| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Swiss Federal Institute for Aquatic Science and Technology, PO Box 611, Überlandstrasse 133, CH-8600 Dübendorf, Switzerland
Correspondence
Thomas Egli
egli{at}eawag.ch
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The GEO accession number for the microarray data reported in this paper is GSE4706.
Four supplementary tables with additional gene expression data are available with the online version of this paper.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
) and E. coli has been extensively investigated in the laboratory with respect to its physiology, biochemistry and genetics. Despite this, surprisingly little is known about how this bacterium grows, competes and survives in its primary (i.e. warm-blooded hosts) (Conway, 1995; Macfarlane & Macfarlane, 1997, 2004) and secondary habitats (i.e. soil and water) (Rozen & Belkin, 2001; Savageau, 1983). In contrast to nutrient-rich laboratory cultures, growth conditions in the aqueous environment, such as surface, ground and drinking water, are much harsher, and concentrations of utilizable carbon sources are typically in the low µg l–1 range (Egli, 1995; Morita, 1993; Münster, 1993). Similarly, access to nutrients in soil is restricted most of the time (Williams, 1985). Hence, growth of heterotrophic microbes in natural environments is carbon/energy-limited most of the time.
It has long been the dogma that cells of E. coli cannot survive in secondary habitats and that they die off quickly. However, much controversial information can be found in the literature on survival and growth of this bacterium in different natural environments (Hazen & Toranzos, 1990). Recent developments in detection techniques indicate longer survival in the environment than observed with the traditional plate count method (Garcia-Armisen & Servais, 2004). For example, growth of E. coli was observed particularly in warmer and polluted fresh water (McFeters, 1990), and Rozen and coworkers have reported that E. coli does not die off quickly, but survives long periods of starvation and can even grow in sea water (Rozen & Belkin, 2001).
There is not much conclusive information available on the physiological state of E. coli in its primary and secondary habitats (Rozen & Belkin, 2001; Savageau, 1983). One of the experimental approaches to understanding the behaviour of this bacterium under carefully controlled nutritional conditions is the use of continuous culture (Egli, 1995; Macfarlane et al., 1998). In particular carbon/energy-limited continuous culture has been used to study the kinetic and physiological properties of micro-organisms during slow growth. Interestingly, during the cultivation of E. coli in glucose-limited continuous culture two phases of adaptation to such growth conditions were detected and they can be identified based on kinetic and physiological properties of the cells (Wick et al., 2001). A first phase, the so-called short-term adaptation phase, follows immediately after the culture has reached the steady state with respect to biomass and residual glucose concentration. This phase lasts no longer than approximately 50–70 generations. Changes in the proteome/cellular composition during this phase occur solely at the physiological level by regulation of gene expression at the transcriptional and translational level. Generally, an increased expression of alternative carbon/energy and nitrogen source transport systems and increased expression of components of stress response regulons are observed (Hua et al., 2004; Ihssen & Egli, 2004; Wick et al., 2001). In strain K12 this phase is characterized by a constant affinity for glucose, with a Ks of 
400 µg l–1, which is a slight improvement compared to growth at glucose-excess concentrations in batch culture (Ks 600 µg l–1). This short-term adaptation phase is followed by a mutant selection phase (50–500 generation times in the chemostat) in which mutants with increased affinity for the growth-limiting glucose are enriched (Wick et al., 2002). This so-called long-term adaptation leads to a 10–15-fold increase in the affinity for glucose. It was shown that this phase is characterized by the selection of regulatory mutants affected in the expression of the mgl/mal system (Notley-McRobb & Ferenci, 1999a, b). These changes have been well documented at the proteome (Wick et al., 2001) and catabolome (Ihssen & Egli, 2005) levels, as well as the genetic level (Notley-McRobb & Ferenci, 1999a, b).
Therefore, the analysis of the transcriptome is an additional tool for an improved understanding of the bacterial response to carbon/energy limitation. The advent of microarray technology has increased the interest in the study of the global gene expression under different environmental conditions. The understanding of global gene expression profiles under different nutritional conditions is very important for the understanding of the growth of bacteria in their natural environment.
The purpose of the work presented here was to identify the genes that are differentially expressed during the adaptation to low glucose concentration in glucose-limited continuous culture as compared to growth with excess glucose during cultivation in batch culture. Special attention was given to the uptake systems and the regulatory proteins involved in the evolution occurring during adaptation to low glucose concentrations. The results obtained in this study are compared with the results obtained from the proteome (Wick et al., 2001) and the catabolome (Ihssen & Egli, 2005), where the same process was documented by different analytical methods. One of the major advantages of the microarray technology is that all genes can be analysed and compared, whereas with the other two methods only a limited set of proteins of physiological information can be obtained. However, for a complete rounded picture one needs all the different methods that give information at complementary levels.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Growth was monitored by three different methods. Cell number was measured by plating on tryptic soy agar (TSA; Biolife) plates after appropriate dilution and, for assessing purity, on ECD-MUG agar plates (Biolife). Dry cell weight was measured by filtration onto 0.22 µm membrane filters (Millipore) and drying to constant weight at 105 °C. Furthermore, culture density was monitored spectrophotometrically by measuring optical density at 546 nm (OD546) in a 1 cm cuvette (Uvikon 930; Kontron).
Cultivation conditions.
For batch cultures of E. coli mineral medium with 1 g glucose l–1 was used. For glucose-limited continuous cultures the concentration of glucose in the feed medium was reduced to 0.1 g l–1 in order to be able to follow the kinetics of adaptation at the level of residual glucose concentrations (Senn et al., 1994). Cultivation in both batch and continuous culture was performed in a computer-controlled glass and stainless steel bioreactor with a 2 litre working volume (MBR). The stirrer speed was set to 800 r.p.m., oxygen concentration was kept over 90 % of air saturation, the temperature was controlled at 37±0.1 °C and pH was maintained at 7.0±0.1 by automatic addition of a mixture of NaOH and KOH (0.5 M each).
For each growth experiment, cultures were always started freshly from the same cryo-vial stored at –80 °C to avoid as much as possible the picking up of mutants due to adaptation to the growth medium. Cells were streaked out first on TSA plates and after 24 h a colony was transferred into 5 ml mineral medium with 4 g glucose l–1. When this first pre-culture had reached an OD546 of 0.3–0.4, 1 ml of the culture was transferred to 100 ml of the same medium. Cultures in bioreactors were inoculated with 60 ml of this second pre-culture when it reached an OD546 of 0.3. Growth was followed by measuring OD546 at regular intervals to confirm that cells were growing at µmax (0.63±0.02 h–1). A sample for transcriptome analysis was taken from the batch exponential phase at an OD546 of 0.4, and immediately after sampling the medium flow was started for continuous cultivation with a dilution rate of 0.3 h–1. Steady state with respect to biomass and residual glucose concentrations was reached approximately 15 h after this manipulation. At 40 and 500 h after switching to the glucose-limited continuous culture mode, 800 ml of culture liquid was withdrawn from the chemostat and immediately cooled on crushed ice, and cells were collected by centrifugation at 4 °C. From the resulting cell pellet RNA was instantaneously isolated (see below). Residual glucose concentrations were monitored during the continuous cultures run; samples were filtered as described earlier (Senn et al., 1994) and analysed by ion chromatography (Ihssen & Egli, 2004). The Ks and the sequence of mutant populations during the adaptation process were simulated using AQUASIM according to Wick et al. (2002).
RNA isolation and synthesis of cDNA.
Total RNA from E. coli cells was isolated as described by Sambrook et al. (1989). RNA in samples was quantified spectrophotometrically by measuring absorbance at 260 nm, and purity (freedom from DNA) was checked by gel electrophoresis. Synthesis of cDNA from RNA was performed with the CyScribe First-Strand cDNA Labelling Kit (Amersham Bioscience). Reverse transcription was performed using 25 µg total RNA (maximally in 10 µl) and 1 µl random nonamer primers. The volume of the assay mixture was adjusted to 11 µl with RNase-free water, then the assay mixture was incubated for 5 min at 70 °C, followed by incubation for 10 min at room temperature to allow the primers to anneal with the RNA. After cooling to room temperature, the reagents for the labelling reaction were added. After the addition of 4 µl 5x CyScript buffer, 2 µl 0.1 M DTT, 1 µl dNTP mixture, 1 µl 0.5 mM either Cy3-labelled or Cy5-labelled dCTP (Amersham Bioscience) and 1 µl CyScript reverse transcriptase (100 U µl–1), the final reaction volume was 20 µl. The control cDNA (from batch-grown cells) and the probe cDNA (from either the 40 h old or the 500 h old chemostat culture) were labelled differently, with the control cDNA always labelled with Cy5. The labelling reaction was performed at 42 °C for 1.5 h, followed by RNA degradation and cDNA purification. The RNA was degraded by addition of 2 µl 2.5 M NaOH; the mixture was then heated at 65 °C for 10 min and subsequently neutralized with 10 µl 2 M HEPES buffer. The control and probe cDNA obtained were pooled and purified on the same column of the MinElute Gel Extraction Kit (Qiagen) to avoid differences in extraction yields.
DNA microarray, hybridization and washing.
Slide microarrays were purchased from MWG-Biotech. The MWG E. coli array contains 4288 gene-specific oligonucleotide probes representing the complete E. coli K-12 genome. The purified cDNA was concentrated to 5 µl and was mixed with 120 µl hybridization buffer (MWG-Biotech), heated to 95 °C for 3 min and cooled on ice for 3 min. The hybridization mixture was then added to the microarray slide and covered with a coverslip. The hybridization slide was incubated overnight at 42 °C. After the hybridization step the slide was washed three times, the first time for 5 min in 2x SSC/0.1 % SDS, the second time for 5 min in 1x SSC, and finally for 5 min in 0.1x SSC. SSC buffer was prepared as a 20x solution containing 0.3 M sodium citrate and 3 M NaCl at pH 7.0. The slides were dried by centrifugation at room temperature for 2 min at 500 g.
Image and data analysis.
Microarray slides were scanned using an Affymetrix 428 Array Scanner. Spot intensities and corresponding background signals were quantified with the Affymetrix Jaguar software version 2.0. Further data analysis was performed with the program GeneSpring from Silicon Genetics. Induction factors were calculated from the Cy3 and Cy5 signal intensities of the spot. Spots with signal intensity below a value of 50 were excluded from the analysis and the minimal induction factor was set to 0.01. The normalization was performed with the 50th percentile distribution of remaining spots after background correction. The mean value of the induction factors of a specific gene was calculated from three replicates. Biological experiments were carried out three times, which provided three biological repeats. Data from the independent experiments were combined; genes that were differentially regulated 
2 and 
0.5 (t-test, P0.05) were defined as being statistically significant.
Real-time RT-PCR.
cDNA was synthesized with SuperScript reverse transcriptase (Invitrogen). The mixture for RNA transcription was prepared for 40 reactions, mixing 0.8 µg total RNA (DNase-digested) with 2 µl random hexamer primers (Invitrogen). The total volume was adjusted to 9 µl with RNase-free water, followed by incubation for 5 min at 65 °C to melt secondary structures, and then cooling for 2 min on ice to anneal the primers. Subsequently, reverse transcriptase reaction components were added, consisting of 0.5 µl RNase inhibitor (Sigma), 4 µl dNTP mixture (2.5 mM each), 2 µl 0.1 M DTT, 4 µl 5x reaction buffer and 0.5 µl SuperScript reverse transcriptase. The reaction mixture was mixed and spun down briefly and was then incubated for 10 min at 25 °C for annealing (90 min at 39 °C) reverse-transcribed mRNA. The reaction was stopped by incubation at 95 °C for 5 min. The total volume was adjusted to 40 µl before starting real-time PCR. Real-time PCR was conducted according to the manufacturer's instructions; the assay contained 12.5 µl SYBR Green PCR Master Mix (Applied Biosystems), 9.5 µl RNase-free water, 1 µl forward primer, 1 µl reverse primer and 1 µl cDNA (20 ng), or the same amount of RNA as negative control. Primers (Microsynth) were designed with the software Primer Express v2.0 (Applied Biosystems) for 12 genes (Table 1), and their optimization (to avoid concatemers or similar problems) was performed by mixing different concentrations of forward and reverse primers. The optimized primer concentration was used to analyse RNA samples. The real-time PCR reaction was conducted using ABI Prism 7000 (Applied Biosystems). The results were normalized using the gene rrsB as endogenous control. Real-time PCR data were analysed with ABI Prism 7000 SDS software version 1.0 according to the System User Bulletin #2 ‘Relative Quantification of Gene Expression’ (P/N 4303859) from Applied Biosystems.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The affinity of the culture for glucose was monitored throughout the continuous culture process and at the two time-points it differed by a factor of more than 10 (Ks from 463±46 µg l–1 to 42±13 µg l–1) (Fig. 1), demonstrating the adaptability of E. coli to low glucose concentrations.
|
). In E. coli growing under glucose-excess conditions, this sugar enters the cell via the porins OmpC/OmpF (outer membrane) and the transport protein PtsG (inner membrane). However, a high-affinity pathway is induced during limited availability of glucose. Interestingly, it is not the PtsG, but transporters from the maltose (mal/lamB) and galactose (mgl/gal) operons which seem to be involved in high-affinity uptake of glucose (Ferenci, 1996). A strongly increased expression of mglB and malE (galactose- and maltose-binding proteins, respectively) in chemostat-grown cells (Wick et al., 2001) confirmed the importance of the mgl and mal gene products for high-affinity glucose uptake. Among the upregulated genes monitored in our study, the transcript levels of the maltose and galactose high-affinity uptake systems were upregulated significantly (Table 2), which is in agreement with earlier proteome data (Wick et al., 2001). Surprisingly, ptsH and crr (not included in Table 4 because their expression factors were >–10, see supplementary data), two genes of the glucose PTS, were slightly downregulated. This result could be a consequence of the fact that the main flux of glucose occurs via the maltose and galactose high-affinity systems under the conditions tested (Death & Ferenci, 1993b; Death et al., 1993a; Ferenci, 1996).
|
|
), due to (an unknown) genetic adaptation. Multiple mal- and mgl-regulatory mutations are selected for and accumulate during growth in glucose-limited chemostats. The selected population is heterogeneous with respect to the location of the mutation but all of the mutations result in an increased affinity for glucose uptake (Notley-McRobb & Ferenci, 1999a, b). However, the major mutations were found in genes of the maltose and galactose operons (Notley-McRobb & Ferenci, 1999a, b), and in rpoS (Notley-McRobb et al., 2002). The maltose and galactose systems were overexpressed after 500 h in a glucose-limited chemostat when compared with cells collected from the physiological adaptation phase. At the mRNA level, the maltose and galactose high-affinity uptake systems were also upregulated, but their expression factor was less pronounced than after 40 h of continuous culture (Table 2
).
Changes in mRNA levels during the physiological (short-term) adaptation
As a result of the physiological adaptation, which was examined here after 40 h of glucose-limited continuous culture, 340 genes (see Supplementary Table S1, available with the online version of this paper) were upregulated by a factor of 2 in comparison to batch culture (Fig. 2a). On the other hand, 502 genes were detected with downregulated expression levels by a factor of –2 (Supplementary Table S3), of which more than 37 % (187 genes) were of miscellaneous (few) or hypothetical (most) function (Fig. 2b). Among the upregulated genes, the transcript levels of 141 genes were significantly elevated above a factor of 5 (Table 3). In this group of genes the maltose and galactose operons were omitted, because these are summarized in Table 2. The expression of 81 genes decreased by a factor of 10 or more (Table 4).
|
|
). Only a few genes of transport systems were downregulated significantly under the conditions tested, and the majority of them are of unknown function.
The physiological response to low glucose concentration led to the upregulation of several genes encoding enzymes of the central metabolism, particularly of enzymes of the TCA pathway and glyoxylate bypass. Among these genes, the highest expression level was found for prpC, encoding methylcitrate synthase and belonging to the prpBCDE operon. Methylcitrate synthase can take over the function of citrate synthase (GltA) and is known to be transcriptionally regulated by cAMP and 
N (Lee et al., 2005). An important gene for the TCA cycle, acetyl-CoA synthetase, which converts acetate to acetyl-CoA, was upregulated to significant levels. Surprisingly, fumarate reductase (FrdABCD), which is known to work under anaerobic conditions (Guest, 1981), was also upregulated significantly, whereas transcription of succinate dehydrogenase (SdhABCD), which is negatively regulated by RpoS (Patten et al., 2004), was downregulated. Two enzymes involved in the glyoxylate shunt were affected differently: whereas malate synthase G (GlcB) was upregulated, isocitrate dehydrogenase (IcdA) was downregulated. The glyoxylate shunt has been proposed to play an important role under glucose-limited conditions (Fischer & Sauer, 2003; Wick et al., 2001). In addition, mRNAs of two aldehyde dehydrogenases were found at high levels: AldA, which is known to function on a broad spectrum of substrates (Limon et al., 1997), and AldB, which is specific for acetaldehyde and is growth phase-dependent (Ho & Weiner, 2005).
As expected, transcription of genes involved in chemotaxis and the motility machinery were upregulated under glucose-limited conditions, with the purpose to allow cells to move and to search for a better environment. Consistent with this, increased transcription of the chemotaxis regulator (CheY) was found together with that of some periplasmic binding proteins that are also involved in chemotaxis towards the transported carbon sources. Also several genes belonging to class II flagella biosynthesis were upregulated, which is in contrast to a recent transcriptome study where these genes were found to be negatively regulated by RpoS (Patten et al., 2004). The upregulation of genes involved in the flagellar machinery in an rpoS mutant might be due to competition, where more 70 can bind to RNA polymerases, resulting in housekeeping genes being transcribed at a higher rate. The competition can involve also F, which is involved in flagellar machinery expression and shows growth-rate-dependent expression (Makinoshima et al., 2003). At a dilution rate of 0.3 h–1 rpoS was upregulated fourfold, which is consistent with the fact that the RpoS protein is expressed at half its maximum level under glucose-limited continuous cultures at this dilution rate (Ihssen & Egli, 2004). Many genes reported to belong to the s regulon in other transcriptome studies (Lacour & Landini, 2004; Patten et al., 2004; Weber et al., 2005) were upregulated under the condition tested. All genes reported earlier to be transcribed in a s-dependent fashion (Patten et al., 2004; Weber et al., 2005) were of hypothetical function, where the most expressed operon of this regulon was the ydcSTUV. YdcS was reported to be highly expressed in the phase of physiological adaptation under the conditions tested in the proteome analysis study (Wick et al., 2001). Many genes encoding stress proteins that are not s-dependent were expressed at higher levels in cells grown for 40 h under glucose-limited as compared to batch conditions. One of them is a cold-shock protein (CspD).
Most of the genes encoding proteins belonging to the protein synthesis apparatus and elongation factors were downregulated. This is consistent with the fact that the specific growth rate of the chemostat culture (0.3 h–1) is only half that in the exponential growth phase in batch cultures (0.63 h–1) (Gourse et al., 1986; Miura et al., 1981). Other operons encoding DNA replication, RNA degradation and purine and pyrimidine nucleotide de novo biosynthesis were also negatively affected by the slow growth in continuous culture. Surprisingly, some genes belonging to the virulence group were downregulated as a physiological response to glucose-limited conditions. Also genes encoding proteins involved in the regulation of antigen production, antigen synthesis, haemolysin and colicin production, as well as colicin uptake, were found in this group. Furthermore, several stress proteins were downregulated under the tested condition, such as chaperones, cold- and heat-shock proteins, damage-inducible proteins and resistance proteins; overall, CspF and YdgF were those most affected. Interestingly, some energy-related genes were also downregulated during slow growth in glucose-limited culture, among them cytochrome synthesis proteins, electron transport complexes and the ATP synthesis machinery.
Changes in mRNA levels during the genetic (long-term) adaptation
After 500 h of cultivation in glucose-limited continuous culture, expression of 83 genes was induced at significant levels (2, Supplementary Table S2) and the major group found was the one consisting of genes with miscellaneous or hypothetical function (Fig. 3a). Genes that were either expressed with an induction factor >5, or had a lower induction factor than 5 but were transcribed at high expression levels in the 40 h old culture, are listed in Table 3. In the evolved population 233 genes were downregulated (Supplementary Table S4), with all of them well distributed in the five groups mentioned (Fig. 3b). Genes with expression levels lower than a factor of 10 in long-term adapted cultures of E. coli under glucose-limited conditions are listed in Table 4.
|
In the evolved population also for the downregulated genes the same general pattern (affecting the same gene classes as found in the physiological adaptation) was found, where the same genes were affected with a higher repression factor compared to the physiological response (Table 4). Interestingly, expression levels of the PTS proteins Hpr and PtsI were strongly reduced. These proteins are common to all PTS sugar uptake systems. This downregulation might be a consequence of the expression of a multitude of high-affinity uptake systems, which confer nutritional flexibility to the catabolome.
Expression levels of RpoS were found to be only half of those detected after 40 h under glucose-limited conditions, although proteins (HdeA, HdeD and YhiE) belonging to the s regulon (Patten et al., 2004; Weber et al., 2005) were highly repressed in evolved cells. However, the genes most affected were members of the protein synthesis apparatus and therefore this effect is probably mainly a consequence of the reduced growth rate.
Real-time RT-PCR
The microarray data were validated using the real-time RT-PCR method. Genes to be confirmed were selected based on the proteome analysis (Wick et al., 2001) and the first microarray data in this study. Generally, the results obtained with this technique confirmed the increase detected with the transcriptome analysis (mglA, malE, rbsB, dppA and katE) and in the proteome study (Wick et al., 2001). Disagreement was apparent between the microarray and real-time PCR data for six genes (mglB, mglC, rpoS, lamB, aldA and ugpB), for which expression levels were higher with the real-time PCR technique (Fig. 4). Data obtained with the real-time PCR are probably more reliable, because this technology is more sensitive for low amounts of mRNA.
|
). Expression levels of the galactose operon were increased by a factor of 3 between 40 and 500 h in glucose-limited continuous culture, and these results are in good agreement with the proteome analysis (Wick et al., 2001). Some discrepancy was obtained with the RpoS-dependent katE gene, which was slightly downregulated in the transcriptome analysis whereas real-time PCR analysis indicated a 10-fold downregulation. Enzyme analysis again showed a decrease of catalase HPII (KatE) specific activity (Visick & Clarke, 1997) that was five times less than that in the starting culture (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Wick et al., 2001). This pattern of protein expression under glucose limitation is not specific for this sugar only, and was reported also for carbon-limited growth in continuous culture with other sugars and complex medium (Ihssen & Egli, 2005), and even for batch growth with unfavourable carbon substrates supporting only low specific growth rates (Liu et al., 2005).
Carbon/energy-limited growth at suboptimal rates, and not the presence of an appropriate substrate, seems to be the major factor for the high expression of different uptake systems (Ihssen & Egli, 2005; Liu et al., 2005; Wick et al., 2001). Cultures of E. coli growing under glucose-limited conditions also express mRNAs for enzymes of central metabolic and respiratory pathways to elevated levels (Hua et al., 2003). The same pattern was observed also with the respiration rates for different substrates (Ihssen & Egli, 2005). This upregulation of genes belonging to transport systems, specific breakdown enzymes and enzymes that facilitate the electron transport from substrate to terminal electron acceptors might be a consequence of the decreased expression of rRNA due to the lower growth rate (Gourse et al., 1986; Kurland & Maaloe, 1962; Liu et al., 2005; Miura et al., 1981). The transcriptome data obtained in this study fit with the catabolic derepression reported earlier using either different ‘global’ techniques, or focusing only on specific genes/proteins under glucose-limited conditions, such as glucose-limited chemostat and fed-batch conditions (Hua et al., 2004; Notley & Ferenci, 1995; Wick et al., 2001; Yoon et al., 2003). Our results concur with the proteome data and periplasmic protein analysis under glucose-limited continuous culture, where DppA, MalE, MglB and RbsB expression levels are upregulated (Ihssen & Egli, 2005; Wick et al., 2001). These and other transporter proteins expressed here at high levels were reported to be under the control of cAMP (Gosset et al., 2004; Salgado et al., 2004; Tan et al., 2001).
The most important global regulator in the adaptation to low carbon and energy sources seems to be cAMP. This global regulator induces most high-affinity transport systems, as well as many starvation proteins (Matin et al., 1989). The cAMP concentration is high in cells growing in glucose-limited chemostat cultures, whereas under glucose-excess conditions cAMP concentrations are low (Notley-McRobb et al., 1997). The derepression of ABC transporters is caused by a combination of increased cAMP concentration and induction by specific sugar endoinducers (Ferenci, 1999a; Tweeddale et al., 1998). For example, the maltose and galactose operon are induced under glucose-limited conditions by the production of intracellular maltotriose (Notley & Ferenci, 1995) and galactose (Ferenci, 1996), respectively.
The expression pattern found during the short-term response, also referred to as physiological adaptation, was observed as well after the long-term (genetic) adaptation to glucose (carbon/energy) limitation. These findings are in agreement with the 2D gel electrophoresis analysis performed under the same conditions (Wick et al., 2001) and catabolome analysis using Biolog as reported by Ihssen & Egli (2005) recently (data not shown). In populations of E. coli evolved under glucose-limited conditions, fewer genes were found with increased or decreased expression.
In addition to the induction of alternative transport systems, the carbon and energy metabolism under glucose-limited conditions was affected seriously. Proteome (Wick et al., 2001) and carbon fluxes analysis (Fischer & Sauer, 2003) indicated the presence of a new metabolic cycle, the PEP-glyoxylate cycle, that appears to increase growth efficiency under glucose-limited conditions. The PEP-glyoxylate cycle, which was known to work only in the presence of acetate or fatty acids (Cronan & LaPorte, 1996), might confer more metabolic flexibility to the cells (Ihssen & Egli, 2005). Isocitrate lyase (AceA) and isocitrate dehydrogenase (IDH) are two key enzymes required for this bypass during growth on acetate or fatty acids as sole carbon and energy sources (Cronan & LaPorte, 1996). IDH is known to be regulated by IDH kinase/phosphatase (AceK), which phosphorylates and inactivates IDH (Koshland et al., 1985). In our study, IDH inactivation was monitored as downregulation of its expression and not by an upregulation of AceK, which is involved in the inactivation of IDH. The second key enzyme of the glyoxylate shunt, isocitrate lyase, was strongly expressed under glucose-limited and fed-batch conditions (Hua et al., 2004; Wick et al., 2001; Yoon et al., 2003), but not in our experiment.
Interestingly, we found some upregulated genes of the citrate cycle that were not induced in previous studies (Hua et al., 2004; Yoon et al., 2003). Methylcitrate synthase, which was the most expressed gene during the physiological adaptation, is thought to take over the function of citrate synthase to produce citrate. The proteins making up fumarate reductase, which is known to be expressed under anaerobic conditions (Cronan & LaPorte, 1996), are very similar to those of succinate dehydrogenase, which acts under aerobic conditions, and they can take over the function of each other (Guest, 1981). Fumarate reductase produces high amounts of superoxide at low succinate concentrations, whereas superoxide formation is low at enhanced concentrations of succinate (Cecchini et al., 2002). Hence, the PEP-glyoxylate cycle, switched on under glucose-limited conditions, produces high concentrations of succinate, and this again might reduce superoxide production by fumarate reductase. An alternative explanation is linked to the expression of RpoS; fumarate reductase expression is positively controlled by s whereas succinate dehydrogenase expression is negatively regulated by this sigma factor (Weber et al., 2005). Many genes belonging to the TCA cycle, except those of the sucCD operon, are downregulated by RpoS (Patten et al., 2004).
These observations might explain overexpression in our microarray data of those genes that are not under negative control of the stress response sigma factor. During the first 40 h in glucose-limited continuous culture at a dilution rate of 0.3 h–1 RpoS is present in high amounts (Ihssen & Egli, 2004) and hence no genes belonging to this regulon were downregulated. However, evolved cells carried an RpoS attenuation phenotype and as a consequence a large part of the s-dependent genes were downregulated. The expression level of rpoS was partially reduced in 500-h-old cells, but the protein translation and stability might be affected also, resulting in the attenuated phenotype. It is well known that the synthesis of RpoS is not limited to gene transcription, but it is affected at all levels starting from mRNA stability via translation, to protein stability and proteolysis (Hengge-Aronis, 2002). The attenuation or the deletion of RpoS was monitored under glucose-limited conditions (Notley-McRobb et al., 2002) and in prolonged cultivation of stationary-phase cells (Farrell & Finkel, 2003). The phenotype resulting from this attenuation or deletion of RpoS exhibited a higher µmax, Glc and an improved Ks, Glc compared to the wild-type strain (Wick et al., 2002). rpoS mutant strains arising under carbon-limited conditions possess a growth advantage and these strains are strongly selected for under such conditions (Notley-McRobb et al., 2002). This phenomenon has also been observed in batch stationary-phase cultures, where they have been referred to as GASP (Growth Advantage in Stationary Phase) mutants (Farrell & Finkel, 2003; Notley-McRobb et al., 2002; Zambrano & Kolter, 1996). Genetic polymorphism in the rpoS genomic region is common in different E. coli populations (Ferenci, 2003). High levels of s and negative effects on transporter expression reduce the competitiveness of the strain in low-glucose environments (King et al., 2004). The attenuation or deletion of s reduces the competition between this sigma factor and 70 (Farrell & Finkel, 2003; Nyström, 2004), which is necessary for an enhanced transcription of genes needed for the uptake and utilization of carbon sources.
The mRNA pools are highly dynamic, and one of the major problems for microarray analysis is to collect mRNA from cells in the same physiological state during independent repeats. Therefore, the experimental design is very important using transcriptome analysis, where a multitude of results are obtained that must be reproduced and validated. Although batch cultures are very easy to perform, they are problematic with respect to sample reproducibility. The requirement for reproducibility can be met by using continuous culture, where cells are in a so-called steady-state.
The main findings of our study were the catabolic flexibility of E. coli and a switch from the TCA pathway to the PEP-glyoxylate cycle in glucose-limited continuous culture. Moreover, this catabolic flexibility and induction of the PEP-glyoxylate cycle were observed also after genetic selection under such conditions. A downregulation of rpoS transcription and of genes belonging to the s regulon was observed in the evolved population, increasing the importance of cAMP as global regulator during adaptation to carbon/energy limitations.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Conway, P. L. (1995). Microbial ecology of the human large intestine. In Human Colonic Bacteria: Role in Nutrition, Physiology, and Pathology, pp. 1–24. Edited by G. R. Gibson & G. T. MacFarlane. Boca Raton, FL: CRC Press.
Cronan, J. E., Jr & LaPorte, D. (1996). Tricarboxylic acid cycle and glyoxylate bypass. In Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, 2nd edn, pp. 206–216. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.
Death, A. & Ferenci, T. (1993b). The importance of the binding-protein dependent Mgl system to the transport of glucose in Escherichia coli growing on low sugar concentrations. Res Microbiol 144, 529–537.[Medline]
Death, A., Notley, L. & Ferenci, T. (1993a). Derepression of LamB protein faciliates outer membrane permeation of carbohydrates into Escherichia coli under conditions of nutrient stress. J Bacteriol 175, 1475–1483.
Egli, T. (1995). The ecological and physiological significance of the growth of heterotrophic microorganisms with mixtures of substrates. Adv Microb Ecol 14, 305–386.
Farrell, M. J. & Finkel, S. E. (2003). The growth advantage in stationary-phase phenotype conferred by rpoS mutations is dependent on the pH and nutrient environment. J Bacteriol 185, 7044–7052.
Ferenci, T. (1996). Adaptation to life at micromolar nutrient levels: regulation of Escherichia coli glucose transport by endoinduction and cAMP. FEMS Microbiol Rev 18, 301–317.[CrossRef][Medline]
Ferenci, T. (1999a). Regulation by nutrient limitation. Curr Opin Microbiol 2, 208–213.[CrossRef][Medline]
Ferenci, T. (2003). What is driving the acquisition of mutS and rpoS polymorphisms in Escherichia coli? Trends Microbiol 11, 457–461.[CrossRef][Medline]
Fischer, E. & Sauer, U. (2003). A novel metabolic cycle catalyzes glucose oxidation and anaplerosis in hungry Escherichia coli. J Biol Chem 278, 46446–46451.
Garcia-Armisen, T. & Servais, P. (2004). Enumeration of viable E. coli in rivers and wastewaters by fluorescent in situ hybridization. J Microbiol Methods 58, 269–279.[Medline]
Gosset, G., Zhang, Z., Nayyar, S., Cuevas, W. A. & Saier, M. H., Jr (2004). Transcriptome analysis of Crp-dependent catabolite control of gene expression in Escherichia coli. J Bacteriol 186, 3516–3524.
Gourse, R. L., de Boer, H. A. & Nomura, M. (1986). DNA determinants of rRNA synthesis in E. coli: growth rate dependent regulation, feedback inhibition, upstream activation, antitermination. Cell 44, 197–205.[CrossRef][Medline]
Guest, J. R. (1981). Partial replacement of succinate dehydrogenase function by phage- and plasmid-specified fumarate reductase in Escherichia coli. J Gen Microbiol 122, 171–179.[Medline]
Hazen, T. C. & Toranzos, G. A. (1990). Tropical source water. In Drinking Water Microbiology, pp. 32–54. Edited by G. A. McFeters. New York: Springer.
Hengge-Aronis, R. (2002). Signal transduction and regulatory mechanisms involved in control of the S (RpoS) subunit of RNA polymerase. Microbiol Mol Biol Rev 66, 373–395.
Ho, K. K. & Weiner, H. (2005). Isolation and characterization of an aldehyde dehydrogenase encoded by the aldB gene of Escherichia coli. J Bacteriol 187, 1067–1073.
Hua, Q., Yang, C., Baba, T., Mori, H. & Shimizu, K. (2003). Responses of the central metabolism in Escherichia coli to phosphoglucose isomerase and glucose-6-phosphate dehydrogenase knockouts. J Bacteriol 185, 7053–7067.
Hua, Q., Yang, C., Oshima, T., Mori, H. & Shimizu, K. (2004). Analysis of gene expression in Escherichia coli in response to changes of growth-limiting nutrient in chemostat cultures. Appl Environ Microbiol 70, 2354–2366.
Ihssen, J. & Egli, T. (2004). Specific growth rate and not cell density controls the general stress response in Escherichia coli. Microbiology 150, 1637–1648.
Ihssen, J. & Egli, T. (2005). Global physiological analysis of carbon- and energy-limited growing Escherichia coli confirms a high degree of catabolic flexibility and preparedness for mixed substrate utilization. Environ Microbiol 7, 1568–1581.[CrossRef][Medline]
King, T., Ishihama, A., Kori, A. & Ferenci, T. (2004). A regulatory trade-off as a source of strain variation in the species Escherichia coli. J Bacteriol 186, 5614–5620.
Koshland, D. E., Jr, Walsh, K. & LaPorte, D. C. (1985). Sensitivity of metabolic fluxes to covalent control. Curr Top Cell Regul 27, 13–22.[Medline]
Kurland, C. G. & Maaloe, O. (1962). Regulation of ribosomal and transfer RNA synthesis. J Mol Biol 4, 193–210.[Medline]
Lacour, S. & Landini, P. (2004). SigmaS-dependent gene expression at the onset of stationary phase in Escherichia coli: function of sigmaS-dependent genes and identification of their promoter sequences. J Bacteriol 186, 7186–7195.
Lee, S. K., Newman, J. D. & Keasling, J. D. (2005). Catabolite repression of the propionate catabolic genes in Escherichia coli and Salmonella enterica: evidence for involvement of the cyclic AMP receptor protein. J Bacteriol 187, 2793–2800.
Limon, A., Hidalgo, E. & Aguilar, J. (1997). The aldA gene of Escherichia coli is under the control of at least three transcriptional regulators. Microbiology 143, 2085–2095.[Abstract]
Liu, M., Durfee, T., Cabrera, J. E., Zhao, K., Jin, D. J. & Blattner, F. R. (2005). Global transcriptional programs reveal a carbon source foraging strategy by Escherichia coli. J Biol Chem 280, 15921–15927.
Macfarlane, G. T. & Macfarlane, S. (1997). Human colonic microbiota: ecology, physiology and metabolic potential of intestinal bacteria. Scand J Gastroenterol Suppl 222, 3–9.[Medline]
Macfarlane, S. & Macfarlane, G. T. (2004). Bacterial diversity in the human gut. Adv Appl Microbiol 54, 261–289.[CrossRef][Medline]
Macfarlane, G. T., Macfarlane, S. & Gibson, G. R. (1998). Validation of a three-stage compound continuous culture system for investigating the effect of retention time on the ecology and metabolism of bacteria in the human colon. Microb Ecol 35, 180–187.[CrossRef][Medline]
Makinoshima, H., Aizawa, S., Hayashi, H., Miki, T., Nishimura, A. & Ishihama, A. (2003). Growth phase-coupled alterations in cell structure and function of Escherichia coli. J Bacteriol 185, 1338–1345.
Matin, A., Auger, E. A., Blum, P. H. & Schultz, J. E. (1989). Genetic basis of starvation survival in nondifferentiating bacteria. Annu Rev Microbiol 43, 293–316.[CrossRef][Medline]
McFeters, G. A. (1990). Drinking Water Microbiology. New York: Springer.
Miura, A., Krueger, J. H., Itoh, S., de Boer, H. A. & Nomura, M. (1981). Growth-rate-dependent regulation of ribosome synthesis in E. coli: expression of the lacZ and galK genes fused to ribosomal promoters. Cell 25, 773–782.[CrossRef][Medline]
Morita, R. Y. (1993). Bioavailability of energy and the starvation state. In Starvation in Bacteria, pp. 1–23. Edited by S. Kjelleberg. New York: Plenum.
Münster, U. (1993). Concentrations and fluxes of organic carbon substrates in the aquatic environment. Antonie van Leeuwenhoek 63, 243–274.[CrossRef][Medline]
Notley, L. & Ferenci, T. (1995). Differential expression of mal genes under cAMP and endogenous inducer control in nutrient-stressed Escherichia coli. Mol Microbiol 16, 121–129.[Medline]
Notley-McRobb, L. & Ferenci, T. (1999a). Adaptive mgl-regulatory mutations and genetic diversity evolving in glucose-limited Escherichia coli populations. Environ Microbiol 1, 33–43.[CrossRef][Medline]
Notley-McRobb, L. & Ferenci, T. (1999b). The generation of multiple coexisting mal-regulatory mutations through polygenic evolution in glucose-limited populations of Escherichia coli. Environ Microbiol 1, 45–52.[CrossRef][Medline]
Notley-McRobb, L., Death, A. & Ferenci, T. (1997). The relationship between external glucose concentration and cAMP levels inside Escherichia coli: implications for model of phosphotransferase-mediated regulation of adenylate cyclase. Microbiology 143, 1909–1918.[Abstract]
Notley-McRobb, L., King, T. & Ferenci, T. (2002). rpoS mutations and loss of general stress resistance in Escherichia coli populations as a consequence of conflict between competing stress responses. J Bacteriol 184, 806–811.
Nyström, T. (2004). Growth versus maintenance: a trade-off dictated by RNA polymerase availability and sigma factor competition? Mol Microbiol 54, 855–862.[CrossRef][Medline]
Patten, C. L., Kirchhof, M. G., Schertzberg, M. R., Morton, R. A. & Schellhorn, H. E. (2004). Microarray analysis of RpoS-mediated gene expression in Escherichia coli K-12. Mol Genet Genomics 272, 580–591.
