Chemoenzymatic synthesis of N-Boc protected (2S,3R)-3-hydroxy-3- methylproline

Article abstract of DOI:10.1016/j.tetasy.2005.05.030

N-Boc protected 3-hydroxy-2-hydroxymethyl-3-methylpyrrolidine was kinetically resolved by acylation in an organic solvent catalyzed by Pseudomonas fluorescens lipase to give the corresponding acetate and residual diol in 14 E. Further selective oxidation afforded the title compound in high diastereo- and enantioselectivity.

Full text of DOI:10.1016/j.tetasy.2005.05.030

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 16 (2005) 2243–2247
Chemoenzymatic synthesis of N-Boc protected
(2S,3R)-3-hydroxy-3-methylproline
*
ˆ
Mansour Haddad and Marc Larcheveque
` ´
Laboratoire de Synthese Organique associe au CNRS, ENSCP, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France
Received 15 April 2005; accepted 24 May 2005
Available online 29June 2005
Abstract—N-Boc protected 3-hydroxy-2-hydroxymethyl-3-methylpyrrolidine was kinetically resolved by acylation in an organic
solvent catalyzed by Pseudomonas fluorescens lipase to give the corresponding acetate and residual diol in 14 E. Further selective
oxidation afforded the title compound in high diastereo- and enantioselectivity.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
synthesis,5 transformation of the chiral pool6 and chem-
ical resolution.7 However, all these methods require
many steps and give low overall yields and there is still
need of a concise method for large-scale production of
this unique component.
Conformationally constrained amino acids have
recently received much attention due to their ability to
act as conformational probes when incorporated into
peptides and peptidomimetics.1 They are also frequently
crucial structural features of biologically active and nat-
ural products.2 Polyoxypeptins A 1 and B 2 are two
recent examples that were isolated from Streptomyces
culture broth.3 Polyoxypeptin A has attracted a great
deal of attention due to its significant apoptosis-induc-
ing activity in human pancreatic carcinoma.4 Structur-
ally, this compound is a unique hexadepsipeptide
containing novel subunits, such as the acyl side-chain
moiety and the (2S,3R)-3-hydroxy-3-methylproline 3
among other non-proteinogenic a-amino acids (Fig. 1).
Herein we report a chemoenzymatic synthesis of the
N-Boc protected amino acid 4 based on the stereoselec-
tive addition of trimethylaluminium on a cyclic b-keto-
ester and a lipase resolution of a diol derived from the
resulting b-hydroxyester.
2. Results and discussion
In order to access a compound that could be easily
usable for further synthetic applications, we decided to
study the resolution of 3-hydroxy-3-methylproline by
enzymatic hydrolysis of a N-protected ester. For the
Various methods have been reported for the synthesis of
this unusual cyclic a-amino acid including asymmetric
Me
O
OH
H
N
N
N
O
R
HO
Me
O
O
O
N
N
H
O
HO
O
Me
N
N
O
Me
CO2H
Et
N
R
H
O
OH
N
i-Pr
OH
Me
i-Pr
OH
3 (R = H)
1 Polyoxypeptin A (R = OH)
2 Polyoxypeptin B (R = H)
4 (R = Boc)
Figure 1.
*
Corresponding author. Fax: +33 (0)1 43 25 7975; e-mail: marc-larcheveque@enscp.fr
0957-4166/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2005.05.030

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ˆ
M. Haddad, M. Larcheveque / Tetrahedron: Asymmetry 16 (2005) 2243–2247
2244
protecting group, Boc was selected on account of its
great utility in peptide synthesis. Moreover, this choice
was also dictated by the successful resolutions recently
reported on structurally related compounds, such as
pyrrolidines,8 proline derivatives9 or cyclic b-amino
acids.10
function of the hydroxyproline rac-6, a result possibly
due to the steric hindrance of the substrate. Attempts
of enzymatic N-acylation of the free amino ester also
proved unsuccessful.
Wanting to keep the Boc, which is a very useful protect-
ing group for the synthesis of complex peptides, we
modified our strategy. In spite of the difficulties fre-
quently observed during the resolution of chiral primary
alcohols,14 we decided to study the resolution of diol 9
derived from ester 6. This was easily obtained in good
yield by LiAlH4 reduction of the ester in THF.
The required protected hydroxy proline was first syn-
thesized by Dieckmann cyclization of the ketoester 5
according the method described by Williams (Fig.
2).11 Although this cyclization mainly afforded the
required (2R*,3S*)-diastereomer rac-6 (rac-6/rac-7: 7/3),
the yield strongly decreased by the concomitant forma-
tion of the b-diketone resulting from the formation and
cyclization of the enolate derived from the keto moiety.
Furthermore, the two diastereomers were difficult to
separate.
In our first attempt we used Candida antarctica lipase B
(CAL B), usually an efficient enzyme for such reactions.
Although the reaction rate was relatively high, the selec-
tivity remained very low and the acylated product was
obtained in an enantiomeric excess of only 9% at 50%
conversion (Table 1, entry 4). We screened other lipases,
but some proved totally inactive (Table 1, entries 1–3).
However, among the fifteen lipases or esterases exam-
ined, we succeeded in finding three lipases that showed
moderate activity (entries 6, 11 and 12). The best result
was obtained with Pseudomonas fluorescens lipase
(Amano AK) with which a 42% conversion of the race-
mic diol was obtained at 35 ꢁC after 48 h to give the cor-
responding acetate in 52% ee. The enantiomeric excesses
were measured by GC with a chiral stationary phase di-
rectly for the acetate or after monoacylation with acetic
or propionic anhydride for the residual diol. In order to
determine the absolute configuration, the residual diol
was oxidized into the corresponding hydroxy acid 4 or
ent-4 and esterified. Comparison of the rotatory power
of this ester with the one previously reported for the
same compound prepared by asymmetric synthesis5c
allowed us to assign a (2R,3S)-configuration to this ester,
thus a (2S,3S)-configuration was given to the residual
diol isolated after resolution (Scheme 2). Alternatively,
the N-Boc protected hydroxyacid was deprotected in
the presence of trifluoroacetic acid to give the known
3-hydroxy-3-methyl proline.
HO
HO
O
Me
Me
N
CO2Et
CO2Et
CO2Et
N
N
Boc
Boc
Boc
rac-7
5
rac-6
Figure 2.
We then decided to develop an alternative synthesis to
access the diastereomerically pure hydroxy ester 6. We
thought that by starting from the known b-ketoester
8,12 it would be possible to gain access to the required
tertiary alcohol by stereoselective addition of an organo-
metallic compound. Taking advantage of a previous
study related to the addition of organometallic com-
pounds to b-ketoesters,13 various methylating reagents
were reacted with 8. Methylmanganese chloride and tri-
chloromethyltitanium were totally inefficient, affording
complex mixtures. In contrast, the use of trimethylalu-
minium at room temperature allowed access to the
(2R*,3S*)-ester 6 as a single isomer in a satisfactory
yield (Scheme 1).
HO
HO
N
HO
N
HO
N
O
Me
Me
Me
Me
Me3Al
lipase
+
CO2Et
CO2Et
N
N
OH
OH
OAc
Boc
Boc
Boc
Boc
Boc
8
rac-6
(2S,3S)-9
(2R,3R)-10
rac-9
Scheme 2.
Scheme 1.
Amongst all the types of enzyme-catalyzed reactions,
hydrolytic transformations involving the cleavage of
the amide or ester bonds with proteases, esterases or lip-
ases were the easiest to perform. In the case of the amino
acids, the presence of two different functions offers the
possibility to use different strategies such as hydrolysis
of esters, acylation of amines or transesterification. Ini-
tially, we attempted to resolve compound 6 by hydroly-
sis of the ester moiety. Although successful results were
reported on structurally related compounds,8–10 we were
unable to find an enzyme able to hydrolyze the ester
For the target molecule, which required a (2S,3R)-con-
figuration, it was necessary to enhance the selectivity
of the resolution in order to isolate the product in an
acceptable enantiomeric excess. Initially we attempted
to modify the acylating reagent. The use of isopropenyl-
acetate, which avoids the formation of acetaldehyde
and frequently enhances the enantiomeric ratio was
unsuccessful (entries 9and 15). Unsatisfactory results
were also obtained with acetic anhydride or mixed i-prop-
yloxycarbonic propionic anhydride (entry 16).16

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ˆ
M. Haddad, M. Larcheveque / Tetrahedron: Asymmetry 16 (2005) 2243–2247
2245
Table 1. Enzyme-catalyzed acylation of diol 9a
Entry
Enzyme
Time (h)
Temp (ꢁC)
Conv (%)
Residual diol: ee (%)
Acetate (ee%)
Eb
1
M. miehie
48
48
48
26
48
26
48
40
48
48
90
48
43
96
48
48
30
30
30
35
30
35
20
25
30
30
35
35
25
20
40
30
0
0
—
—
—
—
—
—
—
—
2
B. cepacia
3
M. javanicus
CAL B
R. oryzae
0
—
4
50
4
99 1.2
—
14
5
70
52
52
33
—
4
—
6
C. cylindracea
C. cylindracea
C. cylindraceac
C. cylindracead
PPL
28
23
34
0
3.6
7
16
15
3.7
3.6
8
9
—
—
—
10
11
12
13
14
15
16
41
29
42
24
35
0
3
20
R. arrhizus
P. fluorescens
P. fluorescens
P. fluorescens
P. fluorescensd
P. fluorescense
50
4.6
74
80
—
45
3.6
3952
15
7.8
14.3
—
47
—
25
20
3.2
a All reactions were carried out by stirring a mixture of substrate (1 mmol) vinylacetate (10 mmol) and catalyst (200 mg) in tert-butylmethyl ether.
b E = ln[eep(1 ꢀ ees)]/(eep + ees)/ln[eep(1 + ees)]/(eep + ees).15
c Acylating agent: Ac2O.
d Acylating agent: isopropenyl acetate.
e Acylating agent: mixed i-propyloxycarbonic propionic anhydride.16
Mixed anhydrides prepared from a longer-chain acid,
such as valeric acid, were unreactive. Changing the sol-
vent also had little effect. Finally, it was the modification
of the temperature, which allowed us to increase the
selectivity; by operating at 20 ꢁC, we succeeded in
obtaining the required (2R,3R)-acetate 10 in a 80% ee
at 35% conversion (entry 14).
lyzed by GC (Chrompack CP-Sil 8CB, 50 m capillary
column) or by TLC (Merck silica gel 60F 254). NMR
spectra were recorded on a Bruker AC at 300 MHz
1
for H and 75.5 MHz for 13C. CDCl3 was used as sol-
vent with TMS as the internal standard. IR spectra were
recorded on a spectrophotometer FT Nicolet 210. Mass
spectra were recorded on a Ribermag R10-10C instru-
ment at 70 eV ionizing voltage; ammonia was used for
chemical ionization. Optical rotations were measured
on a Jasco P-1010 polarimeter. The enantiomeric
excesses were measured on a 25 m Chirasil-DEX CB
column (Chrompack-Varian). Rhizopus arrhizus and
Candida cylindracea (as immobilized in Sol–Gel
AK form) were purchased from Fluka and Pseudomo-
nas fluorescens lipase (Amano AK) from Sigma–
Aldrich.
This acetate was then hydrolyzed to afford the corre-
sponding (2R,3R)-diol 9, which was oxidized into acid
4. The use of Jonesꢀ reagent at low temperature was
not very convenient as it gave a low yield (40%) due
to a partial elimination of the Boc protecting group.
Our best result was obtained when using sodium perio-
date in the presence of a catalytic amount of ruthenium
salt.17 The required acid 4 was then isolated in 85% yield
and 78% ee. The enantiomeric excess was successfully
increased to 99% by recrystallization.
4.2. Ethyl 1-(tert-butoxycarbonyl)-3-oxopyrrolidine-
2-carboxylate 8
3. Conclusion
Under argon, a solution of N-Boc protected 3-(ethoxy-
carbonylmethyl-amino)-propionic acid ethyl ester12c
(12.12 g, 0.04 mol) in toluene (20 mL) was slowly added
to a vigorously stirred suspension of potassium tert-
butoxide (3.36 g, 0.08 mol, 2 equiv) in toluene
(160 mL) cooled to ꢀ5 ꢁC by means of an external
ice–NaCl bath. The reaction mixture was stirred at
ꢀ5 ꢁC for 4 h and then quenched with acetic acid
(4.5 mL, in one portion), followed by a cold solution
of NaH2PO4ÆH2O (22 g) in H2O (220 mL). The layers
were separated and the aqueous layer extracted with
CHCl3 (2 · 100 mL). The combined organic extracts
were washed with pH 7 phosphate buffer (2 · 40 mL),
dried over anhydrous MgSO4 and concentrated in vacuo
to give an oil. The crude product was purified by flash
chromatography on silica gel (cyclohexane/ethyl acetate:
7/3) to afford the compound 8 (6.17 g, 60%) as a colour-
less oil with spectral data identical to those previously
reported.12b
In summary, diol 9 derived from N-protected diastereo-
merically pure 3-hydroxy-3-methylproline was conve-
niently resolved by a vinyl acetate transesterification in
an organic solvent catalyzed by Pseudomonas fluorescens
lipase. Further ruthenium catalyzed oxidation allowed
access to the N-Boc protected (2S,3R)-3-hydroxy-3-
methylproline in high enantiomeric excess.
4. Experimental
4.1. General methods
Products were purified either by distillation or by med-
ium pressure liquid chromatography on a Jobin-Yvon
Modulprep (Kieselgel 60H Merck) or by flash chroma-
tography (Kieselgel 60 Merck: 230–400 Mesh) and ana-

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ˆ
M. Haddad, M. Larcheveque / Tetrahedron: Asymmetry 16 (2005) 2243–2247
2246
4.3. Ethyl (2R*,3S*)-1-(tert-butoxycarbonyl)-3-hydroxy-
3-methylpyrrolidine-2-carboxylate 6
4.5.1. (2S,3S)-1-(tert-Butoxycarbonyl)-3-hydroxy-2-hydr-
20
D
oxymethyl-3-methylpyrrolidine. ½aŠ ¼ þ15.1 (c 3.73,
MeOH), ee: 47%.
Under argon, 2 M trimethylaluminium in toluene
(20 mL, 0.04 mol, 2 equiv) was slowly added to a solu-
tion of the N-Boc protected ketoester 8 (5.14 g,
0.02 mol) in toluene (100 mL) cooled to 0 ꢁC. The solu-
tion was stirred for 3 h at 0 ꢁC and quenched with ice
cold aqueous KHSO4. After extraction with diethyl
ether, the organic layer was dried over MgSO4 and evap-
orated to give a mixture of the starting compound and
the stereochemically pure hydroxyester 6 (4/6). These
compounds were separated on silica gel (cyclohexane/
ethyl acetate: 6/4) to afford the rac-6 ester as a colourless
oil (2.4 g, conv: 60%, yield: 73%). IR (KBr) cmꢀ1: 3450,
2973, 1745, 1706, 1685, 1410, 1190, 1167; 1H NMR
(300 MHz, CDCl3): two conformers at 25 ꢁC; d 1.26 (t,
3H, J = 7 Hz), 1.39and 1.47 (s, 9H); 1.44 (s, 3H); 1.85
(dt, 1H, J = 6.2 and 12.5 Hz); 2.0–2.1 (m, 1H), 3.4–3.5
(m, 1H), 3.55–3.65 (m, 1H), 4.00 and 4.07 (s, 1H); 4.20
(q, 2H, J = 7 Hz); 13C NMR (75.5 MHz, CDCl3): d
14.3, 26.7, 28.2 and 28.3, 38.2 and 38.8, 44.3 and 44.8,
61.0 and 61.1, 68.2 and 68.8, 77.8 and 78.9, 80.0 and
4.5.2. (2R,3R)-2-Acetoxymethyl-1-(tert-butoxycarbonyl)-
20
3-hydroxy-3-methylpyrrolidine 10. ½aŠ ¼ ꢀ18.5 (c 2.0,
D
CH2Cl2), ee: 80%; IR (KBr) cmꢀ1: 3441, 2975, 1743,
1697, 1673, 1394, 1243, 1166; 1H NMR (400 MHz,
CDCl3): d 1.38 (s, 3H), 1.47 (s, 9H); 1.75–1.85 (m,
1H), 1.9–2.0 (m, 1H), 2.05 (s, 3H), 3.41 (dd, 2H,
J = 6.2 and 7.8 Hz), 3.65 (br s, 1H), 4.32 (dd, 1H,
J = 3.5 and 11.3 Hz), 4.46 (dd, 1H, J = 6.0 and
11.3 Hz); 13C NMR (75.5 MHz, CDCl3): d 21.0, 27.8,
28.4, 37.9, 43.9, 62.4, 63.5, 77.3, 80.1, 154.6, 171.0; MS
(CI NH3) m/z 274 (M+1, 73%), 235 (14%), 218
(100%), 174 (33%). Anal. Calcd for C13H23NO5: C
57.13, H 8.48, N 5.12; Found: C 57.21 H 8.55 N 5.07.
4.5.3. (2S,3R)-1-(tert-Butoxycarbonyl)-2-carboxy-3-hydr-
oxy-3-methylpyrrolidine 4. A solution of the acetate
(1.36 g, 5 mmol) in methanol (2 mL) was added to 3 M
aqueous LiOH (10 mL) and stirred for 2 h at room
temperature. The solution was extracted with diethyl
ether and dried over MgSO4. After removing the sol-
vent, the crude diol was isolated as a solid. This was then
oxidized at 0 ꢁC under argon with NaIO4 (4.8 g,
20 mmol) and RuCl3ÆnH2O (26 mg, 0.125 mmol, 2.5%)
in a well-stirred mixture 1/1/1.5 CCl4, CH3CN, H2O
(105 mL). After 4 h, the mixture was extracted with
Et2O (6 · 30 mL) and the organic phase dried over
MgSO4. After removing the solvent, the crude solid
was dissolved in THF and refluxed for half an hour in
the presence of carbon black. The hot solution was fil-
tered on celite to give after removing the solvent a solid
80.2, 153.9and 154.3, 170.4 and 170.6; MS (CI NH )
3
m/z 291 (M+NHþ4 , 32%), 274 (M+H+, 53%), 235
(100%), 218 (18%), 174 (19%).
4.4. (2S*,3S*)-1-(tert-Butoxycarbonyl)-3-hydroxy-2-
hydroxymethyl-3-methylpyrrolidine 9
A solution of ester 6 (1.91 g, 7 mmol) in THF (15 mL)
was added to a suspension of LiAlH4 (319mg,
8.4 mmol, 1.2 equiv) in THF (25 mL) cooled to 0 ꢁC.
After stirring for 2 h, water (319 lL) was added fol-
lowed by 15% NaOH (319 lL) and water again
(957 lL). The mixture was stirred for 2 h and the sus-
pension was filtered. The residual solid was stirred for
15 min with THF (25 mL), filtered again and washed
twice with THF (20 mL). The combined organic phases
were evaporated in vacuo to afford a solid (1.21 g, 80%
yield), which was purified by crystallization in hexane/
i-Prop2O: 9/1. Mp: 90–91 ꢁC; IR (KBr) cmꢀ1: 3408,
2973, 1693, 1666, 1455, 1167; 1H NMR (300 MHz,
CDCl3): d 1.43 (s, 9H); 1.79 (m, 3H), 2.0 (m, 1H),
2.78 (dd, 1H, J = 9.8 and 16.2 Hz), 2.85 (m, 1H), 3.30
(m, 2H), 3.64 (s, 3H), 4.10 (br s, 1H); 13C NMR
(75.5 MHz, CDCl3): d 26.3, 28.4, 38.7, 45.0, 62.3,
67.3, 77.8, 80.1, 156.0; MS (CI NH3) m/z 232 (M+1,
100%), 193 (22%), 176 (19%). Anal. Calcd for
C11H21NO4: C 57.12, H 9.15, N 6.06; Found: C
57.25, H 9.21, N 6.01.
(85% yield), which was recrystallized. White solid: mp
20
D
163–164 ꢁC (i-Prop2O/THF: 3/1); ½aŠ ¼ ꢀ19.5 (c 1.34,
MeOH) ee: 99% (as methyl ester); IR (KBr) cmꢀ1
:
3397, 3014, 1696, 1677, 1414, 1230, 1194; 1H NMR
(300 MHz, CD3COCD3): two conformers at 25 ꢁC; d
1.16 and 1.20 (s, 9H), 1.26 (s, 3H), 1.60–1.65 (m, 1H),
1.80–1.85 (m, 1H), 2.80 (m, 1H), 3.20–3.25 (m, 1H),
3.35–3.40 (m, 1H), 3.76 and 3.77 (s, 1H); 13C NMR
(75.5 MHz, CD3COCD3): d 27.3, 28.5, 38.8 and 39.3,
45.1 and 45.6, 69.2 and 69.7, 78.0 and 79.0, 79.8, 154.5
and 154.9, 171.9; MS (CI NH3) m/z 263 (73%), 246
(21%), 207 (10%), 139(14%), 122 (60%), 88 (100%).
Anal. Calcd for C11H19NO5: C 53.87, H 7.81, N 5.71;
Found: C 53.95, H 7.83, N 5.76.
4.5.4. (2S,3R)-3-Hydroxy-3-methylproline. Acid
4
(245 mg, 1 mmol) was stirred with trifluoroacetic acid
(4 mL) for 2 h at 0 ꢁC. The solvent was then removed
in vacuo and the residue taken up into water (2 mL).
This aqueous solution was passed through a column
of acidic ion-exchange resin (Dowex 50 W). The acid
was eluted with water and then with a 2 M solution
of NH4OH. After lyophilization, pure acid was
obtained as a white solid: mp 198–200 ꢁC (dec);
4.5. Lipase-catalyzed resolution
A solution of diol 9 (1.15 g, 5 mmol) in tert-butylmethyl
ether (55 mL) was stirred at 150 rpm with Pseudomonas
fluorescens (Amano AK) lipase (1 g) and vinyl acetate
(10.0 g, 0.1 mol, 20 equiv) at 20 ꢁC. After four days,
the suspension was filtered through a small pad of Celite
to remove the enzyme and the residual mixture chro-
matographed on silica gel (cyclohexane/ethyl acetate:
4/6) to give acetate 10 (0.48 g, 35%) and the residual diol
(0.73 g, 63%).
20
D
20
D
½aŠ ¼ ꢀ38.6 (c 1.49, H2O) lit.4a ½aŠ ¼ ꢀ41.0 (c 0.40,
1
H2O); H NMR (300 MHz, D2O): d 1.51 (s, 3H), 2.0–
2.1 (m, 2H), 3.3–3.4 (m, 1H), 3.45 (m, 1H,) 3.77 (s,
1H); 13C NMR (75.5 MHz, D2O): d 23.4, 39.0, 42.9,
69.3, 78.0, 170.3.