Cultivars fruit quality traits at harvest
To study the sweet cherry fruit cracking responses, we selected seventeen representative and frequently planted cultivars in Greece. Soluble solids concentration (SSC) of sweet cherry fruits at ripening stage (Fig. 1a) ranged from 11.27 (‘Stella’) to 19.63 (‘Skeena’) Brix’s percentage (Fig. 2a). The acidity of cultivars ranged from 0.51 to 1.4 (%, malate), the lowest and the highest acid concentration was detected for ‘Early Bigi’ and in ‘Tsolakeika’, respectively (Fig. 2b). The color indexes, Chroma and Hue angle of cherries ranged from 3.9 (‘Tragana Edesis’) and 6.6 (‘Regina’) to 39.5 (‘Stella’) and 27.0 (‘Stella’), respectively (Fig. 2c, d). The average weight (g) of ten fruits per cultivar ranged from 60.9 to 123.9, the minimum weight was recorded for ‘Early Bigi’ and the maximum for ‘Krupnoplodnaja’ (Fig. 2e). The relative weight of stem, flesh and skin, expressed as partitioned % percentage of whole cherry fruits, were ranged in stem (Fig. 2f) from 0.8 (‘Samba’) to 1.8 (‘Early Bigi’), in flesh (Fig. 2g) from 82.1 (‘Skeena’) to 88.3 (‘Grace Star’) and in skin (Fig. 2h) from 7.5 (‘Sweet Early’) to 12.8 (‘Skeena’).
Texture properties of the skin tissue
Side skin penetration force among cultivars was ranged from 0.17 (‘Early Bigi’) to 0.47 N (‘Skeena’) (Fig. 3a). Suture skin penetration force was ranged from 0.24 (‘Early Bigi’) to 0.55 N (‘Early Star’). Also, the apical skin penetration force was ranged from 0.24 (‘Early Bigi’ and ‘Tragana Edesis’) to 0.48 N (‘Tsolakeika’). Stem skin penetration force were ranged from 0.15 (‘Sweet Early’) to 0.42 N (‘Tsolakeika’). Finally, the total skin penetration force was ranged from 0.2 (‘Early Bigi’) to 0.46 N (‘Tsolakeika’) (Fig. 3a).
Cracking index and fruit water absorption of cherry cultivars
The application of both cracking methods (Christensen and Waterfall) showed that the ‘Regina’ was the most resistant to skin cracking among the cultivars tested (Fig. 3b). Based on Christensen method, eleven cherry cultivars were evaluated as susceptible to cracking, with cracking index higher than 70% (Fig. 3b). These relative cracking-sensitive cultivars include ‘Early Bigi’, ‘Early Star’, ‘Sweet Early’, ‘Carmen’, ‘Grace Star’, ‘Krupnoplodnaja’, ‘Blaze Star’, ‘Aida’, ‘Ferrovia’, ‘Skeena’ and ‘Lapins’. Using the Waterfall method, six cultivars namely ‘Early Bigi’, ‘Early Star’, ‘Sweet Early’, ‘Carmen’, ‘Grace Star’ and ‘Krupnoplodnaja’ with cracking index higher than 40% were identified (Fig. 3b). Furthermore, sharp absorption of water in ‘Grace Star’, ‘Krupnoplodnaja’, ‘Aida’ and ‘Tsolakeika’ where exceeding 4% of the total fruits weight based on the Christensen method (Fig. 3c). Three cultivars (‘Sweet Early’, ‘Krupnoplodnaja’ and ‘Aida’) with water absorption higher than 2% of total fruit weight was found using the Waterfall method (Fig. 3c). Differences between the two tested methods concerning the type of cracking were also documented (Fig. 3d). For instance, to the cultivars ‘Early Bigi’ and ‘Early Star’ the main type of cracking was near the stem based on Christensen method (Fig. 3d); on the contrary, in the same cultivars the main type of cracking was the apical according to Waterfall method (Fig. 3d).
Cultivar-specific primary metabolic profile of skin tissue
As it is challenging to metabolically monitor the factors involved in fruit cracking, we next studied whether the skin-derived metabolite profiling could be an effective tool to understand this physiological disorder, using GC-MS analysis. Sixty-five polar metabolites in the skin tissue of the seventeen sweet cherry cultivars were quantified prior to the application of both cracking assays (Additional file 3: Supplementary Table S3). In terms of chemical composition, and considering all cultivars analyzed, skin metabolomic profile includes soluble sugars (twenty), sugar alcohols (nine), organic acids (fifteen), amino acids (sixteen) and other compounds (five) (Fig. 4). Among the skin tissue of the 17 different cultivars screened, the highest contents of soluble sugars, organic acids, amino acids, total alcohols were recorded in ‘Skeena’, ‘Lapins’, ‘Tsolakeika’, ‘Skeena’ cultivars respectively. In cherry skin tissue, glucose and fructose, malate, g-aminobutyric acid (GABA), and sorbitol represent the main part of sugars, organic acids, amino acids, total alcohols, respectively (Fig. 4, Additional file 3: Supplementary Table S3). Most metabolites (54 % of them) were detected in all cultivars; with some exceptions that were enriched only in specific cultivars. For instance, the highest levels of maltose were detected in ‘Aida’ and ‘Ferrovia’ while mannobiose was exclusively high in two Greek-originated cultivars: ‘Bakirtzeika’ and ‘Tsolakeika’.
The proportions of individual primary metabolites in skin tissue showed significant variation. Six primary metabolites in the skin of ‘Early Bigi’ (oxalate, xylose, arabinose, ribose, mannitol and myo-inositol) as well as in the skin of ‘Samba’ (putrescine, galactose, gluconate, pantothenate, sucrose and lactose) had the highest concentration, respectively. In addition, ethanolamine, rhamnose, fucose, threose and raffinose were most abundant in ‘Stella’. Cultivar ‘Early Star’ was particularly rich in phenylethanolamine, serine, threonine and aucubin (from other compounds) while ‘Bakirtzeika’ displayed high levels of alanine, phosphorate, lactitol and mannobiose. Furthermore, b-alanine, aucubin, maltose, inosose and asparagine had the highest abundance in the skin of cultivars ‘Sweet Early’, ‘Carmen’, ‘Ferrovia’, Tragana Edessis’ and ‘Grace Star’, respectively (Fig. 4, Additional file 3: Supplementary Table S3). Isoleucine, succinate and aucubin were highly accumulated in ‘Krupnoplodnaja’. The cultivar with the highest malate, sorbitol and tyrosine was the ‘Lapins’ while ‘Regina’ showed the highest levels of glycerol and quinate (Fig. 4, Additional file 3: Supplementary Table S3). Much higher levels of proline, glycine, glycerate, arginine, xylitol, arabitol, fucitol and ascorbate were found in the skin of ‘Aida’. Cystothionine, glutaconate, aspartate, 2-oxoglutarate, GABA, pyroglutamate, threonate, cysteine, 2-isopropylmalate and trehalose had the highest concentration in ‘Tsolakeika’. Statistically highest contents of eleven primary metabolites, namely fructose, shikimate, citrate, talose, mannose, glucose, galacturonate, tryptophan, cellobiose, maltitol and maltotriose were measured in ‘Skeena’ (Fig. 4, Additional file 3: Supplementary Table S3).
Skin secondary metabolites profile among cultivars
Another major target of interest in this study was the patterns of metabolites associated with secondary metabolism in the various cherry cultivars. The UPLC–MS/MS analysis confirmed the presence of thirty-five polyphenolic compounds in the skin samples (Additional file 4: Supplementary Table S4); these secondary metabolites correspond to six anthocyanins and twenty-nine other polyphenols (Fig. 5). Generally, the ‘Krupnoplodnaja’ had the lowest content of total polyphenols contrary to ‘Tsolakeika’, which presented significant enrichment in these compounds (Fig. 5, Additional file 4: Supplementary Table S4). Also, the content of skin anthocyanins was higher in ‘Sweet Early’ and lower in ‘Carmen’ (Fig. 5, Additional file 4: Supplementary Table S4).
Significant differences in the polyphenolic composition of the cultivars were noted (Fig. 5). For instance, the polyphenolic compounds vanillin, vanillic acid, naringenin, catechin, procyanidin B1, taxifolin (syn. dihydroquercetin), kaempferol-3-O-glucoside, quercetin-3-O-galactoside and rutin (syn. quercetin-3-O-rutionisde) had the highest concentration in the skin of ‘Early Bigi’. The ‘Sweet Early’ skin contained significantly higher levels of protocatechuic acid, quercetin, cyanidin-3-O-glucoside, cyanidin-3-O-galactoside, cyanidin-3-O-arabinoside and cyanidin-3-O-sambinoside. Also, neochlorogenic acid, cryptochlorogenic acid, chlorogenic acid, epicatechin, kaempferol-3-O-rutinoside and isorhamnetin-3- O-rutinoside were more abundant in skin of ‘Tsolakeika’ (Fig. 5, Additional file 4: Supplementary Table S4). The cultivar ‘Krupnoplodnaja’ exhibited the highest concentration in p-coumaric acid, caffeic acid and ferulic acid. Very high levels of 5-dihydrobenzoic acid, phloridzin and procyanidin B2+B4 were measured in ‘Regina’. In addition, the cultivars ‘Early Star’ and ‘Samba’ had the highest abundance in ellagic acid and peonidin-3-O-galactoside as well as in dihydrokaempferol and arbutin, respectively (Fig. 5, Additional file 4: Supplementary Table S4). The cultivars ‘Carmen’, ‘Tragana Edessis’ and ‘Stella’ contained considerably higher amounts of naringenin, cyanidin-3-O-rutinoside and quercetin-3,4-O-diglucoside (Fig. 5, Additional file 4: Supplementary Table S4).
Physiological traits and skin metabolites in relations to cracking
To elucidate the connection between skin metabolites and cracking in different cultivars, we conducted a correlation analysis. Highly positive correlations were observed between water absorption and skin cracking index assessment using both Christensen and Waterfall methods, the statistical significance was cultivar-specific as depicted at Figure 6a. Furthermore, a strong negative correlation was detected for skin samples between cracking index and two physiological traits skin weight and penetration force around the fruit stem (Fig. 6a). The metabolites sucrose, total soluble sugars (Fig. 6b), galacturonate (Fig. 6b), glycerol, mannitol, myo-inositol (Fig. 6b) were negatively correlated with cracking assessed with both tested methods. Interestingly, the compound fucose (Fig. 6b) and taxifolin (Fig. 6c) displayed the strongest positive correlation with the cracking index, as evaluated by both assays. Also, negative correlation was recorded between the metabolite’s xylose, arabinose, ribose (Fig. 6b), pantothenate (Fig. 6b), phloridzin, epicatechin, procyanidin B1 and procyanidin B2+B4 (Fig. 6c) with cracking index using the Christensen method. On the other hand, positive correlation was detected between asparagine and cracking following Christensen method (Fig. 6b). Using the Waterfall method, negative correlation among fructose, mannose, glucose, trehalose (Fig. 6b), total alcohols (Fig. 6b) and cracking index was determined. In addition, the cracking index based on Waterfall method was positively correlated with phenylethanolamine (Fig. 6b). It was worth noting that the two methods showed a strong positive correlation regarding the cracking index (Fig. 6d).