Peach (Prunus persica L. Batsch cv. Hujingmilu) fruit, flowers, and leaves were obtained from the Melting Peach Research Institute of Fenghua, Zhejiang Province, China. Peach fruit were harvest at four stages (S1, S2, S3, and S4) according to previous study (Botton et al., 2016; Wang et al., 2016), representing the first fast growth (34 days after bloom, DAB, fruit weight = 5.69 ± 0.37 g), endocarp lignification (stone hardening, 71 DAB, 45.29 ± 0.63 g), the second fast growth (94 DAB, 113.93 ± 2.71 g), and mature stage (ready for harvest, 108 DAB, 207.48 ± 2.08 g), respectively. Peach fruit growth curve and photos of four stages were shown in Supplementary Figure S1. In our experiment, fruit with firmness = 26.70 ± 5.16 N, total soluble solids = 11.18 ± 0.68 ^oBrix, lightness = 65.80 ± 0.44, hue angle = 100.99 ± 0.30 and chroma = 32.31 ± 0.22 were harvested at mature stage. After harvested at the S4 stage, fruit were allowed to ripen up to 6 days (20°C, 90–96% relative humidity), and were sampled at 3 days (S4 + 3d) and 6 days (S4 + 6d). At each sampling time, peach flesh tissues were collected. Vegetative tissue samples were taken from full expanded mature leaves, and flowers were collected at full blossom. For UV-B treatment, mature peach fruit were randomly divided into two groups, and were stored in climatic chambers without any natural light. One group fruit were exposed to irradiation of UV-B (280–315 nm) for 6 and 48 h at 20°C and relative humidity 90–96%. UV-B lamp tubes (Luzchem Research, Inc., Gloucester, ON, Canada) provided 1.50 w/m² at fruit height (approximately 50 cm under the lamps). Control fruit were covered with aluminum foil to avoid exposure to light, and were placed next to the samples undergoing UV-B irradiation according to Suzuki et al. (2015). After treatment, slices of peel tissue (∼1 mm thick) were carefully separately. At each time point, three biological replicates with five fruit each were harvested. Peach tissues were immediately frozen in liquid nitrogen and then stored at -80°C until further analysis.

A feature sequence of UGTs, the 44-amino acid conserved sequence of the PSPG box, was used as a query to search against peach genome database at the Phytozome (v11.0²) with BLASTP program to identify peach UGTs. Information of peach UGTs, including ORFs, description, and chromosome distribution were obtained from the same database. The distribution of UGT genes on the chromosomes was visualized with the MapChart (v2.3). The identification of signal peptide in UGT sequences was performed by SignalP 4.1 Server³.

The predicted amino acid sequences of UGT genes were aligned using the neighbor-joining (NJ) method in ClustalX v2.0 program. Phylogenetic tree was constructed with FigTree v1.4.2 program.

The peach UGT intron map was constructed by determining the intron length, splice sites, phases, and positions in the genome. The exon–intron structure and intron phases were acquired by the online Gene Structure Display Server 2.0⁴. The intron can be inserted anywhere in the transcript, they are located either between codons or within codons which termed as intron phases, and the phases as well as location usually stay unchanged for a long time. Specifically, intron phases were determined as follows: introns positioned between two codons were defined as phase 0, introns positioned between first and second base of codon were defined as phase 1, and introns positioned between second and third base were defined as phase 2 (Barvkar et al., 2012).

RNA was extracted according to Zhang et al. (2006), and quality was monitored by gel electrophoresis and A260/A280. Libraries for high-throughput Illumina strand-specific RNA-seq were prepared as described previously (Zhang et al., 2016). Three biological replicates for various organs, fruit development and ripening stages and treatment were prepared. Transcript profiles for selected peach UGTs were obtained and expressed as heatmaps.

Glycosylated bound volatiles were extracted according to Yauk et al. (2014) with modifications. Twenty grams of frozen samples were ground in liquid nitrogen, and were homogenized with water. After centrifuging for 20 min at 13,000 g, the supernatant was used as crude extract. Isolation of glycosidic precursors was conducted by using SPE LC-18 resins (CNW, Duesseldorf, Germany), which was previously preconditioned with methanol and distilled water. Eliminating free volatile compounds was performed by washing with dichloromethane, and the bound fraction was eluted with methanol. The bound volatile compounds were enzymatic hydrolyzed by adding AR2000 (Rapidase, Séclin, France) at 37°C for 48 h. The released volatiles were extracted with solid phase micro extraction, followed by identification using gas chromatography–mass spectrometry (GC–MS).

Gas chromatograph (Agilent 7890N) coupled with Agilent 5973C mass spectrophotometer (Agilent, Palo Alto, CA, USA) were applied for identification of volatile compounds according to methods described by Zhang et al. (2010). Helium was used as a carrier gas at a flow rate of 1.0 mL min^-1. The column effluent was ionized by electron ionization (EI) at energy of 70 eV with the transfer temperature of 250°C and the source temperature of 230°C. Mass scanning was done from 35 to 350 m/z with a scan time of 7 scans per second. Volatile compounds were identified by comparing with their EI mass spectra to NIST/EPA/NIH Mass Spectral Library (NIST-08 and Flavor) and retention time of authentic standards (Sigma-Aldrich, St. Louis, MO, USA) when available. Semi-quantitative determination of compounds was performed using the peak area of the internal standard as a reference based on total ion chromatogram (TIC) and calculated based on standard curves of authentic compounds.

Construction of expression plasmids of peach UGT genes were carried out by cloning the full-length ORFs with the N-terminal His-tag into the pET6xHN expression vector (Clontech, Palo, CA, USA). Primers for cloning were listed in Supplementary Table S1. The identity of the cloned gene was confirmed by sequencing of the complete insert. Recombinant proteins were heterologous expressed in E. coli BL21 (DE3) pLysS (Promega, Madison, WI, USA). The transformed cells were precultured at 37°C overnight in 20 mL Luria-Bertani (LB) medium containing 100 μg mL^-1 ampicillin, and then inoculated into 1 L LB medium containing the same antibiotics. The culture was grown at 37°C at 150 rpm to an OD (600 nm) = 0.5 to 0.6, and the protein expression was induced with 1 mM isopropylthio-β-galactoside (IPTG). The culture was incubated at 16°C at 150 rpm for 20 h. The cells were harvested by centrifugation (4,000 g, 4°C, and 15 min) and resuspended in 1× PBS (1.37 M NaCl, 26.8 mM KCl, 20.3 mM Na₂HPO4, and 17.6 mM KH₂PO4, pH 7.2–7.4). After storing at -80°C overnight, the cells were disrupted by sonication. The supernatant was obtained by centrifugation (10,000 g, 4°C, and 30 min), and then purified using TALON Spin column (Clontech, Palo, CA, USA) following the manufacturer’s instructions.

Activity assays of recombinant peach UGTs were carried out in reactions using purified protein at final concentration at 0.4 μg μl^-1. Reactions were performed in a buffer containing 100 mM Tris-HCl buffer (pH 7.5, and 2.0 mM DTT), 1.0 mM UDP-glucose (Sigma-Aldrich, St. Louis, MO, USA) and 1.0 mM substrate in a volume of 200 μl. Enzyme assays were incubated at 30°C for 16 h and terminated by addition of 1 μl 24% (v/v) TCA. The reaction mixture was extracted with ethyl acetate and then were evaporated to dryness. The glucosides were dissolved in methanol and analyzed by Agilent 1290 Infinity LC System (binary pump G4220A, diode array UV/VIS detector G4212A; Agilent Technologies, USA) coupled with a SunFire C18 analytical column (5 μm, 4.6 mm × 250 mm; Waters, USA). The column was operated at a temperature of 25°C. The chromatograms were obtained and detected between 200 and 400 nm. High performance liquid chromatography (HPLC) were performed at a flow rate of 1 mL min^-1 with 100% water as solvent A and 100% acetonitrile as solvent B. The injection volume of samples was 20 μl. The column was firstly equilibrated with 10% solvent B, and eluted with a linear gradient program from 30 to 70% solvent B for 15 min, then washed with 100% solvent A for 15 min. The mass spectrometry analyses were performed by an Agilent6460 triple quadrupole mass spectrometer equipped with an ESI source (Agilent Technologies, USA) that operated in negative ionization mode. The scan range was 100–1000 m/z, and the nebulizer pressure was set as 45 psi. Identification of glucoside was based on the HPLC retention time and mass spectrometry spectral data. The glucoside of 2-phenylethanol was quantified by absorption at 210 nm, and the m/z value was 283 [M-H]^-.

MultiExperiment Viewer (version 4.6.0) was applied for heatmap of peach UGT genes transcript abundance and for gene clusters construction. Correlation analysis between UGT genes transcript levels and contents of glycosylated bound volatile compounds was performed by MetaboAnalyst 3⁵.

A BLASTP research of peach genome (Phytozome v11.0) was performed using UGT conserved PSPG box sequence as a query. A total of 168 peach UGTs having lengths of 150–616 amino acids were identified. A phylogenetic tree of peach UGT genes was constructed by aligning the full-length amino acid sequences of the peach UGTs with functionally characterized plant UGTs, including Arabidopsis, maize and UGTs from tomato, citrus, grapevine, apple, kiwifruit and strawberry (Figure 1). These peach UGTs were phylogenetically divided into 16 groups, including 14 groups (A–N) that were identified in Arabidopsis (Li et al., 2001). Two peach UGTs that were identified to be involved in metabolism of anthocyanin (Cheng et al., 2014) were clustered in group F, namely Prupe.1G091100 (PpUGT78A1) and Prupe.1G091000 (PpUGT78A2). Other UGTs responsible for glycosylation of anthocyanin were also located in group F, including strawberry FaGT1 (Griesser et al., 2008), kiwifruit F3GT1 (Montefiori et al., 2011), and grapevine VvGT5 and VvGT6 (Ono et al., 2010).

Distribution of plant UGTs in phylogenetic groups was summarized in Table 1. A total of 17 groups were detected in plants. Except for Zea mays (maize), no plant UGTs was detected in group Q. It is worth noting that group G had 34 UGTs from peach and 40 from apple, while 6–20 in other plants (Table 1). Moreover, group M contained 14 UGTs from peach and 15 in apple, compared with 1–5 in other species. These results indicated groups G and M may play important roles in the Rosaceae family such as apple and peach.

To provide an overview of location of peach UGT genes, genomic distribution of each UGT on the genome is shown in Figure 2. There are 164 UGTs distributed across all eight chromosomes of peach (Figure 2), while the left four UGTs located on scaffolds, including Prupe.I000900, Prupe.I001000, Prupe.I001100, and Prupe.I002600 (Supplementary Table S2). Different members of UGTs were observed for each chromosome. There were 39 UGTs on chromosome 01, followed by 36 members on chromosome 06, and 30 on chromosome 03 (Figure 2). Only three UGTs were observed on chromosome 04.

Considering that peach UGTs could be separated into 16 groups, distribution of these groups on the chromosome were investigated (Figure 2 and Supplementary Figure S2). Group G consisting of 34 peach UGTs had 15 members located on chromosome 1 and 14 on chromosome 03. For group D, 19 members were identified in peach (Table 1), with 7 and 12 UGTs located on chromosomes 07 and 08, respectively. On the contrary, UGTs of group E are randomly distributed across seven chromosomes (01–07) (Figure 2 and Supplementary Figure S2).

To investigate the evolutionary relationships within peach UGT gene family, the exon–intron organization was analyzed. Among the 168 peach UGTs, 72 had no introns, 82 contained one intron each in them (Table 2). For the remaining 14 UGTs, nine had two introns, four had three introns, and one had five introns. For UGT groups, the largest numbers of genes losing introns was observed for group E with 23 members, followed by 16 in group D and 13 in group M. A total of 31 (91%) UGTs in group G contained one intron, followed by 12 in group L (Table 2).

A total of 96 peach UGTs containing intron sequences was observed. After mapping the introns to the amino acid sequences alignment, at least 10 independent intron insertion events were observed. These insertion events are serially numbered as I-1 to I-10 according to their positions (Figure 3). The high conserved introns were observed for I-5 (intron 5), containing 66 (69%) peach UGTs belonging to group A, D–K, N, and P. Among these groups, all members of group K, I, and N contained the intron 5. For group G, 32 out of 33 UGTs had high conserved intron 5. The intron 6, was predominantly observed in group L.

Among the total 118 introns detected in peach UGTs sequences, 25, 84, and 9 were in phase 0, 1, and 2, respectively (Figure 3 and Supplementary Table S3). For the high conserved intron 5, only one was in phase 0, one was in phase 2, and phase 1 accounted for 97% of all introns (Figure 3). For intron 6, 16 out of 18 UGTs were in phase 0. This suggests that the high conserved introns were in the same intron phase.

RNA-seq was carried out to examine the expression pattern of 168 peach UGTs in flowers, leaves, and during fruit development and ripening. A total of 45 (27%) UGTs showed the highest levels of transcripts in leaf (Figure 4). For group D, 12 out of 19 UGTs were predominantly expressed in leaf. For peach flower, totally 54 (32%) UGTs had the highest expression. All members of groups F and N, and more than half members of group A, C, H, J, and P accumulated the most abundance of UGTs transcripts in flower (Figure 4). Tissue-specific expression of Prupe.1G091100 and Prupe.1G091000 in group F were consistent with previously study, where these genes were involved in biosynthesis of peach flower anthocyanin (Cheng et al., 2014). A total of 60 UGTs had the highest abundance in developing fruit and post-harvest ripening. These genes primarily expressed in fruit were from 13 groups of peach UGT except for group C, F, and N (Figure 4).

Among fruit-specific expression of these peach UGT genes, transcript levels of 19 (35%) members in peach was the highest at S1 (the first fast growth), followed by 11 at S2 (stone hardening), nine at S3 (the second fast growth), and 10 at S4 (mature stage). After fruit harvested at S4, expression of peach UGTs usually tended to decrease during shelf-life at ambient temperature (Figure 4). Considering that group G had the largest number of UGT genes in peach, expression patterns during fruit ripening was further analyzed. Five members of UGTs showed the highest transcript levels at S1 stage, including Prupe.1G520000, Prupe.6G189800, Prupe.6G189900, Prupe.1G520300, and Prupe.8G063500. Transcript levels of three UGTs peaked at S4 stage during fruit development, while one member further accumulated during post-harvest ripening at S6 stage (Figure 4).

Abiotic stresses such as UV-B irradiation can affect production of secondary metabolites (Kaling et al., 2015), some of which such as phenylpropanoid are primarily presented as glycosylated form due to UGTs activity (Roy et al., 2016). To test if UV-B treatment could alter expression of UGT genes, transcript profiles were investigated in peach fruit (Figure 5). After 6 h irradiation, transcripts of group F Prupe.1G091100 was induced. Similar pattern was also observed for Prupe.1G090500 and Prupe.2G324700 in group F, five members in groups E and M each, four members in groups H, K, and L each. Transcript levels of 65 (39%) UGT genes were significantly induced after 48 h UV-B treatment, including nine members in Group D, 14 in Group E, 13 in Group G, 5 in Group J, and 10 in Group L (Figure 5).

It has been reported that glycosylated bound volatile was a potential source for aroma formation, and formation of these non-volatile compounds were catalyzed by UGTs via transferring activated sugar molecules to volatile compounds (Schwab et al., 2015). A total of 26 glycosylated volatiles were identified in peach fruit during development and ripening (Figure 6). These chemicals composed of glycosylated aldehydes, esters, ketones and alcohols, including methyl salicylate, C6 alcohols, linalool and 2-phenylethanol. The detected glycosylated volatiles are clustered into three groups based on changes in content during fruit development and ripening. The first group consisted of 12 compounds, including 2-phenylethanol, which had the highest content at S1 stage (Figure 6). For the second group of glycosylated volatiles, content of benzyl alcohol and eugenol increased and peaked at S2 or S3 stage. Glycosylated linalool and terpeniol belong to the third group, which had the highest content at S5 stage. To mine UGT genes that are associated with glycosylated bound volatile formation, a correlation analysis between levels of transcripts and contents of volatiles was carried out (Supplementary Figure S3). A total of 128 (76%) members had positive correlation with the glycosylated volatiles detected in peach fruits.

2-phenylethanol was an important volatile compound derived from phenylalanine that contribute to fruit overall flavor quality (Tieman et al., 2006), and then was used to further analysis. There are eight peach UGTs showing significant positive correlation with 2-phenylethanol (Figure 7), therefore, their enzyme activity was analyzed. Recombinant proteins were obtained via heterologous overexpressing these eight UGTs in E. coli. As shown in Figure 8, Prupe.6G008000, Prupe.6G190100, and Prupe.6G189900 could transfer the UDP-glucose to 2-phenylethanol. Regarding the other five UGTs, no activity was observed for recombinant proteins toward 2-phenylethanol. These results suggested that these three UGTs Prupe.6G008000, Prupe.6G190100, and Prupe.6G189900 were likely to be associated with formation of 2-phenylethyl β-D-glucoside in peach fruit. Subcelluar localization of these three active UGTs was predicated using SignalP program, and no signal peptides were observed. These observation is consistent with the general assumption that plant UGTs are cytoplasmic enzymes (Li et al., 2001).