Georgian Grapevine Cultivars: Ancient Biodiversity for Future Viticulture

Grapevine (Vitis vinifera) is one of the most widely cultivated plant species of agricultural interest, and is extensively appreciated for its fruits and the wines made from its fruits. Considering the high socio-economic impact of the wine sector all over the world, in recent years, there has been an increase in work aiming to investigate the biodiversity of grapevine germplasm available for breeding programs. Various studies have shed light on the genetic diversity characterizing the germplasm from the cradle of V. vinifera domestication in Georgia (South Caucasus). Georgian germplasm is placed in a distinct cluster from the European one and possesses a rich diversity for many different traits, including eno-carpological and phenological traits; resistance to pathogens, such as oomycetes and phytoplasmas; resistance to abiotic stresses, such as sunburn. The aim of this review is to assess the potential of Georgian cultivars as a source of useful traits for breeding programs. The unique genetic and phenotypic aspects of Georgian germplasm were unraveled, to better understand the diversity and quality of the genetic resources available to viticulturists, as valuable resources for the coming climate change scenario.

the only species extensively used in the global wine industry, covering approximately 7.4 Mha in 2018, and producing more than 77.8 mt of grapes (wine, table and dried grapes) and a world wine trade worth around EUR 32 billion 1 . Regions of its cultivation are located roughly between the 35th and 55th northern parallels and between the 25th and 35th southern parallels, in areas with average annual temperatures between 10 and 20 • C. These environments are characterized by the alternation of a favorable growing season and an unfavorable cold one. However, the cold winters are not too intense (minimum temperatures range between −10 and 15 • C) and the favorable season (average temperature higher than 10 • C) is long enough (>200 days) for grapes to ripen (Gladstones, 1992).
Viticulture depends on environmental resources (i.e., climate and soil conditions) in terms of yields and quality (van Leeuwen and Darriet, 2016). The current climatic phase, characterized by the increase of average global temperature, has led to changes in the environmental conditions of agricultural areas that need to be tackled with suitable tools, in a context of adaptation and mitigation. Due to the socio-economic impact of the wine sector in Europe and around the world, over recent years, there has been an increase in work aiming to study the impact of climate change on viticulture (Hannah et al., 2013;Morales-Castilla et al., 2020). Santos et al. (2020) proposed a list of measures to be adopted in viticulture to face with the climate change. The list divided the measures in two categories: the short-term adaptation strategies and the long-term adaptation strategies. The shortterm strategies include crop cultural practices and techniques to delay ripening time, plant protection against extreme heat, irrigation, pest and disease control and soil management. Among the long-term strategies, there are: change in training systems, varietal/clonal and rootstock selection and vineyard relocation. Breeding programs for new varieties which will be better able to perform in the environmental conditions expected in the future could be one of the most promising solutions, although this strategy is included in the long-term ones. An appropriate cultivar selection reduces the inputs required for plant management, increasing the sustainability of production. Great sources of biodiversity in the V. vinifera species have been recently found in its domestication cradle, located in Georgia (South Caucasus) (Imazio et al., 2013) (Figure 1).
Georgia counts 48,000 hectares of vineyards and a production of wine and table grapes of 159,000 and 8,000 tons, respectively (see text footnote 1). In 2015, about 100 M liters of wine were produced, 80% of them obtained by white and 20% from red berry grapes. More than 90% of the 2015 production was supported by Kakheti region, producing mainly white and red wines from Rkatsiteli and Saperavi grapes, in the ratio 7:3.
The aim of this review is to assess the potential of Georgian cultivars as sources of useful traits for new breeding programs, aiming to face the future challenges that await viticulture worldwide. To do this, we reviewed the particular genetic and phenotypic aspects (such as berry traits and resistance to pathogens) of Georgian germplasm, in the hope of better understanding the diversity and quality of the genetic resources 1 http://www.oiv.int/ available to viticulturists, coming directly from the origin of domestication.

SOUTH CAUCASUS, THE FIRST GRAPEVINE DOMESTICATION CENTER
Vitis vinifera is indigenous to Eurasia and it is suggested that the ancestors of the first Vitis genus appeared about 65 million years ago (Olmo et al., 1995). Nowadays, V. vinifera species includes both cultivated (V. vinifera subsp. sativa) and wild (V. vinifera subsp. silvestris) subspecies, the latter considered the progenitor of subspecies sativa . Its domestication process seems to be strongly linked to the alcoholic and gustative superiority of its fermented juice (the wine) in comparison to that of other fleshy fruits (fruit wines), although it is not well known which process predated the other (Terral et al., 2010). The main changes driving grapevine domestication were identified in the flower morphology (appearance of hermaphrodite flowers), larger berry size, higher berry sugar content, a wide range of berry color and aromatic content, characters which ensure yield, quality and a greater sugar content for a better fermentation . The major questions about grapevine domestication process are related to the number of events occurred, single event versus multiple events, and the geographical location where these events took place. For a vine domestication center to be born, different conditions need to occur. Among these, there is a strong awareness in practicing and developing viticulture by entire peasant villages. To bring out such a situation, many factors have to converge: territories with a (relatively) high population density, with stable settlements and in positions at crossroads of trade flows and cultural trends (Forni, 2012). It would be reasonable to expect that such situation could have occurred in several areas, differing in chronology and level of development. The most accredited hypothesis suggests that V. vinifera was domesticated from its wild form in the South Caucasus, between the Caspian and Black Seas, around 6,000-5,800 BC, and then spread throughout Europe and Mediterranean areas thanks to the spread of civilizations (McGovern et al., 2017). Recently, Zhou et al. (2019), proposing a four-state domestication process for grapevine, date the beginning of this process around 20,000 years ago, when South Caucasian human populations started to manage and harvest the local wild populations (Stage 1). In the same region around 8,000 years ago, the humans started with the conscious or unconscious selection of desirable phenotypical traits (Stage 2), although this transition is not well documented yet. Another force driving the transition from Stage 1 to Stage 2 is the bottleneck. Nevertheless, genetic evidences showed that grapevine did not experience a severe bottleneck (Myles et al., 2011;Zhou et al., 2017), making the conscious or unconscious selection as a unique force shaping the genetic diversity of grapevine during the domestication process. Stage 3 consists on spreading of newborn crop in new locations and the consequence of local domestication or introgression events. Reviewing the most comprehensive studies on grapevine genetic population, it turned out that an East-to-West grapevine gene flow after the first domestication process occurred, with some evidence FIGURE 1 | Map of Georgia and location of 12 Georgian wine-growing regions: 1 -Abkhazeti; 2 -Samegrelo; 3 -Guria; 4 -Adjara; 5 -Svaneti; 6 -Lechkhumi; 7 -Racha; 8 -Imereti; 9 -South Kartli; 10 -Inner Kartli; 11 -Lower Kartli; 12 -Kakheti. Image is obtained by Google Earth. Pin indicates Tbilisi position. Map on the right reports the Winkler classification based on yearly average Winkler index calculated for the period 1994-2013 in Georgia (Caucasus). The analysis is limited to the areas below 1250 m above sea level. Description of Winkler indices: (I-) GDD (Growth Degree Days) < 850, viticultural climate is very cool, vinicultural aptitude is very early ripening grapes for fresh and fruity wines or sparkling wine bases. (I) GDD 850-1400, viticultural climate is cool, vinicultural aptitude is early ripening grapes for fresh and fruity wines or sparkling wine bases. (II) GDD 1400-1650, viticultural climate is temperate cool, vinicultural aptitude is early ripening grapes for wines to be aged. Medium ripening grapes for white or red wines ready to drink. (III) GDD 1650-1950, viticultural climate is temperate, vinicultural aptitude is medium ripening grapes for white or red wines ready to be aged. (IV) GDD 1950-2200, viticultural climate is temperate warm, vinicultural aptitude is late ripening grapes for white or red wines ready to be aged. (V) GDD 2200-2700, viticultural climate is hot, vinicultural aptitude is late ripening grapes for bodied red wines to be aged. (V+) GDD > 2700, viticultural climate is very hot; vinicultural aptitude is very late ripening grapes for bodied red wines to be aged.
of putative secondary domestication centers along the main migration routes due to genetic relationships between wild and cultivated accessions, especially in the Mediterranean Basin and Central Asia (Grassi et al., 2003;Arroyo-García et al., 2006;Myles et al., 2011;Zhou et al., 2017;Riaz et al., 2018). The coexistence of wild populations together with domesticated ones is often and the bidirectional gene flow (wild-to-cultivated and cultivated-towild) has been well documented (De Andrés et al., 2012;Ekhvaia et al., 2014;Riaz et al., 2018;D'Onofrio, 2020;Maraš et al., 2020), supporting the occurrence of secondary domestication events from local wild populations or introgression events. These events, the geographical origin and human usage were found to strongly shape the genetic structure of grapevine germplasm . Another aspect, although less investigated, is the role of wild Vitis species in the sativa domestication process. It seems that wild Vitis species have contribute to the current structure of grapevine germplasm (Zhou et al., 2019). The last stage proposed by Zhou et al. (2017) (Stage 4) takes into account the modern breeding programs, a relative recent event occurred over the last few hundred years and led to the birth of so-called anthropic crossings, with the aim of satisfying specific requirements.

GEORGIAN TERRITORY, CLIMATE, AND GRAPEVINE PRODUCTION
Georgia is a large basin of the mid latitudes, bordered by the Greater Caucasus in the North and the Lesser Caucasus in the South, and opening toward the Black Sea in the West and toward the Caspian depression in the East (Figure 1). Those geographical features strongly characterize the climate of its 12 wine-growing regions that, following the Köppen -Geiger classification (Köppen and Geiger, 1936), are characterized by profoundly different climatic conditions, ranging from hot summer continental climates to warm summer continental or hemiboreal climates, that translate into different classes of the Winkler index (Figure 1).
In relation to the climatic conditions of each wine-growing region, the Georgian varietal assortment is strongly differentiated as well, being adapted to a very wide range of cold and summer stresses ( Table 1). Worldwide wine-growing regions experiencing the same climatic conditions of Georgia may provide benefit by this so differentiated varietal spectrum.
It is interesting to highlight that, in 1994 Georgia faced an abrupt rise in temperatures, similarly to what happened in Western Europe in the late 1980s (Reid et al., 2016), with 1987 as the most likely year of change (Mariani et al., 2012). This delay could be explained by the progressive dilution of the Atlantic circulation signal as it moves into the European continent (Cola et al., 2017). The increase of temperature determined an advance in grapevine phenology, which was more significant at the higher altitudes, where more favorable thermal conditions were established. On the other hand, at lower altitudes the phenological advance was partially depleted by the increase of super-optimal thermal conditions (increasing the occurrence of stress conditions during ripening). For instance, in the case of the widely diffused cultivar Rkatsiteli, the average advance of veraison was less than 6 days for the 250-500 m asl elevation belt and around 18 days for the 750-1000 m one (Cola et al., 2017).
In parallel, it is worth noticing the high variability in the plant phenology among Georgian cultivars, both in the sprouting date and in the ripening period. A delayed budburst period could represent an avoidance mechanism against spring frosts. Considering Georgian cultivars, bud swelling of 'Partala'  : The risk of summer stress is expressed as the percentage of the years of the reference period with at least 7 days with maximum temperature above the 35 • C threshold. The classes are: very low (<3%), low (3-5%), medium (5-6.7), high (6.7-10%) and very high (>10%  : The risk of winter frost is expressed as the percentage of the years of the reference period winter minimum temperature below the −15 • C threshold. The classes are: very low (<3%), low (3-5%), medium (5-6.7), high (6.7-10%) and very high (>10%). 4 Spring Frost (1974: The risk of spring frost is expressed as the percentage of the years of the reference period with spring minimum temperature below the −2 • C threshold. The classes are: very low (<3%), low (3-5%), medium (5-6.7), high (6.7-10%), and very high (>10%). 5 N, noir; B, Blanc.
vines was recorded at the end of March, and, thus, a higher susceptibility to spring frost is expected when compared to the other cultivars that sprouted in April . Global warming generally resulted in the increase of cases of temperature above the optimal range (24-26 • C) during summer and in particular during grape ripening (Cola et al., 2020). A delay in the maturation process, obtained through the selection of late-ripening cultivars, could ensure thermal conditions during ripening more suitable for berry metabolism. Maghradze et al. (2012) studied the phenology of Georgian cultivars in northern Italy in comparison to Chardonnay and Cabernet Sauvignon grown in the same area, and they found a relatively late ripening with respect to the reference varieties: nevertheless, a very wide range of variability was observed. Similar results were found in other comparative evaluation carried out in Georgian ampelographic collections and are reported by Maghradze et al. (2014) and Rustioni et al. (2014). Some extreme cases are: early ripening cultivars -Kartuli Saadreo, Meskhuri Mtsvane, Buza, Budeshuri Tsiteli and Daisi; late ripening cultivars -Ojaleshi, Akomshtali, Kamuri, Shavi, Tavkara, Khushia Shavi, Satsuravi, Maghlari Tvrina, Mtevandidi, Argvetula, Dziganidzis Shavi, Adanasuri, Mamukas Vazi, Otskhanuri Sapere, Gorula, Saperavi Meskhuri, Ghrubela and Shavtita. The same results were obtained when comparing the phenological timing of Georgian varieties internationally grown. The phenological model developed by Mariani et al. (2013) for Cabernet Sauvignon and Chardonnay and adapted to the Georgian varieties Saperavi, Rkatsiteli, Mtsvane Kakhuri (Cola et al., 2017) was applied to a long time series of daily temperature (Perugia-Italy. 1990-2019). Figure 2 shows the late phenological timing of Georgian varieties (average values are shown).

THE AMPELOGRAPHIC COLLECTIONS OF GEORGIAN GERMPLASM
To date, fifty grapevine varieties are recommended for cultivation in Georgia. Most of them (37) (7 accessions). Although the number of accessions is high, only a limited number of them were genotyped and phenotyped (see sections "Georgian Germplasm as a Source of Genetic Variability" and "The Phenotypical Characterization of Georgian Germplasm Collections"). This limited number of information makes a not so easy determination of the exact number of autochthonous Georgian varieties. Further efforts are needed to better understand the genetic diversity of this valuable germplasm and to identify synonyms, homonyms and misidentifications.

Georgian Germplasm as a Source of Genetic Variability
Historical information coupled with archeological and palaeobotanical findings pointed to Georgia as a cradle for grapevine domestication (Zohary and Hopf, 2000;McGovern, 2003;McGovern et al., 2017). Molecular analysis produced the same evidence. Genetic diversity of Georgian germplasm was investigated, by both nuclear SSR (simple sequence repeat) (Laucou et al., 2011;Imazio et al., 2013;Ekhvaia et al., 2014) and  Thanks to two European research programs, GrapeGen06 (2007-2010) (Laucou et al., 2011), first, and then COST Action FA1003 (2011-2014) (Failla, 2015), a strong and still active network of scientific collaborations has been developed between European and Georgian researchers, to genetically characterize and preserve the Georgian genetic resources of vines.
All the outcomes about the genetic characterization of Georgian germplasm reported the uniqueness and originality of this germplasm when compared to the European and Central Asian germplasm (Myles et al., 2011;Bacilieri et al., 2013;Imazio et al., 2013;Riaz et al., 2018;De Lorenzis et al., 2019). The Georgian cultivars showed the distinctive features of a domestication center, such as high levels of genetic diversity and heterozygosity, the presence of alleles absent or poorly represented in other countries, and differentiation from the European varieties, clustering in a well-separated branch (as reported in the Figure 3, where SSR and SNP genetic profiles of varieties from France, Georgia, Italy and Spain were re-elaborated to perform a discriminant analysis of principal component, using data published in De Lorenzis et al., 2015, 2019, Laucou et al., 2018, and Riaz et al., 2018. A differentiation inside the germplasm, based on the geographical origin of cultivars, was identified as well: the varieties putatively originated in Kartli and Kakheti (Eastern regions) differ from the ones originating in Abkhazeti, Samegrelo, Guria, Adjara, Imereti, Racha, and Lechkhumi (Western regions). The origin of this subdivision lies in the geographical subdivision of Georgia into two major parts, due to the Likhi Mountains running in a North-to-South direction across Georgia (Imazio et al., 2013;De Lorenzis et al., 2015), confirming that, despite long-standing cultivation, the Georgian cultivars maintain their originality.
Genetic variation provides the foundation for any breeding programs, and natural genetic diversity represented historically the major source of variability for crop improvement and adaptation to changing environmental conditions. Given the uniqueness of Georgian germplasm, its strong link with the regions of origin coupled with the evidence of this country being the center of domestication makes this germplasm very attractive for investigation from the perspectives of phenology, grape phenotype and resistance to biotic and abiotic stresses, as sources of new variability for future breeding programs.

The Phenotypical Characterization of Georgian Germplasm Collections
The collaboration among European and Asian researchers makes feasible the comparisons among ampelographic collections using common protocols. Among them, phenotyping was considered in the framework of the COST Action FA1003 (Rustioni et al., 2014), allowing the description of numerous autochthonous cultivars (Abashidze et al., 2015;Cornea and Savin, 2015;Goryslavets et al., 2015;Maghradze et al., 2015;Ujmajuridze and Mamasakhlisashvili, 2015). This work, finally, produced a general overview of the V. vinifera variability concerning enocarpological traits (Rustioni et al., 2019). Table 3 reports the distribution of the Georgian records with respect to the variability described for the V. vinifera species (data obtained by the reworking of the results published in Rustioni et al., 2014, 2019, and Abashidze et al., 2015. To emphasize some results, showing the differences among the two groups of data, Figure 4 shows the frequency distribution of the Georgian records in comparison with the data collected for the entire V. vinifera species, concerning some specific traits (titratable acidity, percentage of skin, skin phenolic content). Briefly, the main differences highlighted in Table 3 in terms of oenological applications are that Georgian grapes have, with respect to the V. vinifera species population, higher concentrations in both sugars and acids and thicker skins, ensuring acceptable amounts of phenolics despite the lower accumulation per unit of tissue. Details concerning the results of this comparison are discussed in Sections "Fruit Morphology and Technological Quality of Georgian Cultivars" and "Abiotic Stress Adaptations and Secondary Metabolisms." It is worth noting that the phenotypic variability reported is due to the genotype, to the environmental growing conditions, and to their interactions. Thus, further studies will be necessary to discriminate these effects, highlighting the role of genotypes.

Fruit Morphology and Technological Quality of Georgian Cultivars
Despite the wide variability for berry shapes within the species, Georgian grapes generally have round or slightly elongated small berries (Table 3) (OIV, descriptor 223). This is probably due to the ancient traditions that, during millennia of winemaking activities TABLE 3 | Physical dimensions and chemical components distribution of Georgian grapevine germplasm with respect to the variability described for the entire V. vinifera species (reworking of the results published in Rustioni et al., 2014, 2019and Abashidze et al., 2015.

Variable
Sample  Considering technological maturity, Georgian records generally show higher concentrations in both sugars and acids than foreign varieties ( Table 3). In the perspective of climate change, the high sugar content could represent a problem, due to the risk of further increases related to the higher temperatures during the anticipated ripening periods (Keller, 2010;Mira de Orduña, 2010;van Leeuwen and Destrac-Irvine, 2017). The increased sugar concentrations expected in hot ripening conditions may cause growth inhibition or lysis in the yeasts responsible for wine fermentation (Mira de Orduña, 2010). Furthermore, high sugar stress could modify the yeast metabolisms, increasing the accumulation of by-products (such as glycerol and acetic acid) that, together with the increased alcoholic content, affects the wine perceptions of consumers and, thus, it could modify the expected characteristics of traditional wines (Mira de Orduña, 2010). However, the deviation of Georgian data with respect to the species variability concerning the sugar content is rather limited (significance of the difference = 0.073), neither covering the maximum records of the one obtained when analyzing wider grapevine genetic pools ( Table 3) (Rustioni et al., 2019). Furthermore, the expected sugar content increase is usually ascribed to the earlier ripening anticipated in climate changed conditions (Palliotti et al., 2014;Martínez-Moreno et al., 2019), and it is important to remind the prevalence of late ripening cultivars among the Georgian grapevines . Finally, the sugar content seems to be well counterbalanced by the acidity ( Table 3). The distribution of Georgian grapes concerning the titratable acidity, showed a right shift (Figure 4A), demonstrating the ability of these cultivars to keep a high acidic concentration despite the sweetness of the berries. This is a crucial point for viticulture adaptation to climate change, because high temperatures usually cause a decrease in acids, especially due to malic acid degradation (Keller, 2010;Mira de Orduña, 2010;van Leeuwen and Destrac-Irvine, 2017).
In the perspective of climate change adaptation, it is very interesting to note a particular feature of Georgian grapes concerning the proportions among skin, seeds and pulp, at the expense of the latter ( Table 3). The Georgian records shown in Figure 4B, demonstrate an important shift toward thicker skins with respect to the general V. vinifera species. This is due to the lighter berries and heavier skins ( Table 3). Considering the effect of climate change on the berries, a thicker skin could represent a more resistant barrier against the stressful environment. In fact, it has been shown that a possible adaptation to climate change could be related to berry skin thickening (van Leeuwen and Destrac-Irvine, 2017).

Abiotic Stress Adaptations and Secondary Metabolisms
Grape epicuticular waxes also have important protective roles against dehydration (Pangavhane et al., 1999;Di Matteo et al., 2000;Doymaz, 2006;Muganu et al., 2011) and pathogen infections (Marois et al., 1986;Rosenquist and Morrison, 1988;Percival et al., 1993). Furthermore, a study conducted on Georgian cultivars, suggested a possible eco-physiological role of epicuticular waxes in reducing heating stresses by an interaction with infrared radiation . However, a comparison among Georgian cultivars and grape varieties cultivated in other regions is not available and, thus, we should suppose that this mechanism is not exclusive for Georgian cultivars. Excesses of photosynthetically active radiation (PAR) could cause problems in grapes due to chlorophyll overexcitation Rustioni, 2017). Rkatsiteli response to photo-oxidative sunburn was tested by Rustioni et al. (2015). It was considered among the "tolerant cultivars, " as it showed relatively low susceptibility to sunburn (recorded as browning appearance) at all the phenological periods studied. In particular, the correlation between chlorophyll contents and browning symptoms had a high R 2 (0.989), but the slope coefficient (60.2) together with the average Browning Intensity Index (27.5) indicated a light symptom appearance in Rkatsiteli grapes. In photo-oxidative sunburn, browning symptoms appear due to the reactive oxygen species (ROS) scavenging activity of phenolics through their oxidation and consequent polymerization that produce brown pigments (Felicetti and Schrader, 2008;Rustioni et al., 2015). Often, plants face stresses through secondary metabolites, and the crucial role of phenolics against photodamage is well known (Close and McArthur, 2002;Graham et al., 2004;Rustioni, 2017). However, if the substrate for these oxidative polymerizations (phenolics) are in low concentrations, the sunburn browning symptoms could appear less intense: this is likely in the case of Rkatsiteli. Abashidze et al. (2015) reported 404.7 ± 58.3 mg/kg of grapes as average skin phenolics for this cultivar, which falls in the first 10th percentile of the V. vinifera variability concerning this trait (Rustioni et al., 2019) (Table 3). Considering total phenolic compounds, Georgian cultivars appeared to accumulate low amounts of these molecules in skins ( Figure 4C and Table 3), but the difference is still exacerbated by seed phenolics (Table 3). In fact, the average percentage of phenolics arising from seeds is much lower in data coming from Georgia (13.7% in Georgian records in comparison with 20.1% of species characteristics). Of course, considering the eco-physiological role of phenolics, this trait could be considered as a downside of Georgian cultivars. However, in a production perspective, it could be an important advantage.
Climate changes often produce disequilibria in the berry ripening processes, increasing the quantity of phenolic compounds (Keller, 2010;van Leeuwen and Destrac-Irvine, 2017) that, often, do not reach an optimal ripening quality. Unripe phenolics could strongly compromise the wine quality, being involved in the perception of bitterness and astringency (Kontoudakis et al., 2011). Seed phenolics, due to their intrinsic characteristics, are often considered as an undesirable source of defects, so it is true that technologies have been developed to separate seeds to prevent phenolic extractions in wines (Canals et al., 2008) or to artificially ripen them under controlled conditions (Rustioni et al., 2018;VanderWeide et al., 2020). In this perspective, and considering that climate change is expected to make it harder to reach an equilibrated phenolic ripening in grapes, the lower phenolic concentration of Georgian cultivars (especially in the seeds) could be considered as a positive trait to deal with future difficult ripening conditions.
Another important class of phenolic molecules is represented by anthocyanin pigments. Among the 48 native Georgian grapevine varieties described by Tsertsvadze (2012), 21 of them are white grape cultivars, while the other 27 have pigmented berries (22 black, 2 red, 1 gray, and 2 pink). Among the native Georgian grape varieties described by Ketskhoveli et al. (1960), 245 of them are not pigmented (241 white and 4 yellow) grape cultivars, while the other 278 have pigmented berries (221 black, 27 red, 5 gray, and 25 pink). The reflectance spectra of 51 Georgian cultivars, together with other 69 accessions originated from other countries, were studied by Rustioni et al. (2013). Based on this first screening, some of these cultivars were selected to highlight dysfunctions in anthocyanin accumulation: Ubakluri, Ghrubela Kartlis, Rkatsiteli Vardisperi (and Marguli Sapere among the reference cultivars). Ubakluri shows a very light color, due to a very low pigment accumulation. Ghrubela Kartlis, due to the prevalence of peonidin-3-O-glucosides among the anthocyanins, has a gray appearance. Rkatsiteli Vardisperi, with the salmon pink color due to the high proportion of cyanidin-3-O-glucosides, is considered a berry color mutant resulting from a retro-transposon-induced mutation of the Rkatsiteli white-skinned cultivar De Lorenzis et al., 2020). These color peculiarities could be interesting for future selections, especially considering the importance of appearance for table grape markets.
The environmental conditions (e.g., light and temperature) can affect the pigment accumulation in skins and the modulation of the anthocyanin biosynthetic pathway in berries could be considered as a grapevine eco-physiological adaptation mechanism (Keller, 2010;Rustioni et al., 2011;De Lorenzis et al., 2016). Considering anthocyanins (Table 3), Georgian data generally show slightly higher contents of pigments with respect to the V. vinifera species average when expressed as mg/kg of grapes or mg/berry. However, this appears mainly due to the thick Georgian berry skins, and, thus, it is not due to a higher accumulation in this tissue, but to the higher quantity of pigmented tissue itself. In fact, when considering the anthocyanin accumulation in skins, the average Georgian record is 2.77 mg/g of skin, while the species average is nearly twice higher (4.7 mg/g of skin).

RESISTANCE TO GRAPEVINE FUNGAL DISEASES
The grapevine varieties cultivated worldwide belong to the Eurasian grapevine, V. vinifera, and are susceptible, at different levels, to several pathogens (fungi, bacteria, and viruses), while non-vinifera species, from North American and Asian, are resistant to fungi and tolerant to viruses and some bacteria (Oliver and Fuchs, 2011;Armijo et al., 2016). Amongst the various diseases which directly affect grapevines, powdery mildew (caused by the ascomycete Erysiphe necator) and downy mildew (caused by the oomycete Plasmopara viticola) are two of the most important (Bois et al., 2017). Disease management became an unavoidable task for European viticulture in the second half of the nineteenth century, when the two pathogens were introduced into Europe and the European grapevine growers were faced with their destructive effects (Töpfer et al., 2011). The P. viticola introduction was a probable consequence of the massive importation of American grapevine species to be used as rootstock for V. vinifera and contrast the destructive effects of phylloxera, caused by Daktulosphaira vitifoliae, on the Eurasian grapevine species (Granett et al., 2001;Gessler et al., 2011). The search for suitable tools for disease management rapidly became a priority for the viticulturists. The discovery of the efficacy of sulfur and copper in controlling the diseases was a key point, but great attention was also paid to the development of resistant cultivars. The American Vitaceae soon proved to be the best sources of resistance, due to co-evolution with the pathogens, and extensive breeding programs, based on interspecific crosses between American Vitis species (e.g., Vitis riparia, Vitis rupestris, Vitis berlandieri and Vitis labrusca) and V. vinifera, were undertaken at the beginning of the XX century (Gessler et al., 2011). Nevertheless, the interest in searching for resistant plants decreased over time, probably due to the discovery of new fungicides (Russell, 2005), that were widely employed for disease control, and the inheritance of the specific foxy off-flavors from the non-vinifera parent species.
Recently, public concern about sustainability in agriculture and new regulations on plant protection products have renewed the interest of growers in the cultivation of resistant varieties (Merdinoglu et al., 2018). In fact, although viticulture in the whole of the EU only occupies a low percentage of arable land, the industry is responsible for a high use of fungicides to fight downy mildew infections (Eurostat 2 ). Furthermore, studies on the effects 2 http://ec.europa.eu/eurostat/de of CO 2 and temperature on downy and powdery mildews showed that the disease incidence of downy mildew increases with rises in gas and temperature, while an increase in CO 2 did not influence powdery mildew incidence (Pugliese et al., 2011). In view of the coming climate change, that will potentially favor the pathogens' development, it is also important to search for new resistance genes, focusing on alternative species, such as V. vinifera, to the non-vinifera ones.

V. vinifera Resistant Cultivars Against P. viticola
The identification of P. viticola dates back to 1838, when Schweinitz, one of the founders of American mycology, collected the first samples from wild Vitis species in South Carolina (Gessler et al., 2011). In Europe, downy mildew was first reported during 1878 in Bordeaux and then it spread all over the old continent and beyond, reaching Australia and New Zealand between 1919 and 1926 (Emmett et al., 1992). All traditional European grapevine cultivars showed high susceptibility to the pathogen, leading to severe pandemics across Europe (Boso and Kassemeyer, 2008;Gessler et al., 2011). Today, the pathogen is found in warm and humid climates worldwide.
Symptoms of downy mildew (Figure 5) are observable on infected organs as yellowish oily lesions (sometimes red, in black cultivars) on the upper surface of the leaves (Figures 5A,B) followed by sporulation on the underside of the leaf ( Figure 5C); malformations and necrosis on herbaceous shoots and inflorescences (Figures 5D,E); change of color to violet and withering on berries (Figure 5F), that detach from the rachis leaving a dry stem scar (Gessler et al., 2011). The disease negatively impacts grape production at both qualitative and quantitative levels: the loss of photosynthetic tissues limits the sugar amount in berries, that produce low quality wines; the shoot and bunch damage leads to poor yields. Severe infections, in the absence of disease control, can result in total loss of leaves and in some cases, total yield loss (Töpfer et al., 2011;Toffolatti et al., 2018).
Most of the Vitis taxa native to North America are to some extent resistant to P. viticola (Unger et al., 2007). The resistance response to P. viticola results in rapid plant cell death after pathogen recognition and local necrosis induction. This mechanism, known as the hypersensitive response (HR), is an actively triggered procedure initiated by fungal elicitors or other elicitors (Balint-Kurti, 2019) that leads to bursts of production of ROS and nitric oxide (NO). Consequently, the host cells collapse and shrink, hampering the fungal infection (Toffolatti et al., 2016). Cell death is visible to the naked eye as small necrotic spots on plant tissues.
The Georgian grapevine germplasm is characterized by very high genetic diversity, with cultivars differing from major European ones (Imazio et al., 2013). Considering that this high variability could also be a source of resistance to important pathogens, studies have been undertaken to assess the resistance levels of Georgian accessions to P. viticola. The first one, carried out by Bitsadze et al. (2015), showed that 20 accessions were characterized by medium to high levels of resistance to downy mildew in a collection of 61 native Georgian varieties. Given the promising results, it appeared worthwhile to keep screening Georgian germplasm. In Toffolatti et al. (2016), a total of 93 accessions were studied over a period of 3 years in field surveys and in the laboratory. A small group of varieties, including Kamuri Shavi, Mgaloblishvili and Ubakluri, showed low disease severity values, but only Mgaloblishvili showed a strong and constant phenotypical resistance against the pathogen. In Supplementary Table 1, a list of Georgian resistant varieties is reported. Indeed, recent studies on the transcriptome of Mgaloblishvili showed that the cultivar possesses a unique response to P. viticola that is based on the overexpression of genes that are not modulated or downregulated in susceptible (Pinot Noir, a V. vinifera cv) and resistant (Bianca, interspecific hybrid) cultivars (Toffolatti et al., 2018). The resistance mechanism of Mgaloblishvili is based on the overexpression of genes encoding: (i) receptors for pathogen recognition (PAMP -Pathogen Associated Microbial Patterns-receptors) and for damage at the cell wall (DAMP -Damage Associated Microbial Patterns); (ii) an NB-LRR receptor of fungal effectors (named Lr10); (iii) ethylene signaling; (iv) synthesis of terpenes, such as valencene, and flavonoids; and (v) strengthening of cell walls. Besides genes involved in resistance, susceptibility genes were identified as well. Susceptibility genes are essential for plant-pathogen interaction and their disruption leads to resistance, as with mlo gene, whose knockdown is involved in resistance to E. necator (Pessina et al., 2016). The candidate gene related to susceptibility to P. viticola in V. vinifera encodes an LOB domain-containing (LBD) protein (Toffolatti et al., 2020) that has been previously found in the interaction between Arabidopsis thaliana and Fusarium oxysporum (Thatcher et al., 2012). The new genome editing tools, providing several protocols to introduce knockout on target sequences, makes the understanding of plant pathogenresistance mechanism mediated by susceptibility genes a very attractive alternative for the development of durable diseaseresistant varieties (Zaidi et al., 2018).

New Resistant Loci Associated With Resistance to P. viticola in V. vinifera
The investigation of the genetic basis of P. viticola resistance through QTL (Quantitative Trait Loci) analysis on a range of North American and Asian Vitis species has led to the identification of 28 resistance (R) loci (Figure 6). These R loci (designated Rpv for Resistance to P. viticola) confer different degrees of resistance to disease, ranging from partial to total resistance (Dry et al., 2019). The major loci on this list are: (i) Rpv1, identified in Muscadinia rotundifolia, that confers a not total resistance to P. viticola infection and is associated with a gene encoding a TIR-NB-LRR protein (MrRPV1) (Merdinoglu et al., 2003;Feechan et al., 2013); (ii) Rpv2, identified in M. rotundifolia, that confers total resistance to downy mildew and is associated to a cluster of TIR-NB-LRR genes (Dry et al., 2019); (iii) Rpv3, identified in V. labrusca, Vitis lincecumii, V. riparia and V. rupestris, that confers partial resistance to downy mildew (Bellin et al., 2009;Gaspero et al., 2011;Welter et al., 2017); (iv) Rpv8 and Rpv12, identified in V. amurensis, that confer a high resistance to P. viticola infection and are associated with the cluster of genes encoding NB-LRR proteins (Blasi et al., 2011;Venuti et al., 2013); (v) Rpv15, identified in Vitis piasezkii, that confers strong resistance to P. viticola infection (Dry et al., 2019). The other R loci are considered minor loci due to their ability to confer low degrees of resistance and they are only useful when combined with major R loci.
Very recently, new promising downy mildew R loci (Rpv29, Rpv30, and Rpv31) have been identified, through a GWAS (Genome Wide Association Study), in the genetic background of the Georgian V. vinifera germplasm (Figure 6) (Sargolzaei et al., 2020). These new R loci, mapping on chromosome 14, 3 and 16 for Rpv29, 30 and 31, respectively, and conferring from high to very high resistance to downy mildew, seem to be associated with receptors of pathogen effectors, signaling mediated by protein ubiquitination and a cluster of Lr10-like (NB-LRR) effector receptors.

Low Susceptibility of Georgian Grapevine Cultivars to Phytoplasma-Associated Diseases
Flavescence dorée (FD) and Bois noir (BN) are the more important diseases of the grapevine yellows (GY) complex, responsible for severe yield losses in vineyards worldwide (Belli et al., 2010). FD and BN are associated with phytoplasmas, phloem-limited bacteria transmitted by insect vectors (Weintraub and Beanland, 2006). Even if their symptoms were indistinguishable (desiccation of inflorescences, berry shrivel, leaf discolorations, reduction of growth, and irregular ripening of wood), FD and BN are associated with phytoplasmas distinct at both genetic and ecological level (Belli et al., 2010). The FD phytoplasma is efficiently transmitted from grapevine to grapevine by the insect Scaphoideus titanus, which sustains its whole life cycle on Vitis spp. (Oliveira et al., 2019). Consequently, geographic areas hosting large vector populations and FD phytoplasma can be damaged by strong FD epidemics. Due to this aspect, FD phytoplasma is a quarantine pathogen, to be controlled through mandatory measures (Oliveira et al., 2019). On the other hand, BN phytoplasma ('Candidatus Phytoplasma solani') (Quaglino et al., 2013) is occasionally transmitted to grapevine by the insect Hyalesthes obsoletus, a polyphagous vector living preferentially on Urtica dioica (nettle), Convolvulus arvensis (bindweed), and Vitex agnus-castus (chaste tree) (Langer and Maixner, 2004;Kosovac et al., 2016). The epidemiological cycle associated with BN is extremely complex and was recently discovered to include other highly polyphagous insect vectors and a very broad range of secondary wild hosts (Mori et al., 2015;Quaglino et al., 2019). Moreover, the typical management strategies for phytoplasma diseases, based on the control of the vector(s) with insecticides and the removal of infected plants, are not effective against BN. Thus, it is difficult to organize effective prevention and containment measures. An ambitious strategy is based on the selection of plant varieties as the source of resistance-genes for plant breeding programs (Bianco et al., 2019). Unfortunately, none of the Vitis species and V. vinifera varieties studied have been found to be resistant or tolerant to the GY phytoplasmas (Laimer et al., 2009).
Surveys conducted in vineyards of Khaketi and Shida Kartli regions in eastern Georgia highlighted a wide diffusion of BN, while FD was not reported (Quaglino et al., 2014). Moreover, most autochthonous Georgian grapevine cultivars were found to be only mildly symptomatic, maintaining complete berry production, while internationally known cultivars exhibited severe symptoms (Quaglino et al., 2016) (Figure 7). As largely reported for phytoplasma-associated diseases of stone fruits, symptom intensity observed in infected plants can be influenced both by the virulence of the pathogen and the susceptibility level of the plant host (Kison and Seemuller, 2001;Seemüller and Schneider, 2007). Molecular characterization, supported by phylogenetic analyses, revealed that BN phytoplasma strains identified in Georgia constitute a bindweed-related population which is genetically distinct from the one found in central-western Europe. Interestingly, the presence of the same phytoplasma strain in grapevine cultivars showing a range of symptom intensity suggested a low susceptibility of Georgian local cultivars to BN (Quaglino et al., 2016) (Supplementary Table 1). Studies in progress are focusing on (i) identifying genetic traits associated with this low susceptibility to BN in the perspective of improving breeding programs to produce novel tolerant and/or resistant grapevine cultivars; (ii) investigating the susceptibility of Georgian grapevine cultivars to FD.

A BRIEF INTERLUDE ON THE STATUS OF GEORGIAN WILD COMPARTMENT
The V. vinifera subsp. silvestris is considered the progenitor of cultivated species. In the last two decades, an increase interest in preserving wild genetic resources has led to surveys on Georgian land aimed to localize and gather wild grapevine material. The plant material collected in these surveys is summarized in the Supplementary Table 2. These accessions are now partially (more than 100) available in Georgian collections (Saguramo, Skra and other collections) and some other in USDA National Clonal Germplasm Repository of Davis (CA, United States) and in the collection of Milan University. This subspecies is seriously worldwide endangered by human activities, such as urbanization, forest cleaning and setting fires (Arnold et al., 2005). The Georgian one is not an exception. Indeed, very small wild populations have been identified overall the Georgian land (Ocete Rubio et al., 2012). Populations with both male and female individuals were detected, but in Zhinvali and Sabue populations no female individuals were identified. Generally, in the Georgian populations the number of male individuals is higher than the female ones (Supplementary Table 2). Most of the wild Georgian populations showed severe downy and powdery mildew symptoms, although three individuals showed high resistance to P. viticola infection (Supplementary Table 2) (Ocete Rubio et al., 2012;Bitsadze et al., 2015). Nevertheless, remarkable is the absence of symptoms caused by phylloxera in the populations sampled by Ocete Rubio et al. (2012). In the same populations, symptoms caused by two mites, Colomerus vitis and Calepitrimerus vitis, have been observed, although the damages were not serious and appeared to do not affect the viability of the plants.
From the genetic point of view, some of these accessions were genotyped by SSR and SNP molecular markers  Table 2) (Imazio et al., 2013;Ekhvaia et al., 2014;De Lorenzis et al., 2015;Riaz et al., 2018). Results reported in Imazio et al. (2013) andDe Lorenzis et al. (2015) clearly discriminated the wild individuals from the cultivated ones, two subspecies that diverged at least 22,000 years ago (Zhou et al., 2017;Liang et al., 2019). The Georgian accessions were differentiated by European wild accessions and cultivated accessions (Riaz et al., 2018). Interesting, Ekhvaia et al. (2014) identified absence of genetic isolation among some of the analyzed wild populations due to gene flow among them.

(Supplementary
At the phenotypical level, few information is available and further studies need to deeply investigate the enological potential of this compartment. Nevertheless, preliminary results showed that musts obtained by Georgian wild grapes could be added to the must of traditional cultivars to improve the wine color .

WHAT'S NEXT?
Climate change will impact many aspects of human life, the environment, agriculture and food. Regarding viticulture, data available on climate change have already demonstrated impacts on wine growing areas, resulting in changes in grape chemical composition as well as grape phenology. Because of their isolated geographical origin and huge genetic variability, the Georgian grapevine germplasm is of great interest as a worthwhile resource for breeding programs. The Georgian germplasm has distinguished itself by including cultivars characterized by late ripening, which could potentially reduce issues related to excessive temperatures in summertime, distinctive enocarpological traits, which affect the grape and wine quality, specific response to abiotic stresses, such as sunburn, and resistance traits related to biotic stresses, such as P. viticola and phytoplasmas.
Given the reasons stated in this review, the screening and assessment of Georgian germplasm should be promoted at the phenotypical, agronomical, physiological and genetic level. A number of gaps has still to be filled, such as their attitude to abiotic (drought, salinity, iron chlorosis) and biotic stresses, as well as the whole genome analysis of the most performing Georgian cultivars, in order to identify the genetic regions related to such valuable traits. A step toward this direction has been performed by Tabidze et al. (2017), sequencing the whole genome of four major Georgian varieties (Chkhaveri, Saperavi, Meskhetian green, and Rkatsiteli) and releasing information useful to understand the complexity of grape genome and for further comparative analysis. Aside from traditional breeding programs, these invaluable resources could be exploited in breeding programs based on the use of New Breeding Technologies (NBTs), by means of genome editing applied to both resistance and, with even more practical advantages, susceptibility candidate genes to abiotic and biotic stresses. In this way, it will be possible to exploit the valuable traits carried by this unique source of genetic variability for new varieties able to meet the challenges awaiting viticulture in the era of climate change.

AUTHOR CONTRIBUTIONS
GDL and SLT conceived the work. MS, VR, and GDL wrote the introduction and genetic variability section. GC wrote the climate section. LR wrote the section on phenotype. SLT wrote the section on resistance to fungal pathogens. FQ wrote the section on susceptibility to phytoplasma diseases. DM, OF, and PAB critically revised the manuscript. All the authors read and approved the final version of the manuscript.

ACKNOWLEDGMENTS
We would like to thank Dr. Lesley Currah for supporting us on English language editing.  Ketskhoveli et al. (1960) and VIVC. Table includes information related to berry color, origin, usage, spreading, genotyping and resistance to plant disease. Synonyms, homonyms and misidentifications are not verified. SSR, simple sequence repeat; SNP, single nucleotide polymorphism; VIVC, Vitis International Variety Catalogue (https://www.vivc.de/).