An Enantioselective Approach to 4-Substituted Proline Scaffolds: Synthesis of (S)-5-(tert-Butoxy carbonyl)-5-azaspiro[2.4]heptane-6-carboxylic Acid

A catalytic and enantioselective preparation of the (S)-4-methyleneproline scaffold is described. The key reaction is a one-pot double allylic alkylation of an imine analogue of glycine in the presence of a chinchonidine-derived catalyst under phase transfer conditions. These 4-methylene substituted proline derivatives are versatile starting materials often used in medicinal chemistry. In particular, we have transformed tert-butyl (S)-4-methyleneprolinate (12) into the N-Boc-protected 5-azaspiro[2.4]heptane-6-carboxylic acid (1), a key element in the industrial synthesis of antiviral ledipasvir.

More than 110 million people are infected with HCV worldwide, of which 700,000 die annually and chronic infection affects another 80 million. Progression of the disease often leads to cirrhosis or hepatic carcinoma.
Ledipasvir (2) is one of the more effective known anti-viral agents with an activity in the range of picomolar concentrations and, in combination with other antivirals, it is widely used for this infection.
Ledipasvir (2) is one of the more effective known anti-viral agents with an activity in the range of picomolar concentrations and, in combination with other antivirals, it is widely used for this infection.
The preparation of enantioenriched α-amino acid derivatives by alkylation of benzophenone imines of glycine alkyl esters under chiral phase-transfer catalysis (PTC) was first introduced by O′Donnell in 1978 [22]. Subsequent years witnessed impressive improvements reported by the groups of Lygo [23,24] and Maruoka [25], among many others. Nowadays, chiral phase-transfer  More than 110 million people are infected with HCV worldwide, of which 700,000 die annually and chronic infection affects another 80 million. Progression of the disease often leads to cirrhosis or hepatic carcinoma.
Ledipasvir (2) is one of the more effective known anti-viral agents with an activity in the range of picomolar concentrations and, in combination with other antivirals, it is widely used for this infection.
The preparation of enantioenriched α-amino acid derivatives by alkylation of benzophenone imines of glycine alkyl esters under chiral phase-transfer catalysis (PTC) was first introduced by O′Donnell in 1978 [22]. Subsequent years witnessed impressive improvements reported by the groups of Lygo [23,24] and Maruoka [25], among many others. Nowadays, chiral phase-transfer catalysis is a well-established strategy that has been applied to a large number of synthetic goals using chiral natural or non-natural derived catalysts [22,[26][27][28]. Scheme 1. Synthetic approaches to methylene-and spiro-derivatives of proline.
The preparation of enantioenriched α-amino acid derivatives by alkylation of benzophenone imines of glycine alkyl esters under chiral phase-transfer catalysis (PTC) was first introduced by O Donnell in 1978 [22]. Subsequent years witnessed impressive improvements reported by the groups of Lygo [23,24] and Maruoka [25], among many others. Nowadays, chiral phase-transfer catalysis is a well-established strategy that has been applied to a large number of synthetic goals using chiral natural or non-natural derived catalysts [22,[26][27][28].
We envisaged that the direct enantioselective alkylation of glycine-derived imines with 1,1′bis(halomethyl)cyclopropanes under PTC conditions would give ready access to the 4-spiro cyclopropane proline backbone. Unfortunately, the neopentylic nature of the alkylating reagent prevented any alkylation reaction. Thus, 1,1′-bis(bromomethyl)cyclopropane, 1,1′bis(iodomethyl)cyclopropane (and also 1,1′-bis(mesityloxy methyl)cyclopropane) did not react with neither tert-butyl N-diphenylmethyleneglycinate (7) nor p-chlorophenylmethylene imine 8 in the presence of a commercially available chinchonidine-derived [26,27] catalyst 9 (Scheme 2) using different bases (KOH or CsOH) and solvents (toluene or CH2Cl2) under phase-transfer conditions [26]. Looking for improved reactivity, we turned our attention to the more activated bis allylic alkylating systems (Scheme 3). The reaction of imine 7 with an excess of 3-bromo-2-(bromomethyl)-1-propene (10) in the presence of catalyst 9 (10 mol%) afforded an interesting mono-alkylated product 11. Unexpectedly, we observed that 11 was converted in situ in the NMR tube to the desired methylene proline 12 in 40% overall yield on standing overnight at room temperature. Encouraged by this promising result, the same sequence was performed, but without isolating alkylated intermediate imine 11. Furthermore, CDCl3 was substituted by non-deuterated chloroform in the cyclization. As a result, the product was obtained in a better yield but the excess of alkylating agent 10 promoted a further alkylation of the desired product 12 and dimeric proline 13 ( Figure 2) could be obtained (up to 10% yield) when 5 equivalents of 10 was used. Looking for improved reactivity, we turned our attention to the more activated bis allylic alkylating systems (Scheme 3). The reaction of imine 7 with an excess of 3-bromo-2-(bromomethyl)-1-propene (10) in the presence of catalyst 9 (10 mol%) afforded an interesting mono-alkylated product 11. Unexpectedly, we observed that 11 was converted in situ in the NMR tube to the desired methylene proline 12 in 40% overall yield on standing overnight at room temperature.
We envisaged that the direct enantioselective alkylation of glycine-derived imines with 1,1′bis(halomethyl)cyclopropanes under PTC conditions would give ready access to the 4-spiro cyclopropane proline backbone. Unfortunately, the neopentylic nature of the alkylating reagent prevented any alkylation reaction. Thus, 1,1′-bis(bromomethyl)cyclopropane, 1,1′bis(iodomethyl)cyclopropane (and also 1,1′-bis(mesityloxy methyl)cyclopropane) did not react with neither tert-butyl N-diphenylmethyleneglycinate (7) nor p-chlorophenylmethylene imine 8 in the presence of a commercially available chinchonidine-derived [26,27] catalyst 9 (Scheme 2) using different bases (KOH or CsOH) and solvents (toluene or CH2Cl2) under phase-transfer conditions [26]. Looking for improved reactivity, we turned our attention to the more activated bis allylic alkylating systems (Scheme 3). The reaction of imine 7 with an excess of 3-bromo-2-(bromomethyl)-1-propene (10) in the presence of catalyst 9 (10 mol%) afforded an interesting mono-alkylated product 11. Unexpectedly, we observed that 11 was converted in situ in the NMR tube to the desired methylene proline 12 in 40% overall yield on standing overnight at room temperature. Encouraged by this promising result, the same sequence was performed, but without isolating alkylated intermediate imine 11. Furthermore, CDCl3 was substituted by non-deuterated chloroform in the cyclization. As a result, the product was obtained in a better yield but the excess of alkylating agent 10 promoted a further alkylation of the desired product 12 and dimeric proline 13 ( Figure 2) could be obtained (up to 10% yield) when 5 equivalents of 10 was used. Encouraged by this promising result, the same sequence was performed, but without isolating alkylated intermediate imine 11. Furthermore, CDCl 3 was substituted by non-deuterated chloroform in the cyclization. As a result, the product was obtained in a better yield but the excess of alkylating agent 10 promoted a further alkylation of the desired product 12 and dimeric proline 13 ( Figure 2) could be obtained (up to 10% yield) when 5 equivalents of 10 was used. Several conditions for the initial alkylation were explored. Best yields were obtained when a slight excess of dibromide 10 was used in mixtures of toluene/CH2Cl2 and KOH as a base. Reaction temperature affected yield and enantioselectivity. Yield decreases dramatically below -20 °C, whereas a significant enantioselectivity loss was observed above 0 °C. The highest yield of 12 (71%) was obtained with 2.5 equivalent of 10 at -20 °C for 7 h (Scheme 4). Under these conditions, Several conditions for the initial alkylation were explored. Best yields were obtained when a slight excess of dibromide 10 was used in mixtures of toluene/CH 2 Cl 2 and KOH as a base. Reaction temperature affected yield and enantioselectivity. Yield decreases dramatically below -20 • C, whereas a significant enantioselectivity loss was observed above 0 • C. The highest yield of 12 (71%) was obtained with 2.5 equivalent of 10 at -20 • C for 7 h (Scheme 4). Under these conditions, enantioselectivity was determined by HPLC analysis of the N-benzyloxycarbonyl derivative of 12 and showed a 95:5 enantiomeric ratio (e.r.). Under the same conditions, slightly worse enantioselectivity (93:7 e.r.) and lower yield (38%) was obtained when 4 equivalents of 10 were used. Regarding intramolecular alkylation, it can be carried out in dichloromethane but chloroform was a more effective solvent [28]. Several conditions for the initial alkylation were explored. Best yields were obtained when a slight excess of dibromide 10 was used in mixtures of toluene/CH2Cl2 and KOH as a base. Reaction temperature affected yield and enantioselectivity. Yield decreases dramatically below -20 °C, whereas a significant enantioselectivity loss was observed above 0 °C. The highest yield of 12 (71%) was obtained with 2.5 equivalent of 10 at -20 °C for 7 h (Scheme 4). Under these conditions, enantioselectivity was determined by HPLC analysis of the N-benzyloxycarbonyl derivative of 12 and showed a 95:5 enantiomeric ratio (e.r.). Under the same conditions, slightly worse enantioselectivity (93:7 e.r.) and lower yield (38%) was obtained when 4 equivalents of 10 were used. Regarding intramolecular alkylation, it can be carried out in dichloromethane but chloroform was a more effective solvent [28]. With enantioenriched 12 in hand, the nitrogen atom was protected with the Boc group to afford protected amino acid 6. Basic hydrolysis of 6 easily afforded N-Boc 4-methylene proline 5 without loss of enantiopurity (checked by HPLC analysis of the corresponding benzyl ester). As far as we know, this is the first catalytic enantioselective preparation of these unsaturated prolines [29].
Subsequently, we proposed the preparation of amino acid 1. Direct cyclopropanation using the classical Simmons-Smith reaction (ZnEt2, CH2I2) of 12, 6 or 5 led to complex mixtures with significant amounts of deprotected products, probably due to the necessary acidic media [14]. Better results were achieved by addition of dibromocarbene species to the double bond followed by debromination. Several sequences of cyclopropanation/debromination/deprotection procedures were explored based on the results reported for Me esters [15] (or the free acid) [9] analogues in the literature (Scheme 5).
As far as cyclopropanation is concerned, better yields (73%) were obtained in the cyclopropanation of ester 6 with CBr3CO2Na rather than in the addition of dibromocarbene arising from bromoform on the corresponding carboxylic acid 5 (40%). Dibromo derivatives (from 5 and 6) were not isolated but were used as a mixture of diastereoisomers in the next step. In the debromination step, hydrogenolysis gave better results than radical reduction with tris(trimethyl)silane (83% vs. 53%). Therefore, according to our results, the best way to synthesize target molecule 1 is: (i) dibromopropanation of 6 to afford a crude mixture of diastereoisomers, (ii) ester deprotection under basic conditions to afford dibromo acids (80% yield) and, (iii) final debromination with hydrogen to furnish only one stereoisomer. The overall yield of the sequence was 49%. The enantiopurity of Boc-amino acid 1 was checked by chiral HPLC of the corresponding benzyl ester. With enantioenriched 12 in hand, the nitrogen atom was protected with the Boc group to afford protected amino acid 6. Basic hydrolysis of 6 easily afforded N-Boc 4-methylene proline 5 without loss of enantiopurity (checked by HPLC analysis of the corresponding benzyl ester). As far as we know, this is the first catalytic enantioselective preparation of these unsaturated prolines [29].
Subsequently, we proposed the preparation of amino acid 1. Direct cyclopropanation using the classical Simmons-Smith reaction (ZnEt 2 , CH 2 I 2 ) of 12, 6 or 5 led to complex mixtures with significant amounts of deprotected products, probably due to the necessary acidic media [14]. Better results were achieved by addition of dibromocarbene species to the double bond followed by debromination. Several sequences of cyclopropanation/debromination/deprotection procedures were explored based on the results reported for Me esters [15] (or the free acid) [9] analogues in the literature (Scheme 5).
As far as cyclopropanation is concerned, better yields (73%) were obtained in the cyclopropanation of ester 6 with CBr 3 CO 2 Na rather than in the addition of dibromocarbene arising from bromoform on the corresponding carboxylic acid 5 (40%). Dibromo derivatives (from 5 and 6) were not isolated but were used as a mixture of diastereoisomers in the next step. In the debromination step, hydrogenolysis Molecules 2020, 25, 5644 5 of 10 gave better results than radical reduction with tris(trimethyl)silane (83% vs. 53%). Therefore, according to our results, the best way to synthesize target molecule 1 is: (i) dibromopropanation of 6 to afford a crude mixture of diastereoisomers, (ii) ester deprotection under basic conditions to afford dibromo acids (80% yield) and, (iii) final debromination with hydrogen to furnish only one stereoisomer. The overall yield of the sequence was 49%. The enantiopurity of Boc-amino acid 1 was checked by chiral HPLC of the corresponding benzyl ester.

General Considerations
Compounds 7-10 are commercially available. 1 H (400 MHz) and 13 C (101 MHz) nuclear magnetic resonance (NMR) spectra were registered using a Varian Mercury 400 MHz NMR spectrometer. CDCl 3 (99.9%) was used as solvent while SiMe 4 (TMS) was used as reference. Coupling constants (J) are expressed in Hz and chemical shifts in parts per million (ppm). Signal multiplicities are reported using the following abbreviations: s (singlet), d (doublet), t (triplet), q (quadruplet), m (multiplet), and br (broad). More complicated signals are described as a combination of the indicated abbreviations: e.g., dd (doublet of doublets), dt (doublet of triplets). Two-dimensional experiments (NOESY, HSQC) were performed to confirm the signal assignments. ESI (+) or (-) spectra were acquired either on an LC/MSD-TOF instrument or on a ZQ mass spectrometer, utilizing a mixture of H 2 O:CH 3 CN (1:1, v/v) as the eluent.
An analytical sample of benzyl ester or 5 for HPLC analysis purposes was prepared: LiHDMS (1 M in THF, 0.07 mmol) was added to a sample of enantioenriched benzyl ester from 5 (10 mg, 0.035 mmol) in anhydrous THF (1 mL) at -78 • C under N 2 . After 10 min, the solution was warmed to 0 • C and quenched with pH 7 buffer solution (1 mL). The mixture was partitioned with DCM (2 mL) and the aqueous layer was extracted with additional DCM (2 × 2 mL). The combined organic phases were dried over anhydrous MgSO 4 . The solvent was evaporated to obtain a crude benzyl ester. Flash chromatography (SiO 2 , hexanes:EtOAc 90:10) was used to yield the pure racemic benzyl ester as a colorless oil (8 mg).

Synthesis of Compound 1
CBr 3 CO 2 Na (2.230 g, 7.00 mmol) was added to a solution of compound 6 (900 mg, 3.18 mmol) and Bu 4 NBr (21 mg, 0.064 mmol) in anhydrous CH 2 Cl 2 (15 mL) at rt under N 2 . The mixture was heated in an oil bath at 70 • C for 2.5 h and an additional amount of CBr 3 CO 2 Na (304 mg, 0.954 mmol) was added and the reaction was heated at 70 • C for an additional 3 h. A solution 7:3 of tert-butyl methyl ether and hexane (20 mL) was added at rt. The mixture was filtered through Celite ® . The organic layer was washed with water (20 mL) and brine (20 mL), and then dried over anhydrous MgSO 4 . The solvent was evaporated and the residue was filtered through a short silica column (90:10 hexanes/AcOEt) to obtain crude dibromo tert-butyl ester as an yellowish oil (6:4 mixture of diastereomers, 1.056 g, 73%). HRMS (ESI) (m/z) [M + Na] + calcd. for C 16 H 25 Br 2 NNaO 4 476.0048, found 476.0047.
A sample of the crude tert-butyl ester (892 mg, 1.96 mmol) was stirred in a mixture of MeOH (20 mL), THF (10 mL), H 2 O (10 mL) and aqueous KOH (10 mL, 50%) at rt for 48 h. The organic solvents were evaporated and the residue was acidified with HCl 2M to pH = 2. AcOEt (30 mL) was added, the organic layer was separated, and the aqueous layer was extracted with AcOEt (2 × 25 mL). The combined organic phases were dried over anhydrous MgSO 4 , the solvent was evaporated and the mixture was filtered through SiO 2 (CH 2 Cl 2 :MeOH 90:10) to afford the dibromo acid as a brownish solid (80%). HRMS ( The dibromo acid (620 mg) and KOH (554 mg, 30 mmol) in isopropanol (15 mL) were stirred under N 2 until the mixture became homogeneous. An amount of 10% Pd/C (132 mg, 0.122 mmol) was added and the resulting mixture was stirred in an oil bath at 40 • C under an H 2 atmosphere for 48 h. The cold mixture was filtered through Celite ® and the pad was washed with water (20 mL). The organic solvents were removed and tert-butyl methyl ether (20 mL) and 2M aq. HCl (20 mL) were added to the remaining solution. The phases were decanted and the aqueous phase was further washed with tert-butyl methyl ether (2 × 20 mL). The combined organic phases were dried over anhydrous MgSO 4 . The solvent was evaporated to obtain crude compound 1. Flash chromatography (SiO 2 , 95:5 CH 2 Cl 2 /MeOH) of the crude afforded 1 as a solid (311 mg, 83%).
(S)-5-(tert-butoxycarbonyl)-5-azaspiro [2.4] A sample of commercial enantioenriched compound 1 (S isomer, Fluorochem) was also transformed in the same way into the benzyl ester and used as a standard in HPLC experiments. Moreover, a sample of racemic benzyl ester was also prepared: LiHDMS (1 M in THF, 0.30 mmol) was added to a solution of benzyl ester arising from commercial 1 (94 mg, 0.30 mmol) in anhydrous THF (5 mL) at -78 • C under N 2 . After 10 min, the solution was warmed to 0 • C and quenched with pH 7 buffer solution (5 mL). The mixture was partitioned with DCM (10 mL) and the aqueous layer was extracted with additional DCM (2 × 5 mL). The combined organic phases were dried over anhydrous MgSO 4 . The solvent was evaporated to obtain a crude benzyl ester. Flash chromatography (SiO 2 , hexanes:EtOAc 90:10) was used to afford the pure racemic benzyl ester as a colorless oil (90 mg, 95%).

Synthesis of Compound 14
(Me 3 Si) 3 SiH (166 mg, 0.67 mmol) and AIBN (3.6 mg, 0.022 mmol) were added to a solution of the crude dibromo tert-butyl ester arising from 6 (101 mg, 0.22 mmol) in anhydrous toluene at 0 • C under an N 2 atmosphere. The mixture was stirred and heated in an oil bath at 90 • C for 18 h. The solvent was removed and the crude was purified by flash chromatography (SiO 2 , 9:1 hexane/AcOEt) to afford 15 as an yellowish oil (34 mg, 53%).

Hydrolysis of Compound 14
An aqueous solution of KOH (50%, 1 mL) was added to a solution of tert-butyl ester 14 (50 mg, 0.17 mmol) in MeOH (2 mL), THF (1 mL) and water (1 mL). The mixture was stirred 24 h at rt. The organic solvents were evaporated and the residue was acidified with HCl 2M to pH = 2. The mixture was extracted with AcOEt (10 mL), the organic layer was separated, and the aqueous layer was extracted with additional AcOEt (2 × 10 mL). The combined organic layer was dried over anhydrous MgSO 4 and the solvent was evaporated to afford almost pure compound 1 (34 mg, 82%). M.p. 95-6 • C.

Cyclopropanation of Compound 5
Aqueous NaOH (7.6 mL, 50%) was added dropwise to a solution of acid 5 (250 mg, 1.1 mmol), BnN(Me) 3 Cl (14 mg, 0.077 mmol) and CHBr 3 (834 mg, 3.3 mmol) in toluene (3 mL) at rt. The mixture was heated in an oil bath to 60 • C for 28 h. Toluene was evaporated and CH 2 Cl 2 (10 mL) was added. The organic layer was separated, washed with water and HCl 2M and dried over anhydrous MgSO 4 . After filtration through a short pad of silica, the solvent was evaporated to obtain a solid mixture of crude dibromide diastereomers (176 mg, 40%) which was used in the next step. Crude dibromide (176 mg) and KOH (148 mg, 0.80 mmol) were stirred in isopropanol at rt until the solid dissolved. An amount of 10% Pd/C (35 mg, 0.033 mmol) was added and the mixture was stirred under H 2 atmosphere at 40 • C in an oil bath for 48 h. The suspension was cooled down and filtered through Celite ® . The pad was washed with water (6 mL) and the volatiles were removed. The residue was partitioned by addition of 2M aq. HCl (6 mL) and tert-butyl methyl ether (6 mL). The aqueous phase was extracted with tert-butyl methyl ether (2 × 6vmL). The combined organic phases were dried over anhydrous MgSO 4 and the solvent was evaporated to afford compound 1 (88 mg, 83%).

Conclusions
In summary, we described herein the first enantioselective and catalytic approach to the useful protected amino acids 1 and 5 as well as their methylene precursor 12. The key step was the alkylation of the glycine-derived imine 7 with allylic dibromide 10 under phase transfer conditions in the presence of 10% of a commercially available chinchonidine-derived catalyst. Funding: This research was funded by Esteve Química SA, the "Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y la Nutrición" (CIBEROBN) and the Spanish Ministerio de Economía y Competitividad (CTQ2018-098630-B-100 and SAF2017-83813-C3-1-R).