|Home | About | Journals | Submit | Contact Us | Français|
The phosphotransferase system (PTS), encompassing EI, HPr, and assorted EII proteins, uses phosphoenolpyruvate to import and phosphorylate sugars. A paralog of EIIA of the sugar PTS system known as ptsN has been purported to regulate organic nitrogen source utilization in Escherichia coli K-12. Its known biochemical function, however, relates to potassium homeostasis. The evidence for regulation of organic nitrogen source utilization by ptsN is based primarily on the defective growth of ΔptsN mutants on amino acid nitrogen sources and other nutrient combinations. These observations were made with E. coli strains MG1655 and W3110, which carry a nonfunctional version of ilvG. There are three isozymes that effectively catalyze the first committed step of branched-chain amino acid biosynthesis, but ilvG is unique for doing so effectively across a range of potassium concentrations. Here we show that all of the nutrient utilization phenotypes attributed to ptsN are manifested selectively in strains lacking functional ilvG. We conclude that the ptsN gene product does not regulate organic nitrogen source utilization as previously proposed.
The phosphotransferase system (PTS) of Escherichia coli catalyzes the transport and concomitant phosphorylation of numerous carbohydrates via a phosphorelay utilizing phosphoenolpyruvate and occupies a central role in cellular regulation. The enzymes EI and HPr are common to all PTS sugars, while enzyme II (EII) varieties are substrate specific and consist of three functional units—EIIA, EIIB, and EIIC. Over 15 years ago, two genes paralogous to HPr and fructose-specific EIIA were identified within the rpoN operon of E. coli (13). The genes are cotranscribed with the nitrogen-responsive sigma factor gene rpoN (15), but their null and phenocopy mutants exhibited nitrogen source-dependent growth defects independent of rpoN (13). For these reasons, the two genes were classified as the “nitrogen PTS” (Ntr PTS) and dubbed ptsN (EIIANtr) and ptsO (NPr).
To date, more than a dozen studies have focused on the Ntr PTS in E. coli, using combinations of genetics, biochemistry, and structural biology to interrogate its role in cellular homeostasis, and from this, two independent groups have produced strong evidence that EIIANtr modulates K+ uptake (12). It was first shown that EIIANtr allosterically inhibits low-affinity potassium transport by direct interactions with TrkA (9, 10). It was later demonstrated that EIIANtr activates the transcription of the high-affinity potassium transport system KdpFABC through the KdpDE response regulator (11). However, attempts to reconcile these biochemical functions of ptsN with a regulatory role in nitrogen utilization have failed to produce a compelling mechanistic link.
Studies seeking to resolve a connection between ptsN-mediated potassium uptake and nitrogen metabolism zeroed in on ilvBN, the genes for acetohydroxyacid synthase I (AHASI), which catalyzes the first committed step of branched-chain amino acid biosynthesis (Fig. 1). Both AHASI activity and the transcription of ilvBN appear to be inhibited by potassium (10). MG1655 ptsN knockout strains exhibit reduced growth in minimal medium containing large amounts of potassium, a defect resulting from elevated levels of intracellular potassium, since the ptsN gene product, EIIANtr, allosterically inhibits TrkA, a component of the low-affinity (high-flux) potassium uptake system (9). Thus, in the ptsN knockout, accumulated potassium ions reduce AHASI (ilvBN) activity and thereby limit growth by creating an isoleucine pseudoauxotrophy, which may be overcome by adding isoleucine (Fig. 1).
The MG1655, W3110, and MC4100 strains carry an inactive AHAS isozyme, AHASII, encoded by the ilvG and ilvM genes, which are transcribed together (Fig. 1) (1, 3, 16). The addition of valine to the medium completely inhibits the growth of MG1655 or W3110, regardless of the presence of ptsN, presumably through feedback inhibition of AHASI and AHASIII (10). However, the addition of leucine or isoleucine merely retards the growth of a ptsN knockout to a greater extent than that of the parental MG1655 strain. This occurs through feedback inhibition of AHASI and AHASIII combined with the mutant's dysregulated intracellular potassium levels, which synergistically inhibit AHASI (Fig. 1) (9). After observing the sensitivity of ΔptsN strains to potassium and leucine, a hypothesis for the role of ptsN in nitrogen metabolism was proposed: Since ptsN can control intracellular K+, modulation of K+ could tailor intracellular leucine concentrations and thus the activity of Lrp, the leucine-responsive regulatory protein, which activates and represses numerous genes in nitrogen metabolism.
In this work, we demonstrate the importance of the fact that all of the in vivo studies of the Ntr PTS to date have utilized E. coli strains MG1655 and W3110 or their derivatives (7–11, 13). We show that characteristic phenotypes previously ascribed to the ptsN knockout are not manifested in a genetic background containing functional ilvG, as AHASII is not inhibited by potassium and is less completely feedback inhibited by the amino acid end products than is AHASI or AHASIII. Neither K+ nor leucine addition inhibits the growth of the ilvG+ ΔptsN strain. Moreover, in an ilvG+ strain, knockout of ptsN does not alter growth across a range of carbon, nitrogen, sulfur, and phosphorus sources. Thus, the absence of ilvG function in MG1655 and W3110 confounds experimental evidence linking the Ntr PTS to nitrogen metabolism and raises the possibility that the ptsN gene is linked to nitrogen metabolism only through its genomic location.
The strains used in this work are listed in Table 1. The ΔptsN::Kanr and ΔilvG::Kanr alleles originated from the Keio deletion collection (2) and were transferred into the relevant background by transduction using P1 phage. Kanamycin resistance cassettes were removed by the pCP20 plasmid method and cured by growth at 42°C (5). Removal of ptsN was confirmed by PCR with gene-specific primers.
All cultures were grown at 37°C. K200 medium consists of Gutnick minimal medium (6) containing 0.4% glucose and the specified nitrogen source (NH4Cl if not otherwise indicated) at 10 mM. Equimolar Kn medium, which contains n mM K+ ions, was created by making medium in which all of the potassium salts in Gutnick medium were replaced with the corresponding sodium salt (K0) and combined with K200 medium in the necessary proportion. All media were sterilized by filtration through 0.45-μm HNW filters (Millipore).
For leucine inhibition tests, exogenous l-leucine (Sigma) from a freshly prepared, filter-sterilized stock solution was added to a final concentration of 10 mM in K10 medium with 0.4% glucose and 10 mM NH4Cl. Overnight K10 cultures were diluted 1:50, and growth was assayed by absorbance measurements. Absorbance measurements were obtained using a Biotek Synergy II plate reader in a 96-well format at 600 nm using clear, flat-bottom plates. To facilitate comparison, growth curve data for Fig. 2 and and33 were adjusted to equate optical densities at 600 nm (OD600) of 0.05 and 0.3, respectively, to 0 h.
To test the impact of potassium ion concentration on growth, overnight K10 cultures were diluted 1:100 in K1, K10, K20, K50, K100, or K200 medium with 0.4% glucose and 10 mM NH4Cl, and absorbance measurements were acquired as described above. Growth rates and their errors were computed with Microsoft Excel 2003 from linear fits of log-transformed ODs.
Individual colonies from LB medium plates were used to inoculate overnight cultures into K200 or K10 medium containing 0.4% glucose and the indicated nitrogen source at 10 mM with 2.5 mM NH4Cl to allow adaptation to the poorer nitrogen sources while preventing selection of fast-growing mutants. Cultures were diluted 1:50 in K200 or K10 medium containing 0.4% glucose and the matching indicated nitrogen source at 10 mM, and growth was monitored with a plate reader as specified above. Representative samples from two independent experiments each with 4 replicates were used to score the relative growth rate in each medium according to the slope and final OD. l-Amino acid sources unable to support growth were tested, but the results have been omitted from Table 2.
Growth measurements on Phenotype MicroArray plates PM1 to PM8 were performed by Biolog in Hayward, CA. NCM3722 and a PCR-verified NCM3722 ΔptsN strain were inoculated into rich stab cultures and shipped on ice. Growth tests were performed aerobically at 37°C. Consensus results were produced from duplicate experiments by averaging growth measurement at 24 or 48 h according to the manufacturer's guidelines.
We hypothesized that the inhibitory effect of exogenous leucine in MG1655 ΔptsN mutants resulted from the lack of a functional ilvG gene. To test this, we first grew E. coli K-12 strains with and without intact ilvG and/or ptsN in minimal medium and in minimal medium supplemented with l-leucine.
As previously documented, the MG1655 (ilvG-defective) strain grew more slowly than NCM3722 (ilvG+) in minimal medium and the MG1655 ΔptsN strain exhibited an increased sensitivity to leucine (Fig. 2) (10). On the contrary, the deletion of ptsN in the NCM3722 background had no effect on growth kinetics in minimal medium or with leucine supplementation (Fig. 2). Finally, these two phenotypes of the MG1655 ΔptsN strain—slower growth in minimal medium and growth inhibition by leucine—were recapitulated by deletion of ilvG in an NCM3722 ΔptsN background (NCM3722 ΔilvG ΔptsN) (Fig. 2).
NCM3722 ΔilvG ΔptsN cultures do not exhibit the same growth kinetics as MG1655 ΔptsN, probably due to the pyrimidine pseudoauxotrophy of MG1655, which causes slightly slower growth in minimal medium. This minor pyrimidine defect is due to the polar effect of the rph-1 allele—present in MG1655 but not in NCM3722—that reduces the expression of the cotranscribed pyrimidine biosynthesis gene pyrE (16). In spite of these minor differences, it is clear that loss of ptsN creates a leucine inhibitory phenotype only if ilvG function is also absent.
MG1655 ΔptsN mutant cells exhibit decreasing growth rates with increasing potassium concentrations, from 0.78 h−1 at concentrations lower than 1 mM K+ to 0.06 h−1 at 120 mM K+ (9). We hypothesized that a functional ilvG gene would prevent this growth inhibition by maintaining isoleucine biosynthesis in the presence of high potassium concentrations (Fig. 1). To test this, we measured the growth rates of E. coli K-12 strains with and without intact ilvG and/or ptsN in medium with potassium concentrations between 1 and 200 mM with equal osmolarity (sodium in place of potassium).
No significant growth rate differences were observed between NCM3722 and NCM3722 ΔptsN, which both appear relatively agnostic to the salts in this range (Table 3, first and second columns). However, knocking out ilvG in the NCM3722 ΔptsN background manifests the inhibitory effect of potassium (Table 3, third column). At 1 mM K+, the growth rate of NCM3722 ΔilvG ΔptsN is approximately 40% lower than NCM3722 ΔptsN and decreases to 0.08 h−1 at 200 mM K+, which is over 90% lower than that of NCM3722 ΔptsN.
After finding that functional ilvG prevents both leucine and potassium growth inhibition of NCM3722 ΔptsN, we explored the effects of ptsN knockout on the utilization of organic nitrogen sources in an ilvG+ strain. These effects had previously been studied exclusively in backgrounds without functional ilvG (10, 13). Additionally, previous investigations relied on colony size measurement, which is a less sensitive method than the OD measurement approach employed in this work. No major differences in nitrogen source utilization were observed in medium with a low potassium concentration (K10 medium; Table 2, columns 3 and 4). This lack of differences contrasts with prior investigations, which found that numerous growth defects resulted from ptsN deletion (10, 13).
Since potassium ions have been shown to reduce the activity and expression of AHASI (9, 10) and recently have been hypothesized to regulate sigma factor selectivity in an MG1655 ΔptsN strain (8), we investigated whether a 20-fold increase in the medium potassium concentration at a constant osmolarity (K200 medium) would alter the growth of an NCM3722 ΔptsN strain utilizing various nitrogen sources. Although potassium increased the growth rate and cell yield obtained with many nitrogen sources, the condition-specific changes for NCM3722 and its ptsN deletion strain were similar (Table 2, columns 1 and 2). This is unlike the previously reported defects in aspartate, glutamine, glutamate, lysine, and arginine growth associated with a W3110 ΔptsN strain grown at 150 mM potassium (13).
Finally, further investigation of the growth phenotypes of the NCM3722 wild-type and ΔptsN mutant strains using Biolog (Hayward, CA) Phenotype MicroArray plates PM1 to PM8 (the complete set of nutrient sources tested is listed in Fig. S1 in the supplemental material), which compare growth across over 700 carbon, nitrogen, phosphorus, and sulfur sources, revealed no differences between the two strains in both replicates (Fig. 3; see Fig. S2 in the supplemental material). Previously, these same Phenotype MicroArray plates were used to show differences between the MG1655 and MG1655 ΔptsN strains, which were not reproduced when NCM3722 and its ΔptsN mutant derivative were compared. This large set of growth conditions clearly demonstrates that knockout of ptsN does not cause a specific or generalized growth defect in an ilvG+ strain.
In addition to nitrogen utilization, MG1655—but not NCM3722—exhibits slow growth in glycerol minimal medium, in part due to a mild pyrimidine defect and in part due to an unknown locus (16). Although the mechanism is not known, loss of ptsN aggravates this minor glycerol defect and/or interacts with another locus to render MG1655 ΔptsN unable to grow with glycerol or succinate as the sole source of carbon (10). However, the loss of ptsN does not alter the growth rate of NCM3722 during glycerol utilization (Fig. 3, plate 1, location B3).
In this study, we compared the growth characteristics of several E. coli K-12 strains and their Ntr PTS-defective derivatives. We found that the previously reported K+- and leucine-induced growth inhibition phenotypes of a ΔptsN mutant strain are not manifested in an ilvG+ strain. Finally, we used an unbiased set of growth conditions—Phenotype MicroArray plates—to show that NCM3722 ΔptsN does not exhibit defects across a wide variety of media. Our results argue that loss of EIIANtr does not cause defects in organic nitrogen source utilization in a truly prototrophic K-12 strain.
Existing evidence for a link between the Ntr PTS and nitrogen metabolism hinged on the defects in branched-chain amino acid metabolism that were manifested in ΔptsN derivatives of MG1655 or W3110. Lee and coworkers (9) reasoned that since EIIANtr allosterically inhibits the low-affinity potassium transporter TrkA and K+ inhibits leucine biosynthesis in MG1655 ΔptsN, EIIANtr modulates nitrogen metabolism by way of the leucine-responsive transcriptional regulator Lrp. Recently, it has also been suggested that K+ alters nitrogen metabolism through sigma factor selectivity (8). We extended the growth studies of alternate nitrogen sources and found that NCM3722 ΔptsN does not exhibit a deficit in organic nitrogen source utilization across a broad range of substrates at various potassium concentrations with glucose as the carbon source (Table 2). Taking these new findings into account, it can be concluded that EIIANtr is not a physiological regulator of nitrogen source utilization by cells.
Since the discovery of ptsN and ptsO, it has been thought that carbon source quality alters ptsN activity (12). Shortly after the discovery of ptsN and ptsO, investigators identified ptsP (EINtr) as an EI paralog and a third member of the Ntr PTS by measuring in vitro phosphotransferase activity between EINtr and NPr and between NPr and EIIANtr (14, 17). In vitro cross talk of the Ntr PTS with the carbohydrate PTS was also detected (17), which indicated that the presence of PTS carbohydrates might affect the nitrogen metabolism regulatory activity of EIIANtr, as a phosphorylated form of EIIANtr neither inhibits TrkA nor activates KdpDE (12). However, the broad coverage of our growth tests (Table 2) and the Phenotype MicroArray plates (Fig. 3), which together contain both PTS carbon sources and non-PTS carbon sources crossed with NH4 and organic nitrogen sources, found no growth deficit in an NCM3722 ΔptsN strain. The lack of any defect across these nutrient combinations indicates that relief of catabolite repression or changes in EIIANtr phosphorylation do not result in situations where loss of ptsN decreases the growth rate. Thus, unlike the reports for a W3110 ptsN-null mutant (13), to which the Ntr PTS largely owes its name, carbon source quality does not affect organic nitrogen source utilization in an ilvG+ ΔptsN strain.
What may be said about the role of EIIANtr is that it is a regulator of K+ uptake. It is possible that the EIIA-like structure is designed to enable regulation of K+ uptake to respond to the presence of PTS carbohydrates such as glucose. In this way, the carbohydrate environment may be linked to the Km and Vmax of potassium uptake. The GAF domain of PtsP—atypical of PTS proteins—may also shift the phosphorylation state of EIIANtr to effect potassium uptake, but no modulators of its activity have been found to date (12). The ptsN and ptsO genes, along with two predicted genes of unknown function, hpf and yhbJ, are located in the rpoN operon, which is believed to be constitutively expressed (15). Recently, the ptsN homolog in Salmonella was found to negatively control Salmonella pathogenicity island 2 through interactions with SsrB (4).
Outstanding questions concerning the Ntr PTS are numerous. For instance, neither the activator of EINtr nor the terminal phosphate acceptor of EIIANtr~P has been found. Also, the evolutionary advantage of potassium uptake regulation by the Ntr PTS is not clear. When addressing these questions, it is advisable to use a suitable ilvG+ strain which does not suffer from nonphysiological inhibitory effects and allows access to growth conditions that do not sustain the growth of ilvG-defective ΔptsN strains.
M.L.R. and this work are supported by the National Science Foundation Graduate Research Fellowship under grant DGE-0646086. J.D.R. is supported by the National Science Foundation CAREER Award MCB-0643859, Department of Energy grant DE-SC0002077, the Air Force Office of Scientific Research grant FA9550-09-1-0580, and the National Institutes of Health grant P50GM071508.
†Supplemental material for this article may be found at http://jb.asm.org/.
Published ahead of print on 29 April 2011.