Intermediates for the synthesis of 4-substituted proline derivatives

Article abstract of DOI:10.1055/s-0029-1219372

A chemoenzymatic synthesis of a series of 2-hydroxy-5-nitro-4-substituted esters is described that uses two biotransformations in a single-pot process in which a kinetic resolution/reduction occurs. The products are valuable intermediates for the preparation of 4-substituted prolines and 5-substituted 3-hydroxypiperidinones as illustrated by the preparation of (2R,4R)-4- methylproline and (3S,5R)-3-hydroxy-5-methylpiperidinone. Georg Thieme Verlag Stuttgart · New York.

Full text of DOI:10.1055/s-0029-1219372

LETTER
539
Intermediates for the Synthesis of 4-Substituted Proline Derivatives
I
ntermediates
t
u
ubstitute
d
Pr
a
oline
D
eriv
r
atives t R. Crosby,a Richard B. Sessions,b Christine L. Willis*a
a
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, BS8 1TS, UK
Fax 44(117)9298611; E-mail: chris.willis@bristol.ac.uk
Department of Biochemistry, University of Bristol, University Walk, Bristol, BS8 1TD, UK
b
Received 9 January 2010
Dedicated to Professor Gerry Pattenden FRS, in celebration of his 70th birthday
mercially available Candida rugosa lipase (CRL) was
used to hydrolyse the ester to keto acid (R)-6, which is
required for recognition in the active site of a second
enzyme, lactate dehydrogenase from Bacillus stearother-
Abstract: A chemoenzymatic synthesis of a series of 2-hydroxy-5-
nitro-4-substituted esters is described that uses two biotransforma-
tions in a single-pot process in which a kinetic resolution/reduction
occurs. The products are valuable intermediates for the preparation
of 4-substituted prolines and 5-substituted 3-hydroxypiperidinones mophilus (BS-LDH), which catalyses reduction to 2S-hy-
as illustrated by the preparation of (2R,4R)-4-methylproline and
droxy acid 7. Esterification of 7 with diazomethane gave
8.5 The biotransformations were readily performed on a
(3S,5R)-3-hydroxy-5-methylpiperidinone.
Key words: biotransformations, molecular modeling, piperidin-
one, proline, resolution
gram scale in a single-pot process using a further enzyme,
formate dehydrogenase (FDH), in the presence of sodium
formate,6 to recycle the expensive co-factor NADH re-
quired for the reduction. (2S,4S)-Diastereomer 9 was pre-
4-Substituted prolines are components of a variety of nat- pared in 55% yield under the same conditions from (S)-2-
ural products and biologically active compounds. For ex- keto ester (S)-5.5
ample, trans-4-methyl-L-proline (1) has been isolated
from extracts of apple,1 and the ethyl and propyl ana-
logues 2 and 3 are components of antibiotic compounds
such as lincomycin (Figure 1).2 In addition to the naturally
occurring compounds, synthetic 4-substituted proline de-
rivatives have found application as pharmaceuticals, e.g.
Fosinopril (4), is a powerful ACE inhibitor.3 Often, these
heterocycles and their derivatives are prepared either from
commercially available 4-hydroxy L- and D-prolines or
from pyroglutamic acid.4 Our goal was to develop a flex-
ible approach to the enantioselective synthesis of 4-substi-
tuted prolines that could be readily adapted to incorporate
multiple isotopic labels for metabolic studies. Herein, we
describe the use of biotransformations to prepare 2-hy-
droxy esters that are valuable precursors of the target pro-
line derivatives.
O2N
O2N
CRL
O
O
CO2Me
CO2H
(R)-5
(R)-6
BS-LDH,
NADH, FDH, NaHCO2
O2N
O2N
OH
OH
CH2N2
CO2Me
CO2H
8 85% overall yield
7
O2N
O2N
O
OH
CO2Me
CO2Me
(S)-5
9 55% yield
R
Scheme 1
CO2H
R
N
O
Hydroxy esters 8 and 9 have potential value as building
blocks for the synthesis of nitrogen heterocycles, and our
studies began with an investigation into the use of
(2S,4R)-2-hydroxy-5-nitro ester 8 as a precursor to cis-4-
methyl-D-proline (11; Scheme 2). The proposed approach
involved conversion of the alcohol into a good leaving
group, followed by intramolecular attack by an amine
generated by reduction of the nitro group. In so doing, it
was important to ensure that the stereochemical integrity
at the C-2 position was not compromised during the cycli-
sation. Firstly, alcohol 8 was converted into tosylate 10
using tosyl chloride in pyridine (Scheme 2). Subsequent
hydrogenation over platinum(IV) oxide led to reduction
of the nitro functionality and spontaneous cyclisation to
give the corresponding pyrrolidine ester. The crude mate-
rial was hydrolysed using 6 M HCl in dioxane and, fol-
CO2H
O
P
N
Ph
O
O
H
4
1 R = Me
2 R = Et
3 R = Pr
O
4 R = c-Hex
Figure 1
Previously, we have shown that methyl (R)-4-methyl-5-
nitro-2-oxopentanoate [(R)-5] may be converted into
(2S,4R)-hydroxy ester 8 in 85% yield (Scheme 1). Com-
SYNLETT 2010, No. 4, pp 0539–0542
Advanced online publication: 10.02.2010
DOI: 10.1055/s-0029-1219372; Art ID: D01110ST
© Georg Thieme Verlag Stuttgart · New York
0
1.
0
3.
2
0
1
0

540
S. R. Crosby et al.
LETTER
lowing purification by ion-exchange chromatography, the bulky side-chains at C-4 would show an enhanced kinetic
target cis-4-methyl-D-proline (11) was isolated in 78% resolution in favour of the 4R-enantiomer compared to
1
overall yield from 2-hydroxy ester 8. The H- and 13C substrates with a methyl group at C-4.
NMR spectra of 11 revealed the presence of a single dia-
stereomer and the specific rotation {[a]D +82 (c 0.8,
H2O)}, was in accord with the literature value7 {[a]D
+85.2 (c 1.68, H2O)}, confirming that the stereochemical
integrity at C-2 had been maintained. When the nitro
group of (2S,4R)-2-hydroxy ester 8 was reduced without
prior derivatisation of the hydroxyl group, spontaneous
cyclisation occurred giving (3S,5R)-3-hydroxy-5-meth-
ylpiperidinone (12) in 98% yield.8
O2N
O2N
OH
OTs
CO2Me
TsCl
pyridine
CO2Me
8
10
87% yield
i, H2, PtO2
ii, HCl
H2, PtO2
OH
O
CO2H
N
N
H
H
11 90% yield
98% yield
12
Scheme 2
Our next goal was to apply the biotransformations to ac-
cess a range of synthetic intermediates that could be used
to generate prolines with different groups at C-4. Interest-
ingly, we have found previously that reduction of racemic
keto acid 6 with BS-LDH, followed by methylation, gave
a 3:1 mixture of diastereomeric products 8 and 9, rather
than the expected 1:1 ratio (Scheme 3).5 Thus, both an
asymmetric reduction and resolution had occurred.
Figure 2 Modelling of the active site of BS-LDH containing a) (R)-6
and b) (S)-6
Next, a series of 2-keto esters [with ethyl (19), propyl (20)
and phenyl (21) groups at C-4] was prepared as substrates
for the biotransformations (Scheme 4). Horner–Wads-
worth–Emmons chain extension of the parent aldehydes
gave the analogous E-unsaturated esters; subsequent con-
jugate addition of nitromethane, in the presence of DBU,
proceeded in good yield. The required one-carbon ho-
mologation of esters 13, 14 and 15 to the analogous 2-oxo
esters was achieved by applying the method described by
Wassermann and co-workers.10 Acid hydrolysis of the
ester, followed by conversion of the resultant acid into the
corresponding acid chloride and coupling with b-keto-
cyanomethylenetriphenylphosphorane in the presence of
bis(trimethylsilyl)acetamide (BSA) as a proton scaven-
ger, gave 16, 17 and 18. Ozonolysis of the crude b-keto-
phosphoranes gave the target 2-oxo esters in good yield.
This synthetic route was readily adapted for the incorpo-
ration of 13C-labels by use of the appropriately labelled al-
dehyde and/or phosphonate {(EtO)2PO13CH213CO2Et}.
Prior to undertaking further synthetic investigations, we
first turned to molecular modelling to gain an understand-
ing of this unusual kinetic resolution/reduction process.
Both 4-methyl-5-nitropentanoic acids (S)-6 and (R)-6
were docked separately into the active site of BS-LDH in
silico, with the carboxylate hydrogen bonded to Arg171 of
the protein. Inspection of the energy-minimised complex-
es indicated that the R-enantiomer was the preferred sub-
strate for BS-LDH.9 In this case, the nitro group interacts
with two amino acid residues, Thr246 and Asn101, to give
possible hydrogen bonding, and the methyl side-chain
points towards the flexible active-site loop (Figure 2a). In
contrast, the S-enantiomer has only one such interaction
with an amino acid residue (Thr246) and the methyl group
clashes with the NADH (Figure 2b). Hence, it would be
reasonable to expect that 2-keto-5-nitro acids with more
O2N
O2N
O2N
O
OH
i, BS-LDH
ii, CH2N2
OH
+
CO2H
CO2Me
CO2Me
8
6
9
3:1, 60% overall yield
Scheme 3
Synlett 2010, No. 4, 539–542 © Thieme Stuttgart · New York

LETTER
Intermediates for 4-Substituted Proline Derivatives
541
(EtO)2POCH2CO2Et
BuLi
the 4-methyl-2-oxo ester 5. 2-Oxo ester 20 contains the
more bulky 4-propyl side-chain and, whilst the lipase-
mediated hydrolysis proceeded smoothly in two hours, the
BS-LDH catalysed reduction was much slower than for
substrates with either a 4-methyl or 4-ethyl side-chain; af-
ter five days, a 6:1 mixture of 2-hydroxy esters 24 and 25
was isolated from 20 in only 15% yield. 4-Phenyl-2-oxo
ester 21 was also hydrolysed cleanly, but the resultant 2-
keto acid was a very poor substrate for BS-LDH and no re-
duction was apparent even after one week.
CO2Et
RCHO
R
R = Et, 74%
R = Pr, 90%
R = Ph, 92%
CH3NO2,
DBU
O2N
R
O2N
R
O
i, HCl
CN
CO2Me
ii, SOCl2
iii, Ph3PCHCN,
BSA, CH2Cl2
PPh3
To prepare the D-enantiomers of the target 5-nitro-2-hy-
droxy esters, commercially available D-lactate dehydro-
genase from Staphylococcus epidermidis (SE-LDH) was
used in place of BS-LDH. The series of 2-keto esters was
subjected to lipase hydrolysis and reduction catalysed by
SE-LDH, and a similar resolution/reduction was observed
for keto esters 5, 19 and 20 as with the BS-LDH but, this
time, in favour of the 4S-enantiomer (Scheme 6). Again,
no reduction of the 4-phenyl-2-keto acid was apparent.
13 R = Et, 87%
14 R = Pr, 78%
15 R = Ph, 78%
16 R = Et
17 R = Pr
18 R = Ph
O3, MeOH,
CH2Cl2
O2N
R
O
CO2Me
It was clear from these results that the BS- and SE-LDH
catalysed reductions of 5-nitro-2-keto acids bearing either
a propyl or phenyl side-chain, were not going to be syn-
thetically useful. To overcome this problem, the use of a
further oxido-reductase was investigated. D-Hydroxyiso-
caproate from Lactobacillus delbruekii bulgaricus (LB-
hicDH) has more favourable kinetic parameters than D-
LDHs for 2-keto acids with sterically demanding side-
chains.12 Holbrook and co-workers have used genetic en-
gineering of LB-hicH to prepare a series of mutant en-
zymes.13 They reported a greatly enhanced rate for the
reduction of 2-oxo-4-phenylbutanoic acid to the corre-
sponding (R)-2-hydroxy acid with the H205Q mutant (ex-
change of the His-205 for Gln) of LB-hicDH compared to
19 R = Et, 68% from 13
20 R = Pr, 64% from 14
21 R = Ph, 64% from 15
Scheme 4
Each of the 2-oxo esters was subjected to the one-pot li-
pase hydrolysis/BS-LDH catalysed reduction protocol,
and the products were methylated with TMS-diazo-
methane (Scheme 5).11 4-Ethyl-2-keto ester 19 gave a
50% yield of 2-hydroxy esters 22 and 23 in a 9:1 ratio (de-
termined by 1H NMR spectroscopy) after eight hours. As
predicted from the molecular modelling studies, the major
product was the diastereomer (2S,4R)-22; an enhanced ki-
netic resolution was seen to have occurred compared with
O2N
O2N
O2N
i, CRL
O
OH
OH
ii, BS-LDH
+
R
CO2Me
iii, TMS-CHN2
R
CO2Me
R
CO2Me
22 R = Et
24 R = Pr
19 R = Et
20 R = Pr
23 R = Et
25 R = Pr
9:1, 50% yield
6:1, 15% yield
Scheme 5
i, CRL
O2N
R
O2N
O2N
R
ii, SE-LDH or
LB-hicDH
O
OH
OH
+
CO2Me iii, TMS-CHN2
R
CO2Me
CO2Me
26 R = Me
SE-LDH
LB-hicDH
27 R = Me
1.8:1, 75% yield
1:1, 90% yield
5 R = Me
19 R = Et
20 R = Pr
28 R = Et
29 R = Et
SE-LDH
LB-hicDH
8:1, 52% yield
1:1, 83% yield
30 R = Pr
31 R = Pr
SE-LDH
LB-hicDH
5:1, 14% yield
1:1, 80% yield
32 R = Ph
LB-hicDH 2.5:1, 50% yield
21 R = Ph
33 R = Ph
Scheme 6
Synlett 2010, No. 4, 539–542 © Thieme Stuttgart · New York

542
S. R. Crosby et al.
LETTER
(3) Krapcho, J.; Turk, C.; Cushman, D. W.; Powell, J. R.;
LDHs; we found that this genetically engineered oxi-
doreductase has great potential value in synthesis due to
its broad substrate specificity and high catalytic activity.11
Hence, the lipase, LB-hicDH mutant catalysed hydrolysis
reduction of the series of 5-nitro-2-oxo esters was exam-
ined. Interestingly, when substrates bearing methyl, ethyl
or propyl side-chains (5, 19 and 20) were used, there was
no kinetic resolution and a 1:1 mixture of D-hydroxy es-
ters was isolated in good yields (Scheme 6). However,
with the 4-phenyl derivative 21, a 2.5:1 mixture of diaste-
reomers 32 and 33 was obtained. These results illustrate
the broad substrate specificity of the LB-hicDH mutant
and its potential value in organic synthesis.
DeForrest, J. M.; Spitzmiller, E. R.; Karanewsky, D.;
Duggan, M.; Rovnyak, G. J. Med. Chem. 1988, 31, 1148.
(4) See for example: (a) Murphy, A. C.; Mitova, M. I.; Blunt,
J. W.; Munro, M. H. G. J. Nat. Prod. 2008, 71, 806.
(b) Thottahill, J. F.; Moniot, J. L.; Mueller, R. H.; Wong,
M. K. Y.; Kissick, T. P. J. Org. Chem. 1986, 51, 3140.
(c) Moody, C. M.; Young, D. W. Tetrahedron Lett. 1994, 35,
7277. (d) Del Valle, J. R.; Goodman, M. J. Org. Chem.
2003, 68, 3923. (e) Heindl, C.; Hübner, H.; Gmeiner, P.
Tetrahedron: Asymmetry 2003, 14, 3153.
(5) Crosby, S. R.; Hateley, M. J.; Willis, C. L. Tetrahedron Lett.
2000, 41, 397.
(6) Shaked, Z.; Whitesides, G. M. J. Am. Chem. Soc. 1980, 102,
7104.
(7) Dalby, J. S.; Kenner, G. W.; Sheppard, R. C. J. Chem. Soc.
1962, 4387.
In conclusion, we have shown that (2R,4R)-4-methylpro-
line (11) and (3S,5R)-3-hydroxy-5-methylpiperidinone
(12) can be prepared in good yield from methyl (2S,4R)-4-
methyl-5-nitro-2-hydroxypentanoate (8). A chemoenzy-
matic synthesis of a series of 2-hydroxy-5-nitro-4-substi-
tuted esters is described that uses two biotransformations
in a single-pot process in which a kinetic resolution/reduc-
tion occurs. Molecular modelling has provided an insight
into the unusual transformation occurring in the active site
of BS-LDH. Whilst the kinetic resolution/reduction has
synthetic potential, it is more efficient to prepare single
enantiomers of the 5-nitro-2-oxo esters as substrates for
the biotransformations. The C-4 stereogenic centre of the
targets may be generated in excellent yield and stereocon-
trol through addition of acylated camphorsultams to ni-
troalkenes (Scheme 7).14
(8) Piperidinone 12: white solid; mp 149 °C (MeOH); [a]D
+10.6 (c 0.5, MeOH); IR (Nujol): 3336, 2929, 1755, 1290
cm–1; 1H NMR (300 MHz, D2O): d = 1.03 (d, J = 6.5 Hz, 3 H,
5-CH3), 1.53 (q, J = 12.6 Hz, 1 H, 4-Hax), 2.12–2.20 (m, 3 H,
OH, 5-H and 4-Heq), 3.22 (dd, J = 11.0, 8.5 Hz, 1 H, 6-HH),
3.31 (ddd, J = 11.0, 5.6, 2.0 Hz, 1 H, 6-HH), 4.26 (dd, J =
12.6, 5.9 Hz, 1 H, 3-H); 13C NMR (75 MHz, CD3OD): d =
19.1, 28.8, 39.2, 50.0, 68.5 (C-3), 175.6 (C-2); MS (CI): m/z
[MH]+ calcd for C9H12NO2: 130.0865; found: 130.0868.
Tosylate 10: yellow oil; [a]D –19.5 (c 1.0, CHCl3); IR (film):
2959, 1763, 1598, 1555, 1438, 1376 cm–1; 1H NMR (300
MHz, CDCl3): d = 1.05 (d, J = 7.0 Hz, 3 H, 4-CH3), 1.89–
1.96 (m, 2 H, 3-H2), 2.38 (s, 3 H, CH3), 3.66 (s, 3 H, OCH3),
4.26 (dd, J = 12.2, 7.3 Hz, 1 H, 4-HH), 4.33 (dd, J = 12.2, 7.3
Hz, 1 H, 4-HH), 4.96 (dd, J = 7.3, 6.0 Hz, 1 H, 2-H), 7.36 (d,
J = 8.0 Hz, 2 H, ArH), 7.83 (d, J = 8.0 Hz, 2 H, ArH); 13
C
NMR (75 MHz, CDCl3): d = 17.5, 21.6, 28.8, 35.5, 52.7,
74.9, 79.5, 128.0, 128.9, 129.8, 145.5, 168.5; MS (CI): m/z
[MH]+ calcd for C14H20NO7S: 346.0960; found: 346.0964.
(9) The crystal structure coordinates of a ternary complex of BS-
LDH containing NADH and the substrate analogue oxamate,
were taken from the protein data bank (1LDN). Oxamate
was replaced in the structure by the (4R)-2-keto acid (R)-5
and the conformation of the rest of the molecule was
adjusted to best dock in the remaining space in the active
site. This process was repeated for the S-enantiomer [(S)-5]
of the substrate. Both complexes were soaked in a 5Å layer
of water and energy-minimised using DISCOVER v 2.95
and the cvff force field in SGI challenge L.
NO2
O
O
NO
2, TiCl4
i,
R
N
R
N
ii, 20% NH4F (aq)
S
S
O2
O2
R = Me, 85%
R = i-Pr, 80%
R = Ph, 73%
Scheme 7
(10) Wassermann, H. H.; Ho, W.-B. J. Org. Chem. 1994, 59,
4364.
Acknowledgment
(11) For a general procedure for conducting the biotransforma-
tions, see: Sutherland, A.; Willis, C. L. J. Org. Chem. 1998,
63, 7764.
We are grateful to the EPSRC for funding (S.R.C.) and Ms Rebecca
Ward for the preparation of 2-keto ester 19.
(12) Alvarez, J. A.; Gelpi, J. L.; Johnsen, K.; Bernard, N.;
Delcour, J.; Clarke, A. R.; Holbrook, J. J.; Cortés, A. Eur. J.
Biochem. 1997, 244, 203.
(13) Bernard, N.; Johnsen, K.; Gelpi, J. L.; Alvarez, J. A.; Ferain,
T.; Garmyn, D.; Hols, P.; Cortés, A.; Clarke, A. R.;
Holbrook, J. J.; Delcour, J. Eur. J. Biochem. 1997, 244, 213.
(14) Clare, J. E.; Willis, C. L.; Yuen, J.; Lawrie, K. W. M.;
Charmant, J. P. H.; Kantacha, A. Tetrahedron Lett. 2003, 44,
8153.
References and Notes
(1) Hulme, A. C.; Arthington, W. Nature 1952, 170, 659.
(2) Hoeksema, H.; Bannister, B.; Birkenmeyer, R. D.; Kagan,
F.; Magerlein, B. J.; MacKellar, F. A.; Schroeder, W.;
Slomp, G.; Herr, R. R. J. Am. Chem. Soc. 1964, 86, 4223.
Synlett 2010, No. 4, 539–542 © Thieme Stuttgart · New York