Chapter 1

Electron Transfer in Proteins

Marcus theory in biological electron transfer

Electron transfer (ET) is central to biological organisms' ability to manage energy. For example, the respiratory chains of many different types of bacteria consist of a series of electron transfer steps. The initial event in photosynthesis is the transformation of light energy into a charge separation across a membrane. And, in oxidative phosphorylation in mitochondria, electron transfer events are coupled to proton transfers that build up the chemiosmotic gradient which drives ATP synthesis. The way the amino acids and cofactors of these protein complexes control electron flow has been the subject of much theoretical and experimental work (Marcus and Sutin 1985; Beratan et al. 1991; Moser et al. 1992; Winkler and Gray 1992; Evenson and Karplus 1993; Farid et al. 1993; Friesner 1994; Stuchebrukhov 1996).

The most generally useful theoretical framework for thinking about electron transfer is Marcus theory (Franzen et al. 1993). Marcus's essential insight (Marcus 1993) is that the equilibrium positions of the nuclei around the donor and acceptor atoms are different before and after electron transfer and that, according to the Franck-Condon principle, nuclear rearrangements occur on much longer time scales than electronic changes. So, in order to conserve the total energy of the system, electron transfer must proceed either by pre-equilibration to a nuclear configuration whose energy is the same for the DA and D+A- electronic configurations (thermal ET reactions) or by addition of light energy to make the energy of the DA complex equal to that of a D+A- complex with the DA equilibrium nuclear configuration (photoinduced ET reactions). The energy that must be supplied to overcome the lack of prior rearrangement of the nuclei is the reorganization energy, l, defined as the charge transfer energy for a system where theDGŚ for the reaction is zero (self-exchange reactions). (See figure 1.)

The DGŚ for an ET reaction is the difference in the reduction potentials of the donor and acceptor. According to semiclassical Marcus theory, the energies of the products and reactants can be plotted as a function of an abstract reaction coordinate which takes into account the positions of all nuclei relevant to the system (the donor and acceptor plus a solvation sphere). The energies for the products and reactants, plotted as a function of the reaction coordinate, form two parabolas. The point at which they cross gives the activation energy,DG*, which is a function of the driving force for the reaction, -DGŚ, and l, the nuclear reorganization barrier to the reaction: ∆G* = (∆GŚ +l)2/4l. Taking a semiclassical approach, one can describe the rate of electron transfer by the following equation:

kET = 2 exp-

where (HAB)2 is the electronic coupling, and the exponential term is the Franck-Condon or nuclear energy term.

An unusual prediction of the Marcus equation, the so-called inverted effect, comes from the form of the exponential term in the nuclear (Franck-Condon) factor, exp-(DGŚ+l)2/4lkBT. This indicates that as the driving force for the reaction increases (DGŚ becomes more negative), the activation barrier, DG*, decreases until, at -DGŚ= l, it reaches zero. (See figure 2.) Then as the driving force increases further (DGŚ becomes even more negative), the activation energy increases again so that kET decreases with increasing driving force. (See figure 3.) The presence of this inverted region indicates that biological systems must modulate -DGŚ and l simultaneously to achieve fast electron transfer. Interestingly, there has been some suggestion that this inverted effect might be put to use in a situation where one wants a reaction to be inefficient. The inverted effect may help limit the rate of the non-productive charge recombination in the photosynthetic reaction center (Moser et al. 1992).

Proteins modulate DGŚ and l in various ways. In order to conserve energy during ET, the driving force should be as small as possible so little chemical potential energy is lost in the transfer. In biological systems such as cytochrome c oxidase that mediate a series of ET reactions, the reduction potentials of successive electron acceptors are often only slightly higher than the potential of the upstream donor. To get high ET rates, -DGŚ should be almost equal to l. These two factors dictate that, to achieve high ET rates, the reorganization energy, l, should be small too.

It is often useful to partition l into inner sphere and outer sphere components and examine each of these separately. The inner sphere contributions to l are generally changes in equilibrium bond lengths and bond angles around the donor and acceptor. In systems such as plastocyanin and azurin, crystal structures of the oxidized and reduced proteins indicate the linner is small because the positions of the metal ligands change only slightly (Guss and Freeman 1983; Guss et al. 1986; Baker 1988; Shepard et al. 1990). In most cases, the largest contribution to louter is the reorientation of solvent dipoles to align with the new electric fields around D and A. Many ET proteins minimize louter by burying the electron transfer centers under a layer of low dielectric protein matrix.

For a series of reactions differing only in the driving force of the reaction, -DGŚ, as the driving force increases (becomes more negative) and approaches l, the activation energy, DG*, drops to zero. As the ET reaction becomes activationless, it is controlled only by factors which determine the frequency of electron transfer once the nuclei have reached the required intermediate configuration. In the adiabatic limit, this transmission coefficient is 1; electrons are transferred every time the nuclear configuration is correct. In nonadiabatic situations, such as long-range ET in biological systems, the transmission coefficient is determined by the electronic coupling between the donor and acceptor, (HAB)2. This electronic coupling, the overlap between the donor and acceptor wavefunctions, is determined by the match in energies of these wavefunctions and the way they attenuate over the distance between the two sites. The way in which the protein medium affects the electronic coupling between donor and acceptor has been the subject of much theoretical and experimental work (Winkler and Gray 1992; Farid et al. 1993; Franzen et al. 1993; Bjerrum et al. 1995; Stuchebrukhov 1996).

Protein control of electronic coupling

A compilation by Dutton and coworkers (Moser et al. 1992) of intramolecular protein ET rates at a variety of driving forces in natural and model systems suggests that many of the observed ET rates are near their predicted limit (kET ≈ kmax, -∆GŚ = l). This would indicate that many biological ET reactions are controlled not by the nuclear/energetic factors but by the electronic coupling between the donor and acceptor (HAB)2. For the reactions included in his analysis (mainly ET rates for various components of the photosynthetic reaction center and a few ruthenium-labeled heme proteins) Dutton asserts that distance is the primary controlling factor in biological ET and that the system is well modeled by a simple exponential decay of the orbital overlaps with increasing distance, kET = kETŚ exp(-b(r-rŚ)), where kETŚ is the ET rate at close contact distance (~1013 s-1) and b ≈ 1.4 ╩-1. (See figure 1.4.) This value of b is similar to that measured for ET reactions in a frozen organic glass of methyltetrahydrafuran, b=1.2 ╩-1, intermediate between the b predicted for ET through a vacuum, 2.8 ╩-1, and that measured from a series of covalently coupled small molecule DA pairs, 0.7 ╩-1.

Dutton's data span 12 ╩ and eight orders of magnitude in ET rates and would seem to offer a broad range of conditions to observe long-range ET within ordered, biological systems. However, when the data set is expanded to include charge recombination rates at even longer distances within the photosynthetic reaction complex, the simple distance dependence model breaks down (Franzen et al. 1993). Using Dutton's distance decay parameter (b = 1.4 ╩-1), one would predict that, with a separation distance of 43 ╩, the charge recombination rate between the first heme of the bound cytochrome subunit and the menaquinone (CH1+ QA-) would be unobservably slow (10-6 s-1) but the measured rate is 2 s-1. Boxer and coworkers showed that the observed rate could be explained by including the ion pairs for the forward ET reactions as mediating states in a superexchange formalism describing the charge recombination reaction. Thus, even within the photosynthetic reaction center, ET rates appear to have a complex distance dependence that is sensitive to the details of the intervening medium.

Inconsistencies in the dependence of kET on distance have led to a search for alternative representations of the electronic coupling factors controlling biological ET. In the Beratan-Onuchic pathways model, the transferring electron is conceptualized as localized in the series of one electron molecular orbitals linking the donor and acceptor (Beratan et al. 1987; Onuchic and Beratan 1990). The tunneling matrix element, TDA (analogous to the electronic coupling element HAB), is the sum of the individual pathway tunneling elements, tDA, for all physical pathways between the donor and acceptor in the protein. Each individual tunneling element is the product of the bonded and nonbonded interactions along the pathway (tDA = prefactor PeCPeHPeS). Covalent bonds provide the strongest coupling between atoms in the pathway (eC = 0.6). Hydrogen bonds are thought to provide reasonable, perhaps somewhat weaker, coupling between the atoms involved. The original formulation weighted hydrogen bonds as two covalent bonds with an adjustment in the coupling if the hydrogen bond were significantly longer or shorter than average. However, measurements of ET across a hydrogen-bonded interface in a porphyrin model system suggest the coupling is better than for the same distance bridged by covalent bonds (eH = 0.51 rather then 0.62) (de Rege et al. 1995). Recent measurement of ET rates in the b-barrel protein azurin confirm that, in proteins, the electronic coupling through a hydrogen bond may be similar to that through covalent bonds (Regan et al. 1995; Langen et al. 1996).

Sometimes the through bond (covalent and hydrogen bonds) path between a donor and acceptor in a protein is excessively long, even though the two might be relatively close when the direct, through-space distance between them is measured. In this situation, even though ET through a vacuum is much slower than ET mediated by bridging molecular orbitals, it is sometimes better to include a disadvantageous through-space jump in a proposed ET pathway to 'straighten out' the meandering of the protein structure. Through-space jumps are usually included as a penalty parameter times the difference in length of the jump and the length of a normal covalent bond (eS = 0.6 exp(-1.7(r-1.4))).

Onuchic, Beratan, and coworkers have developed algorithms to find the best - most strongly electronically coupled - ET pathways in proteins (Betts et al. 1992). One of the early successes of the pathways model of biological ET was to explain the very similar ET rates within cytochrome c 's with ruthenium labels appended at different distances from the porphyrin (Wuttke et al. 1992). (See figure 5.) It has also been used to map sites in cytochrome c and azurin that are predicted to be more poorly coupled to the metal center than would be expected from the straight-line distance. In the most recent of the pathways programs, collections of closely-related (largely redundant) pathways have been consolidated and dealt with as 'tubes' of pathways that exhibit positive and negative interferences (Regan et al. 1993; Regan et al. 1995). This was done, in part, to deal explicitly with the situations such as is found in myoglobin where there is not one clearly best path; several paths give similar total couplings (Casimiro et al. 1993; Langen et al. 1996). However, the existence of multiple pathways brings up the question of whether those pathways provide alternative routes for ET (constructive interference) or are dissipating dead-ends (destructive interference). Since this interference is a quantum phenomenon, examining this requires a more detailed description of the propagation of the donor wave function by the bridge than is provided by the simple product of couplings. The use of Green's functions allows examination of these interferences without requiring a full description of system (Regan et al. 1993); Skourtis, 1994 #91.

Recent advances in computing power and numerical methods have enabled some researchers to attempt more detailed calculations of the electronic coupling within proteins. Marcus and Siddarth (Siddarth and Marcus 1993) developed an artificial intelligence method which allowed them to identify the amino acids that contribute to electronic coupling between a donor and acceptor in a protein system. This simplification of the system allowed them to use a superexchange method to calculate the coupling between donor and acceptor through these amino acids. One of the advantages of the more detailed superexchange methodology is that it allows one to calculate absolute rates and electronic couplings. In the Beratan-Onuchic model, the initial coupling of the donor (or acceptor) into the bridge is not treated explicitly so one may only compare electron transfer rates within systems of related proteins. In comparing observed and calculated ET rates for a series of cytochrome c and myoglobin mutants, Siddarth and Marcus obtained very smooth correlations between their calculated absolute rates and experimentally determined ET rates; however, for reasons that were unclear, the slopes of their plots of calculated vs. observed rates differ from each other and from the desired slope of 1.

Donor-bridge coupling

In the Beratan-Onuchic model, as stated above, the initial coupling of the donor (or acceptor) into the bridge is not treated explicitly and is usually assumed to be the same for the proteins and pathways that are being compared. The reason for this lies more in the utility of the simplification than a true conviction of its validity. In the heme proteins, cytochrome c and myoglobin, treating all links equally ignores the anisotropic nature of the porphyrin ring and the possibility of very real differences in coupling to the iron via an axial ligand rather than through the more delocalized orbitals of the conjugated porphyrin ring system. A recent paper calculating electronic coupling using a superexchange methodology explicitly addressed how big an effect one might expect to see in electron transfer reactions coupled through different Fe orbitals in the heme protein cyt c (Stuchebrukhov and Marcus 1995). Stuchebrukhov and Marcus found that the symmetry of the donor and acceptor metal ions imposed selection rules on the tunneling pathways that can be used. They predict the coupling from the Fe t2g to the porphyrin ligand orbitals will be of p symmetry; the correlation of experimental coupling with matrix elements calculated using an s orbital as the donor and acceptor states was much worse than any of the calculations using the Fe t2g orbitals. In addition, while all three t2g orbitals of the Ru label contributed approximately equally to all of the predicted ET routes, the contribution of each Fe t2g to a particular ET matrix element was highly dependent on its orientation relative to the Ru label (variations between 2 and 20 fold in individual contributions to the matrix elements were calculated).

While the anisotropy of the coupling into the cytochrome heme iron is intuitively satisfying, the anisotropic nature of couplings of b-strands to the copper center in azurin is less apparent from casual inspection of the X-ray structure. (See figure 6.) But a combination of spectroscopic evidence and self consistent field Xa (SCF-Xa) calculations (modeling the analogous blue copper center in plastocyanin) indicate that the coupling of the copper ion to its ligands is very unequal (Gerwith and Solomon 1988; Lowery and Solomon 1992). The UV-Vis absorption spectrum of type I (blue) copper centers is dominated by the unusually strong interaction of the cysteine 112 sulfur with the copper ion. The extinction coefficient (indicative of the oscillator strength of the interaction) of this 625 nm ligand to metal charge transfer (LMCT) band is ~6,000 M-1cm-1 (compared to an e of ~50 M-1cm-1 in small molecule copper complexes) (Solomon et al. 1980). The CysS-Cu2+ bond is unusually short (2.25 ╩ in WT azurin (Nar et al. 1991)) and is thought to be highly covalent in character. By contrast, the histidine ligands, which are both respectable metal ligands with bond distances of 2.03 ╩ and 2.11 ╩, contribute only a small amount (~4 %) to the ground state wave function (Gerwith and Solomon 1988). The difference in metal-ligand coupling between the Cys and His ligands was invoked to explain the rates of electron transfer to two different sites on plastocyanin. The hydrophobic patch on the surface centered near His37 mediates electron transfer to neutral and anionic complexes; ET to cationic complexes, on the other hand, mainly proceeds after their association with the acid patch centered around Tyr83. The rates of ET to complexes bound at these two sites are remarkably similar (~104-105 M-1s-1) despite the difference in the Cu to surface distances (6 ╩ for the hydrophobic patch vs. 13 ╩ for the acid patch). Using Newton's relationship between ligand covalency and the electronic coupling through that ligand (Newton 1988), Lowery et al. calculated that the difference in coupling through the strong Cys84 ligand would balance the rate enhancement expected for ET to the closer site (Lowery et al. 1993).

The fourth conserved ligand in most type I copper sites is a methionine. This methionine ligand is somewhat puzzling. Though it is found in most type I sites, the length of the Cu-Met bond (2.8 ╩ in plastocyanin (Guss et al. 1992), 3.1 ╩ in azurin (Nar et al. 1991)) and the fact that in plastocyanin its interaction with the Cu ion cannot be seen by EXAFS (Scott et al. 1982), called into question its role as a copper ligand. SCF-Xa calculations by Solomon and coworkers demonstrate that, while a weaker interaction, the methionine sulfur does interact with the Cu2+ ion in blue copper sites; they estimate the Cu-Met bond to be about 30% of a normal ligand-metal bond (Lowery and Solomon 1992). Site saturation mutagenesis has shown that methionine is not needed in order to form a blue copper site (Chang 1991), though the observation that most mutant sites are not fully occupied by Cu2+ suggest the methionine may contribute to the stability of the site (Karlsson et al. 1991).

In an attempt to directly probe the relative abilities of the Cys and Met ligands to mediate electron transfer, Langen introduced histidines at several sites on the b-strands leading from Cys112 and Met121 in P. aeruginosa azurin (Langen et al. 1995; Langen et al. 1996). Electron transfer rates to Ru(bpy)2Im labels at these sites have been analyzed by several different methods. Regan et al. modified their Green's function pathways method to explicitly include a term describing the initial coupling from the Cu into the bridge (Regan et al. 1995). Comparing the case in which all the Cu-ligand couplings are weighted equally with one in which the relative couplings match the estimates from Solomon's SCF-Xa calculations of the Cu2+ HOMO (Cys = 1.0: Met =0.3: His = 0.1: His = 0.1), he showed that both give physically reasonable estimates for the DA coupling.

The ET rate to the label at 126, at the far end of the Met b-strand, does not attenuate as sharply as one might expect given its distance and the expected weak coupling of the all covalent bond path through the Met ligand. This is readily explained if one considers the effects of multiple constructively interfering paths. An electron leaving the Cu1+ center could either leave through the weakly coupled Met ligand and follow an entirely covalent pathway to the Ru3+ acceptor or it could leave by the strongly coupled Cys sulfur, travel through part of the Cys b-strand and then cross via one of several interstrand hydrogen bonds to reach the acceptor. (See figure 7.) The number of possible interstrand crossings increases down the strand so the existence of alternative pathways would affect the ET rates to labels at the far ends of the strands more strongly to the nearby His122 label.

Stuchebrukhov and colleagues have used several different computational methods at the extended Hĺckel level of approximation to model electron transfer in azurin (Stuchebrukhov 1996; Diazadeh et al. 1997). In each case, to limit the computational expense they first 'prune' the protein to select the most important subset of amino acids on which to perform more rigorous calculations. In a direct comparison, they show that exact diagonalization of the Hamiltonian, perturbation theory, and their tunneling currents method all are in excellent agreement (Diazadeh et al. 1997). Because their tunneling currents method gives the flux of current between two atoms, one can estimate not just the net result of the coupling, but can see the effects of interfering currents. This enables one to determine the contribution of individual bonds to the electron transfer process and to address questions such as the importance of hydrogen bonds.

In their examination of ET in labeled azurins, they used a donor wavefunction for the blue copper center that contained a strong contribution from the Cys-S and weaker contributions from the other ligands (in accordance with Solomon's studies of the blue copper site). At both the acceptor and donor sites they saw strong circular currents. For electron transfer to a Ru(bpy)2Im label at His122 they saw the majority of the tunneling current flow from the Cu, through the Met ligand, down the peptide backbone through His122, and into a Ru t2g orbital. This is what one might intuitively expect and is substantially the same as is predicted by a pathway model. The more interesting result is for the tunneling currents to His126. For this residue, the pathways model shows paths through both the Met and Cys strands with the possibility of crossing between them via the series of backbone hydrogen bonds; the relative contributions of these pathways depends on the parameterization for hydrogen vs. covalent bonds and the choice of donor-bridge coupling through the two very different sulfur ligands. The tunneling currents model shows substantial current flux through both strands; each b-strand is a strongly coupling backbone linked strongly to the metal at one end and weakly to the metal at the other. But the Met strand, despite its weak ligation to the Cu1+ donor, carries three times as much current as the Cys strand. Most interestingly, the flux is in opposite directions along the two strands, toward the Ru acceptor along the Met strand and away from it via the Cys strand. This is an example of destructive interference in which the existence of the coupling through the Cys strand decreases the overall rate of electron transfer. In contrast with the pathways model, the hydrogen bonds between the two strands carry no flux. At least in this system with parallel covalent ET routes, hydrogen bonds are not major contributors to electron transfer.

If destructive interference between the two ligand containing strands in azurin explains the slower rates of electron transfer down the Met strand, would changing the nature of the Met ligand so it could compete more effectively for flux out of the Cu1+ center increase overall rate of electron transfer? Site-saturation mutagenesis has replaced the M121 ligand with all 19 other natural amino acids (Chang et al. 1991; Karlsson et al. 1991). Not only are all of these substitutions possible, but they make remarkably little difference in the properties of the type I center. The UV-Vis spectra of most of the M121X mutants of the Pseudomonas aeruginosa azurin remain remarkably similar to that of wild type azurin with the major difference being increases in a peak near 420 nm. This peak is not seen in the WT azurin but has been observed in the spectra of other blue copper proteins and has been discussed as a possible indication of distortion in a type I site (Lu et al. 1993). The spectra of two of the mutants, M121E and M121K, show an interesting pH dependence in the relative heights of these two peaks. The absorption spectra of the M121E mutant at low and high pH are shown in figure 8. At low pH, the normal absorbance around 600 nm dominates, but at high pH, when the ligand would be presumed to be deprotonated, this peak diminishes in intensity and shifts to the blue (lmax = 570 nm) while the 420 nm peak increases in intensity. Interpretation of this spectroscopic change as the addition of an axial ligand is supported by EXAFS studies of the M121E mutant which show an additional oxygen at 1.9 ╩ at pH 8.0 (Strange et al. 1996). (For comparison, at pH 4.0, there is not fifth ligand in the M121E coordination sphere; the WT center shows a sulfur at 3.04 ╩.) (See figure 9.)

This thesis sets out to examine how the changes in the azurin blue copper center evidenced by these spectroscopic changes affect the functional properties of the M121E azurin mutant. As described in the next chapter, ruthenium labels (Ru(II)(bpy)2(Im)) were placed either at the naturally occurring H83 or at an introduced histidine at position 122 and ET rates to these labels were measured at high and low pH using photoinduced and flash/quench laser techniques. Chapter 3 discusses the rates obtained and their interpretation within the context of the spectroscopy and function of blue copper centers. Chapter 4 discusses a related project, initial attempts to characterize the electron transfer dynamics of the CuA center from cytochrome c oxidase using similar ruthenium labeling and transient-absorbance laser spectroscopy techniques to obtain intramolecular electron transfer rates.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.1 Energy diagram for a self-exchange electron transfer reaction. The horizontal axis is an abstract reaction coordinate that represents the changes in position of the nuclei in the system during the course of the electron transfer. The vertical axis is the free energy of the system. The parabola on the left represents the energy of the reactants (D-A+), the parabola on the right, the energy of the products (D+-A). The reorganization energy, l, is the energy required for an instantaneous electron transfer to give D+-A product but with the bond distances and bond angles that are the equilibrium configuration for the reactants.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.2 As in Figure 1, the energy is depicted on the vertical axis, while the horizontal axis represents the abstract reaction coordinate. In the three diagrams, the increasing driving force for the reaction is shown by the product curve shifting downward.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.3 Marcus parabola showing the relationship between driving force (-∆GŚ) and the log of the electron transfer rate.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.4 Plot of the maximal rate of electron transfer vs. the edge to edge distance between donor and acceptor for electron transfer reactions within the photosynthetic reaction center (open circles) and in select protein model systems (filled triangles). The solid line is the expected rate of electron transfer for a system with a close-contact electron transfer rate of 1013 s-1 and a uniform rate decay, b=1.4 ╩-1. Data from (Moser et al. 1992).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.5 The maximal rates of electron transfer for a series of ruthenium-modified cytochrome c proteins are plotted against (a) the edge-to-edge distance between the metal centers and (b) the s-tunneling lengths predicted by the Beratan-Onuchic pathways model (Wuttke et al. 1992).

In (a) the solid line denotes the best fit to the data, which gives b=0.66 ╩-1 and a close contact ET rate of 1.6 x 108 s-1. The dotted line is the expected change in rate with distance if the cytochromes behaved like the steroid compounds of Closs and Miller (Closs and Miller 1988), b=1.0 ╩-1 and a close contact ET rate of 3 x 1012 s-1 at 3 ╩ separation. The dashed line is the distance dependence found by Dutton and coworkers for other protein ET systems (Moser et al. 1992), b=1.4 ╩-1 and a close contact ET rate of 1 x 1013 s-1 at 3 ╩ separation.

In (b) the log of the ET rate is plotted vs. the tunneling length as predicted by the Beratan/Onuchic model; the through bond coupling was converted to ╩ using an average bond length of 1.4 ╩. The solid line represents the best fit to the data (including the estimate of a close contact ET rate of 3 x 1012 s-1 at 3 ╩ separation). The kmax falls off at a rate of 0.71 ╩-1.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.6 The type I copper center of Pseudomonas aeruginosa azurin (Nar et al. 1991). The copper ion is ligated by a trio of in-plane ligands, C112 (to the left and behind the Cu ion in this view), H117 (left front), and H46 (right). M121 (above) and the carbonyl oxygen of G45 (below and to the right) provide weaker axial interactions. The Cu ion is dark blue; the rest of the heavy atoms are colored according to the standard CPK scheme: carbon: gray, nitrogen: blue, oxygen: red, sulfur: yellow.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.7 The b-strands leading away from the azurin copper center via the Cys112 and Met121 ligands are shown as their peptide backbone; the 121-126 strand is on the left, the 112-107 strand on the right. The 5 hydrogen bonds linking the two strands are shown in white.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.8 UV-Vis spectra of M121E at pH 4.5 and pH 8.0.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.9 Comparison of Cu-ligand distances for WT and M121E Pseudomonas aeruginosa azurin obtained by EXAFS and X-ray crystallography (Nar et al. 1991; Murphy et al. 1993; Strange et al. 1996; Karlsson et al. 1997).

 

 

 

 

 

 

 

 

 

 

 

 

 

   

EXAFS

   

X-ray

 
 

WT

M121E

WT

M121E

 

8.0

4.0

8.0

5.5

9.0

6.0

Residue

           

Cys 112 S

2.12

2.15

2.21

2.25

2.26

2.11

His 117 N

1.93

1.93

2.00

2.03

2.04

2.02

His 46 N

1.93

1.93

2.00

2.11

2.09

2.02

Gly 45 C=O

2.79

2.86

2.86

2.97

2.95

3.42

Met 121 S

or Glu 121 O

3.04

-

1.90

3.15

3.12

2.21

 

 

 

References

Baker, E. N. (1988). ˘Structure of azurin from Alcaligenes denitrificans. Refinement at 1.8 ╩ resolution and comparison of the two crystalographically independent molecules.÷ J. Mol. Biol. 203: 1071-1095.

Beratan, D. N., Betts, J. N. and Onuchic, J. N. (1991). ˘Protein electron transfer rates set by the bridging secondary and tertiary structure.÷ Science 252: 1285-1288.

Beratan, D. N., Onuchic, J. N. and Hopfield, J. J. (1987). ˘Electron tunneling through covalent and noncovalent pathways in proteins.÷ J. Chem. Phys. 86(8): 4488-4498.

Betts, J. N., Beratan, D. N. and Onuchic, J. N. (1992). ˘Mapping electron-tunneling pathways - an algorithm that finds the minimum length maximum coupling pathway between electron-donors and acceptors in proteins.÷ J. Am. Chem. Soc. 114: 4043-4046.

Bjerrum, M. J., Casimiro, D. R., Chang, I.-J., Di Bilio, A., Gray, H. B., Hill, M. G., Langen, R., Mines, G. A., Skov, L. K., Winkler, J. R. and Wuttke, D. S. (1995). ˘Electron transfer in ruthenium-modified proteins.÷ J. Bioenerg. Biomemb. 27(3): 295-302.

Casimiro, D. R., Wong, L. L., Colon, J. L., Zewert, T. E., Richards, J. H., Chang, I. J., Winkler, J. R. and Gray, H. B. (1993). ˘Electron transfer in ruthenium zinc porphyrin derivatives of recombinant human myoglobins - analysis of tunneling pathways in myoglobin and cytochrome cJ. Am. Chem. Soc. 115: 1485-1489.

Chang, T. K., Iverson, S. A., Rodrigues, C. G., Kiser, C. N., Lew, A. Y. C., Germanas, J. P. and Richards, J. H. (1991). ˘Gene synthesis, expression, and mutagenesis of the blue copper proteins azurin and plastocyanin.÷ Proc. Natl. Acad. Sci. USA 88: 1325-1329.

Chang, T. K.-Y. (1991). Gene Synthesis, Expression, and Mutagenesis of Azurin. California Institute of Technology.

Closs, G. L. and Miller, J. R. (1988). ˘Intramolecular long-distance electron transfer in organic molecules.÷ Science 240: 440-446.

de Rege, P. J. F., Williams, S. A. and Therien, M. J. (1995). ˘Direct evaluation of electronic coupling mediated by hydrogen bonds: implications for biological electron transfer.÷ Science 269: 1409-1413.

Diazadeh, I., Gehlen, J. N. and Stuchebrukhov, A. A. (1997). ˘Calculation of electronic tunneling matrix element in proteins: comparison of exact and approximate methods for Ru-modified azurin.÷ J. Chem. Phys. 106(13): 5658-5666.

Evenson, J. W. and Karplus, M. (1993). ˘Effective coupling in biological electron transfer: exponential or complex distance dependence?÷ Science 262: 1247-1249.

Farid, R. S., Moser, C. C. and Dutton, P. L. (1993). ˘Electron transfer in proteins.÷ Curr. Biol. 3: 225-233.

Franzen, S., Glodstein, R. F. and Boxer, S. G. (1993). ˘Distance dependence of electron-transfer reactions in organized systems: the role of superexchange and non-Condon effects in photosynthetic reaction centers.÷ J. Phys. Chem. 97: 3040-3053.

Friesner, R. A. (1994). ˘Comparison of theory and experiment for electron transfers in proteins: where's the beef?÷ Curr. Biol. 4: 339-343.

Gerwith, A. A. and Solomon, E. I. (1988). ˘Electronic structure of plastocyanin: excited state spectral features.÷ J. Am. Chem. Soc. 110: 3811-3819.

Guss, J. M., Bartunik, H. D. and Freeman, H. C. (1992). ˘Accuracy and precision in protein structure analysis: restrained least-squares refinement of the structure of poplar plastocyanin at 1.33 ╩ resolution.÷ Acta. Cryst. B48: 790-811.

Guss, J. M. and Freeman, H. C. (1983). ˘Structure of oxidized poplar plastocyanin at 1.6 ╩ resolution.÷ J. Mol. Biol. 169: 521.

Guss, J. M., Harrowell, P. R., Murata, M., Norris, V. A. and Freeman, H. C. (1986). ˘Crystal structure analysis of reduced poplar plastocyanin at six pH values.÷ J. Mol. Biol. 192: 361.

Karlsson, B. G., Nordling, M., Pascher, T., Tsai, L.-C., Sj¸lin, L. and Lundberg, L. G. (1991). ˘Cassette mutagenesis of Met121 in azurin from Pseudomonas aeruginosaProtien Eng. 4(3): 343-349.

Karlsson, B. G., Tsai, L.-C., Nar, H., Sanders-Loehr, J., Bonander, N., Langer, V. and Sj¸lin, L. (1997). ˘X-ray structure determination and characterization of the Pseudomonas aeruginosa azurin mutant Met121Glu.÷ Biochem. 36: 4089-4095.

Langen, R., Chang, I.-J., Germanas, J. P., Richards, J. H., Winkler, J. R. and Gray, H. B. (1995). ˘Electron tunneling in proteins: coupling through a b-strand.÷ Science 268: 1733-1735.

Langen, R., Colon, J. L., Casimiro, D. R., Karpishin, T. B., Winkler, J. R. and Gray, H. B. (1996). ˘Electronic tunneling in proteins: role of the intervening medium.÷ JBIC 1: 221-225.

Lowery, M. D., Guckert, J. A., Gebhard, M. S. and Solomon, E. I. (1993). ˘Active-site electronic structure contributions to electron-transfer pathways in rubredoxin and plastocyanin: direct vs. superexchange.÷ J. Am. Chem. Soc. 115: 3012-3013.

Lowery, M. D. and Solomon, E. I. (1992). ˘Axial ligand bonding in blue copper proteins.÷ Inorg. Chim. Acta 198-200: 233-243.

Lu, Y., Roe, J. A., Gralla, E. B. and Valentine, J. S. (1993). Metalloprotein ligand redesign: characterization of copper-cysteinate proteins derived from yeast copper-zinc superoxide dismutase. Bioinorganic Chemistry of Copper. New York, Chapman & Hall. 64-77.

Marcus, R. A. (1993). ˘Electron transfer reactions in chemistry: theory and experiment (Nobel lecture).÷ Angew. Chem. Int. Ed. Engl. 32: 1111-1121.

Marcus, R. A. and Sutin, N. (1985). ˘Electron transfers in chemistry and biology.÷ Biochim. Biophys. Acta 811: 265-322.

Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S. and Dutton, P. L. (1992). ˘Nature of biological electron transfer.÷ Nature 355: 796-802.

Murphy, L. M., Strange, R. W., Karlsson, B. G., Lundberg, L. G., Pascher, T., Reinhammer, B. and Hasnian, S. S. (1993). ˘Structural characterization of azurin from Pseudomonas aeruginosa and some of its methionine-121 mutants.÷ Biochem. 32: 1965-1975.

Nar, H., Messerschmidt, A., Huber, R., van de Kamp, M. and Canters, G. W. (1991). ˘Crystal structure analysis of oxidized Pseudomonas aeruginosa azurin at pH 5.5 and pH 9.0. A pH-induced conformational transition involves a peptide bond flip.÷ J. Mol. Biol. 221: 765-772.

Newton, M. D. (1988). ˘Electronic-structure analysis of electron-transfer matrix-elements for transition-metal redox pairs.÷ J. Phys. Chem. 92: 3049-3056.

Onuchic, J. N. and Beratan, D. N. (1990). ˘A predictive theoretical-model for electron-tunneling pathways in proteins.÷ J. Chem. Phys. 92: 722-733.

Regan, J. J., Di Bilio, A. J., Langen, R., Skov, L. K., Winkler, J. R., Gray, H. B. and Onuchic, J. N. (1995). ˘Electron tunneling in azurin: the coupling across a b-sheet.÷ Chem. & Biol. 2: 489-496.

Regan, J. J., Risser, S. M., Beratan, D. N. and Onuchic, J. N. (1993). ˘Protein electron-transport - single versus multiple pathways.÷ J. Phys. Chem. 97(50): 13083-13088.

Scott, R. A., Hahn, J. E., Doniach, S., Freeman, H. C. and Hodgson, K. O. (1982). ˘Polarized X-ray absorption spectra of oriented plastocyanin single crystals. Investigation of methionine-copper coordination.÷ J. Am. Chem. Soc. 104: 5364-5369.

Shepard, W. E. B., Anderson, B. F., Lewandoski, D. A., Norris, G. E. and Baker, E. N. (1990). ˘Copper coordination geometry in azurin undergoes minimal change on reduction of copper(II) to copper (I).÷ J. Am. Chem. Soc. 112: 7817-7819.

Siddarth, P. and Marcus, R. A. (1993). ˘Correlation between theory and experiment in electron-transfer reactions in proteins: electronic couplings in modified cytochrome c and myoglobin derivatives.÷ J. Phys. Chem. 97: 13078-13082.

Solomon, E. I., Hare, J. W., Dooley, D. M., Dawson, J. H., Stephens, P. J. and Gray, H. B. (1980). ˘Spectroscopic studies of stellacyanin, plastocyanin, and azurin. Electronic structure of the blue copper sites.÷ J. Am. Chem. Soc. 102: 168-178.

Strange, R. W., Murphy, L. M., Karlsson, B. G., Reinhammar, B. and Hasnian, S. S. (1996). ˘Effect of pH and ligand binding on the structure of the Cu site of the Met121Glu mutant of azurin from Pseudomonas aeruginosaBiochem. 35: 16391-16398.

Stuchebrukhov, A. A. (1996). ˘Tunneling currents in electron transfer reaction in proteins. II. Calculation of electronic superexchange matrix element and tunneling currents using nonorthoganal basis sets.÷ J. Chem. Phys. 105(24): 10819-10829.

Stuchebrukhov, A. A. and Marcus, R. A. (1995). ˘Theoretical study of electron transfer in ferrocytochromes.÷ J. Phys. Chem. 99: 7581-7590.

Winkler, J. R. and Gray, H. B. (1992). ˘Electron transfer in ruthenium-modified proteins.÷ Chem. Rev. 92: 369-379.

Wuttke, D. S., Bjerrum, M. J., Winkler, J. R. and Gray, H. B. (1992). ˘Electron-tunneling pathways in cytochrome-cScience 256: 1007-1009.