In this work, we used RNA-Seq data to investigate the transcriptomic response of resistant and susceptible common bean genotypes during their interaction with X. phaseoli pv. phaseoli. Importantly, our work provided novel whole genome sequence data of two parental lines of a reference population (BAT93 x JaloEEP558) used in many genetic and genomic studies, including mapping of disease resistances [65–73].
A global trend was that the proportion of DEGs was higher in the susceptible genotype (31%) than in the resistant genotype (20%). This result highlights that susceptibility results in a larger reprogramming of gene expression than resistance, which was not surprising since a similar trend was observed in many other studies describing plant transcriptomic responses to diverse plant pathogens including fungi, bacteria and viruses [72–78]
In the core transcriptome, the induction of genes from the CIPK and CDPK families suggests that both genotypes are able to perceive X. phaseoli pv. phaseoli through Ca2+ signaling [79–82]. Following this hypothesis, the susceptibility to CBB observed in JaloEEP558 would result from inhibition of PTI by bacterial effectors rather than from non-detection by PRRs. In accordance with this, the susceptible genotype displayed a large repression of genes putatively involved in PTI such as RLKs, as well as defense response and NLR genes. This indicates that successful bean colonization by X. phaseoli pv. phaseoli is linked to suppression of plant defenses and reflects a potential involvement of bacterial effectors.
Pathogenicity assays showed that X. phaseoli pv. phaseoli switched from a biotrophic to a necrotrophic stage 15 DPI. The defense response to biotrophic pathogens is usually regulated by SA while defense responses to necrotrophic pathogens classically involve JA and ethylene [13]. Here, a global induction of the SA pathway was linked to resistance, suggesting that an adapted SA response against biotrophic pathogens is effective in BAT93 48 h after infection by X. phaseoli pv. phaseoli. On the other hand, susceptibility was linked to the induction of the JA/ethylene pathway and repression of the SA pathway. In this view, it would be interesting to further study the transcriptomic response of common bean at different stages of X. phaseoli pv. phaseoli colonization, including necrotic tissues.
The specific suppression of photosynthesis, sugar metabolism and other chloroplast-associated genes observed in BAT93 reflects a rather classical defense response. Indeed, similar trends have been described using transcriptomics in different interactions between plants and bacteria [83–85], and integration of transcriptomic data from different pathosystems led to the hypothesis that suppression of photosynthesis is part of the plant adaptive immunity [86]. In addition, several studies have shown that incompatible interaction is linked to a decrease in photosynthetic activity [87–90] and that inhibition of photosynthesis often leads to the accumulation of reactive oxygen species [91]. Supporting this, it is interesting to note that Xanthomonas citri pv. citri is able to counter the decrease in photosynthesis by mimicking a plant natriuretic peptide, leading to suppression of resistance in citrus leaves [92].
The plant cell wall plays an important role in plant immunity, both as a physical barrier against pathogens and by releasing signaling compounds known as DAMPs when altered [93]. Successful pathogens are usually able to degrade the plant cell wall to access nutrients. In JaloEEP558, specific induction of genes involved in cell wall modification, such as xyloglucan endotransglycosylases and glycosyl hydrolases, suggests that remodeling of cell wall occurred in the susceptible genotype [94, 95]. In accordance with this, early CBB phenotype usually corresponds to water-soaking symptoms involving the softening and loosening of the cell wall. In contrast, cell wall modification genes were repressed in BAT93, which could suggest that cell wall rigidification occurred in the resistant genotype, thus preventing bacterial progression. Modification of the cell wall is tightly linked to cell size and shape and to morphogenesis of plant organs [96]. Interestingly, transcription factors from the LOB and ovate families were found both among the most differentially expressed genes between BAT93 and JaloEEP558, and among the DEGs located in QRLs to CBB.
LOB is a family of plant-specific transcription factors with key roles in the regulation of plant organ development [97, 98]. Induction of CsLOB1 by Xanthomonas in Citrus sinensis triggers cell expansion and is required for symptom development [99]. Induction of LOB genes can induce genes involved in cell wall modification such as pectin methylesterase inhibitors [100]. Here, the strong induction of LOB genes in JaloEEP558 could contribute to the induction of downstream genes involved in cell wall modification and to the development of symptoms. The LOB gene Phvul.005G049700 is located on chromosome 5 and colocates with a QRL linked to marker D1081, explaining 15% of the phenotypic vartiation [49]. Therefore, Phvul.005G049700 could putatively act as a negative regulator of resistance to CBB.
On the other hand, the ovate family member Phvul.009G057100 was strongly induced in the resistant genotype and repressed in the susceptible genotype. This gene could positively contribute to CBB resistance as it colocates with a QRL on chromosome 9 that is linked to marker D0157, and explains 13% of the phenotypic variation [49]. Consistent with this hypothesis, ovate family members act as transcriptional repressors involved in the suppression of cell growth and elongation as well as in the regulation of secondary cell wall and vascular development [62–64]. However, to our knowledge, the ovate family has not so far been described as playing any role in plant-pathogen interaction.
Altogether, our analyses pointed out different molecular pathways appearing important for either promoting disease in the susceptible genotype or triggering immunity in the resistant genotype. The genes involved in these pathways were scattered throughout the whole common bean genome, which reflects the complexity of CBB resistance. In particular, large clusters of dozens of NLR genes exist at common bean subtelomeres [101–104] but the 34 NLR repressed in JaloEEP558 following X. phaseoli pv. phaseoli infection did not correspond to the specific repression of one of these clusters. Thus, no particular locus was unveiled as being responsible for CBB resistance. To summarize, resistance was linked to suppression of photosynthesis and sugar metabolism and induction of the SA pathway, while susceptibility was linked to downregulation of plant defenses and upregulation of the JA/ethylene pathways and AP2/ERF transcription factors as well as a large induction of genes involved in cell wall modification.
Xanthomonas bacteria possess transcription activator-like effectors (TALE) that are type III effectors able to induce the transcription of genes by specifically binding to the promoter of plant susceptibility genes and recruiting the transcription machinery [105]. Nine different TALE-encoding genes and alleles have been described in CBB agents [106]. Strain CFBP6546R used in this study bears tal19I and tal18H and in silico prediction of susceptibility genes has been done [106, 107]. However, no predicted targets appeared as induced in our transcriptomic data for these two TALEs (not shown).
Interestingly, most of the pathways induced in the susceptible genotype or repressed in the resistant genotype were previously described as being induced by TALEs to promote disease [108]. For example, the best-characterized TALE targets so far are SWEET genes that encode sugar exporters presumably providing nutrients for the pathogen [109–112]. SWEET gene induction by TALE has been described in the interaction of Xanthomonas with rice [112–118] cassava [119] and cotton [120]. According to our results, it is tempting to speculate that the suppression of photosynthesis and sugar metabolism observed in the resistant genotype could lead to the reduction of sugar production by the plant cells, thus contributing to resistance by depriving bacteria of sugar. Interestingly, one SWEET gene (Phvul.002G203600) was induced specifically in the susceptible genotype, but it seems that this induction is not due to the action of TALEs.
Other TALE targets include different types of TFs from the AP2/ERF [121], bHLH [122, 123] or LOB [99, 124, 125] families, which is reminiscent of what was observed here in the susceptible genotype. Interestingly, these targets are often linked to cell wall reorganization and modification of the plant cells shape. In pepper, the TALE AvrBS3 from Xanthomonas campestris pv. vesicatoria induces the bHLH transcription factor UPA20, which leads to the hypertrophy of leaf cells [122]. In tomato, AvrHah1 from X. gardneri targets another bHLH TF whose induction upregulates the expression of a pectate lyase responsible for the apparition of water soaking symptoms [123]. In citrus, the induction of CsLOB1 by different TALEs from X. citri pv. citri or X. citri pv. aurantifolii is required for the apparition of canker symptoms due to hyperplasia and rupture of the epidermis in infected tissue [99, 124, 125]. Here, the JaloEEP558-specific induction of LOB homologs and other genes involved in cell wall degradation and modification suggests that X. phaseoli pv. phaseoli employs mechanisms similar to what was observed in tomato and citrus. The parallel observed between our transcriptomic data and TALE targets in different plant species suggests that TALE evolution was driven by the necessity to target pre-existing susceptibility hubs in plants. This also raises the question as to whether AP2/ERF, bHLH, pectate lyases, or LOB homologs could constitute direct targets for TALEs from X. phaseoli pv. phaseoli in common bean. Thus, searching for X. phaseoli pv. phaseoli TALE targets in common bean would bring valuable knowledge on the molecular interactions between common bean and X. phaseoli pv. phaseoli.