INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The gram-positive, endospore-forming, bacterium Bacillus subtilis contains four different c-type cytochromes, which are all membrane anchored (2, 49, 52). The heme domain of these cytochromes is located on the outer surface of the cytoplasmic membrane. B. subtilis does not require cytochrome c for aerobic or anaerobic growth under laboratory conditions, and the physiological role of cytochromes of this type in the bacterium is not well understood.

The trademark of c-type cytochromes is that they contain protoheme IX covalently bound to the protein via thioether linkages (1, 29). Cytochrome c synthesis, i.e., the formation of the covalently bound heme, occurs on the outer (periplasmic) side of the cytoplasmic membrane in bacteria. In gram-negative bacteria, this biosynthetic process is assisted by several membrane-bound and periplasmic proteins (see reference 19, 27, and 44 for reviews). B. subtilis ccdA, resB, and resC are hitherto the only genes that have been shown experimentally to be required for cytochrome c synthesis in a gram-positive bacterium (21, 36). Genes encoding B. subtilis CcdA orthologues are present in members of the domains Bacteria (such as Mycobacterium tuberculosis, Helicobacter pylori, Haemophilus influenzae, Treponema pallidum, and cyanobacteria) and Archaea and also in plants (encoded by chloroplast genomes), but experimental data on the gene product in these organisms is not available. Very recently, CcdA was found in Rhodobacter capsulatus and shown to be involved in cytochrome c synthesis (8).

B. subtilis CcdA is an integral membrane protein of 228 or 235 amino acid residues (36). The exact function of this protein in the cell is not known, but it is required for a late step in the cytochrome c maturation pathway, after heme and apocytochrome have been transported across the cytoplasmic membrane (35). The amino acid sequence of CcdA is similar to that of the central part of DsbD (also named DipZ) of Escherichia coli, a protein of 489 residues which is thought to transfer reducing equivalents to disulfide isomerase(s) in the periplasm (6, 25, 31). The C-terminal part of DsbD, which has no similarity to CcdA, contains a thioredoxin-like sequence motif, -CysXaaYaaCys-. Cysteine residues in this motif and those that are invariant in DsbD and CcdA have been shown by site-specific mutagenesis to be functionally important (8, 39). It is notable that the genomes of some bacteria, for example H. influenzae and M. tuberculosis, contain genes for both CcdA and DsbD proteins whereas others, such as E. coli and B. subtilis, contain genes for only one of the two proteins (36). The ccdA gene, positioned at 164° on the B. subtilis chromosomal map, is cotranscribed with two downstream genes, yneI and yneJ (20) (originally named orf120 and orf160, respectively [36]) (Fig. 1). YneI is most probably a single-domain response regulator. The sequence of this 120-residue protein is very similar to that of CheY and Spo0F, for which the three-dimensional structures are known (22, 40). YneJ is a predicted 160-residue integral membrane protein without clear similarity to any protein sequence available in the databases. YneI and YneJ are not required for cytochrome c biogenesis, and no clear difference in phenotype compared to the wild type has been observed with yneI or yneJ insertion mutants (36).

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Map of the ccdA-yneI-yneJ region in the wild-type B. subtilis chromosome. Transcription initiation sites, as indicated from primer extension analysis, and a termination site are indicated by hooked arrows and a hairpin symbol, respectively. DNA fragments cloned in pDG1728 and analyzed for promoter activity (double arrows) and mRNA species detected previously (36) by Northern blot analysis of strains containing plasmid pCCD2 are also shown. DNAs contained in plasmids pCCD2 and pLTS29, used in complementation studies, and plasmids pLTS32, pLTS34, and pLTS35, used in the construction of strains with deletions, are indicated by bars. The dashed lines in the latter DNAs indicate an approximately 1-kb fragment (not drawn to scale) harboring the ble resistance gene (an arrow indicates the relative orientation of the gene). Restriction enzyme cleavage sites: C, ClaI; E, EcoRI; Ev, EcoRV; H, HindIII; K, KpnI; P, PstI; S, SacI.

Sporulation in B. subtilis, i.e., the conversion of the vegetative cell into a spore, is a process characterized by ordered gene expression and complex morphological changes (10, 41). After the formation of an asymmetrically positioned septum between the mother cell and the forespore, the transition from one developmental stage to the next is governed by four sigma factors, ^F and ^G (forespore specific) and ^E and ^K (mother cell specific). Gene expression in the forespore and the mother cell is coordinated by intercompartment communication, where the appearance of an active sigma factor in one compartment is dependent on the activity of an earlier sigma factor in the other compartment. The end result is a spore much more resistant to heat and chemicals than is the vegetative cell.

To better understand the function of CcdA in B. subtilis and possibly find a role for the YneI and YneJ proteins in the cell, in this study we have analyzed the transcriptional organization of the ccdA-yneI-yneJ gene cluster and the expression of the three genes during growth. We demonstrate that CcdA is an integral membrane protein whose cellular concentration increases at the transition from exponential growth to stationary phase. Strains with ccdA deleted were found to be deficient, but not completely blocked, in the synthesis of spores with normal properties. This defect in sporulation was investigated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. B. subtilis strains and plasmids used in this work are presented in Table 1. E. coli strains JM83 [ara (lac-proAB) strA thi-1 (80 lacZM15)] (50), MM294 (supE44 hsdR endA1 thi) (30), and XL1-Blue [supE44 hsdR17 recA1 endA1 gyrA46 thi relA1 lac F' proAB⁺ lacI^q lacZM15 Tn10 (Tet^r)] (4) were used for the propagation of plasmids.

Media and general growth of bacteria. E. coli strains were grown on Luria agar plates or in Luria broth medium (34). B. subtilis strains were grown on tryptose blood agar base (TBAB) plates (Difco), Difco sporulation (DS) medium (16) [0.8% (wt/vol) Bacto nutrient broth (Difco), 0.1% (wt/vol) KCl, 0.012% (wt/vol) MgSO₄ · 7H₂O, 0.5 mM NaOH, 1 mM Ca(NO₃)₂, 10 µM MnCl₂, 1 µM FeSO₄) plates, or Spizizen minimal medium (38) plates supplemented with required growth factors (10 mg/liter) and with 0.5% (wt/vol) glucose, succinate, or lactate as the carbon source. Nutrient sporulation medium with phosphate (NSMP) (11) or DS medium was used for liquid cultures, which were grown at 37°C in Erlenmeyer glass flasks (culture volume, 1/10 the volume of the flask) with indentations, on a rotary shaker at 200 rpm. For detection of -galactosidase activity on TBAB plates, 80 mg of 5-bromo-4-chloro-3-indolyl--D-galactosidase (X-Gal) per liter was included in the medium. The following antibiotics were used when appropriate: ampicillin, 75 mg/liter; chloramphenicol, 12.5 mg/liter (E. coli) or 3 to 5 mg/liter (B. subtilis); erythromycin, 5 mg/liter; neomycin, 5 mg/liter; phleomycin, 1.2 mg/liter; spectinomycin, 150 mg/liter; and tetracycline, 25 mg/liter (B. subtilis).

Construction of plasmids and deletion mutants. Plasmid pLTS29, containing the ccdA gene, was constructed by first moving the 1.35-kbp SacI-EcoRI fragment from pLTS17 to pUC18. The 1.12-kbp SalI-ClaI fragment from the resulting plasmid, pLTS28, was isolated, and the ends of the fragment were made blunt by treatment with Klenow enzyme in the presence of deoxynucleoside triphosphates. Finally, the fragment was ligated to pHP13 cleaved with SmaI.

Construction of ccdA-lacZ, yneI-lacZ, and yneJ-lacZ transcriptional fusions. The regions upstream of ccdA (bp 1367 to 1682), yneI (bp 2257 to 2484), and yneJ (bp 2694 to 2920) were amplified by PCR using pCCD2 as template (Fig. 1). The ccdA region was amplified using Taq DNA polymerase (Roche Molecular Biochemicals) and the primer pair 5' CGGAATTCCTGACTGAGCTCTATCG plus 5' CGGGATCCATGATTTGACATTCCTTCAAG. The yneI and yneJ regions were amplified using Pwo DNA polymerase (Roche Molecular Biochemicals), and the primer pairs 5'-CGGAATTCTGAAGTGGATAAGGAAGAAC plus 5'-CGGGATCCAACAATCGATTTCCACAG and 5'-CGGAATTCTTAACCTTTGATCCTAAAGC plus 5'-CGGGATCCGGAGTGTTGATACTATATAC, respectively. EcoRI and BamHI restriction sites were added via the primers (underlined). The amplified fragments were cut with EcoRI and BamHI and ligated into pBluescript SK() or KS(). The complete sequence of each cloned DNA fragment was determined to exclude errors introduced by the PCR. The EcoRI and BamHI fragments were then moved to plasmid pDG1728, resulting in plasmids pLJJ7, pLJJ8, and pLJJ9 (Table 1). The lacZ gene of pDG1728 contains the Shine-Dalgarno sequence of the B. subtilis spoVG gene to provide efficient translation of the lacZ gene in B. subtilis. The obtained plasmids and pDG1728, as a negative control, were linearized by digestion with ScaI and integrated, by means of transformation and a double-crossover recombination event, into the amyE locus of B. subtilis strains 1A1, LU60A1, LU62A1, and LU63A1. The desired spectinomycin-resistant transformants obtained were confirmed by erythromycin sensitivity and the lack of -amylase activity and, in appropriate cases, also for defective cytochrome c synthesis by N,N,N',N'-tetramethyl-p-phenylenediamine (TMPD) oxidation staining.

Antisera and immunoblot analysis. Polyclonal antisera against CcdA and YneJ were obtained by immunizing New Zealand White rabbits with synthetic peptides conjugated to keyhole limpet hemocyanin. The peptides used, YITGVSMDDVKTEK and YRKLHNELQSSNIQMN, correspond to residues 32 to 45 of CcdA and 145 to 160 of YneJ, respectively. Custom peptide synthesis and production of antisera were carried out by Neosystems.

-Galactosidase activity measurements. To determine -galactosidase activity in cells, growth medium was inoculated with cells from an exponentially growing culture. Samples were withdrawn at intervals during the growth of the batch culture, immediately frozen in liquid nitrogen, and then stored at 20°C. The samples were thawed in cold water, and the cells were collected by centrifugation for 1 min at 20,000 × g in an Eppendorf centrifuge. The cell pellets were suspended in AB buffer (60 mM K₂HPO₄, 40 mM KH₂PO₄, 100 mM NaCl). A 175-µl volume of a suspension of an appropriate cell concentration was mixed with 35 µl of 4-methylumbelliferyl--D-galactoside (MUG), 0.4 g/liter in dimethyl sulfoxide, and incubated at 23°C on a water bath. At intervals, 50-µl aliquots were withdrawn and diluted in 2.45 ml of AB buffer, and the fluorescence was immediately determined using a Shimadzu RF-5301PC spectrofluorometer with an excitation wavelength of 366 nm (3-nm slit) and an emission wavelength of 445 nm (10-nm slit). The fluorescence of solutions of 4-metylumbelliferone of known concentrations in AB buffer were used to determine the amount of MUG hydrolyzed in the samples. Rates of MUG hydrolysis were obtained from linear plots of product concentration versus time. Specific -galactosidase activities were calculated for the cell concentration estimated from the optical density at 600 nm (OD₆₀₀) of the original sample taken from the culture.

Spore assays. Cultures were grown in 25 ml of NSMP (pH 6.5) at 30°C in a 250-ml Erlenmeyer flask for 2 days. The spore titer was determined by subjecting the cultures to heat, chloroform, or lysozyme treatment. Routinely, 5-ml samples were heated at 80°C for 10 min (unless stated otherwise) or vigorously mixed with 0.6 ml of chloroform for 10 s, or a small volume was diluted 100-fold in minimal salts solution [80.4 mM K₂HPO₄, 44.1 mM KH₂PO₄, 0.8 mM MgSO₄, 15.1 mM (NH₄)₂SO₄, 3 mM sodium citrate] and incubated with or without lysozyme (0.5 g/liter) at 30°C for 30 min. Serial dilutions were plated on TBAB plates and incubated overnight at 37°C. Dilutions of nontreated samples were plated in parallel.

Electron microscopy. Cells were grown in NSMP medium (pH 6.5) at 30°C in a 250-ml Erlenmeyer flask for 3 days. Samples (10 ml) were harvested and washed in fresh medium once. The cell pellet was treated with 2.5% glutaraldehyde, fixed in OsO₄, and dehydrated in ethyl alcohol, and the sample was finally impregnated in plastic and sections were analyzed by electron microscopy.

Other methods. B. subtilis chromosomal DNA was isolated by the method of Marmur (24). Plasmid DNA was isolated by using a kit (Maxi Prep [Qiagen] or Jet Prep Plasmid Miniprep [Genomed]). Total RNA from B. subtilis was purified using the RNeasy kit (Qiagen). Primer extension analysis was done as described previously (36) using oligonucleotides complementary to the proximal part of yneI (bp 2609 to 2593, 5'-CAGCTGCTTGTTCACCG) and yneJ (bp 3028 to 3012, 5'-AACAGGCTTGTCAGAGG) (base pair numbering refers to the published sequence [36], accession no. X87845). E. coli was transformed by electroporation and B. subtilis was grown to competence essentially as previously described (18). Oxidation of the artificial electron donor TMPD by colonies on TBAB plates was assayed by the soft-agar overlay method (13). -Amylase production was assayed by growth on TBAB plates containing 1.5% (wt/vol) starch and then staining colonies with KI-I₂ solution. Membranes from B. subtilis strains were isolated from osmotically lysed cells essentially as described by Hederstedt (17) and stored at 80°C. Light absorption spectroscopy was carried out as described previously (36). The protein concentration of membrane fractions was determined using the bicinchoninic acid protein assay (Pierce Chemical Co.) with bovine serum albumin as the standard.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Promoters for the yneI and yneJ genes. By using Northern blot analysis, ccdA-yneI-yneJ, yneI-yneJ, and yneJ mRNAs have previously (36) been detected (Fig. 1). The ccdA gene is preceded by an apparent ^A-type promoter, and the initiation site for transcription is positioned about 100 bp from the translational start codon (36). Plasmid pCCD2 is a low-copy-number plasmid containing the ccdA-yneI-yneJ gene cluster and flanking DNA (Fig. 1 and Table 1). Primer extension analysis, with total RNA extracted from B. subtilis strain 3G18/pCCD2 grown in NSMP, identified the 5' end of the yneI-yneJ and yneJ mRNAs (data not shown and Fig. 2). The indicated transcription initiation sites are located in intergenic regions and are both positioned 33 bp from the translation start codon (ATG) for the respective gene. An apparent 10 sequence of a ^A-type promoter, but no corresponding 35 sequence, is present in the promoter regions of yneI and yneJ (Fig. 2).

View larger version (26K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequence of promoter regions with transcription initiation sites (+1), as determined by primer extension analysis. Apparent A-type -35 and -10 promoter sequences, translation start codons, and stop codons are indicated in bold type. There are two possible translation start codons for ccdA, as discussed previously (36). Putative ribosome binding sequences are underlined. The numbering of nucleotides is that used in reference 36.

Promoter activities during growth. DNA fragments predicted to contain the three promoter regions were amplified by PCR and cloned in front of the promoterless lacZ gene in plasmid pDG1728. The ccdA, yneI, and yneJ fragments contain 224, 202 and 215 bp, respectively, of sequence upstream of +1 and the leader sequence up to (but not including) the ribosome binding sequence (Fig. 1 and 2). Each lacZ transcriptional fusion was integrated in single copy into the amyE locus on the chromosome of B. subtilis 1A1 (parental strain), LU60A1 (ccdA), LU62A1 [(ccdA-yneI-yneJ)], and LU63A1 [(yneI-yneJ)] (Table 1; see Materials and Methods for details).

View larger version (21K): [in this window] [in a new window]   FIG. 3.   -Galactosidase activity of B. subtilis strains grown in NSMP. , OD600; , specific enzyme activity. Note the different scales for enzyme activity in the panels. The strains contain the ccdA, yneI, or yneJ promoters fused to a promoterless lacZ gene and integrated into the chromosome in one copy. See Table 1 for a full description of strains.

CcdA and YneJ protein profiles. The subcellular localization and the steady-state levels of CcdA and YneJ in B. subtilis were analyzed using immunoblot with antisera against oligopeptides corresponding to amino acid residues 32 to 45 of CcdA (assuming ATG to be the start codon [Fig. 2]) and the C-terminal 16 residues of YneJ, respectively. Extracts were prepared from cells grown in NSMP and harvested 1 h before the end of the exponential growth phase (T₁), at the end of the exponential growth phase (T₀), and 1 h into the stationary phase (T₊₁), respectively. Typical growth curves are shown in Fig. 3.

View larger version (54K): [in this window] [in a new window]   FIG. 4.   Immunoblot analysis of membrane proteins from different B. subtilis strains, using CcdA antiserum (upper panel) and YneJ antiserum (lower panel). The blots are overexposed to demonstrate the presence of small amounts of CcdA and YneJ antigen. Strain LU62A1 has the ccdA-yneI-yneJ gene cluster deleted, LU60A1 has ccdA deleted, and 1A1 is the parental strain. Plasmids pLTS29 and pCCD2 carry the ccdA gene and the ccdA-yneI-yneJ gene cluster, respectively. The designations 1, 0, and +1 indicate that membranes were isolated from cells harvested 1 h before, at the time, and 1 h after exponential growth ended, respectively. The bands in the top panel seen just above CcdA and in the lower part of the gel are due to cross-reactivity with unrelated antigens and were apparent only on overexposed films.

Phenotype of strains deficient in ccdA, yneI, and yneJ. B. subtilis strains derived from 1A1 and containing deletions in the ccdA, yneI, and yneJ gene cluster (Fig. 1) were tested for growth on TBAB and minimal plates with glucose, succinate, or lactate as the carbon source. No major differences in growth were seen between the various mutants and the parental strain after incubation at 37°C overnight. Incubation of colonies on TBAB or minimal-succinate plates at room temperature for 2 days, however, resulted in lysis of LU60A1 (ccdA) and LU62AI [(ccdA-yneI-yneJ)]. Colonies of strains 1A1 and LU63A1 [(yneI-yneJ)] did not lyse on the plates even after >1 week of incubation at room temperature.

CcdA-deficient mutants show reduced sporulation efficiency. The ability of ccdA mutant strains to form spores was analyzed by growing cultures in NSMP for 2 days at 30°C and then determining the fraction of heat- and chloroform-resistant cells in the cultures (Tables 2 and 3). For the parental strain, 1A1, and the yneI-yneJ deletion mutant, LU63A1, the sporulation efficiency was >85%. Strains LU60A1 and LU62A1, with ccdA and ccdA-yneI-yneJ, respectively, deleted, showed a sporulation efficiency of 1 to 6%. This number is much greater than that obtained with a strain completely blocked in an early step in sporulation such as the SpoIIIG-deficient strain LUA14 (Table 3). The spore yields per volume of culture were very low for the CcdA-deficient strains compared to the parental strain because of the cell lysis of the former strains that occurs in the stationary phase. Incubation of spore-containing cultures at 80°C for 10, 15, and 30 min did not reveal any difference in heat sensitivity between spores of strain 1A1 and LU62A1 (data not shown). The results indicate that CcdA-deficient mutants can synthesize spores with normal properties at only a low efficiency.

Sporulation-specific sigma factor activity in a CcdA-deficient mutant. To determine if CcdA is important for the sporulation-specific sigma factor cascade, we analyzed the expression of different promoter-lacZ fusions integrated as a single copy into the amyE locus of B. subtilis strains with and without an intact ccdA gene. The spoIID, sspB, and cotC promoters in the transcriptional fusions we used are dependent on ^E, ^G, and ^K, respectively, for activity. ^F-dependent gene expression was analyzed by the use of strains LUA17 and LU627, which do not synthesize ^G and produce a variant of ^F that acts like ^G, mediating transcription of the sspB promoter-lacZ fusion.

View larger version (19K): [in this window] [in a new window]   FIG. 5.   K-dependent expression of a cotC-lacZ gene fusion in the chromosome of the wild-type control strain LUA13 (solid diamonds) and CcdA-deficient strain LU623 (open diamonds) during growth in DS medium. The square symbols show OD600 (solid, LUA13; open, LU623). Time zero is the end of the exponential growth phase.

Morphology of mutant spores. Electron microscopy of samples from sporulating cultures of strain 1A1 and LU62A1 showed more cell lysis in the latter strain. Spore-like structures, often enclosed by the mother cell, were seen in samples of both strains (Fig. 6 and electron micrographs not shown). Those of LU62A1, corresponding to the nonmotile cells seen by light microscopy, showed the main features of B. subtilis spores, i.e., a cortex, a lamellar inner coat, and an electron-dense outer coat. We could not identify a consistent structural difference in the morphology of the spores of 1A1 and LU62A1, except that the cytoplasm of the latter strain appeared more granular. From the sporulation efficiencies of the strains (Tables 2 and 3), less than 10% of the LU62A1 spores observed by electron microscopy would be heat-resistant cells. Judging from the relative lysozyme resistance, the poor light diffraction properties, and the heat and chloroform sensitivity of most spores of CcdA-deficient mutants, we consider it likely that synthesis of the cortex and/or coat and dehydration of the spore are defective.

View larger version (128K): [in this window] [in a new window]   FIG. 6.   Electron micrographs of two representative B. subtilis LU62A1 cells from a culture grown for sporulation. Bars, 0.4 µm.

Complementation of sporulation defects by the ccdA gene on plasmids. The deficiency in spore synthesis of strains LU60A1 and LU62A1 was complemented by the ccdA gene or the ccdA-yneI-yneJ gene cluster on plasmids, i.e., by pLTS29 and pCCD2 (data not shown). These plasmids also restored cytochrome c synthesis, as determined by the ability of colonies on TBAB plates to oxidize TMPD. Plasmid pLTS100, a high-copy-number plasmid carrying the ccdA gene, complemented the defect in cytochrome c synthesis in strain LU60A1 and caused a ca. fivefold overproduction of CcdA protein compared to that in LU62A1/pCCD2 (immunoblot not shown). In contrast to strains containing the low-copy-number plasmids pLTS29 or pCCD2 or the vector pGDV1, those containing pLTS100 gave rise to small colonies and were unstable. We do not know if this apparent toxicity of pLTS100 is due to functional activity of CcdA or physical disturbance of the membrane caused by the increased amounts of this integral membrane protein.

Isolation of strains with suppressor mutations. Microcolonies (papilla) appeared within lysed colonies of strain LU62A1 on TBAB plates incubated at room temperature for >1 week. Cells from these microcolonies formed normal-sized colonies when streaked on TBAB plates and oxidized TMPD, albeit more poorly than did 1A1 colonies. These clones were also phleomycin resistant and contained the same ccdA-yneI-yneJ::ble deletion-substitution as in strain LU62A1, as determined by Southern blot analysis of chromosomal DNA. Thus, the clones contain suppressor mutations. The suppressor-containing strains showed wild-type sporulation efficiency (80 to 90%), and light absorption spectroscopy of membranes isolated from one clone, LU62A1R^#3, confirmed the presence of c-type cytochromes (spectra not shown). Strain LU62A1R^#3 and other suppressor-containing strains differ from LU62A1 in that they do not develop competence. The suppressor mutations can be moved to other strains by transformation and are not linked to the ccdA locus on the chromosome. The results show that the suppressor mutation(s) at one or several loci can restore all known cell defects caused by CcdA deficiency. Despite several attempts, using different strategies and isolates, we have so far not been able to identify any of the suppressor mutations.

c-type cytochromes are not required for sporulation. The four c-type cytochromes in B. subtilis, cytochrome c₅₅₀, cytochrome c₅₅₁, the cytochrome c subunit of the bc complex, and subunit II of the cytochrome caa₃ oxidase, are encoded by the cccA, cccB, qcrC, and ctaC genes, respectively (2, 49, 52). Respiration-defective B. subtilis mutants are generally sporulation deficient (43). Defective sporulation of CcdA-deficient strains could therefore be a result of the total lack of cytochrome c. To determine this, we constructed strain LUT36, which has the structural genes for all four cytochromes c deleted. LUT36 was found to have close to normal sporulation efficiency (>60%), demonstrating that cytochrome c is not important for sporulation to occur in B. subtilis.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this work we demonstrate that sporulation in B. subtilis can occur in the absence of the four known c-type cytochromes and that the ccdA gene is required for efficient sporulation. Our available experimental data, taken together, strongly suggest that CcdA deficiency affects a late step in spore synthesis, probably synthesis of the cortex and/or coat and dehydration of the spore.

The ccdA gene is cotranscribed with the yneI and yneJ genes from a promoter upstream of ccdA (36), and in this work we demonstrate that there are also promoters for the synthesis of yneI-yneJ and yneJ mRNAs. These last two promoters show a different temporal pattern of activity during growth of batch cultures compared to the promoter for ccdA. The expression from the ccdA promoter is not regulated by CcdA, YneI, or YneJ, since the absence or overproduction of these three proteins did not affect ccdA promoter activity. Genes for proteins that physically interact are in bacteria often clustered and present in a conserved order on the chromosome (7). B. subtilis genes for functionally closely related proteins, for example enzymes of a biosynthetic pathway, are generally arranged in operons (20). There is, however, no evidence suggesting that the function of the YneI or YneJ protein in the cell is related to that of CcdA. Strains lacking or overexpressing yneI and yneJ are proficient in cytochrome c and spore synthesis. Moreover, the ccdA gene in other gram-positive bacteria of known genome sequence (5) is not flanked by genes corresponding to yneI or yneJ. The B. subtilis cheY and spo0F genes encode single-domain response regulators very similar to YneI, but neither of these genes are flanked by a gene for a membrane protein like YneJ. Although transcription of ccdA, yneI, and yneJ rely on a common terminator located after the yneJ gene, we conclude that the three genes probably encode functionally unrelated proteins. The transcriptional organization of the gene cluster does not cause a problem in the cell, because the promoters in front of the genes progressively increase in strength in steps of approximately 1 order of magnitude.

Cytochromes of the c type are present in exponentially growing B. subtilis cells and increase in concentration together with other cytochromes at the end of exponential growth (45). Transcription of structural genes for c-type cytochromes and ccdA also increases at this growth phase or at the entry into stationary phase in NSMP medium (2, 51). Exponentially growing cells are expected to contain some CcdA protein, since cytochrome c is present. It might be of functional importance to maximize the expression of ccdA when the cell approaches stationary phase, since CcdA is needed for efficient spore synthesis.

The hydrophobicity profile, combined with sequence comparisons and application of the positive-inside rule (47) and topology studies using protein fusions (8, 39), indicates that the CcdA protein has six -helical transmembrane segments and the C-terminal end exposed on the outer side of the cytoplasmic membrane. Two cysteine residues, located far apart in the primary sequence, are conserved in CcdA sequences (references 8 and 36 and our unpublished data) and are functionally important (8). The amino acid sequence similarity between B. subtilis CcdA and E. coli DsbD and the importance of both proteins for a late step in cytochrome c synthesis (6, 33, 35) suggest that CcdA, together with a thioredoxin-like protein, has a function similar to that of DsbD in the transfer of reducing equivalents across the cytoplasmic membrane (28, 31). If so, the defects observed in CcdA-deficient strains might be explained by inefficient disulfide bond isomerization in proteins localized on the outer surface of the membrane.

Some cortex or coat proteins and extracellular protein factors with a function in spore coat or cortex synthesis possibly contain one or more essential disulfide bonds (reference 9 and references therein). Efficient formation of these bonds and cross-linking of proteins in the coat could be mediated by the activity of CcdA. In the absence of CcdA, proteins with several cysteine residues and located on the outer surface of the membrane would be more commonly misfolded or only slowly folded into a functional state. The resulting small amounts of functional protein in the intermembrane space between the mother cell and the forespore might limit the synthesis of heat-resistant spores, which would be observed as a reduced sporulation efficiency compared to that of wild-type strains.