Lessons from Nature: Unraveling Biological C—H Bond Activation

Abstract

The cleavage of unactivated C—H bonds is one of the most challenging reactions in chemical biology. Metalloenzymes have evolved that efficiently perform these transformations with exquisite control of selectivity; however, a proposed requirement is the generation of highly reactive intermediates that could be lethal. A thermodynamic argument involving the putative reactive species is outlined, whereby the interplay between two tunable parameters, redox potential and pK_a, may be the key to sustainable function. In addition, factors that control these parameters are also described, including hydrogen-bonding networks found within protein active sites. Synthetic examples are used to corroborate these ideas.

Introduction

This Opinion discusses current ideas of how the active sites of proteins can modulate highly reactive intermediates to perform chemically challenging reactions. The concepts described are rooted in inorganic and physical organic chemistry, but clearly pertain to chemical biology. The reaction in question is the cleavage of C—H bonds, one of the most difficult chemical transformations [i, ii], yet an essential biological process for survival. The thermodynamic stability of most C—H bonds is often cited as a reason for their low reactivity: many biologically relevant substrates have C—H bond dissociation energies of greater than 90 kcal/mol. To overcome this thermodynamic barrier an arsenal of enzymes have evolved that catalyze the cleavage, most having active sites containing at least one metal center. The paradigm system is the cytochrome P450s, whose hydroxylase components have active sites composed of a heme center with a single Fe-thiolate linkage [iii, iv]. There is also a growing family of non-heme iron enzymes that catalyze C—H bond cleavage—these systems usually contain active sites with either a single iron complex [v] or, in the case of soluble methane monooxygenase, a dinuclear iron center [vi, vii]. In the last 50 years an enormous amount of effort has gone into determining the mechanistic details of their function; while great progress has been made there are still many unanswered questions. A central challenge is determining the structure of competent oxidants responsible for the cleavage events, and elucidating their properties that lead to efficient, and in most cases selective, cleavage of C—H bonds.

A common pattern among these metalloenzymes is that dioxygen binding and activation must precede C—H bond cleavage. The mechanistic implication is that a metal-O₂ derived species is involved in the oxidation of substrates. However, the activation of dioxygen is a complicated process, with the identity of intermediates along the reaction paths still in dispute. In many protein systems, it is proposed that O₂ activation leads to high valent metal-oxo species [4, 5], which serve as the key species in the oxidation of substrates. A possible reaction sequence that follows this concept is summarized in Figure 1, in which a radical species and a M—OH compound are initially formed.

Figure 1.

A possible mechanism for metal-oxo mediated C—H bond cleavage (A), thermodynamic cycle describing the BDE_OH (B), and the relationship between redox potential and pK_a for a metal-oxo species in the cleavage of a C—H bond in methane with BDE_C-H = 104 kcal/mol (C).

Generation of Strong Oxidants within a Protein Active Site

There are obvious evolutionary benefits in coupling O₂ activation with C—H bond cleavage, leading to the success of aerobic life. However, an underlying problem is how a highly reactive form of the active site can exist without destroying the protein [viii]. The dioxygen-derived metal species proposed in Figure 1A exist long enough to cause irreversible damage to essential components of the protein. These deleterious effects are often found in synthetic systems and are the major obstacles for the development of useful synthetic oxidation catalysts using O₂ as a substrate. It is well known that active sites impose structural constraints around metal centers to promote efficient catalysis, yet the details into how proteins can withstand the production of powerful oxidants are still lacking. We describe two approaches that could work in synergy to combat oxidative destruction by metal-oxo centers: tuning the basicity of the metal-oxo unit and hydrogen-bonding networks that are proximal to the active site metals. These approaches are illustrated with synthetic systems, in which structure-function relationships have been established.

Tuning the Reactivity of the Metal-Oxo Moiety: The pK_a Effect

The oxidative power of compounds is often discussed within the context of redox potentials, whereby stronger oxidants have higher reduction potentials. This metric is important but is not the only parameter needed to evaluate the oxidative ability of metal-oxo complexes. Mayer has elegantly described the use of thermodynamic cycles to evaluate the ability of metal-oxo complexes to cleave C—H bonds [ix, x]: these analyses examine the O—H bond dissociation energies (BDE_OH) of metal-hydroxo (MOH) complexes formed after the initial cleavage event (Figure 1A). The relevance to C—H bond cleavage is readily apparent: for the scheme in Figure 1 to be thermodynamically viable, the energy required for cleaving the C—H bond must be similar to that produced on forming the MO—H bond.

A thermodynamic cycle for the treatment of this reaction is shown in Figure 1B, from which the BDE_OH equation can be derived: BDE_OH = 23.06 E° + 1.37 pK_a + C [xi]. The constant (C) is included in the analysis to account for the solvation of the hydrogen atom and is dependent on solvent and reference for the redox potential. One can see that both a one-electron reduction potential and a pK_a value are needed to evaluate the BDE_OH. The inclusion of the pK_a (that is, a gauge of the basicity of an oxo ligand) is an important outcome because it is another tunable parameter that affects the reactivity of metal-oxo complexes. The effect is illustrated graphically (Figure 1C) for the cleavage of a C—H bond in methane [xii], which has a bond dissociation energy of 104 kcal/mol. Note that at low pK_a values the redox potentials required are prohibitory to maintain function: for instance, the metal-oxo unit would irreversibly oxidize key amino acids (e.g. tryptophan). However, as the basicity of the oxo ligand increases there is a decrease in the redox potential to values compatible with protein systems.

Thermodynamic cycles similar to the one in Figure 1B have been utilized to understand the interplay between basicity and redox potential of a metal-oxo species in the cleavage of a C—H bond. Note that two key metal-oxo complexes are presented in the cycle and the basicities of oxo ligands in each species could make important contributions. For instance, some have suggested that the basicity of the oxo ligand in the one-electron reduced species (green path in Figure 1B) plays an important role in this process. The most relevant example is from the work of Green, who argues that the redox potentials in P450s are much lower than previously thought, preventing oxidative damage of the active site [xiii]. Specifically, he proposes that the moderate redox potentials in compound I, the iron-oxo radical species thought to cleave C—H bonds, are offset by the relatively high basicity of the oxo ligand of compound II (the one-electron reduced form of compound I). He further contends that ligation of the cysteine thiolates to the iron centers is partly responsible for this effect [xiv]. Support for the basicity of a thiolate-ligated compound II species was revealed recently by resonance Raman spectroscopy [xv].

Recent reports on well-characterized synthetic manganese-oxo complexes support this idea. Examples developed by Goldberg using Mn^V-oxo species containing the porphyrinoid-like ligands, corralozines, illustrate this point [xvi]. [Mn^V(TBP₈Cz)(O)] (Figure 2A) can cleave the X—H (X=C or O) bonds in a variety of substrates, a reactivity attributed to the basicity of the oxo ligand in the one-electron reduced species [Mn^IV(TBP₈Cz)(O)]⁻. A pK_a value of 15 in acetonitrile has been estimated for [Mn^IV(TBP₈Cz)(OH)]⁺, the conjugate acid of [Mn^IV(TBP₈Cz)(O)]⁻, which compensates the modest one-electron reduction potential of +0.24 vs NHE measured for [Mn^V(TBP₈Cz)(O)].

Figure 2.

Structure of [Mn^V(TBP₈Cz)(O)] (A) and [Mn^IIIH₃buea(O)]²⁻ (B).

Synthetic examples have also been reported showing that the basicity of the higher valent metal-oxo species can dictate reactivity (blue path in Figure 1B). A striking example of this is found in the chemistry of [Mn^IIIH₃buea(O)]²⁻ (Figure 2B, M = Mn^III), the rare example of a terminal oxo ligand bonded to a manganese(III) ion [xvii]. This complex has an extremely low reduction potential of less than −1.3 V vs NHE. However, [Mn^IIIH₃buea(O)]²⁻ behaves as an oxidant, cleaving C—H bonds in a variety of substrates. The mechanism leading to this reactivity is still under investigation, but preliminary indicators point toward the basicity of the oxo ligand in [Mn^IIIH₃buea(O)]²⁻ because of its unusually high pK_a of 28.3 measured in DMSO.

Tuning the Reactivity of the Metal-Oxo Moiety: Non-Covalent Interactions

The above discussion illustrates how adjusting redox potential and pK_a can affect the reactivity and durability of metal-based oxygenases. From a broader viewpoint, metalloproteins accomplish chemical transformations with efficiencies, selectivities, and speeds that exceed what is currently possible in synthetic systems. The causes for these differences are just emerging, but clearly factors that govern basic physical properties, such as redox properties and basicity, are crucial aspects. One factor is hydrogen bond (H-bond) networks within protein active sites. Information derived from structural biology has indicated that many metalloproteins, especially those involved in oxidative processes, have (H-bond) networks within their active sites that directly interact with the oxo (or hydroxo) ligands. For example, the structure of compound I of cytochrome C peroxidase determined by X-ray diffraction methods shows a H-bond network involving a structural water molecule and active site amino acid residues that directly interact with the iron-oxo unit (Figure 3A) [xviii]. We have advocated that these types of non-covalent interactions arising from protein-induced microenvironments can have profound impact on function [xix, xx].

Figure 3.

Active site structures of cytochrome C peroxidase (A, 1ZBZ) and human MnSOD (B, 2ADQ) highlighting the H-bonding networks surrounding the metal centers. Gln69 refers to FeSOD.

Evidence for these microenvironmental effects is found in chemical biology and synthetic chemistry. The iron and manganese superoxide dismutases (SODs) dramatically illustrate this effect: these proteins have nearly identical active site structures (Figure 3B) with extensive H-bonding networks surrounding coordinated water molecules [xxi, xxii, xxiii]. It has been shown that interruption of the networks in both SODs causes major changes in function. For instance, substitution of His30 in human MnSOD with a series of amino acids results in a ten-fold diminution in k_cat, even though this residue is more than 5 Å from the manganese center [xxiv]. In another set of detailed structure-function studies, Miller has found that H-bonding networks regulate the protonation state of the bound water molecule (H₂O or OH⁻), which in turn affects the redox potential. Moreover, she showed that a series of single amino acid substitutions at Gln69 in FeSOD produces changes in redox potentials that span 0.660 V [xxv, xxvi]. Substantial alterations in redox potential have also been observed in synthetic systems, whereby changes in the number of H-bonds to copper-chloride units alter reduction potentials by nearly 0.200 V (Figure 5A) [xxvii]. Correspondingly, a series of zinc-hydroxo complexes with varied H-bond networks can change the pK_a values over two units. These pK_a changes correlate with functional changes of approximately 10⁴ in the rate of hydrolysis of phosphate esters [xxviii, xxix].

To investigate these effects in synthetic compounds, we introduced systems that create intramolecular H-bond networks proximal to metal centers [19, xx]. This design has allowed isolation of new types of metal-oxo and their corresponding metal-hydroxo analogs, as discussed above for the manganese-oxo complex (see Figure 2B, M = Mn). In another case, we have prepared the iron-oxo complex, [Fe^IIIH₃buea(O)]²⁻, in which the oxo ligand is derived from the cleavage of dioxygen (Figure 2B, M = Fe). Numerous studies with this complex show that there are three intramolecular H-bonds involving the oxo ligand and the NH group of the [H₃buea]³⁻ ligand [xvii, xxx, xxxi]. Our findings on [Fe^IIIH₃buea(O)]²⁻ and related compounds indicate that H-bonds can regulate key properties of metal-oxo complexes, such as the metal-oxo bond strength, oxo basicity, and redox potentials.

We have elaborated on our initial system to develop a series of compounds, whereby the H-bonding networks can be varied as shown in Figure 4. This design accommodates site-directed modifications of the microenvironment around the metal centers without altering the metal coordination geometry. Therefore, we can systematically address how physical and functional properties change as a function of increasing H-bond donors. In one study, we showed that a decrease in H-bond donors leads to incremental increases in Fe—O bond lengths for a series of Fe^III—OH complexes; these structural changes are correlated to decreases in the strength of the O—H bond [xxxii]. A series of cobalt complexes illustrated the effects of H-bonds on metal-mediated O₂ activation/C—H bond cleavage [xxxiii]. The complexes with H-bond donors are able to bind and activate dioxygen, albeit with different efficiencies. However, the compound without H-bonding capabilities does not bind O₂, even though it has vacant coordination sites and redox properties sufficient for binding/activation.

Figure 4.

Example of a modular synthetic system containing varied H-bonding networks.

Summary

Advances in understanding the cleavage of C—H bonds have been achieved through interdisciplinary studies on a variety of systems. The crucial question of how proteins can maintain function while producing highly reactive intermediates can be understood within a thermodynamic framework that relies on the interplay between basicity and redox potential. Tuning of these two parameters could lead to reactive species with depressed redox potentials that can co-exist with other elements present in active sites. Proximal H-bonds control these basic properties in proteins and are now being used in synthetic systems to discover new functions.