Copper(II)-Mediated [¹¹C]Cyanation of Arylboronic Acids and Arylstannanes

Abstract

A copper-mediated method for the transformation of diverse arylboron compounds and arylstannanes to aryl–[¹¹C]-nitriles is reported. This method is operationally simple, uses commercially available reagents, and is compatible with a wide variety of substituted aryl- and heteroaryl substrates. This method is applied to the automated synthesis of high specific activity [¹¹C]perampanel in 10% non-decay corrected radiochemical yield.

Graphical Abstract

Carbon-11 is a radioisotope that is commonly used for positron emission tomography (PET) imaging.¹ The introduction of carbon-11 into PET radiotracers is particularly challenging due to its very short half-life (t_1/2 = 20 min).² A number of methods have been developed for [¹¹C]-radiolabeling, and some of the most attractive involve the late-stage introduction of a [¹¹C]CN substituent (Scheme 1).³ [¹¹C]Cyanide offers an advantage because it can be readily generated from [¹¹C]CO₂.^1,2 Additionally, the nitrile functionality is common in bioactive molecules⁴ and can also be rapidly transformed into other important functional groups, including amides, carboxylic acids, and amines (Scheme 1).

Scheme 1.

Diversification of [¹¹C]Nitrile Substrates.

There are currently two major methods for the [¹¹C]cyanation of aromatic and heteroaromatic substrates. The first uses aryl halide precursors in combination with a Pd catalyst to engage [¹¹C]cyanide in aryl–CN cross coupling.^5–8 These reactions often provide high yields, exhibit broad scope, and proceed under mild conditions. However, Pd is relatively toxic;⁹ furthermore, the Pd-aryl intermediates and phosphine ligands required for these reactions can be challenging to handle, particularly in the context of automated radiochemical synthesis.¹⁰

A second [¹¹C]radiocyanation method involves the reaction of aryl halides with [¹¹C]CuCN (i.e., the Rosenmundvon Braun reaction).^11–13 This transformation offers the advantages of operational simplicity (i.e., no phosphine ligands, no requirement to preform organometallic intermediates, no need to remove palladium) and the relatively low toxicity of Cu.⁹ However, it suffers from low yields, modest scope, and forcing reaction conditions (often requiring temperatures of 150–250 °C).² We sought to develop an alternative Cu-mediated [¹¹C]radiocyanation that would leverage the advantages, while addressing the limitations of the existing Cu method. Our laboratory has recently reported that arylboron compounds and arylstannanes undergo Cu-mediated [¹⁸F]radiofluorination with [¹⁸F]KF.^14–17 We hypothesized that changing the nucleophile from fluoride to cyanide under similar reaction conditions might enable a mechanistically analogous [¹¹C]radiocyanation reaction (Scheme 2).¹⁸ We demonstrate here that the combination of Cu(OTf)₂, pyridine, and [¹¹C]KCN effectively promotes the [¹¹C]radiocyanation of diverse aryl boronic acids, aryl boronate esters, aryl trifluoroborates, and arylstannanes under relatively mild conditions. We show that this transformation is compatible with a wide variety of functional groups and of aryl/heteroaryl substrates. Furthermore, it is readily translated to the automated synthesis of the radiotracer [¹¹C]perampanel.¹⁹ Notably, while this work was underway, the team of Hooker, Vasdev, and Liang published a related method for the Cu-mediated [¹¹C]cyanation of arylboronic acids.²⁰

Scheme 2.

Inspiration from Cu-Mediated [¹⁸F]Fluorination of Arylorganometallics

Our initial studies focused on establishing the feasibility of the Cu-mediated cyanation of aryl organometallic reagents under conditions analogous to those for our [¹⁸F]radiofluorination reactions.¹⁵ Using Cu(OTf)₂ and pyridine in DMA at 100 °C, the reaction of 4-methoxyphenyl tributylstannane (1-SnBu₃) with KCN afforded 4-methoxybenzonitrile (1) in 28% yield in 1 h, as determined by ¹H NMR spectroscopy (Table 1, entry 1). Comparable results were obtained with the corresponding boronic acid (1-B(OH)₂) and trifluoroborate (1-BF₃K) substrates (32% and 26% yields, respectively), while the boronate ester afforded a lower yield (6%). However, the yield with the boronate ester could be increased to 21% by the addition of 1.2 equiv of KF. This additive likely promotes transmetalation via the formation of a borate intermediate. Notably, no fluorinated product was observed upon the addition of KF.

Table 1.

Initial results with KCN and optimization with [¹¹C]KCN.

entry	[M]	changes from standard	product (yield or RCC)
Results with KCNa
1	1-SnBu3	none	1 (28%)
2	1-B(OH)2	none	1 (32%)
3	1-BF3K	none	1 (26%)
4	1-Bpin	none	1 (6%)
5	1-Bpin	KF (1.2 equiv)	1 (21%)
Results with [11C]KCNb
6	1-SnBu3	[11C]KCN in DMA	[11C]1 (42%)
7	1-SnBu3	DMA prep, H2Oc	[11C]1 (41%)
8	1-SnBu3	none	[11C]1 (66%)
9	1-B(OH)2	none	[11C]1 (79%)
10	1-BF3K	none	[11C]1 (93%)
11	1-Bpin	none	[11C]1 (76%)

Standard conditions: substrate (10 μmol, 1 equiv), Cu(OTf)₂ (2 equiv), pyridine (15 equiv), DMA (1 mL, 10 mM).

^aKCN (2 equiv), 100 °C, 1 h. Yield determined by ¹H NMR spectroscopy with 1,3,5-trifluorobenzene as an internal standard.

^b[¹¹C]KCN in H₂O, 100 °C, 5 min. Reported values indicate radiochemical conversion (RCC) of determined by radio-TLC (n ≥ 2).

^cH₂O (0.2 mL, 17% v/v).

We next translated this method to [¹¹C]radiocyanation using anhydrous [¹¹C]KCN. Gratifyingly, the product ([¹¹C]1) was formed in 42% RCC, as determined by radio-TLC with identity confirmed by radio-HPLC (entry 6). Subsequent studies revealed that anhydrous conditions (which were essential for the analogous [¹⁸F]radiofluorination with [¹⁸F]KF) are not required for [¹¹C]radiocyanation. For example, a comparable 41% RCC was obtained upon the addition of exogenous water (0.2 mL, 17% v/v) to the anhydrously prepared [¹¹C]KCN reaction mixture. Furthermore, an even higher yield (66% RCC, entry 8) was observed when the [¹¹C]KCN was prepared and used directly as an aqueous solution. This modification decreased the overall synthesis time by 4 min (approximately 20% of the ¹¹C half-life). The improved RCC under these conditions is likely due to increased solubility of [¹¹C]KCN.

Using this aqueous [¹¹C]KCN preparation, we next evaluated the radiocyanation of other aryl organometallic substrates. As summarized in Table 1, entries 8–11, various arylboron compounds afforded comparable and/or improved RCCs relative to their stannane counterpart, with the trifluoroborate (1-BF₃K) giving the best result (93% RCC). Notably, KF was not required for the radiocyanation of arylboronate ester (1-Bpin), likely due to the decreased amount of KCN available in the reactions.

The scope of this transformation was further evaluated using a variety of organoboron and organostannane substrates (denoted by the footnotes in Figure 1). These studies showed that electron-donating ([¹¹C]1–4), -neutral ([¹¹C]5), and -withdrawing ([¹¹C]6–7) substituents on the aromatic ring are all well-tolerated. Ortho-substituted substrates underwent [¹¹C]radiocyanation in comparable yields to their unsubstituted counterparts (compare [¹¹C]8–10).²⁰ In addition, carbonyl groups ([¹¹C]11–14) were compatible with the reaction conditions.¹⁹ Significantly, precursors containing unprotected benzoic acid ([¹¹C]12) and phenol ([¹¹C]4) substituents also afforded modest to excellent yields. Aryl bromides were tolerated at various sites around the phenyl ring ([¹¹C]16–18), and could serve as handles for further elaboration of the products. Pyridine derivatives and related nitrogen heterocycles also underwent [¹¹C]radiocyanation in moderate to high yields ([¹¹C]19–25).

Figure 1.

Substrate Scope. Reported values indicate radiochemical conversion (RCC) of determined by radio-TLC (n ≥ 2). General conditions: substrate (0.01 mmol, 1 equiv), Cu(OTf)₂ (2 equiv), pyridine (15 equiv), [¹¹C]KCN in H₂O (0.1 mL), DMA (10 mM), 100 °C, 5 min. [M] = ^aSnBu₃, ^bB(OH)₂, ^cBF₃K, ^dBpin.

Overall, the scope of this transformation is broader, and many of the RCCs are higher than those of previously reported methods for the [¹¹C]radiocyanation of aromatic substrates.^8,20 As an example, our method affords quinoline product [¹¹C]20 in 71% RCC from 20-B(OH)₂. For comparison, recently reported Cu-mediated [¹¹C]radiocyanation conditions provide 18% RCC for the same substrate,²⁰ while a Pd-mediated method affords 46% RCC from the analogous aryl bromide.⁸ Notably, subjecting 20-B(OH)₂ to our related Cu-mediated fluorination affords only trace amounts of product (<10% ¹⁹F NMR yield),²¹ demonstrating that this cyanation reaction also has improved scope relative to fluorination. Finally, it is noteworthy that a number of the products in Figure 1 have not been labeled with [¹¹C]nitrile before, including [¹¹C]3–4, 6, 8–9, 12, 14–15, 17–19, 22, 25, and 27.

As a final demonstration of this method, we pursued the automated, clinical-scale synthesis of [¹¹C]perampanel. Perampanel, an FDA-approved drug for epilepsy, has been [¹¹C]radiolabeled once before using a Pd-mediated method with an aryl bromide precursor to afford [¹¹C]26 in 40% RCC (manual, radio-TLC) and 9.7% RCY (isolated, non-decay corrected).⁸ Our manual method provided 90 ± 2% RCC of [¹¹C]26 from the arylboronate ester using approximately 1 mCi of [¹¹C]KCN per reaction. The synthesis was scaled to 450 mCi of [¹¹C]KCN, and the [¹¹C]radiocyanation and subsequent HPLC purification of [¹¹C]26 were conducted using an automated radiosynthesis module. Without further optimization, this procedure afforded [¹¹C]26 in 10.4 ± 0.4% non-decay corrected radiochemical yield (RCY; Scheme 3). The fully automated synthesis lasted approximately 32 min from the end of bombardment.

Scheme 3.

Automation of [¹¹C]Perampanel.

General conditions: 26-Bpin (0.01 mmol, 1 equiv), Cu(OTf)₂ (2 equiv), pyridine (15 equiv), [¹¹C]KCN, DMA, 100 °C, 5 min. Radiochemical yield (RCY) determined by isolated material after preparative-HPLC (n = 2). QC was performed to confirm the correct product was formed and purify was >95%. NDC = non-decay corrected. DC = decay corrected. SA = specific activity.

In conclusion, this letter describes a Cu-mediated [¹¹C]radiocyanation of diverse aryl organometallic reagents. This method is compatible with a wide range of substrates, including those containing carboxylic acids, phenols, aldehydes, and heterocycles. This method is also amenable to automation on a clinically relevant scale, as demonstrated in the synthesis of [¹¹C]perampanel.