|Home | About | Journals | Submit | Contact Us | Français|
The two-component signal transduction system (TCS) BarA/UvrY activates transcription of CsrB and CsrC noncoding RNAs, which act by sequestering the RNA-binding global regulatory protein CsrA. Here, we show that the metabolic end products formate and acetate provide a physiological stimulus for this TCS and thus link posttranscriptional regulation by the Csr system to the metabolic state of the cell.
The BarA/UvrY two-component signal transduction system (TCS) of Escherichia coli consists of the histidine sensor kinase BarA and its cognate response regulator UvrY (19). BarA (bacterial adaptive response) is a member of the subclass of tripartite sensor kinases (10, 17), whereas UvrY is a typical response regulator of the FixJ family (16). A key role of this TCS and its homologues in other Gram-negative bacteria, such as BarA/SirA of Salmonella, ExpS/ExpA of Erwinia, VarS/VarA of Vibrio, and GacS/GacA of Pseudomonas species, is to activate expression of noncoding RNAs, e.g., CsrB and CsrC in E. coli (13, 20, 29). These small regulatory RNAs possess repeated sequence elements that allow them to interact with multiple copies of the RNA binding protein CsrA and thus prevent its interaction with mRNA targets (1). Because CsrA somehow activates csrB and csrC expression via the BarA-UvrY TCS and thereby feedback inhibits its own activity in the cell, this regulatory circuitry appears to serve as a homeostatic device for fine-tuning CsrA activity (23, 24). In this model, BarA-UvrY is envisioned to provide a set point adjustment mechanism. Finally, deletion of the barA or uvrY gene affects CsrA activity, which regulates central and secondary carbon metabolism, biofilm development, motility, peptide uptake, hfq expression, and virulence gene expression in E. coli and other species (1-4, 6, 11, 18, 21, 26-28).
Despite the important roles played by the BarA/UvrY system, the physiological signals that regulate the activity of the BarA sensor kinase and thereby control the activity of the UvrY response regulator have not been identified for any of the orthologous TCSs. It has been suggested that BarA may respond to products or conditions of the host organism during an infection (31), as many of the target genes are involved in pathogenesis. However, this system is active in the absence of any host organism (18, 24). Moreover, a solvent-extractable extracellular fraction, capable of inducing a moderate activation of GacS/GacA-regulated targets, was isolated from stationary-phase cultures of Pseudomonas fluorescens and suggested to contain a signal for GacS (BarA) (9). In addition, tricarboxylic acid (TCA) cycle function was recently reported to somehow influence GacS/GacA signaling (25). Here, we show that formate and acetate, but also other short-chain fatty acids, provide a physiological stimulus for the BarA sensor kinase.
Previously, we demonstrated that a pH lower than 5.5 provides an environment that does not allow activation of the BarA/UvrY signaling pathway (15). However, in an exploratory experiment we observed that, despite the low pH, the presence of glucose in the growth medium leads to the activation of the BarA/UvrY TCS, as judged by the csrB-lacZ reporter expression (Fig. (Fig.1A).1A). Curiously, activation of the system was observed approximately 2 h after addition of glucose (Fig. (Fig.1A),1A), suggesting that glucose per se does not act as a direct stimulus. Therefore, we hypothesized that an end product of the glucose metabolism might be responsible for the activation of the BarA and/or UvrY protein. To test this hypothesis, the effects of physiological concentrations of succinate, acetate, ethanol, lactate, and formate on the activity of BarA/UvrY were examined. We observed that addition of succinate, ethanol, or lactate to the growth medium did not affect csrB-lacZ expression, whereas addition of formate and acetate resulted in an immediate activation of reporter expression (Fig. (Fig.1B).1B). It has to be noted that addition of the weak acids formate and acetate to the growth medium led to a drastic decrease in growth rate of the cells (Fig. (Fig.1B).1B). To find out whether the observed increase in reporter expression was due to the decrease in growth rate, benzoic acid was used to mimic the weak-acid-dependent growth inhibition. Although a strong growth rate inhibition was achieved by addition of benzoate to the growth medium, only a minor but significant increase of reporter expression was observed (Fig. (Fig.1C).1C). Thus, it appears that formate and acetate provide a stimulatory signal for the BarA/UvrY TCS.
We wondered whether the above-mentioned carboxylic acids act directly on BarA and/or UvrY. To this end, the csrB-lacZ reporter was transduced into a ΔbarA mutant strain and a ΔuvrY mutant strain (24) by P1vir transduction, and their levels of formate- and acetate-dependent reporter expression were monitored. Formate was not able to activate csrB-lacZ expression in either the ΔbarA or the ΔuvrY mutant strain (Fig. 2A and B), indicating that formate acts through the BarA sensor kinase. On the other hand, acetate was able to activate reporter expression in the ΔbarA mutant strain but not in the ΔuvrY strain (Fig. 2A and B), suggesting that acetate can act through UvrY. This is in agreement with a previous study (12) reporting that acetate activates SirA (UvrY) in Salmonella enterica serovar Typhimurium via the formation of acetyl phosphate (acetyl-P), a high-energy compound that can be used by SirA and many other response regulators (see reference 30 and references therein) to autophosphorylate. Indeed, inactivation of the acetyl-P synthetic pathways, by deletion of the acetate kinase (ackA) and phosphotransacetylase (pta) genes that reversibly convert acetate and acetyl coenzyme A (acetyl-CoA) to acetyl-P (30), in the ΔbarA mutant strain abolished the acetate-dependent activation of reporter expression (Fig. (Fig.2C).2C). Thus, acetyl-P-dependent phosphorylation of UvrY appears to be, at least in part, responsible for the acetate effect on csrB-lacZ expression.
Nonetheless, the kinetics of acetate-dependent reporter expression were significantly slower in the ΔbarA mutant strain than in the wild-type strain (Fig. (Fig.2B).2B). We therefore considered that acetate might also act through BarA. This was tested by examining the effect of acetate on the expression of csrB-lacZ in the ackA pta mutant strain. In this strain, acetate addition to the growth medium resulted in the immediate activation of reporter expression (Fig. (Fig.2D),2D), although with a loss of amplitude. Thus, acetate, like formate, can also act through BarA.
Considering that csrB-lacZ expression is activated in cells grown at pH 7.0 or at pH 5.0 in the presence of glucose, but not at pH 5.0 in the absence of glucose (Fig. (Fig.3A),3A), and assuming that acetate may act as a direct stimulus for the BarA/UvrY TCS, we reasoned that acetate and/or formate should accumulate only under the former conditions. To explore this possibility, we determined the extracellular concentrations of formate and acetate in aerobic cultures grown at pH 7.0 and at pH 5.0 with or without glucose. Formate was not detected under these growth conditions (data not shown), consistent with its preferential accumulation under anaerobic conditions (22). On the other hand, acetate accumulated in the cultures that were grown at pH 7.0 or pH 5.0 with glucose, reaching a maximum of ~6 mM or 9 mM, respectively, but did not accumulate at pH 5.0 in the absence of glucose (Fig. (Fig.3B).3B). To distinguish between the possible acetyl-P-dependent UvrY activation and acetate-dependent BarA activation, we examined the effect of glucose on csrB-lacZ expression in the acetyl-P-defective ackA pta mutant strain. In this case, activation of reporter expression still occurred (Fig. (Fig.3A).3A). Also, the concentration of extracellular acetate was found to increase with time, reaching a maximum of ~7 mM (Fig. (Fig.3B).3B). Because acetate synthesis relies on the ackA pta gene products, which reversibly convert acetyl-CoA to acetate, but also on pyruvate oxidase (poxB), which converts pyruvate to acetate (5), we generated an ackA pta poxB triple mutant strain and tested the effect of glucose on reporter expression and acetate production. In this case, activation of reporter expression failed to occur (Fig. (Fig.3A)3A) and there was no significant increase in the extracellular concentration of acetate (Fig. (Fig.3B).3B). Thus, the onset of reporter expression was strictly correlated with the increase in the extracellular concentration of acetate.
To find out whether the carboxylate group is essential for BarA activation, we tested the effects of various molecules with structures similar to that of formate or acetate on the activation of csrB-lacZ expression. Esterified derivates, such as methyl formate, methyl acetate, and ethyl acetate, failed to activate reporter expression (Fig. (Fig.4A).4A). On the other hand, phenylacetic acid, a compound containing an aromatic ring at the aliphatic tail of acetate, activated reporter expression. The lower degree of activation achieved by phenylacetate than by acetate and formate might be due to effects of the bulky phenyl ring on binding to BarA, although this remains to be determined.
The carboxylate group is shared by formate, acetate, and other short-chain fatty acids, including propionate, butyrate, valerate, and caproate, which possess unbranched aliphatic tails of various lengths (3 to 6 carbons). Therefore, we asked whether these short-chain fatty acids were also able to activate the BarA/UvrY TCS. Indeed, addition of any of these compounds to the growth medium resulted in the immediate activation of the BarA/UvrY signaling cascade (Fig. (Fig.4B).4B). Moreover, the effect of these compounds on the expression of csrB-lacZ was eliminated in a ΔbarA mutant strain (Fig. (Fig.4B),4B), implying that they all act through the BarA sensor kinase. It is noteworthy that the level of reporter expression was inversely proportional with the length of the aliphatic tail.
Our findings demonstrate that aliphatic carboxylic acids, such as formate, acetate, propionate, and others, provide a physiological stimulus for the BarA sensor kinase of the BarA-UvrY TCS of Escherichia coli. This activity did not require acetyl-P formation, although acetyl-P may activate signaling of this TCS through the response regulator UvrY. The rapid response to the stimuli and the fact that these compounds are not readily converted to any common intermediary metabolite suggest that their related chemical structures permit them to signal directly to BarA, although this conclusion requires biochemical studies with the purified protein. Whether BarA homologs of other species also signal via carboxylic acids is not yet clear, although this scenario is compatible with the finding that the status of the TCA cycle somehow influences GacS signaling in pseudomonads (25).
As mentioned earlier, CsrA indirectly activates transcription of its RNA antagonists CsrB and CsrC via the BarA-UvrY TCS. Furthermore, CsrA activates glycolysis through the Embden-Meyerhoff-Parnas pathway, a major source of acetate (21). Thus, our present findings appear to bridge the gap in our understanding of this circuitry and imply that CsrA alters carbon flux into the physiological activator of the BarA sensor kinase (23, 24, 29).
Acetate is a predominant metabolite of E. coli and the most abundant fatty acid in the intestinal environment of this bacterium, where levels can reach severalfold higher than those that were required for the BarA signaling reaction (12). Acetate tends to accumulate in growth media near the transition to stationary phase, as the cells expend glucose or other acetogenic substrates, and is thereafter assimilated via the acetyl-CoA synthetase reaction (30). This shift from acetate dissimilation to assimilation is sometimes referred to as the acetate switch. Its effect on acetyl-P levels and thus autophosphorylation of a variety of bacterial TCSs has been proposed to influence numerous cellular functions (30). Our present findings imply an additional broad signaling role for acetate, as BarA/UvrY and the Csr system have far-reaching regulatory effects on bacterial physiology, metabolism, and virulence.
We thank Claudia Rodriguez for technical assistance, F. Bolivar for the poxB::Cmr mutant, and A. Gómez Puyou for helpful discussions.
This work was supported by grants 37342-N from the Consejo Nacional de Ciencia y Tecnología (CONACyT); IN221106/17 from DGAPA-PAPIIT, UNAM; CRP/MEX08-02 from the International Centre for Genetic Engineering and Biotechnology (ICGEB); R01 GM059969 from the National Institutes of Health; and FLA-MCS-004949 from the University of Florida CRIS project.
Published ahead of print on 29 January 2010.