Reprogrammed peptidoglycan elongation reveals plasticity in bacterial growth modes

Reviewer #1 (Public review):

This is an interesting paper, with the primary finding being that localizing MreB or PBP2 to the cell poles in E. coli is primarily demonstrated via an aggregate formed by expressing M. xanthus MreB.

I have 2 main concerns:

(1) First, the authors should clarify and adjust their interpretation of FDAA incorporation: As written, the authors interpret FDAA incorporation as being caused by incorporation during PG polymerization. However, in E. coli, FDAA incorporation does not result from the elongation of PG strands or their initial 4-3 crosslinking by DD transpeptidases following polymerization, but rather by the remodeling of L,D-transpeptidases.

Thus, it is not accurate to refer to the FDAA incorporation as PG elongation, but rather the modification of crosslinks from 4-3 to 3-3 crosslinks at that location. To claim a link to PG polymerization, other experiments or substantial explanations are needed.

(E. coli Cells Incorporate FDAAs by L,D-TPases in a Growth Independent Manner) - https://doi.org/10.1021/acschembio.

(2) Second, the evidence provided that this is polar elongation is not sufficient to prove elongation. In all images claiming polar growth, the FDAA focus appears as a single spot, which corresponds to the MreB aggregate visible in bright-field. That might indicate incorporation, but it does not demonstrate polar elongation. To prove this, the authors should do different-length pulses of FDAAs and demonstrate an increasing length of labeled PG along the cell. Cells with one focus should show increasing length; polar foci should elongate from both ends. I find the current 2-color labeling insufficient, as BADA labels the entire cell.

Small points:

(3) Lines 350- 302: "In this case, as the nonpolar region is no longer the growth zone, the established cylindrical PG structure is sufficient to maintain cell width, where MreB filaments become nonessential." Does polar elongation give robustness to rod shape? The authors should include an analysis of cell width and its variation within a cell and between cells.

(4) In the abstract: "This reprogrammed growth mode bypasses the requirement for MreB filaments, highlighting a plasticity of the Rod system that suggests polar elongation may have emerged through the evolutionary loss of MreB." This argument should not be made without evolutionary analysis or reference to such work indicating this is the case.

(5) 365-367 "PG-depleted spheroplasts can spontaneously regenerate rod shape through curvature-dependent localization of MreB filaments [21, 57]". This should be amended. The Billing paper did indeed study spheroplasts, but the Hussain paper used teichoic acid-depleted cells that still had a cell wall.

Reviewer #2 (Public review):

Summary:

Based on observations of localisation of MreBEc at the poles within an aggregate-like structure, upon heterologous expression of MreBMx, the authors set out to investigate how this non-canonical localisation of MreBs leads to a reprogramming of peptidoglycan synthesis to the poles. This is analogous to the polar growth observed in phyla which are not dependent on dispersed growth of PG, but only at the poles, and are MreB independent.

The authors proceed to establish that PG synthesis is MreB-dependent, Rod enzyme-dependent, and requires the prior establishment of a pole.

Strengths:

(1) It is a very interesting idea to design experiments to demonstrate reprogramming of non-polar to polar growth based on the observation of localisation of a heterologously expressed MreB.

(2) The experiments to demonstrate the factors that determine polar growth and the observation of the PG in each of these experimental situations are convincing.

(3) I find the observation of an extra layer of PG in the heterologously expressed system very intriguing. It will be interesting to see if this layer merges with the other PG layer at some stage or branches from the non-polar growth near the poles.

Weaknesses:

(1) It is not clear what exactly the identity of the polar aggregates is and how much of this activity is an artefact of partially functional MreBs.

(2) I find it intriguing that the localisation and growth are predominantly at one pole only. It is unclear to me how this can be reconciled with growth and shape maintenance, and an increase in length and width. Is the increase in length and width a consequence of misshapen cells that are bulged in the absence of a normal PG layer?

(3) The authors do not follow up on the observations in the first figure on the length and width changes and the extra peptidoglycan layer (which I feel are the most interesting aspects), and how this can be connected to the polar growth observed in the later sections of the manuscript.

(4) The claim that this could be a precursor of an MreB-independent polar growth mechanism appears to be a bit far-fetched, because the system is still dependent on having an established pole for PG synthesis to occur in the new place.

Reviewer #3 (Public review):

Summary:

Most rod-shaped bacteria grow by one of two mechanisms: growth from the pole or growth from the midcell. It is rare for a single species to utilize both modes of growth, although a few examples do exist. Here, the authors have artificially induced E. coli cells to grow from the poles, either by expressing mreB from Myxoccocus xanthus in E. coli, leading to the mislocalization of MreB to the poles in large aggregates, or by forcing the localization of major cell wall synthesis proteins to the cell pole. The fact that cells switched modes of growth suggests an evolutionary pathway from midcell to polar growing cells as well as suggests that there might be unknown conditions in nature when cells may switch growth modes.

Strengths:

(1) The authors use a strain that has replaced mreB with a functional fluorescent version at the native site. This eliminates any effects of having two copies of mreB. Because MreB is fluorescently tagged, they can monitor its localization when mreB from M. xanthus is expressed in E. coli. They notice that MreBec now forms bright polar foci and that there appear to be changes to the cell wall at the pole.

(2) D-amino acids are specific to the cell wall, and fluorescent versions (FDAA) have been used to mark sites of new cell wall insertion. The authors use these FDAAs to determine how cell wall synthesis correlates to MreB and if that changes when MreBmx is expressed. Again, there is pretty clear evidence that cell wall synthesis follows MreB localization to the pole.

(3) MreB itself does not synthesize the cell wall, but localizes the proteins, such as PBP2, that do. Using a published method to force proteins to the pole, the authors show that when they target PBP2 to the pole, they can phenocopy the polar growth seen when MreB is polar. Interestingly, these cells become resistant to A22, a drug that targets MreB, suggesting that localized growth at the pole does not require MreB and is sufficient to maintain rod shape.

Weaknesses:

(1) The authors do not show what the poles of control cells look like, making it difficult to determine if there is a change when MreBmx is expressed. However, the localization of both MreBec and MreBmx clearly forms bright foci at the pole.

(2) While more quantification is needed, the authors show some evidence that RodZ, an MreB interaction partner, is needed for this polar growth, as cells lacking rodZ still form foci at pole-like regions when MreBmx is in the cell; however, these cells remain spherical and do not elongate from these foci.

When MreB is deleted, and cells become spherical, the authors were unable to cause the polar growth mode. They suggest that this is due to the lack of a preexisting pole; however, experimental evidence to test this is missing.

Conclusion:

Overall, the authors do a good job of showing that E. coli can grow with a polar method rather than a midcell method of cell wall insertion. It is unclear why MreBec forms at poles when MreBmx is present and even if this MreB is functional. The foci look similar to inclusion bodies, which are normally aggregates of misfolded proteins that migrate to the poles. Past work has shown that when MreB is more polarly localized, branches form, which is not seen here. Importantly, the authors also show that there is feedback between the localization of MreB and PBP2 as both appear to regulate the localization of the other.

Author response:

Public Reviews:

Reviewer #1 (Public review):

This is an interesting paper, with the primary finding being that localizing MreB or PBP2 to the cell poles in E. coli is primarily demonstrated via an aggregate formed by expressing M. xanthus MreB.

I have 2 main concerns:

(1) First, the authors should clarify and adjust their interpretation of FDAA incorporation: As written, the authors interpret FDAA incorporation as being caused by incorporation during PG polymerization. However, in E. coli, FDAA incorporation does not result from the elongation of PG strands or their initial 4-3 crosslinking by DD transpeptidases following polymerization, but rather by the remodeling of L,D-transpeptidases.

Thus, it is not accurate to refer to the FDAA incorporation as PG elongation, but rather the modification of crosslinks from 4-3 to 3-3 crosslinks at that location. To claim a link to PG polymerization, other experiments or substantial explanations are needed.

(E. coli Cells Incorporate FDAAs by L,D-TPases in a Growth Independent Manner) - https://doi.org/10.1021/acschembio.

We appreciate the reviewer for bringing up this important question. We will address this concern by further explaining our results in the revised manuscript.

However, we respectfully disagree with the reviewer for two reasons:

First, the paper by Kuru et al. (mentioned by the reviewer) revealed that (exact quote): “Our in vitro and in vivo data unequivocally demonstrate that these bacteria incorporate FDAAs using two extra cytoplasmic pathways: through activity of their D, D-transpeptidases, and, if present, by their L, D-transpeptidases…These mechanistic findings enabled development of a new, FDAA-based, in vitro labelling approach that reports on subcellular distribution of muropeptides, an especially important attribute to enable the study of bacteria with poorly defined growth modes” (Kuru et al., 2019).

Thus, the paper by Kuru et al. identified two FDAA labeling patterns, a D, D-transpeptidase-dependent, concentrated labeling for PG growth and an L, D-transpeptidase-dependent, growth-independent labeling for PG modification along the entire cell envelope (Kuru et al., 2019). The polar FDAA foci we presented do not match the reported pattern of L, D-transpeptidase-dependent incorporation.

Second, we provided the evidence in Fig. 3b that the polar FDAA labeling is due to the activity of PBP2, a D, D-transpeptidase, because mecillinam that inhibits PBP2 is sufficient to abolish polar FDAA incorporation.

(2) Second, the evidence provided that this is polar elongation is not sufficient to prove elongation. In all images claiming polar growth, the FDAA focus appears as a single spot, which corresponds to the MreB aggregate visible in bright-field. That might indicate incorporation, but it does not demonstrate polar elongation. To prove this, the authors should do different-length pulses of FDAAs and demonstrate an increasing length of labeled PG along the cell. Cells with one focus should show increasing length; polar foci should elongate from both ends. I find the current 2-color labeling insufficient, as BADA labels the entire cell.

We appreciate this comment and totally agree with the reviewer. The wide BADA labeling band could indeed come from L, D-transpeptidases. We will follow the reviewer’s recommendation to address this comment with additional staining experiments.

Small points:

(3) Lines 350- 302: "In this case, as the nonpolar region is no longer the growth zone, the established cylindrical PG structure is sufficient to maintain cell width, where MreB filaments become nonessential." Does polar elongation give robustness to rod shape? The authors should include an analysis of cell width and its variation within a cell and between cells.

We appreciate this comment. Judging from the bright-field images we presented, we believe that polar elongation does give robustness to rod shape. We will follow the reviewer’s recommendation and provide the said analysis.

(4) In the abstract: "This reprogrammed growth mode bypasses the requirement for MreB filaments, highlighting a plasticity of the Rod system that suggests polar elongation may have emerged through the evolutionary loss of MreB." This argument should not be made without evolutionary analysis or reference to such work indicating this is the case.

We appreciate this comment and will remove this statement from our revised manuscript.

(5) 365-367 "PG-depleted spheroplasts can spontaneously regenerate rod shape through curvature-dependent localization of MreB filaments [21, 57]". This should be amended. The Billing paper did indeed study spheroplasts, but the Hussain paper used teichoic acid-depleted cells that still had a cell wall.

We thank the reviewer for pointing out this mistake. We will amend this statement in our revised manuscript.

Reviewer #2 (Public review):

Summary:

Based on observations of localisation of MreBEc at the poles within an aggregate-like structure, upon heterologous expression of MreBMx, the authors set out to investigate how this non-canonical localisation of MreBs leads to a reprogramming of peptidoglycan synthesis to the poles. This is analogous to the polar growth observed in phyla which are not dependent on dispersed growth of PG, but only at the poles, and are MreB independent.

The authors proceed to establish that PG synthesis is MreB-dependent, Rod enzyme-dependent, and requires the prior establishment of a pole.

Strengths:

(1) It is a very interesting idea to design experiments to demonstrate reprogramming of non-polar to polar growth based on the observation of localisation of a heterologously expressed MreB.

(2) The experiments to demonstrate the factors that determine polar growth and the observation of the PG in each of these experimental situations are convincing.

(3) I find the observation of an extra layer of PG in the heterologously expressed system very intriguing. It will be interesting to see if this layer merges with the other PG layer at some stage or branches from the non-polar growth near the poles.

Weaknesses:

(1) It is not clear what exactly the identity of the polar aggregates is and how much of this activity is an artefact of partially functional MreBs.

We appreciate this comment and will address it by further explaining our results in the revised manuscript. While we can only say that the polar aggregates resemble inclusion bodies, we do believe that they cause polar PG growth because in the cells that express MreB_Mx, polar PG growth does not occur at the poles that lack MreB aggregates.

(2) I find it intriguing that the localisation and growth are predominantly at one pole only. It is unclear to me how this can be reconciled with growth and shape maintenance, and an increase in length and width. Is the increase in length and width a consequence of misshapen cells that are bulged in the absence of a normal PG layer?

We appreciate this comment. Judging from the bright-field images we presented, we believe that polar elongation does not generate bulges and is thus sufficient for maintaining rod shape. We will follow the reviewer’s recommendation to clarify this.

(3) The authors do not follow up on the observations in the first figure on the length and width changes and the extra peptidoglycan layer (which I feel are the most interesting aspects), and how this can be connected to the polar growth observed in the later sections of the manuscript.

We appreciate this comment. We believe that the thickened PG patches are integral parts of the polar PG, rather than an extra layer, which is, however, technically challenging to prove. Thus, we will relay on fluorescence microscopy to visualize polar PG growth.

(4) The claim that this could be a precursor of an MreB-independent polar growth mechanism appears to be a bit far-fetched, because the system is still dependent on having an established pole for PG synthesis to occur in the new place.

We appreciate this comment and will remove such speculations from our revised manuscript.

Reviewer #3 (Public review):

Summary:

Most rod-shaped bacteria grow by one of two mechanisms: growth from the pole or growth from the midcell. It is rare for a single species to utilize both modes of growth, although a few examples do exist. Here, the authors have artificially induced E. coli cells to grow from the poles, either by expressing mreB from Myxoccocus xanthus in E. coli, leading to the mislocalization of MreB to the poles in large aggregates, or by forcing the localization of major cell wall synthesis proteins to the cell pole. The fact that cells switched modes of growth suggests an evolutionary pathway from midcell to polar growing cells as well as suggests that there might be unknown conditions in nature when cells may switch growth modes.

Strengths:

(1) The authors use a strain that has replaced mreB with a functional fluorescent version at the native site. This eliminates any effects of having two copies of mreB. Because MreB is fluorescently tagged, they can monitor its localization when mreB from M. xanthus is expressed in E. coli. They notice that MreBec now forms bright polar foci and that there appear to be changes to the cell wall at the pole.

(2) D-amino acids are specific to the cell wall, and fluorescent versions (FDAA) have been used to mark sites of new cell wall insertion. The authors use these FDAAs to determine how cell wall synthesis correlates to MreB and if that changes when MreBmx is expressed. Again, there is pretty clear evidence that cell wall synthesis follows MreB localization to the pole.

(3) MreB itself does not synthesize the cell wall, but localizes the proteins, such as PBP2, that do. Using a published method to force proteins to the pole, the authors show that when they target PBP2 to the pole, they can phenocopy the polar growth seen when MreB is polar. Interestingly, these cells become resistant to A22, a drug that targets MreB, suggesting that localized growth at the pole does not require MreB and is sufficient to maintain rod shape.

Weaknesses:

(1) The authors do not show what the poles of control cells look like, making it difficult to determine if there is a change when MreBmx is expressed. However, the localization of both MreBec and MreBmx clearly forms bright foci at the pole.

We appreciate this comment and totally agree with the reviewer. We will provide reference images in the revised manuscript.

(2) While more quantification is needed, the authors show some evidence that RodZ, an MreB interaction partner, is needed for this polar growth, as cells lacking rodZ still form foci at pole-like regions when MreBmx is in the cell; however, these cells remain spherical and do not elongate from these foci.

These results strongly support our conclusions. In the cells that express MreB_Mx, MreB_Mx causes MreB_Ec to mislocalize, and mislocalized MreB_Ec recruits Rod enzymes (RodA and PBP2) to poles through the connector protein RodZ. Here when we delete rodZ, while MreB_Ec still forms aggregates, it is unable to recruit Rod enzymes to those aggregates.

In contrast, when we directly relocalize PBP2 to cell poles through the PopZ tag, polar PG growth can bypass the requirement for MreB_Ec filaments (Fig. 4).

When MreB is deleted, and cells become spherical, the authors were unable to cause the polar growth mode. They suggest that this is due to the lack of a preexisting pole; however, experimental evidence to test this is missing.

We appreciate this comment. We plan to localize PBP2 to cell poles in an mreB depletion strain to test if cells still remain rods when mreB is depleted.

Conclusion:

Overall, the authors do a good job of showing that E. coli can grow with a polar method rather than a midcell method of cell wall insertion. It is unclear why MreBec forms at poles when MreBmx is present and even if this MreB is functional. The foci look similar to inclusion bodies, which are normally aggregates of misfolded proteins that migrate to the poles. Past work has shown that when MreB is more polarly localized, branches form, which is not seen here. Importantly, the authors also show that there is feedback between the localization of MreB and PBP2 as both appear to regulate the localization of the other.

Because CFP-labeled MreB_Ec is still fluorescent in the aggregates, we believe that some MreB_Ec molecules are still correctly folded there, at least on the surfaces of those aggregates.

Reference

Kuru, E., Radkov, A., Meng, X., Egan, A., Alvarez, L., Dowson, A., Booher, G., Breukink, E., Roper, D.I., Cava, F., Vollmer, W., Brun, Y., and VanNieuwenhze, M.S. (2019) Mechanisms of incorporation for D-amino acid probes that target peptidoglycan biosynthesis. ACS Chem Biol.