Seasonal changes in rosemary species: A chemotaxonomic assessment of two varieties based on essential oil compounds, antioxidant and antibacterial activities

Rosemary (Rosmarinus officinalis L.) is a popular herb in cooking, traditional healing, and aromatherapy. This study was conducted to evaluate the effects of meteorological conditions plant growth stage and genetic factors on the yield, quantitative and qualitative composition, on the antioxidant and antimicrobial activities of rosemary essential oil from two Tunisian locations (El Fahs and Matmata) during two successive years. The composition of the essential oils obtained by hydrodistilation from rosemary plants were carried out annually using GC and GC/MS. Results showed the the main constituents were camphor (18.2–28.1%), 1,8-cineole (6.4–18.0%), α-pinene (9.7–13.5%), borneol (4.4–9.5%), and camphene (5.1–8.7%). The principal component and heatmapper analyses showed group segregation of the two studied varities based on major essential oil compounds. Additionally, in vitro antimicrobial and antioxidant activities showed that rosemary essential oils had an important ability in scavenging DPPH, as well as a higher bactericidal effect. The seasonal variation, growth stage and genetic pools seemed to be a factors of significant variation of the composition, antimicrobial and the antioxidant activities of the rosemary essential oils. These finding would be taken to use the chemotaxonomy tools to develop a program for Rosmary protection conservation and identification based on essential oil composition.


Extraction of essential oil
One hundred grams of dried leaves from each sample were submitted to hydro-distillation for 3 h, using a Clevenger-type apparatus. Essential oils were recovred directly and stored in sealed vials protected from light at 4˚C until analyses.

Essential oil analyses
Gas chromatography (GC). The rosemary oil analysis was conducted using A Hewlett-Packard 5890 series II gas chromatograph equipped with HP-5MS capillary column 30 m × 0.25 mm i.d., film thickness 0.25 μm; Hewlett-Packard) and connected to a flame ionization detector (FID). The column temperature was programmed at 50˚C for 1 min, then 7˚C/ min to 250˚C, and then left at 250˚C for 5 min. The injection port temperature was 240˚C and that of the detector 250˚C (split ratio: 1/60). The carrier gas was helium (99.995% purity) with a flow rate of 1.2 mL/min and the analysed sample volume was 2 μL. Percentages of the constituents were calculated by electronic integration of FID peak areas, without the use of response factor correction. Mean percentage of compounds in R. officinalis L. essential oils represented the average calculated on five individuals (n = 5). Retention indices (RI) were calculated for separate compounds relative to (C 7 -C 25 ) n-alkanes mixture (Aldrich Library of Chemicals Standards) [24].
Gas chromatography/Mass spectrometry (GC/MS). The isolated volatile compounds were analysed by GC/MS, using an Agilent Technologies 6890 N gas chromatograph. The Ua: Upper arid, Usa: sub-humid a Bioclimatic zones were defined according to Emberger's (1996)  fused HP-5MS capillary column (the same as that used in the GC/FID analysis) was coupled to an Agilent Technologies 5973B mass-spectrometer (Hewlett-Packard, Palo Alto, CA, USA). The oven temperature was programmed at 50˚C for 1 min, then 7˚C/min to 250˚C, and then left at 250˚C for 5 min. The injection port temperature was 250˚C and that of the detector was 280˚C (split ratio: 1/100). The carrier gas was helium (99.995% purity) with a flow rate of 1.2 mL/min. The mass spectrometer conditions were as follow: ionization voltage, 70 eV; ion source temperature, 150˚C; electron ionization mass spectra were acquired over the mass range 50-550 m/z. Identification of the essential oil compounds was based on a comparison of retention indices (RIs) and computer mass spectra library (Wiley 275). The retention indices were determined relative to the retention times for a series of n-alkanes (C 7 -C 25 ) [24] using linear interpolation and to those previously reported in the literature [10,25].

Antioxidant activity
The antioxidant activity was assessed by 1,1-diphenyl-2-picrylhydrazyl (DPPH) test. The DPPH radical-scavenging activity of essential oils were measured as previously described by Dinis et al. [26] with small modifications. A mixture consisting of 1 mL of methanol and 3 mL of DPPH solution was used as the control. The percentage inhibition of DPPH radical was calculated according to the following formula: % inhibition = [(AB − AA)/AB] � 100, where AB and AA are the absorbance values of the control and of the sample, respectively. The radicalscavenging activity of samples was expressed as the IC50 (μg/mL) reflecting the EO concentration would inhibit 50% of DPPH radical.

Antibacterial activity
Bacterial strains. Antibacterial activities of R. officinalis L. essential oils were tested against 4 strains of bacteria: two Gram-negative (Escherichia coli and Salmonella typhimurium) and two Gram-positive (Bacillus cereus and Staphylococcus aureus).
Antimicrobial activity assay. The antimicrobial activity of oils was determined through the disc-diffusion method according to the modified method described by Freney et al. [27]. Briefly, the bacterial suspension used to inoculate Petri dishes had a turbidity of approximately 0.5 McFarland standards. The oils were suspended in DMSO solvent at a concentration of 20 μg/mL. Then, filter paper discs (6 mm diameter) were separately impregnated with 10 μL of the different oils and put on the surface of the inoculated plates (90 mm). The Petri dishes were left at + 4˚C for 2 h to facilitate the diffusion of essential oils in agar and then incubated at 37˚C for 24 h for bacterial strain. DMSO (10%) was used as a negative control. Gentamicin (10 μg/disk) was used as a positive control. Antimicrobial activity was assessed by measuring the inhibition zone around each disk. All experiments were carried out in triplicate.

Statistical analysis
The 31 compounds identified of the essential oil was checked by a descriptive statistical analysis using the SPSS software for Windows™ (version 11.5). The percentage of compound were transformed using the arcsine transformation to improve the distribution property. However, this transformation did not yield satisfactory results for 18 variables. Therefore, for compounds having skewed distributions, a nonparametric one-way analysis of variance Kruskal-Wallis test was performed. The chemical population structure was assessed by Linear Discriminant Analysis (LDA). Duncan's multiple range test (p < 0.05) was used to compare the averages of essential oil yields and the DPPH radical-scavenging activities among populations. Multivariate analyses were used based on essential oil compounds, yield and antoxidant activity including pricipal component analysis (PCA) and heatmap online analysis (http://www. heatmapper.ca/). A heatmap is a graphical representation of data using a color-coding system from lower to highest values. The Rosmary essential oil samples were classifed based on average linkage and using Euclidian distance.

Essential oil yield according to population locations and phenological stages
Sixten spontaneous populations of R. officinalis populations 1 (var. troglodytorum) and populations 2 (var. typicus) were collected during the vegetative, flowering and fruiting stages at two years from each studied region. These collecting sites belonged to two different micro-edaphoclimatic environmental conditions. Populations 1 is located at the South West of Tunisia (Matmata) in upper arid climate characterized by a mean rainfall of 100-200 mm/year. The population 2 is localized at the North West of Tunisia (El Fahs) in upper semi-aride climate characterized by a main rainfall of 400-500 mm/year (Fig 1). The monthly means of temperature and precipitation at the sampling locations were shown in Fig 1. The altitudes ranged from 220 m (population 2) to 600 m (population 1) ( Table 1). Matmata is drier than El Fahs throughout the year. The vegetations of the collection sites are influenced by the Mediterranean climate, which has less rainfall in the summer. The yield of essential oils at growth stage varied from 1.91% to 3% during 2011-2013 (Table 2). It's affected by meteorological conditions, years and phenological stage in each region. The average yield of rosemary essential oil was highest in early summer for both sampling locations having a value of 3% in June 2012 (Fr2) for the var. troglodytorum collected from the Matmata region and 2.17% in June 2011 (Fr1) for the var. typicus from El-Fahs region.
The var. troglodytarum from Matmata region showed the highest EO yield at the fuiting stage from April to June distinguished by a decrease in precipitation, since the average precipitation was 6.8 mm in April 2012, absent in May 2012 and in June 2012 and the average temperature was around 25˚C.
The highest EO yield in fruiting phase can be explained by the lower average precipitation Sotomayor et al. [28] analysed the effect of water level on the quality of essential oil in the thymol chemotype of Thymus zygis subsp. gracilis and established that the highest amount of essential oil was produced under the lowest (30%) level of watering. It has been reported that lower amounts of moisture increase the yield of essential oil in phenolic chemotypes of other species from Lamiaceae family: for example, the carvacrol chemotype of wild oregano (Origanum vulgare) accumulated more essential oil under lower rainfall [29].
On the other hand, June 2011 in both sampling locations, coincided with the highest temperature compared to other years. Temperature can influence the accumulation of essential oil in different bearing plants both positively and negatively [30]. Analogous results were also obtained through investigations on other species of the Lamiaceae family. The essential oil yield was negatively correlated with higher temperature in wild Feoniculum vulgare from Iran [31], and positively correlated in a Eucalyptus species from Brazil [32]. The phenolic (thymol and carvacrol) chemotypes of T. vulgaris growing in natural habitats occur predominantly at hot and dry sites [33,34].
In conclusion, whatever the phenological stage, the troglodytorum var. from Matamta station showed the highest EO yield particularly in the fruiting phase. This highest yield was recorded also for the same variety troglodytorum compared to other Tunisian varities studied by Zaouli et al. [10].

Variation of essential oil composition according to population locations and phenological stages
The EO composition was mainly investigated using both GC and GC/MS techniques and the percentages of the identified compounds were listed in Table 2. Thirty-one compounds representing 93.1 to 99.91% of the total essential oil were identified At all stages, monoterpenes (hydrocarbons and oxygenated) were dominant in R. officinalis EO independently to the  [35]. Referring to obtained results (Fig 1 and Table 2) we noted that the higest amounts of 1,8-cineole were accumulated at vegetative and floraison stages in the two R. officinalis varieties. While the camphor concentration was higher at vegetative and fruiting stages. Variation in the essential oil quantity and the predominant compounds at different phenological stages is a characteristic of essential oil-bearing plants [36].
During all growth stages the same genotype can synthesize oils with different composition. This criterion would be used to define different chemotypes confirms the opinion that the essential oils composition depends on the time of collect. Also, for the definition of chemotypes it is not enough to base this on a chemical analysis of oil from one phenophase only [35].
On the other hand, in El Fahs region, the content of 1.8-cineole in EOs of var. typicus was significantly higher compared to troglodytorum variety. At all development stages the content of camphor was higher in troglodytorum compared to typicus variety EO. The difference not only in the major compounds but also in the minor compounds.
In addition, the results also showed that the percentage of camphor and α-pinene have significantly decreased, sometimes up to half during the second years. An increase in the percentage of 1.8 cineol, borneol and camphene, was revealed at Vg2 and Fl2. The disappearance of some compounds such as myrcene, p-cymene, alpha-phellandrene, ɣ-Terpinene and α-terpinolene at Vg2 and Fl2 in both varieties was also noted ( Table 2). It has been noted that the icrease of some compounds amount in essential oils are correlated by changes in the amounts of other compounds [34,37]. These significant differences between chemical compositions of EO in the two studied varieties can be attributed to genetic, and the geographic origin factors. Indeed, geographical, climatic and pedological characteristics of habitats explained the significant variation of EO from studied samples The geographical distribution of different chemotypes of R. officinalis essential oils are largely due to the environmental characteristics and the stages of ontogenesis as revisousely reported [35]. Vaiciulyte et al. [36] added that photosynthetically, active solar radiation and sunshine duration affect the amount of essential oil and major compounds of Thymus pulegioides.
Previous studies on the chemical composition of rosemary oil showed that the main components were camphor/1,8-cineole/α-pinene, and the intraspecific chemical variability between plants belonging to different geographical areas was noted [38][39][40][41]. Our results are in agreement with previous study carried out on rosemary species collected from Spain and characterized by 1,8-cineol and camphor as dominant compounds [42]. Likewise, these two compounds have also been described as the most dominant components in essential oils of Tunisian rosemary species [10]. On the other hand, Ojeda-Sana et al. [43] determined that rosemary essential oil collected from Argentina have α-pinene or myrcene as the main compound. The present results, showed that the geographical origin, altitude and seasons would be a source of variation the composition of rosemary essential oil composition as noted previsoulely on the quality and quantity of Iranian rosemary EO characterized by the 1.8-cineole (5.32-28.29%), camphor (1.58-25.32%) and α-pinene (14.19-21.43%) as the main constituents. Iranian accessions also exhibited chemical variability for other major compounds such as borneol, camphene, bornyl acetate. In addition, authors founded a positive and negative correlations between major constituents and environmental factors.
If we consider the same genetic origin and the micro-edaphoclimatic environmental conditions, we noted that the annual differences of meteorological conditions were the main source of variation in yield of the essential oils and/or of their major compounds. Therefore, the meteorological conditions including rainfall, temperature can influence on EOs and theirs major compounds in R. officinalis. With this possible variation the necessity of specifying the composition of rosemary essential oil according to its geographical origin and harvest date is a crucial step to conduct biological activities the essential oils.

Biological activities
DPPH free radical scavenging assay. The concentration of major compounds in rosemary EO showed seasonal variation and a significant relationship with precipitation and temperature in each sampling locations. The seasonal variation affected the chemical composition of EOs and could influence the antioxidant and antibacterial activities.
The antioxidant activity of EOs rosemary collected from the two sampling locations at differents phenologic stage (Fig 2) for two years were evaluated by DPPH (Fig 3). Significant variations (P < 0.05) were observed in antioxidant activities of rosemary EO according to the geographic origin and seasonal variations. Based on DPPH assay, the determined values ranged from 2.94 to 5.32 μg/mL for the oils extracted from El-Fahs and between 1.59 and 5.62 μg/ mL from Matmata. Rosemary EO is well known by its high antioxidant capacity [44]. The highest antiradical activity of rosemary essential oils was detected at Matmata area (var. troglodytorum) (IC 50 average = 1.59 μg/ mL) at Fr2 (June 2012, early summer), followed by the OE extracted at Vg1 (October 2012) (IC 50 average = 2.79 μg/ mL). From the upper semi-arid of var. typicus (El-Fahs), the best activity was detected in EOs extracted during the post-flowering stage (Fr1) (IC 50 average = 2.94 μg/mL). In addition, the low antioxidant activities of EOs from the two collecting regions were recorded at Vg2 stage (IC 50 average = 5.62 μg/mL for troglodytorum variety and IC 50 average = 5.32 μg/mL for typicus variety) (Fig 3). The antiradicalaire activity of R. officinalis essential oil R. officinalis had a positive relationship with aridity. According to Table 2 and the Fig 1, the upper arid (Ua) region includes the R. officinalis var. troglodytorum characterized by a high antioxidant activity compared to the upper semi arid (Usa) region includes the R. officinalis var. typicus. Matmata has a hot climate, while El Fahs has a moderately hot climate. Theses results corroborate with other studies [10]. It could be deduced also that the difference observed in the antioxidant activity level of the two varieties may be due to the variation of the two major compounds contents including the 1.8-cineole and camphor. Indeed, the antioxidant activity increased with camphor and conversely with 1.8-cineole. Compared to other native Mediterranean plants, rosemary can withstand prolonged drought by avoiding damage to its photosynthetic organs. Seasonal variation is associated with certain changes in soil moisture and temperature, which may lead to variations in the biosynthetic pathways of primary and secondary metabolites [45]. The maximum trapping capacity of the oils collected during the flowering phase for the two sampling locations could be explained by the richness of EO in 1.8 cineol, camphene, borneol and camphor compounds [46]. Although camphor and 1,8 cineole were reported as the principal antioxidant in rosemary EO [47] and, also, other compounds as the α-pinene, β-pinene, and 1,8-cineole Values are given as mean ± SD (n = 3). For the same region and bacteria, means followed by the same letter did not share significant differences at p < 5% (Duncan test).

Antimicrobial activity
Antimicrobial activity of rosmery essential oils was tested against four common bacteria pathogens Results showed that the rosmary essential oils exhibited efficient effect against all tested microorganisms ( Table 3). The R. officinalis L. is an important medicinal plant and its essential oil characterized by noteworthy antimicrobial activity [37,39,50].
The rosemary samples from Matmata showed the against E. coli than those from El Fahs at Vg1 (Sep 2011). This activity against E. coli with 28.05 mm as inhibition zone diameter was superior to those previously reported of R. officinalis EO from Tunisia (18.17 mm) [10] and from Iran (17 mm) [9]. It has been demonstrated that hydrocarbons and oxygenated monoterpenes in the essential oils are able to destroy cellular integrity, and thereby inhibit respiration and ion transport processes. This is strongly supported by the effects of different essential oils components on outer membrane permeability in Gram-negative bacteria. Most studies investigating on the action of essential oils against food spoilage organisms and food borne pathogens agree that, essential oils are slightly more active against Gram-positive than Gramnegative bacteria [51].
The variation of the antimicrobial activitiy between the all investigated essential oils samples can be attributed to their chemical composition, in particular to their abundant compounds including the 1.8-cineole, camphor, and camphene. The same suggestion was mentioned by Zaouali et al. [10] of two varieties of Tunisian rosemary (typicus and troglodytorum). However, other authors confirm that the antimicrobial effect of rosemary can not be explained only by the presence of a single substance in large amounts, but by the synergy of several components in smaller amounts [38,42,43,49].
The rosemary EO sampled in this study showed variations in chemical composition mainly in their major compounds, suggesting the variation of thier biological activities (antioxidants and antimicrobial) according the season. Therefore, the antioxidants and antimicrobial potential of rosemary EO changed during the sampling period. The changes in chemical composition and the biological activities during different seasons have been previousely reported [3,29,45].

Principal component analysis
The principal component analysis (PCA) was applied using the essential oils data. This method establishes mathematical criteria that allow similarities between samples or clusters to be expressed quantitatively. The PCA plot showed the segregation of the two varieties, typicus and troglodytorum (Fig 4). The two first principal component expressed 71,85% of total variation. The variety typicus was divided on two groups according to the growth stage and growing season. All plant growth stages for the first season (Vg1, Fl1, Fr1) and the fruiting stage of second season constituted one group. This groupe was formed based on their content on broneol and α-terpineol. The typicus variety from two growth stages (Fl2 and Fr2) of the second season were grouped based on their content of 1,8-cineole and terpinen-4-ol. The second groups were formed by troglodytorum variety at all growth stages in two the seasons (Fig 4 and Table 4).
Heatmapper analysis showed the same groups obtain by PCA (Fig 5 and Table 4). The color reflects highest (yellow) and lowest (blue) values using color score as shown in Fig 5. The troglodytorum (2 and 4) and typicus (10 and 8) varieties were closely at vegetative and flowering stage from second seasons. The 1,8-cineole, α-pinene, camphene and camphor compounds were the highest compound values having yellow color.

Conclusion
The genetic, locality, developement stages and seasons influenced significantly the composition, antimicrobial and antioxidant activities of studied rosemary essential oils. Rosemary EO antioxidants and antibacterial activities were correlated with major compounds. Based on the relative concentrations of the major components in rosemary oils, the multivariate analyses including PCA and heatmapper analyses, two chemotypes were defined. This finding would be used as a criterion for selecting the season and harvest area of rosmery to extract essential oils having crucial potentialities.