To estimate the minimal gene set required to sustain bacterial life in nutritious conditions, we carried out a systematic inactivation of Bacillus subtilis genes. Among Ϸ4,100 genes of the organism, only 192 were shown to be indispensable by this or previous work. Another 79 genes were predicted to be essential. The vast majority of essential genes were categorized in relatively few domains of cell metabolism, with about half involved in information processing, one-fifth involved in the synthesis of cell envelope and the determination of cell shape and division, and one-tenth related to cell energetics. Only 4% of essential genes encode unknown functions. Most essential genes are present throughout a wide range of Bacteria, and almost 70% can also be found in Archaea and Eucarya. However, essential genes related to cell envelope, shape, division, and respiration tend to be lost from bacteria with small genomes. Unexpectedly, most genes involved in the Embden-Meyerhof-Parnas pathway are essential. Identification of unknown and unexpected essential genes opens research avenues to better understanding of processes that sustain bacterial life.

To study the functions of the uncharacterized open reading frames identified in the Bacillus subtih genome, several vectors were constructed t o perform insertional mutagenesis in the chromosome. All the pMUTlN plasmids carry a lac2 reporter gene and an inducible Pspac promoter, which is tightly regulated and tan be induced about 1000-fold. The integration of a pMUTlN vector into the target gene has three consequences: (1) the target gene is inactivated; (2) lac2 becomes transcriptionally fused t o the gene, allowing its expression pattern t o be monitored; (3) the Pspac promoter controls the transcription of downstream genes in an IPTG-dependent fashion. This last feature is important because B. subti/is genes are often organized in operons. The potential polar effects generated by the integration of the vectors can be alleviated by addition of IPTG. Also, conditional mutants of essential genes can be obtained by integrating pMUTlN vectors upstream of the target gene. The vectors are currently being used for systematic inactivation of genes without known function within the B. subtilis European consortium. pMUTlN characteristics and the inactivation of eight genes in the resA-serA region of the chromosome are presented.

Little is known about the genes and enzymes involved in sulfur assimilation in Bacillus subtilis, or about the regulation of their expression or activity. To identify genes regulated by sulfur limitation, the authors used twodimensional (2D) gel electrophoresis to compare the proteome of a wild-type strain grown with either sulfate or glutathione as sole sulfur source. A total of 15 proteins whose synthesis is modified under these two conditions were identified by matrix-assisted laser desorption/ionization time of flight (MALDI TOF) mass spectrometry. In the presence of sulfate, an increased amount of proteins involved in the metabolism of C 1 units (SerA, GlyA, FolD) and in the biosynthesis of purines (PurQ, Xpt) and pyrimidines (Upp, PyrAA, PyrF) was observed. In the presence of glutathione, the synthesis of two uptake systems (DppE, SsuA), an oxygenase (SsuD), cysteine synthase (CysK) and two proteins of unknown function (YtmI, YurL) was increased. The changes in expression of the corresponding genes, in the presence of sulfate and glutathione, were monitored using slot-blot analyses and lacZ fusions. The ytmI gene is part of a locus of 12 genes which are co-regulated in response to sulfur availability. This putative operon is activated by a LysR-like regulator, YtlI. This is the first regulator involved in the control of expression in response to sulfur availability to be identified in B. subtilis.

No abstract

Structures consisting of a genetic marker (erythromycin or kanamycin resistance, thymidylate synthetase) flanked by 3.4-kilobase direct repeats (pBR322 sequences) were inserted in 12 different locations of the Bacilus subtilis chromosome. Recombination between the repeats was followed by the loss of the genetic marker. Recombination frequencies found in different locations varied from 1.2 x 10-5 to 40 x 10-5 per cell generation. Such differences were highly significant (P < 0.001).Direct repeats, single and multiple, are maintained rather stably in the thyB region of the Bacillus subtilis chromosome, which 'indicates that recombination occurs infrequently at this location (10,17). In contrast, multiple direct repeats are not maintained stably in the pheA region of the chromosome (28). Some instability of multiple repeats was also reported in the amyE region (8). The last two observations, in line with previous reports of unstable maintenance of multiple genes in the Escherichia coli chromosome (5,20), raised the possibility that in the thyB region of the B. subtilis chromosome recombination occurs with an unusually low frequency and that elsewhere in the chromosome the frequency is higher. To test this hypothesis, we measured the recombination frequency between identical duplications generated at 12 different sites of the B. subtilis chromosome. At all sites the recombination frequency was low and generally close to that measured previously in the thyB region. Significant differences between certain sites were nevertheless detected. MATERIALS AND METHODSBacterial strains and plasmids are listed in Table 1. Enzymes were commercial preparations and were used as specified by the suppliers. Plasmid and chromosomal DNA were prepared by previously described methods (2, 7). Competence induction and transformation of E. coli and B. subtilis cells was described before (3, 15). To select resistant E. coli cells, ampicillin (Ap), 50 ,ug/ml, and erythromycin (Em), 250 ,ug/ml, were used. Chloramphenicol (Cm), 3 to 5 ,ug/ml, kanamycin (Km), 5 pLg/ml, erythromycin, 0.3 pug/ml and trimethoprim (Tp), 30 ,ug/ml (in the presence of thymine [50 ,ug/ml]) were used for selecting resistant B. subtilis cells.Labeled probes were prepared by nick translation (21) and used as described before (23).E. coli HVC45 and HVC1001 and B. subtilis SB202 and HVS496 were used as hosts for plasmid constructions. HVS496 was obtained as follows. First, plasmid pHV438 (16), which is composed of pBR322, the Cmr gene from pC194, the B. subtilis thyB gene, and a B. subtilis chromosomal segment of unknown genetic content (named X) was cleaved with PvuII, ligated, and introduced into E. coli cells.This treatment deleted 700 base pairs (bp) from the thyB gene and gave the plasmid named pHV482. This plasmid was cleaved at its unique Sall site and used to transform B. RESULTSExperimental strategy and strain construction. To measure recombination frequencies in the B. subtilis chromosome, we generated a structure consisting of direct repeats flanking a genetic ...

Direction of DNA entry in Bacillus subtilis competent cells was studied using molecules in which only one of the two strands was radioactively labelled. The label was either distributed homogeneously or was localized in a small region of the strand, in the centre or at one of the ends. Regardless of the distribution and the position of the label, similar amounts of radioactivity were taken up by the cells exposed to the labelled molecules. This suggests that DNA enters B. subtilis either by two different uptake systems having opposite polarities, or by a single non-polar system.

The efficiencies of intermolecular recombination at 12 different locations on the Bacilus subtilis chromosome were determined by transforming competent cells with a nonreplicative plasmid. The efficiencies varied by only about threefold but were significantly different (P < 0.05 by a chi-square test) for -20%o of the locations. The recA4 gene product is required for recombination, and the add4 gene product appears to affect the variation in a site-specific way.Directly repeated sequences are known to recombine with a rather low frequency in the Bacillus subtilis chromosome (15,22). However, significant variations were observed when the identical repeats were localized at different sites along the chromosome, the reported values being between 1.2 x 10-5 and 4 x 10-4 per generation for 3.4-kb-long, pBR322-derived, repeated sequences (26). This suggests that signals which affect intramolecular recombination might be present in the B. subtilis chromosome. The nature of these putative signals is not known, but it was observed that recombination could be stimulated 10-to 100-fold by integrating an active, or even partially active, plasmid replication origin into the vicinity of the repeats (23,27).In this work, we compared the efficiencies of intermolecular recombination at 12 chromosomal locations, using an assay based on transformation of competent cells. All values were fairly similar, the extremes differing by only about threefold, and no correlations between the relative efficiencies of intermolecular recombination and the previously determined frequencies of intramolecular recombination (26) at the same locations were observed. However, pairwise comparisons indicated significant differences (P 5 0.05 by a chi-square test) between the efficiencies of intermolecular recombination at -20% of chromosomal locations. The threefold difference between the extreme values was no longer observed in strains carrying a mutated addA gene and therefore lacking the RecBCD nuclease (1, 2). MATERIALS AND METHODSBacterial strains and plasmids are listed in Table 1. Enzymes were commercial preparations and were used according to the suppliers' instructions. Chromosomal DNA from B. subtilis and plasmid DNA from Escherichia coli were prepared by previously described methods (5, 13). Induction of competence and transformation of E. coli and B. subtilis cells were described previously (3,8,20). E. coli transformants were selected on 50 pg of ampicillin per ml, and B. subtilis transformants were selected on 3 ,ug of chloramphenicol per ml, 0.3 pg of erythromycin per ml, or 10 ,ug of kanamycin per ml. 2P-labeled radioactive DNA probes were prepared by nick translation and used as described by Southern (25). Plasmid pHV489 was constructed by inser-* Corresponding author. tion of the 1.9-kb HindIII fragment of plasmid pLS54 (6), which contains a part of the B. subtilis g1tA gene, into the unique HindIII site of plasmid pHV1025. RESULTSExperimental strategy and strain description. To compare efficiencies of intermolecular recombination at d...