Preparation of (S)-1-cyclopropyl-2-methoxyethanamine by a chemoenzymatic route using leucine dehydrogenase

Article abstract of DOI:10.1021/op2003562

(S)-1-Cyclopropyl-2-methoxyethanamine is a key chiral intermediate for the synthesis of a corticotropin-releasing factor-1(CRF-1) receptor antagonist. Resolution of the racemic amine by transaminase from Vibrio fluvalis gave a 38% yield of the S-amine wit

Full text of DOI:10.1021/op2003562

Article
pubs.acs.org/OPRD
Preparation of (S)-1-Cyclopropyl-2-methoxyethanamine by a
Chemoenzymatic Route Using Leucine Dehydrogenase
William L. Parker, Ronald L. Hanson,* Steven L. Goldberg, Thomas P. Tully, and Animesh Goswami
Chemical Development, Bristol-Myers Squibb, One Squibb Drive, New Brunswick, New Jersey 08903, United States
ABSTRACT: (S)-1-Cyclopropyl-2-methoxyethanamine is a
key chiral intermediate for the synthesis of a corticotropin-
releasing factor-1(CRF-1) receptor antagonist. Resolution of
the racemic amine by transaminase from Vibrio fluvalis gave a
38% yield of the S-amine with 53% ee. Resolution by lipase-
catalyzed acylation provided the S-amine in 35% yield with 91%
ee. With limited success of these resolution approaches, an
efficient chemo-enzymatic route to (S)-1-cyclopropyl-2-me-
thoxyethanamine was devised starting from methylcyclopropyl ketone. Permanganate oxidation of the ketone gave
cyclopropylglyoxylic acid, which was converted to (S)-cyclopropylglycine by reductive amination using leucine dehydrogenase
from Thermoactinomyces intermedius with NADH cofactor recycling by formate dehydrogenase from Pichia pastoris. Both
enzymes were cloned and expressed in recombinant E. coli. (S)-Cyclopropylglycine obtained from enzymatic reductive amination
was isolated as the N-Boc derivative and converted to the desired amine by reduction, methylation, and deprotection to give (S)-
1-cyclopropyl-2-methoxyethanamine in 62% overall yield from cyclopropylglyoxylic acid, with no detectable R-enantiomer.
INTRODUCTION
RESULTS AND DISCUSSION
Resolution of Racemic 1-Cyclopropyl-2-methoxy-
■
■
Chiral amines are used for the synthesis of many pharmaceuti-
cally interesting compounds. Among the enzymatic processes
used to prepare chiral amines are acylation or deacylation using
lipases or proteases,1,2 or oxidation using amine oxidases.3
Enantioselective acylation combined with chemical racemiza-
tion of the amines1,2,4 and enantioselective oxidation combined
with nonselective chemical reduction of the imine product3
have allowed dynamic resolutions giving >50% yields in some
cases. Amine transaminases have also been used for the
preparation of chiral amines.5−7 The S-specific transaminases
are more common than R-specific enzymes and can be used to
convert a racemic mixture to the R-enantiomer and a ketone.
An R-specific enzyme has been described,8 and some R-specific
transaminases are commercially available. Prochiral ketones
have also been transaminated to chiral amines when conditions
have been devised to overcome the unfavorable equilibrium for
ethanamine (8). An amine transaminase from Bacillus
megaterium expressed in Escherichia coli SC16578 was
previously used for conversion of racemic 1-cyclopropylethyl-
amine to the R-amine by selective transamination of the S-
enantiomer.7 Since (R)-cyclopropylethylamine has the same
arrangement of cyclopropyl and open-chain substituents as S-
amine 3, this transaminase was tested for resolution of racemic
1-cyclopropyl-2-methoxyethanamine (8). The enzyme had low
enantioselectivity favoring the undesired transamination of S-
amine with pyruvate as the amino group acceptor, giving an ee
of 43% for the undesired R-enantiomer.
A screening kit of 17 amine transaminases (Codexis) was
also tested for resolution of racemic amine 8 with pyruvate as
the amino group acceptor following the screening procedure of
the manufacturer. ATA-114 gave 30.5% ee of the S-amine,
ATA-117 gave 5.6% ee of the S-amine and ATA-113 gave 57%
ee of the R-amine. The remaining 14 enzymes gave no reaction.
Transaminase from Vibrio fluvalis gave a 38% yield (with a
maximum 50%) of S-amine with 53% ee.
this reaction.9,10
(S)-1-Cyclopropyl-2-methoxyethanamine (7) was required
for synthesis of a corticotropin-releasing factor-1(CRF-1)
receptor antagonist and was previously prepared using an
asymmetric Strecker reaction to introduce the chirality.11,12
Enzymatic approaches were explored to improve atom
economy, avoid the use of cyanide, and develop an efficient
“green” process. This report describes an efficient chemo-
enzymatic route for the preparation of (S)-cyclopropylglycine
by reductive amination of cyclopropylglyoxylic acid, conversion
to the Boc derivative, reduction to the glycinol, O-methylation,
and deprotection to give the desired amine.
Strains were selected from soil samples able to grow on the
R-enantiomer of amine 7 as sole nitrogen source. However
when supplied with the racemic amine, the isolates consumed
the R-amine, but preferentially used the S-amine, therefore
increasing the ee of R-amine. Two D-amino acid oxidases, six L-
amino acid transaminases, and two D-amino acid transaminases
had no effect on racemic amine 8. An N-Cbz cleaving enzyme13
Received: December 5, 2011
Published: February 2, 2012
© 2012 American Chemical Society
464
dx.doi.org/10.1021/op2003562 | Org. Process Res. Dev. 2012, 16, 464−469

Organic Process Research & Development
Article
Scheme 1. Resolution of racemic amine 8 using Cal B lipase
enantioselective for N-α-Cbz-L-amino acids had no effect on
Cbz-protected racemic amine 8.
phenylalanine dehydrogenase from T. intermedius which gave
94% ee.
Leucine dehydrogenase from T. intermedius expressed in E.
coli SC16591 was chosen for preparative purposes. Comparing
extract activities with 5 mM α-ketoisovalerate (3282 U/mL)
versus 5 mM cyclopropylglyoxylic acid (86 U/mL), leucine
dehydrogenase is 38-fold more active with the natural substrate.
Formate dehydrogenase from Pichia pastoris expressed in E. coli
SC16556 was used for regeneration of NADH. The
construction of the recombinant strains, production of cells
by fermentation, and preparation of extracts containing the
enzyme activities are described in the Experimental Section.
The pH optimum for leucine dehydrogenase with cyclo-
propylglyoxylic acid as substrate was found to be 9.0 (Figure 1).
Resolution of racemic amine 8 by acylation using ethyl
caprate 9 and Cal B lipase in MTBE gave a 35% solution yield
of S amine 7 with 91.3% ee (Scheme 1). This approach was
previously used to prepare S-amines,14 but in this case the
selectivity was not sufficient to achieve the necessary ee (>99%)
with good yield.
After initial attempts at resolution of the racemic amine with
amine transaminases, lipases and other enzymes met with
limited success, a completely different approach using the
enzymatic reductive amination of a keto acid to an amino acid
was pursued to give the desired enantiomer. We have
previously prepared several unnatural S-amino acids from the
corresponding keto acids using amino acid dehydro-
genases15−18 and explored this approach as an alternative
route to the chiral amine. The chemoenzymatic route to the
chiral amine 7 is outlined in Scheme 2.
Scheme 2. Route for conversion of cyclopropymethyl ketone
(1) to (S)-1-cyclopropyl-2-methoxyethanamine (7)
Figure 1. pH optimum for leucine dehydrogenase. The assay mixture
contained 5 mg/mL keto acid 2 potassium salt (32.8 mM), 83 mg/mL
(1.31 M) ammonium formate, 0.1 mM dithiothreitol, and 0.2 mM
NADH. It was adjusted to the indicated pH at 40 °C.
However, since the pH optimum for P. pastoris formate
dehydrogenase has been reported to be 6.5−7.5,20,21 the
reaction was run at pH 8. The apparent Km for NADH for
leucine dehydrogenase with 33 mM cyclopropylglyoxylic acid at
pH 8 was found to be 0.023 mM. Reactions were run with 1
mM NAD, although the low Km indicates a lower concentration
should suffice. Reported Km’s for NAD for formate
dehydrogenase are 0.14 mM20 and 0.27 mM.21 Batches with
1, 25, and 50 g inputs of keto acid 2 potassium salt (using 10%
keto acid salt concentration as described in the Experimental
Section) gave 99−100% solution yield of (S)-cyclopropyglycine
with no detectable R-enantiomer.
Conversion of S-Amino Acid 3 to S-Amine 7. A small
quantity of S-amino acid 3 was isolated and recrystallized.
Treatment of the enzymatic reaction mixture with Boc
anhydride gave N-Boc-amino acid 4. Reduction of 4 to Boc-
(S)-cyclopropylglycinol 5 with various borane and mixed
anhydride/borohydride protocols met with varying degrees of
success, but reduction with sodium bis(2-methoxyethoxy)-
aluminum hydride22 gave superior results. Boc-(R)-cyclo-
propylglycinol has been reported.23 Attempts to methylate
the alcohol function with alkylating agents and base in
homogeneous solution did not give high selectivity for O-
versus N-alkylation. However, phase-transfer methylation
provided a highly selective means of producing methyl ether
6 in the presence of a Boc-protected primary amine.
Deprotection with trifluoroacetic acid gave the desired (S)-1-
Preparation of Cyclopropylglyoxylic Acid (2). An initial
preparation of cyclopropylglyoxylic acid 2 was obtained by
addition of cyclopropyl magnesium bromide to diethyl oxalate,
followed by hydrolysis of the resulting ethyl ester to give the
acid. A better method, oxidation of cyclopropyl methyl ketone
(1) with potassium permanganate, was subsequently employed.
Preparation of cyclopropylglyoxylic acid 2 by oxidation of
cyclopropyl methyl ketone 1, conversion to racemic cyclo-
propyl glycine, and enzymatic resolution of the Boc-protected
methyl ester has been described.19
Reductive Amination of Cyclopropylglyoxylic Acid.
Leucine dehydrogenases from Bacillus sphaericus15 and from
Thermoactinomyces intermedius cloned into E. coli SC16591,
recombinant phenylalanine dehydrogenase from T. intermedius
(original or modified version),18 commercially available
glutamate dehydrogenase from beef liver, alanine dehydrogen-
ase from Bacillus subtilis, and phenylalanine dehydrogenase
from Rhodococcus species converted the keto acid precursor to
(S)-cyclopropyglycine (3). The R-amino acid was produced
using each of four R-amino acid dehydrogenases (D-AADH-
101, -102, -103, -104 from Codexis). All of the amino acid
dehydrogenases gave product with 100% ee, except modified
465
dx.doi.org/10.1021/op2003562 | Org. Process Res. Dev. 2012, 16, 464−469

Organic Process Research & Development
Article
cyclopropyl-2-methoxyethanamine 7 in 62% overall yield from
cyclopropylglyoxylic acid, with no detectable R-enantiomer.
using L-valine as an HPLC standard (correcting for molecular
weight). There was no detectable R-enantiomer.
A 57-g portion of the reaction mixture (pH 8.4) containing
3.8 g of the amino acid 3 was adjusted to pH 2 (H2SO4) to
precipitate protein, adding n-BuOH as necessary to break the
copious foam. After 1 h the mixture was heated to 80 °C,
cooled to room temperature, and filtered. The filtrate was
adjusted to pH 7 (NaOH), concentrated in vacuo to a thick
slurry (24 g), and filtered. The resulting solid (dry weight 4.2 g)
was recrystallized from 20 mL of water to give 1.5 g of 3, mp
CONCLUSION
■
Enzymatic resolution approaches to preparation of amine 7 did
not give sufficiently high enantiomeric purity. An alternative
approach using an S-amino acid dehydrogenase to introduce
the chirality followed by reduction of the Boc-protected amino
acid, then methylation and deprotection gave the desired
amine. Since R-amino acid dehydrogenases gave the R-amino
acid, it is expected that this method could also be used to
prepare the R-amine.
1
>300 °C. The H NMR spectrum in D2O was in agreement
with published data.25
Isolation of Boc-(S)-cyclopropylglycine (4). A 509-g
batch of enzymatic reaction mixture derived from 45.9 g (302
mmol) of cyclopropylglyoxylic acid 2 potassium salt was
adjusted to pH 10.00 (NaOH), and 143 g of di-tert-butyl
dicarbonate melt (650 mmol) was added. The mixture was
stirred vigorously and the pH maintained between 9.7 and 10.1
for 5 h. Stirring at room temperature was continued for another
16 h. The reaction mixture, 700 mL, pH 9.6, was adjusted to
pH 7.9 (H2SO4) and stirred with 700 mL of isopropyl acetate,
and the pH was adjusted to 2.0 (H2SO4). Diatomaceous earth
was added and the mixture filtered. The filtrate was adjusted to
pH 8.4 (NaOH) and the upper phase (containing Boc-NH2)
discarded. The lower phase (∼1000 mL) was stirred with 500
mL of isopropyl acetate and the pH adjusted to 2.0 (H2SO4).
The lower phase was discarded and the upper concentrated in
vacuo to 265 g (about 62 g of product and 230 mL of isopropyl
acetate). After seeding and crystallizing overnight at room
temperature, 460 mL of heptane was added over 1 h and
crystallization continued with ice-bath cooling for 1 h. Filtration
and drying gave 53.7 g of Boc-(S)-cyclopropylglycine 4, mp
116−117.6 °C, 83% yield. Concentration of the mother liquor
gave an additional 8.2 g of product 4, mp 108−113.0 °C.
Reduction of Boc-(S)-cyclopropylglycine (4) to Boc-
(S)-cyclopropylglycinol (5). A solution of 2.02 g (9.39
mmol) of N-Boc-(S)-cyclopropylglycine 4 in 10 mL of dry
THF (DriSolv, EMD Chemicals) was added dropwise over 15
min to 9.7 mL of a 66% solution of sodium bis(2-
methoxyethoxy)aluminum hydride in toluene (Red-Al, 33
mmol) under nitrogen with stirring and cooling in an ice
bath. After an additional 2 h, 2.5 mL (62 mmol) of MeOH was
added dropwise, releasing 630 mL of hydrogen (very vigorous
evolution). Finally, a solution of 20 g of potassium sodium
tartrate tetrahydrate (Rochelle’s salt) in 20 mL of water was
added, releasing an additional 40 mL of hydrogen. The mixture
was stirred at room temperature for 15 min, and the upper of
three phases was separated. The combined lower phases were
extracted with three 10-mL portions of isopropyl acetate. The
combined organic extract (upper phase) was washed with two
10-mL portions of 5% NaHCO3 and two 10-mL portions of
water. The extract was concentrated in vacuo, giving 1.64 g of
solid, which was warmed gently in vacuo until it just melted.
After cooling, the viscous liquid was scraped thoroughly with a
seeded spatula and warmed a little while the solid crystallized in
vacuo, giving 1.62 g (8.07 mmol, 86% yield) of 5, mp 62−64.5
°C. This material was essentially pure and satisfactory for
further use without additional purification. Product from a
different preparation (reduction of the mixed anhydride from
isobutylchloroformate with sodium borohydride) was recrystal-
lized from heptane to give 5 as crystalline solid (ultrafine
needles) melting at 65.5−66.5 °C. 1H NMR (400 MHz,
DMSO-d6): δ 0.13 (m, 1H), 0.25 (m, 2H), 0.37 (m, 1H), 0.82
EXPERIMENTAL SECTION
■
Resolution of Amine 8 by Transaminase. A solution
containing 10 mg racemic amine 8, 10 mg sodium pyruvate,
and 0.02 mL ω-transaminase from V. fluvalis (Julich) in 1 mL
0.1 M potassium phosphate buffer pH 7.5 was incubated at 28
°C. After 160 h, 43% of the amine remained with 53% ee of the
S-enantiomer.
Resolution of Amine 8 Using Cal B Lipase. To 1 mL of
methyl tert-butyl ether solution containing 2 mg of racemic
amine 8 were added ethyl caprate 9 (20 μL) and Cal B lipase
from Candida antarctica (50 mg, from Codexis), and the
mixture was shaken at 500 rpm and 25 °C. After 120 h, 35% of
the amine remained with 91.3% ee of the S enantiomer 7.
Preparation of Cyclopropylglyoxylic Acid (2). Oxida-
tion of cyclopropyl methyl ketone 1 with potassium
permanganate was done essentially as described by Basnak
and Farkas24 except that additional KMnO4 was used to
consume unreacted ketone detected by HPLC. A solution of
140 g (1.66 mols) of cyclopropyl methyl ketone 1 and 5.65 g of
K2CO4 in 1.12 L of water was heated to 50 °C, and a solution
of 526 g (3.33 mol) KMnO4 in 11.4 L of water was added over
54 h. After addition was complete, the temperature was
maintained at 50 °C for 3 h. At this point only a trace of
unreacted cyclopropyl methyl ketone 1 remained, and the
solution yield of cyclopropylglyoxylic acid 2 by HPLC was 75%.
To remove the excess KMnO4, 250 mL of 3.5% H2O2 was
added, completely discharging the deep purple color. The
solution was cooled to 20 °C and filtered to remove MnO2.
The filtrate was concentrated in vacuo to a solid residue.
Recrystallization from methanol (3 L) gave 128 g, 50% yield, of
the potassium salt of cyclopropyl glyoxylic acid 2 as a nacreous
solid, mp 271.5−273.0 °C (dec). The mother liquor contained
61 g of additional product which was not recovered.
Preparation of (S)-Cyclopropylglycine (3). Cyclopropyl-
glyoxylic acid 2 potassium salt (50 g, 0.328 mol), ammonium
formate (41.5 g, 0.657 mols) and water (408 mL) were added
to a jacketed 1-L reactor. Magnetic stirring was used to dissolve
the solids, and the temperature was maintained at 40 °C. The
solution was adjusted to pH 8.0 by addition of 4.6 mL of 5 N
NaOH. Dithiothreitol (77 mg, 0.5 mmol), NAD (331 mg, 0.5
mmol), formate dehydrogease (814U), and leucine dehydro-
genase (2157 U measured with cyclopropylglyoxylic acid 2 as
substrate) were added. The stirred solution was maintained at
40 °C, and pH 8.00 was maintained by addition of 5 N NaOH
from a pH stat. The initial concentration of keto acid was 100
mg/mL. After 2 h the concentration was 67 mg/mL, and after
17 h cyclopropylglyoxylic acid 2 was not detectable. The
solution yield was 37.43 g (S)-cyclopropylglycine 3 (99.0%),
466
dx.doi.org/10.1021/op2003562 | Org. Process Res. Dev. 2012, 16, 464−469

Organic Process Research & Development
Article
(m, 1H), 1.37 (s, 9H), 2.97 (m, 1H), 3.38 (m, 2H), 4.52 (t, 1H,
J = 5.6 Hz, rapid exchange with methanol-d4), 6.10 (br, 0.1 H,
slow exchange with methanol-d4, minor rotamer), 6.47 (br d,
0.9H, J = ∼7 Hz, slow exchange with methanol-d4).
(Invitrogen) at 37 °C for 2 h in a 40 μL final volume. The
entire reaction was electrophoresed on a 1.0% TAE agarose gel,
and the amplified fragment was excised for purification using
the Qiagen Gel Isolation Kit.
Phase Transfer O-Methylation of N-Boc-(S)-cyclo-
propylglycinol 5. A solution of 1.47 g of Boc-(S)-cyclo-
propylglycinol 5 (7.30 mmol) and 4.8 mg of benzyltriethy-
lammonium chloride (0.021 mmol) in 14 mL of dichloro-
methane (DCM) was cooled in an ice bath and stirred, and 3.5
mL of 50% NaOH (65 mmol) was added followed by 1.04 mL
of dimethyl sulfate (11.0 mmol). The mixture was stirred with
cooling for 0.5 h and then at room temperature for 17 h. HPLC
indicated that about 1% of the starting material remained. To
destroy the remaining dimethyl sulfate, 1 mL of 15 M
ammonium hydroxide was added, and the mixture was stirred
for 1 h at room temperature. The upper (organic) phase was
separated, washed with water. and concentrated gently (140
mm, 20 °C) in vacuo to avoid volatilization of the product. The
residue was redissolved in ethyl acetate and concentrated (40
mm, 20 °C) twice to remove water azeotropically, giving 2.3 g
of crude 6, as an oil.
Preparation of O-Methyl-(S)-cyclopropylglycinol (7).
Crude N-Boc-O-methyl-(S)-cyclopropylglycinol (6), 2.3 g, was
dissolved in 7 mL of TFA. After 10 min the solution was
concentrated in vacuo, giving 2.71 g of oil. The oil was
dissolved in 10 mL of DCM and shaken with 14 mL of 5 M
NaOH. The lower (organic) phase was separated and stirred
with 2 mL of water, and the pH (12) was adjusted to 5−6 with
1.00 N HCl (6.80 mL, sharp end point). The basic aqueous
phase was extracted with an additional 2 mL of DCM, the
extract was added to the first extract mixture, and the pH was
readjusted to 5−6 with 1 N HCl (0.20 mL). The upper phase
was separated and concentrated in vacuo, giving 1.06 g of crude
HCl salt of 7, mp 192.5−193.0 °C. Recrystallization from a
mixture of dichloromethane (in which the salt is soluble) and
isopropyl acetate gave 0.97 g of O-methyl-(S)-cyclopropylgly-
cinol 7, HCl salt (hygroscopic), mp 193.0−194.0 °C, yield 88%
from Boc-(S)-cyclopropylglycinol. Anal. Calcd for C6H14ClNO:
C, 47.52; H, 9.30; N, 9.23; Cl 23.38. Found: C, 47.49; H, 9.58;
N, 9.43; Cl, 23.22.
The digested PCR fragment was ligated to NdeI + BamHI-
digested expression vector pBMS2004 in a 5:1 molar ratio using
the FastLink ligation kit (Epicentre) for 15 min at room
temperature. DNA was precipitated by addition of 10 volumes
of 1-butanol and pelleted by centrifugation for 5 min, 13,500g.
The liquid was removed and the pellet washed with 200 μL
70% EtOH before drying in a SpeedVac for 3 min with heat.
The pellet was resuspended in 4 μL water to which 40 μL of
electrocompetent DH10B cells (Invitrogen) was added.
Electroporation was carried out in a 0.2-cm gap cuvette
(Invitrogen) at 25 kV, 25 μF, 200 Ω. SOC medium (960 μL)
was added, and the suspended cells were shaken at 37 °C, 225
rpm for 1 h. Cells were concentrated by centrifugation at
13,500g, and most of the medium was removed. Cells were
resuspended in the remaining liquid and spread for single
colonies on LB agar plates containing 50 μg/mL of kanamycin
sulfate (Sigma), which were then incubated at 37 °C for 20 h.
Correct ligation was confirmed by PCR of Km colonies using
primers included in the original amplification reaction.
Cloning of the P. pastoris Formate Dehydrogenase
(PPFDH) Gene. Initial isolation and cloning of the PPFDH
gene into expression vector pBMS2000 was previously
described.18 The gene27 was subsequently modified by
preparing a synthetic version whose codons usage reflects
that found in E. coli (“Ecobias”) and included an amino acid
change at amino acid 245 from serine to alanine (S245A).
These two modifications led to substantial increases in titer and
stability of the recombinant PPFDH protein.
Heterologous Expression of TILDH. Plasmid DNA was
prepared from a culture confirmed to contain pBMS2004-
TILDH using the Fast Plasmid kit (Eppendorf) and trans-
formed into electrocompetent E. coli BL21 cells as described
above. A single kanamycin-resistant colony was inoculated into
MT5-M2 medium containing 50 μg/mL kanamycin and grown
at 30 °C, 250 rpm, for 20 h. The optical density of the culture
at 600 nm was measured and diluted to OD = 0.3 into 25 mL
MT5-M2 Km medium in a 250-mL flask. The flask was
incubated at 30 °C, 250 rpm, until the OD600 nm was 0.8−1.0.
The culture was divided into two 10-mL aliquots in 250-mL
flasks and IPTG to 50 μM or 1 mM added from a 1 M sterile
stock in water. The flasks were placed on a 30 °C shaker for an
additional 22 h. Cultures were transferred to a 15-mL
disposable polypropylene tube, and cells were pelleted in a
Beckman JA 5.3 swinging bucket rotor for 7 min, 5,000g, 23 °C.
Medium was removed, and the cells were resuspended in a
buffer consisting of 50 mM sodium phosphate buffer pH 7.0 +
1 mM dithiothreitol (DTT). Cells were pelleted as above, and
buffer was removed. The wet cell weight was recorded, and
cells were resuspended in 10 vol of the same buffer. An 0.8-mL
aliquot of the cell suspension was placed into a 1.5-mL
microfuge tube and lysed by sonication using a Fisher Sonic
Dismembrator, (setting “15” with microtip; 3 × 15 s). The
samples were centrifuged 13,500g for 5 min, and the lysate was
placed in a new centrifuge tube. Overexpression of a soluble
40,000 Da protein was confirmed by SDS-polyacryamide
electrophoresis of a 1-μL aliquot of the sonicate followed by
staining using the Simply Blue reagent (Invitrogen). Slightly
better expression was observed after induction with the 1 mM
IPTG, but growth of the culture was greatly retarded as
The preparation of O-methyl-(S)-cyclopropylglycinol 7 by an
alternate route has been reported.11,12
Cloning of the T. intermedius Leucine Dehydrogenase
(TILDH) Gene. A polymerase chain reaction (PCR) was
carried out using primers homologous to the 5′ and 3′ ends of
the published TILDH DNA sequence26 that also contained
restriction endonuclease recognition sites for NdeI and BamHI,
respectively. The reaction used T. intermedius chromosomal
DNA as template and Z-Taq DNA polymerase (Takara) with
included buffer and dNTP solution. Thermocycling conditions
were the followiong: 94 °C, 2 min (1 cycle); 94 °C, 30 s; 50 °C,
30 s; 72 °C, 1 min (30 cycles); 72 °C, 5 min (1 cycle). A small
aliquot of the reaction was electrophoresed on a 1.0% TAE
agarose gel to verify amplification of a 1.1 kb fragment. The
remainder of the reaction was extracted with an equal volume
1:1 phenol/CHCl3 and centrifuged at 13,500g for 2 min. The
upper layer was removed to a new microfuge tube and DNA
precipitated by addition of 0.1 vol 3 M sodium acetate pH 4.8
and 2 vol. 100% EtOH. DNA was pelleted by centrifugation at
13,500g for 5 min. The pellet was washed with 200 μL 70%
EtOH, the liquid was removed, and the tube was placed in a
SpeedVac unit for 3 min with heat. DNA was resuspended in
34 μL water and digested with 7.5 U each NdeI and BamHI
467
dx.doi.org/10.1021/op2003562 | Org. Process Res. Dev. 2012, 16, 464−469

Organic Process Research & Development
Article
compared to the lower level of inducer tested. Further
optimization indicated that 100 μM IPTG gave the preferred
level of growth and activity.
Preparation of Formate Dehydrogenase from E. coli
SC16556 and Leucine Dehydrogenase from E. coli
SC16591. Cells were suspended in 50 mM potassium
phosphate buffer pH 7 containing 1 mM dithiothreitol at a
20% w/v concentration. Cells were disrupted by two passages
through a microfluidizer. Poly(ethyleneimine) (PEI, Sigma)
was added to the extracts to 0.25% w/v. After 30 min on ice to
allow coagulation of debris, the extracts were clarified by
centrifugation for 12 min at 15000g. The supernatants were
stored at −20 °C until use. There was no loss of activity for
either enzyme after more than 4 months storage at −20 °C.
The formate dehydrogenase activity was 32.6 U/ml. Leucine
dehydrogenase activity was 3282 U/mL measured with α-
ketoisovalerate as substrate and 86 U/mL measured with
cyclopropylglyoxylate as substrate. Activities used for the
preparative batches are reported in units with 5 mM
cyclopropylglyoxylic acid as the substrate.
Enzyme Assays. Assays were done using 1-cm path length
cuvets. The leucine dehydrogenase assay contained in 1.0 mL at
40 °C: 0.2 mM NADH, 5 mM sodium α-ketoisovalerate or 5
mM potassium cyclopropylglyoxylate, 0.75 M NH4OH adjusted
to pH 8.75 with HCl. Absorbance decrease was monitored at
340 nm. Blanks contained enzyme but no keto acid (to correct
for background NADH oxidation). Enzyme activity units were
calculated as μmol/min based on the rates of absorbance
change.
Growth of E. coli BL21(pBMS2004-TILDH) SC16591
Cells Expressing Leucine Dehydrogenase. Two frozen
vials of E. coli SC16591 were thawed, and the contents were
transferred to two 500-mL flasks containing 100 mL of MT5-A
medium (2.0% Yeastamin, 4.0% glycerol, 0.6% Na2HPO4, 0.3%
KH2PO4, 0.125% ammonium sulfate, and 0.005% Km sulfate
(filter-sterilized). The flasks were incubated at 30 °C and 250
rpm for 24 h, after which 5 mL was transferred to four 4-L
flasks containing 1 L of the same medium. These second-stage
flasks were then incubated at 30 °C and 230 rpm for 22 h. A
sufficient quantity (1.4 L) to yield a starting OD600 of ∼0.1
AU/cm in 100 L of MT5-M2 medium (2.0% Quest Hy-Pea,
1.85% Tastone 154, 4.0% glycerol, 0.6% Na2HPO4, 0.125%
ammonium sulfate, 0.04% Ucon LB625, and 0.005% Km
sulfate, filter-sterilized) was then pooled and transferred to the
fermentor. Process conditions for the 147-L fermentor
containing 100 L of medium were as follows: 30 °C; no pH
control; agitation, 500 rpm; aeration, 100 lpm; back pressure,
10 psi; foam controlled by the addition of UCON LB625 on
demand.
At an OD600 of ∼1.6 AU/cm, which occurred 4.5 h following
inoculation, a filter-sterilized IPTG solution was added to a final
level of 100 μM. The run was completed after 20 h, at which
time OD600 was ∼40 AU/cm, and off-gas CO2 had declined
from a peak of 2.0% to 0.28%. The whole broth was cooled to
4−10 °C, after which the cells were recovered by centrifugation
and washed with 50 mM pH 7 phosphate buffer. A total of 4.3
kg of cell paste was recovered and stored at −75 °C.
The formate dehydrogenase assay contained in 1.0 mL at 40
°C: 1 mM NAD, 100 mM sodium formate, 100 mM potassium
phosphate buffer, pH 8.0. Absorbance increase was monitored
at 340 nm. Blanks contained no enzyme. Enzyme activity units
were calculated as μmol/min based on the rates of absorbance
change.
Growth of E. coli JM110(pBMS2000-Ecobias
FDHS245A) SC16556 Cells Expressing Formate Dehy-
drogenase. Plasmid pBMS2000-Ecobias FDHS245A was
transformed into E. coli strain JM110 by electroporation as
described above, forming strain SC16556. Two frozen vials of
E. coli SC16556 were thawed, and the contents were transferred
to two 500-mL flasks containing 100 mL of MT5 medium
(2.0% Yeastamin, 4.0% glycerol, 0.6% Na2HPO4, 0.3%
KH2PO4, 0.125% ammonium sulfate, 0.0246% magnesium
sulfate (added separately), and 0.005% Km sulfate (filter-
sterilized). Flasks were incubated at 30 °C and 250 rpm for 8 h,
followed by a 5 mL transfer of broth to four 4-L flasks
containing 1 L of the same medium. These second-stage flasks
were incubated at 30 °C and 230 rpm for 28 h and then pooled
and transferred to a fermentor containing 250 L of MT5-M2
medium (2.0% Quest Hy-Pea, 1.85% Tastone 154, 4.0%
glycerol, 0.6% Na2HPO4, 0.125% ammonium sulfate, 0.04%
Ucon LB625, and 0.005% Km sulfate, the latter filter-sterilized
and added separately). The resulting OD600 at the start of the
run was ∼0.034 AU/cm. Process conditions for the 380-L
fermentor containing 250 L of medium were as follows: 30 °C;
no pH control; agitation, 550 rpm; aeration, 250 lpm; back
pressure, 10 psi; foam controlled by the addition of UCON
LB625 on demand.
HPLC Methods. For the Cal B resolution of racemic amine,
the amount of remaining amine and enantiomeric composition
were determined by conversion to the dansyl derivative28
followed by analysis on C18 and chiral columns. Analysis on a
YMC pack Pro C18 150 × 4.6 mm 3 μm column used a
gradient of solvent A (0.05% TFA in water:methanol 80:20)
and solvent B (0.05% TFA in acetonitrile:methanol 80:20)
starting from 0% B to 100% B in 20 min at a flow rate of 1 mL/
min at 40 °C, and UV detection at 220 nm whereby the dansyl
derivative of 1-cyclopropyl-2-methoxyethanamine showed a
peak at 12.5 min. Analysis on a Chiralpak AS-RH, 150 × 4.6
mm, 5 μm column (Chiral Technologies Inc.) was carried out
using an isocratic mixture of 76% solvent A and 24% solvent B
at a flow rate of 0.5 mL/min at 30 °C and detection by UV at
220 nm. The peaks for the dansyl derivatives of the S- and R-
enantiomers of 1-cyclopropyl-2-methoxyethanamine were at
20.2 and 23.6 min, respectively.
To measure the ee of (S)-1-cyclopropyl-2-methoxyethan-
amine 7 for all other experiments, Marfey’s reagent was used to
generate diasteromeric derivatives. A sample of 10 μL
containing about 0.1 mg amine, 8 μL 1 M NaHCO3, and 40
μL 1% w/v Marfey’s reagent (N-α-[2,4-dinitrophenyl-5-
fluorophenyl]-L-alanine amide, Pierce) in acetone were
combined and heated for 1 h at 40 °C. The samples were
cooled to room temperature; then 8 μL of 1 N HCl and 934 μL
of 40% acetonitrile/water were added, and the solutions were
vortexed and filtered into HPLC vials. Samples were analyzed
with a YMC Pak Pro C18 15 cm × 0.46 cm 3 μm column. The
mobile phase was 35% acetonitrile/65% water (each containing
0.05% trifluoroacetic acid), flow rate was 1 mL/min, detection
was at 340 nm, temperature was 40 °C, and injection volume
Three hours following inoculation, a filter-sterilized IPTG
solution was added to a final level of 50 μM. The run was
completed after 20 h, at which time OD600 was 56 AU/cm, and
off-gas CO2 was still climbing. The whole broth was cooled to
4−10 °C, after which the cells were again recovered by
centrifugation and washed with 50 mM pH 7 phosphate buffer.
A total of 13.3 kg of cell paste was recovered and stored at −75
°C.
468
dx.doi.org/10.1021/op2003562 | Org. Process Res. Dev. 2012, 16, 464−469

Organic Process Research & Development
Article
Krishnananthan, S.; Wong, H.; Qian-Cutrone, J.; Schartman, R.;
Denton, R.; Lodge, N. J.; Zaczek, R.; Macor, J. E.; Bronson, J. J. J. Med.
Chem. 2009, 52, 7653.
(12) Leahy, D. K.; Li, J.; Sausker, J. B.; Zhu, J.; Fitzgerald, M. A.; Lai,
C.; Buono, F. G.; Braem, A.; de Mas, N.; Manaloto, Z.; Lo, E.; Merkl,
W.; Su, B.-N.; Gao, Q.; Ng, A. T.; Hartz, R. A. Org. Process Res. Dev.
2010, 14, 1221.
(13) Nanduri, V. B.; Goldberg, S.; Johnston, R; Patel, R. N. Enzyme
Microb. Technol. 2004, 34, 304.
(14) Goswami, A.; Guo, Z.; Parker, W. L.; Patel, R. N. Tetrahedron:
Asymmetry 2005, 16, 1715.
was 20 μL. Retention times were S-enantiomer (7), 17.1 min;
R-enantiomer, 21.2 min.
Marfey’s reagent was used as described above to determine
the ee of (S)-cyclopropylglycine 3. Samples were analyzed with
a YMC Pak Pro C18 15 cm × 0.46 cm 3 μm column. The
mobile phase was 32% acetonitrile/68% water (each containing
0.05% trifluoroacetic acid), flow rate was 1 mL/min, detection
was at 340 nm, temperature was 40 °C, and injection volume
was 20 μL. Retention times were: S-enantiomer (3), 8.7 min; R-
enantiomer, 14.4 min
(15) Hanson, R. L.; Singh, J.; Kissick, T. P.; Patel, R. N.; Szarka, L. J.;
Mueller, R. H. Bioorg. Chem. 1990, 18, 116.
(16) Hanson, R. L.; Schwinden, M. D.; Banerjee, A.; Brzozowski, D.
B.; Chen, B.-C; Patel, B. P.; McNamee, C. G.; Kodersha, G. A.;
Kronenthal, D. R.; Patel, R. N.; Szarka, L. J. Bioorg. Med. Chem. 1999,
7, 2247.
(17) Hanson, R. L.; Howell, J. M.; LaPorte, T. L; Donovan, M. J.;
Cazzulino, D. L.; Zannella, V.; Montana, M. A.; Nanduri, V. B.;
Schwarz, S. R.; Eiring, R. F.; Durand, S. C.; Wasylyk, J. M.; Parker, W.
L.; Liu, M. S.; Okuniewicz, F. J.; Chen, B.-C.; Harris, J. C.; Natalie, K.
J.; Ramig, K.; Swaminathan, S.; Rosso, V. W.; Pack, S. K.; Lotz, B. T.;
Bernot, P. J.; Rusowicz, A.; Lust, D. A.; Tse, K. S.; Venit, J. J.; Szarka,
L. J.; Patel, R. N. Enzyme Microb. Technol. 2000, 26, 348.
(18) Hanson, R. L.; Goldberg, S. L.; Brzozowski, D. B.; Tully, T. P;
Cazzulino, D.; Parker, W. L.; Lyngberg, O. K.; Vu, T. C.; Wong, M. K.;
Patel, R. N. Adv. Synth. Catal. 2007, 349, 1369.
(19) Larionov, O. V.; de Meijere, A. Adv. Synth. Catal. 2006, 348,
1071.
(20) Hou, C. T.; Patel, R. N.; Laskin, A. I.; Barnabe, N. Arch.
Biochem. Biophys. 1982, 216, 296.
(21) Allais, J. J.; Louktibi, A.; Baratti, J. Agric. Biol. Chem. 1983, 47,
2547.
(22) Juszczyk, P.; Kasprzykowska, R.; Kolodziejczyk, A. S. Lett. Pept.
The ee of 3 was also determined with a Regis Davankov
Ligand Exchange 15 cm × 0.46 cm chiral column. Samples of
0.02 mL were diluted with 0.98 mL water and placed in a
boiling water bath for 2 min. For quantitation, L-valine was used
as a standard, assuming the same molar absorbance for the
copper complexes of the amino acids. The mobile phase was
0.5 mM CuSO4 in water, flow rate was 1 mL/min, detection
was at 235 nm, temperature was 40 °C, and injection volume
was 10 μL. Retention times were: S-enantiomer (3) 2.2 min, R-
enantiomer 4.9 min.
Cyclopropylglyoxylic acid (2) was determined in the same
samples using a YMC Pak Pro C18 15 cm × 0.46 cm 3 μm
column. The mobile phase was 5% acetonitrile/95% water
(each containing 0.05% trifluoroacetic acid), flow rate was 1
mL/min, detection was at 220 nm, temperature was 40 °C, and
injection volume was 10 μL. The retention time for 2 was 4.5
min.
AUTHOR INFORMATION
■
Corresponding Author
*Telephone: (732)-227-6222. Fax: (732)-227-3994. E-mail:
ronald.hanson@bms.com.
Sci. 2003, 10, 79.
(23) Lind, K. E.; Cao, K.; Lin, E. Y.-S.; Nguyen, T. B.; Tangonan, B.
T.; Erlanson, D. A.; Guckian, K.; Simmons, R. L.; Lee, W.-C.; Sun, L.;
Hansen, S.; Pathan, N.; Zhang, L. Preparation of pyridinonyl PDK1
inhibitors. PCT Int. Appl. WO 2008005457 A2 2008
(24) Basnak, I.; Farkas, J. Collect. Czech. Chem. Commun. 1975, 40,
1038.
(25) Hallinan, K. O.; Crout, D. H. G.; Errington, W. J. Chem. Soc.,
Perkin Trans. 1 1994, 3537.
(26) Ohshima, T.; Nishida, N.; Bakthavatsalam, S.; Katoaka, K.;
Takada, H; Yoshimura, T.; Esaki, N.; Soda, K. Eur. J. Biochem. 1994,
222, 305.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Jun Li for preparation of the initial batch of
cyclopropylglyoxylic acid, and David Leahy for helpful advice.
Racemic amine 8 was synthesized by chemists from the
Chemical Development Department of Bristol-Myers Squibb.
(27) Basch, J.; Franceschini, C.; Liu, S. W.; Chiang, S.-J. Genetically
Stable Plasmid Expressing PDH and FDH Enzymes. U.S. Patent Appl.
2010/0261236 A1 2010
(28) Johnston, R. M.; Chu, L. N.; Liu, M.; Goldberg, S. L.; Goswami,
A.; Patel, R. N. Enzyme Microb. Technol. 2009, 45, 484.
REFERENCES
(1) Gotor, V. Biorg. Med. Chem. 1999, 7, 2189.
(2) Gotor-Fernandez, V.; Gotor, V. Curr. Org. Chem. 2006, 10, 1125.
(3) Carr, R.; Alexeeva, M.; Dawson, M. J.; Gotor-Fernandez, V.;
Humphrey, C. E.; Turner, N. J. ChemBioChem 2005, 6, 637.
(4) Gastaldi, S.; Escoubet, S.; Vanthuyne, N.; Gil, G.; Bertrand, M. P.
Org. Lett. 2007, 9, 837.
■
(5) Stirling, D. I.; Zeitlin, A. L.; Matcham, G. W. Enantiomeric
Enrichment and Stereoselective Synthesis of Chiral Amines. U.S.
Patent 4950606, 1990.
(6) Stirling, D. I. In Chirality in Industry; Collins, A. N., Sheldrake, G.
N., Crosby, J., Eds.; John Wiley: Chichester, UK, 1992; pp 209−222.
(7) Hanson, R. L.; Davis, B. L.; Chen, S; Goldberg, S. L.; Parker, W.
L.; Tully, T. P.; Montana, M. A.; Patel, R. N. Adv. Synth. Catal. 2008,
350, 1367.
(8) Iwasaki, A.; Yamada, Y.; Kizaki, N.; Ikenaka, Y.; Hasegawa, J.
Appl. Microbiol. Biotechnol. 2006, 69, 499.
(9) Hohne, M.; Kuhl, S.; Robins, K.; Bornscheuer, U. T.
̈
̈
ChemBioChem 2008, 9, 363.
(10) Shin, J.-S.; Kim, B.-G. Biotechnol. Bioeng. 1999, 65, 206.
(11) Hartz, R. A.; Ahuja, V. T.; Zhuo, X.; Mattson, R. J.; Denhart, D.
J.; Deskus, J. A.; Vrudhula, V. M.; Pan, S.; Ditta, J. L.; Shu, Y.-Z.;
Grace, J. E.; Lentz, K. A.; Lelas, S.; Li, Y.-W.; Molski, T. F.;
469
dx.doi.org/10.1021/op2003562 | Org. Process Res. Dev. 2012, 16, 464−469