The Regulatory C-Terminal Domain of Subunit ε of F_oF₁ ATP Synthase Is Dispensable for Growth and Survival of Escherichia coli^,

Abstract

The C-terminal domain of subunit ε of the bacterial F_oF₁ ATP synthase is reported to be an intrinsic inhibitor of ATP synthesis/hydrolysis activity in vitro, preventing wasteful hydrolysis of ATP under low-energy conditions. Mutants defective in this regulatory domain exhibited no significant difference in growth rate, molar growth yield, membrane potential, or intracellular ATP concentration under a wide range of growth conditions and stressors compared to wild-type cells, suggesting this inhibitory domain is dispensable for growth and survival of Escherichia coli.

F_oF₁ ATP synthases are ubiquitous enzymes that synthesize ATP using a transmembrane electrochemical potential of protons or proton motive force (PMF) generated by the respiratory chain across the cytoplasmic membrane of bacteria, the thylakoid membrane of chloroplasts, or the mitochondrial inner membrane (4, 5, 37). The enzyme consists of two parts: membrane-embedded F_o subcomplex (a complex of subunits a, b, and c in bacteria) and hydrophilic F₁ subcomplex (composed of subunits α, β, γ, δ, and ε). The enzyme is also known as a molecular motor, which is composed of the stator subcomplex (α, β, δ, a, and b) and the rotor subcomplex (γ, ε, and c), and its rotation is coupled to ATP synthesis and proton flow across the membrane (20, 31, 52). The reaction of the enzyme is reversible; ATP is hydrolyzed into ADP and inorganic phosphate, the rotor subcomplex rotates in reverse, and protons are extruded to the periplasmic side, resulting in the generation of PMF. Although some bacteria utilize the reverse reaction under particular conditions, the primary function of F_oF₁ ATP synthase is generation of ATP from the PMF. Therefore, the direction of the activity of F_oF₁ ATP synthase is regulated to avoid wasteful ATP hydrolysis.

Subunit ε in bacterial F_oF₁ has been known to be an intrinsic inhibitor of F₁ and F_oF₁ complex (18, 21, 23) and is proposed to have a regulatory function (10, 11, 42). Although the inhibitory effects of subunit ε vary among species, in general, ε inhibits ATP hydrolysis activity while repressing ATP synthesis activity to a lesser degree (14, 27). This regulatory function of the ε subunit is mediated almost exclusively by the C-terminal region of ε, which is comprised of two antiparallel α-helices (18, 49, 50). Biochemical and crystallographic studies have revealed that the C-terminal helices can adopt two different conformations (34, 46, 47, 48). In the retracted conformation, the α-helices form a hairpin-like structure and sit on the N-terminal β-sandwich domain of the ε subunit. When the ε subunit exhibits an inhibitory effect, it adopts a more extended conformation in which the C-terminal α-helices extend along the γ subunit, which composes the central stalk. It has also been shown that basic, positively charged residues on the second α-helix of the ε subunit interact with negatively charged residues in the DELSEED segment of subunit β to exert the inhibitory effect (12).

Escherichia coli mutants deleted in the entire ε subunit exhibit a reduced growth rate and growth yield, and this effect is proposed to be a result of a deficiency in assembly of the F_o and F₁ complexes (21). The N-terminal β-sandwich domain of the ε subunit is responsible for the assembly of F_o and F₁ and is therefore important for efficient coupling between proton translocation through F_o and ATP synthesis/hydrolysis in F₁ (15, 39). Deletion of the ε subunit leads to dissociation of the F_oF₁ complex and wasteful ATP hydrolysis by free (cytoplasmic) F₁ and dissipation of PMF through free F_o (21, 22, 51).

While the importance of the entire ε subunit in the whole-cell physiology of E. coli is fairly well established, the role of the regulatory C-terminal region of ε has received little attention and warrants investigation to determine if the regulatory functions (e.g., inhibition of ATP hydrolysis) observed in vitro are manifested in the physiology of E. coli under various growth conditions. To address this question, we constructed isogenic E. coli mutants that were deleted in the C-terminal region of ε subunit (ε^DC) and used these strains to compare physiological properties of wild-type versus ε^DC cells under a wide range of environmental conditions and stressors.

The C-terminal domain of the ε subunit regulates ATPase activity in inverted membrane vesicles of E. coli.

To elucidate the proposed regulatory function of the C-terminal helices of subunit ε in the whole-cell physiology of E. coli, chromosomal unmarked and marked (kanamycin resistance cassette) nonpolar deletions of the C-terminal domain of the ε subunit were constructed (Table 1). All chromosomal mutants were constructed using pWM97, as previously described (29). To construct the ε^DC mutant lacking the C-terminal region of the ε subunit (deleted from residue 94 to 139), designated strain NEMG1 (atpC′ Δ94-139) (Table 1), the following primer pairs were used to amplify (wild-type DNA) a region 700 bp upstream of the atpC start codon (viz. 5′-TTTGAGCTCTGACAACATCTATCGTTACACCCTGGCCGG-3′, SacI) and 279 bp into the coding region of atpC (viz. 5′-TTTGCTAGCTTACGCTTCGTCGAGATCCTGGCCGCGAATTGCGG-3′, including the stop codon and an NheI restriction site). This produced an amplicon of 978 bp (fragment 1). A second PCR product (fragment 2, 721 bp) was generated that includes the stop codon of atpC (viz. 5′-TTTGCTAGCTAACACCGGCTTGAAAAGCACAAAAGCCAGTCTGG-3′ and an NheI site) and 700 bp downstream of this stop codon (viz. 5′-TTTTCTAGAACGCTGGAGTGTTTCGACAGAGATCAGCGGC-3′, including an XbaI restriction site). The two products were ligated (fragment 3) by NheI digestion and cloned into the SacI-XbaI site of pWM97 to generate pNT01 (Table 1). To generate the ε mutant lacking the C-terminal region of the ε subunit using a kanamycin resistance (Km^r) cassette, designated strain NEMG2 [atpC[prime] (94-139)::Km^r] (ε^DC::Km) (Table 1), the Km^r cassette was cloned from pUC4K and inserted into the NheI site of fragment 3 and then cloned into the SacI-XbaI site of pWM97 to generate pNT02 (Table 1). All plasmid constructs were confirmed by DNA sequencing. pWM97 and its derivatives were maintained in JM109(λ-pir) because the replicon of this plasmid, oriR_R6Kγ, requires the Π protein for replication (30). The resultant pWM97 derivatives (pNT01 and pNT02) were electroporated into EMG2 cells (wild type), which are unable to maintain pWM97. After incubation of the transformants in Luria-Bertani (LB) medium, merodiploids in which the plasmid was integrated into the chromosome were isolated by selecting for ampicillin resistance. Lethality in LB with 5% sucrose (the sacB gene) was confirmed. The merodiploid was incubated in LB without ampicillin and sucrose to induce the second homologous recombination event, in which the integrated plasmid region is recombined out. The recombinant was selected by resistance to 5% sucrose and ampicillin sensitivity. All mutants were confirmed by PCR and Southern hybridization. To construct the complementing atpC construct (pNT03) (Table 1), DNA fragments containing the promoter of the atp operon (Patp), to drive atpC expression from the native promoter, and the atpC structural gene and Shine-Dalgarno box were amplified from E. coli EMG2 chromosomal DNA by PCR using the following primer pairs: atpIF (5′-TTTGAATTCTCTGTGTCGCTCGTGAGTCGAAACGTTGC-3′) (EcoRI) and atpIR (5′-TTTCTCGAGTTACCCTTTGTTGTTAATTACAGCCGG-3′) (XhoI); atpCF (5′-TTTCTCGAGCGCCTTAATCGGAGGGTGATATGGC-3′) (XhoI) and atpCR (5′-TTTGTCGACTTACATCGCTTTTTTGGTCAACTCGATAACGCGC-3′) (SalI). The fragments were ligated at the XhoI site and cloned into the EcoRI-SalI site of pHSG575 (Table 1). The construct pNT03 (Patp-atpC) was confirmed by DNA sequencing and transformed into strain NEMG1 (ε^DC) to generate NEMG3 (AtpC⁺) (Table 1).

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Genotypea	Source	Reference
EMG2 (εWT)	Wild type	CGCSb	2
NEMG1 (εDC)	EMG2 atpC′ Δ94–139	This study	This study
NEMG2 (εDC::Km)	EMG2 atpC′ (94–139)::Kmr	This study	This study
NEMG3 (AtpC+)	NEMG1 (εDC) harboring pNT03 Patp-atpC, Cmr	This study	This study
JM109(λ-pir)	recA1 endA1 gyrA96 thi hsdR17 supE44 relA1 Δ(lac-proAB) F′[traD36 proAB lacIqlacZΔM15] λ-pir	Lab stock	33
pWM97	oriRR6KγoriTRP4, Ampr, sacB lacZα	NBRPc	29
pNT01	pWM97 harboring atpC′ Δ94–139	This study	This study
pNT02	pWM97 harboring atpC′ (94–139)::Kmr	This study	This study
pHSG575	oripSC101, Cmr, lacZα Plac	Lab stock	44
pNT575	pHSG575, ΔPlac	This study	This study
pNT03	pNT575, Patp-atpC	This study	This study
pUC4K	oripBR322, Ampr Kmr(Tn903), lacZα	Lab stock	45

^aKm^r, kanamycin resistance; Amp^r, ampicillin resistance; Cm^r, chloramphenicol resistance; Plac, promoter of lac operon; Patp, promoter of atp operon.

^bCGCS, The Coli Genetic Stock Center, Yale University.

^cNBRP, National BioResource Project, National Institute of Genetics, Japan.

The properties of the E. coli ε^DC strain were compared to those of its isogenic wild-type parent strain (ε^WT) (Fig. 1). Strains were grown in LB broth (36), and membrane vesicles of E. coli were prepared as described by Suzuki et al. (40, 43). ATP hydrolysis activity was measured using three different assays: an ATP regeneration assay detecting liberation of ADP, detection of inorganic phosphate, and detection of protons using phenol red based on previously published assays (11, 16, 41, 42). The reactions were initiated by the addition of 3 μg membrane vesicles into 1 ml of assay mixture. Where indicated, lauryldimethylamine oxide (LDAO) was used at a final concentration of 0.3%. Inverted membrane vesicles of the ε^DC mutant exhibited 1.3-fold-higher levels of ATPase activity than those of the ε^WT strain with increasing concentrations of ATP (Fig. 1A). LDAO, a nonionic detergent, has been shown to activate F₁ and F_oF₁ and to prevent interaction of the ε subunit with the β subunit, leading to a disappearance of the inhibitory effect of the ε subunit (9, 25). When LDAO was included in the ATPase assay, both ε^DC and ε^WT were activated by LDAO, but the activities of ATP hydrolysis were comparable (Fig. 1A), suggesting that the ATP hydrolysis activity of the α₃β₃γ complex is not impaired in the ε^DC strain.

Fig. 1.

Enzymatic activities of ε^WT and ε^DC in inverted membrane vesicles of E. coli. (A) ATP hydrolysis activities of ε^WT (closed circles and closed triangles [0.3% LDAO]) and ε^DC (open circles and open triangles [0.3% LDAO]) strains as a function of ATP concentration measured at 37°C using the ATP-regenerating assay as described in Materials and Methods. (B) Left panel: ATP hydrolysis activities of ε^WT and ε^DC strains, measured by the liberation of inorganic phosphate (NP assay) or liberation of ADP using the ATP-regenerating assay (PK assay). Where indicated, ADP (NP assay) or P_i (PK assay) was added at 0.5 mM or 2 mM, respectively. Right panel: ATP hydrolysis activities of ε^WT and ε^DC strains, measured using the phenol red assay with or without the addition of 0.5 mM ADP or 2 mM P_i. (C) Effect of P_i on ATP-dependent proton translocation was determined at 37°C by the quenching of ACMA. (D) ATP synthesis of ε^WT and ε^DC strains at pH 8.0 and 37°C with 0.5 mg of inverted membrane vesicles preincubated for 2 min with 2.5 mM NADH with stirring before the reaction was initiated using 0.75 mM ADP and 2.5 mM P_i. Each point is the result of three experimental replicates, and the standard error associated with these measurements is shown.

Mg-ADP inhibition of F_oF₁ ATP hydrolysis activity is a well-described regulatory mechanism (13, 53). A combination of Mg-ADP and inorganic phosphate (P_i) has also been reported to inhibit the ATP hydrolysis activity of F_oF₁ (8). To determine what effect the ε subunit has on this inhibition by ADP and P_i, ATP hydrolysis activity in the presence of these compounds was measured in the ε^DC and ε^WT strains (Fig. 1B). Three different ATPase assays were used for these experiments; liberation of inorganic phosphate (P_i) in the presence of ADP (NP assay), liberation of ADP in the presence of inorganic phosphate (PK assay), and proton-pumping activity in the presence of ADP and P_i (phenol red). Using these ATP hydrolysis assays, the ε^DC strain displayed consistently higher rates of ATPase activity than the ε^WT strain, and this was unaffected by the presence of high concentrations of P_i (2 mM) (Fig. 1B). In contrast, ADP inhibited both ε^DC and ε^WT ATP hydrolysis activity (Fig. 1B). A third assay using phenol red as an indicator of ATPase activity (proton pumping) was used to determine what effect ADP and P_i had in combination on ATPase activity (Fig. 1B, right panel). Using this assay, the ε^DC strain displayed consistently higher rates of ATPase activity, which were not significantly affected by P_i, than did the ε^WT strain. ADP inhibited the ATP hydrolysis activity of both the ε^DC and ε^WT strains, and a combination of ADP and P_i had a synergistic inhibitory effect on the ε^WT strain but no greater effect than that of ADP alone for the ε^DC strain (Fig. 1B, right panel). These data demonstrate that ADP-P_i strongly inhibits ATP hydrolysis by the ε^WT strain but not the ε^DC strain, demonstrating that this inhibition requires the presence of the C-terminal helices of the ε subunit. P_i with ADP has been reported to modulate ATP hydrolysis activity and the coupling efficiency of F_oF₁ from E. coli (8). Although there has been no direct evidence that subunit ε binds P_i, its trypsin sensitivity changes depending on the presence of P_i (28).

Previous work has reported that an E. coli ε mutant defective in the C-terminal region of ε (i.e., truncated after Gln-87) was uncoupled in proton-pumping activity, suggesting a role for the ε subunit as a coupling factor (6). To address this hypothesis, ATP-dependent proton-pumping activities were measured in ε^DC and ε^WT inverted membrane vesicles using 9-amino-6-chloro-2-methoxyacridine (ACMA) quenching as described previously (40, 43). The composition of the reaction mixture was 10 mM HEPES-KOH (pH 7.5), 100 mM KCl, 5 mM MgCl₂, 300 ng/ml ACMA, 1 mM ATP. ACMA quenching was reversed by the addition of the uncoupler carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP; 1 μg/ml). The ATP-dependent proton-pumping activities were comparable between the ε^DC and ε^WT strains (Fig. 1C). When proton-pumping activity was measured in the presence of increasing inorganic phosphate concentrations, the proton-pumping activity of the ε^WT strain was inhibited to a greater extent than that of the ε^DC strain (Fig. 1C), consistent with the phenol red data in Fig. 1B. Our data are consistent with the notion that despite deletion of the C-terminal region of ε, the enzyme remains coupled for proton pumping. It should be noted that the mutant EMG2 atpC′ Δ94-139 used in our study is truncated from residue 94 of the ε subunit, versus the Gln-87-truncated ε mutant previously reported to be uncoupled (6).

To determine the role that the C-terminal region of the ε subunit plays in ATP synthesis activity driven by the PMF, ATP synthesis measurements were performed using inverted membrane vesicles energized with NADH (Fig. 1D). NADH-driven ATP synthesis activity was measured by using a luciferin-luciferase system as previously described (43). The reaction was carried out at 37°C. The activity that synthesized 1 μmol of ATP per min was defined as 1 unit. The rate of ATP synthesis was approximately 20% higher for the ε^DC strain than for the ε^WT strain (Fig. 1D). The explanation for the higher rate of ATP synthesis in the ε^DC strain is not clear, but the data suggest that the C-terminal domain of the ε subunit is not required for efficient coupling between the PMF and ATP synthesis.

Role of ε subunit in vivo: growth kinetics of the ε^DC strain versus those of the ε^WT strain in batch and continuous growth.

The in vitro assays suggested that the ε^DC F_oF₁ complex is moderately elevated in ATPase activity and is less sensitive to regulation by ADP-P_i. If the C-terminal domain of the ε subunit is playing a crucial role in regulating and/or coupling F_oF₁ ATP synthase activity in vivo, this may be manifested in the growth kinetics of E. coli at high and low growth rates. To investigate this hypothesis, we measured a range of energetic properties for the ε^DC and ε^WT strains growing in batch (fast growth, 25-min doubling time in LB medium) and carbon-limited continuous culture (slow growth, 6.9-h doubling time) in nutrient-rich LB medium and minimal medium (see Tables S1, S2, and S3 in the supplemental material).

When the ε^DC and ε^WT strains were grown in batch culture using LB medium, no significant differences were observed in the maximum specific growth rate, final optical density at 600 nm (OD₆₀₀), external pH, dry weight, membrane potential (Δψ), or intracellular ATP concentration (see Table S1 in the supplemental material). The same was true for cells grown in M9 minimal medium with the following carbon and energy sources: glucose and succinate (see Table S1). Results with acetate as the sole carbon and energy source were also comparable (data not shown). For the determination of Δψ, cells (three 1-ml samples) from exponential phase (batch cultures) and steady state (continuous culture) were taken directly from the flask or fermenter and added immediately to a glass tube containing [³H]methyltriphenylphosphonium iodide ([³H]TPP⁺)(30 to 60 Ci/mmol; 2.4 nM, final concentration). After incubation for 5 min at 37°C, the cultures (900 μl in triplicate) were filtered rapidly through a 0.45-μM cellulose-acetate filter (Sartorius) as adapted from Zilberstein et al. (54). The filters were washed twice with 2 ml of 100 mM LiCl and dried for 60 min at 40°C. Filters were resuspended in 2 ml of scintillation liquid, and counts per min (cpm) were determined using an LKB Wallac 1214 Rackbeta liquid scintillation counter (PerkinElmer Life Sciences). The intracellular volume (2.8 ± 0.5 μl mg of protein⁻¹) was determined previously (38). The ΔΨ across the cell membrane was calculated from the uptake of [³H]TPP⁺ according to the Nernst relationship. Nonspecific [³H]TPP⁺ binding was estimated from cells which had been treated with carbonyl cyanide m-chlorophenylhydrazone (CCCP) (250 μM) for 20 min prior to addition of [³H]TPP⁺. ATP was extracted from cells by perchloric acid and KOH-NaHCO₃ treatment after separation of the cells from the growth medium as previously described (24) and frozen at −70°C. Prior to analysis, samples were thawed and potassium perchlorate was removed by centrifugation (13,000 × g, 5 min, 4°C). The ATP concentration was determined by the luciferin-luciferase method (26). Light output was immediately measured with a luminometer (model FB12; EG&G Berthold) using ATP as a standard.

Under anaerobic conditions, there were no significant differences between the ε^DC and ε^WT strains grown on either glucose (fermentative conditions) or glycerol-nitrate (anaerobic respiration) (see Table S2 in the supplemental material). For anaerobic batch culture, rubber-sealed Hungate test tubes with 10 ml of growth medium were used under a nitrogen headspace (O₂ free). For anaerobic growth with 40 mM glycerol and 20 mM nitrate, 1 μM Na₂MoO₄ and 1 μM Na₂SeO₃ were added to M9 minimal medium (36) with additional iron (10 μM FeSO₄), and methylene blue (1.5 μg/ml) was used as an indicator of anaerobiosis. Anaerobic cultures were inoculated (0.1%) using a hypodermic syringe and grown at 37°C without shaking.

During batch growth, the rate of substrate consumption is often greater than that required for growth, i.e., the catabolic and anabolic rates are not precisely matched and energy spilling mechanisms may occur (35). However, in carbon-limited continuous culture at low growth rates, the catabolic and anabolic rates are more tightly coupled (7, 35), and therefore factors that might result in energy wasting/spilling (e.g., unregulated ATP hydrolysis) will be greater and reflected in reduced cell yields. For carbon-limited continuous-culture experiments, a 250-ml culture vessel with a working volume of 150 ml of M9 minimal medium was used (36) with an appropriate carbon source (i.e., glucose, succinate, or acetate, each at a 0.2% final concentration). All carbon sources were confirmed to be limiting for growth at the concentrations used in this study. After inoculation (0.1% inoculum), the culture was left in batch mode, maintaining 50% air saturation, until the OD₆₀₀ reached 80% of the steady-state OD₆₀₀ and then switched to continuous mode at a dilution rate of 0.10 h⁻¹, corresponding to a doubling time of 6.9 h. The chemostat was then left running for at least four volume changes before analysis of energetic parameters (membrane potential, ATP content, and substrate utilization) and cell growth (dry weight) was performed. To monitor culture growth, samples were withdrawn aseptically when required, and the OD₆₀₀ was measured (1-cm light path length). To calculate the molar growth yield (g [dry weight] of cells/g of substrate consumed) on glucose and succinate, cells (50 ml in triplicate) were harvested by filtration, and the filters (0.45 μm; Millipore) were dried at 105°C until a constant dry weight was achieved on consecutive days. Cell-free supernatants were assayed for succinate and glucose. Succinate and glucose were determined using the following detection kits: 10176281035 for succinate and 10716251035 for glucose (Roche).

When the ε^DC and ε^WT strains were grown in carbon-limited continuous culture at a dilution rate of 0.1 h⁻¹, the steady-state OD₆₀₀, cell dry weight, molar growth yield, membrane potential, and intracellular ATP were comparable between the two strains during growth on glucose or succinate (see Table S3 in the supplemental material). The results for acetate were also comparable (data not shown). These data indicate that prominent regulatory features of the ε subunit identified in vitro (e.g., inhibition of ATPase activity, uncoupling of proton translocation, and ATP synthesis) do not appear to play a significant role during growth of E. coli on a range of substrates (fermentable versus nonfermentable), aerobic or anaerobic growth, and high or low growth rate in minimal medium.

Role of the ε subunit during cellular stress.

During the facultative lifestyle of E. coli, various stressors (e.g., low pH, reactive oxygen species, and nutrient deprivation/stationary phase) are encountered, and intracellular ATP is essential for mounting an appropriate physiological response. Initial experiments focused on the growth rate and final OD₆₀₀ of the ε^DC and ε^WT strains grown under a number of environmental stresses and nutrient limitations (Table 2). The only significant difference observed between the ε^DC and ε^WT strains was during growth in low-salt medium (Table 2, 1 mM NaCl or 1 mM KCl). The doubling time of the ε^DC mutant was 50% slower than that of the ε^WT strain, and this effect was specific for Na⁺ or K⁺ ions (Table 2). No differences were noted in the MICs between the ε^DC and ε^WT strains for the following ionophores and protonophores: nigericin (K⁺/H⁺), monensin (Na⁺/H⁺), valinomycin (K⁺), CCCP (H⁺), and dinitrophenol (data not shown). We tested the effect of various salt concentrations (0 to 100 mM NaCl) on ATP hydrolysis activity of inverted membrane vesicles for the ε^DC and ε^WT strains and found no significant effect on ATPase activity in this concentration range (data not shown).

Table 2.

Effects of stress on growth of ε^WT and ε^DC strains

Stressora	Growth of strainb
	εWT strain		εDC strain
	μmax (h−1)	Final OD	μmax (h−1)	Final OD
Nutrient limitation
0.05% glucose	0.82	0.69	0.82	0.71
0.05% succinate	0.72	0.28	0.70	0.28
1 mM phosphate	0.56	0.88	0.56	0.91
2 mM nitrogen	0.51	0.85	0.51	0.88
1 mM NaCl	0.72	0.92	0.33	0.80
1 mM KCl	0.69	0.82	0.39	0.75
20 μM MgCl2	0.69	0.82	0.71	0.87
2 μM CaCl2	0.69	0.81	0.69	0.81
External pH
pH 5.0	0.17	0.68	0.13	0.66
pH 9.0	0.20	0.70	0.16	0.67
Temp
40°C	0.58	0.94	0.62	0.91
20°C	0.51	0.88	0.51	0.87
Additive
250 mM NaCl	0.78	0.91	0.83	0.88
250 mM KCl	0.19	0.57	0.19	0.62

^aAll experiments were performed in M9 succinate (0.2%) minimal medium, except where glucose was the nutrient limitation, under the experimental conditions listed.

^bAll values reported are the means of data from two to four experimental replicates, and the experimental error associated with these values was less than 15%. μ_max, maximum specific growth rate.

The cell viability characteristics of strains NEMG1 (ε^DC), EMG2 (ε^WT), and NEMG3 (complemented ε^DC) were studied during nutrient starvation in nongrowing cultures (Fig. 2A). Bacterial cell viability was monitored by cell counts based on CFU per ml of culture, where serial dilutions of bacterial cell culture in phosphate-buffered saline were plated on LB agar plates supplemented with appropriate antibiotics. When the cell viability of all three strains was monitored following 24 h of growth in succinate-containing minimal medium (stationary-phase cultures), no significant differences were noted in cell viability even after 28 days (Fig. 2A). To add greater stress on the cells, competition experiments between all strains were performed, and cell viability was monitored (Fig. 2B). For these experiments, we used the following combinations of strains to allow enumeration of individual strains through kanamycin resistance/sensitivity: ε^WT strain versus ε^DC::Km^r strain and NEMG1 (ε^DC) strain versus NEMG3 (pNT03-complemented ε^DC) strain. In minimal medium, the population of the ε^DC, ε^DC::Km^r, and ε^WT strains were comparable after 28 days, suggesting that the ε^DC and ε^DC::Km^r mutants were not compromised in fitness relative to the ε^WT strain. When the same experiment was performed at low salt concentrations (i.e., 1 mM NaCl), the ε^DC and ε^DC::Km^r mutants showed a 1-log decrease in cell survival compared with the ε^WT and complemented ε^DC (pNT01) strains (Fig. 2C), suggesting an inability to compete with the ε^WT strain under these conditions.

Fig. 2.

Cell viability of ε^WT and ε^DC strains in stationary-phase batch cultures. (A) Viability of EMG2 (ε^WT), NEMG1 (ε^DC), and NEMG3 (ε^DC + pNT03 AtpC⁺) cells as monocultures in M9 succinate (0.2%) minimal medium at 37°C after 24 h of growth (stationary phase). (B and C) Viability of cocultures EMG2 (ε^WT) versus NEMG2 (ε^DC::km) and NEMG1 (ε^DC) versus NEMG3 (ε^DC + pNT03 AtpC⁺) in M9-succinate medium (B) or in M9-succinate minimal medium containing 1 mM NaCl (C). Cultures were mixed in mid-log phase and grown for a further 24 h (stationary phase), and cell viability was determined after 28 days at 37°C. For all experiments, cellular viability was determined by measuring CFU by plating out appropriate dilutions on LB agar, supplemented with kanamycin where indicated, and enumerating them after 1 day of growth at 37°C. Results are shown as the means ± SEM for technical replicates of each time point.

Conclusions.

A large number of studies have reported that the ε subunit has a significant role in F_oF₁ assembly, coupling, and regulation of enzyme activity. The C-terminal domain of the ε subunit is proposed to act as a regulator and/or coupling factor of ATPase activity, but no studies have been performed to determine if these regulatory mechanisms operate in vivo and how important the ε subunit is in the physiology of E. coli. No significant differences were noted in growth rate, molar growth yield, membrane potential, or intracellular ATP concentration between the ε^WT and ε^DC strains under a wide range of growth conditions tested and when cells were subjected to various stressors (e.g., nutrient limitation, low and high pH, and temperature). One exception was the impaired survival and growth of the ε^DC strain in low-salt (i.e., Na⁺ or K⁺ ions) medium. The physiological explanation for this remains unclear, but intracellular concentrations of Na⁺ and K⁺ ions are precisely controlled by various antiporters (e.g., Na⁺/H⁺ and K⁺/H⁺) (3, 32) and the K⁺-translocating Kdp-ATPase (1), and maintenance of such gradients at low salt levels may perturb energy generation via the ATP synthase, particularly in an ε^DC mutant.

The general lack of ε-subunit-mediated regulation of ATP synthase/ATPase activity in vivo may reflect the extraordinary metabolic flexibility (e.g., aerobic and anaerobic respiration and fermentation) of E. coli. Subunit ε of E. coli has been shown to bind ATP, but the affinity is low (dissociation constant [K_d] = 22 mM) (50). This is in contrast to obligate aerobes like Bacillus species, where the K_d is in the range of 0.67 to 2.2 mM (17, 19). Experiments have not been performed to determine a physiological role for subunit ε in the growth of Bacillus species and other obligate aerobes, and this is the focus of ongoing studies.