Nickel/Brønsted Acid-Catalyzed Chemo- and Enantioselective Intermolecular Hydroamination of Conjugated Dienes.

A novel nickel/Brønsted acid-catalyzed asymmetric hydroamination of acyclic 1,3-dienes has been established. A wide array of primary and secondary amines can be transformed into allylic amines with high yields and high enantioselectivities under very mild conditions. Moreover, our method is compatible with various functional groups and heterocycles, allowing for late-stage functionalization of biologically active complex molecules. Remarkably, this protocol exhibits good chemoselectivity with respect to amines bearing two different nucleophilic sites. Mechanistic studies reveal that the enantioselective carbon-nitrogen bond-forming step is reversible.

Introduction

Chiral class="Chemical">amines represeclass="Chemical">nt a privileged pharmacophore aclass="Chemical">nd are preseclass="Chemical">nt iclass="Chemical">n a myriad of class="Chemical">natural products aclass="Chemical">nd drugs (Figure 1A) (Fraclass="Chemical">ncotte aclass="Chemical">nd Liclass="Chemical">ndclass="Chemical">ner, 2006, Lough aclass="Chemical">nd Waiclass="Chemical">ner, 2002, class="Chemical">n class="Chemical">Nugent, 2010). Therefore, organic chemists have made considerable efforts toward their synthesis during the last decade (Grogan, 2018, Li and Zhang, 2014, Nugent and El-Shazly, 2010, Patil et al., 2018, Robak et al., 2010). Among them, asymmetric hydroamination of unsaturated C-C bonds serves as an efficient and powerful tool in organic synthesis, particularly hydroamination using free amines (Aillaud et al., 2007, Clement and Jerome, 2017, Dondoni, 2015, Hannedouche and Schulz, 2013, Hannedouche and Schulz, 2018, Hii, 2006, Huang et al., 2015, Huo et al., 2019, Jerome, 2018, Müller et al., 2008, Patel et al., 2017, Pirnot et al., 2016, Reznichenko and Hultzsch, 2016, Zi, 2009, Zi, 2011). In this context, transition-metal-catalyzed intermolecular asymmetric hydroamination of allenes (Berthold and Breit, 2018, Berthold et al., 2019, Cooke et al., 2012, Dion and Beauchemin, 2011, Lin et al., 2019, Parveen et al., 2017, Xu et al., 2016), alkynes (Athira et al., 2018, Liu et al., 2011, Lutete et al., 2004, Patil et al., 2006, Xu et al., 2019), and conjugated dienes (Adamson et al., 2017, Dion and Beauchemin, 2011, Lin et al., 2019, Löber et al., 2001, Park and Malcolmson, 2018, Xiong et al., 2018, Yang and Dong, 2017, Zhou and Hartwig, 2008) has been extensively studied (Figure 1B). Nevertheless, the use of noble transition metals such as rhodium and palladium are often mandatory (Adamson et al., 2017, Aillaud et al., 2007, Athira et al., 2018, Berthold et al., 2019, Berthold and Breit, 2018, Clement and Jerome, 2017, Cooke et al., 2012, Dion and Beauchemin, 2011, Dondoni, 2015, Hannedouche and Schulz, 2013, Hannedouche and Schulz, 2018, Hii, 2006, Huang et al., 2015, Huo et al., 2019, Jerome, 2018, Lin et al., 2019, Liu et al., 2011, Löber et al., 2001, Lutete et al., 2004, Müller et al., 2008, Park and Malcolmson, 2018, Parveen et al., 2017, Patel et al., 2017, Patil et al., 2006, Pirnot et al., 2016, Reznichenko and Hultzsch, 2016, Xiong et al., 2018, Xu et al., 2016, Xu et al., 2019, Yang and Dong, 2017, Zhou and Hartwig, 2008, Zi, 2009, Zi, 2011); in addition, these methods suffer from limited amine scope (Adamson et al., 2017, Dion and Beauchemin, 2011, Lin et al., 2019, Löber et al., 2001, Park and Malcolmson, 2018, Xiong et al., 2018, Yang and Dong, 2017, Zhou and Hartwig, 2008), as well as excessive quantities of the unsaturated substrate are always required to achieve a high level of efficiency (Adamson et al., 2017, Dion and Beauchemin, 2011, Lin et al., 2019, Löber et al., 2001, Park and Malcolmson, 2018, Yang and Dong, 2017, Zhou and Hartwig, 2008).

Figure 1

Reaction Design

(A) Representative drugs containing chiral amines.

(B) Toward chiral allylic amines by asymmetric intermolecular hydroamination.

(C) Ni-catalyzed asymmetric hydrofunctionalization.

(D) Nickel/Brønsted acid-catalyzed chemo- and enantioselective intermolecular hydroamination of conjugated dienes.

(A) Representative drugs containing chiral n class="Chemical">amines.

(B) Toward chiral n class="Chemical">allylic amines by asymmetric iclass="Chemical">ntermolecular hydroamiclass="Chemical">natioclass="Chemical">n.

(C) n class="Chemical">Ni-catalyzed asymmetric hydrofuclass="Chemical">nctioclass="Chemical">nalizatioclass="Chemical">n.

(D) n class="Chemical">Nickel/Brøclass="Chemical">nsted acid-catalyzed chemo- aclass="Chemical">nd eclass="Chemical">naclass="Chemical">ntioselective iclass="Chemical">ntermolecular hydroamiclass="Chemical">natioclass="Chemical">n of coclass="Chemical">njugated class="Chemical">n class="Chemical">dienes.

In recent years, research toward class="Chemical">nickel-catalyzed oxidative additioclass="Chemical">n with X-H (X = C, O …) boclass="Chemical">nds has become a hot theme owiclass="Chemical">ng to earth-abuclass="Chemical">ndaclass="Chemical">nce of class="Chemical">n class="Chemical">nickel and its great potential in oxidative addition (Ananikov, 2015, Tasker et al., 2014, Wang, 2016; Figure 1C). Significant progress has been made in the asymmetric hydrofunctionalization of alkenes through nickel-catalyzed reactions (Bezzenine-Lafollee et al., 2017, Cai et al., 2019, Chen and Lu, 2018, Cheng et al., 2018, Cheng et al., 2019, Diesel et al., 2018, Diesel et al., 2019, Donets and Cramer, 2013, Li et al., 2018, Li et al., 2019a, Lv et al., 2018, Richmond and Moran, 2018, Woźniak and Cramer, 2019, Xiao et al., 2016, Xiao et al., 2018, Zhang et al., 2019). Chiral centers are generally induced via a carbon-carbon bond-forming process, involving the direct oxidative addition of C-H bonds (Cai et al., 2019, Cheng et al., 2018, Cheng et al., 2019, Diesel et al., 2018, Diesel et al., 2019, Donets and Cramer, 2013, Li et al., 2019a, Lv et al., 2018, Woźniak and Cramer, 2019, Zhang et al., 2019) or an external stoichiometric reductant, such as alcohol (Chen et al., 2019) or hydrosiloxane (Ahlin and Cramer, 2016). However, nickel-catalyzed asymmetric hydrofunctionalization of unsaturated compounds involving a carbon-heteroatom bond formation has not been studied much (Tran et al., 2019). As an extension of our studies with nickel-catalyzed carbon-carbon bond formations (Li et al., 2019b, Wang et al., 2019), we turned our attention to carbon-heteroatom bonds. Inspired by the recent reports on metal/Brønsted acid dual catalysis (Adamson et al., 2017, Dion and Beauchemin, 2011, Han et al., 2018, Kathe and Fleischer, 2019, Lin et al., 2019, Liu and Feng, 2018, Löber et al., 2001, Park and Malcolmson, 2018, Yang and Dong, 2017, Zhou and Hartwig, 2008), we have developed a novel, room temperature nickel/Brønsted acid-catalyzed asymmetric hydroamination using conjugated dienes as a limiting reagent (Figure 1D). This protocol can transform a wide array of primary and secondary amines into allylic amines in high yields with excellent enantioselectivities. Significantly, good regio-, chemo-, and enantioselectivity have been achieved using amines bearing potentially competitive nucleophilic sites. It is noteworthy that the nickel-catalyzed racemic hydroamination of cyclic dienes has only been reported by the Hartwig group before, wherein they also demonstrated the challenge for the development of an enantioselective variant (Pawlas et al., 2002).

Results

Optimization Reaction Conditions

We initiated this study by choosing class="Chemical">phenyl-1,3-diene (1a) aclass="Chemical">nd class="Chemical">n class="Chemical">morpholine (2a) as model substrates. Ligand evaluations were conducted using Ni(COD)2 as the precatalyst and TsOH⋅H2O as a cocatalyst. As shown in Figure 2, a series of bisphosphine ligands were examined; the 1,2-hydroamination product 3a (Wang et al., 2014) was obtained in a moderate yield with a low enantiomeric excess (ee) when chiral BINAP (L1) or SEGPHOS (L2) was used, which demonstrated the feasibility of this hydroamination reaction. Unfortunately, (S)-SKP (L3), (R)-SDP (L4), and (R)-DIOP (L5) as ligand were not effective for this transformation, although (S,S)-BDPP (L6), a flexible bisphosphine ligand, yielded 3a in an excellent yield, but with low enantioselectivity (23% ee). However, both high yields and enantioselectivities were achieved by (R,S)-DuanPhos (L7). To our delight, excellent ee (95% ee) was obtained when (S,S)-Me-DuPhos (L8), as a more rigid ligand, was used. In addition, the Brønsted acid cocatalyst can also affect the efficiency and enantioselectivity of this hydroamination reaction. Further studies demonstrated that the desired product can also be obtained in high yields without a decrease in enantioselectivity when switching the acid cocatalyst to phenylphosphonic acid (A3) or phthalic acid (A4). To easily weighout, we selected A4 as cocatalyst. Moreover, control experiments indicated that both nickel catalysts and the Brønsted acids were crucial to the success of this reaction. Notably, no other regioisomers were detected in these reactions.

Figure 2

Reaction Optimization

Reactions were conducted at 0.2 mmol scale, see Supplemental Information for reaction details. See also Tables S1–S3.

Substrate Scope Study

With the optimal conditions in hand, we shifted our attention to investigate the generality of this class="Chemical">Ni-catalyzed asymmetric hydroamiclass="Chemical">natioclass="Chemical">n reactioclass="Chemical">n. Utiliziclass="Chemical">ng 1a, we exclass="Chemical">n class="Chemical">amined the scope of the amines. As illustrated in Figure 3, a series of primary amines bearing various functional groups produced the corresponding hydroamination products 3b-3l with good to excellent yields with excellent enantioselectivities. Notably, (R)-(+)-1-Phenylethylamine, a chiral amine, also gave the hydroamination product in a moderate yield with an excellent diastereomeric ratio (dr > 20:1, 3m). In addition to the aliphatic amines, primary arylamines were also suitable for the reaction to generate the chiral amine products with excellent enantioselectivities, albeit in lower yields under the current reaction conditions. It is noteworthy that the aryl bromide is compatible with this nickel-catalyzed reaction (3p). To assess the practicality of this asymmetric hydroamination reaction, a gram-scale experiment was conducted. When the reaction of 1a with 2g was performed on a 5 mmol scale, it still was able to furnish 3g without loss of reaction efficiency and optical enantioselectivities, even in the presence of 1 mol % catalysts.

Figure 3

Scope of Primary and Secondary Amines

Reactions were conducted at 0.2 mmol scale, see Supplemental Information for reaction condition details. aReactions were conducted at 5 mmol scale. b12 h; c36 h; d48 h. See also Scheme S3.

Scope of Primary and n class="Chemical">Secondary Amines

class="Chemical">Next, the scope of class="Chemical">n class="Chemical">secondary amines was tested. Various secondary cyclic amines afforded the chiral allylic amines in both remarkable yields and enantioselectivities (3a-3v). Moreover, acyclic secondary amines were also able to produce the desired hydroamination products with excellent enantioselectivities under the same reaction conditions (3w-3aa). Interestingly, although catalytic amount of Brønsted acid was used as a cocatalyst, amines containing other nitrogen atoms did not affect this asymmetric transformation (3j and 3v). Additionally, a series of functional groups, including ethers (3i and 3a), esters (3l), thioethers (3q), terminal alkenes (3h and 3w), and heterocycles such as furan (3f) and pyrimidines (3v), all were well tolerated in this reaction.

Subsequently, the scope of class="Chemical">1,3-dienes was studied. A set of class="Chemical">n class="Chemical">aryl-substituted linear 1,3-butadienes were examined with both primary and secondary amines under the optimal conditions. As shown in Figure 4, both electron-rich and deficient substituents did not affect the efficiency or enantioselectivity. Alkyl-substituted butadienes were also capable of producing the Markovnikov hydroamination products (3ai, 3aj, 3ar, 3as, and 3at) in excellent yields with an excellent ee value. Notably, no other regioisomers were detected in these reactions. Furthermore, the hydroamination product (3au) could also be synthesized from 1,3-cyclohexadiene, albeit in low yields and enantioselectivity under the current conditions.

Figure 4

Scope of Conjugated Dienes

Reactions were conducted at 0.2 mmol scale, see Supplemental Information for reaction condition details. See also Scheme S3.

Scope of Conjugated n class="Chemical">Dienes

As we have highlighted earlier, both primary and secondary class="Chemical">alkyl and aryl amines caclass="Chemical">n produce satisfactory results iclass="Chemical">n this class="Chemical">n class="Chemical">nickel/Brønsted acid-catalyzed reaction. We were curious about the chemoselectivity when using one substrate containing two different nucleophilic sites. Guided by this idea, a set of more complex amines were tested under the optimal conditions and the results have been displayed in Figure 5. With aminoethanol, only the 1,2-hydroamination product (3av) was isolated with an excellent yield and ee value. Notably, the less sterically encumbered primary amine was found to be more reactive than the secondary amine when N-benzylethylediamine was used (3aw). Interestingly, the acidic phenol did not affect the amination (3ax), and the hydroamination reaction of the aryl amine (3ay) was not affected by the presence of an alcohol. Moreover, a single isomer with both excellent ee and yield could be obtained from tryptamine (3az). Finally, high chemoselectivity was shown at the aliphatic amine part when 4-aminobenzylamine was used (3ba). Collectively, these results suggest that this nickel-catalyzed reaction exhibits good chemoselectivity toward hydroamination and also demonstrates the potential of this method in the late-stage diversification of biomolecules.

Figure 5

Substrates Containing Two Nucleophilic Sites

Reactions were conducted at 0.2 mmol scale, see Supplemental Information for reaction condition details. See also Scheme S3.

Substrates Containing Two n class="Chemical">Nucleophilic Sites

Discussion

Mechanism Study

To get more details of this transformation, a preliminary mechanistic investigation was conducted. In Hartwig's reaction, a reversible class="Chemical">carbon-nitrogen boclass="Chemical">nd formatioclass="Chemical">n was observed. To determiclass="Chemical">ne if this pheclass="Chemical">nomeclass="Chemical">noclass="Chemical">n also exists iclass="Chemical">n our reactioclass="Chemical">n, class="Chemical">n class="Chemical">amine exchange experiments were performed first. When the enantioenriched 3t and stoichiometric morpholine were subjected to the optimal conditions, both 3t and 3a were detected (Scheme 1-1). A similar phenomenon was also observed in the reaction of 3t with a primary amine (Scheme 1-2). This reversible effect was also found when a primary amine-based product was used (Schemes 1-3 and 1-4). These findings strongly suggested that a reversibility of carbon-nitrogen bond formation was involved in this reaction. These results are in consistence with Hartwig's results (Pawlas et al., 2002) but inconsistent with the results of Mazet's conditions (Tran et al., 2019).

Scheme 1

Amine Exchange Experiment

(1) Exchange experiment of secondary amine-based product (3t) with secondary amine (morpholine).

(2) Exchange experiment of secondary amine-based product (3t) with primary amine (furfuryl amine).

(3) Exchange experiment of primary amine-based product (3k) with secondary amine (morpholine).

(4) Exchange experiment of primary amine-based product (3k) with primary amine (furfuryl amine).

Data are represented as mean value of three times; see also Scheme S5.

n class="Chemical">Amine Exchaclass="Chemical">nge Experimeclass="Chemical">nt

(1) Exchange experiment of secondary class="Chemical">amine-based product (3t) with secoclass="Chemical">ndary class="Chemical">n class="Chemical">amine (morpholine).

(2) Exchange experiment of secondary class="Chemical">amine-based product (3t) with primary class="Chemical">n class="Chemical">amine (furfuryl amine).

(3) Exchange experiment of primary class="Chemical">amine-based product (3k) with secoclass="Chemical">ndary class="Chemical">n class="Chemical">amine (morpholine).

(4) Exchange experiment of primary class="Chemical">amine-based product (3k) with primary class="Chemical">n class="Chemical">amine (furfuryl amine).

Furthermore, a decrease in enantioselectivity over time has been observed in the class="Chemical">palladium-catalyzed hydroamiclass="Chemical">natioclass="Chemical">n reactioclass="Chemical">ns (Löber et al., 2001, Pawlas et al., 2002). To determiclass="Chemical">ne if this pheclass="Chemical">nomeclass="Chemical">noclass="Chemical">n also exists iclass="Chemical">n our reactioclass="Chemical">n, time course experimeclass="Chemical">nts were coclass="Chemical">nducted for both primary aclass="Chemical">nd class="Chemical">n class="Chemical">secondary amines (Figure 6). To our surprise, significant racemization was observed for the reaction with a secondary amine (Figure 6A), whereas there was nearly no alteration of enantioselectivity in a reaction with a primary amine (Figure 6B). Moreover, similar results were also obtained switching A4 to A3.

Figure 6

Reaction Profiles

(A) Time course experiments of secondary amine.

(B) Time course experiments of primary amine.

Data are represented as mean value of three times; see also Scheme S6 and Figure S246.

(A) Time course experiments of secondary n class="Chemical">amine.

(B) Time course experiments of primary n class="Chemical">amine.

Finally, based on precedent studies (Adamson et al., 2017, Dion and Beauchemin, 2011, Lin et al., 2019, Löber et al., 2001, Park and Malcolmson, 2018, Xiong et al., 2018, Yang and Dong, 2017, Zhou and Hartwig, 2008) and the above-mentioned findings (see Supplemental Information for more results), a mechanistic profile is proposed for this transformation. As illustrated in Scheme 2, the reaction is initiated by a class="Chemical">Ni(0) species (I), which uclass="Chemical">ndergoes oxidative additioclass="Chemical">n to form a class="Chemical">n class="Chemical">Ni(II)-H species (II). Subsequently, a 1,3-diene migratory insertion leads to the formation of a π-allylNi(II) intermediate (III). The hydroamination product 3 is ultimately generated from the π-allylNi(II) complex by an amine nucleophilic attack (McDonald et al., 2011), accompanied by releasing of a Ni(0) species and regeneration of the acid cocatalyst.

Scheme 2

Proposed Mechanism

Conclusion

In summary, we have developed a novel class="Chemical">nickel aclass="Chemical">nd Brøclass="Chemical">nsted acid-cocatalyzed asymmetric hydroamiclass="Chemical">natioclass="Chemical">n reactioclass="Chemical">n. The choice of chiral class="Chemical">n class="Chemical">bisphosphine ligand and the use of a suitable Brønsted acid in catalytic amount are crucial to the success of this transformation. This protocol allows access to a series of enantiopure secondary and tertiary allylic amines from linear conjugated dienes and free amines. This method provides high enantioselectivity and a broad substrate scope for the synthesis of various chiral amines. Importantly, a set of complex amines have been accomplished with excellent chemo- and enantioselectivity in this system. The good functional group tolerance and the scalability demonstrates the potential of this method in the synthesis of enantiopure amines. Mechanistic studies indicate that the C-N bond formation is a reversible step. Moreover, racemization over time exists in the reaction with secondary amines but not for primary amines. We believe this chemistry will greatly benefit medicinal chemistry and further reaction development.

Limitations of the Study

The disubstituted n class="Chemical">diene was class="Chemical">not suitable iclass="Chemical">n this methodology.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.