A new family of aspartyl phosphate phosphatases targeting the sporulation transcription factor Spo0A of Bacillus subtilis

Authors

  • Marta Perego

    1. Division of Cellular Biology, Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, CA 92037, USA.
    Search for more papers by this author

*For correspondence. E-mail mperego@scripps.edu; Tel. (+1) 858 784 7912; Fax (+1) 858 784 7966.

Abstract

The initiation of the sporulation developmental pathway in Bacillus subtilis is controlled by the phospho-relay, a multicomponent signal transduction system. Multiple positive and negative signals are integrated by the phosphorelay through the opposing activities of histidine protein kinases and aspartyl phosphate phosphatases. Three members of the Rap family of phosphatases (RapA, RapB and RapE) specifically dephosphorylate the Spo0F∼P response regulator intermediate, while the Spo0A∼P transcription factor is specifically dephosphorylated by the Spo0E phosphatase and, as shown here, the newly identified YnzD and YisI proteins. The products of the YnzD and YisI genes are highly homologous to Spo0E and define a new family of phosphatases with a distinct signature motif in their amino acid sequence. As negative regulators of the developmental pathway, YnzD and YisI inhibit spore formation if over-expressed, while a chromosomal deletion of their coding sequences results in increased sporulation frequency. Transcription of the ynzD, yisI and spo0E genes is differentially regulated and generally induced by growth conditions antithetical to sporulation. Negative signals interpreted by aspartyl phosphate phosphatases appear to be a common mechanism in Gram-positive spore-forming microorganisms.

Introduction

Two-component regulatory systems are a major means by which bacteria recognize and respond to environmental, metabolic and cell cycle signals (for reviews see Hoch and Silhavy, 1995). The prototypical two-component pathway is minimally composed of a histidine protein kinase that autophosphorylates upon signal reception, and a response regulator whose activation occurs by transfer of the phosphoryl group from a histidine of the kinase to an aspartic acid of the response regulator.

In all signal transduction systems, a reversible protein phosphorylation mechanism is essential to accurately interpret signalling information. In two-component systems, turnover of phosphorylated response regulators is generally carried out by means of intrinsic autodephos-phorylation or by dephosphorylation by histidine kinases acting alone or in conjunction with ancillary proteins (Russo and Silhavy, 1991; Lukat et al., 1992; P. Jiang et al., 2000; Pioszak, Jiang and Ninfa, 2000). It is now apparent, however, that in prokaryotes as well as eukaryotes, dephosphorylation may also be carried out by auxiliary protein phosphatases. A well-defined example of the resulting complex interplay between protein kinases and protein phosphatases is provided by the phosphorelay signal transduction system for sporulation initiation in Bacillus subtilis (Perego, 1998).

The phosphorelay is a more complex version of the typical two-component system in that differential signals activate multiple histidine kinases to autophosphorylate and then transfer their phosphoryl group to the intermediate response regulator Spo0F (Burbulys, Trach and Hoch, 1991; M. Jiang et al., 2000). Spo0F∼P is the substrate for a phosphotransferase, Spo0B, that transfers the phosphoryl group to the Spo0A response regulator and transcription factor. Spo0A∼P acts both as a repressor of certain genes for vegetative growth and an activator of genes required for sporulation initiation (Strauch and Hoch, 1993). As the level of Spo0A∼P in the cell is the factor determining the decision to either grow or sporulate, its production must be carefully regulated in order to ensure the most appropriate cellular response to any given signal.

The complexity of the phosphorelay provides multiple points of entry for regulatory mechanisms that affect the phosphorylation level of the system. A major regulatory role is played by aspartyl phosphate phosphatases that dephosphorylate response regulators and allow signals contrary to sporulation to have an impact on the phosphorelay and its output product Spo0A∼P (Perego, 1998). The Spo0F∼P intermediate is the target of the RapA, RapB and RapE members of the Rap family of phosphatases. These are known to be differentially activated by physiological processes alternative to sporulation, e.g. competence development induces RapA and RapE, while vegetative growth conditions induce RapB (Perego et al., 1994; Jiang, Grau and Perego, 2000). The phosphorylated active form of the Spo0A transcription factor is subject to deactivation by the Spo0E phosphatase, whose transcription and enzymatic activity is affected by signals that are still unknown (Perego and Hoch, 1991; Ohlsen, Grimsley and Hoch, 1994).

There are many similarities between the enzymatic activities and roles of Rap phosphatases and Spo0E, despite having no significant homology in the primary amino acid sequence. Purified proteins specifically induce dephosphorylation of their targets, Spo0F∼P or Spo0A∼P, with no apparent cross-reactivity and no direct effect on the remaining members of the phosphorelay, kinases and Spo0B. Deletion of the rap genes or spo0E gene results in increased sporulation efficiency, while overproduction of these genes results in inhibition of sporulation. Therefore, these phosphatases act as negative regulators of the developmental process and provide a means for additional signals to have an impact on the phosphorylation level of the phosphorelay pathway.

Two additional phosphatases active on the phosphorelay have been identified and characterized in this study. The product of the ynzD and yisI genes are highly homologous to Spo0E and show specific phosphatase activity towards Spo0A∼P. Therefore, in addition to the Rap family of phosphatases which is comprised of 11 members, the YnzD, YisI and Spo0E proteins identify a new family of phosphatases and reveal additional regulatory elements that the phosphorelay integrates into the cell's decision to grow or sporulate.

Results

Identification of yisI and ynzD as inhibitors of sporulation initiation

Screening of a Bacillus subtilis chromosomal library constructed in the shuttle vector pHT315 allowed the identification of negative regulators of the phosphorelay on the basis of inhibiting sporulation when overexpressed on a multicopy plasmid. The KipI inhibitor of the kinase A histidine kinase and the RapE aspartyl phosphate phosphatase of the Spo0F response regulator were previously identified using this method (Wang et al., 1997; Jiang, Grau and Perego, 2000).

From the same screening an additional clone, pRM85, was isolated carrying a 4.3 kb fragment ranging from nucleotide 1,147,887 to nucleotide 1,152,180 in the B. subtilis genome sequence database (Kunst et al., 1997). The presence of plasmid pRM85 in the wild-type strain JH642 resulted in a severe sporulation-deficient phenotype (∼40-fold less spores than in the wild-type strain carrying the vector pHT315) (Table 1). As the region cloned in pRM85 contained several open reading frames (Fig. 1A), a subcloning strategy in pHT315 was designed in order to pinpoint the determinant of the sporulation phenotype. The fragment subcloned in pRM85V and pYISI1 was found to be both necessary and sufficient to confer sporulation deficiency when present in multicopy in strain JH642. Plasmids pRM85-1.2, pRM85-1.9, pRM85-1.5 and pYISI23 did not inhibit sporulation when present in multicopy in the wild-type strain. As the fragment cloned in pYISI1 contained a single intact open reading frame, yisI, and its promoter region, it was concluded that overexpression of the YisI protein inhibited sporulation.

Table 1.  Sporulation efficiency of B. subtilis strain carrying YisI and YnzD multicopy plasmids.a
PlasmidbInsertViable countSpore count% of spores
  • a.

    Representative of two independent experiments.

  • b. Strains JH642 harbouring the indicated plasmids were grown for 33 h in Schaeffer's sporulation medium containing 25 μg ml −1 erythromycin.

  • c.

    4.3 kb chromosomal region.

pHT315Vector3.5 × 1081.0 × 10828.6
pRM85 yisI regionc3.8 × 1082.6 × 1060.68
pYISI1 yisI 7.1 × 1085.9 × 1040.008
pYISI23 yisI promoter3.4 × 1088.8 × 10725.9
pYNZD17 ynzD 3.0 × 1084.0 × 1030.001
pYNZD13 ynzD promoter3.4 × 1089.9 × 10729.1
Figure 1.

Restriction map of the chromosomal regions containing the yisI and ynzD loci. Fragments cloned in plasmids used in this study are indicated by lines. Restriction sites in parenthesis are not unique. A: yisI. B: ynzD.

A database homology search for the yisI gene product revealed the presence, in the B. subtilis genome, of an additional gene, ynzD, encoding a protein with 32% identical residues to YisI. Whether or not overexpression of YnzD could also result in inhibition of sporulation was then investigated. A chromosomal fragment carrying the ynzD gene and its putative promoter region was amplified by polymerase chain reaction (PCR) reaction from chromosomal DNA of strain JH642 and cloned in the multicopy vector pHT315 giving plasmid pYNZD17 (Fig. 1B). Competent cells of JH642 transformed with pYNZD17 gave rise to sporulation-deficient colonies. However, the multicopy plasmid pYNZD13 carrying the ynzD promoter region and a truncated ynzD coding sequence did not affect sporulation (Table 1). Therefore, the YnzD protein, as well as YisI, acts as an inhibitor of sporulation when overexpressed.

In order to investigate the role of the chromosomal copy of yisI and ynzD in sporulation, gene inactivation experiments were carried out. The chloramphenicol and kanamycin resistance cassettes were integrated in the ynzD and yisI genes, respectively, via a double cross-over event by means of plasmids pYNZD19 and pYISI24 (Fig. 1A and B). Derivatives of JH642 carrying single or double inactivations were then assayed for sporulation efficiency. The results shown in Table 2 indicated that, while ynzD and yisI did not significantly affect sporulation when inactivated singly, an additive phenotype was obtained in the double mutant strain that consistently manifested a sporulation efficiency higher than the wild-type strain. Taken together, these results suggested that the yisI and ynzD gene products act as negative regulators of the sporulation process.

Table 2.  Sporulation efficiency of yisI and ynzD deletion mutant strains.a
StrainRelevant genotypeViable countSpore count% of spores
  • a.

    Representative of three independent experiments. Cells were grown for 30 h in Schaeffer's sporulation medium.

JH642wild type3.4 × 1082.2 × 10864.7
JH11877 yisI 2.7 × 1081.8 × 10866.6
JH11876 ynzD 2.4 × 1081.6 × 10866.6
JH11881 yisI, ynzD 2.5 × 1082.0 × 10880.0
JH12745 spo0E 2.2 × 1081.8 × 10881.8

YisI and YnzD are homologues of the Spo0E phosphatase

The search in the database previously mentioned also revealed a strong similarity of YisI and YnzD to the Spo0E aspartyl phosphate phosphatase that specifically dephosphorylates the Spo0A∼P response regulator of the phosphorelay (Fig. 2) (Perego and Hoch, 1987). YisI (MW6536) and YnzD (MW6547) are small proteins of 56 and 57 amino acids in length respectively. Spo0E is a protein of 85 amino acids and a truncated Spo0E mutant protein (Spo0E94), lacking the carboxy terminal 25 amino acids, is hyperactive in dephosphorylating Spo0A∼P (Ohlsen, Grimsley and Hoch, 1994). The similarity of YisI and YnzD with Spo0E extends over the 59 amino acids corresponding to the Spo0E 94 mutant form and consist of 29% and 34% of identical residues, respectively, while conserved residues range between 12% and 15% (Fig. 2). Particularly striking is the sequence conservation of a pentapeptide (-SQELD-) within the sequence of the three proteins. The -SQELD-conserved sequence is flanked by two conserved amino acid residues upstream and two hydrophobic residues downstream that seems to constitute a signature sequence for Spo0E-like phosphatases. This motif is indeed highly conserved among Spo0E orthologues identified in other spore-forming Gram-positive bacteria (M. Perego and J. A. Hoch, unpublished data) and appears to be essential for dephosphorylation activity (S. Stephenson and M. Perego, unpublished observations).

Figure 2.

Amino acid sequence alignment of the YisI, YnzD and Spo0E phosphatases. Identical and conserved residues are indicated by asterisks and colons respectively. The dots indicate residues identical or conserved between Spo0E and either YisI or YnzD. The underlined Q and W residues in the Spo0E sequence are the sites of the spo0E94 and spo0E11 mutations, respectively, resulting in Spo0E C-terminal truncated proteins with hyperphosphatase activity.

YnzD and YisI dephosphorylate the Spo0A∼P response regulator

The strong similarity of YnzD and YisI with Spo0E and their negative regulatory role in sporulation, reminiscent of the role of Spo0E itself, suggested that these newly identified proteins were also acting as phosphatases of Spo0A∼P.

In order to carry out in vitro biochemical assays, the YnzD and YisI proteins were overexpressed and purified from Escherichia coli cells in N-terminal His-tagged modified forms. Purified YnzD and YisI proteins were then assayed for phosphatase activity against the phosphorelay in which Spo0A∼P was formed as a result of a series of transphosphorylation reactions involving KinA, Spo0F and Spo0B. As shown in Fig. 3A, increasing concentrations of YnzD or YisI inhibited the accumulation of Spo0A∼P. Omitting Spo0A from the reaction (Fig. 3B) showed that the inhibitory activity of YnzD and YisI did not extend to the other reactions. These results suggested that the target of YnzD and YisI was Spo0A∼P, while KinA, Spo0F and Spo0B were not affected.

Figure 3.

A. Spo0A∼P dephosphorylation by YisI and YnzD. Spo0A (5 µM), was phosphorylated in a phosphorelay reaction containing KinA (0.2 µM), Spo0F (5 µM), Spo0B (1 µM) and [γ-32P]-ATP as described in Experimental procedures. YnzD or YisI were then added at the following concentrations: 1 µM in lanes 2 and 7; 5 µM in lanes 3 and 8; 10 µM in lanes 4 and 9; 15 µM in lanes 5 and 10; 20 µM in lanes 6 and 11. Lane 1 is the control reaction without phosphatases.

B. YnzD and YisI do not dephosphorylate the other components of the phosphorelay. A phosphorelay reaction was carried out containing KinA (0.2 µM), Spo0F (5 µM) and Spo0B (1 µM) as described in Experimental procedures. YnzD and YisI were then added at the concentrations described in A.

In order to demonstrate a direct phosphatase activity of YnzD and YisI on phosphorylated Spo0A, 32P-labelled Spo0A∼P was isolated from the remaining components of the phosphorelay reaction and incubated with the YisI or YnzD proteins as described in Experimental procedures. A time-course of Spo0A∼P dephosphorylation by YisI and YnzD is reported in Fig. 4. YnzD and YisI were found to dephosphorylate Spo0A∼P with an apparently similar initial rate. The control reaction containing Spo0A∼P alone was significantly more stable. The presence of Mg2+ ions seemed to stimulate Spo0A dephosphorylation two- to fivefold in the presence of YnzD and YisI while only 25% increase was observed in the autodephosphorylation reaction.

Figure 4.

Dephosphorylation of purified Spo0A∼P by YisI and YnzD. 32P-labelled Spo0A∼P was purified as described in Experimental procedures and incubated with YisI and YnzD proteins for the time indicated in the figure. Samples were analysed on a 15% SDS/glycine/polyacrylamide gel and quantified with the PhosphorImager and ImageQuant software (Molecular Dynamics). The level of remaining Spo0A∼P in the reaction is expressed as pixel values. Reactions carried out with MgCl2 are indicated by solid symbols, while those carried out without MgCl2 are indicated by open symbols. (○) and (●), Spo0A∼P alone; (□) and (▪), Spo0A∼P and YisI; (▵) and (▴), Spo0A∼P and YnzD.

These experiments demonstrate that the YisI and YnzD protein homologues of the Spo0E phosphatase also act by specifically dephosphorylating the Spo0A∼P response regulator and thus they define a new family of aspartyl phosphate phosphatases in the signal transduction system for sporulation initiation.

Transcription analysis of the ynzD and yisI promoters

Previous work established the spo0E promoter as a prototype of Spo0A-AbrB transcriptional control (Perego and Hoch, 1991). Derepression of spo0E transcription at the end of exponential phase is the result of the Spo0A∼P-dependent repression of the abrB gene transcription, as AbrB is a direct repressor of the spo0E promoter (Strauch et al., 1989). Thus, the synthesis of the spo0E gene product is controlled by Spo0A through the action of the AbrB regulator.

In order to determine whether a similar regulatory mechanism controlled ynzD and yisI, the transcription of these genes was investigated by means of transcriptional fusion constructs to the lacZ gene of E. coli. In Schaeffer's sporulation medium, the ynzD promoter was active mainly during exponential growth and this activity decreased three- to fourfold by the time the culture reached the transition to stationary phase (To). An spo0A mutation did not affect ynzD expression but an abrB mutation consistently increased the transcription level (1.5- to twofold) without affecting the temporal regulation (Fig. 5A). Consensus sequences for σA-containing RNA polymerase were identified upstream the GTG start codon, supporting the conclusion that ynzD transcription is mainly vegetative. Curiously, the −35 sequence is adjacent to the −35 consensus sequence of the divergently transcribed ccdA promoter (Schiött, von Wachenfeldt and Hederstedt, 1997), but a co-regulatory mechanism has so far eluded detection.

Figure 5.

Transcription regulation of the yisI and ynzD promoters. Cultures for β-galactosidase assays were grown in Schaeffer's sporulation medium. Time points were taken hourly before and after the transition (0) from exponential growth to stationary phase. A: yisI promoter (plasmid pYISI 1); B: ynzD promoter (plasmid pYNZD 12). Symbols: (●) JH642 wild type; (▪) JH646 spo0A; (▴) JH12546 spo0A, abrB.

The yisI promoter was mainly active at the end of the exponential phase and during the first hour of stationary phase. Its induction was abolished by an spo0A mutation and it was not restored by an abrB mutation. This suggested that yisI transcription may be dependent upon Spo0A for direct activation. The presence of an Spo0A box (TGACGAA) overlapping a putative −35 promoter consensus sequence supports this notion (data not shown).

These results indicate that each member of the Spo0E family of phosphatases is subject to a differential transcription regulatory mechanism that results in a temporally distinct appearance of the phosphatase activity during cell growth in sporulation medium.

Physiological conditions antithetical to sporulation induce expression of the Spo0E family of phosphatases

Physiological conditions antithetical to sporulation, such as vegetative growth and competence, are known to control the expression of the members of the Rap family of phosphatases that dephosphorylate the Spo0F∼P response regulator of the phosphorelay (Perego, 1998). With the identification of the Spo0E family of phosphatases acting on Spo0A∼P, the question arose of whether these additional phosphatases also reflected the need to recognize and interpret differential physiological signals to have an impact on the phosphorelay.

Experiments described above already indicated differential regulatory and temporal controls in the expression of YnzD, YisI and Spo0E in sporulation growth conditions, with the ynzD and yisI genes being transcribed mainly during the vegetative phase of growth and the transition state, respectively, and the spo0E gene being induced approximately 1 h before the transition to stationary phase and sporulation (Perego and Hoch, 1987; 1991). Therefore, we assayed gene expression in growth conditions that did not promote sporulation, but rather promoted competence development, extensive growth and carbon catabolism repression of sporulation. Cultures were grown in the following media: Spizizen minimal medium for competence development (MT), Penassay broth (PY), Luria–Bertani broth (LB) and Schaeffer's medium with the addition of 0.5% of glucose. As shown in Fig. 6, differential levels of derepression were observed with each promoter in the various growth conditions. The strongest derepression was observed with the ynzD promoter in media supplemented with 0.5% glucose (MT and Schaeffer's with glucose 0.5%) (15-fold overexpression at T2 compared with the level of transcription in Schaeffer's sporulation medium). Transcription was induced 60 min before the transition to stationary phase, a time that corresponded to the turn-off of this promoter in sporulation growth condition (Fig. 6D). Growth in LB or PY also induced ynzD transcription at T-1 but only an eight- and fivefold increase, respectively, was observed over the level obtained in Schaeffer's medium (Fig. 6A). yisI was maximally induced by growth in LB and PY (approximately fourfold compared with the expression in Schaeffer's sporulation medium) (Fig. 6B), while media containing 0.5% glucose only induced a low constitutive level of expression (Fig. 6E). A still different pattern of expression was obtained with the analysis of the spo0E promoter which seemed to be deregulated by growth in competence medium but not necessarily by growth in the presence of glucose 0.5%, as inferred by the low effect detected in Schaeffer's medium supplemented with glucose (Fig. 6F). Furthermore, growth in PY induced Spo0E expression approximately twofold and growth in LB prevented the turn-off of the promoter generally occurring between T1 and T2. Timing of induction of spo0E, however, was not significantly affected in these two media (Fig. 6C).

Figure 6.

Transcription of yisI, ynzD and spo0E in non-sporulating growth conditions. Strains JH11858 (yisI–lacZ) (B and E) JH11844 (ynzD–lacZ) (A and D) JH12567 (spo0E–lacZ) (C and F) (Perego and Hoch, 1991) were grown at 37°C and samples were collected hourly before and after the transition (0) between exponential and stationary phase. Symbols (●) Schaeffer's sporulation medium (SM); (▴) Luria–Bertani broth (LB); (▾) Spizizen minimal medium (MT); (▪) Penassay broth (PY); (◆) Schaeffer's sporulation medium with glucose 0.5%.

These results once again support the conclusion that each member of the Spo0E family of phosphatases is differentially regulated at the level of gene transcription as they respond differently to various growth conditions. In general, however, growth conditions that do not promote sporulation result in a higher level of transcription of all the phosphatases belonging to the Spo0E family.

Discussion

We have identified two new aspartyl phosphate phosphatases that, with the Spo0E protein previously identified, constitute a new family of phosphatases negatively regulating the sporulation initiation pathway in Bacillus subtilis. The YisI and YnzD phosphatases, like Spo0E, specifically dephosphorylate the Spo0A∼P response regulator and key transcription factor in the phosphorelay signal transduction system.

Amino acid sequence analyses show a significant level of homology between Spo0E and YisI or YnzD. The homology extends over the first 54 amino terminal residues of Spo0E as YisI and YnzD are of a smaller size (56 and 57 amino acids, respectively, compared with the 85 amino acids of Spo0E). Indeed, a Spo0E mutant protein truncated at residue 59 (spo0E94) (Fig. 2) was originally identified as an hyperactive phosphatase whose presence in the cell resulted in inhibition of sporulation (Perego and Hoch, 1987; Ohlsen, Grimsley and Hoch, 1994). A second mutation resulting in an identical phenotype, spo0E11, was identified as a stop codon at residue 72 (Fig. 2) (Perego and Hoch, 1987). The observation that YisI and YnzD resemble the Spo0E hyperactive protein in size, raises the question of their relative specific activity. In preliminary kinetic experiments (unpublished data), the purified Spo0E94 mutant protein is approximately threefold more active as a phosphatase than the wild-type Spo0E and sixfold more active than YisI and YnzD. This lower specific activity of YisI and YnzD compared with Spo0E94 may explain, in part, why a spo0E94 mutant strain is sporulation-deficient while the expression of active yisI and ynzD genes does not impair the sporulation process. In fact, only overproduction of yisI and ynzD from a replicative plasmid, presumably present at 15 copies per cell, results in inhibition of sporulation. Indeed, the maximal level of transcription of yisI and ynzD is approximately seven- and fivefold lower, respectively, than the maximal level of transcription of spo0E in sporulation growth conditions. Analogously, while a deletion of spo0E results in an elevated sporulation efficiency, yisI and ynzD deletions affect spore formation only when combined (Table 2).

Also quite distinct are the timing and the regulatory mechanisms controlling the transcription of these genes. We previously showed that spo0E is induced at the end of the exponential phase as repression by AbrB is relieved owing to accumulation of Spo0A∼P and its transcription continues for 2 h into sporulation (Perego and Hoch, 1991). The YisI protein is expressed mainly during the transition phase between exponential and stationary phase, its induction is dependent upon the presence of an intact Spo0A gene and is not affected by AbrB repression. A direct interaction of Spo0A with the yisI promoter is suggested by the presence of a Spo0A box overlapping the putative −35 consensus sequence. Transcription of the ynzD gene is somewhat unusual. In fact, it's transcription in sporulation growth conditions is limited to the early exponential phase and is not affected by an spo0A mutation but increases in a double mutant spo0A–abrB.

Therefore, in sporulation growth conditions, Spo0E appears to be the major negative regulator acting on Spo0A∼P. However, different media affect the transcription of yisI, ynzD or spo0E in different ways. In particular, growth conditions that promote growth and do not induce sporulation significantly stimulate transcription of the phosphatase genes. This medium-dependent expression suggests that each phosphatase responds to differential signals that ultimately have an impact on the cell's decision to initiate sporulation. Similarly, the members of the Rap family of phosphatases, RapA, RapB and RapE, that specifically dephosphorylate the Spo0F∼P intermediate of the phosphorelay are transcriptionally controlled by physiological conditions that are antithetical to sporulation, such as growth and competence to DNA transformation (Mueller, Bukusoglu and Sonenshein, 1992; Jiang, Grau and Perego, 2000). A direct effect of the ComP–ComA two-component system on phosphatase gene transcription was not seen with any of the phosphatase-coding genes described here, as suggested by the lack of ComA binding sites in the promoter regions (Nakano, Xia and Zuber, 1991; Dubnau et al., 1994). However, a growth condition that promotes competence development strongly stimulates the transcription of ynzD and spo0E and has a noticeable effect on yisI. This could be as a result of an indirect effect of the Com system perhaps controlling a regulator(s) of transcription acting on the phosphatase coding genes. Alternatively, the effect may be mediated by an element present in the medium. For example, glucose at a concentration of 0.5% may be responsible for the strong induction of ynzD transcription and the increased transcription of yisI in MT medium and Schaeffer's medium. The transcription of spo0E, in contrast, is not necessarily induced by glucose as its rate of transcription in Schaeffer's containing glucose does not differ from the rate observed in plain Schaeffer's. In MT medium, however, spo0E transcription is highly induced from the early hours of exponential phase, suggesting that a mechanism exists to overcome the repression exerted by AbrB. The nature of this mechanism is unknown at this time as well as whether a specific transcription regulator(s) exists that acts on the promoter region of the yisI and ynzD genes. Finally, it should be noted that mechanisms regulating the activity of these phosphatases may be more important than transcriptional controls.

The YisI (6.5 kDa), YnzD (6.5 kDa), and Spo0E (9.6 kDa) phosphatases are small proteins relative to the average size of the Rap phosphatases (approximately 44.5 kDa) that specifically dephosphorylate Spo0F∼P. Consistently, no sequence similarity was observed between the two families of proteins despite the structural similarities between the respective targets, Spo0A∼P and Spo0F∼P. Also, no sequence similarities were detected with the CheZ protein known to stimulate dephosphorylation of the CheY response regulator of the chemotaxis signal transduction system in E. coli (Mutoh and Simon, 1986; Hess et al., 1988). As is also true for CheZ, the molecular basis of the mechanism of action of Rap or Spo0E phosphatases has not been elucidated. They could act as phosphatases catalytically hydrolysing the acyl phosphate bond, or as allosteric effectors changing the conformation of their target protein, therefore accelerating the autodephosphorylation reaction (Lukat, Stock and Stock, 1990; Silversmith, Appleby and Bourret, 1997; Tzeng et al., 1998; Boesch, Silversmith, and Bourret, 2000). The YisI- and YnzD-dependent dephosphorylation of Spo0A∼P is significantly enhanced by Mg2+ ions as is the autodephosphorylation reaction. Spo0E-dependent dephosphorylation of Spo0A∼P is also dependent on Mg2+ as a cofactor (S. Stephenson and M. Perego, unpublished observations) and so is the Spo0F∼P dephosphorylation by the RapA and RapB phosphatases (Tzeng et al., 1998; S. Ishikawa and M. Perego, unpublished observations). These observations point to an allosteric mechanism for the Spo0E and Rap proteins in dephos-phorylating Spo0A∼P and Spo0F∼P respectively.

It has been reported that the CheZ phosphatase activity on CheY∼P may be regulated by other components of the chemotaxis pathway; interaction with a short form of the CheA histidine kinase, CheAs, or oligomerization with CheY∼P may affect CheZ activity (Blat and Eisenbach, 1994; 1996;Wang and Matsumura, 1996). The activity of the RapA and RapE phosphatases is inhibited by the PhrA and PhrE pentapeptides, respectively, as part of a regulatory circuit that allows the cell to make the decision whether to remain in vegetative growth or initiate the sporulation process (Perego and Hoch, 1996; Jiang, Grau and Perego, 2000). To date, there is no evidence for auxiliary proteins involved in modulating the activity of Spo0E, YnzD and YisI. It is, however, intriguing that the carboxy-terminal truncated forms of Spo0E found in the spo0E11 and spo0E94 mutants are hyperactive in dephosphorylating Spo0A∼P. This suggested that the C-terminal 25 residues of Spo0E may have an inhibitory role. This could be achieved by means of an intramolecular mechanism, by interaction with an unknown auxiliary protein acting as regulator or by targeted proteolysis. The YnzD and YisI proteins lack the C-terminal inhibitory region, suggesting that perhaps their presence in the cell is mainly controlled by transcriptional mechanisms that do not necessitate additional regulatory elements.

With the identification of YisI and YnzD, the total number of aspartyl phosphate phosphatases acting on the phosphorelay is now raised to six. This is the only known bacterial signal transduction system whose complexity is depicted by multiple kinases (KinA, B, C, D and E) (M. Jiang et al., 2000) providing input entries for activating signals, and multiple phosphatases providing access to negative regulatory signals. Notably, the sporulation histidine kinases do not possess any specific phosphatase activity towards the response regulators that results in production of inorganic phosphate as often observed in other histidine kinases (M. Jiang et al., 2000). Dephos-phorylation, or resetting of the system is therefore the function acquired by the Rap and Spo0E phosphatases, perhaps as an evolutionary adaptation mechanism able to regulate the complexity of the sporulation pathway. It is indeed intriguing that Rap and Spo0E orthologues are found so far only in spore-forming Gram-positive organisms and none has yet been found in Gram-negative bacteria, archaea or eukaryotes (M. Perego and J. A. Hoch, unpublished data).

Experimental procedures

Bacterial strains and growth conditions

The Bacillus subtilis strains used in this study are listed in Table 3. Sporulation assays were carried out in Schaeffer's sporulation medium (Schaeffer, Millet and Aubert, 1965), cells were grown for the time indicated and then treated with CHCl3 before plating onto Schaeffer's agar plates. Cultures for β-galactosidase assays were grown in Schaeffer's sporulation medium, Luria–Bertani broth (LB), Bacto Antibiotic medium 3 (Penassay broth) or Spizizen minimal medium (MT) (Anagnostopoulos and Spizizen, 1961) as previously described (Ferrari et al., 1988). β-galactosidase activity was expressed in Miller Units (Miller, 1972). Antibiotics were used at the following concentrations: chloramphenicol 5 µg ml−1, erythromycin 25 µg ml−1, kanamycin 2 µg ml−1, spectinomycin 50 µg ml−1.

Table 3. Bacillus subtilis strains used in this study.a
StrainRelevant genotype
  • a.

    All strains are a derivative of JH642 and therefore carry the trpC2 and phe1 auxotrophic markers.

JH642wild type
JH646 spo0A12
JH12546 spo0A12 abrB::Tn917
JH12745 spo0E::cat
JH11843 amyE::ynzD–lacZ (pYNZD6)
JH11844 ynzD:: ynzD–lacZ (pYNZD12)
JH11857 amyE::yisI–lacZ (pYISI22)
JH11858 yisI:: yisI–lacZ (pYISI17)

E. coli DH5α was used for plasmid construction and propagation.

Plasmid constructions

The vectors used in this study were the following: pHT315, a multicopy E. coliB. subtilis shuttle vector (Arantes and Lereclus, 1991); pJM783, an isotopic lacZ transcriptional fusion vector; pDH32, an ectopic (amyE) transcriptional fusion vector; pJM114, a Km cassette vector derived from pBluescript (Stratagene); pJM105A, a Cm cassette vector, derived from pBluescript (Perego, 1993).

Plasmid pRM85 was isolated from a B. subtilis chromosomal library constructed in the multicopy shuttle vector pHT315. Library construction has been described previously (Wang et al., 1997). This plasmid was subject to nucleotide sequence analysis at the 5′ and 3′ terminal end in order to identify the cloned chromosomal region on the B. subtilis genome database. Plasmids pRM85-1.9, pRM85-1.2, pRM85-1.5 and pRM85V were derived from pRM85 by subcloning fragments into pHT315. The fragments carried by plasmids pYISI 1, pYISI 17, pYISI 23 and pYISI 24 were generated by polymerase chain reaction (PCR) amplification of JH642 chromosomal DNA using oligonucleotides that introduced appropriate restriction sites at the 5′ and 3′ ends of the fragments. The fragment carried by pYISI 1 was cloned in the BamHI site of pHT315. The fragment carried by pYISI 17, pYISI 22 and pYISI 23 were cloned as EcoRI–BamHI in pJM783, pDH32 and pHT315 respectively. Plasmid pYISI 24 is a derivative of pJM 114 carrying a BamHI–EcoRI fragment and a XmnI–XhoI fragment cloned upstream and downstream of the kanamycin cassette, respectively. Plasmid pYNZD13 and pYNZD17 were obtained by cloning the PCR amplified chromosomal region carrying the ynzd locus as a EcoRI–BamHI fragment in pHT315. Plasmids pYNZD6 and pYNZD12 carried the same fragment of pYNZD13 cloned in the EcoRI–BamHI sites of pDH32 and pJM783, respectively. Plasmid pYNZD19 is a derivative of pJM105A in which an XhoI–HindIII fragment and a EcoRI–BamHI fragment were cloned upstream and downstream of the chloramphenicol cassette respectively.

Protein expression and purification

Fragments carrying the yisI and ynzd coding sequence were generated by PCR amplification from chromosomal DNA of JH642 using oligonucleotides that introduced a NdeI site at the 5′ end and a BamHI site at the 3′ end. The fragments were cloned in the pET16b expression vector (Novagen), therefore generating an extension of 10 histidine codons at the 5′ end of the gene. Constructs were verified by sequence analysis. Expression of the YisI and YnzD proteins was obtained in the E. coli BL21 (DE3) plysS and BL21 (DE3) strains respectively. Induction was carried out at OD600 of 0.6 with 2 mM isopropyl-β-d-thiogalactopyranoside. Induction was carried out for 2.5 h at 37°C. The proteins were purified by affinity chromatography on Ni-NTA agarose (Qiagen) using a buffer containing 50 mM Tris-HCl pH 8.5, 300 mM KCl, 0.1 mM PMSF (phenylmethylsulphonylfluoride) and 5 mM imidazole. A step elution method with imidazole at 20 mM, 50 mM, 100 mM and 200 mM was used to wash the columns and elute the proteins. Proteins were then dialysed in 25 mM Tris-HCl pH 8.5, 300 mM KCl, 1 mM EDTA and concentrated using Centriprep 3 concentrators (Amicon). Purification of KinA, Spo0F, Spo0B and Spo0A was as previously described (Grimsley et al., 1994; Zapf, Hoch and Whiteley, 1996; Zhou et al., 1997; Grimshaw et al., 1998).

Purification of 32P-labelled Spo0A∼P

A Spo0A protein (40 µM) modified to contain six histidine residues at the C-terminal end (unpublished data) was labelled in a 250 µl reaction mixture containing KinA (0.3 µM), Spo0F (1 µM), Spo0B (1 µM), ATP (900 µM) and [γ32P]-ATP (900 µCi) (6000 Ci mM−1, NEN) in the standard phosphorylation buffer (see below). After 1 h incubation at room temperature, the reaction was mixed with 0.5 ml of Ni-NTA agarose (Qiagen) equilibrated with binding buffer containing 500 mM Tris-HCl and 300 mM KCl. After 2 h incubation at 4°C, the resin was packed in a minicolumn and washed with 25 ml of binding buffer. Two additional washes with binding buffer containing 10 mM and 100 mM imidazole were carried out before eluting the protein in buffer containing 50 mM Tris-HCl, 50 mM KCl, 1 mM EDTA and 200 mM imidazole. Fractions (500 µl) were collected and 10 µl of each were run on a 15% SDS-acrylamide gel. The gel was directly exposed to a Kodak X-Omat RP film and the peak fraction used in dephosphorylation assays after determining protein concentration by the Bradford-based Bio-Rad protein assay (Bio-Rad).

Dephosphorylation assays

The ability of YisI and YnzD to dephosphorylate Spo0A∼P was assayed in a phosphorelay reaction containing KinA (0.2 µM), Spo0F (5 µM), Spo0B (1 µM), Spo0A (5 µM), ATP 300 µM) and [γ32P]-ATP (3.4 µCi ml−1; 6000 Ci mM−1, NEN). The reaction was carried out for 1 h at room temperature in buffer containing 50 mM EPPS (N-(2 hydroxyethyl) piperazine – N′-(3-propanesulfonic acid)) pH 8.5, 20 mM MgCl2, 0.1 mM EDTA, 5% glycerol. The purified YisI and YnzD proteins were then added for 30 min at room temperature at the concentrations shown in the figures. Reactions were stopped by the addition of loading buffer and run on 15% SDS/glycine/polyacrylamide gels at constant current (25 mAmp) for 1.5 h. Gels were immediately exposed to Kodak X-Omat RP films overnight at −80°C. Gels were then exposed to a PhosphorImager screen (Molecular Dynamics) and analysed using the ImageQuant software.

Dephosphorylation of purified Spo0A∼P was carried out in a 60 µl volume of 50 mM EPPS, 5% glycerol in the presence or absence of 40 mM MgCl2. Spo0A∼P was at 0.6 µM while YisI and YnzD were added at 2 µM final concentration. The reaction was carried out at room temperature and 15 µl aliquots were withdrawn at the times indicated in the figure. The reactions were stopped by the addition of loading dye and analysed on 15% SDS/glycine/polyacrylamide gel as described above.

Acknowledgements

This research was supported, in part, by Public Health Service grant GM55594 from the National Institute of General Medical Sciences, National Institutes of Health. The Stein Beneficial Trust supported, in part, oligonucleotide synthesis. This is publication 13845-MEM from The Scripps Research Institute. Acknowledgement is made to Dr Roberto Grau for the construction of pRM plasmids and to Dr Shirley Hanley for the construction of plasmid pYISI 1.

Ancillary