Fenton-like Chemistry by a Copper(I) Complex and ${\text{H}}_2{\text{O}}_2$ Relevant to Enzyme Peroxygenase $\text{C}-\text{H}$ Hydroxylation

Abstract

Lytic polysaccharide monooxygenases have received significant attention as catalytic convertors of biomass to biofuel. Recent studies suggest that its peroxygenase activity (i.e., using ${\text{H}}_2{\text{O}}_2$ as an oxidant) is more important than its monooxygenase functionality. Here, we describe new insights into peroxygenase activity, with a copper(I) complex reacting with ${\text{H}}_2{\text{O}}_2$ leading to site-specific ligand–substrate $\text{C}-\text{H}$ hydroxylation. ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ (${\text{TMG}}_3\text{tren}=1,1,1-\text{Tris}\{2-\text{[}{N}^2-(1,1,3,3-\text{tetramethylguanidino})]\text{ethyl}\}\text{amine}$) and a dry source of hydrogen peroxide, ${\left(o-{\text{Tol}}_3\text{P}=\text{O}\cdot {\text{H}}_2{\text{O}}_2\right)}_2$ react in the stoichiometry, ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}+{\text{H}}_2{\text{O}}_2\to {\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)\right]}^{+}+{\text{H}}_2\text{O}$, wherein a ligand $N$-methyl group undergoes hydroxylation giving ${\text{TMG}}_3\text{tren-OH}$. Furthermore, Fenton-type chemistry (${\text{Cu}}^{\text{I}}+{\text{H}}_2{\text{O}}_2\to {\text{Cu}}^{\text{II}}-\text{OH}+\cdot \text{OH}$) is displayed, in which (i) a $\text{Cu}(\text{II})-\text{OH},(\cdot \text{OH})$ complex could be detected during the reaction and it could be separately isolated and characterized crystallographically and (ii) hydroxyl radical ($\cdot \text{OH}$) scavengers either quenched the ligand hydroxylation reaction and/or (iii) captured the $\cdot \text{OH}$ produced.

Graphical Abstract

INTRODUCTION

Oxidative degradation of biomass such as chitin and cellulose is known¹ to be carried out by bacterial and fungal lytic polysaccharide monooxygenases (LPMOs) which comprise a component of the carbohydrate active enzymes (CAZy) family.^2,3 These mononuclear copper enzymes enable active-site chemistry in the oxidation of recalcitrant polysaccharide C1 and/or C4 $\text{C}-\text{H}$ bonds (see the diagram in the SI⁴) possessing bond dissociation energies of ~101−104 kcal/mol.⁵ Thus, there is considerable potential to generate biofuels in a sustainable manner, by utilizing LPMOs to break down plentiful biomass materials.^1c,2c

Earlier studies revealed a classical monooxygenase activity for LPMOs (Scheme 1a)^1c,2a,6 which possess a mononuclear Cu active site with a tridentate T-shaped coordination, having protein-derived ligation from 2 His residue imidazole N’s plus a primary amine $\left(-{\text{NH}}_2\right)$ derived from the N-terminal His; the latter comprises a chelate, referred to as the “His brace”.^1b,3b,7 In a monooxygenase reaction cycle,⁸ a cupric-superoxide $\left({\text{Cu}}^{\text{II}}\left({\text{O}}_2^{\cdot -}\right)\right)$ species could form via initial ${\text{O}}_2$-interaction with a copper(I) center.^2b,8b,9 This could directly do $\text{HAA}$ or, following electron and/or proton transfers would lead to a ${\text{Cu}}^{\text{II}}$-(hydro)peroxide entity that is further transformed into the key species which would affect the difficult hydrogen-atom abstraction ($\text{HAA}$) reaction (e.g., a copper(II)-oxyl $\left({\text{Cu}}^{\text{II}}-\text{O}\cdot \right)$ species).^1d,3a,8b,10

Scheme 1.

LPMO Reaction Scheme (a) Monooxygenase and (b) Peroxygenase Reaction Pathway; (c) Proposed Mechanisms Relevant to the LPMOs Cu-Site, Processing ${\text{H}}_2{\text{O}}_2$

However, in fact, recent biochemical−biophysical studies^10i,11 detail that LPMOs also are widely functional as peroxygenases and that ${\text{H}}_2{\text{O}}_2$ is faster reacting with reduced copper(I) LPMOs than is molecular oxygen. The peroxygenase biochemistry (Scheme 1b) is found to lead to observable protein damage resulting in lower product yields and loss of reaction selectivity in comparison to the ${\text{O}}_2$-mediated monooxygenase reactivity. Also, computational studies support the viability of LPMO peroxygenase activity.^10h,12 Scheme 1c provides mechanistic pathways which have been proposed or can be considered, for the enzyme ligand−copper(I) $\text{ion}/{\text{H}}_2{\text{O}}_2$ chemistry leading to substrate hydroxylation. The likely reactive species capable of $\text{HAA}$ for these difficult substrates are (i) a hydroxyl radical ($\cdot \text{OH}$) produced by copper Fenton chemistry, $\left({\text{Cu}}^{\text{I}}+{\text{H}}_2{\text{O}}_2\to {\text{Cu}}^{\text{II}}-\text{OH}+\cdot \text{OH}\right)$ (and see below),¹³ (ii) a ${\text{Cu}}^{\text{II}}-\text{O}\cdot ·$ species which may be directly generated from a ${\text{Cu}}^{\text{I}}+{\text{H}}_2{\text{O}}_2$ reaction with release of ${\text{H}}_2{\text{O}}_2$; a related route, that has been suggested, could be if the $\cdot \text{OH}$ moiety produced in the Fenton-like reaction, abstracts an H-atom from the ${\text{Cu}}^{\text{II}}$-hydroxide moiety, ${\text{Cu}}^{\text{II}}-\text{OH}+\cdot \text{OH}\to {\text{Cu}}^{\text{II}}-\text{O}\cdot +{\text{H}}_2\text{O}$ (Scheme 1c)^8b,12,14 and (iii) a high-valent $\text{Cu}(\text{III})$ species,¹⁵ possibly a ${\text{Cu}}^{\text{III}}{(\text{OH})}_2$ complex (not shown in Scheme 1) derived from direct homolytic cleavage of ${\text{H}}_2{\text{O}}_2$ in its interaction with ${\text{Cu}}^{\text{I}}$. The reactive species ${\text{Cu}}^{\text{III}}{(\text{OH})}_2$ would affect substrate $\text{C}-\text{H}$ bond $\text{HAA}$, with one hydroxide bound to copper accepting the proton and then producing ${\text{H}}_2\text{O}$, leaving behind a ${\text{Cu}}^{\text{II}}$-hydroxide species and the substrate carbon radical ($\text{R}\cdot$); rebound,¹⁶ ${\text{Cu}}^{\text{II}}-\text{OH}+\text{R}\cdot \to {\text{Cu}}^{\text{I}}+\text{R}-\text{OH}$, would complete a catalytic cycle.

Several recent biochemical studies^10i,17 on the ${\text{Cu}}^{\text{I}}/{\text{H}}_2{\text{O}}_2$ enzyme reaction reveal generation of protein radicals, via one-electron oxidation of a Tyr and Trp residue near the active site. Solomon and co-workers¹⁰ⁱ could demonstrate direct ${\text{Cu}}^{\text{II}}$ hydroxide formation concomitant with protein radical formation, potentially derived from the $\cdot \text{OH}$ generated and subsequent reactivity. These observations suggest that scenarios (i) or (ii) (see above) may apply, wherein a copper-mediated Fenton reaction initially occurs in LPMOs. Supporting computational results have been published.^10h,12

In the Fenton reaction (with ${\text{Fe}}^{\text{II}}$ or ${\text{Cu}}^{\text{I}}$),^13a–c,18 the particular situation present (e.g., pH in aqueous media, ligand identity) dictates whether $\cdot \text{OH}$ or a high-valent metal-oxo complex forms (e.g., ${\text{Fe}}^{\text{IV}}=\text{O}$);¹⁹ under physiological conditions, carbonate radical anion (${\text{CO}}_3{}^{\cdot -}$) is present, rather than $\cdot \text{OH}$.^13a–c It is well known that iron- or copper ion-mediated Fenton chemistry effect biological substrate $\text{metal-ion}/{\text{H}}_2{\text{O}}_2$ oxidative damage to peptides or nucleic acids, where $\cdot \text{OH}$ may be generated and react in a site-specific (or localized) manner,²⁰ including possibly in LPMOs.¹² Hydrogen peroxide (or ${}^{-}\text{OOH}$) can reduce copper(II) complexes,^10d,21,22 yielding cuprous ions left to react with any excess ${\text{H}}_2{\text{O}}_2$ present, leading to Fenton chemistry. Furthermore, a recent report indicates conditions where ${\text{H}}_2{\text{O}}_2$ reduction of copper(II) coordination complexes is observed; this can occur in situations where the ligand which is binding to the metal ion strongly favors copper(I) (e.g., 2,9-dimethyl-1,10-phenanthroline vs 1,10-phenanthroline), and $\cdot \text{OH}$ is formed if water is present.²³

More broadly, it has been recently suggested that nature may control metal-ion active site oxidative chemistries by utilizing the Fenton reaction in a “constructive manner”.²⁴ It should also be noted that the hydroxyl radical may be generated by photolysis of water at metal/alloy surfaces (or even at the water−gas surface of water microdroplet)²⁵ and in a controlled manner be utilized for organic oxidations including conversion of methane to methanol,²⁶ removal of contaminants in water purification,²⁷ and chemistry applied to bleaching;²⁸ it may even be applied to cancer therapies.²⁹

Here, we illuminate details concerning a chemical system involving a copper-coordination complex, where an LPMO-type peroxygenase reaction is found to occur. Complex ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ reacts with “dry” ${\text{H}}_2{\text{O}}_2$,³⁰ according to Scheme 2, where stoichiometric hydroxylation (i.e., formal insertion of an ‘O’-atom) of one of the twelve (12) outer ligand methyl groups occurs:

$$\begin{gathered} {\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}+{\text{H}}_2{\text{O}}_2 \\ \to {\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)\right]}^{+}+{\text{H}}_2\text{O} \end{gathered}$$

Scheme 2.

Complex ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ Reacts with Dry ${\text{H}}_2{\text{O}}_2$ to Afford ${\text{Cu}}^{\text{I}}$ Complex Product Where a Ligand Methyl Group Has Been Hydroxylated, in Accordance with the Peroxygenase Pathway Postulated in LPMOs; See Text for Further Explanation

This is a peroxygenase reaction; as ${\text{Cu}}^{\text{I}}$ is left as a final product, a potentially catalytic system is established. As described in this report, our conclusion is that this peroxygenase reaction proceeds via Fenton-type chemistry with copper. Among the experimental observations supporting our supposition, are that a ${\text{Cu}}^{\text{II}}$-hydroxide intermediate could be detected (see below) and that an $\cdot \text{OH}$ reactive species (or an equivalent) could be quenched and/or captured.

RESULTS AND DISCUSSION

${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ (Scheme 2) possesses a tripodal tetradentate ${\text{N}}_4$ ligand with strong (highly basic) alkylamine donor groups, thus having some similarity to the nitrogenous ligand environment found at the LPMO Cu-active sites. Complex 1 is known to reversibly bind molecular oxygen giving ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)\left({\text{O}}_2^{\cdot -}\right)\right]}^{+}$,³¹ and it was previously observed that under specific oxidizing conditions, an alkoxide−copper(II) complex ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ could be isolated and structurally (X-ray) characterized (Figure 1a);³² this observation suggested that a ligand methyl group had undergone hydroxylation.

Figure 1.

(a) ChemDraw representation of ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\mathrm{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ based on its crystallographic determination.³² (b) UV−vis spectral changes (over 1 h) when ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ reacts with three equiv ${\text{H}}_2{\text{O}}_2$ in MeTHF at −70 °C. (c) X-band EPR spectrum (red) {$g_{\perp }=2.27$ ($A\perp =82\text{G}$) and $g_{\Vert }=1.99$ ($A_{\Vert }=82\text{G}$)} and simulation (black) of complex 2 in frozen MeTHF at 20 K. (d) Time-resolution of $\text{CSI-MS}$ spectrum for the formation of 2 upon addition of 3 equiv ${\text{H}}_2{\text{O}}_2$ to a solution of 1 at −70 °C, 10 s (red) after injection, then at 54 s (green), and finally at 120 s (black), which is identified as pure alkoxide complex 2.

The experimental observations in that study led to our suggestions that the most likely reactive species which effected the ligand hydroxylation was a ${\text{Cu}}^{\text{II}}$-hydroperoxide, generated (i) directly from ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{2+}+{\text{H}}_2{\text{O}}_2(\text{aq})+\text{base}$, or (ii) by 1-hydroxy-2,2,6,6-tetramethyl-piperidine $(\text{TEMPO-H})$ reductive protonation of the superoxide complex ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)\left({\text{O}}_2^{\cdot -}\right)\right]}^{+}$, or (iii) by reduction of ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{2+}$ and/or ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)\left({\text{O}}_2^{\cdot -}\right)\right]}^{+}$ effected by phenols which were added. We also speculated that a ${\text{Cu}}^{\text{II}}(-\text{OOH})$ could undergo $\text{O}-\text{O}$ cleavage leading to product, via a ${\text{Cu}}^{\text{II}}-\text{O}\cdot$ species, since the reaction of (1) with PhIO also yielded the hydroxylated ligand alkoxide ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}$. However, as discussed and referenced above (Introduction), ligand-copper(II) complexes can be reduced with hydrogen peroxide, and we have ourselves observed such reactivity which appeared to lead to ${\text{Cu}}^{\text{I}}/{\text{H}}_2{\text{O}}_2$ Fenton chemistry.^10d Could reduction of copper(II) to copper(I) in the presence of hydrogen peroxide be involved in that 2008 study?

Thus, we thought to take advantage of this chemical system and explore new chemistry with 1 where we employ Fenton chemistry conditions that might relate to the peroxygenase chemistry in LPMOs, as described in the Introduction. Would addition of hydrogen peroxide to the cuprous complex lead to ligand methyl group hydroxylation and if so, could mechanistic aspects be investigated?

Here, in testing ${\text{Cu}}^{\text{I}}/{\text{H}}_2{\text{O}}_2$ reactivity, the alkoxide-copper(II) complex 2 was indeed formed in the reaction of 1 with three equiv dry ${\text{H}}_2{\text{O}}_2$ (via use of 1.5 equiv ${\left(o-{\text{Tol}}_3\text{P}=\text{O}\cdot {\text{H}}_2{\text{O}}_2\right)}_2$)³³ in 2-methyltetrahydrofuran (MeTHF) at −70 °C (Scheme 3). A dry solid material source of ${\text{H}}_2{\text{O}}_2$ allows for careful stoichometric additions as well as use of organic solvents and cryogenic reaction conditions. Observed in the $\mathbf{1}/{\text{H}}_2{\text{O}}_2$ reaction was a change from colorless to the green compound 2 $\left\{{\lambda }_{\max }\left(\epsilon ,{\text{M}}^{-1}{\text{cm}}^{-1}\right):420(500),875(270)\text{nm}\right\}$ (Figure 1b). A frozen solution EPR spectrum of the reaction solution (Figure 1c) showed, as previously observed, for ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$,³⁴ a reverse axial signal typical of $\text{Cu}(\text{II})$ in a trigonal bipyramidal environment. As was determined previously using ESI-MS,³² we here also confirmed the formation of alkoxide complex ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ employing cold spray ionization mass spectrometry $\text{(CSI-MS)}$; 2 is characterized by a peak at $m/z518.3$ (calcd $m/z518.3$; Figure 1d).

Scheme 3.

Reaction of Complex ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ with 3 equiv ${\text{H}}_2{\text{O}}_2$ Leads to ${\text{Cu}}^{\text{II}}$-Alkoxide Complex (2)

Time resolution of the reaction was achieved by quickly injecting ${\left(\text{o}-{\text{Tol}}_3\text{PO}\cdot {\text{H}}_2{\text{O}}_2\right)}_2$ into a −70 °C solution of 1 into the prechilled $\text{CSI-MS}$ instrument.⁴ The mass spectra clearly show peaks due to ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)(\text{OH})\right]}^{+}$ ($m/z=520.3$) at 10 and 54 s. This diminishes as an increasing amount of ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}$ ($m/z=518.3$) forms; the final product alkoxide builds up as the ${\text{Cu}}^{\text{II}}$-hydroxide intermediate disappears. Since we have not quantitatively determined instrument response factors for the hydroxide vs alkoxide complexes, strictly speaking we can only say that the hydroxide complex (3) forms first. At 120 s, the $\text{CSI-MS}$ signal is essentially pure ${\text{Cu}}^{\text{II}}$-alkoxide 2; the $m/z520$ peak is exactly the intensity expected and observed for authentic 2,⁴ possessing a normal isotope distribution pattern (the effect of ${}^{63}\text{Cu}{/}^{65}\text{Cu}$ isotope abundance). In fact, we show stronger and clearer evidence for initial formation of ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)-(\text{OH})\right]}^{+}$ (3) in other experiments, see below.

At this stage of experiments, the above results suggest:

$$\begin{gathered} {\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})+{\text{H}}_2{\text{O}}_2 \\ \to {\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)(\text{OH})\right]}^{+}(\mathbf{3})+\cdot \text{OH} \end{gathered}$$ (1)

one of the reaction sequences described above and indicated in Scheme 1, essentially the classical Fenton reaction (with copper(I)). Independently, we could generate copper(II)−hydroxo complex ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)(\text{OH})\right]}^{+}(\mathbf{3})$⁴ and characterize experimentally its structure via single-crystal X-ray crystallography (Figure 2).⁴ Unlike the alkoxide complex ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$, no prominent charge-transfer band is apparent for 3;⁴ however, it does display a reverse-axial EPR spectrum ($g_{\perp }=2.22$ ($A_{\perp }=70\text{G}$) and $g_{\parallel }=1.99$ ($A_{\parallel }=85\text{G}$)) and a prominent parent ion peak at $m/z$ of 520.3 in $\text{CSI-MS}$.⁴

Figure 2.

Displacement ellipsoid plot (30% probability level) of one of the two crystallographically independent ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)(\text{OH})\right]}^{+}$ cations (3) at 110(2) K. Hydrogen atoms and lattice solvent molecules are omitted for clarity except for a hydrogen atom on the hydroxo ligand. The hydroxo O-atom is H-bonded to two partially occupied crystal lattice water molecules (not shown); $\text{O}1\text{A}\cdots \text{O}1\text{W}$ (H-bonding) = 2.771 Å (gray, C; white, H; blue, N; red, O; green, Cu).

If the reaction in eq 1 occurs, or even if the products of $\mathbf{1}+{\text{H}}_2{\text{O}}_2$ are $\left({\text{Cu}}^{\text{II}}-\text{O}\cdot +{\text{H}}_2\text{O}\right)$ or ${\text{Cu}}^{\text{III}}{(\text{OH})}_2$ (see above), the $\cdot \text{OH}$ (formally) would attack one of the ligand methyl groups in order to proceed to the alkoxide product, 2. We now present experiments whose results suggest that this is likely the case. When excess ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ is reacted with the ${\text{H}}_2{\text{O}}_2$ reagent, ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})/{\text{H}}_2{\text{O}}_2=5:1$ (via use of 0.1 equiv ${\left(o-{\text{Tol}}_3\text{P}=\text{O}\cdot {\text{H}}_2{\text{O}}_2\right)}_2$), we see from Figure 3a that these reaction conditions do not lead to the observation of the 420 nm UV − vis band associated with alkoxide ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ (i.e., that shown in Figure 1b). This reaction of excess 1 with ${\text{H}}_2{\text{O}}_2$, conditions such that only one (1) equiv ${\text{H}}_2{\text{O}}_2$ would react with one molecule of 1, reveals that essentially no $\text{Cu}\left(\text{II}\right)$ is produced; we observe only ~5% of expected EPR signal intensity which would be due to the presence of a full equivalent of $\text{Cu}\left(\text{II}\right)$ complex (Figure 3b).³⁵

Figure 3.

(a) UV−vis spectral changes of 1 with 0.2 equiv ${\text{H}}_2{\text{O}}_2$ in MeTHF at −70 °C. (b) X-band EPR spectrum of authentic complex 2 (red) and the product solution obtained with 0.2 equiv ${\text{H}}_2{\text{O}}_2$ added to 1 (blue) in frozen MeTHF at 20 K. (c) Matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF MS) spectrum; the reaction of excess 1 and ${\text{H}}_2{\text{O}}_2$ after metal ions were removed by treatment with DIMPI and $\text{KCN}/{\text{CD}}_3\text{CN}$. (d) CSI-MS spectra for the reaction of 1 and 0.2 equiv ${\text{H}}_2{\text{O}}_2$ in MeTHF at −70 °C. The times indicated in the various panels indicate the number of seconds or minutes following sample injection. Also, see the text.

Thus, these results, for reaction conditions where ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})/{\text{H}}_2{\text{O}}_2=5:1$, reveal that:

$$\begin{gathered} \mathbf{5}{\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}+\mathbf{1}{\text{H}}_2{\text{O}}_2 \\ \to \mathbf{4}{\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}+\mathbf{1}{\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}-\text{OH}\right)\right]}^{+} \\ \left(+\mathbf{1}{\text{H}}_2\text{O}\right) \end{gathered}$$

i.e., the oxygenation (by ${\text{H}}_2{\text{O}}_2$) of the ${\text{Cu}}^{\text{I}}$-bound TMG₃tren ligand in ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$, to give hydroxylated ligand ${\text{TMG}}_3\text{tren-OH}$ as a copper(I) complex (following rebound; see also, below), occurs via a peroxygenase stoichiometry, the reaction described in Scheme 2.

However, to further confirm these conclusions, it is required that we show that ligand hydroxylation has occurred, i.e., ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)\right]}^{+}$ is a product. This is, in fact, the case. For the ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})/{\text{H}}_2{\text{O}}_2=5:1$ reaction, the product mixture was quenched at −70 °C with 2,6-dimethyl phenyl isocyanide (DIMPI, as a strong copper(I) specific ligand), the solvent was removed, and the reaction mixture was warmed to RT and then extracted with $\text{KCN}/{\text{CD}}_3{\text{NO}}_2$.⁴ $\text{MALDI-TOF}\text{MS}$ analysis shows that the most intense peak present is due to unreacted ligand ${\text{TMG}}_3\text{tren}$ (Figure 3c: $m/z441.3$, $\left\{\left({\text{TMG}}_3\text{tren}\right)+{\text{H}}^{+}\right\}$ (calcd $m/z441.3$)) which was present in excess. The other major product is one where the methyl group of one ligand has been converted to a $-{\text{CH}}_2\text{OH}$ $({\text{TMG}}_3\text{tren-OH)}$ functionality and in amounts closely correlating with the quantity of ${\text{H}}_2{\text{O}}_2$ added, $m/z479.2$, $\left\{\left({\text{TMG}}_3\text{tren-OH}\right)+{\text{Na}}^{+}\right\}$ (calcd $m/z479\text{.3}$; Figure 3c). The ${\text{TMG}}_3\text{tren-OH}$ peak has very close to 1/4 of the intensity as the peak due to unhydroxylated ligand, ${\text{TMG}}_3\text{tren}$. Thus the reaction yields are very high, appearing to be nearly quantitative since with the limited amount of ${\text{H}}_2{\text{O}}_2$ present, only one out of 5 mole-equiv of ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ can undergo conversion to ${\text{TMG}}_3\text{tren-OH}$.

Additional $\text{CSI-MS}$ based experiments with these reaction conditions where ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})/{\text{H}}_2{\text{O}}_2=5:1$ provide very strong evidence for the Scheme 2 sequence of reactions, i.e., that ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)(\text{OH})\right]}^{+}(\mathbf{3})$ is the initially formed species (as an intermediate). By contrast to the reaction conditions with excess hydrogen peroxide, i.e., the data shown in Figure 1, here ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren}\right)(\text{OH})\right]}^{+}(\mathbf{3})$ ($m/\text{z}=520.3$) is formed in a highly persistent manner (Figure 3d), lasting for many minutes prior to the start to observing alkoxide ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ formation ($m/\text{z}=518.3$; Figure 3d, from 7 min after sample injection, on). It should be emphasized that formation of hydroxide complex 3 implies that the hydroxyl radical must be forming concomitantly (also see Scheme 2).

Experimental observations that further support our characterization of this peroxygenase system (Scheme 2) are:

1. The ${\text{TMG}}_3\text{tren}$ ligand has been hydroxylated prior to formation of the final ${\text{Cu}}^{\text{II}}$-alkoxide complex, supporting the reaction as given by eq 1 (vide supra). When the ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})/{\text{H}}_2{\text{O}}_2=1:3$ is quenched prior to alkoxide ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}\text{(}\mathbf{2}\text{)}$ formation (based on following UV−vis changes up to where the 420 nm absorption just starts to be observable), a similar workup and analysis of organics reveal that high yields (>95%) of ${\text{TMG}}_3\text{tren-OH}$ are obtained. The prominent ion peak at $m/z$ of 457.2 is assigned to $\left\{\left({\text{TMG}}_3\text{tren-OH}\right)+{\text{H}}^{+}\right\}$ (calcd $m/z457.3$; Figure 4a). Only a trace peak for the starting initial unhydroxylated ligand, ${\text{TMG}}_3\text{tren}$ ($m/z441.3$), is observed in the $\text{MALDI-TOF}\text{MS}$ spectrum (Figure 4a).

2. With the excess dry ${\text{H}}_2{\text{O}}_2$ added, we observed additional products of ligand oxygenation, including the overoxidized aldehyde product. Following workup of the reaction mixture containing ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ and utilizing the DIMPI procedure to remove copper ions (vide supra), mass spectrometric analysis of the organics present reveals that together with a small amount of un-oxidized/oxygenated ${\text{TMG}}_3\text{tren}$, several ligand oxidized types are present (Figure 4). They are (i) the ligand hydroxylated alcohol $\mathbf{\text{L-OH}}$ $\left\{\left[{\text{TMG}}_3\text{tren}-\left({\text{CH}}_3\right)\text{N}-{\text{CH}}_2\text{OH}\right]+{\text{Na}}^{+}\right\}$ ($m/z479.2$, calcd $m/z479.3$; Figure 4b), (ii) the $\text{N}-\text{H}$ species arising from ${\text{TMG}}_3\text{tren-}\left({\text{CH}}_3\right)\text{N}-{\text{CH}}_2\text{OH}$ $N$-dealkylation, $\mathbf{\text{L-NH}}$, $\left\{\left[{\text{TMG}}_3\text{tren}-\left({\text{CH}}_3\right)\text{N}-\text{H}\right]+{\text{K}}^{+}\right\}$ ($m/z465.2$, calcd $m/z465.3$; Figure 4b), (released formaldehyde is observed),⁴ and (iii) a small amount of overoxidized aldehyde species $\mathbf{\text{L-CHO}}$ $\left\{\left[{\text{TMG}}_3\text{tren-}\left({\text{CH}}_3\right)\text{N}-\text{C}\left(O\right)\text{H}\right]+{\text{K}}^{+}\right\}$ ($m/z493\text{.2}$, calcd $m/z493.3$; Figure 4b).

Figure 4.

$\text{MALDI-TOF}\text{MS}$ spectrum; metal ions removed by treatment with DIMPI and $\text{KCN}/{\text{CD}}_3\text{CN}$. (a) Prior to, or (b) after, the formation of ${\text{Cu}}^{\text{II}}$-alkoxide species with excess of ${\text{H}}_2{\text{O}}_2$ added to 1. (c) Oxidized products in the reaction of 1 and excess of ${\text{H}}_2{\text{O}}_2$.

To provide still further evidence for this Fenton-like chemistry, we sought to identify the presence of $\cdot \text{OH}$ (or its equivalent) by employing trapping reagents and/or external substrates which have $\text{C}-\text{H}/\text{O}-\text{H}$ bonds (Scheme 4). Inclusion of ten (10) equiv 2,4,6-tri-$t$-butylphenoxyl radical $(\text{T}t\text{BuArO}\cdot )$ with solutions of ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ prior to addition of ${\text{H}}_2{\text{O}}_2\left(\mathbf{1}/{\text{H}}_2{\text{O}}_2=5:1\right)$ quells the peroxygenase type ligand hydroxylation chemistry (Scheme 2); little or no alkoxide complex 2 is formed (UV−vis criterion). We deduce that $\cdot \text{OH}$ produced by the $\mathbf{1}/{\text{H}}_2{\text{O}}_2$ reaction reacts with excess $\text{T}t\text{BuArO}\cdot$ present, and elimination of isobutylene (formed in 59% yield) as well as additional documented phenolic chemistry³⁶ gives 2,6,-di-$t$-butyl-1,4-hydroquinone (as explained in the SI), which is detected in GC−MS as 2,6-ditert-butyl-1,4-benzoquinone formed in 20% yield based on copper (so effectively ~100%).⁴ This implies capture of “$\cdot \text{OH}$” in near quantitative yields. However, addition of only two equiv. $\text{T}t\text{BuArO}\cdot$ gave only an ~25% yield of the benzoquinone; the efficiency of trapping goes up as the quantity of added trapping agent is increased. Related experiments with excess ${\text{H}}_2{\text{O}}_2$ and monitored by EPR spectroscopy are also consistent with our conclusions (Figure S5).⁴

Scheme 4.

Capture/Trapping of a Hydroxyl Radical ($\cdot \text{OH}$) Derived from ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ Reactivity with Hydrogen Peroxide^a

^a Partial or nearly full inhibition of the peroxygenase chemistry where 1 is converted to alkoxide ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\mathrm{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$ occurs. Also, see the text.

For an experiment where 10 equiv trityl radical were added (as Gomberg’s dimer) (Scheme 4), again no alkoxide complex 2 formed (UV−vis criterion). Here, the $\cdot \text{OH}$ released from the copper complex $(\mathbf{1})/{\text{H}}_2{\text{O}}_2$ reaction would be trapped by the trityl radical to directly form triphenylcarbinol; this was generated in 18% yield. As was mentioned above, due to the stoichiometry of this reaction, this 18% yield is very high, as 20% is the theoretical maximum. Again, when only a limited amount of added Gomberg’s dimer is used (2 equiv), the trapping efficiency is only 3% based on the amount of copper and the stoichiometry of reaction employed $\left(\mathbf{1}/{\text{H}}_2{\text{O}}_2=5:1\right)$. Excess amounts of added 2,6-di-$t$-butyl-4-methoxyphenol or xanthene were also observed to “capture” the $\cdot \text{OH}$ generated in reactions of 1 with ${\text{H}}_2{\text{O}}_2$ (Scheme 4), through $\text{HAA}$ to produce 2,6-di-$t$-butyl-4-methoxy phenoxyl radical and xanthone, respectively.⁴ See Table S3 for details/yields for the trapping/quenching experiments.

It is interesting to survey a number of recently published LPMO biomimetic studies.^22,37 In those reports involving ligand−copper(II) complex reactions with added hydrogen peroxide and an oxidizable substrate, Simaan and Hitomi,^37a Kaizer,²² Itoh,^37b Castillo,^37c,d Cowan,^37e and their co-workers have utilized mono- or poly-nuclear ${\text{Cu}}^{\text{II}}$-complexes, some with a His-Brace like ligand. Added ${\text{H}}_2{\text{O}}_2$ (aq) likely leads to ${\text{Cu}}^{\text{II}}\text{-OOH}$ moiety and to oxidation of ligands (e.g., $\text{ACC}$ oxidase substrate analogs) or exogenous glucose derivatives (e.g., as polysaccharides or surrogates). However, neither a specific ${\text{O}}_2$ reduced-derivative (e.g., $\cdot \text{OH}$) nor a metal-based strong oxidant (e.g., a ${\text{Cu}}^{\text{II}}\text{-oxyl}$) has been yet identified. It is notable, however, that Simaan and Hitomi,^37a and Kaizer²² provide evidence that excess ${\text{H}}_2{\text{O}}_2$ at some stage effects cupric ion reduction (via $\text{Cu}-\text{O}$ heterolytic cleavage of the presumed ${\text{Cu}}^{\text{II}}-\text{OOH}$ moiety)^10c,10,21b and the real oxidant species is something like “${\text{Cu}}^{\text{I}}-\text{OOH}$.” This latter hypothesis points to Fenton-like reactivity.

Based on these experimental results, we can establish plausible reaction pathways for the ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ and dry ${\text{H}}_2{\text{O}}_2$ reaction which leads to $\text{C}-\text{H}$ activation in an overall peroxygenase reaction, shown in Scheme 5 for the differing stoichiometries tested experimentally. The most likely initial reaction is formation of a ${\text{Cu}}^{\text{II}}$-hydroxide complex (3) plus a $\cdot \text{OH}$ species. The latter reactive entity performs $\text{HAA}$ from a ligand methyl group, producing water and a ligand carbon radical; subsequent rebound from the ${\text{Cu}}^{\text{II}}$-hydroxide gives ${\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)$.³⁸ This reaction mechanism was evaluated and is further supported, by density functional theory (DFT) calculations on the full complex and its reaction with ${\text{H}}_2{\text{O}}_2$.

Scheme 5.

Proposed Courses of Reaction of ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$ with Varying Amounts of ${\text{H}}_2{\text{O}}_2$^a

^a (Upper): $\mathbf{1}/{\text{H}}_2{\text{O}}_2=5:1$ and the chemistry shown is for that one complex which reacts with ${\text{H}}_2{\text{O}}_2$ in a stoichiometric manner. (Lower): excess ${\text{H}}_2{\text{O}}_2$ relative to complex (1) produces the same hydroxylated ligand complex ${\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)$; however the excess oxidant present leads to further chemistry (far right).

Figure 5 shows the calculated reaction coordinate based on the proposed mechanism in Scheme 5, top (see the SI for computational details). In the initial structure (Figure 5, 0), the ${\text{H}}_2{\text{O}}_2$ associates with the complex through van der Waals interactions but does not bind directly to the $\text{Cu}$ ($\text{Cu}-\text{O}$ distance: 3.27 Å). The reaction proceeds through homolytic cleavage of the ${\text{H}}_2{\text{O}}_2$ forming a ${\text{Cu}}^{\text{II}}-\text{OH}$ and $\cdot \text{OH}$ that is 14.9 kcal/mol downhill in $\Delta G$ (Figure 5, 2) through a low barrier of $\Delta G^{‡}=3.0\text{kcal}/\text{mol}$ (Figure 5, 1). Immediately after homolytic cleavage, the resulting $\cdot \text{OH}$ (Figure 5, 2a) is not properly oriented to abstract an H atom from the ligand methyl group and must reorient to the proper conformation (Figure 5, 2b) to perform $\text{HAA}$ from the $\text{C}-\text{H}$ bond. This rearrangement involves a small increase in the $\text{O}\cdots \text{O}$ distance (2.25 Å in 2a to 2.56 Å in 2b) and a rotation of the $\cdot \text{OH}$ fragment; this proceeds through a low barrier of 0.9 kcal/mol (Figure 5, 2a–2b). From 2b, the $\cdot \text{OH}$ performs $\text{HAA}$ from the ligand methyl $\text{C}-\text{H}$ bond with almost no barrier, $\Delta G^{‡}=0.35\text{kcal}/\text{mol}$ (Figure 5, 3), producing a water molecule and the ligand methyl radical. This $\text{HAA}$ step is further downhill by 23.1 kcal/ mol in $\Delta G$ (Figure 5, 4). Finally, the methyl radical rebound occurs with the highest barrier in this process, $\Delta G^{†}=7.0\text{kcal}/\text{mol}$ (Figure 5, 5), due to the significant steric reorganization of the complex to reach this transition state. The ${\text{Cu}}^{\text{I}}$ hydroxylated ligand complex product 6 (i.e., ${\text{Cu}}^{\text{I}}({\text{TMG}}_3\text{tren-OH})$) is 28.1 kcal/mol downhill from the previous step and 66.5 kcal/mol downhill from the starting structure. Each step in the proposed mechanism is thermodynamically favorable, with very low barriers for $\text{O}-\text{O}$ cleavage and $\text{HAA}$, and a reasonable, limiting barrier for the rebound hydroxylation. Furthermore, the $\cdot \text{OH}$ reorientation to a conformation conducive to $\text{HAA}$ from the ligand would result in a finite lifetime for the ${\text{Cu}}^{\text{II}}-\text{OH}$ and $\cdot \text{OH}$, consistent with the observation of species 3 by $\text{ESI-MS}$ (Figures 2 and 3) and the radical trapping results. Thus, the calculations in Figure 5 show that the proposed mechanism in Scheme 5 is thermodynamically and kinetically feasible, and consistent with the experimental results presented above.

Figure 5.

DFT-calculated reaction coordinate for homolytic ${\text{H}}_2{\text{O}}_2$ cleavage, subsequent $\text{HAA}$ and rebound ligand hydroxylation by ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}$. Optimized structures and singlet energies are shown for each species. Thermodynamics are calculated at −70 °C.

Scheme 2 and the upper part of Scheme 5 represent a first round of a peroxygenase catalytic cycle, as $\text{Cu}(\text{I})$ is regenerated and can accept a new substrate (here, a new unhydroxylated ligand). However, when excess ${\text{H}}_2{\text{O}}_2$ is present (Scheme 5, bottom), ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)\right]}^{+}$ is oxidized by hydrogen peroxide to a cupric form, subsequently leading to ${\text{TMG}}_3\text{tren-OH}$ deprotonation and formation of alkoxide ${\left[{\text{Cu}}^{\text{II}}\left({\text{TMG}}_3{\text{trenO}}^{-}\right)\right]}^{+}(\mathbf{2})$. $N$-Dealkylation can otherwise occur (vide supra), producing formaldehyde plus a ${\text{Cu}}^{\text{II}}\left({\text{TMG}}_3\text{tren-}\left({\text{CH}}_3\right)\text{N}-\text{H}\right)$ species, as we observe experimentally.

CONCLUSIONS

In this study, we have provided considerable new insights into site-specific Fenton-type peroxygenase chemistry, quite likely relevant to LPMOs³⁹ and perhaps also to copper-dependent $p$-methane monooxygenases ($p\text{MMOs}$).^24,40 Using a synthetic analog ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}(\mathbf{1})$, ${\text{Cu}}^{\text{I}}/{\text{H}}_2{\text{O}}_2$ reactions occur. Our experimental results indicate that this leads to a cupric-hydroxide plus hydroxyl radical as suggested in the study on a LPMO by Solomon and co-workers;¹⁰ⁱ subsequent $N$-methyl group hydroxylation occurs leaving behind ${\text{Cu}}^{\text{I}}$. The generation of a $\cdot \text{OH}$ intermediate (or possibly a ${\text{Cu}}^{\text{III}}{(\text{OH})}_2$ or ${\text{Cu}}^{\text{II}}-\text{O}\cdot$ species) was demonstrated via capture or quenching with radical scavengers or external substrates. The proposed reaction mechanism is further determined to be thermodynamically and kinetically feasible by DFT reaction coordinate calculations. The overall reaction, ${\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren}\right)\right]}^{+}+{\text{H}}_2{\text{O}}_2\to {\left[{\text{Cu}}^{\text{I}}\left({\text{TMG}}_3\text{tren-OH}\right)\right]}^{+}$ $\left(+{\text{H}}_2\text{O}\right)$, is consistent with LPMO peroxygenase catalytic behavior. This study provides a fresh perspective on Fenton-like copper chemistry and previously proposed mechanisms and nature of key intermediates in peroxygenase reactivity, including LPMOs.