The genus Salvia is the largest genus in the Lamiaceae family, including over 900 species spread all over the world. All organs of the Salvia plants contain essential oils (EO), the main components of which are cyclic, acyclic, and aromatic monoterpenoids with the predominance of one or several components [1,2]. Salvia officinalis L. (Common sage, Lamiaceae) leaves are used for diseases of the throat and mouth disorders, minor wounds, skin disorders, and gastrointestinal disorders. Sage has been shown to have significant antibacterial and anti-inflammatory effects [3]. Salvia officinalis essential oil has been implemented to treat diseases like the respiratory, digestive and nervous systems, heart and blood circulation, endocrine, and metabolic diseases. In addition, sage EO has been shown to have antioxidant, carminative, antispasmodic, antiseptic, and astringent properties [4–6].

The herbal drug of European Pharmacopoeia Salvia officinalis folium contains more than 15 ml/kg of EO for the whole leaves and not less than 10 mL/kg for the cut raw material in calculation to the anhydrous drug [7,8]. Salvia officinalis EO content has been varied from 0.1% to 2.8% [9–14]. In aerial parts of S. officinalis more than 120 components of the EO have been discovered. The raw material contains up to 3% EO, the dominant components of which are monoterpenoids: 1,8-cineole (1%–15%), camphor (5%–20%), α-thujone (10%–60%) and β-thujone (4%–36%); sesquiterpenes: β-caryophyllene, α-humulene, and viridiflorol [7]. In addition, borneol, pinene, camphor, elemene, ledene, were found in the Salvia EO [15–20].

Its pharmacological activity largely depends on the composition of the EO, which is inherent in the chemotype of the plant. Adapting to various environmental conditions, sage S. officinalis synthesizes different groups of biologically active substances that help it survive, forming stable characteristics of the chemical composition of the plant, the so-called chemotypes. First of all, adaptive substances are represented by terpenoids and phenolic compounds. EO has a very variable composition depending on the harvesting time, genetics, climate, seasonality, environment, and other factors [21–24]. Phenolic substances, amino acids, and monosaccharides in the composition of S. officinalis were also studied [2,25,26].

The effect of drought on the accumulation of cineole, α-thujone, β-thujone, and camphor in the sage leaves was established (the content of monoterpenes in plants that received a sufficient amount of moisture was compared with those that were in conditions of limited water supply—70% of the optimal). Studies have shown that in arid conditions, sage leaves accumulate a significantly higher concentration of monoterpenes (approximately 33%) than those plants cultivated under optimal irrigation conditions [27].

One of the key terpenes in the S. officinalis EO is thujone, whose contents due to its toxicity should be regulated. Thujone is a neurotoxic terpen and chemotypes with its low content should be preferred. The amount of thujone has to be specified in the given product and its daily exposure has to be below 6.0 mg [8]. The S. officinalis aerial parts have been used in traditional medicine and cookery for centuries. Its leaves are approved for use in the European Union as a coloring, category N₂, with preliminary restrictions on the content of α- and β-thujones in the product (0.5 mg/kg) [28]. In the USA, Sage leaves are permitted for use in food and are recognized as safe (21 CFR 182.10 and 182.20) [29].

Also, S. officinalis leaves contain diterpene bitter principles, triterpenes, steroids, rosmarinic acid (up to 3.3%), flavonoids, and tannins [3,21,30–32].

Previously we studied the content of S. officinalis EO from several countries [33]. The purpose of this work is to determine the EO composition in commercial samples of S. officinalis leaves from nine countries to establish the variability of the content of their components and to identify possible chemotypes of this species with a focus on toxic thujone.

The identified constituents in the leaves EO of the nine S. officinalis samples from different countriesare gained in Table 1. The EO yields in the studied samples were 10.0–24.8 mL/kg (Table 2), which corresponded in all cases to the minimum standard (10 mL/kg) of European Pharmacopoeia for the cut drug [8].

High variation coefficients of the predominant compounds (>1) demonstrated that their content strongly differs from samples to samples. Low variation coefficients (0.56–0.75) are typical for p-cymene, α-terpinene, γ-terpinene, terpinolene, (E)-β-caryophyllene and caryophyllene oxide. Trace amounts (<0.05%) of α-thujene, (E)-β-caryophyllene, and caryophyllene oxide were detected in one sample, (Z)-β-ocimene in three samples, myrtenol in seven samples, and thymol and spathulenol in five samples of the studied EOs.

Twenty-five compounds, representing 80.1%–93.6% of the total EO, were identified in the nine studied sage EO. Such a rather large range presented in Table 1 indicates variability in the EO composition of S. officinalis. All identified components have been previously found in the S. officinalis EO [17,19,22,23]. In all the studied samples, monoterpenoids dominate (70.1%–85.4%), much less sesquiterpenoids (6.9%–13.6%) and the least aromatic compounds (0.3%–1.7%). All EOs are characterized by a strong negative correlation (r = −0.88) between the content of monoterpenoids and sesquiterpenoids (Fig. 1), where symbols are individual. To perform scatter plots (or correlation fields), the application package Statistica of Microsoft Excel was used.

The amount of α-thujone (18.0%–43.0%) and β-thujone (3.0%–8.5%), 1,8-cineole (5.5%–13.0%), bornyl acetate (≤2.5%), camphene (1.5%–7.0%) and camphor (4.5%–24.5%), α-humulene (≤12.0%), α-pinene (1.0%–6.5%), limonene (0.5%–3.0%), and linalool+linalyl acetate (≤1.0%) in EOs for medicinal uses are regulated by ISO 9909:1997 [34]. This standard covers EO production methods, quality requirements and evaluation criteria to ensure its purity and potency. The requirements for the content of α-pinene, camphene and α-humulene are met by all the studied samples (Table 2). The content of 1,8-cineole is significantly higher than the upper limit of normalization in the EO sample from Greece (45.3%) and slightly higher in Ukraine (13.7%). At the same time, the sample of oil from Greece differs from others in its low content of α-thujone (3%) and β-thujone (1.5%), which is significantly below the lower limit of normalization of the content of these compounds. The content of β-thujone exceeds the upper limit of normalization in EO samples from France, Ukraine, and Georgia and is 9.4%, 11.6%, and 11.4%, respectively. In the sample of EO from Austria, the content of camphor and bornyl acetate is above the norm—29.3% and 2.7%, respectively. Limonene, linalool, and linalyl acetate are absent in the studied samples.

The main components in the nine studied EOs were 1,8-cineole (8.3%–45.3%), α-thujone (3.0%–34.0%), сamphor (11.3%–29.3%), β-thujone (1.5%–11.6%), viridiflorol (1.1%–10.4%), camphene (2.6%–6.8%), α-pinene (1.3%–6.4%), borneol (1.8%–5.0%), β-pinene (0.3%–4.9%), (E)-β-caryophyllene (tr.–4.9%), myrcene (0.7%–4.2%), α-humulene (0.4%–6.4%), bornyl acetate (0.1%–2.7%) (Table 2).

In the seven studied EO samples from Estonia, England, France, Hungary, Belgium, Ukraine and Georgia α-thujone (18.6%–34.0%) is the main component. Previously scientific publications also indicate that α-thujone is the dominant component in S. officinalis EOs from Turkey [23], Bulgaria [14], Mexico and California [20], Georgia [35], Romania [19,36], Albania [37], Algeria [38], France and Hungary [33,36], Brazil [39], Ukraine, Belgium, Moldova, and Estonia [33]. High concentrations of β-thujone were reported in EO samples from Turkey [40], Sudan [41], Uzbekistan [42], Portugal and Czech Republic [37], but in the studied EOs there were less amount of it. The high concentrations of β-thujone were just observed in the EOs from Ukraine, Geogia and France. So, it is common for variations in these main chemical components in analyzes of EOs from the same plant species that were cultivated in different countries. Thujones are neurotoxic and their amount are key points in the standardization of the S. officinalis EO [43–45]. The European Union, the USA and other countries have restrictions on the content of α- and β-thujones in products [28,29]. In the USA the addition of pure thujone to food is prohibited and its content must be less than 6.0 mg per day [8]. Therefore, for farms cultivating medicinal plants, it is advisable to recommend Salvia spp. seeds from chemotypes with a low thujone content.

In the studied EO from Greece 1,8-cineole (45.3%) dominates. The dominance of 1,8-cineole in the essential oil from Greece is close to the literature data in EO samples from Jordan [9], Egypt [12], Albania [20], Iran [46], Greece [33], and Poland [47].

In the studied EO from Austria сamphor (29.3%) is the predominant component. Previously the camphor dominance in S. officinalis EOs from Morocco [10,11], Tunisia [16], Romania [19] and Sudan [4] is confirmed.

A high content of camphor (19.2%–19.3%) was found in the EO samples from France, Estonia, and Belgium; 1,8-cineole (13.0%–13.7%)–samples from Hungary, Georgia and Ukraine; β-thujone (9.4%–11.6%)–samples from France, Ukraine and Georgia; α-pinene (5.1%–6.4%)–samples from England, Estonia, Hungary, Belgium and Greece; camphene (5.9%–6.8%)–samples from Austria, Belgium and Greece; β-pinene (2.4%–4.9%)–samples from Estonia, Hungary and Greece; mircene (4.2%)–sample from Greece; borneol (4.7%–5.0%)–samples from Hungary, England and Austria; bornyl acetate (2.1%–2.7%)–samples from Estonia, England and Austria; (E)-β-caryophyllene (2.7%–4.9%)–samples from Ukraine, Hungary and Greece; α-humulene (5.3%–6.4%)–samples from Estonia and England; viridiflorol (7.9%–10.4%)–samples from Hungary, Belgium and Ukraine. Viridiflorol was a principal compound in many samples published previously [48,49].

Our results indicate strong positive correlations between the content of α- and β-thujone (r = 0.73) (Fig. 2); between the content of 1,8-cineole and β-pinene (r = 0.81) (Fig. 3) and a negative correlation between 1,8-cineole and the sum of α- and β-thujone (r = −0.82) (Fig. 4). In these diagrams all symbols are individual. Having received the analysis diagram, we did not determine the value that was far from the totality of data and that needed to be removed. In biology and other natural sciences, a significant (strong) correlation is considered to be a value between 0.3 and 1.0 [50,51].

The statistics show a strong correlation between content of the biologically active substances and Pearson coefficients, which confirms this. The positive strong correlation is evidenced the conjugated biosynthesis and accumulation of these substances in S. officinalis leaves. Our research shows genotypic connections of these substances. There are some publications about correlations between terpenoids, phenolic compounds and ecological minds of production, that testify about their adaptation powers [52–55].

The results obtained by us (Table 2) show that if we take into account the content of four components, the samples we studied correspond to 8 chemotypes (CT): CT1 – α-thujone > camphor > 1,8-cineole > β-thujone (samples from Estonia and Georgia); CT2 – α-thujone > camphor > 1,8-cineole > α-humulene = α-pinene (sample from England); CT3 – α-thujone > camphor > 1,8-cineole > viridiflorol (sample from Hungary); CT4 – α-thujone > camphor > viridiflorol > 1,8-cineole (sample from Belgium); CT5 – α-thujone > camphor > β-thujone > 1,8-cineole (sample from France); CT6 – α-thujone > 1,8-cineole > camphor > β-thujone (sample from Ukraine); CT7 – camphor > α-thujone > 1,8-cineole > camphene (sample from Austria); CT8 – 1,8-cineole > camphor > camphene > α-pinene (sample from Greece). Previously according to the content of dominant components, S. officinalis EOs can be divided into different chemotypes. Tucker and Maciarello described five groups of sage chemotypes based on four principal constituents: (1) camphor > α-thujone > 1,8-cineole > β-thujone; (2) camphor > α-thujone > β-thujone > 1,8-cineole; (3) β-thujone > camphor > 1,8-cineole > α-thujone; (4) 1,8-cineole > camphor > α-thujone > β-thujone; and (5) α-thujone > camphor > β-thujone > 1,8-cineole [32].

Jug-Dujaković et al. [39] divided sage leaves by chemotypes, based on the content of 8 main components (α-thujone, β-thujone, camphene, borneol and bornyl acetate, camphor, 1,8-cineole, β-pinene). The authors concluded that the first major component separates populations high in thujone from populations rich in camphor, while the second component separates populations rich in α-thujone from populations rich in β-thujone. They distinguish three chemotypes of S. officinalis populations: (A) α-thujone > camphor > 1,8-cineole > β-thujone; (B) β-thujone > α-thujone > camphor ≈ 1,8-cineole; and (C) camphor > α-thujone > 1,8-cineole > camphene ≈ borneol. The results of our research show that none of the studied EO samples can be attributed to the β-thujone chemotype.

Craft et al. [20] used the content of 26 EO components for cluster analysis and established the presence of 5 main sage chemotypes based on the content of two dominant compounds. They believe the most typical is the α-thujone > camphor > 1,8-cineole chemotype of sage. Of the samples studied by us, CT4 (samples from Estonia, Georgia, Hungary and England) correspond to this type.

In European countries, EO raw materials, in particular S. officinalis, are cultivated for the needs of industry (pharmaceutical, food, etc.,) and are usually supplied by specialized farms for the cultivation of medicinal herbs. This raw material is grown according to strictly regulated conditions (GAСP) [56–59]. The collection period and cultivation conditions are regulated, so the seeds are the main and key factor that affect the quality of the raw material. It is usually standardized, thus the information about their chemotypes is especially important. The farmer is responsible for the quality of raw materials, for compliance with the regulatory document, but they, of course, do not analyse the EO composition, which was done in our work. Depending on the size of the country, the number of such farms may vary, but the issue of seed supply is not so varied. Therefore, taking this into account, the obtained data are of practical importance and will allow to make a targeted choice regarding chemotypes with low content of thujone and high concentration of other target terpenes.

The EO yields in the studied commercial sage leaves from nine countries corresponded to the minimum standard of European Pharmacopoeia for the cut drug. S. officinalis EOs were rich in thujones, camphor, 1,8-cineole, viridiflorol, α-humulene, camphene, and α-pinene. Toxic thujones are found in almost all analyzed samples. Based on these results eight chemotypes of S. officinalis were established. Considering the three components, the samples from Estonia, Georgia, Hungary, and England correspond to the most typical chemotype of 1,8-cineole, camphor, and α-thujone. The obtained results create prospects for purposeful choice of the chemotypes with low concentrations of toxic thujone and high content of other target terpenes.