Metabolomics Study of Browning Mechanism in Eggplant Fruits

Background: Fresh-cut fruits and vegetables is a emerging type of fruits and vegetables processing products for consumers to eat immediately or for the catering industry. Enzymatic browning is one of the crucial problems compromising the flavor and texture of fresh-cut fruit and vegetables. Eggplant is a common vegetable, which is favored by consumers. Accordingly, we used an untargeted metabolomics approach based on liquid chromatography-mass spectrometry (LC-MS) to explore the browning mechanism in peeled eggplant ( Solanum melongena L . ). Results: Metabolomics revealed several hundred differential metabolites, including lipids, phenols, sugars and fatty acids. The content of these metabolites changed dynamically as the peeled time increased. The content of polyphenols, especially chlorogenic acid, increased significantly, suggesting that the main substrate for enzymatic browning in eggplant is chlorogenic acid. Furthermore, all the differential metabolite were mapped to KEGG pathway, revealing significant differences in linoleic acid metabolism, tyrosine metabolism glutathione metabolism, pentose phosphate pathway, tropane, piperidine and pyridine alkaloid biosynthesis, phenylpropanol metabolism and glycosylphosphatidylinositol(GPI)-anchor biosynthesis over time. Therefore, we speculate that some metabolic pathways that are closely connected with respiration, glycolysis, ATP synthesis, and phenolic synthesi are disturbed after peeling, under the action of enzymes, eventually leading to browning. Conclusions: We established an untargeted metabolomics method based on LC-MS technology to explain the mechanism of eggplant browning, which may lay the foundation for better understanding the mechanism of browning during the fruits

and vegetables deeply processing, and furnish new ideas and perspectives for understanding fruit and vegetable browning in the future.

Background
Metabonomics was first proposed by Nicholson et al. [1] in 1999 and interpreted to the quantitative analysis of metabolic response of an organism subjected to pathophysiological stimulation or genetic changes. Fiehn [2] proposed in 2002 that metabolomics is a qualitative and quantitative analysis method for total metabolites in the body of an organism. Metabolomics is a burgeoning omics technology, after genomics and proteomics, which can be divided into target analysis, metabolic profiling, and fingerprint analysis, based on the scope of the study objectives and metabolic coverage. These four levels can be divided into target metabolomics and non-target metabolomics. Metabolomics is widely used in drug screening and toxicity evaluation, drug quality control, clinical marker discovery, early diagnosis, therapeutic function genomics, plant breeding, microbial engineering, nutrition, environmental toxicology and other research fields [3]. A variety of metabolites, such as organic acids, sugars, lipids, amino acids and secondary metabolites as alkaloids, flavonoids, terpenes, can be monitored by qualitative analysis of metabolites based on gas chromatography or liquid chromatography in conjunction with mass spectrometry [4]. With the deepening of plant metabolomics research, we have a better understanding of plant growth regulation, development and metabolic processes and molecular mechanisms of stress response. As well as, plant metabolomics has become a forceful tool for probing physiology and biology all aspects of plant [5]. There are many factors that may affect the metabolic capacity of plant defenses such as the interaction between biological and abiotic stresses. In the study of metabolomics of sorghum root under nitrogen stress demonstrated the biosynthetic capacity of salicylic acid was impaired [6]. In the experiment of different germination capacity of pollen grains of two clones of Chinese fir (Cunninghamia lanceolata (Lamb) Hook) using metabolomics method, it was revealed that in the expression of metabolites across and between clones at all germination stages had significant differences, enriched in 14 metabolic pathways, including flavonoid and flavonol biosynthetic pathways, protein biosynthetic pathways and glycerophospholipid metabolism [7]. In order to explore the mechanism of climate and development regulating the metabolism and antioxidant system of date palm leaves, Du et al.
[8] used the non-targeted metabolomics method based on LC-MS to find that young leaves may be more responsive to the environmental change.
Solanum melongena L., commonly known as eggplant, belongs to solanaceae plants, is one of the largest genus of solanaceae plants, there are more than 1550 species.
It is a common vegetable, rich in polyphenols, dietary fiber, vitamins and other nutrients, and provides a variety of health benefits such as lowering blood lipids, protecting the liver and as an antioxidant, so it is favored by consumers [9].
However, browning of eggplant occurs easily after peeling, which affects its sensory quality and nutritional value, leading to a decline in edibility and commodity. It is well known that the main factors affecting the browning of fruits are water loss on the fruit surface, biomolecular oxidation, electrolyte leakage and anthocyanin degradation [10]. Browning is caused by the destruction of cell structure during the peeling process, resulting in the release of PPO and phenolic substrates. With the occurrence of enzymatic reactions, phenols are catalyzed into quinones and finally polymerizated into brown or black substances. These substances accumulate on the surface of the slices, compromising the flavor of fresh-cut fruits and vegetables [11].
In this study, we selected 24 eggplant samples, which were divided into three groups for metabolic analysis. Eight biological replicates were made for each group.
An untargeted metabolomics method based on LC-MS technology combined with qualitative analysis, multivariate statistical analysis and metabolic pathway analysis in differential metabolites was used to explore the browning mechanism in eggplant, so as to offer a feasible theoretical foundation for the processing and production of fresh-cut fruit and vegetable, and provide a reference for further investigation into the restrain of browning in other fresh-cut products.

Base peak ion current check
We firstly conducted a visual inspection of the base peak ion current of all samples and found that signal peak capacity and retention time repeatability were good in all samples. Figures 1 and 2 showed the base peak chromatogram of the quality control samples in TIC+ and TIC-mode, respectively.

Establish a data matrix
Metabolomics analysis was performed using the pulp of peeled eggplant after 0 (CK), 3 and 5 min. To carry out pattern recognition, the raw data were preprocessed by the metabolomics processing software progenesis QI, including peak recognition, peak alignment, integration, baseline filtering, retention time correction, normalization. In the end, a data matrix was constructed, which include the sample name, mass-to-charge ratio (m/z), retention time (RT), ion model, and metabolites.
There are 15133 m/z features in this matrix, including 7038 m/z in TIC+ and 8095 m/z in TIC-(Additional file 1: Table S1). This matrix was gone to the subsequent analysis.

Multivariate analysis
In the data matrix, the data were disposed by multivariate analysis tools (PCA, PLS-DA and OPLS-DA). PCA, PLS-DA and OPLS-DA scoring plots and validation plots of the OPLS-DA models were built for the three contrastive groups: 3min/CK, 5min/CK and 3min/5min (Fig. 3). Table 1 showed all the parameters of these models. A confidence interval is a measure of confidence, usually the bigger the better. R2X represents the cumulative interpretation rate of the multivariate statistical analysis modeling, which generally requires R2X > 0.4, indicating that the model is credible.
R2 and Q2 are the parameters of response sequencing test, used to measure whether the model is overfitted. External validation generally requires Q2 < 0 to avoid over-fitting. Internal validation generally requires R2, > 0.5; The closer R2 is to 1, the better the model. In our results, all values for Hotelling's T2 were 95% , Q2 were < 0 (Q2 = −0.849 to −0.741), R2X were >0.4, and R2 were > 0.5, indicating that the model is reliable.

Differential metabolites
According to the VIP > 1 for the first principal component in the OPLS-DA, and pvalue < 0.05 were the criteria for screening differential metabolites, we identified the differential metabolites among the three contrastive groups (Additional file 2: Table S2). 256 differential metabolites were detected in the 3min/CK comparative group, 357 in the 5min/CK comparative group and 333 in the 3min/5min comparative group. Among these metabolites, 49 were common to all three groups and 119 were common between 3min/CK and 5min/CK. However, 253 metabolites were the non-repeated differential metabolites between any two groups; 114, 58 and 81 metabolites were exclusive to the 3min/CK, 5 min/CK and 3min/5min comparative groups, respectively (Fig. 4).
The 119 common differential metabolites from the 3min/CK and 5min/CK groups were further divided into three sorts represented by A, B and C, as indicated in Additional File 3: Table S3 based on STC. Furthermore, the metabolites in these three categories were grouped into eight types of compounds ( Table 2). The relative expression of 33 metabolites selected from the three categories (A, B, C) is shown in the heatmap (Fig.5 ).
Category A included 68 metabolites. These were mainly lipids, fatty acid and carbohydrates ( Table 2). The contents and expression levels of these metabolites are increased gradually. The metabolites of category B comprised 40 compounds, most of which were fatty acids and lipids ( Table 2). The contents and expression levels of these metabolites decreased with increase of the peeled time. There were 11 components in category C, mainly lipids (Table 2). Interestingly, their contents and expression levels decreased at 3 min, but are increased at 5 min. However, they were lower than those at 0 min (CK) as a whole.

Metabolite content changes
Eight phenols associated with browning were selected from the data matrix. The content levels of the selected metabolites at different treatment intervals shown in KEGG pathway analysis.
The metabolites in categories A, B and C were mapped, respectively, using the KEGG database onto the KEGG pathways with the following results (Fig. 7).
The metabolites of category A were mainly enriched in 18 metabolic pathways: Linoleic acid metabolism, Pentose phosphate pathway, Nitrogen metabolism, Glycosylphosphatidylinositol (GPI)-anchor biosynthesis, Autophagy-other, Taurine and hypotaurine metabolism, Galactose metabolism, Aminoacyl-tRNA biosynthesis, Arginine biosynthesis, Arginine and proline metabolism, Alanine, aspartate and glutamate metabolism, C5-Branched dibasic acid metabolism, Glutathione metabolism, Butanoate metabolism, Histidine metabolism, ABC transporters, Glyoxylate and dicarboxylate metabolism, and Glycerophospholipid metabolism. Of these, Glycosylphosphatidylinositol (GPI)-anchor biosynthesis, Linoleic acid metabolism, Autophagy-other and Nitrogen metabolism showed extremely significant differences in the 3min/CK and 5min/CK two comparative groups at the p < 0.01 level of significance, while the others revealed a significant difference (p < 0.05). There was a remarkable difference (p < 0.05) in the metabolites of category B that were only enriched in one metabolic pathway, Tropane, piperidine and pyridine alkaloid biosynthesis. Finally, the metabolites of category C were enriched in Tyrosine metabolism and had a remarkable difference (p < 0.05) similarly.

Discussion
Metabolomics is a part of systems biology and can uncover relationships between metabolites and physiology/pathology by qualitative or quantitative analysis of metabolites. Enzymatic browning of fruits is an extremely complex process. Phenols, as important substrates of enzymes, play a key role in the browning that occurred during the fruits and vegetables deeply processing. process of fruits and vegetables [12], are abundant and widely distributed in various kinds of plants. Existing research demonstrates that PPO mainly exists in the cytoplasmic membrane and cell walls, and phenols usually accumulate in vacuoles [13]. Fruit tissue structure and regionalization of the cell space are damaged by fresh-cut operations, destroying partitions between PPO and substrate polyphenols and allowing contact between enzymes and substrates for catalysis by PPO of the transformation of contact the transformation of, phenolic compounds into quinone compounds [14]. Melanins are accumulated through secondary enzyme reactions, inducing the enzymatic browning reaction [15][16][17][18]. Liu et al. [19] studied the reaction mechanism of enzymatic browning during potato processing, and analyzed the correlation between substrates and browning. Their results showed that phenolic compounds are closely related to browning, and chlorogenic acid is an important substrate leading to potato enzymatic browning. The phenolic compounds identified in banana peels in recent years were mainly dopamine, followed by chlorogenic acid and coumarin, with dopaminerase oxidation being the main factor for banana browning [20].
Chlorogenic acid and epicatechin are the phenolic compounds with the highest contents in apple fruits [21], while the substrates in mango are mainly coffeic acid and ferulic acid [22]. The substrates for litchi peel browning belong to the collated catechenol compounds, which can rapidly interact with PPO to form browning showed that peach phenols are chrologenic acid, neo-chlorogenic acid, protocatechuic acid. Mishra et al. [26] also showed that browning of eggplant was related to the content increased of polyphenols. In this study, chlorogenic acid, dopamine, caffeic acid, cinnamic acid and vanillin were found in eggplant fruits, and the metabolite analysis showed that the content of chlorogenic acid, dopamine, caffeic acid and cinnamic acid was increased after peeling; the content of chlorogenic acid accumulated at an extremely significant level (p = 0.0025), but the content of flavonoids such as soybean vanillin did not change significantly, suggesting that the main substrate of enzymatic browning in eggplant is chlorogenic acid.
PPO is involved in the tyrosine metabolism pathway. Tyrosine can produce Ldopamine under the action of PPO, which is converted to dopamine quinone, leading to the occurrence of browning [27]. In this study, the tyrosine metabolism pathway was the common metabolic pathway showing significant differences between the 3min/CK and 5min/CK comparison groups. It is worth noting that the vanillylmandelic acid content decreased at 3 min after peeling, but was rising again at 5 min, although it was still lower than that in the CK group. This is because after the eggplant was peeled, vanillylmandelic acid was oxidized by PPO through tyrosine metabolism and converted into quinones, and finally polymerized into brown substances, leading to browning.
In addition, phenylalanine ammonialyase (PAL) plays a momentous role in regulating the production of phenolic substances in the phenylalanine pathway and is an important enzyme regulating browning [28]. L-phenylalanine rooted in the Shikimate pathway generates trans-cinnamic acid via PAL ammonia solution reaction. The trans-cinnamic acid then enters into phenylpropanol metabolism and generates intermediates including coumaric acid, ferulic acid and mustard acid that can be further converted to coumarin and chlorogenic acid, and also can form CoA esters.
Finally, these are converted to flavonoids, lignin, alkaloids and other secondary metabolites [29][30]. In this study, we detected the phenylpropanol metabolism pathway in the 3min/CK comparison, with the accumulation of L-tryptophan and chlorogenic acid up-regulated. The biosynthesis pathway for flavonoids was found in the 3min/CK and 5min/CK comparisons, where the participation of quercetin was down-regulated. Mechanical injury by peeling can significantly increase PAL activity, and phenols are changed by the phenylpropanol metabolism pathway and flavonoids biosynthesis pathway. At the same time, the membrane system is destructed, and the regional distribution of enzymes and substrates is disrupted, allowing phenols to be catalyzed by PPO under action of reactive oxygen species (ROS), eventually generating brown substances.  [31][32][33]. In this study, glutathione metabolism was found in all three groups, and the content of L-glutamate involved in this pathway increased, while the content of threonine decreased. Therefore, we hypothesized that in exist of H 2 O 2 , phenols and flavonoids are catalyzed by POD after eggplant is peeled and aggregate into brown substances causing fruit browning. Alternatively, glutathione in fruits and vegetables is oxidized, through glutathione metabolism, and the content of L-glutamic acid is increased while the content of threonine is decreased, which accelerating aging and affecting the flavor and nutrition of fruits and vegetables.
Furthermore, the analysis of metabolites gave a comprehensive overview of the active pathways after eggplant is peeled. Glycosylphosphatidyl-inositol anchored protein(GPI), is an integral protein involved in the endoplasmic reticulum and plasma membranes on GPI [34][35]. Like all fatty acids, Linoleic acid can be perceived as a material source of cellular energy [36]. Linoleic acid is a part of membrane phospholipid, which is an essential structural part to sustain a particular level of membrane fluidity. And in vivo it can be esterified to come into being lipids such as cholesterol esters and triacylglycerols [37]. Moreover, when it is released from membrane phospholipids, can be enzymatically oxidized into all kinds of derivatives participate in cell metabolism and signal transduction [38]. In addition, we also found that the content of some alkaloids, such as L(-)-Nicotine, enoxacin and neocnidilide, were decreased as peeled time increased, through tropane, piperidine and pyridine alkaloid biosynthesis. The pentose phosphate pathway provides substrates for essential biosynthetic pathways [39]. In this study, linoleic acid metabolism, the glycosylphosphati-dylinositol (GPI)-anchor, Tropane, piperidine and pyridine alkaloid biosynthesis, and pentose phosphate pathways at the three different treatment stages after peeling were significantly different. Therefore, we hypothesize that these metabolic pathways may play an important role in the browning of fruits and vegetables.

Conclusions
The present investigation clearly demonstrated that metabolomics analysis is a good tool for explaining the browning mechanism of eggplant by detecting changes in the expression of metabolites and differences in metabolic pathways after peeling. There were pronounced changes in metabolites in eggplant for at 3 min and 5 min after peeling. Several hundred differential metabolites were identified, including lipids, phenols, sugars and fatty acids etc. The metabolite contents changed dynamically as the peeled time increased. Furthermore, by analyzing the KEGG pathway of these differential metabolites, we found that some metabolic pathways were perturbed after peeling, such as linoleic acid metabolism, glutathione metabolism, pentose phosphate pathway, tropane, piperidine and pyridine alkaloid biosynthesis, phenylpropanol metabolism and tyrosine metabolism, and these showed significant differences over time. The identified metabolites and metabolic pathways may have specific effects on various aspects of fruit and vegetable browning. We established an untargeted metabolomics method based on LC-MS technology to explain the mechanism of eggplant browning, which provides new ideas and perspectives for understanding fruit browning in the future.

LC-MS analysis
The analytical instrument of this experiment is The UHPLC system that is composed of liquid-mass coupling system consisted of AB Sciex Triple TOF 5600 high resolution mass spectrometer. It was appiled to analyze metabolic profiles in both TIC+ and TIC-modes.
Chromatographic conditions were as follows: Chromatographic column: ACQUITY UPLC BEH C18 (100 mm × 2.1 mm, 1.7 m); column temperature: 45℃; mobile phase: A-water (containing 0.1% formic acid), B-acetonitrile (containing 0.1% formic acid); flow rate: 0.4 mL/min; injection volume: 5 µL. The elution gradient is shown in Table 3. Mass spectrometry conditions are shown in Table 4. In addition, in order to assess the repeatability of the whole analysis process, we added one QC sample per 8 analysis samples.

Data preprocessing and statistical analyses
The original data were disposed using progenesis QI made in Waters Corporation from Milford of USA. Then, we constructed a data matrix that was inputed into XCMS software (version 14.0, Umetrics, Umeå, Sweden), in which PCA, PLS-DA and OPLS-DA were executed. The 95% confidence interval that defines the change of the model is represented by the Hotelling's T2 region. It is represented as an ellipse in the rating diagram of the model. The reliability of the model is illuminated by the terms R2X or R2Y and Q2. R2X represents the cumulative interpretation rate of the multivariate statistical analysis modeling. R2Y represents the cumulative interpretation rate in the Y-axis model. Therefore, R2X or R2Y indicate the goodness of fit. R2 and Q2 are the parameters of response sequencing test, used to measure whether the model is overfitted, so as to represent predictability, which can be calculated through the cross-validation process. To prevent overfitting of the model, seven fold cross-validation and 200 RPT were applied to evaluate the quality of the model.

Identification of differential metabolites
The difference metabolites between groups were screened out using the methods of multi-dimensional analysis and one-dimensional analysis. The screening criteria were VIP >1 for the first principal component of the OPLS-DA model, and p < 0.05 3min/CK and 5min/CK groups were further divided into three sorts represented by A, B and C Additional file 4:  Table 1 The parameters for the assessment of these models.     The TIC-total ion flow diagram of QC Venn diagram of the differential metabolites among three comparative groups Heatmap shows the differences in the expression levels of metabolites The differences in phenols content at the three treatment stages The differential metabolic pathways (a) The differential metabolic pathways that be enriched

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.