A Csi‐miR3954b‐CsCE70/CsPLA1 Regulatory Module Enhances Citrus Fruit Resistance to Penicillium digitatum by Promoting Phenolic Acid and Flavonoid Biosynthesis

ABSTRACT

Green mould is one of the major postharvest diseases affecting citrus. Although evidence indicates that microRNAs (miRNAs) participate in plant defence responses, their roles and regulatory mechanisms in citrus resistance to green mould remain unclear. Small RNA sequencing identified several defence‐associated miRNAs involved in citrus resistance to Penicillium digitatum. Among them, transient overexpression screening identified Csi‐miR3954b as the most effective miRNA, reducing disease incidence by 24.4% compared with the control at 5 days post‐inoculation (dpi). Subsequent transient overexpression and short tandem target mimic (STTM)‐mediated suppression assays demonstrated that Csi‐miR3954b positively regulates citrus resistance to green mould, with transient overexpression reducing disease incidence by 22.7% and STTM‐mediated suppression increasing it by 9.6% at 5 dpi. Degradome sequencing, transient assays in Nicotiana benthamiana leaves and expression analysis identified CsCE70 and CsPLA1 as the direct targets of Csi‐miR3954b. Transcriptome analysis showed that CsCE70 and CsPLA1 mainly regulated phenolic acid and flavonoid biosynthesis, respectively. Transient overexpression of CsCE70 and CsPLA1 increased disease incidence by 17.4% and 10.4%, respectively, compared with the control at 5 dpi, suggesting that both genes negatively regulated citrus resistance to P. digitatum. Silencing these genes enhanced its resistance, accompanied by the accumulation of phenolics and flavonoids in the peel. In general, the results showed that the Csi‐miR3954b–CsCE70/CsPLA1 regulatory modules enhanced citrus resistance to P. digitatum by promoting the biosynthesis of phenolic acids and flavonoids. These results provide new insights into miRNA‐mediated postharvest disease resistance and potential molecular strategies for improving citrus postharvest protection.

Csi‐miR3954b targets CsCE70 and CsPLA1, promoting the synthesis of phenolic acids and flavonoids, thereby enhancing citrus resistance to postharvest green mould.

1. Introduction

Citrus is among the most extensively cultivated fruit crops worldwide and constitutes a large component of the global horticultural market. Unfortunately, their market value is frequently lost by postharvest fungal infections (Bhatta 2022). The green mould caused by Penicillium digitatum is the most common and devastating postharvest disease of citrus during storage and transportation (Cheng et al. 2020). P. digitatum typically infects the fruit by means of superficial injuries, and then quickly invades with its mycelium, causing a general rotting process of tissues (Costa et al. 2019; Droby et al. 2010; González‐Candelas et al. 2010). Citrus peel at this stage acts as the primary barrier to pathogen invasion and plays an essential role during initiation of the defence response (Droby et al. 2003; Romanazzi et al. 2016). Thus, understanding the regulation involved in the defence response will be necessary for enhancing its resistance against postharvest diseases.

Currently, postharvest diseases are mainly prevented and controlled through the use of fungicides such as imidazole and propyl. Nevertheless, the rise in the number of fungicide‐resistant strains and consumer's concerns on chemical residue raises questions over their long‐term viability (Cheng et al. 2020; Li et al. 2021; Liu et al. 2023). Accordingly, biocontrol based on antagonistic yeast has received ever‐increasing interest. Among these yeasts, Pichia galeiformis has been extensively studied. It suppresses P. digitatum through direct antagonism while also inducing host resistance (Chen et al. 2020; Chen et al. 2021; Liao et al. 2025). Previous studies have shown that P. galeiformis activates the phenylpropanoid pathway in citrus by upregulating genes encoding key biosynthetic enzymes and promoting the accumulation of defence‐related secondary metabolites, including phenolics, flavonoids and lignin (Chen et al. 2021). However, whether these defence responses involve post‐transcriptional regulation mediated by microRNAs (miRNAs) remains unknown.

miRNAs are small endogenous non‐coding RNAs that regulate gene expression through sequence‐specific interaction with target transcripts. In plants, their most common function is to guide the cutting or translation inhibition of target mRNAs, thus affecting a wide range of biological processes (Bartel 2004; Voinnet 2009). A large number of studies have shown that miRNA participates in plant growth and response to environmental stress (Katiyar‐Agarwal and Jin 2010; Liu et al. 2022). miR393 and miR160 enhance plant immunity by targeting components of auxin signalling pathways (Huang et al. 2019; Navarro et al. 2006). In fruit, miRNA is also related to the regulation of development, maturity and stress response. In strawberry (Fragaria ananassa), suppression of miR396 enhances abscisic acid (ABA) biosynthesis and accelerates fruit ripening (Chen, Yan, et al. 2025; Chen, Zhang, et al. 2025). In grapes ( Vitis vinifera ), it has been reported that miR395 and miR827 enhance resistance to pathogens during preharvest growth by controlling sulphur metabolism and styrene biosynthesis (Luo et al. 2024; Xiang et al. 2025). Apart from the preharvest phase, there is also postharvest biological evidence for the involvement of miRNAs in fruit defence. In lychee ( Litchi chinensis ), the miR159–GAMYB and miR828–TT2 regulatory modules are closely associated with resistance against postharvest downy blight (Yin et al. 2023). In summary, these findings show that miRNA is involved in many aspects of fruit physiology, including defence response.

The biological functions of miRNA depend on the genes they regulate, because miRNA itself does not encode functional proteins. In plants, these targets include transcription factors, signal transduction components, resistance genes and enzymes involved in secondary metabolism (Deng et al. 2024; Zhao et al. 2025). Through their diverse targets, miRNA can regulate both immune signalling and structural defence in plants. Members of miR482/2118 superfamily regulate immune responses against Verticillium dahliae and Phytophthora infestans by targeting NBS‐LRR resistance genes (Jiang et al. 2018; Yang et al. 2015). This regulatory pattern has also been reported in postharvest fruit. In apple ( Malus domestica ), miR482‐mediated cleavage of the resistance gene MdTNL1 compromises resistance to Alternaria alternata (Liu et al. 2024). In addition to immune signalling, miRNA can also modulate the formation of physical barriers. For instance, in pear ( Pyrus communis ), miR397 regulates a laccase gene that is implicated in lignin biosynthesis, thus regulating fruit mechanical barrier against fungus invasion (Yang et al. 2023). miRNA can be involved in the regulation of other defence‐related secondary metabolites accumulation as well. In Populus, miR156 regulates SPL transcription factors to affect the biosynthesis of flavonoids and related compounds (Wang et al. 2020). Moreover, an miRNA can target many different genes. One example is the miR160 of Arabidopsis that targets the auxin response factors ARF10, ARF16 and ARF17 (Mallory et al. 2005). These studies show that the key role of miRNA is to connect gene regulation with plant defence metabolic and structural pathways.

Although increasing evidence from other plant systems has highlighted the importance of miRNA‐mediated regulation in plant immunity, the roles and regulatory mechanisms of miRNAs in citrus defence remain to be fully elucidated. High‐throughput sequencing has identified numerous conserved and citrus‐specific miRNAs in citrus, and functional studies have revealed their roles in various biological processes, including development and fruit ripening (Song et al. 2010; Xu et al. 2010). For example, the citrus‐specific miR3954 family is involved in flowering regulation, while the miR156‐SPL modules regulate flowering and fruit ripening through hormone‐related pathways (Chen, Zhang, et al. 2025; Liu, Ke, et al. 2017; Zhang, Xu, et al. 2025). Despite these advances, previous studies on citrus miRNAs have mainly focused on developmental processes, whereas whether and how miRNAs regulate citrus resistance to postharvest fungal pathogens and the corresponding functional targets remain largely unknown.

Plant resistance to pathogens usually depends on metabolic pathways, which can not only produce antibacterial compounds but also strengthen structural barriers (Wang et al. 2022; Yang et al. 2024). Among various defence‐related metabolic pathways, the phenylpropanoid network is central to metabolism across species and tissues, playing an important role in fruit defence. Phenolic acids and flavonoids generated via this pathway are capable of limiting colonization by pathogens and maintaining cellular redox balance (Agati et al. 2012; Hu et al. 2023). Lignins deposited from products of this pathway may reinforce cell walls and reduce pathogen entry (Tiwari et al. 2025). Other derivatives, including coumarins and stilbenes, also play a role of antibacterial antitoxin in many plant–pathogen systems (Ortiz and Sansinenea 2023; Stringlis et al. 2019). In citrus fruit, the infection of P. digitatum induces the transcription of phenylpropanoid biosynthesis genes and promotes the accumulation of phenolic compounds in the peel (Ballester et al. 2013; González‐Candelas et al. 2010). However, the mechanism of controlling the metabolic flux between phenolic acid and flavonoid biosynthesis pathway in postharvest fruit tissue is still not clear.

To elucidate how miRNAs contribute to citrus resistance against postharvest fungal pathogens, this study investigated miRNA‐mediated regulation of resistance to P. digitatum in postharvest citrus fruit. Small RNA sequencing was performed to identify differentially expressed miRNAs associated with the interaction among citrus fruit, P. galeiformis and P. digitatum . Among these differentially expressed miRNAs, the candidate miRNA with the strongest positive effect on citrus resistance was selected for further investigation. Candidate miRNA targets were identified through degradome sequencing, followed by functional validation using transient expression assays in Nicotiana benthamiana leaves and citrus fruit. The functions of the identified target genes in citrus resistance were further evaluated through transient overexpression and silencing assays. Transcriptome analysis and metabolite profiling were conducted to characterise the downstream metabolic pathways influenced by the identified miRNA–target modules. This study identifies a miRNA–target regulatory module contributing to citrus resistance against postharvest fungal pathogens.

2. Results

2.1. Small RNA Profiling Reveals Differentially Expressed Conserved miRNAs Under P. galeiformis ‐Associated Resistance Conditions in Citrus Fruit

The reliability of the small RNA sequencing data was first assessed. Hierarchical clustering analysis showed good correlation among biological replicates; samples inoculated with P. digitatum combined with P. galeiformis (PP) clustered separately from samples inoculated with P. digitatum with sterile water (PH) (Figure 1a). The length of most miRNAs was between 20 and 24 nucleotides. Among them, miRNA with 21 nucleotides was the most abundant (108), followed by miRNA with 22 nucleotides (31) (Figure 1b). Nucleotide composition analysis showed that most length categories were obviously biased towards U (Figure 1c). Differentially expressed miRNAs between PP and PH treatments were identified to explore miRNA responses associated with the presence of P. galeiformis during P. digitatum infection. Four conserved miRNAs showed significant expression changes, including three upregulated and one downregulated miRNAs, with log₂ (fold change) values ranging from −1.53 to 3.11 (Figure 1d, Table 1). To further prove that they were indeed miRNAs, their precursor structures were predicted. All four precursors formed a typical stem–loop hairpin structure with a low minimum free energy (MFE) value, which was consistent with the typical miRNA precursor (Figure 1e). Finally, reverse transcription‐quantitative PCR (RT‐qPCR) analysis at 24 h post‐inoculation validated the expression pattern of these miRNAs and confirmed the trend observed in small RNA sequencing (Figure 1f).

FIGURE 1.

Conserved miRNAs differentially expressed in Citrus sinensis fruit after inoculation with Pichia galeiformis and either sterile water (PH, control) or Penicillium digitatum (PP, treatment). (a) Correlation heatmap. (b) Length distribution of conserved miRNAs. (c) The nucleotide preference of conservative miRNA of different lengths. (d) Volcano plot of differentially expressed miRNAs (|log₂FC| ≥ 1 and FDR ≤ 0.05). (e) Predicted pre‐miRNA stem–loop structures using RNAfold. (f) Reverse transcription‐quantitative PCR validation of four selected miRNAs at 24 h post‐inoculation, normalized to PH. Data were presented as means ± SD; **p < 0.01 (Student's t‐test).

TABLE 1.

List of significantly differentially expressed conserved miRNAs between inoculation with Pichia galeiformis and Penicillium digitatum (PP) and inoculation with P. digitatum and sterile water (PH, control).

miRNA ID	Mature	Length (nt)	Chromosome	log2 (fold change) (PP/PH)	FDR	Regulation
miR827	UUAGAUGACCAUCAACAAACA	21	JH999145.1	1.01	4.1 × 10−2	Up
miR9560	UCAUAUUUGCUCCACCACCUGUGG	24	Chr3	3.11	5.73 × 10−4	Up
miR166f	UCGGACCAGGCUUCAUUCCCU	21	Chr3	1.75	4.43 × 10−5	Up
miR3954b	ACCGUGUUUCUCUGCCCAAUC	21	Chr9	−1.53	6.65 × 10−3	Down

2.2. Functional Analysis Identifies miRNAs Differentially Expressed During P. digitatum Infection in Citrus Fruit

To identify differentially expressed miRNAs involved in citrus resistance against P. digitatum , Csi‐miR166f, Csi‐miR827, Csi‐miR3954b and Csi‐miR9560 were transiently overexpressed (OE) in citrus fruit before P. digitatum inoculation. RT‐qPCR analysis confirmed successful overexpression, with transcript levels of all four miRNAs significantly higher than those in fruit infiltrated with the empty vector control (Figure 2a). Representative disease phenotypes observed during storage were shown in Figure 2b. Fruit overexpressing Csi‐miR3954b consistently exhibited reduced disease development compared with the control from 3 to 5 days post‐inoculation (dpi). Csi‐miR827 delayed disease progression during the early storage period, while Csi‐miR166f and Csi‐miR9560 showed limited and transient effects on disease development. These observations were supported by disease incidence measurements. At 5 dpi, OE‐miR3954b fruit showed a disease incidence of 70.8%, compared with 95.2% in the control, representing a 24.4% reduction. OE‐miR827 fruit displayed moderate resistance, with a disease incidence of 86.3%, whereas OE‐miR166f showed a weaker reduction and OE‐miR9560 did not significantly reduce disease incidence compared with the control (Figure 2d). Lesion diameter measurements further confirmed these trends (Figure 2e). At 4 dpi, Csi‐miR3954b, Csi‐miR827 and Csi‐miR9560 significantly suppressed lesion growth relative to the control, whereas Csi‐miR166f showed no inhibition. By 5 dpi, only Csi‐miR3954b maintained a markedly significant reduction in lesion expansion, with lesion diameters reaching 95.3 mm compared with 130.0 mm in the control (26.7% reduction). Together, these results indicated that Csi‐miR3954b was identified as the strongest positive regulator of citrus resistance against P. digitatum among the tested miRNAs and was therefore selected for subsequent mechanistic studies.

FIGURE 2.

Effects of transient overexpression of conserved miRNAs on green mould resistance in Citrus sinensis ‘Xiacheng’ at 25°C. (a) Relative expression levels of miRNAs at 2 days post‐infiltration, normalized to control (set as 1). (b) Representative fruit images, (c) disease incidence and (d) lesion diameter at 3–5 days of storage after transient miRNA overexpression and Penicillium digitatum inoculation. Empty pCAMBIA2300 vector served as control. Data were presented as means ± SD. Statistical significance was determined using one‐way ANOVA followed by Dunnett's multiple comparisons test (*p < 0.05, **p < 0.01).

2.3. Csi‐miR3954b Positively Regulates Citrus Resistance to P. digitatum

To determine whether Csi‐miR3954b is involved in citrus resistance to P. digitatum , OE and short tandem target mimic (STTM)‐mediated suppression assays were performed in citrus fruit. The STTM3954b construct was designed to suppress endogenous Csi‐miR3954b (Figure 3a). RT‐qPCR confirmed that transient overexpression significantly increased the transcript level of Csi‐miR3954b, whereas STTM significantly reduced it (Figure 3b). Representative disease phenotypes observed during storage were shown in Figure 3c. Fruit treated with OE‐miR3954b showed delayed disease development, while STTM3954b‐treated fruit exhibited accelerated disease development relative to the control. During 3 to 5 dpi, OE‐miR3954b fruit showed significantly lower disease incidence than the control, with disease incidence of 60.3% at 5 dpi, compared with 83.0% in control fruit (22.7% reduction). In contrast, disease incidence in STTM3954b fruit increased to 92.6% at 5 dpi, representing 9.6% higher than the control (Figure 3d). Similarly, OE‐miR3954b fruit developed smaller lesions (60.0 mm) than the control (75.4 mm) at 5 dpi, whereas STTM3954b fruit had larger lesions (80.9 mm) (Figure 3e). These results confirmed that Csi‐miR3954b positively regulates citrus resistance to green mould.

FIGURE 3.

Functional analysis of Csi‐miR3954b in regulating green mould resistance in Citrus sinensis ‘Qicheng’ at 25°C. (a) Schematic of the short tandem target mimic (STTM) construct for Csi‐miR3954b silencing. (b) Relative expression levels of Csi‐miR3954b at 2 days post‐inoculation (dpi) (normalized to control = 1). (c) Representative fruit images, (d) disease incidence and (e) lesion diameter at 3–5 dpi. Empty pCAMBIA2300 vector served as control. Data are presented as means ± SD. Statistical significance was determined using one‐way ANOVA followed by Dunnett's multiple comparisons test (*p < 0.05, **p < 0.01).

2.4. Csi‐miR3954b Directly Cleaves CsCE70 and CsPLA1 in Citrus Peel

To identify the potential target of Csi‐miR3954b, prediction was combined with experimental validation. Sequence comparison showed that Csi‐miR3954b had strong complementarity with the transcripts of CsCE70 and CsPLA1 (Figure 4a,h). Consistent with this prediction, the degradation sequencing detected clear cutting features at the predicted binding site, highlighted by the red dots of the peak and the marked cutting position (Figure 4b,i). The interactions were further verified by a transient expression experiment in N. benthamiana. As shown in Figure 4c,j, reporters carrying wild‐type or mutated target sequences were used to evaluate the interaction between Csi‐miR3954b and its predicted targets. In N. benthamiana leaves, the luminescence signal of the region co‐expressed with pre‐miR3954b and CsCE70‐ or CsPLA1‐LUC was significantly weaker than that of the region co‐expressed with the mutation of the target site, while the luminescence signal strength of the latter was not significantly different from that of the empty vector expression region (Figure 4d,k). The co‐expression of pre‐miR3954b with CsCE70‐ or CsPLA1‐LUC significantly reduced the activity of LUC, while the mutation of the target site attenuated this inhibitory effect (Figure 4e,l). To assess physiological relevance, the regulatory relationship was examined in citrus peel. Transient overexpression of Csi‐miR3954b reduced the transcription levels of CsCE70 and CsPLA1, while STTM3954b led to their upregulation (Figure 4f,g,m,n). Taken together, these results showed that Csi‐miR3954b directly targeted CsCE70 and CsPLA1 in citrus.

FIGURE 4.

Validation of CsCE70 and CsPLA1 as targets of Csi‐miR3954b. (a, h) Predicted pairing between Csi‐miR3954b and the transcripts of CsCE70 (a) and CsPLA1 (h) based on psRNATarget analysis. (b, i) Degradome sequencing showing cleavage signals at the predicted target sites (red dots). (c, j) Schematic representation of the dual‐luciferase reporter constructs. Mutated constructs were generated by introducing multiple mismatches within the predicted cleavage site (positions 10–11) to disrupt complementarity. (d, k) Representative luciferase (LUC) luminescence images from Nicotiana benthamiana leaves co‐infiltrated with pre‐miR3954b and reporter constructs containing the wild‐type or mutated target sequences. (e, l) Relative LUC activity measured in leaves expressing the indicated effector and reporter constructs. (f, g, m, n) Relative expression levels of CsCE70 and CsPLA1 in citrus fruit transiently overexpressing (f, m) Csi‐miR3954b or (g, n) STTM3954b. Relative expression levels were normalized to the control (set to 1). Data are presented as means ± SD. Different lowercase letters indicate significant differences among treatments (one‐way ANOVA with Tukey's test, p < 0.05). Asterisks indicate significant differences compared with the corresponding control; **p < 0.01 (Student's t‐test).

To further characterize the potential functions of the two target genes, protein annotation and domain analyses were performed. CsCE70 was annotated as a probable carbohydrate esterase At4g34215‐like protein and was classified into the carbohydrate esterase 6 (CE6) family, containing SGNH hydrolase‐related and SASA domains. Based on homologous protein annotation, CsCE70 was assigned an EC classification of EC 3.1.‐.‐. In contrast, CsPLA1 was annotated as a phospholipase A1‐Igamma1 protein and contained conserved α/β hydrolase and Lipase_3 domains. Based on homologous protein annotation, CsPLA1 was assigned an EC classification of EC 3.1.1.‐. Neither CsCE70 nor CsPLA1 was predicted to encode a transcription factor. Subcellular localization prediction suggested that CsCE70 and CsPLA1 may localize to the cytoplasm and chloroplast, respectively (Table S3).

2.5. CsCE70 and CsPLA1 Are Key Targets of Csi‐miR3954b in Regulating Citrus Fruit Resistance

To investigate whether CsCE70 and CsPLA1 functioned as targets of Csi‐miR3954b in regulating citrus resistance to P. digitatum , transient overexpression assays were performed in citrus fruit. The effective transient overexpression of Csi‐miR3954b and its targets CsCE70 and CsPLA1 was confirmed by RT‐qPCR (Figure 5a,e). Phenotypically, OE‐miR3954b significantly suppressed disease severity and progression in fruit, whereas OE‐CsCE70 had the opposite effect. Co‐expression of Csi‐miR3954b with CsCE70 or CsPLA1 resulted in disease phenotypes that were intermediate between those observed for miR3954b overexpression and single‐gene overexpression fruit (Figure 5b,f). Measurements of disease incidence and lesion diameter confirmed these observations. At 5 dpi, disease incidence in fruit co‐expressing Csi‐miR3954b and CsCE70 was 75%, falling between Csi‐miR3954b overexpression alone (58.3%) and CsCE70 overexpression (83.3%), with the control at 73.3% (Figure 5c). Lesion diameter followed a similar pattern. Lesion diameter in the co‐expression group was intermediate (75 mm) relative to Csi‐miR3954b alone (69.4 mm), CsCE70 alone (97.4 mm) and the control (89.2 mm) (Figure 5d). Similarly, at 5 dpi, fruit co‐expressing Csi‐miR3954b and CsPLA1 showed a disease incidence of 70%, which was lower than that in CsPLA1‐overexpressing fruit (81.7%) but higher than that in fruit overexpressing Csi‐miR3954b alone (58.3%). In comparison, the incidence in control fruit reached 73.3% (Figure 5g). At the same time, fruit co‐expressing Csi‐miR3954b and CsPLA1 showed lesion diameters intermediate between those observed in Csi‐miR3954b overexpressing fruit and CsPLA1 overexpressing fruit. The former showed a decrease in lesion diameter, while the latter showed larger lesion diameter compared to the control group (Figure 5h). These results suggested that CsCE70 and CsPLA1 partially offset the effect of Csi‐miR3954b, supporting that they are functionally relevant targets of Csi‐miR3954b.

FIGURE 5.

Effects of transient co‐overexpression of Csi‐miR3954b and its target genes on green mould resistance in Citrus sinensis ‘Xiacheng’ at 25°C. (a, e) Relative expression levels of Csi‐miR3954b and its target genes (CsCE70 or CsPLA1) at 2 days post‐inoculation (dpi) following transient overexpression of Csi‐miR3954b, target genes, or their co‐overexpression. (b, f) Representative fruit images, (c, g) disease incidence and (d, h) lesion diameter at 3–6 dpi. Relative expression levels were normalized to the control (set as 1). Empty pCAMBIA2300 vector served as control. Data are presented as means ± SD. Different lowercase letters indicate significant differences among treatments (one‐way ANOVA with Tukey's test, p < 0.05). Asterisks indicate significant differences compared with the corresponding control; **p < 0.01 (Student's t‐test).

2.6. Transcriptomic Analysis Reveals That CsCE70 and CsPLA1 Negatively Regulate Distinct Branches of Phenylpropanoid Biosynthesis in Citrus Peel

Transcriptome analysis was performed to investigate the downstream metabolic changes associated with CsCE70 and CsPLA1 overexpression in citrus peel. Compared with the control, 377 differentially expressed genes (DEGs) were identified in OE‐CsCE70 fruit, including 221 upregulated and 156 downregulated genes, whereas 254 DEGs were identified in OE‐CsPLA1 fruit, including 180 upregulated and 74 downregulated genes (|log₂FC| > 1, FDR < 0.05) (Figure 6a,e). KEGG enrichment analysis revealed significant enrichment of the phenylpropanoid biosynthesis pathway among the DEGs in both OE‐CsCE70 and OE‐CsPLA1 fruits (Figure 6b,f). GO analysis supported this, showing enrichment in metabolic processes, catalytic and oxidoreductase activities, terms related to secondary metabolite, and phenolic compound and flavonoid biosynthesis (Figure 6c,g). Pathway models highlighted branch‐specific regulation. In OE‐CsCE70 fruit, several key genes involved in phenolic acid biosynthesis were downregulated, including arogenate dehydratase (ADT), hydroxycinnamoyl transferase (HCT), and caffeic acid O‐methyltransferase (COMT), which contribute to the formation of ferulic acid and chlorogenic acid (Figure 6d). The suppression of these phenolic acid biosynthesis‐related genes upon CsCE70 overexpression suggested that CsCE70 negatively regulates this metabolic branch. In OE‐CsPLA1 fruit, chalcone synthase (CsCHS), cytochrome P450 monooxygenases (CsCYP75B), and O‐methyltransferases (CsOMT) involved in flavonoid biosynthesis were also transcriptionally downregulated (Figure 6h). The suppression of these flavonoid biosynthesis‐related genes upon CsPLA1 overexpression suggested that CsPLA1 negatively regulates this metabolic branch. Together, these results indicated that CsCE70 and CsPLA1 act as negative regulators of distinct branches of phenylpropanoid biosynthesis, with CsCE70 primarily suppressing the phenolic acid branch and CsPLA1 primarily suppressing the flavonoid branch.

FIGURE 6.

Transcriptomic profiling of Citrus sinensis ‘Xiacheng’ fruit following transient overexpression of CsCE70 or CsPLA1. (a, e) Volcano plots of differentially expressed genes (DEGs) in OE‐CsCE70 or OE‐CsPLA1 versus control fruit (|log₂FC| > 1, FDR < 0.05). (b, f) KEGG and (c, g) GO enrichment analyses of DEGs. (d, h) DEGs involved in (d) phenolic acid or (h) flavonoid biosynthesis pathways. Empty pCAMBIA2300 vector served as control. Squares represent individual genes; green indicates significantly downregulated DEGs relative to control (white squares).

2.7. CsCE70 and CsPLA1 Negatively Regulate Citrus Resistance to P. digitatum

To determine the roles of CsCE70 and CsPLA1 in citrus resistance to P. digitatum , OE and virus‐induced gene silencing (VIGS)‐mediated silencing assays were performed in citrus fruit. OE‐CsCE70 significantly increased its transcript level, whereas VIGS‐CsCE70 reduced expression (Figure 7a,e). OE‐CsCE70 fruit showed increased susceptibility, with disease incidence reaching 96.3%, 17.4% higher than the control (79.0%) at 5 dpi (Figure 7b). Lesion diameter increased to 96.8 mm, 10.7% higher than the control (87.5 mm) (Figure 7c). VIGS‐CsCE70 fruit showed increased resistance, with incidence reduced to 74.3%, 15.2% lower than the control (89.6%) at 5 dpi (Figure 7f), and lesion decreased to 64.6 mm, 19.0% smaller than the control (79.8 mm) (Figure 7h). Similarly, OE‐CsPLA1 fruit showed increased susceptibility, with incidence of 90.3%, 10.4% higher than the control (81.8%) at 5 dpi (Figure 7j). Lesion diameter increased to 90.6 mm, 22.2% higher than the control (74.2 mm) (Figure 7k). In contrast, VIGS‐CsPLA1 fruit showed reduced disease with incidence of 75.7%, 11.8% lower than the control (87.4%) at 5 dpi (Figure 7n). Lesion diameter of VIGS‐CsPLA1 fruit was 67.8 mm, 11.2% smaller than the control (76.3 mm) (Figure 7o). Representative symptoms were consistent with these observations (Figure 7d,h,l,p), indicating that both CsCE70 and CsPLA1 negatively regulate citrus resistance to green mould.

FIGURE 7.

Effects of transient overexpression or virus‐induced gene silencing (VIGS)‐mediated transient suppression of CsCE70 and CsPLA1 on green mould resistance in Citrus sinensis ‘Jincheng 447’ at 25°C. (a, e) Relative expression levels of CsCE70 at 2 days post‐inoculation (dpi) following (a) transient overexpression or (e) transient VIGS‐mediated suppression of CsCE70. (b, f) Disease incidence, (c, g) lesion diameter, and (d, h) representative fruit images for transient overexpression or suppression of CsCE70 at 3–5 dpi. (i, m) Relative expression of CsPLA1 at 2 dpi following (i) transient overexpression or (m) VIGS‐mediated suppression. (j, n) Disease incidence, (k, o) lesion diameter, and (l, p) representative fruit images for transient overexpression or suppression of CsCE70 at 3–5 dpi. Relative expression was normalized to the control (set as 1). Empty pCAMBIA2300 and TRV2 vectors were used as controls for overexpression (OE) and VIGS assays, respectively. Data are presented as means ± SD; *p < 0.05, **p < 0.01 (Student's t‐test).

2.8. Csi‐miR3954b Enhances Postharvest Resistance via Repression of CsCE70 and CsPLA1‐Mediated Phenolic Acid and Flavonoid Biosynthesis

To further examine whether the Csi‐miR3954b–CsCE70/CsPLA1 regulatory module affects phenolic acid and flavonoid biosynthesis, the transcript levels of key structural genes identified by transcriptome analysis and the accumulation of their downstream metabolites were quantified during storage. OE‐miR3954b in fruit significantly upregulated genes involved in phenolic acid and flavonoid biosynthesis, whereas STTM‐mediated suppression showed the opposite effect. The expression of CsCOMT, CsADT, CsHCT, CsCHS, CsOMT and CsCYP75B was significantly higher in OE‐miR3954b fruit than in the control at both 2 and 4 dpi. STTM3954b reduced their transcript levels in citrus fruit. Consistent with the transcriptional changes, OE‐miR3954b fruit accumulated higher levels of phenylpropanoid‐derived metabolites. At 4 dpi, the contents of phenolic acids, including total phenolics, ferulic acid and chlorogenic acid, were 27%–39% higher than those in the control, whereas the contents of flavonoids, including total flavonoids, quercetin, hesperidin and tangeretin, were 34%–66% higher. In contrast, the contents of both phenolic acids and flavonoids in STTM3954b fruit were 12%–43% lower than those in the control (Figure 8a). Transient overexpression of CsCE70 suppressed the phenolic acids branch, reducing the expression of CsCOMT, CsADT and CsHCT and decreasing total phenolic acid, ferulic acid and chlorogenic acid by 34%–42% (Figure 8b). Conversely, transient silencing of CsCE70 increased gene expression and the accumulation of these metabolites by 36%–51% (Figure 8c). Similarly, transient overexpression of CsPLA1 downregulated CsCHS, CsOMT and CsCYP75B, resulting in 39%–53% reductions in total flavonoid, quercetin, hesperidin and tangeretin contents (Figure 8d). Transient silencing of CsPLA1 produced the opposite effect, leading to a 43%–70% increase in both gene expression and metabolites accumulation (Figure 8e). Collectively, these findings suggested that Csi‐miR3954b enhances the accumulation of phenolic acids and flavonoids, whereas its targets CsCE70 and CsPLA1 act to suppress these metabolic pathways. This supported a negative regulatory relationship within the Csi‐miR3954b–CsCE70/CsPLA1 module that modulates defence‐related secondary metabolism.

FIGURE 8.

Transcript levels of key phenolic acid and flavonoid biosynthesis pathway genes and associated metabolites accumulation in citrus fruit. (a) Expression of CsCOMT, CsADT, CsHCT, CsCHS, CsOMT, CsCYP75B and contents of total phenolic acid, ferulic acid, chlorogenic acid, total flavonoid, quercetin, hesperidin and tangeretin in fruit transiently overexpressing or silenced for Csi‐miR3954b. (b, c) Expression of CsCOMT, CsADT, CsHCT and contents of total phenolic acid, ferulic acid and chlorogenic acid in fruit (b) transient overexpressing and (c) transient silencing CsCE70. (d, e) Expression of CsCHS, CsOMT, CsCYP75B and contents of total flavonoid, quercetin, hesperidin and tangeretin in fruit (d) transiently overexpressing and (e) transiently silenced for CsPLA1. Empty pCAMBIA2300 and TRV2 vectors were used as controls for overexpression (OE)/short tandem target mimic (STTM) and virus‐induced gene silencing (VIGS) assays, respectively. Relative values were normalized to the control (set to 1). Data are presented as means ± SD. Statistical significance was determined using Student's t‐test for comparisons between two groups (b–e) and one‐way ANOVA followed by Dunnett's multiple comparisons test for comparisons among multiple groups with a control group (a). *p < 0.05, **p < 0.01.

3. Discussion

In this study, we showed that Csi‐miR3954b positively regulates resistance to P. digitatum in citrus fruit. Examples from different crops and pathogens have shown that miRNA is a key post‐transcriptional regulatory factor for plant immunity. Its role covers immune receptor signal transduction, transcription network, and downstream defence responses. The deletion of miR472 in Pseudomonas aeruginosa enhances disease resistance by removing the inhibition of the NLR target gene, while the reduction of miR396 activity offers plants a wider resistance to necrotic fungi (Soto‐Suárez et al. 2017; Vasseur et al. 2024). In wheat, the miR171a–TaSCL6‐1 module enhances the resistance to Puccinia triticina by regulating transcription factors downstream of immune signalling (Guo et al. 2026). On the other hand, in rice, osa‐miR395 acts at the metabolic level as well as promoting antibacterial defence via sulphate metabolism (Yang et al. 2022). The above cases exemplify the multifunctionality of miRNA‐mediated control. Taking this into account, recent work on citrus has started exploring the role of miRNAs in response to diseases. Yet most of this work has been done with expression profiling from either leaves or the entire plant, and only few studies have been functionally validated. Several miRNAs have previously been shown to be responsive to bacterial infection whose predicted targets are enriched for the MAPK and hormone‐related signalling pathways. The overexpression of miR171b enhances the resistance of citrus to huanglongbing (HLB) by targeting SCARECROW‐LIKE transcription factors (Bilal et al. 2024; Lv et al. 2023). Csi‐miR3954 has previously been confirmed to participate in the regulation of citrus flowering time through phasiRNA‐related networks and present dynamic expression during fruit development (Liu, Ke, et al. 2017; Liu, Wang, et al. 2017; Mi et al. 2022). Nevertheless, the roles of miRNA‐mediated regulation in citrus resistance to P. digitatum remain largely unexplored.

In order to explore its potential molecular mechanism, we have studied how Csi‐miR3954b affects its target genes CsCE70 and CsPLA1, as well as the related metabolic pathways of phenolic acids and flavonoids. Functional analysis showed that CsCE70 negatively regulates phenolic acid metabolism by suppressing key biosynthetic genes such as CsADT, CsHCT and CsCOMT, thereby reducing the accumulation of phenolic compounds with antibacterial activity. In contrast, CsPLA1 negatively regulates flavonoid metabolism by suppressing core biosynthetic genes such as CsCHS, CsCYP75B and CsOMT, thereby reducing flavonoid accumulation and weakening citrus resistance. The simultaneous targeting of these two genes suggests that Csi‐miR3954b might coordinate the balance between phenolic acid and flavonoid biosynthesis. The distinct effects of CsCE70 and CsPLA1 on these two metabolic branches may be explained by their predicted biochemical characteristics. CsCE70 was predicted to encode a carbohydrate esterase‐like protein belonging to the CE6 family and containing SGNH hydrolase‐related and SASA domains. In plants, carbohydrate esterase‐related proteins have been reported to participate in the modification of cell wall‐associated carbohydrates and influence cell wall properties (Gou et al. 2012). Ferulic acid, a representative hydroxycinnamic acid, can be ester‐linked to cell wall polysaccharides (Ralph 2010). Therefore, alterations in cell wall‐associated processes may influence phenolic compound accumulation, consistent with the dynamic role of cell wall remodelling in plant stress responses (Tenhaken 2015). Thus, the predicted carbohydrate esterase‐related characteristics of CsCE70 may provide a possible explanation for its association with phenolic acid accumulation. In contrast, CsPLA1 is annotated as a phospholipase A1‐Iγ1 protein and contains conserved α/β hydrolase and Lipase_3 domains. In plants, plastid‐localized PLA1 proteins have been reported to participate in phospholipid hydrolysis and the generation of lipid‐derived signalling molecules (Ellinger et al. 2010; Wang et al. 2018). Lipid‐derived signalling pathways, including jasmonate signalling, have been reported to regulate flavonoid metabolism in a context‐dependent manner (Qi et al. 2011; Wasternack and Hause 2013; Zhao et al. 2023). Therefore, the predicted lipid signalling‐related characteristics of CsPLA1 may provide a possible explanation for its association with flavonoid metabolism during citrus defence responses. Together, these findings suggest that the two target genes may influence distinct branches of secondary metabolism through their predicted biochemical functions. The findings highlight the capability of plant miRNAs to regulate multiple targets involved in complex defence networks. Such a dual target regulation has been reported previously, that is, miRNA can finely tune the expression of multiple genes to optimize the defence response (Silvestri et al. 2024). By targeting CsCE70 and CsPLA1, Csi‐miR3954b may alleviate the inhibitory effects of these negative regulators, thereby facilitating phenolic acid and flavonoid biosynthesis during P. digitatum infection.

Phenolic acid and flavonoid biosynthesis are the two major branches in the phenylpropanoid pathway. The phenylpropanoid pathway is a key part of the secondary metabolism of plants, participating in phenolic acid, lignin and flavonoid biosynthesis. These compounds enhance the defence ability of plants by strengthening cell walls and exerting antibacterial metabolic effects (Zhang, Fan, et al. 2025; Zhang, Xu, et al. 2025). In resistant genotypes, the accumulation of defence‐related metabolites usually increases (Adjei et al. 2021; Ražná et al. 2022; Singh 2025). A similar pattern has also been observed in Cucumis metuliferus , which is resistant to Meloidogyne incognita. The activation of the phenylpropanoid pathway and the increase of CmC4H expression further enhanced its resistance (Wang et al. 2026). Although the current direct evidence on miRNA‐mediated regulation of phenylpropanoid metabolism is still limited, studies in Arabidopsis thaliana and Solanum lycopersicum have shown that miRNA can affect the accumulation of secondary metabolites by regulating MYB transcription factors (Camargo‐Ramírez et al. 2018; Guan et al. 2014; Sharma et al. 2016). Our research results highlight the potential of multiple targets in the metabolic pathways related to miRNA regulation, providing a potential tool to improve plant disease resistance.

Although this study has demonstrated that the Csi‐miR3954b–CsCE70/CsPLA1 regulatory module participates in improving postharvest citrus fruit resistance to P. digitatum , there are still some mechanistic limitations. Although Csi‐miR3954b was identified as a resistance‐associated miRNA through comparison between P. galeiformis‐pretreated and control fruit, the upstream regulatory mechanisms controlling Csi‐miR3954b expression remain unclear. In addition, miRNAs‐mediated regulation is usually affected by genetic background and environmental conditions (Mangrauthia et al. 2017; Jin et al. 2024; Qiao et al. 2021). Future studies should determine whether Csi‐miR3954b‐mediated resistance is conserved across different citrus genotypes, storage conditions, and pathogenic fungi. Beyond these mechanistic questions, the potential application of miRNA‐mediated resistance also deserves further investigation. Currently, postharvest disease management largely relies on external approaches, including edible coatings and preservation treatments, to maintain fruit quality and extend storage life (Xu et al. 2024). Compared with these conventional strategies, miRNA‐mediated regulation represents a potential endogenous approach by enhancing intrinsic fruit resistance. Future studies should further evaluate whether miRNA‐based strategies can also limit disease development after pathogen infection, which may provide additional insights into their potential application in postharvest disease management. Finally, because immune‐related miRNAs also affect plant growth and fruit quality, the potential impact of miR3954b on postharvest citrus fruit quality should be carefully evaluated (Shen et al. 2024; Zhao et al. 2025).

4. Experimental Procedures

4.1. Plant Materials and Pathogen Inoculation

Citrus fruits of three varieties, Citrus sinensis ‘Jincheng 447#’, C. sinensis var. brasiliensis and C. sinensis ‘Xiacheng’, were harvested from orchards in Beibei, Fengjie and Zhong districts of Chongqing, China, respectively. Fruit of uniform size and free of visible defects were selected. The fruit were surface‐disinfected in 2% sodium hypochlorite for 2 min, rinsed with running tap water, air‐dried and wiped with 75% ethanol before use.

The P. digitatum strain used in this study was previously isolated and identified in our laboratory from naturally infected Feng navel oranges collected in Beibei, Chongqing, China, and was preserved as a laboratory strain for subsequent studies (Liu, Ke, et al. 2017; Liu, Wang, et al. 2017). The strain was reactivated on potato dextrose agar (PDA) plates before inoculation. After incubation at 25°C for 7 days in the dark, conidia were harvested with sterile distilled water (SDW). The suspension was filtered through four layers of sterile gauze to remove mycelial debris. The concentration of the conidial suspension was determined using a haemocytometer and adjusted to 1 × 10⁵ spores mL⁻¹ with SDW before inoculation. Two wounds were made at the equator of each fruit on opposite sides using sterile pipette tips, generating punctures of approximately 3 × 3 mm. Agrobacterium tumefaciens harbouring the indicated constructs or the empty vector was infiltrated into the wounds of citrus fruit for transient expression. After 24 h, an equivalent wound was made on the right side and inoculated with 10 μL of P. digitatum, after which the fruit were stored at 25°C to promote disease development and symptom evaluation. Peel tissues that were not inoculated with P. digitatum were collected at 2 and 4 days post‐infiltration and immediately frozen in liquid nitrogen for subsequent analyses.

4.2. Small RNA Sequencing

The small RNA sample used for sequencing is taken from the experiment described by Wei et al. (2023). Sequencing was carried out on the Illumina NovaSeq6000 platform of Beijing Baimaike Biotechnology Co. Ltd. (Beijing, China), and strictly followed the standard operating procedures. The original sequence was first processed to remove the joint sequence and low‐quality base, and only the sequence with a length of 18 to 30 nucleotides was retained for subsequent analysis. Sequencing quality evaluation adopted a variety of indicators, including Q20 and Q30 values, GC content and sequence repetition level. Finally, Bowtie was used to compare the high‐quality sequence with the reference genome, by comparing with the Silva database, GtRNAdb, Rfam and Repbase databases, identifying and removing the corresponding ribosomal RNA (rRNA), transfer RNA (tRNA), small nuclear RNA (snRNA), small nucleon RNA (snoRNA), other non‐coding RNA and repeated sequence reads. The remaining unannotated reads were identified by comparing the known miRNA with the miRBase database and the reference genome. Finally, the abundance of miRNA was quantified by comparing the sequencing reads back to the respective precursor sequences.

4.3. Evaluation of Citrus Resistance to Green Mould Mediated by miRNAs and Their Target Genes

The precursor sequence of candidate miRNAs and the full‐length coding sequences of their corresponding target genes were cloned into the pCAMBIA2300 vector, and the expression vectors were constructed. The STTM sequence that specifically inhibited miR3954b was inserted into the same carrier, and target gene silencing vectors were constructed following the previously reported method (Yan et al. 2012; Chen et al. 2023; Chen, Zhang, et al. 2025; Wang et al. 2024). Primers used for amplification are listed in Table S1. Then, the constructed vectors were transformed into A. tumefaciens EHA105 and transiently expressed in citrus fruit using a modified protocol based on Chen et al. (2023). For transient overexpression and STTM‐mediated miRNA inhibition assays, fruits infiltrated with the empty vector pCAMBIA2300 and subsequently inoculated with P. digitatum were used as the control group. For VIGS assays, fruits infiltrated with the TRV2 empty vector and subsequently inoculated with P. digitatum were used as the control group. Each treatment consisted of three independent biological replicates, with 10 fruits included in each replicate (3 × 10 fruits in total). Disease incidence and lesion diameter were recorded for each fruit.

4.4. Prediction of miRNA Targets

Candidate target genes of miR3954b were predicted using psRNATarget v. 2 (2017 release; http://plantgrn.noble.org/psRNATarget/) with default parameters, following the approach described by Marin et al. (2022).

Degradome libraries were constructed using total RNA extracted from citrus peel samples collected at 2 days after inoculation with the Csi‐miR3954b transient overexpression construct or empty vector control. Sequencing was provided by Hangzhou LC‐Bio Technologies Co. Ltd. The company followed its standard process. In short, the RNA fragment of polyadenosine acidification was captured and the 5′ RNA joint was connected. Then reverse transcription and PCR amplification were carried out to generate degradation group labels. After sequencing on the Illumina platform, the original sequence was trimmed and mass filtered. The obtained high‐quality sequences were compared to the C. sinensis reference genome (v3.0; http://citrus.hzau.edu.cn/index.php). The cut sites indicating the miRNA–mRNA interaction were extracted from the matching label. Based on its comparison results, the potential target genes were identified with the annotated genome sequence.

4.5. Validation of miRNA Targets

4.5.1. Transient Assays in N. benthamiana Leaves

Transient dual LUC reporter assays were carried out in N. benthamiana leaves as a heterologous transient expression system to evaluate the repression of target genes by Csi‐miR3954b (Liu et al. 2014). Empty reporter and effector vectors were obtained from the Li Tian group, Institute of Wheat Research, Chinese Academy of Sciences. Target sequences and the Csi‐miR3954b precursor were inserted into these vectors according to the method outlined in Jian et al. (2022). A. tumefaciens EHA105 carrying the constructs was cultured, resuspended, incubated for 2–3 h and co‐infiltrated into leaves. Leaf samples were collected at 48 h post‐infiltration, and LUC activities were measured as the ratio of firefly (LUC) to Renilla (REN) luciferase.

4.5.2. Total RNA and Small RNA Extraction and cDNA Synthesis From Citrus Peel

Citrus peel tissues collected at different time points were ground under liquid nitrogen. Total RNA was extracted using the OmniPlant RNA Kit (with DNase I), and small RNA was extracted using the SteadyPure Small RNA Kit (Accurate Biology). The extracted total RNA was used for degradome library construction, transcriptome sequencing and gene expression analysis, whereas small RNA was used for small RNA sequencing and miRNA expression analysis. cDNA was synthesized from mRNA with the PrimeScript RT Reagent Kit (Takara) and miRNA with the FastKing RT (with gDNase) Kit (Tiangen Biotech). All cDNA samples were stored at −20°C for downstream analyses.

4.5.3. RT‐qPCR

RT‐qPCR was performed following Chen et al. (2023), with gene‐specific primers listed in Table S2. Relative expression levels of mRNAs and the four selected Csi‐miRNAs were calculated using the 2^−ΔΔCt method with three independent replicates per sample and normalized to Actin and U6 as internal reference genes.

4.6. Bioinformatic Analysis of CsCE70 and CsPLA1 Proteins

The gene sequences of CsCE70 and CsPLA1 were obtained from the NCBI database. Functional annotations of these proteins were obtained from NCBI and InterPro databases. Conserved domains were identified using the NCBI Conserved Domain Database (CDD) and InterPro database. The enzyme commission (EC) classifications were assigned based on homologous proteins identified in the UniProt database. Transcription factor prediction was performed using the Plant Transcription Factor Database (PlantTFDB). Subcellular localization prediction was performed using WoLF PSORT.

4.7. Evaluation of Citrus Fruit Resistance to Green Mould Following Co‐Overexpression of Csi‐miR3954b and Its Targets

The precursor of Csi‐miR3954b and the full‐length coding sequences of CsCE70 or CsPLA1 were individually cloned into the pCAMBIA2300 vector, and the resulting constructs were introduced into A. tumefaciens EHA105. Treatment groups consisted of fruit transiently co‐expressing Csi‐miR3954b with each target gene, whereas control groups included fruit infiltrated with the empty vector or single‐gene overexpression constructs. Disease incidence and lesion diameters were measured and recorded for all fruit.

4.8. Transcriptomics Sequencing

Total RNA was extracted from citrus peel tissues surrounding the Agrobacterium infiltration sites, including the flavedo and part of the albedo layers but excluding pulp tissues, at 2 days after transient expression of CsCE70 or CsPLA1 without subsequent P. digitatum inoculation for transcriptome sequencing analysis. Libraries were prepared and sequenced by Gene Denovo Biotechnology Co. (Guangzhou, China). Reads were aligned to the C. sinensis genome (v3.0) for transcript quantification, and gene expressions were measured as fragments per kilobase of exon per million mapped reads (FPKM) and transcripts per million (TPM) (Trapnell et al. 2010). Differential expression analysis was conducted with DESeq2, and genes with |log₂(FC)| > 1 and false discovery rate (FDR) < 0.05 were considered differentially expressed.

4.9. Determination of Total Phenolics, Total Flavonoids, and Representative Phenolic Acids and Flavonoids in Citrus Peel

Citrus peel samples were collected on the 2nd and 4th days after infiltration for quantitative analysis of total phenolics, total flavonoids and selected phenolic and flavonoid metabolites, including ferulic acid, chlorogenic acid, quercetin, hesperidin and tangeretin. The determination method of total phenolics and flavonoids referred to the method of Meng et al. (2025) and was slightly modified.

High‐performance liquid chromatography with diode array detection (HPLC‐DAD) was used to quantitatively analyse phenolics and flavonoids in citrus peel. The extraction and quantification methods referred to the previously reported methods, and were modified according to the type of each compound (Belajová and Suhaj 2004; Mattila and Kumpulainen 2002). For phenolics, 1 g of peel tissue was extracted with 80% methanol, subjected to ultrasonication, and centrifuged. The supernatant was hydrolysed and acidified, and then extracted with an ether/ethyl acetate mixture. For flavonoid compounds, 1 g of tissue was extracted with 8 mL of 70% methanol, subjected to ultrasonication for 30 min, and centrifuged at 12,000 g for 15 min at 4°C.

Phenolic acids and flavonoids were analysed using a Shimadzu LC‐20 HPLC system equipped with a diode array detector (DAD) and an Agilent SB‐C18 column (4.6 × 250 mm). The column temperature was maintained at 40°C, with a flow rate of 0.8 mL min⁻¹ and an injection volume of 10 μL. For phenolic acids analysis, the mobile phase consisted of 2% acetic acid (solvent A) and methanol (solvent B) with the following gradient programme: 0–10 min, 5%–30% B; 10–25 min, 30%–50% B; 25–35 min, 50%–70% B; and 35–50 min, 70%–5% B. Detection was performed at 320 nm. For flavonoid analysis, 0.5% acetic acid (solvent A) and acetonitrile (solvent B) were used with the following gradient programme: 0–4 min, 20% B; 4–16 min, 20%–35% B; 16–32 min, 35%–75% B; and 32–39 min, 75%–20% B. Detection was performed at 278 nm. Quantification was conducted using the external standard method based on calibration curves generated with authentic standards.

4.10. Statistical Analysis

All data are presented as mean ± SD from at least three independent replicates. Statistical significance was determined using an independent samples t‐test (*p < 0.05, **p < 0.01) for comparisons between two groups. For comparisons among multiple groups, one‐way ANOVA followed by Tukey's multiple comparison test (p < 0.05) was used when all groups were compared with each other, whereas one‐way ANOVA followed by Dunnett's multiple comparisons test (*p < 0.05, **p < 0.01) was performed when each treatment group was compared with a control group. All statistical analyses were performed using IBM SPSS Statistics 26 (SPSS Inc.). Graphs were generated using GraphPad Prism 8.0 (GraphPad Software Inc.).

Author Contributions

Jixin Tian: writing – original draft, investigation, data curation, software, formal analysis, conceptualization, methodology, validation, visualization, writing – review and editing. Xiaoquan Gao: methodology, writing – review and editing, data curation. Jialin Chen: validation, writing – review and editing. Kuo Meng: validation, writing – review and editing. Ou Chen: writing – review and editing, supervision, methodology. Wenjun Wang: methodology, supervision. Lanhua Yi: funding acquisition. Kaifang Zeng: funding acquisition, conceptualization, resources, supervision, project administration, data curation, methodology.

Funding

This study was supported by the National Natural Science Foundation of China (32272376), the Chongqing Natural Science Foundation Innovation and Development Joint Fund Project, China (CSTB2024NSCQ‐LZX0063).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Primer sequences used for constructing vectors.

Table S2: Reverse transcription‐quantitative PCR primer sequences.

Table S3: Functional annotation and predicted subcellular localization of CsCE70 and CsPLA1.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.