Tailored UV-A Irradiation and Callus Selection Enable Distinct Flavonoid Profile Production in Grape Cell Cultures

Abstract

Plant cell culture represents a sustainable platform for the production of high-value natural products. Although ultraviolet A (UV-A) radiation is established as an inducer of phenylpropanoid metabolism, its precise regulatory role in downstream flavonoid biosynthesis within grape cells remains unclear. Using red and white-type callus derived from Vitis vinifera L. cv. Cabernet Sauvignon berry skins, we investigated the effects of UV-A treatments with two durations (45 min and 90 min) on flavonoid biosynthesis. Metabolite profiling demonstrated that UV-A predominantly promoted proanthocyanidin accumulation in white-type callus, while stimulating the global flavonoid pathway in a dose-dependent manner in red callus. Transcriptional analysis identified structural genes potentially governing flavonoid product channeling in both callus types under UV-A exposure. Weighted Gene Co-expression Network Analysis (WGCNA) constructed light-responsive regulatory modules, uncovering potential mechanisms coordinating flavonoid pathway gene expression in response to UV-A. These findings demonstrate how the interaction of callus-type and UV-A shapes flavonoid metabolic flux, providing insights into the regulation of plant cell culture metabolites.

1. Introduction

Flavonoids, a diverse group of polyphenolic compounds, are widely distributed in the plant kingdom and play critical roles in plant physiology, including pigmentation, UV protection, and defense against biotic and abiotic stresses [1,2,3]. Beyond their ecological functions, flavonoids have garnered significant attention in the food industry due to their natural colorants, antioxidant properties, and health-promoting effects [1,4,5]. Among these, proanthocyanidins (PAs) are known for their cardiovascular protective effects [6], anthocyanins for their role in visual health and antioxidation [7], and flavonols for their anti-inflammatory activities [8]. However, conventional methods of flavonoid extraction from field-grown crops are increasingly scrutinized due to environmental concerns such as pesticide residues and variability in product quality [9,10,11,12]. Plant cell culture systems offer a promising alternative, enabling standardized, sustainable production without the risks associated with agricultural contaminants [13,14]. Despite these advantages, the application of plant cell cultures for flavonoid production is hindered by challenges in the targeted regulation of flavonoid subgroups [15]. Current strategies often result in a heterogeneous mixture of flavonoids, limiting the ability to tailor production for specific food applications [16,17,18]. This highlights a critical research gap: the lack of precise methodologies to direct the biosynthesis of desired flavonoid products, such as anthocyanins or flavonols, in plant cell systems.

In fruits, UV radiation has been shown to significantly affect the phenylpropanoid pathway, leading to the upregulation of genes associated with flavonoid and stilbene biosynthesis. For instance, postharvest UV-C treatment specifically induces resveratrol accumulation by activating the stilbene synthase pathway [19,20,21], while both UV-B and UV-C irradiation increase resveratrol levels in harvested berries by up to threefold and twofold, respectively [22]. Moreover, UV-B exposure enhances antioxidant capacity through increased hydrogen peroxide content and specifically promotes flavonol biosynthesis in berry skins as a protective mechanism against UV damage [23,24,25]. Notably, blocking UV radiation reduces flavonol concentrations without affecting PAs composition, underscoring the specific regulatory role of UV in flavonol synthesis in grapevine (Vitis vinifera L.) [24]. Beyond grapes, UV-A has been reported to positively influence the nutritional quality of other crops. In lettuce, UV-A irradiation increases chlorophyll, soluble protein, soluble sugars, vitamin C, flavonoids, polyphenols, and anthocyanin contents [26,27], while in blueberries, it promotes the accumulation of PAs and anthocyanins [28]. These findings suggest that UV-A may serve as a non-thermal, environmentally relevant stimulus for targeted metabolic engineering of health-promoting compounds in horticultural crops.

Transcriptomic studies further support the notion that UV radiation regulates a broad array of metabolic and developmental processes. In grape berries, solar UV exposure enhances ripening-related transcriptional responses and phenolic accumulation, indicating that natural UV components are essential for optimal fruit development and quality formation [29,30]. Additionally, dynamic acclimation to UV radiation involves rapid and sustained changes in secondary metabolites such as tocopherols, polyamines, and flavonoids, reflecting an adaptive strategy to maintain redox homeostasis under fluctuating light conditions [31]. UV-A exposure promotes the formation of acylated anthocyanins, which are key pigments that enhance color stability in wines. This process involves the upregulation of genes such as O-methyltransferase (VviOMT) and glutathione S-transferase 4 (VviGST4). The transcriptional activation is likely controlled by MYB transcription factor A1 (VviMYBA1) [32]. Concurrently, structural genes including Chalcone synthase 2 (VviCHS2) and glutathione S-transferases (VviGSTs) show increased expression under elevated UV radiation [33]. This enhanced expression further amplifies the metabolic flux through the phenylpropanoid pathway.

The perception of UV radiation is mediated by photoreceptors such as UV resistance locus 8 (UVR8), which detects both UV-B and portions of UV-A spectra [34,35]. Upon activation, UVR8 initiates a signaling cascade culminating in the stabilization of key basic leucine zipper (bZIP) transcription factors, notably elongated hypocotyl 5 (HY5) and HY5-homolog (HYH) [36,37]. These factors act as central integrators of light signals, directly binding to promoters of flavonoid structural genes—such as chalcone synthase (CHS), flavanone 3-hydroxylase (F3H), dihydroflavonol 4-reductase (DFR), and leucoanthocyanidin dioxygenase (LDOX)—to activate their expression [36,38,39]. Field experiment revealed that Arabidopsis UVR8 is responsible for leaf development and the accumulation of flavonols and hydroxycinnamic acid derivatives [40]. In the presence of phototropin 1, UVR8 mediates flavonoid accumulation in response to low UV-A and blue light [41]. Notably, HY5/HYH abundance is markedly reduced under blue-light and UV-deprived conditions, underscoring their indispensability in light-mediated flavonoid regulation [42].

Despite these advances, critical gaps persist regarding tissue-specific and dose-dependent responses to UV-A in grapevines. For example, while UV-C radiation enhances flavanol synthesis by activating MYB transcription factor PA1 (VviMYBPA1), which directly binds leucoanthocyanidin reductase 1 (VviLAR1) and anthocyanidin reductase (VviANR) promoters to upregulate enzyme activity and metabolite accumulation [43]. The analogous mechanisms for UV-A remain poorly defined. Moreover, the interplay between UV-A perception and downstream MYB transcription factors—such as VviMYBA1, VvMYBPA1—may involve combinatorial regulation with HY5/HYH [44], yet empirical validation in grapes is lacking. Additionally, the role of UV-A in modulating acyltransferase genes and glutathione transferases, which facilitate anthocyanin transport and stabilization, warrants further investigation.

Given this growing body of evidence, there is a critical need to better understand how UV-A radiation influences flavonoid profiles in economically important crops like grapevine. Although significant progress has been made in the study of the mechanism of plant response to UV radiation, especially in the aspects of UV-B perception and signal transduction pathways, there are still important gaps in the systematic understanding of how plants coordinate to respond to UV-A radiation at different organizational levels. Specifically, the dose threshold of UVA-induced metabolic reprogramming, at which light intensity is sufficient to trigger significant physiological changes, is still unclear. At the same time, the differential response mechanism of tissues with or without anthocyanins to UV-A is also lacking in-depth analysis, which limits our understanding of the local light protection strategy of plants. In addition, how UV-A interacts with known optical signal networks to regulate the transcriptional reprogramming of the flavonoid biosynthesis pathway is still a key issue that has not been fully explored. The existence of these problems highlights the necessity of integrating plant photobiological response from the molecular to the tissue level.

In order to fill the above knowledge gaps, this study aims to systematically analyze the dynamic regulatory network of plant metabolism and gene expression under UV-A radiation. Understanding these phenotype-specific and UV-A dose-dependent regulatory networks is essential to harness callus-derived cell factories for sustainable production of high-value flavonoids, such as PAs, anthocyanins, and flavonols. In light of these considerations, this study proposes a novel approach to optimize flavonoid production in grape callus cultures by synergistically integrating UV-A irradiation with specific color types of callus tissues. We hypothesize that regulated modulation of UV-A dosage, combined with the inherent metabolic potential of distinct grape callus tissues, can selectively enhance the biosynthesis of anthocyanins or flavonols. This study reveals a callus-type–specific and UV-A dose-dependent reprogramming of flavonoid biosynthesis in grape: red-type callus preferentially accumulates anthocyanins under low UV-A, whereas white-type callus channels metabolic flux toward proanthocyanidins in a high-dose-responsive manner. This differential metabolic allocation, not previously documented in grape cell cultures, provides new insight into how cellular context shapes light-driven phenylpropanoid branching. This strategy aims to address the current limitations in targeted flavonoid production and contribute to the development of efficient, sustainable biotechnological platforms for functional food applications.

2. Materials and Methods

2.1. Cell Cultures and UV-A Treatment

The callus induction was performed according to established protocols as previously described [45]. Briefly, grape berries (Vitis vinifera L. cv. Cabernet Sauvignon) were harvested at the E-L 33 developmental stage from the Shangzhuang Experimental Station of China Agricultural University, Beijing, China. After collection, berries were stored at 4 °C and transported to the laboratory within 2 h. The pedicels were removed, and berries were washed thoroughly under running tap water, followed by surface sterilization in a laminar flow hood using 75% (v/v) ethanol for 30 s and 0.8% sodium hypochlorite solution with Tween 20 for 15 min. Sterilized berries were rinsed with sterile distilled water, and the exocarp was aseptically excised into pieces (approximately 5 × 5 mm) and placed onto induction medium. The callus induction and subculture medium was based on the B5 basic medium (3.21 g/L, pH 5.8 to 6.0) according to the method of [45,46] with slight modifications, supplemented with 20 g/L of sucrose, 0.25 g/L of casein enzymatic hydrolysate, 0.1 mg/L of α-naphthaleneacetic acid (NAA), 0.2 mg/L of kinetin (KT), and 3 g/L of phytogel. The cultures were maintained under a 16/8 h (light/dark) photoperiod at 25 ± 1 °C, with a light intensity of 1000–2000 lx provided by cool white fluorescent lamps. The first subculture was carried out 40 days after initial inoculation, followed by subsequent subcultures every 25 days. Building on this established system, we subsequently obtained two distinct callus phenotypes during subculture. A reddish-type capable of light-dependent anthocyanin accumulation and a white-type deficient in anthocyanin synthesis under identical conditions (Figure 1). These lines were stabilized through selective propagation and used for all subsequent experiments.

To assess the dose-dependent responses of white (W) and red (R) grape calli, tissues were exposed to UV-A radiation for 0 min (control, Ctrl), 45 min (low UV-A, L), and 90 min (high UV-A, H). The UV-A illumination system was equipped with PHILIPS TL-D 15W ACTINIIC BL lamps (emission spectrum: 350–400 nm; peak wavelength: 365 nm), mounted 40 cm above the culture plates. The irradiance was calibrated to 3 μW/cm², resulting in total fluences of 8.1 mJ/cm² (45 min) and 16.2 mJ/cm² (90 min) for the low- and high-dose treatments, respectively. Control samples received no UV-A exposure (0 μW/cm²). Lamps were arranged to ensure uniform irradiance coverage over all samples. All irradiation procedures were performed under sterile conditions with the open lid of Petri dishes to avoid spectral interference. Each treatment was replicated three times biologically. Following UV-A exposure, the callus tissues were kept in the dark for 48 h to promote flavonoid accumulation. Samples were flash-frozen in liquid nitrogen and stored at −80 °C until metabolite and RNA extraction.

2.2. Flavonoids Profiling Using LC/MS

To analyze the composition of flavonoids (flavonols, flavan-3-ols and anthocyanins), callus tissues were ground in liquid nitrogen and subsequently freeze-dried. For the extraction of flavonols and anthocyanins, 50 mg of the freeze-dried powdered sample was extracted with 1 mL of 50% (v/v) methanol/water for 20 min by ultrasonication (40 kHz, 20 min, 4 °C). The homogenate was centrifuged at 8000× g for 5 min (Eppendorf 5424R, Eppendorf AG, Hamburg, Germany), and the supernatant was collected for analysis. To extract free flavan-3-ols, 50 mg of the freeze-dried sample powder was thoroughly shaken in 1 mL of 70% acetone aqueous solution containing 0.5% (w/v) ascorbic acid. The mixture was then centrifuged at 8000× g for 10 min at 4 °C, and this extraction process was repeated three times. Combined supernatants (400 μL) were evaporated under nitrogen gas and then resuspended in 200 μL of methanol containing 1% HCl (v/v) and 200 μL of 0.2 mol/L sodium acetate aqueous solution. For the cleavage experiment of polymeric flavan-3-ols, 50 mg of dry powder was suspended in phloroglucinolysis buffer (0.5% (w/v) ascorbic acid, 0.3 mol/L HCl, and 50 g/L phloroglucinol in methanol) and incubated for 20 min at 50 °C. The reaction was terminated by adding 200 mM sodium acetate, followed by centrifugation at 10,000× g for 15 min at 4 °C. Flavonoid profiling was performed using an Agilent 1200 HPLC system coupled to a 6410 Triple Quadrupole (QqQ) mass spectrometer. Separation was achieved on a Poroshell 120 EC-C18 column (150 × 2.1 mm, 2.7 μm, Agilent Technologies) maintained at 35 °C. Targeted LC/MS analysis of flavonoid components was carried out using the multiple reaction monitoring (MRM) technique, with specific methods based on our previous studies [47].

2.3. Determination of PA Content

PAs accumulation in grape callus tissues was visualized using p-dimethylaminocinnamaldehyde (DMACA) staining following the published methods [48]. Three biological replicates were applied for each set of samples. Soluble and insoluble PAs were extracted and quantified as described by the methods previously [49], and spectrophotometric quantification was performed at 640 nm and 550 nm using a VICTOR Nivo multimode plate reader (PerkinElmer). (+)-Catechin (Sigma, Saint Louis, MO, USA) and procyanidin B1 (Sigma, Saint Louis, MO, USA) were used as standards, respectively.

2.4. RNA Extraction, Library Construction, and Sequencing

Total RNA was isolated from calli using the Universal Plant Total RNA Extraction Kit (BioTeke, Beijing, China). RNA integrity was verified using an Agilent Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, USA), with all samples exhibiting RNA Integrity Numbers (RIN) ≥ 8.0. Quantification was performed via NanoDrop ND-2000 spectrophotometry (Thermo Fisher Scientific, Waltham, MA, USA). Strand-specific cDNA libraries were prepared using the NEBNext Ultra II RNA Library Prep Kit (New England Biolabs, Ipswich, MA, USA) according to the manufacturer’s protocols. Sequencing was conducted on an Illumina NovaSeq 6000 platform (Illumina Inc., San Diego, CA, USA) at Novogene Biotech Co., Ltd. (Beijing, China), generating 150-bp paired-end reads with a target depth of 40 million reads per sample. The raw sequencing reads generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under the Bioproject accession number PRJNA1357836.

2.5. Transcript Assembly and Quantification of Gene Expression Level

Raw sequencing data were filtered to remove low-quality reads. Adapter sequences and poly-N sequences were also removed. This process generated clean reads. Clean reads were aligned to the Vitis vinifera L. reference genome (version V3, https://www.grapegenomics.com/). Alignment used HISAT2 software (version 2.0.5) with default settings. Transcript assembly was performed using StringTie (version 1.3.3b). This included reference-based assembly and novel transcript prediction. Gene read counts were obtained using FeatureCounts (version 1.5.0-p3). Gene expression levels were calculated as the fragments per kilobase of transcript per million mapped reads (FPKM) values. Heatmaps visualizing gene expression levels were generated using the integrated heatmap tool within the TBtools software platform [50].

2.6. Differential Expression Analysis and Gene Functional Enrichment

Differentially expressed genes (DEGs) were identified using the R package DESeq2 (1.20.0). Raw read counts from RNA-seq data were used as input, and the analysis incorporated median-of-ratios normalization, dispersion estimation, and fitting of a negative binomial generalized linear model. The Benjamini–Hochberg procedure was applied to adjust p-values for multiple testing, yielding adjusted p-values (padj). Genes with padj < 0.05 and |log₂(fold change)| > 1 were defined as statistically significant DEGs. It should be noted that all statistical inferences were based on raw count data using DESeq2’s negative binomial framework, which appropriately accounts for overdispersion in RNA-seq data and does not rely on assumptions of normality or homoscedasticity required by classical parametric tests (e.g., t-test or ANOVA). FPKM values reported in figures are provided solely for visualization purposes and were not used in any statistical testing.

Functional enrichment analysis of DEGs was performed using TBtools (v2.371) [50], with Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways considered significantly enriched at padj < 0.05. Results were visualized as bar plots and bubble diagrams via the Hiplot platform (https://hiplot.org). All analyses included three biological replicates per group to ensure statistical robustness.

2.7. Gene Co-Expression Analysis

Weighted gene co-expression network analysis (WGCNA) was conducted using the R package WGCNA [51] to explore gene modules significantly associated with specific flavonoid metabolites. Prior to network construction, lowly expressed genes were filtered out by retaining only those with raw read counts ≥ 10 in at least 50% of the samples, resulting in 17,099 genes for downstream analysis. To meet the normality and homoscedasticity assumptions required for Pearson correlation-based network inference, variance-stabilized transformed (VST) expression values were generated from the raw count matrix using DESeq2. The soft-thresholding power (β = 12) was selected based on the scale-free topology criterion (scale-free topology fit index R² > 0.90). Hierarchical clustering was performed with a cut height of 0.40 for module detection, and modules containing fewer than 20 genes were merged into the “grey” (unassigned) module to ensure biological relevance and statistical robustness.

Module–metabolite associations were assessed by calculating Pearson correlation coefficients between module eigengenes and flavonoid metabolite levels. Modules exhibiting significant correlations (|r| > 0.7 and p < 0.05) were selected for functional characterization. Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed on genes within these modules using TBtools [50], with terms considered significant at padj < 0.05.

Hub genes were identified based on the module connectivity (kME > 0.8) and gene significance (GS > 0.5) for the traits of interest. Co-expression subnetworks were constructed for differentially expressed transcription factors (TFs) that met hub criteria and showed strong connections to structural genes involved in flavonoid biosynthesis. These interactions were visualized using Cytoscape (v3.8.2).

2.8. Flavonoid-Related Transcription Factor Annotation

Transcript factor (TF) genes in DEGs were identified using BLASTN v2.13.0 (e-value ≤ 1 × 10⁻⁵) against the Vitis vinifera L. genome database in Plant Transcription Factor Database (PlantTFDB) V5.0. For further annotation of the identified transcription factor genes, BLASTX comparison was performed against the joint database composed of the NCBI non-redundant database and the experimentally verified flavonoid transcription factor sequences manually collected based on the published literature. The above analysis was carried out using the NCBI BLAST standalone version 2.16.0+ program with default parameters.

2.9. Statistical Analysis

Statistical analyses were performed using SPSS Statistics v26.0 (IBM Corp). Prior to conducting parametric tests, the assumptions of normality and homogeneity of variances (homoscedasticity) were rigorously evaluated for each dataset. Normality was assessed by examining the distribution of model residuals using the Shapiro–Wilk test, while homoscedasticity was verified using Levene’s test. All metabolite datasets satisfied both assumptions (Shapiro–Wilk p > 0.05; Levene’s p > 0.05). For comparisons involving three levels of UV-A treatment (control, low, high) within the same callus type (white-type or red-type), a one-way analysis of variance (ANOVA) was performed, followed by Duncan’s multiple range test (α = 0.05) for post hoc pairwise comparisons. For direct two-group comparisons between white-type and red-type calli under identical UV-A conditions (e.g., CKW vs. CKR), an independent two-tailed Student’s t-test was applied. Statistical significance was defined as p < 0.05. Data visualization was conducted with GraphPad Prism v9.0.0 (GraphPad Software).

3. Results

3.1. Compositional Analysis of Flavonoids in Grape Calli Under UV-A Treatment

The flavonoid composition in grape cells depends on both genetic variation and environmental stimuli. In a previous study, we developed a callus system derived from the berry skins of Vitis vinifera L. cv. Cabernet Sauvignon, which was used to investigate the interaction between flavonoid accumulation and environmental factors [46,52]. Originally, the anthocyanin pathway can be activated under light in grape calli (red-type). During the subculture of this system, we obtained a callus line (white-type) that fails to synthesize anthocyanins under light exposure (Figure 1). To assess exactly how UV-A signal affects the flavonoid channeling in these grape cells, we then subjected the pre-light-treated red- and white-type grape calli to UV-A exposure with three durations: 0 min (control), 45 min (low UV-A, L), and 90 min (high UV-A, H). Within 48 h post UV-A treatment, no obvious color changes were observed in all the calli (Figure 1). We further measured the flavonoid composition of these calli using the LC/MS method developed by our group [47,53]. The results showed that the levels of anthocyanin and flavonol in red-type calli were at least twice those in white-type calli, respectively (Figure 2). In white-type calli, anthocyanin content decreased with increasing UV-A exposure dose. As for red-type calli, anthocyanin content peaked under low UV-A treatment, and with high UV-A treatment, it dropped to a level similar to that of the non-UV-treated control (Figure 2a). Despite the difference in flavonol content between the two types of calli, their trend of flavonol accumulation in response to UV-A was quite similar—high UV-A could slightly promote flavonol synthesis, whereas low UV-A to some extent suppressed flavonol accumulation (Figure 2b). Soluble and insoluble PAs were quantified via the DMACA and Butanol-HCl methods, respectively. Results showed that UV-A could induce PA biosynthesis in both callus types. Specifically, the contents of soluble and insoluble PAs in white-type calli were positively correlated with UV-A treatment duration. While the soluble PA content in red-type calli followed the same trend as that in white-type calli, the insoluble PA content in red-type calli increased under short-duration UV-A exposure but failed to increase further with prolonged exposure. Unlike anthocyanins and flavanols, soluble PA content in white-type calli was slightly higher than that in red-type calli under each condition. For insoluble PAs, while their content in white-type calli was only half that in red-type calli under control conditions, this difference was diminished with UV-A treatment (Figure 3b). Regarding the compositions of anthocyanins, flavonols, and flavan-3-ols, their backbones are primarily based on cyanidin, quercetin, and catechin/epicatechin, respectively. This indicates that the metabolic flux of flavonoids containing two hydroxyl groups (-OH) is predominant (Figure 2).

3.2. UV-A Induced Gene Expression in Flavonoid Pathway

Transcriptome analysis was performed with Illumina RNA-Seq to investigate the response of gene expression to UV-A treatment in the grape callus system. Transcripts were assembled from pair-end reads based on the grape PN40024.V3 genome, and their expression levels were normalized by Fragments Per Kilobase Million (FPKM). Quality control of RNA-seq data confirmed high stability, with biological replicates showing pairwise Pearson correlations ≥ 0.95 (Supplementary Figure S1). Differentially expressed genes (DEGs) between two given conditions were identified using DESeq2 (|log₂(FC)| > 1, padj < 0.05).

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed that DEGs were significantly enriched in phenylpropanoid-flavonoid biosynthesis, photosynthesis, and circadian rhythm pathways, which were mainly affected by UV-A treatment in both red- and white-calli in a dose-dependent manner. Correspondingly, we identified four differentially expressed phenylalanine ammonia-lyase (PAL) gene isoforms under UV-A treatment, whose expression levels showed reciprocal complementary changes with increasing UV-A dose. A similar pattern was observed for 4-coumarate: CoA ligase (4CL) genes, whereas this was not the case for cinnamate-4-hydroxylase (C4H). Instead, C4H exhibited increased expression across both calli types upon UV-A exposure, with the highest FPKM values detected in high UV-A treatments. Together, these findings indicate fine-tuning of flux channelling toward the upstream steps of the flavonoid pathway. From chalcone synthase (CHS) genes extending to the anthocyanin and PAs biosynthetic branches, our results demonstrated that while UV-A upregulated the transcripts of a range of flavonoid genes, some of these genes had higher transcript levels in red-type calli than in white-type calli. In general, the expression levels of anthocyanin and PA genes peaked under low UV-A in red-type calli, whereas their expression increased consistently with increasing UV-A dose in white-type calli. This may partially explain why low UV-A treatment yielded more anthocyanin than high UV-A treatment in red-type calli, and why PA levels in white-type calli increased consistently with increasing UV-A dose. Recently, Glutathione S-transferase 4 (GST4) identified in grapevine has been found to exhibit the activity of converting the intermediate product of ANS into anthocyanidin [54,55]. Notably, the transcript level of GST4 in white-type calli is much lower than that in red-type calli, suggesting that this discrepancy constitutes a critical determinant of the differential anthocyanin accumulation between the two callus types. The anthocyanin and PAs pathways compete for substrate channeling. Consequently, PA accumulation is enhanced in white-type calli, stemming from the failure of products at the competitive node to proceed downstream of ANS. Instead, these substrates are sequestered and diverted by LAR proteins. For the anthocyanin modification genes, anthocyanin O-methyltransferase (AOMT) and anthocyanin acyltransferase (AAT), their expression levels are consistently lower in white-type calli than in red-type counterparts. This indicates a global synergistic effect in the gene regulation of the anthocyanin metabolic pathway under UV-A treatment. Although flavonols act as photoprotectants by compensating for anthocyanins in white grape cultivars under high light, their accumulation in white-type calli cultivars was not higher than that in red-type calli cultivars in our study. This finding may be attributed to the fact that UV-A is not a key environmental signal for activating flavonol biosynthesis. (Figure 4b).

To identify the potential regulators that govern the flavonoid biosynthesis under UV-A in grape calli, we analyzed the expression patterns of known transcriptional regulators of the phenylpropanoid-flavonoid pathway (Figure 4c). Among the MYB transcription factors, VviMYB5a and VviMYB5b showed moderate but consistent upregulation in response to UV-A treatment across both callus types, functioning as known transcriptional activators in the flavonoid biosynthetic pathway [56,57]. The expression of VviMYBPA1, a key transcriptional activator of PA biosynthesis [58], was induced by low UV-A treatments but exhibited reduced expression in high UV-A treatments compared to the low UV-A samples. The UV-A-induced accumulation of PAs was concurrent with the activation of their repressor genes, VvMYBC2-L1, VvMYBC2-L2, and VvMYBC2-L3 [59]. Notably, VvibHLH93, the repressor in the PA pathway [60], was down-regulated by UV-A in both white and red-type calli, concomitant with the activation of the expression of PA structural genes. MYB24 acts as a flavonol biosynthesis activator, compensating for inadequate anthocyanin in white grape varieties under high light [61]. Notably, MYB24 induction was absent in white-type calli yet present in red-type calli, which partially results in the differential expression of FLS across the two callus types. Consistent with the activation of flavonoid biosynthesis, GO enrichment analysis revealed that UV-A exposure—particularly at high intensity—significantly upregulated genes involved in defense responses, cell wall remodeling, and transcriptional regulation, while low-dose treatment primarily affected photosynthesis-related components (Supplementary Figure S2). UV-A treatments induced several glutathione S-transferase (GST) genes, while expression of core ROS-scavenging enzymes (SOD, CAT, APX) remained largely unchanged (Figure S3). This coincides with strong upregulation of flavonoid pathway genes (e.g., VviCHS, VviFLS, VviANS) and increased accumulation of corresponding flavonoids—such as flavonols in white callus and anthocyanins in red callus (Figure 2 and Figure 3)—which are known to function as endogenous antioxidants. The data suggest that UV-A elicits a coordinated response in which mild redox signaling activates flavonoid biosynthesis as a primary defense, potentially reducing the need for enzymatic ROS detoxification.

3.3. Expression Patterns of Light Signaling Receptors and Components in Grape Calli Under UV-A Treatment

Exactly how the UV-A signal is sensed by the grape cell is not fully understood. Transcriptomic analysis revealed that UV-A treatment modulated the transcript levels of key light-signaling genes in both red- and white-type grape calli. These included photoreceptors such as cryptochrome 2 (CRY2), zeitlupe (ZTL1), zeitlupe 2 (ZTL2), UV resistance locus 8 (UVR8), and phototropin 1 (PHOT1); components of the ubiquitin ligase complex, including constitutive photomorphogenic 1 (COP1) and suppressor of phytochrome A (SPA1); as well as downstream transcription factors HY5-homolog (HYH) and elongated hypocotyl 5 (HY5) (Figure 5).

VviCRY2, a blue light receptor whose expression has been correlated with anthocyanin accumulation in grape berries under light stress [62,63,64], exhibited a significant increase in expression in LW calli but decreased under high UV-A in white calli, suggesting a dose-responsive expression pattern. In contrast, VviCRY2 remained highly expressed in red-type calli across UV-A treatments, a pattern that coincides with their inherent pigmentation and elevated baseline flavonoid content. VviZTL1 showed minimal variation across treatments, while VviZTL2 displayed a consistent dose-responsive increase, a pattern consistent with involvement in UV-A-responsive circadian or flavonoid-associated processes. The canonical UV-B photoreceptor VviUVR8 showed modest induction under low UV-A, consistent with reports of UVR8-mediated responses to longer UV wavelengths in grapevine [65,66]. VviPHOT1, involved in phototropin-mediated blue light perception [67], decreased under high UV-A, potentially reflecting feedback inhibition under prolonged irradiation.

The E3 ubiquitin ligase COP1 and its cofactor SPA1, which form a complex to degrade HY5 in darkness and under low light [68,69,70], exhibited coordinated upregulation with increasing UV-A dose. This transcriptional co-induction is consistent with a potential role for COP1/SPA1 in modulating HY5 stability under high UV-A, which could influence the trade-off between growth and flavonoid accumulation—a hypothesis supported by studies in other plant systems [65,71].

VviHY5 and VviHYH, transcription factors previously associated with light signaling and anthocyanin accumulation in grapevine [65,66,71,72], displayed a consistent upregulation in response to UV-A, particularly in the high UV-A treatments, suggesting activation of light-responsive transcriptional networks. This aligns with previous reports linking HY5 to anthocyanin biosynthesis and photoprotection in grape and other plant species [73]. Collectively, the data show that short-term UV-A exposure is associated with differential expression of blue-light and UV-A photoreceptors, concurrent with changes in COP1/SPA1 transcript levels and HY5/HYH accumulation—patterns consistent with conserved light-signaling dynamics observed in grape [65,66]. The stronger transcriptional responses in red-type calli coincide with the enhanced flavonoid and anthocyanin accumulation observed in parallel metabolite analyses (Figure 2).

3.4. Co-Expression Network of Photoreceptors and Flavonoid Genes

Weighted gene co-expression network analysis (WGCNA) revealed significant correlations between specific gene modules and flavonoid metabolites in grape calli subjected to UV-A irradiation. The yellow module exhibited strong positive correlations with flavonols, PA extension units, and trihydroxylated anthocyanins, while the blue module showed negative associations with most of these compounds. Notably, the turquoise module was strongly linked to soluble PAs but negatively correlated with other flavonoid types. These findings suggest that distinct transcriptional programs underlie the biosynthesis of different flavonoid subclasses in grape calli.

To link co-expression modules to specific metabolic outcomes, we calculated Pearson correlations between module eigengenes and flavonoid subgroups (Figure 6f). The yellow module exhibited the strongest positive correlation with soluble proanthocyanidin (PA) levels (r = 0.89, p < 0.001) and was significantly enriched in VviLAR1 and VviANR—key structural genes for PA biosynthesis (Figure 4b). Conversely, the turquoise module showed a strong association with anthocyanin accumulation (r = 0.82, p < 0.001) and contained VviUFGT and VviMYBA1, consistent with their established roles [74] in anthocyanin production in grape. Functional enrichment confirmed the yellow module as a hub for integrated regulation: it clustered genes encoding flavonoid pathway enzymes (PAL1/7, CHS, CHI, F3’H, DFR, LAR1, ANR), MYB regulators (MYBPA1, MYBPAR), light-signaling components (CRY2, HYH, CN-PDE1), and photosynthesis-related genes (FNR, PsbO, ELIP1). This co-expression pattern is consistent with a hypothetical model in which UV-A-responsive CRY2 expression is associated with MYB–LAR1 module activity and PA accumulation—a relationship observed in light-exposed grape tissues [62,63]. In contrast, blue-module hubs (ANS, MYB5a/b) were correlated with trihydroxylated anthocyanins and photoreceptors (PhyA, UVR8), whereas turquoise-module genes (bHLH93) showed association with anthocyanin modification and stress responses.

The yellow module exhibited high connectivity, with CRY2 and HYH as central hubs. CRY2 showed strong co-expression with flavonol biosynthetic genes (VviFLS1, VviGST4) and transcription factors, while HYH was highly correlated with VviLAR1 and VviANR—a pattern consistent with their potential involvement in flavan-3-ol and PA metabolism, as suggested by co-expression analyses in grape berries [61,63]. This module features numerous uncharacterized TFs, suggesting novel regulators in flavonol metabolism. In contrast, the blue module highlighted a sparser network focused on anthocyanin regulation, where UVR8 exhibited strong co-expression with VviANS, consistent with their coordinated induction under UV treatment in previous grape studies [65,66]. COP1 is associated with repressor TFs (VviMYBC2-L3), potentially forming a COP1-MYBC2-L3-ANS repression circuit. VviMYB5a/b acted as integrators linking light signals to anthocyanin synthesis, consistent with the module’s correlation with trihydroxylated anthocyanins. The turquoise module featured a streamlined network centered on HY5, which dominantly connected to soluble PAs upstream pathway genes (Vvi4CL, VviC4H) and transporter/enzyme genes (VviGST1, VviGST2, VviGST3). HY5 showed dominant connectivity to upstream phenylpropanoid genes (Vvi4CL, VviC4H) and glutathione transferases (VviGST1–3), and its co-expression with PHOT1 aligns with light-responsive expression patterns reported in grapevine [61,64]. However, regulatory roles remain speculative without biochemical validation. Overall, these co-expression networks highlight candidate gene–metabolite associations that may contribute to UV-A-associated flavonoid reprogramming in grapevine. The proposed light-flavonoid crosstalk axes (e.g., CRY2/HYH with flavonols/PAs; UVR8 with anthocyanin-related genes) are consistent with transcriptomic trends in field-grown grapes exposed to solar UV [29,62,65].

4. Discussion

This study provides a comprehensive dissection of the molecular mechanisms underlying UVA-induced flavonoid metabolic reprogramming in grape (Vitis vinifera L.) callus systems. By leveraging genetically similar red-type (high-flavonoid) and white-type (low-flavonoid) tissues, we effectively eliminated the confounding effects of cultivar variation, revealing that UV-A radiation elicits dose-specific and cell-line-specific effects in flavonoid biosynthesis. Notably, the white-type calli, characterized by their pronounced accumulation of PAs with minimal synthesis of other flavonoids, emerge as a promising dedicated bioreactor for the sustainable production of these commercially valuable compounds. These responses are governed by a complex interplay of photoreceptors, transcription factors (TFs), and structural genes, fundamentally advancing our understanding of how plants fine-tune secondary metabolism in response to specific wavelengths of solar radiation. Our findings demonstrate that UV-A perception is not a linear stimulus-response process but rather involves dynamic, dose-dependent reprogramming of competing flavonoid branches, orchestrated through a sophisticated regulatory framework that integrates light signaling with metabolic priming and resource allocation trade-offs.

A central observation of this study is the inherently biphasic nature of flavonoid responses to UV-A intensity. In white-type calli, characterized by low basal flavonoid levels, low UV-A unexpectedly suppressed flavonol and anthocyanin accumulation. This was in stark contrast to the robust activation of defense-related pathways observed under high UV-A, which resulted in a significant 44% increase in total flavonols (Figure 2b) and a marked surge in PAs (Figure 3). This pattern strongly suggests a threshold effect in photomorphogenic signaling, where only sufficiently intense UV-A stress is capable of activating protective flavonoid branches. Transcriptomic analysis provided mechanistic support: the upregulation of flavonol synthase genes VviFLS5 and VviFLS1 under high UV-A conditions directly correlated with the accumulation of quercetin derivatives (Figure 4 and Figure 6), reinforcing their established functional role in flavonol biosynthesis [74,75,76]. Conversely, the downregulation of VviFLS1 under low irradiance provides a clear molecular explanation for the suppression of flavonol synthesis, highlighting the exquisite sensitivity of pathway regulation to gene dosage. In red-type calli, which possess a pre-primed metabolic state with high flavonoid content, the response was distinctly different. Here, low UV-A preferentially stimulated anthocyanin accumulation, yielding a substantial 42.5% increase (Figure 2a). Notably, this induction was particularly pronounced in acylated anthocyanin forms. This biochemical modification is well-known to confer enhanced stability through steric hindrance, which impedes nucleophilic attack and subsequent hydration-induced degradation [77]. This stability is mechanistically underpinned by the coordinated induction of key genes: VviGST4, a potential PA transporter and the final enzyme to produce anthocyanidin [54,55,78] and the acyltransferases VviGAT1 and VviGAT2 [79,80], which catalyze the addition of acyl groups. These processes are critical for long-term pigment stabilization. However, high UV-A attenuated this anthocyanin enhancement, despite a partial restoration of flavonol synthesis. This shift in metabolic priority can be attributed to the downregulation of VviDFR (dihydroflavonol 4-reductase) (Figure 4b), a pivotal enzyme at the metabolic branchpoint directing flux toward both anthocyanins and PAs. Studies have shown that UV radiation can stimulate the acylation of anthocyanins in Bovale Grande grapes, especially the formation of coumaroylated glucosides [81]. This indicates that excessive UV-A stress triggers a strategic reallocation of resources away from pigmentation and towards more generalized, broad-spectrum antioxidant defenses like flavonols and PAs.

The surge in soluble PAs within white-type calli under high UV-A stress correlated directly with the upregulation of VviLAR1 and VviANR (Figure 4), which encode key enzymes in the PA biosynthetic pathway [53,58,82]. Similar light-induced effects were also verified under UV-C conditions, and the expression of VviLAR1, VviLAR2 and VviANR was significantly enhanced [43]. Our co-expression network analysis (WGCNA) revealed that this response is not an isolated event but is part of a highly coordinated program. The induction of VviGSTs was observed within the same significant co-expression module as VviLAR1 and VviANR (Figure 7). Indeed, the WGCNA analysis proved instrumental in unveiling the central regulatory network governing this system. Our WGCNA analysis revealed co-expression modules that associate photoreceptor-encoding genes (VviUVR8, VviCRY2), light signaling components (VviCOP1, VviHY5), and structural genes of the flavonoid pathway (Figure 7). In grapevine, VviUVR8 has been demonstrated to physically interact with VviCOP1 under UV-B exposure [41,83,84], and transcriptomic studies have reported that VviCRY2 expression is positively correlated with UV-A-responsive flavonoid accumulation in berry skins [62,67,72,85]. Notably, recent work in ‘Red Globe’ and Cabernet Sauvignon grapes also observed coordinated upregulation of CRY2 with anthocyanin and flavonol biosynthetic genes under light stress, suggesting a potential role in photomorphogenic regulation of phenylpropanoid metabolism [62,63,64]. However, direct evidence for UVR8 sensing UV-A or for CRY2 functioning as a bona fide UV-A photoreceptor in grape remains lacking. Therefore, the regulatory relationships inferred from our co-expression network should be interpreted as hypothetical associations, consistent with orthologous pathways in Arabidopsis but requiring functional validation—such as yeast two-hybrid, electrophoretic mobility shift assays (EMSAs), or transient reporter assays—in grape systems.

Our data are consistent with a model in which UV-A dose and the inherent metabolic state of the callus jointly influence the partitioning of flavonoid flux. Specifically, low-dose UV-A coincided with elevated anthocyanin accumulation (Figure 2a) predominantly in red-type calli—a tissue already primed for pigment synthesis—whereas high-dose UV-A was associated with enhanced production of flavonols (Figure 2b) and proanthocyanidins (Figure 3), particularly in white-type calli. This pattern parallels observations in field-grown berries, where sunlight-dependent cultivars show stronger anthocyanin induction than non-pigmented genotypes under moderate UV exposure [86,87,88]. The differential response may reflect a resource allocation strategy shaped by the baseline metabolic configuration of each callus type. While our callus system provides a controlled platform for dissecting UV-A-mediated metabolic regulation, its flavonoid yields could be further enhanced by integrating UV-A with other elicitation strategies (e.g., phytohormones or nutrient modulation) in future bioprocess optimization.

Our findings resonate with—but also extend—knowledge from UV-B studies in grapevine. UV-B irradiation robustly induces flavonols and anthocyanins via the canonical UVR8–COP1–HY5/HYH signaling module, which transcriptionally activates VviFLS4, VviCHS1, and other phenylpropanoid genes [29,65,66,89]. The strong co-expression of VviUVR8 and VviHY5 with flavonoid genes in our dataset aligns with this established framework [65,66]. Intriguingly, both prolonged UV-B [29] and high-dose UV-A (this study) are associated with attenuated anthocyanin accumulation despite sustained induction of upstream pathway genes, suggesting a shared stress-responsive reprogramming at high irradiance. This convergence may involve HY5/HYH as integrators that prioritize general antioxidant defenses (e.g., flavonols, PAs) over specialized pigments under severe stress—a trade-off also implied in transcriptome-metabolome analyses of UV-exposed berries [87]. The apparent dose-saturation effect for anthocyanin synthesis under UV-A [90] further hints at an optimal irradiance window for pigment induction, beyond which metabolic resources may be redirected toward more efficient ROS scavengers. Additionally, given that UV signaling intersects with the circadian clock in plants [91] and that core clock components (e.g., ZTL2) were differentially expressed in our calli (Figure 5), it is conceivable that temporal gating influences flavonoid output. However, such hypotheses require direct testing in systems with intact circadian rhythms.

The use of callus cultures in this study was instrumental in its success. By employing genetically identical tissues, we eliminated the confounding effects of cultivar genetic variation. Furthermore, by studying undifferentiated cells, we minimized the influence of developmental and systemic variables that are prevalent in whole-plant studies (e.g., berry phenophases and source-sink relationships). This approach allowed us to isolate UV-A-specific effects with unprecedented clarity, offering a powerful platform for uncovering core mechanistic principles. The resource reallocation trade-offs we observed—where high UV-A prioritizes defensive PAs and flavonols in non-primed tissues (Figure 2 and Figure 3)—might be obscured in a whole-plant context by competing metabolic demands. These results establish a foundation for future studies aiming to manipulate flavonoid biosynthesis through light quality in grape, though translation to bioreactor systems remains untested.

While our study provides a robust framework, it also opens several key avenues for future research. A primary goal will be to dissect the nature of UVR8-CRY2 crosstalk, including whether these photoreceptors form heterodimers under specific UV-A doses [67] and how their interaction fine-tunes the activity and stability of HY5 and HYH. Genetic approaches, such as generating single and double photoreceptor mutants in the callus system, could provide definitive insights. Furthermore, the role of VviMYB24 [29,92,93], a hub gene in the turquoise module, warrants detailed investigation. This transcription factor may integrate circadian signals [94] with UV-A stress to gate flavonoid flux, thereby optimizing resource allocation. Its potential as a master regulator makes it a prime target for biotechnological intervention. Finally, metabolic engineering of acyltransferases (VviGATs) presents a tangible application. Although only a single time point (48 h post-UV-A) was analyzed, the strong concordance between induced flavonoid levels (Figure 2 and Figure 3) and upregulated biosynthetic gene expression (Figure 4b) demonstrates that this window effectively captures a key phase of UV-A-driven metabolic reprogramming. Overexpression of these genes could enhance anthocyanin stability under high irradiance, extending the effectiveness of low-dose UV-A treatments in vineyard settings. Addressing these prospects will not only advance fundamental knowledge of plant environmental sensing but also enable innovative agricultural and biotechnological applications. While this study focused on a single cultivar to establish a clear causal relationship, future work will benefit from expanding this approach to diverse grape varieties to assess the broader applicability of UV-A treatment for tailored flavonoid production.

5. Conclusions

In conclusion, this study generates testable hypotheses about light-regulated flavonoid channeling in grape callus, which potentially participates in allocating flavonoid products in response to UV-A in grape cells. The potential mechanism involves a dose-dependent switch in photoreceptor dominance and downstream transcriptional networks that orchestrate a trade-off between pigment production (driven by anthocyanins) and antioxidant defense (driven by flavonols and PAs). The use of an isogenic callus system was crucial for dissecting these complex interactions, and the insights generated provide a foundation for future studies aiming to dissect how light signaling fine-tunes flavonoid pathway branching in grape, with implications for understanding tissue-specific metabolic regulation.