Effects of a gerF (lgt) Mutation on the Germination of Spores of Bacillus subtilis

Abstract

One of the proteins of the membrane-bound receptors that recognize individual nutrients that trigger germination of spores of Bacillus subtilis contains the recognition sequence for diacylglycerol addition to a cysteine residue near the protein's N terminus. B. subtilis spores lacking the gerF (lgt) gene that codes for prelipoprotein diacylglycerol transferase exhibited significantly slowed germination in response to nutrient germinants as found previously, but germination of gerF spores with a mixture of Ca²⁺ and dipicolinic acid or with dodecylamine was normal, as was the spontaneous germination of gerF spores lacking all nutrient germinant receptors. The deleterious effects of the gerF mutation on nutrient germination were highest on germination triggered by the GerA nutrient receptor and were less so (but still significant) on germination triggered by the GerB nutrient receptor. However, there was little, if any, effect on GerK nutrient receptor-mediated spore germination. As predicted from the latter results, replacement by alanine of the cysteine residue to which diacylglycerol is thought to be added to these nutrient receptors had a large effect on GerA receptor function, less effect on GerB receptor function, and little, if any, effect on GerK receptor function.

Spores of Bacillus species are metabolically dormant and can remain so for long periods. However, these dormant spores can rapidly return to vegetative growth via the processes of spore germination and then outgrowth (14, 15, 23). Spores normally initiate germination in response to nutrient germinants, such as amino acids and sugars. These nutrient germinants are sensed by receptors in the spore's inner membrane that interact with nutrients in a stereospecific manner, and this receptor-nutrient interaction triggers subsequent events in spore germination (5, 14, 15, 16, 22, 23). There are three functional nutrient receptors in Bacillus subtilis spores, encoded by the homologous gerA, gerB, and gerK operons (termed gerA operon homologs), and gerA gerB gerK spores do not germinate in response to nutrients (14, 15, 16, 20, 21). The gerA operon homologs are tricistronic, and the three proteins (termed A, B, and C) encoded by each operon probably interact to form the functional nutrient receptor. All three proteins are essential for GerA nutrient receptor function, at least GerBA and GerBC are essential for GerB nutrient receptor function, and at least GerKC is required for GerK nutrient receptor function (2, 7, 8, 10, 14, 15, 16, 20, 23, 32). The nutrient receptors encoded by the different gerA operon homologs interact with different nutrients and with the GerA nutrient receptor recognizing l-alanine, while the GerB and GerK nutrient receptors are both required for germination with a mixture of l-asparagine, d-glucose, d-fructose, and K⁺ (AGFK) (8, 14, 15, 16, 20). The mechanism whereby nutrients cause activation of these receptors and subsequent events in spore germination is not known. Spore germination in response to nutrients is also stimulated in some fashion by the GerD protein (14, 31).

As noted above, the nutrient receptors are present in the spore's inner membrane. The A and B members of each nutrient receptor complex are quite hydrophobic proteins, with a number of putative membrane-spanning regions (10, 14, 15, 16). Thus, it is likely that these are integral membrane proteins. In contrast, the C proteins of the nutrient receptors are rather hydrophilic but have an N-terminal signal sequence followed by a consensus sequence for diacylglycerol addition at a specific cysteine residue. Lipid addition at this cysteine residue presumably allows signal peptidase II to remove the signal peptide through cleavage just prior to the modified cysteine residue (2, 10, 12, 14, 15, 16, 29). The gerD gene product also contains a signal sequence followed by a recognition sequence for diacylglycerol addition (10, 29, 31). However, the location of GerD in spores is not known, although there is a suggestion that this protein is in the spore's integument (cortex, outer membrane, and coat) fraction (16). It has been suggested that diacylglycerol addition is important in nutrient receptor and GerD localization and/or function (2, 14, 15, 16). In support of this latter suggestion, the gerF gene, mutation in which markedly reduces nutrient-induced spore germination (13), is identical to the lgt gene that encodes prelipoprotein diacylglycerol transferase, the enzyme that catalyzes the transfer of diacylglycerol to a cysteine residue in bacterial membrane prelipoproteins (24). Despite the identification of gerF as lgt, we will continue to refer to this gene as gerF in this work. The gerF gene encodes the only known enzyme of this type in B. subtilis, and gerF vegetative cells lack all detectable membrane lipoproteins (12, 24). Surprisingly, gerF null mutants are viable and grow relatively normally, although they are defective in protein secretion (12, 24). Even in the absence of diacylglycerol addition, those prelipoproteins that have been examined are still membrane localized, possibly through their signal peptide that is not removed efficiently in the absence of lipid addition (1, 9, 12, 29). While there are notable defects in the function of some membrane lipoproteins in gerF cells (1, 3, 12), other lipoproteins function normally (9, 12). Indeed, as noted above, gerF cells grow relatively normally, and their sporulation is also relatively normal (12, 24). Cells lacking signal peptidase II, in which lipid addition is normal but the signal peptide is not removed, also grow normally, and their sporulation and spore germination are also normal (29). Clearly membrane lipoproteins, presumably including GerD and the spore's nutrient receptors, can tolerate large variations in their precise N-terminal structure.

In addition to nutrient-induced germination, B. subtilis spores exhibit three other germination responses. One is a low level of spontaneous germination seen with spores lacking all nutrient receptors (21). The level of spontaneous spore germination varies from 0.05 to 0.3%/day depending on the spore preparation, although the reason for this variation is not known. The second type of germination response is the rapid germination of spores by a 1:1 chelate of Ca²⁺ and pyridine-2,6-dicarboxylic acid (dipicolinic acid [DPA]) (19, 21). This chelate is present at extremely high levels in the spore's central region or core, and triggering of spore germination by Ca²⁺-DPA does not involve the nutrient receptors but rather the activation of enzymatic hydrolysis of spore cortex peptidoglycan (19, 21). The third type of germination response is that to cationic surfactants, with dodecylamine being one member of this class of compound (25, 26, 27). Recent work has shown that B. subtilis spores lacking nutrient receptors germinate normally with dodecylamine, and this compound appears to trigger spore germination by somehow causing the release of spore DPA (27).

Given our knowledge of spontaneous and Ca²⁺-DPA- and dodecylamine-induced spore germination, it appeared likely that a gerF mutation would not affect these three processes. Similarly, because lack of diacylglycerol addition can have no or very little effect on the activities of some membrane lipoproteins, as noted above, it seemed possible that the activity of only some of the spore's nutrient receptors would be affected by a gerF mutation. Finally, if a gerF mutation does have effects on the function of specific nutrient receptors, it should be possible to mimic these effects by changing the appropriate cysteine codon in the C cistrons of the appropriate gerA operon homolog to a codon for alanine, a residue to which diacylglycerol cannot be added. This work reports the analysis of these various possibilities.

MATERIALS AND METHODS

B. subtilis strains used and their construction

The B. subtilis strains used in this work are derived from strain 168 and are listed in Table 1. The mutations in the ger genes are deletion mutations in which the majority of the coding sequence has been deleted and replaced by an antibiotic resistance marker (21, 24). All plasmid cloning was in Escherichia coli.

TABLE 1.

B. subtilis strains used

Strain	Genotype	Source or referencea
L16220	ΔgerF::ermC	Anne Moir (24)
PS832	Wild-type	Laboratory stock
FB20	ΔgerA::spc	21
FB60	ΔgerB::cat	21
FB61	ΔgerA::spc ΔgerB::cat	21
FB62	ΔgerD::spc	pFE139→PS832
FB68	ΔgerK::erm	21
FB87	ΔgerB::cat ΔgerK::erm	21
FB104	ΔgerA::spc ΔgerB::cat ΔgerK::tet	pFE225→FB61
FB118	ΔgerB::cat ΔgerK::tet	pFE225→FB60
PS3301	ΔgerF::ermC	L16220→PS832
PS3407	ΔgerD::spc ΔgerF::ermC	PS3301→FB62
PS3408	ΔgerA::spc ΔgerB::cat ΔgerF::ermC	PS3301→FB61
PS3409	ΔgerB::cat ΔgerF::ermC ΔgerK::tet	PS3301→FB118
PS3418	ΔgerK::tet	FB104→PS832
PS3419	ΔgerF::ermC ΔgerK::tet	PS3301→PS3418
PS3420	ΔgerA::spc ΔgerK::tet	FB104→FB20
PS3421	ΔgerA::spc ΔgerF::ermC ΔgerK::tet	PS3301→PS3420
PS3434	ΔgerA::spc ΔgerF::ermC	PS3301→PS3424
PS3425	ΔgerB::cat ΔgerF::ermC	PS3301→FB60
PS3427	ΔgerA::spc ΔgerB::cat ΔgerF::ermC ΔgerK::tet	PS3301→FB104
PS3602	gerACC18::spc	pACC18→PS832
PS3603	gerACA18::spc	pACA18→PS832
PS3604	gerBCC20::spc	pBCC20→PS832
PS3605	gerBCA20::spc	pBCA20→PS832
PS3606	gerKCC21gerKB::spc	pKCC21→PS832
PS3607	gerKCA21gerKB::spc	pKCA21→PS832
PS3608	ΔgerA::cam	pΔgerA→PS832
PS3609	ΔgerA::neo	pCm::Nm→PS3608
PS3610	ΔgerA::neo ΔgerB::cat ΔgerK::erm	PS3609→FB87
PS3611	gerACC18::spc ΔgerB::cat ΔgerK::erm	PS3602→FB87
PS3612	gerACA18::spc ΔgerB::cat ΔgerK::erm	PS3603→FB87
PS3613	ΔgerA::neo gerBCC20::spc ΔgerK::erm	PS3604/FB68→PS3609
PS3614	ΔgerA::neo gerBCA20::spc ΔgerK::erm	PS3605/FB68→PS3609
PS3615	ΔgerA::neo ΔgerB::cat gerKCC21gerKB::spc	PS3606/FB60→PS3609
PS3616	ΔgerA::neo ΔgerB::cat gerKCA21gerKB::spc	PS3607/FB60→PS3609
PS3617	ΔgerKC gerKB::spc	pKCΔ→PS832

^aStrains constructed in this work were prepared by transformation with plasmid DNA and/or chromosomal DNA from the strain(s) to the left of the arrow into the strain to the right of the arrow.

The ΔgerA::neo B. subtilis strain (PS3609) was constructed as follows. A region of the gerA operon from base pairs (bp) +97 to +3844 relative to the gerAA translation start site was amplified by PCR (all primer sequences are available upon request) by using EXL DNA polymerase (Stratagene, Cedar Creek, Tex.). The PCR primers used introduced BamHI and SphI sites in the upstream and downstream ends, respectively, of the target sequence. The PCR product was cloned between the BamHI and SphI sites in plasmid pUC19, and the recombinant plasmid was digested with PstI, removing 2.9 kb of internal gerA operon sequence and leaving ∼0.4 kb of gerA DNA on either side of the single PstI site remaining with the plasmid backbone. After the PstI ends were filled in and dephosphorylated with calf intestinal phosphatase, the digested plasmid was ligated with a 1.2-kb chloramphenicol resistance (Cm^r) cassette that had been excised from plasmid pFE109 (20) with BamHI and PstI and made blunt ended. The resultant recombinant plasmid (pΔgerA) was introduced into B. subtilis PS832, and one Cm^r transformant, termed PS3608, was shown by PCR to have arisen by a double crossover event involving the two gerA operon regions flanking the cam gene. The cam gene in strain PS3608 was then exchanged for a neo gene by transformation with plasmid pCm::Nm (28), and the correct chromosomal structure of one Cm^s Km^r transformant, termed PS3609, was confirmed by PCR.

The ΔgerK::tet plasmid, pFE225, was derived from plasmid pFE143 that carries ermC flanked by regions of gerK (21). The ermC marker from plasmid pFE143 was removed by digestion with XbaI and ClaI and replaced with an XbaI-ClaI fragment containing the tet marker from plasmid pDG1919 (4).

The ΔgerD::spc plasmid, pFE139, used to disrupt the gerD locus was constructed as follows. DNA flanking the 5′ end of gerD (bp −85 to +193 relative to the gerD translation start site) was PCR amplified from genomic DNA. The PCR product had BamHI and EagI sites at its ends that had been introduced with the PCR primers, and these were used to insert the PCR product into plasmid pJL74 (11) upstream of the plasmid's spc marker; the resulting plasmid was called pFE137. DNA flanking the 3′ end (bp +376 to +588 relative to the gerD translation start site) of gerD was similarly amplified, except the PCR product was flanked by HindIII and EcoRV sites and was cloned into the TA vector pCR2.1 (Invitrogen, Carlsbad, Calif.), giving plasmid pFE138. The HindIII-EcoRV fragment from plasmid pFE138 was then introduced between the same sites in plasmid pFE137. The resulting plasmid, pFE139, contains the spc marker flanked by DNA from the 5′ and 3′ ends of gerD, and transformation of B. subtilis strains to spectinomycin resistance (Sp^r) by a double crossover event (confirmed by Southern blot analysis) resulted in ΔgerD strains.

For construction of the strains with the appropriate cysteine codon in gerAC changed to an alanine codon, the region between bp −299 and +1621 relative to the gerAC translation start site was amplified by PCR with PfuUltra DNA polymerase (Stratagene), and the blunt-ended fragment was cloned in plasmid pCR4Blunt-TOPO (Invitrogen). An SmaI site was generated in the resultant recombinant plasmid 204 bp downstream of the gerAC translation stop codon and ≥50 bp downstream of the operon's transcription terminator by the QuikChange site-directed mutagenesis kit (Stratagene). The resultant recombinant plasmid was digested with SmaI, dephosphorylated, and ligated with the 1.1-kb Sp^r cassette from plasmid pJL74 (11) that had been excised with BamH1 and PstI and made blunt ended. The resultant recombinant plasmid (pAC^C18) was introduced into B. subtilis PS832, and one Km^s Sp^r transformant that had arisen by a double crossover event with the part of the gerA operon following the spc cassette was termed PS3602. Cysteine codon 18, TGC, in gerAC in this recombinant plasmid was also changed to an alanine codon, GCA, by site-directed mutagenesis as described above, and this plasmid (pAC^A18) was again introduced into B. subtilis PS832; one Km^s Sp^r transformant was termed PS3603. PCR under restrictive conditions confirmed that both PS3602 and PS3603 had the expected codons at position 18 of gerAC. In addition, the region from bp −299 to +517 relative to the gerAC translation start site was PCR amplified from PS3603 DNA and sequenced, as was done for all strains derived from strain PS3603.

A region around gerBC from bp −398 to +1600 relative to the gerBC translation start site was PCR amplified and cloned in plasmid pCR4Blunt-TOPO (Invitrogen), an SmaI site was inserted 94 bp downstream of the gerBC translation stop site and ≥40 bp downstream of the gerB operon's transcription terminator, and a spc cassette was inserted in the SmaI site as described above. The resulting plasmid (pBC^C20) and its variant (pBC^A20) in which codon 20, TGC, was changed to GCA (coding for alanine) were introduced into B. subtilis PS832, giving the Km^s Sp^r strains PS3604 and PS3605. PCR under restrictive conditions confirmed that these strains had the expected codons in position 20. In addition, the region from bp −398 to +541 relative to the gerBC translation start site was PCR amplified from PS3605 DNA and sequenced, as was done for strains derived from strain PS3605.

A region around gerKC from bp −201 to +2700 relative to the gerKC translation start site (note that this region encompasses gerKB, which is downstream of gerKC) was PCR amplified and cloned in plasmid pCR4Blunt-TOPO (Invitrogen), an HpaI site was inserted 123 bp downstream of the gerKB translation stop site and ≥20 bp downstream of the gerK operon's transcription terminator, and an spc cassette was inserted in the HpaI site as described above. The resulting plasmid (pKC^C21) and its variant (pKC^A21) in which codon 21, TGC, was changed to GCA (coding for alanine) were introduced into B. subtilis PS832, giving the Km^s Sp^r strains PS3606 and PS3607. PCR under restrictive conditions confirmed that these strains had the expected codons in position 21 of gerKC, and the region from bp −201 to +527 relative to the gerKC translation start site was PCR amplified from PS3607 DNA and sequenced, as was done for strains derived from strain PS3607.

For construction of a strain with an in-frame deletion of gerKC, plasmid pCR4Blunt-TOPO carrying the gerK operon region from bp −201 to +2700 relative to the gerKC translation start site described above was prepared in E. coli GM1634 (dam) and digested with ClaI. This digestion removes a 1-kb fragment encompassing ∼85% of the gerKC coding sequence, and religation of the plasmid backbone and associated gerK sequence maintains the reading frame of the remaining gerKC coding sequence, ensuring the ultimate expression of the downstream gerKB cistron. The religated plasmid, termed pKCΔ, was isolated in a dam⁺ E. coli strain and was used to transform B. subtilis PS832 to Sp^r. The Sp^r transformants were further screened for Km^s, and Km^s Sp^r transformants were screened by colony-directed PCR with the primers used to confirm the identity of the codon in position 21 of gerKC described above. One clone that gave no PCR fragment in this screen was termed PS3617. PCR amplification of the mutant gerKC gene with PfuUltra DNA polymerase (Stratagene) by using primers ∼130 bp upstream and 110 bp downstream of the ClaI site in gerKC and determination of the sequence of the PCR fragment confirmed the presence of the gerKC deletion and that the reading frame of the remaining gerKC coding sequence had been maintained.

Preparation of spores.

Spores were prepared by growth at 37°C on 2× SG medium (17) agar plates without antibiotics. Plates for preparation of spores carrying the gerF mutation also routinely contained 1 mM isopropyl-β-d-thiogalactoside (IPTG) to ensure expression of genes downstream of and likely cotranscribed with gerF (24). However, we found no differences in the germination behavior of gerF spores prepared with or without IPTG (data not shown). Spores were harvested, cleaned, and stored as described previously (17, 18), and all spores used were free (>98%) of growing cells, germinated spores, or cell debris. Where the germination behavior of spores of large numbers of strains was to be compared (see Table 2 or 3), spores of the different strains were prepared at the same time with the same batch of medium.

TABLE 2.

Germination of spores of various strains with or without a gerF mutation^a

Strain (genotype)	CFU at OD600b	CFU/day without Ca2+-DPA:CFU/day with Ca2+-DPAc
PS832 (wild type)	1.4 × 108	1.0
PS3301 (gerF)	1.2 × 108	0.13
FB20 (gerA)	1.4 × 108	0.49
FB62 (gerD)	1.4 × 108	0.82
PS3407 (gerD gerF)	1.4 × 108	0.09
PS3424 (gerA gerF)	1.2 × 108	0.14
FB60 (gerB)	1.3 × 108	0.82
PS3425 (gerB gerF)	1.2 × 108	0.08
PS3418 (gerK)	1.3 × 108	1.15
PS3419 (gerF gerK)	9 × 107	0.01
FB61 (gerA gerB)	1.5 × 108	0.06
PS3408 (gerA gerB gerF)	9 × 107	0.07
PS3420 (gerA gerK)	9 × 107	0.52
PS3421 (gerA gerF gerK)	8 × 107	0.02
FB118 (gerB gerK)	1.5 × 108	0.88
PS3409 (gerB gerF gerK)	1.2 × 108	0.003
FB104 (gerA gerB gerK)	1.5 × 108	0.002d
PS3427 (gerA gerB gerF gerK)	8 × 107	0.002

^aSpores were prepared and heat activated, and the viable colonies formed in 24 h at 37°C on LB medium plates plus 1 mM IPTG and with or without prior incubation in Ca²⁺-DPA were determined as described in Materials and Methods. Values shown are the average of duplicate determinations on two independent spore preparations. Individual values differed by less than 20% from the average value.

^bValues for CFU were measured after 1 day and were determined after preincubation with Ca²⁺- DPA.

^cValues are the ratio of CFU measured after 1 day (CFU/day) given by spores plated without Ca²⁺-DPA treatment divided by the CFU/day obtained from spores preincubated with Ca²⁺-DPA.

^dPrevious work (20) has shown that with spores of this strain, this slow rate of spore germination is relatively constant over at least 8 days.

TABLE 3.

Nutrient germination of spores with a cysteine→alanine replacement at the putative diacylglycerol-modified cysteine residue in the C proteins of nutrient receptors^a

Strain (genotype)	CFU at OD600
PS834 (wild type)	1.5 × 108
PS3301 (gerF)	3 × 107
PS3602 (gerACC18)	1.4 × 108
PS3603 (gerACA18)	4 × 107
PS3604 (gerBCC20)	1.4 × 108
PS3605 (gerBCA20)	1.5 × 108
PS3606 (gerKCC21)	1.3 × 108
PS3607 (gerKCA20)	1.4 × 108
PS3610 (gerA gerB gerK)b	6 × 104
PS3611 (gerACC18gerB gerK)	4 × 107
PS3612 (gerACA18gerB gerK)	2 × 105
PS3613 (gerA gerBCC20gerK)	4 × 107
PS3614 (gerA gerBCA20gerK)	5 × 106
PS3615 (gerA gerB gerKCC21)	1.5 × 106
PS3616 (gerA gerB gerKCA21)	1.4 × 106

^aSpores were prepared and heat activated, and CFU were given in 24 h at 37°C on LB medium plates were determined as described in Materials and Methods. Values shown are the average of duplicate determinations on two independent spore preparations. Individual values differed by less than 25% from the average value.

^bThis value for the spontaneous germination of spores lacking all nutrient receptors is significantly lower than that determined for the spore preparations analyzed for Table 2. Such differences have been seen previously between different spore preparations and are the reason why all spores whose germination was to be compared were prepared together.

Measurement of spore germination and protein analysis.

Spores at an optical density at 600 nm (OD₆₀₀) of 10 to 20 in water were routinely heat shocked for 30 min at 70°C and then cooled on ice prior to spore germination with nutrients, but this heat shock was not necessary for spore germination with dodecylamine. None of the ger mutations used has any effect on spore heat resistance (13, 19, 21, 22, 24, 31, and data not shown). Spore germination with nutrients in liquid was carried out at 37°C with spores at an initial OD₆₀₀ of 1 in either Luria-Bertani (LB) medium (21) with 4 mM l-alanine or in 10 mM Tris-HCl (pH 8.6) with 10 mM l-alanine or at an initial OD_{600 nm} of 1.5 in 50 mM KPO₄ (pH 7.4)-2.5 mM l-asparagine-500 μg of d-glucose/ml-500 μg of d-fructose/ml (AGFK). Germination in liquid media was monitored by either (i) measuring the OD₆₀₀ of the culture, because this value falls by ∼50% when spores germinate fully (LB medium plus l-alanine); (ii) measuring DPA release, one of the earliest events in germination (23), by centrifugation of 1 ml of culture and measuring the OD₂₇₀ of the supernatant fluid (see below) (AGFK); or (iii) examination of the culture in the phase-contrast microscope (LB medium plus l-alanine). Spores were used at an initial OD₆₀₀ of 1.5 when DPA release was used to monitor spore germination so that the OD₂₇₀ in the supernatant fluid after full DPA release would be ∼0.35. In experiments examining spore germination with nutrients over many hours, spores at an OD₆₀₀ of 1 in water were diluted in water, aliquots were spread on LB medium plates, the plates were incubated at 37°C for appropriate times, and colonies were counted. This assay measures not only spore germination but also spore outgrowth and cell growth, but the mutations examined in this work have no significant effects on cell growth (12, 24) or spore outgrowth (see below). Thus, this assay is, in effect, a measure of spore germination.

For analysis of spore germination with a mixture of Ca²⁺ and DPA, spores at an OD₆₀₀ of 1 were incubated for 1 h at 24°C in 60 mM CaCl₂-60 mM DPA-20 mM Tris-HCl (pH 8.6), aliquots were diluted and spotted on LB agar plates, and colonies were counted after incubation for 24 h at 37°C as described above (21). In this assay spores can germinate either with Ca²⁺-DPA during the 1-h incubation or with nutrients on the LB plates. Consequently, this assay is most informative when spores with poorly functional or nonfunctional nutrient receptors are tested.

For analysis of spore germination with dodecylamine, spores at an OD₆₀₀ of 1.5 were incubated at 45°C in 1 mM dodecylamine-10 mM Tris-HCl (pH 7.5). At various times 1-ml aliquots of the culture were centrifuged in a microcentrifuge, and the OD₂₇₀ of the supernatant fluid was measured, as was the total soluble material in spores absorbing at 270 nm (27). The release of material absorbing at OD₂₇₀ in this incubation is a reflection of the release of DPA from spores (27).

The inner membrane fraction was isolated from spores of strains PS832 (wild type) and PS3301 (gerF), samples of membrane protein were run on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, proteins were transferred to a polyvinylidene difluoride membrane (Immobilon), and GerBA was detected by Western blot analysis with a polyclonal antiserum as described previously (22).

RESULTS

Germination of gerF spores with nutrients

Analysis of the germination of spores of B. subtilis in liquid medium showed that wild-type spores germinated rapidly in a rich medium supplemented with l-alanine (Fig. 1). In contrast, gerF spores exhibited only minimal germination in this medium (Fig. 1). However, examination of the culture by phase-contrast microscopy showed that a small percentage (∼2%) of the gerF spores did germinate during the 2-h period of this experiment and further showed that the outgrowth of these germinated spores appeared normal (data not shown). Presumably, the outgrowth of these few spores is the reason for the slight increase (albeit significant, as seen in multiple experiments) in the OD₆₀₀ of the gerF culture (Fig. 1). These observations suggested that the gerF spores were not absolutely defective in spore germination but only initiated germination slowly in response to nutrients. That at least some of the gerF spores could initiate germination was shown further by incubating gerF and wild-type spores at 37°C on LB medium agar plates (21) and counting the colonies produced after 24 h. The gerF spores gave 2.1 × 10⁷ CFU at OD₆₀₀, while wild-type spores gave 1.4 × 10⁸ CFU at OD₆₀₀.

FIG. 1.

Germination of wild-type and gerF spores in liquid medium. Heat-activated spores of strains PS832 (wild-type) or PS3301 (gerF) were incubated at 37°C in LB medium plus 4 mM l-alanine, and the optical densities were measured at various times. Symbols: ▴, wild-type spores; •, gerF spores.

To demonstrate that more than 15% of the gerF spores could germinate, albeit slowly, heat-activated wild-type and gerF spores were spread on 150-mm-diameter LB plates (20), the plates were incubated at 37°C, and the colonies appearing every 24 h over 4 days were enumerated (Fig. 2). This experiment should assess only spore germination, because gerF spores outgrow normally and gerF cells also grow normally. With wild-type spores, >97% of the colonies that appeared over 4 days arose after only 24 h (Fig. 2). In contrast, gerF spores continued to generate significant numbers of colonies over all 4 days of the experiment, since at the end of 4 days ∼50% of the spores spread on the plates had germinated and formed colonies but only ∼15% had germinated and formed colonies after 1 day (Fig. 2). Presumably all the gerF spores would have germinated if the plates had been incubated further, but the overgrowth of the colonies from spores that had germinated earlier made this analysis impractical.

FIG. 2.

Germination of wild-type, gerF, and gerF gerK spores on plates. Heat-activated spores of strains PS832 (wild type), PS3301 (gerF), and PS3419 (gerF gerK) were spread on 150-mm-diameter LB plates as described in Materials and Methods with no additions (wild type) or with 1 mM IPTG-1 μg of erythromycin/ml-25 μg of lincomycin/ml (gerF and gerF gerK). The plates were then incubated at 37°C, and colonies were formed every 24 h and counted for 4 days. Approximately 100 wild-type spores were spread per plate, 250 gerF spores were spread per plate, and 5,000 gerF gerK spores were spread per plate. The values given are the averages of the percentage of colonies formed in each 24-h period on each of two plates, with the total colonies formed in 4 days set at 100%. After 4 days >98% of the wild-type spores had given rise to colonies, 50% of the gerF spores had given rise to colonies, and 0.8% of the gerF gerK spores had given rise to colonies. The open bars give the results with wild-type spores, the solid bars give the results with gerF spores, and the stippled bars give the results with gerF gerK spores.

Germination of gerF spores with Ca²⁺-DPA or dodecylamine.

Although the experiments described above indicated that gerF spores do germinate in response to nutrients, these spores are clearly defective in nutrient-induced germination. As noted above, this was not unexpected, because one of the proteins making up each of the spore's three nutrient receptors is likely subject to diacylglycerol addition, a modification absent in a gerF mutant. In contrast to nutrient-induced spore germination, Ca²⁺-DPA-induced germination does not involve the nutrient receptors (19, 21) and, thus, might be unaffected by the gerF mutation. To test this prediction, spores were first preincubated for 60 min at room temperature with Ca²⁺-DPA, and then aliquots were spotted on LB medium plates (Table 2). After 24 h of incubation at 37°C, both wild-type and gerF spores gave essentially the same CFU at OD₆₀₀ on these plates (Table 2), indicating that the gerF mutation does not notably interfere with Ca²⁺-DPA-induced spore germination.

Dodecylamine germination of B. subtilis spores also does not require the spore's nutrient receptors, as dodecylamine appears to trigger germination by stimulating spores to release DPA, with the resultant DPA-less spores then continuing through subsequent steps in germination (27). Although the mechanism whereby dodecylamine stimulates spore DPA release is not understood, this could conceivably involve the action of some membrane lipoprotein. However, the rates and extents of spore DPA release triggered by dodecylamine were essentially identical with those of both wild-type and gerF spores (Fig. 3 and data not shown). As expected based on previous results (26), the DPA release triggered by dodecylamine was also essentially identical to that of spores of strain PS3610 that lacks all functional nutrient germinant receptors (Fig. 3). DPA release was used in these experiments to measure spore germination as opposed to the OD₆₀₀, because dodecylamine can cause clumping of spores and this can significantly affect the OD₆₀₀ even in the absence of spore germination.

FIG. 3.

Germination of wild-type, gerF, and gerA gerB gerK spores with dodecylamine. Spores of strains PS832 (wild type), PS3301 (gerF), and PS3427 (gerA gerB gerK) were germinated with dodecylamine, and DPA release was measured as described in Materials and Methods. Symbols: ▴, PS832 (wild-type); ▪, PS3301 (gerF); and •, PS3610 (gerA gerB gerK).

Germination of gerF spores lacking GerD or nutrient receptors.

While the gerF mutation had no effect on spore germination induced by Ca²⁺-DPA or dodecylamine, this mutation caused a defect in nutrient-induced spore germination. Membrane lipoprotein function in gerF cells varies from normal to very subnormal (1, 3, 9, 12). Consequently, nutrient receptor and GerD function might or might not be affected by the lack of diacylglycerol addition. To further examine this topic as well as the effect of the gerF mutation on the slow spontaneous germination seen with spores lacking all nutrient receptors (21), the germination of wild-type and gerF spores lacking GerD or nutrient receptors was assessed with or without pretreatment of spores with Ca²⁺-DPA (Table 2). Differences between spores of the different strains used in this experiment should be due only to differences in spore germination (see Materials and Methods). Strikingly, a gerF mutation had no effect on the low level (∼0.2%/day with these spore preparations) of spontaneous germination of spores of a gerA gerB gerK mutant (Table 2). In addition, neither a gerA nor a gerB mutation significantly (≤2-fold) exacerbated the effect of a gerF mutation on spore germination (Table 2), suggesting that the GerA and GerB nutrient receptors require diacylglycerol addition for function. This was shown further as the gerF mutation reduced the germination of gerB gerK spores to that of gerA gerB gerK spores, while the germination of gerA gerK spores was reduced ∼25-fold by the gerF mutation, although it remained significantly above that of gerA gerB gerK spores (Table 2). In contrast to the large effects of a gerF mutation on GerA and GerB receptor function, there was little to no effect on GerK receptor function, as gerA gerB and gerA gerB gerF spores exhibited essentially identical levels of germination (Table 2). Underscoring the large effects of the gerF mutation on GerA and GerB receptor function and the lack of effect on GerK receptor function, loss of gerK alone had little effect on spore germination, but lack of the GerK receptor greatly increased the effect of the gerF mutation, as germination of gerF gerK spores fell to only ∼5-fold above that of gerA gerB gerK spores. However, while gerF gerK spores exhibited only a small amount of germination in 24 h, this low rate of germination was constant over at least 4 days (Fig. 2), similar to what had been found with gerF spores or previously seen with gerA gerB gerK spores over 8 days (21). Surprisingly, gerD spores exhibited no significant defect in germination as assessed by the colonies formed in 24 h on LB medium plates, and gerD gerF spores gave results similar to those with gerF spores (Table 2). Thus, the gerD mutation appears to have little effect on spore germination in a rich medium.

Germination of spores with nutrient receptors lacking the cysteine residue essential for diacylglycerol addition.

The results presented in the previous section strongly indicated that diacylglycerol is essential for the function of the GerA nutrient receptor, less essential for GerB receptor function, and not essential for GerK receptor function, at least under the germination conditions tested. If the GerK receptor is indeed in the spore's inner membrane, as seems likely based on the analogy with the location of the homologous GerA and GerB receptors (5, 22) and the hydrophobicity of GerKA and GerKB, then the lack of effect of the loss of diacylglycerol addition on GerKC function suggests that GerK lacking diacylglycerol remains associated with the spore's inner membrane, and further that the unmodified GerKC is functional. No antibodies are available against the proteins encoded by the gerK operon, so we could not assess the localization of modified and unmodified GerKC. However, we did use Western blot analysis to analyze the inner membranes of gerF spores (strain PS3301) for GerBA (see Materials and Methods). This analysis showed that wild-type and gerF spores had essentially identical levels of GerBA in their inner membranes (data not shown). Because the A, B, and C proteins of the nutrient receptors are thought to function in a complex (14, 15, 16, 20, 22), this latter finding is consistent with GerBC also being present in the inner membrane of gerF spores, although by no means does it prove this point.

To further assess to role of diacylglycerol in the function of the spore's nutrient receptors, we prepared spores carrying only a single nutrient receptor and with the cysteine in the N-terminal regions of the C proteins of these receptors in sequences with a good match to the consensus sequence for recognition and lipid addition by GerF (residues 18, 20, and 21 in GerAC, GerBC, and GerKC, respectively) (2, 14, 15, 16, and see below) changed to alanine. The germination of spores with these altered nutrient receptors was then compared to the germination of otherwise isogenic spores that retained the cysteine residues. Strikingly, this cysteine-to-alanine change almost completely eliminated GerA receptor function, as germination of spores in which the only nutrient receptor was GerA and with GerAC^A18 (strain PS3612) was only marginally higher than that of spores lacking all nutrient receptors (strain PS3421) (Table 3). In addition, spores containing GerAC^A18 as well as functional GerB and GerK (strain PS3603) germinated at ≤2% the rate of isogenic spores containing GerAC^C18 (strain PS3602) in l-alanine in liquid, as determined by observation with the phase-contrast microscope (data not shown). In contrast, replacement of the crucial cysteine in GerBC by alanine only decreased germination on LB medium plates of spores whose only nutrient receptor was GerB (strain PS3614) ∼10-fold, while the analogous alanine-for-cysteine change in GerKC had no effect on germination on LB medium plates of spores whose only nutrient receptor was GerK (strain PS3616) (Table 3). These latter effects seen on LB medium plates were also seen in liquid AGFK medium, when rates of spore germination were determined by measuring rates of DPA release, as otherwise wild-type spores with gerBC^A20 (strain PS3605) germinated only 15% as rapidly as spores with gerBC^C20 (strain PS3604) while spores with gerKC^A21 (strain PS3607) germinated almost as rapidly (within 25%) as spores with gerKC^C21 (strain PS3606) (data not shown). However, deletion of gerKC (strain PS3617) reduced the rate of spore germination in AGFK, as determined by measuring rates of DPA release, to ≤2% of that of wild-type spores (data not shown). That loss of GerKC abolishes the function of the GerK receptor has also been shown previously (7).

DISCUSSION

The results in this communication allow a number of conclusions. First, neither spontaneous spore germination, Ca²⁺-DPA-induced spore germination, nor dodecylamine-induced spore germination requires the spore's nutrient receptors, as has been suggested previously (19, 21, 27). Second, spontaneous, Ca²⁺-DPA- and dodecylamine-induced spore germination either do not require proteins subject to diacylglycerol modification or this modification is not required for the function of these proteins. This conclusion is not definitive, because many membrane proteins normally subject to diacylglycerol modification retain function in its absence (1, 9, 12). Third, because a gerF mutation or replacement of the cysteine residues in the GerA and GerB receptors predicted to be subject to diacylglycerol addition either eliminated (GerA) or reduced (GerB) the function of these receptors, GerAC and GerBC are subject to diacylglycerol addition. Fourth, while gerF spores are defective in nutrient-induced spore germination, this process is not absent, it is only slowed. It is GerA and, to a lesser extent, GerB receptor function that is defective in gerF spores, while GerK receptor function appears to be largely responsible for the nutrient-induced germination of gerF spores. Unfortunately, the reason why having the GerK receptor alone leads to slowed nutrient-induced spore germination is not clear; the GerK receptor responds in some fashion to glucose (8), but other nutrients to which GerK responds are not known. It is also not clear if lower levels of functional GerK receptor will reduce the germination of spores containing only GerK, and thus we cannot be sure that GerKC without diacylglycerol is fully functional.

Why the GerK receptor, in particular the GerKC component, would be tolerant of the lack of diacylglycerol addition is not clear. One possibility is that the GerK receptor does not require GerKC for function. However, as with gerAC and gerBC mutations that abolish GerA and GerB function in short-term spore germination, mutation of gerKC also abolishes GerK function as shown previously and in the present work (1, 7, 16, 31). Deletion of gerAC also had the identical effect on long-term spore germination on LB medium plates as did deletion of the gerA operon, including the effects of combination with a gerF and other ger mutations (data not shown). This finding, as well as the homology between the C proteins of GerA homologs, indicates that GerKC is required for GerK receptor function. A second possibility is that GerAC and GerBC may localize poorly, if at all, to the inner spore membrane in the absence of lipid addition, and thus the complete receptor will not be assembled and may be degraded. Perhaps GerKC is more relaxed in its requirement for lipid addition for localization. However, this seems unlikely, because B. subtilis prelipoproteins not modified by diacylglycerol addition are membrane localized (1, 9, 12) and because GerBA is localized normally in gerF spores. A third possibility, and the one that seems likely, is that lipid addition and signal peptide cleavage are essential, or nearly so, for the function of the GerA and GerB receptors but not for GerK receptor function. While it is possible that GerKC is not subject to diacylglycerol modification, GerKC does have a good match to the recognition sequence for diacylglycerol addition near the protein's N terminus (25 and see below). However, some membrane lipoproteins retain normal function even in the absence of diacylglycerol addition and signal peptide cleavage (9, 12), and perhaps GerKC retains function in the absence of lipid addition while GerAC and GerBC do not. One possible reason for the difference in response of GerKC, GerAC, and GerBC to a gerF mutation comes from comparison of the N-terminal amino acid sequences of these three proteins (Fig. 4). Of the three proteins, GerKC has the longest hydrophobic region prior to the cysteine residue to which diacylglycerol is added (Fig. 4, underlined residue). Perhaps it is this longer hydrophobic N-terminal region, one not likely to be removed by signal peptidase II in a gerF mutant (1, 9, 12, 29), that allows the proper functioning of GerKC and thus the GerK receptor even when GerKC has not undergone lipid addition. In contrast, GerAC has the shortest hydrophobic N-terminal region of the C proteins of nutrient receptors, which may be why GerA receptor function is most dependent on lipid addition; GerBC, which retains some function in the absence of lipid addition, has an N-terminal region with a hydrophobicity between that of the N-terminal regions of GerAC and GerKC (Fig. 4). Obvious gerF homologs are present in other Bacillus species, including B. anthracis. Interestingly, the C proteins of the GerH and GerS germinant receptors of B. anthracis spores, both of which have roles in spore germination with amino acids and purine ribonucleosides (6, 30), have hydrophobic N-terminal regions that are as short as that of GerAC. This leads to the prediction that amino acid-purine ribonucleoside germination of B. anthracis spores will be strongly dependent on GerF function.

FIG. 4.

Comparison of the N-terminal amino acid sequences of GerAC, GerBC, and GerKC. The sequences are of the initially translated proteins (10), although it is not known if the initiating methionine residue is removed posttranslationally. The cysteine residue at which diacylglycerol is most likely added is underlined, and hydrophobic residues N terminal to this cysteine residue are in boldface type. Normally the lipid-modified protein is cleaved by signal peptidase II before the cysteine residue, but only after lipid addition.