Combination of RNA-Seq transcriptomics and iTRAQ proteomics reveal the mechanism involved in fresh-cut yam yellowing

The aim of this study was to examine the regulation of transcriptomics and proteomics related to the yellowing of fresh-cut yams after storage. The comparison of yellow fresh-cut yam (YFY) vs. white fresh-cut yam (control) revealed 6894 upregulated and 6800 downregulated differentially expressed genes along with 1277 upregulated and 677 downregulated differentially expressed proteins. The results showed that the total carotenoids, flavonoids, and bisdemethoxycurcumin in YFY were higher than in the control due to the significant up-regulation of critical genes in the carotenoid biosynthesis pathway, flavonoid biosynthesis pathway, and stilbenoid, diarylheptanoid, and gingerol biosynthesis pathway. In addition, the tricarboxylic acid cycle and phenylpropanoid biosynthesis were both enhanced in YFY compared to the control, providing energy and precursors for the formation of yellow pigments. The results suggest that the synthesis of yellow pigments is regulated by critical genes, which might explain the yellowing of fresh-cut yam after storage.

www.nature.com/scientificreports/ exceeding 10%, reads of low-quality sequences, and duplicates. A total of 48,301,943 clean reads were obtained for transcriptome assembly. Q20 is the percentage of bases number which is higher than 20 in the total bases number, and its ratio is 97.79%. Q30 is the percentage of bases number which is higher than 30 in the total bases number, and its ratio is 93.39%. Both Q20 and Q30 were higher than 90%. The GC content is the percentage of G and C bases out of the total number of bases; the GC content was approximately 50%. These results demonstrate that the sequencing level meets the quality needs. Supplementary Fig. S1a shows a correlation heatmap depicting the correlations between the three replicate samples based on the normalized expression results. The correlation coefficients between control samples were 96.3% (control 1 and control 2), 96.5% (control 1 and control 3), and 96.9% (control 2 and control 3). For the YFY samples, the correlation coefficients were 97.1% (YFY1 and YFY2), 97.0% (YFY1 and YFY3), and 97.5% (YFY3 and YFY2). The correlations indicate high sample repeatability. However, low correlation coefficients were obtained between the control and YFY, indicating that yellowing had a large effect on the samples. Principal component analysis (PCA) was used to explore the dataset and reveal hidden data patterns. After dimension reduction analysis, relative coordinate points were observed in the principal component plane. The distance between sample points represents the degree of similarity between the samples, with a shorter distance corresponding to greater similarity. In Supplementary Fig. S1b, the points of YFY samples are clustered together, indicating that YFY had the same characteristics. In contrast, the points of the control and YFY are scattered with respect to each other in Supplementary Fig. S1b with large separation distances, indicating that the control samples (control 1, control 2, and control 3) were significantly different from the YFY samples (YFY1, YFY2, and YFY3). More importantly, the process of yellowing greatly changed the characteristics of fresh-cut yam.
Protein quality control. The high-resolution mass spectrometry results were statistically and qualitatively analyzed to reflect the quality of the detection results and data. A total of 866,811 spectra were produced from the iTRAQ experiments. By analyzing these spectra, we identified 175,705 known spectra, 45,580 peptides, 23,092 proteins, and 8459 protein groups. The spectrum utilization (i.e., the percentage of identified spectra out of the total spectra) was approximately 20% ( Supplementary Fig. S2a), demonstrating that the chosen database was appropriate. The distribution of peptide length is shown in Supplementary Fig. S2b. Most peptides contained 5-20 amino acids, indicating that the enzymes had been sufficiently hydrolyzed. Supplementary Fig. S2c shows the distribution of protein mass. Proteins with masses of 1-21 kDa were the greatest in number (2251) followed by proteins with masses of 21-41 and 41-61 kDa. Peptide matching errors were within ± 20, and most of the errors were close to 0 ( Supplementary Fig. S2d).
The results of the protein repetitive analysis are shown in Supplementary Fig. S2e,f. The coefficient of variation (CV) values for the control and YFY groups were both less than 0.1. The proportion of variation level between 10 and 20% accounted for the major part, demonstrating the high repetition in each group. The PCA results are shown in Supplementary Fig. S2g. Based on the score plot, the YFY samples were considerably different from the control samples; the YFY samples plotted on the negative side of the X axis, while the control samples plotted on the positive side. The results indicate that the YFY samples were quite different from the control samples, which can be attributed to the important effects of yellowing.

Analysis of DEGs and DEPs.
The software RNA-Seq by Expectation-Maximization (RESM) was used to analyze the expression levels of genes and transcripts using transcripts per million as a quantitative index. The number of transcripts was taken as the unit of calculation instead of the number of spliced segments. Factors such as the transcript length and the number of genes expressed in samples were considered comprehensively. A total of 21,906 transcripts were expressed in the control and YFY samples (Fig. 2a). DESeq2, a differential expression algorithm, was used for functional analysis to dig DEGs in different transcriptome groups. To generate resource of DEGs associated with the yellowing of fresh-cut yam, three independent biological replicates were produced for the transcriptome analysis of each group to determine the changes in gene expression for control and YFY. Transcripts found by DESeq2 with adjusted P values less than 0.05 were categorized as differentially expressed. Volcano plots were drawn to express the abundances and distributions of DEGs and DEPs. As shown in Fig. 2b, the distribution of DEGs and DEPs was in screening threshold dimensions. The abscissa was the multiple change value of DEGs/DEPs between two samples, which was obtained by dividing the expressions of the YFY and control samples. The ordinate was the P value of the DEGs/DEPs. A higher P value corresponds to a more significant differential expression. The abscissa and ordinate values are presented logarithmically in  Fig. 2b represents a specific gene/protein; the red points represent significantly upregulated genes/proteins, the green points represent significantly downregulated genes/proteins, and gray points represent non-significantly different genes/proteins. After mapping all the genes, we could conclude that the green points  GO and KEGG pathway enrichment analysis of DEGs. We used Web Gene Ontology (WEGO) software for Gene Ontology (GO) functional classification to understand the distribution of genes at a macro level (Fig. 3a). According to the biological process, the DEGs that mapped to "cellular process" and "metabolic process" constituted a high proportion in yellowing fresh-cut yam, revealing high metabolic activity. This might reflect the biosynthesis and accumulation of bioactive secondary metabolites in yellowing fresh-cut yam. Two terms of "membrane" and "membrane part" occupied the highest part in cellular component category. Most of the DEGs were annotated as "binding" and "catalytic activity, " suggesting high catalytic activity in yellowing fresh-cut yam. Additionally, many genes were found to be related to "single-organism process" in the biological process area; "cell" and "cell part" in the cellular component; and "transporter activity" in the area of molecular function. In contrast, few genes were classified into the "cellular component organization or biogenesis, " "organelle part, " and "structural molecule activity" groups ( Fig. 3a). Pathway enrichment analyses of the DEGs based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database were performed to study the gene interactions with each other during biological functions. Figure 3b shows a scatter plot for the top 20 KEGG enrichment results. Most genes were clustered in the phenylpropanoid biosynthesis and plant hormone signal transduction categories. The expressions of DEGs related to flavonoid biosynthesis and amino sugar and nucleotide sugar metabolism were higher in the YFY group compared to in the control group. Supplementary Table S1 shows the annotation pathways of DEGs in YFY vs. control. Of the total 63,357 annotated genes, 13,694 DEGs were produced, which enriched in 127 pathways, while 25 pathways were significant enriched. The main pathways (e.g., phenylpropanoid biosynthesis; flavonoid biosynthesis; and other glycan degradation pathways) affiliated with metabolism in the first category played an important role in the comparison of the YFY and control groups. These results demonstrate that different genes were produced and expressed, inducing the formation of yellow pigments in YFY. Figure 4a shows a visual overview of metabolism obtained using MapMan. This figure shows the different functional categories that passed the cutoff (q value < 0.05 q and greater than two-fold change) for differential expression. In Fig. 4a, red represents upregulated genes, while blue represents downregulated genes. Compared to the control samples, the YFY samples exhibited higher expressions of genes related to ascorbate and glutathione metabolism, OPP, TCA and amino acids but lower expressions of genes related to light reaction and starch metabolism. The secondary metabolism contained numerous genes related to the synthesis of flavonoids, phenylpropanoids, and phenolics. The YFY group exhibited higher expressions of most genes related to flavonoids, phenylpropanoids, and phenolics. Genes in the cell wall and lipids showed different changes; some were upregulated, while some were downregulated. The results show that the yellowing of fresh-cut yam led to the accumulation of flavonoids, phenylpropanoids, and phenolics along with the degradation of starch and tetrapyrrole. The transformation of these substances may be an important reason for the yellowing.

RT-PCR analysis.
To further verify the results from the RNA-Seq work, seven transcripts of each sample were selected for quantitative RT-PCR analysis. The IDs of these genes are as follows: TRINITY_DN24846_ c0_g2 (phenylalanine ammonia lyase), TRINITY_DN14388_c0_g1 (phenylalanine ammonia-lyase-like), TRIN-ITY_DN24631_c0_g1 (4-coumarate-CoA ligase-like 1), TRINITY_DN24631_c0_g3 (4-hydroxycinnamoyl-CoA ligase 3), TRINITY_DN20568_c1_g1 (cytochrome P450 CYP73A100), TRINITY_DN17620_c0_g1 (trans-cinnamate 4-monooxygenase), and TRINITY_DN24846_c0_g1 (PAL5). These genes were found to be differentially expressed in the transcriptome of yellowing fresh-cut yam. The results of the RT-PCR analysis are shown in Fig. 4b. The trends in the expressions of all the unigenes from quantitative RT-PCR were consistent with the RNA sequencing analysis (Fig. 4b). These results demonstrate that the transcriptomic profiling data accurately reflected the response of fresh-cut yam to yellowing.

GO and KEGG pathway enrichment analysis of DEPs.
According to certain criteria (e.g., function, expression level, and expression differences), the protein list was obtained and analyzed. Some proteins related to the yellowing of fresh-cut yam were extracted, and their functions and expressions were studied.
GO enrichment was used to analyze the key cellular process between the control and YFY samples. The distribution of proteins and their functional classifications are shown in Fig. 5a. Under the category of biological www.nature.com/scientificreports/ www.nature.com/scientificreports/ process, the majority of the DEPs were classified as "metabolic process" and "cellular process," response to stimulus and regulation of biological process. The "cell, " "cell part, " and "organelle" categories were predominant www.nature.com/scientificreports/ within the cellular component category, and most genes were annotated in the "catalytic activity" and "binding categories. The comparative transcriptomic and proteomic data show that several cellular processes, including "metabolic process, " "cellular process, " and "catalytic activity" were affected by changes in key regulators at both the transcript and protein levels in our fresh-cut yam samples. Compared to the control samples (Fig. 5b), the YFY samples were enriched in most DEPs related to flavonoid biosynthesis; flavone and flavonol biosynthesis; phenylalanine biosynthesis; and stilbenoid, diarylheptanoid, and gingerol biosynthesis. These results reveal that different proteins were produced and expressed, and the enrichment in proteins involved in different pathways was basically consistent with the transcriptome results.  www.nature.com/scientificreports/ Transcriptome and proteome correlation analysis. The Venn diagram in Fig. 6a presents the correlations in mRNA and protein differences among samples. The expression patterns of related molecules were studied at the transcription and protein levels. The analysis identified 13,790 DEGs and 1954 DEPs. Among which, 1012 genes were differentially expressed, and their corresponding translated proteins were also differentially expressed. Figure 6b shows the results of the GO correlation analysis, indicating that correlated proteins and genes were annotated to three biological processes: "metabolic process, " "cellular process, " and "single-organism process. " The correlated proteins and genes that mapped to "cell, " "cell part, " "organelle, " "membrane, " and "membrane part" constituted a high proportion of proteins and genes in the cellular component category. Furthermore, most of these correlated proteins and genes were annotated to the "binding" and "catalytic activity" molecular function categories. Indeed, as shown by the integrated comparative transcriptomic and proteomic data, several cellular processes were affected by changes in key regulators at both the transcript and protein levels in fresh-cut yam during yellowing, resulting in high metabolic activity, catalytic activity, and the accumulation of bioactive secondary metabolites in yellowing fresh-cut yam. KEGG enrichment of correlation data are shown in Fig. 6c. The correlated proteins and genes were both significantly annotated to flavonoid biosynthesis; phenylpropanoid biosynthesis; alpha-Linolenic acid metabolism; stilbenoid, diarylheptanoid, and gingerol biosynthesis; fatty acid degradation; ubiquinone and other terpenoid-quinone biosynthesis; phenylalanine metabolism; flavone and flavonol biosynthesis; and tyrosine metabolism. The P values of all these pathways were less than 0.1. The changes in these pathways might be critical factors in the yellowing of fresh-cut yam.

Discussion
Vegetable color is generally accepted as one of the most important quality parameters and plays a decisive role in evaluating the quality of the vegetable product at the point of sale. In the present study, storage at 25 °C for 24 h decreased the L* and a* values but increased the b* value of fresh-cut yam, indicating yellowing (Fig. 1). According to our preliminary results, we extracted the yellow pigment in yellowing fresh-cut yam and identified its components. The yellow pigment was found to be composed primarily of bisdemethoxycurcumin, although other unknown substances were also present. To understand the yellowing mechanism of fresh-cut yam after storage at a molecular level, we compared the transcriptome and proteomic profiles of YFY and control samples. The important aspects of the physiological process related to yellowing are discussed below. The tricarboxylic acid cycle is the basic metabolic pathway shared by most organisms. The tricarboxylic acid cycle is the central hub of metabolism, participating in both catabolism and anabolism 15 . The tricarboxylic acid cycle also provides precursors for biosynthetic pathways 16 such as the synthesis of acetyl-CoA, which has been implicated in the synthesis of secondary metabolites including flavonoids (Fig. 7a), stilbenoids, and isoprenoids. Acetyl-CoA is synthesized by two enzyme systems, the pyruvate dehydrogenase complex and acetyl-CoA synthetase systems 17 . In this study, pyruvate dehydrogenase complex (TRINITY_DN13442_c0_g1) and acetyl-CoA synthetase (TRINITY_DN19742_c0_g1) were respectively upregulated by 2.8 and 10.5 times in YFY compared to the control; the expressions of proteins TRINITY_DN15840_c2_g1_i7_m.43526 and TRINITY_DN13441_ c0_g1_i1_m.148133 were also upregulated, resulting in the enhancement of acetyl-CoA metabolism. In addition, acetyl-CoA was carboxylated by acetyl-CoA carboxylase to form malonyl-CoA, the substrate for the biosynthesis of bisdemethoxycurcumin. The expression of acetyl-CoA carboxylase (TRINITY_DN9689_c0_g1) in YFY was 6.8 times higher than in the control. Therefore, the expressions of crucial genes related to acetyl-CoA biosynthesis and accumulation were significantly enhanced, providing precursors for the formation of yellow pigments (Fig. 7a-c). Furthermore, the tricarboxylic acid cycle can also meet most requirements for cell energy via the complete oxidation of acetyl-CoA. The activity of citrate synthase is fundamental to combine acetyl-CoA with oxaloacetate to form citrate 15 . In this study, the expressions of genes and proteins involved in citrate synthase (TRINITY_DN6100_c0_g1 and TRINITY_DN26760_c1_g2_i5_m.177726) were both significantly higher in YFY samples than in control samples. Therefore, the crucial genes/proteins involved in the oxidation of acetyl-CoA were upregulated, providing energy for the formation of yellow pigments.
The branches of the phenylpropanoid biosynthesis pathway give rise to thousands of compounds 18 . As the beginning of metabolic pathway of each branch (Fig. 7d), the phenylpropanoid biosynthesis pathway can provide precursors for the synthesis of yellow pigments. P-coumaroyl-CoA and cinnamoyl-CoA are the critical precursors of flavonoid biosynthesis and stilbenoid, diarylheptanoid, and gingerol biosynthesis 19,20 . The initial three steps in the pathway are catalyzed by phenylalanine ammonia lyase, cinnamate 4-hydroxylase, and 4-coumaroyl CoA-ligase, which are mandatory and provide the basis for all subsequent branches and resulting metabolites 21 . Phenylalanine ammonia lyase is one of the most well studied enzymes in the biosynthetic pathways of secondary metabolism. Phenylalanine ammonia lyase catalyzes the first step in the general phenylpropanoid pathway along with the deamination of phenylalanine to yield trans-cinnamic acid. Subsequently, cinnamate 4-hydroxylase and 4-coumarate CoA ligase catalyze the conversion of cinnamic acid to p-coumaroyl-CoA and 4-coumaroyl-CoA, respectively, which serve as intermediates and precursors for different phenylpropanoid compounds including flavonoids 22,23 and curcuminoids. In this study, the expression levels of phenylalanine ammonia lyase (TRINITY_ DN24846_c0_g1), cinnamate 4-hydroxylase (TRINITY_DN20568_c1_g1), and 4-coumaroyl CoA-ligase (TRIN-ITY_DN19105_c0_g1) were significantly higher in YFY compared to the control. The DEGs can be correlated to the changes in the protein levels of phenylalanine ammonia lyase (TRINITY_DN27047_c0_g2_i5_m.279346), cinnamate 4-hydroxylase (TRINITY_DN20568_c1_g1_i13_m.285108), and 4-coumaroyl CoA-ligase (TRIN-ITY_DN27047_c0_g2_i5_m.279346). This result indicates that phenylpropanoid biosynthesis was enhanced during storage, consistent with the high level of yellow pigments (Fig. 7a-c). The changes in the expressions of genes and proteins involved in phenylpropanoid biosynthesis were proportional to the degree of yellowing, further verifying that phenylpropanoid compounds play important roles in the yellowing of fresh-cut yam. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Flavonoids are a series of natural compounds that are widely distributed in plants and play important roles in diverse biological processes. For example, flavonoids impart various colors as they are major pigmentation factors 24 . In this study, the level of total flavonoids was analyzed in control and YFY samples (Fig. 7a). The YFY samples had a higher flavonoid content (0.22 mg/mL) than the control (0.19 mg/mL), indicating that total flavonoids increased during yellowing. Chalcone synthase, chalcone isomerase, and flavonoid 3′,5′-hydroxylase, are key enzymes in flavonoid biosynthesis and control the flux of flavonoids. In YFY, chalcone synthase (TRINITY_ DN15220_c2_g1), chalcone isomerase (TRINITY_DN21557_c0_g1), and flavonoid 3′,5′-hydroxylase (TRINITY_ DN15219_c0_g1) were respectively upregulated by 37, 25, and 16 times compared to the control, in agreement with the observed changes in protein levels. The expressions of chalcone synthase (TRINITY_DN15220_c2_g2_ i4_m.227203) and chalcone isomerase (TRINITY_DN21557_c0_g1_i4_m.293684) in YFY were also upregulated by 4 and 7 times compared to the control, respectively. Flavonoid 3′-hydroxylase is also considered to be critical and rate-limiting enzyme. In this study, flavonoid 3′-hydroxylase (TRINITY_DN24846_c0_g1) was upregulated www.nature.com/scientificreports/ by 28 times in YFY compared to the control. The upregulated expressions of chalcone synthase, chalcone isomerase, flavonoid 3′,5′-hydroxylase, and flavonoid 3′-hydroxylase correspond directly to the increased content of total flavonoids in YFY compared to the control. Flavonoids such as eriodictyol and naringenin have been shown to lead to the yellowing of fresh-cut Chinese water chestnut 25,26 . However, the specific flavonoid substances responsible for the yellowing of fresh-cut yam merit further study. The characteristic yellowish color of turmeric is attributed to the presence of polyphenolic pigment curcuminoids. The stilbenoid, diarylheptanoid, and gingerol biosynthesis pathway is responsible for the biosynthesis of curcumin, demethoxycurcumin, and bisdemethoxycurcumin, which has an intense yellow color. In the present study, bisdemethoxycurcumin was not detected in the control, while YFY contained 1.507 μg/mL bisdemethoxycurcumin (Fig. 7c), and its content was proportional to the degree of yellowing. This further verifies that bisdemethoxycurcumin plays an important role in the yellowing of fresh-cut yam. DEGs/DEPs were significantly enriched in the stilbenoid, diarylheptanoid, and gingerol biosynthesis pathway. Feruloyl-CoA, an intermediate of curcuminoids, was formed under the catalysis of shikimate O-hydroxycinnamoyltransferase, 5-O-(4coumaroyl)-d-quinate, 3′-monooxygenase, and caffeoyl-CoA O-methyltransferase. In this study, the relative expressions of shikimate O-hydroxycinnamoyltransferase (TRINITY_DN22259_c0_g1, TRINITY_DN12166_ c0_g3_i1_m.298158) and 5-O-(4-coumaroyl)-d-quinate 3′-monooxygenase (TRINITY_DN19359_c4_g1, TRINITY_DN19359_c4_g1_i3_m.150690) were upregulated in YFY. Moreover, the expression of caffeoyl-CoA O-methyltransferase (TRINITY_DN24830_c0_g2, TRINITY_DN21539_c3_g1_i3_m.294149) was approximately three and seven times higher in YFY than in the control. Curcumin synthase (TRINITY_DN8031_c0_g1) catalyzes formation of bisdemethoxycurcumin from p-coumaroyl-CoA, which was 1.8 times higher in YFY than in control. The synthesis of bisdemethoxycurcumin was enhanced in YFY compared to the control, indicating that these compounds were partly responsible for the yellowing of fresh-cut yam; this conclusion is supported by our recent results indicating that the yellow pigment was composed primarily of bisdemethoxycurcumin (73.7%) 6 .
Fruit color is related to the accumulation of carotenoids and other pigments, which play a major role in the yellow and orange pigmentation of maize kernel 27 . Due to the strong effect of carotenoids on fruit color, changes in the content of total carotenoids were analyzed in this study (Fig. 7b). The carotenoid content in the control (0.04 mg/g) was lower than that in YFY (0.06 mg/g), indicating that the content of total carotenoids increased during the yellowing of fresh-cut yam. Phytoene synthase, the first catalytic enzyme in carotenoid synthesis, was previously reported to control the synthesis of carotenoids in chromoplasts and determine the colors of mature pepper fruits 28 . In the present study, the expressions of phytoene synthase gene and protein (TRINITY_DN6245_c0_g1 and TRINITY_DN19316_c0_g1_i2_m.256713, respectively) eight and three times higher in YFY compared to the control, respectively, resulting in a significantly higher content of phytoene in YFY compared to control. Phytoene acts as a precursor to the formation of lutein, which is bright yellow. Lutein could be formed via the catalysis of α-carotene with beta-ring hydroxylase 29 . In this study, the expression of beta-ring hydroxylase (TRINITY_DN26312_c0_g1) in YFY was 12 times higher than in the control. Lycopene is the critical intermediate in the formation of lutein. 15-Cis-phytoene desaturase (TRINITY_DN18297_c0_g1, TRINITY_DN18297_c0_g1_i10_m.312058) and prolycopene isomerase (TRINITY_DN27028_c0_g1), which were upregulated in YFY compared to the control, catalyzed the formation of lycopene. These results suggest that the carotenoid biosynthesis pathway was enhanced in YFY, which might be critical for the accumulation of yellow substances.
According to the analysis above, in addition to bisdemethoxycurcumin and flavonoids, carotenoids might also contribute to the yellowing of fresh-cut yam. Based on the energy and precursors provided by the tricarboxylic acid cycle and phenylpropanoid biosynthesis, the yellow pigments were formed by the flavonoid biosynthesis, stilbenoid, diarylheptanoid, and gingerol biosynthesis, and carotenoid biosynthesis pathways (Fig. 7d).
Uniformly sized yams without physical damage or disease were selected for experiment. The fresh yams were washed, peeled using a sharp, stainless-steel peeler, and chopped into small slices with thicknesses of 5 mm. The slices were then disinfected using 100 mL of NaClO solution (50 mg/L, pH 6.5) for 2 min. Gauze was used to wipe the water off the surface of each yam slice. The fresh-cut yam slices were packed in polyethylene plastic bags (70 ± 0.5 g) and preserved at 25 °C at 97% humidity for 24 h. The yams stored for 0 h were used as the control. After storage, the yam turned yellow (YFY). All materials were stored at − 80 °C until further analysis or used immediately.
Color analysis. The external color of fresh-cut yam was measured using a chroma meter (Konica Minolta R-400 colorimeter; Minolta Co., Osaka, Japan), and measurements were taken on opposite sides of each slice. L* describes lightness (L* = 0 indicates black, L* = 100 indicates white), a* describes red-green intensity (a* > 0 indicates red, a* < 0 indicates green), and b*describes blue-yellow intensity (b* > 0 indicates yellow, b* < 0 indicates blue). The instrument was calibrated by a standard white and black board, in accordance with CIE specifications. 3 points were taken to be measured in each side of yam, and the mean for three pieces of fresh-cut yam in each replication.
Analysis of total flavonoids in fresh-cut yam. The content of total flavonoids was investigated using a commercial kit (Solarbio, Beijing, China). Frozen tissue (0.5 g) was homogenized using a pestle and mortar in liquid nitrogen with 3 mL of extracting solution. The mixture was transferred to 10 mL centrifuge tube and extracted for 30 min. After centrifuging at 12,000 rpm for 10 min at 25 °C, the supernatant was collected as a source of flavonoids. Absorbance readings at 470 nm (UV-1800, Shimadzu, Japan) were taken and rutin was www.nature.com/scientificreports/ used for a standard curve. The content of total flavonoids in the supernatant was measured accurately using a detection kit for flavonoids in plants (Solarbio, Beijing, China) according to the manufacturer's instructions. Three biological repeats were applied in both YFY and Control.
Analysis of total carotenoids in fresh-cut yam. The content of total carotenoids was investigated using a commercial kit (Solarbio, Beijing, China). Frozen tissue (0.5 g) was homogenized using a pestle and mortar in liquid nitrogen with 1 mL of distilled water and 20 mg reagent 1 under dark condition, and then transferred to a 10 mL centrifuge tube. The mixture was extracted for 3 h in the absence of light. During the process of extraction, the centrifuge tube was shaken twice to fully extract. After the tissue residue turning nearly white, which was in the bottom of tube, the mixture was centrifuged at 4000 rpm for 5 min. The supernatant was collected as a source of carotenoids. Absorbance readings at 440 nm (UV-1800, Shimadzu, Japan) were taken. The content of total carotenoids in the supernatant was measured accurately with a detection kit for carotenoids in plants (Solarbio, Beijing, China) according to the manufacturer's instructions. Three biological repeats were carried out in YFY and Control.
Analysis of bisdemethoxycurcumin in fresh-cut yam. The content of bisdemethoxycurcumin was measured and quantified according to Zhao et al. 6 . Three biological repeats were carried out in both YFY and Control.

RNA extraction.
In the transcriptome analysis, three biological replicates of each treatment were used.
Total RNA was extracted from the yam tissue using TRIzol Reagent according the manufacturer's instructions (Invitrogen, Carlsbard, CA, USA), and genomic DNA was removed using DNase I (TaKara, Japan). The integrity and purity of the total RNA was then determined using a bioanalyzer (Agilent 2100 Bioanalyzer, Agilent Technologies, Inc., Santa Clara CA, USA) and quantified using an ND Library preparation and sequencing. Library preparation and sequencing were begun as described previously 30 .
De novo assembly and annotation. The raw paired-end reads were trimmed and controlled for quality using the default parameters of SeqPrep (https:// github. com/ jstjo hn/ SeqPr ep) and Sickle (https:// github. com/ najos hi/ sickle). The clean data were then used for de novo assembly with Trinity (http:// trini tyrna seq. sourc eforge. net/) 31 . To retrieve the function annotations of the assembled transcripts, the transcripts were searched against the NCBI protein nonredundant (NR), string, and KEGG databases using BLASTX to identify the proteins with the highest sequence similarity to the given transcripts; typical cut-off E-values of less than 1.0 × 10 −5 were used. The BLAST2GO (http:// www. blast 2go. com/ b2gho me) 32 program was used to obtain the GO annotations of unique assembled transcripts for describing biological processes, molecular functions, and cellular components. Metabolic pathway analysis was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG, http:// www. genome. jp/ kegg/). www.nature.com/scientificreports/ were centrifugated at 12,000×g at 4 °C for 20 min, the protein supernatants were mixed with equal volumes of Tris-saturated phenol and vortexed for 10 min at 4 °C. After centrifugation at 12,000×g at 4 °C for 20 min, the phenol phase was added to an equal volume of BPP and vortexed for 10 min at 4 °C. After centrifugation at 12,000×g at 4 °C for 20 min, the phenol phase was added to a five-fold volume of pre-cooled methanolic ammonium acetate solution to precipitate protein for 12 h at − 20 °C. After centrifugation at 12,000×g at 4 °C for 20 min, the pellet was rinsed twice with 90% acetone and then dried under vacuum. The acetone-dried powder was resuspended in protein lysis buffer containing 8 M urea, 1% SDS, and an appropriate amount of protease inhibitor. The mixture was treated ultrasonically for 20 min, and after centrifugation at 12,000×g at 4 °C for 20 min, the concentration of protein was determined using the bicinchoninic acid (BCA) method with a BCA Protein Assay Kit (Thermo, MA, USA). Protein quantification was performed according to the kit protocol.

Differential expression analysis and functional enrichment.
Protein digestion and TMT labeling. Protein was digested with trypsin (Promega, Madison, WI, USA) and incubated at 37 °C with a mass ratio of 1:50 (enzyme/substrate) overnight. The iTRAQ labeling was performed according to the manufacturer's illustration (Applied Biosystems, Sciex, Foster City, CA, USA). All labeled peptides were pooled, desalted, and vacuum-dried 37 .
High-pH RPLC separation. The pooled samples were fractionated using an ultraperformance liquid chro-  Protein identification. With ProteomeDiscoverer (Thermo Scientific, Version 2.2), analysing raw data against transcriptome was conducted. The searching of MS/MS was according to the criteria: Two missed cleavages in the trypsin digesting were allowed, with a mass tolerance of 0.05 Da (10 ppm). Choosing TMT of the N-terminus and lysine side chains of peptidases and carbamido methylation of cysteine to fix modification. A false discovery rate (FDR) threshold of less than 1% were used. www.nature.com/scientificreports/