Model studies of azide binding to functional analogues of CcO

Article abstract of DOI:10.1021/ic702294n

N₃^- binding to a functional model of CcO is investigated in its Fe³⁺, Fe³⁺Cu⁺, and Fe ³⁺Cu²⁺ forms. A combination of EPR and FTIR indicates that N₃^- binds in a bridging mode in the bimetallic sites and signature N₃^- bands are identified for several forms of N₃^- binding to the site. The presence of the distal metal increases the binding affinity of N₃^-. This bridging enables antiferromagnetic interaction between the two metal centers in the Fe³⁺Cu²⁺ state, which results in an EPR-silent ground state.

Full text of DOI:10.1021/ic702294n

Inorg. Chem. 2008, 47, 2916-2918
Model Studies of Azide Binding to Functional Analogues of CcO
James P. Collman,* Abhishek Dey, Richard A. Decréau, and Ying Yang
Department of Chemistry, Stanford UniVersity, Stanford, California 94305
Received November 21, 2007
N3- binding to a functional model of CcO is investigated in its
Fe3+, Fe3+Cu+, and Fe3+Cu2+ forms. A combination of EPR and
FTIR indicates that N3- binds in a bridging mode in the bimetallic
sites and signature N3- bands are identified for several forms of
N3- binding to the site. The presence of the distal metal increases
the binding affinity of N3-. This bridging enables antiferromagnetic
interaction between the two metal centers in the Fe3+Cu2+ state,
which results in an EPR-silent ground state.
a and CuA reduced) are yet unexplored. Synthetic biomi-
metic model complexes provide a controlled environment
for studying these key interactions that take place in a protein
active site. Thus, there is a need for a systematic study of
N3- binding to the CcO active site model that will serve as
a reference for analyzing spectroscopic data obtained in
protein active sites.
Several synthetic CcO models have been reported in the
5
-
past decade, and one of them has been used to model N3
binding.6 Unfortunately, in the absence of FTIR data, the
presence of multiple N3 -bound forms, and the lack of O2
-
Cytochrome c oxidase (CcO) is the terminal electron donor
to oxygen in mitochondria, reducing it to H2O. This generates
a proton gradient that drives ATP synthesis.1 The O2
reduction occurs at the heterobimetallic active site of CcO
consisting of a heme a3 and a CuB center within 5 Å.2a–c
The reaction mechanism of this fundamentally important
reduction activity of this complex, it is hard to correlate those
results to the ones obtained in CcO. Recently, one synthetic
model complex has been shown to have O2-reducing activity,
under physiological conditions, comparable to the parent
enzyme (Scheme 1).7a–c In this study, we show that this
-
functional model can be used to investigate N3 binding to
-
enzyme and its interactions with small molecules (e.g., N3 ,
the CcO active site. We use EPR and FTIR techniques to
characterize different modes of N3- binding that helps gain
insight into the origin of noncompetitive inhibition of CcO
by these anionic ligands.8
CO, and NO) have been a focus of major research for several
decades.1,3 In particular, the nature of its interaction with
-
its inhibitor N3 has been investigated using Fourier trans-
form (FTIR), electron paramagnetic resonance (EPR), and
The monometallic Fe3+ complex can be synthesized by
oxidizing the Fe2+ complex with ferrocinum tetrafluoroborate
(Fc+) in dichloromethane (CH2Cl2). The mixed-valent (in
the active site) Fe3+Cu+ state has not been well characterized
in CcO9 or in any other model systems.7b It can be obtained
by adding 1 equiv of Cu2+ to a Fe2+ complex in a CH2Cl2
solution, whereupon FeII gets oxidized to Fe3+ and the
resulting Cu+ binds to the distal pocket of the model (Figure
S1 in the Supporting Information). It can also be synthesized
by adding 1 equiv of Cu+ to a Fe3+ complex. The addition
resonance Raman techniques.4a–c It is generally accepted that
-
one or two N3 ’s can bind either in a bridging or in a terminal
manner to the resting oxidized state of the active site
(Fe3+Cu2+). However, unambiguous assignment of the
spectroscopic data is generally complicated by spectroscopic
features from heme a and CuA centers (also present in the
-
enzyme). Furthermore, the role of the CuB center in N3
-
binding to the CcO active site and the possibility of N3
binding to a possible mixed-valent form (i.e., Fe3+Cu+ or
Fe2+Cu2+ forms of these complexes, not Fe3+Cu2+ with heme
*
To whom correspondence should be addressed. E-mail:
(5) (a) Liu, J.-G.; Naruta, Y.; Tani, F.; Chishiro, T.; Tachi, Y. Chem.
Commun. 2004, 120. (b) Liu, J.-G.; Naruta, Y.; Tani, F. Angew. Chem.,
Int. Ed. 2005, 44, 1836. (c) Kim, E.; Kamaraj, K.; Galliker, B.; Rubie,
N. D.; Moenne-LoCcOz, P.; Kaderli, S.; Zuberbuhler, A. D.; Karlin,
K. D. Inorg. Chem. 2005, 44, 1238.
jpc@stanford.edu.
(1) Ferguson-Miller, S.; Babcock, G. T. Chem. ReV. 1996, 96, 2889.
(2) (a) Iwata, S.; Ostermeier, C.; Ludwig, B.; Michel, H. Nature 1995,
376, 660. (b) Yoshikawa, S.; Shinzawa-Itoh, K.; Nakashima, R.;
Yaono, R.; Yamashita, E.; Inoue, N.; Yao, M.; Fei, M. J.; Libeu, C. P.;
Mizushima, T.; Yamaguchi, H.; Tomizaki, T.; Tsukihara, T. Science
1998, 280, 1723. (c) Fei, M. J.; Yamashita, E.; Inoue, N.; Yao, M.;
Yamaguchi, H.; Tsukihara, T.; Shinzawa-Ito, K.; Nakashima, R.;
Yoshikawa, S. Acta Crystallogr., Sect. D 2000, 56, 529.
(3) Collman, J. P.; Boulatov, R.; Sunderland, C. J.; Fu, L. Chem. ReV.
2004, 104, 561–88.
(6) Dallacosta, C.; Alves, W. A.; da Costa Ferreira, A. M.; Monzani, E.;
Casella, L. Dalton Trans. 2007, 2197.
(7) (a) Collman, J. P.; Sunderland, C. J.; Boulatov, R. Inorg. Chem. 2002,
41, 2282. (b) Collman, J. P.; Decréau, R. A.; Yan, Y.; Yoon, J.;
Solomon, E. I. J. Am. Chem. Soc. 2007, 129, 5794–5795. (c) Collman,
J. P.; Devaraj, N. K.; Decréau, R. A.; Yang, Y.; Yan, Y.-L.; Ebina,
W.; Eberspacher, T. A.; Chidsey, C. E. D. Science 2007, 315, 1565.
(8) N3- does not bind to the fully reduced active form of CcO. Petersen,
L. C. Biochim. Biophys. Acta 1977, 460, 299.
(4) (a) Vamvouka, M.; Muller, W.; Ludwig, B.; Varotsis, C. J. Phys.
Chem. B 1999, 103, 3030. (b) Tsubaki, M.; Yoshikawa, S. Biochem-
istry 1993, 32, 174. (c) Li, W. B.; Palmer, G. Biochemistry 1993, 32,
1833.
(9) Babcock, G. T.; Vickery, L. E.; Palmer, G. J. Biol. Chem. 1978, 253,
2400.
2916 Inorganic Chemistry, Vol. 47, No. 8, 2008
10.1021/ic702294n CCC: $40.75  2008 American Chemical Society
Published on Web 03/25/2008

COMMUNICATION
Figure 1. Absorption spectra of Fe3+, Fe3+Cu+, and Fe3+Cu2+ states and
Figure 2. EPR spectra of the Fe3+, Fe3+Cu+, and Fe3+Cu2+ states and
their N3--bound forms collected at 77 K in frozen CH2Cl2. The upper inset
shows the low-field region, and the lower inset shows the low-spin Fe3+
signals in the high-field region (three g’s are indicated by vertical lines).
their N3--bound forms.
Scheme 1. Representation of the Functional Model Complex in its
Fe3+ (No M2), Fe3+Cu+ (M2 ) Cu+), and Fe3+Cu2+ (M2 ) Cu2+
States (Metals Have Triflate Counterions)
)
-
The small EPR signal in the Fe3+Cu2+N3 complex is due to some
-
unconverted Fe3+N3
.
EPR data indicate that the azide binding to the monometallic
high-spin Fe3+ complex (Figure 2, solid black line and upper
inset) leads to a low-spin Fe3+ center (Figure 2, dashed black
line and lower inset) as suggested above by the blue shift of
the Soret band. Binding of N3- to the mixed-valent Fe3+Cu+
complex, which has the same EPR signal as the monometallic
Fe3+ complex, also leads to the growth of a new low-spin Fe3+
signal (Figure 2, green line and lower inset). The g values of
this signal are different from those obtained from the Fe3+N3-
complex. This implies that the azide binding to the low-spin
Fe3+ centers in these are different. The fully oxidized Fe3+Cu2+
state of the complex has a high-spin Fe3+ signal at 1150 G
(Figure 2, red line; 4.5 K EPR data in Figure S2 in the
Supporting Information) and a Cu2+ signal at 2600-3400 G.
Spin quantification of this Cu2+ signal against a standard at 77
K accounts for >95% of the sample, indicating that the Fe3+
and Cu2+ centers are magnetically uncoupled. This indicates
that in the fully oxidized Fe3+Cu2+ state there is no bridging
ligand between the two paramagnetic centers.11 On the other
hand, binding of N3- leads to disappearance of the Cu2+ and
Fe3+ EPR signals (Figure 2, dashed red). Thus, N3- binds as a
of 2 equiv of Cu2+ to the Fe2+ complex or 1 equiv of Fc+ to
the Fe3+Cu+ complex generates the Fe3+Cu2+ complex.
Binding N3- to the monometallic Fe3+ porphyrin complex
in CH2Cl2 shifts the Soret band from 415 to 424 nm, and
the Q band shifts from 527 to 534 nm (Figure 1, solid black
and dashed black lines). For the Fe3+Cu+ complex (Figure
1, solid green and dashed green lines), the Soret shifts from
406 to 423 nm upon N3 binding, while in the Fe3+Cu2+
-
state (Figure 1, solid blue and dashed blue lines), the Soret
shifts from 404 to 422 nm. These changes in the absorption
features upon N3- binding to Fe3+, Fe3+Cu+, and Fe3+Cu2+
states indicate that N3- ligates to the Fe3+ center in all three
cases. In all of the above cases, the red shift in the Soret
band is indicative of a change of the spin state of Fe3+ from
1
bridging ligand between the S ) /2 Fe3+ (indicated by the
absorption and EPR of the Fe3+N3- complex) and S ) 1/2 Cu2+
centers. This enables antiferromagnetic coupling between these
sites, leading to an EPR-silent ground state. Thus, this model
successfully reproduces the bridging interactions observed in
most CcO’s; i.e., while in its resting Fe3+Cu2+ state, there is
no bridging ligand and the heme a3 and CuB can be bridged by
an N3- ligand.2a,b The bridging of the two metals with a single
N3- implies that, even with the lack of crystallographic data, it
is safe to conclude that the distances between the two metals
in this model must be ∼6 Å as observed in the CcO active
site.
high spin to low spin upon N3 binding.6,10
-
Quantitative addition of N3- to these complexes indicates
that the monometallic Fe3+, the mixed-valent Fe3+Cu+, and
the fully oxidized Fe3+Cu2+ require 4, 2, and 1 equiv of N3
-
for complete binding, respectively. This is consistent with
cooperativity between the two metals in this bimetallic active
site model, where the presence of the distal metal and a
subsequent increase in its charge enhance anionic ligand
binding to the heme Fe.6 This could be due to the bridging
FTIR has been used extensively to study the interaction
-
10
nature of the N3 ligand and is evaluated below.
-
of N3- with ferric porphyrin models. The FTIR of the N3
bound Fe3+ complex shows a N3- asymmetric stretch at 2010
cm-1 (Figure 3, blue line), which corresponds to a low-spin
(10) (a) Byers, W.; Cossham, J. A.; Edwards, J. O.; Gordon, A. T.; Jones,
J. G.; Kenny, E. T. P.; Mahmood, A.; McKnight, J.; Sweigart, D. A.;
et al. Inorg. Chem. 1986, 25, 4767. (b) Neya, S.; Hada, S.; Funasaki,
N. I.; Umemura, J.; Takenaka, T. Biochim. Biophys. Acta 1985, 827,
157.
(11) EPR on the fully oxidized Fe3+Cu2+ species obtained in wet CH2Cl2
also shows the same EPR, indicating the lack of a bridging ligand.
Inorganic Chemistry, Vol. 47, No. 8, 2008 2917

COMMUNICATION
cm-1 in the fully oxidized Fe3+Cu2+N3- complex (Figure 3,
red line), where, from EPR, N3- bridges the Fe3+ and Cu2+
centers. With 15N3 , the 2068 cm-1 band shifts to 2054 and
-
2063 cm-1 and the 2090 cm-1 shoulder shifts to 2083 cm-1
(Figure S3 in the Supporting Information).
In summary, the biomimetic models studied here indicate
that N3- can bind in the Fe3+Cu+ (mixed-valent) as well as
Fe3+Cu2+ (resting oxidized) states of the CcO active site.
The presence of the distal metal increases the binding affinity
-
for N3 because of its bridging interaction. This enhances
N3 interaction with the Fe3+Cu+ and Fe3+Cu2+ states,
-
making it an efficient inhibitor of CcO. This bridging
between the metals results in an antiferromagnetically
coupled ground state in the fully oxidized state of the active
Figure 3. FTIR spectrum of the azide-bound complexes of the Fe3+
,
Fe3+Cu+, and Fe3+Cu2+ complexes (Y axis arbitrarily scaled). The shoulder
-
around 2010 cm-1 in both the Fe3+Cu+ and Fe3+Cu2+N3 complexes is
-
due to some unconverted Fe3+N3 (Figure S4 in the Supporting Informa-
-
site. The terminal N3 binding is characterized by an IR
tion).
vibration at 2010 cm-1, while the bridging N3- is character-
ized by vibrations at 2048 and 2068 cm-1 for the mixed-
valent and fully oxidized states, respectively. These results
should provide benchmark features for the analysis of FTIR
data obtained in a CcO active site.
Fe3+N3 species, as indicated by EPR data. Using a single
-
N15-substituted N3 (i.e., 15NNN-; Figure S3 in the Sup-
-
porting Information), two bands are observed at 1990 and
1998 cm-1.4a,c,12 N3 binding to the mixed-valent Fe3+Cu+
-
complex shows a major IR band at 2048 cm-1 (Figure 3,
green line). This band is 38 cm-1 shifted relative to the N3
-
Acknowledgment. This research has been funded by the
NIH under Grants 5R01 GM-17880-35 and 5RO1 GM69658-
4. Somdatta Ghosh and Prof. E. I. Solomon are acknowl-
edged for their help in EPR data collection. R.A.D. is
thankful for a Lavoisier Fellowship. Dr. A. Hosseini is
thanked for helpful discussions.
band in the Fe3+N3 complex. This is consistent with a
-
bridging N3- between the Fe3+ and Cu+ centers where more
electron density is shifted from its N-N π*-type highest
occupied molecular orbitals. This also parallels the small
differences in the g values in the EPR spectra of the Fe3+N3
-
complex and the mixed-valent Fe3+Cu+N3 complex. With
-
N15-substituted azide, this 2048 cm-1 shifts to 2033 cm-1
(Figure S3 in the Supporting Information). The νassym are
further blue-shifted to 2068 cm-1 with a shoulder at 2090
Supporting Information Available: The 4.5 K EPR spectra of
the Fe3+ complexes, the kinetic traces of the distal metal binding,
and the FTIR data in solution and with 15NNN azide. This material
is available free of charge via the Internet at http://pubs.acs.org.
(12) A single 15N-substituted azide should give rise to two νassym modes of
N3- because it could bind with either the 15N or 14N atom of the azide.
IC702294N
2918 Inorganic Chemistry, Vol. 47, No. 8, 2008