Ketoreductases in the synthesis of valuable chiral intermediates: Application in the synthesis of α-hydroxy β-amino and β-hydroxy γ-amino acids

Article abstract of DOI:10.1016/j.tet.2003.10.109

A general method for the synthesis of β-alkyl α-hydroxy β-amino and α- and γ-alkyl substituted β-hydroxy-γ- amino acids is described. The synthesis of all three classes of amino acids proceeds through a common chiral alcohol intermediate that is generated

Full text of DOI:10.1016/j.tet.2003.10.109

Tetrahedron 60 (2004) 663–669
Ketoreductases in the synthesis of valuable chiral intermediates:
application in the synthesis of a-hydroxy b-amino and b-hydroxy
g-amino acids
*
Spiros Kambourakis and J. David Rozzell
BioCatalytics Inc., 129 North Hill Ave., Suite 103, Pasadena, CA 91106, USA
Received 8 August 2003; revised 5 September 2003; accepted 17 October 2003
Abstract—A general method for the synthesis of b-alkyl a-hydroxy b-amino and a- and g-alkyl substituted b-hydroxy-g-amino acids is
described. The synthesis of all three classes of amino acids proceeds through a common chiral alcohol intermediate that is generated from a
pro-chiral ketone diester via the action of a nicotinamide-dependent ketoreductase. Regioselective chemical or enzymatic hydrolysis
followed by rearrangement under Hofmann or Curtius conditions gives the final amino acid products. High yields of single diastereomers of
the final amino acids are obtained. Amino acids with both natural and unnatural alkyl substituents can be accessed using this methodology.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
In addition, few if any of these methodologies are generally
useful for the synthesis of each of the individual
diastereomers of a range of statine analogs, and most are
not useful for preparing compounds with alkyl substitutions
derived from non-naturally occurring amino acids.4–6
Gaining control over the stereochemistry of the chiral
carbons bearing both the amino and the alcohol groups at
reasonable cost and high enantiomeric purity is the key to
the successful production of these important chemical
intermediates.
Amino alcohols, and in particular a-hydroxy b-amino acids
(1, Fig. 1) and b-hydroxy g-amino acids (2, 3, Fig. 1), also
known as statines, are valuable chiral compounds that are
present in a variety of natural and synthetic compounds
possessing a spectrum of biological activities. These key
structural units are found in molecules showing several
types of pharmacological activity including protease
inhibitors, anti-neoplastic agents, antibacterials and anti-
cancer drugs.1 The significance of these molecules is also
evident given the number of research publications and
books dedicated to their synthesis.1–6
Herein we describe a chemo-enzymatic method for the
synthesis of individual diastereomers of a-hydroxy b-amino
and b-hydroxy g-amino acids and analogs that are shown in
Figure 1. Using non-chiral, readily available starting
materials, a range of vicinal amino alcohols can be
produced. The key step is a diastereoselective enzymatic
reduction that generates two chiral centers in a single step,
setting the absolute stereochemistry in the molecule. The
remaining chemical steps are straightforward and can be
Some methods for the synthesis of such compounds are
based on resolutions of ester derivatives using lipases and
give low yields.5 A number of elegant chemical synthetic
routes have been reported, however, most require multi-step
reaction sequences, chiral catalysts or starting materials, air-
sensitive reaction conditions, or unstable intermediates.1–3
Figure 1. Targeted amino alcohols and amino acids.
*
Corresponding author. Fax: þ1-626-3563999; e-mail address: skambourakis@biocatalytics.com
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.10.109

664
S. Kambourakis, J. D. Rozzell / Tetrahedron 60 (2004) 663–669
Figure 2. Synthesis of ketones and enzymatic reduction.
Figure 3. Synthesis of each mono ester can be achieved by either chemical or enzymatic techniques.
Figure 4. Both the cyclic carbamates or the free amines can be obtained from the rearrangement of the mono-acids.
carried out with readily available reagents. In the basic
synthetic method, a ketodiester 4 (Fig. 2) is reduced
diastereoselectively to form a key hydroxydiester inter-
mediate 5 in high yield by the action of a ketoreductase
enzyme. Both chiral centers present in the final statines and
their analogs are generated in this enzymatic step. Regio-
selective chemical or enzymatic hydrolysis of either ester
group of the key intermediate diester, followed by
rearrangement under Hofmann or Curtius conditions,
gives the final amino alcohol products (Figs. 3 and 4). In a
recent report, we described the synthesis of statine and
phenylstatine (R¼isobutyl, 10a and R¼benzyl, 10c, Fig. 4)
using this methodology.7 We now expand the method to
include the synthesis of a range of statine analogs including
a-hydroxy b-amino acids bearing natural and unnatural
alkyl substitutions (Fig. 1).
Fig. 2) is accomplished in a straightforward fashion via the
mono-alkylation of diethyl 1,3 acetonedicarboxylate with
the appropriate alkyl halide. The homo-statine 4b and
homo-phenylstatine 4d (Table 1) precursors were syn-
thesized by the alkylation of diethyl 1,3 acetonedicarboxyl-
ate with 1-iodo-3-methyl butane and 1-bromo-2-phenyl
ethane, respectively, using methods based on previously
reported methodology for the synthesis of 4a and 4c (Table
1).7 The analogous compound 2-benzyl diethyl oxaloacetate
(4e, Fig. 2) was synthesized by a Claisen condensation of
ethyl hydrocinnamate and diethyl oxaloacetate using an
adaptation of a reported method.8 As an illustration of the
generality of the method, we describe herein two
compounds with natural side chains (4a and 4c) and two
compounds bearing non-natural alkyl substitutions (4b, 4d,
Table 1).
To identify the appropriate ketoreductase for a desired
reaction, all ketodiester substrates were screened for reduc-
tion in the presence of each of 10 different commercially-
available ketoreductases (KRED-10,000 Ketoreductase
Screening Set, BioCatalytics Inc, Pasadena, CA, USA)
using the protocol provided by the manufacturer. At least
one enzyme was shown to be active with every ketone
2. Results and discussion
2.1. Synthesis of substituted ketone diesters 4 (Fig. 2) and
enzymatic reduction
Synthesis of 2-substituted diethyl ketoglutarates (4a–d,

S. Kambourakis, J. D. Rozzell / Tetrahedron 60 (2004) 663–669
665
Table 1. Reductions of ketone substrates using KRED enzymes
diethyl malate and benzyl bromide according to reported
methodology.9
Substrate
Yield (A/B/C/D)a
Cell-free enzymatic reaction conditions were developed to
carry out the enzymatic reductions. Nicotinamide co-factor
was required by all ketoreductases, but with efficient
recycling methods, the amount of added co-factor is
normally less than 1 mol% relative to starting ketone. All
the ketoreductases used in this study require NADPH as the
co-factor for hydride transfer. We therefore used a recycling
system based on reducing NADPþ to NADPH with glucose
and the commercially available enzyme glucose dehydro-
genase (GDH-101, BioCatalytics Inc, Pasadena, CA, USA).
Glucose is an inexpensive, water-soluble reductant, and the
immediate reaction product of the oxidation of glucose,
gluconolactone, spontaneously hydrolyses to gluconic acid
under the reaction conditions employed, making the overall
process essentially irreversible. The pH of the reaction
mixture was maintained at 6.7–6.8 using a pH-stat.
KRED101
KRED108
.95% (0/5/0/95)b
O
4a
4b
85% (0/13/0/87)
95% (0/99/0/0)b
35% (95/5)c
,5%
2.2. Regioselective hydrolysis and rearrangement
.90% (0/99/0/0)d
We have previously shown that the two ester groups of 5a
and 5c (Fig. 3) were hydrolyzed chemically with different
reaction rates.7 As a result, mild basic chemical hydrolysis
gave preferably the mono acids 6a and 6c (Fig. 3) while
complete hydrolysis to the mono acid followed by
esterification in ethanol gave mono-acids 8a and 8c
(Fig. 3).7 This methodology has been successfully applied
to all diethyl 2-alkyl 3-hydroxy glutarates 5a–d as well as
the diethyl 2-benzyl 3-hydroxy succinate 5e (Fig. 3).
a
A/B/C/D represents the ratio of diastereomers. For statine (4a) and
phenylstatine (4c): A: (3S,4S); B: (3R,4R); C: (3R,4S); D: (3S,4R).7
Reactions described in Ref. 7.
The ratio of diastereomers SSþRR/SRþRS is reported. No base-line
separation of all diastereomers was achieved by either chiral GC or
HPLC.
b
c
d
Major diastereomer: (2R,3S).
Both regio- and stereo-selectivity can also be obtained when
a lipase is utilized for the hydrolysis of 5 (Fig. 3). Twenty-
four lipases (ICR Screening Set, BioCatalytics, Inc.
Pasadena, CA, USA) were screened and enzymes were
identified that were active against every diester 5, giving
compounds 6 as the only or the preferred product. In some
cases, enzymatic hydrolysis of the second ester yielding the
diacid was observed when the reactions were allowed to
continue for very long periods or when large excesses of
enzyme were used. When a diastereomeric mixture of
alcohols was utilized as a starting material for enzymatic
hydrolysis, in addition to regioselectivity, stereoselectivity
can also be obtained. Such an effect was previously shown
in the hydrolysis of a diastereomeric mixture of 2-methyl
3-hydroxy glutarate and ICR 112.7 Similarly, lipases ICR
113 and ICR 114 selectively hydrolyzed the major
component [(2R,3S)] of a diastereomeric mixture [(2R,3S)/
(2S,3S), 9:1] of diethyl 2-benzyl-3-hydroxy succinate 5e,
giving mono-acid 6e (n¼0, Fig. 3) in 50–60% yield.
substrate; the results from the best enzyme for each
ketodiester are shown in Table 1.
As described earlier, the KRED-catalyzed diastereoselective
reduction generates two chiral carbons during this reaction,
and as a result, the absolute stereochemistry of the final
statines and their analogs is determined at this point. For
most reactions, the diastereomeric purity of the products is
high, .90%. In cases where the optical purity is lower than
desired, it can be further improved by stereoselective
enzymatic hydrolysis using a lipase enzyme (vide infra).
For statine and phenylstatine precursors (4a and 4c
respectively, Table 1) the absolute stereochemistry of the
enzymatic reduction was identified earlier after rearranging
to the final amine (10a and 10c) and comparing it with
authentic material that was synthesized chemically.7 For the
homo-statine 4b (Table 1) no standards and no stereo-
selective chemical syntheses to either 5b or 10b were
available. However, based on the chiral GC retention time
of the alcohol 5b (Table 1) and the known diastereoselec-
tivity of the reaction catalyzed by KRED101 on ketodiester
5a (Table 1), we assume that the major diastereomer of the
reduction of 5b and KRED101 is (3S, 4R). For the homo-
phenylstatine precursor 5d we were unable to obtain
baseline separation of all diastereomers using either chiral
GC or HPLC chromatography, and as a result, only the
diastereomeric ratio is reported (4d, Table 1). Finally, the
absolute stereochemistry for the reduction of 4e by KRED
108 was assigned using chiral HPLC by comparison with
authentic material that was synthesized from R- and S-
As shown earlier for statine (5a) and phenylstatine (5c)
precursors and now confirmed for the rest of compounds in
Table 1, rearrangement of the free alcoholic acids 8 (Fig. 4)
under either Curtius or Hofmann conditions gave the cyclic
carbamates 12.7 Formation of these products was shown to
be independent of both the rearrangement reagents and the
reaction conditions. The simplest and most straightforward
one step synthesis of carbamates 12 was shown to be the
heating of the free acid 8 with diphenylphosphoryl azide
(DPPA) in toluene.7 We now show that the opposite cyclic
carbamates 11 can be obtained in good yields (50–70%)

666
S. Kambourakis, J. D. Rozzell / Tetrahedron 60 (2004) 663–669
when instead of 8, the less hindered mono-acids 6a–d were
utilized for rearrangement with DPPA. We therefore
conclude that this is a general method for the preparation
of cyclic carbamates 11 and 12 from the corresponding
mono-acids 6 and 8 (Fig. 4).
Synthesis of diethyl 2-benzyl diethyl oxaloacetate 4e
(Table 1, Fig. 2) was achieved by a modification of standard
methodology.8 Ethyl hydrocinnamate (1 mL, 5.6 mmol)
was dissolved in THF (10 mL) and the solution was cooled
at 218 8C before butyl lithium (2.9 mL, 2.5 M in hexanes,
7.3 mmol) was slowly added. After stirring at this
temperature for 10 min, diethyl oxalate (0.8 mL,
8.4 mmol) was added through a syringe and the reaction
was slowly left to reach room temperature overnight. The
reaction mixture was then added to an ice-cold water
(50 mL, 0.15 M HCl) solution and extracted twice with
diethyl ether (30 mL£2). The combined organic layers were
back-extracted with brine, dried with Na2SO4 and evapor-
ated to dryness. After silica gel purification a clear oil of
2-benzyl ethyl hydroxy succinate (0.9 g) (4e, Table 1) was
isolated (yield, 58%). NMR verification of the products was
performed after enzymatic or chemical reduction (vide
infra).
Formation of the amino acids 9 and 10 was accomplished
after the alcohol was protected as an acetate prior to the
rearrangement. Under these reaction conditions, the corre-
sponding amides of 8a–d were prepared and rearranged
to the free amine 10a–d using bis[(trifluoroacetoxy)iodo]
benzene [(CF3CO2)2IPh]. This methodology was utilized
earlier for the synthesis of statine (10a) and phenylstatine
(10c).7 Besides the b-hydroxy g-amino acids 10a–d (n¼1,
Fig. 4), the 2-benzyl a-hydroxy b-amino acid 10e (n¼0,
R¼Bz, Fig. 4) was prepared using this methodology in
similar overall yield. Therefore, this reaction sequence of
acetyl protection of the alcohol followed by rearrangement
is a general method for the preparation of statine and
analogues 10 (Fig. 4).
4.2. Enzymatic reduction to alcohols 5 (Fig. 2)
The previous rearrangement conditions using bis[(trifluoro-
acetoxy)iodo]benzene were shown to be inadequate for the
preparation of the primary amines 9 (Fig. 4) from the
corresponding amides of 6. This rearrangement was
achieved using lead tetracetate [Pb(OAc)4] according to a
previously described procedure.10
The Ketoreductase Screening Set (KRED-10,000; Bio
Catalytics, Inc., Pasadena, CA, USA) containing 10 dif-
ferent ketoreductases was screened to determine the best
enzyme for the diastereoselective reduction of every 2-alkyl-
3-ketoglutarate diethyl ester. In addition to the 10
ketoreductases both NADPH co-factor and glucose
dehydrogenase (GLDH) are products available from
BioCatalytics. Individual reactions containing each keto-
reductase (2 mg/mL), NADPH (5 mM), NaCl (100 mM),
DMSO (2.5 or 5% v/v) each substrate (25 mM), glucose
(100 mM) and glucose dehydrogenase (GLDH, 2 mg/mL)
for co-factor recycling were prepared in a phosphate-
buffered (1 mL, 300 mM, pH 6.5) solution. The reactions
were incubated at 37 8C overnight before they were
extracted with ethyl acetate and analyzed by GC or HPLC
chromatography.
3. Conclusions
This report describes a general method for the synthesis of
statines and a number of analogs. An advantage of the
methodology reported here is that in most cases each of the
four diastereomers can be produced in high stereochemical
purity. All compounds are produced from non-chiral,
readily available starting materials, and the reactions are
performed at room temperature and pressure.
Larger-scale enzymatic reductions (0.5–2 g of ketone) were
prepared according to previously described conditions.7 The
optical purity of each product was determined by chiral GC
chromatography using a ChiralDex column (Chiral Tech-
nologies, 130 8C, 2 min and then to 180 8C, 0.5 or 1 8C/min)
after each enzymatically-produced alcohol was derivatized
asthetriflateester.Theenantiomericpuritywasalsoexamined
using chiral HPLC analysis (column: CHIRALCEL OD-RH
from Chiral technologies, H2O/CH3CN, v/v 60:40, 0.6 mL/
min). All products were oily compounds and were analyzed
by 1H NMR.
From the cloned commercially-available ketoreductases
that we have screened to date against compounds 4a–e, two
commercially available ketoreductases (KRED-101,
KRED-108, BioCatalytics, Inc. Pasadena, CA, USA) gave
products with good to excellent diastereoselectivity
(Table 1). We plan to further broaden the applications for
this general method by the development of new keto-
reductase enzymes. These will be reported in due course.
4. Experimental
4.2.1. Compound 5b. From KRED-101 reduction 1H NMR
(400 MHz, CDCl3): d¼0.87þ0.88 (dþd, J¼3.20 Hz, 6H,
CH2CH2CH(CH3)2), 1.17 (m, 2H, CH2CH2CH(CH3)2),
1.27þ1.29 (tþt, J¼3.60 HzþJ¼3.99 Hz, 6H, CO2CH2-
CH3), 1.55 (m, 2H, CH2CH2CH(CH3)2), 1.75 (m, 1H,
CH2CH2CH(CH3)2) 2.47þ2.51 and 2.55þ2.58 (dþd
J¼9.19 Hz and dþd J¼3.99 Hz first dþd overlaps with
multiplet, 3H, CH2CO2CH2CH3 and CHCO2CH2CH3), 4.2
(pent, J¼7.19 Hz, 5H, CO2CH2CH3, CHOH).
4.1. Synthesis of ketones 4 (Fig. 2)
The methodology that was published earlier for the
synthesis of 4a and 4c (Table 1) was utilized for the
synthesis of homo-statine 4b and homo-phenylstatine 4d
precursors.7 These products can be enzymatically reduced
to the alcohol without any further purification, or they can
be purified using silica gel chromatography (Hex/EtOAc,
v/v, 8:2) giving 88% yield (3.6 g) for 1-iodo-3-methyl-
butane and 65% for 1-bromo-2-phenylethane. NMR veri-
fication of the products was performed after enzymatic or
chemical reduction (vide infra).
1
4.2.2. Compound 5d. From NaBH4 reduction H NMR
(400 MHz, CDCl3): d¼1.25 (two overlapping triplets ,2:1
ratio, J¼7.19 Hz, 3H, CO2CH2CH3), 1.31 (two overlapping

S. Kambourakis, J. D. Rozzell / Tetrahedron 60 (2004) 663–669
667
triplets ,2:1 ratio, J¼7.19 Hz, 3H, CO2CH2CH3), 1.9þ2.05
(mþm, 2H, CH2CH2Ph), 2.55 (m, 3H, CH2CO2CH2CH3
and CHCO2CH2CH3), 2.7 (m, 2H, CH2CH2Ph), 4.15þ4.2
(qþq J¼7.19 Hz, and a multiplet, 5H, CO2CH2CH3,
CHOH), 7.18þ7.25 (mþm, 5H, CH2CH2Ph).
4.2.3. Compound 5e. From KRED-108 reduction 1H NMR
(400 MHz, CDCl3): d¼1.20 (t, J¼7.19 Hz, 3H, CO2CH2CH3),
1.27 (t, J¼7.19 Hz, 3H, CO2CH2CH3), 3.00 (m, 1H,
CHCO2CH2CH3), 2.95þ3.23 (mþasymmetric triplet, 2H,
CH2Ph), 4.15þ4.22 (mþm, 5H, CO2CH2CH3, CHOH), 7.3
(m, 5H, CH2Ph).
(Fig. 4) were obtained. Following the same
rearrangement conditions described earlier,7 6a (reduced
by KRED-101, hydrolyzed by ICR116) was mixed in
toluene with diphenylphosphoryl azide (1.1 equiv. DPPA)
and triethylamine (1.1 equiv.) and the solution was heated
at 85 8C for 1 h before the temperature was lowered to
60 8C where it was stirred overnight. Product isolation
and purification gave an oily compound 11a in 55–60%
yield.
4.4.1. Compound 11a. 1H NMR (400 MHz, CDCl3):
d¼0.90þ0.92 (dþd, J¼4.0 Hz, 6H, CH2CH(CH3)2), 1.12
(m, 1H, CH2CH(CH3)2), 1.27 (t, J¼6.8 Hz, 3H, CHCO2-
CH2CH3), 1.60 (m, 2H, CH2CH(CH3)2), 2.78 (m, 1H,
CHCO2CH2CH3), 3.32þ3.34 and 3.61 (dþd and t, J¼7.20,
8.79 Hz, 2H, CH2NH–CO–OCH), 4.2 (m, 2H, CO2CH2-
CH3), 4.75 (q, J¼7.80 Hz, 1H, CH2NH–CO–OCH). 13C
NMR (CDCl3, 100 MHz) d¼14.16 (s, CHCO2CH2CH3),
21.55 and 23.36 (s and s, CH2CH(CH3)2), 26.06 (s, CH2CH
(CH3)2), 36.06 (s, CH2CH(CH3)2), 43.78 (s, CHCO2CH2CH3),
48.02 (s, CHCO2CH2CH3), 61.06 (s, CH2NH–CO–OCH),
77.18 (s, CH2NH–CO–OCH), 159.74 (s, CH2NH–CO–
OCH), 172.42 (s, CHCO2CH2CH3).
The spectra of 5a and 5c have been reported earlier.7
4.3. Hydrolysis to the mono acid 6 (Fig. 3)
Hydrolysis to the mono acid 6 (Fig. 3) using mild chemical
or enzymatic conditions has been described earlier for the
compounds 5a and 5c.7 The same conditions were applied to
the rest of the compounds of Table 1 including 2-benzyl-3-
hydroxy succinate 5e which was hydrolyzed to the mono-
acid 6e (R:benzyl, n¼0, Fig. 3) by either chemical or
enzymatic methods. Under chemical conditions, 2-benzyl-
3-hydroxy succinate 5e (0.15 g, 0.53 mmol) was added to an
aqueous/ethanol (2 mL, 8:2, v/v) solution containing NaOH
(0.042 g, 1.1 mmol) and the mixture was stirred for 45 min
at room temperature until complete reaction to the mono
acid was detected by HPLC analysis. Acidification with HCl
(pH ,2) and extraction with EtOAc (10 mL, £2) gave after
solvent evaporation the mono-acid 6e (0.11 g, 0.44 mmol,
83% isolated yield). The same mono-acid was obtained by
Treatment of 6c with ICR125 and rearrangement with
DPPA gave 11c in 60% yield.
4.4.2. Compound 11c. 1H NMR (400 MHz, CDCl3):
d¼1.24 (t, J¼7.19 Hz, 3H, CO2CH2CH3), 2.9þ3.00 (mþm,
3H, CHCO2CH2CH3, CH2Ph), 3.45þ3.62 (tþt, J¼7.59þ
8.79 Hz, 2H, CH–O–CO–NH–CH2), 4.10 (q, J¼7.19 Hz,
2H, CO2CH2CH3), 4.81 (m, 1H, CH–O–CO–NH–CH2),
6.02 (s, 1H, CH–O–CO–NH–CH2), 7.25 (m, 5H, CH2Ph).
13C NMR (100 MHz, CDCl3): d¼13.99 (s, CO2CH2CH3),
33.62 (s, CHCO2CH2CH3), 43.68 (s, CH2Ph), 51.56 (s,
CO2CH2CH3), 61.13 (s, CH–O–CO–NH–CH2), 75.95 (s,
CH–O–CO–NH–CH2), 128.93þ128.67þ128.96þ137.39
(s, CH2Ph), 159.33 (s, CH–O–CO–NH–CH2), 171.36 (s,
CO2CH2CH3).
1
enzymatic hydrolysis using ICR125, since H NMR, GC
and HPLC analysis of the chemical and enzymatic
hydrolysis products were identical. Under the enzymatic
reaction conditions, 2-benzyl-3-hydroxy succinate 5e
(0.15 g, 0.53 mmol) was dissolved in 0.3 mL of DMSO
and was mixed with a potassium phosphate solution
(200 mM, pH 7, 10 mL) also containing lyophilized
ICR125 lipase (10 mg). After stirring at 37 8C for 2–3 h,
the mixture was acidified with HCl (2 M) to pH ,2.5 and
extracted with EtOAc (5 mL, £2). Solvent drying with
Na2SO4 and evaporation to dryness gave pure mono-acid 6e
as an oil in 90% isolated yield.
Protection of the free alcohol prior to the rearrangement was
necessary to synthesize the free amines 9. In an illustration
of the methodology 9a and 9c were prepared. Protection of
the alcohol as an acetate ester was accomplished by reacting
it with acetic anhydride (Ac2O, 1.1 equiv.) in the presence
of catalytic amounts (1–2% mol/mol) of trimethylsilyl
trifluoromethanesulfonate (TMSOTf) as described pre-
viously.7 Under these conditions, the alcohols of 6a and
6c was acetylated in 85–90% isolated yield. In a typical
rearrangement reaction, O-acetylated carboxylic acid of 6c
(0.3 g, 0.97 mmol) was dissolved in CH2Cl2 (4 mL,
containing one drop of DMF), oxalyl chloride (0.1 mL,
1.2 mmol) was added and the reaction was stirred at room
temperature for 30 min. The solvent was then evaporated;
the oily residue was redissolved in THF and cooled in an ice
bath for 5 min before ammonia gas was bubbled for 5 min
through the solution under vigorous stirring. The reaction
was left stirring at 4 8C for 3 more hours in a tightly closed
vessel, before it was diluted with 10 mL EtOAc and
extracted once with aqueous acid (1 N HCl) until neutral
solution was obtained. After washing with brine, the organic
layer dried with Na2SO4 and evaporated to dryness. The oily
amide residue redissolved in 10 mL tert-butanol containing
4.3.1. Compound 6e. 1H NMR (400 MHz, CDCl3): d¼1.20
(t, J¼7.19 Hz, 3H, CO2CH2CH3), 3.00þ3.05 and 3.21
(dþd, J¼9.99 Hz and m, 3H, CH2Ph and CHCO2CH2CH3),
4.15 (m, 3H, CO2CH2CH3, CHOH), 7.25 (m, 5H, CH2Ph).
The hydrolysis of 5b under mild chemical or enzymatic
conditions follows the same regio-selectivity, giving the less
1
hindered acid 6b. This selectivity was confirmed by H
NMR analysis after making 8b by the standard hydrolysis/
esterification sequence and rearranging it to 12b (vide
infra). The hydrolysis selectivity and product characteriz-
ation of 5a and 5c has been reported earlier.7
4.4. Rearrangement of mono acids 6 to the amines 9 and
11 (Fig. 4)
When rearrangement reactions were performed without
prior protection of the alcohol, cyclic carbamates 11

668
S. Kambourakis, J. D. Rozzell / Tetrahedron 60 (2004) 663–669
one drop of SnCl4, lead tetraacetate [Pb(OAc)4, 0.53 g,
1.2 mmol] was added and the reaction was stirred at 80 8C
for 4 h. At the end of the reaction tert-butanol was
evaporated, product was redissolved in Et2O (15 mL) and
the solution was filtered through celite. After evaporation of
Et2O and silica gel purification the O-acetylated N-Boc
protected compound 9c was obtained (0.22 g, 60% yield
from the O-acetylated carboxylic acid) as a white solid. The
same reaction sequence gave 9a as a white solid in 55%
overall yield.
4.5.2. Compound 12e. 1H NMR (400 MHz, CDCl3):
d¼1.28 (t, J¼7.19 Hz, 3H, CO2CH2CH3), 2.89þ2.92 and
3.03þ3.06 (dþd and dþd, J¼8.0, 5.20 Hz, 2H, CH2Ph),
4.12 (m, 1H, CH–NHCO–O–CH), 4.25 (q, J¼7.19 Hz, 2H,
CO2CH2CH3), 4.68 (d, J¼4.80 Hz, 1H, CH–NHCO–O–
CH), 5.67 (s, 1H, CH–NHCO–O–CH), 7.2þ7.35 (mþm,
5H, CH2Ph). 13C NMR (100 MHz, CDCl3): d¼14.04
(CO2CH2CH3), 41.70 (CH2Ph), 57.02 (CO2CH2CH3),
62.27 (CH–NHCO–O–CH), 77.23 in between the three
CDCl3 peaks (CH–NHCO–O–CH), 127.52þ129.08þ
129.21þ135.21 (CH2Ph), 157.33 (CH–NHCO–O–CH),
168.52 (CO2CH2CH3).
4.4.3. Compound 9c. 1H NMR (400 MHz, CDCl3): d¼1.12
(t, J¼7.19 Hz, 3H, CO2CH2CH3), 1.46 (s, 9H, C(CH3)3),
2.09 (s, 3H, CHOCOCH3), 2.9 (m broad, 3H, CHCO2C2H5
and CH2Ph), 3.45þ3.5 (m broadþm broad, 2H, CH2-
NHBoc), 4.02 (q, J¼7.19 Hz, 2H, CO2CH2CH3), 4.75 (m,
broad, 1H, CHOCOCH3), 5.17 (m, broad, 1H, CH2NHBoc),
7.2 (m, 5H, CH2Ph). 13C NMR (100 MHz, CDCl3) d¼14.08
(s, CO2CH2CH3), 20.90 (s, CHOCOCH3), 28.35 (s, CO2C
(CH3)3), 34.41 (s, CHCO2C2H5), 41.73 (s, CO2C(CH3)3),
49.54 (s, CH2Ph), 60.68 (s, CO2CH2CH3), 73.44 (s,
CHOCOCH3), 79.75 (s, CH2NHBoc), 126.60þ128.46þ
128.89þ138.26 (all s, CH2Ph), 155.89 (s, CO2C(CH3)3),
170.11 (s, CO2CH2CH3), 171.90 (s, CHOCOCH3).
Protection of the free alcohol of 8e as an acetate
using TMSOTf catalyst and rearrangement to the free
amine using [bis(trifluoroacetoxy)iodo]benzene [(CF3-
CO2)2PhI] was performed as described earlier.7 Under
these conditions 8e (0.3 g, 1.2 mmol) was protected and
rearranged giving 10e (0.15 g, 47% yield from 8e) as a
white solid.
4.5.3. Compound 10e. 1H NMR (400 MHz, D2O): d¼1.27
(t, J¼7.19 Hz, 3H, CO2CH2CH3), 2.32 (s, 3H, CHOCOCH3),
3.14 (m, 2H, CH2Ph), 4.25 (q, overlaps with m, 3H,
CO2CH2CH3 and CHNH2), 5.17 (d, J¼3.20 Hz, 1H,
CHOCOCH3), 7.4 (m, 5H, CH2Ph). 13C NMR (100 MHz,
D2O): d¼15.95 (s, CO2CH2CH3), 22.60 (s, CHOCOCH3),
37.96 (s, CH2Ph), 55.23 (s, CO2CH2CH3), 66.98 (s,
CHOCOCH3), 73.10 (s, H2NCH), 130.83þ132.13þ
132.23þ136.97 (s, CH2Ph), 171.25 (s, CHOCOCH3),
174.95 (s, CO2CH2CH3).
4.4.4. Compound 9a. 1H NMR (400 MHz, CDCl3):
d¼0.88þ0.90 (dþd, J¼4.00 Hz, 6H, CH2CH(CH3)3), 1.25
(t, J¼7.19 Hz, 3H, CO2CH2CH3), 1.45 (s, 9H, C(CH3)3),
1.52þ1.65 (m broadþm broad, 2H, CH2CH(CH3)3), 2.05 (s,
3H, CHOCOCH3), 2.75 (m, 1H, CHCO2C2H5), 3.35þ3.45
(m broadþm broad, 2H, CH2NHBoc), 4.1 (q, J¼7.19 Hz,
2H, CO2CH2CH3), 4.7 (s, broad, 1H, CHOCOCH3), 5.1 (m,
broad, 1H CH2NHBoc).
4.5. Hydrolysis to the mono acids 8 and rearrangement
to 10 and 12 (Fig. 4)
References and notes
1. Some representative publications: (a) Leanna, M. R.; DeMattei,
J. A.; Li, W.; Nichols, P. J.; Rasmussen, M.; Morton, E. Org.
Lett. 2000, 2, 3627. (b) Nicolau, K. C.; Dai, W. M.; Guy, R. K.
Angew. Chem., Int. Ed. Engl. 1994, 33, 15. (c) Liang, B.;
Richard, D. J.; Portonovo, P. S.; Joullie, M. M. J. Am. Chem.
Soc. 2001, 123, 4469. (d) Rich, D. H.; Sun, E. T. O. J. Med.
Chem. 1980, 23, 27. (e) Li, W. R.; Ewing, W. R.; Harris, B. D.;
Joullie, M. M. J. Am. Chem. Soc. 1990, 112, 7659. (f)
Marciniszyn, Jr. J.; Hartsuck, J. A.; Tang, J. J. Biol. Chem.
1976, 251, 7088. (g) Jouin, P.; Poncet, J.; Dufour, M. N.;
Pantaloni, A.; Castro, B. J. Org. Chem. 1989, 54, 617. (h) Sun,
C. I.; Chen, C. H.; Kashiwada, Y.; Wu, J. H.; Wang, H. K.;
Lee, K. H. J. Med. Chem. 2002, 45, 4275. (i) Hom, R. K.;
Fang, L. Y.; Mamo, S.; Tung, J. S.; Guinn, A. C.; Walker,
D. E.; Davis, D. L.; Gailunas, A. F.; Thorsett, E. D.; Sinha, S.;
Knops, J. E.; Jewett, N. E.; Anderson, J. P.; Varghese, J.
J. Med. Chem. 2003, 46, 1799.
Synthesis of statine 10a, phenylstatine 10c as well as the
cyclic carbamates 12a and 12c from the corresponding
mono-acids 8 has been described earlier.7 For the formation
of 12b and 12e the same reaction sequence was applied.
Using the same reaction sequence 12e was synthesized from
5e (n¼0, Fig. 4) in 67% isolated yield. The only difference
is that esterification of the di-acid 7e was performed at room
temperature, since a small amount (5–10%) of diester is
forming under the esterification conditions.
4.5.1. Compound 12b. 1H NMR (400 MHz, CDCl3):
d¼0.89þ0.91 (dþd, J¼1.6 Hz, 6H, CH2CH2CH(CH3)2),
1.22 (m, 2H, CH2CH2CH(CH3)2), 1.28 (t, J¼7.19 Hz, 3H,
CO2CH2CH3), 1.60 (m, 3H, CH2CH2CH(CH3)2), 2.65þ2.7
and 2.78þ2.82 (dþd and dþd, J¼6.80, 6.40 Hz, 2H,
CH2CO2C2H5), 3.51 (q, J¼5.19, 7.19 Hz, 1H, CH–
NHCO–O–CH), 4.20 (q, J¼7.19 Hz, 2H, CO2CH2CH3),
4.58þ4.60þ4.62 (dþdþd, J¼4.80 Hz, 1H, CH–NHCO–
O–CH), 5.8 (s, broad, 1H, CH–NHCO–O–CH). 13C NMR
(100 MHz, CDCl3): d¼14.14 (s, CO2CH2CH3), 22.34 and
22.54 (s and s, CH2CH2CH(CH3)2), 27.92 (s, CH2CH2
CH(CH3)2), 33.15 (s, CH2CH2CH(CH3)2), 34.06 (s, CH2
CH2CH(CH3)2), 39.42 (s, CH2CO2C2H5), 57.84 (s, CO2
CH2CH3), 61.13 (s, CH–NHCO–O–CH), 78.02 (s, CH–
NHCO–O–CH), 158.58 (s, CH–NHCO–O–CH), 169.32
(s, CO2CH2CH3).
2. Reviews: (a) Liu, M.; Sibi, P. Tetrahedron 2002, 58, 7991. (b)
In Enantioselective Synthesis of b-Amino Acids; Juaristi, E.,
Ed.; Wiley-VCH: New York, 1997.
3. (a) Kim, B. M.; Bae, S. J.; So, S. M.; Yoo, H. T.; Chang, S. K.;
Lee, J. H.; Kang, S. Org. Lett. 2001, 3, 2349. (b) Adrian, Jr.
J. C.; Barkin, J. L.; Fox, R. J.; Chick, J. E.; Hunter, A. D.;
Nicklow, R. A. J. Org. Chem. 2000, 65, 6264. (c) Dinh, T. Q.;
Armstrong, R. W. J. Org. Chem. 1995, 60, 8118. (d) Yuste, F.;
Diaz, A.; Ortiz, B.; Sanchez-Obregon, R.; Walls, F.; Ruano,
J. L. G. Tetrahedron: Assymetry 2003, 14, 549.

S. Kambourakis, J. D. Rozzell / Tetrahedron 60 (2004) 663–669
669
4. Raddatz, P.; Radunz, H. E.; Schneider, G.; Schwartz, H.
Angew. Chem., Int. Ed. Engl. 1988, 27, 426.
7. Kambourakis, S.; Rozzell, J. D. Adv. Synth. Catal. 2003, 345,
699.
5. (a) Patel, R. N.; Banerjee, A.; Howell, J. M.; McNamee, C. G.;
Brzozowoski, D.; Nanduri, V.; Thottathil, J. K.; Szarka, L. J.
Tetrahedron: Asymmetry 1993, 4, 2069. (b) Gou, D. M.; Liu,
Y. C.; Chen, C. S. J. Org. Chem. 1993, 48, 1287.
8. Floyd, D. E.; Miller, S. E. Organic Syntheses, Collect. Vol. 4,
p 141.
9. Seebach, D.; Aebi, J.; Wasmuth, D. Organic Syntheses,
Collect. Vol. 63, p 109.
6. Patel, R. N.; Banerjee, A.; Ko, R. Y.; Howell, J. M.; Li, W. S.;
Comezoglu, F. T.; Partyka, R. A.; Szarka, L. J. Biotechnol.
Appl. Biochem. 1994, 20, 23.
10. Baumgarten, H. E.; Smith, H. L.; Staklis, A. J. Org. Chem.
1975, 40, 3554.