SIRT2 and lysine fatty acylation regulate the transforming activity of K-Ras4a

Essential revisions:

The general findings are, in principle, of interest to the readership of eLife. However, the following points should first be addressed:

1) The two different assays for localization of K-Ras4a versus K-Ras4a-G12V raise questions about the actual role of fatty acylation in regulating these two proteins. The modest effects in Figure 4 are based on analysis of punctate signals. However, in Figure 6C, there is not much intracellular punctate localization for K-Ras4a-G12V, and A-Raf looks colocalized only with membrane-bound K-Ras4a-G12V. These results seem to suggest that there is a difference in fatty acylation levels between K-Ras4a and K-Ras4a-G12V. The authors should account for these discrepancies and provide information on the fatty acylation levels of K-Ras4a versus K-Ras4a-G12V.

We agree with the reviewers that K-Ras4a-G12V (Figure 6C) exhibits weaker intracellular punctate localization than K-Ras4a (Figure. 4A) (P < 0.01 by two-way ANOVA analysis). However, K-Ras4a and K-Ras4a-G12V possess comparable total and lysine fatty acylation levels (Figure 4—figure supplement 1C), suggesting that the difference in their subcellular localization is not due to different levels of fatty acylation. It is possible that the activation state of K-Ras4a influences its dynamics on plasma membrane and endomembranes, the mechanism for which remains to be elucidated in the future. We have added discussion on this in the revised manuscript (Discussion, fifth paragraph).

We apologize for not clearly showing the intracellular colocalization of K-Ras4a-G12V and A-Raf. We have added a Merge 2 column in the revised Figure 6C to show magnified regions in Merge 1 and to better demonstrate the endomembrane colocalization of K-Ras4a-G12V and A-Raf. The figure legend for Figure 6C has been revised accordingly.

2) The authors have primarily used K-R mutants as "loss of Lys-fatty-acylation" alleles. However, the multiple Arg substitutions have the potential to introduce alternative modes of membrane binding independent of fatty-acylation. Indeed, in several figures (Figure 5, Figure 5—figure supplement 1, Figure 6—figure supplement 1E), the effects of the 3KR substitution seemed largely independent of SIRT2. It would be important for the authors to include activity and location data on Lys -> Ala mutants, especially for K-Ras4a, in order to support the conclusions.

3KR mutation may not be solely due to lack of lysine fatty acylation. Therefore, we think it is critical to utilize K-Ras4a-3KR mutant in combination with SIRT2 KD to study the function of K-Ras4a lysine fatty acylation. If a biological effect is due to lysine fatty acylation, SIRT2 KD should significantly affect K-Ras4a WT but not 3KR mutant in the most ideal case. This is indeed what we observed for the endomembrane localization (Figure 4A), transforming activity (Figure 5) and A-Raf binding (Figure 6A and B).

Recently, Tsai et al.¹ and Zhao et al.² showed that the lysine to uncharged glutamine mutation (3KQ) decreased transforming activity of K-Ras4a-G12V, suggesting that positive charge is favorable for K-Ras4a activity. Based on these studies, the lysine to alanine mutation (KA), which does not preserve the positive charge, may add more complexity to the membrane binding modes of K-Ras4a than the KR mutation. Any effect caused by the KA mutation may be due to lack of lysine fatty acylation, lack of positive charge, or alternative modes of membrane binding independent of fatty acylation, which will make the interpretation of the results more complicated and therefore will not further support our conclusions. In contrast, with the KR mutant, we only need to consider whether the effect is due to lysine fatty acylation or alternative modes of membrane binding independent of fatty acylation. In other words, we think the KR mutant is the closest to the WT without lysine fatty acylation (i.e. with SIRT2 present) and thus the most logical to use for our purpose.

3) The effects of SIRT2 knockdown in Figures 4 and 5, as well as in Figure 2E are small and thus not entirely convincing. This may be due to an incomplete knockdown, for example as seen in Figure 2E. More significant effects that may be obtainable by knocking out SIRT2 would make a more convincing case for the role of SIRT2 in regulating K-Ras4a.

As the reviewers suggested, we indeed tried to compare the lysine fatty acylation and subcellular localization of K-Ras4a in Sirt2 wildtype (WT) and knockout (KO) mouse embryonic fibroblast (MEF) cells. Lysine fatty acylation level for K-Ras4a-WT in the KO cells was 1.56-fold of that in the WT cells, whereas lysine fatty acylation of K-Ras4a-3KR was similar in WT and KO cells (Figure 3—figure supplement 1C and D, subsection “Mapping the fatty acylated lysine residues regulated by SIRT2”, last paragraph in the revised manuscript). In Sirt2 WT MEF cells, K-Ras4a-3KR displayed slightly but significantly more intracellular punctate localization than K-Ras4a-WT. K-Ras4a-WT showed significantly less intracellular punctate localization in KO cells than in WT cells, whereas K-Ras4a-3KR exhibited comparable level of punctate localization in WT and KO cells (Figure 4—figure supplement 1B, subsection “Lysine fatty acylation regulates subcellular localization of K-Ras4a”, first paragraph). Collectively, these effects of Sirt2 KO in MEF cells are consistent with that of SIRT2 KD in HEK293T cells, which further strengthens our conclusion that SIRT2-mediated lysine defatty-acylation regulates K-Ras4a intracellular localization. However, we did not observe more significant effects of Sirt2 KO in MEF cells than that of SIRT2 KD in HEK293T cells, which is likely because we had over 95% of knockdown efficiency (based on quantification of SIRT2 signal intensity in Figure 2C and E) with shSIRT2-#2 in HEK293T cells.

We have also attempted to examine the transforming activities of K-Ras4a-G12V and -G12V-3KR in Sirt2 WT and KO MEF cells. However, introduction of both K-Ras4a-G12V and -G12V-3KR led to oncogene-induced senescence in both WT and KO cells. Considering that Sirt2 KO shows similar effects on K-Ras4a lysine fatty acylation and subcellular localization to SIRT2 KD, we believe that our current data obtained from 3T3 cells with near 95% SIRT2 KD efficiency (Figure 6A) is significant enough to support our conclusion.

References

1) Tsai, F.D., Lopes, M.S., Zhou, M., Court, H., Ponce, O., Fiordalisi, J.J., Gierut, J.J., Cox, A.D., Haigis, K.M. & Philips, M.R. K-Ras4A splice variant is widely expressed in cancer and uses a hybrid membrane-targeting motif. Proc Natl Acad Sci U S A (2015).

2) Zhao, H., Liu, P., Zhang, R., Wu, M., Li, D., Zhao, X., Zhang, C., Jiao, B., Chen, B., Chen, Z. & Ren, R. Roles of palmitoylation and the KIKK membrane-targeting motif in leukemogenesis by oncogenic KRAS4A. J Hematol Oncol 8, 132 (2015).