INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

^B was discovered in 1980 by Haldenwang and Losick and was the first alternative sigma factor of Bacillus subtilis identified (19). However, the function of the regulon controlled by ^B remained a matter for speculation until 1993, when it was shown to be involved in the cellular response to stress. It has subsequently been demonstrated that expression of a large number of genes is induced in a ^B-dependent manner by such different stimuli as heat shock, ethanol, acid, and salt stress, and starvation for oxygen, phosphate, and glucose (6, 7, 10, 11, 21, 22, 40, 43). Since induced expression of more than 50 genes is absolutely dependent on ^B, it was tempting to assume that the gene products perform essential adaptive functions in B. subtilis. However, earlier studies had shown that a null sigB mutant strain was apparently not impaired in sporulation or response to stress compared to the wild type (9, 10, 21, 23, 39). This is unusual, since starving or stressed B. subtilis cells devote a considerable amount of their residual protein-synthesizing capacity to the synthesis of the members of the ^B regulon (8). This apparently anomalous result has prompted a concerted effort to identify the genes of the ^B regulon, to elucidate the function(s) of their gene products, and to establish their contributions to the cellular response to stress and starvation conditions.

Among others, genes encoding a catalase (katE), a nonspecific DNA-binding and protecting protein (dps), and an osmotically activated proline uptake system (opuE) have been shown to be subject to the control of ^B in B. subtilis (5, 13, 44). Interestingly, in Escherichia coli genes like katE, dps, and proP, whose gene products perform similar functions, are subject to a RpoS-dependent regulation (28, 30, 34). RpoS directs the expression of a large group of genes whose expression is induced following starvation and stress (20, 33). Null mutations in rpoS result in a loss of stationary-phase-induced resistance against heat, acid, or oxidative stress and impairment in the ability to survive prolonged periods of starvation (20, 26, 27, 29). Therefore, it was tempting to speculate that ^B-dependent stress proteins may provide the stressed or starved B. subtilis cell with a general multiple-stress resistance similar to that provided by the RpoS-dependent proteins of E. coli. This view is supported by the demonstration that sigB mutants are impaired in stationary-phase-induced resistance to oxidative stress, like E. coli rpoS mutants (4, 12). Recently, the Dps protein of B. subtilis has been shown to play a crucial role in the development of this nonspecific starvation-mediated resistance to oxidative stress (5). However, it is necessary to identify and investigate the physiological roles of additional general stress proteins to further support this hypothesis. Identification of general stress proteins by N-terminal sequencing (3, 8, 40, 41) has greatly benefited from the recent release of the complete sequence of the B. subtilis genome (25). In this paper we present an investigation of two genes, yklA and ykzA, identified in B. subtilis during the genome-sequencing project, which show homology to osmC from E. coli. The expression profile of osmC in E. coli is complex (16, 18). It has two independent promoters, which provide medium-, growth phase-, and stress-dependent expression. One of the promoters is partially controlled by RpoS. Although there are two genes in B. subtilis which have homology to osmC, the composite expression profile of the two genes (medium-, growth phase-, and stress-dependent expression) is very similar to that of the E. coli gene. In addition, expression of the homolog YkzA is directed by ^B, the B. subtilis stress sigma factor.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and growth conditions. The B. subtilis wild-type strain 168 and its isogenic sigB mutant strain ML6 (trpC2 sigB::HindIII-EcoRV::cat [21]) were cultivated at 37°C under vigorous agitation in Luria broth (LB) (31) or a synthetic medium described previously (37). The synthetic medium had to be used in order to achieve glucose limitation and in the experiments involving salt stress, in order to avoid the protective effects of osmoprotective substances present in LB. Stresses and starvation were imposed as described previously (40, 42).

Cloning, sequencing, and construction of the corresponding mutants. B. subtilis chromosomal DNA for sequencing was isolated from strain 168 by chromosomal walking with the integrating plasmid pDIA5304 as previously described (24). E. coli TP611 (recBC hsdRM cya610 pcn) was used for cloning large chromosomal DNA fragments (15). E. coli TG1 [K-12 (lac pro) supE thi hsdR F' traD36 proAB lacI lacZM15] was used for subcloning and for preparation of sequencing templates. E. coli strains were routinely grown in LB. The isolated chromosomal DNA from B. subtilis was sequenced by a shotgun strategy and by a directed approach with oligonucleotides. Plasmid DNA (30 µg in 120 µl of Tris-EDTA buffer) was randomly sheared either by sonication (Braun Labsonic sonicator) (seven pulses, 0.22 cycles, 0.25 W) or by using DNase I in the presence of manganese (35). Sequencing reactions were carried out with fluorescent dye primer sequencing kits (GENPAK, Polegate, East Sussex, England) according to the manufacturer's instructions. Reactions were resolved on an ABI automated sequencer model 373A. Gaps in the sequence were filled with custom-synthesized oligonucleotides (PCR-Mate DNA synthesizer; Applied Biosystems) and the Applied Biosystems dye terminator sequencing kit.

RNA isolation and analysis of transcription. RNA was isolated with RNeasy cartridges from Qiagen as described previously (42). Northern blot analysis, hybridization, and quantification of specific hybrids were performed as described by Scharf et al. (36). For the preparation of the digoxigenin-labeled RNA probes, a DNA fragment encompassing the two osmC-homologous genes yklA and ykzA was amplified from chromosomal DNA of the wild-type strain 168 with the synthetic oligonucleotides UV114 (5'-GAGAGGATCCGTGAATAGCGGGGTAATG-3') and UV115 (5'-GAGAATCGATGTCCGACACCAAAAAACATC-3'). The PCR fragment was digested with BamHI/ClaI and cloned into pBluescript KS() digested with the same enzymes. The resulting plasmid, pUV321, was digested with HincII and religated, yielding plasmid pUV521. pUV521 can be used for the production of a digoxigenin-labeled, yklA-specific RNA probe with T3 RNA polymerase after linearization with BamHI. Digestion of pUV321 with BamHI/BglII and religation of the remaining plasmid resulted in pUV520, which is devoid of yklA and can be used for the preparation of digoxigenin-labeled, ykzA-specific RNA probe with T3 RNA polymerase after linearization with SphI.

Two-dimensional protein gel electrophoresis. Protein extracts were prepared by passage through a French press after cells had been harvested on ice. Equal amounts of protein (300 µg) were loaded. The proteins were separated with immobiline dry strips, pH 4 to 8, in the first dimension on the multiphor apparatus supplied by Pharmacia, equilibrated, loaded onto 12.5% polyacrylamide gels, and separated according to their molecular masses with the InvestigatorTM electrophoresis system of ESA Inc. (Chelmsford, Mass.).

Enzyme assays. Expression of lacZ was measured as described by Ferrari et al. (14) with the following modifications: activity units are expressed in nanomoles per minute per microgram of protein and the cells were lysed for 25 min in Z buffer containing 10 µg of lysozyme per ml, 1 mM dithiothreitol, 0.00025% Triton X-100, and 1 µg of DNase I per ml.

Computer sequence analysis. Sequence alignment and editing were performed with the XBAP program of the STADEN package. Conceptual translation of the sequence and other sequence analyses were performed with the NIP program of the STADEN package. The GenBank database was accessed with ACNUC (17), and homology searches of the database were performed with the TBLASTN program (1). Multiple sequence alignments were performed with CLUSTAL W (38).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning and sequencing of the chromosomal region containing yklA and ykzA. Two homologs of osmC from E. coli were identified during the B. subtilis genome-sequencing project (25). They are called yklA and ykzA and are arranged as shown in Fig. 1 at approximately 105° on the chromosome. Their sequence can be found in GenBank entry AJ002571. Both genes are transcribed in the direction of chromosomal replication and are separated by ykmA, which is transcribed in the opposite direction. There is a putative ^A promoter positioned upstream of yklA and a putative ^B promoter located upstream of the ykzA gene. The paralogs YklA and YkzA differ in size by five amino acids (141 and 136, respectively) and are 49% identical (67% similar) to each other. Both YklA and YkzA are approximately 28% identical (42% similar) to OsmC from E. coli.

View larger version (7K): [in this window] [in a new window]   FIG. 1.   Schematic diagram of chromosomal organization in the region containing the yklA and ykzA genes. Open reading frames are indicated by arrows, and their positions indicate whether they are encoded on the top (above) or the bottom (below) DNA strand. Terminators are indicated by "lollipops" on the top (above) or the bottom (below) strand. The terminator between yklA and ykmA is bidirectional.

Transcriptional regulation of yklA and ykzA. Expression of the osmC gene from E. coli is subject to osmotic and growth phase-dependent regulation, which are partially dependent on the presence of the stress and stationary-phase sigma factor RpoS. Therefore, we wanted to determine if one or both of the B. subtilis homologous genes yklA and ykzA are similarly regulated by osmotic stress or starvation. Total RNA was prepared from exponentially growing cells and from bacteria which had been treated with sodium chloride or which entered the stationary phase as a result of the exhaustion of glucose. An analysis of the ykzA mRNA level revealed that the expression of this gene was strongly and rapidly induced by salt stress (Fig. 2). The induction was transient, reaching a maximum between 9 and 12 min after the imposition of stress. ykzA was also induced during the exhaustion of glucose, but clearly the level of induction was less pronounced than during salt stress. Induction by stress was not confined to salt stress. A similar very strong and transient induction of ykzA was also measured following heat shock and ethanol stress (Fig. 2). Therefore, ykzA belongs to the group of general stress genes induced by multiple stimuli in B. subtilis. When the same RNA preparations were probed with a yklA-specific probe, we failed to detect significant changes in the expression of yklA in response to any of the stimuli examined (Fig. 2).

View larger version (46K): [in this window] [in a new window]   FIG. 2.   Levels of yklA- and ykzA-specific mRNA before and after the imposition of different stresses. RNA was prepared from the wild-type strain 168 at the times indicated on the x axes. Specific RNAs were detected with digoxigenin-labeled RNA probes, and the intensities of the signals were quantified with a laser densitometer as described previously (36). The amount of RNA present during exponential growth was set to one. The induction ratios of yklA (solid bars) and ykzA (shaded bars) at the different time points are displayed. All stresses were applied at time zero with the exception of glucose limitation, where zero indicates the point at which the culture ceased to grow.

View larger version (70K): [in this window] [in a new window]   FIG. 3.   Northern blot analysis of ykzA. Equal amounts of total RNA (10 µg) prepared from the wild-type 168 or the isogenic sigB mutant (ML6) before (co) and at different times (in minutes) after exposure to stress were separated on denaturing gels, transferred onto a positively charged nylon membrane, and hybridized with digoxigenin-labeled RNA probe specific for ykzA.

View larger version (45K): [in this window] [in a new window]   FIG. 4.   Mapping of the 5' ends of the yklA (A) and ykzA (B) mRNA during growth and after exposure to stress. Total RNA was isolated from the wild-type strain and its isogenic sigB mutant during exponential growth (co), 10 min after the imposition of the different stresses (h, heat shock; e, ethanol stress; s, salt stress), and 30 or 40 min after the limitation of glucose (cl1 and cl2, respectively). The primer extension analysis was performed as described in Materials and Methods. The 5' ends of the transcripts were determined by comparison with a DNA-sequencing ladder generated with the same primer and run in parallel on the same gel (lanes A, C, G, and T).

Expression of yklA and ykzA in minimal and rich medium during the growth cycle and after induction by salt. Two strains (BFS1816 and BFS1818), mutated in yklA and ykzA, respectively, were constructed by integration of pMutin004 and pMutin005 into the chromosome of B. subtilis. The pMutin-derived plasmids contained an internal fragment of the yklA and ykzA open reading frames, respectively, cloned immediately upstream of a promoterless lacZ gene, allowing the expression of each gene to be examined. The strains grew and sporulated normally both on minimal medium and on medium containing 0.3 M NaCl. Expression of yklA-lacZ and ykzA-lacZ was examined in nutrient broth and minimal medium throughout the growth cycle. When cells harboring yklA-lacZ were grown in nutrient medium, -galactosidase activity reached approximately 20 U during exponential growth and decreased slightly at the onset of the stationary phase (Fig. 5A). When these cells were grown in minimal medium, -galactosidase levels rose during the early period of the growth cycle. This accumulation reached a plateau of approximately 90 U by the midpoint of the growth cycle (T₃), and this level was maintained for the remainder of the growth cycle (Fig. 5A). Addition of 0.3 M NaCl to exponentially growing cells containing yklA-lacZ did not affect the -galactosidase activity (data not shown), confirming the results of the slot blot analysis, which showed that expression of yklA is not responsive to osmotic stress (Fig. 2).

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Expression of the yklA-lacZ (A) and ykzA-lacZ (B) transcriptional fusions during the growth cycle in minimal medium or nutrient broth and after exposure to salt stress. Cells were grown as described in Materials and Methods, and samples were taken every 30 or 60 min as indicated. (A) solid symbols indicate growth of B. subtilis BFS1816, and open symbols indicate -galactosidase activity. Circles, nutrient medium; squares, minimal medium. (B) Solid squares represent growth of B. subtilis BFS1818 in minimal medium, and open squares indicate -galactosidase activity. The influence of the addition of salt during exponential growth on the accumulation of -galactosidase is also indicated (open triangles). The point of salt addition is indicated by an arrow. OD550, optical density at 550 nm.

Identification of YkzA on two-dimensional protein gels; level of YkzA during exponential growth and after imposition of stress. We have determined the N-terminal sequences of general stress proteins of B. subtilis by microsequencing (3, 8, 40). When comparing these sequences with the sequences of YklA and YkzA, we discovered that the N-terminal sequence of the general stress protein Gsp17o (ALFTAKVTAR GGRAAHITSD D) matched the amino acid sequence deduced from the ykzA DNA sequence (with the exception of the alanine residue at position 15). Therefore, the ATG codon at position 34145 of the DNA sequence is indeed the start codon of ykzA, and the N-terminal formyl-methionine is subsequently removed. Two-dimensional protein gel electrophoresis of crude protein extracts from cells harvested during exponential growth or after imposition of stress was used to show that levels of YkzA significantly increased following heat, salt, and ethanol stress and that the stress sigma factor ^B was required for this increase to occur (Fig. 6). The intensity of a reference spot corresponding to the ribosomal protein RplJ did not increase during the same time interval (Fig. 6).

View larger version (102K): [in this window] [in a new window]   FIG. 6.   Level of YkzA in the wild type (wt) and the sigB mutant (sigB) before (co) and after the imposition of stresses (heat, ethanol [EtOH], and NaCl). Bacteria were grown in LB (co, heat, and EtOH) or minimal medium (co M and NaCl) and harvested during exponential growth, 60 min after the imposition of heat shock or ethanol, or 90 min after exposure to NaCl. Crude protein extract (300 µg) was separated by two-dimensional gel electrophoresis and stained with Coomassie brilliant blue R-350. Besides YkzA, GsiB is indicated as an additional B-dependent stress protein and RplJ is labeled as a vegetative marker protein. The identities of the labeled proteins were verified by microsequencing or mass spectrometry.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The expression profiles of two B. subtilis genes, yklA and ykzA, whose products are highly similar to the general stress protein OsmC first identified in E. coli, have been presented. The profiles are complex, with expression of the genes being growth phase, medium, and stress dependent. Expression of yklA is directed by a ^A-type promoter, and it is maximally expressed during exponential growth. The expression of yklA is four- to fivefold higher in minimal medium than in a rich medium, and it is not induced by stress.

Expression of ykzA, in contrast, is directed by a promoter which requires ^B, the so-called stress sigma factor of B. subtilis, for initiation of transcription. Expression of ykzA is not medium dependent and is very low throughout exponential growth. However, expression of ykzA is rapidly induced by salt and ethanol stress, heat shock, and starvation. Therefore, the expression patterns, although complex, appear complementary, with yklA being medium responsive whereas ykzA is stress and starvation responsive.

It is instructive to compare the organization and expression of osmC from E. coli with those of yklA and ykzA from B. subtilis. There is only one osmC gene in E. coli, whose expression is directed by two overlapping but apparently independent promoters (16, 18). In contrast, B. subtilis has two osmC homologs, each expressed from a single promoter. The osmCp₁ from E. coli is recognized by the housekeeping sigma factor Sigma70. Similarly, the yklA gene from B. subtilis seems to be recognized by the housekeeping sigma factor ^A. Expression of osmC from the osmCp₂ promoter and expression of ykzA are directed by the stress sigma factors RpoS and ^B, respectively. Induction of these genes, which can be effected by a variety of stresses, is dependent on these sigma factors. Expression directed by osmCp₂ also increases when the cells enter stationary phase, as does expression of ykzA in B. subtilis. Although only the ^B-dependent promoter of ykzA is induced following salt stress in B. subtilis, both promoters of osmC of E. coli are salt responsive.

Nevertheless, despite the overt differences in gene organization between the two bacteria, the similarity of the composite expression profiles is striking. It is evident that both bacteria must regulate the level of the general stress protein with great precision, and expression must be responsive to growth, medium, and stress conditions. In E. coli this is achieved by the expression of one gene being directed by two promoters, whereas in B. subtilis it is achieved by having two genes each directed by a single promoter.

Our data clearly demonstrate that there are three RNA transcripts produced upon induction of ykzA. All three transcripts begin at the ykzA promoter. The major transcript is 0.5 kb in length, which is consistent with transcription ceasing at the putative terminator located immediately downstream of ykzA. The lengths of the other two transcripts are consistent with transcription proceeding through the ykzA terminator and ending at the putative terminators for the ykoA and metC genes, respectively, which are expressed from the strand opposite to ykzA. Therefore, it is evident from our data that there is a disparity between the strength of the ykzA promoter and that of the terminator.

Despite our extensive knowledge of the organization and control of expression of osmC, yklA, and ykzA, the functions of the proteins are unknown. It is evident that they play roles in cellular response to a variety of stressful conditions, but their precise functions remain to be established. No obvious phenotype is observed, even under stressful conditions, when osmC is inactivated in E. coli or when either gene is inactivated in B. subtilis. The occurrence of osmC-homologous genes among bacteria does not correlate with any bacterial group or ecological niche and so does not shed any light on its function. There are now 11 members of this gene family distributed among the following bacteria: E. coli (one copy), B. subtilis (two copies), Mycoplasma pneumoniae (one copy), Mycoplasma genitalium (one copy), Acinetobacter calcoaceticus (one copy), Xanthomonas campestris (one copy), Pseudomonas aeruginosa (two copies), and Deinococcus radiodurans (two copies). However, no member of this gene family has been identified in the complete genome sequences of Haemophilus influenzae, Helicobacter pylori, or Synechocystis sp. An alignment of the 11 proteins reveals four regions which are absolutely conserved (data not shown): (i) a glycine residue near the amino terminus, (ii) an NPEQ/EXL motif, (iii) a CF motif, and (iv) an AXXXCPXS motif. These motifs do not show similarity to any other motif in the database. However, conservation of the two cysteine residues is interesting, suggesting that perhaps the protein contains a disulfide bond which is required for activity. Alternatively, it may bind a metal ion or may participate in maintaining disulfide bonds in other proteins, i.e., a type of disulfide bond chaperone. At least the ohr gene of X. campestris, which is a member of the osmC family, is required for protection against organic hydroperoxides (32). A phylogenetic analysis of the eleven proteins (Fig. 7) shows that they can be grouped into three families: (i) the E. coli family, which includes, besides osmC, one each of the D. radiodurans and P. aeruginosa genes; (ii) the Mycoplasma family; and (iii) a family containing yklA and ykzA from B. subtilis, one each of the D. radiodurans and P. aeruginosa genes, and the genes of A. calcoaceticus and X. campestris. The interesting feature of this tree is that yklA and ykzA are more closely related to each other than to any other member of the family. In contrast to P. aeruginosa and D. radiodurans, which also have two copies of osmC, the paralogs fall into distinct phylogenetic groups. This suggests that the duplicated genes in B. subtilis have not evolved to fulfill separate functions within the cell. Instead we propose that the duplication provides a mechanism for the Bacillus cell to regulate OsmC levels in response to a wide range of environmental and nutritional stimuli by placing each copy of the gene under the control of different but complementary expression signals. In E. coli, the environmental and nutritional conditions under which osmC is expressed are very similar to those in Bacillus. However, the mechanism through which this is achieved differs in that expression of a single gene is directed by two different but independent promoters.

View larger version (17K): [in this window] [in a new window]   FIG. 7.   Unrooted phylogenetic tree of the relationships between the 11 members of the osmC gene family produced from a multiple alignment by using the CLUSTAL W program. The data were bootstrapped 1,000 times, and the values are indicated on the horizontal axes. Drad, D. radiodurans; Paer, P. aeruginosa; Ecol, E. coli; Xant; X. campestris; Acal, A. calcoaceticus; Bsub, B. subtilis; Mpne, M. pneumoniae; Mgen, M. genitalium.

The complexity of osmC gene expression in E. coli and B. subtilis suggests that it plays an important role in the response of the cells to stress. Since protection of stressed or starving cells from oxidative stress seems to be a premier function of the ^B regulon, we are currently investigating whether YkzA and/or YklA is involved in establishing a protective resistance, as does Ohr of X. campestris (32).