The use of widely targeted metabolite profiling to reveal the senescence changes in postharvest ‘Red Globe’ (Vitis vinifera)grape berries

XU Teng-fei,YANG Xin,ZHANG Meng,GUO Shui-huan,FU Wen-jing,ZHOU Bi-jiang,LlU Yu-jia,MA Hai-jun,FANG Yu-lin,YANG Gang,MENG Jiang-fei

1 College of Horticulture,Northwest A&F University,Yangling 712100,P.R.China

2 College of Enology,Northwest A&F University,Yangling 712100,P.R.China

3 College of Biological Science and Engineering,North Minzu University,Yinchuan 750021,P.R.China

4 Shenzhen Camartina Wine Co.,Ltd.,Shenzhen 518031,P.R.China

Abstract Changes in the metabolites of table grapes (Vitis vinifera) reportedly occur during postharvest senescence.The aim of this study was to determine the metabolomic differences in postharvest table grapes (‘Red Globe’) after being subjected to different senescence periods.To this end,we used widely targeted metabolomics based on ultra-performance liquid chromatography and tandem mass spectrometry.A total of 135 differential metabolites were identified.During postharvest senescence,the levels of most differential flavonoids (e.g.,pelargonidin 3-O-glucoside,quercetin-3-O-glucoside,and cyanidin 3-O-glucoside) and L-aspartic acid decreased,while the levels of phenolic acids (e.g.,trans-4-hydroxycinnamic acid methyl ester) and pantothenol increased.During early and late senescence,the levels of most differential lipids,especially LysoPC,as well as those of nucleotides and their derivatives,such as uridine,decreased and increased,respectively.Collectively,the findings of this study provide fundamental insights into the reasonable control of table grape fruit postharvest senescence and lay a solid foundation for further research.

Keywords:table grape, berry,postharvest senescence,widely targeted metabolites

1.lntroduction

Grapevines are fruit trees which are cultivated worldwide.The global grape plantation area was forecasted to reach 7.45 Mha in 2018 (OIV 2019).Currently,grapes are mainly consumed fresh or processed into raisins,wine,and juice.Most table grapes are not consumed immediately after harvest but are stored and made available in the market at the appropriate time.However,once grape clusters are harvested and separated from the plants,they enter senescence.Indeed,grapes are subject to postharvest senescence,exhibiting characteristics such as severe moisture loss,berry softening,off-flavor development,and decay that are caused mainly byBotrytiscinerea,which reduces the commodity value and suitability of grapes for consumption (Niet al.2016),thereby leading to reduced postharvest and shelf lives.

Fruit ripening and senescence are inevitable and irreversible processes in the plant life cycle (Dinget al.2015).Fruits are generally divided into climacteric and nonclimacteric types based on the patterns of respiration and ethylene production.The ethylene content and respiration of climacteric fruit increase sharply during ripening as CO2evolution.In contrast,non-climacteric fruit do not undergo these changes and normally emit ethylene in markedly decreased levels (Bottonet al.2019).Furthermore,a series of metabolic changes related to carbohydrates,lipids,organic acids,and secondary metabolites takes place in fruit after harvesting (Pedreschiet al.2009).During fruit senescence,the sugar and organic acid contents decrease gradually with the extension of storage time because they are consumed during fruit respiration.The sucrose,fructose,and glucose contents of postharvest apples decline owing to the action of sucrose synthase and invertase (Zhuet al.2013).Sunet al.(2013) showed that the levels of succinic acid,γ-aminobutyric acid,and glutamine increase,but the level of 2-oxoglutaric acid decreases,in postharvestCitrusfruit.Additionally,the organic acid and hydrogen peroxide contents,and the superoxide dismutase and peroxidase activities,are obviously associated with senescencerelated physiological processes,indicating that organic acids could be used to measure the senescence level of postharvestCitrusfruit.During postharvest senescence,the phenolic and lignin contents of fruit decrease and increase,respectively (Caiet al.2006;Zenoniet al.2020).However,it is unclear whether any other metabolites are closely associated with postharvest senescence.

Metabolomics refers to the analysis of all metabolites in an organism and allows for the simultaneous measurement of all metabolites in a given biological system (Dixon and Strack 2003).In recent years,metabolomic approaches have been widely used to explore aging-related mechanisms in fruit,as reported for melons,oranges,pears,cherries,bananas,and blackberries (Bernillonet al.2013;Dinget al.2015;Yuanet al.2017;Karagianniset al.2018;Xuet al.2018;Kimet al.2019;Sainiet al.2019).However,senescence differs depending on the fruit tree species because of chemical composition differences.Thus,the underlying mechanisms of senescence are unique among fruit types.To date,there have been no studies on the dynamic changes of endogenous metabolites within the integrated grape berry during postharvest senescence.

Recently,widely targeted metabolomics technology based on ultra-performance liquid chromatography and tandem mass spectrometry (UPLC-MS/MS) has been used to quantify metabolites in a high-throughput manner.This technology combines the advantages of accurate measurements in terms of targeted metabolomics with the wide metabolite coverage conferred by the non-targeted metabolomics (Chenet al.2013).The aim of the present study was to characterize the metabolic changes during the natural senescence of postharvest grape berries by using a widely targeted UPLC-electrospray ionization (ESI)-MS/MS-based metabolomics approach with the application of multivariate statistical tools.The metabolomics approach used in this study will help in establishing a direct correlation between the senescence characteristics and metabolite expression during the postharvest senescence of table grapes,thereby contributing to the development of new strategies to slow the rate of grape senescence.

2.Materials and methods

2.1.Plant materials and fruit sampling process

Grape clusters of ‘Red Globe’ (Vitisvinifera) table grapes were harvested at commercial maturity from an organic vineyard in Weinan City (Shaanxi Province,China).Grapes of uniform size,shape,and color,and free from disease and mechanical damage,were transported to the laboratory.Each grape cluster (approximately 40 berries) was weighed and labeled,and then stored at 20°C with a relative humidity of 80%.Six clusters were collected at 0,4,8,12,16,20,and 24 d postharvest and weighed every 4 d to calculate the water loss rate.Then,the berries of the six clusters were collected and mixed.Half of them were used immediately for the analysis of berry firmness,and the other half were frozen in liquid nitrogen and immediately stored at -80°C for further metabolome analysis.All experiments were performed in biological triplicates.

2.2.Detection of grape berry firmness

A total of 15 berries were randomly selected and cut along the latitude of the berry forming a kerf of approximately 1 cm2.Then,the probe of a fruit firmness meter (GY-4;Zhejiang Top Instrument Co.,Ltd.,Hangzhou,China) was vertically inserted into the berry kerf to a depth of 10 mm,while avoiding the seeds,to detect the hardness.The measurement results were recorded in N.

2.3.Sample preparation and extraction for metabolome analysis

A total of 30 grape berries randomly selected at 0,12,and 24 d postharvest (recorded as RG0,RG12,and RG24,respectively) were deseeded and ground in liquid nitrogen using a blender for 30 s.A 100-mg sample of the freezedried powder was incubated in a 1.5-mL centrifuge tube with 0.6 mL of 70% aqueous methanol at 4°C for 8 h.The mixture was centrifuged at 10 000×g for 10 min at 4°C and the supernatant liquor was absorbed and filtered prior to UPLC-MS/MS analysis.

2.4.UPLC-ESl-MS/MS-based widely target metabolomics

UPLC-ESI-MS/MS measurements were performed on the Shim-pack UFLC SHIMADZU CBM30A system(Shimadzu,Kyoto,Japan) coupled to the Applied Biosystems 4500 QTRAP (Applied Biosystems,Foster City,CA,USA).Chromatographic separation was achieved on a Waters ACQUITY UPLC HSS T3 C18 (1.8 μm,2.1 mm×100 mm) (Waters Corporation,Milford,MA,USA).Eluent A was ultrapure water with 0.04% acetic acid,and eluent B was acetonitrile with 0.04% acetic acid.Sample measurements were performed with a gradient program that employed starting conditions of 95% A and 5% B.Within 10 min,a linear gradient to 5% A and 95% B was run,and that composition was maintained for 1 min.Subsequently,the composition was adjusted to 95% A and 5% B within 0.10 min and maintained for 2.9 min.In addition,the column temperature was 40°C with a 4-μL injection volume.The metabolites separated by liquid chromatography were entered directly into the mass spectrometer for detection.

The MS conditions were as follows:turbo spray;550°C for source temperature;5 500 V (positive ion mode)/-4 500 V(negative ion mode);50,60,and 30 psi for ion source gas I,gas II,and curtain gas,respectively;multiple reaction monitoring (MRM) mode;and 5 psi nitrogen for collision gas.A specific set of MRM transitions was monitored for each period according to the metabolites eluted within that period.

2.5.Screening and function annotation of differential metabolites

Orthogonal partial least squares-discriminant analysis(OPLS-DA) was used to initially screen for metabolites with differences during postharvest senescence.Differential metabolites could be further screened by combining theP-values or fold changes in the univariate analysis.Differentially expressed metabolites were filtered by variable importance in projection (VIP) ≥1 and absolute log2FC(fold change) ≥1 or ≤-1.The data were log transformed(log2) and their means were centered before OPLS-DA.To avoid overfitting,a permutation test (200 permutations)was performed.

The identified metabolites were annotated using the KEGG Compound Database,and the annotated metabolites were mapped to the KEGG Pathway Database.Pathways with significantly expressed metabolites were mapped and subjected to metabolite set enrichment analysis (MSEA),and their significance was analyzed through hypergeometric testP-values.

2.6.Statistical analysis

Principal component analysis (PCA),OPLS-DA,and hierarchical cluster analysis (HCA) were performed using the R package (www.r-project.org).PCA was used to recombine the original variables into new,mutually independent variables through orthogonal transformations.HCA was performed on the accumulation patterns of metabolites between different samples.

3.Results

3.1.Water loss rate and berry firmness change

The grape berry firmness and water loss rate changes during‘Red Globe’ postharvest senescence are shown in Fig.1.With the extension of senescence time,the water loss rate of ‘Red Globe’ grape clusters gradually increased,whereas grape berry firmness increased during the first 4 d,and then decreased during the final 20 d.The pattern between water loss and firmness fits the features.

Fig.1 Changes in the water loss rate (A) and berry firmness (B) during ‘Red Globe’ grapes postharvest senescence.Error bars represent SD (n=3).Different letters indicate significant differences according to Duncan’s multiple range test (P＜0.05).

3.2.Data quality assessment of the UPLC-ESl-MS/MS results

During analysis,quality control (QC) was performed by inserting the mixture of the sample extracts after each group of 10 test samples,to monitor the reproducibility of the unit operations during the analysis process.The total ion current (TIC) is the spectrum obtained by continuously summing the intensities of all ions in the mass spectrum at each time point.By using TIC,the accuracy and reproducibility of metabolite detection were confirmed according to Appendices A and B,satisfactory accuracy and reproducibility were achieved.In the MRM mode,the ion flux spectrum of each extracted substance can be determined by the multi-substance extracted ion chromatogram (XIC).The various metabolites were detected and shown by the variant colors of the mass spectral peaks.The peak area of each sample was calculated and represented the relative abundance of the corresponding substance (Appendix C).The MultiQuant Software was used to perform an integration and calibrationof the detected substance-specific peaks.

3.3.Qualitative and quantitative metabolites

We performed qualitative and quantitative analyses of the metabolites in the postharvest grape samples based on the KEGG database,MetWare database(MWDB),and MRM.A total of 424 metabolites were detected:57 amino acids and derivatives,31 nucleotides and derivatives,29 organic acids,17 free fatty acids,28 saccharides and alcohols,7 dihydroflavones,5 dihydroflavonols,3 phenolamines,38 phenolic acids,7 glycerol esters,12 anthocyanins,32 flavonoids,52 flavonols,11 flavanols,3 phosphatidylcholines,2 lignans,13 lysophosphatidylcholines,6 lysophosphatidylethanolamines,1 sphingolipid,2 tannins,9 triterpenes,8 alkaloids,9 vitamins,2 coumarins,2 isoflavones,1 plumerane,10 proanthocyanidins,6 stilbenes,and 21 other metabolites (Appendix D).

3.4.Principal component analysis

The PCA showed that the variability in each grape berry sample was small.The samples obtained during the same sampling session had similar metabolic characteristics.Therefore,our metabolomics results were stable and reproducible.Additionally,the separation trends among all treatments were obvious,suggesting significant metabolic differences during postharvest senescence (Fig.2-A).The three senescence periods were characterized by distinct metabolisms,and the effect of senescence time on the metabolism of postharvest ‘Red Globe’ berries was obvious(Fig.2-B-D).

Fig.2 Principal component analysis (PCA) of all identified metabolites during ‘Red Globe’ grapes postharvest senescence.A,overall score scatter plot;B to D,PCA score maps of the RG0 vs.RG12,RG0 vs.RG24,and RG12 vs.RG24,respectively.RG0,RG12,and RG24 indicate 0,12 and 24 d postharvest,respectively.

3.5.OPLS-DA

To further maximize the distinctions between grape samples,we performed OPLS-DA,which is a more conclusive way to find differential metabolites.This method allows for theXmatrix information to decompose intoYcorrelation and irrelevance (OPLS-DA).In this method,the differential variables are extracted by eliminating the irrelevant differences.R2Yrepresents the interpretation rate of the model on theYmatrix,andQ2is the prediction ability of the model on theYmatrix.When the values ofR2YandQ2are closer to 1,the model is more stable and reliable.In general,the model is effective whenQ2＞0.5.In this study,R2YandQ2were greater than 0.99 and 0.50,respectively,in each pair group (Fig.3-A-C),thereby demonstrating that this model was effective.The OPLS-DA model was confirmed through 200 alignment experiments.The horizontal line refers to theR2andQ2of the original model,whereas the red and blue dots represent theR2′ andQ2′ after replacement,respectively.These results showed that the current model was meaningful.Additionally,these differential metabolites could be screened on the basis of the VIP value in the subsequent analysis (Fig.3-D-F).

Fig.3 Orthogonal partial least squares-discriminant analysis (OPLS-DA) scores and permutation verification of all identified metabolites during ‘Red Globe’ grapes postharvest senescence.A-C,scores of the OPLS-DA model with RG0 vs.RG12,RG0 vs.RG24,and RG12 vs.RG24,respectively.D-F,OPLS-DA permutation analysis model verification charts of RG0 vs.RG12,RG0 vs.RG24,and RG12 vs.RG24,respectively.RG0,RG12,and RG24 indicate 0,12 and 24 d postharvest,respectively.

3.6.HCA and volcano plot

To evaluate differences in the characteristics of the ‘Red Globe’ berries postharvest senescence that led to the metabolite changes,HCA was used to analyze intratreatment homogeneity and inter-treatment variability.The extension of senescence increased the metabolite differences (Fig.4-A).The points in the volcano map represent all of the detected metabolites from two different senescence periods of the grape berries.Metabolites with a VIP of ≥1 and log2FC of ≥1 or ≤-1 were selected.Based on this method,no differences were detected for most metabolites.The metabolites that met the aforementioned conditions showed significant differences.Among these differential metabolites,the number of metabolites with decreasing levels was greater than the number with increasing levels for RG0vs.RG12 and RG0vs.RG24,while the number of metabolites with increasing levels was greater than the number with decreasing levels for RG12vs.RG24 (Fig.4-B-D).

Fig.4 Hierarchical cluster analysis (HCA),volcano plot,and Venn diagram of differential metabolites during ‘Red Globe’ grapes postharvest senescence.A,hierarchical cluster of the differential metabolites among 0 (RG0),12 (RG12) and 24 (RG24) d postharvest.B-D,volcano plots of RG0 vs.RG12,RG0 vs.RG24,and RG12 vs.RG24,respectively.E,Venn diagram of the differential metabolites.

3.7.Statistical analysis of differential metabolites

Among the 424 metabolites,135 were identified as differential metabolites during the postharvest senescence of ‘Red Globe’ grapes (Appendix E).These were separated into nine categories:amino acids and their derivatives,nucleotides and their derivatives,lipids,saccharides and alcohols,flavonoids,vitamins,phenolic acids,organic acids,and others (Table 1).A total of 50 differential metabolites were verified in the RG12 samples compared with those in the RG0 samples;the levels of 39 of them decreased,while those of the other 11 increased.A total of 82 differential metabolites were verified in the RG24 samples (relative to RG0);and the levels of 66 decreased while those of 16 increased.In addition,68 differential metabolites were confirmed in the RG24 samples (relative to RG12);and the levels of 23 decreased,while those of 45 increased.It is worth noting that,among all the differential metabolite classes,flavonoids were significantly more abundant during the entire senescence process,while the levels of most other metabolites decreased (Table 1;Appendix E).Furthermore,eight metabolites were common in all three postharvest periods examined,namely,xanthine,trans-4-hydroxycinnamic acid methyl ester,methylp-coumarate,O-phospho-L-serine,pantothenol,(R)-pantetheine,glutathione in its reduced form,and procyanidin A2(Fig.4-E).

Table 1 Numbers of differential metabolites in ‘Red Globe’ grapes that were tested at 0,12,and 24 d postharvest,denoted as RG0,RG12,and RG24,respectively

The metabolites that decreased the most in RG0 wereO-phospho-L-serine,allopurinol,glutathione in its reduced form,peonidin 3,5-O-diglucoside chloride,and 12,13-epoxy-9Z-octadecenoic acid (EODE) in sequence (log2FC＜-2.5);while the metabolite that increased the most was methylp-coumarate (log2FC reached 9.34) (Fig.5-A and B).The log2FC values of 10 decreasing metabolites from RG24 were below -4.00 compared with those of RG0,including kaempferol-3-arabinopyranoside and gallate catechin gallate which reached -13.11 and -10.86,respectively(Fig.5-C and D).Between RG12 and RG24,the metabolites with increased levels exhibited relatively small fold changes(maximum log2FC only 2.66),although the number of metabolites with increased content was greater than the number with decreased content;and the metabolite that decreased the most was gallate catechin gallate (log2FC reached -12.26) (Fig.5-E and F).

Fig.5 Fold change bar diagrams (A,C and E) and variable importance in projection (VIP) score plots (B,D and F) of the top ranked differential metabolites.A and B,RG0 vs.RG12.C and D,RG0 vs.RG24.E and F,RG12 vs.RG24.RG0,RG12,and RG24 indicate 0,12 and 24 d postharvest,respectively.

To further analyze the changes of these differential metabolites during postharvest senescence,we standardized and centralized the contents of all differential metabolites and then performed aK-means clustering analysis,which allowed us to divide the metabolites into six subclasses(Fig.6-A-F).Among them,subclass 5 contained the highest number of metabolites (42 differential metabolites),accounting for approximately one-third of the total.The metabolite content of this subclass decreased gradually as senescence increased (Fig.6-E).Next in order was subclass 4,which contained 28 metabolites.The content of this metabolite subclass mainly decreased from 0 to 12 d postharvest,while there were no significant changes over the next 12 d (Fig.6-E).It is also worth noting that subclass 6,which included pantothenol and 13 other metabolites,exhibited increased metabolite content during the senescence period (Fig.6-F).

Fig.6 K-means clustering analysis of differential metabolites in the six subclasses.RG0,RG12,and RG24 indicate 0,12 and 24 d postharvest,respectively.

3.8.Functional annotation and enrichment analysis of the differential metabolites

To screen and analyze the differential metabolic pathway of‘Red Globe’ grape berries during the fruit senescence periods,the metabolites that differed between RG0 and RG12,RG0 and RG24,and RG12 and RG24 were subjected to KEGG enrichment analysis.Three categories could be assigned to all the metabolites:genetic information processing,environmental information processing,and metabolism.According to the functional annotation results (Fig.7-A-C),the differential metabolites during the senescence of postharvest ‘Red Globe’ grapes were mainly involved in metabolism,especially metabolic pathways and the biosynthesis of secondary metabolites.Some metabolites were also involved in genetic information processing(aminoacyl-tRNA biosynthesis) and environmental information processing (ABC transporters),but both of these types of metabolites accounted for less than 20% of the total.Further enrichment analysis showed some differences in the metabolic pathways involved in the three different senescence phases.For the first 12 d,the metabolites were mainly enriched in purine metabolism;phenylalanine metabolism;pantothenate and CoA biosynthesis;and cysteine,methionine,and caffeine metabolism.Additionally,these metabolites were largely enriched in many other metabolic pathways,including flavonoid metabolism,biosynthesis of secondary metabolites,glycerophospholipid metabolism,and glycosylphosphatidylinositol (GPI)-anchor biosynthesis in the next 12 d (Fig.8-A-C).

Fig.7 Functional annotation of differential metabolites during the postharvest senescence of ‘Red Globe’ grapes.A-C,RG0 vs.RG12,RG0 vs.RG24,and RG12 vs.RG24,respectively.RG0,RG12,and RG24 indicate 0,12 and 24 d postharvest,respectively.The numbers on column indicate the amount and proportion of the differential metabolites,respectively.

Fig.8 Functional enrichment analysis of differential metabolites during the postharvest senescence of ‘Red Globe’ grapes.A-C,RG0 vs.RG12,RG0 vs.RG24,and RG12 vs.RG24,respectively.RG0,RG12,and RG24 indicate 0,12 and 24 d postharvest,respectively.

4.Discussion

Globally,grapes are among the most important fruit species,owing to their economic and nutritional values.Generally,table grape berries are not consumed immediately after harvest.Instead,there is usually a time lapse between harvest and consumption.The loss of firmness or softening of fruits is a very important quality parameter.Berry firmness is determined by the thickness of the outer mesocarp with the opaque colored flesh that contains turgid cells (Du Plessis 2008).The softening is due to cell wall degradation induced by several enzymes that are synergistically activated.Interestingly,grape berry firmness presents a slight rise in the initial stages of postharvest senescence,then drops steeply (Fig.1).The slight rise in berry firmness may be due to the structural changes of turgid cell walls caused by pectate lyase,etc.,which makes their arrangement higher in density and more compact within a certain time-frame at the beginning of harvesting.After that,the hydrolyzation of cell wall polysaccharides finally leads to the loosening of cell walls and loss of cell cohesion that are associated with fruit softening.

During postharvest senescence,grape berries are not only subjected to water loss and physiological softening,but they also undergo a series of chemical metabolism changes,such as changes in sugars,organic acids,proteins,lipids,and phenolic and aroma compound.In the present study,135 differential metabolites were identified from a total of 424 metabolites quantified from ‘Red Globe’ grape berries at 0,12,and 24 d postharvest.These differential metabolites included polyphenolics,lipids,amino acids and their derivatives,and nucleotides and their derivatives.

Polyphenolics constituted the class that included the most differential metabolites identified during the postharvest senescence of ‘Red Globe’ grapes.In grapes,these are generally classified as phenolic acids(e.g.,hydroxybenzoic acid and hydroxycinnamic acid),flavonoids (e.g.,flavanols,flavonols,and anthocyanins),and stilbenes (Menget al.2012).Polyphenols have high antioxidant activity and are associated with a reduced risk of a number of cardiovascular diseases and cancers (Oueslatiet al.2012).Therefore,they are used in functional foods,cosmetics,and medicines (Maisuthisakulet al.2007).In this study,the levels of most flavonoids decreased (Table 1),including those of anthocyanins (cyanidin 3-O-galactoside,peonidin 3-O-glucoside,cyanidin 3-O-rutinoside,and peonidin 3,5-O-diglucoside),flavonols (quercetin-3-Oα-L-rhamnopyranoside,quercetin 3-O-glucoside,and kaempferol 7-O-rhamnoside),and flavanols (gallate catechin gallate and gallocatechin 3-O-gallate).These compounds mainly exist in the grape skins and seeds and greatly affect grape color (Ferrandino and Guidoni 2010).Flavanols are the basic building blocks of proanthocyanidins,and their monomers and oligomers contribute bitterness to the astringency of grape berries and red wine (González-Manzanoet al.2004).Flavonols seem to contribute to the bitterness and color (Obreque-Slieret al.2010).Therefore,the decrease in the anthocyanin and flavonol contents may indicate a decrease in the intensity of grape skin color;however,flavanols may form proanthocyanidins after polymerization (the proanthocyanidin A1 and A2 contents increased during senescence),which could affect grape bitterness and the astringent taste during grape postharvest senescence.Furthermore,the nutritional value of grapes may gradually weaken owing to the decreases in the levels of these flavonoids.

It is worth noting that,unlike the flavonoids,the number of phenolic acid differential metabolites (including methylp-coumarate,trans-4-hydroxycinnamic acid methyl ester,andp-coumaric acid-O-glycoside with increased levels was greater than the number of metabolites with decreased levels.Flavonoids participate in the construction of compounds with a C6-C3-C6carbon framework in their basic structure.These compounds are derivatives of chromanes or chromenes,such as flavans,flavones,flavonols,and anthocyanidins(Yonekura-Sakakibaraet al.2019),whereas phenolic acids are types of aromatic acid compounds containing a phenolic ring and an organic carboxylic acid (Neilson and Ferruzzi 2012).Therefore,during the postharvest senescence of grapes,these flavonoids may be degraded to phenylacetic and phenylpropionic acids (Hollman 2009;Murotaet al.2018) under the catalysis of related enzymes from the grape or some microbe,which leads to a decrease in flavonoid and an increase in phenolic acid contents.

Cellular membrane systems play an important role in cellular and subcellular compartmentalization.In the process of enzymatic browning,cellular membrane systems can prevent contact between substrates and result in enzymatic alterations in the properties of the cell membrane composition during fruit ripening and postharvest senescence (Lurie and Ben-Arie 1983).During ripening and senescence,the desaturation levels of the fatty acyl chains of membrane phospholipids increase,which results in their peroxidation.Zhanget al.(2010) showed that the expression of a few desaturase isomers is closely related to the continuous flux of linoleate and linoleate substrates of the lipoxygenase pathway duringAmygdaluspersicafruit ripening.The initial excision of fatty acyl chains can result in the accumulation of lysophospholipids in dates,which display an early ripening profile.The late accumulation of lysophospholipid degradation products or their derivatives (phosphorylated head groups,free head groups,monoacylglycerols,other lysophosphatidic acids,and lysophospholipid derivatives) in dates results in a late ripening profile (Whitakeret al.2001;Coulonet al.2012;Dibounet al.2015).Therefore,in the present study,at the beginning of the 12 d postharvest period,the levels of all the differential lipids (most of which were unsaturated fatty acids,such as glyceryl linoleate,cis-10-heptadecenoic acid,and 12,13-EODE) decreased.On the contrary,during the second 12-d period,the levels of most differential lipids(most of which were lysophospholipid derivatives,such as LysoPE 18:2 (2n isomer),LysoPC 15:0,and LysoPE 16:0(2n isomer)) increased.In addition,pericarp browning and fruit disease appearance in postharvest longans are correlated with damage to the cellular membrane system,which entails cellular membrane phospholipid degradation and peroxidation of unsaturated fatty acids (Yiet al.2009;Linet al.2016;Wanget al.2018).The degradation and peroxidation of membrane lipids could cause damage to the structural integrity of cellular membranes and result in the disruption of cellular compartmentalization,thereby leading to the enzymatic browning process,owing to the contact of phenolic compounds with peroxidase and polyphenol oxidase (Linet al.2017;Wanget al.2018).Therefore,lipid metabolism also plays an important role in ‘Red Globe’ fruit postharvest softening and browning.

Amino acids are not only the building blocks of proteins but also serve as precursors to some secondary metabolites,such as flavonoids,lignins,and phytoalexins (Yamakawa and Hakata 2010;Jinet al.2019).In the present study,the levels of most of the differential amino acids,including L-pyroglutamic acid,glutathione in its reduced form,oxidized glutathione,L-aspartic acid,and L-glutamic acid,but not 5-aminovaleric acid and L-asparagine anhydrous,decreased during postharvest senescence.Proteolysis has been shown to occur with a concomitant increase in free amino acid levels in the early phase of some postharvest crops(Matsumoto and Ikoma 2012;Yanet al.2019).However,in the present study,this phenomenon was not observed;free amino acid levels may increase within 12 d after harvest.In addition,both the oxidized and reduced glutathione levels decreased during the grape postharvest period.Glutathione and ascorbic acid are multifunctional metabolites playing important roles in the redox balance through the ascorbateglutathione cycle,suggesting that more reactive oxygen species were generated with aging,which would have caused further damage to the grape berry cells.

Finally,the levels of pantothenol,a noteworthy metabolite identified in postharvest grapes,increased with senescence time.Therefore,it can be considered a characteristic component of grape berry senescence.Panthenol(pantothenol) is the alcohol analog of pantothenic acid(vitamin B5) and is,thus,a provitamin of B5.In organisms,it is quickly oxidized to pantothenate.Currently,it is mainly used in healthcare and cosmetic products,because it may suppress the growth of some harmful microbes (Chohnanet al.2014).

Senescence of fruits involves several genetic networks where the phytohormone ethylene plays a key role,together with other hormones,integrating different signals and allowing the onset of conditions favorable for stage progression and organ longevity (Iqbalet al.2017).The free auxin increases during senescence and stimulates ethylene biosynthesis through its inductive action on the expression of the key enzyme 1-aminocyclopropane-1-carboxylic acid synthase (ACS).ABA can be a trigger for ethylene production and influence fruit ripening and senescence(Zhanget al.2009).Independently from ethylene,ABA can also increase all hydrolases to enhance softening (Lohaniet al.2004).Therefore,besides cell wall degradation,these hormones also enhance the fruit respiration rate,which may lead to the metabolite changes that were identified in the present study.However,further research should be undertaken to reveal the exact role of hormones in the postharvest senescence of grape fruit.

5.Conclusion

In the present study,we utilized widely targeted metabolomics using a UPLC-ESI-MS/MS detection platform to analyze the metabolic differences of ‘Red Globe’ grapes under postharvest senescence.Flavonoids included the highest number of differential metabolites,the content of which decreased as the postharvest senescence duration increased;and the same trend was observed for amino acids and their derivatives.Therefore,flavonoid and amino acid metabolism play key roles in grape fruit postharvest senescence.However,the levels of phenolic acids,procyanidins,and pantothenol,which are characteristic constituents of grapes,increased with increasing grape berry senescence duration.The lipid and nucleotide metabolite contents decreased during the early senescence phase but increased during the late senescence phase.A total of 27 flavonoids (e.g.,pelargonidin 3-O-glucoside,quercetin-3-O-glucoside,cyanidin 3-O-glucoside),4 phenolic acids (e.g.,trans-4-hydroxycinnamic acid methyl ester),L-aspartic acid and pantothenol can be considered as potential marker metabolites during the postharvest senescence of ‘Red Globe’ grape fruit.

Acknowledgements

The authors would like to thank Wuhan Metware Biotechnology Co.,Ltd.,China for providing the Metabolome Test Platform.This work was supported by the Natural Science Foundation of China (31801833 and 31801811),the Innovation Capability Support Programs of Shaanxi Province,China (2020KJXX-035),the China Postdoctoral Science Foundation (2019M653771 and 2019T120953),the Fundamental Research Funds for the Central Universities of China (2452019016),and the China Agriculture Research System of MOF and MARA (CARS-29-zp-6).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm