Elisa Pedone, Dan Lazzarini Rocca, Lorena Postiglione, Francesco Aulicino, Sandra Montes-Olivas, Diego di Bernardo, Maria Pia Cosma, Lucia Marucci bioRxiv 404699 doi: https://doi.org/10.1101/404699

Overview

Pedone et al. developed a system for simultaneous control of gene expression and protein stability in living mammalian cells. This system enables fine, fast and reversible control of intracellular protein levels in time and space. They tested their system in a mouse embryonic stem cell line, which is difficult to handle, using a major signalling protein (beta-catenin). They also developed a mathematical model to fit experimental results (for prediction of protein levels), as well as an in silico platform for automated control of protein induction based on microfluidics and microscopy.

The paper is very well written. The figures are crisp and clear and the organization of information has been taken good care of. There is a good consistency between the author's statements in the Abstract and Main text, and claims in the abstract are proven.

The paper is highly technical and these reviewers are not fully qualified to analyze the quality of the mathematical model. We see potential for their system to address biological questions, and look forward to it being employed in this way. The system combines two powerful tools and successful demonstrates an application to a challenge in synthetic biology. One possibly application would be to examine how disregulation of protein homeostasis can be critical for protein misfolding disorders; a short study on an aggregation-prone protein could increase highly the impact of the paper and make available and interesting tool for a wider range of researchers. We also wondered whether the range of activation and degradation is sufficient for controlling a diversity of systems, or it they are valid only for particular cases.

We note that the study describes use of the Student t-test without correction for multiple comparison. However, all relevant differences between samples are certainly large enough to survive such a correction.

Minor points & points of discussion:

Doubts were raised about the suitability of the term “destabilization”, because what the TMP/DD system does is regulate the degradation of the protein. We are conscious that “destabilisation” is commonly used in the bibliography for this sort of system. However, interfering with protein homeostasis adds a very relevant biological variable to experiments.
The system presents the limitations of all protein tags, e.g. that some proteins do not accept tags in the N-terminal or C-terminal domains, and must be empirically tested for suitability. B-catenin is a protein widely described in the literature, but the system may not be so flexible for less known proteins. Also, qualitative similarity between B-catenin distributions may not prove equivalent functionality of the fusion protein.
There could be some discussion of how/whether cell autofluorescence was accounted for in analysis.
The results for DDrtTa are interesting from a synthetic biology standpoint; we wonder whether DDrtTa would make much sense a feedback system aiming for more rapidly achieving protein homeostasis since increasing the rate the activator will turn off is counterbalanced by increasing the time it takes to turn on.
One can imagine genetic networks for which the "age" of a protein is as important as its concentration because post-translational modification, localization, finding binding partners, etc takes time. A system that achieves more rapid control of protein concentration by reducing average protein lifetime may not be ideal for interrogating this kind of network.

Note that (1) all of these minor points are meant to inspire discussion and not things that need to change for the article to be scientifically sound in our opinion and (2) we apologize to the authors if we have misread/misinterpreted any aspect of their work and welcome any response to this review.