nitrite reductase
Wijma, Hein Jakob
Citation
Wijma, H. J. (2006, February 9). Catalytic mechanism and protein engineering of
copper-containing nitrite reductase. Retrieved from https://hdl.handle.net/1887/4302
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/1887/4302
Copper
i
n
Bi
ol
ogi
cal
El
ectron
Transfer
1.1 Electron transfer in nature and this thesis
In nature,redox reactions are essentialparts of photosynthesis,respiration,and the production of secondary metabolites. M any of the enzymes that catalyze these reactions employ metal ions that have multiple stable redox-states (iron, copper,nickel, manganese, cobalt, molybdenum, and tungsten). Following iron, copper is the second most common metalin redox enzymes.Furthermore,while allthe mentioned metals catalyze reactions in which both electrons and atoms are transferred,only iron and copper are known to facilitate pure electron transfer reactions. Copper-containing nitrite reductase (NiR) contains both functions as itbinds copper in an electron transferring type-1 site and in a catalytic type-2 site,making itsuitable as a modelsystem for both.
In general, the study of enzymes has resulted in many medical and technological applications, and it is unlikely that we have reached the end of this development. This in itself would be enough already to justify further research.Furthermore, the engineering of enzymes can introduce novel functions into enzymes. NiR is relevant from medical and environmental perspectives while related proteins have similar importance and some of them find commercialapplication (see Section 2.1).The goalof this thesis is to investigate catalysis and electron transfer in copper-containing nitrite reductase and to see how these functions can be altered by protein engineering.
This chapter proceeds with an introduction to type-1 copper sites and the related CuAsites.Allcopper electron transfer sites belong to one of these two classes and CuAsites
are essentially dinuclear versions of the mononuclear type-1 sites. Therefore, any investigation of a CuA site is relevant to the understanding of type-1 copper and the other
Figure 1: Type-1 copper site variation and the CuAsite
The identity of the ligands is indicated in panel A except when denoted otherwise. Depicted are the type-1 Sites of (A) laccase from Trametes versicolor (1), 1GYC; (B) plastocyanin from Dryopteris crassirhozoma (2) 1KDJ; (C) nitrite reductase from Alcaligenes faecalis S-6 (3) 1SJM; (D) umecyanin (stellacyanin group) from horseradish (4) 1X9R; (E) azurin from Pseudomonas aeruginosa (5), 1JVL; and (F) the CuA site of nitrous oxide reductase from
1.2 Structure and Spectral Properties of Type-1 and CuA sites
All type-1 copper sites (Figure 1A-E) contain a single copper atom bound by 2 histidine and one cysteine ligand; these ligands are absolutely conserved. Additionally, weaker ligands can be present like methionine (Figure 1B/C), glutamine (Figure 1D), or both a methionine and a carbonyl oxygen from the backbone (Figure 1E). The geometry of type-1 sites is variable, ranging from trigonal planer (Figure 1A/E), via trigonal pyramidal (Figure 1B/E), to distorted tetrahedral (Figure 1C). It is under debate whether the various degrees of distortion contribute to the functional properties of type-1 copper sites (7-9). Type-1 copper sites can store and transfer a single electron, the stable redox states of the copper atom being Cu1+ and Cu2+.
CuA sites are the dinuclear equivalents of the type-1 sites (Figure 1F). The ligands
in a CuA site (2 His, 2 Cys, 1 Met, one backbone carbonyl oxygen) are the same as found in
type-1 sites. As yet, no variation in the ligand sphere of CuA sites has been reported other
than that the backbone oxygen may derive from different residues (a tryptophan in nitrous oxide reductase, a glutamate in cyt c oxidase). The stable redox states of a CuA site are the
doubly reduced [Cu1+Cu1+] and the mixed-valence state [Cu1.5+Cu1.5+].
Type-1 and CuA sites have spectral properties in the UV/visible and EPR region
that are of value to the enzymologist, as these properties can be used to discern oxidized from reduced enzyme and to monitor the binding of ligands to or near the copper. As no CuA containing protein is used in this thesis research, there is no need to discuss the
wavelength (nm) 300 400 500 600 700 800 900 E x ti n c ti o n c o e ff ic ie n t (M -1 c m -1) 0 2000 4000 6000 8000 260 280 300 320 340 0 1000 2000 3000 4000 Oxidized Reduced Reduced - Oxidized
Figure 2: Spectra of oxidized and reduced pseudoazurin
In the main panel, the thick line is of the spectrum of oxidized pseudoazurin and the thin line is the spectrum of reduced pseudoazurin. The inset shows the difference spectrum (same units on the axis).
The reduced type-1 site has a strong absorption band around 280 nm. In Figure 2 the UV/visible spectrum of both oxidized and reduced pseudoazurin is shown with the difference spectrum in the inset. The difference spectrum for pseudoazurin is similar in shape and intensity to the difference spectrum for other type-1 site containing proteins (12-15) and is similar to that of Cu1+-thiolate complexes (16). Thus, the reduced type-1 site has significant more absorption in the UV-region than the oxidized type-1 site and no absorption in the visible region.
B (Tesla)
2.4 2.6 2.8 3.0 3.2 3.4 3.6
Type-1 Type-2
Figure 3: X-band EPR spectrum of native NiR. For experimental conditions, see chapter 6.
1.3 Origin, classification, and function of blue and purple copper-containing proteins
In nature, type-1 copper sites occur both in small (≅ 14 kDa) electron transferring proteins (cupredoxins) and in larger enzymes. In the enzymes, the type-1 site facilitates electron transfer from the surface of the enzyme to one or more enzymatically active sites. CuA sites occur in membrane-located enzymes (nitrous oxide reductase, cyt c oxidase)
where they have the same function as type-1 sites. All type-1 sites and CuA sites are found
in the same cupredoxin domain (19) and may have evolved from a common ancestor that occurred approximately 3 × 109 years ago (20). A copper protein that consists of a cupredoxin domain but differs from the cupredoxins is nitrosocyanin (21). This protein contains a copper site with a ligand sphere different than the type-1 site (one H2O, two His,
Figure 4: Evolutionary tree of cupredoxins and cupredoxin domains
Distance scale is shown as a bar. AdAzurin, from Alcaligenes denitrificans; PaAzurin, from Pseudomonas aeruginosa; AxAzurin, from Alcaligenes xylosoxidans; NgAzurin, from Neisseria gonorrhoeae; Auracyanin is from Chloroflexus aurantiacus, Nitrosocyanin is from Nitrosomonas europaea; HdNiR domain is the extra domain of the NiR from
Hyphomicrobium denitrificans; Rusticyanin is from Thiobacillus ferrooxidans; RvStellacyanin is from Japanese lacquer tree Rhus vernicefera; mavicyanin is from summer squash Cucurbita pepo; cusacyanin is from cucumber Cucumis sativus and also known as basic blue protein; CsStellacyanin is the stellacyanin from cucumber; Umecyanin is from
To illustrate the functional differences between the different classes of cupredoxins, a clustal W (24, 25) based evolutionary tree has been constructed, which is presented in Figure 4. The distances in the clustal W based tree agree with the differences in properties. E.g., amongst the plastocyanins, which function in photosynthesis, the angiosperm proteins (from seed plants) are closer to green algae and to cyanabacteria than to ferns (spore bearing vascular plants). The fern plastocyanin is indeed the most unusual one in properties such as surface-charge and dependence of reduction potential on pH (26); the cyanobacterial and green algae differ less in properties from seed plants (27).
The name plastocyanin refers to the origin from the chloroplast and the blue colour. Other proteins received their names along similar lines: halocyanin, obtained from a halophilic organism (28, 29); rusticyanin, involved in iron respiration (30); sulfocyanin, involved in sulfur respiration (31); amicyanin, involved in amine oxidation (32); azurin, sky blue protein; pseudoazurin, originally misidentified as azurin (32), etc. Unfortunately, pseudoazurin is now sometimes (33) interpreted as identical in properties to azurin, which it is not. Pseudoazurin is amongst other functions involved in electron transfer in denitrification (see chapter 2), azurin is possibly involved in electron transfer to cyt c peroxidase under oxidizing stress (34) and in the induction of host cell apoptosis by pathogenic organisms (35). For the group of phytocyanins it is not clear what their function is (36, 37).
1.4 Electron Transfer Theory and Redox Proteins
One goal of this thesis is to improve our understanding of how the redox properties of the copper vary and are adapted to make the rates of electron transfer match physiological needs. With a proper understanding, it should become possible to better engineer electron transfer at a molecular level for technological purposes. In this section, the theory of electron transfer is described as far as needed to understand the work in this thesis.
Electron transfer is modelled by Marcus theory (41, 42), which is an extension of transition state theory. Marcus was the first to realize that for electron transfer to occur the reactants and products of an electron transfer reaction need to have equal energy (the electron transfer itself is too fast for energy to be exchanged with the surroundings [Franck-Condon principle]). To be able to calculate the activation energy (ǻG#) for electron transfer, Marcus grouped the potential energy of the reactants and the products into single harmonic wells (see Figure 4). The intersection point, where the potential energy of reactant and product is equal, is then the transition state and its energy can be calculated (equation 1)
ǻG# = (ǻG0+Ȝ)2/4Ȝ (1) from the driving force of the reaction (ǻG0) and the reorganization energy (Ȝ). The reorganization energy is the energy needed to displace, prior to electron transfer, all the nuclei from their lowest energy coordinates in the initial redox state to their future lowest energy coordinates. Thus, for
RED OX k OX RED
B
A
B
A
+
→
ET+
(2) reaction 2, the reorganization energy is the potential energy of the reduced species A, when its nuclei are displaced to the lowest energy coordinates of the oxidized state (ȜAREDĺOX)reaction coordinate G ib b s f re e e n e rg y reactants products ground state reactants Transition State ground state products ∆G# −∆G0 λ
Figure 5: Activation energy for an electron transfer reaction
The thin lines are the energies of reactants and products as indicated in the figure. The thick line is the reaction path from the ground state of the reactants to the ground state of the products. Indicated in the diagram is the driving force ǻG0, Ȝ, and ǻG# (height equals the intersection point versus the ground state of the products). The figure is explained further in the text.
The rate of the electron transfer reaction depends on the fraction of reactants of which the potential energy equals the activation energy. This fraction depends on absolute temperature (T) and the gas constant (R) according to equation 3.
(
)
¸¸ ¸ ¹ · ¨¨ ¨ © § ∆ + − RT G ETe
RT
k
λ λπ λ
4 2 04
1
~
(3)resulting in slower electron transfer due to a higher driving force (Marcus “inverted region”, see Figure 6). The existence of this Marcus inverted region has been verified experimentally (44). Thus, driving force and reorganization energy influence the rate of electron transfer in a predictable manner.
The reorganization energy is often quoted in eV (1 eV = 96.485 kJ mol-1) so that it can be directly compared to the driving force expressed as the difference in midpoint potential between the species. Reported values of Ȝ for one redox species are, by convention, those of the reaction of this redox species with itself (equation 4).
RED OX k OX RED
A
A
A
A
+
→
ET+
(4) The reorganization energy of a cross-reaction (ȜAB) is the average of the reorganizationenergies of the species involved (ȜAB = 0.5×ȜAA + 0.5×ȜBB). For type-1 copper centers, the
reorganization energy is often far larger than the driving force of the reactions, and strongly influences the rate of electron transfer.
-∆G0 (V) 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 kET ( s -1) 100 101 102 103 104 105 106 107 108 109 0.4 eV 0.8 eV 1.2 eV
Figure 6: Electron transfer rate versus driving force and reorganization energy
The rates are calculated from equation 6, using a H2AB of 0.048 which is a good estimate for
Bond changes contribute to the reorganization energy in the protein. This contribution is known as Ȝi (inner sphere reorganization energy). The inner sphere
reorganization can be calculated (equation 5)
¦
∆
=
j j j iF
R
25
.
0
λ
(5)with knowledge of the force-constants (F) of all (j) the bonds between the nuclei involved and the difference (ǻR) between their equilibrium positions in the oxidized and reduced state (42). Importantly, equation 5 shows that a protein can lower the Ȝi by lowering a) the
force constants and by b) limiting the change of equilibrium positions.
Furthermore, upon electron transfer the polar solvent around the protein also reorganizes due to the change of charge on the redox cofactor (41). This is known as the outer sphere reorganization energy (Ȝo). To minimize Ȝo, many redox cofactors are shielded
from the polar solvent by burying the cofactor inside the protein. It is possible that upon complex formation between redox proteins the solvent is shielded even more, resulting in lower reorganization energy than when they are free in solution (47).
The electron transfer occurs by tunneling, which is described by quantum mechanics. The rate of the tunneling is determined by the electronic coupling (H2
AB)
between the reactants electron wave functions with that of the products electron wave functions (equation 6, in which
!
is Planck’s constant).2
2
~
AB ETH
k
!
π
(6) Thus, the electron transfer rate (kET) is predicted by equation 7.(
)
¸¸ ¸ ¹ · ¨¨ ¨ © § ∆ + −=
RT G AB ETe
RT
H
k
λ λπλ
π
2 4 2 04
2
!
(7)When the protein matrix in between the redox centers is treated as an amorphous material (48, 49), equation 7 can be replaced by the semi-empirical equation 8,
Log10 (kET)= 13 – 0.6D -3.1 (ǻG0+ Ȝ)2/Ȝ (8)
in which D is the distance between the redox centers. As shown by equation 8, the rate of electron transfer decreases approximately exponentially with distance. In a more detailed model, H2
AB is calculated for a pathway of electron transfer in which every bond,
hydrogen-bond, or through space jump is assigned a semi-empirical decay factor (50). As an example, the calculated rate of electron transfer for NiR with the electronic overlap as calculated by (45, 46) is shown in Figure 6. In general, the predictions both from equation 8 and from the pathway model indicate that for electron transfer to occur at physiological rates (≅ 103 s-1),
Up to here, we assumed that the electron transfer was intramolecular. When the electron transfer reaction is intermolecular, then one must take into account that the species need to form a complex prior to electron transfer. For the work in thesis, it is relevant that the second order rate constant for the reaction of A with B (reaction 9),
RED OX k OX RED
B
A
B
A
+
→
AB+
(9)is related (41) to driving force via the “cross-relation” (equation 10). AB AB BB AA AB
k
k
K
f
k
=
(10)In the cross-relation, the factor fAB typically approximates unity (41), kAA and kBB are the
second-order rate constants for the electron self exchange reaction, and KAB is related to the
driving force for the reaction according to equation 11. ¸ ¸ ¹ · ¨ ¨ © § ∆−
=
RT G ABe
K
0 (11) In chapter 8, from equations 7, 10, and 11, a new equation was derived that relates the difference in Ȝ between two nearly identical protein variants to their rate of reaction with synthetic electron donors.1.5 Outline to this Thesis
The goals of this PhD research were to investigate the catalytic mechanism of NiR (chapter 3-5) and to explore whether it was feasible to modify the type-1 site with external ligands or allosteric effectors (chapter 6-8) in such a way that its redox function was not compromised. Chapter 2 provides a critical review on NiR. The approach to investigate the catalytic mechanism was in the order that Johnson (53) suggested (first equilibrium, then steady-state, and finally pre steady-state experiments). First the equilibrium constant for the reaction with nitrite and the physiological electron donor pseudoazurin, and the rates of the forward and reverse reactions were studied (chapter 3). Unexpectedly, at physiological pH the equilibrium favours the nitrite rather than the nitric oxide while above pH 7.5 the reverse reaction is faster than the forward reaction. Subsequently, steady-state techniques were used to determine the order of substrate binding and active site reduction (chapter 4). It turned out that nitrite can bind both to the oxidized and to the reduced type-2 site, for a different view see (54). Finally, the known (54) reduced inactive state of NiR was studied by pre-steady state voltammetry (chapter 5), and it was shown that it in slow equilibrium with a reduced active state.
The first step in engineering the type-1 site of NiR was to remove the C-terminal histidine ligand by site-directed mutagenesis (resulting in the H145G/A variants) and to investigate whether imidazole like ligands could take over the function of the histidine. The imidazole ligands restored the electron transfer to the oxidized type-1 site (chapter 6), which they do not fully do in the analogous azurin H117G variant (55). However, once reduced the type-1 site did not bind an external ligand, resulting in a conformation with a high reduction potential that was unable to donate electrons to the type-2 site.
