Physiological and redox status properties of a new late-ripening triploid hybrid mandarin: a seasonal field evaluation

In Corsica, citrus cultivation represents a major economic sector, largely based on high-quality products such as ‘Clementine de Corse’ and ‘Pomelo de Corse’, both of which hold a Protected Geographical Indication (PGI). However, the harvesting periods of these cultivars create a recurrent production gap between winter and spring, during which imported citrus fruits dominate the market, increasing competitive pressure on local producers. The development of late-ripening citrus varieties adapted to Mediterranean pedoclimatic conditions therefore represents a key strategy to strengthen local production and extend market availability. In this context, the present study evaluates the seasonal physiological and antioxidant responses of a novel triploid mandarin hybrid (designated W9) grafted onto C35 citrange, in comparison with two reference diploid cultivars, clementine and pomelo (‘Star Ruby’ grapefruit), grown under open-field conditions Seasonal parameters were monitored across four seasons, including gas exchange (P net, g s, E, iWUE), chlorophyll fluorescence (F v /F m), oxidative stress markers (H 2 O 2, MDA, proline), ascorbate redox components (Asa, DHA, Asa/DHA), and major antioxidant enzymes (SOD, CAT, APX, DHAR). Principal component analysis revealed a strong seasonal structuring of physiological responses, driven primarily by photosynthetic performance and redox metabolism. All genotypes exhibited reduced photosynthetic activity during the cold season. The triploid W9 consistently maintained lower H 2 O 2 and MDA levels, particularly during spring, its harvest period. This was associated with a higher Asa/DHA ratio and increased activities of antioxidant enzymes, notably CAT, APX, and DHAR. These results suggest that triploidy at the scion level confers improved redox homeostasis without drastically altering photosynthetic capacity, thereby enhancing tolerance to chilling during the winter-to-spring transition. Overall, triploid W9 emerges as a promising late-ripening citrus genotype for Mediterranean production systems, with potential benefits for fruit quality, harvest stability, and reinforcement of the Corsican citrus value chain.

Citrus (Citrus spp.) is one of the main cultivated fruit crops in the subtropical regions and is highly appreciated by consumers. According to FAO data, world citrus production is around 165.6 Mt in 2024 (FAOSTAT, 2026). China and Brazil are the major producers, representing 30% and 11%, respectively, while Europe accounts for 6.4%, with Spain accounting for 50% of this production. In contrast to global production, which follows a linear progression of 2 Mt/year, European production has stagnated since the 2000s.

The Mediterranean basin is historically characterized by a cool climate with mild, frost-free winters and predominantly acidic soils, providing favorable conditions for citrus cultivation. In Corsica, an island located between southern France and Italy, citrus production is of major economic importance, with traditional crops such as clementine (Citrus clementina Hort. ex Tan) and pomelo (Citrus paradisi Macf., ‘Star Ruby’ grapefruit), which together represent the majority of citrus production. More than 80% of clementines and 67% of pomelos are marketed under Protected Geographical Indication (PGI) (AGRESTE, 2025). The Corsican citrus economy is largely based on the high-quality products resulting from long-standing traditional expertise. The proximity to markets and fresh quality attributes, such as the presence of leaves at harvest, contribute to a positive image of Corsican citrus fruits. In 2024, annual production reached 47,000 tons for clementines and 7,680 tons for pomelos (AGRESTE, 2025). Citrus production continues to expand compared to the previous year, with 50 ha and 14 ha of new plantations added to the existing 1,707 ha of clementine and 216 ha of pomelo orchards, respectively (AGRESTE, 2025).

However, since the expansion of industrialization, anthropogenic activity has expanded continuously, gradually impacting the environment. This climate change represents one of the most critical challenges for global agriculture. Different abiotic stresses, such as increasing temperatures, prolonged drought periods, soil salinity modification or extreme weather events are altering the balance of agroecosystems and thus directly threatening the productivity and sustainability of perennial crops (IPCC, 2023). Even if greenhouse gas emissions could be drastically reduced, the long-term inertia of the climate system ensures that environmental disturbances will continue for decades or centuries. The Mediterranean basin is no exception to this trend. Projected rising temperatures pose a significant challenge to citrus production, as these crops rely on low-temperature exposure to support optimal fruit maturation and peel degreening. In fact, low temperatures (especially cold nights) allow the degradation of chlorophyll in the flavedo and stimulate the synthesis of carotenoids, which gives the vibrant orange or yellow color of citrus fruits (Purvis and Barmore, 1981; Mitalo et al., 2022). Moreover, climate change and increasing abiotic stresses are altering phenology, fruit size and plant physiology, with direct negative effects on agronomic behavior. In order to face this climatic shift, varietal improvement is essential to enhance tolerance and maintain fruit quality (Santini et al., 2013; Lourkisti et al., 2020). In citrus orchards, the rootstock selection influences scion development, fruit quality, and resilience against environmental conditions. Recent studies have highlighted how distinct citrus rootstocks exhibit highly specialized physiological, biochemical, and anatomical adaptations when subjected to extreme abiotic constraints, such as severe drought periods (Shafqat et al., 2021) or elevated heat stress (Ur Rehman et al., 2024). Polyploidy, a common biological phenomenon (Chen, 2007; Soltis and Soltis, 2009), has emerged as a genetic strategy to improve stress tolerance and broaden phenotypic diversity in fruit crops (Allario et al., 2013; Sobhanian et al., 2026). Polyploid plants harboring additional chromosome sets often display thicker leaves, altered stomatal density, enhanced water-use efficiency, and improved reactive oxygen species (ROS) management (Oustric et al., 2017; Tossi et al., 2022). More specifically, tetraploid rootstocks have been extensively studied for their role in drought and salinity resistance (Allario et al., 2013; Oustric et al., 2017, 2021; Wei et al., 2020) or HLB resistance (Grosser et al., 2015). In addition, genetic improvement strategies targeting the scion have been developed to address both agronomic performance and fruit quality. In this context, controlled hybridization between diploid and tetraploid citrus parents has enabled the development of numerous triploid genotypes that are mostly seedless and display enhanced physiological stability under stress (Aleza et al., 2012; Ramakrishnan et al., 2025). Triploidy has gained increasing attention in breeding programs because seedlessness is a highly appreciated organoleptic characteristic for consumers. Traditionally, the scion is selected for its capacity to produce high-quality fruit, whereas the rootstock dictates agronomic performance. Crop tolerance to abiotic stress involves complex root-to-shoot communication and is, for instance, known to be largely rootstock-driven. This is evidenced by the positive influence of Carrizo citrange rootstocks on Cleopatra mandarin scion compared to other rootstocks (Balfagón et al., 2022). However, triploid scion citrus hybrids such as ‘Safor’ (Cuenca et al., 2010), ‘Tacle’ (Recupero et al., 2005) or ‘Alborea’ (Cuenca et al., 2020) have demonstrated promising agronomic traits. These attributes suggest that triploidy at the scion level plays a significant role in overall plant physiology and stress tolerance in citrus (Lourkisti et al., 2020; Ramakrishnan et al., 2025). Nevertheless, the physiological and biochemical consequences of polyploidy at the scion level still remain poorly documented. Scion polyploidy represents a major varietal improvement strategy to create highly tolerant crops and qualitative fruits to face both climatic and economic pressure.

In this specific context, the Corsican citrus economy faces strong competition from countries where production volumes are considerably higher. This is particularly visible between the end of the Clementine harvest (december–january) and the beginning of the Pomelo harvest (may–june) when imported fruits flood the local market. This gap represents a low-production period tending to increase due to climate change. In this context, the development of late-ripening varieties adapted to local climatic and soil conditions could address several key challenges: (i) extending the harvesting period, (ii) reinforcing production identity through quality schemes, (iii) reducing pressure on currently cultivated varieties and (iv) alleviating the negative effects of global climate change.

To meet these objectives, a hybridization program between Fortune mandarin (Citrus reticulata Blanco) and Ellendale tangor (Citrus reticulata Blanco x Citrus sinensis (L.) Osb.) was initiated in 2000s, and the resulting hybrids were planted in an experimental orchard (The INRA-Cirad research center in San-Giuliano, Corsica) in 2006 (Ahmed et al., 2020). The Fortune mandarin, used as the female parent, was selected for its desirable pomological traits and its ability to generate a high proportion of triploid offspring (Aleza et al., 2010), while Ellendale tangor was chosen for its superior organoleptic quality (Simón-Grao et al., 2014) and its late fruit maturity (April). Prior investigations demonstrated that these 16-years-old triploid citrus hybrids exhibited an enhanced tolerance under natural chilling conditions (Lourkisti et al., 2020). Furthermore, these triploid hybrids showed a greater capacity to recover from water stress deficit (Lourkisti et al., 2021a, 2022). They displayed improved carboxylation efficiency and restored water-related parameters, likely associated with polyploidy-induced anatomical and cytological changes (Lourkisti et al., 2021b). While these two studies provided the first insights into the physiological and biochemical responses of triploid hybrids under natural and controlled stress conditions, they also highlighted the need to evaluate their behavior under natural field conditions over growing seasons. Evaluating their performance under real agronomic conditions is indeed essential to understand how these new varieties adapt to Mediterranean pedoclimatic constraints. Based on these previous findings, the present study focused on a triploid hybrid, named W9, resulting from the same breeding program and parental cross as the hybrids previously investigated. This genotype was selected by scientific experts and producers, through sensory and pomological evaluation. The W9 genotype was grafted onto C35 citrange rootstock (Citrus sinensis ‘Ruby Blood’ x Poncirus trifoliata), which is widely used for clementine grafting (Hussain et al., 2013), and planted in a cultivated plot to evaluate its performance under commercial production conditions. C35 citrange is well known to confer resistance to Phytophthora and tristeza virus (Cameron and Soost, 1986), but it also provides tolerance to natural chilling conditions (Oustric et al., 2017) and modulates scion defenses against abiotic stresses such as heat or drought (Zandalinas et al., 2017; Balfagón et al., 2022).

Accordingly, the aim of the present study was to evaluate the physiological and biochemical behavior of this novel triploid mandarin variety ‘W9’, grafted onto the diploid rootstock C35 citrange, under open-field conditions. Photosynthetic parameters (P net, E, g s, iWUE, Fv/Fm) and biochemical markers related to redox status (MDA, H 2 O 2, proline, Asa, DHA, tAsa, Asa/DHA, SOD, CAT, APX, DHAR) were compared with those of two reference diploid cultivars, Clementine and Pomelo, grafted onto the same rootstock. Given the perennial nature of citrus trees and the strong seasonal modulation of physiological activity and redox metabolism (Ribeiro and MaChado, 2007), a four-season monitoring was conducted to characterize genotype-specific responses under fluctuating natural environmental conditions. This approach also assesses the potential advantages conferred by triploidy at the scion level, particularly regarding redox homeostasis under field conditions.

Plant material and experimental design

The experiment was conducted on three citrus varieties cultivated in an orchard (Aléria, Corsica 42.13° N, 9.47° E): diploid Clementine (Citrus clementina Hort ex. Tan), diploid Pomelo (Citrus paradisi Macf. ‘Star Ruby’ grapefruit) and triploid W9 (Citrus reticulata Blanco x (Citrus reticulata Blanco x Citrus sinensis (L.) Osb)), all grafted onto a diploid citrange C35 rootstock (Citrus sinensis ‘Ruby Blood’ x Poncirus trifoliata). The trees used were 8-years-old, chosen according to uniform health status and canopy size. The same trees were used for all sampling periods (3 Clementine, 3 Pomelo and 3 W9). Physiological measurements and leaf sampling were conducted at four different time points representing the four annual seasons during sunny days: Summer (3 rd of July, 2023, Tmean: 25.1 °C), Autumn (18 th November 2024, Tmean: 18.5 °C), Winter (February 2 nd, 2025, Tmean: 9.3 °C) and Spring (April 4 th, 2025, Tmean: 14.3 °C) (Table 1). For each variety, physiological measurements and the harvesting of fully expanded mature leaves were conducted from four distinct, southeast-exposed canopy zones (bottom, center-left, center-right, and top). For each zone, leaves and measurements from three different trees were pooled to form one composite biological replicate, resulting in four biological replicates per variety.

Photosynthetic measurements

Gas exchange parameters were measured on intact leaves between 9:00 and 11:00 using a portable infrared gas analyser (e.g., LC-PRO-SD, ADC BioScientific, UK) (Lourkisti et al., 2020). Net photosynthetic rate (P net), stomatal conductance (g s), and transpiration rate (E) were recorded under a constant photosynthetically active radiation (PAR) of 1400 μmol m − ² s − ¹, leaf temperature of 25 °C, and ambient CO 2 concentration around 380 ppm. Intrinsic water use efficiency (iWUE) was calculated as the ratio between P net and g s (Supplementary Figure 2). Chlorophyll fluorescence parameters were obtained using a pulse-amplitude-modulated fluorimeter (e.g., OS1p, Opti-Sciences, USA). The maximum quantum efficiency of PSII (Fv/Fm) was measured after 30 minutes of dark adaptation.

Biochemical analyses

Four biological replicates were analysed for each genotype or treatment, each consisting of a pool of fully expanded leaves rapidly frozen in liquid nitrogen, ground to a fine powder with a Mixer Mill MM 500 Control (Retsch, Verder, Germany), and stored at −80 °C until use. Furthermore, three technical replicates were performed for each biological replicate.

Malondialdehyde (MDA) content was quantified following the modified thiobarbituric acid protocol adapted for citrus leaves (Hodges et al., 1999; Santini et al., 2013). Briefly, 80 mg of fine leaf powder was homogenized in 2 mL of 80% (v/v) ethanol. After 15 min 3000 x g centrifugation at 4 °C, suspension was split in two new tubes with 20% trichloroacetic acid (TCA) or 0.65% TBA + 20% TCA (w/v), respectively. Absorbance was recorded at 440, 535, and 600 nm with a V-630 spectrophotometer (Jasco Inc., Tokyo, Japan).

Where OD nm+/-TBA represents the optical density at wavelength nm (with or without the addition of thiobarbituric acid (TBA), 0.0571*(OD 440+TBA – OD 600+TBA corresponds to the correction factor for sucrose interference, and 157,000 (L.mol -1.cm -1) is the molar extinction coefficient of the MDA-TBA complex. To express the final concentration relative to the plant tissue fresh weight (FW), the following dilution and sample mass factor was applied (40 mg of fresh sample tissue extracted per mL, yielding a multiplication factor of 25).

Hydrogen peroxide (H 2 O 2) concentration was determined spectrophotometrically using the xylenol orange method (Jiang et al., 1991) using a PeroxiDetect (Sigma, Aldrich, St. Louis, MO, États-Unis) kit with 150 mg of fresh powder was homogenised in 600 µL distilled water. Then, a 30 min centrifugation at 13000 x g at 4 °C was done. Absorbance was read at 560 nm with a 96-well plate reader Multiskan SkyHigh™ (Thermo Scientific, Waltham, MA, United States).

The free proline content was determined spectrophotometrically using a microplate colorimetric assay (Carillo et al., 2011). Briefly, 40 mg of fine leaf powder was extracted in 1 mL of 70% ethanol, and centrifuged at 15,000 × g for 15 min at 4 °C. The resulting supernatant or standard solutions were mixed with a reaction mixture containing 1% (w/v) ninhydrin in 60% (v/v) acetic acid and 20% ethanol. After microplate homogenization, the reaction was carried out at 95 °C for 20 min in a dry block heater, followed by immediate cooling on ice. Absorbance was recorded at 520 nm using a microplate reader, and proline concentrations were quantified against a standard curve.

Total ascorbate (tAsa), reduced ascorbate (Asa) and dehydroascorbate (DHA) contents were determined spectrophotometrically using a microplate colorimetric assay (Stevens et al., 2008). Fine leaf powder (150 mg) was extracted in 600 µL of ice-cold 6% (w/v) TCA and centrifuged at 13,000 × g for 15 min at 4 °C. For tAsa determination, the supernatant was incubated with 5 mM dithiothreitol (DTT) at 37 °C for 20 min to reduce DHA, followed by the addition of 0.5% (w/v) N-ethylmaleimide (NEM) for 30 s to neutralize excess DTT. Color development was initiated in all wells by adding 80 µL of a color reagent mixture (1:1 solution A: solution B; where solution A contained 4.6% TCA, 31% H 3 PO 4, and 0.6% FeCl 3, and solution B contained 4% (w/v) 2,2’-bipyridyl in 70% ethanol). After incubation at 37 °C for 50 min, absorbance was measured at 550 nm. Ascorbate concentrations were calculated against a standard curve, and the DHA pool was calculated as the difference between total and reduced ascorbate. Asa, tAsa and DHA content were expressed in µmol g -1 FW.

Antioxidant enzyme activities

Enzymatic extracts were prepared in ice-cold phosphate buffer (0.1 M, pH 7.5) and activities were quantified in the resulting supernatants (Santini et al., 2013). Protein concentrations were determined using the Bradford method (Bradford, 1976). All enzymatic activities (CAT, APX, and DHAR) were expressed as units per milligram of soluble protein (U mg − ¹ protein), where one unit corresponds to the conversion of 1 µmol of substrate per minute under the assay conditions (i.e., H 2 O 2 degradation for CAT; ascorbate oxidation for APX; and dehydroascorbate reduction for DHAR). For superoxide dismutase (SOD), activity was expressed as U mg − ¹ protein with one SOD unit corresponding to the amount of enzyme required to inhibit the photochemical reduction of nitro blue tetrazolium (NBT) by 50%, following the classical inhibition-based assay.

Statistical analyses

Four biological replicates per condition were used for all analysed variables. Data are presented as means ± SE. To assess the effects of season and variety on each physiological and biochemical parameter, two-way analyses of variance (ANOVA) were conducted using R software. Prior to analysis, the normality of model residuals was assessed using the Shapiro-Wilk test, and homogeneity of variances was verified using Levene’s test. When these assumptions were met, a classical two-way ANOVA was performed on raw data. In cases where normality or homoscedasticity was violated, data were log-transformed to stabilize variance and normalize residuals before conducting the two-way ANOVA. For both parametric and log-transformed models, Tukey’s Honestly Significant Difference (HSD) post-hoc tests were applied to control the family-wise error rate across multiple pairwise comparisons. Significance groups were represented by distinct letters: lowercase letters for comparing varieties within each season, and uppercase letters for comparing seasonal evolution within each variety. In all cases, different letters indicate significant differences at p < 0.05. Although statistical significance was determined on log-transformed data, when necessary, all graphical representations and tables display the original (non-transformed) means and SE for biological relevance.

A principal component analysis (PCA) was conducted using the FactoMineR package to explore multivariate patterns across variables, following the approach described by Lourkisti et al. (2020). Variables were centered and scaled prior to ordination.

All statistical analyses and graphical representations were performed in R (version 4.3.2) using the ggplot2, emmeans, and patchwork packages.

Multivariate analysis of physiological and biochemical responses

A Principal Component Analysis (PCA) was conducted to summarize the overall variability among physiological and biochemical parameters across the three citrus genotypes and four sampling seasons. The first two principal components explained 56.74% of the total variance (PC1: 39.86%, PC2: 16.88%), indicating a good internal structure within the dataset (Figure 1A). PC1 was positively correlated with photosynthetic parameters (P net, gs, E, Fv/Fm) and negatively correlated with ascorbate redox components (Asa, tAsa and DHA) (Figure 1A). PC2 was positively correlated with CAT, Proline, or Asa/DHA ratio, and negatively correlated with SOD. Variables such as DHAR, MDA or H 2 O 2 exhibited low contribution scores to the first two principal components.

Temperature recordings showed that winter and spring were the coldest periods, while summer and autumn displayed temperatures above 18 °C (Table 1). Accordingly, the projection of individuals (Figure 1A) separated samples into 6 clusters with clusters 1, 2, and 6 located on the left side (cold period) and clusters 3, 4, and 5 located on the right side (warm period). In winter, Pomelo is clustered with Clementine (autumn), but is actually located on the left side of PC1. Indeed, summer and autumn samples are located on the positive side of PC1, associated with high P net, g s, and Fv/Fm. On the other hand, winter and spring samples occupied the negative side of PC1, characterized by lower photosynthetic activity and high activation of the ascorbate-related redox system. Interestingly, in spring, Clementine showed a marked association with DHA (Figure 1B), while W9 (cluster 2) was spatially close to reduced ascorbate (Asa) (Figure 1B). Indeed, for cluster 2, a positive Z-score was observed for Asa, whereas a negative Z-score was observed for H 2 O 2 and MDA (Figure 1B). On the other hand, no clear evidence of enhanced photosynthetic performance was observed for the triploid W9. It is worth noting that, during winter, Pomelo displayed an original distribution across PC1 and PC2 compared to Clementine and W9. In fact, compared to both, Pomelo did not display a negative correlation with photosynthetic parameters during the cold period (Figure 1B). Overall, PCA highlighted that seasonal variation was the major factor structuring physiological responses and revealed a genotype-specific antioxidant metabolism for the triploid hybrid W9 during the winter-to-spring transition.

Seasonal patterns of leaf gas exchange and chlorophyll fluorescence

All genotypes displayed a clear seasonal modulation of photosynthetic performance (Figure 2). For all varieties, P net, E and g s dropped significantly between summer/autumn and winter/spring (Figures 2A–C). The trend observed for Fv/Fm was not as evident, although there was a significant decrease for all varieties between summer and winter-spring (Figure 2D). Pomelo maintained the highest P net, g s and E values across almost all seasons (P net reached up to 13.28 ± 0.20 µmol CO 2 m − ² s − ¹ in autumn). Triploid W9 and Clementine displayed similar overall patterns, though triploid W9 tended to maintain slightly higher g s and E in autumn and spring. Overall, physiological parameters showed seasonal variation, with higher values in summer and autumn compared to winter and spring.

Differential accumulation of oxidative stress markers across seasons

Oxidative stress markers showed variability across seasons (Figure 3). Regarding MDA, autumn had the highest values compared to the other seasons (Figure 3A). MDA content in triploid W9 remained unchanged between summer and winter (28.35 ± 0.92 vs 26.13 ± 1.06), which was not the case for Clementine and Pomelo. However, MDA content dropped significantly in spring regarding the three varieties with a marked decrease concerning W9. Overall, W9 had the lowest MDA content, except during summer. This is associated with reduced H 2 O 2 concentration for W9 in autumn and spring compared to Pomelo and Clementine. Although H 2 O 2 did not follow the same seasonal pattern as MDA, the three varieties showed a similar ranking for both variables. When triploid W9 displayed a significantly lower H 2 O 2 content than Clementine and Pomelo, MDA content was also reduced, except in winter. In contrast, Proline content was overall higher in autumn and spring with no clear differences between varieties (Figure 3C; Supplementary Table 2). Overall, triploid W9 exhibited lower H 2 O 2 and MDA contents compared with the diploid varieties, depending on the season.

Modulation of ascorbate redox status and homeostasis

Total ascorbate content (tAsa) gradually increased from summer to spring except for Pomelo (Table 2). Clementine and triploid W9 displayed similarly low concentrations of reduced ascorbate (Asa) across all seasons except spring, during which triploid W9 exhibited the highest Asa concentration (6.43 ± 0.68 µmol g − ¹ FW) (Table 2). Dehydroascorbate (DHA) content was higher in clementine, while the reduced form was predominant for triploid W9 (Table 2). In Clementine, the Asa/DHA ratio dropped from 1.40 ± 0.05 in summer to 0.69 ± 0.02 in winter but did not change for triploid W9 at the same periods (1.63 ± 0.11 vs 1.86 ± 0.24). Pomelo showed intermediate Asa/DHA ratio values (>1) compared to clementine in all seasons. Overall, seasonal variation in ascorbate metabolism differed among genotypes, and triploid W9 showed the highest Asa content and Asa/DHA ratio in spring (Table 2).

Triploid W9 showed better ROS management and redox potential recycling in early spring

SOD activity did not differ significantly among varieties in most seasons, except in summer, when Clementine exhibited the highest SOD activity (Figure 4A). CAT activity remained stable from summer to winter (Figure 4B). In contrast, Pomelo and triploid W9 showed greater CAT activity than clementine in winter and spring (Figure 4B). APX activity did not differ drastically among seasons in all varieties (Figure 4C). While Pomelo had the highest APX activity in winter (0.32 U.mg -1 protein), triploid W9 showed higher APX activity (0.35 U.mg -1 protein) than Pomelo (0.27 U.mg -1 protein) and Clementine (0.28 U.mg -1 protein) in spring. DHAR activity decreased in Pomelo and Clementine during winter and remained lower than in the triploid W9, which showed higher DHAR activity in winter and spring (Figure 4D). Overall, triploid W9 displayed increased CAT, APX and DHAR activities, especially in spring, compared to diploid counterparts.

Chilling stress alters photosynthetic performance and oxidative balance in citrus

The main purpose of the present study was to investigate the behavior of a new citrus triploid hybrid, W9, under natural seasonal fluctuation in open-field conditions. The performance of the triploid W9 was compared to the two reference diploid crops widely cultivated in Corsica, Clementine and Pomelo, all grafted onto diploid C35 citrange rootstock. Seasons analysed were clustered in two, with summer-autumn displaying the highest mean temperature (25.1 °C and 18.5 °C) and winter-spring the lowest (9.3 °C and 14.3 °C). Moreover, spring was marked by relatively low minimal temperature (9.2 °C) compared to autumn (15.6 °C) (Table 1). Usually, summer and spring cluster together, while autumn is close to winter (Ribeiro and MaChado, 2007). This pattern aligns with typical Mediterranean climatic fluctuations, which are frequently characterized by chilling temperatures during early spring (Lionello et al., 2006). Furthermore, the spring measurement point was located early in the season, which also explains the cooler temperatures observed. Citrus trees are optimally cultivated between 22 °C and 30 °C (Oustric et al., 2017). It is important to note that below 13 °C, citrus growth and development can be altered (Ribeiro and MaChado, 2007). Low temperatures cause alterations in photosynthesis and redox homeostasis (Asada, 1999; Mittler, 2002) and subsequent disruption of photosynthetic machinery (Asada, 2006; Ensminger et al., 2006).

The principal component analysis (PCA) established that seasonal climatic fluctuations represent the primary driver structuring the overall physiological and biochemical responses of the studied citrus genotypes (Figure 1A). A clear spatial segregation was observed along the first principal component (PC1) between warm and cold periods. The warm period, located on the positive side of PC1, was characterized by positive Z-scores for photosynthetic parameters (Fv/Fm, P net, g s, and E), which were inversely correlated with the cold period on the negative side of PC1 (Figure 1B). Interestingly, the PCA revealed six distinct clusters instead of the anticipated four seasonal groups (Figure 1A). The extra clustering beyond the expected seasonal groups is biologically credible because citrus responses are often structured not just by environment, but by genotype and ploidy, with multivariate analyses separating genotypes into distinct groups (Oustric et al., 2019; Bonnin et al., 2023, 2024). This variation is predominantly projected onto the negative side of PC1 (corresponding to cold-period sampling points). These results strongly suggest that low temperatures act as a major catalyst for physiological and biochemical differentiation, as previously shown (Allen and Ort, 2001), amplifying genotype-dependent responses during cold stress. Overall, these findings highlight a seasonal metabolic shift from a photosynthesis-dominant profile in warmer months to an ascorbate-dependent antioxidant response during colder periods, particularly pronounced in the triploid W9.

The maximum photochemical quantum yield (Fv/Fm) is a marker of photochemical efficiency and is typically optimal around 0.83 (Urban et al., 2017). This is consistent with summer recordings (Clementine: 0.79; Pomelo: 0.81 and W9: 0.80) (Figure 2D). In contrast, Fv/Fm ratios in winter and spring declined significantly for all varieties compared to summer, falling below 0.75. Notably, triploid W9 showed a significant drop as early as autumn compared to the other crops. This decline in Fv/Fm is usually attributed to photo-inhibition observed during cold periods, resulting in reduced PSII efficiency (Genty et al., 1989). Citrus plants typically adapt to cold by closing their stomata to prevent transpiration and water loss (Wilkinson et al., 2001) leading to g s decline. This reduction in g s likely limited CO 2 diffusion into the leaf, thereby constraining carbon assimilation and contributing to the observed decrease in photosynthetic rates as widely reported under abiotic stress conditions. This limitation of CO 2 availability directly impacts carbon fixation processes, particularly through the regulation of key enzymes such as Rubisco, which likely contributes to the decrease in photosynthetic rates (Baker et al., 2007; Chaves et al., 2009). Accordingly, we observed a marked reduction in P net, g s, and E during winter and spring in all genotypes, likely explained by shortened daylight and low temperatures. These recordings are in accordance with gas exchange measurements (P net, g s and E) in the literature, where higher values are observed during warm periods than cold ones (Genty et al., 1989; Wu et al., 2025). Moreover, it is known that citrus trees show relatively low Pnet with maximum values around 13 µmol m -2 s -1 (Ribeiro and MaChado, 2007; Chen et al., 2021). Interestingly, while Pomelo maintained higher P net, g s and E, Clementine and triploid W9 stayed at the same level except in spring when triploid W9 showed enhanced gas exchanges (Figures 2A–C). It is well established that polyploidization induces anatomical changes, such as bigger stomata with lower density (Hennig et al., 2015; Hias et al., 2017; Corneillie et al., 2019). We previously described a triploid-induced alteration at the anatomical and cytological levels of hybrids from the same crossing as triploid W9 (Lourkisti et al., 2021b). However, these anatomical changes do not necessarily enhance g s, as stomatal number has a greater influence than size (Franks and Beerling, 2009). This may explain why in triploid W9, g s values were low compared to Pomelo (Figure 2C). More importantly, during the winter-to-spring transition, light intensity progressively increases, while temperatures (Tmin = 4.5 °C in winter and 9.2 °C in spring, Table 1) remain low. During that period, when g s is kept at a low level (Figure 2C), photo-oxidation occurs and induces ROS production (Baker and Rosenqvist, 2004; Azzabi et al., 2012), which can induce cytological damage like photochemical apparatus alteration.

Triploid W9 relies on an efficient antioxidant machinery to maintain low ROS content during winter-to-spring transition

Among ROS, H 2 O 2 has a relatively high lifespan and diffusion capacity, and participates in cellular signaling and gene regulation (Vestergaard et al., 2012). When accumulated in excess, H 2 O 2 can induce polyunsaturated fatty acid peroxidation and MDA formation, which is a useful indicator of oxidative damage (Luna et al., 2000). Even though this study took place under natural, unstressed conditions, MDA greatly varied with temperature (Figure 3A). The present data showed that MDA content peaked during autumn and decreased progressively in winter and spring. Notably, the triploid W9 displayed the lowest levels of both MDA and H 2 O 2 during the spring transition, which was associated with increased catalase activity (Figure 4B). Catalase is one of the principal H 2 O 2 -scavenging enzyme and greater catalase activity in citrus is commonly associated with reduced oxidative damage (Sofo et al., 2015; Liu et al., 2024). Enhanced CAT activity in triploid W9 may reflect both triploid-associated dosage effects and genotype-specific regulatory mechanisms, because polyploidy can modify gene output and CAT expression is known to be complexly regulated across plant species (Sarcheshmeh et al., 2023). Interestingly, MDA and H 2 O 2 values did not follow the same pattern as previously described for Clementine and other triploid hybrid from the same crossing as W9 (Lourkisti et al., 2020). In this previous study, highest MDA contents were observed during the cold period. The differences might come from different soil conditions or crop’s age-related effect. Nonetheless, the capacity of the triploid W9 to maintain lower oxidative markers points toward an enhanced control of ROS accumulation and may contribute to its physiological stability during the critical late-winter harvesting period. This ROS management observed in triploid W9 was associated with ascorbate metabolism modifications. While the diploid Clementine is associated with the oxidized form of ascorbate (DHA) in spring, reflecting an oxidative pressure, the triploid W9 segregates with reduced ascorbate (Asa) within Cluster 2 (Figure 1A). This specific positioning is functionally validated by the negative Z-scores for both H 2 O 2 and MDA within this cluster (Figure 1B), providing evidence of enhanced control over oxidative stress in the triploid hybrid during its specific period. This specialized antioxidant buffering capacity likely minimizes cellular lesions and allows the tree to navigate cold-induced stress more efficiently. H 2 O 2 scavenging is mainly mediated by the Ascorbate-Glutathione cycle (AsA-GSH cycle) and also by Fenton and Haber Weiss reaction (Asada, 2006). In this study, reduced ascorbate (Asa) content and the Asa/DHA ratio followed contrasting patterns among the three varieties (Table 2). In accordance with PCA analysis, Clementine exhibited the lowest Asa/DHA ratio, coupled with the highest DHA levels, suggesting an inability to fully compensate for ROS production, as evidenced by its high MDA content (Figure 3A). In contrast, the triploid W9 maintained a superior redox status, characterized by significantly higher Asa levels and low DHA accumulation, particularly during the winter-to-spring transition. This stability is likely driven by the functional synergy between APX and DHAR. While APX acts as the primary enzyme for H 2 O 2 detoxification, its activity consumes Asa and generates monodehydroascorbic acid (MDHA) and subsequently DHA. DHAR restores redox potential by efficiently reducing DHA back into Asa. Thus, DHAR ensures that the Asa pool remains available for APX-mediated detoxification. Here, especially during the winter-to-spring transition, the triploid W9 maintained significantly higher APX (Figure 4C) and DHAR (Figure 4D) activities. These results highly support the fact that triploid W9 can withstand oxidative stress efficiently through ascorbate metabolism.

In plants, Asa is mainly produced through the Smirnoff-Wheeler pathway (Vargas et al., 2024). Here, Clementine and triploid W9 maintained the highest tAsa content during autumn, winter and spring compared to Pomelo (Table 2). However, Clementine had a higher tAsa pool than triploid W9, except in spring, when tAsa is equivalent. This supports the idea that triploid W9 does not necessarily produce more Asa through the Smirnoff-Wheeler pathway but rather promotes Asa recycling through DHAR activity. It is worth noting that, beyond its antioxidant function, Asa can promote thermal dissipation (Saga et al., 2010; Smirnoff, 2018) when day light exposure increases during cold periods. Thus, in triploid W9, the enhanced ascorbate recycling may not only favour ROS degradation but also prevent their initial formation in the winter-to-spring transition.

These results indicate that triploid W9 displays improved antioxidant regulation under cold conditions. This superior performance could derive from increased gene dosage, a larger size (or an increased number) of cell compartments (Lourkisti et al., 2021b), or more efficient compartmentalization of redox reactions, a trait previously proposed for triploid perennials (Ramakrishnan et al., 2025). By maintaining redox homeostasis, triploid W9 may preserve cellular integrity and sustain the metabolic vigor required for late-ripening cultivars. This stands in contrast to more sensitive diploid genotypes, which showed higher oxidative marker levels and lower antioxidant defenses during the same period.

Proline is an amino acid mainly involved in plant responses to abiotic stress, like drought or salinity (Hasanuzzaman et al., 2020; Renzetti et al., 2025). It acts as an osmoprotectant but can also participate in ROS scavenging (Bauduin et al., 2022). In this study, we observed seasonal variations in proline content, but without clear differences between cultivars (Figure 3C; Supplementary Table 2), except for Pomelo in summer, which showed a significantly higher accumulation than Clementine and W9. Interestingly, a similar trend was observed for iWUE during the summer (Supplementary Figure 2). This suggests a feedback mechanism where the increased iWUE in Pomelo is driven by a reduction in stomatal conductance (Figure 2C) consequent to a decrease in leaf water potential, the latter potentially triggering proline biosynthesis. This type of osmotic and stomatal response has already been described in Citrus (Zaher-Ara et al., 2016) and Prunus (Jiménez et al., 2013). On the other hand, there is no evidence of proline-mediated ROS scavenging in spring. During this period, proline levels did not differ between W9, Pomelo, and Clementine (Figure 3C), even though the antioxidant defenses of W9 were upregulated. This indicates that the low-temperature tolerance of W9 in spring relies more on enzymatic or other redox pathways than on proline accumulation.

Integrative view: redox regulation as the driver of cold tolerance in Triploid W9

Altogether, the present results suggest that cold tolerance in the triploid hybrid W9, during the winter-to-spring transition, is primarily caused by redox plasticity rather than photosynthetic capacity. This period coincides with W9’s specific harvest window, during which fruit maturation reaches its peak. In citrus, maturing fruit acts as a significant metabolic sink, demanding high resource mobilization and inducing systemic oxidative stress. The capacity of triploid W9 to maintain redox homeostasis during this critical phase is therefore a key physiological trait.

At the genetic level, the enhanced redox balance of triploid W9 may come from its polyploid background. In fact, several studies have shown that polyploidy can reshape stress perception and ROS management pathways in citrus (Oustric et al., 2017). It has been shown that cold-tolerant genotypes tend to upregulate detoxification enzymes, including SOD, CAT and APX, while sensitive ones show a decline in their activities (Mutlu et al., 2013; Xu et al., 2013).

In addition to polyploidy-related effects, complex scion-rootstock interactions have to be considered (Schwarz et al., 2010). Even if all varieties in this study were grafted on the same rootstock and grown in identical soil, the specific combination of scion and rootstock can influence systemic signaling. In fact, recent evidence clearly demonstrates a much more complex and bidirectional communication between root and shoot (Asif et al., 2025). As an example, phytohormones like abscisic acid (ABA), central to both stomatal regulation and antioxidant gene expression, may be differentially transported or perceived depending on the scion’s ploidy and genetic heritage. Moreover, the scion influences rootstock physiology as well (Lu et al., 2020). Thus, we cannot exclude an effect coming from Citrange C35 (rootstock) - triploid W9 (scion) combination. Future studies involving a broader range of rootstocks with varying degrees of vigor and cold tolerance would be essential to determine if the superior redox homeostasis of W9 is a constitutive trait of the triploid scion or a result of specific scion-rootstock synergistic interactions.

Regarding its genetic heritage, previous work showed that Fortune mandarin is sensitive to water deficit but highly tolerant to cold (Lourkisti et al., 2020, 2021), whereas Tangor Ellendale exhibits cold sensitivity due to ETR-mediated ROS production (Lourkisti et al., 2020). Specifically, the high ETR observed in Ellendale during the cold period generates a surplus of electrons that cannot be utilized by the Calvin cycle. Instead, these electrons are channeled into the Mehler reaction, producing an excess of ROS, particularly H 2 O 2 (Hasanuzzaman et al., 2020). The triploid W9’s strong antioxidant response under cold conditions likely derives from its Fortune mandarin parent. Chilling tolerance in citrus is consistently associated with enhanced ROS-scavenging capacity (Oustric et al., 2017; Asif et al., 2025), whereas cold-sensitive genotypes like Tangor Ellendale, show excess ROS accumulation when photosynthetic electron flow exceeds carbon assimilation. Furthermore, this inherited tolerance could be amplified by the triploid status of W9. A ‘gene dosage effect’ may enhance the expression of key antioxidant genes, such as CAT, APX, and DHAR, leading to the superior redox stability observed during the winter-spring transition. This is consistent with findings that triploid citrus hybrids exhibit enhanced antioxidant defences and lower oxidative damage compared to their diploid parents (Lourkisti et al., 2022). Consequently, we can hypothesize that ROS tolerance was transmitted via the female parent (Fortune), while Tangor Ellendale probably influenced fruit quality and the late-harvest phenology (Ahmed et al., 2020).

Agronomic and economic implications for the Corsican citrus sector

The Corsican citrus industry is currently dominated by the Clementine, which has been shown to be sensitive to chilling (high H 2 O 2 and MDA, low Fv/Fm), water deficit, and low recovery potential (Lourkisti et al., 2020, 2021b, 2022). The present study corroborated previous results as Clementine displayed a higher DHA content than triploid W9 and Pomelo. Regarding Pomelo (‘Pomelo de Corse’, ‘Star Ruby’ grapefruit), only one study assessed its physiological parameters under field condition onto different rootstock (Çimen et al., 2016). Then the present study provides the first seasonal assessment of its antioxidant profile under natural Corsican pedoclimatic conditions. The elevated photosynthetic performance recorded for pomelo, especially during spring (Figure 2), resulted in a distinct clustering pattern in the PCA (Figure 1A, cluster 6). This underscores the effective physiological acclimation of this cultivar to Corsican pedoclimatic conditions.

While the inherent genetic differences between Clementine, Pomelo and W9 represent a limitation for direct physiological comparison, situating W9 against these two reference crops was essential to evaluate its commercial potential. These data indicate that W9’s triploidy does not change photosynthetic performance, while it may contribute to optimizing ROS metabolism, particularly in spring. Furthermore, from an economic perspective, W9 represents a strategic lever for Corsican producers. By maintaining metabolic vigor and fruit quality during a period when traditional cultivars (Clementine and Pomelo) are absent, W9 effectively bridges the recurrent production gap between winter and spring. This late-ripening capability would reduce the market’s reliance on imported citrus and strengthen the local economy.

This study assessed the seasonal physiological and biochemical modifications of the novel, late-ripening triploid citrus hybrid ‘W9’ under open-field Mediterranean conditions. The triploid hybrid W9 limited oxidative stress during cold periods through efficient antioxidant regulation, highlighted by robust DHAR-mediated ascorbate recycling, while maintaining photosynthetic capacity. These traits reflect the combined influence of its genetic background and triploid status, likely amplifying antioxidant capacity via gene dosage effects. Concurrently, while previous studies addressed the photosynthetic behavior of the Pomelo, ‘Star Ruby’ grapefruit, this work provides its first seasonal antioxidant and redox profile characterization under natural field conditions, revealing an effective acclimation driven by robust photosynthetic performance and osmotic adjustments. From agronomic and economic perspectives, combining Pomelo, W9, and Clementine in Corsican orchards offers a strategic lever to bridge the recurrent winter-to-spring production gap. This diversification ensures a continuous production calendar and a steady supply of high-quality, seedless fruits to the market.