Suppression of Engulfment Defects in Bacillus subtilis by Elevated Expression of the Motility Regulon

Abstract

During Bacillus subtilis sporulation, the transient engulfment defect of spoIIB strains is enhanced by spoVG null mutations and suppressed by spoVS null mutations. These mutations have opposite effects on expression of the motility regulon, as the spoVG mutation reduces and the spoVS mutation increases σ^D-directed gene expression, cell separation, and autolysis. Elevating σ^D activity by eliminating the anti-σ factor FlgM also suppresses spoIIB spoVG, and both flgM and spoVS mutations cause continued expression of the σ^D regulon during sporulation. We propose that peptidoglycan hydrolases induced during motility can substitute for sporulation-specific hydrolases during engulfment. We find that sporulating cells are heterogeneous in their expression of the motility regulon, which could result in phenotypic variation between individual sporulating cells.

A key step in Bacillus subtilis sporulation is engulfment, which mediates a dramatic rearrangement of the two cells required for sporulation, from two adjacent daughter cells to a sporangium in which one daughter cell (the forespore) is completely enclosed within the cytoplasm of another (the mother cell). This transformation is mediated by the migration of the mother cell membrane around the forespore, which is ultimately released into the mother cell cytoplasm (Fig. 1A). Engulfment depends on three conserved proteins, SpoIID, SpoIIM, and SpoIIP (1, 5, 17, 30), and is facilitated by the less-conserved SpoIIB protein (19, 23). These proteins participate in septal thinning (1, 5, 17, 19, 23, 30), during which septal peptidoglycan is degraded by the SpoIID peptidoglycan hydrolase and perhaps other enzymes (1). While spoIID, spoIIM, and spoIIP mutants are completely defective in septal thinning and membrane migration (Fig. 1B), a spoIIB mutant shows slow and uneven septal thinning but ultimately completes engulfment and produces spores (Fig. 1C). Two additional proteins have been proposed to play a role in septal thinning: SpoVG, whose absence exacerbates the engulfment defect of spoIIB (19, 23), and SpoVS, whose absence suppresses a spoIIB spoVG double mutant, allowing engulfment to proceed (27). Here, we provide evidence that the spoVS mutation causes decreased expression of early-sporulation genes and increased expression of motility genes and that motility-specific genes, likely peptidoglycan hydrolases, can suppress septal thinning defects.

FIG. 1.

Engulfment in B. subtilis. (A) Asymmetric division produces the small forespore and large mother cell. Septal thinning (step i) commences at the septal midpoint and proceeds towards the edges, followed by membrane migration up (step ii) and around (step iii) the forespore, until the membrane meets and fuses (step iv) to release the forespore into the mother cell cytoplasm. (B) In the absence of SpoIID, SpoIIM, or SpoIIP, septal thinning is blocked, and the forespore ultimately bulges into the mother cell. (C) In the absence of SpoIIB, septal thinning is slowly and uneven, forming a transient bulge, although engulfment is ultimately completed. In the absence of both SpoIIB and SpoVG, membrane migration is blocked.

To investigate the mechanism of spoVS-mediated suppression of spoIIB spoVG, we used an in vivo membrane fusion assay that employs two membrane stains, the membrane-impermeable FM 4-64 and the membrane-permeable stain MitoTracker Green (strains of B. subtilis used in this study are shown in Table 1) (29). During engulfment in the wild type, the engulfing membrane smoothly curves around the forespore (Fig. 2A, arrow and arrowhead), with the septal membrane staining twice as brightly as other regions, due to the presence of two membranes (24). Ultimately, the migrating membrane meets and fuses to release the forespore into the mother cell cytoplasm. After membrane fusion, FM 4-64 is excluded from the forespore, resulting in a green forespore enclosed within the red mother cell (Fig. 2A, double arrowhead). In spoIIB spoVG, the growing forespore pushes through the unthinned septum, forming a bulge that is typical of septal thinning defective mutants (Fig. 2C to D). In the spoIIB spoVG spoVS triple mutant, sixfold-fewer bulges were observed than in the spoIIB spoVG mutant (6% versus 35% at t_3.0) (Fig. 2G and H; Table 2), and engulfment was complete in 21% of spoVS spoIIB spoVG sporangia versus 0% of spoIIB spoVG sporangia at t₄ (Fig. 2; Table 2). The spoVS mutation also weakly suppressed spoIID and spoIIM strains, reducing the frequency of bulges twofold (from 64% to 31% in spoIIM and from 45% to 21% in spoIID) but not supporting the completion of engulfment or increased spore production (see Fig. S1 and Table S1 in the supplemental material). These results suggest that the spoVS mutation is a general suppressor of septal thinning defects.

TABLE 1.

Bacillus subtilis strains used in this study^a

Strain	Genotype	Reference or source
PY79	Wild type	35
KP52	ΔspoIIB::erm spoVG::Tn917	19
KP84	spoIIAC-lacZ	28
KP174	Δ(spoIIAA-AC)::spec	2
KP343	ΔspoIIB::erm	19
KP548	spoVG::Tn917::spec	23
KP535	ΔspoVS::spec	27
KP787	ΔspoIIB::erm spoVG::Tn917 ΔspoVS::spec	This study
KP798	ΔsinR::neo	6
KP799	ΔsinR::neo ΔspoVS::spec	This study
KP812	flgMΔ80	21
KP813	sigD::pLM5	10
KP814	ΔspoIIB::erm spoVG::Tn917::spec	This study
KP815	ΔspoIIB::erm spoVG::Tn917::spec flgMΔ80	This study
KP818	Phag(−UP)-lacZ	6
KP819	flgMΔ80 Phag(−UP)-lacZ	This study
KP820	sigD::pLM5 (cmr) Phag(−UP)-lacZ	This study
KP821	ΔspoVS::spec, Phag(−UP)-lacZ	This study
KP822	ΔsinR::neo Phag(−UP)-lacZ	This study
KP823	spoVG::Tn917, Phag(−UP)-lacZ	This study
KP826	ΔspoIIB::erm spoVG::Tn917 ΔspoVS::spec lytABC::neo	This study
KP827	ΔspoIIB::erm spoVG::Tn917 ΔspoVS::spec lytD::tet	This study
KP828	ΔspoIIB::erm spoVG::Tn917 ΔspoVS::spec lytABC::neo lytD::tet	This study
KP829	flgMΔ80 sigD::pLM5	This study
KP830	ΔspoVS::spec sigD::pLM5	This study
KP930	spoIIAC-lacZ::tet rvtA11::spec	This study
KP931	spoIIAC-lacZ::tet rvtA11::spec ΔcotE::cm ΔspoVS::spec	This study
KP932	spoIIAC-lacZ::cm ΔabrB::tet	This study
KP933	spoIIAC-lacZ::cat ΔabrB::tet, ΔspoVS::spec	This study
KP934	spoIIAC-lacZ::cat flgMΔ80	This study
KP936	spoIIAC-lacZ::cat ΔspoVS::spec	This study
KP937	spoIIAC-lacZ::cat ΔsinR::neo	This study
KP938	spoIIAC-lacZ::cat spoVG::Tn917	This study

^aAll strains are PY79 derivatives except those with flgMΔ80, which are JH642 derivatives.

FIG. 2.

Effect of the spoVS mutation on engulfment. Sporulation was induced by resuspension (33), and samples taken 2 h (t₂) (A, C, E, G, and I) and 4 h (t₄) (B, D, F, H, and J) later were stained with FM 4-64 (red) and MitoTracker Green (green) as previously described (24, 29). (A and B) Wild-type (PY79) sporangia after septation (arrow) and during engulfment (arrowhead). The septum stains approximately two times more brightly than the cytoplasmic membrane, because it contains two parallel membranes, while the engulfing membrane stains approximately three times more brightly, because it contains three membrane layers (Fig. 1A) (24). After membrane fusion (double arrowhead), FM 4-64 is excluded from the forespore (29), which is stained only with MitoTracker Green. (C and D) The spoIIB spoVG double mutant (KP52) shows sporangia with flat polar septa or bulges (arrow). (E and F) The spoVS mutant (KP535) is delayed in polar septation but engulfs normally (E, arrow) and completes membrane fusion (F, arrow). (G and H) The spoIIB spoVG spoVS triple mutant (KP787) exhibits few bulges (G, arrow) and completes membrane migration (H, arrow) and fusion (H, arrowhead). (I to J) The spoIIB spoVG flgMΔ80 triple mutant (KP815) shows few bulges and completes engulfment (I, arrow).

TABLE 2.

Effect of spoVS and flgM mutations on polar septation, engulfment, and spore formation

Genotype	Spores/ml	% Sporangia (t2)a	% Sporangia with bulges (t3)b	% Sporangia fused (t4)
Wild type (PY79)	4 × 108	55 (775)c	0 (347)	84 (248)
spoVS	1 × 106	6 (1,342)	0 (547)	46 (623)
spoIIB	2 × 108	45 (1,670)d	37 (699)	7 (345)
spoVG	9 × 108	41 (391)d	0 (301)	40 (251)
spoIIB spoVG	9 × 104	55 (1,211)	35 (1,073)	0 (401)
spoIIB spoVG spoVS	2 × 106	1 (312)	6 (1,270)	21 (1,459)
spoIIB spoVG flgMΔ80	2 × 106	30 (573)	5 (761)	9 (474)
spoIIB spoVG spoVS lytABCe	5 × 105	NDf	13 (615)	35 (764)
spoIIB spoVG flgMΔ80 lytABCe	3 × 105	NDf	14 (336)	34 (126)

^aPercent sporangia is determined by dividing the number of sporangia by the total number of total cells (in parentheses) in the fields scored.

^bPercent sporangia with bulges or fused is calculated by dividing the number of sporangia with the indicated phenotype by the total number of sporangia (in parentheses).

^cThe number in parentheses indicates the total number of cells scored (vegetative cells and sporangia).

^dData for percent sporangia at t₂ are from Perez et al. (23) and are included for comparison only.

^eThe additional deletion of lytD had no effect on suppression (data not shown).

^fND, not determined.

We noted that spoVS strains showed fewer polar septa than the wild type (Fig. 2E and F; Table 2). This was due to decreased expression of early sporulation genes such as spoIIAC (Fig. 3F) and spoIIE (see Fig. S3 in the supplemental material), which require the Spo0A∼P transcription factor that governs entry into sporulation (7). Expression of Spo0A∼P-dependent genes is regulated by kinases and phosphatases that modulate the level of phosphorylated Spo0A (22, 32) and by various transcription factors that inhibit Spo0A∼P-dependent gene expression (7-9). We found that the defect in polar septation and Spo0A∼P-dependent gene expression could be substantially rescued by a sinR mutation, which eliminates one of these transcription factors, but not by spo0Asad, which encodes a Spo0A protein active without phosphorylation (11), rvtA11, which encodes a Spo0A protein that can be phosphorylated by an alternative kinase (16), or abrB or spo0J-soj mutations, which eliminate repressors of Spo0A∼P-dependent gene expression (25, 34; see Fig. S3 in the supplemental material). These results suggest that the spoVS mutant has increased SinR activity.

FIG. 3.

Cell separation, autolysis, and Spo0A∼P- and σ^D-directed gene expression. (A and B) FM 4-64 membrane staining showing the wild type (PY79) growing in chains (A) and spoVS (KP535) growing as single cells (B). (C) Electron micrograph showing a vegetative PY79 cell dividing in the middle (arrow) and still connected to its sister cell by septal peptidoglycan in the progress of being split by autolytic enzymes (arrowhead); samples were prepared as previously described (23). (D and E) Sodium azide-induced autolysis during vegetative growth, measured by taking samples from a growing culture, adding sodium azide to a concentration of 0.05 M, and following the optical density at 600 nm over time at 37°C with continued aeration (3). The wild type (squares), sigD (open diamonds; KP813), spoVS (circles; KP535), and flgMΔ80 (open triangle; KP812) trains sare shown. (E) The following strains are shown: wild type (squares); sigD (filled diamonds; KP813); spoVS (circles; KP535); flgMΔ80 (triangle; KP812), flgM sigD (open triangle; KP829), and spoVS sigD (open circle; KP830). (F) Spo0A∼P-dependent expression of spoIIAC-lacZ in the wild type (squares; KP84), spoVS (circles; KP936), sinR (open circles; KP937), and spoVS sinR (open squares). Samples were harvested at the indicated time after the initiation of sporulation by resuspension at 37°C. (G) σ^D-directed expression of hag-lacZ during a resuspension sporulation in the wild-type (squares; KP818), spoVG (diamonds; KP823), spoVS (circle; KP821), and flgMΔ80 (triangle; KP819) strains.

SinR is a transcription factor that directly regulates biofilm formation (13), inhibits Spo0A∼P-dependent gene expression, and activates σ^D-dependent motility genes (4, 6, 13, 18, 26), perhaps indirectly (13). Thus, if the spoVS strain has elevated SinR activity, it should also show increased expression of motility genes. Indeed, the spoVS strain showed approximately twofold elevated expression of the flagellin gene (hag-lacZ) relative to the wild type, while the spoVG mutation (enhancer of spoIIB) showed approximately fivefold reduced σ^D activity (Fig. 3G). This suggests that SpoVS directly or indirectly governs SinR activity, similar to two other recently described proteins, YlbF and YmcA (13). Strains lacking these proteins share some phenotypes with spoVS strains, with continued SinR activity, a failure to form biofilms, and growth that is slightly slower than that of the wild type (spoVS data are not shown) (13).

It is unclear if SpoVS directly or indirectly modulates SinR or σ^D activity, and it remains possible that spoVS mutation has additional direct or indirect effects on gene expression. We therefore tested if elevated σ^D activity was sufficient to suppress spoIIB spoVG by inactivating the anti-sigma factor for σ^D, FlgM (21). As expected, flgM strains have increased σ^D-directed gene expression (Fig. 3G). Further, like the spoVS mutation, the flgM mutation increased spore production by spoIIB spoVG ∼20-fold (from 9 × 10⁴ to 2 ×10⁶) (Table 2), reduced the frequency of bulges ∼7-fold (from 35% to 5%), and increased the proportion of sporangia that completed the final step of engulfment, membrane fusion, from 0% to 9% (Fig. 2I and J; Table 2). Thus, increased σ^D activity is sufficient to suppress the engulfment defect of spoIIB spoVG to a similar extent as the spoVS mutation.

B. subtilis cells often grow in short chains of cells that remain connected by septal peptidoglycan (Fig. 3A and C), which ultimately is split by peptidoglycan hydrolases to allow daughter cell separation (31). We noted that spoVS strains rarely grew in chains (Fig. 3B), suggesting an increased activity of peptidoglycan hydrolases involved in cell separation. Indeed, the σ^D regulon includes genes that encode peptidoglycan hydrolases that mediate cell separation (14, 15, 20), which likely facilitates motility by the generation of single cells (3). The overexpression of peptidoglycan hydrolases in spoVS and flgM strains was confirmed by an autolysis assay, which measures the lysis of bacteria whose growth has been arrested (3, 12). Both strains lysed faster than the wild type; after 4 h, optical density at 600 nm was reduced by 60% in the wild type versus 84% in spoVS and 94% in flgM (Fig. 3D). The elevated autolysis of flgM and spoVS strains was eliminated by a sigD mutation, which inactivates σ^D (Fig. 3E), and by a lytABC mutation, which eliminates the major σ^D-directed autolysin (see Fig. S2 in the supplemental material). The lytABC mutation also reduced but did not eliminate spoVS- and flgM-mediated suppression of spoIIB spoVG, increasing bulges and reducing spore formation to ∼25% of suppressed levels (Table 2). We propose that spoVS and flgM suppress septal thinning defects by increasing expression of peptidoglycan hydrolases involved in cell separation and speculate that several such hydrolases are required for full suppression.

The increased σ^D activity in spoVS and flgM strains could be due to either an increased σ^D activity in all sporangia or an increased fraction of sporangia with σ^D activity. To test these possibilities, we used immunofluorescence microscopy to visualize Spo0A and σ^D activity in sporulating cultures. To detect Spo0A activity, we used antibodies to σ^F, the production of which requires Spo0A∼P; to detect σ^D activity, we used antibodies to β-galactosidase produced from the hag-lacZ fusion. After 2.5 h of sporulation, σ^D activity was observed in 16% of wild-type sporangia (cells containing a sporulation septum visualized by FM 4-64 membrane staining) (Fig. 4A; Table 3), compared to 36% of spoVS sporangia and 72% of flgMΔ80 sporangia (Fig. 4B to C; Table 3). Thus, spoVS and flgM strains have two to four times more sporangia expressing motility genes than the wild type, which would likely result in an increased number of sporangia with elevated levels of peptidoglycan hydrolases involved in cell separation. Clearly, even a wild-type population of sporulating cells displays heterogeneity in the activity of the motility regulon.

FIG. 4.

Immunofluorescence microscopy to examine Spo0A∼P- and σ ^D-directed gene expression. Sporulation was induced by resuspension, and samples were harvested at t_2.5 and processed for immunofluorescence microcopy (23), staining with membrane stain FM 4-64 (red, left) and the DNA stain 4′,6′-diamidino-2-phenylindole

(DAPI; blue) and using antibodies specific for σ^F (magenta) and for β-galactosidase produced by hag-lacZ (green). The wild-type strain (KP818) (A), the spoVS strain (KP821) (B), the flgMΔ80 strain (KP819) (C), and the negative control strain KP174 Δ(spoIIAA-AC) which has a deletion in the gene encoding σ^F and no lacZ fusion (D) are shown.

TABLE 3.

Expression of σ^D and Spo0A∼P-directed genes at t_2.5 of sporulation

Genotype	Cell morphologya	Percentage in class	Percentage of cells in morphological class expressing:
			σF	hag-lacZ	Both	Neither
Wild type (244)c	Vegetativeb	41	46	10	4	39
	Sporulating	59	81	0	16	3
spoVS (185)c	Vegetativeb	85	9	46	27	18
	Sporulating	15	57	0	36	7
flgMΔ80 (171)c	Vegetativeb	71	2	69	26	2
	Sporulating	29	26	0	72	2

^aVegetative cells do not contain a polar septum, while sporulating cells are defined as sporangia that contain a polar septum in any stage of engulfment.

^bCells containing σ^F in the vegetative cell category likely have initiated sporulation, but have not yet synthesized the polar septum. Indeed, in many of these cells the chromosome was rearranged into the axial filament structure that precedes polar spetation.

^cTotal cells scored.

These results demonstrate an unanticipated link between expression of the σ^D-directed motility regulon and the ability of sporulating cells to complete the phagocytosis-like process of engulfment during sporulation. Apparently, the continued expression of σ^D-directed genes during sporulation allows engulfment in the septal thinning defective spoIIB spoVG strain, perhaps by providing additional hydrolases that contribute to septal thinning. Because there is cell-to-cell variation in the level of motility gene expression in individual sporulating cells, these observations might explain why there is also cell-to-cell variation in the phenotypes of engulfment mutants, with variations in whether they divided at the second site in the mother cell and whether they have a flat or bulged septum (24). We anticipate the discovery of additional examples in which a cell's proteomic “memory” of past gene expression alters its behavior during subsequent transcriptional responses or developmental pathways.