Characterisation of yeast microbial fuel cell with the yeast Arxula adeninivorans as the biocatalyst

Biosensors and Bioelectronics 26 (2011) 3742–3747

Contents lists available at ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

Characterisation of yeast microbial fuel cell with the yeast Arxula adeninivorans as the biocatalyst Nicholas D. Haslett a , Frankie J. Rawson b , Frèdèric Barriëre c , Gotthard Kunze d , Neil Pasco e , Ravi Gooneratne a , Keith H.R. Baronian f,∗ a

Faculty of Agriculture and Life Sciences, P.O. Box 84 Lincoln University, Lincoln 7647, Canterbury, New Zealand Department of Chemistry, University of Canterbury, Private Bag 4800, Christchurch, New Zealand Université de Rennes 1, CNRS UMR 6226, Sciences Chimiques de Rennes, Equipe MaCSE, France d Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstr. 3, D-06466 Gatersleben, Germany e Lincoln Ventures Ltd., P.O. Box 133, Lincoln, Christchurch 7640, New Zealand f School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand b c

a r t i c l e

i n f o

Article history: Received 1 December 2010 Received in revised form 21 January 2011 Accepted 9 February 2011 Available online 16 February 2011 Keywords: Microbial fuel cell Yeast Mediated transfer Mediator-less transfer Secreted redox molecule

a b s t r a c t Yeast microbial fuel cells have received little attention to date. Yeast should be ideal MFC catalyst because they are robust, easily handled, mostly non-pathogenic organisms with high catabolic rates and in some cases a broad substrate spectrum. Here we show that the non-conventional yeast Arxula adeninvorans transfers electrons to an electrode through the secretion of a reduced molecule that is not detectable when washed cells are first resuspended but which accumulates rapidly in the extracellular environment. It is a single molecule that accumulates to a significant concentration. The occurrence of mediatorless electron transfer was first established in a conventional microbial fuel cell and that phenomenon was further investigated by a number of techniques. Cyclic voltammetry (CV) on a yeast pellet shows a single peak at 450 mV, a scan rate study showed that the peak was due to a solution species. CVs of the supernatant confirmed a solution species. It appears that, given its other attributes, A. adeninivorans is a good candidate for further investigation as a MFC catalyst. © 2011 Published by Elsevier B.V.

1. Introduction Most microbial fuel cell development has focused bacterial cells as the biocatalyst and many reports are available on their construction, function and performance (Bullen et al., 2006; Davis and Higson, 2007; Logan et al., 2006; Logan, 2009). Reports on microbial fuel cells that use eukaryote cells are however rare in the literature. Bennetto (1990) described a Saccharomyces cerevisiae fuel cell for use in school science. The fuel cell used methylene blue as the anode mediator and potassium ferricyanide as the cathode electron acceptor but was described as ‘sluggish’. Significantly, all other publications from Bennetto involve prokaryote cells as the biocatalyst. S. cerevisiae was also described as a ‘sluggish biocatalyst’ and the fuel cell as ‘lethargic’ by Wilkinson (2000). These fuel cells are also limited in that the maximum potential of the cell is restricted by the difference between the E0 values (vs Ag/AgCl at pH 7) of methylene blue (−0.23 V) and ferricyanide (0.28 V). Although most early yeast MFC research was performed with S. cereviseae, it is not an ideal organism. An important prereq-

∗ Corresponding author. Tel.: +64 3 364 2987x7029; fax: +64 3 364 2590. E-mail address: [email protected] (K.H.R. Baronian). 0956-5663/$ – see front matter © 2011 Published by Elsevier B.V. doi:10.1016/j.bios.2011.02.011

uisite for any MFC is the ability of the organism or consortia of organisms to completely oxidise the substrate to make available the maximum number of electrons. Yeast can have a fully respiratory or fully fermentative catabolism and in some cases a mixed respiratory-fermentative catabolism. S. cerevisiae operates a mixed fermentation/respiration mode even in the presence of oxygen, the natural terminal electron acceptor for the electron transport chain. In S. cereviseae the ratio of fermentation to respiration varies between strains but is approximately 80:20, respectively. Further in some yeast the production of ATP in TCA and electron transport/chemiosmosis is not maximal, e.g. S. cereviseae produces a total of 14 ATP per glucose molecule in mitochondrial process, which is well short of the typical net 28–30 achieved by most aerobes. Most of the substrate energy thus remains in the end product of fermentative pathway. The situation from the MFC perspective is worse when a soluble terminal electron acceptor is not available, for example in anaerobic conditions and fermentation becomes the sole catabolic pathway that produces ATP. Theoretically only about 5% of the energy available in glucose is converted to ATP and the majority of the energy and thus the number of electrons available through oxidation of the substrate, remains in the products of fermentation. Furthermore the absence of functional mitochondria in anaerobic conditions removes the pathways required for non-

N.D. Haslett et al. / Biosensors and Bioelectronics 26 (2011) 3742–3747

conventional substrates such as pentoses, sugar alcohols, organic acids, aliphatic alcohols, hydrocarbons and aromatic compounds (Walker, 1998). Lipids are ultimately degraded by glycolysis and via the glyoxalate cycle, and amino acids are degraded to ammonium and glutamate, which is either catabolised via fermentation or the TCA cycle. In both cases the inhibition of mitochondria will limit the catabolism of lipids and amino acids. Although the total production of electrons per unit time may remain approximately the same in both anaerobic and aerobic conditions, it is achieved by higher throughput of the substrate in fermentation. Yeast fuel cells have received renewed attention recently, but in general, yeast MFCs still have a lower power output than bacterial fuel cells. Hubenova and Mitov (2010) report that a Candida melibiosica mediated MFC with methylene blue in the anode and ferricyanide in the cathode gave power densities of up to 0.18 W m−3 . In another report they describe how the modification of a carbon felt electrode with surface nickel nanostructures significantly increased the power output to 720 mW m−2 , which is much higher than the 36 mW m−2 that was achieved with a bare carbon felt electrode and the 370 mW m−2 achieved with the bare carbon felt electrode and methylene blue as a mediator (Hubenova et al., 2010). Gunawardena et al. (2008) explored various combinations of water (O2 ), methylene blue and potassium ferricyanide as mediators and electron acceptors in a S. cerevisiae microbial fuel cell. It generated a maximum power of 146.71 ± 7.7 mW m−3 with a maximum open circuit potential of 383.6 ± 1.5 mV. The maximum operational efficiency of the fuel cell was 28 ± 1.8% which occurred when methylene blue was the mediator in both anode and cathode compartments. Chiao et al. (2006) report 2.3 nW cm−2 for micro machined MFC using S. cerevisiae as the biocatalyst. Walker and Walker (2006) using S. cerevisiae report <5 ␮W cm−2 . Prasad et al. (2007) report that a mediator-less fuel cell achieved a maximum power density of 2.9 W m−3 using a polyaniline (PANI)–Pt composite coated graphite electrodes and Hansenula anomala (Pichia anomala) as the biocatalyst. Graphite-felt and plain graphite electrodes had maximum power densities of 2.34 W m−3 and 0.69 W m−3 , respectively. Halme and Zhang (1995) constructed an indirect S. cerevisiae fuel cell. The yeast was grown on glucose in a fermenter and the supernatant withdrawn to the anode of the fuel cell. This work showed that a soluble redox molecule(s) is secreted into the growth medium by the yeast. Power output from mature cells (80 h) was 120 ␮W cm−3 with less power available from late log phase cells (12 h). Higher power densities are possible, however, and Ganguli and Dunn (2009) report that a methylene blue mediated S. cerevisiae MFC gave a power density of 1.5 W m−2 , which is similar to well performing bacterial MFCs of ∼1.5–2.5 W m−2 but significantly less than the maximum of 6.86 W m−2 reported by Fan et al. (2008). The selection of a yeast as biocatalyst for use in a MFC should be made carefully because the modes of catabolism vary greatly between species and under different conditions in the same species. In this study Arxula adeninivorans (syn. Blastobotrys adeninivorans – Kurtzman and Robnett, 2007) was selected because of its physiological properties of growth at high temperature (up to 48 ◦ C), pH tolerance (growth from pH 2 to pH 10), salinity tolerance (up to 20% NaCl) and its ability to use very broad substrate range (Böer et al., 2009). These characteristics are all desirable when using waste of variable composition and properties as the substrate. Potassium permanganate was selected as the cathode electron acceptor for convenience and because it has reduction reactions with very high E0 values:

3743

Under acidic conditions the reduction half-reactions are: + − MnO− 4 + 4H + 3e → MnO2 + 2H2 O

MnO− 4

+

+ 8H + 5e− → Mn2+ + 4H2 O

E 0 = 1.68 V

(1)

E 0 = 1.51 V

(2)

A third half-reaction occurs in alkaline conditions: − − MnO− 4 + 2H2 O + 3e → MnO2 + 4OH

E 0 = 0.60 V

(3)

The use of permanganate as a cathode electron acceptor has been more fully discussed by You et al. (2006). 2. Materials and methods 2.1. Cells A. adeninivorans LS3 was obtained from the yeast collection of the Leibniz Institute of Plant Genetics and Crop Plant Research (IPK) Gatersleben, Germany. LS3 is a dimorphic strain and was cultured at 37 ◦ C so that it was in the yeast form. S cerevisiae NCTC 10716 was obtained from the ESR yeast collection, Porirua, New Zealand. 2.2. Cultivation and preparation of cells A. adeninivorans LS3 was cultivated aerobically in indented flasks at 37 ◦ C for 24 h in Yeast Extract Peptone Dextrose (YEPD) broth (peptone 20 g L−1 , yeast extract 10 g L−1 and glucose 20 g L−1 , pH 6.5). Cells were harvested by centrifugation (4500 rcf, 8 min, 10 ◦ C), washed twice in phosphate buffer pH 7 and then re-suspended in phosphate buffered saline (PBS, 0.05 M K2 HPO4 /KH2 PO4 , 0.1 M KCl) pH 7 at OD600 = 2.5. S. cerevisiae was prepared for use identically except that the incubation temperature was 30 ◦ C. Cells for cyclic voltammetry (CV) were cultivated and harvested as above except that the cell pellet was resuspended in a small volume of PBS to form a thick cream. The yeast cream supernatant was prepared by centrifuging the cream in a microcentrifuge at 13,000 rpm to separate the extracellular fluid from the cells. 2.3. Reagents Potassium hexacyanoferrate (III) (K3 FeCN6 ; Merck pro analysis) was dissolved in distilled water at 0.5 M, 2,3,5,6-tetramethyl-1,4phenylenediamine (TMPD, Aldrich) was dissolved in 96% ethanol at 20 mM, and potassium permanganate (KMnO4 ) was dissolved in distilled water at 0.5 M. 2.4. Microbial fuel cell The fuel cell is based on the design of Bennetto et al. (Delaney et al., 1984) with additional ports so that it could be operated in continuous mode. The body was fabricated from polycarbonate and the electrodes made from carbon fibre cloth (275 g m−2 , fibre diameter 6 ␮m). The electrode geometric surface area was 0.001018 m−2 (geometric electrode volume 0.000001018 m−3 ). They were cleaned before first use by ultrasonic agitation in isopropyl alcohol, acetone and ether, each successively for 5 min followed by drying and rinsing in distilled water. A proton exchange membrane (Nafion membrane 115, DuPont, San Diego, USA) separated the anode and cathode compartments (both 15 mL). Anode and cathode volumes were 10 mL plus additions where stated. All fuel cell experiments were with suspended cells and were conducted in a shaking water bath at 40 ◦ C, 80 rpm. The external circuit included a 100 ohm resistor and potential was measured with a digital multimeter at 5 min intervals.

3744

N.D. Haslett et al. / Biosensors and Bioelectronics 26 (2011) 3742–3747

2.5. Voltammetry

ferricyanide

1000

ferricyanide, glucose

2.6. Retention of TMPD Cells were incubated for 1 h at 37 ◦ C with TMPD and ferricyanide, or with ferricyanide alone, both in the presence and absence of glucose. Analysis of the supernatant was by linear sweep voltammetry (LSV). At the end of the experiment, cells were washed and stored in PBS at 4 ◦ C until next use. For the next experiment, fresh ferricyanide and glucose were added to replicate the day 1 experimental conditions and samples were incubated for 1 h at 37 ◦ C and again analysed by LSV. This procedure was repeated on days 6 and 9. 3. Results 3.1. Retention of TMPD by A. adeninivorans cells Fig. 1 shows that there is a rapid decrease in the double mediator response while the single mediator responses remain almost static reflecting the continuous production of NADH/NADPH by catabolism. We thus attribute the decline in the double mediator response to a loss of the lipophilic mediator from the cells and not to any decline in catabolism.

ferricyanide, TMPD

i (nA)

100

3.3. Mediator-less fuel cell power output There is a clear difference in the power output from S. cerevisiae and A. adeninivorans fuel cells operating in mediator-less mode

10

0.1 1

2

3

4

5

6

7

8

9

10

time (days) Fig. 1. Retention of TMPD by A. adeninivorans cells. Cells were incubated for 1 h at 37 ◦ C with TMPD and ferricyanide, or with ferricyanide alone, both with and without glucose and the supernatant was analysed by LSV. At the end of the experiment cells were washed and stored in PBS at 4 ◦ C until next use. For the next experiment fresh ferricyanide and glucose were added to the cells to replicate their supply in the day 1 experiment, incubated for 1 h at 37 ◦ C and the supernatant analysed by LSV. This procedure was repeated on days 6 and 9. The difference in the ‘with’ and ‘without’ glucose trials confirms the responses are from catabolism. (Error bars represent ±1SD of the mean, n = 9.)

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 -0.01 0 -0.02

KMnO4

10

20

30

40

50

time (min)

Fig. 2. Detection of mediatorless transfer of electrons from A. adeninivorans to the electrode. The fuel cell was run for 15 min without a cathode electron acceptor. Addition of KMnO4 to the cathode at 15 min results in a surge of current followed by decay to a level above that seen in the first 15 min. Control shows a small peak on addition of KMnO4 followed by a decay to the 0–15 min levels. Electrode geometric surface area 0.001018 m−2 .

(Fig. 3). Although the maximum growth temperature for S. cerevisiae is usually given as 37 ◦ C, catabolism operates as usual up to 45 ◦ C (Walker and Walker, 2006) and its low power output is not due to the elevated temperature in the MFC.

3.2. Mediator-less electron transfer from A. adeninivorans to the fuel cell anode

0.035

power density (Wm-2)

The fuel cell was set up with organisms in PBS in the anode and PBS only in the cathode and allowed to run for 20 min. At 20 min the PBS was withdrawn from the cathode and replaced with KMnO4 0.5 M in PBS. A control without cells was also run. Fig. 2 shows that on the addition of KMnO4 to the cathode there is a surge in power through the external circuit followed by a decay to a constant level that is above the 0–20 min level. The control shows a smaller surge in power after the addition of KMnO4 at 20 min but it is followed by a decay to pre 20 min levels.

ferricyanide, TMPD, glucose

1

power density (wm-2)

Voltametry was performed using an eDAQ potentiostat and eDAQ power lab 2/20 with eCHEM data acquisition software. Linea Sweep Voltammetry (LSV) used a three-electrode system comprising of a Pt micro-disc electrode (100 ␮M diameter), a Pt auxillary electrode and a Ag/AgCl reference electrode. Scans were performed from +425 mV to +100 mV at a scan rate of 10 mV s−1 with the values reported as the mean steady state current recorded at +400 mV from three LSVs. The Pt working electrode was polished on Lecloth with 0.05 ␮m alumina before each measurement. Cyclic voltammograms were performed using a 3 mm glassy carbon electrode, Ag/AgCl reference electrode and a platinum wire auxiliary electrode. The glassy carbon electrode was mounted vertically with the electrode surface uppermost. The electrode body was wrapped in ‘Sellotape’ to form a well 2 mm deep above the electrode surface. The reference and auxiliary electrodes entered the well at angles from above. Cells were allowed to settle onto the electrode surface before executing analysis. Total volume above the electrode was 100 ␮L comprising 75 ␮L of cell cream, 10 ␮L of 0.1 M glucose and 15 ␮L of PBS buffer. CVs of YEPD and growth supernatant were the controls with 25 ␮L of PBS added to 75 ␮L of sample. The scan rate was 100 mV s−1 scanning from −0.5 V to 0.8 V and from 250 to 650 mV at a scan rate of 50 mV s. Scan rate studies were performed at intervals from 2 mV s−1 to 400 mV s−1 .

A. adeninivorans S. cerevisiae

0.03

Acellular control

0.025 0.02 0.015 0.01 0.005 0

0

5

10

15

20

25

30

time (min) Fig. 3. Power output in mediator-less S. cerevisiae and A. adeninivorans MFCs. Cell OD600 = 2.5, temp. 37 ◦ C, rotated at 80 rpm, external resistance 100 . Electrode geometric surface area 0.001018 m−2 . (Error bars represent ±1SE of the mean, n = 9.)

N.D. Haslett et al. / Biosensors and Bioelectronics 26 (2011) 3742–3747

3745

60

i (nA)

40

20

A. adeninivorans A. adeninivorans + glu S. cerevisiae S. Cerevisiae + glu 0 0

24

48

72

time (h) Fig. 4. Incubation of S. cerevisiae and A. adeninivorans with the hydrophilic mediator, ferricyanide, with and without glucose (glu). (Error bars represent ±1SE of the mean, n = 9.) The difference in the ‘with’ and ‘without’ glucose trials confirms the responses are from catabolism.

Two experiments were undertaken to investigate the results obtained in Section 3.3. First an incubation of S. cerevisiae and A. adeninivorans with the hydrophilic mediator ferricyanide in the presence and absence of glucose shows that the response of both organisms is essentially the same and this result persisted over the length of time a batch MFC may operate (Fig. 4). The response to glucose demonstrates that this is a signal generated by catabolism. The second, cyclic voltammetry of S. cerevisiae and A. adeninivorans, was performed to search for the presence of redox signals, either from the cells in PBS or from a small volume of supernatant from a large number of cells. Typical voltammograms obtained for cell creams of A. adeninivorans and S. cerevisiae are shown in Fig. 5a. A. adeninivorans Cvs have an irreversible oxidation peak at +0.42 V but an equivalent peak was not observed in CVs of S. cereviseae yeast cream. Fig. 5b shows a plot of the faradaic current for the peak at +0.42 V obtained with A. adeninivorans cell cream at varying scan rates. The peak currents obtained were proportional to the square root of scan rate indicating that magnitude of the peak current is under diffusion control and it can be inferred from this finding that the peak must be from a solution species. This was confirmed by performing CVs on the cell cream (cells in PBS) and on the supernatant from the cream. This was done at time 0 i.e. as soon as the cream was produced and at 1 h. The peak was not present in either the cream or the supernatant at 0 h but was present in both the cream and the supernatant at 1 h indicating that the molecule accumulated in the extracellular buffer over time and that significant direct electron transfer from the cell to the electrode did not occur. Accumulation of the molecule in the supernatant continued during an18 h incubation. The peak was larger in the supernatant than in the cream suggesting that the cells block some of the electrode area in the cream CVs. The 1 h supernatant anodic peak potential was 0.45 V with an anodic peak current at 4.4 ␮A and is shown in Fig. 6.

Fig. 5. (A) CVs of cell creams of A. adeninivorans (A) and S. cerevisiae (B) from −500 mV to 800 mV (vs Ag/AgCl) at 100 mV s−1 . A peak for A. adeninivorans is visible at 420 mV. (B) A plot of peak currents obtained for Arxula cell cream versus the square root of scan rate. Scan rates were from 2 to 400 mV s−1 .

The issue of the use of mediators in continuous flow MFCs has been widely discussed. For example, Schroeder (2007), states in his MFC review that due to a number of severe disadvantages, the mediated MFC ‘has, with the exception of some fundamental research, been generally abandoned’. The main problem is the need for regular addition of the exogenous mediator, ‘which is technologically unfeasible and environmentally questionable’. The maximum power density we have observed with TMPD as the anodic mediator and KMnO4 as the cathode reductant in the A. adeninivorans MFC is 1.03 ± 0.06 W m−2 which is close to the highest yeast MFC output reported by Ganguli and Dunn (2009).

4. Discussion We have used a system comprising lipophilic and hydrophilic mediators to detect large substrate dependant signals in yeast (Baronian et al., 2002). Initially the system was simply transferred to our MFC investigation but in the process, the hydrophilic mediator was discarded because it reduced the potential of the MFC and no disadvantages were noticed by it absence.

Fig. 6. Cyclic voltamogram of the supernatant from A. adeninivorans cell cream centrifuged in a microfuge, 13,000 rpm for 10 min. The scan was from 250 to 650 mV at 50 mV s−1 .

3746

N.D. Haslett et al. / Biosensors and Bioelectronics 26 (2011) 3742–3747

This suggested that an investigation into the retention of mediator by the cell should be made because if the lipophilic mediator remained associated with the cell, the problems associated with the use of mediators in continuous flow MFCs referred to above, may be at least partially negated. We attempted to simulate a continuous flow MFC and Fig. 1 shows that there is a dramatic reduction of current in the lipophilic trial over a short period of time while the signals from the ferricyanide only trials remained constant over the whole experiment. It seems likely that the lipophilic mediator was rapidly lost to the environment resulting in the loss of current. Ferricyanide accesses the membrane surface tPMETs only and the responses indicate that the cells are functioning normally during the experimental period and the larger signals detected when glucose was included confirmed that the signals were from catabolism. As stated in Section 1, the power outputs reported for most yeast MFCs are comparatively low. The low power output of these mediator-less yeast MFCs has been attributed to low catabolic rates, however, the high currents seen in the mediated yeast MFC, and similar growth rates suggest the catabolic rates in yeast and mesophilic bacteria are similar and the very low power outputs are probably due to the difficulty in accessing intracellular electrons. The result of the TMPD experiment led us to investigate mediator-less transfer of electrons from the yeast cell to the electrode. Because most eukaryote redox molecules are located intracellularly and cell membrane tPMETs are probably shielded from contact with an electrode by a 100–200 nm thick rigid wall that surrounds the cell membrane, it seemed that the only possible method of mediator-less transfer would be via a secreted redox molecule. Recent studies, however, have arrived at different conclusions regarding mediator-less electron transfer to an electrode. The majority of reports support the idea that electron transfer from the cell to the electrode is by some kind of direct contact between the cell and the electrode. Ducommun et al. (2010) constructed a dual cathode and dual anode MFC. The anode A had contact with the cells in a glucose water medium whereas anode B was separated from the cells by dialysis membrane. Anode B could however interact with soluble redox molecules that the cells may produce. They report that the anode that had contact with the cells produced current whereas the anode in contact with deionised water did not, indicating that a soluble redox species was not diffusing from the cells through the dialysis membrane to the 2nd anode. They speculate that electrons are either transferred by direct contact of the cell with the electrode or that a short-lived electro-active molecule is released at the time of the impact. Prasad et al. (2007) also support the notion of direct transfer from the cell surface to the electrode. They immobilized P. anomala cells by two different methods; physical adsorption and entrapment of cells on a gold electrode and covalent linkage of cells to a gold electrode modified with a self-assembled monolayer of cystamine. They report that CVs with these electrodes gave peaks that corresponded to the potential of redox enzymes that were isolated from the cell membrane (lactate dehydrogenase and ferrireductase). Conversely Hubenova and Mitov (2010) report that Candida melibiosica gave a small current with no added mediator and concluded that it was probably from secreted metabolites acting as electron shuttles and provided evidence for this type of transfer in their subsequent publication (Hubenova et al., 2010). This study attempted to resolve the question of the origin of the mediator-less MFC signal seen with A. adeninvorans (Fig. 3). S. cerevisiae was added as a reference organism because of the results reported for that yeast by Ducommun et al. (2010) and Halme and Zhang (1995). A. adeninivorans and S. cerevisiae performed differently in the mediator-less MFC. The difference in their performance

was investigated by first checking the magnitude of reduction of a hydrophilic mediator by both species. Fig. 4 shows that the reduction of ferricyanide is essentially the same by both species, both with and without glucose, which suggests that the quantity of electrons available to ferricyanide at the cell membrane surface is similar and may also indicate that catabolic rates of the two species in these conditions are similar (Fig. 5a). The higher current seen in A. adeninivorans in the MFC must therefore be from a species that does not interact with ferricyanide. The next check was to perform a CV of cell pellets of both species. A peak can be seen at 0.42 V vs Ag/AgCl pH 7 in the A. adeninivorans scan, but there is no corresponding peak observable in the S. cerevisiae scan (Fig. 5b). Ganguli and Dunn (2009) also did not report any detectable peaks in CVs of S. cereviseae. The half wave potential for ferricyanide vs Ag/AgCl is 0.28 V at pH 7 and thus it would not oxidise the A. adeninivorans molecule and it would not have been detected in the ferricyanide investigation above. A scan rate study of the yeast cream (Fig. 6) showed that the molecule was soluble. The removal of cells from the yeast cream by centrifugation at 13,000 rpm produced a supernatant that had a larger peak, suggesting that removal of cells permitted more of the molecule to interact with the electrode. The peak potential was independent of pH within the range of pH 4–12 but it did shift positively at lower pHs. That it was not detected at time 0 but it was present at 1 h indicates that the molecule was being produced and secreted over time. Its concentration in the yeast cream supernatant was estimated to be 0.1 mM. The evidence presented above is an unequivocal demonstration that Arxula exports an electro-active molecule and it appears that it provides most of the current in the Arxula MFC. The absence of a peak in the S. cerevisiae scans and the small signals it produces in an MFC also supports this notion. Given previous evidence and that presented here, it is possible that the transfer of electrons from yeast cells to the anode is both by the secretion of redox molecules and by direct electron transfer. It is also likely the contribution of each type of transfer to a yeast fuel cell output will vary with species and probably with cell cultivation and MFC conditions.

5. Conclusion Yeast should make good biocatalysts for MFCs; they are robust, some have high levels of tolerance to variations in environmental conditions e.g. pH, salinity, temperature, most are non-pathogens, and many have high growth rates. This work sought to clarify some of the problems that have been seen in yeast MFCs. Low power output is related to the difficulty in accessing catabolic electrons and not an inherent slow rate of catabolism. An A. adeninivorans MFC was shown to have a higher power output than a S. cerevisiae MFC and this difference was demonstrated to be due to the production of an extracellular redox molecule by A. adeninivorans.

Acknowledgements This work was funded by: The Marsden Fund (The Royal Society of New Zealand), The Dumont d’Urville NZ/France S&T Support Programme (New Zealand Ministry of Research, Science and Technology and French Ministry for Foreign Affairs). The Foundation for Research, Science and Technology NZ (DET biotechnologies project, PROJ-13838-NMTS-LVL). NDH thanks the Tertiary Education Commission, NZ, for a NZ Enterprise PhD scholarship. We thank Vincent Oriez for his assistance with the ferricyanide reduction experiments.

N.D. Haslett et al. / Biosensors and Bioelectronics 26 (2011) 3742–3747

References Baronian, K.H.R., Downard, A.J., Lowen, R.K., Pasco, N., 2002. Appl. Microbiol. Biotechnol. 60, 108–113. Bennetto H. P., 1990. National Centre for Biotechnology Education, www.ncbe.reading.ac.uk. Böer E., Steinborn G., Florschütz K., Körner M., Gellissen G., Kunze G., 2009, in: Satyanarayana, T., Gotthard K., (Eds.), Yeast Biotechnology: Diversity and Applications, 615–634. Bullen, R.A., Arnot, T.C., Lakemen, J.B., Walsh, F.C., 2006. Biosens. Bioelectron. 21, 2015–2045. Chiao, M., Lam, K.B., Lin, L., 2006. J. Micromech. Microeng. 16, 2547. Davis, F., Higson, S.P.J., 2007. Biosens. Bioelectron. 22, 1224–1235. Delaney, G.M., Bennetto, H.P., Mason, J.R., Thurston, C.F., 1984. J. Chem. Technol. Biotechnol. 34B, 13–27. Ducommun, R., Favre, M.-F., Carrard, D., Fischer, F., 2010. Yeast 27, 139–148. Fan, Y., Sharbrough, E., Liu, H., 2008. Environ. Sci. Technol. 42, 8101–8107.

3747

Ganguli, R., Dunn, B.S., 2009. Fuel Cells 9 (1), 44–52. Gunawardena, A., Fernando, S., To, F., 2008. Int. J. Mol. Sci. 9, 1893–1907. Halme, A., Zhang, X., 1995. Reprints of the 6th International Conference on Computer Applications in Biotechnology, 165–170, ISSN: 0080423779. Hubenova, Y., Mitov, M., 2010. Bioelectrochemistry 78, 57–61. Hubenova, Y.V., Rashkov, R.S., Buchvarov, V.D., Arnaudova, M.H., Babanova, S.M., Mitov, M.Y., 2010. Ind. Eng. Chem. Res., doi:10.1021/ie1000949. Kurtzman, C.P., Robnett, C.J., 2007. FEMS Yeast Res 7, 141–151. Logan, B.E., 2009. Nat. Rev. Microbiol. 7, 375–381. Logan, B.E., Hamelers, B., Rozendal, R., Schrˆder, U., Keller, J., Freguia, S., Aelterman, P., Verstraete, W., Rabaey, K., 2006. Environ. Sci. Technol. 40, 5181–5192. Prasad, D., Arun, S., Murugesan, M., Padmanaban, S., Satyanarayanan, R.S., Berchmans, S., Yegnaraman, V., 2007. Biosens. Bioelectron. 22, 2604–2610. Walker, A.L., Walker, C.W., 2006. J. Power Sources 160, 123–129. Walker, G.M., 1998. Yeast Physiology and Biotechnology. Wiley, Chichester. Wilkinson, S., 2000. Auton. Robot. 9, 99–111. You, S., Zhao, Q., Zhang, J., Jiang, J., Zhao, S., 2006. J. Power Sources 162, 1409–1415.