JBIC Journal of Biological Inorganic Chemistry

, Volume 22, Issue 2, pp 237–251

A personal perspective on the discovery of dioxygen adducts of copper and iron by Nobumasa Kitajima

Commentary

DOI: 10.1007/s00775-016-1432-1

Cite this article as:
Fujisawa, K. J Biol Inorg Chem (2017) 22: 237. doi:10.1007/s00775-016-1432-1
Part of the following topical collections:
  1. 60 Years of Oxygen Activation

Abstract

Abstract

Transition metal–dioxygen complexes have fascinated bioinorganic and inorganic chemists for over half a century. The late Nobumasa Kitajima was one of the very successful researchers in this field. Despite his short career (40 years old), he made many important contributions. This Commentary highlights his important accomplishments and how they have impacted subsequent work in this area.

Graphical Abstract

Keywords

Copper Dioxygen X-ray structure Hydrotris(pyrazolyl)borate Model complex 

Abbreviations

bppa

Bis(6-pivalamide-2-pyridylmethyl)(2-pyridylmethyl)amine

3,5-iPr2pzH

3,5-Diisopropyl-1-pyrazole

N-Et-HPTB

N, N, N’, N’-tetrakis(1-ethylbenzimidazolyl-2-methyl)-1,3-diamino-2-propanol

L0

Hydrotris(3,5-dimethyl-1-pyrazolyl)borate anion

L1

Hydrotris(3,5-diisopropyl-1-pyrazolyl)borate)borate anion

L3

Hydrotris(3-tertialy butyl-5-isoporpyl-1-pyrazolyl)borate anion

L5

Hydrotris(3,5-diphenyl-1-pyrazolyl)borate anion

L10

Hydrotris(3-adamantyl-5-isoporpyl-1-pyrazolyl)borate anion

mCPBA

m-chloroperbenzoic acid

oxyhemocyanin

Oxygenated hemocyanin

PhIO

Iodosylbenzene

Ph-bimp

2.6-Bis[bis{2-(1-methyl-4,5-diphenylimidazolyl)methyl}aminomethyl]-4-methylphenolate

TMG3tren

1,1,1-Tris{2-[N2-(1,1,3,3-tetramethylguanidino)]ethyl}amine

Tp

Hydrotris(pyrazolyl)borate

TPA

Tris{(2-pyridyl)methyl}amine

XYL–O–

α,α’-Bis{N, N-bis(2-pyridylethyl)amino}-m-xylene-2-olate

Introduction

At the Professor Edward I. Solomon Award symposium dinner at Roy’s restaurant in San Diego on March 14, 2016, Professor Lawrence Que asked me, “I am planning about Special Oxygen Issue of Journal of Biological Inorganic Chemistry for the late Professor Osamu Hayaishi.” My answer was “Yes, I will contribute to this special issue”, since I had a chance to meet Professor Hayaishi with Professor Que at the 14th Takeda Science Foundation Symposium on Bioscience ‘The 50th Anniversary of Oxygenases—Advances and Reflections’ held in Kyoto Hotel Okura on April 10–11, 2006 [1]. Moreover, in my Inorganic Class for undergraduate students, I taught the important methodology of using 18O2 and H218O as tracers independently reported in 1955 by Professors Howard S. Mason and Osamu Hayaishi. Mason and co-workers demonstrated the incorporation of an oxygen atom from molecular oxygen into 3,4-dimethylphenol to form 4,5-dimethylcatechol catalyzed by mushroom phenolase (Scheme 1 top) [2], while Hayaishi and his co-workers found the two oxygen atoms of molecular oxygen to be incorporated into catechol upon its oxidative cleavage to cis,cis-muconic acid by pseudomonad pyrocatechase (Scheme 1 bottom) [3].
Scheme 1

Enzymatic oxygen incorporation reactions using 18O2 as a tracer discovered by Professor Mason (top) [2] and by Professor Hayaishi (bottom) [3]

I was a graduate student of the late Nobumasa Kitajima, who in his short 10-year career made many important contributions to our understanding of oxygen activation by metalloproteins. I joined his group as a Master course student in April 1987 in the Research Laboratory of Resources Utilization, Tokyo Institute of Technology and had the opportunity to work together with Mr. Takayuki Koda and Ms. Chisato Fujimoto on biomimetic copper chemistry. Our focus was on the use of hydrotris(pyrazolyl)borates (Tp, Scheme 2) as ligands for making copper complexes. Mr. Koda worked on the known dimethyl-substituted Tp ligand (denoted as L0), while Ms. Fujimoto investigated the novel diphenyl-substituted Tp ligand (denoted as L5). I worked on the novel diisopropyl-substituted Tp ligand (denoted as L1), which had a few positive features: no tendency to form the dead-end [Tp2M] complex, better solubility in non-coordinating solvents compared with L5, and formation of neutral [MII(Tp)X] complexes that made it easy to separate complexes from salts. The Tp ligands were invented by the late Professor Swiatoslaw Trofimenko and have found many applications in the fields of inorganic and coordination chemistry [4, 5]. His first publication on boron–pyrazole chemistry was in 1966 [6], so 2016 marks the 50th anniversary of the discovery of scorpionate ligands [7]. Professor Trofimenko further developed the utility of his ligand family by the introduction of 3-tert-butyl and 3-phenyl substituents in 1986 [8, 9], which we found to be useful as well for our chemistry [10]. In this minireview, I summarize Professor Kitajima’s achievements in synthesizing copper, iron, and manganese complexes that react with dioxygen and the novel O2-derived species that were identified from this effort. Moreover, our 18O-isotope experiments were inspired by the original experiments of Professors Mason and Hayaishi that gave birth to this field.
Scheme 2

Hydrotris(pyrazolyl)borate (Tp) ligands discussed in this article

Copper–dioxygen complexes

The nature of the copper–dioxygen interaction was one of the challenging themes in the copper bioinorganic chemistry field in the 1980s. An important goal was to determine the O2 binding mode of oxyhemocyanin, the oxygenated form of the copper-containing O2 carrier found in arthropods and mollusks [11, 12, 13, 14, 15, 16, 17, 18], the striking physicochemical properties of which attracted the attention of inorganic chemists. It has two intense absorption bands at ca. 350 and 580 nm that give rise to its bright blue color. It is also diamagnetic at room temperature, a property subsequently shown to arise from strong antiferromagnetic coupling between two copper(II) ions (−2 J > 600 cm−1).

At that time, only the structure of deoxyhemocyanin from Panulirus interruptus (lobster) had been solved [19, 20], which revealed two copper(I) centers, each with three histidine ligands. Therefore, our choice of the Tp ligand system was a reasonable starting point for making hemocyanin model complexes. The only copper–dioxygen complexes in the literature at that point were [CuII(L0)(O2)], reported by Dr. Jeffery S. Thompson [21], and \( \left[ {{\text{Cu}}^{\text{II}}_{ 2} \left( {{\text{XYL}}{-}{\text{O}}{-}} \right)\left( {{\text{O}}_{ 2} } \right)} \right]\left( {{\text{PF}}_{ 6} } \right) \), reported by Professor Kenneth D. Karlin [22, 23].

Professor Kitajima wanted to develop a new synthetic route to make copper–dioxygen complex different from that used by Thompson [21], which used the copper(I)–ethylene complex ([CuI(L0)(C2H4)]) as the starting material (Scheme 3). He obtained a moderately stable reddish purple product with a relative intense absorption band at 524 nm (600 M−1cm−1). This oxygenated product was assigned as the copper(II)–superoxido complex based on a relative weak IR band at 1015 cm−1, which was shifted to 1060 cm−1 upon 18O substitution. However, no dioxygen-uptake experiments were carried out to determine the copper/dioxygen stoichiometry, which could be used to distinguish between a Cu:O2 (superoxide formation) and a Cu2:O2 (μ-peroxide formation) species.
Scheme 3

Oxygenation reaction of [CuI(L0)(C2H4)] [21]

Using a different synthetic strategy, Mr. Koda used the (μ-oxido) dicopper(II) complex as a precursor, which was obtained from the reaction of iodosylbenzene (PhIO) with [CuI(L0)(PPh3)] (Scheme 4) [24, 25, 26, 27, 28]. The [{CuII(L0)}2(μ-O)] complex was readily converted to a bis(μ-hydroxido) dicopper(II) species [{CuII(L0)}(μ-OH)]2, whose structure was determined [27]. The subsequent reaction with hydrogen peroxide afforded a complex that was diamagnetic at −40 °C based on NMR spectroscopy [25, 27] and had two characteristic absorption bands at 338 (20800 M−1 cm−1) and 530 nm (840 M−1 cm−1) (Fig. 1 top). Its resonance Raman spectrum (λex = 514.5 nm; −40 °C) revealed a ν(O–O) feature at 731 cm−1 from H216O2, which shifted to 691 cm−1 upon 18O-labeling [27]. A 1:2:1 peak pattern was observed with mixed-labeled H2O2, indicating that peroxide moiety was symmetrically coordinated between two copper(II) ions (Fig. 1 bottom). These vibrational data resembled those reported for oxyhemocyanin, leading Professor Kitajima to entitle his 1988 manuscript “An accurate synthetic model of oxyhemocyanin” [25].
Scheme 4

Generation of μ-peroxido dicopper(II) complex with L0 and its reactivity [27]

Fig. 1

Top UV–vis spectrum in CHCl3 at −14 °C (left) of the reaction product from [{CuII(L0)}2(μ-O)] and H2O2. Bottom (A) corresponding resonance Raman spectra in CHCl3 at −40 °C for the complex obtained from a 1:2:1 mixture of H216O2/H216O18O/H218O2 and (B) simulated spectrum assuming a band intensity ratio of 1:2:1 with identical widths of 18 cm−1. Adapted with permission from [27]. Copyright 1991 American Chemical Society

During the period of my Master course, Professor Kitajima had his first treatment for cancer at Keio University Hospital from January to March in 1988, so Mr. Koda and I wanted to get good results to cheer him up. At that time, I had already made the corresponding CuI(L1) complex, and its reaction with O2 led to the formation of a purple-colored solid (Scheme 5). Resonance Raman experiments in collaboration with Professor Teizo Kitagawa (Institute for Molecular Science) (Fig. 2 top) showed a spectrum very similar to that from the species from the reaction of CuII(L1) with H2O2 [28, 29, 30, 31], so it appeared that the same species could be formed by the reaction of CuI(L1) with O2 or the reaction of CuII(L1) with H2O2.
Scheme 5

Generation of μ-peroxido dicopper(II) complex with L1 and its reactivity [30]

Fig. 2

Resonance Raman spectra in acetone at −40 °C of dioxygen reaction product from [CuI(L1)] complex (top) and molecular structure of [{CuII(L1)}2(μ-O2)] (bottom) (30% probability thermal ellipsoids). Adapted with permission from [30]. Copyright 1992 American Chemical Society

In parallel experiments with the L5 ligand, Ms. Fujimoto synthesized and characterized the [CuI(L5){OC(CH3)2)] complex and obtained a reddish purple-colored complex from its reaction with O2 in CH2Cl2 at −78 °C (Scheme 6) [30]. This species exhibited almost the same properties with corresponding complexes with L0 and L1. Although this complex could be isolated as a very stable solid, it was not soluble in any solvents after completing its formation.
Scheme 6

Generation of μ-peroxido dicopper(II) complex with L5 [30]

During this effort, Professor Karlin reported the first X-ray structure of a copper(II)–dioxygen complex showing a trans-μ-1,2-peroxido dicopper(II) core, [{CuII(TPA)}2(O2)](PF6)2 [32]. This complex was EPR silent at 77 K and exhibited visible absorption bands at 525 nm (11500 M−1 cm−1) and 590 nm (sh, 7600 M−1 cm−1) and a ν(O–O) Raman feature at 832 cm−1 that downshifted to 788 cm−1 upon 18O2 substitution [32, 33, 34]. The O–O bond length was found to be 1.432(6) Å, consistent with a peroxide ligand. This result took us by surprise, but Professor Kitajima encouraged us by saying, “It was all right; the UV–Vis spectrum was quite different from that of oxyhemocyainin”. This result made Mr. Koda and me work harder to obtain the structures of our own complexes, but we were unable to get crystals. Moreover, I had to focus on completing my Master course thesis.

By that time, I had obtained the bis(μ-hydroxido) dicopper(II) complex [{CuII(L1)}(μ-OH)]2 [35], which turned out to be a very useful precursor for a number of complexes such as acylperoxido copper(II) complex [CuII(L1)(mCPBA)] [35], alkylperoxido copper(II) complexes [CuII(L1)(OOR)] (R = cumyl or tert-butyl) [36, 37], blue copper protein model complexes [CuII(L1)(SR)] (R = C6F5 or Ph3C) [38, 39, 40, 41, 42, 43], and (μ-carbonato) dicopper complex [{Cu(L1)}2(μ-CO3)] [44] (Scheme 7).
Scheme 7

Schematic drawing of the reactions of bis(μ-hydroxido) dicopper(II) complex [35, 36, 37, 38, 39, 40, 41, 42, 43, 44]

After my Masters research presentation in February of 1989, I focused on the recrystallization of the μ-peroxido dicopper(II) complex [{CuII(L1)}2(μ-O2)], which could be obtained by treatment of [{CuII(L1)}(μ-OH)]2 with H2O2 in CH2Cl2 at −50 °C. This complex exhibited very similar physicochemical properties as the μ-peroxido dicopper(II) complexes ligated by L0 and L5, in being diamagnetic at −10 °C by NMR, having two characteristic absorption bands of 349 (21000 M−1 cm−1) and 551 nm (790 M−1 cm−1), and exhibiting a resonance-enhanced Raman ν(O–O) feature at 741 cm−1, which shifted to 719 cm−1 with 16O18O and 698 cm−1 with 18O2 (Fig. 2 top) [28, 29, 30]. At last after many trials, we were able to acquire X-ray diffraction data for the μ-peroxido complex at the end of March 1989. Professor Kitajima wanted to know the structure as soon as possible. However, solving this structure was not so easy, and then he went home before the initial structural determination. When I saw him the next morning, he told me “I could not sleep last night, as I got an idea for the initial solution at home, so I came here at 3 am.”, and showed me the preliminary result of a μ-η2:η2-peroxido structure between two copper(II) ions [{CuII(L1)}2(μ-η2:η2-O2)] (Fig. 2 bottom). We were very happy with this result, which could rationalize its diamagnetism and resonance Raman features.

Unfortunately, my job in a chemical company began in a few days, so Professor Kitajima had to complete the calculations by himself. Moreover, he then presented this exciting structure in the annual meeting of Chemical Society Japan in early April 1989, where he met Professor Stephen J. Lippard. Professor Lippard then contacted Professor Karlin to ask him to allow Professor Kitajima to present his results in the next ICBIC-4 meeting in Cambridge, MA in July 1989. Professor Kitajima submitted these results as a JACS communication before this ICBIC. During the revision process, Professor Lippard in his role as JACS Associate Editor suggested a different choice of space group from P\( \bar{1} \) to C2/c, which was applied for the revised calculations [29]. This change in the space group was also required for solving the structure of [{CuII(L1)}2(μ-η2:η2-S2)] [45, 46].

Professor Kitajima met Professor Solomon after his ICBIC-4 presentation and started our long collaboration. Professor Solomon suggested that the greater σ donation in the side-on structure would strengthen peroxide bonding compared to the end-on isomer and lead to the greater stabilization of the π*σ levels as shown in Fig. 3 [31, 47]. Moreover, the HOMO level would be additionally stabilized in the side-on structure compared to the end-on one due to the π acceptor interaction of the peroxide σ* orbital. This simple molecular orbital picture explained quite well the significantly greater thermal stability of the side-on bridged complex, which was stable below about −20 °C, compared to the end-on bridged dimer complexes, which were unstable above −70 °C.
Fig. 3

Energy level diagrams of the orbitals involved in copper(II) peroxide bonding in a end-on bridged and b side-on bridged models. Adapted with permission from [31]. Copyright 1992 American Chemical Society

After 1 year of company experience, I returned to Professor Kitajima’s laboratory to pursue my doctoral degree and continue my exploration of Cu(L1) chemistry. Having obtained the crystal structure of a (μ-η2:η2-peroxido) dicopper(II) complex, we set our sights on isolating a superoxido copper(II) complex, a species common to all hypothesized reaction mechanisms proposed for non-coupled binuclear copper enzymes such as dopamine β-hydroxylase, peptidylglycine α-hydroxylating monooxygenase, and insect tyramine β-monooxygenase [18]. Despite considerable efforts, no X-ray structure of such a complex had been reported at that time. The reaction of [CuII(L1)(Cl)] with KO2 in acetone/CH2Cl2 yielded a purple-colored powder with an absorption maximum of 555 nm (800 M−1 cm−1). Unpublished resonance Raman results revealed a ν(O–O) feature at 1090 cm−1 that shifted to ~1030 cm−1, clearly indicating this purple-colored complex to be different from the structural characterized μ-peroxido complex. This observation was made in a footnote in ref 26. However, after our careful experiments, it found that this purple-colored complex was just μ-peroxido dicopper(II) complex. In this reaction, KO2 acts as reducing agent as well as O2 source from a small amount of H2O in the solvents (Scheme 5).

I then switched to the bulkier L3 ligand with a tert-butyl substituent to see whether I could get better results (Scheme 2). Treatment of [CuII(L3)Cl] with KO2 in DMF/CH2Cl2 afforded the colorless [CuI(L3)(dmf)] adduct (Scheme 8). This copper(I) complex did not react with O2 in DMF containing solvent unlike the corresponding complex of the L1 ligand, which afforded the μ-peroxido dicopper(II) complex in the reaction with KO2 in CH2Cl2/DMF (Scheme 5). However, [CuI(L3)(dmf)] readily reacted with O2 at lower temperature in CH2Cl2 (Scheme 8) to yield a reddish brown-colored species with characteristic absorption bands at 352 (2300 M−1 cm−1), 510 nm (sh, 230 M−1 cm−1), and 660 nm (90 M−1 cm−1). This new complex showed 1:1 Cu:O2 stoichiometry in dioxygen-uptake experiments, diamagnetism at −40 °C by NMR, and a ν(O–O) feature at 1112 cm−1 that shifted to 1160 cm−1 with 18O2 substitution from IR measurements [48, 49]. Furthermore, this complex exhibited reversible O2 binding behavior (Fig. 4 top). The crystal structure of this adduct was solved to show a Cu(η2-O2) moiety with an O–O bond length of 1.22(3) Å (Fig. 4 bottom). The large uncertainty in the O–O bond distance was due to the poor crystal quality and the collection of the data at room temperature that gave rise to very high isotropic thermal parameters.
Scheme 8

Generation of superoxido copper(II) complex with L3 [48]

Fig. 4

UV–Vis spectral change of reversible dioxygen binding behavior of [CuI(L3)(dmf)] at room temperature. A and C under 1 atom O2 atmosphere and C under an Ar atmosphere (top). Molecular structure of [CuII(L3)(O2)] (50% probability thermal ellipsoids). Hydrogen atoms are omitted for clarity (bottom). Adapted with permission from [48]. Copyright 1994 American Chemical Society

Interestingly, the isolated [CuII(L3)(O2)] complex readily lost its O2 ligand by a simple change in atmosphere to argon, unlike the isolated [{CuII(L1)}2(μ-η2:η2-O2)] complex (Fig. 5, 2). However, in the latter case, loss of O2 was observed in the presence of MeCN, resulting in the formation of [CuI(L1)(MeCN)] (Fig. 5, 3). This reduced complex then easily reacted with dioxygen to yield the μ-peroxido dicopper(II) complex when redissolved in CH2Cl2 (Fig. 5, 4) [50], demonstrating the importance of the coordinating solvent in this binding equilibrium.
Fig. 5

UV–Vis spectral change of the reversible dioxygen binding behavior of [{CuII(L1)}2(μ-O2)] in the absence and presence of coordinating solvents such as MeCN

After my move to Tsukuba in 1998, I resumed the collaboration with Professor Solomon on Cu–O2 adducts [49]. We repeated the resonance Raman experiments on [CuII(L3)(O2)] and observed results that suggested contamination of the [CuII(L3)(O2)] species with some fraction of the dimerized [{CuII(L3)}2(μ-O2)] (~10% from the UV–vis spectrum). To minimize this dimerization reaction even in solution, an even bulkier ligand was required and I synthesized L10 with adamantyl groups in place of the tert-butyl groups in L3 (Scheme 2). The [CuII(L10)(O2)] complex exhibited diamagnetism at −78 °C, three characteristic bands at 452 nm (300 M−1 cm−1), 700 nm (sh, 40 M−1 cm−1), and 975 nm (20 M−1 cm−1) in CH2Cl2 at −78 °C, and a ν(O–O) feature at 1043 cm−1 that downshifted to 984 cm−1 upon 18O2 substitution (Fig. 6). In contrast to this side-on superoxido binding mode, a copper(II) complex with an end-on bound superoxide, [CuII(O2)(TMG3tren)], was more recently structurally characterized by Professor Siegfried Schindler [51] and exhibited somewhat different properties (r(O–O) = 1.280(3) Å; ν(O–O) = 1117 cm−118O = −58 cm−1)).
Fig. 6

UV–Vis absorption spectrum of [CuII(L10)(O2)] in CH2Cl2 at −70 °C (solid line) with Gaussian resolved individual transitions (dashed lines). Overlaid are the resonance Raman profiles of the 1043 cm−1 vibration mode (ν(O–O), filled circle) (top). Resonance Raman spectra of [CuII(L10)(O2)] excited at 482.5 nm in CH2Cl2 (bottom). Adapted with permission from [49]. Copyright 2003 American Chemical Society

On the other hand, using [CuII(L3)(OH)] as precursor [52], we were able to obtain the corresponding hydroperoxido copper(II) complex [CuII(L3)(OOH)], which could serve as a model for the active oxygen species in non-coupled binuclear copper enzymes [18]. This complex exhibited characteristic visible absorption features at 598 nm (1530 M−1 cm−1) and 830 nm (360 M−1 cm−1) [37]. Resonance Raman spectra measured in tert-butylbenzene solvent revealed two vibrations at 843 and 624 cm−1, which shifted to 799 and 607 cm−1 upon 18O2 substitution, respectively (Fig. 7) [37]. These are features similar to those reported by Professor Hideki Masuda (Nagoya Institute of Technology) for [CuII(bppa)(OOH)], the crystal structure of which provided metrical parameters for a CuII-η1-OOH unit [53].
Fig. 7

Resonance Raman spectra of [CuII(L3)(OOH)] derivatives in tert-butylbenzene glass at 77 K. Sharp derivatives and negative features are artifacts due to subtraction of the tert-butylbenzene solvent bands. Adapted with permission from [37]. Copyright 2000 American Chemical Society

In summary, we were able to characterize structurally many copper–dioxygen complexes such as [{CuII(L1)}2(μ-O2)], [CuII(L1)(OOC(CH3)2Ph)], and [CuII(L3)(O2)] and to characterize reactive copper–dioxygen complexes such as [CuII(L1)(mCPBA)] and [CuII(L3)(OOH)]. Moreover, we are now underway to characterize more reactive species such as products of the reaction with protons and other activation methods. In the next section, other metal–dioxygen complexes discovered in Professor Kitajima’s laboratory are discussed.

Iron and manganese–dioxygen complexes

Professor Kitajima’s successful strategy for modeling the oxyhemocyanin active site was also effective in obtaining similar complexes for iron and manganese. The [FeII(L1)(κ2-O2CPh)] precursor was synthesized and observed to react with dioxygen reversibly in non-coordinating solvents to yield a moderately stable green-colored dioxygen adduct (Scheme 9) [54, 55]. This species exhibited resonance Raman bands at 876 cm−1 and 418 cm−1, which, respectively, shifted to 827 and 409 cm−1 upon 18O2 substitution, leading to their assignment as ν(O–O) and ν(Fe–O) modes of a metal–peroxo complex. Despite our persistent efforts, we were unable to obtain diffraction quality crystals. However, Professor Lippard was subsequently successful in obtaining the structure of the corresponding O2 adduct with PhCH2COO as the carboxylate ligand. This structure revealed a 1,2-peroxido-bridged diferric center that was supported by two bidentate carboxylate bridges, [{FeIII(L1)}2(μ-OOCCH2Ph)2(μ-O2)] (Fig. 8) [56], with an O–O bond distance of 1.408(9) Å and an Fe–O (peroxide) bond length of 1.885(12) Å. In the same year 1996, two other structures of (μ-1,2-peroxido) diferric complexes were reported, namely \( [{\text{Fe}}^{\text{III}}_{2}({{\text{Ph}} - {\text{bimp}}})({{\text{O}}_{2}})({\text{PhCOO}})]{({{\text{BF}}_{4}})_{2}} \) by Professor Masatatsu Suzuki (Kanazawa University) [57] and \( {\left[ {{\text{Fe}}^{\text{III}}_{ 2} \left( {{\text{N}} - {\text{Et}} - {\text{HPTB}}} \right)\left( {{\text{O}}_{ 2} } \right)\left( {{\text{OPPh}}_{ 3} } \right)_{ 2} } \right]\left( {{\text{BF}}_{ 4} } \right)_{ 3} } \) by Professor Que [58]. A detailed spectroscopic and theoretical analysis of [{FeIII(L1)}2(μ-OOCR)2(μ-O2)] complexes was published by Professor Solomon in 1998 [59].
Scheme 9

Generation of iron(III)–dioxygen complex [54]

Fig. 8

Molecular structure of [{FeIII(L1)}2(μ-OOCCH2Ph)}2(μ-O2)]. Adapted with permission from [56]. Copyright 1996 American Chemical Society

Professor Kitajima also expanded his research efforts into manganese–dioxygen chemistry. Starting with the bis(μ-hydroxido) dimanganese(II) complex [{MnII(L1)}(μ-OH)]2, Mr. Masahisa Osawa and co-workers obtained three complexes of bioinorganic interest (Scheme 10) [60, 61, 62]. When this complex was reacted with H2O2 in the presence of 2 eq of 3,5-iPr2pzH at room temperature, dark brown-colored crystals could be obtained after recrystallization at −20 °C from acetonitrile. This structure revealed a side-on bound peroxido ligand with an O–O bond distance of 1.428(7) Å (Fig. 9).
Scheme 10

Dioxygen-derived dimanganese (III) complexes [60, 61, 62]

Fig. 9

Molecular structure of [MnIII(L1)(O2)(3,5-iPr2pzH)]. Adapted with permission from [62]. Copyright 1994 American Chemical Society

(μ-Oxo)dimanganese(III) complexes were obtained when the dimanganese(II) precursor was reacted with either dioxygen or potassium permanganate, affording dark red and deep blue crystals. X-ray crystallography of the dark red crystals showed a bis(μ-oxo)dimanganese(III) complex that resulted from a 2-e-oxidation of the precursor (Fig. 10 top) [60]. This complex was formed nearly quantitatively from oxidation by permanganate and was also the major product from the dioxygen reaction. The minor product obtained under the latter conditions was a deep blue-colored complex with a (μ-oxo) dimanganese(III) core. Interestingly, one of the isopropyl groups on each of the supporting L1 ligands had been oxidized to an alkoxide that then coordinated to the manganese ions, i.e., [{MnIII(L1O)}2(μ-O)] (Fig. 10 bottom) [61]. FD-MS spectroscopy showed that both alkoxido oxygens derived from the same dioxygen molecule. The Mn···Mn distances changed from 3.314(1) Å in [{MnII(L1)}(μ-OH)]2 to 2.696(2) Å in [{MnIII(L1)}(μ-O)]2 and 3.530(4) Å in [{MnIII(L1O)}2(μ-O)] [60, 61].
Fig. 10

Molecular structures of [{MnIII(L1)}(μ-O)]2 [60] (top) and [{MnIII(L1O)}2(μ-O)] (bottom) [61]. Adapted with permission from [60, 61]. Copyright 1991 American Chemical Society

Concluding remarks

On January 6, 1995, Professor Kitajima attended the doctoral defense of one of his students, Mr. Shiro Hikichi but had a terrible cold. The next day he went to the Institute for Molecular Science in Okazaki to attend the JSPS meeting. The following morning, he was found dead in his hotel room, an unexpected end to a short but highly productive scientific career. Since that time, I have conducted research at Tokyo Institute of Technology (until 1998), the University of Tsukuba (1998–2010), and Ibaraki University on April 1, 2010. I have many fruitful collaborations with Professor Ed Solomon [40, 43, 46, 49, 63, 64, 65, 66, 67], Professor Larry Que [68, 69, 70] and Professor Ken Karlin [46, 65, 67], as well as with Ed Solomon’s former postdocs, Professor Nicolai Lehnert (University of Michigan) [50, 71, 72, 73, 74, 75, 76, 77] and Professor Robert K. Szilagyi (Montana State University) [77, 78, 79, 80].

Acknowledgements

I would like to thank the authors whose names are found in the cited references for their research efforts and contributions. I am grateful for support from the Japan Society for the Promotion of Science (JSPS) (K.F. 25109505) and the Iwatani Naoji Foundation’s Research Grant. I also thank Professor Larry Que for critical reading of this manuscript. Finally, I would like to express my thanks to our collaborators.

Copyright information

© SBIC 2017

Authors and Affiliations

  1. 1.Department of ChemistryIbaraki UniversityMitoJapan