Supplementary material

Ibrahim Sifi^a, Nadhir Gourine^a∗, Emile M. Gaydou^b and Mohamed Yousfi^a

Email: sifi_ibrahim@yahoo.fr

Affiliation: Laboratoire des Sciences Fondamentales

Address: Laboratoire des Sciences Fondamentales à l’Université Amar TÉLIDJI de Laghouat, Route de Ghardaïa, BP37G, (03000), Algérie.

Email: n.gourine@mail.lagh-univ.dz

gourine.nadir@gmail.com

Affiliation: Laboratoire des Sciences Fondamentales

Address: Laboratoire des Sciences Fondamentales à l’Université Amar TÉLIDJI de Laghouat, Route de Ghardaïa, BP37G, (03000), Algérie.
Emile M. Gaydou

Email: emile.gaydou@univ-cezanne.fr

Affiliation: UMR CNRS 6263, Équipe AD2EM (Phytochimie), Institut des Sciences Moléculaires de Marseille.

Address: Université Paul Cézanne, Faculté des Sciences et Techniques de Saint-Jérôme, Avenue Escadrille Normandie Niémen, Case 461, Marseille Cedex 20, France.

Email: yousfim8@gmail.com

Affiliation: Laboratoire des Sciences Fondamentales

Address: Laboratoire des Sciences Fondamentales à l’Université Amar TÉLIDJI de Laghouat, Route de Ghardaïa, BP37G, (03000), Algérie.

Laboratoire LSF, B.P. 37G, Université Amar TÉLIDJI, Laghouat, (03000), Algeria.

Tel.: +213 29 90 00 66

Fax: +213 29 90 00 66

E-mail address: n.gourine@mail.lagh-univ.dz

gourine.nadir@gmail.com

The essential oils of unripe galls (from male and female plants), of a total number of 52 samples of Pistacia atlantica collected from different regions in Algeria were analysed by GC/MS and GC. The yields of the extraction of the essential oil by hydrodistillation varies from low to high values (0.08-1.89% v/w). The results of both methods of Principal Component Analysis (PCA) and Hierarchical Ascendant Classification (HAC) revealed the presence of two different chemotypes: α-pinene chemotype and α-pinene/sabinene/terpinen-4-ol chemotype.

In Greece the fruits of P. atlantica are used for tanning and as fodder for cattle and since they contain oil, it is used for soap making . The fruits of P. atlantica which smell like mastic are also occasionally chewed by local inhabitant (for mouth flavouring). From the bark of the wood, a resin is collected for laquer production and it is also used in popular medicine “as an antiseptic to wounds, etc.” . The oleoresin of P. atlantica var. mutica, known as “Turk terebinth gum,” is used to make chewing gum in Iran. This plant has also been used traditionally in the treatment of peptic ulcer and as a mouth freshener . In Algeria and Morocco, P. atlantica is important because it is the source of mastic gum, an exudates which strengthens gums, deodorizes breath and combats coughs, chills and stomach diseases . Furthermore, P. atlantica along with Ferula persica, Paronychia argentea are the three plants which are widely recommended by the herbalists and used for their hypoglycaemic activity in Jordan . In Algeria, the galls of P. atlantica are used as an embalming agent by rural inhabitant. They are also known in Arabic as ‘‘afse’’ and are edible and sold in markets.

Pure authentic samples myrcene, β-phellandrene, globulol and phytol were purchased from Aldrich. Linalool, terpinen-4-ol, α-terpineol, were purchased from Fluka.

Fresh unripe galls, from both male and female trees of P. atlantica Desf., were randomly collected from the bottom of the trees branches during the period of the end of July and the beginning of August 2010. Two regions in Algeria were selected. The first region Aïn-Oussera belongs to the town of Djelfa (39 samples, among them 13 were from male trees), while the second region belongs to the town of Laghouat (13 samples). For this region of Laghouat, we have selected three locations of the plant collections: the location of south Laghouat (2 samples, male and female), the location of Kheneg (3 male samples and 4 female samples) and the location of Tilghemt (2 female, 2 male). The chosen numbers of samples collected from each region were roughly proportional to the availability of the plant in each one of these regions.

The locations of Aïn-Oussera, Kheneg, south of Laghouat and Tilghemt “Hassi R’mel” are located at 200, 400, 440 and 500km, respectively south Algiers, the capital of Algeria. The location of Aïn-Oussera (latitude 35°33' (N); longitude 02°31' (E); altitude 649m) was characterized by annual precipitation of 25mm and mean summer temperature of 37.8 °C. The locations of Laghouat and south of Laghouat (latitude 33°47' (N); longitude 02°52' (E); altitude 750m) were characterized by annual precipitation of 18mm and mean summer temperature of 41.4 °C. Finally the location of Tilghemt (latitude 32°95' (N); longitude 03°18' (E); altitude 774m) was characterized by annual precipitation of 16mm and mean summer temperature of 42.3°C. All these locations have same the kind of soil which is the clay soil. This soil is locally known as “Daya”.

For comparison purpose, mature galls (dark brown color) were also collected from the regions of Aïn-Oussera and Laghouat. In addition, other samples were purchased from local markets of medicinal plants in the town of Laghouat (Algeria).

The plant material was identified by staff at the herbarium of National Superior School of Agronomics (ex. INA), Algiers, Algeria. A voucher specimen (PAUG-52S/08/10) was deposited in the herbarium of the Fundamental Sciences Research Laboratory at Laghouat University.

2.3. Preparation of samples

The essential oils were obtained by hydrodistillation using a Clevenger apparatus for almost 2 hours, time which was necessary for the complete extraction. The obtained essential oils were dried over anhydrous sodium sulphate, and stored at +4°C until analyzed.

A CP-Varian 3800 gas chromatograph was used with a flame ionization detector (FID), and a UB-Wax fused silica capillary column (60m×0.32mm, 0.25μm film thickness). Oven temperature was programmed from 50°C to 250°C at a rate of 3°C/min and held at 250°C for 10 min. Injector and detector temperatures were set at 250°C and 260°C, respectively. Helium was the carrier gas at a flow rate of 1mL/min. Diluted samples (1:100 v/v, in n-pentane) of 1 μL were injected manually using split ratio of 80:1.

The GC/MS analyses were performed on an Agilent 6890 GC machine equipped with a capillary column DB-Wax (30 m × 0.25 mm, 0.25 μm film thickness) and coupled with a Quadrupole CMSD 5973 detector in electron ionization mode (EI 70 eV). Helium was the carrier gas, at a flow rate of 1 mL/min. Injector and MS transfer line temperatures were set at 250 and 220°C, respectively. Oven temperature was programmed from 50°C to 250°C at a rate of 3°C/min and held at 250°C for 10 min. Diluted samples (1:100 v/v, in n-pentane) of 1 μL were injected manually using split ratio of 80:1.

Linear retention indices were calculated relative to linear homologous series of n-alkanes (C₈–C₄₀). The identifications of the components were based on the comparison of their mass spectra with those of Wiley and NIST (National Institute of Standards and Technology) libraries, as well as by comparison of their retention indices with those of the values of a home made database and those of components of known oils and by co-injections of some pure authentic samples.

The results shows that the main compounds were not the same for the overall analyzed samples (male and female tree) and coming from the different locations i.e. the major compound is sometimes being a minor compound and vice-versa. In order to investigate the differences between the essential oil samples of the different regions of our collections, we have chosen two methods: Principal Component Analysis (PCA) and Hierarchical Ascendant Classification (HAC). The data set, composed of 52 samples and 14 variables (main compounds), were subjected to multivariate statistical analysis for chemotype determination.

For the PCA method, loading factors for principal axes F1 and F2 with rotation Varimax are given in Figure 1S. The factor loading of the 14 variables on axes F1 and F2 (circle of correlations), using PCA of the chemical composition obtained by gas chromatography of the 52 essential oil samples represented in Figure 1S, shows strong correlations with some components of the essential oils. Tricyclene is strongly correlated with camphene (R=0.992), sabinene is highly correlated with both p-cymene (R=0.839) and terpinen-4-ol (R=0.868) and p-cymene is also highly correlated with terpinen-4-ol (R=0.861). On the other hand, α-pinene is reversibly highly correlated with terpinen-4-ol (R= −0.766), sabinene (R= −0.766) and p-cymene (R= −0.71). Axis F1, which represents 33.92% of the total information, is highly negatively correlated with α-pinene and in good correlation with both tricyclene and camphene, and positively highly correlated with terpinen-4-ol, sabinene and p-cymene. This axis is also in good correlation with γ-terpinene and isocineole. This axis separates the samples into two point clouds (Figure 1). The F2 Axis, which represents 18.84% of the total information, is positively and highly correlated with only one compound which is spathulenol, and negatively correlated with limonene, β-pinene, δ-3-carene and myrcene. This last axis can note separates the clouds of samples (Figure 1). The α-pinene chemotype was found to be more abundant in comparison with sabinene/terpinen-4-ol chemotype, since the number of the samples of this cluster represents 59.6% of the total samples used.

The second method used was HAC (Hierarchical Ascendant Classification). The cluster analysis was performed using the Ward’s technique. Similarly to PCA method, the result of HAC showed and confirmed the existence of the same two different clusters (I and II) within the essential oils of the individuals of P. atlantica (Fig. 2). The distance measuring the dissimilarity was very important between the two identified clusters, which mean that the two chemotypes could be easily distinguished. Taking into account the chemical content range, we can consider two groups or chemotypes, as shown in Figure 2S. These two chemotypes could be easily differentiated upon a three different components: α-pinene, sabinene and terpinen-4-ol. The table 1S gives the data for each chemotype. Finally, it was detected only one high value of unidentified component “unknown 5” with a percentage of 18.4% which was the opposite of the rest of samples where its content didn’t exceed 3.3%.