De novo-designed metallopeptides with type 2 copper centers: Modulation of reduction potentials and nitrite reductase activities

Article abstract of DOI:10.1021/ja406648n

Enzymatic reactions involving redox processes are highly sensitive to the local electrostatic environment. Despite considerable effort, the complex interactions among different influential factors in native proteins impede progress toward complete understanding of the structure-function relationship. Of particular interest is the type 2 copper center Cu(His)₃, which may act as an electron transfer center in peptidylglycine α-hydroxylating monooxygenase (PHM) or a catalytic center in copper nitrite reductase (CuNiR). A de novo design strategy is used to probe the effect of modifying charged amino acid residues around, but not directly bound to, a Cu(His)₃ center embedded in three-stranded coiled coils (TRI-H)₃ [TRI-H = Ac-G WKALEEK LKALEEK LKALEEK HKALEEK G-NH₂]. Specifically, the peptide TRI-EH (=TRI-HK22E) alters an important lysine to glutamate just above the copper binding center. With a series of TRI-EH peptides mutated below the metal center, we use a variety of spectroscopies (EPR, UV-vis, XAS) to show a direct impact on the protonation equilibria, copper binding affinities, reduction potentials, and nitrite reductase activities of these copper-peptide complexes. The potentials at a specific pH vary by 100 mV, and the nitrite reductase activities range over a factor of 4 in rates. We also observe that the affinities, potentials, and catalytic activities are strongly influenced by the pH conditions (pH 5.8-7.4). In general, Cu(II) affinities for the peptides are diminished at low pH values. The interplay among these factors can lead to a 200 mV shift in reduction potential across these peptides, which is determined by the pH-dependent affinities of copper in both oxidation states. This study illustrates the strength of de novo protein design in elucidating the influence of ionizable residues on a particular redox system, an important step toward understanding the factors that govern the properties of this metalloenzyme with a goal of eventually improving the catalytic activity.

Full text of DOI:10.1021/ja406648n

Article
pubs.acs.org/JACS
De Novo-Designed Metallopeptides with Type 2 Copper Centers:
Modulation of Reduction Potentials and Nitrite Reductase Activities
Fangting Yu, James E. Penner-Hahn, and Vincent L. Pecoraro*
Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109-1055, United States
S
* Supporting Information
ABSTRACT: Enzymatic reactions involving redox processes
are highly sensitive to the local electrostatic environment.
Despite considerable effort, the complex interactions among
different influential factors in native proteins impede progress
toward complete understanding of the structure−function
relationship. Of particular interest is the type 2 copper center
Cu(His)3, which may act as an electron transfer center in
peptidylglycine α-hydroxylating monooxygenase (PHM) or a
catalytic center in copper nitrite reductase (CuNiR). A de
novo design strategy is used to probe the effect of modifying
charged amino acid residues around, but not directly bound to, a Cu(His)3 center embedded in three-stranded coiled coils (TRI-
H)3 [TRI-H = Ac-G WKALEEK LKALEEK LKALEEK HKALEEK G-NH2]. Specifically, the peptide TRI-EH (=TRI-HK22E)
alters an important lysine to glutamate just above the copper binding center. With a series of TRI-EH peptides mutated below
the metal center, we use a variety of spectroscopies (EPR, UV−vis, XAS) to show a direct impact on the protonation equilibria,
copper binding affinities, reduction potentials, and nitrite reductase activities of these copper−peptide complexes. The potentials
at a specific pH vary by 100 mV, and the nitrite reductase activities range over a factor of 4 in rates. We also observe that the
affinities, potentials, and catalytic activities are strongly influenced by the pH conditions (pH 5.8−7.4). In general, Cu(II)
affinities for the peptides are diminished at low pH values. The interplay among these factors can lead to a 200 mV shift in
reduction potential across these peptides, which is determined by the pH-dependent affinities of copper in both oxidation states.
This study illustrates the strength of de novo protein design in elucidating the influence of ionizable residues on a particular
redox system, an important step toward understanding the factors that govern the properties of this metalloenzyme with a goal of
eventually improving the catalytic activity.
protons converts NO2 to NO and H2O.7,21−25 Thus, a
−
INTRODUCTION
■
structurally related Cu(His)3 site can be differentially utilized
on the basis of the surrounding protein environment.
Understanding the relationship between structure and
Redox-active enzymes are critical in many biological processes,
with metal cofactors accounting for many of the redox-related
reactions in native proteins. These activities include catalyzing
function has always been a major goal of biomimetic studies.
For redox-active metalloproteins, it is obvious that the ligands
coordinated to the metal center directly influence the redox
properties.26,27 However, electrostatic interactions of the redox
center with the surrounding charged groups are now
recognized to influence redox property modulation. Charged
amino acid residues [e.g., glutamate (Glu) and lysine (Lys)]
both in the interior and on the surface of the proteins can
participate in fine-tuning the potentials of redox-active metal
sites.28−31 Furthermore, enzymatic reactions can be highly
sensitive to the local electrostatic environment. Considerable
effort has been devoted to understanding the electrostatic
interactions in native proteins, but because of the complexity of
native systems, it is always challenging to disentangle the
contributions of the potentially influential factors.
important biological processes such as photosynthesis (electron
transfer and water oxidation), respiration, molecular oxygen
reduction, nitrogen fixation, and denitrification.1−5 Specifically,
redox-active copper sites exist in a large number of metal-
loproteins, which, along with iron proteins, play important roles
in electron transfer, activation and transport of dioxygen, and
the metabolism of other small molecules.6−16 Within the
metallobiosphere, copper proteins are extremely important,
with functions ranging from pure electron transfer to
multielectron redox catalysis. Type 2 copper centers comprise
a broad class of mononuclear sites that again serve both
catalytic and redox functions. Two specific enzymes are of
interest to us as they contain Cu(His)3 structures. The first,
peptidylglycine α-hydroxylating monooxygenase (PHM), has a
Cu(His)3 center (CuH) whose role is to donate an electron to
the catalytic CuM site that activates dioxygen.17−20 The second,
copper nitrite reductase (NiR), is an essential component of
dissimilatory nitrite reduction using a type 1 Cu center to
donate an electron to the Cu(His)3 active site, which along with
Received: June 30, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 1. Peptide Sequences Used in This Study
a
The color code used throughout corresponds to a change of charge compared to a single TRI-H peptide. A difference of −2 charges is given in red,
b
c
−1 in magenta, +1 in purple, and +2 in blue. The C-terminus is capped by an NH2 group and the N-terminus by an acetyl (Ac) group. Δcharge is
defined as the difference in total charge of the specified 3SCC versus (TRI-H)3 assuming all Glu are fully deprotonated and all Lys are fully
protonated. Polyglutamate or polylysine based ligands can exhibit a wide range of pKa values when multiple side chains of the same type occur in the
same peptide. Therefore, one should recognize that nonintegrally charged peptides can exist due to incomplete acid−base chemistry.
A de novo design strategy provides an opportunity to
generate biologically relevant models to understand structure−
function relationships using a minimal construct.32−35 With a
simplified polypeptide sequence, the local environment of
metal centers may be modified rationally in order to investigate
the important factors that govern the properties and functions
of a specific site. Previously, we reported a functional copper
nitrite reductase model embedded in a de novo-designed three-
stranded coiled coil (3SCC) scaffold, which represents one of
the most efficient model systems for NiR in aqueous
solutions.36 Prior to that work, there were a few cases of
Cu(His)3 sites in de novo-designed peptides with controlled
copper coordination.37−39 Although redox processes have been
demonstrated in a few systems,40,41 none were fully
characterized in both oxidation states. Our system was the
first to have both relatively well-characterized, stable Cu(I) and
Cu(II) oxidation states that exhibits NiR activity under aqueous
conditions. We observed very positive reduction potentials for
these systems, which probably reflected the stabilization of the
Cu(I) state in a trigonal-planar geometry.36 Initial designs
focused solely on the first coordination sphere of copper,
yielding modest catalysts. To design a model with higher rates,
the next logical step is to explore systematically secondary
Figure 1. PyMOL models of (1) TRI-H, (2) TRI-HK22Q, (3) TRI-
EH, (4) TRI-EHE27K, (5) TRI-EHE27Q, (6) TRI-EHK24Q, and
coordination sphere modifications around the copper. Of
course, such studies are potentially relevant to understanding
how solely electron transfer centers as found in PHM are
influenced by the same secondary coordination sphere
(7) TRI-EHK24E from the left to the right based on the crystal
structure of Zn(II)−NHg(II)−S(CSL9PenL23H)3 (PDB code 3pbj).42
modifications. We report herein a series of modified peptides
Peptide Synthesis and Purification. All of the peptides in this
work were synthesized on an Applied Biosystems 433A peptide
synthesizer using standard protocols43 and purified and characterized
as previously reported.44 The apo-peptide solution was prepared by
dissolving purified dry peptides in doubly distilled water or buffer
solutions. The peptide concentration was determined on the basis of
the tryptophan (Trp) absorbance at 280 nm (ε = 5500 M−1 cm−1).45
UV−Vis and Fluorescence Spectroscopy. UV−vis spectra were
collected in airtight quartz cuvettes at 25 °C on a Cary 100 Bio UV−
vis spectrophotometer. Emission spectra were recorded on a
FluoroMax-2 fluorimeter at 25 °C.
pKa Determination for Cu(II)−Peptide Complexes. The
deprotonation of Cu(II)−peptide complexes was reflected in the
UV−vis spectra. The UV−vis spectra of a solution containing 0.27
mM Cu−3SCC and 0.03 mM apo-3SCC [to ensure that >99% of the
Cu(II) was bound to the peptide] under various pH conditions were
collected and pKa was fitted to eq 1. The protonation equilibrium
corresponding to n protons can be described as follows:
(Table 1) based on the parent peptide TRIL2WL23H (TRI-H)
that address how the local environment, particularly charged
residues, influence the redox properties and catalytic behaviors
of these designed copper proteins. Models of Cu(I)(TRI-H)3
and the other peptides used are given in Figure 1.
EXPERIMENTAL METHODS
■
The sequences of the peptides used in this work are shown in Table 1,
with each heptad repeat given as abcdefg. Variations were made at the
b, e, and g positions of the heptad in addition to the L23H mutation.
General Procedures. The copper−peptide complexes were
prepared by adding a solution of tetra(acetonitrile)copper(I)
tetrafluoroborate or copper(II) chloride into a (buffered) solution of
apo-peptide. Since the Cu(I)−peptide complexes undergo slow
oxidation under aerobic conditions, all of the reactions involving
Cu(I) or nitrite reduction were carried out in an inert atmosphere box.
B
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
n
Ka
A sample of ∼1 mM Cu(II)(3SCC)3 was prepared in a 50 mM buffer
solution, and the pH was adjusted by adding concentrated HCl or
KOH solution to the sample. An extra 0.5 mM apo-3SCC was added
to ensure that the free Cu(II) was less than 0.1%, and 50% glycerol
was used as a glassing agent. The EPR parameters were extracted by
fitting the collected spectra to the simulated spectra using the EPR
data fitting software SpinCount.
NiR Activity Reaction Rates. The rates were determined as
previously reported for Cu(TRI-H)3.36 NaNO2 (30 mM) was mixed
with 0.18 mM Cu(II)(3SCC) and 0.09 mM apo-3SCC [to ensure that
over 99% of the Cu(II) was bound to the peptide]. The reaction was
initiated by adding 6 equiv of sodium ascorbate (NaAsc) with respect
to Cu(II). The consumption of NaAsc was monitored by UV−vis
spectroscopy. The control was a mixture of 30 mM NaNO2 and 0.09
mM apo-3SCC with the same amount of NaAsc. The rates of the
reaction were calculated as 2 times the difference in rates of NaAsc
consumption (decrease of absorbance at 251 nm) for the sample and
the control. The production of nitric oxide was examined using the
method previously reported.36
EHn XoooY nH+ + En−
The total molar extinction coefficient at a particular wavelength can be
written as
εE + εEH ·10n(pK −pH)
a
n
εtotal
=
1 + 10n(pK −pH)
a
(1)
where εE and εEH stand for the extinction coefficients of individual
n
species in different protonation states.
Determination of Binding Constants. Cu(I) Binding Constants.
The affinities of peptide trimers (pep3) to Cu(I) were determined by
competitive binding assays with disodium bathocuproindisulfonate
(Na2BCS, BCS) or disodium bacinchoninate (Na2BCA, BCA) as a
competitive chelator in 50 mM potassium phosphate buffer. The
competition equilibrium for the titration experiment can be expressed
by the following chemical equation:
Cu−pep3 + 2L ⇄ CuL2 + pep3
RESULTS
where L = BCS or BCA. Starting with 40 μM Cu(I)−pep3 and 40 μM
apo-pep3, small aliquots of a 5.00 mM BCS or BCA stock solution
were added into the system. Transfer of Cu(I) from Cu−pep3 to CuL2
was tracked from the absorbance at 483 nm for L = BCS (ε = 13 000
M−1 cm−1) or 562 nm for L = BCA (ε = 7900 M−1 cm−1).46−48
Titrations were carried out in triplicate, and the dissociation constants
were fitted using the spectrophotometric and/or potentiometric data
fitting program Hyperquad 2006 and the reported dissociation
constants for Cu(I)L2.46−48 Titrations were performed at both pH
7.4 and pH 5.8 to investigate how the affinities change in relation to
pH. At pH 5.8, the concentration of the protonated form of the ligand
BCS was taken into account.47,48
■
pKa of Cu(II)−Peptide Complexes. For the parent peptide
TRI-H, the appearance of a broad band with a maximum
absorbance at 644 nm was observed as the pH of a
Cu(II)(TRI-H)3 solution was increased from 2.92 to 5.52
(Figure 2A). This spectrum has previously been suggested to be
due to Cu(II)(His)3(OH2)1−2
above 7.6, a new set of spectra containing an isosbestic point at
571 nm was observed (Figure 2B). The absorbance at 644 nm
decreased while a new band at 514 nm grew in intensity. From
a fit of the absorbance at 644 nm to eq 1, a pKaw value
corresponding to the changes in the absorption spectra
36
.
When the pH was raised
Cu(II) Binding Constants. As reported previously, the Trp emission
at 350 nm decreases upon Cu(II) binding to the (His)3 pocket.36
A
stock solution of CuCl2 was added to ∼800 nM 3SCC in 50 mM
buffer solution. Triplicate titrations were carried out, and the data were
fitted with Hyperquad 2006.
X-ray Absorption Spectroscopy (XAS). A 1 mM Cu(I)(TRI-
EH)3 solution was prepared by adding 0.8 equiv (with respect to
peptide trimer) of tetra(acetonitrile)copper(I) tetrafluoroborate
solution to the peptide buffer solution in an inert atmosphere box.
The pH of the solution was adjusted to 8.5 using small aliquots of
concentrated KOH. Glycerol was added as a glassing agent. The
homogeneous 50% aqueous glycerol solution was then transferred to a
sample cell and frozen at liquid nitrogen temperature. Data were
collected at the Stanford Synchrotron Radiation Lightsource (SSRL)
on beamline 7-3 with a Si(220) double-crystal monochromator, a flat
Rh-coated harmonic rejection mirror, and a 30-element Ge array
detector. An Oxford Instruments liquid helium cryostat was used to
keep the samples below 10 K during data collection. Data were
collected as fluorescence excitation spectra and normalized to the
incident intensity measured with a N2-filled ion chamber. Data were
measured with steps of 0.25 eV in the X-ray absorption near-edge
structure (XANES) region (1 s integration time) and k = 0.05 to 13.5
Å−1 in the extended X-ray absorption fine structure (EXAFS) region
(1−20 s integration, k3-weighted). Energies were calibrated by
assigning the lowest-energy inflection point of a copper metal foil as
8980.3 eV. An initial E0 value of 9000 eV was used to convert data to k
space, and the background was removed using a three-region cubic
spline. The EXAFS data were analyzed using EXAFSPAK49 with ab
initio amplitude and phase parameters calculated using FEFF 9.0.50
Imidazole outer-shell scattering was modeled using phase and
amplitude parameters calculated for an idealized Cu(imid)4 structure
with Debye−Waller factors defined as suggested by Dimakis and
Bunker;51 the imidazole was refined as a rigid group with distances and
Debye−Waller factors linked to the refined Cu−N parameters, so
there were only three variable parameters (RCu−N, σ2Cu−N, and ΔE0).
Electron Paramagnetic Resonance (EPR) Spectroscopy. X-
band EPR spectra were collected on a Bruker EMX electron spin
resonance spectrometer with a Varian liquid nitrogen cryostat at 77 K.
Figure 2. pH titration spectra of Cu(II)(TRI-H)3 at (A) lower pH
(2.92−7.45) and (B) higher pH (7.60−10.80).
C
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
described above was extracted (Table 2); this value
corresponds to a one-proton deprotonation process.
the corresponding Cu(II) affinities are higher for the original
parent peptide. In an effort to isolate the effects, peptides solely
with modifications at position 22, which is much closer to the
copper site than position 27 (Figure 1), were examined. At pH
7.4, for the series TRI-EH, TRI-HK22Q, and TRI-H,
corresponding to Δcharge = −6, −3, and 0, respectively, the
Cu(I) affinity changed only modestly (a factor of 5 less stable
for TRI-HK22Q and essentially identical for TRI-EH) and the
Cu(II) affinities were essentially invariant. At pH 5.8, the
mutated peptides TRI-HK22Q and TRI-EH have very similar
affinities and reduction potentials (Table 4), although they have
higher Cu(I) affinities and lower Cu(II) affinities compared to
the parent peptide TRI-H. It is worth noticing that TRI-H
behaves quite differently compared to the mutated peptides at
pH 5.8, possibility because of the lower pKa’s of the mutated
peptides; however, it appears that Δcharge of the residues at
position 22 (one residue above the binding site, His23) does
not have a dramatic influence on the affinities and redox
properties of the peptide. On the basis of these observations,
we chose to examine a series of peptide substitutions below the
copper-binding site (positions 24 and 27) using the K22E-
derived peptides, as this would allow for the widest range in
Δcharge (from 0 to −12).
As shown in Figure 5A, the Cu(I) affinities for the series of
peptides containing a K22E substitution are related to the total
charge around the copper binding site at both pH 5.8 and pH
7.4. As the total negative charge is increased, going from Δcharge
= 0 (TRI-EHE27K) to Δcharge = −12 (TRI-EHK24E), the
Cu(I) affinity decreased by about 2 orders of magnitude under
both pH conditions. However, the Cu(II) affinities seemed to
be relatively immune to the change in total charge. As shown in
Figure 5B, at both pH 5.8 and pH 7.4, the change in Cu(II)
affinities toward the peptides was minimal at a particular pH
across the series of peptides with different total charges. It is
interesting to note that unlike Cu(I), whose affinities remained
essentially the same for a particular peptide at different pH with
dissociation constants ranging from picomolar to femtomolar,
Cu(II) was destabilized by at least 2 orders of magnitude when
the pH decreased from 7.4 to 5.8, as shown in Figure 5B. The
difference between Cu(I) and Cu(II) can be described more
clearly when we compare the free energies of binding (section 4
in the Supporting Information) at a particular pH (Figure 6).
While reduction potentials for metalloproteins and small
molecules are often determined by direct electrochemical
methods, one may extract the same information by determining
the protein binding affinities of the metals in both oxidation
states and applying the Nernst equation. This approach has the
added benefit that it allows one to track directly the behavior of
interest (e.g., reaction rates) not only against the reduction
potential but also to see whether correlations associated with a
specific oxidation level exist. The reduction potentials of the
Cu(II)/Cu(I)−peptide systems were calculated using the
methodology shown in Scheme 1 and eq 2, where the value
EC°u,aq = 0.159 V vs NHE was utilized (Table 4).36,52 Since the
free energy of binding for Cu(I) varies with amino acid
substitution more than that of Cu(II) (Figure 6), the change in
the calculated reduction potential is dominated by the variation
of the Cu(I) affinity. This is true under both pH conditions;
however, the change in potential between pH conditions is due
to the stabilization of Cu(II) upon acidification. The linear
regressions for the trends following amino acid substitutions
resulted in two parallel lines, the distance between which was
about 100 mV (Figure 7). This value corresponds to a 1 H+ per
a
Table 2. pKa Values and Spectroscopic Parameters of
Selected Cu(II)−Peptides
λmax/nm (ε/M−1 cm−1
)
peptide
pKaE
pKaw
pH 5.8
pH 7.4
TRI-H
TRI-
8.53(2)
644 (143)
644 (143)
671 (79)
659 (89)
657 (81)
9.59(15)
671 (80)
659 (110)
665 (101)
EHE27K
TRI-EH
TRI-
6.33(4)
6.76(6)
9.86(5)
9.81(5)
EHK24E
a
The two pKa values correspond to two different deprotonation
processes. pKaE is associated with Glu22 and pKaw is associated with
the Cu(II)-bound water.
Interestingly, the pH titration of Cu(II)(TRI-EH)3 resulted
in pKa values distinctively different from that of the parent
peptide (Table 2). Again, in the pH region 2.97−5.57, the
increase in the absorbance at 659 nm is indicative of binding of
copper to the peptide and is reminiscent of the parent peptide
(Figure S1 in the Supporting Information). However, when the
pH was raised from 5.33 to 7.87, the absorbance at 659 nm
decreased slightly (Figure 3), resulting in a pKaE value of 6.33.
Figure 3. pH titration spectra of Cu(II)(TRI-EH)3 between pH 5.33
and pH 7.87.
This deprotonation event fitted best to a 1.65(23) proton
process (Figure S1B), which suggested that the nature of this
deprotonation might be different from the one for Cu(II)(TRI-
H)3. Similar effects were also observed for Cu(II)(TRI-
EHK24E)3 (Figure S2). The Cu(II) d−d band evolved
differently upon pH increase for different peptides within the
pH region of interest to us (Figure 4). In summary, we
observed the appearance of a Cu(II) d−d band when Cu(II)
was bound to different peptides, and the transformations of the
spectra were pH-dependent because of various deprotonation
events.
Cu(I) and Cu(II) Affinities, Free Energies of Binding,
and Calculated Reduction Potentials. Dissociation con-
stants for the complexation of either Cu(I) or Cu(II) at pH 5.8
and 7.4 for each of the peptides studied in this work are
provided in Table 3. The first point to recognize is that the
reversal of the position of the Lys and Glu groups in going from
TRI-H to TRI-EHE27K leads to a marked change in the Cu(I)
affinity (a factor of 100 in favor of the reversed peptide), while
D
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
■
●
○
Figure 4. Changes in Cu(II) d−d band absorbance when Cu(II) was bound to ( ) TRI-H (Δcharge = 0); ( ) TRI-EHE27K (Δcharge = 0); ( ) TRI-
□
EH (Δcharge = −6); and ( ) TRI-EHK24E (Δcharge = −12). The inset shows differences in the evolution of the Cu(II)−peptide d−d band, revealing
different pKa values (same symbols).
Table 3. Cu(I) and Cu(II) Dissociation Constants at pH 5.8 and pH 7.4
Kd/M
Cu(I)
Cu(II)
peptide
ΔCharge
pH 5.8
pH 7.4
pH 5.8
pH 7.4
a
TRI-H
−
−3
0
3.09(66) × 10−12
2.37(27) × 10−14
1.36(56) × 10−14
2.42(118) × 10−14
4.17(19) × 10−14
5.38(323) × 10−13
3.52(9) × 10−13
2.00(61) × 10−13
4.12(114) × 10−14
2.07(27) × 10−15
1.76(41) × 10−14
1.12(43) × 10−13
8.06(509) × 10−14
4.23(292) × 10−13
4.04(83) × 10−8
4.44(59) × 10−7
3.47(104) × 10−7
1.26(32) × 10−7
7.38(324) × 10−7
2.18(104) × 10−7
4.92(225) × 10−7
8.69(112) × 10−9
3.80(106) × 10−9
1.45(71) × 10−9
3.76(72) × 10−9
6.39(258) × 10−9
3.73(203) × 10−9
4.98(173) × 10−9
TRI-HK22Q
TRI-EHE27K
TRI-EHE27Q
TRI-EH
−3
−6
−9
−12
TRI-EHK24Q
TRI-EHK24E
a
The data for TRI-H are from ref 36 and were collected at pH 5.9 and 7.4.
Table 4. Calculated Reduction Potentials Based on Cu(I)
and Cu(II) Affinities
differences between the copper sites. The EXAFS data for
Cu(I)(TRI-EH)3 (Figure 8B,C) support this structural
interpretation. The Fourier transform is dominated by a peak
at R ≈ 1.5 Å typical of Cu−N or Cu−O nearest-neighbor
scattering, with pronounced outer-shell scattering typical of
rigid ligands such as histidine. Quantitative fitting (Table 5)
confirmed this structure. The nearest-neighbor distance of 1.93
Å is typical of three-coordinate Cu(I), and the outer-shell
scattering can be modeled using three imidazole ligands; these
results are quantitatively consistent with structural parameters
reported previously for Cu(I)(TRI-H)3.36 Attempts to fit the
data with a Cu−O shell in addition to or in place of one of the
Cu−imidazole interactions did not give any improvement in
the fit (Table S6 in the Supporting Information).
E°(Cu−pep3)/mV
peptide
ΔCharge
pH 5.8
pH 7.4
a
TRI-H
−
−3
0
402(8)
433(8)
TRI-HK22Q
TRI-EHE27K
TRI-EHE27Q
TRI-EH
589(5)
450(10)
504(13)
474(8)
597(13)
556(14)
587(11)
491(20)
522(12)
−3
−6
−9
−12
440(14)
435(21)
400(20)
TRI-EHK24Q
TRI-EHK24E
a
For TRI-H, calculations were based on the affinities at pH 5.9 and
7.4.
EPR Analyses of Cu(II)−Peptide Complexes. EPR
spectra of Cu(II)(TRI-H)3 were collected from pH 5.87 to
pH 9.14, and they showed a transition of A∥ and g∥ from pH
7.80 to pH 9.14 (section 3 in the Supporting Information). The
EPR parameters of Cu(II)(TRI-EHE27K)3, Cu(II)(TRI-EH)3,
and Cu(II)(TRI-EHK24E)3 at both pH 5.8 and 7.4 were in the
same range as those of the parent peptide complex (Table 6).
Specifically, g⊥ is smaller than g∥ and A∥ is in the range of 17−
19 mT for all of these complexes, indicative of a type 2 copper
center with a high coordination number (4 or 5), which is
consistent with the observed d−d bands in the UV−vis spectra.
electron change in potential, which is consistent with the
difference of 1.6 pH units for these measurements.
K
d,Cu(I)(pep)3
2.303RT
°
°
ECu(pep) = ECu,aq
−
log
3
nF
Kd,Cu(II)(pep)
3
(2)
XAS of Cu(I)(TRI-EH)3. The XANES spectra of Cu(I)(TRI-
EH)3 and Cu(I)(TRI-H)3 (Figure 8A) both show a modest-
intensity peak at ∼8983 eV; this is typical of Cu(I) in an
approximately trigonal environment.53 The near identity of
these spectra shows that there are only minor structural
E
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 6. Free energies of binding for Cu(I) and Cu(II) at (A) pH 5.8
and (B) pH 7.4. Peptides: (1) TRI-EHE27K; (2) TRI-EHE27Q; (3)
TRI-EH; (4) TRI-EHK24Q; (5) TRI-EHK24E.
Figure 5. (A) Cu(I) dissociation constants and (B) Cu(II)
dissociation constants with respect to Δcharge at pH 5.8 and pH 7.4.
Peptides: (1) TRI-EHE27K; (2) TRI-EHE27Q; (3) TRI-EH; (4)
TRI-EHK24Q; (5) TRI-EHK24E. It should be noted that the y axes
have log scales.
Scheme 1
NiR Activity. The Cu−peptides exhibited NiR activity with
different rates (Figure 9), all of which are pH-dependent.
Specifically, as the total negative charge was increased from 0 to
−12, the rates of the reaction increased about 4-fold. Moreover,
the rates are also correlated to the calculated reduction
potentials at pH 5.8 (Figure 10): as the calculated reduction
potentials increase, the rates decrease. The turnover numbers
(TONs) for these reactions, expressed as moles of electrons per
mole of copper, were obtained. The TONs in 1 h are linearly
associated with the local negative charge (Figure S6). As Δcharge
increased from 0 (TRI-EHE27K) to −12 (TRI-EHK24E), the
TONs at pH 5.8 increased by one. For all of the NiR reactions,
we did not observe any degradation of the enzyme 1.5 h after
the reaction was initiated. Nitric oxide production was
confirmed for the peptides showing the lowest and highest
rates [Cu(TRI-EHE27K)3 and Cu(TRI-EHK24E)3, respec-
tively] at pH 5.8 (section 8 in the Supporting Information).
DISCUSSION
Figure 7. Calculated reduction potentials in relation to the changes of
the local charge at pH 5.8 and pH 7.4. Peptides: (1) TRI-EHE27K;
(2) TRI-EHE27Q; (3) TRI-EH; (4) TRI-EHK24Q; (5) TRI-
EHK24E.
■
The environment that surrounds redox sites plays a critical role
in determining the redox potentials and pathways for electron
flow in proteins. However, in most cases, multiple factors exert
influence on the redox potential simultaneously, complicating
the understanding of the structure−function relationship of a
particular system. Conceptually, we can think of the energy
associated with the redox process as the sum of the inner-
sphere contributions, which directly point to the energy needed
to add or remove an electron from the redox site, and outer-
sphere contributions, which are related to the interaction of the
redox site with the protein and solvent environment during the
redox process.54 In reality, more often than not a single change
F
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 8. (A) XANES of Cu(I)(TRI-H)3 at pH 7.4 (black line) and Cu(I)(TRI-EH)3 at pH 8.5 (red line). (B) EXAFS of Cu(I)(TRI-EH)3 at pH
8.5 (experimental data, solid line; fit, dashed line). (C) Fourier transform of the EXAFS data in R space (experimental data, solid line; fit, dashed
line).
a
Table 5. EXAFS Fitting Parameters for Cu(I)(TRI-EH)3
first shell
R/Å
σ2/Å2
outer shells
R/Å
σ2/Å2
3 Cu(I)−N(1)
1.93
0.008
3 Cu−C(1)
3 Cu−C(2)
3 Cu−N(2)
3 Cu−C(3)
2.91
2.95
4.05
4.06
0.012
0.012
0.010
0.010
a
ΔE0 = −11.94 eV, R (distance), and σ2 (Debye−Waller factor) were
the only freely variable parameters. The coordination number was held
constant at 3, and the outer-shell parameters were calculated assuming
a rigid imidazole ring. For a comparison of different models, see Table
S6.
Table 6. EPR Parameters of Cu(II)(3SCC)3 at pH 5.8 and
pH 7.4
pH 5.8
○
Figure 9. pH dependence of nitrite reduction rates. Peptides: ( )
□
TRI-EHK24E (Δcharge = −12); ( ) TRI-EHK24Q (Δcharge = −9);
(◐) TRI-EH (Δcharge = −6); ( ) TRI-EHE27Q (Δcharge = −3); (*)
peptide
g⊥
g∥
A⊥/mT A∥/mT Aiso/mT
●
Cu(II)(TRI-H)3
2.05
2.06
2.06
2.06
2.28
2.27
2.27
2.28
0.64
0.68
0.36
0.18
18.58
17.36
18.33
18.33
6.62
6.24
6.35
6.23
TRI-EHE27K (Δcharge = 0).
Cu(II)(TRI-EHE27K)3
Cu(II)(TRI-EH)3
that mutating the valine (Val) residue into a charged or polar
residue such as Glu, aspartate (Asp), or asparagine (Asn) leads
to a decrease in the midpoint redox potential of recombinant
myoglobin at 25 °C, which demonstrated that changing the
electric field around the redox center would result in a
substantial change in the redox thermodynamics of a particular
system.30 Specifically, the substitution of Val68 close to the
distal site of the heme center with Glu or Asp led to a decrease
of about 200 mV in the reduction potential, whereas Val68Asn
resulted in an 80 mV decrease. Furthermore, the Dutton and
Gibney groups showed that by varying the burial of heme in the
hydrophobic core, heme peripheral substituents, heme-hydro-
phobic amino acids, and so on, they were able to modulate
rationally the heme-protein redox potential over a range of a
few hundred millivolts.55−58 More recently, Lu and co-workers
Cu(II)(TRI-EHK24E)3
pH 7.4
peptide
g⊥
g∥
A⊥/mT A∥/mT Aiso/mT
Cu(II)(TRI-H)3
2.05
2.05
2.06
2.05
2.27
2.27
2.27
2.27
0.64
1.07
1.07
0.71
18.58
17.50
18.15
17.68
6.62
6.55
6.76
6.37
Cu(II)(TRI-EHE27K)3
Cu(II)(TRI-EH)3
Cu(II)(TRI-EHK24E)3
around the active site leads to a cascade of events that
eventually modifies the redox properties, but it is challenging to
track down the specific role of each event.
A significant amount of work has been done to understand
the redox property modulations on systems containing heme
cofactors or cupredoxin sites. It was reported as early as 1989
G
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
upon the deprotonation of Glu at e positions with a pKa of 4.0−
4.5.35,62 This behavior has been noted for several peptides of
the TRI family, but most notably for TRI-H.36 Second, the
copper binding ligand can undergo deprotonation of the
cationic imidazolium form, which does not bind copper, to the
neutral imidazole, which binds to copper.63 An 1H NMR
titration of (TRI-H)3 showed that this deprotonation event
occurs over the pH range 5.5−8.0 in the absence of metal.36
Most likely, the presence of a Lewis acid shifts the imidazolium
to imidazole equilibrium in favor of the neutral metal binding
ligand, as has been reported for both Zn and Cu complexes
with (TRI-H)3.36,42 Furthermore, under higher pH conditions
the imidazole may be converted to the negatively charged
imidazolate form, which also binds to copper efficiently. It is
generally believed, however, that this equilibrium occurs at
much higher pH values than investigated here. Additional
carboxylic acid deprotonation events would be expected when
Lys is substituted by Glu at position 22 and/or 24 for the
mutated peptides. One would expect this deprotonation
chemistry to occur in the general range of the pre-existing
carboxylic acids at the e positions of the peptide. Under very
high pH conditions (again above pH 8), deprotonation of the
Lys side chain may become relevant, which could disrupt the
3SCCs; however, the conditions used in this study were
unlikely to enter this basic range. Finally, when copper is
bound, and if a water is bound to the copper, deprotonation of
the Cu(II)-bound water could occur in the basic pH region.
An examination of the pH-dependent trends in Figure 4
demonstrates that TRI-H, TRI-EHE27K, TRI-EH, and TRI-
EHK24E all exhibit a sharp rise in absorbance moving from pH
3 until the spectra level out by pH ∼5. Two major processes
occur over this pH range. The first is the conversion of the
2SCCs to 3SCCs through deprotonation of Glu. This
conversion for the apo-peptide is not reflected in these
spectroscopic changes. The second process is the binding of
Cu(II) to the 3SCCs, which is directly reflected by these
spectra, as they arise from the d−d transition of Cu(II).
Previous analysis of this copper site for Cu(II)(TRI-H)3 using
EPR spectroscopy suggested that this system is probably a five-
coordinate Cu(II) with three coordinated imidazoles and two
solvent molecules.36 The UV−vis spectra of all these peptides
are consistent with the same chromophore until at least pH 6.0.
As the pH is raised to neutral conditions, peptides containing
additional Glu (TRI-EH, and TRI-EHK24E) show an
additional pH-dependent conversion (pKaE ≈ 6.5). This
process could be due to (A) deprotonation of the copper-
coordinated water; (B) complexation of Glu22 to the copper;
or (C) deprotonation of Glu22, leading to the formation of a
hydrogen bond with either a coordinated His, a water H-
bonded to a coordinated His or a bound water. We disfavor the
explanations that Glu22 binds directly to Cu(II) or forms a
hydrogen bond with coordinated water, as the orientation of
the Glu on the helix places it too far from the metal center to
interact in these ways. This observation leaves the most likely
explanations for the pKaE in the 6−7 range for peptides
containing the K22E mutations as deprotonation of metal-
bound water or deprotonation of the glutamic acid leading to a
H-bond formation between Glu22 and His23 (including a
water bridging the two amino acids).
Figure 10. Rates at pH 5.8 in relation to the calculated reduction
potentials at pH 5.8. Peptides: (1) TRI-EHE27K; (2) TRI-EHE27Q;
(3) TRI-EH; (4) TRI-EHK24Q; (5) TRI-EHK24E.
demonstrated that the redox potential of a cupredoxin site in
azurin can be tuned by modifying two important secondary
coordination sphere interactions: hydrophobicity and H-
bonding. Their work demonstrated an astounding shift in
redox potentials, in some cases by as much as 700 mV, which
extended even beyond the natural redox potential range of the
protein.59,60 To our current knowledge, there have been no
systematic studies of the redox property modulations of the
Cu(His)3 site in a protein environment, despite its importance
in the electron transfer or catalytic functions in native proteins.
Taking advantage of a de novo protein design approach, we
took a known functional peptide, Cu(TRI-H)3, and modified
the local charge in a stepwise manner in order to observe
changes in the reduction potentials and NiR activity of a series
of peptides (Table 1). The rationale of the design was first to
invert the charged residues Glu27 and Lys22, yielding TRI-
EHE27K. Relative to the parent peptide TRI-H, the change in
the overall charge (denoted as Δcharge) for this peptide is 0.
While this double mutation led to a significant shift in potential,
we observed that there was no perturbation in the reduction
potentials between copper complexes of TRI-H and TRI-EH at
pH 7.4. We used the K22E mutation (TRI-EH scaffold) to
evaluate a broader range of charge effects than would be
available with TRI-H. We changed the residues of TRI-EH in a
stepwise manner below the Cu(His)3 site from positively
charged residues to neutral and then negatively charged
residues. Since three peptide strands self-assemble in aqueous
solution above pH 5,61 the resulting series of peptides have
Δcharge values of 0, −3, −6, −9, and −12. None of the residues
involved in the mutations (Lys22, Lys24, and Glu27 in TRI-H)
are inside the helix bundle, nor can they directly coordinate to
the copper center (according to our PyMOL models shown in
Figure 1). Hence, we did not expect to see first-coordination-
sphere effects on the properties of the copper center. This de
novo design strategy allowed us to vary the charge systemati-
cally in order to probe the subtle influence of surface charged
residues on the stability constants of Cu(I) and Cu(II) protein
forms and, subsequently, upon the redox properties and
catalytic activities of these type 2 copper centers.
We first characterized the binding of the Cu(II) ion to the
peptide as a function of pH using UV−vis spectroscopy, as
shown in Figures 2−4. Before consideration of these spectra, it
is worth recognizing that there are multiple pH-dependent
reactions that can occur between pH 2.5 and 8.0 that can affect
the chemistry in the system. First, the conversion of two-
stranded coiled coils (2SCCs) to 3SCCs is known to occur
As shown in Figure 2B, Cu(II)(TRI-H)3, which does not
contain the K22E modification, exhibits a second significant
absorption change at higher pH with a pKaW of 8.53. This is
thought to be associated with the deprotonation of the Cu(II)-
H
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
bound water into a Cu(II)-bound hydroxide because the blue
shift upon this deprotonation is in accordance with the
prediction of the same process in an equatorially coordinated
Cu(II) complex.64 Spectral perturbations over this range were
also observed in the EPR spectra of Cu(II)(TRI-H)3. As the
pH increased from 7.80 to 9.14, g∥ decreased from 2.27 to 2.26
while A∥ increased from 18.58 to 19.29 mT (Table S2). This
observed trend is consistent with the reported case of copper-
bound water deprotonation at the type 2 copper center in the
mercury derivative of laccase.65 This piece of evidence,
combined with the transformation of the UV−vis spectra in
this pH region, leads us to believe that the pKaw at 8.53 for
Cu(II)(TRI-H)3 corresponds to a Cu(II)-bound water
deprotonation, producing a Cu(His)3(OH)(OH2)0−1 species.
Also consistent with the UV−vis spectra, the Cu(II)(TRI-H)3
EPR spectra stay unchanged in the pH region 5.8 to 7.4,
indicating that Cu(II) is likely to maintain a Cu(His)3(OH2)1−2
coordination with three quasi-in-plane imidazoles and one or
two water molecules.36
In contrast, for the TRI-EH series of peptides, we observed
different EPR parameters at pH 5.8 and 7.4. For the “charge-
flipped” peptide Cu(II)(TRI-EHE27K)3, g∥ stayed the same
while A∥ increased by 0.14 mT and the isotropic hyperfine
splitting constant Aiso increased by 0.31 mT (Table 6). Similar
changes in g∥ and Aiso occurred for Cu(II)(TRI-EH)3. For
Cu(II)(TRI-EHK24E)3, however, both g∥ and A∥ decreased
significantly while A⊥ increased, leading to an increased Aiso. In
summary, the Aiso of the TRI-EH series of copper−peptide
complexes increased when pH rose from 5.8 to 7.4, indicating a
higher spin density at the copper center. The changes in the
other EPR parameters are dissimilar to those of Cu(II)(TRI-
H)3 in a higher pH region (7.8 to 9.1), which implies that this
equilibrium (corresponding to pKaE ≈ 6.5) is probably not a
Cu(II)-bound water deprotonation.
was thought that the Cu(II) state would be more greatly
stabilized by higher negative charge, decreasing the reduction
potential of the system. While the reduction potentials (Table
4) were made less positive by increasing the local negative
charge, contrary to our expectation, this shift occurred as a
result of Cu(I) destabilization rather than Cu(II) stabilization.
These results are most clearly visualized by comparing the free
energies of binding shown in Figure 6.
In contrast, Rorabacher and co-workers have reported a
series of potentially tetradentate copper complexes that vary
neutral nitrogen or sulfur donors over the range N4 to NS3. The
reduction potentials of these compounds were strongly
influenced by the stability constants of the Cu(II) oxidation
level rather than those of the Cu(I) complexes.66,67 In
particular, an increase in the stability constants of the Cu(II)
complexes by as much as 104−105-fold was observed upon
substitution of thioether sulfurs for unsaturated nitrogens; at
the same time, the stability constants of the Cu(I) complexes
were maintained in the range 1014−1016 M−1. These changes
led to differences in the reduction potentials that spanned 1.5
V.66 The redox potentials were also observed to be pH-
dependent below pH 5 (but invariant above this pH) because
of the acid−base chemistry of the ligand.68 Important
differences between these models and our system include the
choice of ligands (neutral thioethers and pyridyl groups), the
lack of potential H-bond acceptors (such as glutamate), the lack
of charge variation across the series, and the possibility of a
tertiary amino nitrogen atom that could serve as a fourth ligand.
Hence, our system provides a more biologically relevant
example to assess how electrostatic and H-bonding influence
copper protein reduction potentials.
As shown in Figure 5, the relative Cu(I) or Cu(II) affinity
change across the series of designed protein mutants is
independent of pH, as demonstrated by the invariant slopes
of the lines for the two oxidation states. This statement is not
intended to suggest that pH is not important in this system, as
marked effects on both the Cu(II) stability and reduction
potential are seen. As was shown, the Cu(I) stabilities are more
sensitive to protein modification [Δ(ΔGaCu(I)) ≈ 2 kcal/mol vs
Δ(ΔGaCu(II)) ≈ 0.7 kcal/mol], whereas one sees that the Cu(II)
Considering the combined observations, we assign the pKaE
at ∼6.5 for some of the TRI-EH peptides to a protonation
equilibrium of a glutamate side chain (probably at position 22),
which can become a hydrogen-bond acceptor for the imidazole
proton at position 23 (or a water bridging the two amino
acids). This H-bond leads to the variance of both the UV−vis
and EPR spectral parameters of the Cu(II) center. Consistent
with this assignment, when the same pH titration was carried
out for Cu(II)(TRI-HK24E)3, where the Glu was placed at
position 24 instead of position 22, we did not observe a
protonation equilibrium in this low pH range (Figure S4 and
Table S1). For the “charge-flipped” peptide Cu(II)(TRI-
EHE27K)3, one might have expected it to behave more
similarly to the other TRI-EH peptides having a lower pKaE;
however, models suggest that Glu22 can interact with Lys27 in
TRI-EHE27K, which would weaken the H-bonding interaction
of Glu22 with His23. This putative H-bond would lead to a less
acidic Cu(II), which would increase the pKaw for the Cu(II)-
bound water deprotonation in these peptides. Correspondingly,
we observed a higher pKaw (9.6−9.8) for the TRI-EH series of
peptides, which we assign as the deprotonation of the Cu(II)-
bound water (Table S1).
affinities are more sensitive to pH adjustment [(ΔGaCu(I)
)
pH 7.4
− (ΔGaCu(I)
)
≈ 0.5 kcal/mol vs (ΔGaCu(II)
)
−
pH 5.8
pH 7.4
(ΔGaCu(II)
)
≈ 3.5 kcal/mol]. Thus, the trend is that as
pH 5.8
the negative charge around the copper center is increased via
protein modification, Cu(I) is destabilized and the reduction
potential becomes less positive. In contrast, as the pH is
lowered, the Cu(II) affinity decreases and the reduction
potential becomes more positive (Table 4). Within the series
there is shift in potential of ∼100 mV in going from pH 7.4 to
pH 5.8 for any peptide, providing a slope of ∼60 mV/pH unit,
which indicates that the change in potential is dependent upon
a single proton per electron. The differences between the
copper affinities lead to the two parallel lines shown in Figure 7.
Thus, by controlling the peptide sequence and the pH, one can
systematically vary the reduction potential in this system by
nearly 200 mV.
At pH 5.8 and pH 7.4, we determined the binding affinities
of Cu(I) and Cu(II) to the series of peptides prepared in this
work in order to evaluate the reduction potentials and the
impact of protein mutations at both oxidation levels (Table 3
and Figures 5 and 6). These peptides were prepared in order to
examine the effect of increased local negative charge on the
stabilities of Cu(I) and Cu(II) within the hydrophobic core. It
The observed trends can be explained by the change in H-
bonding and salt-bridging interactions around the copper
center upon protein mutations under different pH conditions.
The mutation of Lys at position 22 into a Glu would likely lead
to an H-bonding interaction between the carboxylic acid/
carboxylate from Glu and the imidazole from His (or a water
bridging the two amino acids), and therefore, we see pH-
I
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
dependent Cu(II) affinities and reduction potentials. Variations
of the other charged residues would alter the salt-bridging
interactions, leading to differences in the electron-donating
abilities of the imidazoles or structural changes to the Cu(I)
center. Consistent with this idea, we see that TRI-H exhibits a
modest shift to a less positive potential (Δ = −30 mV) in going
from pH 7.4 to pH 5.9 while all of the peptides with Gln or Glu
at position 22 show significant shifts to more positive potentials
(Δ = +60 to +140 mV) as the pH is decreased (Table 4). One
possible H-bonding system that supports both the pH trends in
Cu(II) binding and the impact of side-chain mutations of the
Cu(I) affinities is provided in section 9 in the Supporting
Information.
Alteration of the peptide in the layer above the copper site
(position 22) has virtually no effect on the reduction potential
of the system (at pH 7.4: TRI-H, E° = 433 mV; TRI-HK22Q,
E° = 450 mV; TRI-EH, E° = 440 mV). In general, with
peptides containing the K22E modification (i.e., those where
mutations are made to TRI-EH), one observes that increasing
the side-chain positive charge below the metal site causes
significant stabilization of the Cu(I) oxidation level, while the
Cu(II) state is mildly stabilized by the same mutation. Thus, if
we take TRI-EH as the parent peptide, replacing Lys at
position 24 with Gln and Glu progressively destabilizes Cu(I)
binding, whereas converting Glu27 into Gln or Lys yields much
higher Cu(I) affinities. If this were solely an electrostatic effect,
one might expect the opposite trend for the Cu(I) affinity (as
more positive charge around the positive metal should
destabilize binding) and that a greater effect would be observed
for the Cu(II) level (as charge repulsion would be greater).
These observations suggest that an alternative explanation is
required. Our best explanation is that this may again reflect H-
bonding and salt bridging between the different surface
residues. In the case of Cu(I)−peptide complexes, the
dominant factor may be structural perturbations that occur at
the metal center.
In order to form an effective geometry for directional H-
bonding and salt-bridging interactions, the imidazoles may need
to reorient to minimize the conformational changes on the
peptide backbone, leading to a more trigonal-planar or
tetrahedral Cu(I) structure. Our present best explanation for
the decreased Cu(I) affinity as the H-bond is formed is based
on this geometric argument rather than electronic factors. It
was previously shown that Cu(I) in TRI-H forms a distorted
trigonal-planar structure.36 We examined the parent peptide for
the series of peptides reported here: TRI-EH, forming a direct
comparison to TRI-H. XANES data revealed that the Cu(I)
coordination number did not change when we substituted a
Glu for the Lys at position 22, which is consistent with the fact
that the affinity of Cu(I) in this pH range does not change
much in going from TRI-H to TRI-EH. The subtle differences
of the edge structure are consistent with small changes in the
symmetry of the site. For example, the small decrease in the
intensity of the peak at 8983.4 eV would be consistent with a
change from a more T-shaped geometry to a more trigonal-
planar structure,53 consistent with our predictions. The EXAFS
fits suggest relatively large disorder in the Cu(I) site. The fitted
σ2 values for Cu(I)(TRI-EH)3 are smaller than those found for
Cu(I)(TRI-H)3,36 indicating that there is somewhat less
variation in the Cu−N distances for Cu(I)(TRI-EH)3. This is
consistent with the suggestion, based on the XANES, that the
Cu(I) site in Cu(I)(TRI-EH)3 is somewhat more symmetric
than that in Cu(I)(TRI-H)3.
It is likely that a similar, possibly even more rigid trigonal
plane exists for TRI-EHE27K because of the formation of an
interhelical H-bonding network connecting His23, Glu22, and
Lys27 from the adjacent strands, which would hold the copper-
coordinating His ligand in a relatively constrained environment
(Figure S9). This is of great relevance to the rack-induced
binding hypothesis proposed for blue copper electron transfer
centers in native proteins.69 This hypothesis suggests that the
protein matrix is responsible for maintaining a proper metal
coordination environment and fine-tuning the redox potential
to facilitate electron transfer processes. Of great interest to us is
a report that demonstrates the importance of a hydrogen bond
close to the type 1 copper center in plastocyanin, which
contributes to the rigidity of the coordinating environment of
the type 1 copper center.70 The higher reduction potential of
Cu(TRI-EHE27K)3 is reflective of the rigidity at the copper
centers that favors a trigonal Cu(I). This can explain the
difference between our system and Rorabacher’s complexes
mentioned earlier. In our case, the Cu(II) state is more
structurally dynamic while Cu(I) is more rigid. As a result, the
modulation of reduction potentials is dominated by the
variance in the Cu(I) state. In Rorabacher’s case, the small-
molecule scaffold has a similar level of rigidity for both
oxidation states, so the stability constants of the higher
oxidation state are more dramatically influenced by the number
and nature of the donor atoms, the chelate ring size, and the
general ligand topology, while little variation is observed for the
lower oxidation states.
For TRI-EHK24E, while Glu22 forms a hydrogen bond with
His23 (or a water bridging the two amino acids), the charge
repulsion between Glu22, Glu24, and Glu27 on the
neighboring helices might force the imidazoles to adopt a
tetrahedral geometry. Such perturbations in geometry are not
required for the Cu(II) site, as it is believed that this ion is
already five-coordinate with the imidazoles no longer planar.
For Cu(II), the influence of salt-bridging and H-bonding effects
are not as pronounced under a specific pH condition. This
point may be understood most simply as follows: Cu(II) is
much more Lewis acidic, which should lead to acidification of
the coordinated imidazole proton. Thus, regardless of
competing salt bridges, the Glu22 substitution may lead to a
hydrogen bond of sufficient stability not to be diminished
significantly by salt-bridging effects and protein mutations.
We next examined how side-chain modifications would affect
the known catalytic activities of this system. The nitrite
reductase (NiR) activity of these peptides is expressed in terms
of the rate of ascorbate consumption. The influence of the
reduction potential on the NiR rate is clearly revealed by an
examination of the rates at a constant pH. At pH 5.8, for the
TRI-EH series of peptides, the higher the reduction potential,
the lower the rate (Figure 10). The reduction potential changes
from 597 mV (TRI-EHE27K) to 491 mV (TRI-EHK24Q),
leading to a 3.5-fold increase in rate [from 1.07(65) × 10−6 to
3.75(76) × 10−6 M min−1]. While there is an impressive linear
correlation between the rates and the reduction potentials, one
must use caution to ascribe this effect to electron transfer since
our studies show that the rapid reduction of Cu(II) to Cu(I) by
ascorbate cannot be rate-limiting. More likely, the significant
reorganization energy due to the change in coordination
geometry from Cu(I) [three-coordinate] to Cu(II) [five-
coordinate] limits the observed rates.
The nitrite reduction rates also increased markedly with
decreasing pH, which was observed for both the parent
J
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
peptide36 and the mutated peptides (Table S4). The native
CuNiR is also reported to exhibit pH-dependent activities that
are proposed to be closely related to both the electron transfer
process and the nitrite binding step at the catalytic site.23 In our
case, since we do not have a type 1 electron transfer site and the
reduction of the Cu(II) by ascorbate is much faster than the
catalytic reaction, the observed pH dependence is likely
correlated to steps during the nitrite reduction catalytic cycle.
Since the full chemical reaction requires 2 equiv of protons, it is
not surprising to see that the reaction rates are higher at a lower
pH. Quantitatively, when the pH decreases from 7.0 to 5.8, the
proton concentration in the solution [H+] increases about 16-
fold; if the rate of the reaction were directly proportional to
[H+]2, it should have increased by 256-fold. Instead, an increase
in rate of ∼10-fold was observed for TRI-H. The rates for the
TRI-EH series of peptides similarly increased by 10−16-fold
when the pH changed from 7.0 to 5.8 (Table S4). If one fits a
plot of log(initial rate) versus pH for TRI-H to the expression
log(initial rate) = a − b·pH, the b coefficient, which
corresponds to the proton order of the reaction, is 0.85.
Thus, while the stoichiometric reaction requires two protons,
the rate-limiting step is dependent on a single proton. The
second proton is necessary once the hydroxide leaving group is
formed.24,71,72
For all of the TRI-EH peptides, when the pH decreased from
7.4 to 5.8, the reduction potential increased by about 100 mV.
If this were the only factor influencing the reaction rate, the
NiR rate should have decreased. It is worth noticing that the
rates are very low at pH 7.0, so the NiR rates were analyzed
only at pH 5.8. As a result, for a specific peptide, when the pH
decreases from 7.0 to 5.8, the participation of a proton in the
rate-limiting step directly enhances the rate, while the
significant structural change upon conversion from Cu(I) to
Cu(II) (reflected as an increased reduction potential) impedes
the reaction. This is consistent with the trends observed when
comparing the peptides at a fixed proton concentration.
Combining these two opposite effects, we observed a 10−16-
fold rate increase when the pH decreased from 7.0 to 5.8.
It should be noted that the NiR rate versus reduction
potential correlation at a constant pH should be interpreted
with caution, as it compares only peptides that contain a Glu at
position 22. If one compares the parent peptide TRI-H and the
“charge-flipped” peptide TRI-EHE27K at pH 5.8, the situation
is different. The reduction potential of TRI-EHE27K is about
200 mV more positive than that of TRI-H, but its NiR rate is
only 2 times less than that of the parent peptide. Using the
correlation in Figure 9, one would have expected a 6-fold
differential between these two peptides if reduction potential
alone were the determining factor. In general, despite the fact
that the TRI-EH peptides have much more positive reduction
potentials, they have, as a group, greater than (or approximately
equal) activities compared with TRI-H. For example,
comparing the rates of Cu(TRI-H)3 and Cu(TRI-EH)3 at
pH 5.8 (Table S4) shows two peptides with similar activities
but strongly divergent reduction potentials. This observation
points to the importance of the Glu at position 22 for
enhancing the rates and suggests that the other substitutions
made in this series moderate the beneficial impact of this Glu22
on catalysis.
peptidic framework by stepwise modifications that influence the
local charge. These modifications vary the deprotonation
equilibria of the Cu(II)−peptide complexes, which have a
direct impact on the Cu(II) affinities to the apo-peptide under
different pH conditions. At a specific pH, the influence of the
mutations is reflected by the linear correlation of the Cu(I)−
peptide affinity and Δcharge. Reduction potentials were
calculated from the Cu(I) and Cu(II) affinities to the apo-
peptides. For a particular peptide, the difference in the
reduction potentials under different pH conditions originates
from the change in Cu(II) affinity; at a specific pH, the
difference in the reduction potentials for the TRI-EH series of
peptides is dominated by the variance in the Cu(I) affinities,
which could be the result of the structural perturbation of
Cu(I) coordination upon the alteration of charged residues.
Moreover, all of these metalloenzymes exhibit pH-dependent
NiR activities. For a specific peptide, the NiR rates are
governed by the proton concentration and the reduction
potentials under different pH conditions. At a specific pH, the
reduction potentials correlate with the observed rates, but other
factors, such as the formation of the substrate−enzyme
complex, formation of the product−enzyme complex, and
product release, may also play a role in determining the NiR
rates. In particular, the data implicate a key role for a Glu at
position 22 in enhancing catalysis. This is likely due to either a
hydrogen bond to His23 (or a water bridging the two amino
acids) or an interaction with the copper-bound substrate.
This study demonstrates that even in highly simplified
scaffolds, attempts to change the local charge around the active
site may cause multiple interactions that can vary the redox
properties and activities in distinct ways. Nevertheless, this de
novo design approach allows us to begin to peel the onion layer
by layer. Although we still have not drawn a complete picture of
the chemistry in this series of peptides, we have demonstrated
our ability to isolate different factors (pH, affinities, reduction
potentials, etc.) and to compare and analyze the effects in a
multidimensional way. In conclusion, this work showcases the
strength of de novo protein design to anchor the in-depth
understanding of fundamental interactions that modulate the
redox properties and NiR activities for a type 2 copper center,
which provides interesting insights into future work on model
chemistry in general.
ASSOCIATED CONTENT
* Supporting Information
■
S
Detailed data fitting process, calculations, additional character-
izations, and PyMOL models. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
vlpec@umich.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
V.L.P. thanks the National Institutes of Health for financial
support of this research (ES012236). F.Y. thanks the
Department of Chemistry of the University of Michigan for a
research fellowship. Portions of this research were carried out at
the Stanford Synchrotron Radiation Lightsource, a Directorate
of SLAC National Accelerator Laboratory and an Office of
CONCLUSION
We have reported here a systematic approach of modulating the
reduction potentials and NiR activities within a self-assembled
■
K
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(37) Ghadiri, M. R.; Case, M. A. Angew. Chem., Int. Ed. Engl. 1993,
32, 1594−1597.
Science User Facility operated for the U.S. Department of
Energy Office of Science by Stanford University.
(38) Tanaka, T.; Mizuno, T.; Fukui, S.; Hiroaki, H.; Oku, J.-I.;
Kanaori, K.; Tajima, K.; Shirakawa, M. J. Am. Chem. Soc. 2004, 126,
14023−14028.
REFERENCES
■
(1) Prabhulkar, S.; Tian, H.; Wang, X.; Zhu, J.-J.; Li, C.-Z. Antioxid.
Redox Signaling 2012, 17, 1796−1822.
(39) Kiyokawa, T.; Kanaori, K.; Tajima, K.; Koike, M.; Mizuno, T.;
Oku, J. J.-I. I.; Tanaka, T. J. Pept. Res. 2004, 63, 347−353.
(40) Hong, J.; Kharenko, O. A.; Ogawa, M. Y. Inorg. Chem. 2006, 45,
9974−9984.
(41) Hong, J.; Kharenko, O. A.; Fan, J.; Xie, F.; Petros, A. K.; Gibney,
B. R.; Ogawa, M. Y. Angew. Chem., Int. Ed. 2006, 45, 6137−6140.
(42) Zastrow, M. L.; Peacock, A. F. A.; Stuckey, J. A.; Pecoraro, V. L.
Nat. Chem. 2012, 4, 118−123.
(43) Fmoc Solid Phase Peptide Synthesis: A Practical Approach; Chan,
W. C.; White, P. D., Eds.; The Practical Approach Series, Vol. 222;
Oxford University Press: New York, 2000.
(44) Farrer, B. T.; Harris, N. P.; Valchus, K. E.; Pecoraro, V. L.
Biochemistry 2001, 40, 14696−14705.
(45) The Protein Protocols Handbook, 2nd ed.; Walker, J. M., Ed.;
Humana Press: Totowa, NJ, 2002.
(46) Xiao, Z.; Loughlin, F.; George, G. N.; Howlett, G. J.; Wedd, A.
G. J. Am. Chem. Soc. 2004, 126, 3081−3090.
(47) Xiao, Z.; Wedd, A. G. Nat. Prod. Rep. 2010, 27, 768−789.
(48) Xiao, Z.; Brose, J.; Schimo, S.; Ackland, S. M.; La Fontaine, S.;
Wedd, A. G. J. Biol. Chem. 2011, 286, 11047−11055.
(49) George, G. N. EXAFSPAK; Stanford University: Stanford, CA,
2000; available at http://www-ssrl.slac.stanford.edu/exafspak.html.
(50) Ankudinov, A.; Rehr, J. Phys. Rev. B 1997, 56, R1712−R1716.
(51) Dimakis, N.; Bunker, G. Phys. Rev. B 2002, 65, No. 201103.
(52) Atkins, P.; Overton, T.; Rourke, J.; Weller, M.; Armstrong, F.;
Hagerman, M. Inorganic Chemistry, 5th ed.; Freeman: New York, 2010.
(53) Kau, L. S.; Spira-Solomon, D. J.; Penner-Hahn, J. E.; Hodgson,
K. O.; Solomon, E. I. J. Am. Chem. Soc. 1987, 109, 6433−6442.
(54) Perrin, B. S.; Ichiye, T. J. Biol. Inorg. Chem. 2013, 18, 103−110.
(55) Gibney, B. R.; Rabanal, F.; Reddy, K. S.; Dutton, P. L.
Biochemistry 1998, 37, 4635−4643.
(56) Shifman, J. M.; Gibney, B. R.; Sharp, R. E.; Dutton, P. L.
Biochemistry 2000, 39, 14813−14821.
(57) Gibney, B. R.; Isogai, Y.; Rabanal, F.; Reddy, K. S.; Grosset, A.
M.; Moser, C. C.; Dutton, P. L. Biochemistry 2000, 39, 11041−11049.
(58) Gibney, B. R.; Dutton, P. L. Adv. Inorg. Chem. 2001, 68, 409−
456.
(2) Fedurco, M. Coord. Chem. Rev. 2000, 209, 263−331.
(3) Groves, J. T. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3569−3574.
(4) Umena, Y.; Kawakami, K.; Shen, J.-R.; Kamiya, N. Nature 2011,
473, 55−60.
(5) Castiglione, N.; Rinaldo, S.; Giardina, G.; Stelitano, V.;
̀
Cutruzzola, F. Antioxid. Redox Signaling 2012, 17, 684−716.
(6) Klinman, J. P. Chem. Rev. 1996, 96, 2541−2562.
(7) Averill, B. A. Chem. Rev. 1996, 96, 2951−2964.
(8) Crichton, R. R.; Pierre, J. L. Biometals 2001, 14, 99−112.
(9) Solomon, E. I.; Szilagyi, R. K.; DeBeer George, S.; Basumallick, L.
Chem. Rev. 2004, 104, 419−458.
(10) Rosenzweig, A. C.; Sazinsky, M. H. Curr. Opin. Struct. Biol.
2006, 16, 729−735.
(11) MacPherson, I. S.; Murphy, M. E. P. Cell. Mol. Life Sci. 2007, 64,
2887−2899.
(12) Solomon, E. I.; Sundaram, U. M.; Machonkin, T. E. Chem. Rev.
1996, 96, 2563−2606.
(13) Solomon, E. I.; Ginsbach, J. W.; Heppner, D. E.; Kieber-
Emmons, M. T.; Kjaergaard, C. H.; Smeets, P. J.; Tian, L.; Woertink, J.
S. Faraday Discuss. 2011, 148, 11−39.
(14) Solomon, E. I.; Hadt, R. G. Coord. Chem. Rev. 2011, 255, 774−
789.
(15) Solomon, E. I.; Chen, P.; Metz, M.; Lee, S.-K.; Palmer, A. E.
Angew. Chem., Int. Ed. 2001, 40, 4570−4590.
(16) Hannan, J. P.; Busch, J. L. H.; Breton, J.; James, R.; Thomson, A.
J.; Moore, G. R.; Davy, S. L. J. Biol. Inorg. Chem. 2000, 5, 432−447.
́
(17) Chufan, E. E.; Prigge, S. T.; Siebert, X.; Eipper, B. A.; Mains, R.
E.; Amzel, L. M. J. Am. Chem. Soc. 2010, 132, 15565−15572.
(18) Evans, J. P.; Blackburn, N. J.; Klinman, J. P. Biochemistry 2006,
45, 15419−15429.
(19) Kline, C. D.; Mayfield, M.; Blackburn, N. J. Biochemistry 2013,
52, 2586−2596.
(20) Prigge, S. T.; Eipper, B. A.; Mains, R. E.; Amzel, L. M. Science
2004, 304, 864−867.
(21) Libby, E.; Averill, B. A. Biochem. Biophys. Res. Commun. 1992,
187, 1529−1535.
(59) Marshall, N. M.; Garner, D. K.; Wilson, T. D.; Gao, Y.-G.;
Robinson, H.; Nilges, M. J.; Lu, Y. Nature 2009, 462, 113−116.
(60) New, S. Y.; Marshall, N. M.; Hor, T. S. A.; Xue, F.; Lu, Y. Chem.
Commun. 2012, 48, 4217−4219.
(61) Peacock, A. F. A.; Iranzo, O.; Pecoraro, V. L. Dalton Trans.
2009, 2271−2280.
(22) Moura, I.; Moura, J. J. Curr. Opin. Chem. Biol. 2001, 5, 168−175.
(23) Pinho, D.; Besson, S.; Brondino, C. D.; de Castro, B.; Moura, I.
Eur. J. Biochem. 2004, 271, 2361−2369.
(24) Tocheva, E. I.; Rosell, F. I.; Mauk, A. G.; Murphy, M. E. P.
Science 2004, 304, 867−870.
(25) Jacobson, F.; Pistorius, A.; Farkas, D.; De Grip, W.; Hansson,
(62) Dieckmann, G. R.; McRorie, D. K.; Lear, J. D.; Sharp, K. A.;
DeGrado, W. F.; Pecoraro, V. L. J. Mol. Biol. 1998, 280, 897−912.
(63) Sundberg, R. J.; Martin, R. B. Chem. Rev. 1974, 74, 471−517.
(64) Prenesti, E.; Daniele, P. G.; Prencipe, M.; Ostacoli, G.
Polyhedron 1999, 18, 3233−3241.
O.; Sjolin, L.; Neutze, R. J. Biol. Chem. 2007, 282, 6347−6355.
̈
(26) Gray, H. B.; Malmstrom, B. G.; Williams, R. J. P. J. Biol. Inorg.
̈
Chem. 2000, 5, 551−559.
́
(27) Saab-Rincon, G.; Valderrama, B. Antioxid. Redox Signaling 2009,
11, 167−192.
(65) Tamilarasan, R.; Mcmillin, D. R. Biochem. J. 1989, 263, 425−
429.
(28) Matsuda, T.; Warshel, A. Biochemistry 1986, 25, 1675−1681.
(29) Kuila, D.; Fee, J. A. J. Biol. Chem. 1986, 261, 2768−2771.
(30) Varadarajan, R.; Zewert, T. E.; Gray, H. B.; Boxer, S. G. Science
1989, 243, 69−72.
(66) Ambundo, E. A.; Deydier, M.; Grall, A. J.; Aguera-vega, N.;
Dressel, L. T.; Cooper, T. H.; Heeg, M. J.; Ochrymowycz, L. A.;
Rorabacher, D. B. Inorg. Chem. 1999, 38, 4233−4242.
(67) Rorabacher, D. B. Chem. Rev. 2004, 104, 651−697.
(68) Bernardo, M. M.; Heeg, J.; Schroeder, I. R. R.; Ochrymowycz, L.
A.; Rorabacher, D. B. Inorg. Chem. 1992, 31, 191−198.
(31) Hellwig, P.; Behr, J.; Ostermeier, C.; Richter, O. M.; Pfitzner,
U.; Odenwald, A.; Ludwig, B.; Michel, H.; Mantele, W. Biochemistry
̈
1998, 37, 7390−7399.
(32) Baltzer, L.; Nilsson, H.; Nilsson, J. Chem. Rev. 2001, 101, 3153−
(69) Gray, H. B.; Malmstrom, B. G. Comments Inorg. Chem. 1983, 2,
̈
3163.
203−209.
(33) DeGrado, W. F.; Summa, C. M.; Pavone, V.; Nastri, F.;
Lombardi, A. Annu. Rev. Biochem. 1999, 68, 779−819.
(34) Bryson, J. W.; Betz, S. F.; Lu, H. S.; Suich, D. J.; Zhou, H. X.;
O’Neil, K. T.; DeGrado, W. F. Science 1995, 270, 935−941.
(35) Ghosh, D.; Pecoraro, V. L. Inorg. Chem. 2004, 43, 7902−7915.
(36) Tegoni, M.; Yu, F.; Bersellini, M.; Penner-Hahn, J. E.; Pecoraro,
V. L. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 21234−21239.
(70) Dong, S.; Ybe, J. A.; Hecht, M. H.; Spiro, T. G. Biochemistry
1999, 38, 3379−3385.
(71) Antonyuk, S. V.; Strange, R. W.; Sawers, G.; Eady, R. R.;
Hasnain, S. S. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 12041−12046.
(72) Tavares, P.; Pereira, A. S.; Moura, J. J. G.; Moura, I. J. Inorg.
Biochem. 2006, 100, 2087−2100.
L
dx.doi.org/10.1021/ja406648n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX