Headspace Solid-Phase Micro-Extraction Versus Hydrodistillation of Volatile Compounds from Leaves of Cultivated Mentha Taxa: Markers of Safe Chemotypes

Various mint taxa are widely cultivated and are used not only for medicinal purposes but also in cosmetic and industrial applications. The development of new varieties or cultivars of mint generates difficulties in their correct identification and safe use. Volatile organic compounds (VOCs) from the leaves of seven different taxa of the genus Mentha obtained by hydrodistillation (HD) and headspace solid-phase microextraction (HS-SPME) were analyzed using gas chromatography–mass spectrometry (GC-MS). Principal component analysis (PCA) was also performed. Comparative GC-MS analysis of the obtained extracts showed similarity in the major compounds. PCA data allowed the separation of two groups of chemotypes among the analyzed mints, characterized by the abundance of piperitenone oxide and carvone. Two out of seven analyzed taxa were not previously examined for VOC profile, one was examined only for patent application purposes, and six out of seven were investigated for the first time using the HS-SPME technique. The presented analysis provides new data on the abundance and qualitative characterization of VOCs in the studied mint plants and on the safety of their use, related to the possibility of the presence of potentially toxic components. HS-SPME is a valuable method to extend the characterization of the VOC profile obtained by hydrodistillation.


Introduction
Main volatile organic compounds (VOCs), synthesized by various aromatic plants, are used in many areas of everyday life, including medicine, food, and cosmetics. They can be found in many forms, from freshly harvested or dried plants to extracts and herbal preparations, which may include essential oils (Eos). VOCs are complex mixtures of mainly terpenes, such as mono-and sesquiterpenes and phenylpropane derivatives [1]. Monoterpenes are one of the groups commonly found in Eos. These compounds can be classified as hydrocarbons and oxygen derivatives such as alcohols, phenols, aldehydes, ketones, acids, or terpene oxides. They are responsible for the aroma, flavor, and biological properties of the plants and the preparations derived from them, e.g., essential oils [2][3][4]. Some of these properties can be beneficial therapeutically (expectorant, antiseptic, diuretic, digestive stimulants, etc.), but some may have toxic effects [5].
The next most abundant compounds present in plant volatile fractions, in addition to monoterpenes, are sesquiterpenes. They exhibit a wide range of biological activities, such as anti-inflammatory, antioxidant, antibacterial, antifungal, antiviral, and antineoplastic, and they broaden the potential uses of plants that contain them, including those from the genus Mentha [6]. Most terpenes are classified as safe natural chemicals. Some of the biological properties of monoterpenes and sesquiterpenes, such as antifungal and antimicrobial, have been shown to be the result of their disruption of the cell membrane or cytoplasmic structures as a consequence of oxidative stress, while others can assist radical scavengers and are also known as antioxidant molecules [2,7].
Taxa of the genus Mentha are among the most frequently cultivated and used aromatic plants, but their systematics is quite complex. In particular, new cultivars created mainly for food or cosmetic purposes are difficult to correctly identify and to evaluate for their safety. They are mainly developed to obtain specific organoleptic properties, such as odor or taste. Thus, the composition of their VOCs is not always known, and it is not clear whether the compounds they contain are safe for the consumers' health. Mint leaves are well known for the presence of VOCs characterized by a wide variety of monoterpenes and sesquiterpenes, as well as polyphenolic compounds including flavonoids, e.g., eriocitrin, luteolin and their glycosides, and phenolic acids such as rosmarinic acid [8,9].
Various methods are used to extract VOCs from plant materials such as hydrodistillation (HD), supercritical fluid extraction (SFE), and solid-phase micro-extraction (SPME) techniques. HD is a well-known and widely used process to obtain the essential oils from aromatic plants for both laboratory and industrial purposes. It has the advantage of relatively low cost and no chemical contamination of the extract. However, some volatile compounds may be lost at high extraction temperatures or are temperature-sensitive and can be degraded, which has a significant impact on the quality of the essential oil obtained in this way [10]. SPME has been routinely used for the analysis of VOCs in headspace (HS) analysis in laboratory studies and does not require organic solvents. The main differences between the HS-SPME procedure and hydrodistillation are the abundance and qualitative variations of extracted components from plant materials. These depend on the process conditions such as SPME fiber, extraction temperature, and time [11].
The aim of this study was to identify the most suitable method to investigate VOCs in mint leaves and to provide qualitative and abundance data on the major compounds, as well as potentially toxic ones. The analyses were carried out by comparing the volatile profile from leaves of seven different cultivated Mentha taxa, obtained using two methods, headspace solid-phase microextraction (HS-SPME) and hydrodistillation (HD), followed by gas chromatography-mass spectrometry analysis (GS-MS). Two out of seven analyzed taxa were not examined previously for VOCs profile (M. × piperita var. officinalis f. pallescens Camus 'Swiss', M. × carinthiaca), one was examined only for patent application purposes (M. dulcia citreus Hillary's Sweet Lemon), and six out of seven were investigated for the first time using HS-SPME. The statistics included the evaluation of the collected data using principal component analysis (PCA). Figure 1 presents the quantitative content of the EOs in the analyzed samples determined using the HD method and expressed in mL/kg of dry plant matter. The highest essential oil content was observed for M. × piperita var. officinalis f. pallescens Camus 'Swiss' (M1) (16.5 mL/kg) and the lowest for M. longifolia var. schimperi (M6) (0.9 mL/kg). In M. × carinthiaca (M4), essential oil content was 10 mL/kg, and, in the other species, it ranged from 1.1 mL/kg in M. longifolia (M7) to 7.7 mL/kg in M. × piperita f. citrata 'Grapefruit' (M2).

GC-MS Analysis of the Volatile Compounds Obtained Using HD and HS-SPME
The EOs obtained using the HD method were subjected to quantitative and qualitative analysis via GC-MS and the results are presented in Table 1.

GC-MS Analysis of the Volatile Compounds Obtained Using HD and HS-SPME
The EOs obtained using the HD method were subjected to quantitative and qualitative analysis via GC-MS and the results are presented in Table 1. Menthol is generally recognized as safe (GRAS), and its toxic effects on humans have rarely been reported [12,13]. Among the analyzed Eos, only the oil from M. × piperita var. officinalis f. pallescens Camus 'Swiss' (M1) was characterized by p-menthone as the main compound (41%) and by menthol (28.19%).
Another monoterpene often found in taxa of the genus Mentha is linalool, widely applied in the cosmetic, food, and pharmaceutical industries. In an animal toxicology study, it was shown to cause up to a fourfold increase in serum aspartate aminotransferase enzyme levels and a twofold decrease in body weight in animals when fed with 5% linalool. An increase in this enzyme can lead to myocardial necrosis, liver damage, skeletal muscle damage, or pancreatitis [7]. Linalool (40.43%) was the dominant constituent of the EO from M. × piperita f. citrata (M3), which is consistent with the literature [14].
Pulegone and menthofuran, which may have hepatotoxic effects, are found not only in M. pulegium but also in smaller amounts in other taxa of the genus Mentha [19]. The European Medicines Agency, in a public statement on the use of herbal medicinal products containing pulegone and menthofuran, described the toxicity of this compounds and recommended limit values for herbal medicinal products that contain them [20]. According to the toxicological conclusions of the EMA's statement, the target organs for pulegone and menthofuran are the liver and kidney. Several chemotypes of M. longifolia can be distinguished including the piperitenone-dominated chemotype [21,22]. Piperitenone (38.92%), piperitenone oxide (32.88%) and cis-piperitone oxide (21.05%) were present in the highest amounts in M. longifolia var. schimperi (M6), while, in the EO from M. longifolia (M7), only piperitenone oxide was detected as the predominant compound (84.66%). Other studies identified pulegone as the major EO compound from this taxon, as well as isomenthone or carvone [23][24][25][26]. Menthofuran was not detected in any analyzed Eos obtained by HD, while pulegone was detected only in M. × piperita var. officinalis f. pallescens Camus 'Swiss' (M1) (0.62%). In this taxon, pulegone (2.14%) was also identified by Mogosan et al. [17]. The EO from the leaves of M. longifolia was characterized by pulegone as the main component by Sayed et al. [27].
Thymol, 1,8-cineole (syn. eucalyptol), and limonene are very often used in the food and cosmetic industries because of their characteristic organoleptic properties. However, they also show various pharmacological properties: mucolytic, antibacterial, antiviral, antifungal, and anti-biofilm [28][29][30]. Thymol is one of the main constituents of the essential oils of the genus Thymus; however, it can also be found in small amounts in the genus Mentha The results of the GC-MS analysis of the volatile fractions obtained using HS-SPME are presented in Table 2. HS-SPME parameters were optimized with respect to the overall number of identified compounds. The standard deviation achieved was <10% for triplicate measurements. To the best of our knowledge, VOCs of taxa M1-M5 and M7 were investigated for the first time using the HS-SPME technique. Cordero's study showed that piperitenone oxide was found to be predominant among VOCs of M. longifolia (M7) leaves, which was also confirmed by the performed analysis [35]. Najafian et al. identified pulegone as the major component of VOCs obtained using the HS-SPME method from this taxon [23]. Comparing GC-MS results of mints VOCs obtained using HD and HS-SPME, it was found that their main components were the same, and they differed only in their abundance. Some smaller variations were observed in the profile of the constituents. Menthofuran VOCs obtained using HS-SPME, 1,8-cineole was present, but in lower amounts compared to the HD, except for M. longifolia var. schimperi (M6), where this compound was determined only using HS-SPME. Similar amounts of (Z)-β-and (E)-β-ocymenes were observed for each taxon. Higher cis-piperitone oxide abundance was detected using HS-SPME for M. longifolia (M7). In M. longifolia var. schimperi (M6) and M. longifolia (M7), β-bourbonene was absent. In all analyzed taxa, trans-caryophyllene was detected. Germacrane D was not present in M. longifolia var. schimperi (M6).  Selected chromatograms of the sample M5 are presented in Figure 2.  Table 2, and the numbers on (B) correspond to Table 1.

Principal Component Analysis
The datasets containing volatile constituents determined via GC-MS analyses of the fractions obtained using hydrodistillation (EO) and HS-SPME were subjected to PCA. The data in both cases demonstrated some natural clustering into three groups (Figures 3 and  Table 2, and the numbers on (B) correspond to Table 1.

Principal Component Analysis
The datasets containing volatile constituents determined via GC-MS analyses of the fractions obtained using hydrodistillation (EO) and HS-SPME were subjected to PCA. The data in both cases demonstrated some natural clustering into three groups (Figures 3 and 4). In the case of the data obtained for the EO, the first two PCA factors explained 67. 3% of variance between the samples and the data obtained for the headspace extracts; the main two factors explained 65.0% of variance in the dataset. The variables providing the greatest contribution to the first two principal components were the following for EO: carvone (accounting for more than 45% and 35% of components 1 and 2, respectively) and piperitenone oxide (accounting for more than 50% and more than 40% of factors 1 and 2, respectively), along with p-menthone, linalool, and menthol (accounting in total for about 20% of factor 2). In the case of HS-SPME data, the variables providing the greatest contribution to the first two principal components were carvone (accounting for more than 70% and 15% of components 1 and 2, respectively) and piperitenone oxide (accounting for more than 20% of component 1 and nearly 60% of component 2), along with linalool, accounting for more than 10%, and p-methone, cis-dihydrocarvone, and menthol, accounting for less than 5% each of component 2. In Figures 3 and 4, the first five compounds exhibiting major contribution to the variance are presented. The samples naturally formed three groups: chemotypes differentiated by the abundance of piperitenone oxide, carvone or linalool, and menthol, p-menthone, and other related compounds. The clustering was similar based on both EO and HS-SPME data.

Plant Material
The

Plant Material
The leaf samples of seven cultivated mints were obtained from the collections at the Botanical Garden of Medicinal Plants ( Wrocław, Poland), Wroclaw Medical University

Essential Oil Hydrodistillation (HD)
From dried leaves of seven taxa of the genus Mentha, EOs were obtained using a modified pharmacopeial hydrodistillation method according to the monograph of Menthae piperitae folium in the European Pharmacopoeia [37]. Briefly, 20.0 g of dried and crushed plant material was added to a 500 mL flask with 200 mL of water. It was connected to a Deryng apparatus with a water cooler and heated in a laboratory electric heating mantle with thermal controller at the boiling point of the water. Distillation was carried out for 2 h. No xylene was used, but pentane and diethyl ether (1:2, v/v) were trapped in a stack. The results obtained from three replicates are presented as mean values.

Headspace Solid-Phase Microextraction (HS-SPME)
An automated PAL RSI SPME holder (CTC Analytics AG, Zwingen, Switzerland) was used containing DVB/CAR/PDMS fiber (Divinylbenzene/Carboxen/Polydimethylsiloxane, 50/30 µm, Supelco Co., Bellefonte, PA, USA). Each sample of the leaves was cut in small pieces, and~2 g was placed in a 20 mL clear screw top vial and hermetically sealed with PTFE/silicone septum. The automated PAL RSI was programmed as follows: conditioning SPME fiber (according to Supelco Co. Instructions, Bellefonte, PA, USA); sample agitation (agitation speed, 250 rpm; agitator on time, 5 s; agitator off time, 2 s; equilibration, 30 min at 40 • C; VOC extraction, 40 min without agitation; injection of the fiber in the GC injector for desorption, 7 min. The results are presented as mean values of three replicates.

GC-MS Analysis
An Agilent Technologies (Palo Alto, CA, USA) gas chromatograph model 7820A equipped with a mass selective detector (MSD) model 5977E (Agilent Technologies, Palo Alto, CA, USA) was used. An HP-5MS capillary column (Agilent J & W GC column, Palo Alto, CA, USA; 5%-phenyl-methylpolysiloxane; 30 m × 0.25 mm, 0.25 µm film thickness) was used for GC-MS analysis. The GC conditions were as follows: split ratio, 1:50; oven programmed, 2 min at 70 • C, with the temperature increased at the rate of 3 • C/min to 200 • C and held isothermal for 15 min; injector temperature, 250 • C; detector temperature, 300 • C; carrier gas, He (velocity: 1 mL/min). The MSD (EI mode) was operated at 70 eV, and the mass range was 30-300 average mass units (amu). Firstly, 1 µL of diluted essential oil (10 µL of the oil in 1 mL of pentane) was manually inserted with a syringe into the GC injector. The identification of the VOCs was based on the comparison of their retention indices (RIs) determined relative to n-alkanes (C9-C25) with those reported in the literature and their mass spectra were determined via comparison with the spectra from Wiley 9 (Wiley, New York, NY, USA) and NIST 17 (D-Gaithersburg) mass spectral libraries. The percentage composition of the samples was computed from the GC peak areas using the normalization method (without correction factors). The average component percentages were calculated from three GC-MS analyses.