Numbers of Individual Nutrient Germinant Receptors and Other Germination Proteins in Spores of Bacillus subtilis

Abstract

Germination of dormant Bacillus subtilis spores with specific nutrient germinants is dependent on a number of inner membrane (IM) proteins, including (i) the GerA, GerB, and GerK germinant receptors (GRs) that respond to nutrient germinants; (ii) the GerD protein, essential for optimal GR function; and (iii) SpoVA proteins, essential for the release of the spore-specific molecule dipicolinic acid (DPA) during spore germination. Levels of GR A and C subunit proteins, GerD, and SpoVAD in wild-type spores were determined by Western blot analysis of spore fractions or total disrupted spores by comparison with known amounts of purified proteins. Surprisingly, after disruption of decoated B. subtilis spores with lysozyme and fractionation, ∼90% of IM fatty acids and GR subunits remained with the spores' insoluble integument fraction, indicating that yields of purified IM are low. The total lysate from disrupted wild-type spores contained ∼2,500 total GRs/spore: GerAA and GerAC subunits each at ∼1,100 molecules/spore and GerBC and GerKA subunits each at ∼700 molecules/spore. Levels of the GerBA subunit determined previously were also predicted to be ∼700 molecules/spore. These results indicate that the A/C subunit stoichiometry in GRs is most likely 1:1, with GerA being the most abundant GR. GerD and SpoVAD levels were ∼3,500 and ∼6,500 molecules/spore, respectively. These values will be helpful in formulating mathematic models of spore germination kinetics as well as setting lower limits on the size of the GR-GerD complex in the spores' IM, termed the germinosome.

INTRODUCTION

Spores of various Bacillus species are dormant and resistant to harsh environmental conditions but can revert to growing cells through the process of spore germination, which is triggered by specific nutrients (1, 2). Spores of many Bacillus and Clostridium species are significant agents of disease and food spoilage, and thus, there is much interest in efficient methods for spore eradication. Since spores lose their resistance properties upon germination, it is possible that efficient germination may be one component of a spore eradication strategy. The spore is composed of a number of layers, some of which play important roles in germination (1). One such layer is the inner membrane (IM), a lipid bilayer with extremely low permeability in which lipid probes and presumably proteins are immobile prior to germination (3). The IM also contains a number of proteins crucial for spore germination, including the nutrient germinant receptors (GRs) that trigger spore germination in response to specific nutrient germinants (4–9).

Bacillus subtilis, the model organism for bacterial spore germination, has three major GRs, GerA, GerB, and GerK. GerA responds to l-alanine or l-valine, and GerB and GerK together respond to a mixture of l-asparagine, d-glucose, d-fructose, and K⁺ (AGFK) (2, 10, 11). These GRs colocalize in a cluster in the IM termed the germinosome, and this clustering appears to be important for efficient GR function (12). GerA, GerB, and GerK are homologous GRs each encoded by tricistronic operons, and each contain A, B, and C subunits (2, 10, 11). The A and B subunits are almost certainly integral IM proteins with multiple transmembrane domains, whereas the C subunit is largely hydrophilic and is attached to the IM by a diacylglycerol anchor (2, 13, 14). Disruption of any cistron encoding a GR's A, B, or C subunit eliminates germination with that GR's corresponding germinant or germinant mixture, indicating that all three GR subunits are essential for that GR's function. The GerD protein is also essential for rapid GR-dependent spore germination and for germinosome assembly, and this protein is also present in the spore's IM, where it is presumably anchored by a covalently attached diacylglycerol moiety that is also present in GR C subunits (9, 12, 13). Loss of this diacylglycerol anchor eliminates both GerD and GR C protein function, consistent with these proteins' inner membrane location, and almost certainly on the outer leaflet of the inner membrane (13, 15, 16). However, how GRs recognize specific nutrients, how nutrient recognition triggers the downstream events in spore germination, and how GerD influences GR function are largely unknown.

Analyses of a number of engineered strains of B. subtilis have shown that rates of spore germination with various nutrient germinants can be significantly altered when the cognate GR(s) is overexpressed (7, 17). For example, l-valine germination is significantly increased when GerA is overexpressed, and these elevated GerA levels inhibit the GerB/GerK pathway's response to AGFK, although levels of GerBC and GerKA are unchanged (7). In addition, (i) decreases in GR subunit levels by sporulation in nutrient-poor medium are accompanied by lowered spore germination rates with nutrient germinants, (ii) spores that are termed superdormant and germinate extremely slowly with nutrient germinants have very low GR levels, and (iii) spores lacking the regulatory protein SpoVT have higher GR subunit levels and higher rates of nutrient germination (6, 18, 19). These data strongly suggest that GR levels are a major factor determining rates of spore germination.

Since, as noted above, GR levels are almost certainly a major factor determining rates of spore germination with nutrients, it will be imperative to know the numbers of GR molecules in a wild-type spore in order to formulate mathematic models of spore germination and spore germination kinetics. In addition, knowledge of the GRs' subunit stoichiometry seems likely to be important for an eventual understanding of overall GR structure and function. The average number of GerBA molecules per spore was determined more than 10 years ago by Western blot analysis of the IM fraction of wild-type spores (5), but at that time, no antisera against other GR subunits were available. Consequently, in current work, we have determined the average numbers of molecules per spore of six different germination proteins, GerAA, GerAC, GerBC, GerKA, and GerD, as well as the SpoVAD protein involved in the movement of a major spore small molecule, dipicolinic acid (DPA), across the spore's IM early in germination. The numbers determined in this work may ultimately be useful for (i) determining GR subunit stoichiometry, (ii) formulating predictive mathematic models of how GRs function in spore germination, and (iii) setting approximate limits on the size of the germinosome in the spore's IM.

MATERIALS AND METHODS

Spore preparation and purification.

The B. subtilis strains used in this work are isogenic derivatives of strain PS832, a prototrophic laboratory 168 strain. Specific strains used were (i) PS533 (wild type) carrying plasmid pUB110 encoding kanamycin resistance (10 μg/ml) (20), (ii) PS3478 (PsspD::gerK) (17) in which the gerK operon has been placed under the control of the moderately strong forespore-specific promoter of the sspD gene (PsspD), and (iii) PS4255 (PsspD::gerA PsspB::gerB ΔgerK). The latter strain was constructed by using chromosomal DNA from strain FB58 (PsspB::gerB), resistant to spectinomycin (Sp^r) (100 μg/ml) (5) and with the gerB operon under the control of the strong forespore-specific sspB promoter (PsspB), to transform strain PS3726 (PsspD::gerA ΔgerK) (17) with selection for resistance to spectinomycin (100 μg/ml). Spores of all strains were prepared at 37°C on 2× Schaeffer's glucose medium agar plates without antibiotics and harvested and purified as described previously (21, 22). The spore preparations used were essentially free (>98%) of growing or sporulating cells, germinated spores, and cell debris, as seen by phase-contrast microscopy.

Preparation of spore fractions.

Dormant B. subtilis spores (∼15 mg [dry weight]) of various strains were routinely chemically decoated to remove the outer membrane and much spore coat protein and disrupted by bursts of sonication after lysozyme treatment, and the disrupted spores were fractionated as previously described (5, 7, 18, 19, 23). Briefly, after low-speed centrifugation of the disrupted spores for 5 min at 16,000 × g, the supernatant fluid was saved and termed the low-speed supernatant fluid (LS) fraction. The low-speed pellet obtained after lysozyme treatment and sonication that contained residual coat material and most likely some degraded peptidoglycan was washed twice by low-speed centrifugation; the final pellet was termed the integument (I), and the pooled washes were termed the wash (W) fraction. To isolate the IM fraction (M), the LS fraction was centrifuged at 100,000 × g for 1 h at 4°C. The supernatant fluid was termed the soluble fraction (S), and the IM pellet (M) was solubilized in a small volume of TEP buffer (50 mM Tris-HCl [pH 7.4], 5 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride) containing 1% Triton X-100. To solubilize IM protein that was not completely removed from the I fraction, this fraction was suspended in Laemmli sample buffer (Bio-Rad, Hercules, CA) plus 5% 2-mercaptoethanol and incubated at 23°C for 1 to 2 h. The mix was then bath sonicated for 1 to 2 min and centrifuged at 23°C and 16,000 × g for 5 min, and the supernatant fluid was stored frozen for later Western blot analysis. In one experiment, decoated spores (∼6 mg [dry weight]) were disrupted by five 1-min periods of mechanical agitation at 4°C with 0.1-mm-diameter zirconia-silica beads (∼2 g) in a Mini-Beadbeater (BioSpec Products, Bartlesville, OK) in 1 ml of TEP buffer containing 1 μg each of RNase A and DNase I and 20 μg of MgCl₂. Spore breakage was >80%, as determined by phase-contrast microscopy, and the extract was fractionated as described above. In yet another experiment, ∼20 mg (dry weight) of intact wild-type spores was broken by dry rupture with glass beads, and the powder was extracted and fractionated as described above and previously (9).

Preparation of the total spore lysate.

Dormant wild-type B. subtilis spores (∼20 mg [dry weight]) were chemically decoated, treated with lysozyme, and then disrupted with bursts of sonication in TEP buffer, as described above. After the sonication treatment, the total disrupted spore samples were incubated for 2 h at 23°C in TEP buffer containing 1% SDS and 150 mM 2-mercaptoethanol or a 1:1 mixture of TEP buffer with Laemmli sample buffer (final concentration of 1% SDS plus 2.5% 2-mercaptoethanol). After incubation, the samples were bath sonicated for 2 min and centrifuged at 23°C and 16,000 × g for 5 min, and the supernatant fluid was saved as the total lysate (L). The pellet fraction isolated from the total lysate was resuspended again in buffer containing SDS and 2-mercaptoethanol, incubated for another 2 h at 23°C, bath sonicated for 2 min, and centrifuged at 23°C and 16,000 × g for 5 min, and the supernatant fluid was saved as the washed lysate (WL) fraction.

Western blot analysis.

In individual experiments, aliquots from the same amounts of various spore fractions (termed 1×) were run on SDS-polyacrylamide (11 to 15%) gel electrophoresis (PAGE) gels and subjected to Western blot analysis with antisera specific for various germination proteins, as described previously (6, 7, 18, 19). In some cases, after the use of one antiserum, the polyvinylidene difluoride (PVDF) Western blot membranes were stripped with Restore Western blot stripping buffer (Thermo Scientific, Rockford, IL) for 15 to 20 min at 37°C and then probed with antiserum against another germination protein.

Protein quantitation by Western blot analysis.

Quantitation by Western blot analysis was done with M or LS fractions from wild-type spores and the lysate (L) from wild-type spores that were analyzed on the same Western blot with different amounts of the purified (>95%) B. subtilis proteins used to raise various antisera. These proteins were obtained and purified as previously described (6, 7, 18, 19, 24) and were (i) amino acids (aa) 2 to 239 of GerAA (27.1 kDa), (ii) GerAC without its signal sequence (40.5 kDa), (iii) GerBC without its signal sequence (40.4 kDa), (iv) aa 39 to 276 of GerKA (26.7 kDa), (v) GerD without its signal sequence (18.4 kDa), and (vi) full-length SpoVAD (36.0 kDa). The concentrations of these proteins were determined by their absorbance at 280 nm and their molecular extinction coefficient at this wavelength. The intensities of all bands on Western blots were quantitated by using ImageJ.

For calculation of the average numbers of germination proteins per spore, the intensity of the band given by a germination protein in fractions or lysates from wild-type spores was compared to the band intensities from known amounts of the appropriate purified germination protein described above. The amount of the purified protein equivalent to the amount of the corresponding protein in the spores' LS or M fractions and in the total spore L fraction was divided by this protein's molecular weight, and the resultant mole value was multiplied by Avogadro's number and then divided by the number of spores corresponding to the amount of the LS, M, or L fraction analyzed to give the number of molecules of the protein per spore in the LS, M, or L fraction. Values calculated in the M or LS fraction were then corrected for the percentages of the various germination proteins in the M, S, W, and I fractions determined with spores from strain PS4255 (PsspD::gerA PsspB::gerB ΔgerK) or PS3478 (PsspD::gerK) (Table 1). Values calculated in the L fraction were also corrected for the percentages of the various proteins in the L and WL fractions determined with wild-type spores, although amounts in the WL fractions were invariably minimal (GerAA and GerBC, ≤10%; GerAC, GerKA, GerD, and SpoVAD, <1%). All of the latter values were determined by two or three analyses of at least two independent batches of spores of each strain examined.

Table 1.

Percentages of different germination proteins in various spore fractions^a

Germination protein	% of total protein in spore fraction (range)
	M	S	I	W
GerAAb	6 (5–7)	0.4 (0.3–0.4)	92 (89–94)	2 (1–3)
GerACb	6 (4–8)	1 (0.6–2)	91 (87–94)	2 (1–2)
GerBCb	5 (4–6)	0.4 (0.3–0.5)	94 (92–95)	2 (1–2)
GerKAc	6 (4–7)	1 (0.2–2)	92 (89–95)	1 (0.5–2)
GerDb	19 (15–22)	55 (52–58)	21 (15–26)	6 (5–7)
SpoVADb	2	88 (83–92)	≤4	≤5 (2–7)

^aRelative levels of germination proteins in spore fractions were determined by Western blot analyses of fractions from dormant PS4255 (PsspD-gerA PsspB-gerB ΔgerK) or PS3478 (PsspD-gerK) spores, as described in Materials and Methods. All values are the averages from two or three determinations on two independent spore preparations, and the ranges of these values are shown in parentheses.

^bRelative levels of GerAA, GerAC, GerBC, GerD, and SpoVAD were determined in fractions from PS4255 spores.

^cRelative levels of GerKA were determined in fractions from PS3478 spores.

Analysis of fatty acids in spores and spore fractions.

Ninety-four milligrams (dry weight) of spores was chemically decoated as previously described (25). After decoating, 20 mg of dry spores was fractionated, as described above, to isolate the I and M fractions. Twenty mg of dry decoated spores and lyophilized I and M fractions from 20 mg dry decoated spores were saponified and subjected to direct fatty acid methyl ester analysis (Direct-FAME) by Microbial ID, Newark, DE. Details of the procedure for sample workup and Direct-FAME are available from Microbial ID.

RESULTS

Levels of GR subunits in the M fraction from wild-type spores.

Since GR subunits have been localized to the IM (5, 7, 12, 18, 19), the IM pellet (M fraction) was isolated from wild-type spores to compare levels of GR subunits in this fraction to known amounts of purified GerAA protein. As expected (6, 7), the M sample gave a GerAA band at ∼54 kDa in Western blot analysis with the GerAA-specific antiserum (Fig. 1A). In this particular experiment, the band from ∼0.5 ng of the purified truncated GerAA protein used to generate the antiserum against GerAA and the GerAA band in the M fraction from 2 × 10⁸ wild-type spores had the same intensities, as determined by ImageJ. Similarly, in one experiment, the intensity of the ∼41-kDa band of GerAC in the M fraction from 2 × 10⁸ spores was equal to that of the band from ∼1 ng of the purified GerAC antigen (Fig. 1B). Analogous experiments were carried out to determine the relative amounts of the GerBC and GerKA proteins in the same M fraction from wild-type spores (data not shown). All of these analyses were carried out two or three times and with two or more independent spore preparations.

Fig 1.

Western blot analysis of GerAA and GerAC levels in the M fraction of wild-type spores. The M fraction from wild-type (PS533) spores was isolated, and an aliquot from 2 × 10⁸ spores was run on SDS-PAGE gels with known amounts of purified GerAA (A) or GerAC (B) antigens, as described in Materials and Methods. Note that the purified GerAC antigen was not diacylglycerylated. Proteins on the gel were then subjected to Western blot analysis with antisera against either GerAA (A) or GerAC (B). The intensity of the GerAA band in the spores' M fraction at ∼54 kDa was compared to the intensities of known amounts of purified GerAA antigen of ∼27 kDa, and the intensity of the spores' GerAC band in the M fraction from 2 × 10⁸ spores was compared to the intensities of known amounts of the ∼41-kDa GerAC antigen. Note that the GerAC protein from spores is diacylglycerylated, while the GerAC antigen is not. MW, molecular weight markers (in thousands).

Percentages of GR subunits in various spore fractions and numbers per spore.

While the results described above could be used to determine the average numbers of molecules of GR subunits per wild-type spore, it was essential to first determine what percentages of GR subunits were recovered in the M fraction. Most of this initial analysis was carried out with spores of strains PS3478, which overexpresses the GerK GR, and PS4255, which overexpresses the GerA and GerB GRs and lacks GerK. Spores of these strains were used because having elevated levels of GR subunits made quantitation of GR subunit levels in the spores' low-speed supernatant (LS) and integument (I) fractions easier, since these two fractions have very large amounts of protein. This analysis showed that ∼85% of all GR subunits in the LS fraction were recovered in the M fraction (Fig. 2, Table 1, and data not shown). These recoveries are consistent with previous work examining the distribution of the GerBA subunit in spore fractions (5). Presumably, the small amounts of GR subunits in the soluble (S) fraction are in small membrane fragments that do not pellet readily during ultracentrifugation.

Fig 2.

Western blot analysis of levels of germination proteins in the LS, S, and M fractions of spores overexpressing GerA and GerB. The LS, S, and M fractions from PS4255 spores (PsspD-gerA PsspB-gerB ΔgerK) were isolated, and fractions from various amounts of spores (1× is from the same amount of spores in all fractions) were run on SDS-PAGE gels and subjected to Western blot analysis with specific antisera, as described in Materials and Methods. The strips for GerAA and SpoVAD are from the same Western blot where the PVDF membrane was first probed with anti-GerAA serum and then stripped and probed with anti-SpoVAD serum. MW, molecular weight markers (in thousands).

Surprisingly, Western blot analysis of the I fraction revealed that levels of GR subunits in this fraction were ≥10-fold higher than in the M fraction from spores of strains PS4255 (GerAA, GerAC, and GerBC) and PS3478 (GerKA) (Fig. 3 and data not shown). This was also the case when the distribution of GerAA and GerBC was examined in decoated wild-type (PS533) spores disrupted either by lysozyme or mechanical disruption alone or when vortexing vigorously was substituted for sonication of lysozyme-disrupted spores (data not shown). In contrast to the high levels of GR subunits in the I fraction, the pooled washes (W) of the I fraction had ≤25% of the amount of the GR subunits present in the M fraction (Fig. 3, Table 1, and data not shown).

Fig 3.

Western blot analysis of levels of GR proteins in the LS, I, and W fractions of spores of various strains. Various amounts of protein in the LS, I, and W fractions from dormant spores (1× is from the same amount of spores in all fractions) of strain PS4255 (PsspD-gerA PsspB-gerB ΔgerK) (A) or PS4255 (PsspD-gerA PsspB-gerB ΔgerK) for GerAC and PS3478 (PsspD-gerK) for GerKA (B) were subjected to Western blot analysis with various antisera, as described in Materials and Methods. MW, molecular weight markers (in thousands).

The percentages of GR subunits in various spore fractions and the levels of GR subunits per spore in the M fraction now allowed calculation of the number of GR subunits per spore based on analyses of two independent spore preparations (Table 2). The numbers of GerAA and GerAC GRs per spore are approximately the same, and therefore, these proteins are likely at a 1:1 ratio in the GerA GR. These numbers further indicate that the GerA GR is likely the most abundant GR in spores and that the GerB and GerK GRs are present at similar levels in spores.

Table 2.

Numbers of various germination proteins in wild-type spores^a

Protein	No. of germination protein molecules/spore
	Fractionated spores	Total disrupted spores	Avgb
GerAA	1,020	1,000	1,010
GerAC	1,370	1,150	1,260
GerBC	600	800	700
GerKA	670	790	730
GerD	3,150	3,800	3,480c
SpoVAD	8,820	4,040	6,430d

^aNumbers of germination protein molecules per spore were determined by two to three Western blot analyses of various fractions from two or more independent batches of spores or from two to three Western blot analyses of the L fraction from total disrupted wild-type (PS533) spores, as described in Materials and Methods.

^bAll values shown are averages of the numbers determined from the fractionated and total disrupted spores and are ±20%.

^cThis value is similar to the value determined previously (9).

^dThis value is ∼3-fold lower than what was determined previously (8) (see Discussion).

Distribution of GerD and SpoVAD in various spore fractions.

Based on previous results, GerD and SpoVAD are more abundant than GerBA in wild-type spores (5, 8, 9). Therefore, detection of these proteins should not be as challenging in spores where only GRs are overexpressed, because GerD and SpoVAD levels remain at wild-type levels in such strains (7). In contrast to GR subunits, GerD and SpoVAD were much more abundant in the S fraction than in the M fraction from both PS4255 (Fig. 2 and Table 1) and PS3478 (data not shown) spores as well as wild-type spores disrupted with lysozyme with or without sonication or by mechanical breakage alone (data not shown). The high levels of GerD and SpoVAD in the S fraction might be considered surprising, since there are data indicating that both GerD and at least one SpoVA protein are in the IMs of intact spores (8, 9, 12, 26). However, previous work has also found significant amounts of these proteins in the S fraction of disrupted spores (8, 9, 26). While the particular disruption procedure used significantly alters the distribution of IM proteins in various fractions, it is notable that the distribution of GerD in the I, M, and S fractions found in the current work is not too different from that seen previously when spores were broken only by mechanical agitation in liquid with glass beads prior to fractionation (26).

Since high levels of GerD and SpoVAD were found in the S fraction, the LS fraction rather than the M fraction was used to quantitate the numbers of these molecules in wild-type (PS533) spores. Quantitative Western blot analyses showed that the GerD band in the LS fraction from 4 × 10⁷ spores had the same intensity as the band from ∼3 ng of purified GerD antigen (Fig. 4A), while the SpoVAD band in the LS fraction from 1.3 × 10⁷ spores had the same intensity as the band from ∼7 ng of purified SpoVAD antigen (Fig. 5A).

Fig 4.

Western blot analysis of levels of GerD in fractions from spores of various strains. (A) An aliquot of the LS fraction from 4 × 10⁷ wild-type (PS533) spores was subjected to Western blot analysis along with known amounts of purified GerD using antiserum against GerD, as described in Materials and Methods. Note that the endogenous GerD protein migrates lower than the expected 21 kDa, perhaps due to its covalent modification. The intensity of the GerD band at ∼17 kDa in the LS fraction was compared to the intensities of the bands given by various amounts of the purified GerD antigen. Note that the purified GerD antigen is not diacylglycerylated. (B) Various amounts of protein from the I, LS, and W fractions from dormant PS4255 spores (PsspD-gerA PsspB-gerB ΔgerK) (1× is from the same amount of spores in all fractions) were subjected to Western blot analysis with antiserum against GerD, as described in Materials and Methods. MW, molecular weight markers (in thousands).

Fig 5.

Western blot analysis of the levels of SpoVAD in various fractions of spores of several strains. (A) An aliquot of the LS fraction from 1.3 × 10⁷ wild-type (PS533) spores was subjected to Western blot analysis along with known amounts of purified SpoVAD using anti-SpoVAD serum, as described in Materials and Methods. The intensity of the SpoVAD band at 36 kDa in the LS fraction (the upper band denoted by the asterisk) was compared to the intensities of bands given by various amounts of the purified SpoVAD. (B) Various amounts of protein from the I, LS, and W fractions from dormant PS4255 spores (PsspD-gerA PsspB-gerB ΔgerK) (1× is from the same amount of spores in all fractions) were subjected to Western blot analysis with antiserum against SpoVAD, as described in Materials and Methods. MW, molecular weight markers (in thousands).

Since most GR subunits were found in the spores' I fraction and with minimal amounts in the W fraction, we also determined the levels of GerD and SpoVAD in these fractions from PS4255 spores. Compared to levels in the LS fraction from the same amount of spores, GerD levels in the W and I fractions were ∼1 and 20%, respectively, of what was detected in the LS fraction (Fig. 4B and Table 1), and very similar results were obtained when wild-type spores were analyzed (data not shown). Even lower levels of SpoVAD were detected in the W and I fractions than in the LS fraction (Fig. 5B and Table 1), and again, this was the case with wild-type spores (data not shown). The percentages of GerD and SpoVAD in the M, S, W, and I fractions (Table 1) and the results from Western blot analyses noted above were then used to calculate the numbers of GerD and SpoVAD molecules per spore, based on analyses of several independent spore preparations (Table 2). These values determined for GerD are relatively similar to the value determined previously for this protein in wild-type B. subtilis spores (9). However, the SpoVAD value is ∼3-fold lower than that determined previously (8) (see Discussion).

Levels of fatty acids in various spore fractions.

A surprising result of the analyses in the current work was that ∼90% of various GR subunits were found in the I fraction from disrupted spores, with only ∼10% in the M fraction. An obvious question is thus whether the GR subunits in the I fraction are in inner membrane fragments that have not been sheared off peptidoglycan and associated proteins during spore disruption or whether these GR subunits are in a much more different state than those in the M fraction. To obtain some information that might allow a decision between these possibilities, the levels of various fatty acids in decoated spores and their I and M fractions were determined following hydrolysis of phospholipids and esterification of the resultant free fatty acids (Table 3); note that the spores' outer membrane was removed in the decoating procedure prior to spore disruption. This analysis generally found a very similar composition of various individual fatty acids in decoated spores and in the M and I fractions, and the composition of individual fatty acids in decoated spores was relatively similar to what was obtained previously (27). However, only about 10% of the total fatty acids in decoated spores were found in the M fraction, with ∼90% in the I fraction (Table 3). These data strongly suggest that the majority of the IM remained with the I fraction after spore fractionation and furthermore that the IM associated with the I fraction has a fatty acid content similar to that in the M fraction.

Table 3.

Fatty acids in various fractions of wild-type spores^a

Fatty acid	% of total fatty acid
	Decoated spores	M fraction	I fraction
14:0 iso	2.8	2.8	2.8
14:0	0.3	<0.1	0.4
15:0 iso	14.8	13.1	14.8
15:0 anteiso	40.1	37.4	39.4
15:0	0.1	<0.1	0.4
16:0 N alcohol	0.4	<0.1	0.4
16:0 iso	11.4	11.9	11.2
17:0 iso	12.5	12.9	12.3
17:0 anteiso	13.4	14.5	13.0
18:0 iso	0.3	<0.1	0.3
18:0	0.4	2.8	1.0
Total/sporeb	100	10	90

^aPercentages of various fatty acids and the levels of total fatty acid in decoated spores and the M and I fractions from the same amount of decoated wild-type (PS533) spores were determined as described in Materials and Methods.

^bThese values are expressed relative to the amount in the decoated spores, which was set at 100.

Levels of GR subunits determined in lysates from disrupted wild-type spores.

Since the vast majority of the IM remained in the I fraction after spore fractionation, as suggested by fatty acid analysis, the total spore lysate from decoated wild-type spores after lysozyme disruption and sonication was also used to analyze the numbers of germination proteins per spore. For these analyses, SDS and 2-mercaptoethanol were added to disrupted spores after lysozyme treatment and sonication to recover all IM proteins prior to Western blot analyses. Analysis of GerAA in the total lysate (L) and the washed total lysate (WL) showed the expected band at ∼54 kDa, the intensity of which was compared to known amounts of purified GerAA protein (Fig. 6A). In this experiment, the GerAA band in the L fraction from 1 × 10⁸ spores had the same intensity as ∼4.5 ng of the purified GerAA protein, while the WL fraction from the same amount of spores had only minimal amounts of GerAA (2% of that in the L fraction). The GerAC band in the total lysate from 1 × 10⁸ spores also had the same intensity as ∼8 ng of purified GerAC protein (Fig. 6B). Similar analyses of GerBC, GerKA, GerD, and SpoVAD were also used to determine these proteins' levels in spore lysates (data not shown). Again, there were minimal if any of the latter proteins in the WL fraction; indeed, the WL fraction contained only minimal amounts of protein, as determined by Coomassie blue staining (data not shown). Overall, the values for numbers of germination proteins per wild-type spore determined from analyses of total spore lysates generally agreed reasonably well with those calculated from analyses of these proteins in the M, S, I, and W fractions (Table 2). The values for each protein from analyses of both total and fractionated spore lysates were averaged to give the final values of GR subunits, GerD, and SpoVAD per wild-type spore (Table 2).

Fig 6.

Western blot analysis of GerAA and GerAC levels in the total lysate of wild-type spores. The lysate (L) and washed lysate (WL) fractions from wild-type (PS533) spores were isolated, and an aliquot from 1 × 10⁸ spores was run on SDS-PAGE gels with known amounts of purified GerAA (A) or GerAC (B) antigens and subjected to Western blot analysis with antisera against either GerAA (A) or GerAC (B), as described in Materials and Methods. The intensity of the GerAA band in the spores' L and WL fractions at ∼54 kDa was compared to the intensities of known amounts of purified GerAA antigen of ∼27 kDa, and the intensity of the spores' GerAC band in the L and WL fractions from 1 × 10⁸ spores was compared to the intensities of known amounts of GerAC antigen (∼41 kDa). Note that the GerAC protein from spores is diacylglycerylated, while the purified GerAC antigen is not. MW, molecular weight markers (in thousands). No sample was loaded onto the lane designated sp in panel B.

DISCUSSION

The average number of molecules of GerBA in wild-type spores determined previously was 24 to 40 molecules per spore (5), while values found for GR subunits in the current work were each many hundreds of molecules per wild-type spore. This large discrepancy is almost certainly because in the previous work, only GerBA in the M fraction was measured, while in the current work, ∼90% of the GR subunits were actually found in the I fraction from disrupted spores along with ∼90% of the IM, as indicated by the presence of ∼90% of total decoated spore fatty acids in the I fraction. Presumably, this is because the IM was not well sheared off the integument, even though the great majority of the spores were well disrupted by the lysozyme disruption method used. It was also notable that a similar distribution of GR proteins between the M and I fractions was found when decoated spores were disrupted by mechanical breakage alone or when vortexing was substituted for the sonication of lysozyme-disrupted spores (data not shown).

That the large percentage of fatty acids found in the I fraction are indeed from the spores' IM is suggested not only by the presence of similar percentages of GR proteins in the I fraction but also by (i) the absence of the outer membrane in the I fraction, since this membrane and its associated proteins are removed by decoating (23), and (ii) the very similar fatty acid compositions in the I and M fractions. GR proteins as well as GerD and SpoVAD are also absent from the outer membrane, since they are not removed from spores by decoating treatments (4, 5, 8, 26). In addition, at least GerBA remains in the spores' IM after germination and on into outgrowth (5). GR A and B subunits are predicted to have multiple transmembrane domains (13, 15), and GR C subunits have an N-terminal cysteine with a diacylglycerol moiety (14); these structural features further suggest that GR proteins are present in the IM. In addition, the localization by fluorescence microscopy of fluorescent fusions of a number of GR proteins as well as GerD and a SpoVA protein in the IM region just outside the spore core (12) is further evidence for the location of these proteins in the IM.

In contrast to the much higher levels of GR proteins found in the current work, the value for GerD in spores was relatively similar to that determined previously, as this value was also determined by using whole spore extracts (9). However, the value determined for the number of SpoVAD molecules per spore in the current work was significantly lower than that determined previously (∼15,000 molecules per spore (8). While we have no complete explanation for the latter discrepancy, we have observed that samples of purified SpoVAD stored frozen can often begin to precipitate, especially if frozen and thawed. If this had taken place during the course of the initial work, it could have significantly altered the amounts of purified SpoVAD run on Western blots. In addition, at times, there was a slightly lower-molecular-mass band just below the 36-kDa SpoVAD band in various spore fractions (Fig. 5A and B). With longer exposure times or high levels of SpoVAD, it was difficult to distinguish between the two bands because they merge, making the SpoVAD-specific band appear more intense. At present, the identity of this lower-molecular-mass band is not clear, although its intensity was clearly lower than that of the 36-kDa SpoVAD band.

The distribution of GerD and SpoVAD was different from that of the GRs, which was perhaps not unexpected, since (i) GerD is relatively hydrophilic and may be held in the IM primarily by its lipid anchor, (ii) significant amounts of GerD were seen in the S fraction of spores previously (26), and (iii) SpoVAD has no obvious hydrophobic segments suitable for membrane anchoring (28). Consequently, SpoVAD and GerD may not be as tightly bound to the IM as are GR subunits and thus may be easily lost from the IM when spores are disrupted and fractionated. Indeed, epitope-tagged GerD is released from either the I or M fraction with 0.5 M NaCl (9). However, epitope-tagged GerD and native SpoVAD were detected in the IM of dormant spores, when spores were broken by mechanical means in either the dry (GerD) or wet (SpoVAD) state (8, 9, 26). Unfortunately, dry breakage of spores can do much damage to proteins, including GR proteins, such that their immunological detection by polyclonal sera is difficult (K.-A. V. Stewart and P. Setlow, unpublished data). GerD is also found primarily in the S fraction in germinated spores (9), and since lysozyme disruption of spores mimics germination in some ways, it was possible that this was why so much GerD was found in the S fraction of lysozyme-disrupted spores. However, even in dormant spores mechanically disrupted in liquid, a large amount of GerD is found in the S fraction (26) (data not shown). Sonication of lysozyme-disrupted spores was also shown in the current work not to alter the distribution of GerD and SpoVAD between various spore fractions significantly.

Although the GRs' C subunits are also anchored to the IM by their lipid moiety, as is GerD, it is possible that these otherwise hydrophilic GR C subunits are strongly held in the M fraction by their association with the GRs' A and B subunits, which are integral IM proteins. Indeed, a number of amino acid substitutions in the B. subtilis GerAA and GerAB proteins results in loss of GerAC from total spore extracts (29, 30). Wild-type GerAC in the IM also appears likely to be essential for GerAA assembly in the IM, as loss of GerAC from the M fraction in spores lacking GerF, which encodes prelipoprotein diacylglycerol transferase, also results in loss of GerAA from this fraction (7). In contrast, associations between GerD and GR proteins may be weaker than the associations between various GR subunits.

If the previously reported average value of 32 GerBA molecules per spore (5) is recalculated assuming that 90 to 95% of GerBA is present in the I fraction and with only ∼5% in the M fraction (Table 1), this average value becomes 640 molecules per spore, not far from the value of ∼700 molecules per spore determined in the current work for GerBC. Together with the nearly 1:1 ratio of GerAA to GerAC proteins determined in the current work, these findings suggest that the A and C subunits of GRs are most likely present in a 1:1 ratio. The ∼1.5-fold-higher levels of GerA GR subunits than GerB subunits or GerKA in spores are also consistent with the higher levels of β-galactosidase accumulated in spores from a transcriptional gerA-lacZ fusion than from a comparable gerB-lacZ fusion and also with the ∼2-fold-higher levels of gerA mRNA than of gerB and gerK mRNAs (31–33). The finding that GerA is the most abundant GR is also consistent with the major effect that loss of this GR alone has on the germination of B. subtilis spores by a pressure of ∼150 MPa, which triggers spore germination solely via direct activation of the spores' GRs (34). In addition, GerA being the most abundant of the B. subtilis spores' GRs may at least in part explain the much greater inhibitory effects of GerA overexpression (∼8-fold) on germination via either GerB plus GerK or a modified GerB termed GerB* than the effects of GerB or GerB* overexpression (4- to 8-fold) or GerK overexpression (∼2-fold) on germination via GerA (7).

With these numbers of germination proteins per spore now determined, we can begin to think not only about the assembly of subunits in a GR and how this relates to that GR's function but also about how these molecules assemble with GerD into the germinosome. The assembly of this germination protein complex is dependent on GerD, although GerD's exact role in this assembly and in GR-dependent germination is not known (12). From the spore's average number of molecules of GerD and total GRs, the GerD/GR ratio is ∼1.4:1, and possibly, this ratio is optimal for germinosome assembly and GR function. It will certainly be of interest to examine the effects of changes in the GerD/GR ratio on GR function and germinosome assembly by overexpressing or decreasing GerD expression. Knowledge of the total numbers of molecules of GerD plus GRs in spores also allows calculation of the minimal size of the germinosome. If the germinosome is assumed to be a homogeneous sphere containing 2,500 GR and 3,500 GerD molecules, this sphere is predicted to have a diameter of ∼100 nm.

It was previously observed that spores can exhibit heterogeneity even within an isogenic population, with the most striking example of this heterogeneity being the small percentage of spores in populations that have very low GR levels and are superdormant for nutrient germination (6). All the reasons for the large heterogeneity in numbers of GR molecules in individual spores in populations are not known. However, there are multiple factors that affect GR numbers per spore, including the richness of the sporulation medium, as well as levels of regulatory proteins, such as SpoVT and DklA (18, 19, 35), that modulate the transcription of the operons that encode GRs. Certainly, variations in any or all of these factors as well as stochastic variations could lead to significant variability in numbers of GR molecules per spore. One challenge in achieving a thorough understanding of spore germination, in particular the heterogeneity in spore germination kinetics, is to develop a mathematical model for spore germination kinetics. Knowledge of numbers of molecules of various GRs per spore as well as how GR numbers vary in spore populations will be important for generating such mathematic models, and this work is in progress.