Deciphering the Electron Transport Pathway for Graphene Oxide Reduction by Shewanella oneidensis MR-1

ABSTRACT We determined that graphene oxide reduction by Shewanella oneidensis MR-1 requires the Mtr respiratory pathway by analyzing a range of mutants lacking these proteins. Electron shuttling compounds increased the graphene oxide reduction rate 3- to 5-fold. These results may help facilitate the use of bacteria for large-scale graphene production.

The dissimilatory metal-reducing bacterium Shewanella oneidensis strain MR-1 has been extensively investigated for its ability to use insoluble substrates such as iron and manganese oxide minerals as terminal electron acceptors (9). The Mtr respiratory pathway is used by MR-1 to facilitate transfer of electrons from the interior of the cell to these external terminal electron acceptors (2,6,12,18). In addition to naturally occurring insoluble electron acceptors, it has been reported that MR-1 is able to reduce synthetic graphene oxide (GO), forming layered stacks of graphene (17,21). Because large-scale graphene production typically involves reduction of GO by the use of toxic chemicals (10), microbial reduction of GO to graphene by bacteria could offer an alternative approach that is rapid, cheap, and environmentally friendly. Understanding the molecular mechanisms of GO reduction by microbes may facilitate their use in the commercial-scale production of graphene.
It has been shown that both direct contact and electron shuttling are involved in the reduction of insoluble substrates by MR-1 (11,16). The Mtr pathway is required for the reduction of metal oxide minerals and electrodes (5,6), suggesting that it may also be able to transfer electrons to GO. At least five primary protein components have been identified in this pathway: OmcA, MtrC, MtrA, MtrB, and CymA. Current models of electron transfer in MR-1 assume that electrons from carbon source oxidation are passed via the menaquinone pool to the inner-membrane-anchored c-type cytochrome CymA. These electrons are then transferred to a periplasmic c-type cytochrome, MtrA, and eventually to outermembrane-anchored c-type cytochrome MtrC, which interact with each other via an integral outer membrane protein, MtrB (7). Two paralogs of MtrC, encoded by omcA and mtrF, can compensate for the loss of MtrC to reduce a variety of substrates (3,6). The activity of the Mtr respiratory pathway on insoluble substrates is accelerated by electron shuttles (5,19), either produced in situ by MR-1 (the riboflavin and flavin mononucleotide flavins) (13,20) or added exogenously to cultures such as 9,10-anthraquinone-2,6-disulfonic acid (AQDS) (5).
A recent investigation of GO reduction by MR-1 revealed the importance of some proteins in the Mtr pathway, such as MtrA and MtrB (17). However, CymA was found to be dispensable (17). This result was unexpected given that CymA is critical to extracellular respiration, acting as a linkage between the quinone pool and subsequent respiratory pathways in the periplasm and outer membrane (14,15). Here we describe a comprehensive genetic analysis of the relative importance of proteins in the Mtr respiratory pathway and test the role of electron shuttles in GO reduction, with some contradictory results presented with regard to CymA. The results presented here extend our knowledge of how Shewanella catalyzes the reduction of insoluble substrates and may have broader implications for biologically mediated graphene production.
The strains used in this study have been previously described a Corrected OD is the result of subtracting the culture OD from that of bacterial cells, which was extrapolated based on a standard curve of OD 600 and number of CFU. The correlation between the XPS results and the OD 600 values (Spearman correlation coefficient, 0.94) indicates that the corrected OD 600 is a faithful quantification for GO reduction. (5,6,8). Strains were grown in Shewanella basal medium (SBM) (8) anaerobically at 30°C with shaking at 220 rpm. Balch anaerobic tubes (1) were prepared in an anaerobic chamber and sealed with butyl rubber stoppers. Lactate (15 mM) and GO (0.8 mg/ml) were added as the sole electron donor and electron acceptor, respectively. Graphite oxide sheets were synthesized from graphite powder by the method of Hummers and Offeman (10). The aqueous graphite oxide solution was sonicated vigorously for 2 h to facilitate the separation of GO sheets (21). For the GO reduction assay, overnight aerobic LB cultures were washed in SBM and used for inoculation at a dilution of 1:1,000 after normalization based on the optical density at 600 nm (OD 600 ). Solutions of riboflavin and AQDS, a synthetic analogue of the redox-active moieties in humic acids, were prepared in distilled water, filter sterilized, and brought to a final concentration of 12 M. High concentrations of electron shuttles were used to maximize any potential effect on GO reduction. Cell growth was monitored by CFU determination by plating the cells onto LB agar plates and incubating them aerobically.
GO reduction was inferred by the change in OD 600 after correcting for bacterial cells (corrected OD 600 ), which was extrapolated based on a standard curve of OD 600 and number of CFU. By following GO reduction by MR-1 over time by both X-ray photoelectron spectroscopy (XPS) and OD 600 determination, we confirmed that the corrected OD 600 is a faithful quantification of GO reduction, providing a convenient method for quantifying GO reduction in a simpler and more precise manner than that of XPS (Table 1). For XPS sample preparation, cultures at different time points were collected, washed in 80% ethanol, 1 N HCl, and distilled water as described previously (17), and deposited and dried on silicon wafers. Table 1 shows a strong correlation, with a Spearman rank correlation coefficient of 0.94, between the relative percentage of C1s (the binding energy of electrons in the 1s atomic orbital of carbon) determined by XPS and the corrected OD 600 . Impaired GO reduction. To investigate the role of the Mtr respiratory pathway in GO reduction, we tested GO reduction by mutants defective in the Mtr pathway. Because both iron(III) oxide and GO are insoluble at circumneutral pH, we anticipated that the electron transport pathways would be sim-ilar. A steady increase in the amount of reduced GO was observed with the wild-type MR-1 strain for at least 40 h, whereas GO reduction was affected by various degrees in the mutant strains (Fig. 1). In contrast to a recent report by Salas et al. (17), we found that CymA is critical for GO reduction, since a strain lacking cymA retained only minimal GO reduction capability (Fig. 1A). Similarly, strains lacking either MtrA or MtrB also showed ϳ5-fold decrease in GO reduction activity (Fig. 1A), suggesting that MtrA and MtrB are also important for GO reduction (17).
Our cymA mutant result is consistent with how the Mtr respiratory pathway is currently thought to function (18), with CymA providing electrons from the menaquinone pool. An incorrect cymA mutant strain may explain the difference between our observations and those of Salas et al. (17). A recent report by Cordova et al. (4) demonstrated that cymA-null mutants could be suppressed by upregulation of sirCD, which could also explain the discrepancy.
To test potential terminal reductases for GO, we focused on three possible outer membrane c-type cytochromes: MtrC, OmcA, and MtrF (Fig. 1B). While a mutant strain lacking MtrF reduced GO only slightly more slowly than the wild type (85%), both the individual omcA and mtrC mutants showed significant decreases (about 50%) in GO reduction rate. A double mutant lacking both MtrC and OmcA reduced GO at a much lower rate (16%), similar to that of the triple mutant lacking MtrC, OmcA, and MtrF. Controls lacking cells or electron donor (lactate) showed only a minimal GO reduction rate (ϳ5%) during the course of incubation ( Fig. 1A; and data not shown). These results suggest that MtrC and OmcA are the primary terminal reductases for GO by MR-1 and that another minor GO reduction pathway may exist in parallel to the Mtr pathway. Acceleration by electron shuttles. To investigate whether an excess of extracellular electron shuttles could enhance GO reduction, we carried out the GO reduction experiment with the addition of riboflavin or AQDS. It should be noted that these compounds were added in addition to the normal level of flavins released by the bacteria. All strains tested naturally secreted flavins to ϳ1 M when grown in minimal medium (data not shown). We quantified the degree to which these compounds increased, i.e., accelerated, the GO reduction rate when they were added exogenously to wild-type cultures. When using an excess (12 M) of riboflavin or AQDS, we observed increases in the GO reduction rate of 2.7-and 4.7-fold, respectively (Fig. 2). AQDS enhanced GO reduction to a greater degree than riboflavin, suggesting that it reacts more quickly with GO. This observation is in contrast to the behavior of these electron shuttles in iron(III) oxide reduction (5,19). The significant acceleration of GO reduction in the presence of AQDS may have biotechnological implications for microbial graphene production.
We also tested the GO reduction of mutants defective in the Mtr pathway in the presence of additional electron shuttles (Fig. 3). Mutants of the Mtr pathway did not reduce GO at the same rate as the wild type in the presence of riboflavin or AQDS, indicating that this pathway is involved in the reduction of the electron shuttles during GO reduction. We attribute the decreases in OD 600 at later time points for some strains to the increase of graphene particle aggregation observed in older cultures (data not shown). While the relative GO reduction capability of the Mtr mutants with riboflavin or AQDS follows that of endogenous electron shuttles, it is worth noting that mtrB and mtrA mutants recovered significant levels of GO reduction activity in the presence of excess AQDS or riboflavin. In contrast, cymA mutant strains failed to recover any GO reduction activity in the presence of an electron shuttle, suggesting that the inner membrane cytochrome CymA is a key component in electron transfer to GO. Additionally, both the omcA mtrC double mutant and the omcA mtrC mtrF triple deletion mutant showed very little GO reduction activity even in the presence of AQDS or riboflavin, suggesting that the use of electron shuttles relies on these outer membrane cytochromes in MR-1.
Concluding remarks. We found that GO reduction by S. oneidensis MR-1 is catalyzed primarily by the Mtr respiratory pathway. While MtrA, MtrB, and CymA are important for GO reduction, outer membrane multiheme cytochromes MtrC and OmcA can provide partial compensation in the absence of one another, whereas MtrF is dispensable (see Fig. S1 in the supplemental material). Our CymA result contradicts a previous report concluding that CymA is not required for GO reduction by S. oneidensis (17). Electron shuttles, including riboflavin and AQDS, were able to significantly accelerate the GO reduction rate of MR-1, in a manner that was dependent on the Mtr pathway. This work expands the range of known substrates that can be reduced by S. oneidensis and may lead to engineered strains that can be used in the commercial production of graphene.