Water, a crucial component of living organisms, plays vital roles in plant growth, development, maturation, ripening, and senescence (Hou et al., 2019; H. et al., 2020; Lufu et al., 2020; Ha et al., 2021). Typically, water accounts for 70–90% of the fresh weight of most horticultural products (Canet, 1988). Notably, citrus fruits are among the most popular horticultural commodities in both Chinese and global markets (Lu et al., 2023). The peel of citrus fruits, which comprises the flavedo and albedo, contains essential oils, pigments, and cellulose, and its average water content is around 75% of its fresh weight (Czech et al., 2020). The pulp — the edible portion of citrus fruits — consists of succulent segments filled with juice sacs and has a water content that is nearly 90% of its fresh weight (Lu et al., 2023). During postharvest storage, fresh horticultural products are highly prone to water loss through continuous transpiration. This loss, which often exceeds 90% of the total weight loss, is frequently synonymous with mass loss (Lufu et al., 2020). Excessive water loss (e.g., exceeding 5% of the initial weight of orange fruit) can inflict irreversible damage on cellular physiological and metabolic activities. This damage manifests as wilting; shriveling; and the loss of glossiness, firmness, texture, and flavor, rendering the fruit more susceptible to abiotic and biotic stresses. This ultimately leads to senescence and renders the fruit unmarketable (Grierson and Wardowski, 1978; Ben-Yehoshua and Rodov, 2002; Paniagua et al., 2013; Cronjé et al., 2017; Wei et al., 2019).

The physiological and biochemical characteristics of plant cells, such as their water potential and the levels of osmotic solutes, play a pivotal role in regulating water loss (Landrigan et al., 1996; Hou et al., 2019; Ghosh et al., 2021). Water potential, which is positively correlated with water content, acts as a direct driver of water flow (Lockley et al., 2021). Previous studies have demonstrated that the water potential of citrus peel decreases during storage, leading to increased water loss in fruits. Severe dehydration and abrupt changes in water potential can also induce physiological impairments in the peel (Cronjé et al., 2017). Similarly, research on apples with watercore revealed a steep water potential gradient from the normal outer parenchyma to the watercore region, which led to sustained water transport toward the watercore region (Wada et al., 2021). Under conditions of low water potential, plant cells maintain osmotic pressure by increasing the levels of osmotic solutes, including amino acids, soluble sugars, proteins, inorganic ions (such as K⁺, Ca²⁺, and Mg²⁺), and other osmoprotectants like proline and betaine. This adaptive response enhances water absorption and can also reduce water loss (Ozturk et al., 2021; Wang et al., 2022). Postharvest fruits experience water shortage stress because they cannot receive water from the parent tree. However, little research has been conducted to investigate the changes in water potential and osmolytes in both the peel and pulp of postharvest citrus fruit, as well as their relationship with water loss.

Aquaporins (AQPs) are membrane channels belonging to the major intrinsic proteins (MIPs) superfamily. AQPs facilitate the transport of water, hydrogen peroxide, small neutral solutes, and gas molecules across biological membranes (Groszmann et al., 2017; Luang and Hrmova, 2017). In plants, AQPs play a pivotal role in short-distance transmembrane water transport, and plants can modify their water status by regulating the number or activity of AQPs under water stress and other abiotic stress conditions (Kapilan et al., 2018; Singh et al., 2020). AQPs are categorized into five major subfamilies based on the sequence similarity of their coding genes and their subcellular localization: plasma membrane intrinsic proteins (PIPs), tonoplast intrinsic proteins (TIPs), nodulin26-like intrinsic proteins (NIPs), small basic intrinsic proteins (SIPs), and uncategorized intrinsic proteins (XIPs) (Singh et al., 2020). While PIPs, TIPs, NIPs, and SIPs are commonly found in major higher plant species, XIPs are absent in monocots and the Brassicaceae family (Maurel et al., 2015). Previous research identified 34 AQP genes in the genome of sweet orange, including 11 PIPs, nine TIPs, eight NIPs, three SIPs, and three XIPs (Wei et al., 2019). Some of these genes were found to exhibit differential expression in the roots and leaves of two orange cultivars that had contrasting drought-tolerance activities. However, the expression patterns of these AQP genes during the ripening and storage of fruit were not explored.

In recent years, accumulating evidence from research on horticultural crops has highlighted the important roles of AQPs in fruit growth, development, and ripening especially under the water deficit condition. The changes of AQPs gene expression and activity under water stress will help plants to defense the dehydration by regulating water transport (Martins Cde et al., 2015; Wei et al., 2019). For instance, the ectopic expression of PIP1;3 from the apple genome has been shown to increase fruit size and enhance drought tolerance in transgenic tomatoes (Wang et al., 2017). Meanwhile, the transcriptome analysis of pomegranate fruit revealed that the transcript abundance of PgrPIPs in this fruit was significantly higher than that of other subfamilies. Moreover, the mRNA levels of PgrPIP1.3, PgrPIP2.8, and PgrSIP1.2 exhibited a significant linear relationship with water accumulation in the seed coat (edible part), which suggested that these genes may contribute to cell expansion in the outer seed coat (Liu et al., 2021). Similarly, papaya fruits with higher expression of TIPs and SIPs have been found to exhibit larger mesocarp cells and intercellular spaces (Burns et al., 2023). AQPs exhibit tissue- and cultivar-specific expression patterns, although their expression also depends on the stage of fruit development and ripening (Merlaen et al., 2018; Quirante-Moya et al., 2022). In a previous study, 20 AQP genes were found to show differential expression patterns between ‘Red Fuji’ and ‘Golden Delicious’ apples, which have contrasting ripening behaviors (Lv et al., 2023). During the fruit development stage, expression of AQPs genes in peel and pulp was inconsistent, and most of them were down-regulated toward ripening (Yunhao et al., 2019; He et al., 2022). However, some AQP genes have specifically been found to be up-regulated during ripening, such as FaPIP1;1, FaNIP1;1, and FaPIP2;1 in strawberry fruit (Mut et al., 2008; Alleva et al., 2010; Molina-Hidalgo et al., 2015), GjTIP in Gardenia jasminoides (Gao and Guo, 2013), and MaPIP2-10, MaPIP1-6, MaPIP2-7, MaSIP1-1 in banana (Hu et al., 2015), and PaPIP1;4 in sweet cherry (Breia et al., 2020), This suggests the potential roles of these genes in enhancing fruit crispness and firmness and preventing cracking (Burns et al., 2023).

According to some studies, in the postharvest stage, the high expression of certain AQPs may be detrimental to water maintenance in fruits. For instance, increased water loss in postharvest tomatoes induced by exogenous melatonin was associated with the up-regulation of SlPIP12Q, SlPIPQ, SlPIP21Q, and SlPIP22 (Sun et al., 2015). A similar phenomenon was observed in pears, where accelerated water loss was linked to the increased degradation of cell wall polysaccharides and AQP gene expression (Ma et al., 2023). Notably, the overexpression of the CsNIP5;1 gene in citrus fruit and callus tissues was found to inhibit other AQP genes (such as CsPIP1;1, CsPIP2;4, and CsTIP1;2) and alleviate water loss (Zhang et al., 2021). Nevertheless, given the large number of AQPs present in plant cells, their tissue-specific expression, and their functional redundancy, the regulatory mechanisms of AQPs in the context of postharvest water loss in fruits remain poorly understood.

To explore the key factors affecting postharvest water loss in citrus fruits, we focused on the physiological differences in water status between the peel and pulp of Newhall navel orange (Citrus sinensis (L.) Osbeck) fruits during a 105-day storage. The differences of water physiology indexes such as the water potential, water content, weight loss rate and osmolyte accumulation, and the spatial and temporal expression levels of all CsAQP genes were examined regularly. additionally, the relationship between CsAQP expression and water loss was analyzed in detail. Overall, the results of this study will provide a clue for controlling postharvest water loss of citrus fruits.

Water is a major component of fruits and vegetables and directly influences their sensory quality, including their firmness, crispness, glossiness, flavor, stress resistance, and shelf life (Landrigan et al., 1996; Frenkel and Hartman, 2012; Eboibi and Uguru, 2018). In the model fruits tomato and strawberry, water accounts for as high as 92.5% (Hou et al., 2019) and 90.90% (Akhtar and Rab, 2015) for the fresh weight, respectively. Unlike these berry fruits, citrus fruits belong to the group of hesperidium fruits and have a thick leathery peel and succulent pulp. The water content in citrus fruits varies considerably among different parts of the fruit and also among different fruit varieties. The pulp of orange, mandarin, grapefruit, pummelo, and lemon — the most commonly cultivated citrus varieties in the world — contains 82.30–87.90%, 80.70–88.60%, 76.90–91.60%, 89.10%, and 88.98% water, respectively (Mbogo et al., 2010; Lu et al., 2023). Meanwhile, the water content of the peel — which is an inedible component of citrus fruits — varies between 68.1% and 80.4% and is lower than the water content of the pulp (Czech et al., 2020). In line with previous reports, in the present study, the water content in the peel and pulp of Newhall navel orange fruit was found to be around 83.76–85.03% and 70.36–73.54% during storage, respectively ( Figure 1C ). Moreover, the water content of the pulp did not change obviously during storage, while that of the peel increased during the first 60 d and decreased thereafter (P < 0.05) ( Figure 1C ). A similar phenomenon was observed in samples from the year 2017 within 84 d of storage ( Supplementary Figure 1 ). It has been proposed that the decrease in water content may induce fruit ripening (Frenkel and Hartman, 2012). Harvested citrus fruit can be stored for as long as 2–6 months, and excess ripening during this prolonged storage is detected based on the deterioration of internal quality using indicators such as total soluble solids (TSS), titratable acidity (TA) and TSS/TA values. Research on the shelf life of orange fruit has indicated that the critical stage of postharvest quality decline occurs at approximately 60 days postharvest (Frenkel and Hartman, 2012; Ma et al., 2021), consistent with the trend of water content changes in the peel ( Figure 1C ). Consequently, we speculate that, in addition to sugar and acid levels, the peel water content could also serve as an indicator of quality and physiological stage in citrus fruit during postharvest storage. Similar methodologies have been applied for the prediction of strawberry fruit quality based on optical properties (Xie et al., 2021).

In contrast to growing fruits, which absorb water from the parent plant, harvested fruits face water deficiency and are highly susceptible to water loss during storage. The loss of 3–10% of the initial weight renders most fruits and vegetables unmarketable (Ben-Yehoshua and Rodov, 2002), and citrus fruits are no exception. Desiccation due to water loss results in weight loss and affects the appearance, firmness, and commercial value of fruits while also limiting their shelf life. Studies have demonstrated that, besides carbon loss through respiration, water loss attributable to transpiration and respiration accounts for over 95% of the total weight loss in fruits (Bovi et al., 2018; Lufu et al., 2019; Xanthopoulos, 2021). Hence, water loss is frequently used interchangeably with weight loss in research on postharvest fruits. In our study, weight loss was measured to monitor the water loss from navel oranges during storage. We found that the water loss increased with prolonged storage, but this increase was not uniform. Specifically, the rate of water loss was accelerated during the later storage period ( Figures 1A, B ), with the average weight loss rate per 15 days increasing from 0.45% at 0–30 d to 0.60% and 0.90% at 45–60 d and 90–105 d, respectively ( Figure 1C ). This indicated that water loss in harvested oranges was not only due to simple physical diffusion but actually represented a physiological water transmission process that was affected by intracellular factors, such as water potential, solute potential, and turgor (Hou et al., 2021).

The water potential within plant cells, as indicated by their water content (Lockley et al., 2021) is a critical driving force for water movement across cellular membranes. In line with previous findings (Czech et al., 2020), in this study, water content and potential of Newhall navel orange is consistently higher in pulp than that in peel during storage ( Figures 1C, D ), which leads to a sustain water potential gradient between the pulp and peel ( Figure 1E ). Studies on watercore apples have shown that a steeper water potential gradient, extending from the normal outer parenchyma to the watercore region, establishes a sustainable mechanism for water transport to the watercore (Wada et al., 2021). Similarly, the presence of a water potential gradient between the peel and pulp tissues of Newhall navel oranges may generate a continuous driving force for water movement from the pulp to the peel, ultimately resulting in water loss from the surface of the fruit. Furthermore, similar to the minimal changes in water content observed in the pulp, the water potential of the pulp remained relatively stable during storage, whereas the peel water potential demonstrated a sharp decline from 60 d onwards. Previous studies on ‘Marsh’ grapefruit and ‘Navelate Navel’ oranges have revealed a significant positive relationship between peel water potential decline and fruit water loss (Alferez et al., 2010; Cronjé et al., 2017). In our study, the notable increase in the water potential gap between the peel and pulp from 60 to 105 d ( Figure 1E ) may explain the accelerated rate of water loss observed after 60 d ( Figures 1A, B ). Moreover, correlation analyses conducted on both peel and pulp samples ( Figure 6 ) further demonstrated the positive relationship between the water potential gradient and cumulative weight loss. Based on the above analysis, we can infer that the sharp changes in water potential in the peel ( Figure 1D ) influence the water dynamics in postharvest citrus fruits, determining the rate of water loss during storage.

Solute potential, a critical aspect of water potential in plant cells, is determined by the presence of osmolytes, such as soluble sugars, proteins, organic acids, and various inorganic ions (Ozturk et al., 2021; Wang et al., 2022). In this study, we observed that soluble sugars (sucrose, glucose, and fructose), organic acids (citric acid, malic acid, etc.), soluble protein, and mineral elements were significantly more abundant in the pulp than in the peel ( Figures 2A–H , 3A–K ). Generally, the total concentration of organic and inorganic osmotic solutes is slightly higher in the pulp (averaging about 71.51 mg/g FW) than in the peel (averaging about 68.41 mg/g FW). Throughout the storage period, the concentration of sugars and acids decreased in the pulp, with notable declines occurring around 30–60 d ( Figures 2A–G ). This decrease likely led to an increase in the osmotic and water potential within the pulp. Correlation analysis revealed that soluble sugar and citric acid concentrations were negatively correlated with the water potential gradient and cumulative weight loss ( Figure 6 ). However, the water content and potential in the pulp remained relatively stable during storage ( Figures 1C, D ). Sucrose and citric acid are the most abundant compounds in citrus pulp (Zhou et al., 2018). Sucrose metabolism plays a crucial role in determining the soluble sugar concentration and osmotic potential of ripening fruit (Balibrea et al., 2006). Additionally, during storage, the degradation of hexoses and organic acids via glycolysis and the tricarboxylic acid cycle (Luo et al., 2021) further adds water to the pulp. Therefore, the degradation of sugars and acids in pulp cells appears to be essential for maintaining the water potential and water balance in Newhall navel oranges during postharvest storage ( Figures 1C, D ). In contrast, in our study, the osmolytes in the peel tissues showed no notable changes, unlike the water potential. The water potential of the peel sharply declined from -1.48 MPa before 60 d to -2.79 MPa after 90 d, while only the amounts of soluble protein and mineral elements increased in the peel, although the sugar and acid concentrations fluctuated over time ( Figures 2A–G , 3L ). These results align with previous observations demonstrating that the osmotic potential in citrus peels does not undergo significant changes unlike the water potential (Alferez et al., 2010). Similarly, in previous studies, fruits such as apricots, doughnut peaches, and nectarines treated with foliar sprays of Ca+B and Ca+Si did not show changes in osmotic potential despite increases in water potential and turgor potential (Quirante-Moya et al., 2022). In our study, soluble protein exhibited a positive correlation with the water potential gradient in the peel ( Figure 6 ). Hence, its increase may contribute to a lowered water potential in the peel, enlarging the water potential gap between the peel and the pulp and accelerating water transmission from the pulp to the peel during storage.

AQPs, water channels belonging to a major superfamily of intrinsic proteins, facilitate water transport across cellular membranes along the water potential gradient (Kapilan et al., 2018). AQP genes are typically categorized into five major subfamilies: PIPs, TIPs, NIPs, SIPs, and XIPs (Singh et al., 2020). In the sweet orange genome, 34 AQP-encoding genes have been identified. Our investigation revealed that 91.17% (31 out of 34) of these CsAQPs were expressed in postharvest Newhall navel orange fruit, including ten CsTIPs, nine CsNIPs, six CsPIPs, three CsSIPs, and three CsXIPs ( Figure 4 ). This proportion closely resembles the expression rate of AQPs in ripening Lycium chinensis (97.37% or 37 out of 38) (He et al., 2022) and developing bananas (85.11% or 40 out of 47) (Hu et al., 2015). However, it surpasses the expression rates observed in ripening apples (47.62% or 20 out of 42) (Lv et al., 2023) and developing watermelons (34.29% or 12 out of 35) (Lin et al., 2019). This variance may be due to the temporal, spatial, and cultivar-specific expression patterns of AQPs.

Numerous studies have highlighted the differential expression and activity of various AQP genes across various fruits, tissues, and developmental and ripening stages under water deficit that induced by unproper environmental conditions (Merlaen et al., 2018; Wei et al., 2019; Quirante-Moya et al., 2022). Environmental stresses such as drought and salinity stress can rapidly reduce the rate of water transport across the membrane, while plants control water flow inside and outside the cell by regulating the abundance and activity of aquaporins to maintain water balance (Kapilan et al., 2018). In citrus fruits, aquaporin expression may be influenced by several factors, including plant growth conditions, developmental stages, tissue types, and the duration and intensity of stress. An analysis of the transcription profile of CsAQP genes in the roots and leaves of drought-tolerant (HJ) and drought-sensitive (HH) cultivars under drought treatment revealed that most CsPIPs and CsTIPs were down-regulated in roots of both treated cultivars. In addition, the down-regulation of CsPIP1;2, CsTIP3;2, and CsNIP2;1 in roots and up-regulation of CsNIP1;1, CsNIP1;3, CsNIP4;1, and CsNIP5;1 in leaves revealed obvious differences between tolerant and sensitive cultivars during drought (Rodríguez-Gamir et al., 2011; Martins Cde et al., 2015). Most CsAQPs exhibited higher expression levels in the pulp than in the peel and showed a decreasing trend during fruit development and ripening processes. For instance, in sweet cherry, AQP genes were more highly expressed in the flesh than in the skin, with 16 out of 25 AQPs being down-regulated during later developmental stages (Yunhao et al., 2019). Similarly, during Lycium barbarum fruit ripening, 24 AQPs were down-regulated, while three were up-regulated (He et al., 2022). Studies on postharvest mandarin, sweet orange, and pummelo fruits have also revealed the down-regulation of many AQPs and the up-regulation of a few AQPs in the pulp (NIP1;2, NIP5;1, PIP2;2, and TIP3;1) (Ding et al., 2015). Consistent with observations in sweet cherry (Yunhao et al., 2019), CsAQPs exhibited higher expression in the pulp than in the peel in our study, with most (19 out of 27) being down-regulated in the peel (indicated by green circles in Figure 5 ). However, 24 out of 31 CsAQPs showed up-regulation in the pulp at one or more test stages (indicated by red circles in Figure 5 ). Particularly noteworthy was gene cluster I, which includes CsTIP1;3, CsTIP2;2, CsTIP2;3, CsTIP5;1, and CsTIP6;1 in the TIP subfamily; CsNIP2;1, CsNIP2;2, and CsNIP6;1 in the NIP subfamily; CsSIP1;1, CsSIP1;2, CsXIP1;1, and CsXIP 1;2. This gene cluster exhibited expression 10–130 times higher levels in the pulp at 60 d than at 0 d (see Figures 4 , 5 ). A study in pomegranate fruit showed that eight PgrPIP members exhibited relatively high levels of transcript accumulation in both the inner and outer seed coats, with the transcription levels of PgrPIP1.3, PgrPIP2.8, and PgrSIP1.2 showing a significant linear relationship with water accumulation in the juicy outer seed coat (Liu et al., 2021). Additionally, in Lycium barbarum fruit, most LbAQP genes were found to be down-regulated during ripening and showed a negative correlation with the relative water content of the fruit (He et al., 2022). Results from PCA ( Figure 7 ) also indicated that genes in Cluster I, which are highly expressed in the pulp, were significantly positively correlated with pulp water content.

The efficiency of water transmembrane movement is influenced by both AQP expression levels and enzyme activity (Maurel et al., 2015; Kapilan et al., 2018; Rabeh et al., 2024). Several studies have established a connection between AQP gene expression and water loss in postharvest fruits. For instance, one study reported that in tomato fruit, the rate of water loss increased by 8.3% under 50 µM melatonin treatment, and the expression of SlPIP12Q, SlPIPQ, SlPIP21Q, and SlPIP22 was up-regulated 2- to 3-fold (Sun et al., 2015). In pear fruit, the PbPG enzyme accelerated water loss by degrading cell wall polysaccharides, thereby expanding the intercellular layer and AQP-encoding gene expression (Ma et al., 2023). Additionally, repressing the expression of CsPIP1;1, CsPIP2;4, and CsTIP1;2 alleviated water loss in citrus fruits (Zhang et al., 2021). The results of correlation analysis and PCA indicated that AQPs in Clusters I and III were positively correlated with average and cumulative weight loss, as well as water indexes such as water content, water potential, water potential gradient, and osmolytes in the pulp samples at 30, 60, 90, and 105 d along the PC1 direction ( Figure 7 ). Notably, the pulp sample at 60 d formed a distinct cluster separate from the other pulp samples. This suggests that the AQPs that are abundantly expressed in the pulp, especially at 60 d, may facilitate efficient water transport from the pulp to the peel, thereby playing a crucial role in regulating fruit weight loss.

Based on the aforementioned analysis, we can conclude that (as displayed in Figure 8 ) in Newhall navel orange fruit, water content and potential are higher in the pulp than in the peel and show minimal changes in the pulp during storage. Conversely, in the peel, water content increases before 60 d postharvest, and this is followed by a sharp decrease in the water potential after 60 d. This significant change in the water potential gradient between the peel and pulp after 60 d may contribute to an accelerated rate of water loss during later storage periods. Additionally, while osmolyte concentration decreases in the pulp, it increases in the peel over time. The expression pattern of CsAQPs exhibits a tissue-specific mode, with most genes being down-regulated in the peel and up-regulated in the pulp during storage. Notably, 16 specific CsAQPs are up-regulated in the pulp, particularly at 60 d postharvest, and are significantly positively correlated with weight loss, demonstrating their role in facilitating water transportation. Overall, these findings suggest that the differential regulation of dynamics between the pulp and peel, along with tissue-specific patterns of CsAQP gene expression, contribute to the modulation of water loss during the postharvest storage of Newhall navel orange fruit.