Lipases in β-dipeptide synthesis in organic solvents

Article abstract of DOI:10.1021/ol0623163

A number of β-dipeptides were prepared by two-step lipase-catalyzed reactions where N-acetylated β-amino esters were first activated as 2,2,2-trifluoroethyl esters with Candida antarctica lipase B (CAL-B). The activated esters were then used to acylate a

Full text of DOI:10.1021/ol0623163

ORGANIC
LETTERS
2006
Vol. 8, No. 24
5593-5596
Lipases in
Organic Solvents
â-Dipeptide Synthesis in
Xiang-Guo Li and Liisa T. Kanerva*
Department of Pharmacology, Drug DeVelopment and Therapeutics/Laboratory of
Synthetic Drug Chemistry and Department of Chemistry, UniVersity of Turku,
Lemminkaisenkatu 5C, FIN- 20520 Turku, Finland
lkanerVa@utu.fi
Received September 20, 2006
ABSTRACT
A number of
â
-dipeptides were prepared by two-step lipase-catalyzed reactions where N-acetylated
â
-amino esters were first activated as
2,2,2-trifluoroethyl esters with Candida antarctica lipase B (CAL-B). The activated esters were then used to acylate a
â-amino ester in the
presence of Candida antarctica lipase A (CAL-A) in dry Et2O or i-Pr2O.
Proteases (also called peptidases) are widely used as catalysts
for enzymatic peptide synthesis, although many synthesis
limiting facts (enzymatic hydrolysis of the formed peptide
chain, enzymatic instability under nonaqueous conditions
where hydrolysis could be prevented, and often limited sub-
strate specificity to certain coded amino acids) exist.1 On
the other hand, lipases are relatively stable and show wide
substrate specificity, and although they catalyze peptide bond
formation, they do not possess amidase activity which is
essential for splitting peptide bonds. Thus far, the use of
lipases in peptide synthesis has been limited to PPL (porcine
pancreatic lipase)- and CRL (lipase from Candida rugosa)-
catalyzed reactions of N-protected amino esters as electro-
philes and free amino esters as nucleophiles.2 These lipase-
catalyzed methods are typically based on the use of coded
or noncoded R-amino acids. In the case of PPL, some doubts
have been expressed due to possible contamination by
proteases.2b
pects.3 Contrary to R-peptides, in vivo tests with proteolytic
enzymes have shown â-peptides to be completely stable
toward proteolysis owing to very high metabolic stability.3a
Peptidomimetic activities of some â-peptides are well
recognized. A â-tetrapeptide (N-Ac-â3hThr-â2hLys-â3hTrp-
â3hPhe-NH2) is an example.3b It forms a â-turn similar to
that in a peptide hormone somatostatin where R-L-Trp and
R-L-Lys residues situate in the turn. Accordingly, the above-
mentioned â-tetrapeptide nicely imitates the R-peptidic
ligand, acting as a potent agonist at the human somatostatin
hsst4 receptor.
Research interest in our laboratory has focused on using
lipases as chiral catalysts in the enantioselective N-acylation
(2) (a) Margolin, A. L.; Klibanov, A. M. J. Am. Chem. Soc. 1987, 109,
3802. (b) West, J. B.; Wong, C.-H. Tetrahedron Lett. 1987, 28, 1629. (c)
Kawashiro, K.; Kaiso, K.; Minato, D.; Sugiyama, S.; Hayashi, H.
Tetrahedron 1993, 49, 4541. (d) So, J.-E.; Kang, S.-H.; Kim, B.-G. Enzyme
Microb. Technol. 1998, 23, 211. (e) Zhang, L.-Q.; Zhang, Y.-D.; Xu, L.;
Li, X.-L.; Yang, X.-C.; Xu, G.-L.; Wu, X.-X.; Gao, H.-Y.; Du, W.-B.;
Zhang, X.-T.; Zhang, X.-Z. Enzyme Microb. Technol. 2001, 29, 129. (f)
Maruyama, T.; Nakajima, M.; Kondo, H.; Kawasaki, K.; Seki, M.; Goto,
M. Enzyme Microb. Technol. 2003, 32, 655.
(3) (a) Seebach, D.; Beck, A. K.; Bierbaum, D. J. Chem. BiodiVers. 2004,
1, 1111. (b) Gademann, K.; Kimmerlin, T.; Hoyer, D.; Seebach, D. J. Med.
Chem. 2001, 44, 2460. (c) Fu¨lo¨p, F.; Martinek, T. A.; To´th, G. K. Chem.
Soc. ReV. 2006, 36, 323.
During the past few years, interest in the enantiomers of
â-amino acids and further in â-peptides has increased
tremendously from both pharmaceutical and chemical as-
(1) Bordusa, F. Chem. ReV. 2002, 102, 4817.
10.1021/ol0623163 CCC: $33.50
© 2006 American Chemical Society
Published on Web 10/26/2006

of â-amino esters with achiral esters in organic media.4 In
these studies, the two Candida antarctica lipases, lipases A
(CAL-A) and B (CAL-B), revealed an interesting behavior,
CAL-A being a highly chemo- and enantioselective N-
acylation catalyst.4c,e On the other hand, CAL-B gave
reactions both at the amino (N-acylation) and ester (inter-
esterification) functions of an amino ester, chemoselectivity
varying with the structure of an amino ester and an achiral
acyl donor.4b,d,e
Thus far, enzymatic methods have not been used for the
preparation of â-peptides. Inspired by the above results, we
have now initiated studies of lipase-catalyzed â-peptide
synthesis, where CAL-B allows the activation of a N-
protected amino ester 2 as the 2,2,2-trifluoroethyl ester 3
(Scheme 1). This product is then used without purification
Table 1. Used Amino Esters
compd
R1
R2
R3
ee(R) (%)
ee(S) (%)
1a
1b
1b
1c
1d
1e
H
H
H
Me
Ph
H
H
Et
Et
Et
Et
Me
Me
Me
Me
H
H
H
88
95
99
99
68
99
99
99
(enantiomeric excess values, ee, are given in Table 1), as
previously described.4b-e From these, 1a, rac-1b-d, (R)- and
(S)-1b, and (S)-1c were N-protected with the aid of acetic
anhydride, leading to the corresponding substrates 2 for
interesterification. Accordingly, one of the substrates, 2 (0.1
M), was incubated with CAL-B (20-50 mg mL-1, Novozym
435) in the presence of 2,2,2-trifluoroethyl butanoate (0.2
M) in dry diethyl ether (Table 2). tert-Butyl methyl ether
Scheme 1. CAL-B-Catalyzed Interesterification Leading to
Activated N-Acylated Amino Esters
Table 2. CAL-B-Catalyzed Transformation of 2a-d into
3a-d with 2,2,2-Trifluoroethyl Butanoatea
entry substrate (ee) time (h) ee(S)-2 (%) ee(R)-3 (%) C (%)
1
2
3
4
5
6
7
8
2a
rac-2b
(S)-2b (99%)
(R)-2b (95%)
rac-2c
rac-2c
(S)-2c (99%)
rac-2d
24
48
48
48
3
24
24
106
60
41
99
90 (R)
50
64
62
99 (S)
99
90
57
40b
40b
40b
36
to acylate the amino group of 1 in the presence of CAL-A
(Scheme 2). For this purpose, the amino esters 1a-e (Table
1) were prepared in both racemic and enantiomeric forms
53
40
3c
99
2 (R)
99 (S)
73 (S)
a 2 (0.1 M) + 2,2,2-trifluoroethyl butanoate (0.2 M) in diethyl ether.
Enzyme: CAL-B (Novozym 435, 20 mg mL-1). b CAL-B (40 mg mL-1).
c CAL-B (50 mg mL-1).
Scheme 2. CAL-A-Catalyzed Dipeptide Synthesisa
(TBME) was also a possible solvent. After 24-48 h, the
transformation into the corresponding 3 tended to stop at an
equilibrium at 40-60% conversion. Because the interest-
erification rate of rac-2d was extremely slow (Table 2, entry
8), 3d was not used for further preparative purposes. The
preferred enantioselection of CAL-B for rac-2 is shown in
Scheme 1.
For the CAL-B-catalyzed interesterification of rac-2b-d
with 2,2,2-trifluoroethyl butanoate, enantioselectivities are
relatively low (Scheme 1, Table 2). Due to the equilibrium
nature of the interesterification, clear drops in ee values are
(4) (a) Liljeblad, A.; Kanerva, L. T. Tetrahedron 2006, 62, 5831. (b)
Gedey, S.; Liljeblad, A.; Fu¨lo¨p, F.; Kanerva, L. T. Tetrahedron: Asymmetry
1999, 10, 2573. (c) Gedey, S.; Liljeblad, A.; La´za´r, L.; Fu¨lo¨p, F.; Kanerva,
L. T. Tetrahedron: Asymmetry 2001, 12, 105. (d) Gedey, S.; Liljeblad, A.;
La´za´r, L.; Fu¨lo¨p, F.; Kanerva, L. T. Can. J. Chem. 2002, 80, 565. (e)
Solyma´r, M.; Liljeblad, A.; La´za´r, L.; Fu¨lo¨p, F.; Kanerva, L. T. Tetrahe-
dron: Asymmetry 2002, 13, 1923.
a Due to various combinations of 1 and 3 in the obtained
dipeptides, R1 () R1*), R2, and R3 () R3*) have the definition
originally given for 1a-e in Table 1.
5594
Org. Lett., Vol. 8, No. 24, 2006

Table 3. CAL-A-Catalyzed Dipeptide Synthesisa
c
[R]23
D
entry
acyl donor (ee)
nucleophile (ee)
time (h)
dipeptide
yieldb (%)
(10-1deg cm2 g-1
)
ee (%)
1
2
3
4
5
6
7
8
9
10
11
12
13d
14d
15d
16d
17
18
19
2a
3a
3a
3a
3a
3a
3a
3a
rac-1d
1e
rac-1b
24
7
no reaction
4a
(S)-4b
(R)-4b
(S)-4b
(R)-4c
(R)-4c
(S)-4c
(S)-4d
(R)-4d
(S)-4d
(R)-4e
(R,R)-4f
(R,S)-4f
(S,R)-4f
(S,S)-4f
(R)-4g
92
87
75
82
82
77
76
83
61
85
90
70
85
61
82
89
20
26
23
24
44
44
20
21
22
10
81
75
80
72
8
+6.0
-21.1
+17.8
+12.9
+18.9
-18.9
-40.4
+43.4
-43.4
-33.5
-22.5
-103.2
+103.2
+22.5
+13
37
77
74
68
99
99
92
99
99
62
99
99
99
99
57
(R)-1b (88%)
(S)-1b (68%)
rac-1c
(R)-1c (99%)
(S)-1c (99%)
rac-1d
(R)-1d (99%)
(S)-1d (99%)
1a
(R)-1d (99%)
(S)-1d (99%)
(R)-1d (99%)
(S)-1d (99%)
1a
3a
3a
3a
(R)-3b (62%)
(R)-3b (99%)
(R)-3b (99%)
(S)-3b (99%)
(S)-3b (99%)
(R)-3c (57%)
(R)-3c (57%)
(S)-3c (57%)
rac-1c
rac-1c
168
168
no reaction
no reaction
a 3a-c (0.1 M) + 1a-e (0.2 M) in diethyl ether. Enzyme: CAL-A preparation (Chirazyme L5, lyo., Boehringer-Mannheim, 30 mg mL-1 of the preparation
on Celite containing 20% (w/w) of the original enzyme powder).5 b Isolated yield calculated from an acyl donor. c (c 1.0, CHCl3). d DIPE as solvent.
evident when high conversion rates are reached. This is
clearly shown for rac-2c, where ee(R)-3c drops from a value
of 90% to 57% while the respective ee(S)-2 increases only
from 50% to 64% when the reaction proceeds from 36% to
53% conversion (Table 2, entries 5 and 6). It has not been
possible to affect the equilibrium position by changing the
2,2,2-trifluoroethyl butanoate concentration from 0.2 to 0.3
M. In the activation step, low enantioselectivity can be seen
as a benefit, allowing both enantiomers of 2 to be turned
into the 2,2,2-trifluoroethyl ester in a reasonable amount of
time when the enantiomers are available. However, the use
of racemic 2 in interesterification is less beneficial because
the reaction then produces enantiomerically enriched products
3 (Table 2).
Interesterification as shown in Scheme 1 was stopped by
filtering off CAL-B at conversions shown in Table 2. After
evaporation of the excess 2,2,2-trifluoroethyl butanoate (and
the solvent), the resulting mixture of one of the unreacted
ethyl esters 2 and the formed 2,2,2-trifluoroethyl ester 3 (0.1
M) was dissolved in diethyl ether or in diisopropyl ether
(DIPE) without further purification. One of the amino esters
1 (0.2 M) and the CAL-A preparation (30 mg mL-1,
containing 20% (w/w) of the lipase and 12% (w/w) of
sucrose on Celite)5 were added to start the peptide synthesis
(Scheme 2, Table 3). Reactions were stopped when 3 was
totally consumed. The obtained dipeptide 4 and the unreacted
1 and 2 were purified by column chromatography on silica
eluting with dichloromethane containing methanol (3%
(v/v)). As a crucial point, 2 was shown to be unable to acylate
the amino group of 1 in the presence of CAL-A (Table 3,
entry 1), and the separation of 1 and 2 from the dipeptide
made their reuse possible. The secondary enzymatic hy-
drolysis of the formed peptide 4 was not detected. On the
other hand, enzymatic hydrolysis of 3 by the water in the
seemingly dry CAL-A preparation and the competitive
chemical peptide formation were observed. However, the
proportion of the hydrolyzed 3 was only approximately 5%.
For the formation of 4f, dry DIPE was used as a solvent
since the reaction in diethyl ether was very slow.
Alkyl activation in 3 exposes the carbonyl group to a
chemical peptide bond formation in addition to the enzymatic
one. The nature of a nucleophile affects both chemical and
enzymatic reactivities. Accordingly, a chemical reaction
alone was responsible for the dipeptide formation when 1a
or 1e was a nucleophile (Table 3, entries 2, 12, and 17). In
the formation of dipeptide (S)-4d (ee ) 92%, entry 9), the
maximal proportion of a chemical transformation was
approximately 20% in 20 h. This was detected by GC (a
Chrompack CP-Chirasil-L-Val column, 25 m) following the
disappearance of 3a and rac-1d with time without and in
the presence of the CAL-A preparation. For a chemical
transformation, 20% is an overestimation because the
competitive situation between chemical and enzymatic reac-
tions lacks in the absence of the enzyme. It is clear that
CAL-A recognizes the stereostructures of the added nucleo-
philes (entries 3, 6, and 9). In the case of CAL-A, the S-enan-
tiomer of â2-substituted rac-1b was slightly more reactive
than the R-counterpart, while the R-enantiomer was faster
for the CAL-B-catalyzed interesterification. In â3-substituted
1c and 1d, the same stereostructures were recognized by
CAL-A (acylation) and CAL-B (interesterification). This is
in accordance with our earlier observations.4b-e
Thus far, CAL-A has been used only in cases where
activated achiral esters have reacted with a racemic nucleo-
(5) Kanerva, L. T.; Sundholm, O. J. Chem. Soc., Perkin Trans. 1 1993,
2407.
Org. Lett., Vol. 8, No. 24, 2006
5595

phile, recognizing the stereostructure of the nucleophile (the
â-amino group of the â-amino ester).4 In the present work,
the acyl donors, in addition to nucleophiles, can be chiral
(Table 3, entries 12-19). When enantiopure substrates 1d
and 3b are used in the CAL-A-catalyzed dipeptide formation
the resulting product is a pure stereoisomer as shown in Table
3 (entries 13-16). According to these results, the formation
of R,S- and S,S-stereoisomers reaches higher conversions in
a shorter period of time than their antipodes do. The present
results indicate that â2-substituted 3b is accepted as an acyl
donor by CAL-A, whereas the corresponding â3-substituted
acyl donor is unreactive (entries 13-16 vs 18 and 19).
Evidently, the reactions of (R)-3b (entry 12) and (R)-3c (entry
17) with 1a were not enzyme catalyzed.
Lipase-catalyzed reactions (RCONu1 + Nu2H ) Nu1H +
RCONu2) are generally accepted to proceed through a so-
called ping-pong bi-bi mechanism where the first product
(Nu1H) leaves the active site in the formation of an acyl-
enzyme intermediate (RCO2Enz). This intermediate then
reacts with an added nucleophile (Nu2H), leading to the
liberation of another product RCONu,2 and the free enzyme
is ready for another catalytic cycle.4 On the basis of this
mechanism, the terms interesterification (a reaction between
two esters) and transesterification (a reaction between an ester
and an alcohol) are often used as synonyms because in both
cases an alcohol serves as Nu2H. We found it necessary to
differentiate the terms and to point out that the nucleophile
(Nu1H) does not necessarily leave the active site (as expected
on the basis of the mechanism) but rather may stay there
and react as Nu2H when such an enzyme forms an acyl-
enzyme intermediate with 2, leading to the formation of
2,2,2-trifluoroethyl ester 3. As an experimental basis to this
conclusion, 2,2,2-trifluoroethanol which is formed from the
butanoate ester through the CAL-B-catalyzed formation of
an acyl enzyme intermediate is able to form products 3,
whereas a separately added 2,2,2-trifluoroethanol is not. We
have observed similar behavior previously.4
In summary, we have demonstrated a two-step lipase-
catalyzed kinetic approach for the preparation of a number
of â-dipeptides under mild reaction conditions (at room
temperature (23 °C) and without the use of coupling
reagents). The method is based on an activation step
(formation of a N-protected amino ester as a 2,2,2-trifluo-
roethyl ester) through CAL-B-catalyzed interesterification
followed by the CAL-A-catalyzed N-acylation of an amino
ester without purifying the products while replacing the lipase
with another. In peptide synthesis, the amino ester was used
in excess, and the unreacted counterpart was easily separated
and thus recyclable.
Acknowledgment. This work was supported by the
Academy of Finland (Grant No. 109743 to L.K.).
Supporting Information Available: Experimental pro-
cedures and spectral data. This material is available free of
charge via the Internet at http://pubs.acs.org.
OL0623163
5596
Org. Lett., Vol. 8, No. 24, 2006