Oh no! I thought the flag meant post & accidentally reported your thoughtful feedback to moderators instead of replying. Hopefully they figure it out (no obvious way to contact them or "unflag"

Comment Figure 2C → please include indication of statistical significance Figure 3C → please include indication **of statistical significance Figure 6A → please include indication of statistical significance Figure 8B → please include indication of statistical significance Figure S1B → please include indication of statistical significance Figure S3B → please include indication of statistical significance

Response Easy to add

Comment For your overexpression experiments, do the overexpressed proteins have a tag? It would be helpful to have Western blot data showing that the particular proteins are actually being overexpressed. I think the phenotypes that you observe are very compelling so I don’t doubt the conclusions. Western blot data would just provide some additional confirmation that you are actually achieving overexpression of UppS, MraY, and BcrC.

Response The proteins are untagged. For the UppS and BcrC the cell shortening occurs with addition of inducer, , so strong indication expression is occurring. A western would provide information about degree of overexpression, but we don’t think is necessary to support conclusion drawn. Do you think there is an alternative possibility that needs to be excluded? We note that in another preprint (https://www.biorxiv.org/content/10.1101/2022.02.03.479048v1) the authors delete the native uppS in their inducible Phy-uppS strain (Fig S4) and at 100 uM IPTG (10X less than what we used in experiment) the cells have wt growth on LB plates, so we at least know the Phy-uppS is functional and made (or they would die!). We are introducing the uppS deletion into our strain to see if we can identify a concentration of IPTG that doesn’t affect cell growth but still induces shortening.

For MraY, the result is negative, so you are spot on – it is impossible to tell if due to lack of overexpression from data shown. We only know the strain is correctly made from sequencing. We will investigate if there is an antibody or functional fusion available. The reason we were not sure was worth doing is because the MraY reaction is reversible (15131133). This means that without a phenotype, there is no simple way to know the reaction can even be pushed forward even if the overexpression is confirmed (more negative data). We actually overexpressed some other proteins that act downstream (MraY, MurJ, AmJ) and they were also negative for shortening. Probably we should remove the negative data or reword to make the caveats of the negative result clear.

Question Based on your data, there are definitely differences in gene expression when you compare cells grown in media with and without magnesium. Because the majority in cell length increase occurs in such a short time though (the first 10min), I was wondering if you think that some or most of it is not due to gene expression?

Response The shortening is even faster than 10 min (not only statistically significant, but also obvious qualitatively if we mount immediately after adding Mg2+ ). We did not include the first timepoint because original purpose was to check everything was ready with microscope – did not expect shortening so fast! We can definitely add that data in. When we saw, we tried to capture the transition on pads, but going from culture to pad seems to stress the cells too much in the small window where the cool stuff happens. Since growth rate doesn’t appear to be a big factor in those initial divisions, we might be able to grow at lower temp and shift to pads for adjustment period before adding Mg2+. Did not play with it much due to lack of resources atm, but a flowcell setup would probably be best. In short, we think rapid divisions right after transition do not require transcription or translation. It really “smells” more like a biophysical thing.

Question Do you have any hypotheses what is most likely to be affected by magnesium? Do you think if the membrane may be affected?

Response We have a lot of hypotheses, but all speculative. There could be an extracytoplasmic enzyme involved in envelope synthesis is sensitive to Mg2+ availability, and that at lower concentrations, its activity is affected. There is some old literature with membrane preps that suggests PG synthesis requires higher Mg2+ than teichoic acid synthesis. If Und-P is limiting, higher Mg2+ may shift make the pool more available to make the septum. Tingfeng initially hypothesized there might be a receptor/signal mechanism but has not been able to identify one. Und-P seems to be important, but “availability” is not just pool, but how fast (and where!) the flipping across the membrane occurs. If Und-PP needs to be dephosphorylated to Und-P before being flipped back to cytoplasmic side, anything that effects the PP/Pi equilibrium would be predicted to affect the reaction rate, with lower Pi (in periplasm or pseudoperiplasm in case of G+) favoring the dephosphorylation. Cell wall associated Mg2+ could shift equilibrium to be more favorable for a Und-PP phosphatase more closely associated with the divisome. I could go all day… In short, we don’t know enough!

Question Why do you think less magnesium activates this program of less division and more elongation? Additionally why is abundant magnesium activating a program of increased cell division and less elongation? Do you think there is some evolutionary advantage, especially considering how important magnesium is for ATP production?

Response In the window we looked at, the elongation rate is constant (not less or more) and only the division frequency changes. Some bacteria (like Caulobacter and to lesser extent E. coli) clearly elongate and divide simultaneously, so model there is some competition for substrate (like Lipid II) makes sense. Septators like Bacillus seem to delineate the two processes more, though we have found conditions where even Bacillus invaginates during division, so it’s not absolute. Like eukaryotic cells, bacterial undoubtedly have mechanisms not only commit to a round of DNA replication when there is some signal that resources are sufficient. Clearly with some bacteria, this is not the case with cell division. The alternative would be that every cell cycle there is an opportunity to divide if some threshold of something(s) is reached. There is a hypothesis from Mtb literature that it may be GTP, but it’s not at all clear that is sufficient. In yeast, size at cell division is affected by perturbing 1-C pool.

Question Related to this previous question, I also wonder if this magnesium-dependent phenotype would extend to other unicellular organisms, may be protists or algae? That would be a really exciting direction to explore!

Response It’s a great question – lots to do! We didn’t even look at another Gram-positive, but we plan to. It’s trickier to limit Mg2+ in Gram-negatives (see 27471053 – we tried Bsub homolog for those wondering – it’s not responsible for phenotype we see).

Question Regarding the zinc and manganese experiments, why do you think they lead to additional phenotypes compared to magnesium? Do you have any hypotheses?

Response We have hypotheses, but if my (Jen’s @rosh_ba) twitter engagement is any indication, way too speculative for public consumption at present. Need acquire preliminary data/write grant.

Question Regarding your results that Lipid I availability may be a major a problem for the cell division in the absence of magnesium, do you think that is due to effects magnesium has on the enzymes directly, or do you think magnesium affects the substrate availability/conformation by coordinating the phosphate groups? Or something else, may be membrane conformation?

Response Several proteins involved in envelope synthesis (like UppS) are Mg2+ dependent enzymes. But at least for intracellular players, levels of Mg2+ should be more than high enough to support enzyme activity (0.8 – 3.0 mM is Bsub range I recall off top of head). Could have impact extracytoplasmically by lowering pool sponged into the cell wall, but intuition (for what that is worth) is that it is not the coordination of an enzyme with a metal that is impacted rather the equilibrium with other ions like Pi and H+ and their impact on net ATP synthesis. Lots to think about and do, and no simple answers. When Tingfeng started project idea was to find mechanism – didn’t realize we were asking “How does the cell work?” Turned out to be a bit much for a dissertation project :)