Molecular Characterization and Ex Situ Conservation of Wild Grapevines Grown in the Area Around the Neolithic Settlement of Dikili Tash, Greece

Abstract

Dikili Tash is a Neolithic settlement that lies next to the ruins of the ancient city of Philippi on the north-eastern part of Greece. A recent archaeological excavation has unearthed charred grapevine pips and pressings together with two-handed clay cups, jugs, and jars that date to 4300 BC. The majority of the pips were found to be Vitis vinifera ssp. sylvestris. Natural populations of this species have been localized in the valley surrounding Dikili Tash and also on Mt Pangaion and Mt Lekani, which flank the valley. Fifty-one samples from these modern populations have been analyzed using microsatellites on twenty microsatellite loci, and a dendrogram has been constructed showing the genetic closeness of the samples analyzed. Cuttings from all the vines analyzed are currently rooted and grown in the Hellenic Agricultural Organization—DIMITRA (ELGO-DIMITRA) greenhouse facilities in Lykovryssi (Athens) with the aim to, eventually, be transplanted in the grapevine, thus establishing the first V. sylvestris ex situ conservation site in Greece.

1. Introduction

Grapevine (Vitis vinifera L.) is cultivated around the globe and provides the primary plant material for the production of wine, a product of immense financial and cultural importance. The species is subdivided into the cultivated grapevine, V. vinifera spp. sativa, and its wild progenitor, V. vinifera ssp. sylvestris (hereafter V. sylvestris or wild grapevines). Wild grapevines are dioecious plants found on an altitude that spans from the sea level to 1200 m; they tend to inhabit sides where water is adequate, such as river banks, and climb on the surrounding vegetation. Their mature leaves are often three- or five-lobed, the bunch is often small, cylindrical, and sparse, while the berries are small, globular, and dark in color, although white-skinned berries have also been reported [1]. Primary domestication of V. vinifera is believed to have occurred in the greater Caucasus area, where archeological evidence supporting the first wine making practices have been discovered [1]; the possibility, however, of multiple domestication centers around the Mediterranean basin represents an alternative possibility that cannot be ruled out [2]; Greece is such a place, where either a primary or secondary domestication may have occurred [2].

Grapevine cultivation and production of wine represents a long standing practice in Greece. Archeological findings, such as heavily burnt grape pips and pressings together with two-handled pottery cups, jugs, and jars dating to 4300 BC, have been unearthed at the Neolithic settlement of Dikili Tash, in the north-eastern part of the country (Figure 1a,b), making the surrounding area the oldest site in Europe that both cultivation of grapevines and production of wine have occurred [3]. Extraordinarily, this discovery came to support the ancient myth according to which on the slopes of Mt Pangaion, god Dionysos, the god of wine and the pleasures that come with its consumption, was raised and partying with his followers, the Meanands and the Satyrs. Eventually, god Dionysos came to be in conflict with king Lykourgos of Edones, the local tribe dwelling in the valley north of Mt Pangaion. As the consequence of this conflict, god Dionysos was expelled from this area migrating to the south of the Balkan Peninsula, where he became one of the most popular gods of the ancient Greek world. Prior to this migration, however, he got his revenge by transforming Dryas, king Lykourgos’s son, to a vine. As the king was meant to prune the vine, he slaughtered his own offspring [4]. According to an interpretation of the myth, god Dionysos and king Lykourgos represent two basic and opposing social forces: on one side, the god stands for liberation and ecstasy, and the danger of losing control of yourself upon consumption of wine, whereas on the other side the king represents the obligatory obedience to State’s laws for the smooth function of the society [4].

Upon the discovery of the grape pips in Dikili Tash, a crucial question was raised as whether these pips were the wild (Vitis vinifera ssp. sylvestris) or the domesticated (V. vinifera ssp. vinifera) form of grapes, since this would imply a greater investment of time and effort by humans cultivating grapevine in contrast to the effortless exploitation of the wild grape type [5]. Recent morphometric analysis of 368 Late Neolithic Dikili Tash pips classified nearly all of the analyzed pips (366 out of 368) as being of the wild type [6]. This finding strongly implies a direct link between the charred grapevine seeds discovered in Dikili Tash and the seeds of modern wild grapevines; however, conclusive evidence could be generated upon the direct molecular analysis of the ancient DNA extracted from the charred archeological material. At the moment, efforts to recover sufficient intact DNA from charred archeological material have not yet provided promising results [7].

The area around Dikili Tash—Mt Pangaion, Mt Lekani, and the valley that lies between them—is a unique place, in the sense that scientific research enlightens and interprets an ancient myth: according to the myth, the god of wine, god Dionysos, was raised on the slopes of Mt Pangaion, whereas modern archeological research first unearthed grape pips and pressings dating to the Neolithic Period, and later found that these grapevine materials are of the wild grapevine type. Our original working hypothesis was whether the wild grapevine still exists in this area.

The only study regarding the distribution and partial characterization of wild grapevines in Greece was published in the early 1960s [8] reporting the occurrence of V. vinifera ssp. sylvestris as part of the native flora in the region that extends from south of Mt Olympus to the river Evros—this area includes the central and eastern parts of Macedonia as well as part of Thrace (Figure 1a). In the Logothetis study [8], the following types of vines were distinguished: (i) six types of male vines (M1: found in the greater area of Mt Olympus; M2: found from Mt Olympus to the river Evros; M3: around the eastern parts of Kerdyllia; M4: around Mt Olympus and also in Chalkidiki; M5: very common from Mt Olympus to the river Evros; M6: from Mt Olympus to the river Evros), (ii) three types of female vines (F1, F2, and F3: in all cases, from Mt Olympus to the river Evros), and (iii) two hermaphrodite vines (H1: found in the eastern parts of Chalkidiki; H2: found in the central parts of Chalkidiki) (Supplementary Data in Figure S1); in addition, the Pafsanias vine [9,10], an extraordinarily old vine grown in the central part of Peloponnese, has also been included for comparison in the Logothetis study (Supplementary Figure S1).

The international bibliography on V. sylvestris has been exponentially increased over the last two decades: from 26 in the 20th century to more than 240 in the 21st century—46 studies from 2020 onwards (https://www.ecpgr.org/working-groups/vitis/sylvestris, as accessed on 29 May 2025). Initially, the focus was on characterization of individuals in natural populations in terms of classical and molecular ampelography; lately, however, such studies have been extended from in-depth studies on genealogical analyses [11], disease resistance [12,13], yeast-associated studies [14], genomics (review: [15]), population genetic studies [16], endophytic diversity [17], and the wild grapevine virome [18] to wine making [19,20], archeobotany [21], and ancient DNA [22].

Here, we report the localization and molecular characterization of natural wild grapevine populations on Mt Pangaion, Mt Lekani, and the in-between Dikili Tash valley. To our knowledge this is the first report on this research topic referring to modern Greece, although these samples have also been incorporated in the Röckel et al. report [16].

2. Materials and Methods

Numerous expeditions from 2017 to 2023 on Mt Pangaion and Mt Lekani as well as in the valley that lies between them identified a number of natural populations of V. sylvestris at an altitude that spans between 150 m to nearly 800 m. Most of the wild vines have been found either by the road or in a close proximity (up to 150 m) in small groups of two to four individuals, although scattered single vines have also been located (Figure 1b; the areas where sampling was performed are shown, while the exact geographic coordinate data for each sample is not shown in order to protect the natural populations from excessive tourist interest). Leaf shape varied greatly among the vines (Figure S2).

Young and healthy leaves of a total of fifty-one individual plants were collected, kept on ice, and finally stored at −20 °C until their further laboratory treatment (within 1–2 months upon collection). Although it was intended to collect only one sample from each vine, due to thick vegetation, more than one sample from the same vine might have been collected. A wild sample that was collected from Phthiotis (a region about 500 km south of the Dikili Tash area; Figure 1a) was also incorporated in the current analysis.

Genomic DNA was extracted from 100 to 130 mg of the frozen leaf tissue using the commercially available NucleoSpin Plant II kit (Macherey–Nagel, Düren, Germany) according to manufacturer’s instructions. The integrity of the extracted genomic DNAs was checked by electrophoresis on agarose gels, while the concentration was estimated by using a Quawell (Q3000 UV–Vis Spectrophotometer, Quawell Technology Inc., San Jose, CA, USA) spectrophotometer.

At the Hellenic Agricultural Organisation DIMITRA (ELGO-DIMITRA) in Lykovryssi (Athens), Polymerase Chain Reactions (PCRs) with nine polymorphic markers (VVS2 [23], VVMD5, VVMD7 [24], VVMD25, VVMD28, VVMD27, VVMD32 [25], VrZAG62, VrZAG79 [26]) that had been proposed by the GrapeGen06 [27] plus VrZAG67 [26], were performed as described in Merkouropoulos et al. [28]. This analysis was also repeated at the Julius Kühn-Institut (JKI)—Institute for Grapevine Breeding Geilweilerhof in Siebeldinge (Germany)—to ensure compatibility between the two laboratories, and was, later, extended using the following markers, VVIN16, VVIN73, VVIP60, VVMD21, VVMD24, VVIB01, VVIH54, VVIQ52, VVIV37, VMC1B11, and VVIP31, as described in Margaryan et al. [29], to enlarge the dataset. Both laboratories used the same references and the same allelic ladders.

Data analysis, sizing, and genotyping were performed using the GeneMapper (version 4.0) software program. The GenAlEx 6.501 software program [30] was used for statistical analyses. A UPGMA dendrogram was constructed using the MEGA4 software program [31].

During the period of March–April of the years 2022 and 2023, the cuttings from all the samples that had been molecularly analyzed were also been planted and rooted in pots, and are currently grown in the ELGO-DIMITRA greenhouse premises at Lykovryssi. A safety copy of this wild vine collection will be established in the ELGO-DIMITRA premises at Thermi, Thessaloniki (Figure 1a), in the following years.

Genetic diversity indices, including allelic richness (Na), effective alleles (Ne), expected heterozygosity (He), and Shannon’s information index (I), were calculated using GenAlEx v6.5 [30]. The genetic differentiation among populations was quantified using pairwise Fst values based on Nei’s genetic distance [32], computed using Arlequin v3.5 [33]. To further explore genetic relationships, a principal coordinate analysis (PCoA) was performed using GenAlEx v6.5 [30].

To assess the genetic structure and potential admixture among the wild grapevine genotypes collected from the greater Dikili Tash region, a model-based Bayesian clustering analysis was conducted using the program STRUCTURE version 2.3.4 [34]. The analysis was based on multilocus SSR data obtained from 20 microsatellite loci. The admixture model with correlated allele frequencies was applied, assuming that individuals may have mixed ancestry. The number of assumed genetic clusters (K) was set from K = 1 to K = 10, with 10 independent runs per K value to ensure consistency across replicates. Each run consisted of a burn-in period of 100,000 iterations, followed by 100,000 MCMC repetitions. All simulations were performed using the parallel-processing version of STRUCTURE to reduce computation time. To determine the most likely number of genetic clusters, the Evanno method [35] was applied. The outputs were analyzed using STRUCTURE HARVESTER [36], which calculates the rate of change in the log probability of data between successive K values (ΔK). STRUCTURE bar plots were visualized and aligned using CLUMPAK [37].

The genetic relationships between the investigated genotypes were analyzed using a discriminant analysis of principal components (DAPC), implemented in R/adegenet [38].

3. Results and Discussion

In the current research article, the detection of fermented grape biomarkers, in the Neolithic archaeobotanical findings that led to the conclusion that wine making using V. sylvestris grapes was a common practice 4300 years B.C. in the Dikili Tash area [3], was the starting point to explore the hypothesis whether vines of V. sylvestris still occur in the wider area. Research expeditions localized natural populations of wild grapevines either in scattered individuals or in small groups and samples were collected and analyzed by microsatellites. This is the first report to focus on V. sylvestris populations in Greece since the early 1960s, apart from the single Pausanias vine (or “Pafsanias vine”), which attracted the attention of local and foreign research groups in the recent past [9,10,11].

Microsatellite analyses were performed using a total of 20 microsatellite markers (Supplementary Table S1) and a dendrogram was finally constructed (Figure 2) showing the genetic relation of the samples analyzed.

The genetic diversity analysis revealed significant variation among the populations. Mt Pangaion-Route A displayed the highest genetic diversity, with Na = 6.55, Ne = 3.77, and He = 0.708, indicating a genetically diverse population. Mt Lekani populations also exhibited high heterozygosity values (He > 0.70), suggesting a strong genetic base. In contrast, some populations, particularly those in isolated locations, exhibited lower He values, implying potential genetic drift or historical bottlenecks. Shannon’s information index (I) ranged from 0.451 to 1.491, with the highest values observed in Mt Pangaion-Route A (Figure 3a).

Pairwise Fst values, used to quantify genetic differentiation, ranged from 0.008 to 0.541, indicating varying degrees of genetic structuring among populations. Low Fst values (0.008–0.051) were observed between Mt Pangaion-Route A and Mt Lekani-Route B, suggesting close genetic relationships and possible gene flow between these two populations. On the contrary, higher Fst values (>0.30) were observed between certain populations in Mt Lekani and more geographically distant sites, suggesting limited genetic exchange. A heatmap visualization of Fst values confirmed these trends, with some populations forming closely related genetic groups, while others exhibited moderate differentiation (Figure 3b).

The PCoA (Figure 3c) further supports the genetic structuring of populations, with the Outgroup (“Phthiotis” sample) forming a distinct cluster separate from the main populations. The first two coordinates explain 20.11% of the total genetic variance, with Coordinate 1 accounting for 10.76% and Coordinate 2 for 9.35%. The populations from Mt. Pangaiοn and Mt. Lekani routes clustered together, while individuals from Around Dikili Tash population depicted some degree of admixture.

To gain deeper insight into the population structure of the wild grapevine populations around Dikili Tash, we conducted a model-based clustering analysis using STRUCTURE (Figure 4). The number of genetic clusters (K) tested was from 1 to 10, and the optimal K was inferred using the ΔK method by Evanno et al. [35]. The highest ΔK value was observed at K = 2, indicating the presence of two major genetic groups among the samples. Nevertheless, STRUCTURE plots for K = 3 and K = 4 also revealed sub-structuring and signs of admixture in several individuals, particularly from Mt Pangaion routes and the area surrounding the archeological site. These findings are consistent with the DAPC and PCoA analyses (Figure 5), supporting the hypothesis of both genetic differentiation and gene flow across the sampled populations.

Mt Pangaion-Route A and Mt Lekani-Route B exhibit high genetic diversity, likely due to historical gene flow and environmental stability. In contrast, some populations in Mt Lekani and more geographically distant sites show signs of genetic differentiation, likely driven by isolation and localized adaptation. These patterns are consistent with prior studies on wild grapevines in Greece, where the populations show genetic structuring influenced by geographic barriers and historical environmental changes.

The low Fst values between certain Mt Pangaion and Mt Lekani populations suggest genetic connectivity, possibly due to shared ancestry or ongoing gene flow. However, the moderate-to-high Fst values between certain populations highlight potential genetic drift effects. The genetic clustering results, confirmed by UPGMA and PCoA, indicate the presence of three major genetic clusters.

Hence, the populations with high heterozygosity (e.g., Mt Pangaion-Route A and Mt Lekani-Route B) are crucial for maintaining genetic diversity and should be prioritized in conservation efforts. The observed genetic structure suggests that these populations may require conservation strategies to prevent the loss of genetic variation.

A discriminant analysis of principal components (DAPC) showed that the V. sylvestris samples of the wider Dikili Tash area formed a distinct cluster in the lower left of the diagram (Figure 5a) far from the rootstocks cluster on the right of the horizontal axis, and far from the cultivar cluster on the upper left of the diagram. A DACP on Mediterranean V. sylvestris populations grouped the Dikili Tash populations close to Croatia and Bosnia and Herzegovina populations (Figure 5b).

The cuttings from all the vines that the samples have been collected and analyzed in the current work were also collected and are currently grown in the ELGO-DIMITRA green house facilities, until they were strong enough to be transplanted in the open field. By these means, the first ex situ V. sylvestris conservation site will be established in Greece, providing the necessary plant material for future work. It is noteworthy that upon sample collection optical observations and photographs were taken of the collected leaves. The leaf shape of some vines was similar to the leaf shape of some wild grapevine types reported by Logothetis [8]; for example, types M5a, M6, F1, and F2 have been recognized. A proper ampelographic description of the analyzed vines will be performed on the vines that will be established in the ex situ collection in ELGO-DIMITRA premises. In addition, characterization especially in response to abiotic factors and also experimental work aiming to evaluate the oenological potential of the grapes from the identified wild genotypes have already been initiated.

In the meantime, numerous V. vinifera spp. sylvestris samples have been collected from other sites of the country and are currently analyzed at ELGO-DIMITRA and also at JKI as part of a large European set of unique genetic profiles of V. vinifera spp. sylvestris samples aiming to explore the genetic diversity and population structure of this subspecies in a broad context [16].

The Dikili Tash V. sylvestris populations were found to be distinct from the cultivars and the rootstocks since all grouped in a separate cluster (Merkouropoulos on the resource at https://www.ecpgr.org/working-groups/vitis/sylvestris, as accessed on 29 May 2025), and close to Balkan populations. Considering the results presented by Perko et al. [39], it seems that the Balkans constitute a cluster distinct from the eastern (Armenia, Georgia, Azerbaijan), southeast (Israel) or western (France, Spain) populations; oddly, Slovenian populations are grouped closer to the French and Spanish ones.

4. Concluding Remarks

The current research is the first report in recent times concerning localization of V. sylvestris in Greece. Biostatistical analysis of the microsatellite data shows that the Greek populations identified in the north-east part of the country, on Mt Pangaion, on Mt Lekani, and in the valley that occurs between these mountains, are genetically close with possible gene flow to have occurred between them, although in some geographically distant sites such genetic exchange was rather limited. Our data is comparable to other Balkan populations, such as those from Croatia and Bosnia-Herzegovina. Conservation efforts should urgently be set in order to prevent genetic erosion.