PpYUC11, a strong candidate gene for the stony hard phenotype in peach (Prunus persica L. Batsch), participates in IAA biosynthesis during fruit ripening

Highlight PpYUC11 is a strong candidate gene for the control of the stony hard phenotype in peach.


Introduction
Peach (Prunus persica L. Batsch) is a climacteric fruit in which the burst of ethylene production occurs at the ripening stage (Tonutti et al., 1991(Tonutti et al., , 1997. Peach fruits generally soften rapidly after harvest, causing difficulties in postharvest handling and quality maintenance during storage and transportation (Yoshioka et al., 2010). Improving the fruit texture quality of peaches is one of the most important goals of cultivar development and selection, superseded only by appearance and harvest season (Gallardo et al., 2012). Based on the characteristics of fruit firmness and textural changes during ripening, peach cultivars can be classified into three groups: melting (M), nonmelting (NM), and stony hard (SH) flesh types (Bailey and French, 1932;Yoshida, 1976;Haji et al., 2005). The M phenotype shows a prominent softening at advanced stages of ripening and develops a complete melting texture, whereas NM-textured peaches soften slowly when overripe and never melt. Peaches of the SH type, the focus of the present study, have very firm and crisp flesh at ripening (both on-and offtree), although they change colour normally and contain a high level of soluble solids (Haji et al., 2001(Haji et al., , 2004. Endopolygalacturonase (endo-PG), the enzyme responsible for cleaving polygalacturonic acid chains (pectins) from the cell wall, is thought to be responsible for the M/NM flesh texture (Lester et al., 1996;Callahan et al., 2004). This trait is controlled by the 'melting' locus, where the melting (M) is dominant over the non-melting (m) (Bailey and French, 1932). A high level of ethylene production (in the late ripening phase) characterizes the M and NM varieties. This phase, however, is absent in the SH phenotype which is thought to contribute to the inhibition of fruit softening in stony hard peaches because exogenous ethylene softens the flesh effectively; thus, the fruit softens more rapidly when a higher concentration of ethylene is applied (Haji et al., 2003;Hayama et al., 2003Hayama et al., , 2006. Treatment of SH fruits with 1-aminocyclopropane-1-carboxylic acid (ACC), the immediate precursor of ethylene, stimulated the fruits to synthesize ethylene and to soften (Haji et al., 2003). These results indicate that a low activity of ACC synthase is responsible for the inhibition of ethylene production and that ACC oxidase and the ethylene receptor function normally in SH peaches (Haji et al., 2003). Genetic analysis suggested that the Stony hard trait is controlled by a single recessive gene (hd) (Yoshida, 1976) and is inherited independently of the M/m flesh locus (Haji et al., 2005).
It has been assumed that transcription suppression of PpACS1 in the SH phenotype is responsible for its low ethylene production ; however, mRNA transcripts of PpACS1 have been detected in senescing flowers and wounded leaves as well as in immature fruits and coldtreated mature fruits of the SH phenotype Begheldo et al., 2008). These results indicate that stress conditions (wounding or low temperature) seem to be effective in overcoming PpACS1 inhibition that has been ascribed in SH peaches to a disruption of a transcriptional factor specifically activated at ripening . A recent study suggested that the low IAA concentrations in SH peaches may contribute to the suppression of PpACS1 because the application of exogenous 1-naphthalene acetic acid (NAA), a synthetic auxin, to stony hard peaches induced a high level of PpACS1 expression, the production of a large amount of ethylene, and softening (Tatsuki et al., 2013).
Multiple mechanisms regulating auxin-homeostasis have been recognized in plant organs, such as auxin metabolism (biosynthesis and conjugation) and carrier-dependent intercellular and intracellular auxin transport (Rosquete et al., 2012). In many cases, de novo IAA biosynthesis, primarily from tryptophan, is the most important source of free auxin in the plants. Genetic and biochemical studies have revealed that, in Arabidopsis and maize, auxin biosynthesis is controlled by a two-step pathway: tryptophan is first converted into indole-3-pyruvate (IPA) by the TAA family of aminotransferases, and, subsequently, IAA is produced from IPA by the YUCCA family of flavin monooxygenases (Mashiguchi et al., 2011). In addition to de novo synthesis, plants produce active IAA by hydrolysing IAA conjugates. Three major forms of auxin conjugates exist in higher plants, including IAA-amino acid conjugates, IAA-sugar conjugates, and IAA-peptide conjugates (Ludwig-Müller, 2011). Group II of the GRETCHEN HAGEN3 (GH3) family facilitates IAA to amino acids (Staswick et al., 2005) whereas the IAA-LEUCINE RESISTANT 1 (ILR1)-like family of amidohydrolases hydrolyse IAA-amino acid conjugates to free IAA (Bartel and Fink, 1995;Davies et al., 1999;LeClere et al., 2002). Our knowledge of the metabolism of IAA-sugar conjugates and IAA-peptide conjugates is far less detailed than our knowledge of the metabolism of IAA-amino acid conjugates. UDP glucosyltransferases, such as UGT84B1 in Arabidopsis (Jackson et al., 2001) conjugate IAA to glucose. The conversion back to IAA of the presumed storage form of auxin, IBA, is catalysed by the action of peroxisomal β-oxidation enzymes, the IBRs (INDOLE-3-BUTYRIC ACID RESPONSE) (Zolman et al., 2008). In addition, the spatio-temporal distribution of auxin depends on cell-to-cell auxin transport. A few studies have demonstrated that AUXIN RESISTANT 1 (AUX1)-mediated auxin influx and PIN-FORMED 1 (PIN1)-mediated auxin efflux played a key role in the regulation of organ-level auxin transport (Bennett et al., 1996;Gälweiler et al., 1998;Müller et al., 1998;Li et al. 2011;Friml et al., 2002a, b). The low IAA concentration at the lateripening stage of stony hard peaches would be determined by the result of one or the interaction of more mechanisms above (see Supplementary Fig. S1 at JXB online).
The above introduction indicates that auxins are involved in peach softening and that stony hard peach cultivars do not soften due to low IAA concentrations. To date, the molecular mechanism underlying this phenomenon has remained a mystery. Recently, digital gene expression profiling was conducted using deep-sequencing in melting/stony hard flesh peaches, and the results provide opportunities for investigating genes involved in IAA concentrations during peach ripening. It is reported that one gene in particular, PpYUC11, appears to be an excellent candidate gene for the stony hard phenotype in peaches, and it is proposed that the intronic TC microsatellite sequence variant controls the expression of this gene. Our study is not only helpful for understanding the molecular mechanism of IAA accumulation during peach ripening but also useful for the marker-assisted breeding of new peach cultivars with better preservability.
Ethylene production, flesh firmness, and soluble solids content (SSC) Ethylene production and flesh firmness were measured as described previously (Zeng et al., 2015). After the measurement flesh firmness, part of the mesocarp was squeezed and SSC was measured with an Atago digital refractometer (Atago, Tokyo, Japan).

Extraction and UPLC-MS/MS analysis of IAA in the peach mesocarp
Extraction and purification of IAA in mesocarp samples were performed according to a previously described method (Zourelidou et al., 2009, Tatsuki et al., 2013, with minor modifications. To determine the IAA content in each sample, the UPLC-MS/MS analyses were carried out as previously described by Fu et al. (2012).

RNA extraction and quantitative real-time PCR
To validate the differential expression identified by DGE-seq or to detect the expression of other genes, quantitative reverse-transcription PCR (qRT-PCR) was performed as described previously (Zeng et al., 2015). Eleven auxin-homeostasis-related genes with different expression patterns during the late ripening stage between the normal and stony hard flesh phenotypes were chosen for validation and the specific primers used to quantify gene expression levels are listed in Supplementary Table S1 at JXB online. The housekeeping genes have been described by Tatsuki et al. (2013). All gene expression analyses were performed with three independent biological replicates.

Phylogenetic analysis
The amino acid sequences of the YUCCA family of flavin monooxygenases from Prunus persica and Arabidopsis thaliana were used for phylogenetic analysis. Multiple sequence alignment was carried out using ClustalX version 1.83 (Thompson et al., 1997) and adjusted manually as necessary and the phylogenetic tree was generated by the Neighbor-Joining method using MEGA version 5 (Tamura et al., 2011). Bootstrap values were calculated from 1 000-replicate analyses.
Digital gene expression profiling (DGE) library preparation, Illumina RNA-sequencing, and data processing Total RNA was extracted using a Plant Total RNA Isolation Kit (Sangon, China) according to the manufacturer's instructions. RNA concentrations were measured using a NanoPhotometer spectrophotometer (IMPLEN, CA, USA), and the integrity was confirmed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). DGE libraries were generated using the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, USA) following the manufacturer's recommendations and index codes were added to attribute sequences to each sample. Each library's quality was then assessed on the Agilent Bioanalyzer 2100 system. The library preparations were sequenced on an Illumina HiSeq 2000 platform. For gene expression analysis, the number of expressed tags was calculated and then normalized to RPKM (Reads Per Kilobase of exon model per Million mapped reads) (Mortazavi et al., 2008).

Retrieval of auxin-homeostasis-related gene sequences
Auxin-homeostasis-related gene sequences of Arabidopsis thaliana were identified from the Arabidopsis Information Resource (TAIR, https://www.arabidopsis.org/, last accessed: 18 August 2015) and auxin-homeostasis-related gene sequences of peach were obtained through the retrieval of Prunus persica v1.0 assembly at the Genome Database for Rosaceae (Verde et al., 2013) (GDR, http://www. rosaceae.org/, last accessed: 18 August 2015) or by BLASTP search at the National Center for Biotechnology (http://www.ncbi.nlm.nih.gov/ blast/, last accessed: 18 August 2015). All sequences and their accession numbers are provided in Supplementary Table S3 at JXB online.

Detection of intronic TC microsatellite variations
Genomic DNA was extracted using the modified CTAB procedure (Cheng et al., 1997). The concentration of DNA was determined with a NanoDrop 1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA) and diluted to 25 ng μl -1 to carry out PCR amplification reactions. The DNA samples were then used as PCR templates for microsatellite region amplification. The intronic TC microsatellites were amplified using primers: 5′-CTATCTGGTATATAAGCTGAAACG-3′, 5′-ACCTTTTT AGTTATTTCACCACAG-3′. The PCR reactions were performed using KOD-Plus-DNA polymerase (Toyobo, Osaka, Japan) according to the manufacturer's instructions. The DNA amplification products were loaded on to 10% polyacrylamide gels. The gels were silver stained according to the protocol described by Bassam et al. (1991).

Fruit phenotype during ripening: ethylene production, flesh firmness, and soluble solids content
Prunus persica fruits of 'Goldhoney 3' and 'Yumyeong' cultivars were picked at the desired ripening stages: transition from stage S3 (second exponential growth phase) to stage S4 (climacteric). The two cultivars have different fruit flesh texture: 'Goldhoney 3' is a melting type that softens rapidly during ripening, and 'Yumyeong' is a stony hard type that remains firm during ripening. Fruits from the end of stage S3 to stage S4 were measured for ethylene production, flesh firmness, and soluble solids content (°Brix, an approximate measure of the mass ratio of dissolved solids, mostly sucrose, to water in fruit juices) (Fig. 1). Ethylene production was low in 'Goldhoney 3' fruit before 110 DAFB (stage S3) and it sharply increased upon reaching stage S4 whereas, in 'Yumyeong', ethylene production was sustained at a low level during the same period (Fig. 1A). Flesh firmness decreased substantially in 'Goldhoney 3' between 110 and 130 DAFB, but it decreased moderately in 'Yumyeong' during the same period (Fig. 1B). In 'Goldhoney 3' and 'Yumyeong', SSC was 11 °Brix at 105 DAFB; increased gradually to a peak value of 15 or 14 °Brix at 122 DAFB for the two cultivars, respectively; and then gradually decreased; the maximum SSC attained in 'Goldhoney 3' was higher than that of 'Yumyeong' (Fig. 1C). The similarity of these SSC time series is consistent with the fact that both cultivars have an identical maturity stage in Zhengzhou.
Depending on phenotypic parameters, four sampling points covering the climacteric stage at 110, 115, 122, and 127 DAFB (named S3, S4 I, S4 II, and S4 III, respectively, and marked with arrows in Fig. 1C) were selected for IAA concentration and digital gene expression (DGE) analysis.

IAA concentration and digital gene expression (DGE) analysis of auxin-homeostasis-related genes
Previous studies have shown that IAA levels significantly increased in a melting flesh cultivar at the climacteric stage, although the IAA concentration in the mesocarp tissue of a stony hard peach cultivar were low and did not increase (Tatsuki et al., 2013). Before the expression patterns of auxin-homeostasis-related genes in both cultivars during fruit ripening were analysed, IAA concentrations were measured with the same samples at stages S3, S4 I, S4 II, and S4 III in 'Goldhoney 3' and 'Yumyeong' (Fig. 1C). Whereas in 'Goldhoney 3' fruits, the IAA levels sharply increased from stage S3 to stage S4 III, the IAA levels did not increase, and the final IAA concentration was 1.049 ng g -1 fresh weight in 'Yumyeong' fruits ( Fig. 2D).
To obtain the expression profiles of auxin-homeostasisrelated genes for the melting peach 'Goldhoney 3' and the stony hard peach 'Yumyeong' during the late ripening stages, total RNAs were isolated from 'Goldhoney 3' and 'Yumyeong' mesocarps at stages S3, S4 I, S4 II, and S4 III (Fig. 1C). Differential gene expression profiling was performed using high-throughput Tag To estimate the gene expression levels, the read counts were transformed into RPKM (Mortazavi et al., 2008); RPKM >1 is defined as the threshold of significant gene expression. The full list of normalized expression of 28 702 peach genes is shown in Supplementary Table 2 at JXB online.
In total, 51 auxin-homeostasis-related genes were identified from the peach genome: 12 auxin biosynthesis genes (9 YUC genes and 3 TAA genes), 27 auxin conjugation genes (9 ILL genes, 8 GH3 genes, 3 UGT genes, and 7 IBR genes), and 12 auxin transport genes (8 PIN genes and 4 AUX1 genes). Based stony hard cultivar 'Yumyeong' fruit from the late stage of the second exponential growth phase (S3) to the climacteric stage (S4). As shown in (C), the S3, S4 I, S4 II, and S4 III time points, corresponding to 110, 115, 122, and 127 DAFB, respectively, were selected for further study. For ethylene production, the results are the mean ±SE of measurements for at least three individual experiments; for flesh firmness or SSC, the results are the mean ±SE of measurements for at least five fruits. Asterisks indicate statistically significant differences compared with 'Yumyeong' at a similar stage in development (days after full bloom) using Student's t test (*P <0.05, **P <0.01). ns indicates that there were no significant differences compared with 'Yumyeong'. (This figure is available in colour at JXB online.) on our DGE data, 34 genes out of 51 total auxin-homeostasis-related genes have significant gene expression (RPKM >1) in at least one sampling point ( Fig. 2A; see Supplementary  Table S3 at JXB online), constituting the subset of candidate genes involved in controlling IAA concentration in peach fruit during the ripening stage. To identify the gene or genes putatively involved in IAA accumulation during melting peach ripening, the expression profiles of these auxin-homeostasisrelated genes were compared with two peach ACS and two endopolygalacturonase (endo-PG) genes by DGE ( Fig. 2A). In 'Goldhoney 3', the expression level of ACC synthase gene PpACS1 (ppa004774m) increased towards harvest time along with other ripening-related genes [ppa006839m (endo-PG); ppa006857m (endo-PG)]. By contrast, the expression level of PpACS1 in the stony hard peach stayed low and did not increase during ripening ( Fig. 2A, B; see Supplementary  Table S3at JXB online). A total of 34 auxin-homeostasisrelated genes were significantly expressed and their expression profiles were hierarchically clustered with that of PpACS1 and endo-PG ( Fig. 2A). Interestingly, six auxin-homeostasisrelated genes were found to cluster closely with PpACS1: one PIN gene (ppa002944), one IBR gene (ppa10700m), one ILL gene (ppa005951m), two GH3 genes (ppa003134m and ppa003126m), and one YUC gene (ppa008176m) ( Fig. 2A). Among these, two genes-the YUC gene (ppa008176m) and one of the GH3 genes (ppa003126m)-showed an identical expression pattern to PpACS1 for both cultivars (Fig. 2C). Taking the gene functions into account, the YUC gene (ppa008176m) may regulate IAA concentration during the fruit ripening stage of the melting flesh peach 'Goldhoney 3'.
To confirm the gene expression changes observed by DGE analysis, quantitative PCR analysis was performed on a selection of differentially expressed auxin-homeostasis-related genes. Overall, the quantification of 11 auxin-homeostasisrelated genes by qRT-PCR exhibited close agreement with DGE-seq results (see Supplementary Fig. S2 at JXB online).

The effects of NAA on the expression of auxinhomeostasis-related genes in 'Yumyeong'
The hierarchical cluster analysis shows that a couple of auxinhomeostasis-related genes showing a similar expression pattern to PpACS1 in the DGE data for melting flesh peach ( Fig. 2A). Previous studies have shown that a lot of auxin metabolic genes are auxin-response genes (Peer et al., 2004;Terol et al., 2006). Although IAA cannot be accumulated normally in stony hard peach cultivars, the auxin-homeostasis-related genes in SH flesh peaches should have a similar expression pattern (with exogenous auxin treatment) as the normal peach fruits. The different responses to exogenous auxin treatment of the candidates most likely reflect whether the gene's expression is required for IAA accumulation or merely the result of a high IAA concentration. To understand the transcriptional changes in these candidate genes during the ripening stage of melting flesh peaches, the effect of exogenous auxin on the expression of auxin-homeostasis-related genes of stony hard peaches was examined. Mature fruit of 'Yumyeong' were treated with the synthetic auxin NAA. The expression levels of auxin-homeostasis-related genes that were found to cluster closely with PpACS1 in the DGE analysis were examined (Fig. 3). The expression levels of these genes did not increase in control fruits; expression levels of the ILL gene (ppa005951m), the PIN gene (ppa002944m), and the two GH3 genes (ppa003134m; ppa003126m) increased in NAAtreated 'Yumyeong' fruits after 2 d or 4 d of NAA treatment. By contrast, the expression of the YUC gene (ppa008176m) was not affected by exogenous NAA application to the fruits of 'Yumyeong' (Fig. 3).

Structural features and polymorphic loci analysis of ppa008176m (PpYUC11)
Based on the observation that the expression patterns of ppa008176m (named PpYUC11, Fig. 7) and IAA concentrations of the two cultivars (Fig. 2) showed high levels of consistency, a detailed analysis was undertaken of the PpYUC11 locus variations using the hypothesis that PpYUC11 structural variation may cause differential expression of PpYUC11 between the two cultivars. PpYUC11, contains four exons, located in pseudomolecule 6 and is near the centromeric regions in the peach genome sequence (Fig. 4A, B). The PpYUC11 gene was re-sequenced from a set of homozygous cultivars. The analysed regions covered the entire transcribed region as well as the 1.7 kb sequence upstream of the start codon (the promoter region). At least 12 polymorphic loci (Fig. 4B, P1-P12). were found in the PpYUC11 gene using a combination of cloning and direct Sanger sequencing of PCR products: within the promoter region, a 25 bp insertion/ deletion (

PpYUC11 allele distribution in peach germplasm
The sequencing of a PCR-amplified DNA fragment from a set of homozygous cultivars confirmed that the distribution of all PpYUC11 variations can be found in both flesh types, with the exception of the TC dinucleotide microsatellite (see Supplementary Fig. S3B at JXB online). To investigate the relationship between the TC microsatellite genotype of PpYUC11 and the phenotype of flesh texture (normal or stony hard), the distribution of the polymorphic loci was analysed in 36 normal and seven stony hard flesh cultivars (Table 1). To maximize genetic diversity, this set included nine American cultivars, 24 Chinese cultivars, nine Japanese cultivars, and one Korean cultivar. In our results, (TC) n microsatellites were found in at least four different lengths (the repeat number 'n' of Hd 1 , Hd 2 , Hd 3 , and hd are 29, 26, 24, and 20, respectively) in these peach cultivars (Fig. 4C). The diversity of microsatellite length variants exhibited dramatic differences between the 36 normal and the seven stony hard flesh peach cultivars. All seven stony hard peaches were homozygous for hd. By contrast, all of the normal peach cultivars had at least one Hd allele (Hd 1 , Hd 2 or Hd 3 ). These results are summarized in Table 1.

Transcriptional analysis of PpYUC11 alleles
The TC microsatellite identified to be responsible for the stony hard flesh phenotype appears to have an effect on transcriptional regulation of the gene. Direct Sanger sequencing was used to estimate the transcript levels of both alleles of PpYUC11 in the mesocarp tissue of stage S4 in several peach varieties that are heterozygous at the TC microsatellite locus. For instance, sequencing of genomic DNA showed that 'Goldhoney 3' carries two different alleles at its PpYUC11 locus: Hd (n=24), thymidine nucleotide (T) existing at position 2364 bp, and hd with a cytidine nucleotide (C) at position 2364 bp within the fourth exon of the PpYUC11 gene (Fig. 5A). By sequencing the PCR products, it was found that, whereas this SNP is heterozygous in the genomic DNA of 'Goldhoney 3' (both C and T), by contrast, it is homozygous in the cDNA in the mesocarp of S4 fruit (Fig. 5B), supporting the hypothesis that the TC microsatellite is responsible for the transcriptional activity of PpYUC11 and for IAA accumulation in ripening fruits and that PpYUC11 is the candidate gene controlling the stony hard phenotype in peach.

Co-ordinated variation of PpYUC11 expression, IAA concentration, PpACS1 expression, ethylene production, and fruit firmness in ripening peaches
A total of 12 peach cultivars with different genotypes in the PpYUC11 SSR loci were selected and used to investigate the relationship between PpYUC11 expression, IAA concentration, PpACS1 expression, ethylene production, and fruit firmness in the mesocarp of fruits during the ripening transition stage (S3-S4). These cultivars belong to five genotypes: 'Xia Cui' and 'CN16' is hd/hd; 'Arctic Star' and 'Goldhoney 1' is Hd 1 /Hd 3 ; 'CN13', 'Elberta', 'Kurakato Wase', 'Spring Honey', and 'Spring Snow' is Hd 3 /Hd 3 ; 'Bao Lu' and 'Matsumori' is Hd 2 /Hd 2 ; and 'CN9' is Hd 1 /Hd 1 (Fig. 6A). For each cultivar, three samples from stage S3 to stage S4 were collected for further study. Overall, PpYUC11 and PpACS1 transcripts were highly expressed in the mesocarp of normal fleshed S4 fruits, but were either low or undetectable in the mesocarp of stony hard or S3 fruits (Fig. 6B, C). These expression profiles were accompanied by similar IAA content profiles or ethylene production in the mesocarp of these peach cultivars (Fig. 6B, C). Flesh firmness decreased towards harvest time along with ethylene emissions (Fig. 6D). These findings further confirmed that PpYUC11 may regulate the IAA content in the mesocarp during the fruit ripening stage and may lead to different ethylene emissions and firmness of the peach fruits.

YUCCA flavin monooxygenase family genes in peach
The peach genome contains nine predicted YUC-like genes (Table 2; Fig. 7A). Based on the DGE analysis data, both PpYUC10 (ppa007054m) and PpYUC11 (ppa008176m) are enriched in the fruit ripening stage of 'Goldhoney 3', but PpYUC11 cannot be detected in 'Yumyeong' fruit during the same period ( Fig. 2; see Supplementary Table S3 at JXB online). The expression of the other seven YUC-like genes in this family (PpYUC1, 2, 3, 4, 6, 8, and 9) could not be detected in selected tissues of the two cultivars (see Supplementary  Table S3 at JXB online).
To uncover the roles for each YUC gene in plant development, the varied expression profiles of the nine YUC genes were analysed during vegetative and reproductive development of 'Goldhoney 3' using qRT-PCR (Fig. 7C). PpYUC1, the gene with the highest sequence identity to PpYUC4 in peach (Fig. 7A), shows expression patterns similar to that of PpYUC4 in the stem, leaf, flower, and seed (Fig. 7C). Among the nine predicted YUC genes, the expression of PpYUC1, PpYUC4, PpYUC6, PpYUC8, and PpYUC10 could be detected in almost all 12 tissues, of which PpYUC1, PpYUC4, PpYUC6, and PpYUC8 accumulated preferentially in the vegetative tissues (stem and leaf) and seeds, and young fruit. PpYUC10 were also enriched in the developing fruit (stages S1 through S4) (Fig. 7B). For PpYUC3, however, mRNA accumulation was detected only in the leaves. PpYUC9 showed expression only in the leaves and seeds of stage S1, whereas PpYUC2 was detected only in the leaves and flowers. Interestingly, PpYUC11 was expressed at various stages in the seeds of 'Goldhoney 3' and 'Yumyeong' (Fig. 7B, C). However, only in the melting flesh fruit 'Goldhoney 3' did PpYUC11 show obvious tissue-specific expression in S4 fruits (Fig. 7B). This result is consistent with the hypothesis that PpYUC11 is responsible for IAA accumulation in the flesh of fruits at the ripening stage and is a candidate gene controlling the stony hard phenotype.

Discussion
Fruit firmness is essential for efficient harvesting, handling, marketing, and storage (Byrne et al., 2012). Due to the potential for the development of firmer peaches with tree-ripe flavour and longer storage life, the inheritance and physiological characteristics of a stony hard flesh texture have been extensively investigated for many years (Yoshida, 1976;Haji et al., 2003Haji et al., , 2005Tatsuki et al., 2006Tatsuki et al., , 2013Begheldo et al., 2008). Because IAA concentration may control ethylene production (Tatsuki et al., 2013) and, in turn, the process of fruit softening in normal flesh peaches, our study investigated the expression patterns of auxin-homeostasis-related genes in the melting flesh peach 'Goldhoney 3' and stony hard peach 'Yumyeong' during the fruit ripening stage.
The distribution of auxin in plant organs is tightly controlled through synthesis, inactivation, and transport, and key regulatory steps and regulation mechanisms controlling auxin homeostasis have been extensively reviewed (Rosquete et al., 2012;Korasick et al., 2013;Sauer et al., 2013). Differential accumulation patterns of IAA in 'Goldhoney 3' and 'Yumyeong' became evident from late stage S3, and reached a maximum at stage S4 III, by which time 'Goldhoney 3' fruits had accumulated approximately 10-fold more IAA than 'Yumyeong' fruits (Fig. 2D). Accordingly, the two cultivars showed strikingly different developmental expression patterns of auxin-homeostasis-related genes ( Fig. 2A). However, the differential expression of several auxin-homeostasis-related genes appeared to be the effect, rather than the cause, of the differential IAA accumulation. For instance, in 'Goldhoney 3', the up-regulation of GH3 (ppa003134m), GH3 (ppa003126m), and PIN (ppa002944m) at late ripening stages appears to be caused by the high IAA levels in this phenotype because the expression levels of these genes increased in NAA-treated mature 'Yumyeong' fruit ( Fig. 3), which is consistent with previous findings that GH3 and PIN are auxin-responsive genes (Peer et al., 2004;Terol et al., 2006). Previous genetic studies demonstrated that the flavin monooxygenases of the YUC family are key enzymes in Trp-dependent auxin biosynthesis (Cheng et al., 2007). In the present study, a YUCCA (YUC) flavin the monooxygenase gene, PpYUC11, showed an identical expression pattern to both PpACS1 (Fig. 2C) and IAA accumulation (Fig. 2D) in both 'Goldhoney 3' and 'Yumyeong', suggesting that the low IAA concentrations at the late ripening stage of stony hard peaches may result from the suppressed expression of PpYUC11.
An examination of 43 peach genotypes allowed the identification of 12 polymorphisms within the PpYUC11 gene as well as in the 1.7 kb upstream flanking sequences: (i) a 25 bp indel and two SNPs upstream of the start codon; (ii) a 19 bp indel, a TC microsatellite, and two SNPs located in intron 1; (iii) an adenine nucleotide (A) indel and two SNPs located in intron 2; and (iv) two nonsense mutation SNPs in exon 3 and exon 4. Remarkably, the presence of a (TC) 20 microsatellite (an allele named hd) appears to be associated with stony hard cultivars when present in a homozygous state, whereas normal-fleshed (M or NM) cultivars possess at least one other allele (Hd 1 , Hd 2 or Hd 3 ) at the PpYUC11 locus (Table 1). The transcriptional analysis of the two alleles of PpYUC11 in the heterozygous variety 'Goldhoney 3' and the co-ordinated variation analysis of PpYUC11 expression, IAA concentration, PpACS1 expression, ethylene emission, and fruit firmness in ripening peaches lends support to the hypothesis that PpYUC11 is a strong candidate gene for the control of the stony hard phenotype in peach.
Microsatellites, or simple sequence repeats (SSRs), represent a unique type of tandemly repeated genomic sequences (iterations of 1-6 bp nucleotide motifs), which are abundantly distributed across genomes and demonstrate high levels of allele polymorphism (Chistiakov et al., 2006). SSRs typically represent selectively neutral DNA markers. However, numerous lines of evidence have demonstrated that microsatellites are non-randomly distributed across protein coding regions, 3'-UTRs, 5'-UTRs, and introns because microsatellites can affect chromatin organization, the regulation of gene activity, recombination, DNA replication, the cell cycle, and the mismatch repair (MMR) system (see the review by Li et al., 2004). For instance, a tetranucleotide polymorphic microsatellite located in the first intron of the tyrosine hydroxylase gene regulates its gene expression (Meloni et al., 1998). In the present study, both the (TC) 24 and (TC) 20 forms of a TC dinucleotide polymorphic microsatellite localized in the first intron of PpYUC11 were found heterozygously in the genomic DNA of 'Goldhoney 3', but Sanger sequencing of the mesocarp tissue cDNA during ripening recovered only the transcript containing the (TC) 24 microsatellite (Hd), indicating that the intronic TC microsatellite of PpYUC11 may regulate the gene expression during ripening and lead to IAA accumulation. However, the causal relationship between the TC microsatellite polymorphisms and the transcriptional level remains to be formally proved. As a whole, our results strongly suggest that the (TC) 20 microsatellite (hd) interferes with PpYUC11 transcription.
Quantitative RT-PCR analysis revealed that PpYUC11 is preferentially expressed in the seeds of both flesh types ('Goldhoney 3' and 'Yumyeong') during all developmental stages but is only expressed in the mesocarp of stage S4 in the melting flesh cultivar 'Goldhoney 3' (Fig. 7B), suggesting that there is another regulatory element, perhaps a transcription factor, that interacts with the hd locus to control the transcription of PpYUC11 in both seed and flesh during different stages of fruit development. Depending upon the existence or non-existence of this transcription factor in flesh, PpYUC11 will be selectively expressed in stage S4; by contrast, PpYUC11 will be constitutively expressed in other tissues, such as seed. Based on this study's results, both PpYUC10 and PpYUC11 are enriched in the fruit ripening stage of 'Goldhoney 3', but PpYUC11 cannot be detected in 'Yumyeong' fruit in the same period ( Fig. 2; see Supplementary  Table S3 at JXB online), suggesting that PpYUC10 is responsible for the biosynthesis of the remaining basic IAA levels in 'Yumyeong' and that PpYUC11 is responsible for the high concentration of IAA in 'Glodhoney 3'.
Auxin has long been considered a negative regulator for fruit ripening, as the content of auxin is low at the initiation of ripening (Nitsch, 1952). The decline of auxin content before the initiation of ripening has been suggested to be a prior condition for fruit ripening (Given et al., 1988;Chen et al., 1999;Purgatto et al., 2002;Böttcher et al., 2010) because the application of exogenous auxins has been shown to delay the ripening of fruits (Cohen, 1996;Purgatto et al., 2002;Fabbroni et al., 2006). It has been suggested that the low auxin levels in the ripening fruit tissues may be controlled by ripeningassociated GH3 genes, which are involved in auxin conjugation (Böttcher et al., 2011;Schaffer et al., 2013). It is interesting that, in contrast to most fruit species, in peach fruits, the concentration of auxin increases suddenly just before ripening and coincides with climacteric ethylene production (Tatsuki et al., 2013). Auxin induces the expression of genes encoding ACC synthase (Tatsuki et al., 2013). The observations from the current study indicate that, in peach, the flavin monooxygenase gene PpYUC11 may regulate the concentration of IAA at the late ripening stage. A higher expression level and stronger up-regulation of PpYUC11 at the transition stage from maturation to the ripening stage (S3 to S4 III) were correlated with IAA concentration and PpACS1 activation in 'Goldhoney 3' and another ten normal-fleshed cultivars (Figs 2, 6). By contrast, it is possible that, in 'Yumyeong', 'Xia cui', and 'CN16', less efficient auxin biosynthesis in the mesocarp due to the very low level of PpYUC11 gene expression may contribute to the reduced IAA level and the stony hard phenotype (Figs 2, 6). The contrasting expression profiles in melting and stony hard flesh peaches of a YUCCA flavin monooxygenase (PpYUC11) during peach fruit ripening are consistent with the hypothesis that a low IAA concentration in ripening stony hard peaches could result from the suppression of IAA biosynthesis, the acceleration of IAA inactivation and degradation, the transition from free IAA to IAA storage forms, or the inhibition of IAA transport from biosynthetic tissues (Tatsuki et al., 2013). The phylogeny was constructed using the Neighbor-Joining method and a bootstrap test with 1 000 iterations using MEGA5 software, and alignments were generated with ClustalW. (B) Quantitative RT-PCR analysis of PpYUC11 in fruit and seed of stage S1, S2, S3, and S4 of P. persica cv. 'Goldhoney 3' and 'Yumyeong'. (C) Quantitative RT-PCR analysis of YUC genes in different tissues of P. persica cv. 'Goldhoney 3'. The expression of nine peach YUC genes in stem, stem tips, leaf, flower, fruits of stage S1, S2, S3, and S4; and seeds of stage S1, S2, S3, and S4 were analysed. The steady-state levels were normalized to actin. Data are means ±SE of three individual experiments, each performed in triplicate. (This figure is available in colour at JXB online.) Peaches originated and diversified from China, and Chinese peach germplasm has had a great impact on breeding research in other countries (Li et al., 2013). After the introduction of 'Chinese Cling' ('Shanghai Suimitsuto') as parents in the early 20th century, Japan selected out 'Hakuto' (Yamamoto et al., 2003;Xu et al., 2006) and the USA released the well-known cultivar 'Elberta'. Both 'Hakuto' and 'Elberta', selected from seedlings of 'Chinese Cling', were extensively used as parents for further breeding of modern cultivars (Scorza et al., 1985;Aranzana et al., 2012). 'Hakuto' was reported to be heterozygous (Hd/hd) for the stony hard gene (Yoshida, 1976), accordingly, 'Hakuto' is heterozygous (Hd 2 /hd) in the intronic TC microsatellite locus in our study (Table 1). The genotype of 'Elberta' at the TC microsatellite locus is Hd 3 /Hd 3 , which explains the lack of any stony hard peach cultivars in modern American history (Goffreda et al., 1999). In our study, the genotype of 'Chinese Cling' was found to be Hd 1 /Hd 3 , a result that is in conflict with the genotype of the progeny 'Hakuto' (Hd 2 /hd). It has been suggested that 'Chinese Cling' was not a single but, rather, a group of cultivars (Wang and Zhuang, 2001); if so, only a subset of this group of cultivars may have carried the hd allele, and the 'Chinese Cling' that was tested may not have been from this subset. Similar examples are 'Jing Yu' (hd/hd), a well-known stony hard cultivar from China, selected from among the crosses of 'Okubo' (Hd/hd)×'Okitsu' (Hd/hd) and 'Qin Wang' (hd/hd, stony hard), which is a new peach cultivar selected from the seedlings of 'Okubo' (Hd/hd). The pedigrees of all these cultivars are consistent with the hypothesis that the intronic TC microsatellite of PpYUC11 is the hd locus.
Stony hard peaches are characterized by the absence of both ethylene production and softening in mature fruits and are expected to be used as a genetic source for breeding new table peaches (Haji, 2014). However, stony hard fruits are often very difficult to distinguish from NM or very firm, unripe M phenotypes in the field, thereby making reliable selection difficult (Bassi and Monet, 2008;Byrne et al., 2012). Genotyping the TC microsatellite of PpYUC11 can now be used for the marker-assisted breeding of new cultivars of the stony hard flesh type.

Conclusions
In this study, DGE analysis of auxin-homeostasis-related genes in melting flesh and stony hard peaches during fruit ripening has been reported. This strategy allowed several auxinhomeostasis-related genes to be identifed showing a similar expression pattern to the ACC synthetase gene PpACS1. The expression pattern, predicted function, and tissue distribution make PpYUC11 an excellent candidate gene for stony hard type peaches. It is proposed that a TC microsatellite in its first intron is responsible for regulating its gene expression in the mesocarp. In addition, the marker developed on this sequence polymorphism provides a convenient molecular tool with which to discriminate normal/stony hard flesh cultivars or individuals in breeding populations. Our data indicate that a TC microsatellite in PpYUC11 is the molecular basis of stony hard flesh peaches, and our results provide a basis for detailing breeding programmes for peach fruit texture improvement.

Supplementary data
Supplementary data can be found at JXB online. Supplementary Fig. S1. Schematic representation of the auxin homeostasis pathway in normal or stony hard flesh peaches. Supplementary Fig. S2. qRT-PCR validation of 11 differentially expressed genes as detected by DGE analysis. Supplementary Fig. S3. Alignment of the PpYUC11 in melting and stony hard flesh cultivars.