Chemo- and Enantioselective Acyl Transfers by Lipases and Acylase I: Preparative Applications in Hydroxymethylpiperidine Chemistry

Article abstract of DOI:10.1002/adsc.200303018

The lipase- and acylase I-catalysed acylation of bifunctional 2- and 3-hydroxymethylpiperidines (1 and 2) and the alcoholysis of the corresponding diacylated counterparts 7 and 8 have been studied. Lipase AK from Pseudomonas fluorescens allowed the preparative-scale resolution of 7 in neat butanol at 50% conversion whereas the 3-regioisomer 8 reacted with negligible enantioselectivity (E = 7). The lipase- and acylase I-catalysed acylations of 1 and 2 in organic solvents proceeded with low enantioselectivity. On the other hand, more than 90% of 1 and 2 were transformed to amino esters 3 and 4, respectively, in a highly chemoselective O-acylation with 2,2,2-trifluoroethyl butanoate in the presence of Candida antarctica lipase A in TBME. The O-protected product 3 in the reaction mixture was readily available to another transformation at the piperidine nitrogen although 3 was not separated due to intramolecular O→N acyl migration. The tendency for acyl migration was much less significant in the case of 4.

Full text of DOI:10.1002/adsc.200303018

FULL PAPERS
Chemo- and Enantioselective Acyl Transfers by Lipases and
Acylase I: Preparative Applications in Hydroxymethylpiperidine
Chemistry
Katri Lundell, Petka Lehtinen , Liisa T. Kanerva*
Laboratory of Synthetic Drug Chemistry and Department of Chemistry, University of Turku, Lemmink‰isenkatu 2,
FIN-20520 Turku, Finland
Fax: ( ‡ 358)-2-3337955, e-mail: lkanerva@utu.fi
Received: January 23, 2003; Accepted: February 24, 2003
Abstract: The lipase- and acylase I-catalysed acyla- acylation with 2,2,2-trifluoroethyl butanoate in the
tion of bifunctional 2- and 3-hydroxymethylpiperi- presence of Candida antarctica lipase A in TBME.
dines (1 and 2) and the alcoholysis of the correspond- The O-protected product 3 in the reaction mixture
ing diacylated counterparts 7 and 8 have been studied. was readily available to another transformation at the
Lipase AK from Pseudomonas fluorescens allowed piperidine nitrogen although 3 was not separated due
the preparative-scale resolution of 7 in neat butanol to intramolecular O!N acyl migration. The tendency
at 50% conversion whereas the 3-regioisomer 8 for acyl migration was much less significant in the
reacted with negligible enantioselectivity (E ˆ 7). case of 4.
The lipase- and acylase I-catalysed acylations of 1
and 2 in organic solvents proceeded with low
enantioselectivity. On the other hand, more than Keywords: acyl migration; chemoselective O-acyla-
90% of 1 and 2 were transformed to amino esters 3 tion; enantioselective alcoholysis; N-heterocyclic ami-
and 4, respectively, in a highly chemoselective O- no alcohols; lipase catalysis
Introduction
of 3-hydroxymethylpiperidine 2.[10,11] The enzymatic
preparation of the enantiomers of 2-hydroxymethylpi-
Piperidine alkaloids are typical naturally occurring peridine 1 was less satisfactorily described.[5 9]
compounds which exhibit an extensive range of bio-
In addition to enantioselectivity, chemoselectivity is
logical activities.[1] Piperidine is the nitrogen analogue of characteristic to many enzymes and can be highly
a pyranose ring, and as such its derivatives are potential applicable in synthetic chemistry when polyfunctional
inhibitors of enzymes working on glycosidic bonds.[2,3] molecules are involved. In the case of amino alcohols,
The above cases suffice to emphasise the importance of chemoselective O-acylation is a challenge which, in the
producing piperidine-based structures for synthetic case of usual chemical methods calls, for prior protec-
pharmaceuticals and natural products.
tion of the amino group, specific N-acylations being
For a compound to be biologically and pharmaceuti- more easily controlled. The lipase-catalysed O-acyla-
cally active enantiopurity is of pivotal importance. tion of w-amino alcohols [HO(CH2)nNH2] was previ-
Asymmetric chemical methods to substituted piperi- ously well recognised although fast spontaneous O!N
dines utilising the chiral pool, chiral chemical catalysts acyl migration was shown to lead to the formation of the
and chiral auxiliaries were previously reviewed.[4] The N-acylated final product with n ˆ 2 and 3.[16,17] On this
preparation of small enantiopure building blocks, such basis, preferred O-acylation can also be suggested for
as amino alcohols, using enzymatic kinetic resolution or the enantioselective acylation of 2-amino-1-butanol by
desymmetrisation of a meso compound was also describ- PPL although only the formation of the N-monoacy-
ed.[2,3,5 15] Accordingly, lipases, pig liver esterase and lated and N,O-diacylated products were reported.[18] It is
acylase I were used for the enantioselective O-acylation common to the above-mentioned O!N acyl migrations
of N-protected 2- and 3-hydroxymethylpiperidines (1 that the amino group situates at the position b or g to the
and 2) and for the hydrolysis of the corresponding N- formed ester function. Accordingly, intramolecular acyl
protected amino esters.[5 11] Due to the relatively low migration by normal addition-elimination mechanism
ˆ
enantioselectivities (E 23 70 depending on condi- can proceed through five- and six-membered rings,
tions), kinetic control and repeated resolutions were respectively, leading to the formation of the energeti-
typically described in order to prepare both enantiomers cally more favourable amide product.
790
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/adsc.200303018
Adv. Synth. Catal. 2003, 345, 790 796

Chemo- and Enantioselective Acyl Transfers by Lipases and Acylase I
FULL PAPERS
Table 2. Enzymatic acylation of 5 (R ˆ Me, 0.05 M) with
2,2,2-trifluoroethyl butanoate (0.1 M; R ˆ Pr) in TBME at
room temperature; reaction time 2 h.
Enzyme
Conversion [%]
E
Acylase I
CAL-A
CAL-B
Lipase AK
Lipase PS
PPL
26
95
78
43
83
20
2
1
2
45
3
2
Scheme 1. Compounds 1, 3, 5 and 7 represent 2-substituted
and 2, 4, 6 and 8 3-substituted piperidines.
methods. However, the demand for high enantioselec-
tivity can turn to an insuperable challenge when a
primary alcohol acts as a racemic nucleophile. A
In the present work, the enzymatic acylation of commonly proposed reason is that the reaction centre
piperidine derivatives 1, 2 and 5 and the alcoholysis of is not at the asymmetric centre. In the present work,
the N-acylated amino esters 7 and 8 (R ˆ Me) were enzymatic O-acylation of 1 and 2 and the following
studied in the presence of lipases (EC 3.1.1.5) and intramolecular O!N acyl migration in the formation of
acylase I
(N-acylamino
acid
amidohydrolase, 5 and 6 provide a possibility for double enzymatic
EC 3.5.1.14) under non-aqueous conditions (Scheme 1). acylation at the primary alcohol function in the forma-
The focus has been on developing a method that allows tion of 7 and 8 as final products, respectively.
the preparation of the enantiomers of 2-hydroxyme-
Acylase I, CAL-A, CAL-B, lipase PS, lipase AK and
thylpiperidine 1 in a single step kinetic resolution. PPL were screened for the chemo- and/or enantiose-
Another focus has been on the studies of enzymatic lective acylation of 1 with 2,2,2-trifluoroethyl butanoate
chemoselectivity as a means to protect the primary (R ˆ Pr) in tert-butyl methyl ether (TBME). The
alcohol group in 1 and 2, leaving the secondary amino number of enzymes in the screening was restricted to
group free. As the third topic, acyl migration has been those with the most potential on the basis of the previous
considered.
results for the hydrolysis of N-protected 3-hydroxyme-
thylpiperidine and on our experience for the resolutions
of 2-amino-1-phenylethanols, 1-phenyl-1,2-ethanediol,
and other primary alcohols.[11 13,19 22] The results of the
enzyme screening are shown in Table 1. All the possible
acylation products 3, 5 and 7 were observed for CAL-A
(traces of 7 are evidently due to the fast acylation of 5),
CAL-B, lipase AK and lipase PS while the diacylated
Results and Discussion
Enantio- and Chemoselective Enzymatic Acylation
Highly enantioselective acylation of racemic nucleo- product 7 was not detected for acylase I and PPL in the
philes (alcohols, amines, etc.) with alkyl-activated reaction times of 2 hours. CAL-A and lipase AK as the
irreversible acyl donors in the presence of hydrolytic most interesting enzymes were also screened for the
enzymes is one of the most fascinating resolution acylation of 2 under the same conditions. The product
Table 1. Relative amount [%]/ee [%] for the enzymatic acylation of 1 (0.05 M) and 2 (0.05 M) with 2,2,2-trifluoroethyl
butanoate (0.1 M; R ˆ Pr) in TBME at room temperature.
Acylation of 1[a]
Acylation of 2[b]
Enzyme
1
3
5
7
2
4
6
8
Acylase I
CAL-A
CAL-B
Lipase AK
Lipase PS
PPL
88/3
4/2
21/0
19/43
4/0
87/7
12/22
96/0
72/0
75/13
93/0
12/63
0.1/28
0
3/34
3/37
1/66
1/54
0
0.3/0
4/34
3/89
2/54
0
44/14
17/7
55/17
75/10
1/43
2/6
6/14
[a]
Reaction time 2 h.
Reaction time 24 h.
[b]
Adv. Synth. Catal. 2003, 345, 790 796
asc.wiley-vch.de
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
791

Katri Lundell et al.
FULL PAPERS
Table 3. Chemoselective O-acylation of 1 (0.05 M) with 2,2,2-trifluoroethyl butanoate (0.1 M; R ˆ Pr) in organic solvents by
CAL-A preparation (75 mg/mL) at the point where the maximal amount of 3 is observed at room temperature.
Relative amount [%]
Row
Solvent
Time [h]
Conv. [%]
1
3
5
7
1
2
3
4
5
6
7
8
9
THF
toluene
8
2
4
48
8
2
2
2
49
86
83
90
84
83
96
97
83
92
14
17
10
16
17
4
3
17
8
81
80
88
62
79
93
96
83
55
2
1
1
12
2
1
0
0
8
3
2
1
10
2
2
1
0
29
t-amyl alcohol
acetonitrile
dichloromethane
diisopropyl ether
TBME
TBME[a]
TBME[b]
[a]
2,2,2-Trifluoroethyl pentanoate as an acyl donor.
2,2,2-Trifluoroethyl acetate as an acyl donor.
[b]
distribution was similar although the efficiency of allows considerable intramolecular O!N acyl migra-
enzymatic O-acylation was less compared to the acyla- tion and leads to low maximal yield for 3 (row 4). The
tion of 1. As a disappointment, enzymatic N-acylations reactions proceed smoothly especially in toluene, t-amyl
of 1 and 2 by CAL-A did not occur although the enzyme alcohol, diisopropyl ether and TBME (rows 2, 3, 6 and
smoothly catalysed the highly S-enantioselective N- 7).
acylation of methyl pipecolinate.[23]
For further optimisation, the effect of an acyl donor
In the O-acylation of 1 with 2,2,2-trifluoroethyl was tested for the O-acylation of 1 in TBME as the best
butanoate (R ˆ Pr) in TBME enzymatic enantioselec- solvent and CAL-A as a catalyst (Table 3; rows 7 9).
tivities are generally negligible (Table 1). Lipase AK The reaction with 2,2,2-trifluoroethyl acetate (R ˆ Me)
with ee1 ˆ 43% at 81% conversion and with highly is slow and leads to low maximal yield (relative amount
enantiopure product 7 (ee7 ˆ 89%, row 4) shows the 55%) for 3 (row 9). The reaction with 2,2,2-trifluoroethyl
highest enantioselectivity. Double enzymatic acylation butanoate (R ˆ Pr) in TBME clearly gives the best
at the primary alcohol function is now evident, the result, allowing an almost complete transformation of 1
second acylation step from 5 to 7 being more enantio- to 3 in 2 hours (row 7). In the acylation of 2 under the
selective. Encouraged by this and because the formation same conditions, the formation of the maximal amount
of product 5 was slow, the present enzymes were (90%) of 4 takes 48 hours. The extent of acyl migration
screened for the acylation of acetate 5 (R ˆ Me) with with time now is relatively low as deduced on the basis of
2,2,2-trifluoroethyl butanoate (R ˆ Pr) in TBME (Ta- the low content of products 6 and 8. This deduction is
ˆ
ble 2). Lipase AK as the enantioselective (E 45) lipase justifiable also on conformational basis because the
smoothly allowed the formation of (S)-7 (RN-acyl ˆ Me ester group at the position 3 must be axial for the
and RO-acyl ˆ Pr; row 4). However, there is a drop in intramolecular O!N acyl migration in the formation of
ee5 ˆ 71% at 55% conversion (after 24 hours) to ee5 ˆ 6 through a six-membered ring.
59% at 59% conversion (after 48 hours) with the
Acylation at the 2-hydroxymethyl group of 1 proceeds
simultaneous drop in ee7 from 91% to 44%. As an to the maximum (97%) in 2 hours and O!N acyl
evident reason, the enzymatic reaction of the formed migration from 3 to 5 and the further acylation to 7 are
(S)-7 with the unreacted enantiomerically enriched significant at long reaction times. Due to the acyl
primary alcohol (R)-5 (rather than with 2,2,2-trifluoro- migration the efforts to purify butanoate 3 by column
ethanol as R1OH) to (S)-5 ruins the resolution.
chromatography failed, resulting in the product mixture
According to the results in Table 1, highly chemo- [1 (7%), 3 (68%) and 5 (25%)] rather than pure 3.
selective O-acylations of 1 and 2 with 2,2,2-trifluoroeth- However, the purification of 3 is not necessary in
yl butanoate (R ˆ Pr) in TBME are evident. From the preparative chemistry because the desired transforma-
screened enzymes CAL-A as the most effective was tion at the piperidine nitrogen can be performed
chosen for further studies. Accordingly, solvent effects immediately in the reaction mixture when 1 has reacted
on the CAL-A-catalysed O-acylation of 1 with 2,2,2- to 3. In order to show that this proposal is justifiable, the
trifluoroethyl butanoate (R ˆ Pr) were studied. The enzymatically produced ester 3 (R ˆ Pr) was quantita-
data at the point where the maximum amounts of the tively transformed to diacylated product 9 (corresponds
ester product 3 are observed are given in Table 3 (rows to 7 with RN-acyl ˆ Me and RO-acyl ˆ Pr) by the chemical N-
1 7). O-Chemoselectivity is excellent in every solvent acylation with acetic anhydride immediately after the
including acetonitrile where a slow enzymatic reaction removal of the enzyme and before product separation
792
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 790 796

Chemo- and Enantioselective Acyl Transfers by Lipases and Acylase I
FULL PAPERS
by column chromatography as shown in Experimental
Section.
In the presence of CAL-A, the amounts of 1 (^) and 7
*
( ) increase in a practically superimposable manner
when the initial amount of 7% for 1 is taken into account
in the data points with time for curve ^ (Figure 1b). This
leads to an assumption that 1 is formed as the result of
the enzymatic hydrolysis of 3 by the water present in the
Acyl Migration
It is well documented that acids and bases as well as high added enzyme preparation. An equal amount of an acyl
temperature enhance acyl migration.[24] In order to enzyme
intermediate [PrCO2CH2Ser(CAL-A);
study the effect of CAL-A on acyl migration, the Scheme 2] is formed, acylating 5 in the formation of 7.
~
reaction of 1 with 2,2,2-trifluoroethyl butanoate (R ˆ With this assumption, the sum of curves (^) and ( ) in
Pr) in TBME in the presence of CAL-A was stopped Figure 1b gives curve (D) in Figure 1a, approximating
after 2 hours by filtering off the enzyme. After purifi- the disappearance of 3 in the presence of CAL-A for
cation on silica a 2,2,2-trifluoroethyl butanoate-free reasons other than enzymatic hydrolysis. With the same
&
mixture [1 (7%), 3 (68%) and 5 (25%)] was obtained, reasoning, curve ( ; Figure 1a) which equals the sum of
&
*
containing 3 (R ˆ Pr) as the only possible acyl donor. curves 5 ( ) and 7 ( ) in Figure 1b represents the
The mixture was divided into two parts and to one of amount of 5 when acyl migration has taken place in the
them a new portion of CAL-A was added. The relative presence of CAL-A. The closely complementary be-
&
proportions of compounds 1, 3, 5 and 7 in the absence haviour of curve 3 (D) to 5 ( ) proposes O!N acyl
(Figure 1a, filled signs) and in the presence of CAL-A migration as the main reason for the faster disappear-
(Figure 1b) were then analysed with time by the GC ance of 3 and formation of 5. The mechanism of this is
method.
unclear. As one possibility, some acidic or basic groups
of the enzyme molecule are involved. As another
possibility the acyl-enzyme intermediate acylates the
nitrogen in the liberating 1 (ready at the active site) and
leads to the formation of 5 (Scheme 2). The reaction
with the alcohol function of 1 only leads back to 3. The
previous results for the N-acylation of methyl pipeco-
linate by CAL-A support this conclusion.[23]
Enantioselective Enzymatic Alcoholysis
After unsuccessful asymmetric acylation, the present
enzymes were screened for the alcoholysis of 7 (R ˆ Me)
with methanol (0.2 M) in TBME. The results in Table 4
show considerable reactivity except in the case of
acylase I. Lipases generally leave the amide bond
unreactive and accordingly the formation of 5 (R ˆ
Me) as the only new reaction product is expected.
However, PPL is an exception, leading to signs of
enantiopure 1. Because of the slow reaction this was not
ˆ
studied more thoroughly. Lipase AK (E 45) and lipase
PS (E 10) showed highest enantioselectivity and the
ˆ
former was chosen for further consideration. Moreover,
enantioselectivities in the cases of acetylated (R ˆ Me;
row 4) and butanoylated (R ˆ Pr; row 5) 7 were the same
and the work was continued with the acetylated
substrate. In order to optimise the resolution conditions
various methanol concentrations as well as other
alcohols in the place of methanol were studied. The
results in Table 5 show decreasing reactivity (conversion
reached at a certain time) with increasing alcohol
concentration. Butanol as a nucleophile is more favour-
able than the lower analogues and enantioselectivity
increases with increasing butanol concentration (rows 6,
8 and 9) or by decreased temperature (row 7). Finally,
ˆ
Figure 1. Acyl migration (R ˆ Pr) in TBME at room temper-
ature, (a) without CAL-A and (b) in the presence of CAL A;
1 (^),3 ( ), 5 ( ) and 7 ( ) as observed and 3 (~), 5 ( ) as
&
&
~
*
calculated.
the highest enantioselectivity (E
100, row 9) was
Adv. Synth. Catal. 2003, 345, 790 796
asc.wiley-vch.de
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
793

Katri Lundell et al.
FULL PAPERS
Scheme 2. Intramolecular and enzymatic acyl transfers from 3 to 5.
Table 4. Enzyme (75 mg/mL) screening for the alcoholysis of
7 (R ˆ Me; 0.05 M) with methanol (0.2 M) in TBME at room
temperature; reaction time 24 h.
Table 5. Effect of alcohol concentration for the alcoholysis of
7 (R ˆ Me, 0.05 M) by lipase AK preparation (75 mg/ml) in
TBME and in neat butanol at room temperature.
Relative amount [%]/ee [%]
Alcohol
Concen-
tration [M]
Time
[h]
Conv.
[%]
E
Enzyme
7
5
1
E
MeOH
MeOH
MeOH
PrOH
0.2
0.4
0.6
0.2
0.6
0.2
0.2
0.6
neat
24
24
24
24
24
24
24
24
192
38
29
22
37
27
40
28
33
54
45
40
30
47
36
49
72
57
100
Acylase I96/2
64/16
3/3
62/52
65/50
4/33
36/29
97/1
38/93
35/95
57/67
14/18
1
2
1
45
45
10
CAL-A
CAL-B
Lipase AK
Lipase AK[a]
Lipase PS
PPL
PrOH
BuOH
BuOH[a]
BuOH
BuOH
43/84
84/2
2/100
[a]
R ˆ Pr.
[a]
Temperature 88C.
observed when the reaction was performed in neat
butanol. When the lipases were screened for the
alcoholysis of 8 (R ˆ Pr) the highest enantioselectivity
ˆ
of only E 7 was observed for the lipase AK-catalysed
reaction in neat butanol. Accordingly, for the resolution
of 3-hydroxymethylpiperidine derivatives the present
method is poor compared to the previously published
[10,11]
ˆ
acylation and hydrolysis methods (E 23 70).
To this end, the usability of alcoholysis for the
resolution of 7 (R ˆ Me) was confirmed by performing
a gram-scale resolution in neat butanol in the presence
of lipase AK as shown in the Experimental Section
Scheme 3. Lipase AK-catalysed enantioselective alcoholyses
of 7 and 8.
(Scheme 3). The unreacted (R)-7 and the alcohol enantiomer reacts faster by various enzymes. This
product (S)-5 were separated from the resolution reveals the S enantiomer to be more reactive in all
mixture after the reaction was stopped at 50% con- lipase-catalysed acylations and alcoholyses of 2-substi-
version. The absolute configurations of this work are tuted substrates 1, 5 and 7, acylase I preferring the
based on the comparison of the measured optical reactions of the (R)-enantiomer. Interestingly, lipases
rotations of 5 and 7 to those found in literature.[8] After AK and PS were previously proposed to proceed S
knowing the absolute configurations in the case of lipase enantioselectively for the acylation of N-Boc protected
AK catalysis, the peak positions of the enantiomers in 2.[11] Expecting the same enantiopreference for the
the chromatograms can be used to consider which present reactions of the 3-regioisomer, an opposite
794
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 790 796

Chemo- and Enantioselective Acyl Transfers by Lipases and Acylase I
FULL PAPERS
porcine pancreatic lipase (PPL) were obtained from Sigma.
Hydroxymethylpiperidines 1 and 2 were products from
Aldrich. Diacylated products were prepared by the reaction
of 1 and 2 with acetic (R ˆ Me) or butanoic (R ˆ Pr)
anhydrides in the presence of 4-dimethylaminopyridine
(DMAP) and triethylamine. The products were purified by
eluting with petroleum ether/acetone (gradient 3/1 to 1/3) on
silica. N-Acylated products 5 and 6 were obtained in the same
way but using stoichiometric control. 2,2,2-Trifluoroethyl
carboxylates (RCO2CH2CF3; R ˆ Me, Pr and Bu) were pre-
pared from the corresponding acid chloride and 2,2,2-trifluoro-
ethanol. Butanol, methanol and the solvents were of the
highest analytic grade from Aldrich, Lab Scan Ltd, Riedel-de
stereochemical structure then is more reactive than in
the case of the 2-isomers (Scheme 3).
Conclusion
The present work describes the usefulness of lipase AK
for the resolution of racemic diacylated 2-hydroxymeth-
ˆ
yl piperidine 7 in neat butanol with E 100. The benefit
of this method is the simple one-step reaction resulting
in the two enantiomers, N-acylated amino alcohol (S)-5
with 92% and less reactive (R)-7 with 93% ee, at 50%
conversion. The alcoholysis of 7 is favourable over the
acylation of N-acylated amino alcohol 5 because the
reverse enzymatic acylation of 7 by unactivated butyl
acetate (another product of enzymatic alcoholysis) is
not likely under the reaction conditions whereas the
acylation method suffers from the enzymatic alcoholysis
of the produced (S)-ester 7 with the enantiomerically
enriched (R)-5 or with R1OH (another product of
enzymatic acylation), leading to enantiopurity drops
with time. The regioisomer 8 in butanol reacts in a less
»
Haen and J. T. Baker.
The progress of reactions and the ee values of the unreacted
substrates and the formed products were followed by taking
samples (0.1 mL) at intervals and analysing them by gas
chromatography on a Chrompack CP-Chirasil-DEX CB
column (25 m). For good baseline separation the unreacted
amino/hydroxy group of 1 6 in the sample was acetylated with
acetic anhydride for enzymatic butanoylation and pentanyo-
lation and with butanoic anhydride for enzymatic acetylation
in the presence of pyridine containing DMAP (1%). The
ˆ
determination of E was based on equation E ln[(1 eeS)/(1 ‡
ˆ
eeS/eeP)]/ln[(1 ‡ eeS)/(1 ‡ eeS/eeP)] with c eeS/(eeS ‡ eeP) as
ˆ
enantioselective manner (E 7).
based onthe original equations ofChenet al.[26] Theequationis
valid for irreversible acyl transfers when no side reactions of
the substrates or products is observed. 1H NMR and 13C NMR
spectra were recorded in CDCl3 on a Jeol Lambda 400 or a
Bruker 200 spectrometer with tetramethylsilane as an internal
standard. Mass spectra (MS) were taken on Agilent GC-MS
instrument equipped with 5973 network mass selective detec-
tor and 6890 N network GC-system. Optical rotations were
measured using a Jasco DIP-360 polarimeter.
The present work describes highly chemoselective O-
acylation of 2- and 3-hydroxymethylpiperidines 1 and 2
in TBME in the presence of lipases. CAL-A as a catalyst
and 2,2,2-trifluoroethyl butanoate (R ˆ Pr) as an acyl
donor proved to be especially useful in this respect.
Enzymatic O-acylation competes with intramolecular
O!N acyl migration. Accordingly, relatively high
enzyme contents (75 mg/mL of the enzyme preparation
corresponding to 15 mg/mL of commercial enzyme)
were used to favour enzymatic acylation. It was shown
that O!N acyl migration is somewhat faster in the
presence of CAL-A than in the absence of the enzyme.
By this simple and mild method, 90 98% of 1 (and 2)
can be transformed to amino ester 3 (and 4), allowing
the protection of the primary hydroxy group of the
piperidine derivative. Another reaction can then be
Enzymatic Reactions
The reactions were typically performed as small-scale experi-
ments at room temperature (23 248C) where 2,2,2-trifluor-
oethyl butanoate (0.1 M) was added in the solution (0.05 M) of
1, 2 or 5 in an organic solvent. The alcoholysis of 7 and 8
(0.05 M) was performed with an alcohol (0.2 0.6 M) in TBME
directed to the free secondary amino group. As an or in a neat alcohol. The commercial enzyme or enzyme
preparation (75 mg/mL) was added in order to start the
reaction.
example, the chemical N-acylation of 3 by acetic
anhydride was performed after the removal of the
enzyme.
Chemoselective Gram-Scale Preparation of (rac)-3
Experimental Section
CAL-A preparation (75 mg/mL) was added to 1 (0.60 g,
5.2 mmol) and 2,2,2-trifluoroethyl butanoate (1.57 mL,
10.4 mmol) in TBME (104 mL). After 2 h the enzyme was
filtered off at 93% conversion. The isolated product mixture
contained 1, 3, 5 and 7 (R ˆ Pr) with relative proportions of 6.1,
93, 0.2 and 0.7%, respectively. In order to prevent O!N acyl
migration the chemical N-acylation of the produced 3 (R ˆ Pr)
was performed immediately by acetic anhydride (10.4 mmol)
in the presence of triethylamine (5.2 mmol) and DMAP. The
product ratio remained constant. Purification by column
chromatography (ethyl acetate/hexane, 1/1) yielded racemic
N-acetyl-2-butanoyloxymethylpiperidine (9, 0.98 g, 4.3 mmol)
General Remarks
Lipases AK (Pseudomonas fluorecens) and PS (Pseudomonas
cepacia) were purchased from Amano Pharmaceuticals and
CAL-A (Candida antarctica lipase A, Chirazyme L5) and
À
CAL B (Candida antarctica lipase B, Chirazyme L2) from
Boehringer-Mannheim. CAL-A and lipases AK and PS were
immobilised on Celite in the presence of sucrose as described
before,[25] the final preparation containing 20% (w/w) of the
lipase. Acylase I from Aspergillus melleus (0.49 U/mg) and
Adv. Synth. Catal. 2003, 345, 790 796
asc.wiley-vch.de
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
795

Katri Lundell et al.
FULL PAPERS
containing 0.6% of 7 (R ˆ Me) and 0.1% of N-butanoylaceta-
mide formed from 5. In GC analysis the retention times for the
enantiomers of 9 were 37 and 38 while the corresponding N-
butanoylacetamide gave the retention times 30 and 31.
[2] H.-J. Altenbach, G. Blanda, Tetrahedron Asymmetry
1998, 9, 1519 1524.
[3] X. Liang, A. Lohse, M. Bols, J. Org. Chem. 2000, 65,
7432 7437.
(Rac)-9 (corresponds to 7 with RN-acyl ˆ Me and RO-acyl ˆ Pr):
1H NMR (CDCl3, 258C): d ˆ 0.9 (t, 3H, CH3CH2),1.5 1.7 (m,
6H, CH2), 1.5 (m, 2H, CH3CH2CH2), 2.3 (m, 2H,
CH3CH2CH2CO), 2.1 (s, 3H, NCOCH3), 2.6 (t, 1H,
NCHCH2O), 3,2 (t, 2H, NCH2CH2), 4.0 4.6 (m, 2H,
NCHCH2O); 13C NMR (CDCl3, 258C): d ˆ 13.5 (CH3CH2),
18.2 (CH3CH2CH2), 19.3 (CH3CH2CH2), 21.8 (CH3CO), 24.9
and 25.1 (CH2CH2CH2), 26.2 (CH2CH), 42.4 (CH2N), 46.3
(NCHCH2O), 61.6 (CH2OCOPr), 170.0 and 173.2 (CO); MS:
[4] P. D. Bailey, P. A. Millwood, P. D. Smith, Chem. Com-
mun. 1998, 633 640.
¬
¬
[5] F. Sanchez-Sancho, B. Herradon, Tetrahedron Asymme-
try 1998, 9, 1951 1965.
¬
[6] B. Herradon, S. Valverde, Synlett 1995, 599 602.
¬
[7] M. Ors, A. Morcuende, M. I. Jimenez-Vacas, S. Valverde,
¬
B. Herradon, Synlett 1996, 449 451.
[8] E. J. Toone, J. B. Jones, Can. J. Chem. 1987, 65, 2722
2726.
[9] G. Asensio, C. Andreu, J. A. Marco, Tetrahedron Lett.
1991, 32, 4197 4198.
‡
M (calcd. for C12H21NO3): 227 (227.30).
[10] B. Wirz, W. Walther, Tetrahedron Asymmetry 1992, 3,
1049 1054.
Gram-Scale Resolution of 7(R ˆ Me)
Lipase AK preparation (75 mg/mL) was added to 7 (R ˆ Me;
1.00 g, 5.0 mmol) in butanol (100 mL) to start the reaction.
After 382 h the enzyme was filtered off at 50% conversion.
Purification by column chromatography (acetone/ petroleum
ether 1/1) yielded (S)-5 (0.30 g, 1.9 mmol, ee5 ˆ 92%, [a]2D0:
À 54.9 (c 1, CHCl3) compared to [a]20 (lit.)[8]: ‡ 20.5 (c 4.0,
CHCl3; R enantiomer at 45% ee) and D(R)-7 (0.48 g, 2.4 mmol,
ee7 ˆ 93%) containing 3% of amide according to the GLC
method. (R)-7 was further purified by enzymatic alcoholysis
resulting in the product with ee7 ˆ 97% and [a]2D0: ‡ 56.8 (c 1,
CHCl3) compared to [a]20 (lit.)[8]: À 18.3 (c 9.6, CHCl3 ) for the
S enantiomer at 47% eeD.
[11] A. Goswami, J. M. Howell, E. Y. Hua, K. D. Mirfakhrae,
M. C. Soumeillant, S. Swaminathan, X. Qian, F. A.
Quiroz, T. C. Vu, X. Wang, B. Zheng, D. R. Kronenthal,
R. N. Patel, Org. Process Res. Dev. 2001, 5, 415 420.
[12] B. Danieli, G. Lesma, D. Passarella, A. Silvani, J. Org.
Chem. 1998, 63, 3492 3496.
[13] N. Toyooka, Y. Yoshida, Y.Yotsui, T. Momose, J Org.
Chem. 1999, 64, 4914 4919.
[14] R. Chenevert, G. M. Ziarani, M.-P. Morin, M. Dasser,
Tetrahedron Asymmetry 1999, 10, 3117 3122.
[15] R. Chenevert, G. M. Ziarani, M.-P. Morin, J. Org. Chem.
1999, 64, 3178 3180.
[16] N. Chinsky, A. L. Margolin, A. M. Klibanov, J. Am.
Chem. Soc. 1989,111, 386 388.
[17] L. T. Kanerva, M. Kosonen, E. V‰nttinen, T. T. Huuhta-
nen, M. Dahlqvist, Acta Chem. Scand. 1992, 46, 1101
1105.
[18] V. Gotor, R. Brieva, F. Rebolledo, J. Chem. Soc. Chem.
Commun. 1988, 957 958.
[19] A. Liljeblad, E. V‰nttinen, L. T. Kanerva, Chirality 1999,
11, 432 439.
1
(S)-5 (R ˆ Me): H NMR (CDCl3, 258C): d ˆ 1.3 1.9 (m,
6H, CH2), 2.1 (s, 3H, CH3CO), 2.7 (t, 1H, NCHCH2O), 3.2 (t,
2H,NCH2CH2), 3.4 4.2 (m, 2H, CH2OH), 4.8 (OH);
13
C NMR: d ˆ 21.8 (CH3CO), 25.0 and 25.5 (CH2CH2CH2),
26.1 (CH2CH), 50.8 (CH2N), 55.6 (NCHCH2O), 61.6 (CH2OH)
and 170.0 (CO); MS: M (calcd. for C8H15NO2): 157 (157.21).
‡
1
(R)-7 (R ˆ Me): H NMR (CDCl3, 258C): d ˆ 1.3 1.7 (m,
6H, CH2), 2.0 (s, 3H, CH3CO), 2.1 (s, 3H, NCOCH3), 2.6 (t, 1H,
NCHCH2O), 3.1 (t, 2H, NCH2CH2), 4.0 4.7 (m, 2H,
13
NCHCH2O); C NMR (CDCl3, 258C): d ˆ 20.7 (NCOCH3),
21.7 (CH3CO2), 25.1 and 25.6 (CH2CH2CH2), 26.1 (CH2CH),
42.4 (CH2N), 46.4 (NCHCH2O), 61.8 (CH2OCOMe), 169.9
[20] P. Virsu, A. Liljeblad, A. Kanerva, L. T. Kanerva,
Tetrahedron Asymmetry 2001, 12, 2447 2455.
[21] E. V‰nttinen, L. T. Kanerva, Tetrahedron Asymmetry
1996, 7, 3037 3046.
‡
and 170.8 (CO); MS: M (calcd. for C10H17NO3): 199 (199.25).
[22] K. Lundell, L. T. Kanerva, Tetrahedron Asymmetry 1995,
6, 2281 2286.
Acknowledgements
[23] A. Liljeblad, J. Lindborg, A. Kanerva, J. Katajisto, L. T.
Kanerva, Tetrahedron Lett. 2002, 43, 2471 2474.
[24] G. Fodor, J. Kiss, J. Am. Chem. Soc. 1950, 72, 3495 3497.
[25] L. T. Kanerva, O. Sundholm, J. Chem. Soc. Perkin Trans.
1 1993, 2407 2410.
This work is supported by TEKES (Technology Development
Center in Finland).
References
[26] C.-S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294 7299.
[1] G. M. Strunz, J. A. Findlay, in The Alkaloids, (Ed.: A.
Bossi), Vol. 26, Academic Press, New York, 1985,
pp. 89 174.
796
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 790 796