Biocatalytic trifluoromethylation of unprotected phenols

Organofluorine compounds have become important building blocks for a broad range of advanced materials, polymers, agrochemicals, and increasingly for pharmaceuticals. Despite tremendous progress within the area of fluorination chemistry, methods for the direct introduction of fluoroalkyl-groups into organic molecules without prefunctionalization are still highly desired. Here we present a concept for the introduction of the trifluoromethyl group into unprotected phenols by employing a biocatalyst (laccase), tBuOOH, and either the Langlois' reagent or Baran's zinc sulfinate. The method relies on the recombination of two radical species, namely, the phenol radical cation generated directly by the laccase and the CF3-radical. Various functional groups such as ketone, ester, aldehyde, ether and nitrile are tolerated. This laccase-catalysed trifluoromethylation proceeds under mild conditions and allows accessing trifluoromethyl-substituted phenols that were not available by classical methods.

Both elements of these chemobio-transformations (i.e. laccases, Langolis' and Baran's reagents) are well known, although there are few examples where laccases are used for synthetically useful reactions. The novelty stems from bringing the two together and using the biocatalyst to obtain regioselectivity that may not be possible using the existing methods.
A small number of substrates are shown to undergo trifluoromethylation with moderate yields. While in some cases mixtures of regioisomers are formed (table 2 entries 4 and 5), it is a clear that there is a preference for one regioisomer (4:1 to 10:1). Entry 6 does give a single regioisomer (2f), however, and the authors do a good job of explaining the regioselectivities by calculating transition state energies. The mechanism they propose based on these calculations is also plausible.
In order to put their work in context, they compare their chemobio-transformation with existing non-enzymatic trifluoromethylation strategies and they show that their method is complementary providing different regioselectivity.
Overall I am supportive of publication. Trifluoromethylation is a very important transformation, particularly in pharma. While there have been a number of synthetic methods developed to affect such transformations, this is still challenging chemistry. The possibility of using an enzyme to trifluoromethylate phenols is novel and will be of considerable interest to those in the synthesis and biocatalysis communities in academia as well as industry.
There are a few issues that the authors should address to improve the paper: 1) The six examples provided in table 2 do provide a proof -of-principle. However the substrates are structurally similar and the paper would benefit from some more examples including greater structural diversity -e.g. phenols with different substituents, biaryls, naphols, anilines... (within the timescale for publication).
2) Is the comparison with literature methods (Scheme 3) fully exhaustive? There are other methods such as copper/Togni reagent etc. It would be more convincing if the chemobiotransformations were compared with a wider range of non-enzymatic approaches (within the timescale for publication).
3) The yields are moderate at the moment. What is the reason for this and why is the yield of entry 6 (2f) based upon recovered starting material? 4) Compound characterisation is thorough. However there are no HRMS reported for any compound, which would normally be required.
Reviewer #2 (Remarks to the Author): In my view, this submission would be appropriate for consideration for dissemination in Nature Communications, subject to the modifications discussed in the detailed review (attached).

Reviewer #3 (Remarks to the Author):
This communication describes a new protocol for the introduction of the CF3 group into unprotected phenols, based in the recombination of two radical species (phenol radical cation and CF3-radical). The phenol radical is formed by a biocatalyst (laccase using O2 as oxidant), whereas the CF3-radical is generated from trifluoromethanesulfinate (TFMS) salts and tBuOOH. DFT calculations have been performed to account for the regioselectivity of the radical recombination process. Although Na-TFMS has already been used as CF3• source for C-H trifluoromethylation of heterocycles (ref. 38), the present approach is innovative regarding both the biocatalytic generation of the other radical and the substrates covered (unprotected phenols). The mechanistic study dealing with the radical recombination is sound. The paper is potentially suitable for Nature Communications, but an important aspect of this study, the generation of CF3• from TFMS and tBuOOH requires further elaboration. The authors state initially that the two radical are formed via two independent pathways, but later on they state that Cu(I) present in laccase is required to react with tBuOOH to give tBuO•, which in turn reacts with TFMS to generate the CF3-radical. The authors should deeply explore this half of the trifluoromethylation, both experimentally and theoretically, to account for the formation of the CF3-radical.

Reviewer #4 (Remarks to the Author):
This is an excellent paper which describes an original cascade combining a laccase with a radical trifluoromethylation. Considering the broad interest in trifluoromethylation and the lack of broad synthetic methods, this is a very significant contribution which will attract broad interest. The paper is carefully designed and the control experiments and calculations co nfirm that the hypothesis that leads to trifluoromethylation of phenols in indeed valid The regioselectivity issue is an interesting one. While symmetric tetrasubstituted arenes yield a single regioisomer, less substituted phenols tend to afford meta-substitution products. In all cases however, the para position of the phenol is blocked by a carbonyl-bearing moiety. What happens if this group is absent? Does the phenol dimerize? What other funtional groups are tolerated by this cascade? The authors state the two independent pathways operate in a cooperative fashion: what do they mean with this? Overall, an outstanding paper which requires only very minor revisions before acceptance  (2f), however, and the authors do a good job of explaining the regioselectivities by calculating transition state energies. The mechanism they propose based on these calculations is also plausible.
In order to put their work in context, they compare their chemobio-transformation with existing non-enzymatic trifluoromethylation strategies and they show that their method is complementary providing different regioselectivity.
Overall I am supportive of publication. Trifluoromethylation is a very important transformation, particularly in pharma. While there have been a number of synthetic methods developed to affect such transformations, this is still challenging chemistry. The possibility of using an enzyme to trifluoromethylate phenols is novel and will be of considerable interest to those in the synthesis and biocatalysis communities in academia as well as industry.
There are a few issues that the authors should address to improve the paper: Within the three months until the deadline for submission of the revision we tested various compounds and selected two addition compounds for scale up to identify the product. It turned out that also another functional group like nitrile moiety is tolerated. Additionally a substrate without a para substitutent was scaled up. This data was added to the manuscript as well as information about substrates which were not converted at all under the reaction conditions investigated or which resulted in complex product mixtures.
Moreover, the reaction system tolerated ketone-, ester-aldehyde as well as nitrilefunctionalities emphasizing the mildness of the reaction. Interestingly nitrogen containing substrates like indol or 4-aminoacetophenone were not converted at all under the reaction conditions investigated, while other substrates like sesamol, 5,6,7,8-tetrahydro-2-naphthol, 2naphthol or meta-dimethylamino acetophenone resulted in complex product mixtures. a ~4% double trifluoromethylation at C2 and C6 was observed; b isolated yield of the pure regio-isomer 2d-meta; c overall isolated yield of both regio-isomers 2e-ortho and 2e-meta. n. p. = not performed.

2) Is the comparison with literature methods (Scheme 3) fully exhaustive? There are other methods such as copper/Togni reagent etc. It would be more convincing if the chemobio-transformations were compared with a wider range of non-enzymatic approaches (within the timescale for publication).
As an extension for the literature method the Togni reagent was also tested for substrate 1a.
The text and Figure 6 were consequently extended.
As a third method the Togni reagent 32,46 was employed for substrate 1a (Fig. 6, bottom); although the substrate was completely converted, product 2a was only a minor product (8%) while two nonidentified main products were detected, which did not contain any CF 3 -group. Thus, the laccase trifluoromethylation reported here is clearly complementary to literature methods tested. +58% +34% of two non-identified main products Figure 6. Trifluoromethylation methods for electron rich arenes from literature tested for comparison with the here presented laccase/tBuOOH concept. One method (top) involved silver as metal while the second and third methods (middle and bottom) were metal free.

3) The yields are moderate at the moment. What is the reason for this and why is the yield of entry 6 (2f) based upon recovered starting material?
To be concise the yield upon recovered starting material was removed and the table adapted. The isolated yields we report range for the Baran reagent between 31% and 62%, which is in the range of various other (non enzymatic) publications; E.g.in Baran's nature paper the yields range from 35 to 79%, having 89% in one single case and in the PNAS paper the yields range in general between 33 to 78% (a single best value is 96%). Higher yields may be obtained by controlling an undesired side reaction, namely the aryl-aryl coupling of two phenol radicals avoiding di-or oligo/polymerization; this may be achieved ensuring that the CF3 radical is present in excess (e.g. by additional electrochemical reactions, see ref. 42.

4) Compound characterisation is thorough.
However there are no HRMS reported for any compound, which would normally be required.
We have now performed HRMS measurements. The additional data was added to the SI. At "General information" details about the equipment and settings were added: HRMS were recorded on an HPLC-TOF   Following also the comment of the first reviewer additional substrates were tested and the scope of functional group tolerated was extended (e.g. nitrile group). We also added nonsubstrates or substrates leading to complex product mixtures. The substrate scope will depend always whether the laccase can oxidise the phenolic substrate; indeed protein engineering may help to adapt the redox potential and allow to transform other substrates.
Also, to enhance the paper, it is recommended that the authors add reference to the following examples of the use of laccase for C---C and C---O bond---forming reactions: ( While paper (1) describes the increase of sulfomethylation reactivity by 33%, thus it reports a C-C bond formation, the second paper deals with hydroxylation via a mediator, which is out of scope of this paper. Thus only the reference for paper (1) was added as reference 38.