Starch content was determined according to the protocol described (Blanpied and Silsby, 1992). Six apple fruit were sliced horizontally to make a disk covering the whole apple core, cortex, and peel. The exposed side of the apple was dipped into starch–iodine solution for 2 min. Flesh containing starch turn into blue–black color and were indexed from 1 to 8, where index 1 indicated high starch with dark blue–black color covering core and cortex tissue, and index 8 indicated no starch, with no color.

Chlorophyll content was measured by portable delta absorbance (DA) meter (Sinteleia, Bologna, Italy). Delta absorbance meter measures the difference in absorbance of chlorophyll a, at 670 and 720 nm and expresses as index of absorbance difference (I_AD = A670–A720) (Costamagna et al., 2013). I_AD was measured for total 10 apples on two opposite sides of each fruit. Delta absorbance meter has been used previously to determine harvest time in apple (DeLong et al., 2014).

The response to exogenous ethylene was performed at all developmental stages (S1–S5, S6/H). Ethylene (10 ppm) was supplied by injecting into airtight container, containing 10 fruit, for 24 h at 20°C. Following treatment, fruit were placed individually into 2 L jars for 1 h at 20°C, and the ethylene production and respiration rates were monitored as described above. In parallel, non-treated fruit were also examined similarly.

For the determination of TSSs and TA content, each fruit (10 fruit per stage) was peeled and cut from opposite sides to make wedge-shaped slices. Slices obtained from each fruit were pooled and grounded, and the extracted juice was used for further analysis. Total soluble solid content was determined by digital refractometer PR-1 (Atago, Tokyo, Japan), and TA content was determined by the titration of 2 mL juice to 0.1 N NaOH (pH 8.2) using Dosimat 665 (Metrohm, Switzerland) with 678 EP/KF processors, and were expressed as percentage of malic acid.

Juice content (expressible juice) in the fruit was determined on a cylinder of 1 × 1 cm in length and diameter from the fruit cortex tissue, having the approximate weight of 2 g. The cylinder was weighed and compressed in 5 mL syringe, the extract was collected into Eppendorf tube, and centrifuged at 10,000 × g for 15 min. The amount of expressible juice was calculated as percent of original cylinder weight following the protocol of Lill and Van der Mespel (1988).

Fruit firmness was measured on 10 fruit, by penetration using a Agrosta^®14 Motorized Digital Penetrometer or by deformation using Universal testing machine Inspekt table blue 5 kN (Hegewald & Peschke MPT GmbH, Germany). For penetrometer measurement, each fruit was peeled from two opposite side, and a probe (8 mm in diameter) was used to puncture the tissue for 40 ms. For deformation, force was applied to deform the apple by 5% of its circumference. The force applied for both, penetration or deformation was expressed in Newton (N).

Autoluminescence emission is generated from spontaneous photons emission by oxidation of lipid molecules from the tissue (Birtic et al., 2011). This was determined by In-Vivo Imaging Systems (IVIS, PerkinElmer, MA, United States) on six fruit, sampled at S6/H, R1, and R2. Before acquiring the images, fruit were kept in dark for 24 h to avoid any photon excitation. All the imaging parameters were standardized for apple fruit, and were kept the same for all measurements. Autoluminescence from the samples was acquired using the following settings: f/stop = 1.2, binning = large, exposure time = 25 min, excitation = block, emission > 600 nm. Luminescence images showing emission in photons/s/cm²/steradian were captured and quantized from whole apples.

Total RNA was extracted using Spectrum^TM Plant Total RNA Kit (Sigma–Aldrich). The DNA contamination from RNA samples was removed by using TURBO DNA-free^TM kit (Ambion, Life Technology), subsequently, c-DNA was prepared using the Verso cDNA Synthesis Kit (Thermo Scientific), and was used for further analysis. Gene expression analysis was performed by quantitative reverse transcriptase-PCR (qRT-PCR) containing cDNA, forward and reverse primers, and Fast SYBR^TM (Applied Biosystems) in a 10 μL reaction volume. Reactions were performed into StepOnePlus Real-Time PCR System (Applied Biosystem) using reaction condition of 40 cycles for 10 s at 95°C, 15 s at 60°C, and 20 s for 72°C, and results were analyzed by StepOne Software. The relative expression levels of the targeted genes were calculated by either 2^-ΔΔC_t or 2^-ΔC_t method, using actin as housekeeping gene. Primers were designed using Primer3Plus and are listed in Supplementary Table 2.

High throughput gene expression analysis was performed using Biomark HD System (Fluidigm, United States). The Fluidigm 48.48 dynamic array chip was used following the manufacturer’s ADP37 Fast GE¹ protocol, which allows 2304 simultaneous real-time PCR gene expression. Primer specificity and reference genes were validated prior to analysis. Pre-amplification of cDNA was performed on 1.25 μL of 50 ng μL^-1 samples using Fluidigm PreAmp Master Mix (Fluidigm, PN 1005581), and 2.7 μL of each pre-amplified cDNA was mixed with 3 μL of SsoFast EvaGreen Supermix with Low Rox (BioRad, PN 1725211) and to 0.3 μL of 20× Binding Dye Sample Loading Reagent (Fluidigm, PN 1001388). Individual primer pairs (50 μM) in a 1.08 μL volume mixed with 3 μL Assay Loading Reagent (Fluidigm, PN 85000736) and 1.92 μL of Low TE. Total 5 μL of each sample mix or each assay mix was then pipetted into individual sample inlet in the 48.48 Dynamic Array chip, and an (IFC) controller MX (Fluidigm) to prime the chip. The loaded chip was placed in the BioMark system for PCR at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. Following each reaction in a specific inlet, the PCR amplification curve was generated and chip was imaged. The dynamic array raw data were analyzed with the Fluidigm Real-Time PCR Analysis software. The gene expression was calculated using the 2^-ΔC_T method, following normalization with actin. Heatmap was prepared by MultiExperiment Viewer, MeV v4.9 software, using expression profile obtained from 2^-ΔC_t, where hierarchical clustering of genes was based on the Spearman correlation, allowing genes clustering according to their expression patterns.

Statistical analysis of data was performed by Tukey’s HSD pairwise comparison test at p ≤ 0.05, using JMP 5.0.1a statistical software (SAS Institute Inc., NC, United States).

We compared the quality parameters of “Anna” to “Galaxy” and “GD” at commercial harvest (S6/H) and following postharvest storage of 1 (R1) or 2 months (R2) (Figure 1). All cultivars achieved approximately similar size and weight at their commercial harvest (Supplementary Figure S1). Starch was highest at harvest of “Anna” in comparison to “Galaxy” and “GD.” For all three cultivars, the content of starch declined at R1–R2, showing no starch at R2 (Figure 1A).

A significant increase in TSS from H to R2 was observed only for “Galaxy” and the level was higher than in other cultivars (Figure 1B). “Anna” contained the highest TA content (Figure 1C), it decreased significantly from H to R2, with the major decline in “Anna,” compared to “GD” and “Galaxy.”

Expressible juice content was high at harvest for all cultivars, and declined significantly at R1 and R2 for “Anna” and “GD” (Figure 1D). “Anna” produced higher expressible juice at harvest than the other cultivars, showing a drastic decline at R1 (by 81%) and further at R2 (by 91%), whereas in “GD” decline was only by 36–53% at R1–R2, and “Galaxy” exhibited a marginal reduction of 18–30% at R1–R2. The juice levels following storage were high in “GD” and “Galaxy,” and therefore these cultivars remained juicier compared to “Anna” which developed mealiness after storage.

Fruit firmness which was measured, either by deformation (Figure 1E) or penetration (Supplementary Figure S2) was highest at H and declined at storage, R1 (56%) and R2 (70%) in “Anna” (Figure 1E). Comparatively, “Galaxy” and “GD” maintained fruit firmness with a reduction of 20–22% and 20–42% at R1–R2, respectively. Similar high reduction in firmness in “Anna” compared to the other cultivars was measured by penetration (Supplementary Figure S2).

The oxidative stress status of apple fruit was determined by measuring autoluminescence exerted from in vivo oxidation of lipids (Figure 1F). “Anna” and “GD” displayed high autoluminescence at harvest compared to “Galaxy.” Autoluminescence was enhanced significantly during storage for “Anna,” R1 (251%) and R2 (95%), but only by 168% and 148% for “Galaxy” at R1 and R2, respectively. Comparatively, “GD” showed a 12% reduction at R1 and 15% increment at R2.

Rates of ethylene (C₂H₄) and CO₂ (respiration) production were determined at H, R1, and R2 for all three cultivars. Ethylene production rate increased following storage in all three cultivars where “Anna” produced significantly higher level of ethylene at harvest and following storage, compared to “Galaxy” and “GD” (Figure 2A). In parallel, “Anna” produced higher CO₂ also at H, than the other two cultivars (Figure 2B). However, during storage, respiration rate of “Anna” exhibited a peak production at R1, while in the other two cultivars the respiration rate incremented gradually.

Genes of MdACS and MdACO families are available in National Center for Biotechnology Information (NCBI) from several apple cultivars. However, it was not always clear to which chromosome location these sequences referred. We determined the chromosome location and the peptide length of all the available accessions (Supplementary Tables 1A,B). Total 19 genes were identified for MdACS (1, 3A, 3B, 3C, 4, 5A, 5B, and 6–17), and 7 for MdACO (1–7) in the Genome Database for Rosaceae (GDR)². A wide range expression analysis of all these genes was performed during fruit developmental, at harvest and following storage, in “Anna,” “Galaxy,” and “GD.” Six genes, from each family, MdACS (MdACS1, 3a, 5B, 6, 8, and 9) and MdACO (MdACO1-5, 7) were expressed in either of the apple cultivars (Supplementary Tables 1A,B).

The major ethylene biosynthesis gene, MdACS1, MdACO1, and MdACO7 were expressed significantly higher during harvest and storage, in comparison to the other genes (Figure 3). At harvest the expression of MdACS1 and MdACO1 was lowest in “Anna”; however, their expression increased dramatically during storage (R1–R2) compared to “Galaxy” and “GD.” Examining the 5′ region of MdACS1 revealed a 489 nucleotide section in “Anna,” which is a typical feature of MdACS1-1 allele, related to higher gene expression and ethylene production (Supplementary Figure S3). MdACO7 expressed the highest in “Anna,” and increased from H to R2, in contrast to both “Galaxy” and “GD” where expression was low.

Different developmental stages of all three cultivars; “Anna,” “Galaxy,” and “GD” were collected to follow changes in respiration, ethylene production rate, response to exogenous ethylene, expression of genes related to ethylene biosynthesis and response, in addition to developmental regulators. Fruit were harvested according to their size (S1–S6/H) (Figure 4) and the chlorophyll and starch content were monitored at these stages. Chlorophyll content was expressed as index of absorbance difference (I_AD) (Figure 5A). The I_AD value of “Anna” remained similar throughout S1–S5, but declined significantly at S6. Both “Galaxy” and “GD” exhibited an earlier decline in I_AD at S3 and further at S6. At S6, a major decline was observed by 27, 84, and 36% for “Anna,” “Galaxy,” and “GD,” respectively, therefore, S6 was considered as the transition stage (Breaker, B), and indeed fruit were commercially harvested (H) at this stage. Starch content in the fruit was expressed as starch–iodine index (Figure 5B). In all cultivars, the starch content reaching the highest levels at S4 and in “Galaxy” and “GD” it remained high also at S5. In “Anna” slight decline was appeared at S5, which remained same at S6. On contrary, the levels of starch decline dramatically at S6 in both “Galaxy” and “GD.” Interestingly, “Anna” maintained higher starch levels during all developmental growth, compared to “Galaxy” and “GD.”

Since ethylene is known to enhance ethylene production in system II, but not in system I (Liu et al., 2015), we examined the respiration and ethylene production rates with (Figures 6B,D) or without (Figures 6A,C) exogenous ethylene treatment at S1–S6/H/B in all three cultivars. Non-treated “Galaxy” and “GD” fruit exhibited negligible ethylene production rate during development, while “Anna” had low ethylene production rate at S3 which was maintained in subsequent stages (Figure 6A). Both “Galaxy” and “GD” showed similar respiration rate featuring high rate at S1 with a decline at S3 and remaining negligible thereafter. In “Anna,” on the other hand, respiration rate remained high throughout development (Figure 6C).

Exogenous ethylene at all developmental stages induced the CO₂ and ethylene production rates only in “Anna.” The ethylene production rate exhibited 6–14-fold increment for all stages and respiration rate exhibited an increase of 95–160% at S1–S2 and 12–17% at S3–S5. On the other hand, ethylene production increased in “Galaxy” only at S1 and S6, and in “GD” just at S6 (Figures 6B,D).

Since pre-climacteric ethylene and fruit developmental factors/mechanisms might affect fruit ripening (Alexander and Grierson, 2002; Giovannoni, 2004), the homolog of genes belonging to MdACS and MdACO families, encoding ETRs, regulatory components within the ethylene response pathway and few of the developmental regulators acting upstream of ethylene were identified in apple (Table 1). The expression pattern of these genes was determined during fruit development (S1–S6/H/B).

Among the ethylene biosynthesis genes, the highest increase in expression during fruit development occurs in MdACS3a and MdACO7, mainly in “Anna” (Figures 7A,B). “Anna” showed a rise at S3, with lower expression in other stages for MdACS3a, whereas “Galaxy” exhibited minor changes only at S3. MdACO7 expressed throughout fruit development (S1–S6/H/B) of “Anna” with a peak at S3; however, both “Galaxy” and GD showed a lower change in expression of MdACO7.

The comprehensive gene expression profiling during fruit development (S1–S6/H/B) and also during storage (R1–R2) is presented in Figure 8 (and also in Supplementary Figure S4). The hierarchical clustering of genes was based on the Spearman correlation and was clustered into six distinct clades (I–VI) according to their expression pattern and levels (Figure 8). Clade IV containing genes, MdACS5B, MdTAGL1, and MdERF2 were highly expressed in “GD.” The expression of the genes of Clades III, V, VI A, and VI C was similar for all cultivars, except MdFUL and MdACO3 which were higher in “GD” and “Galaxy” than in “Anna.” On the other hand, the expression of genes of Clade II (MdCTR1, MdSBP, MdSBP2, MdSBP7, MdETR1, MdHB-1, MdAP2likeERF/TOE3, MdAP2) and VI B (MdERF5, MdEIN2, MdEIL3, MdCNR) was higher in the two cultivars “Galaxy” and “GD” than in “Anna.” In contrary, clusters I included the genes MdACO2 and 4 which were expressed higher in “Anna” than “Galaxy” and “GD.”

The comparative storage capacity analysis of “Anna,” “Galaxy,” and “GD” apple fruit at their commercial harvest (S6) was most likely performed at a similar developmental stage. This stage might be similar to the “breaker” in tomato and it is supported by the fact that in all cultivars a major decline in chlorophyll content as expressed in I_AD was observed at stage 6 (Figure 5). The I_AD measurement has been established previously to identify the appropriate harvest time for different apple cultivars (Costamagna et al., 2013). In addition, auto-stimulatory response to exogenous ethylene of all cultivars at this stage further supports the notion that all cultivars were in a similar developmental stage (Figure 6). At this stage only “Galaxy” and “GD” expressed the Polygalacturonase (PG1) transcript (Figure 8), while it increased at storage in “Anna.” It seems that developmentally “Anna” might be slightly less mature than the others. Indeed, at this stage “Anna” contained higher starch than the other cultivars, which also might suggest lower maturity, but at least for starch, it has been suggested that in few cases its higher levels are not indicative of lower maturation (Watkins et al., 1993). Despite “Anna” having a lower PG1 expression at harvest (Figure 8), its storage capacity was inferior to the other cultivars.

Juiciness and firmness determined at harvest and following storage confirmed the poor fruit quality of “Anna.” Although “Anna” at harvest was the juiciest among the cultivars, during storage, it lost its juiciness more than the other cultivars. It was suggested that lack of juiciness in apple is related to cell separation, preventing the release of cell content (Tu et al., 2000) and absorption of juice into pectin gel. More recently, it was suggested that it is related to cell wall degrading enzyme pectin methyl esterases (PMEs) which exhibited lower expression in mealy apple fruit throughout development in comparison to non-mealy fruit (Segonne et al., 2014), possibly leading to decreased cell-to-cell adhesion. A drastic declined in firmness occurred in “Anna,” further emphasizing the poor storage capacity of this cultivar. Our results are in accordance with the finding suggesting that early season apple cultivars are more prone to softening, compared to late season cultivars (Wiersma et al., 2007). Since “Anna” is grown at a lower altitude than the other cultivars, an environmental effect on this storage capacity cannot be excluded.

Low storage capacity of “Anna” was also reflected in high autoluminescence photon emission, which was highest in “Anna,” compared to “Galaxy” and “GD” following storage (Figure 1F). Since autoluminescence appears due to lipid oxidation under stress conditions (Birtic et al., 2011), we suggest that “Anna” was under oxidative stress. It is possible that the low storage capacity and the increase in autoluminescence of “Anna” resulted from higher respiration and ethylene production rates. High respiration is responsible for fast metabolism and a decline in fruit acidity, resulting from consumption of organic acid (Etienne et al., 2013). Indeed, TA was reduced significantly in “Anna” during storage where respiration rate was highest (Figure 2).

High ethylene production in “Anna” following storage coincided with higher expression of MdACO1, 2, 4, 7, and MdACS1 (Figure 3). These genes, most likely, are involved in system II ethylene biosynthesis, as suggested for MdACS1 (Dandekar et al., 2004) and MdACO1 (Schaffer et al., 2007). We identified that “Anna” was homozygous MdACS1-1/1-1 (Supplementary Figure S3), which was well correlated with the existence of homozygous MdACS1-1/1-1 in higher ethylene producing apples (Sunako et al., 1999) and in early season apple cultivar (Oraguzie et al., 2004).

Examining respiration and ethylene production rates throughout fruit development (S1–S5; Figure 6) revealed that “Anna” exhibited higher levels in comparison to “Galaxy” and “GD.” Furthermore, in response to exogenous ethylene treatment, “Anna” showed an ethylene-dependent positive feedback regulation with the induction of both ethylene and CO₂ production throughout fruit development (Figure 6). Conversely, both “Galaxy” and “GD” remained unaffected by external ethylene at all stages prior to S6, indicated that system I operates prior to transition stage in these cultivars. Therefore, these results suggested the existence of a system II-like ethylene biosynthesis in “Anna,” where ethylene production is under auto-stimulatory control (Barry et al., 2000; Liu et al., 2015). A similar, system II-like ethylene biosynthesis has been reported in non-climacteric citrus fruit at young stage (Katz et al., 2004); however, unlike in “Anna” it was restricted to an early period of fruit development.

The higher ethylene and respiration in “Anna” was accompanied by higher expression of major ethylene biosynthesis genes, MdACSO2, 4, and 7 and particularly MdACS3a during early fruit development. It has been reported that the cultivars’ specific expression of MdACS3a, and existence of specific allele of this gene leads to high ethylene production in different apple cultivars (Wiersma et al., 2007; Wang et al., 2009; Varanasi et al., 2011; Bai et al., 2012).

In this study, we also examined genes within the ethylene response pathway and upstream transcription factors which control the ripening response (Figure 8 and Supplementary Figure S4). Ethylene receptors and CTR kinase are negative regulators of the ethylene response (Liu et al., 2015). MdETR1 showed a lower expression in all cultivars throughout development in comparison to MdETR2 which fits with the observation that MdETR2, but not MdETR1 is induced by ethylene. Nevertheless, MdETR1, 2, and MdCTR1 exhibited lower expression in “Anna” in comparison to the other cultivars, during fruit development. The contribution of these negative regulators at the pre-breaker stage to “Galaxy” and “GD” fruit quality is still not clear.

Homologs of negative (AP2, MADS1) or positive (NOR, RIN, TAGL1, FUL1/2, CNR) regulators of tomato ripening (Giovannoni, 2004; Karlova et al., 2014) were identified in apple, and most of them were expressed similarly in all three apple cultivars except, MdFUL, MdAP2, and MdCNR. These genes exhibited lower expression throughout development of “Anna,” in comparison to the other cultivars (Figure 8 and Supplementary Figure S4). Fruitful (FUL) might have different function in apple and tomato, since in tomato reduced expression of FUL1/2 inhibited mainly lycopene production, but not ethylene production, and in apple MdMADS2 (MdFUL) was suggested to be involved in maintaining the apple fruit firmness (Cevik et al., 2010). Since expression of MdFUL was lower in “Anna” than in the other cultivars, we suggest that lower expression might be related to lower fruit firmness/higher mealiness (Figure 1 and Supplementary Figure S2), but this should be further investigated. Lower expression of the AP2 homolog enhanced ripening in tomato (Chung et al., 2010), and it is possible that it acts similarly in apple. The expression of AP2 was low during early development (stages 2 and 3; 50–75 days after full bloom) of the cultivars “Galaxy” and “GD,” but increased later, however that of “Anna” remained low throughout development. Similar expression to that of “Galaxy” and “GD” was observed in the “Mondial Gala” apple cultivar (Costa et al., 2010). The expression pattern of FUL fits with its function as negative regulator. Colorless non-ripening is a SQUAMOSA SBP, critical for tomato fruit ripening, and the mutant cnr, exhibiting lower expression of the gene (Manning et al., 2006), has reduced ethylene production and mealy fruit, due to reduction in cell-to-cell adhesion (Thompson et al., 1999). Accordingly, we suggest that low expression of MdCNR in “Anna” might be responsible for the reduced expressible juice content and development of mealiness after harvest (Figure 1D). However, lower expression of CNR in “Anna” is also associated with higher ethylene production and faster firmness loss (Figures 1, 2). Since the nature of the mutation in “Anna” is still not clear it might be possible that additional processes are affected in “Anna.”

“Anna” also exhibited low expression of other SBP, MdSBP, MdSBP2, and MdSBP7. Few SBP transcription factors are known to bind the promotor of many genes responsible for maintaining copper homeostasis within the cell (Yamasaki et al., 2009) or assembly of mitochondrial electron transport chain complex IV subunits (Garcia et al., 2014). Moreover, SBP are also responsible for maintaining the levels of two isoforms of superoxide dismutase (SOD), iron SOD FeSOD, and copper SOD (CuSOD) under oxidative stress (Nagae et al., 2008). Since the expression of MdSBP, MdSBP2, and MdSBP7 was lower in “Anna” than in “GD” and “Galaxy,” we postulate that these genes might be responsible for improper mitochondrial functioning or electron flow in this cultivar, which ultimately leads to high respiration rate (Figure 6). This might explain “Anna” higher autoluminescence, indicative of oxidative stress (Figure 1F).

Recently, the chromosome location of early bud break of “Anna” has been discovered including several SNPs in several genes (Trainin et al., 2016), however, so far, it is not clear if the low storage capacity of “Anna” is also localized to the same site. Taken together, this study provides new understanding on pre-climacteric events in “Anna” that might affect its ripening behavior and storage capacity.