3.Antimicrobial activity.The antibacterial activity was evaluated by the broth microdilution method coupled to the rapid p-iodonitrotetrazolium chloride (INT) colorimetric assay as previously described. 18The microorganisms used were clinical isolates and included three Gram-positive bacteria (Enterococcus faecalis, Listeria monocytogenes and methicillinresistant Staphylococcus aureus) and five Gram-negative bacteria (Escherichia coli, Klebsiella pneumoniae, Morganela morganii, Proteus mirabilis and Pseudomonas aeruginosa).The minimum inhibitory concentration (MIC) was determined through the colorimetric microbial viability based on the reduction of the INT dye (0.2 mg mL −1 ).The minimum bactericidal concentration (MBC) was evaluated by plating the content of the microwells without coloration in the MIC assay.Different antibiotics were used as negative controls, namely ampicillin and imipenem for Gram-negative bacteria, and vancomycin and ampicillin for Gram-positive bacteria.

Statistical analysis
All the analyses were performed at least in triplicate with results being expressed as mean ± standard deviation (SD).SPSS Statistics software (IBM SPSS Statistics for Windows, v. 23.0.) was used to analyse the existence of differences among the two extracts by applying Student's t-test at a 5% confidence level.

Chemical characterization of L. officinale leaves and stems
The results of the nutritional value, free sugars and organic acid composition obtained for the analysis of the aerial edible parts of lovage (leaves and stems) are shown in Table 1.Carbohydrates were the most abundant macronutrient, followed by proteins, with lipids being present in a minor amount.Comparing the obtained results with those reported for other commonly consumed species of the Apiaceae family, such as parsley (Petroselinum crispum), coriander (Coriandrum sativum) and celery (Apium graveolens), it can be seen that the values of the proximate composition are very similar, since all show a high amount of moisture, and carbohydrates and lipids as the major and minor groups of macronutrients. 19,20owever, considering the values calculated for dry mass, while parsley, coriander and lovage have identical contents of carbohydrate (51.5 g per 100 g, 47.1 g per 100 g and 53.4 g per 100 g, respectively) and protein (24.2 g per 100 g, 27.3 g per 100 g, and 28.4 g per 100 g, respectively), celery is a bit less similar as it presents a higher carbohydrate content (65.0 g per 100 g) and a lower protein content (15.0 g per 100 g) compared to lovage.As expected, due to the high moisture content, the energetic value calculated for the edible parts of L. officinale was considerably low (38 kcal per 100 g fw).When searching the USDA food composition database using the words "lovage" or "levisticum" no results were obtained.Additionally, from the literature consulted, no information was found regarding the nutritional composition of lovage, with data being reported for the first time in the present study.
L. officinale presented a low content of free soluble sugars, with only 4 compounds being detected (Table 1).Among those, the predominant compounds were sucrose and glucose, followed by trehalose and smaller amounts of arabinose.
Organic acids have been evaluated in previous studies regarding different vegetables as they are known to affect the organoleptic characteristics of vegetables, such as their flavour (mainly acidity) and appearance, besides playing important roles in plant metabolism and in microbial interactions. 21oreover, organic acids can also influence consumers' acceptability, being frequently used in the food industry as acidifying compounds.A total of six organic acids were identified in the edible aerial parts of L. officinale (Table 1), with oxalic acid being the predominant compound, followed by quinic, citric and malic acids.Regarding oxalic acid, known to promote kidney stones when taken in high amounts, compared to other vegetables, lovage presents a moderate content of oxalic acid (0.46 g per 100 g fw) being lower than that found in parsley raw leaves (1.7 g per 100 g), but higher than those in celery (0.19 g per 100 g) and coriander (0.01 g per 100 g). 20he fatty acid profile is shown in Table 2, where it can be observed that a total of 19 fatty acids were identified, with polyunsaturated fatty acids (PUFA) representing the major group, followed by saturated fatty acids (SFA) and monounsaturated fatty acids (MUFA).The edible aerial parts of lovage presented a predominant composition of unsaturated fatty acids, being particularly rich in omega-3 PUFA, mainly due to the high content of α-linolenic acid (48.7%).L. officinale leaves and stems also showed considerable amounts of linoleic (28.5%) and palmitic (10.0%) acids.To the best of our knowledge, this study reports for the first time the fatty acid profile of this vegetable species.Despite lipids being present in small amounts (Table 1), lovage presents an interesting fatty acid profile concerning omega-3 and also essential fatty acids.Several studies demonstrated that a deficient intake of these fatty acids can lead to various problems including dermatitis, immunosuppression, cardiac dysfunction, promotion of the development of some degenerative diseases and acceleration of the process of aging. 22In this sense, the consumption of L. officinale can be a contribution to a healthy diet owing to its favourable fatty acid profile (Table 2).
The results regarding vitamin E are also shown in Table 2. Vitamin E works in vivo as an antioxidant, preventing the spread of free radical damage caused by peroxyl radicals, particularly protecting PUFA.This prevention is possible because such radicals react 1000 times faster with vitamin E than with PUFA. 23Only two isoforms of tocopherols (α-and γ-) were identified and quantified in the leaves of L. officinale (Table 2), with α-tocopherol being the predominant compound.The low content of tocopherols is most possibly related to the low amount of lipids in the studied sample.
Considering that aromatic plants are generally characterized by a strong taste and an intense aroma, they are used to aromatize and enhance the flavor of dishes and frequently allow the use of a reduced amount of sodium chloride, therefore imparting health benefits beyond their composition in bioactive compounds.Due to its intense and pungent organoleptic characteristics, which somehow reminds us of the flavor of some commercial seasonings, lovage extracts were once used as a raw material in the production of condiments to impart a seasoning-like flavor 6 and is still used by some consumers, in the plant form, as a healthier substitute of commercial seasonings.Therefore, the volatiles profile of the essential oil obtained by hydrodistillation of the aerial parts of the lovage sample grown in Portugal was studied, and is listed in Table 3.The GC-MS analysis allowed identification of 99.1% of the compounds, corresponding to a total of 44 identified compounds.Monoterpene hydrocarbons and oxygenated monoterpenes were the two main groups, each representing approximately 37% of the identified compounds, with phthalides also present in high amount (24.3%).α-Terpinyl acetate was found to be the major compound (33.6%),followed by p-cymene (20.5%), (Z)-ligustilide (22.2%), β-phellandrene (4.7%) and myrcene (4.2%).When comparing the obtained composition, in terms of most abundant compounds, with data previously published for the volatile composition of L. officinale grown in Iran, 9,24,25 Moldavia, 26 Lithuania, 27 and Estonia, 28 in general, the chemical profile of the Portuguese grown sample was considerably different.Despite the similarity to other oils that also presented α-terpinyl acetate as the major compound, a higher content has been previously reported in the aerial parts (52.4%), 24 leaves (55.8% and 49.7% to 70%), 27,28 and stems (49.1% to 69.0%) 27 of lovage.Other studies reported different major compounds, namely pentyl cyclohexa-1,3-diene (28.1%) 9 and γ-terpinene (14.5%). 25Additionally, in the sample grown in Portugal, the second most abundant compound was found to be p-cymene (20.5%), which has not been identified or has been reported in minor amounts (ranging from 0.1% to 4.3%) in the previous studies.In the present work, different phthalides were detected, with the most abundant being (Z)-ligustilide (22.2%).7][28] Phthalides are a relatively small group of

View Article Online
natural compounds confined to various plant families and some genera of fungi.Among the different plant families for which studies have been performed to characterize these compounds, the Apiaceae family stands out. 29Different biological properties, including antioxidant activity, antihyperglycemic activity, and analgesic and neurological effects, have been ascribed to these compounds, in particular to (Z)-ligustilide.Moreover, phthalides have been reported as bioactive constituents of different plant species used in traditional medicine, such as Angelica sinensis (Chinese name: Danggui), Angelica acutiloba (Japanese name: Toki), and Ligusticum porteri (Hispanic name: Oshá). 29

Chemical characterization of L. officinale extracts
Considering that the leaves and stems of lovage are used in gastronomy for flavoring dishes, being frequently added in stews and in meat marinades, besides the chemical characterization of the aerial parts of lovage, they were also subjected to

Food & Function Paper
This two different extractions (decoction and hydroethanolic).These extracts, besides being intended for the bioactivity assays, were also evaluated regarding their composition in hydrophilic compounds, namely soluble sugars, organic acids and phenolic compounds.
Regarding the quantification of free sugars (Table 4), arabinose was the major sugar present followed by sucrose, trehalose and glucose, the main free sugars also present in the dry plant.The hydroethanolic extract presented higher amounts compared to the decoction extract ( p < 0.05); this was expected due to the extraction solvent used.
Regarding organic acids, all the compounds identified in the plant were present in the decoction while citric acid was not detected in the hydroethanolic extract.In comparison with the hydroethanolic extract, the decoction allowed the extraction of a significantly higher amount of total organic acids ( p < 0.05), the difference regarding the malic acid content being particularly noticeable, which was about 4× higher in the aqueous extract (Table 4).
The results of the phenolic compound analysis carried out on the two extracts are shown in Fig. 1 and Table 5.    5.
][34] Quercetin-3-O-rutinoside was the most abundant flavonoid in both extracts, though it was present in a higher amount in the hydroethanolic extract.This is possibly related to the low solubility of this compound in aqueous solutions, which can also affect is bioavailability. 35Still, quercetin-3-O-rutinoside, also known as rutin, has been shown to exhibit a wide range of biological activities mainly related to its antioxidant, antimicrobial, anti-inflammatory, anti-thrombogenic and anticancer activities. 36Regarding the phenolic acid group, 3-Ocaffeoylquinic and cis-5-O-caffeoylquinic acids were the main compounds in the decoction and hydroethanolic extracts, respectively.These compounds are frequently abundant in the human diet and have been linked to reduced risks of developing different chronic diseases mainly due to their antioxidant and anti-inflammatory properties. 37,38The compound tentatively identified as maclurin-3-C-glucoside, a benzophenone derivative previously reported in honeybush, 34 was also present in considerable amounts in both extracts.Although no bioactive properties have been ascribed so far to this C-glycosylated compound, the aglycone maclurin has been reported to exhibit antioxidant activity, anti-cancer effects and suppression of enzymatic browning. 39,40espite the similarity of the qualitative profiles in both types of extracts, significant differences were observed regarding the total amount of phenolic compounds and each individual compound ( p < 0.05), with a higher extraction yield being achieved when using a mixture of ethanol : water as the extraction solvent.Previous data on the phenolic composition of methanolic extracts of L. officinale leaves have been recently reported by Złotek et al. 42 The authors reported the tentative identification of a total of 14 phenolic compounds, some being identical to the ones also reported in this work, such as 4-O and 5-O-caffeoylquinic acids and rutin.Nevertheless, other compounds described by Złotek et al. 42 such as apterin, caffeic, sinapic and p-coumaric acids were not found in the present work.

Bioactive properties of L. officinale hydroalcoholic and decoction extracts
The bioactive properties of plant extracts evaluated in vitro are known to be directly related to their chemical composition and concentration of bioactive compounds. 43Although aromatic plants are generally consumed in small quantities, when ingested regularly they can be a potential source of beneficial compounds, thus playing a relevant role in the human diet.Therefore, this study also comprised the in vitro evaluation of different biological properties, namely antioxidant, cytotoxic and antimicrobial activities.
In this study, five different methods were applied to access the antioxidant properties of the two prepared extracts (Table 6), namely DPPH free radical uptake, reducing power, inhibition of β-carotene bleaching, inhibition of lipid peroxidation (TBARS) and inhibition of oxidative hemolysis (OxHLIA).In general, the hydroalcoholic extract showed superior activity since it presents the lowest EC 50 values for most assays (with the exception of TBARS), which may be related to its higher concentration in phenolic compounds compared to that of the decoction.
In the OxHLIA assay, erythrocytes are subjected to hemolysis by the action of both hydrophilic and lipophilic radicals, therefore being considered an in vitro cell-based antioxidant assay.By observing the data presented in Table 6, it can be noticed that both extracts showed better results regarding antihemolytic activity in comparison with Trolox, a water-soluble analog of vitamin E widely used as an antioxidant in biochemi-

Food & Function Paper
This cal assays, thus demonstrating very interesting antioxidant properties.
Regarding the evaluation of in vitro cytotoxic activity in tumor cells, only the hydroalcoholic extract presented activity against HepG2 cells (Table 6).Bogucka-Kocka et al. 44 evaluated the cytotoxic potential of hydroethanolic extracts obtained from defatted fruits of L. officinale against seven leukaemia human cell lines and two normal cell lines (human T cells and human B cells) and found that the extract strongly affected two leukemic cell lines (C8166 and J45), being also highly significant in inducing apoptosis.Nevertheless, the authors also found that the extract caused a significant decrease of viable cells for the two normal cell lines (H9 and WICL).In the present study, the highest tested concentration of both extracts prepared from the leaves and stems of L. officinale did not inhibit the growth of porcine hepatic cells (PLP2), which demonstrates the absence of hepatotoxicity and also points to a low potential of causing damage in non-tumor cells.
Regarding the antibacterial activity results (Table 7), both extracts showed the capacity of inhibiting the growth of all tested microorganisms, with the exception of the hydroethanolic extract against P. mirabilis for which the highest tested concentration (20 mg mL −1 ) did not exhibit any activity.Still, none 1.2 μg mL −1 (MCF-7), 1.0 μg mL −1 (NCI-H460), 0.91 μg mL −1 (HeLa), 1.1 μg mL −1 (HepG2) and 2.3 μg mL −1 (PLP2).  of the extracts, irrespective of the tested concentration, showed bactericidal activity against the assayed microorganisms.In general, the decoction extract exhibited the best results as it showed the lowest MIC values for all tested bacteria, with the exception of P. aeruginosa.Overall, the two extracts were more efficient against Gram-positive bacteria.This observation has been previously reported by several authors and is most possibly related to the fact that this group of microorganisms presents a less complex cell wall when compared to Gram-negative bacteria that have a selective outer membrane.Moreover, Gram-negative bacteria frequently present multidrug resistance pumps (efflux pumps), which may also explain the reduced antimicrobial activity observed in vitro for several extracts and plant-derived compounds against this type of bacteria. 45,46In the present study, the extracts of lovage, in particular the decoction, presented interesting results as they were also able to inhibit the growth of different Gram-negative bacteria, although requiring higher MIC values compared to those of Gram-positive bacteria.Interestingly both extracts inhibited P. aeruginosa, with a lower MIC value being obtained for the hydroethanolic extract (Table 7).P. aeruginosa is an opportunistic pathogen frequently associated with nosocomial infections, and is being considered an emerging threat to public health due to its intrinsic resistance to several antibiotics.Moreover, infections caused by P. aeruginosa are becoming increasingly difficult to treat as the number of multidrug resistant strains is growing worldwide. 47,48Since numerous strains have been reported as being resistant to several commonly used classes of antibiotics, 48 it has been suggested that medicinal plant extracts/compounds may provide a promising approach as adjuvant therapy, as several studies showed that the antibiotic activity can be potentiated when in combination with phytochemicals. 46,47,49It should be noted that in the present study, the tested microorganisms were clinical isolates resistant to different antibiotics.

Conclusion
In this work, a comprehensive study on the nutritional, chemical and bioactive properties of the edible aerial parts (leaves and stems) of a Portuguese grown sample of L. officinale was performed.Similar to other aromatic plants, lovage is shown to be a low caloric food as it presents a high content of moisture.Regarding the composition of hydrophilic compounds in the leaves and stems of lovage, chromatographic analysis showed that sucrose and oxalic acid are the most abundant free sugar and organic acid, respectively.Moreover, a total of 7 phenolic compounds were detected and tentatively identified in the two types of extracts prepared (decoction and hydroethanolic extract), with 5-O-caffeoylquinic and 3-O-caffeoylquinic acid being the most abundant phenolic acids and quercetin-3-O-rutinoside and maclurin-3-C-glucoside the major flavonoids.
Besides the presence of these bioactive compounds, lovage also showed an interesting fatty acid profile, as it is rich in polyunsaturated fatty acids, in particular having α-linolenic acid, an omega-3 essential fatty acid, as a major compound.In the lipidic fraction, two tocopherol isoforms, α and γ, were also detected.The aroma profile of the essential oil hydrodistilled from this aromatic plant showed that it was rich in monoterpenes, with α-terpinyl acetate being the most abundant volatile, followed by p-cymene.Interestingly, a high content of phthalides, which have been associated with health benefits, was also observed.
Regarding the bioactive potential of this plant, both extracts showed interesting properties, and it is worth mentioning that they showed better activity in the OxHLIA test when compared to the antioxidant used as the control (Trolox).The extracts were able to inhibit the growth of all tested bacteria, and were in general more active against Gram-positive bacteria.Regarding cytotoxicity against tumor cell lines, it was found that only the hydroethanolic extract showed activity against liver cancer cells.Moreover, both extracts did not show toxic effects against the q non-tumor liver cell line (PLP2), suggesting that they are safe for human use.
In brief, an extensive and detailed study was performed revealing that lovage has an interesting composition from the nutritional point of view being also a source of several bioactive compounds; therefore its inclusion as a seasoning/flavoring agent in different dishes should be promoted.
-ESI/MS n analysis, namely retention time, λ max in the UV-vis region, molecular ions and main fragment ions observed in MS 2 , are detailed in Table 5 and were used for the tentative identification of compounds and respective quantification.The compounds trans 5-O-caffeolylquinic acid (chlorogenic acid, peak 4), quercetin 3-O-rutinoside ( peak 6) and kaempferol 3-O-rutinoside ( peak 7) were positively identified according to their retention time, mass spectra and UV-Vis characteristics by comparison with commercial standards.Peaks 1, 2 and 3 were tentatively identified as 3-O-caffeolylquinic acid, 4-O-caffeolylquinic acid and cis-5-O-caffeolylquinic acid, taking into account the hierarchical key fragmentation pattern described by Clifford et al. 31 Peak 5 presented a pseudomolecular ion [M − H] − at m/z 423, being tentatively identified as maclurin-3-C-glucoside based on the UV spectrum and

Fig. 1
Fig.1Phenolic profile of leaves of Levisticum officinale recorded at 280 nm (A) and 370 nm (B).Peak numbers as defined in Table5.

Table 3
Chemical composition of the essential oil extracted from the edible aerial parts (leaves and stems) of L. officinale by hydrodistillation (mean ± SD, n = 3) a LRI, linear retention index determined on a SH-RXi-5ms fused silica column relative to a series of n-alkanes (C8-C40).bLinearretention index reported in the literature (Adams, 2017).30cRelative% is given as mean ± SD, n = 3.

Table 7
Antimicrobial activity of the extracts obtained from the edible aerial parts (leaves and stems) of