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
Chapter
7
A
Rearrangi
ng
Li
gand
Enabl
es
Al
l
osteri
c
Control
of
Catal
yti
c
Acti
vi
ty
i
n
Copper-Contai
ni
ng
Ni
tri
te
Reductase
This chapter is submitted to J.Mol.Biol.,Hein J.W ijma,Iain MacPherson,Maxime Alexandre, Rutger E.M.Diederix,Gerard W .Canters,MichaelE.P.Murphy,and Martin Ph.Verbeet
Abstract
In Cu-containing nitrite reductase from Alcaligenes faecalis S-6 the axial methionine ligand of the type-1 site was replaced (M150G) to make the copper atom accessible to external ligands that might affect the enzyme’s catalytic activity. The type-1 site optical spectrum of M150G (A460/A600 = 0.71) differs significantly from that of the
native nitrite reductase (A460/A600= 1.3).The midpointpotentialof the type-1 site of nitrite
reductase M150G (EM = 312 ± 5 mV versus hydrogen) is higher than that of the native
enzyme (EM= 213 ± 5 mV).M150G has a lower catalytic activity (kcat= 133 ± 6 s-1) than
the wild-type nitrite reductase (kcat = 416 ± 10 s-1). The binding of external ligands to
M150G restores spectral properties, midpoint potential (EM < 225 mV), and catalytic
activity (kcat = 374 ± 28 s-1). Also the M150H (A460/A600 = 7.7, EM = 104 ± 5 mV, kcat =
0.099 ± 0.006 s-1) and M150T (A460/A600 = 0.085, EM = 340 ± 5 mV, kcat = 126 ± 2 s-1)
Introduction
Nature often uses copper to mediate electron transfer in biological redox chains. For this purpose the copper is incorporated in a protein scaffold in a mononuclear so-called type-1 site or in a closely related dinuclear CuA site (213). These sites can be found
throughout the kingdoms of life, from archaea to humans. In photosynthesis and respiration small type-1 site containing proteins (cupredoxins) shuttle electrons between larger enzymes. In enzymes, type-1 (or CuA) sites enable electron transfer between catalytic sites
and external electron donors. These enzymes are often involved in respiration (nitrite reductase, cytochrome c oxidase) or in the conversion of metabolites (multi-copper oxidases). They also find industrial application (laccase). In this work, we show that protein engineering can introduce allosteric control over the type-1 site of copper-containing nitrite reductase (NiR) from Alcaligenes faecalis S-6.
The physiological role of NiR is the dissimilatory reduction of nitrite (NO2- + 2H+
+ e-→ NO + H2O) (56) although NiR does catalyze bidirectionally; at pH 8 the kcat of the reverse reaction is higher than the kcat of the forward reaction (chapter 3). NiR is a
homotrimer, in which each subunit contains a type-1 copper site that transfers electrons from a physiological electron donor to a type-2 copper site that is located deeper inside the enzyme (77, 78, 81). The type-2 copper forms the active site together with a water network and an Asp-His pair, that bind the nitrite and donate protons (83-86).
In a type-1 site, two histidines and one cysteine bind the copper; these three ligands are very strongly conserved. In addition, one or two weaker binding, axial ligands can be present. A methionine or a glutamine can serve as the fourth (axial) ligand and sometimes a fifth axial ligand, in the form of a backbone carbonyl oxygen from glycine, can bind on the opposite side (214). The two-histidines/one-cysteine ligand set results in unique spectroscopic properties of the oxidized type-1 site (Cu2+; the Cu1+ state is spectroscopically silent). All characterized type-1 copper sites have a unique small hyperfine splitting in their EPR spectra (214). Furthermore, strong absorption bands at approximately 600 nm and often also around 460 nm result in a blue or green color, depending mostly on the binding geometry of the weaker axial ligands. In this manuscript, we will refer to these absorption bands as 460 and 600 nm bands also when they are slightly shifted.
copper site in azurin (55, 184, 186, 189, 197, 215-217). By using the enzyme nitrite reductase, it is possible to monitor the electron transfer function which is necessary for the catalytic activity. For NiR, we earlier found (chapter 6) that when this approach was applied to the C-terminal histidine ligand, catalytic activity was lost because the midpoint potential of the type-1 site was altered too much, also in the presence of external ligands. Because the axial ligands less drastically influence the midpoint potential of the type-1 site than the equatorial ligands (102, 133, 207, 214, 218-222), we investigated whether in such an axial cavity variant the electron transfer function could be better restored by external ligands. This question was not addressed in earlier reports (187, 216, 223) about the addition of external ligands to axial cavity variants in azurin.
We replaced the original methionine in NiR by a glycine (M150G) to create space for an external ligand such as an imidazole or an alcohol. As controls for these ligands we prepared the M150H and M150T variants. Optical spectroscopy was used to monitor the binding of exogenous ligands since it is a reliable method to monitor changes in type-1 Cu site geometry (214). Furthermore, optical spectroscopy is selective for the type-1 site since the absorption peak of the type-2 site (İ790nm = 100 M-1 cm-1, chapter 6) is too weak to
influence the results. External ligands were found to bind near the engineered type-1 site and to restore the midpoint potential of this site and the catalytic activity of the enzyme. Crystal structures show that the external ligands act as allosteric effectors, their binding causes a surprising structural rearrangement in which Met62 becomes part of the type-1 copper-site.
Materials and Methods Materials
For crystallography and for activity assays a gel-filtration step was added as the last step of the purification of NiR as described in chapter 3. The Cu-content determined with bicinchoninic acid (148) was 1.9 for wt NiR, 1.7 for NiR M150T, 1.7 for NiR M150H (quoted numbers are per monomer), and 1.0 for pseudoazurin. For NiR M150G the Cu-content varied between 1.7-2.1 per batch; a batch with a Cu-Cu-content of 1.9 was used for the activity assays and for crystallization.
Spectroscopy and assays
The spectrophotometer was a Perkin Elmer Instruments Lambda 800. Prior to measuring spectra, samples were spun down at 16,000g for 10 minutes to remove small quantities (<5%) of aggregated protein that in the case of NiR can produce a scattering contribution comparable in intensity to the absorption spectrum of the type-1 site. NiR M150G (50 µM) was titrated with ligands in 50 mM MOPS pH 7.0. After correction for dilution both the increase of absorbance (A) at 460 nm, and the decrease at 600 nm were least-squares fitted assuming a single binding site (A = ANoLigand + ∆A × [L]/(Kdox+ [L], in
which L is the free ligand concentration). For all the assays in the presence of ligands, the total ligand concentration exceeded the protein concentration at least 10-fold and is therefore taken as equal to the free ligand concentration.
Activity assays were carried out by monitoring the oxidation of pseudoazurin as described in chapter 3. The concentrations of the electron donor pseudoazurin (275-325 µM) and the electron acceptor nitrite (5 mM) were saturating. The concentration of NiR was typically 1 nM. The buffer for activity assays was always 50 mM MOPS pH 7.0. Whenever using volatile compounds, the cuvette was sealed with a PTFE stopper. All reported activities were calculated from initial rates. Apparent dissociation constants (Kdapp)
were obtained from a least-squares fit of activity (v) versus ligand concentration to v =
vNoLigand + ∆v × [L]/(Kdapp+ [L]. The meaning of Kdapp will be explained in the discussion
section.
Potentiometric titrations
redox mediator while the scan range was 400-800 nm. The absorption of oxidized NiR M150H (30 µM) exceeded that of the phenazine methosulfate tenfold.
The recorded spectra were integrated using a routine written in Igor Pro (WaveMetrics Inc., see supplementary material for the routine code). For base line correction this routine approximated the scattering contributions (due to aggregated protein) either by a linear approximation or by a method described elsewhere (225). There was no need to correct for the type-2 site contribution since the absorption of the type-2 site in this part of the spectrum is 30 times lower than that of the type-1 site (chapter 6). The integrated absorbance versus potential was fitted to the Nernst equation with the number of electrons held at one. Therefore, the midpoint potential versus ligand concentration was fitted to equation 1 (226),
EM = EMNL – (RT/F)ln[Kdred×(Kdox +[L])/(Kdox×(Kdred+[L]))] (1)
in which EMNL is the midpoint potential without ligand, [L] denotes the free ligand
concentration, Kdox and Kdred are the ligand dissociation constants from the oxidized and
reduced type-1 site respectively, R is the gas constant, F is the Faraday constant and T is the absolute temperature. Because the ligand concentration far exceeded the protein concentration, [L] was set equal to the total ligand concentration. The midpoint potential of the type-1 site with the external ligand bound (EML) was calculated from equation 2.
EML= EMNL – (RT/F)ln[Kdred/Kdox] (2)
Structure determination
Met150Gly crystals were grown at room temperature by the hanging drop vapor diffusion method. The crystallization conditions were 10 mM sodium acetate pH 4.5, 2 mM zinc acetate, 2 mM cupric sulfate, 60-100 mM ammonium sulfate, and 4-10% poly(ethylene glycol) 6000. A stock protein concentration of 35 mg/ml in 10 mM Tris pH 7 was used. These conditions resulted in blue crystals that grew in an orthorhombic lattice (space group P212121). Once grown, crystals were soaked in mother liquor containing either 2 mM
dimethylsulfide (DMS) or 200 mM acetamide until they turned from blue to green indicating an alteration of the type-1 copper site. Crystals were then transferred to mother liquor supplemented with glycerol as a cryoprotectant and either DMS or acetamide. DMS-soaked crystals were looped into a cryostream (Oxford Cryo Systems) for home source diffraction studies using a MAR345 detector and Rigaku RU-300 x-ray generator. Acetamide-soaked crystals were looped and immersed in liquid nitrogen for data collection using a MAR345 detector at the Stanford Synchrotron Radiation Laboratory (beamline 7-2). Both DMS and acetamide-soaked crystals diffracted to greater than 1.8 Å resolution and diffraction data was processed with DENZO (192).
DMS and acetamide-soaked M150G crystals contain the NiR trimer in the asymmetric unit. A 1.4 Å resolution structure of nitrite-soaked wt NiR (3) was used as the starting refinement model after removal of the Met150 side-chain, nitrite and selected waters. The structures were refined using REFMAC (227) with 5-7% of the data set aside for calculation of the free R-factor. Fo – Fc difference maps were used to locate the
Table 1: Crystallographic Data Collection and Refinement Statistics Crystal M150G Dimethylsulfide M150G Acetamide cell dimensions (Å) a = 61.97 b = 103.0 c = 146.0 a = 61.40 b = 102.4 c = 146.3 resolution (Å) 1.80 (1.85-1.80)A 1.60 (1.64-1.60) r-merge 0.068 (0.292) 0.098 (0.320) {I}/{σ(I)}B 22.1 (6.43) 10.8 (3.15) Completeness (%) 86.5 (93.8) 83.0 (82.0) unique reflections 76078 (8146) 100959 (9877) working R-factor 0.166 0.177 free R-factor 0.199 0.209 rmsd bond length (Å) 0.009 0.008 overall B-factor (Å2)C 19.8 25.9 water molecules 1165 1158 PDB entry code
A Values in parenthesis are for the highest resolution shell. B
{I}/{σ(I)} is the average intensity divided by the average estimated error in intensity. C B-factors are an average from all
three monomers.
Results
Spectral Characterization and Binding of External Ligands
Purified NiR M150G appeared to the eye as blue, unlike wt NiR which is green. A UV/Vis spectrum (Figure 1A) shows that the blue color is the result of a change in the relative contributions of the absorption bands around 460 and 600 nm (NiR wt İ460 = 2900
M-1 cm-1, İ589 = 2200 M-1 cm-1; NiR M150G ε457 = 2000, ε600 = 2800 M-1 cm-1). Two
additional mutants were produced as controls, one for “strong axial interaction” (imidazole side-chain in M150H) and one for “weak axial interaction” (alcohol side-chain in M150T). For NiR M150H and NiR M150T the visible spectrum did differ significantly from the wt NiR spectrum (Figure 1B). In NiR M150H the 460 nm band has gained in absorption and is shifted to significantly shorter wavelengths, while a weak absorption is visible at 547 nm (İ439= 4600 M-1 cm-1, İ547 = 600 M-1 cm-1). For M150T almost all absorption is present in
the 600 nm band (İ460= 400 M-1 cm-1, İ602 = 4700 M-1 cm-1). As a result of the spectral
wavelength (nm ) 300 400 500 600 700 800 e x ti n c ti o n c o e ff ic ie n t (M -1 c m -1 ) 0 500 1000 1500 2000 2500 3000 wavelength (nm ) 300 400 500 600 700 800 e x ti n c ti o n c o e ff ic ie n t (M -1 c m -1 ) 0 1000 2000 3000 4000 5000 W T M 150G M 150H M 150T A B
Figure 1: Optical Spectra of native and mutant nitrite reductases.
(A) NiR wt and NiR M150G.(B) NiR M150H and M150T.Notice the different verticalscale in both panels.For experimentaldetails see materials and methods.
To study ligand binding to the oxidized type-1 site of M150G we monitored the optical spectrum upon addition of different compounds (Figure 2A). Ligands of a great variety all caused a stronger absorption at 460 nm, and a weaker absorption at 600 nm. Isosbestic points were observed, and the titration data could be fit to a single binding site, indeed (Figure 2B, Table 2). Although the used compounds included potentially strong axial ligands (e.g. imidazole/acetamide), weak axial ligands (e.g. propanol), and ligands of similar strength as a methionine (e.g. dimethylsulfide), all resulting spectra were similar
(Figure 2C) and reminiscent of wt NiR (Figure 1A). Ratios of A460 over A600 were about the
same: for example for acetamide-saturated NiR M150G ε458 = 2900, ε600 = 1800 M-1 cm-1,
and A460/A600 = 1.6 (for wt A460/A600 = 1.3) and quite different from M150H (A460/A600 =
7.7) and M150T (A460/A600 = 0.085). The only exception was the spectrum resulting from
wavelength (nm) 400 500 600 700 800 900 1000 e x ti n c ti o n (e v e ry t ic k = 1 0 0 0 M -1 c m -1 ) wavelength (nm) 300 400 500 600 700 800 e x ti n c ti o n c o e ff ic ie n t (M -1 c m -1 ) 0 500 1000 1500 2000 2500 3000 acetamide (mM) 0 200 400 600 e x ti n c ti o n c o e ff ic ie n t (M -1 c m -1 ) 1500 1750 2000 2250 2500 2750 3000 A B C dimethylsulfide propanol imidazole pyridine acetonitrile formamide
Figure 2: Titration of NiR M150G with external ligands
(A) Effect of acetamide on the optical spectrum of NiR M150G. Arrows indicate the direction of the spectral changes occurring upon subsequent additions of acetamide. (B) The absorption at 600 (open triangles) and 460 nm (filled circles) plotted versus acetamide concentration. The lines are from fits to a single binding site as described in the materials and methods. (C) Optical spectra, shifted vertically with respect to each other, of NiR
M150G with several external ligands. The concentrations were: dimethylsulfide, 151 mM (=
7 × Kd); propanol, 1940 mM (= 5.5 × Kd); imidazole, 250 mM (= 5 × Kd); pyridine, 30 mM (=
wavelength (nm)
300 400 500 600 700 800 900a
b
s
o
rp
ti
o
n
0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 time (min) 0 120 240 360 a b s o rp ti o n a t 4 3 0 n m 0.10 0.12 0.14 0.16 Table 2: Affinity constants of allosteric effectors for the type-1 site of NiR M150GExternal Ligand Kdox (mM) dimethylsulfide 21 ± 4 ethylmethylsulfide 14 ± 6 formamide 172 ± 40 acetamide 71 ± 3 imidazole 52 ± 3 ethanol 805 ± 142 propanol 353 ± 52 acetonitrile 71.5 ± 7.3 pyridine 2.6 ± 0.2 4-methylthiazole 22 ± 3 nitrite 157 ± 27
All the ligands in this Table displayed isosbestic points during titration of the type-1 spectrum. For details see materials and methods.
Figure 3. The optical spectrum of NiR M150G-imidazole versus time
Spectra of M150G were recorded every 10 minutes in the presence of imidazole (260 mM =
5 × Kd) at 25°C. Arrows indicate the direction of the spectral change. The inset shows the
For imidazole bound M150G, a further change of the optical spectrum was observed on a longer time-scale (Figure 3). After 6 hours the visible spectrum was stable, the 460 band had shifted to a significantly shorter wavelength (431 nm) and the A460/A600
ratio had increased (Figure 3). When the imidazole was removed by dialysis, the original spectrum of M150G without ligands was observed (results not shown). For other ligands (acetamide, formamide, DMS) no time dependence of the spectrum was observed, not even over a period of weeks at room temperature.
Midpoint potential
Midpoint potentials were determined to define the driving force for the electron transfer function of the type-1 sites. The midpoint potential of the type-1 site of wt NiR was found to be 213 ± 5 mV versus NHE (Figure 4). For NiR M150H the midpoint potential was extremely low with 104 ± 5 mV. The midpoint potentials of M150T (340 ± 5 mV) and M150G without ligands (312 ± 5 mV) were higher than that of the wt NiR.
Potential versus SHE (mV)
0 100 200 300 400 fr a c ti o n o x id is e d (% ) 0 20 40 60 80 100 M150H WT M150G M150T
Figure 4: Redox-titrations of NiR variants
To determine the midpoint potential of NiR M150G with external ligand bound, we measured the dependence of the midpoint potential on the ligand concentration for acetamide and pyridine. Figure 5A shows the dependence expected for a redox-site for which the binding of a ligand affects the midpoint potential (equation 1). Increasing concentrations of the external ligand lowered the midpoint potential (Figure 5B/C) and the midpoint potential levels off at the highest ligand concentrations. However, at these higher concentrations the midpoint potential of the wt NiR is significantly increased. Apparently, at the high ligand concentrations, non-specific effects come into play causing an increase in midpoint potential of wt NiR, and potentially the smaller decrease in midpoint potential of NiR M150G. Therefore, the value obtained for the midpoint potential with ligand bound (EMLsee equation 1 and 2) was interpreted as an upper-limit. With acetamide bound to NiR
M150G the fitted midpoint potential was < 225 mV versus NHE, and the Kdox was 157 ± 13
mM. For pyridine bound NiR M150G the midpoint potential was < 245 mV, and the Kdox
was 3.0 ± 0.5 mM. Thus, the midpoint potential of the type-1 site with ligand did not differ substantially from that of the wt NiR.
Figure 5: Midpoint potential of NiR versus ligand concentration
(A) The solid line is a theoretical curve calculated from equation 1 with Kdox = 1 mM, Kdred = 1
M, and EMNL = 0 mV, T = 298 K. The dashed lines equal the midpoint potential of the
redox-site without ligand (EM NL
= 0 mV), saturated with ligand (EM L
= -177 mV), and the slope in between the dissociation constants. (B) Midpoint potential of NiR M150G (open circles) and NiR wt (closed circles) versus acetamide concentration. The thick gray line indicates the midpoint potential of wt NiR without ligand. The thin line is a fit of the midpoint potential of M150G to equation 1. (C) Midpoint potential of NiR M150G and NiR wt versus pyridine concentration, legend as in panel B.
Activity
The type-1 site of nitrite reductase is essential for catalytic activity; thus, the electron transfer function of type-1 site variants can be assessed by comparison of the catalytic activity of the enzyme variant with that of the wt NiR. Catalytic activity was measured with the physiological electron donor pseudoazurin (Table 3). NiR M150H had 4 orders of magnitude less activity than the wt NiR. The catalytic activities of NiR M150T and NiR M150G without ligands were one third of that of the wt NiR.
Figure 6. Effect of external ligands on catalytic activity
Wt NiR: open circles, NiR M150G: triangles, the gray line is a visual reference to the catalytic activity of native NiR in the absence of external ligands. (A) Rate of catalytic turnover versus dimethylsulfide concentration. The thin dark line is the least-squares fit that yielded the apparent dissociation constant and the maximum activity (see Materials and Methods section for details). (B) Activity versus acetonitrile concentration.
Table 3: Catalytic activity of NiR variants NiR variant Saturated with external ligand kcatsat (s-1) Kdapp (mM) WT - 416 ± 10 -M150H - 0.099 ± 0.006 -M150T - 126 ± 2 -M150G - 133 ± 6 -M150G dimethylsulfide 373 ± 22 34 ± 7 M150G ethylmethylsulfide > 292 >73 M150G formamide > 255 >4000 M150G acetamide 374 ± 28 1068 ± 295 M150G acetonitrile > 390 A ND
For details of the activity measurements see Materials and Methods. A Also the native NiR increased in
activity (Figure 6B). A lower limit means that it was not possible to saturate the activity of NiR M150G with ligand. ND: not determined.
Table 4: Metal ligand geometry of type-1
sites in wt Nitrite Reductase and in M150G A
native AfNiR M150G DMS M150G acetamide Distances (Å) Axial – Cu 2.48 2.39 2.37 95 – Cu 2.07 2.13 2.13 136 – Cu 2.22 2.21 2.24 145 – Cu 2.06 2.07 2.04 Cu – NSN B 0.64 0.57 0.63 Angles (°) 136 – Cu – 95 129 129 126 136 – Cu – Axial 106 97 100 Axial – Cu – 95 89 95 95 Axial – Cu – 145 133 130 129 θC 64 67 71
A The numbers 95, 136, and 145 in the left column refer to the
Nδ of His95, the Sγ of Cys136, and the Nδ of His145. Axial
refers to the Sδ of Met150 in the wt NiR (1SJM) and the Sδ of
Met62 in the M150G structures. Sigma values (standard deviations determined from average values of three monomers in the asymmetric unit) amount to less than 5 % for bond angles and less than 3 % for bond distances. B This
is the distance between the Cu atom and the NSN plane determined by the ligand atoms of residues His95/Cys136/His145. C
θ is the dihedral angle between the planes through 136-Cu-Axial Ligand and the plane through 95-Cu-145.
The activity of NiR M150G could be increased by the addition of exogenous ligands (Figure 6). Acetamide and dimethylsulfide (DMS) restored the activity up to the level of wt NiR. The dependence of activity versus ligand concentration could be fit to a one ligand binding equilibrium. The resulting apparent dissociation constant was in all cases higher than the Kd for binding to the oxidized type-1 site (Table 2 and 3). For
ethylmethylsulfide and formamide it was not possible to saturate NiR M150G with ligand; however, activity doubled over a concentration range in which the wt NiR had constant activity (Table 3).
Structure
Superposition of wt NiR to the DMS and acetamide-bound structures of M150G reveals that these small molecules displace the side-chain of Met62, a residue near the type-1 copper site that is non-coordinating in the wt structure (Figure 7). The remarkable finding is that in its new position Met62 adopts a conformation that allows its Sδ atom to take up a
new position that is similar to the Met150 Sδ position in the native wt structure. DMS and
acetamide bind M150G NiR at nearly the same position, roughly 6 Å from the protein surface. The DMS sulfur is 0.46 Å from the Sδ atom of Met62 in wt NiR. The DMS is in an
orientation analogous to that of the Met62 thioether that it displaces and too far from the Cu-atom (4.5 Å) to be a ligand (Figure 7). Acetamide is slightly further away from the Cu-
Figure 7: Crystal structures of the type-1 copper sites
Identical views are given for panel A, B and C. Foreground: Ala61-Phe64, His95.
Background: Cys136, Trp144, His145, and Gly150 (mutant) or Met150 (wt). The type-1
copper is a sphere. The σA weighted 2Fo – Fc electron density maps are contoured at 1ı. (A)
atom (5.0 Å) and forms hydrogen bonds to two buried water molecules. The acetamide N and O atoms could not be unambiguously assigned. The two buried water molecules are located in a 5 Å deep tunnel that connects to the surface and also is present in the wt and M150G-DMS structures.
The displacement by DMS and acetamide of the Met62 side-chain is accomplished by a 115° rotation of the Ȥ1torsional angle, a 25° rotation of Ȥ2, and a 59° rotation of Ȥ3. The
atomic positions of the Met62 backbone shift only slightly (0.03 Å rms), but the ij torsional angle rotates 27°. As a result of all torsional changes, the Met62 sulfur moves 4.5 Å to bind to the type-1 copper at a position that overlaps that of the Met150 Sδ in wt NiR (Figure 7).
The geometries of the type-1 sites are almost identical to that of wt NiR (Table 4). No other structural perturbations were observed surrounding the type-1 copper site.
A second acetamide molecule is modeled in the active site solvent channel, 7.3 Å from the type-2 copper. In the DMS-bound structure, additional density is present at the substrate binding site of the type-2 copper. This density is modeled as water but may be DMS or a degradation product.
Discussion
Axial Ligand Binding and Spectroscopy
In the crystal structures, the external ligands dimethylsulfide and acetamide do not bind to the Cu atom, but instead they displace Met62 which is coordinated to the type-1 copper. The crystals are grown below pH 5 and data was collected at liquid nitrogen temperature, so a different conformation could prevail in solution at pH 7. This possibility could be excluded by optical spectroscopy.
For type-1 copper sites, the absorbance band at 600 nm originates from ʌ overlap between the copper dx2-y2 and the sulfur orbitals, the 460 nm band originates from pseudo-ı
overlap between the same orbitals (214). The A460/A600 ratios in blue copper proteins reveal
variations in these overlaps. In the case of a trigonal site, such as in azurins, the dx2-y2
orbital overlaps almost solely with the two histidines and the cysteine, resulting in almost pure ʌ overlap with the cysteine. In tetrahedrally distorted type-1 sites like in the nitrite reductase of Alcaligenes faecalis S-6 (78, 228), the d-orbital overlaps with the axial methionine (stronger axial interaction). This change of orientation produces an increased A460/A600 ratio (210, 211), and a shift to shorter wavelengths (214) of both absorption
bands. The change in orientation of the dx2-y2 orbital can be quantified by the dihedral angle
The effects of strong versus weak axial interaction on the optical spectrum of a type-1 copper site can be seen in two examples: NiR M150H (strong) and M150T (weak). The optical spectrum of M150H has a very high ratio of A460/A600, and peaks that are
shifted to shorter wavelengths, while M150T has a very low A460/A600 ratio. For NiR
M150G as purified the A460/A600 ratio is closer to that of the wt NiR than to M150T.
Although this could be explained since also other factors than the nature of the axial ligand (211, 228, 229) can affect the electronic properties of the type-1 site, conceivably the type-1 site of NiR M150G may partly bind another ligand (such as Met62).
Crystallography of the NiR M150G variant indicates that Met62 and not the added exogenous ligands (DMS/acetamide) bind to the type 1 copper, and optical spectroscopy confirms the crystallographic result. Binding of dimethylsulfide and ethylmethylsulfide to NiR M150G restores the spectroscopic properties to those of wt NiR, which is expected if either these thioether compounds bind directly or alternatively Met62 binds to the Cu atom. For acetonitrile, ordinary alcohols (which mimic threonine), imidazoles (which mimic histidine), acetamide (which mimics glutamine) essentially the same spectra are observed as with dimethylsulfide and ethylmethylsulfide. This result is incompatible with direct binding of these groups to the Cu-atom and rather points to similar Cu-sites in all these experiments. Crystallographic observations correlate well to the solution optical properties. Not only were the ligand-soaked crystals green, also the θ dihedral angles found in the crystal structures, which correlates with the A460/A600 ratio, are similar for the wt and the
two M150G-ligand structures (Table 4). All these results indicate that the bound compounds affect the Cu-site structure in the same indirect manner by causing Met62 to bind to the Cu. Thus, the added compounds may be considered as allosteric effectors.
Long-term incubation of NiR M150G with imidazole resulted in spectra indicative of a different axial ligand. A peak shift to shorter wavelengths, accompanied by an increase in the A460/A600 ratio, suggests stronger axial interaction due to direct copper coordination
by imidazole, similar to NiR M150H. Incubation with formamide significantly shifted the A460 peak also, indicating that formamide does bind directly, at least partially, to the type-1
Midpoint Potential and Catalytic Activity
The M150T mutation changed the midpoint potential by +127 mV with respect to the wt protein, which resembles the shift of +107 mV observed for Rhodobacter sphaeroides NiR M182T (133). For NiR M150G, the change in midpoint potential (+99 mV) resembles the change observed for Alcaligenes xylosoxidans NiR M144A (+74 mV (230)) and azurin M121A (+63 mV (222)). For NiR M150H, the shift in midpoint potential (–109 mV) is similar to the shift of –100 mV for Alcaligenes denitrificans azurin M121H (231). The observed variations in the midpoint potential are in line with the idea that stronger axial interaction lowers the midpoint potential of the type-1 site (214). The higher midpoint potential of NiR M150T and NiR M150G with respect to the wt may partly explain the lower catalytic activity since it will hinder the electron transfer to the type-2 site. In A. xylosoxidans NiR M144A, the electron transfer rate from the type-1 to type-2 Cu site is indeed tenfold decreased (45). In Achromobacter cycloclastes NiR M150Q (change – 127 mV), the electron transfer rate from pseudoazurin to NiR had decreased below the detection limit (219), which is reminiscent of the low activity of our NiR M150H (change – 109 mV). Thus, in a qualitative sense the catalytic activities of our NiR variants vary in agreement with the changes in midpoint potentials.
To determine the midpoint potential of NiR M150G with an allosteric effector bound, we tried to saturate both the oxidized and reduced type-1 sites with ligand (otherwise an average midpoint potential with and without ligand bound is measured according to equation 1). Assuming a simple scheme, the Kdox obtained from potentiometric
titration should be identical to that obtained from direct ligand titration, which for pyridine is indeed the case. The calculated EM < 225 mV versus NHE with acetamide as the
Met62 S N Cu N S Met62 S N Cu N S Met62 S N Cu N S Met62 S N Cu N S T state R state
T-Ligand State R-Ligand state M150 cavity Ligand Ligand Keq Ligand Ligand M62 cavity
Scheme 1. Allosteric control of nitrite reductase M150G
The two nitrogen atoms and one sulfur atom that coordinate the copper represent His95,
His145, and Cys136. The side-chain of Met62 is indicated. The position of empty cavities are represented by solid spheres, those of ligand-filled cavities is shown by broken circles. Met62 cavity refers to the space left after its rearrangement with respect to the native structure,Met150 cavity denotes the cavity created by deletion ofthis residue.
Allosteric Control
The structures of NiR M150G show that the allosteric effectors restore the geometry of the native type-1 site (Table 4), in agreement with the restored electron transfer activity. In all naturally existing type-1 sites, either a methionine or an aromatic residue is present at the homologous position of Met62 (232, 233). The native function of this residue is to sandwich the axial ligand together with an aromatic residue that is also conserved. Thus, the binding of the aromatic pyridine and of DMS, the latter of which appears to be bound in the same way as the methionine side-chain of Met62 is packed in the wt NiR, could be viewed as a restoration of the native situation. However, the successful restoration of catalytic activity by acetonitrile and acetamide shows that also chemical groups which do not naturally occur at this position can act as allosteric effectors.
The difference between Kdapp and Kdox (Table 2 and 3) can be explained since the
allosteric effector binds with lower affinity to the reduced type-1 site (Figure 5). Under the turnover conditions in which the Kdapp is measured, the type-1 site needs to accept electrons
from pseudoazurin and donate them to the type-2 site. If the reduced type-1 site needs to bind the external ligand for efficient electron transfer to the type-2 site, the Kdapp will be a
weighted average between Kdox and Kdred.
The only two ligands (imidazole, and formamide) that are observed in the T-ligand state are similar in that both are expected to bind Cu(II) with higher affinity than a thioether group (234). Conversely, alcohols are expected to bind weaker to Cu(II) than a thioether group, and indeed do not bind to the Cu of either azurin M121G or M121A (187). This observation suggests that some of the ligands like ethanol do not bind to the type-1 copper in NiR M150G because the thioether group of the Met62 has greater affinity for Cu(II). Furthermore, if the ligand has higher affinity for the cavity left by Met62, than for Cu(II), then also the R-state is favored over the T-state. However, considering that imidazole-bound M150G needs hours to go from R to T- state (Figure 3), at longer time-scales some of the other ligands in Table 1 could favor the T-state as well (kinetic limitation).
Supplementary material of chapter 7
The following Igor procedure processes spectra as produced by UV/W inLab Lambda 800/900
(PerkinElmer, Inc.). The spectra should be of the form NAME#01.sp, NAME#02.sp… NAME#25.sp and must have been saved as ASCII files. W ithin Igor an appropriate new path (P_NAME) containing the location of the spectra must be created. Then the command is of the form:
Load_SpectrB(“P_NAME”, “NAME#”, 1, 25). Load_SpectrB fits to a real scattering baseline. By
modifying in line16 the command “Loadwave/Q/G /A /L={2,86,291,1,1}” any spectrum (saved as
ASCII) can be loaded and processed.
function Load_SpectrB(PathName, FileName, StartNumber, EndNumber) String FileName, PathName
Variable StartNumber, EndNumber Variable index=StartNumber
String FolderName="S_"+FileName+"_"+num2istr(StartNumber)+"_"+num2istr(EndNumber) NewDataFolder $(FolderName)
SetDataFolder $(FolderName)
Make/O/N=(EndNumber-StartNumber+1) Abs_Data
Display Do if(index<10) String FileNameCurr=FileName+"0"+num2istr(index)+".sp" else FileNameCurr=FileName+num2istr(index)+".sp" endif
Loadwave/Q/G /A /L={2,86,291,1,1} /P=$(PathName) FileNameCurr
Rename wave0, $("S_"+FileName+num2istr(index)) W ave SpecNameCurr=$("S_"+FileName+num2istr(index)) SetScale/I x 800,510,"", SpecNameCurr
Duplicate/O SpecNameCurr, $("S_"+FileName+num2istr(index)+"_Res"), Abs_Value W ave SpecNameCurr_Res=$("S_"+FileName+num2istr(index)+"_Res")
Make/D/N=2/O W _Coef
W _Coef[0] = {1,1}
FuncFit/Q Scatter_real W _Coef SpecNameCurr /R=SpecNameCurr_Res
AppendToGraph SpecNameCurr, SpecNameCurr_Res Abs_Value=abs(SpecNameCurr_Res)
Variable Sum_Curr=area(Abs_Value, 510,800) Abs_Data[(index-StartNumber)]=Sum_Curr
index+=1 while(index<=EndNumber)
KillW aves/Z Abs_Value
Rename Abs_Data $(FolderName) Display $(FolderName)
End
Function Scatter_real(w,y) : FitFunc W ave w
Variable y
return w[0]*1E5/(y^2+w[1]*1E5)
