Abstract
The plant cuticle plays several roles with agronomic relevance such as controlling water loss or protecting against ultraviolet (UV) radiation and mechanical damage. In tomato fruit cuticle, phenolics are chiefly the flavonoid chalconaringenin and a small portion of phenolic acids. They confer the cuticle mechanical resistance as well as thermal properties and UV photoprotection. Nine QTLs related to the accumulation of cuticle phenolics in tomato fruit were previously detected in a Solanum lycopersicum × S. pimpinellifolium recombinant inbred population. Confocal Raman microscopy has been employed to separate the effect of these cuticle QTLs on the two phenolic fractions, showing that most QTL lines preferentially affected chalconaringenin accumulation, with only two QTL lines (ph8.1 and ph12.1) having an equal effect on both chalconaringenin and phenolic acids. A QTL pyramiding strategy has been employed to study the combined effect of different loci and has revealed a complex array of interactions. Approximately 80% of the double and triple QTL lines displayed significant epistatic interaction. Notably, a QTL combination between chromosomes 5 and 12 that increases the percentage of cuticle phenolics beyond single QTL lines was identified. CHALCONE SYNTHASE (CHS1 and CHS2) expression analyses have shown that CHS1 seems to be the main enzyme responsible for the synthesis of chalconaringenin that is accumulated within the cuticle. Additionally, an S. pimpinellifolium genomic region in chromosome 5 that favors CHALCONE ISOMERASE1 expression in the peel of ripening tomatoes without affecting chalconaringenin accumulation in the cuticle was identified, raising the possibility of an increase in fruit of downstream flavonols without compromising cuticle phenolics.
1 Introduction
At the plant surface, in contact with the environment, the cuticle plays numerous roles regarding plant protection against biotic and abiotic stresses. At the same time, it has an effect on fruit and vegetable quality affecting traits such as water loss and fungal growth during postharvest and, in tomato, fruit color and gloss (; ; Petit et al., 2013; ; Wu et al., 2023). The cuticle can be regarded as a modification of the cell wall, and it is chiefly composed of a polyester matrix named cutin, waxes, phenolics, and polysaccharides derived from the cell wall (Reynoud et al., 2022). Phenolics are mainly hydroxycinnamic acids, especially p-coumaric acid, which has been reported in all the cuticles analyzed so far and seems to be esterified to the cutin polyester (Rautengarten et al., 2012; Renault et al., 2017). Although the amount of cuticle phenolics tends to be low, in some species, they are present as a notable fraction (; Reynoud et al., 2022; ; ). Additionally, in tomato fruit, the flavonoid chalconaringenin has also been identified within the phenolic fraction (). Tomato cuticle is characterized by the accumulation of phenolic acids (p-coumaric and p-hydroxybenzoic acids) throughout fruit development, while chalconaringenin only accumulates during ripening (, ). Recently, confocal Raman microscopy (CRM) has been employed to identify spectral bands differentially assigned to chalconaringenin and phenolic acids and to study the topochemistry of phenolics within the tomato fruit cuticle (Reynoud et al., 2022; ). Thus, it was reported that during tomato fruit growth, phenolic acids are mainly located as a thin layer close to the outer side of the cuticle while, during ripening, phenolic acids and flavonoids are preferentially incorporated to the inner side of the cuticle ().
The extensive work carried out in tomato cuticle has allowed several functions to be ascribed to the phenolic fraction. They have been shown to confer biomechanical resistance to the cuticle, especially during ripening when cell walls are being degraded (, ). Additionally, incorporation of the yellow-orange colored flavonoid naringenin chalcone during ripening contributes to the characteristic red color of tomatoes (; ; ). Cuticle enrichment in phenolics, rather than the amount of phenolics itself, has recently been shown to increase the glass transition temperature, modifying the thermal properties of the cuticle (, ). Finally, cuticle phenolics have been shown to provide photoprotection against ultraviolet-B (UV-B) in developing fruits and also UV-A in ripe fruits (; ). Hence, understanding the genetic basis behind natural variation for these compounds brings the possibility of improving several biophysical traits that have relevant agronomical implications including reduced cracking, improved photoprotection, or more biomechanically resistant cuticles under specific storage temperature conditions.
Numerous genes involved in the biosynthesis and regulatory network of cuticle deposition have been identified by means of reverse genetics (Yeats and Rose, 2013; ), and more recently, the genetic basis behind the natural variability for some cuticle traits has been explored (; ; ; Popovsky-Sarid et al., 2017; Rett-Cadman et al., 2019; ; ; Yang et al., 2022). However, despite the relevance of phenolics in some cuticle roles of potential interest for plant breeding, only one work has addressed the analysis of the natural variability for this trait (). Additionally, phenolic compounds, especially flavonoids, are known for their health-promoting characteristics. Thus, in tomato, several studies have reported strategies to enhance the phenolic content of the fruit, thanks to the fact that the metabolic and regulatory pathways of phenolics are quite known (; ; ; Yang et al., 2022). However, although phenolic compounds are known to accumulate in the peel of tomato fruit, only in a few instances has the potential effect of these strategies at the cuticle level been studied. Thus, overexpression of the transcription factors ROSEA and DELILA in tomato fruit did not affect the phenolic content or the mechanical properties of the cuticle (). However, characterization of the mutation colorless fruit epidermis (y) responsible for pink tomatoes showed an almost absence of phenolics in the cuticle and a mechanically weaker cuticle (). Similarly, silencing CHALCONE SYNTHASE, the enzyme responsible for the synthesis of chalconaringenin, or ARLEQUIN/TOMATO AGAMOUS-LIKE1, a transcription factor involved in fruit development, caused a significant reduction of cuticle phenolics and mechanical properties of the tomato fruit cuticle (; ).
Quantitative traits are characterized by a complex genetic makeup involving several genes with small effects. Understanding how genes interact within a given genetic background and investigating how their accumulation affect the phenotype are crucial for plant breeding purposes. In a previous work, the combined analysis of recombinant inbred line (RIL) and introgression line (IL) populations between the domesticated tomato Solanum lycopersicum “MM” and a wild species, Solanum pimpinellifolium, accession “TO-937”, allowed the identification of nine additive QTLs and some epistatic effects related to the accumulation of phenolics in the tomato fruit cuticle, and a candidate gene analysis of some QTLs was carried out (). QTLs with a positive effect on cuticle phenolics were identified for both parents. The potential use of QTLs with desirable traits in a breeding program can strongly benefit from understanding the interactions between different QTLs in order to identify the most beneficial combinations. Therefore, in this work, epistatic interactions of phenolic-related loci were investigated by means of QTL pyramiding in double and triple QTL lines. This analysis will additionally improve our understanding of the pathways involved in phenolic accumulation within the tomato fruit cuticle.
2 Materials and methods
2.1 Plant material and QTL pyramiding
Eight lines harboring QTLs for the percentage of cuticle phenolics were selected from a 52 IL population derived from the recurrent backcross of Solanum pimpinellifolium L. “TO-937” (TO-937) into Solanum lycopersicum L. “Moneymaker” (MM) background (). These lines had TO-937 introgressions for each of the QTLs (QTL lines), and the size of each introgression was reported in . Seven of these lines had a single cuticle phenolic QTL while the remaining one harbored two QTLs together (ph7.1+ph7.2). The eight lines were crossed among them to pyramid QTLs. Additionally, two subILs that separate ph7.1 and ph7.2 were also employed and crossed with the IL harboring the highest additive effect QTL, ph1.1 (). Thus, a total of 30 F1s were generated (Supplementary Table 1). A minimum of 20 F1 plants from each cross were grown in a polyethylene greenhouse and self-pollinated, and F2 seeds were collected. Between 100 and 200 plantlets from each F2 population were tested to identify plants combining TO-937 alleles for the different QTLs. DNAzol (Life Technologies, USA) was used to isolate DNA from these F2 populations. Flanking markers spanning each QTL, together with the peak marker, were reported in . Peak markers were used to design primers for plant selection (Supplementary Table 2). Single nucleotide polymorphisms (SNPs) were employed to select the desired combinations and resolved using high-resolution melting (HRM) analysis.
Forty plants of MM and TO-937 and 20 plants per QTL line (single, double, or triple) were grown in a polyethylene greenhouse at the Estación Experimental La Mayora, CSIC, Spain. Because of the high number of double and triple QTL lines, the study was divided and carried out in 3 years. Each year, MM, TO-937, and the single QTL lines were grown and studied together with a selected group of double and triple QTL lines. A random four-block design was employed to study the effect of cuticle phenolic QTL pyramiding. Flower trusses were vibrated twice a week to ensure that fruit set. Flowers were tagged at anthesis and fruits were harvested at three distinct physiological stages: immature green [15 days after anthesis (daa)], mature green (MG; 30–35 daa) and red ripe (RR; 55 daa). Sample size consisted of 100 fruits per genotype (200 for MM) at 15 daa and 40 fruits per genotype (80 in the case of MM) for the subsequent stages. Following harvest, all fruits were immediately processed for cuticle isolation. Additionally, a subset of fruits was collected at breaker stage (40–45 daa) and their epicarp tissue was collected and flash-frozen for gene expression analyses.
2.2 Cuticle isolation and characterization
Tomato fruits were halved and immersed in a 50 mM aqueous solution of sodium citrate (pH 3.7) with a mixture of 1% w/v cellulase and 1% w/v pectinase and 1 mM NaN3 to prevent microbial growth. Samples were incubated at 37°C with constant agitation. After 1 week, peels were detached from the rest of the fruit tissue and subsequently incubated in fresh solution to allow total enzymatic disruption of the remaining cellular tissues. After 7–10 days, the solution was refreshed and samples were incubated for at least another week. Isolated cuticles were air dried and stored under dry conditions. Cuticle extraction was carried out at three stages of development—immature green, MG, and RR—for MM, TO-937, and all the QTL lines (single, double, and triple).
The amount of cuticle was determined gravimetrically from flat pieces of cuticle of known surface area. Ten measurements per line, each corresponding to pieces of different RR fruits collected from different plants, were carried out. This analysis was carried out in MM, the single QTL lines that carried out overlapping QTLs for phenolics and amount of cuticle (ph4.1, ph5.1, ph5.2, and ph12.1) (), and the 22 double and triple QTL lines that contained them.
Cuticle color was analyzed with a Konica Minolta CR-400 colorimeter in the CIE 1976 L*a*b* and CIE L*C*h°color spaces using illuminant D65. Twenty-five pieces of cuticle were measured for TO-937 and single QTL lines, and 50 pieces of cuticle were measured for MM control. Each piece of cuticle corresponded to a different fruit collected from either 12 plants (TO-937 and single, double, and triple QTL lines) or 25 plants (MM). Additionally, CIE DE2000 color difference equation (Sharma et al., 2005) was employed to assess cuticle color differences of the double and triple QTL lines with the parental line MM.
Total cuticle phenolics were estimated at three stages of development (immature green, MG, and RR) using a UV–VIS spectrophotometer (Pharmacia Biotech, NJ, USA) after cutin depolymerization for 24 h in a 1% w/v of NaOH in methanol at 65°C with agitation (). The fraction of phenolics present in the cuticle, expressed as percentage of cuticle phenolics, was estimated at each developmental stage from five biological samples per QTL line (single, double, and triple) and TO-937, whereas for MM control, 10 samples were studied. Each biological sample corresponded to pieces of cuticles isolated from at least five fruits collected from four different plants.
Raman spectroscopy analyses of isolated cuticles from MM, TO-937, and the individual QTL lines were carried out using an NRS 5100 (Jasco) Raman dispersive spectrometer coupled with a confocal microscope. The 785-nm laser was employed to minimize autofluorescence in the confocal Raman microscope. Samples were focused with a 100× objective [numerical aperture (NA) = 0.90]. A CCD detector cooled by the Peltier system was employed as detector. The laser power was set at 11.8 mW and the integration time was set to 30 s. Twenty scans were accumulated per sample. The internal side of isolated cuticles was analyzed since accumulation of flavonoids and phenolic acids during ripening has been shown to be differentially located across the tomato fruit cuticle (). Band deconvolution and spectral analyses were carried out using the software OMNIC (Thermo Scientific, USA). Five to seven samples per line, each corresponding to a cuticle piece from a fruit collected from a different plant, were analyzed.
2.3 Expression analyses
RNA extraction from fruit epicarp at breaker stage was carried out with Trizol® (Life Technologies, USA). Genomic DNA was removed by treating with RNase-free DNase, and RNA was cleaned with Nucleospin RNA clean-up (Macherey-Nagel, Germany). First-strand cDNA synthesis was carried out with the Super Script III First-Strand Synthesis Super Mix for qRT-PCR according to the manufacturer’s instructions (Invitrogen, USA). Three biological replicates corresponding to pools of fruit epicarp from five different plants were analyzed for TO-937 and the single QTL lines and six for MM. Relative transcript amount of the different genes was measured by RT-qPCR using SsoAdvanced™ SYBR® Green Supermix (Bio-Rad, USA). The ΔΔCt method modified to account for primer efficiency and several endogenous genes (Vandesompele et al., 2002) was employed. The control genes employed were CAC (Solyc08g006960), EXP (Solyc07g025390), and SAND (Solyc03g115810) (). For each biological replicate, three technical replicates were performed for MM, and TO-937 sequences of CHS1 (Solyc09g091510), CHS2 (Solyc05g053550), and CHI1 (Solyc05g010320) were compared prior to primer design to avoid regions with mismatches. Primers employed and their efficiencies are presented in Supplementary Table 3.
2.4 Statistical analyses
IBM SPSS Statistics software was employed to carry out statistical analyses. T-tests were employed to compare the percentage of cuticle phenolics between the parental lines MM and TO-937. One-way analysis of variance (ANOVA) and Tukey-b tests were used to compare QTL lines and the parental line MM. Different letters indicate significant differences at p < 0.05. One-way ANOVA and Dunnett tests were employed to compare color and gene expression differences in TO-937 and single QTL lines against MM control. Two-way fixed-effects ANOVA with the two QTLs as main effects and MM and TO-937 homozygous alleles as the levels of those effects was employed to study the interaction between pairs of QTLs. For each combination, significance of the epistatic interaction was calculated as the p-value associated to the statistical double interaction between the two main effects. Also, two-way fixed-effects ANOVA was employed to study the effect of the interaction between genotype and season. Data are expressed as means ± standard error (SE).
3 Results
3.1 Effect of single QTL lines in the phenolic composition of the cuticle
Comparison of parental lines showed that TO-937 exhibited a significantly lower percentage of cuticle phenolics throughout development (Figure 1A) (p = 4 × 10−6 immature green, p = 2 × 10−5 MG). These differences were exacerbated at RR (p = 5 × 10−18), when TO-937 cuticle remained almost colorless compared to the orange color observed in MM (Figure 1B). At early stages of development, the phenolic acids p-coumaric and p-hydroxybenzoic acids accumulate in the cuticle, while at RR, the flavonoid chalconaringenin is also incorporated, conferring the observed orange color to the cuticle of MM (Figure 1C; , ). The colorless nature of the ripe cuticle of TO-937 (Figure 1B) can be considered an indication of little to no accumulation of chalconaringenin. In a previous work (), eight ILs harboring QTLs with a positive (ph4.1, ph5.1, ph5.2, and ph12.1) and a negative (ph1.1, ph7.1 + 7.2, ph8.1, and ph12.2) effect on percentage of cuticle phenolics were identified. However, cuticle phenolics include phenolic acids and chalconaringenin, which can only be fully extracted from the cuticle after cutin depolymerization. The harsh chemical conditions needed to depolymerize the cutin matrix causes chalconaringenin degradation, preventing its individual quantitation (). Since there is also an increase in phenolic acids during ripening, the difference in phenolic compounds between MG and RR cannot be considered a proxy for the chalconaringenin fraction. Thus, QTLs identified at RR could be associated with changes in either phenolic acids or chalconaringenin, or both. To overcome this limitation, CRM was employed to study the cuticles of the parental and single QTL lines.
Figure 1
Figure 2A shows Raman spectra within the 1,680–1,480 cm−1 region of RR isolated cuticles of parental lines together with the lines harboring QTLs for the percentage of cuticle phenolics. Deconvolution of this spectral region identified the contribution of four bands (Figure 2B), and their assignations according to the literature are shown in Supplementary Table 4. The band around 1,620 cm−1 was assigned to a phenolic fraction enriched in phenolic acids, mainly p-coumaric acid, while the 1,550 cm−1 band was associated to chalconaringenin (). The other two bands, around 1,585 and 1,602 cm−1, could be assigned to the interaction between phenolic compounds, either phenolic acids or chalconaringenin. A spectral shift in the band assigned to phenolic acids was observed between parental lines, while in MM, this band was present at 1,624 cm−1, and in TO-937, it was located at 1,633 cm−1. Interestingly, in the lines with QTLs for reduced percentage of cuticle phenolics (ph1.1, ph7.1 + 7.2, ph8.1, and ph12.2), located in the graph between TO-937 and MM, this band displayed an intermediate position between the parental lines, whereas in the lines with QTLs with a positive effect (ph4.1, ph5.1, ph5.2, and ph12.1, located above MM in the graph), the band showed the same position as MM (Figure 2A). This band shift is associated with changes in the chemical environment. Figure 2C shows the ratio between the areas of the 1,550 and 1,620 cm−1 bands (A1550/A1620), which can be considered an indication of the chalconaringenin/phenolic acids ratio. TO-937 exhibited an outstanding sixfold reduction of this ratio compared to MM. This low ratio, combined with the reduced percentage of total phenolics (Figure 1), indicates a preferential decrease in chalconaringenin, with phenolic acids being the main contributors to the phenolic domain present in the cuticle of TO-937. In a similar fashion to TO-937, QTL lines with a negative effect on cuticle phenolics, ph1.1, ph7.1 + 7.2, and ph12.2, showed a lower A1550/A1620 ratio compared to MM. On the other hand, QTL lines with a positive effect on cuticle phenolics, ph4.1, ph5.1, and ph5.2, displayed a higher A1550/A1620 ratio compared to MM, which is indicative of a preferential accumulation of chalconaringenin over phenolic acids. Interestingly, QTL lines ph8.1 and ph12.1, which harbor, respectively, QTLs with a negative and a positive effect on the percentage of total phenolics, showed A1550/A1620 ratios similar to MM. This indicates a proportional reduction (ph8.1) and increase (ph12.1) of phenolic acids and chalconaringenin.
Figure 2
Chalconaringenin confers orange color to the tomato fruit cuticle. Cuticle color analyses of the parental and single QTL lines showed remarkable differences in all color parameters: lightness (L*), saturation (chroma, C*), and tone (hue, h°) (Figure 2D, Supplementary Table 5). TO-937 cuticle color exhibited increased L* and h° and reduced C* compared to MM. The differences in color saturation and tone could be attributed to the differential accumulation of chalconaringenin, while the lightness increase could be partially a consequence of the lower amount of cuticle (). Similarly, all single QTL lines with reduced percentage of total phenolics showed a significant decrease (p < 0.001) in color saturation (C*) and a yellower tone (higher h° values). However, of the lines with QTL for enhanced cuticle phenolics, only ph4.1 displayed a significant effect on cuticle color, with increased saturation (p = 0.001) and lower hue (more orange) (p = 0.0004) than MM.
Chalconaringenin accumulation in ripe tomatoes is due to the combination of enhanced CHALCONE SYNTHASE (CHS) expression during ripening together with CHALCONE ISOMERASE (CHI) expression below the limit of detection, thus limiting the accumulation of downstream flavonoids (). Indeed, two CHS genes (CHS1 and CHS2) have been reported to be expressed in tomato epicarp and their expressions related to chalconaringenin incorporation to the cuticle of ripe tomatoes (). Although none of these genes locate within any of the QTL regions, the reported changes in total phenolics indicate that their expressions may be altered in some of the QTL lines. Thus, CHS and CHI expression levels were analyzed in the epicarp of fruits at breaker stage (Figures 2E–G). Comparison of parental lines showed a significant reduction of CHS1 (p = 9 × 10−17) and CHS2 (p = 2 × 10−18) expression in TO-937 compared to MM (Figures 2E, F; Supplementary Figure 1). Out of the four QTL lines with a positive effect on total phenolics, only ph12.1 showed expression level of both CHS genes higher than MM, in agreement with the increased percentage of total phenolics reported for this line (). Despite the fact that ph4.1 also increased total cuticle phenolics, expression levels of CHS1 (p = 3 × 10−4) and CHS2 (p = 9 × 10−8) were significantly lower than MM. The remaining positive-effect QTLs, ph5.1 and ph5.2, showed values similar to MM. On the other hand, QTL lines with a negative effect on total phenolics displayed CHS expression lower than MM. The lowest values were detected in ph1.1, the QTL with the highest additive effect, and the line that displayed the highest reduction of chalconaringenin/phenolic acid ratio (Figure 2D). Although slight changes between CHS1 and CHS2 expression were observed in several QTL lines, in general, both CHS genes exhibited a similar behavior, with the exception of ph7.1 + 7.2, where CHS1 expression was significantly reduced (p = 0.002) while CHS2 expression remained similar to MM.
Regarding CHI1, its expression during ripening was extremely low in MM, especially in comparison to CHS1 and CHS2 (Supplementary Figure 1), as expected from the literature (), but a surprising 22-fold increase was observed in TO-937 (Figure 2G). CHI1 expression in the QTL lines showed values similar to MM or lower. However, a remarkable 13-fold increase was detected in ph5.2, similar to the high CHI1 expression detected in TO-937 (Figure 2G).
3.2 Interaction between QTLs for percentage of cuticle phenolics
Thirty double and triple QTL lines were obtained through crossing single QTL lines followed by marker-assisted selection (Supplementary Table 1). Six and eight of them corresponded to combinations of QTLs with a positive and a negative effect, respectively, while 16 lines accumulated QTLs with an opposite effect (positive and negative) on cuticle phenolics. Figure 3 shows the percentage of cuticle phenolics at three stages of development for all the double and triple QTL lines, and their corresponding single QTL lines, expressed as fold changes relative to MM control. At RR, only one of the six lines pyramiding QTLs with a positive effect, ph5.2 + 12.1, showed a notable increase in cuticle phenolics compared to MM and the respective parental QTL lines. Of the eight lines pyramiding QTLs with a negative effect on cuticle phenolics, the lowest percentage was obtained with the triple QTL lines ph7.1 + 7.2 + 8.1 and ph7.1 + 7.2 + 12.2. It is surprising that ph1.1, despite having the highest additive effect in the RIL population, not only showed a small effect in the IL population (), but also had little contribution to cuticle phenolics in the double and triple lines pyramiding negative effect QTLs. The negative effect of ph1.1 was only observed in the combination ph1.1 + 12.1, where ph1.1 reduced the positive effect of ph12.1.
Figure 3
The above-mentioned results point to little additive effect and the presence of epistatic interactions between these QTLs. Figure 4 shows a heat map of the probability values of epistatic interactions for each of the double and triple QTL lines. As it can be clearly observed, at stages prior to ripening, most of the crosses involving ph5.1 and/or ph5.2 showed significant epistatic interactions, whereas the other QTL lines showed no epistatic interaction. At RR, however, over 86% of the double and triple QTL lines studied displayed a significant epistatic interaction. These results show that variation in cuticle phenolics is governed not only by additive QTL effects but also by strong non-additive interactions.
Figure 4
Figure 5 shows examples of the different interactions detected in the double and triple QTL lines. With the exception of ph7.1 + 7.2+ph8.1, with only additive effect, all the lines pyramiding QTLs for reduced cuticle phenolics showed epistatic interactions (Supplementary Figure 2). The triple QTL line ph7.1 + 7.2+ph12.2 displayed a combination of epistatic and additive effects, and the reduction in phenolics caused by ph12.2 was more evident when ph7.1 + 7.2 were present as MM alleles, as it can be seen by the different slopes. Epistatic interactions were observed between ph12.2 and ph1.1 or ph8.1, with the double QTL line showing similar values to ph12.2 in both cases. Accumulation of ph1.1 with ph8.1 or ph7.2 did not have an effect on cuticle phenolics, since the percentage of cuticle phenolics was similar to both parental lines. However, interaction of ph7.1 with ph1.1 showed that the double QTL line increased its percentage in phenolics compared to the single QTL lines, but without reaching the values of MM control. A similar effect was observed in the triple QTL line ph1.1+ph7.1 + 7.2, but in this case, percentage of cuticle phenolics was intermediate between parental QTL lines.
Figure 5
Regarding lines pyramiding QTLs with a positive effect on cuticle phenolics, only ph5.1+ph5.2 showed no effect. This was expected, since these single QTL lines only showed effect at stages prior to RR (Supplementary Figure 3). However, they did show epistatic interaction with ph12.1 and ph4.1 modifying their effects. Thus, ph5.1 interacted with ph4.1 and ph12.1 cancelling their effects. A similar behavior was observed in the interaction between ph4.1 and ph12.1, with the double QTL line displaying a lower percentage of cuticle phenolics than ph4.1. Conversely, ph5.2 did not modify the effect of ph4.1 and had a positive effect increasing the percentage of cuticle phenolics in combination with ph12.1.
Out of the 16 lines pyramiding QTLs with positive and negative effects on cuticle phenolics, 13 showed epistatic interactions (Supplementary Figures 4, 5). Double QTL lines ph12.1+ph8.1 and ph12.1+ph12.2 only displayed additive effects, with the former cancelling their opposite effects and resulting in values similar to MM and the latter displaying values intermediate between the single QTL lines, but lower than MM control. Similarly, a percentage of cuticle phenolics intermediate between parental QTL lines was observed in six other combinations (ph4.1+ph7.1 + 7.2, ph5.1+ph7.1 + 7.2, ph12.1+ph7.1 + 7.2, ph4.1+ph8.1, ph4.1+ph12.2, and ph5.2+ph12.2) with the double/triple QTL lines displaying values similar to or above MM control. The triple QTL line ph5.2+ph7.1 + 7.2 showed values similar to ph7.1 + 7.2, indicating that ph5.2 did not modify the effect of the negative QTL. An opposite behavior was observed in the double QTL lines pyramiding negative effect QTLs ph1.1 and ph8.1 with positive effect QTLs ph5.1 and ph5.2, as well as in ph12.2+ph5.1 and ph1.1+ph4.1. In all of these double lines, the negative effects of the single QTL were canceled. The percentage of cuticle phenolics was similar to the parental QTL line with positive effects.
Analysis of epistatic interactions for cuticle phenolics at IG and MG in the double and triple QTL lines involving ph5.1 and ph5.2 are presented in Supplementary Figures 6-9. With a few exceptions, the positive increase in phenolics caused by ph5.1 or ph5.2 was cancelled at both immature green and MG by the other QTLs.
The four QTLs that increase the percentage of cuticle phenolics overlapped to some extent with QTLs related to the amount of cuticle (). In three cases (ph4.1, ph5.1, and ph12.1), the associated QTL for the amount of cuticle (cm) showed an opposite effect to that observed for percentage of cuticle phenolics (Supplementary Figure 10). Therefore, it was possible that the increase in percentage of cuticle phenolics in these single QTL lines was caused by a reduction in the amount of cuticle. However, determination of the amount of cuticle in the double and triple QTL lines involving these four QTLs showed no clear relationship between the amount of cuticle and percentage of cuticle phenolics (Supplementary Figure 11).
As was mentioned previously, chalconaringenin is responsible for the orange color of the ripe tomato cuticle. Cuticle color was measured in MM and the double and triple QTL lines in order to determine color changes with respect to MM control using the CIE DE2000 (ΔE) equation. A clear relationship was observed between ΔE and fold change (relative to MM) in the percentage or amount of cuticle phenolics (Supplementary Figure 12). In both cases, but especially for the amount of phenolics, data were grouped following two different lines. As the relative amount or percentage of cuticle phenolics increased approaching 1 (values similar to MM), the differences in color decreased. The second group of data showed that as cuticle phenolics continued to increase (values higher than MM), color differences increased again. This is a clear indication that phenolics, especially their amount, influence the color of tomato fruit cuticle. The highest color differences of the double/triple QTL lines corresponded to negative × negative interactions, specifically those that caused the highest reduction in phenolics. Positive × positive interactions also caused color change, but it was not as manifest as the negative effects.
4 Discussion
4.1 Deciphering the complexity behind the QTLs for cuticle phenolics
Numerous strategies are currently available for dissecting the genetic basis behind phenotypic variation within natural populations. However, phenotypic variability is often quantitative, with several genes with variable effects involved, influenced by the environment and genetic factors. In this sense, cuticle biosynthesis has been reported to be modified by environmental conditions, as well as phenolic accumulation in plants (; ). Although genotype × environment interactions can influence phenotype (), results of the parental and single QTL lines showed similar effects within seasons (Supplementary Table 6) and were consistent with those previously reported (). A genotype × season interaction analysis showed a significant effect of the environment on the percentage of cuticle phenolics between winter and summer (p = 10−6), while no differences were detected between both summer seasons. However, this study across seasons was not carried out with the double and triple QTL lines. Thus, the epistatic effects reported here should be validated on different seasons prior to its use in breeding programs. It is worth mentioning that the percentage of cuticle phenolics did not change between seasons in TO-937 (Supplementary Table 1). This could be indicative of a low, constitutive level of cuticle phenolics, or that the effect of environmental conditions can vary among cultivars and/or in red-fruited wild species. Given the impact of climate change on crop growth and yield, this is a topic that deserves further study.
Many analyses of natural variability focus on major effects, yet, uncovering epistatic (non-additive) interactions is important for plant breeding since it allows the identification of the most beneficial interactions that produce a desirable phenotype within a given genetic background (). In the last years, numerous works have highlighted the importance of epistasis in traits such as inflorescence branching, plant height, flower sex determination, or metabolite accumulation (Shen et al., 2014; Soyk et al., 2017; Zebell et al., 2025; Zunjare et al., 2025). In this work, of the 30 double/triple QTL lines studied, only 7 (3 positive × positive and 4 negative × negative combinations) did not show an effect on the percentage of cuticle phenolics due to QTL accumulation, displaying values similar to the single QTL lines. Regarding epistasis, 80% of the studied crosses exhibited some degree of epistatic interaction, which is well above the only one previously identified using specialized software (). Moreover, the two QTLs with the highest percentage of variance within the RIL population, ph1.1 and ph8.1, only displayed a small reduction in cuticle phenolics in the IL population (), revealing epistatic interactions within the genetic background. These results underscore the practical implications of these analyses for crop engineering through the identification of advantageous genetic combinations and deepen our understanding on genetic interactions yielding similar phenotypic effects.
Cryptic genetic variation can have considerable effect on a phenotype under specific environmental conditions or different genetic backgrounds (). It is interesting to note that, although TO-937 displayed a very low percentage of phenolics within the cuticle, four QTLs with a positive effect of the TO-937 allele were identified (). However, two of these QTLs, ph5.1 and ph5.2, when placed in the MM background did not show any effect on cuticle phenolics at the RR stage, only at previous stages of development, leading to the idea that their effect might solely be attributed to minor changes in cuticle phenolics during growth, without a significant effect during ripening (). However, Raman spectroscopy analysis of ph5.1 and ph5.2 cuticles at RR showed an altered ratio of chalconaringenin/phenolic acid band ratio, which, together with the increased CHI1 expression level at breaker in ph5.2, is indicative of an effect on cuticle phenolics despite not being identifiable as percentage of cuticle phenolics at RR. Moreover, the combination of ph5.1 and ph5.2 with other QTLs allowed their effects on cuticle phenolics to be manifested, but only in specific genetic combinations. This points out to a negative interaction between MM background and ph5.1 and ph5.2 TO-937 alleles that buffered their effects.
Figure 6 summarizes the interactions found between QTLs for cuticle phenolics. The number of QTLs with relatively small effects, together with the complexity of the interactions among them and the presence of QTLs with positive and negative effects, hampered the identification of a genetic combination that reduced the percentage of cuticle phenolics to levels similar to those of TO-937. Out of the six positive × positive QTL combinations, the only combination of TO-937 introgressions that allowed an increase in cuticle phenolics compared to the single QTL lines was ph5.2+ph12.1, while the others showed no interaction or a negative one. However, none of the double negative × negative combinations was able to reduce the percentage of cuticle phenolics beyond the single QTL lines. This could only be achieved with two triple QTL combinations ph7.1 + 7.2+ph8.1 and ph7.1 + 7.2+ph12.2, none of which included ph1.1, the QTL with the highest additive effect on the RIL population. Indeed, a positive epistatic effect arose from the combination of ph1.1 and ph7.1, since a partial cancelation of their negative effects, which deserves further study, was observed. Most of the positive × negative QTL combinations showed an intermediate behavior between the single QTL lines that, in some cases, led to a mutual cancelation of their effects; however, in five of such combinations, only the negative effect was cancelled, whereas in ph1.1+ph12.1, the opposite occurred.
Figure 6
4.2 Implications of cuticle phenolics in the improvement of desirable traits
None of the structural genes involved in phenolic acid and flavonoid biosynthesis were located within any of the phenolic QTL regions reported () suggesting that the genes involved may be related to phenolic transport to the cuticle and/or vacuole, glycosylation, or regulation of the phenylpropanoid metabolic pathway, as it is the case of MYB12, the candidate gene for ph1.1 (; ; ). Hence, further dissection of the phenolic fraction of the parental and single QTL lines using CRM and expression analyses of genes involved in flavonoid biosynthesis has allowed us to delve into specific characteristics of cuticle phenolic accumulation that could lead to a better approach to identify candidate genes behind these genomic regions. Hence, by targeting the inner cuticle side, the phenolic fraction accumulated during ripening could be studied. The highest reduction in CHS expression and chalconaringenin/phenolic acid ratio observed in ph1.1 is in accordance with this QTL displaying the highest additive effect in the RIL population, but does not explain its moderate reduction in total cuticle phenolics (). Silencing MYB12, the candidate gene behind this QTL, altered the phenylpropanoid pathway, affecting both chalconaringenin and phenolic acid levels within the cuticle (). This suggests that the TO-937 allele of MYB12 has a clear effect on chalconaringenin accumulation within the cuticle, but little effect on phenolic acids, thus explaining the low reduction in total cuticle phenolics exhibited by ph1.1. It is possible then that the effect on phenolic acids was only measurable when MYB12 is silenced. On the other hand, ph7.1 + 7.2 and ph12.2, the QTL lines that displayed the highest negative effect on total cuticle phenolics, had a more equilibrated effect on both chalconaringenin and phenolic acids, positing that their highest reduction in total cuticle phenolics is the consequence of an effect in both phenolic fractions. A special case is ph4.1, a QTL with increased percentage of cuticle phenolics, a high A1550/A1620 band ratio indicative of a chalconaringenin enriched fraction, but low CHS1 and CHS2 expression. This seems to point to a preferential transport of chalconaringenin to the cuticle.
Identification of genomic regions that enrich the cuticle in phenolic compounds and/or increase the fraction of chalconaringenin are important for improving the thermal and mechanical properties of the cuticle. Recently, variability for the glass transition temperature (Tg) in RR tomatoes has been reported, which was associated with the percentage of cuticle phenolics (). Thus, it was postulated that a higher percentage of cuticle phenolics increases Tg due to their contribution to the hydrogen bond network within the cuticle. MM has a low Tg, approximately 8°C, which implies that, above this temperature, the cuticle exhibits a mechanically weaker behavior and higher permeability (). On the other hand, chalconaringenin has been shown to increase cuticle’s elasticity, conferring mechanical resistance (; ). Hence, increasing the percentage of cuticle phenolics could be a strategy to improve the mechanical resistance and water barrier properties during storage and on the vine. In this sense, molecular markers herein employed to identify QTLs could be used to introduce TO-937 genomic regions such as ph4.1 or ph12.1 within the domesticated tomato and generate an ideotype with a cuticle enriched in phenolic compounds for breeding purposes.
Tomato CHS1 and CHS2 have notable sequence similarity and comparable expression profile during ripening, which has precluded the study of their individualized effects (). The specific reduction in CHS1 expression, without affecting CHS2 levels, in combination with the significant effect on cuticle color observed in ph7.1 + 7.2, suggests that CHS1 could be responsible for the synthesis of chalconaringenin that is transported to the cuticle. It has been reported that CHI1 is not expressed in the peel of ripening domesticated tomatoes, leading to the accumulation of chalconaringenin in ripe fruits (). However, CHI1 transcript has been detected in the peel of green-fruited wild tomato species (Willits et al., 2005). Hence, CHI1 expression during ripening in TO-937, an accession of the red-fruited species S. pimpinellifolium, is remarkable. This could posit that the lack of chalconaringenin in TO-937 fruit cuticle is due to the restoration of the flavonoid metabolic pathway leading to the accumulation of glycosylated flavonols in fruits, as it was reported after overexpression of a Petunia CHI gene in tomato (). Thus, CHI expression would reduce the amount of chalconaringenin available to be transported and render the cuticle colorless. However, the high levels of CHI1 expression during ripening in the QTL line ph5.2, together with the orange color of its ripe fruit cuticle, and a percentage of cuticle phenolics similar to MM, indicate otherwise. CHI1 expression in ripening fruit peel was low compared to that of CHS1 and CHS2; hence, the cuticle phenolic phenotype identified in ph5.2 is probably due to a low rate of chalconaringenin conversion to its flavone isomer (Figure 1C). Nevertheless, this genomic region could be of interest to increase flavonol content in fruits without compromising the biophysical properties that chalconaringenin accumulation confers to the cuticle and could have beneficial agronomical implications. Thus, identification of candidate genes for ph5.2 and ph12.1 and understanding their negative interaction with MYB12 (ph1.1) would not only allow us to better comprehend the regulatory pathway of phenolic accumulation but also identify strategies to improve their accumulation within the cuticle.
Statements
Data availability statement
Data generated and analyzed in this study is available in the Zenodo repository, 10.5281/zenodo.21534257.
Author contributions
RB: Data curation, Investigation, Writing – review & editing, Writing – original draft, Formal analysis. MU: Formal analysis, Data curation, Investigation, Writing – original draft. JM: Investigation, Formal analysis, Writing – review & editing. AG: Writing – review & editing, Methodology, Investigation, Formal analysis. AH: Funding acquisition, Project administration, Data curation, Writing – review & editing, Validation. RF-M: Supervision, Conceptualization, Funding acquisition, Writing – review & editing, Resources. ED: Conceptualization, Project administration, Data curation, Funding acquisition, Writing – original draft.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This work was funded by project PID2024-160692OB-I00 of the MCIN/AEI/10.13039/501100011033 Spain, co-financed by ERDF “A way of making Europe”. MU was supported by the EMERGIA DGP_EMEC_2023_00020 grant from Junta de Andalucía, Spain. JCMA was recipient of the predoctoral contract PRE2019-087788 from MCIN/AEI.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2026.1863369/full#supplementary-material
References
1
AdatoA.MandelT.Mintz-OronS.VengerI.LevyD.YativM.et al. (2009). Fruit-surface flavonoid accumulation in tomato is controlled by a SlMYB12-regulated transcriptional network. PloS Genet.5, e1000777. doi: 10.1371/journal.pgen.1000777
2
AlkanN.FortesA. M. (2015). Insights into molecular and metabolic events associated with fruit response to post-harvest fungal pathogens. Front. Plant Sci.6, 889. doi: 10.3389/fpls.2015.00889
3
BallesterA. R.MolthoffJ.de VosR.Lintel HekkertB.OrzaezD.Fernández-MorenoJ. P.et al. (2010). Biochemical and molecular analysis of pink tomatoes: deregulated expression of the gene encoding transcription factor SlMYB12 leads to pink tomato fruit color. Plant Physiol.152, 71–84. doi: 10.1104/pp.109.147322
4
Barraj BarrajR.SegadoP.Moreno-GonzálezR.HerediaA.Fernández-MuñozR.DomínguezE. (2021). Genome-wide QTL analysis of tomato fruit cuticle deposition and composition. Hortic. Res.8, 113. doi: 10.1038/s41438-021-00548-5
5
BarrantesW.López-CasadoG.García-MartínezS.AlonsoA.RubioF.RuizJ. J.et al. (2016). Exploring new alleles involved in tomato fruit quality in an introgression line library of Solanum pimpinellifolium. Front. Plant Sci.7, 1172. doi: 10.3389/fpls.2016.01172
6
BenítezJ. J.González MorenoA.Guzmán-PuyolS.Heredia-GuerreroJ. A.HerediaA.DomínguezE. (2022). The response of tomato fruit cuticle membranes against heat and light. Front. Plant Sci.12, 807723. doi: 10.3389/fpls.2021.807723
7
BenítezJ. J.López-CasadoG.SegadoP.HerediaA.DomínguezE. (2026). Cuticle changes during tomato fruit postharvest: implications of natural variability in biophysical properties. Postharvest Biol. Technol.235, 114178. doi: 10.1016/j.postharvbio.2026.114178
8
BorisjukN.HrmovaM.LopatoS. (2014). Transcriptional regulation of cuticle biosynthesis. Biotechnol. Adv.32, 526–540. doi: 10.1016/j.bioteChadv.2014.01.005
9
ButelliE.TittaL.GiorgioM.MockH. P.MatrosA.PeterekS.et al. (2008). Enrichment of tomato fruit with health-promoting anthocyanins by expression of select transcription factors. Nat. Biotechnol.26, 1301–1308. doi: 10.1038/nbt.1506
10
CohenS. D.KennedyJ. A. (2010). Plant metabolism and the environment: implications for managing phenolics. Crit. Rev. Food Sci. Nutr.50, 620–643. doi: 10.1080/10408390802603441
11
CooperM.van EeuwijkF. A.HammerG. L.PodlichD. W.MessinaC. (2009). Modeling QTL for complex traits: detection and context for plant breeding. Curr. Opin. Plant Biol.12, 231–240. doi: 10.1016/j.pbi.2009.01.006
12
DavuluriG. R.van TuinenA.FraserP. D.ManfredoniaA.NewmanR.BurgessD.et al. (2005). Fruit-specific RNAi-mediated suppression of DET1 enhances carotenoid and flavonoid content in tomatoes. Nat. Biotechnol.23, 890–895. doi: 10.1038/nbt1108
13
DomínguezE.EspañaL.López-CasadoG.CuarteroJ.HerediaA. (2009). Biomechanics of isolated tomato fruit cuticles during ripening: the role of flavonoids. Funct. Plant Biol.36, 613–620. doi: 10.1071/FP09039
14
DwivediS. L.Heslop-HarrisonP.AmasJ.OrtizR.EdwardsD. (2024). Epistasis and pleiotropy-induced variation for plant breeding. Plant Biotechnol. J.22, 2788–2807. doi: 10.1111/pbi.14405
15
EspañaL.Heredia-GuerreroJ. A.Reina-PintoJ. J.Fernández-MuñozR.HerediaA.DomínguezE. (2014a). Transient silencing of CHALCONE SYNTHASE during fruit ripening modifies tomato epidermal cells and cuticle. Plant Physiol.166, 1371–1386. doi: 10.1104/pp.114.246405
16
EspañaL.Heredia-GuerreroJ. A.SegadoP.BenítezJ. J.HerediaA.DomínguezE. (2014b). Biomechanical properties of tomato fruit cuticle during development are modulated by changes in the relative amount of their components. New Phytol.202, 790–802. doi: 10.1111/nph.12727
17
Expósito-RodríguezM.BorgesA. A.Borges-PérezA.PérezJ. A. (2008). Selection of internal control genes for quantitative real-time RT-PCR studies during tomato development process. BMC Plant Biol.8, 131. doi: 10.1186/1471-2229-8-131
18
Fernández-MorenoJ. P.Levy-SamohaD.MalitskyS.MonforteA. J.OrzaezD.AharoniA.et al. (2017). Uncovering tomato quantitative trait loci and candidate genes for fruit cuticular lipid composition using the Solanum pennellii introgression line population. J. Exp. Bot.68, 2703–2716. doi: 10.1093/jxb/erx134
19
GibsonG.DworkinI. (2004). Uncovering cryptic genetic variation. Nat. Rev. Genet.5, 681–690. doi: 10.1038/nrg1426
20
GiménezE.DomínguezE.PinedaB.HerediaA.MorenoV.LozanoR.et al. (2015). Transcriptional activity of the MADS Box ARLEQUIN/TOMATO AGAMOUS-LIKE1 gene is required for cuticle development of tomato fruit. Plant Physiol.68, 1036–1048. doi: 10.1104/pp.15.00469
21
González MorenoA.DomínguezE.MayerK.XiaoN.BockP.HerediaA.et al. (2023). 3D (x-y-t) Raman imaging of tomato fruit cuticle: microchemistry during development. Plant Physiol.191, 219–232. doi: 10.1093/plphys/kiac369
22
González MorenoA.de CózarA.PrietoP.DomínguezE.HerediaA. (2022). Radiationless mechanism of UV deactivation by cuticle phenolics in plants. Nat. Commun.13, 1786. doi: 10.1038/s41467-022-29460-9
23
GosneyB. J.PottsB. M.O'Reilly-WapstraJ. M.VaillancourtR. E.FitzgeraldH.DaviesN. W.et al. (2016). Genetic control of cuticular wax compounds in Eucalyptus globulus. New Phytol.209, 202–215. doi: 10.1111/nph.13600
24
HullH. M.MortonH. L.WharrieJ. R. (1975). Environmental influences on cuticle development and resultant foliar penetration. Bot. Rev.41, 421–452. doi: 10.1007/BF02860832
25
HuntG. M.BakerE. A. (1980). Phenolic constituents of tomato fruit cuticles. Phytochemistry19, 1415–1419. doi: 10.1016/0031-9422(80)80185-3
26
KaplanY.WangY.ManasherovaE.CohenH.GinzbergI. (2024). Metabolic and gene-expression analyses reveal developmental dynamics of cutin deposition in pomegranate fruit grown under different environmental conditions. Plant Physiol. Biochem.215, 108991. doi: 10.1016/j.plaphy.2024.108991
27
Kögel-KnabnerI.de LeeuwJ. W.TegelaarE. W.HatcherP. G.KerpH. (1994). A lignin-like polymer in the cuticle of spruce needles: implications for the humification of spruce litter. Org. Geochem.21, 1219–1228. doi: 10.1016/0146-6380(94)90165-1
28
LinM.MatschiS.VasquezM.ChamnessJ.KaczmarN.BaseggioM.et al. (2020). Genome-wide association study for maize leaf cuticular conductance identifies candidate genes involved in the regulation of cuticle development. G3: Genes Genomes Genet.10, 1671–1683. doi: 10.1534/g3.119.400884
29
ManasherovaE.FeygenbergO.MaurerD.AlkanN.CohenH. (2025). Postharvest storage affects the chemical composition of the avocado fruit cuticle in a cultivar-specific manner. Food Chem.493, 145837. doi: 10.1016/j.foodchem.2025.145837
30
MuirS. R.CollinsG. J.RobinsonS.HughesS.BovyA.Ric De VosC. H.et al. (2001). Overexpression of petunia chalcone isomerase in tomato results in fruit containing increased levels of flavonols. Nat. Biotechnol.19, 470–474. doi: 10.1038/88150
31
OfnerI.LashbrookeJ.PlebanT.AharoniA.ZamirD. (2016). Solanum pennellii backcross inbred lines (BILs) link small genomic bins with tomato traits. Plant J.87, 151–160. doi: 10.1111/tpj.13194
32
ParsonsE. P.PopopvskyS.LohreyG. T.LüS.Alkalai-TuviaS.PerzelanY.et al. (2012). Fruit cuticle lipid composition and fruit post-harvest water loss in an advanced backcross generation of pepper (Capsicum sp.). Physiol. Plant146, 15–25. doi: 10.1111/j.1399-3054.2012.01592.x
33
PetitJ.BresC.JustD.GarciaV.MauxionJ. P.MarionD.et al. (2013). Analyses of tomato fruit brightness mutants uncover both cutin-deficient and cutin-abundant mutants and a new hypomorphic allele of GDSL lipase. Plant Physiol.164, 888–906. doi: 10.1104/pp.113.232645
34
Popovsky-SaridS.BorovskyY.FaigenboimA.ParsonsE. P.LohreyG. T.Alkalai-TuviaS.et al. (2017). Genetic and biochemical analysis reveals linked QTLs determining natural variation for fruit post-harvest water loss in pepper (Capsicum). Theor. Appl. Genet.130, 445–459. doi: 10.1007/s00122-016-2825-9
35
RautengartenC.EbertB.OuelletM.NafisiM.BaidooE. E.BenkeP.et al. (2012). Arabidopsis Deficient in Cutin Ferulate encodes a transferase required for feruloylation of omega-hydroxy fatty acids in cutin polyester. Plant Physiol.158, 654–665. doi: 10.1104/pp.111.187187
36
RenaultH.AlberA.HorstN. A.Basilio LopesA.FichE. A.KriegshauserL.et al. (2017). A phenol-enriched cuticle is ancestral to lignin evolution in land plants. Nat. Commun.8, 14713. doi: 10.1038/ncomms14713
37
Rett-CadmanS.ColleM.MansfeldB.BarryC. S.WangY.WengY.et al. (2019). QTL and transcriptomic analyses implicate cuticle transcription factor SHINE as a source of natural variation for epidermal traits in cucumber fruit. Front. Plant Sci.10, 1536. doi: 10.3389/fpls.2019.01536
38
ReynoudN.GeneixN.PetitJ.D’OrlandoA.FanuelM.MarionD.et al. (2022). The cutin polymer matrix undergoes a fine architectural tuning from early tomato fruit development to ripening. Plant Physiol.190, 1821–1840. doi: 10.1093/plphys/kiac392
39
SharmaG.WuW.DalalE. N. (2005). The CIEDE2000 color-difference formula: implementation notes, supplementary test data, and mathematical observations. Color Res. Appl.30, 21–30. doi: 10.1002/col.20070
40
ShenG.ZhanW.ChenH.XingY. (2014). Dominance and epistasis are the main contributors to heterosis for plant height in rice. Plant Sci.215-216, 11–18. doi: 10.1016/j.plantsci.2013.10.004
41
SoykS.LemmonZ. H.OvedM.FisherJ.LiberatoreK. L.ParkS. J.et al. (2017). Bypassing negative epistasis on yield in tomato imposed by a domestication gene. Cell.169, 1142–1155. doi: 10.1016/j.cell.2017.04.032
42
VandesompeleJ.De PreterK.PattynF.PoppeB.Van RoyN.De PaepeA.et al. (2002). Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol.3, research0034. doi: 10.1186/gb-2002-3-7-research0034
43
WillitsM. G.KramerC. M.PrataR. T. N.de LucaV.PotterB. G.SteffensJ. C.et al. (2005). Utilization of the genetic resources of wild species to create a nontransgenic high flavonoid tomato. J. Agric. Food. Chem.53, 1231–1236. doi: 10.1021/jf049355i
44
WuX.ChenY.ZhuJ.ZhangN.WeiY.JiangS.et al. (2023). 1-Methylcyclopropene reduces postharvest water loss by modulating cuticle formation in tomato fruit. Postharvest Biol. Technol.206, 112564. doi: 10.1016/j.postharvbio.2023.112564
45
YangH.ZouY.LiX.ZhangM.ZhuZ.XuR.et al. (2022). QTL analysis reveals the effect of CER1-1 and CER1-3 to reduce fruit water loss by increasing cuticular wax alkanes in citrus fruit. Postharvest Biol. Technol.185, 111771. doi: 10.1016/j.postharvbio.2021.111771
46
YeatsT. H.RoseJ. K. C. (2013). The formation and function of plant cuticles. Plant Physiol.163, 5–20. doi: 10.1104/pp.113.222737
47
ZebellS. G.Martí-GómezC.FitzgeraldB.CunhaC. P.LachM.SemanB. M.et al. (2025). Cryptic variation fuels plant phenotypic change through hierarchical epistasis. Nature64, 984–992. doi: 10.1038/s41586-025-09243-0
48
ZunjareR. U.MuthusamyV.ChandG.ChauhanP.RathoreA. S.KatralA.et al. (2025). Genetic evidence of recessive epistasis among sex determining silkless1 (sk1) tasselseed1 (ts1) and tasselseed2 (ts2) genes and its utility in maize (Zea mays L.) breeding. Plant Breed.144, 97–106. doi: 10.1111/pbr.13224
Summary
Keywords
CHI, CHS, epistasis, phenolics, plant cuticle, Raman spectroscopy, tomato
Citation
Barraj Barraj R, Urrutia M, Mateos del Amo JC, González Moreno A, Heredia A, Fernández-Muñoz R and Domínguez E (2026) QTL pyramiding uncovers epistatic interactions that modify cuticle phenolics in tomato. Front. Plant Sci. 17:1863369. doi: 10.3389/fpls.2026.1863369
Received
23 April 2026
Revised
03 July 2026
Accepted
07 July 2026
Published
03 August 2026
Volume
17 - 2026
Edited by
Rebecca Grumet, Michigan State University, United States
Reviewed by
Johann Petit, Institut National de recherche pour l’agriculture, l’alimentation et l’environnement (INRAE), France
Fikru Kenea, Dilla University, Ethiopia
Updates
Copyright
© 2026 Barraj Barraj, Urrutia, Mateos del Amo, González Moreno, Heredia, Fernández-Muñoz and Domínguez.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Eva Domínguez, edominguez@eelm.csic.es
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.