RETRACTED: Pyrophosphate modulates plant stress responses via SUMOylation

Essential revisions:

Title: The title should include "Arabidopsis thaliana" (or plants).

We have changed the title to “Pyrophosphate modulates plant stress responses via SUMOylation”. We are admittedly not entirely happy with this as we feel that our results for yeast and the in vitro assay with the human E1 and E2 enzymes suggest that the inhibitory effect of PPi is not limited to plants. We hope that our Abstract along with the keywords are sufficient to get attention from non-plant researchers.

Main concerns:

For the following items, additional experiments/analyses are required:

1) Figure 1B-C: please provide appropriate statistical tests with p-values to show (in Figure 1B) that the leaf numbers analyzed give a sound difference between Col-0 and fugu5-1 after cold-acclimation (chi-square test?). The percentage given does not show that the population is large enough to conclude that the difference seen is significant.

Statistical analysis has been added.

2) Provide AVP1 gene expression data in response to cold/heat/salt (using the RNA samples of Figure 2A). Data/information on AVP1 gene expression in response to cold/heat/salt stress was deemed essential to address the following points:

"The authors argue that gene expression of the PPase genes is dynamically controlled in response to cold (subsection “PPi controls cold-acclimation via SUMOylation”, first paragraph, Figure 2A), but in the Introduction the authors refer to work by Segami et al. (2018) that revealed that the quadruple mutant ppa1 ppa2 ppa4 ppa5 does not display a major phenotype: This suggests that the gene expression of these PPa genes has nothing to do with keeping the [PPi]cyto low and rather PPi levels are primarily controlled by AVP1. The dynamic response of PPase genes to cold is thus a weak argument for 'control' and not substantiated by phenotypic data on these 4 genes (see the aforementioned paragraph)."

We agree with the reviewer that the dynamic response of sPPase genes to cold alone is not a strong argument for control and have thus included additional data to strengthen our conclusion. The quadruple mutant ppa1 ppa2 ppa4 ppa5 does indeed not display a major growth phenotype, however in the newly added Figure 1—figure supplement 3 we show that prior to cold acclimation it is less freezing tolerant than the wildtype. The fact that V-PPase activity has been shown to increase during cold acclimation (Schulze et al., Kriegel et al.) can explain that after acclimation freezing tolerance of the quadruple sPPase mutant is comparable to wt. In addition, we have included qRT-data for AVP1 in the revised Figure 2 showing that the cold-induced increase of AVP1 activity is not due to transcriptional up-regulation. Finally, publicly available gene expression data under cold, heat and salt stress confirm that AVP1 is not upregulated.

3) Provide evidence that the fugu5-1 mutation impacts the gene expression pattern of the PPa1/2/4/5 genes in response to cold (using the RNA samples of Figure 2B and 2C). This would reveal whether the gene expression pattern of these four genes in response to cold is altered in the Fugu5 mutant and if their gene expression is thus impacted by the increased PPi levels in the fuga5-1 mutant at 22C/4C. The presented increase in the PPi levels in the fugu5 mutation (Figure 1G) suggests that induction of these PPa genes does not occur as a compensatory mechanism of a rise of the PPi levels due to the fugu5 mutation.

As requested by the reviewer, we have included qRT data for the expression of sPPase genes in fugu5 in the revised Figure 2. As expected, the cold response of several PPa genes is reduced indicating that they are under control of the ICE1/CBF-transcriptome.

The following points are unclear and should be revised in the text/figures. Although some of these excerpted reviewer statements include experiments, it is the author's discretion as to whether to do the experiments, or to modify the text. An explanation of the resolution should be included in the response to reviewers.

4) In Figure 1—figure supplement 2G, authors provide data about pyrophosphatase activity. The activity in fugu5 was similar to that in Col-0. According to Figure 1—figure supplement 1, V-PPase activity in fugu5 was reduced. How is V-PPase activity in fugu5 expression UBQ:PP5a-GFP? And why did authors use UBQ:PP5a-GFP, instead of UBQ:AVP1? More clear information is required.

Figure 1—figure supplement 2G shows measurements of soluble PPase whereas Figure 1—figure supplement 1 shows PPi hydrolysis by membrane-integral H⁺-PPase in microsomal extracts. The fugu5 mutant lacks the membrane-integral H⁺-PPase AVP1 so that PPi hydrolysis by membrane-integral H⁺-PPase is absent even when the soluble PPase PPa5 is overexpressed.

5) Please provide a brief argument/observation on why the PPa5-GFP protein levels are much higher in the fugu5-1 mutant (#6, #8) than in the wt plants (Col-0, #8, #10) (Figure 1—figure supplement 2F, subsection “Lack of V-PPase activity impairs cold acclimation”, second paragraph). Apparently, the fugu5 mutation stabilizes the PPa5 protein levels, which is relevant as it suggests that increased [PPi] levels in the cytosol/nucleus promotes the protein stability of PPa5 (reduced PPa5 turnover?). A similar thing might be happening with the native PPa proteins (so the regulation is at PPa gene as well as at the PPa protein level?). Are the transgene expression levels comparable in the 4 transgenic lines (RT-qPCR for GFP amplicon and/or PPa5 amplicon)? At the same time, Figure 1G suggests that the soluble PPase enzymatic activity does not differ between the Col-0 and fugu5-1 backgrounds (all 4 transgenic lines belong all to group 'b'). Does this comparable soluble PPase activity signify that the PPa5-GFP chimera is partially inactive in a fugu5-1 background for an unknown reason?

We have consistently observed that PPa5-GFP accumulates to higher levels in the fugu5-1 mutant than in the wt in all the lines that we analysed. We have so far not analysed this any further but are planning to do this in future experiments. PPa5-GFP is expressed under control of the UBI10-promotor in both genotypes and we thus assume that we are indeed facing an effect on protein stability but we will of course have to demonstrate that transgene expression is not affected. The fact that soluble PPase activity does not differ might indeed indicate yet another level of regulation and we will follow up on this. We have modified the text in the manuscript accordingly.

6) In Figure 1B and 1C, statistical analyses are required. And freezing tolerance of UBQ:AVP1 seems to be increased, compared to Col-0, but why is freezing tolerance of UBQ:PPa5/Col-0 similar to that of Col-0? If PPi hydrolysis is important, UBQ:PPa5/Col-0 may exhibit tolerance to freezing.

Statistical analysis was added for Figure 1B and 1C. As UBQ:PPa5 might not reduce PPi levels to the same extent as UBQ:AVP1, it might not be sufficient to improve freezing tolerance. We are currently testing transgenic lines that express PPa5 under -inducible promotors and hope to report on their freezing tolerance in the near future.

7) According to Figure 2A, there are at least 4 sPPase. But why did authors use PPa5 for overexpression? How about overexpression of other sPPase in fugu5? Do other sPPase rescue cold sensitivity of fugu5?

Please see response to point 5.

8) According to phenotype and SUMOylation assay (Figure 1 and 4), PPi level looks important to regulate SUMOylation and phenotype. Thus, how do plants exhibit when PPi is applied by a foliar spray? Application of PPi by a foliar spray to Col-0 interferes SUMOylation and decrease tolerance to freezing and heat stresses, right? Check phenotype and SUMOylation after application of PPi to Col-0.

We have not tried foliar spraying of PPi but we have tried to grow plants in the presence of PPi in the media. However, as PPi is not a very stable molecule and is not likely to be taken up, we gave up on these experiments as we did not observe any noticeable effects.

9) A potential weak point is that other processes that might be inhibited by PPi may contribute to the physiologic phenotypes (cold stress tolerance etc.). The authors discuss formation of UDP-glucose as one process that may not contribute to the observed phenotypes. However, the Discussion should give more room to the possibility of contribution of other processes not considered or investigated in the manuscript. In particular, ubiquitin activation uses an enzyme similar to SAE, which might also be inhibitable by PPi. Ubiquitin conjugation almost certainly contributes to stress responses, as well.

The Discussion has been modified as follows:

“It remains to be determined if accumulation of PPi has a similar effect on ubiquitination. However, reduced ubiquitination would stabilize ICE1 and would result in increased cold tolerance. As we cannot exclude that the altered sugar metabolism of fugu5 directly contributes to freezing tolerance or indirectly impinges SUMOylation during cold acclimation, we extended our analysis to the heat stress response.”