|Home | About | Journals | Submit | Contact Us | Français|
It is well established that respiratory organisms use proton motive force to produce ATP via F-type ATP synthase aerobically and that this process may reverse during anaerobiosis to produce proton motive force. Here, we show that Shewanella oneidensis strain MR-1, a nonfermentative, facultative anaerobe known to respire exogenous electron acceptors, generates ATP primarily from substrate-level phosphorylation under anaerobic conditions. Mutant strains lacking ackA (SO2915) and pta (SO2916), genes required for acetate production and a significant portion of substrate-level ATP produced anaerobically, were tested for growth. These mutant strains were unable to grow anaerobically with lactate and fumarate as the electron acceptor, consistent with substrate-level phosphorylation yielding a significant amount of ATP. Mutant strains lacking ackA and pta were also shown to grow slowly using N-acetylglucosamine as the carbon source and fumarate as the electron acceptor, consistent with some ATP generation deriving from the Entner-Doudoroff pathway with this substrate. A deletion strain lacking the sole F-type ATP synthase (SO4746 to SO4754) demonstrated enhanced growth on N-acetylglucosamine and a minor defect with lactate under anaerobic conditions. ATP synthase mutants grown anaerobically on lactate while expressing proteorhodopsin, a light-dependent proton pump, exhibited restored growth when exposed to light, consistent with a proton-pumping role for ATP synthase under anaerobic conditions. Although S. oneidensis requires external electron acceptors to balance redox reactions and is not fermentative, we find that substrate-level phosphorylation is its primary anaerobic energy conservation strategy. Phenotypic characterization of an ackA deletion in Shewanella sp. strain MR-4 and genomic analysis of other sequenced strains suggest that this strategy is a common feature of Shewanella.
Shewanella oneidensis strain MR-1 is a nonfermentative, facultative anaerobe which respires various substrates, including oxygen, soluble metals, insoluble iron and manganese oxide minerals, electrodes, and organic compounds (8, 12, 18, 22). Other bacteria with the ability to respire electrodes and oxide minerals, such as Geobacter and Geothrix, oxidize acetate to carbon dioxide (4, 7, 9), consistent with these organisms generating ATP primarily from oxidative phosphorylation rather than substrate-level phosphorylation. Yet, an examination of metabolic end products and a variety of central metabolism and flux analyses of MR-1 show that acetate is the major product under anaerobic conditions (18, 27, 29, 31). The general anaerobic metabolism model for MR-1, as depicted in Fig. Fig.1,1, has key features of glycolysis via the Entner-Doudoroff pathway as well as acetyl coenzyme A (acetyl-CoA) flux toward acetate anaerobically via phosphate acetyltransferase (Pta) and acetate kinase (AckA) (27, 29, 31). High-performance liquid chromatography (HPLC) studies in our lab and others have shown that pyruvate may be excreted during lactate utilization both aerobically and anaerobically (30, 31), and MR-1 has not been shown to maintain significant flux through the tricarboxylic acid (TCA) cycle under anaerobic conditions (31).
Characterization studies of proton motive force (PMF) in MR-1 have not definitively determined whether the source of anaerobic proton pumping or translocation is electron transport, ATP synthase, or metabolite transport (13, 19). Myers et al. demonstrated that anaerobic MR-1 cells starved of electron acceptor generate PMF in response to fumarate addition (19). However, the directionality of the ATP synthase (i.e., generation of ATP or ATPase to pump protons) was not characterized. Previous work has confirmed that proteorhodopsin (PR), a light-dependent, proton-pumping integral membrane protein, can be used in MR-1 to supplement PMF (13). However, the observed increase in PMF in wild-type cells expressing PR did not result in higher optical densities (ODs) or in a higher growth rate. Though all known bacteria depend on PMF, whether MR-1 uses that PMF for ATP production or uses ATP to help generate PMF under anaerobic conditions has yet to be determined.
To examine ATP production in MR-1, growth on carbon sources that offer various amounts of substrate-level-derived ATP and reducing equivalents (NADH, formate, or quinones) in their oxidation was characterized. Two carbon sources entering central metabolism at different locations are N-acetylglucosamine (NAG) and lactate, which enter before and after glycolysis, respectively (Fig. (Fig.1)1) (24, 36). Both are oxidized to acetate and carbon dioxide anaerobically, though lactate yields one ATP and two reducing equivalents per molecule, while NAG yields three ATPs and four reducing equivalents per molecule. The differences in ATP yields derived from utilization of NAG versus lactate, combined with modification of those yields through gene deletions, allowed for characterization of ATP production in MR-1.
The goal of this work was to elucidate the primary source of ATP generation under anaerobic conditions in MR-1. Data presented here support a model of anaerobic metabolism where substrate-level phosphorylation is the primary mechanism for ATP generation and where some amount of the ATP pool is used to generate PMF. Paradoxically, the most diverse respiratory organism characterized to date (8, 12, 22) does not generate ATP from electron transport reactions and PMF. Our finding highlights a critical difference in metabolic strategies between Shewanella and other organisms that are able to reduce insoluble substrates, such as Geobacter and Geothrix.
Strains used in this study are listed in Table Table1.1. Aerobic and anaerobic growth (in sealed, nitrogen-flushed tubes) was assayed as previously described (1, 10). Minimal medium consisted of 0.225 g K2HPO4, 0.225 g KH2PO4, 0.46 g NaCl, 0.225 g (NH4)SO4, and 0.117 g MgSO4·7H2O per liter, buffered to a pH of ~7.2 (using NaOH) with 10 mM and 100 mM HEPES for aerobic and anaerobic cultures, respectively. HEPES buffer was used to avoid potential pH shifts associated with anaerobic growth or its by-products. Vitamin mix (5 ml/liter), 5 ml/liter modified mineral mix, and 0.01% Bacto Casamino Acids were added to minimal media to stimulate growth. Carbon sources were routinely at a concentration of 20 mM or as indicated. Lactate (40 mM) was used in aerobic overnight cultures of the Δatp strain due to low biomass yield. The vitamin mix consisted of 0.002 g biotin, 0.002 g folic acid, 0.01 g pyridoxine-HCl, 0.005 g riboflavin, 0.005 g thiamine, 0.005 g nicotinic acid, 0.005 g pantothenic acid, 0.0001 g vitamin B12, 0.005 g p-aminobenzoic acid, and 0.005 g thioctic acid per liter in water and was neutralized and filter sterilized prior to use. Mineral mix was as previously described (16) and modified by adding 3.0 g MgSO4·7H2O, 0.5 g MgSO4·H2O, 1.0 g NaCl, and 0.1 g CaCl2·2H2O and by excluding Na2SeO4. Anaerobic cultures were given fumarate as the electron acceptor at a maximum concentration of 40 mM to avoid growth inhibition observed at higher concentrations (data not shown). For testing other terminal electron acceptors, Luria broth (LB) was mixed 1:1 with minimal medium containing lactate (20 mM final concentration); next, dimethyl sulfoxide (DMSO; 20 mM), sodium nitrate (40 mM), Fe(III) citrate (20 mM), or trimethylamine N-oxide (TMAO; 20 mM) was added, 100 μl of aerobic LB overnight cultures was added, and then the mixture was flushed with N2 gas for 10 min. When necessary to maintain a plasmid, 50 μg ml−1 kanamycin was added to the growth medium.
Strains and plasmids used with proteorhodopsin are listed in Table Table1.1. Strain construction, verification, growth, and light conditions were as previously described (13). Growth was characterized anaerobically using minimal medium with 10 μM retinal, 50 μg ml−1 kanamycin, and lactate and fumarate at the concentrations indicated in Fig. Fig.55.
Deletion strains and complementation vectors were made as previously described (11, 26) using the plasmids and strains listed in Table Table1.1. Deletion strains were verified using PCR and phenotypic analysis. Complementation vectors were verified by sequencing the cloned insertion. For deletion construction, upstream regions (~1 kb) were amplified using primers named with “UF” and “UR” and downstream regions were amplified using primers named with “DF” and “DR.” Primers used in this study are listed in Table Table2.2. Restriction enzymes were chosen based on sequences of amplified regions (as indicated in Table Table2).2). PCR-amplified regions and pSMV3 were digested with the appropriate endonucleases and ligated in 3-way ligations. The method for mutagenesis of ackA in Shewanella sp. strain MR-4 was slightly different in that a single primer set was designed to amplify ackA and ~1 kb flanking on both sides. This PCR product was cloned into a TOPO TA cloning vector (Invitrogen Technologies, Carlsbad, CA). Primers were then designed to amplify outward from the ends of the cloned ackA gene (delLeft, delRight). This PCR product was religated to itself and transformed into Escherichia coli UQ950; the new insert, lacking ackA, was digested and cloned into pSMV3 using SpeI and used as previously described. Complementation plasmids were made using PCR-amplified regions (Table (Table2)2) and pBBR1MCS-2 digested with the appropriate endonucleases and ligated. The endogenous Plac promoter in pBBR1MCS-2 drove expression of genes on complementation plasmids.
Cells were grown aerobically overnight in minimal medium with 20 mM lactate, washed once with fresh medium, resuspended to a final optical density of 0.5 (600 nm) in minimal medium containing 20 mM lactate and 40 mM fumarate, and transferred to sealed anaerobic tubes containing medium preflushed with nitrogen for several minutes to remove oxygen. Initial time points were taken for all samples, and wild-type experiments were allowed to proceed for 5 h while mutant strains were incubated overnight (22 h) prior to sampling to account for less overall activity exhibited by the mutant strains. Samples were centrifuged at 16,000 × g for 5 min to pellet cells and remove particulates and then stored at −20°C for analysis.
HPLC was performed using a Shimadzu system (Shimadzu, Columbia, MD), consisting of an SCL-10A system controller, an LC-10AT pump unit, an SIL-10AF automatic injection unit, a CTO-10Avp oven with a Bio-Rad HPX-87H cation-exchange column, and two detectors in series, an SPD-10Avp UV-visible-light detector and an RID-10A refractive index detector. The oven was maintained at 42°C, and a flow rate of 0.4 ml/min of 5 mM H2SO4 was used for the mobile phase.
To examine ATP generation via substrate-level phosphorylation in MR-1, strains were made with in-frame deletions of ackA (ΔackA) and pta (Δpta), as well as a double deletion (ΔackA Δpta). ackA and pta were deleted separately and in combination to account for potential regulation effects posed by the presence or absence of acetyl phosphate (35). Strains were initially tested with lactate to use growth as a proxy for ATP generation. We predicted that loss of a single ATP from a pathway producing only one ATP and two reducing equivalents would be fatal to a substrate-level-based system. When all three deletion strains and the wild type were tested in minimal medium under anaerobic conditions with lactate as the electron donor and fumarate as the electron acceptor, only the wild type grew, in agreement with our prediction (Fig. (Fig.2A).2A). The growth defects of mutant strains lacking ackA and pta were also complemented by providing wild-type copies of each gene (Fig. (Fig.2B).2B). ackA and pta mutant strains were also defective in anaerobic growth with lactate and DMSO, TMAO, or nitrate as the electron acceptor and were unable to reduce Fe(III) citrate (Table (Table33).
Though growth was not observed for the ΔackA, Δpta, or ΔackA Δpta strain, we expect that they retain some ability to consume substrate, and monitoring this rate and quantifying products can provide additional insight into central metabolism. HPLC analysis of anaerobic culture supernatants confirmed that lactate was metabolized in deletion strains. Cells were identified as metabolically active under these conditions through the conversion of lactate to pyruvate or acetate and fumarate to succinate (Table (Table4).4). The mutant strains are metabolically active yet are unable to grow, consistent with the inability to produce sufficient ATP. Discrepancies in the electron balance are suspected to be due to the large inocula, since dead cells would have been carried over and metabolized.
If growth defects in deletion strains on lactate under anaerobic conditions were caused by eliminating substrate-level-derived ATP gained from lactate, using a higher-substrate-level ATP-yielding carbon source, like NAG, may provide additional insights into ATP generation. As depicted in Fig. Fig.1,1, oxidation of NAG to acetate and CO2 produces three ATPs per molecule through substrate-level phosphorylation and four reducing equivalents. Anaerobic growth in minimal medium with NAG as the carbon source and fumarate as the electron acceptor of mutant strains lacking ackA and/or pta was decreased but not eliminated (Fig. (Fig.3).3). The limited growth observed with this carbon source is likely due to ATP generated via the Entner-Doudoroff pathway.
If substrate-level phosphorylation is the primary means of conserving energy anaerobically, a strain lacking the ATP synthase operon (Δatp) should remain viable and eliminate the possibility of ATP generation via PMF. The entire ATP synthase operon (SO4746 to SO4754) was deleted to avoid futile cycling of protons or ATP through a partial complex. We tested the Δatp mutant under aerobic conditions to determine if oxidative phosphorylation was the primary form of aerobic energy conservation. As expected, cultures of the Δatp mutant grew poorly under aerobic conditions in minimal medium using either NAG (Fig. (Fig.4A)4A) or lactate (data not shown). Complementation of the Δatp mutant with the wild-type operon restored growth in aerobic cultures (Fig. (Fig.4B).4B). Surprisingly, Δatp mutant cells grown anaerobically in minimal medium using NAG had higher growth rates than wild-type MR-1 and maximum optical densities (Fig. (Fig.4C),4C), suggesting that the ATP synthase may consume ATP under these slow-growth conditions. The Δatp mutant in minimal medium using lactate under anaerobic conditions exhibited a minor defect in growth rate and maximum optical density (Fig. (Fig.4D).4D). Mutants lacking ATP synthase grown anaerobically with either NAG or lactate oxidized the carbon source to acetate and reduced fumarate to succinate, as observed via HPLC analysis of supernatants (data not shown). The anaerobic-growth phenotypes of the Δatp mutant are consistent with substrate-level phosphorylation and not PMF-driven phosphorylation as the primary ATP source under these conditions.
To gain insight regarding the directionality of ATP synthase under anaerobic conditions, we supplemented PMF artificially through the introduction of PR. Previous work demonstrated that PR generates PMF in MR-1 when exposed to light in the presence of retinal, an essential cofactor for the enzyme (13). Therefore, the addition of PR to the Δatp mutant should at least partially restore anaerobic growth on lactate if the growth defect was due to suboptimal PMF. PR expression strains in the MR-1 and Δatp mutant backgrounds were tested for growth in the presence and absence of light to supplement and control for PMF, respectively. When exposed to light, strains lacking ATP synthase but expressing PR had a significant increase in biomass yield (50% increase in maximum optical density) when grown anaerobically with lactate and fumarate (Fig. (Fig.5).5). As previously characterized (13), the wild-type growth rate and maximum optical density were not significantly affected by the presence or absence of light.
We hypothesized that under anaerobic conditions, MR-1 relies on substrate-level phosphorylation as the primary means of energy conservation. A major prediction of this hypothesis is that anaerobic growth using lactate as a carbon source would be significantly hindered for the ΔackA, Δpta, and ΔackA Δpta strains due to loss of the primary ATP-producing reaction. Growth experiments with mutants defective in ackA and/or pta verified this prediction (Fig. (Fig.2).2). NAG offers higher ATP yields than lactate, though only one net ATP would be gained by strains lacking ackA and/or pta (Fig. (Fig.1).1). During NAG utilization under anaerobic conditions, these strains had strong defects in both growth rate and growth yield (Fig. (Fig.3).3). Expression studies by Beliaev et al. (3) and Charania et al. (6) have shown that the nag cluster, a gene cluster containing the NAG utilization genes in MR-1 (36), is downregulated anaerobically. We hypothesize that the decreased NAG utilization rate associated with the downregulation of nag genes and the resulting decrease in growth rate allow for PMF dissipation. The depleted PMF causes ATP to be consumed by ATP synthase to pump protons to help reestablish the gradient. This idea was supported by increased growth of the Δatp mutant on NAG under anaerobic conditions (Fig. (Fig.4C4C).
We determined ATP synthase directionality through restoration of the Δatp mutant growth defect by heterologously expressing PR and exposing these cultures to light (Fig. (Fig.5).5). Using light energy, PR directly contributes to PMF by pumping protons across the cytoplasmic membrane (2, 17). The growth defect observed is due to either (i) the inability to generate sufficient levels of ATP or (ii) the inability to generate sufficient PMF. Introduction of PR partially restored the growth of the Δatp strain in the presence of light, indicating that the strain was not operating at optimal PMF since proton pumping cannot contribute directly to ATP generation in this mutant background. This result is consistent with ATP synthase working in reverse to contribute to PMF under anaerobic conditions in MR-1.
We acknowledge the alternative conclusion that PR could create an artificially high PMF resulting in an increase in proton-dependent transport substrates, such as lactate (23). However, we excluded this possibility because no increase in growth was observed for the wild type with PR, which would experience the same effect. Our observations are consistent with work showing that anaerobically grown cells starved of electron acceptors produce PMF in response to fumarate (19). Electron acceptor limitation of cells limits central metabolic flux and substrate-level phosphorylation. Based on the anaerobic-growth phenotype of the Δatp mutant using lactate and restoration of growth with PMF supplementation, we know that some amount of PMF must be generated by both the electron transport chain and the reversed ATP synthase. If the electron transport chain did not significantly contribute to PMF under anaerobic conditions, the Δatp strain should exhibit a more severe growth defect on either NAG or lactate, because PMF is required for metabolite transport and cellular maintenance. Therefore, when fumarate was introduced to cells limited in electron acceptors, we propose that metabolism resumed, ATP was made via substrate-level phosphorylation, and PMF was generated through both the reversed function of ATP synthase and electron transport.
Additional work in our lab has shown that deletion of acetate kinase (ackA) in Shewanella sp. strain MR-4 (an isolate from the Black Sea ) results in a similar inability to grow under anaerobic conditions using either glucose or lactate as the electron donor (see Fig. S1 in the supplemental material). It has also been established that acetate is the primary end product during anaerobic growth of numerous Shewanella strains, as seen through metabolite characterization (7, 15, 25, 27). Genomic analysis of sequenced strains of Shewanella shows the acetate kinase/phosphotransacetylase cluster to be conserved throughout the genus (see Table S1 in the supplemental material). Our characterization of MR-1 and MR-4, together with metabolite and genomic analysis, suggests that reliance on substrate-level phosphorylation is likely a trait shared throughout Shewanella.
How does electron transport directly contribute to PMF in Shewanella? Though we have treated all reducing equivalents the same for simplicity, there are notable differences in their oxidation, and NADH and quinones may differ in proton-pumping/translocation ability depending on the dehydrogenase. In addition, work with E. coli has demonstrated that both electron transport driven by formate oxidation and metabolite transport are potential generators of PMF (28, 32). The notion that electron transport contributes to PMF in Shewanella is supported by HPLC analysis of mutant cultures (ΔackA, Δpta, and ΔackA Δpta strains) tested anaerobically with lactate and fumarate, which remain metabolically active and continue electron transport, as seen through fumarate reduction (Table (Table4).4). However, these mutant strains are unable to grow anaerobically with lactate and fumarate (Fig. (Fig.2)2) or any other anaerobic electron acceptor tested (Table (Table3),3), suggesting that electron transport is not sufficient to support ATP generation requirements of the cell under anaerobic conditions. Moreover, the Δpta strain supplemented with PR was still unable to grow in the presence of light (data not shown), consistent with the lack of ATP generated via substrate-level phosphorylation as the primary reason for growth inhibition. We assert that electron transport is not the primary means of energy conservation used by MR-1 during anaerobic growth.
It is intriguing to consider the possibility that bacteria from the genus Shewanella, known to exhibit the widest diversity of respiratory capacity described to date (8, 12, 21, 33), do not directly conserve energy from this process. Many bacteria use substrate-level phosphorylation for the bulk of ATP production during anaerobic fermentation (34) rather than electron transport to conserve energy since they must balance redox reactions by reducing their carbon source, often at the cost of the substrate. The energetic strategy of MR-1 is a blend of respiratory and fermentative pathways, suggesting that the extensive repertoire of electron acceptors used by MR-1 serves not only to balance internal redox pools but also to prevent the need for less efficient fermentative pathways.
This work was supported by a Discovery Grant from the Institute on the Environment (IonE), the Institute for Renewable Energy and the Environment (IREE), the Cargill Higher Education Initiative (CHEI) at the University of Minnesota, and the National Science Foundation (grant CBET-0756296).
We thank Daniel Bond (University of Minnesota) for interesting and helpful discussions and the anonymous reviewers for their insightful suggestions.
Published ahead of print on 16 April 2010.
#Supplemental material for this article may be found at http://jb.asm.org/.