The extra virgin olive oils used in this work came from eleven different Algerian varieties of olives: Aghenfas, Akerma, Blanquette de Guelma, Bouchouk Soummam, Bouricha, Chemlal, Ferkani, Limli, Neb Djemel, Tabelout and Chemlal Tazmalt.

Fruits were manually harvested from an orchard located in Takerietz (Bejaia), during the 2008 and 2009 seasons, with the exception of the Chemlal Tazmalt variety which was harvested in the Tazmalt region (Bejaia) in east-central Algeria. Collection, transportation and processing of the olive samples were carried out rapidly and with care. Therefore, only fresh drupes without damage were selected.

Virgin olive oil samples were obtained using a laboratory oil mill (Levi-Deleo-Lerogsame), consisting of three basic elements: a hammer crusher, thermo-beater (mixer) and a pulp centrifuge. The olive fruits were milled in the hammer crusher, and then the olive paste was kneaded for 30 min with the addition of warm water (50 mL of water was added to 920 g of paste). After the vertical centrifugation, the oil was collected and left to stand. The oil samples were stored in amber glass bottles at 4 °C in the dark without headspace until analysis.

Acidity and peroxide values were determined following the analytical methods described by the International Union of Pure and Applied Chemistry (IUPAC, 1979). K₂₃₂ and K₂₇₀ were determined according to the analytical methods described by the International Olive Oil Council (IOOC, 1996).

A Liquid–liquid extraction system was carried out in the same extraction conditions, reported by Tsimidou et al. (1992) with some modifications. Briefly, 50 g of oil were dissolved in 50 mL of hexane, the solution was extracted successively three times with 30 mL of an MeOH/water (80:20, v/v) solution. The extracts were combined and washed twice with 50 mL of hexane to remove the residual oil. The hexane phase was discarded, and the methanolic solutions were concentrated and dried using rotary evaporator under vacuum at 40 °C. Finally, the residue was re-dissolved in a solution of MeOH/water (80:20, v/v).

The total phenolic content of the extract was determined by the Folin-Ciocalteu spectrophotometric method at 765 nm, using gallic acid as calibration curve (Favati et al., 1994).

According to Bendini et al. (2003), 0.5 mL of phenolic extract obtained from olive oil were dissolved in 5 mL of MeOH-water (1:1, v/v); a mixture of 4 mL of the solution and 1 mL of a 5% solution of sodium molybdate dihydrate in MeOH-water (1:1, v/v) was shaken vigorously. After 15 min, the absorbance at 370 nm was measured using caffeic acid for the calibration curve with a glass cuvette.

A solution of internal standard (1 mL of 0.015 mg·mL^-1 of syringic acid in water/MeOH (20:80 v/v) was added to a sample of virgin olive oil (2 g). The mixture was shaken (30 s) and 5 mL of extraction solution water and MeOH (20:80 v/v) were added. The obtained mixture was shaken for 1 min, extracted for 15 min in an ultrasonic bath and then centrifuged at 5000 r.p.m. (2500 g) for 25 min at T= 20 °C (NGD., 2010). The upper phase was filtered with a 0.45 µm PVDF syringe filter. 20 µL of the filtered solution were analyzed by HPLC with a UV detector at 280 nm. The HPLC system used consisted of a C18 Spherisorb ODS-2 reverse column (5 µm, 250 mm, 4.6 mm). Elution was performed at a flow rate of 1 mL·min^-1 following a gradient composed of a mixture of water and orthophosphoric acid (99.8:0.2 v/v) (solvent A), MeOH (solvent B) and acetonitrile (solvent C): from 96% (A) – 2% (B) – 2% (C) to 0% (A) – 50% (B) – 50% (C) in 60 min. The last gradient was kept for 10 min. The successive gradient was: from 0% (A) – 50% (B) – 50% (C) to 96% (A) – 2% (B) – 2% (C) in 2 min and then kept for 10 min.

The identification of phenolic compounds was performed by HPLC-MS. The main phenolic compounds were identified by comparison with relative retention times and UV spectra of pure standards.

The data reported were subjected to analysis of variance using the Statistica 5.5 package (StatSoft 97 edition). Where statistical differences were noted, differences among packages were determined using the ANOVA/MANOVA following the Newman-Keuls test. Significance was defined at p<0.05.

For the majority of the oils analyzed, the mean values of the quality parameter (acidity, peroxide index, and specific ultraviolet absorbance) fell within the ranges allowed by the IOC norms (2003) for the extra-virgin olive oil category (free acidity ≦ 0.8%, peroxide value ≦20 meq·kg-¹, K₂₇₀≦0.22 and K₂₃₂≦2.50).

The peroxide value and acidity of the studied oil were in the ranges of 3.25–8.75 meq·kg–¹ and 0.05–0.23 (% oleic acid), respectively. K₂₇₀ was between 0.107 and 0.206 and K₂₃₂ was in the range of 1.155 to 2.461.

As shown in Table 1, the amount of total phenols and O-diphenols in the analyzed oils varies significantly (p≦0.05) among the different cultivars. The results show that the amounts of total phenols and O-diphenols are highly variety-dependent. The average folin-ciocalteun content of total phenols in the samples analyzed was 218.66 mg·kg^-1, although a wide range of concentrations was observed, from 115 mg/kg (Chemlal Tazmalt variety) up to 420.95 mg·kg^-1 (Neb Djemel variety). The extract from the Neb Djemel variety had the highest contents of phenols and O-diphenols (420.95 and 29.67 mg·kg^-1, respectively), whereas, the extracts from the Chemlal Tazmalt variety (115.73 and 7.54 mg·kg^-1, respectively) and Limli variety (123.43 and 5.68 mg·kg^-1, respectively) had the lowest (Table 1). Our results are higher than those found by Taniglan et al. (2007) for several Turkish olive oil varieties. A significant positive correlation between total phenol and O-diphenol contents was observed (r=0.88, p<0.05). These results are in accordance with the correlation coefficient reported by Cerretani et al. (2006).

According to the classification of Montedoro et al. (1992), the total phenol contents determined by HPLC (Table 2) can be classified as high for Neb Djemel, Blanquette de Guelma and Bouricha, mid-range for Akerma, Chemlal, Chemlal de Tazmalt, Ferkani, Limli and Tabelout and low for Aghenfas and Bouchouk Soummam.

The identified phenolic compounds are mainly secoiridoid derivatives, phenolic alcohols, flavonoids, and lignans. The main simple phenols found in the analyzed olive oils were hydroxytyrosol and tyrosol. The concentration of tyrosol varied from 2 to 28 mg·kg^-1 in all analyzed samples, which was higher than that of hydroxytyrosol (0.5–5 mg·kg^-1). These results agree with previous data found in Tunisian varieties (Haddada et al., 2008) and Spanish and Greek varieties grown in northern Tunisia (Allalout et al., 2009). Tyrosol was the main phenolic component present in the Blanquette de Guelma variety, with 28 mg·kg^-1 (Table 2).

The most abundant secoiridoids of the olive oil analyzed were secoiridoid derivatives of oleuropein and ligstroside; these results agree with previous data found for Spanish varieties (Brenes et al., 2000; Oliveras-Lopez et al., 2007). The content of oleuropein and ligstroside derivatives ranged from 50 to 26 mg·kg^-1 in Aghenfas oils to 276 and 173 mg·kg^-1 in Neb Djemel oils, respectively.

The amount of derivatives of oleuropein in all the analyzed samples was higher than that of the derivatives of ligstroside except for the Akerma, Limli and Tabelout varieties which showed higher contents of the derivatives of ligstroside (44, 62 and 124 mg·kg^-1, respectively). Arslan, (2012) noted that oleuropein was the most abundant phenolic in the olive oil of Turkish varieties.

Flavonoids are present in considerable amounts with concentrations varying from 10 mg·kg^-1 for the Limli variety to 44 mg·kg^-1 for the Aghenfas variety. These results were higher than those reported by Ilyasoglu et al. (2010). The main flavonoid compounds quantified in our study were luteolin and apigenin. Considerable quantitative differences in their concentrations among the different varieties were noted. The luteolin amount ranged from 6 mg·kg^-1 for the Tabelout variety to 38 mg·kg^-1 for the Akerma variety. Apigenin was present at a high level in the Chemlal Tazmalt variety (14 mg·kg^-1) compared to the others.

Lignans were also present in significant amounts in all the analyzed samples at concentrations ranging from 12 mg·kg^-1 in the Bouchouk Soummam variety to 51 mg kg^-1 in the Bouricha variety. Although the variety Neb Djemel showed a high level of total polyphenols (537 mg kg^-1), it contains only 21 mg·kg^-1 of lignans, on the other hand, the Chemlal Tazmalt variety, which is less rich in total polyphenols (257 mg·kg^-1) showed a high content of lignans (41 mg·kg^-1).

Lignans are the main component of the phenolic fraction of the olive seed and are practically absent in the pulp, leaves, and limbs, therefore their presence in the oil must be due to the breaking of the stones when the olives are crushed (Owen et al., 2000).

The presence of elenolic acid in olive oil in its free form is due to the hydrolysis of oleuropein, ligstroside, and related glucosidic compounds. The Neb Djemel variety was distinguished by a high concentration of elenolic acid (68.4 mg kg^-1), the lowest value was noted for the Blanquette de Guelma variety (13.3 mg·kg^-1). This compound was found at a low concentration compared to the varieties cultivated in Chile and Spain (GarcÍa-GonzÁlez et al., 2010) which did exceed 115 mg·kg^-1.

Recently, an interest in the oxidized form of virgin olive oil phenols has significantly increased, especially in relation to determination of freshness/ageing status (Rovellini and Cortesi, 2002). The ratio between oxidized and natural forms is the key for a good quality judgment (Rovellini, 2008). A high amount of oxidized phenols was noted in the Chemlal Tazmalt variety (67 mg·kg^-1).

The concentration of oxidized elenolic was very low in all varieties and was completely absent in the Chemlal Tazmalt variety, the highest values were noted for the Aghenfas, Tabelout varieties.

The different olive oils extracts at different concentrations (0.12, 0.25, 0.5 and 1 mg·disc^-1) were screened for their antimicrobial activity against B. subtilis, S. aureus and L. innocua (Gram-positive), E. coli, P. aeruginosa, and K. pneumoniae (Gram-negative) following the agar diffusion method (Table 4). The minimal inhibitory concentration (MIC) values for bacteria (Table 5) were determined. Standard antibiotics (Gentamycin, Chloramphenicol and Ofloxacine) were used, and are also mentioned in Table 6. The solvents used for extraction were also used for dissolving the extracts, and all the solvent controls did not show any activity (data not shown).

As shown in Table 4, the results of the in-vitro antimicrobial activity assay showed that the phenolic extracts from the Bouricha and Blanquette de Guelma varieties possessed broad antibacterial activity against the microorganisms tested. The zone of inhibition varied from 6 to 26 mm and from 0.6 to 21 mm, respectively. The extracts from the Chemlal Tazmalt and Tabelout varieties only showed activity against S. aureus. Whereas the others eight extracts displayed a moderate activity against S. aureus and B. subtilis. Our results are in concordance with those reported by Medina et al. (2006) who observed higher activity in virgin olive oils and none in other edible vegetable oils (corn, sunflower, soybean, rapeseed and cotton). The authors attributed this activity of olive oil to the phenolic compounds; the dialdehidic form of decarboxymethyl oleuropein and ligstroside aglycons, hydroxytyrosol, and tyrosol were the phenolic compounds that statistically correlated with bacterial survival. A recent study by Karaosmanoglu et al. (2010) showed different antimicrobial activities among extra virgin olive oil, refined olive oil and canola oil against three microorganisms (E. coli, S. enteritidis and L. monocytogenes).This activity was also related to the high concentrations of phenolic compounds found in virgin olive oils. The oil contained many phenolic compounds; an increase in their overall effect can be related to a synergistic interaction.

The effect of each individual phenolic compound, isolated by HPLC, against L. monocytogenes, showed that the dialdehydic form of decarboxy methy-ligstroside presented a potent antimicrobial activity. These results are in agreement with ours because the Algerian olive oil varieties presented high percentages of oleuropein and ligtroside derivatives.

Gentamycine, chloramphenicol and ofloxacine, which were used as a positive control, had inhibition zones ranging from 28 to 38 mm. S. aureus, with high sensitivity to some extracts, was also sensitive to antibiotics at a concentration much lower than methanolic extracts.

The MICs reported in Table 6 show quantitative evidence of the antimicrobial activity of different extracts the olive oil varieties. The MICs of the extracts varied from 0.6 to 2 mg·mL^-1 on most of the tested bacteria. In fact, the growth of S.aureus was significantly inhibited (0.6-1.6 mg·mL^-1). The lowest MIC value (0.6 mg·mL^-1) was observed for the extract from Bouricha on S.aureus. These values of MIC are higher than those reported by Pereira et al. (2006) for the phenolic compound extract from Portuguese table olives. Nevertheless, they are in accordance with the results reported by TunÇel and Nergis (1993) on the MIC of ferulic, vanillic, and caffeic acids and tyrosol against S. aureus, which were determined as 0.6, 0.55, 0.4 and 0.6 mg.mL^-1, respectively, by the agar dilution method.

The MICs of phenolic compounds that inhibit bacterial growth are much higher than those found in olive oils. In the study by Bisignano et al. (1999), minimum inhibitory concentrations of hydroxytyrosol against Staphylococcus aureus (penicillin susceptible) and Staphylococcus aureus (penicillin-resistant) were determined; they ranged from 3.9–31.25 µg.mL^-1 and 3.9–31.25 µg.mL^-1, respectively. These concentrations of this phenolic compound are about 1000 times higher than the levels of those in extra virgin olive oil, EVOO.

Among the bacteria used in the present study, S. aureus (Gram-positive) showed a higher sensitivity to different extracts with the largest zone of inhibition and lowest MIC (Table 4 and Table 6). This was consistent with the results from previous reports, which suggested a higher susceptibility of food-borne pathogenic Gram-positive bacteria to polyphenols compared with Gram-negative bacteria (Brenes et al., 2006). This observation can be attributed to differences in the structure of the bacteria cell wall. The less complex the structure of the cell wall is in the Gram-positive bacteria, the more permeable it is to anti-microbial compounds (Papadopoulou et al., 2005).

Although the anti-microbial effects of various phenolic extracts are well documented, the mechanisms of action of these products and their components are not fully understood. Polyphenols are thought to cause surface activity that damages the membranes of bacterial cells, to inhibit its enzymes, or to interfere with the production of certain amino acids necessary for bacterial growth (Cowan, 1999). Antibacterial activity of olive oil's phenolic compounds is due to the presence of the ortho-diphenolic system (catechol) (Bisignano et al., 1999). Therefore, the position and number of hydroxyl groups on the phenol group are thought to be related to their relative toxicity to microorganisms, with evidence that increasing hydroxylation results in an increase in anti-microbial activity. Oleuropein was significantly less toxic to bacterial cells than hydroxytyrosol; one can speculate that the glycosidic group of oleuropein might render the unable to penetrate cell membranes or to reach the target site.