**1. Introduction**

Essential oils comprise a mixture of secondary metabolites, which are biosynthesized by aromatic plants as natural protectants [1]. The role of essential oils is not restricted to protection as they also offer many therapeutic benefits to humans that can exceed the benefits provided by the dried herbs on their own [2]. Recently, they have become well known as a part of traditional medicine for the treatment of a plethora of human ailments, in aromatherapy, as well as in spices with high nutritive value [3]. In addition, many essential oils as well as plant extracts have shown significant antioxidant potential [4–6]. New sources of medicinal agents that are e ffective and safe as well as selective has recently become the main target in drug discovery. Medicinal plants in general, and their volatile constituents in particular, act as a very important sources for the production of a huge number of biologically active agents, which are attractive chemical leads that are promising therapeutic agents for the alleviation of many ailments [7,8]. Many biological activities have been ascribed to the volatile constituents obtained from a variety of plants such as antinociceptive, anticancer, antiphlogistic, antiviral, antioxidant, antimicrobial, antimycotic, antiparasitic and insecticidal activities [9]. Moreover, the volatile constituents of plants are highly popular in the food, cosmetic and pharmaceutical industries because of their broad acceptance by consumers, relative safety, and their potential multipurpose effect [10,11].

The Apiaceae family is well-known for its rich aromatic plants, which are categorized under approximately 112 genera and nearly 316 species. Anise, chervil, celery, coriander, cumin, caraway, dill, fennel, ferula and galabanum are significant members of this family and they are characterized by their notable odor owing to the presence of considerable amounts of essential oils or the oleoresin predominant in their di fferent organs [3]. These plants are widely used for culinary purposes either for their aroma or as nutrients [12].

*Ferula* constitutes the third largest genus in the Apiaceae family with nearly 180 species. The members of this genus are very popular for their essential oils, which are recognized as having many biological activities including antibacterial, antifungal, antiviral, antispasmodic, anticonvulsant, and antioxidant activity as well as having high nutritive value [13,14].

This study aimed to investigate the contents of the essential oil from six *Ferula* species growing in Uzbekistan, namely, *F. caratavica* (*Fc*), *F. kuchistanica* (*Fk*), *F. pseudoreoselinum* (*Fp*), *F. samarcandica* (*Fs*), *F. tenuisecta* (*Ft*) and *F. varia* (*Fv*) using GC analyses. Discrimination of these species was carried by coupling the data obtained from GC-analyses with chemometrics employing unsupervised pattern recognition techniques represented by principal component analysis (PCA). Furthermore, the antioxidant potential of the di fferent essential oil samples using di fferent assays, namely, 2,2-azino-bis (3-ethylbenzothiazoline-6-sulphonic acid) (ABTS), cupric reducing antioxidant capacity (CUPRAC), ferric reducing power (FRAP), and the phosphomolybdenum (PM) assay were evaluated *in vitro*. In addition, an evaluation of the possible enzymatic inhibitory activities of essential oils against tyrosinase and α-amylase was done using standard in vitro bioassays.

#### **2. Results and Discussion**

#### *2.1. Qualitative and Semi-quantitative Determinations by GC-MS and GC-FID*

The di fferences in the composition of the essential oils obtained from the aerial parts of *Fc*, *Fk*, *Fp*, *Fs*, *Ft* and *Fv* were detected both qualitatively and semi-quantitatively using GC-MS and GC-FID analyses, respectively. All of the essential oils are yellow in color and possess a characteristic odor. Characterization of the essential oils using GC analyses revealed the presence of 106 metabolites (Table 1, Figures 1 and 2) that account for 92.10, 96.43, 87.43, 95.95, 92.90 and 89.48% of *Fc*, *Fk*, *Fp*, *Fs*, *Ft* and *Fv* whole essential oils, respectively. Twenty-nine compounds were detected in *Fc* with α-pinene (21.17%), 10,13 docosadienoic acid methyl ester (15.20%), β-caryophyllene oxide (13.23%) and caryophyllene (10.88%) representing the predominant compounds. Meanwhile, thirty-nine compounds were identified in *Fk* essential oil with α-pinene (36.79%) and verbenol (8.49%) being the major compounds. In *Fp*, forty-five compounds were characterized with 4-terpineol (16.28%), α-pinene (10.99%), β-myrcene (6.04%), β-caryophyllene oxide (5.69%), p-cymen-8-ol (5.36%) and spathulenol (5.34%) as the main metabolites in the oil. Furthermore, 15 compounds were determined in *Fs* oil with the main compounds, palmitic acid, β-myrecene, heptacosane, octacosane, hexacosane and pentcosane accounting for 39.09, 10.75, 10.27, 9.60, 8.99 and 6.29%, respectively. For *Ft*, 62 compounds were detected of which α-pinene (42.0%), camphene (8.34%) and α-cadinol (8.14%) exist in high percentages in the oil. Finally, 25 compounds were identified in the *Fv* oil with 10,13 docosadienoic acid methyl ester (69.61%) constituting the major component (Figure 3). From the data shown in Table 1, it was concluded that monoterpenes are the predominate class of essential oil metabolites in *Fc*, *Fk* and *Ft*, where they represents 24.90, 42.91 and 61.95%, respectively, while oxygenated monoterpenes are the dominant class of metabolites in *Fp* (35.60%), and they also exist in a high percentage in *Fk* (34.82%). On the contrary, fatty acids are highly predominate in *Fs* and *Fv* and account for 82.55 and 79.84%, respectively.


**Table 1.** Composition of volatile oil in the aerial parts of *F. caratavica* (*Fc*), *F. kuchistanica* (*Fk*), *F. pseudoreoselinum* (*Fp*), *F. samarcandica* (*Fs*), *F. tenuisecta* (*Ft*) and *F. varia* (*Fv*).

**Table 1.** *Cont.*


Compounds were identified based on a comparison of the compounds' mass spectral data and retention indices with those of the NIST Mass Spectral Library (December 2011), the Wiley Registry of Mass Spectral Data, 8th edition and by comparison with the authentic standard (AU). The content (%) was calculated using the normalization method based on the GC-FID data generated from the average of three independent chromatographic runs.

**Figure 1.** GC-MS chromatograms of *F. caratavica* (**A**), *F. kuchistanica* (**B**) and *F. pseudoreoselinum* (**C**).

**Figure 2.** GC-MS chromatograms of *F. samarcandica* (**A**), *F. tenuisecta* (**B**) and *F. varia* (**C**).

**Figure 3.** Main secondary metabolites in the *Ferula* species.

#### *2.2. Chemometric Analysis*

It is extremely difficult to identify the qualitative and quantitative differences between the *Ferula* species under evaluation with the naked eye. So, the data obtained from GC analyses were subjected to unsupervised pattern recognition chemometric analysis utilizing PCA to improve the visualization of these differences. The results of the PCA, as represented by the obtained score plot shown in Figure 4A effectively discriminated the six *Ferula* species into five clusters along the first component (PC1) and the second component (PC2) that account for 57% and 30%, respectively, or 87% of the total variance. From the obtained results, it is obvious that both *Fk* and *Ft* are very closely related to each other as they are gathered together in one cluster in the lower left quadrant. However, PC1 successfully discriminated between *Fk* and *Ft* with negative values of PC1 as they are located in the lower left quadrant and *Fc* and *Fv*, which show positive values of PC1 are located in the lower right quadrant. Meanwhile, PC2 significantly discriminated between *Fk* and *Ft*, which show negative values of PC2 as they are located in the lower left quadrant and between *Fs* and *Fp*, which show positive values of PC2 and are located in the upper left quadrant. Furthermore, both PC1 and PC2 significantly discriminated between *Fc* and *Fv*, which show positive values for PC1 and negative values for PC2 as they are located in the lower right quadrant and between *Fs* and *Fp*, displaying negative values for PC1 and positive values for PC2 as they are located in the upper left quadrant. The major discriminatory signals are α-pinene, 10,13-docosadienoic acid methyl ester and palmitic acid as revealed in the loading plot shown in Figure 4B.

The Pearson correlation coefficient (r) between the essential oil contents of different studied samples indicated that *Fc* had a highly significant positive correlation with *Ft* (r = 0.71), *Fk* (r = 0.58), *Fv* (r = 0.47) and *Fp* (r = 0.35), while a non-significant negative correlation was observed between *Fc* and *Fs* (the highest correlations were observed between *Ft* and *Fk* (r = 0.89, *p* < 0.001), between *Fc* and *Ft* (r = 0.71, *p* < 0.001), and between *Fc* and *Fk* (r = 0.58, *p* < 0.001) as seen in Table 2. These data indicate that three samples, *Ft*, *Fk*, and Fc have highly similar essential oil content.

**Figure 4.** Score plot (**A**) and loading plot (**B**) of GC data obtained from *F. caratavica*, *F. kuchistanica*, *F. pseudoreoselinum*, *F. samarcandica*, *F. tenuisecta* and *F. varia* essential oil analyses using principal component analysis (PCA). In the loading plot, compounds are given numbers as in Table 1 where the major discriminatory signals are α-pinene **(4),** palmitic acid **(95)** and 10,13-docosadienoic acid methyl ester **(102)**.

**Table 2.** The Pearson correlation matrix of the essential oils content of different samples.


The data is represented as the r value of the correlation coefficient and \*\*\* is the level of significance, *p* < 0.001.

## *2.3. Biological Evaluation*

#### 2.3.1. Antioxidant Potential of Different *Ferula* Species

The antioxidant potential of the different essential oil samples was performed in vitro using the 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), the cupric ion reducing antioxidant capacity (CUPRAC), The ferric reducing antioxidant power (FRAP) and the phosphomolybdenum method (PM) assays. The results displayed in Table 3 reveal that most of the samples showed

considerable antioxidant potential in the performed assays. *Fc* (41.36 mgTE/g oil) exhibited the most antioxidant activity in ABTS assays, followed by *Fk* (29.12 mgTE/g oil) and *Ft* (28.03 mgTE/g oil). However, in CUPRAC assay, *Fp* (289.45 mgTE/g oil) showed the most superior antioxidant potential followed by *Ft* (278.87 mgTE/g oil) and *Fk* (120.43 mgTE/g oil). Furthermore, *Ft* exhibited the most significant antioxidant power in both FRAP and PM assays with antioxidant activity equivalent to 136.81 mgTE/g oil and 78.66 mmolTE/g oil, respectively, followed by *Fp*, which showed antioxidant potential of 121.64 mgTE/g oil and 50.86 mmolTE/g oil in FRAP and PM assays, respectively. Thus, it can be concluded that the essential oil from both *Ft* and *Fp* exhibited the most notable antioxidant properties as evidenced by their pronounced activities in most of the performed antioxidant assays, followed by *Fc*. α-Pinene, the predominant compound in *Ft* and *Fp* has previously been shown to possess notable antioxidant activity [15]. Additionally, the significant antioxidant activity found in this study, which can be interpreted as a result of the synergistic action between the di fferent components that exist in the oils, was in accordance with that previously reported for many other *Ferula* species such as *F. microcolea*, *F. orantalis* and *F. communis.* Various mechanisms can be used to interpret antioxidant potential including the prohibition of chain initiation, peroxide decomposition, obstruction of continual hydrogen removal as well as the scavenging of free radical and uniting transition metal ion catalysts [3,16,17]. Additionally, α-pinene, the main constituent in both *Ft* and *Fp*, has previously been shown to be a potent antioxidant in both DPPH and FRAP assays, displaying EC50 values equal to 310 and 238 μg/mL, respectively [18].

**Table 3.** Antioxidant activities of the essential oil samples of *Ferula* species using the 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), the cupric ion reducing antioxidant capacity (CUPRAC), The ferric reducing antioxidant power (FRAP) and the phosphomolybdenum method (PM) assays.


Values are reported as mean ± S.D of three parallel measurements. TE: Trolox equivalents. Different superscripts (a–f) indicate significant differences in the tested *Ferula* species (*p* < 0.05).

#### 2.3.2. Tyrosinase and α-Amylase Inhibitory Potential

Tyrosinase enzyme is an oxidase enzyme containing copper that assists in the completion of the first two steps of mammalian melanogenesis, which leads to undesirable hyperpigmentation. Thus, the search for e ffective tyrosinase inhibitors has recently become vital so that they can be incorporated in cosmetics for e ffective skin whitening and to counteract hyperpigmentation [19]. *Fk* showed the most e ffective tyrosinase inhibitory potential, which was estimated as 119.67 mgKAE/g oil followed by *Fv*, which showed an inhibitory potential equivalent to 118.42 mgKAE/g oil, where KAE is a Kojic acid equivalent, a potent tyrosinase inhibitory drug. *Fv* oil is rich in 10,13 docosadienoic acid methyl ester, a polyunsaturated fatty acid, which greatly accounts for its promise as a tyrosinase inhibitor [20]. The underlying tyrosinase inhibitory mechanism mainly relies on the essential oils being rich in components that possess a hydrophobic portion that competitively inhibits the active sites of tyrosinase enzyme with subsequent interference of melanin synthesis. This inhibition may be achieved via interaction with Cu+<sup>2</sup> that exists in the active sites of tyrosinase in addition to the prohibition of tautomerization to dopachrome triggered by the oil, which behaves as a reducing agen<sup>t</sup> and blocks of the oxidation reaction during the formation of melanin intermediates during the conversion of tyrosinase/DOPA into melanin, thus reducing skin pigmentation [21].

The α-amylase enzyme is critical in assisting in the catalysis of the first steps in the conversion of starch into maltose, and subsequently to glucose [22,23]. Nowadays, α-amylase inhibitors are used in therapeutic approaches to counteract hyperglycemia. *Fp* and *Fv* exhibited the most potent α-amylase inhibitory potential as evidenced by their pronounced inhibitory activity, which was equivalent to 2.61 and 1.40 mmol ACAE/g oil, respectively, in which ACAE is the acarbose equivalent, a potent α-amylase inhibitor (Figure 5). 4-Terpineol as well as α-pinene, which predominate the essential oil of *Fp*, were previously reported to possess considerable α-amylase inhibitory activity [24]. Similarly, the potent α-amylase inhibitory potential is mainly due to the synergistic action between the different components, which is in accordance to different previously reported studies that confirmed the α-amylase inhibitory effect of different terpenes and different *Ferula* species such as *F. gummosa* essential oil [24,25].

**Figure 5.** In vitro tyrosinase inhibition (**A**) and α-amylase inhibition (**B**) of the essential oil of different *Ferula* species, *F. caratavica* (*Fc*), *F. kuchistanica* (*Fk*), *F. pseudoreoselinum* (*Fp*), *F. samarcandica* (*Fs*), *F. tenuisecta* (*Ft*) and *F. varia* (*Fv*). Different letters (a–f) indicate significant differences in the tested *Ferula* species (*p* < 0.05).

#### **3. Materials and Methods**

#### *3.1. Plant Material*

Aerial parts (flowers, leaves and stems) of *F. caratavica* Regel & Schmalh (N2004), *F. pseudoreoselinum* (Regel & Schmalh.) Koso-Pol., p.p. (N1489), *F. tenuisecta* Korovin (N1488) were collected from the Tashkent region of Uzbekistan. *F. varia* (Schrenk ex Fisch., C.A.Mey. & Avé-Lall.) Trautv. (N1407) was collected from the Bukhara region (Uzbekistan), while *F. kuchistanica* Korovin (N1425) and *F. samarcandica* Korovin (N1919) were collected from the Samarkand region of Uzbekistan. The plants were collected during the flowering stage in June–July 2018. Their taxonomic authentication was accomplished by Dr. A. Nigmatullaev at the Institute of the Chemistry of Plant Substances (Tashkent, Uzbekistan).

#### *3.2. Preparation of Essential Oil Samples*

All the plant materials were air-dried in the shade for 7 days at room temperature and powdered using a mortar and pestle to ge<sup>t</sup> particles of a uniform, reduced size. Preparation of the essential oil samples was achieved by hydrodistillation of the air-dried aerial parts of the different *Ferula* species, *F. caratavica* (*Fc)*, *F. kuchistanica (Fk)*, *F. pseudoreoselinum* (*Fp)*, *F. samarcandica* (*Fs)*, *F.* tenuisecta *(Ft)* and *F. varia* (*Fv)* for 2 h by Clevenger-type apparatus. Anhydrous Na2SO4 was used to dehydrate the prepared essential oils, yielding 0.4, 0.7, 0.3, 0.3, 0.8 and 0.5 % v/w of dry weight for *Fc*, *Fk*, *Fp*, *Fs*, *Ft* and *Fv*, respectively. Then the various oil samples were maintained at −30 ◦C in dark-colored stoppered glasses until their analyses [26,27].

#### *3.3. GC-FID and GC-MS Analyses*

A Shimadzu GC-17A gas chromatograph (Shimadzu Corporation, Kyoto, Japan) with an FID detector and DB-5 fused-bonded cap column (Phenomenex; 29 m × 0.25 mm i.d., film thickness 0.25 μm; Torrance, California, USA) was utilized for the semi-quantitative determination of the different components of the essential oils using the normalization method to ge<sup>t</sup> the relative percentage of each component and applying GC-FID data that is highly sensitive using GC solution® software ver. 2.4 (Shimadzu Corporation, Kyoto, Japan). The areas under the peaks (AUP) were determined using three independent runs where the total area is considered as 100%. Meanwhile, the Shimadzu GC-2010 plus gas chromatograph (Shimadzu Corporation, Kyoto, Japan) supplied with Rtx-5MS (Restek, Bellefonte, PA, USA) in addition to a quadrupole mass spectrometer was used for the identification of the essential oil different metabolites. Instrument settings were adjusted according to what was previously reported [28,29]. The Wiley Registry of Mass Spectral Data 8th edition, NIST MassSpectral Library (December 2011), and previously reported data were employed to confirm the identity of the compounds and the retention indexes were calculated to corroborate the identification of the volatile compounds [30,31].

#### *3.4. Chemometric and ANOVA Analysis*

To examine the differences between the essential oils' components prepared from different *Ferula* species, the data collected from the different GC-MS spectra were subjected to chemometric analysis of unsupervised pattern recognition represented by PCA, which was processed by employing Unscrambler 9.7 (CAMO SA, Oslo, Norway) [28,32]. Meanwhile, other statistical analyses used for biological assessment were performed using ANOVA assay (with Tukey's test, significant value: *p* < 0.05) and Xlstat 2017 software.

#### *3.5. Biological Evaluation*

#### 3.5.1. Determination of the Antioxidant Potential

The antioxidant activity of the different essential oil samples from different *Ferula* species was evaluated using ABTS, CUPRAC, FRAP and PM assays. These assays were performed following the methods described by Mamadalieva et al. [33]. The antioxidant activities were reported as Trolox equivalents and the samples were analyzed in triplicate.

#### 3.5.2. Determination of Enzyme Inhibitory Effects

The possible inhibitory potential of the essential oil samples was investigated against tyrosinase and α-amylase enzymes using standard in vitro bioassays as previously reported by Mamadalieva et al. [33] in which all the samples were analyzed in triplicate. Results are expressed in mgKAE/g oil for tyrosinase inhibitory activity and in mmol ACAE/g oil for α-amylase inhibition.
