Trifluoromethylation of pyridines through skeletal rearrangement

Abstract

We report a skeletal rearrangement of pyridines that enables functionalisation of the 2 and 3-positions, via Zincke ketone intermediates. The ring-opened pyridines undergo selective reaction with cheap and readily available acylating agents, setting up an alternative ring closure pathway to afford rearranged structures incorporating groups such as CF₃, CF₂H, and CO₂Et at the pyridine C2 position. The process enables highly selective C2 over C4 trifluoromethylation of pyridines, a reaction that challenges conventional Minisci methods. Sequential functionalisation of the Zincke intermediate is further demonstrated, to access triply-substituted pyridine products from mono-substituted starting materials.

---

Introduction

Heteroarene ring-opening/ring-closing rearrangements have a rich history in synthesis, enabling skeletal re-organisations to be exploited as strategic disconnections. The Dimroth rearrangement from 1909 is archetypal (Scheme 1A),¹ first observed on amino-triazoles, and since expanded to numerous nitrogenated aromatic systems. General features consist of an initial ring-opening, typically via pericyclic reaction or nucleophilic addition, to an open-chain intermediate (e.g. 2). A σ-bond rotation then enables ring closure via an alternative atom, with a resultant switching of substituent positions. For the Dimroth rearrangement, the heteroarene undergoes overall single-atom exchange (N to N), illustrated in Scheme 1B for a process-scale synthesis of a drug candidate developed by Goundry and co-workers.² Other well-known examples include the Cornforth rearrangement of acyloxazoles,³ a thermal pericyclic reaction proceeding through a nitrile-ylide intermediate that leads to a double atom switch (C and O, elegantly applied to macrocyclic peptide contraction by the Yudin lab, Scheme 1C),⁴ and the Kost–Sagitullin pyridine recyclisation to give arene products⁵ (N=C switching for C=C in this recent example from Kano, Morofuji and co-workers, Scheme 1D).⁶

Scheme 1 (A–D) Heteroarene ring-closing/ring-opening rearrangement. (E) Zincke imine reaction and proposed Zincke skeletal rearrangement idea.

The Zincke reaction of pyridiniums is a classic heteroarene ring-opening/ring-closing reaction, but it does not typically lead to skeletal rearrangement (Scheme 1E). Rather, the original reaction enables exchange of the NR moiety, whereby the open-chain Zincke intermediate 12 undergoes amine exchange and re-closure to preserve the initial pyridine skeleton connectivity.⁷ Recent advances in Zincke chemistry have shown it is possible to functionalise the open-chain intermediate 12 in new ways, which on re-closure give valuable 3- and 4-substituted pyridine derivatives 13 that are difficult to obtain directly.^8–16 This approach has been extended to several recent skeletal-editing studies of pyridines, where the addition of nucleophile ring-opening/ring-closing (ANRORC) mechanism sets up a transformation into a different heterocycle.^17–20 Our work in this area^9,14,18 prompted us to consider a functionalisation – rearrangement mode, whereby a pyridine is opened to the Zincke intermediate 15, and functionalised such that an alternative closure pathway is created. Such a rearrangement could create new ways of installing pyridine functionality via skeletal reorganisation.²¹

The idea is set out in Scheme 1E, showing an electrophilic acylating agent such as TFAA reacting regioselectively with the Zincke imine 15. This could then set up an alternative ring-closing pathway, via the newly introduced trifluoroacyl carbonyl group. Under the ammonia ring-closure conditions typical for pyridine reformation, preferential imine formation could take place at the newly introduced ketone group to give the skeletally re-organised pyridine 17. The original C2 atom is now external to the ring at the C3 position, and the acylating agent has introduced a new CF₃ group at C2. Direct trifluoromethylation at the pyridine C2 position has received considerable attention in the literature, given the relevance of both azines and the CF₃ group to drug design.²² The proposed route would offer a novel, economic alternative to Minisci-based and related methods, which often employ expensive sources of CF₃ and struggle to discriminate between the C2 and C4 positions.

Results and discussion

We began by examining trifluoromethylation of the Zincke imine 15a derived from 2-phenyl pyridine 14a, using trifluoroacetic anhydride (TFAA) (Table 1). Acylation of Zincke imines has yet to be described in the literature, although two methods for Vilsmeier formylation were recently disclosed: The Fu lab demonstrated formylation of related streptocyanines derived from 3- and 4-substituted pyridines, and the Gryko lab reported formylation of 2-arylpyridines via Zincke imines to access rearranged ketone products.^15,23,24 Our initial trials with TFAA were low-yielding, but instructive; we were pleased to observe the desired trifluoroacetylated product 16a, but only at low conversions. Further, when 16a was heated with NH₄OAc, we observed equal amounts of both pyridine regioisomers 17a (rearranged) and 18. Electrophilic addition had taken place exclusively at the 3-position, essential for the postulated reaction to work, but the presence of the NTf imine appeared problematic in terms of both ring-closure and overall efficiency of the acylation. Switching to the Zincke ketone 15b, available through simple hydrolysis, solved both problems. When 15b was treated with TFAA in THF at room temperature, followed by heating with NH₄OAc in EtOH, we were delighted to observe the formation of the rearranged 2-trifluoromethylpyridine 17a in a very good 80% yield. A short optimisation study of temperature and reaction time led to no further improvements (Table 1), so we took these conditions forward to investigate reaction scope.

Table 1 Reaction development

Entry	Starting material	Variation	Yield of 17a (%)
1	15a	None	<10
2	15b	None	80
3	15b	Without DMAP	0
4	15b	60 °C in step 1	77
5	15b	0 °C in step 1	80
6	15b	Overnight step 1	80

---

Initially, we were interested in examining the scope of acylating agents that could react with the Zincke ketone to access differentially functionalised pyridines at C2 (Scheme 2). We were pleased to see that both the electrophilic reagents (F₂HCCO)₂O and (F₂ClCCO)₂O were successful, delivering the CF₂H and CF₂Cl substituted pyridines 17b and 17c in good yield. Both the CF₂H and CF₂Cl groups are valued in medicinal and agrochemistry for their lipophilic H-bonding and halogen-bonding properties, respectively,²⁵ and are typically introduced via Minisci radical reactions. While these methods are well developed for the generation of the respective radicals,²⁶ selectivity for the 2 over the 4-position is not generally possible, and mixtures of pyridines are generated.

Scheme 2 Substrate scope. Conditions: (A) electrophile addition step: Zincke ketone (0.2 mmol), acylating agent (1.1 equiv.), DMAP (20 mol%), triethylamine (2 equiv.), THF (0.1 M), r.t., 2 h. (B) Ring-closing step: NH₄OAc (10 equiv.), EtOH (0.05 M), 60 °C, 2–4 h.

We could also introduce the perfluoro congeners C₂F₅ and C₃F₇ with a similarly effective acylation step, but lower regioselectivity in the pyridine recyclisation (overall yields of 55% and 46%). It appears that the greater steric bulk of the perfluoroalkyl groups diminishes the selectivity between ketone groups in the ammonia cyclisation step. We were also successful with oxalyl reagents to access the picolinate derivatives, 17f and 17g, introducing ester functionality to complement the perfluoroalkyl building blocks. The conditions were effectively scaled up to gram scale for Zincke ketone 15b (3 mmol), delivering the rearranged product 17a in an identical 80% yield. Less electrophilic acyl donors were ineffective in the reaction, with simple Ac₂O, AcCl and BzCl undergoing no reaction with the Zincke ketone.

With the scope of acylation and the C2 substituent established, we moved on to examine the C3 aroyl group installed through the rearrangement (Scheme 3). We used trifluoroacetic anhydrides, difluoroacetic anhydrides, and chlorodifluoroacetic anhydrides as the model electrophilic agents, with the Zincke ketones derived from 2-aryl substituted pyridines. The reaction was generally tolerant of various aryl functionality; Zincke ketones with an electron-neutral and electron-donating phenyl substituent worked effectively (17h, 17i and 17j), along with extended conjugated substituents such as vinyl (17k) and phenyl groups (17l). The structure of this biphenyl derivative was confirmed by X-ray crystallography (CCDC 2487303), illustrating the CF₃ substituent forcing the pyridyl group to twist out of planarity with the 3-aroyl substituent. Halogenated, disubstituted, and electron-withdrawing Zincke ketones were likewise found to undergo rearrangement in a good yield (17m and 17n). Zincke ketones derived from 2,4-disubstituted pyridines were also found to react effectively with TFAA to furnish more densely functionalised pyridine products (17o and 17p); however, Zincke ketones derived from 2,5-disubstituted pyridines were not viable in the TFAA reaction. Next, we synthesised a variety of heteroaryl-substituted Zincke ketones to study the C3 acylation with trifluoroacetic anhydride. Thiophene and furan-substituted Zincke ketones were good substrates (17q and 17r), along with the Zincke ketone derived from a substituted bipyridine to give the pseudo-symmetric C3 acylated compound (17s). Increasing the complexity, we used tricyclic (hetero)aromatic oxacycle and azine, such as dibenzofuran-, carbazole- and benzofuropyridine-substituted Zincke ketones, with trifluoroacetic anhydride.

Scheme 3 Substrate scope. Conditions: (A) electrophile addition step: Zincke ketone (0.2 mmol), acylating agent (1.1 equiv.), DMAP (20 mol%), triethylamine (2 equiv.), THF (0.1 M), r.t., 2 h. (B) Ring-closing step: NH₄OAc (10 equiv.), EtOH (0.05 M), 60 °C, 2–4 h. ^a Trifluoroacetylation step was performed at 0 °C.

We were pleased to see that each reacted well with TFAA, delivering C3 acylated heterocycles (17t, 17u, and 17v). These groups are valuable in medicinal and material chemistry due to their extended conjugation and photophysical properties. The structures of these valuable compounds were confirmed by X-ray crystallography (CCDC 2478546 and CCDC 2515793). We were also able to derivatise the aryl motif of the diabetes medication canagliflozin, giving 17w in 45% yield. Zincke ketones from simple 2-alkyl-substituted pyridines did not react with TFAA (see SI for a list of challenging substrates).

We next moved on to investigate the viability of para-substituted pyridines (Scheme 4). Trifluoroacylation was feasible on the Zincke aldimine derivative 15 directly, with ring closure giving a ca. 1 : 2 mixture in favour of the non-rearranged product 20a, representing direct meta-trifluoroacylation. As expected, the electrophilic aldimine is competitive with the trifluoroacyl group in the ring-closure step, leading to both constitutional isomers. We could isolate the skeletal-rearranged products 19a–19f as the minor components, incorporating a trifluoromethyl substituent at C2 and with the original C2 carbon now rearranged into the 3-formyl group. The formyl group presents a versatile vector for further functionalisation, complimenting the aroyl group installed from 2-substituted pyridines.

Scheme 4 Substrate scope. Conditions: (A) electrophile addition step: Zincke imine (0.2 mmol), trifluoroacetic anhydride (1.2 equiv.), DMAP (20 mol%), triethylamine (2 equiv.), THF (0.1 M), 40 °C, 2 h. (B) Ring-closing step: NH₄OAc (10 equiv.), EtOH (0.05 M), 60 °C, 2–4 h. *The rest of the mass balance corresponds to the non-rearranged product.

The sequence could be conducted as a telescoped process with no chromatographic purification of intermediates, starting from 2-phenyl pyridine 14a (Scheme 5). The Zincke ketone was made in a one-pot procedure, followed by simple work-up and concentration. The crude material was then subjected to trifluoroacetylation and NH₄OAc treatment to give the final pyridine product 17a in 60% overall yield. The reaction transfers an aryl group from the pyridine 2-position, a very accessible position to functionalise with numerous analogs available commercially, to the more challenging 3-position as an aroyl unit. We could further demonstrate the derivatizations of this 3-aroyl group to access diarylmethane derivatives, versatile building blocks with widespread application in chemistry and materials. Methylenation with Tebbe's reagent gave the 1,1-diarylethylene 21, and we could reduce the ketone with NaBH₄ in butanol to access the diarylmethanol 22. We synthesised 3-benzoyl picolinic acid 23 by alkaline hydrolysis of our rearrangement picolinate product 17g, with X-ray analysis of the α-iminocarboxylic acid structure showing a unit cell consisting of two molecules – one neutral and the other in zwitterionic form.

Scheme 5 Telescoped process and synthetic utility.

The nucleophilicity of the Zincke intermediate enables functionalisations that would not be possible directly on the pyridine. We wondered if we could further expand the viability of our skeletal rearrangement via sequential electrophilic addition, to achieve two C–H functionalisations. The Zincke intermediate is nucleophilic at both C3 and C5, with preference for C3 generally being observed for simple electrophiles. An initial C3-selective trifluoroacetylation followed by the addition of a second electrophile could access sequentially meta-substituted pyridines, following ring closure with NH₄OAc. We investigated this idea through trifluoromethylation of Zincke ketone derived from 2-phenylpyridine 14a at C3 to set up the skeletal rearrangement, followed by treatment with either NFSI, NCS, NBS, or NIS as described by McNally.^8,10 Ring closure in the usual way then gave each of the 2-trifluoromethyl-3-aroyl-5-halo pyridines (24a–d) in good overall yield, with the sequential Zincke ketone functionalisation taking place in a one-pot, telescoped process (Scheme 6).

Scheme 6 Polyfunctional pyridine synthesis with overall yields from 2-phenylpyridine. Conditions for halogenation-trifluoromethylation: (A) Zincke ketone formation: 2-phenylpyridine (0.2 mmol), Tf₂O (1.0 equiv.), EtOAc (0.1 M), −78 °C, 30 min, then Bn₂NH (1.2 equiv.), collidine (1.0 equiv.), 30 min, concentrate. Then K₂CO₃ (2.0 equiv.), THF : H₂O (4/1, 0.1 M), 100 °C, 18–24 h. (B) Trifluoroacylation: Zincke ketone (0.2 mmol), trifluoroacetic anhydride (1.1 equiv.), DMAP (20 mol%), triethylamine (2 equiv.), THF (0.1 M), r.t., 2 h. (C) Halogenation step: N-halosuccinamide (1.1 equiv.), r.t., 1 h. (D) Ring-closing step: NH₄OAc (10 equiv.), EtOH (0.05 M), 60 °C, 2–4 h. ^aNFSI (1 equiv.) was used, stirred at r.t. overnight. Conditions for arylation-trifluoromethylation: conditions as previous for Zincke ketone formation, trifluoroacylation (performed at 0 °C), and ring closing. Arylation step: Zincke ketone (0.2 mmol), aryl halide (1.2 equiv.), Pd(OAc)₂ (10 mol%), K₂CO₃ (2 equiv.), THF (0.1 M), 100 °C, overnight, followed by work-up and chromatography.

We were interested in extending this sequential pyridine functionalisation to C–C bond formation, using our previously developed C4-selective Heck-arylation method.⁹ We found that the trifluoroacetylated Zincke ketone 16b (Table 1) performed poorly under the Pd-catalysed C4 arylation conditions, with the trifluoroacetyl group hindering the Heck arylation step. Altering the reaction sequence to begin with the Heck reaction, however, was productive (Scheme 6), with the arylated Zincke ketone being purified by chromatography. Subsequent trifluoroacetylation tolerated the flanking aryl substituent at C4, enabling in situ ring closure with NH₄OAc to access the 2,3,4-trisubstituted pyridines containing both aryl and heteroaryl groups (25a–d). X-Ray analysis of 25a and 25d confirmed the structures and showed that the three arene units are successively orthogonal to one another, projecting substituents into distinct areas of structural space.

Conclusions

In summary, we have developed a novel pyridine ring-opening/ring-closing skeletal rearrangement that enables a number of C2-functionalisations. The process introduces electron-poor groups such as CF₃, CF₂H, and CO₂Et to the 2-position exclusively using cheap and readily available acylating agents, complementing unselective Minisci-type radical approaches. The facility of Zincke ketones to undergo sequential functionalisation enables both 2,3,5- and 2,3,4-tri-substituted pyridines to be synthesised from 2-substituted pyridine starting materials, for both halogenation and arylation reaction classes, generating versatile di(hetero)arylmethane products. Further applications of this skeletal rearrangement are underway in our laboratory.

Author contributions

H. W., N. B. and M. F. G. conceived and designed the experiments. H. W. and N. B. performed the experiments and analysed the data. H. W., N. B. and M. F. G. wrote the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

Details on the procedures and the corresponding datasets reported in this study are provided in the paper and its supplementary information (SI) files, and are available from the corresponding author on request. Supplementary information: all preparative procedures, characterisation data, and NMR spectra. See DOI: https://doi.org/10.1039/d6sc04438g.

CCDC 2487297, 2515794, 2487303, 2515794, 2478546, 2515793, 2487299, 2475756, 2475750, 2475751 and 2475672 contain the supplementary crystallographic data for this paper.^27a–j

Acknowledgements

We thank EPSRC and the Chinese Scholarship Council for funding (iCAT CDT EP/S023755/1 studentship to NB, CSC scholarship to HW). Dr Avantika Hasija (University of Manchester) is thanked for X-ray crystallography.

Notes and references

1. O. Dimroth, Justus Liebigs Ann. Chem., 1909, 364, 183–226.
2. W. R. F. Goundry, K. Boardman, O. Cunningham, M. Evans, M. F. Jones, K. Millard, R. Rozada-Sanchez, Y. Sawyer, P. Siedlecki and B. Whitlock, Org. Proc. Res. Dev., 2017, 21, 336–345.
3. J. W. Cornforth, E. Fawaz, L. J. Goldsworthy and R. Robinson, J. Chem. Soc., 1949, 1549.
4. S. Huh, G. J. Saunders and A. K. Yudin, Angew. Chem., Int. Ed., 2023, 62, e202214729.
5. R. S. Sagitullin, S. P. Gromov and A. N. Kost, Tetrahedron, 1978, 34, 2213–2216.
6. T. Morofuji, S. Nagai, A. Watanabe, K. Inagawa and N. Kano, Chem. Sci., 2023, 14, 485–490.
7. T. Zincke, Justus Liebigs Ann. Chem., 1904, 330, 361–374.
8. B. T. Boyle, J. N. Levy, L. de Lescure, R. S. Paton and A. McNally, Science, 2022, 378, 773–779.
9. H. Wang and M. F. Greaney, Angew. Chem., Int. Ed., 2024, e202315418.
10. M. A. Hart, B. J. H. Uhlenbruck, J. N. Levy and A. McNally, J. Am. Chem. Soc., 2025, 147, 18406–18411.
11. S. Li, J. Tang, Y. Shi, M. Yan, Y. Fu, Z. Su, J. Xu, W. Xue, X. Zheng, Y. Ge, R. Li, H. Chen and H. Fu, Nat. Commun., 2024, 15, 7420.
12. K. F. Szabó, P. Banachowicz, A. Powała, D. Lunic, I. F. Ardoiz and D. Gryko, Nat. Commun., 2025, 16, 5072.
13. W. C. Hartley, A. Huerta, J. C. Negredo and O. Boutureira, Org. Chem. Front., 2025, 10, 1039D5QO00533G.
14. Z. Liu and M. F. Greaney, Angew. Chem., Int. Ed., 2025, 64, e202512321.
15. Y. Shi, M. Yan, D. Yang, J. Zhao, X. Jin, S. Li, Y. Ge, W. Xue, J. Xu, X. Zheng, R. Li, H. Chen and H. Fu, Org. Lett., 2025, 27, 6403–6408.
16. L. Qi, Z. Chen, C. Aïssa and J. Xiao, Chem.–Eur. J., 2025, 31, e202404043.
17. B. J. H. Uhlenbruck, C. M. Josephitis, L. De Lescure, R. S. Paton and A. McNally, Nature, 2024, 631, 87–93.
18. A. Conboy and M. F. Greaney, Chem, 2024, 10, 1940–1949.
19. M. Yan, Y. Shi, C. Lv, S. Huang, S. Li, D. Yang, J. Zhao, Z. Su, W. Jiang, W. Xue, J. Xu, X. Zheng, R. Li, H. Chen and H. Fu, Nat. Commun., 2025, 16, 7133.
20. S. Li, Y. Shi, J. Tang, M. Yan, S. Huang, T. Dou, S. Wang, X. Jin, Z. Su, W. Jiang, J. Xu, X. Zheng, R. Li, H. Chen, W. Xue and H. Fu, Nat. Commun., 2025, 16, 7112.
21. Selected recent examples of pyridine functionalisation: (a) B. Ramesh and M. Jeganmohan, Org. Lett., 2017, 19, 6000–6003; (b) L. Zhang, J. Yan, D. Ahmadli, Z. Wang and T. Ritter, J. Am. Chem. Soc., 2023, 145, 20182–20188; (c) C. Li, Z. Yan, B. Wang, J. Li, W. Lyu, Z. Wang, N. Jiao and S. Song, Chem, 2024, 10, 628–643; (d) C.-Y. Cai, S.-J. Chen, R. R. Merchant, Y. Kanda and T. Qin, J. Am. Chem. Soc., 2024, 146, 24257–24264; (e) Z.-C. Li, D. Tian, Z.-Q. Wang, P. Yuan and H. Wu, J. Am. Chem. Soc., 2025, 147, 30050–30059; (f) Z.-H. Chen, L. Liu, Y.-B. Wang, H. Luo, Z.-L. Tang, X.-Y. Zhou and X.-C. Wang, J. Am. Chem. Soc., 2025, 147, 36882–36889; (g) C. Li, X. Li, J. Li, Z. Wang, D. Ouyang, N. Jiao and S. Song, Nat. Synth, 2025, 5, 36–45; (h) J. Zhang, G. Wu, Y. Zhao, X. Liu, H. Zhu, A. Studer, X. Qi and Q. Cheng, Nat. Synth, 2025, 5, 27–35.
22. Y. Kuninobu, Chem. Rec., 2023, 23, e202300003.
23. P. Banachowicz, A. Powała, K. Szabó, O. Danylyuk and D. Gryko, Org. Lett., 2025, 27, 13350–13355.
24. For direct pyridine meta-formylation via a dearomatizing approach, see: P. Xu, S.-M. Guo and A. Studer, CCS Chem., 2025, 7, 2996–3004.
25. S. Vaas, M. O. Zimmermann, D. Schollmeyer, J. Stahlecker, M. U. Engelhardt, J. Rheinganz, B. Drotleff, M. Olfert, M. Lämmerhofer, M. Kramer, T. Stehle and F. M. Boeckler, J. Med. Chem., 2023, 66, 10202–10225.
26. Y. Fujiwara, J. A. Dixon, R. A. Rodriguez, R. D. Baxter, D. D. Dixon, M. R. Collins, D. G. Blackmond and P. S. Baran, J. Am. Chem. Soc., 2012, 134, 1494–1497.
27. (a) CCDC 2487297: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2ph7dv; (b) CCDC 2515794: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2qfwnq; (c) CCDC 2487303: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2ph7l1; (d) CCDC 2478546: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2p6435; (e) CCDC 2515793: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2qfwmp; (f) CCDC 2487299: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2ph7gx; (g) CCDC 2475756: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2p3735; (h) CCDC 2475750: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2p36xy; (i) CCDC 2475751: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2p36yz; (j) CCDC 2475672: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2p34dc.

---

Footnote

† These authors were contributed equally.

---