Anthocyanin fingerprint of different genotypes of wild grapes ( Vitis vinifera spp. sylvestris (Gmelin) Hegi)

2.1 Plant material

Grapes of 75 different wild grape accessions, corresponding to 75 different genotypes on the basis of previous results [28], collected from different natural populations in Spain and Southern France, and preserved in El Encin Germoplasm Bank (Alcalá de Henares, Madrid, Spain), were sampled in 2008. This germoplasm bank is located 35 km from central Madrid, in a Mediterranean climate environment. Each natural population was identified by two letters and one or two numbers, and each genotype was identified with the population code and an additional number. Tables 1 and 2 summarize the geographic origin of those genotypes and the natural population where they were collected; 31 genotypes came from Northern Spain (Asturias, Cantabria, Castile and Leon, Basque Country and Navarre) and the French Basque Country, and 44 genotypes from Southern Spain (Andalusia, Castilla-La Mancha and Extremadura). Only two plants of each accession, grafted on 110R and trained to cordon Royat, were available, and four to six clusters were collected for each sample. All the samples were collected at the proper time, after measuring Brix and tritatable acidity of must.

Deionized water was purified with a Milli-Q water system (Millipore, Bedford, Massachussets, USA) before use. HPLC-gradient grade acetonitrile was obtained from Merck (Darmstadt, Germany), and analytical reagent grade trifluoroacetic from Sigma-Aldrich (Tres Cantos, Spain). All other chemicals (analytical-reagent grade) were purchased from Panreac (Mollet del Vallès, Spain). Standards of several anthocyanins were prepared from fresh red grape skins, as previously described in the literature [31]; its identity was elucidated by HPLC-MS, using a procedure described previously [32]. Briefly, HPLC/MS analyses were performed using a Hewlett-Packard 1100 system with a PDA UV-Vis coupled to a mass spectrometer furnished with an ESI interface. Stationary phase, mobile phase, oven temperature and injection volume were the same used for HPLC analysis of anthocyanins (see below). MS parameters were as follows: capillary voltage 4000 V; fragmenter ramped from 90 to 120 V; drying gas temperature 325 °C; gas flow (N₂) 12 mL/min. The instrument operated in positive ion mode, scanning from m/z 50 to 2000 at a scan rate of 1.47 s/cycle.

2.3 Anthocyanins extraction

After sampling, a set of 50 berries from each accession was randomly selected and weighted. Grape skins were separated from pulps and seeds and submitted to sequential extraction with different solvents (methanol, 80% methanol, 50% methanol, deionized water, and 75% acetone), as described elsewhere [33]; extracts remained at –20°C until analysis.

2.4 HPLC analysis

A liquid chromatograph consisting of a 600 quaternary pump, a 717 automatic injector, a TC2 controller for a column oven, a 996 photodiode array detector and a Millennium32 workstation (Waters, Milford, Mass., USA) was used for HPLC analysis of anthocyanins. Separation was performed with a Waters Nova-Pak C18 steel cartridge (3.9×250 mm), filled with 5μm particles, and furnished with a Waters Sentry Nova-Pack C18 guard cartridge (20×3.9 mm). Both were thermostated at 55 °C. Water/acetonitrile (95:5) adjusted to pH 1.3 with trifluoraectic acid (solvent A), and water/acetonitrile (50:50) adjusted to pH 1.3 with trifluoracetic acid (solvent B) were used as mobile phases; applying a linear gradient elution (0.8 mL/min flow rate) with the following program: 0 min, 15% B; 20 min, 35% B; 30 min, 50% B; 36 min, 50% B; 41 min, 100% B; 46 min, 100% B; 47 min, 15% B. Samples (20μL) were injected in triplicate. Spectra were recorded between 250 and 600 nm every second, with a bandwidth of 1.2 nm. Samples, standard solutions and mobile phases were filtered before analysis through a 0.45μm pore size membrane filter. HPLC analysis indicates that, after 42 minutes, 16 different anthocyanins were separated and quantified. Table 3 shows name, abbreviation, number of peak and retention time of the different compounds considered.

Total anthocyanins were determined in grape skins extracts, using a procedure described previously [34], using a Boeco S-22 UV/Vis spectrophotometer. The analyses were carried out by triplicate. This data were only used to confirm some evidences, as mentioned in subsections 3.2 and 3.3.

2.6 Statistical analysis

One-way ANOVA and principal component analysis (PCA) were performed using Statgraphics Centurion XIV statistical package (Statistical Graphics Corp., Warrenton, VA, USA).

The absence of acylated anthocyanins is a very strange trait in cultivated grape varieties [18, 19], and it seems that only a reduced number of wild grape genotypes present that trait, which must be associated to genes codifying acyltransferases, that are not present or are not expressed in those genotypes. Eight genotypes (six from six wild populations located in the North of Spain and in the French Basque Country, and other two from two wild populations located in the South of Spain) did not contain acylated anthocyanins. Their detailed anthocyanin fingerprint is displayed in Tables 4 and 5. Moreover, other genetic differences, which are related to the extent of substitution in B ring and to the extent of methylation of B-ring hydroxyl groups can be observed. These differences are shown in Table 6. First, trisubstituted anthocyanins, or Dp-derived anthocyanins (Dp-3-gl, Pt-3-gl and Mv-3-gl) predominate (>70%) in five genotypes (FR-2.2, LE-1.2, NA-3.2bis, SS-3.5bis and J-2.1); on the other hand, disubstituted anthocyanins, or Cy-derived anthocyanins (Cy-3-gl and Pn-3-gl) are prevalent (>40%) in genotypes BI-1.3bis, O-1.9 and H-1.3. This difference should be related to differences in the expression of genes related to B-ring hydroxylation of biosynthetic precursors, that are those genes that codify F3’H and F3’5’H enzymes [36]. Secondly, the expression of genes related to methylation of B-ring hydroxyl groups by O-methyltransferases (OMT) differs among genotypes [37]; thus, methylated anthocynins (Pt-3-gl, Pn-3-gl and Mv-3-gl) varies from 82.32% (LE-1.2) to 36.75% (SS-3.5bis). Differences in B-ring hydroxylation and in methylation of B-ring hydroxyl groups are reflected, at some extent, in the major anthocyanin present in each genotype. Thus, Dp-3-gl predominated in genotype SS-3.5bis, with a low intensity of methylation and a high proportion of trisubstituted anthocyanins, but also in a genotype with a medium to high intensity of methylation and a high proportion of trisubstituted anthocyanins (FR-2.2). On the other hand, Pn-3-gl was the major anthocyanin in genotypes with a high proportion of disubstituted anthocyanins and with a high intensity of methylation (BI-1.3bis, O-1.9 and H-1.3). Finally, Mv-3-gl was more abundant in the other three genotypes (LE-1.2, NA-3.2bis and J-2.1), which present a high proportion of trisubstituted anthocyanins and a high intensity of methylation. Thus, it should be possible to consider three different groups of genotypes lacking acylated anthocyanins. Four genotypes (FR-2.2, LE-1.2, NA-3.2bis and J-2.1) presented a high proportion of trisubstituted anthocyanins and high intensity of methylation; one genotype (SS-3-5bis) contained a high proportion of trisubstituted anthocyanins and low intensity of methylation; finally, three genotypes (B-1.3bis, O-1.9 and H-1.3) presented a high proportion of disubstituted anthocyanins and high intensity of methylation.

To conclude, the absence of acylated anthocyanins implies that acyltransfereases are not present or are not expressed, but this trait is independent of the regulation of the remainder steps of the anthocyanins biosynthetic pathway. Analysis of expression of the F3’H, F3’5’H and OMT genes indicated that there are important differences among genotypes [37]. Thus, the prevalence of the expression of F3’H over F3’5’H would lead to a higher proportion of Cy-derivatives [36].

Most genotypes examined contain acylated anthocyanins, as it could be expected, because previous studies on the anthocyanin fingerprint of cultivated grapes have shown that most cultivars contain acylated anthocyanins [13, 17–22]. Tables 4 and 5 shows the detailed anthocynin fingerprint of those genotypes Nevertheless, the extent of acylation is quite variable, as well as the ratio of acetylated anthocyanins to p-coumarylated anthocyanins. This fact suggests that acylation of monoglucosides is a complex phenomena, controlled by several genes. Recently, two weak QTLs were detected for p-coumaric acid acylation on cv. Grenache parental map, but candidate genes have not been clearly identified [38]. Taking into account all these questions, several subgroups of genotypes may be distinguished, considering four factors: the abundance of trisubstituted (Dp-derived) anthocyanins, the extent of methylation of B-ring hydroxyl groups, the extent of acylation, and the ratio of acetylated to p-coumarylated anthocyanins.

Seven genotypes presented a very high proportion of disubstituted (Cy-derived) anthocyanins; more than 50% anthocyanins derived from Cy. This fingerprint is characteristic of pink and rosé cultivars [19], but the external appearance of berries correspond to that shown by red cultivars; their content of total anthocyanins determined by spectrophotometry was higher than 300 mg/kg grapes (data not shown). It has been described that the transcriptomic analysis revealed that metabolic pathways involved in polyphenol synthesis were significantly altered in genes related to the biosynthesis and transport of phenolics compounds [39]. In this way, expression of MybA1 and UFGT genes were associated with anthocyanin content. Regarding the differences in the anthocyanin profile are more related to the expression of F3'5'H, F3'H and the OMT genes [39] For that reason, our results pointed out that the differences in the anthocyanin profile are more related with the gene expression in the biosynthesis pathway [37]. Table 7 displays a reduced anthocyanin fingerprint of those genotypes. Three different subgroups may be considered. One genotype (CA-6.1) presented a very high proportion of methylated anthocyanins (>80%) and a very low amount of acylation (<5% of acylated anthocyanins), predominating p-coumarylated anthocyanins over acetylated anthocyanins. Two genotypes (CC-1.5 and CO-5.1) contained a medium proportion of methylated anthocyanins (about 55%) and a low amount of acylation (<5 %), but p-coumarylated anthocyanins predominate. Finally, four genotypes (CA-11.4, FR-1.4bis, SS-6.5bis and SS-7.1bis) contained a low proportion methylated anthocyanins (<40%) and usually a low amount of acylation (<15%), predominating acetylated anthocyanins over p-coumarylated anthocyanins, except in the case of SS-7.1bis.

Other nine genotypes presented a relatively high proportion of disubstituted anthocyanins: 30–50% anthocyanins derived from Cy. Of course, the external appearance of berries resembled red cultivars, and the content of total anthocyanins was usually higher than 1000 mg/kg grapes. A reduced anthocyanin fingerprint of those genotypes is displayed in Table 8, and two different subgroups may be considered. The first subgroup includes five genotypes (CO-5.5, H-6.1, NA-2.4bis, SS-3.2bis and VI-2.1bis) presented a medium proportion of methylated anthocyanins (40–60%. Two of them (CO-5.5 and VI-2.1bis) present a low amount of acylation (<10% of acylated anthocyanins), predominating p-coumarylated anthocyanins over acetylated anthocyanins, but the other three genotypes (H-6.1, NA-2.4bis and SS-3.2bis) contained a high amount of acylated anthocyanins (>10%), predominating acetylated anthocyanins over p-coumarylated anthocyanins. The second subgroup contains four genotypes (CA-13.3, CA-13.6, CO-5.4 and H-4.1), which presented a high proportion of methylated anthocyanins (>60%) and p-coumarylated anthocyanins predominated over acetylated ones; nevertheless, the extent of acylation was quite variable: very low for genotypes CO-5.4 and H-4.1, relatively high (>10%) for CA.13.3 and CA-13.6.

Twenty-five genotypes presented a very high amount of trisubstituted anthocyanins (>85%), derived from Dp; Table 9 displays a reduced fingerprint of them. In these genotypes, the extent of methylation was quite variable (between 50 and 95%); also were quite variable the extent of acylation (between 3 and 29%) and the ratio of acetylated to p-coumarylated anthocyanins (from 0.31 to 3.50). To understand differences among this group of genotypes, principal component analysis (PCA) was performed, considering five different variables (the proportions of Dp-derived anthocyanins, methylated anthocyanins, acylated anthocyanins, acetylated anthocyanins and p-coumarylated anthocyanins). Three principal components (PC) with eigenvalues > 0.98, that explain 90.506% of variance, were obtained. PC1, which explains 46.995% of total variance, was mainly affected by acylated, acetylated and p-coumarylated anthocyanins, PC2 (explaining 23.775% of total variance) by Dp-derived anthocyanins, and PC3 (explaining 19.735% of total variance) by methylated anthocyanins. Figure 1 shows that three different groups of genotypes may be clearly distinguished in the plane defined by PC1 and PC 2. A first group (A1), at left, with seven genotypes containing <11% acylated anthocyanins; a second group (A2), in the center, with nine genotypes containing 14–21% acylated anthocyanins, and a third group (A3), at right, with the nine remaining genotypes, containing > 23% acylated anthocyanins. Of course, each group contains genotypes that differ in the proportion of methylated anthocyanins (e.g., genotypes BA-1.1 and CA-9.7 belong to the first group, but contain 60.13% and 93.27% of methylated anthocyanins). Moreover, several genotypes contained a high proportion of p-coumarylated anthocyanins, up to 15%.

OPEN IN VIEWER

Fig.1

Groups of genotypes with a very high amount of Dp-derived anthocyanins (>85%) in the plane defined by principal components PC1 and PC2.

Finally, other 26 genotypes presented a high amount of Dp-derived anthocyanins (70–85%); a reduced fingerprint of them is shown in Table 10. In this group, the extent of methylation was quite variable (between 45 and 85%); also were quite variable the extent of acylation (between 2 and 37%) and the ratio of acetylated to p-coumarylated anthocyanins (between 0.36 and 4.38). PCA was performed to understand differences among genotypes, considering the five different variables above mentioned. Two PCs with eigenvalues > 1.0, that explain 74.892% of variance, were obtained. PC1, which explains 52.971% of total variance, was mainly affected by acylated, acetylated and p-coumarylated anthocyanins, and PC2 by methylated, Dp-derived and acetylated anthocyanins. Figure 2 shows that three different groups of genotypes may be distinguished in the plane defined by PC1 and PC2. A first group (B1), at right, with eight genotypes containing > 22% acylated anthocyanins; a second group (B2), in the center, with eight genotypes containing 13–22% acylated anthocyanins, and a third group (B3), at left, with the ten remaining genotypes, containing <13% acylated anthocyanins. Of course, each group contains genotypes that differ in the proportion of non-methylated anthocyanins (e.g., genotypes CR-1.1 and NA-1.4bis belong to the third group, but contain 78.42 and 44.15% of methylated anthocyanins, respectively).

OPEN IN VIEWER

Fig.2

Groups of genotypes with a high amount of Dp-derived anthocyanins (70–85%) in the plane defined by principal components PC1 and PC2.

In conclusion, four different groups of genotypes containing acylated anthocyanins may be distinguished, Differences into those groups are mainly related with the predominance of Dp-derived or Cy-derived anthocyanins, but also other traits, like the extent of methylation, the extent of acylation and the ratio of acetylated to p-coumarylated anthocyanins differ among genotypes within a group. These differences may be explained by the different expression level of the structural genes and transcriptional factors in the anthocyanins biosynthesis pathway.

Table 11 displays the geographical distribution of five great groups of wild grapes genotypes. As can be noted, there are notable differences on the prevalence of the five groups of genotypes if two geographical groups of genotypes (Northern Spain and Southern Spain) are considered. First, the lacking of acylated anthocyanins is more prevalent in genotypes collected in populations from the Northern regions of Spain and the contiguous French Basque Country (19.35% vs. 4.35% in Southern Spain genotypes). Secondly, the number of genotypes which contain a high (30–50%) or very high (>50%) amount of Cy-derived anthocyanins is slightly higher in samples collected in populations from the Southern regions of Spain (22.73% vs. 19.36% in Northern Spain). Finally, genotypes with a high (70–85%) or very high (>85%) content of Dp-derived anthocyanins are more abundant among those collected in Southern Spain (72.73% vs. 61.29% in Northern Spain genotypes). As it has been pointed out, the 75 genotypes under study have grown under the same conditions. Thus, these data have led us to hypothesize that evolution of the anthocyanin fingerprint of wild grapes has been affected by the climatic conditions of the different environments where wild populations are located, as genotypes from Northern Spain have evolved under Oceanic climate conditions, whereas genotypes from Southern Spain have evolved under Mediterranean climate conditions.

To evaluate the impact of different climatic conditions in Northern and Southern Spain on the anthocyanin fingerprint of wild grapevines, several parameters related to that fingerprint have been compared for genotypes from wild populations of both geographical regions (Table 12). Data point out that there is a considerable range of variation for the five parameters related to anthocyanin fingerprint that have been considered, as could be expected taking in account the different phenotypic groups above described. Nevertheless, there are several important differences between the two geographical groups of genotypes. Fist of all, genotypes from Southern Spain usually contain a higher proportion of methylated, acylated and p-coumarylated anthocyanins than genotypes from Northern Spain. Moreover, in genotypes from Southern Spain the proportion of p-coumarylated anthocyanins is usually higher than the proportion of acetylated anthocyanins, and as a consequence usually present a ratio of acetylated to p-coumarylated anthocyanins lower than 1, as many grape cultivars considered of Spanish origin [17, 20–22]. On the other hand, genotypes from Northern Spain behave opposite; so, the ratio of acetylated to p-coumarylated anthocyanins is higher than 1, as it has been reported in many French cultivars originated in South West France [17, 20–22], a region close to Northern Spain and French Basque Country wild grapevine populations.

Table 12 also displays ranges and mean values of total anthocyanins in genotypes from Southern and Northern Spain. As can be noted, genotypes from Northern Spain present, as a mean, a higher amount of anthocyanins than genotypes from Southern Spain (2384 and 1708 mg/kg grapes, respectively), despite both groups of genotypes present quite similar ranges of variation. As it is well known, the accumulation of anthocyanins in grapes, that take place after veraison, is affected, at a great extent, by day-night thermal contrast [9, 10]. This factor can be considered neutral in our study, as all genotypes grew in the same environment. Thus, differences observed in the anthocyanin content can be considered of genetic nature. The most probable explanation is that genotypes from Northern Spain have evolved in Oceanic climate environments. In these environments grape veraison takes place at the end of summer and day-night thermal contrast is smaller than in the Mediterranean climate environments in which evolved those genotypes collected in Southern Spain. Thus, it is probable that wild grapes in Northern Spain have evolved to accumulate enough anthocyanins capable of attracting birds and other animals to facilitate the dispersion of seeds, despite the limiting weather conditions for anthocyanin accumulation. Thus, when genotypes from Northern Spain grow in a warmer environment, like that of El Encin Gemoplasm Bank, which corresponds to a Mediterranean climate, the accumulation of anthocyanins may be very high, as has been observed in our study. This implies that some genotypes from Northern Spain present genetic traits that allow an intense accumulation of anthocyanins during berry maturation. That is probably the case of genotypes NA-1.1bis, NA-2.4bis, O-1.1bis, S-1.6bis and SS-7.2bis. In these genotypes, the total content of anthocyanins was higher than 3000 mg/kg grapes; only two genotypes from Southern Spain (SE-1.1 and SE-1.5) presented a similar content of anthocyanins. This unique profile in wildgrape berries could be due to a conserved response system for environmental stress that could be driven by cooler and wetter climatic conditions [37]. At the molecular level, such adaptation involves activation of cascade(s) of gene modulations and synthesis of defense metabolites as the anthocyanins. Therefore, anthocyanins often appear transiently at specific developmental stages and may be induced by a number of environmental factors including visible and UVB radiation, cold temperatures and water stress. The subsequent production and localization of anthocyanins in different plant tissues may allow the plant to develop resistance to a number of environmental stresses.