Heme-copper/dioxygen adduct formation relevant to cytochrome c oxidase: spectroscopic characterization of [(6L)FeIII-(O2(2-))-CuII]+.

Abstract

In the further development and understanding of heme-copper dioxygen reactivity relevant to cytochrome c oxidase O(2)-reduction chemistry, we describe a high-spin, five-coordinate dioxygen (peroxo) adduct of an iron(II)-copper(I) complex, [((6)L)Fe(II)Cu(I)](BArF(20)) (1), where (6)L is a tetraarylporphyrinate with a tethered tris(2-pyridylmethyl)amine chelate for copper. Reaction of 1 with O(2) in MeCN affords a remarkably stable [t(1/2) (rt; MeCN) approximately 60 min] adduct, [((6)L)Fe(III)-(O(2) (2-))-Cu(II)](+) (2) [EPR silent; lambda(max)=418 (Soret), 561 nm], formulated as a peroxo complex based on manometry (1:O(2)=1:1; spectrophotometric titration, -40 degrees C, MeCN), mass spectrometry {MALDI-TOF-MS: (16)O(2), m/z 1191 ([((6)L)Fe(III)-((16)O(2) (2-))-Cu(II)](+)); (18)O(2), m/z 1195}, and resonance Raman spectroscopy (nu((O-O))=788 cm(-1); Delta(16)O(2)/(18)O(2)=44 cm(-1); Delta(16)O(2)/(16/18)O(2)=22 cm(-1)). (1)H and (2)H NMR spectroscopy (-40 degrees C, MeCN) reveals that 2 is the first heme-copper peroxo complex which is high-spin, with downfield-shifted pyrrole resonances (delta(pyrrole)=75 ppm, s, br) and upfield shifted peaks at delta= -22, -35, and -40 ppm, similar to the pattern observed for the mu-oxo complex [((6)L)Fe(III)-O-Cu(II)](BAr(F)) (3) (known S=2 system, antiferromagnetically coupled high-spin Fe(III) and Cu(II)). The corresponding magnetic moment measurement (Evans method, CD(3)CN, -40 degrees C) also confirms the S=2 spin state, with mu(B)=4.9. Structural insights were obtained from X-ray absorption spectroscopy, showing Fe-O (1.83 A) and Cu-O (1.882 A) bonds, and an Fe...Cu distance of 3.35(2) A, suggestive of a mu-1,2-peroxo ligand present in 2. The reaction of 2 with cobaltocene gives 3, differing from the observed full reduction seen with other heme-Cu peroxo complexes. Finally, thermal decomposition of 2 yields 3, with concomitant release of 0.5 mol O(2) per mol 2, as confirmed quantitatively by an alkaline pyrogallol dioxygen scavenging solution.

Full text

Page 1

ORIGINAL ARTICLE
Reza A. Ghiladi Æ Hong-wei Huang
Pierre Moe ¨ nne-Loccoz Æ Jay Stasser
Ninian J. Blackburn Æ Amina S. Woods
Robert J. Cotter Æ Christopher D. Incarvito
Arnold L. Rheingold Æ Kenneth D. Karlin
Heme-copper/dioxygen adduct formation relevant to cytochrome
c oxidase: spectroscopic characterization of [(6L)FeIII-(O22?)-CuII]+
Received: 21 March 2004/ Accepted: 21 October 2004/Published online: 4 December 2004
? SBIC 2004
Abstract In the further development and understanding
of heme-copper dioxygen reactivity relevant to cyto-
chrome c oxidase O2-reduction chemistry, we describe a
high-spin, five-coordinate dioxygen (peroxo) adduct of
an iron(II)-copper(I) complex, [(6L)FeIICuI](BArF20)
(1), where6L is a tetraarylporphyrinate with a tethered
tris(2-pyridylmethyl)amine chelate for copper. Reaction
of 1 with O2in MeCN affords a remarkably stable [t1/2
(rt; MeCN)?60 min] adduct, [(6L)FeIII-(O22-)-CuII]+(2)
[EPR silent; kmax=418 (Soret), 561 nm], formulated as a
peroxo complex based on manometry (1:O2=1:1; spec-
trophotometric titration, ?40 ?C, MeCN), mass spec-
trometry {MALDI-TOF-MS:16O2, m/z 1191 ([(6L)FeIII-
(16O22?)-CuII]+);
Raman spectroscopy (m(O-O)=788 cm–1; D16O2/18O2=
44 cm–1; D16O2/16/18O2=22 cm–1).
spectroscopy (?40 ?C, MeCN) reveals that 2 is the first
heme-copper peroxo complex which is high-spin, with
downfield-shifted pyrrole resonances (dpyrrole=75 ppm,
s, br) and upfield shifted peaks at d= ?22, ?35, and
?40 ppm, similar to the pattern observed for the l-oxo
complex [(6L)FeIII-O-CuII](BArF) (3) (known S=2 sys-
tem, antiferromagnetically coupled high-spin FeIIIand
CuII). The corresponding magnetic moment measure-
ment (Evans method, CD3CN, ?40 ?C) also confirms
the S=2 spin state, with lB=4.9. Structural insights
were obtained from X-ray absorption spectroscopy,
showing Fe–O (1.83 A˚) and Cu–O (1.882 A˚) bonds, and
an Fe...Cu distance of 3.35(2) A˚, suggestive of a l-1,2-
peroxo ligand present in 2. The reaction of 2 with co-
baltocene gives 3, differing from the observed full
reduction seen with other heme-Cu peroxo complexes.
Finally, thermal decomposition of 2 yields 3, with con-
comitant release of 0.5 mol O2per mol 2, as confirmed
quantitatively by an alkaline pyrogallol dioxygen scav-
enging solution.
18O2, m/z 1195}, and resonance
1H and
2H NMR
Keywords Heme-copper Æ Iron(II)-copper(I) complex Æ
Peroxo complex Æ Mass spectrometry Æ
Resonance Raman spectroscopy Æ Dioxygen adduct Æ
Model compound
Introduction
As a member of the heme-copper oxidase superfamily,
cytochrome c oxidase (CcO) utilizes electrons provided
sequentially by cytochrome c to catalyze the 4e?/4H+
reduction of dioxygen to water [1, 2]. This membrane-
bound multimetallic enzyme couples this exergonic
Electronic Supplementary Material Supplementary material is
available in the online version of this article at http://dx.doi.org/
10.1007/s00775-004-0609-1
R. A. Ghiladi Æ K. D. Karlin (&)
Department of Chemistry,
The Johns Hopkins University,
Charles & 34th Streets, Baltimore,
MD 21218, USA
E-mail: karlin@jhu.edu
Tel.: +1-410-5168027
Fax: +1-410-5168420
H. Huang Æ P. Moe ¨ nne-Loccoz Æ J. Stasser Æ N. J. Blackburn
Environmental & Biomolecular Systems,
OGI School of Science & Engineering,
Oregon Health & Science University,
Beaverton, OR 97006, USA
A. S. Woods Æ R. J. Cotter
Department of Pharmacology and Molecular Sciences,
The Johns Hopkins School of Medicine,
Baltimore, MD 21205, USA
C. D. Incarvito Æ A. L. Rheingold
Crystallography Laboratory,
Department of Chemistry,
University of Delaware,
Newark, DE 19716, USA
Present address: A. L. Rheingold
Chemistry Department, University of California,
San Diego, La Jolla, CA 92093, USA
J Biol Inorg Chem (2005) 10: 63–77
DOI 10.1007/s00775-004-0609-1

Page 2

process to the movement of four additional protons
across the mitochondrial membrane, resulting in a net
translocation of four charges. The proton pumping
mechanism of CcO generates the membrane potential
used by ATP synthase to drive the formation of ATP
from ADP, and, in light of its biologically important
relevance, has been the subject of numerous and intense
biophysical and spectroscopic investigations.
Confirming earlier biochemical and molecular bio-
logical suppositions are recent X-ray crystallographic
studies [3, 4, 5, 6, 7, 8, 9, 10, 11], which show that di-
oxygen reduction occurs at an active site consisting of a
heme a3-CuBheterobinuclear center with an iron–copper
separation of 4.5–5.2 A˚, depending on the protein
derivative (Fig. 1). As is generally but not fully agreed
upon, dioxygen binding and reduction occurs as follows
[12]. An initial CuB-O2 interaction [1] is followed by
transfer of O2to heme a3, forming an oxy ferrous (FeIII-
superoxo) observable intermediate (A). The next inter-
mediate detected has already undergone O–O cleavage
[12, 13], giving the oxy-ferryl FeIV=O species (P) and
CuBII-OH, with either tyrosinate or tyrosyl radical
(Tyr•), depending on the enzyme form used for the O2
chemistry. Rapid O–O cleavage seems an attractive
mechanistic feature as such a reaction course bypasses
the formation of potentially cytotoxic (if leaked) super-
oxide (as O2?or HO2) or peroxide (as HO2?or H2O2)
products. However, (hydro)peroxide metal-bridged or
associated FeIII-O-O(H)...CuBtransients are discussed,
based on structural, spectroscopic, and theoretical con-
siderations [7, 14, 15]. The post-translationally formed
[16, 17] amino acid His-Tyr cross-link (Fig. 1) may be
key in the reduction mechanism as a hydrogen atom
(i.e., proton plus electron) donor, and/or as an impor-
tant element of the proton-pumping mechanism [18, 19,
20]. Subsequent electron/proton transfer to P results in
reduction of Tyr• or protonation of CuBII-OH to give
intermediate F, with FeIV=O and CuBII-OH2compo-
nents. Modified descriptions of P and F exist [21, 22, 23].
Further reduction-protonation gives an ‘‘oxidized’’
FeIII...CuIIintermediate O, which is the same or similar
[24] to the as-isolated ‘‘resting’’ enzyme.
While these considerable advances in understanding
the fundamental processes performed by CcO have oc-
curred, there has also been intense interest from inor-
ganic chemists concerned with providing coordination
chemistry insights into CcO heme a3-CuB physical
properties and reactivity [25, 26, 27]. Considerable effort
has focused on structural-spectroscopic modeling of the
oxidized ‘‘resting’’ state and the generation of FeIII-X-
CuIIcomplexes, notably with X=CN?, RCO2?, oxo
(O2?), and OH?[25, 27]. Dioxygen reactivity studies
have recently undergone considerable advancement, as
approached either via electrocatalytic O2reduction with
heme-Cu assemblies [26, 27], or by the generation of O2
adducts of discrete heme-Cu complexes or heme and Cu-
ligand components [28, 29, 30, 31, 32, 33, 34, 35, 36, 37].
Recent highlights are given in Fig. 2. Collman et al. [36]
have generated a complex I with a heme-superoxide
(FeIII-O2–) adduct which is stable in the presence of the
neighboring Cu(I) complex; this mimics the initial O2
adduct A intermediate in the CcO O2-reduction cycle.
Likewise, we have spectroscopically and/or kinetically
observed solutions containing both Fe-O2and CuIspe-
cies which exist prior to forming l-peroxo FeIII-(O22–)-
CuIIproducts [30, 32, 34, 35]. An exciting development
is Naruta and co-workers [37] crystal structure deter-
mination of binuclear complex II, demonstrating that an
Fig. 1 Rasmol depiction of the reduced heme a3-CuB binuclear
active site from bovine cytochrome c oxidase. Fe...Cu=5.19 A˚.
Coordinates (1OCR) were obtained from the Protein Data Bank
(Brookhaven, NY)
Fig. 2 Diagrams of recently prominent heme-copper dioxygen
adducts: I is a heme-superoxo (FeIII-O2–)-CuIcompound, while
II–IV are heme (FeIII)-peroxo-CuIIcomplexes, II being a structur-
ally characterized compound. See text for further discussion and
references
64

Page 3

g2:g1FeIII-(O22–)-CuIIcoordination is a viable structure,
at least in a high-spin complex. Our group has carried
out the solution characterization of a number of O2
adducts of reduced heme and Cu components, including
high-spin
l-peroxo FeIII-(O22–)-CuII
[(F8TPP)FeIII-(O22?)-CuII(TMPA)]+
[(F8TPP)FeIII-(O22?)-CuII(LMe2N)]+
{TMPA=tris(2-pyridylmethyl)amine, LMe2N=N,N-bis
[2-(2-(4-dimethylamino)pyridyl)ethyl]methylamine}.
combination, studies on these complexes reveal that the
O–O bond strength (as indicated by mO–O) can be dra-
matically lowered (by ?40–50 cm–1) by changing the
peroxo binding from a tetradentate to tridentate Cu
chelate [35].
Thus, such investigations involving systematic studies
with changes in ligand environment provide insights into
the O2-binding process and properties of resulting
adducts in a heme/Cu environment, as mentioned. To
these ends, and to elicit more control by reducing the
probability of intermolecular O2-binding chemistry, we
have also been studying heme/Cu/O2chemistry utilizing
heterobinucleating ligands such as6L (Scheme 1). As a
close analogue of [(F8TPP)FeIII-(O22?)-CuII(TMPA)]+
(III) (Fig. 2), we reported the preliminary characteriza-
tion of the first high-spin heme/Cu l-peroxo complex,
[(6L)FeIII-(O22–)-CuII]+(2), which forms from the reac-
tion of dioxygen with the fully reduced compound
[(6L)FeIICuI]+(1) (Scheme 1). Complex 2 thermally
transforms to l-oxo species [(6L)FeIII-O-CuII]+(3) [30].
The focus of this report is to provide a detailed char-
acterization of this system, especially new insights into
the spectroscopy, structure, and chemical behavior of
[(6L)FeIII-(O22–)-CuII]+(2).
complexes
[34]
(IV)
(III) and
[35]
In
Materials and methods
Synthetic details and spectroscopic characterization
have been previously reported for [(6L)FeIICuI]+(1) and
[(6L)FeIII-(O22–)-CuII]+(2); however, additional experi-
mental procedures, as well as updated protocols for
NMR, EPR, and resonance Raman spectroscopic and
MALDI-TOF spectrometric
found in the Supplementary material.
Reagents and solvents used were purchased from
commercial sources and were of available reagent quality
unless otherwise stated. All solvents were distilled under
argon prior to use. Dichloromethane (CH2Cl2), aceto-
nitrile (MeCN), and heptane were distilled directly from
CaH2. Propionitrile (C2H5CN, EtCN) was first distilled
over P4O10, then refluxed and distilled over CaH2. Tet-
rahydrofuran (THF) was distilled from Na/benzophe-
none. Preparation and handling of air-sensitive materials
was carried out under an argon atmosphere using stan-
dard Schlenk techniques. Solvents and solutions were
deoxygenated by either freeze–pump–thaw cycles (5·), or
by bubbling argon (>25 min) directly through the
solution. Solid samples were stored and transferred, and
samples for spectroscopic characterization were pre-
pared, in an MBraun LabMaster 130 inert atmosphere
(<1 ppm O2, <1 ppm H2O) glovebox under a nitrogen
atmosphere. O2used was of ultra-pure grade (99.994%;
WSC, Baltimore, Md., USA).
investigations, can be
Dioxygen-uptake/spectrophotometric titration
of [(6L)FeIICuI](BArF20) (1)
In the glovebox, to an air-free cuvette assembly with
side-arm stopcock and female 14/20 joint, was added a
precise quantity of 1 (1 mg in 5 g MeCN) prepared via
stock solution and serial dilution. A 14/20 syringe needle
adapter (Chemglass, CG-1036) equipped with three
septa [rubber septum (Chemglass, CG-3022-05), flat
Teflon septum spacer (Chemglass, MW-104-01), and
inverted rubber septum (Aldrich, Z12436-2)] were wired
together and used to seal the cuvette from the atmo-
sphere. The cuvette assembly was then removed from the
glovebox, placed into the modified low-temperature UV-
visible spectrophotometer, and cooled to ?40 ?C.
A commercially available 1.01±0.02 mol% O2 in
argon gas mixture was slowly passed through a two-side-
armed Schlenk flask equipped with rubber septum and
oil bubbler, which allowed for continuous flow of the
1% O2gas mixture at atmospheric pressure. Hamilton
gas-tight syringes (250 and 1000 lL) equipped with
three-way syringe purge valves (for argon purging of the
syringe and needle) were used to transfer exact molar
quantities of dioxygen to the cooled cuvette assembly.
Typical transfer ratios included 0.5 equiv dioxygen
(650 lL O2 per 1 mg of 1), followed by 0.25 equiv
(325 lL) and three successive 0.1 equiv (130 lL) addi-
tions. Each aliquot of dioxygen was bubbled directly
Scheme 1
65

Page 4

into the solution via the syringe needle, and was
accompanied by gentle shaking of the cuvette assembly
and UV-visible spectroscopic monitoring. When no
further spectral changes were detected after each suc-
cessive addition of dioxygen, the next partial equivalent
was added. In this manner, it was established that full
formation of the oxygen intermediate of 1 occurred be-
tween 0.95 equiv (1235 lL 1.01% O2 in argon) and
1.05 equiv (1365 lL 1.01% O2in argon) of dioxygen.
No spectral changes were observed from direct injection
of dioxygen (>100 equiv) after the total 1.05 equiv were
added. Furthermore, a second cuvette filled with an
identical amount of the same stock solution of 1 used for
the titration was directly injected with a large excess of
dioxygen, and the spectrum obtained for this was nearly
superimposable (with respect to kmaxand absorbance) to
that obtained from the spectrophotometric titration.
Finally, a control experiment was performed in which 1
was injected a total of five times with 500 lL of oxygen-
free argon in a manner identical to that above, with no
detectable changes in the spectrum of 1 observed.
Dioxygen evolution of [(6L)FeIII-(O22?)-CuII]
(BArF20) (2)
An alkaline pyrogallol solution was used to detect di-
oxygen evolution upon decomposition of [(6L)FeIII-
(O22–)-CuII](BArF20) (2) to [(6L)FeIII-O-CuII](BArF20)
(3), and was prepared in the following manner. A 100-
mL Schlenk flask was specially modified via attachment
of a quartz cuvette, to which was added pyrogallol
(4.0 g), 25 mL of a deoxygenated 50% KOH (aq) solu-
tion, and a stir bar. The faint yellow solution was placed
under vacuum, its UV-visible spectrum recorded, and
transferred to the glovebox.
A calibration curve of absorbance (400 nm) versus
O2(moles) was determined in the following manner. In
the glovebox, the alkaline pyrogallol solution was
capped with the multiple-septa apparatus used in the
dioxygen uptake experiment (see above), and then
transferred outside the glovebox. Using a 2.5-mL
Hamilton gas-tight syringe equipped with a three-way
syringe purge valve connected to the argon line of a
Schlenk manifold, 1.5 mL of air was added by direct
injection into the alkaline pyrogallol solution using a
thin gauge needle. Rapid stirring accompanied by
periodic shaking of the Schlenk cuvette flask over a
period of 20 min followed, after which the UV-visible
spectrum of the pyrogallol solution was recorded. This
process was repeated three more times, such that a
total of 6.0 mL air (0.250 mmol O2 total) had been
added to the alkaline pyrogallol solution. Upon addi-
tion of each 1.5 mL aliquot, the solution turned darker,
such that at the end of the experiment a golden brown
solution had formed. By plotting the absorbance (at
400 nm) versus moles of O2added, a known calibration
curve was established. This process was repeated a
second time, thereby creating a calibration curve based
on two independent runs: absorbance (400 nm)=
(0.0716·mL O2)+0.026.
Dioxygen evolution of the heme-peroxo-Cu complex
was determined in the following manner. In the glove-
box, to a 25-mL Schlenk flask equipped with a rubber
septum was added 20 mg of 1 dissolved in 8 g MeCN.
The solution was transferred outside the glovebox,
where it was cooled in a dry-ice/MeCN cold bath. The
solution was oxygenated by direct bubbling of ultra-
high-purity O2via a syringe needle, and allowed to stand
in the cold bath for several minutes to ensure complete
formation of the of [(6L)FeIII-(O22–)-CuII](BArF20)
complex. The rubber septum was replaced with a glass
stopper, and the entire solution was degassed via five
freeze–pump–thaw cycles to remove excess oxygen.
After the last freeze–pump–thaw cycle, the Schlenk flask
was placed under an inert atmosphere and carefully
transferred back into the glovebox. Using hose clamps, a
short length of vacuum tubing was used to attach
the sidearm of the Schlenk cuvette flask containing the
alkaline pyrogallol solution (under vacuum) to the
sidearm of the 25-mL Schlenk flask containing the heme-
peroxo-Cu complex (under inert atmosphere). After
several hours at room temperature, the side-arm stop-
cock valves were opened, and the headspace of the 25-
mL Schlenk flask was drawn into the evacuated Schlenk
cuvette flask containing the alkaline pyrogallol, which
slowly turned from a faint yellow to a golden brown in
color. The two flasks were allowed to stir overnight,
open to each other via the vacuum tubing connecting
their side-arm stopcocks. This ensured complete reaction
of any evolved dioxygen from the [(6L)FeIII-(O22–)-
CuII](BArF20) decomposition pathway with the alkaline
pyrogallol solution.
X-ray absorption data collection and analysis
XAS data were collected at the Stanford Synchrotron
Radiation Laboratory (SSRL) on beam line 7.3 (BL 7.3)
operating at 3.0 GeV with beam currents between 100
and 50 mA. An Si220 monochromator with 1.2-mm slits
was used to provide monochromatic radiation in the 7–
10 keV energy range. Harmonic rejection was achieved
by detuning the monochromator by 50% at the end of
the scan. The samples were measured as frozen glasses in
acetonitrile at 11–14 K in fluorescence mode using a 13-
element Ge detector. To avoid detector saturation, the
count rate of each detector channel was kept below
110 kHz while the rise in fluorescent counts through the
edge was kept below 20 kHz per channel. Under these
conditions, no dead-time correction was necessary. The
summed data for each detector channel were then in-
spected, and only those channels that gave high-quality
backgrounds free from glitches, drop outs, or scatter
peaks were included in the final average. Data averaging
and background subtraction were carried out using the
program EXAFSPAK [George GN (1998) http://www-
ssrl.slac.stanford.edu/exafspak.html]. Data analysis was
66

Page 5

performed using the program EXCURV98 as previously
described [38]. Single and multiple scattering pathways
were estimated for the outer-shell interactions of pyr-
role, pyridine, and acetonitrile groups. Metrical details
of multiple scattering pathways were initially derived
using idealized geometries of the pyrrole, pyridine, and
acetonitrile groups, but these were subsequently allowed
to vary by up to 5% during refinement. The goodness of
fit was estimated using a fitting parameter F defined as:
F2¼1
N
X
n
i¼1
k6Datai? Modeli
ðÞ2
ð1Þ
Results and discussion
Synthesis and spectroscopic characterization
of [(6L)FeIICuI](BArF20) (1)
Full synthetic experimental details are provided in the
Supplementary material. The6L ligand is composed of a
TMPA-like Cu chelate appended to the periphery of a
fluorinated tetraphenylporphyrin. The use of the tripo-
dal tetradentate TMPA unit was justified given that the
reaction of O2with [(TMPA)Cu(MeCN)]+complexes is
very well understood in a variety of solvents [39, 40, 41,
42, 43, 44]. Similarly, F6TPPH2-OH was chosen as the
porphyrin to which the TMPA unit would be tethered
since the dioxygen reactivity of the nearly identical
(F8TPP)FeIIcomplex is also known [45]. Thus, the6L
heterobinucleating ligand is composed of mononuclear
Cu and Fe components which have independently
known kinetics and thermodynamics of dioxygen reac-
tivity. Furthermore, the
formation of a heterobinuclear metal complex such that
the copper and iron metal centers are held in close
proximity to each other to facilitate an intramolecular
heme-Cu/O2reactions.
6L ligand architecture allows
Dioxygen reactivity of [(6L)FeIICuI]+(1)
in MeCN: UV-visible spectroscopy
As reported earlier [46], addition of dioxygen to the re-
duced complex [(6L)FeIICuI]+(1) in THF at room
temperature immediately gives the l-oxo complex
[(6L)FeIII-O-CuII]+(3) (confirmed by UV-visible and
NMR spectroscopies), in an O–O bond cleavage reac-
tion which is crudely biomimetic, since the l-oxo atom
(deprotonated water) was shown to be derived from
dioxygen. However, when MeCN or acetone are em-
ployed as reaction solvents, UV-visible spectroscopic
monitoring (MeCN, room temperature) revealed that
prior to forming the l-oxo complex [(6L)FeIII-O-CuII]+
(3) [kmax (?, M?1cm?1)=438 (Soret: 147,000), 556
(13,500) nm], bubbling of dioxygen directly through a
solution of [(6L)FeIICuI]+(1) [kmax(?, M?1cm?1)=423
(Soret: 225,000), 529 (21,000), 554 (sh) nm] leads to the
irreversible formation of a new complex (2) [kmax(?,
M?1cm?1) = 418 (Soret: 126,000), 561 (12,500) nm]
(Fig. S1) having remarkable thermal stability [t1/2
(rt)?60 min in MeCN], which we have formulated [46]
as the peroxo-level dioxygen adduct [(6L)FeIII-(O22–)-
CuII]+(2) (Scheme 1). When oxygenation was per-
formed at low temperature, e.g. ?40 ?C for MeCN, the
stability of 2 was increased to over 24 h.
In contrast to the room temperature irreversible di-
oxygen-binding behavior of 1 in MeCN, the mononu-
clear heme-only and copper-only analogues (separately)
exhibit substantially different O2 reactivity. At room
temperature, the heme-only complexes (6L)FeII[47] and
(F8TPP)FeII[45] both react with dioxygen without for-
mation of detectable dioxygen adducts [slow formation
of the decomposition ferric-hydroxy products, PFeIII-
OH (UV-visible: 408–412 (Soret) and 566–71 nm), is
observed]. Furthermore, room temperature oxygenation
of [(TMPA)Cu(MeCN)](ClO4) also results in immediate
decomposition [48]. Thus, it appears from the UV-visi-
ble spectroscopic study that [(6L)FeIICuI]+(1) not only
possesses markedly different oxygenation behavior from
its mononuclear analogue components, but that the
presence of both metal centers leads to the stabilization
of the resulting dioxygen adduct.
X-ray crystallography was also used to confirm 3 as
the intramolecular l-oxo product formed directly from
the reduction of dioxygen by [(6L)FeIICuI]+(1) (Fig. 3).
The structure of 3 exhibits a nearly linear Fe–O–Cu
moiety, very short Fe–O (1.748 A˚) and Cu–O (1.853 A˚)
bond distances, and a typical high-spin structure with Fe
out of the porphyrin plane by 0.5711 A˚. An —Fe–O–Cu
angle of 173.0? and a Fe...Cu separation of 3.600 A˚were
also found. These values are all in very close agreement
with the previously reported [46] structure for [(6L)FeIII-
O-CuII]+(3) formed in an acid–base reaction from the
addition of a cupric salt to (6L)FeIII-OH in the presence
of an organic base. The very minor differences between
Fig. 3 ORTEP diagram (30% ellipsoids) of [(6L)FeIII-O-CuII]+(3).
Selected bond lengths (A˚) and angles (?): Fe–O1=1.748(3), Cu–
O1=1.853(3), Fe–N (N1–N4)=2.102–2.134, Cu–N5=2.382(4),
Cu–N6=1.972(5), Cu–N7=1.975(6), Cu–N8=2.089(4), —Fe–O–
Cu=173.0(3)
67

Page 6

these two structures exist due to changes in the solvent
system employed for crystallization and in the counter-
ion present1. Additional details (crystal data and struc-
ture refinement) can be found in the Supplementary
material.
O2-uptake stoichiometry of [(6L)FeIICuI]+(1)
For all spectroscopic experiments described herein,
oxygenation of [(6L)FeIICuI]+(1) was performed under
conditions of a large molar excess of dioxygen. How-
ever, it was imperative to determine the stoichiometry of
the O2?uptake reaction in order to confirm the 1:1 di-
oxygen-adduct formulation of [(6L)FeIII-(O22–)-CuII]+
(2). Given the large amounts of material required by a
traditional manometric experiment, a more convenient
UV-visiblemonitored spectrophotometric
using sub-molar equivalents of dioxygen was employed
to determine the overall reaction stoichiometry of di-
oxygen uptake (Fig. 4).
At 230 K, [(6L)FeIICuI]+(1) exhibits UV-visible
features at 414 (sh) and 423 (Soret) nm. Syringe addition
of a 0.5 equiv aliquot of dioxygen (prepared from a
1.01% O2 in argon gas mixture, see Materials and
methods section) results in the partial loss of the 423 nm
Soret band (Fig. 4a), while an additional injection of
0.25 equiv O2 aliquot (0.75 equiv O2 total) led to a
further loss of the 423 nm Soret intensity with con-
comitant with the broadening of the 414 nm feature
(Fig. 4b). Three successive injections of 0.1 equiv O2
each (corresponding to 0.85, 0.95, and 1.05 equiv total
O2added) yielded a progressively sharper and more in-
tense Soret band at 417 nm, such that the final spectrum
obtained (at 1.05 equiv total O2added), with kmaxat 417
and 561 nm, matches without discrepancy the spectrum
ofauthentic[(6L)FeIII-(O22?)-CuII]+
230 K). Bubbling the solution with a large excess
(>100 equiv) of dioxygen showed no further spectral
changes, indicative of a final dioxygen-uptake stoichi-
ometry of 0.95–1.05 equiv O2per [(6L)FeIICuI]+(1).
titration
(2) (MeCN,
Homonuclear dioxygen adducts ruled out
While this stoichiometry is consistent with an intramo-
lecular O2adduct, it alone does not necessarily rule out
the possibility of a heme-only [PFeIII-(O2–) superoxo:
414–418 (Soret), 536–537 nm; or PFeIII-(O22–)-FeIIIP
intermolecular peroxo: 414–418 (Soret), 535–536 nm]
adduct [45, 47]. Copper-only [(L)CuII-(O2–) superoxo:
410 nm; or (L)CuII-(O22–)-CuII(L): 525, 610 nm] [39, 42,
48] adducts could also form upon oxygenation of 1.
However, given the large spectral differences between the
homonuclear dioxygen adducts and 2, these possibilities
were excluded. Additional support for this premise came
from resonance Raman spectroscopy (see below).
Mass spectrometry of [(6L)FeIII-(O22–)-CuII]+(2)
MALDI-TOF mass spectrometry proved to be very
useful for the study and characterization of the reduced
andoxygenated complexes
‘‘empty-tether’’ complex (6L)FeIIgave a parent ion at
1095 [m/z (M+H+)+], a loss of 17 (hydroxide) from the
(6L)FeIII-OH precursor (which exhibits m/z 1112; see
Materials and methods section). Cuprous ion insertion
described herein. The
Fig. 4 UV-visible spectrophotometric titration of O2uptake for
[(6L)FeIICuI]+(1) in MeCN at ?43 ?C
1Extremely minor differences exist between the two structures of
complex 3. These are due to the use of different counterions and
solvent systems employed for crystallization. Here, we crystallized
3 after direct reaction with dioxygen, employing the BarF20–anion
[i.e., B(C6F5)4–] and an MeCN/toluene solvent system, yielding
[(6L)FeIII-O-CuII][B(C6F5)4]Æ3toluene. Previously, 3 was formed
from an acid–base assembly reaction using a THF/heptane solvent
system and employing the (so-called) BArF–anion, yielding
[(6L)FeIII-O-CuII][B(C8H3F6)4], with 4–8 disordered heptane mol-
ecules (and perhaps some THF) which were not located, but in-
ferred by refinement with PLATON/SQUEEZE. That structure
gave Fe–O=1.750(4), Cu–O=1.848(4), and Fe...Cu=3.586 dis-
tances (A˚), and —Fe–O–Cu of 171.1(3)?
68

Page 7

into the copper chelate of (6L)FeIIyielded [(6L)FeII-
CuI]+(1), concomitant with an increase in mass (as
expected) to 1159 [m/z (M?BArF20–)+]2. Furthermore,
1 showed a decrease in the fragmentation pattern of the
pyridyl arms of the Cu chelate as compared to the
‘‘empty-tether’’ complexes of (6L)FeIIand (6L)FeIII-OH,
consistent with the insertion of copper into the tethered
TMPA moiety. Mass spectra were also obtained for the
16O2 and
CuII]+(2) (Fig. 5). These were carried out by first oxy-
genating [(6L)FeIICuI]+(1) with either16O2or18O2in
MeCN at ?40 ?C, followed by evaporation of the
resulting dioxygen adduct onto the MALDI-TOF-MS
sample holder at room temperature.
Oxygenation of [(6L)FeIICuI]+
(Fig. 5a) resulted in loss of the 1159 m/z peak, and the
formation of three new peaks at higher m/z. Of these,
the largest intensity peak (1175 m/z) corresponds to the
decomposition product [(6L)FeIII-O-CuII]+(3), possibly
formed in the mass spectrometer, but more likely in-
curred as a result of the time required for sample
preparation and handling at room temperature. As
expected, oxygenation with18O2gas caused a shift in
the mass of this peak by 2 amu (1177 m/z). While the
l-oxo complex was the largest intensity peak, a strong
peak at 1191 m/z, corresponding to [(6L)FeIII-(O22–)-
CuII]+(2), was also observed. This peak also shifted
upon18O2substitution, to 1195 m/z, a shift of 4 amu
(Fig. 5b, corresponding to the difference in mass be-
tween bound16O2and18O2)3. The least intense of the
isotopically sensitive peaks appeared at 1208 m/z (16O2
oxygenation), which shifted by 6 amu to 1214 m/z
upon18O2substitution. While this peak does not cor-
respondtoeither[(6L)FeIII-(O22–)-CuII]+
[(6L)FeIII-O-CuII]+(3), it is an oxygenation product
whose mass indicates three additional oxygen atoms
over that of the starting complex 1. While we are un-
sure about the identity of the species corresponding to
this 1208 m/z peak, it may represent a transient inter-
mediate formed during the decomposition of 2 (form-
ing 3), or it might represent a fleeting species not
18O2 dioxygen adducts of [(6L)FeIII-(O22–)-
(1) with
16O2
(2)or
relevant to the solution chemistry described here and
formed only in the mass spectrometer.
Resonance Raman spectroscopy
of [(6L)FeIII-(O22–)-CuII]+(2)
The
[(6L)FeIICuI]+(1) with dioxygen gas was characterized
by resonance Raman spectroscopy (Fig. 6) using a 413-
nm excitation coinciding with the Soret absorption
maximum of 2. Spectra from different dioxygen isotope
labeling experiments were normalized with the intense
intermediate formedfromthe reaction of
Fig. 5 MALDI-TOF mass spectra of [(6L)FeIII-(O22–)-CuII]+(2)
using (a)16O2and (b)18O2
Fig. 6 Resonance Raman spectra (413 nm excitation, rt, MeCN
solvent) of [(6L)FeIII-(O22–)-CuII]+(2), formed by oxygenation of
[(6L)FeIICuI]+(1) in MeCN at ?40 ?C using16O2(A),16/18O2(B),
and18O2(C). Difference spectra16O2?18O2(D) and16O2?16/18O2
(E) are also presented
2As a general feature of MALDI-TOF mass spectrometry, the lack
of peak resolution leads to the observed mass numbers being the
averaged one of the cluster peaks, different from the base peak.
This leads to systematic deviations in the observed peak from the
calculated values. For example, an ion of the reduced form
[1?BArF20]+should give the following isotope distribution: m/z
1155 (rel int 6%), 1156 (4), 1157 (100), 1158 (76), 1159 (71), 1160
(40), 1161 (14), 1162 (2). However, we observe one unresolved peak
and its peak maximum was m/z=1159. In a similar way, the16O2
and
m/z=1189 and 1193, respectively. However, the observed data
yielded peaks at m/z=1191.5 and 1195.6, respectively. The
decomposition product 3 is calculated to give m/z=1173 (16O) and
1175 (18O); observed, 1174.9 and 1177.1 The systematic differences
from the calculated values by ?2 mass units could come from the
lack of accuracy and peak resolution on the applied mass spec-
trometric measurement, and we thank a reviewer for bringing this
to our attention
3See footnote 2
18O2adducts derived from 2 should give the base peaks at
69

Page 8

phenyl vibrational mode at 898 cm?1. An16O2(Fig. 6A)
minus18O2(Fig. 6C) difference spectrum (Fig. 6D) sin-
gles out a unique isotope-sensitive band at 788 cm?1,
shifting by ?44 cm?1with18O2. To allow a definitive
assignment of the 788 cm?1feature to a peroxo m(O–O)
stretching vibration rather than a ferryl-oxo m(Fe=O)
mode, the experiments were repeated using a 1:2:1
mixture of
the positive signal at 788 cm?1and its negative
counterpart at 744 cm?1, the isotopic mixture minus
pure
?766 cm?1, assigned to m(16O–18O) (Fig. 6E). The
observation of this third signal excludes the m(Fe=O)
possibility, since only two stretching frequencies would
be expected for a ferryl species. Although the latter band
appears to be composed of two components on either
side of the 766 cm?1frequency, the intensity ratio be-
tween the three m(O–O) of the difference spectrum
(?1:?2:+3) corresponds to the ratio of the different
isotope gas and argues against a splitting of the
m(16O–18O). The absence of a splitting of the mixed-
isotope signal is consistent with a symmetrically bound
peroxide ligand, in either a l-1,2 (end-on) or l-g2:g2
(side-on) geometry, but rules out a l-1,1 bridging mode
where the two oxygen atoms would have quite different
chemical environments (as occurs in the non-heme di-
oxygen carrier hemerythrin [49]). Furthermore, por-
phyrin ligand vibrations consistent with the presence of
a five-coordinate high-spin ferric heme were observed,
suggesting that 2 lacks a sixth axial base ligand.
The m(O–O)of [(6L)FeIII-(O22–)-CuII]+(2) [788 cm?1;
D16O2/18O2=44 cm–1; D16O2/16/18O2=22 cm–1] is in the
range expected for a peroxide-level dioxygen adduct.
Other metal-bound peroxide complexes have been
characterized, including end-on and side-on bound di-
16O2:16O18O:18O2(Fig. 6B). In addition to
18O2
16O2difference spectrum shows a new signal at
oxygen adducts (Table 1): (1) heterobinuclear heme-pe-
roxo-Cucomplexessuch
[(F8TPP)FeIII-(O22–)-CuII(TMPA)]+
[(TMP)FeIII-(O22–)-(5MeTPA)CuII]+
[(PTACN)FeIIICuII-(O22–)]+
ane-capped porphyrinate)
(O22–)]+[PTPA=tris(2-pyridylmethyl)amine-linked TPP]
[29, 33]; (2) homobinuclear g2:g2-peroxo-dicopper(II)
complexes, [{(L)CuII}2-(O22–)]2+
hemocyanin model complexes of Kitajima and co-
workers [51, 52] and those of Karlin and co-workers [53,
54, 55]; (3) mononuclear anionic g2-heme-peroxo com-
plexes, [(P)FeIII-(O22–)]–[56, 57, 58]. The possibility of 2
as a TMPA-based copper-only intermolecular peroxide
adduct can be ruled out, given that the end-on bound
l-1,2-peroxo complex [{(TMPA)CuII}2-(O22–)]2+has a
m(O–O)=832 cm–1(D16O2/18O2=44 cm–1, EtCN) [59].
Unfortunately, no definitive m(O–O)has been assigned for
diheme peroxo (P)FeIII-(O22–)-FeIII(P) complexes; thus,
a comparison to such species is not feasible, but NMR
spectroscopy (see below) rules this possibility out.
Overall, given the wide range of stretching frequencies
available for metal-bound end-on and side-on peroxides,
as well as for peroxide-bridged heme-Cu complexes, a
definitive assignment of the bonding mode of O2 in
[(6L)FeIII-(O22–)-CuII]+(2) is not possible based upon
resonance Raman spectroscopy.
When compared to the analogous ‘‘untethered’’
complex [(F8TPP)FeIII-(O22–)-CuII(TMPA)]+, the m(O–O)
of the peroxide ligand in 2 is 20 cm?1lower in energy,
possibly indicating two different coordination geome-
tries (i.e., end-on versus side-on; Fig. 7), despite their
strong spectral similarities (nearly identical UV-visible
and NMR spectral features). Since metal complexes of
the6L and F8TPPH2porphyrin ligands are very similar
asthose
(III)
inFig. 2,
and
[37],
[34]
(II)
(PTACN=triazacyclonon-
[28],[(PTPA)FeIIICuII-
[50], such as the
Table 1 Resonance Raman
data for selected peroxo-level
dioxygen adducts
aAbbreviations used: PTPA=
tris(2-picolylamine)-linked tet-
raphenylporphyrinate; P5-MeT-
PA=tris[2-(5-methylpyr-
idyl)methyl]amine-linked tetra-
phenylporphyrinate; PTACN=
triazacyclononane-capped tet-
raarylporphyrin; OEP=dian-
ion of octaethylporphyrin;
F20TPP=dianion of tetra-
kis(pentafluorophenyl)por-
phryin; pz¢=3,5-bis(isopropyl)
pyrazolyl; 6-Me3-TPA=tris(6-
methyl-2-pyridylmeth-
yl)amine;N-Et-hptb=N,N,N¢,
N¢-tetrakis(1-ethylbenzimidaz-
olyl-2-methyl)-1,3-diamino-2-p-
ropanol
Complexa
m(16O–16O) (D16O2/18O2,
D16O2/16/18O2) (cm?1)
Ref
[(6L)FeIII-(O22–)-CuII]+(2)
[(F8TPP)FeIII-(O22?)-CuII(TMPA)]+
[(TMP)FeIII-(O22–)-(5MeTPA)CuII]+(II)
[(PTPA)FeIIICuII-(O22–)]+
[(P5-MeTPA)FeIIICuII-(O22–)]+
[(PTACN)FeIIICuII-(O22–)]+
[(PTACN)CoIIICuII-(O22–)]+
[(OEP)FeIII-(O22–)]–
[(F20TPP)FeIII-(O22–)]–
[{(TMPA)CuII}2-(O22–)]2+
[{HB(3,5-iPr2pz)3CuII}2–(O22–)]2+
[CuII2(Nn)-(O22–)]2+
788 (44, 22)
808 (46, 23)
790 (44)
803 (44, 22)
793 (42)
758 (18)
804 (48)
808 (47)
802 (n.a.)
832 (44)
741 (43,22)
n=3: 765
n=4: 751
n=5: 741
730 (39), side-on peroxo
588 (28), bis-l-oxo core
R=NO2: 747 (40)
R=F: 735 (39)
888 (46)
848 (46)
900 (50)
This work
[34]
[37]
[29, 33]
[33]
[28]
[60]
[57]
[56]
[41, 48]
[61]
[53, 62]
[{(MePY2)CuII2}-(O22–)]2+
[54]
[CuII2(R-XYL-H)-(O22–)]2+
[55]
[(l-O2CCH2Ph)2{HB(pz¢)3}2FeIII2-(O22–)]
[(6-Me3-TPA)2(l-O)FeIII2-(O22–)]2+
[(N-Et-hptb)2(OPPh3)FeIII2-(O22–)]2+
[63]
[64]
[65]
70

Page 9

electronically [i.e., m(Fe–O) are similar for FeIII-OH
complexes of both ligands4], we exclude an electronic
effect as the cause of the observed m(O–O)differences. On
the other hand, the binding mode of the peroxide ligands
in [(6L)FeIII-(O22–)-CuII]+(2) and [(F8TPP)FeIII-(O22–)-
CuII(TMPA)]+may be the same, with the variation in
m(O–O)arising from structural differences imposed by the
‘‘tethered’’ and ‘‘untethered’’ ligand architectures. Such
ligand-induced structural variations of FeIII-X-CuII
cores (X=bridging ligand) have been observed previ-
ously for these two systems in their corresponding l-oxo
complexes, [(6L)FeIII-O-CuII]+(3) and [(F8TPP)FeIII-O-
CuII(TMPA)]+[66], resulting in differences observed for
l-oxo proton uptake and dpyrroleversus 1/T (Curie plot)
slopes between the two complexes (reflecting proton
sensitivity to unpaired spin density and Fe...Cu magnetic
interactions).Additional
which affect the m(O–O)of the bound dioxygen adduct are
observed in the Nnseries of dicopper(II)-peroxo com-
plexes (Table 1); there, it was shown that less strain in
the ligand leads to lower m(O–O)of the bridging peroxide
[50, 62]. If the ‘‘tethered’’ ligand imposes a structural
constraint upon the peroxide bridge in 2 as it does for
the oxo bridge of 3, we suggest that this somehow strains
or distorts and weakens the O–O bond, as compared to
the ‘‘untethered’’ [(F8TPP)FeIII-(O22–)-CuII(TMPA)]+
complex, whose Fe-O2-Cu moiety may be free to obtain
a more optimal binding coordination geometry. A fuller
understanding of the structures and bonding in these
complexes will be necessary to explain the origin of the
differences in m(O–O)between 2 and [(F8TPP)FeIII-(O22–)-
CuII(TMPA)]+.
ligand-induced variations
NMR spectroscopy of [(6L)FeIII-(O22–)-CuII]+(2)
Further characterization of [(6L)FeIICuI]+(1) and its
oxygenation products came from
spectroscopy. By selectively deuterating the pyrrole
hydrogens (see Materials and methods section), a com-
plementary2H NMR study was also performed, thereby
providing unambiguous assignment of the pyrrole
hydrogens. This proved advantageous, given the ability
to assign the oxidation level and spin state of hemes
based upon the chemical shift of their pyrrole hydro-
gens. Such assignments are possible due to the wealth of
information pertaining to heme/O2chemistry in the lit-
erature [67, 68], and by our independent and separate
NMR spectroscopic studies of the (6L)FeII[47] and
(F8TPP)FeII[45] complexes.
The variable-temperature
reduced complex [(6L)FeIICuI]+(1) in CD3CN are
shown in Fig. S2 (Supplementary material). At 298 K,
the pyrrole-H resonances of 1 were found centered at
d=32.7 ppm (lB=4.1; Evans method, 293 K), and is
1H and
2H NMR
1H NMR spectra of the
representative of an equilibrium mixture of high- and
low-spin states for the ferrous heme. Upon lowering
the temperature of the sample to 233 K, the chemical
shift of the pyrrole hydrogens displayed anti-Curie
behavior as it shifted upfield to d=10 ppm, indicating
more complete formation of a low-spin six-coordinate
heme. This temperature-dependent behavior is consis-
tent with a spin-equilibrium comprised of a high-spin
(five-coordinate, one axially bound MeCN) structure
favored at higher temperature, and a low-spin (six-
coordinate, twoaxially
favored at lower temperatures. This change in structure
at lower temperature leads to a low-spin ferrous heme,
supported by the lack of a detectable magnetic moment
by Evans method (S=0 spin state) for 1 (at 233 K).
Similar observations of temperature-dependent spin
state (due to temperature-dependent axial solvent
fligation) have been made for the similar heme-only
complex (F8TPP)FeII[45].
Direct injection of O2into a chilled NMR tube con-
taining [(6L)FeIICuI]+(1) (Fig. 8a; MeCN, 233 K) led
to the formation of [(6L)FeIII-(O22–)-CuII]+(2), whose
downfield-shifted pyrrole resonances [dpyrrole(ppm)=76,
81; br, m (Fig. 8b)] are indicative of a high-spin ferric
boundMeCN)structure
Fig. 8
reaction in MeCN (233 K): (a) [(6L)FeIICuI]+(1) prior to addition
of dioxygen; (b) [(6L)FeIII-(O22–)-CuII]+(2); and (c) [(6L)FeIII-O-
CuII]+(3) formed from thermal decomposition of 2. Inserts (right)
depict the corresponding Evans method magnetic susceptibility
measurements as followed by2H NMR spectroscopy; two peaks,
one corresponding to the paramagnetically shifted (downfield)
CD3CN solvent and the other to the capillary reference CD3CN
solvent, are observed
1H NMR spectra of the [(6L)FeIICuI]+(1) oxygenation
Fig. 7 Possible heme-peroxo-Cu geometries present in [(6L)FeIII-
(O22–)-CuII]+(2)
4m(Fe–16OH)(D16O/18O) for (6L)FeIII-OH and (F8TPP)FeIII-OH are
636 cm?1(?29) and 638 cm?1(?29), respectively (MeCN, rt,
442 nm excitation); unpublished results
71

Page 10

heme [67, 68] (as supported by the resonance Raman
study, see above). Furthermore, upfield-shifted peaks
were also seen at d=?23, ?36, and ?40 ppm. We have
previously observed and reported this characteristic
pattern of downfield-shifted pyrrole resonances and
upfield-shiftedpeaksin
(X=O22–, O2–) systems having S=2 spin states,
including[(6L)FeIII-O-CuII]+
method, MeCN, ?40 ?C), with downfield-shifted pyr-
role resonances (dpyrrole=75, 80, 87 ppm), and upfield-
shifted TMPA-chelate pyridyl-H resonances at ?27 (py-
5), ?51 (py-3), and ?57 (py-3¢) ppm (Fig. 8c) [30, 32, 69,
70]. The S=2 spin state arises from the antiferromag-
netic coupling of the S=5/2 heme center to an S=1/2
copper(II) moiety, through the bridging peroxo (in 2) or
oxo ligand (in 3). Supporting an overall S=2 spin state
(accompanied by strong electronic/magnetic coupling)
for 2 is the magnetic susceptibility measurement
(lB=4.9; Evans method, CH3CN, ?40 ?C), corre-
sponding to (within experimental error) the spin-only
value of 4.9 lB for four unpaired electrons. Further
evidence for the antiferromagnetic coupling of the S=5/
2 high-spin ferric heme to the S=1/2 copper(II) comes
from EPR spectroscopy (MeCN, 77 K), which shows a
silent spectrum for [(6L)FeIII-(O22–)-CuII]+(2).
Confirmation of the pyrrole resonance assignments
came from2H NMR spectroscopy on pyrrole-deuterated
[(6L-d2)FeIICuI]+(1-d2). In an identical oxygenation
experiment to that described above for the
study (CH3CN, 233 K), a downfield-shifting of the
pyrrole resonances for [(6L-d2)FeIII-(O22–)-CuII]+(2-d2)
[Fig. 9b: dpyrrole(ppm)=75, 80] upon oxygenation of
[(6L-d2)FeIICuI]+(1-d2) [Fig. 9a: dpyrrole(ppm)=10.2]
was observed. Warming the sample to room temperature
similar(P)FeIII-X-CuII
(3)(lB=5.1,Evans
1H NMR
for several hours, followed by re-cooling to 233 K, re-
sulted in the formation of [(6L-d2)FeIII-O-CuII]+(3-d2)
[Fig. 9c: dpyrrole(ppm)=75, 78, 85, 87].
Furthermore, variable-temperature NMR spectros-
copy was able to establish, over a very limited temper-
ature range (228–288 K), the linearity of the Curie plot
for the pyrrole hydrogen resonances in 2 (see Supple-
mentary material), consistent with the predominance of
a single spin state in [(6L)FeIII-(O22–)-CuII]+(2), rather
than a temperature-dependent spin-equilibrium. The
slight deviation from linearity at the coldest temperature
(228 K) can be attributed to the spectrum being ob-
tained close to the freezing point of the MeCN solvent.
Attempts at obtaining the spectrum of [(6L)FeIII-(O22–)-
CuII]+(2) at temperatures higher than 288 K resulted in
significant overlap of the pyrrole resonances with those
from the decomposition l-oxo product [ (6L)FeIII-O-
CuII]+(7), making definitive pyrrole resonance assign-
ments for 2 unreliable.
The analogous ‘‘untethered’’ complex [(F8TPP)FeIII-
(O22–)-CuII(TMPA)]+also exhibits similar NMR spec-
tral features (MeCN, 233 K) to those found for
[(6L)FeIII-(O22–)-CuII]+(2), with the same pattern of
downfield-shifted pyrrole resonances at d=68 ppm and
upfield-shifted pyridyl-H peaks at d=?11 and ?20 ppm
being observed [34]; again, this is indicative of an S=2
spin system (supported by lB=5.1; Evans method,
MeCN, 233 K). This is not surprising, given the degree
of similarity between these two systems, and was ex-
pected considering the l-oxo complexes [(6L)FeIII-O-
CuII]+(3) and [(F8TPP)FeIII-O-CuII(TMPA)]+also
show a high degree of NMR spectral similarity. By
contrast, the NMR spectra of two other heme-O2-Cu
complexes have been reported: [(PTACN)FeCu-(O2)]+
[28] and [(4L)FeIII-O2-CuII]+[31]. Their corresponding
NMR spectra were reported to be diamagnetic, pre-
sumably due to the presence of a strong field axial ligand
in both cases: (1) in [(PTACN)FeCu-(O2)]+, the NMR
spectrum was obtained in pyridine-d5; (2) [(4L)FeIII-O2-
CuII]+contains a tethered pyridine moiety (axial histi-
dine mimic) capable of binding to the heme. Owing to
the lack of a strong field axial ligand, it was possible to
observe for the first time in [(6L)FeIII-(O22–)-CuII]+(2)
the existence of a five-coordinate high-spin ferric heme
in an Fe-O2-Cu heterobinuclear system.
EXAFS spectroscopy of [(6L)FeIII-(O22–)-CuII]+(2)
Structural insights were obtained from an EXAFS
spectroscopic study on frozen solutions (MeCN solvent)
of [(6L)FeIICuI]+(1) and [(6L)FeIII-(O22–)-CuII]+(2);
experimental details and fits can be found in the Sup-
plementary material. The best fit for the Fe center of 1
was obtained with four N atoms (pyrrole) at 1.98 A˚
(typical for a ferrous heme), with two additional N or O
atoms at 2.11 A˚, suggesting axial solvent ligation by two
MeCN molecules. This structure was predicted from the
1H NMR spectroscopic study (see above), which showed
Fig. 9
tion reaction in CH3CN (233 K). The sharp peak at d=1.94 ppm
corresponds to the natural abundance of deuterium in the MeCN
solvent
2H NMR spectra of the [(6L-d2)FeIICuI]+(1-d2) oxygena-
72

Page 11

a pyrrole resonance (d =10.1 ppm; MeCN, 233 K), in
agreement with the presence of a low-spin six-coordinate
ferrous heme at low temperature (lB=0; Evans method,
MeCN, 233 K).
Upon oxygenation in acetonitrile forming 2, a new
Fe–ligand interaction was observed at 1.83 A˚(Fig. 10),
consistent with the formation of an Fe–O bond; an
additional Fe-N/O shell is also observed at 1.99 A˚, ei-
ther representative of a coordinated MeCN molecule or
of an additional Fe–O bond. Moreover, the Cu atom
was detected at 3.409(8) A˚, a feature not observed in 15.
An increase in the Fe–Nhemedistances to 2.08 A˚ was
observed, and is to be expected given the radius of ir-
on(III) is smaller than that of iron(II). The Cu-edge data
for 2 also show a new short Cu–ligand distance of
1.88 A˚, suggestive of a single Cu–O bond. However, the
M...M separation as determined from the Cu edge is
poorly resolved, with much less definition and certainty
than the data obtained from the iron edge to detect Cu:
the least-squares fitting parameter for Cu, utilizing a
fixed M...M distance of 3.4 A˚(as the more accurate Fe
edge suggests), is only slightly worse (0.59 instead of
0.56, within error of the measurement) than that ob-
tained for the best-fit 3.29 A˚distance.
In comparison to the closely related l-oxo com-
plexes [(6L)FeIII-O-CuII]+
CuII(TMPA)]+(Table 2), both of which possess near
linear FeIII-O-CuIIcores, peroxo complex 2 has a sig-
nificantly (?0.3 A˚) shorter Fe...Cu separation. Fur-
thermore, while the Cu–O bond distance of 2 is only
?0.03 A˚greater than those of the l-oxo complexes, the
Fe–O distance is ?0.09 A˚
complexes, where the linear Fe-O-Cu unit elicited a
strong multiple scattering (MS) interaction, simulations
of the peroxo complex 2 could not accommodate any
MS contribution from the Fe-O-O-Cu unit, as expected
for a bridging peroxo.
One interpretation of the data suggests that each
metal center is bound to a single oxygen atom, indicative
of an end-on l-1,2 peroxide ligand in 2 (Fig. 7), with an
additional MeCN (weakly) coordinated to the heme-Fe.
Although the Fe...Cu separation is unusually short at
3.35 A˚(average of Fe and Cu edge data) in 2, a number
of l-1,2-peroxo dinuclear complexes with short M...M
separations of ?2.5–3.5 A˚have been reported (Table 2),
and support this as a possible Fe-O2-Cu geometry. A
second plausible interpretation is that the peroxide
bridge is orientated in a similar fashion as the closely
related structure from Naruta’s group [37], structure II
in Fig. 2, with its g2:g1-FeIII-(O22–)-CuIIcoordination
(Fig. 7c). The two Fe–O/N and one Cu–O distances
derived from this EXAFS spectroscopic study are in
good agreement with the crystallographically deter-
mined M–O bond lengths of structure II. However, the
very significant difference in M...M distances between
(3) and [(F8TPP)FeIII-O-
longer. Unlike the l-oxo
the two structures (>0.55 A˚) argues against this inter-
pretation, and favors an end-on l-1,2 peroxide ligand
in 2.
Thermal rearrangement of [(6L)FeIII-(O22?)-CuII]+(2)
to [(6L)FeIII-O-CuII]+(3)
The question arises as to the fate of the second oxygen
atom lost upon the transformation of the l-peroxo
complex [(6L)FeIII-(O22?)-CuII]+(2) [kmax=418 (So-
ret), 561 nm] to the l-oxo complex [(6L)FeIII-O-CuII]+
(3) [kmax=438 (Soret), 556 nm]. Does 2 undergo
O-atom transfer to a substrate, i.e. solvent, exhibiting
monooxygenase activity? Or, as an alternative, is
dioxygen liberated during this decomposition process?
To answer the latter, a dioxygen-scavenging alkaline
pyrogallol (1,2,3-trihydroxybenzene) solution was used
to detect evolved dioxygen gas; its reaction with O2
causes an immediate change in absorption from near
colorless to intense dark brown. This O2-dependent
absorbance change was quantified using UV-visible
spectroscopy (see Materials and methods section),
leading to a calibration curve which correlates the
absorption of an alkaline pyrogallol solution to the
number of moles O2it has scavenged.
Fig. 10 Experimental versus simulated Fourier transforms and
EXAFS (insets) for complex 2: (a) Fe-edge data: 4 N (pyrrole)
2.08 A˚, 1 O (peroxo) 1.83 A˚, 1 O/N (solvent) 1.99 A˚, 1 Cu 3.41 A˚;
(b) Cu-edge data: 3 N (pyridine, solvent) 2.02 A˚, 1 O (peroxo)
1.88 A˚, 1 Fe 3.29 A˚. Simulations included MS contributions from
the outer shell C atoms of four pyrrole and two pyridine rings at
the Fe and Cu edges, respectively
5The more distant O(peroxide) atom (that bound to Cu) is not
resolvable in the EXAFS becasue it is a low Z scatterer and will be
obscured also by 2nd shell carbons of the porphyrin ring
73

Page 12

When a dioxygen-free solution of alkaline pyrogallol
was exposed to a solution of [(6L)FeIII-(O22?)-CuII]+(2)
as the only possible source of O2(excess O2removed), a
slow and gradual change in absorption of the pyrogallol
was observed, from colorless to yellow, and then to
golden brown. The absorption of the calibrated alkaline
pyrogallol solution after exposure to the O2liberated by
the thermal decomposition of 2 indicated that 0.39–0.43
equivalents of dioxygen were experimentally observed to
be evolved.
Thus, with the formation of 3 being accompanied by
the evolution of nearly 0.5 equivalents of dioxygen gas,
the fate of the second oxygen atom lost in the trans-
formation of l-peroxo complex 2 to l-oxo complex 3
was determined. This decomposition process is highly
reminiscent of several known peroxo-bridged diiron
systems which decay to l-oxo decomposition products
via liberation of dioxygen. Balch and co-workers [77, 78]
proposed that decomposition of their heme-peroxo-
heme adducts proceeded through an irreversible O–O
bond cleavage step (rate limiting), affording a ferryl
species, PFeIV=O, which ultimately rearranges in a
multi-step decay mechanism to yield the l-oxo diiron
product and releasing O2 (as detected by mass spec-
trometry):2 PFeIII-(O22–)-FeIIIP fi 2 PFeIII-O-FeIII-
P+O2. Additionally, Lippard and co-workers [79, 80]
have studied a series of non-heme alkoxo-bridged di-
iron(II) complexes (as synthetic analogues of ribonu-
cleotide reductase and soluble methane monooxygenase)
whose oxygenation yields l-peroxo dioxygen adducts.
They established the liberation (by manometric means)
of a half-equivalent of O2upon the decomposition of
these l-peroxo dioxygen adducts to l-oxo products.
Here, they propose two possible decomposition path-
ways yielding such a result: (1) formation of a tetraoxide
intermediate6similar to the decay of secondary and
tertiary alkyl peroxide radicals, which then transforms
to the final end-product and liberates dioxygen; and (2) a
nucleophilic attack of one peroxide species on the other,
followed by a disproportionation which forms O2. While
the thermal transformation of 2 to 3 occurs without
detection of an intermediate, this does not preclude the
presence of a highly reactive species (i.e., ferryl, cupryl,
tetraoxide, etc.) whose formation yields the decomposi-
tion products in an analogous mechanism to those de-
scribed above for the homonuclear (di)iron systems.
Cobaltocene reduction of [(6L)FeIII-(O22–)-CuII] (2)
Chemical confirmation of the presence of a peroxo level
dioxygen adduct in [(6L)FeIII-(O22–)-CuII]+(2) came
from a UV-visible spectroscopic titration experiment
with cobaltocene (Fig. S3).
[kmax=418 (Soret), 561 nm; MeCN, 233 K] in which
excess dioxygen was removed by successive freeze–
pump–thaw cycles (5·, thaw temperature of 233 K) was
added 2 equiv of Co(Cp)2(as four sequential 0.5 equiv
aliquots) under an argon flow. After each cobaltocene
Toa solutionof2
aAbbreviations used: OEP=dianion of octaethylporphyrinato;
tren=tris(2-aminoethyl)amine; Me5dien=permethyl-diethylenetri-
amine;bpmp=2,6-bis[bis(2-pyridylmethyl)aminomethyl]-4-meth-
ylphenolato; Ph-bimp=2,6-bis[bis{2-(1-methyl-4,5-diphenylimi-
dazolyl)methyl}aminomethyl]-4-methylphenolate; pz¢=3,5-bis(iso-
propyl)pyrazolyl; 6-Me3-TPA=tris(6-methyl-2-pyridylmethyl)amine;
N-Et-hptb=N,N,N¢,N¢-tetrakis(1-ethylbenzimidazolyl-2-methyl)-1,
3-diamino-2-propanol; bpman=2,7-bis[bis(2-pyridylmethyl)ami-
nomethyl]-1,8-naphthyridine; pz=pyrazolyl
bM1=iron; M2=copper
Table 2 Structural comparisons of [(6L)FeIII-(O22–)-CuII]+(2) with several bridged heme-copper assemblies and l-peroxo diiron and
dicopper adducts
Complexa
M1–O (A˚)b
M2–O (A˚)b
Mn+...Mn+(A˚)Ref
[(6L)FeIII-(O22–)-CuII]+(2) 1.83(1)1.882(4)3.409(8) Fe-edge
3.294(17) Cu-edge
3.915
This work
[(TMP)FeIII-(O22–)-(5MeTPA)CuII]+(II)1.890
2.031
1.746
1.78
1.740
1.87
1.74
1.93
1.893
1.944
1.863
1.881
1.85
1.880
?2
?1.95
1.852
1.927
1.915
2.657
1.851
1.84
1.856
1.89
1.84
1.99
1.863
1.864
1.862
1.877
1.85
1.880
?2
?1.95
1.852
1.903
[37]
[(6L)FeIII-O-CuII]+(3)
[(5L)FeIII-O-CuII]+
[(F8TPP)FeIII-O-CuII(TMPA]+
[(F8TPP)FeIII-OH-CuII(TMPA]2+
[(OEP)FeIII-O-CuII(Me6tren)]+
[(OEP)FeIII-OH-CuII(Me5dien)]2+
[(benzoato)(bpmp)CoIII2-(O22–)]2+
[(benzoato)(Ph-bimp)FeIII2-(O22–)]2+
[(Ph2PO2)(Ph-bimp)CoIII2-(O22–)]2+
[(l-O2CCH2Ph)2{HB(pz¢)3}2FeIII2-(O22–)]
[(6-Me3-TPA)2(l-O)FeIII2-(O22–)]2+
[(N-Et-hptb)2(OPPh3)FeIII2-(O22–)]2+
M ferritin
[{(bpman)CuII}2-(O22–)]2+
[{(TMPA)CuII}2-(O22–)]2+
[{(HB(3,5-iPr2pz)3)CuII}2-(O22–)]2+
3.586
3.40
3.596
3.66 (av)
3.58
3.89
3.151
3.327
3.231
4.007
3.14
3.462
2.53
2.84
4.359
3.560
[46]
[66]
[69]
[66]
[71]
[71]
[72]
[73]
[74]
[63]
[64]
[65]
[75]
[76]
[40]
[51]
6This intermediate was ruled out when mass spectrometric analysis
of the headspace gas revealed no mixed-isotope dioxygen gas was
evolved upon the decomposition of a mixture of16O2and18O2l-
peroxo adducts
74

Page 13

addition, an immediate change in the UV-visible spec-
trum was observed, corresponding to a loss of 2 and
formation of 3 [kmax=438 (Soret), 556 nm]. Full for-
mation of the oxo product 3 was achieved after addition
of 2 equiv Co(Cp)2, whereas fewer equivalents gave a
mixture of 2 and 3, while excess cobaltocene was unable
to further reduce 3 back to the starting FeII/CuIcomplex
1. This differs from the cobaltocene reactivity observed
by Collman and co-workers [60, 81], whose heme-pe-
roxo-Cu systems undergo full 4e–reduction back to the
corresponding iron(II)/copper(I) starting complexes.
This difference in reactivity may be attributed to the
overall spin state of the peroxo complexes. The low-spin,
and hence easier to be reduced, systems of Collman and
co-workers are able to undergo full reduction back to
the FeII/CuIstarting complex, whereas the high-spin,
and thus harder to reduce, peroxo-level dioxygen adduct
of [(6L)FeIII-(O22–)-CuII]+(2) can only undergo cobal-
tocene reduction to the extremely stable l-oxo complex
[(6L)FeIII-O-CuII]+(3), despite the presence of a large
excess of reductant.
Dioxygen reactivity of [(6L)FeIICuI]+(1)
in other solvents
While the bulk of the characterization of [(6L)FeIII-
(O22–)-CuII]+(2) was performed in acetonitrile, other
reaction solvents have been utilized to study the oxy-
genation chemistry of 1, with details provided in the
Supplementary material. Although 2 forms in acetone
and THF, as followed by UV-visible and NMR spec-
troscopies, stabilization of the heme-superoxo precursor
is further enhanced in these solvents, making its detec-
tion readily observable under normal benchtop spec-
troscopic monitoring. One other difference is that the
stability of 2 in THF solvent is much lower than in
acetone or MeCN at room temperature (but equally
stable at low temperature), and immediately yields 3
under these conditions.
Conclusions
Our investigations into the dioxygen reactivity of
[(6L)FeIICuI]+(1) reveal the formation of a heme-Cu/O2
complex formulated as [(6L)FeIII-(O22–)-CuII]+(2). This
peroxo-level dioxygen adduct, as confirmed by reso-
nance Raman spectroscopy, forms upon addition of one
equivalent dioxygen (per FeII/CuIheterobinuclear cen-
ter), and was found to be the first example of what is
now a growing number of high-spin heme-peroxo-Cu
adducts in the literature. The result of this unique S=2
spin state for 2 (as confirmed by Evans method magnetic
moment measurements) is the very distinctive pattern of
1H and
presence of downfield-shifted pyrrole resonances and
upfield-shifted pyridyl protons, which has become the
2H NMR spectroscopic features, namely the
signature of the S=2 spin state of antiferromagnetically
bridged peroxo- or oxo-heme-Cu complexes of the form
PFeIII-X-CuII[X=O22–, O2–]. A second consequence of
the high-spin nature of [(6L)FeIII-(O22–)-CuII]+(2)
manifests itself in a difference in chemical reactivity, with
respect to reduction by cobaltocene, when compared to
other heme-peroxo-Cu complexes in the literature.
While 2 immediately undergoes a 2e–reduction to
[(6L)FeIII-O-CuII]+(3) (MeCN solvent, 233 K) when
exposed to an excess of Co(Cp)2, the peroxo systems of
Collman and co-workers undergo complete 4e?reduc-
tion back to their deoxygenated FeII/CuIforms. The
low-spin heme-peroxo-Cu complexes thus appear to be
more easily reduced than their high-spin counterparts.
X-ray absorption spectroscopy, while not absolutely
definitive, suggests a l-1,2-peroxo ligand geometry in
[(6L)FeIII-(O22–)-CuII]+(2), whose structure is similar
(yet distinctive) to l-oxo bridged heme-Cu assemblies.
This geometry is supported by resonance Raman spec-
troscopy, which showed the presence of a symmetrically
bound peroxide ligand. Furthermore, decomposition of
2 to [(6L)FeIII-O-CuII]+(3) occurs with concomitant
release of 0.5 equiv O2, thus establishing the fate of all
oxygen atoms present in the thermal decomposition
reaction of 2.
Our investigations into the dioxygen reactivity of
heme-Cu systems which form l-peroxo complexes such
as [(6L)FeIII-(O22–)-CuII]+(2) have shown that the
spectroscopy and chemical reactivity of this high-spin
peroxo complex differ from its low-spin FeIII-O2-CuII
counterparts. Many questions arise from this work. If
a bridging peroxide species is identified as a fleeting
intermediate in the mechanism of dioxygen reduction
by cytochrome c oxidase, is it a low-spin or high-spin
species? More pertinent to what can truly be deter-
mined from chemical modeling studies is the question
of what further chemistry occurs when S=2 high-spin
species such as 2 are subjected to electron/proton
sources; does O–O cleavage occur? Can a high-spin
(S=2) heme-peroxo-Cu complex be converted to a
low-spin complex, either by addition of a strong axial
base donor (i.e., imidazole) or through changes in the
ligand and resulting electronic structure of the starting
FeII/CuIcomplex? If so, what is the subsequent reac-
tivity of such species? Further efforts to elucidate such
basic aspects of heme/copper dioxygen chemistry are
in progress.
Acknowledgements We are grateful to the National Institutes of
Health (K.D.K.,GM60353;
GM18865) for support of this research.
R.J.C.,GM54882;P.M.-L.,
References
1. Ferguson-Miller S, Babcock GT (1996) Chem Rev 96:2889–
2907
2. Michel H, Behr J, Harrenga A, Kannt A (1998) Annu Rev
Biophys Biomol Struct 27:329–356
75

Page 14

3. Iwata S, Ostermeier C, Ludwig B, Michel H (1995) Nature
376:660–669
4. Tsukihara T, Aoyama H, Yamashita E, Tomizaki T, Yamag-
uchi H, Shinzawa-Itoh K, Nakashima R, Yaono R, Yoshikawa
S (1995) Science 269:1069–1074
5. Tsukihara T, Aoyama H, Yamashita E, Tomizaki T, Yamag-
uchi H, Shinzawa-Itoh K, Nakashima R, Yaono R, Yoshikawa
S (1996) Science 272:1136–1144
6. Ostermeier C, Harrenga A, Ermler U, Michel H (1997) Proc
Natl Acad Sci USA 94:10547–10553
7. Yoshikawa S, Shinzawa-Itoh K, Nakashima R, Yaono R,
Yamashita E, Inoue N, Yao M, Jei-Fei M, Libeu CP, Mizu-
shima T, Yamaguchi H, Tomizaki T, Tsukihara T (1998) Sci-
ence 280:1723–1729
8. Harrenga A, Michel H (1999) J Biol Chem 274:33296–33299
9. Soulimane T, Buse G, Bourenkov GP, Bartunik HD, Huber R,
Than ME (2000) EMBO J 19:1766–1776
10. Abramson J, Riistama S, Larsson G, Jasaitis A, Svensson-Ek
M, Laakkonen L, Puustinen A, Iwata S, Wikstrom M (2000)
Nat Struct Biol 7:910–917
11. Svensson-Ek M, Abramson J, Larsson G, Tornroth S, Brze-
zinski P, Iwata S (2002) J Mol Biol 321:329–339
12. Babcock GT (1999) Proc Natl Acad Sci USA 96:12971–12973
13. Fabian M, Wong WW, Gennis RB, Palmer G (1999) Proc Natl
Acad Sci USA 96:13114–13117
14. Proshlyakov DA, Pressler MA, Babcock GT (1998) Proc Natl
Acad Sci USA 95:8020–8025
15. Blomberg MRA, Siegbahn PEM, Wikstro ¨ m M (2003) Inorg
Chem 42:5231–5243
16. Das TK, Pecoraro C, Tomson FL, Gennis RB, Rousseau DL
(1998) Biochemistry 37:14471–14476
17. Okeley NM, Van der Donk WA (2000) Chem Biol 7:R159–
R171
18. Michel H (1998) Proc Natl Acad Sci USA 95:12819–12824
19. Zaslavsky D, Gennis RB (2000) Biochim Biophys Acta
1458:XXX–XXX
20. Wikstro ¨ m M, Verkhovsky MI (2002) Biochim Biophys Acta
1555:128–132
21. Szundi I, Van Eps N, Einarsdo ¨ ttir O (2003) Biochemistry
42:5074–5090
22. Van Eps N, Szundi I, Einarsdo ¨ ttir O (2003) Biochemistry
42:5065–5073
23. Kitagawa T (2000) J Inorg Biochem 82:9–18
24. Verkhovsky MI, Jasaitis A, Verkhovskaya ML, Morgan JE,
Wikstro ¨ m M (1999) Nature 400:480–483
25. Kim E, Chufan EE, Kamaraj K, Karlin KD (2004) Chem Rev
104:1077–1133
26. Collman JP, Boulatov R, Sunderland CJ, Fu L (2004) Chem
Rev 104:561–588
27. Collman JP, Boulatov R, Sunderland CJ (2003) Porphyrin
Handbook 11:1–49
28. Collman JP, Herrmann PC, Boitrel B, Zhang X, Eberspacher
TA, Fu L, Wang J, Rousseau DL, Williams ER (1994) J Am
Chem Soc 116:9783–9784
29. Sasaki T, Nakamura N, Naruta Y (1998) Chem Lett 351–352
30. Ghiladi RA, Ju TD, Lee D-H, Moe ¨ nne-Loccoz P, Kaderli S,
Neuhold Y-M, Zuberbu ¨ hler AD, Woods AS, Cotter RJ, Karlin
KD (1999) J Am Chem Soc 121:9885–9886
31. Kopf M-A, Karlin KD (1999) Inorg Chem 38:4922–4923
32. Kopf M-A, Neuhold Y-M, Zuberbu ¨ hler AD, Karlin KD (1999)
Inorg Chem 38:3093–3102
33. Naruta Y, Sasaki T, Tani F, Tachi Y, Kawato N, Nakamura N
(2001) J Inorg Biochem 83:239–246
34. Ghiladi RA, Hatwell KR, Karlin KD, Huang H, Moe ¨ nne-
Loccoz P, Krebs C, Huynh BJ, Marzilli LA, Cotter RJ, Kaderli
S, Zuberbu ¨ hler AD (2001) J Am Chem Soc 123:6183–6184
35. Kim E, Helton ME, Wasser IM, Karlin KD, Lu S, Huang H,
Moe ¨ nne-Loccoz P, Incarvito CD, Rheingold AL, Honecker M,
Kaderli S, Zuberbu ¨ hler AD (2003) Proc Natl Acad Sci USA
100:3623–3628
36. Collman JP, Sunderland CJ, Berg KE, Vance MA, Solomon EI
(2003) J Am Chem Soc 125:6648–6649
37. Chishiro T, Shimazaki Y, Tani F, Tachi Y, Naruta Y, Karas-
awa S, Hayami S, Maeda Y (2003) Angew Chem Int Ed
42:2788–2791
38. Blackburn NJ, Rhames FC, Ralle M, Jaron S (2000) J Biol
Inorg Chem 5:341–353
39. Karlin KD, Kaderli S, Zuberbu ¨ hler AD (1997) Acc Chem Res
30:139–147
40. Jacobson RR, Tyekla ´ r Z, Karlin KD, Liu S, Zubieta J (1988) J
Am Chem Soc 110:3690–3692
41. Karlin KD, Wei N, Jung B, Kaderli S, Zuberbu ¨ hler AD (1991)
J Am Chem Soc 113:5868–5870
42. Karlin KD, Wei N, Jung B, Kaderli S, Niklaus P, Zuberbu ¨ hler
AD (1993) J Am Chem Soc 115:9506–9514
43. Karlin KD, Lee D-H, Kaderli S, Zuberbu ¨ hler AD (1997) Chem
Commun 475–476
44. Zhang CX, Kaderli S, Costas M, Kim E, Neuhold Y-M, Karlin
KD, Zuberbu ¨ hler AD (2003) Inorg Chem 42:1807–1824
45. Ghiladi RA, Kretzer RM, Guzei I, Rheingold AL, Neuhold Y-
M, Hatwell KR, Zuberbu ¨ hler AD, Karlin KD (2001) Inorg
Chem 40:5754–5767
46. Ju TD, Ghiladi RA, Lee D-H, van Strijdonck GPF, Woods AS,
Cotter RJ, Young VG Jr, Karlin KD (1999) Inorg Chem
38:2244–2245
47. Ghiladi RA, Karlin KD (2002) Inorg Chem 41:2400–2407
48. Tyekla ´ r Z, Jacobson RR, Wei N, Murthy NN, Zubieta J,
Karlin KD (1993) J Am Chem Soc 115:2677–2689
49. Stenkamp RE (1994) Chem Rev 94:715–726
50. Mirica LM, Ottenwaelder X, Stack TDP (2004) Chem Rev
104:1013–1045
51. Kitajima N, Fujisawa K, Fujimoto C, Moro-oka Y, Hashim-
oto S, Kitagawa T, Toriumi K, Tasumi K, Nakamura A (1992)
J Am Chem Soc 114:1277–1291
52. Kitajima N, Moro-oka Y (1994) Chem Rev 94:737–757
53. Liang H-C, Karlin KD, Dyson R, Kaderli S, Jung B,
Zuberbu ¨ hler AD (2000) Inorg Chem 39:5884–5894
54. Obias HV, Lin Y, Murthy NN, Pidcock E, Solomon EI, Ralle
M, Blackburn NJ, Neuhold Y-M, Zuberbu ¨ hler AD, Karlin KD
(1998) J Am Chem Soc 120:12960–12961
55. Pidcock E, Obias HV, Zhang CX, Karlin KD, Solomon EI
(1998) J Am Chem Soc 120:7841–7847
56. Selke M, Sisemore MF, Valentine JS (1996) J Am Chem Soc
118:2008–2012
57. Burstyn JN, Roe JA, Miksztal AR, Shaevitz BA, Lang G,
Valentine JS (1988) J Am Chem Soc 110:1382–1388
58. Chufa ´ n EE, Karlin KD (2003) J Am Chem Soc 125:16160–
16161
59. Baldwin MJ, Ross PK, Pate JE, Tyekla ´ r Z, Karlin KD, Solo-
mon EI (1991) J Am Chem Soc 113:8671–8679
60. Collman JP, Fu L, Herrmann PC, Zhang X (1997) Science
275:949–951
61. Kitajima N (1993) In: Karlin KD, Tyekla ´ r Z (eds) Bioinorganic
chemistry of copper. Chapman & Hall, New York, pp 251–263
62. Pidcock E, Obias HV, Abe M, Liang H-C, Karlin KD, Solo-
mon EI (1999) J Am Chem Soc 121:1299–1308
63. Kim K, Lippard SJ (1996) J Am Chem Soc 118:4914–4915
64. Dong Y, Zang Y, Shu L, Wilkinson EC, Que L Jr (1997) J Am
Chem Soc 119:12683–12684
65. Dong Y, Yan S, Young VG Jr, Que L Jr (1996) Angew Chem
Int Ed Engl 35:618–620
66. Obias HV, van Strijdonck GPF, Lee D-H, Ralle M, Blackburn
NJ, Karlin KD (1998) J Am Chem Soc 120:9696–9697
67. Walker FA, Simonis U (1993) In: Berliner LJ, Reuben J (eds)
Biological magnetic resonance. Plenum Press, New York, pp
133–274
68. Walker FA (2000) In: Kadish KM, Smith KM, Guilard R (eds)
The porphyrin handbook. Academic Press, San Diego, pp 81–
184
69. Karlin KD, Nanthakumar A, Fox S, Murthy NN, Ravi N,
Huynh BH, Orosz RD, Day EP (1994) J Am Chem Soc
116:4753–4763
70. Nanthakumar A, Fox S, Murthy NN, Karlin KD (1997) J Am
Chem Soc 119:3898–3906
76

Page 15

71. Kauffmann KE, Goddard CA, Zang Y, Holm RH, Mu ¨ nck E
(1997) Inorg Chem 36:985–993
72. Suzuki M, Ueda I, Kanatomi H, Murase I (1983) Chem Lett
185–188
73. Ookubo T, Sugimoto H, Nagayama T, Masuda H, Sato T,
Tanaka K, Maeda Y, Okawa H, Hayashi Y, Uehara A, Suzuki
M (1996) J Am Chem Soc 118:701–702
74. Seo J, Sung N-D, Hynes RC, Chin J (1996) Inorg Chem
35:7472–7473
75. Hwang J, Krebs C, Huynh BH, Edmondson DE, Theil EC,
Penner-Hahn JE (2000) Science 287:122–125
76. He C, DuBois JL, Hedman B, Hodgson KO, Lippard SJ (2001)
Angew Chem Int Ed 40:1484–1487
77. Chin D-H, La Mar GN, Balch AL (1980) J Am Chem Soc
102:4344–4350
78. Balch AL (1992) Inorg Chim Acta 198–200:297–307
79. Feig AL, Becker M, Schindler S, van Eldik R, Lippard SJ
(1996) Inorg Chem 35:2590–2601
80. DuBois JL, Mizoguchi TJ, Lippard SJ (2000) Coord Chem Rev
200:443–485
81. Collman JP, Fu L, Herrmann PC, Wang Z, Rapta M, Bro ¨ ring
M, Schwenninger R, Boitrel B (1998) Angew Chem Int Ed
37:3397–3400
77