TY - JOUR
AU - Costa, Fabrizio
AB - Abstract Fruit ripening is a complex physiological process in plants whereby cell wall programmed changes occur mainly to promote seed dispersal. Cell wall modification also directly regulates the textural properties, a fundamental aspect of fruit quality. In this study, two full-sib populations of apple, with ‘Fuji’ as the common maternal parent, crossed with ‘Delearly’ and ‘Pink Lady’, were used to understand the control of fruit texture by QTL mapping and in silico gene mining. Texture was dissected with a novel high resolution phenomics strategy, simultaneously profiling both mechanical and acoustic fruit texture components. In ‘Fuji×Delearly’ nine linkage groups were associated with QTLs accounting from 15.6% to 49% of the total variance, and a highly significant QTL cluster for both textural components was mapped on chromosome 10 and co-located with Md-PG1, a polygalacturonase gene that, in apple, is known to be involved in cell wall metabolism processes. In addition, other candidate genes related to Md-NOR and Md-RIN transcription factors, Md-Pel (pectate lyase), and Md-ACS1 were mapped within statistical intervals. In ‘Fuji×Pink Lady’, a smaller set of linkage groups associated with the QTLs identified for fruit texture (15.9–34.6% variance) was observed. The analysis of the phenotypic variance over a two-dimensional PCA plot highlighted a transgressive segregation for this progeny, revealing two QTL sets distinctively related to both mechanical and acoustic texture components. The mining of the apple genome allowed the discovery of the gene inventory underlying each QTL, and functional profile assessment unravelled specific gene expression patterns of these candidate genes. Apple fruit texture, cell wall genes, ethylene, in silico gene mining, mechanical and acoustic components, phenomics, QTL mapping, transcription assay Introduction Fruit development and ripening is a programmed physiological process unique to plants that has been intensively studied by the scientific community because of the importance of fruits in the human diet (Giovannoni, 2001; Moore et al., 2002, 2005; Alba et al., 2005). Amongst all the biochemical and physiological evolution of traits occurring during ripening, change in texture is one of the most evident variation (Giovannoni, 2004). Texture change is interdependently co-ordinated by a wide range of cell-wall enzymes acting on both the middle lamella and the primary cell wall that, together with alteration of the turgor pressure, cause the weakening of the cell wall structure leading to seed dispersal and the final conversion of unripe hard fruit into edible soft and crispy fruit (Brummel, 2006; Saladié et al., 2007; Thomas et al., 2008). The primary cell wall is mainly composed of a series of polysaccharides as well as structural proteins and phenolic compounds. The polysaccharides are degraded by several enzymes (Brummell and Harpster, 2001), which normally increase their activity during the maturation phase. Fruit texture is nowadays recognized as a combination of different features, all dependent on the anatomical properties of the primary cell wall. Force applied to the apple cortex breaks the chemical bonding of either the middle lamella or the cell wall, resulting in a mealy or crispy apple flesh texture characteristic, respectively (Andani et al., 2001; Duizer, 2001). Cell wall disassembly has been described as a complex physiology due to the synergic action of several enzymes activated in a time-specific fashion to degrade the cell wall polysaccharide network. The complexity of this physiology is further illustrated by the recent discovery that, in the genomes sequenced to date, almost 10% of the entire gene inventory is devoted to cell wall metabolism (McCann and Rose, 2010). Functional surveys to profile cell wall gene dynamics over apple fruit maturation and ripening have recently been published employing different platforms (Jannsen et al., 2008; Soglio et al., 2009; Costa et al., 2010a). The contribution of each gene in the trait control can be quantitatively estimated by a QTL mapping approach (Patterson, 1998; Asins, 2002; Collard et al., 2005; Moose and Mumm, 2008; Hospital, 2009). Specific genomic regions controlling fruit texture in apple have already been reported, targeting major QTLs positioned on linkage groups (LG) 1, 10, and 15, and co-localized with the Md-ACS1, Md-ACO1, Md-Exp7, and Md-PG1 candidate genes (Harada et al., 2000; Costa et al., 2005, 2008, 2010b; Zhu and Barritt, 2008). Recently, three elements derived from the FRUITFULL and SHATTERPROOF of Arabidopsis thaliana orthologous genes have been mapped on other linkage groups, such as 6, 9, and 14 (Cevik et al., 2009), further suggesting apple texture to be a complex trait. Fruit texture, besides its biological relevance, is also a fundamental aspect in the definition of general fruit quality, together with appearance, flavour, and nutritional properties (Bourne, 2002). Fruit texture has the capacity to influence general consumer appreciation (Harker et al., 2003). From an anatomical point of view the propagation of the cell wall disruption, together with the internal turgor pressure, generates an expanding sound pressure wave perceived by the human senses as crispness (Kilcast, 2004). This feature, related to the integrity and rigidity of the cell wall, is also physiologically associated with other important aspects of the fruit quality, such as juice release and fruit freshness (Hampson et al., 2000). A major goal in modern breeding programmes is the selection of a novel cultivar able to overcome rapid flesh decay (Matas et al., 2009). To achieve these results, precise QTL mapping performed with high resolution phenotyping may represent the key strategy to link the perceived quality with the physiological structure of the plant cell wall (Waldron et al., 2003), as well as to target the genomic regions involved in the control of fruit texture. In this work, a comprehensive QTL mapping is described for fruit texture in apple, distinguishing both the mechanical and acoustic texture components. Based on our knowledge, the acoustic response has never been instrumentally characterized for QTL mapping purposes and we are not aware of any references regarding the identification of the genomic regions or candidate gene discovery involved in the control of the dissected texture sub-traits. The anchoring of the QTL genomic intervals to the assembled genome (Velasco et al., 2010) allowed the discovery of a novel gene set that could be exploited in the future to gain a new insight into an understanding of cell wall physiology in apple. Materials and methods Plant material Two full-sib progenies, derived by crossing the high texture quality apple cv. ‘Fuji’ (common maternal parent) with cv. ‘Delearly’ (poor texture cultivar) and cv. ‘Pink Lady’ (high texture cultivar) had 94 seedlings each, were located in the same block at the experimental orchard of FEM (Foundation Edmund Mach, Trento, Italy), and maintained in situ following standard technical management. Each individual was grafted onto M9 rootstock and, at the time of the first harvest, the plants were 10 years old. Total genomic DNA was isolated from young leaf tissue using the Qiagen DNeasy Plant kit. DNA quantity and quality was measured spectrophotometrically with a Nanodrop ND-8000® (Thermo Scientific, USA). Fruit texture assessment The optimal harvest time for the two progenies, occurring from August to the beginning of November, was instrumentally determined through the use of a vis/NIR (near infrared spectrometer) DA-meter (Costa et al., 2009), collecting the samples within an IAD range of 1–1.4. After harvesting, fruit samples were stored at 2 °C in a controlled temperature cellar for 2 months prior to assessment. High resolution phenotyping was carried out by measuring the apple texture components in three consecutive years (2008, 2009, and 2010). For the first two years both populations were assessed, while for the last experimental season the phenotyping was carried out only for ‘Fuji×Delearly’ (‘Fj×Del’). ‘Fuji’, is affected by biennial bearing and, because of that, the yield of 2010 for ‘Fuji×Pink Lady’ (‘Fj×PL’) was insufficient for genetic investigation and so this year of analysis was not included in the study. Mechanical and acoustic signatures were detected and measured using a texture analyser TA-XTplus coupled with an AED (acoustic envelop device; Stable Micro System Ltd., Godalming, UK). TA-XTplus is a microprocessor-controlled texture analysis system which measures force, distance, and time providing three-dimensional analysis. The probe carrier contains a very sensitive load cell (5 kg for this study) with an electronic overload protection. Mechanical measurements were operated with a 4 mm flat probe, speed of 100 mm min−1 and auto-force trigger at 5 g. Distance was expressed as % of compression (strain). An acoustic device (AED) connected to the instrument allowed the additional and simultaneous collection of the second pressure released during the sample’s fracturing. Sample preparation, instrument setting, and parameter characterization are fully described in Costa et al. (2011). Mechanical and acoustic assessments were performed in an isolated room avoiding any possible external noise interference. Texture profiles were analysed with an ad hoc compiled macro (operated with Exponent v.4 software, Stable Microsystems) which allowed the definition of three parameter categories: (i) mechanical: yield force (end of the initial slope), maximum force, final force, mean force, area, force linear distance, Young’s modulus of elasticity, and number of force peaks; (ii) acoustic: acoustic linear distance, number of acoustic peaks, maximum and mean acoustic pressure; (iii) force direction: Δforce and force index (as difference and ratio, respectively, between the yield force and the final force). For each sample, 20 measurements were performed, composed of five technical (samples obtained from the same fruit) and four biological replicates (samples obtained from different fruit). Molecular marker genotyping To perform genetic mapping and QTL calculation, two types of genetic markers were used: SSR and SNPs. The microsatellite markers used to anchor the two maps to the reference (Maliepaard et al., 1998; Liebhard et al., 2002) were selected based on their chromosome position as shown in the HiDRAS web-site (www.hidras.unimi.it). SSR primers were assembled in new triplex (three primer pairs in multiplex; see Supplementary Table S1 at JXB online) and labelled with three different fluorochromes (6-FAM, HEX, and NED). PCR reactions were performed in a final volume of 20 μl with 5 ng DNA, 10× buffer, 0.25 mM dNTPs, 0.075 μM forward labelled and reverse primers and 0.625 U of Eppendorf®Taq polymerase. Amplification of the thermal conditions started with a denaturation at 94 °C for 2 min followed by 10 cycles of denaturation at 94 °C for 30 s, annealing at 58 °C for 30 s, extension at 72 °C for 1 min; 15 cycles of denaturation at 94 °C for 30 s, annealing at 57 °C for 30 s, extension at 72 °C for 1 min, and a final round of 10 cycles of denaturation at 94 °C for 30 s, annealing at 56 °C for 30 s, extension at 72 °C for 1 min, finishing with a final extension at 72 °C for 5 min. The three consecutive rounds of annealing temperatures allowed the amplification of all the triplex employing the same thermal setting, avoiding specific optimization for each set. Single SSR PCR mix contained 5 ng of DNA, 10× buffer, 0.25 mM dNTPs, 0.12 μM of forward labelled and reverse primers and 0.5 U of Eppendorf®Taq polymerase in a final volume of 12.5 μl. Amplification thermal setting started with 94 °C for 2 min followed by 32 cycles of denaturation at 94 °C for 30 s, annealing at 58 °C for 30 s, extension at 72 °C for 45 s, and a final extension at 72 °C for 5 min. SSR-anchored to functional candidate genes were also positioned on both populations. Sequences used for SSR screening regarding the polygalacturonase (Md-PG1) and the alcohol acyltransferase (Md-AAT6) genes were retrieved from literature (Costa et al., 2010b; Dunemann et al., 2011, respectively). The other elements were derived by a combined homologous/heterologous investigation (Costa et al., 2010a). Sequences for MADS-RIN and NOR were obtained from the TED database (Tomato Expression Database, http://ted.bti.cornell.edu; Fei et al., 2006), and used to query the apple genome via the BlastN algorithm. The SSR based on candidate genes were tested following the strategy of the labelled tail (Schuelke, 2000), where a M13, SP6 or T7 sequence (tail) was synthesized at the 5' of each forward primer, acting as annealing site for the fluorescent probe. The PCR reaction mix for the tailed SSR were carried out in a final volume of 25 μl containing 5 ng of DNA, 10X buffer, 0.2 mM dNTPs, 0.2 μM reverse primers, 0.04 μM forward primers, 0.16 μM labelled primer, and 0.75 U of Eppendorf®Taq polymerase. The amplification condition was 94 °C for 90 s, 30 cycles of denaturation at 94 °C for 30 s, annealing at 58 °C for 30 s, extension at 72 °C for 1 min, 10 cycles of denaturation at 94 °C for 30 s, annealing at 53 °C for 30 s, extension at 72 °C for 1 min, and a final extension at 72 °C for 7 min. Fragments analysis was performed with an ABI PRISM® 3730 capillary sequencer (Applied Biosystems by Life Technologies) in a final mix of 0.3 μl of PCR product, 9.67 μl formamide, and 0.03 μl of 500-LIZ denaturated for 3 min at 95 °C. Fragment sizing was operated with GeneMapper® software (Applied Biosystems, by Life Technologies). Both maps were saturated using SNP markers (see Supplementary Table S2 at JXB online) high throughput genotyped with two genomic platforms: Golden Gate assay by Illumina (384 SNP chip), and SNPlex Technology by Applied Biosystems (as described in Pindo et al., 2008) testing 9 SNPset of 48 SNPs/each. The SNPs were identified during an early assembly draft (4× sequencing depth) of the ‘Golden Delicious’ genome sequencing project, as described in detail in Micheletti et al. (2011). RNA isolation and transcription analysis Total RNA was isolated from fruit collected at harvest of four cultivars, ‘Golden Delicious’ (used as reference), ‘Delearly’, ‘Fuji’, and ‘Pink Lady’ using the Plant total RNA kit (Sigma) with a modified protocol. Total RNA was initially purified from gDNA with a Deoxyribonuclease I, Amplification Grade (Invitrogen®), then complementary DNA was synthesized by SuperScript® VILO cDNA Synthesis kit. In order to choose the best housekeeping gene, six different genes were tested: ubiquitin, two actin genes, cortex genes: 8283:1:a, 4592:1:a, and Efα1 (primer sequences available in Li and Yuan, 2008; Dal Cin et al., 2009; Botton et al., 2011). RT PCR was performed using an Applied Biosystems (AB) 7000 Sequence Detection System machine in a final volume of 12.5 μl containing 1 μl of properly diluted cDNA, 6.25 μl of Platinum SYBR green qPCR Super-Mix_UDG, 0.25 μl of Rox Reference Dye, and 0.2 μM of forward and reverse primers. Amplification conditions were 50 °C for 2 min, 95 °C for 2 min, and 40 cycles of denaturation at 95 °C for 15 s, annealing, and extension at 60 °C for 1 min. The best housekeeping genes (8283:1:a and 4592:1:a) were selected by geNorm software (Vandesompele et al., 2002), while real-time data were analysed by using LinRegPCR v12.9.0. The transcript levels of 11 specific genes: Md-PG1, Md-PG5, Md-NOR, Md-RIN, Md-Pel, Md-ACS1, Md-ACO1, Md-Exp, Md-XET, Md-XXT, and Md-XEG was investigated. Besides Md-PG1, where primers were obtained from Costa et al. (2010b), the primer sequences for the other 10 genes were retrieved from the apple genome contigs using Primer3 software (http://primer3.sourceforge.net/). Designed primers were then tested by performing an electronic PCR (http://www.ncbi.nlm.nih.gov/projects/e-pcr/) with the apple in silico predicted transcriptome (see Supplementary Table S3 at JXB online). Ethylene analysis Ethylene production was monitored over post-harvest ripening for a period of 15 d. Starting from harvest, ethylene was measured at day 1, 6, 10, and 15. For each measurement data point, three fruit/cultivars were closed in a sealed glass of 5.0 l for 1 h. From the headspace, 10 ml of air were sampled with a syringe and injected in a gas chromatographer (Carlo Erba Instruments GC6000) equipped with a flame ionization detector (FID). Statistical computation and data analysis For both progenies separated parental and combined genetic maps were constructed using JoinMap®4.0 (Van Ooijen, 2006). A LOD of 5 and a recombination frequency of 0.45 were used in order to define linkage groups, and genetic distances between markers were calculated using the Kosambi mapping function. Linkage groups were visualized using MapChart 2.1 (Voorips, 2002) and numbered from 1 to 17 (according to Malieppard et al., 1998; Silfverberg-Dilworth et al., 2006; www.hidras.unimi.it). QTL analysis carried out to detect genomic regions associated with textural components was performed using MapQTL®6 (Van Ooijen, 2009). Initially, the genomic regions with potential QTL effects were identified employing the Interval Mapping (IM) algorithm. To take over the role of other QTLs and to minimize the residual variance, markers coincident with the highest LOD value were selected as cofactors and further implemented in the MQM computation. The threshold established after running 1000 permutation was set at LOD=3, defined as the average of the type I error α=0.05 for all parameters across the 17 linkage groups for both mapping populations. QTL-LOD profiles are shown in a heatmap produced by Harry Plotter software, a stand-alone program written in Java. Harry Plotter has been developed internally at FEM as a new tool to visualize genome and genetic map features. Multivariate statistical Principal Component Analysis (PCA) was computed with STATISTICA software v7. Microsatellite motives were identified in the ‘Golden Delicious’ genomic assembled contigs through Sputnik, an algorithm for searching repeated nucleotide pattern (http://espressosoftware.com/sputnik/index.html). ‘Golden Delicious’ information related to gene ID and contigs is available at the FEM data warehouse (http://genomics.research.iasma.it) and GDR (Genomic Database for Rosaceae; http://rosaceae.org). Genes were annotated through the Uniprot database (http://www.uniprot.org) and further compared with the information available at the GDR database. Results and discussion Experimental design Fruit from the two progenies were assessed after 2 months of cold storage, in order to study the contribution of the two different genetic backgrounds in the fruit texture control. Moreover, from previous reports it has been emphasized that, after 2 months of cold storage, the physiological ripening evolution is maximized (Kouassi et al., 2009) enhancing therefore the trait variability, a fundamental requirement for the QTL intervals resolution. For the three years of phenotypic assessments all the parameters investigated showed a Pearson correlation from 0.57 to 0.9, resulting in a consistent quantitative trait distribution (data not shown). Texture physiology dissection, and combined acoustic-mechanical profiling Fruit from both progenies were harvested at the same physiological ripening stage, objectively determined by a vis/NIR portable spectrometer (DA-meter). The complex texture phenotype was assessed with the texture analyser TA-XTplus which produced a different texture profile for the three parental cultivars as well as for the two progenies (Fig. 1). ‘Fuji’ and ‘Pink Lady’ greatly differ from ‘Delearly’ in terms of general texture performance, as demonstrated by the different mean value for the maximum force: 11.95, 13.30, and 6.14 N; number of force peak: 22.23, 22.05, and 6.11; number of acoustic peaks: 107.23, 71, and 5.83 and mean acoustic pressure: 49.7, 45.7, and 39.65 dB, respectively. Fig. 1. View largeDownload slide Combined acoustic-mechanical texture profiles. The nine panels represent: (a) ‘Delearly’; (b) ‘Pink Lady’; (c) ‘Fuji’; (d, e, f) three seedlings of the ‘Fj×Del’ population; (g, h, i) three seedlings of the ‘Fj×PL’ population. For each graph the black line represents the mechanical force displacement profile scaled on the y primary axis (Newton), while the grey line is the acoustic profile scaled on the y secondary axis (dB). On the x-axis is shown the 90% deformation (strain). Fig. 1. View largeDownload slide Combined acoustic-mechanical texture profiles. The nine panels represent: (a) ‘Delearly’; (b) ‘Pink Lady’; (c) ‘Fuji’; (d, e, f) three seedlings of the ‘Fj×Del’ population; (g, h, i) three seedlings of the ‘Fj×PL’ population. For each graph the black line represents the mechanical force displacement profile scaled on the y primary axis (Newton), while the grey line is the acoustic profile scaled on the y secondary axis (dB). On the x-axis is shown the 90% deformation (strain). The overall variability of the texture sub-traits represented by the set of 14 parameters for the two populations is shown in the two-dimensional PCA plot (Fig. 2A). The first principal component (PC1), describing 55.39% of the entire phenotypic variability, together with the second principal component (PC2), accounting for an additional 20.29%, discriminated the orientation of the mechanical component from the acoustic signature (Fig. 2B), confirming the results previously obtained from a large apple collection (Costa et al., 2011). The PCA plot for the two progenies showed a bimodal distribution and, in fact, the variability detected in ‘Fuji×Delearly’ (‘Fj×Del’) was mainly distributed along PC1, with the extreme values represented by the two parental cultivars. The scenario observed in ‘Fuji×Pink Lady’ (‘Fj×PL’) was different, where the transgressive distribution observed for this progeny was more oriented towards PC2, with the seedlings exceeding the values observed for the two parental cultivars. Fig. 2. View largeDownload slide (A) Principal component analysis plot showing the general texture variability of the two mapping populations explained by the first two components. In the figure ‘a’ and ‘b’ are for ‘Fj×Del’ and ‘Fj×PL’, respectively, while the three parental cultivars are indicated by full name. (B) PCA parameter projection with ‘a’ and ‘b’ indicating the mechanical and acoustic parameters respectively, while ‘c’ is for the parameters related to the force direction. Fig. 2. View largeDownload slide (A) Principal component analysis plot showing the general texture variability of the two mapping populations explained by the first two components. In the figure ‘a’ and ‘b’ are for ‘Fj×Del’ and ‘Fj×PL’, respectively, while the three parental cultivars are indicated by full name. (B) PCA parameter projection with ‘a’ and ‘b’ indicating the mechanical and acoustic parameters respectively, while ‘c’ is for the parameters related to the force direction. Genetic mapping To unravel the highly co-ordinated cell wall physiology leading to the textural properties of apple during post-harvest ripening, two genetic maps were de novo developed and assembled. To construct the map scaffolds 734 markers were totally positioned (671 SNPs and 63SSRs; see Supplementary Tables S1 and Supplementary Data at JXB online). The SNP transferability between the reference cv. ‘Golden Delicious’ and the parental varieties was of 38.97%, 50.38% and 50.95% for ‘Fuji’, ‘Pink Lady’, and ‘Delarly’, respectively (in agreement with Micheletti et al., 2011). The saturated linkage maps were anchored with SNPs (newly identified during the assemblage of the early apple genome) to the reference published maps with a set of 38 SSRs (Liebhard et al., 2002; Silfverberg-Dilworth et al., 2006) out of which 30 were assembled in novel triplex. Multiplex have been already published so far (Patocchi et al., 2009) requiring, however, several optimizations. The advantage of the set presented here is the common amplification protocol, which allowed a more standardized and efficient mapping. In both maps, new microsatellite markers were also positioned within contigs where specific functional genes were identified. Genes related to ethylene synthesis/perception, cell wall metabolism, and transcription factors were selected from a heterologous microarray experiment designed to highlight a common set of differentially expressed genes during the ripening of both apple and tomato (Alba et al., 2005; Costa et al., 2010a). The genetic maps for both populations were initially computed for single parental meiosis and then the assembled maps were further used for QTL mapping purposes. For the ‘Fj×Del’ population the 494 segregating markers were assembled in 17 linkage groups (see Supplementary Fig. S1 at JXB online) with a final length of 1053.24 cM, and an averaged density of 2.28 cM between markers. In ‘Fj×PL’ 487 markers were assembled in a map with a length of 1470.8 cM (see Supplementary Fig. S2 at JXB online) and an averaged density of 3.25. To test the synteny between the two genetic maps and the apple genome, the genetic position of the 16 SSR anchored to candidate genes was compared with their physical position on the apple genome (see Supplementary Fig. S3 at JXB online). The SSR-anchored to candidate genes were located on 9 chromosomes (2, 5, 8, 9, 10, 13, 14, 15, and 16), showing a consistent position beside three elements (Md-ACS1 and two MYB TFs). This inconsistency can be assigned to a low sequencing coverage of these genomic regions, leading to an erroneous genome assembling. QTL detection and candidate gene mapping Mechanical and acoustic data were used in a QTL analysis which allowed the identification of 22 total QTLs associated to texture sub-traits, 12 mapped in ‘Fj×Del’ (Fig. 3) and 10 in ‘Fj×PL’ (Fig. 4), respectively. QTL computed with the IM algorithm and further validated by MQM reported a LOD values ranging from 3.11 to 10.86 and accounting for an explained variance from 10% up to 49%. Fig. 3. View largeDownload slide QTL-ATLAS for ‘Fuji×Delearly’ representing the heat-map QTL profiles detected in nine linkage groups. Numbers above each heat-map bar refers to texture parameters (described in the box on the left side), while ‘a’ and ‘b’ are for Interval Mapping and the MQM algorithm, respectively. The markers linked to candidate genes are indicated in red text. The Heatmap colour scale, going from black (LOD=0) to white (LOD ≥8) is shown in the box on the right. The scale on the right side of each heat map cluster shows the linkage distance in cM. Fig. 3. View largeDownload slide QTL-ATLAS for ‘Fuji×Delearly’ representing the heat-map QTL profiles detected in nine linkage groups. Numbers above each heat-map bar refers to texture parameters (described in the box on the left side), while ‘a’ and ‘b’ are for Interval Mapping and the MQM algorithm, respectively. The markers linked to candidate genes are indicated in red text. The Heatmap colour scale, going from black (LOD=0) to white (LOD ≥8) is shown in the box on the right. The scale on the right side of each heat map cluster shows the linkage distance in cM. Fig. 4. View largeDownload slide QTL-ATLAS representing the heat-map QTL profile for ‘Fuji×Pink Lady’. Letters and numbers are the same as for ‘Fuji×Delearly’ (Fig. 3). Markers linked to candidate genes are shown in red. The scale on the right side of each heat map cluster shows the linkage distance in cM. Fig. 4. View largeDownload slide QTL-ATLAS representing the heat-map QTL profile for ‘Fuji×Pink Lady’. Letters and numbers are the same as for ‘Fuji×Delearly’ (Fig. 3). Markers linked to candidate genes are shown in red. The scale on the right side of each heat map cluster shows the linkage distance in cM. QTLs were detected based on a LOD threshold of 3, obtained after averaging the value corresponding to α=0.05 for all traits over the 17 LGs (see Supplementary Fig. S4 at JXB online). The QTL mapping investigation carried out in the ‘Fj×Del’ progeny enabled the detection of a QTL set related to texture components located on nine linkage groups (1, 5, 6, 10, 12, 14, 15, 16, and 17, Fig. 3). From the general QTL-ATLAS it has emerged that LG10 represents a hot spot for texture control, most of the QTLs computed for the three years being associated with the texture parameters clustered on this chromosome (beside Δforce), with LOD values spanning from 4.35 to 10.97, and explaining from 20% to 49% of the total phenotypic variance (Table 1). QTLs identified by IM were, for the most part, also confirmed by the MQM algorithm, exceptions being made for force linear distance, mean force, and Young’s modulus (in 2009). This QTL cluster co-located with the Md-PG1 gene (mapped in this study as Md-PG1SSR, a new microsatellite targeted approximately at 3 kb upstream of the starting codon; see Supplementary Table S1 at JXB online), an ethylene-dependent candidate cell wall gene (Costa et al., 2010a, b). Table 1. QTL-ATLAS for ‘Fj×Del’ population Fj×Del      2008      2009      2010      2008      2009      2010            IM      IM      IM      MQM      MQM      MQM      Trait  LG  Marker  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  Initial force  5  GDsnp02834                          5.5  5.48  18.7          10  GDsnp02183        7.66  7.66  34.3  6.02  6.02  41.3                      10  GDsnp01764                    4.37  4.16  18.7                10  Md-PG1-SSR  4.37  4.01  23.8  7.66  6  28  6.05  4.55  33.2  4.37  4.01  23.8  6.37  6  28  4.99  4.55  33.2    12  GDsnp00747                          4.13  4  14.2        Max force  5  GDsnp02834                          5.97  5.97  18.8          10  GDsnp02183  4.23  4.12  24.3  9.25  9.25  39.8  6.39  6.39  43.2                      10  GDsnp01764                    4.2  4.03  18.3                10  Md-PG1-SSR  4.23  3.96  23.5  9.25  7.22  32.7  6.39  4.7  34  4.2  3.96  23.5  7.22  7.22  32.7  5.03  4.7  34    12  GDsnp01984                          4.51  4.51  14.7        Final force  5  GDsnp02834                          5.99  5.97  18.8          10  GDsnp02183        10.86  10.86  44.9  6.9  6.9  45.7                      10  Md-PG1-SSR  3.85  3.69  22.1  10.86  8.98  38.9  6.9  5.32  37.6  3.85  3.69  22.1  9.2  8.98  38.9  5.55  5.32  37.6    12  GDsnp01984                          4.51  4.51  14.7        Mean force  5  GDsnp02834                          5.14  5.12  18.8          10  GDsnp02183        10.01  10.01  42.2  6.76  6.76  45                      10  Md-PG1-SSR  4.28  4.03  23.9  10.01  8.03  35.6  6.76  51.4  36.6  4.28  4.03  23.9        5.55  5.14  36.6    12  GDsnp01984                          4.09  4.09  12.9        Area  5  GDsnp02834                          5.13  5.13  15.9          10  GDsnp02183  3.9  3.7  22.1  9.91  9.91  41.9  6.77  6.77  44.5                      10  GDsnp01764                    3.9  3.59  16.9                10  Md-PG1-SSR  3.9  3.56  21.4  9.91  7.93  35.3  6.77  6.72  36.1  3.9  3.56  21.4  8.31  8.31  35.3  5.47  5.06  36.1    12  GDsnp01984                          3.97  3.97  12.7        Force linear  5  GDsnp02834                                      distance  10  GDsnp02183        8.76  8.76  38.1  6.51  6.51  43.8        8.76  8.76  38.1          10  Md-PG1-SSR  7.86  7.86  41.7  8.76  6.85  31.3  6.51  4.41  32.3  7.86  7.86  41.7        4.59  4.41  32.3    16  GDsnp00200                          4.63  4.63  13.9        Young's modulus  10  GDsnp00355        4.35  4.35  20.2                            10  Md-PG1-SSR        4.35  3.61  18                            10  GDsnp02183              3.75  3.75  28.3                      14  Md-Le-MADS  4.41  4.41  25.8  3.74  3.74  18.5  3.75  2.39  19.1  4.41  4.41  25.8                14  Md-Pel              3.75  2.4  19.2        3.11  2.86  10          15  GDsnp02823        4.82  4.82  23.2              3.86  3.86  13.2          17  GDsnp01098        3.74                  3.74  3.74  12.8        No force peaks  1  Hi02b10                          4.02  3.96  11.9          10  GDsnp02183        9.05  9.32  39.1  6.61  6.61  44.3        9.32  9.05  39.1          10  Md-PG1-SSR        9.05  8.87  38.5  6.61  5.67  39.5              5.67  5.67  39.5    10  GDsnp1961  4.47  4.47  26.1              4.47  4.47  26.1                15  GDsnp02823                          5.54  5.54  16          15  Md-ACS_1                          5.54  4.94  14.5          15  GDsnp01687                                5.67  4.93  21.4  Δforce  6  GDsnp00166        4.03  4.04  15.6              4.03  4.03  19.8          14  Md-Le-MADS  3.47  3.05  18.8  3.29  3.29  16.5  3.13  2.9  22.6                      14  GDsnp02021              3.13  3.13  24.2                      15  Md-ACS1  3.07  3.07  19.1  3.8  3.8  18.88                          Forve index  10  Md-PG1-SSR        4.98  4.94  23.7              4.98  4.94  23.7          14  Md-MADS/        3.54  3.55  17.7                              Md-Pel                                        15  Md-ACS_1                          4.17  4.16  15.6          17  GDsnp00328        4.88  4.88  23.5              4.64  4.53  16.8        Acoustic linear  1  GDsnp00631        3.71  3.73  18.4                          Distance  10  GDsnp02183  8.3  8.3  42.5  10.45  10.45  43.6  7.65  7.65  49.2                      10  Md-PG1-SSR  8.3  6.37  34.6  10.45  8.97  38.8  7.65  5.83  40.3  6.36  6.36  34.6  8.97  9.87  38.8  5.83  5.83  40.3    15  GDsnp02823                          5.04  5.04  14.8          15  GDsnp01687                                5.83  4.61  20  No acoustic  1  GDsnp01027        3.98  3.98  20.7                          peaks  10  GDsnp02183  5.72  5.72  41.6  7.97  7.97  37.2  7.24  7.24  47.3  7.15  6.55  31.8                10  Md-PG1-SSR  5.72  3.93  30.9  7.97  6.41  31.2  7.24  5.1  36.4  7.15  3.93  30.9  6.41  6.41  12.17  5.34  5.34  24    15  GDsnp02823                          4.81  4.81  16.8  5.34  5.1  36.4  Max acoustic  1  GDsnp01470        3.77  3.77  22.3                          pressure  10  GDsnp02183  3.77  3.77  25.5  10.97  10.97  49  8.03  8.03  50.9  3.77  3.77  25.5                10  Md-PG1-SSR        10.97  9.79  45.2  8.03  5.92  40.8        9.85  9.79  45.2  5.92  5.92  40.8    15  GDsnp02823                          4.44  4.44  13.1        Mean acoustic  1  GDsnp01470        4.12  4.12  22.3                          pressure  10  GDsnp02183        8.3  8.3  39.9  5.76  5.76  40                      10  Md-PG1-SSR  2.89  2.89  20.2  8.3  7.91  38.5  5.76  4.64  33.7  2.89  2.89  20.2  5.31  5.31  21.6  5.76  4.64  33.7    16  GDsnp00200                          4.21  4.21  12.8        Fj×Del      2008      2009      2010      2008      2009      2010            IM      IM      IM      MQM      MQM      MQM      Trait  LG  Marker  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  Initial force  5  GDsnp02834                          5.5  5.48  18.7          10  GDsnp02183        7.66  7.66  34.3  6.02  6.02  41.3                      10  GDsnp01764                    4.37  4.16  18.7                10  Md-PG1-SSR  4.37  4.01  23.8  7.66  6  28  6.05  4.55  33.2  4.37  4.01  23.8  6.37  6  28  4.99  4.55  33.2    12  GDsnp00747                          4.13  4  14.2        Max force  5  GDsnp02834                          5.97  5.97  18.8          10  GDsnp02183  4.23  4.12  24.3  9.25  9.25  39.8  6.39  6.39  43.2                      10  GDsnp01764                    4.2  4.03  18.3                10  Md-PG1-SSR  4.23  3.96  23.5  9.25  7.22  32.7  6.39  4.7  34  4.2  3.96  23.5  7.22  7.22  32.7  5.03  4.7  34    12  GDsnp01984                          4.51  4.51  14.7        Final force  5  GDsnp02834                          5.99  5.97  18.8          10  GDsnp02183        10.86  10.86  44.9  6.9  6.9  45.7                      10  Md-PG1-SSR  3.85  3.69  22.1  10.86  8.98  38.9  6.9  5.32  37.6  3.85  3.69  22.1  9.2  8.98  38.9  5.55  5.32  37.6    12  GDsnp01984                          4.51  4.51  14.7        Mean force  5  GDsnp02834                          5.14  5.12  18.8          10  GDsnp02183        10.01  10.01  42.2  6.76  6.76  45                      10  Md-PG1-SSR  4.28  4.03  23.9  10.01  8.03  35.6  6.76  51.4  36.6  4.28  4.03  23.9        5.55  5.14  36.6    12  GDsnp01984                          4.09  4.09  12.9        Area  5  GDsnp02834                          5.13  5.13  15.9          10  GDsnp02183  3.9  3.7  22.1  9.91  9.91  41.9  6.77  6.77  44.5                      10  GDsnp01764                    3.9  3.59  16.9                10  Md-PG1-SSR  3.9  3.56  21.4  9.91  7.93  35.3  6.77  6.72  36.1  3.9  3.56  21.4  8.31  8.31  35.3  5.47  5.06  36.1    12  GDsnp01984                          3.97  3.97  12.7        Force linear  5  GDsnp02834                                      distance  10  GDsnp02183        8.76  8.76  38.1  6.51  6.51  43.8        8.76  8.76  38.1          10  Md-PG1-SSR  7.86  7.86  41.7  8.76  6.85  31.3  6.51  4.41  32.3  7.86  7.86  41.7        4.59  4.41  32.3    16  GDsnp00200                          4.63  4.63  13.9        Young's modulus  10  GDsnp00355        4.35  4.35  20.2                            10  Md-PG1-SSR        4.35  3.61  18                            10  GDsnp02183              3.75  3.75  28.3                      14  Md-Le-MADS  4.41  4.41  25.8  3.74  3.74  18.5  3.75  2.39  19.1  4.41  4.41  25.8                14  Md-Pel              3.75  2.4  19.2        3.11  2.86  10          15  GDsnp02823        4.82  4.82  23.2              3.86  3.86  13.2          17  GDsnp01098        3.74                  3.74  3.74  12.8        No force peaks  1  Hi02b10                          4.02  3.96  11.9          10  GDsnp02183        9.05  9.32  39.1  6.61  6.61  44.3        9.32  9.05  39.1          10  Md-PG1-SSR        9.05  8.87  38.5  6.61  5.67  39.5              5.67  5.67  39.5    10  GDsnp1961  4.47  4.47  26.1              4.47  4.47  26.1                15  GDsnp02823                          5.54  5.54  16          15  Md-ACS_1                          5.54  4.94  14.5          15  GDsnp01687                                5.67  4.93  21.4  Δforce  6  GDsnp00166        4.03  4.04  15.6              4.03  4.03  19.8          14  Md-Le-MADS  3.47  3.05  18.8  3.29  3.29  16.5  3.13  2.9  22.6                      14  GDsnp02021              3.13  3.13  24.2                      15  Md-ACS1  3.07  3.07  19.1  3.8  3.8  18.88                          Forve index  10  Md-PG1-SSR        4.98  4.94  23.7              4.98  4.94  23.7          14  Md-MADS/        3.54  3.55  17.7                              Md-Pel                                        15  Md-ACS_1                          4.17  4.16  15.6          17  GDsnp00328        4.88  4.88  23.5              4.64  4.53  16.8        Acoustic linear  1  GDsnp00631        3.71  3.73  18.4                          Distance  10  GDsnp02183  8.3  8.3  42.5  10.45  10.45  43.6  7.65  7.65  49.2                      10  Md-PG1-SSR  8.3  6.37  34.6  10.45  8.97  38.8  7.65  5.83  40.3  6.36  6.36  34.6  8.97  9.87  38.8  5.83  5.83  40.3    15  GDsnp02823                          5.04  5.04  14.8          15  GDsnp01687                                5.83  4.61  20  No acoustic  1  GDsnp01027        3.98  3.98  20.7                          peaks  10  GDsnp02183  5.72  5.72  41.6  7.97  7.97  37.2  7.24  7.24  47.3  7.15  6.55  31.8                10  Md-PG1-SSR  5.72  3.93  30.9  7.97  6.41  31.2  7.24  5.1  36.4  7.15  3.93  30.9  6.41  6.41  12.17  5.34  5.34  24    15  GDsnp02823                          4.81  4.81  16.8  5.34  5.1  36.4  Max acoustic  1  GDsnp01470        3.77  3.77  22.3                          pressure  10  GDsnp02183  3.77  3.77  25.5  10.97  10.97  49  8.03  8.03  50.9  3.77  3.77  25.5                10  Md-PG1-SSR        10.97  9.79  45.2  8.03  5.92  40.8        9.85  9.79  45.2  5.92  5.92  40.8    15  GDsnp02823                          4.44  4.44  13.1        Mean acoustic  1  GDsnp01470        4.12  4.12  22.3                          pressure  10  GDsnp02183        8.3  8.3  39.9  5.76  5.76  40                      10  Md-PG1-SSR  2.89  2.89  20.2  8.3  7.91  38.5  5.76  4.64  33.7  2.89  2.89  20.2  5.31  5.31  21.6  5.76  4.64  33.7    16  GDsnp00200                          4.21  4.21  12.8        In the table are reported the traits, the linkage group on which the QTLs have been identified, the marker and the candidate gene closest to the LOD peak, the maximum LOD for the QTL, the LOD corresponding to the marker/candidate gene, and the % of explained variability calculated with both IM and MQM. View Large The only parameter not associated with LG10 was Δ force, which, instead, was mapped on linkage group 6, 14, and 15. QTLs for force index were mapped on linkage groups 14, 15, and 10, but, in the latter case, it showed the lower statistical value. This observation is consistent with the projection of the parameters shown in the PCA (Fig. 2B) where these two parameters are differentially oriented, suggesting therefore a different genetic control. On linkage group 14, two candidate genes were mapped in the QTL intervals associated with the two force directional parameters, Md-RIN (an orthologue of the tomato MADS-RIN transcription factor, Vrebalov et al., 2002; Seymour et al., 2011) and the cell wall enzyme pectate lyase (Md-Pel), a gene known to be specifically over-expressed during ripening and involved in the fruit softening process, as documented for strawberry (Jiménez-Bermùdez et al., 2002; Marín-Rodríguez et al., 2003). Md-RIN was interestingly associated with Young’s modulus (or elasticity module) for the three years of assessments in ‘Fj×Del’. This result suggests that this transcription factor also plays an important physiological role for apple as it influences the rigidity of the cell wall. From the functional assessment carried out on the three parental cultivars (plus ‘Golden Delicious’ used as the reference), a different transcript accumulation was observed for Md-Pel, which was more abundant in ‘Delearly’ with respect to the other cultivars (Fig. 5). Fig. 5. View largeDownload slide Ethylene dynamics and real-time qPCR for three parental cultivars and ‘Golden Delicious’ (reference). Ethylene dynamics is expressed in μl kg−1 h−1 plotted on the y-axes, while on the x-axes the days of measurements are reported. The grey line is for ‘Delearly’, the black line is for ‘Golden Delicious’, the black dotted line is for ‘Fuji’, and the black dashed line is for ‘Pink Lady’. Real-time histograms show the expression analysis of 11 genes: Md-PG1, Md- PG5, Md-ACS1, Md-ACO1, Md-Exp, Md-Pel, Md-RIN, Md-NOR, Md-XET, Md-XXT, and Md-XEG. ‘Delearly’ is visualized in dark grey, ‘Golden Delicious’ in black stripes, ‘Fuji’ in light grey, and ‘Pink Lady’ in white. Bars show the mean standard error. X-axes: cultivars (Del, ‘Delearly’; GD, ‘Golden Delicious’; Fj, ‘Fuji’; PL, ‘Pink Lady’). Y-axes: relative expression level. Housekeeping genes used as reference were 8283:1:a and 4592:1:a. Fig. 5. View largeDownload slide Ethylene dynamics and real-time qPCR for three parental cultivars and ‘Golden Delicious’ (reference). Ethylene dynamics is expressed in μl kg−1 h−1 plotted on the y-axes, while on the x-axes the days of measurements are reported. The grey line is for ‘Delearly’, the black line is for ‘Golden Delicious’, the black dotted line is for ‘Fuji’, and the black dashed line is for ‘Pink Lady’. Real-time histograms show the expression analysis of 11 genes: Md-PG1, Md- PG5, Md-ACS1, Md-ACO1, Md-Exp, Md-Pel, Md-RIN, Md-NOR, Md-XET, Md-XXT, and Md-XEG. ‘Delearly’ is visualized in dark grey, ‘Golden Delicious’ in black stripes, ‘Fuji’ in light grey, and ‘Pink Lady’ in white. Bars show the mean standard error. X-axes: cultivars (Del, ‘Delearly’; GD, ‘Golden Delicious’; Fj, ‘Fuji’; PL, ‘Pink Lady’). Y-axes: relative expression level. Housekeeping genes used as reference were 8283:1:a and 4592:1:a. This functional pattern was consistent with the ethylene dynamics observed for this set of cultivars. ‘Delearly’, in fact, at harvest, was already in the climacterium phase (producing 34.05 μl kg−1 h−1 of ethylene). ‘Golden Delicious’ is also a cultivar known to produce a large amount of ethylene, but its burst was detected at the end of the observation period, and so, at harvest (0.18 μl kg−1 h−1 of ethylene), the expression of Md-Pel is still at the minimum level. As additional proof of the tight ethylene-dependent expression of Md-Pel, in ‘Fuji’ and ‘Pink Lady’, the transcript accumulation of this gene was at the minimum level, consistent with the low ethylene levels detected (1.36 and 10.82 μl kg−1 h−1, respectively). The main gene involved in the ethylene pathway, Md-ACS1, was mapped on LG15 as already reported by Costa et al. (2005). For this gene, as well as for Md-ACO1, the transcript accumulation was consistent with the different ethylene accumulation measured amongst the cultivars. ‘Delearly’, which presented the faster ripening physiology and the highest ethylene production, showed the highest expression for these two genes. Traits such as Young’s modulus and acoustic-related parameters were associated with QTL on linkage group 15 with a LOD range between 3.8 and 5.5 (13.10–16.8% of explained variance), however, the last passed the threshold only after MQM computation, hypothesizing possible epistatic effects. For force index, another QTL was also identified (LOD of 4.88 and 16.8% of variance) located on linkage group 17, further confirmed by MQM. On this linkage group, MQM evidence was also found for the presence of a QTL for the Young’s modulus trait, in the proximity of a candidate gene mapped and related to a WIZZ transcription factor. A specific set of parameters particularly related to the mechanical displacement profile were significantly associated with linkage group 5, with a LOD value ranging from 5.14 to 8.76 (15.9–38.10% of expressed variance). The QTL cluster associated with these traits (identified only after MQM) closely mapped to the Md-NOR gene, a NAC transcription factor orthologue of nor (non-ripening), a tomato mutant that fails to ripe normally (Giovannoni, 2004). Among the cultivars tested Md-RIN (mapped on LG14) and Md-NOR (mapped on LG5) did not show any differential expression. The similar functional pattern of these two genes suggests that the different ripening physiology observed among the varieties depends mainly on the gene acting in the final steps of the ripening pathway. For this class of parameters a possible second QTL was located further in the chromosome, but with a lower significant value. This second QTL was also closely mapped with two other candidate genes, a MYB transcription factor, and the homoeologous PG gene of Md-PG1 (here named Md-PG5 and sharing 86% of sequence identity with Md-PG1). The different association observed for these two genes can be explained by the duplicated nature of the apple genome, which has led to the formation of several chromosome doublets, of which 5–10 is one of the most evident cases of chromosome homology. It is worth noting that Md-PG5 is much less associated with fruit texture than Md-PG1, suggesting that, during the genome duplication event, this gene has lost its functionality (Wagner, 1998; Sanzol, 2010), as also suggested by the qPCR assessment. The comparison of the expression related to these two genes showed a clear different pattern. Md-PG1 is in fact over-expressed in ‘Delearly’, consistent with the texture performance and ethylene production observed in this study at harvest time (Fig. 5). No significant functional variation was instead observed among the cultivars for Md-PG5. The same set of parameters (except the force linear distance) was also associated with QTLs identified on LG12, with a LOD range of 3.97–4.51 and expressed variance between 12.7% and 14.2%. By MQM computation, a QTL was mapped on LG16 related to the mean acoustic pressure with a LOD of 4.63 and 13.9% of variance. A different scenario was thereafter offered by the second population ‘Fj×PL’ that, with respect to ‘Fj×Del’, showed a transgressive segregation along the PC2 which explained 20% of the total texture variability. According to the different variable orientation showed by the PCA (Fig. 2B), the phenotypic variability observed in ‘Fj×PL’ determined the identification of a smaller number of chromosomes associated with textural parameters. In fact, in ‘Fj×PL’ the QTLs were detected on a total number of seven linkage groups (1, 3, 7, 8, 11, 12, and 16; Table 2). The main difference observed between the two QTL-ATLAS profiles is that on ‘Fj×PL’ (Fig. 4) most of the QTLs were mapped on LG16 rather than LG10, where no QTL was detected in this case. For both algorithms, QTLs associated with all the mechanical parameters, plus only two acoustic parameters (max and mean acoustic pressure), with a LOD interval between 3.32 and 6.68 and explaining up to 34.6% of the total variance, were mainly detected on linkage group 16. All the textural parameters related to the acoustic signature, together with the number of force peaks (that, according to Costa et al., 2011 is more correlated with the acoustic phenomena), were associated with a QTL identified on LG12 (LOD of 3.11–4.48 and 13.6–24.4% of variance). Table 2. QTL summary for the ‘Fj×PL’ population Fj×PL      2008            2009                  IM      IM      MQM      MQM      Trait  LG  Marker  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  Initial force  16  GDsnp00444        5.29  5.25  31.1        5.29  5.25  31.1    16  GDsnp00578  4.28  3.8  23.3        4.28  3.8  23.3        Max force  16  GDsnp01866        5.61  5.54  32.5                16  GDsnp00071                    5.61  5.07  30.2    16  GDsnp01244  4.17  3.79  23.2        4.17  3.79  23.2        Final force  8  GDsnp01598                    3.75  3.41  14.5    11  GDsnp01640        3.49  3.49  21.9                16  GDsnp00444        5.68  5.58  32.6        6.68  5.58  32.6    16  GDsnp01244  3.58  3.36  20.9        3.58  3.36  20.9        Mean force  11  GCsnp01640        3.37  3.37  21.2                16  GDsnp01866        6.01  5.92  34.3        6.01  5.92  34.3    16  GDsnp00071  4.26  4.23  25.2        4.26  4.23  25.2        Area  16  GDsnp01866  3.82  3.34  20.8  6.04  5.95  34.4        5.98  5.95  34.4    16  GDsnp01329              3.77  3.77  18.3        Force linear distance  11  GDsnp01640        3.31  3.26  20.6                16  GDsnp00444        6.26  6  34.6        6.26  6  34.6    16  GDsnp01866  3.88  3.88  24.1        3.88  3.88  24.1        Young's modulus  3  GDsnp01982        3.78  3.78  23.5        3.78  3.78  23.5    3  GDsnp01969  4.62  4.62  27.5        4.62  4.62  27.5        No force peaks  12  GDsnp00747        3.3  2.45  15.9        4.67  3.56  18.2    16  GDsnp00444  3.48  3.48  21.6        3.48  3.48  21.6        Δforce  1  GDsnp01035                    3.96  3.96  18.4    7  GDsnp00247        4.01  4.01  24.7        4.01  4.01  24.7  Force index  3  GDsnp01937        3.4  3.37  21.3                7  GDsnp00247        3.3  3.3  20.9              Acoustic linear distance  12  CH04G04        3.77  3.75  23.4        3.77  3.75  23.3    16  GDsnp00444  3.23  3.23  20.2                    No acoustic peaks  12  CH04G04        4.48  3.95  24.4                12  GDsnp02537                    3.83  3.83  23.8    12  GDsnp00318              3.48  3.48  16.4          16  GDsnp00444  3.86  3.86  23.6        3.86  3.86  23.6        Max acoustic pressure  12  GDsnp00747                    3.27  2.81  13.6    16  GDsnp00578        3.94  3.1  19.7                16  GDsnp00444  2.65  2.65  16.9              3.94  3.94  24.4  Mean acoustic pressure  12  GDsnp00747        3.11  2.93  18.7        3.74  3.39  16.9    16  GDsnp00578        3.32  3.32  20.9        3.32  3.32  20.9  Fj×PL      2008            2009                  IM      IM      MQM      MQM      Trait  LG  Marker  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  LOD max  LOD marker  %  Initial force  16  GDsnp00444        5.29  5.25  31.1        5.29  5.25  31.1    16  GDsnp00578  4.28  3.8  23.3        4.28  3.8  23.3        Max force  16  GDsnp01866        5.61  5.54  32.5                16  GDsnp00071                    5.61  5.07  30.2    16  GDsnp01244  4.17  3.79  23.2        4.17  3.79  23.2        Final force  8  GDsnp01598                    3.75  3.41  14.5    11  GDsnp01640        3.49  3.49  21.9                16  GDsnp00444        5.68  5.58  32.6        6.68  5.58  32.6    16  GDsnp01244  3.58  3.36  20.9        3.58  3.36  20.9        Mean force  11  GCsnp01640        3.37  3.37  21.2                16  GDsnp01866        6.01  5.92  34.3        6.01  5.92  34.3    16  GDsnp00071  4.26  4.23  25.2        4.26  4.23  25.2        Area  16  GDsnp01866  3.82  3.34  20.8  6.04  5.95  34.4        5.98  5.95  34.4    16  GDsnp01329              3.77  3.77  18.3        Force linear distance  11  GDsnp01640        3.31  3.26  20.6                16  GDsnp00444        6.26  6  34.6        6.26  6  34.6    16  GDsnp01866  3.88  3.88  24.1        3.88  3.88  24.1        Young's modulus  3  GDsnp01982        3.78  3.78  23.5        3.78  3.78  23.5    3  GDsnp01969  4.62  4.62  27.5        4.62  4.62  27.5        No force peaks  12  GDsnp00747        3.3  2.45  15.9        4.67  3.56  18.2    16  GDsnp00444  3.48  3.48  21.6        3.48  3.48  21.6        Δforce  1  GDsnp01035                    3.96  3.96  18.4    7  GDsnp00247        4.01  4.01  24.7        4.01  4.01  24.7  Force index  3  GDsnp01937        3.4  3.37  21.3                7  GDsnp00247        3.3  3.3  20.9              Acoustic linear distance  12  CH04G04        3.77  3.75  23.4        3.77  3.75  23.3    16  GDsnp00444  3.23  3.23  20.2                    No acoustic peaks  12  CH04G04        4.48  3.95  24.4                12  GDsnp02537                    3.83  3.83  23.8    12  GDsnp00318              3.48  3.48  16.4          16  GDsnp00444  3.86  3.86  23.6        3.86  3.86  23.6        Max acoustic pressure  12  GDsnp00747                    3.27  2.81  13.6    16  GDsnp00578        3.94  3.1  19.7                16  GDsnp00444  2.65  2.65  16.9              3.94  3.94  24.4  Mean acoustic pressure  12  GDsnp00747        3.11  2.93  18.7        3.74  3.39  16.9    16  GDsnp00578        3.32  3.32  20.9        3.32  3.32  20.9  View Large A quite clear separation between the two components in ‘Fj×PL’ was observed in the data obtained in 2009 (LG16 more specific for mechanical and LG12 for acoustic parameters, respectively) and consistency was found with the variable projection showed in the PCA plot, where PC2 separates and distinguishes the mechanical from the acoustic components. QTL mapping on linkage group 10 was already shown in previous work focused on texture fruit quality (Maliepaard et al., 2001; Liebhard et al., 2003; Kenis et al., 2008), while QTLs mapped on linkage groups 12 and 16 were identified only in Liebhard et al. (2003) and Kenis et al. (2008). QTLs were simultaneously found for these three linkage groups in King et al. (2000, 2001), where additional texture sub-traits (sensorially evaluated) were considered. These results are fully consistent with our computations, not only for the common set of linkage groups identified as being associated with the QTLs related to texture, but also for the LOD magnitude. It is worth noting that, from the analysis of the estimated mean of the quantitative trait distribution associated with the genotypes of both populations, a specific genetic control of the texture phenotype resulted (see Supplementary Fig. S5 at JXB online). In the ‘Fj×Del’ population, the role of one allele of ‘Delearly’ in the control of the distribution of the texture sub-traits was clear (see Supplementary Fig. S5 at JXB online). For the second population, ‘Fj×PL’, a more distinct and specific effect regarding the two components seems to be present. For mechanical traits, one allele of ‘Fuji’ was responsible for most of the quantitative distribution, while acoustic parameters appears to be more controlled by a specific combination of two alleles, one from ‘Fuji’ and one from ‘Pink Lady’, respectively. From the analysis of the QTL profiles it is clear that fruit texture is one of the most quantitatively inherited traits in apple, due to a time-specific activation of a high number of genes. This comprehensive dissection of texture allowed the identification of the major regions controlling texture, which can be further exploited for advanced genetic and functional investigation in order to unravel the fruit texture machinery and to create new markers valuable for advanced breeding selection. The re-sequencing of the Md-XET gene (2276 bp) among the three parental cultivars enabled the identification of a set of 40 SNPs (see Supplementary Table S6 at JXB online). The segregation of this polymorphisms within the ‘Fj×PL’ progeny can be involved in the genetic control of the acoustic traits. These SNPs might be novel interesting markers useful for apple breeding, and further validation steps will be required in order to validate their possible real association with texture phenotype. Gene mining To date, there is a large body of evidence reporting that specific gene function can also be unravelled by QTL mapping approach (Moose and Mumm, 2008), emphasized by the co-location of candidate genes involved in biochemical pathways related to the trait of interest into the QTL intervals (Stevens et al., 2007). In this view, the apple genome was exploited for a post-genomic in silico positional discovery of the gene set underlying the QTL intervals identified here (see Supplementary Tables S4 and Supplementary Data at JXB online). For the two major LGs targeted in the two mapping populations (LG10 for ‘Fj×Del’ and LG16 for ‘Fj×PL’) significant gene categories have been annotated, identifying elements involved in cell wall metabolism and the ethylene synthesis/signalling pathway, as expected due to the tight relationship existing between the level of ethylene produced and the rate of cell wall modification in apple (Harada et al., 2000; Costa et al., 2005; Wakasa et al., 2006; Janssen et al., 2008). Beside these two classes, several transcription factors have also been identified within the QTL intervals, validating the results proposed by White (2002), Bartley and Ishida (2003), and Costa et al. (2010a) that cell wall metabolism is a physiological process controlled by a complex regulatory network. It is also interesting to note that, by comparing the gene inventory annotated among these three linkage groups, both common and specific gene families, were identified. Genes encoding for pectinesterases were commonly present in all the QTLs targeted to linkage groups 10, 12, and 16. Xyloglucan endotransglycosylase and polygalacturonase were mainly present on LG10 and 12. Cell wall enzymes such as pectate lyase, pectin acetylesterase, and xyloglucan xylosyltransferase were specifically identified on LG10, 12, and 16, respectively (see Supplementary Tables S4 and Supplementary Data at JXB online). A cluster of α expansin was only discovered on LG12. This class of genes are thought to cause a reversible disruption of the hydrogen bonding between cellulose microfibrils and the polysaccharide matrix, which may be one of the initial steps of the cell wall loosening process (Cosgrove, 2000). The functional activation and ethylene relationship of the expansin gene was also confirmed here. A gene of this cluster was, in fact, highly accumulated in ‘Delearly’ and ‘Golden Delicious’ (characterized by the high ethylene level and strong firmness decay during ripening) than ‘Fuji’ and ‘Pink Lady’ (showing low ethylene production and low softening). The early activation of this gene is clearly represented by ‘Golden Delicious’ which showed the maximum expression at harvest. All these cell-wall-modifying proteins, together with the regulation of water loss, are thought to influence the fruit-softening process (Brummell and Harpster, 2001; Saladié et al., 2007). Two major associated phenomena are depolymerization of the pectin network through a hydrolytic cleavage of homogalacturonan (a major component of the middle lamella) by polygalacturonase action, and the endocleavage of the hemicellulosic glycan matrices (of which xyloglucan is the most abundant in dicots) by the action of the hemicellulases (Campbell and Braam, 1999; Brummell, 2006). While still poorly understood, it seems that depolymerization of xyloglucan may act as a major contributor in the reduction of the cell wall turgidity (Rose and Bennett, 1999; Saladié et al., 2006). This theory finds consistency with the results showed in the QTL-ATLAS reported for the two populations (Figs 3, 4). In ‘Fj×Del’ progeny the high texture variability resulted in a major QTL cluster coinciding with the position of Md-PG1 gene. This gene, acting on the middle lamella, can be the causal factor determining the difference between the mealy and the high texture apple. However, this is not sufficient to discriminate mechanical and acoustic behaviour, so Md-PG1 is, in fact, only indistinctly associated with all the parameters defined and related to the different textural components. Sequencing of the genomic region harbouring Md-PG1SNP highlighted that ‘Pink Lady’ shares the same allelotype as ‘Fuji’ (T/T), while ‘Delearly’ showed a heterozygous state (G/T) as ‘Mondial Gala’, where the ‘G’ was described as the causal allele for the loss of firmness (Costa et al., 2010b). The homozygousity of the ‘T’ allele for both parents of the second population may be the reason that not a single QTL was identified on linkage group 10 in the ‘Fj×PL’ population. Two relevant QTLs were instead detected in this progeny, on LG16 and LG12, where a xyloglucan transglycosylase was annotated. In this scenario, the enzymatic action on the middle lamella could be less evident because of the absence of the Md-PG1SNP-G allele, enabling the possible action of Md-XET in reducing the rigidity of the cell wall. The hypothesis that solubilization and depolymerization of pectic and hemicellulosic polymers caused by these two enzymes may act in concert during the fruit softening process found consistency in the study of Hiwasa and collegues (2004) carried out on three different pear cultivars. The observed distinct physiology was associated with specific gene expression patterns. In cv. ‘La-France’, the transcript abundance of Pc-PG1 and Pc-PG2 was high and the expression of Pc-XET1 slightly increased during ripening, resulting in a remarkable softening and melting texture. In cv. ‘Nijisseki’, the gradual decrease in fruit firmness and mealy texture was attributed to the low and absent expression of Pc-PG2 and Pc-PG1, and to the constitutive expression of Pc-XET1. In cv. ‘Yali’, the small change in firmness and the crispy phenotype was explained by the observed lack of a detectable endo-PG activity and the constitutive expression of Pc-XET1. The effect of Md-XET in regulating texture change in fruit was also supported by the functional assessment that was carried out for xyloglucan endotransglycosilase (Md-XET), xyloglucan-6-xylosyltransferase (Md-XXT) and xyloglucan endoglucanase (Md-XEG). Amongst these, only Md-XET showed a distinct expression pattern, with its major transcript accumulation in ‘Fuji’ (crispy type of apple). The role of this gene may be the causal event of the acoustic segregation observed in the ‘Fj×PL’ population, detectable for the absence of Md-PG1 expression. The interplay between polygalacturonase and xyloglucan endotransglycosylase may regulate the crispy and mealy phenotype in apple (Brummell, 2006). Nowadays, crispness is a priority in worldwide breeding programmes, so functional markers based on these two candidate genes may be fundamental for a marker assisted breeding towards the selection of new superior apple cultivars. Conclusion Fruit texture is a feature composed of several sub-traits determined by the activity of different enzymes encoded by multi-gene families and regulated by a transcriptional network. This complex physiology and genetic control is shown, in this work, by the large number of significant QTLs targeted in the two mapping populations, co-located to several cell wall structural and regulatory genes. The use of a highly sophisticated texture analyser (TA-XTplus) allowed a comprehensive phenotypic dissection of the texture components. Thanks to this advanced phenomics tool, the highest number of texture QTLs reported up to now has been discovered and mapped, unravelling new genomic regions and genes with possible important effects in the control of texture. The choice of the two mapping populations characterized by a divergent but very distinct textural behaviour enabled the assessment of a wide textural variability. In ‘Fj×Del’, several QTLs have been identified, with the major ones located on chromosome 10 and, coincident with Md-PG1, indistinctly associated with all the textural parameters. The new SSR-anchored to this gene can be used as a valuable and efficient marker to improve fruit texture in apple. Moreover, in this experimental design, the use of the ‘FjxPL’ population was of great value because its transgressive segregation (essential for breeding) has shed light on other loci specifically involved in the control of either the mechanical or the acoustic components. The knowledge of the genetic loci controlling texture traits, together with the annotation of the genes underlying the QTL intervals, can give an insight into a better understanding of the complex physiology regulating the dynamics of fruit texture during apple ripening. The authors are grateful to Andrey Zharkikh (Myriad Genetics, UT, USA) and Michela Troggio (FEM) for the identification and selection of SNP regions, Chiara Ferrandi and Pietro Piffanelli (PTP, Italy) for SNP genotyping by the Golden Gate assay, and Massimo Pindo for his help with the SNPlex assay. The authors wish to thank Pierluigi Magnago and his team (FEM) for plant maintenance and Marco Fontanari (FEM) for his support in phenotyping. Finally, the authors thank Justin Lashbrooke for proofreading this manuscript. This work was supported by the Post-doc project CANDI-HAP granted by the Autonomous Province of Trento. References Alba R,  Payton P,  Fei ZJ,  McQuinn R,  Debbie P,  Martin GB,  Tanksley SD,  Giovannoni JJ.  Transcriptome and selected metabolite analyses reveal multiple points of ethylene control during tomato fruit development,  The Plant Cell ,  2005, vol.  17 (pg.  2954- 2965) Google Scholar CrossRef Search ADS PubMed  Andani Z,  Jaeger SR,  MacFie HJH.  Mealiness in apples: towards a multi-lingual consumer vocabulary,  Journal of Food Science ,  2001, vol.  66 (pg.  872- 878) Google Scholar CrossRef Search ADS   Asins MJ.  Present and future of quantitative trait locus analysis in plant breeding,  Plant Breeding ,  2002, vol.  121 (pg.  281- 291) Google Scholar CrossRef Search ADS   Bartley GE,  Ishida BK.  Developmental gene regulation during tomato fruit ripening and in-vitro sepal morphogenesis,  BMC Plant Biology ,  2003, vol.  3 pg.  4  Google Scholar CrossRef Search ADS PubMed  Botton A,  Eccher G,  Forcato C, et al.  Signaling pathways mediating the induction of apple fruitlet abscission,  Plant Physiology ,  2011, vol.  155 (pg.  185- 208) Google Scholar CrossRef Search ADS PubMed  Bourne MC. ,  Food texture and viscosity: concept and measurement ,  2002 2nd edn San Diego Academic Press Brummell DA.  Cell wall disassembly in ripening fruit,  Functional Plant Biology ,  2006, vol.  33 (pg.  103- 119) Google Scholar CrossRef Search ADS   Brummell DA,  Harpster MH.  Cell wall metabolism in fruit softening and quality and its manipulation in transgenic plants,  Plant Molecular Biology ,  2001, vol.  47 (pg.  311- 340) Google Scholar CrossRef Search ADS PubMed  Campbell P,  Braam J.  Xyloglucan endotransglycosylases: diversity of genes, enzymes and potential wall-modifying functions,  Trends in Plant Science ,  1999, vol.  4 (pg.  361- 366) Google Scholar CrossRef Search ADS PubMed  Cevik V,  Ryder CD,  Popovich A,  Manning K,  King GJ,  Seymour GB.  A FRUITFULL-like gene is associated with genetic variation for fruit flesh firmness in apple (Malus domestica Borkh.),  Tree Genetics and Genomes ,  2009, vol.  6 (pg.  271- 279) Google Scholar CrossRef Search ADS   Collard BCY,  Jahufer MZZ,  Brouwer JB,  Pang ECK.  An introduction to markers, quantitative trait loci (QTL. mapping and marker-assisted selection for crop improvement: the basic concepts,  Euphytica ,  2005, vol.  142 (pg.  169- 196) Google Scholar CrossRef Search ADS   Cosgrove DJ.  Loosening of plant cell walls by expansins,  Nature ,  2000, vol.  407 (pg.  321- 326) Google Scholar CrossRef Search ADS PubMed  Costa F,  Alba R,  Schouten H, et al.  Use of homologous and heterologous gene expression profiling tools to characterize transcription dynamics during apple fruit maturation and ripening,  BMC Plant Biology ,  2010, vol.  10 pg.  229  Google Scholar CrossRef Search ADS PubMed  Costa F,  Cappellin L,  Longhi S, et al.  Assessment of apple (Malus×domestica Borkh.) fruit texture by a combined acoustic-mechanical profiling strategy,  Postharvest Biology and Technology ,  2011, vol.  6 (pg.  21- 28) Google Scholar CrossRef Search ADS   Costa F,  Peace CP,  Stella S,  Serra S,  Musacchi S,  Bazzani M,  Sansavini S,  Van de Weg WE.  QTL dynamics for fruit firmness and softening around an ethylene-dependent polygalacturonase gene in apple (Malus×domestica Borkh.),  Journal of Experimental Botany ,  2010, vol.  61 (pg.  3029- 3039) Google Scholar CrossRef Search ADS PubMed  Costa F,  Stella S,  Van de Weg WE,  Guerra W,  Cecchinel M,  Dalla Via J,  Koller B,  Sansavini S.  Role of the genes Md-ACO1 and Md-ACS1 in ethylene production and shelf life of apple (Malus domestica Borkh),  Euphytica ,  2005, vol.  141 (pg.  181- 190) Google Scholar CrossRef Search ADS   Costa F,  Van de Weg WE,  Stella S,  Dondini L,  Pratesi D,  Musacchi S,  Sansavini S.  Map position and functional diversity of Md-Exp7, a new putative expansin gene associated with fruit softening in apple (Malus×domestica Borkh.) and pear (Pyrus communis),  Tree Genetics and Genomes ,  2008, vol.  4 (pg.  575- 586) Google Scholar CrossRef Search ADS   Costa G,  Noferini M,  Fiori G,  Torrigiani P.  Use of Vis/NIR spectroscopy to assess fruit ripening stage and improve management in post-harvest chain. GBS Journals, Fresh Produce 3,  Special Issue ,  2009, vol.  1 (pg.  35- 41) Dal Cin V,  Velasco R,  Ramina A.  Dominance induction of fruitlet shedding in Malus×domestica (L. Borkh): molecular changes associated with polar auxin transport,  BMC Plant Biology ,  2009, vol.  9 pg.  139  Google Scholar CrossRef Search ADS PubMed  Duizer L.  A review of acoustic research for studying the sensory perception of crisp, crunchy and crackly textures,  Trands in Food Science Technology ,  2001, vol.  12 (pg.  17- 24) Google Scholar CrossRef Search ADS   Dunemann F,  Ulrich D,  Malysheva-Otto L,  Weber WE,  Longhi S,  Velasco R,  Costa F.  Functional allelic diversity of the apple alcohol acyl-transferase gene MdAAT1 associated with fruit ester volatile contents in apple cultivars,  Molecular Breeding doi ,  2011  10.1007/s11032-011-9577-7 Fei Z,  Tang X,  Alba R,  Giovannoni J.  Tomato expression database (TED): a suite of data presentation and analysis tools,  Nucleic Acids Research ,  2006, vol.  34 (pg.  D766- D770) Google Scholar CrossRef Search ADS PubMed  Giovannoni J.  Molecular biology of fruit maturation and ripening,  Annual Review of Plant Physiology and Plant Molecular Biology ,  2001, vol.  52 (pg.  725- 749) Google Scholar CrossRef Search ADS PubMed  Giovannoni JJ.  Genetic regulation of fruit development and ripening,  The Plant Cell ,  2004, vol.  16 (pg.  S170- S180) Google Scholar CrossRef Search ADS PubMed  Hampson CR,  Quamme HA,  Hall JW,  MacDonald RA,  King MC,  Cliff MA.  Sensory evaluation as a selection tool in apple breeding,  Euphytica ,  2000, vol.  111 (pg.  79- 90) Google Scholar CrossRef Search ADS   Harada T,  Sunako T,  Wakasa Y,  Soejima J,  Satoh T,  Niizeki M.  An allele of the 1-aminocyclopropane-1-carboxylate synthase gene (Md-ACS1) accounts for the low level of ethylene production in climacteric fruits of some apple cultivars,  Theoretical and Applied Genetics ,  2000, vol.  101 (pg.  742- 746) Google Scholar CrossRef Search ADS   Harker FR,  Gunson FA,  Jaeger SR.  The case for fruit quality: an interpretive review of consumer attitudes, and preferences for apples,  Postharvest Biology and Technology ,  2003, vol.  28 (pg.  333- 347) Google Scholar CrossRef Search ADS   Hiwasa K,  Nakano R,  Hashimoto A,  Matsuzaki M,  Murayama H,  Inaba A,  Kubo Y.  European, Chinese, and Japanese pear fruits exhibit differential softening characteristics during ripening,  Journal of Experimental Botany ,  2004, vol.  55 (pg.  2281- 2290) Google Scholar CrossRef Search ADS PubMed  Hospital F.  Challenges for effective marker-assisted selection in plants,  Genetica ,  2009, vol.  136 (pg.  303- 310) Google Scholar CrossRef Search ADS PubMed  Janssen BJ,  Thodey K,  Schaffer RJ, et al.  Global gene expression analysis of apple fruit development from the floral bud to ripe fruit,  BMC Plant Biology ,  2008  8 Jiménez-Bermúdez S,  Redondo-Nevado J,  Muñoz-Blanco J,  Caballero JL,  López-Aranda JM,  Valpuesta V,  Pliego-Alfaro F,  Quesada MA,  Mercado JA.  Manipulation of strawberry fruit softening by antisense expression of a pectate lyase gene,  Plant Physiology ,  2002, vol.  128 (pg.  751- 759) Google Scholar CrossRef Search ADS PubMed  Kenis K,  Keulemans J,  Davey MW.  Identification and stability of QTLs for fruit quality traits in apple,  Tree Genetics and Genomes ,  2008, vol.  4 (pg.  647- 661) Google Scholar CrossRef Search ADS   Kilcast D.  Texture in food Vol. 2. Solid foods,  2004 Cambridge Woodhead Publishing Limited King GJ,  Maliepaard C,  Lynn JR, et al.  Quantitative genetic analysis and comparison of physical and sensory descriptors relating to fruit flesh firmness in apple (Malus pumila Mill.),  Theoretical and Applied Genetics ,  2000, vol.  100 (pg.  1074- 1084) Google Scholar CrossRef Search ADS   King GJ,  Lynn JR,  Dover CJ,  Evans KM,  Seymour GB.  Resolution of quantitative trait loci for mechanical measures accounting for genetic variation in fruit texture of apple (Malus pumila Mill.),  Theoretical and Applied Genetics ,  2001, vol.  102 (pg.  1227- 1235) Google Scholar CrossRef Search ADS   Kouassi AB,  Durel CE,  Costa F, et al.  Estimation of genetic parameters and prediction of breeding values for apple fruit-quality traits using pedigreed plant material in Europe,  Tree Genetics and Genomes ,  2009, vol.  5 (pg.  659- 672) Google Scholar CrossRef Search ADS   Li JG,  Yuan RC.  NAA and ethylene regulate expression of genes related to ethylene biosynthesis, perception, and cell wall degradation during fruit abscission and ripening in ‘Delicious’ apples,  Journal of Plant Growth Regulation ,  2008, vol.  27 (pg.  283- 295) Google Scholar CrossRef Search ADS   Liebhard R,  Gianfranceschi L,  Koller B,  Ryder CD,  Tarchini R,  Van de Weg E,  Gessler C.  Development and characterisation of 140 new microsatellites in apple (Malus×domestica Borkh.),  Molecular Breeding ,  2002, vol.  10 (pg.  217- 241) Google Scholar CrossRef Search ADS   Liebhard R,  Koller B,  Gianfranceschi L,  Gessler C.  Creating a saturated reference map for the apple (Malus×domestica Borkh.) genome,  Theoretical and Applied Genetics ,  2003, vol.  106 (pg.  1497- 1508) Google Scholar CrossRef Search ADS PubMed  Maliepaard C,  Alston FH,  van Arkel G, et al.  Aligning male and female linkage maps of apple (Malus pumila Mill.) using multi-allelic markers,  Theoretical and Applied Genetics ,  1998, vol.  97 (pg.  60- 73) Google Scholar CrossRef Search ADS   Maliepaard C,  Sillanpaa MJ,  van Ooijen JW,  Jansen RC,  Arjas E.  Bayesian versus frequentist analysis of multiple quantitative trait loci with an application to an outbred apple cross,  Theoretical and Applied Genetics ,  2001, vol.  103 (pg.  1243- 1253) Google Scholar CrossRef Search ADS   Marín-Rodriguez MC,  Smith DL,  Manning K,  Orchard L,  Seymour GB.  Pectate lyase gene expression and enzyme activity in ripening banana fruit,  Plant Molecular Biology ,  2003, vol.  51 (pg.  851- 857) Google Scholar CrossRef Search ADS PubMed  Matas AJ,  Gapper NE,  Chung MY,  Giovannoni JJ,  Rose JKC.  Biology and genetic engineering of fruit maturation for enhanced quality and shelf-life,  Current Opinion in Biotechnology ,  2009, vol.  20 (pg.  197- 203) Google Scholar CrossRef Search ADS PubMed  McCann M,  Rose J.  Blueprints for building plant cell walls,  Plant Physiology ,  2010, vol.  153 (pg.  365- 365) Google Scholar CrossRef Search ADS PubMed  Micheletti D,  Troggio M,  Zharkikh A,  Costa F,  Malnoy M,  Velasco R,  Salvi S.  Genetic diversity of the genus Malus and implications for linkage mapping,  Tree Genetics and Genomes ,  2011  doi:10.1007/s11295-011-0380-8 Moore S,  Payton P,  Wright M,  Tanksley S,  Giovannoni J.  Utilization of tomato microarrays for comparative gene expression analysis in the Solanaceae,  Journal of Experimental Botany ,  2005, vol.  56 (pg.  2885- 2895) Google Scholar CrossRef Search ADS PubMed  Moore S,  Vrebalov J,  Payton P,  Giovannoni J.  Use of genomics tools to isolate key ripening genes and analyse fruit maturation in tomato,  Journal of Experimental Botany ,  2002, vol.  53 (pg.  2023- 2030) Google Scholar CrossRef Search ADS PubMed  Moose SP,  Mumm RH.  Molecular plant breeding as the foundation for 21st century crop improvement,  Plant Physiology ,  2008, vol.  147 (pg.  969- 977) Google Scholar CrossRef Search ADS PubMed  Paterson AH.  Molecular dissection of complex traits. CRC Press LLC,  1998 Patocchi A,  Fernandez-Fernandez F,  Evans K, et al.  Development and test of 21 multiplex PCRs composed of SSRs spanning most of the apple genome,  Tree Genetics and Genomes ,  2009, vol.  5 (pg.  211- 223) Google Scholar CrossRef Search ADS   Pindo M,  Vezzulli S,  Coppola G,  Cartwright DA,  Zharkikh A,  Velasco R,  Troggio M.  SNP high-throughput screening in grapevine using the SNPlexTM genotyping system,  BMC Plant Biology ,  2008, vol.  8 pg.  12  Google Scholar CrossRef Search ADS PubMed  Rose JKC,  Bennett AB.  Cooperative disassembly of the cellulose-xyloglucan network of plant cell walls: parallels between cell expansion and fruit ripening,  Trends in Plant Science ,  1999, vol.  4 (pg.  176- 183) Google Scholar CrossRef Search ADS PubMed  Saladie M,  Matas AJ,  Isaacson T, et al.  A re-evaluation of the key factors that influence tomato fruit softening and integrity,  Plant Physiology ,  2007, vol.  144 (pg.  1012- 1028) Google Scholar CrossRef Search ADS PubMed  Saladié M,  Rose JKC,  Cosgrove DJ,  Catalá C.  Characterization of a new xyloglucan endotransglucosylase/hydrolase (XTH) from ripening tomato fruit and implications for the diverse modes of enzymic action,  The Plant Journal ,  2006, vol.  47 (pg.  282- 295) Google Scholar CrossRef Search ADS PubMed  Sanzol J.  Dating and functional characterization of duplicated genes in the apple (Malus domestica Borkh.) by analyzing EST data,  BMC Plant Biology ,  2010, vol.  10 pg.  87  Google Scholar CrossRef Search ADS PubMed  Schuelke M.  An economic method for the fluorescent labeling of PCR fragments,  Nature Biotechnology ,  2000, vol.  18 (pg.  233- 234) Google Scholar CrossRef Search ADS PubMed  Seymour GB,  Ryder CD,  Cevik C,  Hammond JP,  Popovich A,  King GJ,  Vrebalov J,  Giovannoni JJ,  Manning K.  A SEPALLATA gene is involved in the development and ripening of strawberry (Fragaria×ananassa Duch.) fruit, a non-climacteric tissue,  Journal of Experimental Botany ,  2011, vol.  62 (pg.  1179- 1188) Google Scholar CrossRef Search ADS PubMed  Silfverberg-Dilworth E,  Matasci CL,  van de Weg WE, et al.  Microsatellite markers spanning the apple (Malus×domestica Borkh.) genome,  Tree Genetics and Genomes ,  2006, vol.  2 (pg.  202- 224) Google Scholar CrossRef Search ADS   Soglio V,  Costa F,  Molthoff JW,  Weemen-Hendriks WMJ,  Schouten HJ,  Gianfranceschi L.  Transcription analysis of apple fruit development using cDNA microarrays,  Tree Genetics and Genomes ,  2009, vol.  5 (pg.  685- 698) Google Scholar CrossRef Search ADS   Stevens R,  Buret M,  Duffe P,  Garchery C,  Baldet P,  Rothan C,  Causse M.  Candidate genes and quantitative trait loci affecting fruit ascorbic acid content in three tomato populations,  Plant Physiology ,  2007, vol.  143 (pg.  1943- 1953) Google Scholar CrossRef Search ADS PubMed  Thomas TR,  Shackel KA,  Matthews MA.  Mesocarp cell turgor in Vitis vinifera L. berries throughout development and its relation to firmness, growth, and the onset of ripening,  Planta ,  2008, vol.  228 (pg.  1067- 1076) Google Scholar CrossRef Search ADS PubMed  Van Ooijen JW. ,  Joinmap® 4, software for the calculation of genetic linkage maps in experimental populations ,  2006 Wageningen, Netherlands Kyazma BV Van Ooijen JW. ,  MAPQTL® 6, Software for the mapping of quantitative trait loci in experimental populations of diploid species ,  2009 Wageningen, Netherlands Kyazma BV Vandesompele J,  De Preter K,  Pattyn F,  Poppe B,  Van Roy N,  De Paepe A,  Speleman F.  Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes,  Genome Biology ,  2002, vol.  3 pg.  7  Google Scholar CrossRef Search ADS   Velasco R,  Zharkikh A,  Affourtit J, et al.  The genome of the domesticated apple (Malus×domestica Borkh.),  Nature Genetics ,  2010, vol.  42 (pg.  833- 839) Google Scholar CrossRef Search ADS PubMed  Vorrips RE.  MapChart: software for the graphical presentation of linkage maps and QTLs,  Journal of Heredity ,  2002, vol.  93 (pg.  77- 78) Google Scholar CrossRef Search ADS PubMed  Vrebalov J,  Ruezinsky D,  Padmanabhan V,  White R,  Medrano D,  Drake R,  Schuch W,  Giovannoni J.  A MADS-box gene necessary for fruit ripening at the tomato ripening-inhibitor (Rin) locus,  Science ,  2002, vol.  296 (pg.  343- 346) Google Scholar CrossRef Search ADS PubMed  Wagner A.  The fate of duplicated genes: loss or new function?,  Bioessays ,  1998, vol.  20 (pg.  785- 788) Google Scholar CrossRef Search ADS PubMed  Wakasa Y,  Kudo H,  Ishikawa R,  Akada S,  Senda M,  Niizeki M,  Harada T.  Low expression of an endopolygalacturonase gene in apple fruit with long-term storage potential,  Postharvest Biology and Technology ,  2006, vol.  39 (pg.  193- 198) Google Scholar CrossRef Search ADS   Waldron KW,  Parker ML,  Smith AC.  Plant cell walls and food quality,  Comprehensive Reviews in Food Science and Food Safety ,  2003, vol.  2 (pg.  101- 119) Google Scholar CrossRef Search ADS   White PJ.  Recent advances in fruit development and ripening: an overview,  Journal of Experimental Botany ,  2002, vol.  53 (pg.  1995- 2000) Google Scholar CrossRef Search ADS PubMed  Zhu YM,  Barritt BH.  Md-ACS1 and Md-ACO1 genotyping of apple (Malus×domestica Borkh.) breeding parents and suitability for marker-assisted selection,  Tree Genetics and Genomes ,  2008, vol.  4 (pg.  555- 562) Google Scholar CrossRef Search ADS   © The Author [2011]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org
TI - Comprehensive QTL mapping survey dissects the complex fruit texture physiology in apple (Malus x domestica Borkh.).
JF - Journal of Experimental Botany
DO - 10.1093/jxb/err326
DA - 2011-11-25
UR - https://www.deepdyve.com/lp/oxford-university-press/comprehensive-qtl-mapping-survey-dissects-the-complex-fruit-texture-hNA0I94708
SP - 1107
EP - 1121
VL - 63
IS - 3
DP - DeepDyve
ER -