A First Description of the Phenolic Profile of EVOOs from the Maltese Islands Using SPE and HPLC: Pedo-Climatic Conditions Modulate Genetic Factors

Achieving economic sustainability in the olive oil production sector is a challenge. This is particularly so for small scale producers who are faced with pressing, production and marketing costs that relative to overall sales, minimise profits. In this study we aimed to describe the phenolic profile of extra virgin olive oils (EVOOs) derived from the Maltese islands. The polar fractions from EVOOs from nine indigenous (six Bidni and three Malti), one historically acclimatized tree (Bajda), 12 locally-grown but foreign cultivars and 32 foreign EVOOs were extracted using SPE (solid phase extraction), separated using HPLC analysis at 280 nm and 320 nm and identified using mass spectrometry. Application of ANOVA and Tukey post hoc hypothesis testing for analysis of variance on the peak areas identified a significantly higher concentration of p-coumaric acid, tyrosol acetate, DHPEA-EDA and oleocanthal in EVOOs derived from indigenous or historically acclimatized cultivars. Imported but locally grown cultivars showed differences when compared to the same cultivar grown in other countries, confirming that pedo-climatic conditions modulate genetic factors.


Introduction
The inhabited Maltese Islands consist of a set of three small islands, just under 316 km 2 in area (Malta 247.5 km 2 , Gozo 67.1 km 2 and Comino 2.8 km 2 ), isolated in the central Mediterranean Sea [1]. Typical of the region bordering the Mediterranean basin, the islands exhibit conditions favorable for olive (Olea europaea L. subsp. Europaea) cultivation. Historical records and cultural references abound, with olive pollen having been reported in samples from the Neolithic and Phoenician periods [2], a Bronze Age pit [3] as well as from a detailed study of a core from the Burmarrad region [4]. Remnants of olive cultivation may be evidenced by the remains of 'rustic villas' where processing of olives occurred during the Roman occupation [5], while today towns and valleys still bear place-names associated with olives and cultivation, e.g., Zejtun, Zebbug and Wied Ghajn Zejtuna. Despite the optimal geo-climatic conditions on the islands, down-sizing and abandonment of the sector in response to rationalization of land use for maximal financial or agronomic returns has been a feature of the industry [6].
The local olive culture industry offers an interesting case-study as the revival may be traced to a recent commercial initiative, the Project for the Revival of the Indigenous Olive (PRIMO). This focused on the grafting of the 'Bidni' tree where 30,000 trees were subsequently planted. The 'Bidni' variety is named after the place where the trees were found growing (Bidnija, Malta). The tree shows vigorous growth and is characterized by a small, almost elliptical, violet drupe and is markedly resistant to the fruit-fly Bactrocera oleae. The small-sized drupe is difficult to harvest but results in an oil of excellent quality [6]. Other trees that form part of the nascent industry are the Maltese olive 'Malti' and the white olive or 'Bajda'. The former shows vigorous growth, with a violet drupe that is also elliptical but blunt at the lower extremity. The tree is highly productive and the drupe is used both for oil and table purposes. The 'Bajda' olive tree is smaller and produces a pale green drupe and is mostly used for the table. The project PRIMO has served as a strong impetus and over the last two decades, olive trees such as Carolea, Frantoio, Leccino, Pendolino and the table varieties Uovo di Piccione and Bella di Spagna have become widely cultivated [7]. Currently olive groves cover a total of 128.5 hectares [7] equivalent to 8.88% of permanent crop cover or 1.3% of total agricultural land.
Achieving economic sustainability for the olive oil sector is a challenge [8][9][10], but more so locally, given the low availability of agricultural land for olive tree cultivation and the severe conflict over land use. Absorption of production, harvesting, pressing and marketing costs from a small volume of high quality olive oil to generate sustainable profits is also difficult. A niche market based on strong product valorization may permit premium olive oil prices and help to offset some of the additional costs identified above. Honojosa-Rodriguez et al. [10] argues that reduction of production cost is also an essential step and that a strategy focused on achieving quality can help to achieve 'competitive and marketing advantages'.
The ability to distinguish olive cultivars is a critical preliminary step for traceability, valorization and breeding programs [11] where genetic sequences are often favored over biochemical markers as they are independent of terroir. DNA based methods are effective in allowing a traceability chain to be established between tree and oil [12]. In a first approach, random amplified polymorphic DNA (RAPD) and microsatellite analysis [11,13] were used to identify marker sequences specific to local varieties. Both studies showed that indigenous olive trees occupied a distinct group. Presently further studies using a larger number of local varieties are ongoing. As yet, there are no papers in the published domain describing the organoleptic, biochemical and nutritional characteristics of Malta-derived olive oils with the exception of a published description of their antioxidant properties [14]. Characterization of these properties would help to identify uniqueness, while preventing adulteration or mislabeling of local extra virgin olive oils (EVOOs) while ensuring the growth of the local industry.
Several methods have been developed to characterize phenols from olive oils including high pressure liquid chromatography (HPLC) separation in conjunction with UV-vis, mass spectrometry (MS) and nuclear magnetic resonance (NMR) [15][16][17][18]. Extraction of phenolic compounds may be achieved using liquid-liquid extraction (LLE) or solid phase extraction (SPE) [19]. LLE is reported to yield a higher recovery of secoiridoids under certain conditions, but is more laborious and uses higher volumes [19]. SPE methods are well characterized and include conditioning of cartridges with hexane, application of small volumes of olive oil, followed by washing with hexane to remove non polar fractions and elution with methanol or acetonitrile [20,21]. Many separation methods on HPLC have been described using a variety of solvents and separation profiles [22][23][24].
In this study the aim was to describe the phenolic profile of Maltese-derived extra virgin olive oils (EVOOs), in so doing leading to future variable-characterization that may help to establish their authenticity and reducing the potential for adulteration with cheaper oils (both olive and seed oils) or mislabeling by foreign imported oil.

EVOO Sample Collection and Extraction
In this paper specific metabolites belonging to the polyphenolic class of compounds that could be used as biochemical markers were studied. The EVOOs for this study were derived from nine indigenous (six Bidni and three Malti), one historically acclimatized tree (Bajda) and 12 locally-grown but foreign cultivars. Indigenous and foreign but locally grown EVOOs were collected directly from olive presses following expert identification of drupes between late September to November of the harvest seasons 2014-2016 (bar 1 sample dating to 2013). Thirty-two foreign EVOOs were also purchased over the same time period 2014-2016 and included in the study (Table S1). These oils were sought with the demarcation EVOO as well as the protected designation of origin (PDO) in order to ensure traceability of the product. Samples were collected from different oil producers so as to cover a representative sample of the Maltese islands in terms of pedological conditions and pressing equipment, all olive presses used cold press technique. EVOOs were stored at 4 • C in the absence of light prior analysis, i.e., stored using standard methods in line with methods used in published papers were such methods are recorded, to assure the integrity of the analysis, then preheated to 35 • C in a water bath for an hour and mixed to ensure homogeneity. Phenolic compounds from monocultivar EVOOs were extracted using solid phase extraction (SPE) on diol-bonded phase cartridges (Analytical Columns SolGel-1ms™, Croydon, England). Of EVOO samples 3.5 g were applied under vacuum (not more than 15 mmHg) to the cartridge, washed with hexane to remove apolar components (2 × 6 mL) and twice with hexane/ethyl acetate (85:15, v/v; 8 mL) based on the method described in Mateos et al. (2001) [19]. The phenolic fraction was eluted with methanol (20 mL) filtered and concentrated to half of the initial volume under vacuum and then to dryness using a nitrogen evaporator at 45 • C to ensure complete removal of solvent.

Polyphenol Separation
All solvents were HPLC grade. HPLC analysis was carried out on an HPLC (Waters, Milford, MA, USA) using a 20 µL injection volume on a Symmetry ® C18 analytical column (250 × 4.6 mm i.d. The column temperature was set to 35 • C whilst the sample chamber was set to 10 • C to prevent phenolic degradation. UV detection was carried out at 280 nm and 320 nm. Compounds were identified using peak identification through retention time followed by analysis of molecular fragments by MS. Method transfer from HPLC was carried to UPLC MS/MS.

UPLC MS/MS Analysis
An m/z scan was carried out for every eluting peak, which allowed detection of the parent molecular ion. For each of these a separate multiple reaction monitoring (MRM) experiment to identify the most stable fragment followed using IntelliStart™ function in MassLynx™ (Waters, Milford, MA, USA). The UPLC system consisted of an AcQuity TMUPLC equipped with Waters binary pump system (Milford, MA, USA) using an AcQuityUPLC TM BEH C18 column (1.7 µm, 100 mm × 2.1 mm i.d.). During the analysis, 1.7 µL of sample was injected, the column was kept at 35 • C and the flow rate was 0.2 mL/min using 2% acetic acid as solvent A and 1:1 methanol/acetonitrile as solvent B. The elution program was as follows 80% The UPLC system was coupled to a TUV detector AcQuity UPLCTM and a TQDTM mass spectrometer equipped with a Z-spray electrospray ionization (Waters, Milford, MA, USA). The software used was MassLynx 4.1. Ionization was performed by electro-spray (ESI) in the negative mode and the data was collected in the selected ion recording mode (SIR) mode.
For the detection and identification of phenolic compounds present in olive oils, a phenolic sample was infused at a rate of 50 µL/min for 10 min. Nitrogen was used as the nebulizing and desolvation gas. The MS conditions were capillary potential 3.0 kV, extractor cone voltage 30 V, RF lens voltage 0.1 V, source temperature 130 • C, desolvation temperature 300 • C and desolvation gas flow rate 500 L/h. Data were collected by use of the MassLynx 4.0 software resident in a personal computer. In order to confirm the identity of the compounds MRM transitions parameters were optimized by using the Intellistart tool of the MassLynx software, which consisted in automatically detecting the major fragments and optimizing cone voltages and collision energies.

Results
The phenolic profile of indigenous Maltese EVOOs (Bidni, Malti) and foreign but locally grown EVOOs is presented, where a summary of the peaks identified using SPE followed by HPLC analysis at 280 nm and 320 nm and MS are presented in Table 1 (see also Supplementary Material Figure S1 and Table S2). A total of 32 peaks were observed where p-coumaric acid (peak 6), Tyrosol acetate (peak 10), DHPEA_EDA (peak 15), oleocanthal (peak 17) and an unknown, possibly an aldehydic form of oleuropein glycone (peak 28) were identified in EVOOs derived from indigenous or locally grown but foreign trees. Conversely, phenolic acids such as p-hydroxybenzoic acid, o-coumaric acid and syringic acid were present either at very low or undetectable concentrations. With respect to the Italian EVOOs luteolin (peak 20), followed by pinoresinol (peak 19) and an unknown, possibly decarboxymethyl-ligstroside aglycone (peak 21) featured prominently, while in the French EVOOs an unknown compound possibly 10-hydroxy-oleuropein aglycone (peak 26) and another unknown compound, possibly an isomer of the former compound (peak 27) were distinguished. Pinoresinol (peak 19) and the oleuropein aglycone 3,4 DHPEA-EA were noticeable in the Greek EVOOs at 320 nm. Similarly peak 23 identified as a possible hydroxy-decarboxymethyl-ligstroside aglycone was observed in the Spanish EVOOs. A univariate statistical analysis on the distinct peak areas showed that EVOOs derived from locally grown cultivars and indigenous cultivars had a significantly higher concentration of the polyphenol p-coumaric acid (peak 6, significant at the 95% confidence level) when compared to EVOOs derived from other Mediterranean countries (Table 2). Likewise, application of an ANOVA and Tukey post hoc hypothesis testing for the analysis of variance on the peak areas of tyrosol acetate (peak 10) showed a significantly higher level in EVOOs derived from indigenous or historically acclimatized cultivars such as Bidni, Malti and Bajda, when compared to EVOOs derived from foreign varieties. Similar results were also obtained on comparing DHPEA-EDA (peak 15), oleocanthal (p-HPEA-EDA) (peak 17) and an aldehydic form of oleuropein aglycone (peak 28). Superscript letters in the same column represent statistically distinct homogeneous subsets as determined by an ANOVA and post hoc Tukey analysis at a 5% confidence level, same letters followed by an * indicate homogeneous subsets at a 10% confidence level.

Discussion
In this paper a first description of the phenolic profile of Maltese derived EVOOs using SPE and HPLC is presented. Phenolic profiles of Maltese EVOOs (both indigenous and foreign but locally grown) derived during the three harvest seasons of 2014-2016 were studied. 3,4 DHPEA-EDA, and oleocanthal (pHPEA-EDA) were markedly present in the indigenous EVOOs rather than in the foreign varieties. As these EVOOs were extracted using the same mechanical and recovery processes, the presence of these compounds appears to reflect a marked occurrence in the indigenous EVOOs. The tyrosol/hydroxytyrosol ratio was observed to be similar to EVOOs from other Mediterranean countries for other varieties [25]. The presence of these phenols is attributed to the partial hydrolysis of their derivatives, while the higher antioxidant activity of hydroxytyrosol when compared to tyrosol make it more susceptible to degradation [26]. The presence of pinoresinol was not noticeable, while trans-cinnamic acid was observed to co-elute with 1-acetoxypinoresinol. These compounds are the main components of the phenolic fraction, which is derived from olive seed and are noticeably absent from the drupe pulp [27]. Hence their presence in oil reflects crushing and admixing of the olive pits during the crushing process rather than inherent variations in the drupe and hence in the EVOO.
Direct comparison with the foreign EVOOs is difficult either because the phenolic profile for a particular EVOO is not available in the published literature or due to differences in the extraction processes, separation methods and elution profiles used. For example the UV detection in this study was carried out at 280 nm and 320 nm in line with the absorption of polyphenols. In this study the Italian EVOOs were rich in the flavonoid luteolin that is known to increase with fruit ripening and pinoresinol but these latter phenolic compounds were negligibly present in local EVOOs. The relatively low levels of oleuropein in all EVOOs tested are correlated to the action of β-glucosidases that result in the formation of secoiridoid aglycones.
The results confirm the organoleptic qualities reported for certain Maltese EVOOs. The presence of the phenolic compound in Table 1 such as oleocanthal supports the known pungent, burning sensation associated with the Bidni variety [23].
Notably the HPLC chromatograms of foreign and indigenous/local EVOOs (Figure 1) showed diversity in terms of a quantitative variation in their phenolic profile rather than a qualitative one (for example tyrosol acetate, p-coumaric acid). Variations in polyphenol expression may arise due to the harvesting period and ripeness, which reportedly affects an increase in hydroxytyrosol, tyrosol and luteolin, the extraction system [28], oxygen concentration and temperature during the malaxation process [29][30][31][32], removal of stones [33] storage conditions [34] and not least irrigation regimes [35]. However there are important factors that affect the phenolic fingerprint and are independent of processing, such as the combination of climate and pedological factors [36] as well as cultivar genotype. Compounds such as 3,4 DHPEA-EA, the aldehydic form of oleuropein aglycone and particularly DHPEA-EDA are produced by the activation of endogenous β-glucosidases the expression of which reflects cultivar genotype. These compounds that have been ascribed protective functions and contribute to EVOO bitterness [37][38][39] were markedly present in both indigenous and foreign but local varieties. As these phenolic compounds were recorded across indigenous as well as in foreign but locally derived EVOOs (e.g., Carolea) this indicates that the pedo-climatic conditions contributed strongly to the expressed phenolic phenotype, particularly since the ripening stages, i.e., the harvesting period, the conditions of oil extraction (uniquely due to few oil presses) and irrigation treatment based on a low water regime typical for the arid island were uniform [39]. Interestingly low water levels are correlated with oils with increases in DHPEA-EDA and 3,4 DHPEA-EA as well as elevated bitterness and pungency and this may explain their marked presence in indigenous and acclimatized EVOOs [39]. Ultimately the presence of these biochemical markers in the locally derived EVOOs offers an encouraging indication that a tool derived from variable data derived from multiple separation and structural pattern recognition techniques in conjunction with DNA markers may be possible [39][40][41]. This will have important implications on developing methods to ensure authenticity and allowing sustainability contribution of the olive industry to the economy.

Conclusions
This paper achieves a profile based on the separation of polyphenols in EVOOs; for both indigenous and foreign but locally derived EVOOs from the Maltese islands. Evidence that ultimately certain polyphenols were dependent on the pedo-climatic conditions and not solely on the genetic component of the originating olive tree is also presented.

Conclusions
This paper achieves a profile based on the separation of polyphenols in EVOOs; for both indigenous and foreign but locally derived EVOOs from the Maltese islands. Evidence that ultimately certain polyphenols were dependent on the pedo-climatic conditions and not solely on the genetic component of the originating olive tree is also presented.
Supplementary Materials: The following are available online at http://www.mdpi.com/2077-0472/9/5/107/s1, Figure S1: The recorded selected ions pertaining to different compounds, Table S1: The cultivars used in this study and their country of origin, Table S2: MRM method development for the most common minor compounds present in EVOO.

Conflicts of Interest:
The authors declare no conflicts of interest.