Induction of the Pho Regulon Suppresses the Growth Defect of an Escherichia coli sgrS Mutant, Connecting Phosphate Metabolism to the Glucose-Phosphate Stress Response

Abstract

Some bacteria experience stress when glucose-6-phosphate or analogues like α-methyl glucoside-6-phosphate (αMG6P) accumulate in the cell. In Escherichia coli, the small SgrS RNA is vital to recovery from glucose-phosphate stress; the growth of sgrS mutants is strongly inhibited by αMG. SgrS helps to restore growth in part through inhibiting translation of the ptsG mRNA, which encodes the major glucose transporter EIICB^Glc. While the regulatory mechanism of SgrS has been characterized, little is known about how glucose-phosphate stress connects to other aspects of cell physiology. In the present study, we discovered that mutation of pitA, which encodes the low-affinity transporter of inorganic phosphate, partially suppresses the αMG growth defect of an sgrS mutant. Induction of the stress response was also reduced in the sgrS pitA mutant compared to its sgrS parent. Microarray analysis suggested that expression of phosphate (Pho) regulon genes is increased in the sgrS pitA mutant compared to the sgrS parent. Consistent with this, we found increased PhoA (alkaline phosphatase) activity in the sgrS pitA mutant compared to the sgrS strain. Further, direct induction of the Pho regulon (in a pitA⁺ background) also resulted in partial suppression of the sgrS growth defect. The suppression was reversed when Pho induction was prevented by mutation of phoB, which encodes the Pho transcriptional activator. Deletion of individual Pho structural genes in suppressed strains did not identify a single gene responsible for suppression. Altogether, this work describes one of the first studies of glucose-phosphate stress physiology and suggests a novel connection of carbon and phosphate metabolism.

INTRODUCTION

Bacterial small RNAs (sRNAs) are important regulators of cell function and play key roles in everything from metabolism and biofilm formation to virulence and the responses to many environmental stresses (17, 65). sRNAs regulate expression through a wide array of mechanisms, such as protein sequestration, conformational changes in mRNA (riboswitches), and base pairing with mRNA targets, the latter of which generally leads to enhancement of translation (in the case of positive regulation) or translational repression and/or mRNA degradation (in the case of negative regulation) (for reviews, see references 17, 50, and 65). Although many mechanisms of regulation by sRNAs are well understood, much remains to be learned about the biological functions of many of these regulatory responses and their effects on the physiology of the cell (46).

The SgrS sRNA is vital to the recovery of Escherichia coli from glucose-phosphate stress, a condition that occurs when the phosphorylated form of glucose or certain nonmetabolizable glucose analogues like α-methyl glucoside (αMG) and 2-deoxyglucose accumulate in the cell and hinder growth (14, 54, 55). In E. coli, both glucose and αMG can be brought into the cell and phosphorylated by the phosphoenolpyruvate phosphotransferase system (PTS) transporter EIICB^Glc, which is encoded by ptsG. EIICB^Glc itself is activated via phosphorylation by EIIA^Glc (encoded by crr), the penultimate member of a phosphorelay cascade dedicated to transporting sugars into the cell (5, 52). When sugar-phosphates amass in the cell, sgrS expression is induced by the transcriptional regulator SgrR in response to an unknown stress signal (55, 56). Both SgrR and SgrS are crucial for the ability to deal with glucose-phosphate accumulation, as sgrS and sgrR mutants are inhibited for growth during stress (40, 55). SgrS has been shown to aid in recovery from stress by preventing synthesis of the glucose transporter EIICB^Glc, stopping accumulation of sugar-phosphates (26, 40, 55). SgrS negatively regulates ptsG posttranscriptionally through specific base pairing with ptsG mRNA, resulting in translation inhibition and subsequent degradation of ptsG transcript by the RNase E degradosome (26, 39, 40). Recently, a second regulatory target of SgrS was described, the PTS transporter EIIABCD^Man encoded by manXYZ. EIIABCD^Man is able to transport mannose and, with a lower affinity, glucose and 2-deoxyglucose into the cell. During stress, SgrS negatively regulates the levels of manXYZ transcript through translational inhibition in a manner similar to that of ptsG (45). As with EIICB^Glc, prevention of stressor accumulation in the cell appears to be the primary consequence. Interestingly, the two transporters differ in their affinity for stress-causing glucose analogues: EIICB^Glc has a higher affinity for αMG than 2-deoxyglucose, and the opposite is true for EIIABCD^Man (1, 10, 19, 44, 49). Accordingly, deletion of ptsG leads to decreased stress and reduced induction of the sgrS stress response during growth with αMG, while a ΔmanXYZ mutant exhibits decreased stress and sgrS induction during growth with 2-deoxyglucose (45). Thus, regulation of both targets allows SgrS to cope with stress caused by different sugar-phosphates. In addition to its function as an sRNA, SgrS encodes a small protein, SgrT, which inhibits activity of the EIICB^Glc protein and, as such, also limits transport of stressor sugar-phosphates into the cell (58).

The regulatory mechanism by which E. coli deals with glucose-phosphate stress is therefore well described. In contrast, little is known about the cause of stress or how the glucose-phosphate stress response affects aspects of cell physiology beyond carbon transport (46, 54). To begin to identify other cellular factors that impact the glucose-phosphate stress response, we performed a screen for transposon insertion mutations that suppress the growth defects of sgrS and sgrR mutants during glucose-phosphate stress. Here, we describe a mutation in pitA that partially rescues the growth defect of an sgrS mutant. pitA encodes one of two major inorganic phosphate (P_i) transporters of E. coli. PitA is a low-affinity transporter; when P_i is plentiful, it is cotransported into the cell by PitA, along with divalent cations (6, 25, 47, 48, 57, 67). One of the cations transported by PitA is Zn²⁺; while pitA is always expressed at some level, its transcription is upregulated in response to the combination of low P_i and high Zn²⁺ concentrations (6, 25). pstSCAB encodes the other major P_i transporter, which has a high affinity for P_i. PstSCAB is part of the phosphate (Pho) regulon activated by the two-component system PhoBR and induced during growth in phosphate starvation conditions (47, 48, 53, 60, 62, 67). PstSCAB is also active in a pitA mutant, supporting growth on P_i (18, 21, 48). Although E. coli possesses another low-affinity P_i transporter, PitB, it does not support growth in a pst pitA mutant on an inorganic source of phosphate (18, 20, 48). In this study, we characterize the effects of the pitA mutation on the glucose-phosphate stress response and demonstrate that induction of the Pho regulon can rescue an sgrS mutant from glucose-phosphate stress. We also discuss possible physiological consequences of this novel connection between phosphate and carbon metabolism.

MATERIALS AND METHODS

Bacterial strain construction.

E. coli strains and plasmids used in these experiments are listed in Table 1. Most strains are derived from E. coli MG1655 or DJ480 (D. Jin, National Cancer Institute [NCI]). Mutant alleles were introduced into strains by P1 phage transduction except where otherwise noted. Unless indicated differently, mutations were verified by PCR using GoTaq polymerase (Promega, Madison, WI) according to the manufacturer's instructions. Strain GR102 was constructed by moving the pitA::Tn5 mutation isolated from the transposon insertion-deletion mutant screen (see below) to sgrS mutant host CS123, which contains the sgrS1 mutation and is defective in both ptsG regulation and the ability to recover from glucose-phosphate stress (59). Deletion-insertion alleles of the pitA, pitB, poxB, pstC, phoB, and uhpT loci (as well as those of the Pho regulon members listed in Table 3) containing kanamycin (kan) cassettes flanked by FLP recombination target (FRT) sites were moved into the indicated backgrounds from the Keio collection of single-gene mutations in the wild-type strain background BW25113 (3). The ΔpitA::FRT-kan-FRT allele was moved into Δlac backgrounds DJ480 (wild type), CS123, and CV700 (ΔsgrR::cm) (55) to create, respectively, strains GR103, GR100, and GR105. ΔpitA::FRT-kan-FRT was also moved into lac⁺ strains MG1655 (wild type) and CS141 (ΔsgrS; constructed by C. Wadler in our lab) to make GR151 and GR150 for use in inducer exclusion assays (see below). Strain GR161 was constructed by moving the ΔpstC::FRT-kan-FRT allele into CS123. The kanamycin resistance cassettes were removed from GR100 and GR161 using FLP-mediated site-specific recombination (13), resulting in strains GR101 and GR165. The FRT-kan-FRT alleles of pitB, poxB, and uhpT were each then introduced into strain GR101 to create strains GR107, GR108, and GR109, respectively. Strain GR166 was constructed by moving the ΔphoB::FRT-kan-FRT allele into GR165. The mutants listed in Table 3 each were constructed in an identical fashion in sgrS1 ΔpitA::FRT (GR101) and sgrS1 ΔpstC::FRT (GR165) backgrounds.

Table 1.

E. coli strains and plasmids used in this study

Strain or plasmid	Description or relevant characteristic(s)	Source or reference(s)
E. coli strains
DH5α	E. coli general cloning strain	Invitrogen (Carlsbad, CA)
MG1655	Wild-type E. coli K-12	D. Jin (NCI)
DJ480	MG1655 ΔlacX74	D. Jin (NCI)
CS123	DJ480 sgrS1(G176C,G178C)	59
CV700	DJ480 ΔsgrR::cm	55
BAH100	DJ480 λattB::PsgrS-lacZ	45, 51
CL109	CS123 λattB::PsgrS-lacZ	This study
CS141	MG1655 ΔsgrS setA::cm	This study
GR100	CS123 ΔpitA::FRT-kan-FRT	This study
GR101	CS123 ΔpitA::FRT	This study
GR102	CS123 pitA::Tn5; Kmr	This study
GR103	DJ480 ΔpitA::FRT-kan-FRT	This study
GR105	CV700 ΔpitA::FRT-kan-FRT	This study
GR106	CS123 ΔpitB::FRT-kan-FRT	This study
GR107	GR101 ΔpitB::FRT-kan-FRT	This study
GR108	GR101 ΔpoxB::FRT-kan-FRT	This study
GR109	GR101 ΔuhpT::FRT-kan-FRT	This study
GR150	CS141 ΔpitA::FRT-kan-FRT	This study
GR151	MG1655 ΔpitA::FRT-kan-FRT	This study
GR159	GR103 λattB::PsgrS-lacZ	This study
GR160	GR101 λattB::PsgrS-lacZ	This study
GR161	CS123 ΔpstC::FRT-kan-FRT	This study
GR165	CS123 ΔpstC::FRT	This study
GR166	GR165 ΔphoB::FRT-kan-FRT	This study
Plasmids
pRL27	Kmr; donor for Tn5 transposon mutants	33
pZE21-MCS1	Kmr Tcr; tetracycline-inducible PLtetO-1 promoter	35
pZEGR1	pZE21-MCS1 + pitA (∼1.6-kb insert)	This study

Table 3.

Effects of individual Pho regulon gene mutations on sgrS pitA and sgrS pstC suppression

Genea	Operonb	Encoded functionc	Cause of suppression?d
amn	amn	AMP nucleosidase	No
argP (iciA)	argP	Arginine transport transcription factor	No
bfr	bfd-bfr	Bactoferrin	No
cspI	cspI	Cold shock-like protein	No
dps	dps	Global regulator (stationary phase, oxidative stress)	No
dusB (yhdG)	dusB-fis	Putative dehydrogenase	No
eda	edd-eda	Aldolase	No
gadB	gadBC	Glutamine decarboxylase (acid stress)	No
gadC (xasA)	gadBC	Acid sensitivity protein	No
mcbA (ybiM)	mcbA	Protein involved in colanic acid production	No
phnD	phnCDEEFGHIJKLMNOP	Phosphonate transport	No
phnK	phnCDEEFGHIJKLMNOP	Phosphonate metabolism	No
phoA	phoA-psiF	Alkaline phosphatase	No
phoE	phoE	Outer membrane protein	No
phoH	phoH	Conserved ATP binding protein	No
potE	speF-potE	Putrescine transport protein	No
psiE (yjbA)	psiE	Predicted protein	No
psiF	phoA-psiF	Predicted protein	No
sra (rpsV)	bdm-sra	30S ribosomal subunit protein S22	No
tktB	talA-tktB	Transketolase	No
ugpB	ugpBAECQ	Glycerol-3-phosphate transport (periplasmic binding)	No
ydfI	ydfI	Putative oxidoreductase	No
yibD	yibD	Predicted glycosyl transferase/regulator	No
ytfK	ytfK	Conserved uncharacterized protein	No

^aAlternate names are shown in parentheses. Genes in the Pho regulon were identified using data from references 4, 24, and 27.

^bBased on data from reference 27.

^cProtein function based on data from references 4 and 27.

^dDoes deletion mutation in an sgrS pitA or sgrS pstC mutant background abrogate suppression of sgrS growth defect in LB with αMG to induce glucose-phosphate stress?

The P_sgrS-lacZ transcriptional reporter fusion (45, 51) was inserted into the λattB chromosomal site of Δlac strains CS123 (sgrS1), GR103 (ΔpitA::FRT-kan-FRT), and GR160 (sgrS1 pitA::FRT) to yield strains CL109 (constructed by C. Lloyd in our lab), GR159, and GR160, respectively.

Construction and expression of ectopic pitA vector.

A 1,591-nucleotide (nt) region containing the wild-type pitA gene (as well as 33 nt upstream and 58 nt downstream of the pitA coding sequence) was amplified by PCR of chromosomal DNA using the primers O-PitAApaF (5′-NNNGGGCCCCGCGTTCATGTCCTCAAAATGG-3′ [the engineered ApaI restriction site is underlined]) and O-PitABamR (5′-NNNGGATCCGCGCACTATGTCACAATCTGAAG-3′ [the engineered BamHI restriction site is underlined]) (Integrated DNA Technologies, Inc., Coralville, IA) and Phusion high-fidelity DNA polymerase (Finnzymes, New England BioLabs, Ipswich, MA) according to the manufacturer's instructions. The correct sequence was verified at the Biotechnology Center of the University of Illinois at Urbana-Champaign. Using the engineered ApaI and BamHI restriction sites, the amplified pitA product was cloned into the vector pZE21-MCS1, which has a tetracycline-inducible P_LtetO-1 promoter (35) and was provided to us by J. Cronan. The resulting plasmid, pZEGR1, was transformed by electroporation into the cloning strain DH5α (Invitrogen, Life Technologies, Carlsbad, CA). Either pZEGR1 or pZE21-MCS1 (as a negative control) were transformed by electroporation into the wild-type DJ480, the sgrS1 parent CS123, and the sgrS1 ΔpitA::FRT mutant GR101.

Media and growth conditions.

Luria-Bertani (LB) medium (36) was used to culture bacteria at 37°C unless otherwise indicated. To induce glucose-phosphate stress, 0.5% αMG was added to media except where otherwise described. M63 minimal medium (36) was used to test growth on particular carbon sources (0.2% glucose or 0.4% glycerol) at indicated times, in both the absence and presence of αMG to induce stress. For inducer exclusion assays, cultures were grown in TB (Bacto Tryptone) medium (BD, Franklin Lakes, NJ). Kanamycin was added to media at a concentration of 25 μg ml⁻¹ where described to maintain plasmids or select for mutants. To induce expression of P_LtetO-1-pitA in plasmid pZEGR1, anhydrotetracycline (Sigma-Aldrich, St. Louis, MO) was added at a concentration of 100 ng ml⁻¹.

For growth experiments, strains grown overnight were subcultured and normalized to an optical density at 600 nm (OD₆₀₀) of approximately 0.02 in fresh LB medium. Strains were then grown until they reached an OD₆₀₀ of approximately 0.1, at which time each culture was split in two, and αMG was added to one of the two cultures to induce stress. The OD₆₀₀ was used to monitor growth for 7 h. Because overnight growth during glucose-phosphate stress cannot be reliably measured due to risk of suppressor mutations, growth at 24 h was measured qualitatively by monitoring individual colony size on plate media with and without αMG.

Construction and screening of Tn5 mutant library.

The original sgrS pitA::Tn5 mutant was identified in a screen for transposon insertion mutants that suppress the growth defects of sgrR and sgrS mutants during sugar-phosphate stress. CV700 (ΔsgrR::cm) and CS123 (sgrS1) mutants were subjected to Tn5 transposon insertion random mutagenesis by transformation with plasmid pRL27 (33) provided to us by W. Metcalf. Libraries of Tn5 transposon insertion mutants were selected based on resistance to kanamycin, and pools of the mutants were screened on LB medium with 0.5% αMG to induce stress. Mutants that formed colonies larger than their sgrR or sgrS parent mutants were isolated. The locations of transposon insertion mutations were determined by cloning (33) and sequencing of the mutated regions at the Biotechnology Center of the University of Illinois at Urbana-Champaign or ACGT, Inc. (Wheeling, IL).

β-Galactosidase assays.

Strains containing the P_sgrS-lacZ transcriptional fusion were grown overnight and subcultured in fresh LB medium to a normalized OD₆₀₀ of approximately 0.02. At an OD₆₀₀ of approximately 0.1, 0.01% αMG was added to induce stress; in this experiment, a lower concentration of αMG was used because expression of the P_sgrS-lacZ fusion is highly sensitive and easily saturated at higher αMG concentrations (51). Samples were taken at indicated times and subjected to Miller assay as previously described (36). Briefly, samples were suspended in Z-buffer, and reactions were executed at 28°C using 4 mg/ml 2-nitrophenyl β-d-galactopyranoside as a substrate and 1 M Na₂CO₃ to stop the reaction (36).

Alkaline phosphatase (PhoA) assays.

Alkaline phosphatase (PhoA) assays used to monitor PhoA activity generated from expression of the native chromosomal gene were performed as described previously (8) with the modifications mentioned. Briefly, strains were cultured overnight and subsequently subcultured and normalized to an OD₆₀₀ of approximately 0.02 in fresh LB medium. At an OD₆₀₀ of approximately 0.1, the first sample was taken; in the case of cultures being subjected to stress, αMG was first added. A second sample was taken 1 h later. Samples were suspended in Tris-HCl (pH 8.0), and PhoA activity was measured at 37°C using 8 mg/ml 4-nitrophenylphosphate (Thermo Fisher Scientific, Waltham, MA) as a substrate and 1 M K₂HPO₄ to stop the reaction (8).

Inducer exclusion assays.

Inducer exclusion assays were performed as described previously (22, 58), with the mentioned modifications. Overnight cultures of strains containing the native chromosomal lacZ gene were subcultured and normalized to an OD₆₀₀ of approximately 0.02 in fresh TB medium supplemented with 0.2% lactose and 0.5% αMG to induce glucose-phosphate stress. Samples of cultures were taken at the indicated time points, and β-galactosidase activity was measured by Miller assay as described previously (36).

RESULTS

Mutations in pitA partially suppress the growth defect of an sgrS mutant during glucose-phosphate stress.

To identify potential connections of the glucose-phosphate stress response to other aspects of cell physiology, libraries of ΔsgrR::cm (55) and sgrS1 (59) mutants with random kanamycin-resistant Tn5 transposon insertion mutations (33) were screened for suppression of parent growth defects (based on the larger colony sizes of mutants compared to the parent) during glucose-phosphate stress induced by αMG. The vast majority of suppressor mutations were located in ptsG or crr, which presumably result in suppression by preventing the uptake of αMG (45) (data not shown). One mutation, in pitA, encoding the P_i transporter (47, 48, 67), was found to partially rescue the growth defect of the sgrS parent. pitA is divergently oriented from both the upstream gene yhiN [predicted to encode a flavin adenine dinucleotide (FAD)/NAD(P)-binding oxidoreductase], and the downstream gene uspB (encoding the universal stress protein B) (15, 27), making it unlikely that the suppressor phenotype was due to any polar effects. The suppression effect of pitA was confirmed by transducing both the original pitA::Tn5 mutation and a second, independent, kanamycin-resistant ΔpitA::FRT-kan-FRT deletion mutation (3) into the sgrS1 parent background. Both reconstructed sgrS pitA mutants exhibit the same partial suppression phenotype as the original: in the presence of αMG, they grow significantly better than the sgrS parent, although not as well as wild-type E. coli. The sgrS pitA mutants exhibit this suppression both in liquid culture (Fig. 1A) and on plates after 24 h of growth (Fig. 1B). An sgrR pitA mutant exhibits similar suppression to that of sgrS pitA, indicating that the suppression effect of the pitA mutation is not specific to sgrS (see Fig. S1A in the supplemental material).

Fig 1.

Mutation of pitA partially suppresses the sgrS mutant defect during growth in αMG. (A) Wild-type (DJ480; squares), sgrS1 (CS123; circles), and sgrS1 ΔpitA::FRT-kan-FRT (GR100; triangles) strains were grown in liquid LB medium in the absence of 0.5% αMG (− αMG) or presence of 0.5% αMG (+ αMG) to induce glucose-phosphate stress. αMG was added to the indicated cultures at an OD₆₀₀ of approximately 0.1, and OD₆₀₀ was monitored over time. A representative example is shown (n = 3). (B) Complementation of sgrS pitA suppression. Growth of wild-type (WT; DJ480) sgrS1 (CS123), and sgrS1 ΔpitA::FRT (GR101) strains carrying either vector pZE21-MCS1 (left plate; negative control) or pZE21-MCS1 with a wild-type copy of pitA under the control of the P_LtetO-1 promoter (right plate). Strains were grown for 24 h on solid LB medium containing 100 ng ml⁻¹ anhydrotetracycline (aTc) to induce P_LtetO-1 expression and 0.5% αMG to induce stress.

To test whether mutating pitA provided an overall benefit during glucose-phosphate stress, we also constructed a pitA mutation in the wild-type sgrS⁺ background. However, the pitA mutation did not measurably change the growth of an sgrS⁺ strain in the presence of αMG (see Fig. S1B in the supplemental material). In addition, we eliminated the possibility that the other low-affinity P_i transporter, pitB, had an effect similar to pitA by showing that mutating pitB did not result in suppression in the sgrS background (Fig. S1C). This was expected, given the relatively minor role in phosphate transport attributed to pitB (18, 20, 48).

To verify the causality of the pitA mutation for the partial suppression of the sgrS growth defect, a wild-type copy of pitA was ectopically expressed from the tetracycline-inducible P_LtetO-1 promoter in plasmid pZE21-MCS1 (35). This construct was introduced into the wild type, the sgrS1 parent, and an sgrS1 pitA mutant. When the strains were grown on αMG to induce glucose-phosphate stress, expression of P_LtetO-1-pitA resulted in loss of sgrS pitA mutant suppression, confirming that the pitA mutation is responsible for the partial rescue of the sgrS growth defect (Fig. 1B). Expression of P_LtetO-1-pitA does not affect the growth of the wild type or the sgrS parent during stress, which was expected, given that both strains have wild-type chromosomal copies of pitA. In addition, overexpression of pitA does not affect recovery of the wild-type strain from stress (Fig. 1B; data not shown).

The pitA mutation reduces induction of the glucose-phosphate stress response.

Consistent with their inability to recover from glucose-phosphate stress, sgrS mutants experience greater stress than the wild type does, as reflected by higher levels of P_sgrS-lacZ expression during growth in the presence of αMG (51). Since mutation of pitA partially rescues the glucose-phosphate growth defect of the sgrS strain, it was hypothesized that stress response induction is lower in the sgrS pitA mutant than in the sgrS strain. To investigate the effect of mutating pitA on stress response signaling, we introduced a P_sgrS-lacZ transcriptional fusion (45, 51) at the λattB chromosomal sites of the Δlac wild-type, sgrS, pitA, and sgrS pitA strains and monitored the level of stress response induction during growth with αMG. In contrast to their respective wild-type and sgrS parent strains, expression of P_sgrS-lacZ was reproducibly lower in both pitA and sgrS pitA mutants, suggesting that pitA mutants experience less stress (Fig. 2). Consistent with the partial suppression of the sgrS growth defect, the sgrS pitA mutant displayed an intermediate level of P_sgrS-lacZ induction lower than its sgrS parent but higher than the wild type (Fig. 2). (The sgrS mutant exhibited increased P_sgrS-lacZ expression compared to the wild type [Fig. 2], as observed previously [51].)

Fig 2.

Expression of P_sgrS-lacZ in pitA mutants during growth in αMG. Wild-type (BAH100; squares), sgrS1 (CL109; circles), ΔpitA::FRT-kan-FRT (GR159; diamonds), and sgrS1 ΔpitA::FRT (GR160; triangles) strains containing chromosomal copies of P_sgrS-lacZ were grown in LB medium to an OD₆₀₀ of approximately 0.1, at which point 0.01% αMG was added to induce P_sgrS-lacZ expression. β-Galactosidase activity was monitored at 0, 15, 45, and 60 min after the addition of αMG. Error bars indicate standard deviations (n = 3).

pitA-mediated suppression is not due to decreased EIICB^Glc activity.

In the most common class of suppressor mutants (ptsG and crr), inhibition of EIICB^Glc (PtsG) expression or activity likely prevents uptake of the stressor αMG and curtails induction of the SgrRS stress response (45). Since P_sgrS-lacZ expression is decreased in sgrS pitA mutants (Fig. 2), it was hypothesized that mutation of pitA might relieve stress by somehow decreasing EIICB^Glc activity and reducing uptake of αMG. To test this hypothesis, we indirectly measured PtsG activity using an inducer exclusion assay, which previously was used to examine the effects of SgrS and SgrT on EIICB^Glc activity (58).

EIIA^Glc (Crr) activates EIICB^Glc through phosphorylation. When EIICB^Glc is actively transporting glucose, EIIA^Glc is primarily dephosphorylated. Dephosphorylated EIIA^Glc prevents the activity of certain other transporters like lactose permease, and this negative regulation is referred to as inducer exclusion (11, 22). When lactose permease is inhibited via inducer exclusion, expression of lactose-inducible genes such as lacZ is also repressed. Thus, a decrease in lacZ expression reflects an increase in EIICB^Glc activity and vice versa. The activity of LacZ (β-galactosidase activity) therefore can be used as an indirect reporter for EIICB^Glc activity.

We monitored EIICB^Glc activity during glucose-phosphate stress using β-galactosidase activity as a reflection of inducer exclusion. β-Galactosidase activity was measured in lac⁺ wild-type, sgrS, pitA, and sgrS pitA strains. Cells were grown in lactose and the glucose analogue αMG to induce stress. Wild-type cells were expected to display high LacZ activity due to induction of SgrS and subsequent inhibition of EIICB^Glc (PtsG) expression/activity. Because the sgrS mutant is unable to inhibit EIICB^Glc expression/activity, LacZ activity was expected to be low. If sgrS pitA suppression is the result of decreased EIICB^Glc activity and reduced αMG uptake, we predicted that sgrS pitA would exhibit higher LacZ activity than its sgrS parent during growth with αMG and lactose.

As expected, wild-type cells had high β-galactosidase activity, indicating the wild-type stress response was able to abrogate EIICB^Glc activity (Fig. 3A). In contrast, both sgrS and sgrS pitA mutants exhibited much lower levels of β-galactosidase activity compared to the wild type, reflecting the lack of a functional stress response to turn off EIICB^Glc activity. Surprisingly, EIICB^Glc activity was found to be slightly but reproducibly increased (i.e., decreased β-galactosidase activity) in the sgrS pitA mutant compared to its sgrS parent during stress (Fig. 3A). These results strongly suggest that sgrS pitA suppression is not caused by decreased EIICB^Glc activity leading to less αMG entering the cell.

Fig 3.

Effects of pitA mutation on lacZ expression (inducer exclusion) as a measure of EIICB^Glc activity during stress. (A and B) Wild-type (MG1655; black bars), ΔsgrS (CS14; white bars), ΔpitA::FRT-kan-FRT (GR151; vertically lined bars), and ΔsgrS ΔpitA::FRT-kan-FRT (GR150; gray bars) were grown in TB medium supplemented with 0.2% lactose and 0.5% αMG to induce stress, and β-galactosidase activity was monitored at the indicated time points. Higher β-galactosidase activity indicates lower EIICB^Glc activity. (B) β-Galactosidase activity of the ΔpitA mutant is shown at earlier time points when expression differences were observed (see the text for details).

Like the wild type, the (sgrS⁺) pitA mutant exhibited high levels of β-galactosidase activity at 3.5 h (1,506 ± 623 Miller units for the wild type compared to 1,572 ± 264 for the pitA mutant; n = 3), which is consistent with the fact that the wild-type SgrRS response of both strains is able to repress EIICB^Glc activity. However, at earlier time points (presumably before EIICB^Glc activity was fully turned off by the SgrRS response), the pitA mutant exhibited higher EIICB^Glc activity (lower β-galactosidase activity) than the wild type (Fig. 3B). Thus, both the sgrS pitA and pitA mutants appear to have increased EIICB^Glc activity compared to their parent strains.

To determine whether the observed increase in EIICB^Glc activity in the pitA mutants had an effect on growth, pitA and sgrS pitA mutants, along with their wild-type and sgrS parents, were grown on M63 minimal medium with glucose as the sole carbon source, with and without αMG to induce stress. Consistent with enhanced EIICB^Glc activity in these strains, pitA and sgrS pitA exhibited slightly increased growth compared to the parent strains, even in the presence of αMG (Table 2). In contrast, growth of pitA mutants on a non-PTS carbon source, glycerol, was not enhanced compared to parental strains in the presence or absence of αMG (Table 2). (The growth-defective phenotype of sgrS mutants is more severe on minimal medium containing glycerol than on LB or minimal media containing PTS carbon sources for reasons that are not yet clear [51]. The pitA mutation does not suppress the growth defect of the sgrS mutant on minimal medium containing glycerol in the presence of αMG.) A growth advantage of pitA and sgrS pitA mutants was not observed in the absence of glucose-phosphate stress in the rich medium LB, which contains only small amounts of glucose (Table 2; Fig. 1A). Taken together, these results are consistent with an increase in EIICB^Glc transporter activity in the pitA and sgrS pitA mutants.

Table 2.

Growth of pitA mutants on selected carbon sources

Straina	Growth of strain onb:
	LBc	LB + αMGc	M63 + 0.2% glucosed	M63 + Glu + αMGd	M63 + 0.4% glycerold	M63 + Gly + αMGd
Wild type	+	+	+	+	+	+
sgrS mutant	+	−	+	+/−	+	−e
sgrS pitA mutant	+	+/−	++	++	+	−e
pitA mutant	+	+	++	++	+	+

^aStrains DJ480 (wild type), CS123 (sgrS mutant), GR100 (sgrS pitA mutant), and GR103 (pitA mutant) were examined.

^bThe growth of the wild type and pitA mutants on selected carbon sources was determined by qualitative evaluation of bacterial colony size as follows: ++, size was increased relative to that of the wild-type control; +, size was indistinguishable from that of wild type; +/−, size was reduced compared to that of wild type; −, growth was strongly inhibited. Glu, 0.2% glucose; Gly, 0.4% glycerol; αMG, 0.5% αMG.

^cColony size 24 h after inoculation on LB plates with or without 0.5% αMG to induce glucose-phosphate stress.

^dColony size 48 h after inoculation on M63 minimal medium plates with or without 0.5% αMG to induce glucose-phosphate stress.

^eStill no growth detected after 96 h.

sgrS pitA suppression is not due to activity of the organic phosphate transporter UhpT or metabolic flux to pyruvate oxidase PoxB.

In a pitA mutant, P_i primarily enters the cell through the high-affinity transporter and Pho regulon member PstSCAB (18, 20, 25, 48, 66). However, P_i may also be able to enter pitA cells through the transporter UhpT (25, 43, 68). Normally, UhpT takes up glucose-6-phosphate in exchange for excretion of P_i. However, UhpT also has a low affinity for P_i and can work in the opposite direction, although it cannot support growth of a pst pitA mutant when grown on P_i (18, 25, 43, 48, 68). We tested the hypothesis that the pitA mutation suppressed the αMG growth defect of the sgrS mutant by promoting UhpT-dependent efflux of αMG-6-phosphate. We reasoned that if this were true, an sgrS pitA uhpT mutant should lose the suppression phenotype and demonstrate an sgrS-like growth defect in the presence of αMG. Instead, the sgrS pitA uhpT mutant displayed a suppression phenotype indistinguishable from the sgrS pitA mutant—i.e., it still grew better than the sgrS parent (data not shown). This result suggests that the UhpT transporter is not responsible for pumping out αMG-6-phosphate to yield the suppressor phenotype.

Under aerobic, low phosphate conditions, E. coli glucose metabolism switches from using pyruvate dehydrogenase (Pdh) to pyruvate oxidase (PoxB) (37). If mutating pitA simulates low phosphate conditions, the postglycolytic flux could be routed through PoxB. Since there are indications that glucose-phosphate stress may be related to an imbalance in glycolytic metabolites (54), we hypothesized that a switch to PoxB metabolism in the sgrS pitA mutant might relieve stress. However, an sgrS pitA poxB mutation still suppresses the αMG growth defect as well as sgrS pitA (data not shown), ruling out this possibility.

The phosphate (Pho) regulon is induced in the sgrS pitA mutant.

To search for other possible causes of sgrS pitA suppression, we performed microarray analysis of an sgrS pitA mutant to examine differences in gene expression compared to its sgrS parent during growth with αMG to induce stress. Expression of several members of the Pho regulon, including but not limited to phoE, phoR, and pst genes, was significantly increased between 1.6- and 5.5-fold in the sgrS pitA mutant compared to its sgrS parent (see Table S1 in the supplemental material). These results represent the most striking differences in expression observed for these strains and are consistent with previous reports that, in the absence of pitA, PhoBR-activated PstSCAB is responsible for transporting P_i (18, 20, 25, 48). We therefore hypothesized that increased expression of one or more Pho-regulated genes in the sgrS pitA mutant might be important for the suppression phenotype.

We first confirmed that the Pho regulon is induced in sgrS pitA and pitA mutants by measuring the activity of regulon member PhoA, alkaline phosphatase. As expected, the sgrS pitA and pitA mutants exhibited a low level of PhoA induction, between 2.5- and 6-fold that of the wild-type and sgrS strains (Fig. 4). PhoA was induced in both pitA mutants at similar levels in the absence (Fig. 4A) or presence (Fig. 4B) of glucose-phosphate stress. The fold changes in PhoA activity in the pitA mutants correlate with the fold changes in Pho gene expression in the sgrS pitA microarray analysis (see Table S1 in the supplemental material). In addition, it is worth noting that stress did not result in the induction of PhoA activity in wild-type E. coli (Fig. 4B), suggesting that Pho regulon induction is not required as part of the wild-type response to glucose-phosphate stress.

Fig 4.

Activity of PhoA (alkaline phosphatase) in pitA mutants. Wild-type (DJ480; black bars), sgrS1 (CS123; white bars), ΔpitA::FRT-kan-FRT (GR103; light gray bars), and sgrS1 ΔpitA::FRT-kan-FRT (GR100; dark gray bars) strains were grown in LB medium in the absence (A) or presence (B) of 0.5% αMG to induce glucose-phosphate stress. αMG was added to the indicated cultures at an OD₆₀₀ of approximately 0.1, and alkaline phosphatase activity was assayed both at that time and 1 h later. Error bars indicate standard deviations (n = 3).

Individual Pho regulon members are not responsible for sgrS pitA suppression.

If an individual member of the Pho regulon is responsible for sgrS pitA suppression, mutating that gene in the sgrS pitA background should result in loss of suppression when grown with αMG. We constructed insertion-deletion mutations in verified PhoBR-regulated genes (24, 62), as well as several genes putatively regulated by PhoBR based on transcriptome studies (4) in the sgrS pitA mutant strain (Table 3). We then monitored growth of each triple mutant for loss of suppression of the sgrS mutant growth defect on αMG. However, none of the 24 Pho genes individually deleted in sgrS pitA mutants were found to be responsible for suppression (Table 3). An sgrS pitA pstC mutant had an overall growth defect regardless of the presence of αMG (data not shown). This was expected because pitA pst mutants have an overall growth defect when grown with P_i as the sole source of phosphorus (18, 48). However, the growth of the sgrS pitA pstC mutant was not further impaired by the addition of αMG, so it does not appear that increased expression of the PstSCAB transporter is responsible for the suppression observed in the sgrS pitA mutant.

Direct induction of the Pho regulon is sufficient to partially suppress the glucose-phosphate growth defect of an sgrS mutant.

Although individual Pho members were not found to be the cause of sgrS pitA suppression, induction of all or part of the Pho regulon might still be responsible. To examine this possibility, we directly induced the entire Pho regulon in the sgrS mutant pitA⁺ background by introducing a pstC::FRT mutation. PstSCAB has long been known to inhibit activation of PhoBR through an unknown mechanism, so mutations in pst result in high levels of constitutive Pho regulon expression (62). We confirmed induction of the Pho regulon in the sgrS pstC mutant by measuring PhoA activity. With (Fig. 5B) or without (Fig. 5A) αMG to induce stress, PhoA activity was much higher in the sgrS pstC mutant than in either the wild-type or sgrS strain (Fig. 5). The level of PhoA induction in the sgrS pstC mutant was much higher (approximately 200- to 400-fold compared to the wild type) than that of the sgrS pitA mutant (approximately 3- to 6-fold) (Fig. 5). However, this was not unexpected, considering that PstSCAB is an inhibitor of PhoBR activation. To verify that the observed increase in PhoA activity was due to induction through PhoBR, we introduced a phoB::FRT-Kan-FRT insertion-deletion mutation into the sgrS pstC mutant (phoB encodes the transcriptional activator of the Pho regulon). As expected, deleting phoB in the sgrS pstC background resulted in a decrease in PhoA activity to the wild-type and sgrS parent levels (Fig. 5).

Fig 5.

Activity of PhoA (alkaline phosphatase) in Pho regulon mutants. Wild-type (DJ480; black bars), sgrS1 (CS123; white bars), sgrS1 ΔpitA::FRT (GR101; gray bars), sgrS1 ΔpstC::FRT (GR165; hatched bars), and sgrS ΔpstC::FRT ΔphoB::FRT-kan-FRT (GR166; vertically lined bars) strains were grown in LB medium in the absence (A) or presence (B) of 0.5% αMG to induce glucose-phosphate stress. αMG was added to the indicated cultures at an OD₆₀₀ of approximately 0.1, and alkaline phosphatase activity was assayed both at that time and 1 h later. Error bars indicate standard deviations (n = 3).

To test whether direct induction of the Pho regulon in the sgrS pstC mutant resulted in suppression of the sgrS glucose-phosphate growth defect, growth of wild-type, sgrS, sgrS pitA, sgrS pstC, and sgrS pstC phoB strains was measured with and without αMG to induce stress. Like the sgrS pitA mutant, the sgrS pstC mutant exhibited partial suppression of the growth defect of the sgrS parent both in culture (Fig. 6A) and after 24 h on plates (Fig. 6B), although not to wild-type levels. However, the growth defect suppression shown by the sgrS pstC mutant was not quite as robust as that of the sgrS pitA mutant (Fig. 6). The sgrS pstC strain also appeared to exhibit a slight growth defect in the absence of αMG but was still able to partially rescue the sgrS defect when grown with αMG. As with sgrS pitA, mutation of individual Pho genes in the sgrS pstC background failed to abrogate suppression (Table 3). When the entire Pho regulon was repressed by mutating phoB in the sgrS pstC mutant, suppression was lost, resulting in a growth defect similar to that of the sgrS parent strain (Fig. 6). Taken together, these results indicate that all or part of the Pho regulon is likely responsible for this suppression of the sgrS αMG growth defect.

Fig 6.

Induction of the Pho regulon partially suppresses the sgrS mutant growth defect in αMG. (A and B) Wild-type (WT) (DJ480; squares), sgrS1 (CS123; circles), sgrS1 ΔpitA::FRT (GR101; triangles), sgrS1 ΔpstC::FRT (GR165; diamonds), and sgrS1 ΔpstC::FRT ΔphoB::FRT-kan-FRT (GR166; crosses) strains were grown in liquid LB medium in the absence (A) or presence (B) of 0.5% αMG to induce glucose-phosphate stress. αMG was added to the indicated cultures at an OD₆₀₀ of approximately 0.1, and OD₆₀₀ was monitored over time. A representative example is shown (n = 4). (C) Growth at 24 h of strains from panels A and B on solid LB medium with (right plate) or without (left plate) 0.5% αMG to induce stress.

DISCUSSION

The regulatory mechanisms by which the SgrR-SgrS response aids in recovery from glucose-phosphate stress are well characterized, but the cause of stress and the physiological connections of the response to the rest of the cell are largely unknown. Here, we establish a novel connection between the glucose-phosphate stress response and phosphate metabolism. We demonstrate that mutations in the P_i transporter pitA result in the partial suppression of an sgrS mutant growth defect during stress (Fig. 1). (Figure 7 summarizes the effects of mutating pitA in the sgrS mutant.) An sgrS pitA mutant also experiences decreased levels of stress, as measured by lower levels of sgrS transcriptional activation during growth in αMG (Fig. 2). To our knowledge, this represents the first cellular physiological link to the glucose-phosphate stress response beyond carbon transport and metabolism. This phenotype also illustrates the most significant rescue of the sgrS growth defect observed in a suppressor mutant thus far, with the exception of mutations that directly prevent uptake of stressor sugar-phosphates via the EIICB^Glc (PtsG) PTS transporter (e.g., ptsG and crr) (45). We eliminated the possibility that sgrS pitA suppression likewise is due to decreased EIICB^Glc transport activity and uptake of the stressor αMG; in fact, an sgrS pitA mutant actually exhibits a slight increase in EIICB^Glc activity (Fig. 3). Suppression was also not caused by efflux of the stressor αMG by the UhpT transporter or by an alteration in central metabolism by pyruvate oxidase PoxB. However, the results of microarray analysis (see Table S1 in the supplemental material) suggested that the Pho regulon is induced to a low level in an sgrS pitA mutant, which was confirmed by PhoA alkaline phosphatase assays (Fig. 4). We found that, as in the sgrS pitA strain, direct induction of the Pho regulon in an sgrS pstC mutant also results in partial suppression of the sgrS growth defect, and repression of the Pho regulon in an sgrS pstC phoB mutant abrogated that suppression (Fig. 6). Taken together, these results indicate that induction of the Pho regulon can aid the recovery of an sgrS mutant from glucose-phosphate stress and are consistent with the notion that sgrS pitA suppression is also due to Pho regulon induction.

Fig 7.

Effects of mutating pitA in an sgrS mutant. The absence of the low-affinity inorganic phosphate transporter encoded by pitA (indicated by the large black X through the light gray cylinder) results in several physiological consequences (denoted by white arrows) for an sgrS mutant (indicated by the large black X through the double helix). Deletion of pitA leads to a slight increase in activity of the glucose/αMG (gray hexagons) transporter EIICB^Glc (dark gray ovals). In addition, induction of the glucose-phosphate stress response is decreased, as evidenced by a decrease in expression of sgrS-lacZ. In the absence of PitA, inorganic phosphate (white circles containing P) is brought into the cell by the high-affinity transporter PstSCAB (complex of gray ovals and cylinders), which is part of the phosphate (Pho) regulon. Mutation of pitA leads to induction of Pho regulon gene expression (indicated by wavy black lines under double helix), which partially suppresses the growth defect of an sgrS mutant during glucose-phosphate stress through an as-yet-unknown mechanism.

sgrS pitA and sgrS pstC mutants exhibit similar but not identical growth defect suppression: the growth of the sgrS pstC mutant in the presence of αMG is better than that of the sgrS parent but appears to be not quite as robust as growth of the sgrS pitA mutant (Fig. 6). While it is possible that another, as-yet-unidentified factor could account for this difference, it seems unlikely for several reasons. First, we ruled out other, more obvious potential causes of sgrS pitA suppression described above (e.g., decreased EIICB^Glc transport activity). Second, the two independent mutants both exhibit Pho regulon induction and partially suppress the sgrS growth defect to similar levels. Furthermore, for both sgrS pitA and sgrS pstC, suppression could not be attributed to an individual member of the Pho regulon (Table 3), which is also consistent with all or part of the regulon being required for suppression. Finally, sgrS pitA and sgrS pstC mutants exhibit different levels of Pho induction, which could conceivably account for the slight difference in suppression. PhoA activity in the sgrS pstC mutant is approximately 200- to 400-fold that of the wild type, while that of the sgrS pitA mutant is 3- to 6-fold (Fig. 5), indicating that a moderate level of Pho expression could be most advantageous to recovery from stress. Indeed, there is some evidence in at least one other strain of E. coli, avian pathogenic O78 strain χ7122, to support the notion that intermediate levels of Pho regulon induction are optimal for growth under certain conditions. phoB was found to be expressed by strain χ7122 during the course of infection (12). However, in another study, the virulence of χ7122 mutants was found to decrease as Pho induction was increased, implying that moderate regulation of the level of Pho induction is important for the ability to cause disease (7, 9). Regardless of the reason for the slight difference in the level of suppression between sgrS pitA and sgrS pstC mutants, it is clear that direct induction of the Pho regulon in the sgrS pstC mutant helps sgrS mutants, which are normally unable to recover from glucose-phosphate stress, to at least partially overcome stress-associated growth inhibition, and turning off Pho expression in sgrS pstC phoB results in loss of this suppression (Fig. 6).

What is less clear is why Pho regulon induction relieves stress. The pleiotropic nature of the physiological changes caused by Pho induction makes it difficult to ascribe cause of suppression. In addition, PhoA activity is not induced in wild-type E. coli in response to glucose-phosphate stress caused by αMG (Fig. 4), suggesting that the Pho regulon may not be required as a “normal” part of the stress response. Moreover, cells experience glucose-phosphate stress regardless of external P_i concentration; while LB and M63 media contain high P_i levels, stress also occurs in defined media with low P_i concentrations (data not shown). Thus, changes in P_i acquisition per se are unlikely to be responsible for suppression. Nevertheless, it may not be surprising that Pho induction aids in recovery from glucose-phosphate stress, given that the Pho regulon has been shown to play an integral role in the responses to other stresses in E. coli, such as oxidation, acid stress, and pathogenesis (see the recent review by Crépin et al. [9]).

While the mechanism of suppression of the sgrS glucose-phosphate stress defect is not immediately apparent, it is possible that induction of the Pho regulon confers a general growth advantage to the sgrS mutant during stress; indeed, the fact that no individual Pho genes were found to be responsible for suppression (Table 3) supports this notion. (However, given that transcriptome studies have revealed uncharacterized, putative members of the Pho regulon [4], we cannot rule out the possibility that an as-yet-unknown Pho member is responsible for suppression.) Such a growth advantage could perhaps be due, for example, to alteration of available intracellular phosphate pools such as ATP or polyphosphate (31, 41, 62), enhancing uptake of a particular nutrient or alleviating a metabolic imbalance that occurs during glucose-phosphate stress. Improved growth during stress could also be caused by the slight increase in EIICB^Glc (PtsG) transporter activity, which was observed in pitA and sgrS pitA mutants (Fig. 3). Although the underlying cause of enhanced EIICB^Glc activity is unclear, it is likely influenced by availability of phosphate pools. Certainly, the phosphorylation state of EIICB^Glc, which is determined by the PTS phosphorelay cascade (52), directly impacts the ability to take up substrates. The first member of the PTS cascade is phosphoenolpyruvate (PEP), and the ratio of PEP to pyruvate in the cell is thought to be a major determinant of the EIIA^Glc phosphorylation state and thus the activity of EIICB^Glc (23). Interestingly, the ratio of PEP to pyruvate may also play a role in glucose-phosphate stress. While the cause of stress is unknown, available evidence suggests that it may be due to depletion of glycolytic intermediates that occurs during accumulation of sugar-phosphates (30, 38, 46, 54). In addition, one type of sugar-phosphate stress has been shown to inhibit growth at least in part by depleting cellular PEP pools (32). The fact that SgrR regulates AlaC (56), a glutamic-pyruvic transaminase capable of generating pyruvate (28), may reflect the cell's effort to modulate PEP/pyruvate ratios. Another type of sugar-phosphate stress that occurs when cells accumulate UDP-galactose is also characterized by a metabolic imbalance (16, 34). It was recently shown that growth inhibition of galE mutant strains in the presence of galactose (with high intracellular UDP-galactose levels) was not caused by phosphosugar accumulation but rather by depletion of UTP and subsequently CTP (34). The growth inhibition was mostly reversed by supplementation with exogenous pyrimidines. It is reasonable to hypothesize that an analogous rescue of metabolic imbalance could be achieved genetically. Thus, pitA and pstC mutations and subsequent induction of the Pho regulon may be altering metabolism sufficiently to alleviate the presumed metabolic imbalance experienced by stressed sgrS mutants, and future research will explore this possibility.

Other studies have reported intersections between phosphate and carbon metabolism pathways. Expression of at least one member of the Pho regulon, psiE, is repressed by the cyclic AMP (cAMP)-CRP (cAMP receptor protein) complex under low glucose conditions (29, 61). cAMP-CRP was also recently shown to indirectly activate expression of sgrS during glucose-phosphate stress by increasing activity of SgrR (51). In the absence of phoR, phoB expression is activated by acetyl-phosphate (64) and by the catabolite regulatory kinase CreC in response to glucose (2, 63). Another connection between phosphate and carbon metabolism is the transcriptional regulator KdgR. KdgR negatively affects the expression of both sgrS (again by altering SgrR activity during glucose-phosphate stress) (51) and Pho regulon gene eda, which encodes an aldolase that converts 2-keto-3-deoxy-6-phosphogluconate to triose-3-phosphate and pyruvate (42). This tantalizing overlap and the connections between the Pho regulon and glucose-phosphate stress reported here underscore the idea that “no stress response is an island.” Similar future studies examining factors, such as the effect of PEP and pyruvate metabolism on stress, are likely to identify other physiological connections, which could help reveal the biological role(s) of the glucose-phosphate stress response.