Uptake of self-secreted flavins as bound cofactors for extracellular electron transfer in Geobacter species†

Department of Applied Chemistry, The Univ Tokyo 113-8656, Japan. E-mail: hashimoto@ Interdisciplinary Research Organization, Kiyotake, Miyazaki 889-1692, Japan Departments of Earth Sciences and Biol California, Los Angeles, CA 90089, USA Biofunctional Catalysts Research Team, Science, 2-1 Hirosawa, Wako, Saitama 35 @riken.ac.jp † Electronic supplementary information (E DOI: 10.1039/c3ee43674h Cite this: Energy Environ. Sci., 2014, 7, 1357

Geobacter species are among the most efficient current-producing bacterial species, yet their electron-transfer mechanisms have been scarcely investigated at the molecular level. Here, we provide evidence that Geobacter cells secrete and utilize riboflavin as a bound-cofactor in outer-membrane c-type cytochromes. This finding highlights the potential roles of riboflavin as a major electron carrier in current production.
Electron transfer from cell metabolic systems to exterior solid substrates, termed extracellular electron transfer (EET), is an intriguing aspect of microbial respiration. 1,2 In anaerobic environments, particularly biolms, EET is a terminal step of catabolism and is involved in redox sensing 3 and intercellular 4 / interspecies 5 energy transfer. It is also a fundamental process in microbial communities involved in energy production (e.g., microbial fuel cells), 6 bioremediation of waste waters and contaminated sediments, 7 and anaerobic pipeline corrosion. 8 Over the last decade, several mechanisms for microbial EET have been proposed, including indirect electron transfer via redox-active organic electron shuttles 9,10 and direct electron transfer by c-type cytochromes (c-Cyts) located in the outer membrane (OM) 11 or on nanometer-scale bacterial laments. 12,13 In indirect EET, microbial cells are able to perform electron transfer without the necessity of direct contact with solid-phase electron acceptors. Thereby, as the surface area of the solids is limited, EET via self-secreted, naturally occurring or articially supplemented soluble electron shuttles appears to be a major pathway for electron transfer to solids. 10 However, such soluble redox molecules are apparently not involved in the EET process of Geobacter sulfurreducens, 14 which is the most efficient current-producing microorganism characterized to date. This conclusion was reached aer it was shown that the exchange of supernatants to fresh medium in an electrochemical reactor did not impair current production by G. sulfurreducens as it did in other microorganisms. 14 Recently, we reported that cell-secreted redox molecules, such as avins, have a high affinity for OM c-Cyts with reduced hemes in Shewanella oneidensis MR-1. 15 This avin binding to OM c-Cyt scaffolds was found to facilitate a one-electron redox reaction via semiquinone, resulting in a 10 3 -to 10 5 -fold enhancement of the EET rate compared to free avin. 15 Based on this nding, we speculated that Geobacter cells might also use self-secreted redox molecules as redox cofactors to promote EET. If true, the excreted molecules should not act as a diffusion-based electron

Broader context
Geobacter sulfurreducens is an iron-reducing bacterium that has a signicant content of c-type cytochromes (c-Cyts) in an outer-cell membrane (OM) and bacterial laments, drawing keen attention as a model microorganism for the research of microbial fuel cells and bioremediation technology. Concerned with the ability of Geobacter to transport electrons to electrodes in microbial fuel cells, processes termed extracellular electron transport (EET), a great deal of research has been focused on identifying the molecular mechanisms behind the bacterial current generation. Herein, we report the rst evidence that Geobacter utilizes selfsecreted avins as a redox cofactor in OM c-Cyts. Using a highly sensitive voltammetry technique, the key redox signal for the Geobacter EET was identied in vivo. Experiments using a mutant unable to produce OM c-Cyts, together with the spectroscopic and LC-MS analyses, revealed that the noble redox signal is assigned to the bound riboavin associated with OM c-Cyts containing reduced hemes. As the heme redox state reects a balance between the electron input from respiration and the output by EET, the present study not only signicantly contributes to the molecular understanding of EET, but also highlights the microbial capability of utilizing self-secreted riboavin as a regulator for intracellular redox homeostasis.
shuttle in G. sulfurreducens, and thereby the supernatant exchange could not impair current production by G. sulfurreducens. 14 This possibility is supported by the fact that, similar to S. oneidensis MR-1, 16 G. sulfurreducens have genes for both avin biosynthesis and secretion ( Fig. S1 †), although their functions have not been conrmed.
In this paper, we report that G. sulfurreducens secrete avin that contributes to EET at the cell/electrode interface, highlighting the crucial role of avin redox cycling for efficient EET in this species.
To directly examine the secretion of avin by G. sulfurreducens, a supernatant solution of anaerobically grown cultures was subjected to spectroscopic measurements (Fig. 1). Geobacter cells were grown in a dened medium supplemented with acetate (20 mM) as an electron donor and fumarate (80 mM) as an electron acceptor. As shown in Fig. 1a and b, both the peak intensities of emission at 525 nm and excitation at 370 and 440 nm increased when the growth curve of cells was in a sigmoidal phase. The emission and excitation spectral peaks were identical to those observed in the spectra of ribo-avin (RF) and avin mononucleotide (FMN) solutions without cells ( Fig. S2a and b †), suggesting that G. sulfurreducens cells secrete avin species during growth. In support of these data, mass chromatography patterns of the cell culture supernatant measured in selective ion monitoring (SIM) mode at m/z 375.35 identied that RF was present at a concentration of 100 nM, although FMN at m/z 455.34 was under the detection limit ( Fig. 1c and S3 †). These ndings provide evidence that G. sulfurreducens secrete RF at concentrations comparable to that observed in cultures of anaerobically grown S. oneidensis MR-1. 17 The contribution of secreted RF for EET was examined by adding a concentrated RF solution during the electrochemical cultivation of G. sulfurreducens. Current production (I c ) was measured using a cell suspension with an optical density at l ¼ 600 nm (OD 600 ) of 0.2 cultured with 10 mM acetate as the sole electron source on an indium-tin oxide (ITO) electrode at a poised potential of +0.4 V (vs. SHE) (ESI †). At several time points during the course of microbial current production, differential pulse (DP) voltammetry was conducted (Fig. 2a). Three peaks were observed in DP voltammograms aer 46 h (Fig. 2b), but only the redox peak current at À0.2 V (I À0.2 ) increased with time, whereas the peak current of the other redox peaks decreased inversely with time (Fig. 2b). When 1.0 mM RF was added to Geobacter cells on the electrode during current production under the same conditions as above, the I c rapidly increased by $10% (Fig. S4 †). Furthermore, as the addition of RF also caused an increase in I À0.2 (Fig. S5 †), it is possible to assign the redox peak at À0.2 V to the redox cycling of RF. This result also indicates that Geobacter cells utilize self-secreted RF for EET, even before the addition of exogenous RF.
We plotted I À0.2 against I c to quantify the contribution of the redox species at À0.2 V to EET. As shown in Fig. 2, I c exhibited a positive correlation with I À0.2 , as a tted line passed through the point of origin with a high correlation coefficient (r 2 ¼ 0.998). Importantly, this trend was observed both before and aer the addition of RF, and even aer the addition of malonic acid as a metabolic inhibitor (Fig. 2d and S4 †). These results conrm the assignment of the peak potential (E p ) at À0.2 V to the redox cycling of RF, and demonstrate that respiratory current generation by G. sulfurreducens is strongly coupled with the amount of electrochemically active RF. In addition, FMN exhibited a similar effect on the I c and peak current to that seen with RF, but at a different E p (Fig. S6 †). Upon the addition of FMN, both the peak current at À0.175 V (I À0.175 ) and I c increased, and a  tted line for the plot of I À0.175 and I c passed through the point of origin (Fig. S6 †), demonstrating the capability of G. sulfurreducens to utilize FMN as an efficient electron carrier. Together, these data provide evidence for the involvement of RF and/or FMN as important electron carriers at the interface between Geobacter cells and ITO electrodes.
The measured E p in the DP voltammogram of RF and FMN solutions in the absence of cells signicantly differ from those observed in Geobacter biolms (Fig. S7 †), suggesting that both avins alter their redox properties, as is reported for avins bound to OM c-Cyts in S. oneidensis MR-1. 15 Assuming one of the c-Cyts located on bacterial laments or embedded in the OM interacts with avins as bound cofactors in G. sulfurreducens, insight into the location of c-Cyts that bind RF is important to determine the predominant EET pathway in this species. Fig. 3 shows scanning electron microscopy (SEM) images of G. sulfurreducens cells attached to the ITO electrode aer 50 h of electrode cultivation. Neither lament-like assemblages nor multilayer biolms were formed on the electrode surface, corresponding to a previous report where thick biolms of G. sulfurreducens cells with Cyt-bound laments required more than 4 days of electrochemical cultivation. 18 Therefore, it appears that the EET process mediated by the redox cycling of avins observed in Fig. 2 is the consequence of the activation of c-Cyts located primarily on the OM surface, as opposed to those located on the surface of conductive bacterial laments.
To examine the specic interactions between avins and OM c-Cyts, we used a mutant strain (DomcBEST) of G. sulfurreducens lacking OmcB, OmcE, OmcS, and OmcT, which are the major multi-heme c-Cyts localized on the OM of Geobacter. 19 When cultured at +0.4 V in the presence of 10 mM acetate, current production by the DomcBEST strain was highly impaired (Fig. S8 †). In addition, the peak position and I p of RF displayed a negative shi of 75 mV and 20-fold decrease, respectively, compared to those of the wild-type (WT) strain ( Fig. 4a and S9 †). Under FMN-supplemented conditions, the E p of DomcBEST also exhibited a 50 mV negative shi in the DP voltammogram and a large I p decrease compared to that of WT (Fig. S10 †). These signicant effects on E p observed for the DomcBEST strain demonstrate that both FMN and RF associate with OmcB, OmcE, OmcS, or OmcT c-Cyt proteins to enhance the rate of EET in WT G. sulfurreducens cells. Such specic association of RF with OM c-Cyts of G. sulfurreducens cells was also conrmed by the uorescence analysis of the OM fraction, 20 where less avin contents were observed for the OM fraction of DomcBEST cells compared with that of WT (Fig. S11 †). However, in the system without omcBEST genes, the RF peak current at À0.125 V and I c still showed a positive correlation with a comparable slope value to WT (Fig. S9c †). This result suggests that, in addition to omcBEST proteins, other types of OM c-Cyts in the Geobacter WT strain associate with RF to mediate EET (Fig. 4). This situation differs from the EET process reported for S. oneidensis MR-1, which is unable to bind FMN in the absence of a single binding protein (MtrC). 14 This nding is also consistent with the wider variety of c-Cyts encoded in the genome of G. sulfurreducens compared to that of S. oneidensis MR-1. 21 Together, these results conrm that RF and FMN associate with OM c-Cyts of G. sulfurreducens and serve as redox cofactors (Fig. 4), and that Geobacter is more exible with respect to avin uptake and binding mechanisms than S. oneidensis MR-1. In addition, because the peak current for bound-RF decreased when I c was reduced by the addition of a metabolic inhibitor (Fig. 2d), reduced hemes in OM c-Cyts of G. sulfurreducens play an important role in the interaction with avins, as has also been reported in MR-1. 15

Conclusions
We have provided the rst experimental evidence for the involvement of self-secreted avin in EET in G. sulfurreducens.
As the avin uptake mechanism in G. sulfurreducens appears to be similar to that of S. oneidensis MR-1, 14 and that a wide range of microorganisms possesses homologous OM c-Cyt protein complexes to these two species, 2,22 avin may be a universal factor for efficient EET at bacteria/solid interfaces. In addition, if reduced hemes in OM c-Cyts govern the interaction with avin, the intracellular redox homeostasis of Geobacter cells could be maintained by the release and binding of avin, as is observed in MR-1, 15 given that the oxidation state of hemes in OM c-Cyts reects the balance between the electron input from respiration and the output by EET. Our present experiments were focused on monolayers of cells, where conductive laments (nanowires) were not present; thus, it will be of great interest to examine the interaction of avins with conductive laments in thick biolms, where these laments may also play a role in intercellular EET. Recent theoretical analysis of charge ow along conductive laments highlighted the importance of a multi-step hopping transport mechanism with charge localizing sites separated by less than 1 nm and reorganization energies lower than those known in biology. 13 As arrays of OM c-Cyts such as OmcS expressed on the conductive laments of G. sulfurreducens have been implicated as charge carriers, 23 a specic avin association with c-Cyts could also play an extensive role in multi-step charge hopping through thick biolms. Further studies on the uptake of self-secreted avins as the boundcofactor of OM c-Cyts will serve for the mechanistic understanding of the complex EET reactions and also for optimization of microbial fuel cells 6 and the bioremediation of waste waters and contaminated sediments. 7