Differential Expression Analysis of Genes Related to Flesh Color in Hylocereu polyrhizus and Hylocereu undatus

Pitaya is a burgeoningly tropical fruit, there are two common varieties, red peel with red pulp and red peel with white pulp. To investigate the reason of different colors in two kinds of pitaya, explore differencially expressed genes (DEGs), in this study, we profiled transcriptome in flesh of two varieties (Hylocereu polyrhizus and Hylocereu undatus) in green stage and mature stage respectively. 53 240 reads with high quality were obtained. Analysing the DEGs, we found there were bigger differences in different stage rather than different variety. Gene ontology (GO) functional analysis and KEGG pathway analysis were adopted in R1 VS R3 and W3 VS W3. Through GO analysis, we found many genes enriched in protein binding transcription factor activity in molecule founction. Through KEGG pathway analysis, we found most genes were enriched in biosynthesis of secondary metabolism, amino acid and nucleic acid metabolism, and tyrosine metabolism. In addintion, tyrosine was the precursor of betalaines. In conclusion, the development of Hylocereu polyrhizus accompanying by the synthesis of a large amount of betalaines, tyrosine played a key role as a precursor, and the synthesis of this pigment required a mass of amino acids, enzymes, and transcription factors. Betalaines synthesis pathway has not cleared yet, this study explored key genes related to betalaines synthesis and provided useful information for optimizing the betalaines synthesis pathway, which was the basis of later experiments.

present, gas chromatography-mass spectrometry (GC-MS) technology has been used to study the metabolic components of the peel and pulp of red flesh pitaya. It is found that Hylocereu polyrhizus will accumulate a large amount of betalanins (Wu et al., 2019) during mature period. Betalanins is a kind of water-soluble pigment, which has two forms: betaine (purplish red) and betaxanthin (orange) (Duan et al., 2017). Both of them are synthesized from tyrosine, which is a precursor. Betalanins has good effects in health, such as antioxidation, anticancer, scavenging free radicals and reducing blood lipids, but it will easily degrade under the conditions of light, heat, oxygen and metal ions (Fe 3+ , Cu 2+ , Al 3+ ). When pH is 4-5, it is the most stable, and exists in many plants such as amaranth, Chenopodiaceae, Jasminaceae and Phytolaccaceae (Chen et al., 2013). Betalanins is composed of beet aldehyde acid and imine of cyclodopa, in which the conjugated double bond system forms its chromophoric group. The conjugated double bond of cyclic dopa in aromatic group makes the absorbance of betaine 50~70nm higher than that of beet flavin. The difference of pigment content may be an important factor leading to the difference among varieties (Wu et al., 2019). The biosynthesis of this natural plant pigment is more complex. At present, the studies on its physiological activity at home and abroad have been more thorough, but the metabolic pathway and molecular mechanism are still not clear. For many years, academia has speculated that glucosyltransferase (GT) and tyrosinase (TYR) are involved in the metabolic pathway of betaine (Wyler et al., 1984;Gandía and García, 2013;Sakuta, 2014), but it has not been strongly verified. In the squama of dragon fruit, vigs (virus-mediated gene silencing) technique was used to make the breaking stop and change color to yellow by silencing HmB5GT or HmHCGT2. After detecting the content of betaine and betaxanthin respectively, it was found that they all decreased, among which the content of betaine decreased more significantly (Xie et al., 2020).
Transcriptome sequencing technology can determine the amount of all mRNA transcribed by a certain tissue or cell of a species in a specific period, which can be used as a basis to determine the expression of some genes in this period, which can provide an important reference for gene transcription in a certain state. At present, it has been widely used in the field of molecular biology, and it is a convenient and fast technical method (Wilhelm and Landry, 2009). In this experiment, Hylocereu polyrhizus (R) and Hylocereu undatus (W) were used as research materials. Through transcriptome sequencing and follow-up analysis, the DEGs in the pulp of two different types of pitaya fruit was compared, and the related gene function and metabolic pathway enrichment analysis were carried out. The DEGs were screened out, and the metabolic differences of pigments in red and white pitaya fruits were preliminarily analyzed. Through the functional verification experiment of the screened genes, we can provide reliable information for the synthesis and metabolism of sugar betalanins.  Table 2). Through the statistical analysis of the length of the assembled sequence (Figure 1), excluding the head and tail, the length distribution of the sequence is uniform, which shows that the assembly result is reliable.    (Figure 4), 6 498 genes were found to be associated with general functional prediction and 3 224 genes associated with post-translational modification, protein transport and molecular chaperone. The top 10 categories showed that these proteins were mainly related to gene transcription, translation, protein processing and intracellular transport, and 705 genes were related to secondary metabolites.
GO (Geneontology) functional annotation analysis ( Figure 5) enriched genes in biological processes, cellular components and molecular functions in all samples, in which more genes were enriched in cellular components and less in molecular functions. In the biological process, 7 892 genes are enriched in the metabolic process, 7 637 genes in the intercellular process, 5 774 genes in the single organism process, and 2 165 genes in biological regulation and response to stimulus. These processes are closely related to the metabolism of pigment. Among the cellular component, it can be seen that cells or part of cells, organelles or part of organelles, and membranes or partial membranes are enriched into more genes; in molecular function, 6 496 genes are enriched in catalytic activity, 6 065 genes are enriched in binding function, and 6 065 genes are enriched in transport activity. In addition, some genes are enriched in the binding activities of proteins and transcription factors.  (Table 2). According to the criteria of FDR < 0.05 and | log2FC | > 1, pairwise comparison was used to filter among the four samples (Table 3). It can be seen that the DEGs of the same variety in different periods are more than those of different varieties in the same period. In R1 VS R3, a total of 10 077 genes were significantly expressed, of which 6 651 were up-regulated and 3 426 down-regulated (Table 3). GO (GeneOntology) analysis of DEGs ( Figure 6) showed that four parts were obviously enriched to more genes in molecular function, catalytic activity, connection, transport activity, nucleic acid binding activity with transcription factor, which had an important relationship with pigment metabolism. In the KEGG metabolic pathway analysis (Figure 7), more genes were enriched in the secondary metabolite synthesis pathway, amino acid and nucleic acid metabolic pathway and tyrosine metabolic pathway. In the comparative analysis of W3 VS R3, we got 7 067 DEGs, of which 2 714 were up-regulated and 4 353 down-regulated (Table 3). Based on the progressive GO enrichment analysis of all DEGs (Figure 8), we found that the up and down regulated genes were basically enriched in the same functional region, similar to R1 VS R3 in category, and enriched in catalysis, transport, binding, transcription factor activity, and protein binding. Three up-regulated genes were significantly enriched in antioxidant function (Unigene 0 022 428; Unigene 0 023 466; Unigene 0 037 666), and no down-regulated genes were enriched in this molecular function. Further enrichment analysis of KEGG metabolic pathways was carried out on 7 067 DEGs (Figure 9). It was found that in addition to

Verification of transfer group data by qRT-PCR
Based on the analysis of KEGG metabolic pathway, this study selected the key enzyme genes involved in betalanins metabolic pathway or upstream pathway. Genes' ID were used to obtain the sequence data of related annotated genes. Through the KEGG database, the key enzyme genes and other key genes in the related metabolic pathway were screened out. With the RPKM as the reference, 18 transcripts were randomly selected for qRT-PCR verification (Table 4), with UBQ as the internal reference gene (Table 5). The results showed that the expression patterns of the 18 transcripts were consistent with the RNA-Seq data ( Figure 10). These results show that the transcriptome data of this study are reliable and can be studied further.

Disscussion
Pulp's conversion of color is a complex process, which requires the regulation of various genes, enzymes and transcription factors (Hatlestad et al., 2015), as well as light, temperature and other natural factors. As the key substance, betalanins is synthesized from tyrosine ( According to the quantitative results, this study found that the transcriptional Unigene 0017646 of tyrosine permease isomer 1 related to tyrosine transport was up regulated in R3, but not significantly. This expression pattern is similar to the transcriptional Unigene 0051695. The L-Dopa catalyzed by Unigene 0051695 encoding DOD, TYR is the substrate (Hua et al., 2016) of DOD. Unigene0004253 and Unigene0006446 can encode and synthesize CytP450. These two transcripts are up-regulated in R3, and the Cyclo-Dopa catalyzed by them is necessary for the synthesis of beet pigment in sugar beet (Hatlestad et al., 2012). GT can transfer carbohydrate groups to form betaine. We randomly selected four GT-related transcripts for verification, but only one showed an up-regulation trend in R3, which was not consistent with the prediction of this study, indicating that GT has a more complex expression pattern and needs further research. With the development of Hylocereu polyrhizus, the gene expression level of some key enzymes related to the synthesis of betalanins increased, betalanins gradually accumulated in the flesh, and the flesh turned red, but there was no betalanins in Hylocereu undatus for the expression of these key genes was too low, its color did not change (Hua et al., 2016).   The genes of all samples were analyzed by GO, and it was found that more genes would be enriched in catalytic activity, ligation, transport activity, nucleic acid and transcription factor binding activity, indicating that pigment metabolism needs a variety of enzymes and acting factors to regulate together. The pairwise comparison of four samples (W1 VS W3, R1 VS R3, W1 VS R1, W3 VS R3) showed that there were more DEGs in the same variety at different stages (R1 VS R3, 10 077), which indicated that the difference of flesh color was more closely related to the fruit development stage. With the growth and development of Hylocereu polyrhizus, some key genes related to color may play a role in promoting the synthesis of betalanins. Therefore, in this study, GO and KEGG analysis of R1 VS R3 showed that most of the enriched genes were closely related to enzyme activity, and more genes were enriched in secondary metabolic pathway and tyrosine metabolic pathway. Betalanins is a secondary metabolite synthesized from tyrosine (Duan et al., 2017). There may be key genes related to fruit color development in R1 VS R3. We also analyzed different mature samples (W3 VS R3) and found that some genes were enriched in phenylalanine metabolic pathway, flavonoid and flavonol metabolic pathway. Phenylalanine is the precursor of tyrosine synthesis (Zhao et al., 2017), while flavonoids and flavonols are closely related to the synthesis of flavonoid pigments. The color composition of pitaya is more complex, not only betalanins, there may also be flavonoid pigments such as anthocyanins (Zeng, 2017). Recent studies have shown that there may be competition between betalanins and anthocyanin metabolic pathway in pitaya. Anthocyanin contributes to the development of pitaya's color partly (Fan et al., 2020).
To sum up, among the four genes related to tyrosinase or tyrosine synthesis (Unigene0017646 may be related to tyrosinase synthesis; Unigene0023505 may be related to prephenylalanine aminotransferase synthesis; Unigene0028648 is related to phenylalanine hydroxylase synthesis; Unigene0039772 is related to preaniline dehydrogenase synthesis. All three are key enzymes in upstream metabolism of tyrosine synthesis). Tyrosine permeability enzyme isomer 1 encoded by Unigene0017646 contributes to the transport of tyrosine and provides raw materials for the synthesis of beet pigment; two (Unigene0004253 and Unigene0006446) are related to the formation of cytochrome P450 (CytP450), and their expression patterns are consistent with previous studies. (Hua et al., 2016); two (Unigene0025394 and Unigene0019344) are related to the synthesis of flavonoids. Seven (Unigene0003029, Unigene0003132, Unigene0012439, Unigene0047695, Unigene0037073, Unigene0031027 and Unigene0015977) were related to the synthesis of glucosyltransferase (GT) and one (Unigene0051695) was related to the synthesis of DOPA dioxygenase (DOD). In some Caryophylla plants, some genes in the anthocyanin biosynthesis pathway are deleted, which triggers the betalanins biosynthesis pathway. DOD is the key enzyme for the formation of chromogenic groups in the sugarbeet pigment metabolism pathway, and it is also the key enzymes related to the expression of key genes synthetising betalanins (Xu et al., 2013;Wu et al., 2019). One (Unigene0048533) belongs to the ERF (ethylene response factor, ERF) family gene of ethylene response factor, ERF is closely related to plant fruit ripening (Ma et al. 2017), and one (Unigene0042389) is related to the synthesis of colorless anthocyanin reductase. We can select the genes with higher multiple of difference and more significantly expressed in R3 samples for more in-depth study.

Collection and processing of test materials
The experimental materials were collected from the dragon fruit plantation in Liguo Town, Dongfang City, Hainan Province (19°09'N, 108°64'E), and the tree age was 5 years old. Two varieties' Vietnam White flesh and "Jindu No.1" (Red flesh) were selected, and the fruit samples of green stage (15 days after pollination) and mature stage (30 days after pollination) were collected and marked as W1 (white flesh in green stage), W3 (white flesh in mature stage), R1 (red flesh in green stage) and R3 (red flesh in flesh stage). They were cutted into pieces and packed in 50 mL centrifuge tube in laboratory and stored in -80°C refrigerator after liquid nitrogen quick freezing.

Experimental steps for RNA-Seq
The pulp tissue samples of green and mature fruit were sent to Shenzhen Chideo Biological Company for transcriptome sequencing based on IlluminaHiSeqTM 2 000. After the total RNA was extracted from the sent samples, all mRNA, was enriched by magnetic beads method, and then the collected transcripts were decomposed into short fragments of 200~700nt with decomposition buffer. Using them as templates, random primers and reverse transcriptase were used to synthesize cDNA. After PCR amplification with suitable fragments, the complete piyata library was obtained, and the library was sequenced by IlluminaHiSeqTM 2 000.