1
Anatomy of a Redox-Active Ligand

Marine Desage-El Murr

Université de Strasbourg, Institut de Chimie (UMR 7177), 1 rue Blaise Pascal, 67081 Strasbourg cedex, France

1.1 Introduction

Over the past decades, redox-active ligands have transitioned from chemical curiosities to staples in current coordination systems. Their versatile nature allows them to encompass molecular strategies ranging from the design of single-molecule magnets to enlarging chemical reactivity through productive electronic communication between the ligand scaffold and a metal center.

Learning more about this subject takes us on a tour of its origins, which are profoundly rooted in biological systems [1] although cooperativity between metal and ligand was first observed in the laboratory in 1966 by Harry Gray on metal-dithiolate synthetic complexes with no biological or biochemical motivation [2]. No connection is at first made between these findings and the biochemical or biological fields, and it almost seems as if the two evolved in complete independence for some time before being conceptually reunited for a few decades.

Ligand participation has always been well known to coordination chemists, but the specific case of redox-active ligands implies a strong orbital overlap and proximity in energy levels between ligand and metal, resulting in complexes with molecular orbitals located on both metal center and ligand. This goes beyond the established electronic action of ligands in the Green formalism (also known as the Covalent Bond Classification Method), in which molecular orbitals can be metal- or ligand-based but imply less orbital mixing [3]. The electronic participation of the ligands then becomes so strong that the matter is sometimes the subject of controversy regarding the relative implications of metal and ligand [4].

1.2 Biological Inspiration: From the Enzyme to the Flask, a Continued Journey

1.2.1 Radicals in Biological Systems

Since a seminal report by Ehrenberg and Reichard on an amino-acid-based radical [5], later identified as a tyrosine radical [6], the implication of radicals in biological functions has become a topic of renewed interest. The most prevalent biologically relevant radicals are based on glycine, cysteine, tyrosine, and tryptophan residues [7] and all play pivotal roles in enzymatic catalytic functions, as reviewed by Stubbe and van der Donk [8]. These functions encompass a broad range of enzymes from ribonucleotide reductases (class-I RNRs), in which a diiron(III)-tyrosyl-radical cofactor is involved in the catalytic conversion of nucleotides into deoxynucleotides [9] to prostaglandin H synthase (PGHS), which also relies on an iron-tyrosyl radical system to effect the conversion of arachidonic acid (AA) into prostaglandin endoperoxide G2 (PGG2), a key step in the biosynthesis of the prostaglandin family. This ubiquity extends to numerous enzymes, cytochrome c peroxidase, DNA photolyase, galactose oxidase (GAO), and amine oxidases being just a few of them.

The specific role of radicals in these enzymes is to perform chemical transformations in a controlled fashion, a notion that is often not readily associated with radical reactivity. In this case, the metalloenzymes harness the reactivity of radicals to avoid possible shortcomings, such as uncontrolled side reactions. Extensive tuning of the redox properties by the environment within the protein is responsible for the very different kinds of chemical reactivities that can be accomplished by these radicals. Indeed, it is well known that the pKa and reduction potential of a specific amino acid residue can vary dramatically depending on its surroundings through electronic and H-bonding effects. The protein-based radical is generated by the action of a neighboring metallocofactor and then proceeds to catalyze a chemical transformation, and can do so through multiple turnover sequences before regenerating the initial (non-radical) form of the protein residue.

1.2.2 Redox Cofactors: Electrons and Protons in Metalloenzymatic Systems

The field of research dealing with redox-active ligands has roots in the natural redox cofactors found in (metallo)enzymes and involved in electron transfer [10, 11]. Three major families are involved in electron and proton transfer in natural systems: quinones, nicotinamides, and flavins. These redox cofactors are mostly derived from amino acids through post-translational modifications and thus offer a wider structural landscape than the original amino acids. A few of the more elaborate structures are presented in Figure 1.1, especially in the quinone family [12]. Interestingly, active research in this area has unraveled new cofactors such as a nickel-pincer nucleotide (NPN) coenzyme [13], which provides fruitful inspiration for chemists to design catalytic systems [14, 15].

Figure 1.1 Three main families of redox cofactors and their redox behavior.

Source: Desage-El Murr [10]/with permission of John Wiley & Sons.

Among the most studied biocatalytic systems with built-in redox-active ligands is the GAO enzyme, which performs the two-electron oxidation of alcohols into aldehydes using molecular O2 as an external oxidant. Established crystal structures have proved the catalytic site of this enzyme to feature a copper (II) ion embedded in a ligand framework consisting of two histidine imidazoles and two tyrosine phenolate residues (Scheme 1.1) [16]. Initial activation of the catalyst by single-electron oxidation generates the active form of the complex, which can then oxidize the primary alcohol. Overall, this is a two-electron process in which one of the electrons is accepted by the tyrosyl radical and the second electron by the copper center, which is consequently reduced to Cu(I). In this example, the Tyr272 moiety acts as an electronic storage unit, accepting one electron and allowing the active site to perform a two-electron redox process with the copper participating formally only by one electron (? d.o.Cu = 1).

The distinctive mechanism and electronic structure of this enzyme have attracted much attention from chemists and opened the way to the design of several GAO mimic systems [17, 18]. Unsurprisingly, alcohol oxidation under aerobic conditions has thus been one of the first reactivities to be reported with redox-active complexes of 3Dmetals [19]. The GAO's unique catalytic system is both a starting point and a proof of concept for strategies aiming to explore new oxidative reactivities of metal centers bound to redox-active ligands [20]. These redox-active ligands are able to stabilize one electron through their radical ligand form and have been successfully applied to the oxidation of simple or activated alcohols. Selected representative structures (Figure 1.2) include salen derivatives and iminosemiquinones.

Scheme 1.1 The galactose oxidase reactivity (top) and detailed catalytic cycle (bottom).

Figure 1.2 Selected copper complexes for alcohol oxidation inspired by GAO.

Examples of this redox-active behavior are, of course, not confined to copper and extend to most naturally abundant 3D metals. The uncertainty in the oxidation state of metal centers in biological systems is a prevalent topic, and P450 compound I (P450-I) was found to have an Fe(IV) oxidation state and an oxidized radical ligand, while the formal oxidation state would be Fe(V) [21].

1.3 Chemical History: Puzzle-Solving for Coordination Chemists

1.3.1 The Curious Case of Metal Complexes with Dithiolene Ligands

The chemistry of coordination complexes of 3D metals with dithiolene ligands led to the discovery of fully synthetic systems exhibiting ligand non-innocence. The intriguing electronic structure of these complexes was studied by three groups [22-25], and these seminal contributions established the possibility of storing electrons on the ligands. This behavior implies that the metal's spectroscopic redox state is different from its formal theoretical redox state, which has strong implications for the chemical and electronic reactivity of the complexes [26]. This field, its founding contribution, and its consequences have been reviewed by Eisenberg and Gray, who have provided an in-depth account of the pivotal role of this early family of redox-active ligands [27] (Scheme 1.2).

Scheme 1.2 Redox activity in metal (M) dithiolene complexes.

An early definition of what an innocent ligand is was provided by Jorgensen, who stated that "ligands are innocent when they allow oxidation states of the central atoms to be defined." [28] This definition suggests that, for specific ligands, electronic transfer can be favored at the ligand rather than at the metal. This rather unusual behavior in coordination chemistry is observed in cases where ligands and metals have inverted energy levels compared to the more classic situation where the metal has the lowest...