Evaluation and Origin Discrimination of Two Monocultivar Extra Virgin Olive Oils, Cultivated in the Coastline Part of North-Western Greece

: Extra virgin olive oil (EVOO) quality and authenticity are important and challenging factors nowadays for the assurance of consumers’ protection, prevention of unfair competition, and disruption of the national economy by a false declaration of origin. Hence, the recognition of EVOO authenticity is of great interest in terms of commercial and quality aspects. The objective of this study was to evaluate and discriminate monovarietal extra virgin olive oils of the two dominant olive cultivars, Lianolia Kerkyras and Koroneiki, produced in the coastline part of Western Greece, based on their chemical characteristics, followed by statistical and chemometric analysis in order to proﬁle for the ﬁrst time the typical characteristics of Lianolia Kerkyras as well as to identify possible markers for authenticity purpose. A total of 104 olive oil samples were collected. Both cultivars had an overall high quality proﬁle as far as their basic qualitative parameters (free fatty acid, peroxide value, and UV spectrometric indices) are concerned. A higher concentration in the mono-unsaturated oleic acid characterize olive oils of cv. Koroneiki compared to cv. Lianolia Kerkyras, while a clearly higher concentration in the poly-unsaturated linoleic acid was observed in olive oils of cv. Lianolia Kerkyras. In addition, olive oil samples of cv. Koroneiki showed a clear lower total sterols concentration with a percentage of 40.9% not surpassing the required EU Regulatory limit of 1000 mg / kg, an observation which strengthens previous published results of our research group and depicts an overall “intrinsic characteristic” of cv. Koroneiki. As far as the proﬁle of the individual sterols is concerned, Lianolia Kerkyras samples exhibited higher mean value for the total sterol content as well as for β -sitosterol, the major phytosterol in olive oils, compared to the relative values of Koroneiki. Signiﬁcant di ﬀ erences in the sterolic and fatty acid composition of the examined olive oil samples were shown by means of statistical analysis demonstrating a strong botanical e ﬀ ect and depicting that those compositional markers can be suggested as possible authenticity tools.


Introduction
Extra virgin olive oil (EVOO) is cherished as a fundamental ingredient not only in the Mediterranean diet but also internationally due to its proven health-promoting effects and nutritional

Olive Oil Sampling
One hundred and four (N = 104) virgin olive oil samples were taken during the harvesting period of 2019-2020 from the coastline part of Western Greece. In more detail, sixty (60) olive oil samples of Lianolia Kerkyras and forty four (44) samples of Koroneiki cultivar were originated from the following regional units: Preveza area (at 38 • 57 33.35 N and 20 • 25 06.18 E)Parga area (at 39 • 17 07.20 N and 20 • 24 01.80 E) and Thesprotia area (at 39 • 29 36.60 N and 20 • 22 42.76 E) according to Google Earth version 7.0.1.8244-beta, Google, Inc., USA). Similar climatic conditions were evident at all regions, characterized mainly by mild climate with temperatures rarely falling below freezing. Summer months are hot, and the weather is the typical Mediterranean weather with not so much rain but lots of sunlight. Olive fruits were derived from conventional agriculture and harvested between mid-November to February according to specific technical details provided by the responsible agriculturists in the framework of the project. Olive fruits were transferred to local oil mills, equipped with 2-phase centrifugal systems (decanters) in all cases and the olive oil was extracted within 24 h by using the same protocol for all samples and as such they were not treated as variables. This involved washing and crushing of olive fruits followed by olive paste malaxation at 27-29 • C for 35-45 min and decanting. The resulting olive oil samples were transferred to 1 L air-tight dark-green glass bottles and stored at 4 • C in the fridge for analysis. All examined parameters were determined in duplicate.

Analysis of the Quality Parameters
Free fatty acid, peroxide value, and UV absorption characteristics (K 232 and K 268 ) were determined as described by the analytical methods of Regulation EEC/2568/91 of the European Commission and later amendments [25].

Analysis of the Sterolic Profile
The oils, with the addition of α-cholestanol as internal standard (Sigma, St. Louis, MO, USA), were saponified with KOH in an ethanolic solution and the unsaponifiable matter was then extracted with diethylether. Separation of the different alcoholic compounds fractions took place by thin-layer chromatography (TLC) on a basic silica gel plate (Fluka, St. Louis, MO, USA). The addition of the appropriate silylation reagent (Alfa Aesar GmbH & Co., Karlsruhe, Germany) made possible the transformation of the recovered fractions into trimethyl-silyl-ethers. Analysis was performed by capillary column gas chromatography as described by EEC/2568/91 regulation, Annexes V [25].

Analysis of the Fatty Acid Profile
The principle of the method is based on the conversion through trans-esterification of the fatty acids into fatty acid methyl esters (FAME) and extraction with n-hexane. Analysis of the individual fatty acids was carried out according to Regulation EEC/2568/91, Annex IV with later amendments.

Statistical Analysis
The findings were presented as mean values ± standard deviation (SD). The use MINITAB 18 software (MINITAB, INC. State College, PA, USA) was applied to analyze the samples. In addition, minimum and maximum values of the samples, mean, and standard deviation (SD) were collected. In the theory of statistics, an indication of statistical dispersion is provided by the difference between the biggest and smallest values (range). Moreover, in this study, the statistical mean differences were evaluated based on the statistical tool analysis of variation (ANOVA). A statistical significance level was set at p < 0.05. The tests of normality and homogeneity for the variances were carried out, and we found that both conditions are satisfied. The principal component analysis (PCA) was also conducted to investigate the association of the two monocultivars with the chemical characteristics analyzed.

Quality Parameters of the Examined Olive Oils
Injuries caused to the olive fruits from insect attacks and fungal diseases, improper olive harvest practices, and extraction methods as well as poor storage conditions of the extracted olive oil are the main factors affecting the quality characteristics of the produced olive oil [26]. Hence, categorization of the collected monovarietal olive oils took place by means of determination of free fatty acids, peroxide value, and spectrophotometric absorption. Table 1 shows that all analyzed samples obtained from the two examined cultivars in the coastline part of Western Greece belong to the highest quality category of "extra virgin olive oil" since they satisfy the specifications set by EU Regulation 2568/91 [25]. More specifically, the mean free fatty acid was 0.24% and 0.27%, respectively for Koroneiki and Lianolia Kerkyras olive oils. Likewise, the mean peroxide value for cv. Koroneiki olive oils was 6.64 meqO 2 kg −1 whereas for cv. Lianolia Kerkyras the mean peroxide value was 5.21 meqO 2 kg −1 . Similarly, both monovarietal olive oils had K 232 and K 268 mean values quite below the limit set by the EU Regulation 2568/91. The results depict that both cultivars had an overall high quality profile in that crop year as far as their basic qualitative parameters is concerned. 0.14 ± 0.01 0.11-0.19 0.14 ± 0.02 0.11-0.21

Fatty Acid Profile of the Two Monocultivar Olive Oils
Fatty acid profile plays an important role in the quality and characterization of an olive oil as its composition reflects the nutritional properties of an olive oil [27]. Several researchers have reported that among other major components, fatty acids composition seems to represent a possible tool for varietal characterization and authentication [28][29][30][31][32][33]. Table 2 shows the mean fatty acid composition of the analyzed monovarietal olive oils. As it is shown, all fatty acids identified were found in the normal range expected for the extra virgin olive oil category for both monocultivars. With respect to the mono-unsaturated oleic acid (C18:1), olive oils of cv. Koroneiki presented a higher concentration with a mean value of 75.07% compared to cv. Lianolia Kerkyras (69.55%). Moreover, the saturated stearic acid (C18:0) concentration was higher for cv. Koroneiki with a mean value of 2.51% compared to the concentration of 2.04% for cv. Lianolia Kerkyras. On the other hand, olive oils of cv. Lianolia Kerkyras presented a clearly higher concentration of the poly-unsaturated linoleic acid (C18:2) with a mean value of 10.40% compared to the Koroneiki olive oils (6.43%) as well as a higher concentration in palmitic acid (14.46%). Regarding the comparison of the contents of the different lipids in the two varieties, the different flesh to stone ratio is characteristic of the olives themselves. The flesh to stone ratio is 3-4 to 1 in cv. Lianolia Kerkyras and 5 to 1 in cv. Koroneiki, and the stone is relatively large in the latter, according to the catalogue of "Apulian and Greek Olive Varieties". This anatomical characteristic combined with the fact that the olive seeds are richer in linoleic acid and beta-sitosterol than the olive pulp [34] could explain the higher content of these components in cv. Lianolia Kerkyras. Moreover, the discrimination of the two varieties can also be aided and explained by the 18:1/18:2 ratio which indicates that the ratio in cv. Koroneiki is double that of cv. Lianolia Kerkyras. Values are expressed as means ± standard deviation (SD). n.s = not-significant. The p < 0.05 was set at the level of statistical significance.
Variability in fatty acid composition between the two monocultivar olive oil samples led to the performance of an analysis of variance (ANOVA) in order to assess their differences. Table 2  The effect of cultivar in the fatty acid profile of the two monocultivar olive oils originating from the coastline region of north-western Greece is depicted from the results of this work proving the usefulness of fatty acid composition for this varietal discrimination. Those results are in agreement with our previously published data as well as other relevant studies, demonstrating that fatty acid profile plays a crucial role in the classification of virgin olive oils according to their cultivar [28][29][30][31][32][33]. Finally, according to our findings, this local Greek olive variety has the tendency of exhibiting higher concentrations in the poly-unsaturated omega-6 linoleic acid and lower concentration in the mono-unsaturated omega-9 oleic acid compared to the most systematically cultivated variety of cv. Koroneiki. Of course, further in-depth research for more crop years is necessary for the adequate characterization of the fatty acid profile of Lianolia Kerkyras olive oils.

Sterolic Profile of the Two Monocultivar Olive Oils
Phytosterols and triterpenic dialcohols belong to the unsaponifiable fraction of olive oil and constitute one of its minor components with an important health beneficial impact [35][36][37]. Many researchers have shown that each variety has a characteristic sterol "fingerprint", revealing that the sterolic profile can be used as a reliable indicator with a high discrimination potential for olive oil classification [38][39][40][41][42][43].
Taking into account the unexplored chemical characteristics of Lianolia Kerkyras, we employed the present study to determine and compare the sterolic profile of the Koroneiki and Lianolia Kekryras olive oils obtained from the coastline region of north-western Greece. The percentage of individual sterols as well as the concentrations of total sterols for the examined monovarietal olive oil samples are presented in Table 3. In general, total sterols concentration and individual sterols content of cv. Lianolia Kerkyras olive oil samples comply with the up to date EU legislation [18]. In contrast, olive oil samples of cv. Koroneiki showed lower concentration in total sterols with a mean value of 1020.8 mg/kg compared to olive oils of cv. Lianolia Kerkyras (1343.7 mg/kg). More precisely, 40.9% of the analyzed samples of cv. Koroneiki did not exceed the defined limit of 1000 mg/kg for total sterols (EEC Regulation 2568/91) as illustrated in Figure 1. This observation is in agreement with results of our previous publication regarding the tendency of low total sterol concentration in Koroneiki olive oils of the southern Peloponnese, enhancing and clearly depicting an "intrinsic characteristic" for Koroneiki cultivar [44].
The European Commission Regulation (EEC 2568/91) imposes limits or ranges for each type of sterol and total sterols based on the natural levels found in traditional olive oil types [25]. Sterol profiles outside these limits, in combination with other chemical parameters, could theoretically suggest that the oil is not authentic. However, a number of cases have found olive oils which naturally exceed or subceed the limits for sterols [45,46]. Analysis of virgin olive oils from cv Koroneiki in completely different geographical regions such as Crete and Australia has also shown a tendency of low total sterol concentration [47,48]. Thus, low mean value in the concentration of total sterols may depict a "special and intrinsic characteristic" for Koroneiki cultivar in general, which has to do with the cultivar itself [44].
results of our previous publication regarding the tendency of low total sterol concentration in Koroneiki olive oils of the southern Peloponnese, enhancing and clearly depicting an "intrinsic characteristic" for Koroneiki cultivar [44].
The European Commission Regulation (EEC 2568/91) imposes limits or ranges for each type of sterol and total sterols based on the natural levels found in traditional olive oil types [25]. Sterol profiles outside these limits, in combination with other chemical parameters, could theoretically suggest that the oil is not authentic. However, a number of cases have found olive oils which naturally exceed or subceed the limits for sterols [45,46]. Analysis of virgin olive oils from cv Koroneiki in completely different geographical regions such as Crete and Australia has also shown a tendency of low total sterol concentration [47,48]. Thus, low mean value in the concentration of total sterols may depict a "special and intrinsic characteristic" for Koroneiki cultivar in general, which has to do with the cultivar itself [44].  Values are expressed as means ± standard deviation (SD). n.s = not-significant. The p < 0.05 was set at the level of statistical significance.
With respect to the profile of the individual sterols, Lianolia Kerkyras olive oils samples showed a higher mean value for the major phytosterol β-sitosterol (89.21%) and for sitostanol (0.69%) compared to the relative values of Koroneiki olive oil samples (Table 3). Moreover, Lianolia Kerkyras exhibited lower mean values for the most abundant sterols, namely Δ-5-avenasterol (4.31%),  With respect to the profile of the individual sterols, Lianolia Kerkyras olive oils samples showed a higher mean value for the major phytosterol β-sitosterol (89.21%) and for sitostanol (0.69%) compared to the relative values of Koroneiki olive oil samples (Table 3). Moreover, Lianolia Kerkyras exhibited lower mean values for the most abundant sterols, namely ∆-5-avenasterol (4.31%), campesterol (3.48%), stigmasterol (0.49%), as well as for total erythodiol content (1.43%). Although no previous reported data for the sterolic profile of cv. Lianolia Kerkyras is available, to the best of our knowledge, our results depict that this local Greek olive variety presents higher percentage mean values in β-sitosterol and total sterols compared to the most-known and exploited Koroneiki variety. The health-promoting aspects of β-sitosterol have been reported in several studies, recognizing its great hypocholesterolemic function in opposing the intestinal absorption of cholesterol and therefore contributing to the prevention of many diseases, including various types of cancer (colon, prostate, and breast) [49,50].
Comparison of the two monocultivar olive oils according to their sterolic profile, as shown in Table 3, by means of the calculated p-value, shows it to be in most cases close to 0.00 (p ≈ 0.00), indicating a strong botanical effect. Thus, the dataset of individual and total sterols can enable the classification of the examined olive oils according to their cultivar and indicate them as a possible compositional marker for olive oil authentication. Future in-depth research by comparing the sterolic profile of olive oils derived from more olive cultivars would be advisable.

Chemometric Analysis
Principal component analysis (PCA) was carried out for the confirmation and strengthening of the classification of the examined monovarietal olive oils according to the cultivar. In order to restrict initial variables to a small number of new variables (known as principal components), the principal component analysis (PCA) is used to describe most of the original variations. The main goal of the key factor analysis is to identify the associated variables. The PCA score plot of Koroneiki versus Lianolia Kerkyras olive oils, according to their fatty acid and sterolic data set, is shown in Figure 2a,b respectively. In more detail, as shown in Figure 2a, the majority of K-points (corresponding to cv. Koroneiki) point to the left part of PC1, indicating that K has significant negative loads on component 2. Moreover, it can be observed that L-points (corresponding to cv. Lianolia Kerkyras) are shown on the right part of PC1, meaning that L has significant positive loads on component 1. Therefore, both K and L regions are independent and the data collection on fatty acids is not similar, impacting the two main regions. In accordance, the PCA score plot of Koroneiki vs. Lianolia Kerkyras olive oils' sterolic profile resulted in the creation of two separate clusters as shown in Figure 2b. In that case, we noted that L points are grouped very close to each other, depicting that the variability among Lianolia Kerkyras olive oils is very small compared to that of Koroneiki samples. A low variability for L points reveals that the samples appear to be very similar to the mean for the Lianolia Kerkyras, and therefore are not affected by any external factors.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 7 of 11 campesterol (3.48%), stigmasterol (0.49%), as well as for total erythodiol content (1.43%). Although no previous reported data for the sterolic profile of cv. Lianolia Kerkyras is available, to the best of our knowledge, our results depict that this local Greek olive variety presents higher percentage mean values in β-sitosterol and total sterols compared to the most-known and exploited Koroneiki variety. The health-promoting aspects of β-sitosterol have been reported in several studies, recognizing its great hypocholesterolemic function in opposing the intestinal absorption of cholesterol and therefore contributing to the prevention of many diseases, including various types of cancer (colon, prostate, and breast) [49,50]. Comparison of the two monocultivar olive oils according to their sterolic profile, as shown in Table 3, by means of the calculated p-value, shows it to be in most cases close to 0.00 (p ≈ 0.00), indicating a strong botanical effect. Thus, the dataset of individual and total sterols can enable the classification of the examined olive oils according to their cultivar and indicate them as a possible compositional marker for olive oil authentication. Future in-depth research by comparing the sterolic profile of olive oils derived from more olive cultivars would be advisable.

Chemometric Analysis
Principal component analysis (PCA) was carried out for the confirmation and strengthening of the classification of the examined monovarietal olive oils according to the cultivar. In order to restrict initial variables to a small number of new variables (known as principal components), the principal component analysis (PCA) is used to describe most of the original variations. The main goal of the key factor analysis is to identify the associated variables. The PCA score plot of Koroneiki versus Lianolia Kerkyras olive oils, according to their fatty acid and sterolic data set, is shown in Figure 2a,b respectively. In more detail, as shown in Figure 2a, the majority of K-points (corresponding to cv. Koroneiki) point to the left part of PC1, indicating that K has significant negative loads on component 2. Moreover, it can be observed that L-points (corresponding to cv. Lianolia Kerkyras) are shown on the right part of PC1, meaning that L has significant positive loads on component 1. Therefore, both K and L regions are independent and the data collection on fatty acids is not similar, impacting the two main regions. In accordance, the PCA score plot of Koroneiki vs. Lianolia Kerkyras olive oils' sterolic profile resulted in the creation of two separate clusters as shown in Figure 2b. In that case, we noted that L points are grouped very close to each other, depicting that the variability among Lianolia Kerkyras olive oils is very small compared to that of Koroneiki samples. A low variability for L points reveals that the samples appear to be very similar to the mean for the Lianolia Kerkyras, and therefore are not affected by any external factors. In general, a discrete separation between the two cultivars was detected by applying the PCA algorithm to the data set of fatty acids and sterols. The obtained results are in accordance with other In general, a discrete separation between the two cultivars was detected by applying the PCA algorithm to the data set of fatty acids and sterols. The obtained results are in accordance with other relevant studies concentrating on olive oil major and minor compounds as effective tools for studying olive oil authentication. For example, fatty acid and triglyceride composition data have shown significant potential for olive oil classification according to cultivar [29][30][31][32]. Likewise, clear differences have been observed in the content of the fatty acid as well as phenolic content of Tunisian olive varieties [33]. In addition, according to Lorenzo et al., botanical discrimination of olive oils can be achieved by examining the variables of stearic acid, campesterol, total sterols, and oxidative stability [42]. Another research group has recently used both sterol and phenolic fingerprints to discriminate Tunisian and Italian EVOOs, outlining their potential for authenticity evaluations [43].
Finally, a combined PCA was performed using the variables of the fatty acid compositional and sterolic data (a total of 28 variables in 104 observations). As shown in the score plot of Figure 3, a complete separation according to the cultivar was achieved. Hence, it can be concluded that the discrimination of cv. Koroneiki and Lianolia Kerkyras samples in terms of cultivar could occur based on both fatty acid and sterolic profile data.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 11 relevant studies concentrating on olive oil major and minor compounds as effective tools for studying olive oil authentication. For example, fatty acid and triglyceride composition data have shown significant potential for olive oil classification according to cultivar [29][30][31][32]. Likewise, clear differences have been observed in the content of the fatty acid as well as phenolic content of Tunisian olive varieties [33]. In addition, according to Lorenzo et al., botanical discrimination of olive oils can be achieved by examining the variables of stearic acid, campesterol, total sterols, and oxidative stability [42]. Another research group has recently used both sterol and phenolic fingerprints to discriminate Tunisian and Italian EVOOs, outlining their potential for authenticity evaluations [43]. Finally, a combined PCA was performed using the variables of the fatty acid compositional and sterolic data (a total of 28 variables in 104 observations). As shown in the score plot of Figure 3, a complete separation according to the cultivar was achieved. Hence, it can be concluded that the discrimination of cv. Koroneiki and Lianolia Kerkyras samples in terms of cultivar could occur based on both fatty acid and sterolic profile data.

Conclusion
In the present study, we evaluated and profiled for the first time the chemical characteristics of Lianolia Kerkyras olive oils, a local Greek olive cultivar cultivated exclusively in the north-western coastline part of Greece as well as compared its chemical properties with those of the most wellknown Greek olive cultivar, cv. Koroneiki. The high differentiation potential of sterols and fatty acid compositional data, as efficient authenticity tools for origin discrimination, was confirmed by employment of statistical analysis tools. The study and discrimination among local olive varieties is particularly important in order to preserve biodiversity and maintain the advantages of local varieties so as to promote and strengthen Greek olive sector. Τhe obtained results not only lead to valuable information about the studied monocultivar olive oils, but can also contribute in the future to the establishment of a continuously enriched "Greek Authentic Olive Network" of indigenous, local, and less known and exploited monovarietal olive oils produced in Greece. Further in-depth research in combination with more examined parameters (e.g., sensory analysis, phenolic profile) and pioneering chemometric tools could secure the authenticity, traceability, and therefore higher commercial presence of Greek local olive varieties.

Conclusions
In the present study, we evaluated and profiled for the first time the chemical characteristics of Lianolia Kerkyras olive oils, a local Greek olive cultivar cultivated exclusively in the north-western coastline part of Greece as well as compared its chemical properties with those of the most well-known Greek olive cultivar, cv. Koroneiki. The high differentiation potential of sterols and fatty acid compositional data, as efficient authenticity tools for origin discrimination, was confirmed by employment of statistical analysis tools. The study and discrimination among local olive varieties is particularly important in order to preserve biodiversity and maintain the advantages of local varieties so as to promote and strengthen Greek olive sector. The obtained results not only lead to valuable information about the studied monocultivar olive oils, but can also contribute in the future to the establishment of a continuously enriched "Greek Authentic Olive Network" of indigenous, local, and less known and exploited monovarietal olive oils produced in Greece. Further in-depth research in combination with more examined parameters (e.g., sensory analysis, phenolic profile) and pioneering chemometric tools could secure the authenticity, traceability, and therefore higher commercial presence of Greek local olive varieties.