CHEMICAL COMPOSITION OF ASTER ALBANICUS DEG . ( ASTERACEAE ) ESSENTIAL OIL : TAXONOMICAL IMPLICATIONS

The composition of essential oil isolated from the areal parts of Aster albanicus Deg, an endemic species of the central Balkans, was analyzed. In total, 111 compounds were identified, representing 98% of the essential oil. The essential oil was dominated by sesquiterpene (69.3%) and monoterpene hydrocarbons (15.9%), with germacrene D as the most abundant compound (34.7%). Several multivariant statistical methods (HCA, NJ, PCoA) were deployed to infer the relation between A. albanicus and other species belonging to this genus. Taxonomical implications are discussed.


INTRODUCTION
The genus Aster L. (Asteraceae) is comprised of ca.180 Eurasian species, 17 in SE Africa and 1 in N America (Mabberley, 2008).In the flora of Europe, ca. 30 species are recognized.Aster albanicus Deg. is an endemic species growing wild in the mountain border between Serbia and Albania (Tutin et al., 1976).Traditional treatments place most species in the genus Aster L. According to the latest analyses of morphology, chloroplast DNA RFLP and ITS sequence data, as well as karyotype studies, species of Aster are polyphyletic and members of a number of very distinct phylads within the tribe (Lane, 1982;Guy L. Nesom, 1994;Semple et al., 1996;Semple et al., 2001;Noyes and Rieseberg, 1999;Brouillet et al., 2001).
Essential oils have been previously used in the chemotaxonomy of conifers (Šarac et al., 2013;Rajčević et al., 2013) and angiosperms (Harborne and Turner, 1984;Pérez et al., 2000;Skaltsa et al., 2001;Maggio et al., 2012).Thus, further chemical investigation of Aster including essential oil (terpenoids) composition and distribution (Tsankova and Bohlmann, 1983;Bohlmann et al., 1985;Chung et al., 1993) show potential chemotaxonomical significance.The composition of the essential oil of Aster albanicus Deg. was not previously investigated.Our results in combination with the available data might prove helpful in future infrageneric classificatory schemes of this genus.

Plant material
Aerial parts were collected from the population of Aster albanicus Deg.(Asteraceae) growing wild on Mt.Rogozna, north of Kosovska Mitrovica, during the autumn 2007-2008.A voucher specimen has been deposited at the Herbarium of the Institute of Botany, University of Belgrade, Faculty of Biology (BEOU).Crude essential oil was obtained by 2-h simultaneous distillation-extraction (SDE) in a Likens-Nickerson-type apparatus (Likens and Nickerson, 1964;Chaintreau, 2001) from dried areal

CHEMICAL COMPOSITION OF ASTER ALBANICUS DEG. (ASTERACEAE) ESSENTIAL OIL: TAXONOMICAL IMPLICATIONS
parts of the plant (50 g).The volatiles were collected in CH 2 Cl 2 .

GC/GC-MS (gas chromatography-mass spectrometry)
Analysis was performed on an Agilent 7890A GC system equipped with 5975C MSD and FID, using DB-5 MS column (30 m × 0.25 mm × 0.25 μm).Injection volume was 1 μL and injector temperature was 220°C with a 10:1 split ratio.Carrier gas (He) flow rate was 1.0 ml/min at 210°C (constant pressure mode).Column temperature was linearly programmed in a range of 60-240°C at a rate of 3°C/ min.The transfer line was heated at 240°C.The FID detector temperature was 300°C.EI mass spectra (70 eV) were acquired in m/z range of 30-550.A library search and mass spectral deconvolution and extraction were performed using NIST AM-DIS (Automated Mass Spectral Deconvolution and Identification System) software version 2.64.113.71, using retention index (RI) calibration data analysis parameters with a "strong" level and 10% penalty for compounds without an RI.The retention indices were experimentally determined using the standard method involving retention times of n-alkanes, injected after the essential oil under the same chromatographic conditions.The search was performed against our own library, containing 4972 spectra.Percentage (relative) of the identified compounds was computed from GC peak area.

Statistical procedures
These are indicated in the results and figure captions article.a ) Kovats indices obtained experimentally using the standard method involving retention times of n-alkanes, injected after the essential oil under the same chromatographic conditions.b ) Contents are given as percentages of the total essential oil composition; tr: trace (0.05<tr<0.10%);compounds with contents <0.05% are not listed; c ) Others: aliphatic hydrocarbons, aliphatic aldehydes and alcohols, aliphatic acids and their esters and aldehydes, aromatic ester + aliphatic acid, alkyl aromatic alcohols, aryl esters of aromatic acid.
The genus Symphyotrichum belongs to the subtribe Symphyotrichinae G. L. Nesom (1994) and Doellingeria is an unplaced genus.On the other hand, the genus Aster belongs to the subtribe Asterinae (Cass.)Dumort ( 1827) of the tribe Astereae Cass.(1819) (G.L. Nesom and Robinson, 2007).In relation to other investigated species, the essential oils of species belonging to Aster s.l.varied, from those dominated by monoterpenes to those dominated by sesquiterpenes.There was no evident genus-related pattern in this domination.Furthermore, the composition of the essential oil of A. albanicus has the most similarities with A. ageratoides, even though sesquiterpenes dominated the former and monoterpenes the latter.
To infer similarities between samples, several statistical methods were deployed using the presence/absence of compounds in the essential oil (without taking into account the average abundances of the components).Cluster analyses using dif-   According to all available data and the present results, we concluded that the relationship of Aster s.str.and other genera belonging to Astereae is unclear, and the systematics within the subtribe Asterinae in the sense of Nesom ( 2007) is still uncertain.Even though D. scabra was always separated, Aster s.str.and the species belonging to other genera, i.e.Kalimeris, Miyamayomena, Symphyotrichum, always grouped together, forming mixed subclades.However, more extensive study of the essential oil of Aster species, which should take into account both compounds and their relative abundances in essential oils, is necessary to enhance resolution and improve systematics within the subtribe Asterinae.ferent distance measures (Euclidean, Jukes-Cantor, Gower) and UPGMA resulted in the same trees, with high bootstrap support (Fig. 1).The neighborhood joining tree with D. scabra used as an outgroup gave similar results, grouping A. albanicus and A. ageratoides close together and the rest of the species in a related subclade (Fig. 2).Parsimony analysis using the branch-and-bound algorithm and Wagner optimization resulted in several trees, from which the optimal one was chosen (i.e. the one with the highest bootstrap support) (Fig. 3).All cluster analyses gave similar results.D. scabra was always in a separate clade from the other species.Furthermore, A. albanicus and A. ageratoides were always grouped close together, next to all other Aster species, including species from other genera (e.g.Symphyotrichum, Kalimeris etc.).Species from genus Symphyotrichum never formed a separate group.

Fig 1 .
Fig 1. Cluster analysis based on presence/absence of components in essential oil of Aster s.l.species (Euclidean distances, UPGMA).Fig 2. Neighbor joining tree based on presence/absence of components in essential oil of Aster s.l.species.

Fig 2 .
Fig 1. Cluster analysis based on presence/absence of components in essential oil of Aster s.l.species (Euclidean distances, UPGMA).Fig 2. Neighbor joining tree based on presence/absence of components in essential oil of Aster s.l.species.

Fig 3 .
Fig 3. Parsimony analysis based on presence/absence of components in essential oil of Aster s.l.species.

Table 1 .
Relative abundances of compounds in essential oil of Aster albanicus Deg.obtained by SDE.