- Research article
- Open Access
Transcriptional activity around bacterial cell death reveals molecular biomarkers for cell viability
© Kort et al; licensee BioMed Central Ltd. 2008
- Received: 22 July 2008
- Accepted: 06 December 2008
- Published: 06 December 2008
In bacteriology, the ability to grow in selective media and to form colonies on nutrient agar plates is routinely used as a retrospective criterion for the detection of living bacteria. However, the utilization of indicators for bacterial viability-such as the presence of specific transcripts or membrane integrity-would overcome bias introduced by cultivation and reduces the time span of analysis from initiation to read out. Therefore, we investigated the correlation between transcriptional activity, membrane integrity and cultivation-based viability in the Gram-positive model bacterium Bacillus subtilis.
We present microbiological, cytological and molecular analyses of the physiological response to lethal heat stress under accurately defined conditions through systematic sampling of bacteria from a single culture exposed to gradually increasing temperatures. We identified a coherent transcriptional program including known heat shock responses as well as the rapid expression of a small number of sporulation and competence genes, the latter only known to be active in the stationary growth phase.
The observed coordinated gene expression continued even after cell death, in other words after all bacteria permanently lost their ability to reproduce. Transcription of a very limited number of genes correlated with cell viability under the applied killing regime. The transcripts of the expressed genes in living bacteria – but silent in dead bacteria-include those of essential genes encoding chaperones of the protein folding machinery and can serve as molecular biomarkers for bacterial cell viability.
- Colony Form Unit
- Heat Shock Response
- Competence Gene
- Subtilis Cell
- Colony Form Unit Count
Ever since the pioneering work by Louis Pasteur and Robert Koch at the end of the nineteenth century, the detection of viable bacteria has been carried out by cultivation and enumeration of colony forming units (CFU). Practically all judgments on hygiene, food safety, conservation treatments, drinking water quality, infections of pathogens, efficacy of disinfectants and antibiotics are based on growth on solid agar medium followed by CFU counts. However, the assessment of cell viability on agar plates is laborious, it requires at least an overnight incubation and the outcome yields little information on bacterial physiology. Besides, the assessment of CFU counts is limited to bacteria that are readily culturable under laboratory conditions and even when they are, the failure of bacteria to reproduce on an agar plate does not necessarily imply that they are metabolically inactive or were inactive at the time of sampling [1, 2].
To overcome the shortcomings of CFU enumeration pointed out above, a number of alternative, cultivation-independent methods have been applied over the years to get hold of indicators for bacterial cell viability. The most commonly employed methods include fluorescence-based assays for enzymatic activity, electron transport and membrane permeability and molecular approaches for the detection of rRNA or specific mRNA molecules. The fluorescence-based assays provide signals that can be related to known cell properties, which can be assessed at the cellular rather than population level by the use of fluorescence microscopy and flow cytometry. The molecular approaches include microarrays and real-time PCR to select and quantify specific RNA molecules. The latter methods have been generally looked upon as providing indicators of specific aspects of bacterial physiology rather than indirect measures for bacterial cell viability valid under a wide range of lethal stress conditions. However, the feasibility of cultivation-independent biomarkers under a number of well-defined conditions as indirect indicators for cell viability is still a matter of debate [1, 3, 4].
In this study, we evaluated a number of commonly applied cultivation-independent methods under accurately defined conditions by a gradual exposure of the Gram-positive model bacterium Bacillus subtilis to heat stress. We systematically determined the effects of heat exposure on two cultivation-dependent measures, the ability to form colonies on agar plates, expressed in CFU counts, and outgrowth of the heat-exposed bacteria diluted in liquid cultures, expressed in the time delay to reach mid log phase. Furthermore, we determined a number of cultivation-independent measures, including the permeability of the bacterial membrane for fluorescent compounds, the stability of ribosomal RNA molecules and the presence of specific gene transcripts, as identified with microarrays and quantitative PCR methods. The results presented here reveal clear correlations between viability and cultivation-independent indicators in B. subtilis exposed to lethal heat stress. Notably, the genome wide expression profiles of heat-exposed B. subtilis reveal a coherent transcriptional program including known heat shock responses [5, 6] and the rapid expression of sporulation and competence genes, the latter only known to be expressed in the stationary growth phase [7, 8]. This and other forms of coordinated gene expression in the bacterial cell continued even after cell death, in other words after all bacteria permanently lost their ability to reproduce.
Strains and culture conditions
B. subtilis strains used in this study
B. subtilis strains
trpC2/parental strain in this study
Bacillus Genetic Stock Center
food product isolate
food product isolate
Daniel Kearns 
DS323 → 168 1A1 (sigD)
PB344 → 168 1A1 (sigB)
AH46 → 168 1A1 (sigF)
AH74 → 168 1A1 (sigE)
AH70 → 168 1A1 (sigK)
An overnight culture of B. subtilis was diluted 20× in TSB to a volume of 600 ml and incubated at 30°C at 100 rpm until OD600 reached 0.5. Subsequently, portions of 40 ml of the preculture of B. subtilis were exposed to seven different temperature regimes in 250-ml bottles in a reciprocal-shaking water bath at 65°C. Duplicate samples were taken when the temperature in the bottles reached 40, 50, 57, 58, 59, 60, and 62°C (at incubation times of approximately 100, 200, 300, 320, 380, 500 and 720 s after transfer, respectively). Immediate quenching of metabolic activity in the samples was performed with cold methanol, according to described methods . Briefly, 40-ml culture samples were sprayed in 160-ml of a stirred solution of 60% methanol, 66.7 mM HEPES, pH 6.5 that was kept at – 45°C in an ethanol bath. Besides RNA-isolations, the heat-treated cultures were directly subjected to outgrowth analysis to monitor the recovery time and flow cytometric analysis, as further described below. The outgrowth was carried out in 10× dilution series by OD measurements in time at 37°C in TSB by use of the Microbiology Reader Bioscreen C (Oy Growth Curves Ab Ltd, Finland). The time delay in reaching the mid log phase (OD wideband = 0.5) compared to the 30°C-control sample was used as a measure for the increased recovery time from heat injury for each specific exposure.
The RNA isolation protocol was adapted from an extraction method described previously . In brief, cells were disrupted in a 2-ml screw-cap tube in the presence of 100-μm zirconium beads and SDS/phenol for 60 seconds in the Minibeadbeater 8™ (Biospec Products). After chloroform extraction, nucleic acids were precipitated, vacuum-dried, dissolved in distilled water and incubated for 15 min in the presence of DNase I (Qiagen). Finally, RNA was purified by phenol/chloroform extraction and ethanol precipitation as verified by absorption properties in the 200–300 nm region by the use of the ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, Delaware, USA). RNA MW profiles were verified with the 2100 Electrophoresis Bioanalyzer (Agilent Technologies). The RNA 6000 ladder (Ambion Inc, Austin TX, USA) consisting of 6 transcripts at a concentration of 20 ng/μl, ranging from 0.2 to 6.0 kb in length was used as a marker. In addition, a fast migrating "reference marker" compound was added to all samples for software alignment of all electropherograms within one LabChip run (Agilent Technologies).
Construction of B. subtilis microarrays
The B. subtilis microarrays were constructed by the Micro-Array Department (MAD), University of Amsterdam. An oligonucleotide library of 65-mers (Sigma-Genosys Ltd.), which covered 4100 open reading frames of the B. subtilis 168 genome , was spotted in duplicate on glass slides with a Generation III microarray spotter (Molecular Dynamics, Sunnyvale, California, USA) using Lucidea Spotting Pins (GE Healthcare) as described previously .
Labelling of cDNA
Fluorescently labelled cDNA was prepared from 12.5 μg of total RNA by random hexamer pd(N)6-primer (Roche, Mannheim, Germany) polymerization using Superscript II reverse transcriptase (Life Technologies). Concentrations of nucleotides in labelling reaction mixture were 0.4 mM dATP, dGTP, dCTP and 0.2 mM dTTP. The final concentration of Cy3-dUTP or Cy5-dUTP (GE Healthcare) was 0.1 mM. Unincorporated Cy-dye-labelled dUTP, dNTP's, primers and salts were removed by purification with AutoSeq G50 columns (GE Healthcare).
Hybridization and image analysis
Microarray slides were incubated for 45 min at 42°C with prehybridization solution (1% bovine serum albumin, 5× SCC and 0.1% SDS, filtered), washed three times with milliQ water, and dried by the use of a nitrogen flow. The Cy3-labelled cDNA from untreated cells sampled at 30°C, was mixed with Cy5-labelled cDNA from heat-treated cells for all hybridizations. After 2 min of denaturation at 95°C, hybridizations of microarray slides were performed overnight at 42°C in 40 μl of EasyHyb buffer (BioCat, Heidelberg, Germany) with both labelled cDNAs and 2.5 mg ml-1 yeast tRNA (final concentration). Microarray slides were washed at room temperature for 10 sec in 1× SSC/0.2% SDS at 37°C, 10 sec in 0.5× SSC at 37°C, and twice for 10 min in 0.2× SSC at room temperature. Slides were dried by the use of a nitrogen flow and scanned with a ScanArray 5000 laser scanner (Perkin Elmer Life Sciences). The TIFF-images were quantified with the software package ImaGene version 5.6.1 (BioDiscovery, El Segundo, CA, USA).
Data processing and normalization
The spot intensities of B. subtilis oligonucleotide microarrays were processed by the use of Microsoft Office Excel 2003 SP2. The intensities of 8200 spots were calculated by extraction of the median background (B) from the median signal (S) and values for S-B < 1 were set to 1. The raw Cy5/Cy3 ratios (r) were calculated from the remaining background-corrected median signals. The total intensity normalization was carried out by the normalization factor N, defined as the average Cy5 (S-B) divided by the average Cy3 (S-B)). The normalization factor was determined from spot intensities with S/B > 2 for both Cy5 and Cy3 channels in order to avoid bias resulting from the degradation of the major of the transcripts at temperatures from 57–62°C. The normalised Cy5/Cy3 ratios (R = r/N) with S/B > 2 for either the Cy5 or Cy3 channel were calculated. Ratios from duplicate spots on the array and 2 independent experiments were averaged and 2log transformed. Next, all gene IDs with less than 2 observations out of 8 conditions (30, 40, 50, 57, 58, 59, 60, 62°C) were removed from the data set, leaving 3850 out of 4100 gene IDs (94%). Expression profiles were analyzed with K-means clustering into 4 groups with the Pearson correlation as distance metric by the use of Multiple Experiment Viewer MeV v4.1 . Subsequently, the 2log ratios of spot intensities were selected which showed 5 or more useful ratios out of 8 temperature conditions, comprising a total of 3562 ratios (87%). In addition, all determined transcription levels for each gene at each temperature condition were analyzed by the use of the T-profiler http://www.science.uva.nl/~boorsma/t-profiler-bacillusnew, a tool that uses the t-test to score changes in the average activity of predefined groups of genes [16, 17]. Group definitions were obtained from the DBTBS http://dbtbs.hgc.jp/ver4/, the database of transcriptional regulation in B. subtilis, release 4.1 .
Microarray data accession number
Microarray data were deposited in the Gene Expression Omnibus (GEO) database http://www.ncbi.nlm.nih.gov/geo/ under accession number GSE12372.
Heat-treated B. subtilis cell suspensions (OD600 = 0.5) in the temperature range of 40–62°C were directly labelled with propidium iodide (Molecular Probes Inc, Eugene, OR, USA) at 2 μM final concentration and incubated for 2 min at room temperature. For calibration purposes a sample treated for 10 min at 96°C (100% PI-permeable) was mixed with an untreated sample in relative amounts of 0, 20, 40, 60, 80 and 100% PI-permeable cells. For flow-cytometric analysis the cell-suspensions were diluted 40-fold in filtered phosphate-buffered saline solution and directly measured on a Coulter Epics XL-MCL flow cytometer (Beckman Coulter, Mijdrecht, The Netherlands) operating an argon laser (488 nm). For each sample, at least 10,000 cells were analyzed. Fluorescent signal data were collected in the FL4 detector channel, and the photomultiplier voltage was set at 917 V for all measurements. Data were captured and analyzed using the EXPO32 analysis software (Beckman Coulter). Figures were prepared for publication using MicroCal Origin (version 7.5).
Labelling was carried out with fluorophores Syto9 and propidium iodide (PI) according to methods described in the LIVE/DEAD BacLight kit (Molecular Probes Inc.) . Subsequently, 10-μl aliquots of the cultures were embedded in 1% (w/v) final concentration of agarose and deposited on the surface of a microscope slide. Samples were analyzed by the use of the BX61 TRF fluorescence microscope (Olympus Inc.).
The extraction of RNA was carried out on a cell suspension of 750 μl of bacterial cells, which were quenched immediately after heat exposures as described above in detail. For cDNA synthesis, 1 μg of the isolated total RNA was vacuum dried and incubated in 5 μl Random primer mix, consisting of 3 μl 5U μl-1 RNAsin (Promega) and 2 μl 2 μg μl-1 pd(N)6 random primer (Roche) for 5 min at 70°C. After cooling down for 10 min, 8.5 μl RT mix was supplemented: 1 μl RNAsin (10 U μl-1, Promega); 2.5 μl First Strand Buffer 5× (Invitrogen); 1 μl 0,1 M DTT (Invitrogen); 3 μl 2 mM dNTP's (Invitrogen) and 1 μl Superscript II enzyme (200 U μl-1, Invitrogen). The RT reaction was carried out for 60 min at 42°C. Subsequently, to degrade the RNA, 2.5 μl of 2.5 M NaOH was added to the reaction mix and incubated for 15 min at 65°C. This was followed by addition of 2.5 μl 2.5 M NaAc and 10 μl milliQ. The cDNA was purified by the use of a G50 Autoseq column, dried in vacuum and resuspended in 25 μl prior to the Q-PCR reactions.
List of primer probe combinations for Q-PCR
Construction of sigB, sigD, sigF, sigE and sigK mutants
The DNA recipient strain B. subtilis strain 1A1 was cultured overnight in Spizizen salts minimal medium, containing per liter 14 g K2HPO4, 6 g KH2PO4, 2 g (NH4)2SO4, 1 g Na3C6H5O7 (sodium citrate) and 0.4 μg MgSO4, 0.2 g casamino acids, 5 g glucose (Tritium Microbiology) and 80 μg L-tryptophan (Sigma Aldrich Inc.). The overnight cultures were diluted 10 times in Spizizen salts minimal medium, and grown for 3 more hours under the same conditions. Subsequently, one volume unit of starvation medium (same as Spizizen salts minimal medium, but without casamino acids and L-tryptophan) was supplemented and growth was continued for two more hours. Competent cells were incubated for 15 min at 37°C in the presence of isolated genomic DNA from sigma factor mutants as listed in Table 1 and plated on selective agar media. The transformed sigF, sigE and sigK strains were checked for the loss of ability to sporulate after overnight growth in nutrient agar (Tritium Microbiology) supplemented with 0.5% (w/v) CaCl2.H2O and 0.5% (w/v) MnSO4. The sigD strain was checked for loss of motility by microscopy.
Heat exposure regime
Colony forming units
Culture-dependent and independent viability indicators
2log ratios μ-array
2log ratios q-PCR
CFU( 10 log)
Outgrowth in liquid cultures
The recovery of bacteria from each temperature exposure was determined in dilute liquid cultures in triplicate by monitoring the time delay in reaching the mid log phase (OD wideband = 0.5) compared to the 30°C-control sample. Small increases in the average lag time were observed (0.2 and 0.8 hours) for temperature exposures of 40 and 50°C (Table 3). However, their significance remains uncertain as standard deviations for observed lag times ranged from 0.2 to 1.5 hours. A more drastic and significant change in the lag time of approximately 4 hours occurred at a temperature of 57°C, while at this temperature no significant loss of CFU was observed (Figure 1B). Apparently, the cells are able to fully recover from temperature treatments up to 57°C during an extended incubation in growth medium at 37°C. For more stringent temperature treatments in the range of 58 to 62°C the lag time ranged from 8 to 12 hours (Table 3). The relatively large increase in lag time from 4 to 8 hours for temperature treatments of 57 and 58°C coincides with the first loss of CFU at the latter temperature. Thus, for temperatures from 40 to 57°C the increase in lag time exclusively results from cellular recovery processes, while the increase in lag time at 58°C and higher results from both cellular recovery processes as well as the loss of viable cells in the culture. These results clearly show that the time delay to reach the mid log phase is a more sensitive measure to assess the effect of a heat exposure than CFU counts.
The effect of heat on B. subtilis cells was also studied with a number of cultivation-independent methods, including the fluorescence-based BacLight kit (Molecular Probes Inc.). This kit employs two nucleic acid stains, the green-fluorescent SYTO9 and the red-fluorescent propidium iodide stain, which differ in their ability to permeate bacterial cells. The cellular membrane is in general permeable to SYTO9, while propidium iodide penetrates only bacteria with damaged membranes. Propidium iodide reduces SYTO9 fluorescence when both dyes are present together. Thus, bacteria with intact membranes fluoresce green, while bacteria with damaged membranes fluoresce red . Microscopic inspection confirmed that this system can be applied to stain heat-exposed B. subtilis cells (data not shown). However, we found that the fluorescence of the SYTO9 fluorescent dye was relatively unstable under the conditions used, as the green fluorescence signal slowly decayed, while the fluorescence of the propidium iodide dye remained stable. Therefore, we subjected bacteria that were stained only with propidium iodide to flow-cytometric analysis. A calibration curve of mixtures of living and heat-killed bacteria (10 min at 96°C) showed a perfectly linear correlation between the percentage of heat-exposed bacterial cells (0, 20, 40, 60, 80 and 100%) and the number of events with relatively high red fluorescence intensity (data not shown). The heat-exposed bacterial cell samples were stained with propidium iodide and analyzed by flow-cytometry. The three-dimensional representation in Figure 1C shows the number of cells plotted as a function of fluorescence intensity in the red FL4 channel for each temperature treatment. Clearly, the heat-induced decrease of non-fluorescent cells coincides with an increase of membrane-permeable red-fluorescent cells all with Gaussian-shaped population distributions. The effect is most prominent at 62°C (apart from the control experiment at 96°C), at which 25% of the cell population is permeable to the propidium iodide dye. A comparison of CFU counts with data from propidium iodide stained bacteria (Figure 1B, C and Table 3) shows that the membrane integrity is a much more insensitive measure for the applied heat regime than the ability to form colonies. In the range of 59–62°C the majority of bacteria do still contain an intact membrane, while more than 99.9% of these bacteria lost their ability to form colonies. Apparently, heat-induced cell death precedes the loss of membrane integrity.
Ribosomal RNA stability
Genome-wide expression profiles
Heat-induced expression of gene clusters
Heat effects on transcriptional regulation
Besides the classical heat shock responses, we identified a number of novel responses to heat stress. Notably, the response induced at the lowest temperature, already maximal at 40°C, includes the expression of genes under control of σD, the transcription factor for motility. This response is only present at 40–50°C; it involves among others the fla/che operon, the lytA-C genes for autolysis and genes encoding methyl accepting chemotaxis proteins mcpA-B. Surprisingly, the heat-induced transcriptional response also includes a number of sporulation and competence genes, described in more detail below. One of the most prominent responses at high temperature was the derepression of large regulons involved in carbon metabolism, controlled by the CcpA (average of 109 genes), AraR (regulator of the arabinose operon; average of 12 genes) and IolR (regulator of the inositol metabolism; average of 13 genes) transcription factors. Note that both the inositol and arabinose operons are controlled by CcpA, a global regulator that directs the carbon flow (Figure 6).
We observed a comparable partial activation for genes involved in competence development. Competence is another stationary-phase differentiation process leading a cell-type to allow uptake of foreign DNA. The transcription factor for competence is ComK, which is post-translationally activated by the product of the comS gene, which is part of large operon that encodes surfactin biosynthesis and is in turn under control of the ComP/ComA two component system . Again only a minor sub-set of competence related genes was involved, including comS, srfAB, srfAC, srfAD, phrA and rapA under control of ComA and addA, addB, gidA, gidB, comFA, comGG, thdF, yyaA controlled by ComK (Figures 5 and 7).
Validation of transcript levels by quantitative PCR
Microarray results have been validated by the use of quantitative PCR. We selected the 23S rRNA gene as an internal standard based on its stability under the experimental conditions used in this study (Figure 2). In addition, we selected the cotJA gene, which encodes a spore coat protein, under control of transcription factor σE, which is known to be active during sporulation in the mother cell. Furthermore, the dnaK gene, which encodes a class I heat shock gene, a chaperone, under control of the transcriptional repressor HrcA. Finally, the gspA gene, which encodes the general stress protein GspA and is a general reporter gene for activity of the general stress response transcription factor σB . The Q-PCR analysis shows a good match between levels of heat-induction resulting from microarray and Q-PCR analysis of three typical heat-induction profiles (compare Figures 5, 7 and Table 3). The Q-PCR results of gspA transcripts confirm the transient induction of the σB-controlled genes at 40 and 50°C, the more stable induction of dnaK transcripts at elevated temperatures and the gradually increasing levels of cotJA transcripts as a representative of the σE regulon.
Assessment of cotJA transcript levels in σ-factor mutants
Heat-induced expression levels of cotJA in B. subtilis σ-factor mutants
168 1A1 (wt)
In this study we subjected the Gram-positive bacterium B. subtilis to a lethal temperature regime and analyzed the effects of the treatment in a comparative analysis using a number of cultivation-dependent and independent methods. The application of cultivation-independent methods as an alternative to assess bacterial viability is highly valuable for a number of reasons. A rapid measurement of the physiological state of the bacterium indicating whether it survived a certain processing condition (heat, high pressure, antimicrobial compounds) would save enormous amount of time and labour, currently lost in the classical routine of plating and counting. The timely and laborious plating method renders high-throughput screenings an almost impossible mission, although the recently reported miniaturization of the culture plate to a micro-Petri dish holds great promise for the future .
Any cultivation-based method results by definition in an estimate of the survivors after a treatment under predefined growth conditions, which are usually very different from those present during the actual processing conditions. A rapid measurement of the physiological state of the bacterium would be beneficial, as it reflects the direct cellular response towards processing conditions and could yield information on the successful survival strategy of the bacterium. This could allow a rational design for preferred adaptations in processing conditions aimed at a more efficient bacterial inactivation for harmful microbes or survival for beneficial ones.
The need for cultivation-independent methods to assess bacterial cell viability is most evident when microbes only grow in their natural habitat and colony formation is absent or extremely slow in any of the available growth media. Likewise, the enumeration of colonies is not applicable for bacteria present in dense populations of multiple species for which no selective media are available. In those cases the highly selective PCR-based assessment of specific gene transcripts, as further discussed below, can be applied to assess cell viability of one bacterial species in a complex population.
The workflow in this study starts from the assessment of bacterial viability upon exposure to a lethal heat treatment with conventional methods, including the enumeration of colonies, the determination of outgrowth curves and the widely used dead/live stain in the Bac Light kit . These experiments set a clear reference for molecular studies discussed below and roughly establish the 40 to 57°C temperature regime as inhibitory and sub-lethal, as judged from the extension in the lag phase and CFU counts (Table 3). The membrane integrity as judged from the dead/live stain is clearly a less sensitive indicator for the subtle heat regime applied here than colony enumeration or lag phase extension (Table 3), but more stringent temperature treatments do lead to a 100% dead (propidium-iodide permeable) cell population (Figure 1C).
The electrophoretic analysis of total RNA extracts clearly indicates a much higher heat tolerance of 23S rRNA molecule for the applied heat regime than 16S rRNA molecule, as one can assume that constant level of 23S rRNA cannot exclusively result from higher 23S rRNA synthesis rates (Figure 2). The difference in heat susceptibility of 23S and 16S rRNA molecules is commonly observed among prokaryotes, including B. subtilis, Staphylococcus aureus and Salmonella typhimurium [27–29]. The results clearly show that B. subtilis is able to fully recover from the observed heat-induced rRNA degradation, as significant cell death only occurs from 58°C exposures, while a major part of 16S rRNA has been degraded at 57°C (Table 3). Therefore, the level of 16S rRNA is not a very good indicator for cell viability under these conditions.
Additional file 1: Time-lapse movie of heat-induced expression profiles. The time-lapse movie of expression profiles was constructed by the use of Windows Movie Maker version 5.1 (Microsoft Inc.). All genes were sorted in the gene order present on the B. subtilis genome and presented in 32 columns of 96 genes each. Only those genes that contained a detectable transcription level at 30°C were analyzed. This filter led to a total of 75% of the genes (3072 out of 4100 genes). The 29 frames used for the movie consist of 8 measurements from 0–12 minutes and 21 interpolations. The parameters time, temperature and CFU counts were inserted in the microarray images, which were generated with the TIGR MultiExperimentViewer version 4.0.01. The 2log ratios were colour-coded, ranging from -3 (green) to +3 (red). Note that there is one region throughout the whole series of profiles with nearly identical expression levels to the reference profile at 30°C. This region contains 67 ORFs from rplK to rpsL and encodes at least 47 different ribosomal proteins. (WMV 2 MB)
In addition, gene clusters were identified containing the classic heat shock operons class I and class III (h, a, Figure 5), while the class II heat shock or σB regulon was only identified in the T-profile analysis as a result of the widespread distribution of genes from this regulon over the genome. The cluster of genes known to be highly induced upon stress is the mobile genetic element ICEbs1, which up regulation occurs upon global DNA damage (cluster d) . This fact and the up regulation of the cluster of pyrimidine genes (cluster g) may be part of the cell's response towards heat-induced DNA damage. The gene clusters (e, i, j) which are all involved in catabolic carbon metabolism may indicate the cell's attempt to deal with energy stress and generate ATP. Most surprising is the up regulation of a number of genes, known to be involved in stationary phase processes sporulation (cluster f) and competence development (cluster c). As their activation is usually strictly controlled by dedicated transcription regulators, we tested whether the observed up regulation of inner spore coat genes (cotJ operon) is dependent on sporulation-specific transcription factors. We confirmed that the heat-induced expression of the cotJ operon is dependent on sigF and sigE, but not on sigK genes, in agreement with the current model of the sporulation signal transduction cascade (Table 4). The activation of the sporulation sigma factors is normally coupled to the course of morphogenesis during spore differentiation. It is not clear at this point what mechanism causes the heat-induced short-cut in part of the sporulation program and whether the expressed gene products play an alternative role in heat resistance of the mother cell.
The observed cellular responses suggest that heat stress causes a wide variety of physiological stress. Which type of damage leads to cell death is not clear, but the microarray data indicate that the transcriptional machinery is at least still intact in cells that completely lost the ability to reproduce. This can be concluded from the results that show increasing transcript levels of most genes in the clusters a-j at temperatures from 58–62°C, whereas the majority of the cells died at these temperatures (compare Table 3 and Figure 5).
Next, we defined a number of criteria to identify transcripts that could act as biomarkers for cell viability: (i) high expression levels in living bacteria, (ii) relatively low expression levels in heat-inactivated bacteria, and (iii) genes that are widely conserved throughout the bacterial kingdom. We selected a set of 4 genes, all under the control of the HrcA regulator, which matched these criteria; they include grpE, dnaK, groES and groEL (see Figure 7). We observed that the above criteria for expression levels of these genes also matched in microarray studies on other bacterial species, where the cause of death was different: Lactobacillus plantarum exposed to ultra high-pressure treatments and Staphylococcus aureus exposed to biocides (data not shown).
Both dnak and groE operons are preceded by a SigA-type promoter, which ensures transcription of these genes under growth conditions without stress (in line with the dnaK expression levels in table 3), as protein folding is also needed under these conditions. A fascinating question involves the mechanism for the reproducible rapid decrease of these transcripts in bacteria that lost the ability to grow. In living bacteria, the CIRCE element (controlling inverted repeat of chaperone expression), located between the transcriptional and translational start sites of both dnaK and groESL operons , is involved in the regulation of both operons. In the current model the GroE chaperonin, the activator of the HrcA repressor, which binds to CIRCE, is titrated away in the presence of unfolded proteins in an ATP-dependent process, leading to inactivation of HrcA and enhancement of transcription of dnaK and groESL operons . Thus, the down regulation or degradation of these transcripts in excess of unfolded proteins asks for an extension of this model. No matter what regulatory mechanisms, the identification of 4 transcripts of chaperone genes as biomarker for viability needs to be further validated for a wide range of environmental conditions to determine the lethal conditions for which they are bona fide.
This study contains a systematic comparison between cultivation-based assessment of cell viability and cultivation-independent indicators for cell viability by sampling and analyzing heat-exposed B. subtilis cells going through the transition from culturable to unculturable cells. Accordingly, some viability indicators appear more sensitive than the conventional enumeration of colony forming units (CFU's), while others appear less sensitive measures: during heat stress the extension of lag time for outgrowth, the degradation of 16S rRNA and the induction of classic heat shock genes precede the loss of CFU's, while the loss of membrane permeability and the degradation of 23S rRNA occur after the loss of CFU's. The genome-wide transcriptional response indicates that heat stress causes a wide variety of physiological stresses. Which type or types of cell damage lead to cell death is not clear, but the microarray data indicate that the transcriptional machinery is at least still intact, as a subset of genes is still expressed in cells that lost the ability to reproduce.
The heat-induced transcriptional program includes known heat shock responses as well as the rapid expression of a small number of sporulation and competence genes, the latter only known to be active in the stationary growth phase. Further research needs to assess whether the activation of a small number of genes in these well-studied developmental programs is a result of deregulation due to heat damage or a meaningful, physiological event fitting in the cell's strategy to cope with heat stress. Notably, the transcription of a limited number of genes correlated with cell viability under the applied killing regime. The transcripts of the expressed genes in living bacteria – but silent in dead – include those of essential genes encoding chaperones of the protein folding machinery and may serve as molecular biomarkers for bacterial cell viability.
The authors thank Stanley Brul, Adriano Henriques and Daniel Kearns for generous strain donations. Tiemen van Eijndthoven, Michel Ossendrijver and Jojet Staal are gratefully acknowledged for their experimental contributions to this work. We thank the 4 anonymous referees for reviewing the manuscript.
- Barer MR, Harwood CR: Bacterial viability and culturability. Adv Microb Physiol. 1999, 41: 93-137.PubMedView ArticleGoogle Scholar
- Nystrom T: Not quite dead enough: on bacterial life, culturability, senescence, and death. Arch Microbiol. 2001, 176: 159-164.PubMedView ArticleGoogle Scholar
- Postgate JR: Death in microbes and macrobes. The survival of vegetative microbes. Edited by: Gray TRG, Postgate JR. 1976, Cambridge: Cambridge University Press, 1-19.Google Scholar
- Postgate JR: Viable counts and viability. Methods Microbiol. 1976, 1: 611-628.View ArticleGoogle Scholar
- Hecker M, Schumann W, Volker U: Heat-shock and general stress response in Bacillus subtilis. Mol Microbiol. 1996, 19: 417-428.PubMedView ArticleGoogle Scholar
- Helmann JD, Wu MF, Kobel PA, Gamo FJ, Wilson M, Morshedi MM, Navre M, Paddon C: Global transcriptional response of Bacillus subtilis to heat shock. J Bacteriol. 2001, 183: 7318-7328.PubMedPubMed CentralView ArticleGoogle Scholar
- Hamoen LW, Venema G, Kuipers OP: Controlling competence in Bacillus subtilis: shared use of regulators. Microbiology. 2003, 149: 9-17.PubMedView ArticleGoogle Scholar
- Stragier P, Losick R: Molecular genetics of sporulation in Bacillus subtilis. Annu Rev Genet. 1996, 30: 297-341.PubMedView ArticleGoogle Scholar
- Kort R, O'Brien AC, van Stokkum IH, Oomes SJ, Crielaard W, Hellingwerf KJ, Brul S: Assessment of heat resistance of bacterial spores from food product isolates by fluorescence monitoring of dipicolinic acid release. Appl Environ Microbiol. 2005, 71: 3556-3564.PubMedPubMed CentralView ArticleGoogle Scholar
- Oomes SJ, van Zuijlen AC, Hehenkamp JO, Witsenboer H, Vossen van der JM, Brul S: The characterisation of Bacillus spores occurring in the manufacturing of (low acid) canned products. Int J Food Microbiol. 2007, 120: 85-94.PubMedView ArticleGoogle Scholar
- de Koning W, van Dam K: A method for the determination of changes of glycolytic metabolites in yeast on a subsecond time scale using extraction at neutral pH. Anal Biochem. 1992, 204: 118-123.PubMedView ArticleGoogle Scholar
- Pieterse B, Jellema RH, Werf van der MJ: Quenching of microbial samples for increased reliability of microarray data. J Microbiol Methods. 2006, 64: 207-216.PubMedView ArticleGoogle Scholar
- Kunst F, Ogasawara N, Moszer I, Albertini AM, Alloni G, Azevedo V, Bertero MG, Bessieres P, Bolotin A, Borchert S, et al: The complete genome sequence of the gram-positive bacterium Bacillus subtilis. Nature. 1997, 390: 249-256.PubMedView ArticleGoogle Scholar
- Keijser BJ, Ter Beek A, Rauwerda H, Schuren F, Montijn R, Spek van der H, Brul S: Analysis of temporal gene expression during Bacillus subtilis spore germination and outgrowth. J Bacteriol. 2007, 189: 3624-3634.PubMedPubMed CentralView ArticleGoogle Scholar
- Saeed AI, Sharov V, White J, Li J, Liang W, Bhagabati N, Braisted J, Klapa M, Currier T, Thiagarajan M, et al: TM4: a free, open-source system for microarray data management and analysis. Biotechniques. 2003, 34: 374-378.PubMedGoogle Scholar
- Boorsma A, Foat BC, Vis D, Klis F, Bussemaker HJ: T-profiler: scoring the activity of predefined groups of genes using gene expression data. Nucleic Acids Res. 2005, 33: W592-595.PubMedPubMed CentralView ArticleGoogle Scholar
- Ter Beek A, Keijser BJ, Boorsma A, Zakrzewska A, Orij R, Smits GJ, Brul S: Transcriptome analysis of sorbic acid-stressed Bacillus subtilis reveals a nutrient limitation response and indicates plasma membrane remodeling. J Bacteriol. 2008, 190: 1751-1761.PubMedView ArticleGoogle Scholar
- Makita Y, Nakao M, Ogasawara N, Nakai K: DBTBS: database of transcriptional regulation in Bacillus subtilis and its contribution to comparative genomics. Nucleic Acids Res. 2004, 32: D75-77.PubMedPubMed CentralView ArticleGoogle Scholar
- Boulos L, Prevost M, Barbeau B, Coallier J, Desjardins R: LIVE/DEAD BacLight: application of a new rapid staining method for direct enumeration of viable and total bacteria in drinking water. J Microbiol Methods. 1999, 37: 77-86.PubMedView ArticleGoogle Scholar
- Hartigan JA, Wong MA: A K-means clustering algorithm. Applied Statistics. 1979, 28: 100-108.View ArticleGoogle Scholar
- Mogk A, Homuth G, Scholz C, Kim L, Schmid FX, Schumann W: The GroE chaperonin machine is a major modulator of the CIRCE heat shock regulon of Bacillus subtilis. Embo J. 1997, 16: 4579-4590.PubMedPubMed CentralView ArticleGoogle Scholar
- Zuber U, Schumann W: CIRCE, a novel heat shock element involved in regulation of heat shock operon dnaK of Bacillus subtilis. J Bacteriol. 1994, 176: 1359-1363.PubMedPubMed CentralGoogle Scholar
- Eichenberger P, Fujita M, Jensen ST, Conlon EM, Rudner DZ, Wang ST, Ferguson C, Haga K, Sato T, Liu JS, Losick R: The program of gene transcription for a single differentiating cell type during sporulation in Bacillus subtilis. PLoS Biol. 2004, 2: e328-PubMedPubMed CentralView ArticleGoogle Scholar
- Wang ST, Setlow B, Conlon EM, Lyon JL, Imamura D, Sato T, Setlow P, Losick R, Eichenberger P: The forespore line of gene expression in Bacillus subtilis. J Mol Biol. 2006, 358: 16-37.PubMedView ArticleGoogle Scholar
- Antelmann H, Bernhardt J, Schmid R, Hecker M: A gene at 333 degrees on the Bacillus subtilis chromosome encodes the newly identified sigma B-dependent general stress protein GspA. J Bacteriol. 1995, 177: 3540-3545.PubMedPubMed CentralGoogle Scholar
- Ingham CJ, Sprenkels A, Bomer J, Molenaar D, Berg van den A, van Hylckama Vlieg JE, de Vos WM: The micro-Petri dish, a million-well growth chip for the culture and high-throughput screening of microorganisms. Proc Natl Acad Sci USA. 2007, 104: 18217-18222.PubMedPubMed CentralView ArticleGoogle Scholar
- Miller LL, Ordal ZJ: Thermal injury and recovery of Bacillus subtilis. Appl Microbiol. 1972, 24: 878-884.PubMedPubMed CentralGoogle Scholar
- Sogin SJ, Ordal ZJ: Regeneration of ribosomes and ribosomal ribonucleic acid during repair of thermal injury to Staphylococcus. J Bacteriol. 1967, 94: 1082-1087.PubMedPubMed CentralGoogle Scholar
- Tomlins RI, Ordal ZJ: Precursor ribosomal ribonucleic acid and ribosome accumulation in vivo during the recovery of Salmonella typhimurium from thermal injury. J Bacteriol. 1971, 107: 134-142.PubMedPubMed CentralGoogle Scholar
- Auchtung JM, Lee CA, Monson RE, Lehman AP, Grossman AD: Regulation of a Bacillus subtilis mobile genetic element by intercellular signaling and the global DNA damage response. Proc Natl Acad Sci USA. 2005, 102: 12554-12559.PubMedPubMed CentralView ArticleGoogle Scholar
- Kearns DB, Losick R: Cell population heterogeneity during growth of Bacillus subtilis. Genes Dev. 2005, 19: 3083-3094.PubMedPubMed CentralView ArticleGoogle Scholar
- Boylan SA, Redfield AR, Brody MS, Price CW: Stress-induced activation of the sigma B transcription factor of Bacillus subtilis. J Bacteriol. 1993, 175: 7931-7937.PubMedPubMed CentralGoogle Scholar
- Henriques AO, Beall BW, Roland K, Moran CP: Characterization of cotJ, a sigma E-controlled operon affecting the polypeptide composition of the coat of Bacillus subtilis spores. J Bacteriol. 1995, 177: 3394-3406.PubMedPubMed CentralGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.