Plant Extract Valorization of Melissa officinalis L. for Agroindustrial Purposes through Their Biochemical Properties and Biological Activities

Lemon balm (Melissa officinalis L.) is one of the rare medicinal plants in Tunisia. It was found only in two sites in the north of Tunisia with a small number of plants. The study of germination under the NaCl and PEG effect showed that Tunisian lemon balm seeds were sensitive to saline and osmotic stress. Morphological and biochemical characterizations of Tunisian M. officinalis were performed. Results showed that the Tunisian populations presented plants with long, broad leaves and weak branching. The major constituent in leaf essential oil was germacrene-D with a percentage ranging from 29.17 to 24.6%, and the major fatty acids were polyunsaturated fatty acids, linoleic acid, ranging from 73.93 to 66.74%. The phenolic content of M. officinalis extract varied significantly among origins which could explain the high variation in antiradical scavenging activity. The evaluation of allelopathic activities showed that the extract of the lemon balm leaves presented an allelopathic effect with the majority of the tested seeds.


Introduction
During the last two decades, research in herbal medicine has become one of the greatest scientific concerns [1]. In the socioeconomic context, the study of plants can lead to the achievement of adequate and low-cost therapeutic responses, combining proven scientific efficacy and cultural acceptability [2]. Located in the Mediterranean basin with great climatic variations from North to South, Tunisia presents a favorite terrain for the development of a flora rich in medicinal and aromatic species. More than 500 species out of a total of 2,200 (about 25% of the total flora) are considered for therapeutic use [3]. Melissa officinalis L. or lemon balm is a very rare medicinal species in a spontaneous area in Tunisia, classified among the 48 species identified as endemic rare and endangered according to the IUCN classification [4]. It has been encountered in few sites with a reduced number of plants in the forest region of Kroumirie and Mogods, having the geographical coordinates 8°-8°30′ East for the longitude and 36°15′-36°45 ′North for the latitude [5]. Lemon balm is sought after for its lemony scent as well as for its many therapeutic activities. Since ancient times, it is used in cases of nervousness and minor sleep disorders, as well as in case of gastrointestinal disorders such as flatulence and abdominal pain, especially antidepressants [6]. In recent years, other uses have been studied, particularly in the field of plant protection products [7]. It is empirical that these properties have been attributed to it [8].
In Tunisia, there are no reports conducted on M. officinalis. erefore, the aims of this study were to investigate, for the first time, the germination rate under abiotic stress and the morphological and chemical characterization and to evaluate their antioxidant and allelopathic activities which may provide data for suitable conditions for cultivation of the best populations and for their agroindustrial exploitation.

Plant Material.
e plant material was botanically characterized by Dr Ben Brahim N. (Laboratory of Science and Agricultural Techniques, National Agricultural Research Institute of Tunisia (INRAT)) according to the morphological descriptions in the Tunisian flora. Several surveys were carried out before and after flowering (May and August) on the Tunisian territory according to the data of the geographical distribution of El Mokni et al. [9]. Tabarka and Nefza (Figure 1) cover almost the entire area of M. officinalis in Tunisia. Tunisian seeds were harvested from sites found in northern Tunisia. French seeds of M. officinalis were provided by the National Institute for Agricultural Research, France, and the German seeds were provided by the Institute for Food and Resource Economics (ILR), University of Bonn.
e introduced seeds were used as references.

Study of Germination under Abiotic Stress.
Seed germination is an essential process in the development of the plant. It is influenced by many abiotic factors such as salt and drought stress, which are perhaps the two most important abiotic constraints limiting plant development [10,11]. In order to eliminate the germination inhibitors, the seeds, which had the same age, were sterilized in 0.5% sodium hypochlorite solution for 1 min and then washed with distilled water solution. Germination was performed in incubators with a photoperiod of 12 h (Sylvania white fluorescent lamps, 25 mM·m −2 ·s −1 photons with photosynthetically active radiation). Each treatment consisted of three replicates of 50 seeds. e seeds were germinated in NaCl solutions (0; 50; 100; 200; and 300 mM) and PEG 600 (polyethylene glycol) (0; −2.3; −4.6; −9.3; −13.9 bar) under the optimum temperature of 20°C [12]. e seeds were considered germinated during the appearance of radicals [13]. Sprouted seeds were counted and eliminated every two days over a 14-day period [14,15].

Culture.
e seeds of M. officinalis were grown on an experimental site at the National Institute of Agronomic Research of Tunisia (Ariana, 36°50′, 10°11′E), at 10 m altitude, with alkaline soil (pH � 8.7), and 450 mm of rain and upper semiarid bioclimatic stage. e experimental setup included lines spaced 1m and 50 cm apart between the plants. Each line comprised 10 individuals at the rate of three replicates for each origin. e average temperature was 25°C; the average monthly temperature varied between 15°C in January and 35°C in August. e plants were developed under biological conditions (no pesticide and fertilizer, weeds were eliminated manually, and additional irrigation if necessary was made). e cultivation was followed until the collection of the seeds. Eleven morphological  characters were measured from 26 plants for each population. ese characters relate to vegetative and reproductive developments. e choice of these characters was inspired by previous work done on several Lamiaceae plants such as Lavandula species [16].

Isolation of Essential Oil and GC/MS Analysis.
e essential oils were obtained from 100 g (dry weight) of plant material using a Clevenger-type apparatus for 3 h. e hydrodistillation was performed in three replicates, and the oils were stored at 4°C until analysis by GC/MS. e average oil yields were estimated on the basis of the dry weight of the plant material. e oils were analyzed with a Hewlett-Packard 6890N/5975B inert GC-MSD system (Agilent, USA) equipped with two cap. columns, a HP-INNOWAX (30 m × 0.25 mm i.d., film thickness 0.25 μm), and a HP-5MS (30 m × 0.25 mm i.d., film thickness 0.25 μm) column (J&W Scientific, USA). e oven temperature was programmed isothermal at 50°C for 1 min, then increased from 50 to 250°C at 28/min, and finally held isothermal at 250°C for 15 min (injector temperature, 250°C; ion source temperature, 230°C; carrier gas, He (high purity 99.99%; 1.2 ml·min −1 ); injection volume, 1 μl; split ratio, 100 : 1). e electron impact ionization mode was used with an ionization voltage of 70 eV. Total ion chromatograms were obtained over the scan range of 30-800 a.m.u. in the full-scan acquisition mode, and the compounds were identified using the NIST05 and Wiley 7 databases with a resemblance percentage above 85%. Semiquantitative data were calculated from the GC peak areas without using correction factors and were expressed as relative percentage (peak area %) of the total volatile constituents identified. Retention indices (RIs) were determined for all the detected compounds based on the retention times (tr) of a homologous series of n-alkanes (C8-C30) [17,18].

Fatty Acid Methyl Ester Preparation.
Triplicate samples of 1 g of M. officinalis leaves were subjected to lipid extraction using a modified version of the Bligh and Dyer [19] method. us, leaf samples were kept in boiling water for 5 minutes and then ground manually using a mortar and pestle; chloroform/methanol mixture (2 : 1 v/v) was used for lipid extraction. After washing using fixation water, the organic layer containing lipids was recovered and dried under a nitrogen stream. Total fatty acids (TFAs) of total lipids were transmethylated using sodium methoxide solution (3% in methanol) [20]. Methyl heptadecanoate (C17 : 0) was used as an internal standard. e fatty acid methyl esters (FAMEs) obtained were subjected to GC analyses.

Phenolic Content and DPPH Radical Scavenging Assay.
M. officinalis extracts were prepared with two solvents of ethanol and water. 20 g of plant powder was stirred in the presence of 200 ml of solvent for 72 h and then filtered. e filtrate was evaporated in the rotavapor [21]. e yield was calculated by the following formula: % yield � mass of the extract obtained mass of plant material before the extraction × 100.
(1) e total phenolic content of ethanolic and aqueous extracts was determined according to the method described by Lowman and Box and slightly modified by Moghaddam and Miran [22,23]. About 100 μL of each sample was mixed with 46 mL of distilled water and 1 mL Folin-Ciocalteu reagent.
e mixture was thoroughly shaken and left for 3 min before adding 2.9 mL of Na2CO3 (2%). After incubation for 2 hr in dark, the absorbance was measured at 760 nm. A standard curve of gallic acid was prepared. Total phenolic contents were expressed as milligrams gallic acid equivalents per gram of the essential oil (mg GAE/g DW) through the calibration curve with gallic acid. All samples were analyzed in triplicate. e antioxidant activity was determined using the 1,1diphenyl-2-picrylhydrazyl (DPPH) radical scavenging method according to Tohidi and Rahimmalek [24]. About 100 mL of each sample at different concentrations (between 50 and 800 μg/mL) was added to 2,500 mL DPPH ethanolic solution.
e mixtures were shaken vigorously and then placed in the dark for 30 min. e absorbance of the solutions was measured at 517 nm. All samples were analyzed in triplicate. Butylated hydroxytoluene (BHT) was used as Journal of Chemistry positive control. e antiradical activity was expressed as IC 50 (in mg/mL) which was defined as the concentration of the sample required to inhibit the formation of DPPH radicals by 50%. e percentage inhibition of the DPPH radical was calculated according to the following equation: where A control is the absorbance of the DPPH solution without sample solution and A sample is the absorbance of the sample after 30 min at 517 nm.

Allelopathic
Activity. e extracts of lemon balm have been studied for their herbicidal effect. ese extracts were tested on four seeds: radish (Raphanus sativus L.), fenugreek (Trigonella foenum-graecum L.), wild chicory (Cichorium intybus L.), medicago (Medicago polymorpha L.). e germination tests are carried out in an oven set at 25°C. For each test, four concentrations were used (0; 1; 3; and 5 mg/ml). e percentage of germination was observed daily for five days. Measurements of root length and shoot length were reported only at the end of the test (total germination of controls). e results represent the average of three replicates of 50 seeds for each treatment. To ensure their germination capacity, they were sterilized with 70% ethanol for 30 seconds, washed with sterile water, and then soaked in sodium hypochlorite for 20 minutes. Germination indices were determined by counting the number of seeds germinated at 24 hour intervals.

Statistical
Analysis. All parameters were statistically described to determine means, standard deviations, and coefficients of variation by population. ANOVA (SAS version 9 software) was performed using the Duncan 5% test for each parameter. Principal component analysis (PCA) is one of the factor analysis methods that reduce data by defining the main axes (or principal components) from the monitored parameters. e observations are represented in relation to the main axes. Group analysis is performed by the ascending hierarchical classification (ACH) method. e graph representing the classification is a dendrogram of dissimilarity with standardized distances representing the nearest population in homogeneous groups. is analysis was performed using JMP ® 11.0 statistical software (SAS Institute, Inc., Cary, NC) with component analysis procedure.

Study of Germination under Abiotic Stress.
e seed germination under different concentrations of sodium chloride showed that the germination rates of Tunisian origins and those introduced decrease when the sodium chloride concentration increases (Table 1). Exposure to osmotic conditions during the germination shows that Tunisian seeds are clearly sensitive to variations in water potential.
e germination rate of Tunisian M. officinalis seeds decreases slightly at −2.3 bars and disappears from −4.6 bars. On the other hand, introduced seeds are relatively more tolerant (Table 2). Seeds from Germany have a germination rate above 40% at −4.6 bars. Introduced seeds, from Germany and France, show similar responses for the two treatments, tolerating the variation of NaCl and PEG than the Tunisian lemon balm seeds. We have little data about the tolerance of lemon balm seeds to abiotic stress. Germination of Tunisian seeds is studied for the first time in order to analyze the influence of salinity and osmotic potential on germination. Significant differences were observed between Tunisian and introduced seeds. Although the data were obtained from seeds sprouted in Petri dishes, the result may be related to in situ performance. Seeds from Tabarka and Nefza are more sensitive to salinity and to the osmotic potential, which explains the reduced number of M. officinalis plants in situ, where climatic constraints exert a natural selection pressure [25]. e germination rate of the lemon balm seeds studied is similar to the germination rate observed by Chartzoulakis and Klapaki, which demonstrated that 50 mM of salinity in the external environment delays germination in some plants but did not reduce the percentage of germination observed at the end of the experiment [26]. Similar results have been reported for Atriplex griffithii and Cressa cretica [27,28]. Indeed, Iranian lemon balm seeds presented a germination rate that reaches 90% at 0 bar and 81% at −2 bar [29]. e reduction of germination rate under salt stress and their toxic and osmotic effects involved in germination have been demonstrated by several authors [25,30]. In the case of Tunisian lemon balm seeds based on the results obtained, it is suggested that the osmotic effect of NaCl is responsible of the disturbance of seed germination. So, the osmotic effect, which leads to reduced water absorption, may be responsible for the inhibition of germination seeds from Tunisia.

Morphological Characterization.
e morphological characterization of M. officinalis from the four origins, Tabarka, Nefza, Germany, and France, was addressed by statistical analysis carried out on 11 agro-morphological characters measured on plants cultivated under homogeneous conditions. One-way analysis of variance (population effect) showed significant differences for most of the traits studied (Table 3). To understand the variance sources of M. officinalis, a principal component analysis (PCA) was performed by grouping together the seven significant morphological characters in the first two axes describing 79.3% of the total variation (Figure 2(a)). e plot drawn along the two axes showed three distinct groups of plants (Figure 2(b)). e first group consisted of the plants of the German population, which was characterized by plants with good vegetative development presenting the tallest plants (43.34 cm), the most branched, and the highest number of leaves and flowers. e Tunisian populations (Tabarka and Nefza) formed the second group, which was characterized by plants with long, broad leaves and weak branching. e third group was formed by the plants from France population, which was distinguished by a weak vegetative development. Hierarchical analysis based on morphological characters using the Euclidean square method allowed establishment of the relationship between lemon balm of different origins.
e morphological study showed a large phenotypic diversity between the populations for the majority of the agromorphological characters. is indicated the existence of a wide range of genetic variability for the traits studied and highlighted the potential from genetic improvement using such a gene pool. Sari and Ceylan also observed a great variability of morphological characters in 11 populations of M. officinalis from different regions of Turkey and European countries [31]. is may be due to cross pollination of lemon balm.

Essential Oil.
e M. officinalis samples cultivated under the climatic conditions of the INRAT yielded a small amount of essential oil. e oils were analyzed by GC/MS. e chemical composition of the four M. officinalis from different origins was reported in Table 4. Chemical analysis of the essential oil samples conducted according to their retention time (RT) revealed the presence of 42 compounds, representing about (86.11%, 83.1%, 96.72%, and 71.88%) of the total oils obtained from Nefza, Tabarka, Germany, and France, respectively. In addition to the differences in the essential oil yield, the GC/MS analyses revealed qualitative and quantitative differences in the composition between the oils of the four origins. GC/MS analysis showed that the oils of the German population were characterized by the presence of 39.31% of geranial and 27.71% neral, which were the dominant components, and 12.23% of β-caryophyllene. e sesquiterpene caryophyllene oxide (27.06%) was found to have the highest value in the French population, which exhibited lower levels (7.12-4.29%, respectively) of geranial and neral. Germacrene-D (32.08-27.06%) was the highest in the Tunisian samples from Tabarka and Nefza, together with the sesquiterpene caryophyllene (16.4-14.7%, respectively). Considering the fact that minor components of the essential oils were less important for a comparison, only the major components representing more than 2% of the essential oil composition were taken into account to illustrate the variation of the components content from different origins [17].
In order to investigate the similarity and relationship between essential oil compositions of our studied samples, a hierarchical cluster analysis (HCA) was performed based on the components. e dendrogram of HCA classified the oils of the Tunisian populations of Tabarka and Nefza in the same group. is group was characterized by the predominance of germacrene-D as the major compound and emphasized the distinctiveness of the French oil, rich in caryophyllene oxide. e third group was formed by the German oil characterized by citral (neral and geranial) (Figure 3).

Fatty Acids.
To the best of our knowledge, the foliar fatty acid composition of M. officinalis was reported herein for the first time. e proportion of the majority of fatty acids did not show differences according to the origin of samples (Table 5). Melissa officinalis leaves were characterized by a high proportion of polyunsaturated fatty acid (PUFA), linoleic acid (73.93-66.74%), versus (16.25-13.32%) saturated ones (SFA), oleic acid, and (6.29-4.26%) of monounsaturated (MUFA) palmitic acid. e comparison between different lemon balm evidenced a similarity, at least with reference to the presence of the main fatty acid constituents. It is noteworthy that previous findings showed that the genera Satureja, Origanum, and ymus of the Lamiaceae family had some minor variations in fatty acid composition, which were dominated by the chemotypes of linoleic acid, palmitic acid, and linolenic acid [32].

Total Phenolic Content and Antioxidant Activity.
Total phenolic contents in lemon balm leaves from different origins are shown in Table 6. e amounts of total phenolics in M. officinalis varied significantly among populations. Ethanolic extract from Tunisian populations Tabarka and Nefza showed the highest amount of phenolic compounds (0.94 mg GAE/g DR and 0.87 mg GAE/g DR, respectively). However, the lowest content was observed in the aqueous extract from German population (0.2 mg GAE/g DR) (Figure 4). e extraction yield was influenced by the polarity of the solvent; according to Bourgou, the water (5.2) is more polar than ethanol (9.0), which is why the yield ethanol extraction is lower than the water [33]. When comparing values in this study to the literature, there was considerable variation in total phenolic contents reported of lemon balm from different origins. Boneza and Niemeyer indicated that the average total phenolic contents for the five lemon balm cultivars from the USA ranged from 5.50 ± 1.03 mg GAE/g DW to 26.87 ± 1.93 mg GAE/g DW [34]. Wojdyło et al. and Dastmalchi and Dorman reported that unspecific lemon balm showed total phenolic contents from 0.13 to 269 mg GAE/g DW [35,36]. Boneza and Niemeyer noted much higher total phenolic contents for the lemon balm cultivars, with values ranging from 359 mg GAE/g DW for the "Lorelei" cultivar to 427 mg GAE/g DW for "Gold Leaf" [34].
is variability in lemon balm total phenolic contents across studies was therefore most likely due to differences in plant      Table 7. ere was significant variation in free radical scavenging activities among populations, and the inhibitory concentrations IC50 ranged from 123,29 to 541,81 mg/mL    In accordance with the data obtained from this study, Aissi et al. reported the effect of geographic origin on antioxidant activity of essential oils [37]. Indeed, these variations were probably due to differences in phenolic compound content [38,39].

Journal of Chemistry
In fact, phenolics, due to their hydroxyl groups that allow them to donate hydrogen to DPPH free radicals, were considered as the major factor contributing to antioxidant activity of plants [40]. Several studies reported by Hammoudi et al., Habellah et al., and Tlili et al. showed a correlation between total phenolic content and antioxidant activity [41][42][43]. Cultivar also influenced DPPH antioxidant capacity (P � 0.002) with "Lemonella" (45.2 ± 11.2 μmol TEAC/g DW) having significantly lower DPPH free radical scavenging than all other cultivars in the study. "Lime" balm also had the highest DPPH antioxidant capacity, 189.5 ± 19.7 μmol TEAC/ g DW. Samples having higher phenolic levels typically exhibit greater antioxidant capacity [44].

Allelopathic Activity.
e inhibitory effect of the aqueous extract of Melissa officinalis from different origins on seed germination was clearly visible from the lower concentrations. In the case of Medicago, the aqueous extract of lemon balm had the highest inhibition levels (Table 8).
e ethanol extract was tested only on the seeds of medicago and chicory, given the low yield of extract. e results showed that the rate of germination was strictly concentration dependent ( Figure 5). e study of IC 50 of ethanolic extract from the four origins showed close concentrations with chicory seeds. However, with the medicago seeds, the ethanol extract from Tunisian M. officinalis showed the lowest IC 50 . To follow the effect of the ethanolic extract of Tunisian Lemon balm on the development of seedlings, the length of the root and the shoot were measured under the effect of different concentrations. e longest root was recorded in untreated seeds, and the shortest length was recorded in seeds treated with 5 mg/ml. For the shoot, the ethanolic extract of lemon balm significantly reduced its length depending on the concentration in both species. Variance analysis showed significant differences (P < 0.05) between treatments (Tables 9 and 10). From a physiological point of view, germination begins with the imbibitions of the seed and ends with the beginning of growth marked by the lengthening of the root [45]. e germination of a seed can take place only if certain favorable conditions are met: oxygen, temperature, and water. Moreover, it is well known that natural substances produced by plants are able to delay or inhibit seed germination and seedling growth. is explains the effects of aqueous and ethanolic extracts of M. officinalis on the germination of certain seeds. e results obtained from the germination tests showed the existence of an allelopathic phenomenon under experimental conditions. is provided evidence that some plants contain chemical compounds with herbicidal activity. e aqueous extract of lemon balm from Tunisia was shown as the most inhibitor against the majority of seeds tested. e ethanolic extract showed a greater allelopathic power against the Medicago seeds than on the chicory seeds. e allelopathic effect may depend on the species of seeds tested, which was provided by    Serghini et al. [46], where sunflower extract (Helianthus annus L) had an effect on the germination of Orobanche ramose (Phelipanche ramose L.) but had no effect on the germination of Orobanche cernua Wallr. e results obtained corroborated with those found by Kato-Noguchi and Ino in which certain fractions of a hydroacetonic extract of M. officinalis inhibited the germination and the growth of roots and shoots of Amaranthus caudatus L., Lepidium sativum L., Digitaria sanguinalis L., Phleum pratense L, Lactuca sativa L., and Lolium multiflorum Lam. [6]. e aqueous extract is the most effective with an IC 50 being 0.14 mg/ml.

Conclusion
Exploration of native plant genetic resources is one of the main objectives of the Medicinal and Aromatic Plants Program. It also aims to evaluate and conserve rare spontaneous plants, genetic resources, and especially threatened species in their natural environment. e seeds of Tunisian lemon balm were the most sensitive to salinity and the variation of the osmotic pressure, which explains the reduced number of plants in situ. e morphological study showed that Tunisian lemon balm was characterized by small plants with a weak branching. e aqueous and ethanolic extracts of this species had an interesting allelopathic activity, which allowed the use of lemon balm in the field of phytosanitary products as weed killers.

Data Availability
All data generated or analyzed during this study are included in this published article.

Conflicts of Interest
e authors declare that there are no conflicts of interest. Means in each line followed by the same letters are not significantly different (P > 0.05). Means in each line followed by the same letters are not significantly different (P > 0.05).