Improving cassava bacterial blight resistance by editing the epigenome

The study addresses whether targeted DNA methylation of a pathogen effector binding element (EBE) can block induction of a host susceptibility (S) gene and thereby reduce disease symptoms without impairing plant growth. In many plant–pathogen systems, bacterial TAL effectors activate S genes by binding specific promoter sequences. Editing S genes to avoid such activation can increase resistance, but direct gene mutations may compromise normal plant development because S genes often have essential roles. Because DNA methylation can inhibit TAL binding and modulate gene expression, the authors hypothesized that directing cytosine methylation to the TAL20 EBE in the cassava MeSWEET10a promoter would prevent its ectopic induction during infection by Xanthomonas phaseoli pv. manihotis (Xam) and reduce cassava bacterial blight (CBB) symptoms. The work aims to test this epigenome editing strategy and evaluate its specificity, effects on gene activation, disease phenotypes, and plant growth.

Background literature indicates that TAL effectors from Xanthomonas and Ralstonia bind predictable EBEs to activate host S genes, a common virulence strategy. Prior studies showed that genome editing (e.g., TALENs/CRISPR) of S genes or their promoters can confer disease resistance in crops. Epigenetic regulation, particularly cytosine DNA methylation established via the plant RNA-directed DNA methylation (RdDM) pathway, can influence gene expression and inhibit TAL binding. Targeted methylation tools have fused CRISPR-based systems or artificial zinc-fingers (ZFs) with methylation machinery such as DRM2 or DMS3, with ZF approaches offering compact transgenes and PAM-independent targeting. For cassava, CBB is widespread and no strong broad-spectrum resistance genes are currently available; TAL20 is a major virulence determinant that induces MeSWEET10a, a SWEET family sugar transporter required for water-soaking and full virulence. The MeSWEET10a promoter is unmethylated in wild type, suggesting that targeted methylation at the TAL20 EBE could block induction and reduce disease.

• In vitro binding assay: Electrophoretic mobility shift assay (EMSA) using purified 6xHis-TAL20 from Xam668 tested binding to unmethylated vs methylated EBE oligonucleotides corresponding to the MeSWEET10a promoter.
• Construct design: A synthetic zinc-finger (ZF) DNA-binding domain targeting the TAL20 EBE sequence (5'-CGCTTCTCGCCCATCCAT-3') was fused to Arabidopsis DMS3, an RdDM factor effective for de novo methylation targeting. A ZF-only construct served as a negative control.
• Transgenic lines: Cassava plants (varieties including 60444 and TME419) were generated expressing DMS3-ZF (independent lines e.g., #133 and #204) or ZF-only control (e.g., #216). Transgene expression was confirmed by western blotting (anti-FLAG) and SDS-PAGE/Coomassie staining.
• Targeted methylation assessment: Amplicon-based bisulfite sequencing (ampBS-seq) of a 233 bp region encompassing the EBE within the MeSWEET10a promoter quantified CG/CHG/CHH methylation. Whole-genome bisulfite sequencing (WGBS) evaluated on-target specificity, flanking (±5 kb) regions, genome-wide methylation levels, genebody methylation profiles, and differentially methylated regions (DMRs) relative to wild type.
• Off-target analysis: Potential ZF off-target sites were predicted genome-wide with Cas-OFFinder allowing up to 4 mismatches (1542 sites). Methylation levels at these sites plus ±50 bp flanks were compared between DMS3-ZF lines and wild type; enrichment was contrasted with random control regions.
• Infection and expression assays: Leaves were infiltrated with Xam wild type or XamΔTAL20 (mock as control). MeSWEET10a expression at 48 h post-infiltration was measured by RT-qPCR normalized to GTPb and PP2A4 reference genes.
• Disease phenotyping: Water-soaked lesion area and intensity (grayscale-based quantification) were measured 4 days post-infiltration and compared across genotypes and treatments. Bacterial population assays assessed in planta bacterial load. Plant growth phenotypes were evaluated in greenhouse and field settings.
• WGBS library prep and analysis details: DNA library construction (Kapa HyperPrep, bisulfite conversion, NovaSeq 6000 sequencing), mapping with BSMAP allowing one mismatch, duplicate removal (SAMtools), conversion rate estimation via chloroplast genome, methylation calling per cytosine, chromosome-level methylation in 100 kb bins, and DMR calling (p<0.01; ΔCG≥0.4, ΔCHG≥0.2, ΔCHH≥0.1).

• EMSA showed TAL20 binds the unmethylated EBE oligo, but binding to the methylated oligo was inefficient; methylated competitor oligo did not outcompete binding even in excess, confirming methylation blocks TAL20-EBE interaction in vitro.
• In vivo targeting with DMS3-ZF induced high levels of ectopic CG methylation at and near the ZF target site within the MeSWEET10a promoter, including the only two CGs within the EBE; ZF-only controls showed no methylation at the EBE.
• WGBS confirmed strong de novo methylation within the target region, with methylation in ±5 kb flanks comparable to controls, indicating local specificity. Genome-wide CG/CHG/CHH methylation levels and genebody methylation profiles were similar between DMS3-ZF lines and controls.
• Off-target evaluation: 1542 predicted ZF off-target sites showed no significant methylation differences between DMS3-ZF and wild type. DMR analysis detected two hypermethylated DMRs at the targeted MeSWEET10a region; most other DMRs were hypomethylated, consistent with tissue culture-associated global methylation decreases.
• Functional impact: Upon Xam infection, DMS3-ZF plants displayed a significant lack of MeSWEET10a induction compared to controls, across independent lines and experiments (suppression did not correlate with DMS3-ZF expression level).
• Disease symptoms: DMS3-ZF plants had reduced water-soaked lesion area; lesion areas in DMS3-ZF plants infected with Xam WT were comparable to infections with XamΔTAL20. Water-soaking intensity (relative grayscale) was also reduced in DMS3-ZF plants.
• Plant performance and bacterial growth: DMS3-ZF plants maintained normal growth and development in greenhouse and field assessments; no significant differences in bacterial population sizes were observed compared with controls.

Targeted cytosine methylation of the TAL20 EBE in the cassava MeSWEET10a promoter effectively prevents TAL20 binding and in vivo transcriptional activation of this S gene during infection. This directly addresses the hypothesis that epigenetic modification of an EBE can block pathogen effector-driven host susceptibility. The approach reduces hallmark CBB symptoms (water-soaking lesion size and intensity) to levels similar to infections lacking TAL20, while preserving normal plant growth, indicating that essential physiological roles of MeSWEET10a under non-infected conditions remain uncompromised. Specificity analyses (WGBS, off-target predictions) show methylation changes confined to the intended locus, supporting the safety and precision of ZF-DMS3-mediated epigenome editing in cassava. Collectively, the results demonstrate epigenome editing as a viable, potentially generalizable strategy to enhance disease resistance by modulating accessibility of S gene promoters to pathogen effectors.

The study establishes an epigenome editing strategy for crop improvement by directing DNA methylation to a pathogen effector binding site to block S gene activation. A ZF-DMS3 fusion efficiently and specifically methylated the TAL20 EBE in the cassava MeSWEET10a promoter, preventing its induction during Xam infection and reducing CBB symptoms without detectable growth penalties. This provides a proof-of-concept for engineering resistance via targeted promoter methylation. Future work could assess heritability and stability of targeted methylation across generations, evaluate efficacy across diverse Xam strains and additional EBEs/S genes, optimize delivery systems and regulatory acceptance (e.g., transgene-free approaches), and quantify agronomic performance and durability under field conditions.

• The resistance mechanism targets TAL20-mediated activation and may be strain-specific; efficacy against Xam isolates that use different effectors/targets was not evaluated.
• Heritability and long-term stability of targeted methylation across vegetative cycles or sexual generations were not reported.
• While specificity appeared high, the study noted global hypomethylation likely due to tissue culture; broader epigenomic consequences of transformation/regeneration could confound interpretation.
• Bacterial population sizes were not significantly reduced despite symptom attenuation, indicating partial resistance that mitigates symptoms rather than restricting pathogen growth.
• Field performance, yield impacts, and multi-season durability were not comprehensively tested.