Chemical Characterization and Antibacterial Activity of Essential Oil of Medicinal Plants from Eastern Serbia

The objective of this study was to evaluate wild growing Satureja kitaibelii, Thymus serpyllum, Origanum vulgare, Achillea millefolium and Achillea clypeolata with respect to their essential oil (EO) content, composition and antimicrobial activity. The five species were collected at Mt. Rtanj and the village of Sesalac, Eastern Serbia. The main EO constituents of Lamiaceae plants were p-cymene (24.4%), geraniol (63.4%) and germacrene D (21.5%) in Satureja kitaibelii, Thymus serpyllum and Origanum vulgare ssp. vulgare, respectively. A. millefolium EO had multiple constituents with major ones being camphor (9.8%), caryophyllene oxide (6.5%), terpinen-4-ol (6.3%) and 1,8-cineole (5.6%), while the main EO constituents of A. clypeolata were 1,8-cineole (45.1%) and camphor (18.2%). Antimicrobial testing of the EO showed that Staphylococcus aureus (Gram-positive) was more sensitive to all of the tested EOs than Escherichia coli (Gram-negative). S. kitaibelii EO showed the highest antimicrobial activity against both tested bacterial strains. This is the first study to characterize the EO composition and antimicrobial activity of these five medicinal species from Eastern Serbia in comparison with comprehensive literature data. The results can be utilized by the perfumery, cosmetics, food and pharmaceutical industries, but also for healing purposes in self-medication.


Introduction
The plants of the Lamiaceae and Asteraceae families are widely distributed medicinal plants throughout the world and have been used since ancient times for medicine and food. Historically, medicinal plants from these families have been used for flavoring, food preservation and medicinal purposes, due to their bioactive properties [1][2][3]. Today, pharmaceutical companies are interested in the wide range of beneficial properties of these plants in order to develop modern herbal remedies that would be used either as a replacement or supplement to conventional medicines and for prevention of illnesses [4][5][6]. In addition, medicinal plants are valuable raw material for perfumery and cosmetics, but also for healing purposes in self-medication [7][8][9], as well as in organic agriculture [10][11][12].

Antibacterial Activity
According to the assay, S. aureus was more sensitive to the tested EOs than E. coli. However, S. kitaibelii expressed the most potent activity against both G-positive and G-negative bacteria, followed by T. serpyllum and A. clypeolata, while O. vulgare ssp. vulgare showed the weakest antibacterial activity. A. millefolium showed weak activity against E. coli but was very potent against S. aureus (Table 2). Further investigations will be focused on studying major constituents of S. kitaibelii, T. serpyllum, O. vulgare ssp. vulgare, A. millefolium and A. clypeolata EOs, such as geraniol, 1,8-cineole, p-cymene, germacrene D, nerol, camphor, sabinene, limonene, trans-caryophyllene, etc., to evaluate concentrations of the components that could be responsible for the antibacterial effect. Additional tests with a larger number of bacteria regarding the synergic potential of EOs will be implemented in our future investigations.

Satureja kitaibelii
S. kitaibelii EO constituents from this study and from literature reports [18,[36][37][38][39][40] are shown in Table 3. The classification between them according to the content of chemical compounds was performed using HCA analysis (Figure 1a). This analysis showed that differences in EO composition could be separated into several potential clusters (chemotypes): geraniol, p-cymene and limonene, with high abundance of linalool and borneol, which is in accordance with a previous study [18].  Taking into account the chemical content range, the correlation network based on components from S. kitaibelii EO is shown in Figure 1b. There were strong positive correlations between p-cymene and limonene (r = 0.75; p < 0.01), as well as between α-pinene and sabinene hydrate (r = 0.76; p < 0.01). However, the strongest negative correlation was between geraniol and compounds such as p-cymene (r = −0.86 p < 0.01), limonene (r = −0.74; p < 0.01) and borneol (r = −0.66; p < 0.01).
The major constituent of extracts and EO used in traditional medicines as an antimicrobial agent is p-cymene [41]. However, in this study, p-cymene exhibited the weakest antibacterial activity against S. aureus and E. coli [42]. Contrarily, limonene exhibited moderate antibacterial activity against these bacteria [43], while linalool exhibited strong antimicrobial activity [44]. It is well known that many plants exert their beneficial effects through the additive or synergistic action of several chemical compounds acting at single or multiple target sites [45]. The results from this study suggest that the combination effects of these compounds (p-cymene, limonene and linalool) in S. kitaibelii EO had Taking into account the chemical content range, the correlation network based on components from S. kitaibelii EO is shown in Figure 1b. There were strong positive correlations between p-cymene and limonene (r = 0.75; p < 0.01), as well as between α-pinene and sabinene hydrate (r = 0.76; p < 0.01). However, the strongest negative correlation was between geraniol and compounds such as p-cymene (r = −0.86 p < 0.01), limonene (r = −0.74; p < 0.01) and borneol (r = −0.66; p < 0.01).

Locality
Ref.   Data in Figure 2a suggests the presence of 11 potential chemotypes in T. serpyllum: geraniol, citronellol, linalool, α-terpinyl acetate, trans-nerolidol, trans-nerolidol + caryophyllene oxide, caryophyllene oxide, carvacrol, thymol, thymol + carvacrol and multiple-component chemotypes. Previous research on T. vulgaris described genetically distinct chemotypes that can be distinguished on the basis of the dominant monoterpene produced in the glandular trichomes. It was established that the monoterpene variations in T. vulgaris plants may represent an adaptive strategy in relation to the environmental variations, as the different chemotypes showed different geographic and locality distribution [59].
Taking into account the chemical content range, the correlation network based on components from T. serpyllum EO is shown in Figure 2b. There were strong negative multiple correlations between sabinene and β-caryophyllene, caryophyllene oxide and germacrene D (p < 0.01), but all these three compounds are in positive correlations (p < 0.01).
Previous research reported that geraniol exhibited good antimicrobial activity and, in combination with antibiotics, would have substantial therapeutic potential against S. aureus and E. coli infections [60]. Although geraniol and nerol are geometric isomers, they demonstrated equal activities [61]. Taking into account the moderate antibacterial activity against S. aureus and E. coli, EOs of T. vulgaris (with geraniol and nerol) could be used as herbal supplements to conventional therapy.
An antimicrobial effect of synthetic antibiotic in combination with germacrene D showing growth inhibition on E. coli and S. aureus [77] is noted. Conversely, 1,8-cineole derivatives displayed significant antibacterial activity [78], while trans-caryophyllene displayed moderate antibacterial activity [79]. Sabinene exhibited prominent antibiofilm properties against E. coli and S. aureus providing a novel and effective alternative/complementary approach to counteract chronic infections and the transmission of diseases in clinical settings [80]. The weak antibacterial activity of O. vulgare ssp. vulgare EO could be attributed to low concentrations of bioactive compounds and their inability to exhibit activity. An antimicrobial effect of synthetic antibiotic in combination with germacrene D showing growth inhibition on E. coli and S. aureus [77] is noted. Conversely, 1,8-cineole derivatives displayed significant antibacterial activity [78], while trans-caryophyllene displayed moderate antibacterial activity [79]. Sabinene exhibited prominent antibiofilm properties against E. coli and S. aureus providing a novel and effective alternative/complementary approach to counteract chronic infections and the transmission of diseases in clinical settings [80]. The weak antibacterial activity of O. vulgare ssp. vulgare EO could be attributed to low concentrations of bioactive compounds and their inability to exhibit activity.

Achillea millefolium
A. millefolium represents a polyploidic complex of hardly distinguishable species, subspecies, forms and hybrids [81]. However, morphological, chemical and molecular traits as well as PCA analysis showed that the terpenoid variation can be used to explore biogenetic pathways [82]. The
Taking into account the chemical content range, a correlation network based on components from EO of this species is shown in Figure 4b. The strongest negative multiple correlations were found between β-caryophyllene and terpinene-4-ol (r = −0.61; p < 0.01) and the strongest positive correlations were between β-caryophyllene and chamazulene (r = 0.56; p < 0.01).
The A. millefolium EO showed antimicrobial activity in vitro against Streptococcus pneumoniae, Clostridium perfringens, C. albicans, C. krusei, Mycobacterium smegmatis and Acinetobacter lwoffii [88]. Additionally, in vitro antibacterial activity against nine Gram positive and negative bacteria (S. epidermidis, S. aureus, B. cereus, E. faecalis, E. coli, P. aeruginosa, K. pneumoniae, S. typhimurium and Shigella dysenteria) demonstrated that A. millefolium EO can potentially be used for controlling certain bacteria that cause many infectious diseases, but its effectiveness varied in different regions because of the differences in EO composition [87].

Achillea clypeolata
The main compounds in the A. clypeolata EO from Mt. Rtanj (collected in July 1996) were 1,8-cineole (38.6%) and camphor (19.9%) [95]. The concentration of 1,8 cineole in A. clypeolata in this study was a bit higher. This variation in the chemical composition could be attributed to the weather conditions during the year, collection time, population and exposition. Because this is an endemic species in the Balkan region, there were very few studies on its EO composition. A. clypeolata of Serbian origin contained 1,8-cineole as the dominant compound [95,96], and 1,8-cineole as a potential chemotype of this species was confirmed by this study. However, A. clypeolata grown in a botanical garden in Italy could be characterized as β-pinene chemotype [97].
Taking into account the chemical content range, a correlation network based on components from EO of this species is shown in Figure 5b. There were strong negative correlations between terpinen-4-ol and β-pinene as well as β-pinene and borneol. Positive correlations are noted between terpinen-4-ol and borneol, as well as multiple correlation between allo-aromadendrene, epi-α-cadinol and α-cadinol.
A. clypeolata has not been investigated thoroughly yet, with the exception of its antioxidant and antimicrobial properties [96,98]. In this study, the EO of A. clypeolata exhibited the strongest activity against K. pneumoniae and P. aeruginosa, even stronger than the antibiotic Ampicilin, which was used as the standard for comparison. A. clypeolata EO showed lower activity against Gram-positive S. aureus, but still stronger than Ampicilin, while E. coli was the most resistant to the oil [96]. Indeed, previous study on Cinnamomum longepaniculatum leaf EO has shown that its compounds have excellent antibacterial activities, and the antibacterial mechanism of 1,8-cineole against E. coli and S. aureus might be attributed to its hydrophobicity [99].  A. clypeolata has not been investigated thoroughly yet, with the exception of its antioxidant and antimicrobial properties [96,98]. In this study, the EO of A. clypeolata exhibited the strongest activity against K. pneumoniae and P. aeruginosa, even stronger than the antibiotic Ampicilin, which was used as the standard for comparison. A. clypeolata EO showed lower activity against Gram-positive

Plant Material
The plant species were collected at full flowering stage in July 2018. Satureja kitaibelii Wierzb. ex Heuff., Origanum vulgare L. ssp. vulgare, Achillea millefolium L. sensu lato and A. clypeolata Sibth. & Sm were collected on Mt. Rtanj and Thymus serpyllum L. was collected in the village of Sesalac. The aboveground parts were harvested by cutting them manually at around 2-3 cm above the soil surface and were then dried in a shady well-aerated place to a constant mass.

Essential Oil (EO) Extraction
Clevenger apparatus was used to extract the essential oil from the air-dried aboveground parts of each sample in three replications. According to the European Pharmacopoeia, 30 g each of the drug plants from the Lamiaceae family (S. kitaibelii, T. serpyllum and O. vulgare ssp. vulgare) and 400 mL of water and 20 g each of cut drug from the Asteraceae family (A. millefolium and A. clypeolata) and 500 mL of water were placed separately in 1000 mL round bottom flasks and distilled for 3 h [100].

Essential Oil (EO) Analysis
Gas chromatography-mass spectrometry (GC-MS) was performed using an HP 5890 coupled with an HP 5973 MSD fitted with an HP-5MS capillary column (conditions in detail described by Aćimović et al. [101]). The components were identified based on their linear retention index relative to C8-C32 n-alkanes, in comparison with data reported in the literature (Wiley and NIST databases). The percentage (relative) of the identified compounds was computed from the GC peak area.

Association among Chemical Compounds
The association of the chemical compounds of EOs was estimated using the Spearman non-parametric correlation coefficient. In order to further analyze and represent the associations among the chemical compounds, a correlation network graph [102] was used. In this type of graph, the chemical compound variables are represented by the nodes, which are connected by the edge whose width is directly proportional to the strength of the correlation coefficient. The correlation network graph is easier to interpret than a numerical correlation matrix, and the pattern of the correlations, i.e., the clusters of the correlated variables, can be visually identified [103].
A data set, composed of 16 S. kitaibelii, 54 T. serpyllum, 53 O. vulgare ssp. vulgare, 63 A. millefolium and 4 A. clypeolata samples and 13 variables (the main EO constituents), were depicted using a correlation network graph for visualization correlations between chemical compounds from EOs (Spearman's rank order correlation coefficient, r). Non-significant correlations were removed from the correlation network graph.

Antibacterial Activity
The antimicrobial activities of the tested EOs were investigated using American Type Culture Collection test strains of Escherichia coli (ATCC 8739) and Staphylococcus aureus (ATCC 25923). Strains were cultured on Tryptone Soya Agar (TSA) and incubated at 37 • C for 24 h. Isolated colonies were picked and transferred to 5 mL of Tryptone Soy Broth (TSB) and incubated at 37 • C for 18 h. The density of the suspensions used for tests was adjusted to 0.5 Mc Farland units (~1-2 × 10 8 CFU/mL) using a densitometer DEN-1 and standard plate counts. Efficacy of EOs on microorganisms was determined according to the CLSI (2018) [104] with slight modifications [105].
Tested EO was dissolved with sterile water supplemented with 0.05% Tween 80 and added to 96-well microtiter plates at concentrations from 200 to 0.781 µL/mL (the final concentration in microtiter plates was from 20 to 0.078 µL/mL). The 160 µL in Mueller-Hinton Broth (MHB) were also added to each well, and in the end, 20 µL of overnight bacterial cultures suspensions were inoculated. Plates were incubated at 37 • C for 24 h. After incubation, 20 µL of the resazurin solution (0.01%) were added to each well, and the plates were further incubated at 37 • C for 24 h in darkness. A change of color from blue (oxidized-resazurin remained unchanged) to pink (reduced) indicated the growth of bacteria.
Referring to the results of the minimal inhibitory concentration (MIC), the wells showing a complete absence of growth were identified and 100 µL of the solutions from each well were transferred to plate count agar plates (PCA) and incubated at 37 • C for 24 h. The minimal bactericidal concentration (MBC) was defined as the lowest concentration of the EOs at which 99.9% of the inoculated microorganisms were killed.

Statistical Analysis
Experimental results were expressed as mean ± standard deviation, with 10 repetitions for microbiological analyses. Analysis of variance (ANOVA) for comparison of sample means and a post-hoc Tukey's HSD test was used to analyze variations in observed parameters among the samples.
Hierarchical cluster analysis (HCA) was used to evaluate intra-and interpopulation variability and differentiation of EO constituents of S. kitaibelii, T. serpyllum, O. vulgare ssp. vulgare, A. millefolium and A. clypeolata in samples collected in different locations and/or taken from literature reports. Data was analyzed using StatSoft Statistica 12.

Conclusions
This is the first study to characterize EO composition and antimicrobial activity of the five medicinal species from Eastern Serbia against comprehensive literature data. This study demonstrated that S. kitaibelii plants from Mt. Rtanj belonged to the p-cymene chemotype, T. serpyllum plants belonged to the geraniol chemotype, while the O. vulgare ssp. vulgare plants belonged to the germacrene D chemotype. The A. millefolium were recorded as a multicomponent chemotype, while A. clypeolata belonged to 1,8-cineole (45.1%) chemotype. Furthermore, the S. kitaibelii EO was demonstrated to be a promising agent against S. aureus and E. coli bacterial strains. The chemical composition of studied EOs particularly focuses on the main EO constituents, which are assumed to be responsible for the observed antibacterial activity. Further investigations will be focused on studying major constituents of these EOs to evaluate concentrations of the components that could be responsible for the observed effect. Additional tests with larger numbers of bacteria regarding the synergic potential of EOs will be implemented in our future investigations.