Dimers of Azurin as model systems for electron transfer
Jongh, Thyra Estrid de
Citation
Jongh, T. E. de. (2006, September 12). Dimers of Azurin as model systems for electron
transfer. Retrieved from https://hdl.handle.net/1887/4554
Version:
Corrected Publisher’s Version
License:
Licence agreement concerning inclusion of doctoral thesis in the
Institutional Repository of the University of Leiden
Downloaded from:
https://hdl.handle.net/1887/4554
3
Electron self-exchange in N 42C-BM M E
azurin dim ers as a function of
48 C h a p te r 3
Abstract
The electron self-exchange (e.s.e.) reaction in a covalent homodimer of N42C-BM M E azurin was investigated as a function of temperature in an attempt to determine the reorganization energy O. Rates of intramolecular e.s.e. were obtained by simulation of the redox sensitive Val31 1HJ2 resonance in NM R spectra recorded
DWWHPSHUDWXUHVEHWZHHQDQG.$W7.DORZHUOLPLWIRUN eseintra was
found of 2x104 s-1 which is consistent with previous reports. At lower temperatures
49 C h a p te r 3
Introduction
Azurin is a small blue copper protein that mediates electron transfer (ET) in certain gram-negative bacteria. In vitro it can react with a variety of electron donors and acceptors, including various redox proteins as well as small organic and inorganic molecules. The azurin isolated from P. aeruginosa has become a classical model protein in the study of intramolecular ET.[1-10] The electron self-exchange reaction of
azurin, on the other hand, has served as a model reaction for interprotein ET.[11-14]
Here we report on the analysis of the electron self-exchange reaction in a covalently crosslinked dimer of azurin as a function of temperature. Crosslinking of redox complexes is a means to simplify the interpretation of ET rates since it eliminates the contributions of complex formation and dissociation. Furthermore, co-crystallization of normally short-lived redox complexes has proved to be difficult and crosslinking can be a useful way to stabilize these complexes. A covalent homodimer of azurin was previously constructed by van Amsterdam et al. through site-directed introduction of a surface exposed cysteine residue.[13] This mutation was introduced
at position 42 (N42C), situated close to the so-called hydrophobic patch (HP). This HP is thought to form the main site of interaction between azurin molecules during the electron self-exchange process.[15-18] In order to allow sufficient flexibility
for the complex to adopt a conformation suitable for electron self-exchange, the introduced cysteine groups were reacted with a flexible, bifunctional spacer; bis-maleimidomethylether (BMME). The crystal structure of the N42C-BMME dimer displayed a conformation very reminiscent of the wild type crystal packing in which the HP areas on the individual moieties are close together and the copper centres are positioned at a distance of ~14 Å from each other.[13;19]
ET over such long distances is usually described by the Marcus theory for non-adiabatic ET, according to which ET is a thermally activated process involving the reorganizational energy (O) and the free energy difference between reactants and products ('G0):[20]
50 C h a p te r 3
As reactants and products of an electron self-exchange reaction are identical, in these cases 'G0=0. The reorganizational energy, O, is comprised of an inner-
and outer sphere term. The inner sphere term (Oin) represents the energy that is required to reorganise the ligands during ET. The outer sphere term (Oout) reflects the rearrangement of the second sphere residues and, most importantly, of the solvent shell of the protein. For the blue copper proteins such as azurin Oin tends to be small, owing to its particular ligand set, and the largest contribution to the total reorganizational energy will come from Oout.[2;21] In general, the pronounced
dependence of kET on 'G0 and O allows adaptation of the redox properties of
proteins to their function. W hen -'G0 = O, k
ET is at its maximum and becomes
largely independent of temperature.
Several experimental and quantum chemical studies on wild type azurin and some of its variants have established values for O between 0.6 and 0.8 eV.[2;22-24] Values for
O have, however, not yet been determined in crosslinked complexes of azurin. Due to the conformational ‘locking’ of the complex one can expect O to be significantly affected, in particular the outer sphere contribution. To determine such possible effects we have analyzed kese in N42C-BMME azurin over a temperature range from 312 to 272K.
In studies of the electron self-exchange reaction of this dimer, presented by van Amsterdam et al., fast intramolecular exchange was observed at 312K but only a lower limit could be determined for the rate of this reaction (keseintra
x 104 s
-1).[13] The method used here to obtain electron self-exchange rates relies on the line
broadening of certain redox sensitive NMR resonances as a result of electron self-exchange.[25] However, at exchange rates well above the point at which coalescence
occurs, i.e.:
exchange-51 C h a p te r 3
induced line broadening are significant as compared to the natural line width.
Materials & Methods
N MR sam ple preparation
N42C azurin was produced, purified and crosslinked according to protocol.[13] Two
samples of 1 mM protein each were prepared in 25 mM potassium phosphate (KPi) in 99.9% D2O pH* 8.5 (*pH uncorrected for the deuterium isotope effect). One of the samples was kept fully oxidized while the other was reduced by incubation at RT with 3 molar equivalents dithiotreitol (DTT). Once full reduction was achieved, apparent from the complete loss of the intense blue colour, DTT was removed from the sample by washing with deaerated 25 mM KPi/D2O pH* 8.5. Partially oxidized samples for NMR were obtained by mixing of oxidized and reduced protein in the required ratio. After mixing the samples were left to equilibrate by intermolecular electron exchange for at least 10 minutes prior to measurement.
N MR
1H NMR spectra with a spectral width of 13.98 ppm were recorded on a Bruker
Avance DMX 600 MHz spectrometer equipped with a Eurotherm B-VT2000 thermostat. Free induction decays were accumulated in 32 K memory and Fourier-transformed using a squared sine window function. Each series of measurements was started at 312K and the temperature was then lowered to 303, 295, 284 and 272K consecutively. In between measurements, the sample was allowed to equilibrate to the new temperature for 30 minutes. All 1H spectra were calibrated against an
internal reference of 200 PM 3-(trimethylsilyl)propionate-d4 (TSP). D ata analysis
Homodimeric systems can experience two types of e.s.e.: intermolecular exchange between copper sites on different dimers and intramolecular exchange within the dimer itself.[13] Using the MEX software developed by Bain and Duns, the rates
with which these processes occur can be extracted from experimentally observed line broadening effects.[26] To this end, firstly the line widths and positions of the
52 C h a p te r 3
can be analyzed simultaneously. Based on the exchange rates set for the individual processes, the overall line shape can then be calculated by the MEX program. In the case of the N42C-BMME azurin dimer, the Val31 methyl 1HJ2 NMR resonance
was analysed at ca. 80%, 65% and 40% oxidation. Spectra were recorded also of the fully oxidized (OO) and fully reduced (RR) dimers and the obtained line widths and chemical shift positions were used as input parameters in the simulation procedure. Additional input parameters are the relative populations of each of the defined redox sites, i.e.: 1) an oxidized (ox) site in OO, 2) a reduced (red) site in RR, 3) ox in RO and 4) red in RO. Their populations were derived from the degree of oxidation of the sample given by A628nm relative to a fully oxidized sample under the assumption that the reduction potentials for the two sequential half-reactions for reduction of the dimer are equal. UV-Vis spectra of the NMR samples were recorded on a Perkin Elmer lambda 800 spectrophotometer using a special sample holder equipped with fibre-optics (Hellma). Using the dual display mode in XWINNMR (Bruker Biospin), the simulated and actual spectra were overlayed and the exchange rates were manually iterated until simulations were acquired that best resembled the experimental traces.
Results & Discussion
As observed previously, the position of the Val31 methyl 1HJ2 resonance of
N42C-BMME azurin in the fully oxidized (OO) dimer is shifted compared to that of the fully reduced (RR) dimer [Figure 3.1].[14] At intermediate levels of oxidation, the
resonance is split into 3 signals, the outer two of which are the signals arising from the OO and RR dimers. The signal in between originates from the population of semi-reduced (RO) dimers engaged in fast intramolecular e.s.e..[13] In all spectra
53 C h a p te r 3
Fitting of the Val31 1HJ2 resonances of the RR and OO species at different
temperatures shows that at lower temperatures line broadening occurs, implying an increase of the rotational correlation time Wc [Figure 3.2]. The simulations of the spectra of the partially oxidized samples indicate that, consistent with previous reports on N42C-BMME at 313K, the dimer is engaged exclusively in intramolecular Figure 3.2: Line widths (rad s-1) of the
Val31 1HJ2 resonance of fully oxidized
(Ŷ DQGIXOO\UHGXFHG ¹ 1&%00( azurin as a function of temperature (K). The line widths were determined from simulation of the experimental spectra. Figure 3.1: Region of the 1H NMR spectra of N42C-BMME azurin showing the Val31 methyl resonance at 272, 284, 293, 303 and 312K. The numbers to the left of the traces denote the percentage of total oxidized protein present in the samples. The upper, black, traces show the experimental traces. The simulated traces are shown below in grey. The values for keseintra that were used to generate the simulated
traces are indicated below the spectra. These should be considered only as lower limits for keseintra
54 C h a p te r 3
exchange and intermolecular e.s.e. is essentially absent at all temperatures (keseinter <
103 M-1s-1) . At 312K and 303K a lower limit of k
eseintraê4 s-1 was determined,
ZKLFKLVFRPSDWLEOHZLWKWKHORZHUOLPLWRIê 4 s-1 that was determined by van Amsterdam et al. at 313K.[13] At this rate the line broadening of the Val31 1HJ2
resonances of the oxidized and reduced forms, induced by electron self-exchange, becomes insignificant compared to the line width in the absence of exchange. At temperatures below 293K the lines of the different Val31 resonances all become severely broadened. The rate at which exchange induced line broadening effects are significant compared to the natural line width therefore also substantially decreases, so that at T < 293K keseintraFRXOGRQO\EHGHWHUPLQHGWRDORZHUOLPLWRIê 3
s-1. Furthermore, the line broadening ‘washes out’ the structure of the exchange
pattern and the spectral resolution becomes insufficient for accurate determination of keseintra.
The temperature dependent e.s.e. behaviour of WT azurin has been investigated by Groeneveld et al. under a range of buffer, salt and pH conditions using the redox sensitivity of the His35 and His46 1H resonances.[11;27] Under conditions very
similar to those used in the current study (20 mM KPi, pH 9.0), WT azurin showed Arrhenius behaviour between 323 and 288 K and kese is on the order of 106 M-1s-1.
At temperatures below 288 K the spectral resolution was considered insufficient for accurate determination of kese. Between 323 K and 288 K no indications were found for temperature dependent gating in WT azurin. The strong sensitivity of the line widths in the NMR spectra of N42C-BMME azurin on temperature similarly does not permit us to determine the temperature dependency of kese in the crosslinked complex or the determination of O.
Conclusions
This chapter describes an attempt to determine the reorganizational energy O, associated with the intramolecular electron self-exchange reaction within the covalent N42C-BMME azurin dimer. Rates of electron self-exchange were determined by simulation of the redox sensitive Val31 1HJ2 NMR resonance in partially oxidized
55 C h a p te r 3
partially oxidized samples becomes broad and unstructured. Under these conditions, large variations of kese result in relatively small changes in line shape and values for kese could not be determined with sufficient accuracy for determination of O.
References
[1.] M. J. Bjerrum, D. R. Casimiro, I.-J. Chang, A. J. Di Bilio, H. B. Gray, M. G. Hill, R. Langen, G. A. Mines, L. K. Skov, J. R. Winkler, D. S. Wuttke, Journal of B ioenergetics and B iomembranes 1995, 27 295-302.
[2.] A. J. Di Bilio, M. G. Hill, N. Bonander, B. G. Karlsson, R. M. Villahermosa, B. G. Malmstrom, J. R. Winkler, H. B. Gray, Journal of the American Chemical Society 1997, 119 9921-9922.
[3.] N. M. Kostic, R. Margalit, C. M. Che, H. B. Gray, Journal of the American Chemical Society 1983, 105 7765-7767.
[4.] J. R. Winkler, H. B. Gray, Chemical Review s 1992, 92 369-379. [5.] O. Farver, I. Pecht, Israel Journal of Chemistry 1981, 21 13-17. [6.] O. Farver, I. Pecht, Biophysical Chemistry 1994, 50 203-216.
[7.] O. Farver, L. K. Skov, T. Pascher, B. G. Karlsson, M. Nordling, L. G. Lundberg, T. Vänngård, I. Pecht, Biochemistry 1993, 32 7317-7322.
[8.] O. Farver, L. K. Skov, G. Gilardi, G. van Pouderoyen, G. W. Canters, S. Wherland, I. Pecht, Chemical Physics 1996, 204 271-277.
[9.] O. Farver, I. Pecht, Journal of Biological Inorganic Chemistry 1997, 2 387-392.
[10.] O. Farver, I. Pecht, Proceedings of the National Academy of Sciences of the U nited States of America 1989, 86 6968-6972.
[11.] C. M. Groeneveld, G. W. Canters, E uropean Journal of Biochemistry 1985, 153 559-564.
[12.] C. M. Groeneveld, S. Dahlin, B. Reinhammar, G. W. Canters, Journal of the American Chemical Society 1987, 109 3247-3250.
[13.] I. M. C. van Amsterdam, M. Ubbink, O. Einsle, A. Messerschmidt, A. Merli, D. Cavazzini, G. L. Rossi, G. W. Canters, Nature Structural Biology 2002, 9 48-52.
56 C h a p te r 3
[15.] G. van Pouderoyen, G. Cigna, G. Rolli, F. Cutruzzola, F. Malatesta, M. C. Silvestrini, M. Brunori, G. W. Canters, European Journal of Biochemistry 1997, 247 322-331.
[16.] G. van Pouderoyen, S. Mazumdar, N. I. Hunt, H. A. O. Hill, G. W. Canters, European Journal of Biochemistry 1994, 222 583-588.
[17.] M. van de Kamp, G. W. Canters, C. R. Andrew, J. Sanders-Loehr, C. J. Bender, J. Peisach, European Journal of Biochemistry 1993, 218 229-238. [18.] M. van de Kamp, R. Floris, F. C. Hali, G. W. Canters, Journal of the American
Chemical Society 1990, 112 907-908.
[19.] H. Nar, A. Messerschmidt, R. Huber, M. van de Kamp, G. W. Canters, Journal of Molecular Biology 1991, 221 765-772.
[20.] R. A. Marcus, N. Sutin, Biochimica et Biophysica Acta 1985, 811 265-322. [21.] E. Fraga, M. A. Webb, G. R. Loppnow, Journal of Physical Chemistry 1996,
100 3278-3287.
[22.] H. B. Gray, B. G. Malmstrom, R. J. P. Williams, Journal of Biological Inorganic Chemistry 2000, 5 551-559.
[23.] M. H. M. Olsson, U. Ryde, B. O. Roos, Protein Science 1998, 7 2659-2668. [24.] L. K. Skov, T. Pascher, J. R. Winkler, H. B. Gray, Journal of the American
Chemical Society 1998, 120 1102-1103.
[25.] A. D. Bain, Progress in Nuclear Magnetic Resonance Spectroscopy 2003, 43 63-103.
[26.] A. D. Bain, G. J. Duns, Canadian Journal of Chemistry-Revue Canadienne de Chimie 1996, 74 819-824.
5
