Screening of antioxidant activity and volatile compounds composition of Chamerion angustifolium (L.) Holub ecotypes grown in Lithuania

Since biological activity of medicinal plants is dependent on cultivation area, climatic conditions, developmental stage, genetic modifications and other factors, it is important to study flora present in different growing sites and geographical zones. This study was focused on screening of antioxidant activity of C. angustifolium harvested in six different locations in Lithuania. The total contents of phenolic compounds, flavonoids and 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging activity were evaluated by spectrophotometric methods. A correlation between radical scavenging activity and total phenolic compounds content was observed (correlation coefficient 0.98). HPLC with online post-column DPPH radical scavenging reaction detection was used for the separation of extracts. Oenothein B, rutin and one unidentified compound were predominant. Volatile compounds were analysed using solid-phase microextraction coupled with gas chromatography–mass spectrometry. Based on the analysis of volatiles, all samples were classified into two chemotypes: (I) with predominant α- and β-caryophyllenes and (II) with predominant anethole.

Our study was focused on the screening of antioxidant activity of C. angustifolium (L.) collected in different locations in Lithuania. To our knowledge, this is the first study of the content of phenolic compounds, flavonoids and their DPPH free radical scavenging activity, also volatile compounds composition of different C. angustifolium ecotypes growing in Lithuania. Furthermore, to our knowledge, there is no data published concerning the volatile compounds composition of C. angustifolium yet, and there is a very scare data on comparative chemometric taxonomy according to the volatile and non-volatile biologically active compounds of any plant published in the literature.
C. angustifolium, called willow herb, is widely used in traditional medicine, and many researchers justify the usefulness of the plant to the human health. It was found that extracts of C. angustifolium inhibit growth of human prostatic epithelial cells ; influence the expression of oestrogen receptor α and β mRNAs (Kujawski et al. 2010); possess antimicrobial activity against Staphylococcus aureus (including methicillin-resistant S. aureus), Bacillus subtilis, Escherichia coli (including p-fimbriae positive strain), Pseudomonas aeruginosa, Proteus mirabilis, Candida albicans, C. tropicalis, C. dubliniensis and Saccharomyces cerevisiae (Kosalec et al. 2013); have analgesic properties ; modulate phagocyte functions (Schepetkin et al. 2009); reduce lipid oxidation (Rey et al. 2005); exhibit antifungal activity (Webster et al. 2008); possess the immune enhancing properties (Ramstead et al. 2012); and distinguish by antioxidant activity (Kiss et al. 2011). In spite of the studies performed, it should be kept in the mind that phytochemical composition of the plant is dependent on the geographic origin (Pirbalouti et al. 2011).

Spectrophotometric evaluation of antioxidant activity
Literature data about the total amount of phenolic compounds and flavonoids contents in C. angustifolium (L.) extracts are scarce. Jürgenson et al. (2012) analysed vegetational variation of phenolic compounds in C. angustifolium and determined 27.0-85.0 mg/g of polyphenols (expressed as chlorogenic acid equivalents) and 0.1-2.4 mg/g of flavonoids; amount varied depending on the collection time and part of the plant. Kiss et al. (2011) evaluated the amount of oenothein B and flavonoids in three Epilobium species and found that C. angustifolium accumulates 225.8 mg/g of oenothein B and 13.4 mg/g of flavonoids. Rutin was used as a standard for all spectrophotometric measurements in our study, so all data are expressed in mg of rutin per 1 g of dry material. Total phenolic content in C. angustifolium samples varied from 90.5 to 144.5 mg/g, while total flavonoids content was more than five times lower and varied between 6.7 and 22.9 mg/g. Medium correlation between the total phenolic content and total flavonoid content was observed (correlation coefficient 0.65). RSA in tested samples ranged from 110.9 to 174.2 mg/g (Figure 1). RSA showed strong correlation with total phenolic content (correlation coefficient 0.98) and a bit lower with total flavonoids content (correlation coefficient 0.71). The highest contents of phenolic compounds and flavonoids contained samples collected in Užutrakis forest (2CA) and in Kaunas Botanical garden (6CA); these samples also possessed the highest RSA.
Samples 4CA and 7CA distinguished by the lowest amount of flavonoids (6.7 and 8.8 mg/g, respectively); both samples were harvested in Panara, just collected in different vegetation phase, i.e. during the intensive growth stage (4CA) and butonisation stage (7CA). There was no significant difference between the total phenolic content and RSA in these samples. In this case, vegetation phase did not have significant impact on the total content of phenolic compounds, total content of flavonoids and RSA, although studies of other plants show that phytochemical composition is greatly dependent on the plant vegetation phase (Wang & Lin 2000;Kaškonienė et al. 2013).
Spectrophotometric evaluation of C. angustifolium (L.) ecotypes showed that samples possess different antioxidant activity: different harvesting location, i.e. soil composition, precipitation, temperature, and sun shining days in that location, may have the highest impact on the total phenolic and flavonoid contents and radical scavenging activity.

Online HPLC-DPPH · evaluation of antioxidant activity
Online HPLC-DPPH radical scavenging assay allows to determine compounds responsible for the radical scavenging activity in DPPH model system, because not all the compounds may scavenge DPPH radicals (the chromatographic profile of tested C. angustifolium (L.) extracts is presented in Figure S1). Five compounds were identified in C. angustifolium extracts, i.e. oenothein B, 3,4-dihydroxybenzoic, chlorogenic and caffeic acids and rutin. Caffeic and chlorogenic acids were also found in C. angustifolium by Tóth et al. (2009). All the identified compounds showed radical scavenging activity (negative peaks in Figure S1); oenothein B distinguished by the highest activity.
The quantitative comparison of identified compounds was performed according their HPLC peak areas (Figure 2). According to the peak area of oenothein B, tested samples were in the following order: 6CA > 2CA > 5CA > 3CA > 7CA > 4CA ~ 1CA. Oenothein B composed from 36.2 to 59.5% of the total HPLC peak area. Rutin was also predominant in the samples and composed from 14.0 to 26.6% of all compounds in the extracts. HPLC-DPPH data coincide with the spectrophotometric data -the smallest UV chromatogram peaks obtained for 4CA and 7CA samples correlate with the lowest amount of flavonoids evaluated by spectrophotometric method. The correlation coefficient between the peak area, i.e. amount, of oenothein B and DPPH free radical scavenging activity, was 0.90.

Volatile compounds composition
Literature data about the volatile compounds composition of C. angustifolium are also scarce. To our knowledge, this is the first study of C. angustifolium volatile compounds composition grown in Lithuania. Volatile compounds composition of tested samples are listed in Table 1. C. angustifolium is not rich in quantity and diversity of the volatile compounds: 11 different compounds were identified in the samples. Anethole and caryophyllenes (α-and β-) were identified in all samples. Caryophyllenes (α and β) were predominant in three samples (1CA, 2CA and 5CA) and composed of 54.0 to 78.5% of the essential oils; anethole composed more than 51% of the oils in 4CA and 7CA samples, which were harvested in Panara. Sample 3CA (from Aleksotas) distinguished by relatively high amount (8.5%) of cis-arbusculone.
Caryophyllene-rich (60.9%) rhizome oil of Zingiber nimmonii showed significant inhibitory activity against the fungi, Candida glabrata, C. albicans and Aspergillus niger and the bacteria B. subtilis and P. aeruginosa; however, no activity was observed against the fungus Fusarium oxysporum (Sabulal et al. 2006). It was found that β-caryophyllene possesses antigenotoxic activity (Di Sotto et al. 2010), anaesthetic activity (Ghelardini et al. 2001) and anticancer activity (Legault & Pichette 2007). Although evaluation of biological activities of the essential oils of C. angustifolium was not under the scope of this study, according to the above-mentioned published data, the essential oils could be also therapeutically valuable.

Classification of the samples
To estimate the similarity between any pair or groups of samples, a linkage calculation procedure (hierarchical clustering) was used. The calculated distances according to the volatiles composition and antioxidant activity are plotted in the dendrograms (Figure 3). The hierarchical cluster classification approach according to the volatiles composition distinguished two big groups of the samples, each group being represented by the samples with predominant compounds, i.e. caryophyllenes or anethole (Figure 3(a)). Almost the same classification was obtained using data of antioxidant activity (Figure 3(b)), i.e. only the sample 1CA appeared in another group.
This analysis revealed that the composition of C. angustifolium herb harvested in different locations differs greatly. The study showed that ecotypes of C. angustifolium grown in Lithuania significantly differ in the total phenolic compounds and flavonoid content and in radical scavenging activity. According to the predominant volatile compounds, investigated C. angustifolium samples can be divided into two groups: with predominant anethole (3CA, 4CA and 7CA) and with predominant α-and β-caryophyllenes (1CA, 2CA, 5CA and 6CA). Almost the same classification into two groups was obtained analysing radical scavenging activity spectrophotometric data. Only a minor difference, i.e. transition of 1CA from one group to another, was observed, which confirms existence of ecotypes and correlation between classification of plants according volatile and non-volatile secondary metabolites.

Plant material
Seven samples of C. angustifolium were collected in six different locations in Lithuania from May till July in the year 2012 (Table S1). Samples were identified by prof. O. Ragažinskienė. Notes: 1CA -sample collected in Kazlų Rūda forest, 2CA -Užutrakis forest, 3CA -Aleksotas, 4CA -Panara (collected at growing phase), 5CA -Švenčionys, 6CA -Kaunas Botanical Garden, and 7CA -Panara (collected at butonisation phase), Bold values indicate two dominant compounds in the sample. Table S1) were deposited at the Herbarium of the Kaunas Botanical Garden of Vytautas Magnus University. The aerial part of C. angustifolium herb was cut and dried at room temperature in a well-ventilated shadow place. 100 mg of dried material was extracted with 4 mL 75% methanol for 24 h in TiterTek shaker (Germany). Extracts were filtered through a paper filter and 0.22 μm pore size disposable membrane filter. The extracts were stored at +4 °C and used for spectrophotometric and HPLC analysis.

Spectrophotometric analysis
Evaluation of the total content of phenolic compounds in C. angustifolium extracts was performed using the Folin-Ciocalteu method described by Kaškonienė et al. (2015). The total content of flavonoids in C. angustifolium extracts was evaluated by the aluminium trichloride colorimetric assay (Kaškonienė et al. 2015). The DPPH free radical activity of C. angustifolium extracts was evaluated using 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radical (Kaškonienė et al. 2015). All measurements were compared to a calibration curve of rutin solutions and expressed as mg of rutin equivalents per gram of herb. Rutin was selected as a standard in all spectrophotometric assays for comparative reasons. Experiments were performed in triplicates.

Online HPLC-DPPH radical scavenging assay
The assay was performed with some modifications as described by Stankevičius et al. (2011). Separations were performed on a LiChroSpher RP-18e, 5 μm 12.5 × 0.4 cm column and 0.5 × 0.4 cm pre-column (Merck, Germany), packed with C18 phase with particle size of 5 μm. Detection was carried out at 254 nm wavelength. The eluents were (A) water with 0.05% trifluoroacetic acid and (B) methanol with 0.05% trifluoroacetic acid. The gradient program was following: from 0 to 30 min B increased from 20 to 95%; from 30 to 37 min kept 95%; from 37 to 38 min decreased from 95 to 20%; from 38 to 48 min kept 20%. A linear binary gradient was used at a flow rate of 0.8 mL/min, and injection volume was 10 μL.
The 0.01 M DPPH · reagent was prepared in acetonitrile/methanol/sodium phosphate buffer (0.1 M pH 5.5) at a ratio of 25:25:50 at the beginning of each day of analysis and kept protected from light. The DPPH · reagent was degassed before use. The flow of DPPH · reagent was 0.75 mL/min. The components were identified by comparison of their retention times to those of authentic HPLC grade standards (3,4-dihydroxybenzoic acid, gallic acid, chlorogenic acid, catechin, caffeic acid, rutin, ferulic acid, trans-p-coumaric acid, apigenin, and kaempferol from different suppliers) under analysis conditions. A 10-min equilibrium time was allowed before injections.

Solid-phase microextraction with GC-MS analysis
Volatile compounds composition was determined using solid-phase microextraction (SPME) coupled with GC-MS. Extraction of C. angustifolium volatiles was performed on 65 μm PDMS/ DVB (polydimethylsiloxane/divinylbenzene) Stable Flex fibre (Supelco, Bellefonte, USA). 50 mg of samples was added into 10 mL vials and thermostated for 15 min at 50 °C. The analysis was carried out using a GC-MS system (GCMS-QP2010, Shimadzu, Tokyo, Japan). A Restec (Bellefonte, USA) RTX-5MS (30 m × 0.25 mm i.d. × 0.25 μm film thickness) GC column was used. The oven temperature gradient was started at 30 °C and raised to 200 °C at 5 °C/min, and then raised to 280 °C at 20 °C/min and was held for 2 min. Helium (99.999%, AGA Lithuania) was used as carrier gas with a constant flow rate of 1.2 mL/min. The injector temperature was kept at 230 °C in split mode (1:10). The mass detector was operated in electron impact mode (70 eV). The ion source and interface temperatures were set at 220 and 260 °C correspondingly. Identification of compounds was performed according to their mass spectra (NIST v1.7). Positive identification was assumed when good matches (90% and more) of mass spectra were achieved.

Statistical analysis
All experiments were repeated in triplicate. All values are expressed as the mean ± standard deviation. Standard deviations were calculated using spreadsheet software (Excel ® ). The correlation coefficients were calculated using MS Excel ® software (CORREL statistical function). Hierarchical cluster analysis was performed using PASW Statistics 18 software.

Conclusions
This analysis revealed that the composition of C. angustifolium herb harvested in different locations differs greatly. The study showed that ecotypes of C. angustifolium grown in Lithuania significantly differ in the total phenolic compounds and flavonoid content and in radical scavenging activity. Total phenolic compounds content ranged from 90.5 to 144.5 mg/g (expressed as rutin equivalents); flavonoid content varied between 6.7 and 22.9 mg/g (expressed as rutin equivalents). Radical scavenging activity was 110.9-174.2 mg/g (expressed as rutin equivalents) in the tested samples.
According to the predominant volatile compounds, C. angustifolium samples can be classified into two groups: with predominant anethole and predominant α-and β-caryophyllenes.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by Science Council of Lithuania [grant number MIP 084/2012 (ONKOFITAS)].