Flotillin-mediated membrane fluidity controls peptidoglycan synthesis and MreB movement
Abstract
The bacterial plasma membrane is an important cellular compartment. In recent years it has become obvious that protein complexes and lipids are not uniformly distributed within membranes. Current hypotheses suggest that flotillin proteins are required for the formation of complexes of membrane proteins including cell-wall synthetic proteins. We show here that bacterial flotillins are important factors for membrane fluidity homeostasis. Loss of flotillins leads to a decrease in membrane fluidity that in turn leads to alterations in MreB dynamics and, as a consequence, in peptidoglycan synthesis. These alterations are reverted when membrane fluidity is restored by a chemical fluidizer. In vitro, the addition of a flotillin increases membrane fluidity of liposomes. Our data support a model in which flotillins are required for direct control of membrane fluidity rather than for the formation of protein complexes via direct protein-protein interactions.
eLife digest
Every living cell is enclosed by a flexible membrane made of molecules known as phospholipids, which protects the cell from harmful chemicals and other threats. In bacteria and some other organisms, a rigid structure known as the cell wall sits just outside of the membrane and determines the cell’s shape.
There are several proteins in the membrane of bacteria that allow the cell to grow by assembling new pieces of the cell wall. To ensure these proteins expand the cell wall at the right locations, another protein known as MreB moves and organizes them to the appropriate place in the membrane and controls their activity. Previous studies have found that another class of proteins called flotillins are involved in arranging proteins and phospholipid molecules within membranes. Bacteria lacking these proteins do not grow properly and are unable to maintain their normal shape. However, the precise role of the flotillins remained unclear.
Here, Zielińska, Savietto et al. used microscopy approaches to study flotillins in a bacterium known as Bacillus subtilis. The experiments found that, in the presence of flotillins, MreB moved around the membrane more quickly (suggesting it was more active) than when no flotillins were present. Similar results were observed when bacterial cells lacking flotillins were treated with a chemical that made membranes more ‘fluid’ – that is, made it easier for the molecules within the membrane to travel around. Further experiments found that flotillins allowed the phospholipid molecules within an artificial membrane to move around more freely, which increases the fluidity of the membrane.
These findings suggest that flotillins make the membranes of bacterial cells more fluid to help cells expand their walls and perform several other processes. Understanding how bacteria control the components of their membranes will further our understanding of how many currently available antibiotics work and may potentially lead to the design of new antibiotics in the future.
Introduction
The shape of a bacterium is predominantly defined by the structure of its peptidoglycan. Although there is a great variety in bacterial shapes, the overall chemistry of peptidoglycan is very similar between bacteria and thus the shape of peptidoglycan is primarily determined by the temporal and spatial regulation of peptidoglycan synthesis. In rod-shaped bacteria, peptidoglycan synthesis is thought to be mediated by two protein assemblies, the elongasome and the divisome, that synthesise peptidoglycan along the long axis and across the division plane of the cell, respectively (Typas et al., 2012; Zhao et al., 2017). These complexes contain a set of proteins required for the final steps of synthesis and translocation of the peptidoglycan precursor, LipidII, from the inner to the outer leaflet of the cytoplasmic membrane, and proteins that incorporate LipidII into peptidoglycan. These include SEDS (Shape, Elongation, Division and Sporulation) proteins that can perform glycosyl transferase reactions (Cho et al., 2016; Meeske et al., 2016; Taguchi et al., 2019), and Penicillin Binding Proteins (PBPs) that are divided in class A PBPs (aPBPs) that catalyse both glycosyl transferase and transpeptidase reactions, class B PBPs (bPBPs) that only catalyse transpeptidase reactions and low molecular weight PBPs that modify peptidoglycan, as well as hydrolases (Zhao et al., 2017; Morales Angeles and Scheffers, 2017).
Coordination of these complexes is linked to cytoskeletal elements, MreB (-like proteins) for the elongasome and FtsZ for the divisome. In models, the cytoplasmic membrane is often depicted as a passive environment in which these machineries are embedded. However, it is becoming clear that the structure of the membrane plays a critical role in the coordination of peptidoglycan synthesis (Strahl and Errington, 2017). Inward membrane curvature serves as a localisation trigger for MreB and the elongasome, and enhanced local synthesis at bulges straightens out the membrane sufficient to convert spherical cells to a rod shape (Hussain et al., 2018; Ursell et al., 2014). In Bacillus subtilis, the motion of MreB along the membrane is associated with elongasome activity (Domínguez-Escobar et al., 2011; Garner et al., 2011), and the velocity of MreB patches is related to growth rate (Billaudeau et al., 2017), indicating that MreB motion can be used as a marker for elongasome activity. Interestingly, MreB localises to and organises regions of increased membrane fluidity (RIF) (Strahl et al., 2014), which in turn is linked to the presence of LipidII, which favours a more fluid membrane and promotes local membrane disorder (Ganchev et al., 2006; Witzke et al., 2016). Inhibition of LipidII synthesis by genetic or chemical means results in a dissolution of membrane structures observed with the dye FM 4–64 and release of MreB from the membrane (Domínguez-Escobar et al., 2011; Garner et al., 2011; Muchová et al., 2011; Schirner et al., 2015).
Next to RIFs, membrane regions of decreased fluidity have been identified in bacteria (Strahl and Errington, 2017; Bramkamp and Lopez, 2015; Lopez and Koch, 2017). These so-called functional membrane microdomains (FMMs) are thought to be organised by the bacterial flotillin proteins, are enriched in isoprenoid lipids (García-Fernández et al., 2017; López and Kolter, 2010), and can be found in so-called Detergent Resistant Membrane (DRM) fractions of the membrane. Since the formulation of the FMM hypothesis, FMMs have been linked to many processes, such as protein secretion, biofilm formation, competence and cell morphology (Mielich-Süss and Lopez, 2015; Mielich-Süss et al., 2013; Bach and Bramkamp, 2013; Dempwolff et al., 2012). Cell morphology defects are linked to cell wall synthesis, and analysis of the protein content of Bacillus subtilis DRMs identified several PBPs, MreC and other proteins involved in cell wall metabolism as well as the two flotillins, FloA and FloT (López and Kolter, 2010; Bach and Bramkamp, 2013; Yepes et al., 2012). FloA is constitutively expressed, whereas FloT is expressed primarily during stationary growth, cell wall stress and sporulation (Schneider et al., 2015a; Huang et al., 1999; Nicolas et al., 2012). Super resolution microscopy showed that the flotillins and other proteins found in DRMs do not colocalise and have different dynamics (Dempwolff et al., 2016), so it is unlikely that FMMs are regions in the membrane that offer a favourable environment in which these membrane proteins are continuously present and active. Recently, the hypothesis has been put forward that FMMs/flotillins form a platform for the formation of functional protein oligomers, as work in Staphylococcus aureus showed that multimerisation of Type 7 secretion systems and PBP2a depends on FMMs (Lopez and Koch, 2017; García-Fernández et al., 2017; Mielich-Süss et al., 2017).
Here, we have analysed the role of flotillins in peptidoglycan synthesis in B. subtilis. Our results show that, at high growth rates, flotillins control membrane fluidity in a manner that is critical for peptidoglycan synthesis and MreB dynamics, but have no effect on PBP oligomerisation. This results in a new model for flotillin function in the physical organisation of membranes during fast growth.
Results
Absence of flotillins shifts peptidoglycan synthesis to division Septa
In previous studies, a double deletion of floA/floT was either reported to suffer severe shape defects and perturbed membrane structure (Dempwolff et al., 2012), or to not have strong shape defects but with a change in the overall lipid ordering of the membrane (Bach and Bramkamp, 2013). We grew wild type and ΔfloAT strains and analysed exponentially growing cells. We did not observe striking shape defects but did see an increase in median cell length and distribution of cell lengths in the absence of flotillins (Figure 1A,G). To look at effects on peptidoglycan synthesis, we labelled cells with HADA, a fluorescent D-Alanine analogue that reports on sites of active peptidoglycan synthesis (Kuru et al., 2012), and with fluorescent vancomycin (Van-FL), which labels LipidII and peptidoglycan containing pentapeptide side chains (Daniel and Errington, 2003; Morales Angeles et al., 2017). This revealed a significant accumulation of peptidoglycan synthesis stains at division septa in the ΔfloAT strain (Figure 1A–C). To look at membrane structure, cells were labelled with FM4-64, Nile-Red and DiI-C12, which are lipid dyes that accumulate in zones enriched in fluid lipids (Strahl et al., 2014). Again, the stains accumulated at the septa in the ΔfloAT strain, which also showed some accumulation of FM4-64 and DiI-C12 in patches, suggesting that the more fluid regions of the membrane are coalescing into larger regions (Figure 1A,D–F). The HADA, FM4-64 and Nile-Red measurements were repeated using a wild type strain expressing endogenous GFP, allowing simultaneous imaging of both strains on the same slide, and gave similar results, confirming that the observed signal increase is not due to variation between microcopy experiments (Figure 1—figure supplement 1A,B). In this mixed-strain experiment, Nile-Red labelling at the lateral membrane was the same between wild type and ΔfloAT strains, indicating that there is no difference in dye diffusion between the strains (Figure 1—figure supplement 1D). Inspection of the septa by electron microscopy revealed that there was no difference between the thicknesses of the septa between the wild type and ΔfloAT strain, ruling out that the increase in signal was due to formation of thicker septa (Figure 1—figure supplement 1C). The shift of peptidoglycan synthesis to the division site could hint at stress in the overall peptidoglycan synthesis route. This was confirmed by growing cells at a sublethal concentration of fosfomycin, which limits synthesis of LipidII (Kahan et al., 1974), but that does not impact growth rate at the concentration used. This resulted in bulging cells and some lysis, which was exacerbated in the ΔfloAT strain (Figure 1—figure supplement 1F,G). It should be noted that the peptidoglycan synthesis stress caused by fosfomycin is not the same as the stress caused by the absence of flotillins, as the phenotypes of wild type cells with sublethal fosfomycin are quite distinct from ΔfloAT cells without fosfomycin.

Accumulation of peptidoglycan synthesis and membrane material at division sites in a flotillin mutant.
(A) Morphology of the exponentially growing wild type (WT) and ΔfloAT strains labelled with HADA, fluorescent Vancomycin (Van-FL), FM 4–64, Nile Red, and DiI-C12. Scale bar: 5 μm. (B–F) Peak intensity of HADA (B), Van-FL (C), Nile Red (D), FM4-64 (E) and DiI-C12 (F) labelled division sites of the cells shown in (A). Cells from each strain (n ≥ 100, except E, n = 60) were analysed using the ObjectJ macro tool PeakFinder followed by statistical analysis with Prism. Significant differences are based on the two-tailed Mann-Whitney test (*p<0.05; **p<0.01). (G) Distribution of the cell length of the strains analysed in (A). Statistical analysis of the data (n = 100, two tailed Mann-Whitney test, *p<0.05) was performed with Prism, resulting in box plot graphs.
-
Figure 1—source data 1
Fluorescence intensity and cell length measurements.
- https://cdn.elifesciences.org/articles/57179/elife-57179-fig1-data1-v1.xlsx
We ruled out that the peptidoglycan synthesis stress was caused by a change in the folding or complex formation by PBPs in the absence of flotillins, as there were no differences in the overall PBP-profiles of Bocillin-FL labelled wild type or flotillin deletion strains (Figure 1—figure supplement 2A). PBP complex formation was analysed using a combination of Native-PAGE and SDS-PAGE with Bocillin-labelled membrane fractions (Trip and Scheffers, 2016) and showed that various PBPs can be found in a high-MW complex (notably PBPs 1, 2, 3 and 4), but that complex formation is similar in the ΔfloAT strain (Figure 1—figure supplement 2B). Also, none of the five functional GFP-PBPs examined changed their localisation in the ΔfloAT strain (Figure 1—figure supplement 2C). Overall, the data suggest that in the absence of flotillins, peptidoglycan synthesis is affected and relatively increased at division septa, with a concomitant accumulation of membrane dyes that are indicative of higher membrane fluidity.
The absence of both flotillins and PBP1 causes a severe phenotype, linked to a loss of membrane fluidity
We reasoned that a non-lethal defect in septal peptidoglycan synthesis could reveal more about the role of flotillins and constructed a flotillin mutant that lacks PBP1, a bifunctional glycosyl transferase/transpeptidase that is required for efficient cell division (Scheffers and Errington, 2004). Simultaneous deletion of pbp1, floA, and floT resulted in strong filamentation and delocalisation of peptidoglycan synthesis as well as membrane dyes to patches (Figure 2A, Figure 2—figure supplement 1A). Deletion of single flotillin genes and PBP1 had similar, albeit less severe effects (Figure 2—figure supplement 1B,C). To exclude the possibility that an alteration of peptidoglycan modification resulted in the delocalisation of HADA and Van-FL, we used D-Alanine-D-Propargylglycine (D-Ala-D-Pra), a clickable dipeptide analogue which is exclusively incorporated into peptidoglycan via LipidII (Sarkar et al., 2016). D-Ala-D-Pra incorporation was delocalised in the ∆pbp1ΔfloAT strain, indicating that peptidoglycan synthesis itself is delocalised (Figure 2—figure supplement 1D). So far, our experiments were done with fast growing cells and Lysogeny Broth (LB) as the growth medium. Strikingly, none of the mutant strains had an apparent phenotype when cultivated in Spizizen’s minimal medium (SMM, Figure 2B), and peptidoglycan synthesis and lipid dyes were no longer accumulating at division sites in the ΔfloAT strain (Figure 2—figure supplement 2). SMM has a higher Mg2+ concentration, which is known to rescue various cell shape mutations by inhibition of cell wall hydrolysis (Dajkovic et al., 2017). However, the increase in Mg2+ was not sufficient to explain the reversal of phenotype as cells grown on LB supplemented with Mg2+ (6 mM, concentration in SMM, or 20 mM) still displayed the elongated phenotype with delocalised peptidoglycan synthesis (Figure 2—figure supplement 3). This indicated that the phenotypes associated with the absence of flotillins are growth-rate and/or nutrient related.

Cell morphology and cell wall synthesis localisation is dependent on growth conditions.
Morphology of the WT, ΔfloAT, Δpbp1, and ∆pbp1ΔfloAT strains grown in (A) rich (LB), (B) minimal (SMM) medium, and in (C) rich medium with membrane fluidising conditions (0.1% benzyl alcohol, LB+BnOH). Cells were labelled with HADA, and aberrant cell shape and peptidoglycan synthesis are indicated with arrowheads. Panels on the right indicate corresponding cell length distributions (n ≥ 100). Distributions were analysed using Dunn’s multiple comparison tests after Kruskal–Wallis. Statistically significant cell length distribution classes (p<0.001) are represented as letters above each graph – in B and C there were no significant differences. Scale bar: 4 µm.
-
Figure 2—source data 1
Cell length measurements.
- https://cdn.elifesciences.org/articles/57179/elife-57179-fig2-data1-v1.xlsx
Next, we determined lipid packing order in the different strains using the fluorescent dye Laurdan, a reporter for flotillin-mediated lipid ordering (Bach and Bramkamp, 2013). LB-grown cells lacking flotillins displayed an increased generalised polarisation (GP) (Bach and Bramkamp, 2013), indicative of an overall increase in ordered lipid packing in the membrane, but the effect of flotillins on membrane ordering completely disappeared when cells were grown on SMM (Figure 3). The resolution obtained with Laurdan does not allow the detection of local differences in fluidity between the lateral membrane and the septa, but does report on overall lipid ordering. Overall, lipid order was increased in cells grown on SMM compared to LB (Figure 3), whereas the absence of PBP1 had no significant effect on membrane fluidity, also not when combined with flotillin deletions (Figure 3). The changes in lipid ordering were not due to changes in the overall fatty acid composition of the membranes - the ratios of C17/C15 side chains and iso/anteiso fatty acids, which are indicative of fluidity (Strahl et al., 2014), were identical for wild type and ΔfloAT strains grown on LB, and very similar for cells grown on SMM (Figure 3—figure supplement 1).

Flotillins increase overall membrane fluidity at high growth rate.
Changes in overall membrane fluidity were assessed by Laurdan microscopy in cells grown on LB (A), SMM (B) and LB+BnOH (C). Micrographs show colour-coded generalised polarisation (GP) maps in which red indicates regions of decreased fluidity (scale bar: 4 µm). Correspondent theoretical GP measurements in the graphs vary from −1 (more fluid) to 1 (less fluid). Significant statistical differences according to Dunn’s multiple comparison tests after Kruskal–Wallis are represented as letters above each graph in panel (A). Data labelled ‘A’ are significantly different from data labelled ‘B’; data with the same letter are not significantly different. No statistically significant difference was observed for the data in panels (B) and (C) (p<0.001; n ≥ 150, two biological replicates).
-
Figure 3—source data 1
GP measurement.
- https://cdn.elifesciences.org/articles/57179/elife-57179-fig3-data1-v1.xlsx
Restoring membrane fluidity rescues normal peptidoglycan synthesis
The GP values indicated that membranes are more ordered when cells are grown on minimal medium, and this suggests that the flotillin-associated increase in overall membrane fluidity is important for cell shape control at high growth rates. This was tested by growing the strains lacking flotillins and PBP1 on LB in the presence of benzyl alcohol, an extensively used membrane fluidiser that increases membrane hydration due to disordering of membrane structure (Konopásek et al., 2000). Notably, the addition of benzyl alcohol increased membrane fluidity to similar extents in the wildtype and the mutant strains (see Figure 3C), but did not affect the growth rates of the strains (Figure 2—figure supplement 4). The increase in membrane fluidity restored normal cell length and normal peptidoglycan synthesis patterns to the pbp1/floA/floT strain (Figure 2C).
In B. subtilis, the rate of growth and of peptidoglycan synthesis is linked to the speed of MreB movement – in minimal media, the speed of MreB patches is reduced compared to the speed in rich media (Billaudeau et al., 2017). Analysis of the movement of a fully functional mRFPruby-MreB fusion (Domínguez-Escobar et al., 2011) by time lapse TIRF (Total Internal Reflection Fluorescence) microscopy, confirmed that MreB patch mobility is higher in cells grown on LB than in cells grown on SMM, with MreB speeds similar to those reported previously (Billaudeau et al., 2017; Figure 4, Figure 4—videos 1 and 2). Strikingly, in the absence of flotillins, MreB patch mobility was notably decreased in cells grown on LB, while in SMM grown cells MreB patch mobility was independent of the presence of flotillins (Figure 4, Figure 4—videos 3 and 4). Fluidising the membrane with benzyl alcohol, which does not alter the growth rate, almost completely restored MreB mobility in LB grown cells (Figure 4, Figure 4—videos 5 and 6). These results indicate that the MreB patch mobility is not only controlled by growth rate, but also by membrane fluidity. Thus, in fast growing cells with decreased membrane fluidity there is a decrease in elongasome mediated peptidoglycan synthesis, reflected by the reduction of MreB mobility. This fits with an observed increase in peptidoglycan synthesis at the division site which may act as a compensatory mechanism.

MreB speed is linked to membrane fluidity.
(A) The MreB speed in different strain backgrounds and growth conditions was analysed by time-lapse TIRF microscopy. Scatter plot of the speed of patches obtained from individual tracks in 5 different cells are represented per fusion and condition. Average speeds are shown; error bars indicate the standard deviation. Significant statistical differences according to Dunn’s multiple comparison tests after Kruskal–Wallis are represented (p<0.001). (B) Representative kymographs showing fast and slow moving patches of mRFPruby-MreB in B. subtilis cells lacking endogenous mreB (WT) or mreB and floAT (ΔfloAT). See Figure 4—videos 1–6 for corresponding raw image series.
-
Figure 4—source data 1
MreB patch mobility measurements determined by TIRFM.
- https://cdn.elifesciences.org/articles/57179/elife-57179-fig4-data1-v1.xlsx
Flotillin increases fluidity of model membranes in vitro
To assess whether the influence of flotillins on membrane fluidity is direct, we determined the membrane fluidity of model membranes with purified flotillin using solid-state NMR (ssNMR). 2H ssNMR is a biophysical tool that assesses lipid mobility in native-like model membranes on the atomic level, by monitoring the carbon-deuterium order parameter of a deuterated lipid along the acyl chain (here POPC-d31) (Molugu et al., 2017; Legrand et al., 2019). We purified B. subtilis FloT and tested the impact of FloT on the membrane, when reconstituted in POPC-d31 liposomes (Schematically depicted in Figure 5A). FloT decreases the spectral width of the 2H quadrupolar splitting, reflecting an increase in motion on the atomic scale (Figure 5B). The 2H spectrum encodes the local order parameter SCD of the carbon-deuterium in absence and in presence of FloT. Strikingly, FloT has an important impact on the order parameter along the entire acyl chain. It is remarkable that the protein significantly decreases the order parameter SCD, reaching even the inner carbon atoms of the acyl chain, indicating a different packing behaviour and increased membrane fluidity upon interaction with FloT (Figure 5B). The strong fluidising effect described for FloT is notably different from the effects other proteins have when reconstituted into liposomes, such as plant remorins (Legrand et al., 2019) or the membrane binding peptide of the nonreceptor tyrosine kinase Src (Scheidt and Huster, 2009). The anisotropic lineshape of the 31P spectra indicates that the membrane is in the lamellar phase as expected for POPC at the chosen temperature (298K) (Huster, 2014). Upon interactions with FloT the lamellar phase remains intact with formation of a few smaller objects, indicating that the overall liposome structure is not affected and that its phase is maintained (Figure 5—figure supplement 1).

Lipid ordering of FloT probed by 2H solid-state NMR.
(A) Wide-line 2H spectra of POPC-d31 liposomes with or without FloT at a lipid-to-protein molar ratio of 25:1 acquired at 298 K. (B) Effect of FloT on the C-2H order parameters of the PC acyl chain. De-Pake-ing and simulations were applied on the 2H solid-state NMR spectra to determine accurately individual quadrupolar splittings. Order parameters of POPC-d31 acyl chain were derived from experimental quadrupolar splittings and plotted as a function of the labelled carbon position. Insert: schematic depiction of a liposome with added FloT which attaches to the membrane via a hairpin loop (Bach and Bramkamp, 2015).
Discussion
Our data provide evidence that flotillins play a direct role in controlling membrane fluidity and that membrane fluidity is critical for peptidoglycan synthesis at certain growth conditions. In vitro, flotillins enhance the fluidity of a model membrane, and in vivo, the membranes of fast growing flotillin-mutant cells are less fluid even though the fatty acid composition in these cells is identical. Therefore, we propose that the effect of flotillins on membrane fluidity is direct, through a change in the packing behaviour of the lipids resulting in an efficient separation of states of liquid ordered and disordered lipid domains in the membrane bilayer (Bach and Bramkamp, 2013). We found that membrane fluidity is not solely a function of temperature, but also of growth conditions. In vivo, flotillins may also recruit specific, more rigid lipids, such as hopanoids and carotenoids (Bramkamp and Lopez, 2015; García-Fernández et al., 2017; López and Kolter, 2010) which have been found in association with FMMs, and whose synthesis could be growth condition dependent. The predominantly physical role in membrane organisation for flotillins fits with our observation that adding a chemical fluidiser is sufficient to restore MreB dynamics and cell shape to fast growing cells that lack flotillins. We propose that in fast-growing cells on rich medium, flotillin-mediated control of membrane fluidity is critical and sufficient to allow essential membrane bound processes, such as peptidoglycan synthesis, to proceed normally.
A sufficiently fluid membrane is necessary for the efficient recruitment and movement of MreB, and provides a more favourable environment for the peptidoglycan precursor LipidII (Hussain et al., 2018; Ursell et al., 2014; Schirner et al., 2015). It has recently been shown that modulation of either MreBCD or PBP1 levels is sufficient to alter the shape of B. subtilis cells (Dion et al., 2019), underscoring the importance of both systems. In the absence of flotillins, the activity of the MreBCD component is strongly reduced – as evidenced by the reduction of MreB speed – and the overall rigidity of the membrane is increased. This results in a less favourable environment for the peptidoglycan precursor LipidII, which prefers more liquid, disordered membrane phases (Ganchev et al., 2006; Witzke et al., 2016; Calvez et al., 2019). Our data indicate that the reduction in elongasome activity, which does not impact the growth rate itself, is compensated by increased peptidoglycan synthesis activity around division sites in flotillin mutants, which is sufficient to keep the overall cell shape intact, although cells are elongated. The accumulation of lipid dyes indicative of increased fluidity at division sites is in line with a recent study that showed phases of different fluidity in Streptococcus pneumoniae membranes, with more fluid membranes and LipidII localising at midcell (Calvez et al., 2019) where the membrane is most bent. Our findings are also in agreement with the recent observation that B. subtilis cells elongate and lose organisation of MreB when membrane fluidity is decreased by altering the membrane fatty acid composition (Gohrbandt, 2019) (H. Strahl, personal communication). It could very well be that the shift of fluidity towards the septum is only relative as the overall fluidity of the membrane is decreased in the absence of flotillins. This is yet to be determined, as the resolution of Laurdan imaging does not allow conclusive statements about local fluidity changes at the septum. The observation that reduced MreB mobility and therefore altered lateral cell wall synthesis lead to accumulation of Van-FL and HADA staining at the septum is not immediately conclusive. Since septal PG synthesis is MreB independent in B. subtilis, a direct effect of MreB seems unlikely. Rather, a reduction of overall membrane fluidity in a flotillin knock-out might impair LipidII dynamics within the membrane. MurG is the enzyme that catalyses the final step of LipidII synthesis. There are several reports that MurG localises to the septum in different organisms (Aaron et al., 2007; Mohammadi et al., 2007). Thus, it seems likely that the septum is a place of increased LipidII synthesis and a change in membrane fluidity would create problems for LipidII molecules to diffuse away from their insertion site, resulting in reduced lateral PG synthesis and MreB mobility. Alternatively, the reduced MreB mobility and reduced lateral PG synthesis lead to a reduced LipidII consumption at the lateral wall, and the excess LipidII is used by the septal PG synthesis machinery, thereby leading to an increase in midcell PG. It remains to be tested which of these possibilities is responsible for the observed phenotype. Nevertheless, in both cases the increase in cell wall staining at the septum would be indicative of a higher local synthesis activity.
It may seem paradoxical that cells elongate when elongasome activity is reduced, but one has to remember that a large amount of the peptidoglycan synthesis contributing to elongation of bacterial cells is actually taking place at midcell, before the ingrowth of the septum (Aaron et al., 2007; Pazos et al., 2018; Varma and Young, 2009). A relative increase in peptidoglycan synthesis at future division sites makes the activity of PBP1 critical and explains why its deletion has such a dramatic effect in cells lacking flotillins. Restoring fluidity using a chemical fluidiser allows the MreBCD component to again efficiently drive peptidoglycan synthesis during elongation, which is sufficient to suppress the flotillin mutant phenotype. The net effect of this is that cells lacking PBP1 and flotillins grown with benzyl alcohol behave as cells that only lack PBP1, which is quite similar to wild type. At low growth rates, there is no difference between wild type cells and cells lacking flotillins with respect to membrane fluidity, and the speed of MreB is similar between the two cell types. Thus the deletion of flotillins does not exacerbate the phenotype of cells lacking PBP1. The reason for the change in membrane fluidity between cells grown on rich or minimal medium is not yet clear – it does not seem to be caused by a large shift in the fatty acid composition of the membranes. Various factors could play a role, such as the synthesis of specific lipids (hopanoids, isoprenoids) on either type of medium, but also protein crowding, which is higher in membranes of fast-growing cells than in slow-growing cells (Szenk et al., 2017). It will be an important future challenge to establish the cause for this difference. An overall rigidification of the membrane may also lead to retardation of processes which require membrane modifications such as division and sporulation, which is indeed observed in B. subtilis flotillin mutants (Dempwolff et al., 2012; Donovan and Bramkamp, 2009).
One of the proposed roles for flotillin proteins is that they form a ‘platform’ that transiently interacts with membrane proteins that need to oligomerise into functional complexes (Lopez and Koch, 2017). We tested this hypothesis for B. subtilis PBPs by comparing their localisation and oligomerisation in wild type and flotillin mutant strains. Although we were capable of detecting a high MW complex containing various PBPs (notably PBP1, 2a, 2b, 3 and 4), the complex was not dependent on the presence of flotillins. We also note that PBP1 was present in the complex, as well as present in a large smear in the first dimension native gel, which would explain why PBP1 was detected by mass spectrometry analysis of a native PAGE band containing FloA (Schneider et al., 2015a). PBP5, on the other hand, was not part of the high MW complex, which fits with its role in processing of the terminal D-Ala from stem-peptides that have not been cross-linked, which it exercises over the entire surface of the cell (Kuru et al., 2012). Although we cannot exclude that flotillins may affect PBPs that are not easily detected by Bocillin-FL, our results do not provide any evidence for a role for flotillins in the oligomerisation of PBPs in B. subtilis. This extends the finding of the Graumann lab that found either transient or no colocalisation between flotillins and other proteins present in DRM fractions (Dempwolff et al., 2016). Although it is obvious that peptidoglycan synthesis is altered in the absence of flotillins, our data strongly suggest that the basis for this alteration is in the physical organisation of the membrane rather than inefficient formation of divisome or elongasome complexes in the absence of flotillins, because flotillin mutants strains show no synthetic phenotype on minimal medium, and the defects on rich medium can be reverted by chemically fluidising the membrane.
In conclusion, our data provide a new model for flotillin function in the physical organisation of membranes during fast growth. The observation that flotillins differentially affect the membrane in different growth conditions also explains the diversity of phenotypes described for flotillin mutants in the literature.
Materials and methods
Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
---|---|---|---|---|
Strain, strain background (Escherichia coli) | BL21(DE3) | Thermo Fisher Scientific | EC0114 | Chemically competent cells |
Strain, strain background (Bacillus subtilis) | BB001 | Bach and Bramkamp, 2013 | trpC2 yqfA::tet | |
Strain, strain background (Bacillus subtilis) | BB003 | Bach and Bramkamp, 2013 | trpC2 yuaG::pMUTIN4 yqfA::tet | |
Strain, strain background (Bacillus subtilis) | DB003 | Donovan and Bramkamp, 2009 | trpC2 yuaG::pMUTIN4 | |
Strain, strain background (Bacillus subtilis) | RWBS5 | Domínguez-Escobar et al., 2011 | trpC2 amyE::spc Pxyl-mrfpruby-mreB | |
Strain, strain background (Bacillus subtilis) | PS832 | Popham and Setlow, 1995 | Prototrophic revertant of 168 | |
Strain, strain background (Bacillus subtilis) | 2082 | Scheffers et al., 2004 | trpC2 pbpD::cat Pxyl–gfp–pbpD 1–510 | |
Strain, strain background (Bacillus subtilis) | 2083 | Scheffers et al., 2004 | trpC2 ponA::cat Pxyl–gfp–ponA 1–394 | |
Strain, strain background (Bacillus subtilis) | 2085 | Scheffers et al., 2004 | trpC2 dacA::cat Pxyl–gfp–dacA 1–423 | |
Strain, strain background (Bacillus subtilis) | 3105 | Scheffers et al., 2004 | trpC2 pbpC::cat Pxyl-gfp–pbpC 1–768 | |
Strain, strain background (Bacillus subtilis) | 3122 | Scheffers et al., 2004 | trpC2 pbpB::cat Pxyl-gfp-pbpB 1–825 | |
Strain, strain background (Bacillus subtilis) | 3511 | Scheffers and Errington, 2004 | trpC2 ponA::spc | |
Strain, strain background (Bacillus subtilis) | 4042 | Lages et al., 2013 | trpC2 pbpA::cat Pxyl-mkate2-pbpA 1−804 | |
Strain, strain background (Bacillus subtilis) | 4056 | Morales Angeles et al., 2017 | trpC2 dacA::kan | |
Strain, strain background (Bacillus subtilis) | 4059 | This work | trpC2 dacA::cat Pxyl-gfp–dacA 1–423yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4064 | This work | trpC2 dacA::kan yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4090 | This work | trpC2 ponA::spc yuaG::pMUTIN4 | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4091 | This work | PS832 ponA::spc yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4092 | This work | trpC2 ponA::spc yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4095 | This work | trpC2 ponA::cat Pxyl-gfp–ponA 1–394 yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4099 | This work | trpC2 pbpB::cat Pxyl-gfp-pbpB 1–825yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4102 | This work | trpC2 pbpA::cat Pxyl-mkate2-pbpA 1−804 yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4108 | This work | trpC2 pbpD::cat Pxyl-gfp–pbpD 1–510yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4122 | This work | trpC2 pbpC::cat Pxyl-gfp–pbpC 1–768yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4128 | This work | trpC2 ponA::spc pbpD::cat Pxyl-gfp–pbpD 1–510 | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4129 | This work | trpC2 ponA::spc yuaG::pMUTIN4 pbpD::cat Pxyl-gfp–pbpD 1–510 | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4070 | This work | trpC2 mreB::kan amyE::spc Pxyl-mrfpruby-mreB | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4076 | This work | trpC2 mreB::kan amyE::spc Pxyl-mrfpruby-mreB yuaG::pMUTIN4 yqfA::tet | Scheffers lab |
Strain, strain background (Bacillus subtilis) | 4259 | This work; Veening et al., 2009 | trpC2 amyE::PrrnB-gfp | Scheffers lab |
Other | Bocillin | Thermo Fisher Scientific | BOCILLIN FL Penicillin, Sodium Salt | 5 µg/ml |
Other | HADA | Synthesised as described (Morales Angeles et al., 2017) | 7-hydroxycoumarin 3-carboxylic acid-amino-D-alanine | 50 µM |
Other | Vancomycin-FL | Sigma-Aldrich and Molecular Probes (Zhao et al., 2017) | Van-FL | 1:1 mixture of Vancomycin and BODIPYFL Vancomycin (Zhao et al., 2017), final concentration 1 µg/ml |
Other | Nile Red | Thermo Fisher Scientific | 5H-Benzo[α]phenoxazin-5-one, 9-(diethylamino)- 7385-67-3 | 0.5 µg/ml |
Other | 16:0-d31-18:1 PC | Avanti | 860399 | Phospholipids |
Other | Laurdan | Sigma-Aldrich | 6-Dodecanoyl-N,N-dimethyl-2-naphthylamine | - |
Other | Benzyl alcohol | Sigma-Aldrich | Benzyl alcohol | - |
Other | DiI-C12 | Thermo Fisher Scientific | 1,1'-Didodecyl-3,3,3',3'-Tetramethylindocarbocyanine Perchlorate | 2.5 µg/ml |
Other | FM4-64 | Thermo Fischer Scientific | (N-(3-Triethylammoniumpropyl)−4-(6-(4-(Diethylamino) Phenyl) Hexatrienyl) Pyridinium Dibromide) | 0.5 µg/ml, Invitrogen FM 4–64 Dye |
Software, algorithm | Prism 5 | 1992–2010 GraphPad Software | RRID:SCR_002798 | - |
Software, algortithm | ImageJ 1.52p/FIJI | Wayne Rasband – National Institutes of Health, USA | RRID:SCR_002285 | Free software |
Software, algorithm | SPSS | SPSS | RRID:SCR_002865 | software |
Software, algorithm | NMR Depaker 1.0rc1 software | [Copyright (C) 2009 Sébastien Buchoux] | software | |
Software, algorithm | Bruker Topspin 3.2 software | Bruker | RRID:SCR_014227 | software |
B. subtilis strains and growth conditions
Request a detailed protocolAll B. subtilis strains used in this study are derived from strain 168 and are listed in the Key resources table. Construction of new strains was based on natural competence of B. subtilis (Harwood and Cutting, 1990). Gene integration or deletion was validated by colony PCR whereas the expression and localisation of the fluorescent fusions was additionally validated by microscopy. Cells were grown either in LB Lennox (5 g/L yeast extract; 5 g/L NaCl; 10 g/L tryptone) (Lennox, 1955) or Spizizen minimal medium (SMM) (Anagnostopoulos and Spizizen, 1961) supplemented with 1% glucose, at 37°C and 200 rpm, unless indicated otherwise. Induction of the Pxyl promoter was triggered by addition of 0.2–0.5% xylose. Cell cultures were supplemented with spectinomycin (50 µg/ml), tetracycline (10 µg/ml), chloramphenicol (5 µg/ml), kanamycin (5 µg/ml), erythromycin (1 µg/ml), benzyl alcohol (BnOH, 0.1%) or magnesium sulphate (MgSO4, 6–20 mM) when necessary.
Growth curves
Request a detailed protocolGrowth experiments were performed either manually or automatically with a PowerWave 340 microplate reader (BioTek Instruments, U.S.A). Strains were pre-cultured overnight in 3 ml LB or SMM medium at 37°C with shaking at 200 rpm. Next, stationary or late-exponentially cells were diluted with fresh LB or SMM medium (supplemented when necessary), and cell densities (OD600) were measured every 1 hr when monitored manually or every 10 min when monitored automatically, for a total time of 7–22 hr.
Fluorescence microscopy
Request a detailed protocolFor standard fluorescence microscopy, exponentially growing cells were immobilised on microscope slides covered with a thin film of 1% agarose (w/v) in water or the appropriate medium. For TIRFM, agarose pads were mounted using Gene Frames (1.7 × 2.8 cm chamber, 0.25 mm thickness, 125 µL volume) from ThermoScientific. Standard fluorescence microscopy was carried out using an Axio Zeiss Imager M1 fluorescence microscope (EC Plan-Neofluar 100x/1.30 Oil Ph3 objective) equipped with an AxioCam HRm camera and an Nikon-Ti-E microscope (Nikon Instruments, Tokyo, Japan) equipped with Hamamatsu Orca Flash 4.0 camera.
For Laurdan and TIRFM experiments, a Delta Vision Elite microscope (Applied Precision, GE Healthcare) equipped with an Insight SSI Illumination, an X4 Laser module, a CoolSnap HQ (Zhao et al., 2017) CCD camera and a temperature-controlled chamber set up at 37 ⁰C was used. Laurdan images were taken with an Olympus UplanSApo 100x/1.4 oil objective. TIRFM image series were taken using an Olympus UAPO N 100X/1.49 TIRF objective and a 561 nm laser (50 mW, 100% power). Data processing was performed with softWoRx Suite 2.0 Software.
Visualisation of cell wall synthesis
Request a detailed protocolPeptidoglycan (PG) synthesis was assessed by labelling the cells with HADA (7-hydroxycoumarin 3-carboxylic acid-amino-D-alanine) (Kuru et al., 2015), Van-FL (Daniel and Errington, 2003) or D-Ala-D-Pra (Sarkar et al., 2016).
HADA: synthesised as described (Morales Angeles et al., 2017). Overnight cultures of B. subtilis strains were diluted 1:100 into fresh LB medium or LB medium supplemented with 0.1% (w/v) of benzyl alcohol (BnOH), a membrane fluidiser. Cells were grown until exponential phase, a sample of 1 ml of culture was spun down for 30 s, 5000 × g and the cell pellet was resuspended in 25 µl of fresh pre-warmed LB or LB containing 0.1% (w/v) BnOH. HADA was added to a 50 µM final concentration. Cells were incubated for 10 min in the dark (37 ⁰C, 200 rpm) and then washed twice in 1 ml PBS buffer (58 mM Na2HPO4, 17 mM NaH2PO4, 68 mM NaCl, pH 7.3) to remove the excess of unbounded HADA. Cells were spun down again and resuspended in 25 µl of the appropriate medium and 2 µl of cells were mounted on 1% agarose slides before visualisation. Visualisation of HADA patterns (excitation: 358 nm/emission: 461 nm) under fluorescence microscopy from two biological replicates and cell length measurements were taken from at least 100 cells each strain/treatment.
Van-FL: a 1:1 mixture of vancomycin (Sigma Aldrich) and BODIPYFL Vancomycin (Molecular Probes) at a final concentration 1 µg/ml was used to label cells for 5–10 min at room temperature.
D-Ala-D-Pra: synthesised as described (Sarkar et al., 2016). 1 ml of cell cultures was pelleted and resuspended in 50 µl of PBS buffer. Dipeptide was added to a final concentration of 0.5 mM following with 5 min incubation at room temperature. Cells were fixed by adding 70% ethanol and incubated for minimum 2 hr in −20°C. Next, cells were washed twice with PBS in order to remove unattached peptides, and resuspended in 50 µl of PBS. The D-Ala-D-Pra was subsequently labelled via a click reaction with fluorescent azide (20 μM) that was incubated for 15 min at room temperature with addition of copper sulphate (CuSO4, 1 mM), tris-hydroxypropyltriazolylmethylamine (THPTA, 125 μM) and ascorbic acid (1.2 mM). The sample was washed twice with PBS and resuspended in 50 µl of the same buffer.
Laurdan staining and GP calculations
Request a detailed protocolLaurdan (6-Dodecanoyl-N, N-dymethyl2-naphthylamine, Sigma-Aldrich) was used to detect the liquid ordering in the membrane, as decribed (Bach and Bramkamp, 2013), with modifications. Cells were grown in LB or SMM medium until late exponential phase. Laurdan, dissolved in dimethyformamide (DMF), was added at 10 µM final concentration and cells were incubated for 10 min in the dark at 37 °C, 200 rpm. Cells were then washed twice in PBS buffer supplemented with 0.2% (w/v) glucose and 1% (w/v) DMF, and resuspended in fresh prewarmed appropriate medium. Laurdan was excited at 360 ± 20 nm, and fluorescence emission was captured at 460 ± 25 nm (exposure time: 500 ms) and at 535 ± 25 nm (exposure time: 1 s) (Strahl et al., 2014). The image analysis including the generation of GP maps was carried out using Fiji Software (Schindelin et al., 2012) in combination with the macro tool CalculateGP designed by Norbert Vischer (http://sils.fnwi.uva.nl/bcb/objectj/examples/CalculateGP/MD/gp.html). The GP values were measured for at least 100 individual cells after background subtraction, from two biological replicates.
Other fluorescent membrane probes
Request a detailed protocolB. subtilis cell membranes were probed with Nile Red (0.5 µg/ml), FM4-64 (0.5 µg/ml) or DiI-C12 (2.5 µg/ml). To this end, an overnight culture was grown in appropriate antibiotic, diluted 1:100 in LB medium supplemented with DiI-C12 followed by growth until exponential phase. Membranes were probed with Nile Red or FM4-64 for 5 min at room temperature after reaching exponential phase. The stained cells were washed three times in prewarmed LB medium supplemented with 1% DMSO before visualisation under fluorescence microscopy.
TIRF time lapse microscopy
Request a detailed protocolTime-lapse TIRFM movies were taken in two independent experiments for each strain and condition. To this end, overnight cultures of strains grown in LB medium supplemented with the appropriate antibiotic were diluted 1:100 in medium containing 0.5% (w/v) xylose and grown until exponential phase. All experiments were performed inside the incubation chamber set to 37 °C, no longer than 10 min after taking the sample. The cells were imaged over 30 s with 1 s inter-frame intervals in a continuous illumination and ultimate focus correction mode. The single particle tracking analyses and kymographs were done using Fiji Software (Schindelin et al., 2012) in combination with the MTrackJ (Meijering et al., 2012) and MicrobeJ plugins (Ducret et al., 2016).
Bocillin labelling
Request a detailed protocolCells were grown until an OD600 of 0.4–0.5, and washed twice with PBS. Next, samples were resuspended in 50 μl PBS containing Bocillin-FL (5 µg/ml) and incubated at room temperature for 10 min. Subsequently cells were harvested, lysed by sonication and cell-free extracts were prepared. Samples, equalised for culture OD, were prepared with SDS-PAGE sample buffer and run on a 12% SDS-PAGE gel. Fluorescent bands were visualised using a Typhoon Trio (GE Healthcare) scanner.
Isolation of membranes
Request a detailed protocolMembrane isolation was adapted from Schneider et al., 2015b. Briefly, cells were grown until an OD600 of 0.4–0.5, cell fractions were collected and resuspended in PBS with Lysozyme (1 µg/ml), EDTA (5 mM), 1/10 tablet cOmplete protease inhibitor (Roche), and DNAse (5 µg/ml) and incubated for 30 min on ice. Samples were sonicated, cell which did not lyze were spun down (8000 rpm, 2 min, 4°C), and the supernatant fraction was centrifuged at 4°C and 40000 rpm for 1 hr. The membrane pellet was dissolved in ACA750 buffer (750 mM aminocaproic acid, 50 mM Bis-Tris, pH 7.0) to a final protein concentration of 1 μg/μl. Membranes were solubilised overnight at 4°C in 1% (w/v) dodecylmaltoside (DDM) and either used directly or stored at −20°C.
Blue native PAGE (BN-PAGE)
Request a detailed protocolThe experiment was performed as described (Trip and Scheffers, 2016). Samples were prepared by mixing sample buffer (0.1% Ponceau S, 42.5% Glycerol) with solubilised membranes in a 1:3 ratio. Samples were resolved on a mini-PROTEAN TGX Stain-Free gradient gel (4–15%, BioRad) using cathode (50 mM Tricine and 15 mM BisTris), and anode (50 mM BisTris pH 7.0) buffers. The Novex NativeMark Unstained Protein Standard marker was used as a Mw marker.
Second dimension SDS PAGE (2D SDS-PAGE)
Request a detailed protocolA lane of interest was excised from the Native-PAGE gel and immobilised horizontally on top of a SDS-PAGE gel (5% stacking, 12% resolving). The excised fragment was flanked with a piece of Whatman paper soaked with PageRuler Prestained Protein Ladder. The gel fragment to be resolved in the second dimension was topped with a mix of 1% (w/v) LowTemperature agarose, 0.5% (w/v) SDS and bromophenol blue. After the agarose had solidified, standard SDS-PAGE electrophoresis was performed.
TEM
Request a detailed protocolCultures were harvested by centrifugation and a small amount of pellet was placed on a copper dish. A 400 copper mesh grid and a 75 µm aperture grid was placed on top of the cells to create a thin layer. The sandwiched cells were plunged rapidly into liquid propane. Sandwiches were then disassembled and placed on frozen freeze-substitution medium containing 1% osmium tetroxide, 0.5% uranyl acetate and 5% water in acetone. Cells were dehydrated and fixed using the rapid freeze substitution method (McDonald, 2014). Samples were embedded in epon and ultrathin sections were collected on formvar coated and carbon evaporated copper grids and inspected using a CM12 (Philips) transmission electron microscope. For each strain 70 random septa were imaged with pixel resolution of 1.2 nm. Using ImageJ the cell wall thickness for each septum was measured at 4 places from which the average was taken.
Statistical analysis
Request a detailed protocolEach set of micrographs to be analysed was imaged with the same exposure time. For the septum intensity analysis of HADA, FM4-64 and Nile Red, the wildtype strain (expressing GFP) and the ΔfloAT strains were mixed, labelled and imaged on the same agarose pad. Intensity of the fluorescently labelled septa was measured using the ObjectJ macro tool PeakFinder (https://sils.fnwi.uva.nl/bcb/objectj/examples/PeakFinder/peakfinder.html) (Vischer et al., 2015). A perpendicular line was drawn across the septal plane, the background intensity was removed resulting in a maximum peak intensity. The number of septa compared was indicated for every individual experiment. Populations were compared using the non-parametric Mann-Whitney test. The null hypothesis was tested with the p value of 0.05. The statistical analyses and their graphical representation (box plots) were generated with GraphPad Prism 8.1 (San Diego, California, USA). Box plots show the median and the interquartile range (box), the 5th and 95th percentile (whiskers). Laurdan fluorescence generalised polarisation, cell length and MreB speed statistical analyses were performed using Kruskal-Wallis with Dunn's multiple comparison post-hoc test.
Fatty acid composition analysis
Request a detailed protocolThe fatty acid composition of B. subtilis wild-type cells and the flotillin/PBP mutants was analysed with gas chromatography as fatty acid methyl esters. Cells for the analyses were grown at 37°C in LB or SMM until mid-exponential (OD600 ~0.5), harvested (6000 rpm, 10 min, 4°C) and washed with 100 mM NaCl. Next, the cells were freeze dried at −50°C, 0.012 mbar for a minimum of 18 hr. All analyses were carried out on biological duplicates by the Identification Service of the DSMZ, Braunschweig, Germany.
Sample preparation for solid-state NMR
Request a detailed protocolFloT was essentialy purified as described (Bach and Bramkamp, 2013), in solubilised form, and stored in buffer A (50 mM Tris HCl pH 7.5, 150 mM NaCl, 5 mM MgCl2) supplemented with 0.05% Triton X-100.
Liposomes containing POPC-d31 were prepared by mixing appropriate lipid powders in organic solvents (chloroform/methanol, 2:1 ratio). Solvents were evaporated under a flow of N2 to obtain a thin lipid film. Lipids were rehydrated with ultrapure water before lyophilisation over night. The lipid powder was hydrated with an appropriate amount of buffer A with 10% glycerol and homogenised by three cycles of vortexing, freezing (liquid nitrogen, −196°C, 1 min) and thawing (40°C in a water bath, 10 min). This protocol generated a milky suspension of micrometer-sized multilamellar vesicles. FloT was solubilised in Buffer A supplemented with 0.05% Triton X-100 and added to preformed liposomes and incubated for 1 hr at room temperature. A dialysis step was then performed against Buffer A at 4°C under agitation to remove the detergent. Samples were centrifuged at 100,000 g at 4°C for 1 hr to pellet the proteoliposomes. 2H solid-state NMR spectra were recorded of liposomes in the presence or absence of FloT at a lipid/protein ratio of 25:1 at 298 K.
Solid-state NMR
Request a detailed protocol2H NMR spectroscopy experiments were performed using a Bruker Avance III 500 MHz WB (11.75 T) spectrometer. They were recorded on 2H-labelled POPC at 76.77 MHz with a phase-cycled quadrupolar echo pulse sequence (90°x-t-90°y-t-acq). Acquisition parameters were as follows: spectral window of 500 kHz for 2H NMR spectroscopy, p/2 pulse width of 3.90 ms for 2H, interpulse delays (t) were of 40 µs, recycled delays of 1.3 s for 2H; 3000 and 8000 scans were used for 2H NMR spectroscopy on liposomes and liposomes with FloT, respectively. Spectra were processed using a Lorentzian line broadening of 300 Hz for 2H NMR spectra before Fourier transformation from the top of the echo. Samples were equilibrated for 30 min at a given temperature before data acquisition. All spectra were processed and analysed using Bruker Topspin 3.2 software. Spectral moments were calculated for each temperature using the NMR Depaker 1.0rc1 software [Copyright (C) 2009 Sébastien Buchoux]. Orientational order parameters (SCD) were calculated from experimental quadrupolar splittings (DnQ) as described in Huster, 2014. For 31P ssNMR, we applied a static Hahn spin echo sequence at the 31P frequency of 162 MHz on a 400 MHz (9.4T) Bruker Avance III HD spectrometer, with a 90◦ pulse of 8 μs, a delay of 40 μs, a recycle delay of 5 s, a spectral window of 400 ppm and a number of scans of 4000 and 3400 was used on liposomes and liposomes with FloT, respectively. Spectra were processed using a Lorentzian line broadening of 100 Hz.
Data availability
All data generated or analysed during this study are included in the manuscript and supporting files.
References
-
Requirements for transformation in Bacillus subtilisJournal of Bacteriology 81:741–746.https://doi.org/10.1128/JB.81.5.741-746.1961
-
Flotillins functionally organize the bacterial membraneMolecular Microbiology 88:1205–1217.https://doi.org/10.1111/mmi.12252
-
Contrasting mechanisms of growth in two model rod-shaped BacteriaNature Communications 8:15370.https://doi.org/10.1038/ncomms15370
-
Exploring the existence of lipid rafts in BacteriaMicrobiology and Molecular Biology Reviews 79:81–100.https://doi.org/10.1128/MMBR.00036-14
-
Lipid phases and cell geometry during the cell cycle of Streptococcus pneumoniaeFrontiers in Microbiology 10:351.https://doi.org/10.3389/fmicb.2019.00351
-
Super resolution fluorescence microscopy and tracking of bacterial flotillin (Reggie) Paralogs provide evidence for Defined-Sized protein microdomains within the bacterial membrane but absence of clusters containing Detergent-Resistant proteinsPLOS Genetics 12:e1006116.https://doi.org/10.1371/journal.pgen.1006116
-
Solid-state NMR spectroscopy to study protein–lipid interactionsBiochimica Et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids 1841:1146–1160.https://doi.org/10.1016/j.bbalip.2013.12.002
-
The mechanism of action of fosfomycin (phosphonomycin)Annals of the New York Academy of Sciences 235:364–386.https://doi.org/10.1111/j.1749-6632.1974.tb43277.x
-
Cold shock in Bacillus subtilis: different effects of benzyl alcohol and ethanol on the membrane organisation and cell adaptationBiochimica Et Biophysica Acta (BBA) - Biomembranes 1464:18–26.https://doi.org/10.1016/S0005-2736(99)00240-0
-
In Situ probing of newly synthesized peptidoglycan in live bacteria with fluorescent D-amino acidsAngewandte Chemie International Edition 51:12519–12523.https://doi.org/10.1002/anie.201206749
-
Nanodomain clustering of the plant protein remorin by Solid-State NMRFrontiers in Molecular Biosciences 6:00107.https://doi.org/10.3389/fmolb.2019.00107
-
Exploring functional membrane microdomains in Bacteria: an overviewCurrent Opinion in Microbiology 36:76–84.https://doi.org/10.1016/j.mib.2017.02.001
-
Functional microdomains in bacterial membranesGenes & Development 24:1893–1902.https://doi.org/10.1101/gad.1945010
-
Methods for cell and particle trackingMethods in Enzymology 504:183–200.https://doi.org/10.1016/B978-0-12-391857-4.00009-4
-
Molecular mechanisms involved in Bacillus subtilis biofilm formationEnvironmental Microbiology 17:555–565.https://doi.org/10.1111/1462-2920.12527
-
Concepts and methods of Solid-State NMR spectroscopy applied to biomembranesChemical Reviews 117:12087–12132.https://doi.org/10.1021/acs.chemrev.6b00619
-
Pentapeptide-rich peptidoglycan at the Bacillus subtilis cell-division siteMolecular Microbiology 104:319–333.https://doi.org/10.1111/mmi.13629
-
BookBacillus: Cellular and Molecular BiologyFreiburg: Caister Academic Press.https://doi.org/10.21775/9781910190579
-
Changes of lipid domains in Bacillus subtilis cells with disrupted cell wall peptidoglycanFEMS Microbiology Letters 325:92–98.https://doi.org/10.1111/j.1574-6968.2011.02417.x
-
In vivo probe of lipid II-Interacting ProteinsAngewandte Chemie International Edition 55:8401–8404.https://doi.org/10.1002/anie.201603441
-
PBP1 is a component of the Bacillus subtilis cell division machineryJournal of Bacteriology 186:5153–5156.https://doi.org/10.1128/JB.186.15.5153-5156.2004
-
Fiji: an open-source platform for biological-image analysisNature Methods 9:676–682.https://doi.org/10.1038/nmeth.2019
-
Lipid-linked cell wall precursors regulate membrane association of bacterial actin MreBNature Chemical Biology 11:38–45.https://doi.org/10.1038/nchembio.1689
-
The actin homologue MreB organizes the bacterial cell membraneNature Communications 5:3442.https://doi.org/10.1038/ncomms4442
-
Bacterial membranes: structure, domains, and functionAnnual Review of Microbiology 71:519–538.https://doi.org/10.1146/annurev-micro-102215-095630
-
From the regulation of peptidoglycan synthesis to bacterial growth and morphologyNature Reviews Microbiology 10:123–136.https://doi.org/10.1038/nrmicro2677
-
A mechanism for cell cycle regulation of sporulation initiation in Bacillus subtilisGenes & Development 23:1959–1970.https://doi.org/10.1101/gad.528209
-
Don't let sleeping dogmas lie: new views of peptidoglycan synthesis and its regulationMolecular Microbiology 106:847–860.https://doi.org/10.1111/mmi.13853
Decision letter
-
Tâm MignotReviewing Editor; CNRS-Aix Marseille University, France
-
Gisela StorzSenior Editor; National Institute of Child Health and Human Development, United States
-
Leendert HamoenReviewer; University of Amsterdam, Netherlands
In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.
Acceptance summary:
In bacteria, Flotillins organize so-called micro-domains in the bacterial membrane that have been proposed to act as protein scaffolds for the local recruitment of protein complexes, including cell wall synthetic proteins. In this work, Zielinska et al. provide evidence for an alternative mechanism whereby Flotillins instead function to regulate fluidity of the membrane, indirectly affecting cytoskeletal dynamics and thus cell wall synthesis. In the future, it will be important to determine whether the proposed mechanisms are mutually exclusive or whether Flotillins indeed perform both functions.
Decision letter after peer review:
[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]
Thank you for submitting your work entitled "Membrane fluidity controls peptidoglycan synthesis and MreB movement" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.
Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered for publication in eLife at this stage.
While the reviewers agree that the work conveys a potentially interesting message with regards to a new model of Flotillin function, they both indicate that the results are yet too preliminary to fully disprove the existing hypothesis that Flotillins form a physical platform for the recruitment of cell wall synthetic complexes. Specifically, the reviewers insist on two key points that need further investigations to strengthen the conclusions:
i) The quantification of Peptidoglycan and/or precursor accumulation at the septa is not supported by the EM images and Laurdan staining experiments. Comparison of fluorescence intensities between conditions is particularly delicate, requiring very precise microscope calibrations and possibly imaging of distinct cell genotypes in the same field of view. The conclusions that PG accumulation and membrane fluidity increase at septa in the floAt mutant is central to the authors' hypothesis and should be verified by a complementary approach.
ii) The benzyl alcohol rescue experiment is currently the main piece of evidence that challenges the previous model, but is it sufficient? Effects other than restoring membrane fluidity alone could explain growth restoration.
The policy at eLife is to reject manuscripts that cannot be revised within a three-months’ time period. If you think that you can address the comments in full and given the level of interest, we would be inclined to reconsider a profoundly revised manuscript, note however that it would be processed as a new submission and would have to go through another round of review.
Reviewer #1:
Zielinska and co-workers show here that flotillins affect overall membrane fluidity in Bacillus subtilis and that this explains their effect on the activity of MreB and peptidoglycan synthesis. They discovered this phenomenon by observing differences in cell wall synthesis at division sites and related differences in lipid staining in a flotillin deletion mutant. An important discovery was that they could exacerbate the phenotype by combining it with a deletion in the non-essential peptidoglycan synthase PBP1. What completed their experiments was the fact that they could restore the phenotype by fluidizing the cell membrane with an alcohol.
This is an interesting paper that highlights the importance of membrane fluidity, the role of bacterial flotillins in this, and the fact that this physiological characteristic changes under different growth conditions and can be adjusted by specific membrane proteins.
I agree with their overall conclusion but some conclusions seem to me incorrect (outlined below). The paper is clearly written, and experiments generally well performed. However, the microscopic figures are far too small. Even when displayed on a large screen many cellular details are not visible, e.g. Laurdan staining that does not show any clear septa. Firstly, a paper should be readable when printed, secondly, microscopic images should be clear so that phenotypic differences are clearly observable. This is currently not the case and should be improved.
I do not agree with the conclusion in the Discussion (second paragraph) that membrane fluidity and peptidoglycan synthesis is higher at septa in order to compensate the absence of flotillins. This reasoning does not make sense because what should be compensated for? The overall fluidity still changes. It is also a bit difficult how this can be compensated, in other words, how can the fluidity be increased at septa? Moreover, if there would be more peptidoglycan synthesis, then cells would divide faster, thus be shorter, since the authors show that septa do not become wider (Supplementary Figure 1). But a floAT deletion strain is not shorter (Figure 2). This discrepancy might be related to the possibility that the authors have not measured the septal intensities in Figure 1 correctly. That is, they should have compared the intensities of septa with that of the lateral wall, because it might well be that a floAT deletion strain has an overall more permeable cell wall that facilitate fluorophore diffusion. In fact the laurdan images in Figure 3 even suggest this. Maybe the authors measured the relative increase in intensity at septa, but I was unable to determine this from their experimental description. Moreover, and importantly, their laurdan images show no evidence for an increase of fluidity in septa. This seems to contradict their own argumentation.
Reviewer #2:
The experiments are well done and for the most part convincing. The problem is that their conclusions are overstated. The claim that all effects are explained by a direct action of flotillins on membrane fluidity rather than primarily affecting the assembly or activity of a PBP is not supported by the data. In fact most of the experiments (except potentially the benzyl alcohol addition, see below) can be explained by the floAT double mutant affecting primary septal PG synthesis or turnover. This does not mean that the manuscript is not suited for eLife, the data should just be interpreted differently.
1) One could state that all of the observed effects are due to floAT control of a septal PBP or septal PG remodeling. First, the claim that there is more lipidII/PG precursors in Figure 1 in the float mutants is based on a small difference and is weak. The signal is barely above background and HADA does not mark sites of PG synthesis, rather it marks sites of PG remodeling/turnover, as it depends on the exchange of the labeled D-ala in lipid II on the extracytoplasmic side. Moreover, the fact that floAT cells depend strongly on PBP1 for cell division, provides genetic evidence that the double mutant is under septal PG stress, which would probably be observable by applying PG stress in the cytoplasm (with Fosfomycin for example) and extracellularly with β-lactams or glycopeptides or bacitracin or A22.
If PBP biosynthetic activity is reduced (or PG degradation increased) in the floAT mutants, would this not lead to a (slight) accumulation of lipid II at the septum as well, and this could explain the observation in Figure 1? Also, this effect on PG stress could easily explain the reduction in MreB movement. Finally, my understanding was that an increase in lipidII would INCREASE membrane fluidity, but they show in Figure 3 that fluidity is reduced in the floAT mutant.
2) Growth rate effect: I disagree that the difference between LB and SMM can be pinned on growth rate. The cellular metabolism is completely different in minimal vs. rich medium so how can the authors rule out that a change in biosynthetic pathways does not affect lipid composition, PG (bactoprenol or lipidII) synthesis, isoprenoids or simply certain membrane proteins, all of which could alter membrane fluidity. The growth rate argument does not hold here, and it could only be proven if the authors stick to LB and lower growth rate there, ideally by reducing temp, but there could be complications on fluidity arising from a temperature shift. Inducing ppGpp would also be possibility, but this could also reprogram cell metabolism and make the interpretation ambiguous. Perhaps use diluted LB?
3) The benzyl alcohol addition experiment is the best (and in my opinion only) argument for their claim on membrane fluidity, but how do we know that this compound directly affects the membrane, rather than acting indirectly, like SMM, changing metabolism? How quick are the changes observed after addition of BA? Are the changes in fluidity and MreB localization still observed when cells are pre-treated with protein synthesis inhibitors and /or cerulenin (FA inhibitor)? By the way, are FRAP measurements of membrane-tethered GFP consistent with a change in fluidity?
The bottom line is: I'm not a fan of the previous flotillin model either, but I don't think the authors provide compelling evidence in favor of their model. In fact, their experimental data does not completely exclude the previous model about flotillins facilitating PG assembly/turnover machines either.
[Editors’ note: further revisions were suggested prior to acceptance, as described below.]
Thank you for submitting your article "Flotillin mediated membrane fluidity controls peptidoglycan synthesis and MreB movement" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Gisela Storz as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Leendert Hamoen (Reviewer #2).
The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.
We would like to draw your attention to changes in our revision policy that we have made in response to COVID-19 (https://elifesciences.org/articles/57162). Specifically, we are asking editors to accept without delay manuscripts, like yours, that they judge can stand as eLife papers without additional data, even if they feel that they would make the manuscript stronger. Thus the revisions requested below only address clarity and presentation.
Overall the study is solid, convincing, well presented and certainly suitable for publication in eLife. Nevertheless, some clarifications should be provided before the manuscript is accepted for publication.
1) It is not clear why a reduction in MreB mobility in the membrane of floAT cells grown in LB would lead to increased vancomycin-FL and HADA staining at the septum. This relationship is poorly accounted for in the manuscript. MreB is not at the septum, so why would lower mobility of MreB cause an increase in septal lipid II? This observation is thus circumstantial. MreB and septal lipid II could be both affected because of reduced fluidity, but those events do not have to be linked:
– Perhaps membranes are less fluid and if lipid II is inserted at the septum, it is less likely to diffuse away in floAT cells because of the reduced fluidity?
– Or there is just a preference for lipid II to accumulate there because the membrane is curved and fluidity at the curvature is affected in the absence of FloAT? If that's the case, what does the additional deletion of pbp1 really tell us (although it is interesting)?
2) To clarify these issues, these questions should be discussed and it could be easier for the reader if the order of the results were turned around and tell the story more linearly. For example, rather than going from PG studies to membrane fluidity studies, the manuscript could present the established part first (NMR, fluidity, MreB) and then argue that reduced MreB mobility prompted imaging with HADA, Vanco-FL and finally check the pbp1floAT triple mutant. The current read takes a circular path (starting with PG and ending with MreB) and again there is no compelling causality between MreB movement and septal PG presented here.
https://doi.org/10.7554/eLife.57179.sa1Author response
[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]
While the reviewers agree that the work conveys a potentially interesting message with regards to a new model of Flotillin function, they both indicate that the results are yet too preliminary to fully disprove the existing hypothesis that Flotillins form a physical platform for the recruitment of cell wall synthetic complexes. Specifically, the reviewers insist on two key points that need further investigations to strengthen the conclusions:
i) The quantification of Peptidoglycan and/or precursor accumulation at the septa is not supported by the EM images and Laurdan staining experiments. Comparison of fluorescence intensities between conditions is particularly delicate, requiring very precise microscope calibrations and possibly imaging of distinct cell genotypes in the same field of view. The conclusions that PG accumulation and membrane fluidity increase at septa in the floAt mutant is central to the authors' hypothesis and should be verified by a complementary approach.
We have carried out an additional series of microscopy experiments in which the distinct genotypes were imaged in the same field of view using two lipid dyes and one peptidoglycan synthesis dye. These results (more detail below) are completely in line with our initial findings and have now been included in the manuscript.
We would also like to note that the EM imaging and Laurdan staining are not suitable for the detection of precursor accumulation (or absence thereof) – and we make no statements that would suggest this. EM is used to measure septum thickness as a thicker septum might result in accumulation, Laurdan is used to determine overall membrane fluidity in cells and does not have the resolving power in our setup to say anything about differences at the septum.
We would like to emphasize that the main conclusion of this paper is the function of flotillins in fluidizing the membrane and the resulting effect on membrane coupled processes of which we use cell wall synthesis as one example.
ii) The benzyl alcohol rescue experiment is currently the main piece of evidence that challenges the previous model, but is it sufficient? Effects other than restoring membrane fluidity alone could explain growth restoration.
We fully agree that additional proof was necessary to prove the exact effects of flotillins on membranes. We have now included a biophysical experiment using state of the art solid state NMR that shows that the flotillin FloT alone is sufficient to fluidize membranes in model liposomes. This completely different approach provides an additional line of evidence that flotillins on their own are required and sufficient to fluidize the membrane.
Reviewer #1:
Zielinska and co-workers show here that flotillins affect overall membrane fluidity in Bacillus subtilis and that this explains their effect on the activity of MreB and peptidoglycan synthesis. They discovered this phenomenon by observing differences in cell wall synthesis at division sites and related differences in lipid staining in a flotillin deletion mutant. An important discovery was that they could exacerbate the phenotype by combining it with a deletion in the non-essential peptidoglycan synthase PBP1. What completed their experiments was the fact that they could restore the phenotype by fluidizing the cell membrane with an alcohol.
This is an interesting paper that highlights the importance of membrane fluidity, the role of bacterial flotillins in this, and the fact that this physiological characteristic changes under different growth conditions and can be adjusted by specific membrane proteins.
I agree with their overall conclusion but some conclusions seem to me incorrect (outlined below). The paper is clearly written, and experiments generally well performed. However, the microscopic figures are far too small. Even when displayed on a large screen many cellular details are not visible, e.g. Laurdan staining that does not show any clear septa. Firstly, a paper should be readable when printed, secondly, microscopic images should be clear so that phenotypic differences are clearly observable. This is currently not the case and should be improved.
We thank the reviewer for the overall very positive comments. We have made adjustments to the figures to make them clearer. For easy reading we have inserted the figures in the text, but high resolution files are now also available that would be used for the published version.
We also want to point out that microscopic images showing laurdan measurement do not aim to highlight individual cells or even subcellular region (e.g.) septa. The resolution is not good enough to make strong claims about membrane fluidity measured by laurdan GP in distinct membrane regions. We provide an overview in a heat map coloration to support the calculated values shown in individual graphs. The micrographs allow the reader to quickly judge about the effect of flotillins deletions or benzyl alcohol addition on the overall membrane fluidity. We feel the provided images serve this idea perfectly.
I do not agree with the conclusion in the Discussion (second paragraph) that membrane fluidity and peptidoglycan synthesis is higher at septa in order to compensate the absence of flotillins. This reasoning does not make sense because what should be compensated for? The overall fluidity still changes. It is also a bit difficult how this can be compensated, in other words, how can the fluidity be increased at septa?
We agree that we cannot draw a definitive conclusion on the increase of fluidity at the septum in a flotillin deficient strain. Therefore, we have rephrased the statement in the Discussion to clearly indicate that it is speculative. We do note, however, that it is in line with other recent findings from the Zapun lab on increased fluidity at the division site.
The statement now reads:
“Our data indicate that the reduction in elongasome activity, which does not impact the growth rate itself, is compensated by a shift to increased peptidoglycan synthesis activity around division sites in flotillin mutants, which is sufficient to keep the overall cell shape intact, although cells are elongated. […] This is yet to be determined, as the resolution of Laurdan imaging does not allow conclusive statements about local fluidity changes at the septum.”
Moreover, if there would be more peptidoglycan synthesis, then cells would divide faster, thus be shorter, since the authors show that septa do not become wider (Supplementary Figure 1). But a floAT deletion strain is not shorter (Figure 2).
We respectfully disagree – increased peptidoglycan synthesis around the septum does not necessarily mean that all the peptidoglycan is contributing to the septum and allows faster division. In fact a recent preprint from the Deckers-Hebestreit and Strahl labs in which membrane fluidity is lowered by changing the fatty acid composition also shows a disruption of MreB activity (and thus lateral wall synthesis) while cells are slightly elongated (personal communication H. Strahl, length measurements were not provided in the preprint but do show elongation. Preprint: Biorxiv 10.1101/852160). It may seem paradoxical that cells elongate by synthesis around the division site but this is not uncommon, as is mentioned in the Discussion and backed up with various references.
This discrepancy might be related to the possibility that the authors have not measured the septal intensities in Figure 1 correctly. That is, they should have compared the intensities of septa with that of the lateral wall, because it might well be that a floAT deletion strain has an overall more permeable cell wall that facilitate fluorophore diffusion.
We have tackled this issue in two ways – first, we repeated a series of experiments with two membrane dyes and one PG dye using a wild type strain expressing GFP, which allowed mixing of the two strains and direct comparison of septal signal intensities to exclude the possibility that differences in microscopy setup cause the measured intensity difference (new Figure 1—figure supplement 1A, B). Second, we have excluded the possibility that diffusion is different between the strains. This cannot be done with a peptidoglycan synthesis dye as these could also be differently distributed because of differences in peptidoglycan. Thus, we compared intensities at the lateral wall for the membrane dye Nile Red, in the new experiment with the mixed strains. No difference in lateral intensity was observed indicating that there is no difference in fluorophore diffusion (new Figure 1—figure supplement 1D).
In fact the laurdan images in Figure 3 even suggest this. Maybe the authors measured the relative increase in intensity at septa, but I was unable to determine this from their experimental description. Moreover, and importantly, their laurdan images show no evidence for an increase of fluidity in septa. This seems to contradict their own argumentation.
Laurdan is used to determine overall membrane fluidity in cells and does not have the resolving power in our setup to say anything about differences at the septum. We have added a statement to clarify this to the Discussion (second paragraph).
Reviewer #2:
The experiments are well done and for the most part convincing. The problem is that their conclusions are overstated. The claim that all effects are explained by a direct action of flotillins on membrane fluidity rather than primarily affecting the assembly or activity of a PBP is not supported by the data. In fact most of the experiments (except potentially the benzyl alcohol addition, see below) can be explained by the floAT double mutant affecting primary septal PG synthesis or turnover. This does not mean that the manuscript is not suited for eLife, the data should just be interpreted differently.
1) One could state that all of the observed effects are due to floAT control of a septal PBP or septal PG remodeling. First, the claim that there is more lipidII/PG precursors in Figure 1 in the float mutants is based on a small difference and is weak. The signal is barely above background and HADA does not mark sites of PG synthesis, rather it marks sites of PG remodeling/turnover, as it depends on the exchange of the labeled D-ala in lipid II on the extracytoplasmic side. Moreover, the fact that floAT cells depend strongly on PBP1 for cell division, provides genetic evidence that the double mutant is under septal PG stress, which would probably be observable by applying PG stress in the cytoplasm (with Fosfomycin for example) and extracellularly with β-lactams or glycopeptides or bacitracin or A22.
We appreciate the remark. We have carried out an additional control as suggested by the reviewer by growing the cells with sublethal amounts of fosfomycin to apply PG stress while keeping growth rate constant. This stress has an effect – such as bulging in wildtype cells. We have included an image (Figure 1—figure supplement 1F) that shows that the flotillin mutant also suffers from the fosfomycin stress. Importantly, if the flotillin deletion would lead to general PG stress such as caused by fosfomycin, the fosfomycin treated wildtype cells should look like the flotillin deletion without fosfomycin – which is not the case.
We included a remark to state that flotillin deletion does cause peptidoglycan synthesis stress. Other stressors were not included as they would impact on membrane processes (bacitracin, glycopeptides) or cause a potential accumulation of precursor (betalactams).
If PBP biosynthetic activity is reduced (or PG degradation increased) in the floAT mutants, would this not lead to a (slight) accumulation of lipid II at the septum as well, and this could explain the observation in Figure 1? Also, this effect on PG stress could easily explain the reduction in MreB movement.
It is true that the reduction of the activity of some specific PBPs (2a and H) can lead to reduced MreB movement, but it is not evident that this of itself would result in increased LipidII at the septum – one could also argue that the unprocessed LipidII accumulates at the lateral wall. But a general reduction in biosynthetic activity or increase of degradation would also be expected to affect septum formation.
Finally, my understanding was that an increase in lipidII would INCREASE membrane fluidity, but they show in Figure 3 that fluidity is reduced in the floAT mutant.
We have tried to explain this point better in the text – the overall fluidity of the membrane is reduced, but a shift resulting in a relative local increase of disordered lipids and LipidII at the septum would still keep the membrane more fluid at the septum relative to the rest of the cell.
2) Growth rate effect: I disagree that the difference between LB and SMM can be pinned on growth rate. The cellular metabolism is completely different in minimal vs. rich medium so how can the authors rule out that a change in biosynthetic pathways does not affect lipid composition, PG (bactoprenol or lipidII) synthesis, isoprenoids or simply certain membrane proteins, all of which could alter membrane fluidity. The growth rate argument does not hold here, and it could only be proven if the authors stick to LB and lower growth rate there, ideally by reducing temp, but there could be complications on fluidity arising from a temperature shift. Inducing ppGpp would also be possibility, but this could also reprogram cell metabolism and make the interpretation ambiguous. Perhaps use diluted LB?
We agree with the reviewer that there are alternative explanations for the differences observed between LB and SMM. We compared these to growth rate because in the B. subtilis literature the same point is made about growth rate and MreB velocity, which our paper shows may have a different cause – but we do observe a correlation between growth rate, fluidity and MreB velocity. We have modified our remarks to make it clear that growth rate or growth conditions determine fluidity. We agree that this needs to be investigated more carefully but think that is beyond the scope of the current paper as various causes need to be investigated such as the presence of specific lipids in the membrane in certain growth conditions, or differences in crowding of the membrane by membrane proteins during various growth conditions. We have tried the suggested experiment with diluted LB (Author response image 1). This experiment revealed that we could lower growth rate by diluting LB up to 5 times (stronger dilutions led to excessive lysis) – but not to the growth rate observed with SMM (Author response image 1B). Importantly, at all these conditions cell dimensions were still completely different. Cells grown on (diluted) LB are wider and longer than SMM grown cells (Author response image 1A). Also, at all LB dilutions, the phenotype of the triple mutant remains – elongated cells and delocalized cell wall synthesis (Author response image 1A).
3) The benzyl alcohol addition experiment is the best (and in my opinion only) argument for their claim on membrane fluidity, but how do we know that this compound directly affects the membrane, rather than acting indirectly, like SMM, changing metabolism?
The action of benzyl alcohol as a membrane fluidizer is well established in the literature. It acts by increasing membrane hydration, as has been shown in experiments on B. subtilis (Konopasek et al., 2000). We have clarified this in the text. But we agree that this was the sole experiment indicating a link between flotillins, membrane fluidity and cell wall synthesis (e.g. MreB dynamics). We therefore thought about a direct proof to show that flotillins really fluidize membranes. We teamed up with Dr. Birgit Habenstein in Bordeaux and used purified and reconstituted FloT for solid-state-NMR experiments (see answer below).
How quick are the changes observed after addition of BA? Are the changes in fluidity and MreB localization still observed when cells are pre-treated with protein synthesis inhibitors and /or cerulenin (FA inhibitor)? By the way, are FRAP measurements of membrane-tethered GFP consistent with a change in fluidity?
We appreciate the suggested experiments but have added a different, in vitro, experimental approach that provides an independent line of evidence that flotillin proteins are capable of fluidizing membranes. Using solid state NMR to determine lipid mobility in liposomes with and without reconstituted flotillin we show that the addition of FloT is sufficient to fluidize the liposomal membranes, to a much larger extent than observed for other membrane inserted proteins. This experiment has now been added to the paper in an additional section and Figure 5.
The bottom line is: I'm not a fan of the previous flotillin model either, but I don't think the authors provide compelling evidence in favor of their model. In fact, their experimental data does not completely exclude the previous model about flotillins facilitating PG assembly/turnover machines either.
We do hope that with the series of new data including the in vitro characterization of FloT, we have now proven the fluidizing effect of flotillins.
[Editors’ note: what follows is the authors’ response to the second round of review.]
Overall the study is solid, convincing, well presented and certainly suitable for publication in eLife. Nevertheless, some clarifications should be provided before the manuscript is accepted for publication.
1) It is not clear why a reduction in MreB mobility in the membrane of floAT cells grown in LB would lead to increased vancomycin-FL and HADA staining at the septum. This relationship is poorly accounted for in the manuscript. MreB is not at the septum, so why would lower mobility of MreB cause an increase in septal lipid II? This observation is thus circumstantial. MreB and septal lipid II could be both affected because of reduced fluidity, but those events do not have to be linked:
– Perhaps membranes are less fluid and if lipid II is inserted at the septum, it is less likely to diffuse away in floAT cells because of the reduced fluidity?
– Or there is just a preference for lipid II to accumulate there because the membrane is curved and fluidity at the curvature is affected in the absence of FloAT? If that's the case, what does the additional deletion of pbp1 really tell us (although it is interesting)?
We agree with some of the possible explanations provided. We have included a new paragraph in the Discussion to clarify this issue, in which some of the possibilities raised are included, and in which the relation between MreB and septal staining is further clarified and supported with additional references (Discussion, second paragraph).
2) To clarify these issues, these questions should be discussed and it could be easier for the reader if the order of the results were turned around and tell the story more linearly. For example, rather than going from PG studies to membrane fluidity studies, the manuscript could present the established part first (NMR, fluidity, MreB) and then argue that reduced MreB mobility prompted imaging with HADA, Vanco-FL and finally check the pbp1floAT triple mutant. The current read takes a circular path (starting with PG and ending with MreB) and again there is no compelling causality between MreB movement and septal PG presented here.
We have discussed a reorganization of the text with Dr. Mignot – we found that this would cause other problems with the organization of the text and Dr. Mignot agreed that changing the organization was not a must.
https://doi.org/10.7554/eLife.57179.sa2Article and author information
Author details
Funding
Nederlandse Organisatie voor Wetenschappelijk Onderzoek (Vidi 864.09.010)
- Dirk-Jan Scheffers
Deutsche Forschungsgemeinschaft (BR 2915/7-1)
- Marc Bramkamp
Deutsche Forschungsgemeinschaft (INST 86/1452-1)
- Marc Bramkamp
Nederlandse Organisatie voor Wetenschappelijk Onderzoek (721.014.008)
- Anna KH Hirsch
European Research Council (starting grant 757913)
- Anna KH Hirsch
Deutscher Akademischer Austauschdienst (PhD fellowship DAAD-GSSP)
- Abigail Savietto
Fundação para a Ciência e a Tecnologia (PhD fellowship SFRH/BD/78061/2011- POPH/FSE/FCT)
- Anabela de Sousa Borges
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Acknowledgements
We thank Henrik Strahl for discussions and sharing of unpublished data, Rut Carballido-Lopez for strain RWBS5 and Luiza Morawska and Oscar Kuipers for the PrrnB-gfp plasmid.
This work was funded by NWO grant 864.09.010 (DJS), DFG grants BR 2915/4–1; INST 86/1452–1 (MB), ERC starting grant 757913; NWO grant 721.014.008 (AKHH); PhD fellowships DAAD-GSSP to AS; and SFRH/BD/78061/2011- POPH/FSE/FCT to ASB.
Senior Editor
- Gisela Storz, National Institute of Child Health and Human Development, United States
Reviewing Editor
- Tâm Mignot, CNRS-Aix Marseille University, France
Reviewer
- Leendert Hamoen, University of Amsterdam, Netherlands
Version history
- Received: March 24, 2020
- Accepted: June 12, 2020
- Version of Record published: July 14, 2020 (version 1)
Copyright
© 2020, Zielińska et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 3,654
- Page views
-
- 472
- Downloads
-
- 24
- Citations
Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.
Download links
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Further reading
-
- Cell Biology
- Microbiology and Infectious Disease
The relative positions of viral DNA genomes to the host intranuclear environment play critical roles in determining virus fate. Recent advances in the application of chromosome conformation capture-based sequencing analysis (3 C technologies) have revealed valuable aspects of the spatiotemporal interplay of viral genomes with host chromosomes. However, to elucidate the causal relationship between the subnuclear localization of viral genomes and the pathogenic outcome of an infection, manipulative tools are needed. Rapid repositioning of viral DNAs to specific subnuclear compartments amid infection is a powerful approach to synchronize and interrogate this dynamically changing process in space and time. Herein, we report an inducible CRISPR-based two-component platform that relocates extrachromosomal DNA pieces (5 kb to 170 kb) to the nuclear periphery in minutes (CRISPR-nuPin). Based on this strategy, investigations of herpes simplex virus 1 (HSV-1), a prototypical member of the human herpesvirus family, revealed unprecedently reported insights into the early intranuclear life of the pathogen: (I) Viral genomes tethered to the nuclear periphery upon entry, compared with those freely infecting the nucleus, were wrapped around histones with increased suppressive modifications and subjected to stronger transcriptional silencing and prominent growth inhibition. (II) Relocating HSV-1 genomes at 1 hr post infection significantly promoted the transcription of viral genes, termed an ‘Escaping’ effect. (III) Early accumulation of ICP0 was a sufficient but not necessary condition for ‘Escaping’. (IV) Subnuclear localization was only critical during early infection. Importantly, the CRISPR-nuPin tactic, in principle, is applicable to many other DNA viruses.
-
- Cell Biology
Impaired spermatogenesis and male infertility are common manifestations associated with mitochondrial diseases, yet the underlying mechanisms linking these conditions remain elusive. In this study, we demonstrate that mice deficient for the mitochondrial intra-membrane rhomboid protease PARL, a recently reported model of the mitochondrial encephalopathy Leigh syndrome, develop early testicular atrophy caused by a complete arrest of spermatogenesis during meiotic prophase I, followed by degeneration and death of arrested spermatocytes. This process is independent of neurodegeneration. Interestingly, genetic modifications of PINK1, PGAM5, and TTC19 – three major substrates of PARL with important roles in mitochondrial homeostasis – fail to reproduce or modify this severe phenotype, indicating that the spermatogenic arrest arises from distinct molecular pathways. We further observed severe abnormalities in mitochondrial ultrastructure in PARL-deficient spermatocytes, along with prominent electron transfer chain defects, disrupted coenzyme Q (CoQ) biosynthesis, and metabolic rewiring. These mitochondrial defects are associated with a germ cell-specific decrease in GPX4 expression leading arrested spermatocytes to ferroptosis – a regulated cell death modality characterized by uncontrolled lipid peroxidation. Our results suggest that mitochondrial defects induced by PARL depletion act as an initiating trigger for ferroptosis in primary spermatocytes through simultaneous effects on GPX4 and CoQ – two major inhibitors of ferroptosis. These findings shed new light on the potential role of ferroptosis in the pathogenesis of mitochondrial diseases and male infertility warranting further investigation.