A Microenvironment for Shewanella oneidensis MR-1 Exists within Graphite Felt Electrodes

Shewanella oneidensis MR-1 employs various methods of Extracellular Electron Transport (EET) to insoluble terminal electron acceptors, including graphite felt anodes of BioElectrochemical Systems (BESs). Fibers in the felt form a three dimensional meshwork within which microbes can form biofilms, as well as occupy the interstitial spaces as planktonic cells. Our results indicate that these interstices generated by the meshwork create a novel microenvironment where planktonic cells grow to higher density under certain conditions. When incubated anaerobically with 18 mM lactate and 30 mM fumarate, planktonic cell counts within the electrodes are ∼10-fold higher than bulk planktonic cell counts; a phenomenon we termed the “interstitial felt effect”. Upon lowering both lactate and fumarate concentrations 10-fold, while bulk planktonic cell counts are stable, the interstitial felt effect disappears. This effect reappears when lactate and fumarate concentrations are lowered another 10-fold. Cyclic voltammetry experiments did not reveal any modification of the graphite fibers within the electrodes. A mutant strain lacking the primary flavin transporter gene, bfe, also expresses the interstitial felt effect. The interstitial planktonic microenvironment of electrically inert polyurethane sponge did not demonstrate this phenomenon. This study provides new insights into interactions of microbes with electrode materials that may help improve overall BES performance. © The Author(s) 2017. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0161703jes] All rights reserved.

Shewanella oneidensis MR-1 is a Gram-negative facultative gamma-proteobacterium with remarkable respiratory versatility.It can respire several soluble terminal electron acceptors including oxygen, fumarate, dimethylsulfoxide (DMSO), and, insoluble terminal electron acceptors such as iron and manganese oxides, which can be attributed to some of the 41 c-cytochromes encoded by the genome. 1,2he ability of bacteria to transfer electrons to insoluble substrates is accomplished through a process called Extracellular Electron Transport (EET). 3,4S. oneidensis MR-1 has been proposed to use direct and indirect mechanisms for EET: direct contact via outer membrane decaheme c cytochromes, [5][6][7] and via nanowires -periplasmic and outer membrane extensions to which decaheme c cytochrome complexes are localized; [8][9][10] and indirectly via secreted flavin molecules. 11,12This EET capability of S. oneidensis MR-1 places it in the group of electrochemically active bacteria (EAB).Other examples from this group include delta-proteobacteria such as Geobacter metallireducens, phototrophic bacteria like Rhodopseudomonas palustris, and even a yeast, Pichia anomala. 13EAB can serve as redox catalysts in bioelectrochemical systems (BESs), which typically consist of an anodic and a cathodic chamber separated by a proton permeable membrane.Various types of BESs can be constructed, depending on the process and desired outcome. 14In microbial fuel cell (MFC) type BESs, EAB in the anodic chamber catalyze an oxidation reaction, for example oxidation of a specific carbohydrate, generating electrons, which are deposited at the anode; protons diffuse through the proton permeable membrane to the cathodic compartment.The electrons travelling through an external circuit can be harvested to reduce oxygen, and combine with protons to form water. 14 While decades of research in the field of microbial electrosynthesis have led to significant improvements in laboratory MFC design and overall performance, practical large-scale applications of this technology have not yet been realized.Multiple factors affect the final output of an MFC; Babanova et al., 15 conducted statistical analyses that identified and evaluated different factors influencing MFC performance and their level of contribution to the final output, which included z E-mail: sfinkel@usc.eduvarious aspects of operational design, electrode compartments, and the electrochemical methods used.Some important parameters that correlated with maximal current output include anode material, high geometrical surface area of the electrodes and membrane, and substrate contents.Indeed, understanding the metabolic and physiological interactions of microbes with electrode surfaces has always been a key challenge in perfecting them as effective electrocatalysts.
MFCs require the use of electrode materials that have good electrical conductivity, are biologically compatible, chemically stable, and relatively inexpensive.Non-corrosive metals and carbon based materials have been commonly used in MFCs.These include stainless steel, platinum, and titanium electrodes; carbon paper and cloth; and, graphite fiber brushes, sheets, granules and felt. 13,16Large surface area is one of the most important factors determining the performance of an electrode, and graphite felt can provide a high three dimensional surface area owing to its multiple layers of fibers.The relatively loose meshwork provides ample space for the microbes to migrate planktonically into and out of the electrode.The interstitial planktonic microenvironment thus formed may affect the stability of biofilms formed within an MFC graphite felt electrode.Here, we characterize growth of microbes in the planktonic medium of the graphite felt interstitial environment, and compare it to the growth attributes of cells in the bulk planktonic environment of the surrounding medium (Fig. 1).Analysis of different growth conditions with respect to electron donor and acceptor availability indicate that S. oneidensis MR-1 cells prefer the interstitial felt environment under certain conditions.An understanding of these preferences may help us to better design electrodes where microbes can serve as more efficient electrocatalysts.

Materials and Methods
Bacterial strains and growth conditions.-Weutilized a spontaneous rifampicin-resistant mutant of Shewanella oneidensis MR-1 (SFS1002, referred to as WT) as our wild type, and a bfe mutant (SFS1004, bfe) which lacks the primary inner membrane transporter required for efficient flavin mononucleotide (FMN) transport from the cytoplasm to the periplasm. 17Cultures were initiated by transferring cells directly from a frozen 20% glycerol stock into 6 ml of Luria-Bertani broth (Difco) in a test tube and incubated overnight with aeration in a TC-7 rolling drum (New Brunswick Scientific, Edison, NJ) at 30 • C. To begin experiments, the overnight grown cells were introduced into 25 ml M1 medium (50 mM PIPES, 85 mM NaOH, 28.04 mM NH 4 Cl, 1.34 mM KCl, 4.35 mM NaH 2 PO 4 • H 2 O), supplemented with trace minerals, vitamins, and amino acids as described previously, 6 except that the vitamin supplement was prepared without riboflavin.The cells were diluted 1:1,000 (vol:vol) in borosilicate anaerobic crimp test tubes (18 mm × 150 mm), resulting in very little head space.The medium was also supplemented with lactate and fumarate as the electron donor and acceptor, respectively, the concentrations of which were changed according to experimental requirements.Graphite felt (GF-S6-06, Electrolytica Inc, NJ, USA) was subsequently introduced into the tubes, which were then sealed using rubber stoppers and aluminum crimps (Fig. 1).The graphite felt sections were thus maintained at Open Circuit Potential (OCP).The graphite felt was acid treated before sterilization; the felt was cut into 1 × 1.5 × 0.6 cm pieces, soaked in 90% ethanol for 30 minutes, washed 8-10 times in deionized water and left stirring overnight in 1 M HCl.The felt sections were then washed 8-10 times in deionized water, autoclaved and stored in M1 medium buffer (pH 7) solution prior to use.Electrically inert polyurethane sponge (pore size 0.83 μm, Jaece Identi-plugs, Fisher Scientific, USA) sections of the size 1 cm x 1.5 cm x 1 cm were prepared in a manner similar to the graphite felt sections, except they were treated with 1 M HCl for 3 hours.All cultures were incubated statically for up to three days at 30 • C. Porosity of the graphite felt (35.6% ± 2.5) and sponge (42% ± 3.6) materials are similar, as calculated based on the total and pore volumes of statically incubated sections of either material (data not shown) and both materials have pore sizes on the order of ∼400 to ∼800 μm 3 .

Monitoring of cell growth and survival.
-Viable cell counts were determined to monitor cell growth and survival by serial dilution of cells sampled from the cultures, followed by plating on LB agar plates. 18The limit of detection in all experiments was >1,000 CFU/ml.All experiments were performed using biological triplicates (n ≥ 3 for all experimental conditions).Except for the planktonic-only controls, three replicates were sacrificed at each time-point to determine planktonic and felt interstitial medium cell counts.Aliquots for the planktonic controls were withdrawn using a sterile 18 G needle and syringe.For the felt interstitial medium cell counts, the felt was wicked to remove excess medium, then pressure was applied to the felt between gloved fingers to transfer a sample of the interstitial medium into a sterile Eppendorf tube.The average volume of medium removed from each felt sample was 400 μl (±100 μl), and from each sponge section was about 800 μl (±100 μl).Biofilm cells from the felts were obtained by mild sonication using a sonicator probe (Branson Digital Sonifier Model 102 C), to ensure cell detachment and minimal lysis.Prior to sonication, felts were rinsed with and resuspended in M1 buffer, the solution was then subjected to four sonication bursts of 3 seconds each at 10% amplitude.
Cyclic voltammetry.-Theculture media with felt sections were transferred from the anaerobic tubes to electrochemical cells of the same volume to allow optimal electron transfer between the analyte and electrodes.The felt, which served as the working electrode, was attached to a titanium wire to facilitate transfer of electrons.A platinum wire counter electrode and a reference electrode (Ag/AgCl wire within a 1M KCl solution) completed the circuit.The samples were continually flushed with N 2 gas to maintain an anaerobic environment.All voltammetry experiments were performed using a Gamry Ref 600 potentiostat, which can detect currents in the range of 60 pA -600mA (Gamry Inc).A scan rate of 10mV/s was used to generate cyclic voltammograms between −0.6V to 0.3V (vs Ag/AgCl).

Comparison of planktonic cell density in the bulk medium versus within the graphite felt interstitial medium.
-Viable planktonic cell counts for the bulk medium and interstitial medium of graphite felt held at Open Circuit Potential (OCP) were compared over a range of lactate and fumarate concentrations using our WT strain (Fig. 2A-2C).In M1 medium containing 18 mM lactate and 30 mM fumarate, our 'standard' condition, interstitial planktonic cell counts within the graphite felt matrix were consistently ∼10-fold higher than those of the surrounding bulk medium (∼5.0 × 10 8 CFU/ml vs. ∼5.0× 10 7 CFU/ml; Fig. 2A).When the lactate and fumarate concentrations were decreased 10-fold to 1.8mM and 3mM respectively, the interstitial felt effect disappeared, with viable planktonic counts in the felt being the same density as the surrounding bulk environment; ∼5.0 × 10 7 CFU/ml (Fig. 2B).Upon decreasing the lactate and fumarate concentrations another order-of-magnitude to 0.18 mM and 0.3 mM, respectively, while the overall cells counts dropped 10-fold as compared to standard M1 medium, the interstitial felt effect reappeared (Fig. 2C); planktonic cell densities within the electrode felts were again ten times greater than in the bulk medium.
Biofilm cell counts of OCP graphite felt electrodes.-Whileremoving the interstitial medium removes virtually all planktonic cells within the felt environment, the biofilm cells remain attached to the felt electrodes.In order to quantify the biofilm cells, low intensity sonication was performed, which removes most, if not all, attached cells without lysis (Ribbens and Finkel, unpublished observations and Ref. 18).Under all three electron donor and acceptor conditions, the biofilm cell counts were 100-fold lower than their corresponding interstitial planktonic cell counts (Table I).

Comparison of cell density in bulk medium versus interstitial sponge medium.
-Non-conductive polyurethane sponge was tested for the same phenomenon using identical growth and incubation conditions, as it is also a crosslinked structure that would allow the formation of an interstitial planktonic microenvironment within the bulk environment.Under no conditions was the interstitial effect observed between planktonic cell counts of bulk and interstitial medium microbes.At 18 mM lactate + 30 mM fumarate concentrations, the bulk medium cell counts of cultures grown with non-conductive sponges and electrode felts were similar (∼5.0 × 10 7 CFU/ml vs. ∼5.0× 10 7 CFU/ml; Fig. 2A and 2D).However, while interstitial planktonic counts in the graphite felt were increased 10-fold, planktonic cells counts in the interstices of the non-conductive sponge were about the same as the bulk medium cell counts (∼5.0 × 10 7 CFU/ml; Fig. 2D).At the other two concentrations of lactate and fumarate, the interstitial planktonic sponge cell counts were also similar to their corresponding bulk medium cell counts (Figs.2E and 2F).
Felt effect phenomenon in a bfe mutant.-S.oneidensis MR-1 has been shown to use flavin mediated mechanisms for EET. 11,12The flavin exporter protein, Bfe, transports FAD across the inner membrane to the periplasm, where it can be incorporated into proteins as a co-factor, or can be hydrolyzed to FMN and AMP by UshA. 17MN then diffuses across the outer membrane, and can spontaneously lose a phosphate group to form riboflavin (RF).Both FMN and RF have been demonstrated to serve as cofactors for outer membrane c-cytochromes. 19,20To determine the involvement of flavins in the interstitial felt effect, viable planktonic cell counts both in bulk medium and the graphite felt interstitial medium were determined for the flavin exporter mutant bfe, using conditions identical to the WT strain.Under all three lactate and fumarate conditions, both bulk and interstitial felt planktonic cell counts were similar to the WT (data not shown).Essentially, the interstitial felt effect was observed for the bfe mutant as well.

Cyclic voltammetry (CV) analysis.-CV analysis was carried out
to identify any differences in redox activity within the graphite felts under different concentrations of electron donor and acceptor, for both the WT and bfe mutant strains.In this technique, an electrical potential range is scanned cyclically in forward and reverse directions to investigate the presence of redox couples.A voltammogram is obtained by plotting the resulting current versus the potential at the working electrode (in this case the graphite felt incubated with the cells), against the constant potential of a reference electrode.The voltammogram of the WT cells grown in standard M1 medium and harvested on Day 1 exhibits the double sigmoidal curve typical of S. oneidensis MR-1 EET under turnover conditions, i.e. in the presence Table I.Planktonic versus biofilm cell counts of the WT strain grown with graphite felt electrodes at OCP. Biofilm cell counts of the OCP graphite felt electrodes are 100-fold less than their corresponding interstitial planktonic cell counts, for all three lactate and fumarate concentrations tested.The interstitial planktonic cell counts are in turn 10-fold more than the bulk medium counts, at the highest and lowest lactate and fumarate concentrations, i.e. the interstitial felt effect was observed.This is not the case for the intermediate lactate and fumarate concentration.Standard deviation represents n = 6.  of an electron donor (Fig. 3A,). 20The first wave at ∼ −0.45V and the second wave onset at ∼0.0 V (vs Ag/AgCl) indicate EET-facilitated reduction of the electrode by a flavin dependent mechanism, and direct contact of outer membrane multiheme c-cytochromes, respectively.Similar double sigmoidal voltammograms with flavin and multiheme c-cytochrome waves were obtained for Day 1 bfe cells grown in the same medium and incubation conditions, although with a higher current output (Fig. 3B).The reduced presence of flavin peaks in these samples indicates that the strain is not entirely lacking in flavin secretion.After the interstitial medium was removed and replaced with fresh medium, CV was again performed.Now, voltammograms are similar to the abiotic control with minimal to no flavin and ccytochrome peaks (Figs.3A and 3B).This indicates that the biofilm cells that are still present on the electrodes do not contribute to electrode reduction under turnover conditions.Additionally, it appears that neither the interstitial planktonic cells, nor the flavins and ccytochromes, adhere strongly to the felt fibers, since the voltammograms of the electrodes after rinsing show no redox couples.The CV analysis thus indicates that all detectable EET to the graphite felt electrodes is actually mediated by the presence of microbes in the interstitial medium, and not the biofilm cells.CV analysis of the graphite felt sections of both the wild type and bfe mutant strains grown in M1 medium with 1.8 mM lactate and 3 mM fumarate, resulted in voltammograms similar to those of the abiotic control (Figs.3C and 3D).Although the WT exhibits a small flavin peak, the signature flavin and c-cytochrome peaks were not observed for bfe.Addition of excess electron donor i.e. lactate, also did not enable any EET-facilitated reduction of the electrode (not shown).The voltammograms of both strains grown with 0.18 mM lactate and 0.3 mM fumarate were also similar to the abiotic control (data not shown).

Discussion
Shewanella oneidensis MR-1 is one of the primary model organisms for studying mechanisms underlying EET, a process crucial for biogeochemical cycling of minerals and metals in the environment, 1 biotechnological applications such as utilizing microbially catalyzed fuel cells for bioremediation, and renewable energy generation. 14lthough the performance of BESs has improved over the years, large-scale practical applications nonetheless face several challenges.These include understanding the metabolic diversity required for breaking down complex fuel substrates, mediating biochemical interactions with electrode interfaces, and, technological challenges that encompass various factors such as long-term stability in the presence of electrode potential losses. 21Here, we discuss an open circuit biological system, wherein we examine the interaction of S. oneidensis MR-1 cells with graphite felt electrodes and variations in electron donor and acceptor concentrations.Experiments such as these are helping us to tease apart factors contributing to metabolic interactions of bacterial cells with graphite felt electrode material that may then be applied to operational BES systems like MFCs.
Not only have a number of different anode materials and modification techniques been tested in various kinds of MFCs, anode material and surface area have been demonstrated to influence current output by S. oneidensis MR-1. 22,23Our study adds another factor for consideration while selecting an anode material for a MFC design.The multiple stacked layers of fiber meshwork in graphite felt versus other graphite electrode materials like cloth, fiber brushes, or rods, enables the felt to contain small amounts of medium within it; ∼400 μl within 1 cm × 1 cm × 0.6 cm felt sections in our system.The resultant porosity of the graphite felt likely enables cells to migrate in and out of the felt creating a third novel environment, namely the interstitial planktonic microenvironment, apart from the widely recognized bulk planktonic and electrode biofilm environments (Fig. 1).Furthermore, our results demonstrate that the cells prefer the interstitial microenvironment to the bulk planktonic environment under certain conditions.When the cells are grown under our standard conditions with 18 mM lactate and 30 mM fumarate as electron donor and acceptor, respectively, the interstitial felt microenvironment has ∼10-fold higher density of planktonic cells than the bulk planktonic medium, in a volume that is ∼65 times smaller.This is an unusual phenomenon, as an equilibration of cell concentrations between both types of planktonic environments would be expected if there was no preference.Moreover, biofilm cell concentrations, which were measured after removal of the interstitial medium always decrease as the interstitial planktonic cell density decreases, whether or not the "felt effect" is observed.
When graphite felt electrodes were incubated under standard conditions and analyzed using cyclic voltammetry, a double sigmoidal curve was observed indicative of biotic EET-dependent electrode reduction, under turnover conditions, using both c-cytochrome and flavin related mechanisms.When the interstitial medium was removed from the felt, including most of the interstitial planktonic cells, but leaving the biofilm cells intact, the subsequent cyclic voltammogram did not exhibit any electrode reduction, with no peaks for either flavins or c-cytochromes.This indicates that the planktonic microbes residing in the interstices of the felt are performing all detectable electrochemical activity and that they do not physically modify the felt in any manner.The biofilm cells still present on the felt fibers are either not a major contributor to electrode reduction in this system, or must work intimately with the interstitial cells to reduce the electrode with an EET-dependent mechanism.Further, planktonic microbes no longer prefer this interstitial microenvironment when the lactate and fumarate concentrations are reduced 10-fold; the cell concentrations of the bulk and interstitial planktonic environments are similar, and biofilm cell concentrations are 100-fold lower than the interstitial planktonic cell concentrations.The cyclic voltammograms of these felts resembled those of the abiotic controls.Decreasing lactate and fumarate concentrations another 10-fold to 0.18 mM and 0.3 mM, respectively, restores the interstitial felt effect, though overall cell counts drop 10-fold as compared to the standard medium conditions, again indicating preference for this microenvironment.The cyclic voltammograms of the felt sections under these conditions also resemble those of the abiotic medium controls.It is plausible that in our experimental set-up, planktonic cells prefer the felt interstitial microenvironment either under conditions of nutrient "excess" or "dearth", perhaps to move away from metabolic by-product buildup in the former case, or to aid each other in nutrient uptake in the latter case.Whether the planktonic cells are continually migrating in and out of the felt, or the two populations differ from each other physiologically and grow differently in their respective environments, remains to be determined.
Examination of a bfe mutant deficient in flavin secretion indicated that flavins might not be a major contributor to the interstitial felt effect.However, their role in this phenomenon cannot be entirely eliminated as the mutant still secretes small amounts of flavins, as observed using cyclic voltammetry.Significantly, this phenomenon was not observed when cells were grown with electrically inert polyurethane sponge sections instead of graphite felt, which also allow the formation of a planktonic interstitial microenvironment within the bulk planktonic medium.It should be noted that since porosity of these materials are similar and the pore sizes are significantly larger than the bacterial cells (which are on the order of 1 μm 3 ), it is most likely that the felt effect is more dependent on the conductivity of electrode materials, making it an additional factor to take into consideration while designing BESs.
Biffinger et al. 24,25 have demonstrated in Shewanella catalyzed operational MFCs containing carbon coated titanium anodes that current output can be correlated to electron donor and biofilm formation, moreover, that planktonic cells may be responsible for current output to a greater degree than biofilm cells.The latter observation was also made by Lanthier et al. 26 using S. oneidensis driven MFCs containing graphite rod anodes, wherein more planktonic biomass was recovered than from the biofilms at the conclusion of the experiments.Dolch et al. 27 similarly reported that cells do not need to be localized to the anodes to perform EET; they observed higher current densities from S. oneidensis catalyzed MFCs using graphite felt anodes as compared to graphite foil anodes, in spite of lower biofilm coverage on graphite felt.Given the three dimensional architecture of graphite felt, and in light of the observations made by us and others, it is likely that current output from MFCs containing graphite felt anodes could be dependent on interstitial planktonic biomass to a higher degree than surface attached biofilms, which in turn, according to data presented here, could be significantly more dependent on nutrient availability than previously appreciated.
Data from the experimental system described here indicates that microbes in open circuit BESs with graphite felt electrodes survive in at least three environments: as cells in the bulk planktonic medium, within biofilms on the electrode surface, and, within the interstitial felt environment as planktonic cells; a previously unclassified environment that we identified in this study.The felt interstitial environment appears to be favored by S. oneidensis MR-1 cells under certain electron donor and acceptor concentrations.Therefore, modulating nutrient concentrations could directly affect current outputs and overall performance of a BES by tipping the balance between cells inside and outside the electrode environment.This work highlights the need for a deeper understanding of the interactions and physiological responses of microbes to the combination of electrode material and carbon and energy sources provided.

Figure 1 .
Figure 1.Experimental set-up and a single layer of the felt interstitial microenvironment: The cells were cultivated in an anaerobic tube containing M1 medium with 18 mM lactate and 30 mM fumarate and a section of graphite felt.The meshwork of graphite fibers allows the cells to live planktonically within the electrode (dark gray cells), as well as form a biofilm (light gray cells).

Figure 2 .
Figure 2. Bulk Planktonic (light gray bars), interstitial felt planktonic (dark gray bars) and interstitial sponge planktonic (white bars) cell counts for the WT strain (A) WT grown in standard M1 medium with graphite felt.(B) WT grown with 1.8 mM lactate and 3 mM fumarate with graphite felt.(C) WT grown in M1 medium with 0.18 mM lactate and 0.3 mM fumarate with graphite felt.(D) WT grown in standard M1 medium with sponge.(E) WT grown in M1 medium with 1.8 mM lactate and 3 mM fumarate with sponge.(F) WT grown in M1 medium with 0.18 mM lactate and 0.3 mM fumarate with sponge.All cultures started at a density of ∼5.0 × 10 5 CFU/ml.Error bars represent standard deviation for n ≥ 3.

Figure 3 .
Figure 3. Cyclic voltammetry analysis on OCP graphite felt electrodes with cells grown in M1 medium (light gray) for a day, graphite felt electrodes with one day old interstitial medium replaced with fresh M1 medium (dotted line), and abiotic medium control (black).Three biological replicates of the light gray curves are shown in the insets, to show reproducibility.(A) WT grown in standard M1 medium (18 mM lactate + 30 mM fumarate).The first peak at ∼ −0.45 V and second peak onset at ∼0.0 V (vs.Ag/AgCl) indicate electrode reduction by flavin mediated and outer membrane c-cytochrome mechanisms respectively.On replacing the interstitial medium in these electrodes with fresh M1 medium, electrode reduction via these mechanisms was not observed.(B) bfe cells grown in standard M1 medium.A double sigmoidal curve similar to the WT strain was observed, with greatly reduced flavin secretion.(C) WT grown with 1.8 mM lactate and 3 mM fumarate.(D) bfe grown with 1.8 mM lactate and 3 mM fumarate.In both (c) and (d), one day old graphite felt electrodes resemble the abiotic control.Voltammograms represent one example of n = 6.