| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Philipps-University Marburg, Department of Biology, Laboratory for Microbiology, D-35032 Marburg, Germany
2 Max-Planck-Institute for Terrestrial Microbiology, D-35043 Marburg, Germany
3 Ernst-Moritz-Arndt-University, Medical School, Laboratory for Functional Genomics, Walther-Rathenau-Str. 49A, D-17487 Greifswald, Germany
Correspondence
Uwe Völker
voelker{at}uni-greifswald.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Supplementary material is available with the online version of this paper.
These authors contributed equally to this work.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
) that in extreme cases can induce the formation of a highly resistant endospore (Sonenshein, 2000). Changes in temperature constitute a key factor that influences cell growth and survival in the soil. Under laboratory conditions, B. subtilis is able to sustain growth in a temperature range from approximately 11 °C (Nichols et al., 1995) to 52 °C (Holtmann & Bremer, 2004). Adaptation responses of the cell are triggered when the temperature is rapidly increased or decreased within this temperature range (Hecker et al., 1996; Weber & Marahiel, 2002), or when the cells are cultivated for prolonged periods close to the upper or lower temperature limits (Brigulla et al., 2003; Holtmann et al., 2004).
The changes in gene expression that occur following a rapid increase in temperature are collectively known as the heat-shock response (Hecker et al., 1996; Schumann et al., 2002). Heat-shock proteins help the cell to refold temperature-damaged proteins via chaperones and to degrade thermally denatured polypeptides via proteases (Schumann et al., 2002). A sudden increase in growth temperature also triggers a rapid induction of a large general stress regulon (Hecker & Völker, 2001; Price, 2002) that is controlled by the alternative transcription factor SigB (Benson & Haldenwang, 1993; Völker et al., 1994). Induction of the SigB-dependent general stress regulon provides B. subtilis cells with a multiple, non-specific and pre-emptive stress resistance that includes protection against a growth-inhibiting heat shock (54 °C) (Hecker & Völker, 2001; Höper et al., 2005; Völker et al., 1999).
Heat-shock proteins such as GroEL, GroES, DnaK and GrpE are also detected at an increased level when B. subtilis is propagated close to its upper temperature limit (Holtmann et al., 2004). This growth condition also triggers a sustained high-level expression of the entire general stress regulon (Holtmann et al., 2004). Effective thermoprotection of B. subtilis can also be accomplished by the uptake of exogenously provided compatible solutes, such as glycine betaine and the amino acid glutamate (Holtmann & Bremer, 2004).
A sudden drop from the standard growth temperature (37 °C) down to 15 °C is generally used in the laboratory to elicit the cold-shock response of B. subtilis (Graumann & Marahiel, 1996; Weber & Marahiel, 2002). Such a cold shock imposes constraints on the B. subtilis cell that are different from those triggered by heat shock. A rapid decrease in temperature hampers the translational capacity of the cell due to the formation of increasingly stable mRNA secondary structures and diminished ribosome activity, results in slower protein folding and reduced protein activity, and impairs the fluidity of the cytoplasmic membrane, thereby negatively affecting transport processes (Mansilla et al., 2004; Weber & Marahiel, 2002). B. subtilis mounts different sets of cellular defence reactions to cope with these constraints.
One hallmark of the cold-shock response of B. subtilis is the high-level production of so-called cold-shock proteins (CspB, CspC, CspD) (Graumann et al., 1996). These proteins act as RNA chaperones that improve the protein-synthesis capacity of the cell by specifically binding to single-stranded nucleic acids, thereby probably reducing the formation of mRNA secondary structures. They are essential for effective protein synthesis and cell viability at both low and optimal growth temperatures (Graumann et al., 1997). Double csp deletion strains display multiple phenotypes, such as cell lysis upon entry into the stationary growth phase, aberrant nucleoid structure, and impairment in sporulation (Graumann et al., 1997; Weber et al., 2001b).
Rigidification of the cytoplasmic membrane at low temperature can be prevented in B. subtilis by either the isoleucine-dependent de novo synthesis of branched-chain fatty acids (Klein et al., 1999) or the desaturation of pre-existing fatty acids by a cold-shock-induced, membrane-bound phospholipid desaturase (Des) (Aguilar et al., 1998; Weber et al., 2001a). Both pathways result in reduced attraction between adjacent fatty acid chains, thereby appropriately adjusting membrane fluidity to the low-temperature environment (Mansilla et al., 2004). Expression of the gene (des) encoding the phospholipid desaturase is controlled by a two-component regulatory system, DesK/DesR (Aguilar et al., 2001). The membrane-embedded sensor kinase DesK monitors membrane fluidity and controls, via the phosphorylation of its cognate response regulator DesR, the transient induction of the structural gene for the 
5-lipid desaturase upon a temperature downshift (Aguilar et al., 1999, 2001). This two-component system seems to exclusively control transcription of the des gene (Beckering et al., 2002), and hence does not function as a general cold-shock sensory and regulatory device for B. subtilis. Recent genome-wide transcriptional profiling studies of cold-shocked B. subtilis cells have revealed many additional changes in gene expression, but the impact of these changes on cold adaptation is largely unexplored (Beckering et al., 2002; Kaan et al., 2002).
The drastic and rapid temperature downshifts used under laboratory conditions to provoke a cold-shock response in B. subtilis are unlikely to occur frequently in nature. In the upper layers of the soil, gradual changes in temperature prevail, and the cells have to adapt to and continue to grow under chill-stress conditions. This facet of low-temperature adaptation by B. subtilis is less well characterized than the cold-shock response. A proteomic study by Brigulla et al. (2003) of cells growing at low temperature (15 °C) revealed a delayed but sustained induction of the SigB-dependent general stress regulon. Mutations in the structural gene for the central regulator SigB profoundly affect growth under chill-stress conditions (Brigulla et al., 2003) and also reduce stationary-phase survival of cold-adapted cells (Mendez et al., 2004). Sensitization against growth at low temperature is also observed in a set of mutants lacking selected general stress proteins (Höper et al., 2005). Furthermore, the compatible solute glycine betaine not only provides protection against high growth temperature (Holtmann & Bremer, 2004) but also greatly improves growth of B. subtilis under chill-stress conditions, and can even rescue the growth defect of a sigB mutant at 15 °C (Brigulla et al., 2003). An additional aspect of chill adaptation by B. subtilis has recently been discovered by Mendez et al. (2004), who have detected a novel role for the Spo0A response regulator in cellular viability and stress survival at low temperature that is independent of its central role as the master switch in the initiation of sporulation (Hoch, 1995).
To foster our understanding of the adaptational response of the cell to continued growth at low temperature, we used a combined transcriptomic and proteomic approach to evaluate the changes that occur in B. subtilis cells cultivated at low temperature, so as to define the chill-stress stimulon on a global scale. Transcriptional analysis revealed that major physiological changes occur in the B. subtilis cell when it is actively growing under chill-stress conditions, affecting approximately 14 % of all B. subtilis genes. Our proteome analysis of chill-adapted cells suggests a major contribution of post-transcriptional regulation phenomena in the adaptation reaction of B. subtilis to growth at low temperature.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
) and several mutant derivatives. Mutations were transferred into the B. subtilis 168 strain background by transformation (Harwood & Cutting, 1990). We used two different sigB mutant derivatives of strain 168 in this study. Strain BSM29 (Brigulla et al., 2003) carries the sigB2 : : spc allele, and strain BSM27 was constructed by transferring the sigB : : HindIII-EcoRV : : cat allele from strain ML6 (Igo et al., 1987) into the 168 background and by selecting chloramphenicol-resistant transformants. The sigF (spoIIA) mutant was constructed by transforming strain 168 with chromosomal DNA isolated from strain BSG03 (spoIIA : : spc amyE : : pGK30-gsiB : : gfp cat86; U. Völker, unpublished data) and selecting for spectinomycin-resistant transformants, thereby obtaining strain IBB1. The sigB sigF double mutant IBB2 was generated by transforming strain BSM27 with chromosomal DNA from strain IBB1 and screening for colonies that simultaneously exhibited resistance to chloramphenicol and spectinomycin. The sigB sigF sigG triple-mutant strain BSG27 was constructed in two steps. First, competent cells of the sigB mutant strain BSM27 were transformed with chromosomal DNA of the sigG mutant strain Marb04 (Steil et al., 2005). Selection for spectinomycin-resistant colonies yielded strain BSG26 (sigB : : HindIII-EcoRV : : cat spoIIIG : : spc). In a second step, competent cells of this sigB sigG double mutant were transformed with chromosomal DNA of strain MO1073 (spoIIAC : : erm) (Stragier et al., 1988), and erythromycin-resistant colonies were selected. This yielded strain BSG27. The presence of the resistance cassettes in the structural genes for the different sigma factors was verified by PCR. For drug-resistance selection in B. subtilis, antibiotics were used at the following final concentrations: chloramphenicol, 5 µg ml–1; spectinomycin, 200 µg ml–1; and erythromycin, 1 µg ml–1.
Media and growth conditions.
B. subtilis strains were maintained on Luria–Bertani (LB) medium containing appropriate antibiotics; plates were incubated at 37 °C. Bacteria were routinely grown with vigorous agitation (220 r.p.m.) in an aerial shaker in Spizizen's Minimal Medium (SMM) with 0·5 % (w/v) glucose as the carbon source, L-tryptophan (20 mg l–1) and a solution of trace elements (Harwood & Archibald, 1990). Pre-cultures of B. subtilis strains were inoculated from exponentially growing overnight cultures that had been propagated in SMM to a final OD578 of 0·1. For the cultures exclusively propagated at 37 °C, cells were harvested after mixing the samples with an equal volume of frozen killing buffer (20 mM NaN3, 20 mM Tris/HCl, pH 7·5, 5 mM MgCl2) by centrifugation (10 min at 6000 g at 4 °C) when the cultures had reached an OD578 of 1·0. These cells were used either for the isolation of total RNA for DNA array experiments or for the isolation of crude protein extracts for proteome studies. For the growth of B. subtilis at low temperature, pre-cultures were propagated to an OD578 of 0·5 at 37 °C, diluted to an OD578 of 0·1 and subsequently transferred to the lower growth temperatures (15 or 16 °C) indicated in the individual experiments. The growth regimen applied allowed growth of the cultures up to an OD578 of at least 6, and cultures were propagated at the lower temperature for at least three generations before harvesting. At 72±8 h after the transfer to the lower temperature, when the OD578 had reached 1·0, aliquots of the cultures were harvested by mixing the samples with an equal volume of frozen killing buffer and by subsequent centrifugation for 10 min at 6000 g and 4 °C. The pellets were stored at –80 °C until used for RNA or protein preparation.
Cell lysis and RNA isolation.
Total RNA of B. subtilis was isolated after mechanical disruption of frozen cells in a Micro-Dismembrator (B. Braun Biotec Int., Melsungen, Germany) as described by Hauser et al. (1998), using the modifications introduced by Petersohn et al. (2001).
Northern blot analysis.
Aliquots of the total RNA prepared for the DNA macroarray experiments were used for Northern blot analysis of the transcription of individual B. subtilis genes. RNA electrophoresis on agarose gels, RNA transfer by diffusion onto a Nylon membrane (NY13N; Schleicher & Schuell), hybridization to gene-specific probes, and signal detection were performed as described by Holtmann et al. (2003). Digoxigenin (DIG)-labelled anti-sense RNA probes were generated by in vitro transcription using a StripEZ-kit (Ambion, Inc.) and gene-specific PCR products as templates. In each of the PCR reactions with chromosomal DNA prepared from B. subtilis strain 168, one of the DNA primers carried the sequence of the T7 promoter. The PCR fragment was subsequently used for in vitro RNA synthesis with commercially available T7 RNA polymerase (Ambion, Inc.). These reactions yielded hybridization probes internal to the structural genes of the following size: spo0A (621 nt), spoIIIG (272 nt), sspE (229 nt), yhcM (429 nt), bofC (455 nt), spoIVA (602 nt), cotE (475 nt), yjbX (638 nt), spoIIID (272 nt), rapA (609 nt), phrA (118 nt), yoaA (465 nt), yoxB (539 nt), yoxC (272 nt), ywcI (283 nt), sacT (548 nt), dps (351 nt), oppA (665 nt), oppF (661 nt), wapA (450 nt), yxxG (331 nt), htrA (501 nt), ldh (505 nt), lctP (525 nt), spo0M (682 nt) and fabHB (713 nt). The nucleotide sequences of the DNA primers used for the generation of the PCR products are available upon request.
Preparation of labelled cDNA, array hybridization and DNA macroarray regeneration.
Prior to the cDNA labelling, the overall integrity of the total RNA preparation was analysed using a Bioanalyser 2100 capillary electrophoresis system (Agilent Technologies). cDNA synthesis and labelling were performed exactly as described by Steil et al. (2003). Transcriptional profiling was carried out with commercially available Panorama B. subtilis DNA macroarrays (Sigma Genosys Ltd) that carry duplicate spots of PCR products representing 4107 B. subtilis genes. cDNA denaturation, probe hybridization and washing of the filters were performed as described by Petersohn et al. (2001), with the modifications detailed by Steil et al. (2003). Hybridized arrays were then exposed to storage phosphor screens (Molecular Dynamics) for 2 to 4 days, and subsequently scanned with a Storm 840/860 phosphorimager (Molecular Dynamics) at a resolution of 50 µm and a colour depth of 16 bit. Bound cDNA was stripped off the DNA macroarray membranes by three washing cycles involving a short (1 min) washing step with 250 ml of boiling buffer (5 mM sodium phosphate, pH 7·5, 0·1 % SDS) and an incubation in 250 ml of fresh buffer at 95 °C for 20 min.
Data analysis of the transcriptional profiling experiments.
Data analysis followed a three-step procedure. First, the ArrayVision software (Version 6.1, Imaging Research, St Catherines, ON, Canada) was used for the quantification of the hybridization signals after direct import of the phosphorimager files. The analysis yielded the artefact-removed volumes (ARVol) and background values, calculated from the median of a line surrounding each group of eight spots on the array. These data were then used in a second step in Microsoft Excel to calculate for every spot on the array a quality score that reflected the ratio between the signal intensity and the background intensity (see supplementary material). This quality score was utilized to identify the hybridization signals close to the detection limit, thereby avoiding artificially high induction ratios for those genes. Data normalization and data analysis were done in a third step with GeneSpring version 7.2 (Agilent Technologies). Gene expression was considered to be changed when three criteria were fulfilled: (i) expression of the gene had to exceed the background signal level by a threshold determined as described in the supplementary material; (ii) changes in expression of the gene had to be statistically significant, as defined in a statistical group comparison of the values of the selected conditions with a parametric test (Welch t test; P value cut-off 0·01) and a Benjamini and Hochberg false discovery rate correction with a P value cut-off of 0·01, as defined in the GeneSpring software package; and (iii) the change in expression had to exceed a factor of two. Calculations of ratios were done with means of the four parallel samples described below. For transcriptional profiling, mRNA was prepared from four independent cultivations of B. subtilis cells and then used for independent cDNA synthesis and DNA array hybridizations. The Panorama B. subtilis DNA macroarrays contained duplicated DNA samples for each of the 4107 B. subtilis genes, therefore the processing of four independent samples for each growth condition yielded eight data points for the calculation of signal intensities for each gene. The potential functions of the encoded proteins were inferred from the SubtiList database (http://genolist.pasteur.fr/SubtiList/) (Moszer et al., 2002).
Two-dimensional differential in-gel protein electrophoresis (2D-DIGE).
Cell pellets were resuspended in Tris/EDTA buffer (10 mM Tris/HCl, pH 7·5, 1 mM EDTA) and bacteria were disrupted by sonication. After removal of cell debris by centrifugation (20 000 g for 30 min at 4 °C), the protein concentration of the crude protein extracts was determined according to the method of Bradford (1976). Prior to the two-dimensional protein gel electrophoresis (2-DE), protein extracts were labelled with Cy-dyes according to the manufacturer's instructions (GE Healthcare, Freiburg, Germany). Briefly, after adjustment of the pH of the protein extract to 8·5 with 50 mM NaOH, 150 µg of each protein extract was labelled with 1200 pmol (400 pmol µl–1) of Cy5 or Cy3 dye. Furthermore, aliquots (75 µg) of each of the four samples [wild-type strain 168 grown at 37 °C, wild-type strain 168 grown at 16 °C, strain BSM29 (sigB) grown at 37 °C and strain BSM29 (sigB) grown at 16 °C] were mixed and labelled with 2400 pmol (400 pmol µl–1) of Cy2 dye, generating a pooled standard of all proteins present in the crude protein extracts. Labelling of 150 µg of protein extract from each experimental condition allowed the separation of three replicates by 2-DE. 2D-DIGE gels were loaded with 50 µg of the Cy2-labelled pooled standard, a Cy5-labelled sample and a Cy3-labelled sample. In order to allow direct comparison of protein patterns of wild-type cells of strain 168 grown at 37 °C (Cy5 labelled) and 16 °C (Cy3 labelled), the corresponding protein extracts were separated together with the pooled standard. An additional set of three gels was used to separate protein extracts of cells of the sigB mutant strain BSM29 grown at 37 °C (Cy5 labelled) or 16 °C (Cy3 labelled) together with the pooled standard. Mixed Cy-dye-labelled protein extract (150 µg) was added to a rehydration solution containing 8 M urea, 2 M thiourea, 2 % (w/v) CHAPS, 28 mM DTT, 1·3 % (v/v) Pharmalytes, pH 3–10, and bromphenol blue. Immobilized pH gradient (IPG) strips (GE Healthcare) with a pH range from 4 to 7 were rehydrated in the protein-containing solution for 24 h under low-viscosity paraffin oil. Subsequently, the IPG strips were subjected to isoelectric focusing with the following voltage/time profile: a linear increase from 0 to 500 V for 1000 Vh, 500 V for 2000 Vh, a linear increase from 500 to 3500 V for 10 000 Vh, and a final phase of 3500 V for 35 000 Vh. IPG strips were incubated consecutively for 15 min each in equilibration solutions A and B [solution A: 50 mM Tris/HCl, pH 6·8, 6 M urea, 30 % (v/v) glycerol, 4 % (w/v) SDS, 3·5 mg DTT ml–1; solution B, as solution A, but 45 mg iodoacetamide ml–1 instead of DTT]. In the second dimension, proteins were separated on 12·5 % SDS-PAGE gels with the Dodecan Electrophoresis System (Bio-Rad) at 2 Watts per gel. After each gel had been scanned with separate laser excitation for each of the three Cy-dyes with a Typhoon scanner (GE Healthcare), analysis of the 2-DE images was performed with version 3.3 of the Delta2D software package (Decodon GmbH, Greifswald, Germany). All gel images (three dyes by six gels) were matched with the Delta2D software and fused to a master gel; subsequently, spot detection was performed for this fusion gel. The spot mask of the fusion gel was then transferred to each of the three individual 2-DE images (Cy2, Cy3, Cy5) of the six gels. After background subtraction, spot volumes were calculated with the Delta2D software and normalized both to the corresponding image (relative expression of each spot compared to the total spot volume of the image) and the pooled Cy2 standard of each individual spot (ratio of the particular sample spot volume to the Cy2 spot volume of the internal standard). The resulting values, which were normalized for image and pooled standard (percentage volume as defined by the Delta2D software) were exported to the GeneSpring software version 7.2 (Agilent Technologies). Protein levels were considered to be changed when two criteria were fulfilled: (i) changes in the level of the protein had to be statistically significant, as defined in a statistical group comparison of the values of the selected conditions with a parametric test (Welch t test, P value cut-off 0·05), as defined in the GeneSpring software package; and (ii) the change in level had to exceed a factor of two. Calculations of ratios were done with averages of the three parallel samples described above.
Preparation of peptide mixtures for matrix-assisted laser desorption ionization (MALDI)-MS.
After the imaging with the Typhoon imager, the 2-DE gels were stained with colloidal Coomassie brilliant blue. After matching spot patterns of gels stained with Coomassie brilliant blue and Cy-dye, proteins were excised from stained 2-DE gels using a spot cutter (Proteome Works, Bio-Rad) with a picker head of 1·5 mm diameter, and transferred into 96-well microplates loaded with 100 µl Lichrosolv water per well. Digestion with trypsin and subsequent spotting of peptide solutions onto the MALDI targets were performed automatically in the Ettan spot handling workstation (GE Healthcare) using a modified standard protocol that has been described elsewhere (Eymann et al., 2004). The matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) measurement of spotted peptide solutions was carried out on a 4700 Proteomics Analyser (Applied Biosystems), as described by Eymann et al. (2004). MS/MS analysis was performed for the three strongest peaks of the MS spectrum. For one main spectrum, 20 subspectra with 125 shots per subspectrum were accumulated using a random search pattern. The internal calibration was automatically performed as one-point calibration if the mono-isotopic arginine (M+H)+ m/z at 175·119 or lysine (M+H)+ m/z at 147·107 reached a signal-to-noise (S/N) ratio of at least 5. The peak lists were created using the ‘peak to mascot’ script of the 4700 Explorer software (Applied Biosystems) with the following settings: mass range from 60 Da to a mass that was 20 Da lower than the precursor mass; peak density of 10 peaks per 200 Da; minimal area of 100 and maximal 100 peaks per precursor; minimal S/N ratio of 5. Database searches employed a B. subtilis-specific database using the Mascot search engine (Matrix Science Ltd, London, UK). The identification of a protein spot was considered significant if the Mowse score exceeded a value of 49, which corresponds to a P value of 0·05. Potential protein functions were inferred from the SubtiList database (http://genolist.pasteur.fr/SubtiList/) (Moszer et al., 2002).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
). The expression pattern of each of the cultures grown in parallel was highly reproducible, and on average showed Pearson correlation coefficients of 0·972 (data not shown). In contrast, the expression pattern of samples grown at 15 °C differed significantly from those of cultures grown at 37 °C (Pearson correlation coefficients of 0·861). A group of 279 genes displayed at least a twofold statistically significant induction in cultures grown at 15 °C, and 301 genes displayed at least a twofold statistically significant repression under chill-stress growth conditions (Fig. 1A). The 279 genes whose transcription is induced in the cold are listed in Table 1, together with their predicted function as currently annotated in the SubtiList database (Moszer et al., 2002). We also list in this table known or suggested regulatory proteins for these genes. Detailed information on the 301 genes whose activity is reduced at low growth temperature is available in Supplementary Table 2.
|
|
demonstrate that in each case the pattern observed in the DNA array analysis was confirmed in the Northern blot analysis, thereby supporting our identification of differentially expressed genes via the DNA array technique.
|
of chill-adapted B. subtilis cells revealed the induction of the SigB-dependent general stress response at low temperature (15 °C), and also demonstrated the chill induction of the sigB operon, which encodes the master regulator (SigB) and several regulators of its own activity (RsbV, RsbW, RsbX) (Kalman et al., 1990). We therefore mined our transcriptome data for chill induction of members of the SigB regulon that had previously been defined through three independent transcriptional profiling studies (Helmann et al., 2001; Petersohn et al., 2001; Price et al., 2001). Each of these SigB profiling studies assign between 120 and 190 genes to the SigB regulon, raising the potential size of this regulon to almost 300 members. However, only a limited number (approx. 25 %) of potential SigB-dependent genes are found in all three studies (Helmann et al., 2001; Petersohn et al., 2001; Price et al., 2001).
Comparing our list of chill-induced genes (Table 1
) with the data published by Petersohn et al. (2001) on the make up of the SigB regulon, we found 97 potential SigB-controlled genes among the chill-induced genes of B. subtilis. When the data of Helmann et al. (2001) and Price et al. (2001) were further considered, we found 82 and 92 SigB-controlled genes, respectively, among the low-temperature-induced genes. Taking all data into account, and removing the redundancy between the three studies, we now can assign approximately 40 % (112 of 279 genes) of the B. subtilis chill-induced genes to the SigB-dependent general stress regulon. Thus, the proteome study of Brigulla et al. (2003) and the data presented here conclusively demonstrate that chill stress is a potent inducer of the SigB regulon, and that the SigB-dependent general stress response is an integral part of the adaptation of the B. subtilis cell to growth at low temperature. This view is supported by the findings that a sigB mutant cannot effectively grow at low temperature (Brigulla et al., 2003), that loss of SigB reduces the stationary-phase viability of cold-adapted cells (Mendez et al., 2004), and that the disruption of individual members of the SigB regulon often leads to a cold-sensitive phenotype (Höper et al., 2005). Of the 96 mutants with defects in individual general stress genes tested by Höper et al. (2005), 35 displayed a low-temperature-sensitive phenotype, and of these 35 mutated genes, 22 were found to be chill-induced in this study (Table 1). Many of the 112 SigB-dependent genes that are induced by chill stress have no assigned physiological function (Table 1). Even if one considers only those genes with a putative function, one cannot readily deduce how the products of these genes might contribute to the adaptation of B. subtilis to low temperature. It is thus obvious that detailed genetic, biochemical and physiological studies are required to reveal how these proteins are integrated into the low-temperature-adaptation network.
We find it worth commenting that we observed a different degree of matching between the chill-induced genes identified in this study (Table 1) and the assignment of the SigB regulon by the three previous transcriptional profiling studies (Helmann et al., 2001; Petersohn et al., 2001; Price et al., 2001). Petersohn et al. (2001) discriminate between general stress genes whose stress induction is exclusively dependent on SigB activity and those genes that are also subjected to SigB-independent multiple stress induction. For the former group, 89 % (89 out of 100 genes) of the genes (Fig. 1B) were observed in this study to be induced by growth at low temperature, whereas for the latter group, only 33 % (8 out of 24 genes) were found to be so induced. Of the SigB-dependent genes proposed by Helmann et al. (2001) and Price et al. (2001), 65 and 48 %, respectively, were recovered if inspected for chill inducibility. This observation does not necessarily reflect quality differences in the different array studies, but rather appears to stem from the use of different DNA arrays. We and Petersohn et al. (2001) used the same type of commercially available DNA macroarray (Sigma Genosys Ltd), whereas Helmann et al. (2001) and Price et al. (2001) made use of different custom-synthesized DNA arrays.
Chill stress triggers partial induction of the SigF, SigE and SigG sporulation regulons
Further inspection of the chill-induced B. subtilis genes (Table 1) led to the discovery that a substantial portion of these genes have previously been assigned by transcriptional profiling studies to regulons of the sporulation response (Eichenberger et al., 2003, 2004; Fawcett et al., 2000; Feucht et al., 2003; Steil et al., 2005). Contrary to that which has been observed for the SigB-dependent general stress regulon, for which most of the SigB-dependent genes are induced by chill stress (Fig. 1B), only a subgroup (15 %, 67 of 439 genes) of sporulation-specific genes (Steil et al., 2005) is induced by chill stress (Fig. 1C). After initiation of sporulation (Hoch, 1995), the sporulation developmental programme progresses through the sequential induction of four regulons that are controlled by the alternative transcription factors SigF, SigE, SigG and SigK (Stragier & Losick, 1996). Based on the assignment of sporulation genes to the individual sporulation regulons, as described by Steil et al. (2005), we found 18 chill-induced genes among the 55 SigF regulon members, 28 among the 154 SigE regulon members, and 25 among the 113 SigG regulon members. None of the 132 SigK-controlled genes listed by Steil et al. (2005) was found among the low-temperature-induced genes identified by us. Comparison of the chill-induced genes defined in this study (Table 1) with other transcriptional profiling data that assess the sporulation response of B. subtilis (Eichenberger et al., 2003, 2004; Fawcett et al., 2000; Feucht et al., 2003), assigned 14 additional chill-induced genes to the sporulation programme (Table 1). The observation that only 81 of the 743 potential sporulation genes (Eichenberger et al., 2003, 2004; Fawcett et al., 2000; Feucht et al., 2003; Steil et al., 2005) were also chill induced (Table 1) indicates that the full sporulation programme is not implemented in cells growing at low temperature. This is consistent with the observation that spore formation is drastically reduced (approx. 1000-fold) in cells grown at 15 °C in comparison with cells growing at 37 °C (A. Bashir & U. Völker, unpublished results).
Partial induction of the sporulation regulons by chill stress raised the question of whether the enhanced transcription of these subgroups of the various spo regulons is dependent on the regular sporulation-specific transcription factors or is mediated by some unknown regulatory pathway. We chose three members (cotE, spoIVA, yjbX) of the chill-induced subgroup of the SigE regulon and performed Northern blot analysis on total RNA isolated from cells grown at 15 and 37 °C. As expected from the DNA array analysis (Table 1), transcription of each of these genes was induced by chill stress, and this induction was also observed in a sigB mutant (Fig. 3A). In contrast, chill induction of the cotE, spoIVA and yjbX genes was not detected in a mutant that lacks the first sporulation-specific sigma factor SigF (Fig. 3A). Taken together, these observations indicate that low-temperature induction of the small chill-induced fraction of the SigE regulon does indeed require a functional SigF protein.
|
). However, transcription still occurred in a sigF mutant when grown at low temperature (Fig. 3B). This persisting chill induction in the sigF mutant could perhaps be mediated by overlapping promoter recognition by SigB, SigG and SigF (Amaya et al., 2001; Helmann & Moran, 2002). Chill induction of all three genes (yhcM, sigG, sspE) was still observable in a sigB sigF double mutant, but was lost upon the introduction of a defective sigG gene into the sigB sigF double-mutant strain (Fig. 3B). The dependence of low-temperature induction of the subgroups of the spo regulons on the sigma factors SigF and SigG provides evidence that regular signal transduction in the sporulation pathway is required to achieve this chill induction.
Low-temperature induction of genes involved in the fine-tuning of Spo0A activity
Interestingly, the list of chill-stress-induced B. subtilis genes also contained genes encoding proteins involved in fine-tuning the phosphorylation status and activity of the master regulator (Spo0A) of the sporulation response and stationary-phase adaptation. The rapA and phrA genes displayed chill induction of 7·4- and 4·9-fold, respectively (Table 1). Northern blot analysis confirmed this induction and revealed cotranscription of the rapA and phrA genes under chill-stress conditions (Fig. 4B).
|
). This group of sequence-related proteins regulate protein activity either by acting as protein phosphatases that target the phosphorylated forms of response regulators of two-component systems (Perego et al., 1996; Perego, 1999) or by specifically binding to regulatory proteins (Bongiorni et al., 2005; Core & Perego, 2003; Ogura et al., 2003). The activity of seven of the eleven Rap proteins, including RapA, is regulated by specific pentapeptide inhibitors; these are generated through the export of the Phr protein from the cell, the proteolytic processing of this precursor protein and the subsequent reimport of a Phr-derived pentapeptide through the Opp oligopeptide transport system (see Fig. 4A) (Perego, 1997, 1999). The genes encoding the OppABC transport system (Perego et al., 1991) were also induced by chill stress, as revealed by our transcriptional profiling data (Table 1) and the subsequent Northern blot and dot-blot analysis (Fig. 4B). Low-temperature induction was also observed for a second oligopeptide ATP-binding cassette (ABC)-type transport system, App (Table 1).
RapA, RapB and RapE specifically mediate the dephosphorylation of the Spo0F
P intermediate of the sporulation phosphorelay that controls the phosphorylation status of the Spo0A response regulator (Jiang et al., 2000; Perego et al., 1994). In contrast to low-temperature induction of the rapA and phrA genes, the expression of the structural genes for the RapB and RapE proteins and the PhrE peptide did not change with decreasing temperature. None of the other eight rap genes was induced by chill stress; in fact, the genes for rapF and rapH were repressed.
The coordinated induction of the RapA/PhrA systems and the Opp oligopeptide transport system in chill-adapted cells indicates that the phosphorylation status of the Spo0A regulatory protein might need fine-tuning to ensure the survival of B. subtilis at low temperature. In this context, it is interesting to note that Mendez et al. (2004) have recently observed a dramatic (10 000-fold) decrease in stationary-phase survival of cold-adapted B. subtilis cells that are defective in Spo0A. These authors have also shown that the requirement for Spo0A for cellular survival at low temperature is independent of the transition-state regulator AbrB and does not rely on the ability of the cells to sporulate (Mendez et al., 2004). In their study, Mendez et al. (2004) report induction of a transcriptional spo0A–lacZ fusion after a temperature shift from 37 to 20 °C. In contrast, our DNA array data did not show transcriptional induction of the spo0A gene in cells that were continuously cultured at 15 °C. We found by Northern blot analysis of total mRNA isolated from these chill-adapted cells that the spo0A mRNA was present at a level comparable to that in cells growing exponentially at 37 °C (Fig. 5A). However, when we assessed and quantified the level of the Spo0A protein in the B. subtilis wild-type strain 168 grown at 37 or 16 °C by a proteomic approach (see below), we found an accumulation of Spo0A in the low-temperature-adapted cells (Fig. 5B, C). This chill-induced accumulation of Spo0A was also apparent in an isogenic sigB mutant of B. subtilis strain 168 (Fig. 5C). Accumulation of the Spo0A protein has also been observed by Mendez et al. (2004) by Western blot analysis of wild-type cells shifted to 20 °C.
|
). Eight genes (sigE, sigF, spoIIAB, spoIIE, yocH, yttP, ycgM, ykuL) (Table 1) have been shown to be subjected to positive control by the Spo0A regulator (Molle et al., 2003a), some of which (sigE, sigF, spoIIAB, spoIIE) are involved in controlling the onset of sporulation and sporulation-specific gene expression (see above).
The transcriptional profiling data revealed a 19-fold and a 23-fold chill induction of the paiA and paiB genes, respectively, which are cotranscribed as an operon, as assessed by Northern blot analysis (G. Streubel & E. Bremer, unpublished results). Overproduction of the PaiA and PaiB proteins from a multicopy plasmid has been shown elsewhere to reduce the formation of extracellular enzymes, including the alkaline serine protease AprE and the neutral metalloprotease NprE, and to trigger inhibition of sporulation at 37 °C (Honjo et al., 1990). However, we found that the structural genes for the AprE and NprE proteases were moderately upregulated (Table 1) in chill-adapted B. subtilis cells, despite the strong induction of the paiAB operon, perhaps indicating that PaiAB exert their effects at the post-transcriptional level.
The list of chill-induced genes contained eight genes that code for additional regulatory proteins: YdeB, YisR, YmfC, DegA, LmrA, SacT, SacY and SigL. The alternative transcription factor SigL has previously been implicated in cold-shock adaptation of B. subtilis by Beckering et al. (2002), who identified the SigL-dependent transcriptional activator homologue YplP as a cold-shock-induced protein. These authors also demonstrate that a mutant strain with a gene disruption of yplP can only ineffectively recover from a cold shock. B. subtilis contains 36 histidine kinases and 34 response regulators of two-component regulatory systems (Fabret et al., 1999). None of the genes encoding these two-component regulatory systems was induced by chill stress; in fact, the resD/resE and degS/degU genes were repressed.
In contrast to the general reduction in metabolism of chill-adapted cells that is described below, we noted specific induction of a few biosynthetic operons. We observed chill induction of five genes (argB, argD, argG, argH, carA) (Table 1) of the arginine biosynthetic pathway; the remaining four genes (argC, argF, argJ, carB) of this biosynthetic route were also slightly chill-induced, but did not pass the cut-off criteria adopted in this study. Interestingly, one of the operons (ureA, ureB, ureC) of the urea cycle that uses arginine as a substrate for ammonia generation was chill induced as well (Table 1). Furthermore, we observed induction of part of the tryptophan biosynthetic pathway: trpB, trpD and trpF were clearly induced by chill stress (Table 1), whereas trpA, trpC and trpE exhibited a moderate induction by chill stress, but did not fulfil our selection criteria.
Chill induction of the excision repair genes uvrA, uvrB and uvrC was also observed (Table 1). These three genes encode an excision nuclease complex that is involved in the repair of ultraviolet-light-induced pyrimidine dimers in DNA, indicating that there might be a greater need for DNA excision repair in cells grown at low temperature.
Low-temperature-mediated repression of gene expression
Our transcriptional profiling data revealed that the transcription of 301 genes was repressed in B. subtilis cells grown at low temperature. Considering the potential functions of these genes, the transcriptional data indicate a general reduction in the metabolic activity of B. subtilis cells cultivated at low growth temperature. This is consistent with the strongly slowed growth of chill-adapted B. subtilis cells (Brigulla et al., 2003). Among the processes that we found reduced in low-temperature-grown cells (15 °C) was the synthesis of components of the ribosome (Fig. 1F). The catabolic processes glycolysis, oxidative phosphorylation and ATP synthesis were all strongly reduced (Fig. 6), a finding which is compatible with a reduced energy demand of the cell during slow growth at low temperature. Anabolic activity was also affected, as exemplified by the overall reduction of the transcription of genes encoding enzymes involved in purine, pyrimidine, haem and fatty acid synthesis (Fig. 7). Of particular interest in this context is the reduced level of fatty acid synthesis, indicating that the cell has fully adjusted the lipid composition of the cytoplasmic membrane to sustain growth at the low temperature. This is consistent with our finding that the des gene encoding the 5-lipid desaturase was not induced during low-temperature growth. On the contrary, it was strongly upregulated following a cold shock (Aguilar et al., 1998; Beckering et al., 2002; Kaan et al., 2002). The list of chill-stress-repressed genes contains approximately 50 additional genes with diverse metabolic functions, and another group of 59 genes of thus far undefined function.
|
|
), the expression of 32 genes was reduced in the cold to a statistically significant extent (Fig. 1E). The reduction of the expression of chemotaxis and motility genes in general (Fig. 1E) suggests that the ability of cells to actively swim is reduced or abolished when B. subtilis is cultivated at low temperature. DegUP is a major repressor of the motility genes in B. subtilis (Amati et al., 2004). By matching our expression data with the assignments of the DegS/DegU-controlled regulon of Ogura et al. (2001) and Mäder et al. (2002), we were able to ascribe 28 of the 32 chill-repressed genes with a predicted function in chemotaxis and motility to the DegS/DegU regulon. The influence of the DegS/DegU two-component system on chill-repressed genes is not confined to the group of chemotaxis and motility genes, but extends to 35 additional cold-repressed genes.
The SigW regulon of B. subtilis has been functionally implicated in cell-wall-associated processes and the adaptation to alkaline shock (Cao et al., 2002; Huang et al., 1998; Wiegert et al., 2001). Based on the definition of the SigW regulon structure of Huang et al. (1998) and Wiegert et al. (2001), we found 26 of the 62 SigW regulon members to be repressed at low temperature (Fig. 1D). Interestingly, among these were the structural genes for the central regulator of this regulon (sigW) and its repressing anti-sigma factor (rsiW) (Schobel et al., 2004). Furthermore, a partial repression of the comK regulon (31 of 164 genes) (Berka et al., 2002), controlling the development of competence for DNA uptake by B. subtilis, was observed in cells grown at low temperature. Among the chill-stress-repressed genes were four additional genes (cggR, hbs, yvbA, yrxA) that code for proteins involved in the regulation of gene expression.
A temperature-dependent gene expression was observed for the ytrABCDEF and bkdR-ptb-pcd-buk-lpdV-bkdAA-bkdAB-bkdB gene clusters. While both gene clusters have previously been shown to be induced following a cold shock (Beckering et al., 2002; Nickel et al., 2004), our transcriptome analysis of fully chill-stress-adapted B. subtilis cells revealed their repression. The ytrABCDEF gene cluster encodes a repressor (ytrA) and the components of an ABC transporter for acetoin utilization (Yoshida et al., 2000). The bkdR-ptb-pcd-buk-lpdV-bkdAA-bkdAB-bkdB gene cluster encodes enzymes that catalyse the degradation of branched-chain amino acids, which are necessary for the maintenance of membrane fluidity at low temperatures; its transient induction subsequent to cold shock is primarily due to stabilization of its mRNA following temperature downshift (Nickel et al., 2004).
The opuAA-opuAB-opuAC operon that encodes the components of an ABC transport system (OpuA) for the compatible solute glycine betaine (Kempf & Bremer, 1995) is repressed at low temperature. However, the OpuA transporter makes an important contribution to the overall uptake of glycine betaine (Bremer, 2002) when B. subtilis cells accumulate this compatible solute as a cryoprotectant (T. Hoffmann & E. Bremer, unpublished results). Thus, in chill-adapted cells, appropriate functioning of the OpuA ABC transport system might be ensured by post-transcriptional regulation or by control of its activity.
Proteome analysis of chill-adapted B. subtilis cells
Low-temperature incubation of micro-organisms imposes severe constraints on translation due to the increased formation of mRNA secondary structures. One of the protective adaptation responses of the B. subtilis cell to cold shock specifically circumvents this problem by synthesizing cold-shock proteins that act as RNA chaperones (Graumann & Marahiel, 1996; Weber & Marahiel, 2002) or by inducing the synthesis of RNA helicases that are involved in the unwinding of double-stranded RNA regions (Beckering et al., 2002; Jones et al., 1996). Thus, post-transcriptional regulatory processes are particularly important in adjusting translation and thereby ensuring appropriate protein levels for the low-temperature adaptation of microbial cells.
We therefore supplemented our transcriptional profiling approach for low-temperature-grown B. subtilis cells with a proteomic study. For this purpose, we grew the B. subtilis wild-type strain 168 at either 37 or 16 °C to mid-exponential phase (OD578 of 1) and prepared from these cells total soluble protein extracts for analysis by 2D-DIGE. Matching of the protein patterns by the Delta2D software allowed the assignment of 1085 protein spots. Of these protein spots, 491 exhibited statistically significant differences (twofold) between cells grown at 37 and 16 °C: 246 protein spots were chill-induced and 245 were chill-repressed. The dual-channel image overlay of the protein profiles that is presented in Fig. 8 gives an impression of the massive temperature-dependent changes in the cytosolic protein profile of B. subtilis. We identified 109 induced and 88 repressed protein spots by mass spectrometry that were subsequently allocated to 86 and 60 distinct protein species, respectively. Detailed information on the identification of proteins by mass spectrometry and the quantification of the protein levels is given in Supplementary Tables 3 and 4.
|
) and the proteins found at increased levels in low-temperature-grown cells (Supplementary Table 3) revealed an overlap of 28 entries. Consistent with the strong induction of the SigB regulon (Fig. 1B) and the preponderance of its members among the chill-induced genes (Table 1), 20 of the 28 entries were allocated to the SigB-dependent general stress response (Fig. 8). As expected, these members of the SigB regulon were not found as chill-induced proteins when a proteome analysis was carried out with a sigB mutant derivative of the B. subtilis wild-type strain 168. Examples of the comparison of the wild-type strain 168 and its isogenic sigB mutant are shown in Fig. 9(A). The remaining eight proteins that were chill induced and whose structural genes displayed transcriptional induction at low temperature were UvrB, SpoIVA, MsrA, UreC, Mfd, YjcH, ArgD and PrkA (Fig. 8). 
