Chapter 2

Chapter 2: The pH-dependent redox inactivation of amicyanin as studied by rapid - film voltammetry

Abstract The redox properties of the blue copper protein amicyanin have been studied with slow and rapid scan protein-film . At slow scan rates, when the system is able to equilibrate with the electrode potential, the reduction potential of amicyanin depends on pH in a sigmoidal manner. The data are analysed by assuming that a single titrating group with a redox state dependent pKa is responsible for the pH-dependence of red ox the reduction potential, which gave pKa = 6.3, pKa £ 3.2 at 22 °C. Voltammetry at higher scan rates shows that the low-pH form of amicyanin is not oxidised directly. Instead, oxidation occurs only after conversion to the high-pH form. Simulations show that this conversion is achieved with a rate constant > 750 s-1 at 25 °C. In order to slow the coupled reaction down, the experiments were performed at 0 °C, at which a rate constant of 35 ± 20 s-1 was determined. In conjunction with evidence from NMR, the data are discussed within a scheme that incorporates protonation and dissociation of the copper coordinating histidine 96 in the reduced form.

27 Protein film voltammetry on amicyanin

Introduction This paper deals with the dynamics of the pH-dependent conformational change that occurs in the active site of the small blue copper protein, amicyanin. Amicyanin functions as the obligatory electron acceptor of methylamine dehydrogenase (MADH) [Husain et al., 1985; Tobari et al., 1981; van Houwelingen et al., 1985; van Spanning et al., 1990], an that is essential when bacteria like Paracoccus versutus are grown on mimimal medium with methylamine as the sole source of carbon, nitrogen and energy [Dennison et al., 1998]. MADH is located at the beginning of the energy generating redox chain and is responsible for the conversion of methylamine into formaldehyde.

As with most blue copper , the single copper ion in amicyanin from P. versutus is anchored to the protein framework by two histidines (His36 and His96), a cysteine (Cys92) and a methionine (Met99) [Romero et al., 1994]. At low pH, the C-terminal histidine ligand (His96) becomes protonated and dissociates from the Cu, a property amicyanin shares with plastocyanin and pseudoazurin [Dennison et al., 1994b; Lommen et al., 1988; Sykes, 1991a]. The accompanying change from four- to three-fold coordination leads to a large increase in reduction potential [Dennison et al., 1996a], illustrating the stabilising effect the three-fold coordination has on the Cu(I) state. NMR experiments with reduced amicyanin indicate that protonation of His96 coincides also with a decrease in the self- exchange rate of amicyanin [Lommen et al., 1990]. It has been proposed that His96 might thus act as a switch that controls the physiological flow of electrons along the MADH redox chain [Guss et al., 1986; Ubbink et al., 1994].

Crystal structures of reduced plastocyanin and pseudoazurin at various pH values showed that upon protonation of the copper-coordinating histidine, the copper ion is displaced towards the plane of the remaining ligands (Sg(Cys), Nd(His) and Sd(Met)), resulting in a trigonal planar geometry [Guss et al., 1986; Vakoufari et al., 1994]. EXAFS studies on amicyanin have led to a similar conclusion [Lommen et al., 1991]. To our knowledge, protonation has never been observed when the metal is oxidised, which signifies a profound decrease of the pKa’s of these histidines.

The kinetics of the dissociation/association of His96 have been studied at 25 °C by 1H NMR [Lommen et al., 1988; Lommen et al., 1990]. It appears that, when dissociated, His96 may occur in two conformations, possibly corresponding to different rotamers of the His96 side chain, similar to that observed in crystals of reduced plastocyanin at various pH

28 Chapter 2 values [Guss et al., 1986]. Since most of the mechanistic information has been obtained indirectly, it is important to complement these data by more direct measurements of redox activity. I have therefore performed cyclic voltammetry measurements on films of amicyanin immobilised on an electrode, where, by varying the pH and the scan rate, interconversion between the protonated and unprotonated forms can be observed directly. Furthermore, the variation in reduction potential can be studied over a larger range of pH values than before, leading to a more complete picture of the redox behaviour. In order to compare the results with the NMR studies, measurements were performed at 25 °C. However, since at this temperature the kinetics are very fast, the system was also studied at a second, lower temperature (0-2 °C) at which more accurate rate constants could be determined.

Methods Amicyanin from P. versutus was obtained from a heterologous expression system in E. coli as described before [Kalverda et al., 1994]. For the cyclic voltammetry studies a mixed buffer system (20 mM 2-[N-cyclohexylamino]-ethanesulfonic acid (CHES), 20 mM N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid] (HEPES), 20 mM 2-[N- morpholino]-ethanesulfonic acid (MES) and 50 mM sodium-acetate) was used. Besides the mixed buffer, the cell solutions contained 50 mM NaCl as supporting electrolyte and 2 mM neomycin sulfate was included as a co-adsorbate to stabilise the protein layer on the electrode. In the absence of neomycin, no electrochemical response was observed. The pH of the cell solutions was adjusted with HCl or NaOH at the temperature used in the experiment and checked again after the measurements.

Analogue cyclic voltammetry was performed using an Autolab electrochemical analyser (Eco-chemie, Utrecht, the Netherlands) equipped with a PGSTAT 20 and a fast analogue scan generator (SCANGEN), in combination with a fast analogue to digital converter (ADC750). The electrochemical cell and pyrolytic graphite ‘edge’ (PGE) working electrode have been described previously20. Fourier transform smoothing and background subtraction were achieved using a computer programme (provided by H.A. Heering) that fits a cubic spline function [Press et al., 1989] to the baseline in regions sufficiently far from the peak and interpolates throughout the peak region. Numerical modeling was carried out as described before [Hirst et al., 1998a].

29 Protein film voltammetry on amicyanin

Protein solutions for preparing films were typically 0.5 mM in 20 mM HEPES, pH 7.0 and 2 mM neomycin. Prior to voltammetric experiments, the PGE working electrode was polished with 3 or 6 micron diamond polish (Kemet Europe B.V.) on a polishing cloth and rinsed with water. A film was applied by adding 1-2 ml of protein solution to the electrode surface with a plastic pipette-tip and retracting the solution shortly after.

For NMR measurements, amicyanin (in 20 mM HEPES, pH 7.0) was reduced with sodium dithionite. The (reducing) buffer was exchanged for D2O and the sample was concentrated to ~5 mM by means of ultrafiltration. The pH was adjusted using 0.1 M NaOD or DCl, but was not corrected for the deuterium isotope effect. NMR spectra were recorded on a Bruker Avance DMX 600 MHz spectrometer and spectra were analysed using standard procedures and Bruker software. Combinations of spectra obtained with a Carr-Purcell-

Meiboom-Gill (CPMG) pulse sequence [90°-t-(180°-2t)n-180°-t] (n=59; t=1ms) and a Hahn-spin-echo (HSE) pulse sequence [90°-t-180°-t] (t=60ms) were used to 0.40 22°C discriminate between singlets, doublets 0.36 and triplets in the aromatic region HE) S 0.32 [Campbell et al., 1975; Carr et al., 1954]. vs (V

/ 0.28

The temperature of the NMR spectrometer 0’ E 0.24 was checked by measuring the HDO shift 0.20 with respect to 3-trimethylsilylpropionate 3 4 5 6 7 8 9 0.40 (TMSP-2,2,3,3-d4). 2°C 0.36 HE) S 0.32 vs (V

Results and discussion / 0.28 0’ E pH-dependence of the reduction potential 0.24

Figure 1 shows the pH-dependence of the 0.20 3 4 5 6 7 8 9 reduction potential of amicyanin as pH determined with protein-film voltammetry Figure 1: Reduction potential (E0’) vs pH of at 2 °C and 22 ºC. At scan rates below 0.5 amicyanin. E0’ was measured by protein film -1 voltammetry, commencing with a reductive V s , the system is able to equilibrate with poise and with a scan rate of 0.1 V s-1. The the electrode potential (vide infra) and the voltammograms were measured at 2°C and 22°C, mixed buffer, 50 mM NaCl and 2 mM oxidative and reductive peaks of the neomycin. The line shows the least squares fit voltammograms are symmetrical at all pH using eq. 1 (see text), with values shown in Table 1. All values are referenced to the values. Note, however, that at these ‘slow’ standard hydrogen electrode (SHE).

30 Chapter 2 scan rates the peak separation (35-40 mV)* and the width at half-height (~100 mV) are greater than the ideal one-electron values (0 and 84 mV at 2 ºC, respectively and 0 and 90 mV at 22 ºC, respectively) [Armstrong et al., 1997; Hirst et al., 1998b]. Further lowering scan rates did not decrease the peak separation or width at half-height.

0’ The data were fitted to Equation 1 [Moore et al., 1990], in which E alk is the limiting value ox of the formal reduction potential at high pH when His96 is not protonated and Ka and red Ka are the proton dissociation constants of His96 in the oxidised and reduced states, respectively.

æ ö K red æ K ox + éH+ ùö ç a ç a ê ú÷ ÷ 0' 0' RT ç è ë ûø ÷ E = Ealk - ln (1) nF ç ox æ red é + ùö ÷ ç Ka ç Ka + H ÷ ÷ è è ëê ûúø ø

red The best fit (see Figure 1) to the data gave the values shown in Table 1. The value of pKa at 22 °C (6.3 ± 0.1) is comparable to the previously reported values of 6.6 [Dennison et al.,

1996a] (in H2O) and 6.8 [Lommen et al., 1988] (in D2O) [Lommen et al., 1990] for amicyanin in solution (both at 25 ºC). The first value was based on electrochemical data in a similar manner as reported here, while the latter value was determined by NMR. Values ox for pKa could not be determined previously due to instability of amicyanin at pH values below 4, and like amicyanin in solution, the protein-film became increasingly unstable below pH 4, as was apparent from the rapid disappearance of the voltammetry signals over successive scans. This is in contrast to the signals at higher pH values which were reversible and stable. Although the experimental data in Figure 1 give a good fit to equation 1, I cannot exclude the possibility that the levelling off that is apparent for the reduction potential at low pH is due instead to protein instability. Therefore, the values for ox pKa are quoted as upper limits.

Voltammograms starting from a reductive poise Figure 2 shows voltammograms of the protein films at pH 7.08 and 0 °C, measured with increasing scan rates and commencing from a reductive poise. At this pH value His96 remains largely unprotonated (even in the reduced form, less than 20% is protonated). Therefore, the peak separation depends mainly on the electron-transfer rate, i.e., higher

* The reduction potentials shown in Figure 1 are determined by taken the average of the peak positions.

31 Protein film voltammetry on amicyanin

13 mV/s 1.3 V/s 130 V/s

1 1 1

pH = 7.08

1 0 0 0

2

-1 -1 -1 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9

E / (V vs SHE) Figure 2: Background-corrected voltammograms recorded for amicyanin, commencing from a reductive poise. In each case the peaks have been normalised to a Nernstian peak height of unity, i.e., the maximum peak intensity of a theoretical (Nernstian) peak is set to 1. The voltammograms were measured at 0°C, mixed buffer, pH 7.08, 0.05 M NaCl, 2 mM neomycin. After a pre-treatment of 20 s at 0.54 V, a reductive poise (0.06 V) of 20 s was applied prior to the scans. The scan rates are indicated in the figure. scan rates produce larger peak separations and greater peak widths. This is illustrated in Figure 3, which shows the oxidative and reductive peak positions as a function of scan rate. The electron-transfer behaviour can be modelled using the Butler-Volmer equations (Eq. 2)

0’ 0’ kred = k0 exp {-anf(E – E )} and kox = k0 exp {(1-a)nf(E – E )} (2)

Here, f is F/RT, i.e., the Faraday constant divided by the temperature and the gas constant, kred and kox are the potential-dependent rate constants for reduction and oxidation, respectively, k0 is the standard (exchange) rate constant at zero over-potential and a is the transfer coefficient, which was set at 0.5 [Hirst et al., 1998b]. The numerical model based on Eq. 2 [Hirst et al., 1998a] predicts that the peaks separate symmetrically about the reduction potential (E0’) as the scan rate is increased. Figure 3 includes the simulation using a finite difference procedure [Hirst et al., 1998a] (solid lines). A good fit was -1 obtained with k0 = 600 ± 100 s . Note that a constant peak separation of 40 mV had to be introduced in the simulation to account for the observed non-ideal peak separation, the origin of which is under further investigation.

32 Chapter 2

Voltammograms starting from a 0.40 oxidative poise 0.35

Figure 4 (top) shows the background HE) 0.30 S

corrected voltammograms of vs 0.25

(V 0.20

amicyanin at pH 7.08 and at 0 °C, at / different scan rates starting from an E 0.15 0.10 oxidative poise. Although the peaks 0.01 0.1 1 10 100 are symmetrical at all scan rates, the Scan rate / (V s-1) reduction potential decreases as the Figure 3: Peak positions (vs SHE) in the voltammograms scan rate is increased (Figure 4 and 5, of amicyanin as a function of scan rate (log scale) at the conditions mentioned in Figure 2, commencing with a top). This in contrast to the reductive poise. The line represents the values predicted by -1 the model (see text) with k0 = 600 s . A constant peak voltammograms that followed of a separation of 40 mV was added to the model. reductive poise (Figure 3). A similar effect was observed for the blue copper protein azurin and was attributed to potential- dependent surface reactions of the PGE electrode [Hirst et al., 1998b]. Recently, I have performed extensive studies (which are not part of this thesis) and on basis of the results I have proposed models that account for these non-ideal observations. The models include multiple protein states, which interconvert slowly compared to the scan rates used. This work will be published elsewhere, while here I did not attempt to include it in the analysis, since the decrease in reduction potential is not directly associated with amicyanin. Instead, simulations were corrected by adding a logarithmic decrease to E0’ with increasing scan rate (thus the correction seems linear in Figure 5, in which the scan rate is shown on a -1 logarithmic scale). The best fit was obtained with k0 = 2000 ± 400 s , although the fit deviates from the experimental data at higher scan rates. This indicates that the logarithmic ‘addition’ does not correct for all the electrochemically coupled processes of the PGE electrode. Also note that the value of k0 obtained from these fits is considerably higher than the value determined when commencing from a reductive poise (600 ± 100 s-1). However, the electron transfer rates are independent from the histidine protonation and, therefore, the discrepancy in k0 does not influence the analysis of the protonation kinetics.

33 Protein film voltammetry on amicyanin

13 mV/s 11.6 V/s 130 V/s

1 1 1

pH = 7.1

2 0 0 0 1

-1 -1 -1 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 1 1 1

pH = 5.8

0 0 0

-1 -1 -1 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 1 1 1

pH = 5.0

0 0 0

-1 -1 -1 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 -0.3 0.0 0.3 0.6 0.9 E / (V vs SHE)

Figure 4: Background-corrected voltammograms recorded for amicyanin, commencing from an oxidative poise. In each case the peaks have been normalised to a Nernstian peak height of unity, i.e. the maximum peak intensity of a theoretical (Nernstian) peak is set to 1. The voltammograms were measured at 0°C and in a mixed buffer with 0.1 M NaCl and 2 mM neomycin. After a pre-treatment of 20 s at 0.06 V, an oxidative poise (0.54 V) of 20 s was applied prior to the scans. The scan rates and pH values are indicated in the figure.

34 Chapter 2

Histidine protonation At lower pH, protonation of His96 is expected to affect the electrochemical behaviour of amicyanin. Figure 4 also shows the background-corrected voltammograms at pH 5.84 and 5.04. At low scan rates (< 0.5 V s-1, Figure 4 left) the system is able to equilibrate with the electrode, as is evident from the symmetrical oxidative and reductive peaks*. By contrast, when the voltammograms at pH 5.84 and 5.04 were acquired at higher scan rates (~ 10 V s-1, Figure 4, middle), the reduction peak remained sharp whereas the oxidation peak became broadened and severely attenuated. At pH 5.04, which is 1.4 pH units below red pKa , the oxidative peak became flattened and broadened to such an extent that the background could not be subtracted reliably. This suggests that the oxidation reaction becomes “gated” by a preceding step, that is, the re-oxidation of amicyanin cannot occur before His96 is deprotonated and the copper ion is re-coordinated by the imidazole ring (see Scheme 1A). Thus, reduced amicyanin with a protonated His96 is ‘redox-inactive’ and the oxidation rate is controlled by the rate of conversion to the ‘redox-active’ unprotonated state.

Figure 4 (right) also shows voltammograms measured at 130 V s-1. At this scan rate oxidative waves are observed at all pH values and the reduction potentials obtained from the signals are similar in all cases. This shows that reoxidation is initiated before His96 is able to protonate and convert to the redox-inactive state: thus, by applying sufficiently fast scan rates, amicyanin is ‘trapped’ in its redox-active state. To confirm this interpretation, the fast-scan experiments were repeated following a reductive poise. At low pH values, amicyanin should now be in its reduced and protonated state and, as expected, the signals were absent for the first cycle. In other words and by contrast with the oxidative poise, amicyanin is trapped in its ‘redox-inactive’ state.

Simulations In Figure 5 the peak positions are depicted as a function of scan rate for the three pH values at 0 °C. In each case the voltammetric cycle has been commenced from an oxidative poise. At low scan rates, the system is reversible and the system equilibrates with the electrode potential throughout the cycle. As the scan rate increases the oxidation becomes gated, and at high scan rates (> 25 V s-1) the potentials of the oxidative peaks decrease. Simulating the data required accounting for the protonation of His96, which was achieved

* A difference in peak height can be detected at the very slow scan rates (< 0.1 V/s) at pH 5.84 and 5.04. This is due to slow loss of the protein film during the scan.

35 Protein film voltammetry on amicyanin by including the (de)protonation rates as given by Equation 3 (see ref. [Hirst et al., 1998a] for more details), in which Cu(I)His96-H+ is ‘redox inactive’, according to Scheme 1A,

+ 96 96 + rateon = kon[H ]{[Cu(I)His ]}surf and rateoff = koff{[Cu(I)His -H ]}surf (3) red Ka = koff/kon

red With pKa = 6.4, k0 = 2000 -1 0’ s and E alk = 231 mV (vide supra) the best fit for the three 0.35 0.30 pH = 7.08 pH values was obtained with 0.25 -1 koff = 35 s (see Figure 5). 0.20 0.15 Although the fit describes the 0.10 general behaviour of the peak 0.05 0.00 positions in Figure 5, the 0.01 0.1 1 10 100 0.35 optimal k determined from off 0.30 pH = 5.84 the data at pH 5.84 (koff = 45 HE) 0.25 S -1 0.20 s ) differed from that vs

(V 0.15 /

determined at pH 5.04 (koff £ E 0.10 25 s-1). An explanation for this 0.05 0.00 may be that there are 0.01 0.1 1 10 100 0.35 additional equilibria that are 0.30 pH = 5.04 not accounted for in Scheme 0.25 2.5 2.0 1A. Therefore, a relatively 0.20 1.5 1.0 large error is introduced in koff 0.15 0.5 0.0

Peak Height (A.U.) 0.01 0.1 1 10 100 = 35 ± 20 s-1, as determined 0.10 by multiple simulations. 0.01 0.1 1 10 100 -1 Scan rate / (V s )

Protonation studied with Figure 5: Peak positions (closed symbols) in the voltammograms NMR of amicyanin as a function of scan rate (log scale) at the conditions 1 mentioned in Figure 4 (commencing with an oxidative poise). The Previous H NMR studies insert shows the normalised peak heights as a function of scan rate -1 [Lommen et al., 1990] (log scale). Due to the 'flattening' of the peaks at scan rates > 5 V s the peak maxima (open symbols) are difficult to determine showed that the protonation/ accurately and exhibit relatively large errors. The lines have been -1 calculated with the model discussed in the text with k0 = 2000 s , -1 red 0’ deprotonation equilibrium at koff = 35 s , pKa = 6.4 and E alk = 231 V. The broken line 25 °C is fast on the NMR indicates where according to the simulation the (normalised) peak height is less than 10% of the peak height obtained at low scan rates 5 -1 -1 time-scale (k-1 = 2.7 x 10 s ). (< 0.5 V s ).

36 Chapter 2

However, after protonation the histidine was observed to occur in two conformations that 2 -1 3 -1 interconvert relatively slowly (k2 = 2.5´10 s ; k-2 = 2.1´10 s at 25 °C). Thus, the complete model for reduced amicyanin involves a three-state equilibrium as shown in Scheme 1B, and only ~10% of the protonated amicyanin is present in the [Cu(I)His96-H+]* state under the conditions used (25 °C, 20 mM phosphate in D2O). The existence of two protonated conformations was postulated in order to explain why the NMR signals broaden upon lowering the pH [Lommen et al., 1990]. Although the [Cu(I)His96-H+]* species could not be observed directly because of its low concentration and the broadness of its His96 CdH signal, the chemical shift of the latter proton was concluded to be in the range between 7.7 and 7.9 ppm on the basis of spectral simulations [Lommen et al., 1990].

Table 1: Equilibrium and rate constants for reaction connected with the protonation of His96 of Amicyanin 0 – 3 °C 22 – 25 °C CV NMRb CV NMRb red a a pKa 6.4 ± 0.1 7.3 6.3 ± 0.1 6.8 ox pKa £ 3.4 ± 0.2 £ 3.2 ± 0.2 Equilibria 0’ E alk (mV) 231 ± 3 221 ± 3 0’ E acid (mV) ³ 395 ± 10 ³ 402 ± 10 k (s-1) 2.7 ´ 105 a Rates alk ® acid for Cu(I) 35 ± 20 < 104 > 750 2.1 ´ 103 a a Values taken from ref. [Lommen et al., 1988] and [Lommen et al., 1990]. b red In D2O with pKa not corrected for the isotope effect.

When comparing the previous NMR with the present CV data I note that the NMR experiments indicate that there is an additional equilibrium involved in the conversion of protonated species into the deprotonated [Cu(I)His96] form (see Scheme 1B) [Lommen et al., 1990]. The first fast equilibrium is the actual protonation, while the second is of unknown origin, although it possibly corresponds to different rotamers of the His96 side chain similar to that observed in crystals of reduced plastocyanin at various pH values [Guss et al., 1986]. I will now elaborate on which of these two equilibria might correspond to the equilibrium as found by cyclic voltammetry.

37 Protein film voltammetry on amicyanin

The first-order conversion rates 0.40 at 25 ºC are k = 2.7 x 105 s-1 0.35 pH = 5.84 -1

HE) 3 -1

S 0.30 and k-2 = 2.1 x 10 s , both of

vs 0.25 which are considerably larger (V 0.20 /

E 0.15 than the rate constant of 35 ± 20 0.10 s-1 determined by voltammetry 0.01 0.1 1 10 100 at 0 °C. In view of these large Scan rate / (V s-1) differences I repeated the CV Figure 6: Peak positions (vs SHE) in the voltammograms of experiments at 25 °C (see amicyanin as a function of scan rate (log scale) at 25°C, buffer mix, pH 5.81, 0.1 M NaCl, 2 mM neomycin, commencing from an Figure 6). Indeed, at this oxidative poise. Contrary to the findings at 0°C (Figure 5, middle, temperature the system appears pH 5.84) the data in this figure show that the system is in thermodynamic equilibrium at scan rates up to 160 V s-1. reversible up to a scan rate of

250 V s-1, indicating an -1 increased koff with a lower limit of koff > 750 s as determined from simulations. Interestingly, the decrease in reduction potential observed upon increasing the scan rate at 0 °C (Figure 5, top) largely disappears at 25 °C and E0’ is almost the same for all scan rates.

It remained to be seen if the CV data at 0 ºC are consistent with NMR when the two sets of experiments are performed under comparable conditions. Therefore I repeated the NMR experiments at 3 ºC. At high pH the [Cu(I)His96] His96 CdH signal is observed at 7.27 ppm. When the pH* is lowered this peak broadens and moves to a new position at 7.82 ppm (see Figure 7), within the range of values previously identified as being characteristic of the [Cu(I)His96-H+]* species. These observations are consistent with [Cu(I)His96-H+]* being the dominant protonation form under these conditions (the other one being [Cu(I)His96- H+]), with [Cu(I)His96] and the [Cu(I)His96-H+]* species interconverting via a slow exchange process (< 104 s-1). This would be in qualitative agreement with the CV data if the slow step observed in the CV experiments also represents the interconversion of [Cu(I)His96-H+]* and [Cu(I)His96-H+] species. A more elaborate NMR study is necessary in order to check the quantitative agreement between the data stemming from the CV and the NMR experiments.

38 Chapter 2

pH = 10.00

pH = 8.53

pH = 8.14

pH = 7.49

pH = 6.84

Figure 7: 1D NMR spectra of ~5 mM amicyanin (3°C) in D2O at different pH values obtained by using the Hahn Spin Echo pulse sequence.

Combining the NMR and CV data; three reduced states, with two equilibria can be identified. The first equilibrium is rapidly established and involves the actual protonation, and therefore, dissociation from the Cu(I) of His96. The second ‘slower’ process involves a conformation change, after which amicyanin becomes ‘redox-inactive’.

39 Protein film voltammetry on amicyanin

Scheme 1

A B CV NMR Cu(I)His96 0’ E alk. Cu(II)His96 Cu(I)His96 k-1 k1 k0 red pKa k k off on Cu(I)His96 - H+

k k Cu(I)His96 - H+ -2 2

[Cu(I)His96 - H+]*

Conclusion We have shown that ‘rapid’ scan cyclic voltammetry on adsorbed amicyanin-films complements the insight into the redox-activity and kinetics of interconversion between different states of amicyanin that is obtained by NMR. The reduction potential could be determined down to pH 3, which considerably extends the pH range as compared to previous determinations. The pH-dependence of the reduction potential is due to protonation/deprotonation equilibration of His96, for which pKa values are shown in Table 1 for two temperatures. The oxidation of the low-pH form of amicyanin, in which His96 is protonated, is gated by its conversion to the high pH form. The rate of this conversion (koff) was determined to be 35 ± 20 s-1 at 0 °C while, at 25 °C, the kinetics are considerably faster and only a lower limit for koff could be obtained.

40