Expression of the glycolytic gapA operon in Bacillus subtilis: differential syntheses of proteins encoded by the operon

Christoph Meinken, Hans-Matti Blencke, Holger Ludwig and Jörg Stülke

Lehrstuhl für Mikrobiologie, Institut für Mikrobiologie, Biochemie und Genetik der Friedrich-Alexander-Universität Erlangen-Nürnberg, Staudtstrasse 5, D-91058 Erlangen, Germany

Correspondence
Jörg Stülke
jstuelke{at}biologie.uni-erlangen.de


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Glycolysis is one of the central routes of carbon catabolism in Bacillus subtilis. Several glycolytic enzymes, including the key enzyme glyceraldehyde-3-phosphate dehydrogenase, are encoded in the hexacistronic gapA operon. Expression of this operon is induced by a variety of sugars and amino acids. Under non-inducing conditions, expression is repressed by the CggR repressor protein, the product of the promoter-proximal gene of the operon. Here, it is shown that the amount of glyceraldehyde-3-phosphate dehydrogenase encoded by the second gene of the operon exceeds that of the CggR repressor by about 100-fold. This differential synthesis was attributed to an mRNA processing event that takes place at the 3' end of the cggR open reading frame and to differential segmental stabilities of the resulting cleavage products. The mRNA specifying the truncated cggR gene is quickly degraded, whereas the downstream processing products encompassing gapA are quite stable. This increased stability is conferred by the presence of a stem–loop structure at the 5' end of the processed mRNAs. Mutations were introduced in the region of the cleavage site. A mutation affecting the stability of the stem–loop structure immediately downstream of the processing site had two effects. First, the steady-state transcript pattern was drastically shifted towards the primary transcripts; second, the stability of the processed mRNA containing the destabilized stem–loop structure was strongly decreased. This results in a reduction of the amount of glyceraldehyde-3-phosphate dehydrogenase in the cell. It is concluded that mRNA processing is involved in differential syntheses of the proteins encoded by the gapA operon.


Abbreviations: CAA, Casamino acids; His6, hexahistidine


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
In bacteria, gene expression is most often regulated at the level of transcription. Repressors and activators control the rate of transcription initiation, whereas transcript elongation can be controlled by sophisticated termination/antitermination mechanisms involving either RNA–RNA or protein–RNA interactions (for reviews see Wagner, 2000; Henkin, 2000). This is due to the short half-life of bacterial mRNAs which allows rapid changes in protein synthesis in response to changing environmental conditions. Indeed, recent microarray analyses revealed that about 80 % of all Escherichia coli mRNAs have half-lives of between 3 and 8 min (Bernstein et al., 2002).

Usually, proteins that are involved in a common physiological function are encoded in operons in bacteria. This allows co-ordinated expression of all enzymes of a metabolic pathway or of all proteins for a cellular function. However, some operons encode proteins that are needed in different amounts. Several strategies have evolved to allow the differential synthesis of proteins encoded in one operon. First, translation efficiency may vary among the genes of an operon due to ribosome-binding sites that differ in their match to the 16S rRNA or due to different codon usage strategies (Moszer et al., 1999; Dobrindt et al., 2002; Vellanoweth, 1993). Second, translation of an individual gene of an operon may be inhibited by a specific antisense RNA, as has recently been demonstrated for the E. coli galactose operon (Møller et al., 2002). Third, the primary transcript may be processed to give rise to final mRNAs with different segmental stability which accounts for differential protein synthesis. This was first described for the Rhodobacter capsulatus puf photosynthesis operon and subsequently for the malEFG operon in E. coli, the pap operon in uropathogenic E. coli, the Zymomonas mobilis gap–pgk operon and the Bacillus subtilis dnaK operon (Belasco et al., 1985; Klug et al., 1987; Newbury et al., 1987; Båga et al., 1988; Nilsson et al., 1996; Eddy et al., 1991; Burchhardt et al., 1993; Heck et al., 2000; Homuth et al., 1999).

We are interested in the control of glycolysis in B. subtilis. Transcriptional analyses revealed that genes encoding enzymes that perform irreversible steps are induced by glucose and other sugars while those enzymes that are required for both glycolysis and gluconeogenesis are synthesized both in the presence and absence of sugars (Ludwig et al., 2001). The regulatory mechanisms have been elucidated for the first step of glycolysis, glucose transport and concomitant phosphorylation, and also for the expression of glyceraldehyde-3-phosphate dehydrogenase. Glucose induction of the ptsGHI operon encoding the enzymes of the glucose : phosphoenolpyruvate phosphotransferase system (PTS) is mediated by the transcriptional antiterminator GlcT (Stülke et al., 1997; Langbein et al., 1999). The gapA gene encoding glyceraldehyde-3-phosphate dehydrogenase is the second gene of a hexacistronic operon. Expression of the gapA operon is induced by glucose and other sugars that are catabolized via glycolysis or the pentose phosphate pathway. Induction depends on the pleiotropic regulator of carbon metabolism of B. subtilis, CcpA, and on the specific repressor CggR (Tobisch et al., 1999; Fillinger et al., 2000; Ludwig et al., 2001). While CcpA is required to allow transport of PTS sugars and subsequent formation of the intracellular inducer, CggR seems to repress transcription of the gapA operon in the absence of this inducer. The nature of the inducer has not yet been identified; however, it seems to be a glycolytic intermediate, probably a triose phosphate (Ludwig et al., 2002). CggR is encoded by the first gene of the gapA operon (see Fig. 1a). However, while GapA is one of the prominent proteins in cells of B. subtilis growing in rich media, CggR has not been identified on two-dimensional gels (Tobisch et al., 1999; Büttner et al., 2001). This may be due to an endonucleolytic mRNA processing event that takes place at the very end of the cggR open reading frame (ORF). The resulting mRNAs exhibit different segmental stabilities: while the truncated cggR transcript is extremely unstable, the monocistronic gapA and the pentacistronic gapA pgk tpi pgm eno mRNAs have half-lives of about 3–4 min (Ludwig et al., 2001).



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 1. Potential RNA secondary structures for the mRNA processing site of the B. subtilis gapA operon in the wild-type and different mutants strains as calculated by the Zuker algorithm. (a) Genetic map of the hexacistronic gapA operon and the predicted secondary structures for the wild-type mRNA processing site. The mRNA cleavage site is indicated by an arrow. The cggR stop codon is indicated by the frame. The calculated {Delta}G° energy values of the upstream stem–loop and the downstream stem–loop amount to -12·5 kcal mol-1 and -9·9 kcal mol-1, respectively. (b) Predicted structure for the mRNA processing site in the gapA-P1 mutant strain GP552. This allele carries mutations of nucleotides at positions -34 (C->U) and -43 (C->G) relative to the cggR stop codon. The {Delta}G° energy value of the upstream stem–loop was calculated to be -11·3 kcal mol-1. (c) Predicted structure for the mRNA processing site in the gapA-P2 mutant strain GP554. This allele carries mutations of nucleotides at positions +3 (C->G) and +4 (C->G) relative to the cggR stop codon. The calculated {Delta}G° energy value of the downstream stem–loop if present at all amounts to -1 kcal mol-1.

 
The mechanism by which the processing of the primary transcripts of gapA and other operons occurs in B. subtilis is currently unknown. While RNase E is the main factor of endonucleolytic mRNA processing in proteobacteria such as E. coli and R. capsulatus (Mudd et al., 1988; Nilsson et al., 1996; Braun et al., 1998; Jäger et al., 2001), this enzyme is not present in B. subtilis and other Gram-positive bacteria with a low G+C content. However, E. coli RNase E is active on mRNAs that are subject to processing in B. subtilis, as has been shown for the thrS leader mRNA. It was therefore proposed that there is an enzymic activity similar to RNase E in B. subtilis (Condon et al., 1997). Studies of ribonucleolytic activities in B. subtilis identified RNase III, RNase M5 and RNase P. While the former enzymes are active on double-stranded RNA substrates, RNase P cleaves single-stranded RNA immediately adjacent to double-stranded regions (Wang & Bechhofer, 1997; Hansen et al., 2001; Condon et al., 2002). Since cleavage of the gapA operon primary transcripts occurs in an AU-rich single-stranded region surrounded by potential stem–loop structures it is tempting to speculate that the postulated RNase E-like enzyme might be required for processing of the two gapA operon primary mRNAs.

In this study, we analysed the role of the stem–loop regions surrounding the cleavage site by site-directed mutagenesis. One mutant resulted in a drastic reduction of processing efficiency and in accumulation of primary transcripts. Interestingly, this mutation, which simultaneously changed the secondary structure at the 5' end of the stable processing product, had a strong impact on the stability of the final transcripts. mRNA cleavage and different stability of the final processing products are major factors accounting for a more than a 100-fold difference in protein amounts between the CggR repressor protein and glyceraldehyde-3-phosphate dehydrogenase, the key enzyme of glycolysis, even though both proteins are encoded in the same operon in B. subtilis. The mutation that destabilized the messages containing the gapA gene also resulted in a reduced synthesis of the gene product glyceraldehyde-3-phosphate dehydrogenase. Most probably, mRNA processing serves to generate stable mRNAs required for glycolytic enzyme synthesis from a very unstable precursor.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and growth conditions.
The B. subtilis strains used in this study are listed in Table 1. E. coli DH5{alpha} (Sambrook et al., 1989) was used for cloning experiments and protein expression. B. subtilis was grown in CSE minimal medium supplemented with auxotrophic requirements (at 50 mg l-1) (Faires et al., 1999). Carbon and nitrogen sources were added as indicated. E. coli was grown in Luria–Bertani (LB) medium and transformants were selected on plates containing ampicillin (100 µg ml-1). LB and SP plates were prepared by the addition of 17 g Bacto agar l-1 (Difco) to the medium.


View this table:
[in this window]
[in a new window]
 
Table 1. B. subtilis strains used in this study

 
DNA manipulation.
Transformation of E. coli and plasmid DNA extraction were performed using standard procedures (Sambrook et al., 1989). Restriction enzymes, T4 DNA ligase and DNA polymerases were used as recommended by the manufacturers. DNA fragments were purified from agarose gels using the Nucleospin extract kit (Macherey & Nagel). Pfu DNA polymerase was used for the PCR as recommended by the manufacturer. DNA sequences were determined using the dideoxy chain termination method (Sambrook et al., 1989). Chromosomal DNA of B. subtilis was isolated as described by Kunst & Rapoport (1995).

Transformation and characterization of the phenotype.
B. subtilis was transformed with chromosomal DNA according to the two-step protocol (Kunst & Rapoport, 1995). Transformants were selected on SP plates containing spectinomycin (Spc; 100 µg ml-1), kanamycin (Km; 10 µg ml-1) or chloramphenicol (5 µg ml-1). Quantitative assays of lacZ expression in B. subtilis were performed with cell extracts using ONPG as the substrate (Kunst & Rapoport, 1995).

Construction of B. subtilis strains with mutations of the mRNA processing site.
First, we constructed plasmid pGP231 containing the cggR gapA chromosomal region. For this purpose, the region was amplified using the primers CM1 (5'-CCTATACATTTTGGATCTTTGC) and CM2 (5'-GTTAAGTACTTTTGCAAACGGCG), digested with HindIII and EcoRV and cloned into pBluescript SKII+ (Stratagene) cut with the same enzymes.

Mutations of the processing site of the gapA operon were introduced by a modified PCR protocol, the combined chain reaction (CCR) (Bi & Stambrook, 1998). Primers HMB5 (5'-CGATCAATCCCCATGGGTC) and CM3 (5'-GATCGCGTTCTGCAGAAA) were used as outer primers. The gapA-P1 and gapA-P2 mutations were present on the oligonucleotides CM4 (5'-CGGTTCTGGTCACAGATGAAGGAG) and CM6 (5'-GGATGAATAATCGGTCAATATAAATAT), respectively. These primers were phosphorylated and allowed ligation of the nascent elongation product initiated from HMB5. The resulting CCR products carrying the mutations were cut with NcoI and PstI and cloned into pGP231 digested with the same enzymes. The resulting plasmids were pGP232 and pGP234.

To transfer the mutations from the plasmids to the B. subtilis chromosome, we made use of strain GP550. This strain was constructed by exchanging the kanamycin-resistance gene linked to the lacZ fusion of GP311 by the chloramphenicol-resistance gene of pAC5 (Martin-Verstraete et al., 1992). Strain GP550 carries a deletion of the cggR gene. Transformation of GP550 with chromosomal DNA of GP311 (cggR–lacZ aphA3) resulted in the formation of kanamycin-resistant colonies exhibiting a dark-blue colour on CSE plates containing X-Gal in the absence of glucose. Simultaneous transformation of GP550 with both chromosomal DNA of GP311 carrying the cggR–lacZ fusion and plasmids containing a functional cggR gene gave rise to white colonies on the same plates at low frequency. These white colonies result from a double homologous recombination at the cggR locus and in the restoration of a functional chromosomal cggR gene. The plasmids pGP232 and pGP234 allowed us to introduce the mutations of the processing site concomitantly with the functional cggR gene. The resulting strains were B. subtilis GP552 (gapA-P1) and GP554 (gapA-P2). The presence of the mutations was verified by sequencing of the chromosomal cggR–gapA region.

Northern blot analysis.
To isolate high-quality RNA suited for the detection of long transcripts and their precursors, RNA was prepared by the modified ‘mechanical disruption protocol’ described by Hauser et al. (1998). The cells were harvested at the exponential phase. For RNA preparation, 25 ml of cells were used. After mechanical cell disruption, the frozen powder was immediately resuspended in 3 ml lysis buffer [4 M guanidine isothiocyanate; 0·025 M sodium acetate, pH 5·3; 0·5 % (w/v) N-laurylsarcosine]. Subsequently, total RNA extraction with acid phenol solution and Northern blot analysis were carried out as described by Homuth et al. (1997). For the analysis of mRNA stability, rifampicin was added to exponentially growing cultures (final concentration 100 µg ml-1) and samples were taken at the time points indicated. Digoxigenin RNA probes specific for cggR, gapA and pgk (Ludwig et al., 2001) were used to detect the corresponding transcripts according to the instructions of the manufacturer (DIG RNA labelling kit and detection chemicals; Roche Diagnostics). The sizes of the RNA molecular size markers (Gibco-BRL) were 9·49, 7·46, 4·40, 2·37, 1·35 and 0·24 kb. Quantitative evaluation of mRNA amounts was performed using the TINA software (version 2.0).

Overexpression and purification of CggR.
To overexpress CggR and GapA fused to a hexahistidine (His6) sequence at the N terminus, plasmids pGP705 and pGP704, respectively, were constructed as follows. DNA fragments corresponding to the cggR and gapA ORFs were amplified by PCR using chromosomal DNA of B. subtilis 168 and the primer pairs HMB1 (5'-ACGCGTCGACATGAACCAGTTAATACAAGCTC)/HMB2 (5'-CCCAAGCTTTCATTATTCATCCCTTAATAACTTCTT) and HMB3 (5'-ACGCGTCGACATGGCAGTAAAAGTCGGTATTA)/HMB4 (5'-CCCAAGCTTTCATTAAAGACCTTTTTTTGCGATGT). The PCR products were digested with SalI and HindIII, and the resulting fragments were cloned into the expression vector pWH844 (Schirmer et al., 1997) cut with the same enzymes.

For overexpression of the recombinant proteins, E. coli DH5{alpha} was transformed with plasmids pGP704 or pGP705 and cultivated in LB medium. Expression was induced by the addition of IPTG (final concentration 1 mM) to exponentially growing cultures (OD600 of 0·8). The crude extracts were passed over a Ni2+ HiTrap chelating column (Pharmacia) followed by elution with an imidazole gradient. The Bio-Rad dye-binding assay was used to determine protein concentration. BSA was used as the standard.

Western blot analysis.
Purified His6–GapA and His6–CggR were used to generate rabbit polyclonal antibodies (Eurogentec). For Western blot analysis, B. subtilis cell extracts were separated on 12·5 % SDS-PAGE gels. After electrophoresis, the proteins were transferred to a PVDF membrane (Bio-Rad) by electroblotting. GapA and CggR were detected with polyclonal antibodies. Antibodies were visualized by using anti-rabbit IgG-AP secondary antibodies (Chemikon International) and the CDP* detection system (Roche Diagnostics). For the determination of the cellular levels of CggR and GapA, the procedure described by Homuth et al. (1999) was used. Crude extracts were prepared from an exponentially growing culture and simultaneously the number of cells was determined by plate counting. Cells were disrupted by using a French press [20 000 p.s.i. (1·378x108 Pa)]. Defined amounts of purified proteins served as an internal calibration standard in Western blots to determine the cellular levels of CggR and GapA. For the analysis of in vivo protein stability, transcription and translation were stopped by adding rifampicin (40 µg ml-1) and erythromycin (300 µg ml-1) to growing cultures. Samples were taken at the time points indicated. Quantitative evaluation of protein amounts was performed using the TINA software (version 2.0).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Construction and analysis of strains containing point mutations in potential stem–loop structures surrounding the mRNA processing site
The processing site of the 7·2 and 2·2 kb primary transcripts of the gapA operon was mapped to a site in the very 3' end of the cggR ORF that is surrounded by two potential stem–loop structures (Ludwig et al., 2001) (see Fig. 1). In E. coli, such stem–loop structures are thought to be involved in endonucleolytic cleavage by the RNase E (Rauhut & Klug, 1999; Grunberg-Manago, 1999). Since the requirements for RNase recognition and cleavage have not yet been studied in B. subtilis we wished first to address the role of the two stem–loop structures. We chose to introduce mutations to destabilize the structures. However, since the first stem is part of the cggR coding region, we had to introduce a mutation that did not interfere with the CggR amino acid sequence. The position of the two mutations of the processing site, designated gapA-P1 and gapA-P2, and their influence on the potential RNA structure as deduced from the algorithm of Zuker (1989) is shown in Fig. 1. The stem upstream of the cleavage site has a free energy of 12·5 kcal mol-1 (52·3 kJ mol-1) in the wild-type strain. The stability of this structure is only weakly affected by the gapA-P1 mutation [{Delta}G°=-11·3 kcal mol-1 (-47·3 kJ mol-1)] (Fig. 1b). In contrast, the effect of the gapA-P2 mutation on the stability of the second stem is rather large [{Delta}G°=-9·9 kcal mol-1 (-41·2 kJ mol-1) in the wild-type vs -1 kcal mol-1 (-4·184 kJ mol-1) for the mutant structure] (Fig. 1c). Therefore, it seems reasonable to expect that the gapA-P2 mutation will have a significant impact on the processing event and/or the stabilities of the gapA operon mRNAs.

The mutations were introduced as described in Methods. Briefly, the {Delta}cggR mutant strain GP550 was simultaneously transformed with chromosomal DNA of a strain carrying a cggR–lacZ fusion and a plasmid carrying the cggR gapA region with the desired point mutation. Selection for transformants carrying the cggR–lacZ fusion resulted in a large number of colonies that expressed the fusion in the absence of glucose due to the cggR deletion in the recipient strain. However, few transformants had experienced recombination of the plasmid as well and formed white colonies on X-Gal plates in the absence of glucose. These bacteria had acquired both the cggR gene and the mutations of the processing site.

To test the effect of the mutations of the processing site on the steady-state mRNA levels, we performed Northern blot analyses with RNA isolated from cells grown in CSE minimal medium in the presence of glucose and casein hydrolysate to induce expression of the gapA operon. With a probe specific for cggR, we detected two main transcripts in both the wild-type strain GP303 and the gapA-P1 mutant strain GP552. These were the 2·2 kb primary transcript and the truncated 1·0 kb cggR processing product. In addition, the 7·2 kb hexacistronic full-length transcript was visible as a faint band (Fig. 2b, see Fig. 2a for an interpretation of the transcripts). These findings were in good agreement with previous observations (Ludwig et al., 2001). In the gapA-P2 mutant strain GP554, however, the amounts of the 7·2 and 2·2 kb primary transcripts were strongly increased as compared to the wild-type strain (Fig. 2b). Since the ratio of the mRNA species was shifted towards the primary transcripts, the gapA-P2 mutation might indeed affect the processing of the primary mRNAs of the gapA operon. This idea was supported by Northern blot analyses with the gapA probe. Again, the transcript patterns were very similar for the wild-type and the gapA-P1 strains: the two major mRNA species were the pentacistronic 6·2 kb and the monocistronic 1·2 kb processing products. In addition, the 2·2 kb primary transcript was present in these strains. In contrast, major changes in the transcript pattern were observed for the gapA-P2 mutant strain GP554. As observed with the cggR probe, the amount of the primary transcripts was increased. Moreover, the amounts of the two processing products that were the most prominent mRNAs in the wild-type strain were drastically reduced in strain GP554 (Fig. 2b). The sizes of the processed transcripts are similar in the wild-type and gapA-P2 mutant strains. We can, however, not exclude the possibility that the processing event occurs at a position different from the site mapped in the wild-type strain (Ludwig et al., 2001). The increased amounts of the primary transcripts in GP554 might result from increased transcription initiation at the promoter of the gapA operon or from changes in mRNA processing and stability. To distinguish between these possibilities, we determined the effect of the mutations on the expression of a cggRlacZ fusion. Similar expression patterns in all three strains and only slightly increased amounts of {beta}-galactosidase in GP554 were observed (data not shown). Moreover, the amount of CggR, the product of the promoter-proximal gene of the gapA operon, was similar in the wild-type and gapA-P2 mutant strains (see below, Fig. 6a). We may, therefore, conclude that the gapA-P2 mutation affecting the stem structure immediately downstream of the mRNA processing site interferes with proper processing of the primary transcripts.



View larger version (39K):
[in this window]
[in a new window]
 
Fig. 2. The effect of mutations of the mRNA processing site on the steady-state mRNA levels encoded by the gapA operon in B. subtilis. (a) Transcriptional organization of the glycolytic gapA operon. The lengths of the various transcripts deduced from Northern blot analysis are indicated, and the thickness of the arrows represents their relative abundance within the cells (Ludwig et al., 2001). (b) Northern blot analysis of the gapA operon using probes specific for cggR and gapA. RNA was isolated from B. subtilis GP303 (wild-type) and from the mutant strains GP552 (gapA-P1) and GP554 (gapA-P2) grown in CSE minimal medium with 0·5 % glucose and 0·1 % CAA. Total RNA was isolated and separated by electrophoresis in a 0·9 % agarose gel and after blotting, nylon membranes were hybridized to riboprobes specific for cggR and gapA. Note that the probes cross-hybridized with the 16S and 23S rRNAs. The sizes of 16S rRNA and 23S rRNA are indicated by arrows to the left. Five micrograms of RNA were applied per lane.

 


View larger version (31K):
[in this window]
[in a new window]
 
Fig. 6. Quantification of intracellular amounts of CggR and GapA. To quantify the amounts of CggR (a) and GapA (b), different concentrations of purified His6 proteins and 5 µg of crude extracts of B. subtilis GP303 (wild-type) and the gapA-P2 mutant strain GP554 were separated on 12·5 % SDS-PAGE gels and transferred to PVDF membranes. CggR and GapA were detected by rabbit polyclonal antibodies.

 
Effects of the gapA-P2 mutation of the mRNA processing site on transcript stability
The primary transcripts of the gapA operon have a short half-life of less than 2 min. Similarly, the 1·0 kb processing product corresponding to the truncated cggR gene is highly unstable (Ludwig et al., 2001). We asked whether there is an altered stability of the gapA operon primary transcripts as a consequence of the gapA-P2 mutation. To address this question, the decay of the mRNAs in the wild-type and the gapA-P2 mutant strains was compared directly. RNA was isolated from the bacteria after treatment with rifampicin to inhibit transcription initiation.

Using the cggR probe, we observed a rapid decay of all three transcripts as described previously (Fig. 3a). In the wild-type strain GP303, both primary transcripts exhibited half-lives of about 1 min (Fig. 3c). The truncated cggR processing product had a half-life of 1·5 min. In the gapA-P2 mutant strain GP554, the steady-state level of the primary transcripts was increased. However, as observed with the wild-type, the primary transcripts disappeared rapidly, and the half-lives of the transcripts were unaffected by the mutation (see Fig. 3c). Thus, the gapA-P2 mutation of the processing site has no detectable effect on the stability of the primary transcripts, whereas it exhibits a major effect on the stability of the processed transcript (see below).



View larger version (31K):
[in this window]
[in a new window]
 
Fig. 3. Effect of the gapA-P2 mutation on the stability of the gapA operon mRNAs. RNA was prepared from B. subtilis GP303 (wild-type) and the isogenic mutant strain GP554 (gapA-P2) grown in CSE minimal medium with 0·5 % glucose and 0·1 % CAA before (0) and at indicated times after the addition of rifampicin (final concentration 100 µg ml-1). After electrophoresis and blotting, the nylon membranes were hybridized to riboprobes specific for cggR (a) and pgk (b). For strain GP303, 10 µg RNA were applied per lane, whereas only 5 µg RNA were loaded for the mutant GP554. The relative amounts of the 7·2 kb primary transcript (c) and of the 6·2 kb processing product (d) in the wild-type strain GP303 ({circ}) and the gapA-P2 mutant strain GP554 ({bullet}) were plotted against the time to determine mRNA half-lives.

 
A probe specific for pgk encoding the third gene of the operon was used for two reasons. First, this probe detects the hexacistronic primary transcript as well as the pentacistronic 6·2 kb message. Second, a tetracistronic pgk tpi pgm eno transcript is initiated upstream of pgk and served as an internal control (Ludwig et al., 2001; see Fig. 3b). As observed with the cggR probe, the 7·2 kb primary transcript was much more abundant in the gapA-P2 mutant strain, but it disappeared rapidly in both the wild-type and mutant strain (Fig. 3c). As observed with the gapA probe, the amount of the promoter-distal processing product (6·2 kb) was greatly reduced in the gapA-P2 mutant strain GP554. Moreover, this mRNA was quickly degraded in the gapA-P2 mutant strain, whereas it was quite stable in the wild-type (half-life>5 min). As determined for the primary transcripts, the half-life of this 6·2 kb mRNA was found to be 1 min in GP554 (see Fig. 3b, 3d). This finding indicates that the altered structure at the 5' end of the pentacistronic processing product results in faster decay of the processed transcript (see Discussion). In contrast, the half-life of the tetracistronic 4·7 kb mRNA (about 4·5 min) was not significantly affected by the gapA-P2 mutation (see Fig. 3b).

Induction profile of CggR and glyceraldehyde-3-phosphate dehydrogenase
Although encoded by the same operon, the CggR repressor and the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase are needed by the cell under very different conditions: while CggR has to repress expression of the gapA operon in the absence of glucose, glyceraldehyde-3-phosphate dehydrogenase (GapA) is needed as a key enzyme of glycolysis if the sugar is present. To elucidate this obvious contradiction, we analysed the cellular amounts of the CggR and GapA proteins. To get the tools for this purpose, the cggR and gapA genes were fused to a sequence encoding an N-terminal His6 sequence. The resulting fusion proteins were expressed, purified and used to generate polyclonal antibodies as described in Methods.

The induction pattern of the two proteins was determined after growth of the bacteria in CSE minimal medium in the presence or absence of glucose and casein hydrolysate as synergistic inducers of the gapA operon. Expression of the proteins was assayed by Western blot analysis. As shown in Fig. 4, both proteins were barely detectable in wild-type bacteria grown under non-inducing conditions. However, strong signals were observed upon induction. This finding is in good agreement with the previous observation of glucose-inducible synthesis of GapA (Tobisch et al., 1999) and with the regulation of the promoter of the gapA operon (Ludwig et al., 2001). In addition, we assayed the cellular level of GapA and CggR in a cggR deletion strain (GP311). While GapA expression was constitutive in this strain, there was no CggR present. These results confirm that CggR is the factor that prevents expression of the gapA operon in the absence of glucose, and demonstrate that the cggR deletion in GP311 has no polar effect on gapA expression.



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 4. Induction profiles of CggR and GapA. Crude extracts were prepared from B. subtilis GP303 (wild-type) and from the {Delta}cggR mutant strain GP311 grown in CSE minimal medium with (+) or without (-) 0·5 % glucose and 0·1 % CAA. After electrophoresis in a 12·5 % SDS-PAGE gel and transfer onto a PVDF membrane, CggR and GapA were detected by rabbit polyclonal antibodies raised against the respective proteins from B. subtilis. Five micrograms of protein extract were applied per lane.

 
The mRNA processing site of the primary transcripts was mapped to the 3' end of the cggR ORF resulting in a truncated mRNA that does not possess a stop codon. Such truncated mRNAs pose a problem to the translation machinery and are often resolved by trans-translation: the tmRNA adds a tag to the truncated protein and allows ribosome rescue. The tagged proteins are then degraded. In B. subtilis, the ATP-dependent ClpX/ClpP protease is thought to degrade such marked proteins (Gillet & Felden, 2001; Wiegert & Schumann, 2001). Therefore, we asked whether CggR might be a target of tmRNA-mediated tagging and degradation. First, we analysed the stability of the CggR protein. For this purpose, B. subtilis 168 was grown in CSE medium containing glucose and casein hydrolysate. Transcription and translation were stopped by the addition of rifampicin and erythromycin, respectively, and protein extracts were prepared and subjected to Western blot analysis (Fig. 5). The amount of CggR in the cells remained constant for at least 40 min, suggesting that CggR is a stable protein. To test the effect of the components of the tagging and degradation system, we studied the stability of CggR in ssrA and clpP mutant strains, which are defective in tmRNA and the protease component of the ClpXP protease, respectively. In both strains, the amounts and stability of CggR were not changed as compared to the wild-type strain (data not shown). This mechanism of trans-translation may therefore not be involved or not play a major role in CggR translation.



View larger version (10K):
[in this window]
[in a new window]
 
Fig. 5. Stability of CggR in vivo. Protein extracts were prepared from B. subtilis 168 grown in CSE minimal medium with 0·5 % glucose and 0·1 % CAA in the absence (0) and at different times (5, 10, 20, 40 and 60 min) after the addition of erythromycin and rifampicin (final concentrations 100 µg ml-1 and 40 µg ml-1, respectively). CggR was detected using CggR-specific antibodies as described in the legend to Fig. 4. Five micrograms of protein extract were applied per lane.

 
Effect of mRNA processing on the differential intracellular amounts of CggR and GapA
As shown above, CggR and GapA syntheses are induced simultaneously as expected for proteins encoded in one operon. However, one can assume that a small amount of CggR may be sufficient for efficient repression of the gapA operon, whereas large quantities of GapA are certainly required for glycolysis. To test this hypothesis, we determined the intracellular ratio of the two proteins and the influence of mRNA processing on their relative amounts. To do this, we prepared protein extracts from exponentially growing cells of the wild-type strain B. subtilis GP303 and the isogenic processing site mutant strain GP554 and performed a quantitative Western blot analysis (Fig. 6). In the wild-type strain, 5 µg protein of crude extract contained substantially less than 1 ng of CggR and about 110 ng of GapA. This corresponds to less than 230 molecules of CggR and about 25 000 molecules of GapA per cell. In the gapA-P2 mutant strain, the amount of CggR remained unaffected whereas the expression of GapA was reduced to about 11 000 molecules per cell. Thus, in the wild-type strain there is a more than 100-fold excess of glyceraldehyde-3-phosphate dehydrogenase as compared to the CggR repressor. This drastic difference can presumably be attributed to a large extent to the differential segmental stabilities of mRNAs containing the two genes of the gapA operon. This idea is supported by the correlation of reduced half-lives of the transcripts specifying gapA in the gapA-P2 mutant strain and the concomitant reduction of GapA amount in the cell. Additional factors like codon usage and ribosome binding may be involved in differential expression of CggR and GapA (see Discussion).


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The proteins encoded by the B. subtilis gapA operon fulfil different functions and are required under different conditions: the promoter-proximal gene, cggR, encodes the repressor of the operon. CggR-dependent repression occurs in the absence of the inducers of the operon, i.e. sugar phosphates and amino acids (Fillinger et al., 2000; Ludwig et al., 2001). Thus, CggR is needed under conditions when the operon is repressed, but if the operon is induced there is no need for CggR. Glyceraldehyde-3-phosphate dehydrogenase encoded by gapA is specifically involved in glycolysis but not in gluconeogenesis (Fillinger et al., 2000). Thus, gapA expression is required and occurs only in the presence of sugars. In contrast, the distal four genes of the gapA operon encode enzymes catalysing reversible reactions involved in both glycolysis and gluconeogenesis. This is possibly due to the presence of an internal promoter upstream of pgk which allows constitutive expression of these four genes (Ludwig et al., 2001).

The different expression of cggR and gapA results at least partially from endonucleolytic processing of the primary transcripts and subsequent differential segmental mRNA stability (Ludwig et al., 2001). In this work, we provide evidence for this hypothesis. The determination of the number of CggR and GapA molecules per cell reveals an about 100-fold molar excess of GapA as compared to CggR. A similar ratio was observed for products of the B. subtilis dnaK heat-shock operon, HrcA and DnaK. As described here for the gapA operon, the first gene of the dnaK operon encodes the repressor, HrcA, and the hrcA mRNA is cleaved off the primary transcript (Homuth et al., 1999). Thus, endonucleolytic processing and different segmental mRNA stability may be a common mechanism for operons that include their own regulator. Examples for such an arrangement include the B. subtilis fruRBA, gnt and rbs operons, encoding enzymes necessary for fructose, gluconate and ribose utilization, respectively (Reizer et al., 1999; Fujita & Fujita, 1987; Woodson & Devine, 1994).

Three endonucleolytically active RNases have been identified in B. subtilis, namely RNase III, RNase P and RNase M5. The known substrate specificities as well as direct evidence obtained with a rnmV mutant deficient for RNase M5 suggest that none of these enzymes may be involved in the processing of the primary transcripts of the gapA operon. As suggested by a previous study, these transcripts may rather be cleaved by any functional equivalent of RNase E postulated to be present in B. subtilis (Condon et al., 1997). To characterize the mRNA processing site of the primary transcripts, we performed a site-directed mutagenesis of the site. To the best of our knowledge, this is the first study reporting the effects of mutations of a processing site in the natural sequence context in B. subtilis. RNase E recognizes and cleaves single-stranded AU-rich regions. These regions are often stabilized by adjacent stem–loop structures (McDowell et al., 1994). Our results indicate that destabilization of the stem–loop region immediately downstream of the processing site of the gapA operon primary transcripts affects processing of the mRNA; however, it is only affected to a small extent (see Fig. 2). Moreover, an alteration of the nucleotide sequence of the stem–loop upstream of the processing site which does not significantly interfere with secondary structure formation does not affect cleavage of the primary transcripts at all.

To result in differential expression of two proteins encoded on the same primary transcripts one has to postulate not only a processing event but also differential segmental stabilities of the processing products. Indeed, the monocistronic cggR mRNA is highly unstable, whereas the monocistronic gapA and the pentacistronic gapA pgk tpi pgm eno processing products are rather stable (Ludwig et al., 2001; Fig. 3). It was suggested that mRNA decay in bacteria is initiated at the 5' end of the mRNA. Specifically, single-stranded regions at the 5' end allow initial attack by endonucleases and subsequently degradation of the fragmented mRNA by exonucleases (Grunberg-Manago, 1999; Rauhut & Klug, 1999; DiMari & Bechhofer, 1993). The very short single-stranded region upstream of the predicted stem–loop structure at the 5' end of the stable processed mRNAs of the gapA operon might confer stability to these RNAs. A stem–loop structure at the 5' end is also implicated in the stabilization of the B. subtilis aprE mRNA (Hambraeus et al., 2000, 2002). This idea is strongly reinforced by the observation that the cleaved transcript of a mutant strain specifying an extended 5' single-stranded region (GP554) exhibits a drastically reduced stability as compared to the wild-type (see Fig. 3b). Thus, processing of the primary transcripts upstream of gapA may be an important prerequisite to obtain stable transcripts from which the glycolytic enzymes can be translated.

There is only a weak correlation between mRNA stability and protein levels of CggR and GapA as demonstrated by the still existing 50-fold excess of GapA over CggR in a gapA-P2 mutant strain. In addition to the differential segmental mRNA stability, at least two further factors may contribute to the huge difference in the expression of CggR and GapA. First, the codon usage of the two genes is different. The gapA gene belongs to the class II of highly expressed genes, whereas cggR belongs to the class I of the majority of genes (Moszer et al., 1999). Second, the Shine–Dalgarno sequence of gapA (AAGGAAGG) corresponds exactly to the consensus sequence, whereas the ribosome-binding site of cggR (AAGGAACG) differs at one position from the consensus. Beside the lower efficiency of translation initiation of cggR this may be one determinant of the short half-life of the monocistronic cggR transcript, as it was also shown for the B. subtilis aprE gene (Hambraeus et al., 2002).

An interesting feature of the processing of the primary transcripts of the gapA operon is the position of the processing site which is located in the cggR ORF resulting in a truncated cggR mRNA (Ludwig et al., 2001). It is well established that bacterial ribosomes are stalled on truncated mRNAs and that these complexes are resolved by tmRNA-mediated trans-translation. The truncated proteins are tagged by a specific Ala-rich sequence that is recognized by proteases (Gillet & Felden, 2001). Since only few natural substrates of tmRNA-tagging are known in B. subtilis (Fujihara et al., 2002), we tested whether this mechanism might apply to the translation of CggR from the truncated message. Our findings clearly show that CggR is a very stable protein. However, there was no indication for tmRNA-dependent tagging and degradation of the presumptive tagged protein. The failure to observe the tagging may result from the extreme instability of the truncated cggR mRNA. The portion of tagged CggR may therefore be so small (if there is any) that it was not detectable by our methodology. In agreement with this study, trans-translation of CggR was also not detected in a previous study aimed at the identification of proteins that are subject to this tagging and degradation system (Fujihara et al., 2002).


   ACKNOWLEDGEMENTS
 
We are grateful to Wolfgang Hillen for continuous encouragement. We wish to thank Georg Homuth and Fritz Titgemeyer for helpful discussions. Ulf Gerth and Thomas Wiegert are acknowledged for providing strains. This work was supported by the DFG priority programme Regulatorische Netzwerke in Bakterien and by grants from the Fonds der Chemischen Industrie to J. S.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Båga, M., Göransson, M., Normark, S. & Uhlin, B. E. (1988). Processed mRNA with differential stability in the regulation of E. coli pilin gene expression. Cell 52, 197–206.[CrossRef][Medline]

Belasco, J. G., Beatty, J. T., Adams, C. W., von Gabain, A. & Cohen, S. N. (1985). Differential expression of photosynthetic genes in Rhodopseudomonas capsulata results from segmental differences in stability within a polycistronic transcript. Cell 40, 171–181.[Medline]

Bernstein, J. A., Khodursky, A. B., Lin, P. H., Lin-Chao, S. & Cohen, S. N. (2002). Global analysis of mRNA decay and abundance in Escherichia coli at single-gene resolution using two-color fluorescent DNA microarrays. Proc Natl Acad Sci U S A 99, 9697–9702.[Abstract/Free Full Text]

Bi, W. & Stambrook, P. J. (1998). Site-directed mutagenesis by combined chain reaction. Anal Biochem 256, 137–140.[CrossRef][Medline]

Braun, F., Le Derout, J. & Régnier, P. (1998). Ribosomes inhibit an RNase E cleavage which induces the decay of the rpsO mRNA of Escherichia coli. EMBO J 17, 4790–4797.[Abstract/Free Full Text]

Burchhardt, G., Keshav, K. F., Yomano, L. & Ingram, L. O. (1993). Mutational analysis of segmental stabilization of transcripts from the Zymomonas mobilis gappgk operon. J Bacteriol 175, 2327–2333.[Abstract]

Büttner, K., Bernhardt, J., Scharf, C. & 7 other authors (2001). A comprehensive two-dimensional map of cytosolic proteins of Bacillus subtilis. Electrophoresis 22, 2908–2935.[CrossRef][Medline]

Condon, C., Putzer, H., Luo, D. & Grunberg-Manago, M. (1997). Processing of the Bacillus subtilis thrS leader mRNA is RNase E-dependent in Escherichia coli. J Mol Biol 268, 235–242.[CrossRef][Medline]

Condon, C., Rourera, J., Brechemier-Baey, D. & Putzer, H. (2002). Ribonuclease M5 has few, if any, mRNA substrates in Bacillus subtilis. J Bacteriol 184, 2845–2849.[Abstract/Free Full Text]

DiMari, J. F. & Bechhofer, D. H. (1993). Initiation of mRNA decay in Bacillus subtilis. Mol Microbiol 7, 705–717.[Medline]

Dobrindt, U., Piechaczek, K., Schierhorn, A., Fischer, G., Hecker, M. & Hacker, J. (2002). Influence of the leuX-encoded on the regulation of gene expression in pathogenic Escherichia coli. J Mol Microbiol Biotechnol 4, 205–209.[Medline]

Eddy, C. K., Keshav, K. F., An, H., Utt, E. A., Mejia, J. P. & Ingram, L. O. (1991). Segmental message stabilization as a mechanism for differential expression from the Zymomonas mobilis gap operon. J Bacteriol 173, 245–254.[Medline]

Faires, N., Tobisch, S., Bachem, S., Martin-Verstraete, I., Hecker, M. & Stülke, J. (1999). The catabolite control protein CcpA controls ammonium assimilation in Bacillus subtilis. J Mol Microbiol Biotechnol 1, 141–148.[Medline]

Fillinger, S., Boschi-Muller, S., Azza, S., Dervyn, E., Branlant, G. & Aymerich, S. (2000). Two glyceraldehyde-3-phosphate dehydrogenases with opposite physiological roles in a nonphotosynthetic bacterium. J Biol Chem 275, 14031–14037.[Abstract/Free Full Text]

Fujihara, A., Tomatsu, H., Inagaki, S., Tadaki, T., Ushida, C., Himeno, H. & Muto, A. (2002). Detection of tmRNA-mediated trans-translation products in Bacillus subtilis. Genes Cells 7, 343–350.[Abstract/Free Full Text]

Fujita, Y. & Fujita, T. (1987). The gluconate operon gnt of Bacillus subtilis encodes its own transcriptional negative regulator. Proc Natl Acad Sci U S A 84, 4524–4528.[Abstract]

Gerth, U., Krüger, E., Derré, I., Msadek, T. & Hecker, M. (1998). Stress induction of the Bacillus subtilis clpP gene encoding a homologue of the proteolytic component of the Clp protease and the involvement of ClpP and ClpX in stress tolerance. Mol Microbiol 28, 787–802.[CrossRef][Medline]

Gillet, R. & Felden, B. (2001). Emerging views on tmRNA-mediated protein tagging and ribosome rescue. Mol Microbiol 42, 879–885.[CrossRef][Medline]

Grunberg-Manago, M. (1999). Messenger RNA stability and its role in control of gene expression in bacteria and phages. Annu Rev Genet 33, 193–227.[CrossRef][Medline]

Hambraeus, G., Persson, M. & Rutberg, B. (2000). The aprE leader is a determinant of extreme mRNA stability in Bacillus subtilis. Microbiology 146, 3051–3059.[Abstract/Free Full Text]

Hambraeus, G., Karhumaa, K. & Rutberg, B. (2002). A 5' stem–loop and ribosome binding but not translation are important for the stability of Bacillus subtilis aprE leader mRNA. Microbiology 148, 1795–1803.[Abstract/Free Full Text]

Hansen, A., Pfeiffer, T., Zuleeg, T., Limmer, S., Ciesiolka, J., Feltens, R. & Hartmann, R. K. (2001). Exploring the minimal substrate requirements for trans-cleavage by RNase P holoenzyme from Escherichia coli and Bacillus subtilis. Mol Microbiol 41, 131–143.[CrossRef][Medline]

Hauser, N. C., Vingrom, M., Scheideler, M., Krems, B., Hellmuth, K., Entian, K. D. & Hoheisel, J. D. (1998). Transcriptional profiling on all open reading frames of Saccharomyces cerevisiae. Yeast 14, 1209–1221.[CrossRef][Medline]

Heck, C., Balzer, A., Fuhrmann, O. & Klug, G. (2000). Initial events in the degradation of the polycistronic puf mRNA in Rhodobacter capsulatus and consequences for further processing steps. Mol Microbiol 35, 90–100.[CrossRef][Medline]

Henkin, T. M. (2000). Transcription termination control in bacteria. Curr Opin Microbiol 3, 149–153.[CrossRef][Medline]

Homuth, G., Masuda, S., Mogk, A., Kobayashi, Y. & Schumann, W. (1997). The dnaK operon of Bacillus subtilis is heptacistronic. J Bacteriol 179, 1153–1164.[Abstract]

Homuth, G., Mogk, A. & Schumann, W. (1999). Post-transcriptional regulation of the Bacillus subtilis dnaK operon. Mol Microbiol 32, 1183–1197.[CrossRef][Medline]

Jäger, S., Fuhrmann, O., Heck, C., Hebermehl, M., Schiltz, E., Rauhut, R. & Klug, G. (2001). An mRNA degrading complex in Rhodobacter capsulatus. Nucleic Acids Res 29, 4581–4588.[Abstract/Free Full Text]

Klug, G., Adams, C. W., Belasco, J., Doerge, B. & Cohen, S. N. (1987). Biological consequences of segmental alterations in mRNA stability: effects of deletion of the intercistronic hairpin loop region of the Rhodobacter capsulatus puf operon. EMBO J 6, 3515–3520.[Abstract]

Kunst, F. & Rapoport, G. (1995). Salt stress is an environmental signal affecting degradative enzyme synthesis in Bacillus subtilis. J Bacteriol 177, 2403–2407.[Abstract]

Langbein, I., Bachem, S. & Stülke, J. (1999). Specific interaction of the RNA-binding domain of the Bacillus subtilis transcriptional antiterminator GlcT with its RNA target, RAT. J Mol Biol 293, 795–805.[CrossRef][Medline]

Ludwig, H., Homuth, G., Schmalisch, M., Dyka, F. M., Hecker, M. & Stülke, J. (2001). Transcription of glycolytic genes and operons in Bacillus subtilis: evidence for the presence of multiple levels of control of the gapA operon. Mol Microbiol 41, 409–422.[CrossRef][Medline]

Ludwig, H., Rebhan, N., Blencke, H.-M., Merzbacher, M. & Stülke, J. (2002). Control of the glycolytic gapA operon by the catabolite control protein A in Bacillus subtilis: a novel mechanism of CcpA-mediated regulation. Mol Microbiol 45, 543–553.[CrossRef][Medline]

Martin-Verstraete, I., Débarbouillé, M., Klier, A. & Rapoport, G. (1992). Mutagenesis of the Bacillus subtilis ‘-12, -24’ promoter of the levanase operon and evidence for the existence of an upstream activating sequence. J Mol Biol 226, 85–99.[Medline]

McDowall, K. J., Lin-Chao, S. & Cohen, S. N. (1994). A+U content rather than a particular nucleotide order determines the specificity of RNase E cleavage. J Biol Chem 269, 10790–10796.[Abstract/Free Full Text]

Møller, T., Franch, T., Udesen, C., Gerdes, K. & Valentin-Hansen, P. (2002). Spot 42 RNA mediates discoordinate expression of the E. coli galactose operon. Genes Dev 16, 1696–1706.[Abstract/Free Full Text]

Moszer, I., Rocha, E. P. & Danchin, A. (1999). Codon usage and lateral gene transfer in Bacillus subtilis. Curr Opin Microbiol 2, 524–528.[CrossRef][Medline]

Mudd, E. A., Prentki, P., Belin, D. & Krisch, H. M. (1988). Processing of unstable bacteriophage T4 gene 32 mRNAs into a stable species requires Escherichia coli ribonuclease E. EMBO J 7, 3601–3607.[Abstract]

Newbury, S. F., Smith, N. H. & Higgins, C. F. (1987). Differential mRNA stability controls relative gene expression within a polycistronic operon. Cell 51, 1131–1143.[CrossRef][Medline]

Nilsson, P., Naureckiene, S. & Uhlin, B. E. (1996). Mutations affecting mRNA processing and fimbrial biogenesis in the Escherichia coli pap operon. J Bacteriol 178, 683–690.[Abstract]

Rauhut, R. & Klug, G. (1999). mRNA degradation in bacteria. FEMS Microbiol Rev 23, 353–370.[CrossRef][Medline]

Reizer, J., Bachem, S., Reizer, A., Arnaud, A., Saier, M. H., Jr & Stülke, J. (1999). Novel phosphotransferase system genes revealed by genome analysis – the complete complement of PTS proteins encoded within the genome of Bacillus subtilis. Microbiology 145, 3419–3429.[Abstract/Free Full Text]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Schirmer, F., Ehrt, S. & Hillen, W. (1997). Expression, inducer spectrum, domain structure, and function of MopR, the regulator of phenol degradation in Acinetobacter calcoaceticus NCIB8250. J Bacteriol 179, 1329–1336.[Abstract]

Stülke, J., Martin-Verstraete, I., Zagorec, M., Rose, M., Klier, A. & Rapoport, G. (1997). Induction of the Bacillus subtilis ptsGHI operon by glucose is controlled by a novel antiterminator, GlcT. Mol Microbiol 25, 65–78.[Medline]

Tobisch, S., Zühlke, D., Bernhardt, J., Stülke, J. & Hecker, M. (1999). Role of CcpA in regulation of the central pathways of carbon catabolism in Bacillus subtilis. J Bacteriol 181, 6996–7004.[Abstract/Free Full Text]

Vellanoweth, R. L. (1993). Translation and its regulation. In Bacillus subtilis and Other Gram-Positive Bacteria. Biochemistry, Physiology, and Molecular Biology, pp. 699–711. Edited by A. L. Sonenshein, J. A. Hoch & R. Losick. Washington, DC: American Society for Microbiology.

Wagner, R. (2000). Transcription Regulation in Prokaryotes. Oxford: Oxford University Press.

Wang, W. & Bechhofer, D. H. (1997). Bacillus subtilis RNase III gene: cloning, function of the gene in Escherichia coli, and construction of Bacillus subtilis strains with altered rnc loci. J Bacteriol 179, 7379–7385.[Abstract]

Wiegert, T. & Schumann, W. (2001). SsrA-mediated tagging in Bacillus subtilis. J Bacteriol 183, 3885–3889.[Abstract/Free Full Text]

Woodson, K. & Devine, K. M. (1994). Analysis of a ribose transport operon from Bacillus subtilis. Microbiology 140, 1829–1838.[Abstract]

Zuker, M. (1989). Computer prediction of RNA structure. Methods Enzymol 180, 262–288.[Medline]

Received 28 October 2002; revised 25 November 2002; accepted 26 November 2002.