Technical CommentsPosttranslational Modifications

Response to Comment on “SUMO deconjugation is required for arsenic-triggered ubiquitylation of PML”

See allHide authors and affiliations

Sci. Signal.  09 Aug 2016:
Vol. 9, Issue 440, pp. tc2
DOI: 10.1126/scisignal.aad9777


Arsenic trioxide chemotherapy cures acute promyelocytic leukemia by inducing the ubiquitylation of an oncogenic fusion protein containing promyelocytic leukemia protein (PML) subsequent to modification of PML by SUMO1 and SUMO2. We proposed that the SUMO switch at Lys65 of PML enhanced subsequent SUMO2 conjugation to Lys160 and consequent RNF4-dependent ubiquitylation of PML. Ferhi et al. note differences between their experimental system and ours regarding the outcome and mechanisms of SUMO-dependent PML signaling. When confronted by apparently contradictory data, it is appropriate to drill down to where the differences could lie.

It seems that a concern of Ferhi et al. (1) is that arsenic did not trigger promyelocytic leukemia protein (PML) degradation in our experimental systems. Degradation of PML is an end point used by some scientists; however, we note that although all PML isoforms are ubiquitylated after arsenic treatment, they are not all degraded. For example, although PML-V disappears after 24 hours of treatment, other isoforms—including the isoform PML-IV used widely by the community and used in our paper—persist after 24 hours of treatment, as demonstrated previously (2). This is why we chose to use ubiquitylation, not disappearance, as an experimental readout. One wonders whether the cell lysis and extraction methods may influence interpretations because we showed in Fig. 4C [in (3)] that endogenous PML does not disappear from total cell material after arsenic treatment but does indeed seem to vanish from Triton-soluble material. However, we may have caused confusion by including the word “degraded” in the abstract, which was meant to summarize briefly the field, and the abstract is the only place throughout the article where we use this term to describe outcomes of arsenic treatment of PML.

We, too, were surprised that arsenic enhanced PML-IV abundance; however, the data were made clear in our paper using multiple experiments and two independent cell lines stably transfected with PML-IV. The possibility that this reflected the observed increases in SUMO or ubiquitin conjugates in response to arsenic was ruled out because each experiment contained internal controls—comparisons between SUMO or SUMO mutant transfections, or SENP knockdown—demonstrating that the observed ubiquitylation changes do not result from different expression amounts.

Another concern was that our study used cells stably transfected with PML-IV for most of our experiments. Again, this is in line with common practice in the field, and we were inspired by the excellent cell biochemistry by leaders in the field, including the authors of the Technical Comment, to pursue exactly this avenue of experimentation (2, 47). Our experiments were carefully controlled, and we provided extensive repeats, quantitation, and statistical assessment. We totally agree that ectopic expression is not equal to endogenous expression; however, if one is going to analyze a specific isoform, there is really no other way to go. Today, we would assume that most laboratories will be pursuing CRISPr technology to tag or alter coding sequences of endogenous proteins; however, this is going to be quite difficult in the PML system where the mechanism of splicing that generates multiple isoforms is at play. We look forward to developments in this area.

Ferhi et al. claim that within their experimental setting, they could still observe a response of PML to arsenic treatment after mutation of Lys65 in PML. We have no argument with this observation. Mutation of Lys65 in PML does not preclude the interaction between UBC9 and the RING domain of PML, and Lys160 can potentially still be conjugated by SUMO2. In our paper (3), we showed that upon mutation of this residue, the ubiquitin signal could still be propagated but less efficiently. In our experience, protein modification is not all-or-none (cell biochemistry is not binary; it is measurable); therefore, there are quantifiable degrees of modification that can result in a particular biology. Indeed, this is exactly the point that we are making with our paper. Moreover, in our discussion, we clearly proposed that the switch in SUMO paralog preference on PML Lys65 could simply be a way to shift the SUMO conjugation site preference, promoting SUMO2 conjugation on Lys160. Once Lys65 is modified by SUMO1, UBC9 defaults to conjugate SUMO2 on Lys160 and propagates the signal.

Finally, Ferhi et al. (1) question the requirement of deSUMOylation to promote the arsenic response on PML, a major finding of our work. Regrettably, they do not provide a solid alternative hypothesis. We have clearly demonstrated the requirement of deSUMOylation to drive the ubiquitin signal with two orthogonal approaches: nondeconjugatable SUMO mutants and knockdown of multiple proteases. The experiments were performed in two different cell lines, and the caveats of the approach were clearly explained in the article.

In conclusion, our work demonstrates that coordinated SUMO conjugation and deconjugation is required for SUMO-dependent ubiquitylation of PML after arsenic exposure.


View Abstract

Navigate This Article