Nickel-Catalyzed Deaminative Cyanation: Nitriles and One-Carbon Homologation from Alkyl Amines

Abstract

A nickel-catalyzed deaminative cyanation of Katritzky pyridinium salts has been developed. When coupled with formation of the pyridinium salt from primary amines, this method enables alkyl amines to be converted to alkyl nitriles. A less toxic cyanide reagent, Zn(CN)₂, is utilized, and diverse functional groups and heterocycles are tolerated. The method also enables a one-carbon homologation of alkyl amines via reduction of the nitrile products, in addition to many other potential transformations of the versatile nitrile group.

Graphical Abstract

An alkyl nitrile is one of the most common groups in bioactive molecules, ranking as the 15^th most common functional group and found in 5.2% of bioactive molecules (Scheme 1A).¹ This prevalence is due to nitriles’ effective bonding in protein active sites.² A nitrile is also a versatile precursor for a wide variety of functional groups.³ Because of this utility, elegant methods have been developed for nitrile incorporation. These methods typically employ alkyl halides as substrates and install the nitrile via either an S_N2 or transition metal-catalyzed reaction with a cyanide source (Scheme 1B).⁴ Because of the sensitivity of alkyl halides to substitution or elimination under many reaction conditions, their use mandates strategic timing of this cyanation sequence, often at the beginning of the synthesis.

Scheme 1. Alkyl nitriles and cyanation reactions.

^aPrices of bulk quantities (≥1 kg) from MilliporeSigma on 5/23/2021

In contrast to an alkyl halide, an alkyl amine offers tremendous flexibility in synthetic strategy; an amine can be functionalized early or, via suitable protection, carried through multiple synthetic steps and revealed for a late-stage transformation.⁵ In addition, amines are ubiquitous and diverse, spanning simple starting materials of exceptional availability and variety to complex pharmaceutical intermediates and APIs.^5–6 We envisioned that a deaminative cyanation would be possible via a Katritzky pyridinium salt.⁷ Starting from an alkyl amine 1, condensation with the commercially available pyrylium salt 2 gives pyridinium 3,⁸ which would deliver alkyl nitrile 4 via a nickel-catalyzed cyanation (Scheme 1C). Use of a relatively low toxicity cyanide source, typically enabled by a transition metal-catalyzed approach,⁹ would further increase the utility of this method. Not only would this strategy enable complex, drug-like alkyl amines to be converted into alkyl nitriles with potential bioactivity against new targets, it would also offer the opportunity for a one-carbon homologation of alkyl amines. As shown by the comparison of β-alanine and γ-aminobutyric acid, the addition of a single methylene can have dramatic impact on the bioactivity and cost (Scheme 1D).¹⁰ However, a one-carbon homologation of a lead compound would typically require complete re-synthesis. Using a cyanation/reduction sequence enables the original amine to be directly converted to the homologated product. Notably, this new amine product could then be used in the suite of deaminative reactions that have been developed since 2017.^{7, 11}

We selected the cyanation of pyridinium 3a for reaction development. Inspired by the work of Liu and co-workers using Zn(CN)₂ as a less toxic cyanide source in reactions with alkyl halides, we selected Zn(CN)₂ as our coupling partner.^4d Using the bidentate bipy ligand, which is commonly employed in nickel-catalyzed reactions of pyridinium salts, poor reactivity was observed (Table 1, entry 1). In contrast, the use of Xantphos provided 72% yield (entry 2). Notably, Liu also used Xantphos in their cyanation.^4d Yield was further improved by the addition of zinc halide salts, with 87% yield observed when ZnBr₂ was used (entries 3 and 4). Because the Ni/Xantphos catalyst system is unprecedented in deaminative couplings of pyridinium salts and unusual in Ni^I/III catalysis, we investigated the use of related ligands. Although P-alkyl Xantphos derivatives completely failed (entry 5), related P-aryl ligands with large bite angles could be used (entries 6 and 7). These results are consistent with Diao’s observation that the large bite angle of Xantphos can stabilize the necessary Ni(I) intermediate, and that bulky P-alkyl groups prevent productive cross-coupling.¹² In addition, control experiments demonstrated that both nickel and ligand are required (entries 8 and 9); competitive S_N2 displacement does not occur at this reaction temperature for either primary or secondary alkylpyridinium salts (see 6 in Scheme 2 for primary example).

Table 1.

Optimization^a

entry	[Ni]	ligand	additive	yield (%)b
1	NiCl2	bipy	none	<5
2	NiCl2	Xantphos	none	72
3	NiCl2	Xantphos	ZnCl2	80
4	NiCl2	Xantphos	ZnBr2	87
5	NiCl2	t-Bu-Xantphos	ZnBr2	<5
6	NiCl2	DPE-Phos	ZnBr2	52
7	NiCl2	dppf	ZnBr2	47
8	NiCl2	none	ZnBr2	<5
9	none	Xantphos	ZnBr2	<5

^aConditions: 3a (0.10 mmol), Zn(CN)₂ (1.0 equiv), [Ni] (10 mol %), ligand (12 mol %), Et₂Zn (0.4 equiv), additive (1.0 equiv), DMSO (0.2 M), 80 °C, 16 h, unless otherwise noted.

^bDetermined by ¹H NMR analysis using 1,3,5-trimethoxybenzene as internal standard.

Scheme 2. Scope^a.

^a Conditions: 3 (1.0 mmol), Zn(CN)₂ (1.0 equiv), NiCl₂ (10 mol %), Xantphos (12 mol %), Et₂Zn (0.40 equiv), ZnBr₂ (1.0 equiv), DMSO (0.2 M), 80 °C, 16 h, unless otherwise noted. ^b 0.1 mmol scale. Yield determined by ¹H NMR analysis using 1,3,5-trimethoxybenzene as internal standard. ^c Without ZnBr₂. ^d 2.1:1 dr. ^e 24 h.

Under these optimized conditions (see Table 1, entry 4), we observed broad scope across an array of Katritzky pyridinium salts. Both primary and secondary alkylpyridinium salts were compatible (Scheme 2). Various functional groups are well tolerated, including acetal (6), benzonitrile (7), carbamate (9, 11, 13), tertiary amine (10, 13), and ester (15, 16) groups. Excellent heterocycle incorporation is exemplified by pyridine 8, azetidine 9, pyrrolidine 10, piperidines 11 and 12, pyrimidine 12, and piperazine 13. Notably, a reactive benzylic pyridinium salt could also be utilized in this cyanation with good yield (14). Additionally, pyridinium salts with β-electron withdrawing groups often suffer elimination under basic reaction conditions.^7a However, these types of pyridinium salts worked well in this cyanation, as shown by nitrile 16.

Generality of this methodology was demonstrated by late-stage functionalization of natural products, pharmaceuticals, and pharmaceutical intermediates (Scheme 2). The cyanation of terpene-derived amines, (−)-cis-myrtanylamine (17) and pinanamine (18) proceeded in acceptable yields. The pyridinium salts of amine intermediates in the syntheses of the antidepressant drug agomelatine,¹³ Lipitor^®,¹⁴ and mosapride¹⁵ were efficiently converted to nitriles 19, 20, and 21 in excellent yields. Ester derivatives of the muscle spasticity treatment, baclofen (22),¹⁶ and bleeding disorder treatment, tranexamic acid (23),¹⁷ were also effective as pyridinium substrates. Finally, the cyanation of the pyridinium salt of the diterpene amine, leelamine (24),¹⁸ proceeded in moderate yield. Additional functional group robustness was also noticed in these examples, including ketal 20, morpholine 21, and aryl chloride 22.

We have also demonstrated the use of this cyanation in an efficient one-carbon homologation of amine 25 to Boc-protected amine 26 (Scheme 3). We first tried a one-pot procedure, in which we simply added the reagents for each step sequentially to the same reaction vessel. However, pyridinium formation is incompatible with the use of DMSO as solvent, and the cyanation was sensitive to EtOH, which would typically be used for pyridinium formation. In addition, the presence of DMSO in the nitrile reduction diminishes yield by competitive reduction. We found that much higher yields and cleaner reactions were achieved if a simple filtration and concentration procedure followed pyridinium formation and cyanation. Under these conditions, 44% overall yield was achieved in the one-carbon homologation of 25 to 26, which compares favorably to the 53% overall yield observed with careful purification after each step. This protocol gives medicinal chemists another useful tool to homologate alkyl primary amines efficiently.

Scheme 3.

One-carbon homologation

With respect to the reaction mechanism, key steps likely include formation of a (Xantphos)Ni(CN) species, single-electron transfer (SET) from a Ni(0) or Ni(I) intermediate to the alkylpyridinium salt to generate an alkyl radical, and either inner or outer sphere attack of that radical on a nickel cyanide intermediate to ultimately deliver product. The role of Et₂Zn appears to be predominantly as a reductant (see Supporting Information), and the formation of both diastereomers of 15 are consistent with an alkyl radical intermediate. Further studies are ongoing to establish why Xantphos is particularly effective, as well as the role of the ZnBr₂ additive.

In summary, we have developed a nickel-catalyzed cyanation via deamination of Katritzky pyridinium salts. A broad scope of primary and secondary alkyl amine precursors, easy preparation of pyridinium salts, and practical cyanation conditions with a less toxic cyanide source makes this method a powerful tool to discover new molecules, particularly nitriles and nitrile derivatives, including homologated amines.