From July to October 2022, grape plants with vigorous growth and healthy appearance were selected, and disease-free berries were randomly harvested during the expanding (July-August), veraison (August-September), and rpening (September-October) stages. Skin-pulp adherence was measured in the laboratory, and paraffin sections were prepared for microscopic observation. The harvested grape skins and flesh tissues were rapidly frozen at -80 °C for subsequent analysis of polysaccharide content and the activity of enzymes related to polysaccharide degradation. The 12 table grape varieties selected for this study included ‘Zaoheibao’ (V vinifera), ‘Qiuhongbao’ (V vinifera), ‘Summer Black’ (V vinifera×V labrusca), ‘Black Balado’ (V vinifera), ‘Jinghongbao’ (V vinifera), ‘Lihongbao’ (V vinifera), ‘Flame Seedless’ (V vinifera), ‘Crimson Seedless’ (V vinifera), ‘Wanheibao’ (V vinifera), ‘Wuhecuibao’ (V vinifera), ‘Thompson Seedless’ (V vinifera), and ‘Hutai No.8’ (V vinifera×V labrusca) All of these berries were collected from the Grape Germplasm Resources Nursery of the Horticultural Station of Shanxi Agricultural University(37°23′N 112°32′E, altitude 830 m).

The tops of individual grape berries were removed, and cylindrical sections (10 mm thick), including both skin and pulp, were taken from the equatorial region. The skin was gently scraped with a razor blade and folded outward by approximately 5 mm. The skin–pulp adherence force was measured using a digital dynamometer (ZMF-3, Weidu, Zhejiang) equipped with a jig (HJJ-001, Edelberg, Zhejiang). The outer rind was clamped using the jig, and the pulp was pulled vertically downward with forceps until complete separation occurred. The maximum value recorded by the dynamometer was taken as the skin-pulp adherence force (N), replicated ten times in total.

The vertical pericarp was cut from the equatorial region of the fruit into 0.5 cm × 0.5 cm × 0.5 cm pericarp squares with pulp, and was fixed using FAA (containing 90 mL 70% alcohol, 5 mL glacial acetic acid, and 5 mL 38% formaldehyde). This was followed by paraffin sectioning, ethanol gradient dehydration, xylene transparency, paraffin embedding, safranin O-fast green staining, and sealing with Canada gum. The sections were then placed under an ordinary light microscope to select the sections for observation and photographed using an Olympus DP71 (Japan) microimaging system.

The cell wall content was determined according to the method described by Zhang et al. (2021). Approximately 0.3 g of peel and pulp tissue was weighed, ground, mixed with 1 mL of 80% ethanol, and heated in a water bath at 95 °C for 20 minutes. After cooling to room temperature, the mixture was centrifuged at 4,000 × g for 10 minutes at 25 °C, and the supernatant was discarded. The pellet was washed once with 1.5 mL of 80% ethanol and acetone, then vortexed for 2 minutes. It was then soaked in 1 mL of 90% DMSO solution for 15 hours to remove starch. After soaking, the sample was centrifuged at 4,000 × g for 10 minutes at 25 °C, and the supernatant was discarded. The resulting pellet was dried at 45 °C to a constant weight to obtain the cell wall material. Weighing 0.1 g of dried cell wall material of cell wall, the cellulose and hemicellulose contents were determined by 72% concentrated sulfuric acid hydrolysis and 2 mol/L hydrochloric acid hydrolysis (Zhao et al., 2021), respectively, in mg·g^-1 FW. The extraction of protopectin, chelator-soluble pectin, and water-soluble pectin was carried out by the method of Qi et al. (2015), and the content was determined by carbazole-sulfuric acid colorimetric method in μg·g^-1 FW. Cellulase, xylanase and polygalacturonase activities were determined by DNS colorimetric method (Lo’ay and EI-Khateed, 2018). Salicin hydrolysis method was used for the determination of β-glucosidase activity (Su et al., 2017). Pectin lyase enzyme activity was determined with reference to the method of Sathiyaraj et al. (2011). Determination of xyloglucan endotransglycosylase and pectin methylesterase activities was carried out by the method of Ye (2014). The DNS colorimetric method (Tian et al., 2020) was used for the determination of β-mannanase activity. The ONPG colorimetric method (Zhang et al., 2021) was used for the determination of β-galactosidase activity. Determination of α-L-arabinofuranosidase activity was carried out by p-nitrophenyl arabinofuranoside method (Li et al., 2023). The results were repeated three times.

Microsoft Excel 2016 was used for data organization; SPSS 19.0 was used for significance analysis, Pearson’s correlation analysis and cluster analysis of the experimental data; Duncan’s multiple comparison Method was used to determine significant differences (p<0.05). XLSTAT, The Unscrambler X 10.4 was used for partial least squares regression analysis; OriginPro 2021, Adobe Photoshop CC 2018 was used for plotting.

In the process of fruit development, the skin-pulp adherence of ‘Flame Seedless’, ‘Qiuhongbao’, ‘Wanheibao’, ‘Jinghongbao’, ‘Hutai No.8’, ‘Crimson Seedless’, and ‘Black Balado’ showed a decreasing trend. ‘Thompson Seedless’, ‘Summer Black’, and ‘Zaoheibao’ showed an increasing-then-decreasing trend, while ‘Wuhecuibao’ and ‘Lihongbao’ showed a decreasing-then-increasing pattern ( Figure 1 ). However, for the five varieties, skin-pulp adherence at maturity was lower than during the expansion stage. The reduction in skin-pulp adherence among different grape varieties ranged from 6.4% to 52.4%, suggesting that peelability may develop progressively during fruit maturation. Statistical analysis classified grape skin-pulp adherence into four significantly distinct levels. Among them, ‘Flame Seedless’ exhibited the lowest skin-pulp adherence. This was followed by ‘Thompson Seedless’, ‘Wuhecuibao’, and ‘Summer Black’, all significantly lower than ‘Qiuhongbao’, ‘Wanheibao’, ‘Jinghongbao’, ‘Hutai No.8’, ‘Lihongbao’, ‘Zaoheibao’, and ‘Crimson Seedless’. During the expansion, veraison, and ripening stages, ‘Black Balado’ had significantly higher skin-pulp adherence than other varieties. Cluster analysis of skin-pulp adherence during ripening divided the ten Eurasian grape varieties into two categories: The first group included easy-peel varieties, including ‘Flame Seedless’, ‘Thompson Seedless’, ‘Wuhecuibao’, ‘Wanheibao’, ‘Zaoheibao’, ‘Jinghongbao’, ‘Lihongbao’, ‘Qiuhongbao’, and ‘Crimson Seedless’. The second group comprised hard-peel varieties, such as ‘Black Balado’ ( Figure 2 ).

During fruit development ( Figure 3a ), the pericarp cell volume of ‘Thompson Seedless’, ‘Summer Black’, ‘Qiuhongbao’, ‘Wanheibao’, ‘Hutai No.8’, ‘Lihongbao’, ‘Zaoheibao’, and ‘Crimson Seedlees’ gradually increased. Their transitioned from a densely packed and intact morphology to a looser arrangement ( Figure 3b ). The pericarp thickness of ‘Summer Black’, ‘Hutai No.8’, ‘Lihongbao’, ‘Zaoheibao’ and ‘Crimson Seedless’ gradually decreased from the expansion stage to the ripening stage. At the expansion stage, pericarp cells of ‘Flame Seedless’, ‘Thompson White’, ‘Qiuhongbao’, ‘Jinghongbao’ and ‘Black Balado’ were loosely arranged. As fruit development progressed, the connections between adjacent cells weakened, and pulp cell separation increased. Meanwhile, the flattened and neatly arranged pericarp cells observed in ‘Wanheibao’, ‘Hutai No.8’, ‘Lihongbao’, ‘Zaoheibao’ and ‘Crimson Seedless’ may be associated with their more robust cell walls.

During fruit development, the cell wall polysaccharide content in the 12 grape varieties showed a decreasing trend ( Figures 4a, b ). Notably, the decline of cell wall material content in the pulp (30.3%~64.8%) was more pronounced than that in the skin (23.9%~51.4%). The reduction in cell wall material content in the pericarp was greater from the expansion stage to the veraison stage (5.5%~49.5%) than from the veraison stage to the ripening stage (7.2%~47.8%). The contents of protopectin and chelator-soluble pectin showed the most rapid decline. Among the 12 varieties, the chelator-soluble pectin content decreased by 87.8%~97.7% in the skin, and by 73.7%~94.6% in the pulp. The reduction in water-soluble pectin content was greater in the skin (84.3%~92.5%) than in the pulp (43.1%~87.8%).

During fruit development, the activities of xylanase, xyloglucan endotransglycosylase, β-mannanase, polygalacturonase, pectate lyase, pectin methylesterase, β-galactosidase, α-L-arabinofuranosidase in the skin and pulp of different varieties exhibited an overall increasing trend ( Figures 5a, b ). Among them, cellulase and β-glucosidase activities showed a smoother change, with an initial increase followed by a decrease. The activities of xylanase, xyloglucan endotransglycosylase and polygalacturonase increased more significantly. From the expansion to the ripening stage, xylanase activity increased by 0.37-2.55 times in the skin and 0.01-1.84 times in the pulp; xyloglucan endotransglycosylase activity by 0.38~2.37 times in the pericarp and 0.42~2.33 times in the pulp; and polygalacturonase activity increased 0.21-2.85 times in the skin and 0.58-2.43 times in the pulp. Pectate lyase activity increased rapidly by 16%-43% before the veraison and more slowly 1%-11% after the veraison. Pectin methylesterase and β-galactosidase exhibited higher activity increases during both the expansion-veraison and veraison-ripening stages. In grape skin, the activities of pectin methylesterase and β-galactosidase increased by 0.69-1.07 and 0.21-0.55 times, respectively, and in grape pulp by 0.29-1.53 and 0.05-1.02 times, respectively. Compared with the increase in the skin (1%~341%), α-L-arabinofuranosidase activity showed a more substantial increase in the pulp (85%~365%). The increased activities of these enzymes related to cell wall polysaccharide degradation likely facilitated polysaccharide breakdown, contributing to the development of easy-peeling traits in grapes.

Pearson correlation analysis was conducted between skin–pulp adherence and 16 polysaccharide-related indices. The results indicated that skin–pulp adherence was positively correlated with the content of cell wall polysaccharides in both skin and pulp, and negatively correlated with the activities of enzymes involved in their degradation ( Table 1 ). Specifically, skin–pulp adherence was highly positively correlated with total cell wall material, cellulose, hemicellulose, protopectin, chelator-soluble pectin, and water-soluble pectin in the skin. It showed significant negative correlations with cellulase, xylanase, xyloglucan endotransglycosylase, β-mannanase, pectin methylesterase, and β-galactosidase, as well as with β-glucosidase, polygalacturonase, and α-L-arabinofuranosidase. However, the correlation with pectin lyase was not significant ( Table 1 ). These findings suggest that low cell wall polysaccharide content in the skin and pulp, combined with high enzymatic activities related to polysaccharide degradation, may contribute to reduced skin–pulp adherence. Moreover, a significant negative correlation was observed between the content of cell wall polysaccharides and the activity of related degrading enzymes indicating that the degradation of cell wall polysaccharides is likely promoted by increased enzymatic activity; cellulase activity showed positive correlations with β-glucosidase, xylanase, xyloglucan endoglycosyltransferase, polygalacturonase, pectin lyase, pectin methylesterase, β-galactosidase, and α-L-arabinofuranosidase, but not with β-mannanase suggesting a potential synergistic relationship between cellulase and these enzymes in facilitating the development of the easy-peeling trait in grapes ( Figure 6 ).

In this study, the cell wall polysaccharide content in the pulp was lower than that in the skin. However, the absolute difference between the correlation coefficients of skin- and pulp-dreived polysaccharide content with skin-pulp adherence was small. This suggests that the decline in skin-pulp adherence may be more influenced by the activity of cell wall polysaccharide degradation related enzymes and other factors that alter polysaccharide levels. Although the content of cell wall polysaccharides varied between the skin and pulp, a universally strong feedback regulation of polysaccharide content on the activity of enzymes related to cellular polysaccharide degradation may not exist.

In the correlation loading plot, the shortest linear distance among the 32 indicators was observed between “17” (pulp cell wall material) and “A” (skin-pulp adherence). This suggests that a reduction in pulp cell wall material may be closely associated with decreased skin-pulp adherence ( Figure 7 ). In contrast, “4” (skin protopectin), “5” (skin chelator-soluble pectin), and “6” (skin water-soluble pectin) were located farthest from “A” in the first and fourth quadrants, suggesting that changes in skin pectin content may exert less influence on skin-pulp adherence.

Based on the 32 measured indices related to cell wall polysaccharides, the skin-pulp adherence of different grape varieties at various stages was evaluated using principal component analysis ( Figure 8 ). “L” (‘Black Balado’-expanding stage) and “K” (‘Crimson Seedless’-expanding stage) were located in the first quadrant and farthest from the origin, indicating that these two varieties exhibited the highest skin-pulp adherence during the expanding stage, consistent with the measured values. Using the lowest observed values on the horizontal and vertical axes as the origin, point “X” (‘Black Balado’-veraison stage) showed lower skin-pulp adherence than points “B” (‘Thompson Seedless’-expanding stage), “F” (‘Wanheibao’-expanding stage), “J” (‘Zaoheibao’-expanding stage), “E” (‘Qiuhongbao’-expanding stage), “G” (‘Jinghongbao’-expanding stage), “H” (‘Hutai No.8’-expanding stage), and “I” (‘Lihongbao’- expanding stage). This suggests that the decrease in skin-pulp adherence may be smaller than the decrease in cell wall polysaccharide content. In addition, cell wall polysaccharide content is a key factor influencing variation in skin-pulp adherence. However, other factors contributing to skin-pulp adherence remain to be further investigated. The variation in skin-pulp adherence among grape varieties may be closely linked to differences in total cell wall material or polysaccharide content during the expansion stage. This may reflect that varieties with higher adherence possess more active polysaccharide synthesis pathways and elevated activities of related enzymes.

In both the skin and pulp, the standardized regression coefficients of cellulase, β-glucosidase, xylanase, xyloglucan endoglycosyltransferase, and β-mannanase were negative, while those of polygalacturonase, pectin lyase, pectin methyl esterase, β-galactosidase, and α-L-arabinofuranosidase were positive. However, polygalacturonase, pectin lyase, pectin methyl esterase, β-galactosidase, and α-L-arabinofuranosidase exhibited negative correlation coefficients with skin-pulp adherence force ( Figure 9 ). This suggests that the five enzymes involved in pectin degradation play a relatively minor role in regulating skin-pulp adherence compared to cellulase, β-glucosidase, xylanase, xyloglucan endoglycosyltransferase, and β-mannanase. Since the linear distances between pulp cellulose, pulp cell wall cell wall material and skin-pulp adherence were the shortest in the correlation loadings plot, the absolute values of the standardized regression coefficient for pulp cellulase (-0.018) was 37.5% of that for pulp xylanase (-0.048) and 20.2% of that for pulp β-mannanase (-0.089). respectively. It suggests that the degradation of hemicellulose, which is involved in cellulose cross-linking, may be one of the factors constituting the decrease in the cell wall material content of cellulose and cell wall.

Principal component scores of 12 grape varieties at different ripening stages were predicted using cell wall polysaccharide content as negative indicators and enzymes related to cellular polysaccharide degradation activities as positive indicators ( Table 2 ). At ripening stage, the peelability ranking was as follows: ‘Flame eedless’, ‘Thompson Seedless’, ‘Wuhecuibao’, ‘Qiuhongbao’, ‘Summer Black’, ‘Lihongbao’ ‘Hutai No.8’ ‘Jinghongbao’, ‘Zaoheibao’ ‘Wanheibao’, ‘Crimson Seedless’, ‘Black Balado’. During the expansion stage, the varieties were ranked as follows: ‘Wuhecuibao’, ‘Flame Seedless’, ‘Thompson Seedless’, ‘Summer Black’, ‘Wanheibao’, ‘Qiuhongbao’, ‘Zaoheibao’, ‘Jinghongbao’, ‘Hutai No.8’, ‘Crimson Seedless’, ‘Lihongbao’, and ‘Black Balado’. ‘Qiuhongbao’, ‘Lihongbao’, and ‘Hutai No.8’ showed lower scores during the expansion stage and higher scores at maturity, while ‘Wanheibao’, ‘Zaoheibao’, and ‘Wuhecuibao’ exhibited the opposite trend ( Table 3 ). This suggests that these six varieties underwent more substantial changes in cell wall polysaccharide content and cell wall polysaccharide degradation related enzymes activity during development, resulting in significant variation in peelability.