Electroactive Multilayer Assemblies of Bilirubin Oxidase and Human

Electroactive Multilayer Assemblies of Bilirubin Oxidase and Human Cytochrome C Mutants: Insight in Formation and Kinetic Behavior. Franziska Wegerich...
0 downloads 0 Views 5MB Size
Download PDF
ARTICLE pubs.acs.org/Langmuir

Electroactive Multilayer Assemblies of Bilirubin Oxidase and Human Cytochrome C Mutants: Insight in Formation and Kinetic Behavior Franziska Wegerich,† Paola Turano,‡ Marco Allegrozzi,‡ Helmuth M€ohwald,† and Fred Lisdat§,* †

Max Planck Institute of Colloids and Interfaces, Am M€uhlenberg 1, 14424 Potsdam-Golm, Germany; University of Florence, 50019 Sesto Fiorentino (FI), Italy; § Wildau University of Applied Sciences, Biosystems Technology 15745 Wildau, Germany ‡

bS Supporting Information ABSTRACT: Here, we report on cytochrome c/bilirubin oxidase multilayer electrodes with different cytochrome c (cyt c) forms including mutant forms of human cyt c, which exhibit different reaction rates with bilirubin oxidase (BOD) in solution. The multilayer formation via the layer-by-layer technique and the kinetic behavior of the mono (only cyt c) and biprotein (cyt c and BOD) multilayer systems are studied by SPR and cyclic voltammetry. For the layer construction, sulfonated polyaniline is used. The only cyt c containing multilayer electrodes show that the quantity of deposited protein and the kinetic behavior depend on the cyt c form incorporated. In the case of the biprotein multilayer with BOD, it is demonstrated that the catalytic signal chain from the electrode via cyt c to BOD and oxygen can be established with all chosen cyt c forms. However, the magnitude of the catalytic current as well as the kinetic behavior differ significantly. We conclude that the different cytochrome c forms affect three parameters, identified here, to be important for the functionality of the multilayer system: the amount of molecules per layer, which can be immobilized on the electrodes, the cyt c self-exchange rate, and the rate constant for the reaction with BOD.

’ INTRODUCTION The construction of protein multilayer films on electrodes by the layer-by-layer technique is a popular research area especially in the field of biosensors because this is one possibility to drastically increase the surface concentration of the biological recognition element in a controlled way.1-6 Synthetic polyelectrolytes are often applied as building blocks7-11 but also other materials such as DNA6,12,13 or nanoparticles14-16 can be used to incorporate biological compounds within the films. Communication within such assemblies however is an issue because the electron transfer distance drastically increases with the number of deposited layers.17 Thus, often shuttle molecules are used to transfer the information of a redox conversion at the protein molecule to the electrode.6,18,19 The small heme protein cytochrome c (cyt c) is the first redox protein that could be successfully incorporated in a fully redoxactive multilayer assembly using negatively charged sulfonated polyaniline (PASA) as a counter polyelectrolyte.20,21 Charge transfer occurs via electron hopping between the heme centers, which are embedded in a protein shell mainly positively charged at neutral pH.22 A biosensor for superoxide radicals based on this multilayer architecture was reported with a much higher sensitivity in comparison to a cyt c monolayer system.23 r 2011 American Chemical Society

A more sophisticated approach is the embedment of other enzymes in these layers to form a biprotein artificial signal chain.2 Cyt c can act here as electron transfer protein that can conduct the electron transfer between the electrode and the enzyme without any external mediator. First assemblies with cyt c and xanthine oxidase still relied on the generation of an internal shuttle molecule.24 However, reaction partners of cyt c such as sulfite oxidase25,26 or laccase27 allow the construction of electrode assemblies without the need of any diffusing shuttle molecule. This was shown for the multicopper oxidase bilirubin oxidase (BOD) for the first time.28 Cyt c can act as electron donor instead of the natural substrate bilirubin and can thus mediate the catalytic oxygen reduction of BOD.29 This could be successfully applied in a multilayer assembly on gold.28 In this biprotein electrode, electrons are transferred from the gold electrode to the cyt c monolayer and then via the cyt c molecules in the different layers to BOD where finally the four electron reduction of oxygen takes place. It could be shown that the oxygen reduction at the enzyme becomes the rate determining step at low scan rates. However, it is found that at higher scan Received: December 14, 2010 Revised: February 10, 2011 Published: March 14, 2011 4202

dx.doi.org/10.1021/la104964z | Langmuir 2011, 27, 4202–4211

Langmuir rates the supply of electrons to the enzyme becomes important. One possibility for a closer examination of this aspect is to alter the protein-protein reaction rate by protein engineering. Mutational studies of proteins are often used to understand the reaction with its partners or to figure out binding sites. It is known that mutations of cyt c can affect the reaction with its natural reaction partners for example the reaction with cyt c oxidase.30 It could be shown that the lysine residues near the heme edge of cyt c are important for the electron transfer to cyt c oxidase. Human cyt c mutants were also applied as recognition elements in a superoxide sensor with a higher sensitivity due to a higher reaction rate of the cyt c mutants with the oxygen radical.31 In this study, we want to investigate different forms of cyt c including mutants with an extra introduced positively charged lysine residue with respect to the reaction rate with BOD and their electrochemical behavior in a multilayer assembly. Thus, the study is also directed to a deeper understanding of the mechanism of electron transfer within multilayer assemblies based on direct electron exchange between immobilized protein molecules.

’ MATERIALS AND METHODS Materials. Horse-heart cytochrome c, Bilirubin oxidase from Myrothecium verrucaria (EC 1.3.3.5), lysozyme, 11-mercapto-1undecanoic acid (MUA), 11-mercapto-1-undecanol (MU), 3-mercapto-1-propanol, potassium hydroxide, isopropyl β-D-1thiogalactopyranoside (IPTG), ferrous sulfate (FeSO4), and ferriycyanide were provided by Sigma-Aldrich (Taufkirchen, Germany). Potassium dihydrogen phosphate (KH2PO4), dipotassium hydrogen phosphate (K2HPO4), disodium sulfite, sodium dihydrogen phosphate (NaH2PO4), disodium hydrogen phosphate (Na2HPO4), and potassium chloride were purchased from Merck (Darmstadt, Germany). Gold wire electrodes with a diameter of 0.5 mm were provided by Goodfellow (Bad Nauheim, Germany). All solutions were prepared in 18 MΩ Millipore water (Millipore, Eschborn, Germany). Buffers. For the preparation of the sodium phosphate buffer (50 mM and 10 mM), Na2HPO4 or NaH2PO4 was used with the pH adjustment using sodium hydroxide or phosphoric acid. The 10 mM sodium phosphate buffer containing 1 M KCl was adjusted for pH after KCl addition. The potassium phosphate buffers (0.5 mM and 5 mM) were made with K2HPO4 or KH2PO4 salts and the pH was adjusted using hydrochloric acid or potassium hydroxide, respectively. Protein Mutant Preparation. The plasmid pET21a-CCHLhCYC provided by Chuang and co-workers32 carries the human cyt c gene, and the CCHL genes of the yeast heme lyase.33 This plasmid was used for introducing additional mutations and is further referred as wild-type (WT). Mutations were introduced using the QuickChange site-directed mutagenesis kit from Stratagene (La Jolla, USA) following the provided protocol. Gene sequencing was done to confirm the mutations. Human cyt c was expressed and purified using adapted protocols from the literature.32,34 It was coexpressed with heme lyase and cytochrome c maturation proteins. Competent BL21(DE3)C41 E. coli cells were used for transformation of the plasmid DNA with the mutated cyt c gene and the yeast heme lyase. These cells contain an additional pEC86 plasmid, which expresses the bacterial cyt c maturation genes ccmABCDEFGH under the control of the Tet promoter.35 The cells were cultivated in minimal media M9 supplemented with minerals,

ARTICLE

vitamins, and glycerol in a shaker at 37 °C and 180 rpm until OD600 reached 1.0. The expression started after addition of IPTG (1 mM) and FeSO4 (100 mg/L) and the culture was incubated for 72 h at 30 °C and 60 rpm. After harvesting by centrifugation, the cells were lysated using lysozyme and sonication. Purification of cyt c, involved two chromatography steps, first, the supernatant of the centrifuged lysate was loaded onto a 5 mL SP Sepharose cationic exchange column (GE Healthcare, Sweden) and eluted with a linear NaCl gradient (0-500 mM) in 50 mM sodium phosphate buffer, pH 6.8. Pertinent fractions were determined by SDS-PAGE, and those containing cyt c were concentrated using an Amicon ultra centrifugal filter device with a molecular weight cutoff of 5000 kDa (Millipore, USA). The sample was then loaded onto a 120 mL Hi Load 16/60 Superdex 75 prep grade column (GE Healthcare, Sweden) and eluted with 50 mM sodium phosphate buffer, pH 6.8. Fractions were monitored by UV/vis spectroscopy and pooled together to obtain a final cyt c sample. SPR Measurements. A Biacore X and T100 (Biacore AB, Sweden) was used to monitor the multilayer formation by SPR spectroscopy. A flat gold sensor chip was cleaned by a 10 min incubation with 96% (w/w) H2SO4/30% (w/w) H2O2 (3:1) solution. Then the chip was incubated in a MUA/MU (1:3) ethanolic solution for 72 h, and subsequently washed with Millipore water and installed into the Biacore flow system. The following solutions (all in 0.5 mM potassium phosphate buffer, pH 5) were injected subsequently with a flow rate of 1 μL/min: buffer, cyt c (5 min), PASA (5 min), and cyt c (5 min). The latter two steps were repeated four times. Electrochemical Measurements. Cyclic voltammetry was performed with an Autolab PGSTAT 20 and a μAutolab Type II potentiostat (Metrohm, Germany). Amperometric measurements were conducted on a model 720 A potentiostat from CHI Instruments (Austin, TX, USA). For the electrochemical studies, a custom-made 1-ml measurement cell, a Ag/AgCl/1 M KCl reference electrode with a potential of þ0.237 V versus NHE (Biometra, G€ottingen, Germany), and a platinum counter electrode was used. Gold wire electrodes were cleaned following an established protocol:36 First, the wires were boiled in 2.5 M KOH for four hours, then rinsed with water, and finally put in concentrated (98% (w/w)) H2SO4 at least overnight. Just before modification, the electrodes were incubated for 10 min in concentrated HNO3 (65% (w/w)). The immersion depth of the electrodes during the measurements was 4 mm. The experiments were all carried out at room temperature. Reaction of Cytochrome C with Bilirubin Oxidase in Solution. MUA/MU modified gold electrodes were used to investigate the reaction of cyt c with BOD in solution. Cyclic voltammetry 10 μM cyt c prepared in air saturated 50 mM sodium phosphate buffer pH 7.0 was used. The scan rate variation was conducted from 4 to 16 mV/s for BOD concentrations of 50 to 350 nM. Construction of the Cyt C/BOD Multilayer Electrodes. The multilayer assembly was performed with a MUA/MU modified gold electrode with a cyt c monolayer adsorbed (immersion in 20 mM cyt c for 2 h). The electrodes were incubated up to 6 times alternating in a 2 mg/mL PASA solution and a 20 μM cyt c solution (or for the BOD multilayers a protein mixture of 20 μM cyt c and 200 nM BOD) for each 10 min. Between the incubation steps the electrodes were dipped 10 times in 0.5 mM potassium phosphate buffer pH 5. The solutions were also prepared with 4203

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

ARTICLE

Figure 2. Cyclic voltammogram of MUA/MU modified gold electrode in air-saturated 50 mM phosphate buffer pH 7.0 in presence of 10 μM horse heart or N70K cyt c, scan rate 4 mV/s, dashed line: without BOD, straight line: with 200nM BOD.

Figure 1. NMR structure of human cyt c with mutation sites depicted as green sticks. The heme group is presented with blue spheres with the heme iron as red sphere in the center. The image is built with the software Pymol (http://pymol.sourceforge.net) using the PDB file 1J3S. The first conformer of the family of NMR structures is taken.

this buffer. The electrochemical response of the electrodes was controlled by recording cyclic voltammograms in air saturated 5 mM potassium phosphate buffer pH 7. The oxidation peak currents and the reduction peak or catalytic currents were evaluated at fixed potentials and as absolute current values (without applying a baseline). NMR Characterization of the Cyt C/BOD Interaction in Solution. Bilirubin oxidase from Myrothecium verrucaria (EC 1.3.3.5) purchased from Sigma-Aldrich (Taufkirchen, Germany), has a nominal purity of 15%. The protein was therefore purified by a Superdex 75 HR 10/30 column, equilibrated and eluted with 20 mM phosphate buffer pH 6.8. The final purity was checked by SDS-PAGE, and then the protein was lyophilized. Titration of 0.44 mM 15N human cyt c in 20 mM phosphate buffer, pH 6.8 with unlabeled lyophilized BOD was followed through 1H-15N HSQC, up to a cyt c/BOD ratio of 1:1. The possible interaction between the two proteins was monitored through chemical shift changes of the signals from the backbone amide moieties of cyt c. The NMR spectra were acquired at 300 K using a Bruker Advance spectrometer operating at a proton frequency of 800 MHz, equipped with a cryoprobe. NMR spectra were processed with Topspin and analyzed with the program Cara.

’ RESULTS AND DISCUSSION Selection and Evaluation of Human Cytochrome C Mutants. We have chosen human wild-type cyt c and seven human

cytochrome c mutants to be investigated as electron donor for bilirubin oxidase (BOD) in comparison with the horse heart cyt c which was used in former studies.29 In each mutant, a single neutral or negatively charged amino acid is exchanged by a positively charged lysine. The mutation sites are surface exposed

and are located on one site of the molecule with respect to the heme edge (Figure 1). Because electrostatic interactions are often involved in protein-protein interactions as for example for the well studied partners cyt c and cyt c oxidase,37 we aim to investigate whether an introduced extra positive charge in form of the amino acid lysine may alter the reaction of cyt c with BOD. The recently published crystal structure of BOD (PDB file 3gbd) demonstrates a large structural similarity to laccases38 and several hints exist that the electron entrance occurs near to the T1 copper site of BOD.29,39,40 However, because no model of the intermolecular complex was available by the time of the experiments, electrostatic interactions of BOD with cyt c can only be assumed. The human wild-type protein and their mutants are expressed in E.coli and then purified.31 The electrochemical properties of the cyt c forms such as the redox potentials are determined (Table S1 of the Supporting Information). No significant alterations for the determined properties are found among the different proteins. Also, the structural integrity is confirmed with 1H NMR spectroscopy for the oxidized and reduced species. Reaction of the Different Cytochrome C Forms with BOD in Solution. Cyclic voltammetry is used to examine the catalytic reaction of BOD and cyt c in solution. For the first time, it can be shown that also human cyt c reacts with BOD. Although the structure of both forms is superimposable the sequence homology between the human and the horse heart cyt c is 78% with 12 surface amino acids differing.41 A representative cyclic voltammogram of the catalytic reaction of a mutant and horse heart cyt c in comparison is shown in Figure 2. The different magnitude of the catalytic reduction current for the same concentrations indicates that the reaction between N70K and BOD seems to be less effective than for horse cyt c. The electron pathway of the system can be described as follows: cyt c is reduced at the MUA/MU modified gold electrode and then donates its electron probably to the T1 site copper ion where normally bilirubin as natural substrate gets oxidized. Then an intramolecular electron transfer to the T2/T3 trinuclear cluster takes place. Here, subsequently four electrons are transferred to molecular oxygen.42-44 Because cyt c is reoxidized by BOD there is a regeneration of cyt c available for 4204

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

ARTICLE

reduction at the gold electrode. To quantify the differences for the homogeneous electron transfer reaction between BOD and cyt c the second-order reaction rate constant k is determined using the method of Nicholson and Shain.45 Here, first the pseudo first-order kinetic constant kcat is calculated by quantifying the ratio of the catalytic currents (Icat) and the diffusion limited currents (Idiff without BOD) at various scan rates (ν) and for different BOD concentrations and using the relations given in eq 1. λ ¼ 0:1992ðI cat =I diff Þ2 ¼ kcat ðζRT=nFÞ1=ν

ð1Þ

ζ represents the ratio of the stoichiometric coefficients of the reaction between the mediator (here cyt c) and the enzyme, which is equal to four since four molecules of cyt c are oxidized during reduction of one molecule of oxygen. Plotting kcat versus the BOD concentration, the reaction rate constant k can be derived from the slope using eq 2. kcat ¼ k½BOD

ð2Þ

The different second-order reaction rate constants for horse heart cyt c, human wild-type cyt c, and their mutants are presented in Table 1. It can be seen that there is already a twofold difference between horse heart cyt and the human wildtype (0.65 and 1.4 μM-1 s-1 respectively). Table 1. Second-Order Reaction Rate Constants for the Reaction of Cytochrome C (Mutants) with BOD in Solution, Calculated According to Nicholson and Shain (1965) with MUA/MU Modified Gold Electrodes in Air-Saturated 50 mM Phosphate Buffer pH 7.0 in the Presence of 10 μM Cyt C and 50-700 nM BOD (the Experimental Error Was in the Range of 0.2 μM-1 s-1) k (μM-1/s-1)

mutant unmutated cyt c single human lysine mutants

horse heart

0.65

human wild-type

1.40

A50K

0.15

A51K

0.25

E66K

0.58

N70K G77K

0.25 1.30

I81K

0.75

I85K

0.30

The higher reaction rate found for the human wild-type might originate from the 12 amino acids differing between both species. Rodríguez-Roldan et al. have reported on differences in the interaction of cyt c oxidase with human cyt c compared to Arabidopsis and horse cyt c: The second-order rate constant for the human form is by 76% higher compared to the horse cyt c form.46 None of the mutants have a higher reaction rate constant than the human wild-type. Among the mutants an introduction of a lysine at position 50 has the most decreasing effect on the reaction rate (0.15 μM-1 s-1). The horse form of the protein with an asparagine at position 50 has also a decreased rate constant (0.65 μM-1 s-1) compared with the human form (1.4 μM-1 s-1). Also the mutant A51K has a reduced reaction rate (0.25 μM-1 s-1). It is spatially located close to position 50 (Figure 1) and the effect may be explained considering that this area plays a role for the interaction with BOD. Introduction of a lysine at position 77, does not influence the reaction rate significantly. There is a decreasing effect upon introduction of a lysine residue going from position 50 to 51 and then to 77 (Figure 1). Introduction of a Lys at position 66 reintroduces a positive charge in this area of the human form, which does not have a lysine at position 61 as the horse form. Indeed the reaction rate of E66K is almost the same as that of horse heart cyt c. The behavior of N70K, I81K, and I85K where a hydrophobic amino acid has been replaced show a decreased rate constant and this indicates that the presence of hydrophobic residues at these positions is important for the interaction with BOD. N70 and I81 have already been identified in former studies to be important for the interaction and electron transfer reactions of cyt c with other proteins.41,47 The newly available pdb structure file of BOD gives the possibility to analyze the amino acid environment of the T1 copper site. Positively and negatively charged patches as well as hydrophobic patches can be found around this area (Figure S1 of the Supporting Information). This situation is different from that of many other known reaction partners of cyt c where often a negatively charged patch interacts with positively charged lysine residues around the exposed heme edge of cyt c (e.g., ref 48). Chemical shift changes provide a highly sensitive tool for identifying the residues that play a role in interprotein interactions. NMR chemical shift perturbations of backbone amides in wild-type human cyt c in the presence of BOD reveal that the two proteins form detectable amounts of an adduct. Residues, whose chemical shift values are affected by the presence of the partner

Figure 3. Left: SPR experiment showing the formation of the cyt c-PASA assembly for four bilayers on a gold chip modified with MUA/MU, the black bars indicate the time of cyt c flow and the gray bars the time of PASA flowing over the chip, in between buffer is flushing over the chip (a - G77K, b wild-type, c - horse heart, d - N70K). Right: Mean SPR signal for the different cyt c forms for each layer of cyt c and PASA calculated from three independent measurements (Layer 0 corresponds to a cyt c monolayer). 4205

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir molecule, when mapped on the surface of cyt c, are located in different regions, namely 2, 11, 29, 50, 55, 66, 68, 69, 74, 84, 93, 95, 100, 102, 103. Their distribution on the protein surface does not appear consistent with the presence of a single interaction area between the two proteins. The involvement of residues 50, 55, 66, 68, 69, 74, and 84 of the wild-type (identified by NMR) supports the observed effects with the investigated mutants on the rate constant due to positively charged residues at positions 50, 51, 61, 66, 70, 77, 81, and 85. This may support the view that electron transfer occurs through different sites on cyt c surface. The different rate constants provide the motivation to examine some of the cyt c forms also in the immobilized state in a multilayer assembly with BOD. Together with G77K we have chosen also N70K with a large decrease of the reaction rate constant along with horse heart cyt c and human wild-type cyt c to be investigated for comparison as electron donor protein in a layered architecture with BOD. Multilayers with Different Forms of Cytochrome C and PASA. In a first approach, multilayer systems consisting only of different cyt c forms are investigated (schematically shown on the left side of Figure 5). The ability of multilayer formation with sulfonated polyaniline (PASA) is studied and then the electrochemical properties are examined. Study of Multilayer Formation by SPR. The formation process of the different assemblies of cyt c and the polyelectrolyte is

Figure 4. Cyclic voltammograms of a human wild-type monolayer MUA/MU modified gold electrode and of a 4 bilayer electrode with human wild-type cyt c and PASA in each layer, scan rate 100 mV/s, 5 mM phosphate buffer pH 7.0.

ARTICLE

examined by surface plasmon resonance (SPR) spectroscopy. Here cyt c and PASA are deposited on a planar MUA/MU modified SPR-Chip in a flow system to study the mass adsorption during the built-up. The formation process for a four bilayer assembly is given in Figure 3. The human and horse wild-type proteins show rather similar mass deposition during the multiple adsorption steps. However, the mutant N70K exhibits a slightly enhanced immobilization, whereas for the G77K mutant the assembly is rather weak. The different charge distribution on the protein surface might alter the immobilization with the sulfonated polyaniline. However, the drastic decrease in adsorption of G77K compared to the other proteins cannot be fully understood since NMR investigations reveal a rather undisturbed overall structure.31 Despite G77K all sensograms show an exponential increase of the deposited mass during the multilayer assembly. This has already been observed in former studies and is related to partial interpenetration of the layers and an increased surface roughness.49 Electrochemical Behavior. Because the SPR experiments reflect only the amount of immobilized cyt c molecules, the functional properties are analyzed by cyclic voltammetry. For all human cyt c forms, electro-active multilayer electrodes can be constructed. Representative cyclic voltammograms measured for a monolayer electrode and after the assembly of four bilayers of the human wild-type and PASA are given in Figure 4. One can see that oxidation and reduction peaks increase with the number of layers. This means that also the cyt c molecules in the outer layers are in contact with the electrode. Several electron transfer steps have to be considered here (right side in Figure 5): The electron transfer from the gold electrode via the MUA/MU SAM-layer to the cyt c monolayer and then to the cyt c molecules of the subsequent layers. The latter step can be either realized by an interprotein electron transfer only between the cyt c molecules, with PASA being just a stabilizing compound or by PASA being an active part in the electron transfer chain, which can conduct the electrons from one cyt c molecule to another. Several arguments have been collected to support the idea of electron exchange between the cyt c molecules as the dominating mechanism.2 A closer examination of the transfer steps can be performed by a kinetic analysis with a systematic scan rate variation for all multilayer electrodes with the four different proteins. The resulting cyclic voltammograms for horse heart cyt c and the mutant N70K for four different scan rates are presented in parts a

Figure 5. Schematic representation of a Au-MUA/MU-cyt c-(PASA-cyt c)4 multilayer coated electrode (left) and the basic electron transfer steps (right). 4206

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

ARTICLE

Figure 6. Cyclic voltammetry of Au-MUA/MU-cyt c-(PASA-cyt c)4 multilayer electrodes for horse heart cyt c (left) and the human mutant N70K (right) measured at different scan rates in 5 mM phosphate buffer pH 7.0. The current values of the voltammograms b) and d) are normalized to 5 mV/s.

and c of Figure 6. For the mutant N70K, we find that the decrease of the voltammetric peaks with decreasing scan rate is much smaller than for horse heart cyt c. The difference becomes even more obvious when normalization of the current values according to the scan rate is performed. This can be done because there is a linear relation of the current with the scan rate for the immobilized species.50 The cyclic voltammograms normalized in this way are shown in parts b and d of Figure 6. For horse heart cyt c, it is clearly visible that the redox behavior is almost not changing during the scan rate variation in this range. For N70K, the behavior is considerably different. Although for the slow scan rate of 5 mV/s the current reaches the same level as for horse heart cyt c, the voltammetric peaks and thus the amount of electrode-addressable cyt c decreases strongly for higher scan rates. Here, apparently the scan rate becomes too fast, so that not all molecules are converted during the voltammetric scan. Quantifying the peak area of the voltammograms for all mutants gives the electro-active surface coverage Γ at the different scan rates and is depicted in Figure 7. This specifies once more the different electrochemical behavior of the protein multilayers. The horse heart cyt c multilayer electrode exhibits a behavior as already described:20 Up to about 80 mV/s the amount of electrode-addressable cyt c remains stable and then decreases slightly. This suggests that beyond 80 mV/s a scan rate

Figure 7. Electro-active surface coverage Γ for the four different cyt c multilayer electrodes at different scan rates (cyclic voltammetry, 5 mM phosphate buffer pH 7). Lines are drawn to guide the eyes of the reader.

is reached which is in the range of the cyt c - cyt c electron transfer rate (step 2 in the right scheme of Figure 5) so that not all cyt c molecules immobilized in the multilayer assembly can be oxidized or reduced, respectively, in the time period of a voltammetric sweep. For the human wild-type cyt c electrode, 4207

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

ARTICLE

although the behavior at 80-100 mV/s is rather similar compared with horse cyt c, the electro-active amount increases further at slow scan rates. This indicates a slower self-exchange rate of the immobilized protein. For the mutant N70K, this behavior is even more pronounced. At a scan rate of 80 mV/s, only a small amount of cyt c molecules can be detected. The immobilized cyt c molecules in the layers become only visible at very low scan rates. To rule out that the electron transfer from the electrode to the monolayer (step 1 in the right scheme of Figure 5) is the limiting step, the ks values of the cyt c monolayer electrodes are determined (for horse heart, wild-type and N70K). No major decrease is observed (55 ( 5, 64 ( 6, and 45 ( 3 s-1, respectively). The slight decrease for N70K cannot cause the drastic decrease of cyt c molecules contactable at higher scan rates. The experiments with the G77K mutant reflect the results of the SPR investigations. Much less protein can be adsorbed in the multilayer structure and thus detected electrochemically. However, also here the electro-active amount is decreasing with increasing scan rate, with a similar relative change compared to the human wild-type, indicating also a decreased self-exchange rate for this mutant. It is known that the electron self-exchange rate constant for structurally very similar cytochromes c can span over several orders of magnitudes.51 Hence, it is likely that also the rate constants of the horse, the human and the mutant forms differ. To get an estimation of the self-exchange rate kex for the different cyt c forms in the multilayer assembly, the Dahms-Ruff equation has been used, assuming an intermolecular electron hopping process between the cyt c molecules in the multilayer film:52-54 kex ¼ 6Deffective =ðδ2 ½cyt cÞ

ð3Þ

The effective diffusion coefficient Deffective is calculated according to the Randles-Sevcik equation55 by evaluating the peak currents at different scan rates. For horse heart cyt c, a Deffective of 9  10-13 cm2 s-1 is calculated. The approximate distance between the adjacent redox centers δ is estimated to be 2.6 nm and a reasonable cyt c concentration of 76 mM is used. With these values, a selfexchange rate constant of 9.4  102 M-1 s-1 for horse heart cyt c is calculated. It has to be mentioned here that this model is not based on the assumption of diffusing cyt c molecules because they are immobilized, but on the idea of hopping electrons between the cyt c sites. From the analysis of the peak current versus the scan rate, both plots (linear and square root) result in linear dependencies of similar quality (Figure S3 of the Supporting Information), which gives the background for the use of the diffusion model here, however the calculation has to be seen as a rough estimation of the self-exchange rate. Literature values for kex of horse heart cyt c in solution measured with NMR are in the range of 102 M-1 s-1 to 105 M-1 s-1 depending on the ionic strength.51,56 Despite the assumed high ionic strength within the multilayer film due to the polyelectrolyte, the calculated value here is quite low. This can be explained by a reduced mobility of cyt c within the film (mainly rotational). The local concentration of cyt c might be influenced by the specific protein incorporated. This can be particularly relevant for the mutant G77K where the SPR binding analysis shows less ability for assembly formation. However, because an exact determination is not possible we assume for the calculations that the concentration of cyt c is the same for all investigated biprotein multilayer systems. For the human wild-type and the mutants N70K and G77K, kex is thus calculated with the same approach

giving values of 2.5  102 M-1 s-1, 0.3  102 M-1 s-1, and 0.6  102 M-1 s-1, respectively. These estimated rate constants are below the constant of the horse form and reflect the observations and suppositions described above. The slight decrease of the selfexchange rate of the human wild-type protein compared with horse heart cyt might be caused by the fact that some of the amino acids that are different between horse heart cyt and human cyt c are situated in the interaction area identified by NMR. This influences the protein-protein interaction and thereby the electron transfer rate. The decreased self-exchange rate for the two mutant proteins can be understood as an increase of electrostatic repulsions between two cyt c molecules due to the extra positive charge since there are studies indicating that the protein-protein contact region is centered on the heme crevice.57 For the two mutants investigated here, the mutation site is near to this contact area and may modulate the protein recognition process. In an earlier study the importance of surface charges for the selfexchange rate has been shown by chemically modifying single lysine side chains (which are concentrated around the heme crevice) or by binding anions at specific sites near the partially exposed heme edge.56 The dependence of the formal potential and the peak separation on the scan rate can be found in Table 2. The peak separation increases for all protein electrodes with increasing scan rate which reflects the not fully reversible character of the system. For the mutant N70K the increase at low scan rates is much steeper compared to the other multilayer electrodes. This observation supports once more the reduced efficiency of electron transfer within the N70K multilayer assembly. The formal potentials of the cyt c forms in the multilayer assembly are varying only slightly. By evaluating the peak shape of the normalized cyclic voltammograms at small scan rates (Figure 6), one observes an additional voltammetric peak at about þ75 mV for horse cyt c containing assemblies. This can be attributed to the reduction of PASA as it has been shown before.49 For electrodes of the human form, this is much weaker and cannot be observed for the two mutants (Figure S4 of the Supporting Information). Overall the investigations show that the different cyt c proteins exhibit not only a different adsorption behavior during the multilayer assembly but also a different kinetic electrochemical behavior with a limiting interprotein electron transfer. However, electrochemically active multilayer electrodes can be constructed not only with horse heart cyt c but also with at least two human forms of the redox protein. Multilayers of Different Cytochrome C Forms with BOD and PASA. To build up new enzyme containing assemblies with the four different cyt c proteins, the cyt c monolayer electrodes are incubated alternatingly with solutions of PASA and a cyt c/BOD mixture allowing an immobilization of BOD in multiple layers.28 The resulting multilayer electrode is schematically shown on the left side of Figure 8. In the right scheme of Figure 8, the relevant electron transfer steps are depicted. Two additional steps have to be considered: the electron transfer from cyt c to the enzyme BOD (step 3) and from BOD to molecular oxygen (step 4). With all four constructed cyt c-BOD multilayer electrodes, a catalytic current can be observed at low scan rates (Figure 9). Hence, a successful electron transfer from the electrode via cyt c and BOD to oxygen is achieved. However, both mutant multilayer electrodes show a significantly smaller catalytic activity to oxygen reduction 4208

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

ARTICLE

Table 2. Formal Potential Ef and Peak Separation ΔEp for the Cyt C Multilayer Assembly with the Four Different Cyt C Forms at Various Scan Rates (Results from Three Independent Electrodes) scan rate (mV/s)

Ef (mV) horse heart

Ef (mV) human wild-type

Ef (mV) G77K

Ef (mV) N70K

20

-12

-16

-9

-8

40 80

-5 -6

-15 -18

-12 -12

-10 -16

scan rate (mV/s)

ΔEp (mV) horse heart

ΔEp (mV) human wild-type

ΔEp (mV) G77K

ΔEp (mV) N70K

20

7

23

32

59

40

15

36

40

59

80

45

60

51

46

Figure 8. Schematic representation of a Au-MUA/MU-cyt c-(PASA-cyt c/BOD)n multilayer coated electrode (left) and the relevant electron transfer steps (right).

Figure 9. Cyclic voltammetry of Au-MUA/MU-cyt c-(PASA-cyt c/BOD)4 multilayer electrodes, scan rate 2.5 mV/s, 5 mM phosphate buffer, air saturated, pH 7.

compared to the human and horse wild-type. To evaluate the behavior in more detail, a scan rate variation is also conducted here as it was done for the layered electrodes without BOD. Cyclic voltammograms for three different scan rates for the four different cyt c forms normalized to 5 mV/s are presented in Figure S4 of the Supporting Information. Beside the different magnitude of the reduction current at 5 mV/s, it is also obvious that with higher scan rates the catalytic current disappears. Although this holds for all four cyt c forms, the behavior differs.

For instance, at 40 mV/s for horse heart cyt c still a large impact of the catalytic reaction on the cyclic voltammogram can be seen, whereas for N70K the catalytic current disappears already at this scan rate. Because the catalytic current at low scan rate follows the oxygen concentration, as it has already been shown with horse heart/BOD multilayers in a previous study,28 the oxygen conversion of BOD is the limiting step in the signal chain. However, at higher scan rates the catalytic current disappears and the system behaves similar to a multilayer assembly with only cyt c because the scan rate is here too high to allow a transfer of the electrons to the BOD. Because the reaction rate for the immobilized form of the proteins is not known, the limiting process cannot be definitely identified, but the decrease in electro-active amount found for the only cyt c containing systems indicates that the cyt c-cyt c electron transfer (step 2) is rate limiting. A deeper understanding of the kinetics of the electron transfer processes is provided by the evaluation of the different normalized peak current values for reduction and oxidation presented in Figure 10. The increasing reduction current with decreasing scan rate clearly demonstrates the catalytic process at low scan rates. At higher scan rates, the catalytic current disappears. However, the current does not remain at a certain level but shows the tendency of further decrease. This behavior is less pronounced for horse cyt c and reflects the behavior of multilayers only consisting of cyt c as analyzed before. Here, the amount of electrode addressable horse cyt c decreases less steeply than for the other multilayer electrodes. 4209

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

ARTICLE

Figure 10. Normalized cathodic (Ired, left) and anodic (Iox, right) peak current values of Au-MUA/MU-cyt c-(PASA-cyt c/BOD)4 multilayer coated electrodes measured at different scan rates in air saturated 5 mM phosphate buffer pH 7. The current values are normalized to 5 mV/s by dividing the values by the factor [(actual scan rate)/(5 mV/s)]. Current values are derived from three independent electrodes. Lines are drawn to guide the eyes of the reader.

Regarding the efficiency of oxygen reduction by the multilayer electrode, the currents for the human wild-type and horse heart/ BOD multilayers do not differ significantly, although the reaction rate constant of the human form is determined to be higher. This is probably due to a slightly decreased cyt c self-exchange rate of human cyt c. The G77K/BOD multilayer electrode shows a clearly decreased catalytic efficiency compared to all other electrodes. To a large extent this is because only a small amount of protein can be adsorbed but also the decreased self-exchange rate contributes to this behavior. Also, the N70K/BOD multilayer electrode exhibits a decreased efficiency with respect to the human and horse heart cyt c. However, here the immobilization process is not hindered (Figure 3). Thus, for this mutant/BOD electrode step 2 and step 3 are limiting the catalytic current: The rate constant for the reaction with BOD is decreased and already at rather low scan rates the range of the cyt c self-exchange rate is reached. This pictures the reason why only a small reduction peak current can be seen in the left side of Figure 10. However, it has to be noted that the amount of BOD molecules embedded within the layers cannot be given exactly. We assume that the amount of BOD immobilized goes with the amount of cyt c adsorbed without the presence of BOD. The normalized oxidation peak currents plotted against the scan rate, given in the right diagram of Figure 10, also reflect well the described behavior. It can be seen that for all mutants at low scan rates the cyt c oxidation current disappears. With higher scan rates, the cyt c redox peak becomes visible and so the peak current increases. However, the oxidation current does not remain at a certain level because with further enhanced scan rates the electrode addressable amount of cyt c decreases, and so does the oxidation peak current. The different magnitudes of the maximum oxidation peaks reflect the order of differences found for the cyt c self-exchange rate.

’ CONCLUSIONS In this study, different forms of cyt c are investigated with respect to their behavior as electron donors to the enzyme BOD in solution and incorporated in multilayer assemblies with and without BOD. It is shown that the species of cyt c as well as single point mutations in the human form of cyt c can influence the reaction rate with BOD in solution. Hints are found that highlight the importance of hydrophobic interactions between both reaction partners.

It is demonstrated that electroactive monoprotein multilayers can be constructed also with other cyt c forms. Hence, this kind of assembly is not limited to horse heart cyt c. However, it is observed that the amount of protein deposited by the layer-bylayer approach can be negatively influenced by an introduction of a lysine as in the case of the mutant G77K and thus can limit the current density. The distinct kinetic behavior of the multilayer system with different cyt c forms shows that a decreased cyt c selfexchange rate can be also a limiting factor. This corroborates a dominating role of the redox protein for the electron transfer through the system. Multilayer electrodes with BOD and different forms of cyt c show also a diverse catalytic behavior for the signal chain from the electrode via cyt c to BOD and molecular oxygen. Overall, three different parameters are identified, which influence the behavior of the BOD/cyt c multilayers: the amount of molecules per layer which can be immobilized on the electrodes, the cyt c self-exchange rate and the rate constant for the reaction with BOD. Finally, our study illustrates that protein engineering is a helpful instrument to study protein-protein reactions as well as the electron transfer mechanisms of complex multilayer systems.

’ ASSOCIATED CONTENT

bS

Supporting Information. Table of redox potential of cytochrome c, figures of cyclic voltammetry and oxidation peak current of compounds in this article. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: fl[email protected], phone: þ49 (0)3375 508-456, fax: þ49 (0)3375 508-578.

’ ACKNOWLEDGMENT Access to Research Infrastructure activity in the sixth Framework Program of the EC (contract no. RII3-0261454, EU-NMR) and the International Max Planck Research School on Biomimetic Systems is acknowledged for partial support. We thank also Prof. Ulla Wollenberger for fruitful discussions. 4210

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211

Langmuir

’ REFERENCES (1) Tang, Z.; Wang, Y.; Podsiadlo, P.; Kotov, N. Adv. Mater. 2006, 18, 3203–3224. (2) Lisdat, F.; Dronov, R.; M€ohwald, H.; Scheller, F. W.; Kurth, D. G. Chem. Commun. 2009, 274–83. (3) Ariga, K.; Nakanishi, T.; T J. Nanosci. Nanotechnol. 2006, 6, 2278–2301. (4) Caruso, F.; Niikura, K.; Furlong, D. N.; Okahata, Y. Langmuir 1997, 13, 3427–3433. (5) Zhao, W.; Xu, J.-J.; Chen, H.-Y. Electroanalysis 2006, 18, 1737–1748. (6) Rusling, J. F.; Hvastkovs, E. G.; Hull, D. O.; Schenkman, J. B. Chem. Commun. 2008, 141–54. (7) Decher, G.; Eckle, M.; Schmitt, J.; Struth, B. Curr. Opin. Colloid Interface Sci. 1998, 3, 32–39. (8) Song, S.; Liu, H.; Guo, X.; Hu, N. Electrochim. Acta 2009, 54, 5851–5857. (9) Lutkenhaus, J. L.; Hammond, P. T. Soft Matter 2007, 3, 804. (10) Calvo, E. J.; Etchenique, R.; Pietrasanta, L.; Wolosiuk, A.; Danilowicz, C. Anal. Chem. 2001, 73, 1161–1168. (11) Calvo, E. J.; Flexer, V.; Tagliazucchi, M.; Scodeller, P. Phys. Chem. Chem. Phys. 2010. (12) Lvov, Y. M.; Lu, Z.; Schenkman, J. B.; Zu, X.; Rusling, J. F. J. Am. Chem. Soc. 1998, 120, 4073–4080. (13) Sarauli, D.; Tanne, J.; Sch€afer, D.; Schubart, I. W.; Lisdat, F. Electrochem. Commun. 2009, 11, 2288–2291. (14) Katz, E.; Willner, I. ChemPhysChem. 2004, 5, 1084–104. (15) Bonk, S. M.; Lisdat, F. Biosens. Bioelectron. 2009, 25, 739–44. (16) Tang, H.; Chen, J.; Nie, L.; Kuang, Y.; Yao, S. Biosens. Bioelectron. 2007, 22, 1061–7. (17) Ma, H.; Hu, N.; Rusling, J. F. Langmuir 2000, 16, 4969–4975. (18) Moraes, M. L.; Rodrigues Filho, U. P.; Oliveira, O. N.; Ferreira, M. J. Solid State Electrochem. 2007, 11, 1489–1495. (19) Shi, L.; Lu, Y.; Sun, J.; Zhang, J.; Sun, C.; Liu, J.; Shen, J. Biomacromolecules 2003, 4, 1161–7. (20) Beissenhirtz, M. K.; Scheller, F. W.; St€ocklein, W. F. M.; Kurth, D. G.; M€ohwald, H.; Lisdat, F. Angew. Chem., Int. Ed. 2004, 43, 4357–60. (21) Beissenhirtz, M. K.; Kafka, J.; Sch€afer, D.; Wolny, M.; Lisdat, F. Electroanalysis. 2005, 17, 1931–1937. (22) Scott, R.; Mauk, G. Cytochrome C: A Multidisciplinary Approach; Univ Science Books, 1996. (23) Beissenhirtz, M. K.; Scheller, F. W.; Lisdat, F. Anal. Chem. 2004, 76, 4665–4671. (24) Dronov, R.; Kurth, D.; M€ohwald, H.; Scheller, F.; Lisdat, F. Electrochim. Acta 2007, 53, 1107–1113. (25) Spricigo, R.; Dronov, R.; Rajagopalan, K. V.; Lisdat, F.; Leimk€uhler, S.; Scheller, F. W.; Wollenberger, U. Soft Matter 2008, 4, 972–978. (26) Dronov, R.; Kurth, D. G.; M€ohwald, H.; Spricigo, R.; Leimk€uhler, S.; Wollenberger, U.; Rajagopalan, K. V.; Scheller, F. W.; Lisdat, F. J. Am. Chem. Soc. 2008, 130, 1122–3. (27) Balkenhohl, T.; Adelt, S.; Dronov, R.; Lisdat, F. Electrochem. Commun. 2008, 10, 914–917. (28) Dronov, R.; Kurth, D. G.; M€ohwald, H.; Scheller, F. W.; Lisdat, F. Angew. Chem., Int. Ed. 2008, 47, 3000–3. (29) Dronov, R.; Kurth, D.; Scheller, F.; Lisdat, F. Electroanalysis 2007, 19, 1642–1646. (30) D€opner, S.; Hildebrandt, P.; Rosell, F. I.; Mauk, A. G.; Walter, M.; von; Buse, G.; Soulimane, T. Eur. J. Biochem. 1999, 261, 379–91. (31) Wegerich, F.; Turano, P.; Allegrozzi, M.; M€ohwald, H.; Lisdat, F. Anal. Chem. 2009, 81, 2976–2984. (32) Jeng, W.-Y.; Chen, C. Y.; Chang, H. C.; Chuang, W. J. J. Bioenerg. Biomembr. 2002, 34, 423–431. (33) Pollock, W. B.; Rosell, F. I.; Twitchett, M. B.; Dumont, M. E.; Mauk, A. G. Biochemistry 1998, 37, 6124–31. (34) Rivera, M.; Walker, F. A. Anal. Biochem. 1995, 230, 295–302. (35) Arslan, E.; Schulz, H.; Zufferey, R.; K€unzler, P.; Th€ony-Meyer, L. Biochem. Biophys. Res. Commun. 1998, 251, 744–7.

ARTICLE

(36) Ge, B.; Lisdat, F. Anal. Chim. Acta 2002, 454, 53–64. (37) Rieder, R.; Bosshard, H. J. Biol. Chem. 1980, 255, 4732–4739. (38) Mizutani, K.; Toyoda, M.; Sagara, K.; Takahashi, N.; Sato, A.; Kamitaka, Y.; Tsujimura, S.; Nakanishi, Y.; Sugiura, T.; Yamaguchi, S.; Kano, K.; Mikami, B. Acta Crystallogr., Sect. F: Struct. 2010, 66, 765–70. (39) Shleev, S.; Tkac, J.; Christenson, A.; Ruzgas, T.; Yaropolov, A. I.; Whittaker, J. W.; Gorton, L. Biosens. Bioelectron. 2005, 20, 2517–54. (40) Shleev, S.; Elkasmi, a; Ruzgas, T.; Gorton, L. Electrochem. Commun. 2004, 6, 934–939. (41) Banci, L.; Bertini, I.; Rosato, A.; Varani, G. J. Biol. Inorg. Chem. 1999, 4, 824–837. (42) Tsujimura, S.; Kano, K.; Ikeda, T. J. Electroanal. Chem. 2005, 576, 113–120. (43) Shimizu, A.; Kwon, J. H.; Sasaki, T.; Satoh, T.; Sakurai, N.; Sakurai, T.; Yamaguchi, S.; Samejima, T. Biochemistry 1999, 38, 3034–42. (44) Xu, F.; Shin, W.; Brown, S.; Wahleithner J. Biochim. Biophys. Acta, Biophys. 1996, 1292, 303–311. (45) Nicholson, R.; Shain, I. Anal. Chem. 1965, 37, 178–190. (46) Rodríguez-Roldan, V.; García-Heredia, J. M.; Navarro, J. A.; Hervas, M.; De La Cerda, B.; Molina-Heredia, F. P.; De La Rosa, M. A. Biochem. Biophys. Res. Commun. 2006, 346, 1108–13. (47) Wallace, C. J.; Clark-Lewis, I. Biochemistry 1997, 36, 14733–40. (48) Pelletier, H.; Kraut, J. Science 1992, 258, 1748–55. (49) Dronov, R.; Kurth, D. G.; M€ohwald, H.; Scheller, F. W.; Friedmann, J.; Pum, D.; Sleytr, U. B.; Lisdat, F. Langmuir 2008, 24, 8779–8784. (50) Wang, J. Analytical Electrochemistry; Wiley-VCH: New York, 2000; Vol. 3, pp 36-39. (51) Simonneaux, G.; Bondon, A. Chem. Rev. 2005, 105, 2627–46. (52) McKenzie, K. J.; Marken, F. Langmuir 2003, 19, 4327–4331. (53) Ruff, I.; Friedrich, V. J. J. Phys. Chem. 1971, 75, 3297–3302. (54) Dahms, H. J. Phys. Chem. 1968, 72, 362–364. (55) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and Applications; John Wiley and Sons: New York, 1980. (56) Concar, D. W.; Hill, H. A. O.; Moore, G. R.; Whitford, D.; Williams, R. J. P. FEBS Lett. 1986, 206, 15–19. (57) Pielak, G. J.; Concar, D. W.; Moore, G. R.; Williams, R. J. P. Protein Eng. 1987, 1, 83–88.

4211

dx.doi.org/10.1021/la104964z |Langmuir 2011, 27, 4202–4211