Borylated cyclobutanes via thermal [2 + 2]-cycloaddition

A one-step approach to borylated cyclobutanes from amides of carboxylic acids and vinyl boronates is elaborated. The reaction proceeds via the thermal [2 + 2]-cycloaddition of in situ-generated keteniminium salts.


Introduction
Small aliphatic rings attract considerable attention in contemporary research. 1For example, the cyclobutane ring is common within modern bioactive compounds 2 and can be found in the structures of at least ten market-approved drugs. 3This motivated substantial development of cyclobutyl boronate chemistry during the past decade owing to the fact that the carbon-boron bond provides an excellent site for functionalization. 4The known approaches to the preparation of cyclobutyl boronates include the C-H activation of cyclobutanes, 5 electrocyclization, 6 functionalization of bicyclo [1.1.0]butanes, 7borylation 8 and hydrogenation of cyclobutenes, 9 along with other methods. 10,11he most frequently used approach is a [2 + 2]-cycloaddition.Strikingly, while the photochemical version of this transformation has been elaborated in numerous studies by Hollis, Bach, Hiemstra, Grygorenko, Yoon, Swierk with Brown, Romanov-Michailidis with Knowles, and Fürstner, 12 the thermal approach remained underdeveloped for unclear reasons (Scheme 1).We found only a single example in the literature on non-catalyzed thermal [2 + 2]-cycloaddition between alkene 1 and ketene 2 (Scheme 1).In 1969, Fish demonstrated that heating this mixture in a sealed vial for 15 days afforded the target cyclobutane 3 in 23% yield. 13Also, recently Brown showed an example of a thermal [2 + 2]cycloaddition between a borylated alkene and an allene that required, however, Lewis acid catalysis. 14n this work, we developed a one-step approach to borylated cyclobutanes by thermal [2 + 2]-cycloaddition between vinyl boronates and in situ-generated keteniminium salts.

Results and discussion
From the pioneering study of Fish (Scheme 1), 13 it seemed reasonable to assume that vinyl boronates are amenable to noncatalytic thermal [2 + 2] cycloadditions; however, a more active partner than a ketene was needed.We turned our attention to keteniminium salts that are known to be more active than ketenes. 15Moreover, the [2 + 2]-cycloaddition of keteniminium salts with alkenes has been reported. 16,17Despite the substantial recent progress in keteniminium chemistry, 15 we found no literature mentioning the reaction of borylated alkenes with keteniminium salts.Initially, we suspected that the conditions for their generation, which typically involved treatment with triic anhydride, 17 do not tolerate the Bpin group and the latter might decompose.Nonetheless, we decided to examine the feasibility of this transformation.The model reaction between N,N-dimethylacetamide and vinyl Bpin did not produce the desired product 4 even at trace amounts when performed in reuxed dichloroethane.Lowering the temperature to 60 °C or increasing the number of keteniminium equivalents from 1.2 to 3, 4, and 5 were as unsuccessful (Scheme 2, part limitations).Only the starting vinyl Bpin along with Scheme 2 Reaction conditions: (i) vinyl Bpin/allyl Bpin (1.0 equiv.),amide (1.2 equiv.),triflic anhydride (1.4 equiv.),collidine or lutidine (1.4 equiv.),1,2-dichloroethane, reflux, and 16 h; (ii) aqueous NaHCO 3 ; (iii) purification (vacuum distillation or column chromatography).The scale of the synthesis: a 100-500 mg; b 1-7 g; c 10-50 g of the isolated product.d Product 25 (d.r.= 7 : 1) was obtained with ca.70% purity.e Additional purification by column chromatography provided the pure product 25 as a single diastereomer in 3% yield.f Product 26 (d.r.= 2 : 1) was obtained with ca.50% purity.g Additional purification by column chromatography provided the pure product 26 as a single diastereomer in 1% yield.X-ray crystal structures of compounds 14-16, 20, and 22 are shown as thermal ellipsoids at a 50% probability level; carbonwhite, oxygenred, boronbrown, and fluorinegreen; hydrogen atoms are omitted for clarity.
unidentied side products was detected in the reaction mixture.We were quite discouraged by the futility of our initial efforts, yet we attempted another reaction involving the homologous N,N-dimethylpropanamide. Serendipitously, the reaction worked out.Activation of the amide with (CF 3 SO 2 ) 2 O/collidine (in situ formation of the keteniminium salt) followed by its reaction with vinyl Bpin produced the desired product 5 (Scheme 2).
Aer short optimization of the reaction conditions, we found that performing the reaction in reuxing dichloroethane for 12 hours produced excellent conversion of the starting vinyl boronate (see Table S1 in the ESI †).Thus, we examined the reaction scope by taking various amide counterparts and obtained borylated cyclobutanes 5-16 in decent yields (Scheme 2).The reaction was found compatible with the presence of the cyclopropyl ring (7), the active chlorine atom (8), and the gem-diuoro motif (10 and 15) in the product substances.Substituted alkenes, CH 2 ]C(Me)-Bpin and CH 2 ]C(Ph)-Bpin, gave the desired borylated cyclobutanes 17-22 as well.Products 5, 7, 8, 12, and 18 were obtained as inseparable mixtures of two diastereomers.The structure of products 14-16, 20 and 22 was conrmed by X-ray crystallographic analysis. 18he developed reaction showed few limitations, however.The keteniminium salt obtained from N,N-dimethylacetamide reacted with substituted vinyl boronates (products 17 and 21) but failed to react with vinyl Bpin (4).Attempts towards the synthesis of compounds 23, 24 and 27 failed as well.Analysis of the reaction mixture revealed either the presence of unreacted vinyl Bpin (23, 27) or the formation of a complex mixture (24).Some products, such as 25 and 26, were obtained in low isolated yields because they required rather tedious purication.The purication led to isolation of a single diastereomer in each case, and the trans-conguration of compounds 25 and 26 was revealed by X-ray crystallography. 18he analogous reaction between amides of carboxylic acids and the homologous CH 2 ]CH-CH 2 -Bpin also produced the desired products 28-31 (Scheme 2).
It is important to note that the reaction demonstrated good performance on milligram, gram, and even multigram scales (10, 12, and 20).When carrying out the reaction on a small scale, we puried products by silica gel column chromatography.On a gram-to-multigram scale, we isolated the products by distillation under reduced pressure, which is more practical.Despite the seeming simplicity of the current approach to borylated cyclobutanes, to the best of our knowledge, none of the obtained products depicted in Scheme 2 has been reported in the literature.
Somewhat unexpectedly, the reaction between b,b-disubstituted vinyl boronate 32 and N,N-dimethylacetamide produced ketone 33 rather than the borylated cyclobutane 34 (Scheme 3).While we did not examine the exact mechanism of this transformation, 19 we found a fairly reasonable explanation for the observed outcome.We reasoned that the bulky C]NMe 2 moiety approached the tertiary rather than quaternary carbon of the amide in the course of the cycloaddition probably due to steric reasons (Scheme 3, proposed explanation).Effectively, this steered the reaction towards the formation of the intermediate compound 35, which is related to the class of unstable a-borylated ketones 20 prone to hydrolytic protodeborylation, thus producing ketone 33.
Next, we performed transformations of the obtained products.For example, despite the failed attempt to direct synthesis of cyclobutane 23 from amide 36 (Scheme 2, limitations), we were able to obtain compound 23 by an intramolecular cyclization of the previously synthesized chloride 8 in 84% yield (Scheme 4a).From ketone 6, the corresponding amino boronate 37 was synthesized in two steps (Scheme 4b).The subsequent Nprotection provided N-Boc amino boronate 38, which represents a useful medicinal chemistry precursor.Analogously, the N-Boc amino boronate 40 was obtained from ketone 9 via amine 39.
In 2021, Qin and colleagues developed an elegant intramolecular coupling towards multi-substituted bicycloalkyl boronates.11a The corresponding starting materials were synthesized in multiple steps.In this work, we obtained ketones 29 and 30 in one step, and we also attempted their cyclization into the desired bicyclo[1.1.1]pentanes41 and 42 (Scheme 4c).Unfortunately, under the original conditions reported by the Qin group, the formation of the desired products was not observed.Our unsuccessful efforts corroborate the seminal conclusion that the presence of an additional substituent at the cyclobutane ring is crucial for the formation of bicyclo[1.1.1]pentanein this reaction.11a Finally, we performed a few other representative modications.The reaction of pinacol boronates 6, 9, and 13 with KHF 2 produced potassium triuoroborates 6a, 9a, and 13a (Scheme 5).The reduction of the ketone group in 13 with NaBH 4 gave alcohol 43.The oxidative cleavage of the Bpin group in the latter substance gave diol 44.We assume that similar modications could be performed with the other obtained Bpin ketones following our protocols.

Conclusions
Here, we elaborated a thermal [2 + 2]-cycloaddition between vinyl boronates and in situ generated keteniminium salts.This practical approach allows the preparation of borylated cyclobutanes in one step.The obtained compounds can be used in the syntheses of various functionalized cyclobutanes.