Exploring the Genetic Networks of HLB Tolerance in Citrus: Insights Across Species and Tissues

Abstract

Huanglongbing (HLB), caused mainly by Candidatus Liberibacter asiaticus (CLas), is a devastating disease threatening citrus production worldwide, leading to leaf mottling, fruit deformation, and significant yield losses. This study generated a comprehensive co-expression network analysis using RNA-seq data from 17 public datasets. Weighted gene co-expression network analysis (WGCNA) was applied to identify gene modules associated with citrus species, tissue types, and days post-infection (DPIs). These modules revealed significant enrichment in biological pathways related to stress responses, metabolic reprograming, ribosomal protein synthesis, chloroplast and plastid function, cellular architecture, and intracellular transport. The results offer a molecular framework for understanding HLB pathogenesis and host response. By elucidating module-specific functions and their correlation with species- and tissue-specific responses, this study provides a robust foundation for identifying key genetic targets. These insights facilitate breeding programs focused on developing HLB-tolerant citrus cultivars, contributing to the long-term sustainability and resilience of global citrus production.

Keywords: HLB; RNA-seq analysis; citrus varieties tolerance; gene co-expression network.

Figure 1

Comparison of alignment efficiency using HISAT2 and STAR aligners with C. clementina v1 and C. sinensis DHSO v3 reference genomes. Boxplots show the percentage of aligned reads across different Bioprojects. The red line marks the 60% alignment threshold, below which samples were excluded from downstream analyses. HISAT2 and STAR aligners were tested, and samples with less than 60% alignment (highlighted by the red line) were discarded to ensure data quality.

Figure 2

PCA of 230 RNA-seq samples after ComBat-seq batch correction. (A) PC1 vs. PC2 shows tissue-specific clustering. (B) PC2 vs. PC3 further separates tissues, confirming effective batch correction.

Figure 3

Selection of soft-thresholding power and module merging in the co-expression network analysis. (A) Scale-free topology fit index (R²) across soft-thresholding powers. The red numbers indicate the powers tested; the red line marks the R² = 0.9 reference threshold. (B) Mean connectivity for each power. Together, these plots guided the selection of power 7 as a balance between scale-free topology and network connectivity. (C) Module merging process, reducing module count to 47 based on eigengene correlation.

Figure 4

Module–trait relationships: Heatmap showing the correlation between module eigengenes and biological traits of interest, including Citrus species, Tissue, Treatment, and DPI. Each cell contains the correlation coefficient (r) and associated p-value, with color intensity reflecting the strength and direction of the correlation (red for positive and blue for negative correlations).

Figure 5

Functional enrichment analysis of key genes in the turquoise module using ClueGO. The analysis revealed significant enrichment in pathways related to metabolic reprograming, signal transduction, and active transport, suggesting their involvement in adaptive or response mechanisms. These functional categories may underline species-specific responses to HLB in citrus. Genes in red are associated with these enriched terms.

Figure 6

ClueGO enrichment network for the plum1 module showing functional groups related to zinc ion binding, dynein complex, and transcription. Genes in red are associated with these enriched terms.

Figure 7

Functional enrichment of the floralwhite module linked to DPI in HLB-infected Citrus spp., highlighting its involvement in transport, oxoacid lyase activity, organization of protein-containing complexes, vesicle docking, and the auxin-activated signaling pathway.

Figure 8

Upset diagram of differentially expressed genes: (A) upregulated genes; (B) downregulated genes. Consistent gene expression patterns are shown across different traits (DPI, Citrus species, and Tissue), revealing possible common pathways. Comparisons are labeled as follows: p203307_fr_im, p203307_fr_m, and p203307_lf_m represent orange fruit and leaf samples mature and immature; p417324-56-dpi, p417324-126-dpi, etc., indicate infected orange leaves at different days post-infection; p557834_T_vs_W, p557834_TL_vs_W, and p557834_TL_vs_T compare thornless, wild-type, and thorny grapefruit plants; p629966-0-dpi, p629966-3-dpi, and p629966-9-dpi represent orange feeder roots post-infection; p739186-fall, p739186-spring, etc., denote symptomatic versus asymptomatic mandarin leaf samples across seasons; and p755969-lim-vs-or compares finger lime and orange leaves. Full details are provided in Supplementary Table S1.