Transcriptome and Metabolome Analyses Reveal Key Genes potentially relative to sucrose and citrate accumulation in wampee (Clausena Lansium Skeels) Fruit During Fruit Advancement and Ripening

Background: The seedless Clausena lansium Skeels, belonged to the Rutaceae Clausena family, is the main commercial variety in China currently. The high citrate contents in mature fruit have seriously restricted the development of the industry. The dynamic changes of sucrose and citrate, as well as the involved key regulatory genes or metabolism pathways during the fruit development of seedless wampee, however, remain unclear. Results: In this study, we used the seedless wampee fruits at various advancement levels as materials and specied a total of 623 compounds by widely targeted metabolome. Metabolome analysis further revealed that sucrose was gradually accumulated throughout the fruit development, while citrate was accumulated rapidly in the transitionary period from fruit expansion to color conversion, and decreased slowly after color conversion period without signicant difference. Moreover, we identied several key differentials expressed genes existed in sucrose as well as citrate metabolism, that mainly encode insoluble acid invertases (IVR), sucrose-phosphate synthase (SPSs), sucrose synthase (SuSy), NAD-dependent malate dehydrogenase (NAD-MDH), trehalose 6-phosphate synthase (TPS), and ATP citrate lyase (ACL), by comparing the transcriptome data of fruit in each stage. Conclusions: We found that sucrose content was mainly regulated by SUS, IVR, SPS, and TPS, the citrate synthesis was mainly regulated by MDH, and the decomposition of citrate was mainly controlled by ACLβ in the acetyl-CoA pathway, suggesting that GABA shunt pathway did not play a decisive role. stages. Therefore, this study mainly focused on the differential mechanisms of metabolites during the fruit development of wampee fruits, especially sucrose and citrate. The transcriptome data showed that the differential genes mainly involved in energy production and transfer (C), amino acids translocation and metabolism (E), secondary metabolites synthesis, translocation and metabolism (Q) during fruit development, suggesting that the metabolite content of sucrose and organic acids in these pathways signicantly changed. The transcriptome and metabolome analyses further showed that sucrose content, which might be mainly regulated by genes such as SUS, IVR, SPS, and TPS, was gradually accumulated throughout fruit development with gradually decreased accumulation amount. While the citrate content was increased rstly and then slowly decreased. Also, its synthesis might be mainly regulated by MDH, and its decomposition might be dominated by the acetyl-CoA pathway, suggesting that the GABA shunt pathway cannot determine citrate content.


RNA-Seq and differentially expressed genes (DEGs) studies
To explore the mechanism of metabolic changes in different stages of seedless wampee fruit, especially the mechanism of sucrose and citrate changes, four transcriptome databases (PDQseq, ZSQseq, CSQseq, WSQseq) were established by RNA-sEq. The signals that were larger than 240 million could be detected in each transcriptome data library. The results showed that a total of 7191618051nt of RNA fragments were detected in PDQ stage, of which GC content accounted for 44.11%, 7392980402nt of for ZSQ stage(44.35% GC content), 7270067923nt for CSQ stage(44.25% GC content), and 7487257723nt for WSQ stage(44.43% GC content) ( Table 4). According to the analysis of annotation database, such as the annotation COG, GO, KEGG, KOG, Pfam Annotation, Swissprot, eggNOG, NR, a total of 82673 unigenes were thoroughly analyzed, and the number of unigenes with a length over 1000 bp was 21781, and the number of unigenes with a length between 300 and 1000 was 34764. All RPKM values of unigene were shown in Supplement 2.
By comparing different transcriptome data library, 2540 down-regulated and 1134 up-regulated differential genes have ascertained among PDQseq and ZSQseq, 837 down-regulated and, 568 up-regulated differential genes were identi ed between ZSQseq and CSQseq, and 272 up-regulated and 837 down-regulated differential genes were identi ed between CSQseq and WSQseq ( Fig. 2A).
The Venn diagram of differential genes between different databases showed the total of 19 up-regulated (Fig. 2B) and 89 downregulated differential genes (Fig. 2C) in different developmental stages.
Based on the co-expression analysis, the total of 5722 differentially expressed genes have divided into four categories, containing 452, 705, 1837 and 2728 genes in each group respectively. Group A included the up-regulated genes (Fig. 3A), and Group B included the down-regulated genes (Fig. 3B), while group C and D included the unchanged genes (Fig. 3C, 3D). Since the changed genes were involved in many metabolism pathways, we classi ed these genes into different pathways by KOG database according to the investigation of genes with up and down regulation, of which energy generation and transfer (C), amino acid metabolism and transport (E), secondary metabolites synthesis, metabolism and transport (Q) were the main classi cations, and the sum of down-regulated genes in the aforementioned trajectories was signi cantly higher than the sum of up-regulated genes, indicating that the content of energy storage substances (sucrose and glucose), amino acids and secondary metabolites ( avonoids) changed signi cantly during fruit development of seedless wampee.
According to the annotated genomic information of the genes that related to sucrose and citrate metabolism (shown in Table 4 and Table 5), we predicted the location information and the ratio of the expression level at each stage (the ratio > 1 means upregulation; the ratio < 1 means down-regulation). Among all the genes, nine genes were selected from the expression pro le to verify by qRT-PCR with three biological replicates for each gene. Linear analysis showed that the correlation between transcriptome data obtained by RNA-seq and gene expression levels acquired through qRT-PCR, indicating that transcriptomic data was authentic, as well as reliable (Fig. 4).

Regulation of sucrose metabolism
Fruit sucrose content mainly depends on sink strength. Sucrose is absorbed and assimilated by various organs, and it depends on the activity of enzymes (e.g., SPS, SUS, IVR [14][15][16][17]) in the sucrose metabolism pathway [18]. Here, the transcriptome data (see Table 4) showed that three SPS genes, ve SUS genes, four IVR genes and four TPS genes could be detected. As shown in Fig. 5, with the maturation of wampee fruit, the sucrose and glucose content gradually augmented, which was in agreement with the cumulative trend of sucrose, glucose in other fruits, such as peach [3], pomelo [4], and watermelon [19].
From PDQ stage to ZSQ stage, the increase of sucrose content was mainly due to the up-regulation of SUS gene, leading to a larger amount of sucrose synthesis compared to its degradation caused by up-regulation of IVR gene expression. SPS can regulate the sucrose synthesis, and SUS is able to catalyze sucrose in order to produce fructose and UDP-glucose [17]. During this stage, SPS (c31711.graph_c0) was up-regulated by 2.11-folds, SUS (c49920.graph_c0, c60779.graph_c0) was up-regulated by 4.5-folds and 30.9-folds, respectively. While, the expression level of IVR and TPS did not change signi cantly. Besides, the sucrose content was increased by 2.05-folds, the glucose content was increased by 1.41-folds, while the Trehalose-6P content did not change. These results indicated that the synthesis rate of sucrose was much higher than the degradation rate. Also, the expression level of HXK, a gene involved in glycolysis, and PEPCK [20], a gene related to the gluconeogenesis, were up-regulated, and the content of Fructose-6P was reduced by 55%, suggesting that the respiration of wampee fruit gradually enhanced from the PDQ stage to ZSQ stage, and the precursors and energy generated by the TCA cycle increased rapidly, providing a material basis and energy for fruit expansion.
The increase rate of sucrose content slowed down during the transitionary period from ZSQ stage to CSQ stage, which might be due to the increased degradation of sucrose caused by upregulation of IVR, SUS and TPS genes. During this period, the upregulation of SUS promoted the decomposition of sucrose into fructose and UDP-glucose, while, the upregulation of TPS contributed to further increasing the Trehalose-6P content. The upregulation of IVR caused the decomposition of sucrose into fructose, which could be further transformed into Fructose-6P via down-regulating FRK. The metabolome database showed that the increase of sucrose (only 66%) and glucose (only 29%) gradually slowed down, while trehalose-6-phosphate, fructose-6phosphate, and glucose-6-phosphate improved by 81%, 55%, and 37%, respectively, suggesting that some sucrose was converted into monosaccharides, which was further converted into other precursors through TCA cycle to provide energy during this period.
During the period from CSQ stage to WSQ stage, sucrose was further reduced which might be due to slower synthesis and further increased consumption rate. During this period, sucrose content, glucose, fructose-6-phosphate, glucose-6-phosphate, trehalose-6phosphate increased by 36%, 33%, 64%, 29%, 21%, respectively. Transcriptomic data exhibited that the expression level of SPS was unchanged and the SUS gene was up-regulated during this period, suggesting that sucrose synthesis was further reduced compared to the stage before maturation. The upregulation of IVR involved in sucrose degradation indicated the increase of sucrose consumption. The expression of IVR increased rapidly from PDQ stage to ZSQ stage, then became stable. It has been reported that the acidic IVR activity of citrus fruit was activated before maturation, and then gradually disappeared due to higher fruit pH (> 5) during the CSQ stage [21][22]. While the pH of seedless wampee fruit has been maintained between pH 3 and 5 throughout the development (Table 1), indicating acidic IVR activity maintained a high activity throughout the development. Hence, degradation amount of sucrose from the CSQ stage to WSQ stage was the same with the period from ZSQ stage to CSQ stage, while the expression of sucrose synthesis-related genes (SPS, SUS) was lower than the period before the CSQ stage, leading to a reduced synthesis (only 36%).
According to the transcriptome and metabolome analysis, sucrose content was mainly regulated by SUS, IVR, SPS and TPS during the maturation of wampee fruit. Among these genes, synthesis related genes, such as SUS and SPS were mainly responsible for the period from PDQ stage to ZSQ stage, and the degradation-related genes, such as IVR and TPS mainly functioned after ZSQ stage.

Regulation of citrate metabolism
To explore the citrate mechanism in seedless wampee fruit, transcriptome and metabolome analyses were used to analyze fruit citrate metabolism-related genes and pathways during fruit development (Fig. 6). From PDQ stage to ZSQ stage, MDH expression increased signi cantly, citrate content increased by 35%, while malic acid content decreased by 66%. These ndings indicated that citrate synthesis and malic acid decomposition were positively correlated with MDH expression, which was most likely due to the upregulation of MDH that promote the malic acid to be oxidized by NAD-mtMDH, and further being converted into citrate via pyruvate and stored in vacuole. Besides, there was no differential expression for citrate synthase gene, suggesting the synthesis of citrate in wampee fruit might be regulated by MDH but not citrate synthase gene [4,23].
After the ZSQ stage, the citrate content decreased slowly, indicating only part of citrate was converted into other substances after transporting out of the vacuole. Generally, the conversion of citrate into other substances was mainly through TCA cycle, glyoxylate cycle, GABA shunt, and acetyl-CoA catabolism [24]. The TCA cycle might not determine the citrate content of wampee fruit. After the ZSQ stage, the malic acid content gradually decreased, the content of fumaric acid did not change signi cantly, and other genes in TCA cycle, such as citrate synthase gene, isocitrate dehydrogenase gene, and cis-aconitate gene, were also not signi cantly differentially expressed. Also, it has been reported that citrate could be transported out of the mitochondria in the TCA cycle [25], implying the TCA cycle was closely related to the decomposition of fruit citrate but cannot determine the citrate content [26][27]. Here, the transcriptome data showed that the key gene ICL was not differentially expressed in glyoxylate cycle pathway, indicating that the degradation of citrate was not involved in this pathway. GABA shunt and acetyl-CoA pathways are the main decomposition trajectories for citrate during fruit ripening such as banana [28], tomato [29], and orange [30]. Here, we found that after ZSQ stage, the expression of ACO and IDH (not detected), key genes in GABA shunt pathway, did not change signi cantly (Table 5), and the content of glutamic acid, glutamine, asparagine in the pathway gradually decreased. These ndings strongly suggested that the GABA shunt pathway might not be the main pathway for citrate decomposition after ZSQ stage. The acetyl-CoA pathway is mainly involved in the decomposition of citrate during fruit maturation [31]. In this pathway, ACL, a key gene in acetyl-CoA pathway, could generate oxaloacetic acid and acetyl-CoA by catalyzing citrate, and nally generate fatty acids, amino acids, and avonoids. Here, the expression of ACLβ was negatively correlated with citrate content, suggesting that the acetyl-CoA pathway might be the main pathway for citrate decomposition after the ZSQ stage (Fig. 6).

Conclusions
The widely targeted metabolome has been widely used in rice [32], citrus [33], and potato [34]. Here, we identi ed and annotated a total of 623 compounds of wampee fruits in different development stages, mainly for avones, saccharides, amino acids and organic acids.
Wampee has been very popular with consumers due to their abundant polyphenol and other bene cial substances, but high citrate content in mature fruit seriously restricts the development of the industry. Nevertheless, there were still little reports on differential metabolisms and genes of seedless wampee fruits in different developmental stages. Therefore, this study mainly focused on the differential mechanisms of metabolites during the fruit development of wampee fruits, especially sucrose and citrate. The transcriptome data showed that the differential genes mainly involved in energy production and transfer (C), amino acids translocation and metabolism (E), secondary metabolites synthesis, translocation and metabolism (Q) during fruit development, suggesting that the metabolite content of sucrose and organic acids in these pathways signi cantly changed. The transcriptome and metabolome analyses further showed that sucrose content, which might be mainly regulated by genes such as SUS, IVR, SPS, and TPS, was gradually accumulated throughout fruit development with gradually decreased accumulation amount. While the citrate content was increased rstly and then slowly decreased. Also, its synthesis might be mainly regulated by MDH, and its decomposition might be dominated by the acetyl-CoA pathway, suggesting that the GABA shunt pathway cannot determine citrate content.

Plant materials
Wampee trees grow in the Wampee Resource Nursery of the Ministry of Agriculture (Guangzhou). All the wampee trees used in this experiment were planted in 2008. The branches were planted with a spacing of 3 to 4 m, and all the lines were heading north to south. Fertilizer and water management and pest control were performed according to the resource nursery management manual of the Ministry of Agriculture. Fruits were harvested during the wampee developing and ripening season in 2018. Each sample constituted 120 fruits from three trees with the same size and color in expansion (PDQ stage, 50DAF), coloring (ZSQ stage, 70DAF), fruit ripening (CSQ stage, 91DAF), and delay harvest (WSQ stage, 100DAF). After the samples were randomly divided into three replicates and stored at -80 ℃ for further use. pH, Total acid (TA) and Total soluble sugars pH was evaluated utilizing a digital hand-held refractometer (pHS-25, Shanghai). Three droplets of juice attained via one fragment were evaluated, and the process was iterated two times to per fruit and with 12 individual fruit repeats. Total soluble sugars(TSS) concentrations were determined by the phenol-sulfuric acid method. Following the aforementioned evaluation, the remained fragments were separated into three categories. Each category of separated fragments was selected from four fruits, and also 5 ml of juice from each one of the categories was diluted well in 20 ml distilled water and then by using 0.1 N NaOH was titrated at

RNA-seq and Annotation
The total amount of utilized RNA per specimen was 3 µg, which was used as an input substance for the preparation of RNA specimen. It is worth to point that the minimum of three biological replications have assembled and blend simultaneously. The act of RNA-seq assembly and sequencing was accomplished via Biomarker Technology Co. (Beijing, China). The sequencing libraries were created through utilizing NEBNext® Ultra™ RNA Library Prep Kit for Illumina® (NEB, USA) and taking advantage of manufacturer's suggestions and criteria codes to allocate sequences to each specimen. The rst strand of cDNA was prepared by utilizing random hexamer primer as well as M-MuLV Reverse Transcriptase(RNase H − ). Further, the preparation of Second strand cDNA has done by utilizing RNase H and DNA Polymerase I. Through the exonuclease/polymerase activities, remaining overhangs were transformed into at ends. Following the adenylation of 3' ends of DNA segments, the adaptor of NEBNext containing hairpin loop con guration was illuminated in order to be prepared for hybridization. In continue, 3 µl USER Enzyme (NEB, USA) has implemented with the scale-selected, ligated adaptor cDNA for 15 min at 37 °C and extended at 95 °C for 5 min prior PCR. Then PCR was exerted within Phusion High-Fidelity DNA polymerase, Universal PCR primers as well as Index (X) Primer. Eventually, the products of PCR have completely puri ed (AMPure XP system) and the quality of the library has evaluated on the system of Agilent Bioanalyzer 2100. The clustering procedure of the index-coded specimen was exerted on a system of cBot Cluster Generation by implementing TruSeq PE Cluster Kit v3-cBot-HS (Illumina) conforming to the manufacturer's instructions.

Investigation of the Differentially Expressed Genes(DEGs)
DESeq R package (1.10.1) was used for a meticulous investigation of differential expression analysis for two conditions/groups.
DESeq is capable to represent statistical routines for ascertaining differential expression in digital gene expression data by utilizing a model founded on the negative binomial distribution. The calculated P values were modi ed by utilizing Hochberg and Benjamini's method for monitoring the rate of incorrect discovery. The genes with an tuned P-value < 0.05 based on DESeq were determined as differentially expressed.
qPCR Validation RNA extraction as well as quality evaluation was accomplished via RNASEq. Reverse transcription was carried out by utilizing HiFi-MMLV cDNA First-Strand Synthesis Kit (Invitrogen). For RT-qPCR, nine genes were picked within particular primers (Supplement 3) created by Primer Premier 5 software. The RT-qPCR was done with a detection system of ABI 7500 Fast Real-Time (Applied Biosystems) utilizing the Ultra SYBR Mix kit (CWBIO, Beijing, China). The system of ampli cation was formed from 10.4 µL Ultra SYBR Premix System II, 0.8 µL of 10 µmol/L upstream primer, 0.8 µL of 10 µmol/L downstream primer, 2 µL template, and sterile distilled water to a whole volume of 20 µL. The program of ampli cation was ful lled for 10 min at 95•C, succeeded by 40 cycles of 95•C for 15 s and, 55•C for 1 min. corresponding quantitative investigation of data was carried out through the 2 −△△CT approach within the reference genes actin-7. Three technical replications were implemented for each specimen to certify the reliability and reproducibility. Statistical investigation of variance (ANOVA) and Duncan's novel multiplex span tests were executed with SPSS Version 16.0 (Chicago, IL, USA). The signi cance level was adjusted to P < 0.05.

Figure 5
Regulation of sucrose metabolism trajectories during wampee fruit development and ripening. The contents of metabolites are exhibited in columns, whilst the expression values of genes are presented circles. The color from white to black displays the expression value from low to high. The expression trends were shown as RPKM.