Green oxidations with laccase–mediator systems

Synthesis of aldehydes

Since the reaction is water-based, we thought that laccase– mediator oxidation would be ideal for the synthesis of hetero-aromatic aldehydes, especially pyridines, since these are often quite water-soluble. This oxidation has been demonstrated using stoichiometric oxonium perchlorate salts derived from 4-acetylamido-TEMPO in dichloromethane [20]. Hence we chose to see if catalytic TEMPO could be used, the oxonium salt being generated by laccase oxidation. Indeed, with a few weight% of enzyme and 10–20 mol% of TEMPO in pH 5 buffer at 30–35°C, excellent yields of aldehydes (A–D, Figure 2) could be obtained from the corresponding carbinols. Several attempts to produce 2-acetylpyridine (E) from 1-pyridin-2-yl ethanol gave only very low yields of product although 4-methoxyacetophenone (F) could be readily produced in 50% plus yield under identical conditions [19]. We concluded that (F) or its precursor must inhibit the enzyme.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Products and yields from laccase–mediator oxidations

2,6-Bis(hydroxymethyl)pyridine (G) can give two products, the mono-aldehyde (H) and/or the dialdehyde (I). Our initial attempts gave approx. 1:1 mixtures of (H) and (I), but with moderate fine tuning (certainly not optimized) the yield of mono-aldehyde (H) could be increased to approx. 45%, with the residue being mainly unchanged diol (G). Thus the laccase–mediator method shows promise for the selective mono-oxidation of diols. In comparison, it has been reported that mono-aldehyde (H) can be produced from (G) in 70% yield using 66 mole equivalents of manganese dioxide in chloroform [21].

For all pyridine carbinols, no oxidation was observed with laccase or mediator alone, and TEMPO was the only successful mediator identified, suggesting that an ionic pathway is required for oxidation of these electron-deficient aromatics. ABTS or HBT gave only a few per cent conversion into products [5].

2-Furan-2-yl methanol and 2-thiophen-2-yl methanol were cleanly oxidized to the corresponding aldehydes (J) and (K) with no major by-products detected by HPLC.

Oxidation of sulphur compounds

Free thiols (RSH) have been shown to be potent inhibitors of laccases, presumably via co-ordination of the thiol to the copper atoms in the enzyme active site [15]. We attempted a laccase–TEMPO oxidation of thioanisole, and saw no sulphoxide or sulphone formation, even with high concentrations of enzyme and mediator. While some inhibition of the laccase was apparent, enough activity remained to suggest that laccase oxidation of thioether-containing molecules may be a possibility. Indeed this proved to be the case, with a high yield (88%) of thioanisaldehyde (L) being produced from the corresponding alcohol (Figure 2). This was a very selective reaction with no sulphur oxidation or Dakin-type (phenols) by-products being produced.

Primary benzyl amines

By analogy with aryl alcohols, we expected primary benzylamines to give aldehydes after oxidation to the imine and subsequent hydrolysis. By and large, this was the case for a range of substrates, with aldehydes (M), (N), (O), (P) and (Q) being produced in 40–100% yield (see Figure 2). Oxidation of benzylamine to yield benzaldehyde (R) was very sluggish and low yielding. Further investigation revealed that the laccase used (Trametes versicolor) was severely inhibited by benzaldehyde, suggesting that product inhibition could be problematic with substrates that are oxidized to produce benzaldehyde. Curiously, T. versicolor was inert to most of the other aldehyde products except for 2-thiophein carboxaldehyde (K), which was a weak inhibitor.

During the oxidation of 4-aminomethylbenzoic acid, apart from the expected product (Q), two by-products were detected and identified by HPLC. These were identified as terephthalic acid (T) and 4-cyanobenzoic acid (S), both formed at approx. 10% yield. This interesting observation shows that the laccase–mediator system is capable of oxidizing the initially formed imine intermediate to a nitrile [22], and further oxidizing the primary aldehyde product to a carboxylic acid [23]. The concentration of (S) did not change after complete consumption of the benzylamine, but the concentration of (T) did increase (reaching ∼40% yield) at the expense of (Q).

Tert-amines

N,N-Diethylaminomethylbenzene has been reported to be quantitatively oxidized to benzaldehyde by T. villosa laccase–TEMPO [15]. Due to potential problems with benzaldehyde inhibition, we examined N,N-diethylaminomethyl-4-trifluoromethylbenezne (U), the corresponding aldehyde being relatively non-toxic to T. versicolor. Rather than aldehyde (O), we observed a surprising dealkylation reaction producing (V) and (W) as initial products, giving approx. 1:1:1 mixtures of all three possible amines, along with small amounts of aldehyde (O) (Figure 2).

Following on from this unexpected result, we briefly examined N,N-dimethylaminomethylbenzene and dibenzylamine with laccase–TEMPO and found similar behaviour, e.g. mixtures of products arising from N-dealkylation and benzaldehyde were detected.

Oxidation of hydroxylamines

We also examined the oxidation of hydroxylamines to produce oximes and nitrones and found that for a range of compounds, oxidized products could be produced in very high yield for primary and secondary hydroxylamines. A free hydroxy group is not necessary for the reaction since O-hydroxylamine ethers also give high yields of products (Figure 2). As opposed to pyridylcarbinols, both ionic and radical pathway mediators (TEMPO and ABTS) could give high product yields.

Our key driver for the study of hydroxylamine oxidation was to look for enantioselectivity in the oxidation. In all cases examined where the hydroxylamine was dosed as a racemate, chiral HPLC/GLC showed no enrichment of a single enantiomer. Thus we conclude that the oxidation occurs away from any chiral influence of the enzyme (N.J. Turner, T. Eve and A. Wells, unpublished work).

Aromatization

Although at a very early stage of development, the laccase– mediator system shows promise in certain aromatization reactions, giving high yields of aromatic compounds from dehydro-aromatics (see Figure 2) for the oxidation of Nifedipine. In this case, a radical pathway is preferred, with TEMPO being an inefficient mediator.

As with the carbinols, oxidation occurred under very mild conditions with catalytic amounts of enzyme and mediator. Typical chemical reagents for this type of transformation are chromates, manganese dioxide, hypervalent iodine reagents and oxygen and cobalt catalysts, so laccase–mediator aromatization does appear to be a green alternative.

Fate of mediators

Product or substrate inhibition is a potential drawback for any enzyme-catalysed process. This can to some extent be addressed by supporting the protein in an inert matrix [24] or modifying the reaction conditions [25]. In order to ensure maximum productivity, it is essential to measure the stability and activity throughout the reaction to ensure that active enzyme is present. With laccase–mediator oxidations, we have found that it is also vital to monitor the concentration of mediator present since this can change throughout the reaction. In some reactions, the TEMPO concentration remains fairly constant and catalytic activity is good. In some cases, the TEMPO concentration decreases rapidly at the start of the reaction and then levels off. Some reactions, although few, consume relatively large amounts of TEMPO, rendering the process uncatalytic! Examination of these reactions by GLC–MS showed that the fate of the TEMPO was predominately deoxygenation (Figure 3), to give 2,2,6,6-tetramethylpiperidine. Looking at the catalytic cycle for TEMPO (Figure 3), there is no requirement for oxygen transfer from TEMPO or its oxonium ion in the catalytic cycle, so the mechanism of deoxygenation is unclear – possibly some interaction and oxidation process with the protein?

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Fate of mediators

In other laccase oxidations, we have been looking at HBT as a laccase mediator. HBT acts as a radical mediator rather than ionic, but a severe factor limiting catalytic productivity is again deoxygenation to benztriazole. Indeed, for recalcitrant substrates in the presence of large amounts of enzyme and HBT, deoxygenation is the predominant process seen. When this occurs, rapid deactivation of the enzyme is also seen, suggesting some interaction of the oxidized mediator with the protein. A similar deoxygenation of HBT has been reported in manganese peroxidase-catalysed oxidations [26].

Mediator stability and unfavourable interactions of the oxidized mediator with the protein can severely limit catalytic efficiency of these oxidation processes.