Study of Inter- and Intra-varietal Genetic Variability in Grapevine Cultivars

Vitis vinifera includes a large number of cultivars that are further distinguished in biotypes and clones, and it is actually hard to differentiate them, even through complex molecular techniques. In this work, the plant materials of 56 putative Sangiovese and 14 putative Montepulciano biotypes, two of the most widespread black-berried Italian cultivars, were collected in different wine-growing areas of Italy distributed in 13 regions, from north to south. Firstly, the samples were analyzed using SSR markers to have proper varietal identification. According to the results, the genotypes belonged to three different cultivars: Sangiovese, Sanforte, and Montepulciano. Subsequently, the samples were investigated using AFLP, SAMPL, M-AFLP, and I-SSR molecular markers to estimate their intra-varietal genetic variability. The DNA marker-based method used turned out to be performing to bring out the geographic differences among the biotypes screened, and it can therefore be considered as a powerful tool available for all the grapevine varieties.


Introduction
Vitis vinifera L. is one of the oldest known fruit-producing crops that, over the centuries, has undergone strong domestication. This process led to obtaining plants with hermaphroditic flowers (self-fertile), good fruitfulness, propagation capacity, and most of all optimal grape quality for fresh consumption or winemaking [1,2]. The taxonomic reference unit for the grapevine is the variety or cultivar (i.e., a contraction of cultivated variety), and it is estimated that the total number worldwide stands between 6000 and 10,000 [3]. Additionally, a grapevine variety has a higher genetic complexity level: it can be further subdivided into biotypes and clones (namely, officially recognized biotypes). A biotype, or clonal line, is generally selected because standing out for some favorable phenotypic outcomes. These intra-varietal special characteristics can be minor changes in morphological traits, such as bunch compactness and canopy thickness, or macroscopic modifications, such as fruit color (e.g., the grey and white variants of Pinot noir) [4][5][6][7]. It is known that such polymorphisms derive from somatic mutations that occur accidentally after reproduction by vegetative propagation [1,8,9]. Moreover, it has been recently ascertained that intra-varietal differences can arise as plant responses to the environment through epigenetic modifications (clone-dependent DNA methylation patterns) that affect gene expression without altering the DNA sequence [10,11].
In Italy, the first wine-producing country in the world [12], Sangiovese represents the main grapevine variety, with a cultivation area of 54,000 Ha in 2017, covering 8% of the total vineyard area [13] and approximately 7.5 million rooted grafts produced in 2020 [14]. This cultivar is the basic grapevine of the major Tuscan D.O.C. (denomination of controlled origin) and D.O.C.G. (denomination of controlled and guaranteed origin) wines, nationally and internationally renowned (i.e., Chianti Classico, Nobile di Montepulciano, and Brunello di Montalcino). Sangiovese has an ancient, well-attested history in the bibliography of the past, and the first reference to this variety as "Sangioveto" has been recently identified in a treaty about agriculture written by Girolamo Gatteschi (also known as Girolamo da Firenzuola) during a period of imprisonment in Florence in 1552 [15]. Before, Soderini was considered the first author who mentioned Sangiovese as "Sangiogheto" in 1590 [16]. Sangiovese, like many other grapevine cultivars [17,18], is known with different local denominations, according to the area of cultivation (e.g., Brunello in Montalcino, Calabrese in Arezzo, Prugnolo gentile in Montepulciano, Morellino in Grosseto, and Pignolo in Siena, just to mention the cases of some Tuscan wine districts) [19]. In addition, Sangiovese is characterized by high morphological variability, a common feature within the most widely cultivated varieties over the centuries [20,21]. It has a broad population of different biotypes [5,22,23] and the highest number of clones officially registered in the Italian Catalogue of Grapevine Varieties (128, plus three further recorded under the name of Prugnolo gentile) [14]. This ambiguity has often caused varietal classification mistakes [24][25][26][27][28][29]. In the past, Sangiovese was subdivided into two main biotypes: "Sangiovese grosso" (with large berries) and "Sangiovese piccolo" (with small berries) due to a different cluster morphology [28,30]. Subsequently, with the coming of DNA analysis, these two main biotypes showed two distinct genetic profiles: the samples belonging to "Sangiovese grosso" were considered as the true-to-type Sangiovese, while "Sangiovese piccolo" was Sanforte, also known with the synonyms of Maiolica, Sangiovese forte, and Sanvicetro [14,31,32]. It is interesting to mention another peculiar case: Prugnolo gentile, traditionally cultivated in the area of Montepulciano (Siena, Tuscany), was registered in the Italian Catalogue of Grapevine Varieties in 1971 as an independent cultivar, but with the introduction of molecular screenings, it turned out to be indisputably Sangiovese [33].
Varietal identification in viticulture has a long tradition. It has always been based mainly on ampelography (from the Greek words ampelos and graphia, meaning vine description) and ampelometry (the measurement of vine organs) [34,35]. More recent observational methods were introduced later, such as biochemical analyses on either phenolic, terpenic, or protein profiles of the grapes [36,37]. However, these laboratory techniques have some limitations because, as already stated, the vines can often undergo changes in the expression of their phenotypic characteristics [38]. Therefore, a molecular approach is essential since only genetic screening can provide an unequivocal fingerprinting and reveal synonyms or mistaken identifications due to empirical hypotheses. For this purpose, microsatellite markers have high discriminative power and are strongly informative [32,39,40], the reason why a standard set of nine simple sequence repeats (SSR) was developed within the Grape-Gen06 European project, and ever since, it has been adopted internationally to discover the identity of a grapevine sample [41]. The use of SSR markers, however, does not allow the detection of intra-varietal differences [6], and thus, other DNA marker-based techniques have been set up to study the genetic variability within Vitis species. The most commonly employed are RAPD (random amplified polymorphic DNA), MSAP (methylation-sensitive amplified polymorphism), RT (retrotransposon-based molecular markers), SNP (singlenucleotide polymorphism), IRAP (inter-retrotransposon-amplified polymorphism), AFLP (amplified fragment length polymorphism), M-AFLP (microsatellites amplified fragment length polymorphism), SAMPL (selective amplification of microsatellite polymorphic loci), I-SSR (inter simple sequence repeat), and ISTR (inverse sequence-tagged repeat), and they can also be combined to obtain much more comprehensive information [42][43][44][45][46][47][48][49][50][51][52].
This research aimed at exploring the genetic variability in a broad pool of biotypes allegedly belonging to Sangiovese, the most widespread, black-berried Italian grapevine cultivars [53], and Montepulciano, which was added as the second most cultivated blackberried variety [53] that shares some important wine-growing districts with Sangiovese. In detail, we collected 78 grapevine samples (70, plus seven reference clones and one reference biotype), empirically categorized as putative Sangiovese (56), belonging both to "grosso" and "piccolo" biotypes, and Montepulciano (14) ( Table 1). The plant material was recovered in 13 Italian regions, from north to south along much of the peninsula (Figure 1). At first, we verified the true-to-type varietal identity using SSR markers. After establishing that the samples clustered in three different cultivars (i.e., Sangiovese, Sanforte, and Montepulciano), we screened them with four additional molecular markers (i.e., AFLP, SAMPL, M-AFLP, and I-SSR) to estimate the intra-varietal genetic differences. Using a method already proven to be performing in Vitis species [54], we wanted to bring out any variations attributable to the geographical origins among the biotypes that may have arisen as an adaptation to the growing environment.

Varietal Identification
According to the results obtained by the SSR marker-assisted screening at 11 microsatellite loci, the 78 samples analyzed can be divided into three variety clusters: 51 Sangiovese  Figure 2 shows the standard bunch at maturity and the standard adult leaf of the three grapevine cultivars revealed by SSR analysis.
By observing the SSR profiles (reported in Table 2), Sangiovese and Sanforte show identical base pair lengths for both alleles in two SSR loci (VVMD5 and VVMD27), and they share 1 SSR allele in 8 SSR loci (VVS2, VVMD7, VrZAG62, VrZAG79, VMC6E1/ISV2, VMC6F1/ISV3, VMC6G1/ISV4, and VMCNG4b9). The two cultivars may appear genetically similar but the first-degree kinship between them can be excluded since there are no shared alleles in one of the tested loci (VVMD28). Looking further at the values in Table 2, Montepulciano shares one allele in eight SSR loci with Sanforte and only five SSR with at least one common locus with Sangiovese.

Intra-Varietal Analyses
The samples belonging to Sangiovese, Sanforte, and Montepulciano grapevine varieties (Table 3)  The data obtained showed that Sangiovese and Sanforte were genetically similar, and it was also found that Montepulciano stands closer to Sangiovese than to Sanforte.

Genetic Diversity
A UPGMA (unweighted pair group method with arithmetic mean) dendrogram based on genetic diversity [57] was obtained using the experimental data ( Figure 3). The dendrogram displays the 78 samples grouped into two main clusters: 1) Sangiovese and Sanforte genotypes; 2) Montepulciano genotypes. In any case, the three grapevine varieties are just as clearly separated; these results are consistent with the previous SSR markerassisted screening (Table 3). Interestingly, several intra-varietal differences related to the area of origin have emerged within each variety cluster. By observing the lower part of the dendrogram, it can be noted that the 15 samples belonging to the Montepulciano cultivar are grouped according to their region of origin. In particular, there is a sharply defined separation between the samples from Abruzzo, where this grapevine variety is predominant with 56% of the total vineyard area [53] (and also the area of origin of the reference clone RAUSCEDO 7), and Tuscany/Marche, where this cultivar is present but less widespread, with only 19% of vineyard area in Marche and less than 0.5% in Tuscany [53].
As intended, the larger cluster is then divided into two sub-clusters: Sangiovese and Sanforte. Sanforte can be discriminated between Tuscan samples and the ones from other regions of Central Italy (Emilia-Romagna, Marche, Abruzzo, Latium). The more complex Sangiovese sub-subcluster brings out pronounced and highly detailed geographical differences. In fact, the 51 samples are split into four main branches: Tuscany, Central Italy, Southern Italy, and Northern Italy. The largest branch is composed of biotypes originating from Tuscany, where this cultivar plays a central role in the regional wine industry and is spread on 64% of the total vineyard area [53]. Within the Tuscan branch, it can be noted that the portion of the province of Grosseto is secluded from the mixed cluster of samples from Siena/Florence. Subsequently, the Central Italy branch (including the regions located in a central position of the Italian peninsula, except for Tuscany) is present and can be partitioned into two main clusters. On one side stands Emilia-Romagna, where Sangiovese is one of the principal varieties linked to Sangiovese di Romagna D.O.C. wine production; on the other side are the samples from other central regions, where Sangiovese is not as strongly widespread (Marche, Umbria, Latium, and Abruzzo). Among the samples from Emilia-Romagna, there is SG-29 (AR), an outlier from Arezzo (a neighboring Tuscan province); the genetic similarity of this outgroup sample with its sub-cluster (SG-31-SG-13) is quite high, more than the average similarity of the whole cluster of Tuscany, a result that is correct both from the statistical and molecular point of view. The Southern Italy branch rightly comprises all the samples from Campania, Basilicata, Calabria, and Apulia. Here too, is inserted an outlier, SG-39 (PT) from Pistoia (Tuscany), which has homogeneity with the samples from the southern regions, where this biotype may have originated. Finally, the Northern Italy branch is divided clearly between the samples from Veneto and those from Trentino Alto Adige/Friuli Venezia Giulia.
The GS matrix estimates and the dendrogram results agree with the principal component analysis (PCA) reported in Figure 4. In the PCA, the first coordinate (C1) captured 46.6% and the second one 37.9% of the variation of the three grapevine genotypes. In particular, Sangiovese samples were well separated from the other cultivars by the coordinate C2, while the coordinate C1 divided Sanforte from Montepulciano, and as for Sangiovese, the samples from Tuscany were distributed according to the geographic origin of the biotypes. As can be seen, the two-dimensional plotting of the centroids showed eight major clusters: Based on the PCA analysis, the most informative markers in sample discrimination were the primer combination of AFLP Pst+AG/Mse+CAA with 31 polymorphic markers (51%), the primer combination of SAMPL As2/Mse+TGG with 32 polymorphic markers (54%), the primer combination M-AFLP I-SSR#02/M+AGG with 26 (55%), and the primer combination of I-SSR (TC) 7 ACGG with 4 polymorphic markers (44%) ( Table S1).
Collectively considering the results, the combination of different plant-specific molecular markers (i.e., AFLP, SAMPL, M-AFLP, and I-SSR) allowed us to discriminate very efficiently the biotypes belonging to the same variety of Vitis vinifera. In particular, the molecular markers used were effective in finding intra-varietal differences since they detect both some repeated regions (different from those of SSR markers) and not-repeated regions adjacent to these repeated regions, thus having a broad view of the grapevine genome. This in-depth genetic characterization has yielded interesting evidence: Sangiovese, Sanforte, and Montepulciano samples were separated quite exhaustively based on their geographical origin (Figures 3 and 4). Genotypes from neighboring areas showed higher genetic similarity compared to the same variety grown in geographically distant regions, highlighting an influence due to environmental stimuli [6,11].  Table 3-Assigned codes.
It is well-known that the grapevines, especially several historically more exploited varieties such as Sangiovese [1,58,59], have peculiar phenotypic plasticity (namely, when a genotype can produce different phenotypes). This ability allows the plant to survive and carry forward fruit maturation even in limiting conditions [58,60], thanks to some adaptation strategies that consist of selective modulation of gene expression [59,61]. For this reason, it is possible to hypothesize that the geographic differences found between the biotypes screened have arisen as an adaptation to the pedoclimatic characteristics of the growing area, and probably also in response to any biotic and abiotic stress tackled by the vines [62]. Once a grapevine biotype has adapted to a specific environment via somatic mutation or epigenetic modifications (which occur without causing changes in the DNA sequence), the phenotype can be permanently altered [59]; the transcriptional regulation activity is transmitted through the cell division and, therefore, can be directly inherited from the mother plant [63]. It is important to note that the performance of a specific cultivar or clone in a viticultural district, defined by the genotype-environment interaction, is a distinctive feature of the wine typicality, being one pillar of the "terroir" concept [10,11,64].
Our findings are in agreement with the results of other research successfully conducted with the same DNA marker-based method on other important grapevine varieties, such as Malvasia Nera di Brindisi/Lecce, Malvasia di Candia, Negroamaro, Primitivo/Zinfandel, Grenache Noir/Garnacha Tinta/Cannonau, and Malvasia Istriana [38,49,65], whose effective identification has always been very confused, like that of Sangiovese biotypes.

Plant Materials
Small portions of a woody branch of 56 putative Sangiovese vines (Vitis vinifera L. subsp. vinifera), belonging both to "Sangiovese grosso" and "Sangiovese piccolo" biotypes, were made available by some farm owners in different wine-growing districts located in 13 different regions along the Italian peninsula ( Figure 1). The 56 Sangiovese samples collected are listed in Table 1 with the local denominations assigned and the sampling site (Italian region, provincial acronym). The list includes six official Sangiovese clones (i.e., five from Tuscany (VCR 4, VCR 108, BF 10, MI-BF 50, and VCR 103) and one from Emilia Romagna, VCR 16) [14], and one biotype (Sanvicetro F59 P6 C2, retrieved in Tuscany and preserved in the vineyard-collection of CREA-Research Center for Viticulture and Enology), that were added as references. Similarly, a total of 14 alleged Montepulciano (Vitis vinifera L. subsp. vinifera) woody branch samples (Table 1) were collected in 3 different regions located in Central Italy (Figure 1; Tuscany, Marche, and Abruzzo). Here too, a reference official clone of Montepulciano (i.e., RAUSCEDO 7, from Abruzzo) was added. Each woody branch was placed in water and sprouted to obtain optimal plant material (fresh young leaves) for subsequent molecular analyses.
The restriction of DNA was performed by EcoRI, PstI, and MseI enzymes with EcoRI, PstI, and MseI adapters (a and b, see Table S1); T4 ligase was used as a ligation enzyme (all the reagents were obtained from New England Biolabs, Ipswich, MA, USA). Preamplification was performed using 5 µL of seven-fold diluted, digested, and ligated DNAs in 20 µL of reaction mixture containing 75 ng of EcoRI+1 (or PstI+1) and MseI+1 primers (one selective nucleotide), 1x PCR buffer (50 mM KCl, 1.5 mM MgCl2, 10 mM Tris-HCl), 10 mM dNTPs, and 1 U of Taq DNA polymerase (Sigma-Aldrich Corp., St. Louis, MO, USA). The AFLP-based and I-SSR analyses were performed using a 6FAM/VIC/NED/PET labeled EcoRI+3 (or PstI+2) primer and an unlabeled MseI+3 primer (three selective nucleotides). Primers and primer combinations are reported in Table S1.
A binary presence or absence (1 vs. 0) matrix was created for AFLP, SAMPL, M-AFLP, and I-SSR markers and for each genotype. Molecular markers were defined by a standard ladder using the GeneMapper software with some reference DNA genotypes and automatically visualized using the software of ABI-3130XL capillary sequencer.

Genetic Similarity
Genetic similarity (GS TOT ) estimates among individuals were calculated in all possible pairwise comparisons using Dice's genetic coefficient by NTSYS software and UPGMA algorithm [38,49]. GS was calculated within (GS W ) and between (GS B ) sample clusters and marker systems (AFLP, SAMPL, M-AFLP, and I-SSR). The cluster analysis of GS (dendrograms, PCA centroids) was performed according to the UPGMA algorithm using the NTSYS software [57]. The molecular profiles of the Sangiovese and Montepulciano clones and the Sanforte biotype (SSR, AFLP, I-SSR, M-AFLP, and SAMPL) were used as a reference only.

Conclusions
The system based on the four AFLP, SAMPL, M-AFLP, and I-SSR molecular markers used turned out to be performing to bring out the geographic differences among the biotypes comprised in this research work. This method can be considered a powerful tool available for all Vitis vinifera cultivars, especially those characterized by high morphological variability. It finds applications where SSR analysis is not sensitive enough, such as biotype intra-varietal screening or if more complex "omics" approaches cannot be used.

Supplementary Materials:
The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/plants11030397/s1, Table S1: Primers and primer combinations: AFLP, SAMPL, and M-AFLP primer combinations and I-SSR primers used to study the intra-varietal genetic variability of Sangiovese, Sanforte, and Montepulciano.