The objects were the leaves of Crocus speciosus M. Bieb harvested at Vytautas Magnus University Botanical Garden in Kaunas, Lithuania in December 2019. Sample were identified and authenticated by Dr. Mykhailenko. Analysis was performed with air-dried raw materials. Voucher specimens (N 2019057) were deposited in the Herbarium of Pharmacognosy Department of the National University of Pharmacy, Kharkiv, Ukraine. HPLC grade methanol and acetonitrile (Sigma-Aldrich GmbH, Switzerland), HPLC grade glacial acetic acid (Fluka Chemie, Switzerland) were used in the analysis work. HPLC grade water was obtained from a water purifying system (Millipore, Bedford, MA, USA). 2, 20-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS, 98%) were purchased from Sigma (Sigma-Aldrich, Steinheim, Germany and St. Louis, MO, USA, respectively). Trolox (98%) was received from Fluka Chemika (Buchs, Switzerland). The reference compounds: chlorogenic acid (95.33%), mangiferin (98.0%), isoorientin (98%), hyperoside (98%), kaempferol (98%), and isoquercetin (94.16%) were purchased from Sigma-Aldrich GmbH (Steinheim, Germany). Other chemicals and solvents were of analytical grade.

Crushed sample was filled with 70% methanol and kept in an ultrasonic bath for 20 min (1:50). The sample was then diluted to the mark in 10 mL flasks and diluted with methanol and filtered through microfilters. Preliminary qualitative analysis of phenolic compounds of C. speciosus leaves was established as described by Morsy [24], by Wilson [25], by TLC analysis [26]. On the chromatograms the zones of compounds were detected by the characteristic fluorescence in UV-light at the wavelength 365 and 254 nm before and after treating the TLC plate with ammonium vapors, 2% alcoholic solution of aluminum chloride; 10% sodium/potassium hydroxide solution; 5% alcoholic solution of diazotized sulfanilic acid (diazo reagent) (it is for testing of coumarins, hydroxycinnamic acids) that allows to obtain bands with brighter fluorescence in UV-light.

Total phenolic content in the methanol extracts of C. speciosus leaves was measured using the Folin–Ciocalteu reagent method as described by Cicco et al. [27]. The absorbance of solutions was measured at 750 nm. The measurements were repeated three times. The standard acid solutions were prepared in the same way as the test solutions, but instead of 1 mL of extract, 1 mL of a solution of known concentration was taken. Using a 70% methanolic solution of gallic acid, five concentrations of gallic acid solutions were prepared: 0.5, 1.0, 1.5, 2.0, and 2.5 mg/mL. The obtained data were evaluated according to the linear regression equation of the end acid calibration graph: y = 0.9068x + 0.0617; R² = 0.9960; y = absorption intensity; x = total phenolic compounds expressed as gallic acid equivalent [mg GAE/g].

Total flavonoid content of C. speciosus leaves was determined using spectrophotometric method as described by Bezruk et al. [28]. C. speciosus extract (0.2 mL) was mixed with 0.2 mL glacial acetic acid 33%, 0.8 mL hexamethylenetetramine 5%, 0.6 mL aluminum (III) chloride 10%, and 2.2 mL of water. The resulting mixture was incubated for 30 min at 70°C. Furthermore, absorbance was measured at 407 nm [29]. The measurements were repeated three times. Reference solutions of five different concentrations were prepared: 0.2, 0.4, 0.6, 0.8 and 1.0 mg/mL. Data were evaluated according to the linear regression equation of a rutin calibration graph: y = 0.9465x – 0.0950; R² = 0.9960; y = absorption intensity; x = total amount of flavonoids expressed as rutin equivalent [mg RE/g].

The separation of phenolic compounds in C. speciosus leaves methanol extracts was carried out using an ACE C18 column (250 mm × 4.6 mm, 5.0 μm) as described by Mykhailenko et al. [30]. The flow rate of elution was 1 mL/min. The solvent system comprised solvent A (0.1% acetic acid in water) and solvent B (acetonitrile). A linear gradient program was applied: 0–8 min, 5–15% B; 8–30 min, 15–20% B; 30–48 min, 20–40% B; 48–58 min, 40–50% B; 58–65 min, 50%; 65–66 min, 50–95% B. The temperature of the column was constant at 25°C. The injection volume of the sample solution was adjusted at 20 μL. An aliquot of 20 μL of 70% methanol extract was injected three times into the HPLC system for analysis. The reference compounds (mangiferin, chlorogenic acid, isoorientin, kaempferol, hyperoside, isoquercitin) were used to prepare the standard solutions at a concentration of 1.0 mg/mL in methanol and were used for calibration. The retention time (R_t), and the UV-spectra of the peaks in the samples were compared with those of the authentic reference compounds.

The compound concentration in the plant was calculated (mg/g) by the following formula:

(1) X(mgg)=S×mst×VSst×m×Vst

where S–phenolic compound peaks average area calculated from the parallel chromatograms of the sample solution; S_st–reference compound peaks average area calculated from the parallel chromatograms of the standard solution; m–powdered raw materials weights in g; m_st–reference compound weights in mg; V–volumetric flask volume of the test extract in mL and V_st–volumetric flask volume of the reference compounds in mL.

Following the United States Pharmacopeia (USP) recommendations, there are various analytical method validation parameters, including the limit of quantification (LOQ), the limit of detection (LOD), linearity, accuracy, and repeatability [31]. The responses’ linearity range of the standards was obtained using ten concentration levels with two injections for each level. The analytes were dissolved in methanol and the stock solutions were prepared. The stock solutions were diluted to a series of appropriate concentrations to construct the calibration curves. All calibration curves were recorded using the solutions of the reference compounds with an injection volume of 2.2 µL. The working solution with the lowest concentration was diluted with methanol to various concentrations. These solutions were then used for the determination of the limits of detection (LOD) and limits of quantification (LOQ) at a signal-to-noise ratio (S/N) of 3 and 10 for each compound. The repeatability was evaluated by analyzing six replicates of each preparation using HPLC (repeatability on the real sample). The main peak areas of two repeated chromatograms were used to calculate the relative standard deviation (RSD).

Separation of the samples’ components was carried out with the ACQUITY H-class UPLC system (Waters, Milford, MA, USA) equipped with ACQUITY UPLC BEH C18 (50 × 2.1 mm, particle size 1.7 µm) (Merck Millipore, Darmstadt, Germany). Gradient elution was performed with 0.1% formic acid water solution (solvent A) and acetonitrile (solvent B), the flow rate at 0.5 mL/min. The following proportions of the solvent system were applied using a linear gradient profile B: Initial 0–3 min, 5%, 3–10 min, 30%, 50%, 10–18 min, 95%, 18–34 min, 5%. Xevo TQD triple quadrupole mass spectrometer detector (Waters) was used to obtain MS/MS data. Positive electrospray ionization was applied with the following settings: Capillary voltage was 1.5 kV, source temperature was 150°C, desolvation temperature was 350°C, with a desolvation gas flow 650 L/h, cone gas flow was 25 L/h. Collision energy and cone voltage were optimized for each compound separately. Collision energy varied in the range from 6 to 20 eV and cone voltage was selected from 8 to 38 V.

Measurement of radical scavenging properties of C. speciosus leaves methanol extracts was carried out according to the method described by Bezruk et al. [32] and Marksa et al. [33]. After PDA detection, the ABTS solution was mixed with the mobile phase carrying the analytes in the reaction coil [32]. Empower Software Chromatographic Manager System (Waters Corporation, Milford, USA) was used to analyze the data. The ABTS post-column chromatograms were detected at the wavelength of 650 nm using Waters 2487 UV/VIS detector (Waters Corporation). Teflon reaction coil of 3 m length, 0.25 mm i.d. was used. The system with the ABTS solution was monitored as follows: temperature set at 50°C and the flow rate of the reagent was set 0.5 ml/min. The standard antioxidant Trolox (0.3995 µmol/g) was used for the preparation of the calibration curves [33]. The value was calculated as μmol Trolox equivalent (TE) for 1 g of dry mass (DM) of the plant material using the following formula: TEAC = c × V/m (μmol/g) where ‘c’ is the Trolox concentration in μM established from the calibration curve, ‘V’ is the plant material extract volume in L, and ‘m’ is the weight (precise) in g [33].

For the spectrophotometry analysis, analysis was conducted as follows: 10 µl of the ethanolic extract of C. speciosus leaves was mixed with 3 ml of working ABTS^•+ stock solution (concentration was 2 mmol/L). The mixture was kept in the dark for 30 min, then its absorption was measured with a spectrophotometer at a wavelength of 734 nm. The calibration graph was created using standard Trolox solutions of 8000 to 24000 µmol/L: y = 0.00003x − 0.00360; R² = 0.9714; y = extent of absorption; x = antioxidant activity. Stock solution preparation: 0.0548 g of ABTS powder was dissolved in 50 ml of purified water. To the resulting solution was added 0.0095 g of K₂S₂O₈. ABTS working solution^•+: Stock solution was diluted with purified water until absorbance at 734 nm measured with a spectrophotometer was 0.8 ± 0.03 [34].

The plant material is weighed with electronic scales from Sartorius AG (Götingen, Germany). The plant material was shredded using a household shredder BOSCH MKM6003 (Gerlingen, Germany). For ultrasonic extraction, an ultrasonic bath Biosonic UC100 Ultrasonic Cleaner Set WUC-A06H (Witeg, Wertheim) was used. 1 cm diameter cells were used for sample analysis (Germany). The Halo DB-20 UV-Vis spectrophotometer (Dynamica GmbH, Switzerland) was used for spectrophotometric studies to determine the total amount of flavonoids, phenolic compounds and antioxidant activity.

The content of the components and antioxidant capacity were expressed as a mean ± SD (standard deviation) of three replicates. The data were processed by using SPSS v25 (IBM, Armonk, NY, USA) software. A statistically significant difference was found when p < 0.05.

The phenolic composition of C. speciosus was studied in the methanol extract of leaves by two-dimensional thin layer chromatography in mobile phase n-butanol–acetic acid–water (4:1:2)–I direction and 15% acetic acid–II direction. The chromatograms were examined on presence of phenolic compounds under daylight and UV light before and after derivatization with chromogenic agents. Substances that had a dark and yellow color under UV light and changed to an intense yellow, yellow-green, yellow-brown or orange under the action of ammonia vapor, 2% zirconium chloride solution and 5% ethanolic sodium hydroxide were classified as flavonoids. Zones with a yellow-brown or dark brown color are classified as flavones (3-and 5-hydroxyflavones) and 5-hydroxyflavanones. That is also confirmed by the boric acid colour identification reaction (yellow colouring) [25]. The reaction with 3% solution of iron (III) chloride gives a green coloration of the solution. That indicates the presence of flavonols or flavones in Crocus leaves. Zones with a blue color on the chromatogram in UV light was belong to derivatives of hydroxycinnamic acids. Chromatograms were treated with ammonia vapor and aluminum chloride reagent. The zones of flavonoid aglycones acquired bright yellow fluorescence. The dark brown zones acquired a yellow-green color, which is characteristic of flavone glycosides. In addition, chromatograms in UV light had spots with a dark yellow fluorescence characteristic for benzo-γ-pyrones. After derivatization with ammonia vapor, these zones turned a bright orange color. After treatment with a chromogenic reagent (AlCl₃), these zones became a green, which confirms their belonging to xanthones. According to literature data on the chemical composition of Crocus genus plants, several classes of biologically active compounds were selected for further identification in plant raw materials by HPLC methods.

The preliminary results of phytochemical screening of C. speciosus leaves extracts showed the presence of aminoacids, proteins, saponins, hydrolysable tannins, triterpenoids. Alkaloids and coumarins were not found in raw material (Tab. 1).

The results are difficult to compare with the literature data, because, the information in the scientific sources prevails for C. sativus [5,6,35,36], as well as, the chemical composition are influenced by the type of plant genotype, extraction procedures, geographical and climatic conditions of growth, and also phenological phases.

Further analysis was carried out for the leaves of C. speciosus, since the literature data on the study of this particular Crocus species are very limited. According to the results of study conducted in 2001 year [35] the total phenolic content in extracts of C. sativus leaves from different parts of the world averaged 8.55%–9.49%. Leaves of C. sativus from Jammu and Kashmir (India) contained 5.62 mg GAE/g of phenolic compounds [37]. In current investigation, the total phenolic content in extracts of C. speciosus leaves were 0.412 ± 0.01 mg GAE/g or 4.12 ± 0.21 µmol/g. The total flavonoid content in C. speciosus leaves was 1.07 ± 0.02 mg RE/g or 10.62 ± 0.42 μmol/g. The total content of phenolic compounds depends not only on the place of production, but also on the various growing conditions, the time of harvesting or the preparation time of the plant raw material [30,32]. For this experiment, the leaves of C. speciosus were harvested in December in VDU botanical garden of Kaunas, Lithuania. The experimental place has a humid continental climate (Dfb in the Köppen climate classification). Average temperatures on the coast are −2.5°C (27.5°F) in January and 16°C (60.8°F) in July. The geographic coordinates of the area are 54°5350 N latitude, 23°5310 EW longitude and 48 m above sea level [38]. The climate of Lithuania is temperate and cool, but the Gulf Stream creates a favorable humid climate for autumn flowering Crocuses.

The spectrophotometric method based on the oxidation of phenolic compounds in an alkaline medium with the Folin–Ciocalteu reagent as a rule used to determine the total phenolic content in plant raw materials [39]. Recent studies [40,41] have shown that the total content of phenols, determined by the Folin–Ciocalteu method, correlates with antioxidant activity of plant extracts. The method of measuring the optical density of solutions at 410–430 nm after adding a solution of aluminum chloride is used to assess the total content of flavonoids [32,42]. However, this reagent oxidizes different groups of phenolic compounds.

The absorption in the area of 405–410 nm interacting with the reagent of aluminum chloride gives not only flavonols (for example, kaempferol, isorhamnetin, rutin), but also flavones (apigenin, luteolin, etc.) and flavonones (naringenin, hesperetin, etc.) with the hydroxyl groups in position 3 and/or 5 [42,43]. Flavonones and flavones with aluminum chloride and form stable complexes. Analysis of literature data [4,11,36,37] showed that plants of Crocus genus may contain naringenin and apigenin and their glycosides. Therefore, the total flavonoids content in terms of rutin in C. speciosus leaves is higher than the total amount of phenolic compounds. The content of total phenols is less using the Folin-Ciocalteu reagent than the total flavonoids content with aluminum chloride. This is probably due to the presence of compounds in Crocus leaves that do not react with the Folin-Chocalteu reagent in the specified UV region. The leaves contain approximately equal amounts of flavonols and phenolic acids (including the identified chlorogenic acid), which are not oxidized by the Folin–Ciocalteu reagent. The assessment of the compounds content can be given precisely for the flavonoids content. This study represents the first report on the determination of total phenolic and flavonoids contents in C. speciosus leaves. Moreover, Crocus leaves are generally regarded as by-products and could be used as an alternative source of bioactive compounds.

A validation study was conducted to demonstrate the applicability of the developed analytical method. The validation was done in terms of specificity, linearity, LOD, LOQ, precision and recovery according to the International Conference on Harmonization [31]. The results are summarized in Tabs. 2 and 3. The regression equation for each reference standard compound, together with the LOD and LOQ values are shown in Tab. 2. All the calibration curves showed acceptable linear regression (r² ≥ 0.999). The overall intra-day and inter-day precision RSDs were not more than 1.3%. The overall stability over 24 h and repeatability were not more than 1.06% for both parameters. The developed analytical method showed excellent precision with overall recovery in the range from 99.58% to 101.91% (RSD ≤ 1.31%) for all compounds. Therefore, the method was precise, accurate and sensitive enough for the simultaneous quantitative evaluation of all compounds in C. speciosus methanol extracts. The specificity is the ability of a method to discriminate the study analytes and other constituents in the sample. It was demonstrated by the separation of the analytes from other interfering compounds. The determination of the main compounds in the tested solutions was done by comparing the retention times of the peaks and UV-spectrum with those of the standard solution. The results showed that the conditions for the fingerprint analysis were repeatable and precise.

Previously, Harborne et al. [11] identified kaempferol in C. speciosus leaves by paper chromatography. Its derivative kaempferol 3-sophoroside was further isolated by the authors from the flowers of C. laevigatus and C. korolkowii. Nørbaek et al. [13,14] found astragalin, kaempferol 3-O-β-D-(2-O-β-D-glucopyranosyl) glucopyranoside, kaempferol 3,4′-di-O-β-D-glucopyranoside, kaempferol 3-O-α-L-(2-O-β-D-glucopyranosyl) rhamnopyranosides, quercetin 3,4′-di-O-β-D-glucopyranoside, quercetin 3-O-β-D-sophoroside, and isoramnetin in petals of C. speciosus. According to our knowledge, no further studies were carried out for the leaves of this Crocus species.

For further investigation the HPLC method was applied. It is the optimal method, which stands to provide a simple and versatile approach to identify and quantitative assessment of secondary metabolites in plant extracts [44]. The identification of compounds occurs by comparing the detected UV spectra and the retention time of the compounds on the chromatogram with the spectra of reference compounds. HPLC analysis of C. speciosus leaves methanol extract showed the presence of six phenolic compounds, namely mangiferin, chlorogenic acid, isoorientin, kaempferol, hyperoside, and isoquercitin (Figs. 1a and 1b). Isoorientin and kaempferol had higher content 3.68 and 1.25 mg/g, respectively. All identified compounds in C. speciosus leaves play an important role as antioxidants. Thus, their presence in the raw material suggests antiradical activity of its extracts [36].

To confirm the identified compounds, further UPLC-MS/MS analysis of these components was performed in negative ion mode, and their retention time (t_R), calculated molecular weights and MS/MS data are shown in Tab. 4, respectively. This method has high sensitivity and therefore allows the identification of substances that are contained in raw materials in a small amount and can elucidate unknown structures based on mass fragmentation pathways of known compounds [45]. All compounds were unambiguously identified with the reference standards comparison. Peak 1 ([M – H]⁻ at m/z 353) was assigned to monocaffeoyquinic acids. The MS spectrum of peak 1 was characterized by the loss of one moiety of caffeic acid, due to the ester bond, which readily dissociated and was identified as 3-O-caffeylquinic acid or chlorogenic acid in comparison with the reference standards. In the MS spectrum peak 2 produced the [M − H]⁻ ion at m/z 421. The MS² spectrum gave the predominant fragment ions at m/z 331 and 301 owing to the neutral loss of C₃H₆O₃ and C₄H₈O₄, respectively. The low intensity product ion at m/z 259 appeared in the MS² spectrum due to the loss of the glucose residue from the [M − H]⁻ ion. The MS³ spectrum of the precursor ion at m/z 301 gave the product ion at m/z 273, 271 and 257 resulting from the loss of CO, CH₂O and CO₂, respectively. The RDA fragmentation reaction was also observed in the MS/MS spectrum and corresponding with literature data [46]. For flavones, Peaks 3 ([M − H]⁻ at m/z 447), 4 ([M − H]⁻ at m/z 463), 5 ([M − H]⁻ at m/z 463) and 6 ([M − H]⁻ at m/z 285) were identified as isoorientin (calculated for C₂₂H₂₀O₁₁, m/z 248), hyperoside (calculated for C₂₁H₂₀O₁₂, m/z 464), isoquercitrin (calculated for C₂₁H₂₀O₁₂, m/z 464) and kaempferol (calculated for C₁₅H₁₀O₆, m/z 286), respectively. They were further confirmed by comparing the retention time and the MS/MS spectra with the corresponding standards.

For determination of antiradical activity in vitro we used the ABTS radical-cation binding method. Pre-study antiradical activity of C. speciosus leaves extract with spectroscopic method showed promising results (150.08 ± 4.5 µmol Trolox/g). Therefore, further analysis was performed with an application of HPLC-ABTS. The antioxidant activity of the identified compounds (TEAC values, Trolox µmol/g) is displayed in Tab. 5. The antioxidant activity of C. speciosus leaves has a potent antioxidant capacity (TEAC 1049.930 µmol/g). The extract possessed antioxidant activity due to the present phenolic compounds (chlorogenic acid, mangiferin, isoorientin, hyperoside, isoquercitin, kaempferol). This could be explained by the fact that phenolic compounds are potent antioxidants due to their high redox potential allowing them to become hydrogen donors and singlet oxygen quenchers. The greatest antioxidant activity was detected in samples according to the presence of mangiferin, isoorientin, and kaempferol (Fig. 2). Other peaks of the compounds are also present on the chromatogram, but they did not reveal antioxidant activity, and precise identification of these compounds was not carried out.

Flavonoids have antioxidant properties that can prevent the oxidative degradation of cell wall lipids and increase the body’s antioxidant capacity [47]. According to the literature data, C. sativus methanol 80% leaves extract has a higher antioxidant activity than C. sativus methanol petal extract. Also, methanolic 80% leaf extract was similar to α-tocopherol in its antioxidant capacity [48,49]. Baba et al. [37] compared the antioxidant activity of extracts of stigma, corms, and leaves of C. sativus. The extract of C. sativus stigma demonstrated the strongest antioxidant activity which could be attributed to its highest content of phenolics and flavonoids. The water leaf extract of C. sativus showed more moderate antioxidant activity (IC₅₀ 482.78 ± 4.81 μg/mL). The antioxidant activity of this extract was lower than ascorbic acid had (IC₅₀ 10.50 ± 0.42 μg/mL) as a positive control substance. In current experiment, the potent antioxidant activity of the leaves extract of C. speciosus is likely to be associated with total phenolic and flavonoid contents.