- Email: [email protected]
Low Potential Catalytic Voltammetry of Human Sulfite Oxidase
Electrochimica Acta 199 (2016) 280–289
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Low Potential Catalytic Voltammetry of Human Sulfite Oxidase Palraj Kalimuthua , Abdel A. Belaidib , Guenter Schwarzb , Paul V. Bernhardta,* a b
School of Chemistry and Molecular Biosciences, University of Queensland, Brisbane 4072, Australia Institute of Biochemistry, Department of Chemistry & Center for Molecular Medicine, Cologne University, Zülicher Str. 47, 50674 Köln, Germany
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 December 2015 Received in revised form 21 January 2016 Accepted 24 January 2016 Available online 27 January 2016
Mediated electrocatalytic voltammetry of human sulfite oxidase (HSO) is demonstrated with synthetic one electron transfer iron complexes bis(1,4,7-triazacyclononane)iron(III) ([Fe(tacn)2]3+) and 1,2-bis (1,4,7-triaza-1-cyclononyl)ethane iron(III) ([Fe(dtne)]3+) at a glassy carbon working electrode. The two synthetic electron acceptors for HSO, differing in redox potential by 270 mV, deliver different driving forces for electrocatalysis. Digital simulation of the catalytic voltammetry was achieved with single set of enzyme-dependent kinetic parameters that reproduced the experimental data across a range of sweep rates, and sulfite and mediator concentrations. Amperometry carried out in a stirred solution with the lower potential mediator [Fe(tacn)2]3+ was optimised and exhibited a linear increase in steady state current in the sulfite concentration range 5.0 106 to 8.0 104 M with a detection limit of 0.2 pM (S/ N = 3). The HSO coupled electrode was successfully used for the determination of sulfite concentration in white wine and beer samples and the results validated with a standard spectrophotometric method. ã 2016 Elsevier Ltd. All rights reserved.
Keywords: enzyme molybdenum sulfite
1. Introduction The molybdenum-dependent sulfite oxidizing enzymes comprise sulfite oxidase (SO) and sulfite dehydrogenase (SDH) [1,2]. SO is found in animals and plants whereas SDH is only found in bacteria [3]. Only the plant SO is a true oxidase while all other sulfite oxidizing enzymes donate electrons to cytochrome c. Vertebrate SOs can use either cytochrome c or dioxygen as an electron acceptor. Only one crystal structure is available for a vertebrate SO (from chicken liver) [4] revealing a 103 kDa homodimer in which each subunit contains a negatively charged small heme b domain at the N-terminus and positively charged larger molybdopterin domain at the C-terminus. The heme accepts electrons from the Mo ion following sulfite oxidation. A flexible connects the Mo and heme domains which are more than 30 Å apart in the crystal structure conformation; a distance too great for electron transfer. Spectroscopic and kinetic studies have demonstrated that the heme b domain swings around to be in proximity to the molybdenum active site where electron transfer (Mo to heme) can take place after sulfite oxidation [5–7]. SO
SO3 2 þ H2 O þ 2ðcytcÞox ! SO4 2 þ 2ðcytcÞred þ 2Hþ
* Corresponding author. E-mail address: [email protected] (P.V. Bernhardt). http://dx.doi.org/10.1016/j.electacta.2016.01.181 0013-4686/ ã 2016 Elsevier Ltd. All rights reserved.
ð1Þ
Human sulfite oxidase (HSO) shares a 68% sequence identity with chicken SO [4]. Among the eukaryotic SOs, HSO has been studied extensively because of its role in the potentially fatal disease SO deficiency [8,9]. The physiological role of SO is to remove toxic sulfite (a product of organo-sulfur compound metabolism) and covert it to chemically inert sulfate. Despite its name the physiological electron acceptor of SO is in fact cytochrome c (Eq. (1)). In the catalytic reaction, SO is active in its fully oxidized state (MoVI) in which molybdenum is coordinated by a cysteine thiolate, the dithiolene group of molybdopterin, and two terminal oxygen atoms as shown in Scheme 1 [7,10–12]. Upon reaction with sulfite, one oxido ligand is transferred to sulfite to give sulfate and the Mo ion is reduced its tetravalent state. Subsequently, hydroxide displaces sulfate, and the removal of this hydroxido ligand proton occurs spontaneously when the Mo ion is reoxidised to its hexavalent state by two cytochrome c molecules. There have been a number of electrochemical investigations of SO and SDH enzymes from different organisms. In these cases the electrode is the ultimate electron acceptor resulting in an anodic catalytic current. Electrons may be transferred directly from the enzyme [2,13–17] or via a mediator which may be synthetic [18– 20] or natural (cytochrome c) [18,21–25]. The dynamics HSO are potentially problematic for efficient electrocatalysis. While the Mo and heme cofactors are separated, the enzyme is unable to be reactivated through reoxidation. It is of interest whether confinement of HSO enzyme to a thin layer at the electrode surface suppresses this motion. Spectroelectrochemistry of HSO showed the FeIII/II redox potential to be +62 mV vs NHE (pH
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
281
Scheme 1. Simplified catalytic cycle of SO reduced forms of enzyme and substrate in red and oridised forms of enzyme and product in blue.
7.5) [6]. At applied electrochemical potentials above this value, the enzyme will be continually reoxidsed and reactivated for sulfite oxidation. To achieve this we employed two artificial electron acceptors; the hexa-amine complexes [Fe(tacn)2]3+ and [Fe(dtne)]3+ (Fig. 1) with redox potentials of +144 and +415 mV vs NHE, respectively which present significantly different overpotentials but are structurally almost the same. The higher FeIII/II redox potential of [Fe(dtne)]3+ is due to the presence of two tertiary amines compared to the all-secondary amine [Fe(tacn)2]3+. It is notable that nonspecific oxidation of sulfite at an electrode (without any enzyme present) is inevitable above ca. +550 mV vs NHE [26] and this places an upper bound on the redox potential of any mediator in a sulfite oxidizing electrochemical system. An additional feature of this study is electrochemical simulation of the experimental voltammetry. Given that the catalytic cycle involves several steps, some chemical reactions between HSO and sulfite/sulfate and some being outer sphere electron transfer reactions between HSO and the mediators, a set of rate constants can be defined (Scheme 2). These rate constants must be able to
reproduce the catalytic voltammetry under a variety of conditions including sweep rate, mediator concentration and sulfite concentration. Finally, amperometry is employed to estimate the lowest detection limit and linear current response for the determination of sulfite in aqueous solution and in the quantification of sulfite in beer and wine samples where it is a commonly found as an additive to combat spoilage from oxidation and microbial activity [27,28]. 2. Experimental 2.1. Materials Human sulfite oxidase (HSO) was purified in E. coli TP1000 as previously described [29]. The iron complexes bis(1,4,7-triazacyclononane) iron(III) bromide ([Fe(tacn)2]Br3) [30] and 1,2-bis (1,4,7-triaza-1-cyclononyl) ethane iron(III) bromide ([Fe(dtne)] Br3.3H2O) [31] were synthesized according to the previous procedures. Sodium sulfite and 5,50 -dithio-bis(2-nitrobenzoic acid) (Ellman's reagent) were purchased from Aldrich and were
3+ HN
NH
HN
3+ HN HN
Fe HN
N Fe
NH N H
[Fe(tacn) 2] 3+ E' (Fe III/II) +144 mV vs NHE
HN
N N H
[Fe(dtne) ]3+ E' (Fe III/II) +415 mV vs NHE
Fig. 1. Molecular structures and redox potentials of the mediators used in this study.
282
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
used as received. The beer and wine samples were purchased from local retail outlets. All other reagents used were of analytical grade purity and used without any further purification. Tris acetate buffer (50 mM) was used for all experiments at pH 8.0. For pHdependent experiments, the mixture of buffers (20 mM MES buffer pH 5.5–6.7, 20 mM Bis-Tris buffer pH 5.8–7.2, 20 mM Tris buffer pH 7.0–9.0, 20 mM CHES buffer pH 8.6–10.0 and 20 mM CAPS buffer pH 9.7–11.1) was used and the desired pH was obtained with dilute acetic acid or NaOH. All solutions were prepared with ultrapure water (resistivity 18.2 MV.cm) from a Millipore Milli-Q system. 2.2. Electrochemical Measurements and Electrode Cleaning The cyclic voltammetry (CV) and chronoamperometry experiments were carried out with a BAS 100B/W electrochemical workstation. A three-electrode system was employed comprising a glassy carbon (GC) disk working electrode, a Pt wire counter, and an Ag/AgCl reference electrode (+196 mV vs NHE). Potentials are cited versus NHE. Experiments were carried on solutions that had been purged with argon gas for 30 min. The GC electrode was polished with 0.50 and 0.05 mm alumina slurry and then rinsed thoroughly with water. Then the electrode was sonicated in water for 10 min to remove adsorbed alumina particles and dried in a nitrogen atmosphere. The electro-active surface area of the GC electrode (A) was determined from the cyclic voltammetry of 1 mM ferrocene methanol [32] in 0.1 M KCl solution at different sweep rates using the Randles-Sevcik equation (Eq. (1)) [33]. ip = (2.69 105)n3/2ADo1/2Con1/2
(1)
The standard diffusion coefficient (Do) of ferrocene methanol is 6.7 106 cm2 s1 [34],ip is the measured current maximum, n is the number of electrons, Co is concentration of analyte (mol cm3), and n is the sweep rate (V s1). The variation of the catalytic current (ilim) as a function of sulfite concentration was fit to Michaelis–Menten kinetics (Eq. (2)) yielding KM,app (the apparent Michaelis constant) and imax (the effective electrochemical turnover number, imax = nFA[HSO]) [35]. h i imax SO2 3 h i ð2Þ ilim ¼ K M;app þ SO2 3 The pH dependence of the catalytic current was modeled by Eq. (3) [11] which is applicable for an active form of the enzyme that is deactivated by either deprotonation of an acid at high pH (pKa1) or protonation of a base at lower pH (pKa2). imax ðpHÞ ¼
iopt 1 þ 10ðpHpK a1 Þ þ 10ðpK a2 pHÞ
ð3Þ
2.3. Enzyme Electrode Preparation A 3 mL droplet of HSO (66 mM) in 50 mM Tris buffer (pH 8.0) was pipetted onto the conducting surface of an inverted, freshly prepared GC working electrode and this was allowed to dry to a film at 4 C. To prevent protein loss the electrode surface was carefully covered with a semi-permeable dialysis membrane (SERVA MEMBRA-CEL, molecular weight cut off 3500 Da), presoaked in water. The dialysis membrane was pressed onto the electrode with a Teflon cap and fastened to the electrode with a rubber O-ring to prevent leakage of the internal membrane solution. The resulting modified electrode was stored at 4 C in 50 mM Tris buffer (pH 8.0) when not in use. The enzyme was confined to a thin layer beneath the membrane while substrate and mediators were able to diffuse across the dialysis membrane.
2.4. Electrochemical Simulation The DigiSim program (version 3.03b) was employed to simulate the experimental cyclic voltammograms [36]. The experimental parameters restrained in each case were the working electrode surface area (0.055 cm2) and the double-layer capacitance (12 mF). Semi-infinite diffusion was assumed and all pre-equilibration reactions were enabled. The apparent redox potential of mediators was determined from control experiments in the absence of enzyme or substrate. The diffusion coefficients of mediators were also obtained in the presence of a dialysis membrane covering the electrode by simulation of the cyclic voltammetry at different sweep rates in the absence of substrate and enzyme to give value of 5 106 cm2 s1. The diffusion coefficients for HSO and substrate were taken to be 5 107 and 5 106 cm2 s1 [26]. These values were kept constant for simulating the various substrate- and mediator-concentration-dependent CVs. The heterogeneous rate constant (k0) was determined from simulating the sweep rate dependence of the anodic peak to cathodic peak separation of mediators (in the absence of HSO) and then held constant thereafter. The only values that were allowed to differ were the rate constants for the outer sphere electron transfer reaction between each mediator and enzyme (k4, k40 , k-4 and k-40 in Scheme 2). It was assumed that k4 = k40 and k–4 = k-40 i.e. oxidation of either the MoV/FeII or MoVI/FeII forms of HSO proceeded at the same rate, which is reasonable given that the heme cofactor is the site of oxidation and its redox potential is known (+62 mV vs NHE) [6]. 2.5. Spectrophotometric Sulfite Determination As a complement to amperometric sulfite determination the results were validated using Ellman's reagent which is cleaved by sulfite to form an organic thiosulfate and 5-mercapto-2-nitrobenzoate stoichiometrically; the latter being determined spectrophotometrically [37]. 3. Results and Discussion 3.1. Electrocatalytic Mechanism of HSO The electrocatalytic mechanism of HSO is illustrated in Scheme 2. The single electron transfer acceptors [Fe(tacn)2]3+ and [Fe(dtne)]3+ used in the present study are synthetic substitutes for cytochrome c and so two consecutive one electron transfer reactions are necessary to regenerate the reduced HSO to its active form after it has been reduced by sulfite. There are also two intramolecular electron transfer (IET) processes. The first IET step occurs when MoIV transfers one electron to the oxidized ferric heme b cofactor. The ensuing MoV/ FeII species transfers an electron to the artificial electron acceptor producing the MoV/FeIII state (rate constant k4). A second IET step generates the MoVI/FeII state, and reduction of a second molecule of mediator (rate constant k40 ) regenerates the fully oxidized MoVI/ FeIII state of the enzyme. The sum of the forward and reverse IET2 steps is known (ket > 400 s1 at pH 7) [38] and this first order reaction is always much faster than the rates of the (second order) outer sphere redox reactions (k4, k40 ) which are slowed down by the low concentrations of mediators and HSO used in this experiment. For this reason we have not included either IET step in our kinetic model i.e. it is assumed to be fast and never rate limiting. The substrate (sulfite) and mediator ([Fe(tacn)2]3+ or ([Fe (dtne)]3+) are under diffusion control while HSO is confined to the small volume under the membrane but still may diffuse within that space. We have assumed that the catalytic reaction follows
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
283
Scheme 2. Mediated Electrochemically Driven Catalysis of HSO.
Michaelis-Menten kinetics comprising substrate binding (k1/k-1), turnover (k2/k-2) and product release (k3/k-3). A simplified double substrate ‘ping-pong’ mechanism is appropriate for this type of catalysis. 3.2. Mediator Voltammetry Interestingly upon introduction of 10 mM of the mediator [Fe (tacn)2]3+ into the electrochemical cell no significant redox response was observed initially (in the first cycle) at the dialysis membrane-covered HSO modified GC electrode (Supporting Information, Fig. S1A). In the second cycle the redox response of [Fe(tacn)2]3+ emerged and increased in current up to about 12 cycles where a consistent waveform was established. The GC/HSO electrode shows a well-defined redox wave centred at ca. +56 mV vs NHE for 10 mM [Fe(tacn)2]3+ with a peak to peak separation of only 42 mV in 50 mM Tris buffer solution (Fig. 2, curve a). There are several important features to note. The increasing current at the GC/HSO electrode as a function of cycle number (Supporting Information, Fig. S1A) indicates that the [Fe (tacn)2]3+ molecules only cross the membrane slowly from the bulk solution i.e. flux across the membrane during the sweep is insufficient to keep pace with depletion of the mediator from the diffusion layer in the initial sweep. Moreover, the establishment of a consistent CV after several cycles indicates that the amount of [Fe (tacn)2]3+ that eventually accumulates under the membrane is sufficient to sustain catalysis i.e. it is not depleted during the sweep (see below). Secondly the peak to peak separation is less than 57 mV but greater than 0 mV, which is intermediate of a response governed by linear diffusion and that seen in a thin layer cell [33]. Furthermore the symmetry of the wave is a hybrid of the tailing waveform characteristic of normal linear diffusion and the symmetrical wave characteristic of a thin layer cell due to the confines of the membrane [33]. It is also apparent that the observed currents are much greater than would be expected for a
10 mM solution on the basis of Eq. (1). It appears that a significant amount of [Fe(tacn)2]3+ is concentrated on the inner side of the dialysis membrane and escape to the bulk solution (10 mM) is slow. If the electrode is transferred to a buffer solution containing no Fe complex then the current response is significantly diminished upon continuous cycling (Supporting information Fig. S1B). On the basis of the maximum peak height attained and Eq. (1) we have estimated the amount maximum concentration of [Fe(tacn)2]3+ under the dialysis membrane to be ca. 120 mM. The redox response was investigated at different scan rates. Both oxidation and reduction currents increase with scan rate from 10 to 100 mV s1 (Supporting information Fig. S2A) but the linear increase of peak height with the square root of sweep rate (R2 = 0.999, with zero
Fig. 2. CVs obtained for 10 mM [Fe(tacn)2]3+ in the absence (a) and presence (b) of 5 mM sulfite at the GC/HSO electrode in 50 mM Tris buffer (pH 8) at a sweep rate of 5 mV s1.
284
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
intercept) is more consistent with a linear diffusion controlled process than a thin layer response (linear increase of current with sweep rate). On balance the mediator redox response is predominantly under diffusion control (Supporting information Fig. S2B). Also of interest is the deviation of the [Fe(tacn)2]3+/2+ redox potential under these conditions (E' +56 mV vs NHE) from that obtained in solution without a membrane and in the absence of HSO (+144 mV). This is not a consequence of the membrane as the CVs of [Fe(tacn)2]3+ alone are the same in the absence or presence of the membrane (Figs. S1C and S1D). The pronounced cathodic shift in the [Fe(tacn)2]3+/2+ redox potential is only seen in the presence of HSO and this is attributed to the formation of a noncovalent (outer sphere) complex with HSO under the membrane. The natural electron acceptor of HSO is the highly positively charged protein cytochrome c which is thought to bind (noncovalently) at a negatively charged surface of HSO adjacent to the heme cofactor. The affinity of the tri-positively charged [Fe(tacn)2]3+ evidently mimics cytochrome c and presumably interacts with the same highly negatively charged surface of HSO and this interaction lowers the redox potential of the mediator. The same analysis was carried out with [Fe(dtne)]3+ and once again the CV behaviour is consistent with a predominantly diffusion controlled response this time at a much higher redox potential (Supporting Information Fig. S3). Again the observed [Fe (dtne)]3+/2+ redox potential (+355 mV) was shifted cathodically from its value in solution (+410 mV) [31]. 3.3. Catalytic Voltammetry Although HSO has two electroactive centers (Mo and heme) no redox responses were observed from either cofactor in the absence (or presence) of sulfite at the GC electrode without mediator present (data not shown). This is not unexpected as direct electrochemistry of HSO has only been observed at chemically modified Au [39], Ag [17] or Sb-doped SnO2 [40] electrodes and only quite weak responses were seen. In the presence of HSO (under the membrane), [Fe(tacn)2]3+ and sulfite (5 mM), a well-defined classic sigmoidal waveform is seen and the limiting anodic current increases by an order of magnitude (Fig. 2, curve b). No cathodic wave is present in this case and obviously the forward and backward sweeps are same when the charging current is taken into account. In a control experiment, we found that the redox response of [Fe(tacn)2]3+ is insensitive to 5 mM sulfite at a bare GC electrode (data not shown) within the
potential window of 200 to +350 mV vs NHE indicating that mediator alone cannot oxidize sulfite. Thus, the sigmoidal voltammetric waveform at the enzyme modified electrode is characteristic of a catalytic homogeneous reaction coupled to heterogeneous electron transfer (EC' mechanism) [33] where sulfite is oxidized enzymatically yielding the reduced form of enzyme (MoIV), which is reoxidized by electro-generated [Fe (tacn)2]3+. 3.4. HSO-Sulfite Reaction The reaction between HSO and sulfite was investigated by varying the sulfite concentration while maintaining a constant concentration of mediator and enzyme. The examples in Fig. 3 show the CVs of the GC/HSO electrode in the presence of 10 mM [Fe (tacn)2]3+ (Fig. 3A) and 20 mM [Fe(dtne)]3+ (Fig. 3B) at a sweep rate of 5 mV s1 in Tris buffer (pH 8). In both cases, the CVs take the form of an asymmetric transient catalytic wave up to 800 mM sulfite with a pronounced anodic peak but no corresponding cathodic current. The peak is due to mass transport limitations where sulfite becomes depleted at the electrode surface due to the rate it is consumed by HSO, which cannot be sustained by diffusion from the bulk solution across the membrane. It is apparent that as the sulfite concentration is increased further, (3.2 mM) the wave increases in magnitude and the sharp transient form of the wave becomes more symmetrical. Ultimately (> 4 mM sulfite), the transient wave becomes sigmoidal where the concentration of sulfite within the reaction layer is constant during the sweep. The sigmoidal waveform is indicative of an electrochemical steady state i.e. the oxidized form of mediator is consumed (by homogeneous reaction with HSOred) at the same rate that it is generated at the electrode surface and mass transport of sulfite from the bulk is fast enough to ensure its concentration is constant under the membrane. CVs at all concentrations of sulfite examined appear in the Supporting Information with [Fe(tacn)2]3+ (Fig. S4A) and [Fe(dtne)]3+ (Fig. S5A) as mediator. The catalytic sulfite oxidation current increased linearly up to 800 and 1600 mM sulfite before saturating at millimolar concentrations. Apparent Michaels constants (KM,sulfite) of 512 mM ([Fe (tacn)2]3+) and 970 mM ([Fe(dtne)]3+) were obtained (Supporting Information, Fig. S6). Of course KM,sulfite should be mediatorindependent so these are not true Michaelis constants and they have little mechanistic relevance other than defining the approximate linear current response of the electrode. The true KM,sulfite value for HSO in solution is 9 mM in reaction with its physiological
Fig. 3. CVs obtained for varying sulfite concentrations in the presence of (A) 10 mM [Fe(tacn)2]3+ and (B) 20 mM [Fe(dtne)]3+ at GC/HSO electrode in 50 mM Tris buffer (pH 8) at a sweep rate of 5 mV s1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
285
Fig. 4. CVs obtained for varying (A) [Fe(tacn)2]3+ and (B) [Fe(dtne)]3+ concentrations in the presence of 4 mM sulfite at GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at a sweep rate of 5 mV s1.
electron acceptor cytochrome c [41]. So utilizing the mass transport limitations presented by the membrane, the linear response of the electrode to sulfite is increased by at least 2 orders of magnitude. We have observed similar observations in other Mo enzyme systems [25,26,42]. 3.5. HSO-Mediator Reaction The HSO-mediator reaction was examined with increasing [Fe (tacn)2]3+ and [Fe(dtne)]3+ concentrations in the presence of a high (constant) concentration of sulfite. Fig. 4A displays the effect of varying the concentration of [Fe(tacn)2]3+ in the presence of 4 mM sulfite at a sweep rate of 5 mV s1. At 2 mM [Fe(tacn)2]3+, a sigmoidal voltammogram is found, which is indicative of an electrochemical steady state; the forward and backward sweeps are the same and the catalytic current is switched on and off in a Nernstian fashion. However, as the concentration of [Fe(tacn)2]3+ increases (to 4 and 6 mM), the waveform becomes asymmetric due to an excess of the oxidized form of mediator being produced at the electrode, which overwhelms the limiting amount of HSOred formed during the sulfite oxidation step. In other words the concentration of [Fe(tacn)2]3+ is no longer at steady state. Fig. 4B displays similar experiments but this time with increasing concentrations of [Fe(dtne)]3+ (6, 12 and 18 mM). The observed sigmoidal wave at 6 mM turns to a transient form upon increasing concentration to 12 and 18 mM [Fe(dtne)]3+. Data collected at all mediator concentrations are shown in Supporting Information Figs. S4B and S5B.
catalysis involving substrate binding [45]. The higher pKa value observed here may be due to Tyr343 deprotonation at high pH which is believed to be close to the active site and involved in Hbonding with the substrate [45]. The pH profile was independent of the direction of titration and catalytic activity was fully restored when the solution pH was returned to its optimal value. Furthermore the voltammetry of both [Fe(tacn)2]3+ and [Fe (dtne)]3+ are pH-independent within this range. The complex [Fe(tacn)2]3+ can be deprotonated but only at much higher pH (pKa 11.7) [46]. 3.7. Electrochemical Simulation In recent years, we have employed digital simulation for a better understanding of the mechanism of mediated enzyme electrochemical reactions [26,47–50]. The objective of the simulation is to obtain the rate constants defined in Scheme 2 that reproduce all voltammetric features over a range of sweep rates, substrate and mediator concentrations. The voltammetric sweep rate is a significant variable to elucidate the kinetics of electrochemical processes coupled with chemical reactions. The DigiSim program enables the same set of kinetic parameters to be optimized to CVs measured across a range of sweep rates, but under an identical set of concentrations (HSO, mediator and sulfite). When the concentrations of mediators and
3.6. pH Dependence The pH dependence of the catalytic sulfite oxidation current at the GC/HSO electrode was explored in the range 5.5 < pH < 11 in 100 mM mixed buffer solution. Fig. 5 depicts the baseline subtracted maximum catalytic current as a function of pH. The actual CVs are provided in the Supporting Information (Fig. S7). The catalytic current exhibits a pH optimum of 8.5 which is similar to that reported for HSO at an osmium redox polymer modified electrode [43] as well as in a solution assay for HSO with its natural electron acceptor cytochrome c [44]. A bell shaped profile obtained by the application of Eq. (3) enabled the two pKa values to be determined (7.2 and 9.8); the lower value defining the protonation constant of a base that switches off catalysis and the higher one being the protonation constant of a base that switches on catalysis. It has been proposed that Tyr343 plays an important role in HSO
Fig. 5. Plot of the pH dependence of the maximum catalytic oxidation current at the GC/HSO electrode with 4 mM sulfite and in the presence of 10 mM [Fe(tacn)]23+ in 100 mM mixed buffer solution at a scan rate of 5 mV s1. The solid curve is obtained from a fit to the experimental points using Eq. (3) (pKa1 9.8 and pKa2 7.2).
286
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
Fig. 6. Experimental (solid lines) and simulated (broken lines) sweep rate dependent CVs obtained for 800 mM sulfite in the presence of (A) 10 mM [Fe(tacn)2]3+ and (B) 20 mM [Fe(dtne)]3+ at GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at different scan rates.
sulfite are varied then ideally the same parameters reproduce CVs measured under those situations as well. Fig. 6 shows experimental and simulated CVs for 800 mM sulfite in the presence of 10 mM [Fe(tacn)2]3+ (Fig. 6A) and 20 mM [Fe(dtne)]3+ (Fig. 6B) as a function of sweep rate. All other sweep rate dependent simulated voltammograms recorded as a function of various mediator and substrate concentrations are given in the Supporting Information (Fig. S8-S13). The same features are well reproduced for both mediators. In Fig. 6A as the scan rate increases from 5 to 50 mV s1, the asymmetric transient CV becomes reversible as electrochemical oxidation and reduction of the mediator is too rapid and the HSO-mediator reaction becomes uncompetitive. A very similar trend observed for [Fe(dtne)]3+ as shown in Fig. 6B and these features are also well reproduced in the simulation. The substrate binding rate constant (k1) is well defined by simulation and changing its value has a major influence on the quality of the fit between experiment and theory. Although k2 is also an important value, and defines the maximum current at HSO saturation, its value is entangled with the concentration of HSO under the membrane (Eq. (2), imax = nFAk2[HSO]). The concentration of HSO under the membrane is only known approximately because the volume under the membrane cannot be measured directly but instead determined by introduction of an known amount of external standard e.g. cytochrome c as reported
previously [25]. In this case k2 is the same as determined for HSO at pH 8.0 [45]. The product dissociation rate k3 value has little influence on the CV in this case if allowed to deviate from its optimal value (values in the range 0.5 to 50 s1 gave the same result here. The k4 values are also accurately determined although the same issues regarding the accurate concentration of HSO under the membrane introduce some uncertainty. The CVs as a function of increasing mediator concentration of [Fe(tacn)2]3+ (1 to 4 mM) and [Fe(dtne)]3+ (2 to 8 mM) in the presence of a saturating (4 mM) sulfite concentration and sweep rate of 5 mV s1 are represented in Fig. 7. An approximately sigmoidal wave is observed at a low concentration of [Fe(tacn)2]3+ (1 mM) and this wave becomes progressively peak-shaped (transient) as the higher concentration of mediator overwhelms the HSO present and the electrochemical steady state of mediator breaks down. The same set of parameters also reproduced CVs measured at various sulfite concentrations. Fig. 8 displays the anodic current response of the GC/HSO electrode as function of sulfite concentration in the presence of 10 mM of [Fe(tacn)2]3+ (Fig. 8A) and 20 mM of [Fe(dtne)]3+(Fig. 8B). At lower concentrations of sulfite (400 mM), the voltammograms took on a reversible transient form due to the excess amount of oxidized mediator at the electrode surface. The wave becomes firstly asymmetric as the
Fig. 7. Experimental (solid lines) and simulated (broken lines) CVs obtained for varying mediator concentration in the presence of 4 mM of sulfite (A) [Fe(tacn)2]3+ and (B) [Fe (dtne)]3+ at the GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at a sweep rate of 5 mV s1.
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
287
Fig. 8. Experimental (solid lines) and simulated (broken lines) CVs obtained for varying sulfite concentration in the presence of (A) 10 mM of [Fe(tacn)2]3+ and (B) 20 mM of [Fe (dtne)]3+ at the GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at a sweep rate of 5 mV s1.
sulfite concentration rises to 1600 mM, where catalysis becomes significant, and tailing is due to sulfite depletion. Finally as the enzyme is saturated with sulfite (> 3 mM) the expected steady state sigmoidal waveform is observed. The catalytic current increased linearly with sulfite concentrations up to 800 and 1600 mM in the presence of [Fe(tacn)2]3+ and [Fe(dtne)]3+, respectively. This discrepancy is due to the different driving forces of the mediators (different k4 values) and the catalytic current attains saturation quickly when the oxidative driving force is small ([Fe(tacn)2]3+) but extends to a much higher value when the driving force is relatively high ([Fe(dtne)]3+). We have reported similar phenomena before in the mediated electrochemistry of the enzyme DMSO reductase [49]. Briefly the more rapid reaction between HSO and [Fe(dtne)2]3+ leads to faster sulfite depletion under the membrane and thus higher concentrations of sulfite are needed to reach the point where the enzyme becomes truly saturated. An additional point of interest is the low redox potential of [Fe(tacn)2]3+/2+ in these experiments (+56 mV vs NHE). This is similar to the heme b redox potential reported for HSO [6,41] but much lower than the natural redox partner cytochrome c (ca. +260 mV). Thus the [Fe(tacn)2]3+:HSO catalytic system operates as an extremely low catalytic potential (200 mV lower than the naturally mediated electrochemical reaction), which is advantageous for avoiding interference from the nonspecific oxidation of other species in solution.
3.8. Analysis of Kinetics Parameters The rate and equilibrium constants defined in Scheme 2 and presented in Table 1 reproduced all our experimental voltammetry carried out at different sweep rates, concentrations of sulfite and both mediators. Importantly there same set of mediator-independent parameters (k1-k3) were used for the two mediators. The accurate determination of multiple parameters is problematic in that some parameters will have no effect on the CV depending on the concentrations of the reactants. That is the rate limiting step in Scheme 2 will vary depending on the conditions. The sulfite binding rate constant determined here (k1 = 106 M1s1) has not been reported for HSO. This value may even be an underestimate due to mass transport limitations set by the membrane. The turnover number (k2 = 25 s1) obtained in the simulations is consistent with experimental value reported by solution assays for wild type HSO with its natural electron acceptor cytochrome c in pH 8 (27 s1) [41]. The larger outer sphere electron transfer rate constant for the HSO:[Fe(dtne)]3+ reaction compared with the HSO:[Fe(tacn)2]3+ reaction is consistent with Marcus theory (log ket / DG2) [51] given that higher redox potential of [Fe(dtne)]3+ delivers a greater driving force. The rate constant (k4) obtained for the reaction between the SO and higher potential mediator [Fe (dtne)]3+ is similar to the value reported for HSO in reaction with cytochrome c (4.0 106 M1 s1) [23]. 3.9. Amperometric Sulfite Determination
Table 1 Kinetic parameters (defined in Scheme 2) from electrochemical simulation. E0 (mV vs NHE)
k4 (M1 s1)a k-4 (M1 s1)b
k1 (M1 s1) k-1 (s1) k2 (s1) k-2 (s1) k3 (s1) k-3 (M1 s1) KM,Sulfite (mM)d a
[Fe(tacn)2]3+ 56 mV Mediator dependent 1.0 104 0.1c
[Fe(dtne)]3+ 355 mV 2.0 106 2c
Mediator independent 1.0 106 20 25 5.0 102 c 5c 1.0 102 c 102
k4 = k40 ; b k–4 = k-40 ; c approximate (simulation not sensitive to this parameter); KM,sulfite = (k2 + k1)/k1.
d
Sulfite is used as a preservative in food and beverages to prevent oxidation and bacterial growth and to control enzymatic reactions during production and storage [27,28]. Nevertheless, sulfite has been regulated since the realization that it may cause asthmatic attacks and allergic reactions in some people [52,53]. Typically, a warning label is required for any food or beverage containing more than 10 ppm (125 mM) sulfite so its accurate measurement in solution is important. Here we were able to achieve sulfite determination using the GC/HSO electrode and the low potential mediator [Fe(tacn)2]3+ in an amperometric experiment. Fig. 9A illustrates the amperometric i-t curve for the catalytic oxidation of sulfite at a GC/HSO electrode (covered with a dialysis membrane) in a homogeneously stirred 50 mM Tris buffer solution (pH 8) at an applied potential of +150 mV vs NHE. An initial baseline current response was stabilized for about 2 min in the presence of 10 mM [Fe(tacn)2]3+ at the GC/HSO electrode to ensure the mediator was able to concentrate under the membrane (vide supra). Upon addition of 10 mM of sulfite to the stirred solution in the
288
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
Fig. 9. (A) An amperometric i t curve obtained for the determination of sulfite at the GC/HSO electrode in stirred 50 mM Tris buffer solution (pH 8). Each increment corresponded to a 10 mM increase in sulfite which was injected at regular intervals of 100 s. The electrochemical cell contained 10 mM [Fe(tacn)2]3+ and the electrode was poised at +150 mV vs NHE. (B) Plot of the steady state current as a function of sulfite in the linear range.
electrochemical cell, the catalytic anodic current increased suddenly and reached a plateau (steady state) within 3 sec. Further 10 mM sulfite increments at intervals of 100 s led to a regular and consistent step in the current. Furthermore, the amperometric current increased linearly with sulfite concentration from 10 to 180 mM and the detection limit [54] was found to be 0.2 pM (S/ N = 3) (Fig. 9B). The obtained detection limit is even lower than we reported with Starkeya novella sulfite dehydrogenase on a 11mercaptoundecanol monolayer modified Au electrode (44 pM) [25]. In separate voltammetry experiments, we found that the anodic catalytic current increased linearly from 5 to 800 mM at +150 mV vs NHE (Supporting Information, Fig. S6). Spricigo et al. reported a sulfite biosensor by co-immobilization of HSO within an osmium redox polymer on a carbon screen-printed electrode [43]. The biosensor operates at +100 mV vs NHE with detection limit (0.5 mM) and linearity (1 to 100 mM). 3.10. Determination of Sulfite in Wine and Beer Samples In order to demonstrate the practical application of the present biosensor, the enzyme modified electrode was used for the determination sulfite concentration in white wine and beer samples. Two beers and one white wine sample were obtained commercially were analyzed for sulfite using the present biosensor and this was validated by the standard spectroscopic method [37] using Ellman's reagent. Beer and wine, as prepared, are acidic (pH 4) and at this pH HSO is inactive as shown in Fig. 5. Therefore, the beer and wine samples were neutralized with dilute NaOH then diluted with Tris buffer solution and analyzed immediately without any other pretreatment. The method of standard additions was employed by injecting known amounts of sulfite to each beer or wine sample within the linear range and measuring the increase in catalytic current which enabled the original sulfite concentration to be Table 2 Determination of sulfite in wine and beer samples using present method and also compared with the standard spectroscopic method. Sample
Present methoda RSD (%) Spectroscopic methoda (mM) (mM)
Beer sample 1 600 Beer sample 2 540 White wine sample 1250 a
Mean of three determinations.
2.2 2.0 1.8
588 532 1240
RSD (%) 2.0 2.2 2.0
determined by back extrapolation to zero current (Supporting information, Fig. S14). Further, we did not observe any interference signals in the beer and wine samples from non-specific oxidation reactions at the electrode. Table 2 shows the results of sulfite determination in beer and wine sample using the present electrochemical biosensor. The obtained results are compared with the standard spectroscopic method (Supporting information, Fig. S15) where sulfite reacts with Ellman's reagent (5,50 -dithio-bis (2-nitrobenzoic acid) to produce an organic thiosulfate and releasing coloured 5-mercapto-2-nitrobenzoate. The sulfite concentration determined using the present electrochemical biosensor is in excellent agreement with the spectroscopic method. The obtained results clearly revealed that the present electrochemical biosensor is suitable for practical applications. 4. Conclusions We have demonstrated the mediated catalytic voltammetry of HSO with two synthetic electron acceptors [Fe(tacn)2]3+ and [Fe (dtne)]3+. The redox potential difference between these two mediators results in different oxidative driving forces for enzyme catalysis. A set of self-consistent rate constants was obtained by simulating the experimental CVs measured at different sweep rates, mediator concentrations and substrate concentrations. An amperometric biosensor was constructed with the lower potential mediator [Fe(tacn)2]3+and it showed linear catalytic response from 5 mM to 800 mM sulfite and lowest detection limit of 0.2 pM (S/ N = 3). As a practical application of the HSO modified electrode, we successfully used it for the amperometric determination of sulfite concentration in beers and wine samples and the results agreed well with values obtained by a standard spectroscopic method. Acknowledgements PVB acknowledges financial support from the Australian Research Council (DP150103345). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. electacta.2016.01.181. References [1] R. Hille, J. Hall, P. Basu, Chem. Rev. 114 (2014) 3963–4038.
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289 [2] K.-F. Aguey-Zinsou, P.V. Bernhardt, U. Kappler, A.G. McEwan, J. Am. Chem. Soc. 125 (2003) 530–535. [3] C. Feng, G. Tollin, J.H. Enemark, Biochim. Biophys. Acta 1774 (2007) 527–539. [4] C. Kisker, H. Schindelin, A. Pacheco, W.A. Wehbi, R.M. Garrett, K.V. Rajagopalan, J.H. Enemark, D.C. Rees, Cell 91 (1997) 973–983. [5] T. Utesch, M. Sezer, I.M. Weidinger, M.A. Mroginski, Langmuir 28 (2012) 5761–5769. [6] K. Johnson-Winters, A.R. Nordstrom, S. Emesh, A.V. Astashkin, A. Rajapakshe, R.E. Berry, G. Tollin, J.H. Enemark, Biochemistry 49 (2010) 1290–1296. [7] S. Emesh, T.D. Rapson, A. Rajapakshe, U. Kappler, P.V. Bernhardt, G. Tollin, J.H. Enemark, Biochemistry 48 (2009) 2156–2163. [8] M.Z. Seidahmed, E.A. Alyamani, M.S. Rashed, A.A. Saadallah, O.B. Abdelbasit, M. M. Shaheed, A. Rasheed, F.A. Hamid, M.A. Sabry, Am. J. Med. Genet. A 136 (2005) 205–209. [9] C. Hoffmann, B. Ben-Zeev, Y. Anikster, A. Nissenkorn, N. Brand, J. Kuint, T. Kushnir, J. Child Neurol. 22 (2007) 1214–1221. [10] A. Rajapakshe, R.A. Snyder, A.V. Astashkin, P. Bernardson, D.J. Evans, C.G. Young, D.H. Evans, J.H. Enemark, Inorg. Chim. Acta 362 (2009) 4603–4608. [11] M.S. Brody, R. Hille, Biochemistry 38 (1999) 6668–6677. [12] R. Hille, Biochim. Biophys. Acta Bioenerg. 1184 (1994) 143–169. [13] S.J. Elliott, A.E. McElhaney, C. Feng, J.H. Enemark, F.A. Armstrong, J. Am. Chem. Soc. 124 (2002) 11612–11613. [14] E.E. Ferapontova, L. Gorton, Bioelectrochemistry 66 (2005) 55–63. [15] T.D. Rapson, U. Kappler, P.V. Bernhardt, Biochim. Biophys. Acta Bioenerg. 1777 (2008) 1319–1325. [16] T.D. Rapson, U. Kappler, G.R. Hanson, P.V. Bernhardt, Biochim. Biophys. Acta Bioenerg. 1807 (2011) 108–118. [17] M. Sezer, R. Spricigo, T. Utesch, D. Millo, S. Leimkuehler, M.A. Mroginski, U. Wollenberger, P. Hildebrandt, I.M. Weidinger, Phys. Chem. Chem. Phys. 12 (2010) 7894–7903. [18] L.A. Coury Jr., B.N. Oliver, J.O. Egekeze, C.S. Sosnoff, J.C. Brumfield, R.P. Buck, R. W. Murray, Anal. Chem. 62 (1990) 452–458. [19] J.T. Spence, C.A. Kipke, J.H. Enemark, R.A. Sunde, Inorg. Chem. 30 (1991) 3011–3015. [20] L.A. Coury, Jr., L. Yang, R.W. Murray, Anal. Chem. 65 (1993) 242–246. [21] E.E. Ferapontova, T. Ruzgas, L. Gorton, Anal. Chem. 75 (2003) 4841–4850. [22] R. Dronov, D.G. Kurth, H. Moehwald, R. Spricigo, S. Leimkuehler, U. Wollenberger, K.V. Rajagopalan, F.W. Scheller, F. Lisdat, J. Am. Chem. Soc. 130 (2008) 1122–1123. [23] R. Spricigo, R. Dronov, K.V. Rajagopalan, F. Lisdat, S. Leimkuehler, F.W. Scheller, U. Wollenberger, Soft Matter 4 (2008) 972–978. [24] R. Spricigo, R. Dronov, F. Lisdat, S. Leimkuehler, F.W. Scheller, U. Wollenberger, Anal. Bioanal. Chem. 393 (2009) 225–233. [25] P. Kalimuthu, J. Tkac, U. Kappler, J.J. Davis, P.V. Bernhardt, Anal. Chem. 82 (2010) 7374–7379. [26] P. Kalimuthu, U. Kappler, P.V. Bernhardt, J. Phys. Chem. B 118 (2014) 7091–7099. [27] C. Ruiz-Capillas, F. Jimenez-Colmenero, Food Chem. 112 (2008) 487–493.
289
[28] S. Vandevijvere, E. Temme, M. Andjelkovic, W.M. De, C. Vinkx, L. Goeyens, L.J. Van, Food. Addit. Contam. Part A Chem. Anal. Control Expo. Risk Assess. 27 (2010) 1072–1083. [29] A.A. Belaidi, G. Schwarz, Biochem. J 450 (2013) 149–157. [30] K. Wieghardt, W. Schmidt, W. Herrmann, H.J. Kueppers, Inorg. Chem. 22 (1983) 2953–2956. [31] K. Wieghardt, I. Tolksdorf, W. Herrmann, Inorg. Chem. 24 (1985) 1230–1235. [32] G.D. Broadhead, J.M. Osgerby, P.L. Pauson, J. Chem. Soc. (1958) 650–656. [33] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications (2001). [34] N. Anicet, C. Bourdillon, J. Moiroux, J.-M. Saveant, J. Phys. Chem. B 102 (1998) 9844–9849. [35] M. Situmorang, D.B. Hibbert, J.J. Gooding, D. Barnett, Analyst 124 (1999) 1775–1779. [36] M. Rudolf, S.W. Feldberg, DigiSim version 3.03b, Bioanalytical System, West Lafayette (2004). [37] R.E. Humphrey, M.H. Ward, W. Hinze, Anal. Chem. 42 (1970) 698–702. [38] K. Johnson-Winters, A.C. Davis, A.R. Arnold, R.E. Berry, G. Tollin, J.H. Enemark, J. Biol. Inorg. Chem. 18 (2013) 645–653. [39] S. Frasca, O. Rojas, J. Salewski, B. Neumann, K. Stiba, I.M. Weidinger, B. Tiersch, S. Leimkuehler, J. Koetz, U. Wollenberger, Bioelectrochemistry 87 (2012) 33–41. [40] S. Frasca, A. Molero Milan, A. Guiet, C. Goebel, F. Pérez-Caballero, K. Stiba, S. Leimkühler, A. Fischer, U. Wollenberger, Electrochim. Acta 110 (2013) 172–180. [41] A. Rajapakshe, K.T. Meyers, R.E. Berry, G. Tollin, J.H. Enemark, J. Biol. Inorg. Chem. 17 (2012) 345–352. [42] P. Kalimuthu, S. Leimkühler, P.V. Bernhardt, J. Phys. Chem. B 115 (2011) 2655–2662. [43] R. Spricigo, C. Richter, S. Leimkuehler, L. Gorton, F.W. Scheller, U. Wollenberger, Colloids Surf. A 354 (2010) 314–319. [44] C.A. Temple, T.N. Graf, K.V. Rajagopalan, Arch. Biochem. Biophys. 383 (2000) 281–287. [45] H.L. Wilson, K.V. Rajagopalan, J. Biol. Chem. 279 (2004) 15105–15113. [46] K. Pohl, K. Wieghardt, W. Kaim, S. Steenken, Inorg. Chem. 27 (1988) 440–447. [47] K.-I. Chen, A.G. McEwan, P.V. Bernhardt, J. Biol. Inorg. Chem. 14 (2009) 409–419. [48] P. Kalimuthu, S. Leimkuhler, P.V. Bernhardt, J. Phys. Chem. B 116 (2012) 11600–11607. [49] K.-I. Chen, A.G. McEwan, P.V. Bernhardt, J. Biol. Inorg. Chem. 16 (2011) 227–234. [50] P. Kalimuthu, J. Heider, D. Knack, P.V. Bernhardt, J. Phys. Chem. B 119 (2015) 3456–3463. [51] C.C. Moser, J.M. Keske, K. Warncke, R.S. Farid, P.L. Dutton, Nature 355 (1992) 796–802. [52] G. Gastaminza, S. Quirce, M. Torres, A. Tabar, S. Echechipia, D. Munoz, d.C.L. Fernandez, Clin. Exp. Allergy 25 (1995) 698–703. [53] W.H. Yang, E.C. Purchase, CMAJ 133 (1985) 865–867 880. [54] H.R. Zare, F. Memarzadeh, M.M. Ardakani, M. Namazian, S.M. Golabi, Electrochim. Acta 50 (2005) 3495–3502.
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Low Potential Catalytic Voltammetry of Human Sulfite Oxidase Palraj Kalimuthua , Abdel A. Belaidib , Guenter Schwarzb , Paul V. Bernhardta,* a b
School of Chemistry and Molecular Biosciences, University of Queensland, Brisbane 4072, Australia Institute of Biochemistry, Department of Chemistry & Center for Molecular Medicine, Cologne University, Zülicher Str. 47, 50674 Köln, Germany
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 December 2015 Received in revised form 21 January 2016 Accepted 24 January 2016 Available online 27 January 2016
Mediated electrocatalytic voltammetry of human sulfite oxidase (HSO) is demonstrated with synthetic one electron transfer iron complexes bis(1,4,7-triazacyclononane)iron(III) ([Fe(tacn)2]3+) and 1,2-bis (1,4,7-triaza-1-cyclononyl)ethane iron(III) ([Fe(dtne)]3+) at a glassy carbon working electrode. The two synthetic electron acceptors for HSO, differing in redox potential by 270 mV, deliver different driving forces for electrocatalysis. Digital simulation of the catalytic voltammetry was achieved with single set of enzyme-dependent kinetic parameters that reproduced the experimental data across a range of sweep rates, and sulfite and mediator concentrations. Amperometry carried out in a stirred solution with the lower potential mediator [Fe(tacn)2]3+ was optimised and exhibited a linear increase in steady state current in the sulfite concentration range 5.0 106 to 8.0 104 M with a detection limit of 0.2 pM (S/ N = 3). The HSO coupled electrode was successfully used for the determination of sulfite concentration in white wine and beer samples and the results validated with a standard spectrophotometric method. ã 2016 Elsevier Ltd. All rights reserved.
Keywords: enzyme molybdenum sulfite
1. Introduction The molybdenum-dependent sulfite oxidizing enzymes comprise sulfite oxidase (SO) and sulfite dehydrogenase (SDH) [1,2]. SO is found in animals and plants whereas SDH is only found in bacteria [3]. Only the plant SO is a true oxidase while all other sulfite oxidizing enzymes donate electrons to cytochrome c. Vertebrate SOs can use either cytochrome c or dioxygen as an electron acceptor. Only one crystal structure is available for a vertebrate SO (from chicken liver) [4] revealing a 103 kDa homodimer in which each subunit contains a negatively charged small heme b domain at the N-terminus and positively charged larger molybdopterin domain at the C-terminus. The heme accepts electrons from the Mo ion following sulfite oxidation. A flexible connects the Mo and heme domains which are more than 30 Å apart in the crystal structure conformation; a distance too great for electron transfer. Spectroscopic and kinetic studies have demonstrated that the heme b domain swings around to be in proximity to the molybdenum active site where electron transfer (Mo to heme) can take place after sulfite oxidation [5–7]. SO
SO3 2 þ H2 O þ 2ðcytcÞox ! SO4 2 þ 2ðcytcÞred þ 2Hþ
* Corresponding author. E-mail address: [email protected] (P.V. Bernhardt). http://dx.doi.org/10.1016/j.electacta.2016.01.181 0013-4686/ ã 2016 Elsevier Ltd. All rights reserved.
ð1Þ
Human sulfite oxidase (HSO) shares a 68% sequence identity with chicken SO [4]. Among the eukaryotic SOs, HSO has been studied extensively because of its role in the potentially fatal disease SO deficiency [8,9]. The physiological role of SO is to remove toxic sulfite (a product of organo-sulfur compound metabolism) and covert it to chemically inert sulfate. Despite its name the physiological electron acceptor of SO is in fact cytochrome c (Eq. (1)). In the catalytic reaction, SO is active in its fully oxidized state (MoVI) in which molybdenum is coordinated by a cysteine thiolate, the dithiolene group of molybdopterin, and two terminal oxygen atoms as shown in Scheme 1 [7,10–12]. Upon reaction with sulfite, one oxido ligand is transferred to sulfite to give sulfate and the Mo ion is reduced its tetravalent state. Subsequently, hydroxide displaces sulfate, and the removal of this hydroxido ligand proton occurs spontaneously when the Mo ion is reoxidised to its hexavalent state by two cytochrome c molecules. There have been a number of electrochemical investigations of SO and SDH enzymes from different organisms. In these cases the electrode is the ultimate electron acceptor resulting in an anodic catalytic current. Electrons may be transferred directly from the enzyme [2,13–17] or via a mediator which may be synthetic [18– 20] or natural (cytochrome c) [18,21–25]. The dynamics HSO are potentially problematic for efficient electrocatalysis. While the Mo and heme cofactors are separated, the enzyme is unable to be reactivated through reoxidation. It is of interest whether confinement of HSO enzyme to a thin layer at the electrode surface suppresses this motion. Spectroelectrochemistry of HSO showed the FeIII/II redox potential to be +62 mV vs NHE (pH
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
281
Scheme 1. Simplified catalytic cycle of SO reduced forms of enzyme and substrate in red and oridised forms of enzyme and product in blue.
7.5) [6]. At applied electrochemical potentials above this value, the enzyme will be continually reoxidsed and reactivated for sulfite oxidation. To achieve this we employed two artificial electron acceptors; the hexa-amine complexes [Fe(tacn)2]3+ and [Fe(dtne)]3+ (Fig. 1) with redox potentials of +144 and +415 mV vs NHE, respectively which present significantly different overpotentials but are structurally almost the same. The higher FeIII/II redox potential of [Fe(dtne)]3+ is due to the presence of two tertiary amines compared to the all-secondary amine [Fe(tacn)2]3+. It is notable that nonspecific oxidation of sulfite at an electrode (without any enzyme present) is inevitable above ca. +550 mV vs NHE [26] and this places an upper bound on the redox potential of any mediator in a sulfite oxidizing electrochemical system. An additional feature of this study is electrochemical simulation of the experimental voltammetry. Given that the catalytic cycle involves several steps, some chemical reactions between HSO and sulfite/sulfate and some being outer sphere electron transfer reactions between HSO and the mediators, a set of rate constants can be defined (Scheme 2). These rate constants must be able to
reproduce the catalytic voltammetry under a variety of conditions including sweep rate, mediator concentration and sulfite concentration. Finally, amperometry is employed to estimate the lowest detection limit and linear current response for the determination of sulfite in aqueous solution and in the quantification of sulfite in beer and wine samples where it is a commonly found as an additive to combat spoilage from oxidation and microbial activity [27,28]. 2. Experimental 2.1. Materials Human sulfite oxidase (HSO) was purified in E. coli TP1000 as previously described [29]. The iron complexes bis(1,4,7-triazacyclononane) iron(III) bromide ([Fe(tacn)2]Br3) [30] and 1,2-bis (1,4,7-triaza-1-cyclononyl) ethane iron(III) bromide ([Fe(dtne)] Br3.3H2O) [31] were synthesized according to the previous procedures. Sodium sulfite and 5,50 -dithio-bis(2-nitrobenzoic acid) (Ellman's reagent) were purchased from Aldrich and were
3+ HN
NH
HN
3+ HN HN
Fe HN
N Fe
NH N H
[Fe(tacn) 2] 3+ E' (Fe III/II) +144 mV vs NHE
HN
N N H
[Fe(dtne) ]3+ E' (Fe III/II) +415 mV vs NHE
Fig. 1. Molecular structures and redox potentials of the mediators used in this study.
282
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
used as received. The beer and wine samples were purchased from local retail outlets. All other reagents used were of analytical grade purity and used without any further purification. Tris acetate buffer (50 mM) was used for all experiments at pH 8.0. For pHdependent experiments, the mixture of buffers (20 mM MES buffer pH 5.5–6.7, 20 mM Bis-Tris buffer pH 5.8–7.2, 20 mM Tris buffer pH 7.0–9.0, 20 mM CHES buffer pH 8.6–10.0 and 20 mM CAPS buffer pH 9.7–11.1) was used and the desired pH was obtained with dilute acetic acid or NaOH. All solutions were prepared with ultrapure water (resistivity 18.2 MV.cm) from a Millipore Milli-Q system. 2.2. Electrochemical Measurements and Electrode Cleaning The cyclic voltammetry (CV) and chronoamperometry experiments were carried out with a BAS 100B/W electrochemical workstation. A three-electrode system was employed comprising a glassy carbon (GC) disk working electrode, a Pt wire counter, and an Ag/AgCl reference electrode (+196 mV vs NHE). Potentials are cited versus NHE. Experiments were carried on solutions that had been purged with argon gas for 30 min. The GC electrode was polished with 0.50 and 0.05 mm alumina slurry and then rinsed thoroughly with water. Then the electrode was sonicated in water for 10 min to remove adsorbed alumina particles and dried in a nitrogen atmosphere. The electro-active surface area of the GC electrode (A) was determined from the cyclic voltammetry of 1 mM ferrocene methanol [32] in 0.1 M KCl solution at different sweep rates using the Randles-Sevcik equation (Eq. (1)) [33]. ip = (2.69 105)n3/2ADo1/2Con1/2
(1)
The standard diffusion coefficient (Do) of ferrocene methanol is 6.7 106 cm2 s1 [34],ip is the measured current maximum, n is the number of electrons, Co is concentration of analyte (mol cm3), and n is the sweep rate (V s1). The variation of the catalytic current (ilim) as a function of sulfite concentration was fit to Michaelis–Menten kinetics (Eq. (2)) yielding KM,app (the apparent Michaelis constant) and imax (the effective electrochemical turnover number, imax = nFA[HSO]) [35]. h i imax SO2 3 h i ð2Þ ilim ¼ K M;app þ SO2 3 The pH dependence of the catalytic current was modeled by Eq. (3) [11] which is applicable for an active form of the enzyme that is deactivated by either deprotonation of an acid at high pH (pKa1) or protonation of a base at lower pH (pKa2). imax ðpHÞ ¼
iopt 1 þ 10ðpHpK a1 Þ þ 10ðpK a2 pHÞ
ð3Þ
2.3. Enzyme Electrode Preparation A 3 mL droplet of HSO (66 mM) in 50 mM Tris buffer (pH 8.0) was pipetted onto the conducting surface of an inverted, freshly prepared GC working electrode and this was allowed to dry to a film at 4 C. To prevent protein loss the electrode surface was carefully covered with a semi-permeable dialysis membrane (SERVA MEMBRA-CEL, molecular weight cut off 3500 Da), presoaked in water. The dialysis membrane was pressed onto the electrode with a Teflon cap and fastened to the electrode with a rubber O-ring to prevent leakage of the internal membrane solution. The resulting modified electrode was stored at 4 C in 50 mM Tris buffer (pH 8.0) when not in use. The enzyme was confined to a thin layer beneath the membrane while substrate and mediators were able to diffuse across the dialysis membrane.
2.4. Electrochemical Simulation The DigiSim program (version 3.03b) was employed to simulate the experimental cyclic voltammograms [36]. The experimental parameters restrained in each case were the working electrode surface area (0.055 cm2) and the double-layer capacitance (12 mF). Semi-infinite diffusion was assumed and all pre-equilibration reactions were enabled. The apparent redox potential of mediators was determined from control experiments in the absence of enzyme or substrate. The diffusion coefficients of mediators were also obtained in the presence of a dialysis membrane covering the electrode by simulation of the cyclic voltammetry at different sweep rates in the absence of substrate and enzyme to give value of 5 106 cm2 s1. The diffusion coefficients for HSO and substrate were taken to be 5 107 and 5 106 cm2 s1 [26]. These values were kept constant for simulating the various substrate- and mediator-concentration-dependent CVs. The heterogeneous rate constant (k0) was determined from simulating the sweep rate dependence of the anodic peak to cathodic peak separation of mediators (in the absence of HSO) and then held constant thereafter. The only values that were allowed to differ were the rate constants for the outer sphere electron transfer reaction between each mediator and enzyme (k4, k40 , k-4 and k-40 in Scheme 2). It was assumed that k4 = k40 and k–4 = k-40 i.e. oxidation of either the MoV/FeII or MoVI/FeII forms of HSO proceeded at the same rate, which is reasonable given that the heme cofactor is the site of oxidation and its redox potential is known (+62 mV vs NHE) [6]. 2.5. Spectrophotometric Sulfite Determination As a complement to amperometric sulfite determination the results were validated using Ellman's reagent which is cleaved by sulfite to form an organic thiosulfate and 5-mercapto-2-nitrobenzoate stoichiometrically; the latter being determined spectrophotometrically [37]. 3. Results and Discussion 3.1. Electrocatalytic Mechanism of HSO The electrocatalytic mechanism of HSO is illustrated in Scheme 2. The single electron transfer acceptors [Fe(tacn)2]3+ and [Fe(dtne)]3+ used in the present study are synthetic substitutes for cytochrome c and so two consecutive one electron transfer reactions are necessary to regenerate the reduced HSO to its active form after it has been reduced by sulfite. There are also two intramolecular electron transfer (IET) processes. The first IET step occurs when MoIV transfers one electron to the oxidized ferric heme b cofactor. The ensuing MoV/ FeII species transfers an electron to the artificial electron acceptor producing the MoV/FeIII state (rate constant k4). A second IET step generates the MoVI/FeII state, and reduction of a second molecule of mediator (rate constant k40 ) regenerates the fully oxidized MoVI/ FeIII state of the enzyme. The sum of the forward and reverse IET2 steps is known (ket > 400 s1 at pH 7) [38] and this first order reaction is always much faster than the rates of the (second order) outer sphere redox reactions (k4, k40 ) which are slowed down by the low concentrations of mediators and HSO used in this experiment. For this reason we have not included either IET step in our kinetic model i.e. it is assumed to be fast and never rate limiting. The substrate (sulfite) and mediator ([Fe(tacn)2]3+ or ([Fe (dtne)]3+) are under diffusion control while HSO is confined to the small volume under the membrane but still may diffuse within that space. We have assumed that the catalytic reaction follows
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
283
Scheme 2. Mediated Electrochemically Driven Catalysis of HSO.
Michaelis-Menten kinetics comprising substrate binding (k1/k-1), turnover (k2/k-2) and product release (k3/k-3). A simplified double substrate ‘ping-pong’ mechanism is appropriate for this type of catalysis. 3.2. Mediator Voltammetry Interestingly upon introduction of 10 mM of the mediator [Fe (tacn)2]3+ into the electrochemical cell no significant redox response was observed initially (in the first cycle) at the dialysis membrane-covered HSO modified GC electrode (Supporting Information, Fig. S1A). In the second cycle the redox response of [Fe(tacn)2]3+ emerged and increased in current up to about 12 cycles where a consistent waveform was established. The GC/HSO electrode shows a well-defined redox wave centred at ca. +56 mV vs NHE for 10 mM [Fe(tacn)2]3+ with a peak to peak separation of only 42 mV in 50 mM Tris buffer solution (Fig. 2, curve a). There are several important features to note. The increasing current at the GC/HSO electrode as a function of cycle number (Supporting Information, Fig. S1A) indicates that the [Fe (tacn)2]3+ molecules only cross the membrane slowly from the bulk solution i.e. flux across the membrane during the sweep is insufficient to keep pace with depletion of the mediator from the diffusion layer in the initial sweep. Moreover, the establishment of a consistent CV after several cycles indicates that the amount of [Fe (tacn)2]3+ that eventually accumulates under the membrane is sufficient to sustain catalysis i.e. it is not depleted during the sweep (see below). Secondly the peak to peak separation is less than 57 mV but greater than 0 mV, which is intermediate of a response governed by linear diffusion and that seen in a thin layer cell [33]. Furthermore the symmetry of the wave is a hybrid of the tailing waveform characteristic of normal linear diffusion and the symmetrical wave characteristic of a thin layer cell due to the confines of the membrane [33]. It is also apparent that the observed currents are much greater than would be expected for a
10 mM solution on the basis of Eq. (1). It appears that a significant amount of [Fe(tacn)2]3+ is concentrated on the inner side of the dialysis membrane and escape to the bulk solution (10 mM) is slow. If the electrode is transferred to a buffer solution containing no Fe complex then the current response is significantly diminished upon continuous cycling (Supporting information Fig. S1B). On the basis of the maximum peak height attained and Eq. (1) we have estimated the amount maximum concentration of [Fe(tacn)2]3+ under the dialysis membrane to be ca. 120 mM. The redox response was investigated at different scan rates. Both oxidation and reduction currents increase with scan rate from 10 to 100 mV s1 (Supporting information Fig. S2A) but the linear increase of peak height with the square root of sweep rate (R2 = 0.999, with zero
Fig. 2. CVs obtained for 10 mM [Fe(tacn)2]3+ in the absence (a) and presence (b) of 5 mM sulfite at the GC/HSO electrode in 50 mM Tris buffer (pH 8) at a sweep rate of 5 mV s1.
284
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
intercept) is more consistent with a linear diffusion controlled process than a thin layer response (linear increase of current with sweep rate). On balance the mediator redox response is predominantly under diffusion control (Supporting information Fig. S2B). Also of interest is the deviation of the [Fe(tacn)2]3+/2+ redox potential under these conditions (E' +56 mV vs NHE) from that obtained in solution without a membrane and in the absence of HSO (+144 mV). This is not a consequence of the membrane as the CVs of [Fe(tacn)2]3+ alone are the same in the absence or presence of the membrane (Figs. S1C and S1D). The pronounced cathodic shift in the [Fe(tacn)2]3+/2+ redox potential is only seen in the presence of HSO and this is attributed to the formation of a noncovalent (outer sphere) complex with HSO under the membrane. The natural electron acceptor of HSO is the highly positively charged protein cytochrome c which is thought to bind (noncovalently) at a negatively charged surface of HSO adjacent to the heme cofactor. The affinity of the tri-positively charged [Fe(tacn)2]3+ evidently mimics cytochrome c and presumably interacts with the same highly negatively charged surface of HSO and this interaction lowers the redox potential of the mediator. The same analysis was carried out with [Fe(dtne)]3+ and once again the CV behaviour is consistent with a predominantly diffusion controlled response this time at a much higher redox potential (Supporting Information Fig. S3). Again the observed [Fe (dtne)]3+/2+ redox potential (+355 mV) was shifted cathodically from its value in solution (+410 mV) [31]. 3.3. Catalytic Voltammetry Although HSO has two electroactive centers (Mo and heme) no redox responses were observed from either cofactor in the absence (or presence) of sulfite at the GC electrode without mediator present (data not shown). This is not unexpected as direct electrochemistry of HSO has only been observed at chemically modified Au [39], Ag [17] or Sb-doped SnO2 [40] electrodes and only quite weak responses were seen. In the presence of HSO (under the membrane), [Fe(tacn)2]3+ and sulfite (5 mM), a well-defined classic sigmoidal waveform is seen and the limiting anodic current increases by an order of magnitude (Fig. 2, curve b). No cathodic wave is present in this case and obviously the forward and backward sweeps are same when the charging current is taken into account. In a control experiment, we found that the redox response of [Fe(tacn)2]3+ is insensitive to 5 mM sulfite at a bare GC electrode (data not shown) within the
potential window of 200 to +350 mV vs NHE indicating that mediator alone cannot oxidize sulfite. Thus, the sigmoidal voltammetric waveform at the enzyme modified electrode is characteristic of a catalytic homogeneous reaction coupled to heterogeneous electron transfer (EC' mechanism) [33] where sulfite is oxidized enzymatically yielding the reduced form of enzyme (MoIV), which is reoxidized by electro-generated [Fe (tacn)2]3+. 3.4. HSO-Sulfite Reaction The reaction between HSO and sulfite was investigated by varying the sulfite concentration while maintaining a constant concentration of mediator and enzyme. The examples in Fig. 3 show the CVs of the GC/HSO electrode in the presence of 10 mM [Fe (tacn)2]3+ (Fig. 3A) and 20 mM [Fe(dtne)]3+ (Fig. 3B) at a sweep rate of 5 mV s1 in Tris buffer (pH 8). In both cases, the CVs take the form of an asymmetric transient catalytic wave up to 800 mM sulfite with a pronounced anodic peak but no corresponding cathodic current. The peak is due to mass transport limitations where sulfite becomes depleted at the electrode surface due to the rate it is consumed by HSO, which cannot be sustained by diffusion from the bulk solution across the membrane. It is apparent that as the sulfite concentration is increased further, (3.2 mM) the wave increases in magnitude and the sharp transient form of the wave becomes more symmetrical. Ultimately (> 4 mM sulfite), the transient wave becomes sigmoidal where the concentration of sulfite within the reaction layer is constant during the sweep. The sigmoidal waveform is indicative of an electrochemical steady state i.e. the oxidized form of mediator is consumed (by homogeneous reaction with HSOred) at the same rate that it is generated at the electrode surface and mass transport of sulfite from the bulk is fast enough to ensure its concentration is constant under the membrane. CVs at all concentrations of sulfite examined appear in the Supporting Information with [Fe(tacn)2]3+ (Fig. S4A) and [Fe(dtne)]3+ (Fig. S5A) as mediator. The catalytic sulfite oxidation current increased linearly up to 800 and 1600 mM sulfite before saturating at millimolar concentrations. Apparent Michaels constants (KM,sulfite) of 512 mM ([Fe (tacn)2]3+) and 970 mM ([Fe(dtne)]3+) were obtained (Supporting Information, Fig. S6). Of course KM,sulfite should be mediatorindependent so these are not true Michaelis constants and they have little mechanistic relevance other than defining the approximate linear current response of the electrode. The true KM,sulfite value for HSO in solution is 9 mM in reaction with its physiological
Fig. 3. CVs obtained for varying sulfite concentrations in the presence of (A) 10 mM [Fe(tacn)2]3+ and (B) 20 mM [Fe(dtne)]3+ at GC/HSO electrode in 50 mM Tris buffer (pH 8) at a sweep rate of 5 mV s1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
285
Fig. 4. CVs obtained for varying (A) [Fe(tacn)2]3+ and (B) [Fe(dtne)]3+ concentrations in the presence of 4 mM sulfite at GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at a sweep rate of 5 mV s1.
electron acceptor cytochrome c [41]. So utilizing the mass transport limitations presented by the membrane, the linear response of the electrode to sulfite is increased by at least 2 orders of magnitude. We have observed similar observations in other Mo enzyme systems [25,26,42]. 3.5. HSO-Mediator Reaction The HSO-mediator reaction was examined with increasing [Fe (tacn)2]3+ and [Fe(dtne)]3+ concentrations in the presence of a high (constant) concentration of sulfite. Fig. 4A displays the effect of varying the concentration of [Fe(tacn)2]3+ in the presence of 4 mM sulfite at a sweep rate of 5 mV s1. At 2 mM [Fe(tacn)2]3+, a sigmoidal voltammogram is found, which is indicative of an electrochemical steady state; the forward and backward sweeps are the same and the catalytic current is switched on and off in a Nernstian fashion. However, as the concentration of [Fe(tacn)2]3+ increases (to 4 and 6 mM), the waveform becomes asymmetric due to an excess of the oxidized form of mediator being produced at the electrode, which overwhelms the limiting amount of HSOred formed during the sulfite oxidation step. In other words the concentration of [Fe(tacn)2]3+ is no longer at steady state. Fig. 4B displays similar experiments but this time with increasing concentrations of [Fe(dtne)]3+ (6, 12 and 18 mM). The observed sigmoidal wave at 6 mM turns to a transient form upon increasing concentration to 12 and 18 mM [Fe(dtne)]3+. Data collected at all mediator concentrations are shown in Supporting Information Figs. S4B and S5B.
catalysis involving substrate binding [45]. The higher pKa value observed here may be due to Tyr343 deprotonation at high pH which is believed to be close to the active site and involved in Hbonding with the substrate [45]. The pH profile was independent of the direction of titration and catalytic activity was fully restored when the solution pH was returned to its optimal value. Furthermore the voltammetry of both [Fe(tacn)2]3+ and [Fe (dtne)]3+ are pH-independent within this range. The complex [Fe(tacn)2]3+ can be deprotonated but only at much higher pH (pKa 11.7) [46]. 3.7. Electrochemical Simulation In recent years, we have employed digital simulation for a better understanding of the mechanism of mediated enzyme electrochemical reactions [26,47–50]. The objective of the simulation is to obtain the rate constants defined in Scheme 2 that reproduce all voltammetric features over a range of sweep rates, substrate and mediator concentrations. The voltammetric sweep rate is a significant variable to elucidate the kinetics of electrochemical processes coupled with chemical reactions. The DigiSim program enables the same set of kinetic parameters to be optimized to CVs measured across a range of sweep rates, but under an identical set of concentrations (HSO, mediator and sulfite). When the concentrations of mediators and
3.6. pH Dependence The pH dependence of the catalytic sulfite oxidation current at the GC/HSO electrode was explored in the range 5.5 < pH < 11 in 100 mM mixed buffer solution. Fig. 5 depicts the baseline subtracted maximum catalytic current as a function of pH. The actual CVs are provided in the Supporting Information (Fig. S7). The catalytic current exhibits a pH optimum of 8.5 which is similar to that reported for HSO at an osmium redox polymer modified electrode [43] as well as in a solution assay for HSO with its natural electron acceptor cytochrome c [44]. A bell shaped profile obtained by the application of Eq. (3) enabled the two pKa values to be determined (7.2 and 9.8); the lower value defining the protonation constant of a base that switches off catalysis and the higher one being the protonation constant of a base that switches on catalysis. It has been proposed that Tyr343 plays an important role in HSO
Fig. 5. Plot of the pH dependence of the maximum catalytic oxidation current at the GC/HSO electrode with 4 mM sulfite and in the presence of 10 mM [Fe(tacn)]23+ in 100 mM mixed buffer solution at a scan rate of 5 mV s1. The solid curve is obtained from a fit to the experimental points using Eq. (3) (pKa1 9.8 and pKa2 7.2).
286
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
Fig. 6. Experimental (solid lines) and simulated (broken lines) sweep rate dependent CVs obtained for 800 mM sulfite in the presence of (A) 10 mM [Fe(tacn)2]3+ and (B) 20 mM [Fe(dtne)]3+ at GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at different scan rates.
sulfite are varied then ideally the same parameters reproduce CVs measured under those situations as well. Fig. 6 shows experimental and simulated CVs for 800 mM sulfite in the presence of 10 mM [Fe(tacn)2]3+ (Fig. 6A) and 20 mM [Fe(dtne)]3+ (Fig. 6B) as a function of sweep rate. All other sweep rate dependent simulated voltammograms recorded as a function of various mediator and substrate concentrations are given in the Supporting Information (Fig. S8-S13). The same features are well reproduced for both mediators. In Fig. 6A as the scan rate increases from 5 to 50 mV s1, the asymmetric transient CV becomes reversible as electrochemical oxidation and reduction of the mediator is too rapid and the HSO-mediator reaction becomes uncompetitive. A very similar trend observed for [Fe(dtne)]3+ as shown in Fig. 6B and these features are also well reproduced in the simulation. The substrate binding rate constant (k1) is well defined by simulation and changing its value has a major influence on the quality of the fit between experiment and theory. Although k2 is also an important value, and defines the maximum current at HSO saturation, its value is entangled with the concentration of HSO under the membrane (Eq. (2), imax = nFAk2[HSO]). The concentration of HSO under the membrane is only known approximately because the volume under the membrane cannot be measured directly but instead determined by introduction of an known amount of external standard e.g. cytochrome c as reported
previously [25]. In this case k2 is the same as determined for HSO at pH 8.0 [45]. The product dissociation rate k3 value has little influence on the CV in this case if allowed to deviate from its optimal value (values in the range 0.5 to 50 s1 gave the same result here. The k4 values are also accurately determined although the same issues regarding the accurate concentration of HSO under the membrane introduce some uncertainty. The CVs as a function of increasing mediator concentration of [Fe(tacn)2]3+ (1 to 4 mM) and [Fe(dtne)]3+ (2 to 8 mM) in the presence of a saturating (4 mM) sulfite concentration and sweep rate of 5 mV s1 are represented in Fig. 7. An approximately sigmoidal wave is observed at a low concentration of [Fe(tacn)2]3+ (1 mM) and this wave becomes progressively peak-shaped (transient) as the higher concentration of mediator overwhelms the HSO present and the electrochemical steady state of mediator breaks down. The same set of parameters also reproduced CVs measured at various sulfite concentrations. Fig. 8 displays the anodic current response of the GC/HSO electrode as function of sulfite concentration in the presence of 10 mM of [Fe(tacn)2]3+ (Fig. 8A) and 20 mM of [Fe(dtne)]3+(Fig. 8B). At lower concentrations of sulfite (400 mM), the voltammograms took on a reversible transient form due to the excess amount of oxidized mediator at the electrode surface. The wave becomes firstly asymmetric as the
Fig. 7. Experimental (solid lines) and simulated (broken lines) CVs obtained for varying mediator concentration in the presence of 4 mM of sulfite (A) [Fe(tacn)2]3+ and (B) [Fe (dtne)]3+ at the GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at a sweep rate of 5 mV s1.
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
287
Fig. 8. Experimental (solid lines) and simulated (broken lines) CVs obtained for varying sulfite concentration in the presence of (A) 10 mM of [Fe(tacn)2]3+ and (B) 20 mM of [Fe (dtne)]3+ at the GC/HSO electrode in 50 mM Tris buffer solution (pH 8) at a sweep rate of 5 mV s1.
sulfite concentration rises to 1600 mM, where catalysis becomes significant, and tailing is due to sulfite depletion. Finally as the enzyme is saturated with sulfite (> 3 mM) the expected steady state sigmoidal waveform is observed. The catalytic current increased linearly with sulfite concentrations up to 800 and 1600 mM in the presence of [Fe(tacn)2]3+ and [Fe(dtne)]3+, respectively. This discrepancy is due to the different driving forces of the mediators (different k4 values) and the catalytic current attains saturation quickly when the oxidative driving force is small ([Fe(tacn)2]3+) but extends to a much higher value when the driving force is relatively high ([Fe(dtne)]3+). We have reported similar phenomena before in the mediated electrochemistry of the enzyme DMSO reductase [49]. Briefly the more rapid reaction between HSO and [Fe(dtne)2]3+ leads to faster sulfite depletion under the membrane and thus higher concentrations of sulfite are needed to reach the point where the enzyme becomes truly saturated. An additional point of interest is the low redox potential of [Fe(tacn)2]3+/2+ in these experiments (+56 mV vs NHE). This is similar to the heme b redox potential reported for HSO [6,41] but much lower than the natural redox partner cytochrome c (ca. +260 mV). Thus the [Fe(tacn)2]3+:HSO catalytic system operates as an extremely low catalytic potential (200 mV lower than the naturally mediated electrochemical reaction), which is advantageous for avoiding interference from the nonspecific oxidation of other species in solution.
3.8. Analysis of Kinetics Parameters The rate and equilibrium constants defined in Scheme 2 and presented in Table 1 reproduced all our experimental voltammetry carried out at different sweep rates, concentrations of sulfite and both mediators. Importantly there same set of mediator-independent parameters (k1-k3) were used for the two mediators. The accurate determination of multiple parameters is problematic in that some parameters will have no effect on the CV depending on the concentrations of the reactants. That is the rate limiting step in Scheme 2 will vary depending on the conditions. The sulfite binding rate constant determined here (k1 = 106 M1s1) has not been reported for HSO. This value may even be an underestimate due to mass transport limitations set by the membrane. The turnover number (k2 = 25 s1) obtained in the simulations is consistent with experimental value reported by solution assays for wild type HSO with its natural electron acceptor cytochrome c in pH 8 (27 s1) [41]. The larger outer sphere electron transfer rate constant for the HSO:[Fe(dtne)]3+ reaction compared with the HSO:[Fe(tacn)2]3+ reaction is consistent with Marcus theory (log ket / DG2) [51] given that higher redox potential of [Fe(dtne)]3+ delivers a greater driving force. The rate constant (k4) obtained for the reaction between the SO and higher potential mediator [Fe (dtne)]3+ is similar to the value reported for HSO in reaction with cytochrome c (4.0 106 M1 s1) [23]. 3.9. Amperometric Sulfite Determination
Table 1 Kinetic parameters (defined in Scheme 2) from electrochemical simulation. E0 (mV vs NHE)
k4 (M1 s1)a k-4 (M1 s1)b
k1 (M1 s1) k-1 (s1) k2 (s1) k-2 (s1) k3 (s1) k-3 (M1 s1) KM,Sulfite (mM)d a
[Fe(tacn)2]3+ 56 mV Mediator dependent 1.0 104 0.1c
[Fe(dtne)]3+ 355 mV 2.0 106 2c
Mediator independent 1.0 106 20 25 5.0 102 c 5c 1.0 102 c 102
k4 = k40 ; b k–4 = k-40 ; c approximate (simulation not sensitive to this parameter); KM,sulfite = (k2 + k1)/k1.
d
Sulfite is used as a preservative in food and beverages to prevent oxidation and bacterial growth and to control enzymatic reactions during production and storage [27,28]. Nevertheless, sulfite has been regulated since the realization that it may cause asthmatic attacks and allergic reactions in some people [52,53]. Typically, a warning label is required for any food or beverage containing more than 10 ppm (125 mM) sulfite so its accurate measurement in solution is important. Here we were able to achieve sulfite determination using the GC/HSO electrode and the low potential mediator [Fe(tacn)2]3+ in an amperometric experiment. Fig. 9A illustrates the amperometric i-t curve for the catalytic oxidation of sulfite at a GC/HSO electrode (covered with a dialysis membrane) in a homogeneously stirred 50 mM Tris buffer solution (pH 8) at an applied potential of +150 mV vs NHE. An initial baseline current response was stabilized for about 2 min in the presence of 10 mM [Fe(tacn)2]3+ at the GC/HSO electrode to ensure the mediator was able to concentrate under the membrane (vide supra). Upon addition of 10 mM of sulfite to the stirred solution in the
288
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289
Fig. 9. (A) An amperometric i t curve obtained for the determination of sulfite at the GC/HSO electrode in stirred 50 mM Tris buffer solution (pH 8). Each increment corresponded to a 10 mM increase in sulfite which was injected at regular intervals of 100 s. The electrochemical cell contained 10 mM [Fe(tacn)2]3+ and the electrode was poised at +150 mV vs NHE. (B) Plot of the steady state current as a function of sulfite in the linear range.
electrochemical cell, the catalytic anodic current increased suddenly and reached a plateau (steady state) within 3 sec. Further 10 mM sulfite increments at intervals of 100 s led to a regular and consistent step in the current. Furthermore, the amperometric current increased linearly with sulfite concentration from 10 to 180 mM and the detection limit [54] was found to be 0.2 pM (S/ N = 3) (Fig. 9B). The obtained detection limit is even lower than we reported with Starkeya novella sulfite dehydrogenase on a 11mercaptoundecanol monolayer modified Au electrode (44 pM) [25]. In separate voltammetry experiments, we found that the anodic catalytic current increased linearly from 5 to 800 mM at +150 mV vs NHE (Supporting Information, Fig. S6). Spricigo et al. reported a sulfite biosensor by co-immobilization of HSO within an osmium redox polymer on a carbon screen-printed electrode [43]. The biosensor operates at +100 mV vs NHE with detection limit (0.5 mM) and linearity (1 to 100 mM). 3.10. Determination of Sulfite in Wine and Beer Samples In order to demonstrate the practical application of the present biosensor, the enzyme modified electrode was used for the determination sulfite concentration in white wine and beer samples. Two beers and one white wine sample were obtained commercially were analyzed for sulfite using the present biosensor and this was validated by the standard spectroscopic method [37] using Ellman's reagent. Beer and wine, as prepared, are acidic (pH 4) and at this pH HSO is inactive as shown in Fig. 5. Therefore, the beer and wine samples were neutralized with dilute NaOH then diluted with Tris buffer solution and analyzed immediately without any other pretreatment. The method of standard additions was employed by injecting known amounts of sulfite to each beer or wine sample within the linear range and measuring the increase in catalytic current which enabled the original sulfite concentration to be Table 2 Determination of sulfite in wine and beer samples using present method and also compared with the standard spectroscopic method. Sample
Present methoda RSD (%) Spectroscopic methoda (mM) (mM)
Beer sample 1 600 Beer sample 2 540 White wine sample 1250 a
Mean of three determinations.
2.2 2.0 1.8
588 532 1240
RSD (%) 2.0 2.2 2.0
determined by back extrapolation to zero current (Supporting information, Fig. S14). Further, we did not observe any interference signals in the beer and wine samples from non-specific oxidation reactions at the electrode. Table 2 shows the results of sulfite determination in beer and wine sample using the present electrochemical biosensor. The obtained results are compared with the standard spectroscopic method (Supporting information, Fig. S15) where sulfite reacts with Ellman's reagent (5,50 -dithio-bis (2-nitrobenzoic acid) to produce an organic thiosulfate and releasing coloured 5-mercapto-2-nitrobenzoate. The sulfite concentration determined using the present electrochemical biosensor is in excellent agreement with the spectroscopic method. The obtained results clearly revealed that the present electrochemical biosensor is suitable for practical applications. 4. Conclusions We have demonstrated the mediated catalytic voltammetry of HSO with two synthetic electron acceptors [Fe(tacn)2]3+ and [Fe (dtne)]3+. The redox potential difference between these two mediators results in different oxidative driving forces for enzyme catalysis. A set of self-consistent rate constants was obtained by simulating the experimental CVs measured at different sweep rates, mediator concentrations and substrate concentrations. An amperometric biosensor was constructed with the lower potential mediator [Fe(tacn)2]3+and it showed linear catalytic response from 5 mM to 800 mM sulfite and lowest detection limit of 0.2 pM (S/ N = 3). As a practical application of the HSO modified electrode, we successfully used it for the amperometric determination of sulfite concentration in beers and wine samples and the results agreed well with values obtained by a standard spectroscopic method. Acknowledgements PVB acknowledges financial support from the Australian Research Council (DP150103345). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. electacta.2016.01.181. References [1] R. Hille, J. Hall, P. Basu, Chem. Rev. 114 (2014) 3963–4038.
P. Kalimuthu et al. / Electrochimica Acta 199 (2016) 280–289 [2] K.-F. Aguey-Zinsou, P.V. Bernhardt, U. Kappler, A.G. McEwan, J. Am. Chem. Soc. 125 (2003) 530–535. [3] C. Feng, G. Tollin, J.H. Enemark, Biochim. Biophys. Acta 1774 (2007) 527–539. [4] C. Kisker, H. Schindelin, A. Pacheco, W.A. Wehbi, R.M. Garrett, K.V. Rajagopalan, J.H. Enemark, D.C. Rees, Cell 91 (1997) 973–983. [5] T. Utesch, M. Sezer, I.M. Weidinger, M.A. Mroginski, Langmuir 28 (2012) 5761–5769. [6] K. Johnson-Winters, A.R. Nordstrom, S. Emesh, A.V. Astashkin, A. Rajapakshe, R.E. Berry, G. Tollin, J.H. Enemark, Biochemistry 49 (2010) 1290–1296. [7] S. Emesh, T.D. Rapson, A. Rajapakshe, U. Kappler, P.V. Bernhardt, G. Tollin, J.H. Enemark, Biochemistry 48 (2009) 2156–2163. [8] M.Z. Seidahmed, E.A. Alyamani, M.S. Rashed, A.A. Saadallah, O.B. Abdelbasit, M. M. Shaheed, A. Rasheed, F.A. Hamid, M.A. Sabry, Am. J. Med. Genet. A 136 (2005) 205–209. [9] C. Hoffmann, B. Ben-Zeev, Y. Anikster, A. Nissenkorn, N. Brand, J. Kuint, T. Kushnir, J. Child Neurol. 22 (2007) 1214–1221. [10] A. Rajapakshe, R.A. Snyder, A.V. Astashkin, P. Bernardson, D.J. Evans, C.G. Young, D.H. Evans, J.H. Enemark, Inorg. Chim. Acta 362 (2009) 4603–4608. [11] M.S. Brody, R. Hille, Biochemistry 38 (1999) 6668–6677. [12] R. Hille, Biochim. Biophys. Acta Bioenerg. 1184 (1994) 143–169. [13] S.J. Elliott, A.E. McElhaney, C. Feng, J.H. Enemark, F.A. Armstrong, J. Am. Chem. Soc. 124 (2002) 11612–11613. [14] E.E. Ferapontova, L. Gorton, Bioelectrochemistry 66 (2005) 55–63. [15] T.D. Rapson, U. Kappler, P.V. Bernhardt, Biochim. Biophys. Acta Bioenerg. 1777 (2008) 1319–1325. [16] T.D. Rapson, U. Kappler, G.R. Hanson, P.V. Bernhardt, Biochim. Biophys. Acta Bioenerg. 1807 (2011) 108–118. [17] M. Sezer, R. Spricigo, T. Utesch, D. Millo, S. Leimkuehler, M.A. Mroginski, U. Wollenberger, P. Hildebrandt, I.M. Weidinger, Phys. Chem. Chem. Phys. 12 (2010) 7894–7903. [18] L.A. Coury Jr., B.N. Oliver, J.O. Egekeze, C.S. Sosnoff, J.C. Brumfield, R.P. Buck, R. W. Murray, Anal. Chem. 62 (1990) 452–458. [19] J.T. Spence, C.A. Kipke, J.H. Enemark, R.A. Sunde, Inorg. Chem. 30 (1991) 3011–3015. [20] L.A. Coury, Jr., L. Yang, R.W. Murray, Anal. Chem. 65 (1993) 242–246. [21] E.E. Ferapontova, T. Ruzgas, L. Gorton, Anal. Chem. 75 (2003) 4841–4850. [22] R. Dronov, D.G. Kurth, H. Moehwald, R. Spricigo, S. Leimkuehler, U. Wollenberger, K.V. Rajagopalan, F.W. Scheller, F. Lisdat, J. Am. Chem. Soc. 130 (2008) 1122–1123. [23] R. Spricigo, R. Dronov, K.V. Rajagopalan, F. Lisdat, S. Leimkuehler, F.W. Scheller, U. Wollenberger, Soft Matter 4 (2008) 972–978. [24] R. Spricigo, R. Dronov, F. Lisdat, S. Leimkuehler, F.W. Scheller, U. Wollenberger, Anal. Bioanal. Chem. 393 (2009) 225–233. [25] P. Kalimuthu, J. Tkac, U. Kappler, J.J. Davis, P.V. Bernhardt, Anal. Chem. 82 (2010) 7374–7379. [26] P. Kalimuthu, U. Kappler, P.V. Bernhardt, J. Phys. Chem. B 118 (2014) 7091–7099. [27] C. Ruiz-Capillas, F. Jimenez-Colmenero, Food Chem. 112 (2008) 487–493.
289
[28] S. Vandevijvere, E. Temme, M. Andjelkovic, W.M. De, C. Vinkx, L. Goeyens, L.J. Van, Food. Addit. Contam. Part A Chem. Anal. Control Expo. Risk Assess. 27 (2010) 1072–1083. [29] A.A. Belaidi, G. Schwarz, Biochem. J 450 (2013) 149–157. [30] K. Wieghardt, W. Schmidt, W. Herrmann, H.J. Kueppers, Inorg. Chem. 22 (1983) 2953–2956. [31] K. Wieghardt, I. Tolksdorf, W. Herrmann, Inorg. Chem. 24 (1985) 1230–1235. [32] G.D. Broadhead, J.M. Osgerby, P.L. Pauson, J. Chem. Soc. (1958) 650–656. [33] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications (2001). [34] N. Anicet, C. Bourdillon, J. Moiroux, J.-M. Saveant, J. Phys. Chem. B 102 (1998) 9844–9849. [35] M. Situmorang, D.B. Hibbert, J.J. Gooding, D. Barnett, Analyst 124 (1999) 1775–1779. [36] M. Rudolf, S.W. Feldberg, DigiSim version 3.03b, Bioanalytical System, West Lafayette (2004). [37] R.E. Humphrey, M.H. Ward, W. Hinze, Anal. Chem. 42 (1970) 698–702. [38] K. Johnson-Winters, A.C. Davis, A.R. Arnold, R.E. Berry, G. Tollin, J.H. Enemark, J. Biol. Inorg. Chem. 18 (2013) 645–653. [39] S. Frasca, O. Rojas, J. Salewski, B. Neumann, K. Stiba, I.M. Weidinger, B. Tiersch, S. Leimkuehler, J. Koetz, U. Wollenberger, Bioelectrochemistry 87 (2012) 33–41. [40] S. Frasca, A. Molero Milan, A. Guiet, C. Goebel, F. Pérez-Caballero, K. Stiba, S. Leimkühler, A. Fischer, U. Wollenberger, Electrochim. Acta 110 (2013) 172–180. [41] A. Rajapakshe, K.T. Meyers, R.E. Berry, G. Tollin, J.H. Enemark, J. Biol. Inorg. Chem. 17 (2012) 345–352. [42] P. Kalimuthu, S. Leimkühler, P.V. Bernhardt, J. Phys. Chem. B 115 (2011) 2655–2662. [43] R. Spricigo, C. Richter, S. Leimkuehler, L. Gorton, F.W. Scheller, U. Wollenberger, Colloids Surf. A 354 (2010) 314–319. [44] C.A. Temple, T.N. Graf, K.V. Rajagopalan, Arch. Biochem. Biophys. 383 (2000) 281–287. [45] H.L. Wilson, K.V. Rajagopalan, J. Biol. Chem. 279 (2004) 15105–15113. [46] K. Pohl, K. Wieghardt, W. Kaim, S. Steenken, Inorg. Chem. 27 (1988) 440–447. [47] K.-I. Chen, A.G. McEwan, P.V. Bernhardt, J. Biol. Inorg. Chem. 14 (2009) 409–419. [48] P. Kalimuthu, S. Leimkuhler, P.V. Bernhardt, J. Phys. Chem. B 116 (2012) 11600–11607. [49] K.-I. Chen, A.G. McEwan, P.V. Bernhardt, J. Biol. Inorg. Chem. 16 (2011) 227–234. [50] P. Kalimuthu, J. Heider, D. Knack, P.V. Bernhardt, J. Phys. Chem. B 119 (2015) 3456–3463. [51] C.C. Moser, J.M. Keske, K. Warncke, R.S. Farid, P.L. Dutton, Nature 355 (1992) 796–802. [52] G. Gastaminza, S. Quirce, M. Torres, A. Tabar, S. Echechipia, D. Munoz, d.C.L. Fernandez, Clin. Exp. Allergy 25 (1995) 698–703. [53] W.H. Yang, E.C. Purchase, CMAJ 133 (1985) 865–867 880. [54] H.R. Zare, F. Memarzadeh, M.M. Ardakani, M. Namazian, S.M. Golabi, Electrochim. Acta 50 (2005) 3495–3502.