A Genome-Wide Identification Study Reveals That HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT Involved in Betalain Biosynthesis in Hylocereus

Betalains are water-soluble nitrogen-containing pigments with multiple bioactivities. Pitayas are the only at large-scale commercially grown fruit containing abundant betalains for consumers. Currently, the key genes involved in betalain biosynthesis remain to be fully elucidated. Moreover, genome-wide analyses of these genes in betalain biosynthesis are not available in betalain-producing plant species. In this study, totally 53 genes related to betalain biosynthesis were identified from the genome data of Hylocereus undatus. Four candidate genes i.e., one cytochrome P-450 R gene (HmoCYP76AD1), two L-DOPA 4,5-dioxygenase genes (HmoDODAα1 and HmoDODAα2), and one cyclo-DOPA 5-O glucosyltransferase gene (HmocDOPA5GT) were initially screened according to bioinformatics and qRT-PCR analyses. Silencing HmoCYP76AD1, HmoDODAα1, HmoDODAα2 or HmocDOPA5GT resulted in loss of red pigment. HmoDODAα1 displayed a high level of L-DOPA 4,5-dioxygenase activity to produce betalamic acid and formed yellow betaxanthin. Co-expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT in Nicotiana benthamiana and yeast resulted in high abundance of betalain pigments with a red color. These results suggested that HmoCYP76AD1, HmoDODAα1, and HmocDOPA5GT play key roles in betalain biosynthesis in Hylocereus. The results of the present study provide novel genes for molecular breeding programs of pitaya.


Introduction
Betalains are a group of hydrophilic nitrogen-containing pigments with multiple functions [1][2][3]. Betalains are usually used as functional additives in pharmaceuticals, cosmetic, and food products due to their antimicrobial effects and natural colorant [4]. In addition, betalains are involved in plant protection against abiotic and biotic stresses [5][6][7][8][9][10]. Therefore, it is of great significance to enhance the betalain contents in plants through genetic transformation.

Gene Identification and Phylogenetic Analyses
Genes associated with betalain biosynthesis were downloaded from NCBI database according to the publications. The GenBank Accession No. are listed in Supplementary  Table S1. Gene sequences were identified by BLAST search of H. undatus genomes [43] using TBtools software [44]. Annotated gene sequences were manually checked and adjusted to remove the repeated sequences and retain the genes with identity equal to or greater than 40%. Each potential gene was further manually examined to ensure the complete conserved domain using the Conserved Domain Database (CDD) (http://www.ncbi.nlm.nih.gov/cdd (accessed on 1 September 2019)). Alignment of multiple sequences was performed with CLUSTAL (Bootstrap value = 1000). The phylogenetic tree was constructed using a neighbors-joining (NJ) method with MEGA7.0 software. The Newick tree was exported from MEGA7.0 and improved with iTOL (https://itol.embl.de/gallery.cgi (accessed on 20 July 2021)).

Chromosomal Distribution Analysis
The location information of genes related to betalain biosynthesis was obtained from the pitaya genome database [43]. The gene location map was constructed using TBtools software [44].

Full-Length cDNA Amplification and Plasmid Construction
Results from qRT-PCR indicated that all four candidate genes were preferentially expressed in pigmented tissues. Differences in nucleotide sequence and amino acids were detected between H. monacanthus and H. undatus. Therefore, the sequences from H. monacanthus were used in the subsequent experiments. Total RNA was extracted using EASYspin Plus Complex Plant RNA Kit (RN53) (Aidlab Biotechnology, Beijing, China) according to the manufacturer's protocol. The cDNAs were synthesized using a PrimeScript TM RT reagent Kit with gDNA Eraser (TaKaRa, Shiga, Japan). The full-length coding sequences of HmoCYP76AD1 (MZ541999), HmoDODAα1 (MZ542000), HmoDODAα2 (MZ542001), and HmocDOPA-5GT (MZ542002) were cloned using Phanta ® Max Super-Fidelity DNA Polymerase (Vazyme, Nanjing, China). Vectors were constructed using ClonExpress ® II One Step Cloning Kit (Vazyme, Nanjing, China) according to the manufacturer's manual. Gene coding sequences were constructed into pEAQ-GFP and pC18 vectors for transient expression. pPZP6k90 vector harboring gene coding sequences was used for genetic transformation. All gene fragments were recombined into a pTRV2 vector for VIGS assay. Single-gene expression vectors (YES2-Ura, pYES3-Trp) and/or double-gene expression vector (pESC-Ura) were used for recombinant expression in yeast. The primers used for gene cloning and vector construction are listed in Supplementary Table S2. The information of all vectors is described in Supplementary Text S3.

Subcellular Localization Assay
The pEAQ-GFP fusion vectors (pEAQ-HmoCYP76AD1, pEAQ-HmoDODAα1, pEAQ-HmoDODAα2, and pEAQ-HmocDOPA-5GT), positive control GFP vector, and the nuclear and plasma membrane RFP marker [49] were transformed into Agrobacterium tumefaciens strain GV3101 (Coolaber, Beijing, China) using the freeze-thaw method. The bacterial suspensions harboring the RFP marker and pEAQ-GFP fusion vectors were mixed with 1:1 ratio and co-infiltrated into the abaxial side of 4-to 6-week-old N. benthamiana plants according to the method of Cheng et al. (2017) [40]. Two to three days after infiltration, the abaxial epidermis of the leaves was subjected to confocal microscopy (Zeiss LSM 7 DUO, Oberkochen, Germany). The fluorescence signals were detected using an emission bandwidth of 488/505-525 nm for GFP and 543/585-615 nm for RFP.

Genetic Transformation of A. thaliana
The constructed plasmids (pPZP6k90-HmoDODAα1 and pPZP6k90-HmoDODAα2) were separately introduced into the A. tumefaciens strain GV3101 (Coolaber, Beijing, China) and transferred to A. thaliana using the floral dipping transformation method [50]. T 0 seeds were harvested, and transgenic lines were selected on MS medium supplemented with 100 mg/L kanamycin and confirmed by PCR analyses. T 3 transgenic homozygous seedlings were used for phenotypic characterization and gene expression analyses.
Ten-day-old A. thaliana seedlings were transferred onto plates containing absorbent paper with 10 mM L-DOPA solution. Sterile water was substituted for the L-DOPA as a negative control. Inflorescences from transformed A. thaliana seedlings were collected and immersed in 10 mM L-DOPA solution or water for 24 h. Betalain pigment production in A. thaliana seedlings and inflorescences were photographed after 24 h treatment.

Transient Expression Assay in N. benthamiana
Transient expression vectors pC18 containing candidate genes (HmoCYP76AD1, HmoDODAα1, HmoDODAα2, HmoB5GT1 and HmocDOPA-5GT) were introduced into A. tumefaciens strain GV3101 (Coolaber, Beijing, China), respectively. Transient expression in N. benthamiana was performed as described previously [22,51]. A. tumefaciens strains were grown overnight in YEP media and resuspended in infiltration buffer (10 mM MES, 10 mM MgCl 2 and 100 µM acetosyringone) to a final OD 600 of 0.2-0.3. Agrobacteria harboring different plasmids were mixed in a 1:1 ratio before infiltration. Pigment deposition in the infected N. benthamiana leaves was analyzed within 2-3 d after co-infiltration. Betalains were extracted and measured as described by Hua et al. (2016a) [35]. 0.1 g fresh leaves were ground and mixed with 1 mL 80% (v/v) aqueous methanol solution. After sonicating for 10 min, samples were centrifuged and kept at room temperature for 20 min in darkness. Then the supernatants were measured through spectrophotometry (Infinite M200, Shanghai, China) at 478 nm for betaxanthins and 538 nm for betacyanins. Three biological replicates were performed for each experiment.

Recombinant Expression in Yeast
Coding sequences of candidate genes related to betalain biosynthesis were recombined into pYES2-Ura (for HmoCYP76AD1 and HmocDOPA-5GT), pYES3-Trp (for HmoDODAα1 and HmoDODAα2) and pESC-Ura (for tandem HmoCYP76AD1 and HmocDOPA-5GT) vectors, respectively. The constructed plasmids were transformed into Saccharomyces cerevisiae strain INVSc1 (Coolaber, Beijing, China) using the polyethylene glycol/lithium acetate (PEG/LiAc) method (Coolaber, Beijing, China). A single gene or different combina-tions of CYP76AD1-DODA and CYP76AD1-DODA-cDOPA5GT were transferred into yeast competent cells. Yeast strains were cultured overnight in SD medium lacking Ura and Trp (SD/SD-Ura -Trp). Single colonies were picked and transferred into liquid SC/SC-Ura -Trp with 2 mM ascorbate and 10 mM L-DOPA, and grown overnight at 200 rpm in a shaking water bath at 30 • C. An aliquot of 1 mL supernatant was transferred to a rotary evaporator to remove water, re-suspended in 100 µL of ddH 2 O, and then detected by spectral analysis (Infinite M200, Shanghai, China).

Statistical Analysis
The experiments were carried out at least three individual biological replicates. Data in figures are the mean ± standard errors (S.E.). Statistical analysis of one-way ANOVA was used to determine the significant differences using SPSS 17.0 software at p < 0.05 and p < 0.01.

Identification and Classification of Candidate Genes Involved in Betalain Biosynthesis
Potential encoding genes involved in pitaya betalain biosynthesis were comprehensively screened by alignment using TBtools (Table S1). These candidate genes were further manually screened according to the completeness of conserved domains. A total of 53 candidate genes related to betalain biosynthesis were identified in the pitaya genome (Table S3), including 15 CYP76AD, 3 DODA and 35 GTs genes.
The genomic full-length sequences of GTs ranged from 1392 bp (HU03G02206.1 and HU07G01550.1) to 21,223 bp (HU04G00198.1). The proteins were mainly distributed between 400 bp and 500 bp in length. The average molecular mass of GTs was ~52.70 kDa except for HU04G00198.1 (96.81 kDa). The encoded proteins were weak acidic protein as supported by their pI. Only five GT proteins had hydrophilicity greater than 0 indicating that most of these GT proteins were hydrophilic (Table S3). The 35 GTs genes had an average of 1.3 exons. Among these GTs proteins, 66% proteins were unstable and 77% contained one exon ( Figure S1).

Expression Patterns of Candidate Genes Involved in Betalain Biosynthesis
The expression patterns of 36 candidate genes (nine CYP76ADs, three DODAs and 24 GTs) involved in pitaya betalain biosynthesis were analyzed during pitaya fruit development by qRT-PCR ( Figure S2). cDNA and amino acid sequences of H. monacanthus and H. undatus were provided in the Supplementary Text S1. Amino acid sequence similarities between H. monacanthus and H. undatus were shown in Supplementary Table S4. As shown in Figure 1A, the pulp and peel color of 'Guanhuahong' pitaya began to turn red at S2 stage and continuously accumulated towards fully red, while the pulp color of 'Guanhuabai' pitaya did not change during fruit maturation. The expression levels of

Gene Characteristics and Structures
The lengths of the genomic and protein sequences, isoelectric point (pI), molecular weight (MW), grand average of hydropathicity (GRAVY), number of exons and instability index of the 53 candidate genes were comprehensively analyzed to obtain more information about the nature of the proteins (Table S3 and Figure S1). The genomic sequence lengths of CYP76ADs ranged from 1796 bp (HU09G01023.1) to 71,367 bp (HU11G01017.1). The protein lengths ranged from 438 amino acids (HU07G00784.1) to 1835 amino acids (HU11G01017.1) with MW 49.88 to 208.48 kDa. The pI ranged from 5.98 (HU07G00784.1) to 9.39 (HU03G00481.1). Twelve proteins had a negative GRAVY score, indicating that they were hydrophilic in nature. Moreover, eight proteins were unstable (Table S3). Most of the full-length CYP76AD genes had two exons, except that HU01G02117.1, HU07G00500.  (Table S3). DODA genes had the same exon/intron structure: three exons and two introns ( Figure S1).
The genomic full-length sequences of GTs ranged from 1392 bp (HU03G02206.1 and HU07G01550.1) to 21,223 bp (HU04G00198.1). The proteins were mainly distributed between 400 bp and 500 bp in length. The average molecular mass of GTs was~52.70 kDa except for HU04G00198.1 (96.81 kDa). The encoded proteins were weak acidic protein as supported by their pI. Only five GT proteins had hydrophilicity greater than 0 indicating that most of these GT proteins were hydrophilic (Table S3). The 35 GTs genes had an average of 1.3 exons. Among these GTs proteins, 66% proteins were unstable and 77% contained one exon ( Figure S1).

Expression Patterns of Candidate Genes Involved in Betalain Biosynthesis
The expression patterns of 36 candidate genes (nine CYP76ADs, three DODAs and 24 GTs) involved in pitaya betalain biosynthesis were analyzed during pitaya fruit development by qRT-PCR ( Figure S2). cDNA and amino acid sequences of H. monacanthus and H. undatus were provided in the Supplementary Text S1. Amino acid sequence similarities between H. monacanthus and H. undatus were shown in Supplementary Table S4. As shown in Figure 1A, the pulp and peel color of 'Guanhuahong' pitaya began to turn red at S2 stage and continuously accumulated towards fully red, while the pulp color of 'Guanhuabai' pitaya did not change during fruit maturation. The expression levels of CYP76AD1 (HU03G00480.1, belonging to CYP76AD1-α) in pigmented tissues were higher than those of the white ones throughout fruit developmental stages ( Figure 1B). Compared with 'Guanhuabai' pitaya, the expression of CYP76AD1 in 'Guanhuahong' pitaya displayed a sharp upward trend with pigmentation ( Figure 1B). However, no close correlation was observed between the changes in genes belonging to the CYP76AD1-β or CYP76AD1-γ clade and the alterations of pigmentation. L-DOPA 4,5-dioxygenase in betalain-producing plants is encoded by DODA gene. The expression patterns of two DODAα genes (HU03G01342.1 and HU08G01572.1) were correlated with pigment accumulation: increasing during red pigmentation, while constantly low in white pulps. Expression levels of HU03G01342.1 were higher than HU08G01572.1 ( Figure 1B). According to the reported betalain-related genes, GTs genes were categorized into B5GT, B6GT, cDOPA5GT, UGT79B6 and UGT7B30 ( Figure 2C). The expression levels of cDOPA5GT (HU07G00240.1) increased gradually with betalain accumulation ( Figure 1B). The expression patterns of CYP76AD1 (HU03G00480.1), DODAα (HU03G01342.1 and HU08G01572.1) and cDOPA5GT (HU07G00240.1) were consistent with betalain accumulation during pitaya fruit development ( Figure 1B), suggesting that they are possibly involved in betalain biosynthesis in pitaya.

Subcellular Localizations
Subcellular localizations of HmoCYP76AD1, HmoDODAα1, HmoDODAα2 and Hmoc-DOPA5GT were validated using the GFP tag. As shown in Figure 4, the HmoCYP76AD1-GFP, HmoDODAα1-GFP, HmoDODAα2-GFP and HmocDOPA5GT-GFP were localized to both the cytoplasm and nucleus in terms of stronger green fluorescent signals. These results are in accordance with previously reported studies [52]. The dual localization signals were coincided with those of the PM marker RFP and the nuclear signal matched with that of the nucleolar marker RFP when the fusion protein was co-expressed with the markers.

VIGS Analyses of Candidate Genes
Gene silencing assay was performed to further elucidate the functions of HmoCYP76AD1, HmoDODAα1, HmoDODAα2 and HmocDOPA5GT genes. Silencing HmoB5GT1 resulted in the loss of betalain accumulation in red scales of pitaya [39]. Therefore, pTRV2 and pTRV2-HmoB5GT1 were separately infiltrated pitaya scales and served as the negative and positive controls. Silencing HmoCYP76AD1, HmoDODAα1, HmoDODAα2 or HmocDOPA5GT resulted in the loss of betalain accumulation, which was evidenced by green scales ( Figure 5A). Significant differences in contents of betacyanins and betaxanthins were detected in the gene-silenced tissues ( Figure 5B). Results from qRT-PCR confirmed that HmoCYP76AD1, HmoDODAα1, HmoDODAα2 and HmocDOPA5GT were silenced ( Figure 5C).
HmocDOPA5GT were validated using the GFP tag. As shown in Figure 4, the HmoCYP76AD1-GFP, HmoDODAα1-GFP, HmoDODAα2-GFP and HmocDOPA5GT-GFP were localized to both the cytoplasm and nucleus in terms of stronger green fluorescent signals. These results are in accordance with previously reported studies [52]. The dual localization signals were coincided with those of the PM marker RFP and the nuclear signal matched with that of the nucleolar marker RFP when the fusion protein was coexpressed with the markers.

VIGS Analyses of Candidate Genes
Gene silencing assay was performed to further elucidate the functions of HmoCYP76AD1, HmoDODAα1, HmoDODAα2 and HmocDOPA5GT genes. Silencing HmoB5GT1 resulted in the loss of betalain accumulation in red scales of pitaya [39]. Therefore, pTRV2 and pTRV2-HmoB5GT1 were separately infiltrated pitaya scales and served as the negative and positive controls. Silencing HmoCYP76AD1, HmoDODAα1, Hmo-DODAα2 or HmocDOPA5GT resulted in the loss of betalain accumulation, which was evidenced by green scales ( Figure 5A). Significant differences in contents of betacyanins and betaxanthins were detected in the gene-silenced tissues ( Figure 5B). Results from qRT-PCR confirmed that HmoCYP76AD1, HmoDODAα1, HmoDODAα2 and HmocDOPA5GT were silenced ( Figure 5C).

Overexpression of HmoDODAα1 and HmoDODAα2 in Arabidopsis and Yeast
To further confirm the role of HmoDODAα1 and HmoDODAα2, we generated transgenic A. thalianas plants constitutively expressing HmoDODAα1 and HmoDODAα2 driven by the 35S promoter ( Figure 6). No pigmentation was observed in the wild-type and transgenic lines harboring HmoDODAα1/2 without L-DOPA ( Figure 6A(a,c,e)). Yellow pigmentation was produced in the 35S::HmoDODAα1 transgenic T 3 generation seedlings and inflorescences with the addition of L-DOPA ( Figure 6A(d,h)). No yellow pigment was detected in the 35S::moDODAα2 transgenic lines after feeding L-DOPA ( Figure 6A(f,i)). The expression levels of HmoDODAα1 and HmoDODAα2 in the transgenic lines were significantly higher than those of wild types ( Figure 6B). Similar results were obtained in yeast (Figure 7). The expression of HmoDODAα1 in yeast supplemented with L-DOPA resulted in yellow cultures ( Figure 7A(b)). The yellow pigments had a peak absorption at about 468 nm wavelength ( Figures 6C and 7B) which was the typical characteristic of betaxanthins [53]. Compared to HmoDODAα1, no visible pigment production in yeast was observed by expressing of HmoDODAα2 with the presence of L-DOPA ( Figure 7A(d)). detected in the 35S::moDODAα2 transgenic lines after feeding L-DOPA ( Figure 6A(f,i)). The expression levels of HmoDODAα1 and HmoDODAα2 in the transgenic lines were significantly higher than those of wild types ( Figure 6B). Similar results were obtained in yeast (Figure 7). The expression of HmoDODAα1 in yeast supplemented with L-DOPA resulted in yellow cultures ( Figure 7A(b)). The yellow pigments had a peak absorption at about 468 nm wavelength ( Figures 6C and 7B) which was the typical characteristic of betaxanthins [53]. Compared to HmoDODAα1, no visible pigment production in yeast was observed by expressing of HmoDODAα2 with the presence of L-DOPA ( Figure 7A(d)).  3.8. Co-Expression of HmoCYP76AD1, HmoDODAα1/α2 and HmoB5GT1/HmocDOPA5GT in N. benthamiana HmoCYP76AD1, HmoDODAα1/ α2, and HmoB5GT1/HmocDOPA5GT were transiently overexpressed in N. benthamiana leaves. No visible pigmentation was detected in the infiltrated area of N. benthamiana leaves when single HmoCYP76AD1, HmoDODAα1, HmoDODAα2, HmoB5GT1 or HmocDOPA5GT was overexpressed ( Figure 8A(a-f)). Coinfiltration of A. tumefaciens containing HmoCYP76AD1 and HmoDODAα1 resulted in light red pigmentation ( Figure 8A(g)). Light red pigment was also observed when HmoCYP76AD1, HmoDODAα1 and HmoB5GT were co-expressed in N. benthamiana ( Figure 8A(j)). Dark red pigmentation was only observed in the infiltrated areas after coexpressing HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT ( Figure 8A(i)). However, no pigmentation was detected in the infiltrated areas when HmoCYP76AD1-DODAα2-cDOPA5GT or HmoCYP76AD1-DODAα2-B5GT1 were co-expressed ( Figure 8A(k,l)). Moreover, there was no pigment accumulation in N. benthamiana leaves when HmoDODAα1 or HmoDODAα2 was co-expressed with HmocDOPA5GT or HmoB5GT1 without Hmo-CYP76AD1 ( Figure S3). The red pigments had a peak absorption at 538 nm wavelength ( Figure 8B) which was the typical characteristic of betacyanins. Co-expression of Hmo-CYP76AD1-DODAα1-cDOPA5GT resulted in the highest abundance of betacyanins in the infiltrated areas compared to very low levels of betacyanin after co-expressions of the other combinations such as HmoCYP76AD1-DODAα1 and HmoCYP76AD1-DODAα2-B5GT1 ( Figure 8C). In addition, the expression levels of these genes significantly increased in the infiltrated areas ( Figure S4).

Recombinant Expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT in Yeast
Recombinant expression assay was performed to further confirm the function of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT genes. No pigment was detected when HmoCYP76AD1, HmoDODAα1 or HmocDOPA5GT was separately expressed in yeast ( Figure 9A(a-c)). Yellow pigments were detected in the yeast cultures when HmoCYP76AD1 was co-expressed with HmoDODAα1 ( Figure 9A(d)). Red pigments were observed in the yeast cultures containing pYES2-HmoDODAα1 and pESC-HmoCYP76AD1-HmocDOPA5GT vectors ( Figure 9A(e)). The peak absorption of red pigments was detected at 538 nm wavelength ( Figure 9B) suggesting that betacyanins are produced with co-expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT in yeast.
Recombinant expression assay was performed to further confirm the function of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT genes. No pigment was detected when HmoCYP76AD1, HmoDODAα1 or HmocDOPA5GT was separately expressed in yeast ( Figure 9A(a-c)). Yellow pigments were detected in the yeast cultures when HmoCYP76AD1 was co-expressed with HmoDODAα1 ( Figure 9A(d)). Red pigments were observed in the yeast cultures containing pYES2-HmoDODAα1 and pESC-HmoCYP76AD1-HmocDOPA5GT vectors ( Figure 9A(e)). The peak absorption of red pigments was detected at 538 nm wavelength ( Figure 9B) suggesting that betacyanins are produced with co-expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT in yeast.
The expression pattern of cDOPA5GT was in accordance with the accumulation of betalains ( Figure 1B) suggesting that cDOPA5GT maybe a potential gene encoding GT in pitaya.

HmoCYP76AD1 Catalyzes Tyrosine to L-DOPA and to Cyclo-DOPA
The functions of CYP76ADα and CYP76ADβ clade homologues have been previously reported in betalain-producing plant species [13,[15][16][17][18][19]. In B. vulgaris, silencing CYP76AD1 alone blocked the production of betacyanins but not betaxanthins, suggesting that CYP76AD1 and CYP76AD6 catalyze the formation of L-DOPA from tyrosine. However, only CYP76AD1 can catalyze the conversion of L-DOPA to cyclo-DOPA [18]. In this study, HmoCYP76AD1 belonging to the CYP76ADα clade was identified in pitaya. Silencing HmoCYP76AD1 resulted in the reduced accumulation of betacyanins and betaxanthins ( Figure 5), suggesting that HmoCYP76AD1 was the enzyme that contributes to L-DOPA formation and the subsequent conversion to cyclo-DOPA in pitaya. Betaxanthins were detected in yeast cultures after coexpression of HmoCYP76AD1 and HmoDODAα1 genes without feeding L-DOPA (Figure 9), suggesting that HmoCYP76AD1 has tyrosine hydroxylase activity for converting tyrosine to L-DOPA. Similar findings were also reported in other studies [17][18][19]. Red pigmentation was detected after expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT in N. benthamiana and yeast culture (Figures 8 and 9), indicating that HmoCYP76AD1 catalyzes two succeeding steps in the pitaya betalain biosynthesis pathway to produce cyclo-DOPA. Similar results were also found in other betalain-producing plant species such as B. vulgaris, C. quinoa, A. tricolor and M. jalapa [13,[16][17][18]22,23].

HmoDODAα1 Exhibits a High Degree of L-DOPA 4,5-Dioxygenase Activity
L-DOPA 4,5-dioxgenase responsible for synthesis of betalamic acid is considered to be an important betalain-producing enzyme. Betalain-pigmented species have multiple paralogues of DODAα genes, but only one of them in each species encodes the protein with a high degree of L-DOPA 4,5-dioxygenase activity [22,24]. Loss of betalains was observed when the gene encoding DODA was silenced in B. vulgaris [13]. In this study, two candidate genes (HmoDODAα1 and HmoDODAα2) were obtained from H. monacanthus. Silencing HmoDODAα1 and HmoDODAα2 in H. monacanthus resulted in betalain reduction ( Figure 5A,C), suggesting that HmoDODAα1 and HmoDODAα2 encode proteins with L-DOPA 4,5-dioxygenase activity. Overexpression of HmoDODAα1 and HmoDODAα2 in A. thaliana, N. benthamiana and yeast revealed that only HmoDODAα1 could synthesize betalamic acid (Figures 6-9). Similar findings were also reported in B. vulgaris, C. quinoa and A. hypochondriacus [16,22]. Silencing DODAα2 caused slight decolouration, but DODAα2 is not active in betalain accumulation when DODAα2 was overexpressed in N. benthamiana and yeast. The reason is that the DODAα2 may not have sufficient levels to influence pigment accumulation [22]. Further experiments such as co-silencing DODAα1 and DODAα2 are necessary to elucidate its role in betalain biosynthesis. These results suggested that HmoDODAα1 encodes a protein with a high degree of L-DOPA 4,5-dioxygenase activity involved in pitaya betalain biosynthesis.

Reconstruction of Heterogenous Production of Betalain Biosynthesis Pathway
Betacyanins are formed through glycosylation by GT [11,12]. In D. bellidiformis, glycosylation on betanidin is directly catalyzed by B5GT and B6GT [12,28,29]. In M. jalapa, A. hypochondriacus and C. quinoa, glycosylation by cDOPA 5-O-GT (cDOPA-5GT) is involved in the betalain biosynthesis [16,25,30]. In our previous study, we found that HmoB5GT1 was involved in the betalain biosynthesis in pitaya [39]. In the present study, low production of betacyanins was observed when HmoCYP76AD1, HmoDODAα1 and HmoB5GT1 were co-expressed in N. benthamiana (Figure 8). In contrast, the co-expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT resulted in high levels of betalain pigments (Figures 8 and 9). These results suggested that HmocDOPA5GT plays an important role in glycosylation during betalain biosynthesis in H. monacanthus.

Conclusions
In summary, this study was the first attempt to identify and validate the key genes involved in pitaya betalain biosynthesis based on the whole genome sequence of H. undatus. Fifteen CYP76AD-encoding genes, three DODA-encoding genes and 35 GTs-encoding genes were obtained from the H. undatus genome. Four candidate genes i.e., HmoCYP76AD1, HmoDODAα1, HmoDODAα2 and HmocDOPA5GT showed upward expression trends, which was consistent with betalain accumulation during pitaya fruit maturation. Furthermore, HmoDODAα1 displayed a high degree of L-DOPA 4,5-dioxygenase activity to produce betalamic acid. Co-expression of HmoCYP76AD1, HmoDODAα1 and HmocDOPA5GT in N. benthamiana and yeast resulted in high abundances of betalain pigments. These results suggested that HmoCYP76AD1, HmoDODAα1, and HmocDOPA5GT play key roles in pitaya betalain biosynthesis. The results of the present study not only shed light on the evolution of betalain biosynthesis pathway in Caryophyllales, but also provide novel genes to genetically improve the fruit quality of pitaya.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/genes12121858/s1, Figure S1: Gene structures of 53 genes. Figure S2: Expression analyses of candidate genes in different fruit developmental stages of pitayas by RT-qPCR. Figure S3: Penotype of N. benthamiana leaves after infected. Figure S4: Gene expression of N. benthamiana leaves after infected. Table S1: Th e accession numbers of genes in this study. Table S2: All primers used in this study. Table S3: 53 candidate genes in pitaya. Table S4: Amino acid sequence similarities between H. monacanthus and H. undatus. Text S1: cDNA and amino acid sequences of genes for qPCR in pitaya. Text S2: cDNA sequences of genes for VIGS in red pitaya. Text S3: Vectors used in this study.