Unexpected Thorpe Reaction of an α-Alkoxynitrile

Abstract

α-Alkoxynitrile 1 in the presence of tris(methylthio)methyllithium 2 at –78°C gave the dimer 5 instead of the expected C₁-elongated product 3. The formation of com-pound 5 is explained in terms of anion formation and self-condensation, a variant of the Thorpe reaction. Scrutinizing the ¹H NMR spectra revealed that the enamine tautomer 5b is predominant over the imine 5a in the solvents investigated.

Introduction

Nitrile 1 (Scheme 1) was obtained en route to acyclic analogues of the C-nucleoside tiazofurin [1]. This compound was deemed to be a good starting material for the synthesis of α-keto ester 4 via in-termediate 3. This assumption was based on analogy with the Grignard reaction of nitriles giving rise to ketones [2] or the Blaise reaction (Reformatsky reaction of nitriles affording β-keto esters) [3].

For the required carboxylate anion synthon different tris(alkylthio or arylthio)methyllithiums, de-rived from tris(alkyl or arylthio)methanes [4], were used successfully in the C₁ elongation of ketones [5] and lactones [6]. Most often [tris(methylthio)methyl]lithium (2) was applied which is stable up to –40°C [7]. The obtained trithioorthoesters can be transformed into the corresponding esters using the common mercury(II) salt-assisted demercaptalisation procedure. In principle, alkyl chloroformates could also be employed as C₁ synthons. However, in the single reported Reformatsky-type reaction of ethyl chloroformate with ethyl acetoacetate in the presence of a zinc-copper couple leading to diethyl 2-hydroxy-2-methylsuccinate, the yield was rather low (20%) [8]. In a recent application, ethyl chloro-formate was allowed to react with methyl phenylglyoxylate in a titanium(IV)-promoted reaction to af-ford ethyl methyl 2-hydroxy-2-phenylmalonate in 60% yield [9].

Results and Discussion

Nitrile 1 was allowed to react with compound 2 at –78°C for 4 h. Contrary to our expectations, compound 3 was not obtained. The isolated substance exhibited surprisingly complicated ¹H NMR spectra in different solvents, according to the microanalysis it contained no sulfur and its elemental composition was identical with that of nitrile 1. Its low-voltage (15 eV) EI mass spectrum showed a molecular ion at m/z 622 (the molecular mass of the starting nitrile is 311) and a fragmentation pattern compatible only with the dimeric structures 5a and 5b. Scrutiny of the ¹H NMR spectrum revealed that the substance exists as an enamine (5b) rather than an imine (5a). The IR spectrum also confirmed the presence of a nitrile group (ν_CN 2180 cm^-1).

The literature search for precedents (Beilstein’s CrossFire 4.0, release BS9902PRPR using Beilstein Commander 4.0) showed that α-alkoxynitriles react normally with Grignard reagents [10] while orga-nolithium [11] and organosodium compounds [12,13] provoke self-condensation. This can be ex-plained by the strong basicity of organolithiums and the result is, just as in the above instance, the spe-cial case of the well-known Thorpe reaction [14] (Scheme 2).

It is noteworthy that even acetonitrile tends to react abnormally with Grignard reagents, e.g. with phenylmagnesium bromide, only 42% of the expected acetophenone has been obtained along with 32% of benzene [15]. Nitriles lacking α-hydrogens react normally and afford ketones.

Experimental

General

To tris(methylthio)methane [4] (0.927 g, 6.0 mmol) dissolved in anhydrous THF (2.0 mL) was added butyllithium (1.26 M solution, 5.20 mL, 6.60 mL) at -78°C under nitrogen. Nitrile 1 [1] (1.40 g, 4.5 mmol) dissolved in anhydrous THF (10.0 mL) was added dropwise to this solution after 20 min and kept at –78°C for 4 h. The reaction mixture was quenched at –60°C with satd. NH₄Cl solution. Af-ter allowing to warm up to room temperature, the mixture was diluted and extracted with chloroform (25 mL). The aqueous phase was extracted with chloroform (2 × 50 mL). The combined organic phases were dried (MgSO₄) and evaporated in vacuo. Chromatographic purification using 20% (v/v) ethyl acetate in hexanes afforded 3-amino-2,4-bis{2-(benzyloxy)-1-[(benzyloxy)methyl]-ethoxy}-2-butene-nitrile (5, 0.56 g, 40%) as a yellow oil.

Spectral Data

IR (KBr, ν, cm^-1): 3450m, 3330m, 3020m, 2910s, 2860s, 2180m, 1680m, 1493m, 1450m, 1090m.

¹H NMR (CDCl₃, 200 MHz, δ, ppm): 3.55 (d, J 6 Hz, 4 H, 2 x CH₂); 3.65 (d, J 6 Hz, 4 H, 2 x CH₂); 3.71 (m, 1 H, CH); 3.94 (m, 1 H, CH); 4.37 (s, 2 H, 4-CH₂); 4.49 (s, 4 H, 2 x PhCH₂); 4.52 (s, 4 H, 2 x PhCH₂); 5.45 (br s, exchangeable with D₂O, 2 H, NH₂); 7.30 (m, 20 H, 4 x C₆H₅).

¹H NMR (DMSO-d₆, 200 MHz, δ, ppm): 3.52 (d, J 5 Hz, 4 H, 2 x CH₂); 3.65 (d, J 5 Hz, 4 H, 2 x CH₂); 3.72 (m, 1 H, CH); 3.88 (m, 1 H, H); 4.20 (s, 2 H, 4-CH₂); 4.45 (s, 8 H, 4 x PhCH₂); 6.00 (br s, exchangeable with D₂O, 2 H, NH₂); 7.30 (m, 20 H, 4 x C₆H₅).

EI-MS (15 eV, m/z, %): 622 (5, M⁺); 531 (1, M-91); 418 (4); 387 (4); 367 [4, M-CH(CH₂OBn)₂]; 331 (6); 292 (9); 278 (24); 220 (14); 181 (75); 102 (88); 91 (100, C₇H₇⁺).

Anal. calcd. for C₃₈H₄₂N₂O₆ (622.750): C, 73.29; H, 6.80; N, 4.50; found: C, 73.49; H, 6.71; N, 4.18.