Consequences of adaptation of TAL effectors on host susceptibility to Xanthomonas

Transcription activator-like effectors (TALEs) are virulence factors of Xanthomonas that induce the expression of host susceptibility (S) genes by specifically binding to effector binding elements (EBEs) in their promoter regions. The DNA binding specificity of TALEs is dictated by their tandem repeat regions, which are highly variable between different TALEs. Mutation of the EBEs of S genes is being utilized as a key strategy to generate resistant crops against TALE-dependent pathogens. However, TALE adaptations through rearrangement of their repeat regions is a potential obstacle for successful implementation of this strategy. We investigated the consequences of TALE adaptations in the citrus pathogen Xanthomonas citri subsp. citri (Xcc), in which PthA4 is the TALE required for pathogenicity, whereas CsLOB1 is the corresponding susceptibility gene, on host resistance. Seven TALEs, containing two-to-nine mismatching-repeats to the EBEPthA4 that were unable to induce CsLOB1 expression, were introduced into Xcc pthA4:Tn5 and adaptation was simulated by repeated inoculations into and isolations from sweet orange for a duration of 30 cycles. While initially all strains failed to promote disease, symptoms started to appear between 9–28 passages in four TALEs, which originally harbored two-to-five mismatches. Sequence analysis of adapted TALEs identified deletions and mutations within the TALE repeat regions which enhanced putative affinity to the CsLOB1 promoter. Sequence analyses suggest that TALEs adaptations result from recombinations between repeats of the TALEs. Reintroduction of these adapted TALEs into Xcc pthA4:Tn5 restored the ability to induce the expression of CsLOB1, promote disease symptoms and colonize host plants. TALEs harboring seven-to-nine mismatches were unable to adapt to overcome the incompatible interaction. Our study experimentally documented TALE adaptations to incompatible EBE and provided strategic guidance for generation of disease resistant crops against TALE-dependent pathogens.

In this well-written article entitled « Consequences of adaptation of TAL effectors on host resistance against Xanthomonas » Teper and Wang investigate the potential and the mechanisms of and through which Xanthomonas TAL effectors repeats evolve to adapt to incompatible DNA target sites within promoters of host susceptibility genes. Here, based on the analysis of PthA4 which is a major virulence TAL effector in Xanthomonas citri (the causal agent of Asian Citrus canker), the authors evaluate the capacity of PthA4 variants harboring from 2 to 9 mismatching repeats preventing for the induction of the PthA4 matching S gene CsLOB1 to evolve new repeat arrays allowing for virulence recovery. The relevance of the project is particularly high knowing that deployment of resistance strategies based on TALEmediated loss-of-susceptibility has mobilized major interest over the last decade in several economically important crops threatened by Xanthomonas species relying on TALE effectors.
Overall I like the results and appreciate the hard work which was elegantly designed and achieved but I am worried about the "too artificial" conditions of the experimental evolution. Because such approach is artificial in essence, a detailed section in the discussion taking a step back on the limitations of the strategy would be more than necessary. I am personally not convinced that TAL repeat changes observed in this assay would occur in a field.

Major comments
As illustrated in Figure 2B, the authors repeatedly alternate phases where the bacteria are grown on synthetic medium versus phases where the bacteria are in the plant leaves. This means that bacterial populations are also evolving on NA medium. How does this influence the overall strategy ? Can one exclude that some adaptive mutations occurring in planta were counter-selected when bacteria bearing them where grown on plates ? Or that mutations have appeared during multiplication on plates and that they were next selected in planta, thus resulting of an adaptation to an artificial medium and not the host tissues ? I see this as an important conceptual gap. I recognize that the experiment is carefully conducted and that results are very interesting and worth being published BUT they should be interpreted with more caution, taking more this strong bias in consideration. In the same line of idea, the assay is based on over-expressing a dTALE from a high copy plasmid which probably favor intra versus inter TAL repeats recombination. I am actually surprised that no modification veseem to has taken place at the pthA4 locus leading for instance to a loss of Tn5 : has this been verified ? Having included as negative control Xcc pthA4:Tn5 bearing an empty vector in the experimental evolution assay would have helped to exclude the potential for other sorts of extra-dTALE modifications to be occurring).
The authors should at least show that they are aware of these caveats and be discussing them. They also should rationalize their choice : why did they have to multiply bacteria on plates between each passage on plant ? Was it on purpose ? For technical aspetcs ?

Minor comments
• Specify in the title that this article is dealing with loss-of-susceptibility mediated resistance.  Figure, a phylogenetic analysis among the 13 allelic variants would be useful and give more weight to this interesting analysis depicting the relative affinity of 13 unique array variants to the pLOB variants generated from a diverse set of Rutaceae plants. A bioinformatic tool called DisTAL could be useful for this as it evaluates phylogenetic distances between repeat arrays: https://doi.org/10.3389/fpls.2015.00545. Fig1D-E tend to experimentally validate the predictions reported in Fig1C. Fig1D is useless because signals are quite weak overall (even for PthA4 on sweet orange), thus requiring a quantitative analysis which is shown in Fig 1E,