ShfP is a member of a vast radiation of bacterial cell-surface calcineurin-like phosphoesterases

ShfP is a 1289 amino acid–secreted protein with an N-terminal signal peptide followed by nine globular domains (Fig. 2A). The first of these is a lamin N-terminal domain (LTD), prototyped by the globular DNA binding and intranuclear interaction domain of the animal nuclear envelope lamins (23). This is followed by seven immunoglobulin (Ig)–like β-sandwich domains and a calcineurin-like phosphoesterase domain inserted into the seventh Ig domain. Other than nuclear lamins, the LTD was previously reported as fused to different phosphoesterase catalytic domains in several bacterial cell-surface proteins and, based on its animal counterpart, is predicted to bind extracellular DNA or chemically analogous cell-surface polysaccharides (23). The Ig domains are similarly predicted to bind cell-surface carbohydrates or mediate adhesion via interaction with other proteins. The calcineurin-like domain of ShfP belongs to a vast radiation of such domains in secreted/membrane-anchored proteins that are found across the bacterial superkingdom (Fig. 2A and fig. S1A). These versions are typically distinguished by multiple fusions to carbohydrate-binding and cell-surface adhesion domains which, in addition to LTD and Ig, include domains such as the concanavalin lectin–like laminin G, fibronectin type III (FN3), S-layer homology, discoidin-like, cadherin, cell-wall binding, and the PQQ β propeller (Fig. 2A). They may also be further fused to other catalytic domains, such as the ZU5 autopeptidase, which is widely represented in proteolytically processed cell-surface proteins, and the phosphodiester glycosidase (Pfam: NAGPA).

Open in a separate window

Fig. 2.

Phylogenetic and domain analysis of ShfP.

(A) A phylogenetic tree of a representative set of extracellular calcineurin-like domains from the homologs of the ShfP family is shown. Select domain architectures of proteins are shown near the branch of the tree in which they occur. Numbers indicate the location on the phylogenetic tree of the selected architectures. The branches are colored according to the phylum as shown in the legend; the branch with the B. subtilis ShfP and its orthologs is highlighted and shown with thicker lines. The color of the circles at the end of each branch denotes the most common domain in that branch as indicated in the legend, while a gray circle denotes other domain associations. Bootstrap values are shown as a star at each node as grayscale (black = 100%). PQQ, version of the Beta-Propeller; FN3 shown here is a specialized version mapping on to the Pfam model Pur_ac_phosph_N; CW, Pfam model CW_binding (cell wall binding domain); CAD, cadherin; SIG, signal peptide; TM, transmembrane helix; LTD, lamin tail domain; IG, immunoglobulin domain; 7TMR, 7TM receptors with diverse intracellular signaling modules extracellular domain 1; ZU5, ZU5 peptidase; NAGPA, N-acetylglucosamine-1-phosphodiester alpha-N-acetylglucosaminidase; SLH, S-layer homology domain; ATRXPA-N, Anthrax protective antigen glycosidase N-terminal domain. (B) Alphafold2-predicted structure of the B. subtilis ShfP calcineurin-like domain with the active sites shown as ball and stick models and the sequence logos of the five conserved motifs of the calcineurin-like domain derived from a multiple sequence alignment of a representative set of ShfP homologs. Letters represent amino acid abbreviations; the height of each letter represents the bit-score of conservation among homologs of the family. (C) Alphafold2-predicted structure of B. subtilis ShfP. Gold, Ig domains; blue, LTD domain; green and purple, calcineurin like domain.

The calcineurin-like superfamily comprises universally distributed phosphoesterase domains that feature a four-layered α/β sandwich fold formed from two repeat units of the IF3-C domain (Fig. 2, B and C) (24). It displays five conserved motifs that bind two divalent metal ions [usually Zn(II) or Fe(II)] that catalyze phosphoester hydrolysis. Members of this superfamily act on a variety of organophosphate substrates, including nucleic acids, nucleotides, phosphate-containing lipid head groups, and phosphorylated serines and threonines in proteins (24). A B. subtilis member of one clade of calcineurin-like phosphoesterases (PhoD) has been previously shown to degrade the phosphodiester linkages in cell-surface teichoic acids that feature glycerol phosphate or ribitol phosphate (25). On the basis of this precedent and the multiplicity of fusions to potential carbohydrate, adhesion, and DNA binding domains, we predict that ShfP is likely to adhere to the cell-surface macromolecular matrix and catalyze the hydrolysis of phosphoester linkages in one of its components such as (lipo)teichoic acid, eDNA, or a phospholipid head group.

ShfP is a forespore intermembrane protein requiring the calcineurin-like and adhesion domains for function

The upstream region of the shfP gene harbors a putative binding site for σ^G, a sporulation-specific transcription factor that is active exclusively in the forespore (Fig. 1D) (26). To test whether shfP is expressed in the forespore during sporulation, the shfP upstream sequence was fused to superfolder–green fluorescent protein (sfGFP) and sfGFP production was visualized by epifluorescence microscopy. We detected fluorescence signal inside the forespore 4 hours following sporulation induction, consistent with the timing and location of σ^G-mediated activation (Fig. 1, E and E′). In addition, deletion of σ^G or sporulation sigma factors (σ^F and σ^E) that are activated earlier than σ^G abrogated the forespore-specific production of GFP (Fig. 1, F to H′), while deletion of the downstream sigma factor (σ^K) did not (Fig. 1, I and I′). Thus, unlike shfA, which is expressed earlier in the mother cell by σ^E (19, 20), shfP is a sporulation gene that is expressed later in the forespore by σ^G. We then fused ShfP to sfGFP and determined its localization during sporulation. Consistent with its predicted signal peptide, 4 hours after induction of sporulation, ShfP-sfGFP localized to the forespore periphery in an otherwise WT cell (Fig. 1, J and J′), and deletion of shfA did not disrupt the localization of ShfP-sfGFP (Fig. 1, K and K′).

We next tested whether the predicted phosphoesterase activity of ShfP is required for function by disrupting conserved active site residues in the calcineurin-like phosphoesterase domain (Fig. 2B). Substituting D927, D963, NHD997-999, H1080, or H1114 and H1116 with Ala did not diminish steady-state levels of ShfP (fig. S1B) but resulted in a partial or total failure of these variants to complement the ΔshfP strain (Fig. 1B, lanes 6 to 10), suggesting that the predicted ShfP phosphoesterase activity is needed for sporulation inhibition. In addition, deleting the LTD displayed a partial defect (Fig. 1B, lane 12), consistent with the predicted role of this region in anchoring ShfP to the cell surface.

The localization pattern of ShfP-sfGFP and the presence of a signal peptide (Fig. 2C) led us to wonder whether ShfP could reside in the intermembrane macromolecular matrix surrounding the forespore (Fig. 1A, “I.M.”). To discern whether ShfP resides between the two membranes, which are spaced ~50 nm apart, we used a super-resolution fluorescence microscopy technique called DNA-PAINT (27). To visualize the membranes, we produced a small membrane–bound amphipathic α helix fused to mCherry before the induction of sporulation and detected this fusion using antibodies directed against mCherry (28). In strains producing this fusion, we coproduced ShfP fused to GFP and detected the fusion using antibodies directed against GFP. Using this technique, the forespore envelope appeared to have >2-fold increase in thickness relative to the membrane surrounding the mother cell, consistent with the presence of two membranes (Fig. 1, L and M, red), and the ShfP-GFP signal (Fig. 1, L and M, cyan) localized as puncta at the center of the fluorescence from the membrane stain at the forespore periphery, consistent with an intermembrane space residence of ShfP. In contrast, SpoIVA-GFP, which is produced in the mother cell and resides on the surface of the outer forespore membrane (29), visualized using the same technique, displayed a biased localization away from the center of the membrane fluorescence (Fig. 1, N and O). Consistent with the intermembrane localization pattern, production of ShfP that did not harbor a signal peptide failed to complement the deletion of shfP (Fig. 1B, lane 11, and fig. S1B). Together, we conclude that ShfP is a sporulation-specific phosphoesterase that is produced in the forespore and translocated into the forespore intermembrane macromolecular matrix where it is likely anchored to the inner forespore membrane.

ShfP produces a diffusible extracellular molecule that inhibits sporulation against which ShfA provides immunity

Spo0A is the master regulator for the entry into sporulation (30). To study the effect of the ShfA-ShfP pathway, we monitored progression through sporulation using a strain that produces GFP under control of a promoter (P_spoIIE) that is activated when levels of phosphorylated Spo0A reach an adequate threshold that is needed to drive the sporulation program (10, 31). We then quantified GFP fluorescence using flow cytometry to monitor individual cells that had initiated sporulation. In an unsynchronized culture, 63 ± 15% of WT cells displayed Spo0A activation (Fig. 3A, “asynchronous”) in early stationary phase, but the fraction of cells displaying Spo0A activation in ΔshfA cells was reduced at this time point (18 ± 16%), although shfA is only expressed after, and is ultimately dependent on, Spo0A itself. This reduction in Spo0A activation was ameliorated in both ΔshfP and ΔshfA ΔshfP strains (Fig. 3A, “asynchronous”).

Open in a separate window

Fig. 3.

ShfP⁺ cells release an extracellular factor to inhibit late-sporulating cells from entering the sporulation pathway.

(A) Percent of Spo0A-positive cells (in cells harboring P_spoIIE-gfp: GFP produced under spo0A control), determined by flow cytometry, in otherwise WT, shfA, shfP, or shfA shfP cells in (left) asynchronous or (right) synchronous cultures. (B and C) Percent of Spo0A-positive shfA shfP cells harboring the P_spoIIE-gfp reporter, determined by flow cytometry, (B) when cocultured with the indicated strain, or (C) when induced to sporulate with cell-free culture supernatant harvested at t = 5 hours after induction of sporulation from the indicated strain. Strains: MF277, CC218, CC219, and CC220. (D and E) Sporulation efficiencies of: (D) shfA shfP cells induced to sporulate in fresh medium (lane 1) or cell-free culture supernatant harvested at the indicated times after induction of sporulation from WT (lanes 2 to 4) or shfP cells (lanes 5 to 7); (E) indicated recipient strains (shfA shfP or shfP) were induced to sporulate for the indicated times, whereupon the cultures were centrifuged and supernatant was replaced with cell-free culture supernatant harvested at the indicated times after induction of sporulation from WT cells. (F) Sporulation efficiency of cells expressing shfP under control of σ^G (lanes 3 and 4) or shfP and catalytically inactive shfP alleles under control of σ^F (lane 5 to 7). Strains: PY79, TD517, TD507, TU36, TU40, TU41, CC134, and CC160.

To explain how deletion of a late sporulation gene can affect activation of an early gene in the pathway, we hypothesized that ShfP, produced by cells that entered sporulation early, somehow signals to cells that have not yet entered sporulation to delay their entry into sporulation. In this model, the presence of ShfA in early-sporulating (ShfP-producing) cells provides immunity from this ShfP-mediated sporulation delay. Since this model demands asynchronous entry into sporulation, we tested it by repeating the experiment with the Spo0A reporter strains, but instead induced sporulation using the resuspension method (32), which synchronizes cells to enter sporulation more uniformly. In a synchronized population of cells, the Spo0A activation defect of the ΔshfA strain was largely eliminated (Fig. 3A, “synchronous”), which was consistent with a model in which early-sporulating cells actively delay sporulation entry in cells that had not yet activated Spo0A.

To directly test this model, we devised a coculture system that contained two different strains: ΔshfA ΔshfP cells harboring the P_spoIIE-gfp reporter to assess the entry into sporulation in a sporulation-competent strain (Fig. 3A, lane 4) that did not harbor the ShfA-ShfP pathway, and different cocultured strains that could influence the strain harboring the reporter. We then analyzed Spo0A activation using flow cytometry. Two hours past the onset of stationary phase, coculturing the ΔshfA ΔshfP cells with strains that produced ShfP (WT, ΔshfA) resulted in an Spo0A activation defect (Fig. 3B, lanes 1 and 2), whereas coculturing the ΔshfA ΔshfP cells with strains that did not produce ShfP (ΔshfP, ΔshfA ΔshfP) permitted Spo0A activation in the ΔshfA ΔshfP reporter strain (Fig. 3B, lanes 3 and 4). To test whether this ShfP influence in trans is mediated by a soluble extracellular factor or cell-cell contact, we repeated the experiment using cell-free supernatant produced by different donor strains instead of coculturing. We harvested cell-free supernatants of cultures of various donor strains 5 hours after induction of synchronous sporulation and induced sporulation of the recipient ΔshfA ΔshfP harboring the P_spoIIE-gfp reporter using these supernatants. After 2.5 hours, we analyzed Spo0A activation in the recipient strain using flow cytometry. Supernatants harvested from cells producing ShfP (WT, ΔshfA) repressed the activation of Spo0A (Fig. 3C, lanes 1 and 2) relative to those harvested from cells that did not produce ShfP (ΔshfP, ΔshfA ΔshfP; Fig. 3C, lanes 3 to 4), indicating the presence of an ShfP-produced soluble sporulation-delaying factor.

To test the timing of production of the soluble molecule, we harvested culture supernatants from WT and ΔshfP cells at 4, 4.5, and 5 hours after induction of synchronized sporulation, lyophilized the sample, concentrated the material ~4-fold by resuspending in fresh sporulation medium, induced sporulation of the ΔshfA ΔshfP recipient strain using this supplemented medium, and measured sporulation efficiency. Cells that were resuspended in material harvested 4 hours after induction of either WT or ΔshfP exhibited a modest decrease in sporulation (Fig. 3D). However, the cells resuspended in material harvested at later time points (coincident with the timing of σ^G-activated expression of shfP), exclusively from WT cells exhibited ~1000-fold reduction in sporulation. Conversely, we induced synchronous sporulation in ΔshfA ΔshfP and ΔshfP strains and, at different time points, replaced the culture supernatants with those derived from WT (Fig. 3E) or ΔshfA ΔshfP (fig. S2) cells at 5 hours (which contains the ShfP-generated product in the WT supernatant). At initial time points, addition of the WT supernatant did not alter sporulation efficiency, but replacing the supernatant any time after t = 3 hours (which coincides with the expression of shfA) resulted in decreased sporulation efficiency in the ΔshfA ΔshfP, but not in the ΔshfP strain (which produces ShfA; Fig. 3E). However, addition of culture supernatant from ΔshfA ΔshfP cells did not alter sporulation efficiencies of the recipient cells at any tested time point (fig. S2). Thus, ShfA activity manifests 3 hours after the induction of sporulation, whereas ShfP activity occurs 5 hours after sporulation induction. We next tested the effect of expressing shfP before shfA by cloning the shfP gene under the control of an earlier forespore-specific transcription factor (σ^F). Complementing the ΔshfP strain with σ^F-expressed shfP (which is before shfA expression) resulted in reduced sporulation efficiency (Fig. 3F, lane 5), whereas expressing the defective shfP^H1080A or shfP^D927A alleles earlier under σ^F control (Fig. 3F, lanes 6 and 7) or expressing shfP at the proper time and compartment under control of a different σ^G promoter (Fig. 3F, lane 4) did not affect sporulation. The data are therefore consistent with a requirement for the earlier production of ShfA to provide a protective benefit against the deleterious effects of the ShfP-generated diffusible product.

Before committing to the sporulation pathway, cells that activate Spo0A early secrete at least two toxins that kill nonsporulating cells (13, 14). To test whether the ShfA-ShfP pathway operates via this cannibalism-dependent pathway, we deleted both pathways and measured sporulation efficiency. Deleting shfA resulted in a sporulation defect that was not corrected by deleting genes that mediate either or both cannibalism pathways (sdpABC and skfABCDEF), whereas deleting the cannibalism pathway alone did not disrupt sporulation (fig. S3A). We therefore conclude that the ShfA-ShfP pathway functions independently of the previously described cannibalism pathway.

ShfP generates extracellular glycerol, which delays entry into sporulation

To identify the ShfP-produced diffusible factor, we sought to purify the sporulation-delaying activity. Separation of cell-free culture supernatants generated by either WT or ΔshfP cultures using size exclusion chromatography revealed several differences in detected peaks (Fig. 4A, top). One fraction (#19) displayed sporulation inhibition activity only in supernatants derived from WT cultures, but not ΔshfP cultures (Fig. 4A, bottom). To determine the structure of the released molecule, we lyophilized the active fraction and used nuclear magnetic resonance (NMR) and mass spectrometry (MS). Analysis of ¹H and ¹³C NMR spectra, along with two-dimensional (2D) correlation spectra, showed signals that could only be assigned to glycerol in this subfraction (Fig. 4, B and C), and the mass spectrum showed a molecular weight of 93, consistent with glycerol. Last, we compared the spectroscopic data with a glycerol standard. The retention time of the active compound and the standard sample (1 mg/ml glycerol) were the same in the extracted ion chromatogram, with both peaks showing a positive mass ion peak for glycerol (Fig. 4D). Quantification of the amount of extracellular glycerol in the sporulation medium 5 hours after the induction of sporulation using a coupled enzymatic reaction revealed that WT cells produce 1.22 ± 0.07 mM glycerol, whereas ΔshfP only produce 0.31 ± 0.10 mM glycerol (Fig. 4E). Consistent with the notion that the phosphoesterase activity of ShfP is required for glycerol generation, cells that produced variants of ShfP with substitutions that disrupted the phosphoesterase domain produced similar amounts of glycerol as the ΔshfP strain (Fig. 4E). Last, the addition of 1.2 mM, but not 0.2 mM, glycerol to the sporulation medium of ΔshfA ΔshfP cells resulted in an ~100-fold sporulation defect (Fig. 4F). Complementation of the ΔshfA ΔshfP strain with shfA at a single chromosomal locus corrected the glycerol-mediated sporulation defect, indicating that shfA provides protection against the extracellular glycerol (Fig. 4G). We conclude that cells entering sporulation early generate glycerol in an ShfP-dependent manner, which inhibits and delays entry into sporulation for cells that have not already initiated the program.

Open in a separate window

Fig. 4.

ShfP-dependent release of glycerol by early-sporulating cells inhibits entry into sporulation.

(A) Above, size exclusion chromatogram of cell-free culture supernatants harvested from WT (gray) or shfP (pink) cells. Below, relative sporulation efficiencies of shfA shfP cells (strain TD507) subjected to sporulate in cell-free culture supernatants harvested from both strains from the above FPLC fractions. Strains: PY79 and TD517. (B) ¹H and (C) ¹³C NMR spectra of active sporulation-inhibitor fraction of purified culture supernatants from WT cells. (D) Extracted ion chromatogram of (black) purified active sporulation-inhibitor fraction and (red) glycerol standard. (E) Concentration of glycerol released into the culture medium of indicated strains. Bars represent mean; data points represent a measure from an independent experiment. Strains: PY79, TD517, CC134, CC254, CC255, and CC258. (F) The addition of glycerol to sporulation medium causes a sporulation defect. Average sporulation efficiency of the ΔshfA ΔshfP (TD507) mutant strain cultured in synchronous sporulation medium with the indicated amount of glycerol added. Sporulation efficiencies are reported as normalized values to that of the recipient strain cultured in the absence of glycerol. Mean and SD of three independent cultures for each condition are shown. (G) Sporulation efficiencies of the ΔshfP (TD517), ΔshfA ΔshfP (TD507), and ΔshfA ΔshfP thrC::shfA (CC221) strains cultured in synchronous sporulation medium in the presence or absence of 1.3 mM glycerol. Sporulation efficiencies are reported as normalized values to that of the ΔshfP (TD517)–recipient strain cultured in the absence of glycerol. Bars represent mean; errors, SD.

The ShfAP pathway promotes heterogeneous entry into sporulation

Although deletion of shfA and shfP did not result in a sporulation efficiency defect, we hypothesized that inhibiting the entry of a subpopulation of cells into sporulation would affect the heterogeneous entry into the sporulation program. To test this, we observed the fraction of cells initiating sporulation in an asynchronous population using fluorescence microscopy to monitor polar septum formation and forespore engulfment. Over time, WT cells displayed a steady increase in the fraction of cells entering sporulation, but at each time point, we observed a relatively larger fraction of ΔshfAP cells that had initiated sporulation (Fig. 5A). Concomitantly, ΔshfAP cells also completed sporulation faster relative to WT cells, as evidenced by their elaboration of bright forespores (as viewed by differential interference contrast optics, indicating dehydration of the spore core) and mature, released spores (fig. S4A). To test whether the larger proportion of WT cells that had not initiated sporulation confers a growth advantage when faced with an influx of nutrients, we initiated asynchronous sporulation and, at different time points, removed an aliquot of the cells, resuspended them in fresh medium to permit regrowth, and measured initial doubling time. At each time point that we examined, we observed that ΔshfA, ΔshfP, and ΔshfAP cultures displayed a slower doubling time relative to WT, indicating that the absence of the ShfAP pathway resulted in a population that was less poised to take advantage of fresh nutrients after the initiation of sporulation, presumably because these mutants harbored a smaller subpopulation of cells that had not entered sporulation (Fig. 5B and fig. S4B). To directly observe population heterogeneity, we monitored Spo0A activity using flow cytometry at various times after sporulation induction in asynchronous cultures of either WT or ΔshfAP cells harboring the P_spoIIE-gfp reporter. Figure 5C shows the Spo0A activation profiles of three independent cultures of WT (gray) and ΔshfAP (red) cells at various time points. Qualitatively, we noticed that the WT cultures displayed multiple populations of cells activating Spo0A, whereas the distribution of ΔshfAP cells appeared more Gaussian. This difference in population heterogeneity was reflected in a higher bimodal coefficient (33) for WT cells (0.45 ± 0.02) compared to ΔshfAP cells (0.37 ± 0.03) at t = 18.5 hours. As a further test, we plotted the flow cytometry data for the three independent cultures of WT and ΔshfAP cells at t = 18.5 hours as a box plot (Fig. 5D) and examined the distribution of statistical outliers. Cultures of ΔshfAP cells showed a roughly equal distribution of statistical outliers at high and low levels of Spo0A activation, whereas outliers in WT cells were heavily skewed toward low activation of Spo0A, suggesting that WT cells maintained a larger population of cells that had likely not committed to enter the sporulation pathway.

Open in a separate window

Fig. 5.

Glycerol secretion enhances phenotypic heterogeneity in sporulating populations.

(A) Fraction of (black) WT or (red) shfA shfP cells committed to sporulation (evidenced by polar septum formation or engulfing membrane) at indicated time points. Bars, mean; data points, measures from independent experiments. Strains: MF277 and CC219. (B) Calculated initial doubling times of indicated strains, relative to WT, induced to sporulate by gradual nutrient deprivation and then resuspended in fresh growth medium at indicated time points. Strains: MF277, CC218, CC219, and CC220. (C) Histograms of fluorescence intensities in WT and shfA shfP cells harboring the P_spoIIE-gfp reporter induced to sporulate by gradual nutrient deprivation and analyzed at the indicated time points using flow cytometry. Strains: MF277 and CC219. (D) 18.5-hour time point in (C) represented as box-and-whisker plots. Black, WT; red, shfA shfP. Box, interquartile range; bars, median; whiskers, range from minimum to lower quartile, and upper quartile to maximum. Statistical outliers are shown as triangle data points. (E) Growth curves, measured by OD₆₀₀ of WT cells grown in minimal medium supplemented with (black) 0 mM, (blue) 0.25 mM, or (gray) 1.25 mM glycerol as the sole carbon source, or (red) glpFK grown in 1.25 mM glycerol. Strains: PY79 and TU148. (F) Fluorescence intensity of indicated strains harboring the P_spoIIE-gfp reporter induced to sporulate by gradual nutrient deprivation. Strains: PY79, MF277, CC220, TU69, TU78, TU80, and TU84. (G) Activation of the KinD-influenced P_epsA promoter in the indicated strains, or in shfP cells in the presence of 1.25 or 0.25 mM glycerol in the medium. Strains: TU87, TU88, TU89, and TU92. (H) Model of early-sporulating cells inhibiting Spo0A activation in late-sporulating cells. ShfP (blue) generates extracellular glycerol that inhibits Spo0A phosphorylation via KinD activation and glycerol metabolism to sustain vegetative growth. Glycerol also inhibits cortex assembly, which is ameliorated by ShfA (red).

Growth conditions for sporulation

B. subtilis PY79 or isogenic mutant derivatives were grown on LB plates (10 g of tryptone, 10 g of yeast extract, and 10 g of sodium chloride per liter; KD Medical) with appropriate antibiotic for single colony isolation. Culture conditions for promoting sporulation were done as previously described (50). Briefly, for asynchronous sporulation assay, isolated colonies were used to inoculate 2 ml of Difco sporulation medium (DSM, KD Medical) and grown at 37°C, rotating at 250 rpm, for spore formation. Assays for measuring sporulation efficiencies (see below) were conducted after 24 hours of incubation, while assays for measuring Spo0A activation, electron microscopy, etc. were done at the indicated times. For synchronous sporulation assay (32), isolated colonies were used to inoculate 2 ml of casein hydrolysate medium (CH; KD Medical) and grown overnight at 22°C rotating at 250 rpm. Overnight cultures were back diluted 1:20 into 20 ml of CH media [to optical density at 600 nm (OD₆₀₀) ~0.1] and grown 2 hours at 37°C rotating at 250 rpm. Cell cultures were harvested by centrifugation (typically ~0.9 ml) and pellets resuspended in 1 ml of resuspension medium (A + B KD Medical) supplemented with threonine (80 μg ml⁻¹; Sigma-Aldrich) for strains lacking a functional thrC locus, with 1 mM isopropyl-β-d-thiogalactopyranoside if strain contains the inducible P_hyperspank promoter, or the indicated concentration of glycerol. The cells were subsequently grown at 37°C rotating at 250 rpm and harvested after 24 hours for sporulation efficiency assay or at the indicated time for other assays.

Sporulation efficiency assay

Sporulation efficiencies were assessed as previously described (51). Briefly, WT and mutant B. subtilis cells were grown synchronously (in resuspension medium) or asynchronously (in DSM) for at least 24 hours at 37°C as described above. Cultures (typically 1- or 2-ml volume) were then exposed to 80°C for 30 min to kill nonsporulating cells. Surviving cells were enumerated by 1/10 serial dilution and plating on LB agar. Viable spores were counted as colony-forming units (CFUs); sporulation efficiencies were reported as a ratio to CFUs recovered from a parallel experiment using WT B. subtilis. Spontaneous suppressor mutants were isolated by enriching for colonies that grew after multiple rounds of heat treatment and growth; mutations were identified by whole-genome sequencing as described previously (21).

β-Galactosidase assay

Beta galactosidase assays were conducted under asynchronous sporulating conditions as described previously (35). Briefly, 2 ml of cultures in DSM were inoculated and incubated at 37°C for 5.5 hours, after which the OD₆₀₀ was measured for each culture. One milliliter of each culture was centrifuged, and the cell pellet was resuspended in 1 ml of Z buffer. Twenty microliters of lysozyme (10 mg ml⁻¹) was added, and cultures were incubated at 37°C for 15 min. Two hundred microliters of ONPG [4 mg ml⁻¹; 2-nitrophenyl β-d-galactopyranoside (Sigma-Aldrich)] was added to each culture, and the color change was monitored, upon which the reaction within each culture were stopped via the addition of 500 μl of 1 M Na₂CO₃, taking note of the time in minutes between adding the ONPG and stopping the reaction. The OD₄₂₀ of each reaction was measured, and the resulting Miller units were calculated and plotted using the following equation: 1000 × OD₄₂₀/[time (min) × volume (ml) × OD_600].

Epifluorescence microscopy

Fluorescence microscopic images of WT and mutant B. subtilis were taken as previously described (52). Briefly, B. subtilis cultures were grown under synchronous or asynchronous sporulation conditions, and the cells were harvested and resuspended in phosphate-buffered saline (PBS, KD Medical) at the indicated times. The cells were then stained with FM4-64 (1 μg/ml; Invitrogen) to visualize membranes and then placed on lysine-coated glass bottom dish (MatTek Corp.) under a 1% agarose pad. The cells were viewed with a DeltaVision Core microscope system (Applied Precision) equipped with an environmental control chamber. Images were captured with a Photometrics CoolSnap HQ2 camera. Seventeen planes were acquired every 0.2 μm at 22°C, and the data were deconvolved using SoftWorx software (GE Healthcare). Control experiments with sporulating strains that did not harbor a gfp fusion indicated that the level of GFP fluorescence was well above the limited background fluorescence of the cells.

Computational analysis of ShfA and ShfP

Homologs for shfA and shfP were identified using sequence profile searches performed with the PSI-BLAST program (53). The searches were run against the nonredundant (NR) protein database of National Center for Biotechnology Information or the sample database compressed by clustering at 50% (nr50) or a custom database of 4210 compete prokaryotic proteomes. Profile-profile searches were conducted using HHpred program (54, 55) with multiple alignments to derive the query hidden Markov model augmented by hits against the nr50 or nr70 databases. They were run against (i) Protein Data Bank (PDB), (ii) Pfam, and (iii) an in-house collection of profiles; the significance of the hits was assessed using the probability percentage and P value of the HHpred hits. Structure inference was performed using Alphafold2 (56). Domain architectures were obtained using a combination of profile searches against Pfam and in-house profiles. Detection of homologs for phyletic pattern correlation analysis was done on a curated set of 4210 complete prokaryotic proteomes. Multiple sequence alignment was built using FAMSA (57) and MAFFT (58) programs and manually corrected on the basis of secondary structure inferred using Alphafold2 (56). Phylogenetic analysis was done using the Le-Gascuel 2008 model in FastTree (59). Phylogenetic trees were rendered with the TreeViewer (60). Structural visualization of the pdb files were carried out using the Mol*viewer (61). The membrane topology was established using deep-learning with the DeepTMHMM program (62) and visualized using MembraneFold (63).

DNA-PAINT

No. 1.5 coverslips were first coated with poly-l-lysine (PLL; Sigma-Aldrich) at room temperature for 10 min. The coverslips were then incubated with FluoSpheres beads (0.1 μm, 1:10⁶ dilution; Life Technologies) for 10 min as fiducial markers. B. subtilis cells were fixed with 4% paraformaldehyde (PFA; Thermo Fisher Scientific) and 0.02% glutaraldehyde directly in the resuspension buffer for 15 min at room temperature, followed by 30 min incubation on ice. The cells were then washed three times with PBS and then resuspend in GTE buffer (50 mM glucose, 20 mM tris-HCl, and 10 mM EDTA). The cells were then incubated with lysozyme (2 mg ml⁻¹) at 37°C for 3 min. The cells were then deposited onto PLL-coated coverslips and incubated for 3 min. After incubation, all the liquid from the sample was removed and the coverslip completely dried at room temperature. The coverslip was then treated with cold (−20°C) methanol for 5 min and followed by −20°C acetone for 30 s. The sample was blocked with blocking buffer (2% bovine serum albumin and 0.1% Triton X-100 in 1× PBS) for 30 min at room temperature. Following blocking, the sample was then incubated with single-domain anti-GFP (1:300 dilution) antibody and anti–red fluorescent protein antibody (1:300) conjugated with docking site overnight (Massive Photonics, Germany). The sample was then washed three times with washing buffer (Massive Photonics, Germany). Imager strands (100 pM) corresponding to different docking sites were added to the sample sequentially for two-color DNA-PAINT super-resolution imaging. Imaging was performed on a custom-built microscope with a Nikon Ti base and a Nikon N-STORM module equipped with a 100× 1.49 NA oil-immersion lens. Images were acquired with a scientific complementary metal-oxide semiconductor camera (Prime 95B, Teledyne Photometrics) with inclined illumination. A 637-nm laser (Coherent, USA) was used as the excitation wavelength. Typically, 7000 frames were collected with exposure time of 250 ms for DNA-PAINT experiments. Images were analyzed and rendered in Picasso (0.6.0) (64). Fluorescence beads were used for drift correction as well as sequential image registration. For image quantification, the spatial relationship between SpoVM-mCherry (used to mark membranes), ShfP-sfGFP, and SpoIVA-GFP in DNA-PAINTING super-resolution image was determined using the same line with line width at five pixels drawn in different channels (red and cyan) to obtain the fluorescence intensity as a function of position. The center position for each protein was quantified through 2D Gaussian fitting. The SD, controlling the spread or width of the curve, obtained from 2D Gaussian fitting, was used to define the width of protein distributions. The mean positions ShfP, SpoIVA, and SpoVM were then normalized through subtracting the corresponding mean position of SpoVM. The graph was then plotted as means ± SD (10 different line profiles from 5 different cells) to show the spatial relationship between SpoVM and ShfP and SpoVM and SpoIVA. The x axis in the graph shows the relative position (nanometer) of ShfP or SpoIVA to SpoVM.

Immunoblotting

Steady state levels of ShfP variants were assessed via immunoblotting as previously described (65). Briefly, B. subtilis and isogenic mutant cells were induced to sporulate in resuspension medium as described above, and after 6 hours of incubation at 37°C, the cells were harvested and resuspended in 500 μl of protoplast buffer [0.5 M sucrose, 10 mM K₂PO₄, 20 mM MgCl₂^, and lysozyme (0.1 mg ml⁻¹) (Sigma-Aldrich)] and incubated at 37°C for 30 min with shaking at 300 rpm. Protoplasts were harvested by centrifugation and lysed by resuspension in 200 μl of PBS and repeatedly passaged through a 20-gauge hypodermic syringe needle (Covidien). Fifteen microliters of the sample was combined with 5 μl of 4× lithium dodecyl sulfate (LDS) sample buffer (NuPAGE), separated by SDS–polyacrylamide gel electrophoresis, and transferred to polyvinylidene difluoride membrane (Novex) using iBlot (Invitrogen). The blots were blocked in 5% skim milk (Carnation) in tris-buffered saline (TBS)/Tween (TBS + 1% Tween 20; Sigma-Aldrich) overnight at 4°C with gentle shaking. The blots were incubated with rabbit antisera raised against purified ShfP or σ^A and detected using anti-rabbit immunoglobulin G StarBright (Bio-Rad) with a ChemiDoc MP imager (Bio-Rad).

Transmission electron microscopy

Sporulating cells were visualized by transmission electron microscopy (TEM) as previously described (21) with slight modifications. Briefly, the cells were allowed to sporulate synchronously in resuspension medium for 5 hours. The cells were harvested by centrifugation, washed with PBS, and resuspended in equal volume of 4% formaldehyde, 2% glutaraldehyde, and in 0.1 M cacodylate buffer. Resuspended cells were then added to the top of a step gradient of five different metrizoic concentrations (70, 60, 50, 40, and 30%) and centrifuged at 40,000g for 60 min at 4°C. Spore-forming cells were found in the middle layers and collected, washed with water, resuspended in fixative, postfixed using 1% osmium tetroxide solution, and then dehydrated sequentially in 35, 50, 75, 95, and 100% ethanol followed by 100% propylene oxide. The cells were infiltrated in an equal volume of 100% propylene oxide and epoxy resin overnight and embedded in pure resin the following day. The epoxy resin was cured at 55°C for 48 hours. The cured block was thin-sectioned and stained in uranyl acetate and lead citrate. The samples were imaged with a Hitachi H7600 TEM equipped with a charge-coupled device camera.

Coculturing

Both reporter and nonreporter strains were grown in 2 ml of DSM at 37°C with shaking at 250 rpm for 4 hour until OD₆₀₀ ~0.8. Cultures were combined by back-diluting into 6 ml of fresh DSM with equal cell numbers for a total OD₆₀₀ of 0.01 (OD₆₀₀ = 0.05 for each strain). The combined cultures were grown overnight at 22°C shaking at 250 rpm. The cultures were then shifted to 37°C and grown for 2 hours past stationary phase (OD₆₀₀ ~2.5). The cells were then harvested for flow cytometry analysis to measure Spo0A activation (see below).

Supernatant harvesting

To collect conditioned cell-free supernatant, cultures were grown overnight in CH medium at 22°C shaking at 250 rpm. Cultures were then back diluted 1:20 (to OD₆₀₀ ~0.1) in CH medium and grown at 37°C for 2 hours shaking at 250 rpm. The cultures were centrifuged and resuspended in an equal volume of resuspension medium (with added threonine if thrC locus disrupted) and incubated at 37°C for 6 hours (or otherwise indicated times). The cultures were then centrifuged for 7 min at 7000g and supernatant collected and filtered through a 3000 Dalton MWCO Amicon Ultra-4 filter (Millipore). Aliquots were stored at −80°C.

For supernatant swapping assays, a culture of the recipient reporter strain was grown overnight in 20 ml of CH medium at 22°C shaking at 250 rpm. The culture was then back diluted 1:20 (final OD₆₀₀ ~0.1) in CH medium and grown for 2 hours at 37°C with 250-rpm shaking until it reached midexponential phase (OD₆₀₀ = 0.4 to 0.6). The culture was then split into equal volumes and centrifuged for 5 min at 7000g. The supernatant was discarded, and the cell pellets were resuspended in conditioned cell-free supernatant. Sporulation was allowed to proceed for 2.25 hours after induction, and the cells were then harvested for flow cytometry analysis to measure Spo0A activation (see below). To determine sporulation efficiencies, the recipient strains were grown overnight in 2 ml of CH medium at 22°C shaking at 250 rpm. The cultures were then back diluted 1:20 (OD₆₀₀ ~0.1) in CH medium and continued to grow at 37°C for 2 hours. Then, 0.8 ml of culture was centrifuged, and the supernatant was removed. Next, 2 ml of frozen supernatant were lyophilized from indicated strains and resuspended in 1 ml of fresh resuspension medium (supplemented with 80 μg/ml of threonine) and used to resuspend harvested recipient strains. The cultures were then incubated at 37°C for 24 hours, shaking at 250 rpm, and sporulation efficiencies were measured as described above.

Flow cytometry

To analyze Spo0A activation of the reporter spo0A-gfp fusion strains, 160 μl of sporulating cultures were centrifuged and the supernatant was removed. Pellets were resuspended in 1 ml of PBS at pH 7.4 (KD Medical) to a final concentration of ~10⁶ to 10⁸ CFU/ml. The samples were then analyzed using a Sony SA3800 flow cytometer. A minimum of 400,000 cells per sample was analyzed using a 488-nm laser and 530/30 band-pass filter to detect GFP signal. The following voltage values were used: 200 V for forward scatter, 250 V for side scatter, and 400 V for GFP channel. The median GFP signal was measured for each strain, and gates were drawn to encompass the GFP signal using isogenic non-gfp–fused cells as controls to ascertain the amount of background fluorescence the cells emitted. Data were analyzed using FlowJo flow cytometry software. The percentage of the total population of cells with GFP signal above the background was determined as a proxy for the Spo0A activation. Three replicates of each fusion strain were analyzed using the above method for each condition, and the reported values of these three independent replicates, along with the mean value, are shown. To measure Spo0A repression activity in culture supernatants, three independent overnight cultures of the spo0A-gfp and the ΔshfA ΔshfP spo0A-gfp reporter strains in LB medium were back diluted in 200 μl of DSM medium to 0.1 OD₆₀₀ in a black-wall and clear-bottom 96-well microplate (PerkinElmer). The plate was then incubated at 37°C with continuous shaking (600 cpm) using a BioTek Synergy H1 microplate reader to record GFP fluorescence, using the 479 excitation and 520 emission settings, and OD₆₀₀ measurements were taken every 0.5 hour. At the indicated time points, the cultures were paused and 50 μl of each culture was diluted in an equal volume of 4% (w/v) PFA in PBS (Thermo Fisher Scientific) solution and stored at 4°C overnight. Flow cytometric analysis was performed on 100 μl of sample from each well in the 96-well plate using the same flow cytometric settings as above.

Fast protein liquid chromatography

Cell-free supernatants stored at −80°C were thawed on ice and filtered through a 0.22-μm membrane to remove precipitants. Then, 1 ml of supernatant was added to a Superdex 30 (GE Healthcare) size exclusion column equilibrated with buffer (25 mM tris at pH 8.0 and 150 mM NaCl) and separated at 0.25 ml min⁻¹ using an AKTA Pure instrument (GE Healthcare). One-milliliter fractions were collected, lyophilized, and resuspended in 1 ml of resuspension buffer to be used on recipient strains to determine sporulation efficiency.

Nutrient replenishment and outgrowth assay

Three independent overnight cultures of each strain grown in LB medium were back diluted in 200 μl of DSM medium to 0.1 OD₆₀₀ in a 96-well flat-bottom microplate (PerkinElmer). The plate was then incubated at 37°C with continuous shaking (600 cpm) using BioTek Synergy H1 microplate reader to record OD₆₀₀ measurements every 0.5 hours. At the indicated time points, cultures were paused and either 2 μl of each culture was used to seed 200 μl of fresh prewarmed liquid LB medium, and cultures were allowed to continue growth for the indicated time; or 100 μl of culture was centrifuged, washed with 100 μl PBS, stained with FM4-64, and imaged using epifluorescence microscopy to determine sporulation stage. The initial doubling time for each subculture was calculated as described previously (66) using the equation: [Ln (2)/(Ln(OD600_T2) − Ln(OD600_T1/T2-T1)].

Alternatively, three independent overnight cultures of the reporter strain were grown overnight in LB medium. The overnight cultures were then back diluted in 200 μl of DSM medium (supplemented with the indicated amount of glycerol) to OD₆₀₀ = 0.1 in a 96-well (black-wall, clear-bottom) microplate (PerkinElmer). The plate was then incubated at 37°C with continuous shaking (600 cpm) using BioTek Synergy H1 microplate reader to record GFP fluorescence, using the 479 excitation and 520 emission settings, and OD₆₀₀ measurements every 0.5 hours.

Glycerol detection

Cell-free supernatants from indicated strains were harvested after 5.5 hours of growth at 37°C in resuspension medium as described above. Glycerol content in each supernatant fraction was quantitated using the Abcam Free Glycerol Assay Kit (Fluorometric, High Sensitivity; Ab174092) following the manufacturer’s instructions.

We thank the members of the K.S.R. laboratory for suggestions and comments on the manuscript; S. Gottesman, G. Storz, A. Khare, V. T. Lee, and S. Wickner for discussions; C. Weiss, E. Kim, M. Tisza, J. Barriga, and C. Ellermeier for strains; R.-C. Hsia, F. Soheilian, and C. Burks (CCR Electron Microscopy Laboratory) for TEM sample preparation and imaging; the CCR Genomics Core Facility for whole-genome sequencing; and R. D. O’Connor for NMR support.