Copper(I)-free Syntheses of [¹¹C/¹⁸F]Trifluoromethyl Ketones from Alkyl- or Aryl Esters and [¹¹C/¹⁸F]Fluoroform

Abstract

Herein, we report a copper(I)-free method for labeling the trifluoroacetyl group with positron-emitting carbon-11 (t_1/2 = 20.4 min) or fluorine-18 (t_1/2 = 109.8 min) as part of our exploration of radiolabeled fluoroforms to access new radiolabeled chemotypes of interest for tracer development. Treatment of alkyl esters and aryl esters, containing electron-donating or electron-withdrawing groups, with [¹¹C/¹⁸F]fluoroform in the presence of strong base, gave [¹¹C/¹⁸F]trifluoromethyl ketones as novel radiolabeling synthons in moderate to high yields within 15 minutes.

Graphical Abstract

The positron emitters, carbon-11 (t_1/2 = 20.4 min) and fluorine-18 (t_1/2 = 109.8 min), are highly important for labeling tracers for positron emission tomography (PET) – a biomedical imaging modality that is now indispensable for research,^1–2 drug development,^3–4 and medical diagnosis.⁵ Carbon-11 and fluorine-18 play complementary roles in PET because of the differences in their half-lives and radiochemical utility. Versatile radiochemistry has been developed for installing carbon-11 into tracers, predominantly at methyl and carbonyl positions,^6–8 whereas chemistry for creating carbon fluorine-18 bonds is now very advanced.^8–9 Nonetheless, there remains a great need and interest in expanding the chemotypes that can be labeled with these positron-emitters.

In drug candidates, a trifluoromethyl group (CF₃) may favorably enhance bioavailability, metabolic stability, and binding selectivity, and may also allow fine-tuning of the lipophilicity for pharmacokinetic benefit.^10–11 The trifluoroacetyl group (COCF₃) is found in many drug molecules (Figure 1)^12–14 and therefore the introduction of a trifluoroacetyl group has attracted considerable synthetic effort.¹⁵ In addition, trifluoromethyl ketones (TFMKs) are very useful synthons for the syntheses of important classes of compounds, such as trifluoromethyl alcohols and amines,^15–18 lactones,^19–21 and heterocycles.²²

Figure 1:

Examples of bioactive small molecules containing a trifluoroacetyl moiety.

[¹¹C]Fluoroform is readily synthesized²³ and converted into reactive [¹¹C]CF₃Cu(I), which has been utilized to produce [¹¹C]trifluoromethylarenes in high molar activities from aryl iodides, boronic acids, diazonium salts,^24–25 or diaryliodonium salts.²⁶ [¹¹C]Trifluoromethylstyrenes have also been produced from arylvinyliodonium salts.²⁷

Several methods have been developed to produce [¹⁸F]fluoroform.²⁸ We have developed a convenient method for gas phase [¹⁸F]fluoroform synthesis in good yield and molar activity.²⁹ As in the case of the copper(I) derivative of [¹¹C]fluoroform, the copper(I) derivative of [¹⁸F]fluoroform, [¹⁸F]CF₃Cu(I) has been used widely for labeling trifluoromethylarenes.²⁸ In this study, we aimed to explore the reactivity of no-carrier-added [¹¹C/¹⁸F]fluoroform in the absence of copper(I) and to expand the chemical space for producing new PET tracers.

Carboxylic acids and their derivatives have been converted readily into trifluoromethyl ketones¹⁵ by using various “CF₃” sources such as CF₃CO₂Na,³⁰ AgCF₃,³¹ TMSCF_3,^32–35 or fluoroform^36–37 (Scheme 1, Eq. 1). Despite these advances, only a few methods are reported for producing [¹¹C/¹⁸F]trifluoromethyl ketones. Prakash et al.³⁸ utilized difluoroenol silyl ethers and carrier-added [¹⁸F]fluorine (¹⁸F-F) in acetonitrile at −30 °C to afford ¹⁸F-labeled aryl trifluoromethyl ketones in low yields and very low molar activities (15–20 MBq/μmol) (Scheme 1, Eq. 2). Gomez et al.³⁹ radiofluorinated 2-bromo-2,2-difluoromethyl amides with [¹⁸F]n-Bu₄NF in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in DMF at 100 °C for 10 minutes to produce [¹⁸F]trifluoroacetamides, but obtained very low yields and low molar activities (8.2 GBq/μmol). More recently, Meyer et al.¹⁴ reported that radiofluorination of 2-halo-2,2-difluoromethyl ketones with [¹⁸F]n-Bu₄NF in the presence of 1,5,7-triazabicyclo-[4.4.0]dec-5-ene (TBD) in DMF at 75 °C for 10 minutes gave the [¹⁸F]aryltrifluoromethyl ketones in moderate to high yields but again with very low molar activities (0.41–1.27 GBq/μmol) (Scheme 1, Eq. 2). Here, we demonstrate the syntheses of both aryl and alkyl ¹¹C/¹⁸F-trifluoromethyl ketones by using no-carrier-added [¹¹C/¹⁸F]fluoroform and a suitable base under mild conditions (Scheme 1, Eq. 3).

Scheme 1:

Methods for synthesizing aryl and alkyl trifluoromethyl ketones.

To address our aims, we started by testing the ¹¹C-trifluoromethylation of methyl benzoate (1a, 50 μmol) with [¹¹C]fluoroform in the presence of KO^tBu (15 μmol) in DMF at RT for 10 minutes, followed by quench of the reaction with 12M HCl (0.1 mL) and heating at 60 °C for 5 minutes. To our delight, this reaction furnished [¹¹C]2,2,2-trifluoro-1-phenylethanone ([¹¹C]2a) in good yield (65%) (Table 1, entry 1). In an attempt to increase the yield, other bases such as P₁-t-Bu, BEMP, and KHMDS (15 μmol each) were investigated. All these bases were ineffective in producing [¹¹C]2a (Table 1, entries 2–4). Increasing the amount of KO^tBu to 45 μmol however increased the yield of [¹¹C]2a substantially to 83% (Table 1, entry 5). Yields were found to be rather insensitive to lower HCl concentrations down to 0.5M (Table 1, entries 6–8). P₁-t-Bu and BEMP (45 μmol) still failed to afford any [¹¹C]2a (Table 1, entries 9 and 10), whereas increasing the amount of KHMDS to 45 μmol gave [¹¹C]2a in 79% yield (Table 1, entry 11) under identical reaction conditions. Moreover, increasing KO^tBu to 60 μmol substantially decreased the yield of [¹¹C]2a to 42% (Table 1, entry 12). but the same amounts of P₁-t-Bu and BEMP (60 μmol) produced no product (Table 1, entries 13 and 14). Highest yield (87%) of [¹¹C]2a was obtained with 60 μmol KHMDS (Table 1, entry 15) and somewhat lower yield (77%) was obtained with 75 μmol KHMDS (Table 1, entry 16). Not surprisingly, 25 μmol of KHMDS failed to afford any [¹¹C]2a (Table 1, entry 17). Because 45 μmol of KO^tBu gave the best yield of [¹¹C]2a and the yield was only a little less than that with 60 μmol KHMDS, we decided to study substrate scope with KO^tBu (45 μmol) as base in DMF at RT for 10 minutes followed by quenching with 4M HCl (0.1 mL) at 60 °C for 5 minutes. We tested several methyl esters for ¹¹C-trifluoromethylation under these conditions (Table 2).

Table 1:

Optimization for the formation of [¹¹C]2,2,2-trifluoro-1-phenylethanone ([¹¹C]2a).^a,b,c

Entry	Base	Amount (μmol)	HCl (M)	Yield (%)
1	KOtBu	15	12	65 ± 10
2	P1-t-Bu	15	12	0
3	BEMP	15	12	0
4	KHMDS	15	12	0
5	KOtBu	45	12	83 ± 3
6	KOtBu	45	4	84 ± 5
7	KOtBu	45	2	81 ± 2
8	KOtBu	45	0.5	75 ± 8
9	P1-t-Bu	45	4	0
10	BEMP	45	4	0
11	KHMDS	45	4	79 ± 2
12	KOtBu	60	4	42 ± 5
13	P1-t-Bu	60	4	0
14	BEMP	60	4	0
15	KHMDS	60	4	87 ± 3
16	KHMDS	75	4	77 ± 9
17	KHMDS	25	4	0

^aReaction conditions: i) 1a (50 μmol), base, DMF (450–500 μL). [¹¹C]Fluoroform was added in DMF (100–150 μL), giving a reaction volume of ~ 600 μL. ii) HCl (0.5‒12M, 100 μL), 60 °C, 5 min.

^bEach decay-corrected yield is from [¹¹C]fluoroform and determined with radio-HPLC. Yield was also measured at least once from the radioactivity of isolated [¹¹C]2,2,2-trifluoro-1-phenylethanone ([¹¹C]2a) as a percentage of the radioactivity collected from the HPLC column; these yields are very similar to the HPLC yields. (Note that ¹¹C-labeled byproducts from the synthesis of [¹¹C]fluoroform were excluded for the calculation of yields).

^cReactions were performed in triplicate (n = 3) and the radio-HPLC yield is shown as mean ± SD for n = 3. [P₁-t-Bu = tert-Butylimino-tris(dimethylamino)phosphorane; BEMP = 2-tert-Butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorine; KHMDS = Potassium hexamethyldisilazide].

Table 2.

Substate scope for aryl- or alkyl-[¹¹C]trifluoromethyl ketones ([¹¹C]2a‒2o) synthesis.^a,b,c

^aReaction conditions: i) 1a–1o (50 μmol), KO^tBu (45 μmol), DMF (450–500 μL), [¹¹C]Fluoroform was added in DMF solution (100–150 μL) to give a reaction volume of ~ 600 μL. ii) 4M HCl (100 μL), 60 °C, 5 min.

^bRadio-HPLC yields (%) are shown in blue;

^cReactions were performed in triplicate (n = 3);

^dKHMDS (45 μmol) was used because 45 μmol KO^tBu gave much lower yield;

^eIsolated yield (n = 2) in green.

The methyl benzoates, 1a‒1j, bearing either an electron-withdrawing or electron-donating group afforded the desired aryl products, [¹¹C]2a‒2j, in mostly moderate to excellent yields (15‒95%), except for methyl 4-nitro benzoate (1i) which gave no yield. We indicate that the iodo- and bromo-substituted compounds, [¹¹C]2e, [¹¹C]2h and [¹¹C]2j, might be used in a metal-catalyzed C‒C coupling or C‒H activation reactions. We measured the molar activity (A_m) of [¹¹C]2e to exemplify the no-carrier-added nature of this radiochemical procedure. [¹¹C]2e was isolated in about 80% decay-corrected yield with 97.5% radiochemical purity and a high molar activity (A_m = 56 ± 6 GBq/μmol; n = 2) at the end of synthesis (60 min from the production of [¹¹C]methane using a 15 μA x 15 min cyclotron irradiation). Such a high molar activity is fully consitent with a NCA procedure. The ortho-methyl substituted ester, 1b, gave only 15% yield of [¹¹C]2b, but the ortho-fluoro benzoate, 1c, produced [¹¹C]2c in excellent yield (82%). These reactions, when performed in the presence of 45 μmol KO^tBu, gave much lower yields of [¹¹C]2b and [¹¹C]2c, respectively. However, much higher yields of [¹¹C]2b and [¹¹C]2c were obtained by using 45 μmol KHMDS. The ¹¹C-trifluoromethylation of α,β-unsaturated esters, 1k‒1l, was also successful in producing the ¹¹C-trifluoromethylketones [¹¹C]2k and [¹¹C]2l, giving 77% and 23% yields, respectively. A very good yield of 69% was obtained for ¹¹C-labeled 3-(2,2,2-trifluoroacetyl)pyridine [¹¹C]2m by ¹¹C-trifluoromethylation of methyl nicotinate (1m). Finally, to extend the scope of ¹¹C-trifluoromethylketones formation, we also tested alkyl esters, 1n‒1o, under optimal ¹¹C-trifluoromethylation conditions (Table 1, entry 5). These reactions produced alkyl [¹¹C]trifluoromethylketones [¹¹C]2n and [¹¹C]2o in moderate to excellent yields of 22% and 82%, respectively.

We then focused on the synthesis of [¹¹C]2,2,2-trifluoro-1-(4-nitrophenyl)ethenone ([¹¹C]2i) because the earlier optimal reaction conditions (Table 1, entry 5) failed to render any product. Previous studies suggest that 2,2,2-trifluoro-1-(4-nitrophenyl)ethenone (2i) is moisture sensitive and therefore, exists as 2,2,2-trifluoro-1-(4-nitrophenyl)ethane-1,1-diol (3i).³¹ Nevertheless, to achieve a successful reaction with methyl 4-nitrobenzoate (1i), we further tuned the reaction conditions (Table 3). ¹¹C-Trifluoromethylation of the methyl ester (1i) was still inefficient with 15 μmol, 45 μmol, and 125 μmol KO^tBu (Table 3, entries 1‒3) or with 25 μmol KHMDS (Table 3, entry 4). We sought to utilize a Weinreb’s amide (1p) to produce [¹¹C]3i because Rudzinski et al. reported that such amides are effective precursors for synthesizing trifluoromethyl ketones.³³ Amide 1p failed to provide [¹¹C]3i in the presence of 25 μmol KHMDS (Table 3, entry 5). However, we were delighted to find that 1p underwent successful ¹¹C-trifluoromethylation in the presence of 45 μmol KHMDS as base to afford [¹¹C]3i in good yield (60%) (Table 3, entry 6). The yield of [¹¹C]3i was improved substantially to 92% by using 60 μmol KHMDS as base (Table 3, entry 7).

Table 3:

Optimization for the formation of 2,2,2-trifluoro-1-(4-nitrophenyl)ethane-1,1-diol ([¹¹C]3i).^a,b

Entry	R	Base	μmol	Yield (%)
1	OMe (1i)	KOtBu	15	0
2	OMe (1i)	KOtBu	45	0
3	OMe (1i)	KOtBu	125	0
4	OMe (1i)	KHMDS	25	0
5	N(OMe)Me (1p)	KHMDS	25	0
6	N(OMe)Me (1p)	KHMDS	45	60 ± 6
7	N(OMe)Me (1p)	KHMDS	60	92 ± 2

^aReaction conditions: i) 1i or 1p (50 μmol), base, DMF (450–500 μL). [¹¹C]Fluoroform was added in DMF solution (~ 100–150 μL) to give a reaction volume of ~ 600 μL. ii) 4M HCl; 60 °C, 5 min.

^bDecay-corrected yields were measured from radio-HPLC chromatograms as mean ± SD for n = 3.

We emphasize that this radiosynthetic method can be adapted easily for use with [¹⁸F]fluoroform when there is a need for radiotracers with longer physical half-life. We verified this potential by achieving a 79% yield of [¹⁸F]2a from 1a, utilizing [¹⁸F]fluoroform under the optimal reaction conditions described in Table 1, entry 6. ¹⁸F-Trifluoromethylation of 1d was also successful in giving a high yield (94%) for [¹⁸F]2d. However, a much lower yield (32%) was obtained for [¹⁸F]2j from 1j under optimal conditions. We speculate that an unknown impurity in the [¹⁸F]fluoroform (which is used in crude form) led to the lower yield, in this case (Table 4).

Table 4.

Reaction scope leading to the formation of aryl- [¹⁸F]trifluoromethyl ketones ([¹⁸F]2)^a,b,c

^aReaction conditions: i) 1a, 1d or 1j (50 μmol), KO^tBu (45 μmol), DMF (450–500 μL); [¹⁸F]Fluoroform was added in DMF solution (~ 100–150 μL) to give a reaction volume of ~ 600 μL. ii) 4M HCl (100 μL), 60 °C, 5 min;

^bRadio-HPLC yields (%) are shown in blue.

^cReactions were performed in triplicate (n = 3).

In summary, we have developed a new method for the copper (I)-free syntheses of no-carrier-added ¹¹C- and ¹⁸F-labeled trifluoromethyl ketones as new radiolabeled chemotypes from the reactions between aryl- and alkyl esters with [¹¹C/¹⁸F]fluoroform in presence of a suitable base under mild conditions. The no-carrier-added radiolabeled products were obtained in moderate to excellent radiochemical yields (15‒95%) within 15 minutes. This methodology can be expanded into a modular method for synthesizing a diverse range of ¹¹C- and ¹⁸F-labeled tracers.