Enzymatic method for the synthesis of blockbuster drug intermediates - Synthesis of five-membered cyclic γ-amino acid and γ-lactam enantiomers

Article abstract of DOI:10.1002/ejoc.200800723

A very efficient enzymatic method was developed for the synthesis of cyclic γ-lactam and γ-amino acid enantiomers, intermediates for drugs with a prominent turnover (e.g., abacavir and carbovir), through the CAL-B-catalysed enantioselective (E > 200) hydrolysis of the corresponding N-Boc protected and unprotected racemic γ-lactams with H₂O in iPr₂O. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200800723

FULL PAPER
DOI: 10.1002/ejoc.200800723
Enzymatic Method for the Synthesis of Blockbuster Drug Intermediates –
Synthesis of Five-Membered Cyclic γ-Amino Acid and γ-Lactam Enantiomers
Eniko˝ Forró[a] and Ferenc Fülöp*[a]
Keywords: Amino acids / Enzyme catalysis / Lactams
A very efficient enzymatic method was developed for the
synthesis of cyclic γ-lactam and γ-amino acid enantiomers,
intermediates for drugs with a prominent turnover (e.g., aba-
cavir and carbovir), through the CAL-B-catalysed enantiose-
lective (E Ͼ 200) hydrolysis of the corresponding N-Boc pro-
tected and unprotected racemic γ-lactams with H2O in iPr2O.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
thesis of active ingredients of drugs and the search for bio-
logically active compounds, several methods have been de-
veloped for the resolution of cyclic γ-amino acids; for exam-
ple, the conventional method of diastereomeric salt forma-
tion is used for resolution. The method leads to pure
enantiomers through a multistep process starting from race-
mic 2-azabicyclo[2.2.1]hept-5-en-3-one.[10]
Besides the conventional resolution methods, enzymatic
resolution processes have also been described. For example,
Csuk et al.[11] present the preparation of the appropriate
Introduction
In the series of cyclic cis γ-amino acids, 4-aminocy-
clopent-2-ene-1-carboxylic acid, 3-aminocyclopentane-1-
carboxylic acid and the enantiomers thereof are known
compounds; owing to their GABA analogue structures,
they possess GABA receptor agonist, partial agonist and
antagonist activities,[1] and they can additionally be used in
the synthesis of peptides with definite nonnatural secondary
structures and for the design and synthesis of self-assemb-
ling peptide nanotubes.[2] (In this specification, the follow-
ing numbering and nomenclature are applied for the com-
pounds: cis-4-aminocyclopent-2-ene-1-carboxylic acid and
its (1S,4R)-(–) and (1R,4S)-(+) enantiomers, cis-3-aminocy-
clopentane-1-carboxylic acid, 2-azabicyclo[2.2.1]hept-5-en-
3-one and 2-azabicyclo[2.2.1]heptan-3-one.)
One of the enantiomers of cyclic cis γ-amino acids,
(1S,4R)-4-aminocyclopent-2-ene-1-carbocyclic acid, is a key
intermediate used in one of the syntheses of purine and
pyrimidine carbonucleosides with significant antiviral ac-
tivity (e.g., abacavir[3] and carbovir[4]),[5] through known
methods, for example by reduction[6] and a subsequent
coupling reaction.[7] Abacavir is a selective and potent re-
verse-transcriptase inhibitor for the treatment of human im-
munodeficiency virus and hepatitis B virus infections in
adults and children.
(1S,4R)-4-(acylamino)cyclopent-2-ene-1-carboxylic
acid
(49.7%ee) and methyl (1R,4S)-4-(acylamino)cyclopent-2-
ene-1-carboxylate (Ͼ99%ee) through pig liver esterase cat-
alysed enantioselective ester cleavage of the racemic methyl
cis-4-acylaminocyclopent-2-ene-1-carboxylate. By these
methods, γ-amino acid enantiomers can be obtained after
nonenzymatic hydrolysis of the starting racemic 2-azabicy-
clo[2.2.1]hept-5-en-3-one in several steps (acylations, esteri-
fication and deacylation) and resolution is carried out by
selective cleavage of the ester groups linked to the cyclopen-
tene ring with esterase or lipases.
Evans et al.[12] described a process for the resolution of
racemic 2-azabicyclo[2.2.1]hept-5-en-3-one by using the lac-
tamase activity of ENZA-1 (Rhodococcus equi NCIMB
41213) and ENZA-20 (Psedomonas solanacearum NCIMB
40249) strains. The (+) and (–) enantiomers of 2-azabicy-
clo[2.2.1]hept-5-en-3-one and the (–) and (+) enantiomers
of cis-4-aminocyclopent-2-ene-1-carboxylic acid were ob-
tained through opening of the lactam ring. Taylor et al.[13]
reported that the use of the more selective and stable
ENZA-25 and ENZA-22 strains with lactamase activity
was more suitable for the resolution of 2-azabicyclo[2.2.1]-
hept-5-en-3-one.
The antipode γ-amino acid enantiomer is also used as a
starting material for substances exerting valuable biological
activity.[8] Further, the saturated analogue 3-aminocyclo-
pentane-1-carboxylic acid is a starting material applied in
the search for biologically active compounds.[9]
Because formation of the appropriate absolute configura-
tion of the asymmetric centres is essential for both the syn-
In this paper, we report the first direct enzymatic method
that is equally applicable for the synthesis of cyclic γ-lactam
and γ-amino acid enantiomers through the CAL-B (lipase
B from Candida antarctica) catalysed enantioselective (E Ͼ
[a] Institute of Pharmaceutical Chemistry, University of Szeged,
6720 Szeged, Eötvös utca 6, Hungary
Fax: +36-62-545705
E-mail: fulop@pharm.u-szeged.hu
Eur. J. Org. Chem. 2008, 5263–5268
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5263

E. Forró, F. Fülöp
FULL PAPER
200) hydrolysis of the corresponding N-Boc protected and (Table 1, entry 13) and lipases PS, AK and PPL at 45 °C
unprotected racemic γ-lactams in an organic solvent (Table 1, entries 14–16) catalysed the hydrolysis of (Ϯ)-2 se-
(Scheme 2).
lectively, but the reaction rates were relatively low. Chira-
zyme -2 (Table 1, entry 11), Novozym 435 (Table 1, en-
try 12) and Lipolase (Table 1, entry 1) were all promising
catalysts at 60 °C, as they directed the hydrolysis selectively
and faster, although with slight differences in reaction rates.
Lipolase was chosen as the enzyme for further studies and
for the gram-scale resolution.
Results and Discussion
Syntheses of (؎)-1 and (؎)-3
γ-Lactam (Ϯ)-1 was prepared from commercially avail-
able 2-azabicyclo[2.2.1]hept-5-en-3-one (Vince lactam[14]
)
[(Ϯ)-2] by catalytic transfer hydrogenation in the presence
of the hydrogen donor cyclohexene,[15] whereas the reaction
of (Ϯ)-2 with di-tert-butyl dicarbonate[16] resulted in desired
N-Boc-protected γ-lactam (Ϯ)-3 (Scheme 1).
Scheme 2. Lipase-catalysed enantioselective ring cleavage of (Ϯ)-1–
(Ϯ)-3.
Scheme 1. Syntheses of (Ϯ)-1 and (Ϯ)-3.
The catalytic activity of the tested Lipolase was affected
by the amount of added H2O to the reaction mixture
(Table 1, entries 2–6). The shortest reaction time (the time
needed to reach 50% conversion) was obtained for
Lipase-Catalysed Enantioselective Ring Cleavage of (؎)-1–
(؎)-3
On the basis of the results of the enantioselective hydro- 0.5 equiv. of added H2O (Table 1, entry 2). As the amount
lyses of carbocyclic β-lactams[17] and β-amino esters[18] by of added H2O was increased, the reaction became progress-
using lipase B from Candida antarctica in an organic sol- ively slower, though E remained high (when the hydrolysis
vent, we started preliminary experiments in iPr2O with en- was performed in H2O, the reaction reached 45% conver-
zyme screening including lipase PS (Pseudomonas cepacia), sion after 19 h; E Ͼ 200). Furthermore, when the reaction
lipase AK (Pseudomonas fluorescens), lipase AY (Candida was performed without the addition of H2O (Table 1, en-
rugosa), Chirazyme -2, Novozym 435 and Lipolase (all try 3), the H2O in the reaction medium (Ͻ0.1%) or present
three are forms of immobilised lipase B from Candida ant- in the enzyme preparation (Ͻ5%) was still sufficient for the
arctica), Chirazyme -5 (lipase A from Candida antarctica) hydrolysis to proceed. On the principle that “the best sol-
and PPL (porcine pancreatic lipase) (Scheme 2). The reac- vent is no solvent”,[19] the hydrolysis of (Ϯ)-2 was also per-
tions were performed for the model compound (Ϯ)-2 with formed in a solvent-free system: a mixture of thoroughly
H2O (1 equiv.) at 45 or 60 °C. Chirazyme -5 at 60 °C powdered racemic 2 (55 mg, 0.5 mmol), Lipolase (150 mg)
Table 1. Conversion and enantioselectivity of hydrolysis of (Ϯ)-2.[a]
[c]
Entry Reaction time
Enzyme
Temp.
[°C]
H2O
[equiv.]
Conv.
[%]
ees[b]
[%]
eep
E
(30 mgmL–1)
[%]
1
2
140 min
140 min
(50 min)
140 min
140 min
140 min
140 min
17 h
50 min
50 min
50 min
140 min
140 min
64 h
Lipolase
Lipolase
60
60
1
0,5
42
50
(26)
50
36
24
11
43
11
29
48
41
40
15
20
10
14
72
Ͼ99
(34)
Ͼ99
55
32
12
74
12
40
86
68
65
17
25
11
16
Ͼ99
Ͼ99
(Ͼ99)
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ200
Ͼ200
(Ͼ200)
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Lipolase
Lipolase
Lipolase
Lipolase
Lipolase
Lipolase
Lipolase
Lipolase
60
60
60
60
30
45
70
80
60
60
60
45
45
45
–
2
5
10
0.5
0.5
0.5
0.5
1
1
1
1
1
Chirazyme -2
Novozym 435
Chirazyme -5[d]
PPL
64 h
64 h
64 h
Lipase PS[d]
Lipase AK[d]
1
[a] 0.05  substrate in iPr2O. [b] According to GC (Experimental Section). [c] According to GC after double derivatisation (Experimental
Section). [d] Contains 20 wt.-% of lipase adsorbed on Celite in the presence of sucrose.
5264
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5263–5268

Enzymatic Method for the Synthesis of Blockbuster Drug Intermediates
and added H2O (0.25 mmol) was vigorously shaken at Table 3. Effect of quantity of Lipolase on hydrolysis of (Ϯ)-2.[a]
60 °C. The reaction led to 47% conversion in 33 h [ees =
[c]
Entry
Lipolase
Conv.
[%]
ees[b]
[%]
eep
E
[mgmL–1]
[%]
88%, eep Ͼ 99%; E Ͼ 200 (ees and eep stand for the enan-
tiomeric excess of the substrates and the products, respec-
tively)].
1
10
15
18
(98)
29
46
39
33
30
49
56
64
Ͼ99
(Ͼ99)
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ200
(Ͼ200)
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
(50 after 20 h)
Next, we analysed the effects of temperature on the
enantioselectivity and reaction rate. The Lipolase-catalysed
hydrolysis of (Ϯ)-2 with 0.5 equiv. of H2O in iPr2O at 30 °C
proved to be a relatively slow reaction but the enantio-
selectivity was high (Table 1, entry 7). When the reaction
was performed at 45 °C (Table 1, entry 8), the reaction rate
increased; upon further increase of the temperature (to 70
or 80 °C), considerably faster reactions were observed
(Table 1, entries 9 and 10) without a drop in the enantio-
selectivities.
Several solvents were tested to study the solvent effect
in the Lipolase-catalysed hydrolysis of (Ϯ)-2 at 60 °C. The
enzyme was practically inactive in Me2CO (Table 2, en-
try 1). The reaction proceeded faster in tBuOMe and n-hex-
ane (Table 2, entries 5 and 7), more slowly in 1,4-dioxane,
Et2O, toluene and CH2Cl2 (Table 2, entries 2, 4, 6 and 8)
and much more slowly in THF (Table 2, entry 3) than that
in iPr2O (Table 1, entry 2), which all afforded high enantio-
selectivities (E Ͼ 200).
2
3
4
5
6
7
8
9
20
30
23
32
28
25
23
33
36
39
30[d]
30[e]
30[f]
40
50
75
[a] 0.05  substrate in iPr2O, with 0.5 equiv. of H2O, at 60 °C, after
50 min. [b] According to GC (Experimental Section). [c] According
to GC after double derivatisation (Experimental Section). [d] Al-
ready used one time. [e] Already used two times. [f] Already used
three times.
Finally, we analysed the reusability of the enzyme: the
hydrolysis of (Ϯ)-2 was tested with Lipolase that had al-
ready been used in 1, 2 or 3 cycles (Table 3, entries 4–6).
The catalytic activities of the tested Lipolase were progress-
ively slightly lowered, though the enantioselectivity was ap-
parently not affected.
The hydrolysis of (Ϯ)-1 under the optimised conditions
revealed excellent enantioselectivity (E Ͼ 200), whereas a
drop in enantioselectivity (E = 51) occurred for the hydroly-
sis of (Ϯ)-3 under these conditions. In order to increase E,
the reactions were performed at lower temperatures, and we
found E Ͼ 200 when the ring cleavage of (Ϯ)-3 was per-
formed at 30 °C.
Table 2. Conversion and enantioselectivity of hydrolysis of (Ϯ)-2.[a]
[c]
Entry
Reaction
time
Solvent
Conv.
[%]
ees[b]
[%]
eep
E
[%]
1
2
3
4
5
6
7
8
36 h
36 h
36 h
50 min
50 min
36 h
50 min
36 h
Me2CO
1,4-dioxane
THF
no reaction
On the basis of the preliminary results, the gram-scale
resolutions of (Ϯ)-1–(Ϯ)-3 were performed with H2O
(0.5 equiv.) in the presence of Lipolase in iPr2O at 30 °C for
(Ϯ)-3 and at 65 °C for (Ϯ)-1 and (Ϯ)-2. The products were
characterised by a very good enantiomeric excess at close
to 50% conversion. The results are reported in Table 4 and
the Experimental Section.
The transformations involving the hydrolysis of γ-lac-
tams 5, 7 and 9 with 18% aqueous HCl resulted in the
enantiomers of γ-amino acid hydrochlorides 10 and 11
(Scheme 3). Adding 22% HCl/EtOH to amino acids 4, 6
and 8, followed by solvent evaporation resulted in enantio-
22
7
28
7
Ͼ99
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ200
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Ͼ99
Et2O
18
35
33
34
20
21
54
49
51
24
tBuOMe
toluene
n-hexane
CH2Cl2
[a] 0.05  substrate in the solvent tested, 30 mgmL–1 Lipolase and
0.5 equiv. of H2O, at 60 °C. [b] According to GC (Experimental
Section). [c] According to GC after double derivatisation (Experi-
mental Section).
The reaction rate for the hydrolysis of (Ϯ)-2 increased pure hydrochloride salts 4·HCl and 6·HCl. Physical data on
with increasing the amount of enzyme (Table 3, entries 1–3 the enantiomers prepared are reported in the Experimental
and 7–9).
Section.
Table 4. Lipolase-catalysed preparative-scale hydrolyses of (Ϯ)-1–(Ϯ)-3.[a]
γ-Amino acid·(4, 6 and 8)
γ-Lactam (5, 7 and 9)
Yield Isomer ee[c] [α]2D5 (CHCl3)
Reaction Conv.
E
Yield Isomer ee[b]
[α]2D5 (H2O)
time [h]
[%]
[%]
[%]
[%]
[%]
(Ϯ)-1
(Ϯ)-2
(Ϯ)-3
91
4
18
50
50
50
Ͼ200
Ͼ200
Ͼ200
42
48
44
1R,3S
1S,4R Ͼ99
1S,4R 96
98
–10.6[d]
–243[f]
47
46
44
1R,4S Ͼ99
1S,4R Ͼ99
+158[e]
+549[g]
+187[i]
–40.8[h]
1S,4R
97
[a] 30 mgmL–1 enzyme in iPr2O, 0.5 equiv. of H2O, 60 °C for (Ϯ)-1 and (Ϯ)-2, and 30 °C for (Ϯ)-3. [b] Determined by GC after double
derivatisation (Experimental Section). [c] According to GC (Experimental Section). [d] c = 0.35. [e] c = 0.45. [f] c = 0.34. [g] c = 0.26.
[h] c = 0.25. [i] c = 0.28.
Eur. J. Org. Chem. 2008, 5263–5268
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5265

E. Forró, F. Fülöp
FULL PAPER
rier-fixed lipase B from Candida antarctica) was purchased from
Roche Diagnostics Corporation. Lipase PS (Pseudomonas cepacia)
and lipase AK (Pseudomonas fluorescens) were purchased from
Amano Pharmaceuticals, whereas lipase AY (Candida rugosa) was
obtained from Fluka. Racemic 2-azabicyclo[2.2.1]heptan-3-one
[(Ϯ)-2] was purchased from Acros and used after recrystallisation
from iPr2O. The solvents were of the highest analytical grade.
In a typical small-scale experiment, racemic γ-lactam (0.05  solu-
tion) in an organic solvent (1 mL) was added to the lipase tested
(10, 20, 30, 40, 50 or 75 mgmL–1). H2O (0, 0.5, 1, 2, 5 or 10 equiv.)
was added. The mixture was shaken at 30, 45, 60, 70 or 80 °C. The
progress of the reaction was followed by taking samples from the
reaction mixture at intervals and analysing them by gas chromatog-
raphy. The ee values for the unreacted γ-lactam enantiomers were
determined by gas chromatography on a Chromopack Chiralsil-
Dex CB column (25 m) (120 °C for 25 min Ǟ 160 °C) {temperature
rise 10 °Cmin–1; 140 kPa; retention times [min], 5: 29.22 (antipode:
28.82); 7: 26.74 (antipode: 26.12); 9: 27.79 (antipode: 28.04)},
whereas the ee values for the γ-amino acids produced were deter-
mined by using a gas chromatograph equipped with a chiral col-
umn after double derivatisation with (i) CH2N2[21] (Caution! Deriv-
atisation with diazomethane should be performed under a well-venti-
lated hood!); (ii) Ac2O in the presence of 4-dimethylaminopyridine
and pyridine (Chromopack Chirasil-Dex CB column, 120 °C for
25 min Ǟ 160 °C) {temperature rise 10 °Cmin–1; 140 kPa; retention
times [min], 4: 32.78 (antipode: 33.19); 6: 31.02 (antipode: 31.65);
8: 36.66 (antipode: 36.41)}.
Scheme 3. Hydrolyses of (+)-5, (+)-7 and (+)-9.
The absolute configurations were proved by comparing
the [α] values with the literature data (Experimental Sec-
tion), except in the case of 5, when unsaturated γ-lactam
enantiomer (1S,4R)-7 was reduced catalytically in the pres-
ence of cyclohexene as a hydrogen donor to (1R,4S)-5. It
should be mentioned that in all these reactions the same
stereochemical demands are fulfilled; only the sequence of
priority of the substituents on the substrates differs.
Conclusions
The present enzymatic procedure is equally applicable for
the ring cleavage of N-Boc protected and unprotected γ-
lactams, resulting in γ-amino acid and γ-lactam enantio-
mers (Ն96%ee), blockbuster intermediates. The CAL-B-ca-
talysed highly enantioselective reactions (E Ͼ 200) with
added H2O (0.5 equiv.) in iPr2O at 30 or 60 °C afforded the
products in good chemical yields (Ն42%). Further advan-
tages of this method are the facts that the amino group
need not necessarily be protected and the products can be
easily separated. The amino acid enantiomers produced [4
and 6, Ն98%ee, where (1S,4R)-4-aminocyclopent-2-ene-1-
carboxylic acid (6) is used in the synthesis of Carbovir] pre-
cipitates from organic solvents, whereas they are soluble in
H2O, so that they can be easily washed off the surface of
the enzyme: they can be extracted in H2O. Unreacted γ-
lactam enantiomers 5 and 7 (Ͼ99%ee) can be isolated from
organic solvent and submitted to acidic hydrolysis to form
the corresponding γ-amino acid enantiomers (10 and 11;
Ն97%ee). The separation of the N-Boc protected γ-lactam
and γ-amino acid (8 and 9; Ն96%ee) necessitates column
chromatography. An advantage of this process is that in-
stead of the lactamases, which are not readily available, it
uses lipases, which are commercially available, stable, have
high enantioselectivity and can be used on an industrial
scale.[20]
Optical rotations were measured with a Perkin–Elmer 341 polari-
1
meter. H and 13C NMR spectra were recorded with a Bruker Av-
ance DRX 400 spectrometer. Melting points were determined with
a Kofler apparatus.
2-Azabicyclo[2.2.1]heptan-3-one [(؎)-1]: Palladium-on-carbon
(1.2 g) was added to 2-azabicyclo[2.2.1]hept-5-en-3-one [(Ϯ)-2] (2 g,
18.34 mmol) dissolved in a mixture of MeOH (60 mL) and cyclo-
hexene (16 mL). The mixture was heated at reflux for 3 h, and the
catalyst was filtered off. After evaporation, white crystalline 2-aza-
bicyclo[2.2.1]heptan-3-one [(Ϯ)-1] was recrystallised from iPr2O
[1.8 g, 89%; m.p. 69–74 °C]. 1H NMR (400 MHz, CDCl3, 25 °C,
TMS): δ = 1.38–1.94 (m, 6 H, 3ϫCH2), 2.74 (s, 1 H, CHCO), 3.89
(m, 1 H, CHNH), 5.88 (br. s, 1 H, NH) ppm. C6H9NO (111.14):
calcd. C 64.84, H 8.16, N 12.60; found C 64.51, H 8.16, N 12.46.
N-(tert-Butoxycarbonyl)-2-azabicyclo[2.2.1]hept-5-en-3-one [(؎)-3]:
To a solution of 2-azabicyclo[2.2.1]hept-5-en-3-one [(Ϯ)-2] (1 g,
9.17 mmol) in dioxane (10 mL) was added H2O (10 mL) and a 1 
solution of di-tert-butyl dicarbonate (1.74 g, 10.08 mmol) dissolved
in 1,4-dioxane (10 mL) dropwise at 0 °C. The solution was stirred
for 4 h at room temperature, after which the dioxane was evapo-
rated off and the aqueous solution was acidified with a 10% solu-
tion of H2SO4 (pH = 2.5). The solution was extracted with EtOAc
(3ϫ15 mL), and the organic phase was dried (Na2SO4), filtered
and the solvents evaporated to dryness. The resulting white crystal-
line
N-(tert-butoxycarbonyl)-2-azabicyclo[2.2.1]hept-5-en-3-one
[(Ϯ)-3] was recrystallised from iPr2O [1.4 g, 73%; m.p. 60–65 °C].
1H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 1.37–1.42 (overlap-
ping of the s of 9 H, 3ϫCH3 and m of 2 H, CH2), 2.04–2.06 (d, J
= 8.6 Hz, 1 H, CHCO), 2.26–2.28 (d, J = 8.6 Hz, 1 H, CHNH),
6.72–6.97 (m, 2 H, CHCH) ppm. C11H15NO3 (209.24): calcd. C
63.14, H 7.23, N 6.69; found C 63.32, H 7.18, N 6.77.
Experimental Section
Materials and Methods: Lipolase (lipase B from Candida antarc-
tica), 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 purchased from Sigma–Aldrich. CAL-A (lipase
A from Candida antarctica) and Novozym 435 as an immobilised
lipase (lipase B from Candida antarctica) on a macroporous acrylic
resin were purchased from Novo Nordisk. Chirazyme -2 (a car-
Gram-Scale Resolution of 2-Azabicyclo[2.2.1]heptan-3-one [(؎)-1]:
To a solution of racemic 1 (1.00 g, 9.00 mmol) dissolved in iPr2O
(60 mL) was added Lipolase (2 g, 30 mgmL–1) and H2O (81 µL,
4.5 mmol), and the mixture was shaken in an incubator shaker at
5266
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5263–5268

Enzymatic Method for the Synthesis of Blockbuster Drug Intermediates
60 °C for 91 h. The reaction was stopped by filtering off the enzyme
at 50% conversion. The solvent was evaporated, and residue
(1R,4S)-5 crystallised out (470 mg, 47%); the solid was recrystal-
lised from iPr2O {[α]2D5 = +158 (c = 0.45, CHCl3); m.p. 78–81 °C;
Ͼ99%ee}. The filtered off enzyme was washed with distilled H2O
(3ϫ15 mL), and the H2O was evaporated off to yield crystalline γ-
amino acid (1R,3S)-4 (488 mg, 42%). [α]2D5 = –10.6 (c = 0.35, H2O)
{ref.[22] [α]2D1 for (1S,3R) amino acid = +7.0 (c = 2, H2O); m.p.
Ͼ260 °C with decomposition (recrystallised from H2O and
Me2CO), 98%ee}.
graphic separation of the above products was performed on silica
[Silica gel 60 (0.063–0.200 mm), purchased from Merck KGaA],
with elution with ethyl acetate/n-hexane (1:3).
When 22% HCl/EtOH (4 mL) was added to 8 (100 mg) and the
solvent was evaporated off (1S,4R)-6·HCl (57 mg, 79%) was
formed. [α]2D5 = –101.6 (c = 0.4, H2O); m.p. 201–204 °C (EtOH/
Et2O); 95%ee.
1
9: H NMR (400 MHz, [D6]DMSO, 25 °C): Data were similar to
those for (Ϯ)-3. C11H15NO3 (209.24): calcd. C 63.14, H 7.23, N
6.69; found C 62.90, H 7.28, N 6.67.
When 22% HCl/EtOH (4 mL) was added to 4 (200 mg) and the
solvent was evaporated off (1R,3S)-4·HCl (225 mg, 88%) was ob-
tained. [α]2D5 = –10.8 (c = 0.6, H2O). M.p. 177–180 °C (EtOH/Et2O).
99%ee.
5: 1H NMR (400 MHz, CDCl3, 25 °C, TMS): Data were similar to
those for (Ϯ)-1. C6H9NO (111.14): calcd. C 64.84, H 8.16, N 12.60;
found C 64.66, H 8.19, N 12.51.
4: 1H NMR (400 MHz, D2O, 25 °C, TMS): δ = 1.79–2.26 (m, 6 H,
3ϫCH2), 2.79–2.84 (m, 1 H, CHCOOH), 4.85 (m, 1 H, CHNH2)
ppm. C6H11NO2 (129.16): calcd. C 55.80, H 8.58, N 10.84; found
C 55.61, H 8.60, N 10.99.
8: 1H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 1.15–1.19 (m, 2 H,
CH2), 1.37 (s, 9 H, 3ϫCH3), 1.67–1.72 (m, 1 H, CHCOOH), 2.37–
2.42 (m, 1 H, CHNH), 4. 48 (br. s, 1 H, COOH), 5.70–5.82 (m, 2
H, CHCH), 6.85 (br. s, 1 H, NH) ppm. C11H17NO4 (227.26): calcd.
C 58.14, H 7.54, N 6.16; found C 57.91, H 7.60, N 6.06.
6·HCl: 1H NMR (400 MHz, D2O, 25 °C, TMS): Data for 6·HCl
obtained from 8 were similar to those for 6·HCl obtained from 6.
C6H9NO2·HCl (163.60): calcd. C 44.05, H 6.16, N 8.56; found C
44.04, H 6.00, N 8.48.
Hydrolyses of (+)-5, (+)-7 and (+)-9: Enantiomeric 5 (200 mg,
1.79 mmol), 7 (200 mg, 1.83 mmol) or 9 (100 mg, 0.47 mmol) was
dissolved in 18% HCl (10 mL) and heated at reflux for 2 h. The
solvent was evaporated off, and the product was recrystallised from
EtOH and Et2O, which afforded white crystals of 10 {241 mg, 81%;
m.p. 175–177 °C; [α]2D5 = +10.7 (c = 0.5, H2O); 97%ee} or 11 {from
7: 268 mg, 90%; m.p. 208–209 °C; [α]2D5 = +111.1 (c = 0.35, H2O);
Ͼ99%ee; from 9: 64 mg, 82%; m.p. 202–205 °C; [α]2D5 = +100.7 (c
= 0.3, H2O); 95%ee}.
1
4·HCl: H NMR (400 MHz, D2O, 25 °C, TMS): δ = 1.77–2.40 (m,
6 H, 3ϫCH2), 2.97–3.01 (m, 1 H, CHCOOH), 3.72–3.75 (m, 1 H,
CHNH2) ppm. C6H11NO2·HCl (165.62): calcd. C 43.51, H 7.30, N
8.46; found C 43.59, H 7.40, N 8.65.
Gram-Scale Resolution of 2-Azabicyclo[2.2.1]hept-5-en-3-one [(؎)-
2]: The procedure described above was used. The reaction of race-
mic 2 (1.00 g, 9.16 mmol) and H2O (82 µL, 4.58 mmol) in iPr2O
(45 mL) in the presence of Lipolase (1.5 g, 30 mgmL–1) at 60 °C
afforded unreacted (1S,4R)-7 {461 mg, 46%; [α]2D5 = +549 (c = 0.26,
CHCl3), ref.[23] [α]2D5 for (1R,4S) lactam = –513 (c = 0.52, CHCl3);
m.p. 97–100 °C (iPr2O) ref.[23] m.p. 89–90 °C; Ͼ99%ee} and γ-
amino acid (1S,4R)-6 {559 mg, 48%; [α]2D5 = –243 (c = 0.34, H2O);
m.p. Ͼ260 °C (decomp.; H2O/Me2CO); Ͼ99%ee} in 4 h.
(1S,3R)-10: 1H NMR (400 MHz, D2O, 25 °C, TMS): Data were
similar to those for (1R,3S)-4·HCl. C6H11NO2·HCl (165.62): calcd.
C 43.51, H 7.30, N 8.46; found C 43.50, H 7.17, N 8.21.
(1R,4S)-11 (obtained from 7 and 9): Data were similar to those for
(1S,4R)-6·HCl. C6H9NO2·HCl (163.60): calcd. C 44.05, H 6.16, N
8.56; found (from 7) C 44.08, H 6.16, N 8.45; found (from 9) C
43.86, H 6.01, N 8.52.
When 22% HCl/EtOH (5 mL) was added to 6 (200 mg) and the
solvent was evaporated off (1S,4R)-6·HCl (223 mg, 87%) was ob-
tained. [α]2D5 = –110.9 (c = 0.55, H2O); m.p. 205–208 °C (EtOH/
Et2O); Ͼ99%ee.
Acknowledgments
1
7: H NMR (400 MHz, CDCl3, 25 °C, TMS): δ = 2.23–2.42 (m, 2
H, CH2), 3.23 (s, 1 H, CHCO), 4.34 (m, 1 H, CHNH), 5.54 (br. s,
1 H, NH), 6.67–6.80 (m, 2 H, CHCH) ppm. C6H7NO (109.13):
calcd. C 66.04, H 6.47, N 12.84; found C 66.30, H 6.51, N 12.75.
6: 1H NMR (400 MHz, D2O, 25 °C, TMS): δ = 2.00–2.54 (m, 2 H,
CH2), 3.50–3.52 (m, 1 H, CHCOOH), 4.33–4.35 (m, 1 H, CHNH2),
5.94–6.28 (m, 2 H, CHCH) ppm. C6H9NO2 (127.14): calcd. C
56.68, H 7.13, N 11.02; found C 56.89, H 7.10, N 10.92.
The authors acknowledge the receipt of grants from the Hungarian
Scientific Research Fund (OTKA) (K 71938 and T 049407) and a
Bolyai Fellowship for E.F.
[1] R. D. Allan, H. W. Dickenson, J. Fong, Eur. J. Pharmacol.
1986, 122, 339–348; D. L. Crittenden, A. Park, J. Qiu, R. B.
Silverman, R. K. Duke, G. A. R. Johnston, M. J. T. Jordan, M.
Chebib, Bioorg. Med. Chem. 2006, 14, 447–455.
[2] M. Ordonez, C. Cativiela, Tetrahedron: Asymmetry 2007, 18,
3–99.
[3] a) J. Weiss, N. Weis, N. Ketabi-Kiyanvash, C. H. Storch, G.
Caroline, W. E. Haefeli, Eur. J. Pharmacol. 2008, 579, 104–109;
b) S. Ramanathan, G. Shen, J. Hinkle, J. Enejosa, B. P. Kear-
ney, J. Acquir. Immune Defic. Syndr. 2007, 46, 160–166; c) L. J.
Waters, G. Moyle, S. Bonora, A. D’Avolio, L. Else, S. Mand-
alia, A. Pozniak, M. Nelson, B. Gazzard, D. Back, M. Boffito,
Antivir. Ther. 2007, 12, 825–830; d) M. A. Wainberg, J. L. Mar-
tinez-Cajas, B. G. Brenner, Fut. HIV Ther. 2007, 1, 291–313; e)
A. Kleemann, J. Engel, B. Kutschher, D. Reichert, Pharmaceu-
tical Substances, Thieme, Stuttgart, 2001, pp. 1–4.
1
6·HCl: H NMR (400 MHz, D2O, 25 °C, TMS): δ = 2.05–2.71 (m,
2 H, CH2), 3.73 (m, 1 H, CHCOOH), 4.40 (m, 1 H, CHNH2),
5.97–6.24 (m, 2 H, CHCH) ppm. C6H9NO2·HCl (163.60): calcd. C
44.05, H 6.16, N 8.56; found C 44.28, H 5.90, N 8.56.
Gram-Scale Resolution of N-(tert-Butoxycarbonyl)-2-azabicy-
clo[2.2.1]hept-5-en-3-one [(؎)-3]: The procedure described above
was used. The reaction of racemic 3 (500 mg, 2.39 mmol) and H2O
(21 µL, 1.18 mmol) in iPr2O (30 mL) in the presence of Lipolase
(0.9 g, 30 mgmL–1) at 30 °C afforded unreacted (1S,4R)-9 {223 mg,
44%; [α]2D5 = +187 (c = 0.28, CHCl3), ref.[24] [α]2D5 for (1R,4S) lactam
= –189 (c = 0.89, CH2Cl2); m.p. 77–80 °C (iPr2O); 96%ee} and γ-
amino acid (1S,4R)-8 {238 mg, 44%; [α]2D5 = –40.8 (c = 0.25, H2O),
ref.[24] [α]2D5 for (1S,4R) amino acid = –40.3 (c = 2.1, CH2Cl2); m.p.
130–132 °C (H2O/Me2CO); 96%ee} in 18 h. Column chromato-
[4] a) N. Venhoff, B. Setzer, K. Melkaoui, U. A. Walker, Antivir.
Ther. 2007, 12, 1075–1085; b) B. G. Roy, P. K. Jana, B. Achari,
S. B. Mandal, Tetrahedron Lett. 2007, 48, 1563–1566; c) M.
Eur. J. Org. Chem. 2008, 5263–5268
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5267

E. Forró, F. Fülöp
FULL PAPER
Freiria, A. J. Whitehead, W. B. Motherwell, Synthesis 2005, 18,
3079–3084.
[14]
[15]
[16]
[17]
S. Daluge, R. V. Robert, J. Org. Chem. 1978, 43, 2311–2320;
M. L. Peterson, V. Robert, J. Med. Chem. 1990, 33, 1214–1219.
E. Forró, J. Árva, F. Fülöp, Tetrahedron: Asymmetry 2001, 12,
643–649.
J. A. Campbell, W. K. Lee, H. Rapoport, J. Org. Chem. 1995,
60, 4602–4616.
a) E. Forró, F. Fülöp, Org. Lett. 2003, 5, 1209–1212; b) E.
Forró, F. Fülöp, Chem. Eur. J. 2006, 12, 2587–2592.
[5] a) J. Melroy, V. Nair, Curr. Pharm. Des. 2005, 11, 3847–3852;
b) P. K. Sharma, V. Nair, Recent Res. Developments Org. Chem.
2000, 4, 53–86; c) S. M. Roberts, N. M. Williamson, Curr. Org.
Chem. 1997, 1, 1–20; d) P. S. Hervey, C. M. Perry, Drugs 2000,
60, 447–479; e) J. Melroy, V. Nair, Curr. Pharm. Des. 2005, 11,
3847–3852; f) P. S. Hervey, C. M. Perry, Drugs 2000, 60, 447–
479.
[6] S. M. Daluge, M. T. Martin, B. R. Sickles, D. A. Livingston,
Nucleos. Nucleot. Nucl. 2000, 19, 297–327.
[7] a) M. T. Crimmins, Tetrahedron 1998, 54, 9229–9272; b) M.
Ferrero, V. Gotor, Chem. Rev. 2000, 100, 4319–4347.
[8] D. E. Levy, M. Bao, D. B. Cherbavaz, J. E. Tominson, D. M.
Sedlock, C. J. Homcy, R. M. Scarborough, J. Med. Chem.
2003, 46, 2177–2186.
[9] C. Evans, R. McCague, S. M. Roberts, A. G. Sutherland, J.
Chem. Soc. Perkin Trans. 1 1991, 656–657.
[10] H. Nohira, K. Hara, K. Nagashima, T. Hayashibara, Eur. Pat.
Appl., EP 590685 A1, 1994, 9 pp.
[11] R. Csuk, P. Dörr, Tetrahedron: Asymmetry 1994, 5, 269–276.
[12] C. T. Evans, S. M. Roberts, M. Stanley, Eur. Pat. Appl. EP
424064 A17, 1991, 7 pp..
[13] S. J. C. Taylor, R. McCague, R. Wisdom, C. Lee, K. Dickson,
G. Ruesroft, F. O’Brien, J. Littlechild, J. Bevan, S. M. Roberts,
C. T. Evans, Tetrahedron: Asymmetry 1993, 4, 1117–1128.
[18]
[19]
E. Forró, F. Fülöp, Chem. Eur. J. 2007, 13, 6397–6401.
E. Forró, F. Fülöp, Tetrahedron: Asymmetry 2008, 19, 1005–
1009.
[20] R. Vaidyanathan, L. Hesmondhalgh, S. Hu, Org. Process Res.
Dev. 2007, 11, 903–906.
[21] a) F. Arndt, Org. Synth. 1943, 2, 165; b) J. A. More, D. E. Reed,
Org. Synth. 1973, 5, 351.
[22] M. J. Milewska, T. Polon´ski, Tetrahedron: Asymmetry 1994, 5,
359–362.
[23] K. Tanaka, M. Kato, F. Toda, Heterocycles 2001, 54, 405–410.
[24] S. J. C. Taylor, R. McCague, R. Wisdom, C. Lee, K. Dickson,
G. Ruecroft, F. O’Brien, J. Littlechild, J. Bevan, S. M. Roberts,
C. T. Evans, Tetrahedron: Asymmetry 1993, 4, 1117–1128.
Received: July 21, 2008
Published Online: September 30, 2008
5268
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5263–5268