*3.3. Gene Expression Analysis*

The expression profiles of the selected genes did not follow the same trend among the cultivars. In most genotypes, the mRNA levels of TYRD were quite low and remained almost constant as the fruit development proceeded (Figure 3A). An exception was represented by cultivars Sivigliana and Gentile dell'Aquila, in which this gene was highly expressed at the first drupe sampling (T1) and then drastically dropped at the second sampling, albeit with no significance. On the contrary, in cultivars Semidana and Tortiglione, TYRD mRNA levels remarkably increased from the second to the third sampling, thus reaching the highest values when drupes were almost dark and mature.

In some cultivars, the CuAO expression pattern did not notably change during fruit maturation, while in other genotypes, such as Semidana and Bambina, a slight increase was observed up to the third sampling time. Interestingly, the CuAO expression level was noteworthy in one of the Apulian cultivars, i.e., Cima di Melfi, not for its general trend but specifically for its abundance during the entire drupe maturation process, since it was extremely and significantly different (higher) in comparison to that detected in all other cultivars (Figure 3B).

In the majority of genotypes, ALDH was almost exclusively expressed during the T3 stage of fruit development, with a remarkable increase starting at the turning phase of drupes, in particular in the cases of Gentile dell'Aquila and Cima di Melfi, and with a less pronounced increment observed in the cases of Bambina, Semidana and Sivigliana. On the contrary, ALDH expression levels in the remaining cultivars were very low (Figure 3C).

**Figure 3.** Expression profiles of the analyzed genes involved in the first step of the phenol biosynthetic pathway during fruit development (T1 = yellow-green drupes, T2 = turning drupes, T3 = almost dark drupes). The mRNA levels were determined by qRT-PCR and relatively expressed as ΔΔCt. (**A**) TYRD = tyrosine/dopa decarboxylase; (**B**) CuAO = copper amine oxidase; (**C**) ALDH = alcohol dehydrogenase. Data are means (*n* = 3) ± standard deviation. Different letters indicate significant differences according to one-way ANOVA followed by Fisher's LSD post-hoc test (*p* = 0.05).

#### **4. Discussion**

The identification of minor olive cultivars with favorable agronomic and nutraceutical features among the available and still uncharacterized Italian germplasm, and their fine characterization with different approaches, could effectively contribute to both the valorization and the spread of these cultivars worldwide. The possibility of their introduction in the olive-growing sector is reinforced by the availability of a molecular fingerprint, that can provide a crucial tool for the identification of these genotypes and the traceability of their oils. The use of microsatellite markers in this study has allowed not only a fine characterization of the considered genotypes, but also clarification of the phylogenetic relationships among them and assessment of an identification key. The Unweighted Neighbor Joining dendrogram mostly separated the genotypes according to their geographic origin as reported in many other studies about the evaluation of the genetic diversity of the Italian olive germplasm [5,64]. Surprisingly, only one of the analyzed genotypes has escaped this kind of clustering; indeed, the Gentile dell'Aquila cultivar, belonging to the autochthonous cultivars of Abruzzo region, grouped with all the Sardinian genotypes instead of with the other cultivars from Central Italy (Figure 1). The explanation of this result can be traced back to the complex migration history of the olive. In 2015, Dìez and colleagues [65] demonstrated the existence of three different gene pools (Q1, Q2 and Q3) in the Mediterranean area and the presence of a broad 'mosaic' group as a mixture of Q1 and Q2. The Italian cultivars considered in that research mainly belonged to the Q2 group. Anyway, the authors postulated that the wide occurrence of the 'mosaic' cultivars might be indicative of admixture events. Thereby, both the outcrossing nature of *O. europaea* ssp. *europaea* as well as the exchanges related to human migration over the centuries [64,66] could explain a certain level of genetic admixture between Gentile dell'Aquila and the Sardinian cultivars or their origin from a common wild ancestor.

Over the last decades, virgin olive oil (VOO) became the symbol of good nutritional habits. This role has been mostly linked to some features such as the high content of MUFA and of phenolic compounds, which allow us to recognize VOO as a functional food [67]. As proof of that, the European Food Safety Authority (EFSA) has permitted the declaration "Olive oil polyphenols contribute to the protection of blood lipids from oxidative stress" when olive oil contains a minimum of "5 mg hydroxy-tyrosol and its derivatives per 20 g of olive oil" [68,69]. However, VOO phenolic content not only plays an important nutritional role but also has a great impact on the organoleptic features and on the stability of the product [60]. Given the wide spectrum of features directly linked to olive phenolics, it is clear how important the understanding of their synthesis and regulation is.

Polyphenol accumulation has been reported to be a complex process considerably varying among genotypes, tissues, developmental stages and in response to different agronomic and environmental conditions [70–72]. In accordance with other studies, there was a component of the variability of the phenolic composition observed among the analyzed monovarietal virgin oils that mostly depended on the genotype. On average, the Apulian cultivars generally showed the highest values of the detected compounds at the complete maturation stage of drupes, followed by the Sardinian cultivars. This could also be related to the different cultivation areas and environmental temperatures of the considered groves, since the climates of Apulia and Sardinia are generally similar to each other and usually milder than that of Central Italy (Abruzzo). As reported by [73], the pathways related to olive phenol compounds could be differentially modulated on the basis of altitude and temperatures, phenols biosynthesis being prolonged in temperate areas.

Quantitative differences in phenolic compounds were highlighted both among the analyzed cultivars and among fruits of the same cultivar at different ripening stages. With regard to total phenols (TPC), a sharp decrease is generally and mainly observed between June and September, as result of the oleuropein hydrolysis during drupe maturation (high β-glucosidase activity) [74,75]. Indeed, the very early stages of drupe formation are dominated by secoiridoids (above all oleuropein), while in the last ripening stages, simple phenols and flavonoids become the major components. In our study, the drupe collecting period between October and December did not allow us to highlight such remarkable differences in total phenol content, this period being characterized by a higher stability of phenol metabolism (Table 3; Figure 2D). Thereby, with the only exception of cultivars Tortiglione and Oliva Rossa, that showed a sharp decrease from T2 to T3 stage, the variation of TPC was minimal in all cultivars during the fruit maturation period considered in this study. An intermediate behavior was registered in oils obtained by Sardinian cultivars, that showed a slight, but significant, decrease only at the overripe (T4) stage.

On the contrary, the content of single compounds was very variable among cultivars and sampling times. Secoiridoid derivatives, i.e., aglycon forms of the secoiridoid glucosides usually formed during oil extraction by the enzymatic hydrolysis of oleuropein, demethyloleuropein and ligstroside [76], were reported to be the most abundant phenols in monovarietal olive oils [72]. Among those, oleacein (3,4-DHPEA-EDA) is an important compound, structurally similar to oleocanthal and with comparable pharmacological properties [77]. In our study, the oleacein levels considerably differed among all genotypes, progressively decreasing as fruit maturation proceeded (Table 3; Figure 2C). Indeed, the higher values were found at the T2 time-point, although with remarkable differences among cultivars (in the range between 2 and 63 mg/kg). As a result of an opposite behavior to this general trend, oils obtained by fully ripe drupes (T4) of the Apulian cultivar Cima di Melfi and of the Abruzzo cultivar Gentile dell'Aquila were those with the highest oleacein concentration. As already mentioned, variation in the VOOs' phenolic composition can be ascribable to several factors, including the technological steps of olive oil extraction, which is far from being merely a mechanical process. Indeed, starting from drupe crushing, several enzymes are set free whose activities are strongly influenced by the genotype, the maturity degree and the environmental conditions (time, temperature, atmosphere) [78,79]. Moreover, secoiridoid derivatives, such as 3,4-DHPEA-EDA, are among the most affected olive phenols during the extraction process. For example, in Tuscan olive oils, oleacein was found only a long time after milling, likely due to the activity of different esterases [80]. Thus, linearity between drupe genetic expression and VOO phenols could be difficult to see and the identification of clear markers of maturity is a complex task [81].

The content of minor phenols such as tyrosol and hydroxy-tyrosol was also evaluated in the considered genotypes, due to their relevance as bioactive molecules with beneficial properties on human health [16,82]. Hydroxy-tyrosol is a simple alcohol conjugated to form oleuropein derivatives, usually highly expressed in young olive fruits [42,43]. In accordance with other reports, the hydroxy-tyrosol amount decreased from semi-green to nearly ripe drupes (with the exception of Corsicana, Oliva Rossa and Gentile dell'Aquila), but then an interesting increase was again observed in the final developmental stage, i.e., in totally black fruits (Table 3; Figure 2B). This result is generally attributed to oleuropein catabolism rather than to new hydroxy-tyrosol biosynthesis, as a consequence of the recycling of all the valuable molecules of complex compounds and their conversion to simple phenols [42,44,83]. Moreover, the oleuropein catabolism in the mature fruit is necessary to reduce the bitter taste typically associated with this compound.

Changes in tyrosol content with ripening did not follow a clear pattern among the studied varieties. In some cases (such as Semidana and Cima di Melfi) an opposite trend with the oleacein content was found. These variations could be supposed to be linked to the genetic expression of the enzymes in the biosynthetic pathway (Figure 3). Indeed, tyrosol could be initially synthesized—and mostly found—in free form, while subsequently it is used as a precursor for oleacein production (Table 3; Figure 2A). However, as already reported [81], finding a straightforward relationship between gene expression and phenotype is very difficult, as the interaction with agronomic and pedoclimatic conditions

should be also taken into account. Moreover, the content of phenolic compounds depends also on the activity of other enzymes such as the olive *β*-glucosidases (not considered in this study) which strongly affect their release in the oily phase. Different patterns of tyrosol accumulation with respect to the cultivar and the maturity degree have been already observed by other authors [84].

When this study started, the exact polyphenol biosynthetic pathway was still not clarified in detail; therefore, among genes annotated in phenol biosynthesis, for expression analysis our choice fell on three genes (TYRD, CuAO and ALDH) on the basis of the involvement of the corresponding encoded enzymes in the first steps of this biosynthetic pathway and in relation to the compounds detected in VOOs under characterization. Tyrosine is referred to as the precursor of the secoiridoid class [85]; in particular, the biosynthesis of tyrosol and hydroxy-tyrosol has been proposed to proceed through the formation of tyramine or, otherwise, it can follow the path including L-DOPA and dopamine as intermediates (Figure 4). Initially, an enzyme similar to a tyrosine/DOPA decarboxylase (TYRD) was recognized to be responsible for the conversion of both tyrosine in tyramine and L-DOPA in dopamine [40,86] (Figure 4A). Recently, these two conversions have been clarified to take place by the action of two different enzymes, i.e., a tyrosine decarboxylase (TDC) in the first way and a DOPA decarboxylase (DCC) in the parallel path (Figure 4B) [44]. Moreover, in the previous proposed version, the subsequent steps leading to the conversion of tyramine to tyrosol and dopamine to hydroxy-tyrosol required two enzymes called CuAO and ALDH [87]. Now, the presence of other intermediates in both paths and the action of an additional enzyme, that is, a phenylacetaldehyde reductase (PAR), have been recognized (Figure 4B) [44]. The reaction of the conversion of tyrosol to hydroxy-tyrosol and vice versa seems to involve a polyphenol oxidase (PPO) [42]. Since most of the genes responsible for secoiridoid formation have been identified in a restricted number of species, some steps of the biosynthesis of oleuropein are still unclear, even though the pathway from geranyl diphosphate to secologanin has been elucidated [42,44].

The relative expression of the selected genes was monitored during three stages of fruit maturation, at T1, T2 and T3 sampling times. The expression analysis showed the general tendency of these mRNA levels to decrease or remain more or less stable in semi-green fruits and then to increase, slightly or significantly in a few cases, in semi-dark drupes (Figure 3). These results are in line with the biochemical characterization of monovarietal oils at T2 and T3 stages, which proved that at these time-points the tyrosol and hydroxytyrosol contents were progressively diminishing, following the same pattern as the genetic expression data. Some exceptions to this trend have been surprisingly observed for few varieties and in particular for TYRD and ALDH genes. Indeed, these two genes were highly expressed, respectively, in the Tortiglione and Semidana genotypes, and in Cima di Melfi and Gentile dell'Aquila, with a remarkable increase in their transcript levels from T2 to T3 time-points. These unexpected but interesting results, not yet reported in other olive cultivars, to our knowledge, certainly require further investigation. In order to clarify the role of TYRD and ALDH genes in these cultivars, their expression profile should be investigated at more developmental stages of drupes during more consecutive years. Protein expression experiments as well as enzymatic activity assays could also help us to understand the involvement of these genes in polyphenol metabolism of these minor cultivars. At the moment, we could hypothesize that these genotypes are effectively characterized by a high level of gene expression even towards the end of drupe maturation, contrarily to that reported in this study and, more in general, in the literature. It is also plausible that the corresponding enzymes are conveyed to other metabolisms or that some post-translational control takes place and makes these enzymes unfunctional. Exceptions apart, the general agreement of qRT-PCR results with the biochemical outcomes confirmed that hydroxy-tyrosol biosynthesis is under a transcriptional control, as also reported by other authors [40,42–44].

**Figure 4.** Simplified representation of the biosynthetic pathway of some important olive polyphenols. The previously identified enzymes and reactions (**A**) as well as the recently identified ones, with additional steps in the biosynthesis (**B**), are reported. Abbreviations: 3,4-DHPEA-EDA: elenolic acid linked to 3,4-dihydroxyphenyl ethanol (3,4-DHPEA); 4-HPA: 4-hydroxyphenylacetic acid; 3,4-DHPA: 3,4-dihydroxyphthalic acid; TYRD: tyrosine/DOPA decarboxylase; PPO: polyphenol oxidase; CuAO: copper amine oxidase; ALDH: alcohol dehydrogenase; DCC: DOPA decarboxylase; PAR: phenylacetaldehyde reductase. Adapted from [40,44].

#### **5. Conclusions**

Unveiling the genetic potentialities of minor cultivars/genotypes could open the doors to their valorization. Besides the safeguarding of local varieties, the information provided in this work could be useful for breeding programs with selection of genotypes that, while being resilient and better adapted to specific environmental conditions, could also exploit their potentiality in terms of accumulation of bioactive compounds with important consequences on the nutritional, sensory and stability properties of VOO. With a view to an eco-sustainable and low-input agriculture, the introduction of minor cultivars may offer interesting features that, together with the extraction technology, could address the needs of a worldwide market with growing interest towards VOOs. Thus, the results and information obtained in this study are encouraging for the valorization of the still poorly explored Italian germplasm, with important both economic and scientific consequences.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/foods10061371/s1, Figure S1: Example of SSR electropherograms, Table S1: Microsatellite markers used in this study, Table S2: List of primers utilized for qRT-PCR analysis.

**Author Contributions:** Conceptualization, W.S., C.M. and F.C.; investigation, S.G., I.M., G.S., F.F., P.C. and A.P.; writing—original draft preparation, W.S.; writing—review and editing, W.S., G.S., F.F., C.D.D.M., P.C., A.P., C.M. and F.C.; supervision, W.S., C.M. and F.C.; funding acquisition, F.C., W.S. and C.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was mainly funded by the Italian AGER 2 Project, grant no. 2016-0105, and it was in part supported by the CNR project NUTR-AGE (FOE-2019, DSB.AD004.271) and the Apulia Region Project "Approcci di Next Generation Sequencing per l'analisi di variabilità e di espressione genica in genotipi di olivo autoctoni pugliesi" L.R. 28 December 2018 n. 67, art. 37.

**Conflicts of Interest:** The authors declare that they have no competing financial or personal interests that could influence the work reported in this paper.

#### **References**

