A mutational block in the early stages of the glycolytic pathway facilitates the degradation of the ptsG mRNA encoding the major glucose transporter IICBGlc in Escherichia coli. The degradation is RNase E dependent and is correlated with the accumulation of either glucose-6-P or fructose-6-P (Kimata et al., 2001, EMBO J 20: 3587–3595; Morita et al., 2003, J Biol Chem 278: 15608–15614). In this paper, we investigate additional physiological effects resulting from the accumulation of glucose-6-P caused by a mutation in pgi encoding phosphoglucose isomerase, focusing on changes in gene expression. The addition of glucose to the pgi strain caused significant growth inhibition, in particular in the mlc background. Cell growth then gradually resumed as the level of IICBGlc decreased. We found that the transcription of the cps operon, encoding a series of proteins responsible for the synthesis of colanic acid, was markedly but transiently induced under this metabolic stress. Both genetic and biochemical studies revealed that the metabolic stress induces cps transcription by activating the RcsC/YojN/RcsB signal transduction system. Overexpression of glucose-6-P dehydrogenase eliminated both growth inhibition and cps induction by reducing the glucose-6-P level. Mutations in genes responsible for the synthesis of glucose-1-P and/or dTDP-glucose eliminated the activation of the Rcs system by the metabolic stress. Taken together, we conclude that an increased synthesis of dTDP-glucose activates the Rcs phosphorelay system, presumably by affecting the synthesis of oligosaccharides for enterobacterial common antigen and O-antigen.
Bacterial cells take up a number of sugars either by the phosphoenolpyruvate (PEP):sugar phosphotransferase system (PTS) or by non-PTS transport systems. The incorporated sugars are metabolized primarily by the Embden–Meyerhof glycolytic pathway and by the pentose–phosphate pathway to produce numerous intermediary metabolites and energy in cells (Fraenkel, 1996). It is well established that the expression of proteins responsible for the uptake and metabolism of sugars is regulated at the level of transcription by either specific and/or global transcription factors. The activity of transcription factors is modulated by the availability of individual and/or other sugars. In the case of glucose, the expression of the ptsG gene encoding the major glucose transporter, IICBGlc, is regulated by two global control systems at the level of transcription initiation (Kimata et al., 1997; 1998; Plumbridge, 1998; 2002). First, it is under positive control by CRP-cAMP, hence ptsG expression is absolutely dependent on this complex. Secondly, the transcription of the ptsG gene is negatively regulated by a global repressor, Mlc. We and others have demonstrated that external glucose induces ptsG transcription by modulating the Mlc-mediated regulatory pathway (Kimata et al., 1998; Plumbridge, 1998). The way in which glucose regulates Mlc activity is unique because dephosphorylated IICBGlc is involved in the sequestration of Mlc at the membrane (Lee et al., 2000; Tanaka et al., 2000; Nam et al., 2001).
In addition to the transcriptional control, a second level of control of ptsG expression exists at the level of mRNA degradation in response to the glycolytic flux. Namely, the stability of ptsG mRNA is dramatically reduced by metabolic blocks at early stages of glycolysis (Kimata et al., 2001). The mechanism of stimulation of ptsG mRNA degradation is largely unknown except that this phenomenon is dependent on RNase E and associated with elevated levels of hexose phosphates such as glucose-6-P and/or fructose-6-P (Kimata et al., 2001; Morita et al., 2003). However, the physiological relevance of this feedback regulation of ptsG expression is clear because the downregulation of ptsG expression through mRNA degradation certainly prevents glucose uptake to avoid too much accumulation of hexose phosphates that are potentially toxic to cells (Lee and Cerami, 1987; Singer et al., 1991; Levi and Werman, 2001).
It is interesting to investigate how the metabolic stress that leads to the accumulation of hexose phosphates affects cell physiology including gene expression. In this study, we report a new finding regarding the physiological consequences produced by metabolic perturbation in the glycolytic pathway, namely induction of the cps operon encoding a series of proteins responsible for the synthesis of colanic acid (Stevenson et al., 1996). We demonstrate that mutational blocks at early stages in the glycolytic pathway cause significant growth inhibition in the presence of glucose when cells carry an additional mutation in the mlc gene. Under this metabolic stress, the transcription of the cps operon was markedly induced. Both genetic and biochemical studies revealed that the metabolic stress activates the Rcs signal transduction system. Mutations in genes responsible for the synthesis of glucose-1-P and/or dTDP-glucose eliminated the activation of the Rcs system by the metabolic stress. We conclude that perturbation in the synthesis of polysaccharides via the dTDP-glucose pathway somehow activates the RcsC/YojN/RcsB phosphorelay system. The present study has illuminated an additional aspect regarding the physiological effects of a metabolic block in the glycolytic pathway.
Physiological effects of the metabolic block produced by a pgi mutation
Phosphoglucose isomerase encoded by pgi catalyses the interconversion of glucose-6-P and fructose-6-P. It is located at the first junction of different pathways for glucose metabolism (Fig. 1). We are interested in investigating how glucose uptake affects cell physiology when the glycolytic pathway is blocked by the pgi mutation. The major glucose transporter IICBGlc encoded by ptsG is essentially undetectable in pgi cells in the absence of glucose because ptsG transcription is repressed by Mlc (Fig. 2B, lane 1). The addition of glucose causes two consequences regarding the expression of IICBGlc in pgi cells. First, it induces the initiation of ptsG transcription by relieving the repression by Mlc. Secondly, it facilitates the RNase E-dependent degradation of ptsG mRNA. The combined outcome of these two effects is a low level of expression of IICBGlc upon addition of glucose (Fig. 2B, lanes 2–4). Thus, the downregulation of IICBGlc at the mRNA degradation step in response to the metabolic block plays an important role in modulating the extent of glucose uptake to an appropriate level to avoid too much accumulation of potentially toxic metabolic intermediates. As a result, the growth of pgi cells was only slightly affected by the presence of external glucose (Fig. 2A, left). When the mlc gene is inactivated, the ptsG gene is highly transcribed even in the absence of glucose. Thus, a high level of IICBGlc was detected in the pgi mlc cells growing in TB medium without glucose (Fig. 2B, lane 5). When these cells find glucose in the medium, an efficient glucose uptake should occur as a result of the presence of a high level of IICBGlc. This would lead to a higher accumulation of glucose-6-P and other sugar intermediates that may affect cell physiology. In fact, the addition of glucose caused a marked but transient inhibition of cell growth in pgi mlc cells (Fig. 2A, right). The inhibition started within 1 h after the addition of glucose and lasted for several hours. Then, cell growth was gradually resumed. Northern blot analysis showed that the ptsG mRNA is destabilized both in pgi and pgi mlc cells shortly after the addition of glucose (Fig. 2C) as shown previously (Kimata et al., 2001; Morita et al., 2003), whereas Western blot analysis indicated that the level of IICBGlc gradually decreases after the addition of glucose (Fig. 2B, lanes 6–10). Thus, it takes several hours for the pre-existing IICBGlc to be sufficiently reduced through dilution and/or degradation in pgi mlc cells even though there is very little de novo synthesis of IICBGlc after the addition of glucose because of the degradation of ptsG mRNA. We conclude that growth inhibition by glucose is tightly associated with the extent of glucose uptake, which is determined by the IICBGlc level. When IICBGlc was overexpressed in the pgi cells by introducing a plasmid carrying the ptsG gene, growth inhibition by glucose was further enhanced (data not shown). Thus, it is apparent that the accumulation of glucose-6-P is certainly responsible for the inhibition of cell growth. A similar but less significant growth inhibition by glucose was observed when pfkA was disrupted in the mlc background. The pfkA gene encodes phosphofructokinase A, which catalyses the phosphorylation of fructose-6-P by ATP to yield fructose-1,6-biphosphate. Disruption of other glycolytic genes including tpiA (triose-phosphate isomerase), pgk (phosphoglycerate kinase), gpm (phosphoglycerate mutase) and pyk (pyruvate kinase) did not cause growth inhibition by glucose in an mlc background. The effect of fda (fructose-1,6-P2 aldolase), gap (glyceraldehyde-3-phosphate dehydrogenase) and eno (enolase) mutations could not be tested because we failed to disrupt these genes.
Identification of a protein induced by the metabolic stress
To gain insight into the cellular events caused by glucose uptake in pgi mlc cells, total cellular proteins were analysed by SDS-PAGE at various intervals after the addition of glucose. The addition of glucose to pgi mlc cells significantly affected the profile of total proteins (Fig. 3A, lanes 2–4). The levels of several proteins increased whereas those of others decreased. Among the proteins induced by glucose, we focused on a 40 kDa band (shown by an arrow in Fig. 3A) because its intensity was increased further in the presence of glucose when IICBGlc was overexpressed in the pgi cells (Fig. 3A, lanes 6 and 7), whereas it was not detectable when glucose was added to mlc or pgi cells (data not shown). Thus, the 40 kDa band appears to represent a protein that is specifically induced in response to the metabolic block at the early stage of glycolysis. This band was excised, digested with trypsin and analysed by matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry. The resulting set of peptide masses was used to search a database for potential matches (Fig. 3B). The protein was identified as GDP mannose dehydratase (GMD) encoded by the gmd that is a part the colanic acid synthetic operon (cps or wca) (Stevenson et al., 1996).
The cps mRNA is induced by the metabolic stress
To understand the mechanism underlying the induction of GMD, cellular RNAs were analysed by Northern blotting using the gmd DNA probe (Fig. 4A). It is known that the cps operon is little expressed under normal laboratory conditions (Gottesman and Stout, 1991; Gottesman, 1995). In fact, the cps mRNA was undetectable before the addition of glucose in pgi mlc cells (Fig. 4B, lane 1). A dispersed RNA signal, which reflects a complicated transcription and/or processing of the large cps mRNA (Stevenson et al., 1996), became detectable gradually after the addition of glucose (Fig. 4B, lanes 2–6). The maximum induction of cps mRNA was observed 2 h after the addition of glucose. The level of cps mRNA started to decline with the recovery of cell growth and was essentially undetectable 4 h after the addition of glucose. The induction of cps mRNA by glucose was also confirmed by another DNA probe corresponding to the wza gene (Fig. 4B, lanes 7–9). When the pgi mlc cells were grown without glucose, no induction of cps mRNA was observed (data not shown). We have already shown that the inhibition of cell growth is associated with the level of IICBGlc. Thus, we conclude that cps expression is induced in response to blocking the glycolytic pathway, and that the magnitude of induction is dependent upon the amount of glucose uptake. It is highly possible that the accumulation of glucose-6-P and/or other metabolites caused by glucose uptake is responsible for the increased expression of cps mRNA.
Activation of the Rcs system is responsible for the induction of the cps operon
The dramatic increase in cps mRNA after the addition of glucose in pgi mlc cells suggests that cps expression is activated by metabolic stress at the transcriptional level. Transcription of the cps operon is regulated by the modified His–Asp phosphorelay system (two-component system), RcsC/YojN/RcsB (Gottesman and Stout, 1991; Gottesman, 1995; Takeda et al., 2001). In this system, RcsC is the transmembrane sensor kinase/phosphatase, and RcsB is the response regulator containing a helix–turn–helix DNA-binding motif (Gottesman and Stout, 1991; Gottesman, 1995). Under certain conditions, stimulation of RcsC is believed to lead to the phosphorylation of RcsB. A histidine-containing phosphotransmitter, YojN, is implicated as an intermediate in this phosphotransfer (Chen et al., 2001; Takeda et al., 2001). The phosphorylated RcsB then activates transcription from the cps operon along with an auxiliary unstable protein RcsA. In order to examine whether the induction of cps mRNA by metabolic stress in pgi mlc cells is dependent on the Rcs system, we inactivated each of the rcsA, rcsB, rcsC or yojN genes in the pgi mlc background and analysed the effects of glucose addition on cell growth and on cps induction in these triple mutant strains. First, we observed that growth inhibition by glucose in pgi mlc cells was essentially unaffected when the rcsA, rcsB, rcsC or yojN mutation was introduced (data not shown). On the other hand, the disruption of rcsB, rcsC or yojN completely abolished the induction of cps mRNA (Fig. 5, lanes 3–5). This implies that the induction of cps mRNA by glucose in pgi mlc cells occurs at the transcriptional level in an Rcs system-dependent manner. The null mutation in rcsA dramatically reduced, but did not abolish, cps induction by glucose (Fig. 5, lane 2). This suggests that activation of the RcsC/YojN/RcsB system is primarily responsible for the induction of the cps operon by metabolic stress.
It is well established that overproduction of RcsA induces cps mRNA by acting together with RcsB in an RcsC-independent manner (Gottesman and Stout, 1991; Gottesman, 1995). In fact, a strong constitutive induction of cps mRNA was observed when the rcsA plasmid was introduced in mlc cells (Fig. 5B). The magnitude of the induction of cps mRNA by glucose in pgi mlc cells is comparable with that in mlc cells harbouring the rcsA plasmid. Thus, the signal produced by metabolic stress in pgi mlc cells is very effective in activating the Rcs system, as in the case of RcsA overproduction. However, there are clear differences between the two conditions. The addition of glucose to pgi mlc cells causes transient induction of cps mRNA along with growth inhibition, whereas overproduction of RcsA permanently induces cps mRNA without growth inhibition (data not shown).
Activation pathway of the Rcs system by metabolic stress is different from that by osmotic shock
Osmotic shock is an environmental signal that can induce cps transcription transiently in an Rcs-dependent manner (Sledjeski and Gottesman, 1996). To investigate a possible link between osmotic shock and metabolic stress by glucose in pgi mlc cells, we examined the induction of cps mRNA by osmotic shock by Northern blotting. It is known that the induction of cps by osmotic shock is more effective at 30°C than at 37°C, presumably because RcsA is more stable at lower temperature (Sledjeski and Gottesman, 1996). In fact, the cps mRNA was only slightly induced by the addition of sucrose to a final concentration of 15% at 37°C (Fig. 6, lanes 6–8), whereas sucrose addition caused a marked induction of cps mRNA at 30°C in pgi mlc cells (Fig. 6, lanes 1–5). It should be noted, however, that the cps mRNA level peaked 30 min after the shock and gradually decreased to the unshocked state. On the other hand, the cps mRNA level peaked about 3 h after the addition of glucose in pgi mlc cells at 30°C (data not shown). Thus, the kinetics of cps mRNA induction by glucose and the temperature effect in pgi mlc cells are clearly different from those in the osmotic shock. These results suggest that the way in which glucose activates the Rcs system in pgi mlc cells is different from that in the case of osmotic shock.
Overexpression of glucose-6-P dehydrogenase relieves the metabolic stress caused by glucose in pgi mlc cells
We found previously that overproduction of glucose-6-P dehydrogenase (Zwf) prevents the accumulation of glucose-6-P, resulting in stabilization of ptsG mRNA in pgi cells (Morita et al., 2003). It is interesting to investigate whether the overproduction of Zwf could also relieve the observed physiological effects of excess uptake of glucose in pgi mlc cells. When Zwf was overproduced by introducing pTM6 carrying the zwf gene into the pgi mlc cells, the accumulation of glucose-6-P was markedly prevented (Fig. 7A). Under this condition, both growth inhibition (Fig. 7B) and induction of the cps operon (Fig. 7C) by glucose were completely eliminated. Thus, the increased flow of the pentose pathway by overproduction of Zwf is sufficient to eliminate the metabolic stress caused by glucose in pgi mlc cells. These results clearly indicate that the accumulation of glucose-6-P triggers events that lead to growth inhibition and the activation of the RcsC/YojN/RcsB system. Interestingly, the pgi zwf mlc triple mutant could not grow in the presence of glucose. The hypersensitivity of the triple mutant to glucose can be explained by an extreme accumulation of glucose-6-P resulting from blockage of both the Embden–Meyerhof and the pentose–phosphate pathways.
A mutational block of the synthesis of glucose-1-P eliminates the activation of RcsC/B system
The fact that cps operon induction is associated with glucose-6-P accumulation in the mutant strains raises the possibility that glucose-6-P itself may act as a signal for cps operon induction. However, this is less likely because the accumulation occurs within several minutes after the addition of glucose (Morita et al., 2003), whereas cps operon induction starts 30 min after glucose addition. This suggests that glucose-6-P itself is not the direct signal, but may lead to the production of a final signal by further metabolism. In addition to the conversion to fructose-6-P by Pgi and to 6-P-gluconolactone by Zwf, glucose-6-P can be converted to glucose-1-P by Pgm (Fig. 1). The above-mentioned results suggest that the overflow in this third pathway of glucose-6-P metabolism could be required for the production of the signal for the activation of the Rcs system because the increased level of glucose-6-P in pgi mlc cells is expected to lead to increased synthesis of glucose-1-P through Pgm. Then, we examined the effect of the inactivation of pgm on cell physiology in the pgi mlc background. Interestingly, the pgm mutation dramatically reduced induction of the cps operon caused by glucose uptake (Fig. 8A, lane 2), whereas it did not affect growth inhibition by glucose (Fig. 8C). The reduction in cps operon expression is not the result of a general reduction in transcription by the pgm mutation because crp expression increased (Fig. 8B, lane 2). This implies that the conversion of glucose-6-P to glucose-1-P is required for activation of the Rcs system, whereas the accumulation of glucose-6-P itself is sufficient to cause inhibition of cell growth. Thus, either the accumulation of glucose-1-phosphate itself or the increased flow in pathways downstream from glucose-1-P may be responsible for activation of the RcsC/YojN/RcsB system.
A mutational block in the synthesis of dTDP-glucose eliminates the activation of the RcsCB system
To specify the metabolic signal leading to the activation of the Rcs system, we examined the effects of mutations in genes responsible for metabolic pathways downstream of glucose-1-P (Fig. 1). Glucose-1-P is a common precursor for the synthesis of nucleoside diphosphate sugars as activated glycosyl donors, ADP-glucose, UDP-glucose and dTDP-glucose (Schulman and Kennedy, 1977). These reactions are catalysed by glucose-1-phosphate adenylyltransferase (glgC), UTP-glucose-1-phosphate uridylyltransferase (galU) and glucose-1-phosphate thymidylyltransferase (rffH/rfbA) respectively. The known function of ADP-glucose in Escherichia coli is the biosynthesis of glycogen, while UDP-glucose is of central importance in the synthesis of the components of the cell envelope and in both galactose and trehalose metabolism. It is used for the biosynthesis of the core oligosaccharide region of lipopolysaccharide (LPS) (Raetz, 1996), capsular polysaccharide (Rick and Silver, 1996) and membrane-derived oligosaccharides (MDO) (Kennedy, 1996). On the other hand, dTDP-glucose is a precursor of sugars found in the O-antigen and enterobacterial common antigen (ECA) (Rick and Silver, 1996). The finding that the pgm mutation could suppress the activation of the Rcs system under metabolic stress prompted us to examine the effects of mutations in each of the three pathways for the synthesis of nucleoside diphosphate sugars on the activation of Rcs system. For this, the glgC, galU and rffH/rfbA genes were disrupted individually in the pgi mlc background and tested for their effect on the induction of the cps operon by glucose. A strong activation of the cps operon by metabolic stress was observed in the glgC pgi mlc triple mutant, indicating that the overflow in the pathway to the synthesis of ADP-glucose and glycogen is not responsible for the production of the signal for the activation of the Rcs system (Fig. 8A, lane 4). Similarly, the galU mutation did not relieve the activation of the cps operon by metabolic stress in the pgi mlc background, indicating that the metabolic perturbation caused by increased synthesis of UDP-glucose is also not responsible for the activation of the Rcs system (Fig. 8A, lane 3). On the other hand, induction of the cps operon by glucose was dramatically reduced when both rffH and rfbA genes were disrupted together to block the synthesis of dTDP-glucose in the pgi mlc background (Fig. 8A, lane 7). The disruption of rfbA or rffH alone did not affect the activation of the cps operon by metabolic stress in the pgi mlc background because the rfbA and rffH genes are redundant (Fig. 8A, lanes 5 and 6). These mutations did not affect crp expression (Fig. 8B). The results strongly suggest that an overflow in the synthetic pathway of ECA and/or O-antigen is responsible for the production of a signal required for the activation of the Rcs system.
Glycolysis is a central metabolic pathway that is responsible for the production of numerous intermediary metabolites and energy in cells (Fraenkel, 1996). Therefore, any mutational and/or biochemical blocks in the pathway are expected to lead to various physiological consequences. This was typically shown in the mutant strain defective in the pgi gene that is located at the first junction of different pathways for glucose metabolism (Fig. 1). For example, the pgi mutation is known to reduce the growth rate when glucose is provided as a sole carbon source, presumably because of the increased demand for NADPH reoxidation and the limited capacity of the pentose pathway (Canonaco et al., 2001). The pgi mutation is also known to cause a dramatic increase in the mutation frequency in the presence of glucose as a result of the accumulation of hexose phosphates (Lee and Cerami, 1987).
A striking recent observation is that mutations in pgi, pfkA or ts8 fda lead to rapid degradation of the ptsG mRNA in an RNaseE-dependent manner as a result of the accumulation of glucose-6-P and/or fructose-6-phosphate (Morita et al., 2003). The role of this regulation is apparently to prevent too much accumulation of potentially toxic metabolic intermediates. In this paper, we demonstrated that the uptake of glucose causes strong but transient growth retardation in pgi mlc cells in which IICBGlc protein is highly expressed before the cells are exposed to glucose. This growth retardation by glucose uptake was abolished as the concentration of IICBGlc in the cell decreased on account of its degradation and/or dilution. This observation clearly illuminates the importance of the feedback regulation of ptsG expression through mRNA degradation.
A central finding in this work is that transcription of the cps operon responsible for the synthesis of capsular polysaccharide, colanic acid, is strongly but transiently induced during the period of growth retardation by glucose uptake in pgi mlc cells. We also showed that accumulation of glucose-6-P triggers the activation of cps operon transcription in the mutant strains. Colanic acid (M-antigen) is a mucoid extracellular polysaccharide produced in many enterobacteria and is believed to play a role in bacterial survival outside the host, in particular in resistance to desiccation (Gottesman, 1995; Rick and Silver, 1996). It is a high-molecular-weight polymer containing six-residue repeat units composed of glucose, galactose, glucuronic acid and fucose (Rick and Silver, 1996; Stevenson et al., 1996). The active precursors of these sugar constituents are nucleotide sugars. Although the biosynthesis of colanic acid is complex and a large number of enzymes are involved in the process, most of these enzymes are encoded by the cps operon composed of 19 open reading frames (ORFs; Stevenson et al., 1996). Although numerous direct and indirect genetic and environmental factors are known to affect the expression of cps, little is known about metabolic signals that lead to activation of the synthesis of colanic acid (Gottesman and Stout, 1991; Gottesman, 1995).
Expression of the cps operon is controlled by a complex regulatory network (Gottesman and Stout, 1991; Gottesman, 1995). Two pathways that can lead to the activation of cps transcription are known. One involves the activation of the RcsC/YojN/RcsB His–Asp phosphorelay system in which a sensor kinase RcsC somehow senses external and/or internal signals leading to the activation (phosphorylation) of RcsB. The phosphorylated RcsB would then activate cps transcription. An alternative pathway involves an increased level of RcsA protein that acts together with RcsB to activate cps transcription. RcsA is limiting in normal cells because it is unstable as a result of its degradation by Lon protease. When sufficient RcsA accumulates, through either stabilization or overproduction of RcsA, cps transcription can be activated even in the absence of RcsC. Another conclusion from this work is that accumulation of glucose-6-P activates cps transcription transiently through the activation of the RcsC/YojN/RcsB His–Asp phosphorelay system. This conclusion was derived from the finding that the activation of cps transcription by metabolic stress was completely abolished by either rcsB or rcsC mutation, while the rcsA mutation allowed residual activation of cps by metabolic stress. It should be noted, however, that rcsA transcription is also activated by the Rcs system (Ebel and Trempy, 1999). This means that the activation of the RcsC/YojN/RcsB pathway leads to the activation of both the cps operon and the rcsA gene, resulting in an increased amount of RcsA protein. In fact, we observed that both rcsA mRNA and RcsA protein were markedly induced upon addition of glucose in pgi mlc cells (data not shown). It is apparent that this positive autoregulation of RcsA plays an important role for effective induction of the cps operon under stress conditions.
We also demonstrated that an increased flow in the pathway for the synthesis of dTDP-glucose resulting from the accumulation of glucose-6-P is responsible for production of the metabolic signal that leads to the activation of RcsC/YojN/RcsB in response to glucose uptake in pgi mlc cells. How does the increased synthesis of dTDP-glucose lead to activation of the RcsC/YojN/RcsB system? The sensor kinase RcsC has a periplasmic domain along with the inner membrane-spanning and cytoplasmic regions and is believed to sense some changes in the periplasm and/or the cytoplasmic membrane (Gottesman and Stout, 1991; Gottesman, 1995). It should be noted that dTDP-glucose is a common precursor for O-antigen and ECA oligosaccharides that are synthesized in the periplasm and transferred to the outer membrane region (Raetz, 1996; Rick and Silver, 1996). It is reasonable to speculate that the increased synthesis of dTDP-glucose affects the biosynthesis of these membrane-associated oligosaccharides, which could in turn be sensed by RcsC even though E. coli K-12 strains lack O-antigen itself because of a mutation in the last rfb gene encoding a rhamnosyltransferase (Raetz, 1996). It remains to be investigated what change actually occurs in the ECA and/or O-antigen biosynthetic pathways when the synthesis of dTDP-glucose is increased and how RcsC senses its perturbation.
In this regard, it should be noted that genetic and/or environmental conditions that are known to induce the cps operon through the RcsC/YojN/RcsB pathway are apparently linked with alterations in envelope composition and/or integrity. For example, a deletion of rfa genes responsible for the synthesis of LPS core was first shown to induce the cps operon in an Rcs-dependent manner (Parker et al., 1992). Similarly, a mutation in the mdoH gene involved in the biosynthesis of MDOs is known to activate the Rcs system (Ebel et al., 1997). In addition, osmotic shock is believed to induce cps expression by affecting the level of MDOs (Sledjeski and Gottesman, 1996; Ebel et al., 1997). Furthermore, overproduction of the DnaJ-like transmembrane protein DjlA (Kelley and Georgopoulos, 1997) and the addition of a cationic amphipathic compound (Conter et al., 2002) were shown to stimulate the expression of genes regulated by the RcsC/YojN/RcsB system. The present results, in combination with previously described results, strongly suggest that alterations in the synthesis and/or composition of any membrane-related oligosaccharides could lead to the activation of RcsC. Finally, it has been found that the Rcs system can be activated when wild-type cells are grown at a low temperature (20°C) in the presence of glucose and in the presence of a high concentration of Zn2+ (Hagiwara et al., 2003). It is possible that glucose and/or Zn2+ activate the Rcs system by affecting the synthesis and/or composition of membrane-related oligosaccharides at low temperature.
Strains, plasmids and growth conditions
The bacterial strains and plasmids used in this study are listed in Table 1. A series of disruption mutants except ΔrcsC and ΔyojN were constructed by the one-step gene inactivation protocol that is based on the high efficiency of the phage λ Red recombinase (Datsenko and Wanner, 2000). The FRT-flanked resistance gene was eliminated using the FLP expression plasmid pCP20 in some cases (Datsenko and Wanner, 2000). The ΔrcsC and ΔyojN alleles were transferred from ST261 (Chen et al., 2001) and SRC122 (Chen et al., 2001) to other backgrounds by P1 transduction. The 3.2 kb SalI–SalI DNA fragment containing the ptsG gene derived from pIT499 (Takahashi et al., 1998) was cloned into the SalI site of pMW119 (Nippon Gene) to construct pMWG21. Cells were grown at 37°C (unless specified) in TB medium (Miller, 1972) supplemented with kanamycin (15 µg ml−1), tetracycline (15 µg ml−1), chloramphenicol (15 µg ml−1) or ampicillin (50 µg ml−1) when needed. Bacterial growth was monitored by determining the optical density at 600 nm.
Table 1. Bacterial strains and plasmids used in this study.
Bacterial cells grown in TB medium containing appropriate antibiotic(s) were harvested under the indicated conditions and suspended in 100 µl of SDS-PAGE loading buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5%β-mercaptoethanol, 0.1% bromophenol blue). The sample was heated at 100°C for 5 min. The indicated amounts of total cellular proteins were subjected to 12% acrylamide−0.1% SDS gel electrophoresis and transferred to Immobilon membrane (Millipore). The polypeptides detected by the antibodies were visualized with the ECL system (Pharmacia). The anti-IIBGlc polyclonal antibodies have been described previously (Tanaka et al., 2000).
Northern blot analysis
Total cellular RNAs were isolated from exponentially growing cells as described previously (Aiba et al., 1981). The RNAs (10 µg) were resolved by 1.2% agarose gel electrophoresis in the presence of formaldehyde and blotted onto Hybond-N+ membrane (Amersham Bioscience) as described previously (Sambrook et al., 1989). The mRNAs were visualized using digoxygenin (DIG) reagents and kits for non-radioactive nucleic acid labelling and detection system (Roche Molecular Biochemicals) according to the procedure specified by the manufacturer. The following DIG-labelled DNA probes were prepared by polymerase chain reaction (PCR) using DIG-dUTP (Roche): 305 bp fragment corresponding to the 5′ region of ptsG; 824 bp fragment corresponding to the gmd; 340 bp fragment corresponding to the wza; and 621 bp fragment corresponding to the crp.
The protein bands stained with Coomassie brilliant blue were cut out from the gel, and a small piece of the 40 kDa band was treated with 125 ng of trypsin (Promega) in 10 µl of 20 mM ammonium bicarbonate for 12 h at 37°C. The dige-sted peptides were eluted with 300 µl of 50% acetonitrile, 5% formic acid and concentrated to 20 µl. Then, the sample was desalted with a zip-tip reverse-phase column, mixed with 1%α-CHCA (α-cyano-4-hydroxycinnamic acid) in 70% acetonitrile and subjected to MALDI/TOF mass spectrometry.
Determination of intracellular glucose-6-P
One millilitre of the culture was centrifuged at 10 000 g for 2 min at room temperature. The pellet was suspended in 100 µl of H2O, and 50 µl of 5 M HClO4 was added immediately and chilled on ice. After adding 100 µl of 2.5 M K2CO3, the mixture was centrifuged at 14 000 g for 10 min at 4°C. The supernatant was used for the glucose-6-P assay. The assay was performed according to the method of Hogema et al. (1999). The intracellular concentration of glucose-6-P was calculated on the assumption that an OD600 of 1.4 corresponds to 109 cells ml−1 (Miller, 1972) and the volume of a cell is 2 × 10−12 ml (Joseph et al., 1982).
We thank Dr T. Mizuno (Nagoya University) for providing the strains and plasmids and for discussion. We also thank Dr J. Plumbridge (IBPC) for discussion and for careful reading of the manuscript. We thank Dr N. Kido (Nagoya University) for discussion. This work was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by the Ajinomoto Co., Inc.