NHC-Copper Mediated Ligand-Directed Radiofluorination of Aryl Halides

Abstract

[¹⁸F]-labeled aryl fluorides are widely used as radiotracers for positron emission tomography (PET) imaging. Aryl halides (ArX) are particularly attractive precursors to these radiotracers, as they are readily available, inexpensive, and stable. However, to date, the direct preparation of [¹⁸F]-aryl fluorides from aryl halides remains limited to S_NAr reactions between highly activated ArX substrates and K¹⁸F. This report describes an aryl halide radiofluorination reaction in which the C(sp²)–¹⁸F bond is formed via a copper-mediated pathway. Copper N-heterocyclic carbene complexes serve as mediators for this transformation, using aryl halide substrates with directing groups at the ortho position. This reaction is applied to the radiofluorination of electronically diverse aryl halide derivatives, including the bioactive molecules vismodegib and PH-089.

Graphical Abstract

Late-stage methods for constructing ¹⁸F–(hetero)aryl bonds are highly valued for the synthesis of positron emission tomography (PET) radiotracers.^1,2 Historically, ¹⁸F-labeled aromatic substrates have most commonly been prepared via S_NAr reactions between electron deficient aryl halide precursors and K¹⁸F (Scheme 1A).^3,4 Aryl halides are particularly attractive radiofluorination precursors because they are abundant, stable, and synthetically accessible. However, the substrate scope of S_NAr (radio)fluorination reactions remains narrow, as resonance electron withdrawing substituents on the aromatic ring are required to stabilize Meisenheimer-type intermediates.^1,5 Furthermore, even with such highly activated substrates, S_NAr pathways often require long reaction times and forcing conditions, which renders them ill-suited for many late-stage radiofluorination applications.^6,7 As such, a key objective for the field is to develop complementary methods for the radiofluorination of (hetero)aryl–halides and pseudohalides.^8,9

Scheme 1.

Strategies for direct fluorination of aryl halides.

Our approach to this challenge has focused on developing Cu-mediated methods for C(sp²)–¹⁸F coupling reactions.^1,2b Recent studies have shown that Cu salts such as Cu(OTf)₂ and Cu(CH₃CN)₄PF₆ mediate the nucleophilic radiofluorination of aryl stannane,¹⁰ aryl boron,¹¹ diaryliodonium,¹² and aryl C–H substrates¹³ with K¹⁸F. In these systems the key C(sp²)–¹⁸F bond is formed via reductive elimination from an organometallic Cu(aryl)(¹⁸F-fluoride) intermediate.^11f,14 This organometallic pathway is mechanistically distinct from an S_NAr reaction. As such, it enables the radiofluorination of a wide scope of electronically diverse aryl groups.

Despite this progress, analogous Cu mediators have proven ineffective at engaging aryl halide substrates in radiofluorination reactions. Two reports have documented the Cu-promoted nucleophilic ¹⁹F-fluorination of aryl halides (e.g., the work of Liu in Scheme 1B). However, both require superstoichiometric AgF as the fluoride source,^15,16 and neither has proven translatable to radiolabeling with ¹⁸F^– (vide infra). This report describes the use of N-heterocyclic carbene (NHC) Cu complexes as mediators for ligand-directed aryl halide radiofluorination (Scheme 1C). The discovery of this transformation in the context of ¹⁹F-fluorination and its subsequent translation to radiofluorination are described in detail.

Our initial studies attempted to translate Liu’s ¹⁹F-fluorination of 2-(2-bromophenyl)pyridine (Scheme 1B) to a radiolabeling protocol. However, as shown in eq. 1, under the standard conditions (with Cu^I(CH₃CN)₄PF₆, Ag¹⁸F, and NBu₄PF₆ in CH₃CN at 120 °C), no trace of product 1-¹⁸F was detected by radio-TLC or radio-HPLC after 0.5 h. Furthermore, no improvement was observed upon variation of the ¹⁸F source, solvent, additives, or temperature (Table S7). We note that, in contrast to the ¹⁹F-fluorination, the radiofluorination reaction requires the use of Ag¹⁸F as the limiting reagent at sub-micromolar concentrations. We hypothesize that this renders Cu^I(CH₃CN)₄PF₆-mediated radiofluorination prohibitively slow relative to the decay of the radionuclide (t_1/2 ~110 min).

Literature reports suggest that aryl-bromide bond activation (via oxidative addition at Cu^I) is likely the slow step in this transformation.^16,17 We reasoned that the introduction of a strongly electron donating NHC ligand at the Cu^I center would accelerate this key step.^18,19 Furthermore, since (NHC)Cu^I(F) complexes can be generated directly from KF,²⁰ this approach should eliminate the requirement for excess AgF. Finally, sterically bulky NHC ligands are known to stabilize Cu^I–fluoride complexes to dimerization or disproportionation,^19,21 which are likely competing decomposition pathways for the Cu mediator.²²

To test this hypothesis, we initially examined the reactivity of a series of (NHC)Cu^I(¹⁹F) complexes with 2-(2-bromophenyl)pyridine (Scheme 2A). As summarized in Table S3, the yield of fluorinated product 1-¹⁹F varied from 3–65% as a function of the structure of the NHC ligand,^19,23 with 1,3-bis-(2,6-diisopropylphenyl)imidazol-2-ylidine (IPr) affording the optimal result. Notably, (IPr)Cu^I(¹⁹F) (A-¹⁹F) is available in nearly quantitative yield from the reaction of (IPr)Cu^I(OTf) (A-OTf) with K¹⁹F (Scheme 2A),²⁰ thus precluding the requirement for Ag salts in this transformation. Importantly, control studies revealed that other group 11 metal salts including Cu^I(CH₃CN)₄PF₆/KF, CuF₂,²⁴ or AgF afforded ≤3% of 1-¹⁹F under otherwise identical conditions (Table S4). Furthermore, no reaction was observed between the aryl bromide substrate and K¹⁹F under these conditions in the absence of copper.

Scheme 2.

NHC-Cu-mediated 19F-fluorination of aryl bromides

(A) Conditions: A-OTf (0.006 mmol, 1 equiv), KF (1.5 equiv), DMF (0.01 M), 140 °C for 30 min, then aryl bromide (0.006 mmol), 140 °C for 21 h. (B) Conditions: A-¹⁹F (0.01 mmol, 1 equiv), aryl bromide (1 equiv), DMF (0.015 M), 140 °C for 21 h. (C) Conditions: A-¹⁹F (0.006 mmol, 1 equiv), aryl bromide (1 equiv), DMF (0.01 M), 140 °C for 21 h. Yields determined by ¹⁹F NMR spectroscopic analysis of crude reaction mixtures.

A time study with A-¹⁹F shows that the fluorination reaction is complete within 2 h at 140 °C and affords 40% yield after just 30 min (Scheme 2B). This suggests the feasibility of achieving radiofluorination with this system. Finally, a preliminary survey of substrates revealed that A-¹⁹F-mediated fluorination has a significantly enhanced scope versus that of Liu’s Cu^I(CH₃CN)₄PF₆/Ag¹⁹F system (Scheme 1B). For instance, the sterically hindered pyridine substrate 2-(2-(bromo)phenyl)-6-methylpyridine was unreactive under Liu’s conditions, but affords 2-¹⁹F in 34% yield with A-¹⁹F as the Cu mediator (Scheme 2C). Similarly, the oxazoline and imine substrates were unreactive under Liu’s conditions, but afford 30% and 37% yield of 3-¹⁹F and 4-¹⁹F, respectively, using A-¹⁹F.²⁵

We next focused on translating these preliminary results to radiofluorination. The reaction of (IPr)Cu^I(OTf) (A-OTf) with 2-(2-bromophenyl)pyridine and K¹⁸F for 30 min at 140 °C in DMF afforded 1-¹⁸F in 10% radiochemical conversion (RCC) as determined by radio-TLC and radio-HPLC (Table 1, entry 1).^26,27 The reaction was optimized by exploring additives that have been shown to enhance yields in other Cu-mediated C(sp²)–¹⁸F coupling reactions (e.g., phase transfer reagents, nitrogen heterocycles, Table 1, entries 2–5).^1,11a,13,28 Of the surveyed additives, 1 equiv of 4-dimethylaminopyridine (DMAP) relative to the aryl bromide precursor proved optimal, affording 1-¹⁸F in 65% RCC.

Table 1.

Cu-mediated radiofluorination of aryl halides.

entry	Additive	RCC (%)
1	none	10
2	Kryptofix	26
3	pyridine	23
4	DBU	30
5	DMAP	65

Conditions: aryl bromide (0.005 mmol, 1 equiv), A-OTf(1 equiv), additive (1 equiv), K¹⁸F, DMF (0.015 M), N₂ atmosphere, 140 °C, 30 min.²⁷ RCC determined by radio-TLC (n ≥ 2).

With these optimized conditions in hand, we next explored the scope of the A-OTf-mediated radiofluorination of aryl halides. As shown in Figure 1, the chloro-, bromo-, and iodo-2-phenylpyridine precursors all reacted to afford 1-¹⁸F in RCCs ranging from 10–65%. In contrast, no ¹⁹F/¹⁸F exchange was detected with 1-¹⁹F under these conditions. It is currently unclear why 1-I affords lower yield than 1-Br; however, this observation is in line with Liu’s results for the Cu-catalyzed [¹⁹F]-fluorination of halophenylpyridines.¹⁶ Substitution on either the pyridine or aryl ring was tolerated to afford products such as 2-¹⁸F, 6-¹⁸F, and 7-¹⁸F. Other nitrogen-donors, including oxazoline, pyrazole, cyclohexyl imine, and mesityl imine, served as effective directing groups, affording 3-¹⁸F, 8-¹⁸F, 4-¹⁸F, and 9-¹⁸F, respectively. The scope of cyclohexyl imine derivatives was most thoroughly explored, as this directing group is straightforward to install and remove starting from readily available benzyaldehyde derivatives. Various substitution patterns on the (hetero)arene ring were well tolerated, affording compounds 10–16-¹⁸F in RCCs ranging from 16–74%. An intramolecular competition reaction between an ortho-chloride and bromide resulted in selective radiofluorination of the bromide to form 13-¹⁸F. This selectivity is consistent with that expected for a metal-mediated activation of a C(sp²)–X bond.²⁹

Figure 1.

Substrate scope of Cu-mediated radiofluorination of aryl halides.

Conditions: aryl halide (0.005 mmol, 1 equiv), A-OTf (1 equiv), DMAP (1 equiv), K¹⁸F, DMF (0.015 M), N₂ atmosphere, 140 °C for 30 min.²⁷ RCC determined by radio-TLC (n ≥ 3). ^aReaction conducted at 160 °C.

Importantly, a variety of control reactions were conducted in these systems. First, the 4-substituted aryl bromides in the pyridine and cyclohexyl imine series were subjected to the reaction conditions. These are electronically similar, but do not benefit from the directing effect. As shown in Figure 1, these substrates did not afford detectable 17-¹⁸F or 18-¹⁸F under the optimized conditions.³⁰ In addition, all of these reactions were conducted in the absence of Cu to test for background S_NAr reactivity. As shown in Table S12, ≤1% of compounds 1–16-¹⁸F were detected under these conditions. Finally, substituting simple Cu^I or Cu^II salts for (IPr)Cu^I(OTf) afforded yields of ≤5% for representative substrates (Table S11), underscoring the central role of the NHC ligand in these transformations.

¹⁸F-analogues of several bioactive molecules could also be accessed using this approach. In a first example, the bromide analogue of vismodegib (19-Br), a basal cell carcinoma treatment,³¹ underwent radiofluorination to afford 19-¹⁸F (Scheme 3). In a second example, ¹⁸F-labeled PH-089 (20-¹⁸F in Scheme 3), an MK-2 inhibitor,³² was synthesized in 5% RCC from the chloride precursor.

Scheme 3.

Radiofluorination of bioactive molecules.

A final set of studies focused on automating the radiosynthesis of 1-¹⁸F using a TRACERLab FX_FN synthesis module. Initial automated studies using 241.1 mCi (8.93 × 10⁹ Bq) of K¹⁸F gave 57 ± 8 % radiochemical yield (RCY; n = 2), demonstrating the compatibility of the method with automation. Further investigations coupled automated synthesis with semi-preparative HPLC purification to afford 1-¹⁸F in 14.3 ± 3.2% RCY (decay-corrected; 119.9 mCi ± 28; n = 2) with good molar activity (1614 ± 353 Ci/mmol; n = 2) and radiochemical purity. While unoptimized, this result demonstrates the potential of this method for PET applications.

In conclusion, we have developed a Cu-mediated protocol for the ¹⁹F- and ¹⁸F-fluorination of diverse aryl halide substrates. Strategic design of the Cu mediator was necessary to achieve the reaction rates/yields required for efficient radiofluorination, and an NHC-ligated Cu complex ultimately proved optimal in this system. A wide scope of nitrogen-containing directing groups and substituted aryl halide derivatives underwent ¹⁸F-fluorination, and the reaction proved effective for the synthesis of biologically relevant molecules such as 19-¹⁸F and 20-¹⁸F. More broadly, this work demonstrates that NHC-type ligands enable new C(sp²)–F coupling reactions at Cu. As such, this work opens up opportunities for designing next-generation Cu mediators for the radiofluorination of currently inert substrates (e.g., aryl halides that lack a directing group).