The Effect of SO₂ Fumigation, Acid Dipping, and SO₂ Combined with Acid Dipping on Metabolite Profile of ‘Heiye’ Litchi (Litchi chinensis Sonn.) Pericarp

Abstract

Sulfur fumigation (SF), acid dipping (HCl treatment, HAT), and their combination (SF+HAT) are common methods for long-term preservation and color protection of litchi. However, their effects on the metabolic profile of the litchi pericarp have not been investigated. SF resulted in a yellowish-green pericarp by up-regulating lightness (L*), b*, C*, and h° but down-regulating total anthocyanin content (TAC) and a*, while HAT resulted in a reddish coloration by up-regulating a*, b*, and C* but down-regulating L*, h°, and TAC. SF+HAT recovered reddish color with similar L*, C* to SF but a*, b*, h°, and TAC between SF and HAT. Differential accumulated metabolites (DAMs) detected in HAT (vs. control) were more than those in SF (vs. control), but similar to those in SF+HAT (vs. control). SF specifically down-regulated the content of cyanidin-3-O-rutinoside, sinapinaldehyde, salicylic acid, and tyrosol, but up-regulated 6 flavonoids (luteolin, kaempferol-3-O-(6″-malonyl)galactoside, hesperetin-7-O-glucoside, etc.). Five pathways (biosynthesis of phenylpropanoids, flavonoid biosynthesis, biosynthesis of secondary metabolites, glutathione metabolism, and cysteine and methionine metabolism) were commonly enriched among the three treatments, which significantly up-regulated sulfur-containing metabolites (mainly glutathione, methionine, and homocystine) and down-regulated substrates for browning (mainly procyanidin B2, C1, and coniferyl alcohol). These results provide metabolic evidence for the effect of three treatments on coloration and storability of litchi.

1. Introduction

Litchi (Litchi chinensis Sonn.) fruit is a famous tropical (or subtropical) fresh fruit worldwide, which is attractive to customers due to its bright red color, unique flavor, and rich nutrition [1]. Due to the high temperature and high humidity in its harvest season (usually from late May to mid-July), litchi is extremely prone to browning and decay and, thus, is subjected to significant loss of commodity value and shelf life [1]. Except for being used widely as a preservative and sanitizing agent to prevent spoilage by microorganisms in fruit juices and syrups, sulfur dioxide and its sustained-release agents were also widely applied in postharvest preservation of fresh fruits, such as grapes [2], litchis [1], and longans [3]. Sulfur/SO₂ fumigation (SF) is the most important fresh-keeping treatment of litchi in international trade to effectively inhibit the rot and browning of fresh litchi, although it causes a short-term dramatic increase of sulfur residues in the pericarp and pulp [1]. However, excessive and prolonged SO₂ fumigation could lead to postharvest decolorization and abscission of grapes and fading of the red color and aggravation of rot in litchis (specifically infected by Penicillium) [4]. In order to solve the decolorization problem caused by SF, acid dipping was used to effectively reduce the pigment loss caused by SF bleaching of the litchi pericarp [4,5].

More importantly, the roles of SF in postharvest fruit preservation depend not only on its direct antimicrobial properties but also on its regulation of physiology, biochemistry, and metabolism of fruits, especially sulfur metabolism and secondary metabolism [3,4,6]. The accumulation of sulfur-containing compounds such as cysteine, hydrogen sulfide, and glutathione was up-regulated after the SF treatment [6]. Polyphenol oxidases (PPOs) were recognized as the key factor in the enzymatic browning of the litchi pericarp, which shows high activity to O-diphenols, such as catechol or 4-methylcatechol [5]. In addition, (-)-epicatechin, the major endogenous polyphenol in the litchi pericarp, was identified as a substrate for PPO [7]. Peroxidase (POD/PRX) might be involved in an anthocyanase-anthocyanin-phenolic-H₂O₂ reaction during litchi enzymatic browning [8]. In a series of studies [9,10,11], laccase was also confirmed to be involved in litchi pericarp browning as it functions in proanthocyanidin and lignin polymerization. pH is an important factor that influences the activity of POD and PPO, which is more effectively inhibited in an acid (especially pH < 3.5) than in an alkaline pH [12,13]. It was reported that the activity of litchi laccase (LcADE/LAC) in vitro was inhibited by Na₂SO₃ and L-cysteine at 0.1 mM, while the enzyme activity and gene expression of LcADE/LAC were strongly inhibited by SO₂-HCl treatment in postharvest litchi pericarp [14]. Moreover, pre-treatment with 0.25% cysteine effectively delayed browning and conserved litchi fruit quality under controlled atmosphere storage (1% O₂ + 5% CO₂, 5 ± 1 °C) [15], while postharvest 0.25% methionine application significantly reduced browning, enhanced antioxidant activities, and maintained litchi fruit quality [16]. The results from above-mentioned documents indicated that SO₂ (including its donors, such as Na₂SO₃ [14], Na₂S₂O₅ [3]), cysteine [15], methionine [16], combined with acidic solution (such as HCl [15], oxalic acid [13]) were effective choices for inhibiting the browning and preservation of fresh litchi after harvest.

As a mature commercial preservation technology, SO₂ fumigation showed remarkable effects to reprogram the metabolic profile of litchis [12], longans [3], and grapes [2]. SO₂ boosts the antioxidant capacity of grape berries by promoting ROS scavenging enzyme activities and the ascorbate-glutathione (AsA-GSH) cycle [2]. Treatment with SO₂-releasing paper resulted in a significantly changed profile of glucosinolate metabolites, which further altered the flavor of longan pulp [3]. However, the metabolic basis underlying the decolorization of the litchi pericarp caused by SF and the restoration of the red color by acid dipping after SF has not been studied. In this study, we compared the metabolite profiles of ‘Heiye’ litchi pericarp treated with SF, HAT, and SF+HAT using widely targeted metabolomics, which was expected to explain the effects of these three treatments on coloration and biochemistry of the litchi pericarp and be helpful for optimizing the SO₂-fumigation preservation of litchi.

2. Materials and Methods

2.1. Materials and Treatments

The ‘Heiye’ fresh litchi fruit used in this experiment was grown following commercial cultivation practices in an orchard in Maoming city (110.95° E, 22.23° N), Guangdong province. Commercial mature fresh litchi fruit with no damage and no disease was harvested and transported to a laboratory within 4 h. The fresh litchi fruit (80 kg) was randomly divided into 4 groups (each 20 kg) which were then treated as follows: Control (CK): 20 kg of fruit of was dipped with 500 mg/L prochloraz for 2 min; SF (SO₂ fumigation): 20 kg of fruit was fumigated with SO₂ (12 g sulfur was burned in a foil box) for 30 min in a 0.3 m³ enclosed room (the dose referenced from [17]); HAT: 20 kg of fruit was dipped in a 4% hydrochloric acid (food graded) solution for 2 min; SF+HAT: 20 kg of fruit was fumigated with SO₂ (12 g sulfur was burned in a foil box) for 30 min in a 0.3 m³ enclosed room, and then dipped in a 4% hydrochloric acid (food graded) solution for 1 min. After air-drying for 30 min, approximately 20 fruits were packed into a polyethylene bag (0.02 mm thick) and stored at 1 ± 0.5 °C. The fruits from the CK, SF, HAT, SF+HAT groups, which were stored and cooled for 6 h (marked as 0 d after harvest, 0 DAH), were used to take photos, determine the total soluble solid (TSS) and titratable acid (TA) content in pulp, pH, relative electrolytic leakage, total flavonoid content (TFC), total phenolics content (TPC), and TAC of the pericarp. Then, the chromatic values of no less than 10 individual fruits were determined. No less than 40 individual fruits from each group were sampled. After being separated, the pulp and pericarp were immediately frozen in liquid nitrogen and kept at −80 °C until used.

2.2. Reagents

Ethanol, gallic acid, rutin, Folin-Ciocalteau reagent, Na₂CO₃, NaNO₂, Al(NO₃)₃, NaOH used in this part were all analytical reagents (supplied by Sinopharm Chemical Reagent Co. Ltd., Shanghai, China). HPLC degraded acetic acid, methanol, and acetonitrile from Merck (Merck KGaA, Darmstadt, Germany) and ultra-pure water prepared by distilled water through a Milli-Q A10 system (Millipore, Milford, MA, USA) were used for UPLC-Q-TRAP analysis.

2.3. Determination of Chromatic Values of Pericarp

Color of the pericarp surface was measured by a color analyzer (KONICA MINOLTA CR-300, Tokyo, Japan). The red to green was expressed as +a* to −a*, the yellow to blue was expressed as +b* to −b*, and brightness was expressed as L*. Each test was performed with thirty individual litchi fruits.

2.4. Determination of pH and Relative Electrolytic Leakage (REL) of Pericarp

The frozen pericarp samples from Section 2.1 were ground into powder by adding liquid nitrogen. Then, 1.0 g of the pericarp sample was mixed with 8 mL of deionized water in a tube and extracted by ultrasonic extraction for 30 min. The volume was fixed to 10 mL with deionized water, and then the pH value was detected by a pH meter (PHS-3C, Shanghai INESA Scientific Instrument CO., Ltd., Shanghai, China). The pH value of the pericarp sample was the measured pH value −1.

The cell membrane permeability of the pericarp revealed by relative electrolytic leakage (REL) was measured according to a method reported by Luo et al. (2019) [1]. Three 0.5 cm diameter disks were punched from the peel of each fruit, and peel disks of ten fruits were collected. After being washed 3 times using deionized water, 10 peel disks were transferred into a 50 mL tube containing 25 mL deionized water. After a dipping for 30 min, the electro-conductibility (D₁) was measured by a conductivity meter (INESA Instrument DDS-307, Shanghai, China). Then, the tube was sealed and bathed in boiling water for 15 min. After an ice-bath cooling, the electro-conductibility (D₂) of the solution was measured. The electro-conductibility in another 50 mL tube containing 25 mL deionized water with no peel disk was measured and set as the control (D₀). Each test was performed with three biological repeats. REL of each sample was calculated according to Equation (1):

REL(%) = (D₁ − D₀)/(D₂ − D₀) × 100%,

(1)

2.5. Determination of TSS and TA

One-half of the pulp of each sample from Section 2.1 was juiced and used to determine the total soluble solid (TSS) and titratable acid (TA) content. Each test was performed with three biological repeats.

2.6. Determination of TFC, TPC, and TAC of Pericarp

The total phenolics content (TPC) was assayed at 760 nm by the Folin–Ciocalteu method using gallic acid as the standard, while the TFC (total flavonoid content) was measured using a modified colorimetric method at 510 nm using a standard curve of rutin according to our previously reported method [18,19]. The total anthocyanin content (TAC) was measured using a colorimetric method [20]. Each test was performed with three biological repeats.

2.7. Widely Targeted Metabolomic Analysis

The frozen pericarp samples from Section 2.1 were crushed using a mixer mill (MM400, Retsch, Verder Shanghai Instruments and Equipment Co., Ltd., Shanghai, China) with a zirconia bead for 1.5 min at 30 Hz. Sample powder (100 mg) was added to 1.0 mL 70% methanol and extracted for 8–12 h (or overnight) at 4 °C. Each sample was vortexed three times during the period to increase the extraction efficiency. After a centrifugation at 10,000× g for 10 min, the supernatant was collected, filtered using a Carbon-GCB SPE Cartridge (250 mg, 3 mL, CNWBOND, ANPEL, Shanghai, China), and each sample was filtrated (SCAA-104, 0.22 μm pore size, ANPEL, http://www.anpel.com.cn/, accessed on 1 November 2024) before LC-MS analysis. Each test was performed with three biological repeats.

The sample (2 μL) was injected and analyzed using an ultra-performance liquid chromatography (Shim-pack UFLC CBM30A system, SHIMADZU, Kyoto, Japan) coupled with tandem ESI-MS/MS (6500 Q-TRAP, Applied Biosystems, Shanghai, China). The UPLC conditions were performed according to the previous reported method [20]: chromatographic column: ACQUITY UPLC HSS T3 (C₁₈, 100 × 2.1 mm i.d., 1.8 µm, Waters Technology (Shanghai) Co., LTD, Shanghai, China); mobile phase A: ultra-pure water containing 0.04% acetic acid, B: acetonitrile containing 0.04% acetic acid; elution steps: min (A, %): 0 (95%) → 11.0 (5%) → 12 (5%) → 12.1 (95%) → 15 (95%); flow rate was 0.40 mL/min and column temperature was 40 °C. The effluent was alternatively connected to the ESI-triple quadrupole-linear ion trap (Q-TRAP)-MS.

Widely targeted metabolites were analyzed by LIT and triple quadrupole (QQQ) scans using a triple quadrupole-linear ion trap mass spectrometer (Applied Biosystems QTRAP 6500, Shanghai, China) [21]. The MS/MS system was equipped with an ESI Turbo Ion-Spray interface and controlled by Analyst 1.6.3 software (AB Sciex, Concord, Ontario, Canada). The parameters for operating ESI source were set as follows: ion source, turbo spray; source temperature 500 °C; ion spray voltage (IS) 5500 V; ion source gas I (GSI), gas II (GSII), curtain gas (CUR) were set at 55, 60, and 25.0 psi, respectively; high collision gas (CAD). Instrument tuning and mass calibration were performed with 10 and 100 μM polypropylene glycol solutions in QQQ and LIT modes, respectively. QQQ scans were acquired as MRM experiments with collision nitrogen gas set to 5 psi. Declustering potential (DP) and collision energy (CE) for individual MRM transitions were conducted with further DP and CE optimization. A specific set of MRM transitions was monitored for each period according to the metabolites eluted within this period.

After the isotope signal and the repetitive signal were removed, the metabolites were identified qualitatively by the secondary spectral information based on the public metabolite database (e.g., MassBank, KNApSAcK…) and the self-built MetWare database (from Metware Biotechnology Co., Ltd., Wuhan, China). Multiple reaction monitoring (MRM) of triple quadrupole mass spectrometry was employed to quantify each metabolite: only the precursor ions of the target substance were screened and ionized in the collision cell to break and form fragment ions. The precursor ions and the characteristic fragment ions were selected by triple quadrupole filtration to make more accurate and repeatable quantitative results [22]. After the integration and correction of chromatography peaks using MultiaQuant software 3.0.3 and performed on each mass spectrometer file, the abundance (relative content) of the corresponding compound (area of each chromatography peak) was calculated.

2.8. Screening of Differential Accumulated Metabolites and KEGG Enrichment Analysis

The rest metabolites were used for orthogonal partial least squares discriminant analysis (OPLS-DA) [23]. OPLS-DA was performed to eliminate the factors unrelated to sample grouping, the errors between samples, and other random errors, and was usually used to maximize the differences between groups. The metabolites with variable importance in projection (VIP) value ≥ 1, p-value (t-test) < 0.05, and |log₂(Fold change)| ≥ 1 were identified as differential accumulated metabolites (DAMs).

2.9. Statistical Analysis

The variance of data (the abundance of each compound) among groups was analyzed using SPSS software package release 18.0 (IBM Corp., Armonk, NY, USA). Multiple comparisons were performed by one-way ANOVAs based on Duncan’s multiple range tests, while paired-samples t-tests were performed to test the statistical significance between two samples. The correlation network graph according to the Pearson correlation test was drawn using the OmicStudio tools accessed on 12 November 2024 at https://www.omicstudio.cn/tool [24].

3. Results

3.1. The Effect of SF, HAT, and SF+HAT on the Chroma Value, Physiological and Chemical Properties of Pericarp, and Quality

As shown in Figure 1, the red on the surface of the litchi pericarp was eliminated, and the yellowish-green background appeared after the SF treatment. The HAT treatment removed the brown spots on the surface of the peel (see the CK group), resulting in an even, bright red color. It is worth noting that the yellow-green background color was still visible on the HAT-treated litchi pericarp. The litchi that was treated by SF and then HAT showed a uniform red surface and background as well as inner peel.

As shown in

Table 1. The effect of SF, HAT, and SF+HAT on the chroma value, TAC, TFC, TPC, pH, REL, TSS, and TA.

Treatment	L*	a*	b*	C*	h°	TAC (mg/g)
CK	40.96 ± 0.29 c	18.22 ± 0.46 b	17.72 ± 0.37 d	25.57 ± 0.24 c	44.34 ± 1.23 c	1.689 ± 0.043 a
SF	55.55 ± 0.57 a	5.64 ± 0.45 c	35.03 ± 0.46 a	35.60 ± 0.48 a	80.91 ± 0.71 a	0.551 ± 0.014 d
HAT	39.51 ± 0.62 d	27.67 ± 0.75 a	19.52 ± 0.50 c	34.08 ± 0.56 b	35.44 ± 1.26 d	1.414 ± 0.036 b
SF+HAT	52.45 ± 0.73 b	19.26 ± 0.88 b	29.41 ± 0.76 b	36.34 ± 0.36 a	56.75 ± 1.77 b	1.163 ± 0.049 c
Treatment	TFC(mg/g)	TPC (mg/g)	pH	REL(%)	TSS (%)	TA (%)
CK	45.735 ± 1.008 b	20.975 ± 0.230 ab	5.28 ± 0.10 a	22.29 ± 1.54 c	16.17 ± 0.13 a	0.141 ± 0.003 a
SF	49.268 ± 1.632 a	22.443 ± 1.607 a	4.23 ± 0.02 b	75.42 ± 3.64 b	15.93 ± 0.07 b	0.147 ± 0.007 a
HAT	50.600 ± 0.398 a	19.580 ± 0.201 c	2.91 ± 0.09 c	96.90 ± 0.39 a	16.17 ± 0.09 a	0.106 ± 0.003 b
SF+HAT	46.664 ± 1.474 b	20.258 ± 0.773 bc	3.05 ± 0.07 c	88.37 ± 9.35 a	15.70 ± 0.10 b	0.131 ± 0.015 ab

Note: The different lowercase letters in each column represented significant differences (p < 0.05) among the samples. TAC, total anthocyanin content; TFC, total flavonoid content; TPC, total phenolic content; REL, relative electrolytic leakage; TSS, total soluble solid; TA, titratable acid.

, the SF treatment resulted in a significant increasing of L*, b*, C* and h° and a significant decreasing of a*, but the HAT treatment resulted in a decreasing of L*, b*, and h° and a significant increasing of a* and C* when compared to the CK. Thus, SF treatment resulted in a yellowish-green coloration of the pericarp. The combined treatment (SF+HAT) showed a similar effect to the SF treatment on the L* and h° values, but showed a similar effect to the HAT treatment on the a* value. It was interesting to note that SF+HAT treatment increased the b* and C* values, which were closer to the level of those in the SF group, but significantly higher than those of the HAT group, although the b* and C* values of the HAT group were higher than the CK. The above results indicated that SF and HAT had different influences on the chromatic value of the litchi pericarp. The influence of the combined treatment (SF+HAT) on different chroma values was affected by the two single treatments to different degrees.

The three treatments significantly changed the total anthocyanin content: SF resulted in a significant decrease in TAC by almost two-fold; HAT resulted in a slight but significant decrease in TAC, while SF+HAT (that is, HAT after SF) recovered the TAC in the pericarp, which was bleached by the SF treatment (Table 1). Both the SF and HAT lead to increased TFC, while HAT resulted in decreased TPC.

All of these three treatments resulted in a significantly lower pH value but a higher ion leakage (REL) of the litchi pericarp, and the two treatments with HAT showed a more dramatic effect (Table 1). The higher REL might result in the reddish inner pericarp of SF+HAT-treated fruit (Figure 1). These three treatments did not show a similar effect on the TSS and TA of litchi pulp as on the pericarp.

3.2. Identification, Quantification, and Classification of Metabolites Detected in the Litchi Pericarp Treated with CK, SF, HAT, and SF+HAT

In order to explore the effects of these three treatments on the metabolite profile, especially color and browning-related metabolites of the litchi pericarp, a widely targeted metabolome was analyzed. As shown in Figure 2a, in total, 649 metabolites categorized into 13 groups (class level 1) and 42 subgroups (class level 2) were detected in the CK and treated litchi pericarp samples. Among the detected metabolites, 401 metabolites were annotated with cpd_id in the KEEG COMPOUND database, but 248 metabolites were not annotated with any cpd_id (Figure 2a).

The largest group of secondary metabolites identified in the pericarp was flavonoids, containing 144 metabolites and accounting for 22.19% of the detected metabolites (Figure 2a,b). The detected flavonoids were comprised of 51 flavonols (15 annotated, 36 unannotated), 40 flavones (13 annotated, 27 unannotated), 14 flavanols (5 annotated, 9 unannotated), 13 dihydroflavones (7 annotated, 6 unannotated), 6 dihydroflavonols (4 annotated, 2 unannotated), 6 flavanones (2 annotated, 4 unannotated), 6 chalcones (5 annotated, 1 unannotated), 4 isoflavones (3 annotated, 1 unannotated), 2 flavan 4-ols (1 annotated, 1 unannotated), and 2 anthocyanidins (2 annotated).

Moreover, 81 lipids (accounting for 12.48% of the detected metabolites) were detected, which were comprised of 31 free fatty acids (24 annotated, 7 unannotated), 13 glycerol esters (2 annotated, 11 unannotated), 22 LPCs (unannotated), 13 LPEs (unannotated), 1 PCs (annotated), and 1 sphingolipid (unannotated). In addition, 76 amino acids and derivatives (accounting for 11.71%, 59 annotated, 17 unannotated), 72 phenolic acids (accounting for 11.09%, 43 annotated, 29 unannotated), 71 organic acids (accounting for 10.94%, 65 annotated, 6 unannotated), 40 saccharides (accounting for 6.16%, 38 annotated, 2 unannotated), and 21 others (accounting for 3.24%, 10 annotated, 11 unannotated) (Figure 2a,b). Another 38 nucleotides and derivatives (accounting for 5.86%), which were comprised of 2 cyclic nucleotides (annotated), 4 deoxyribonucleosides (annotated), 4 deoxyribonucleotides (annotated), 6 purines (annotated), 4 pyrimidines (annotated), 12 ribonucleosides (9 annotated, 3 unannotated), and 6 ribonucleotides (annotated) were detected.

Meanwhile, 33 alkaloids (accounting for 5.08%, 16 alkaloids, 7 phenolamine, 8 plumerane, 1 quinoline alkaloids, 1 tropan alkaloids), 23 tannins (accounting for 3.54%, 12 proanthocyanidins, 4 annotated, 8 unannotated; 11 tannins, 9 annotated, 2 unannotated), 21 lignans and coumarins (7 coumarins, 4 annotated, 3 unannotated; 14 lignans, 3 annotated, 11 unannotated), 17 vitamins and cofactors (4 cofactors, 4 annotated; 13 vitamins, 12 annotated, 1 unannotated), and 12 terpenoids were detected in the samples.

The complete separation among the CK, SF, HAT, and SF+HAT samples was revealed in the principal component analysis (PCA) and indicated the significant difference of metabolite profile among these four sample groups (Figure 2c). This result was confirmed by the changes in each class of metabolites after treatments (Figure 2d). Among the 13 classes of metabolites, most of the lipids, amino acids and derivatives, organic acids, saccharides, nucleotides and derivatives showed an up-regulated (deep red) or a down-regulated (deep blue) accumulation when compared to the CK samples (Figure 2d).

3.3. Significantly Differently Accumulated Metabolites Between the Treated Samples and Control

Based on the OPLS-DA analysis, the metabolites with variable importance in projection (VIP) value ≥ 1 and 2.0 fold-change were identified as the significantly differential accumulated metabolites (DAMs) between the treated samples and control (Figure 3a). As shown in Figure 3a, 90 metabolites and 67 metabolites were, respectively, found to be significantly up-regulated and down-regulated in the SF-treated samples compared to the CK. In addition, 135 up-regulated DAMs and 107 down-regulated metabolites were observed in the HAT-treated samples compared to the CK. Similarly, 132 up-regulated DAMs and 105 down-regulated DAMs were identified in the SF+HAT-treated samples compared to the CK. It was interesting to note that only 66 up-regulated DAMs and 81 down-regulated DAMs were found between the SF- and SF+HAT-treated samples.

Within the differential accumulated metabolites of SF vs. CK, HAT vs. CK, and SF+HAT vs. CK, lipids were the largest class (Figure 3b). In details, the top 5 classes of up-regulated metabolites, including 26 lipids, 17 flavonoids, 15 amino acids, 9 phenolic acids, and 7 nucleotides, and the top 5 classes of down-regulated metabolites, including 16 lipids, 11 phenolic acids, 10 amino acids, 8 organic acids, and 5 alkaloids, were observed in the SF-treated pericarp samples compared to the CK. The top 5 classes of up-regulated metabolites in the HAT-treated pericarp samples compared to the CK were lipids (29), amino acids (23), organic acids (17), phenolic acids(17), and flavonoids (16), while the top 5 classes of down-regulated metabolites in the HAT-treated pericarp were lipids (32), saccharides (18), organic acids (12), amino acids (10), and nucleotides (8). Similarly, the top 6 classes of up-regulated metabolites in the SF+HAT-treated pericarp samples compared to the CK were lipids (31), amino acids (23), organic acids (16), phenolic acids (15), flavonoids (15), and nucleotides (15), while the top 5 classes of down-regulated metabolites in the HAT-treated pericarp were lipids (32), saccharides (21), organic acids (14), phenolic acids (8), and amino acids (6). Compared to the SF-treated samples, 18 phenolic acids, 12 organic acids, 8 flavonoids, 8 nucleotides, and 5 lipids were up-regulated, while 30 lipids, 14 saccharides, 7 organic acids, 7 nucleotides, and 5 amino acids were down-regulated in the SF+HAT-treated samples.

In order to analyze the differences and similarities of DAMs caused by these three treatments, a Venn analysis was performed (Figure 3c). In total, 20 DAMs including 8 up-regulated (6 flavonoids: kaempferol-3-O-(6″-malonyl)-galactoside, luteolin, hesperetin-7-O-glucoside, luteolin-7-O-(6″-malonyl)-glucoside, 3′-O-Methyltricetin-7-O-glucoside, and 2′-Hydroxygenistein; 1 nucleotide: guanine; 1 organic acid: 2-hydroxy-3-phenylpropanoic acid) and 12 down-regulated metabolites (4 phenolic acids: salicylic acid, sinapinaldehyde, tyrosol, 6-O-acetylarbutin; 2 lignans and coumarins: isoscopoletin, umckalin; and cyanidin-3-O-rutinoside) were found to be uniquely significant accumulated in the SF vs. CK, but not in other comparing pairs (Figure 3c and Figure 4a). It was interesting to note that 28 DAMs including 11 up-regulated (3 amino acids, 3 organic acids, 2 saccharides, 2 phenolic acids, 2 nucleotides, and 2 vitamins) and 17 down-regulated metabolites (3 nucleotides, 2 amino acids, 2 lipids, 1 saccharide and 1 organic acid), which were mainly primary metabolites, were found to be uniquely significantly accumulated in the HAT vs. CK (Figure 3 and Figure 4b). Moreover, 18 DAMs including 8 up-regulated (3 amino acids, 3 organic acids, 1 nucleotide, and 1 saccharide) and 10 down-regulated metabolites (3 lipids, 3 saccharides, 1 nucleotide, 1 phenolic acid, 1 flavonoid, and 1 alkaloid), which were also mainly primary metabolites, were found to be uniquely significantly accumulated in the SF+HAT vs. CK (Figure 3c and Figure 4c).

In further, only 9 DAMs were significantly differently accumulated in both the SF vs. CK and the HAT vs. CK, which consisted of 4 up-regulated DAMs (1 amino acid, 3 flavonoids) and 5 down-regulated DAMs (3 amino acids, 1 coumarin, and 1 vitamin) (Figure 3c and Figure 4d). In total, 14 DAMs were significantly differently accumulated in both of the SF vs. CK and SF+HAT vs. CK, which consisted of 9 up-regulated DAMs (3 flavonoids, 2 amino acids, 1 alkaloid, 1 lipid, 1 organic acid, and 1 nucleotide) and 5 down-regulated DAMs (2 phenolic acids, 1 organic acid, 1 lipid, and 1 saccharide) (Figure 3c and Figure 4e). DAMs of a larger number were significantly accumulated in both the HAT vs. CK and SF+HAT vs. CK, which consisted of 44 up-regulated DAMs (9 organic acids, 8 nucleotides, 7 flavonoids, 6 phenolic acids, 5 lipids, 5 amino acids, 3 flavonoids, 1 tannin, 1 saccharide, 1 alkaloid and 1 Others) and 47 down-regulated DAMs (15 saccharides, 13 lipids, 8 organic acids, 3 nucleotides, 2 amino acids, 2 vitamins, 2 Others, and 1 flavonoids) (Figure 3c and Figure 4f).

More importantly, 114 metabolites were common DAMs screened from SF vs. CK, HAT vs. CK, and SF+HAT vs. CK (Figure 3c and Figure 4g). Among them, 65 DAMs were significantly up-regulated in all of the three comparison pairs (SF vs. CK, HAT vs. CK, and SF+HAT vs. CK), which consisted of 23 lipids, 12 amino acids, 9 phenolic acids, 5 flavonoids, 5 saccharides, 4 nucleotides, 2 organic acids, 2 others, 1 alkaloid, 1 tannin, and 1 vitamin, while 40 DAMs were significantly down-regulated in all of the three comparison pairs and consisted of 15 lipids, 5 phenolic acids, 4 alkaloids, 3 amino acids, 3 organic acids, 2 saccharides, 2 others, 2 terpenoids, 2 vitamins, 1 flavonoid, and 1 nucleotide. The remaining 9 DAMs showed different trends in the three comparison pairs: 1 amino acid and 2 organic acids were down-regulated in SF vs. CK and SF+HAT vs. CK, but up-regulated in HAT vs. CK; 1 amino acid and 1 organic acid were down-regulated in SF vs. CK, but up-regulated in HAT vs. CK and SF+HAT vs. CK; 2 lipids, 1 nucleotide, and 1 vitamin were up-regulated in SF vs. CK and SF+HAT vs. CK, but down-regulated in HAT vs. CK (Figure 4g).

The above results indicated that the HAT treatment resulted in more significantly changed metabolites in the litchi pericarp compared with the SF treatment; Most of the DAMs were shared among the three comparison pairs (SF vs. CK, HAT vs. CK, and SF+HAT vs. CK), or between the two comparison pairs HAT vs. CK and SF+HAT vs. CK.

3.4. KEGG Enrichment of DAMs and Important Metabolic Pathways

In total, 97 out of 157 DAMs from SF vs. CK, 163 out of 242 DAMs from HAT vs. CK, and 157 out of 237 DAMs from SF+HAT vs. CK were annotated with cpd_id and used for KEGG pathway enrichment analysis. The results showed that the 97 DAMs from SF vs. CK were significantly enriched into 7 pathways including biosynthesis of phenylpropanoids (map01061), flavonoid biosynthesis (map00941), biosynthesis of secondary metabolites (map01110), glutathione metabolism (map00480), phenylpropanoid biosynthesis (map00940), tryptophan metabolism (map00380), and cysteine and methionine metabolism (map00270) (FDR value < 0.05, Figure 5a).

The 163 DAMs from HAT vs. CK were significantly enriched into 26 pathways and the top 20 of them were biosynthesis of phenylpropanoids (map01061), ABC transporters (map02010); flavonoid biosynthesis (map00941); biosynthesis of plant hormones (map01070); biosynthesis of secondary metabolites (map01110); glutathione metabolism (map00480); purine metabolism (map00230); pentose phosphate pathway (map00030); nicotinate and nicotinamide metabolism (map00760); alanine, aspartate, and glutamate metabolism (map00250); biosynthesis of alkaloids derived from histidine and purine (map01065); glycine, serine, and threonine metabolism (map00260); galactose metabolism (map00052); cysteine and methionine metabolism (map00270); pyrimidine metabolism (map00240); carbon fixation in photosynthetic organisms (map00710); linoleic acid metabolism (map00591); reductive carboxylate cycle (CO₂ fixation) (map00720); phenylalanine, tyrosine, and tryptophan biosynthesis (map00400); and valine, leucine, and isoleucine biosynthesis (map00290) (FDR value < 0.05, Figure 5b).

The 157 DAMs from SF+HAT vs. CK were significantly enriched into 36 pathways and the top 20 of them were biosynthesis of phenylpropanoids (map01061); ABC transporters (map02010); flavonoid biosynthesis (map00941); biosynthesis of secondary metabolites (map01110); biosynthesis of alkaloids derived from histidine and purine (map01065); galactose metabolism (map00052); biosynthesis of plant hormones (map01070); cysteine and methionine metabolism (map00270); pyrimidine metabolism (map00240); alanine, aspartate and glutamate metabolism (map00250); reductive carboxylate cycle (CO₂ fixation) (map00720); linoleic acid metabolism (map00591); glutathione metabolism (map00480); phenylalanine, tyrosine, and tryptophan biosynthesis (map00400); glyoxylate and dicarboxylate metabolism (map00630); pentose phosphate pathway (map00030); glycine, serine, and threonine metabolism (map00260); starch and sucrose metabolism (map00500); purine metabolism (map00230); and glycolysis/gluconeogenesis (map00010) (FDR value < 0.05, Figure 5c).

In order to reveal the common and uniquely enriched pathways among the three comparison pairs (SF vs. CK, HAT vs. CK, and SF+HAT vs. CK), a Venn analysis was conducted. Five pathways, including biosynthesis of phenylpropanoids (map01061), flavonoid biosynthesis (map00941), biosynthesis of secondary metabolites (map01110), glutathione metabolism (map00480), and cysteine and methionine metabolism (map00270), were found to be common enriched pathways among the three comparison pairs (Figure 5d,e).

It was interesting to note that 20 common enriched pathways were observed between HAT vs. CK and SF+HAT vs. CK, but no common enriched pathways were found between SF vs. CK and HAT vs. CK or between SF vs. CK and SF+HAT vs. CK. These 20 common enriched pathways consisted of 6 pathways related to carbohydrate metabolism (map00030 pentose phosphate pathway, map00052 galactose metabolism, map00630 glyoxylate and dicarboxylate metabolism, map00010 glycolysis/gluconeogenesis, map00020 citrate cycle (TCA cycle), and map00500 starch and sucrose metabolism), 4 pathways related to amino acid metabolism (map00250 alanine, aspartate, and glutamate metabolism; map00260 glycine, serine, and threonine metabolism; map00400 phenylalanine, tyrosine, and tryptophan biosynthesis; and map00290 valine, leucine, and isoleucine biosynthesis), 2 pathways related to energy metabolism (map00710 carbon fixation in photosynthetic organisms and map00720 reductive carboxylate cycle (CO₂ fixation), 2 pathways related to nucleotide metabolism (map00230 purine metabolism and map00240 pyrimidine metabolism), 2 pathways related to metabolism of cofactors and vitamins (map00760 nicotinate and nicotinamide metabolism and map00770 pantothenate and CoA biosynthesis), ABC transporters (map02010), biosynthesis of plant hormones (map01070), biosynthesis of alkaloids derived from histidine and purine (map01065), galactose metabolism (map00052), and linoleic acid metabolism (map00591) (Figure 5d,e).

More importantly, only 2 pathways (phenylpropanoid biosynthesis map00940 and tryptophan metabolism map00380) and 1 pathway (zeatin biosynthesis map00908) were uniquely enriched in SF vs. CK and HAT vs. CK, respectively, but 11 pathways (biosynthesis of alkaloids derived from shikimate pathway map01063; ascorbate and aldarate metabolism map00053; biosynthesis of terpenoids and steroids map01062; glycerophospholipid metabolism map00564; benzoate degradation via hydroxylation map00362; biosynthesis of alkaloids derived from terpenoid and polyketide map01066; biosynthesis of alkaloids derived from ornithine, lysine, and nicotinic acid map01064; vitamin B6 metabolism map00750; pyruvate metabolism map00620; C5-Branched dibasic acid metabolism map00660; and sulfur metabolism map00920) were uniquely enriched in SF+HAT vs. CK (Figure 5d,e).

The above results indicated that SF and HAT had significant and different effects on the metabolic profile of the litchi pericarp, while SF+HAT, as the combined treatment of SF and HAT, did not show a simple superimposed effect of SF and HAT on the metabolic profile of the litchi pericarp (Figure 5).

3.5. Correlation Analysis Between Metabolites and Physiological and Biochemical Indexes

In order to screen out the metabolites whose change trend was consistent with that of physiological and biochemical indexes, correlation analysis was performed using the Pearson correlation coefficient to screen the significant pairwise correlations (R > 0.95 and significance > 0.05) (Figure 6). It was interesting to note that b* was significantly positively correlated with the L* value (r = 0.979, p-value = 0.021), while b* was significantly negatively correlated with the TAC (r = −0.954, p-value = 0.046), a* was significantly negatively correlated with the TPC (r = −0.979, p-value = 0.021), and REL was significantly negatively correlated with the pH value (r= −0.950, p-value = 0.0497) (Figure S1). TFC and its significantly correlated metabolites were separated from other networks of physiological and biochemical indexes (Figure 6).

On the one hand, REL was connected to pH and C* values due to the common correlated metabolites. It could be seen in the network that pH was the largest node, which showed significant positive correlation with 74 metabolites (mainly including 26 lipids, 15 saccharides, 8 organic acids, 5 flavonoids, 4 phenolic acids, 3 nucleotides and derivatives, 2 amino acids and derivatives, 2 lignans and coumarins, 2 tannins) and significant negative correlation with 19 metabolites (mainly including 5 lipids, 3 organic acids, 3 saccharides, 2 amino acids and derivatives, 2 phenolic acids). In total, 13 metabolites (including 9 lipids, 1 amino acid and derivatives, 1 flavonoid, 1 phenolic acid and 1 saccharide) showed significant positive correlation and 60 metabolites (mainly including 18 lipids, 6 phenolic acids, 6 saccharides, 6 alkaloids, 5 organic acids, 4 amino acids and derivatives, 4 flavonoids, 3 tannins) showed significant negative correlation with the REL value. Among them, 39 metabolites showed significant correlation with both the REL and pH values.

There were 19 metabolites (mainly containing 8 lipids, 4 nucleotides and derivatives, 3 amino acids and derivatives) that showed significant positive correlation, and 41 metabolites (mainly including 11 phenolic acids, 6 organic acids, 4 alkaloids, 4 flavonoids, 3 tannins) that showed significant negative correlation with the C* value. The above results indicated that lipids were the largest metabolite class, which was correlated with the pH and REL values, while phenolic acids and lipids were the top two metabolite classes, which were correlated with the C* value.

On the other hand, due to the common correlated metabolites, the b* value was connected to TAC and the L* value, and the a* value was connected to TPC and the h° value. Four metabolites (2 phenolic acids, 1 alkaloid, 1 vitamin, and cofactors) showed positive correlation, and 31 metabolites (mainly including 19 flavonoids, 3 terpenoids, 3 amino acids and derivatives, 2 phenolic acids, 2 lignans and coumarins) showed negative correlation with the a* value. In addition, 35 metabolites (mainly including 9 flavonoids, 9 lipids, 6 amino acids and derivatives, 4 nucleotides and derivatives) showed positive correlation, and 7 metabolites (3 phenolic acids, 2 lipids, 1 flavonoid, 1 saccharide) were found to be negatively correlated with the b* value. Moreover, 36 metabolites (mainly 13 flavonoids, 7 lipids, 6 amino acids and derivatives, 3 tannins, 3 nucleotides and derivatives) showed positive correlation and 9 metabolites (mainly 3 phenolic acids, 2 flavonoids) showed negative correlation with the L* value. Similarly, 50 metabolites (mainly 32 flavonoids, 4 amino acids and derivatives, 3 lipids, 3 nucleotides and derivatives) showed positive correlation, and 3 metabolites (1 flavonoid, 1 lipid, 1 vitamin and cofactors) showed negative correlation with the h° value.

3.6. Effect of SF, HAT, and SF+HAT on the Abundance of Sulfur-Containing Metabolites

In total, 18 sulfur-containing metabolites (including 3 glutathione and related metabolites, 5 methionine and related metabolites, 6 cysteine and related metabolites, and 3 other sulfur-containing metabolites) were detected in the control and treated litchi pericarp (Figure 7).

Glutathione was undetectable in the control pericarp but showed high abundance in the treated pericarp (SF > SF+HAT > HAT) (Figure 7a). HAT resulted in a higher level of oxiglutatione (the oxidized form of glutathione), but SF and SF+HAT resulted in lower content of oxiglutatione in the litchi pericarp (Figure 7b). Only SF resulted in a significant increase of S-(Methyl)glutathione (another metabolite related to glutathione) (Figure 7c). Among the methionine and 4 related metabolites, L-methionine sulfoxide showed the highest level in the litchi pericarp. All three treatments resulted in increased L-methionine (Figure 7d). SF and SF+HAT resulted in increased L-homomethionine, but HAT resulted in decreased L-homomethionine (Figure 7e). Similarly, SF and SF+HAT resulted in increased S-Adenosyl-L-methionine (Figure 7f). Only SF+HAT resulted in increased L-methionine sulfoxide (Figure 7g). On the contrary, all three treatments resulted in decreased L-methionine methyl ester (Figure 7h).

Among the cysteine and cysteine-related metabolites, L-cysteine was undetectable in the CK- and SF-treated pericarp (Figure 7i), while L-cysteinyl-L-glycine (Figure 7m) and γ-glu-cys (Figure 7n) were undetectable in the CK. Both HAT and SF+HAT resulted in remarkably higher L-cysteine accumulation (Figure 7i), while all three treatments resulted in a significantly higher accumulation of L-homocystine (Figure 7j), S-(5′-Adenosy)-L-homocysteine (Figure 7l), L-cysteinyl-L-glycine (Figure 7m), and γ-glu-cys (Figure 7n), but down-regulated accumulation of S-Allyl-L-cysteine (Figure 7k). Moreover, the three treatments significantly up-regulated the content of 6-methylmercaptopurine (Figure 7o) and 4-Methyl-5-thiazoleethanol (Figure 7r); both the SF and HAT down-regulated the content of 5′-Deoxy-5′-(methylthio)adenosine (Figure 7p) and biotin (Figure 7q).

3.7. Effect of SF, HAT, and SF+HAT on the Abundance of Anthocyanins and Reported Browning-Related Metabolites

The abundance of the two main anthocyanins in the litchi pericarp was cyanidin-3-O-rutinoside and cyanidin-3-O-glucoside, which were significantly down-regulated by almost 50% in the SF-treated pericarp. The HAT treatment resulted in higher cyanidin-3-O-glucoside in the litchi pericarp than in the CK pericarp, although the TAC in HAT was slightly lower than in the control. The HAT after SF (SF+HAT) treatment significantly reduced the loss of red color and anthocyanin content in the pericarp caused by SF by increasing the content of cyanidin-3-O-rutinoside and cyanidin-3-O-glucoside, which was still lower than the CK- and HAT-treated pericarp (Figure 8a,b).

All three treatments showed no significant effect on the content of epicatechin, which is the most abundant substrate in the litchi pericarp for enzymatic browning (Figure 8c). Similarly, no significant difference in the abundance of EC-EC-EC was observed among the three treated and CK pericarp (Figure 8e). The SF up-regulated but HAT down-regulated the epicatechin-glucoside content in the litchi pericarp (Figure 8d). SF did not significantly influence, but both the HAT and SF+HAT down-regulated the content of catechin (Figure 8f) and CCC (Figure 8g). In contrast, the content of protocatechuic acid-4-O-glucoside was significantly up-regulated in all three treated pericarp samples (Figure 8h).

All three treatments resulted in up-regulated content of gallic acid, another kind of phenolic acid with high abundance in the litchi pericarp (Figure 8i). However, the glucoside form of gallic acid, β-glucogallin, was down-regulated in SF but up-regulated in the HAT-treated litchi pericarp (Figure 8j). In addition, SF did not significantly influence the content of gallocatechin (Figure 8k) and epigallocatechin (Figure 8l), but up-regulated the content of gallocatechin-catechin and gallocatechin-gallocatechin. In contrast, HAT down-regulated the content of GCC but up-regulated the content of GC.

Among the six detected procyanidins, procyanidin A1 was not influenced by the three treatments (Figure 8s); procyanidin B4 (Figure 8o) and B1 (Figure 8t) was down-regulated by HAT while procyanidin B3 (Figure 8p) was up-regulated by SF; the content of procyanidin B2 (Figure 8q) and C1 (Figure 8r) was totally down-regulated byall of the three treatments. Similarly, the abundance of coniferyl alcohol (Figure 8u) was down-regulated by all three treatments, but coniferin (glucoside form of coniferyl alcohol; Figure 8v) was up-regulated by SF. In total, the above results indicated that all three treatments had a significant influence on the abundance of anthocyanins and browning-related metabolites.

4. Discussion

4.1. The Metabolic Basis for the Effect of SF, HAT, and SF+HAT on Coloration and Browning of Litchi Pericarp

As a commercial preservation, SO₂ and a category of its donors were permitted and regularly used in a wide range of food, including wine, dried fruits, and meat products [25]. So far, the use of SO₂ in the fruit and vegetable industry has been indispensable, although its safety concerns have been controversial. Cyanidin-3-O-rutinoside and cyanidin-3-O-glucoside were identified as the main anthocyanins in the litchi pericarp, which contribute to the bright red appearance of litchi fruit [6]. SO₂ treatment rapidly bleaches the litchi pericarp due to the formation of colorless anthocyanin-SO₃H complexes. In this study, a significant decrease of cyanidin-3-O-rutinoside and cyanidin-3-O-glucoside was observed in the SF-treated litchi pericarp. However, no anthocyanin-SO₃H was detected in the SF-treated, HAT-treated, or SF+HAT-treated litchi pericarp. This might be explained by the limitations of detection technology (widely-targeted LC-MS/MS metabolome) or the instability of anthocyanin-SO₃H. HAT (HCl dipping) resulted in a significant increase of cyanidin-3-O-glucoside but not cyanidin-3-O-rutinoside in the litchi pericarp, while HAT after SF (SF+HAT) significantly reduced the loss of cyanidin-3-O-rutinoside and cyanidin-3-O-glucoside caused by SF, which were still lower than in the control. In total, the effect of HAT and HAT after SF on anthocyanins in the litchi pericarp was consistent with the previously reported stabilizing effect on anthocyanins [6].

Flavanols are among the most important groups of secondary metabolites, due to their ubiquity and importance to plant physiology [10]. The monomeric epicatechin (EC), catechin (CT), gallocatechin (GC), epigallocatechin gallate (EGCG), catechin gallate (CG), and oligomeric procyanidin (e.g., procyanidin B2 and procyanidin B3) were proven as the substrates for enzymatic browning in the litchi pericarp [7,8,9,10,11]. Among these metabolites, the SF treatment did not significantly influence the content of EC, CT, GC, epigallocatechin (EGC), catechin-catechin-catechin (CCC), epicatechin-epicatechin-epicatechin (EC-EC-EC), procyanidin A1, procyanidin B4, and procyanidin B1 but up-regulated the abundance of epicatechin-glucoside (EC-glu), protocatechuic acid-4-O-glucoside, gallic acid, GCC, GCGC, procyanidin B3, and coniferin, but down-regulated the abundance of β-glucogallin, procyanidin B2, procyanidin C1, and coniferyl alcohol. This result indicated the protective effect of SF on the special substrates for enzymatic browning, as SO₂ is a strong enzyme inhibitor of laccase [14]. Contrary to the SF, the HAT treatment significantly down-regulated the abundance of EC-glu, CT, CCC, GCC, procyanidin B4, and procyanidin B1, but up-regulated protocatechuic acid-4-O-glucoside (also higher than that in the SF-treated and SF+HAT-treated pericarp), β-glucogallin, and GC. This might be explained that the HAT treatment led to a lower pH than SF (Table 1), and thus, resulted in a stronger acid-catalyzed depolymerization of proanthocyanidins [26]. In total, the influence of SF+HAT on anthocyanins (cyanidin-3-O-rutinoside and cyanidin-3-O-glucoside) and most of the browning-related metabolites (such as EC-glu, CCC, protocatechuic acid-4-O-glucoside, gallic acid, β-glucogallin, GC, GCC, GCGC, and procyanidin B3) showed a superimposed effect of SF and HAT.

In recent years, sever organic acids such as oxalic acid [27,28,29,30], oxalic acid combined with ascorbic acid [31], kojic acid [32], malic acid combined with lycopene [33], phytic acid [14], and gallic acid [34] were reported to show anti-browning or delayed-browning effects on the litchi pericarp during storage. Among these compounds, the abundance of gallic acid (Figure 8i), ascorbic acid (Figure S2), vanillic acid (Figure S2), and ferulic acid (Figure S2) was detected in the litchi pericarp and found to be up-regulated by several folds by all three treatments (SF, HAT, and SF+HAT). This result revealed the preservation effect of SF, HAT, and SF+HAT, that is, they immediately changed the color-related metabolites in the litchi pericarp and promoted the accumulation of compounds that help inhibit browning.

Overall, from a metabolic perspective, SF and HAT, respectively, exert negative and positive regulatory effects on anthocyanin content, while SF+HAT is the result of their neutralization. All three treatments significantly increased the abundance of one of the highest abundant phenolic acids in the pericarp: protocatechuic acid-4-O-glucoside (Figure 8h) and gallic acid (Figure 8i), as well as the most abundant water-soluble antioxidant, VC (Figure S2a). This might be the reason for the enhanced anti-browning ability of the litchi peel in the treatment group, which showed delayed increase of pericarp browning and maintained higher TAC during the long-term storage (Figure S3).

4.2. The Metabolic Basis for the Effect of SF, HAT, and SF+HAT on Sulfur-Containing Metabolites in Litchi Pericarp

Once the SO₂ enters the plant’s cell, the plant conducts detoxification by oxidizing the SO₃²⁻ into SO₄²⁻ (by sulfite oxidase in peroxisome), which is then stored in the vacuoles or converting the SO₃²⁻ into a thiol or other sulfur-containing compounds via a reduction pathway. Our previous results indicated that the key enzymes for sulfite metabolism were up-regulated at both enzymatic and transcriptional levels after SF [1]. In this work, 18 sulfur-containing metabolites were detected in the treated litchi pericarp (Figure 7). Among these sulfur-containing amino acids, methionine [16,17,18], cysteine [16,17,18], and glutathione [18] were reported to show anti-browning or delayed-browning effects on the litchi pericarp during storage. It was interesting to note that glutathione (Figure 7a), methionine (Figure 7d), cysteine (Figure 7i) and most cysteine-related derivatives (homocysteine, Figure 7j; S-(5′-adenosy-homocystine, Figure 7l; L-cysteinyl-L-glycine, Figure 7m; γ-Glu-Cys, Figure 7n) were found to be up-regulated by all three treatments (SF, HAT and SF+HAT). This might help the litchi pericarp enhance the ability to scavenge oxygen free radicals and inhibit the activities of browning-related enzymes. In total, in this study, it was unexpected to find that HAT treatment could also lead to an increase in the abundance of sulfur-containing metabolites (9 out of 18), but the mechanisms involved still need further research.

5. Conclusions

SO₂ fumigation is an important preservation method for the international trade of fresh litchi fruit. Acid dipping was used to inhibit the enzymatic browning and restore anthocyanin levels. The acid dipping combined with SO₂ was applied to restore the red color of the pericarp, which was bleached by SO₂-fumigation alone. However, the metabolic basis underlying the decolorization of the litchi pericarp caused by SF and the restoration of the red color by acid dipping after SF has not been studied. Herein, we compared the effects of SF, HAT, and SF+HAT on the coloration, physiology, and metabolic profiles of ‘Heiye’ litchi, which is a main cultivar used for export in China. SF resulted in a yellowish-green pericarp characterized as up-regulated L*, b*, C* and h° but down-regulated a*, which was associated with lower TAC, cyanidin-3-O-glucoside, and cyanidin-3-O-rutinoside, while HAT resulted in a reddish coloration characterized as up-regulated a*, b*, C* but down-regulated L* and h°, which was associated with higher TAC and cyanidin-3-O-glucoside. HAT after SF (SF+HAT) significantly reduced the loss of reddish color caused by SF through partly recovering the TAC, cyanidin-3-O-glucoside, and cyanidin-3-O-rutinoside. The SF, HAT, and SF+HAT significantly reprogrammed the metabolic profile in the pericarp. Differential accumulated metabolites detected in HAT (vs. CK) (both of the up- and down-regulated DAMs) were more than those in SF (vs. CK), but similar to those in SF+HAT (vs. CK). SF specifically down-regulated the content of cyanidin-3-O-rutinoside, sinapinaldehyde, salicylic acid, and tyrosol, but up-regulated six flavonoids, while HAT resulted in more up-regulated and down-regulated DAMs belonging to lipids. Biosynthesis of phenylpropanoids, flavonoid biosynthesis, biosynthesis of secondary metabolites, glutathione metabolism, and cysteine and methionine metabolism were commonly enriched among the three treatments, which significantly up-regulated sulfur-containing metabolites (mainly glutathione, methionine, and homocystine) and down-regulated substrates for browning (mainly procyanidin B2, C1, and coniferyl alcohol). These results provide metabolic evidence for the effect of the three treatments on coloration and storability of litchi.