Enantiospecific enzyme-catalysed resolution of novel N,N-disubstituted α-amino acid phenolic ester derivatives using pig liver esterase

Article abstract of DOI:10.1039/b008000o

R-Amino acid esters 1, 2 and 3 are novel compounds possessing hypnotic activity. On attempting an asymmetric synthesis of these molecules, racemisation was observed when reacting bis(2-methoxyethyl)amine with α-bromo intermediate 4. In vitro plasma stability studies showed that the R enantiomers had much greater resistance to esterase-mediated degradation than the corresponding S enantiomers. This observation led to the use of commercially available pig liver esterase to prepare 1, 2 and 3 on a multigram scale. The crystal structures of 1 and 2 are reported and confirm R configuration.

Full text of DOI:10.1039/b008000o

View Article Online / Journal Homepage / Table of Contents for this issue
Enantiospecific enzyme-catalysed resolution of novel
N,N-disubstituted ꢀ-amino acid phenolic ester derivatives using
pig liver esterase
1
D. Jonathan Bennett, Kirsteen I. Buchanan, Andrew Cooke, Ola Epemolu, Niall M. Hamilton,
Edward J. Hutchinson* and Ann Mitchell
Organon Research, Newhouse, Lanarkshire, Scotland, UK ML1 5SH
Received (in Cambridge, UK) 3rd October 2000, Accepted 9th January 2001
First published as an Advance Article on the web 24th January 2001
R-Amino acid esters 1, 2 and 3 are novel compounds possessing hypnotic activity. On attempting an asymmetric
synthesis of these molecules, racemisation was observed when reacting bis(2-methoxyethyl)amine with α-bromo
intermediate 4. In vitro plasma stability studies showed that the R enantiomers had much greater resistance to
esterase-mediated degradation than the corresponding S enantiomers. This observation led to the use of
commercially available pig liver esterase to prepare 1, 2 and 3 on a multigram scale. The crystal structures of 1
and 2 are reported and confirm R configuration.
Introduction
N,N-Disubstituted α-amino acid ester derivatives have been
shown to be useful intermediates for the synthesis of novel
β-lactams.1 In addition, there have been a number of reports
describing interesting biological activity of this class of
compounds. Puhl et al.2 recently described N,N-disubstituted
α-amino acid esters as endothelin inhibitors with potential
therapeutic uses in the treatment of cardiovascular and other
disorders. Previous reports outlined the local anaesthetic activ-
ity of aryl ester and amide derivatives3 as well as compound A
Scheme 1 Attempted synthesis of (R)-1 and (R)-2 from (S)-2-
bromobutyric acid. Reagents and conditions: (i) PyBroP, DCM,
iPr2EtN, DMAP, 2,6-dimethoxyphenol or 2,6-dimethoxy-4-methyl-
phenol, RT, 18 h; (ii) bis(2-methoxyethyl)amine, PhMe, 100 ЊC, 2 d.
Whilst the PyBroP (bromotripyrrolidinophosphonium
hexafluorophosphate)-mediated esterification6 worked well,
racemisation was observed during subsequent displacement of
the α-bromo substituent on compound 4 with bis(2-methoxy-
ethyl)amine. Relatively high temperatures were required for the
reaction to proceed and despite varying the reaction conditions,
we were unable to prevent loss of chiral integrity. This result
was disappointing since minimal racemisation was observed
when 1,2,3,4-tetrahydroisoquinoline was used as the nucleo-
phile. It was therefore decided to examine alternative routes to
our desired targets. Preparative chiral HPLC proved useful only
for 2. Inadequate separation of the component enantiomers of
1 and 3, as well as lengthy run times and the large volumes
of solvent waste, rendered this approach unviable to provide
bulk quantities of material.
which was shown to possess general anaesthetic activity.4 More
recently, phenolic esters of general structure B have also been
shown to possess this activity.5 Compared with compound A,
these new leads have improved stability and anaesthetic profiles
such as quick onset, with smooth, short sleeps and fast
recovery. As a result of this research, compound 2 was exam-
ined in clinical trials as a potential new intravenous anaesthetic.
Results and discussion
In order to perform extensive pharmacological testing of target
molecules 1–3, it was necessary to prepare gram-scale quan-
tities of enantiopure material. Generally, the ‘unnatural’ R
enantiomer of this series is the more active anaesthetic, so a
synthesis starting from commercially available (S)-2-bromo-
butyric acid was envisioned as the most direct route to the
desired targets 1 and 2 (Scheme 1).
During a series of stability studies of racemic 2 in blood
plasma at 37 ЊC, a number of interesting observations were
made. Different rates of degradation of 2 occurred in mouse,
rat, dog and human plasma. The degradation was thought to be
due to esterases present in the plasma, with rodent plasma
362
J. Chem. Soc., Perkin Trans. 1, 2001, 362–365
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b008000o

View Article Online
Fig. 1 Degradation of racemic 2 in different blood plasma types over 5 h.
Fig. 2 Degradation of (R)-2 and (S)-2 in different blood plasma types over 5 h.
esterases being the most aggressive (Fig. 1). This theory of
hydrolysis was monitored by chiral HPLC with work-up
esterase-catalysed degradation was given further credence by the
effect of potassium fluoride (KF), a known inhibitor of esterase
activity,7 which dramatically slowed the rate of breakdown.
When the separate enantiomers of 2 were subjected to the
same study, a large difference in degradation rate was observed,
particularly in mouse plasma. As can be seen in Fig. 2, the R
enantiomer was shown to be extremely stable in all plasma
types over a 5 h time-course with essentially no degradation
occurring. In contrast, the ‘natural’ S enantiomer underwent
extensive degradation with only 13% remaining in mouse
plasma after 5 h. The inclusion of potassium fluoride (at 80
mM) reversed the effect, but only for the first 2.5 h. Subsequent
degradation suggested an element of non-specific breakdown
or the presence of some KF-insensitive enzyme. However the
possible contribution of spontaneous breakdown was dis-
counted since on incubation of (S)-2 at pH 7.4 in Tris buffer at
37 ЊC for 5 h, negligible compound loss was observed.
Based on these results, we considered the possibility of using
a commercially available esterase to effect an enantioselective
ester hydrolysis. Pig liver esterase (ple) was chosen since it is
readily available and able to tolerate a wide range of structur-
ally diverse racemic, prochiral and meso substrates.8 However,
the vast majority of substrates reported to be hydrolysed in
high enantioselectivity were methyl or ethyl esters, so we were
interested to see if the bulky substituted phenolic esters would
be tolerated by the enzyme. There are few reports of enzyme-
mediated hydrolyses of phenolic esters in the literature;
Parmar’s work with lipase-catalysed polyphenolic acetates9 is
the most noteworthy, although the products mentioned are not
optically active.
occurring once all the S enantiomer had been consumed.
Yields of the recovered R ester were routinely between 25–
30% (max. 50%) after chromatography on alumina to remove
the substituted phenol released during the hydrolysis of the
S enantiomer.
Addition of 10–20% DMSO is known to improve enantio-
selectivity in ple-catalysed reactions10 and a slight improvement
with substrate 1 was observed as well as an increase in reaction
rate. However, when DMSO is replaced with acetone, the reac-
tion rate is slowed to such an extent that only a 21% ee is
observed after 90 h. Compound 2 was hydrolysed much more
slowly than 1 or 3 such that an ee of 85.6% was observed after
13 d reaction time. However, by increasing the amount of
enzyme used from ~2000 to ~4000 units per mmol substrate
and by performing the reaction with DMSO present, complete
enantioselective hydrolysis was observed within 90 h. Since
both 1 and 2 possessed good solubility in aqueous media,
differential solubility was not thought to be the cause of the
slower hydrolysis of 2. Hence this result indicates either an
unfavourable inductive effect arising from the p-methyl group
of 2, or that this group may be interacting unfavourably with
amino acid residues lining the hydrophobic pocket occupied by
the bulky phenolic ester in the enzyme active site. No enzyme
crystal structure is available to confirm this latter hypothesis,
although a widely accepted model for the prediction of enzyme-
enantiomer complementarity has been proposed by Jones et
al.11 Whilst this cubic-space model pays particular attention to
the dimensions of hydrophobic and polar pockets available for
side chain binding, the precise positioning of the ester group,
other than it being in close proximity to a key serine residue
responsible for the hydrolysis, has not been examined. There
may therefore be an additional hydrophobic pocket available
for accommodating ester groups bulkier than methyl or ethyl.
As shown in Table 1, racemates of 1–3 are hydrolysed enantio-
selectively with the desired R enantiomers being recovered in
>99.5% ee within 90 h at RT in pH 7 phosphate buffer. Each
J. Chem. Soc., Perkin Trans. 1, 2001, 362–365
363

View Article Online
Table 1 Ple-mediated enantioselective hydrolysis of racemic 1–3
Substrate/
mmol
Enzyme/units per
mmol substrate
Vol. phosphate
buffer/ml
Enantiomeric
ratio/h
Ee of product
(%)
Compound
1
1
1
2
2
2
3
7.66
3.83
3.83
5.23
1.41
1.41
7.40
1538
1538
1538
2216
4042
4042
2054
40
90.9:9.1 (42)
100:0 (66)
60.3:39.7 (66)
57.8:42.2 (24)
88.9:11.1 (24)
90.0:10.0 (24)
—
>99.5a
>99.5b
21.0a
35 ϩ 5 DMSO
30 ϩ 10 Me2CO
75
40
35 ϩ 5 DMSO
100
85.6c
99.4a
>99.5a
>99.5d
a After 90 h reaction time. b After 66 h. c After 312 h. d After 72 h.
Fig. 3 X-Ray crystal structures of (R)-1 and (R)-2.
Further refinements to Jones’ model would assist in the explan-
ation of these results. Also a more in-depth study of other
substituted phenolic ester substrates may help in establishing
these activated esters, in conjunction with ple, as general tools
for inducing chiral discrimination.
In order to verify that the R enantiomer was the configur-
ation of the recovered ester from the enzyme hydrolysis
reaction, and indeed the enantiomer which invokes the greater
general anaesthetic activity, crystals of 1 and 2 were grown and
submitted for X-ray analysis. The structures are shown below
(Fig. 3) and confirm R configuration.
The 3D structure of 1 is represented by an orthorhombic
crystal system with a P212121 space group. In contrast, the
p-methyl group of 2 forces the crystal into a monoclinic system
with a P21 space group. The asymmetric units of both com-
pounds consist of two independent complexes of a protonated
parent molecule and a chloride counterion, which is hydrogen
bonded to a water molecule.
available phenol to an appropriate 2-bromoalkylacyl chloride,
displacement of the halide with bis(2-methoxyethyl)amine and
subsequent hydrochloride salt formation.
The racemate of 2 was separated into its component enantio-
mers by chiral HPLC. The sample was run on a ChiracelTM OJ
Daicel column 25 × 2 cm using hexane–isopropyl alcohol–
diethylamine (95:5:0.1) as eluant, a flow rate of 10 cm3 minϪ1
at 30 ЊC with detection at 230 nm. For enzymatic studies and
analytical chiral separation of the racemates of 1, 2 and
3, samples were run on a ChiralpakTM AD Daicel column
25 × 0.46 cm using hexane–ethanol–diethylamine (97:3:0.1)
as eluant, a flow rate of 1 cm3 minϪ1 at 30 ЊC with detection at
230 nm.
For the plasma studies, Tris buffer (39.6 mg Tris HCl, Sigma-
Aldrich) was dissolved in 1 l ultrapure water (Milli-Q quality)
to give a concentration of 250 mM. The pH was adjusted to 7.4
with 1 M sodium hydroxide. A 1 mg cmϪ3 stock solution of 1, 2
or 3 in 50:50 acetonitrile–water was also prepared. For the
assay, the concentration of compound is diluted to 10 µg cmϪ3
in plasma.
Conclusion
742.5 µl of plasma was pre-warmed for 10 minutes in a
shaking water bath at 37 ЊC, then 7.5 µl of the pre-warmed
stock solution of compound was added giving a final concen-
Using commercially available pig liver esterase, we have per-
formed enantioselective hydrolyses on a series of substituted
phenolic esters. The active R enantiomers are recovered in 25–
30% yield (max. 50%) and in >99.5% ee. These reactions have
been scaled up and have been used to obtain multigram quan-
tities of target compounds 1, 2 and 3, required for further
testing as general anaesthetics. The Jones’ model for pre-
dicting enzyme–substrate complementarity does not take into
account any additional stabilisation provided by the binding of
large ester groups such as the substituted phenolic esters used
here. Further studies are recommended to investigate whether
these, or more activated phenolic esters, can be used to probe
suitability of substrates for chiral discrimination by esterases.
tration of 10 µg cmϪ3
.
The plasma was sampled at 0, 0.5, 1 and 2.5 h. The sample
was analysed by taking 100 µl plasma from the incubation tube
into 300 µl acetonitrile containing 10 µg cmϪ3 internal standard.
This was extracted then analysed by chiral HPLC as described
above.
For the enzymatic resolutions, pig liver esterase was pur-
chased as a crude, lyophilised powder from Sigma-Aldrich. The
reactions were performed in pH 7 phosphate buffer made up by
mixing 30.5 cm3 0.2 M disodium hydrogen orthophosphate
(Na2HPO4) and 19.5 cm3 0.2 M sodium dihydrogen ortho-
phosphate (NaH2PO4ؒ2H2O). 2 M Sodium hydroxide solution
was added during the course of the reaction to maintain the pH
at 7.0. Aliquots were taken from the reaction mixture at regular
intervals. To these were added hexane (1 cm3), the layers
agitated and then passed down a short plug of anhydrous
sodium sulfate. The samples were then analysed by chiral
HPLC as described above.
Experimental
General
The racemates of compounds 1, 2 and 3 were prepared by a
three step procedure5 comprising addition of a commercially
364
J. Chem. Soc., Perkin Trans. 1, 2001, 362–365

View Article Online
Melting points were taken with either a Gallenkamp capil-
lary melting point apparatus or a Reichert hot plate apparatus
and are uncorrected. Optical rotations were determined at
room temperature on an Optical Activity Ltd. AA-1000 polar-
imeter. 1H NMR (200 or 400 MHz) spectra were obtained using
a Bruker AM200 or a Bruker DRX400 instrument; chemical
shifts (δ) are relative to tetramethylsilane as internal standard.
Only discrete or characteristic signals are reported. IR spectra
were obtained with a Perkin-Elmer 16PC FT-IR spectrometer.
Accurate masses were measured on a PerSeptive Biosystems
MarinerTM Biospectrometry Workstation using TOF-MS
analysis.
C5D5N) 1.13 (t, 3H), 1.92–2.13 (m, 2H), 2.34 (s, 3H), 3.10–3.21
(m, 2H), 3.23–3.35 (m, 2H), 3.37 (s, 6H), 3.56–3.65 (m, 2H),
3.68–3.81 (m, 8H), 3.82–3.91 (m, 1H), 6.42 (s, 2H); νmax/cmϪ1
(KBr) 2944, 2278, 1768, 1607, 1470, 1248, 1196, 1119 cmϪ1; m/z
(ES) [M ϩ H]ϩ 370; Found: 370.2196. C19H32NO6 requires
370.2227.
(R)-2-[N,N-Bis(2-methoxyethyl)amino]pentanoic acid,
2Ј,6Ј-dimethoxyphenyl ester hydrochloride (3)
[α]D Ϫ5.88 (c 0.20 in MeOH); mp 85.5–86.5 ЊC; δH(CDCl3 ϩ
C5D5N) 1.00 (t, 3H), 1.53–1.67 (m, 2H), 1.88–1.99 (m, 2H),
3.05–3.14 (m, 2H), 3.17–3.28 (m, 2H), 3.37 (s, 6H), 3.55–3.72
(m, 4H), 3.80 (s, 6H), 3.90 (t, 1H), 6.60 (d, 2H), 7.12 (t, 1H),
7.21 (br s, 1H); νmax/cmϪ1 (KBr) 2934, 2113, 1750, 1607, 1484,
1309, 1265, 1174, 1114 cmϪ1; m/z (ES) [M ϩ H]ϩ 370; Found:
370.2196. C19H32NO6 requires 370.2227.
Crystal data
C18H29O6NؒHClؒH2O (1), M = 409.91, monoclinic, P21, a =
7.7225(1), b = 27.3326(1), c = 10.6961(1) Å, β = 103.669(1)Њ,
V = 2193.75(4) Å3, Z = 4, Dc = 1.2411(1) g cmϪ3, λ(Cu Kα) =
1.54184 Å, µ = 18.6 cmϪ1, F(000) = 880, T = 294 K, R = 0.054
for 3561 reflections with I > 2σ(I).
Acknowledgements
C19H31O6NؒHClؒH2O (2), M = 423.93, orthorhombic,
P212121, a = 12.0201(1), b = 14.7647(1), c = 25.6491(1) Å, V =
4552.03(5) Å3, Z = 8, Dc = 1.2372(1) g cmϪ3, λ(Cu Kα) = 1.54184
Å, µ = 18.1 cmϪ1, F(000) = 1824, T = 294 K, R = 0.090 for 5595
reflections with I > 2σ(I).
CCDC reference numbers 153401 and 153402.
See http://www.rsc.org/suppdata/p1/b0/b008000o/ for crys-
tallographic files in .cif format.
We are grateful to B. Montgomery and A. Osprey at Organon
for HPLC and accurate mass determinations respectively. We
would also like to thank A. Schouten, R. Boer and J. Kroon
from the Dept. of Crystal and Structural Chemistry, University
of Utrecht, The Netherlands for the X-ray analysis of
compounds 1 and 2.
References
(R)-2-[N,N-Bis(2-methoxyethyl)amino]butyric acid,
2Ј,6Ј-dimethoxyphenyl ester hydrochloride (1)
1 F. H. van der Steen, H. Kleijn, J. T. B. H. Jastrzebski and G. van
Koten, Tetrahedron Lett., 1989, 30(6), 765.
Pig liver esterase (0.62 g, 19 units mgϪ1) was added to a stirred
solution of racemic 1 as its hydrochloride salt (3.00 g, 7.66
mmol) in phosphate buffer (35 cm3) and DMSO (5 cm3). After
90 h, the system was stirred with MTBE (3 × 100 cm3) for 2 h
and the layers separated. Any gel that formed was dispersed by
filtration through a glass sinter. The combined organic layer
was dried (Na2SO4), filtered and solvent removed. The residue
was then purified by chromatography on alumina using petrol-
eum ether (40:60)–ethyl acetate (2:1) as eluant. This gave the
desired optically pure R enantiomer as its free base (693 mg,
25.5%). The hydrochloride salt was then prepared by passing
dry hydrogen chloride gas through a solution of 1 in diethyl
ether.
2 M. Puhl, J.-C. Zechel, K. Ditrich, H. Hillen, T. Kohl, M. Erhardt,
S. Hergenroeder and C. O. Markert, DE19745146 (Chem. Abstr.,
1999, 130, 267770).
3 G. Brancaccio and A. Larizza, Farmaco, Ed. Sci., 1964, 19, 986;
R. G. N. Lingwood, GB1102011 (Chem Abstr., 1968, 69, 18834);
L. Coscia, P. Causa and G. De Natale, Boll. Chim. Farm., 1968, 107,
310; J. R. Barrio, A. Novelli, V. Ferrari and O. Alonso, J. Med.
Chem., 1971, 14(1), 78.
4 D. G. Bamford, D. F. Biggs, G. E. Lee, A. J. Owen, D. W. Pulsford
and W. R. Wragg, ZA6802928 (Chem. Abstr., 1969, 71, 3391).
5 N. M. Hamilton, WO0005196 (Chem. Abstr., 2000, 132, 137727).
6 E. Frerot, J. Coste, A. Pantaloni, M. N. Dufour and P. Jouin,
Tetrahedron, 1991, 47(2), 259.
7 R. J. Simmonds, S. A. Wood and C. A. James, Analysis for Drugs
and Metabolites Including Anti-infective Agents, ed. E. Reid and
I. D. Wilson, Royal Society of Chemistry, Cambridge, 1990,
vol. 20, pp. 95–102; P. Bouis, G. Taccard and U. A. Boelsterli,
J. Chromatogr., 1990, 526, 447.
[α]D ϩ4.62 (c 0.52 in CHCl3); mp 53.5–54.5 ЊC; νmax
/
cmϪ1 (KBr) 3507, 2946, 2531, 1762, 1606, 1483, 1264, 1174,
1112; δH(CDCl3 ϩ C5D5N) 1.13 (t, 3H), 1.96–2.09 (m, 2H),
3.08–3.17 (m, 2H), 3.20–3.30 (m, 2H), 3.38 (s, 6H), 3.56–3.65
(m, 2H), 3.66–3.75 (m, 2H), 3.76–3.89 (m, 7H), 6.60 (d, 2H),
7.14 (t, 1H); m/z (ES) [M ϩ H]ϩ 356; Found: 356.2060.
C18H30NO6 requires 356.2068.
8 L. Zhu and C. M. Tedford, Tetrahedron, 1990, 46(19), 6587;
C. Tamm, Pure Appl. Chem., 1992, 64, 1187.
9 V. S. Parmar, A. Kumar, K. Bisht, S. Mukherjee, A. K. Prasad,
S. K. Sharma, J. Wengel and C. E. Olsen, Tetrahedron, 1997, 53(6),
2163; K. Bisht, O. D. Tyagi, A. K. Prasad, N. K. Sharma, S. Gupta
and V. S. Parmar, Bioorg. Med. Chem., 1994, 2(10), 1015.
10 F. Bjorkling, J. Boutelje, S. Gatenbeck, K. Hult and T. Norin,
Tetrahedron Lett., 1985, 26, 4957; M. Luyten, S. Muller, B. Herzog
and R. Keese, Helv. Chim. Acta, 1987, 70, 1250.
11 E. J. Toone, M. J. Werth and J. B. Jones, J. Am. Chem. Soc., 1990,
112, 4946; L. Provencher and J. B. Jones, J. Org. Chem., 1994, 59,
2729.
The following compounds were prepared similarly.
(R)-2-[N,N-Bis(2-methoxyethyl)amino]butyric acid,
2Ј,6Ј-dimethoxy-4Ј-methylphenyl ester hydrochloride (2)
[α]D ϩ7.03 (c 0.77 in CHCl3); mp 108–109 ЊC; δH(CDCl3 ϩ
J. Chem. Soc., Perkin Trans. 1, 2001, 362–365
365