Asymmetric division triggers cell-specific gene expression through coupled capture and stabilization of a phosphatase

  1. Niels Bradshaw
  2. Richard Losick  Is a corresponding author
  1. Harvard University, United States

Abstract

Formation of a division septum near a randomly chosen pole during sporulation in Bacillus subtilis creates unequal sized daughter cells with dissimilar programs of gene expression. An unanswered question is how polar septation activates a transcription factor (σF) selectively in the small cell. We present evidence that the upstream regulator of σF, the phosphatase SpoIIE, is compartmentalized in the small cell by transfer from the polar septum to the adjacent cell pole where SpoIIE is protected from proteolysis and activated. Polar recognition, protection from proteolysis, and stimulation of phosphatase activity are linked to oligomerization of SpoIIE. This mechanism for initiating cell-specific gene expression is independent of additional sporulation proteins; vegetative cells engineered to divide near a pole sequester SpoIIE and activate σF in small cells. Thus, a simple model explains how SpoIIE responds to a stochastically-generated cue to activate σF at the right time and in the right place.

https://doi.org/10.7554/eLife.08145.001

eLife digest

An important question in biology is how genetically identical cells activate different sets of genes. This is particularly perplexing for cells that rely on random events to specify the genes they switch on. Normally, cells of a bacterium called Bacillus subtilis divide symmetrically to produce two identical cells that express identical sets of genes. However, B. subtilis cells can also undergo a developmental program to form a spore to help it survive periods of extreme conditions. To do this, first a B. subtilis cell divides asymmetrically by placing the site of division close to a randomly selected end of the cell. This creates a smaller cell that becomes the spore and a larger cell that nurtures the developing spore. Each cell must turn on different genes to play its role in spore development, but how asymmetry in the position of cell division leads to these differences in gene expression has been a longstanding mystery.

Bradshaw and Losick studied a regulatory protein called SpoIIE, which is responsible for switching on genes in the small cell. SpoIIE is made before cells divide asymmetrically, but only accumulates in the small cell. The experiments revealed that an enzyme broke down the SpoIIE protein if it wasn’t in the small cell. This prevented SpoIIE from incorrectly switching on genes before division was completed or in the large cell.

Protection of SpoIIE from being broken down in the small cells was then shown to be linked to the placement of cell division; SpoIIE first accumulates at the asymmetrically positioned cell division machinery and then is transferred to a secondary binding site at the nearby end of the cell. Capture of SpoIIE at the end of the cell was coupled to its stabilization as SpoIIE molecules interacted with one another to form large complexes.

Together these findings provide a simple mechanism to link the asymmetric position of cell division to differences in gene expression. Future studies will focus on understanding how SpoIIE is captured at the end of the cell and how this prevents SpoIIE from being degraded.

https://doi.org/10.7554/eLife.08145.002

Introduction

How genetically identical daughter cells adopt dissimilar programs of gene expression following cell division is a fundamental problem in developmental biology. A common mechanism for establishing cell-specific gene expression is asymmetric segregation of a cell fate determinant between the daughter cells (Horvitz and Herskowitz, 1992; Neumüller and Knoblich, 2009). In polarized cells, intrinsic asymmetry can be inherited from generation to generation. For example, the dimorphic bacterium Caulobacter crescentus localizes certain cell fate determinants to the old cell pole, leading to their asymmetric distribution following division (Iniesta and Shapiro, 2008; Bowman et al., 2011). However, non-polarized cells such as Bacillus subtilis must generate asymmetry de novo, which is passed on to the daughter cells to differentiate.

Bacillus subtilis divides by binary fission to produce identical daughter cells during vegetative growth but switches to asymmetric division when undergoing the developmental process of spore formation (Piggot and Coote, 1976; Stragier and Losick, 1996). To sporulate, cells place a division septum near a randomly chosen pole of the cell (Veening et al., 2008) to create two unequally sized daughter cells with dissimilar programs of gene expression. The smaller cell, the forespore, which largely consists of the cell pole, will become the spore, whereas the larger cell, the mother cell, nurtures the developing spore (Figure 1B). An enduring mystery of this developmental system is how stochastically generated asymmetry initiates dissimilar programs of gene expression in the daughter cells resulting from polar division (Barak and Wilkinson, 2005).

Figure 1 with 1 supplement see all
SpoIIE is compartmentalized in the forespore and activates σF.

(A) Diagram of the pathway for σF activation. Phorphorylated SpoIIAA (AA-P) is dephosphorylated by SpoIIE. Dephosphorylated SpoIIAA (AA) then binds to SpoIIAB (AB) displacing σF and leading to σF-directed transcription. (B) SpoIIE (green), AA, AB and σF are produced in predivisional cells. Prior to completion of asymmetric cell division SpoIIE associates with the polar divisome near one or both cell poles (the pole at which division initiates is chosen randomly). Following completion of cytokinesis, SpoIIE is enriched in the forespore where it dephosphorylates AA-P to activate σF. (C) A montage of images taken every 6 min from a single sporulating cell (strain RL5876) producing SpoIIE-YFP. Cells are oriented with the forespore on the right as in the diagram. A movie of this sporulating cell is provided as Video 1. Scale bar: 0.5 µm.

https://doi.org/10.7554/eLife.08145.003
Video 1
Movie file of the sporulating cell shown in Figure 1C (2fps).
https://doi.org/10.7554/eLife.08145.005

The earliest acting cell-specific regulatory protein in the sporulation program is the transcription factor σF. The σF factor and the proteins that control it – SpoIIAB, SpoIIAA, and SpoIIE – are produced at the onset of sporulation (Gholamhoseinian and Piggot, 1989), but σF is held inactive until the completion of asymmetric cell division, when it turns on gene expression selectively in the forespore (Margolis et al., 1991; Stragier and Losick, 1996) (Figure 1A,B). SpoIIAB is an anti-sigma factor that traps σF in an inactive complex (Min et al., 1993; Duncan and Losick, 1993). Escape from SpoIIAB is mediated by the anti-anti-sigma factor SpoIIAA (Diederich et al., 1994). SpoIIAA is, in turn, activated by SpoIIE, a member of the PP2C family of protein phosphatases (Bork et al., 1996; Levdikov et al., 2011). SpoIIE converts the inactive phosphorylated form of SpoIIAA (SpoIIAA-P) to the active dephosphorylated form (Duncan et al., 1995) (Figure 1A). Dephosphorylation of SpoIIAA-P by SpoIIE is therefore the critical event in activating σF. Understanding how SpoIIE reads out cellular cues to delay dephosphorylation of SpoIIAA-P until after septation and restrict phosphatase activity to the forespore is thus the central challenge in understanding how cell-specific gene transcription is established during sporulation.

SpoIIE consists of three domains: a PP2C phosphatase domain at the C-terminus, a ten-pass transmembrane domain at the N-terminus, and a 270-amino acid central domain (henceforth referred to as the regulatory domain) that is important for regulating SpoIIE compartmentalization and activity (Figure 2B; Arigoni et al., 1999). Prior to asymmetric cell division, SpoIIE localizes to the polar divisome and contributes to its placement (Arigoni et al., 1995; Ben-Yehuda and Losick, 2002). After septation is complete, SpoIIE is found principally in the forespore and to a limited extent at a second polar divisome near the distal cell pole (Figure 1C, Video 1).

Figure 2 with 1 supplement see all
SpoIIE degradation depends on FtsH.

(A) SpoIIE is compartmentalized to the forespore. A single sporulating cell (strain RL5874) is shown following the completion of asymmetric septation. The membrane stained with FM4-64 is shown in red above SpoIIE-YFP and CFP driven by an in frame fusion to the start of the spoIIE open reading frame. Scale bar: 0.5 µm. (B) The domain architecture of SpoIIE. The N-terminal cytoplasmic tail (red), followed by 10 transmembrane-spanning segments, the regulatory region (amino acids 320–589, gray), and the phosphatase domain (amino acids 590–827, green). (C) SpoIIE is degraded during sporulation. Translation was arrested (by addition of 100 µg/ml chloramphenicol) in sporulating cells producing SpoIIE-FLAG (strain RL5877), and samples were withdrawn at the indicated times. SpoIIE was detected by western blot using α-FLAG monoclonal antibody (left). Quantitation of the western (right) fit to a single exponential equation. (D) The genes for spoIIE and ftsH are near each other in the genome with conserved synteny. The diagram shows genomic organization of diverse endospore forming species. Filled boxes indicate genes with spoIIE in green, ftsH in orange, and other genes in gray. (E) SpoIIE degradation requires FtsH, and FtsH mediated degradation requires the N-terminal cytoplasmic tail (TagSpoIIE) of SpoIIE. Translation was arrested in vegetatively growing cells producing SpoIIE-FLAG with an IPTG inducible promoter (strain RL5878 wt, RL5879 ∆ftsH, and RL5880 SpoIIE-∆Tag (residues 11–37 were deleted). Degradation was monitored (left) and quantitated (right) as in panel C. (F) TagSpoIIE is sufficient to target a heterologous protein for FtsH-dependent degradation. Degradation of MalF-TM-FLAG fused at the N-terminus to either TagSpoIIE (wt RL5888, ∆ftsH RL5889) or the first 10 amino acids of SpoIIE (wt RL5890, ∆ftsH RL5891) produced during exponential growth was monitored as in panel C.

https://doi.org/10.7554/eLife.08145.006

Here we describe three interdependent features of SpoIIE that together explain how SpoIIE links polar septation to the cell-specific activation of σF: (1) SpoIIE is proteolytically unstable and is degraded dependent on the AAA+ protease FtsH; (2) SpoIIE is transferred from the polar divisome – the de novo origin of asymmetry – to the proximal cell pole during polar septation; and (3) SpoIIE forms homooligomeric complexes which promote capture at the pole, protection from proteolysis and activation as a phosphatase, thus linking the cues that direct localization of SpoIIE to its stabilization and activation.

Results

SpoIIE is degraded in an FtsH-dependent manner

Although transcription of spoIIE commences in pre-divisional cells and continues in the mother cell following cytokinesis (Fujita and Losick, 2003), SpoIIE protein and activity are restricted to the forespore (Figure 2A). This apparent contradiction led us to consider the possibility that spatially restricted proteolysis contributes to compartmentalization of SpoIIE. Selective stabilization of SpoIIE in the forespore coupled to efficient global degradation would enrich SpoIIE in the forespore despite ongoing transcription of spoIIE in predivisional cells and the mother cell. To investigate this hypothesis, we sought to determine if SpoIIE turns over on a timescale commensurate with σF activation and, if so, to identify the responsible protease.

To detect SpoIIE degradation during sporulation, we monitored the disappearance of SpoIIE with a C-terminal FLAG tag following inhibition of translation with chloramphenicol (Figure 2C). SpoIIE-FLAG was degraded with a half-life of 7 min, demonstrating that SpoIIE is unstable relative to the approximately 1 hr progression of asymmetric cell division and σF activation (Figure 1C) and supporting spatially restricted degradation as a plausible mechanism to compartmentalize SpoIIE.

Next, we sought to identify the protease that degrades SpoIIE. We noticed that the gene (ftsH) for the transmembrane AAA+ protease FtsH is located near spoIIE in the genome with highly conserved synteny (Figure 2D). Furthermore, FtsH is known to degrade transmembrane protein substrates (Akiyama, 2009), making it an attractive candidate protease for SpoIIE. FtsH degrades several proteins that block entry into sporulation and prevent the expression of spoIIE, such as the Spo0A inhibitor Spo0E (Le and Schumann, 2009). We therefore engineered the synthesis of SpoIIE during vegetative growth to bypass the requirement for FtsH in the expression of spoIIE. In exponential phase cells deleted for ftsH, SpoIIE was stable for more than 1 hr after chloramphenicol treatment, whereas in ftsH cells SpoIIE was degraded as rapidly as during sporulation (t1/2 = 7.1 min) (Figure 2E). (SpoIIE instability and its dependence on FtsH was also seen with untagged SpoIIE [Figure 2—figure supplement 1A,B]). We conclude that SpoIIE is degraded in an FtsH-dependent manner. The simplest explanation for this is that SpoIIE is a direct substrate for the protease.

Finally, we attempted to identify the feature or features of SpoIIE that renders it susceptible to proteolysis by FtsH. Truncation of the N-terminal, cytosolic tail of SpoIIE (removal of residues 11 to 37) blocked degradation (Figure 2E, ∆Tag), whereas removal of the regulatory domain or the phosphatase domain or substitution of the transmembrane domain with the first two transmembrane segments of E. coli MalF (MalF-TM) did not impede FtsH-dependent degradation (Figure 2—figure supplement 1C). Additionally, the first 37 amino acids of SpoIIE (TagSpoIIE) were sufficient to confer FtsH-dependent degradation on a heterologous protein, MalF-TM-FLAG (Figure 2F). Therefore, the N-terminal tail of SpoIIE is a tag that is both necessary and sufficient for FtsH-dependent proteolysis.

Degradation restricts SpoIIE and σF activity to the forespore

To test whether SpoIIE degradation is required for compartmentalization of σF activity and SpoIIE, we examined the effect of blocking degradation during sporulation. Here and in the experiments that follow, we removed TagSpoIIE from SpoIIE in cells that were wild type for ftsH to selectively block SpoIIE degradation and circumvent off-target effects from other FtsH substrates had we used an ftsH mutation. Indeed, we observed a dramatic increase in aberrant activation of σF in ∆tag spoIIE cells (Figure 3A). Whereas in wild-type cells σF activity was highly specific for the forespore (less than 2% non-specific activation), ∆tag spoIIE caused non-specific activation of σF in 71% of the cells (Figure 3B). Quantification of σF activity with a lacZ reporter revealed that a strain with ∆tag spoIIE activated σF with a similar time dependence but had 10-fold elevated σF activity (Figure 3C).

Figure 3 with 1 supplement see all
Degradation of SpoIIE is required to compartmentalize SpoIIE and σF activity.

(A) Images of sporulating cells producing SpoIIE-YFP (left, strain RL5876) or ∆Tag SpoIIE-YFP (right, strain RL5892). Top images show CFP (blue) produced under the control of the σF dependent spoIIQ promoter and FM4-64-stained membrane (white); bottom images show SpoIIE-YFP. White arrows indicate cells with uncompartmentalized σF activity and SpoIIE in the mother cell. The contrast for images of SpoIIE-YFP has been adjusted to approximately 5X brighter than for ∆Tag-SpoIIE-YFP for display purposes. Scale bar: 1 µm. (B) Quantification of the forespore specificity of σF activity from hundreds of cells from images as shown in panel A. (C) σF activity was measured during sporulation using a translational fusion of the σF dependent SpoIIQ promoter to LacZ (wt SpoIIE strain RL5893, ∆Tag SpoIIE strain RL5894). Time after initiation of sporulation is indicated, and error bars represent the standard deviation from three biological replicates. (D) Quantification of the dependence of σF activation on asymmetric cell division driven by ∆Tag-SpoIIE. Hundreds of cells from images as shown in panel A were manually assessed for completion of asymmetric cell division and compartmentalization of σF activity. The percent of cells with each pattern of σF activity is indicated for ∆tag-spoIIE cells (values for wt spoIIE cells are indicated in parenthesis).

https://doi.org/10.7554/eLife.08145.008

Activation of σF is tightly coupled to the completion of asymmetric cell division, and SpoIIE mutants have been characterized that uncouple cell division and σF activation (Carniol et al., 2004; Feucht, et al., 2002; Hilbert and Piggot, 2003). In contrast to these other cases of σF mis-activation in predivisional cells, stabilization of SpoIIE led to activation of σF primarily in cells that had completed asymmetric cell division (Figure 3A,D). Thus, stabilization of SpoIIE only partially uncouples σF activation from cell division.

Consistent with the idea that degradation contributes to compartmentalization of σF activity by helping to restrict SpoIIE to the forespore, we observed a striking correlation between elevated levels of ∆Tag-SpoIIE in the mother cell and mis-activation of σF (Figure 3A bottom panels). Together, these data indicate that SpoIIE degradation is required for compartmentalization both of SpoIIE and σF activity.

SpoIIE mutants blocked in compartmentalization and σF activation

How is SpoIIE selectively stabilized in the forespore? We considered two models: (1) FtsH is not active in the forespore, or (2) specific features of SpoIIE stabilize it in the forespore. To test the former possibility, we engineered the production of the model FtsH substrate TagSpoIIE-MalF using a forespore specific, σF-dependent promoter. TagSpoIIE-MalF was rapidly degraded in a manner dependent on the TagSpoIIE (Figure 3—figure supplement 1). Therefore FtsH is active in the forespore, suggesting that SpoIIE is specifically stabilized against FtsH-dependent degradation.

To identify features of SpoIIE required for its accumulation in the forespore, we screened for SpoIIE variants defective in compartmentalization. We created amino acid substitutions of the most highly conserved residues in SpoIIE and tested these variants (and previously described variants) for function in sporulation (Figure 4A, Figure 4—source data 1) and forespore accumulation (Figure 4B). To monitor accumulation in the forespore of each SpoIIE variant, we compiled average profiles of SpoIIE-YFP along the long axis of hundreds of cells that had undergone polar division. Through this analysis, we identified nine variants of SpoIIE (for example SpoIIEK356D) that were absent in the forespore (Figure 4A,B blue) and accumulated to reduced levels (Figure 4C). Variants with normal compartmentalization, in contrast, accumulated at approximately wild-type levels (Figure 4B,C black). Supporting the idea that failure to accumulate in the forespore was due to unrestricted, FtsH-dependent degradation, removal of TagSpoIIE restored these SpoIIE mutant proteins to levels several-fold higher than for wild-type SpoIIE and equivalent to ∆Tag SpoIIE (Figure 4D). We conclude that SpoIIE undergoes a transition in the forespore that protects it from FtsH-dependent proteolysis and that this transition is blocked by amino acid substitutions such as K356D.

Figure 4 with 1 supplement see all
SpoIIE stabilization and localization mutants.

(A) Diagram of SpoIIE mutants with sporulation defects. Variants with localization and σF activation defects are shown above the diagram in blue, and variants with normal localization but defects in σF activation are shown below the diagram in black. (B) Localization of SpoIIE mutants. Hundreds of asymmetrically divided cells were aligned at the forespore pole to generate average profiles of SpoIIE-YFP localization for each SpoIIE mutant (strains RL5895- 5909) with a reference plot (gray) from wild-type SpoIIE-YFP from σF mutant cells (strain RL5910). The dashed line represents the approximate position of the asymmetric septum. Images of representative cells with the mislocalized variant SpoIIEK356D (left, strain RL5895) and forespore-localized SpoIIEQ483A (right, strain RL5904) are shown. (C) Western blots of protein levels in SpoIIE mutant strains shown in panel B probed for SpoIIE-YFP (with α-GFP antibody), CFP produced under the control of a σF-driven promoter, and σA as a loading control. Levels of each SpoIIE variant were normalized to wt SpoIIE (strain RL5876). Error bars represent the standard deviation from three biological replicates. All localization mutants (shown in blue) are different from the σF mutant control with p values less than 0.0025 from a paired t-test. (D) Western blots of SpoIIE compartmentalization mutants with TagSpoIIE removed (strains RL5911-5916, with strain RL5876 as a reference) as in panel C.

https://doi.org/10.7554/eLife.08145.010
Figure 4—source data 1

Sporulation efficiency of SpoIIE mutants.

Cultures of indicated strains with were grown in DSM for 28 hr at 37°C followed by heat killing for 20 min at 85°C. The number of spores was determined by counting viable cells.

https://doi.org/10.7554/eLife.08145.011

Additionally, we tested whether stabilization of the SpoIIE mutant proteins was sufficient to support σF activation independent of their susceptibility to degradation. We found that even when TagSpoIIE was removed, the mutant proteins failed to activate σF (Figure 4D middle panel). A simple unifying model is that the proposed, K356-dependent conformational rearrangement that protects SpoIIE from proteolysis in the forespore is also required to allow it to activate σF.

An allele-specific suppressor of SpoIIEK356D

To investigate the link between stabilization, compartmentalization and activation of SpoIIE, we selected for and isolated several suppressors that restored sporulation to the compartmentalization-defective mutant spoIIEK356D. We chose this mutant because the K356D substitution was located in the regulatory domain of SpoIIE and caused a particularly severe sporulation defect. We isolated intragenic suppressors at two codons in an apparently saturating screen (see Materials and methods, Figure 4—source data 1 and Figure 4—figure supplement 1). One of them, causing a T353I substitution, was allele-specific (it suppressed spoIIEK356D but not spoIIES361F or spoIIEV490K, Figure 4—source data 1). The T353I substitution restored SpoIIEK356D to wild-type protein levels (Figure 4—figure supplement 1B), partially restored restriction of SpoIIEK356D to the forespore (Figure 4—figure supplement 1C), and restored compartment-specific σF activation (Figure 4—figure supplement 1D). The coordinated rescue of these phenotypes by a single amino acid substitution supports our model that a common feature of SpoIIE mediates protection from proteolysis, accumulation in the forespore, and activation of σF.

The other intragenic suppressors of SpoIIEK356D were substitutions at V697 (V697A and V697F), which is located in the phosphatase domain of SpoIIE. V697A had been independently isolated previously and shown to cause premature activation of σF (in the absence of the K356D substitution) (Hilbert and Piggot, 2003). These suppressors were not allele specific; V697A suppressed all other mutants of SpoIIE, including the compartmentalization defective SpoIIES361F mutant and the compartmentalized SpoIIEQ483A mutant (Figure 4—source data 1, [Carniol et al., 2004]). The V697A substitution did not restore compartmentalization or stabilization of SpoIIE. It did restore σF activation but not compartmentalization of σF activity (Figure 4—source data 1 and Figure 4—figure supplement 1). All together, these results suggest that the V697A substitution locks the phosphatase domain in a high activity state, bypassing the activation defect of SpoIIEK356D. Indeed, biochemical experiments showed that the V697A substitution enhanced the activity of SpoIIE in dephosphorylating SpoIIAA-P (Figure 4—figure supplement 1E).

SpoIIE is compartmentalized and stabilized by binding to the cell pole

To further investigate the mechanism of SpoIIE compartmentalization, we took advantage of the compartmentalization-defective variant SpoIIEK353D and revisited the localization of stabilized SpoIIE. To isolate events prior to σF activation, and because certain targets of σF (e.g. spoIIQ) affect the localization of SpoIIE (Campo et al., 2008), we used a mutant lacking σF to analyze the localization of SpoIIE, ∆Tag-SpoIIE, and its K353D mutant derivative (in contrast to the experiment of Figure 3A in which cells were σF+).

Our most striking observation was that ∆Tag-SpoIIE was noticeably enriched at the poles of cells that had not initiated polar division (Figure 5A). Polar enrichment was dependent on stabilization by removal of TagSpoIIE (Figure 5A gray line). Because the forespore is derived from the cell pole, we hypothesized that the pole is a landmark that directs SpoIIE compartmentalization. In support of this idea, polar localization was abolished by the K356D substitution and partially restored by the T353I suppressor (Figure 5A lower panel). Thus, the pole is a cue that directs compartmentalization of SpoIIE, and the same feature(s) of SpoIIE that is required for polar recognition is also required for stabilization and σF activation.

Figure 5 with 1 supplement see all
Stabilized SpoIIE localizes to the cell pole.

(A) Average profiles of SpoIIE-YFP from undivided sporulating cells lacking σF activity are shown (strains RL5910, 5917, 5912, 5918), with a representative cell with ∆Tag-SpoIIE-YFP (strain RL5917) displayed above. (B) Vegetatively growing cells expressing SpoIIE-YFP (strains RL5919-5922) displayed as in panel A; variants as indicated) were imaged 30 min after induction of expression of the FtsZ polymerization inhibitor MciZ. (C) MciZ-expressing cells (ΔdivIVA [strain RL5923] or otherwise wildtype [strain RL5924]) were imaged as in panel B and average profiles of ∆Tag-SpoIIE-YFP were generated from 20 randomly selected cell poles. (D) DivIVA-FLAG was immunoprecipitated with α-FLAG magnetic beads from extracts of sporulating cells expressing SpoIIE variants as indicated (strains RL5925, 5926), and detected by Western blot. The elution (e) shown is 100X concentrated relative to the load (l) and flowthrough (ft) samples. Blots were probed with α-GFP antibody (top two images; lower image in high contrast*) and α-DivIVA antisera (below). Because DivIVA oligomerizes, untagged DivIVA is also co-immunoprecipitated. (E) SpoIIE preferentially localizes to the divisome rather than the pole. Representative cells expressing SpoIIE-YFP and CFP-ZapA are shown. Left images show exponentially growing divICts cells (strains RL5927, 5928), and right images show a sporulating ΔdivIB cell (strain RL5929). Scale bars indicate 0.5 µm in all panels.

https://doi.org/10.7554/eLife.08145.013
Video 2
Movie file of the sporulating cell shown in Figure 1—figure supplement 1A (2fps).

SpoIIE-YFP is shown in grey, and the divisome marked by CFP-ZapA is shown in blue.

https://doi.org/10.7554/eLife.08145.015
Video 3
Movie file of the sporulating cell shown in Figure 1—figure supplement 1B (2fps).

SpoIIE-YFP is shown in grey, the divisome marked by CFP-ZapA is shown in blue, and the membrane marked by MalFtm-mNeptune is shown in red.

https://doi.org/10.7554/eLife.08145.016

We next asked whether SpoIIE has an intrinsic affinity for the cell pole or whether SpoIIE is captured there by features unique to cells undergoing sporulation. To address this question, we engineered the synthesis of SpoIIE-YFP and ∆Tag SpoIIE-YFP in vegetative cells that were blocked in divisome formation through the use of the FtsZ inhibitor MciZ (Handler et al., 2008). We observed that ∆Tag SpoIIE-YFP was enriched at the ends of these cells, and this localization recapitulated the features of polar localization seen during sporulation: polar localization was only observed for stabilized SpoIIE, was blocked by K356D substitution, and was restored by the T353I suppressor (Figure 5B). Therefore, a fundamental, constitutive feature of the cell pole mediates SpoIIE polar localization.

We next sought to identify the feature of the pole that is responsible for SpoIIE localization. DivIVA recognizes the negative curvature of the cell pole and directs the polar localization of several other proteins during growth and sporulation (Lenarcic et al., 2009; Ramamurthi and Losick, 2009). Recently, DivIVA was shown to co-immunoprecipitate with SpoIIE, making it an attractive candidate to anchor SpoIIE to the cell pole (Eswaramoorthy et al., 2014). We found that whereas wild-type SpoIIE co-immunoprecipitated with DivIVA from extracts of sporulating cells (as previously observed), SpoIIEK356D did not (Figure 5D). Additionally ∆Tag SpoIIE-YFP polar localization during vegetative growth was abolished by a divIVA deletion (in the background of a minD deletion to suppress the cell division defect of divIVA deletion) (Figure 5C). Thus, DivIVA directly or indirectly anchors SpoIIE at the cell pole and can do so independently of sporulation. Based on the result with the K356D mutant, we further propose that this anchoring serves to stabilize, compartmentalize and activate SpoIIE, and that these activities are linked through a common feature of SpoIIE.

The divisome competes with the cell pole for binding of SpoIIE

During sporulation, SpoIIE first accumulates at the polar divisome, constricts along with the septum and then is released into the forespore following the completion of cytokinesis (as shown by time-lapse and structured illumination microscopy in Figure 1C, Videos 13, Figure 1—figure supplement 1). This suggests that the divisome competes with the pole for SpoIIE binding and that SpoIIE is not free to associate with the pole until the divisome is disassembled. To investigate this model, we monitored SpoIIE localization in the background of a temperature-sensitive allele of divIC that stalls cell division after divisome formation but before cytokinesis (Levin and Losick, 1994). In this background, SpoIIE-YFP and ∆Tag SpoIIE-YFP localized to the divisome (as visualized with a ZapA-CFP fusion) but not to the cell pole (Figure 5E). Similarly, when cytokinesis was blocked during sporulation by a divIB deletion (Thompson et al., 2006), ∆Tag SpoIIE-YFP localized to the divisome but not to the cell pole (Figure 5E). Therefore, the divisome sequesters and prevents SpoIIE from associating with the cell pole. We conclude that SpoIIE has affinity for two subcellular sites: the divisome, its dominant binding site, and the pole, where it is captured only after release from the divisome after the completion of cytokinesis.

SpoIIE is compartmentalized in vegetative cells engineered to divide asymmetrically

The results discussed above suggest a simple model for how SpoIIE and σF activity are compartmentalized in the forespore. We propose that SpoIIE is sequestered at the asymmetrically positioned divisome and is released and captured at the proximal (forespore) pole when cytokinesis is completed. In support of this idea, cells that cannot synthesize additional SpoIIE molecules in the forespore nonetheless robustly compartmentalize SpoIIE (Figure 5—figure supplement 1). Asymmetric compartmentalization of SpoIIE in the forespore could be achieved by virtue of the close proximity of the divisome and the forespore pole. Weak SpoIIE association with the pole would be compensated for by the small volume of the forespore and reinforced by protection from degradation by FtsH. Finally, any SpoIIE released into the mother cell would be captured at the divisome, preventing capture at the mother cell pole. Thus, a simple model explains how SpoIIE is protected from degradation and compartmentalized in the forespore only after cytokinesis is complete.

The heart of this model is that asymmetric positioning of the division septum is all that is necessary for compartment specific stabilization and activation of SpoIIE. To test this prediction, we sought to compartmentalize SpoIIE in cells that had been engineered to undergo polar division independently of sporulation. To do so, we artificially expressed spoIIE and overexpressed ftsAZ in vegetative cells, which was previously demonstrated to reposition the division septum from the mid-cell to near the pole (Ben-Yehuda and Losick, 2002). To preclude transcription of sporulation-specific genes, we additionally deleted the master regulator for entry into sporulation, spo0A. We then visualized the localization of SpoIIE in these cells. As predicted by our model, cells enriched for SpoIIE were much smaller than average for the entire population (Figure 6A,B). Further, to ask whether this compartmentalization of SpoIIE was sufficient to direct cell-specific activation of σF, we additionally induced synthesis of σF and its regulators SpoIIAA (the anti-anti σF factor substrate for the SpoIIE phosphatase) and SpoIIAB (the anti-σF factor) in a strain harboring a reporter for σF activity. Remarkably, σF was activated with high selectivity in a subpopulation of the minicells (Figure 6A,B). We conclude that polar division is the only feature of sporulation necessary to restrict SpoIIE protein and activity to the small cell and that this is sufficient to explain compartmentalized activation of σF.

Repositioning the septum in vegetative cells is sufficient for compartmentalization of SpoIIE.

(A) In the top image, vegetatively growing cells producing SpoIIE-GFP and overexpressing the ftsZ operon formed minicells enriched for SpoIIE-GFP (strain RL5930). In the lower image, cells additionally expressed the spoIIA operon and harbored a YFP reporter for σF activity (strain RL5931). (B) Quantification of images as shown in panel A. SpoIIE enriched cells (representing 56 out of 3168 total cells) are shown in green in the middle plot and are defined as cells with 2 standard deviations above the mean SpoIIE-GFP intensity. Cells that had active σF were rare in the population (12 cells) and were identified based on detectable levels of YFP fluorescence and their sizes were measured. They are shown in blue at the bottom.

https://doi.org/10.7554/eLife.08145.017

Multimerization of SpoIIE is required for polar anchoring and activation of σF

To gain insight into the molecular mechanism of SpoIIE localization and activation, we expressed and purified the C-terminal cytosolic domain of SpoIIE (residues 320 to 827, SpoIIE320-827) for biochemical characterization. A striking feature of SpoIIE320-827 was that it multimerizes, and analytical ultracentrifugation revealed these multimers to be hexamers and higher order assemblies of hexamers (Figure 7A).

Figure 7 with 1 supplement see all
Multimerization is required for compartmentalization of SpoIIE and σF activation.

(A) Purified soluble SpoIIE320-827 was analyzed by sedimentation velocity analytical ultracentrifugation detected by absorbance at 280nm and fitted to c (s) using Sedfit. Peaks corresponding to the predicted sedimentation coefficient for monomeric SpoIIE, hexamers, and multimers of hexamers were observed. (B) Diagram of the mutations and truncations in SpoIIE analyzed in panel C. (C) Multimerization of SpoIIE variants was analyzed by gel filtration with a 24 ml Superose 6 column. 1 ml fractions from 7-–18 ml of the run were collected, run on SDS-PAGE gels, and stained with SYPRO Ruby. (D) Model for handoff of SpoIIE from the divisome to the adjacent cell pole. SpoIIE (green) initially accumulates at the divisome and constricts along with FtsZ during cytokinesis. Prior to the completion of cell division, SpoIIE is degraded by FtsH through its TagSpoIIE (red). (We cannot distinguish whether association with the divisome protects SpoIIE from proteolysis or if SpoIIE turns over while associated with the divisome.) Upon the completion of cytokinesis, SpoIIE transfers to the adjacent cell pole where multimerization protects it from proteolysis (as depicted by the light orange Tags) and leads to phosphatase activation. We propose that close proximity favors transfer to the immediately adjacent pole and that concentration of SpoIIE in the forespore, which is almost entirely derived from the pole, promotes multimerization.

https://doi.org/10.7554/eLife.08145.018

To identify determinants of multimerization and its role for SpoIIE function, we made serial N-terminal truncations of SpoIIE starting at residue 320 and determined the oligomeric state of each by gel filtration (Figure 7B,C). We found that the 13-amino acid interval from residues 345 to 358 contained a feature required for multimerization; although SpoIIE truncated to residue 345 (SpoIIE345-827) multimerized, all truncations extending to 358 (SpoIIE358-827) and beyond did not (Figure 7C). This region encompasses K356, raising the possibility that K356D might block stabilization and activation of SpoIIE at the pole by blocking multimerization, and that T35I might suppress the phenotypes of K356D by restoring multimermization. Indeed, gel filtration experiments confirmed that SpoIIE320-827, K356D did not multimerize and that T353I partially restored multimerization (Figure 7C). Of all the SpoIIE variants analyzed, the K356D substitution uniquely blocked multimerization, highlighting a key role for the N-terminal region of the regulatory domain in mediating multimerization (Figure 7—figure supplement 1).

In sum, these biochemical experiments in conjunction with the in vivo experiments described above lead us to propose that multimerization of SpoIIE is the critical transition that stabilizes SpoIIE, enabling it to recognize the cell pole and leading to its activation as a phosphatase. First, we found that multimerization is required for stabilization of SpoIIE and compartmentalization of SpoIIE to the forespore (Figure 4B,C—figure supplement 1). Second, removal of TagSpoIIE uncoupled multimerization from degradation, and revealed an additional link between SpoIIE activation and multimerization (Figure 4D). Finally, we found that recognition of the cell pole by SpoIIE also depended on multimerization, even when SpoIIE was synthesized in vegetative cells uncoupled from degradation and σF activation (Figure 5B).

Discussion

A hallmark of sporulation is a process of asymmetric division that creates a septum near one randomly selected pole of the cell. Polar placement of the septum directs the protein phosphatase SpoIIE to activate σF in the resulting forespore. How SpoIIE activates σF at the right time and in the right place has been one of the enduring mysteries of this developmental system. As the end of the cell used for asymmetric division is chosen without regard to whether it is the old or new pole (Veening et al., 2008), the cues that SpoIIE interprets to achieve cell-specific activation of σF must arise de novo, that is, from the position of the septum rather than from preexisting asymmetry. Indeed, our results show that no feature of the sporulation process other than polar placement of the septum is necessary for compartmentalizing SpoIIE and for cell-specific activation of σF (Figure 6). In addition, as transcription of spoIIE commences prior to asymmetric division (Fujita and Losick, 2003) (Figure 2A), SpoIIE must also respond to temporal cues to ensure it is not active prior to the completion of cytokinesis.

Here we have provided evidence for a model in which SpoIIE leverages the asymmetric position of the septum to selectively associate with the adjacent cell pole of the forespore where it is stabilized and activated (Figure 7D). Three key features of our model are as follows:

  1. Capture at the cell pole. SpoIIE initially accumulates at the polar divisome and is handed off to the forespore pole following cytokinesis. Sequential transfer is enforced by preferential binding to the divisome over the cell pole, and we propose that selectivity for the forespore is achieved by the divisome being immediately adjacent to the forespore pole and that the forespore is largely derived from the pole. Additionally, SpoIIE not captured in the forespore would be sequestered at the distal divisome in the newly formed mother cell, preserving asymmetry.

  2. Spatially restricted proteolysis. SpoIIE is degraded by the AAA+ protease FtsH and is selectively stabilized in the forespore. This ensures that SpoIIE does not accumulate or become active prior to asymmetric division or in the mother cell following asymmetric division.

  3. Oligomerization. Polar recognition, protection from FtsH, and activation as a phosphatase are linked by a transition that takes place at the pole and is mediated by oligomerization.

Together these three features provide a simple mechanism for how cues derived from asymmetric cell division restrict SpoIIE to the forespore and couple σF activation to the completion of cytokinesis. At the same time our model raises several unanswered questions important both for understanding sporulation and diverse related biological systems.

How does SpoIIE localize to the cell pole and the divisome? Localization to the divisome depends on FtsZ and FtsA, the earliest assembling proteins to define the divisome (Levin et al., 1997). But whether SpoIIE interacts with these proteins directly, what features of SpoIIE mediate divisome association, and how SpoIIE influences FtsZ polymerization and divisome maturation are unknown. Answering these questions will help us to understand how SpoIIE is transferred from the divisome to the cell pole as well as how SpoIIE influences the position of the division septum. Similarly, localization to the pole depends on DivIVA, which directly senses the shape of the pole and acts as an organizing center for other pole-associated proteins (Lenarcic et al., 2009; Ramamurthi and Losick, 2009). But it is not known whether this interaction is direct or depends on an accessory protein.

How does oligomerization of SpoIIE promote σF activation? Genetic and biochemical evidence are consistent with a model in which stabilization, compartmentalization, and activation of SpoIIE are linked by oligomerization of SpoIIE molecules and that this oligomerization takes place in the forespore after asymmetric division. We cannot exclude the possibility, however, that oligomerization commences earlier in sporulation and that some other unrecognized feature of SpoIIE is additionally required for its transition to a stable and active state in the forespore. Structural information about the organization of SpoIIE oligomers and an in vivo assay for oligomerization may help distinguish between these possibilities and yield new insights into how it contributes to compartment specific σF activation.

How is activation of σF coordinated with the completion of asymmetric cell division? Our model proposes two mechanisms to prevent predivisional activation of σF: First, the features of the forespore (small size, high concentration of cell pole, and proximity to the divisiome) that promote SpoIIE stabilization and σF activation are all emergent properties that depend on completion of cell division. Second, competition between the divisome and cell pole for binding to SpoIIE prevents premature accumulation and activation of σF. Although there has been uncertainty about when SpoIIE is released from the divisome and when asymmetry in SpoIIE compartmentalization is established (Eswaramoorthy et al., 2014; Lucet et al., 2000; Wu et al., 1998), our time-lapse imaging and structured illumination microscopy indicate that SpoIIE constricts along with the FtsZ ring during cytokinesis (Figure 1, Figure 1—figure supplement 1). This is consistent with our model that SpoIIE remains sequestered at the divisome until cytokinesis is completed. In the future it will be important to determine just how association with the divisome prevents SpoIIE from oligomerizing and activating σF.

How is SpoIIE protected from degradation in the forespore? We have shown that the N-terminal tail of SpoIIE is necessary and sufficient for FtsH-dependent degradation (Figure 2E,F) and that stabilization in the forespore is mediated by features of SpoIIE that are required for interaction with the cell pole (Figure 3,4). Additionally, we have presented evidence that multimerization of SpoIIE is required for both stabilization and interaction with the pole. One possibility is that multimerization shields the TagSpoIIE from FtsH as depicted in Figure 7D. Alternatively, as FtsH has been shown to have weak unfoldase activity (Herman et al., 2003), multimerization might render SpoIIE resistant to FtsH unfolding and hence proteolysis. Finally, although we favor the view that SpoIIE is directly recognized by FtsH, it is conceivable that it requires an adaptor as is the case for some substrates of AAA+ proteases (Gottesman, 2003). If so, SpoIIE could be protected from degradation by negative regulation of the adaptor.

How is the phosphatase activity of SpoIIE regulated? Our genetic analysis provides clues for how activation occurs. We found that the V697A substitution locks SpoIIE in a high activity state in vitro, and restores σF activity in mutants defective for compartmentalization and σF activation. V697 is in an active site proximal loop (Levdikov et al., 2011); in many PP2C phosphatases this loop coordinates a third manganese ion that is critical for activity (Su et al., 2011). However, SpoIIE lacks the aspartate that coordinates this manganese, which could indicate that V697A locks the phosphatase in a conformation that compensates for the missing manganese ion. Additionally, we found that the stimulation of phosphatase activity (and binding to the pole) is genetically linked to multimerization: oligomerization, and σF activation were blocked by the substitution K356D and restored by T353I. We therefore speculate that multimerization induces a conformational change that organizes the catalytic center, compensating for the missing manganese and activating the phosphatase. A test of our hypothesis for a multimerization-dependent conformational change in the active site will require reconstituting multimerization-dependent activation of SpoIIE in vitro. Other PP2C phosphatases, such as the tumor suppressor protein PHLPP (Gao et al., 2005), similarly lack the aspartate to coordinate a third magnesium ion, suggesting that our speculation, if correct, could represent a more general regulatory mechanism for PP2C phosphatases.

In summary, we propose that the asymmetrically positioned division machinery – the de novo-generated source of asymmetry – positions SpoIIE to be captured at the adjacent cell pole, triggering σF-directed gene expression in the forespore. Capture at the pole, proteolytic stabilization and stimulation of the phosphatase all depend on oligomerization of SpoIIE. Thus, three interlinked regulatory events are sufficient to explain how SpoIIE exploits a stochastically generated spatial cue to the cell-specific activation of a transcription factor.

Materials and methods

Strains and strain construction

Request a detailed protocol

B. subtilis strains were constructed in PY79 using standard molecular genetic techniques (Harwood and Cutting, 1990). Full details of strain genotypes, and construction are provided in Supplementary file 1. For IPTG dependent expression (Plac), the hyperspank promoter was used (from pDR111a, gift of David Rudner), and for σF dependent expression (PσF), the spoIIQ promoter was used. Constructs were made by Gibson Assembly (New England Biolabs, Ipswitch, MA), and point mutations were introduced using QuikChange mutagenesis (Agilent Technologies).

Isolation of suppressors 

Request a detailed protocol

Suppressors of the spoIIE-K356D mutation were isolated by growing 100 ml cultures of strain RL5895 in DSM sporulation medium at 37°ºC for 28 hr. 11 ml of cells were heat killed at 80°C and used to re-inoculate a new 100 ml DSM culture. Heat killed cells from the second round culture were plated on DSM agar plates. Genomic DNA was prepared from the strain and retransformed to strain RL5875, lacking spoIIE, to confirm linkage to spoIIE, and the spoIIE locus was sequenced. Finally, suppressor mutations were then reconstructed by quickchange mutagenesis. Mutations T353I, V697A, V697F, and pseudorevertants K356T and K356Y were each isolated from multiple independent cultures, suggesting that the screen was near saturation.

Protein degradation and protein levels

Request a detailed protocol

Protein degradation rates were measured by shutting off translation by addition of chloramphenicol (100 µg/ml) to cultures. Samples were removed at indicated timepoints and immediately put on ice. Cells were lysed by mechanical disruption in a FastPrep (MP-BIO, Santa Ana, CA). Western blots were conducted by standard procedures and imaged on a BioRad ChemiDoc imager using chemiluminescence. Antibodies used were polyclonal anti-GFP (Rudner and Losick, 2002), polyclonal anti-σA (Fujita, 2000), polyclonal anti-DivIVA (Eswaramoorthy et al., 2014), and monoclonal anti-FLAG M2 (Sigma Aldrich, St. Louis, MO). Standards were used to determine linearity in each experiment. Immunoprecipitation of DivIVA was performed as published (Eswaramoorthy et al., 2014).

Fluorescence microscopy

Request a detailed protocol

All micrographs were acquired on an Olympus BX61 upright fluorescence microscope with a 100X objective, with the exception of timelapse images taken on a Nikon ti inverted microscope. Cells were immobilized on 2.5% agarose pads made with sporulation resuspension medium. To quantatitatively analyze micrographs, cells were segmented from phase images using either MicrobeTracker (Sliusarenko et al., 2011) or SupperSeggerOpti (Kuwada et al., 2015), and analyzed with custom MatLab scripts (scripts for quantitative image analysis are included as a Source code 1). Cell specificity of σF activity was determined by analyzing the distribution of PσF-CFP along the long axis of cells. SpoIIE localization profiles were calculated for each cell as the normalized ratio of SpoIIE-YFP to FM4-64 along the long axis of the cell. Because SpoIIE is a transmembrane protein, FM4-64 accounts for differences in membrane area along the cell axis. For sporulating cells that had undergone division, division septa were detected using FM4-64 and cells were oriented based on the position of the polar septum. Vegetative cells (strains RL5930 and RL5931) were induced to produce minicells by dilution from a log phase overnight culture to OD 0.05 and addition of 1mM IPTG, and 0.25% xylose as appropriate. Cells were imaged after 4 hr of growth at 37°C. Segmentation was performed based on phase images subtracted for FM4-64 to identify division septa in chained cells. Cells enriched for SpoIIE-GFP were defined as cells with average SpoIIE-GFP intensity greater than two standard deviations above the mean. Structured illumination microscopy was performed on a Zeiss Elyra microscope in the Harvard Center for Biological Imaging.

Biochemistry

Request a detailed protocol

SpoIIE was expressed as an N-terminal Sumo-6His fusion in BL21(DE3) cells following overnight induction with 0.5 mM IPTG at 14°C. Cells were lysed in 50 mM Tris pH 8.5, 200 mM NaCl, 1mM beta-mercaptoethanol and purified using HisTrap columns (GE Healthcare, Pittsburg, PA) eluting with a gradient of imidazole. The Sumo-6His tag was removed by cleavage with Ulp1 (Sumo Protease) followed by Ni-NTA subtraction. For velocity analytical ultracentrifugation, SpoIIE was dialyzed to 20 mM Tris pH8.5, 100 mM NaCl, 2 mM DTT overnight and data was collected at 280 nM spinning at 20,000 RPM at The Biophysical Instrumentation Facility (NSF-0070319) at MIT. Data were fit to a continuous model using SedFit (Schuck, 2000). Gel filtration was conducted on a 24 ml Superose 6 column (GE Healthcare, Pittsburg, PA), loading 100 µl of 1 µM SpoIIE. Phosphatase assays of soluble fragments of SpoIIE lacking the transmembrane domain (SpoIIE320-827 and SpoIIE320-827, V697A) were performed using 32P phosphorylated SpoIIAA (phosphorylated by purified SpoIIAB) as a substrate. Multiple turnover reactions were performed with 0.05 µM SpoIIE and varying concentrations of SpoIIAA-P as indicated. Dephosphorylation of SpoIIAA was detected by TLC chromatography on PEI-Cellulose plates developed in 1 M LiCl, 0.8 M Acetic acid.

References

    1. Gholamhoseinian A
    2. Piggot PJ
    (1989)
    Timing of spoII gene expression relative to septum formation during sporulation of bacillus subtilis
    Journal of Bacteriology 171:5747–5749.
    1. Gottesman S
    (2003)
    Proteolysis in bacterial regulatory circuits
    Annual Review of Cell and Developmental Biology 19:565–587.
  1. Book
    1. Harwood CR
    2. Cutting SM
    (1990)
    Molecular Biological Methods for Bacillus
    Chichester: Wiley-Blackwell.
    1. Levin PA
    2. Losick R
    (1994)
    Characterization of a cell division gene from bacillus subtilis that is required for vegetative and sporulation septum formation
    Journal of Bacteriology 176:1451–1459.
    1. Piggot PJ
    2. Coote JG
    (1976)
    Genetic aspects of bacterial endospore formation
    Bacteriological Reviews 40:908–962.
    1. Ramamurthi KS
    2. Losick R
    (2009) Negative membrane curvature as a cue for subcellular localization of a bacterial protein
    Proceedings of the National Academy of Sciences of the United States of America 106:13541–13545.
    https://doi.org/10.1073/pnas.0906851106
    1. Stragier P
    2. Losick R
    (1996)
    Molecular genetics of sportulation in Bacillus subtilis
    Annual Review of Genetics 30:297–341.

Decision letter

  1. Michael Laub
    Reviewing Editor; Massachusetts Institute of Technology, United States

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for submitting your work entitled "A Handoff Model for How Asymmetric Cell Division Triggers Cell-Specific Gene Expression in Bacillus subtilis" for peer review at eLife. Your submission has been evaluated by Michael Marletta (Senior Editor) and three reviewers, one of whom is a member of our Board of Reviewing Editors. The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision.

The reviewers each felt that this was a clearly written manuscript that provided some new insight in the mechanism by which SpoIIE activates σF in the forespore of developing B. subtilis. However, there were several major concerns about the conclusions and model proposed that limited enthusiasm for the paper. After extensive discussion between the reviewers, the consensus was that the paper is not suitable for eLife unless the model proposed can be further substantiated through additional experiments and analysis, as summarized below. If each of the issues below can be addressed, we would consider a revised manuscript, subject to an additional round of peer review.

Essential revisions:

1) The results in Figure 3 raised several questions. First, the fact that SpoIIE was apparently active in the mother cell simply by eliminating FtsH-dependent degradation implies that the proposed change in SpoIIE oligomerization during spore formation may not be so critical to its activation. Second, it was noted that the authors haven't shown whether the activation that occurs in mother cells depends on cell division. It has previously been reported that overexpressing SpoIIE bypasses the requirement for cell division (Arigoni et al., 1999), which, with other data, led the authors to suggest that compartmentalization of σF activity occurred via the exclusion of an inhibitor (perhaps FtsH) from the forespore, so testing dependence is critical. These two issues should be addressed.

2) Central to the model proposed is the notion that SpoIIE's oligomerization state changes following compartmentalization. The authors must show a change in oligomerization state in vivo during sporulation, e.g. by using sucrose gradients to directly monitor oligomerization.

3) It is odd that stabilizing the mutant SpoIIE variants by removing the N-terminal tag does not restore σF activity. If the only role of oligomerization is to protect SpoIIE from proteolysis, then removal of the degradation tag should bypass the need for oligomerization and restore σF activity, even if it's no longer compartment-specific. This result suggests that the point mutants have independent defects in activating the phosphatase domain, or that oligomerization itself is necessary for activation. These predictions can be tested in part by analyzing the oligomeric state of their mutants in vitro and by showing oligomerization-dependent protease activity in vitro. In sum, it was not clear whether oligomerization of SpoIIE was protecting it against FtsH-dependent degradation or activating phosphatase activity, or both.

4) The authors need to somehow demonstrate more directly that SpoIIE-YFP at the divisome is transferred to the nearby cell pole. Although the genetics and cell biology presented were consistent with this conclusion, a direct demonstration of the "handoff" being postulated is needed, perhaps by photoactivation and single particle tracking. Additionally, it wasn't clear whether SpoIIE was really being "handed off" or whether SpoIIE potentially just relocalizes from the septum to the pole via release, diffusion, and capture. The word "handoff" implies a more active mechanism, so the mechanism and wording should be clarified throughout the paper.

5) The authors need a more thorough discussion of the SpoIIE literature and an attempt to work towards an integrated model. This issue arose in two contexts. First: SpoIIE requires FtsZ to localize to the septum, but not DivIC. Interestingly, in a divICts mutant, there are no septa, but SpoIIE is active, producing dephosphorylated SpoIIAA, but apparently not at a high enough concentration to disrupt the SpoIIAB-σF complex (Carniol et al. 2004, Feucht 2002). How can these results be reconciled with the model here? Second: Figure 1, Figure 7D and the Introduction (fourth paragraph) state that SpoIIE constricts during cell division. This conflicts with recent data in Eswaramoorthy 2014, which used super resolution microscopy to demonstrate that SpoIIE does not constrict together with the divisome, but rather stays at the edge of the septal disk and that SpoIIE is biased towards the forespore side of the septum immediately after cell division and before relocalization to the pole.

Minor points:

1) Does the MalF-TM chimera accumulate properly in the forespore, activate σF, and support wild-type sporulation efficiency? These results would define the contribution of the transmembrane segments to SpoIIE function. If these results have been published previously, they should be mentioned in the Discussion.

2) The authors identified amino acid substitutions in the regulatory domain of SpoIIE which cause sporulation defects and defects in σF activation, but are not mislocalized. The authors should use these mutants to strengthen the prediction that oligomerization is required for SpoIIE localization in the forespore. In particular, mutants that cannot activate σF but are properly localized would be expected to oligomerize properly.

3) The SpoIIE variant C399A has very low SpoIIE levels, mislocalized SpoIIE protein, and nearly undetectable transcription driven by σF, yet it only has a ten-fold decrease in sporulation efficiency. The authors should discuss this apparent discrepancy in phenotypes.

4) Figure 1. In the absence of colocalization with FtsZ or with membrane staining, it is unclear how the SpoIIE localization pattern correlates with division during this timelapse.

5) The frequency with which σF activity is observed in single cells versus in the mother cell of sporangia should be scored and shown.

6) Figure 3. Many of the cells that have uncompartmentalized σF activity lack sporulation septa. Does activation of σF in this case depend on cell division? Or does stabilizing SpoIIE lead to σF activation in predivisional cells? The frequency of the various cell types should be shown.

7) Figure 3. The legend states that "the contrast [for ΔTag-SpoIIE] is approximately 5X brighter for SpoIIE-YFP." The authors must mean that the fluorescence intensity is 5X greater than SpoIIE-YFP.

8) The supplement for Figure 4 should include a few representative sporangia that are stained with FM4-64. The spore titers should be presented in standard scientific notation. 4.2 X108.

9) Figure 6B. The graphs showing the frequency with which σF is activated in minicells does not accurately convey the data. How many total minicells were scored? What fraction had IIE? σF activity? The data would likely best be presented in table format, so that the precise number of the various types of minicells counted is clear. If it is to be presented in a graphical format, the Y-axis should be labeled and the numbers of each cell type presented.

10) Additional images and quantification is necessary for the cell biological data, so that it is clear how often σF is activated before polar septation versus active in the mother cell.

11) Introduction, fourth paragraph: what does 'a central regulatory domain' mean? Is it a domain that regulates SpoIIE or that is involved in regulating SpoIIAA? It isn't clear at this point of the paper – eventually there is some evidence that emerges that this domain may affect multimerization status and hence activity of the phosphatase domain, but I think the authors should clarify much earlier.

12) In the subsection “SpoIIE is degraded by FtsH”, second paragraph: Stability of SpoIIE is monitored using a C-terminal FLAG tag. But this tagging could interfere with stability as protein termini often play critical roles in protease activity. This issue should be addressed by examining native SpoIIE stability somehow and by demonstrating the functionality of the tagged SpoIIE.

13) In the subsection “SpoIIE is degraded by FtsH”, end of third paragraph: What's the evidence that degradation is "likely" direct? Additional evidence is needed or the conclusion should end at SpoIIE being degraded in an FtsH-dependent manner. On a related note: the title of Figure 2 ("SpoIIE is degraded by FtsH") and comments in the Discussion (second and fifth paragraphs) should be adjusted given the data shown.

14) Figure 2C/E: Should be shown on a semi-log plot.

15) In the subsection “SpoIIE mutants blocked in compartmentalization and σF activation”, last paragraph: The evidence here that SpoIIE transitions from a hypothetical "active" state to an "inactive" state is weak. It could be that the compartmentalization defective mutants are also disrupted in phosphatase activity, but that phosphatase activity is constitutive. In other words, the mutants identified may prevent SpoIIE from being degraded and from being a proper phosphatase – I don't see the logic that SpoIIE must be transitioning from one state to another in terms of catalytic activity inside the forespore. The simplest test of the authors' idea is to examine in vitro the phosphatase activity of SpoIIE and the various compartmentalization-defective SpoIIE mutants.

16) In the subsection “An allele-specific suppressor of SpoIIEK356D”, first paragraph: The 'regulatory domain' still remains ill-defined at this stage of the manuscript and its function/role opaque to all but the B. subtilis aficionado.

17) In the subsection “An allele-specific suppressor of SpoIIEK356D”, first paragraph: As above (point number 15), it doesn't seem necessary to invoke the notion that SpoIIE's activity as a phosphatase is regulated. It could be that it is just the proteolysis of SpoIIE that is regulated with spore-specific stabilization.

18) In the subsection “An allele-specific suppressor of SpoIIEK356D”, last paragraph: Did the authors really test full-length SpoIIE phosphatase activity or a fragment of SpoIIE lacking the TM and potentially other domains? I would have guessed the latter, but the Methods section doesn't provide enough details. If it really is full-length protein, how are the multiple TM helices dealt with?

19) Figure 5B: Why does the very bright focus not show up in the line scan? More to the point: this bright focus seems to imply that ΔTag SpoIIE-YFP mainly accumulates away from the pole.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Asymmetric Division Triggers Cell-Specific Gene Expression Through Coupled Capture and Stabilization of a Phosphatase" for further consideration at eLife. Your revised article has been favorably evaluated by Michael Marletta (Senior Editor), a Reviewing Editor, and one reviewer. The manuscript has been greatly improved but there is still one remaining issue that needs to be addressed before acceptance, as outlined below:

The authors tried, at the reviewers’ request, to generate data supporting the notion that SpoIIE makes a change in oligomeric state during sporulation. As noted in their responses, this did not work and apparently both sucrose gradient fractionation and co-IPs from merodiploid strains indicated that SpoIIE is oligomeric, even in the presence of the K356D mutation. Two points here:

1) Did the authors examine oligomerization via sucrose gradients as a function of development to test whether oligomeric state changes, as postulated? This seems just as important to test as the K356D mutant. And if the experiment was done, it should probably be shown or at least discussed in the paper (see next point as well).

2) The paper, as written, leaves the reader with the feeling that the transition in oligomeric state is known with more certainty than it really is e.g. in the Discussion it states that 'We have presented evidence that multimerization of SpoIIE is a critical transformation'. While I definitely agree that there is some evidence in favor of this model, notably the in vitro studies of oligomerization by a truncated SpoIIE coupled with mutagenesis, there is also potentially evidence against this model or at least insufficient evidence (i.e. the sucrose fractionation experiments noted in the responses to the reviewers) to make a strong conclusion. I think the authors should be more explicit in the Discussion about what the evidence is and they should probably cite the outcome of their sucrose fractionation experiments, both with the K356D mutant and as a function of development. On this topic: I'm also still puzzled why the authors are postulating that the multimerization of SpoIIE occurs in the forespore, implying that cell division and compartmentalization are necessary for this transition, and yet mother cells that can't degrade SpoIIE but haven't yet divided can still show some activation of σF (Figure 3). Although predivisional cells only constitute 18% of the total cells showing σF activation in Figure 3, that result still implies, to me, that SpoIIE either multimerizes before septation in some cells or that it doesn't actually undergo a transition. Perhaps I'm being dense here, but this one aspect of the model still doesn't make sense to me and may need a bit of clarification and/or softening of the conclusions and statements about changes in multimerization.

https://doi.org/10.7554/eLife.08145.024

Author response

Essential revisions: 1) The results in Figure 3 raised several questions. First, the fact that SpoIIE was apparently active in the mother cell simply by eliminating FtsH-dependent degradation implies that the proposed change in SpoIIE oligomerization during spore formation may not be so critical to its activation.

Actually, we do conclude that oligomerization is required for both stabilization of SpoIIE and activation of σF. We have modified the text describing Figure 4D (subsection “An allele-specific suppressor of SpoIIEK356D”) and Figure 7C (“Discussion”) to make the basis for this conclusion clearer. Specifically, Figure 4D demonstrates that SpoIIE oligomerization is critical for σF activation even when degradation by FtsH is blocked. Thus, the requirement for oligomerization to stabilize SpoIIE in the forespore can be uncoupled from the requirement for oligomerization to activate σF.

Second, it was noted that the authors haven't shown whether the activation that occurs in mother cells depends on cell division. It has previously been reported that overexpressing SpoIIE bypasses the requirement for cell division (Arigoni et al., 1999), which, with other data, led the authors to suggest that compartmentalization of σF activity occurred via the exclusion of an inhibitor (perhaps FtsH) from the forespore, so testing dependence is critical. These two issues should be addressed.

We thank the reviewers for drawing our attention to the importance of the connection between cell division and σF activation and have added data and text that address this link. Specifically, we have added Figure 3D quantifying mis-activation of σF when FtsH mediated degradation is blocked. The figure shows mis-activation primarily in cells that have completed asymmetric cell division. This supports the conclusion that stabilization of SpoIIE partially uncouples σF activation from cell division. Partial uncoupling of σF activation from cell division by stabilizing SpoIIE is consistent with our model that sequestration of SpoIIE at the divisome prohibits premature oligomerization of SpoIIE and activation of σF. Thus, our model is consistent with the results of Arigoni et al., but rather than FtsH being excluded from the forespore (Figure 3—figure supplement 1), we find that competition between binding of SpoIIE to the divisome or the cell pole helps ensure proper timing of σF activation. We have also added additional text clarifying and expanding this point (subsection “SpoIIE mutants blocked in compartmentalization and σF activation”).

Additionally, we found that blocking cell division with a divICts mutation delays but does not block high σF activity (Author response image 1). A simple explanation for these results is that in most cells ∆Tag-SpoIIE is sequestered at the divisome, delaying σF activation until levels of ∆Tag-SpoIIE saturate the divisome and excess ∆Tag-SpoIIE can oligomerize and activate σF. Predivisional activation of σF could additionally occur in a subpopulation of cells where ∆tag-spoIIE is expressed prior to assembly of the divisome (as in Figure 5A).

2) Central to the model proposed is the notion that SpoIIE's oligomerization state changes following compartmentalization. The authors must show a change in oligomerization state in vivo during sporulation, e.g. by using sucrose gradients to directly monitor oligomerization.

This is the one challenge posed by the reviewers that we failed to address as we now explain. We have performed both sucrose gradient fractionation of SpoIIE and co-immunoprecipitation of SpoIIE from merodiploid strains with differentially tagged SpoIIE constructs. In both cases we detect oligomeric assemblies of SpoIIE even in the presence of the K356D mutation, preventing us from isolating K356D-dependent oligomers. These complexes could be indirect through SpoIIE association with macromolecular complexes such as the divisome, or could be mediated by the transmembrane domain of SpoIIE. Nevertheless, our in vitro studies of SpoIIE oligomerization combined with the unbiased isolation of a suppressor mutation that restored sporulation and oligomerization strongly suggest, we believe, a role for oligomerization in vivo. Furthermore, the additional experiments we performed dissecting the phenotype of the oligomerization mutant and its suppressor provide compelling evidence that oligomerization is required following the completion of asymmetric septation to interact with the cell pole, protect SpoIIE from degradation, and activate σF.

3) It is odd that stabilizing the mutant SpoIIE variants by removing the N-terminal tag does not restore σF activity. If the only role of oligomerization is to protect SpoIIE from proteolysis, then removal of the degradation tag should bypass the need for oligomerization and restore σF activity, even if it's no longer compartment-specific. This result suggests that the point mutants have independent defects in activating the phosphatase domain, or that oligomerization itself is necessary for activation. These predictions can be tested in part by analyzing the oligomeric state of their mutants in vitro and by showing oligomerization-dependent protease activity in vitro. In sum, it was not clear whether oligomerization of SpoIIE was protecting it against FtsH-dependent degradation or activating phosphatase activity, or both.

Based on the results in Figure 4D and Figure 5, we do conclude that oligomerization is required both for σF activation and for stabilization of SpoIIE from degradation by FtsH and have added text to make our reasoning on this point clearer. Responding to reviewer comments here and above in point 1, we now emphasize the connection between oligomerization and σF activation in our text (subsection “An allele-specific suppressor of SpoIIEK356D” and the Discussion). Additionally, we assessed the oligomerization state of other variants of SpoIIE and found that SpoIIEK356D (and truncated variants beginning after amino acid 356) was uniquely defective in oligomerization, consistent with our model that oligomerization is necessary but not sufficient for compartmentalization and activation of σF. These data have now been included as Figure 7—figure supplement 1. We additionally monitored the phosphatase activity of SpoIIE variants in vitro as detailed below in response to point 15.

4) The authors need to somehow demonstrate more directly that SpoIIE-YFP at the divisome is transferred to the nearby cell pole. Although the genetics and cell biology presented were consistent with this conclusion, a direct demonstration of the "handoff" being postulated is needed, perhaps by photoactivation and single particle tracking. Additionally, it wasn't clear whether SpoIIE was really being "handed off" or whether SpoIIE potentially just relocalizes from the septum to the pole via release, diffusion, and capture. The word "handoff" implies a more active mechanism, so the mechanism and wording should be clarified throughout the paper.

In brief, we do and did envision diffusion and capture when we used the word “hand off” and now provide additional data to support this. In light of the confusion it caused, we have added text to clarify this point and we no longer use handoff in the text. Further, and importantly, to demonstrate that SpoIIE is transferred from the divisome to the forespore, we have added Figure 5—figure supplement 1 demonstrating that forespore transcription of spoIIE is not required for compartmentalization. To block forespore transcription of spoIIE, we prevented forespore capture of the spoIIE gene, either by deletion of racA (to reduce the efficiency of chromosome capture in the forespore), or by transplanting the spoIIE gene to the terminus in cells harboring spoIIIE36 mutation (to prevent chromosome transport to the forespore).

5) The authors need a more thorough discussion of the SpoIIE literature and an attempt to work towards an integrated model. This issue arose in two contexts. First: SpoIIE requires FtsZ to localize to the septum, but not DivIC. Interestingly, in a divICts mutant, there are no septa, but SpoIIE is active, producing dephosphorylated SpoIIAA, but apparently not at a high enough concentration to disrupt the SpoIIAB-σF complex (Carniol et al. 2004, Feucht 2002). How can these results be reconciled with the model here?

To further address the link between cell division and σF we have added new data as described above and a paragraph to the Discussion (subsection”Fluorescence microscopy”). Specifically, and as in previous studies, we observe SpoIIE accumulation at the divisome but reduced activation of σF in cells mutant for divIC. This is consistent with IIE being largely sequestered at the divisome and not oligomerized at the cell pole. Carniol et al. (2004) saw partial dephosphorylation of SpoIIAA to levels similar to wt sporulating cells but much lower than SpoIIEV697A. A simple explanation for the lack of σF activation in the Carniol et al. (2004) study is that dephosphorylated SpoIIAA is concentrated in the forespore in wild-type cells, while in the divIC mutant there is still a significant amount of SpoIIAA-P present throughout the cell.

Second: Figure 1, Figure 7D and the Introduction (fourth paragraph) state that SpoIIE constricts during cell division. This conflicts with recent data in Eswaramoorthy 2014, which used super resolution microscopy to demonstrate that SpoIIE does not constrict together with the divisome, but rather stays at the edge of the septal disk and that SpoIIE is biased towards the forespore side of the septum immediately after cell division and before relocalization to the pole.

We have added Figure 1—figure supplement 1 to provide additional evidence of SpoIIE constriction with the divisome and have moved discussion of the constriction of SpoIIE along with the divisome to the Results section. Our results mostly agree nicely with and are complementary to Eswaramoorthy et al. (2014), and we build on their results by providing genetic evidence that association of SpoIIE with DivIVA is required for SpoIIE compartmentalization to the forespore. The major discrepancy as noted is with respect to the co-localization of SpoIIE and FtsZ. The main difference in our experimental approach is that we have taken time lapse images, while Eswaramoorthy et al. looked at static images. Because there is a lag between assembly of the divisome and constriction of the septum, and SpoIIE is released to the forespore while the FtsZ ring is disassembling, it is difficult to accurately stage the static images, which might have obscured the constriction of SpoIIE along with the FtsZ ring. To further document that SpoIIE constricts, we have taken additional time-lapse images of cells with the FtsZ ring and membrane labeled in addition to SpoIIE (Figure 1—figure supplement 1). From these images it is clear that SpoIIE constricts along with the FtsZ ring and is released after the completion of septation. Artifacts from the FM4-64 membrane stain can further complicate efforts to localize SpoIIE at the nascent septum. We imaged SpoIIE during sporulation by structured illumination microscopy, taking advantage of a fluorescent fusion protein to stain the membrane and mark the nascent septum. These images further supported our conclusion that SpoIIE constricts along with the nascent septum.

Minor points:1) Does the MalF-TM chimera accumulate properly in the forespore, activate σF, and support wild-type sporulation efficiency? These results would define the contribution of the transmembrane segments to SpoIIE function. If these results have been published previously, they should be mentioned in the Discussion.

The MalF-TM-SpoIIE chimera does not support wild-type sporulation efficiency and we have added these data to the table in Figure 4—source data 1. Previous studies (King et al. (1999), Carniol et al. (2004)) that used a MalF-TM-SpoIIE chimera had unknowingly removed the FtsH degradation tag. While stabilized MalF-TM-SpoIIE prematurely activates σF and prevents asymmetric cell division, MalF-TM-SpoIIE with the FtsH degradation tag is not stabilized in the forespore and fails to activate σF. Thus, the transmembrane domain plays a role in protecting SpoIIE from proteolysis in the forespore.

2) The authors identified amino acid substitutions in the regulatory domain of SpoIIE which cause sporulation defects and defects in σF activation, but are not mislocalized. The authors should use these mutants to strengthen the prediction that oligomerization is required for SpoIIE localization in the forespore. In particular, mutants that cannot activate σF but are properly localized would be expected to oligomerize properly.

As discussed above in response to essential revision 3, we have assessed the oligomerization of several additional mutants of SpoIIE and found that SpoIIEK356D was uniquely defective in oligomerization (Figure 7—figure supplement 1).

3) The SpoIIE variant C399A has very low SpoIIE levels, mislocalized SpoIIE protein, and nearly undetectable transcription driven by σF, yet it only has a ten-fold decrease in sporulation efficiency. The authors should discuss this apparent discrepancy in phenotypes.

The simplest explanation why the spoIIEC399A mutant has a less severe sporulation defect than other SpoIIE variants is that it supports a low level of σF activation. This is supported by a longer exposure of the western from Figure 4C, which shows a low but significant level of σF-directed expression of a reporter. Additionally, we note that assessment of sporulation efficiency by heat killing is highly sensitive, while assays for IIE levels and σF activity do not accurately report on events in a minor population of cells. Finally, C399A is part of a motif of four invariant cysteines in SpoIIE, and we observe synergistic effects of mutating multiple cysteines in tandem, suggesting that mutation of a single cysteine does not abolish function.

4) Figure 1. In the absence of colocalization with FtsZ or with membrane staining, it is unclear how the SpoIIE localization pattern correlates with division during this timelapse.

To further document that it does and in response to comment 5 above, we have taken new time lapse images colocalizing SpoIIE and FtsZ, which have been included as Figure 1—figure supplement 1 and as supplementary movies. We observe clear constriction of SpoIIE along with FtsZ and the septal membrane.

5) The frequency with which σF activity is observed in single cells versus in the mother cell of sporangia should be scored and shown.

These data have been added as Figure 3D.

6) Figure 3. Many of the cells that have uncompartmentalized σF activity lack sporulation septa. Does activation of σF in this case depend on cell division? Or does stabilizing SpoIIE lead to σF activation in predivisional cells? The frequency of the various cell types should be shown.

These data have been added as Figure 3D.

7) Figure 3. The legend states that "the contrast [for ΔTag SpoIIE] is approximately 5X brighter for SpoIIE-YFP." The authors must mean that the fluorescence intensity is 5X greater than SpoIIE-YFP.

The text of the legend has been clarified to indicate that the contrast has been equalized between the images for display purposes, but that we observe brighter fluorescence signal for ∆Tag-SpoIIE-YFP as expected from the results in Figure 3.

8) The supplement for Figure 4 should include a few representative sporangia that are stained with FM4-64. The spore titers should be presented in standard scientific notation. 4.2 X108.

We have modified the notation of sporulation efficiency in Figure 4—source data 1. Although we have imaged all of the spoIIE mutant strains presented in Figure 4—source data 1, we do not think that images of the defective sporangia would be informative beyond the average traces presented in Figure 4B. We have decided not to include this additional data but we would of course agree to do so if the reviewer considers it important.

9) Figure 6B. The graphs showing the frequency with which σF is activated in minicells does not accurately convey the data. How many total minicells were scored? What fraction had IIE? σF activity? The data would likely best be presented in table format, so that the precise number of the various types of minicells counted is clear. If it is to be presented in a graphical format, the Y-axis should be labeled and the numbers of each cell type presented.

We have added labels indicating the Y-axis scale to the histograms in Figure 6B and have modified the legend to note the total number of cells observed in each case.

10) Additional images and quantification is necessary for the cell biological data, so that it is clear how often σF is activated before polar septation versus active in the mother cell.

These data have been added as Figure 3D.

11) Introduction, fourth paragraph: what does 'a central regulatory domain' mean? Is it a domain that regulates SpoIIE or that is involved in regulating SpoIIAA? It isn't clear at this point of the paper – eventually there is some evidence that emerges that this domain may affect multimerization status and hence activity of the phosphatase domain, but I think the authors should clarify much earlier.

The central regulatory domain is defined physically from sequence analysis (as distinct from the transmembrane domain to the N-terminus and the PP2C phosphatase domain to the C-terminus), and functionally from genetic evidence supporting a role in regulating SpoIIE function. We have modified the text to clearly define our use of the term “regulatory domain” in the Introduction.

12) In the subsection “SpoIIE is degraded by FtsH”, second paragraph: Stability of SpoIIE is monitored using a C-terminal FLAG tag. But this tagging could interfere with stability as protein termini often play critical roles in protease activity. This issue should be addressed by examining native SpoIIE stability somehow and by demonstrating the functionality of the tagged SpoIIE.

We used tagged SpoIIE for degradation experiments due to the fact that C-terminally tagged variants of SpoIIE support sporulation to levels indistinguishable from wild-type (these sporulation efficiency numbers are now included in Figure 4—source data 1), and YFP tagged SpoIIE is compartmentalized in the forespore. Additionally, native SpoIIE antibodies are of poor quality. In response to the reviewers’ suggestion, we conducted degradation experiments with untagged SpoIIE, and an affinity purified SpoIIE antibody. The results are similar to those seen with the tagged protein, and we detected degradation of SpoIIE that was largely dependent on FtsH. These data have been added to Figure 2—figure supplement 1.

13) In the subsection “SpoIIE is degraded by FtsH”, end of third paragraph: What's the evidence that degradation is "likely" direct? Additional evidence is needed or the conclusion should end at SpoIIE being degraded in an FtsH-dependent manner. On a related note: the title of Figure 2 ("SpoIIE is degraded by FtsH") and comments in the Discussion (second and fifth paragraphs) should be adjusted given the data shown.

We agree that “likely” is too strong and have changed the wording throughout the manuscript, and have modified the title to Figure 2. In the Discussion we present as a model that FtsH degrades SpoIIE, and for simplicity and readability we have left this unchanged.

14) Figure 2C/E: Should be shown on a semi-log plot.

Although plotting exponential decay on semi-log axes leads to a linear fit, the linear axes are more intuitive to the general reader, and so we prefer to leave the plots as they are. We have included semi-log plots of the data here for reference.

15) In the subsection “SpoIIE mutants blocked in compartmentalization and σF activation”, last paragraph: The evidence here that SpoIIE transitions from a hypothetical "active" state to an "inactive" state is weak. It could be that the compartmentalization defective mutants are also disrupted in phosphatase activity, but that phosphatase activity is constitutive. In other words, the mutants identified may prevent SpoIIE from being degraded and from being a proper phosphatase – I don't see the logic that SpoIIE must be transitioning from one state to another in terms of catalytic activity inside the forespore. The simplest test of the authors' idea is to examine in vitro the phosphatase activity of SpoIIE and the various compartmentalization-defective SpoIIE mutants.

We acknowledge here and in the text that we do not have direct evidence of activation of SpoIIE phosphatase activity. However, this model is consistent with previous evidence (Feucht et al. 2002) and our analysis of the SpoIIEV697A variant. We have measured the in vitro phosphatase activity of a significant set of IIE mutants that do not support σF activity in vivo and did not detect defects in phosphatase activity for any mutants. This was true for both SpoIIE variants such as K356D that are destabilized and not compartmentalized and for variants such as Q483A that are compartmentalized properly. Importantly mutation of a residue (D628A) that coordinates manganese in the active site abolished phosphatase activity. We suspect that our in vitro phosphatase assay lacks a critical factor for SpoIIE activation (such as DivIVA), and that we are monitoring a low basal level of phosphatase activity. This is further supported by the unphysiologically high manganese concentrations required to detect phosphatase activity for SpoIIE (Km=2mM). It is a major future goal to reconstitute SpoIIE activation in vitro.

16) In the subsection “An allele-specific suppressor of SpoIIEK356D”, first paragraph: The 'regulatory domain' still remains ill-defined at this stage of the manuscript and its function/role opaque to all but the B. subtilis aficionado.

We have modified the text to clearly define our use of the term “regulatory domain” in the Introduction.

17) In the subsection “An allele-specific suppressor of SpoIIEK356D”, first paragraph: As above (point number 15), it doesn't seem necessary to invoke the notion that SpoIIE's activity as a phosphatase is regulated. It could be that it is just the proteolysis of SpoIIE that is regulated with spore-specific stabilization.

See response above to point 15.

18) In the subsection “An allele-specific suppressor of SpoIIEK356D”, last paragraph: Did the authors really test full-length SpoIIE phosphatase activity or a fragment of SpoIIE lacking the TM and potentially other domains? I would have guessed the latter, but the Methods section doesn't provide enough details. If it really is full-length protein, how are the multiple TM helices dealt with?

As noted in the Methods and the legend to Figure 4—figure supplement 1, phosphatase assays were performed with soluble fragments of SpoIIE encompassing amino acids 320-827. We have added a further note to the Methods section to clarify this.

19) Figure 5B: Why does the very bright focus not show up in the line scan? More to the point: this bright focus seems to imply that ΔTag-SpoIIE YFP mainly accumulates away from the pole.

The line scan presented in Figure 5B is an average trace of SpoIIE-YFP fluorescence normalized to FM4-64 from all cells in a population, while the image above is of a representative cell. Some, but not all cells have bright spots of SpoIIE fluorescence that coincide with bright spots of FM4-64 staining. It is not uncommon to observe such artifactual puncta of FM4-64 staining, particularly when cells are stressed, (as they are when cell division is blocked).

[Editors' note: further revisions were requested prior to acceptance, as described below.]

The authors tried, at the reviewers’ request, to generate data supporting the notion that SpoIIE makes a change in oligomeric state during sporulation. As noted in their responses, this did not work and apparently both sucrose gradient fractionation and co-IPs from merodiploid strains indicated that SpoIIE is oligomeric, even in the presence of the K356D mutation. Two points here:1) Did the authors examine oligomerization via sucrose gradients as a function of development to test whether oligomeric state changes, as postulated? This seems just as important to test as the K356D mutant. And if the experiment was done, it should probably be shown or at least discussed in the paper (see next point as well).

Indeed, we did examine oligomerization as a function of developmental stage using sucrose gradients (comparing time points during sporulation and genetic perturbations to arrest development at various stages). Due to the technical limitations mentioned above, these experiments were uninformative either way. As described below, we have added a paragraph that to the Discussion that addresses the uncertainty in when oligomerization commences.

2) The paper, as written, leaves the reader with the feeling that the transition in oligomeric state is known with more certainty than it really is e.g. in the Discussion it states that 'We have presented evidence that multimerization of SpoIIE is a critical transformation'. While I definitely agree that there is some evidence in favor of this model, notably the in vitro studies of oligomerization by a truncated SpoIIE coupled with mutagenesis, there is also potentially evidence against this model or at least insufficient evidence (i.e. the sucrose fractionation experiments noted in the responses to the reviewers) to make a strong conclusion. I think the authors should be more explicit in the Discussion about what the evidence is and they should probably cite the outcome of their sucrose fractionation experiments, both with the K356D mutant and as a function of development.

In our revised manuscript we have added a paragraph to the Discussion (fourth paragraph) that directly discusses these issues and have changed the language in the seventh paragraph. In particular, we discuss an alternative model in which oligomerization commences earlier in sporulation and state that we cannot distinguish the models without an in vivo assay for oligomerization.

On this topic: I'm also still puzzled why the authors are postulating that the multimerization of SpoIIE occurs in the forespore, implying that cell division and compartmentalization are necessary for this transition, and yet mother cells that can't degrade SpoIIE but haven't yet divided can still show some activation of σF (Figure 3). Although predivisional cells only constitute 18% of the total cells showing σF activation in Figure 3, that result still implies, to me, that SpoIIE either multimerizes before septation in some cells or that it doesn't actually undergo a transition. Perhaps I'm being dense here, but this one aspect of the model still doesn't make sense to me and may need a bit of clarification and/or softening of the conclusions and statements about changes in multimerization.

I think the reviewers are overlooking the fact that SpoIIE in these cells accumulates to high levels because we have removed the degradation tag. Figure 5A shows that when stabilized, SpoIIE accumulates at the poles of predivisional cells, which should lead to predivisional activation of σF. Although we cannot make a direct comparison (because σF activation perturbs SpoIIE localization), the frequency with which we observe polar SpoIIE is consistent with the fraction of predivisional cells with σF activity.

https://doi.org/10.7554/eLife.08145.025

Article and author information

Author details

  1. Niels Bradshaw

    Department of Molecular and Cellular Biology, Harvard University, Cambridge, United States
    Contribution
    NB, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    No competing interests declared.
  2. Richard Losick

    Department of Molecular and Cellular Biology, Harvard University, Cambridge, United States
    Contribution
    RL, Conception and design, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    losick@mcb.harvard.edu
    Competing interests
    RL: Senior Editor, eLife.

Funding

National Institutes of Health (GM18568)

  • Richard Losick

Damon Runyon Cancer Research Foundation (DRG 2051-10)

  • Niels Bradshaw

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We dedicate this article to the memory of Patrick J Piggot. We thank T Wilkinson, and I Barak for collaboration and discussions throughout this work, K Ramamurthi and P Eswaramoorthy for discussions and reagents, L Shapiro and J Kardon for valuable comments during manuscript preparation, P Wiggins for code and assistance using SuperSeggerOpti, and A Leech and D Pheasant for assistance with analytical ultracentrifugation.

Reviewing Editor

  1. Michael Laub, Massachusetts Institute of Technology, United States

Version history

  1. Received: April 16, 2015
  2. Accepted: October 13, 2015
  3. Accepted Manuscript published: October 14, 2015 (version 1)
  4. Version of Record published: December 17, 2015 (version 2)
  5. Version of Record updated: April 25, 2017 (version 3)

Copyright

© 2015, Bradshaw 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

  • 2,751
    Page views
  • 551
    Downloads
  • 26
    Citations

Article citation count generated by polling the highest count across the following sources: Scopus, PubMed Central, Crossref.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

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)

  1. Niels Bradshaw
  2. Richard Losick
(2015)
Asymmetric division triggers cell-specific gene expression through coupled capture and stabilization of a phosphatase
eLife 4:e08145.
https://doi.org/10.7554/eLife.08145

Further reading

    1. Computational and Systems Biology
    2. Microbiology and Infectious Disease
    Vanessa Dumeaux, Samira Massahi ... Michael T Hallett
    Research Article Updated

    Candida albicans, an opportunistic human pathogen, poses a significant threat to human health and is associated with significant socio-economic burden. Current antifungal treatments fail, at least in part, because C. albicans can initiate a strong drug tolerance response that allows some cells to grow at drug concentrations above their minimal inhibitory concentration. To better characterize this cytoprotective tolerance program at the molecular single-cell level, we used a nanoliter droplet-based transcriptomics platform to profile thousands of individual fungal cells and establish their subpopulation characteristics in the absence and presence of antifungal drugs. Profiles of untreated cells exhibit heterogeneous expression that correlates with cell cycle stage with distinct metabolic and stress responses. At 2 days post-fluconazole exposure (a time when tolerance is measurable), surviving cells bifurcate into two major subpopulations: one characterized by the upregulation of genes encoding ribosomal proteins, rRNA processing machinery, and mitochondrial cellular respiration capacity, termed the Ribo-dominant (Rd) state; and the other enriched for genes encoding stress responses and related processes, termed the Stress-dominant (Sd) state. This bifurcation persists at 3 and 6 days post-treatment. We provide evidence that the ribosome assembly stress response (RASTR) is activated in these subpopulations and may facilitate cell survival.

    1. Immunology and Inflammation
    2. Microbiology and Infectious Disease
    Taylor J Abele, Zachary P Billman ... Edward A Miao
    Research Article

    Pyroptosis and apoptosis are two forms of regulated cell death that can defend against intracellular infection. When a cell fails to complete pyroptosis, backup pathways will initiate apoptosis. Here, we investigated the utility of apoptosis compared to pyroptosis in defense against an intracellular bacterial infection. We previously engineered Salmonella enterica serovar Typhimurium to persistently express flagellin, and thereby activate NLRC4 during systemic infection in mice. The resulting pyroptosis clears this flagellin-engineered strain. We now show that infection of caspase-1 or gasdermin D deficient macrophages by this flagellin-engineered S. Typhimurium induces apoptosis in vitro. Additionally, we engineered S. Typhimurium to translocate the pro-apoptotic BH3 domain of BID, which also triggers apoptosis in macrophages in vitro. During mouse infection, the apoptotic pathway successfully cleared these engineered S. Typhimurium from the intestinal niche but failed to clear the bacteria from the myeloid niche in the spleen or lymph nodes. In contrast, the pyroptotic pathway was beneficial in defense of both niches. To clear an infection, cells may have specific tasks that they must complete before they die; different modes of cell death could initiate these ‘bucket lists’ in either convergent or divergent ways.