Enzymatic resolution of a secondary amine using novel acylating reagents

Article abstract of DOI:10.1016/j.tetasy.2004.03.026

Phenyl allylcarbonates are useful acylating agents for the enzymatic resolution of 1-methyl tetrahydroisoquinoline.

Full text of DOI:10.1016/j.tetasy.2004.03.026

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 1427–1430
Enzymatic resolution of a secondary amine using
novel acylating reagents
Gary F. Breen*
GlaxoSmithKline, Old Powder Mills, Leigh, Tonbridge, Kent TN11 9AN, UK
Received 20 January 2004; revised 19 March 2004; accepted 19 March 2004
Available online 9 April 2004
Abstract—Phenyl allylcarbonates are useful acylating agents for the enzymatic resolution of 1-methyl tetrahydroisoquinoline.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
2. Results and discussion
As part of the development of YH1885 1, a potential
treatment for GERD (gastro-oesophageal reflux disease)
and duodenal ulcers,1 access to both enantiomers of
Initially, racemic 1-MTQ 2 was screened against a
variety of common lipases and esterases using a Chiro-
KitTM-TE screening kit from Altus Biologics.13 Ethyl
acetate was used as the acyl donor in diisopropyl ether.
Only one hit from these 28 enzymes was obtained, that
using chiroCLEC-CR, a cross-linked enzyme crystal of
Candida rugosa lipase.14 However, under these condi-
tions there was only 8% of amide 3 formed after 48 h
while the amine that remained was racemic. No other
commercially available Candida rugosa lipases showed
any activity (Fig. 1).
1-methyl tetrahydroisoquinoline (1-MTQ)
required.
2
was
N
HCl
F
N
NH
N
N
H
NH
N
1-MTQ
2
YH1885
1
O
rac-2
3
Since racemic 2 was available in large quantities we
decided to investigate whether an enzymatic resolution
was feasible. Biocatalysis has, of course, long been
recognised as an important synthetic tool in the prepa-
ration of numerous chiral intermediates2;3 and many of
these processes are now carried out on an industrial
scale.4 Lipase mediated asymmetric acylation of chiral
primary amines is becoming increasingly common.5
Surprisingly, there are very few examples of the resolu-
tion of secondary amines in the literature,6–12 even
though such building blocks are commonly used in
pharmaceuticals. The resolutions that have been
reported tend to suffer from poor yields or poor enan-
tiomeric excesses.
Figure 1. Reagents and conditions: Candida rugosa lipase, EtOAc,
iPr2O, 48 h.
In the enzyme catalysed acylation of alcohols, many
activated esters are used as acylating agents.15 Enol
esters are the most commonly used because of the fast
rate of enzymatic acylation. The tautomerism of the
enol to ketone means that the process is irreversible and
product work-up is simple. However, many of these
acylating agents cannot be used with amines because of
high background reactions, which leads to low ee values.
Several acyl donors were therefore screened in order to
increase the rate and selectivity (Fig. 2).
Acyl donors 7–11 reacted with 2 under enzymatic
catalysis using a large excess of the acylating reagent.
There was some background reaction using 8 and 10,
* Tel.: +44-1732-372045; fax: +44-1732-372199; e-mail: gary_breen-1
@gsk.com
0957-4166/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.03.026

1428
G. F. Breen / Tetrahedron: Asymmetry 15 (2004) 1427–1430
O
1759 cmꢀ1 (e.g., R ¼ NO2, 2-Br, 3-Br) were too reactive,
Derivatives with a C@O stretch below 1759 cmꢀ1
O
giving rise to a reaction in the absence of enzyme.
,
O
O
5
4
notably 2-Me and 3-MeO (18 and 20) gave excellent
selectivities and allowed the enzyme loading to be
reduced to 10%. The phenol by products appeared to
cause no enzyme inhibition and can be removed easily
with a base wash at the end of the resolution.
O
O
CF3
O
6
7
O
O
O
O
The reaction was also enhanced by controlling the water
levels because it is known that enzyme reactions in
organic media require some water to maintain the con-
formational integrity of the enzyme.22 In diisopropyl
ether it was found that 0.25% w/w water was required to
enable the reaction to proceed efficiently. One way of
achieving this was to add zinc sulfate heptahydrate to
the solution to act as a water buffer giving a thermo-
dynamic water activity aw of 0.6 (where 1 is the activity
of pure water).23 However it was noted that over time,
the zinc salt caused some degradation and that this level
of water reduced the selectivity of the lipase, presumably
due to hydrolysis of the carbamate formed. Several
other methods have been developed for controlling
water levels in organic media.24 It was more convenient
to attain this water activity by using toluene containing
0.05% w/w water.
O
CF3
9
8
O
O
O
O
O
11
10
Figure 2. Acyl donors studied.
but the best results were seen using 11. Amine 2, which
remained after 24 h, was found to be the (S)-enantiomer
with an ee of 96%. The only drawback from this reso-
lution was that high catalyst loading (33% by weight)
was required to drive the reaction to completion. This
was found to be due to the presence of allyl alcohol,
which is formed as a by-product during the reaction.
Doping small quantities of allyl alcohol into the reaction
mixture severely retarded the rate suggesting that allyl
alcohol was acting as an inhibitor. Alcohols are already
known to act as inhibitors in lipase catalysed reac-
tions.16–19 This inhibition is dependent on the desolva-
tion properties of the alcohol and increases as the
hydrophobicity of the solvent increases.20
Because the amount of water present is much less (even
though its activity is the same), there is much less
background hydrolysis with selectivity being excellent.
A solvent screen had previously shown that toluene
containing no added water was less selective than
diisopropyl ether. Moreover, the enzyme loading could
now be reduced to 2%. To keep the water level constant
over the course of the acylation, the reaction was per-
formed in a closed system containing a reaction flask
attached to another containing saturated sodium chlo-
ride solution at 50 ꢁC. This solution had a water activity
of 0.75,25 slightly above the desired value for the acyl-
ation itself, which could possibly be due to a less than
perfect equilibrium between the two flasks. This closed
system reduced the reaction time from 21 to 8 h. These
optimised conditions provided the (S)-amine with a 46%
yield and an ee of 99.6% whilst the (R)-carbamate was
recovered with a 47% yield and an ee of 98.4% (Fig. 3).
In order to overcome this problem, a series of novel allyl
carbonates 12–21 were prepared using substituted phe-
nols (Table 1). These were easily prepared in high yields
by reacting the phenols with allyl chloroformate under
phase transfer conditions. The reactivity of these car-
bonates can be fine tuned by varying the substituent
group R on the phenol portion of the molecule. In
general, the infrared C@O stretch of the carbonates
correlated with its reactivity.21 In this case, it was found
that those carbonates with a C@O stretch above
Table 1. Phenylallyl carbonate acyl donors
NH
O
NH
(S)-2
O
O
R
rac-2
Compound
R
m(C@O)/cmꢀ1
N
O
12
13
14
15
16
17
18
19
20
21
2-NO2
2-Br
1765.2
1762.6
1762.5
1759.9
1759.7
1759.4
1758.8
1758.7
1757.7
1757.5
O
22
3-NO2
H
3-Br
Figure 3. Reagents and conditions: 18 or 20, Candida rugosa lipase,
toluene, 0.05 wt% water.
4-Br
3-MeO
4-I
The initial enzyme screen was repeated using 3-meth-
oxyphenyl-allylcarbonate 18 as the acyl donor in toluene
with 0.05% w/w water. This time every lipase tested
2-Me
4-MeO

G. F. Breen / Tetrahedron: Asymmetry 15 (2004) 1427–1430
1429
showed some activity. Several other commercially
available Candida rugosa lipases were then reacted with
the (R)-enantiomer to leave (S)-2 with >99% ee and
good recovery.26 One enzyme, Subtilisin protease,
showed some selectivity for the (S)-enantiomer under
these conditions to give (R)-2.
NaOH (10 mL), then dried over MgSO4 and evapo-
rated to give the crude carbonate as an oil (88–97%).
The oils were purified by distillation under reduced
pressure.
4.4. Enzyme resolution using 18
1-MTQ (5 g, 34 mmol) and 3-methoxyphenyl allylcar-
bonate (4.65 g, 22.3 mmol) were stirred in toluene
(50 mL) containing 0.05% w/w water. This can be con-
veniently achieved by stirring toluene and excess water
together then separating off the water to leave a water–
saturated toluene layer. ChiroCLEC-CR (100 mg) was
added and the reaction stirred at 30 ꢁC with the flask
connected in a closed system to another flask containing
a saturated sodium chloride solution at 50 ꢁC. The
reaction was followed by HPLC. After 8 h, the enzyme
was filtered off and washed with toluene (10 mL). The
organic layer was washed with 2 M HCl (2 · 25 mL) and
the combined acid was then washed with toluene
(10 mL). The pH of the aqueous layer was then adjusted
to 12 with 10 M NaOH and then extracted with TBME
(2 · 50 mL). The TBME was dried over MgSO4 and
evaporated to give (S)-1-methyltetrahydroisoquinoline
as a colourless oil (2.3 g, 46%, 99.65 ee). 1H NMR
(CDCl3): d 1.46 (d, 3H, J ¼ 6:8 Hz), 1.9 (br s, 1H), 2.73
(dt, 1H, J ¼ 16:3, 4.8 Hz), 2.87 (m, 1H), 3.02 (m 1H),
3.26 (dt, 1H, J ¼ 12:8, 5.0 Hz), 4.10 (q, 1H, J ¼ 6:8 Hz),
7.10 (m, 4H); MS MHþ 148.3 (100), 138.1 (20). The
toluene layer was stirred with 2 M NaOH (25 mL), then
dried over MgSO4 and evaporated to give carbamate 22
as a colourless oil (3.7 g, 47%, 98.4% ee). 1H NMR
(CDCl3): d 1.47 (d, 3H, J ¼ 6:8 Hz), 2.76 (dt, 1H,
J ¼ 16:1, 3.8 Hz), 2.94 (br s, 1H), 3.29 (br m, 1H), 4.16
(br m, 1H), 4.64 (dt, 2H, J ¼ 5:5, 1.5 Hz), 5.22 (dq, 2H,
J ¼ 10:5, 1.5 Hz), 5.33 (dq, J ¼ 17:2, 1.5 Hz), 5.97 (m,
1H), 7.15 (m, 4H); MS MHþ 232.1 (95), 188.2 (100),
145.1 (75).
3. Conclusions
Substituted phenyl allylcarbonates can be used to
increase the rate of acylation of 1-methytetrahydroiso-
quinoline with lipases. It is anticipated that these phenyl
allylcarbonates could be useful in other enzymatic
resolutions where the reactivity and selectivity can be
altered for a particular substrate by altering the substi-
tution on the phenol.
4. Experimental
4.1. General methods
Racemic 1-methyltetrahydroisoquinoline was obtained
from Yuhan Corporation. All other chemicals were
obtained from commercial sources and used without
further purification. NMR spectra were recorded with a
Bruker 400 MHz spectrometer. HPLC analyses were
performed on a Beckman Coulter 126/166 liquid chro-
matograph using a Phenomenex Aqua 150 · 4.6 mm 5 l
C18 column. The eluents used were 0.1% methanesulfo-
nic acid in water and 0.1% methanesulfonic acid in
acetonitrile, at 15–90% organic over 15 min then holding
at 90% organic for 3 min, with solvent flow at 1 mL/min
and UV detector at 210 nm. Chiral HPLC analysis was
performed using a Chiralcel-OD column, isocratic
analysis with 3% hexane in methanol as eluent, 1.5 mL/
min solvent flow and UV detector at 220 nm. MS anal-
yses were performed on a Waters Micromass ZQ.
References and notes
4.2. Screening procedure
1. Lee, J. W.; Chae, J. S.; Kim, C. S.; Kim, J. K.; Lim, D. S.;
Shon, M. K.; Choi, Y. S.; Lee, S. H. Patent WO9605177,
1996.
2. Patel, R. N. Enzyme Microb. Technol. 2002, 31, 804–826.
3. Gotor, V. Org. Proc. Res. Dev. 2002, 6, 420–426.
4. Straathof, A. J. J.; Panke, S.; Schmid, A. Curr. Opin.
Biotechnol. 2002, 13, 548–556.
1-Methyltetrahydroisoquinoline (115 mg) was dissolved
in diisopropyl ether (30 mL). One millilitre of this solu-
tion was added to a series of vials in a ChiroKitTM-TE
screening kit. Ethyl acetate (100 lL) was added to each
vial. The kit was shaken in an orbital shaker at 30 ꢁC for
48 h. A sample was taken from each vial to determine
whether any reaction had taken place.
5. van Rantwijk, F.; Sheldon, R. A. Tetrahedron 2004, 60,
501–519.
6. Morgan, B.; Zaks, A.; Dodds, D. R.; Liu, J.; Jain, R.;
Megati, S.; Njoroge, F. G.; Girijavallabhan, V. M. J. Org.
Chem. 2000, 65, 5451–5459.
7. Chiou, T.-W.; Chang, C.-C.; Lai, C.-T.; Tai, D.-F. Bioorg.
Med. Chem. Lett. 1997, 7, 433–436.
4.3. General procedure for the synthesis of carbonates 12–
21
8. Orsat, B.; Alper, P. B.; Moree, W.; Mak, C.-P.; Wond,
C.-H. J. Am. Chem. Soc. 1996, 118, 712–713.
9. Chen, S.-T.; Chen, S.-Y.; Kao, C.-L.; Wang, K.-T. Bioorg.
Med. Chem. Lett. 1994, 4, 443–448.
The phenol (50 mmol) and tetra-n-butylammonium
chloride hydrate (100 mg, 0.35 mmol) were dissolved in
dichloromethane (40 mL) and stirred with 4 M NaOH
(20 mL) at 0–5 ꢁC. Allyl chloroformate (6 mL,
56.5 mmol) was added slowly. After 1 h, the two layers
were separated and the organic layer washed with 2 M
ꢀ
10. Herradon, B.; Valverde, S. Synlett 1995, 599–602.
11. Gutman, A. L.; Meyer, E.; Yue, X.; Abell, C. Tetrahedron
Lett. 1992, 33, 3943–3946.

1430
G. F. Breen / Tetrahedron: Asymmetry 15 (2004) 1427–1430
ꢀ
12. Asensio, G.; Andreu, C.; Marco, J. A. Tetrahedron Lett.
1991, 32, 4197–4198.
20. Garcıa-Alles, L. F.; Gotor, V. Biotechnol. Bioeng. 1998,
59, 163–170.
13. This kit is no longer commercially available.
14. Margolin, A. L. Trends Biotechnol. 1996, 14, 223–230.
15. Klibanov, A. M. Acc. Chem. Res. 1990, 23, 114–120.
16. Madhav, M. V.; Ching, C. B. J. Chem. Technol. Biotech-
nol. 2001, 76, 941–948.
17. Bousquet-Duboch, M.-P.; Graber, M.; Sousa, N.;
Lamare, S.; Legoy, M.-D. Biochim. Biophys. Acta 2001,
1550, 90–99.
18. Janssen, A. E. M.; Vaidya, A. M.; Halling, P. J. Enzyme
Microb. Technol. 1996, 18, 340–346.
19. Gandhi, N. N.; Sawant, S. B.; Joshi, J. B.; Mukesh, D.
Enzyme Microb. Technol. 1995, 17, 373–380.
21. Takayama, S.; Lee, S. T.; Wong, C.-H. J. Chem. Commun.
1999, 127–128.
22. Khan, S. A.; Halling, P. J.; Bell, G. Enzyme Microb.
Technol. 1990, 12, 453–458.
23. Halling, P. J. Biotechnol. Technol. 1992, 6, 271–276.
24. Wehtje, E.; Kaur, J.; Adlercreutz, P.; Chand, S.; Mattias-
son, B. Enzyme Microb. Technol. 1997, 21, 502–510.
25. Apelblat, A.; Korin, E. J. Chem. Thermodyn. 1998, 30, 59–
71.
26. Other Candida rugosa lipases found to be active under
these conditions are Biocatalysts L034P and Euroform
lipases 1 and 4.