Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Identification of a pathway for electron uptake in Shewanella oneidensis

Subjects

Abstract

Extracellular electron transfer (EET) could enable electron uptake into microbial metabolism for the synthesis of complex, energy dense organic molecules from CO2 and renewable electricity1,2,3,4,5,6. Theoretically EET could do this with an efficiency comparable to H2-oxidation7,8 but without the need for a volatile intermediate and the problems it causes for scale up9. However, significant gaps remain in understanding the mechanism and genetics of electron uptake. For example, studies of electron uptake in electroactive microbes have shown a role for the Mtr EET complex in the electroactive microbe Shewanella oneidensis MR-110,11,12,13,14, though there is substantial variation in the magnitude of effect deletion of these genes has depending on the terminal electron acceptor used. This speaks to the potential for previously uncharacterized and/or differentially utilized genes involved in electron uptake. To address this, we screened gene disruption mutants for 3667 genes, representing ≈99% of all nonessential genes, from the S. oneidensis whole genome knockout collection using a redox dye oxidation assay. Confirmation of electron uptake using electrochemical testing allowed us to identify five genes from S. oneidensis that are indispensable for electron uptake from a cathode. Knockout of each gene eliminates extracellular electron uptake, yet in four of the five cases produces no significant defect in electron donation to an anode. This result highlights both distinct electron uptake components and an electronic connection between aerobic and anaerobic electron transport chains that allow electrons from the reversible EET machinery to be coupled to different respiratory processes in S. oneidensis. Homologs to these genes across many different genera suggesting that electron uptake by EET coupled to respiration could be widespread. These gene discoveries provide a foundation for: studying this phenotype in exotic metal-oxidizing microbes, genetic optimization of electron uptake in S. oneidensis; and genetically engineering electron uptake into a highly tractable host like E. coli to complement recent advances in synthetic CO2 fixation15.

Introduction

Electromicrobial production technologies aim to combine the flexibility of CO2-fixing and C1-assimilating microbial metabolism for the synthesis of complex, energy-dense organic molecules from CO2 and renewable electricity1,2,3,4,5,6. Already, the Bionic Leaf device has demonstrated that technologies of this class could dramatically exceed the efficiency of photosynthesis7,8. However, while highly efficient at lab scale, the Bionic Leaf relies on H2 oxidation to transfer electrons from the electrode to microbes, and the low solubility of H2 in water would pose a significant challenge for scale-up of this and related technologies9.

Extracellular electron uptake (EEU) as an electron source for metabolism could allow engineers to circumvent the scale-up limitations of H2 oxidation. Naturally occurring electroautotrophic microbes can produce acetate and butyrate from CO2 and electricity with Faradaic efficiencies exceeding 90%16. Furthermore, theoretical analysis suggests that the upper-limit efficiency of electromicrobial production of biofuels by EEU could rival that of H2-mediated systems9. However, naturally occurring electroactive organisms capable of EEU suffer from multiple technical drawbacks. Most notably, they have a low-tolerance to high-osmotic-strength electrolytes, requiring the use of electrolytes that confer low electrochemical cell conductivity and thus a low overall energy efficiency. Additionally, they have a poor ability to direct metabolic flux to a single product more complex than acetate or butyrate16. Correcting these problems to take full advantage of EEU’s potential by genetic engineering17 will require extensive knowledge of the genetics of EEU.

Growing evidence suggests that the model electroactive microbe S. oneidensis can couple EEU to the regeneration of ATP and NADH, both essential precursors for biosynthesis13, by reversal of its extracellular electron transfer (EET) pathway (Fig. 1), making it an attractive chassis organism for electromicrobial production. However, EEU machinery in S. oneidensis appears to involve more than just operating the well-characterized EET machinery in reverse13,18. EEU in S. oneidensis can link cathodic current with multiple terminal electron acceptors, including oxygen, which draws into question how electrons transfer between canonically discrete electron-transport chains. Finding this machinery has been hindered by the lack of high-throughput assays for electron uptake and the challenge of developing screens for non-growth-related phenotypes. Even with recent advances in high-throughput electrode arrays19, searching through the thousands of genes in even a single microbial genome by direct electrochemical measurements remains impractical.

Fig. 1: Electron uptake in the model electroactive microbe Shewanella oneidensis MR-1 cannot be fully explained by reversal of its extracellular electron-transfer pathway.
figure 1

The canonical anodic extracellular electron-transport (EET) pathway for electron deposition is shown in light blue and the putative cathodic extracellular electron-uptake (EEU) pathway is shown in pink. Known electron-transfer pathways are denoted with solid lines, while speculated transfer pathways are shown as dashed lines. Two possible mechanisms for transfer of cathodic electrons from the Mtr EET complex to the ubiquinone pool and onto terminal cytochrome oxidases are highlighted. We speculate that two of the proteins identified in this work (SO_0400 and SO_3662) could form part of possible mechanism 2.

To address this, we developed a rapid colorimetric assay to screen all 3667 members of the S. oneidensis whole-genome-knockout collection20,21 (covering ≈99% of all nonessential genes) and characterize the genetics of EEU. The assay relies upon oxidation of the reduced form of the redox dye anthra(hydro)quinone-2,6-disulfonate (AHDSred for the reduced form and AQDSox for the oxidized form) and is coupled to reduction of the anaerobic terminal electron acceptors fumarate and nitrate22,23,24 (Figs. 2 and S1). While AHDSred/AQDSox redox dye assays are not a perfect proxy for EEU and EET, they are capable of identifying many components of the S. oneidensis EET machinery20. While AHDSred/AQDSox does appear to be able to enter the cell, it also appears to be rapidly pumped out by a TolC efflux pump25. We suspect these results in a lower concentration of AHDSred/AQDSox in the interior of the cell than in the exterior solution. As a result, cell-surface proteins, like the well-known Mtr EET complex, are responsible for transferring a detectable fraction of electrons to AQDSox (ref. 20). Thus, we believed it was reasonable to assume that the AHDSred oxidation assay could detect genes involved in EEU. To ensure that genes are involved in EEU with solid surfaces, a subset was tested in electrochemical systems, the gold standard for measuring EEU26,27.

Fig. 2: A genome-wide screen of S. oneidensis finds 149 genes that disrupt electron uptake.
figure 2

All 3712 members of the S. oneidensis whole-genome-knockout collection were screened for electron uptake capability with AHDSred oxidation assays, either fumarate or nitrate as a terminal electron acceptor. In total, 149 genes disrupt AHDSred oxidation with fumarate, nitrate and in many cases both as a terminal electron acceptor (note, only one terminal electron acceptor is used at once). a AHDSred/AQDSox redox reaction is used as a proxy for extracellular electron uptake. AHDSred changes color from orange to clear when oxidized. Electrons are transferred to either a fumarate or nitrate terminal electron acceptor by S. oneidensis. b Blank, no-cell and quasi-wild-type (transposon mutants that contain a kanamycin cassette but have no effect on AHDSred oxidation) controls. The color of the AHDSred dye is recorded photographically and displayed at 1-hour intervals after the start of the experiment by a series of colored circles above each gene. Further information on this assay can be seen in Fig. S1 and “Materials and Methods”. Data shown use fumarate. c The electron-uptake assay associates 149 genes with electron uptake. Electron uptake failure can be explained in 108 cases, but in 41 cases, it fails for unknown reasons, implicating these genes in an uncharacterized electron-uptake process. Full screening results and functional categorizations are shown in Supplementary Data 1. dg AHDSred oxidation assay results are shown for selected mutants containing deletions of genes highlighted in this article that robustly disrupt electron uptake from a cathode (the selection process for these five mutants is shown in Fig. 3) (pink arrow with a red cross through the center) along with gene-disruption mutants for the surrounding genes (purple arrow, with a red line indicating the location of the transposon insertion). All time courses are from experiments using fumarate as a terminal electron acceptor.

Results and discussion

High-throughput electron uptake screen finds 41 genes with unknown function

We identified mutants in 149 coding and intergenic regions in the S. oneidensis genome that slowed or eliminated AHDSred oxidation with fumarate, nitrate, or both terminal electron acceptors (Supplementary Data 1). While O2 is the most useful terminal electron acceptor for electromicrobial production due to its high redox potential and enormous availability, we were unable to design a reliable high-throughput assay that used it. AHDSred is exquisitely sensitive to oxidation by O2, forcing us to use fumarate and nitrate, both of which cannot directly oxidize AHDSred, but instead require S. oneidensis as an intermediate.

Among 149 hits, 18 mutants were slow or failed at oxidation of AHDSred in only the fumarate assay, 50 mutants, in only the nitrate assay, and 81 mutants in both assays. In total, 109 of these mutants were grouped by gene annotation into functional categories that satisfactorily explain the slowing or failure of AHDSred oxidation (Fig. 2c). For example, disruption of the periplasmic fumarate reductase (δfccA; we refer to transposon-disruption mutants with δ, and gene-deletion mutants with Δ) eliminates AHDSred oxidation when using fumarate as a terminal electron acceptor. Detailed time courses of AHDSred oxidation for selected anticipated hits from the genome-wide screen are shown in Fig. S2. Of note, 41 of the AHDSred oxidation-deficient mutants could not be assigned to an established functional category, suggesting that their function might be more directly involved in electron uptake (Fig. 2c). AHDSred oxidation time courses for knockout mutants, where we later observed a cathode phenotype, are shown in Fig. 2d–g, along with those for mutants with disruptions in adjacent genes. Detailed time courses for these mutants are shown in Fig. S3.

Electrochemical measurements confirm robust EEU phenotype of five unknown function mutants

We selected 17 of the 41 “unknown function” S. oneidensis AHDSred oxidation-deficient mutants for further on-electrode testing. These mutants were chosen for annotations that indicated possible redox activity (e.g., δSO_3662), interaction with the quinone pool (e.g., δSO_0362, δSO_0400), along with mutants with no functional annotation. To exclude genes involved in solely in terminal electron-acceptor utilization (the very end of the electron-transport chain), we used a different terminal electron acceptor (O2) than in the AHDSred oxidation screen. The use of O2 also ensures that the genes identified are part of the overlapping electron-uptake pathway, rather than previously unidentified components of fumarate/nitrate reduction. We confirmed this using δfccA, δnapA and δnapG as negative controls as these genes encode anaerobic terminal reductases that we did not expect to disrupt electron uptake using O2 as a terminal electron acceptor. We also selected three positive control mutants of genes known to be involved in EET (δcymA, δmtrA and δmtrC) and one expected negative control based on AHDSred oxidation screen results (δSO_0401). δSO_0401 was chosen as it is adjacent to a hit (δSO_0400) in the AHDSred oxidation assay, but does not itself produce an oxidation phenotype.

Biofilms of each of the mutants were grown on ITO working electrodes in a three-electrode bio-electrochemical system13. For analysis of electron uptake, the working electrodes were poised at −378 mV vs. the standard hydrogen electrode (SHE). Significant negative currents (i.e., electrons flowing from the working electrode to the biofilm/solution) were only observed in the presence of O2 as a terminal electron acceptor. To quantify the amount of negative current due to biological vs. nonbiological processes, the electron-transport chain was inhibited at the end of each experiment with the ubiquinone mimic, Antimycin A and the remaining abiotic current was measured (Fig. S4). Each mutant was tested in at least three replicate experiments.

Most of the 17 mutants of unknown function demonstrate a limited-to-modest change in average electron uptake from the working electrode (Figs. 3a, S5a, S5c, S5d, and Table S1). As expected, mutants that disrupt components of the well-known Mtr EET complex produce significant reductions (p value < 0.05) in electron uptake, except for cymA10,13. Though cymA was previously shown to be important under anaerobic cathodic conditions10, only a small reduction in electron uptake was noted under aerobic conditions, consistent with previous results13. It is plausible that the other unknown genes tested that did not generate a cathodic phenotype play a previously uncharacterized role in one of the other subcategories highlighted in the AHDSred assay rather than electron uptake, such as the reduction of fumarate or nitrate, as opposed to O2.

Fig. 3: Electrochemical measurements confirm robust extracellular electron-uptake phenotype for 5 S. oneidensis mutants identified by our high-throughput screen (δSO_0181, δSO_0400, δSO_0841, δSO_03660 and δSO_03662; (δ indicates a transposon-insertion mutant)).
figure 3

δSO_03662 was implicated in electron uptake by the AHDSred oxidation assay. However, at the time of collecting data for panels a, b, we were unable to retrieve δSO_03662 from the S. oneidensis whole-genome-knockout collection, so later constructed a clean deletion mutant to collect data for panel c. Hence, we talk about four mutants in panels a, b, but 5 in panel c. We were later able to retrieve δSO_3662 mutant and retested it in the AHDSred oxidation assay prior to submitting this article. The δmtrA control mutant in panels a, b shows significant derivation from wild type, but the current range is too small to be drawn clearly, so is marked with arrows. a We measured the cathodic biological current for 18 S. oneidensis mutants that produced AHDSred oxidation failure for unknown reasons, control mutants (δfccA, δmtrA, δmtrC, and δnapG), and wild type (WT). In total, 6 gene-disruption mutants, four of them previously uncharacterized, produced robust reductions in cathodic current. The region corresponding to wild-type currents and one standard deviation is shaded in blue. ANOVA results indicated that there was a significant variation in means of the strains from one another (F-value = 5.94; Pr(>F) = 2.34 × 10−9), while individual comparisons specifically revealed δSO_0181, δmtrA, δSO_0841, δSO_0400, δSO_3660, and δmtrC to all display significantly lower cathodic currents (p < 0.05) than WT. b However, three of the six robust electron-uptake mutants (three of the four previously uncharacterized disruption mutants) do not have a significant effect on anodic current production (electron deposition), the well-characterized phenotype of S. oneidensis. The range of wild-type current values and a standard deviation from the mean is shaded in blue. Analysis of anodic current values (F-value = 5.801, Pr(>F) = 2.10 × 10−10) revealed that of the four previously uncharacterized disruption mutants, only δSO_0841 shows anodic current production different from WT (p < 0.05). c Gene deletion (indicated by Δ) confirms the electron uptake phenotype of five mutants. Complementation of the deleted gene restores electron uptake phenotype. Further information on electrochemical methods can be seen in Materials and methods. Data are available in ref. 53.

Disruption mutants in four genes that code for proteins that are of unknown function (δSO_0181, δSO_0400, δSO_0841 and δSO_3660) demonstrated a highly significant reduction in electron uptake (p value < 0.05) (Fig. 3a). At the time of collecting data for Fig. 3a, b, we were unable to retrieve δSO_03662 from the S. oneidensis whole-genome-knockout collection. To further verify that the reductions in current seen in δSO_0181, δSO_0400, δSO_0841 and δSO_3660 were due to the loss of the disrupted gene and not due to any polar effects, we made a clean deletion for each gene, which all demonstrated a decreased electron-uptake phenotype compared with wild-type (Fig. 3c). We also constructed a clean deletion mutant for SO_3662SO_3662) to collect data for Fig. 3c. ΔSO_3662 (annotated as an inner membrane ferredoxin), demonstrated both an AHDSred oxidation phenotype (Fig. 2g and Fig. S3) and a cathodic phenotype (Fig. 3c). Furthermore, complementation of the knockout mutants with a plasmid encoding the deleted gene restored electron uptake function in all mutants (Fig. 3c).

We were later able to retrieve δSO_3662 mutant and retested it in the AHDSred oxidation assay prior to submitting this article. In addition, while retesting δSO_3660 and δSO_3662, we tested gene-disruption mutants for the surrounding genes and found that these also disrupted AHDSred oxidation (Fig. 2g). However, as of the time of writing, we have not yet validated these mutants by electrochemical measurements.

Electron uptake disruption is not due to biofilm formation failure

Electron-uptake disruption in these five gene-deletion mutants cannot be explained by changes to biofilm production or cell growth rate (Tables S2 and S3). We used protein quantification to ensure that the cathode phenotypes we observed were not caused by deficiencies in biofilm formation due to direct interference with biofilm formation or failure of anode conditioning prior to cathodic current measurement. In an earlier work13, we screened mutants of the known EET pathway (known anodic phenotypes) for cathodic phenotypes. We imaged all the biofilms, performed cell counts, and quantified the total protein from the electrode biofilms13. We observed no evidence of differences between the biofilm-formation capability of the mutants we screened compared with wild-type. While protein measurements were variable, they were no less so than microscopy-count data. Thus, we adopted protein quantification to measure biofilm-cell abundance.

No statistically significant difference in on-electrode protein abundance was seen for any of the mutants tested (Table S1). Furthermore, no statistically significant difference in growth rate between wild-type S. oneidensis and any mutant was observed in minimal media under aerobic or anaerobic conditions, with the exception of a longer lag time observed for complementation mutants containing an extra plasmid and grown in the presence of chloramphenicol and/or kanamycin (Table S2 and Fig. S6).

Four of the five genes are only involved in electron uptake

In addition to electron uptake, we also analyzed the extracellular electron deposition of each of the 24 mutants selected for on-electrode testing (Fig. 3b). Of the five EEU genes identified here (SO_0181, SO_0400, SO_0841, SO_3660 and SO_3662), only disruption of SO_0841 significantly reduces both electron uptake and deposition. The δSO_0841 disruption mutant produces half the positive current of wild-type S. oneidensis, similar to the effect of disruption of mtrA and mtrC (Table S1 and Fig. 3b). Interestingly, no growth phenotype was observed for cells grown on soluble (iron citrate) or insoluble iron (iron oxyhydroxide), suggesting that SO_0841 does not play a significant role in EET to minerals.

Sequence analysis suggests gene functions and shows electron uptake genes are widely distributed across species

Phylogenetic analysis of the five EEU genes identified here suggests that they are broadly conserved across Shewanella species and across numerous clades of the Gammaproteobacteria. Phylogenetic trees are shown in Figs. S7S12. Metadata for trees are attached in Supplementary Data 2.

SO_0841

Sequence analysis suggests that SO_0841 is involved in cell signaling. SO_0841 encodes a transmembrane protein with a 250-amino-acid long periplasmic region and a 250-amino-acid long cytoplasmic domain containing a GGDEF c-di-GMP signaling domain. As a phenotype was only observed on electrodes, and not iron minerals, the role of this protein in traditional EET remains unclear. The GGDEF domain is often used for regulation of biofilm formation and cell motility28; however, disruption of SO_0841 had no on impact on biofilm or cell morphology (Table S1 and Fig. S7).

This suggests a more specific role for SO_0841 in on-electrode EET. SO_0841 also has a broad distribution of homologs ranging across the Proteobacteria (Fig. S8). Though electron-uptake phenotypes have not been tested in a wide range of organisms, homologs of SO_0841 are found in electrochemically active microbes with the capacity for EEU, including Mariprofundus ferrooxydans29, Idiomarina loihiensis30 and Marinobacter species30. Thus, we speculate that this gene may have a conserved role in electron uptake.

The remaining four genes identified all play significant roles in electron uptake, but have no detectable role in electron deposition (Fig. 3a, b).

SO_0181

SO_0181 is predicted to be membrane-associated and contains a putative nucleoside triphosphate-binding and/or hydrolase domain (suggesting that it interacts with ATP or GTP). Furthermore, SO_0181 is located immediately upstream of phyH which encodes an uncharacterized putative oxidoreductase (part of a family of bacterial dioxygenases with unconfirmed activity in S. oneidensis), which also demonstrated an electron-uptake phenotype in the AHDSred oxidation screen (Supplementary Data 1 and Fig. 2e).

There is little bioinformatic support for a direct role of SO_0181 in redox chemistry. However, a role associated with activating or modifying the other redox-active proteins involved in EEU seems feasible but needs to be further investigated.

A phylogenetic tree constructed from the 200 closest relatives of SO_0181 shows a distinct clade specific to the Shewanellae (90–100% sequence identity) (Fig. S9). Closely related clades of SO_0181 are found in the Pseudomonas, Cellvibrio and Hahella genera of the Gammaproteobacteria, though homologous gene clusters are also seen in Beta- (Delftia and Acidovorax) and Delta-proteobacterial (Archangium and Cystobacter) lineages. Notably, several of the genera with close homologs of SO_0181, including Delftia and Pseudomonas, have EET-capable representatives within them.

SO_0400

SO_0400 belongs to a superfamily of quinol-interacting dimeric monooxygenses (dimeric α–β-barrel superfamily SSF54909). Of the homologs of known function, SO_0400 is most closely related to the YgiN quinol monooxygenase in E. coli31. Structural analysis of YgiN suggests that it interacts with the semiquinone state of quinols and suggests the existence of an unusual quinone redox cycle in E. coli31. Additionally, deletion of ygiN in E. coli inhibits the transition between aerobic and anaerobic growth32. Interestingly, deletion of SO_0400 does not inhibit the transition between aerobic and anaerobic growth conditions (and vice versa) in S. oneidensis (Fig. S6). Furthermore, deletion of SO_0400 did not affect the sensitivity of S. oneidensis to oxygen-free radicals using a disc-diffusion assay with hydrogen peroxide (data not shown). These data suggest a possible previously uncharacterized function in this quinol monooxygenase.

SO_0400 has a very broad distribution of close homologs found in the Proteobacteria, Gram-positive Actinobacteria, Bacteriodetes and Archaeal Methanobrevibacter (Fig. S10). Within the Shewanellae, homologs of this quinol monooxygenase are both broadly distributed and tightly clustered in a highly conserved clade, with many homologs exhibiting 95–100% amino-acid-sequence identity. This may speak to the highly conserved function of this gene within the Shewanellae that is possibly distinct from the other homologs observed in other genera.

SO_3660, SO_3662, and an electron uptake operon

The AHDSred oxidation screen points to the existence of an electron-uptake operon in S. oneidensis. Disruption of any of the loci from SO_3656 to SO_3665 causes failure of AHDSred oxidation (Fig. 2g). This putative operon is putatively regulated by SO_3660, annotated as a transcriptional regulator. SO_3662 is annotated as an inner-membrane ferredoxin, supporting a direct role in electron transfer. In-frame deletions of the genes coding for SO_3660 and the putative inner-membrane-bound ferredoxin SO_3662 both disrupt electron uptake (Fig. 3c), but not deposition. Interestingly, the formal potentials quantified for each deletion mutant were statistically indistinguishable (Table S2).

Phylogenetic trees constructed from the 200 closest homologs of SO_3600 and SO_3662 revealed that these genes are highly conserved in Gammaproteobacteria and across numerous (≈100) Shewanella species (Fig. S11 and S12). SO_3662 appears to be highly conserved among the order Alteromodales in particular.

Role of electron uptake genes in S. oneidensis

Like many other facultatively anaerobic microorganisms, including E. coli33, S. oneidensis employs discrete anaerobic and aerobic electron-transport chains. Menaquinone is the dominant quinone used by S. oneidensis under anaerobic conditions where EET is used for mineral respiration23. Conversely, under aerobic conditions, ubiquinone is the dominant quinone used by S. oneidensis34. Furthermore, ubiquinone is important for reverse electron flow to NADH mediated by Complex I (a NADH:ubiquinone oxidoreductase) under cathodic conditions13.

When taken together, the normally discrete machinery of aerobic and anaerobic electron-transport chains and the ability of S. oneidensis to couple reversal of the anaerobic EET pathway to O2 reduction suggests that a specific connection between the two transport chains is likely to exist. However, to our knowledge, the mechanism allowing these organisms to transition between one electron transport chain, or quinone pool to another, is poorly understood.

We outline two possible mechanisms for a connection between the anaerobic EET pathway and the aerobic electron-transport chain in EEU in Fig. 1. First, the EET complex could transfer electrons to CymA, which then reduces menaquinone. Electrons could then be transferred from menaquinone to ubiquinone and into the aerobic electron-transport chain, finally arriving at Complex IV where they reduce O2. This option seems intuitive as under anaerobic conditions using fumarate as an electron acceptor, S. oneidensis was shown to require menaquinone in addition to several components of the EET complex to uptake cathodic currents10. However, knockout of the gene coding for cymA does not disrupt cathodic electron uptake when using O2 as a terminal electron acceptor13.

The lack of involvement of CymA in cathodic electron uptake under aerobic conditions suggests a second option (Fig. 1): that cathodic electrons bypass the menaquinone pool under aerobic conditions.

We speculate that the putative quinol-interacting protein SO_0400 and the putative ferredoxin SO_3662 (and possibly proteins coded by nearby genes) are directly involved in connecting the reverse EET pathway during electron uptake and the aerobic electron-transport chain. Notably, the lack of a phenotype for most of these proteins under anodic conditions supports the hypothesis of a distinct connection between a subset of EET proteins and the aerobic electron-transport chain during EEU (Fig. 1). This work has also highlighted some genes potentially involved in cell signaling or transcriptional responses that may help aid in facilitating electron uptake under specific conditions (SO_0841, SO_0181 and SO_3660).

Though the motivation of this work stemmed from the application of S. oneidensis to electrosynthesis, it is likely that the process of EEU plays a role in Shewanella physiology and ecology. It has been shown that minerals in nature can serve a capacitive or electron-storing function for microbes35. While electron deposition and electron uptake from minerals such as magnetite were shown to function as both sinks and sources for different metabolisms, it is feasible that a single organism with both functionalities could utilize minerals in this way—functionally storing charge akin to a battery. Though iron oxidation has only been demonstrated in a single Shewanella strain36, this could be due to the challenge of distinguishing biologic and abiotic iron oxidation in the absence of growth. As Shewanella are not generally capable of carbon fixation, the process of EEU is unlikely to have evolved as a growth-linked metabolism. However, previous work has linked electron uptake to maintenance of cell biomass or decreasing the rate of cell death, which could suggest a role in allowing cells to conserve energy under nongrowth conditions13. Interestingly, these genes appear to be widely conserved across the Shewanella, as well as other marine Gammaproteobacteria (several of which have also been implicated in EEU). This supports the yet-unexplored ecological role for EEU in sediment and/or marine microbes, though our knowledge of the specific activity and role of this process is still at its inception.

Conclusions

EEU holds significant potential for conversion of CO2 and renewable electricity to complex organic molecules. While potential of unusual phenotypes like this can be limited by a lack of genetic understanding, especially in nonmodel organisms, synthetic biology can greatly expand the possibility for their improvement and application.

We used a whole-genome-knockout collection previously built with the rapid, low-cost knockout Sudoku method to screen the S. oneidensis genome for redox dye oxidation, a proxy for electron uptake. In this work, we have performed detailed electrochemical analyses, focusing on genes encoding proteins of unknown function. We have identified five previously uncharacterized genes in S. oneidensis that are involved in EEU from both solid phase and extracellular donors, coupled to both aerobic and anaerobic terminal electron acceptors.

Identification of these important genes lays the foundation for further genetic characterization of metal oxidation in nature, improvement of EEU in S. oneidensis and for synthetically engineering an electron-uptake pathway into easily engineered or synthetic microbes for powering recent advances in synthetic CO2 fixation15 and EET37 in E. coli.

Methods

Genome-wide AHDSred oxidation screen

The S. oneidensis whole genome knockout collection20,21 was screened for members unable to oxidize the redox dye anthra(hydro)quinone-2,6-disulfonate (AHDSred for the reduced form and AQDSox for the oxidized form)23,24, and subsequently reduce either fumarate22 or nitrate.

Knockout collection construction

The S. oneidensis whole-genome-knockout collection was previously built with the Knockout Sudoku whole-genome-knockout collection construction procedure20,21. Prior to high-throughput screening, the mutant collection was duplicated with a multiblot replicator (Catalog Number VP 407, V&P Scientific, San Diego CA, USA) into 96-well polypropylene plates containing 100 μL of LB with 30 μg mL−1 kanamycin per well. The plates were labeled with barcodes and registration marks for identification in high-throughput analysis. The plates were sealed with an air-porous membrane (Aeraseal, Catalog Number BS-25, Excel Scientific) and grown to saturation overnight (at least 24 h) at 30 °C with shaking at 900 rpm.

AHDSred preparation

Solutions of 25 mM AHDSred for screening experiments were prepared electrochemically. About 200–1000-mL batches of 25 mM AQDSox were prepared by dissolving 10.307 mg of AQDSox powder (Catalog no. A0308, TCI America) per 1 mL of deionized water at 60 °C. The AQDSox solution was then transferred to a three-electrode electrochemical system (Catalog no. MF-1056, BASI Bulk Electrolysis Cell) inside a vinyl anaerobic chamber (97% N2, 3% H2 and <20 ppm O2; Coy Laboratory Products, Grass Lake MI, USA). The system uses an Ag/AgCl reference electrode, a coiled Pt wire counterelectrode inside a fritted counterelectrode chamber and a reticulated vitreous carbon working electrode. The working-electrode potential was maintained at 700 mV vs. Ag/AgCl with a digitally controlled potentiostat (PalmSens, EmStat3). To enhance cell conductivity, 3 M H2SO4 acid was added to the working electrode, allowing the cell current to rise to ≈2 mA. AHDSred reduction is assumed to be complete when the solution is translucent yellow in color, typically after ≈7 h. The pH of the AHDSred stock solution was returned to 7.2 by addition of 3 M NaOH and the color changed from yellow to yellow-orange.

Assay media preparation

Shewanella basal media (SBM) was used for all AHDSred oxidation assays. SBM consists of ammonium chloride (NH4Cl) (0.0086 M), dibasic potassium phosphate (K2HPO4) (0.0013 M), monobasic potassium phosphate (KH2PO4) (0.0017 M), magnesium sulfate heptahydrate (MgSO4.7H2O) (0.0005 M), ammonium sulfate ((NH4)2SO4) (0.0017 M), and HEPES (0.1 M). The media was buffered to pH 7.2 with sodium hydroxide (NaOH) and sterilized by autoclave. Trace-mineral supplement (5 mL L−1; Catalog no. MD-TMS, American Type Culture Collection (ATCC), Manassas, VA, USA) and vitamin supplement (5 mL L−1; Catalog no. MD-VS, ATCC) were added aseptically. No carbon source (e.g., lactate) is added to the SBM to ensure that electrons flowing into the cell are only from AHDSred.

Solutions of 25 mM potassium nitrate and 25 mM sodium fumarate adjusted to pH 7.2 and filter-sterilized were prepared as terminal electron acceptors for the AHDSred oxidation screens. These solutions were transferred to an anaerobic chamber at least the evening prior to any experiment to allow deoxygenation.

AHDSred oxidation screens with nitrate

Aliquots of 10 μL of culture of each mutant from the replicated Shewanella whole-genome-knockout collection were transferred to 96-well polystyrene assay plates filled with 50 μL of SBM using a 96-channel pipettor (Liquidator 96, Mettler-Toledo Rainin LLC, Oakland, CA, USA). The assay plates were transferred to an anaerobic chamber for de-oxygenation for at least 9 h. The cell activity was confirmed by transferring cells from the assay plate to LB after the 9 h resting period in the anaerobic chamber and observing growth.

Following deoxygenation, 40 μL of 1:1 mixture of 25 mM AHDSred and 25 mM potassium nitrate were added to each well of the assay plates with a multichannel pipettor. The final concentration of both AHDSred and potassium nitrate in each well was 5 mM. Each assay plate was immediately photographed after media addition with the macroscope. All plates were repeatedly imaged for at least ≈40 h. After completion of the experiment, photographs were analyzed with the macroscope image analysis software. Mixing instructions for a single well of a 96-well assay plate are shown in Table 1.

Table 1 Solution mixing instructions for AHDSred oxidation assay for each well in a 96-well plate.

AHDSred oxidation screen with fumarate

Aliquots of 1 μL of culture of each mutant from the replicated S. oneidensis whole-genome-knockout collection were transferred to 96-well polystyrene assay plates filled with 59 μL of SBM using a multiblot replicator (Catalog Number VP 407, V&P Scientific, San Diego CA, USA). The assay plates were transferred to an anaerobic chamber for deoxygenation for at least 9 h. The cell activity was confirmed after the resting period by subculturing a small amount of assay media in LB. AHDSred oxidation was found to proceed much faster with fumarate than with nitrate, so the volume of cells added to the assay was reduced to make data collection more manageable.

Following deoxygenation, 40 μL of 1:1 mixture of 25 mM AHDSred and 25 mM sodium fumarate were added to each well of the assay plates with a multichannel pipettor. The final concentration of both AHDSred and sodium fumarate in each well was 5 mM. Each assay plate was immediately photographed after media addition with the macroscope. All plates were repeatedly imaged for at least ≈40 h. After completion of the experiment, photographs were analyzed with the macroscope image analysis software. Mixing instructions for a single well of a 96-well assay plate are shown in Table 1.

Macroscope data acquisition system

An automated photographic data-acquisition system was used to record the results of AHDSred oxidation assays. The device consists of a digital single-lens reflex (DSLR) camera (any member of the Canon Rebel Series) controlled by a macOS computer running a custom data-acquisition program that downloads images to the computer and timestamps them. The camera shutter is controlled by a single-switch foot pedal (vP-2, vPedal), leaving both hands free to manipulate 96-well plates. The camera is mounted to a frame constructed with extruded aluminum rails (T-slot). Barcodes attached to each plate enable images to be automatically sorted. Registration marks on the barcodes allow for identification of well positions and each well to be associated with a specific gene-knockout mutant. The device allows a stack of 200 96-well plates to be imaged in ≈15 min. This process can be repeated immediately, allowing each plate to be quasi-continuously imaged. A set of photographs of three generations of macroscope device, along with photographs taken with the macroscope, are shown in Figs. S13S17.

This work has used two variants of the device: the first to image AHDSred oxidation in transparent 96-well assay plates, and the second to measure bacterial growth in 96-well storage plates covered with an air-porous membrane (Aeraseal, Catalog Number BS-25, Excel Scientific). In the first configuration, the camera is mounted above the AHDSred oxidation-assay plates. Each plate is placed inside a laser-cut acrylic holder and illuminated by an LED light pad from below (A920, Artograph). Barcodes are printed on transparent labels (Catalog no. 5660, Avery) and attached to the top of each plate. In the second configuration, the camera is mounted below the plate and illumination is provided from above with an LED light pad. A white barcode (SIDE-1000, Diversified Biotech, Dedham, MA, USA) is attached to the side of the plate and viewed through a 45° right-angle mirror (Catalog no. 47–307, Edmund Optics, Barrington, NJ, USA).

Analysis of macroscope images

A custom Macroscope Image Analyzer program was developed with Pyzbar38, Pillow39, Numpy40, Matplotlib41 and OpenCV-Python42 using Python3. The program was used to process images taken with the macroscope device and aided in detection of loss-of-function mutants. The program handled the image analysis in four steps: creating barcodes, organizing the images collected with the macroscope by barcodes, collecting the data from the images, and finally presenting the data for analysis.

An additional image-analysis algorithm was developed with SciKit Image43 and SimpleCV44 to test images of 96-well storage plates for cross-contamination and growth-failure events by comparison with the collection catalog. The image-analysis algorithm updated the record for each well in the collection catalog with growth information to assist in rejection of false positives in the AHDSred oxidation screen due to growth failure.

The Macroscope Image Analyzer software is available on GitHub at https://github.com/barstowlab/macroscope-image-analyzer.

As AHDSred is oxidized to AQDSox, it changes color from yellow–orange to clear. Almost all information on the reduction state of the AHDSred/AQDSox dye can be found in the blue-color channel of the assay plate images. At the start of the assay, the intensity of the blue-color channel is low, and the dye is orange. As the AHDSred/AQDSox dye is oxidized and becomes clear, the intensity of the red channel remains approximately constant, with a small increase in green-channel intensity and a large increase in the blue-channel intensity. However, we found reporting the blue channel intensity as a proxy for the AHDSred/AQDSox redox state to be unintuitive.

To aid the reader and experimenter, we used the RGB color data,

$${\overrightarrow{C}} = \left[\begin{array}{c}r\\ g\\ b\end{array}\right]$$

to calculate a single number that represents how “yellow” a well is. The vector overlap (dot product) was calculated between the current color of the well and the most saturated yellow color in the assay photographic dataset.

The reference yellow color,

$${\overrightarrow{Y}}_{\!0} = \left[\begin{array}{c}225\\ 153\\ 0\end{array}\right]$$

was calculated relative to the reference white color,

$${\overrightarrow{W}}_{\!0} = \left[\begin{array}{c}255\\ 255\\ 255\end{array}\right]$$

Thus, the transformed yellow reference,

$${\overrightarrow{Y}}_{0}^{{\prime} }={\overrightarrow{Y}}_{0}-{\overrightarrow{W}}_{0}.$$

The transformed well color, relative to the white reference,

$${\overrightarrow{C}}^{\prime} ={\overrightarrow{C}}-{\overrightarrow{W}}_{0}.$$

The yellow intensity was calculated by normalizing the dot product between the transformed well color and the transformed yellow reference,

$$y=\frac{ {\overrightarrow{C}} ^{\prime} {\overrightarrow{Y}}_{0}^{{\prime} }}{{|{\overrightarrow{Y}}_{0}^{{\prime} }|}^{2}}.$$

The normalized yellow intensity has a maximum value of 1 when the well is yellow and a minimum value of 0 when it is clear.

The time series of colors for each gene shown as colored circles above each gene in Fig. 2 were generated by an algorithm that interpolated the multireplicate average of mean well-center color values for that mutant at 0, 1, 2, 3 and 4 h after the initiation of the oxidation experiment. A visual explanation of this process is shown in Fig. S1. AHDSred oxidation rates reported in Fig. S5 were calculated by a linear fit to the linear portion of the yellow intensity curve with Datagraph (Visual Data Tools).

Bioelectrochemical cell construction and experimental conditions

A 3-electrode electrochemical cell based on a design by Okamoto et al.45 was assembled in-house, with the exception of a salt bridge that was included to contain the reference electrode (Part no. MF-2031, BASi, West Lafeyette, IN, USA). As described, the cell consisted of a working electrode made of ITO- (indium tin oxide) plated glass (Delta Technologies, Ltd., Loveland, CO, USA), a counterelectrode of platinum wire and an Ag/AgCl reference electrode suspended in 1 M KCl (HCH Instruments, Inc., Austin, TX, USA). The reactor volume contained approximately 20 mL of liquid with a working-electrode surface area of 10.68 cm2.

The electrochemical cell was used for chronoamperometry (CA) experiments that measure change in current over time and cyclic voltammetry (CV) experiments that measure current in response to a change in voltage. Both types of experiment were controlled with a 16-channel potentiostat (Biologic, France). In anodic CA experiments, the working electrode was poised at 422 mV vs. SHE. In cathodic CA experiments the working electrode was poised at −378 mV. This voltage minimizes hydrogen production while maximizing the biological cathodic activity. In CV experiments, the working-electrode potential was scanned between +422 mV and −378 mV vs. the SHE at a rate of 1 mV s−1.

Shewanella culturing conditions

Cultures of wild-type S. oneidensis and S. oneidensis mutants were grown from glycerol stocks overnight in Luria Broth (LB) prior to each experiment. Aerobic and anaerobic growth curves and electrochemical experiments for all strains were performed in a Shewanella defined media (SDM)46, which we have found optimal for electrochemical experiments (note that SDM is not the same as SBM used in AHDSred oxidation assays). Aerobic cultures were performed in 50 mL volumes using 10 mM lactate as an electron acceptor shaking at 150 rpm at 30 °C. Kanamycin (Kan) and chloramphenicol (Chl) were added to LB and SDM media at concentrations of 100 μg mL−1 and 34 μg mL−1, respectively, for selection of transposon (Kan), clean gene deletion (Kan), and complementation (Chl + Kan) strains. The same growth conditions were used for anaerobic growth curves, with the exception that the media contained 20 mM fumarate and was purged with nitrogen gas for 10 min in serum vials.

For growth curves, strains were inoculated from overnight at a 100-fold dilution. Optical densities were recorded for triplicate cultures at 600 nm every 2–3 h. For cathodic growth, an overnight culture was backdiluted by a factor of 100 into SDM with 10 mM lactate and grown overnight. The overnight cultures grown in SDM were pelleted and resuspended in fresh SDM to an optical density at 600 nm of 0.1. About 20 mL of the resuspended culture was transferred to the working-electrode chamber of an electrochemical reactor. The reactor was attached to a 16-channel potentiostat (BioLogic) and the culture was anode-conditioned by poising the working electrode at +422 mV vs. SHE12,13. Anaerobic conditions needed to encourage biofilm formation and anodic current generation were maintained by continuous purging with N2.

After approximately 24 h, the reactors were detached from the potentiostat and the media containing planktonic cells was carefully removed to avoid disturbing the biofilm on the working electrode. The reactor was then refilled with 20 mL of fresh carbon-free SDM13. The reactors were then reattached to the potentiostat and the working electrode was cathodically poised at −378 mV vs. SHE. Air was slowly bubbled into the reactors via an aquarium pump, until a steady stream was reached to provide a source of O2.

To determine the portion of the cathodic current due to biological processes, the respiratory inhibitor Antimycin A was added to the electrochemical cell working-electrode chamber to a final concentration of 50 μM. Currents reported are the average difference between the steady-state currents pre and post Antimycin additions. To control for the effects of DMSO (the solvent for Antimycin A), blank injections of DMSO were made to the electrochemical cell and had no impact on current production. To confirm the effect of Antimycin A on biological current, addition of Antimycin A added to sterile minimal media in a reactor was performed and shown to have no impact on current production (Fig. S4).

Protein collection and quantification

Protein quantification was used to assess the total biomass in bioelectrochemical experiments. At the end of an electrochemical experiment, the spent media from the reactor (~20 mL) was collected and the biofilm was scraped from the working electrode. Biomass was centrifuged at 8000 × g and the pellet was resuspended in 2 mL of 10% w/v NaOH. Total protein collected from the reactor was quantified using a Qubit TM Protein Assay Kit (Invitrogen, USA) according to the manufacturer’s protocol.

Biofilm imaging and cell counts

Biofilms from δSO_0841 and wild-type bioelectrochemical reactors were imaged on Nikon TI-E eclipse inverted microscope equipped with UV fluorescence. Cells were stained using the FM 4-64X lipid dye (Molecular Probes, Life Technologies Inc). At least 20 images were taken from each of three reactors for δSO_0841 and wild-type replicate experiments. Cell dimensions and fluorescence intensity of a biofilm cross section were performed using the Nikon NIS-Elements software.

Statistics and reproducibility

Uncertainties in bioelectrochemical measurements were calculated by analysis of at least three replicates for each bioelectrochemical experiment. All statistical analyses were performed in Excel and/or R.

Cathodic midpoint potentials were calculated from cyclic voltammogram (CV) scans. A subsection of the CV scan (between the aforementioned parameters found under electrochemical conditions) was separated into a forward scan (from 222 mV to −322 mV vs. SHE) and a reverse scan (−322 mV to 222 mV). The scan range was chosen to contain the voltage at which maximum current production was achieved. Because linearity could not be assumed in these data to generate a function for a first-derivative analysis, an alternative method was used to analyze the data directly. A cubic smoothing spline (spar = 0.70) was applied to the current data in R to remove noise, but still captures the general trend. An approximate derivative was then taken from these values. From this approximate derivative, the maximum current produced, and its corresponding voltage potential were found for both the forward and reverse scans, which were then averaged to find the midpoint potential. The midpoint potential from each replicate was then pooled and averaged to obtain the reported value.

Anodic and cathodic currents were determined by chronoamperometry (CA). An average of the final 100 data points of each CA scan was taken to determine the average current achieved for each biological replicate. The replicates were then averaged to determine the average current produced for each strain. Biological current was determined by subtracting the average current post antimycin addition from the average cathodic current prior to addition. To determine if the average current of any of the mutants was significantly different from the wild-type, the cumulative current data were compiled in Excel and then fitted to a linear model in order to perform a type-II analysis of variance (ANOVA) test with R. Following the ANOVA test, anodic and cathodic currents for all mutants were compared with wild-type by Tukey’s honestly significant difference (HSD) test to determine if any significant difference existed between them.

Phylogenetic analysis

Phylogenetic trees of relatives of the genes identified in this work were generated by search for homologs, homolog alignment and tree assembly by a maximum likelihood method. Approximately 120–200 homologs for each gene identified in this work were identified with the top homolog hit program that interrogates the Integrated Microbial Genomes and Microbiomes Database (https://img.jgi.doe.gov/)47,48. Homolog sets for each gene were aligned with the Muscle aligner49 with default parameters.

Phylogenetic trees were generated for each homology set by a maximum likelihood method with RAxML 8.2.1150. In total, 100 trees were generated for each set of homologs for bootstrapping. Tree images and taxonomic metadata (Figs. S7S11) were generated and annotated using the interactive Tree of Life (iTOL) program51.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The datasets generated during and analyzed during the current study are available from the corresponding authors (A.R. and B.B.) on reasonable request. As the macroscope device for high-throughput image acquisition evolved organically, a blueprint does not exist. However, should a reader wish to construct one, B.B. and M.B. would be happy to consult.

Code availability

The Macroscope Image Analyzer software is available at https://github.com/barstowlab/macroscope-image-analyzer and in ref. 52. Unfortunately, due to licensing restrictions on the camera-control libraries, we cannot distribute the macroscope data-acquisition software online, but we can share our source code on request.

References

  1. Rabaey, K. & Rozendal, R. A. Microbial electrosynthesis—Revisiting the electrical route for microbial production. Nat. Rev. Microbiol. 8, 706–716 (2010).

    CAS  PubMed  Article  Google Scholar 

  2. Barstow, B. Molecular mechanisms for the biological storage of renewable energy. Adv. Sci. Eng. Med. 7, 1066–1081 (2015).

    CAS  Article  Google Scholar 

  3. Nealson, K. H. & Rowe, A. R. Electromicrobiology: realities, grand challenges, goals and predictions. Microb. Biotechnol. 9, 595–600 (2016).

    PubMed  PubMed Central  Article  Google Scholar 

  4. Adesina, O., Anzai, I. A., Avalos, J. L. & Barstow, B. Embracing biological solutions to the sustainable energy challenge. Chem 2, 20–51 (2017).

    CAS  Article  Google Scholar 

  5. Claassens, N. J., Cotton, C. A. R., Kopljar, D. & Bar-Even, A. Making quantitative sense of electromicrobial production. Nat. Catal. 2, 437–447 (2019).

    CAS  Article  Google Scholar 

  6. Salimijazi, F., Parra, E. & Barstow, B. Electrical energy storage with engineered biological systems. J. Biol. Eng. 13, 1–21 (2019).

    Article  Google Scholar 

  7. Torella, J. P. et al. Efficient solar-to-fuels production from a hybrid microbial-water-splitting catalyst system. Proc. Natl Acad. Sci. USA 112, 2337–2342 (2015).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  8. Liu, C., Colón, B. C., Ziesack, M., Silver, P. A. & Nocera, D. G. Water splitting biosynthetic system with CO2 reduction efficiencies exceeding photosynthesis. Science 352, 1210–1213 (2016).

    CAS  PubMed  Article  Google Scholar 

  9. Salimijazi, F. et al. Constraints on the efficiency of engineered electromicrobial production. Joule 4, 2101–2130 (2020).

    CAS  Article  Google Scholar 

  10. Ross, D. E., Flynn, J. M., Baron, D. B., Gralnick, J. A. & Bond, D. R. Towards electrosynthesis in Shewanella: energetics of reversing the Mtr pathway for reductive metabolism. PLoS ONE 6, e16649 (2011).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  11. Hsu, L., Masuda, S. A., Nealson, K. H. & Pirbazari, M. Evaluation of microbial fuel cell Shewanella biocathodes for treatment of chromate contamination. RSC Adv. 2, 5844–5855 (2012).

    CAS  Article  Google Scholar 

  12. Okamoto, A., Hashimoto, K. & Nealson, K. H. Flavin redox bifurcation as a mechanism for controlling the direction of electron flow during extracellular electron transfer. Angew. Chem. Int. Ed. Engl. 53, 10988–10991 (2014).

    CAS  PubMed  Article  Google Scholar 

  13. Rowe, A. R. et al. Tracking electron uptake from a cathode into Shewanella cells: implications for energy acquisition from solid-substrate electron donors. mBio 9, 1–19 (2018).

    Article  Google Scholar 

  14. Tefft, N. M. & TerAvest, M. A. Reversing an extracellular electron transfer pathway for electrode-driven acetoin reduction. ACS Synth. Biol. 8, 1590–1600 (2019).

    CAS  PubMed  Article  Google Scholar 

  15. Gleizer, S. et al. Conversion of Escherichia coli to generate all biomass carbon from CO2. Cell 179, 1255–1263.e1212 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  16. Prévoteau, A., Carvajal-Arroyo, J. M., Ganigué, R. & Rabaey, K. Microbial electrosynthesis from CO2: forever a promise? Curr. Opin. Biotechnol. 62, 48–57 (2020).

    PubMed  Article  CAS  Google Scholar 

  17. Glaven, S. M. Bioelectrochemical systems and synthetic biology: more power, more products. Microb. Biotechnol. 12, 819–823 (2019).

    PubMed  PubMed Central  Article  Google Scholar 

  18. Rosenbaum, M., Aulenta, F., Villano, M. & Angenent, L. T. Cathodes as electron donors for microbial metabolism: which extracellular electron transfer mechanisms are involved? Bioresour. Technol. 102, 324–333 (2011).

    CAS  PubMed  Article  Google Scholar 

  19. Molderez, T. R., Prévoteau, A., Ceyssens, F., Verhelst, M. & Rabaey, K. A chip-based 128-channel potentiostat for high-throughput studies of bioelectrochemical systems: optimal electrode potentials for anodic biofilms. Biosens. Bioelectron. 174, 112813 (2021).

    CAS  PubMed  Article  Google Scholar 

  20. Baym, M., Shaket, L., Anzai, I. A., Adesina, O. & Barstow, B. Rapid construction of a whole-genome transposon insertion collection for Shewanella oneidensis by Knockout Sudoku. Nat. Commun. 7, 13270 (2016).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  21. Anzai, I. A., Shaket, L., Adesina, O., Baym, M. & Barstow, B. Rapid curation of gene disruption collections using Knockout Sudoku. Nat. Protoc. 12, 2110–2137 (2017).

    CAS  PubMed  Article  Google Scholar 

  22. Lovley, D. R., Fraga, J. L., Coates, J. D. & Blunt-Harris, E. L. Humics as an electron donor for anaerobic respiration. Environ. Microbiol. 1, 89–98 (1999).

    CAS  PubMed  Article  Google Scholar 

  23. Newman, D. K. & Kolter, R. A role for excreted quinones in extracellular electron transfer. Nature 405, 94–97 (2000).

    CAS  PubMed  Article  Google Scholar 

  24. Lies, D. P. et al. Shewanella oneidensis MR-1 uses overlapping pathways for iron reduction at a distance and by direct contact under conditions relevant for biofilms. Appl. Environ. Microbiol. 71, 4414–4426 (2005).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  25. Shyu, J. B. H., Lies, D. P. & Newman, D. K. Protective role of tolC in efflux of the electron shuttle anthraquinone-2,6-disulfonate. J. Bacteriol. 184, 1806–1810 (2002).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  26. Ueki, T. et al. Construction of a geobacter strain with exceptional growth on cathodes. Front. Microbiol. 9, 1512 (2018).

    PubMed  PubMed Central  Article  Google Scholar 

  27. Yee, M. O., Snoeyenbos-West, O. L., Thamdrup, B., Ottosen, L. D. M. & Rotaru, A.-E. Extracellular electron uptake by two methanosarcina species. Front. Energy Res. 7, 29 (2019).

    Article  Google Scholar 

  28. Ryjenkov, D. A., Tarutina, M., Moskvin, O. V. & Gomelsky, M. Cyclic diguanylate is a ubiquitous signaling molecule in bacteria: insights into biochemistry of the GGDEF protein domain. J. Bacteriol. 187, 1792–1798 (2005).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  29. Summers, Z. M., Gralnick, J. A. & Bond, D. R. Cultivation of an obligate Fe(II)-oxidizing lithoautotrophic bacterium using electrodes. mBio 4, e00420–00412 (2013).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  30. Rowe, A. R., Chellamuthu, P., Lam, B., Okamoto, A. & Nealson, K. H. Marine sediments microbes capable of electrode oxidation as a surrogate for lithotrophic insoluble substrate metabolism. Front. Microbiol. 5, 784 (2015).

    PubMed  PubMed Central  Article  Google Scholar 

  31. Adams, M. A. & Jia, Z. Structural and biochemical evidence for an enzymatic quinone redox cycle in Escherichia coli. J. Biol. Chem. 280, 8358–8363 (2005).

    CAS  PubMed  Article  Google Scholar 

  32. Portnoy, V. A. et al. Deletion of genes encoding cytochrome oxidases and Quinol Monooxygenase blocks the aerobic-anaerobic shift in Escherichia coli K-12 MG1655. Appl. Environ. Microbiol. 76, 6529–MG6540 (2010).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  33. Unden, G. & Bongaerts, J. Alternative respiratory pathways of Escherichia coli: energetics and transcriptional regulation in response to electron acceptors. Biochim. Biophys. Acta 1320, 217–234 (1997).

    CAS  PubMed  Article  Google Scholar 

  34. Nealson, K. H. & Little, B. Breathing manganese and iron: solid-state respiration. Adv. Appl. Microbiol 45, 213–239 (1997).

    CAS  Article  Google Scholar 

  35. Byrne, J. M. et al. Redox cycling of Fe(II) and Fe(III) in magnetite by Fe-metabolizing bacteria. Science 347, 1473–1476 (2015).

    CAS  PubMed  Article  Google Scholar 

  36. Philips, J. et al. A novel Shewanella isolate enhances corrosion by using metallic iron as the electron donor with fumarate as the electron acceptor. Appl. Environ. Microbiol. 84, e01154–01118 (2018).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  37. Jensen, H. M. et al. Engineering of a synthetic electron conduit in living cells. Proc. Natl Acad. Sci. USA 107, 19213–19218 (2010).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  38. Newby, A. & Hudson, L. Pyzbar, https://github.com/NaturalHistoryMuseum/pyzbar (2019).

  39. Clark, A. Pillow, https://python-pillow.org (2021).

  40. Harris, C. R. et al. Array programming with NumPy. Nature 585, 357–362 (2020).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  41. Hunter, J. D. Matplotlib: A 2D Graphics Environment. Comput. Sci. Eng. 9, 90–95 (2007).

    Article  Google Scholar 

  42. Mordvintsev, A. & Abid, K. OpenCV-Python Tutorials Documentation, https://media.readthedocs.org/pdf/opencv-python-tutroals/latest/opencv-python-tutroals.pdf (2014).

  43. van der Walt, S. J. et al. scikit-image: image processing in Python. PeerJ 2, e453 (2014).

    PubMed  PubMed Central  Article  Google Scholar 

  44. Demaagd, K., Oliver, A., Oostendorp, N. & Scott, K. Practical Computer Vision with SimpleCV (O’Reilly Media, 2012).

  45. Okamoto, A., Hashimoto, K. & Nakamura, R. Long-range electron conduction of Shewanella biofilms mediated by outer membrane C-type cytochromes. Bioelectrochemistry 85, 61–65 (2012).

    CAS  PubMed  Article  Google Scholar 

  46. Roh, Y. et al. Metal Reduction and iron biomineralization by a psychrotolerant Fe(III)-reducing bacterium, Shewanella sp. Strain PV-4. Appl. Environ. Microbiol. 72, 3236–3244 (2006).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  47. Chen, I. M. A. et al. IMG/M v.5.0: an integrated data management and comparative analysis system for microbial genomes and microbiomes. Nucleic Acids Res. 47, gky901 (2018).

    Google Scholar 

  48. Mukherjee, S. et al. Genomes OnLine database (GOLD) v.7: updates and new features. Nucleic Acids Res. 47, gky977 (2018).

    Google Scholar 

  49. Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  50. Stamatakis, A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313 (2014).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  51. Letunic, I. & Bork, P. Interactive tree of life (iTOL) v4: recent updates and new developments. Nucleic Acids Res. 47, gkz239 (2019).

    Article  CAS  Google Scholar 

  52. Jurgensen, L., Medin, S. & Barstow, B. Macroscope image analyzer. Zenodo, https://doi.org/10.5281/zenodo.5007192 (2020).

  53. Rowe, A. et al. Datasets for identification of a pathway for electron uptake in Shewanella oneidensis. Zenodo, https://doi.org/10.5281/zenodo.5013687 (2021).

Download references

Acknowledgements

We thank L. Jurgensen and S. Medin for programming assistance, B. Pian for experimental assistance, and A.M. Schmitz for review of this paper. This work was supported in the Barstow lab by a Career Award at the Scientific Interface from the Burroughs Welcome Fund, Princeton University startup funds, Cornell University startup funds and by U.S. Department of Energy Biological and Environmental Research grant DE-SC0020179. In the Rowe lab, this work was funded by Air Force Office of Scientific Research grant FA9550-19-1-0305 and University of Cincinnati startup funds. The Baym lab is supported by an award from the David and Lucille Packard Foundation.

Author information

Authors and Affiliations

Authors

Contributions

Conceptualization, A.R. and B.B.; Methodology, A.R., M.B. and B.B.; Investigation, A.R. F.S., L.T., J.S., O.A., I.A., L.K., M.B. and B.B.; Writing—Original Draft, A.R., L. T., F.S., and B.B.; Writing—Review and Editing, A.R., J.S., F.S. and B.B.; Funding Acquisition, A.R. and B.B.; Resources, A.R. and B.B.; Supervision, A.R. and B.B.; Data Curation, A.R., L.T., F.S. and B.B.; Visualization, A.R., F.S., L.T., J.S. and B.B. Formal Analysis, A.R., F.S., L.T., J.S. and B.B.

Corresponding authors

Correspondence to Annette R. Rowe or Buz Barstow.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Communications Biology thanks Sarah Glaven, Michaela TerAvest and the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Calvin Henard, Nicolas Fanget and Christina Karlsson Rosenthal. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Rowe, A.R., Salimijazi, F., Trutschel, L. et al. Identification of a pathway for electron uptake in Shewanella oneidensis. Commun Biol 4, 957 (2021). https://doi.org/10.1038/s42003-021-02454-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1038/s42003-021-02454-x

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing