Endopolygalacturonase Gene Polymorphisms: Asset of the Locus in Different Peach Accessions

Endopolygalacturonase (endoPG) plays a pivotal role in determining peach [Prunus persica L. (Batsch)] fruit characteristics. Different Pp-endoPG genes or allelic variants have been described, characterized by different polymorphisms: insertions-deletions (InDels) and single nucleotide polymorphisms (SNPs). Eighty-five peach accessions (comprising commercial cultivars, F1 progenies of selected crosses, and three haploid seedlings) with different flesh softening patterns (Non Melting: NM; Melting: M; Slow Softening: SS; Stony Hard: SH) were screened by exploiting specific polymorphisms, with the aim to characterize their asset at the endoPG locus and evaluate a potential relationship with fruit flesh texture phenotype. The results of InDel analysis allowed to distinguish, by a simple genotyping procedure, NM flesh phenotypes from the others. Further information arose from this analysis, showing that two Pp-endoPG genes, i.e., Pp-endoPGm (Ppa006839m), involved in the determination of the Melting/Non Melting trait, and Pp-endoPG_M (Ppa006857m), involved in the determination of the Clingstone/Freestone trait, always co-segregate, and that SS Big Top possesses a “null” Pp-endoPG allele. Cleaved Amplified Polymorphic Sequence (CAPS) analysis allowed to preliminarily discriminate the Pp-endoPG variants of the SS and SH accessions considered. The integrated use of the considered polymorphisms in a high number of peach accessions proved useful, by individuating the different gene variants and their combinations, to describe the structure of the endoPG locus in different genotypes.


Introduction
Flesh texture and stone adherence are important factors of the overall quality of the peach [Prunus persica (L.) Batsch] fruit, contributing to the consumer's satisfaction, fruit nutritional features, and postharvest behavior [1] [2].
Peaches are classified on the basis of their different flesh softening behavior during ripening.In Melting (M) flesh, a marked and steep loss of firmness ("melting" phase) occurs in the final stages of ripening.The melting process is delayed in Slow Softening (SS) fruit and undetectable in Non Melting (NM) and Stony Hard (SH) ones [2] [3] [4] [5] [6].The SS phenotype is very interesting, retaining flesh firmness on the tree for longer time than standard M, with full development of sensory qualities, and remarkable keeping quality appreciated by both growers and consumers.The spontaneous occurrence of melting, although a few days later than standard M, allows to group SS fruit into the M phenotype, since the physiological basis of SS is due to a delay in ethylene production [6].
Unfortunately, this phenotype, like the SH one, is of very difficult assessment on the tree when scoring segregating progenies [7].
Fruit flesh texture is affected by biochemical and physiological factors with a major role played by hydrolytic enzymes and other proteins that cooperate to modify the composition and architecture of the cell wall polysaccharides [8] [9].
In particular, peach flesh melting has been related to a strong increase in the expression of Pp-endopolygalacturonase (Pp-endoPG) and accumulation of an active Pp-endoPG protein.In NM fruit, Pp-endoPG is absent or detected at levels lower than in M ones [5] [10] [11] [12].Therefore, Pp-endoPG is considered as a candidate gene for peach flesh softening behavior [13].
In eukaryotic organisms, Insertion-Deletion (InDel) and Single Nucleotide Polymorphisms (SNPs) represent the most abundant DNA mutations ( [14] and references therein) that can modify the amplification patterns of selected DNA sequences as well as the restriction endonuclease recognition sites in PCR amplicons, allowing the development of markers useful for genotyping even in the absence of massive sequencing data ( [15] and references therein).
In peach, availability of different molecular markers allowed to gain information on some 760 Quantitative Trait Loci and Mendelian Trait Loci linked to horticultural and physiological traits such as tree development, pest and disease resistance, flowering, ripening, seed and fruit quality [16] [17].
The availability of the peach genome sequence [18] and the development of SNP genotyping resources [19] [20] offer the opportunity to study at the molecular level the development and inheritance of different phenotypic traits, as recently reported for skin blush [21].
A high degree of allelic diversity has been identified in the peach Pp-endoPG sequence [13]  F a -Ppa006839m (Pp-endoPG m ), the only sequence retrieved in Oro A, shows 97 % identity with its variants Pp-endoPG SH (GenBank ID: HQ891822.1 of SH Yumyeong and GenBank ID: HQ891821.1 of SH Ghiaccio) and Pp-endoPG BT (GenBank ID: HQ891820.1 of SS Big Top), for the presence in the latter ones of one 17-bp intronic deletion.Moreover, the Big Top-specific Pp-endoPG BT sequence shows a peculiar SNP (bp 348) in Exon 1, that does not induce any change in the deduced amino acidic sequence [6].F b -Ppa006857m (Pp-endoPG_M), present in Bolero together with F a -Ppa006839m, is by 34 bp shorter than Pp-endoPG m , for the presence of one additional 17-bp intronic deletion.On this basis, amplification of a selected (1455-1892 bp) Pp-endoPG genomic sequence including these InDels generates fragments specifically referable to F a -Ppa006839m (Pp-endoPG m ; 437 bp), to its variants Pp-endoPG SH and Pp-endoPG BT (both of 420 bp), or to F b -Ppa006857m (Pp-endoPG_M; 403 bp) [6].F a -Ppa006839m has recently been proposed as the main responsible for the Melting trait [23].
In the present work the InDel and the CAPS (Cleaved Amplified Polymorphic Sequence) markers developed from the described polymorphisms have been exploited to screen a total of 85 accessions (commercial cultivars, F 1 offsprings from selected crosses, and three haploids) to describe them for the configuration at the endoPG locus and evaluate a potential relationship with fruit flesh texture phenotype.

Plant Material
The study was conducted on commercial cultivars or on F 1 offsprings of different crosses.Three haploid accessions (P VIN 1 1N from Vineland, P RRL 1 1N from Rutgers Red Leaf, P LOV 3 1N from Lovell) were also considered (Table 1).
Plant material was obtained from a peach germplasm collection grown at the Experimental Orchard "Zabina" (Castel San Pietro Terme, Italy; lat.44˚23'52''N; long.11˚35'22''E), under the weather and climate conditions detailed in Table 2. Trees were grafted on GFF677 rootstock planted at 3.5 × 4.5 m and trained as open vase.Seedlings were planted on their own roots with a spacing of 1 m within and 4 m between rows and trained as slender spindle (one stem with short lateral scaffolds).Pruning was performed yearly and standard cultural practices were applied.

Fruit Flesh Texture Phenotype Scoring
Flesh texture was scored in the orchard by trained personnel on fruit harvested at commercial ripening (onset of the veraison stage), as identified in each  Assessment of the latter parameter by human inspectors' grading is coherent with instrumented measurements [26], and represents a non-destructive method capable to yield information on the time course of fruit texture evolution.
Discrimination between M and SS phenotypes was based upon sensory scoring of flesh texture changes during 4 d of postharvest at room temperature.Due to stunted growth and low fruit set, in a few F 1 offsprings from selected crosses, fruit were not phenotypically classifiable; haploids also had stunted growth and did not produce fruit at all, so the fruit flesh phenotype remained unascertained.

Extraction of Genomic DNA
Genomic DNA from young leaves was used as template for the analysis of the amplification profiles based on major InDel polymorphisms in Intron 3 of the Pp-endoPG sequence [5] [6], and for CAPS analysis [27] based on the SNP 348 of the Pp-endoPG BT variant of F a -Ppa006839 m [6].Extraction was conducted according to Geuna et al. [28].Frozen leaf samples (100 mg fresh weight) were powdered in liquid N 2 by a mortar and a pestle and at 37˚C for 1 h to remove any RNA contamination and precipitated as described above.

InDel Analysis
A ~440 bp sequence of the Pp-endoPG gene region encompassing a few bases of Exon 3 and a wide sequence of Intron 3 comprising the major InDels [6] was amplified with proper primers (Pp-endoPG_InDel For : 5'-GTGCCCTGGTCAGGTAAG-3'; Pp-endoPG_InDel Rev :

InDel Analysis of Different Peach Accessions
Amplification of the cited genomic sequence of Pp-endoPG including the InDels of interest confirmed the presence of two amplicons of apparent length of ~440 bp and ~410 bp in Bolero, consistent with the expected lengths of 437 bp and 403 bp for F a -Ppa006839m (Pp-endoPG m ) and F b -Ppa006857m (Pp-endoPG_M) [5] [6].The M cultivars Contender, Maycrest, Springbelle, Springcrest, Springred, Glohaven, plus two (BO 96016136 and BO 96016094) out of the seven offsprings from the cross Contender (M) × Ambra (M) showed a Bolero-like amplification pattern with two fragments.In Ambra and in five offsprings from Contender × Ambra (BO 96016015, BO 96016018, BO 96016023, BO 96016165, and BO 96016208) an additional amplicon of ~420 bp was observed (Figure 1).
Figure 2 shows the results obtained in NM, SH, and SS accessions.In the NM Oro A, Andross [29], Ionia, and BO 82010554 (offspring of NM Jungerman × NM Loadel), only one amplicon of ~440 bp was retrieved.In the SH accessions Yumyeong, Ghiaccio, Helena Cling, and D 41-62, as well as in SS Big Top, amplification generated only one fragment of ~420 bp, confirming the presence of the Pp-endoPG SH -or Pp-endoPG BT -like variants of F a -Ppa006839m, respectively [6].
The amplification pattern of SS Alitop was similar to that of Bolero.SH 7-28 and the SS Honey Gold, Ruby Rich, Vista Rich, BO 96028059 (F 1 from M Springred × SS Big Top) and BO 96013046 (F 1 from M Bolero × SS Rich Lady) yielded three amplicons (~440 bp, ~420 bp, and ~410 bp).The two haploids P VIN 1 1N and P LOV 3 1N yielded amplification patterns similar to that of Oro A (a single fragment of ~440 bp), whereas the pattern of P RRL 1 1N was similar to that of Bolero (two fragments of ~440 bp and ~410 bp; Figure 3).

Ambra × Big Top and Max 7 × Big Top Crosses: InDel Analysis of F1 Offsprings
Figure 4 shows the amplification patterns obtained in the M Ambra and SS Big Top parents and in 23 F 1 seedlings from their cross (from BO 02002001 to BO 02002006, BO 02002008, BO 02002009, from BO 02002020 to BO 02002026, from BO 02002011 to BO 02002015, from BO 02002017 to BO 02002019).The amplification profiles were similar to that of the Big Top parent (one amplicon of ~420 bp) for eight of them, and to that of the Ambra parent (three amplicons of ~440 bp, ~420 bp, and ~410 bp) for 12. Three accessions showed the same pattern as Bolero (~440 bp and ~410 bp; compare to Figure 1).

Cleaved Amplified Polymorphic Sequence (CAPS) Analysis
In the Pp-endoPG BT variant of F a -Ppa006839m present in Big Top, a peculiar, silent SNP 348 in Exon 1 originates a polymorphic sequence recognized by the BstXI restriction enzyme (Figure 6(b)), suitable for distinguishing between the Pp-endoPG BT and Pp-endoPG SH variants of F a -Ppa006839m (Pp-endoPG m ) [6].
The SNP 348 was exploited to broaden CAPS analysis to a few additional cultivars/accessions phenotypically scored as SS (Alitop, Helena Cling, Honey Gold, Ruby Rich, Vista Rich, BO 96028059, BO 96013046) or SH (D 41-62, 7- (Table 1).BstXI confirmed to be ineffective in Oro A and in Bolero, as it was in all the accessions scored as SH.Concerning the accessions scored as SS, Helena Cling showed, like Big Top, two digestion fragments; Honey Gold, Ruby Rich, Vista Rich, BO 96028059, and BO 96013046 showed two Big Top-like restriction fragments plus an undigested one, whereas BstXI was ineffective in Alitop (Figure 6(c)).

Discussion
The F allele at the endoPG locus on peach chromosome 4 contains two gene sequences, i.e.F a (Ppa006839m) and F b (Ppa006857m), at short distance (32-34 kbp) from each other.Mutations and deletions of these genes determine several allelic variants: F, resulting from F a plus F b ; f, resulting from mutation of F a plus complete deletion of F b ; f1, resulting from F a plus complete deletion of F b ; f null , resulting from deep mutation or complete deletion of both F a and F b .The different alleles F, f, f1, f null at the endoPG locus contribute to the fruit phenotype for stone adhesion and flesh texture [4] [18].In particular, it has been recently suggested that F a -Ppa006839m plays a driving role in the determination of the Melting/Non Melting trait, whereas F b -Ppa006857m seems involved in the determination of the Clingstone/Freestone trait, although it has been speculated that F b -Ppa006857m could exert a pleiotropic effect on flesh melting through a negative feedback control on the transcription of F a -Ppa006839m [23].
From reference peach genotypes, peculiar endoPG sequences have been isolated.In particular, in NM Oro A has been retrieved a single sequence (Pp-endoPG m ) coincident with Ppa006839m, that does not present any deletion were exploited in the present work as tools to describe 85 peach accessions (Table 1).InDel analysis confirmed the co-segregation of   7; Table 1).
Most (31) of the 64 accessions scored as M or SS (all grouped in a general Melting class [6] [7]) were characterized by presence of all the three amplicons (~440 bp, ~420 bp and ~410 bp), referable respectively to F a -Ppa006839m (no deletion), its Pp-endoPG BT/SH -like variants (one 17-bp deletion), and F b -Ppa006857m (two 17-bp deletions).Eighteen accessions showed a pattern with the two amplicons of ~440 bp and ~410 bp, and 15 showed only the amplicon of ~420 bp (Figures 1-4; Table 1).From these amplification profiles, it seems reasonable to deduce that the M and SS accessions can present all the three combinations of amplicons.
In particular, the InDel profiles of M Contender, M Maycrest and M Springcrest (Figure 1; Table 1) appear consistent with their allelic combinations F/F and F/-reported in the literature [30] [31].
The InDel polymorphism, when scored in a cross progeny, appears useful to deduce the allelic combination at the endoPG locus of the parental genotypes, as in the case of Big Top.In fact, a few F 1 individuals of the crosses Ambra × Big Top or Max 7 × Big Top, as well as Alitop [(Flavortop × Snow Queen) × Big Top], lacked the amplicon of ~420 bp (Figure 2, Figure 4 and Figure 5).These results can only be explained by hypothesizing that Big Top is heterozygous at the endoPG locus for the presence of a "null" allele, in addition to the f allele composed only by the Pp-endoPG BT variant of F a -Ppa006839m.The literature proposes for SS Big Top the allelic combinations f/f or f/f null [30] [31]; our results seem to support the latter one (Figure 4 and Figure 5; Table 1).
The coherence of the allelic combinations at the endoPG locus inferred by the InDel analysis with, where available, those reported in the literature (Table 1) suggests that the InDel polymorphism may represent a tool for a simple genotyping of peach accessions.Nevertheless, this polymorphism does not allow to discriminate between the peculiar gene sequences individuated in SS or SH genotypes.
CAPS analysis exploiting the SNP individuated in Exon 1 of the endoPG sequence of SS Big Top integrates the InDel results allowing to distinguish between amplicons of ~420 bp derived from the Pp-endoPG BT -like variant of F a -Ppa006839m and the Pp-endoPG SH -like one [6].All the SS genotypes tested in CAPS analysis proved sensitive to BstXI producing upon cleavage two digestion fragments.In a few accessions, the contemporary presence of an undigested fragment indicated that their allelic asset was heterozygous (F/f; Figure 6(c) and Figure 2).For Ruby Rich, this asset confirms the functional genotype (F/f) proposed in the literature [30] [31] (Table 1).Lack of effect of BstXI in SS Alitop further confirms the inheritance of a possible 'null' Pp-endoPG variant from the Big Top parent.Conversely, the SH genotypes tested were not sensitive to BstXI digestion (Figure 6(c)).
In the climacteric peach fruit the evolution of texture characteristics is regulated by ethylene.The SH trait (hdhd) is related to absence of ethylene production for a mutation in the Pp-ACS1 gene and altered indole-3-acetic acid levels due to regulation of PpYUC11 [6] [32] [33] [34].In SH 7-28, homozygous-recessive for the stony hard gene (hdhd) [35], InDel analysis shows presence of the ~420 bp amplicon (Figure 2) attributable, by CAPS analysis, to a Pp-endoPG SH -like sequence (Figure 6(c)).The contemporary presence of the other two amplicons allows to assign this accession, as far as it concerns Pp-endoPG, to the Stony Hard melting category (hdhd M-) consistent to Haji et al. [36].
Therefore, it can be concluded that although CAPS analysis of the specific Pp-endoPG BT/SH variants seems to allow a preliminary distinction of SS from SH genotypes, conclusive phenotypic attribution of a specific accession will necessarily require the assessment of other activities like fruit ethylene production.

Conclusions
The two molecular markers (InDel and CAPS) based upon Pp-endoPG poly-morphisms tested in the present work may represent useful tools for simple genotyping of peach accessions as far as it concerns the characteristics and different allelic combinations at the endoPG locus.
In particular, the use of the Pp-endoPG InDel marker yields results consistent with the existing knowledge on the polymorphic structure of the endoPG locus in peach, and allows to individuate NM accessions.The CAPS marker developed on the basis of the peculiar SNP 348 of SS Big Top integrates the results of InDel analysis by discriminating the Pp-endoPG variants in SS and SH accessions.
Future development of the work may foresee the testing of a broader number of NM, as well as SS and SH accessions, in order to confirm the potential use of the described InDel and CAPS markers in Marker-Assisted Selection peach breeding programs.

Figure 1 .
Figure 1.InDel analysis of Pp-endoPG in accessions with M fruit flesh texture.Positions and lengths of DNA markers (DirectLoad™ 50 bp DNA Step Ladder; Sigma-Aldrich) are shown on the left.Twenty micrograms DNA were loaded per lane; the amplicons were separated on 3% (w/v) agarose gels.One representative gel from three independent experiments.

Figure 2 .
Figure 2. InDel analysis of Pp-endoPG in accessions with NM, SH or SS fruit flesh texture.Positions and lengths of DNA markers (DirectLoad™ 50 bp DNA Step Ladder) are shown on the left.Twenty micrograms DNA were loaded per lane; the amplicons were separated on 3% (w/v) agarose gels.One representative gel from three independent experiments.

Figure 5
Figure5shows the amplification profiles obtained in the M Max 7 and SS Big Top parents and in 27 F 1 seedlings from their cross (from BO 02004001 to BO 02004027).In Max 7, as well as in eight offsprings, amplification yielded three fragments (~440 bp, ~420 bp, and ~410 bp).Amplification profiles similar to that of Big Top (one fragment of ~420 bp) were found in 13 seedlings, whereas six seedlings yielded Bolero-like (~440 bp and ~410 bp fragments) patterns (compare to Figure1).

Figure 3 .
Figure 3. InDel analysis of Pp-endoPG in three haploid accessions.Positions and lengths of DNA markers (50 bp DNA Step Ladder; Sigma-Aldrich) are shown on the left.Twenty micrograms DNA were loaded per lane; the amplicons were separated on 3% (w/v) agarose gel.One representative gel from three independent experiments.

Figure 4 .
Figure 4. InDel analysis of Pp-endoPG in M Ambra, SS Big Top and in 23 F 1 seedlings.Positions and lengths of DNA markers (DirectLoad™ 50 bp DNA Step Ladder) are shown.NC, not classifiable.Twenty micrograms DNA were loaded per lane; amplicons were separated on 3% (w/v) agarose gel.One representative gel from three independent experiments.

Figure 5 .
Figure 5. InDel analysis of Pp-endoPG in M Max 7, SS Big Top, and in 27 F 1 seedlings.Positions and lengths of DNA markers (DirectLoad™ 50 bp DNA Step Ladder) are shown on the left.NC, not classifiable.Twenty micrograms DNA were loaded per lane; amplicons were separated on 3% (w/v) agarose gel.One representative gel from three independent experiments.

Figure 6 .
Figure 6.CAPS analysis exploiting the SNP 348 of the Pp-endoPG BT variant of F a -Ppa006839m.(a) Schematic representation of the structure of the Pp-endoPG gene.Exons and introns are indicated by black solid bars and by solid lines, respectively; positions of the considered InDels of Intron 3 are also indicated.Arrows and vertical dotted lines define the region (1-972 bp) amplified by the primers used.(b) Lengths (~972 bp) of the amplicons obtained from the F a -Ppa006839m variants of: NM Oro A and M Bolero (Pp-endoPG m ), SH Yumyeong and SH Ghiaccio (Pp-endoPG SH ), and SS Big Top (Pp-endoPG BT ); position of the peculiar SNP 348 of the Pp-endoPG BT variant and predicted lengths (348 bp and 626 bp) of the fragments obtainable by BstXI digestion (scissors symbol) are also indicated (modified from [6]).The Pp-endoPG_M amplicon from F b -Ppa006857m of Bolero is also reported.(c) Restriction patterns obtained in different accessions before (-) and after (D) digestion with BstXI.Positions and lengths of DNA markers (1 Kb Plus DNA Ladder, Invitrogen) are shown on the right.Expected lengths (bp) of the undigested amplicons and of the products of BstXI digestion are shown on the left.Twenty micrograms DNA were loaded per lane; 3% (w/v) agarose.One representative gel from three independent experiments.
F a -Ppa006839m (Pp-endoPG m ) and F b -Ppa006857m (Pp-endoPG_M).In fact, the longest (~440 bp) amplicon from F a -Ppa006839m was always accompanied by the shortest (~410 bp) one from F b -Ppa006857m, as indicated by the simultaneous presence of three amplicons in several accessions and of both amplicons in the haploid P RRL 1 1N

Figure 7 .
Figure 7. Schematic picture of the structure of the Pp-endoPG locus with the different allelic variants and the discussed InDel mutations, as deduced from [4] [5] [6] [18].C: Clingstone; F: Freestone.Exons are indicated by black solid bars and introns by solid lines.

Table 1 .
List of the Prunus persica accessions used, with fruit flesh texture phenotype, pedigree, estimated length of endoPG-derived amplicons, and asset at the endoPG locus as hypothesizable by InDel analysis or, where available, reported in the literature.M: Melting; NM: Non Melting; SH: Stony Hard; SS: Slow Softening.NC, not classifiable.OP, open pollination.-, lack of fruit production.

Table 2 .
Climate and weather conditions recorded in the orchard during the time period of fruit growing season.n. d.: not determined.