Chemometric Analysis Evidencing the Variability in the Composition of Essential Oils in 10 Salvia Species from Different Taxonomic Sections or Phylogenetic Clades

Essential oil (EO) of Salvia spp. has been widely used for culinary purposes and in perfumery and cosmetics, as well as having beneficial effects on human health. The present study aimed to investigate the quantitative and qualitative variations in EOs in wild-growing and cultivated pairs of samples from members in four Salvia sections or three clades, namely S. argentea L. (Sect. Aethiopis; Clade I-C), S. ringens Sm. (Sect. Eusphace; Clade I-D), S. verticillata L. (Sect. Hemisphace; Clade I-B), S. amplexicaulis Lam., and S. pratensis L. (Sect. Plethiosphace; Clade I-C). Furthermore, the natural variability in EO composition due to different genotypes adapted in different geographical and environmental conditions was examined by employing members of three Salvia sections or two phylogenetic clades, namely S. sclarea L. (six samples; Sect. Aethiopis or Clade I-C), S. ringens (three samples; Sect. Eusphace or Clade I-D), and S. amplexicaulis (five samples; Sect. Plethiosphace or Clade I-C). We also investigated the EO composition of four wild-growing species of two Salvia sections, i.e., S. aethiopis L., S. candidissima Vahl, and S. teddii of Sect. Aethiopis, as well as the cultivated material of S. virgata Jacq. (Sect. Plethiosphace), all belonging to Clade I-C. The EO composition of the Greek endemic S. teddii is presented herein only for the first time. Taken together, the findings of previous studies are summarized and critically discussed with the obtained results. Chemometric analysis (PCA, HCA, and clustered heat map) was used to identify the sample relationships based on their chemical classes, resulting in the classification of two distinct groups. These can be further explored in assistance of classical or modern taxonomic Salvia studies.


Introduction
The genus Salvia L. is the largest genus of the Lamiaceae family [1], with approximately 1000 accepted species to date [2], and they are distributed in the temperate, subtropical, and tropical regions from both the Old and New World, extending from Central and South America (above 500 species) to Central Asia and the Mediterranean Region (above 250 species) and East Asia (above 90 species) [3].Taxonomically, the genus Salvia has continuously raised conflicting opinions regarding its classification [1].The genus Salvia has been traditionally divided into four morphologically circumscribed subgenera (Subgen.Salvia, Sclarea, Calosphace, and Leonia), which include 12 sections, comprising members of Sect.
Hymenosphace, Eusphace, and Drymosphace, as well as members of Sect.Horminum, Aethiposis, and Plethiosphace, which occur in the Old World, while members of Sect.Calosphace occur in the New World, and members of Sect.Echinosphace, Pycnosphace, Heterosphace, Notiosphace, and Hemisphace occur in both of them [4].Phylogenetically, the genus Salvia has also been disputed as being monophyletic or polyphyletic [3] or paraphyletic [5].These modern studies have re-circumscribed Salvia to consist of three separate clades (Salvia clades I-III), each with different sister groups [3], or they have embedded five other small genera (Dorystaechas, Meriandra, Perovskia, Rosmarinus, and Zhumeria) in a broadly defined Salvia circumscription [5].In Greece, 26 Salvia taxa (species and subspecies) are found in different geographical areas [6], some of which are single-country endemics such as S. eichleriana Heldr.ex Halácsy and S. teddii Turrill.
With a genus name alluding to strong medicinal power and effective therapeutic properties (Salvia originated from the Latin word "salvare" meaning "to heal") [9] for human health, several Salvia spp.have been used worldwide in traditional medicine since ancient times to treat various diseases like digestive disorders, inflammations/infections of the mouth and throat, respiratory ailments, skin disorders, tuberculosis, infections, etc. [9][10][11][12].In addition, their leaves and EOs have been widely applied for culinary purposes since ancient times (e.g., as spices and flavor additives), in perfumery, and in cosmetics [9].Over the years, many Salvia taxa (species and subspecies) have been studied for their chemical diversity, unveiling a rich reservoir of diversified and specialized products, including volatile compounds (essential oil; monoterpenoids, sesquiterpenes, etc.) and nonvolatile compounds (e.g., terpenoids, flavonoids, and phenolic acids) [12,13].Salvia EOs have attracted increasing interest for their potential beneficial effects on human health, demonstrating to date various pharmacological activities, including strong antimicrobial and anti-inflammatory potential among others [12,14].
According to different pharmacopeias (United States, British, and European), the quality control of EOs from medicinally important plants is essential and involves the evaluation of any qualitative and quantitative modification of their constituents, particularly the principal and active ones [15].It is noteworthy to mention that high variability in the chemical composition of Salvia EOs has been observed, which is usually attributed to various factors such as genotypic variation or different geographical origin, varied environmental conditions, diversified harvesting time, and different types of plant material (e.g., fresh or dried/cultivated or collected from wild populations, different plant parts, etc.), thus resulting in the high chemical polymorphism of the EOs of the studied materials and subjects (e.g., [16][17][18]).
Context-wise, the scope of this study was three-fold; firstly, we aimed to detect variability patterns in EO composition between wild-growing and cultivated pairs of samples from members in four Salvia sections or three clades such as S. argentea L. (Sect.Aethiopis; Clade I-C), S. ringens Sm. (Sect.Eusphace; Clade I-D), S. verticillata L. (Sect.Hemisphace; Clade I-B), S. amplexicaulis Lam., and S. pratensis L. (Sect.Plethiosphace; Clade I-C).Secondly, we examined various samples from wild-growing populations of different species to document natural variability in EO composition due to different geographical and environmental conditions by employing members of three Salvia sections or two clades, namely S. sclarea L. (six samples; Sect.Aethiopis or Clade I-C), S. ringens (three samples; Sect.Eusphace or Clade I-D) and S. amplexicaulis (five samples; Sect.Plethiosphace or Clade I-C).Furthermore, we examined the EO composition of four wild-growing species of two Salvia sections, i.e., S. aethiopis L., S. candidissima Vahl, and S. teddii of Sect.Aethiopis, as well as the cultivated material of S. virgata Jacq. of Sect.Plethiosphace, all belonging to Clade I-C.All results of these investigations were chemometrically evaluated to identify the sample relationships based on their chemical classes.Taken together, the findings of previous studies are summarized and critically discussed with the obtained results.

Chemical Analysis of EOs
A total number of 26 Salvia population samples (20 wild-growing and 6 cultivated) were studied, out of which 24 were collected from Greece and 2 from North Macedonia (Supplementary Materials Table S1).The plants were collected at different wild habitats at diverse altitudes spanning from 82 to 1532 m (Supplementary Materials Table S1).The EO yields of the studied materials (Supplementary Materials Table S1) ranged from 0.50 to 1.50% (v/w, based on the dry weight of the plant material), with the sample of S. candidissima presenting the highest amount (1.50%).In total, 205 compounds were identified in the present analysis (Tables 1-4).A summary of the main chemical compounds (>5.0%) and groups of the investigated Salvia EOs reported in the literature is presented in Supplementary Materials Table S2.In total, 74 publications were retrieved from 1976 to 2023.
Previous studies have investigated the EO of S. argentea from Greece [31], Italy [32], Morocco [32,33], Serbia [34], North Macedonia [35], Tunisia [36,37], and Turkey [38] (Supplementary Materials Table S2).It is worth noting that many differences were observed across the results of these studies.More specifically, oxygenated sesquiterpenes were the predominant chemical class in most of the Mediterranean samples from Italy [32], North Macedonia [32,35], Serbia [32,34], and Tunisia [36], while the samples from Morocco [32,33] and Turkey [38] were characterized by oxygenated monoterpenes (64.1%) and monoterpene hydrocarbons (24.26%), respectively.Regarding the main components of EOs, caryophyllene oxide (37.6%) was identified as the main compound in the samples from North Macedonia, followed by α-copaene (8.5%), humulene epoxide II (6.3%), and β-caryophyllene (6.1%) [32,35], whereas in the sample from Serbia, viridiflorol (32.4%), manool (14.6%), α-humulene (10.7%), and cis-thujone (7.3%) were identified [34].The EOs of the two Tunisian populations [36] were quite similar to the Serbian one.Although the composition of the EO of S. argentea collected in Sicily was rich in oxygenated sesquiterpenes like those from North Macedonia, Serbia, and Tunisia, quite different constituents were found.More precisely, 14-hydroxy-α-humulene (40.1%), 1,3,8-p-menthatriene (12.1%), globulol (7.4%), and β-sesquiphellandrene (5.8%) were the main components of the Sicilian sample, which were not present in the other EOs.The effect of different growth stages on the composition of the S. argentea EO has been also studied in samples collected from Tunisia, reporting observed differences in the chemical classes and constituents among the three developmental stages [37].In the EO from Turkey, sclareol (40.01%), germacrene D (13.90%), β-pinene (11.93%), sclareol oxide (9.65%), and α-pinene (6.59%) were the main compounds [38].Moreover, germacrene D (37.41%) and β-caryophyllene (6.75%) were reported as major constituents in the EO from Greece [31].Although germacrene D was also the principal component in our samples, variations were found in the composition of the total EOs.Overall, our findings presented differences from the results of previous studies regarding the main chemical classes and constituents of EOs, which could be attributed to environmental conditions, harvesting time, the type/origin of the plant material and plant parts examined.It is noteworthy to mention that this study is the first report on S. argentea EOs collected from Greece, investigating two different population samples (cultivated and wild-growing).
Overall, 37 volatile constituents were identified in the EO of S. teddii, representing 94.4% of the total oil (Table 1).The EO was characterized by a high content of (E)caryophyllene (59.6%), followed by caryophyllene oxide (11.2%), germacrene D (6.2%), and (E)-β-farnesene (5.4%).The sesquiterpene fraction was the main group of the compounds (93.9%), of which sesquiterpene hydrocarbons were the prevailing group (80.2%) compared to the oxygenated sesquiterpenes (13.7%).This study reports the composition of S. teddii EO for the first time; thus, no data are available for any comparison.
The results of this study are generally in agreement with those of previous studies examining the EO of S. ringens from Greece [65,66] (Supplementary Materials Table S2).However, there are important variations in terms of the relative percentages of the components.More specifically, 1,8-cineole, α-pinene, bornyl acetate, and β-pinene were the main compounds of the EO in the first study [65]; it is noteworthy to mention that bornyl acetate was found in fewer amounts in all our analyzed samples, apart from src1.In the second study, α-pinene, β-pinene, 1,8-cineole, camphene, and borneol were detected as the dominant components [66].Moreover, the major component of S. ringens EO from Bulgaria was camphor [67] (Supplementary Materials Table S2).In our findings, camphor was only identified in low amounts (1.8-4.5%).In the EO of North Macedonia, 1,8-cineole, camphene, borneol, and α-pinene were the main compounds [68] (Supplementary Materials Table S2).However, our results of src1, which was ex situ-cultivated in Northern Greece after the propagation of cuttings from wild-growing populations in North Macedonia, revealed different dominant components such as bornyl acetate (20.3%), 1,8-cineole (17.9%), camphene (15.6%), borneol (12.9%), α-pinene (9.1%%), and β-pinene (4.8%).It is well established that such differences in the chemical composition might be the result of several factors, including the plant part or developmental stage examined, origin, and the harvesting period [68].
According to our literature survey on previous studies on S. verticillata EOs (Supplementary Materials Table S2), their composition is reported to be characterized by a high level of complexity.More specifically, a previous study [39] investigated the EO of S. verticillata from Greece and reported monoterpenes as the major fraction of the oil with the main constituents of β-pinene (30.7%), p-cymene (23.0%), isopropyl ester of lauric acid (16.8%), α-pinene (7.6%), and (E)-nerolidol (5.2%).It is noteworthy to mention that the isopropyl ester of lauric acid was not found in our samples.The S. verticillata EO from Iran has been extensively studied [18,[69][70][71][72][73][74].These EOs mainly consisted of sesquiterpenes, such as β-caryophyllene, germacrene D, bicyclogermacrene, and α-humulene [74,75].In addition, β-caryophyllene (13.3%) and γ-muurolene (10.3%) were the principal compounds in the EO from former Yugoslavia [30].S. verticillata EO from three different populations in Serbia exhibited some differences among the samples [24,76].Moreover, germacrene D, bicyclogermacrene, and β-caryophyllene were the dominant constituents of the EO of this species from Italy [77].In the EO from Poland, α-pinene (10.72%), camphor (5.23%), and limonene (5.85%) were the main compounds [57].Smekalova et al. [78] investigated the chemical compositions of S. verticillata EO collected from seven locations in the Czech Republic.Consequently, a high chemical polymorphism could be noticed in the S. verticillata EO due to the geographical origin of the studied materials, the processed plant parts examined, and the applied techniques [75,77].Further studies are necessary to better understand the variability in the chemical profile of S. verticillata.
Considering the S. amplexicaulis EO from Poland [57] and Serbia [24,79,80] (Supplementary Materials Table S2), sesquiterpenes were found to be the dominant chemical class in the EOs reported in two of the previous studies [79,80], with detected monoterpene hydrocarbons as the main group in another one [24].Regarding the major constituents, variations were observed among the reported studies, which could be attributed to the different collection areas of the plant materials and the applied techniques.This study reports the composition of S. amplexicaulis EO from different parts of Greece for the first time.
The chemical constituents of the two EOs of S. pratensis subsp.pratensis are presented in Table 4.However, it should be mentioned that one of these samples refers to ex situ-cultivated individuals originally collected from wild-growing populations in North Macedonia (sprc1), which were identified as S. pratensis subsp.bertolonii (Vis) Soó but are included in Table 4 and Supplementary Materials Tables S1 and S2 as S. pratensis subsp.pratensis to adopt the nomenclature of the Plants Of the World Online [81], while another one (spr2) refers to the typical wild-growing population identified as S. pratensis subsp.pratensis.Similar total percentages of 97.4% (sprc1) and 97.9% (spr2) represented by 25 and 19 compounds, respectively, were found in the EOs of these samples.In both samples, sesquiterpene hydrocarbons were the major chemical class, with percentages of 48.8% and 53.6%, respectively (Table 4), thus showing a similar profile in terms of dominant compounds.However, quantitative variations were observed regarding the rest of the main chemical classes.For instance, the second main group of sprc1 was oxygenated sesquiterpenes (29.0%), while in spr2, it was monoterpene hydrocarbons (27.2%).In contrast, similarities were also detectable; in both EOs (sprc1 and spr2), the major compounds were (E)-caryophyllene (45.5%; 46.9%), caryophyllene oxide (25.8%; 10.4%), and sabinene (14.8%; 15.0%).
Anackov et al. [82] have investigated the EOs of S. pratensis subsp.pratensis from Serbia (Supplementary Materials Table S2).Caryophyllene oxide and (Z)-caryophyllene were the dominant constituents in one sample attributed to the EO of S. bertolonii Vis., which is considered a synonym of S. pratensis subsp.pratensis, with oxygenated sesquiterpenes being the most abundant class.However, the EO from the typical S. pratensis subsp.pratensis from a different Serbian region had (E)-caryophyllene (26.4%), (Z)-β-farnesene (6.0%), β-cubebene (5.6%), and epi-bicyclosesquiphellandrene (5.6%) as the major components in the same study, with sesquiterpene hydrocarbons being in higher amounts.In addition, camphene and thujol were found as the main constituents in the S. pratensis EO from Poland [57].Such differences and similarities clearly indicate the need for new multidisciplinary (taxonomic, genetic, chemical, and ecological) investigations that should be carried out in an attempt to shed light on the circumscription of S. pratensis subsp.bertolonii at the subspecies or species level.

Chemometric Analysis
The hierarchical cluster analysis (HCA) (Figure 1) showed the formation of two major groups for the studied Salvia samples.The first group (Group I) comprised 14 samples clustering the studied members of Sect.Aethiopis (saeth, sargc1, sarg2, scad, and sted) together with members of Sect.Plethiosphace (samp1, samp2, samp3, samp4, samp5, sprc1, spr2, and svirg), all belonging to the phylogenetic Clade I-C; however, in the same major group, the cultivated sample of S. verticillata (sverc1) of Sect.Hemisphace was also clustered.The second group (Group II) comprised 12 samples clustering S. sclarea samples of Sect.Aethiopis or Clade I-C in a subgroup (sscl1, sscl2, sscl3, sscl4, sscl5, and sscl6), with S. ringens samples of Sect.Eusphace or clade I-D (src1, sr2, sr3, and sr4) in another subgroup; however, the wild-growing sample of S. verticillata of Sect.Hemisphace or Clade I-B (sver2) and the sample of cultivated S. amplexicaulis of Sect.Plethiosphace or Clade I-C (sampc6) were not tightly clustered with the rest.
Applying additional multivariate analyses in the heatmap analysis combined with HCA with the chemical classes, the color pattern corresponding to different samples of Salvia members varied with color intensity and increased gradually, from lowest to the highest grade (blue indicates low correlations, while red color indicates high correlations).The clustered heatmap (Figure 4) confirmed the abovementioned clustering results for HCA and PCA.  5 and Supplementary Materials Table S1).Mean and 95% confidence intervals are given.Chemical classes marked with asterisks (*) differed statistically significantly in the t-test (p < 0.05).
Applying additional multivariate analyses in the heatmap analysis combined with HCA with the chemical classes, the color pattern corresponding to different samples of Salvia members varied with color intensity and increased gradually, from lowest to the highest grade (blue indicates low correlations, while red color indicates high correlations).The clustered heatmap (Figure 4) confirmed the abovementioned clustering results for HCA and PCA.   5 and Supplementary Materials Table S1).Mean and 95% confidence intervals are given.Chemical classes marked with asterisks (*) differed statistically significantly in the t-test (p < 0.05).5 and Supplementary Materials Table S1).Mean and 95% confidence intervals are given.Chemical classes marked with asterisks (*) differed statistically significantly in the t-test (p < 0.05).
Applying additional multivariate analyses in the heatmap analysis combined with HCA with the chemical classes, the color pattern corresponding to different samples of Salvia members varied with color intensity and increased gradually, from lowest to the highest grade (blue indicates low correlations, while red color indicates high correlations).The clustered heatmap (Figure 4) confirmed the abovementioned clustering results for HCA and PCA.For the arrangement of Salvia species in taxonomic sections and phylogenetic clades, see Table 5.

Plant Materials
The flowering aerial parts of 10 Salvia species were collected from May to July 2021.Specifically, 24 samples originated from Greece and 2 from the Republic of North Macedonia (Figure 5).Among them, 20 samples were obtained from wild-growing populations (from 16 different areas of Greece), while 6 were sourced from ex situ-cultivated samples that originated from wild-growing populations (4 from Greece and 2 from North Macedonia).A detailed catalog of the samples, collection sites, and dates is provided in Supplementary Materials Table S1.Voucher specimens were identified by Dr. Nikos Krigas and are deposited in the Herbarium of the Balkan Botanic Garden of Kroussia (BBGK), Institute of Plant Breeding and Genetic Resources, Agricultural Organization Demeter, together with living plant specimens maintained ex situ (Supplementary Materials Table S1).The sections and clades of the investigated Salvia species are included in Table 5. donia).A detailed catalog of the samples, collection sites, and dates is provided in Supplementary Materials Table S1.Voucher specimens were identified by Dr. Nikos Krigas and are deposited in the Herbarium of the Balkan Botanic Garden of Kroussia (BBGK), Institute of Plant Breeding and Genetic Resources, Agricultural Organization Demeter, together with living plant specimens maintained ex situ (Supplementary Materials Table S1).The sections and clades of the investigated Salvia species are included in Table 5.    S1).

EO Isolation
All collected plant materials were air-dried at room temperature for 10 days and then comminuted.About 20 g from each plant was used, and the EOs were obtained by hydro-distillation in a modified Clevenger apparatus for 3 h, according to the Hellenic Pharmacopoeia [94].GC (gas chromatography) grade n-pentane was used for the collection of the EOs, with the addition of anhydrous sodium sulfate to reduce any moisture.The EOs were subsequently analyzed by GC-MS (gas chromatography-mass spectrometry) and finally stored at −20 • C.

GC-MS Analysis
GC-MS analyses were carried out using a Hewlett Packard 7820A-5977B MSD system operating in EI mode (70 eV), equipped with an HP-5MS fused silica capillary column (30 m × 0.25 mm; film thickness 0.25 µm), and a split-splitless injector.The temperature program was 60 • C at the time of the injection, and then it was raised to 300 • C at a rate of 3 • C/min and subsequently held at 300 • C for 10 min.Helium was used as a carrier gas at a flow rate of 2.0 mL/min.The injected volume of the samples was 1 µL.Each analysis was repeated three times.
Retention index (RI) values were calculated using a linear equation by Van den Dool and Kratz [95] based on a homologous series of n-alkanes from C9 to C24.The identification of the chemical components was based on a comparison of RI values and mass spectra fragmentation patterns with those reported in the NIST/NBS and Wiley libraries, as well as those described by Adams [96] and other literature data.

Chemometric Analysis
We applied hierarchical cluster analysis (HCA), using the centroid clustering method, with interval measuring the squared Euclidean distance and applying standardization to variables with range -1 to 1 to evaluate the variability of samples.We also applied principal components analysis (PCA) to investigate the amount of variability explained.The number of principal components included in the analysis was based on the percent of variability explained (criteria set to over 80%).The data consisted of 11 chemical classes: monoterpene hydrocarbons (MH), oxygenated monoterpenes (OM), sesquiterpene hydrocarbons (SH), oxygenated sesquiterpenes (OS), diterpenes (D), oxygenated diterpenes (OD), hydrocarbonsalcohols (HAlc), hydrocarbons-aldehydes (Hald), hydrocarbons-ketones (HK), hydrocarbonsalkanes (HAlk), and others (OT).Data on the total percentages of these chemical classes were evaluated.The total number of samples was n = 26.All analyses were performed using SPSS software (SPSS v.28 IBM Corporation, 2023 Armonk, New York, USA).

Literature Review
An extensive survey concerning the EO and volatile compounds of each examined Salvia species was conducted separately in widely used scientific databases such as PubMed, Scopus, and Google Scholar using keywords related with EOs and volatile constituents (i.e., Salvia aethiopis essential oil, Salvia aethiopis chemical composition, Salvia aethiopis chemical profile, Salvia aethiopis volatile constituents, Salvia amplexicaulis essential oil, Salvia amplexicaulis chemical composition, Salvia amplexicaulis chemical profile, Salvia amplexicaulis volatile constituents, Salvia argentea essential oil, Salvia argentea chemical composition, Salvia argentea chemical profile, Salvia argentea volatile constituents, Salvia candidissima essential oil, Salvia candidissima chemical composition, Salvia candidissima chemical profile, Salvia candidissima volatile constituents, Salvia pratensis essential oil, Salvia pratensis chemical composition, Salvia pratensis chemical profile, Salvia pratensis volatile constituents, Salvia ringens essential oil, Salvia ringens chemical composition, Salvia ringens chemical profile, Salvia ringens volatile constituents, Salvia sclarea essential oil, Salvia sclarea chemical composition, Salvia sclarea chemical profile, Salvia sclarea volatile constituents, Salvia teddii essential oil, Salvia teddii chemical composition, Salvia teddii chemical profile, Salvia teddii volatile constituents, Salvia verticillata essential oil, Salvia verticillata chemical composition, Salvia verticillata chemical profile, Salvia verticillata volatile constituents, Salvia virgata essential oil, Salvia virgata chemical composition, Salvia virgata chemical profile, and Salvia virgata volatile constituents).In total, 74 publications were retrieved from 1976 to 2023, while articles focused on non-volatile constituents were excluded.

Conclusions
In this study, we provided new information regarding the quantitative and qualitative variation in EOs in wild-growing and cultivated pairs of samples from members in four different Salvia sections or three different clades, thus evidencing the profile of the cultivated samples compared to the ones sourced from the wild.On one hand, such differences and similarities may indicate the natural potential of the examined Salvia species in quantitative and qualitative terms, and this can pave the way for the sustainable exploitation of the cultivated materials; on the other hand, these findings clearly indicate the need for new multidisciplinary (taxonomic, genetic, chemical, and ecological) investigations employing Salvia members.This is especially true for any attempt to shed light on synonymizing S. pratensis subsp.bertolonii with subsp.pratensis or, in contrast, to provide solid evidence regarding the re-circumscription of S. pratensis at the subspecies or species level.Furthermore, in this investigation, we provided documentation regarding the natural variability in EO composition due to different genotypes adapted in different geographical and environmental conditions by employing members of three different Salvia sections or two different phylogenetic clades.
By reviewing published research and through a comprehensive re-assessment of the EO composition of the Salvia species herein studied in terms of taxonomic sections and phylogenetic clades, important variability patterns were discussed, providing evidence regarding the effects of genotypes adapted in different geographical origins and environmental conditions, revealing diversification patterns of EO composition between cultivated and wild-growing samples, and documenting EO variability in the different plant parts examined.
Novelty-wise, this investigation outlined two distinct groups (Group I and Group II) based on the chemical classes among 10 Salvia members of different taxonomic sections or phylogenetic clades, which were quantitatively and qualitatively discerned; these can be further explored with the help of classical or modern taxonomic studies on Salvia.Moreover, this study furnished new data regarding the chemotaxonomy of Salvia members of Sect.Aethiopis and Sect.Plethiosphace of Clade I-C; it represents the first report from Greece on the EO composition of S. aethiopis (Sect.Aethiopis) and S. argentea originated from two different population samples (cultivated and wild-growing, Sect.Aethiopis), as well as the first Greek report for S. amplexicaulis (from different Greek regions) and S. virgata (Sect.Plethiosphace), all phylogenetically belonging to Clade I-C.Rich sclareol content is consolidated for Greek S. sclarea.Additionally, this study exclusively reports the EO composition of S. teddii for the first time (Sect.Aethiopis; Clade I-C).
Regarding the benefits of the investigated Salvia EOs on human health, the relevant studies reviewed in this paper for the specific taxa were found to be limited.Indeed, EOs of some species have not been pharmacologically studied so far.In contrast, S. sclarea EO has been explored for its antimicrobial activity in vitro.However, additional experimental studies should be carried out in order to determine the possible applications of these EOs in human health.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules29071547/s1,Table S1.List of the investigated Salvia species in different sections with provenance, abbreviations used (essential oil code, EO), voucher specimens, and/or living material in ex situ conservation at the Balkan Botanic Garden of Kroussia (BBGK) with IPEN (International Plant Exchange Network) code and yield; Table S2.Overview of essential oil compositions of 10 Salvia species in different sections based on literature sources; Figure S1.Principal component analysis of the major chemical classes by Salvia groups as defined in HCA.
tered.The second group (Group II) comprised 12 samples clustering S. sclarea samples of Sect.Aethiopis or Clade I-C in a subgroup (sscl1, sscl2, sscl3, sscl4, sscl5, and sscl6), with S. ringens samples of Sect.Eusphace or clade I-D (src1, sr2, sr3, and sr4) in another subgroup; however, the wild-growing sample of S. verticillata of Sect.Hemisphace or Clade I-B (sver2) and the sample of cultivated S. amplexicaulis of Sect.Plethiosphace or Clade I-C (sampc6) were not tightly clustered with the rest.

Figure 1 .
Figure 1.Hierarchical cluster analysis of 26 samples of Salvia members of three phylogenetic clades or four taxonomic sections.For the coding of samples, see Tables 1, 2, 3, and 4. For the arrangement of Salvia species in taxonomic sections and phylogenetic clades, see Table5.
Figure 1.Hierarchical cluster analysis of 26 samples of Salvia members of three phylogenetic clades or four taxonomic sections.For the coding of samples, see Tables 1, 2, 3, and 4. For the arrangement of Salvia species in taxonomic sections and phylogenetic clades, see Table5.

Figure 1 .
Figure 1.Hierarchical cluster analysis of 26 samples of Salvia members of three phylogenetic clades or four taxonomic sections.For the coding of samples, see Tables1-4.For the arrangement of Salvia species in taxonomic sections and phylogenetic clades, see Table5.
Molecules 2024, 29, x FOR PEER REVIEW 19 of 28 OD, HK, and OT, while negative loadings were observed for MH, OM, SH, HAlc, HAld, and HAlk.More details on the correlations can be found in the Supplementary Materials FigureS1.

Figure 2 .
Figure 2. Contributions of chemical classes to principal components.

Figure 2 .
Figure 2. Contributions of chemical classes to principal components.

Molecules 2024 , 28 Figure 3 .
Figure 3.Chemical classes of two groups concerning the 26 samples of the 10 studied members of genus Salvia from different taxonomic sections or phylogenetic clades (Table5and Supplementary Materials TableS1).Mean and 95% confidence intervals are given.Chemical classes marked with asterisks (*) differed statistically significantly in the t-test (p < 0.05).

Figure 4 .
Figure 4. Clustered heat map of Salvia samples in chemical classes.For the coding of samples, see Tables 1, 2, 3, and 4. For the arrangement of Salvia species in taxonomic sections and phylogenetic

Figure 3 .
Figure 3.Chemical classes of two groups concerning the 26 samples of the 10 studied members of genus Salvia from different taxonomic sections or phylogenetic clades (Table5and Supplementary Materials TableS1).Mean and 95% confidence intervals are given.Chemical classes marked with asterisks (*) differed statistically significantly in the t-test (p < 0.05).

Molecules 2024 , 28 Figure 3 .
Figure 3.Chemical classes of two groups concerning the 26 samples of the 10 studied members of genus Salvia from different taxonomic sections or phylogenetic clades (Table5and Supplementary Materials TableS1).Mean and 95% confidence intervals are given.Chemical classes marked with asterisks (*) differed statistically significantly in the t-test (p < 0.05).

Figure 4 .
Figure 4. Clustered heat map of Salvia samples in chemical classes.For the coding of samples, see Tables 1, 2, 3, and 4. For the arrangement of Salvia species in taxonomic sections and phylogenetic clades, see Table5.

Figure 4 .
Figure 4. Clustered heat map of Salvia samples in chemical classes.For the coding of samples, see Tables 1-4.For the arrangement of Salvia species in taxonomic sections and phylogenetic clades, see Table 5.

1
Section based on the study of Bentham[93].2Section based on the study of Will & Glassen-Bockhoff[1].

Figure 5 .Figure 5 .
Figure 5. Botanical collections in Greece and the Republic of Northern Macedonia regarding the herein investigated Salvia species (n = 10) projected on Google Earth.Samples and living plant individuals collected from wild-growing populations are indicated with green symbols, while purple symbols indicate the location of ex situ cultivation sites for the plant individuals collected from the Figure 5. Botanical collections in Greece and the Republic of Northern Macedonia regarding the herein investigated Salvia species (n = 10) projected on Google Earth.Samples and living plant individuals collected from wild-growing populations are indicated with green symbols, while purple symbols indicate the location of ex situ cultivation sites for the plant individuals collected from the wild; these are close to Thessaloniki for Salvia pratensis, S. amplexicaulis, and S. ringens or close to Athens for S. argentea, S. verticillata, and S. virgata (see Supplementary Materials TableS1).

Table 5 .
Botanical names, sections, and clades of the 10 investigated Salvia species.

Table 5 .
Botanical names, sections, and clades of the 10 investigated Salvia species.