Pyruvate and Lactate Metabolism by Shewanella oneidensis MR-1 under Fermentation, Oxygen Limitation, and Fumarate Respiration Conditions^,

Abstract

Shewanella oneidensis MR-1 is a facultative anaerobe that derives energy by coupling organic matter oxidation to the reduction of a wide range of electron acceptors. Here, we quantitatively assessed the lactate and pyruvate metabolism of MR-1 under three distinct conditions: electron acceptor-limited growth on lactate with O₂, lactate with fumarate, and pyruvate fermentation. The latter does not support growth but provides energy for cell survival. Using physiological and genetic approaches combined with flux balance analysis, we showed that the proportion of ATP produced by substrate-level phosphorylation varied from 33% to 72.5% of that needed for growth depending on the electron acceptor nature and availability. While being indispensable for growth, the respiration of fumarate does not contribute significantly to ATP generation and likely serves to remove formate, a product of pyruvate formate-lyase-catalyzed pyruvate disproportionation. Under both tested respiratory conditions, S. oneidensis MR-1 carried out incomplete substrate oxidation, whereby the tricarboxylic acid (TCA) cycle did not contribute significantly. Pyruvate dehydrogenase was not involved in lactate metabolism under conditions of O₂ limitation but was required for anaerobic growth, likely by supplying reducing equivalents for biosynthesis. The results suggest that pyruvate fermentation by S. oneidensis MR-1 cells represents a combination of substrate-level phosphorylation and respiration, where pyruvate serves as an electron donor and an electron acceptor. Pyruvate reduction to lactate at the expense of formate oxidation is catalyzed by a recently described new type of oxidative NAD(P)H-independent d-lactate dehydrogenase (Dld-II). The results further indicate that pyruvate reduction coupled to formate oxidation may be accompanied by the generation of proton motive force.

INTRODUCTION

Shewanella oneidensis MR-1 is a facultatively anaerobic, Gram-negative bacterium that generates energy by coupling the oxidation of organic compounds to the reduction of a wide range of electron acceptors, including O₂, fumarate, and Fe(III) (15, 17). The diverse metabolic capabilities of Shewanella species provide a competitive advantage in a range of environments that are subject to spatial and temporal variations in the type and concentration of electron acceptors (for a review, see reference 7). Considered strictly respiratory organisms, oxidative phosphorylation is thought to be the primary pathway for ATP synthesis in shewanellae (18, 29). A recent report, however, demonstrated that S. oneidensis MR-1 could gain energy for survival by fermenting pyruvate (15). Although the physiological significance of this process in S. oneidensis MR-1 is yet to be understood, fermentative metabolism in obligatory respiratory bacteria may represent an important mechanism of survival in the absence of available electron acceptors. Long-term survival via pyruvate fermentation was also reported previously for Pseudomonas aeruginosa (6).

A common metabolic trait displayed by shewanellae is the anaerobic production of acetate when grown on lactate or pyruvate as the sole source of carbon and energy (13, 15, 18, 22), suggesting that some ATP can be produced by substrate-level phosphorylation through the phosphotransacetylase-acetate kinase (Pta-AckA) pathway (28). A recent publication supported the involvement of this pathway in lactate metabolism under anaerobic conditions, as Δack and Δpta mutants of S. oneidensis MR-1 lost the ability to grow with fumarate or Fe(III) citrate as the electron acceptor, whereas an inactivation of the F_oF₁ ATP synthase operon resulted in only a minor growth defect (10). These findings raise an important question regarding the anaerobic energy metabolism of S. oneidensis MR-1, given the essentiality of substrate-level ATP production and the dispensability of oxidative phosphorylation: why is pyruvate fermentation not linked to the organism's growth? Related to that is the question of acetate metabolism in Shewanella, which is utilized only under fully aerobic conditions where O₂ is not growth limiting. Since Shewanella species have a complete tricarboxylic acid (TCA) cycle and are able to oxidize acetate to CO₂ under aerobic conditions (20), it is not clear what underlies their inability to use this substrate anaerobically.

To answer these questions, we quantitatively analyzed the energy conservation and central carbon metabolism of S. oneidensis MR-1 under fermentative, respiratory anaerobic, and respiratory aerobic (O₂-limited) conditions.

MATERIALS AND METHODS

Bacterial strains and growth media.

The strains of S. oneidensis MR-1 used in this study are listed in Table S1 in the supplemental material. Wild-type and mutant strains were routinely cultured at 30°C in tryptic soy broth (TSB) (pH 7.4) (25). Modified M1 medium (11) (designated PFM) with the following composition was used to conduct pyruvate fermentation experiments: 25 to 30 mM sodium pyruvate, 28 mM NH₄Cl, 25 mM NaH₂PO₄·H₂O, 30 mM NaCl, 1 mM MgCl₂·6H₂O, 1.34 mM KCl, 6.8 μM CaCl₂, 1 μM Na₂SeO₄, and 10 ml each of 10× Wolfe's vitamin solution and 10× mineral solution (the pH was adjusted to 7.0 with HCl and NaOH). Anaerobiosis was achieved by extensively purging the medium with O₂-free N₂. The inocula for experiments were generally grown as aerobic batch cultures in TSB at 30°C on a rotary shaker operated at 80 rpm, which were then centrifuged (6,000 × g for 7 min at room temperature), washed once, and resuspended in PFM. An inoculum of the ΔsucB strain was grown in anaerobic TSB supplemented with 30 mM fumarate and 20 mM lactate. Controlled batch and chemostat experiments were conducted by using a modified M1 medium described previously (20). To grow S. oneidensis MR-1 under anaerobic conditions, NaCl was excluded, and sodium fumarate was added as an electron acceptor to final concentrations of 35 mM. Lactate was used as a carbon source at concentrations of 90 mM (controlled batches), 22 mM (O₂-limited chemostats), and 30 mM (fumarate-limited anaerobic chemostats [fumarate was added to a final concentration of 35 mM]). Uncontrolled anaerobic batch cultivation was conducted in the same medium with an elevated concentration of PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)] (30 mM).

Pyruvate fermentation experiments.

All experiments were conducted in anaerobic Balch tubes (26-ml total volume containing 10 ml of PFM) stoppered with butyl rubber stoppers and crimp sealed with aluminum caps. At each sampling, four to six tubes were sacrificed for further analysis. When CO₂ quantification was performed, 0.2 ml of 12.1 M HCl was added to each tube, and tubes were vigorously shaken for 1 min prior to gas chromatography (GC) analysis. At least two independent experiments were done for each tested strain and condition.

Controlled batch experiments.

Controlled batch experiments were performed with 6-liter Bioflo 3000 reactors (New Brunswick Scientific, Edison, NJ) with 3 liters of medium supplemented with 90 mM d,l-lactate at 30°C. Agitation was maintained at 300 rpm, the gas flow rate was maintained at 3 liters/min, and the pH was maintained at 7.0 by the automatic addition of 2 M HCl. Batches were started by the inoculation of bioreactors with 1 ml/liter of a culture grown overnight in TSB and maintained at a dissolved oxygen tension (DOT) of 20% air saturation by automatically changing the ratio of N₂ and air in the gas mix. When cultures reached an optical density at 600 nm (OD₆₀₀) of 0.295 to 0.330, DOT control was turned off, and bioreactors were sparged with a gas mix of N₂ and air (91.5% and 8.5%, respectively). The time point when the DOT reached 0 was considered the start of the O₂-limited phase. Periodically, culture samples (between 5 and 10 ml) were withdrawn for analyses of optical density and organic acids. For experiments with the aceE (pyruvate dehydrogenase [PDH] E1 subunit) deletion mutant, cells were grown in TSB overnight, washed twice with modified M1 medium, resuspended in the same medium, and added to a final OD₆₀₀ of 0.32 to the bioreactor purged with 91.5% N₂ and 8.5% air.

Chemostat cultivation.

Anaerobic chemostat experiments were performed by using 6-liter Bioflo 3000 reactors operated at a 3-liter working volume at 30°C. Reactors were inoculated with 3 ml of S. oneidensis MR-1 cultures grown overnight in TSB and were maintained in a batch mode until the late logarithmic phase. The continuous mode was initiated and maintained at a dilution rate of 0.1 h⁻¹. The pure N₂ flow rate and agitation were kept at 2 liters/min and 250 rpm, respectively. The pH was maintained at 7.0 by the automatic addition of 2 M HCl.

A 15-liter New Brunswick Bioflo 3000 reactor operated at a 7-liter working volume was used to grow O₂-limited chemostat cultures of S. oneidensis MR-1 at a dilution rate of 0.1 h⁻¹. When aerobic steady state was achieved at a 20% DOT (20), the gas mix was changed to 8.5% air and 91.5% pure N₂, at which point the DOT became zero. The O₂-limited process was maintained for the time necessary for the change of at least 5 bioreactor volumes after steady-state conditions were attained. At this point, bioreactors were sampled for final analysis as described previously (20).

Analytical methods.

Organic acids in cultures were analyzed as described previously (19). The evolution of H₂ and CO₂ was measured by use of an HP 5890 series II gas chromatograph with a thermal conductivity detector and a Supelco 1-2392U 60/80 Carboxen 1000 column operated with ultrapure N₂ as a carrier gas at an oven temperature of 180°C to 235°C. Gas samples were removed by using a 1-ml gas-tight locking syringe (SGE Analytical Science Pty. Ltd.). Ultrapure (>99.999%) H₂ and CO₂ gases were used to prepare standards. Headspace and dissolved-in-culture gas concentrations were calculated by using Henry's law constants. OD₆₀₀ measurements were converted to ash-free dry weight (AFDW) biomass using a previously established correlation (20). Enzyme activities in cell extracts were measured according to previously reported protocols (20).

RESULTS

Reconstruction and validation of the pyruvate fermentation pathway.

S. oneidensis MR-1 cells metabolized pyruvate to lactate, acetate, formate, H₂, and CO₂ for over 690 h (Table 1). At 97% carbon recovery, the reaction stoichiometry was that of the amount of pyruvate metabolized equaled to the sum of lactate and acetate produced, indicating that pyruvate was not used for the de novo biosynthesis of cell components, consistent with the absence of growth. Based on the product stoichiometry, previously reported experimental data, and the S. oneidensis MR-1 genome, we have assembled a putative pathway that consists of six enzymatic reactions potentially responsible for the observed metabolic conversions (Fig. 1). In this scheme, pyruvate is oxidized by the pyruvate dehydrogenase (PDH) complex (aceEFG [SO_0424 to SO_0426]) and/or pyruvate formate-lyase (pflB [SO_2912]), resulting in the formation of acetyl coenzyme A (acetyl-CoA) and CO₂ or formate, respectively. The transformation of acetyl-CoA to acetate is coupled to substrate-level ATP production and catalyzed by consecutive acetyltransferase (pta [SO_2916]) and acetate kinase (ackA [SO_2915]) reactions (see reference 28). Formate is converted to H₂ and CO₂ by the combined action of putative formate dehydrogenase and hydrogenase enzymes, for which multiple copies have been identified in the S. oneidensis MR-1 genome (fdnGHI [SO_0101 to SO_0103], fdhA1B1C1 [SO_4508 to SO_4511], fdhA2B2C2 [SO_4512 to SO_4515], hydAB [SO_3920 and SO_3921], and hyaAB [SO_2098 and SO_2099]). hyaAB encodes an [Ni-Fe] hydrogenase (HyaAB), the main enzyme responsible for H₂ production by MR-1 cells fermenting pyruvate (15). Finally, the action of the fermentative lactate dehydrogenase (LDH) encoded by ldhA (SO_0968), which demonstrated NADH-dependent pyruvate reductase activity in vitro (21), is likely to be responsible for the pyruvate conversion to lactate.

Table 1.

Fermentation products and carbon recovery during fermentative pyruvate metabolism by S. oneidensis MR-1

Fermentation product	Value for product at time (h) of:					Carbon recovery (mM)b
	48	168	325	491	692
Pyruvate (mM)a	4.72	8.43	12.4	15.58	17.31	51.93
Lactate (mM)	0.54	1.4	2.8	4.22	5.13	15.39
Acetate (mM)	4.08	6.93	9.74	11.2	12.1	24.2
Formate, mM	1.91	2.61	1.9	1.01	0.39	0.39
H2 (mmol/liter culture)	0.8	1.92	2.6	2.82	3.17	NAc
CO2 (mmol/liter culture)	NA	NA	NA	NA	10.4	10.4
∑acetate + lactate	4.62	8.33	12.54	15.42	17.23
∑carbon in by-products						50.38

^aPyruvate that was consumed.

^bCarbon recovery was calculated for 692 h.

^cNA, not available.

Fig. 1.

Metabolic pathways implicated in pyruvate fermentation by S. oneidensis MR-1. FDH stands for one or more formate dehydrogenases encoded in the MR-1 genome (SO_0101 to SO_0103, SO_4508 to SO_4511, and SO_4512 to SO_4515). LdhA (SO_0968), fermentative lactate dehydrogenase; PDH (SO_0424 to SO_0426), pyruvate dehydrogenase complex; PflB (SO_2912), pyruvate formate-lyase; Pta (SO_2916), phosphotransacetylase; AckA (SO_2915), acetate kinase; Hyd (SO_3920 and SO_3921) and Hya (SO_2098 and SO_2099), [Fe-Fe] and [Ni-Fe] hydrogenases, respectively. The scheme reflects the predicted localizations of the corresponding enzymes.

The predicted pathway was investigated by use of targeted mutagenesis, whereby S. oneidensis MR-1 knockout strains deficient in key pyruvate fermentation genes were evaluated for products of pyruvate metabolism relative to wild-type MR-1. The putative involvement of PDH and pyruvate formate-lyase (PflB) in anaerobic pyruvate utilization was assayed with ΔaceE and ΔpflB mutants, respectively. As shown in Table 2, the ΔpflB mutant strain failed to produce formate and H₂ and had significantly lower pyruvate consumption and acetate production rates than those of the wild type, whereas the ΔaceE strain retained the same end product profile and production rates as those of the wild type (data not shown). A ΔpflB ΔaceE double mutant strain failed to ferment pyruvate, indicating that the S. oneidensis MR-1 genome does not encode any other enzymes able to catalyze pyruvate degradation under the experimental conditions used.

Table 2.

Specific rates of pyruvate consumption and by-product generation by the wild type and selected mutants of S. oneidensis MR-1

By-product	Rate of production (μmol/g AFDW·h) for strain
	Wild type	ΔhydA ΔhyaB	ΔpflB	Δdld-II
Pyruvate	988.8	1,533.8	377.4	3,698.7
Acetate	926.6	1,189.6	200.8	2,904.6
Lactate	173.5	262.7	80.9	76.6
Formate	556.3	794.1	NDa	2,062.2
H2	64.5	ND	ND	226.4

^aND, not detected.

As expected, the ΔackA strain was severely impaired in pyruvate metabolism under anaerobic conditions and produced negligible amounts of acetate and lactate during the first 24 h (data not shown). Similarly, a mutant deficient in both periplasmic hydrogenases (ΔhydAB ΔhyaAB) did not produce H₂ but displayed higher rates of pyruvate consumption as well as higher lactate, acetate, and formate production rates (Table 2). Surprisingly, the deletion of ldhA did not impair lactate production in S. oneidensis MR-1, as the ΔldhA strain generated all the by-products at rates similar to those of the wild type (data not shown). In contrast, mutants with the pflB deletion displayed a significant (>2-fold) decrease in lactate production.

An apparent inconsistency between our putative pathway (Fig. 1) and the gene knockout results is the enzyme catalyzing pyruvate reduction to lactate and the source of reducing equivalents, as mutants lacking LdhA and PDH retained wild-type metabolite profiles. While formate, a product of pyruvate disproportionation catalyzed by PflB, represents an alternative to the NADH source of reducing equivalents for pyruvate reduction, the enzymatic catalyst of this process is not clear. We focused our search for the protein(s) catalyzing pyruvate reduction to lactate on two novel recently described oxidative lactate dehydrogenases of S. oneidensis MR-1, d-lactate-specific (dld-II [SO_1521]) and l-lactate-specific tripartite (lldEFG [SO_1520 to SO_1518]) enzymes (21). Unlike other known NAD(P)H-independent oxidative lactate dehydrogenases, both Dld-II and LldEFG contain ferredoxin domains (21). The involvement of both enzymes in pyruvate reduction was tested with S. oneidensis MR-1 mutants deficient in one or both lactate dehydrogenases (Δdld-II, ΔlldF, and Δdld-II ΔlldF). Two of these mutants, the Δdld-II and Δdld-II ΔlldF mutants, exhibited identical pyruvate fermentation profiles and produced lactate at a rate 2-fold lower than that of wild-type cultures but other end products at rates >3-fold higher than those of the wild type (Table 2). In contrast, the ΔlldF strain produced lactate and other products at the same rate as that for the wild type (data not shown).

These results demonstrate that S. oneidensis MR-1 fermentative pyruvate metabolism is catalyzed by pyruvate formate-lyase, with the oxidation of formate to CO₂ via the action of several different enzymes, including the periplasmic formate dehydrogenase and [Ni-Fe] hydrogenase. The conversion of acetyl-CoA to acetate appears to be the principal ATP-generating pathway during pyruvate fermentation. PflB activity is also critical for the reductive branch of this pathway, which is catalyzed by NAD(P)H-independent d-lactate dehydrogenase. Although not lethal, the inactivation of pflB leads to a dramatic decrease in anaerobic pyruvate conversion rates and the elimination of formate and H₂ production, while pyruvate catabolism is catalyzed by PDH, LdhA, and Pta/AckA.

Energetics of fumarate respiration.

Substrate-level phosphorylation via Pta and AckA (Fig. 1) may also be an important part of energy conservation during the anaerobic respiratory growth of S. oneidensis MR-1, as significant concentrations of acetate were reported previously to accumulate extracellularly (13, 15). To quantitatively assess the energetics of anaerobic growth, we carried out a material balance analysis of an electron acceptor-limited fumarate-reducing chemostat culture supplemented with lactate as the sole carbon and energy source (Table 3). It was calculated that 82% of the consumed lactate was converted to acetate, and the sum of the specific rates of lactate consumption and acetate production was equal to that of fumarate reduction to succinate. This stoichiometry strongly suggests that fumarate reduction is coupled primarily to the lactate dehydrogenase reaction and oxidation of pyruvate to acetyl-CoA, while the TCA cycle and glyoxylate shunt do not appear to significantly contribute to lactate oxidation under fumarate-reducing conditions. This notion is supported by the growth profiles of mutants defective in glyoxylate shunt enzymes, specifically, malate synthase (AceB) and isocitrate lyase (AceA). The growth of these mutants was indistinguishable from that of the wild type, while an α-ketoglutarate dehydrogenase deletion mutant (ΔsucB) was able to grow but at a reduced growth rate and biomass yield compared to those of the wild type (Fig. 2).

Table 3.

Growth parameters measured for S. oneidensis MR-1 chemostat cultures under anaerobic fumarate-reducing conditions at steady state^a

Compound	Concn (mM)	Consumption rate (mmol/g AFDW·h)	Production rate (mmol/g AFDW·h)
Lactate	10.94	11.8	NA
Pyruvate	0.34	NA	0.21
Acetate	15.62	NA	9.5
Formate	ND	NA	NA
Fumarate	0.02	21.1	NA
Succinate	34.6	NA	21.2

^aThe biomass concentration was 0.165 g AFDW/liter, and the yield was 8.66 g AFDW/mol lactate. NA, not available; ND, not detected.

Fig. 2.

Growth dynamics of S. oneidensis MR-1 and selected mutants in anaerobic M1 medium supplemented with 20 mM lactate and 30 mM fumarate. Some error bars are smaller than the symbols.

To calculate the input of AckA-catalyzed substrate-level phosphorylation into the overall energy balance, we compared the amount of ATP required for the production of 1 g AFDW biomass to the amount of excreted acetate. Using the recently developed constraint-based metabolic model of S. oneidensis MR-1 (referred to as iSO783), experimental data on biomass composition, and previously determined growth-rate-dependent and non-growth-rate-dependent ATP requirements (20), we estimated that the production of 1 g AFDW biomass at a growth rate of 0.1 h⁻¹ requires 131 mmol ATP. (To determine the effect of biomass composition on ATP requirements, we varied protein [the most energetically “expensive” polymer] concentrations in the biomass from 0.37 to 0.69 g/g AFDW and found that the ATP requirement changed only 2.3%.) Based on the biomass yield and the stoichiometry of lactate conversion to acetate (Table 3), we determined that substrate-level phosphorylation could generate 95 mmol ATP, or 72.5% of the amount of ATP necessary to synthesize 1 g AFDW when S. oneidensis MR-1 is grown anaerobically on lactate using fumarate as the electron acceptor. The remaining 27.5% of required ATP is likely produced through oxidative phosphorylation coupled to lactate catabolism.

Based on the stoichiometry and known pathways of lactate catabolism, there are three putative reactions that can result in reductant generation, specifically catalyzed by LDH, PDH, and PflB in conjunction with FDH. The role of both oxidative LDH enzymes was previously demonstrated (21). To probe the involvement of PDH and PflB in pyruvate catabolism, we tested the fumarate-dependent anaerobic growth of the corresponding knockout strains. The ΔpflB mutant did not grow anaerobically, nor did it oxidize lactate or reduce fumarate, whereas its aerobic growth on lactate was not impaired (20). The ΔaceE mutant did not lose the ability to grow under fumarate-reducing conditions, but its growth rate and biomass production were significantly decreased (Fig. 2). With respect to the role of FDH, formate (a product of the PflB enzyme reaction) was oxidized by anaerobic resting wild-type cells when fumarate was provided as an electron acceptor at rates of 18 to 25 mmol/g AFDW·h.

Energy metabolism under conditions of O₂ limitation.

Chemostat cultures of S. oneidensis MR-1 were grown on lactate at steady state under conditions of O₂ limitation, and as described above, the amount of excreted acetate was used as a measure of energy conservation via substrate-level phosphorylation. Previously, it was shown that acetate production from glucose by Escherichia coli can be a function of O₂ availability, which is determined essentially by gas mass transfer (1). Therefore, to avoid any ambiguity in calculating the maximal contribution of substrate-level phosphorylation, we defined O₂-limited growth as conditions where the O₂ flux is insufficient to metabolize all provided lactate into biomass or by-products, and therefore, free lactate is present in excess.

Similar to fumarate-reducing cells, pyruvate and acetate were the only metabolites excreted into the medium by the O₂-limited S. oneidensis MR-1 lactate-fed chemostat culture (Table 4). The difference between consumed lactate and the sum of excreted organic acids (3.5 mM) represents the amount of lactate utilized for biomass synthesis or fully oxidized to CO₂. This corresponds to 12.84 mmol lactate per g of AFDW, since the biomass concentration in the bioreactor was 0.275 g AFWD/liter under conditions of steady-state growth (Table 4). The carbon content with this amount of lactate is 38.51 mmol, which is close to the average carbon content (47.5%, or 39.58 mmol C/g AFDW) in bacterial biomass (3). This suggests that under O₂-limited conditions, the bacteria oxidized little lactate to CO₂ via the TCA cycle. Using the constraint-based metabolic model of S. oneidensis MR-1 as described above, we calculated that approximately 33% of the ATP necessary to synthesize 1 g AFDW biomass was generated by substrate-level phosphorylation under O₂-limited conditions. We infer that electron transport-based energy conservation was responsible for the remainder of the ATP requirement.

Table 4.

Growth parameters measured for S. oneidensis MR-1 chemostat cultures under conditions of O₂ limitation at steady state^a

Compound	Concn (mM)	Consumption rate (mmol/g AFDW·h)	Production rate (mmol/g AFDW·h)
Lactate	4.13	6.5	NA
Pyruvate	2.59	NA	0.94
Acetate	11.75	NA	4.3
Formate	ND	NA	NA
Fully oxidized/assimilated lactate	3.53

^aThe biomass concentration was 0.275 g AFDW/liter, and the yield was 15.39 g AFDW/mol lactate. NA, not available; ND, not detected.

To identify the enzymatic catalyst(s) of pyruvate conversion to acetyl-CoA, we compared the growths of the ΔpflB and ΔaceE mutants under conditions of O₂ limitation. As shown in Fig. 3, the growth of the ΔaceE strain was similar to that of the wild type, while the ΔpflB strain was unable to grow under these conditions, although the strain was able to oxidize lactate to acetate albeit at a 4.5-fold-lower rate. No PDH activity was measured in cell extracts prepared from chemostat-grown O₂-limited wild-type cultures, while high specific activity was detected in lactate-limited cells grown aerobically (Table 5). In contrast, PflB activity was present in extracts from O₂-limited cultures (Table 5), providing additional evidence that this enzyme solely facilitated pyruvate degradation to acetyl-CoA under O₂-limited conditions. To test if substrate-level phosphorylation is indispensable for the O₂-limited growth of S. oneidensis MR-1, we evaluated the growth of the ΔackA mutant. Our results show that acetate kinase activity is not required for growth under conditions of O₂ limitation; growth decreased only by 25% relative to that of the wild-type strain despite a significant decrease in levels of acetate production (Fig. 3). The acetate produced under these conditions was likely due to the natural instability of acetyl phosphate (4, 12) or its degradation by acylphosphatase (SO_2253).

Fig. 3.

Effects of gene deletions on S. oneidensis MR-1 growth (A) and acetate accumulation (B) under O₂-limited conditions in M1 medium supplemented with 90 mM sodium lactate as the sole source of carbon. Experiments were run in controlled bioreactors (see Materials and Methods for details). Samples for data presented here were taken at 16.5 to 17 h after the O₂ limitation phase started. Growth was measured as the optical density at 600 nm. The results are representative of a typical experiment. Experiments were repeated twice, and standard deviations did not exceed 8%.

Table 5.

Specific activities of selected enzymes in cell extracts prepared from cultures grown on lactate in carbon-limited fully aerobic and O₂-limited chemostats

Enzyme	Sp act (U/min·mg protein)
	Carbon-limited chemostata	O2-limited chemostat
Pyruvate dehydrogenase	5.85	ND
Pyruvate formate-lyase	ND	0.16
Malate synthase	ND	0.01
Isocitrate lyase	0.085	0.182

^aData from reference 20. ND, not detected.

Although NADH oxidation does not appear to be the main source of electrons for oxidative phosphorylation under O₂-limited conditions, NADH may be a precursor for NADPH, an important carrier of reducing equivalents for some major biosynthetic reactions (16). As we had eliminated PDH-catalyzed pyruvate oxidation as a mechanism for NADH generation, and material balance does not support the functioning of a complete TCA cycle, we hypothesized that the glyoxylate shunt was, in part, a source of NADH under conditions of O₂-limited growth on lactate. In support of this hypothesis, activities of key glyoxylate shunt enzymes, isocitrate lyase and malate synthase, were detected in cell extracts. In contrast, the isocitrate lyase specific activity was 2-fold lower, and no malate synthase activity was found in cells grown in a fully aerobic lactate-limited chemostat (Table 5). Additionally, the ΔaceA strain grew and produced acetate at a rate about 25% lower than that of the wild type (Fig. 3).

DISCUSSION

Bacteria of the genus Shewanella inhabit a wide range of redox-stratified environments and have also been found in association with oxygenic phototrophs in the water column and biofilms (2, 18, 24). Their metabolic versatility has been suggested to be a key reason why these bacteria are able to thrive in such diverse habitats where the electron donor and acceptor vary widely in type and concentration, yet mechanistic and quantitative aspects of energy conservation in Shewanella are poorly understood. Under fully aerobic conditions, S. oneidensis MR-1 did not produce any by-products, other than CO₂, from lactate and therefore generated ATP exclusively by oxidative phosphorylation (14, 20). However, according to our experimental data, conditions of electron acceptor limitation promoted substrate-level ATP generation, where the ATP fraction produced by oxidative phosphorylation varied significantly as a function of the electron acceptor availability and type. Moreover, in the absence of an electron acceptor, S. oneidensis MR-1 exhibited the ability to survive, but not grow, by fermenting pyruvate (15) or serine (our unpublished data).

Based on experimental results, the fermentative pyruvate metabolism of S. oneidensis MR-1 can be subdivided into two modules, the first one generating ATP from acetyl-CoA and the second one catalyzing the disposal of formate. The latter is a product of the pflB-facilitated reaction and can be harmful to cells, as it was shown previously to inhibit growth under anaerobic conditions (15). There are apparently two pathways of formate transformation: disproportionation to H₂ and CO₂ or oxidation coupled to the reduction of pyruvate to lactate. The impairment of either pathway by the deletion of the corresponding genes resulted in increased rates of formate accumulation and pyruvate consumption/acetate production (Table 2), i.e., higher specific rates of ATP synthesis. This increase can be explained by the fact that formate accumulation in the cytoplasm leads to a decrease of the pH and therefore to a partial proton motive force (PMF) collapse across the cytoplasmic membrane (26). It is likely that in order to compensate for the transmembrane charge loss, cells increased rates of ATP synthesis to restore intracellular pH via an ATP-utilizing proton pump such as the F-type ATPase (10, 27).

S. oneidensis MR-1 formate dehydrogenases and hydrogenases are predicted to be localized to the S. oneidensis MR-1 periplasm (23), suggesting that formate disproportionation to H₂ and CO₂ does not result in energy conservation. In contrast, the localization of Dld-II in the cytoplasm suggests that the pyruvate reduction to lactate coupled to formate oxidation may be accompanied by the generation of PMF (Fig. 4). Therefore, the energy metabolism of pyruvate-fermenting S. oneidensis MR-1 cells is likely a combination of fermentative and respiratory pathways, where the source of electrons (pyruvate) can also serve as an electron acceptor.

Fig. 4.

Schematic representation of the proposed spatial localization of enzymes involved in pyruvate reduction. According to the proposed scheme, protons resulting from formate oxidation remain in the periplasm, whereas electrons through the electron transport chain are moved toward Dld-II and used in a reaction of pyruvate reduction, with the latter accompanied by the consumption of two protons in the cytoplasm. This process may therefore lead to PMF generation. Dld-II, oxidative d-lactate dehydrogenase; FDH, formate dehydrogenase.

We determined that the ATP produced by S. oneidensis MR-1 during anaerobic metabolism through fumarate reduction-driven oxidative phosphorylation constituted only 27.5% of the overall energy budget, which is in compliance with the previously established ability of the Δatp mutant to grow anaerobically (10). This fact raises the question of why pyruvate fermentation cannot sustain S. oneidensis MR-1 growth, since most ATP is produced from acetyl-CoA in both fermenting and fumarate-respiring cells. One possibility is that fumarate reduction sustains a significantly higher rate of formate oxidation, thus providing its efficient removal and maintaining transmembrane potential. Another possibility is that the observed formate accumulation by cells fermenting pyruvate negatively affects cellular metabolism, since this compound can be a potent metabolic regulator (9).

The elimination of substrate-level phosphorylation in the ΔackA strain had different consequences for S. oneidensis MR-1 growth under anaerobic versus O₂-limited conditions. The mutant did not grow when fumarate or Fe(III) citrate was supplemented as an electron acceptor (10), whereas its O₂-limited growth was not significantly affected. However, the deletion of ackA dramatically reduced acetate accumulation in O₂-limited cultures, suggesting that oxidative phosphorylation was the main source of ATP for this mutant growing on lactate.

Another finding of this research concerns the role of PflB and PDH in S. oneidensis MR-1 lactate metabolism. Under anaerobic and O₂-limited conditions, PflB is the main enzyme catalyzing pyruvate catabolism in wild-type cells, and the ΔpflB mutant could not grow under either condition. In spite of the inability to grow, the ΔpflB mutant was able to oxidize lactate under conditions of O₂ limitation (Fig. 3) but not with fumarate as the electron acceptor (20). Conversely, the incapacitation of PDH by the deletion of aceE did not affect S. oneidensis growth under conditions of O₂ limitation but severely decreased it when fumarate was the electron acceptor. Taken together, our results suggest that PDH in conjunction with transhydrogenase (16) supplies NADPH for biosynthetic reactions under anaerobic conditions, whereas under conditions of O₂ limitation, the glyoxylate shunt likely serves this purpose.

S. oneidensis MR-1 uses NAD as a redox intermediate when it grows aerobically, but under anaerobic and O₂-limited conditions, formate is likely the main intermediate that serves as an electron donor for respiration. The deletion of ackA shifted S. oneidensis MR-1 metabolism toward ATP production by oxidative phosphorylation under conditions of O₂ limitation but not under conditions of anaerobic growth. These observations combined with the ΔpflB phenotype may reflect the inability of S. oneidensis MR-1 to couple NADH oxidation to fumarate or Fe(III) citrate reduction. The inability to reoxidize NADH may be the metabolic reason for the restricted range of organic compounds supporting S. oneidensis MR-1 growth under conditions of anaerobic respiration. We hypothesize that if the pathway for the catabolism of a given compound involves the generation of NAD(P)H in an amount exceeding the needs of anabolism, then the compound cannot be used as a sole source of carbon and energy by S. oneidensis MR-1 under anaerobic conditions. However, if O₂ is available, then NADH can be reoxidized. This is an important metabolic difference between Shewanella and other anaerobic dissimilatory Fe(III)-reducing bacteria such as Geobacter that likely explains, at least in part, the differences in their environmental distributions. The latter bacteria are able to fully oxidize acetate via the TCA cycle in the absence of O₂ and use the resulting NAD(P)H to reduce electron acceptors and couple this process with energy conservation (5, 8), whereas the former bacteria produce ATP mostly from acetyl phosphate under anaerobic conditions. The ability of Shewanella to inhabit a wide range of environments with variable electron acceptors and donor types and concentrations may be due in part to their flexibility in shifting energy metabolism between oxidative and fermentative pathways of ATP synthesis. Upon a shift from the oxidative to the fermentative mode, formate generation, instead of NADH production, can present a significant advantage, as formate can be excreted and converted to H₂ or consumed by neighboring species. Such excretion can prevent metabolic pathway disruption and increase Shewanella resilience against environmental perturbations.