PREreview of A bacterial type III effector hijacks plant ubiquitin proteases to evade degradation

Brief description:

In this article the authors describe how a Ralstonia solanacearum effector (RipE1) evades degradation by associating with plant ubiquitin proteases, promoting its de-ubiquitination, and enhancing its stability inside plant cells. The authors also provide evidence to support that phosphorylation of RipE1 at key residues is also important to its stability, and that de-ubiquitination and phosphorylation act independently to stabilise RipE1 in planta.

Introduction:

We felt that the rationale of the study was not well-explained in the introduction and at various points throughout the main text, which impacted the storytelling of the paper. For example, it is unclear why the authors chose to study RipE1, over the other 70 Ralstonia effectors. Is it the most essential effector in establishing infection? What is its function? Why did the authors decide to study its stability? Are there unstable mutants that compromise infection? 

The most interesting fact regarding RipE1 that arises from the introduction is that it can be recognised by the plant immune system. It would be nice if the authors expanded on how it is recognised, since this information is known and key for the rest of the paper. Moreover, it would be helpful to expand on the plant defenses that exist to destabilise foreign proteins, to explain why effector stability is important in establishing infection, and therefore worth studying; as well as on the significance of Ralstonia solanacearum infections for Solanaceous crops, such as tomato.

Note that some of the information lacking in the introduction, appears later on in the discussion.

Results:

Overall, most of the evidence provided in the results section is convincing to support the authors’ conclusion, namely that RipE1 associates with NbUCHs to evade protein degradation, and undergoes additional phosphorylation to enhance its stability. However, the flow of this section could be improved (see comments below).

• The sections of the results that are discussing phosphorylation vs. ubiquitination of RipE1 seem disjointed. Results headings go from “RipE1 phosphorylation in plant cells contributes to protein stability” to “RipE1 interacts with ubiquitin carboxyl-terminal hydrolases in plant cells”. The results headings and main figure captions contain the most important findings of the paper, however, they should also flow through rationally, with the previous leading on to the next, and so on. This would help improve the paper’s storytelling. 

• It is unclear why some experiments include all three ubiquitin proteases studied in this paper (NbUCH05/12/15) and others focus on only two (NbUCH12/15) or one of them (NbUCH15).

• The authors show that silencing of NbUCHs triggers HR and that SGT1 is required for this response. They further speculate that: (1) the function of UCHs as negative regulators of immunity requires SGT1, and/or (2) compromised integrity of UCHs is perceived by an SGT1-dependent NLR triggering HR. It was not surprising to us that UCH (non-specific/general) silencing triggers HR, as a lot more proteins would be tagged for degradation. Perhaps it is protein degradation signatures or DAMPs (damage-associated molecular patterns) that are being sensed by SGT1, which would be consistent with the finding that SGT1 is not required for PAMP-sensing as you had previously demonstrated. 

• The rationale behind testing whether HR triggered by NbUCH silencing is mediated by SGT1 and Ptr1 could have been better explained. It was not clear to us why the authors hypothesised that SGT1 would be required for this response, or why it would be mediated by Ptr1 in the absence of RipE1, considering that Ptr1 triggers HR after sensing RipE1, in the presence of NbUCHs.

• We were also wondering whether the stability of RipE1 impacts upon effector recognition and/or virulence and whether the authors had any evidence to comment on this. 

Methodology:

It is clear that the authors put in a lot of work to generate the results presented in the paper, however we believe that their methodology could be improved, especially when it comes to data analysis and presentation (see comments below).

• The authors proceed to quantify Western blots throughout the manuscript and draw conclusions based on this quantification. This is not accurate for several reasons:

  1. Non-linearity: the relationship between signal intensity and protein concentration may not be linear.

  2. Variability: since Western blots are multi-step processes, each subject to error, small inconsistencies in band intensity are common, and not necessarily reflective of protein levels. Moreover, certain proteins that might be expressed and accumulating at good levels might be harder to extract because of their subcellular localisation (e.g. membrane-bound proteins), which might be reducing their signal on the blot.

  3. Background noise: black or white patches may arise on a blot due to nonspecific antibody binding, burning of the membrane, or other factors that are not reflective of protein levels.

  4. Subjectivity: quantification of Western blots involves subjective decisions, like choosing the baseline for background subtraction and the region for signal quantification. For example in Figure 1E background signal varies across the anti-Ub blot, whilst the region chosen for quantification is not clear, especially when the signal is spanning a large area.

  5. Dynamic range: the range of signal detection might not cover the entire range of protein expression levels, leading to saturation of strong signals and poor detection of weak signals.  

  Western blots are great and essential tools to answer biochemical questions, however, for the reasons stated above they should be interpreted qualitatively.

• Some quantitative figures are missing statistical analyses.

• It is not always clear how many replicates the authors included for each HR assay. This should be stated in all sub-Figure legends (it is currently present in some, absent from others). Better still, the authors could use a qualitative visual scale to quantify HR observed in different replicates, or include reps in the supplement, to give an idea of whether HR is variable or consistent across replicates. Also it is not clear how many plants and leaves were infiltrated for each biological replicate. This should be stated in the methods. 

• We appreciated that the authors are going beyond HR assays, in certain cases, to include ion leakage assays that helps quantify the HR response. However, this was not done consistently throughout. The authors could consider adding a visual scale for HR quantification (as recommended above).  

• We appreciated that the authors used three different methods (co-IP, FRET and fluorescence-tagged protein localisation) to demonstrate the interaction between RipE1 and NbUCH15. However, to demonstrate protein interaction through protein localisation, the authors seem to have selected specific regions of the confocal micrographs for quantification (denoted by white boxes) arbitrarily. To reduce bias, we recommend that the authors remove the quantification altogether, or select a greater number of representative and equivalent regions to quantify. Note that visually there seems to be greater overlap between NbUCH15-GFP and the RFP control than with RipE1-RFP. The proteins might still interact despite their signals not completely overlapping, as their interaction will be Ub-dependent (RipE1 will likely dissociate from UCHs upon deubiquitination), and UCHs might also associate with other proteins besides RipE1.

• We appreciated that the authors took the time to provide a quality control for VIGS using RT-PCR. However, since the silencing of each of the NbUCHs was determined to be affecting all of them non-specifically it is misleading that in the figures the authors are labelling each of them as being silenced separately. 

• Key experimental design details are missing from the methods regarding gene silencing experiments. The authors use the Sol Genomics Network to design silencing fragments, however key details such as the fragment size chosen, and the genome database used are missing from the methods. Moreover, an alignment of the fragments used to silence the three different NbUCHs should be deposited in the supplement. At the moment only the primers used to generate these constructs have been included in the paper. This is especially important since off-target, non-specific silencing of the NbUCHs took place.

• Due to off-target gene silencing it would have been preferable if the authors had gone on to test more gene silencing fragments, targeting different regions of the candidate genes, prior to conducting their experiments. If gene silencing is not specific we cannot exclude the possibility that other genes are also silenced, and that observed phenotypes are not necessarily resulting from the silencing of RipE1-interacting ubiquitin proteases. It should be noted that the genome database of the Sol Genomics Network might not be up-to-date, and cross-verifying the fragments chosen against other genomics databases (Ranawaka, Buddhini, et. al., Nature Plant, 2023; or Kurotani, Ken-ichi, et al., Plant and Cell Physiology, 2023) might help explain the off-target silencing observed, and prevent it in future designs. 

Minor points:

• A few of the blots were cropped presumably to facilitate interpretation, however we are not sure in this case if it promotes clarity. For example, in Figure 2B, anti-GFP antibody treatments are cropped to only show bands above 25 and 70 kDa. It would be nicer to include the whole anti-GFP blots from ~100-25 kDa, throughout. Figure size should not be impacted too much by doing so, and it is important to show that tehse proteins were detected on the same blot. 

• The figures could have been better aligned throughout, to avoid large empty spaces.

• In HR assays the authors use an unspecified empty vector (EV) as negative control. It would be nice to indicate which plasmid this refers to per experiment, in the figure legends and/or the methods section. 

• We felt that “hijacks” might be too strong of a word in the title, “associates” or “interacts with” might be more accurate as an alternative considering the results presented. 

Discussion:

We have no specific comments regarding the discussion. We found that this section was more effective than the introduction in presenting the rationale of the study and providing key information that is essential to understand the significance of RipE1 retaining its stability.

Closing remarks:

Overall, the manuscript was easy to follow, however it lacked clear descriptions of the rationale behind the study and each of the key experiments presented in the results. We are confident that the authors could adapt the text and in some cases the analysis of the results, to present a more complete, accurate and interesting story.

Competing interests

The authors declare that they have no competing interests.