Abstract
The copper-catalyzed H–F insertion into α-diazocarbonyl compounds is described using potassium fluoride (KF) and hexafluoroisopropanol. Access to complex α-fluorocarbonyl derivatives is achieved under mild conditions, and the method is readily adapted to radiofluorination with [18F]KF. This late-stage strategy provides an attractive route to 18F-labeled biomolecules.
Methods for the construction of C–F bonds are of high synthetic value due to the broad applications of organo-fluorine compounds as pharmaceuticals, agrochemicals, and materials.1 Additionally, strategies to incorporate fluorine-18 into organic molecules are important for the synthesis of radiotracers for positron emission tomography (PET), a minimally invasive molecular imaging technology used in diagnostic medicine and clinical pharmaceutical research.2 The demand for increasingly complex radiotracers has stimulated the development of late-stage fluorination strategies.3 Although PET applications are frequently cited as a rationale for the identification of these new fluorination techniques, the successful translation of contemporary fluorination approaches to practical and valuable radiofluorination protocols remains a critical challenge.4 In addition to the inherent difficulties associated with forming C–F bonds,5 several unique obstacles must be addressed to translate a fluorination reaction to the 18F-radionuclide counterpart. Methods must use fluoride as a limiting reagent and take place rapidly, owing to the low available concentrations (nM to μM) of the short-lived 18F-radioisotope (half-life = 110 min).2 Additionally, the radiochemical protocol should proceed with high specific activity and be amenable to automation or use by nonspecialists.
To address these challenges, we sought to design a new catalytic method for the synthesis of α-fluorocarbonyl compounds that would enable late-stage fluorination and translate readily to radiosynthesis. Among medicinally relevant fluorinated molecules, the α-fluorocarbonyl motif is prevalent and serves as a versatile building block for the construction of numerous aliphatic fluorine-containing structures (Figure 1A).5 Many highly selective catalytic methods have been identified for the synthesis of α-fluorocarbonyl derivatives using electrophilic fluorine sources.3,6 However, these strategies are not well suited for radiofluorination in a clinical setting because electrophilic fluorination reagents generally suffer from low specific activity.7 The α-[18F]fluorocarbonyl scaffold is featured heavily in PET tracers, in part as a tag for labeling biomolecules such as peptides, which are desirable as PET probes due to their binding specificity.8 Typically, C–18F bonds are formed by nucleophilic substitution with [18F]KF in the presence of phase transfer reagent Kryptofix 2.2.2 (K222). While this approach takes advantage of nucleophilic fluoride as a readily available, high specific activity source of fluorine-18, the harsh conditions necessary for substitution (high temperatures and basicity) are not suitable for the direct radiofluorination of most α-[18F]fluorocarbonyl targets.9 Instead, an indirect method is pursued wherein fluorine-18 is first introduced into a reactive precursor, such as activated ester [18F]NFP, and then into the compound of interest (Figure 1B).2b This prosthetic group strategy necessitates multiple radiochemical steps, highlighting the limitations of classic nucleophilic substitution with [18F]KF. Identification of catalytic schemes for α-fluorocarbonyl synthesis that use nucleophilic fluoride, tolerate biologically relevant functionality, and are amenable to radiofluorination would enable the radiosynthesis of previously challenging or inaccessible PET tracers in a single step. Furthermore, such methods could prove complementary in scope to known electrophilic fluorination reactions.
Figure 1.

(A) α-Fluorocarbonyl-containing bioactive targets. (B) Multistep assembly of complex radiotracers using an α-fluoroacyl prosthetic group. (C) Catalytic (radio)fluorination of α-diazocarbonyl compounds.
Here we report the development of a copper-catalyzed fluorination of α-diazocarbonyl compounds with KF (Figure 1C). This reaction facilitates the direct installation of fluorine-19 and fluorine-18 into advanced substrates under mild and functional-group-tolerant conditions. We chose to investigate α-diazocarbonyl compounds as precursors to α-fluorocarbonyl derivatives since the diazo functionality is known to react rapidly with transition metals to form electrophilic metal carbenoids that can insert into carbon–hydrogen, carbon–heteroatom, and heteroatom–hydrogen bonds under mild conditions.10 Although this approach has become a powerful tool to forge new C–X (X = C, Si, N, O, S) bonds regio- and stereoselectively,11 the catalytic insertion into halogen–hydrogen bonds has not been reported. The Davies group has recently described a silver-catalyzed fluorination of vinyl diazoacetates with Et3N·3HF, but this transformation yields γ-fluoro-α,β-unsaturated carbonyl compounds rather than fluorinating at the carbenoid position.12 Direct fluorination of α-diazocarbonyl substrates may be accomplished with highly reactive and acidic HF-containing reagents.13–16 Only simple, unfunctionalized substrates are tolerated, however, and translation to PET applications is challenging.17
As an alternative to using stoichiometric HF or HF-containing reagents, we envisioned accessing the reactive equivalent from the combination of a mild fluoride reagent and proton source. Initial experiments were conducted using diazoacetate 1a as a model substrate with a copper(I) catalyst and a commercially available bis(oxazoline) ligand L1 (Table 1). Previously, our laboratory reported the enantioselective hydrofluorination of epoxides and aziridines using benzoyl fluoride (PhCOF) and 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) as a latent HF source.18 For the HF insertion chemistry, this combination yielded the desired α-fluoroester 2a (entry 2); however, since [18F]PhCOF is not readily accessible, we evaluated other reagents that offer more convenient sources of fluorine-18. Promisingly, the combination of KF and HFIP provided the desired product in 68% yield (entry1).
Table 1.
Reaction Optimization
| |||
|---|---|---|---|
| entry | conditions | Yield 2aa (%) | Yield 3a (%) |
| 1 | standard conditions | 68b | 2 |
| 2 | PhCOF (2 equiv), HFIP (4 equiv), DBN (20 mol%) instead of KF/HFIP | 64c | 1 |
| 3 | no Cu or ligand | 1 | 4 |
| 4 | no HFIP | 0 | n/a |
| 5 | TFE instead of HFIP | 22 | n/a |
| 6 | phenol instead of HFIP | 7 | n/a |
| 7 | HFIP (3 equiv) | 39 | 35 |
| 8 | HFIP (10 equiv) | 64 | 5 |
| 9 | ligand L2 instead of L1 | 35d | 15 |
| 10 | ligand L3 instead of L1 | 57 | 10 |
| 11 | Rh2(OAc)4 instead of Cu/ligand | 53 | 4 |
| 12 | [Cu(OTf)]2•PhMe instead of [Cu(MeCN)4]PF6 | 62e | 2 |
Determined by 19F NMR, 0.1 mmol scale using fluorobenzene as an external standard.
9% ee, determined by chiral HPLC.
15% yield in the absence of Cu and ligand.
23% ee.
31% ee.
Interestingly, HFIP is necessary for productive chemistry (entry 4). Deuterium-labeling experiments indicate the OH proton of HFIP is incorporated into the product; however, other proton sources with similar pKas are inadequate substitutes (entries 5 and 6), suggesting an additional role for HFIP. Although HFIP is considered a poor nucleophile,19 competitive O–H insertion (3) is observed as the major side product, along with dimerization of the diazo substrate.20 Counterintuitively, the yield and selectivity of the reaction are significantly improved by increasing the equivalents of HFIP (entries 7 and 8). We hypothesize HFIP solvates the otherwise insoluble KF and increasing HFIP concentration alters the solvent coordination environment of fluoride, leading to a more reactive species. This finding appears related to the influence of t-BuOH on reactions with TBAF21 and may prove useful for enhancing the reactivity of other fluorination reactions using KF. The reaction is less efficient with other commercial bis(oxazoline) ligands such as L2 and achiral ligand L3 (entries 9 and 10). Evaluation of other transition metal catalysts indicated that Cu(I) is most effective, as Rh2(OAc)4 delivered 2a in modest yield (entry 11).22 Although the optimal conditions with [Cu(MeCN)4]PF6 and bis(oxazoline) L1 afford 2a in low ee, modest asymmetric induction is achieved using [Cu(OTf)]2·PhMe (31% ee, entry 12), demonstrating for the first time that enantioselective halogen–hydrogen insertion is feasible using transition metal catalysis.
With the optimized conditions in hand, we next explored the scope of this transformation and found that a variety of aryldiazoacetate substrates perform well (Table 2). Substitution around the aromatic ring is well tolerated (2b–d), including at the ortho position, which suppresses substrate dimerization. High yields of the desired products containing σ- and π-electron-withdrawing substituents (2e–k) are obtained, albeit with slightly longer reaction times or higher temperatures (2g–i). Electron-rich α-diazocarbonyl substrates are less effective under the reaction conditions; nevertheless, fluorinated products 2l, 2m, and 2o are furnished in 30%, 46%, and 72% yields, respectively. Notably, many valuable functional groups are tolerated in the reaction, including azides (2k), silyl-protected alcohols (2n), and Lewis basic functionality (2o, 2p). These results are significant as many of these common functional groups are not compatible with current electrophilic α-fluorination methods. Furthermore, this straightforward protocol is readily scalable, as 2i was isolated in 89% yield in a multigram-scale reaction set up on the benchtop.22
Table 2.
Scope of Fluorinationa
|
Yields are the average of two runs, 0.18–0.5 mmol scale.
Reaction run at 50 °C for 24 h.
Reaction run at 45 °C.
Reaction run at rt.
The reaction is also generalizable to other α-diazocarbonyl precursors, including terminal diazoacetates (2q), diazoketones (2r), and diazoacetamides (2s–2y). Acceptor/acceptor α-diazoester compounds, however, are unreactive,16b and α-alkyl-α-diazoacetates predominately undergo elimination to afford α,β-unsaturated esters with only trace amounts of the desired fluorinated products.22 To probe the scope of the reaction further, we investigated the fluorination of α-diazo substrates derived from biologically relevant structures. Importantly, this approach provides access to fluorinated amino acid derivatives (2v, 2w), peptides (2x), and glycosides (2y) bearing unprotected, protic amines (2u) and alcohols (2w). The exceptional functional group tolerance of these examples highlights the mildness of the reaction conditions and the potential for late-stage fluorination
We then sought to demonstrate the utility of this transformation by identifying a practical radiochemical protocol. By designing our reaction with KF, we envisioned this method would be readily adaptable to the direct, late-stage radiofluorination of α-diazocarbonyl compounds simply by applying [18F]KF and K222 to our standard reaction conditions. Typically, radiochemical protocols differ substantially from their synthetic counterparts, and most syntheses employing [18F]KF require high temperatures (>100 °C) for reactivity. Therefore, we were pleased to find substrate 1a underwent radiofluorination in excellent radiochemical conversion (80% RCC) using the combination of [18F]KF-K222 and HFIP at 50 °C (Table 3). Control experiments without copper or HFIP delivered no significant radiochemical conversion,22 and the reaction conditions are remarkably similar to the 19F-protocol.
Table 3.
Radiofluorination Using [18F]KFa
|
RCC determined by radio-TLC and corrected for insoluble activity; identity of product confirmed by radio-HPLC.
Specific activity: 24 mCi/μmol EOB.
Specific activity: 1300 mCi/μmol EOB.
Under these standard conditions, however, [18F]2a was obtained in poor specific activity (24 mCi/μmol), suggesting at least one reagent other than KF contains labile fluoride. Consistent with previous reports,23 we found that HFIP is not a significant fluoride source; yet, in the absence of KF, the hexafluorophosphate anion of the copper catalyst generates small amounts of product.22 Replacing hexafluorophosphate with a perchlorate anion restores specific activity to 1300 mCi/μmol, a value consistent with other no-carrier-added radiofluorinations (Catalyst B in Table 3). Thus, for applications requiring high specific activity, [Cu(MeCN)4]ClO4 should be used in place of [Cu(MeCN)4]PF6.
Based on these results and with insights gained from the 19F-catalytic system, we interrogated the generality of the α-radiofluorination. Complex α-diazocarbonyl substrates were subjected to the procedure with [18F]KF-K222, and trends in the functional group tolerance of the 19F-conditions were recapitulated in the radiofluorination (Table 3). Diazo esters and amides are competent in the reaction, and Lewis basic ([18F]2o, [18F]2t) as well as protic functionality on peptides ([18F]2x) and glycosides ([18F]2y) is compatible. Moreover, known radiotracer N5-[18F]FAO ([18F]2v)24 is readily accessible using this practical approach, as is the homologue of PET probe [18F]FPDA ([18F] 2t).25 Notably, the one-step copper-catalyzed radiofluorination is significantly higher yielding than the synthetic sequences previously reported in the literature: whereas [18F]FPDA required installation of a prosthetic group over four radiochemical steps with 3% overall RCY,25 [18F]2t can be generated in 45% RCC in a single step. Likewise, N5-[18F]FAO, which was previously prepared by SN2 displacement of bromide with 8% RCY,24 can be labeled under our conditions in 39% RCC. This method establishes a strategic alternative for the radiofluorination of complex biomolecules via their diazoacetylated precursors.26
This work introduces a new method for catalytic C–F/18F bond formation enabled by copper catalysis and the use of KF-HFIP as a reactive fluoride source. The approach generates useful α-fluorocarbonyl derivatives under mild reaction conditions and is compatible with a variety of valuable functional groups. Furthermore, by considering translation to fluorine-18 in our design and employing commonly used [18F]KF-K222, we are able to effect the one-step radiofluorination of biomolecules under operationally convenient conditions.
Supplementary Material
ACKNOWLEDGMENTS
We thank Dr. Eric Hostetler (Merck & Co., Inc., West Point, PA) for providing access to radiosynthesis facilities and Ivanny Jacome Ottati for substrate synthesis. Financial support from Janssen Research & Development LLC and NSF (CHE-1565983) is gratefully acknowledged. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-1148900 (E.E.G.) and a F32 Ruth L. Kirschtein NRSA Fellowship under Award No. GM108291 (K.A.C.).
Footnotes
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.6b06770.
Experimental procedures, additional reaction optimization, and spectroscopic data for all new compounds (PDF)
The authors declare no competing financial interest.
REFERENCES
- (1).(a) Gillis EP; Eastman KJ; Hill MD; Donnelly DJ; Meanwell NA J. Med. Chem 2015, 58, 8315. [DOI] [PubMed] [Google Scholar]; (b) Jeschke P ChemBioChem 2004, 5, 570. [DOI] [PubMed] [Google Scholar]
- (2).(a) Cai L; Lu S; Pike VW Eur. J. Org. Chem 2008, 2008, 2853. [Google Scholar]; (b) Miller PW; Long NJ; Vilar R; Gee AD Angew. Chem., Int. Ed 2008, 47, 8998. [DOI] [PubMed] [Google Scholar]
- (3).For recent reviews, see:; (a) Champagne PA; Desroches J; Hamel J-D; Vandamme M; Paquin J-F Chem. Rev 2015, 115, 9073. [DOI] [PubMed] [Google Scholar]; (b) Yang X; Wu T; Phipps RJ; Toste FD Chem. Rev 2015, 115, 826. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Liang T; Neumann CN; Ritter T Angew. Chem., Int. Ed 2013, 52, 8214. [DOI] [PubMed] [Google Scholar]
- (4).(a) Neumann CN; Ritter T Angew. Chem., Int. Ed 2015, 54, 3216. [DOI] [PubMed] [Google Scholar]; (b) Brooks AF; Topczewski JJ; Ichiishi N; Sanford MS; Scott PJH Chem. Sci 2014, 5, 4545. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (5).(a) O’Hagan D Chem. Soc. Rev 2008, 37, 308.18197347 [Google Scholar]; (b) Kirsch P Modern Fluoroorganic Chemistry; Wiley-VCH: Weinheim, 2004. [Google Scholar]
- (6).For an example of α-fluorocarbonyl radiosynthesis using [18F] NFSI, see:; Buckingham F; Kirjavainen AK; Forsback S; Krzyczmonik A; Keller T; Newington IM; Glaser M; Luthra SK; Solin O; Gouverneur V Angew. Chem., Int. Ed 2015, 54, 13366. [DOI] [PubMed] [Google Scholar]
- (7).For a notable example of the high specific activity synthesis of [18F] Selectfluor, see:; Teare H; Robins EG; Kirjavainen A; Forsback S; Sandford G; Solin O; Luthra SK; Gouverneur V Angew. Chem., Int. Ed 2010, 49, 6821. [DOI] [PubMed] [Google Scholar]
- (8).(a) Richter S; Wuest F Molecules 2014, 19, 20536. [DOI] [PMC free article] [PubMed] [Google Scholar]; (b) Cai H; Conti PS J. Labelled Compd. Radiopharm 2013, 56, 264. [DOI] [PubMed] [Google Scholar]
- (9).Liu S; Shen B; Chin FT; Cheng Z Curr. Org. Synth 2011, 8, 584. [Google Scholar]
- (10).Ford A; Miel H; Ring A; Slattery CN; Maguire AR; McKervey MA Chem. Rev 2015, 115, 9981. [DOI] [PubMed] [Google Scholar]
- (11).For leading X–H insertion reviews, see:; (a) Gillingham D; Fei N Chem. Soc. Rev 2013, 42, 4918. [DOI] [PubMed] [Google Scholar]; (b) Zhao X; Zhang Y; Wang J Chem. Commun 2012, 48, 10162. [DOI] [PubMed] [Google Scholar]; (c) Zhu S-F; Zhou Q-L Acc. Chem. Res 2012, 45, 1365. [DOI] [PubMed] [Google Scholar]
- (12).Qin C; Davies HML Org. Lett 2013, 15, 6152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (13).Curtius TJ Prakt. Chem 1888, 38, 396. [Google Scholar]
- (14).(a) Olah G; Kuhn S Chem. Ber 1956, 89, 864. [Google Scholar]; (b) Ynunyants IL; Kisel YM; Bykhovskaya EG Bull. Acad. Sci. USSR, Div. Chem. Sci 1956, 5, 363. [Google Scholar]
- (15).(a) Olah GA; Welch J Synthesis 1974, 1974, 896. [Google Scholar]; (b) Setti EL; Mascaretti OAJ Chem. Soc., Perkin Trans 1 1988, 2059. [Google Scholar]
- (16).(a) Ohno M; Itoh M; Ohashi T; Eguchi S Synthesis 1993, 1993, 793. [Google Scholar]; (b) Pasceri R; Bartrum HE; Hayes CJ; Moody CJ Chem. Commun 2012, 48, 12077. [DOI] [PubMed] [Google Scholar]
- (17).For an example of radiofluorination of trifluoroethyldiazo compounds, see:; Emer E; Twilton J; Tredwell M; Calderwood S; Collier TL; Liégault B; Taillefer M; Gouverneur V Org. Lett 2014, 16, 6004. [DOI] [PubMed] [Google Scholar]
- (18).(a) Kalow JA; Doyle AG J. Am. Chem. Soc 2010, 132, 3268. [DOI] [PubMed] [Google Scholar]; (b) Kalow JA; Doyle AG J. Am. Chem. Soc 2011, 133, 16001. [DOI] [PubMed] [Google Scholar]; (c) Kalow JA; Schmitt DE; Doyle AG J. Org. Chem 2012, 77, 4177. [DOI] [PubMed] [Google Scholar]; (d) Kalow JA; Doyle AG Tetrahedron 2013, 69, 5702. [Google Scholar]
- (19).Azzouzi-Zriba K; Bonnet-Delpon D; Crousse BJ Fluorine Chem 2011, 132, 811. [DOI] [PubMed] [Google Scholar]
- (20).Mass balance consists of a maleate dimer, a fumarate dimer, and an azine dimer. Trace methyl mandelate, resulting from H2O insertion, is also observed, but this side product can be eliminated by thoroughly drying reagents.
- (21).Engle KM; Pfeifer L; Pidgeon GW; Giuffredi GT; Thompson AL; Paton RS; Brown JM; Gouverneur V Chem. Sci 2015, 6, 5293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (22). See SI for details.
- (23).Revunov E; Zhuravlev FJ Fluorine Chem 2013, 156, 130. [Google Scholar]
- (24).Turkman N; Gelovani JG; Alauddin MM J. Labelled Compd. Radiopharm 2011, 54, 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (25).Liu H; Liu S; Miao Z; Jiang H; Deng Z; Hong X; Cheng Z Mol. Pharmaceutics 2013, 10, 3384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (26).Josa-Culleré L; Wainman YA; Brindle KM; Leeper FJ RSC Adv 2014, 4, 52241. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
