A plasmid-borne Rap-Phr system of Bacillus subtilis can mediate cell-density controlled production of extracellular proteases

As a soil micro-organism, the Gram-positive eubacterium Bacillus subtilis has acquired the ability to survive in a complex and continuously changing environment. To this purpose, B. subtilis has developed complex signal transduction systems to sense a wide variety of extracellular stimuli. During its life-cycle, B. subtilis uses two-component regulatory systems, consisting of a sensor kinase and a response regulator (mostly a transcription factor), for competence development (Dubnau et al., 1991), synthesis of peptide antibiotics (Nakano et al., 1991; Marahiel et al., 1993), production of secreted proteolytic enzymes (Kunst et al., 1994) and eventually sporulation (Grossman, 1995). Upon phosphorylation of a response regulator by its cognate sensor kinase, transcription of a specific subset of genes is induced (Fabret et al., 1999; Jiang et al., 2000a). Intracellular response regulators aspartyl phosphatases (Raps) and their antagonistic phosphatase regulators (Phrs) that serve as cell-density signalling molecules (Perego et al., 1994, 1996), can govern fine-tuning of these regulatory systems. The Rap proteins have the ability to dephosphorylate their cognate response regulators, thereby temporally overruling the action of sensor kinases (Gray, 1997; Tzeng et al., 1998).

Most B. subtilis 168 rap genes are transcribed in operons with genes that encode their antagonistic Phr peptides (Kunst et al., 1997). While the Rap proteins remain in the cytoplasm, Phr peptides contain an amino-terminal signal peptide and have the potential to be exported as pro-peptides, most likely via the Sec pathway (Tjalsma et al., 2000). Further extracellular processing results in active Phr pentapeptides with a weakly conserved XRXXT sequence (Meijer et al., 1998; Perego, 1997). After reimport via the oligopeptide permease (Opp) system, Phr peptides specifically inhibit the activity of their cognate Rap protein (Perego, 1997, 1998; Meijer et al., 1998). However, four of the eleven known rap genes are not followed by functional phr genes, suggesting that these are not subject to quorum sensing. Interestingly, rapphr operons were also found on several endogenous rolling-circle plasmids from B. subtilis, but no experiments were carried out to find target genes that can be regulated by this class of extra-chromosomally encoded RapPhr systems (Uozumi et al., 1980; Meijer et al., 1998).

The chromosomal-encoded RapPhr systems orchestrate the sequential programs of global gene expression in a growing culture of B. subtilis cells. During exponential growth (low cell densities), RapC dephosphorylates ComA~P, thereby inhibiting competence development. At the end of the exponential growth phase, RapC is inactivated by the accumulation of the PhrC peptide and competence development is stimulated (Mueller et al., 1992; Roggiani & Dubnau, 1993; Lazazzera et al., 1997, 1999). During the transition phase between exponential and post-exponential growth, sporulation is repressed, due to the stimulation of rapA and rapE expression via ComA~P. This results in high cellular levels of RapA and RapE that together dephosphorylate Spo0F (Jiang et al., 2000b; Perego et al., 1996). During the post-exponential growth phase, when nutrients become limiting, RapA and RapE are antagonized by PhrA and PhrE, respectively, and sporulation will be initiated via Spo0F~P. Although the production of degradative enzymes is known to be increased during high cell densities via the DegSDegU two-component system (Msadek et al., 1990), no involvement of Rap proteins in the repression of protein secretion at low cell densities has been documented thus far. The phosphorylation state of the DegU response regulator, mediated by the sensor DegS, acts as a molecular switch allowing either the development of genetic competence or the production of degradative enzymes (Dahl et al., 1992; Kunst et al., 1994; Kunst & Rapoport, 1995).

In this study, we show that the pTA1060-encoded phosphatase Rap60 can actually repress the production of proteolytic enzymes in B. subtilis 168. Expression of aprE, encoding the major extracellular proteolytic enzyme subtilisin, is strongly repressed in cells containing Rap60. However, co-expression of Phr60 or the addition of synthetic Phr60 peptide restored aprE transcription under these conditions. The regulatory cascade affected by Rap60 comprises, most likely, components of the phosphorelay (Hoch, 1993) and the transition-state regulator AbrB (Strauch & Hoch, 1993). These data show for the first time that endogenous plasmids of B. subtilis contain RapPhr systems that have a function in the cell-density-controlled production of at least one secreted protease. This may be of particular importance for the fitness of B. subtilis that can secrete large amounts of proteins into the medium as an adaptive response to environmental fluctuations.

Plasmids, bacterial strains and media.
Table 1 lists the plasmids and bacterial strains used. TY medium (tryptone/yeast extract) contained Bacto tryptone (1 %), Bacto yeast extract (0·5 %) and NaCl (1 %). Extracellular proteolytic activity was tested on NB (Bacto Nutrient Broth; 0·8 % MgSO₄.7H₂O; 0·025 % KCl, pH 7·0) agar (1·5 %) with 1 % skimmed milk. Schaeffer's sporulation medium (SSM) was prepared as described by Schaeffer et al. (1965). When required, media for Escherichia coli were supplemented with ampicillin (50 µg ml^-1), or erythromycin (100 µg ml^-1). Media for B. subtilis were supplemented with chloramphenicol (Cm; 5 µg ml^-1), kanamycin (5 µg ml^-1), spectinomycin (Sp; 100 µg ml^-1), and/or erythromycin (0·3 µg ml^-1). For induction of the PxylA' promoter, 1 % xylose was added to the growth medium.

Table 1. Plasmids and bacterial strains

DNA and RNA techniques.
Procedures for DNA purification, restriction, ligation, agarose gel electrophoresis and transformation of E. coli were carried out as described by Sambrook et al. (1989). Enzymes were from Boehringer Mannheim. B. subtilis was transformed as described by Bron & Venema (1972). For a high reliability PCR, Pwo polymerase (Roche Diagnostics) was used.

Construction of xylose-inducible rap60 and/or phr60 genes.
Functional analysis of the rap60phr60 operon was performed with three constructs in which rap60, phr60 or both genes were placed under control of the PxylA' promoter on the integration plasmid pX (Kim et al., 1996). To this purpose, PCR reactions with pTA1060 as DNA template were performed with the following primer sets (mismatches with the original sequence of pTA1060 are indicated in lower case; restriction sites are underlined). First, prRap1 (5'-gctctagaTTA AAT TAG GGG AGG AG-3') and prRap2 (5'-cgggatccGGA CCG CTG AAA AAA GACC-3') were used to amplify rap60 (1174 bp). Second, prRap3 (5'-gctctagaGAG ATG GTG TGC GCT CAA AG-3') and prRap4 (5'-cgggatccCAT GGT CGC ACA CAA AG-3') were used to amplify phr60 (246 bp). Third, prRap1 and prRap4 was used to amplify the rap60phr60 operon (1420 bp). After digestion of these fragments with XbaI and BamHI, these were ligated into the SpeI and BamHI sites of pX, resulting in pXR (rap60) and pXP (phr60) pXRP (rap60phr60). Next, B. subtilis 168 was transformed with these plasmids, resulting in the integration of these constructs in the amyE locus by homologous gene replacement (Table 1, Fig. 1). Integration into the amyE locus of transformants was confirmed by the lack of halo formation upon growth on plates containing 1 % starch.

Table 1). The position of the Cm resistance marker (cat) is indicated. (e) Schematic presentation of the aprE locus of B. subtilis AL (aprElacZ). Due to the integration of pMutin2 in the aprE locus of B. subtilis 168, the aprE gene was placed under the control of the IPTG-inducible Pspac promoter and the lacZ gene under the control of the PaprE promoter. Emr, erythromycin resistance marker; T0, T1, T2, transcriptional terminators of pMutin2.

Construction of a transcriptional aprElacZ fusion.
To construct a transcriptional aprElacZ fusion, a PCR reaction with chromosomal DNA of B. subtilis 168 as DNA template was performed with the primers aprE400 (5'-aagaattcGGC GGC CGC ATC TGA TGT-3') and aprE+100 (5'-aaggatccGTT AAC GCA AAC AAC AAG-3'). The resulting fragment, containing the promoter region of aprE, was digested with EcoRI and BamHI and ligated into the corresponding sites of pMutin2, resulting in plasmid pAL. Next, this plasmid was integrated into the aprE locus of B. subtilis 168 by a single crossover event (Fig. 1). Correct integration of pAL was verified by Southern blotting (data not shown). Additional mutations were introduced by transformation of chromosomal DNA of the B. subtilis strains A1 (abrBSp) and S1 (spo0K-Sp). The latter strains were derived from BD1807 and K1566, in which the Cm marker was replaced with the Sp marker of pECE74 (Steinmetz & Richter, 1994), respectively.

Sporulation assay.
The efficiency of sporulation was determined by overnight growth of cells in SSM medium, killing of the cells with 0·1 vol. chloroform and subsequent plating.

ß-Galactosidase assays.
Overnight cultures were diluted to an OD₆₀₀ of 0·1 in fresh TY medium, supplemented with Phr60 pentapeptide when desired. Samples for OD₆₀₀ readings and ß-galactosidase activity determinations were taken at hourly intervals. The assay and the calculation of ß-galactosidase activity (expressed as units ß-galactosidase activity per OD₆₀₀ unit) of the samples were performed as described by Miller (1982). Experiments were repeated at least three times and the relative effects within one experiment were reproducible.

Functional analysis of the pTA1060-encoded RapPhr system
To perform a functional analysis of the pTA1060-encoded RapPhr system, the rap60 and phr60 genes were together, and separately, placed under control of the PxylA' promoter in the amyE locus of B. subtilis 168 (Fig. 1). This approach was used to realize significant production levels of Rap60 and/or Phr60, without possible interference of other pTA1060-encoded proteins. However, we realize that, as the expression system used in this study is different from that of the rap60phr60 systems in B. subtilis strains containing pTA1060, observed activities of Rap60Phr60 do not necessarily reflect the (only) physiological function of this system. As at least some chromosomal-encoded Raps are involved in competence, peptide antibiotic production and sporulation (Perego, 1998), these processes were analysed, according to the standardized assays used in the Bacillus subtilis Functional Analysis' (BSFA) programme (see Schumann et al., 2001). However, no obvious phenotype was found with respect to these post-exponential growth processes in B. subtilis producing Rap60 and/or Phr60 (data not shown). To investigate whether Rap60 is involved in other post-exponential growth processes, the secretion of proteolytic enzymes by strains producing Rap60 and/or Phr60 was analysed. To that purpose, colonies were grown on TY agar plates containing skimmed milk and 1 % xylose for 5 days at 15 °C. As shown in Fig. 2, colonies of cells producing Rap60 showed no halo, which is diagnostic for a severe protease secretion defect. As B. subtilis is known to secrete many proteases (Tjalsma et al., 2000), this finding indicates that Rap60 has a general role in protease production or secretion. In contrast, expression of Phr60 alone had no effect on the halo size, whereas production of Rap60 together with Phr60 had a moderate effect on the halo size on skimmed milk plates compared to that of the parental strain. The latter finding indicates that the amount of Phr60 that is produced under these conditions is not sufficient to completely inhibit the action of Rap60. Notably, these phenotypes were observed at 15 °C, but not at 37 °C (data not shown), suggesting that the impaired protease secretion is a temporal rather than a long-term feature of Rap60-producing cells. Possibly, the time window of the Rap60 action on secreted proteases has elapsed after overnight incubation at 37 °C and can only be seen if cellular processes are slowed down at low temperatures. Together, these findings indicate that Rap60 can repress the production of extracellular proteases, and that Rap60 itself is inhibited by Phr60.

(102K): Fig. 2. The Rap60Phr60 system is involved in protease production. To monitor extracellular protease production, B. subtilis X (parental), B. subtilis RP (expressing Rap60 and Phr60), B. subtilis R (expressing Rap60) and B. subtilis P (expressing Phr60) were grown on TY-agar plates containing 1 % skimmed milk and 1 % xylose for 5 days at 15 °C. The size of the haloes around the colonies reflects the level of proteolytic activity secreted into the medium.

The Rap60Phr60 system can mediate cell-density controlled regulation of the aprE gene, encoding a major extracellular protease
Impaired protease secretion in cells producing Rap60 can be due to either a general secretion defect or repression of the transcription of genes encoding extracellular proteases. To discriminate between these possibilities, the promoter of the aprE gene, encoding the major extracellular protease subtilisin, was genetically fused to the lacZ reporter. At the same time, the intact aprE gene was placed under control of the IPTG-inducible Pspac promoter (Fig. 1). Next, the xylose-inducible rap60 gene (R) was introduced in the amyE locus of the aprElacZ PspacaprE containing strain B. subtilis AL. The resulting strain, B. subtilis ALR (for details see Table 1), was grown on TY agar plates containing skimmed milk and xylose for Rap60 expression. The presence of IPTG did partially restore halo formation under these conditions (data not shown), indicating that the secretion of AprE was not impaired upon Rap60 expression. To monitor the effect of Rap60 expression on aprE transcription, the cellular ß-galactosidase activities of B. subtilis ALR (aprElacZ; xylose-inducible Rap60 expression) were determined during growth at 37 °C in TY medium, with and without xylose. As shown in Fig. 3(a), the expression of Rap60 resulted in about a threefold reduction of aprElacZ transcription. In contrast, B. subtilis ALR grown in medium without xylose (no Rap60 expression) displayed an aprElacZ transcription profile similar to that of the parental strain. Finally, the growth rates, final cell densities and aprElacZ transcription profiles were not significantly affected by the presence of xylose in the culture medium (data not shown), showing that Rap60 acts as a regulator of aprE transcription.

(13K): Fig. 3. The Rap60Phr60 system can mediate cell-density-controlled transcription of aprE. (a) Xylose-dependent repression of aprE by Rap60. Time-courses of the transcription of the aprElacZ gene fusion in the parental strain B. subtilis AL () or B. subtilis ALR (xylose-inducible expression of Rap60) grown in the presence () or absence of 1 % xylose (•) in TY medium at 37 °C were monitored. (b) Cell-density controlled transcription of aprE mediated by the Rap60Phr60 system. Time-courses of the transcription of the aprElacZ gene fusion in the parental strain B. subtilis AL (), B. subtilis ALR ( xylose-inducible expression of Rap60), B. subtilis ALP (•; xylose-inducible expression of Phr60) or B. subtilis ALRP (; xylose-inducible expression of Rap60Phr60), all grown in the presence of 1 % xylose in TY medium at 37 °C, were monitored. ß-Galactosidase activities were determined in U per OD600 unit. Time zero indicates the transition point between the exponential and post-exponential growth phases.

To investigate whether the Rap60Phr60 system can mediate cell-density-controlled regulation of aprE transcription, time-courses of aprElacZ were monitored in the parental strain B. subtilis AL (parental), B. subtilis ALR (xylose-inducible expression of Rap60), B. subtilis ALP (P+; xylose-inducible expression of Phr60), and B. subtilis ALRP (xylose-inducible expression of Rap60Phr60), all grown in the presence of 1 % xylose in TY medium at 37 °C. As shown in Fig. 3(b), the response of aprE expression to changes in cell density is smaller in cells producing Rap60 and Phr60 than in cells of the parental strain. Furthermore, the transcription of aprE shows a smaller change in response to cell-density changes in cells expressing only Rap60 than in cells producing Rap60 and Phr60. Together these observations show that the Rap60Phr60 system can mediate aprE in a cell-density controlled manner. It should be noted that all strains had similar growth curves in this experiment. Finally, Fig. 3(b) shows that production of only Phr60 by the xylose system does not significantly affect the expression of aprE in a strain lacking Rap60. This observation shows that Phr60 affects aprE transcription only in cells containing Rap60.

Synthetic Phr60 pentapeptide is imported by the Opp system and inhibits Rap60
Although the partial restoration of halo formation on skimmed milk plates by the co-expression of Phr60 suggests that this peptide inhibits the activity of Rap60 (Fig. 2), the export and reimport of this peptide from and back into the cell remained unproven. To investigate whether the Phr60 peptide could be imported and antagonize Rap60 activity, the deduced active Phr60 pentapeptide (SRNAT; see Meijer et al., 1998) was synthetically produced. Next, transcription of aprElacZ was monitored in B. subtilis ALR (xylose-inducible Rap60 expression) grown in TY medium with xylose to which 1, 5 or 10 µM Phr60 pentapeptide was added. As shown in Fig. 4(a), the addition of Phr60 caused a concentration-dependent increase in aprElacZ transcription under these conditions. In contrast, the addition of 1 µM Phr60 had no effect on the aprElacZ transcription level in the parental strain 168 (AL). These observations show that extracellularly added Phr60 pentapeptides can fulfil an intracellular role in the inhibition of Rap60.

(13K): Fig. 4. Inhibition of Rap60 by extracellular added Phr60. (a) Time-courses of the transcription of the aprElacZ gene fusion in B. subtilis ALR (xylose-inducible expression of Rap60) grown in TY medium with 1 % xylose at 37 °C, without Phr60 peptide (), or with 1 µM (•), 5 µM (), or () 10 µM Phr60 peptide added to the medium, were monitored. (b) Uptake of Phr60 by the Opp system. Time-courses of the transcription of the aprElacZ gene fusion in the B. subtilis strain ALR carrying the xylose-inducible rap60 gene (•), and B. subtilis ALRS carrying an addition spo0K mutation, without (), or with () 10 µM Phr60 pentapeptide in TY medium containing 1 % xylose at 37 °C. ß-Galactosidase activities were determined in U per OD600 unit. Time zero indicates the transition point between the exponential and post-exponential growth phases.

Chromosomal-encoded Phr peptides, such as PhrC, are imported via the Opp system (Solomon et al., 1996). To verify whether Phr60 uses the same import pathway, uptake experiments were performed using a strain carrying a spo0K mutation. This mutation was first identified as a sporulation mutant that later appeared to contain a disrupted opp operon (Perego et al., 1991; Rudner et al., 1991). The spo0K mutation was introduced in the B. subtilis strain ALR to evaluate the importance of the Opp system for Phr60-dependent aprElacZ transcription. To this purpose, the B. subtilis strain ALR (aprElacZ; xylose-inducible Rap60 expression) carrying the spo0K mutation was grown in TY medium with xylose in the presence or absence of 10 µM Phr60 pentapeptide. As shown in Fig. 4(b), transcription of aprElacZ was not significantly affected by the extracellular addition of Phr60 in this spo0K genetic background, indicating that the uptake of the Phr60 pentapeptide indeed requires the Opp system.

Rap60 acts on a component of the phosphorelay
Previous studies showed that mutations in spo0 genes resulted in a drastic decrease in aprE transcription, while a suppressor mutation in the abrB gene restored high-level transcription of aprE in these strains (Ferrari et al., 1988). As overproduction of Rap60 causes a decrease in aprE transcription, like spo0 mutations, possible targets for Rap60 are Spo0F~P, Spo0B~P and Spo0A~P, which sequentially form the phosphate-transfer chain of the phosphorelay (Hoch, 1993). To investigate whether Rap60 acts on a phosphorylated component of the phosphorelay, the abrB mutation was introduced into the B. subtilis strain ALR (aprElacZ; xylose-inducible Rap60 expression). The rationale of this experiment is that, in the absence of AbrB, decreased levels of Spo0A~P would not have an effect on aprE transcription, thereby creating a phosphorelay bypass' (Ferrari et al., 1988). As shown in Fig. 5(a), the expression of Rap60 in cells grown in TY medium with 1 % xylose had no effect on the level of aprElacZ transcription in the abrB genetic background. Taken together, these observations show that by creating a phosphorelay bypass, the inhibitory effect of Rap60 on aprE transcription is diminished, indicating that Rap60 might indeed dephosphorylate one or more components of the phosphorelay.

(13K): Fig. 5. Rap60 acts on a component of the phosphorelay. (a) No effect of Rap60 on aprE transcription in a phosphorelay-bypass mutant. Time-courses of the transcription of the aprElacZ gene fusion in B. subtilis ALR (xylose-inducible expression of Rap60) grown in the presence () or absence (•) of 1 % xylose, or in B. subtilis ALRA (xylose-inducible expression of Rap60 combined with abrB mutation) grown in the presence () or absence () of 1 % xylose in TY medium at 37 °C were monitored. (b) Rap60 can repress the transcription of spoIIA. Time-courses of the transcription of the spoIIAlacZ gene fusion in the parental strain B. subtilis SL (•), or B. subtilis SLR (xylose-inducible expression of Rap60) grown in the presence () or absence () of 1 % xylose in TY medium at 37 °C, were monitored. ß-Galactosidase activities were determined in U per OD600 unit. Time zero indicates the transition point between the exponential and post-exponential growth phases.

The initiation of sporulation through the phosphorelay results in increased levels of Spo0A~P which acts as a transcription factor to repress certain genes, e.g. abrB, but activates other genes, such as spoIIA (Trach et al., 1991). To further investigate whether Rap60 acts on a component of the phosphorelay, the cellular Spo0A~P levels were diagnosed by measuring transcription of the spoIIA gene in cells with and without Rap60. To that purpose, a spoIIAlacZ fusion was introduced into the B. subtilis strain R (xylose-inducible Rap60 expression). As shown in Fig. 5(b), the expression of Rap60 in B. subtilis SLR cells (spoIIAlacZ; xylose-inducible Rap60 expression) grown in TY medium with 1 % xylose strongly decreased the transcription levels of spoIIA compared to that in cells of the same strain grown in the absence of xylose, or the spoIIA transcription levels in cells of the parental strain. This observation strongly suggests that Rap60 dephosphorylates one or more components of the phosphorelay. About 47 % of the chromosomal B. subtilis genes belong to paralogous gene families (Kunst et al., 1997). In addition, certain B. subtilis strains contain endogenous plasmids providing an additional pool of (duplicated) genes. As exemplified by the sipP gene for a plasmid-encoded type I signal peptidase (SPase), the presence of such endogenous plasmids can provide the cell with backup SPases to prevent potential bottlenecks in protein secretion (Tjalsma et al., 1999). Moreover, the presence of plasmid-borne sip genes allows the cell to use additional mechanisms to regulate its patterns of gene expression, as illustrated by the concerted transcription of sipS, sipT, sipP and the genes for secreted degradative enzymes (Tjalsma et al., 1997, 1998, 1999).

Our present studies suggest that the presence of a plasmid-borne RapPhr system from B. subtilis provides this organism with another mechanism to regulate its patterns of gene expression. However, instead of SPase overproduction, under conditions of high-level protein secretion, this Rap60Phr60 system can actually regulate the production of secreted proteases. As indicated by the results obtained from the experiments on the uptake of Phr60 pentapeptide, this regulation is cell-density controlled and repression of protease production occurs under conditions of low cell densities. This property of the Rap60Phr60 system may be important for the fitness of B. subtilis under natural conditions, as it prevents protease production under nutrient-rich conditions, thereby making more efficient use of the available energy sources. Even though these considerations are based on experiments with B. subtilis 168-derived strains, we believe that they are also relevant for B. subtilis strains carrying plasmids pTA1015, pTA1040 or pTA1060. The latter B. subtilis strains produce capsular poly-γ-glutamate and are used for the fermentation of soy beans to produce natto, a traditional Japanese food product. These B. subtilis (natto) strains and B. subtilis 168 are regarded as different isolates of the same organism and are genetically closely related (Priest, 1993). In this respect, it should be noted that in contrast to B. subtilis (natto) strains, B. subtilis 168 contains a mutation in the degQ promoter causing low DegQ production (Yang et al., 1986; Msadek et al., 1991). As AprE production is to a certain extent regulated by the ComPComA quorum-sensing system via DegQ, the latter mutation causes reduced levels of aprE transcription in B. subtilis 168. Finally, the presence of several other endogenous plasmids with genes for RapPhr systems found in natural B. subtilis soil isolates (our unpublished observations), suggests a general importance of plasmid-borne RapPhr systems in the natural habitat(s) of this organism.

Our data indicate that the Rap60Phr60 system is analogous to the known chromosomal-encoded systems, as schematically depicted in Fig. 6. Cells containing pTA1060 express Rap60 which remains in the cytoplasm and Phr60 which is secreted into the medium. Rap60 is likely to be involved in the dephosphorylation of one of the components of the phosphorelay (Spo0FSpo0BSpo0A), thereby repressing aprE transcription. Previous studies have shown that an abrB mutation restored aprE transcription in spo0 mutants (Ferrari et al., 1988). This indicated that AbrB is a preventer of aprE transcription when Spo0A~P is not available (Strauch & Hoch, 1993). Surprisingly, our results suggest that AbrB is directly, or indirectly, also an activator of aprE transcription, as the expression of aprE decreases when abrB is disrupted (Fig. 5a). As the transcription of aprE is regulated by complex systems involving at least ten known regulators (Kallio et al., 1991; Smith, 1993; Strauch & Hoch, 1993; Bolhuis et al., 2000), the exact role of AbrB in aprE regulation is not completely clear at the moment. Nevertheless, expression of Rap60 did not affect the transcription of aprE in an abrB mutant background (Fig. 5a). In contrast, expression of Rap60 did affect the transcription of spoIIA (Fig. 5b), a well-known target gene of the phosphorelay (Trach et al., 1991). Together, these observations strongly suggest the existence of a Rap60phosphorelayAbrB regulatory cascade involved in aprE transcription. At high cell densities, the extracellular Phr60 concentration increases and is internalized by the Opp system. Finally, these high intracellular Phr60 concentrations cause inhibition of Rap60, more phosphorylated Spo0A, less AbrB and eventually increased AprE production.

(17K): Fig. 6. Model for cell-density controlled regulation of AprE production. Cells containing pTA1060 express Rap60 which remains in the cytoplasm and Phr60 which is secreted into the medium. Rap60 is able to dephosphorylate one of the components of the phosphorelay (Spo0F, Spo0B, or Spo0A), thereby lowering the cellular Spo0A~P levels, resulting in lower aprE transcription, possibly via AbrB. At high cell densities, sufficient Phr60 is accumulated in the medium to be taken up in sufficient amounts by the Opp system. This results in Rap60 inhibition, more phosphorylated Spo0A and eventually increased AprE secretion.

Although Rap60 seems to be acting on a component of the phosphorelay, which is known to govern, in particular, the sporulation process in B. subtilis, we did not observe an effect on sporulation by the expression of Rap60 (our unpublished observations). A possible explanation is that expression of Rap60 causes a level of delay in sporulation that is too small to be detected in our assays. Similarly, the protease production defect upon Rap60 expression on skimmed milk plates was only observed at low temperatures. At 37 °C, the repression of protease production was probably too mild to be seen after overnight growth. Nevertheless, when transcription of aprE or spoIIA were monitored at hourly intervals at 37 °C, a clear repression of the transcription of these genes was observed (Figs 3a and 5b).

The reason for presence of RapPhr systems on the B. subtilis plasmids pTA1015, pTA1040 and pTA1060 (Meijer et al., 1998) is presently not clear. The identification of these systems on endogenous plasmids of B. subtilis strains producing poly-γ-glutamate suggests that RapPhr systems may have a role during the fermentation of soy beans for the production of natto. Therefore, these plasmid-borne rapphr operons might be involved in the cell-density controlled regulation of secreted proteases necessary for producing poly-γ-glutamate (Birrer et al., 1994; Nagai et al., 2000). Alternatively, these plasmid-encoded RapPhr systems might be important for the temporal repression of protease production during natto production. Finally, this system may be important for the, until now unidentified, regulation of other plasmid-borne or chromosomal genes.

At least one further issue that needs to be addressed is whether only plasmid-encoded RapPhr systems are involved in cell-density controlled protease production. As Rap60 seems to be acting on a component of the phosphorelay, other chromosomal-encoded Raps that have been shown to be involved in the phosphorelay might be involved in protease production as well. The latter idea is corroborated by some lines of evidence. First, the most closely related chromosomal-encoded homologue of Rap60 is RapE (our unpublished observations), which is involved in the cell-density controlled dephosphorylation of Spo0F~P (Jiang et al., 2000b). Second, high Phr60 concentrations result in aprE transcription levels that exceed those of the parental strain (our unpublished observations), suggesting that other chromosomal-encoded Raps can be inhibited by the Phr60 pentapeptide. As the Rap60 active pentapeptide (SRNAT) is quite similar to PhrE (SRNVT), a likely candidate for being subject to Phr60 inhibition is RapE. Furthermore, the inhibitory effect of Phr60 on Rap60 can be mimicked by the PhrE peptide, albeit with a much lower efficiency (our unpublished observations). These data indicate that cross-talk may exist between the plasmid- and chromosomal-encoded RapPhr systems. Taken together, these findings are important to increase our insight into how B. subtilis can exploit (plasmid-borne) RapPhr systems for the modulation of extracellular protease production during highly fluctuating environmental conditions.

We thank Dr Eugenio Ferrari (Genencor International, Palo Alto, USA) for the generous gift of Phr pentapeptides, Leendert W. Hamoen, Jan D.H. Jongbloed and other members of the Groningen and European Bacillus Secretion Groups, and the Chromosome Minimization Consortium for stimulating discussions. E.J.K. and S.B. were supported by European Union Grants Bio4-CT98-O250 and QLK₃-CT-99-00413; H.T. was supported by Genencor International (Leiden, The Netherlands).