Referee 2's review:

Summary. The authors seek to understand why the sigma-E protein is essential in normal E. coli that is not stressed. They address this question with two complementary approaches: by surveying the morphological phenotypes that occur when sigma-E is inactivated, and by isolating genes that, when present in multiple copies, can suppress the loss of viability in the absence of functional sigma-E. In the first case, the authors present a logical and thorough progression of experiments to narrow the possible mechanisms that could explain cell death associated with sigma-E inhibition. Their second approach, the isolation of multicopy suppressors, is a classical method for discovering new pathways and for finding unknown functions of known proteins. A strong conclusion from the work is that, whatever the eventual mechanism may be, the loss of viability is related to the proper development of the cell envelope. Much less clear is how either of the two suppressor genes is able to reverse these effects, although the authors present strong evidence regarding the participation of PtsN.

Comments

1. The authors state that the "blebs rarely formed at either the poles or the septum" (p 9 lines 1-2). If this is true, then the cells depicted in Fig. 4 are either exceptions to this rule or contradict the statement. Of the blebs pictured there, four arise at or near the center of a cell (at possible septal sites) and two others arise at a pole or near a pole where one might expect that a minicell division septum might occur. These photographs seem more in line with the interpretation that a defect in cell division might be at work (mentioned as a minor possibility on p 17, last two lines). Because the authors say they have several hundred observations of such blebs, it would seem relatively simple to devise a measuring system (e.g., a percentage of the length of the rod cell) to locate these blebs and generate a histogram of bleb locations. Then they could determine more objectively the relationship of the blebs to potential cell division sites.

2. The sigma-E protein is inactivated rapidly by sequestration by overproducing RseA. However, the length of time required for visible phenotypes to appear (blebbing, lysis) takes from 2.5 to 3 hours (2.5 to 3 generations). This long lag between the loss of sigma-E and the phenotypic effects would seem to indicate that the killing effect is not direct but is indirect (as the authors suggest elsewhere for other reasons - p 19 lines 3-4).

3. The repeated statement (in many places) that sigma-E is "required for maintenance of cell envelope integrity" is pretty vague (as I'm sure the authors know). The "envelope integrity" phrase seems devoid of any explanatory power, except that "the cell breaks down." This could be for any of a number of reasons. Is it possible to find a phrase that describes what is happening without implying a specific mechanism?

4. Related to the above, a breakdown in "cellular integrity" implies that the whole envelope is affected, which the authors clearly show is not the case (e.g., with their nice examination of membrane composition, mentioned on p 16 lines 15-16). This would imply that a much more specific, defined, and strictly localized effect is occurring, instead of a process that changes the overall "envelope integrity."

5. In a similar vein, what does the idea "strengthen the cell envelope" mean (p 20 line 13)?

6. It would be very useful to have an integrated Figure of the pathways being discussed in order to orient the reader during the complex descriptions in the Results and Discussion sections.

Other comments

7. At the end of the legend to Fig. 3 (p 38 lines 7-8), it seems as though labels are incorrect for the control and sigma-E panels. Shouldn't the control be the top panel and the inhibition of sigma-E be at the bottom?

*********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.