N-octanoyldimethylglycine trifluoroethyl ester, an acyl donor leading to highly enantioselective protease-catalysed kinetic resolution of amines

Article abstract of DOI:10.1002/adsc.201100993

The use of N-octanoyldimethylglycine trifluoroethyl ester as acyl donor in the kinetic resolution of aliphatic amines catalysed by proteases led to enantiomeric ratios >200 in most cases. The resolutions mediated by Protex 6L were shown to be much faster

Full text of DOI:10.1002/adsc.201100993

FULL PAPERS
DOI: 10.1002/adsc.201100993
N-Octanoyldimethylglycine Trifluoroethyl Ester, an Acyl Donor
Leading to Highly Enantioselective Protease-Catalysed Kinetic
Resolution of Amines
Severine Queyroy,c,* Nicolas Vanthuyne,b Stꢀphane Gastaldi,a,*
Michꢁle P. Bertrand,a,* and Gꢀrard Gilb,*
a
Aix-Marseille Univ, Institut de Chimie Radicalaire UMR 7273 and CNRS, Institut de Chimie Radicalaire, Equipe CMO,
13397 Marseille Cedex 20, France
E-mail: stephane.gastaldi@univ-amu.fr or michele.bertrand@univ-amu.fr
Aix-Marseille Univ and CNRS, ISM2 UMR 7313, Equipe Chirosciences, 13397 Marseille Cedex 20, France
b
E-mail: gerard.gil@univ-amu.fr
Aix-Marseille Univ and CNRS, Institut de Chimie Radicalaire UMR 7273, Equipe CTM, 13397 Marseille Cedex 20,
c
France
E-mail: severine.girard-queyroy@univ-amu.fr
Received: December 21, 2011; Revised: February 20, 2012; Published online: June 5, 2012
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201100993.
Abstract: The use of N-octanoyldimethylglycine tri- ficient commercially available serine proteases, i.e.,
fluoroethyl ester as acyl donor in the kinetic resolu- alkaline protease, Properase 1600L, and Subtilisin
tion of aliphatic amines catalysed by proteases led to
enantiomeric ratios >200 in most cases. The resolu-
tions mediated by Protex 6L were shown to be much Keywords: amines; enantioselectivity; enzyme catal-
faster than the resolutions achieved with the most ef- ysis; kinetic resolution
Introduction
enantioselectivity of different serine proteases in the
resolution of amines performed with different acyl do-
The importance of amines in the agrochemical and nors.[6b]
pharmaceutical industries has stimulated an enormous
In preliminary investigations,[6b] in which the amine
interest for their production in optically pure form. acylation was achieved with a large series of peptide
Biocatalysis ranks now among the most important of mimics, Protex 6L turned out to be a non-selective
these processes.[1] A panel of enzymatic routes is catalyst. However, it reacted much faster than any
available including the use of hydrolases, transaminas- other tested enzyme. Our attention was focused on
es, amine oxidases or amine reductases. Whereas en- the ability of Protex 6L to become as selective as the
zymatic resolution of amines by lipases has been ex- most selective proteases, i.e., alkaline protease, Prop-
tensively investigated and excellent results have been erase 1600L, and Subtilisin.[7] This goal was reached
reported, literature data concerning their kinetic reso- by using 3d (Figure 1), a peptide mimetic derived
lution (KR) by proteases are by far less numerous.[2,3] from dimethylglycine, as acyl donor. The results of
During the course of our efforts to devise protocols this study are reported herein.
for the chemoenzymatic dynamic kinetic resolution of
amines through the association of a radical racemisa-
tion procedure with an enzymatic kinetic resolution,[4]
we have had to optimise the kinetic resolution of
amines by both lipase CAL-B[5] and a series of serine
proteases.[6] This led us to report the first switchable
DKR process allowing the synthesis of either (R)- or
(S)-amides from racemic amines depending on the
nature of the enzyme.[4c,d] In order to reach this goal,
we have had to investigate the reactivity and the Figure 1. Acyl donors.
Adv. Synth. Catal. 2012, 354, 1759 – 1764
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1759

Severine Queyroy et al.
FULL PAPERS
A computational approach undertaken in order to results obtained with 4-phenyl-2-butyl amine (1a, see
rationalise the specific behaviour of acyl donor 3d is Figure 3) that was used as model amine all along our
detailed in the Supporting Information.
study are reported in Table 1. As shown in entry 4, N-
octanoyldimethylglycine trifluoroethyl ester (3d) was
far more selective than the corresponding trifluor-
oethyl esters derived from either glycine (3a, entry 1),
l-alanine (3b, entry 2), or l-phenylalanine (3c,
Results and Discussion
The most commonly used acyl donors[8] (e.g., ethyl a- entry 3). The time of reaction was too long for the
methoxyacetate[9]), known to be particularly efficient purpose of using it as reagent in a DKR reaction. It
in lipase KR, were tested first. As already reported was totally incompatible with the very short time
they gave disappointing results with Subtilisin as com- needed for sulfanyl radical-mediated racemisation
pared to peptide mimetics which are close to the nat- procedure (commonly 60 to 90 min). Nevertheless,
ural substrates of proteases.[6a]
the enantiomeric ratio was exceptionally high for
A series of peptide mimetics[10] was tested with al- a protease-catalysed amine KR.[2,3] Such an increase
kaline protease, i.e., the commercially available pro- in itself, obtained with a commercially available
tease that was shown to be the best suited to DKR enzyme, deserved further investigations.
purposes, owing to its compatibility with the radical
As shown in Table 2, the resolution of 1a per-
racemisation conditions at room temperature.[11 The formed with Protex 6L in the presence of 3d in 3-
Table 1. Alkaline protease-catalysed KR of 1a at 218C in the presence of 3a–d.
Entry
Acyl Donor
Time
1a ee [%][b]
2aa–2ad ee [%][c]
Conversion (C) [%][d]
E[e]
1
2
3
4
3a
3b
3c
3d
35 min
90 min
40 min
24 h
83
87
75
73
91
94
94
>99.5
48
48
44
42
57[6b]
99[6b]
72[6b]
>200 (this paper)
[a]
Standard procedure: reactions were performed on a 0.5M solution of 1a (0.125 mmol) in 3-methyl-3-pentanol at room
temperature, with 6 mg of coated enzyme (co-lyophilised in different buffers corresponding to optimal pH with Brijꢃ56
and n-octyl a,b-d-glucopyranoside; 4/1/1, w/w/w) and 1 equiv. of acyl donor.
Determined by GC after derivatisation.
[b]
[c]
[d]
[e]
Determined by HPLC.
Calculated according to C=ee /
Enantiomeric ratios were calculated according to E=ln
(eeS+eeP) unless otherwise stated.
[(1ÀC)(1ÀeeS)]/ln
G
E
[(1ÀC)
(1+eeS)].[12]
ACHTUNGTRENNUNG
Table 2. KR of 1a, achieved with different enzymes at 218C.
Entry
Enzyme
Time
1a ee [%][b]
2ad ee [%][c]
Conversion (C) [%]
E
1
2
3
4
alkaline protease
Properase 1600L
Subtilisin Novo
Protex 6L
24 h
96 h
7 days
10 h
73
53
33
97
>99.5
>99.5
>99.5
>99.5
42
35
25
49
>200
>200
>200
>200
[a]
Standard procedure given in Table 1 unless otherwise stated.
Determined by GC after derivatisation.
Determined by HPLC.
[b]
[c]
1760
asc.wiley-vch.de
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1759 – 1764

N-Octanoyldimethylglycine Trifluoroethyl Ester, an Acyl Donor
Table 3. Solvent effect on the KR of 1a, mediated with Protex 6L and 3d.
Entry
Solvent/Concentration
Time
1a ee [%][b]
2ad ee [%][c]
Conversion (C) [%]
E
1
2
3
4
5
6
7
3-MP/0.5M
t-BuOH/0.5M
10 h
72 h
48 h
48 h
40 h
40 h
40 h
97
29
60
83
71
54
73
>99.5
>99.5
68
62
87
49
22
47
57
45
36
44
>200
>200
9
11
29
3-MP/THF (1/1)/0.5M
3-MP/THF (1/1)/0.1M
3-MP/THF (2/1)/0.1M
3-MP/hexane (2/1)/0.1M
3-MP/toluene (2/1)/0.1M
96
93
82
61
[a]
Standard procedure given in Table 1 unless otherwise stated.
Determined by GC after derivatisation.
Determined by HPLC.
[b]
[c]
methyl-3-pentanol (3-MP) was completed in 10 h. It gated led to very low enantiomeric ratios (examples
was much faster than the resolutions carried out with are given in entries 3–7).
any of the other selected proteases (entries 1–4),
The nature of the coating agents also influenced
under the reported conditions. It can be remembered the reactivity (Figure 2). The enantiomeric excess of
that, for the resolution of 1a, Protex 6L led to 46% the residual (R)-amine reached 28% in 6 h, and did
conversion in 10 min (E=46) in the presence of 3a; it not exceed 40% after 10 h of reaction in the absence
led to 46% conversion in 50 min (E=67) in the pres- of coating. It reached 79 or 85% in 6 h when the
ence of 3b; and to 54% conversion in 54 min (E=61) enzyme was coated with Brij 56+n-octyl a,b-d-gluco-
with 3c.[6b]
pyranoside, or PEG3000+n-octyl a,b-d-glucopyrano-
Further improvement of the KR was achieved by side, respectively. After the same period (6 h) the ee
testing the efficacy of a series of organic solvents was very close to 100% with either n-octyl a,b-d-glu-
(Table 3).
copyranoside alone, or with Me-b-cyclodextrin (Me-b-
Proteases are reputedly less stable in organic sol- CD)+n-octyl a,b-d-glucopyranoside, or even with
vents than lipases, and coating of the enzyme was Triton B+n-octyl a,b-d-glucopyranoside.
shown to significantly affect both the rate of reaction
The 1:1 ratio of Me-b-cyclodextrin/n-octyl a,b-d-
and the enantiomeric ratio.[13] These experiments con- glucopyranoside was selected.[14] The ratio of enzyme
firmed the pioneering results of Klibanov,[3a] that is, to the binary coating was also further optimized.
tertiary alcohols led to the highest enantiomeric
According to the data reported in Table 4. the fast-
ratios. However, 3-methyl-3-pentanol (entry 1) was est resolution was observed for an 8:1:1 ratio of
far superior to tert-butanol (entry 2) regarding the re- Protex 6L/Me-b-cyclodextrin/n-octyl a,b-d-glucopyra-
action rate. The mixtures of solvents that were investi- noside, with which the rate was found to be slightly
superior to that registered for the 4:1:1 ratio (entries 3
and 5). The conditions described in entry 6 were
therefore selected for all further experiments.
The reaction profile was confirmed by repeating
the resolution of amine 1a on a 5-mL scale (the plot
of amine ee/time is given in the Supporting Informa-
tion). The yields and the ee were similar (Table 5).
The optimised experimental conditions were then
applied to a large series of amines of various structur-
al features shown in Figure 3 to investigate which
structural parameters were determining for the dis-
crimination of the two enantiomers. The results are
given in Table 5.
The non-catalysed reaction of amines 1a, 1d, and
1e, with 3d, were followed by 19F NMR (the conver-
sion was determined from the signals of trifluoro-
methyl groups of the donor and of released trifluoroe-
thanol, respectively). After 7 days the amount of non-
catalysed acylation was 0.8%, 1.3%, and 0.7%, re-
spectively. It could therefore be neglected.
Figure 2. Influence of Protex 6L coating on the rate of reso-
lution of amine 1a in 3-methyl-3-pentanol. a) In the case of
only one coating agent: Protex 6L/n-octyl a,b-d-glucopyra-
noside=4/1. b) In the case of two coating agents: Protex
6L/coating agent 1/coating agent 2=4/1/1.
Adv. Synth. Catal. 2012, 354, 1759 – 1764
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1761

Severine Queyroy et al.
FULL PAPERS
Table 4. Influence of the enzyme/coating agents ratio on the resolution rate.
Entry
Protex 6L/Me-b-CD/a,b-D-glucopyranoside
Time [h]
1a ee [%][b]
1
2
3
4
5
6
7
8
2/1/1
2/1/1
4/1/1
4/1/1
8/1/1
8/1/1
16/1/1
16/1/1
3
6
3
6
3
6
3
6
54
81
78
>99.5
81
>99.5
40
70
[a]
Standard procedure. as given in Table 1 except for the coating agents.
Determined by GC after derivatisation.
[b]
Table 5. Resolution of amines 1a–r.
Amine
Time [h]
Amine
ee [%][b]
Amide
ee [%][c]
Conversion (C) [%]
E
Yield
Yield
1a
1a[d]
1b
1c
1d
1e
1f
1g
1h
1i
1j
1k
1l
1m
1n
1o
1p
1q
1r
6
7
72
7
11 days
72
6 days
18 days
7 days
7 days
72
8 days
48 h
48 h
19 days
19 days
6 days
72
96
>99.5
38
90
61
60
63
58
91
76
83
95
97
87
68
38
54
97
88
48%
49%
63%
43%
47%
60%
59%
61%
51%
55%
46%
53%
49%
43%
54%
39%
39%
47%
50%
>99.5
>99.5
>99.5
68
63
96
>99.5
>99.5
>99.5
>99.5
78
99
99
74
96
27
39
98
96
48%
49%
21%
54%
44%
35%
33%
34%
43%
39%
51%
46%
46%
51%
38%
55%
53%
45%
44%
49
49
27
57
49
38
39
37
48
43
52
49
49
54
41
59
58
50
48
>200
>200
>200
14
8
200
>200
>200
>200
>200
21
>200
>200
18
103
2.4
4
>200
153
7 days
[a]
See general procedure.
Determined by GC after derivatisation.
Determined by HPLC.
[b]
[c]
[d]
20-fold scale-up.
Unusually high enantiomeric ratios were observed ortho-substituted amine 1n). The saturated equivalent
in most cases. The lowest E factors were noted for of amine 1c, that is, amine 1d behaved similarly.
benzylic amines (with the exception of the hindered There is clearly an effect of the bulk of the substitu-
1762
asc.wiley-vch.de
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1759 – 1764

N-Octanoyldimethylglycine Trifluoroethyl Ester, an Acyl Donor
donors. The gem-dimethyl effect seems to facilitate
the folding of the lipophilic octanoyl chain. This
makes for a more difficult access to the hindered re-
active carbonyl group. Therefore, the gem-dimethyl
group promotes the formation of non-productive con-
formations of the acyl-enzyme complex in which the
apolar chain lies out of the enzyme channel. This con-
formational peculiarity might be responsible for the
slowing down of the reactions performed in the pres-
ence of 3d.
Conclusions
Simple acyl donor engineering, i.e., the use of N-octa-
noyldimethylglycine trifluoroethyl ester, was found to
have a remarkable influence on the enantioselectivity
of protease-catalysed resolution of aliphatic amines.
Enantiomeric ratios >200 are viable and can have
practical application for a large series of amines using
a commercially available enzyme, Protex 6L.
Figure 3. Amine structures.
Experimental Section
ent in position beta relative to nitrogen which is likely
to hinder the approach of the acyl enzyme.
General Experimental Methods and Characterisation
Table 2 shows that very high enantiomeric factors of New Products
can be obtained with different proteases, sharing the
See the Supporting Information
same catalytic triad. Subtilisin Novo (entry 3) was
used as a model enzyme, to approach by docking cal-
culations the affinity of this protease for acyl donors
3a–d. Such an approach of a multistep process leading
to chiral amides from racemic amines should be taken
cautiously but it could give interesting information
about which step is likely to determine the rate of the
resolution process.
Protease Coating Procedure
Octyl a,b-d-glucopyranoside (60 mg) and methyl-b-cyclo-
dextrin (60 mg) were diluted at 608C in 50 mL of buffer
(phosphate 0.1M, pH 7), the solution was then cooled to
room temperature. The Protex 6L liquid extract (500 mg)
was added to 25 mL of this solution. The mixture was rapid-
ly frozen in liquid N2 and lyophilised for 20 h.
Subtilisin Novo had already been used as model in
previous studies.[4b] In the latter, several structures of
Subtilisin Novo were tested, but their docking results
were not significatively different. This is the reason
why only the 2 sni structure of the PDB database was
selected.
General Procedure for the Kinetic Resolution of 1a–r
To a solution of acyl donor (0.187 mmol, 1.5 equiv.) in 3-
methyl-3-pentanol (250 mL), 10 mg of coated protease and
amine (1a–r) (0.125 mmol, 1 equiv.) were added. The result-
ing mixture was stirred at 218C. The kinetic resolution was
monitored by analysing aliquots at different time intervals.
The amide ee was determined by HPLC after dilution in the
eluting solvent. The amine ee was determined by GC after
dilution in diethyl ether and derivatisation in trifluoroaceta-
mide using 1.5 equiv. of N-methyl-bis-trifluoroacetamide.
The different acyl donors were docked into the
enzyme using Autodock 4.0.[15]
The computational details are reported in the Sup-
porting Information. They suggest that the four acyl
donors have rather close affinities (3c exhibits the
strongest one).
The dynamic conformational behaviour of acyl-
enzyme complexes formed from 3a–d was sampled.
Their conformational flexibility was examined by
means of systematic torsion around the different ro-
tatable bonds of the acyl moiety. The partition of con-
formations with a significant contribution in each
series was analysed. In the case of 3d the contribution
of conformations in which the chain is folded is much
more important than in the cases of the other acyl
Supporting Information
Computational studies are detailed in the Supporting Infor-
mation as are also the synthesis of 3d and characterisation
data of all new kinetic resolution products [2ad–2rd and
Boc-(R)-1l, Boc-(R)-1m, Boc-(R)-1n, Boc-(R)-1o, Boc-(R)-
1p, Boc-(R)-1r].
Adv. Synth. Catal. 2012, 354, 1759 – 1764
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1763

Severine Queyroy et al.
FULL PAPERS
[5] M. Nechab, N. Azzi, N. Vanthuyne, M. P. Bertrand, S.
Gastaldi, G. Gil, J. Org. Chem. 2007, 72, 6918.
[6] a) M. Nechab, L. El Blidi, N. Vanthuyne, S. Gastaldi,
M. P. Bertrand, G. Gil, Org. Biomol. Chem. 2008, 6,
3917; b) A.-L. Bottalla, S. Queyroy, N. Azzi-Schue, N.
Vanthuyne, S. Gastaldi, M. P. Bertrand, G. Gil, Tetrahe-
dron: Asymmetry 2009, 20, 2823.
[7] Among the four commercially available enzymes, only
Subtilisin Novo is a solid, the others are provided as
mother liquors.
[8] For a general review on acyl donors, see: U. Hanefeld,
Org. Biomol. Chem. 2003, 1, 2405 and relevant referen-
ces cited therein.
[9] a) F. Balkenhohl, K. Ditrich, B. Hauer, W. Ladner, J.
Prakt. Chem./Chem.-Ztg. 1997, 339, 381; b) F. Balken-
hohl, B. Hauer, W. Ladner, U. Pressler, C. Nꢄbling,
(BASF AG), Patent WO 95/08636, 1993; Chem. Abstr.
1995, 122, 289035.
[10] The interest in using N-acylamino esters as acyl donor
resides in the binding affinity of proteases for these
donors. Their structure is closely related to those of the
natural substrates of proteases. The use of N-protected
amino acid carbamoyl esters as acyl donors was also
shown to accelerate the rate the a-CT-catalysed amida-
tion of chiral amines, see: T. Miyazawa, N. Yabuuchi,
R. Yanagihara, T. Yamada, Biotechnol. Lett. 1999, 21,
725.
Acknowledgements
Gifts of alkaline protease, from Valley Research, of Subtilisin
Novo from Novozymes, and gifts of Protex 6L, and Proper-
ase 1600L from Genencor, are gratefully acknowledged. This
work was supported by the Agence Nationale de la Recher-
che (ANR 06-Blan-0137). The authors thank Dr Hugues-
Olivier Bertrand (Accelrys) for fruitful discussions and Dr
Anne-Lise Bottalla for her contribution. All of the computa-
tions were performed with thanks to the large-scale facilities
of the Centre de Calcul et de Recherche Technologique
(CCRT) of the French Nuclear Agency (CEA) and at the
Centre Regional de Competences en Modelisation Molecu-
laire (CRCMM) in Marseille.
References
[1] M. Hçhne, U. T. Bornscheuer, ChemCatChem 2009, 1,
42–51, and reference cited therein.
[2] a) U. T. Bornscheuer, R. J. Kazlauskas, in: Hydrolases
in Organic Synthesis Regio- and Stereoselective Bio-
transformations, 2nd edn., chap. 6, Wiley-VCH, Wein-
heim, 2006 and references cited therein; b) F. van Rant-
wijk, R. A. Sheldon, Tetrahedron 2004, 60, 501, and ref-
erences cited therein.
[3] For other examples of protease-mediated amine acyla-
tion, see: a) H. Kitagushi, P. A. Fitzpatrick, J. E. Huber,
A. M. Klibanov, J. Am. Chem. Soc. 1989, 111, 3094;
b) A. L. Gutman, E. Meyer, E. Kalerin, F. Polyack, J.
Sterling, Biotechnol. Bioeng. 1992, 40, 760; c) B. Orsat,
P. B. Alper, W. Moree, C.-P. Mak, C.-H. Wong, J. Am.
Chem. Soc. 1996, 118, 712; d) Y. F. Wang, K. Yakolev-
sky, A. L. Margolin, Tetrahedron Lett. 1996, 37, 5317;
e) Y. F. Wang, K. Yakolevsky, B. Zhang, A. L. Margo-
lin, J. Org. Chem. 1997, 62, 3488; f) S. Takayama, W. J.
Moree, C.-H. Wong, Tetrahedron Lett. 1996, 37, 6287.
[4] a) S. Gastaldi, S. Escoubet, N. Vanthuyne, G. Gil, M. P.
Bertrand, Org. Lett. 2007, 9, 837; b) L. El Blidi, M.
Nechab, N. Vanthuyne, S. Gastaldi, M. P. Bertrand, G.
Gil, J. Org. Chem. 2009, 74, 2901; c) L. El Blidi, N. Van-
thuyne, D. Siri, S. Gastaldi, M. P. Bertrand, G. Gil, Org.
Biomol. Chem. 2010, 8, 4165; d) F. Poulhꢁs, N. Van-
thuyne, M. P. Bertrand, S. Gastaldi, G. Gil, J. Org.
Chem. 2011, 76, 7281.
[11] L. Routaboul, N. Vanthuyne, S. Gastaldi, G. Gil, M. P.
Bertrand, J. Org. Chem. 2008, 73, 364.
[12] E values were calculated according to: C.-S. Chen, Y.
Fujimoto, G. Girdaukas, C. Sih, J. Am. Chem. Soc.
1982, 104, 7294.
[13] For a general review, see: F. Bordusa, Chem. Rev. 2002,
102, 4817, and references cited therein.
[14] Keeping in mind a possible application to amine DKR,
this coating agent was preferred to the simple coating
of the enzyme with n-octyl a,b-d-glucopyranoside, be-
cause the so-coated enzyme was stable to irradiation.
[15] a) R. Huey, G. M. Morris, A. J. Olson, D. S. Goodsell, J.
Comput. Chem. 2007, 28, 1145; b) R. Huey, D. S. Good-
sell, G. M. Morris, A. J. Olson, Lett. Drug Des. Discov-
ery 2004, 1, 178; c) G. M. Morris, R. Huey, W. Lind-
strom, M. F. Sanner, R. K. Belew, D. S. Goodsell, A. J.
Olson, J. Comput. Chem. 2009, 30, 2785.
1764
asc.wiley-vch.de
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1759 – 1764