Efficient Enzymatic Routes for the Synthesis of New Eight-membered Cyclic β-Amino Acid and β-Lactam Enantiomers

Abstract

Efficient enzymatic resolutions are reported for the preparation of new eight-membered ring-fused enantiomeric β-amino acids [(1R,2S)-9 and (1S,2R)-9] and β-lactams [(1S,8R)-3, (1R,8S)-3 (1S,8R)-4 and (1R,8S)-7], through asymmetric acylation of (±)-4 (E > 100) or enantioselective hydrolysis (E > 200) of the corresponding inactivated (±)-3 or activated (±)-4 β-lactams, catalyzed by PSIM or CAL-B in an organic solvent. CAL-B-catalyzed ring cleavage of (±)-6 (E > 200) resulted in the unreacted (1S,8R)-6, potential intermediate for the synthesis of enantiomeric anatoxin-a. The best strategies, in view of E, reaction rate and product yields, which underline the importance of substrate engineering, are highlighted.

1. Introduction

Chiral β-amino acids and β-lactams are important compounds because of their pharmacological properties and their use in diverse syntheses. For example, (1R,2S)-2-aminocyclopentanecarboxylic acid (cispentacin) [1], the simplest, naturally occurring carbocyclic β-amino acid, is an antifungal antibiotic, but can also be found in the structures of some natural products, e.g., amipurimycin [2]. Its methylene derivative, (1R,2S)-2-amino-4-methylenecyclopentanecarboxylic acid (icofungipen, PLD-118) [3], in turn, is active in vitro against Candida species. Enantiomeric eight-membered ring-fused β-lactams (1R,8S)- and (1S,8R)-9-azabicyclo[6.2.0]dec-4-en-10-one are potential key intermediates [4] in the syntheses of anatoxin-a [5]. It is a neurotoxic alkaloid, one of the most toxic of the cyanobacterial toxins, but a potent and stereospecific agonist at nicotinic acetylcholine receptors [5,6,7]. A relatively large number of publications deal with its isolation from strains of Anabaena flos aqua, a freshwater blue-green alga, but synthetic methods [8,9,10], also for the preparation of new anatoxin-a homologues [11] have also been described. Cyclic β-amino acids can serve as building blocks for the synthesis of modified peptides with increased activity and stability [12] and with well-defined three-dimensional structures (foldamers) (e.g., β-peptides with possible antibiotic activity) similar to those of natural peptides [13,14]. Additionally, the alkene functionality in molecules is amenable to a range of transformations. Cyclic β-amino acids can also be used in heterocyclic [15,16] and combinatorial [17] chemistry.

In addition to conventional resolution methods for the preparation of enantiomeric β-amino acids and β-lactams, enzymatic strategies have also been described [18,19,20]. Our research group has also devised a number of efficient enzymatic kinetic and sequential kinetic resolution processes (acylations, deacylations and hydrolyses). Most of these methods have been reviewed [21,22,23].

A primary aim of this work was to devise adequate enzymatic strategies for the preparation of valuable new enantiomeric eight-membered carbocyclic β-lactam and β-amino acid derivatives. In view of earlier results on enzymatic acylation of N-hydroxymethyl eight-membered carbocyclic β-lactams [4,24], we first planned to carry out enzymatic acylation of N-hydroymethyl-9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-4] (Figure 1). The extensive investigations of the lipase-catalyzed ring cleavage of unactivated [25,26,27] and activated [28] β-lactams then suggested the possibility of the lipase-catalyzed enantioselective ring cleavage of racemic unactivated 9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-3] and activated (±)-4 and N-hydroxymethyl 9-azabicyclo[6.2.0]dec-4-en-10-one [(±)-6]. A systematic comparison of the efficiency of these methods, with regard to E, reaction rate and yield for the products, was also intended.

2. Results and Discussion

2.1. Synthesis of β-lactams (±)-3–(±)-6

Lactams (±)-3 and (±)-5 were synthesised from cyclooctadiene 1 or 2 by the addition of chlorosulfonyl isocyanate (CSI), according to a slightly modified literature procedure (Scheme 1) [29,30]. Then product lactams were reacted with paraformaldehyde under sonication to form N-hydroxymethyl β-lactams (±)-4 and (±)-6 [4].

2.2. Lipase-Catalyzed O-acylation of (±)-4

On the basis of the earlier results on the enzymatic resolution of N-hydroxymethyl 9-azabicyclo[6.2.0]dec-4-en-10-one [4] and N-hydroxymethyl 9-azabicyclo[6.2.0]decane-4-en-10-one [24], first the acylation of (±)-4 (Scheme 2) was carried out with vinyl butyrate (VB) in the presence of PSIM (Burkholderia cepacia) in diisopropyl ether (iPr₂O) at −15 °C (

Table 1. Enzyme-catalyzed acylation of (±)-4 ^a.

Entry	Enzyme (30 mg mL−1)	Acyl Donor (Equiv)	Solvent	Temp. (°C)	R. Time (Min)	Conv. (%)	ees b (%)	eep b (%)	E
1	PSIM	VB (2)	iPr2O	-15	120	44	76	96	112
2	PSIM	VB (2)	iPr2O	2-3	60	43	70	94	67
3	PSIM	VB (2)	iPr2O	30	10	46	77	90	44
4	PSIM	VB (10)	iPr2O	30	10	51	87	84	32
5	PSIM	VB (10) + Et3N + Na2SO4	iPr2O	30	10 25	45 50	77 91	96 92	114 76
6	PSIM	2,2,2-Trifluoroethyl-butyrate(10)	iPr2O	30	20	49	83	86	34
7	PSIM	VA (10)	iPr2O	30	10	50	81	82	25
8	PSIM	EtOAc(10)	iPr2O	30	240	16	12	62	5
9	PSIM	Ac2O (10)	iPr2O	30	10	52	87	80	25
10	AK c	VB (10)	iPr2O	30	20	49	74	78	17
11	AY c	VB (10)	iPr2O	30	240	25	14	42	28
12	CAL-A c	VB (10)	iPr2O	30	10	29	12	29	2
14	CAL-B	VB (10)	iPr2O	30	50	67	2	1	1
14	PPL	VB (10)	iPr2O	30	120	20	23	91	26
15	PSIM	VB (10)	tBuOMe	30	10	51	87	85	34
16	PSIM	VB (10)	Toluene	30	5	49	89	93	82
17	PSIM	VB (10)	Acetone	30	60	49	86	89	47

^a 0.015 M substrate; ^b According to GC (Experimental Section); ^c Contains 20% (w/w) of lipase adsorbed on Celite in the presence of sucrose.

, entry 1).

When the acylation was performed at higher temperatures, the reaction rate increased with the concomitant decrease in E (entries 2 and 3 vs. 1). As the amount of VB was increased from 2 to 10 equiv., the reaction rate increased while E apparently decreased (entry 4 vs. 3). The addition of a catalytic amount of Et₃N and Na₂SO₄ resulted in a clear increase in E (entry 5 vs. 4).

To further increase the E values, acyl donors, such as 2,2,2-trifluoroethyl butyrate, vinyl acetate (VA), ethyl acetate (EtOAc) and acetic anhydride (Ac₂O) were tested (entries 6–9). Unfortunately, none of the acyl donors tested exerted any beneficial influence on the reaction course. Several solvents have also been tested. When iPr₂O was replaced by tBuOMe, practically no change was observed in the reaction course (entry 15 vs. 4). The same high E but somewhat lower reaction rate was observed in acetone (entry 17), while the best E and fastest reaction were detected in toluene (entry 16). Finally the environmentally less harmful iPr₂O was selected as solvent.

In addition to lipase PSIM, several enzymes, such as lipase AK (Pseudomonas fluorescens), lipase AY (Candida rugosa), CAL-A (lipase A from Candida antarctica), CAL-B (lipase B from Candida antarctica) and PPL (porcine pancreatic lipase) have been investigated (entries 10–14). However, in terms of E and reaction rate, none of them proved to be better than PSIM.

2.3. Lipase-Catalyzed Ring Cleavage of (±)-3–(±)-6

In view of the results on the lipase-catalyzed ring cleavage of carbocyclic β-lactams [25,26,27], the ring opening of (±)-3 was attempted with 1 equiv. of H₂O in the presence of CAL-B (lipase B from Candida antarctica) in iPr₂O at 60 °C (Scheme 3;

Table 2. CAL-B-catalyzed ring cleavage of racemic 3 ^a, 4 ^b, 5 ^a and 6 ^b.

Entry	Substrate	R. Time (h)	Conv. c (%)	ees d (%)	eep e (%)
1	(±)-3	5	43	75	>99
2	(±)-5	5	46	84	>99
3	(±)-4	3	49	96	>99
4	(±)-6	3	50	>99	98

^a 0.015 M substrate, 30 mg mL⁻¹ CAL-B, 0.5 equiv. of H₂O, iPr₂O, 60 °C; ^b 0.015 M substrate, 30 mg mL⁻¹ CAL-B, 0.5 equiv. of H₂O, 1 equiv. of benzylamine, iPr₂O, 60 °C; ^c Calculated from ee_S and ee_P; ^d According to GC (Experimental Section); ^e Determined by using GC, after double derivatization (DD) [31].

, entry 1). When the ring cleavage of the regioisomeric 9-azabicyclo[6.2.0]dec-4-en-10-one [(±)-5] was performed under the same conditions, a similar fast reaction and the same high ee values (>98%) were observed (entry 2).

In further studies, we have probed our very recent results found about the ring cleavage of specially activated lactams [28], where the activating group underwent to a traceless, in situ degradation. Accordingly, the ring cleavage of (±)-4 was attempted with H₂O in the presence of CAL-B and benzylamine to capture formaldehyde in iPr₂O at 60 °C (Scheme 4, Table 2, entry 3). Excellent ee (>99%) characterized formed amino acid (1R,2S)-9 at a conversion close to 50%. The ring cleavage of regioisomeric (±)-6 was also carried out under the same conditions (entry 4), and the same high ee (>98%) for amino acid (1R,2S)-10 and unreacted lactam (1S,8R)-6, potential intermediate in the synthesis of enantiomeric anatoxin-a was observed. In good accordance with the earlier observation that the hydroxymethyl group activates the ring cleavage of lactams, racemic 4 and 6 underwent ring cleavage much faster than their corresponding inactivated counterparts (entry 3 vs. 1 and 4 vs. 2).

On the basis of the preliminary results, preparative-scale resolutions of racemic 3, 4 and 6 were performed under the optimized conditions (footnotes to

Table 3. Preparative-scale resolutions.

				Product				Unreacted substrate
	Reaction Partner	R. Time (Min)	Conv. a (%)	Yield (%)	Isomer	eeP (%)	${\mathbf{\left[}\mathbf{\alpha }\mathbf{\right]}}_{\mathbf{D}}^{\mathbf{25}}$	Yield (%)	Isomer	eeS (%)	${\mathbf{\left[}\mathbf{\alpha }\mathbf{\right]}}_{\mathbf{D}}^{\mathbf{25}}$ (EtOH)
(±)-3 b	H2O	330	50	48	(1R,2S)-9	99 c	−17 d	49	(1S,8R)-3	99 e	−140.6 f
(±)-4 g	VB	10	51	46	(1R,8S)-7	94 e	+39.2 h	44	(1S,8R)-4	96 e	−142.4 i
(±)-4 j	H2O	180	49	47	(1R,2S)-9	>99 c	−17.1 d	48	(1S,8R)-4	98 e	−140.4 f
(±)-6 j	H2O	180	50	47	(1R,2S)-10	99 c	+24.9 k	46	(1S,8R)-6	99 e	−28.7 l

^a Calculated from ee_S and ee_P; ^b 0.015 M substrate, 30 mg mL⁻¹ CAL-B, 0.5 equiv. of H₂O, iPr₂O, 60 °C; ^c Determined by GC after DD (Experimental Section); ^d c = 0.35; H₂O; ^e Determined by GC (Experimental Section); ^f c = 0.5; ^g 0.015 M substrate, 30 mg mL⁻¹ PSIM, 10 equiv. of VB, catalytic amount of Et₃N and Na₂SO_4, iPr₂O 30 °C; ^h c = 0.35; EtOH; ⁱ c = 0.45; ^j 0.015 M substrate, 30 mg mL⁻¹ CAL-B, 0.5 equiv. of H₂O, 1 equiv. of benzylamine, iPr₂O, 60 °C; ^k c = 0.3; H₂O; ^l c = 0.5.

). The results are reported in Table 3 and Experimental Section.

2.4. Further Transformations

Hydrolysis of enantiomeric 3, 4, 6 and 7 with 18% aqueous HCl (Scheme 2, Scheme 3 and Scheme 4) gave the corresponding hydrochloride salts 9·HCl and 10·HCl (ee > 97%), while treatment of enantiomeric 4 and 7 with NH₄OH/MeOH (Scheme 2) resulted in β-lactams (1R,8S)-3 and (1S,8R)-3 (ee ≥ 95%). Catalytic reduction of enantiomeric lactams 3 and amino acids 9 with H₂ or cyclohexene as a hydrogen donor and using palladium-on-carbon furnished saturated lactams (1R,8S)-8 and (1S,8R)-8, and amino acids (1R,2S)-11 and (1S,2R)-11, respectively (Scheme 2 and Scheme 3), without a drop in ee (>96%). Physical data on the enantiomers prepared are reported in

Table 4. Physical data on enantiomers prepared.

Entry	Enantiomers	ee (%)	${\mathbf{\left[}\mathbf{\alpha }\mathbf{\right]}}_{\mathbf{D}}^{\mathbf{25}}$
1	(1R,8S) 3 from (1R,8S)-7	95	+147 (c = 0.5; EtOH)
2	(1S,8R) 3 from (1S,8R)-4	96	−148.7 (c = 0.4; EtOH)
3	(1R,8S) 8 from (1R,8S)-3	98	+17.7 (c = 0.5; CHCl3)
4	(1S,8R) 8 from (1S,8R)-3	96	−17.1 (c = 0.5; CHCl3)
5	(1S,2R)-9·HCl from (1S,8R) 3	99	+19.6 (c = 0.5; H2O)
6	(1S,2R)-9·HCl from (1S,8R) 4	99	+19.6 (c = 0.6; H2O)
7	(1R,2S)-9·HCl from (1R,8S)-7	98	−17.3 (c = 0.35; H2O)
8	(1S,2R)-10·HCl from (1S,8R)-6	97	−15.0 (c = 0.5; H2O)
9	(1R,2S)-11 from (1R,2S)-9	99	+19.2 (c = 0.4; H2O)
10	(1S,2R)-11 from (1S,2R)-9	99	−19 (c = 0.33; H2O)

and Experimental Section.

2.5. Absolute Configurations

Absolute configurations were determined by comparing the [α] values of the enantiomeric compounds with literature data. Specifically, [α] values of enantiomeric compounds 8 (Table 4, entries 3 and 4), obtained from enantiomeric compounds 4 and 7 (Scheme 2), were compared with the literature data of (1S,8R)-9-azabicyclo[6.2.0]decan-10-one {${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −18 (c = 0.5; CHCl₃)} [25]. In a similar way, the [α] values of enantiomeric compounds 11 (Table 4, entries 9 and 10), obtained from enantiomeric compounds 9·HCl (Scheme 3), were compared with the literature data for (1R,2S)-2-aminocyclooctane-1-carboxylic acid {${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +17.8 (c = 0.4; H₂O)} [25]. Thus, the (1R,8S) configuration is determined for ester 9 and the corresponding lactam (3), and the (1S,8R) configuration is assigned to unreacted alcohol 4 and the corresponding lactam (3). The CAL-B-catalyzed ring cleavage of either inactivated lactams 3 and 5 or activated lactams 4 and 6 afforded the corresponding amino acids with (1R,2S) absolute configuration I.

3. Experimental Section

3.1. Materials and Methods

CAL-B (lipase B from Candida antarctica), produced by the submerged fermentation of a genetically modified Aspergillus oryzae microorganism and adsorbed on a macroporous resin (Catalogue no. L4777), and porcine pancreatic lipase (PPL) were from Sigma-Aldrich (St. Louis, MO, USA). CAL-A (lipase A from Candida antarctica) was purchased from Roche Diagnostics Corporation. Lipase PSIM (Burkholderia cepacia, immobilized on diatomaceous earth) was a kind gift from Amano Enzyme Europe Ltd. (Suqian, China) Lipase AK (Pseudomonas fluorescens) was from Amano Pharmaceuticals, and lipase AY (Candida rugosa) was from Fluka. The solvents were of the highest analytical grade. Melting points were determined with a Kofler apparatus. NMR spectra were recorded on a Bruker DRX 400 spectrometer. Optical rotations were measured with a Perkin-Elmer 341 polarimeter. Elemental analyses were performed with a Perkin-Elmer CHNS-2400 Ser II Elemental Analyzer. The syntheses of racemic N-hydroxymethyl 9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-5] and N-hydroxymethyl 9-azabicyclo[6.2.0]dec-4-en-10-one [(±)-6] were performed according to our earlier method [4].

3.2. Typical Small-Scale Enzymatic Experiments

Racemic β-lactam in an organic solvent (0.05 M solution, 1 mL) was added to the lipase tested (30 mg mL⁻¹). The tested acyl donor (2 or 10 equiv.) in acylations or H₂O (1 equiv.) in hydrolyses was added. The mixture was shaken at −15 °C, 2–3 °C, 30 °C or 60 °C. The progress of the reaction was followed by taking samples from the reaction mixture at intervals and analysing them by using a gas chromatograph equipped with a chiral column. The ee values for the unreacted β-lactam enantiomers were determined directly on a Chromopack Chiralsil-Dex CB column [retention times are given in min; 140 °C for 25 min → 190 °C (temperature rise 20 °C∙min⁻¹; 140 kPa, (1S,8R)-3: 27.74 (antipode: 27.56); 140 °C for 7 min → 190 °C (temperature rise 10 °C min⁻¹; 140 kPa, (1S,8R)-5: 12.55 (antipode: 12.12)] or on a Chirasil-L-Val column, [140 °C for 7 min → 190 °C (temperature rise 10 °C/min; 140 kPa), (1S,8R)-4: 14.39 (antipode: 14.61); (1S,8R)-6: 14.40 (antipode: 14.87); (1R,8S)-7: 14.67 (antipode: 14.36)]. The ee values for the β-amino acids produced were determined on a Chirasil-L-Val column, after double derivatization with (i) CH₂N₂ [31] (caution! derivatization with diazomethane should be performed under a well-ventilated hood); (ii) Ac₂O in the presence of 4-dimethylaminopyridine and pyridine [120 °C for 7 min → 190 °C (temperature rise 10 °C min⁻¹; 140 kPa, (1R,2S)-9: 12,41 (antipode: 12.95); (1R,2S)-10: 12.88 (antipode: 14.06)] and after double derivatization with (i) CH₂N₂ and (ii) (PropCO)₂O in the presence of 4-dimethylaminopyridine and pyridine [140 °C for 10 min → 190 °C (temperature rise 10 °C min⁻¹; 140 kPa; (1R,2S)-11: 17.58 (antipode: 17.78)].

3.3. Synthesis of Racemic 9-Azabicyclo[6.2.0]dec-6-en-10-one [(±)-3]

A solution of CSI (2.28 mL, 26.2 mmol) in CH₂Cl₂ (25 mL) was added dropwise to a stirred solution of 1,3-cyclooctadiene (3 mL, 26.2 mmol) in CH₂Cl₂ (25 mL) at 0 °C. The reaction mixture was stirred at room temperature for 21 h, and the resulting liquid was then added dropwise to a vigorously stirred solution of Na₂SO₃ (0.33 g) and K₂CO₃ (7.12 g) in H₂O (50 mL). After stirring the mixture for 3.5 h, the combined organic layers were dried (Na₂SO₄) and, after filtration, concentrated. The resulting white solid, racemic 3 (3.41 g, 86%), was recrystallized from iPr₂O (m.p. 103–106 °C).

The ¹H-NMR (400 MHz, CDCl₃, 25 °C, TMS) δ (ppm) data for (±)-3: 1.46–2.11 (8H, m, 4 × CH₂), 3.34–3.38 (1H, dd, J = 5.34, 12.26 Hz, H-1), 4.57–4.58 (1H, m, H-8), 5.41–5.79 (2H, m, CHCH), 6.01 (1H, bs, NH). Anal. calcd for C₉H₁₄NO: C, 71.49; H, 8.67; N, 9.26; Anal. found: C, 71. 30; H, 8.52; N, 9.23.

3.4. Synthesis of Racemic N-Hydroymethyl-9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-4]

Racemic 3 (3 g, 19.84 mmol) was dissolved in THF (35 mL) followed by adding paraformaldehyde (0.6 g, 19.84 mmol), K₂CO₃ (0.17g, 1.19 mmol) and H₂O (1.1 mL). The solution was sonicated for 4 h, the solvent was evaporated off and the residue was dissolved in ethyl acetate (50 mL). The solution was dried (Na₂SO₄) and then concentrated. The residue was recrystallised from iPr₂O to afford (±)-4 as a white crystalline product (3.2 g, 89%; m.p. 81–84 °C).

The ¹H-NMR (400 MHz, CD₃OD, 25 °C, TMS) δ (ppm) data for (±)-4: 1.45–2.09 (8H, m, 4 × CH₂), 3.30–3.32 (1H, dd, J = 1.89. 5.52 Hz, H-1), 3.33-3.39 (1H, s, OH), 4.60–4.64 (1H, dd, J = 7.78, 11,58 Hz, CH₂OH), 4.69–4.71 (1H, m, H-8), 4.85–4.89 (1H, dd, J = 6.47, 11,66 Hz, CH₂OH), 5.57–5.60 (2H, m, CHCH). Anal. calcd. for C₁₀H₁₅NO₂: C, 66.27; H, 8.34; N, 7.73; found: C, 66.19; H, 8.44; N, 7.76.

3.5. Gram-Scale Resolution of N-hydroxymethyl-9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-4] through Acylation

Racemic 4 (1 g, 5.51 mmol) was dissolved in iPr₂O (40 mL). Lipase PSIM (1.2 g, 30 mg mL⁻¹) then VB (6.29 mL, 55.1 mmol), Et₃N (a few drops) and Na₂SO₄ (0.2 g) were added and the mixture was shaken at 30 °C. After 10 min, the enzyme was filtered off at 51% conversion and iPr₂O was evaporated. The residue was chromatographed on silica (EtOAc:n-hexane 7:3) affording unreacted (1S,8R)-4 [0.44 g, 44%; [α]_D²⁵ = –142.4 (c = 0.45, EtOH); m.p. 79–83 °C (recrystallized from iPr₂O); ee = 96%] and ester (1R,8S)-7 (0.63 g, 46%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +39.2 (c = 35, EtOH); ee = 94%) as a pale-yellow oil.

The ¹H-NMR (400 MHz, CDCl₃, 25 °C, TMS) δ (ppm) data for (1R,8S)-7: 0.93–0.97 (3H, t, J = 7.4 Hz, CH₃), 1.41–1.44 (2H, m, CH₂CH₂CH₃), 1.57–2.33 (10H, m, CH₂CH₂CH₃ and 4 × CH₂ from ring), 3.31–3.32 (1H, m, H-1), 4.59–4.60 (1H, dd, J = 1.58, 3.88 Hz, H-8), 5.15–5.18 (1H, d, J = 11.36 Hz, CH₂OCOPr), 5.20–5.23 (1H, d, J = 11.36 CH₂OCOPr), 5.49–5.81 (2H, m, CHCH). Anal. calcd. for C₁₄H₂₁NO₃: C, 66.91; H, 8.42; N, 5.57; found: C, 66.83; H, 8.60; N, 5.42.

The ¹H-NMR (400 MHz, CDCl₃, 25 °C, TMS) δ (ppm) data for (1S,8R)-4 are similar to those for (±)-4. Anal. found: C, 66.39; H, 8.24; N, 7.73.

3.6. Gram-Scale Resolution of 9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-3] through Hydrolysis

Racemic 3 (0.5 g, 3.3 mmol) was dissolved in iPr₂O (20 mL). CAL-B (0.6 g, 30 mg mL⁻¹) and H₂O (29.7 μL, 1.65 mmol) were added, and the mixture was shaken in an incubator shaker at 60 °C for 5.5 h. The reaction was stopped by filtering off the enzyme at 50% conversion. The solvent was evaporated off and the residue crystallized to give (1S,8R)-3 [240 mg, 48%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −140.6 (c = 0.5; EtOH); m.p. 151–153 °C recrystallized from iPr₂O; ee = 99%]. The filtered-off enzyme was washed with distilled water (3 × 15 mL), and the water was evaporated off, yielding crystalline β-amino acid (1R,2S)-9 [268 mg, 48%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −17 (c = 0.35; H₂O); 266–288 °C with sublimation (recrystallized from H₂O/acetone); ee = 99%].

The ¹H-NMR (400 MHz, CD₃OD, 25 °C, TMS) δ (ppm) data for (1S,8R)-3 were similar to those for (±)-3. Anal. found: C, 71.33; H, 8.57; N, 9.10.

The ¹H-NMR (400 MHz, D₂O) δ (ppm) data for (1R,2S)-9: 1.45–2.24 (8H, m, 4 × CH₂) 2.86–2.90 (1H, m, H-1) 4.42–4.45 (1H, m, H-2) 5.71–6.03 (2H, m, CHCH). Anal. calcd. for C₉H₁₅NO₂: C, 63.88; H, 8.93; N, 8.28; found: C, 63.99; H, 8.84; N, 8.28.

3.7. Gram-Scale Resolution of N-hydroxymethyl 9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-4] through Hydrolysis

(±)-4 (250 mg, 1.37 mmol) was dissolved in iPr₂O (10 mL) followed by the addition of CAL-B (300 mg, 30 mg mL⁻¹), benzylamine (150 μL, 1.37 mmol) and H₂O (12.3 μL, 0.68 mmol) and by shaking the mixture in an incubator shaker at 60 °C for 3 h. The reaction was stopped by filtering off the enzyme at 49% conversion. After solvent evaporation the residue (1S,8R)-4 crystallized out [120 mg, 48%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −140.4 (c = 0.5; EtOH), m.p. 80–83 °C (recrystallized from iPr₂O), ee = 98%]. The filtered enzyme was washed with distilled H₂O (3 × 15 mL), and crystalline β-amino acid (1R,2S)-9 was isolated after evaporation of H₂O. [109 mg, 47%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −17.1 (c = 0.35; H₂O), m.p. 264–265 °C (recrystallised from H₂O/acetone), ee > 99%].

The ¹H-NMR (400 MHz, CD₃OD, 25 °C, TMS) δ (ppm) data for (1S,8R)-4 were similar to those for (±)-4. Anal. found: C, 66.18; H, 8.35; N, 7.73.

The ¹H-NMR (400 MHz, D₂O) data δ (ppm) for (1R,2S)-9 from (±)-4 were similar to those for (1R,2S)-9 from (±)-3 (4.4.). Anal. found: C, 63.95; H, 9.01; N, 8.25.

3.8. Gram-Scale Resolution of N-hydroxymethyl 9-azabicyclo[6.2.0]dec-4-en-10-one [(±)-6] through Hydrolysis

Via the procedure described above, the reaction of racemic 6 (250 mg, 1.37 mmol), benzylamine (150 μL, 1.37 mmol) and H₂O (12.3 μL, 0.68 mmol) in iPr₂O (10 mL) in the presence of CAL-B (300 mg, 30 mg mL⁻¹) at 60 °C afforded amino acid (1R,2S)-10 [108 mg, 47%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +24.9 (c = 0.3; H₂O), lit. [26] = +23.9 (c = 0.3; H₂O), ee = 95%; m.p.: 264–265 °C (recrystallized from H₂O/acetone), lit. [26] 218–220 °C; ee = 99%] and unreacted (1S,8R)-6, [115 mg, 46%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −28.7 (c = 0.5; EtOH), ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −28.5 (c = 1; MeOH), lit. [4] = −27.0 (c = 1; MeOH), ee = 97%; m.p. 48–49 °C (recrystallized from n-hexane), lit. [4] 48–49 °C; ee = 99%] after 3 h.

The ¹H-NMR (400 MHz, D₂O) δ (ppm) data for (1R,2S)-10: 1.81–2.58 (8H, m, 4 × CH₂) 2.79 (1H, m, H-1) 3.70–3.73 (1H, m, H-2) 5.72–5.73 (2H, m, CHCH). Anal. calcd. for C₉H₁₅NO₂: C, 63.88; H, 8.93; N, 8.28; found: C, 63.59; H, 8.88; N, 8.18.

The ¹H-NMR (400 MHz, CDCl₃, 25 °C, TMS) data δ (ppm) for (1S,8R)-6 were similar to those for (±)-6: 1.86–2.21 and 2.37–2.44 (8H, m, 4 × CH₂), 3.28–3.30 (1H, m, H-1), 3.71 (1H, bs, OH), 3.93–3.96 (1H, m, H-8), 4.45–4.48 (1H, d, J = 11.6 Hz, CH₂OH), 4.73–4.79 (1H, d, J = 11.6 Hz, CH₂OH), 5.67–5.71 (2H, m, CHCH). Anal. calcd. for C₁₀H₁₅NO₂: C, 66.27; H, 8.34; N, 7.73; found: C, 66.10; H, 8.28; N, 7.66.

3.9. Synthesis of (1R,8S)-8 and (1S,8R)-8 through (1R,8S)-3 and (1S,8R)-3 from β-lactams (1S,8R)-4 and (1R,8S)-7

Ester (1R,8S)-7 (100 mg, 0.66 mmol) was dissolved in MeOH (10 mL), NH₄OH (1 mL) was added and the mixture was stirred at room temperature for 6 h. The solvent was evaporated off, the residue chromatographed on silica (EtOAc:n-hexane 7:3) providing white crystals of (1R,8S)-3 [48 mg, 80%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +147 (c = 0.5; EtOH); m.p. 152–153 °C (recrystallised from iPr₂O); ee = 95%].

Similarly, (1S,8R)-4 (100 mg, 0.55 mmol) afforded white crystals of (1S,8R)-3 [59 mg, 71%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −148.7 (c = 0.4; EtOH); m.p. 150–153 °C (recrystallised from iPr₂O); ee = 96%].

The ¹H-NMR (400 MHz, CD₃OD, 25 °C, TMS) δ (ppm) data for (1S,8R)-3 and (1R,8S)-3 were similar to those for (±)-3. Anal. found for (1S,8R)-3: C, 71.45; H, 8.57; N, 9.22. Anal. found for (1R,8S)-3: C, 71.55; H, 8.67; N, 9.12.

Palladium-on-carbon (100 mg) was added to (1R,8S)-3 or (1S,8R)-3 (100 mg, 0.66 mmol) dissolved in a mixture of MeOH (10 mL) and cyclohexene (1 mL). The mixture was treated at reflux temperature for 3 h, and then the catalyst was filtered off. After evaporation, (1R,8S)-8 [51 mg, 51%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +17.7 (c = 0.5; CHCl₃); m.p. 104–106 °C (recrystallized from iPr₂O); ee = 98%] or (1S,8R)-8 [47 mg, 47%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −17.1 (c = 0.5; CHCl₃), lit. [25] ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −18 (c = 0.5; CHCl₃)]; m.p. 105–107 °C, lit. [25] m.p. 108–112 °C; ee = 96%] was obtained as white crystalline product.

The ¹H-NMR (400 MHz, CD₃OD, 25 °C, TMS) δ (ppm) data for (1S,8R)-8 were similar to those for (1R,8S)-8: 1.30–2.09 (12H, m, 6xCH₂), 3.03–3.06 (1H, m, H-1), 3.65–3.69 (1H, m, H-8), 5.89 (1H, bs, NH). Anal. found for C₉H₁₅NO, (1S,8R)-8: C, 70.48; H, 9.76; N, 9.14 and for (1R,8S)-8: C, 70.51; H, 9.82; N, 9.08.

3.10. Preparation of (1R,2S)-11 and (1S,2R)-11

Palladium-on-carbon (60 mg) was added to enantiomer (1R,2S)-9 or (1S,2R)-9 (100 mg) dissolved in MeOH (20 mL) and H₂ was bubbled through the system at RT for 6 h. The catalyst was then filtered off and after evaporation white crystalline enantiomers of 2-aminocyclooctane-1-carboxylic acid (1R,2S)-11 [69.1 mg, 68%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +19.2 (c = 0.4; H₂O), lit. [25] ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = +17.8 (c = 0.4; H₂O); m.p. 250–252 °C (recrystallized from H₂O/acetone), lit. [25] m.p. 245–248 °C; ee = 99%] and (1S,2R)-11 [76.2 mg, 75%; ${\left[\alpha \right]}_{\mathrm{D}}^{25}$ = −19 (c = 0.33; H₂O); m.p. 241–245 °C (recrystallized from H₂O/acetone); ee = 99%] were isolated.

The ¹H-NMR (400 MHz, D₂O) data δ (ppm) for (1R,2S)-11 are similar to those for (1S,2R)-11: 1.50–1.92 (12H, m, 6xCH₂) 2.78–2.79 (1H, m, H-1) 3.59–3.63 (1H, m, H-2). Anal. calcd. for C₉H₁₇NO₂: C, 63.14; H, 10.01; N, 8.18; found for (1R,2S)-11: C, 63.20; H, 10.18; N, 8.02; found for (1S,2R)-11: C, 63.14; H, 9.91; N, 8.29.

3.11. Acidic Hydrolyses to β-amino Acid Hydrochlorides

When the enantiomeric lactam at issue (0.2 mmol) was treated with 18% aqueous HCl (5 mL) at reflux temperature, the desired amino acid hydrochloride was obtained, as follows:

(1R,2S)-9·HCl [37 mg, 92%; ${\left[\mathsf{\alpha }\right]}_{\mathrm{D}}^{25}$ = −17.3 (c = 0.35; H₂O); m.p. 218–220 °C (recrystallised from EtOH/Et₂O); ee = 98%] from 50 mg (1R,8S)-7.

(1S,2R)-9·HCl [35 mg, 86%; ${\left[\mathsf{\alpha }\right]}_{\mathrm{D}}^{25}$ = +19.6 (c = 0.5; H₂O); m.p. 214–218 °C (recrystallised from EtOH/Et₂O); ee = 99%] from 30 mg (1S,8R) 3.

(1S,2R)-9·HCl [31 mg, 76%; ${\left[\mathsf{\alpha }\right]}_{\mathrm{D}}^{25}$ = +19.6 (c = 0.6; H₂O); m.p. 219–220 °C (recrystallised from EtOH/Et₂O); ee = 99%] from 36 mg (1S,8R) 4.

(1S,2R)-10·HCl [38 mg, 93%; ${\left[\mathsf{\alpha }\right]}_{\mathrm{D}}^{25}$ = −15.0 (c = 0.5; H₂O), lit. [25] ${\left[\mathsf{\alpha }\right]}_{\mathrm{D}}^{25}$ = −15.9 (c = 0.3; H₂O; m.p. 200–204 °C (recrystallised from EtOH/Et₂O ee = 97%] from 36 mg (1S,8R)-6.

The ¹H-NMR (400 MHz, D₂O) data δ (ppm) for (1R,2S)-9·HCl are similar to those for (1S,2R)-9·HCl: 1.52–2.31 (8H, m, 4 × CH₂) 3.21–3.24 (1H, m, H-1) 4.55–4.58 (1H, m, H-2) 5.76–6.14 (2H, m, CHCH).

The ¹H-NMR (400 MHz, D₂O) data δ (ppm) for (1S,2R)-10·HCl: 1.64–2.33 (8H, m, 4 × CH₂) 2.91–2.98 (1H, m, H-1) 3.66–3.69 (1H, m, H-2) 5.56–5.63 (2H, m, CHCH).

Anal. calcd. for C₉H₁₆ClNO₂: C, 52.56; H, 7.84; N, 6.81; found for (1R,2S)-9·HCl: C, 52.50; H, 7.68; N, 6.80; found for (1S,2R)-9·HCl from (1S,8R) 3: C, 52.68; H, 7.64; N, 6.78; found for (1S,2R)-9·HCl from (1S,8R)-4: C, 52.56; H, 7.66; N, 6.95; found for (1S,2R)-10·HCl: C, 52.47; H, 7.77; N, 6.62.

4. Conclusions

A highly efficient CAL-B (lipase B from Candida antarctica)-catalyzed ring-cleavage reaction (E > 200) of 9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-3)] with 1 equiv. of H₂O in iPr₂O at 60 °C resulted new eight-membered ring-fused β-lactam (1S,8R)-3 and β-amino acid (1R,2S)-9 (ee ≥ 99%). Even more efficient two-step transformation was carried out for the ring cleavage of activated lactams (±)-4 with 1 equiv. of H₂O, in the presence of CAL-B using 1 equiv. of benzylamine in iPr₂O at 60 °C. The ring cleavage of regioisomeric (±)-6, carried out under the same conditions resulted unreacted (1S,8R)-6, potential intermediate for the synthesis of enantiomeric anatoxin-a. These results underline the importance of substrate engineering, since faster reactions were clearly detected when activated lactams vs. their inactivated counterparts were reacted. Advantages of these reactions are the spontaneous degradation of N-hydroxymethyl groups and easy product separation by organic solvent–H₂O extraction.

An indirect enzymatic method, in view of the synthesis of enantiomeric β-amino acids, has also been devised, through a relatively fast acylation of N-hydroxymethyl 9-azabicyclo[6.2.0]dec-6-en-10-one [(±)-4)] with 10 equiv. of VB mediated by lipase PSIM (Burkholderia cepacia) using a catalytic amount of Et₃N and Na₂SO₄ in iPr₂O at 30 °C (E = 114). This route is somewhat less efficient and, in fact, it is a longer procedure to prepare amino acid enantiomers. However, it ensures the simultaneous preparation of new β-lactam enantiomers [(1S,8R)-4 and (1R,8S)-7, and (1S,8R)-3 and (1R,8S)-3, respectively].