In search of elusive high-valent manganese species that evaluate mechanisms of photosynthetic water oxidation

Article abstract of DOI:10.1021/ic7017488

Significant progress in the understanding of biological water oxidation has occurred during the past 25 years. Today we have a somewhat clearer description of the structure of the Mn₄Ca cluster and an idea of the appropriate oxidation states for the enzyme during catalysis. At issue is the mechanism of water oxidation. Depending on one's belief of the manganese ion oxidation levels at the catalytically active S₄ configuration, one can invoke a variety of different processes that could lead to water oxidation. We have suggested that the most likely process is the nucleophilic attack of a water bound to calcium (or manganese) onto a highly electrophilic Mn^v=O center. In this Article, we explore the difficulties of preparing Mn^v in dimeric systems and the even more arduous task of definitively assigning oxidation states to such highly reactive species.

Full text of DOI:10.1021/ic7017488

Inorg. Chem. 2008, 47, 1765-1778
Forum
In Search of Elusive High-Valent Manganese Species That Evaluate
Mechanisms of Photosynthetic Water Oxidation
Vincent L. Pecoraro* and Wen-Yuan Hsieh
Departments of Chemistry and Biophysics, The UniVersity of Michigan,
Ann Arbor, Michigan 48109-1055
Received September 5, 2007
Significant progress in the understanding of biological water oxidation has occurred during the past 25 years.
Today we have a somewhat clearer description of the structure of the Mn4Ca cluster and an idea of the appropriate
oxidation states for the enzyme during catalysis. At issue is the mechanism of water oxidation. Depending on one’s
belief of the manganese ion oxidation levels at the catalytically active S4 configuration, one can invoke a variety
of different processes that could lead to water oxidation. We have suggested that the most likely process is the
nucleophilic attack of a water bound to calcium (or manganese) onto a highly electrophilic MnVdO center. In this
Article, we explore the difficulties of preparing MnV in dimeric systems and the even more arduous task of definitively
assigning oxidation states to such highly reactive species.
Introduction
at best, poorly described. This Inorganic Chemistry Forum
focuses on several aspects of this amazing process ranging from
understanding fully the structure and function of the oxygen-
evolving complex (OEC) of photosystem II (PS II) to chemical
water oxidation systems requiring ruthenium. Our specific
contribution to this collection of articles is to address the role
of small-molecule bioinorganic models for defining the mech-
anism of photosynthetic water oxidation. The interested reader
is referred to reviews for other discussions of small-molecule
manganese complexes and their role in understanding photo-
synthesis.1–3
On paper, water oxidation really is a rather simple reaction.
Take two molecules of water, oxidize them, and release four
protons, as depicted in eq 1.
2H2O f O2 + 4H+ + 4e- (0.89 V ⁄ e- at pH 6) (1)
4H+ + 4e- f 2H2
2H2 + O2 f 2H2O
(2)
(3)
Water oxidation is the engine that drives higher plant and algal
photosynthesis, which ultimately captures nearly all biological
energy on earth.1 Water oxidation also promises to be essential
to any futuristic hydrogen economy because from both resource
and ecological perspectives the ideal cycle for generating and
burning dihydrogen should invoke a step where water is
oxidized to dioxygen (a combination of eqs 1 and 2) and
dioxygen is converted to water (eq 3). Amazingly, despite its
importance to bioenergetics and geopolitical/economic stability,
the mechanism of water oxidation and functional models are,
Bioenergetics of Water Oxidation
While described more fully in other articles in this Forum,
it is worth spending a short time examining the structural
and enzymological features of PS II, the workhorse of higher
plant photosynthesis.1 PS II is embedded within the thylakoid
membrane of the chloroplast. Originally, it was thought that
(2) (a) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. U.S.A. 2006,
103, 15729. (b) Crabtree, G. W.; Lewis, N. S. Phys. Today 2007, 60,
37. (c) Nocera, D. G. Daedalus 2006, 135, 112.
* To whom correspondence should be addressed. E-mail: vlpec@
umich.edu.
(3) (a) Wieghardt, K. Angew. Chem. 1989, 101, 1179. (b) Law, N. A.;
Caudle, M. T.; Pecoraro, V. L. AdV. Inorg. Chem. 1999, 46, 305. (c)
Pecoraro, V. L.; Baldwin, M. J.; Gelasco, A. Chem. ReV. 1994, 94,
807. (d) Mukhopadhyay, S.; Mandal, S. K.; Bhaduri, S.; Armstrong,
W. H. Chem. ReV. 2004, 104, 3981. (e) Wu, A. J.; Penner-Hahn, J. E.;
Pecoraro, V. L. Chem. ReV. 2004, 104, 903. (f) Pecoraro, V. L.
Photochem. Photobiol. 1988, 48, 249.
(1) (a) Wydrzynski, T. Photosystem II: The Light-DriVen Water: Plas-
toquinone Oxidoreductase; Springer: New York, 2005. (b) Tommos,
C.; Babcock, G. T. Acc. Chem. Res. 1998, 31, 18–25. (c) McEvoy,
J. P.; Brudvig, G. W. Chem. ReV. 2006, 106, 4455–4483. (d) Yocum,
C. F.; Pecoraro, V. L. Curr. Opin. Chem. Biol. 1999, 3, 182–187. (e)
Yachandra, V. K.; Sauer, K.; Klein, M. P. Chem. ReV. 1996, 96, 2927.
10.1021/ic7017488 CCC: $40.75  2008 American Chemical Society
Inorganic Chemistry, Vol. 47, No. 6, 2008 1765
Published on Web 03/10/2008

Pecoraro and Hsieh
the reaction center, P680, captured a photon of visible light
(λ ) 680 nm) and was promoted to a virtual state that served
as a reductant of photopigments such as chlorophyll, pheo-
phytin, and quinones. Workers in the ultrafast community4
have now shown that energy transfer occurs at the site of
chlorophyll a (Chlor) and pheophytin D1 (PheoD1), generat-
ing an initial Chlor+PheoD1- state, with the electron transfer-
ring to the quinone and chlorophyll reduction achieved by
P680. P680+ directly serves to oxidize a protein-bound tyrosine,
Yz, which serves as an electron-transfer conduit to the water
oxidation center (a cluster of four manganese, one calcium,
and possibly chloride and carbonate) known as the oxygen-
evolving complex (OEC). A 680 nm photon contains 1.8
eV. This energy must be apportioned between the reductive
reactions necessary for plant growth and the oxidative
reactions that generate protons used for making adenosine
triphosphate. It is thought that approximately 0.6-0.7 V is
necessary for driving the reactions of the “reducing side” of
the enzyme (ultimately generating a doubly reduced quinone
known as QB that transfers the newly generated electrons
from PS II to membrane acceptor proteins that shuttle these
reducing equivalents to photosystem I). Because 0.6 V is
used on the reducing side of the reaction, the oxidizing side
of the process is left with 1.2 eV to oxidize water. Recent
direct determinations of the P680+/P680 potential have given
1.255a and 1.26 V.5b At first blush, this should not be a
problem because, at the slightly acidic pH conditions of the
thylakoid, only 0.89 V/e- is required for this process.
However, one must recognize that photosynthesis must be a
high-quantum-yield reaction and requires at least one inter-
mediate oxidant, Yz+. Therefore, some of this excess energy
must be lost to attain a near-unity quantum yield. It has been
estimated that 100 mV is necessary to drive each oxidation
step (P680+/Yz to P680/Yz+ and Yz+/OECn to Yz/OECn+1).
Lavergne and Diner5b have estimated the S2/S1 potential to
be 1.02 V. Obviously, this means there is only ∼1.0 V of
useful oxidizing power to drive the reaction.
Figure 1. Diagram of the S clock, which describes the five intermediate
oxidation states of the OEC. S0 is the product state after dioxygen has been
released, S1 is the stable oxidation level in the dark, and the transition S3
f S4 f S0 is the step leading to dioxygen production and release. The S4*
state has recently been reported. Reproduced with permission from ref 6.
Figure 2. Cyclic voltammogram for the successive oxidation of [Mn(2-
OHSalpn)]2 complexes illustrating the large energy separation between each
successive dimer oxidation. Reproduced from ref 7.
In the early 1980s, many workers struggled to understand
how such a strongly oxidized cofactor could be found in
biology.3f It was well-known by the late 1960s and early
1970s, based on the pioneering work of Pierre Joliot6a–c and
Bessel Kok6d that had placed the kinetic mechanism of water
oxidation in clear view, that four oxidizing equivalents were
required for water oxidation. Kok described this via the S
cycle shown in Figure 1 (this version of the S clock shows
an S4* state that was uncovered very recently6e). This series
of photochemically driven reactions invokes five OEC redox
intermediates known as states and abbreviated as S0-S4. The
most reduced state is S0, the most oxidized state is S4, and
the transition from S3 f S4 f S0 corresponds to the critical
step where dioxygen is formed and released. As discussed
above, the amount of energy available to oxidize each of
these states is a uniform 1.0 V. With the information
provided, there appeared to be no problem with these
energetics; however, we pointed out in the early 1990s that
there was an even more vexing problem that biology needed
to overcome. That problem was the accumulation of four
oxidizing equivalents on a manganese oxo cluster maintained
within the proper potential range and unable to participate
in unwanted side reactions. As was illustrated by the
successive oxidations of a manganese compound [Mn(2-
OHSalpn)]2, which we described some years ago7 (but which
had been known for other small metal aggregates for many
years before this), each successive oxidation of a metal ion
(4) (a) Groot, M. L.; Pawlowicz, N. P.; Van Wilderen, L. J. G. W.; Breton,
J.; van Stokkum, I. H. M.; van Grondelle, R. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 13087. (b) Ishikita, H.; Biesiadka, J.; Loll, B.;
Saenger, W.; Knapp, E.-W. Angew. Chem., Int. Ed. 2006, 45, 1964.
(5) (a) Grabolle, M.; Dau, H. Biochim. Biophys. Acta, Bioenerg. 2005,
1708, 209. (b) Rappaport, F.; Guergova-Kuras, M.; Nixon, P. J.; Diner,
B. A.; Lavergne, J. Biochemistry 2002, 41, 8518.
(6) (a) Joliot, P. BrookhaVen Symp. Biol. 1966, 418. (b) Joliot, P. Biochim.
Biophys. Acta 1965, 102, 116. (c) Joliot, P.; Joliot, A. Biochim.
Biophys. Acta 1973, 305, 302. (d) Kok, B.; Forbush, B.; McGloid,
M. Photochem. Photobiol. 1970, 11, 457. (e) Haumann, M.; Liebisch,
P.; Muller, C.; Barra, M.; Grabolle, M.; Dau, H. Science 2005, 310,
1019.
(7) (a) Gelasco, A.; Kirk, M. L.; Kampf, J. W.; Pecoraro, V. L. Inorg.
Chem. 1997, 36, 1829. (b) Gelasco, A.; Pecoraro, V. L. Inorg. Chem.
1998, 37, 3301.
1766 Inorganic Chemistry, Vol. 47, No. 6, 2008

ElusiWe High-Valent Manganese Species
in a cluster causes the subsequent oxidation of the cluster to
occur at higher potential (between 250 and 500 mV higher)
unless there is a concomitant loss of the newly accumulated
positive charge. Notice that this concept is different from
examination of a Pourbaix diagram, in which one evaluates
the pH dependence of a single redox couple as a function of
the pH. In this case, we are comparing different redox
couples restricted to the same pH conditions. One might also
argue that a significant structural change of a cluster could
mitigate this principle; however, we would contend that this
significantly structurally changed cluster is really a new
molecule in its own right. For example, the [Mn(2-
OHSalpn)]2 system has approximately 300 mV shifts in
potential each time the manganese ions are oxidized. If the
same principle is applied to the OEC, oxidation of the
manganese ions would engender an approximately 300 mV
shift in potential. Recognizing that the maximum potential
for any S transition can only be ∼1 V, this would require
that the less oxidizing transitions be at 0.7 V (S2 f S3), 0.4
V (S1 f S2), and 0.1 V (S0 f S1). The sum of these potentials
is 2.2 V, far from the 0.89 V/e- (∼3.6 V) necessary for water
oxidation. This discussion has not yet considered the second
significant problem with the potential range.
Figure 3. Comparison of cyclic voltammograms for {[MnIII(2-OH-
Salpn)]2H2O} and {MnIIIMnIV(2-OHSalpn)2OH} illustrating how the po-
tential for a MnIII to MnIII/MnIV transition is nearly at the same potential
2
as a MnIII/MnIV to MnIV2 oxidation when a proton has been removed from
a bound ligand. Adapted from ref 13.
possibly each oxidation has a maximal 1.1 V potential.
Nonetheless, the concerns raised by this analysis are troubling.
The only way to avert the energetic penalty associated with
cluster oxidation is to have a charge-invariant transition or
to have a major cluster structural change (presumably in a
process in which the addition of a ligand cancels the added
positive charge). The simplest way to achieve a charge-
invariant cluster is to lose a proton (alternatives would
include the addition of an anionic ligand such as chloride11
or (bi)carbonate12). This principle is illustrated in Figure 3
using the related complex {[Mn(2-OHSalpn)]2H2O},13 which
has a bound water molecule that can be deprotonated to form
a hydroxide complex. Thus, the potential of {[MnIII(2-
OHSalpn)]2H2O} is within 50–100 mV of the more highly
oxidized {MnIIIMnIV(2-OHSalpn)2OH}. The studies of proton
release in PS II are too numerous to recount here, but it is
well accepted that on at least several steps a proton departs
the vicinity of the manganese cluster upon an S transition.14
Thus, we can conclude that the energetic constraints of the
system require proton loss, anion complexation, or cluster
structure rearrangement during the steps preceding water
oxidation and release. Interestingly, all three of these
possibilities have been suggested at some point in the
photosynthetic literature (although not always with the re-
cognition that such processes may be essential for the enzyme
to progress through its redox cycle).
PS II is formed from subunits that appeared by gene
duplication, suggesting that the redox-active tyrosine Yz
might have a symmetry-related and redox-active tyrosine in
the second subunit. This, in fact, is the case, with the second
tyrosine being given the designation tyrosine-D or YD.8 In
the old photosynthetic literature, Yz and YD are referred to
as signal 2fast and signal 2slow, respectively, based on their
electron paramagnetic resonance (EPR) relaxation behavior.
We now know that the lifetimes are a consequence of the
distance of the tyrosines from the paramagnetic manganese
cluster in the OEC. In addition, it appears that while Yz has
a potential of around 1 V, the oxidation potential of YD may
be significantly less than this (as low as 0.76 V).9 It has
long been known that PS II preparations that are kept in the
dark overnight will convert from S0 to S1, with the subsequent
reduction of YD to YD.10 Why this manganese oxidation
+
occurs and whether it is critical to the normal water oxidation
+
process have been subjects of keen speculation (e.g., YD
may protect the fragile S0 manganese cluster from destruction
by oxidizing it to the less labile S1 form). For us, however,
the raison d’etre of the process is unimportant, but simply
that it occurs is critical. This is because the S0 to S1 oxidation
by YD+ provides a lower limit for the manganese oxidation
potential in S0. This value then is ∼0.6 V. Therefore, if 3.6
V of the oxidizing power is required to oxidize water, if the
first step only provides 0.6 V to this sum, and if the maximal
potential on any transition is at most ∼1 V, then the
remaining transitions must be accumulating the maximum
available potential at each step. Now, of course, there is some
slop in these numbers. Possibly, S0-S1 occurs at 0.7 V, and
(11) Wincencjusz, H.; vanGorkom, H. J.; Yocum, C. F. Biochemistry 1997,
36, 3663.
(12) (a) Dasguptar, J.; van Willigen, R. T.; Dismukes, G. C. Phys. Chem.
Chem. Phys. 2004, 6, 4793. (b) Kovlov, Y. N.; Zharmukhamedov,
S. K.; Dismukes, G. C. Phys. Chem. Chem. Phys. 2004, 6, 4905.
(13) (a) Caudle, M. T.; Riggs-Gelasco, P.; Gelasco, A. K.; Penner-Hahn,
J. E.; Pecoraro, V. L. Inorg. Chem. 1996, 35, 3577. (b) Randall, D. W.;
Gelasco, A.; Caudle, M. T.; Pecoraro, V. L.; Britt, R. D. J. Am. Chem.
Soc. 1997, 119, 4481. (c) Caudle, M. T.; Pecoraro, V. L. Inorg. Chem.
2000, 39, 5831–5837.
(8) Faller, P.; Debus, R. J.; Brettel, K.; Sugiura, M.; Rutherford, A. W.;
Boussac, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 14368.
(14) (a) Junge, W.; Haumann, M.; Ahlbrink, R.; Mulkidjanian, A.; Clausen,
J. Philos. Trans. R. Soc. London, Ser. B 2002, 357, 1407. (b) Haumann,
M.; Mulkidjanian, A.; Junge, W. Biochemistry 1997, 36, 9304. (c)
Lavergne, J.; Junge, W. Photosynth. Res. 1993, 38, 279.
(9) (a) Boussac, A.; Etienne, A. L. Biochim. Biophys. Acta 1984, 766,
576. (b) Vass, I.; Styring, S. Biochemistry 1991, 30, 830.
(10) Styring, S.; Rutherford, A. W. Biochemistry 1987, 26, 2401.
Inorganic Chemistry, Vol. 47, No. 6, 2008 1767

Pecoraro and Hsieh
With the energetics of photosynthetic water oxidation now
in hand, one must consider the two great controversies
presently gripping researchers of the OEC: what are the
cluster oxidation states and what is the cluster structure? We
will first deal with the oxidation states of the manga-
nese clusters in each of the five Kok states. An accurate
knowledge of the manganese oxidation levels, especially in
the product (S0) and reactant (S4) states, is essential for any
serious consideration of mechanism. Given the limited scope
of this paper, we will not cover in detail the arguments for
or against each assignment in detail, rather limiting ourselves
to a statement of the controversies and indicating our
preferred oxidation state assignment. While it might seem
most logical to first consider the lowest oxidation levels for
the four manganese ions in the OEC, found in S0, the product
state resulting after the liberation of dioxygen, we will instead
first examine the oxidation levels of the dark stable S1 and
singly oxidized S2 states. The reason for starting in the middle
of the cycle is that for many years the S1 and S2 states were
the most easily accessed and studied oxidation levels of the
enzyme. Thus, historically these two states were the first with
assigned manganese oxidation levels and have served as the
underpinning for the assignment of all other manganese
oxidation levels in each of the other enzyme oxidation states.
Figure 4. XANES spectra of [Mn(2-OHSalpn)]2 complexes illustrating
how the edge energy increases as the oxidation state of the complex
increases. Spectra from left to right correspond to MnII2, MnIIMnIII, MnIII
and MnIIIMnIV. Reproduced from ref 7.
,
2
to the addition of p character into the d molecular orbital.16
Because XANES provides oxidation state information for
the average of the manganese ions and because the transitions
observed for MnII, MnIII, and MnIV are quite weak, the pre-
edge region has only been used for oxidation or structural
assignment in the extreme cases of multiply bonded metal
oxos (or nitridos).
In contrast, the X-ray edge, basically corresponding to the
photoelectron experiment, looks at the energy required to
promote the 1s core–electron into the continuum. A simpli-
fied, though not exact, way of understanding the relationship
between the oxidation state and edge energy is that the higher
the oxidation level of the ion being probed is, the harder it
will be to remove this core–electron and the higher will be
the energy necessary to remove this 1s electron. This general
trend is shown for a series of isostructural [Mn(2-OH
Salpn)]22- to 1+ complexes in Figure 4.7 Of course, other
factors such as ligand charge, type of bonding, and coordina-
tion number complicate this simple picture; however, the
general trend with manganese is that the higher oxidation
state usually leads to the higher edge energy. However, what
might appear to be a more philosophical question (what is
the edge energy?) has led to significant disagreement over
oxidation state assignment.
Oxidation State Assignments of Manganese in Each S
State
While numerous techniques have been used to determine
the manganese oxidation state, the approach that appears to
be the most generally reliable has been X-ray absorption
spectroscopy (XAS) and, more specifically, X-ray absorption
near-edge structure (XANES) spectroscopy. The following
comments are specifically intended for K-edge XANES
because this technique has been the most exhaustively
employed, although in recent years the L-edge technique has
also been probed to provide oxidation state information.15
The XANES spectrum provides two areas of interest known
as the pre-edge region and the X-ray edge. The pre-edge
region is the part of the spectrum where 1s electrons may
be absorbed into higher lying p or d orbitals. Thus, the pre-
edge region follows normal selection rules and provides
information about the electronic structure of the manganese
ions. Typically, for manganese, one sees weak absorption
due to the 1s f 3d transitions, which are symmetry-
forbidden; however, in certain special cases, such as the
presence of an MndO, one can see significant intensity due
Generally speaking, there are three ways in which edge
energies have been extracted from the XANES spectrum.
The first of these assigns the edge as the first inflection point
in the edge spectrum,17 which can be identified by the first
zero crossing in the derivative spectrum. The second ap-
proach defines the edge energy as the energy at half-height
(15) (a) Messinger, J.; Robblee, J. H.; Bergmann, U.; Fernandez, C.; Glatzel,
P.; Visser, H.; Cinco, R. M.; McFarlane, K. L.; Bellachio, E.; Pizarro,
S. A.; Cramer, S. P.; Sauer, K.; Klein, M. P.; Yachandra, V. J. Am.
Chem. Soc. 2001, 123, 7804. (b) Glatzel, P.; Bergmann, U.; Yano, J.;
Visser, H.; Robblee, J. H.; Gu, W.; de Groot, M. F.; Christou, G.;
Pecoraro, V. L.; Cramer, S. P.; Yachandra, V. K. J. Am. Chem. Soc.
2004, 126, 9946–9959.
(17) (a) Kirby, J. A.; Goodin, D. B.; Wyrdzynski, T.; Robertson, A. S.;
Klein, M. P. J. Am. Chem. Soc. 1981, 103, 5537. (b) Roelofs, T. A.;
Liang, W. C.; Latimer, M. J.; Cinco, R. M.; Rompel, A.; Andrews,
J. C.; Sauer, K.; Yachandra, V. K.; Klein, M. P. Proc. Nat. Acad. Sci.
U.S.A. 1996, 93, 3335. (c) Messinger, J.; Robblee, J. H.; Bergmann,
U.; Fernandez, C.; Glatzel, P.; Visser, H.; Cinco, R. M.; McFarlane,
K. L.; Bellachio, E.; Pizarro, S. A.; Cramer, S. P.; Sauer, K.; Klein,
M. P.; Yachandra, V. J. Am. Chem. Soc. 2001, 123, 7804. (d) Vissier,
H.; Anxolabehere-Mallart, E.; Bergmann, U.; Glatzel, P.; Robblee,
J. H.; Cramer, S. P.; Girerd, J.-J.; Sauer, K.; Klein, M. P.; Yachandra,
V. K. J. Am. Chem. Soc. 2001, 123, 7804. (e) Liang, W. C.; Roelofs,
T. A.; Cinco, R. M.; Rompel, A.; Latimer, M. J.; Yu, W. O.; Sauer,
K.; Klein, M. P.; Yachandra, V. J. Am. Chem. Soc. 2000, 122, 3399.
(16) (a) Weng, T. C.; Hsieh, W. Y.; Uffelman, E. S.; Gordon-Wylie, S. W.;
Collins, T. J.; Pecoraro, V. L.; Penner-Hahn, J. E. J. Am. Chem. Soc.
2004, 126, 8070. (b) Yano, J.; Robblee, J.; Pushkar, Y.; Marcus, M. A.;
Bendix, J.; Workman, J. M.; Collins, T. J.; Solomon, E. I.; DeBeer
George, S.; Yachandra, V. K. J. Am. Chem. Soc. 2007, 129, 12989.
(c) Song, W. J.; Seo, M. S.; DeBeer-George, S.; Ohta, T.; Song, R.;
Kang, M.-J.; Tosha, T.; Kitagawa, T.; Solomon, E. I.; Nam, W. J. Am.
Chem. Soc. 2007, 129, 1268.
1768 Inorganic Chemistry, Vol. 47, No. 6, 2008

ElusiWe High-Valent Manganese Species
of the edge.18 The third approach extracts the energy by
fitting a spline curve through all of the points of the edge.16,19
Each approach has its strengths and weaknesses, which will
not be explored herein. Suffice it to say that each approach
will give slightly different edge energies. We mentioned
above that the average oxidation state of the manganese
cluster can be deduced from the edge energy. To do so
requires that a series of well-characterized model compounds
of established oxidation state have also been probed by
XANES spectroscopy to build a calibration curve. If the
structure of the biological center is known, it is best to use
models that reflect the first coordination sphere ligands and
coordination number of the active site. Separate calibration
curves for each of these three different analytical methods
have been generated using a wide range of models. Regard-
less of the determination method of edge energies, all
workers in the field have agreed that the best formulation
for S1 is a mixture of MnIII and MnIV. A similar conclusion
can be drawn for the S2 state, and each analysis agrees that
the manganese ions in S2 are slightly more oxidized (by one
electron) than those in S1. A more precise assignment from
XANES alone is difficult; however, when EPR spectroscopy
is used in conjunction, a better understanding of oxidation
states emerges.
Figure 5. Comparison of a g ) 2 multiline spectrum of the OEC with that
of the “low-field” g ) 4 signal. The top spectrum is with illumination at
140 K for 30 s, the middle spectrum was warmed to 190 K in the dark for
90 s after illumination at 140 K, and the bottom spectrum is with illumination
at 190 K. All spectra were recorded at 10 K. Reproduced with permission
from ref 21.
The discovery of the multiline EPR spectrum in S2 by
Dismukes and Siderer20 was one of the most significant steps
in garnering a molecular understanding of the manganese
cluster in the OEC. The multiline spectrum, shown as Figure
5, contains at least 19 lines with hyperfine separations
consistent with a spin delocalized over a mixture of MnIII
and MnIV. The nuclear spin of 55Mn is 5/2 so one expects six
hyperfine lines for a mononuclear manganese ion such as
MnII. More than six hyperfine lines is definitive proof that
at least two and, in this case, as many as four manganese
ions are associated within the cluster. In addition, the
localization of the signal at g ) 2 (and the variable-
temperature dependence of the signal) indicates that the
multiline arises from an S ) 1/2 ground state. Because there
are four manganese ions, the only way that one can achieve
an S ) 1/2 ground state is for the cluster to be mixed valent.
The multiline spectrum (and models for it) has been probed
by numerous workers using techniques ranging from con-
tinuous-wave X-band, Q-band, and higher frequencies to
ENDOR and ESEEM methods.22–33 In addition, multiline
signals in the presence and absence of CaII, SrII, inhibitors
such as ammonia, small molecules such as methanol, and
myriad other conditions have been employed to unravel the
secrets of this cluster.34–42
If we take the XANES observations that S2 contains only
MnIII and MnIV and is one electron more oxidized that S1
(24) Teutloff, C.; Kessen, S.; Kern, J.; Zouni, A.; Bittl, R. FEBS Lett. 2006,
580, 3605.
(25) Zheng, M.; Dismukes, G. C. Photosynth. Res. 1992, 34, 144.
(26) Miller, A. F.; Brudvig, G. W. Biochim. Biophys. Acta 1991, 1056, 1.
(27) Beck, W. F.; DePaula, J. C.; Brudvig, G. W. J. Am. Chem. Soc. 1986,
108, 4018.
(28) DePaula, J. C.; Brudvig, G. W. J. Am. Chem. Soc. 1985, 107, 2643.
(29) Beck, W. F.; Brudvig, G. W. Biochemistry 1986, 25, 6479.
(30) Peloquin, J. M.; Campbell, K. A.; Randall, D. W.; Evanchik, M. A.;
Pecoraro, V. L.; Armstrong, W. H.; Britt, R. D. J. Am. Chem. Soc.
2000, 122, 10926.
(31) Peloquin, J. M.; Campbell, K. A.; Britt, R. D. J. Am. Chem. Soc. 1998,
120, 6840.
(32) Randall, D. W.; Sturgeon, B. E.; Ball, J. A.; Lorigan, G. A.; Chan,
M. K.; Klein, M. P.; Armstrong, W. H.; Britt, R. D. J. Am. Chem.
Soc. 1995, 117, 11780.
(33) Britt, R. D.; Peloquin, J. M.; Campbell, K. A. Annu. ReV. Biophys.
Biomol. Struct. 2000, 29, 463.
(34) Britt, R. D.; Zimmermann, J. L.; Sauer, K.; Klein, M. P. J. Am. Chem.
Soc. 1989, 111, 3522.
(18) (a) Ono, T.; Noguchi, T.; Inoue, Y.; Kusunoki, M.; Yamaguchi, H.;
Oyangi, H. J. Am. Chem. Soc. 1995, 117, 6386. (b) Ono, T.; Noguchi,
T.; Inoue, Y.; Kusunoki, M.; Yamaguchi, H.; Oyangi, H. Science 1992,
258, 1335.
(35) Kim, D. H.; Britt, R. D.; Klein, M. P.; Sauer, K. Biochemistry 1992,
31, 541.
(36) Boussac, A.; Rappaport, F.; Carrier, P.; Verbatz, J. M.; Gobin, R.;
Kirilovsky, D.; Rutherford, A. W.; Sugiura, M. J. Biol. Chem. 2004,
279, 22809.
(19) (a) Dau, H.; Iuzzolino, L.; Dittmer, J. Biochim. Biophys. Acta 2001,
1503, 24. (b) Kuntzleman, T.; McCarrick, M.; Penner-Hahn, J.;
Yocum, C. Phys. Chem. Chem. Phys. 2004, 6, 4897. (c) Riggs-Gelasco,
P. J.; Mei, R.; Yocum, C. F.; Penner-Hahn, J. E. J. Am. Chem. Soc.
1996, 118, 2387. (d) Meinke, C.; Sole, V. A.; Pospisil, P. J.; Dau, H.
Biochemistry 2000, 39, 7033.
(37) Tyryshkin, A. M.; Watt, R. W. K.; Baranov, S. V.; Dasgupta, J.;
Hendrich, M. P.; Dismukes, G. C. Biochemistry 2006, 45, 12876.
(38) Britt, R. D.; Campbell, K. A.; Peloquin, J. M.; Gilchrist, M. L.; Aznar,
C. P.; Dicus, M. M.; Roblee, J.; Messinger, J. Biochim. Biophys. Acta
2004, 1655, 158.
(20) Dismukes, G. C.; Siderer, Y. Proc. Natl. Acad. Sci. U.S.A. 1981, 78,
274.
(39) Sivaria, M.; Tso, J.; DIsmukes, G. C. Biochemistry 1989, 28, 9459.
(40) Gilchrist, M. L.; Ball, J. A.; Randall, D. W.; Britt, R. D. Proc. Natl.
Acad. Sci. U.S.A. 1995, 92, 9545.
(21) Casey, J. L.; Sauer, K. Biochim. Biophys. Acta 1984, 767, 21.
(22) Chalot, M. F.; Boussac, A.; Blondin, G. Biochim. Biophys. Acta 2005,
1708, 120.
(41) Carrell, T. G.; Tyryshkin, A. M.; Dismukes, G. C. J. Biol. Inorg. Chem.
2002, 7, 2.
(23) (a) Boussac, A. Biochim. Biophys. Acta 1996, 1277, 253. (b) Boussac,
A. Chem. Phys. 1995, 194, 409.
(42) Campbell, K. A.; Peloquin, J. M.; Pham, D. P.; Debus, R. J.; Britt,
R. D. J. Am. Chem. Soc. 1998, 120, 447.
Inorganic Chemistry, Vol. 47, No. 6, 2008 1769

Pecoraro and Hsieh
basically believes that oxidation of the enzyme occurs on
either a ligand or a protein residue, but not at the manganese
atoms.17 The second group is a proponent of manganese-
centered oxidation on this transition.18,19,44–46 The two
oxidation formulations can thus be summarized as MnIII
MnIV3(substrate)• or MnIIIMnIV3(protein)• for the first model
and MnIV for the second model. The primary basis of this
4
disagreement is associated with the interpretation of the
XANES spectra. As discussed above, different groups use
different methodologies to define the edge energy. While the
differences associated with the different methods are not a
significant issue when looking at any one S state, the
comparison becomes more difficult when making compari-
sons between different methods making substractions from
adjacent S states. Thus, some workers estimate that the
change in edge energy is inconsistent with a manganese-
centered oxidation, whereas other workers obtain larger
edge-energy differences that are more consistent with
manganese-centered oxidation occurring. Compounding these
methodological differences is the fact that sample prepara-
tion is crucial for obtaining high-quality results and that
there is an inherent difficulty in obtaining clean S-state
transitions. Early PS II preparations suffered from having
either double hits (advancement of the system by two S
states with one photon) or misses (despite the addition of
a photon, the enzyme is not oxidized).47 Modern laser
methods minimize the number of misses; however, double
hits still occur. Thus, depending on the quality of the initial
sample, the number of double hits occurring after illumina-
tion and the edge analysis all contribute to the various
discrepancies. To circumvent these problems, some research-
ers have gone to L-edge spectra,17 while others have looked
at various types of EPR spectroscopy of S3 (parallel mode48
because perpendicular-mode EPR is silent, as might be
expected for manganese oxidation). Some models have even
suggested the formation of MnIVdO (oxyl groups) at this
oxidation level.49 We will simply conclude that, in our
opinion, the oxidation level of S3 remains unresolved;
however, our reading of the literature gives us the preference
for manganese-centered oxidation, yielding MnIV4. What can
be said definitively about S3, based on XANES pre-edge
spectra, is that this enzyme form does not contain a
manganese-oxyl such as MnIVdO.16
Figure 6. S1-state parallel-mode EPR spectrum (top). This spectrum is
compared to a tetranuclear MnIV cluster (bottom). Reproduced from ref 43.
4
and combine this with the EPR observation that S2 contains
an odd number of MnIV ions (this necessity comes from the
fact that MnIV is a d3 ion and only an odd number of such
ions can yield a nonintegral spin system), then S2 should be
MnIIIMnIV3. A formulation of MnIII3MnIV is unlikely because
S1 must contain both MnIII and MnIV, and we know from
the above discussion that the transition from S1 to S2 requires
a manganese-centered oxidation. Generally speaking, fits to
ESEEM, ENDOR, and EPR spectra are consistent with this
assignment, although there have been claims that the
MnIII3MnIV oxidation formulation better fits simulations to
the different g ) 2 multilines formed with the native form
and modified multilines.41
Given the assignment of S2 as MnIII3MnIV, then one-
electron reduction of this cluster should yield S1 as
MnIII2MnIV2. Such a non-Kramers ion is not expected to show
an EPR signal at X-band frequencies using perpendicular-
mode detection (the standard cavity alignment); however, it
is possible that a parallel-mode-detected experiment can
allow for observation of an EPR signal from this integral
spin cluster. Figure 6 shows the spectrum that Britt’s group
isolated for the S1 state.42 It too has a multiline structure,
but the resonance is centered at lower field. This OEC S1
signal is compared with the only known spectroscopic model
for this S state in Figure 6.43 The complex {[Mn(2-
OHSalpn)]4O}2+ is thought to contain four MnIV ions. As
such, it has a slightly reduced hyperfine coupling compared
with the enzymatic signal. This observation is consistent with
the S1 state being slightly reduced compared to the model,
which is fully consistent with the assignment of S1 as
Because one cannot define the S3 state conclusively, it
seems a bit premature to provide an assignment for the
(44) Pecoraro, V. L.; Hsieh, W. Y. In Metal Ions in Biological Systems;
Sigel, H., Sigel, A., Eds.; Marcel Dekker: New York, 2000; Vol. 37,
p 429.
(45) Pecoraro, V. L.; Baldwin, M. J.; Caudle, M. T.; Hsieh, W. Y.; Law,
N. A. Pure Appl. Chem. 1998, 70, 925.
MnIII2MnIV
.
2
(46) (a) Vrettos, J. S.; Limburg, J.; Brudvig, G. W. Biochim. Biophys. Acta
2001, 1503, 229. (b) Vrettos, J. S.; Brudvig, G. W. Philos. Trans. R.
Soc. London, Ser. B 2002, 357, 1395. (c) Limburg, J.; Szalai, V. A.;
Brudvig, G. W. J. Chem. Soc., Dalton Trans. 1999, 1353.
(47) (a) Shinkarev, V. Biophys. J. 2005, 88, 412. (b) Jursinic, P. Biochim.
Biophys. Acta 1981, 635, 38.
While the consensus of workers in photosynthesis is in
agreement with these oxidation state assignments, the relative
unanimity disappears when the remaining S states are
discussed. Nowhere is the discussion more vigorous than
when considering the transition from S2 to S3. One camp
(48) (a) Ioannidis, N.; Nugent, J. H. A.; Petrouleas, V. Biochemistry 2002,
41, 9589. (b) Ioannidis, N.; Petrouleas, V. Biochemistry 2002, 41, 9580.
(c) Ioannidis, N.; Zahariou, G.; Petrouleas, V. Biochemistry 2006, 45,
6252.
(43) Hsieh, W. Y.; Campbell, K. A.; Gregor, W.; Britt, R. D.; Yoder, D. W.;
Penner-Hahn, J. E.; Pecoraro, V. L. Biochim. Biophys. Acta 2004,
1655, 149.
(49) (a) Hoganson, C.; Babcock, G. T. Science 1997, 277, 1953. (b)
Tommos, C.; Babcock, G. T. Acc. Chem. Res. 1998, 31, 18.
1770 Inorganic Chemistry, Vol. 47, No. 6, 2008

ElusiWe High-Valent Manganese Species
catalytically active S4 level. Nonetheless, if one wants to
assess mechanistic options for water oxidation, some working
hypothesis for S4 must be forwarded. The same questions
that plagued the S2 f S3 transition again arise: Is the enzyme
oxidation centered on manganese or elsewhere and, if
elsewhere, is the location a protein residue or the substrate?
It is known that oxygen production is concurrent with
rereduction of Yz+; therefore, it seems highly unlikely that
a protein residue (i.e., amino acid) could first be oxidized
followed by substrate or metal oxidation and still follow these
fast kinetics. If oxidation to form S3 yields a substrate radical,
then formation of S4 could consist of a biradical substrate
species (i.e., two oxygen radicals bound to the cluster).
formulation is not without precedent as Armstrong54 dem-
onstrated that MnII, MnIII, and MnIV could coexist in a single
structure. Subsequently, Kessissoglou et al.55 reported a
tetrameric cluster with manganese composition MnII MnIV
3
and we prepared a trinuclear MnIIMnIVMnII system.56 In an
effort to assess whether XANES spectroscopy was capable
of distinguishing oxidation state formulations for redox
isomers, we examined two Mn38+ structures having composi-
tions MnIIMnIVMnII and MnIIIMnIIMnIII 57 and showed that,
at least for trinuclear systems, XANES could distinguish
outside of experimental error the proper oxidation assign-
ments for these molecules.56 A recent study56c suggests that
10+
this is a more difficult proposition when tetranuclear Mn4
complexes are studied (i.e., MnII MnIII vs MnII MnIV). Not
Because we prefer the MnIV formalism for S3, we are left
2
2
3
4
surprisingly, more recent studies have supported the previous
MnIII3MnIV.58 In this case, it may be that both formulations
are correct because the answer that one gets may be related
to how S0 is prepared. Thus, for the purposes of this Article,
we will assume that either MnIII3MnIV or MnIIMnIIIMnIV2 can
exist and that they may even interconvert under appropriate
circumstances.
In summary, the oxidation state assignments of the
manganese ions in the OEC are still controversial. Consensus
supports manganese oxidation from S0 to S1 and from S1 to
S2. There are two schools of thought on enzyme oxidation
beyond S2, with manganese-centered, protein-centered, and
ligand-centered oxidation all proposed. The S1 level is almost
certainly a mixture of MnIII and MnIV, with MnIII2MnIV2 most
likely. Formation of S2 probably yields MnIIIMnIV3. Our
preference that this is to form a MnIV3MnV species in S4,
although computational studies support an MnIV4(substrate)•.
After dioxygen is released, S0 may contain either MnIIMnII-
IMnIV2 or MnIII3MnIV. With this understanding of the single
ion structure of the manganese atoms, we will now consider
the structure of the CaMn4 cluster.
with either a MnIV3MnV or a MnIV4(substrate)• species. The
ligand usually invoked for the latter possibility is a
manganyl-oxo atom50 or a bridging µ-oxo ligand.17 There
have been no experiments that directly probe these three
possibilities; however, computational studies have tended to
support the MnIVO(radical) except for recent QM/MM
studies.51 In some senses, the difference may be semantic
in that a molecular orbital with significant manganese and
oxygen character could be the source of this last electron.
However, there is a subtle difference in the reaction pathway
that would be predicted between these two formulations. This
difference will be discussed in more detail below.
Before leaving manganese oxidation assignments, we must
address the product state of the system, S0. For years, it was
tacitly assumed that if S1 contained MnIII2MnIV2, then the
one-electron-reduced enzyme would best be formulated as
MnIII3MnIV. Such a system should be EPR-active, and yet it
was only within the past decade that a perpendicular-mode
X-band spectrum has been reported (and only in the presence
of methanol).52 The g ) 2 multiline signal for S0 suggests
1
an S ) /2 ground state, and the spacing of the hyperfine
Structure of the Mn4Ca Cluster
lines is consistent with a lower oxidation level than the S2
multiline previously reported. Subsequent analysis of the EPR
spectrum seemed to indicate that MnII could be present in
S0. Further XANES data were also consistent with this
idea.17a,53 If MnII is present in S0, then one must reevaluate
the total manganese oxidation assignment. The simplest
possibility is that S0 could have manganese existing in three
different oxidation levels such as MnIIMnIIIMnIV2. Such a
While there is an extremely rich and interesting scientific
path that has led to our present day understanding of the
cluster responsible for water oxidation, this Article can
provide little more than the briefest summary of this en-
deavor. Interested readers are directed to a series of excellent
(54) Chan, M. K.; Armstrong, W. H. J. Am. Chem. Soc. 1990, 112, 4985.
(55) Afrati, T.; Dendrinou-Samara, C.; Raptopoulou, C. P.; Terzis, A.;
Tangoulis, V.; Kessissoglou, D. P. Angew. Chem., Int. Ed. 2002, 41,
2148.
(50) (a) Lundberg, M.; Siegbahn, P. E. M. Phys. Chem. Chem. Phys. 2004,
6, 4772. (b) Siegbahn, P. E. M.; Blomberg, M. R. A. Philos. Trans.
R. Soc. London, Ser. A 2005, 363, 847. (c) Lundberg, M.; Blomberg,
M. R. A.; Siegbahn, P. E. M. Inorg. Chem. 2004, 43, 264. (d)
Lundberg, M.; Siegbahn, P. E. M. Phys. Chem. Chem. Phys. 2004, 6,
4772. (e) Siegbahn, P.; Crabtree, R. H. J. Am. Chem. Soc. 1999, 121,
117.
(56) (a) Alexiou, M.; Dendrinou-Samara, C.; Karagianni, A.; Biswas, S.;
Zaleski, C. M.; Kampf, J.; Yoder, D.; Penner-Hahn, J. E.; Pecoraro,
V. L.; Kessissoglou, D. P. Inorg. Chem. 2003, 42, 2185. (b) Alexiou,
M.; Zaleski, C. M.; Dendrinou-Samara, C.; Kampf, J.; Kessissoglou,
D. P.; Pecoraro, V. L. Z. Anorg. Allg. Chem. 2003, 629, 2348. (c)
Zaleski, C.; Weng, T.-C.; Dendrinou-Samara, C.; Alexiou, M.;
Kanakaraki, P.; Hsieh, W.-Y.; Kampf, J. W.; Penner-Hahn, J. E.;
Pecoraro, V. L.; Kessissoglou, D. P. submitted to Inorg. Chem.
(57) (a) Kessissoglou, D. P.; Kirk, M. L.; Lah, M. S.; Li, X.; Raptopoulou,
C.; Hatfield, W. E.; Pecoraro, V. L. Inorg. Chem. 1992, 31, 5424. (b)
Kessissoglou, D. P.; Kirk, M. L.; Bender, C. A.; Lah, M. S.; Pecoraro,
V. L. J. Chem. Soc., Chem. Commun. 1989, 84. (c) Li, X.; Kessis-
soglou, D. P.; Kirk, M. L.; Bender, C. A.; Pecoraro, V. L. Inorg. Chem.
1988, 27, 1.
(51) (a) Spoviero, E. M.; Gascon, J. A.; McEvoy, J. P.; Brudvig, G. W.;
Batista, V. S. J. Chem. Theor. Comput. 2006, 2, 1119. (b) Batista,
V. S. Philos. Trans. R. Soc. London, Ser. B,in press..
(52) (a) Messinger, J.; Nugent, J. H. A.; Evans, M. C. W. Biochemistry
1997, 36, 11055. (b) Ahrling, K. A.; Peterson, S.; Styring, S.
Biochemistry 1997, 36, 13148. (c) Messinger, J.; Robblee, J.; Yu,
W. O.; Sauer, K.; Yachandra, V. K.; Klein, M. P. J. Am. Chem. Soc.
1997, 119, 11349.
(53) Yachandra, V. K.; Guiles, R. D.; McDermott, A.; Britt, R. D.; Cole,
J.; Dexheimer, S. L.; Sauer, K.; Klein, M. P. J. Phys., Colloq. 1986,
2, 1121.
(58) (a) Kulik, L. V.; Epel, B.; Lubitz, W.; Messinger, J. J. Am. Chem.
Soc. 2007, 129, 13421. (b) Kulik, L. V.; Epel, B.; Lubitz, W.;
Messinger, J. J. Am. Chem. Soc. 2005, 127, 2392.
Inorganic Chemistry, Vol. 47, No. 6, 2008 1771

Pecoraro and Hsieh
reviews that describe the shifting thoughts from manganese
cubane-like clusters,59 to independent trimer plus mono-
mers,57,60 to dimer of dimers,61 and ultimately to the dangler
model30 (or a general tetranuclear 3 + 1 model44). We also
will not have time to discuss the past controversy on the
location of CaII with respect to the manganese centers.
Instead, we will begin with the thinking in the late 20th
century where the controversy focused on the likelihood of
a dimer of dimers topology versus a less symmetric 3 + 1
or dangler model and then move into the new millennium
studies that have been dominated by crystallographic analysis
of the system.
Earlier, we discussed the importance of XAS (using
XANES) to establish the oxidation states of the manganese
in the OEC. An even greater contribution of this technique
has been the elucidation of the structure of the manganese
center using extended X-ray absorption fine structure
(EXAFS) spectroscopy, which provides element-specific
metrical and coordination number/type information. It has
been known for many years62 that some of the manganese
ions are linked at a 2.7 Å distance while either a Mn-Mn
or Mn-Ca vector at 3.3 Å also existed.62b The 2.7 Å distance
is the hallmark of manganese bridged with µ2- or µ3-oxo
ligands as seen in numerous dimeric and higher nuclearity
systems.63 From these studies and knowledge of the EPR
spectral behavior of the cluster, a variety of structural
possibilities were proposed, with the structure that emerged
as most interesting being the dimer of dimers. If one
examines the literature of the 1990s, it is difficult to find a
water oxidation model that does not invoke, at least in part,
this topology. However, we44,45,64 and others65 were con-
cerned about certain observations that seemed to be incom-
patible with such an orientation of the manganese ions. Chief
among these issues for us was the observation of a second
S2 EPR signal21 at low field (∼g ) 4) that could be formed
under conditions alternative to those responsible for the g
) 2 multiline signal described by Dismukes. It was shown
using EXAFS that the manganese cluster that exhibited this
low-field signal was essentially structurally invariant from
that of the form giving the g ) 2 multiline. Yet, the g ) 4
signal had a ground state of higher spin multiplicity (probably
S ) 5/2).66 Given the oxidation assignments for S2 (MnIIIM-
nIV3) and the observation that all µ2-oxo-bridged manganese
complexes were antiferromagnetically coupled (even those
protonated or alkylated),67 it became clear that a simple dimer
of dimers structure was unlikely, if not impossible. We
realized that a more complex relationship between the
manganese ions was necessary to explain the observed EPR
features while also complying with the EXAFS constraints.
It seemed necessary to us to have a more complex coupling
scheme in order to achieve the higher spin ground state
suggested by the S2 low-field EPR signal. We proposed as
an alternative the 3 + 1 structure shown in Figure 7A.44
At the same time that we were shifting our thinking to
this 3 + 1 topology, Britt’s group was completing a detailed
analysis of the ESEEM spectra of the g ) 2 multiline
signal.30 His group reasoned that the dimer of dimers model
could not adequately explain this EPR signal and proposed
that one must have a special manganese distinct from the
other three manganese ions in order to reproduce properly
the observed spectra. The topology originally proposed to
account for this was a linear orientation of manganese ions,
with a special fourth manganese that became known as “the
dangler”. It should be noted, however, that the arguments
that Britt forwarded are perfectly consistent with the more
three-dimensional model that we had proposed in Figure 7A.
At this point, there were two independent approaches that
were casting doubt on the dimer of dimers model. Of course,
any model derived from the sporting techniques of spec-
troscopy or magnetism are often a bit tenuous, and so it came
as a great relief when the first X-ray structure of the OEC
appeared to generally validate the idea that the manganese
ions were arranged in a 3 + 1 configuration.68 Unfortunately,
there were issues with this structure including its low
resolution and relatively poor modeling of the metal cluster.
In addition, the workers could not address the issue of the
presence or absence of calcium. A more satisfying model
appeared a few years later, when Barber’s group reported a
slightly higher resolution diffraction set that could be refined
with a new model to higher certainty.69 More importantly,
by varying the synchrotron energy, these workers were able
to distinguish the location of the calcium atom with respect
to the manganese atoms. Their model for the OEC is shown
(59) Brudvig, G. W.; Crabtree, R. H. Proc. Natl. Acad. Sci. U.S.A. 1986,
83, 4586. (b) DePaula, J.; Beck, W.; Brudvig, G. W. J. Am. Chem.
Soc. 1986, 108, 4002.
(60) (a) Kessissoglou, D. P.; Butler, W. M.; Pecoraro, V. L. J. Chem. Soc.,
Chem. Commun. 1986, 1253. (b) Kessissoglou, D. P.; Li, X. H.; Butler,
W. M.; Pecoraro, V. L. Inorg. Chem. 1987, 26, 2487.
(61) Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji, I.; Sauer,
K; Klein, M. P. Science 1993, 260, 675.
(62) (a) Kirby, J. A.; Robertson, A. S.; Smith, J. P.; Thompson, A. C.;
Cooper, S. R.; Klein, M. P. J. Am. Chem. Soc. 1981, 103, 5529. (b)
DeRose, V. J.; Mukerji, I.; Latimer, M. J.; Yachandra, V. K.; Sauer,
K.; Klein, M. P. J. Am. Chem. Soc. 1994, 116, 5239.
(63) (a) Cooper, S. R.; Dismukes, G. C.; Klein, M. P.; Calvin, M. J. Am.
Chem. Soc. 1978, 100, 7248. (b) Larson, E.; Lah, M. S.; Li, X.;
Bonadies, J. A.; Pecoraro, V. L. Inorg. Chem. 1992, 31, 373. (c)
Larson, E.; Pecoraro, V. L. J. Am. Chem. Soc. 1991, 113, 3810. (d)
Wang, S. Y.; Folting, K.; Streib, W. E.; Schmitt, E. A.; McCusker,
J. K.; Hendrickson, D. N.; Christou, G. Angew. Chem., Int. Ed. 1991,
30, 305–306. (e) Bashkin, J. S.; Chang, H. R.; Streib, W. E.; Huffman,
J. C.; Hendrickson, D. N.; Christou, G. J. Am. Chem. Soc. 1987, 109,
6502–6504. (f) Hendrickson, D. N.; Christou, G.; Schmitt, E. A.;
Libby, E.; Bashkin, J. S.; Wang, S. Y.; Tsai, H. L.; Vincent, J. B.;
Boyd, P. D. W.; Huffman, J. C.; Folting, K.; Li, Q. Y.; Streib, W. E.
J. Am. Chem. Soc. 1992, 114, 2455.
(66) (a) Haddy, A.; Dunham, Wr.; Sands, R. H.; Aasa, R. Acta Biochim.
Biophys. 1992, 1099, 25. (b) Horner, O.; Riviere, E.; Blondin, G.;
Un, S.; Rutherford, A. W.; Girerd, J.-J.; Boussac, A. J. Am. Chem.
Soc. 1998, 120, 7924.
(67) (a) Baldwin, M. J.; Gelasco, A.; Pecoraro, V. L. Photosynth. Res. 1993,
38, 303. (b) Baldwin, M. J.; Stemmler, T.; Riggs-Gelasco, P. J.; Kirk,
M. J.; Penner-Hahn, J. E.; Pecoraro, V. L. J. Am. Chem. Soc. 1994,
116, 11349. (c) Baldwin, M. P.; Law, N. A.; Stemmler, T. L.; Kampf,
J. W.; Penner-Hahn, J. E.; Pecoraro, V. L. Inorg. Chem. 1999, 38,
4801–4809. (d) Larson, E. L.; Riggs, P.; Penner-Hahn, J. E.; Pecoraro,
V. L. J. Chem. Soc., Chem. Commun. 1992, 102.
(64) Law, N. A.; Kampf, J. W.; Pecoraro, V. L. Inorg. Chim. Acta 2000,
297, 252.
(68) Zouni, A.; Witt, H.-T.; Kern, J.; Fromme, P.; Krauss, N.; Saenger,
W.; Orth, P. Nature 2001, 409, 739.
(65) Kirk, M. L.; Chan, M. K.; Armstrong, W. H.; Solomon, E. I. J. Am.
Chem. Soc. 1992, 114, 10432.
(69) Ferreira, K. N.; Iverson, T. M.; Maghlaoui, K.; Barber, J.; Iwata, S.
Science 2004, 303, 1831.
1772 Inorganic Chemistry, Vol. 47, No. 6, 2008

ElusiWe High-Valent Manganese Species
Figure 7. (A) Proposed structure of the Mn4 cluster based on a magnetic analysis of the OEC in ref 44. (B) ChemDraw representation of the manganese
ions as deduced from the Barber structure for the OEC. Notice that one bond has been broken and one added between manganese ions as compared to that
given in part A. (C) The calcium ion has now been added to the Barber representation (D).
topology so the general metal atom arrangement would not
be expected to match the X-ray structure; however, if one
considers the related 3 + 1 structure44 shown in Figure 7, a
direct comparison can be made. Remarkably, one needs to break
only one bond and remake a new bond from the drawing shown
as Figure 7A to obtain Barber’s topology shown in Figure 7B.
Thus, the structure predicted based on the sporting methods
was in remarkably good accord with these crystallographic
results. Of course, CaII, being diamagnetic, does not contribute
directly to the magnetics and so could not be accounted for in
the Figure 7A model. Its location is shown in Figure 7C, and
one realizes immediately the impressive similarity between the
X-ray structure and our prediction.
At times, it seems as if nothing in photosynthetic research
stands the test of time and so it is, in part, with the
crystallographic analysis. Yachandra et al. carried out a
Figure 8. Mn4Ca cluster derived from the Barber X-ray structure of the
detailed study of the X-ray damage associated with the data
OEC. Reproduced with permission from ref 69.
collection on the X-ray crystals.70 It had been assumed,
though not really stated, that the structures that were
appearing were of the S1 state because it is the most stable
schematically in parts B and C of Figure 7, and electron
densities are given in Figure 8.
form of the enzyme. However, it was shown that, under the
The Barber model places three manganese ions and one
high-intensity radiation necessary to collect diffraction data,
calcium ion in an CaMn3 cubane-like cluster. The fourth
manganese ion appears to be off to the side, “dangling” from
the cluster as predicted by Britt and co-workers. As already
stated, the original dangler model was presented as a linear
(70) Yano, J.; Kern, J.; Irrgang, K. D.; Latimer, M. J.; Bergmann, U.;
Glatzel, P.; Pushkar, Y.; Biesiadka, J.; Loll, B.; Sauer, K.; Messinger,
J.; Zouni, A.; Yachandra, V. K. Proc. Natl. Acad. Sci. U.S.A. 2005,
102, 12047.
Inorganic Chemistry, Vol. 47, No. 6, 2008 1773

Pecoraro and Hsieh
Figure 9. Multiple representations of the Mn4Ca cluster based on polarized EXAFS data. Notice the shift in the position of the calcium ion with respect
to the three manganese ions of the triangle. Also note in this model that a di-µ2-oxo link to the dangler is suggested, which would restrict the Mn-Mn
separation to a shorter distance. Reproduced with permission from ref 71.
there was rapid reduction of the manganese cluster. This
observation, of course, throws into question the conclusions
regarding the specific ligation and distance details from the
crystallographic experiments. Certainly, one would expect
some changes, possibly very large, upon reduction of an
structure no longer is a cubane being somewhat more
distorted). Of course, what is significant to one person71 is
not to another, and so one may read an opposing view of
this statement in a separate contribution in this issue. It should
be noted, however, that the diagram in Figure 7A can as
easily be converted into this EXAFS-derived topology as
the structure proposed by Barber had been.
In summary, the most recent evidence strongly suggests
that the manganese ions are arranged in a Mn4Ca cluster.
The structure is probably something like a Mn3Ca cubane
(or a distorted structure), with a fourth manganese atom
linked to this cluster forming the dangler manganese that
was predicted from spectroscopic and magnetic measure-
ments. In the following section, we will now try to bring
together all of the separate parts so far discussed (energetics,
oxidation states, and structure) to assemble a model for the
reactivity of the OEC and show how synthetic models are
being used to test some of these ideas.
MnIII2MnIV cluster to nearly all MnII. Fortunately, EXAFS
2
experiments use a lower photon flux than is necessary for
X-ray diffraction experiments so that data from this technique
can still be used to assess structure. Furthermore, preparations
of PS II that have been aligned can be used in oriented
EXAFS experiments to propose a polarized EXAFS-derived
model for the OEC.71 This structure is given in Figure 9.
Ironically, these authors have used the coordinates based on
the X-ray structures to help orient the cluster, and a newer,
higher-resolution structure has appeared from the Berlin
group.72 Our reading of this work is that while there are some
important refinements to the structure proposed by Barber
and movement of manganese/calcium atoms, the basic
topology really is not altered dramatically (the Mn3Ca
Model for Photosynthetic Water Oxidation
(71) Yano, J.; Kern, J.; Sauer, K.; Latimer, M. J.; Pushkar, Y.; Biesiadka,
J.; Loll, B.; Saenger, W.; Messinger, J.; Zouni, A.; Yachandra, V. K.
Science 2006, 314, 821.
Instructors in general chemistry often refer to the number
of grains of sand found on the beaches of the world in order
to help students grasp the concept of a mole. An alternative
(72) Loll, B.; Kern, J.; Saenger, W.; Zouni, A.; Biesiadka, J. Nature 2005,
438, 1040.
1774 Inorganic Chemistry, Vol. 47, No. 6, 2008

ElusiWe High-Valent Manganese Species
example could be the number of chemical models that have
been proposed for photosynthetic water oxidation. For
obvious reasons, we will not discuss all of these proposals,
some of which appear in the complementary Articles of this
issue. Instead, we will focus on a suggestion that appeared45
in 1998 and that has been expanded on very thoughtfully by
Brudvig and co-workers.46
For a number of years, we were focused on methods of
generating dioxygen from a dimer of dimers model. Most
promising in this regard were two observations for the reactivity
of [Mn(Salpn)(µ2-O)]2. The first of these was reported73 by
Boucher and Coe in which the acidification of this complex
was reported to generate H2O2 and two MnIII(Salpn)(solvent)
molecules. The second report74 came from our group in
which we demonstrated that hydrogen peroxide underwent
a vigorous catalytic reaction that generated dioxygen and
regenerated the original manganese catalyst. Combining these
two reactions led to a hypothetical system that would make
dioxygen from water. Because of the importance of these
observations, we explored more fully the hydrogen per-
oxide step. First, we demonstrated that the oxo groups of the
complex could be protonated,67 that this led to measurable
changes in the structure and magnetics of the complex, and that
these protons inhibit the catalase reactivity. More importantly,
Law demonstrated in his Ph.D. thesis75 that hydrogen peroxide
was not formed by the addition of two proton equivalents to
[Mn(Salpn)(µ2-O)]2 but rather that the Salpn ligand was oxidized
in the process. This discouraging observation occurred con-
temporaneously with our revised view of the manganese cluster
as a 3 + 1 structure described above.
At this point, we realized that one of the manganese ions
appeared to be structurally “special”. Furthermore, the
Berkeley group was acquiring strong evidence that the
calcium atom was in close proximity to the manganese
cluster. These two thoughts led us to propose a function for
calcium that placed it in the heart of the water oxidation
reaction, as shown in Figure 10. The idea sprang from the
recognition that metalloenzymes containing alkali or alkaline-
earth ions sometimes used these ions to polarize water
molecules in order to facilitate a nucleophilic attack on a
substrate. We reasoned that the special manganese may be
able to reach the MnV oxidation level, which in the presence
of water would likely be found as a MnVdO moiety. This
highly polarized oxo group should be particularly electro-
philic and could, in principle, be attacked by a strongly
nucleophilic hydroxide ion bound to CaII. The result would
be a manganese(III) hydroperoxo species [MnIIIOOH], which
Figure 10. Perpendicular-mode EPR spectra of {MnIV2(2-OHSalpn)2OH}+
at (A) 4.2 K, (B) 10 K, and (C) 16 K. The manganese complex was dissolved
in butyronitrile (2 mM): microwave frequency, 9.432 GHz; power, 20 mW;
modulation amplitude, 10 G; modulation frequency, 100 kHz.
could subsequently decompose with the redox assistance of
the remaining trinuclear manganese assembly to S0 and
dioxygen. Brudvig and co-workers have nicely developed a
detailed mechanism accounting for proton movement and
oxidation through the S clock based on these ideas.46
The next issue for us as coordination chemists is to prepare
molecules that can test aspects of this proposal. Several
reports76–78 have made claims for water oxidation using
manganese model compounds. In these cases, high-valent
manganese has been invoked in the critical oxidation step.
While exciting, clearly more work is needed in this area,
especially given controversies associated with the chemistry.79
Our efforts have focused on preparing manganese compounds
for binuclear MnIVMnV systems, particularly those containing
the MnVdO moiety. We have examined extensively the
chemistry of ring-substituted derivatives of 2-OHSalpn with
manganese.80 Early in this Article,13 we discussed the ability
of this compound to be oxidized through multiple oxidation
levels beginning with the MnII dimer [Mn(2-OHSalpn)]22-
and progressing as high as the MnIII/MnIV complex, [Mn(2-
OHSalpn)]2+. The substitution of water leads to the asym-
metric complexes {[MnIII(2-OHSalpn)]2H2O}, {MnIIIMnIV(2-
(73) Boucher, L. J.; Coe, C. G. Inorg. Chem. 1976, 15, 1334.
(74) Larson, E.; Pecoraro, V. L. J. Am. Chem. Soc. 1991, 113, 7809.
(75) Law, N. Models for the Reactivity of Mn di-µ2-oxo cores in the
Oxygen Evolving Complex. Ph.D. Dissertation, University Microfilm,
1999.
(76) (a) Limburg, J.; Vrettos, J. S.; Liable-Sands, L. M.; Rheingold, A. L.;
Crabtree, R. H.; Brudvig, G. W. Science 1999, 283, 1524. (b) Chen,
H. Y.; Faller, J. W.; Crabtree, R. H.; Brudvig, G. W. J. Am. Chem.
Soc. 2004, 126, 7345. (c) Limburg, J.; Vrettos, J. S.; Chen, H. Y.;
dePaula, J. C.; Crabtree, R. H.; Brudvig, G. W. J. Am. Chem. Soc.
2001, 123, 423.
(77) Yagi, M.; Narita, K. J. Am. Chem. Soc. 2004, 126, 8084.
(78) Poulsen, A. K.; Rompel, A.; MacKenzie, C. J. Angew. Chem., Int.
Ed. 2005, 44, 6916.
(79) Baffert, C.; Romain, S.; Richardot, A.; LePetre, J. C.; Lefebvre, B.;
Deronzier, A.; Collomb, M.-N. J. Am. Chem. Soc. 2005, 127, 13694.
(80) Gelasco, A.; Pecoraro, V. L. J. Am. Chem. Soc. 1993, 115, 7928.
Inorganic Chemistry, Vol. 47, No. 6, 2008 1775

Pecoraro and Hsieh
OHSalpn)2H2O}+, and {MnIIIMnIV(2-OHSalpn)2OH}. We
had independently shown81 that the homolytic bond dis-
sociation energies associated with the complexes were
approximately 90 kcal/mol, suggesting that if we could make
our solutions sufficiently basic, we should be able to
prepare {MnIV2(2-OHSalpn)2OH}+ from {MnIIIMnIV(2-
OHSalpn)2H2O}+ directly by electrochemical methods. Fur-
thermore, we hoped that by using a proton-coupled electron-
transfer process we would be able to access the desired
MnVdO species {MnIVMnVdO(2-OHSalpn)2}+ as well.
The second cyclic voltammogram in Figure 3 illustrates
that the preparation of {MnIV2(2-OHSalpn)2OH}+ from
{MnIIIMnIV(2-OHSalpn)2H2O}+ is readily achieved and we
have isolated a solid that yields the desired physical
properties for this species. The perpendicular-mode EPR
spectra for this complex are shown in Figure 10. To our
knowledge, this was the first reported example of a parallel-
mode EPR spectrum of a MnIV dimer. Like the previously
reported [MnIV(2-OHSalpn)]22+, this complex also exhibits
a perpendicular-mode EPR spectrum because of its weak
antiferromagnetic exchange coupling. XANES spectra also
confirm the assignment of the complex as MnIV because the
edge energy for the resultant solid is 1.5 eV higher than that
of the starting {MnIIIMnIV(2-OHSalpn)2OH}, a difference
consistent with numerous other MnIIIMnIV to MnIV2 oxidations.
Subsequent efforts to prepare the desired {MnIVMnVdO(2-
OHSalpn)2}+ using an electrochemical approach were unsuc-
cessful. We believe this to be true for two reasons. First, the
{MnIV2(2-OHSalpn)2OH}+ is not sufficiently basic to generate
the necessary {MnIV2(2-OHSalpn)2O}2+ precursor under stan-
dard conditions to carry out the metal oxidation. Under more
basic conditions, {MnIV2(2-OHSalpn)2OH}+ converts rapidly
to the tetrameric {[MnIV2(2-OHSalpn)2]2O}2+ whose EPR
spectrum was shown in Figure 6. We next explored alternative
strategies for the preparation of the MnIVMnVdO species.
It is well established that the reaction of certain oxo-transfer
agents such as m-chloroperbenzoic acid (m-CPBA) or hy-
pochlorous acid under basic conditions with MnIII(Salpn) or
MnIII(Salen) causes olefin oxidation to the desired epoxide.82,83
A quick comparison of MnIIISalen and {MnIIIMnIV(2-OH
Salpn)2H2O}+ shown in Scheme 1 illustrates that the MnIII ion
has a similar coordination environment in both structures. We
reasoned that {MnIIIMnIV(2-OHSalpn)2H2O}+ should be capable
of olefin epoxidation, as observed for MnIII(Salen). Furthermore,
though controversial,84–88 it has been suggested that the active
Scheme 1
species of the Salen-based epoxidation reactions is a MnVdO
species. If true, then we should be able to chemically oxidize
{MnIIIMnIV(2-OHSalpn)2H2O}+ by two electrons using an oxo-
transfer agent such as m-CPBA.
We tested whether the first portion of this proposition is
true, that is, that {MnIIIMnIV(2-OHSalpn)2H2O}+ in the
presence of m-CPBA n-morphiline oxide (NMO) is capable
of epoxidizing olefins. {MnIIIMnIV(2-OHSalpn)2H2O}+ was
reacted with m-CPBA in the presence of cyclohexene and
NMO at -78 °C. Cyclohexene oxide was detected by gas
chromatography (GC)-mass spectrometry (MS), whereas no
2-cyclohexen-1-ol or 2-cyclohexen-1-one was detected. Cy-
clohexene oxide is the direct product of the epoxidation via
a concerted two-electron oxo-transfer process. In contrast,
2-cyclohexen-1-ol and 2-cyclohexen-1-one are allylic oxida-
tion products generated through radical processes such as
those shown by Caudle et al. for related systems.13 The
reaction proceeded for 20 turnovers based on a manganese
catalyst. This reactivity could be taken as circumstantial
evidence supporting the transient existence of the desired
MnIVMnVdO. However, a troubling side reaction was
observed, the oxidation of m-chlorobenzoic acid (mCBA)
to m-chlorophenol. We investigated this chemistry with a
known MnVdO complex reported by Collins et al.89 and also
with Mn(Salen) and observed a rapid oxidation of the organic
product, yielding the same m-chlorophenol product. While
an X-ray structure of the MnVHMPAB complex89 is known
and we have shown using XANES spectroscopy that such a
species is formed upon reaction of the MnIII precursor with
m-CPBA, the case of Mn(Salen) is more ambiguous.
Because of the rapid olefin epoxidation, we attempted to
detect the MnIVMnVdO species by adding 1 equiv of m-CPBA
to {MnIIIMnIV(2-OHSalpn)2H2O}+ in the absence of a substrate
at -78 °C. The parallel-mode spectrum that results is identical
with that observed in Figure 6. This is the same product formed
by bulk oxidation in basic conditions, the tetrameric {[Mn(2-
OHSalpn)]4O}2+. One explanation for this tetrameric product
is that a comproportionation reaction between the starting
MnIIIMnIV dimer and the generated MnIVMnVdO species occurs
at near-stoichiometric oxidant concentration.
(81) Caudle, M. T.; Pecoraro, V. L. J. Am. Chem. Soc. 1997, 119, 3415.
(82) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N. J. Am.
Chem. Soc. 1990, 112, 2801.
(83) Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. Soc. 1986,
108, 2309.
(84) Campbell, K. A.; Lashley, M. R.; Wyatt, J. K.; Nantz, M. H.; Britt,
R. D. J. Am. Chem. Soc. 2001, 123, 5710.
(85) Yin, G. C.; McCormick, J. M.; Buchalova, M.; Danby, A. M.; Rodgers,
K.; Day, V. W.; Smith, K.; Perkins, C. M.; Kitko, D.; Carter, J. D.;
Scheper, W. M.; Busch, D. H. Inorg. Chem. 2006, 45, 8052.
(86) Yin, G. C.; Buchalova, M.; Danby, A. M.; Perkins, C. M.; Kitko, D.;
Carter, J. D.; Scheper, W. M.; Busch, D. H. Inorg. Chem. 2006, 45,
3467.
In an effort to surmount this difficulty, we next added a 500-
fold excess of m-CPBA to the reaction solution. Gas evolution
was immediately observed. In experiments that allowed us to
trap and analyze the liberated head gas, we confirmed that CO2
(87) Yin, G.; Danby, A. M.; Kitko, D.; Carter, J. D.; Scheper, W. M.; Busch,
D. H. J. Am. Chem. Soc. 2007, 129, 1512.
(89) (a) Collins, T. J.; Powell, R. D.; Slebodnick, C.; Uffelman, E. S. J. Am.
Chem. Soc. 1990, 112, 899. (b) Collins, T. J.; Gordon-Wylie, S. W.
J. Am. Chem. Soc. 1989, 111, 4511.
(88) Yin, G.; Danby, A. M.; Kitko, D.; Carter, J. D.; Scheper, W. M.; Busch,
D. H. Inorg. Chem. 2007, 46, 2173.
1776 Inorganic Chemistry, Vol. 47, No. 6, 2008

ElusiWe High-Valent Manganese Species
Figure 12. Species distribution plot from the stopped-flow data fitted to a
biexponential model: A f B, 2B f 2C. Stopped-flow conditions:
[{MnIIIMnIV(2-OHSalpn)2H2O}+] ) 0.15 mM; [m-CPB] ) 4 mM, tem-
perature ) 5 °C, collection time ) 1 s. Solid circles correspond to the
concentration of {MnIIIMnIV(2-OHSalpn)2H2O}+, open circles correspond
to the concentration of {MnIV2(2-OHSalpn)2OH}+, and open squares reflect
the concentration of {[Mn(2-OHSalpn)]4O}2+. EPR spectra at designated
time points are added to correlate the EPR behavior of samples as a function
of time. EPR spectral conditions are given in Figure 11.
Figure 11. Parallel-mode EPR spectra of {MnIIIMnIV(2-OHSalpn)2H2O}+
reacted with m-CPBA in acetonitrile at 20 °C. Freeze quench at various
reaction times: (A) 30 ms; (B) 120 ms. (C) Temperature dependence from
4 K (top) to 40 K (bottom). EPR conditions for 4 K: microwave frequency,
9.424 GHz; power, 2 mW; modulation amplitude, 10 G; modulation
frequency, 100 kHz.
that is characteristic of the starting {MnIIIMnIV(2-
OHSalpn)2H2O}+ and a very weak radical signal. Note that one
was being evolved. This observation is identical with that for
the reaction of Mn(Salen) or Mn(HMPAB) with m-CPBA
described above. Mass spectral analysis of the solution dem-
onstrated that m-chlorophenol was also formed, indicating that
the source of CO2 was likely the m-CBA that was oxidized
by the high-valent manganese complex. Under these conditions,
the resultant EPR signal was that of the tetrameric compound,
the growth of which is shown in Figure 11. Such a signal could
be generated by the reaction of a MnIVMnVdO complex with
m-CBA to generate the MnIV dimer and m-CBA radical. The
radical would decompose to CO2 plus the m-chlorophenyl
radical, while the MnIV dimer would dimerize to give the
observed tetramer. We have ruled out the possibility of a stable
phenoxy radical as described by Wieghardt et al.90 because the
characteristic UV–vis bands for such radicals are not observed
in this system.
Once again, all of these results are consistent with the
formation of a MnIVMnVdO complex; however, there has been
no direct spectroscopic detection of the species. Therefore, we
carried out detailed stopped-flow kinetics and rapid freeze-
quench EPR studies on the {MnIIIMnIV(2-OHSalpn)2H2O}+/
m-CPBA system to glean more insight on the early stages of
this reaction. These data are summarized in Figure 12. Analysis
of the stopped-flow data suggested that there were three primary
species, which we will describe as A-C. During the first 100
ms, species A rapidly decays to near zero and species B rapidly
rises to a maximum. The rapid freeze-quench perpendicular-
mode EPR spectrum at 10 ms showed mostly the g ) 2 signal
3
would expect an S ) /2 signal for the MnIVMnVdO species
because d2 metal oxo species are invariably low spin. As the
reaction time increases to 30 ms and then 80 ms, the radical
signal intensifies while the g ) 2 multiline weakens. At the
same time in the low-field region, broad spectral features for
{MnIV2(2-OHSalpn)2OH}+ develop. We also observe a signal
indicative of the tetramer developing from 50 to 120 ms. We
conclude that the three EPR-observable manganese species are
the starting MnIIIMnIV dimer (species A), which converts into
the MnIV dimer (species B) and the radical (mCBA radical),
with species C being the ever-present tetramer. The simplest
explanation from these experiments is that a MnIVMnVdO
species could be formed in low concentration and with an
extremely short half-life (<10 ms); however, we favor the
proposal of Busch et al.88 that the complexes both carry out
their epoxidation chemistry and exist in the absence of a
substrate as MnIV species, possibly with coordinated peroxides,
not as MnVdO.
Summary
The experiments above demonstrate the extreme difficulty
in clearly assessing chemical mechanism in many manganese
redox reactions. All of the evidence is consistent with the
(90) Marlin, D. S.; Bill, E.; Weyhermuller, T.; Bothe, E.; Wieghardt, K.
J. Am. Chem. Soc. 2005, 127, 6095.
Inorganic Chemistry, Vol. 47, No. 6, 2008 1777

Pecoraro and Hsieh
formation of a MnIVMnVdO species; however, the definitive
spectral evidence remains elusive. Such has been the behavior
of PS II biophysics. One obtains evidence that is consistent
with the ghost of the species desired, but always that
intermediate (or maybe best transition state) is far too fleeting
to actually capture. It is ironic that one must come to the
same conclusion with the Salpn system that one draws from
the photosystem, that either a highly reactive and undetect-
able MnVdO exists or some highly reactive and undetectable
MnIV radical is the active agent. Future experiments will try
to push back the time barriers in order to answer these
important questions. Nonetheless, the progress in understand-
ing the chemical aspects of water oxidation has progressed
remarkably, on both the chemical and biological fronts, in
the past 20 years, and one can only hope that similar major
advances await us during the next 2 decades.
Quadrupole GC-MS spectrometer with an HP 5890 gas chromato-
graph equipped with a DB-5 capillary column installed in a splitless
injector was used for these studies. The mass spectrometer is capable
of unit mass resolution.
Stopped-Flow Experiments. The general sample preparations
for the kinetic experiments were as follows: 12.8 mg of {MnIII
MnIV(2-OHSalpn)2}PF6 was dissolved in 3 mL of 150 mM
tetrabutylammonium perchlorate (TBAClO4)/acetonitrile to make
a 5 mM manganese stock solution. This stock solution was diluted
with a 150 mM TBAClO4 acetonitrile solution to make 0.1, 0.15,
0.2, 0.25, and 0.3 mM solutions. A total of 27 mg of m-CPBA
dissolved in 5 mL of a 150 mM TBAClO4/acetonitrile solution was
then added to make a 20 mM stock solution. The stock solution
was diluted with a 150 mM TBAClO4/acetonitrile solution to make
0.1,0.5, 1.0, 2.0, and 4.0 mM solutions. These solutions were kept
at 0 °C prior to reaction. The oxidation of {MnIIIMnIV(2-
OHSalpn)2H2O}+ by m-CPBA was monitored using an Online
Instrument System rapid scanning monochromator (OLIS-RSM)
stopped-flow spectrophotometer equipped with a NESLab RTE 111
refrigerated bath/circulator. The reaction solutions were placed in
two 2-mL syringes separately and submerged in a water bath
maintained at 5 °C. The reactions were scanned at 1000 scans/s
for 1 s. The reactions were monitored at 500 nm. Data were
analyzed using the program SPECFIT.91
EPR Spectroscopy. X-band continuous-wave perpendicular-
mode EPR spectra were collected either on a Bruker EMX200E
spectrometer at the University of Michigan or a Bruker ESP300E
spectrometer at Michigan State University. An Oxford Instruments
ESR-900 continuous-flow crysostat was used to maintain a tem-
perature between 4 and 60 K for both instruments. X-band
continuous-wave parallel-mode detection experiments were col-
lected at Michigan State University using the same experimental
system as that described above. Experimental details for appropriate
spectra are provided in the figure captions.
Rapid-Mix Freeze-Quench Spectroscopy. The freeze-quench
experiments were performed at Michigan State University using
an Update Instruments System 1000 Chemical/Freeze Quench
apparatus (model 715 ram controller and model 1019 syringe ram).
The manganese/m-CPBA solutions were placed in two equal-
volume syringes and kept at room temperature. The syringe contents
were combined in a Wiskind grid mixer (model 1155) at a ram
velocity of 1.25 cm/s. Different reaction times were achieved by
varying the length of the tubing (aging hose) connecting the mixer
to the spray nozzle. The reaction mixture was ultimately sprayed
from a 0.008-in.-diameter spray nozzle into a 4.5-in.-long fused
quartz EPR tube equipped with a funnel filled with cold isopentane
(HPLC grade, Sigma), which quenches the reaction in approxi-
mately 5 ms. The EPR tube and funnel were filled with isopentane
and equilibrated for at least 5 min in a 8 L isopentane bath
maintained at -140 °C with a LakeShore model 340 liquid-nitrogen
bath. The frozen solids of an oxidized manganese solution were
packed into the bottom of the EPR tube by using a precooled
packing rod until a densely packed sample with a height of at least
1 cm was obtained. Time intervals for the quenching reaction of
10, 30, 80, 120, and 200 ms were employed.
Experimental Section
Preparation of Complexes. {MnIIIMnIV(2-OHSalpn)2H2O}+ and
{MnIIIMnIV(3,5-di-t-Bu-2-OHSalpn)2H2O}+ were prepared via lit-
erature procedures.13 {MnIV2(3,5-di-t-Bu-2-OHSalpn)2OH}+ was
prepared by the reaction of 0.53 g (0.4 mmol) [as a PF6 salt]
dissolved in 15 mL of dichloromethane/ether (1:1) with stirring.
The solution was cooled to -40 °C in an acetonitrile/dry ice bath.
Then CeIV(t-Bu4NH)2(NO3)6 (0.48 g, 0.482 mmol) was dissolved
in 5 mL of cold dichloromethane, which was then added to the
manganese solution at -40 °C. The solution was stirred for 10
min at -40 °C, and then it was suction-filtered at low temperature.
The solvent was evaporated and dried at -25 °C under vacuum.
The dark-brown solid was then redissolved in 15 mL of cold ether/
pentane (2:1) at -50 °C. This solution was filtered and dried under
vacuum, and the brown solid was collected and stored below -40
°C. {[Mn(2-OHSalpn)]4O}2+ was prepared either by bulk oxidation
of {MnIIIMnIV(2-OHSalpn)2H2O}+ in acetonitrile in the presence
of a base or by the reaction of {MnIIIMnIV(2-OHSalpn)2H2O}+ with
m-CPBA in acetonitrile at -15 °C.
Electrochemistry. Cyclic voltammograms for the complexes
were measured under identical conditions using a platinum disk
working electrode, an aqueous Ag/AgCl reference electrode, and a
platinum wire counter electrode.
GC-MS. Cyclohexene, cyclohexene oxide, and NMO were
purchased from Aldrich and used without further purification.
m-CPBA was also purchased from Aldrich as a 50–60% pure
sample and then was purified by dissolving the solid in dry
methylene chloride, extracting with a saturated sodium bicarbonate
solution several times, and then drying under vacuum for 1 day.
The peroxy acid versus reduced acid ratio was determined by iodine
titration to be 85%. The general methods for the epoxidation of
organic substrates were that 1 equiv of manganese dimer was
reacted with 100 equiv of cyclohexene and 100 equiv of NMO in
dichloromethane. A total of 50 equiv of m-CPBA dissolved in
dichloromethane was added to this solution at -78 °C. After 20
min, a portion of the mixture was collected and analyzed by
GC-MS. A fixed amount of n-decane was added to each solution
as an internal standard. The same reaction procedures were carried
out at room temperature. Control reactions were carried out in the
absence of a manganese complex. Authentic epoxide product and
allylic oxidation products were used as standards. A Finnigan 4500
Acknowledgment. V.L.P. acknowledges the NIH (Grant
GM39406) for financial support of this work. We thank the
EPR facility, especially Bryan Schmidt, at Michigan State
University for access to and assistance with EPR and freeze-
quench apparatuses.
(91) Stultz, L. K.; Binstead, R. A.; Reynolds, M. S.; Meyer, T. J. J. Am.
Chem. Soc. 1995, 117, 2520.
IC7017488
1778 Inorganic Chemistry, Vol. 47, No. 6, 2008