Screening of yarrow (Achillea millefolium Agg.) populations in Serbia for yield components and essential oil composition

Yarrow (Achillea millefolium Agg.) is well-known medicinal plant, with a wide spectrum of applications and it is one of the most frequently used plant drug in Serbia. In this study, we have observed cultivation and essential oil chemical properties of 28 A. millefolium populations collected from Serbian sites. In the second vegetation, the yield of useful part (upper 15 cm) ranged from 925 – 3630 kg/ha, while the yield of essential oil ranged from 0.40 – 0.82%. The most dominant compounds in monoterpene fraction were β-pinene (max. 36.3%), sabinene (max. 35.7%), 1,8-cineol (max. 26.6%) and borneol (max. 20.2%), while in the sesquiterpene fraction the most abundant compounds were trans-caryophyllene (max. 18.6%) and lavandulyl acetate (max. 18.1%). Among aromatic compounds, the most abundant was chamazulene (max. 29.1%). This screening has shown that only 10 populations out of 28 satisfied official quality requirement of 0.02% of chamazulene in the dried drug. Four populations had higher yield than commercial variety ProA, while one of them had even higher level of chamazulene.


INTRODUCTION
Genus Achillea includes about 120 perennial species, which are mostly present in Euro-Asian region. Plants from this genus are, in wide diversity, spread in Serbian territory, where 19 species have been recognized (Gajić, 1975). Regarding pharmacological properties, the most important species belong to the Millefolium group, which is characterized by wide morphological, cytological and chemical diversity, and it has been botanically systemized in few subspecies whose chromosome numbers are ranging from diploid (2n=18) to octaploid (8n = 72) forms (Nemeth, 2005). Therefore A. millefolium L. sensu lato is the developing evolutionary taxon currently considered as an aggregate of 12 species (Saukel and Langer, 1992). Since this group consists of several species difficult to distinguish which again are polymorphic, the raw material from wild collecting often represents a mixture of several species. According to European Pharmacopoeia (Ph.Eur.8.0., 2013), the main quality parameter is the essential oil content (minimum 2 ml/kg), with not less than 0.02% of proazulenes, expressed as chamazulene. Diploid species (A. asplenifolia Vent. and A. setacea Waldst. et Kit.) and tetraploid species (A. collina Becker ex Reichenb.) are considered pharmaceutically acceptable, while hexaploid species (A. millefolium L. sensu stricto and A. distans Waldst. et Kit. and octaploid species (A. pannonica Scheele.) are considered as chamazulene-free and therefore should be avoided (Chandler et al., 1982). Yarrow (Achillea millefolium Agg.) is well-known medicinal plant, with a wide spectrum of applications. It is one of the most frequently used plant drug in Serbia (Tucakov, 1984). Upper parts of the plant collected during the blooming are recognized as an anti-inflammatory, antinociceptive, spasmolytic, antimicrobial and holagog drug (Nadim et al., 2011;Cavalcanti et al., 2006;Benedek et al., 2008;Ali et al., 2017). Yarrow preparations in the form of infusions, decoctions or fresh juices have been applied against various indications such as anorexia, stomach cramps, flatulence, gastritis, enteritis, internal and external bleeding (Willuhn, 2002;Wichtl, 2004). In this study, we have observed cultivation and chemical properties of 28 A. millefolium populations collected from Serbian sites. The main objective of this research was to evaluate local yarrow populations for yield components and chemical constituents of the essential oil regardless their explicit botanical taxonomy. Furthermore, we tried to explore what would be the yield and quality of only upper 15 cm of plant considering that this part represents "flowering tops" as it is suggested by official quality standard (Ph.Eur.8.0., 2013).  -35). Each collecting site was at least 5 km distant (air distance) from the nearest one. Additionally, seeds of three populations (36-38) were purchased from seed exchange programme (Hohenheimer Gärten) and one commercial variety (ProA, Pharmaplant) has been included (39). All collected populations were sown in the glass-house for seedlings production, but only 28 among them had seeds viable enough to form a sufficient number of seedlings to be included in field cultivation experiment. Details of included populations are listed in Table 1.

Essential oil extraction
The air-dried yarrow plants (useful parts) were grounded, and the volatile oils were obtained by hydrodistillation using a Clevenger-type apparatus according to Procedure I of the Yugoslav Pharmacopoeia IV (Ph.Jug.IV, 1951). The essential oil yield, expressed as a percentage, was calculated on a moisturefree basis. Samples (20 µL) were dissolved in EtOH 96% (2 mL) and kept in the refrigerator until further analysis. A single plant from population 31 was morphologically very distinct from rest of the population. Thus, that plant was analyzed separately and marked as 31a.

GC-FID analysis
The GC-FID analyses were carried out with HP-5890 Series II apparatus (Hewlett-Packard, Waldbronn, Germany) equipped with a split-splitless injector, a flame-ionization detector (FID), and HP-5 capillary column (25 m×0.32 mm i.d., film thickness 0.52 µm). The oven temperature was programmed rising from 40 to 260°C at 4°C/min; injector temperature 250°C; detector temperature 300°C; carrier gas, H 2 (1.0 mL/min). Samples were injected in the amount of 1 µL. Split ratio was 1:5. The relative contents expressed as percentages were obtained from electronic integration of the peak areas measured using FID.

GC/MS analysis
The GC/MS analyses were performed under almost the same analytical conditions as the GC-FID analyses, with HP G 1800C Series II GCD analytical system (Hewlett-Packard, Palo Alto, CA, USA) equipped with HP-5MS capillary column (30 m×0.25 mm i.d., film thickness 0.25 µm). Helium (1.0 mL/min) was used as carrier gas, and the transfer-line temperature (MSD) was heated to 260°C. Mass spectra were acquired in the EI mode (70 eV) over the m/z range 40-450 amu. Samples (1 µL) were injected in split mode (1:30). The identification of the constituents was performed by comparing their mass spectra and retention indices (RIs) with those obtained from authentic samples and/or listed in the NIST/Wiley mass-spectra libraries, using different types of search (PBM/NIST/AMDIS) and available literature data (Hochmuth, 2006;Adams, 2007).

Statistical analysis
Since measurements of a large number of replications in agronomical assay produced data with very wide variance in populations, estimation of differences between groups was inaccessible. Therefore for each population quantile reduction of data has been applied, where only second and third quartiles have been taken for further statistical analysis. The hypothesis that all populations yield useful parts equally has been evaluated with one-way ANOVA and differences among mean values has been tested with post-hoc Duncan's multiple range test. Production differences among populations have been presented graphically (bar plot with standard deviations). The strength of yield components relationships has been estimated through Pearson's correlation coefficients. Essential oil compositions of observed populations have been presented in the table, while their similarity has been accessed through cluster analysis with Euclidian single linkage distance and presented graphically as an unrooted dendrogram. All statistical computing and graphs production was made by R software packages (The R Project for Statistical Computing).

Yield components
In the second vegetation, yarrow populations' yield ranged from 925 -3630 kg/ha ( Figure 1). Our findings differ from previously published papers most probably because we have observed yields of only upper 15 cm of the plants. Giorgi et al. (2005) reported yield in range of 5-12 t/ha, while Dachler and Pelzmann (1999) reported ca. 4.5 t/ha. Since these papers lack in cutting height information, we believe that discrepancies in reported yields and the yield range of our research is result of our intention to estimate dry mass of the "flowering tops". Population 20 had the highest yield, while population 27 had the smallest, 3630 kg/ha and 925 kg/ha, respectively. Most of the populations, accurately 25 of them, yielded more than 2000 kg/ha. Populations 1, 19, 20 and 25 showed the highest yield potential and no statistically significant differences among their mean values has been observed. The smallest yield potential has been recorded in populations 27 and 32. Yield components have been observed through correlation matrix ( Figure 2). Whole plant yield ranged from 20 -300 g/plant and showed the strongest correlation to useful part yield (r = 0.96). This was expected since robustness of the whole plant was inherited by the useful part of the plant trait.
In other words useful part of the plant was only the part of the whole plant with exactly the same number of stems, which is strongly correlated with both traits (r = 0.64 and r = 0.68, respectively). On the other hand, plant height and width traits showed weaker correlation to the useful part of the plant (r = 0.31 and r = 0.45, respectively). This is observations clearly showed that only number of stems, among morphological traits, contributes to useful part yield.

Essential oil yield
The yield of essential oil of the cultivated yarrow populations in our experiment ranged from 0.40 -0.82% (Figure 3). Chamazulene-free populations have been represented in the figure with white bars. Total count of chamazulene-free populations was 11 and at this level we can distinguish which population is not suitable for pharmaceutical purposes according to official standards (Ph.Eur.8.0., 2013). The highest oil yield was observed in population 27 (0.82%), while the lowest yield was recorded in population 14 (0.40%). Majority of oil yield (59% of populations) have been observed in amounts more than 0.5%. These data are accordance than previously reported, where yarrow oil yield ranged from 0.11 to 1.03% (Verma et al., 2017;Spinarova and Petrikova, 2003;Raal et al., 2012;Ghasemi Pirbalouti, 2017). Dachler and Pelzmann (1999) reported yarrow yields up to 0.5%. According to mean values of reported yields, oil content in our samples was above average. Again, this is most probably the result of plant cutting height in our experiment.

Essential oil composition
Chromatogram integration of yarrow essential oils recognized 113 components out of which we were able to identify 96 (Table 2). One of these identifications was tentative (RI = 1679.4).

Constituents a RI b % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m % m/m
Cluster analysis of yarrow essential oil chemical profiles in our experiment revealed branching showed on unrooted dendrogram in Figure 4. Population grouping based on single linkage Euclidian distances has distinguished sub-branch with populations rich in chamazulene (18.3-29.1% of essential oil or over 0.09% in the dry drug). Those populations are 20, 26, 27 and 32, and all of them could be included in the future selection process. On the same branch, sabinene rich populations (pop. 38, 36 and 39) formed sub-branch. Commercial variety ProA (pop. 39) has been grouped on the same branch with chamazulene rich populations, but thanks to its high level of sabinene (19.3%) it has been grouped in sub-branch with sabinene rich populations.

CONCLUSION
Screening of yarrow populations in Serbia revealed valuable information regarding yield of raw material and quality of the essential oil. The yield has been observed in the range of 925 -3630 kg/ha, while essential oil content ranged from 0.40 -0.82%. The most dominant compounds in monoterpene fraction were β-pinene (max. 36.3%), sabinene (max. 35.7%), 1,8-cineol (max. 26.6%) and borneol (max. 20.2%), while in the sesquiterpene fraction the most abundant compounds were trans-caryophyllene (max. 18.6%) and lavandulyl acetate (max. 18.1%). Among aromatic compounds, the most abundant was chamazulene (max. 29.1%). Moreover, this screening has shown that only 10 populations out of 28 satisfied official quality requirement of 0.02% of chamazulene in the dried drug. Four populations (1, 19, 20, 24) had higher yield than commercial variety ProA (39), while one of them (20) had even higher level of chamazulene (23.1%). These results favors cultivation of species with known chemical profile versus wildcrafting of morphologically very similar yarrow plants which do not satisfy quality requirements.