Enantiospecific synthesis and cytotoxicity of 7-(4-methoxyphenyl)-6-phenyl-2,3,8,8a-tetrahydroindolizin-5(1H)-one enantiomers

Article abstract of DOI:10.1016/j.bmc.2008.02.073

An enantiospecific synthesis was developed to generate both enantiomers of 7-(4-methoxyphenyl)-6-phenyl-2,3,8,8a-tetrahydroindolizin-5(1H)-one. A biological assay utilizing the HCT-116 colon cancer cell line to determine the cytotoxicity of these analogs revealed that only the (R)-enantiomer exhibited appreciable cytotoxicity with an IC₅₀ value of 0.2 μM.

Full text of DOI:10.1016/j.bmc.2008.02.073

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 4367–4377
Enantiospecific synthesis and cytotoxicity
of 7-(4-methoxyphenyl)-6-phenyl-2,3,8,8a-
tetrahydroindolizin-5(1H)-one enantiomers
F. Scott Kimball,a Brandon J. Turunen,a Keith C. Ellis,a
Richard H. Himesb and Gunda I. Georga,*
aDepartment of Medicinal Chemistry, University of Kansas, 1251 Wescoe Hall Drive, Lawrence, KS 66045, USA
bDepartment of Molecular Biosciences, University of Kansas, 1200 Sunnyside Avenue, Lawrence, KS 66045, USA
Received 24 December 2007; revised 16 February 2008; accepted 21 February 2008
Available online 4 March 2008
Abstract—An enantiospecific synthesis was developed to generate both enantiomers of 7-(4-methoxyphenyl)-6-phenyl-2,3,8,8a-tet-
rahydroindolizin-5(1H)-one. A biological assay utilizing the HCT-116 colon cancer cell line to determine the cytotoxicity of these
analogs revealed that only the (R)-enantiomer exhibited appreciable cytotoxicity with an IC50 value of 0.2 lM.
Ó 2008 Elsevier Ltd. All rights reserved.
OMe
1. Introduction
OMe
OMe
H
N
The emergence of resistance to common anti-tumor
agents by cancer cells makes it imperative that new com-
pounds, operating through novel modes of action, be
developed. Successful endeavors in this arena could
allow for the identification of new cancer targets that
may not have inherent resistance mechanisms. To this
end, research in our laboratories has focused on the syn-
thesis and biological evaluation of cytotoxic compounds
that act through unknown mechanisms of action. These
pursuits led us to our research into the tyloindicine fam-
ily, originally disclosed by Ali et al.1,2
H
N
OH
OMe
O
OMe
(+/-)-2
1
Figure 1. Tyloindicine I (1) and tetrahydroindolizin-5(1H)-one ( )-2.
cancer and leukemia cell lines, and was effective
in vivo in the mouse hollow fiber assay.3 Driven by this
intriguing biological data, we set out to determine if the
stereochemistry of 2 influenced the biological activity.
In a recent communication, the racemic synthesis of a
tyloindicine I (1) based diaryl-2,3,8,8a-tetrahydroindoli-
zin-5(1H)-one analog library was described (Fig. 1).3
This library evaluated substitution on both the northern
and southern aromatic rings attached to the tetrahydro-
indolizin-5(1H)-one core.4 Our studies revealed that the
lead compound ( )-2 (NSC 707904) was the most cyto-
toxic in the library and that it most likely operates
through an unknown, novel mechanism of action.3
Furthermore, ( )-2 was selectively active toward colon
Employing preparative HPLC,5 the racemate ( )-2 was
separated to provide enantiopure (+)-2 and (ꢀ)-2.6
The pure enantiomers were assayed for cytotoxicity
against the HCT-116 colon cancer cell line, using race-
mic tylophorine as a standard. This assay revealed that
the (ꢀ)-2 enantiomer exhibited submicromolar cytotox-
icity whereas (+)-2 was less than 1% as active (Table 1).
These results reinforced the necessity to develop an
enantiospecific synthesis of (ꢀ)-2 so the absolute stereo-
chemistry of the active enantiomer could be determined.
Keywords: Tetrahydroindolizinones; Enantiospecific synthesis; Colon
cancer; HCT-116.
The selection of enantiomer (S)-2 as the target for syn-
thesis (Scheme 1) was influenced by Ali’s proposal of
the (S)-stereochemistry of tyloindicine I (Fig. 1).2
*
Corresponding author. Tel.: +1612 626 6320; fax: +1612 626 6318;
e-mail: georg239@umn.edu
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.02.073

4368
F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
Table 1. Cytotoxicity of stereoisomers of 2 and tylophorine against
HCT-116 cells
furnished aryl ketone 5. An acidic BOC-deprotection
of ketone 5 formed a stable hydrochloride amine salt,
which was subjected to an EDCI coupling protocol to
supply the aldol precursor 3.
Analog
IC50 (lM)
( )-2
(+)-2
0.39
46
It was crucial to handle and store the amine precursor to
intermediate 3 as the HCl salt, as the free amine rapidly
decomposed. The final step called for a basic intramolec-
ular aldol condensation to deliver title compound 2. To
our disappointment, we obtained the racemate ( )-2,7
rather than the desired (S)-2 enantiomer as suggested
by the lack of optical rotation and confirmed by chiral
HPLC. This result led us to hypothesize that at some
point in the synthesis, the reaction conditions triggered
a retro-Michael–Michael ring opening/closing sequence
that racemized the stereogenic center.
(ꢀ)-2
0.20
0.67
( )-Tylophorine
OMe
OMe
OH
O
H
N
O
N
N
O
t
-Bu
O
O
O
2
3
4
Retro-Michael–Michael reactions are known to be cata-
lyzed by either acid or base. Efforts to identify at what
stage the racemization occurred were hampered by the
inability to resolve the intermediates. Consequently, at-
tempts to revise the synthesis to suppress the racemiza-
tion had to take the mechanism for the retro-Michael–
Michael process into account. The acid-catalyzed race-
mization mechanism is shown in Scheme 3, although
the base-catalyzed racemization would proceed through
a similar mechanism.
Scheme 1. Retrosynthesis of (S)-2.3
Retrosynthetically, title compound (S)-2 can be envi-
sioned as coming from amide 3 via an intramolecular al-
dol condensation. We reasoned that the desired
asymmetry could be effectively derived from a chiral
pool approach using N-BOC-L-homoproline (4) as the
starting substrate. Thus, in turn, amide 3 may be derived
from 4 by Weinreb amide formation, Grignard addition,
and BOC-deprotection/amide bond formation sequence
(Scheme 1).
In this process, an acid-catalyzed carbamate activation
allows for the initial retro-Michael ring opening.
Depending on which a-proton is removed, either 5.1
or 5.2 may result. A simple bond rotation can give rise
to either 5.3 or 5.4. Once p-orbital overlap is re-estab-
lished the intramolecular Michael reaction can proceed.
In that each step is reversible and there are no stereocon-
trolling influences elsewhere in the molecule, discrimina-
tion of the Re and Si faces of the enone is nullified
during the intramolecular Michael reaction and racemi-
zation occurs to provide the racemic mixture. It should
be noted that this is the only one possible scenario. It
seemed feasible that epimerization could have occurred
at any, or all, of the steps along the way.
In a forward sense, our synthesis commenced with the
Weinreb (S)-2-(1-(tert-butoxycar-
bonyl)pyrrolidin-2-yl)acetic acid (4) (Scheme 2). A sub-
amidation
of
sequent Grignard addition into the Weinreb amide
OMe
OH
O
O
a, b
c, d
N
O
O
t
-Bu
N
O
t
-Bu
O
4
Ar
Ar
5
OMe
Bond
Rotation
Acid-catalyzed
Retro-Michael
O
OMe
O
N
H
O
t
H
N
N
H
O
HO
e
O
-Bu
O
t-Bu
N
5.1 trans
5.2 cis
O
(S)-5
O
(+/-)-2
H
Ar
3
O
Michael
Reaction
N
Scheme 2. Initial synthetic attempt toward (S)-2. Reagents and
conditions: (a) N,O-dimethylhydroxylammonium chloride, EDCI,
NMM, dry CH2Cl2, ꢀ15 °C, 12 h (91%); (b) i—Mg0 turnings, cat. I2
crystals, 4-bromoanisol, dry THF, reflux, 1 h (87%); ii—add cooled
Grignard reagent to Weinreb amide intermediate, ꢀ5 °C, 1 h; (c) 4 M
HCl/dioxane, 24 °C, 15 min; (d) phenylacetic acid, EDCI, NMM, dry
O
O
t
N
O
Ar
-Bu
HO
O
t
-Bu
5.3 trans
5.4 cis
(R)-5
CH2Cl2, ꢀ15 °C, 90 min (74% over two steps); (e) 5% KOH(EtOH)
,
Scheme 3. Retro-Michael–Michael racemization mechanism for the
conversion of (S)-5 to (R)-5.
reflux, 5 h (80%).

F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
4369
With this in mind, we turned to the use of a diastereo-
meric derivative of starting material 4, the hydroxypro-
line derivative 6 (Scheme 4). It was reasoned that this
would enable us in several respects. First, this additional
‘chemical reporter’ group would be situated in a locale
in which racemization is unlikely and, as such, would al-
low us to readily track the integrity of our core stereo-
genic center while consequently permitting us to
determine at which step the epimerization occurred. Sec-
ond, if properly chosen, it may provide us a tunable,
conformational bias during this retro-Michael–Michael
process. Lastly, it would allow for practical separation
of epimerized side products. A trait we deemed neces-
sary in this auxiliary group: it must be readily removed
and therefore a traceless reporter/controller.
The free hydroxy analog (2S,4R)-12 exhibited the lowest
yield for this transformation (38%) while both the acyl-
ated and benzylated versions exhibited satisfactory
yields (88% and 73%, respectively). The benzylated ana-
log 14 showed the poorest dr at 60:40 and acylated
(2S,4R)-13 the best dr at 80:20. All of the diastereomeric
mixtures could be resolved through silica chromatogra-
phy to supply pure diastereomers in preparation for the
basic intramolecular aldol condensation (Table 2). This
condensation proceeded to furnish (2R,8aR)-15 in 89%
yield (70:30 dr) when (2S,4R)-12 was employed as the
starting material. (2S,4R)-13 underwent the intramolec-
ular aldol condensation with concomitant acetate-
deprotection/to provide the same (2R,8aR)-15 in a
slightly lower yield (75%) and dr (64:36 dr). The benzy-
lated (2S,4R)-14 provided the condensed product
(2R,8aR)-16 in 44% and 61:39 dr. Analogous to the
acidic BOC-deprotection conditions used earlier in the
synthesis, the basic aldol conditions also triggered race-
mization. The purified single diastereomer (2R,8aR)-15
was re-subjected to the basic aldol conditions to ascer-
tain whether the racemization occurs before or after
the cyclization takes place. A 92:8 dr was observed fol-
lowing exposure to these basic conditions indicating that
the majority of the racemization occurs prior to the
cyclization.
The revised synthesis toward (S)-2 began with the
Arndt-Eistert homologation of (2S,4R)-1-(tert-butoxy-
carbonyl)-4-hydroxyproline (6, Scheme 4). The resultant
homolog 7 was then converted to the Weinreb amide to
allow for Grignard addition of (4-methoxyphenyl)mag-
nesium bromide into the amide to furnish arylketone
8. The acylated intermediate 9 and benzylated interme-
diate 10 were then prepared from 8 to probe the signif-
icance of protecting groups of increasing size during the
retro-Michael–Michael racemization.8 Up to this point,
no epimerization was detected. Next, the BOC-protected
analogs were deprotected and acylated to give the corre-
sponding amides (2S,4R)-12, (2S,4R)-13, and (2S,4R)-
14. After deprotection, ketone 8 afforded the hydrochlo-
ride salt 11 (R = H) in a 77:23 diastereomeric ratio,
revealing that the retro-Michael–Michael process does
occur under the acidic BOC-deprotection conditions.
We also examined the purified syn-diastereomers in the
aldol condensation (Table 3). Compounds (2R,4R)-13
and (2R,4R)-14 were subjected to the basic intramolecu-
lar aldol condensation conditions. The yields and the dr
observed for this transformation were indeed lower than
those for the corresponding anti-substrates (Table 2). As
OMe
O
OH
OH
2
O
O
HO
a
b, c
4
HO
O
O
RO
N
N
N
O
t
-Bu
t
-Bu
t-Bu
O
O
O
7
(2S,4R)-6
8
9
R=H
d
e
R=Ac
10 R=Bn
OMe
OMe
OMe
H
8a
2
RO
2
N
4
h
O
RO
f
g
N
O
RO
O
Cl
NH2
O
(2R,8a
R
)-15 R=H
(2S
(2S
(2S
,4R
,4R
,4R
)-12 R=H
11
(2R,8a
R
)-16 R=Bn
)-13 R=Ac
)-14 R=Bn
Scheme 4. Synthesis of tetrahydroindolizin-5(1H)-one intermediates (R,R)-15 and (R,R)-16. Reagents and conditions: (a) i—TEA, ClCO2Et,
diazomethane, THF, 0 °C, rt, 12 h; ii—PhCO2Ag, TEA, ꢀ25 °C, rt, 12 h, (27%, unoptimized); (b) N,O-dimethylhydroxylammonium chloride, EDCI,
NMM, dry CH2Cl2, 0 °C, rt, 2 h (98%); (c) Mg0 turnings, cat. I2 crystals, 4-bromoanisol, dry THF, (85%); (d) Ac2O, DMAP, pyridine, 0 °C, rt, 1 h
(99%); (e) Ag2O, BnBr, toluene, 50 °C, 3 h (74%); (f) 4 M HCl/dioxane, (for R = H a 77:23 dr was noted; for R = Ac, Bn the dr was not determined);
(g) phenylacetic acid, EDCI, NMM, dry CH2Cl2, (R = H 38%, 74:26 dr), (R = Ac 88%, 80:20 dr), (R = Bn, 73%, 60:40 dr); h) KOHMeOH, reflux (see
Table 2).

4370
F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
Table 2. Intramolecular aldol condensation for anti-analogs
Starting material Conditions Product
Yield (%) dra
OMe
a
OMe
S
H
N
H
N
MeS
(2S,4R)-12
(2S,4R)-13
(2S,4R)-14
KOHMeOH (2R,8aR)-15 89
KOHMeOH (2R,8aR)-15 75
KOHEtOH (2R,8aR)-16 44
70:30
64:36
60:40
HO
O
O
O
a Diastereomeric ratios determined by 1H NMR.
(2R,8aS)-15
(2R,8a
OMe
S)-17
Table 3. Intramolecular aldol condensation for syn-analogs
H
N
Starting material Conditions Product
Yield (%) dra
b
(2R,4R)-13
(2R,4R)-13
(2R,4R)-14
KOHEtOH (2R,8aS)-15 64
KOHMeOH (2R,8aS)-15 73
KOHEtOH (2R,8aS)-16 28
56:44
56:44
50:50
O
a Diastereomeric ratios determined by 1H NMR.
(R)-2
Scheme 6. Synthesis of (R)-2. Reagents and conditions: (a) 95% NaH.
carbon disulfide, MeI, dry THF, 24 °C, 1 h 45 min (85%, 99:1 dr)a; (b)
Bu3SnH, AIBN, dry toluene, reflux, 3 h (85%, 99:1 er)a. aDid not
exhibit the anticipated 100:0 dr and er presumably due to incomplete
resolution of the hydroxy precursor.
expected, the anti-substrates incorporated the more
energetically favorable stereochemistry for the aldol
condensation whereas the syn-substrates were more
prone to racemization. Despite the incorporation of
the larger O-benzyl group, analogs (2S,4R)-14 and
(2R,4R)-14 surprisingly showed the lowest dr for both
the anti- and syn-stereochemical situations respectively.
supply enantiopure (S)-2 in 61% yield. The optical rota-
tion indicated that (S)-2 was the biologically inactive
antipode (+)-2 allowing us to identify the stereochemis-
try of the biologically active (ꢀ)-2 stereoisomer as (R)-2.
Overall, from the commercially available homolog 7, the
synthesis of 15 was achieved in four steps and in 28%
overall yield (via intermediate 12) for the diastereomeric
mixture and in 15% overall yield for the pure diastereo-
mer (2R,8aR)-15. When protected as the acetate, the
synthesis of 15 can be accomplished over five steps
(via intermediate 13) in 54% overall yield as a mixture
of diastereomers of 15 and in 28% overall yield for the
pure diastereomer (2R,8aR)-15. The results for the
O-benzyl analogs led us not to pursue them further.
The synthesis of (R)-2 was achieved by subjecting the
previously isolated minor diastereomer (2R,8aS)-15 to
the same Barton–McCombie deoxygenation protocol.
Thus, (2R,8aS)-15 was treated with NaH, carbon disul-
fide, and iodomethane in dry THF to furnish the S-
methyl xanthate (2R,8aS)-17 (Scheme 6). The synthesis
was successfully completed under radical conditions to
deliver the biologically active enantiomer (ꢀ)-(R)-2.
With a route to the pure (2R,8aR)-15 diastereomer real-
ized, a Barton–McCombie deoxygenation protocol was
followed to provide the target molecule (S)-2.9 As such,
(2R,8aR)-15 was converted into the S-methyl xanthate
(2R,8aR)-17 (Scheme 5). Completion of the synthesis
of (S)-2 was accomplished under radical conditions to
In summary, it was determined that stereochemistry in-
deed plays a significant role in the biological activity for
the lead indolizidinone ( )-2. A stereospecific synthesis
was developed for (S)-2 to determine if this indolizidi-
none, which shares the same stereochemistry with the
natural product tyloindicine I, is the cytotoxic enantio-
mer. This endeavor identified enantiomer (S)-2 as the
biologically inactive (ꢀ)-2 isomer. As a consequence of
this synthetic study, it was determined that the R-stereo-
chemistry was needed for biological activity. As such,
antipode (R)-2 was also generated using this route
although conceptually (2R,4S)-6 could be employed as
the starting substrate to more efficiently prepare
(ꢀ)-(R)-2. Our results suggest that the potentially novel
cancer target involved may have a very ligand-specific
binding pocket and thus be susceptible to potent and
directed inhibition by these specific substrates.
OMe
a
OMe
S
H
N
H
N
MeS
HO
O
O
O
(2R,8aR)-15
(2R,8a
OMe
R)-17
H
N
b
2. Experimental section for Scheme 2
O
(S)-2
2.1. (S)-tert-butyl 2-(2-(4-Methoxyphenyl)-2-oxoethyl)-
pyrrolidine-1-carboxylate (5)
Scheme 5. Synthesis of (S)-2. Reagents and conditions: (a) 95% NaH,
carbon disulfide, MeI, dry THF, 24 °C, 1 h (76%, 100:0 dr); (b)
Bu3SnH, AIBN, dry toluene, reflux, 3 h (61%, 100:0 er).
Step 1: N-BOC-L-homoproline 4 (10.92 g, 47.62 mmol)
was dissolved in CH2Cl2 and cooled to ꢀ15 °C. To this

F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
4371
mixture were added N,O-dimethylhydroxylamineÆHCl
2.2. 1-(4-Methoxyphenyl)-2-(1-(2-phenylacetyl)-pyrroli-
din-2-yl)ethanone (3)
(9.848 g, 101.0 mmol), NMM (11.66 mL, 101.0 mmol),
and EDCI (19.35 g, 101.0 mmol). The reaction solution
was allowed to come to room temperature over 2 h
after which it was again cooled to 0 °C. The reaction
was quenched by the addition of an ice-cold 10%
HCl solution and allowed to stir at this temperature
for 5 min. The reaction mixture was diluted with water
and extracted with CH2Cl2. The combined organic lay-
ers were washed with saturated sodium bicarbonate
and then dried over Na2SO4, filtered, and concentrated.
Silica gel column chromatography (33% EtOAc/hex-
anes) provided (S)-tert-butyl 2-(2-(methoxy(methyl)
amino)-2-oxoethyl)pyrrolidine-1-carboxylate in 91%
yield (11.78 g) as a colorless oil. Rotamers were evident
using NMR spectroscopy: 1H NMR (400 MHz,
CDCl3) d 1.38 (s, 9H), 1.62–1.82 (m, 3H), 1.93–2.00
(m, 1H), 2.32–2.38 (m, 1H), 2.87 (bs, 1H), 3.15 (s,
3H), 3.38 (bs, 2H), 3.64 (s, 3H), 4.15 (bs, 1H); 13C
NMR (100 MHz, CDCl3) d 23.6, 28.9, 31.5, 36.7,
46.8, 54.2, 56.8, 61.5, 79.5, 154.5, 172.7; IR 3564,
2973, 2879, 1694, 1394, 1170, 1110 cmꢀ1; HRMS
Carbamate 5 (0.100 g, 0.313 mmol) was taken into 4 M
HCl/dioxane (2 mL) and stirred at room temperature
for 20 min. The solvent was removed under reduced
pressure and the crude amine salt was allowed to sit un-
der high vacuum for 12 h. The amine salt was then dis-
solved in CH2Cl2 (4 mL) with stirring and cooled to
ꢀ15 °C. Solid phenylacetic acid (0.0512 g, 0.376 mmol),
NMM (0.037 mL, 0.34 mmol), and EDCI (0.0720 g,
0.376 mmol) were added to the reaction vessel and stir-
ring was continued for 2 h. The reaction mixture was di-
luted with water and the organic product was extracted
with CH2Cl2 and dried over Na2SO4. The solvent was
removed in vacuo and the product purified via flash col-
umn chromatography on silica gel (25% EtOAc/hex-
anes) to provide the pure amide 3 in 74% yield
(0.078 g) as a colorless oil: 1H NMR (400 MHz, CDCl3)
d 1.93 (m, 4H), 2.60 (q, 1H), 3.45 (m, 1H), 3.47 (m, 1H),
3.69 (s, 2H), 3.87 (s, 3H), 3.89 (m, 1H), 4.56 (m, 1H),
6.94 (d, J = 8.80 Hz, 2H), 7.34 (m, 5H), 8.11 (d,
J = 8.80 Hz, 2H); 13C NMR (100 MHz, CDCl3) d
24.32, 29.64, 42.49, 42.92, 47.79, 55.64, 55.86, 114.20,
114.26, 127.24, 129.09, 129.33, 129.43, 130.10, 130.67,
(ES+) m/e calcd for [M+H]+ C13H25N2O4: 273.1814,
22
D
found 273.1819; ½aꢁ ꢀ58 (c = 0.76, CHCl3). Step 2:
In a 25 mL two-necked round-bottom flask equipped
with a reflux condenser was placed activated magne-
sium(0) turnings (0.089 g, 3.67 mmol). THF (3 mL)
was added and stirring was commenced. Through the
side neck port was injected p-bromoanisol (0.23 mL,
1.8 mmol) and the resultant solution was heated to-
ward reflux. Before appreciable heating had taken
place, a catalytic amount of iodine (2 crystals) was
added and the solution immediately turned brown
but quickly dissipated back to the originally clear solu-
tion. As the solution was heated near 60 °C, the radical
initiation process catalyzed by iodine commenced and
the reaction solution bubbled violently for 2–5 min.
The Grignard reagent preparation was allowed to pro-
ceed over 1 h and the reaction mixture was then cooled
to ꢀ5 °C. In a separate flask was dissolved (S)-tert-bu-
131.37, 135.14, 163.98, 170.22, 197.96, 203.30; MS (EI)
22
D
[M+H]+ C21H24NO3: m/z 338; ½aꢁ +7.4 (c = 1.0,
CHCl3).
2.3. 7-(4-Methoxyphenyl)-6-phenyl-2,3,8,8a-tetrahydro-
indolizin-5(1H)-one ( )-2)
Compound 3 (0.500 g, 1.5 mmol) was added to 5% (w/v)
potassium hydroxide in 200 proof EtOH (125 mL,
111 mmol) and the reaction mixture was heated to
78 °C for 5 h. After cooling the solvent was removed
in vacuo, and the crude residue was dissolved in CH2Cl2,
washed twice with water and once with brine. The or-
ganic layer was dried over Na2SO4, filtered, and concen-
trated. Silica gel column chromatography (25% EtOAc/
hexanes to 50% EtOAc/hexanes) afforded the desired
product ( )-2 (0.383 g, 80%) as an off-white solid. mp
tyl
2-(2-(methoxy(methyl)amino)-2-oxoethyl)pyrroli-
dine-1-carboxylate (0.100 g, 0.367 mmol) in THF
(1 mL) at ꢀ5 °C. The Grignard reagent solution was
added to the Weinreb amide solution and the reaction
mixture was stirred for 2 h. Ice-cold 10% HCl (1 mL)
was used to quench the reaction mixture to pH 2–4.
EtOAc was used to extract the crude product and the
combined organic fractions were dried over Na2SO4
and condensed under vacuum. Column chromatogra-
phy with silica gel (10% EtOAc/hexanes) was employed
to isolate the desired product 5 in 87% yield (0.102 g)
as a colorless oil. Rotamers were evident by NMR
1
184.8–186.4 °C (acetone); H NMR (400 MHz, CDCl3)
d 1.73 (m, 1H), 1.88 (m, 1H), 2.08 (m, 1H), 2.29 (m,
1H), 2.77 (m, 2H), 3.59–3.73 (m, 2H), 3.73 (s, 3H),
3.94 (m, 1H), 6.46 (d, J = 7.24 Hz, 2H), 6.66 (d,
J = 8.72 Hz, 2H), 7.14 (m, 5H); 13C NMR (100 MHz,
CDCl3) d 23.59, 34.23, 37.94, 45.20, 55.55, 56.00,
113.68, 127.05, 127.98, 130.31, 131.55, 132.70, 132.75,
136.79, 145.11, 159.21, 164.54; IR 1033, 1247, 1434,
1604, 1643, 2925, 2962 cmꢀ1; HRMS (ES+) m/z calcd
for [M+H]+ C21H22NO2: 320.1651, found 320.1635.
1
spectroscopy. H NMR (400 MHz, CDCl3) d 1.48 (s,
9 H), 1.86 (m, 3H), 2.05 (m, 1H), 2.78 (m, 1H), 3.42
(m, 2H), 3.73-3.91 (m, 1H), 3.73 (s, 3H), 4.32 (t, 1H),
6.95 (d, J = 8.68 Hz, 2H), 7.99 (q, 2H); 13C NMR
(100 MHz, CDCl3) d 23.17, 23.93, 28.95, 30.55, 31.65,
43.24, 46.63, 47.04, 54.75, 55.03, 55.85, 79.56, 80.06,
114.13, 130.38, 130.88, 131.11, 154.82, 163.84, 163.97,
197.63, 198.12; IR 1168, 1259, 1396, 1600, 1689,
2877, 2935, 2972 cmꢀ1; HRMS (ES+) m/z calcd for
3. Experimental section for Scheme 4
3.1. N-BOC-L-(4-hydroxy)homoproline (7)
Warning: Large amounts of diazomethane were used for
this transformation. Proper care should be taken when
handling this highly explosive reagent. All glassware used
were free of cracks, scratches or ground-glass joints and a
blast shield was used. N-BOC- -(4-hydroxy)proline
[M+Na] C18H25NO4Na: 341.1681, found 342.1672;
22
D
½aꢁ ꢀ18 (c = 1.0, CHCl3).

4372
F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
(20.00 g, 86.49 mmol) was taken into THF (250 mL)
with stirring and cooled to 0 °C with an ice-bath. The
reaction solution was treated with triethylamine
(12.04 mL, 86.42 mmol) and allowed to react for
15 min to fully deprotonate the carboxylic acid. With
the addition of ethyl chloroformate (8.23 mL,
126 mmol), the resultant anhydride product precipitated
out of solution as a thick white solid. Stirring was con-
tinued for 15 min and then stopped. In a separate flask,
an ice-cold solution of diazomethane was prepared (see
procedure below) and, without stirring, was carefully
decanted into the freshly prepared anhydride reaction
flask using a glass funnel. The reaction solution was
lightly stirred for 4 s then stirring was stopped. The mix-
ture was allowed to warm to room temperature and re-
act overnight. Any unreacted diazomethane was
carefully quenched with acetic acid (0.5 N, 69 mL).
The drop-wise addition of saturated sodium bicarbonate
regulated the solution back to a basic pH 8–9 with gentle
stirring. The organic and aqueous layers were separated.
The organic phase was washed twice each with saturated
sodium bicarbonate and brine then dried over Na2SO4.
The solvent was evaporated under reduced pressure
and placed under high vacuum overnight. The crude
diazo intermediate was taken into water (150 mL) and
THF (1500 mL) and cooled to ꢀ25 °C and stirred for
30 min. The reaction solution turned into a thick slush.
Foil was used to cover the reaction flask so as to exclude
light from the reaction solution and silver benzoate
(2.56 g, 11.2 mmol), dissolved in triethylamine
(40.0 mL, 287 mmol), was then added. The reaction
temperature was allowed to slowly warm to room tem-
perature and the solution was stirred overnight. The
reaction mixture was evaporated to dryness and the res-
idue stirred for 1 h with a saturated sodium bicarbonate
solution (100 mL). The resultant black solution was par-
titioned with water and EtOAc (100 mL). The organic
layer with black suspension was collected and washed
with brine once then twice with saturate sodium bicar-
bonate. The organic phase was set aside and not used
again. All aqueous layers were combined in a large
Erlenmeyer flask with a large stir bar and cooled to
0 °C. EtOAc (100 mL) was added and stirring was com-
menced. The drop-wise addition of 5 N HCl to pH 2 re-
sulted in the emission of carbon dioxide gas. The
organic layer was separated and the aqueous phase
was extracted an additional three times with EtOAc.
The combined organic layers were dried over Na2SO4
and the solvent was removed in vacuo. Flash column
chromatography with silica gel (33% EtOAc/hexanes)
was used to purify homolog 7 from its progenitor
N-BOC-L-(4-hydroxy)proline in 27% yield (10.92 g) as
a colorless oil. At least two purifications were performed
with iodine stain utilized to monitor the product on
TLC. Diazomethane preparation. Warning: Glassware
were free of cracks, scratches, and ground glass joints.
An ice-cold solution of potassium hydroxide (80.00 g,
1425 mmol) in water (200 mL) was prepared and cold
ether (1000 mL) followed by 1-methyl-3-nitro-1-nitroso-
guanidine (MNNG, 50.89 g, 346.0 mmol) were added.
With the addition of MNNG the basic solution turned
clear yellow and bubbled moderately as the reagent dis-
solved and volatile diazomethane was generated. The
solution was reacted for 2 min without stirring. Using
a glass funnel, the mixture was taken into a separatory
funnel and the bottom aqueous layer was collected
and set in a negative pressure hood overnight. The yel-
low organic layer was collected in an ice-cold Erlen-
meyer flask which was layered on the bottom with
potassium hydroxide pellets to absorb any remaining
moisture. The diazomethane solution was used immedi-
ately. The remaining aqueous layer was carefully
quenched the following day with dilute acid. All other
glassware were allowed to sit in a negative pressure hood
and was carefully washed with dilute acid. For product
7, rotameric peaks were evident for NMR spectroscopy
1
methods: H NMR (500 MHz, CDCl3) d 1.41 (s, 9H),
1.79 (m, 3H), 2.03 (m, 1H), 2.28 (q, 1H), 2.76 (bd,
J = 13.65 Hz, 0.5H), 2.93 (bd, J = 15 Hz, 0.5H), 3.31
(bs, 2H), 4.08 (m, 1H), 7.61 (s, 1H); 13C NMR
(125 MHz, CDCl3) d 22.61, 23.31, 28.13, 28.30, 28.45,
30.67, 31.23, 38.52, 39.02, 46.08, 46.48, 53.79, 79.65,
79.88, 154.36, 176.24.
3.2. (2S,4R)-tert-butyl 4-Hydroxy-2-(2-(4-methoxy-
phenyl)-2-oxoethyl)pyrrolidine-1-carboxylate (8)
Step 1: The N-BOC-L-(4-hydroxy)homoproline
7
(3.115 g, 12.70 mmol) was dissolved in CH2Cl2
(180 mL) and cooled to ꢀ15 °C. To this mixture
was added N,O-dimethylhydroxylamine.HCl (2.617 g,
26.93 mmol), NMM (3.006 mL, 27.31 mmol), and
EDCI (5.158 g, 26.93 mmol). The reaction solution
was allowed to come to room temperature over 2 h
after which it was again cooled to 0 °C. The reaction
mixture was quenched by the addition of an ice cold
10% HCl solution and allowed to stir at this temper-
ature for 5 min. The reaction mixture was diluted
with water and extracted with CH2Cl2. The combined
organic layers were washed with saturated sodium
bicarbonate and then dried over Na2SO4, filtered,
and concentrated. Silica gel column chromatography
(33% EtOAc/hexanes) was used to purify the desired
intermediate (2S,4R)-tert-butyl 4-hydroxy-2-(2-(meth-
oxy(methyl)amino)-2-oxoethyl)pyrrolidine-1-carboxyl-
ate in 98% yield (3.556 g) as a colorless oil. Rotamers
were evident using NMR spectroscopy. The alcohol
proton was found to exhibit variable chemical shifts
depending on product concentration: 1H NMR
(400 MHz, CDCl3) d 1.34 (s, 9H), 1.86 (bs, 1H),
2.16 (bs, 1H), 2.40 (bs, 1H), 3.01–3.09 (m, 4H),
3.37–3.51 (bs, 3H), 3.62 (s, 3H), 4.23 (bs, 1H), 4.31
(t, 1H); 13C NMR (100 MHz, CDCl3,
d 13.53,
18.94, 20.82, 28.37, 29.49, 30.45, 31.86, 36.31, 37.31,
40.13, 40.54, 52.70, 54.20, 54.50, 61.08, 64.19, 68.67,
69.15, 79.17, 79.61, 95.94, 154.54, 171.13, 172.07; IR
1120, 1159, 1365, 1400, 1668, 1691, 2935, 2974,
3434; HRMS (ES+) m/e calcd for [M+H]+
22
D
C13H25N2O5: 289.1764, found 289.1759; ½aꢁ ꢀ77
(c = 1.0, CHCl3). Step 2: In a 100 mL two-necked
round-bottom flask equipped with a reflux condenser
were placed activated Mg(0) turnings (5.000 g,
205.8 mmol). THF (50 mL) was added and stirring
was commenced. Through the side neck port was in-
jected 4-bromoanisol (10.6 mL, 193 mmol) and the
resultant solution was heated toward reflux. Before

F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
4373
appreciable heating had taken place,
a
catalytic
calcd for [M+H]+ C20H28NO6: 378.1917, found
378.1905; ½aꢁ ꢀ25 (c = 1.0, CHCl3).
22
D
amount of iodine (4 crystals) was added and the solu-
tion immediately turned brown but quickly dissipated
back to the originally clear solution. As the solution
was heated near 60 °C, the radical initiation process
catalyzed by iodine commenced and the reaction solu-
tion bubbled violently for 2–5 min. The Grignard re-
agent preparation was allowed to proceed over 1 h
and was then cooled to ꢀ5 °C. In a separate flask
was dissolved the intermediate (2S,4R)-tert-butyl 4-hy-
droxy-2-(2-(methoxy(methyl)amino)-2-oxoethyl)pyrroli-
dine-1-carboxylate (4.898 g, 16.99 mmol) in THF
(200 mL) at ꢀ5 °C. The Grignard reagent solution
was added to the Weinreb amide solution and the
reaction mixture was stirred for 2 h. Ice cold 10%
HCl was used to quench the reaction mixture to
pH 2–4. EtOAc was used to extract the crude product
and the combined organic fractions were dried over
Na2SO4 and condensed under vacuum. Column chro-
matography with silica gel (66% EtOAc/hexanes) was
employed to isolate the desired product 8 in 85%
yield (100:0 dr) (4.846 g) as a colorless oil. Rotamers
were evident by NMR spectroscopy: mp 105–109 °C;
1H NMR (500 MHz, CDCl3) d 1.44 (s, 9H), 1.89
(bs, 1H), 2.19 (bs, 1H), 2.80 (bs, 1H), 3.04 (bs,
1H), 3.42 (s, 1.5H), 3.58 (bs, 1H), 3.84 (s, 3.5H),
4.38 (s, 2H), 6.89 (d, J = 8.45 Hz, 2H), 7.93 (bd,
J = 10.85 Hz, 2H); 13C NMR (125 MHz, CDCl3) d
28.42, 39.90, 40.60, 43.02, 44.06, 52.69, 53.27, 54.33,
54.58, 55.32, 68.88, 69.33, 79.43, 79.97, 96.00,
111.44, 113.66, 129.79, 130.01, 130.37, 130.59,
154.67, 163.49, 196.89, 197.605; IR 1170, 1259,
1366, 1401, 1600, 1673, 2359, 2934, 2974,
3.4. (2S,4R)-tert-Butyl 4-(benzyloxy)-2-(2-(4-methoxy-
phenyl)-2-oxoethyl)pyrrolidine-1-carboxylate (10)
Toluene (8 mL) was used to dissolve the carbamate 8
(0.100 g, 0.298 mmol, 100:0 dr) with stirring. Foil was
wrapped around the reaction vessel so as to exclude
light. The addition of benzyl bromide (0.284 mL,
2.39 mmol) and Ag2O (0.691 g, 2.98 mmol) followed
and the mixture was allowed to react at room tempera-
ture for 3 h and 20 min. The reaction mixture was
quenched with water and the organic product was ex-
tracted with EtOAc. The organic phase was dried over
Na2SO4 and filtered to remove any remaining silver re-
agent/salts and was then evaporated under vacuum.
The desired product 10 was isolated using flash column
chromatography with 10% EtOAc/hexanes as mobile
phase in 74% yield (100:0 dr) (0.094 g) as a colorless
oil. Rotamers could be visualized using NMR spectros-
copy: mp 61–62 °C; 1H NMR (400 MHz, CDCl3) d 1.48
(s, 9 H), 2.00 (bs, 1H), 2.30 (bs, 1H), 2.80 (q, 1H), 3.40–
3.89 (m, 4H), 3.87 (s, 3H), 4.41–4.51 (m, 3H), 6.93 (d,
J = 8.88 Hz, 2H), 7.32 (m, 5H), 7.97 (bd, 2H); 13C
NMR (100 MHz, CDCl3, trace impurities) d 23.06,
25.68, 28.93, 32.00, 35.06, 37.10, 38.55, 43.63, 44.65,
51.14, 51.51, 52.36, 53.92, 55.87, 71.30, 76.16, 76.70,
79.94, 80.42, 114.17, 128.09, 128.83, 130.33, 131.00,
131.06, 138.41, 155.03, 163.96, 197.32, 197.93; IR
1112, 1157, 1170, 1259, 1365, 1396, 1600, 1689, 2933,
2974 cmꢀ1; HRMS (ES+) m/e calcd for [M+H]+
22
D
C25H32NO5: 426.2280, found 426.2290; ½aꢁ ꢀ16
3437 cmꢀ1; HRMS (ES+) m/e calcd for [M+H]+
(c = 1.0, CHCl3).
22
D
C18H26NO5: 336.1811, found 336.1783; ½aꢁ ꢀ33
(c = 2.6, CHCl3).
3.5. (2S,4R)-2-(4-Hydroxy-1-(2-phenylacetyl)pyrrolidin-
2-yl)-1-(4-methoxyphenyl)ethanone ((2S,4R)-12) and
(2R,4R)-2-(4-hydroxy-1-(2-phenylacetyl)pyrrolidin-2-yl)-
1-(4-methoxyphenyl)ethanone ((2R,4R)-12)
3.3. (2S,4R)-tert-Butyl 4-acetoxy-2-(2-(4-methoxy-
phenyl)-2-oxoethyl)pyrrolidine-1-carboxylate (9)
Pyridine (2 mL, xs) was used to dissolve substrate 8
(0.091 g, 0.27 mmol, 100:0 dr). With stirring, the reac-
tion mixture temperature was lowered to 0 °C and
DMAP (0.033 g, 0.27 mmol) and acetic anhydride
(0.13 mL, 1.4 mmol) were added. The reaction mixture
was allowed to come to room temperature over 1 h
and the reaction was then quenched with water. EtOAc
was used to extract the organic product. The organic
layer was dried over Na2SO4 and the solvent removed
under reduced pressure. Silica gel column chromatogra-
phy (20% EtOAc/hexanes) was used to isolate the de-
sired product 9 in 99% yield (100:0 dr) (0.101 g) as a
colorless oil. Rotamers were evident using NMR spec-
Carbamate 8 (100:0 dr) (1.904 g, 5.677 mmol) was taken
into 4 M HCl/dioxane (15 mL) and stirred at room tem-
perature for 30 min. With the addition of the acid, the
reaction solution changed colors from a cloudy orange
to a cloudy scarlet within 5 min. The solvent was re-
moved under reduced pressure and the crude amine salt
was allowed to sit under high vacuum overnight and ap-
peared as a magenta solid. The amine salt was then dis-
solved in CH2Cl2 (35 mL) with stirring and cooled to
ꢀ15 °C. Solid phenylacetic acid (0.927 g, 6.81 mmol),
NMM (0.686 mL, 6.25 mmol) and EDCI (1.306 g,
6.812 mmol) were added to the reaction vessel and stir-
ring was continued for 2 h. The orange reaction mixture
was quenched with 10% HCl (2 mL) and diluted with
water. The organic product was extracted with CH2Cl2
three times. As the organic layer was washed twice with
saturated sodium bicarbonate the cloudy orange organic
layer color was observed to change to a clear yellow col-
or. The combined saturated sodium bicarbonate washes
were back extracted twice with CH2Cl2. The combined
organic fractions were dried over Na2SO4. The solvent
was removed in vacuo and the two diastereomeric prod-
ucts were purified from a number of impurities via
1
troscopy: H NMR (400 MHz, CDCl3) d 1.46 (s, 9H),
1.97–2.04 (m, 1H), 2.03 (s, 3H), 2.35 (bs, 1H), 2.28 (q,
1H), 3.54–3.89 (m, 3H), 3.86 (s, 3H), 4.38 (m, 1H),
5.20 (bs, 1H), 6.92 (d, J = 8.52 Hz, 2H), 7.96 (bs, 2H);
13C NMR (100 MHz, CDCl3) d 14.57, 21.40, 21.48,
28.86, 30.06, 37.57, 38.78, 43.16, 44.28, 52.16, 52.72,
53.15, 53.55, 55.84, 60.74, 67.44, 72.29, 72.82, 80.17,
114.14, 130.32, 130.51, 130.88, 154.84, 163.96, 170.97,
171.47, 196.96, 197.47; IR 1170, 1246, 1367, 1398,
1600, 1692, 1739, 2930, 2974 cmꢀ1; HRMS (ES+) m/e

4374
F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
precipitations in 80% EtOAc/hexanes. The precipitate
was collected by filtration to provide the diastereomeric
product 12 in 38% yield (74:26 dr favoring the anti-dia-
stereomer) (0.759 g). The diastereomers were resolved
using flash column chromatography (1.5% methanol/
CH2Cl2). The anti-diastereomer proved to be slightly
more polar than the syn-counterpart. Rotamers could
be visualized using NMR spectroscopy: For (2S,4R)-12
reomer): colorless crystals, mp 108 °C; 1H NMR
(500 MHz, CDCl3) d 1.89 (s, 3H), 2.05 (m, 1H), 2.31
(m, 1H), 2.92 (q, 1H), 3.63 (s, 2H), 3.63–3.69 (m, 2H),
3.83 (s, 3H), 3.88 (q, 1H), 4.59 (m, 1H), 5.19 (s, 1H),
6.90 (d, J = 8.55 Hz, 2H), 7.28 (m, 5H), 7.98 (d,
J = 8.60 Hz, 2H); 13C NMR (125 MHz, CDCl3) d
20.84, 36.24, 41.82, 42.68, 52.74, 53.46, 55.36, 72.53,
113.69, 126.78, 128.61, 128.74, 129.69, 130.51, 134.34,
163.54, 169.91, 170.38, 196.79; IR 1026, 1172, 1226,
1247, 1259, 1311, 1371, 1421, 1600, 1643, 1668, 1737,
2839, 2906, 2935, 3029 cmꢀ1; HRMS (ES+) m/e calcd
1
(major diastereomer): colorless crystals, mp 125 °C; H
NMR (500 MHz, CDCl3) d 2.03 (m, 1H), 2.14 (m,
1H), 2.76 (q, 0.92H, rotameric), 3.06 (d, J = 6.80 Hz,
0.08H, rotameric), 3.49 (d, J = 10.90 Hz, 1H), 3.58 (m,
1H), 3.65 (d, J = 5.2 Hz, 2H), 3.85 (s, 3H), 3.93 (q,
1H), 4.46 (bs, 1H), 4.62 (m, 1H), 6.91 (d, J = 8.80 Hz,
2H), 7.24–7.33 (m, 5H), 7.77 (d, J = 8.65 Hz, 0.16H,
rotameric), 8.01 (d, J = 8.80 Hz, 1.84H, rotameric); 13C
NMR (125 MHz, CDCl3) d 39.27, 42.23, 42.67, 53.78,
55.47, 69.94, 113.81, 126.88, 128.70, 128.98, 129.41,
130.79, 134.61, 163.66, 170.37, 197.34; IR 1178, 1257,
for [M+H]+ C23H26NO5: 396.1811, found 396.1810;
22
½aꢁ +10.0 (c = 1.80, CHCl3). For (2R,4R)-13 (minor dia-
D
stereomer), colorless oil: 1H NMR (500 MHz, CDCl3) d
2.07 (m, 4H), 2.25 (m, 0.8H, rotameric), 2.35 (m, 0.2H,
rotameric), 2.95 (q, 1H), 3.66 (m, 3H), 3.79 (q, 1H), 3.81
(s, 2.4H, rotameric), 3.87 (s, 0.6H, rotameric), 3.97 (q,
1H), 4.67 (t, 0.8H), 4.70 (t, 0.2H), 5.30 (s, 0.8H, rotamer-
ic), 5.35 (s, 0.2H, rotameric), 6.94 (d, J = 8.80 Hz, 2H),
7.28 (m, 5H), 7.85 (d, J = 8.80, 0.4H, rotameric), 8.05
(d, J = 8.75 Hz, 1.6 Hz, rotameric); 13C NMR
(125 MHz, CDCl3) d 21.12, 34.63, 41.88, 42.39, 53.38,
54.37, 55.37, 73.56, 113.73, 126.89, 128.64, 128.96,
129.69, 130.60, 134.10, 163.53, 169.69, 170.26, 197.20;
IR 1170, 1230, 1247, 1375, 1417, 1598, 1643, 1668,
1737, 2840, 2937, 3028, 3060 cmꢀ1; HRMS (ES+) m/e
calcd for [M+H]+ C23H26NO5: 396.1811, found
396.1812.
1311, 1438, 1602, 1618, 1662, 2850, 2923, 3438 cmꢀ1
;
HRMS (ES+) m/e calcd for [M+H]+ C21H24NO4:
354.1705, found 354.1714. For (2R,4R)-12 (minor dia-
stereomer), colorless oil: H NMR (500 MHz, CDCl3)
1
d 1.99 (m, 1H), 2.18 (m, 1H), 3.01 (q, 0.17H, rotameric),
3.20 (q, 0.83H, rotameric), 3.58–3.69 (m, 4H), 3.85 (s,
2.49H, rotameric), 3.87 (s, 0.51H, rotameric), 3.92 (q,
1H), 4.50 (m, 1H), 4.59 (t, 0.83H rotameric), 4.70 (t,
0.17H, rotameric), 6.90 (d, J = 8.85 Hz, 2 H), 7.28 (m,
5H), 7.85 (d, J = 8.80 Hz, 0.4H, rotameric), 8.03 (d,
J = 8.80 Hz, 1.6H, rotameric); 13C NMR (125 MHz,
CDCl3) d 38.08, 42.27, 42.50, 54.57, 55.46, 55.80,
71.27, 113.76, 126.85, 128.97, 129.40, 129.82, 130.83,
134.42, 163.65, 169.99, 198.32; IR 1170, 1259, 1311,
1419, 1598, 1668, 2852, 2923, 3388 cmꢀ1; HRMS
(ES+) m/e calcd for [M+H]+ C21H24NO4: 354.1705,
found 354.1710.
3.7. 2-((2S,4R)-4-(Benzyloxy)-1-(2-phenylacetyl)pyrroli-
din-2-yl)-1-(4-methoxyphenyl) ethanone ((2S,4R)-14) and
2-((2R,4R)-4-(benzyloxy)-1-(2-phenylacetyl)pyrrolidin-2-
yl)-1-(4-methoxyphenyl)ethanone ((2R,4R)-14)
Carbamate 10 (0.654 g, 1.54 mmol, 100:0 dr) was taken
into 4 M HCl/dioxane (10 mL) and stirred at room tem-
perature for 30 min. The solvent was removed under re-
duced pressure and the crude amine salt was allowed to
sit under high vacuum overnight. The amine salt was
then dissolved in CH2Cl2 (30 mL) with stirring and
cooled to ꢀ15 °C. Solid phenylacetic acid (0.252 g,
1.85 mmol), NMM (0.170 mL, 1.58 mmol) and EDCI
(0.351 g, 1.83 mmol) were added to the reaction vessel
and stirring was continued for 2 h. The reaction mixture
was diluted with water and the organic product was ex-
tracted with CH2Cl2 and dried over Na2SO4. The sol-
vent was removed in vacuo and the product purified
via preparative TLC (50% EtOAc/hexanes; multiple
runs) to provide the pure amide 14 in 73% yield (60:40
dr favoring the anti-diastereomer) (0.496 g). The anti-
product could also be obtained via recrystallization
from EtOAc or isopropanol. The yield reflects the dia-
stereomeric mix although the slightly less polar anti-dia-
stereomer and more polar syn-counterpart were
successfully resolved. Rotamers could be visualized
using NMR spectroscopy. For (2S,4R)-14 (major diaste-
reomer): colorless crystals, mp 119 °C; 1H NMR
(400 MHz, CDCl3) d 2.07 (m, 1H), 2.29 (m, 1H), 2.78
(q, 1H), 3.58–3.64 (m, 4H), 3.87 (s, 3H), 3.95 (q, 1H),
4.19 (t, 1H), 4.33 (d, J = 11.88 Hz, 1H), 4.45 (d,
J = 11.96 Hz, 1H), 4.66 (bd, J = 6.44 Hz, 1H), 6.94 (d,
J = 8.80 Hz, 2H), 7.21–7.37 (m, 10H), 8.06 (d,
J = 8.76 Hz, 2H); 13C NMR (100 MHz, CDCl3) d
3.6. (3R,5S)-5-(2-(4-Methoxyphenyl)-2-oxoethyl)-1-(2-
phenylacetyl)pyrrolidin-3-yl acetate ((2S,4R)-13) and
(3R,5R)-5-(2-(4-methoxyphenyl)-2-oxoethyl)-1-(2-phenyl-
acetyl)pyrrolidin-3-yl acetate ((2R,4R)-13)
Carbamate 9 (0.424 g, 1.13 mmol) was taken into 4 M
HCl/dioxane (2 mL) and stirred at room temperature
for 20 min. The solvent was removed under reduced
pressure and the crude amine salt was allowed to sit un-
der high vacuum overnight. The amine salt was then dis-
solved in CH2Cl2 with stirring and cooled to ꢀ15 °C.
Solid phenylacetic acid (0.1823 g, 1.337 mmol), NMM
(0.1356 mL, 1.260 mmol) and EDCI (0.2584 g,
1.348 mmol) were added to the reaction vessel and stir-
ring was continued for 2 h. The reaction mixture was di-
luted with water and the organic product was extracted
with CH2Cl2 and dried over Na2SO4. The solvent was
removed in vacuo and the product purified via prepara-
tive TLC (1.5% methanol/CH2Cl2; 4 runs) to provide
amide 13 in 88% yield (80:20 dr favoring the anti-diaste-
reomer) (0.392 g). The anti-product could also be recrys-
tallized in minimal isopropanol. Although the
purification successfully resolved the slightly less polar
anti-product from the syn-product, the yield reflects
the diastereomeric mix. Rotamers could be visualized
using NMR spectroscopy: For (2S,4R)-13 (major diaste-

F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
4375
1
36.44, 42.80, 43.09, 52.91, 54.36, 55.88, 71.39, 77.06,
114.23, 127.25, 127.98, 128.21, 128.89, 129.12, 129.37,
130.12, 131.25, 135.03, 138.23, 164.08, 170.55, 197.76;
IR 1026, 1099, 1174, 1261, 1602, 1627, 1668, 2837,
2925, 3031, 3060 cmꢀ1; HRMS (ES+) m/e calcd for
colorless oil: H NMR (500 MHz, CDCl3) d 1.70 (m,
1H), 2.18 (q, 1H), 2.61 (t, 1H), 2.75 (q, 1H), 3.56 (d,
J = 13.4 Hz, 1H), 3.65 (s, 3H), 3.71 (m, 1H), 4.05 (br
s, 1H), 4.11 (bt, 1H), 4.47 (s, 1H), 6.58 (d, J = 8.70,
2H), 6.86 (d, J = 8.70 Hz, 2H), 7.07 (m, 5H); 13C
NMR (125 MHz, CDCl3) d 36.28, 41.36, 52.20, 53.15,
54.12, 68.14, 112.29, 125.70, 126.58, 128.74, 130.14,
131.07, 131.18, 135.13, 144.16, 157.87, 163.19; IR
1031, 1178, 1249, 1440, 1510, 1604, 1633, 2929,
3382 cmꢀ1; HRMS (ES+) m/e calcd for [M+H]+
C21H22NO3: 336.1600, found 336.1606; For (2R,8aS)-
15 (syn-diastereomer), colorless oil: 1H NMR
(400 MHz, CDCl3) d 1.85 (m, 1H), 2.51 (m, 1H), 2.75
(q, 1H), 2.94 (t, 1H), 3.67–3.77 (m, 3H), 3.74 (s, 3H),
4.04 (m, 1H), 4.52 (t, 1H), 6.66 (d, J = 8.76 Hz, 2H),
6.94 (d, J = 8.80 Hz, 2H), 7.11 (m, 2H), 7.18 (m, 3H);
13C NMR (125 MHz, CDCl3) d 37.71, 41.24, 52.53,
53.99, 55.14, 69.43, 113.30, 126.75, 127.63, 129.85,
131.08, 132.08, 132.23, 136.22, 145.43, 158.90, 164.00;
IR 1033, 1178, 1249, 1440, 1510, 1604, 1633, 2923,
3382 cmꢀ1; HRMS (ES+) m/e calcd for [M+H]+
C21H22NO3: 336.1600, found 336.1618.
[M+H]+ C28H30NO4: 444.2175, found 444.2173;
22
D
½aꢁ +8 (c = 1, CHCl3). For (2R,4R)-14 (minor diastereo-
1
mer), colorless oil: H NMR (400 MHz, CDCl3) d 2.06
(m, 1H), 2.25 (d, J = 14.08 Hz, 1H), 2.90 (q, 0.2H, rota-
meric), 3.14 (q, 0.8H, rotameric), 3.63–3.76 (m, 4H),
3.87 (s, 3H), 3.94 (q, 1H), 4.19 (bs, 1H), 4.3944.65 (m,
3H), 6.83 (d, J = 8.84 Hz, 0.4H, rotameric), 6.91 (d,
J = 8.88 Hz, 1.6 H, rotameric), 7.26–7.37 (m, 10H),
7.81 (d, J = 8.84 Hz, 0.4H, rotameric), 8.04 (d,
J = 8.88 Hz, 1.6H, rotameric); 13C NMR (100 MHz,
CDCl3) d 34.18, 42.52, 42.92, 53.84, 53.92, 54.96,
55.84, 71.20, 114.12, 127.29, 128.05, 128.34, 128.92,
129.11, 129.48, 130.33, 131.17, 134.58, 137.98, 163.89,
170.00, 198.34; IR 1027, 1170, 1259, 1417, 1598, 1643,
1666, 2839, 2869, 2935, 3029, 3060 cmꢀ1; HRMS
(ES+) m/e calcd for [M+H]+ C28H30NO4: 444.2175,
found 444.2174.
3.8. (2R,8aR)-2-Hydroxy-7-(4-methoxyphenyl)-6-phenyl-
2,3,8,8a-tetrahydroindolizin-5(1H)-one ((2R,8aR)-15)
and (2R,8aS)-2-hydroxy-7-(4-methoxyphenyl)-6-phenyl-
2,3,8,8a-tetrahydroindolizin-5(1H)-one ((2R,8aS)-15)
3.9. (2R,8aR)-2-(Benzyloxy)-7-(4-methoxyphenyl)-6-phe-
nyl-2,3,8,8a-tetrahydroindolizin-5(1H)-one ((2R,8aR)-16)
and (2R,8aS)-2-(benzyloxy)-7-(4-methoxyphenyl)-6-phe-
nyl-2,3,8,8a-tetrahydroindolizin-5(1H)-one ((2R,8aS)-16)
Via aldol condensation using 2% KOHMeOH(xs). The
diastereomerically pure substrate (2S,4R)-12 (0.385 g,
1.09 mmol) was dissolved in 2% KOHMeOH (xs) and re-
fluxed with stirring for 4 h. The solvent was removed un-
der reduced pressure and the residue partitioned
between water and CH2Cl2. A 10% HCl solution was
added to the biphasic system until a pH 4 had been at-
tained to break apart an emulsion. The organic product
was extracted with CH2Cl2 and dried over Na2SO4. The
solvent was removed in vacuo to give product 15 in 89%
yield (70:30 dr favoring the anti-diastereomer) (0.324 g)
without further purification. The yield reflects the diaste-
reomeric mix although the diastereomers were success-
fully resolved. Via aldol condensation of acetate
protected substrate using 5%KOH (xs) in either ethanol
or methanol: The diastereomerically pure acetate pro-
tected substrate (2S,4R)-13 (0.145 g, 0.367 mmol) was
taken in 5% KOH (24 mL, 22 mmol) dissolved in either
ethanol or methanol and refluxed. The ethanolic solu-
tion was allowed to react for 90 min whereas the meth-
anolic solution was allowed to stir for 4 h. The solvent
was removed in vacuo and the resultant residue was par-
titioned between water and CH2Cl2. The organic prod-
uct was extracted with CH2Cl2 and dried over
Na2SO4. Regardless of the alcoholic solvent used, the
solvent was evaporated and the desired product 15
was isolated via preparative TLC (EtOAc) in 75% yield
(64:36 dr favoring the anti-diastereomer) (0.092 g).
When (2R,4R)-13 was used, the ethanolic solution pro-
vided product 15 in 64% yield (56:44 dr favoring the
syn-diastereomer) (0.018 g) and the methanolic solution
generated the product in 73% yield (56:44 dr favoring
the syn-diastereomer) (0.021 g). The yield reflects the
diastereomeric mix although the diastereomers were suc-
cessfully resolved. For (2R,8aR)-15 (anti-diastereomer),
The aldol precursor substrate (2S,4R)-14 (0.168 g,
0.395 mmol, 100:0 dr) was taken into 5% KOHEtOH
(26.53 mL, 23.69 mmol) and refluxed for 1 h and
45 min. The solvent was removed under reduced pres-
sure and the residue partitioned between water and
CH2Cl2. The organic product was extracted with
CH2Cl2 and dried over Na2SO4. The solvent was evap-
orated under vacuum and product 16 was isolated in
44% yield (60:40 dr favoring the anti-diastereomer)
(0.073 g) using preparative TLC (33% EtOAc/hexanes;
3 runs). When (2R,4R)-14 was used, product 16 could
be isolated in 28% yield (0.022 g) as an equal mix of
the syn- and anti-diastereomers. The yield reflects the
diastereomeric mix although the slightly more polar
anti-diastereomer and less polar syn-counterpart were
successfully resolved. For (2R,8aR)-16 (anti-diastereo-
1
mer), colorless oil: H NMR (500 MHz, CDCl3) d 1.71
(m, 1H), 2.38 (q, 1H), 2.65 (q, 1H), 2.76 (q, 1H), 3.65
(s, 3H), 3.68 (q, 1H), 3.79 (d, J = 13.55 Hz, 1H), 4.21
(m, 2H), 4.43 (d, J = 11.75 Hz, 1H), 4.52 (d,
J = 11.70 Hz, 1H), 6.58 (d, J = 8.75 Hz, 2H), 6.86 (d,
J = 8.75 Hz, 2H), 7.02–7.29 (m, 10H); 13C NMR
(125 MHz, CDCl3) d 36.29, 38.90, 49.54, 52.48, 54.10,
69.71, 74.92, 112.22, 125.66, 126.39, 126.58, 126.83,
127.47, 128.82, 130.07, 131.18, 131.23, 135.13, 136.71,
143.66, 157.75, 163.14; HRMS (ES+) m/e calcd for
[M+H]+ C28H28NO3: 426.2069, found 426.2064. For
1
(2R,8aS)-16 (syn-diastereomer), colorless oil: H NMR
(500 MHz, CDCl3) d 1.95 (m, 1H), 2.55 (m, 1H), 2.76
(q, 1H), 2.89 (q, 1H), 3.72 (s, 3H), 3.77 (d,
J = 6.10 Hz, 2H), 4.02 (m, 1H), 4.26 (m, 1H), 4.53 (d,
J = 11.75 Hz, 1H), 4.60 (d, J = 11.70 Hz, 1H), 6.65 (d,
J = 8.70 Hz, 2H), 6.93 (d, J = 8.65 Hz, 2H), 7.09–7.37
(m, 10H); 13C NMR (125 MHz, CDCl3) d 37.59,
39.21, 49.60, 53.73, 55.12, 71.82, 75.77, 113.23, 126.70,

4376
F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
127.61, 127.78, 127.88, 128.51, 129.88, 131.06, 131.84,
132.11, 136.15, 137.70, 145.15, 158.81, 164.03; HRMS
(ES+) m/e calcd for [M+H]+ C28H28NO3: 426.2069,
found 426.2074.
132.70, 132.75, 136.79, 145.11, 159.21, 164.54; IR
1033, 1247, 1434, 1604, 1643, 2925, 2962 cmꢀ1; HRMS
(ES+) m/z calcd for [M+H]+ C21H22NO2: 320.1651,
found 320.1635.
4. Experimental section for Scheme 5
5. Experimental section for Scheme 6
4.1. (2R,8aR)-O-7-(4-Methoxyphenyl)-5-oxo-6-phenyl-
1,2,3,5,8,8a-hexahydroindolizin-2-yl S-methyl carbonodi-
thioate ((2R,8aR)-17)
5.1. (2R,8aS)-O-7-(4-Methoxyphenyl)-5-oxo-6-phenyl-
1,2,3,5,8,8a-hexahydroindolizin-2-yl S-methyl carbonodi-
thioate ((2R,8aS)-17)
(2R,8aR)-15 (0.024 g, 0.072 mmol), dissolved in THF
(2 mL), was added to a solution of 95% NaH (0.0034
g, 0.14 mmol) in THF (0.5 mL) at ambient temperature
and stirred for 5 min. Carbon disulfide (0.301 mL,
5.01 mmol) and iodomethane (0.312 mL, 5.01 mmol)
were added and stirring was continued for an additional
1 h. The reaction solution was condensed under reduced
pressure. The remaining residue was dissolved in mini-
mal CH2Cl2 and applied directly to a preparative silica
TLC plate for purification. A 33% EtOAc/hexanes mo-
bile phase (2 runs) was used to purify the desired xan-
thate (2R,8aR)-17 in 76% yield (100:0 dr) (0.023 g) as a
(2R,8aS)-15 (0.025 g, 0.075 mmol, 99:1 dr), dissolved in
THF (2 mL), was added to a solution of 95% NaH
(0.0035 g, 0.15 mmol) in THF (1 mL) at ambient tem-
perature and stirred for 5 min. Carbon disulfide
(0.314 mL, 5.22 mmol) and iodomethane (0.325 mL,
5.22 mmol) were added and stirring was continued for
1 h. After monitoring the sluggish reaction progress
via TLC, additional NaH (95%, 3 mg) was added and
the reaction mixture was stirred for 45 min. After a total
reaction time of 1 h and 45 min, the reaction solvent was
removed under vacuum to leave a thick moist white so-
lid. The solid was dissolved in minimal CH2Cl2 and ap-
plied directly to a preparative silica TLC plate for
purification. A 33% EtOAc/hexanes mobile phase (2
runs) was used to purify the desired xanthate
(2R,8aS)-17 in 85% yield (99:1 dr) (0.027 g) as a colorless
oil: 1H NMR (500 MHz, CDCl3) d 2.13 (m, 1H), 2.58 (s,
3H), 2.80–2.93 (m, 3H), 3.72 (s, 3H), 3.88 (q, 1H), 4.05–
4.16 (m, 2H), 5.99 (t, 1H), 6.66 (d, J = 8.40 Hz, 2H), 6.94
(d, J = 8.40 Hz, 2H), 7.09 (d, J = 6.90 Hz, 2H), 7.17 (t,
3H); 13C NMR (125 MHz, CDCl3) d 19.32, 37.68,
38.29, 49.89, 53.76, 55.14, 80.45, 113.32, 126.82,
127.65, 129.89, 131.06, 131.82, 131.90, 135.98, 145.25,
158.97, 164.04, 215.28; IR 1033, 1064, 1180, 1207,
1
colorless oil: H NMR (400 MHz, CDCl3) d 2.05 (m,
1H), 2.60 (s, 3H), 2.65 (q, 1H), 2.80 (q, 1H), 2.90 (q,
1H), 3.75 (s, 3H), 3.93 (d, J = 14.64 Hz, 1H), 4.02 (q,
1H), 4.30 (m, 1H), 6.15 (t, 1H), 6.68 (d, J = 8.88 Hz,
2H), 6.96 (d, J = 8.88 Hz, 2H), 7.13 (m, 2H), 7.18 (m,
3H); 13C NMR (125 MHz, CDCl3) d 18.19, 36.23,
38.71, 50.22, 52.80, 54.13, 79.94, 112.34, 125.81,
126.63, 128.85, 130.08, 131.00, 131.12, 134.90, 143.90,
157.95, 163.35, 214.00; IR 1033, 1051, 1080, 1188,
1213, 1249, 1434, 1510, 1604, 1645, 2837, 2925,
2954 cmꢀ1; HRMS (ES+) m/e calcd for [M+H]+
C23H24NO3S2: 426.1198, found 426.1202.
1249, 1434, 1510, 1604, 1645, 2837, 2927, 3053 cmꢀ1
;
4.2. (+)-(S)-7-(4-Methoxyphenyl)-6-phenyl-2,3,8,8a-tet-
rahydroindolizin-5(1H)-one ((+)-(S)-2)
HRMS (ES+) m/e calcd for [M+H]+ C23H24NO3S2:
426.1198, found 426.1203.
To a two necked flask equipped with a reflux condenser
were added toluene (2 mL) and tributyltinhydride
(0.018 mL, 0.067 mmol). Catalytic AIBN (0.0008 g,
0.005 mmol) was added and heating was commenced to-
ward 75 °C. The S-methyl xanthate (2R,8aR)-17
(0.022 g, 0.052 mmol, 100:0 dr) was dissolved in toluene
(3 mL) and slowly added to the reaction mixture. The
reaction mixture was allowed to progress over 3 h.
Although minimal amounts of the starting xanthate
could be visualized with TLC under UV light, the reac-
tion solvent was removed in vacuo. The remaining resi-
due was dissolved in CH2Cl2 and applied directly to a
preparative silica TLC plate. A mobile phase of 66%
EtOAc/hexanes was used to isolate enantiomerically
5.2. (ꢀ)-(R)-7-(4-Methoxyphenyl)-6-phenyl-2,3,8,8a-tet-
rahydroindolizin-5(1H)-one ((ꢀ)-(R)-2)
To a two necked flask equipped with a reflux condenser
were added toluene (2 mL) and tributyltinhydride
(0.021 mL, 0.076 mmol). Catalytic AIBN (0.0009 g,
0.006 mmol) was added and heating was commenced to-
ward 75 °C. The S-methyl xanthate (2R,8aS)-17
(0.025 g, 0.059 mmol, 99:1 dr) was dissolved in toluene
(3 mL) and slowly added to the reaction mixture. The
reaction mixture was allowed to progress over 3 h then
the reaction solvent was removed in vacuo. The remain-
ing residue was dissolved in CH2Cl2 and applied directly
to a preparative silica TLC plate. A mobile phase of 66%
EtOAc/hexanes was used to isolate the indolizidinone
pure indolizidinone (+)-(S)-2 in 61% yield (0.010 g):
22
mp 199.8–201.7 °C ½aꢁ +42 (c = 1.0, CHCl3); colorless
(R)-2 in 85% yield (99:1 er) (0.016 g): mp 199.5–
D
22
crystals, mp 199.8–201.7 °C (acetone); 1H NMR
(400 MHz, CDCl3) d 1.73 (m, 1H), 1.88 (m, 1H), 2.08
(m, 1H), 2.29 (m, 1H), 2.77 (m, 2H), 3.59–3.73 (m,
2H), 3.73 (s, 3H), 3.94 (m, 1H), 6.46 (d, J = 7.24 Hz,
2H), 6.66 (d, J = 8.72 Hz, 2H), 7.14 (m, 5H); 13C
NMR (100 MHz, CDCl3) d 23.59, 34.23, 37.94, 45.20,
55.55, 56.00, 113.68, 127.05, 127.98, 130.31, 131.55,
201.4 °C ½aꢁ ꢀ42 (c = 1.0, CHCl3); colorless crystals
D
mp 199.5–201.4 °C (acetone); 1H NMR (400 MHz,
CDCl3) d 1.73 (m, 1H), 1.88 (m, 1H), 2.08 (m, 1H),
2.29 (m, 1H), 2.77 (m, 2H), 3.59–3.73 (m, 2H), 3.73 (s,
3H), 3.94 (m, 1H), 6.46 (d, J = 7.24 Hz, 2H), 6.66 (d,
J = 8.72 Hz, 2H), 7.14 (m, 5H); 13C NMR (100 MHz,
CDCl3) d 23.59, 34.23, 37.94, 45.20, 55.55, 56.00,

F. S. Kimball et al. / Bioorg. Med. Chem. 16 (2008) 4367–4377
4377
4. For related investigations of the tetrahydroindolizin-
5(1H)-one core see Sharma, V. M.; Adi Seshu, K. V.;
Vamsee Krishna, C.; Prasanna, P.; Chandra Sekhar, V.;
Venkateswarlu, A.; Rajagopal, S.; Ajaykumar, R.; Deevi,
D. S.; Rao Mamidi, N. V. S.; Rajagopalan, R. Bioorg.
Med. Chem. Lett. 2003, 13, 1679.
113.68, 127.05, 127.98, 130.31, 131.55, 132.70, 132.75,
136.79, 145.11, 159.21, 164.54; IR 1033, 1247, 1434,
1604, 1643, 2925, 2962 cmꢀ1; HRMS (ES+) m/z calcd
for [M+H]+ C21H22NO2: 320.1651, found 320.1635.
5. Chiralcel OJ, elution with 20–50% isopropanol in hexanes,
flow rate 5 mL/min. Retention time for (+)-2 of 15 min.
Retention time of 20.8 min for (ꢀ)-2.
Acknowledgments
22
These studies were supported by the National Cancer
Institute (CA 90602 and N01-CM-67259). B.J.T. was
supported by the Department of Defense Breast Cancer
Predoctoral Fellowship (DAMD17-02-1-0435). We
thank Deborah A. Bohlander for conducting the cyto-
toxicity assays and Dr. Dinah Dutta for the synthesis
of ( )-tylophorine (Table 1).10
6. (+)-2: Colorless crystals, ½aꢁD +42 (c = 1.0 CHCl3), mp
199.8–201.7 °C (acetone); (ꢀ)-2: colorless crystals,
22
½aꢁD ꢀ42 (c = 1.0 CHCl3), mp 199.5–201.4 °C (acetone);
( )-2: colorless crystals, mp 184.8–186.4 °C (acetone).
7. The enantiomeric ratio varied between experiments from
50:50 (racemic) to 60:40 (+)-(S)-2:(ꢀ)-(R)-2.
8. Some degree of racemization was noticed in step e, with
the isolation of 5% yield of the O-benzyl, N-benzyl side-
product. This product was isolated in a 60:40 dr suggest-
ing that perhaps the racemization does not require strong
acid to occur. Furthermore, this side-product also indi-
cated, that either the BOC-protected amine or the free
amine, rather than the amine salt, are prone to
racemization.
References and notes
1. Ali, M.; Ansari, S. H.; Grever, M. R. Pharmazie 2001, 56,
188.
2. Ali, M.; Ansari, S. H.; Qadry, J. S. J. Nat. Prod. 1991, 54,
1271.
9. Barton, D. H. R.; Mc Combie, S. W. J. Chem. Soc., Perkin
Trans. 1 1975, 1574.
3. Kimball, F. S.; Tunoori, A. R.; Victory, S. F.; Dutta, D.;
White, J. M.; Himes, R. H.; Georg, G. I. Bioorg. Med.
Chem. Lett. 2007, 17, 4703.
10. Prepared according to reference 4 and Commins, D.
L.; Morgan, L. A. Tetrahedron Lett. 1991, 32, 5919–
5922.