The 'unexpected' epimerization on bicyclic thiazolidine γ-lactam scaffolds

Article abstract of DOI:10.1016/j.tet.2008.02.070

Preliminary findings related to the base induced cyclization and spontaneous epimerization leading to stereochemically defined bicyclic thiazolidine γ-lactam systems are reported.

Full text of DOI:10.1016/j.tet.2008.02.070

Available online at www.sciencedirect.com
Tetrahedron 64 (2008) 4363e4369
www.elsevier.com/locate/tet
The ‘unexpected’ epimerization on bicyclic thiazolidine
g-lactam scaffolds
a
a
Ruzhang Liu a, Yang Wu a, Qiang Li a, Wensheng Liao a, , Shu-Hui Chen , Ge Li ,
*
Juan M. Betancort b, David T. Winn b, David A. Campbell b
a WuXi PharmaTech Co., Ltd, No. 1 Building, 288 FuTe ZhongLu, WaiGaoQiao Free Trade Zone, Shanghai, 200131, PR China
b Phenomix Corporation, 5871 Oberlin Drive, Suite 200, San Diego, CA 92121, United States
Received 28 November 2007; received in revised form 4 February 2008; accepted 22 February 2008
Available online 4 March 2008
Abstract
Preliminary findings related to the base induced cyclization and spontaneous epimerization leading to stereochemically defined bicyclic
thiazolidine g-lactam systems are reported.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
in Figure 1.9e11 The potential use of the conformationally
restricted PheePro surrogate 2b in the design of HIV protease
inhibitors has also been postulated by Baldwin et al. in 1992.12
Prompted by their diverse biological activities and unique
structural features (rigidity imposed by the bicyclic framework
and complexity due to the presence of three stereo centers),13
we became interested in the synthesis of the four stereo-
isomers of bicyclic thiazolidine g-lactams 9ae9d as shown
in Scheme 1 via a base induced final lactam bond formation
reaction. We were also interested in optimizing conditions
for the production of 9a because of its unique structure (pos-
sessing the same relative stereochemistry to that seen in 2b
as shown in Fig. 1).
In this paper, we describe some of the ‘unexpected’ base in-
duced epimerization reactions that we encountered while syn-
thesizing these systems and how we use them to our advantage
for the synthesis of the desired target compounds. In addition,
we present herein some preliminary molecular modeling cal-
culations performed on these bicyclic g-lactam frameworks.
The past decade has witnessed a great deal of research ac-
tivities directed toward the design, synthesis, and evaluation of
peptide and peptidomimetic based therapeutic agents.1 These
efforts have culminated in the discovery of many important
peptidomimetics endowed with desirable biological activity
and toxicological profiles such as renin inhibitors (hyper-
tension),2 HIV protease inhibitors (AIDS),3 rhino virus 3C
protease inhibitors (common cold),4 HCV NS3 protease inhib-
itors (HCV infections),5 beta-secretase inhibitors (Alzheimer’s
disease), and cathepsin K inhibitors (osteoclasts).6,7
Despite the impressive successes achieved by the agents
described above, peptide based therapeutic agents suffer from
poor ADME properties such as poor absorption and extensive
metabolism resulting in low oral bioavailability. This is due in
part to the presence of numerous amide bonds and an overall
lack of structural rigidity.8 To circumvent these deficiencies,
many approaches have been investigated including incorpora-
tion of conformationally constrained peptidomimetics such
as substituted bicyclic thiazolidine lactams. This motif can be
found in melanocyte-stimulating hormone release-inhibiting
factor analogs 2a and 2b and angiotensin II analog 4 as shown
2. Result and discussion
2.1. Synthesis of the key cyclization precursors
The synthetic route used for the preparation of the target
molecules 9a through 9d is shown in Scheme 1. The requisite
* Corresponding author. Tel.: þ86 21 50363913; fax: þ86 21 50463718.
E-mail address: liao_wensheng@pharmatechs.com (W. Liao).
0040-4020/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.02.070

4364
R. Liu et al. / Tetrahedron 64 (2008) 4363e4369
H2N
HN
NH
HO
H
N
H
N
N
HN
O
O
O
O
H
N
H
N
H
N
N
O
O
H2N
HO
N
OH
N
H
NH
O
H2N
H
O
NH
O
O
O
1
O
3
peptide backbone rigidification
peptide backbone rigidification
H
N
H2N
HN
NH
HO
N
O
H
S
O
R
H
N
N
O
N
4
N
N
H
OH
HN
O
O
3
H
N
O
O
5
O
2
H2N
HO
O
H2N
N
H
N
6
O
S
1
8
H
7
N
H
O
H
2a R = i-Pr
2b R = Ph
O
4
Figure 1. Representative bicyclic thiazolidine lactam bearing molecules.
CO2Me
Table 1
COOR
BocHN
NH2
COOMe
O
H2N
Bn
HN
COOMe
BocHN
Bn
i, ii
v or vi
iii, iv
Bn
Bn
COOMe
S
6
5
7
R = H, or 8 R = Me
L-Phenylalanine
O
CO2Me
O
CO2Me
O
CO2Me
O
CO2Me
4
N
3
5
Bn
Bn
6
BocHN
Bn
Bn
N
N
N
2
+
+
+
+
8
BocHN
BocHN
BocHN
S
S
S
S
1
7
H
9c
H
9d
H
9a
H
9b
DBU/Tol.
CO2Me
X
DBU/Tol.
CO
X DBU/Tol.
CO2Me
N
DBU/Tol.
CO2Me
O
O
2Me
O
O
Bn
Bn
Bn
Bn
N
H
N
N
H
+
+
BocHN
BocHN
BocHN
BocHN
S
S
S
S
H
H
9c-r (enantiomer of 9a)
9d-r (enantiomer of 9b)
9a-r
9b-r
Scheme 1. Synthesis of the bicyclic thiazolidine framework. Conditions: (i) PhCHO, MgSO4; (ii) KOtBu, allyl bromide, then 2 N HCl, 88% (two-step); (iii) Boc2O;
(iv) O3, PPh3, 61% (two-step); (v) LiOH, then aq NaHCO3, EtOH, L-CyseOMe, 72%; (vi) NaOAc, AcOH, aq EtOH, L-CyseOMe, 84%.
aldehyde precursor 6 was prepared from L-phenylalanine via
a four-step sequence consisting of imine formation, C-allyl-
ation, N-Boc protection, and ozonolysis of the resulting olefin
according to a modified protocol from Takahashi et al.12e15
Treatment of the aldehyde 6 with LiOH, followed by L-cyste-
ine in the presence of sodium bicarbonate in aqueous ethanol
afforded the methyl ester 7. Alternatively, reacting 6 with
L-cysteine in conjunction with NaOAc and acetic acid yielded
the diester 8.
ratio in a total yield of 82%. While isomers 9a (17%) and
9d (26%) were easily isolated after silica gel column chroma-
tography, isomers 9b (28%) and 9c (11%) were obtained via
preparative HPLC separation. The structures of 9a through
9d were confirmed on the basis of extensive NOE experiments
(see Fig. 2 for details).
In order to increase the yield of isomer 9a, additional cou-
pling conditions were evaluated and the cyclization results are
listed in Table 1. The use of DCC/DMAP (entry B) or HOBt/
EDCI/Et3N (entry D) produced similar ratios and yields for 9a
(16e19%). To our disappointment, a lower yield of 9a (up to
9%) was obtained when ClCO2Bu-i/Et3N (entry C) or PyBOP/
Et3N (entry E) or TBTU/Et3N (entry F) was used as the
coupling reagents.
2.2. Construction of the bicyclic ring system via 7 (acid)
or 8 (ester)
The conversion of the acid 7 to the target molecules 9a
through 9d was carried out under various conditions as shown
in Table 1. According to the procedure of Khalil and Johnson,9
Mukaiyama reagent (entry A) was used to affect cyclization
reaction and provided four respective isomers with similar
In light of the poor yields obtained under basic coupling
conditions, we resorted to the use of TsOH as the coupling re-
agent (entry G). To our satisfaction, a much cleaner product
distribution (only 9a and 9b) and higher yield of 9a (38%)

R. Liu et al. / Tetrahedron 64 (2008) 4363e4369
4365
Table 1
Bicyclic g-lactam isomer distribution under different coupling conditions
Entry
Substrate
Conditions
Isolated product ratio (%)
Total yield (%)
A
B
C
D
E
F
G
H
I
7
7
7
7
7
7
7
8
8
Mukaiyama reagent/TEA/DCMa
DCC/cat. DMAP/DCMc
ClCO2Bu-i/Et3N/DCMc
HOBt/EDCI/Et3N/DCMc
PyBOP/Et3N/DCMc
TBTU/Et3N/DCMc
cat. TsOH/Tol.a
17 (9a); 28 (9b); 11 (9c); 26 (9d)b
19 (9a); 28 (9b); 11 (9c); 19 (9d)d
9 (9a); 4 (9b); 12 (9c); 13 (9d)d
16 (9a); 20 (9b); 18 (9c); 17 (9d)d
1 (9a); 6 (9b); 12 (9c); 12 (9d)d
3 (9a); 11 (9b); 6 (9c); 15 (9d)d
38 (9a); 35 (9b)
82
77
38
71
31
35
73
60
69
cat. TsOH/Tol.a
49 (9a); 11 (9b)
33 (9a); 24 (9b); 9 (9c-r); 3 (9d-r)d
DBU/Tol.e
a
Refluxed overnight.
The yield after separation.
Stirred overnight.
b
c
d
e
The yields of 9b, 9c, 9c-r, and 9d-r were calculated on the basis of chiral HPLC trace.
Base (3 equiv) in refluxing toluene for 2 h, the reaction was conducted in anoxic conditions.
9d
9a
9b
9c
Me
O
S
Me
Me
O
Me
O
O
S
O
O
S
O
O
S
O
O
O
O
H3
O
H3
N
O
H3
H2b
H3
O
N
O
N
H2b
N
H2b
2b
Me
3C
H
N
H2a
O
Me3C
N
H2a
Me3C
N
H7b
H
N
Me3C
O
H2a
H2a
O
H
O
H7b
H
H
H8
H7b
H7a
H8
H7b
H7a
H8
H8
H7a
H7a
Figure 2. Graphical summary of the NOEs observed for the bicyclic thiazolidine lactams 9a through 9d.
were observed. Encouraged by this success, the methyl ester 8
was treated with TsOH in toluene under reflux for 2 h. Once
again, the desired isomer 9a was obtained in 49% yield along
with 11% of its epimer 9b (entry H).
We further reasoned that isomers 9c-r and 9d-r might arise
from base induced C3 epimerization of 9c and 9d, respec-
tively. Consistent with the results revealed by chiral HPLC
analysis, we discovered that the optical rotation value ([a]D)
observed for ‘9a’ was ꢀ76.4 (c 1.0, MeOH), which was lower
than that measured for the optically pure 9a ([a]D ꢀ102.8
(c 1.0, MeOH)). The same trend was also found with ‘9b’
([a]D ꢀ131.3 (c 1.0, MeOH)) and the optically pure sample
9b ([a]D ꢀ170.7 (c 1.0, MeOH)).
2.3. In situ C3 epimerization
Prompted by the improved yield of 9a obtained when the
cyclization reaction was performed under elevated tem-
perature, we decided to carry out the g-lactam ring formation
reaction using DBU in refluxing toluene. However, in this
event, two peaks corresponding to isomers ‘9a’ (42%) and
‘9b’ (27%) were detected by HPLC analysis, respectively.
Upon careful inspection by chiral HPLC analysis, compound
‘9a’ was separated into two peaks with identical mass units.
Likewise, compound ‘9b’ was also a mixture of two com-
pounds with equal mass.
After extensive analysis, we concluded that ‘9a’ was indeed
a mixture of 9a with its enantiomer 9c-r in a 3.6:1 ratio favor-
ing the former. Similarly, compound ‘9b’ was found to be
a mixture of 9b and 9d-r in a ratio of 8:1 favoring 9b. Further-
more, two C-6 isopropyl analogs of ‘9a’ and ‘9b’, namely
‘11a’ (57%) and ‘11b’ (40%) were obtained using condition
I in Table 1. Both of these bicyclo-products were subjected
to crystallization and solid state structure analysis. As shown
below, we found that the crystal obtained from ‘11b’ was a sin-
gle compound 11b. In contrast, a pair of enantiomers (11a and
11c-r) was obtained from ‘11a’. Evidently, the corresponding
stereochemistry (at C-3, C-6, and C-8) found with these
crystals was in good agreement with that deduced from
NOE experiments for 9a and 9b (Fig. 3).
To verify the in situ C3 epimerization event, isomers 9c and
9d were treated with DBU in refluxing toluene. Indeed, these
two isomers were epimerized to their C-3 isomers 9c-r and
9d-r in 93 and 97% yield, respectively. When compounds
9a and 9b were treated under identical reaction conditions,
no C3 epimerization took place. The difference in behavior
observed for the four isomers (9ae9d) under basic treatment
was later rationalized by computational analysis as outlined
in the following section.
Intrigued by this interesting in situ base induced C3 epime-
rization event, we further investigated the aminolysis of 9ae9d
under elevated temperature.9 We discovered that C3 epimeri-
zation did occur and gave rise to the corresponding amide deriv-
atives 10c-r and 10d-r, which were confirmed by 1H NMR and
chiral HPLC analysis, to be the enantiomers of 10a and 10b
obtained via aminolysis of 9a and 9b. (Scheme 2).
2.4. Molecular modeling calculations
To further understand this base mediated in situ C3 epi-
merization observed with 9ae9d, we carried out molecular
modeling calculations using the Cerius2 4.10 software

4366
R. Liu et al. / Tetrahedron 64 (2008) 4363e4369
CO2Me
O
CO2Me
O
CO2Me
O
N
N
H
N
BocHN
BocHN
BocHN
S
S
S
H
11a
H
11b
Figure 3. ORTEP drawings of compounds 11a, 11b, and 11c-r.
11c-r
O
CO2Me
3
O
CONH2
3
CONH2
O
CO2Me
3
O
CONH2
3
O
Bn
3
Bn
NH3 / MeOH
sealed tube;
N
Bn
Bn
Bn
N
NH3 / MeOH
r.t. 2 h
N
N
H
N
S
BocHN
S
BocHN
BocHN
S
S
S
BocHN
BocHN
50 °C, 2 h
H
9c
H
H
H
9a
10c-r
E = -19.5 kcal/mol
10a
10c
E = -14.9 kcal/mol
enantiomers
O
CONH2
3
O
CO2Me
3
O
CONH2
CONH2
3
O
O
CO2Me
NH3 / MeOH
sealed tube;
3
Bn
3
NH3 / MeOH
Bn
Bn
Bn
N
Bn
N
N
N
N
r.t. 2h
S
BocHN
BocHN
BocHN
50 °C, 2 h
S
BocHN
S
S
S
BocHN
H
10d
H
9d
H
H
10b
H
9b
10d-r
E = -24.9 kcal/mol
enantiomers
E = -20.6 kcal/mol
Scheme 2. Base induced epimerization of 10c and 10d.
package,16 which utilizes molecular mechanics methods to ob-
tain minimized energy values. The molecular structures were
built by Cerius2 3D-Sketcher Module, from which CFF17
force field was selected for energy calculations.
The minimized energies for 10c, 10c-r, 10d, and 10d-r
were also calculated using the same molecular modeling pro-
gram. In light of the results highlighted in Scheme 2, it became
evident that compounds 10c and 10d could undergo C3 epime-
rization to yield their corresponding lower energy C3 epimers
10c-r and 10d-r under the reaction conditions previously
outlined.
According to the energy calculations shown in Table 2, iso-
mers 9a and 9b were found to be thermodynamically more sta-
ble than their corresponding C3 epimers 9a-r (by 5.3 kcal/
mol) and 9b-r (by 1.4 kcal/mol). Contrary to this finding, iso-
mers 9c and 9d had relatively higher energy than their corre-
sponding C3 epimers 9c-r (by 5.0 kcal/mol) and 9d-r (by
1.4 kcal/mol), suggesting that 9c and 9d could be converted
to 9c-r and 9d-r under DBU treatment. The chemical reactiv-
ity toward C3 epimerization predicted by the molecular mod-
eling calculations is in agreement with the experimental
findings outlined in Table 1 and Scheme 1.
3. Conclusion
We describe herein a detailed investigation into the thiazo-
lidine g-lactam isomer distribution (9ae9d) obtained under
various cyclization conditions. As a result of this effort, we
found that isomer 9a could be prepared in close to 50% yield
(highest theoretical yield) by treatment of methyl ester 8 with
TsOH in toluene. Furthermore, we discovered a C3 epimeri-
zation event occurred to 9c, 9d as well as to 10c, 10d upon
basic treatment under thermodynamic conditions (e.g., DBU
or ammonia in methanol). This ‘unexpected’ C3 epimerization
is further rationalized and supported by molecular modeling
calculations as well as confirmed by crystal structure analyses
on 11a, 11b, and 11c-r.
Table 2
Minimum energy calculation for thiazolidine g-lactam isomers
Isomer
9a
9b
9c
9d
E (kcal/mol)
Isomer
E (kcal/mol)
9.011
9a-r
14.257
7.102
9b-r
8.507
11.953
9c-r
6.921
11.386
9d-r
9.978

R. Liu et al. / Tetrahedron 64 (2008) 4363e4369
4367
4. Experimental
to reflux. After refluxing overnight, THF was removed
in vacuo. The residue was dissolved in dichloromethane
(300 mL) and cooled to ꢀ78 ꢂC. Then ozone in oxygen (5e
10%) was bubbled through until the color turned gray. Nitro-
gen was bubbled through the reaction mixture to remove extra
ozone and triphenyl phosphine (78.0 g, 0.3 mol) was added
slowly. The reaction was stirred for 1 h. The solvent was
removed under reduced pressure and the residue was purified
by silica gel column chromatography to afford compound 6
4.1. General
Unless otherwise noted, materials were obtained from com-
mercial suppliers and used without further purification. 1H and
13C NMR spectra were recorded on a Bruker Avance 300 or
400 spectrometer at ambient temperature. Chemical shifts
are reported relative to residual solvents. Abbreviations for
1H NMR: s¼singlet, d¼doublet, t¼triplet, m¼multiplet, br¼
broad. Mass spectra were obtained on a Finnigan LCQ mass
spectrometer with ESI source. Melting points were obtained
on a WRR melting point apparatus and are uncorrected. Opti-
cal rotations were measured on a PerkinElmer model 341LC
polarimeter at the 589-nm Na D-line. X-ray diffraction was
performed on BRUKER SMART APEX-CCD.
1
(45.0 g, 61%). H NMR (CDCl3) d 1.44 (s, 9H), 2.96 (d, J¼
13.2 Hz, 1H), 3.06 (d, J¼13.2 Hz, 1H), 3.61 (d, J¼12.6 Hz,
1H), 3.73 (s, 3H), 3.85 (d, J¼12.6 Hz, 1H), 5.55 (br, 1H),
7.00e7.04 (m, 2H), 7.25e7.26 (m, 3H), 9.67 (s, 1H).
4.4. 2-Benzyl-2-(tert-butoxycarbonylamino)-3-((4R)-4-
(methoxycarbonyl)thiazolidin-2-yl)propanoic acid (7)
4.2. Methyl 2-amino-2-benzylpent-4-enoate (5)
To a solution of compound 6 (8.0 g, 25 mmol) in methanol
(120 mL) and water (40 mL), lithium hydroxide hydrate
(4.2 g, 100 mmol) was added. After stirring overnight, the re-
sulting mixture was adjusted to pH¼5e6 using 1 M HCl, and
extracted with ethyl acetate. The combined organic phases
were concentrated to afford the corresponding acid (7.58 g,
99%). To a solution of the above acid (7.58 g, 24.8 mmol) in
EtOH (100 mL) and water (100 mL) was added NaHCO3
(2.5 g, 30 mmol), followed by the addition of L-cysteine hy-
drochloride (4.76 g, 28 mmol). The pH value of the resulting
mixture was adjusted to 6.5 with 1 M NaHCO3. After stirring
overnight, the mixture was concentrated to about one-half
volume, and the pH was adjusted to 6 with 1 M HCl. The mix-
ture was poured into water (100 mL) and extracted with ethyl
acetate (200 mL). The combined organic phases were dried
over Na2SO4, filtered, and concentrated to afford compound
7 (9.12 g, 72%). MS (ESI): 425 (MþþH).
To a suspension of L-phenylalanine hydrochloride (110.0 g,
0.5 mol) in dichloromethane (1000 mL) was added triethyl-
amine (83.5 mL, 0.6 mol). After stirring for 0.5 h, benzaldehyde
(53 g, 0.5 mol) was added dropwise to the mixture, followed by
the addition of magnesium sulfate (120 g, 0.5 mol). The mixture
was stirred overnight. Dichloromethane was removed under va-
cuo and ethyl acetate (1000 mL) was added. After filtration, the
filtrate was concentrated to afford the corresponding imine
(133.72 g, 100%). To a solution of potassium tert-butoxide
(72.8 g, 0.65 mol) in THF (1000 mL) at room temperature
was added a solution of the above imine (133.72 g, 0.5 mol)
in THF (300 mL). After stirring for 0.5 h, allyl bromide (84 g,
0.7 mol) was added dropwise to the resulting mixture. The
mixture was stirred overnight at room temperature. THF was
removed under vacuo and ethyl acetate (1000 mL) was added.
After filtration, the filtrate was concentrated to afford the alkyl-
ated imine (153.66 g, 100%). To a solution of the alkylated
imine (153.66 g, 0.5 mol) in THF (500 mL) was added 3 M
HCl (500 mL, 1.5 mol) slowly at room temperature. After stir-
ring for 1 h, THF was removed in vacuo, and ethyl acetate
(500 mL) and 3 M HCl (150 mL) were added. After separation
of both phases, 300 mL of ethyl acetate were used to extract
away the impurity and then the aqueous layer was adjusted to
pH¼10e12. The aqueous phasewas extracted with dichlorome-
thane (2ꢁ250 mL). The combined organic layers were washed
with brine (100 mL) and concentrated to afford compound 5
(96.75 g, 88%) as a colorless oil, which was used directly to
4.5. (4R)-Methyl 2-(2-benzyl-2-(tert-butoxycarbonyl amino)-
3-methoxy-3-oxopropyl)thiazolidine-4-carboxylate (8)
To a solution of compound 6 (9.6 g, 20 mmol), AcOH
(1.25 mL, 20.9 mmol), and AcONa (3.43 g, 41.8 mmol) in
an ethanol (50 mL) and water (5 mL) mixture was added
L-cysteine hydrochloride (3.93 g, 23.0 mmol). After stirring
overnight, ethyl acetate (150 mL) was added and the organic
layer was washed with brine (50 mL), dried over anhydrous
Na2SO4, filtered, and concentrated to afford compound 8
(7.33 g, 84%). MS (ESI): 439 (MþþH).
1
the next step without further purification. H NMR (CDCl3)
d 2.36 (dd, J¼8.0 and 12.8 Hz, 1H), 2.50e2.75 (br, 2H), 2.79
(dd, J¼6.0 and 12.8 Hz, 1H), 2.82 (d, J¼12.8 Hz, 1H), 3.19
(d, J¼12.8 Hz, 1H), 3.70 (s, 3H), 5.15e5.21 (m, 2H), 5.68e
5.74 (m, 1H), 7.13e7.17 (m, 2H), 7.22e7.29 (m, 3H).
4.6. Methyl 6-benzyl-6-(tert-butoxycarbonylamino)-5-oxo-
hexahydropyrrolo[2,1-b]thiazole-3-carboxylate (9)
4.6.1. Method A
To a solution of the thiazolidine mixture 7 (4.24 g, 10 mmol)
in dichloromethane (400 mL) was added 2-chloro-1-methyl-
pyridinium iodide (Mukaiyama reagent) (3.06 g, 12 mmol),
followed by the addition of triethylamine (2.75 g, 25 mmol).
This resulting mixture was heated to reflux overnight. The reac-
tion mixture was allowed to cool to room temperature and
4.3. Methyl 2-benzyl-2-(tert-butoxycarbonylamino)-4-oxo-
butanoate (6)
A solution of compound 5 (50.0 g, 0.228 mol) and Boc
anhydride (78.0 g, 0.342 mol) in THF (300 mL) was heated

4368
R. Liu et al. / Tetrahedron 64 (2008) 4363e4369
then washed with 10% citric acid (50 mL), 1 M NaHCO3
(50 mL), and brine (50 mL). The concentrated mixture was
purified by silica gel column chromatography and preparative
HPLC to afford compounds 9a (0.68 g, 17%), 9b (1.13 g,
28%), 9c (0.43 g, 11%), and 9d (1.05 g, 26%). [Note: com-
pounds 9b and 9c were purified by HPLC. Analytic HPLC con-
ditions: YMC 150 mmꢁ4.6 mm, 5 mm; mobile phase: 50%
MeOHþ50% H2O (0.05% TFA); flow rate: 1 mL/min; detector:
PDA 207 nm. Retention times for 9a, 9b, 9c, and 9d were
50.15, 55.89, 45.61, 52.44 min, respectively. Preparative
HPLC conditions: AS-V, 250 mmꢁ77 mm, 20 mM; mobile
phase: 4% ethanolþ96% n-hexane; flow rate: 250 mL/min;
detector: PDA 220 nm. Retention times for 9b and 9c were
16.5 and 12.3 min, respectively.]
40.2, 51.2, 55.6, 62.7, 65.3, 78.3, 126.2, 127.2, 129.4, 134.6,
153.2, 167.6, 169.8; HRMS (ESI) calcd for C20H27N2O5S:
407.1641 (MþþH), found: 407.1629.
4.6.1.4. (3R,6R,8R)-Methyl 6-benzyl-6-(tert-butoxycarbonyl-
amino)-5-oxohexahydropyrrolo[2,1-b]thiazole-3-carboxylate
(9d). White solid; TLC Rf ¼0.4 (PE/EA¼5:1, twice); mp
1
71.1e73.2 ꢂC; [a]D þ15.1 (c 1.0, MeOH); H NMR (CD3CN)
d 1.32 (s, 9H, C(CH3)3), 2.37 (dd, J¼5.4 and 13.8 Hz, 1H,
H7b), 2.98 (dd, J¼7.2 and 13.8 Hz, 1H, H7a), 3.00 (d,
J¼14.1 Hz, 1H, CHHPh), 3.11 (d, J¼14.1 Hz, 1H, CHHPh),
3.31 (dd, J¼3.0 and 12.3 Hz, 1H, H2b), 3.59 (dd, J¼7.5 and
12.3 Hz, 1H, H2a), 3.73 (s, 3H, COOCH3), 4.23 (dd, J¼3.0
and 7.5 Hz, 1H, H3), 5.18 (dd, J¼5.4 and 7.2 Hz, 1H, H8),
5.45 (br, 1H, NH), 7.26e7.37 (m, 5H, Ph); 13C NMR
(CD3CN) d 26.7, 34.2, 37.1, 41.3, 51.2, 56.1, 63.3, 63.8,
78.6, 126.2, 127.6, 129.8, 134.8, 153.7, 167.9, 170.8; HRMS
(ESI) calcd for C20H27N2O5S: 407.1641 (MþþH), found:
407.1625.
4.6.1.1. (3R,6R,8S)-Methyl 6-benzyl-6-(tert-butoxycarbonyl-
amino)-5-oxohexahydropyrrolo[2,1-b]thiazole-3-carboxylate
(9a). White solid; TLC Rf ¼0.6 (PE/EA¼5:1, twice); mp
121.5e122.2 ꢂC; [a]D ꢀ102.8 (c 1.0, MeOH); 1H NMR
(MeOD) d 1.42 (s, 9H, C(CH3)3), 2.56 (dd, J¼7.5 and
12.9 Hz, 1H, H7a), 2.79 (dd, J¼6.6 and 12.9 Hz, 1H, H7b),
2.91 (d, J¼12.9 Hz, 1H, CHHPh), 3.07 (d, J¼12.9 Hz, 1H,
CHHPh), 3.19e3.22 (m, 1H, H2b), 3.35e3.39 (m, 1H, H2a),
3.66 (s, 3H, COOCH3), 3.97 (br, 1H, H8), 5.05 (br, 1H, H3),
7.22e7.29 (m, 5H, Ph); 13C NMR (MeOD) d 26.5, 33.0,
39.5, 40.8, 50.8, 57.7, 59.5, 64.8, 78.4, 126.1, 127.2, 129.3,
133.2, 154.1, 168.5, 173.1; HRMS (ESI) calcd for
C20H27N2O5S: 407.1641 (MþþH), found: 407.1631.
4.6.2. Method G
A solution of compound 7 (4.24 g, 10 mmol) and p-tolue-
nesulfonic acid monohydrate (0.19 g, 1.0 mmol) in toluene
(300 mL) was heated to reflux overnight. The solution was
allowed to cool down to room temperature, washed with
NaHCO3 solution (50 mL) and brine (50 mL), and dried
over Na2SO4. After filtration and removal of solvents under
reduced pressure, the residue was purified by silica gel column
chromatography to give compounds 9a (1.56 g, 38%) and 9b
(1.41 g, 35%).
4.6.1.2. (3R,6S,8S)-Methyl 6-benzyl-6-(tert-butoxycarbonyl-
amino)-5-oxohexahydropyrrolo[2,1-b]thiazole-3-carboxylate
(9b). White solid; TLC Rf ¼0.53 (PE/EA¼5:1, twice); mp
107.0e108.8 ꢂC; [a]D ꢀ170.7 (c 1.0, MeOH); 1H NMR
(CD3CN) d 1.39 (s, 9H, C(CH3)3), 2.42 (dd, J¼2.7 and
14.4 Hz, 1H, H7a), 2.64 (dd, J¼7.8 and 14.4 Hz, 1H, H7b),
2.93 (d, J¼13.5 Hz, 1H, CHHPh), 3.03 (d, J¼13.5 Hz, 1H,
CHHPh), 3.07 (dd, J¼7.8 and 11.1 Hz, 1H, H2a), 3.22 (dd,
J¼3.3 and 11.1 Hz, 1H, H2a), 3.71 (s, 3H, COOCH3), 4.96
(dd, J¼3.3 and 7.8 Hz, 1H, H3), 5.16 (dd, J¼2.7 and 7.8 Hz,
1H, H8), 5.50 (br, 1H, NH), 7.22e7.35 (m, 5H, Ph); 13C
NMR (CD3CN) d 26.7, 33.0, 34.7, 41.3, 51.5, 57.6, 61.5,
62.1, 78.9, 126.4, 127.6, 129.9, 134.4, 153.5, 169.1, 173.9;
HRMS (ESI) calcd for C20H27N2O5S: 407.1641 (MþþH),
found: 407.1635.
4.6.3. Method H
Similar procedure to Method G.
4.6.4. Method I
A solution of compound 8 (3.13 g, 7.1 mmol) and DBU
(3.3 g, 22 mmol) in toluene (60 mL) was heated to reflux over-
night. The reaction mixture was allowed to cool down to room
temperature and then neutralized with acetic acid. The solu-
tion was washed with brine (20 mL), dried over Na2SO4,
and filtered. After concentration, the residue was purified by
silica gel column chromatography to give compounds ‘9a’
(1.21 g, 42%; 9a: 33%, 9c-r: 9%) and ‘9b’ (0.80 g, 27%;
9b, 24%, 9d-r, 3%). [Chiral HPLC conditions: AD-H
250 mmꢁ4.6 mm, 5 mm; mobile phase: 15% EtOH (0.1%
TFA)þ85% n-hexane; flow rate: 0.5 mL/min; detector: PDA
207 nm. Retention times detected for 9a, 9b, 9c, 9d, 9c-r,
and 9d-r were 18.51, 34.85, 27.63, 17.24, 25.43, and
23.52 min.]
4.6.1.3. (3R,6S,8R)-Methyl 6-benzyl-6-(tert-butoxycarbonyl-
amino)-5-oxohexahydropyrrolo[2,1-b]thiazole-3-carboxylate
(9c). White solid; TLC Rf ¼0.53 (PE/EA¼5:1, twice); mp
157.3e158.7 ꢂC; [a]D ꢀ25.2 (c 1.0, MeOH); 1H NMR
(CD3CN) d 1.44 (s, 9H, C(CH3)3), 2.50 (dd, J¼8.7 and
12.6 Hz, 1H, H7b), 2.98 (dd, J¼5.7 and 12.6 Hz, 1H, H7a),
3.06 (d, J¼13.5 Hz, 1H, CHHPh), 3.16 (d, J¼13.5 Hz, 1H,
CHHPh), 3.32 (dd, J¼2.4 and 12.6 Hz, 1H, H2b), 3.52 (dd,
J¼7.5 and 12.6 Hz, 1H, H2a), 3.69 (s, 3H, COOCH3), 4.14
(dd, J¼2.4 and 7.5 Hz, 1H, H3), 4.55 (dd, J¼5.7 and 8.7 Hz,
1H, H8), 5.26 (br, 1H, NH), 7.15e7.17 (m, 2H, Ph), 7.25e
7.31 (m, 3H, Ph); 13C NMR (CD3CN) d 26.7, 37.1, 40.0,
4.7. Epimerization experiment
A solution of compound 9c (81 mg, 0.2 mmol) and DBU
(90 mg, 0.6 mmol) in toluene (5 mL) was heated to reflux
overnight. The reaction mixture was allowed to cool down
to room temperature and then neutralized with acetic acid.
The solution was washed with brine (2 mL), dried over

R. Liu et al. / Tetrahedron 64 (2008) 4363e4369
4369
Na2SO4, and filtered. After concentration, the residue was
Acknowledgements
purified by column chromatography on silica gel to give com-
pound 9c-r (75 mg, 93%). The same experiment is carried
out for 9a, 9b, and 9d to give 9a (95%), 9b (94%), and
9d-r (97%).
We shall thank Mr. Qusheng Xu and Yujun Bai for their
help with synthetic work, Ms. Wei Ji for her help with NMR
analysis. We are also indebted to the chemists from Analytic
Department at WuXi PharmaTech for their support.
4.8. Aminolysis experiment
References and notes
A round-bottom flask was charged with compound 9a
(500 mg, 1.2 mmol) and 10 M ammonia solution in methanol
(50 mL), and the mixture was stirred overnight. The solvent
1. Martin-Martinez, M.; Patino-Molina, R.; Garcia-Lopez, M. T.; Gonzalez-
Muniz, R. Mini Rev. Med. Chem. 2004, 4, 669e680.
1
2. Review: Greenlee, W. J. Med. Res. Rev. 1990, 10, 173e236.
3. Review: Abdel-Rahman, H. M.; Al-karamany, G. S.; El-Koussi, N. A.;
Youssef, A. F.; Kiso, Y. Curr. Med. Chem. 2002, 9, 1905e1922.
4. Review: Wang, Q. M.; Chen, S. H. Frontiers of Biotechnology &
Pharmaceuticals; Reiner, Z., Chen, G., Eds.; Science: USA, 2004;
Vol. 4, pp 30e49.
5. Reviews: (a) De Francesco, R.; Tomei, L.; Altamura, S.; Summa, V.;
Migliaccio, G. Antiviral Res. 2003, 58, 1e16; (b) Chen, S. H.; Tan,
S.-L. Curr. Med. Chem. 2005, 12, 2317e2342.
was distilled out to afford compound 10a (465 mg, 99%). H
NMR (CDCl3) d 1.46 (s, 9H), 2.64e2.70 (m, 1H), 3.03e3.09
(m, 2H), 3.19e3.25 (m, 2H), 3.54 (dd, J¼4.5 and 11.4 Hz,
1H), 4.01 (t, J¼9.6 Hz, 1H), 4.80e4.83 (m, 1H), 5.27 (br,
1H), 5.74 (br, 1H), 7.25e7.36 (m, 5H); 13C NMR (CDCl3)
d 28.3, 31.5, 40.9, 42.7, 57.8, 60.6, 66.5, 80.8, 127.6, 128.7,
129.6, 133.5, 154.5, 171.1, 172.0; HRMS (ESI) calcd for
C19H26N3O4S: 392.1644 (MþþH), found: 392.1635.
The same experiment is carried out for 9b to give 10b
(98%). 1H NMR (CDCl3) d 1.38 (s, 9H), 2.48e2.55 (m,
2H), 2.94 (d, J¼12.9 Hz, 1H), 3.13 (d, J¼12.9 Hz, 1H), 3.31
(dd, J¼8.9 and 11.2 Hz, 1H), 3.53 (dd, J¼4.0 and 11.2 Hz,
1H), 4.81e4.83 (m, 1H), 5.08 (br, 1H), 5.14e5.17 (m, 1H),
5.45 (br, 1H), 7.21e7.25 (m, 2H), 7.30e7.36 (m, 3H); 13C
NMR (CDCl3) d 28.5, 31.4, 36.3, 43.4, 57.7, 62.6, 64.1,
81.2, 128.0, 129.2, 130.6, 134.4, 155.1, 171.7, 172.7; HRMS
(ESI) calcd for C19H26N3O4S: 392.1644 (MþþH), found:
392.1626.
6. Review: John, V.; Beck, J. P.; Bienkowski, M. J.; Sinha, S.; Heinrikson,
B. L. J. Med. Chem. 2003, 46, 4625e4630.
7. Barrett, D. G.; Catalano, J. G.; Deaton, D. N.; Hassell, A. M.; Long, S. T.;
Miller, L. R.; Shewchuk, L. M.; Wells-Knecht, K. J.; Willard, D. H.;
Wright, L. L. Bioorg. Med. Chem. Lett. 2004, 14, 4897e4902.
8. (a) Gante, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 1699e1720; (b)
Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1244e1267.
9. Khalil, E. M.; Pradhan, A.; Ojala, W. H.; Gleason, W. B.; Mishra, R. K.;
Johnson, R. L. J. Med. Chem. 1999, 42, 2977e2987.
10. Johannesson, P.; Erdelyi, M.; Lindeberg, G.; Frandberg, P.-A.; Nyberg, F.;
Karlen, A.; Hallberg, A. J. Med. Chem. 2004, 47, 6009e6019.
11. Genin, M. J.; Johnson, R. L. J. Am. Chem. Soc. 1992, 114, 8778e8783.
12. Baldwin, J. E.; Lee, V.; Schofield, C. J. Heterocycles 1992, 34, 903e906.
13. Campbell, D.; Betancort, J.; Winn, D. WO2006017292, 2006.
14. Fuchi, N.; Doi, T.; Harada, T.; Urban, J.; Cao, B.; Kahn, M.; Takahashi, T.
Tetrahedron Lett. 2001, 42, 1305e1308.
A round-bottom flask was charged with compound 9c
(500 mg, 1.2 mmol) and 10 M ammonia solution in methanol
(50 mL) and sealed. It was heated to 60 ꢂC for 2 h. After cool-
ing to room temperature, the solvent was distilled out to afford
compound 10c-r (460 mg, 98%).
15. Dolbeare, K.; Pontoriero, G. F.; Gupta, S. K.; Mishra, R. K.; Johnson,
R. L. J. Med. Chem. 2003, 46, 727e733.
16. Cerius2 Modeling Environment, Release 4.10; Accelrys: San Diego, CA,
USA, 2005.
17. Hagler, A. T.; Ewig, C. S. Comput. Phys. Commun. 1994, 84, 131e155.
The same experiment is carried out for 9d to give 10d-r
(97%).