INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Recent studies have demonstrated that bacterial cells share with eukaryotic cells the ability to actively move macromolecules within their cytoplasm. The rapid separation of the bacterial origin of chromosomal replication and of plasmids just prior to the onset of cell division provided the first evidence that this was the case (14-16, 33, 52-54). The second example was provided by the MinC and MinD proteins of Escherichia coli, which are required for division site selection and which rapidly oscillate from one cell pole to the other (40, 41). An oscillating localization pattern has also been observed for the Bacillus subtilis Soj protein (29, 39), which regulates the onset of sporulation (3, 18). The mechanisms for these dynamic events remain elusive, as bacteria lack a characterized cytoskeleton, although they have distant homologues of tubulin and actin (11, 51). However, these proteins do not appear to be required for either chromosome segregation or MinCD oscillation.

Another example of the movement of macromolecules within a bacterial cell is provided by the spore formation pathway of B. subtilis, in a process known as engulfment. Engulfment mediates a dramatic reorganization of the sporangium, from two adjacent sister cells to a sporangium in which one sister cell lies within the cytoplasm of the other (Fig. 1) (48). Engulfment superficially resembles phagocytosis, as it occurs by the migration of membrane projections from one cell (the mother cell) around the other cell (the forespore) and culminates in the engulfed cell being enclosed in three membranes, the forespore cytoplasmic membrane, a separate membrane derived from the engulfing mother cell membrane, and the outermost mother cell cytoplasmic membrane. However, thus far, there is no evidence for a cytoplasmic contribution to engulfment analogous to the role of actin in assembling membrane projections during phagocytosis and motility in eukaryotic organisms. Indeed, all of the engulfment proteins identified thus far are either integral membrane or exported proteins (48). It therefore seems likely that engulfment will provide a novel mechanism by which phagocytosis-like events can occur.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Initially, a layer of peptidoglycan (gray) lies between the two membranes separating the smaller forespore (top cell) and the larger mother cell (bottom cell; stage IIi). Engulfment starts with the thinning of this peptidoglycan, starting in the middle of the septum (stage IIii); this step is temporally and spatially regulated by SpoIIB (35). Proteins listed above the arrows are essential for engulfment under all conditions, while those below the arrows are not essential for engulfment, either because they contribute only to the efficienty of engulfment (i.e., SpoIIB) or because they are required only under certain culture conditions (i.e., SpoIIQ and two putative proteins under the control of F, CsfY and CsfZ, which are required only when sporulation is induced by nutrient exhaustion). The completion of septal thinning to the edge of the septum requires SpoIID, SpoIIM, and SpoIIP; this step allows the mother cell membrane to migrate up the side of the forespore (stage IIiii), a process that under certain conditions requires an unknown protein under the control of F, CsfY. The mother cell membrane then migrates across the cell pole, in a step that, under certain sporulation conditions, requires either SpoIIQ or a second protein under F control that depends on SpoIIQ for its expression, CsfZ. Ultimately the engulfing membranes meet at the cell pole (stage IIiv) and fuse to fully enclose the forespore (stage III). This membrane fusion event is the final step of engulfment. After synthesis of the asymmetrically positioned septum, the forespore-specific transcription factor F becomes active in the smaller forespore, thereby allowing activation of E in the larger mother cell. Production of inactive G in the forespore required both F activity and E activity. Following the completion of engulfment, G becomes active in the forespore, allowing activation of K in the mother cell. The stages of engulfment are designated by lowercase Roman numerals as previously described (17, 44).

Genetic studies have defined several stages of engulfment (Fig. 1) (17, 36, 48), starting with septal thinning, during which septal cell wall material is either removed or isomerized, resulting in the septal membranes of the mother cell and forespore being very close to one another, rather than being separated by a thick layer of peptidoglycan (Fig. 1, stage IIi to stage IIii). The spatial and temporal regulation of this process is conferred by SpoIIB, a protein made early in sporulation (28, 35). Three proteins expressed in the mother cell are necessary for septal thinning to proceed to the edges of the septum, SpoIID, SpoIIM, and SpoIIP (13, 26, 45). During the next step, the mother cell membrane migrates up the sides of the forespore and across the cell pole until the migrating membrane projections meet and fuse (Fig. 1, stage IIii to stage III). This membrane fusion event is the final step of engulfment and can readily be assessed in living sporangia, allowing genetic dissection of membrane fusion in bacteria (44).

A comparison of engulfment to phagocytosis raises the question of what role the forespore, which is being engulfed, plays in this process. If engulfment is similar to phagocytosis, then the forespore may play a more passive role in engulfment than the mother cell. Thus far, only one forespore-specific engulfment protein has been identified, SpoIIQ, which was reported to be essential for migration of the mother cell membrane across the cell pole at late stages of engulfment (24). SpoIIQ is a membrane protein with a large external domain that exhibits significant identity to endopeptidases, including proteins from Staphylococcus species that cleave peptide cross-bridges in the cell wall. This suggests that SpoIIQ could play an enzymatic role in engulfment, such as breaking the bridges between the forespore membrane and the cell wall, thereby allowing engulfment to proceed.

Here we demonstrate that SpoIIQ is dispensable for engulfment under certain sporulation conditions, although it is required for the expression of some forespore-specific genes normally transcribed by ^F-containing RNA-polymerase. The ^F factor plays a key role in sporulation, as it is the first transcription factor whose activity is cell specific, and it sets into motion the entire cascade of compartmentalized gene expression (Fig. 1). Using a strain that bypasses the normal requirement for ^F to activate the mother cell transcription factor ^E (56), we found that forespore-specific gene expression is dispensable for engulfment under the same conditions permissive for engulfment in spoIIQ mutants. These results raise the possibility that SpoIIQ is not directly involved in engulfment but rather, is involved in forespore-specific gene expression. They also suggest that although the forespore contributes to engulfment under certain conditions, the essential engulfment machinery is expressed in the mother cell, demonstrating one mechanistic similarity between engulfment and phagocytosis: the relative passivity of the cell that is being engulfed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and genetic manipulations. All strains used in this study are derivatives of wild-type strain PY79 (55), and their genotypes are listed in Table 1. The mutations were introduced into PY79 by transformation (8). All lacZ fusions were transformed into their respective host strains with selection for chloramphenicol resistance.

Growth conditions. B. subtilis cells were grown at 37°C and induced to sporulate either by the resuspension method (46) or by nutrient exhaustion in Difco Sporulation Medium (DSM) (42). Samples were taken and analyzed at various times after the start of sporulation. Sporulation efficiency was monitored after 40 h of sporulation, by heating the cultures to 80°C for 10 min and determining the number of heat-resistant spores per milliliter.

Fusion assay. The membrane fusion assay was performed essentially as described previously (44). Briefly, a 0.5-ml sample of each culture was taken, centrifuged, and resuspended in 0.15 ml of the original culture supernatant. A 2-µl volume of the concentrated sample was added to 1 µl of sporulation salts containing a 5-µg/ml concentration of the membrane-impermeable stain N- (3-triethylammoniumpropyl)-4-{6-[4-(diethylamino)phenyl]-hexatrienyl}py-ridinium dibromide (FM 4-64; Molecular Probes, Eugene, Oreg.), a 1-µg/ml concentration of the DNA stain 4',6-diamidino-2-phenylindole dihydrochloride (DAPI), and a 40-µg/ml concentration of the membrane-permeable stain MitoTracker Green FM (Molecular Probes) on a microscope slide and immobilized with a poly-L-lysine-treated coverslip (44). All stains were obtained from Molecular Probes and resuspended as suggested by the manufacturer.

Microscopy and image analysis. To visualize the cells, an Applied Precision (Issaquah, Wash.) optical sectioning microscope (described in reference 37) was used to collect images spaced 0.1 to 0.2 µm apart throughout the specimen. The images were deconvolved using the Delta Vision deconvolution software as previously described (37). Deconvolved images were saved in TIFF format and imported into Adobe Photoshop.

-Galactosidase assay. Samples (1 ml) from sporulating cultures were harvested, briefly centrifuged to pellet the bacteria, and frozen at 70°C. After thawing, the cells were resuspended in Z buffer, lysozyme treated, and permeabilized with Triton X-100, essentially as described previously (38). -Galactosidase activity was measured as described by Miller (31) with o-nitrophenyl--D-galactopyranoside (ONPG) as the substrate.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

SpoIIQ is required for engulfment only under certain conditions. Because SpoIIQ is the only known engulfment protein required for a late stage of engulfment and also the only engulfment protein expressed by ^F and thereby forespore specific, we investigated further its involvement in engulfment using a membrane fusion assay that unambiguously detects sporangia that have completed engulfment (44). Prior to the completion of engulfment, the forespore membrane is still in contact with the external medium and the engulfing membrane remains contiguous with the mother cell cytoplasmic membrane (e.g., Fig. 1, Stage IIiv). Therefore, both membranes are accessible to membrane-impermeable fluorescent membrane stains such as FM 4-64 (Fig. 2A, arrow 1). However, following the membrane fusion event that is the final step of engulfment, both the forespore membrane and the engulfing membrane are isolated from the external environment (e.g., Fig. 1, Stage III) and therefore remain unstained with FM 4-64, which is unable to cross the mother cell cytoplasmic membrane (Fig. 2A, arrow 2). In contrast, the membrane-permeable stain MitoTracker Green FM is able to stain these membranes both before and after the completion of engulfment (Fig. 2B, arrow 2). Only sporangia that have successfully completed engulfment will exclude FM 4-64 from the forespore membranes while retaining MitoTracker Green FM staining of these membranes.

View larger version (52K): [in this window] [in a new window]   FIG. 2.   Fusion assay for the completion of engulfment in the wild-type strain and the spoIIQ-null mutant. Sporulation was induced as indicated, and samples were stained with FM 4-64 (red) (A, E, I, M, Q, and U). MitoTracker Green FM (green) (B, F, J, N, R, and V) and DAPI (blue) (D, H, L, P, T, and X), as described in the text (Materials and Methods). Overlays of FM 4-64 and MitoTracker Green FM are also shown (C, G, K, O, S, and W). (A to D) The wild-type strain (PY79) at t2.0 in a resuspension sporulation. Arrows 1 indicate the forespore of a sporangium at an early stage of engulfment; the mother cell membrane has begun to curve around the smaller forespore. Arrows 2 show a sporangium that has completed engulfment and membrane fusion; FM4-64 fails to stain the two membranes surrounding the forespore (A), while MitoTracker Green continues to stain membranes (B). Arrows 3 indicate a sporangium in which the forespore stains very faintly with FM4-64, but we consider the sporangium to have completed membrane fusion, as the forespore staining is less bright than that of the cytoplasmic membrane, rather than two- to threefold more bright as predicted based on the number of membranes surrounding the forespore (see Fig. 1). This faint staining of the forespore membranes is only rarely observed and could represent a limited exchange of FM 4-64 between the mother cell cytoplasmic membrane and the outer forespore membrane. (E to H) The spoIIQ-null mutant at t2.0 in a resuspension sporulation. Very few sporangia have completed engulfment and membrane fusion, and in most sporangia the membranes have reached, but not traversed, the cell pole (arrows). (I to L) The wild-type strain at t3.0 in a resuspension sporulation; most sporangia have complete membrane fusion (arrows), and the forespore chromosome fails to stain with DAPI, for reasons that are as yet unclear. (M to P) The spoIIQ mutant at t3.0 in a resuspension sporulation. Engulfment and membrane fusion are complete in many sporangia (arrows). (Q to T) The wild-type strain at t3.0 in a DSM nutrient exhaustion. Most sporangia have completed engulfment and membrane fusion (arrows). (U to X) The spoIIQ-null mutant at t3.0 in a DSM nutrient exhaustion. Very few sporangia have completed engulfment. Some sporangia are blocked prior to the migration of the engulging membranes across the cell pole (arrows 2), and sporangia with flat polar septa are also visible (arrows 1). Bar = 2 µm.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   The percentage of wild-type and spoIIQ-null mutant sporangia that fall into various phenotypic classes. Samples from at least two independent experiments at t1.5, t2.0, t3.0, and t4.0 were scored for the percentage of sporangia that are in various stages of engulfment. At least 70 (and usually more than 100) sporangia were scored from each time point. Grey bar, fully engulfed and fused; shaded bar, engulfed but not fused; black bar, curved polar septa; white bar, straight polar septa. (A) Wild-type strain in DSM nutrient exhaustion; (B) spoIIQ mutant in DSM nutrient exhaustion; (C) wild-type in resuspension sporulation; (D) spoIIQ mutant in resuspension sporulation.

^F is required for engulfment only under certain conditions. SpoIIQ is the only engulfment protein identified so far that is under the control of ^F (24, 48); however, our results indicate that SpoIIQ is not essential for engulfment. If, indeed, no other ^F-dependent engulfment protein exists, then mutants defective in ^F-directed gene expression should have the same conditional engulfment defect as the spoIIQ mutant. Previous studies with a mutated form of ^F (spoIIAC561) which alters promoter specificity but allows ^E activation (17, 24) resulted in a late engulfment defect similar to that observed with the spoIIQ mutant; indeed, our results confirmed these studies. In spoIIAC561 mutant sporangia, 3 h after resuspension, the mother cell membrane has migrated up the sides of the forespore, but not across the cell pole (Fig. 4A to C, arrows 1). We also observed a higher percentage of dispores than previously reported (Fig. 4A to C, arrows 2). However, the spoIIAC561 mutation is not ideal for studying the role of ^F in engulfment, because it has pleiotropic effects on the expression of ^F-dependent genes, such that some genes are overexpressed whereas others are underexpressed (20, 25, 34). Indeed, the mutation allows ^F to recognize some ^B promoters (27), so it is possible that certain ^B-dependent genes are abnormally expressed in the forespore, perhaps interfering with engulfment.

View larger version (48K): [in this window] [in a new window]   FIG. 4.   The engulfment defects of F mutants, as assessed by the membrane fusion assay. FM 4-64 (red) (A, D, G, G', J, and M) and MitoTracker Green FM (green) (B, E, H. H', K, and N) staining are shown, as is an overlay of both membrane stains (C, F, I, I', L, and O). (A to C) The spoIIAC561 mutant strain KP51 was induced to sporulate by resuspension, and samples were harvested at t3.0. Arrows 1 point to a sporangium in which engulfment has proceeded to, but not across, the cell pole. Arrows 2 point to a sporangium in which division has occurred at both polar division sites, producing a disporic sporangium. The absence of F-directed gene expression affects engulfment, as shown in panels D to O. Strain KP331 (P spoIIE-spoIIR spoIIAC::kan) was used to allow activation of E in the absence of F. (D to F) Samples harvested at t2.0 in a resuspension sporulation. Arrows 1 show a sporangium with a flat septum, while arrows 2 indicate a sporangium with a curved septum. (G to I and G' to I') Two separate fields showing samples harvested at t3.0 in a resuspension sporulation. Arrows 1 indicate a sporangium in which the engulfing membrane has migrated around the forespore but membrane fusion is not complete. Arrows 2 indicate a sporangium in which membrane fusion is complete, as indicated by the exclusion of FM 4-64 from the forespore membranes, which are stained with the membrane-permeable MitoTracker Green FM. (J to L) Samples harvested at t2,0 in a DSM nutrient exhaustion. The arrows indicate a sporangium with a flat septum. (M to O) Samples harvested at t3.0 in a DSM nutrient exhaustion. Arrows 1 indicate a sporangium with a flat septum, while arrows 2 indicate a sporangium with a curved septum. Several disporic sporangia (those containing two forespores) are also evident; such sporangia are formed in mutants that fail to activate E. The quantitative analysis of these engulfment phenotypes is shown in Fig. 5. Bar = 2 µM.

spoIIAC

View larger version (43K): [in this window] [in a new window]   FIG. 5.   Quantitative analysis of the engulfment phenotype of strain KP311 (spoIIAC PspoIIE-spoIIR) under different sporulation conditions. The numbers shown are the percentages of nondisporic sporangia at given time points within the identical morphology class; bold type indicates the most prevalent phenotypic class for a given time point. Class 1 sporangia have flat septa (Fig. 4J, arrow); class 2 sporangia have slightly curved septa with little or no migration of the membrane towards the pole (Fig. 4D, arrow 2); in class 3 sporangia, the engulfing membrane has moved up the side of the forespore but not across the pole (Fig. 4A, arrow 1); in class 4 sporangia, the migrating membranes have moved around the forespore, but membrane fusion is not complete (Fig. 4G, arrow 1), whereas membrane fusion is complete in class 5 sporangia (Fig. 4G', arrow 2). The frequency of sporangia was 30% at t2.0, 42% at t3.0, and 31% at t4.0 in a resuspension sporulation and 14% at t2.0, 16% at t2.0, and 20% at t4.0 in a DSM nutrient exhaustion. Disporic sporangia comprised between 2 and 6% of the cells in the culture, likely reflecting the inefficient activation of E when SpoIIR production is uncoupled from forespore-specific gene expression. Such sporangia were excluded from the analysis of engulfment phenotypes because dispores lack a mother cell chromosome, which eliminates all mother cell-specific gene expression, thereby preventing engulfment.

spoIIQ-null mutant lacks ^G activity. To determine why SpoIIQ is required for sporulation even when it is not required for engulfment, we investigated the activities of late sporulation-specific sigma factors (^G and ^K) in the spoIIQ-null mutant. Activation of ^G appears to depend on the completion of engulfment (34), so using the sspB-lacZ fusion that is under the control of ^G, we first investigated whether spoIIQ mutants have normal ^G activity when they can complete engulfment. Surprisingly, the spoIIQ mutant, under conditions in which engulfment has been completed, shows a reduction in ^G activity significantly more severe than that seen with a spoIID mutant, which never completes engulfment (Fig. 6A). Similar results were obtained when the protein products of the ^G-dependent sspB and sspA genes, the /-type small, acid-soluble proteins (SASP) were assayed by immunofluorescence microscopy (data not shown). Thus, the spoIIQ null mutant appears to almost completely lack ^G activity.

View larger version (26K): [in this window] [in a new window]   FIG. 6.   The effect of spoIID and spoIIQ mutations on the expression of several F-dependent lacZ fusions. (A to D) Strains containing sspB-lacZ (A), spoIIIG-lacZ (B), sspE(2G)-lacZ (C), and spoIIR-lacZ (D) chromosomal lacZ fusions were induced to sporulate by resuspension sporulation. Samples were taken every 30 min, and -galactosidase activity was measured. , wild-type strain; , spoIID mutant; , spoIIQ mutant. Strains used for panels B through D also contain the spoIIIG mutation to eliminate -galactosidase activity contributed by G. (E) Strain KP234 () [spoIIG1 spoIIIAC(V233A) sspB-lacZ] in which the promoter specificity of F has been changed to that of G. The spoIIIG mutation inactivates G, allowing use of the sspB-lacZ fusion (which is normally dependent on G) to monitor F activity. Strain KP572 () is KP234 plus spoIIQ::spc.

spoIIQ null mutant lacks ^F-directed expression of the gene encoding ^G. Because ^G is under both transcriptional and posttranslational control (19, 30), we investigated whether the lack of ^G activity in a spoIIQ mutant is due to a lack of expression of the gene encoding it, spoIIIG, or to a lack of activation of ^G following completion of engulfment. The spoIIIG gene is initially transcribed by ^F (49, 50); however, ^G is autoregulatory and thus, once activated, can recognize the spoIIIG promoter and maintain its own synthesis (19, 49). We therefore utilized a ^G-null background (spoIIIG1 or spoIIIG::neo) to avoid -galactosidase units contributed by ^G. Using a translational spoIIIG-lacZ fusion that is well recognized by ^F-containing RNA polymerase (49), we found that ^F-directed expression of spoIIIG is abolished in the spoIIQ mutant (Fig. 6B). The same effect is observed with this spoIIIG-lacZ fusion integrated at the amyE locus (data not shown). In contrast, the spoIID mutant showed normal expression of spoIIIG (Fig. 6B), suggesting that the decreased ^G activity of this mutant is due not to a failure to express the protein but rather to a failure to fully activate ^G, possibly as a consequence of a failure to complete engulfment. The spoIIIG gene is unique among ^F-transcribed genes, as it also requires a factor under the control of ^E for its expression (34). We therefore investigated ^E activity in the spoIIQ mutants, using the spoIID-lacZ reporter gene, and found normal ^E activity (not shown) as previously reported (24). Thus, the defect in ^G activity in the spoIIQ mutant appears to be due to the lack of ^F-directed expression of the gene encoding ^G, spoIIIG.

spoIIQ-null mutant affects the expression of some, but not all, ^F-dependent genes. We next investigated if the spoIIQ mutant affected expression of other ^F-dependent genes or if the phenomenon was specific to the spoIIIG promoter. We made use of another ^F-dependent promoter, sspE(2G)-lacZ (50). The sspE promoter is normally recognized by ^G; however, with the substitution of two guanines at the 15 and 16 positions, it can be transcribed by ^F-containing RNA polymerase and is well transcribed in a mutant lacking ^G protein (50). The spoIIQ mutant shows 24% the level of expression of sspE(2G)-lacZ of the wild-type strain, a less severe effect than that seen with expression of spoIIIG (Fig. 6C). A spoIID mutant has slightly decreased expression of sspE(2G)-lacZ relative to the wild-type strain (77% at t_5.0), although the reduction is less than that caused by the spoIIQ mutation (24% at t_5.0) (Fig. 6C). A third promoter affected by the spoIIQ mutation is gpr, a promoter recognized by both ^F and ^G, whose ^F-dependent transcription was therefore monitored in a spoIIIG mutant background. The spoIIQ mutant showed about 45% the level of -galactosidase activity as the wild-type strain at t_4.0 (data not shown). However, several other ^F-dependent promoters were unaffected by the spoIIQ mutation, including spoIIR-lacZ (Fig. 6D), csfA-lacZ, and csfB-lacZ (reference 6 and data not shown). Therefore, the spoIIQ mutation has no effect on expression of some ^F-dependent genes (spoIIR, csfA, and csfB), while it has moderate [gpr and sspE(2G)] or strong (spoIIIG) effects on others.

SpoIIQ is required for the ^F-dependent expression of genes normally under the control of ^G. The decrease in ^F-dependent gene expression in the spoIIQ mutant might be due to a requirement for SpoIIQ to modify the activity of an activator or repressor of certain ^F-dependent promoters, or it might be due to a requirement for SpoIIQ to allow high levels of ^F-directed gene expression. To distinguish between the first and second possibilities, we made use of a strain with a mutation in the gene encoding ^F (spoIIAC V233A), which allows ^F to recognize promoters normally recognized by ^G (27, 43). We inactivated ^G (with a spoIIIG1 mutation) and used a ^G-dependent promoter (sspB-lacZ) to monitor activity of the altered ^F. If SpoIIQ acts via a repressor or activator of transcription of ^F-dependent genes, then it seems unlikely that it would be required for expression of an artificial ^F gene such as sspB-lacZ. However, the spoIIQ-null mutant showed a dramatic decrease in the ^F-directed expression of this gene (Fig. 6E). A similar though somewhat weaker effect is observed with another unnatural promoter (sspE(2G), as described earlier (Fig. 6C). Because SpoIIQ is required for expression from two unnatural ^F promoters, it seems unlikely that SpoIIQ decreases ^F-directed transcription via an activator or repressor of transcription, unless this transcription factor acted on both ^F- and ^G-dependent promoters. It is therefore possible that SpoIIQ is required for the full activation of the very sigma factor required for its expression. However, it is also possible that SpoIIQ does not play a direct role in forespore-specific gene expression but rather is required for some aspect of forespore physiology essential for the expression of certain forespore-specific genes.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The role of forespore-specific gene expression in engulfment has previously been difficult to investigate, due both to the lack of methods by which to assess the completion of engulfment and to the absolute dependence of engulfment on the mother cell transcription factor ^E, whose activity requires the forespore-specific transcription factor ^F. We have recently developed fluorescent membrane stain-based assays for engulfment (37) and for the membrane fusion event that is the final step of engulfment (44). Using these assays, we have investigated the role of forespore-specific gene expression in engulfment, making use of a strain in which ^E is able to become active in the absence of ^F (56) and also the role of the only forespore-expressed engulfment protein described to date, SpoIIQ (24). Our results indicate that forespore-specific gene expression is not essential for engulfment, which, under certain conditions, can be completed without SpoIIQ and also in the complete absence of forespore-specific gene expression (albeit inefficiently in the latter case). These results suggest that the essential components of the engulfment machinery are within the mother cell. Thus, engulfment, which bears a superficial resemblance to phagocytosis, shares certain mechanistic similarities to phagocytosis, as in both processes the engulfed cell (in this case the forespore) plays a relatively passive role, whereas the engulfing cell (in this case the mother cell) plays a more active role.

We did, however, find that under certain conditions, forespore-specific gene expression is essential for engulfment. When sporulation is induced by nutrient exhaustion, the absence of ^F-directed gene expression results in an early engulfment defect, whereas the absence of SpoIIQ results in a late engulfment defect similar to that previously described (24). These results suggest that there is at least one unidentified engulfment protein encoded by ^F, CsfY (controlled by ^F), that is required for the movement of the mother cell membrane up the sides of the forespore only when sporulation is induced by nutrient exhaustion (Fig. 1, stage IIii to stage IIiii) but not by resuspension. Because only a small proportion of the spoIIQ mutant sporangia were able to complete mother cell membrane migration across the cell pole after nutrient exhaustion, it is possible that SpoIIQ is directly required for this late step of engulfment. However, because spoIIQ mutants have defects in ^F-dependent gene expression (as discussed below), it is also possible that unidentified engulfment protein under the control of ^F whose expression depends on SpoIIQ (CsfZ) is conditionally required for this step of engulfment (Fig. 1, stage IIiii to stage IIiv).

The medium dependence of forespore-specific gene expression in engulfment is unanticipated, and may reflect differences in the barriers to the migration of the engulfing membranes under the two conditions used here. It has previously been proposed that bridges between the forespore membrane and the cell wall, such as that formed by lipoteichoic acid, would provide a steric barrier to the engulfing membranes and that such bonds would be removed prior to engulfment (36). Forespore-expressed hydrolytic enzymes would be ideally positioned for this task. The two sporulation conditions used here differ both in medium composition (that used for nutrient exhaustion is a more complex medium) and in the means by which sporulation is induced (in a nutrient exhaustion, the bacteria are gradually starved as they enter stationary phase, whereas in resuspension sporulation, starvation occurs rapidly, as exponentially growing bacteria are resuspended in a salt solution). The two methods could thereby result in an altered cell wall structure, which is modified both in response to medium composition (2, 22, 23) and during the transition to stationary phase (4, 5, 23). The future identification of genes required for engulfment when sporulation is induced by nutrient exhaustion but not by resuspension will help to elucidate the physical basis for the unusual effect of media on the requirement of forespore-specific gene expression in engulfment.

Although SpoIIQ is not essential for engulfment, it is essential for sporulation under all conditions, apparently because it is required for the expression of spoIIIG, the gene which encodes the second forespore-specific sigma factor, ^G. spoIIIG is initially transcribed by ^F-containing RNA polymerase, although the spoIIIG promoter is poorly recognized by ^F, and is later transcribed by ^G-containing RNA polymerase (49). The spoIIQ-null mutation also reduces expression of two other genes transcribed by ^F- and ^G-containing RNA polymerase [gpr and sspE(2G)] but does not affect expression of the genes transcribed solely by ^F-containing RNA polymerase that we examined [spoIIR, csfA, and csfB 6]. It is therefore possible that SpoIIQ governs the activity of a transcription factor that affects expression of promoters transcribed by both ^F and ^G, perhaps by controlling the activity of an activator or repressor of such genes or by allowing the full activation of ^F. However, it is not immediately apparent how SpoIIQ, and extacellular protein, could modify the ^F activation pathway (1, 7, 9, 10, 32) and it remains possible that the spoIIQ mutation adversely affects forespore viability or transcription and indirectly reduces expression of certain forespore specific genes.

However, regardless of the mechanism by which SpoIIQ affects forespore-specific gene expression, our results confirm the existence of two phases of ^F-dependent gene expression (34), with the initial expression of certain ^F-dependent genes such as spoIIR. During the second phase, the remaining ^F-dependent genes, such as spoIIIG are expressed (21, 34). We suggest two possible benefits of dividing ^F-directed gene expression into two phases. First, if SpoIIQ is directly involved in forespore-specific gene expression, it could contribute to the compartmentalization of ^F activity to the forespore by providing a positive feedback loop for ^F-dependent gene expression in the forespore. Second, having two phases of ^F-directed gene expression may serve as a checkpoint to delay production of ^G until the onset of engulfment, which may ensure more effective coupling of ^G and ^K activation to the completion of engulfment.

Does SpoIIQ play an active role in engulfment, or is the conditional engulfment defect of spoIIQ mutants a secondary consequence of decreased forespore-specific gene expression? SpoIIQ is a protein with a large extracellular domain that shows homology to endopeptidases such as lysostaphin, which cleaves peptide cross-bridges in the cell wall. Certainly, this potential enzymatic activity suggests an active role in engulfment, as does the observation that SpoIIQ initially accumulates in the septum and then spreads around the forespore membrane (24). It is possible that SpoIIQ plays a dual role, contributing to membrane migration as well as to forespore-specific gene expression, possibly coupling the full expression of ^F-dependent genes to the onset of engulfment. However, until it is possible either to genetically separate the two functions of SpoIIQ or to achieve full ^F-directed gene expression in the absence of SpoIIQ, this question will remain open.