Amperometric Response from the Glycolytic versus the Pentose

Anal. Chem. 2007, 79, 8919-8926

Amperometric Response from the Glycolytic versus the Pentose Phosphate Pathway in Saccharomyces cerevisiae Cells Christer F. Spe´gel,† Arto R. Heiskanen,† Natalie Kostesha,‡ Ted H. Johanson,‡ Marie-F. Gorwa-Grauslund,‡ Milena Koudelka-Hep,§ Jenny Emne´us,† and Tautgirdas Ruzgas*,⊥

Department of Analytical Chemistry, Department of Applied Microbiology, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden, Institute of Microtechnology, Universite´ de Neuchaˆ tel, Rue Jaquet Droz 1, 2007 Neuchaˆ tel, Switzerland, and Faculty of Health and Society, Malmo¨ University, 20506 Malmo¨, Sweden

The two main metabolic pathways involved in sugar metabolism, i.e., the pentose phosphate pathway (PPP) and the glycolytic pathway (GP), were amperometrically monitored using a double-mediator system composed of menadione and ferricyanide. With the use of the Saccharomyces cerevisiae deletion mutant, EBY44, lacking the gene encoding for the branch point enzyme phosphoglucose isomerize, selective amperometric monitoring of the PPP, mainly producing NADPH, and the GP, mainly producing NADH, could be achieved. It was found that the bioelectrocatalytic current was primarily originating from NADPH. This conclusion was supported by metabolite flux analysis, confirming that, in the presence of menadione, the cells increase the rate of NADPH-producing reactions although these processes might be detrimental to cell survival. The higher rate of in vivo NADPHdependent menadione reduction can be ascribed to the fact that the intracellular NADPH/NADP+ ratio is much higher than NADH/NAD+ as well as that the former ratio is more tightly controlled. This tight control over the cofactor ratios is lost upon cell disintegration as observed from spectrophotometric assays using crude cell extract, and amperometric investigations of permeabilized cells indicate a higher rate of NADH- than NADPH-dependent menadione reduction. These in vitro experiments show a higher activity of NADH-dependent than NADPH-dependent menadione-reducing dehydrogenases in S. cerevisiae cells. Bioelectrochemistry, as a research discipline, is strongly acknowledged by its contribution to the development of amperometric enzyme-based biosensors, e.g., glucose biosensors.1,2 A number of new applications are expected from fundamental studies of electron transfer (ET) reactions of redox enzymes at conducting * Corresponding author. Phone: +46-40-6657431. Fax: +46-40-6658100. E-mail: [email protected]. † Department of Analytical Chemistry, Lund University. ‡ Department of Applied Microbiology, Lund University. § Universite ´ de Neuchaˆtel. ⊥ Malmo ¨ University. (1) Clark, L. C. J.; Lyons, C. Ann. N.Y. Acad. Sci. 1962, 102, 29-45. (2) Wang, J. Sens. Update 2002, 10, 107-119. 10.1021/ac0710679 CCC: $37.00 Published on Web 11/01/2007

© 2007 American Chemical Society

materials of macro3 or nanoscopic4,5 dimensions as well as in enzyme redox hydrogel structures.6 Electrochemical investigation of redox processes in living cells are also carried out. Direct ET between intact living cells and electrodes has been observed and is under growing fundamental interest.7-10 The majority of the electrochemical measurements on living cells, however, address biochemical redox processes by using redox mediators, which shuttle electrons between the electrode and intracellular redox reactions.11-18 Following the progress in cell biology, it is of great interest to understand and demonstrate how electrochemical techniques can be exploited, e.g., to monitor a particular intracellular redox process or defined metabolic and signaling pathway or to assay a specific enzyme under in vivo conditions. Exciting examples in this direction are measurements of enzyme activity in living cells, e.g., hydrogenase activity,15 alcohol dehydrogenase activity,14 and study of detoxification of menadione by cells.16,19 (3) Gorton, L.; Lindgren, A.; Larsson, T.; Munteanu, F. D.; Ruzgas, T.; Gazaryan, I. Anal. Chim. Acta 1999, 400, 91-108. (4) Xiao, Y.; Patolsky, F.; Katz, E.; Hainfeld, J. F.; Willner, I. Science 2003, 299, 1877-1881. (5) Luo, X.; Morrin, A.; Killard, A. J.; Smyth, M. R. Electroanalysis 2006, 18, 319-326. (6) Heller, A. Curr. Opin. Chem. Biol. 2006, 10, 664-672. (7) Compton, R. G.; Perkin, S. J.; Gamblin, D. P.; Davis, J.; Marken, F.; Padden, A. N.; John, P. New J. Chem. 2000, 24, 179-181. (8) Zhang, T.; Cui, C.; Chen, S.; Ai, X.; Yang, H.; Shenb, P.; Peng, Z. Chem. Commun. 2006, 2257-2259. (9) Reguera, G.; McCarthy, K. D.; Mehta, T.; Nicoll, J. S.; Tuominen, M. T.; Lovley, D. R. Nature 2005, 435, 1098-1101. (10) Reguera, G.; Nevin, K. P.; Nicoll, J. S.; Covalla, S. F.; Woodard, T. L.; Lovley, D. R. Appl. Environ. Microbiol. 2006, 72, 7345-7348. (11) Ramsay, G.; Turner, A. P. F. Anal. Chim. Acta 1988, 215, 61-69. (12) Ertl, P.; Unterladstaetter, B.; Bayer, K.; Mikkelsen, S. R. Anal. Chem. 2000, 72, 4949-4956. (13) Baronian, K. H. R.; Downard, A. J.; Lowen, R. K.; Pasco, N. Appl. Microbiol. Biotechnol. 2002, 60, 108-113. (14) Ikeda, T.; Kato, K.; Maeda, M.; Tatsumi, H.; Kano, K.; Matsushita, K. J. Electroanal. Chem. 1997, 430, 197-204. (15) Lojou, E.; Durand, M. C.; Dolla, A.; Bianco, P. Electroanalysis 2002, 14, 913-922. (16) Mauzeroll, J.; Bard, A. J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 78627867. (17) Mayer, D.; Naumann, R.; Edler, L.; Bannasch, P. Biochim. Biophys. Acta 1990, 1015, 258-263. (18) Rabinowitz, J. D.; Vacchino, J. F.; Beeson, C.; McConnell, H. M. J. Am. Chem. Soc. 1998, 120, 2464-2473. (19) Mauzeroll, J.; Bard, A. J.; Owhadian, O.; Monks, T. J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 17582-17587.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007 8919

Figure 1. (A) Oxidative metabolic pathways in S. cerevisiae. Glucose and fructose are metabolized by two biochemical pathways, i.e., the PPP, forming ribulose 5-phosphate (Ru5P), and the GP, forming ethanol or acetate. In the S. cerevisiae strain, EBY44, studied in this paper, the gene encoding phosphoglucose isomerase (PGI) had been deleted. As a consequence of this deletion, fructose is metabolized via the GP mainly producing NADH, whereas glucose is metabolized via the PPP generating NADPH. (B) The nonoxidative pathway couples the PPP to the GP by converting Ru5P into glycolytic intermediates through a series of reactions. Metabolites: G6P, glucose-6-phosphate; F6P, fructose6-phosphate; F1,6P, fructose 1,6-bisphosphate; G3P, glyceraldehyde-3-phosphate; DHAP, dihydroxyacetone phosphate; 1,3BPG, 1,3bisphosphoglycerate; 3PG, 3-phosphoglycerate; 2PG; 2-phosphoglycerate; PEP, phosphoenolpyruvate; Pyr, pyruvate; 6PGL, 6-phosphogluconolactone; 6PG, 6-phosphogluconate; Ru5P, ribulose-5-phosphate; AcAld, acetaldehyde; Glycerol3P, glycerol-3-phosphate; R5P, ribose-5phosphate; X5P, xylulose-5-phosphate; Sh7P, sedoheptulose-7-phosphate; E4P, erythrose-4-phosphate; TCA, tricarboxylic acid (cycle).

The most extensive electrochemical studies of intracellular redox processes have, however, been devoted to the measurement of glucose metabolism with the help of artificial redox mediators.13,18,20,21 Extracellular glucose enters a biological cell through hexose uptake channels, i.e., hexose transporters.22 The intracellular glucose is then metabolized by two major biochemical pathways, i.e., the glycolytic pathway (GP), transferring electrons from glucose onto NAD+, and the pentose phosphate pathway (PPP), which transfers electrons from glucose onto NADP+. The production of NADH and NADPH in the GP and the PPP is illustrated in Figure 1. The respiration processes schematically indicated in the mitochondria (Figure 1) are not included in the consideration in this paper since yeast cells cultivated on glucose or fructose, as is the case in this work, do not developed functioning mitochondria.23 This change in the cellular functions means that NADH produced in the GP is regenerated by the production of ethanol (see Figure 1A), i.e., sugar fermentation. Important is that the fermentation enables a biochemical machinery to control the NAD+/NADH balance without requiring oxygen for NADH oxidation. On the basis of this it is clearly understood that yeast cells will counteract the imbalance in the NAD+/NADH ratio (20) Heiskanen, A.; Yakovleva, J.; Spe´gel, C.; Taboryski, R.; Koudelka-Hep, M.; Emne´us, J.; Ruzgas, T. Electrochem. Commun. 2004, 6, 219-224. (21) Zhao, J.; Yang, Z.; Gong, Q.; Lu, Y.; Yang, Z.; Wang, M. Anal. Lett. 2005, 38, 89-98. (22) Ozcan, S.; Johnston, M. Microbiol. Mol. Biol. Rev. 1999, 63, 554-569. (23) Walker, G. M. Yeast Physiology and Biotechnology, 1st ed.; Wiley: Chichester, U.K., 1998.

8920 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

caused by exogenous mediators by adjusting the outcome of the fermentation products (Figure 1A). The control of the NADP+/ NADPH ratio in yeast cells is more complicated. As can be seen in Figure 1A, the cell can limit production of NADPH by directing sugars (e.g., glucose, fructose) to the GP, i.e., NADH production. For this purpose the cells need PGI, especially when glucose is used as the carbon source. Genetic deletion of the gene encoding for PGI will challenge the cells with overproduction of NADPH. This is exactly the situation that has been faced by the deletion mutant EBY44, which otherwise is similar to the “normal” CEN.PK yeast strain. EBY44 is thus an ideal strain for studying the formation of NAD(P)H in the GP and the PPP by either supplying fructose or glucose to the cells in order to address the question, which of the two cofactors, NADH or NADPH, mainly contributes to the measured current at cell-modified electrodes. As summarized in Figure 2, the transfer of the electrons from NADH and NADPH onto redox mediators (usually lipophilic) proceeds in the course of redox reactions catalyzed by a number of intracellular dehydrogenases. A second hydrophilic mediator is often used to facilitate the ET process between the lipophilic mediators, mainly dissolved in cellular membranes, and the electrode (Figure 2). One of the conclusions from the discussion above is that the both cofactors, NADH and NADPH, contribute to the electrode current. This has been confirmed by experiments where the GP or PPP have been inhibited using soluble inhibitors.18,21 The main conclusions presented in these two papers were, however, very different. Zhao et al.21 demonstrated that the current almost entirely (>90%) originated from NADH produced during glyco-

Figure 2. Schematic representation of whole-cell-mediated bioelectrocatalysis. Glucose is metabolized in the GP and the PPP generating reduced NADH and NADPH. These cofactors are subsequently oxidized by the lipophilic mediator menadione (M) with the help of menadionereducing enzymes (MRE). Finally, the highly water-soluble hexacyanoferrate enhances electron transfer (ET) from the poorly soluble reduced menadione (MH2) to the electrode surface. The electron flow proceeds along a potential gradient controlled by the formal potentials (E°’) of the redox components and the electrode potential (Eappl).

lysis, whereas the contribution of the NADPH-generating PPP was negligible. On the other hand, Rabinowitz et al.18 presented an opposite conclusion stating that NADPH is the predominant electron donor involved in the reduction of artificial redox mediators. When comparing these results it should be kept in mind that Rabinowitz conducted inhibition experiments with a number of different eukaryotic cells and referred to the fact that in these cells the NAD+/NADH ratio might be 105 times larger than the NADP+/NADPH ratio. In both of these papers the same compounds for inhibition of the GP (iodoacetate) and the PPP (epiandrosterone) were used. It would be of great interest to explain why the conclusions in the mentioned two papers contradict each other. However, more important would be to understand and experimentally create conditions under which the current at cell-modified electrodes could mostly be dependent on the produced NADH or NADPH. If such an understanding could be gained the methods based on electrochemistry could facilitate more specific monitoring of intracellular NADH and NADPH fluxes, which turn out to be considerably different in normal as compared to cancer cells.24-26 To electrochemically assay unmodified and genetically engineered cells, as has been done in this paper, is another important way to study how intracellular redox reactions are converted into current at cell-modified electrodes. The advantage of using genetic modification of metabolic pathways in comparison with inhibitorbased approaches is that the changes imposed on the cellular structures are much more defined in the former case, whereas inhibitor-based methods usually affect a number of enzymes, being, as a rule, much broader and less strictly defined. A commonly used inhibitor of the GP, iodoacetate, not only affects glyceraldehyde-3-phosphate dehydrogenase but also other enzymes containing free sulfur groups or disulfide bridges, e.g., the alcohol dehydrogenases as well as some glucose-6-phosphate dehydrogenases. (24) Liu, B.; Rotenberg, S. A.; Mirkin, M. V. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 9855-9860. (25) Kroemer, G. Oncogene 2006, 25, 4630-4632. (26) Pelicano, H.; Martin, D.; Xu, R.-H.; Huang, P. Oncogene 2006, 25, 46304632.

We used the lipophilic quinone, menadione (vitamin K3, 2-methyl-1,4-naphthalenedione, E° ) 83 mV vs Ag/AgCl27), and the hydrophilic redox complex, ferricyanide (E° ) 274 mV vs Ag/ AgCl28), as a double-mediator system to probe the function of the GP and the PPP in unmodified and genetically engineered yeast cells. The genetically unmodified laboratory yeast strain CEN.PK 113-7A [MATa his3-∆1 MAL2-8C SUC2]29 was used as a reference for interpreting the metabolism of the genetically engineered EBY44 [ENY.WA-1A pgi1-1∆::URA3]30 strain. EBY44 lacks the gene encoding for the enzyme phosphoglucose isomerase (PGI in Figure 1), responsible for the interconversion of glucose-6phosphate and fructose-6-phosphate, resulting in decoupling of the GP and PPP. Such a genetic decoupling in combination with the use of different carbon sources enables a novel bioelectrochemical study of the NADH-producing pathway (GP) and the NADPH-generating pathway (PPP) in living cells without using inhibitors or cell disruption. To our best knowledge such experiments have not been performed before. The results extend our knowledge on the origin of amperometric response from metabolic pathways as well as the effect of artificial redox mediators on living cells in general. MATERIALS AND METHODS Chemicals. Menadione, glucose, and sodium alginate were from Sigma Chem. Co. (St. Louis, MO), fructose was from Fluka BioChemica (Buchs, Germany), potassium ferricyanide was from Merck (Darmstadt, Germany), yeast protein extraction reagent (Y-Per) was from Pierce (Rockford, IL), and 99.7% (v/v) ethanol was from Solveco Chemicals AB (Ta¨by, Sweden). All other chemicals were analytical grade and used without further purification. The buffer used for measurements contained 10 mM Tris, 10 mM succinic acid, 10 mM CaCl2, and 100 mM KCl, adjusted to pH 5.0 with KOH. All buffers and stock solutions were filtered (Millex-GS 0.22 µm, Millipore Co., Cork, Ireland) to prevent (27) Alonso, L.; Palmero, S.; Mun ˜oz, E.; Sanllorente, S.; Garcı´a-Garcı´a, M. A. Electroanalysis 2000, 12, 757-762. (28) Kolthoff, I. M.; Lingane, J. J. Polarography, 2nd ed.; Interscience Publishers: New York, 1952. (29) Entian, K.-D.; Kotter, P. Methods Microbiol. 1998, 26, 431-449. (30) Boles, E.; Zimmermann, F. K. Mol. Gen. Genet. 1994, 243, 363-368.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

8921

bacterial contamination and stored at 4 °C for a maximum of 2 days. Ultrapure water was obtained from a Milli-Q water purification system (Millipore, Bedford, WY). Apparatus. The working electrode was a platinum band microelectrode, containing four bands each with a width of 25 µm and a spacing of 25 µm, fabricated on a Si/Si3N4 substrate using thin-film technology.31 The signals from all four microelectrodes were summed and recorded as a single response. The working potential was set versus a Ag/AgCl (KCl, saturated) reference electrode, and a platinum wire served as a counter electrode. Amperometry was carried out using an LC-4C amperometric detector (Bioanalytical Systems Inc., Lafayette, IN), and data was transferred to a PC via an A/D converter interface and visualized using homemade data acquisition software. A constant temperature of 30 °C during the electrochemical measurements was maintained using a water-jacketed electrochemical cell, and oxygen concentration was kept constant by vigorous stirring. Optical density (OD) was determined with a Hitachi U-1100 spectrophotometer (Hitachi Ltd. Tokyo, Japan) using 1 mL cuvettes with 10 mm path length. Yeast Strains. The Saccharomyces cerevisiae strain, CEN.PK 113-7A [MATa his3-∆1 MAL2-8C SUC2],29 was obtained from Dr. P. Ko¨tter (Institute of Microbiology, Frankfurt, Germany) and EBY44 [ENY.WA-1A pgi1-1∆::URA3]30 was a gift from Professor E. Boles (Institute of Microbiology, Frankfurt, Germany). Cell Growth. CEN.PK was kept on rich medium agar plates (10 g/L yeast extract, 20 g/L tryptone, and 15 g/L agar in 100 mM phosphate buffer at pH 6.2) with 20 g/L glucose as carbon source and grown in the same medium without agar. EBY44 was kept on rich medium agar plates (10 g/L yeast extract, 20 g/L tryptone, and 15 g/L agar in 100 mM phosphate buffer at pH 6.2) with 20 g/L fructose and 1 g/L glucose as carbon source and grown in the above-described medium without agar. Colonies were taken from agar plates with sterile cotton-tipped applicators and used for inoculation of 200 mL of rich medium, supplemented with 20 g/L glucose or 20 g/L fructose/1 g/L glucose, in 1000 mL shake flasks. CEN.PK and EBY44 cells were grown for 16 and 36 h, respectively, at 30 °C and 200 rpm. Cells where harvested in the early stationary phase. Cell Immobilization and Electrochemical Measurement. Prior to electrochemical experiments, an aliquot of cells was harvested by centrifugation at 3500 rpm for 10 min (Hermle Labortechnik Z230 table centrifuge, Wehingen, Germany) and washed three times in Milli-Q water. The cells were immobilized on the platinum band microelectrode in a calcium alginate gel according to the procedure previously described,20 with the modification that the cell mass was determined by OD measurement at 600 nm, according to a procedure described elsewhere.32 In brief, OD at 600 nm was measured in the cell suspension after dilution so that the value was in the range of 0.25-0.40. The original OD was then calculated by multiplication with the dilution factor. A suspension with the desired OD was prepared by centrifuging the cell suspension and resuspending it in a smaller volume. A 0.5 mL cell suspension with final OD600 of 70 was gently (31) Fiaccabrino, G. C.; Koudelka-Hep, M. Electroanalysis 1998, 10, 217-222. (32) Potvin, J.; Fonchy, E.; Conway, J.; Champagne, C. P. J. Microbiol. Methods 1997, 29, 153-160.

8922

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

mixed with 0.5 mL of 2% (w/v) sodium alginate. The working electrode was dipped in the resulting suspension, and excess material was removed by gently shaking the electrode, creating a thin film of the yeast/sodium alginate suspension. To fix the yeast/alginate layer, the electrode was immersed into a crosslinking solution (100 mM CaCl2 in water) for 1 min to form a stable film. The platinum band microelectrode with immobilized cells was introduced into the thermostatic electrochemical cell, and a potential of +400 mV was applied. The concentrations of glucose, oxygen, menadione, and ferricyanide at the solution-yeast/ alginate interface were maintained as close as possible to their corresponding bulk concentrations by vigorous stirring. All responses are presented relative to the current (set to unity) detected prior to the addition of the carbon source, i.e., the baseline current generated by, e.g., the detoxification of menadione. All results were recorded as individual real-time responses, and repeated experiments showed that the reproducibility was within a relative standard deviation (RSD) of 8% (n ) 3). Metabolite Flux Analysis. The experimental procedure for metabolite flux analysis is described in detail in Supporting Information S2. In brief, 7.5 g/L (dry weight) yeast was incubated in the presence of the carbon sources, glucose or fructose, with or without (control) the mediators, ferricyanide and menadione. Samples were taken after a 30 min incubation period, which corresponded well to the time of exposure of the cells to the mediator system during the electrochemical measurements. Samples were centrifuged, filtered, and analyzed by highperformance liquid chromatography (HPLC) to assay the metabolic end products ethanol, acetate, and glycerol. The formation or utilization of NADH and NADPH was calculated using Table S2.1 in the Supporting Information. Crude Cell Extract Preparation Using Y-Per. Cells were harvested by centrifugation at 3500 rpm for 10 min at 15 °C and washed twice in Milli-Q water. Crude cell extracts were prepared by lysing the yeast cells with Y-Per. An amount of 0.4 g of wet yeast cells per milliliter Y-Per was incubated for 20 min on a rocking table at ambient temperature. Cell debris was spun down at 14 000 rpm for 5 min at room temperature. The crude cell extract was immediately used for spectrophotometric enzyme assay. Protein concentration was determined by performing a Bradford assay using bovine serum albumin (BSA) as the standard. Crude Cell Extract Preparation Using X-press. The cells were harvested by centrifugation at 3500 rpm for 10 min at 15 °C. The cell pellets were washed twice in Milli-Q water and resuspended in 0.1 M potassium phosphate buffer (PB) pH 7.0 at a ratio of 1:1 (w/v). The resulting suspension was then quickly frozen and kept at -20 °C for 2 h. While still frozen, the cells were then disintegrated in a BIOX X-press, model X-5, from AB Biox (Go¨teborg, Sweden). The cell debris was removed by centrifuging at 14 000 rpm for 10 min at room temperature followed by further centrifugation (14 000 rpm, 10 min, room temperature) of the obtained supernatant. The crude cell extract was immediately used for spectrophotometric enzyme assay. Protein concentration was determined by performing a Bradford assay using BSA as the standard.

Spectrophotometric Assay of Menadione-Reducing Enzyme Activity in Crude Cell Extract. The total menadionereducing capacity of crude cell extracts was determined by following the oxidation of NADPH or NADH through the change in absorbance at 340 nm over 30 s at 30 °C. The assay was performed in PB (100 mM, pH 7.0). The following conditions were used: 50 µg of proteins, 20 µL of NADPH or NADH (7.2 mM), 500 µL of PB containing menadione (200 µM), and PB to adjust the volume to 1 mL. The reaction was initiated by the addition of the cofactor. The crude cell extracts and cofactor solutions were kept on ice during the assay. RESULTS AND DISCUSSION Once a solution consisting of oxidizing menadione-ferricyanide mixture is applied onto living cells the redox mediators will be reduced by intracellular redox compounds. The entire process can be considered as a cellular response to the destabilization of the intracellular redox balance (redox stress) by external redox mediators. It is well understood that external mediators will be mainly reduced by NADH and NADPH cofactors in reactions catalyzed by different dehydrogenases.33-35 To exploit electrochemistry for probing cellular functions, it would be of high interest to understand which of these cofactors mostly contribute to the current, since the biological function of the two cofactors is somewhat different. In general, NADH is the substrate for mitochondrial NADH dehydrogenases involved in aerobic ATP production. NADPH on the other hand is the main energy source for the assimilatory (anabolic) pathways, e.g., the biosynthesis of cellular components.36 These general considerations regarding the biological role of NADH and NADPH might be different depending on the type of cells and their growth conditions. Double-Mediator System. The electrode with a surfacebound layer of alginate-entrapped yeast cells was immersed into 10 mM glucose solution, and the current response was measured after the addition of ET mediators. As can be seen in Figure 3, the amperometric signal at yeast-modified electrodes strongly depends on the chosen external ET mediator and their mixture. Ferricyanide alone is able to generate less than 0.15% (