University of Oxford Crest

Laboratory of Molecular Biophysics
Laboratory Journal 2002
Richard J. Lewis

Return to: Contents.

Richard J. Lewis


The stress response of B. subtilis


Sujit Dutta

In collaboration with Chien-Cheng Chen, Olivier Delumeau and Michael Yudkin, Microbiology Unit, University of Oxford.

Introduction


Figure 1
Figure 1. Schematic of the  sigma B regulon of B. subtilis. ...more
Variations in environmental factors, such as reduced aeration, extremes of temperature and the availability of essential nutrients, restrict microbial growth. As a result, bacteria spend much of their life cycle in stationary phase (Msadek, 1999). The requirement for acclimatisation to the environment has forced bacteria to develop a series of complex adaptive responses to stress. One of the initial responses to stress of some Gram+ bacteria, including Bacillus subtilis, is to synthesise a large and diverse family of sigma B-dependent general stress proteins (Hecker & Volker, 2001). sigma B is an alternative RNA polymerase subunit conferring promoter specificity thus directing gene expression in a defined manner.

sigma B is kept under strict control by the gene products found in the sigB operon, which may be functionally divided into "upstream" and "downstream" modules (Figure 1). Each module comprises an ATP-dependent serine/threonine protein kinase, a phospho-serine/phospho-threonine protein phosphatase, and a "switch" protein substrate for the kinase and phosphatase. In both modules, the kinases also participate in alternative protein:protein recognition events. Together, the binding partner for the kinase and the phosphorylation state of the switch molecules control the activity of sigma B.

1.1 The downstream

module
Figure 2
Figure 2. Native gel electrophoresis ...more
In the downstream module, the kinase RsbW, the substrate for which is the switch protein, RsbV, also forms a protein:protein complex with sigma B, thus acting as an anti-sigma factor. RsbP is a PP2C-type protein phosphatase, which also encodes a PAS sensory domain, whose substrate is the phosphorylated form of RsbV, RsbV-P. RsbP is somehow activated by RsbQ, a hydrolase, and the genes encoding these proteins have been cloned, and overexpressed, but have not yet been purified. In collaboration with Dr. Olivier Delumeau and Professor Michael Yudkin of the Microbiology Unit, we have established the KM for ATP and KI for ADP of RsbW (the values of which suggest that the in vivo enzymatic activity of RsbW is regulated by the intracellular concentration of adeneine nucleotides), the stoichiometry of the RsbW-sigma B and RsbW-RsbV complexes, the affinity of RsbW for its two alternative partners (Figure 2) and the concentration of these three proteins in the cell, before and during stress. Taken together, the data are supportive of a mechanism where the kinetic and equilibrium constants of RsbW regulate the stress response of B. subtilis (Delumeau et al., 2002).
 
 
In addition to the biochemical studies, we have also obtained crystals of RsbW, both in the presence and absence of the non-hydrolysable ATP analogue, AMP-PNP. However, despite appreciable efforts, the quality of these crystals remains poor, and diffract maximally only to low resolution, some 7 Å. We have thus turned our efforts to determine the structures of the protein:protein complexes that RsbW participates in. The over-expression of any sigma factor in E. coli tends to be toxic to the host: this can be rationalized by the high degree of conservation of RNA polymerase subunits between bacterial species, and the probability that core E. coli RNA polymerase uses the heterologous sigma factor to direct physiologically unsupportable patterns of gene transcription. Consequently, sigma factors are notoriously difficult to express in E. coli. RsbV is a poorly-behaved protein in solution, with a tendency to aggregate. Hence we are designing co-expression strategies for RsbW with its partner molecules in the hope that the presence of RsbW in the cell abrogates the sigma B toxicity, and the RsbV aggregation problems.  

1.2 The upstream module

In some respects, the RsbU-RsbS-RsbT components of the upstream module function similarly to their downstream RsbP-RsbV-RsbW homologues. In the presence of RsbS-P, the RsbT kinase from the upstream module forms a protein:protein complex with the RsbU PP2C-type phosphatase of the downstream module, stimulating its phosphatase activity at least 20-fold towards RsbV-P (Kang et al., 1998). The product of this reaction, RsbV, liberates sigma B from its inactive complex with RsbW and induces expression of the general stress proteins. Thus, RsbT links the upstream and downstream modules, by binding to RsbU and stimulating its enzymatic properties towards its substrate.
Figure 3
Figure 3. Trypsinolysis of full-length RsbU. ...more


 Analysis of the amino-acid sequence of RsbU predicts that the C-terminal domain adopts the same fold as PP2C-type phosphatases (Das et al., 1996), whereas the N-terminal domain shares no sequence homology to other proteins, and thus whose function and fold is unknown. Limited proteolytic digestion of intact RsbU yields two fragments of approximate molecular mass 25 and 12 kDa (Figure 3). Amino acid sequencing indicates that trypsin cleaves in the region of Leu112, generating fragments with masses in agreement to those observed by SDS-PAGE. Sequence alignment of RsbU with the RsbX phosphatase, which encodes solely a PP2C-type domain, reveals that trypsin cleavage occurs in RsbU at positions which align closely to the first few amino acids of RsbX, indeed, Leu111 of RsbU aligns to Met1 of RsbX. Based on these results, and also in the absence of crystals of full-length RsbU, fragments of rsbU which encode the N-terminal or C-terminal domains of RsbU have been cloned and the proteins purified. Both of these domains have now been crystallized (Figure 4, 5), with native, orthorhombic crystals of N-RsbU diffracting X-rays to at least 1.6 Å (Dutta & Lewis, 2002). Experimental phasing is underway, by MAD/MIR methods.

        
Figure 4
Figure 5
Figure 4. Photomicrograph of crystals of N-RsbU. ...more
 Figure 5. Photomicrograph of crystals of C-RsbU. ...more

Figure 6
Figure 6. Native gel electrophoresis  ...more
Preliminary solution experiments indicate that the main RsbT-binding determinants in RsbU are found in its N-terminal domain and with colleagues inMicrobiology (C.-C. Chen, O. Delumeau and M. Yudkin) we are measuring a variety of kinetic and equilibrium constants for RsbT and RsbU. RsbT is poorly soluble (<1 mg/ml), and thus we have successfully made co-expression constructs in order to prepare enough of both of these complexes for crystallization trials. The phosphatase activity of the C-terminal fragment of RsbU towards RsbV-P is no longer regulated by the N-terminal domain, and cannot be stimulated by the presence of RsbT (Figure 6). Hence the function of N-RsbU is in molecular recognition and recruitment of the stimulatory factor, the kinase RsbT, and in the (de)regulation of the phosphatase activity of RsbU. Intriguingly, the human pathogen Staphylococcus aureus encodes a sB regulon that is similar to that of B. subtilis, but does not encode an RsbT orthologue, and thus the regulation of RsbU must differ in this organism. This is an interesting difference in physiology between phylogenetically closely-related members of the same bacterial group, which we are investigating biochemically as well as structurally.

   Our in vitro data suggest that in vivo, the N-terminal domain of RsbU plays an important regulatory role. One possibility is that in the absence of N-RsbU the phosphatase activity of C-RsbU could be unchecked, sB would not be sequestered in a non-productive complex with RsbW and the stress response would be permanently induced. It is possible to monitor sB activity in the cell by fusing a reporter gene, lacZ, to a sB-dependent gene, in this case, ctc. Accordingly, in collaboration with O. Delumeau, M. Yudkin and Dr. Chester Price (University of Davis, CA), we have made appropriate in-frame deletions at the 5' end of rsbU and replaced the functional copy of rsbU in the chromosome with this truncated form. The phenotype of this strain of B. subtilis is consistent with our in vitro studies, in that no sB activity can be measured, implying that no induction of the stress response occurs. We conclude that the specific activity of C-RsbU towards RsbV-P is low, and that in the context of full-length RsbU, N-RsbU does not function as an inhibitory domain, but is probably required for both substrate recognition and allosteric activation by kinase-binding.
Figure7
Figure 7. Negatively-stained TEM image of the 'stressosome'. ...more
The presence of RsbR in the upstream module further complicates signal transduction. This protein has an N-terminal domain of unknown function, and a C-terminal domain with similarity to RsbS. RsbR is a substrate for the RsbS-specific kinase, RsbT, but no phosphatase has yet been identified that dephosphorylates RsbR-P. Cell extracts containing RsbR, RsbS and RsbT have been shown to co-elute from a gel filtration column, suggesting that these three proteins form some sort of complex. In collaboration with Dr. Robin Harris (University of Mainz) we have examined negatively-stained samples of this complex by transmission electron microscopy, which reveals its architecture (Figure 7). The complex appears to be a hollow sphere. We assume that RsbT decorates the external surface of the sphere, and that RsbR and RsbS form the scaffold, the interpretation of which is supported by the fact that RsbT has to switch from this set of interactions during stress to activate RsbU, and the observation that RsbR and RsbS can form a similar complex in the absence of RsbT. This RsbR:RsbS complex has been analysed by sedimentation velocity, in conjunction with Dr. David Scott, University of Bristol, and has a sedimentation coefficient of 25S, corresponding to a molecular mass of approximately 0.8 MDa. We are currently pursuing higher resolution structural information on the 'stressosome', which we hope will help to explain why these proteins behave in such a way!


                                

References

 Das, A. K., Helps, N. R., Cohen, P. T. W. & Barford, D. (1996). EMBO J. 45, 209-221.
Delumeau, O., Lewis, R. J. & Yudkin, M. D. (2002). J. Bacteriol. 184, 5583-5589.
Dutta, S & Lewis, R. J. (2002). Acta crystallogr. In press.
Hecker, M. & Volker, U. (2001). Adv. Microb. Physiol. 44, 35-91.
Kang, C. M., Vijay, K. & Price, C. W. (1998). Mol. Microbiol. 30, 189-196
Msadek, T. (1999). Trends in Microbiol 7, 201-207.


2 DNA-binding Studies on the B. subtilis Transcriptional Regulator SpoVT


In collaboration with Tran Cat Dong and Simon Cutting, Royal Holloway University of London.

Gene expression during spore formation in Bacillus subtilis is controlled by the temporal and spatial action of five developmental sigma factors (Errington, 1993). At least three regulons controlled by the action of three of these sigma factors, sE, sK and sG are subject to further modulation brought about by the action of the DNA-binding proteins, SpoIIID (in conjunction with sE), GerE (sK) and SpoVT (sG).  The action of these secondary transcription factors ensures the fine-tuning of developmental gene expression and hence the high fidelity of spore formation.  The transcription factor SpoVT, modulates gene expression controlled by sG only in the forespore chamber of the developing cell (Bagyan et al., 1996).  The forespore is the enclosed chamber of the sporangial cell that is destined to become the dormant endsopore, released from the mother cell chamber.  As a general modulator of forespore-specific gene expression, SpoVT must serve an important role in the final stages of spore development since deletion of spoVT leads to the formation of aberrantly formed spores with defective resistance properties. In order to understand the mode by which SpoVT regulates gene expression, we have begun to study it by genetics, biochemistry and biophysics.
Although it is clear from genetic studies that SpoVT is a transcriptional regulator (Bagyan et al., 1996), in our own DNA-binding studies we cannot demonstrate full length SpoVT binding either specifically or non-specifically to DNA. It is most likely that in vivo, SpoVT requires an interaction with another protein to bind to DNA, this protein could be RNA polymerase itself, or perhaps a small molecule co-factor of unknown composition. The N-terminal 53 residues of SpoVT are believed to encode its DNA-binding function, based on sequence homology to the transition state sentinel, AbrB, and we have been able to demonstrate that N-SpoVT, truncated at residue 53 binds DNA, but non-specifically. Our genetic studies show that the C-terminal domain of SpoVT is essential for the function of SpoVT (Tran et al., 2002), which is presumably required for the interaction with the auxiliary factors that confer DNA-binding specificity to SpoVT, however, we have no clues about its function from its amino acid sequence. Unfortunately, no crystals have yet been grown of SpoVT, or appropriate fragments, the structures of which might help unravel the molecular mechanism of transcriptional regulation by SpoVT.


References


Bagyan, I., Hobot, J. & Cutting, S. M. (1996) J. Bacteriol. 178, 4500-4507.
Errington, J. (1993) Microbiol. Rev. 57, 1-33.
Tran, D. C., Cutting, S. M. & Lewis, R. J. (2002) J. Bacteriol. Submitted

Return to: Contents.

[ Lab. of Molecular Biophysics | University of Oxford | Feedback ]
Last updated: 15-MAY-2003 15:24