Building Up a Piperazine Ring from a Primary Amino Group via Catalytic Reductive Cyclization of Dioximes

Piperazine is one of the most frequently found scaffolds in small-molecule FDA-approved drugs. In this study, a general approach to the synthesis of piperazines bearing substituents at carbon and nitrogen atoms utilizing primary amines and nitrosoalkenes as synthons was developed. The method relies on sequential double Michael addition of nitrosoalkenes to amines to give bis(oximinoalkyl)amines, followed by stereoselective catalytic reductive cyclization of the oxime groups. The method that we developed allows a straightforward structural modification of bioactive molecules (e.g., α-amino acids) by the conversion of a primary amino group into a piperazine ring.

Given the importance of piperazines, numerous synthetic approaches have been developed to access the broad chemical space of these heterocycles [11,12]. A common strategy relies on the functionalization of the parent molecule, which is easily accomplished by the addition of electrophiles at N-atoms and cross-coupling reactions [13][14][15][16]. Direct functionalization at C-atoms in piperazine is challenging (although some progress has been achieved in recent years [17,18]), and the synthesis of carbon-substituted piperazines is often accomplished via cyclization of the corresponding linear diamine precursors [11,12]. Alternative strategies employing the hydrogenation of pyrazines [19], [3+3]-type dimerization of aziridines [20], ring-opening reactions in DABCO derivatives [21], and ring expansion reactions in imidazolines [22] are limited by the specific structure of the starting materials. Overall, the synthesis of polysubstituted piperazines still remains a challenging problem.
For drug design purposes, the construction of piperazine ring 1 from a primary amino group is a useful synthetic strategy (Scheme 1). Since a plethora of pharmaceuticals and natural products contain a primary amino group, this synthetic tool would generate their potentially potent piperazine modifications [23]. Several attempts were undertaken to develop this methodology using 1,5-dihaloamines, diethanolamines, and their derivatives as cyclizing agents (Scheme 1). For example, Watanabe et al. [24] reported a hydrogen borrowing strategy to assemble piperazines from simple amines and N-substituted diethanolamines, yet the yields of the products were low in many cases. Huang and Li et al. [25] developed a one-pot protocol for the conversion of primary amines into piperazines by cyclization 2 of 14 with tosylbis(2-(tosyloxy)ethyl)amine. However, these methods were applied only to the synthesis of C-unsubstituted piperazine derivatives, most likely due to a limited availability of the corresponding substituted diethanolamine derivatives and their reduced reactivity in these reactions. Also, harsh reaction conditions are required, under which complex substrates may not be tolerated. To develop a more general method, we suggested a conceptually different strategy (Scheme 1), which involves the conversion of a primary amine into bis(oximinoalkyl)amine 2 through double Michael addition to nitrosoalkenes (NSA), followed by catalytic reductive cyclization of the dioxime unit to give a piperazine ring. In our recent work, we have shown that a related cyclization of dioximes of 1,5-diones provides an expedient route to piperidines [26].
and their derivatives as cyclizing agents (Scheme 1). For example, Watanabe et al. [24] reported a hydrogen borrowing strategy to assemble piperazines from simple amines and N-substituted diethanolamines, yet the yields of the products were low in many cases. Huang and Li et al. [25] developed a one-pot protocol for the conversion of primary amines into piperazines by cyclization with tosylbis(2-(tosyloxy)ethyl)amine. However, these methods were applied only to the synthesis of C-unsubstituted piperazine derivatives, most likely due to a limited availability of the corresponding substituted diethanolamine derivatives and their reduced reactivity in these reactions. Also, harsh reaction conditions are required, under which complex substrates may not be tolerated. To develop a more general method, we suggested a conceptually different strategy (Scheme 1), which involves the conversion of a primary amine into bis(oximinoalkyl)amine 2 through double Michael addition to nitrosoalkenes (NSA), followed by catalytic reductive cyclization of the dioxime unit to give a piperazine ring. In our recent work, we have shown that a related cyclization of dioximes of 1,5-diones provides an expedient route to piperidines [26]. Scheme 1. Background of this study and the current work [24,25].

Results and Discussion
The first step of the suggested sequence requires the double oximinoalkylation of a primary amine through a Michael reaction with nitrosoalkenes NSA. Due to the instability of NSA, suitable precursors are used to generate these species in situ [27][28][29]. In the previous studies by our group [26,30] and other researchers [31], it has been shown that silylated ene-nitrosoacetals 3 are convenient precursors of NSA providing high chemoselectivity when coupling with various types of nucleophiles, including amines [32][33][34]. Using this methodology, a series of symmetrically substituted dialdo-and diketooximes 2a−o was prepared by the treatment of simple primary amines with 2.1 equiv. of the corresponding ene-nitrosoacetals 3a−c (R 1 = H, Me, Ph, respectively, Scheme 2), followed by methanolysis of the resulting O-silyl ethers. In most cases, dioximes 2 were formed in high yields, except for sterically hindered amines (tert-butylamine and α-phenylethylamine leading to products 2b,k) and α-amino acid esters (products 2g,h). In the latter case, the lower efficiency is likely due to the self-dimerization of α-amino acid esters to diketopiperazines as a side process. Also, in the reaction with L-leucine ethyl ester, a noticeable amount of a mono-addition product was obtained.

Results and Discussion
The first step of the suggested sequence requires the double oximinoalkylation of a primary amine through a Michael reaction with nitrosoalkenes NSA. Due to the instability of NSA, suitable precursors are used to generate these species in situ [27][28][29]. In the previous studies by our group [26,30] and other researchers [31], it has been shown that silylated ene-nitrosoacetals 3 are convenient precursors of NSA providing high chemoselectivity when coupling with various types of nucleophiles, including amines [32][33][34]. Using this methodology, a series of symmetrically substituted dialdo-and diketooximes 2a-o was prepared by the treatment of simple primary amines with 2.1 equiv. of the corresponding ene-nitrosoacetals 3a-c (R 1 = H, Me, Ph, respectively, Scheme 2), followed by methanolysis of the resulting O-silyl ethers. In most cases, dioximes 2 were formed in high yields, except for sterically hindered amines (tert-butylamine and α-phenylethylamine leading to products 2b,k) and α-amino acid esters (products 2g,h). In the latter case, the lower efficiency is likely due to the self-dimerization of α-amino acid esters to diketopiperazines as a side process. Also, in the reaction with L-leucine ethyl ester, a noticeable amount of a mono-addition product was obtained.
for diketooximes, the 2i−n E,E-isomer was often predominant. The stereochemical assignment of oxime groups was performed on the basis of known relationships between the configuration of the C=N bond and chemical shifts of neighboring atoms in 1 H and 13 C NMR spectra (see Supplementary Materials for details on stereochemistry elucidation) [33,34]. The assigned configuration was additionally confirmed by 2D 1 H− 1 H NOESY correlations and 1 JCH coupling constants in the C(N)H unit [35] for dioxime 2a.
Dioximes 2a−o were then subjected to catalytic hydrogenation conditions (Scheme 2). Two heterogeneous catalysts, namely palladium on charcoal (5%-Pd/C) and Raney ® nickel (Ra-Ni), were chosen for this study since these cheap and readily available catalysts often show the best performance in the hydrogenation of oximes [26,36]. It was found that hydrogenation of a model dialdooxime 2a gave the desired piperazine 1a in 44% yield with 5%-Pd/C catalyst at 40 bar H2/50 °C (method A1), while in the reaction with Ra-Ni (method B1), a complex mixture of products formed. The protection of piperazine 1a with a Boc group during the hydrogenation was advantageous in terms of Scheme 2. Assembly of C-unsubstituted and 2,6-disubstituted piperazines 1 from primary amines via double oximinoalkylation followed by reductive cyclization of dioximes 2. For 2,6-disubstituted piperazines (1i−o series), the structure of the major diastereomer is shown. The yields and diastereomeric ratios (d.r.) refer to the isolated products. Ene-nitrosoacetals 3: 3a Typically, dioximes 2 were obtained and characterized as mixtures of E/Z-isomers (the isomeric ratio slightly changed with time and depended on the solvent). For dialdooximes 2a-h, E,Eand E,Z-isomers were formed in comparable quantities, while for diketooximes, the 2i-n E,E-isomer was often predominant. The stereochemical assignment of oxime groups was performed on the basis of known relationships between the configuration of the C=N bond and chemical shifts of neighboring atoms in 1 H and 13 C NMR spectra (see Supplementary Materials for details on stereochemistry elucidation) [33,34]. The assigned configuration was additionally confirmed by 2D 1 H− 1 H NOESY correlations and 1 J CH coupling constants in the C(N)H unit [35] for dioxime 2a.
Dioximes 2a-o were then subjected to catalytic hydrogenation conditions (Scheme 2). Two heterogeneous catalysts, namely palladium on charcoal (5%-Pd/C) and Raney ® nickel (Ra-Ni), were chosen for this study since these cheap and readily available catalysts often show the best performance in the hydrogenation of oximes [26,36]. It was found that hydrogenation of a model dialdooxime 2a gave the desired piperazine 1a in 44% yield with 5%-Pd/C catalyst at 40 bar H 2 /50 • C (method A1), while in the reaction with Ra-Ni (method B1), a complex mixture of products formed. The protection of piperazine 1a with a Boc group during the hydrogenation was advantageous in terms of the product yield and isolation simplicity (cf. yields of products 1a and Boc-1a). Optimized conditions (method A2) were successfully applied to a series of dialdooximes 2a-h that gave the corresponding Boc-piperazines Boc-1 in moderate to good yields. Expectedly, the alkene moiety in dioxime 2d and the benzylic C−N bond in dioximes 2e,f was not tolerated under hydrogenation conditions with Pd/C, leading to piperazines Boc-1d' and Boc-1e'. The reductive debenzylation process was suppressed by using the Ra-Ni catalyst, as demonstrated by a synthesis of chiral piperazine Boc-1f from dioxime 2f derived from α-phenylethylamine by method B2.
Importantly, the developed approach allowed the building up of a piperazine ring in amino acid derivatives, as shown for glycine and L-leucine esters (products Boc-1g and Boc-1h) as examples. We believe that piperazine-modified peptides can be prepared in a similar manner.
In contrast to dialdooximes 2a−h, diketooximes 2i−o underwent the desired reductive cyclization with Ra-Ni catalyst (Scheme 2), while poor conversion was observed with 5%-Pd/C (studies with model substrate 2i). Moreover, protection with Boc 2 O (method B2) was inefficient in this case due to the small reaction rate attributed to a sterically encumbered environment around the nitrogen atom in 2,6-disubstituted piperazines. Thus, protection with more reactive propionic anhydride (method B3) or no protection (method B1) was performed in the hydrogenation of diketooximes 2i−o. Using these procedures, a series of 2,6-disubstituted piperazines 1 or their N-propionyl derivatives EtCO-1 was successfully prepared (Scheme 2). Note that the benzylic C−N bonds remained intact in products EtCO-1k,m, and 1o.
In the reductive cyclization of diketooximes 2i−o, cis-isomers of the corresponding piperazines 1 formed predominantly. In some cases, the process was stereospecific, and only 2,6-cis-isomers were obtained. Note that 2,6-diaryl-substituted piperazines of type 1o have scarcely been described in the literature, and those which have been reported have the 2,6-trans-configuration [18]. So far, the reductive cyclization of dioximes 2 appears to be the only available synthetic route to cis-2,6-diaryl-piperazines [37]. The comparison of NMR spectra of the obtained piperazines with those previously reported for cis-1n [38] and trans-1o [18] confirmed the stereochemical assignment (see Supplementary Materials for details on stereochemistry elucidation).
In the next stage, we aimed to adapt the developed strategy to access unsymmetrical 2-substituted piperazines (Scheme 3a). To accomplish this, a sequential assembly of unsymmetrically substituted dioximes 2 from two different NSA precursors was required. However, the reaction of n-butylamine with an equimolar amount of NSA precursor 3a gave a complex mixture containing products of mono-and bis-addition (4a and 2a, respectively). Selective Michael addition of one equivalent of nitrosoethylene was accomplished only when a significant excess of n-butylamine (12 equiv.) was employed. The monooxime 4a obtained was then used in a reaction with a second NSA precursor 3b,d,e that afforded unsymmetrically substituted dioximes 2p−r. Note that changing the sequence of the addition of NSA residues (i.e., initial addition of 3b to n BuNH 2 followed by addition of 3a) was found to produce dioxime 2p at a lower overall yield. The reductive cyclization of these dioximes was successfully accomplished by hydrogenation with a 5%-Pd/C catalyst according to method A2 to give the corresponding racemic 2-substituted N-Boc-protected piperazines Boc-1p−r in good yields. Subsequent lactamization of the transient γ-aminoester upon heating in toluene produced the desired bicyclic product 5.
The proposed mechanism for the reductive cyclization of dioximes 2 into piperazines 1 is depicted in Scheme 4 [26]. The key stages involved catalytic hydrogenolysis of both N−O bonds to give diimine intermediate I-1, followed by its cyclization to give dihydropyrazine I-2. Subsequent hydrogenation of the C=N bond in I-2, elimination of ammonia, and reduction of dihydropyrazine I-3 afforded the final piperazine product 1. The observed predominant formation of 2,6-cis-isomers of piperazines 1 can be explained by the addition of dihydrogen from the less hindered side of the C=N bond in I-3, opposite to substituent R 2 .

Scheme 4.
Proposed mechanism for the conversion of dioximes 2 into piperazines 1.

Materials and Methods
Full compound characterization, detailed synthetic procedures, and copies of NMR spectra are provided in Supplementary Material. To showcase the utility of the developed strategy, the synthesis of a fused piperazine derivative, namely 1,4-diazabicyclo [4. 3.0]nonane 5, was accomplished (Scheme 3b). The titled compound represents a building block in the synthesis of numerous pharmaceutically relevant molecules, in particular, hNK 1 antagonist Orvepitant [39][40][41][42]. Dioxime 2s was prepared by the sequential addition of NSA precursors 3a and 3f to excess benzylamine (the major fraction of benzylamine that did not react in the first stage could be regenerated).
In the hydrogenation of dioxime 2s over 5%-Pd/C (method A1), reductive cyclization of the dioxime motif and catalytic debenzylation occurred. Subsequent lactamization of the transient γ-aminoester upon heating in toluene produced the desired bicyclic product 5.
The proposed mechanism for the reductive cyclization of dioximes 2 into piperazines 1 is depicted in Scheme 4 [26]. The key stages involved catalytic hydrogenolysis of both N−O bonds to give diimine intermediate I-1, followed by its cyclization to give dihydropyrazine I-2. Subsequent hydrogenation of the C=N bond in I-2, elimination of ammonia, and reduction of dihydropyrazine I-3 afforded the final piperazine product 1. The observed predominant formation of 2,6-cis-isomers of piperazines 1 can be explained by the addition of dihydrogen from the less hindered side of the C=N bond in I-3, opposite to substituent R 2 .
The proposed mechanism for the reductive cyclization of dioximes 2 into piperazines 1 is depicted in Scheme 4 [26]. The key stages involved catalytic hydrogenolysis of both N−O bonds to give diimine intermediate I-1, followed by its cyclization to give dihydropyrazine I-2. Subsequent hydrogenation of the C=N bond in I-2, elimination of ammonia, and reduction of dihydropyrazine I-3 afforded the final piperazine product 1. The observed predominant formation of 2,6-cis-isomers of piperazines 1 can be explained by the addition of dihydrogen from the less hindered side of the C=N bond in I-3, opposite to substituent R 2 .

Scheme 4.
Proposed mechanism for the conversion of dioximes 2 into piperazines 1.

Materials and Methods
Full compound characterization, detailed synthetic procedures, and copies of NMR spectra are provided in Supplementary Material.

Materials and Methods
Full compound characterization, detailed synthetic procedures, and copies of NMR spectra are provided in Supplementary Materials.

General Procedures
General procedure for the synthesis of symmetrically substituted dioximes 2a−m,o. A solution of ene-nitrosoacetal 5 (2.1 mL, 1M in CH 2 Cl 2 ) was added dropwise to a solution of an amine (1 mmol) in CH 2 Cl 2 (1 mL), and the mixture was vigorously stirred at room temperature for 24 h. Then, MeOH (2 mL) was added to the reaction mixture and vigorously stirred for 8 h. Then, the reaction mixture was concentrated under reduced pressure, and the residue was subjected to column chromatography on silica gel.
General procedure for the synthesis of monooximes 4. Amine (10-14 equiv.) was added to a solution of ene-nitrosoacetal 5 in dichloromethane (1 equiv, 1 M in CH 2 Cl 2 ), and the mixture was stirred at room temperature for 24 h. Then, MeOH (5 mL) was added to the mixture and stirred at room temperature for 8 h. The reaction mixture was concentrated under reduced pressure, and the residue was subjected to column chromatography on silica gel (PE:EtOAc = 5:1 → 3:1 → 1:1).
Synthesis of unsymmetrically substituted dioximes 2p−2s. Ene-nitrosoacetal 5 (0.6 mL, 1M in CH 2 Cl 2 ) was added dropwise to a solution of monooxime 4 (0.5 mmol) in CH 2 Cl 2 (1 mL), and the mixture was vigorously stirred at room temperature for 24 h. Then, MeOH (2 mL) was added to the reaction mixture and vigorously stirred for 8 h. Then, the reaction mixture was concentrated under reduced pressure, and the residue was subjected to column chromatography on silica gel.
General procedure for the synthesis of free piperazines 1 with 5%-Pd/C catalyst (method A1). A 5%-Pd/C catalyst (50 mg per 0.5 mmol of 2) was added to a solution of dioxime 2 (1 equiv.) in methanol (0.1 M). The vial was placed in a steel autoclave which was flushed and filled with hydrogen to a pressure of ca. 40 bar. Hydrogenation was conducted at this pressure and 50 • C for 6 h with vigorous stirring. Then, the autoclave was cooled to rt and slowly depressurized, the catalyst was filtered off, and the solution was concentrated under reduced pressure. The residue was subjected to column chromatography on silica gel (eluent EtOAc:MeOH = 3:1).
General procedure for the synthesis of Boc-protected piperazines Boc-1 with 5%-Pd/C catalyst (method A2). A 5%-Pd/C catalyst (50 mg per 0.5 mmol of 2) was added to a solution of dioxime 2 (1 equiv.) and Boc 2 O (3 equiv.) in methanol (0.1 M of 2). The vial was placed in a steel autoclave which was flushed and filled with hydrogen to a pressure of ca. 40 bar. Hydrogenation was conducted at this pressure and 50 • C for 6 h with vigorous stirring. Then, the autoclave was cooled to rt and slowly depressurized, the catalyst was filtered off, and the solution was concentrated under reduced pressure. The residue was subjected to column chromatography on silica gel (eluent PE:EtOAc = 5:1).
General procedure for the synthesis of free piperazines 1 with Raney nickel catalyst (method B1). A suspension of Ra-Ni (ca. 50 mg per 0.5 mmol of 2) in methanol (1 mL) was added to a vial containing a solution of dioxime 2 (1 equiv.) in methanol (0.1 M). The vial was placed in a steel autoclave which was flushed and filled with hydrogen to a pressure of ca. 40 bar. Hydrogenation was conducted at this pressure and 50 • C for 6 h with vigorous stirring. Then, the autoclave was cooled to rt and slowly depressurized, the catalyst was filtered off, and the solution was concentrated under reduced pressure. The residue was subjected to column chromatography on silica gel (eluent PE:EtOAc = 5:1 → 3:1 → 1:1 → EtOAc).
General procedure for the synthesis of Boc-protected piperazines Boc-1 with Raney nickel catalyst (method B2). A suspension of Ra-Ni (ca. 50 mg per 0.5 mmol of 2) in methanol (1 mL) was added to a solution of dioxime 2 (1 equiv.) and Boc 2 O (3 equiv.) in methanol (0.1 M of 2). The vial was placed in a steel autoclave which was flushed and filled with hydrogen to a pressure of ca. 40 bar. Hydrogenation was conducted at this pressure and 50 • C for 6 h with vigorous stirring. Then, the autoclave was cooled to rt and slowly depressurized, the catalyst was filtered off, and the solution was concentrated under reduced pressure. The residue was subjected to column chromatography on silica gel (eluent PE:EtOAc = 5:1).
General procedure for the synthesis of propionyl-protected piperazines EtCO-1 with Raney nickel catalyst (method B3). A suspension of Ra-Ni (ca. 50 mg per 0.5 mmol of 2) in methanol (1 mL) was placed into a vial containing a solution of dioxime 2 (1 equiv.) in methanol (0.1 M). The vial was placed in a steel autoclave which was flushed and filled with hydrogen to a pressure of ca. 40 bar. Hydrogenation was conducted at this pressure and 50 • C for 6 h with vigorous stirring. Then, the autoclave was cooled to rt and slowly depressurized, the catalyst was filtered off, and the solution was concentrated under reduced pressure. The residue was dissolved in dichloromethane (3 mL per 0.5 mmol of 2) and propionic anhydride (3 equiv.), then triethylamine (3 equiv.) and dmap (1 equiv.) were added to the solution. The mixture was placed in a refrigerator (about 0 • C) for 12 h. Then the solution was concentrated under reduced pressure, and the residue was subjected to column chromatography on silica gel (eluent PE:EtOAc = 5:1 → 3:1 → 1:1).

Conclusions
In conclusion, a general approach to the synthesis of piperazines bearing substituents at carbon and nitrogen atoms was developed utilizing primary amines and nitrosoalkenes as synthons. The method relies on the sequential double Michael addition of nitrosoalkenes to amines to give bis(oximinoalkyl)amines followed by catalytic reductive cyclization of oxime groups. The reductive cyclization of diketooximes is stereoselective and results in cisisomers of 2,6-disubstituted piperazines. Importantly, the method developed in this work