INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Asparagine biosynthesis in the gram-positive bacteria has not been studied extensively. We chose Bacillus subtilis as a convenient bacterium for such study, since it is able to grow well in minimal media without asparagine, implying that it possesses efficient asparagine biosynthesis pathways. In addition, the completion of the genome sequencing of this organism (10) should allow the identification of genes which could be involved in asparagine biosynthesis.

The reactions that are catalyzed by asparagine synthetase use either glutamine or ammonia as a nitrogen source, as follows: L-Asp + ATP + NH₃  L-Asn + AMP + PP_i (reaction 1) and L-Asp + ATP + L-Gln  L-Asn + AMP + PP_i + L-Glu (reaction 2). To our knowledge, two families of asparagine synthetase have been reported. One is the AsnA family, represented by AsnA of Escherichia coli, whose members were found in prokaryotes such as E. coli and Klebsiella aerogenes (8, 15). Members of the AsnA family are able to use only ammonia as the amino group donor, as in reaction 1. The other is the AsnB family, represented by AsnB of E. coli, whose members were found in both prokaryotes and eukaryotes (7, 18, 20). Members of the AsnB family are able to use both glutamine and ammonia as the nitrogen donor, but glutamine is preferred. E. coli and K. aerogenes have two asparagine synthetase genes, asnA and asnB, and the presence of either ensures sufficient asparagine biosynthesis, while inactivation of both causes asparagine auxotrophy (8, 15).

Analysis of the genome sequence of B. subtilis predicted three genes encoding glutamine-dependent AsnB-type enzymes but no gene for an ammonia-dependent AsnA-type enzyme. The three genes were designated asnB, asnH, and yisO (10); the last gene is referred to as asnO in this paper. We report here that each of the three genes encodes an asparagine synthetase and describe their expression pattern as well as the study of mutants lacking the three genes individually or in combination, revealing a physiological role for asnB in vegetative cells and for asnO in sporulating cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and media. The bacterial strains used in this study are listed in Table 1. Plasmids pOU71 (11), pBEST-4F and pBEST513, pIC156, and pUC19 (23) were provided by Seiichi Yasuda (Cloning Vector Collection, National Institute of Genetics, Mishima, Japan), Mitsuo Itaya (Mitsubishi Kasei Institute of Life Sciences, Tokyo, Japan), Rozenn Dervyn (Institut National de la Recherche Agronomique, Jouy-en-Josas, France), and Takara Shuzo Co., Ltd. (Ohtsu, Japan), respectively. Plasmid pMUTIN2mcs (19) was provided by Valérie Vagner (Institut National de la Recherche Agronomique, Jouy-en-Josas, France). E. coli cells harboring plasmids were grown on following media containing ampicillin (50 µg/ml): Luria broth (LB) (16) and M9 minimal medium (16) supplemented with asparagine-free Casamino Acids (2 mg/ml) (Difco), thiamine (50 µg/ml), thymine (5 µg/ml), and, when required, asparagine (50 µg/ml). B. subtilis cells were grown on the following media containing appropriate antibiotics when needed (see below): tryptose blood agar base (Difco) supplemented with 0.18% glucose (referred as TBABG), DSM (17), and S6 minimal medium (4) supplemented with tryptophan (50 µg/ml), 0.02% Casamino Acids, and, when required, asparagine (S6 plates were prepared by adding 2.0% Noble agar [Difco] containing no nitrogen source).

Construction of recombinant plasmids. E. coli plasmids pASNB, pASNH, pASNO, and pYXBB, carrying asnB, asnH, asnO, and the genes from yxbB to asnH of B. subtilis, respectively (Fig. 1), were constructed as follows. DNA fragments carrying the entire coding and 5' flanking regions of the genes were amplified by PCR with specific primer pairs and chromosomal DNA of B. subtilis 168 as a template (Fig. 1). All PCR was done with a GeneAmp XL PCR kit (Perkin-Elmer). The specific primer pairs used were as follows (restriction sites are underlined): for pASNB, asnBupB (5'-CGCGGATCCATAGCCGCTTACTGGTTAAG-3') and asnBdnB (5'-CGCGGATCCTGGGTAAATCAATGATGATGG-3'); for pASNH, asnHupE (5'-CCGGAATTCTCGTAAATACCCACACTTGG-3') and asnHdnB (5'-CGCGGATCCATTGCTAATCCCCTAAGTGC-3'); for pASNO, asnOupE (5'-CCGGAATTCTTTCCGTTTCATCCATGCTG-3') and asnOdnB (5'-CGCGGATCCTCTTATTGAAGGAATGCGGG-3'); and for pYXBB, yxbBupE (5'-CCGGAATTCTACAAGGAAGGAGGGAAAAG-3') and asnHdnB (5'-CGCGGATCCATTGCTAATCCCCTAAGTGC-3'). The PCR product for the pASNB construction was trimmed with BamHI and then ligated with pOU71 previously digested with BamHI, and each of the other three products was cleaved with EcoRI and BamHI and then ligated with pOU71 previously digested with EcoRI and BamHI. The ligated DNAs were introduced into E. coli JM109 by transformation to give ampicillin resistance on LB plates. Plasmids in the transformants were extracted, and the identity of each of the PCR products cloned into pOU71 was verified by digesting them with various restriction enzymes.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Genetic organization of the asnB, asnH, and asnO (yisO) regions and DNA stretches amplified by PCR for plasmid and mutant constructions and probe preparations. The genetic organization of the asnB (top), asnH (middle), and asnO (bottom) regions is shown schematically, and a scale bar with nucleotide positions within the whole genome sequence (10) is given for each of the regions. Genes are shown as thick arrows; solid arrows indicate genes within possible transcription units that contain one of the asparagine synthetase genes. DNA stretches amplified by PCR are shown as solid lines, and names of plasmids, mutants, and probes prepared with the respective PCR products are given on the left. The position and orientation of each specific primer are shown with small arrowheads. The length and orientation of each of antibiotic resistance gene cassette ligated with the PCR products are shown as thinner arrows. The antibiotic resistance genes are abbreviated as follows: cat, chloramphenicol acetyltransferase gene; neo, neomycin resistance gene; spc, spectinomycin resistance gene.

Construction of B. subtilis mutant strains. B. subtilis BFS41, BFS55, BFS56, and FU339, carrying transcriptional fusions of ytnA, asnH, asnB, and asnO to lacZ, respectively, were constructed as follows. DNA fragments (approximately 300 bp) corresponding to initial parts of each of the genes were amplified by PCR with specific primer pairs and chromosomal DNA of B. subtilis 168 as a template (Fig. 1). The specific primer pairs used for the constructions were as follows (restriction sites are underlined): for BFS41, ytnAH (5'-CCCAAGCTTTAGGGGAGAAGAAGCATG-3') and ytnAB (5'-CGCGGATCCACCAGTAGTTCCAACCTG-3'); for BFS55, asnHH (5'-CCCAAGCTTCAATAACGCTATTGGGAG-3') and asnHB (5'-CGCGGATCCTGCTGTCCATTTACAAGG-3'); for BFS56, asnBH (5'-CCCAAGCTTTAGGGGTTCAATGATGAC-3') and asnBB (5'-CGCGGATCCTCTCTCAGTTCGATATAG-3'); and for FU339, asnOH (5'-CCCAAGCTTGATTGGAGCTGATGTCAC-3') and asnOB (5'-CGCGGATCCAATGGTGTAGCTATCGCC-3'). Each of the PCR products was trimmed with HindIII and BamHI and was then ligated with pMUTIN2mcs previously digested with HindIII and BamHI. Plasmid pMUTIN2mcs (lacZ lacI amp erm) replicates in E. coli but not in B. subtilis and carries an erythromycin resistance gene that is active in B. subtilis (19). In addition, pMUTIN2mcs carries a promoterless lacZ gene derived from E. coli that can be used as a reporter gene (19). The ligated DNAs were introduced into E. coli C600 by transformation to give ampicillin resistance on LB plates. The identity of each of the PCR products cloned into pMUTIN2mcs was verified by DNA sequencing. The resulting four plasmids, pYTNA1, pASNH1, pASNB5D, and pMASNO, were used to transform B. subtilis 168 to erythromycin (0.3 µg/ml) resistance on TBABG, providing B. subtilis BFS41, BFS55, BFS56, and FU339, respectively. Correct integration of a single copy of each plasmid into the respective genes through a single-crossover event was confirmed by Southern blot analysis. In these strains, each of the target genes was inactivated, and instead lacZ was expressed under the regulation of its upstream sequence.

RNA techniques. Cells of B. subtilis strains were inoculated into DSM to an optical density at 600 nm (OD₆₀₀) of about 0.05 and allowed to grow at 37°C with shaking. The cells were harvested at 1.5 h (in exponential growth), 3 h (at the transition between exponential growth and stationary phase), 4 h (defined as the beginning of sporulation), 5 h (1 h after the beginning of sporulation), 7 h (3 h after the beginning of sporulation), and 9 h (5 h after the beginning of sporulation) after inoculation, at OD₆₀₀s of approximately 0.4, 1.5, 2.1, 2.3, 2.0, and 2.3, respectively. Total cellular RNA was extracted by mixing the cells with glass beads, phenol, and cetyltrimethylammonium bromide and then purified as described previously (22). Alternatively, cells were inoculated into S6 medium and grown as described above to extract RNAs at 3 h (earlier in exponential growth), 5 h (later in exponential growth), and 8 h (in stationary phase) after inoculation, at OD₆₀₀s of approximately 0.5, 1.5, and 3.7, respectively.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Three asparagine synthetase homologs are encoded in the B. subtilis genome. Three genes, asnB, asnH, and asnO (originally named yisO), that encode asparagine synthetase homologs are present in the B. subtilis genome and are located at kb 3126.80, 4097.70, and 1156.80, respectively (10) (Fig. 1). Similarity comparisons among the three gene products and AsnB of E. coli (18) indicated that the three products were paralogous to each other, with the highest similarity between AsnB and AsnO, and also orthologous to AsnB of E. coli to almost the same extent (Fig. 2A). As shown in a multiple alignment of the amino acid sequences of the four proteins (Fig. 2B), they had higher similarity in their N-terminal parts, while the C-terminal parts were less conserved with respect to both sequence and length.

View larger version (61K): [in this window] [in a new window]   FIG. 2.   Similarity among the putative products of asnB, asnH, and asnO of B. subtilis and asnB of E. coli. (A) Similarity among the four gene products. Similarity was calculated by using the FASTA program (14) for each pairing among the four gene products. The FASTA optimized score (boldface) and sequence identity (percentage of overlapping amino acid residues) are shown. The size of each of the gene products is given as amino acid residues (aa). (B) Alignment of the amino acid sequences of the gene products. The amino acid sequence alignment of AsnB, AsnH, and AsnO of B. subtilis (ASNB BSU, ASNH BSU, and ASNO BSU, respectively) and AsnB of E. coli (ASNB ECO) was performed with the CLUSTAL W program (6). Conserved and related amino acid residues are marked with asterisks and dots beneath the sequences, respectively. Gaps introduced within the sequences to optimize the alignment are shown by hyphens. Conserved Cys and Asp residues are boxed (see text).

Expression of the asparagine synthetase genes in B. subtilis. To analyze expression of each of the three asparagine synthetase homolog genes, we constructed transcriptional fusions of each gene with the E. coli lacZ reporter by integration of pMUTIN2mcs into the respective loci on the chromosome via a single-crossover event (see Materials and Methods). -Galactosidase activity in extracts of the constructs grown in a rich sporulation medium, DSM, was then measured (Fig. 3A). The asnB-lacZ fusion was expressed during exponential growth and repressed in stationary phase. Expression of the asnH-lacZ fusion was dramatically elevated at the transition between exponential growth and the stationary phase. The asnO-lacZ fusion was expressed less well, but the expression was nevertheless significant during sporulation. Thus, the three genes in B. subtilis were likely to be expressed differentially during growth in DSM.

View larger version (48K): [in this window] [in a new window]   FIG. 3.   Expression analysis of asnB, asnH, and asnO in B. subtilis cells grown in a rich sporulation medium. (A) Expression of lacZ reporters fused with each of the asparagine synthetase homolog genes. Cells of B. subtilis 168 (wild type), BFS55 (asnH::pMUTIN2mcs), BFS56 (asnB::pMUTIN2mcs), and FU339 (asnO::pMUTIN2mcs) were cultured in DSM. At various intervals, cells in 1 ml of the cultures were harvested, and -galactosidase activity (nanomoles of 2-nitrophenyl--D-galactopyranoside hydrolyzed per minute per OD600 unit) in cell extracts was determined as described previously (22). Activities of the asnB-lacZ fusion in cells of strain BFS56 (left), asnH-lacZ in cells of strain BFS55 (middle), and asnO-lacZ in cells of FU339 (right) are shown as solid squares, and that of endogenous lacZ in cells of strain 168 are shown as open squares. The OD600 for cells is shown as solid and open circles for each of the mutants and the wild type, respectively. (B) Northern analysis. Results of Northern analyses for the asnB (left), asnH (middle), and asnO (right) transcriptions are shown. RNAs were prepared from cells of strain 168 grown in DSM. The RNA samples were taken during exponential growth (lane 1), at the time of transition between exponential growth and stationary phase (lane 2), and at 1 h (lane 3) and 5 h (lane 4) after the beginning of sporulation. Positions of size marker RNAs (Millennium markers; Ambion) are given on the left of each panel. Positions of major transcripts are indicated with arrows.

Complementation of E. coli asparagine auxotrophy by B. subtilis asparagine synthetase genes. It was reported that asnB of E. coli was not able to be cloned into a high-copy-number plasmid in E. coli (18), implying that the same problem might occur in cloning the asparagine synthetase homolog genes of B. subtilis in E. coli. Indeed, asnB of B. subtilis could not be cloned into plasmid pUC19 (data not shown). For cloning of the three genes, therefore, we used a low-copy-number plasmid vector, pOU71, which replicates in E. coli cells at one copy per chromosome (11). A DNA fragment containing the entire asnB coding region as well as its 5' flanking region, probably carrying a promoter (Fig. 1), was cloned into plasmid pOU71 to provide plasmid pASNB. Similarly a DNA fragment containing asnO (Fig. 1) was cloned to provide plasmid pASNO. In the case of asnH cloning, two DNA fragments, which covered only (i) asnH and (ii) the four genes from yxbB to asnH and the 5' flanking region of yxbB (Fig. 1), were cloned, to provide plasmids pASNH and pYXBB, respectively, because the Northern analyses described above suggested that a promoter for the operon containing asnH might be located upstream of yxbB.

View larger version (15K): [in this window] [in a new window]   FIG. 4.   Complementation of the asparagine-auxotrophic phenotype of E. coli ME6279 by expression of asparagine synthetase genes of B. subtilis. E. coli ME6279 cells harboring plasmid pASNB (solid triangles), pASNH (solid squares), pASNO (open circles), pOU71 (solid diamonds), or pYXBB (open diamonds) were precultured in LB with asparagine (50 µg/ml) for 16 h at 37°C with shaking. The cells in each of the precultures were washed once and then inoculated into M9 medium with (left) and without (right) asparagine (50 µg/ml) and allowed to grow at 37°C with shaking. The OD600s for cells were monitored at 1-h intervals. Data from a single experiment are presented. The same experiments were repeated at least three times with similar results.

Deletion analysis of the asparagine synthetase genes. To investigate whether the three genes were involved in asparagine biosynthesis in B. subtilis, the coding regions of three asn genes were deleted via double-crossover events, resulting in marker replacement. A series of mutant strains lacking one, two, or all three genes was constructed (see Materials and Methods). As summarized in Table 2, the resulting strains were cultured in S6 liquid minimal medium to compare their growth rates in the absence and presence of low (50 µg/ml) and high (5 mg/ml) levels of asparagine. It is known that some amino acid-auxotrophic phenotypes (for example, tryptophan- and methionine-auxotrophic phenotypes) can be restored by adding the required amino acid at the low level of 50 µg/ml. Among strains lacking one of the three genes, strains FU341 and FU342, lacking asnH and asnO, respectively, grew as well as the wild-type strain even without asparagine, while strain FU340, lacking asnB, grew more slowly than the wild type. Among strains lacking two of the three genes, strain FU345, lacking both of asnH and asnO, grew as well as the wild type, while strains FU343 (lacking asnB and asnH) and FU344 (lacking asnB and asnO) grew more slowly. In addition, the slow growth of strain FU340 was not well restored even in the presence of the high level of asparagine, and strain FU343 exhibited a tendency to grow more slowly than FU340 in the presence of asparagine. Finally, the growth of strain FU346, lacking all three genes, was even slower, in particular when asparagine was present in the medium.

View larger version (83K): [in this window] [in a new window]   FIG. 5.   Effect on growth of inactivation of asnB, asnH, asnO, and ytnA in B. subtilis. A single colony each of strains 168 (wild type [WT]), FU340 (B), FU341 (H), FU342 (O), FU343 (BH), FU344 (BO), FU345 (HO), and FU346 (BHO) was taken from overnight cultures on TBABG plates containing appropriate antibiotics and spread onto S6 plates with (right) and without (left) asparagine (50 µg/ml). The plates were incubated at 37°C for 36 h, and then colony formation was observed. Data from a single experiment are presented. The same experiments were repeated three times independently with similar results.

View larger version (77K): [in this window] [in a new window]   FIG. 6.   Transcription of asnB, asnH, and asnO in cells of strain 168 grown in minimal medium. Results of Northern analysis of the asnB (left), asnH (middle), and asnO (right) transcriptions are shown. The same specific probes as for the previous experiments (Fig. 3B) were used. RNAs were prepared from cells of strain 168 grown in S6 minimal medium without (lanes 1 to 3) and with (lanes 4 to 6) 5 mg of asparagine per ml. The RNA samples were taken earlier (lanes 1 and 4) and later (lanes 2 and 5) in exponential growth and in stationary phase (lanes 3 and 6). Positions of size marker RNAs are given on the left of each panel. Positions of the detected major transcripts covering each of the entire transcriptional units are indicated with arrows.

The asnO gene is indispensable for sporulation. Since asnO is expressed only during sporulation (Fig. 3A), we examined whether its transcription depended on sporulation sigma factors. For this purpose RNA samples were prepared from sporulating cells harboring different sigma factor mutations (Table 1) and from the wild-type cells, at stages when the respective sigma factors were activated. The samples were then subjected to Northern analysis, targeting the asnO transcript (Fig. 7A). At the stage when sigma-H was activated, the asnO transcript was not detected even in the wild-type cells, indicating that asnO was not transcribed at the initiation of sporulation. At the stage when sigma-E and -F were activated, the transcript was detected in the wild-type cells but not in the sigma-F or -E mutants. At the stage when sigma-K and -G were activated, the transcript was detectable in both mutants and in the wild-type cells. The sigma-F mutant lacks both sigma-F and -E, but the sigma-E mutant must have active sigma-F. We therefore conclude that asnO transcription depends on sigma-E. In addition, primer extension analysis mapped a 5' end of the asnO transcript (Fig. 7B) and allowed us to identify a corresponding promoter sequence as 10 (CATAAACT [the 10 position is underlined]) and 35 (TCAACTA) regions, separated by 14-bp spacer. This promoter is likely recognized by sigma-E RNA polymerase (5).

View larger version (87K): [in this window] [in a new window]   FIG. 7.   Expression of asnO depends on sigma-E. (A) Northern analysis of asnO transcription in sporulating cells of B. subtilis strains. The same specific asnO probe as for the previous experiments (Figs. 3B and 6) was used. RNAs were prepared from cells of B. subtilis strains grown in DSM. The RNA samples were taken from cells of strains 168 and ASK201 (without sigma-H) at the beginning of sporulation (T0) (lanes 1 and 2, respectively); strains 168, ASK203 (without sigma-E), and ASK202 (without sigma-F) at 3 h after the beginning of sporulation (T3) (lanes 3, 4, and 5, respectively); and strains 168, ASK205 (without sigma-K), and ASK204 (without sigma-G) at 5 h after the beginning of sporulation (T5) (lanes 6, 7, and 8, respectively). Positions of size marker RNAs and the asnO transcript (arrow) are given on the left and right, respectively. (B) Primer extension mapping of a 5' end of the asnO transcript. The end-labeled primer (see Materials and Methods) was hybridized to RNA samples prepared from cells of strain 168 grown in DSM and was then extended with Moloney murine leukemia virus reverse transcriptase (Gibco BRL). The RNA samples were taken in exponential growth (lane 1), at the time of transition between exponential growth and stationary phase (lane 2), and at 1 h (lane 3) and 3 h (lane 4) after the beginning of sporulation. A known DNA sequence ladder (lanes G, A, T, and C) made by dideoxy sequencing reactions with the same end-labeled primer was loaded to estimate the size of the extended product. The position of the extended product is indicated by an arrow on the right. Part of the nucleotide sequence (noncoding strand) of the asnO promoter region is shown on the left. The 10 and 35 promoter regions are underlined, and the 5' end corresponding to the transcription start site (+1) is indicated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of the genomic sequence of B. subtilis allowed us to predict three asparagine synthetase gene homologs. These homologs were designated asnB, asnH, and asnO, encoding glutamine-dependent AsnB-type enzymes (Fig. 2). No ammonia-dependent AsnA-type enzyme gene has been detected within the genome. Enteric bacteria, such as E. coli and K. aerogenes, have the two types of asparagine synthetases, and ammonia is their optimal nitrogen source. In contrast, the fastest growth of B. subtilis occurs in medium containing glutamine as the sole nitrogen source (3). Ammonia assimilation in B. subtilis depends on successive reactions involving glutamine synthetase and glutamate synthase, as mutations in either enzyme result in an inability to grow with ammonia as a nitrogen source (2), and glutamate dehydrogenase functions only as a catabolic, not an assimilatory, enzyme in B. subtilis. This indicates that B. subtilis has no other efficient mechanism for ammonia assimilation. These facts might be related to the finding that B. subtilis has no ammonia-dependent AsnA-type enzyme but rather only three glutamine-dependent AsnB-type ones.

Each of the three asparagine synthetase homologs of B. subtilis likely functions as asparagine synthetase, as shown by their complementation of the asparagine-auxotrophic phenotype of E. coli ME6279 (Fig. 4). The analysis of deletion mutants of B. subtilis indicated that asnB is the main gene involved in asparagine biosynthesis in vegetative cells (Table 2 and Fig. 5). It may be relevant that pASNB, carrying asnB, complemented the E. coli auxotrophy most efficiently (Fig. 4). We have no explanation for the fact that the growth rates of B. subtilis mutants lacking asnB fluctuated to a great extent (Table 2). In addition, even a high level of asparagine (5 mg/ml) did not restore the growth of asnB mutants completely (Table 2), and a saturating concentration of asparagine (25 mg/ml) could not either (data not shown). Furthermore, even in the presence of abundant asparagine in the medium, asnB was expressed as constitutively as in the absence of asparagine (Fig. 6). It is possible that not only asparagine itself but also the function of the major asparagine synthetase of the asnB product might be needed to support normal growth in the minimal medium.

The asnB and asnH genes are part of possible longer operons and are followed by ytnA and yxaM, respectively (Fig. 1). These downstream genes could have been inactivated by a possible polar effect of the marker replacements in the asnB and asnH mutants. That possibility was not investigated for yxaM, since there was no discernible phenotype associated with the insertion in the asnH gene. Inactivation of ytnA had no significant effect when asnB was functional (Table 2). However, it is interesting that both ytnA and yxaM are homologous to some of the known amino acid transporter genes (10). If the former was involved in asparagine uptake, the putative polar effect of asnB inactivation might be one of the reasons for the lack of growth restoration by asparagine.

In B. subtilis cells grown in the rich medium, asnH was induced at the transition between exponential growth and stationary phase (Fig. 3). In the minimal medium, not only asnB but also asnH was expressed constitutively, but when the medium was supplemented with asparagine, the expression of asnH was partially repressed (Fig. 6). These results imply that asnH may be repressed by rich nutrients and partially repressed by asparagine. Moreover, in cells grown in the rich medium, the amount of asnH transcript was decreased early in sporulation and then increased again later, suggesting that its regulation is complex. Although homology searches did not suggest any functions for yxbB, yxbA, and yxnB, genes which may be transcribed together with asnH, the three genes might be involved in asnH regulation. However, the expression of asnH alone did not support normal growth in minimal medium without asparagine (Table 2 and Fig. 5). Recently, it was observed that another asnH mutant strain, ASNHd, which was constructed within the framework of the international B. subtilis genome functional analysis project, gave abnormal colonies on high-salt plates and a penicillin-sensitive phenotype (9). The asnH gene might not be as efficient in asparagine biosynthesis as asnB but might be responsible for some function involved in cell surface organization.

The asnO gene was indispensable for sporulation (Table 3). Its transcription depends on sigma-E (Fig. 7), suggesting that the gene was expressed in sporulating mother cells, and the failure of sporulation of the asnO mutants was not restored by adding asparagine to the medium (Table 3). We cannot eliminate the possibility that sporulating mother cells were less efficient in asparagine uptake, but it is also possible that asnO might have an unknown specific role in sporulation besides asparagine biosynthesis.

Finally, the deletion analysis suggested that B. subtilis could have another, minor asparagine biosynthesis pathway independent of the three asparagine synthetases. Because it was impossible to predict candidates for asparagine synthetase genes other than the three described in this paper, the minor biosynthesis might be supported by a novel and unique asparagine synthetase or by some amidotransferases with broad specificity. Further work is required to test these possibilities.