Key words
thiourea catalysis - asymmetric catalysis - diarylmethylamines - organocatalysis - Michael reaction
The pioneering works of the groups of MacMillan[1] and of List and Barbas[2] have triggered interest in the development of organocatalysts, which has evolved into one of the major topics of research in modern synthetic organic chemistry.[3] Many organocatalysts, such as chiral quaternary ammonium salts (Maruoka catalysts),[4] chiral phosphoric acid catalysts (Akiyama–Terada catalysts),[5]
[6] chiral cyclic secondary amine catalysts (Hayashi–Jørgensen catalysts),[7] and a chiral aminothiourea catalyst (the Takemoto catalyst)[8,9] have been developed, and their efficiencies have been well showcased in various asymmetric reactions. Among these organocatalysts, the aminothiourea catalyst, which consists of a chiral diamine moiety and a thiourea portion, has received immense attention for two reasons: (1) the ease of modification of the amine moiety, and (2) its high reliability and easy prediction of its asymmetric induction.[10]
[11] One of the major trends in these catalysts is that an aniline (aromatic amine) moiety is employed as an adjunct to a chiral amine moiety. The aromatic amine moiety is important for the improvement of the hydrogen-bonding ability of the thiourea moiety; however, its rigidity, in other words, its planar structure might be unfavorable for asymmetric induction in some reactions.[12] On the basis of this consideration, we recently focused on the diarylmethylamine unit, namely the so-called ‘butterfly’-type amine unit as a potentially good alternative. Because of the presence of one carbon atom between the aromatic rings and the thiourea moiety, a high flexibility and also the construction of a wide reaction field are expected. In addition, the involvement of additional interactions between a substrate and the catalyst, such as a C–H–π/π–π interaction, which is difficult to achieve in the case of an aniline-type catalyst, is also expected.
It is unfortunate that, despite their potential utility, examples of chiral aminothiourea catalysts having a butterfly-type amine unit are quite limited (Figure [1]). An early example was reported by Itoh and co-workers in 2010,[13a] who found that a catalyst with a chiral butterfly-type amine unit [a (2-hydroxynaphthyl)(phenyl)methyl group] effectively worked as a chiral unit in the asymmetric acyl-Strecker reaction of dihydroisoquinolines, although the selectivity was moderate (46% ee). In 2017, the group of Li and Cheng reported that the use of a simple (diphenylmethyl)amino-group-substituted catalyst with a chiral 1,2-cyclohexanediamine moiety resulted in moderate enantioselectivity (67% ee) in an asymmetric Michael reaction between nitrostyrene and diethyl malonate.[14] In 2016, Soós and co-workers reported that squaramide catalysts with a butterfly unit [a bis(1-naphthyl)methyl unit] exhibited excellent catalytic performance in a Robinson annulation reaction (91% ee).[15] Surprisingly, however, despite the high potential of the butterfly-type amine unit, no detailed investigation of the substituent effect on the aromatic rings of the amine unit has been conducted. Motivated by these circumstances, we investigated the substituent effect with a focus on the electronic effects of the aromatic rings in butterfly-type catalysts, and we found that a 3,5-bis(trifluoromethyl)phenyl-group-substituted catalyst exhibited superior catalytic performance compared with the well-known Takemoto catalyst in some asymmetric reactions.
Figure 1 Butterfly-type aminothiourea catalysts
The preparation of the catalyst 9a, containing a 3,5-bis(trifluoromethyl)phenyl group, is shown as a representative example in Scheme [1]. The coupling reaction between aldehyde 1a and the Grignard reagent 2a, prepared from commercially available 1-bromo-3,5-bis(trifluoromethyl)benzene, followed by mesylation of the resulting alcohol group, gave mesylate 4a in a good chemical yield (60% for the two steps). (Diarylmethyl)amine 6a, which was synthesized from 4a by a two-step azidation/reduction sequence, reacted with the known isothiocyanate 8
[16] to afford 9a in a good chemical yield (82% for the three steps).
Scheme 1 Synthesis of the butterfly-type aminothiourea catalyst 9a
By following this procedure, a series of catalysts 9b–g having various electronic natures of the aromatic rings were synthesized and their performance as catalysts was evaluated in the asymmetric Michael reaction between nitrostyrene 10 and dimethyl malonate as a model reaction (Scheme [2]). In the case of the benzylic-type catalyst 12, product 11a was obtained in 70% chemical yield with 71% ee. The enantioselectivity increased to 82% ee when the simplest butterfly-type catalyst 9b was employed. The use of catalysts 13a and 13b, derived from (R)- and (S)-1-phenylethylamine, respectively,[17] resulted in lower enantioselectivities of the product (77% ee and 74% ee, respectively) than the use of 9b. Although the selectivity toward the product obtained with 9b was lower than that obtained with the Takemoto catalyst 14 (82% ee vs 89% ee), the high potential of the butterfly-type catalyst prompted us to perform a detailed examination of the electronic effects of the aromatic rings on the catalytic activity. An excellent chemical yield (95%) and a high selectivity (86% ee) were realized when catalyst 9c with a 4-(trifluoromethyl)phenyl group was employed, whereas the use of 4-methoxyphenyl-group-substituted catalyst 9d resulted in a low chemical yield (32%) and a moderate enantioselectivity (73% ee). Catalyst 9e with a 3,5-dimethoxyphenyl group exhibited a good catalytic performance (82% yield and 84% ee). Gratifyingly, both the chemical yield and the enantioselectivity were improved to excellent levels (99% yield, 94% ee) when catalyst 9a with a 3,5-bis(trifluoromethyl)phenyl group was employed. It is worth emphasizing that both the chemical yield and the enantioselectivity were superior to those obtained when the Takemoto catalyst 14 was used. Stimulated by these results, we concentrated on the catalysts with highly electron-deficient aromatic rings (fluoroaromatic rings) and we obtained unexpected results. A satisfactory chemical yield (95%) and a high selectivity (90% ee) were achieved in the case of catalyst 9f with a 3,4,5-trifluorophenyl group. In sharp contrast, catalyst 9g with a pentafluorophenyl group afforded adduct 11a in 59% yield with 67% ee, the lowest among the butterfly-type catalysts examined.
Scheme 2 Evaluation of the catalytic performance of butterfly-type catalysts
The substrate scope of the asymmetric Michael reaction was then evaluated (Figure [2]). Various Michael adducts 11b–d possessing an electron-donating group (OMe) or an electron-withdrawing group (F or Br), and naphthyl-type adduct 11e were obtained in excellent enantioselectivities (≥91% ee) when catalyst 9a was employed. These enantioselectivities were higher than those obtained with the catalyst 14.
Figure 2 Substrate scope of the asymmetric Michael reaction
Examination of the above-mentioned catalysts offered helpful hints on the transition-state model of the present catalytic system. The clear correlation between the electronic nature of the aromatic groups and the enantioselectivity suggests the involvement of some interactions, such as a π–π/C–H–π interaction, between the aromatic groups in the catalyst and the substrate. The presence of an internal hydrogen-bonding interaction between the sulfur atom of the thiourea moiety and the hydrogen at the 2-position of the aromatic group, which is proposed for the Takemoto catalyst and the Schreiner catalyst,[18] is implied by considering the low enantioselectivity, even when using the highly electron-deficient catalyst 9g, which does not have hydrogen atoms at appropriate positions. On the basis of these considerations and the precedent of aminothiourea catalysts,[19] we propose the transition-state model for the present reaction shown in Figure [2]. In the same manner as the Takemoto catalyst, a hydrogen-bonding interaction between the nitrostyrene moiety and the conjugated acid of the dimethylamino portion of the catalyst, and two hydrogen-bonding interactions between the two hydrogens of the thiourea moiety and the conjugated base of malonate are involved. The above-mentioned hydrogen-bonding interaction between the sulfur atom of the thiourea moiety and the hydrogen at the 2-position of the aromatic group fixes the position of one of the aromatic rings, that is, it fixes the position of the other aromatic ring to the β-face, which is responsible for some interactions with the nitrostyrene moiety. These interactions give rise to the superior performance of 9a compared with the Takemoto catalyst in terms of enantioselectivity. This proposed transition state explains the slightly lower enantioselectivity of 13b than 13a, as shown in Figure [2]. Whereas a small hydrogen atom was located on the β-face in the case of 13a, a methyl group, which can cause a steric repulsion with a substrate, occupied the same position when 13b was employed and, consequently, a lower enantioselectivity was observed.
Figure 3 Proposed transition-state model
The superior catalytic performance of 9a compared with the Takemoto catalyst was not limited to the asymmetric Michael reaction, as shown in Scheme [3]. In the asymmetric Strecker reaction of trifluoromethyl ketimine 15,[20] the asymmetric amination of the β-keto ester 17,[21] and the asymmetric Friedel–Crafts reaction/lactonization reaction sequence of the benzylidene Meldrum’s acid 20 with 2-naphthol (21),[22] our catalyst 9a exhibited a good performance, and the corresponding adducts were obtained with higher enantioselectivities than those obtained when the Takemoto catalyst was used.
Scheme 3 Examination of the catalytic performance of 9a in other asymmetric reactions
In summary, we have evaluated the electronic effects of the aromatic rings in the diarylmethylamine unit, namely, the butterfly-type amine unit, on the performance of aminothiourea catalysts.[23] Among the aromatic rings examined, the 3,5-bis(trifluoromethyl)phenyl group was the best; an excellent enantioselectivity was realized in the asymmetric Michael reaction between nitrostyrene and dimethyl malonate. It is worth noting that its asymmetry-inducing ability was superior to that of the well-known aminothiourea catalyst, the Takemoto catalyst, and this characteristic was observed in various asymmetric reactions. These catalysts could provide a trump card for the improvement of enantioselectivity in aminothiourea-catalyzed asymmetric reactions. The development of novel asymmetric reactions with these catalysts is underway in our laboratory.
