Integrative analysis of metabolome and transcriptome profiles provides insight into the fruit pericarp pigmentation disorder caused by ‘Candidatus Liberibacter asiaticus’ infection

Background Mandarin ‘Shatangju’ is susceptible to Huanglongbing (HLB) and the HLB-infected fruits are small, off-flavor, and stay-green at the maturity period. To understand the relationship between pericarp color and HLB pathogen and the effect mechanism of HLB on fruit pericarp coloration, quantitative analyses of HLB bacterial pathogens and carotenoids and also the integrative analysis of metabolome and transcriptome profiles were performed in the mandarin ‘Shatangju’ variety with four different color fruits, whole green fruits (WGF), top-yellow and base-green fruits (TYBGF), whole light-yellow fruits (WLYF), and whole dark-yellow fruits (WDYF) that were infected with HLB. Results the HLB bacterial population followed the order WGF > TYBGF > WLYF > WDYF. And there were significant differences between each group of samples. Regarding the accumulation of chlorophyll and carotenoid, the chlorophyll-a content in WGF was the highest and in WDYF was the lowest. The content of chlorophyll-b in WGF was significantly higher than that in other three pericarps. There were significant differences in the total content of carotenoid between each group. WGF and TYBGF pericarps were low in phytoene, γ-carotene, β-cryptoxanthin and apocarotenal, while other kinds of carotenoids were significantly higher than those in WDYF. And WLYF was only short of apocarotenal. We comprehensively compared the transcriptome and metabolite profiles of abnormal (WGF, TYBGF and WLYF) and normal (WDYF, control) pericarps. In total, 2,880, 2,782 and 1,053 differentially expressed genes (DEGs), including 121, 117 and 43 transcription factors were identified in the three comparisons, respectively. The qRT-PCR confirmed the expression levels of genes selected from transcriptome. Additionally, a total of 77 flavonoids and other phenylpropanoid-derived metabolites were identified in the three comparisons. Most (76.65 %) showed markedly lower abundances in the three comparisons. The phenylpropanoid biosynthesis pathway was the major enrichment pathway in the integrative analysis of metabolome and transcriptome profiles. Conclusions Synthesizing the above analytical results, this study indicated that different color pericarps were associated with the reduced levels of some carotenoids and phenylpropanoids derivatives products and the down-regulation of proteins in flavonoids, phenylpropanoids derivatives biosynthesis pathway and the photosynthesis-antenna proteins. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-021-03167-3.

showed markedly lower abundances in the three comparisons. The phenylpropanoid biosynthesis pathway was the major enrichment pathway in the integrative analysis of metabolome and transcriptome profiles.
Conclusions: Synthesizing the above analytical results, this study indicated that different color pericarps were associated with the reduced levels of some carotenoids and phenylpropanoids derivatives products and the downregulation of proteins in flavonoids, phenylpropanoids derivatives biosynthesis pathway and the photosynthesisantenna proteins.
Keywords: 'Shatangju', Pigment, Carotenoid, Flavonoid, Metabolome, Transcriptome, Candidatus Liberibacter asiaticus Background Mandarin 'Shatangju' (Citrus reticulata cv. 'Shatangju') is one of the superior native citrus varieties in Guangdong Province, China. Its fruits are very popular because of its soft and delicate taste, smell, thin pericarp and some other advantageous fruit characteristics [1][2][3]. However, mandarin 'Shatangju' (Citrus reticulata cv. 'Shatangju') is susceptible to Huanglongbing (HLB), caused by Candidatus Liberibacter spp., which has greatly affected 'Shatangju' plantations. The typical symptoms of HLB-infected fruits are that the fruits are small, lopsided, and off-flavor, and their stylar end remains green at the maturity period, such as in sweet orange and mandarin 'Shatangju' [4,5]. However, some fruits of one HLB-infected 'Shatangju' tree are similar to the fruits of HLB-uninfected tree in terms of quality and appearance [6]. In previous research reports, the influence of HLB infection on citrus fruit yield and quality including internal and external quality have been reported [7]. But the influencing mechanism of HLB infection was not well clarified.
The carotenoid biosynthetic pathway is one of the important metabolic pathways in citrus. Its composition and content of carotenoid not only directly determine the color appearance of fruits, but also affect the nutrition. The carotenoids in citrus mainly include α-carotene, β-carotene, violaxanthin, antheraxanthin, lutein, zeaxanthin, α-cryptoxanthin, βcryptoxanthin, phytoene, lycopene and so on [14]. Depending on the cumulative difference of carotenoid, citrus varieties can be broadly divided into four categories: categories I that are rich in β-cryptoxanthin (such as Citrus reticulata Blanco), categories II that are rich in violaxanthin (such as Citrus sinensis Osbeck), categories III that are rich in phytoene (Citrus paradisi Macf) and categories IV that are low in carotenoids (Citrus limon L. Burm) [15,16]. The carotenoids not only are influenced by genetic factors, but also are affected by a variety of environmental factors, such as temperature, light and disease. The accumulation of β-cryptoxanthin in mandarin fruits was due to the high expression of upstream genes and the low expression of downstream hydroxylase genes [17]. The expression of carotenoid cleavage oxygenase (CCD) gene was significantly correlated with the accumulation of violaxanthin [18]. When the expression level of (E/Z)-phytoene, phytofluene and lycopene related synthase gene were high and the lycopene cyclase gene were low, the accumulation of lycopene and its precursors would increase substantially [19]. Although the related genes of carotenoid have been basically perfected, the effect mechanism of Huanglongbing on carotenoid still needs to be further analyzed.
Transcriptome analysis of citrus has successfully been used for identifying genes involved in the synthesis and signal transduction of photosynthesis, carbohydrate metabolism [20], jasmonic acid biosynthesis [21] and flavonoid biosynthesis [22]. Transcriptome profiling was applied to analyze the gene expression of the leaf [22], root [23] and pulp samples [24] that were related to HLB. However, transcriptome analysis of different colored pericarps of mandarin 'Shatangju' fruits associated with HLB was lacking. The application of metabolomics is widely used to investigate various fruits or vegetables and has significantly facilitated the identification of active metabolites and correlated metabolic pathways in various plants [25,26]. The combination of metabolomics and transcriptomics would help to reveal the biosynthetic mechanisms of key metabolic pathways. Therefore, it is feasible to analyze the carotenoid and flavonoid biosynthesis pathways in citrus by applying these two omics technologies.
The fruits color, as an important appearance and quality character of citrus fruits, has been widely studied [12,27]. In this study, the affected mechanism of fruit color by HLB was further clarified in WGF, TYBGF and WLYF fruits by combination of metabolomics and transcriptomics data. According to the metabolome results, the levels of some flavonoid compounds were drastically decreased and even could not be detected, whereas some flavonoids showed high accumulation in WGF, TYBGF and WLYF fruits compared with that in WDYF fruits. Fourteen kinds of carotenoids were identified and quantified in the four kinds of mandarin 'Shatangju' pericarps. Correlation network were mapped to highlight the regulatory genes associated with flavonoid and carotenoid metabolites by combination of genes and metabolites. These findings provide new insight into the molecular mechanisms associated with the biosynthesis and regulation of metabolites under the influence of HLB in mandarin 'Shatangju' pericarps and emphasize the usefulness of the integration method for understanding this process.

Plant materials
Citrus reticulata cv. 'Shatangju' for this study were grown at the citrus research orchard at Institute of Fruit Tree Research, Guangdong Academy of Agricultural Sciences, Guangzhou, Guangdong Province, China (E 113°21'59", N 23°9'15"). We chose ten HLB-infected mandarin 'Shatangju' trees that were selected from twenty-two trees by PCR. All trees were infected by Candidatus Liberibacter asiaticus and fresh fruits were collected from ten HLB-infected trees and mixed together in January 2019. all fruits were classified into four types according to the coloration: whole green fruits (WGF), top-yellow and base-green fruits (TYBGF), whole light-yellow fruits (WLYF), and whole dark yellow fruits (WDYF). The color was detected by a colorimeter (Minolta CR-300, Konica Minolta Investment Ltd, Shanghai, China). Three biological replicates were collected per sample, each with 60 fruits randomly collected from 10 HLB-infected mandarin 'Shatangju' trees. The fruits were transported back to the laboratory, and the pericarps were carefully excised with scalpels, collected, frozen in liquid nitrogen, roughly ground, and kept at -80°C for further research.

Exaction and detection of chlorophyll and carotenoid
The exaction and detection method of chlorophyll was acetone method by spectrophotometer. The exaction protocol of carotenoids was described as follows: plant materials (100 mg fresh weight) were frozen in liquid nitrogen, ground into powder, and extracted with nhexane: acetone: ethanol (2:1:1, V/V/V, Merck, Darmstadt, Germany). The extract was vortexed (30 s), ultrasound-assisted extraction was carried out for 20 min at room temperature, and centrifugation was performed (12,000 rpm for 5 min). The supernatants were collected, and the steps above were repeated. The supernatant was collected twice, then evaporated to dryness under a nitrogen gas stream, and reconstituted in 75 % methanol (V/V, Merck, Darmstadt, Germany). The solution was centrifuged, and the supernatant was collected for LC-MS analysis. Eighteen kinds of carotenoids were used for quantitative analysis. All of the carotenoid standards were purchased from Olchemim Ltd. (Olomouc, Czech Republic) and Sigma (St. Louis, MO, USA). Acetic acid was obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). The stock solutions of standards were prepared at a concentration of 10 mg/ mL in ACN. All stock solutions were stored at -20°C. The stock solutions were diluted with ACN to working solutions before analysis.

UHPLC conditions and APCI-Q TRAP-MS/MS for carotenoids
The detection protocol of carotenoids by UHPLC was described as follows: the carotenoid extracts were analyzed using an LC-ESI-MS An API 6500 Q TRAP LC/MS/MS System was equipped with an APCI Turbo Ion-Spray interface, operated in positive ion mode and controlled by Analyst 1.6.3 software (AB Sciex, USA). The APCI source operation parameters were as follows: ion source, turbo spray; source temperature, 350°C; curtain gas (CUR), 25.0 psi; and collision gas (CAD), medium. DP (declustering potential) and CE (collision energy) for individual MRM transitions was performed with further DP and CE optimization. A specific set of MRM transitions were monitored for each period according to the carotenoids eluted within this period.

Extraction and qualitative and quantitative analysis of metabolites
The extraction and separation of metabolites of mandarin 'Shatangju' fruit pericarps were reported previously [27]. In brief, ca. 100 mg fine powder of per sample was extracted with 70 % methanol overnight at 4 ℃. Then, the sample was centrifuged and filtered to obtain the extract of pericarps before LC-MS analysis. The qualitative and quantitative analyses of metabolites were performed by ESI-Q TRAP-MS/MS by a professional metabolomics company (Metware Biotechnology Co., Ltd., Wuhan, China). The significantly changed metabolites (SCMs) were filtered according to | Log 2 (Fold Change) | ≥1 and p-value < 0.05. The specific procedures refer to Wang et al [27].

RNA extraction and illumina sequencing
The isolation and purification of RNA, construction of cDNA libraries, and sequencing were completed by a professional transcriptome sequencing company (Novogene Bioinformation Technology Co., Ltd., Beijing, China). The filtered reads were mapped to the Citrus sinensis genome(https://www.citrusgenomedb.org/ organism/Citrus/sinensis) using HISAT2 software [31]. Reads Per Kilo Base per Million mapped reads (FPKM) was used as an indicator for gene and transcript level quantification [32]. The differentially expressed genes (DEGs) were determined by the absolute value of log 2-FoldChange (log 2 FC) greater than 1 and false discovery rate (FDR) ≤ 0.05. All DEGs were analyzed by gene ontology (GO) [33] and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment using KOBAS software [34]. The transcription factors were predicted by ITAK software [35].

Quantitative real-time PCR (qRT-PCR) validation
Total RNA extraction was carried out by the polysaccharide polyphenol plant Total RNA Fast Extraction Kit (BioTeke, Beijing, China). Reverse transcription was performed using a PrimeScript II First Strand cDNA Synthesis Kit (Takara, Ishiyama, Japan). Twelve genes were selected for qRT-PCR with specific primers (Supplemental Table S1). The qRT-PCR was performed with a StepOnePlus Real-Time PCR System (Applied Biosystems, USA) using a TB Green SYBR Kit (Takara, Ishiyama, Japan). The amplification system consisted of 10.4 μL of TB Green SYBR Premix System II, 0.4 μL each of 10 μmol/L forward and reverse primer, 2 μL of cDNA template, and sterile distilled water to a total volume of 20 μL. The amplification program was 95 ℃ for 30 s, followed by 40 cycles of 95 ℃ for 5 s and 60 ℃ for 30 s. The 2 −ΔΔct method was performed to analyze relative quantitative data with a reference gene, β-actin. Three technical replicates were carried out for each sample to ensure reproducibility and reliability.

Integrative analysis of metabolome and transcriptome data
Carotenoids and the SCMs in the metabolome and DEGs in the transcriptome were chosen for integrative analysis. SCMs were filtered on the basis of variable importance in the projection (VIP) greater than 1, p-value less than 0.05, and absolute values of Log 2 FC greater than 1 for correlation analysis. Pearson correlation coefficients and p-values were calculated for carotenoids and SCMs and DEGs data integration using the Spearman method.
Statistical analysis SPSS 22.0 was performed for statistical analysis. Data are presented as means ± standard deviations (SD). The levels of statistical significance were analyzed by the least significant difference (p-value < 0.05).

Phenotypic analysis of fruit appearance
The fruits used in the present study all were harvested at the fruit maturity. The development of mandarin 'Shatangju' fruits is very rapid at the last maturity period. The fruits quickly enlarge and color so that they reach their final coloration and size in two months. However, if the trees are infected by the HLB-associated pathogen, Candidatus Liberibacter asiaticus, the fruit development will be affected. Herein, we grouped the fruits collected from HLB-infected trees at the mature stage into four types, namely, whole green fruits (WGF), top-yellow and base-green fruits (TYBGF), whole light-yellow fruits (WLYF), and whole dark-yellow fruits (WDYF, CK), based on their coloration and size (Fig. 1, Supplemental Table S2a). Both WGF and TYBGF fruits can't entirely turn dark-yellow in appearance. The WGF fruits were whole green (H 0 = 104.49, CI (color index, a/b ratio) = -0.29) and the smallest (32.68 × 30.71 mm). TYBGF fruits were bigger (35.62 × 33.77 mm), and the color of the pedicel parts was lighter-yellow (H 0 = 100.3, CI=-0.18) than that in WGF fruits. The vertical and horizontal diameters of WGF and TYBGF fruits both were significantly lower than that of WDYF and the WGF were lowest (Supplemental Table S2a). The WLYF fruits could not completely develop dark-yellow pericarp so that they all were light-yellow (H 0 = 72.72, CI = 0.31), and the WDYF fruits were the darkest yellow fruits (H 0 = 55.12, CI = 0.70) in all samples, but there was no significant difference in the vertical and horizontal diameters between the WLYF and WDYF fruits (Supplemental Table S2a).

Quantitation of the HLB bacterial population in pericarps
By constructing the recombinant plasmids of 16 S rDNA and establishing a standard curve between plasmid copy Fig. 1 Phenotypes of the four-color types of fruits from HLB-infected mandarin 'Shatangju' trees. WDYF (CK): whole dark yellow fruits; WLYF: whole light-yellow fruits; TYBGF: top-yellow and base-green fruits; WGF: whole green fruits number and CT value of qRT-PCR, the equation, y=-3.6448x + 46.822 (R²=0.9926) was established (Fig. 2 A). According to the equation and mean Ct values of qRT-PCR of HLB bacterial pathogen in the four types of pericarps, the population of HLB bacterial pathogens in per nanogram of total DNA were calculated. The results show that the population of Candidatus Liberibacter asiaticus in WGF, TYBGF, WLYF and WDYF fruit pericarps were 9.33 × 10 4 , 7.03 × 10 4 , 5.41 × 10 4 , and 1.03 × 10 4 cells/ng of total DNA, respectively (Fig. 2B). According to significant difference analysis, WGF, TYBGF and WLYF fruit pericarps were all significantly higher than WDYF fruit pericarps in the population of Candidatus Liberibacter asiaticus. The WGF had the highest population of HLB bacterial pathogens and also was most seriously affected in phenotype.

Quantitative analysis of carotenoid in all samples
The total contents of chlorophyll in WGF, TYBGF, WLYF and WDYF showed a gradually decreasing trend, and was 7.69 × 10 − 2 , 5.85 × 10 − 2 , 4.83 × 10 − 2 and 4.17 × 10 − 2 mg/g, respectively (Fig. 3). The content of chlorophyll a in WGF was the highest, followed by TYBGF, with no significant difference between them, and the content of chlorophyll a in WDYF was the lowest. The content of chlorophyll b in WGF was the highest, 3.48 × 10 − 2 mg/g, and showed significant difference with other three types of fruit pericarps. There was no significant difference among other three types of fruit pericarps. Among the four types of fruit pericarps, there were significant differences in the total content of carotenoid. WGF, TYBGF, WLYF and WDYF were 7.31 × 10 − 2 , 8.47 × 10 − 2 , 1.61 × 10 − 1 and 2.48 × 10 − 1 mg/g (Fig. 3).
The content of eighteen kinds of carotenoids were identified and detected but fourteen kinds of carotenoids were identified in WGF, TYBGF, WLYF and WDYF fruit pericarps (Supplemental Table S2bcd and Fig. 4). The identified carotenoids didn't include phytofluene, capsanthin, ε-carotene and capsorubin. The results show that phytoene, lutein, β-carotene, neoxanthin and violaxanthin were the major carotenoid in all samples. Except γ-carotene, phytoene, β-cryptoxanthin, apocarotenal and astaxanthin, the content of other nine kinds of carotenoids in WGF and TYBGF fruits were significantly higher than those in WDYF samples (Fig. 4). The contents of α-carotene, β-carotene, lutein, neoxanthin, zeaxanthin and antheraxanthin in WLYF samples were significantly higher than those in WDYF samples. There was no significant difference between the contents of αcryptoxanthin, β-cryptoxanthin, γ-carotene, phytoene, lycopene and violaxanthin in WLYF samples and that in WDYF. And only apocarotenal in WLYF sample was significantly lower than that in WDYF samples. In addition, astaxanthin did not show any significant difference in all comparisons (Fig. 4).

Differential accumulation of phenylpropanoids derivatives in the three comparisons
To compare the differences of primary and secondary metabolism between the three treatment groups and the control group (CK), the fruit pericarps were subjected to LC-ESI-MS/MS analysis. In the present study, most of significantly changed metabolites (SCMs) were flavonoids and other phenylpropanoid derivatives, accounting for 35.81 %. According to the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, the phenylpropanoid biosynthesis pathway was the most significantly enriched metabolic pathways in all comparisons [27]. The numbers of differently accumulated flavonoids and phenylpropanoid derivatives were 60, 53 and 29 in WGF    genes were detected in the WGF, TYBGF, WLYF and WDYF groups, respectively. A high correlation coefficient (R 2 > 0.99) of gene expression between biological replicates indicated the effectiveness of the data (Fig. 5 A).
With the filter criteria of |log 2 FoldChange| ≥1 and false discovery rate (FDR) < 0.05, there were 2880, 2782, and 1053 differentially expressed genes (DEGs) detected in the three comparisons, WGF vs. WDYF, TYBGF vs. WDYF and WLYF vs. WDYF, respectively, of which 1538, 1468 and 772 genes were upregulated, and 1342, 1314 and 281 genes were downregulated in the three comparisons: WGF vs. WDYF, TYBGF vs. WDYF, and WLYF vs. WDYF, respectively (Supplemental Tables S5, Fig. 5B and D). Venn diagram analysis shows that there were 705 DEGs that were common to the three comparisons (Fig. 5 C). The majority of DEGs in the all comparisons were up-regulated. The greater phenotypic difference of the sample, the more differentially expressed genes of the sample.
The Gene Ontology (GO) analysis was conducted to investigate the roles of HLB-affected genes in the biological process (BP), cell component (CC) and molecular function (MF) terms. The GO analysis of 2880 DEGs in WGF comparison and 2782 DEGs in TYBGF comparison showed some enrichment of five major biological processes, including cellular process, metabolic process and response to stimulus; and 1053 DEGs in WLYF comparison mainly enriched in metabolic process and cellular process by GO analysis. In the molecular function aspect, most of DEGs in the three comparisons all were enriched in catalytic activity and binding. In the cellular component aspect, those DEGs of the three comparisons all were mainly enriched in five categories including cell part, membrane and organelle (Supplemental Fig. S1). Top 20 remarkably enriched GO terms in respect of CC, MF, and BP enriched in the three comparisons were summarized in supporting information (Supplemental Fig. S2). Briefly, several GO terms were associated with photosynthesis process (e.g., GO: 0019684 photosynthesis, light reaction, GO:0009521 photosystem, GO:0009522 photosystem I, GO:0015979 photosynthesis) in the three comparisons. Thus, these GO terms were suggested to play vital roles in the HLBcaused abnormal color of 'Shatangju' mandarin fruits.
KEGG analysis (p-value < 0.05) revealed that all of these DEGs in all comparisons were mainly enriched in 11 metabolic processes. The top three significantly enriched metabolic pathways include photosynthesisantenna proteins, steroid hormone biosynthesis and phenylpropanoid biosynthesis in WGF vs. WDYF, and photosynthesis-antenna proteins, phenylpropanoid biosynthesis and photosynthesis in TYBGF vs. WDYF and WLYF vs. WDYF (Supplemental Table S6 and Supplemental Fig. S3ACE). Therefore, two metabolic pathways, photosynthesis-antenna proteins and phenylpropanoid biosynthesis, were enriched in all comparisons. Among these DEGs detected in WGF vs. WDYF, TYBGF vs. WDYF and WLYF vs. WDYF, we found 41, 40 and 23 glycosyltransferase genes, 44, 41 and 5 methyltransferase genes, and 25, 23 and 10 acyltransferase genes that catalyzed the synthesis of different types of flavonoids. We also detected 12, 12 and 4 GST (glutathione-S-transferase) genes, 17, 18 and 5 ABC transporter genes, 37, 38 and 13 MATE (multidrug and toxic compound extrusion) genes, and 1, 1 and 0 SNARE (soluble NSF attachment protein receptor) genes, which might play important roles in transporting flavonoids (Supplemental Fig. S3BDF).

Integrative analysis of transcriptome and metabolome
We combined the analysis of transcriptome and metabolome data to investigate the association between metabolites and genes involved in the same biosynthesis pathway in 'Shatangju' fruit pericarps. Results showed that p-values of metabolites and genes involved in phenylpropanoid biosynthesis both were less than 0.01 (Supplemental Fig. S4) in the three comparisons. To understand the regulatory mechanism of decreased flavonoids and other phenylpropanoid biosynthesis in the WGF, TYBGF and WLYF fruit pericarps, 60, 53, and 29 SCMs (31, 27 and 11 flavonoids and 29, 26 and 18 other phenylpropanoid-derived metabolites, Supplemental  Table S7) were subjected to analysis using Pearson's Correlation Coefficient. Results showed strong positive or significant negative correlations (R > 0.8 or <-0.8 and p-value < 0.05) between the SCMs and DEGs (Supplemental Table S8). Based on these results, the interaction networks between the SCMs and DEGs (R > 0.9) were organized in the WGF, TYBGF and WLYF comparisons (Supplemental Fig. S5). At the same time, the nine-quadrant model was proposed to delineate the relationship of the SCMs with DEGs in the three comparisons. According to the nine-quadrant analysis, the significantly changed flavonoids and other phenylpropanoid-derived SCMs and DEGs (R > 0.8 or <-0.8) of the associated metabolic pathways were focused mainly in the 1, 3, 7 and 9 quadrants (Supplemental   Fig. S6). The 3 and 7 quadrants indicate that differential patterns of SCMs and DEGs were consistent, but the 1 and 9 quadrants indicate the inconsistency.
Flavonoids play many different roles throughout the plant's life cycle. In this study, a large number of flavonoids were identified and quantified in the WGF, TYBGF, WLYF and WDYF fruit pericarps (Supplemental Table S3). In the three comparisons, fifteen DEGs were annotated in the flavonoid biosynthesis pathway in citrus (Supplemental Table S7). In flavonoids, most SCMs were markedly decreased in WGF, TYBGF and WLYF fruit pericarps relative to those in WDYF fruit pericarps,  Table S3).
Six DEGs (the homologs of CHS, HCT (hydroxy cinnamyl transferase), HIDH (2-hydroxyisoflavanone dehydratase), IF7MAT (isoflavone 7-O-glucoside-6''-O-malonyl transferase) and UGT79B1 genes were screened to validate the transcriptomic results. The qRT-PCR results showed that the expression level of a homologs of CHS (Cs3g20330), two homologs of HCT (Cs6g12870 and Cs6g12880), a homolog of IF7MAT (Cs5g21190) a homolog of HIDH (Cs9g06490) and a homolog of UGT79B1 (Cs1g08110) were lower in WGF, TYBGF and WLYF fruit pericarps than that in WDYF fruit pericarps (Fig. 6). These results were consistent with the RNA-Seq data. Most homolog genes of the flavonoid biosynthesis pathway in citrus showed lower expression levels in WGF, TYBGF and WLYF than that in WDYF fruit pericarps, implying vital function of these enzyme genes in citrus flavonoid biosynthesis.

Discussion
Carotenoids are a large class of natural lipid-soluble pigments and are important for plant including citrus plant [14]. More than 1000 kinds of carotenoids have been identified from natural sources [15]. Carotenoid pigments massively accumulate in citrus fruit. And their contents and compositions are important indexes for the nutritional and commercial quality [41]. Previous studies have shown that the contents of β-carotenoid, lutein, and α-carotene decreased rapidly to a low level, and massive accumulation of β-cryptoxanthin, zeaxanthin and violaxanthin occurred in some varieties during the ripening process [41]. In this study, we identified and determined the contents of fourteen kinds of carotenoids in the four types of mandarin 'Shatangju' fruit pericarps  Table S2, Fig. 4). The highest accumulation of phytoene was in the WLYF group, while the lowest accumulation was in the WGF group (Fig. 4). The α-carotene and lutein were significantly higher in WGF, TYBGF and WLYF than those in WDYF samples. The γ-carotene and β-cryptoxanthin, astaxanthin and apocarotenal in the WDYF group were significantly higher than those in the other three groups. This was consistent with previous studies. This indicated that WGF, TYBGF and WLYF has not reached the full maturity stage. Several TFs have been implicated in the regulation of carotenoid accumulation during fruit ripening, including the MADS-box TFs RIN, TAGL1, and TDR4 [42][43][44][45], the AP2/ERF family TFs SlAP2a and SlERF6 [46,47] and MYB family TFs RCP1 [48]. According to the transcriptome analysis and the correlation analysis between transcription factors and carotenoids, multiple genes, such as AP2/ERF family TFs RAP27 (Cs8g17390) and NAC family TFs NAC22 (Cs8g01860), NAC62 (orange1.1t00590) and NAC45 (Cs7g11190), and  Table S11). These data indicate that these TFs may contribute to the synthesis of carotenoids and be affected by the invasion of HLB. This aspect should be further studied to confirm the effect mechanism.
Carotenoids and flavonoids both are important factors for the pigment of citrus fruits. The transcriptional level of flavonoid related genes was regulated by transcription factors (TFs), such as MYB-bHLH-WD40 (MBW) complex [39,48,49], WRKY [50] and MADS [51]. MYB transcription factor is a decisive regulator in the MBW complex. It positively or negatively regulates structural genes in the flavonoid pathway and maintains the balance of flavonoids in plant organs. The MdMYB1 in apple [52], PpMYB10 in peach [53] and VvMYBA1, VvMYBA2, VvMYBA-pa2 in grape [54,55] belong to the positive regulation of MYB factors and improve the transcription level of flavonoid related genes. FaMYB9/ FaMYB11 [56] and FaMYB1 in strawberry [57], PpMYB18 in peach [58], Vvmybc2-l3 and Vvmybc2-l2 in grape [59], MdMYB16 and MdMYB15L in apple [60,61] were negative regulation of MYB factors. They inhibit the synthesis of anthocyanins. In the present study, the transcriptome profiles of abnormal pigment fruit pericarps and WDYF (CK) fruit pericarps were analyzed. And it was found that some TF families, including MADS, WRKY, MYB and bHLHs, showed significantly differential expression levels. As such, 66.67 % 80.77 and 66.67 % of MYBs were significantly down-regulated in WGF, TYBGF and WLYF comparisons. All of the differentially expressed TFs may be candidate genes that regulated flavonoid biosynthesis in citrus.
Transcriptomics integrated with metabolite analysis revealed that most flavonoids and other phenylpropanoidderived metabolites and some carotenoids (phytoene, γ-carotene, β-cryptoxanthin, astaxanthin and apocarotenal) showed drastically lower abundances in WGF, TYBGF and WLYF than in WDYF (CK) fruits. Based on chemical structures of the flavonoids and other phenylpropanoid-derived metabolites, the putative roadmaps of flavonoids and other phenylpropanoid biosynthesis pathways were drawn. The levels of three flavanones (butein, eriodictyol and phloretin), the common precursors of flavonoids, were decreased in WGF vs. WDYF, especially eriodictyol in all comparisons. It should be noted that five metabolites (6,7dimethoxy-4-methylcoumarin, acetyl-eriodictyol ohexoside, myricetin, tricin 4'-o-β-guaiacylglycerol and p-coumaraldehyde) in all flavonoids and other phenylpropanoid-derived products showed the most obviously decreased abundances in all comparisons, especially myricetin and tricin 4'-o-β-guaiacylglycerol. 6,7-Dimethoxy-4-methyl-coumarin and myricetin are known to have significant antibacterial activity [62]. Therefore, this result may cause WGF, TYBGF and WLYF fruit pericarps to be high in terms of the population of pathogenic bacteria.

Conclusions
In this study, we explored that the population of pathogenic bacteria followed the order WGF > TYBGF > WLYF > WDYF. In regard to chlorophyll and carotenoid abundance, the WGF had the highest total content of chlorophyll and the lowest total content of carotenoid. WGF and TYBGF fruit pericarps were low in phytoene, γ-carotene, β-cryptoxanthin, astaxanthin and apocarotenal, and the levels of other carotenoids were significantly higher than those in WDYF fruit pericarps. And WLYF was only short of apocarotenal. Integrative analysis of transcriptomic and metabolomic data indicated that the influencing mechanisms of HLB infection on the mandarin 'Shatangju' fruit pericarps may involve phenylpropanoid-derived metabolic pathways, and extensive down-regulation genes of photosynthetic-related protein. This work not only provides insights into the molecular and biochemical mechanisms of HLBinfection on the mandarin 'Shatangju' fruit pericarps but may be of significance in uncovering the unknown gene for responding to HLB.