Changes in hyperspectral characteristics during sooty mold disease infection
To illustrate the differences between different canopy layers, an average of the original spectrum was created (Fig. 1). The reflectance trends of the three canopy layers increased in the 397–552 nm wavelength range and declined in the 552–673 nm wavelength range. There was a noticeable variation in the reflectance of reaching the inflection point with canopy layers. The reflectance of all three canopy layers increased exponentially in the wavelength range of 673–800 nm. Based on the spectral characteristics, the average spectral reflectance of the three canopy layers in the designated wavelength range displayed a similar overall pattern. In the spectral characteristics, three inflection spots were located close to the 552, 673, and 800 nm wavelengths.
Transcriptomic analysis and screening of DEGs
Nine cDNA libraries were generated from the leaf samples at three canopy layers: A1, A2, and A3. Transcriptomic sequences were obtained after the removal of adapters and low-quality reads. An average of 44.8 million clean reads per sample were received, with a Q20 base composition percentage exceeding 97.56% and a Q30 base composition percentage exceeding 93.02%. The GC content measured greater than 43.56% (Table S1). A total of 2404 genes showed significantly differential expression under the log2 (fold change) ≥ |2| and p-value < 0.5 (Table S2). The number of differentially expressed genes (DEGs) ranged from 322 (231 upregulated, 91 downregulated) between A3 and A2 to 1994 (1265 upregulated, 729 downregulated) between A3 and Al (Fig. 2A). The number of upregulated DEGs was higher than the downregulated DEGs in A2 and A3 samples, showing that many DEGs were highly expressed in the leaves in down canopy layers. The overlap in gene expression revealed that numerous DEGs were found to be unique to each of the combinations. The most unique DEGs were identified at A3 vs A1 (1254), and the fewest were found at A3 vs A2 (78) (Fig. 2B).
To further explore the functions of DEGs in response to sooty mold disease, Gene Ontology (GO) annotation was conducted. The most significantly enriched terms were oxidoreductase activity, secondary metabolite process, phenylpropanoid catabolic process, lipid biosynthetic process, and cell wall organization or biogenesis (Fig. 2C). An analysis of the enriched pathways in the Kyoto Encyclopedia of Genes and Genomes (KEGG) showed that the most enriched pathways were metabolic pathways, biosynthesis of secondary metabolites, and plant-pathogen interaction (Fig. 2D).
Different expression patterns of DEGs
Under pathogen infection, DEGs exhibit diverse expression patterns in tea plants. A heatmap with GO enrichment was employed to visualize the expression patterns of DEGs in the nine samples. As shown in Fig. 3, many genes are differentially expressed in A1 samples. These DEGs were mainly enriched in carbohydrate binding, oxidoreductase activity, isoprenoid metabolic and biosynthetic processes, and terpenoid metabolic processes. In Cluster II, the DEGs were upregulated in A2 and A3 samples. The genes involved in the homeostatic process, secondary active transmembrane transporter activity, UDP-glycosyltransferase activity, and response to oxidative activity. In A3 samples, numerous genes associated with the secondary metabolic process, phenylpropanoid catabolic process, cell wall biogenesis, and flavonoid metabolic process exhibited induced expression in Cluster III. These findings show an elevated immune response in the A2 and A3 samples. In addition, UDP-glucosyltransferase and oxidoreductase activities are observed in all samples, suggesting their essential roles in the tea plant response to sooty mold disease.
Metabolome profiling of tea leaves infected by sooty mold disease
A multivariate analysis method, PLS-DA, was performed to investigate metabolite changes under sooty mold disease infection. Score plots showed that the differentially expressed metabolites in A1, A2, and A3 were statistically separated (Fig. 4A). In total, 733 metabolites were detected (Table S3). The identified compounds include flavonoids (18.76%), phenolic acids (12.48%), amino acids and derivatives (6.34%), organic acids (5.01%), alkaloids ( 5.01% ), and others (Fig. 4B).
A heatmap was used to visualize the differentially expressed metabolites between different canopy layers infected by SM (Fig. 4C). A1 mainly drives the accumulation of alcohol, lipids (free fatty acids), hydrocarbons, and amino acids and derivatives. Much differently, flavonoids were predominantly upregulated in A2 and A3, particularly in A3. To elucidate the critical metabolites in tea plants that resist SM, scatter plots were used to display the differential metabolites between A2 and A3 as compared to A1 (Fig. 4D). The alkaloids, amino acids and derivatives, and flavonoids were significantly expressed in A2 samples compared with A1. Much differently, alcohol, aldehyde, ester, flavonoids, heterocyclic compounds, hydrocarbons, ketone, lignans, and coumarins were significantly enriched in A3 samples. The results suggest that A3 employs more metabolites to protect against SM than A2.
Correlations between the DEGs and metabolites
The differentially expressed genes and metabolites were used to construct a network of six and three distinct modules with similar expression trends among different samples, respectively (Figs. S1 and S2). The differentially expressed metabolites and genes upregulated in A3 samples, which showed more defense against SM, were selected to construct the network.
WGCNA analysis demonstrates that five hub genes (Dormancy-associated protein, Serine/threonine-protein phosphatase, ABC transporter, Wound-induced protein, and uncharacterized proteins) and two hub metabolites (D-Mannitol and 17-Hydroxylinolenic Acid) have significant relationships with the DEGs and metabolites (Table S4). The hub genes and metabolites were developed into a subnetwork comprising three groups linked to the biosynthesis of secondary metabolites, plant hormone signal transduction, and plant-pathogen interaction (Fig. 5, Table S5). Out of the 98 genes, 40 genes in the biosynthesis of secondary metabolites category included cytochrome P450, gibberellin 20-oxidase, alcohol dehydrogenase, and mannitol dehydrogenase. The high expression of mannitol dehydrogenase was consistent with the enrichment of D-mannitol. The plant-pathogen interaction category includes numerous WRKY, LRR, and PR members. In addition, the genes related to auxin and ethylene show high enrichment in the plant hormone signal introduction category. These results suggested that the tea plant mainly employed genes and metabolites related to the biosynthesis of secondary metabolites, plant hormone signal transduction, and plant-pathogen interaction to defend against SM.
qRT-PCR verification of DEGs
To assess the authenticity and reliability of the transcriptome data and the extent of DEGs. Nine DEGs across the network were selected for validation via qRT-PCR. Of these, Dormancy-associated protein 1 (CSS0035295), uncharacterized protein LOC114300808 (CSS0008624), wound-induced protein (CSS0009105), Serine/threonine-protein phosphatase (CSS0030286), and ABC transporter (CSS0044987) were hub genes. The mannitol dehydrogenase (CSS0031337) and cytochrome P450 (CSS0017190) were involved in the biosynthesis of secondary metabolites. The LRR receptor (CSS0020320) and pathogenesis-related protein (CSS0037023) were related to plant-pathogen interaction. As shown in Fig. 6, qRT–PCR analyses showed the same expression trend for each of the analyzed candidates. These results suggest that the transcriptomic data were accurate and could be used for further functional analysis.