The Membrane Domain of SpoIIIE Is Required for Membrane Fusion during Bacillus subtilis Sporulation

Abstract

During Bacillus subtilis sporulation, SpoIIIE is required for both postseptational chromosome segregation and membrane fusion after engulfment. Here we demonstrate that SpoIIIE must be present in the mother cell to promote membrane fusion and that the N-terminal membrane-spanning segments constitute a minimal membrane fusion domain, as well as direct septal localization.

Bacillus subtilis SpoIIIE and Escherichia coli FtsK are well-characterized members of a conserved family of bacterial proteins involved in chromosome segregation. These proteins have similar architectures, with an N-terminal membrane domain, a variable linker, and a C-terminal cytoplasmic domain containing a Walker-type ATP binding site (6, 7). The membrane domains of FtsK and SpoIIIE localize the proteins to the site of cell division (2, 13, 16), whereas the cytoplasmic domains move along DNA in an ATP-dependent manner (1, 2). FtsK is an essential protein required for cell division as well as chromosome decatenation and segregation; under such circumstances, it likely aligns the recombination sites (4) and activates the Xer recombinase (1). In contrast, SpoIIIE is dispensable for growth but plays two crucial roles in sporulation. First, it completes chromosome partitioning during sporulation, which differs from that of vegetative growth since the asymmetrically positioned sporulation septum initially bisects one chromosome, trapping 30% in the forespore (Fig. 1ii) (14). SpoIIIE acts in the mother cell to rescue this trapped chromosome, exporting the DNA into the forespore (12) in a process which requires a functional ATP binding site (10). Second, SpoIIIE is necessary for the membrane fusion event which releases the forespore into the mother cell cytoplasm at the completion of the phagocytosis-like process of engulfment (Fig. 1iii to v) (10). These two roles of SpoIIIE are distinct and genetically separable, as we have isolated mutations that block DNA translocation but produce only a minor membrane fusion defect (10). It is tempting to speculate that the SpoIIIE/FtsK family of proteins also participates in the membrane fusion event at the completion of cell division, perhaps serving to coordinate chromosome segregation with cell division and thereby preventing chromosome damage that would arise if daughter cell separation preceded the completion of chromosome segregation, a proposal consistent with the late division defect of certain FtsK mutants (3, 5).

FIG. 1.

Stages of forespore engulfment and chromosome translocation. The black circles indicate the localization of SpoIIIE.

The membrane fusion event at the completion of engulfment occurs in the membrane of the migrating mother cell, suggesting that proteins directly involved in membrane fusion should act in the mother cell and localize to the cell pole before membrane fusion. While SpoIIIE localizes to the pole before fusion (10), it is produced prior to polar septation and is predicted to be present in both cells of the sporangium. To test if SpoIIIE is specifically required in the mother cell for membrane fusion, we fused spoIIIE to similarly expressed forespore- and mother cell-specific promoters (P_spoIIQ and P_spoIID, respectively [12]) and tested the ability of these hybrid genes to complement the membrane fusion defects of various spoIIIE mutants (Table 1) using an assay which relies on the membrane-impermeable stain FM 4-64 and the membrane-permeable stain Mitotracker Green (MTG) (10). When these stains were applied to sporangia that had completed membrane fusion, the forespore membranes failed to stain with FM 4-64 and stained only with MTG (Fig. 2A) while the forespores of unfused sporangia were accessible to both stains (Fig. 2A).

TABLE 1.

B. subtilis strains used in this study

Straina	Genotypeb	Reference
PY79	Wild type	15
KP92	spoIIIE36	9
KP141	ΔspoIIIE::spc	9
KP629	ΔspoIIIE::spc amyE::PspoIIIE-spoIIIE-gfpΩcat	12
KP633	spoIIIE36 amyE::PspoIIIE-spoIIIE-gfpΩcat	12
KP634	spoIIIE36, amyE::PspoIIQ-spoIIIE-gfpΩcat	12
KP635	spoIIIE36 amyE::PspoIID-spoIIIE-gfpΩcat	12
KP676	spoIIIEMSS1-4-gfpΩspc	This study
KP677	spoIIIEMSS1-4Ωspc	This study
KP682	spoIIIEMSS1-4Ωspc amyE::PspoIIQ-spoIIIEMSS1-4-gfpΩcat	This study
KP683	spoIIIEMSS1-4Ωspc amyE::PspoIID-spoIIIEMSS1-4-gfpΩcat	This study

^aAll strains are derivatives of PY79.

^bAll plasmids used for amyE integration were derivatives of pDG1662 (8) and had a variety of modifications as described in references 11 and 12.

FIG. 2.

Membrane fusion assay at t₃, showing the membranes stained with FM 4-64 (red) and MTG (green). Arrowheads indicate sporangia that have completed membrane fusion. (A) Wild type (the arrows indicate an unfused sporangium); (B) ΔspoIIIE sporangia; (C) spoIIIE36 sporangia; (D) spoIIIE36 amyE::spoIIIE-gfp sporangia; (E) spoIIIE36 amyE::P_spoIIQ-spoIIIE-gfp sporangia; (F) spoIIIE36 amyE::P_spoIID-spoIIIE-gfp sporangia; (G) spoIIIEMSS1-4 sporangia; (H) spoIIIEMSS1-4 amyE::P_spoIIQ-spoIIIEMSS1-4 sporangia; (I) spoIIIEMSS1-4 amyE::P_spoIID-spoIIIEMSS1-4 sporangia. Scale bar, 2 μm.

Four hours after the initiation of sporulation (t₄), approximately 85% of wild-type sporangia had completed membrane fusion while only 1% of spoIIIE-null sporangia had fused (Table 2; Fig. 2A and B show t3). We tested the effects of cell-specific SpoIIIE expression in both a ΔspoIIIE strain and a spoIIIE36 strain, the latter of which had a less severe membrane fusion defect (24% of sporangia fused by t₄) (Fig. 2C; Table 2) but normal compartmentalization of cell-specific gene expression, unlike the null mutant (9, 14). The maintenance of cell-specific gene expression in the spoIIIE36 background highlighted the differences between mother cell- and forespore-expressed spoIIIE genes. Mother cell-expressed spoIIIE complemented the membrane fusion defect of spoIIIE36 as well as the expression of spoIIIE from its native promoter (79% of sporangia with P_spoIIIE-spoIIIE fused versus 78% with P_spoIID-spoIIIE fused) (Fig. 2D and F; Table 2). In contrast, expression of spoIIIE in the forespore had no effect on membrane fusion (23% of sporangia with P_spoIIQ-spoIIIE fused) (Fig. 2E; Table 2). Similar effects were seen in the spoIIIE-null strain (Table 2), although forespore-produced SpoIIIE supported some membrane fusion (albeit less than mother cell-produced SpoIIIE), likely because of the compartmentalization defect. Thus, SpoIIIE is required in the mother cell to promote fusion of the engulfing mother cell membrane, suggesting that it might be directly involved in membrane fusion.

TABLE 2.

Complementation of the membrane fusion defects of various spoIIIE mutants with forespore- or mother cell-produced SpoIIIE

Gene at amyE	% of sporangia with the indicated spoIIIE background showing membrane fusiona,b
	ΔspoIIIE	spoIIIE36	spoIIIEMSS1-4
None	1	24	13
PspoIIIE-spoIIIE	84	79	ND
PspoIID-spoIIIE	69	78	ND
PspoIIQ-spoIIIE	32	23	ND
PspoIID-spoIIIEMSS1-4	12	ND	24
PspoIIQ-spoIIIEMSS1-4	4	ND	9

^aPercentages of sporangia that had completed membrane fusion at t₄. At least 200 sporangia were scored for each strain. ND, not done.

^bNumbers of heat-resistant spores per milliliter were as follows: ΔspoIIIE P_spoIIIE-spoIIIE, 3 × 10⁸; spoIIIE36 P_spoIIIE-spoIIIE, 3 × 10⁸; ΔspoIIIE P_spoIID-spoIIIE, 2 × 10⁸; spoIIIE36 P_spoIID-spoIIIE, 3 × 10⁸; ΔspoIIIE P_spoIIQ-spoIIIE, 7 × 10⁶; spoIIIE36 P_spoIIQ-spoIIIE, 6 × 10⁶. Heat resistance was assessed 24 h after the onset of sporulation, by heating to 80°C for 20 min. At least two independent cultures were assessed.

Mutagenesis of the ATP binding site in the cytoplasmic domain of SpoIIIE abolishes DNA translocation while only modestly effecting membrane fusion (10), suggesting that this domain may be dispensable for membrane fusion. To test this hypothesis, we deleted the cytoplasmic domain of SpoIIIE, truncating the protein after amino acid 192 and leaving intact the four membrane-spanning segments and a short linker region before green fluorescent protein. This truncated protein (SpoIIIEMSS1-4) was completely inactive for chromosome segregation (data not shown) but was able to support some membrane fusion; by t₄, 13% of spoIIIEMSS1-4 sporangia completed membrane fusion (Fig. 2G, Table 2), compared to 1% of ΔspoIIIE sporangia (Fig. 2B; Table 2). While 13% was significantly lower than the 84% membrane fusion of the wild type, it was more than 10-fold higher than the percent membrane fusion of the spoIIIE null strain. In contrast, strains expressing only the cytoplasmic domain of spoIIIE were defective in both DNA translocation and membrane fusion (12). Overproduction of SpoIIIEMSS1-4 in the mother cells of spoIIIEMSS1-4 sporangia (from P_spoIID-spoIIIEMSS1-4, which should produce about 100-fold more SpoIIIEMSS1-4 than P_spoIIIE [12]) increased the fusion efficiency from 13 to 24% (Fig. 2I; Table 2), whereas overproduction of SpoIIIEMSS1-4 in the forespores (P_spoIIQ-spoIIIEMSS1-4) of spoIIIEMSS1-4 sporangia had no effect (Fig. 2H; Table 2), suggesting that the membrane-spanning segments alone are capable of promoting membrane fusion. The relatively low efficiency of membrane fusion supported by SpoIIIEMSS1-4 is likely caused its inefficient relocalization to the cell pole (2), where membrane fusion occurs (10).

These results demonstrate that SpoIIIE is required in the mother cell to mediate membrane fusion at the completion of engulfment, as well as to translocate the forespore chromosome across the sporulation septum (12). Interestingly, the membrane fusion domain of SpoIIIE comprises the membrane-spanning segments, which are conserved even in distantly related bacteria that do not produce endospores (7). Such bacteria would not require a protein to catalyze membrane fusion at the completion of engulfment, supporting speculation that these proteins might be able to participate in membrane fusion during cell division.