PREreview of Secreted TAL effectors protect symbiotic bacteria from entrapment within fungal hyphae

Morgan Carter and Adam Bogdanove, Section of Plant Pathology and Plant-Microbe Biology, Cornell University, Ithaca, NY

       Richter and colleagues investigate the role of TAL effector-like proteins in a bacterial-fungal symbiosis between Burkholderia (Mycetohabitans) rhizoxinica and Rhizopus microsporus. TAL effectors are proteins found in some plant pathogenic Xanthomonas and Ralstonia species that are injected into plant cells by type III secretion (T3S) and act as eukaryotic transcription factors to transcriptionally activate host genes that contribute to disease. Genes homologous to TAL effectors were identified in a B. rhizoxinica genome a few years ago. They were named bat (Burkholderia TAL) genes (or in one case “burrH”) and investigated for their DNA-binding ability and use in biotechnology, but not for their function in vivo; it was unknown whether the encoded proteins are secreted or if they impact the host. Having posted our own preprint on the native role of these TAL-like proteins (Carter, et al 2020 bioRxiv), which we refer to as Btl (Burkholderia TAL-like) proteins, we were interested to see how our study and this one would inform each other, and what broader conclusions could be drawn in light of both investigations. We were enthused by the use of microfluidic chambers to observe and quantify the morphological dynamics of the interaction as it is established following inoculation of the fungus with the bacterium. This is an elegant experimental platform that overcomes a tricky challenge for this system. We were also excited to see that bat/btl genes are present across all strains the authors sequenced, but were diverse. This result aligns with the Southern blot presented in our paper. However, we did have some concerns with the study. Here, we offer our major comments:

1.Title: The authors should consider removing the word “secreted” from the title of the paper, as they did not investigate this claim, and BAT/Btl proteins lack the N terminal region of TAL effectors that contains their T3S signal.

2. Lack of complementation of the bat mutants: The authors generated and phenotyped individual knockout strains for each of the three bat genes from the type strain of Burkholderia rhizoxinica, HKI 0454 (also called B1). However, because they do not show complementation of any mutant by the corresponding wildtype gene on a plasmid, whether the phenotypes are due to the targeted mutations is uncertain, and the functional claims therefore unfounded.

a. Complementation, if observed, would rule out the possibility that the phenotypes are due to spontaneous mutation that occurred during the long period of culturing under selection and counterselection outside the fungus, or from an off-target recombination concurrent with integration of the suicide plasmid.

3. Claim that mutants are unable to induce host sporulation. The authors claim in the abstract and in the discussion that bat mutants were unable to trigger host sporulation, which contradicts the data in Figure 2 where sporulation still occurred after coculturing; sporulation was not assessed once the symbiosis was established.

4. Mutant phenotypes that are difficult to reconcile with previous findings and with each other. When allowed to re-associate with the fungal host, all 3 mutants were similarly reduced in their ability to trigger sporulation (Figure 2). This result would suggest (if borne out by complementation) that each BAT is required for full induction of sporulation, but none is sufficient, and that they therefore are acting cooperatively. The proteins indeed vary in their predicted DNA binding specificity and were shown previously to differ in their ability to bind DNA; in particular, BAT3, which contains only 6 repeats in the DNA recognition domain, does not bind DNA (de Lange et al., 2014).

a.There is no known example of TAL effectors functioning cooperatively on a single target or within a pathway, as is suggested by each mutant having the same sporulation phenotype. No mechanism was proposed in the manuscript.

b. If BAT proteins function like TAL effectors, one would not expect the bat3 mutant to have any phenotype, since the protein does not bind DNA. We note that the bat3 deletion might have a polar effect on expression of the bat2 gene, which is immediately downstream, that could explain the presence of a phenotype. However, see comment 4c.

c. It is puzzling that although all three Δbat strains had the same reduced sporulation phenotype, Δbat1 and Δbat3 had the same dense accumulation phenotype, while Δbat2 showed the opposite (Figure 4). If one assumes based on the inability of BAT3 to bind DNA that the Δbat3 sporulation phenotype is due to a polar effect on bat2, how could the accumulation phenotypes of the Δbat3 and Δbat2 mutants differ? Is there perhaps partially impaired expression of bat2 in the Δbat3 strain? These puzzling differences underscore the need to carry out genetic complementation tests.

5. Clustering of BAT proteins and interpretation of BAT protein diversity across strains: In Figure 2e, the authors assigned each BAT protein to a category based on the BAT proteins of a single strain using conventional, multiple sequence alignment-based clustering. It is unclear if the intent is to deduce a phylogeny or to show putative functional relationships. Furthermore, it is not clear from the alignments in the supplementary material that equivalent full-length sequences were used for each protein.

a. The approach taken is an imperfect solution to grouping TAL effectors (or BATs) because proteins of the same length or that appear closely related can vary greatly in their DNA binding specificity due to differences across the repeat variable diresidues (RVDs, which collectively determine DNA binding specificity), thus varying in their role and function. For establishing possible functional relationships, RVD sequence alignments have been used (Grau et al., 2016). For phylogenetic relationships, often 5’ and 3’ nucleotide sequences, without the central repeat region, are used.

b. Next, the diversity of BAT proteins across strains was not explored by the authors within the context of the mutant phenotypes. If all 3 BAT proteins from B1 are necessary for successful infection, why do the number of bat genes and RVD sequences differ across Burkholderia strains, with some strains only having an ortholog of one BAT?

c. Finally, the inclusion of B13 and B14 in the analysis is redundant, given that these are the same isolates as B4 and B7, respectively.

6. BAT nomenclature: The clustering together of BAT proteins under groups designated BAT1, 2, and 3 oversimplifies the diversity among these proteins and creates groupings that are incongruous with their anticipated biological functions based on variation in their RVD sequences. We strongly urge studies on these proteins to adopt a more information-rich nomenclature that uniquely identifies each gene to prevent confusion.

a. As with the difficulties of phylogenetic analysis, homology does not necessarily indicate function as TAL effectors with the same number of repeats can vary in RVD sequence and DNA binding specificity, thus target and function.

b. In our study, we employ a naming scheme that conveys both the repeat number and strain number to give each BAT/Btl protein a unique identifier, i.e. Btl19-13 to designate the 19 repeat-long Btl protein of strain B13.

7. Other: The authors should clarify what is meant by a biological and a technical replicate in the microscopy experiments, as this was unclear in this context and impacts interpretation of statistical significance (there were “3 biological replicates” and “16 technical replicates”; one might expect less variation among technical replicates).

References:

de Lange, O., Wolf, C., Dietze, J., Elsaesser, J., Morbitzer, R., and Lahaye, T. (2014). Programmable DNA-binding proteins from Burkholderia provide a fresh perspective on the TALE-like repeat domain. Nucleic Acids Res 42, 7436-7449.

Grau, J., Reschke, M., Erkes, A., Streubel, J., Morgan, R.D., Wilson, G.G., Koebnik, R., and Boch, J. (2016). AnnoTALE: bioinformatics tools for identification, annotation, and nomenclature of TALEs from Xanthomonas genomic sequences. Scientific Reports 6, 21077.