ABSTRACT
SpoIIAA, a phosphorylatable protein, is essential to the regulation of ςF, the first sporulation-specific transcription factor of Bacillus subtilis. The solution structure of SpoIIAA has recently been published. Here we examine four mutant SpoIIAA proteins and correlate their properties with the phenotypes of the corresponding B. subtilis mutant strains. Two of the mutations severely disrupted the structure of the protein, a third greatly diminished the rate of its phosphorylation and abolished dephosphorylation, and the fourth left phosphorylation unaffected but reduced the rate of dephosphorylation about 10-fold.
Early in sporulation, Bacillus subtilis divides asymmetrically to produce a small compartment, the prespore, and a large compartment, the mother cell. Differential gene expression in the prespore and the mother cell depends on sigma factors whose activity is confined to one or other compartment and is limited temporally (8, 13). The first sporulation-specific sigma factor, ςF, becomes active in the prespore soon after asymmetric septation and is responsible (directly or indirectly) for the activation of the remaining sigma factors both in the prespore and the mother cell (reviewed in reference19). The activity of ςF is regulated by three proteins, SpoIIAB, SpoIIAA, and SpoIIE. SpoIIAB is an inhibitor of ςF and also a specific protein kinase, SpoIIAA is the substrate of the kinase, and SpoIIE is a specific protein phosphatase that hydrolyzes phosphorylated SpoIIAA (reviewed in reference 19). In the predivisional cell ςF is kept in an inactive complex with SpoIIAB (3, 5). Two models have been suggested to account for the liberation of active ςF from this complex (7, 15). These models, although differing to some extent from one another, both propose that the activity of ςF depends on the concentration of non-phospho-SpoIIAA in the cell or compartment, which in turn will depend on the balance between the activities of the SpoIIAB kinase and the SpoIIE phosphatase.
The phenotypes of several strains with spoIIAA mutations have been described. However, although the corresponding amino acid changes in SpoIIAA are identified, there is for most of the strains no satisfactory account of the mutant phenotype in molecular terms. One possible explanation for the phenotype might be that the SpoIIAA proteins made by these mutants differ from wild-type SpoIIAA in their response to the SpoIIAB kinase or the SpoIIE phosphatase (or both). Alternatively, the mutant SpoIIAA proteins might differ radically in structure from the wild type and be unable to fold correctly. The solution structure of SpoIIAA, a 117-residue protein, has been published recently (11). It shows a fold consisting of a four-stranded β-sheet and four α-helices, with the site of phosphorylation, Ser-58, at the N terminus of the second helix. In the publication just referred to (11) tentative suggestions were made to account for the phenotype of the known SpoIIAA mutations, proposing that the β3-α2 loop (containing the phosphorylation site), the β2-α1 loop, and the α3 helix were all involved in the interaction between SpoIIAA and SpoIIAB.
In this paper we describe studies of the SpoIIAA proteins present in four mutant strains: strain 42 (SpoIIAAG95D) (18), strain 69 (SpoIIAAG62D) (18), strain 565 (SpoIIAAG20S,S58N) (2), and strain 568 (SpoIIAAG76E) (2). The incidence of spore formation in strain 42 is about 0.1% of that in the wild type, while the other three mutants are completely asporogenous (2, 18). Figure 1 shows the structure of the wild-type protein (11) and the positions of the residues affected in these mutants.
The structure of SpoIIAA, with the residues whose mutations are discussed in the text indicated.
Accumulation of wild-type and mutant SpoIIAA proteins in B. subtilis.One possible explanation of the phenotypes of the mutant strain(s) is that the mutant SpoIIAA protein(s) might be misfolded and thus rendered insoluble. To examine this possibility, we grew the wild type and the four spoIIAA mutant strains of B. subtilis, resuspended them in a sporulation medium, took 1-ml samples 2 h later, and harvested the cells by centrifugation. Cell pellets were resuspended in 330 μl of an ice-cold sonication buffer (50 mM Tris-HCl [pH 7.5], 10% glycerol, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride) and subjected to sonication (2 by 15 sec) in an ice-water bath. Additional phenylmethylsulfonyl fluoride (to bring the concentration to 2 mM) was added. After centrifugation for 5 min at 10,000 × g , the supernatants were subjected to electrophoresis on sodium dodecyl sulfate-polyacrylamide (15%) gels and immunoblotted with purified anti-SpoIIAA antibodies (9). The blots showed that the soluble fraction of wild-type cells and that of the cells carrying SpoIIAAG62D or SpoIIAAG95D contained substantial quantities of SpoIIAA (Fig.2). By contrast, in the soluble fraction of the mutants with SpoIIAAG20S,S58N or SpoIIAAG76E, SpoIIAA was barely detectable (Fig. 2). When these experiments were repeated with sonicated cells that had not been fractionated by centrifugation, mutant SpoIIAA was found in the strains with SpoIIAAG20S,S58N or SpoIIAAG76E, but in both strains the quantity of SpoIIAA present was only 10 to 20% of that in the cells with SpoIIAA+, SpoIIAAG62D, or SpoIIAAG95D (results not shown). (The previously published results in which SpoIIAA was detected in strains with SpoIIAAG20S,S58N or SpoIIAAG76E [2] failed to distinguish between the soluble and insoluble fractions of the cell.) We concluded that when SpoIIAAG20S,S58N and SpoIIAAG76E are made in B. subtilis, a certain quantity of these mutant proteins accumulates in an insoluble form but that most is degraded. We presume that this degradation is by the B. subtilisequivalents of the Escherichia coli proteases that are responsible for the proteolysis of misfolded proteins (10).
Content of wild-type and mutant SpoIIAA proteins detected by immunoblotting in the soluble fraction of cells 2 h after resuspension in sporulation medium. The lanes and the quantities of soluble protein loaded were as follows: lane 1, wild type (1.9 μg); lane 2, strain 42 (0.9 μg); lane 3, strain 69 (1.2 μg); lane 4, strain 565 (1.7 μg); and lane 5, strain 568 (1.6 μg). The arrow shows the position of a SpoIIAA marker.
Overproduction and purification of the mutant proteins.We overproduced the mutant SpoIIAA proteins in E. coli to study their biochemical properties. PCR was used to amplify thespoIIAA gene from the four mutant strains of B. subtilis. Each PCR product was cloned into expression vector pET-3a (Novagen), the DNA sequence was verified, and the mutantspoIIAA was overexpressed in E. coli BL21(DE3) after induction with IPTG (isopropyl-β-d-thiogalactopyranoside). The total quantities of mutant proteins produced were, in all cases, comparable with that of the wild-type protein. The mutant proteins SpoIIAAG95D and SpoIIAAG62D were readily purified from the soluble cell extract by the procedure developed previously in this laboratory, which yields SpoIIAA at a purity greater than 95% (4). By contrast, SpoIIAAG20S,S58N and SpoIIAAG76E accumulated in the overproducing cells in inclusion bodies, and this behavior persisted even when the temperature of growth after induction with IPTG was varied between room temperature and 37°C. SpoIIAAG20S,S58N and SpoIIAAG76E were solubilized from inclusion bodies with 6 M guanidine hydrochloride. The proteins were refolded by a gradual dilution of the guanidine to 0.09 M. After the proteins had been adsorbed on DEAE-Sepharose and eluted with a buffer containing 300 mM NaCl, gel filtration of these two proteins on Superdex 75 showed that both of them were excluded from the gel and therefore that their estimatedMr was greater than 70 kDa. TheMr of the wild-type protein is 13 kDa, and it is known to be monomeric. We concluded that the structures of SpoIIAAG20S,S58N and SpoIIAAG76E had been altered in such a way as to cause the proteins to aggregate. A further indication of this aggregation came from electrophoresis of purified SpoIIAAG20S,S58N and SpoIIAAG76E on a 10% native gel, which showed in each case a ladder of multiple bands.
Accounting for the phenotypes of strain 565 and strain 568.The results suggest that strains 565 and 568 contain little or no soluble SpoIIAA. Lacking an alternative binding partner in these strains, SpoIIAB will bind exclusively to ςF, and ςF activity will be totally inhibited. Strains carrying the doubly mutant protein SpoIIAAG20S,S58N show no detectable expression of the two ςF-dependent genes gprand spoIIIG and are partially complemented by the wild-typespoIIAA gene (2). By contrast the single mutation leading to the replacement S58N is dominant to the wild type, strains with this mutation express gpr and spoIIIG to a small but perceptible extent (2), and a SpoIIAA protein containing the single substitution S58N is soluble and easily purified (12). It is clear therefore that in the doubly mutant strain, it is G20S that causes the protein to be misfolded and deprives it of normal function.
In view of the aggregation of SpoIIAAG20S,S58N and SpoIIAAG76E inE. coli, we were unable to study the biochemical properties of these mutant proteins further. The remainder of the experiments described here concern the phosphorylation and dephosphorylation of SpoIIAAG62D and SpoIIAAG95D, studied with proteins purified from overexpressing strains of E. coli as described above. (Studies of noncovalent binding of the SpoIIAA proteins with SpoIIAB in the presence of ADP [1, 4] revealed no significant difference between SpoIIAA+, SpoIIAAG62D, and SpoIIAAG95D [results not shown].)
Phosphorylation of the mutant proteins catalyzed by SpoIIAB.SpoIIAB catalyzes the phosphorylation of wild-type SpoIIAA by ATP (16) in a reaction in which the phosphate is transferred to Ser-58 of the substrate (17). To see how SpoIIAAG62D and SpoIIAAG95D responded in this reaction, we incubated each of them with SpoIIAB and [γ-32P]ATP and monitored the incorporation of the labelled phosphate into acid-insoluble material, as described earlier (15). As Fig.3A shows, SpoIIAAG95D was indistinguishable from wild-type SpoIIAA in its response to the SpoIIAB protein kinase. We have previously reported that the time course of phosphorylation was biphasic for wild-type SpoIIAA (see reference15), and we note that the same is true of SpoIIAAG95D (Fig. 3A). By contrast, SpoIIAAG62D was phosphorylated very slowly, with an apparent velocity constant about one-half of that for the steady-state phosphorylation of wild-type SpoIIAA (Fig. 3B). It is probably relevant that residue 62 is one turn of an α-helix from S58, the residue that is phosphorylated in this reaction (17). We suggest that the presence of a negatively charged residue at position 62 hampers the phosphorylation of SpoIIAA by SpoIIAB, as predicted previously (11).
Phosphorylation of wild-type and mutant SpoIIAA proteins by SpoIIAB. (A) Wild-type SpoIIAA (squares) and SpoIIAAG95D (triangles). (B) Wild-type SpoIIAA (squares) and SpoIIAAG62D (circles).
Hydrolysis of the phosphorylated mutant SpoIIAA proteins by SpoIIE.A method for overproducing and purifying full-length SpoIIE from E. coli has been developed recently (14). This SpoIIE preparation achieves a rapid and complete hydrolysis of phosphorylated wild-type SpoIIAA (14), a reaction that can be monitored by taking advantage of the difference in mobility between free and phosphorylated SpoIIAA on nondenaturing gels. To see how the SpoIIAA mutants behaved in this assay, we prepared phosphorylated samples of wild-type SpoIIAA and of the two mutant proteins SpoIIAAG62D and SpoIIAAG95D, by incubating each of the three purified proteins with a one-sixth molar ratio of SpoIIAB in the presence of excess ATP and then removing the ATP and SpoIIAB by gel filtration over Superdex 75. We exposed the purified phosphorylated SpoIIAA proteins to SpoIIE in a 30-μl mixture containing 3 μl of a 10× dephosphorylation buffer (1 M Tris-HCl [pH 7.5], 500 mM NaCl, 10 mM dithiothreitol, 10 mM MnCl2) at 30°C and monitored the appearance of the nonphosphorylated products by running the reaction mixtures on 10% nondenaturing polyacrylamide gels.
The results (Fig. 4) show a striking difference in the rate of hydrolysis between the phosphorylated wild-type protein and phosphorylated SpoIIAAG95D (note the difference in the time scale of the experiment with the wild-type protein and the experiments with the mutant proteins). Still more strikingly, no hydrolysis of phosphorylated SpoIIAAG62D could be detected, even after overnight incubation with SpoIIE (Fig. 4C). To confirm that these differences in the rate of hydrolysis were not due to the adventitious presence of small quantities of SpoIIAB kinase in the phosphorylated SpoIIAA preparations, we ran samples of phosphorylated wild-type SpoIIAA and of each of the phosphorylated mutant SpoIIAA proteins on a sodium dodecyl sulfate gel and probed the gel by Western immunoblotting with an anti-SpoIIAB antibody. No SpoIIAB was found, whereas 50 ng of SpoIIAB run in a neighboring lane was easily detected (results not shown). We conclude that the differences in the rate of hydrolysis of the phosphorylated proteins are a genuine characteristic of these proteins. We surmise that the face of the protein on which both G62 and G95 lie (Fig. 1) may be involved in the interaction with SpoIIE and that a negative charge at residue 62 may abolish the reaction in which a phosphate group attached to S58 is cleaved.
Hydrolysis of phosphorylated wild-type and mutant SpoIIAA proteins by SpoIIE. After electrophoresis in nondenaturing conditions the gels were stained with Coomassie blue and destained with 30% methanol–10% acetic acid. In each panel, lane M contains, as a marker, the relevant nonphosphorylated SpoIIAA protein, which is the expected product of the hydrolysis. (A) Wild-type SpoIIAA-phosphate, with samples taken at the times shown between 0 and 90 min. (B) SpoIIAAG95D-phosphate, with samples taken at the times shown between 0 and 20 h. (C) SpoIIAAG62D-phosphate, with samples taken at the times shown between 0 and 20 h.
Accounting for the phenotype of strain 69.The results shown in Fig. 3B suggest that SpoIIAAG62D will be a substrate for the kinase activity of SpoIIAB in the predivisional cell. When SpoIIAB acts on wild-type SpoIIAA, a given quantity of SpoIIAB phosphorylates an equimolar quantity of SpoIIAA in approximately 20 min at 30°C (15) and therefore, presumably, in about 10 min at 37°C. The kcat for SpoIIAB in phosphorylating SpoIIAAG62D is about two times less than that for wild-type SpoIIAA (Fig. 3B), but since the intracellular concentration of SpoIIAA in the predivisional cell is less than that of SpoIIAB (15), we could expect that by the time of asymmetric septation, all of the SpoIIAAG62D protein in the cell will be phosphorylated. Thereafter all the SpoIIAB will be available to inhibit ςF. After asymmetric septation SpoIIAAG62D will all remain in the phosphorylated form, since it is not a substrate for SpoIIE, and therefore no SpoIIAA will be available to liberate ςF from the complex with SpoIIAB. In these respects, SpoIIAAG62D resembles the SpoIIAAS58T protein (carrying a mutation at the phosphorylation site) described by Duncan et al. (6, 7), which is a substrate for phosphorylation by the SpoIIAB kinase but whose phosphorylated form cannot be hydrolyzed by SpoIIE. The complete absence of active ςF from strain 69 would then account for its unequivocally Spo− phenotype.
Accounting for the phenotype of strain 42.In cells with SpoIIAAG95D the SpoIIAA protein is phosphorylated by SpoIIAB as rapidly as in the wild type (Fig. 3A), with the result that, as described for cells with SpoIIAAG62D, ςF will be inhibited in the predivisional cell. After asymmetric septation, the phosphorylated SpoIIAAG95D will become a substrate for hydrolysis by SpoIIE. But since hydrolysis is extremely slow (Fig. 4B), nonphosphorylated SpoIIAA will be made available at a very low rate. Consequently very little of the ςF will be activated—just enough to lead to the formation of a few spores in this mutant strain.
Conclusion.The phenotypes of all four of thespoIIAA mutants studied here can be explained in terms of the three factors mentioned at the beginning of this paper—changes in protein folding, changes in response to SpoIIAB, and changes in response to SpoIIE. The results of the analysis of strains 69 and 42 clearly demonstrate the importance of both phosphorylation and dephosphorylation of SpoIIAA in the regulation of ςF and point the way to detailed investigations of the SpoIIAA protein by means of site-directed mutagenesis.
ACKNOWLEDGMENTS
Daniela Barillà and Isabelle Lucet contributed equally to this work.
We thank Julie Wickson for outstanding technical assistance, D. Comfort for an analysis of the ability of the SpoIIAA structure to tolerate the changes described here, J. Errington for valuable discussions, and R. Losick for helpful comments on the manuscript.
We acknowledge the Biotechnology and Biological Sciences Research Council for financial support.
FOOTNOTES
- Received 10 November 1998.
- Accepted 11 April 1999.
- Copyright © 1999 American Society for Microbiology
